The bug was that full memory barrier was missing in the code that ensures that
a waiter on an InnoDB mutex will not go to sleep unless it is guaranteed to be
woken up again by another thread currently holding the mutex. This made
possible a race where a thread could get stuck waiting for a mutex that is in
fact no longer locked. If that thread was also holding other critical locks,
this could stall the entire server. There is an error monitor thread than can
break the stall, it runs about once per second. But if the error monitor
thread itself got stuck or was not running, then the entire server could hang
infinitely.
This was introduced on i386/amd64 platforms in 5.5.40 and 10.0.13 by an
incorrect patch that tried to fix the similar problem for PowerPC.
This commit reverts the incorrect PowerPC patch, and instead implements a fix
for PowerPC that does not change i386/amd64 behaviour, making PowerPC work
similarly to i386/amd64.
Analysis: InnoDB error monitor is responsible to call every second
sync_arr_wake_threads_if_sema_free() to wake up possible hanging
threads if they are missed in mutex_signal_object. This is not
possible if error monitor itself is on mutex/semaphore wait. We
should avoid all unnecessary mutex/semaphore waits on error monitor.
Currently error monitor calls function buf_flush_stat_update()
that calls log_get_lsn() function and there we will try to get
log_sys mutex. Better, solution for error monitor is that in
buf_flush_stat_update() we will try to get lsn with
mutex_enter_nowait() and if we did not get mutex do not update
the stats.
Fix: Use log_get_lsn_nowait() function on buf_flush_stat_update()
function. If returned lsn is 0, we do not update flush stats.
log_get_lsn_nowait() will use mutex_enter_nowait() and if
we get mutex we return a correct lsn if not we return 0.
When UNIV_INTERN is missing in built-in XtraDB, this causes the
innodb_plugin to call the XtraDB version of the function instead
of its own (seen in --embedded-server test failure in Buildbot).
This in turn causes bad things to happen in case of difference
between XtranDB and innodb_plugin.
MDEV-6483 - Deadlock around rw_lock_debug_mutex on PPC64
This problem affects only debug builds on PPC64.
There are at least two race conditions around
rw_lock_debug_mutex_enter and rw_lock_debug_mutex_exit:
- rw_lock_debug_waiters was loaded/stored without setting
appropriate locks/memory barriers.
- there is a gap between calls to os_event_reset() and
os_event_wait() and in such case we're supposed to pass
return value of the former to the latter.
Fixed by replacing self-cooked spinlocks with system mutexes.
These days system mutexes offer much better performance. OTOH
performance is not that critical for debug builds.
MDEV-6450 - MariaDB crash on Power8 when built with advance tool
chain
InnoDB mutex_exit() function calls __sync_test_and_set() to release
the lock. According to manual this function is supposed to create
"acquire" memory barrier whereas in fact we need "release" memory
barrier at mutex_exit().
The problem isn't repeatable with gcc because it creates
"acquire-release" memory barrier for __sync_test_and_set().
ATC creates just "acquire" barrier.
Fixed by creating proper barrier at mutex_exit() by using
__sync_lock_release() instead of __sync_test_and_set().
Part of this work is based on Stewart Smitch's memory barrier and lower priori
patches for power8.
- Added memory syncronization for innodb & xtradb for power8.
- Added HAVE_WINDOWS_MM_FENCE to CMakeList.txt
- Added os_isync to fix a syncronization problem on power
- Added log_get_lsn_nowait which is now used srv_error_monitor_thread to ensur
if log mutex is locked.
All changes done both for InnoDB and Xtradb
InnoDB: Failing assertion: mutex_own(&buf_pool->LRU_list_mutex)
and
InnoDB: Assertion failure in thread 2868898624 in file buf0lru.c line 1077
InnoDB: Failing assertion: mutex_own(&buf_pool->LRU_list_mutex)
Analysis: Function buf_LRU_free_block might release LRU_list_mutex on
same cases to avoid mutex order problems, we need to take it back
before accessing list.
than with InnoDB plugin
Fix: os0file.h in XtraDB had OS_AIO_N_PENDING_IOS_PER_THREAD 256
when on InnoDB it is OS_AIO_N_PENDING_IOS_PER_THREAD 32. Changed
XtraDB also to use 32.
Analysis: Based on crashed the buffer pool instance identifier is
not correct on block to be freed. Add LRU list mutex holding
on functions calling free and add additional safety checks.
Analysis: Can't disable the error message because you may get database
started with incorrect log file size.
Fix: Thus only improve the error message to give more information
to users.
on select from I_S.INNODB_CHANGED_PAGES
Analysis: limit_lsn_range_from_condition() incorrectly parses
start_lsn and/or end_lsn conditions.
Fix from SergeyP. Added some test cases.
- Table locks now ends with state "After table lock"
- Open table now ends with state "After opening tables"
- All calls to close_thread_tables(), not only from mysql_execute_command(), has state "closing tables"
- Added state "executing" for mysql admin commands, like CACHE INDEX, REPAIR TABLE etc.
- Added state "Finding key cache" for CACHE INDEX
- Added state "Filling schema table" when we generate temporary table for SHOW commands and information schema.
Other things:
Add limit from innobase for thread_sleep_delay. This fixed a failing tests case.
Added db.opt to support-files to make 'make package' work
mysql-test/suite/funcs_1/datadict/processlist_val.inc:
Use new state
mysql-test/suite/funcs_1/r/processlist_priv_no_prot.result:
Updated test result because of new state
mysql-test/suite/funcs_1/r/processlist_val_no_prot.result:
Updated test result because of new state
sql/CMakeLists.txt:
Have option files in support-files
sql/lock.cc:
Added new state 'After table lock'
sql/sql_admin.cc:
Added state "executing" and "Sending data" for mysql admin commands, like CACHE INDEX, REPAIR TABLE etc.
Added state "Finding key cache"
sql/sql_base.cc:
open tables now ends with state "After table lock", instead of NULL
sql/sql_parse.cc:
Moved state "closing tables" to close_thread_tables()
sql/sql_show.cc:
Added state "Filling schema table" when we generate temporary table for SHOW commands and information schema.
storage/xtradb/buf/buf0buf.c:
Removed compiler warning
storage/xtradb/handler/ha_innodb.cc:
Add limit from innobase for thread_sleep_delay. This fixed a failing tests case.
support-files/db.opt:
cmakes needs this to create data/test directory
This is not yet a fix. This is change to print additional information at the point
when this assertion is going to happen. Print as much information about the pages
and index to find out why next page is not a compact format.
file storage.
Analysis: posix_fallocate was called using 0 as offset and len as
desired size. This is not optimal for SSDs.
Fix: Call posix_fallocate with correct offset i.e. current file size
and extend the file from there len bytes.
on debug build when innodb_use_fallocate=1
Analysis: There was merge error that caused additional call to
fil_node_complete_io.
Fixed by removing incorrect call and fixed calculation of extended
file size.
Analysis: When filespace is extended there is first prepare for IO. But if
posix_fallocate is used there was no complete IO causing assertion
at shutdown indicating that all IO is not finished.
Fix: If posix_fallocate is used to extend the filespace, there
is no need to wait for IO to complete, thus we treat this
operation as a read operation. We need to mark IO as
completed or there would be assertion on shutdown at
fil_node_close_file() because all pending IO is not finished.
- Reset static variables that are used to signal "init done" for DBUG, in dbug_end()
- Set string server variables to NULL after memory for the value is freed - avoids double free()
- fix DBUG_ASSERTs that happened during reinitialization.
Change default for innodb_use_fallocate to FALSE, due to bugs in older Linux kernels (posix_fallocate() does not always guarantee that file size is like one specified)
(because it's conceptually wrong. only the user can decide whether the kill is
allowed to leave tables in the inconsistent state, storage engine has no say in that)
KILL now breaks locks inside InnoDB
Fixed possible deadlock when running INNODB STATUS
Added ha_kill_query() and kill_query() to send kill signal to all storage engines
Added reset_killed() to ensure we don't reset killed state while awake() is getting called
include/mysql/plugin.h:
Added thd_mark_as_hard_kill()
include/mysql/plugin_audit.h.pp:
Added thd_mark_as_hard_kill()
include/mysql/plugin_auth.h.pp:
Added thd_mark_as_hard_kill()
include/mysql/plugin_ftparser.h.pp:
Added thd_mark_as_hard_kill()
sql/handler.cc:
Added ha_kill_query() to send kill signal to all storage engines
sql/handler.h:
Added ha_kill_query() and kill_query() to send kill signal to all storage engines
sql/log_event.cc:
Use reset_killed()
sql/mdl.cc:
use thd->killed instead of thd_killed() to abort on soft kill
sql/sp_rcontext.cc:
Use reset_killed()
sql/sql_class.cc:
Fixed possible deadlock in INNODB STATUS by not getting thd->LOCK_thd_data if it's locked.
Use reset_killed()
Tell storge engines that KILL has been sent
sql/sql_class.h:
Added reset_killed() to ensure we don't reset killed state while awake() is getting called.
Added mark_as_hard_kill()
sql/sql_insert.cc:
Use reset_killed()
sql/sql_parse.cc:
Simplify detection of killed queries.
Use reset_killed()
sql/sql_select.cc:
Use reset_killed()
sql/sql_union.cc:
Use reset_killed()
storage/innobase/handler/ha_innodb.cc:
Added innobase_kill_query()
Fixed error reporting for interrupted queries.
storage/xtradb/handler/ha_innodb.cc:
Added innobase_kill_query()
Fixed error reporting for interrupted queries.
Changed implementation os_file_rename() on Windows such as it does not fail if destination file already exists. Now MoveFileEx() with MOVEFILE_REPLACE_EXISTING flag is used, instead of prior MoveFile().
This fixed implementation is better compatible with rename() on POSIX systems.
- Wrong thd uses in Item_subselect, could lead to crash
- Inititalize uninitialized variable in new autoincrement handling code
sql/handler.cc:
More DBUG_PRINT
sql/item_subselect.cc:
Wrong thd uses in Item_subselect, could lead to crash
storage/innobase/handler/ha_innodb.cc:
Initialize variable needed by upper level. This only happens when auto-increment value wraps over.
storage/xtradb/handler/ha_innodb.cc:
Initialize variable needed by upper level. This only happens when auto-increment value wraps over.
create table t1 (a smallint primary key auto_increment);
insert into t1 values(32767);
insert into t1 values(NULL);
ERROR 1062 (23000): Duplicate entry '32767' for key 'PRIMARY
Now on always gets error HA_ERR_AUTOINC_RANGE=167 "Out of range value for column", independent of
store engine, SQL Mode or number of inserted rows. This is an unique error that is easier to test for in replication.
Another bug fix is that we now get an error when trying to insert a too big auto-generated value, even in non-strict mode.
Before one get insted the max column value inserted.
This patch also fixes some issues with inserting negative numbers in an auto-increment column.
Fixed the ER_DUP_ENTRY and HA_ERR_AUTOINC_ERANGE are compared the same between master and slave.
This ensures that replication works between an old server to a new slave for auto-increment overflow errors.
Added SQLSTATE errors for handler errors
Smaller bug fixes:
* Added warnings for duplicate key errors when using INSERT IGNORE
* Fixed bug when using --skip-log-bin followed by --log-bin, which did set log-bin to "0"
* Allow one to see how cmake is called by using --just-print --just-configure
BUILD/FINISH.sh:
--just-print --just-configure now shows how cmake would be invoked. Good for understanding parameters to cmake.
cmake/configure.pl:
--just-print --just-configure now shows how cmake would be invoked. Good for understanding parameters to cmake.
include/CMakeLists.txt:
Added handler_state.h
include/handler_state.h:
SQLSTATE for handler error messages.
Required for HA_ERR_AUTOINC_ERANGE, but solves also some other cases.
mysql-test/extra/binlog_tests/binlog.test:
Fixed old wrong behaviour
Added more tests
mysql-test/extra/binlog_tests/binlog_insert_delayed.test:
Reset binary log to only print what's necessary in show_binlog_events
mysql-test/extra/rpl_tests/rpl_auto_increment.test:
Update to new error codes
mysql-test/extra/rpl_tests/rpl_insert_delayed.test:
Ignore warnings as this depends on how the test is run
mysql-test/include/strict_autoinc.inc:
On now gets an error on overflow
mysql-test/r/auto_increment.result:
Update results after fixing error message
mysql-test/r/auto_increment_ranges_innodb.result:
Test new behaviour
mysql-test/r/auto_increment_ranges_myisam.result:
Test new behaviour
mysql-test/r/commit_1innodb.result:
Added warnings for duplicate key error
mysql-test/r/create.result:
Added warnings for duplicate key error
mysql-test/r/insert.result:
Added warnings for duplicate key error
mysql-test/r/insert_select.result:
Added warnings for duplicate key error
mysql-test/r/insert_update.result:
Added warnings for duplicate key error
mysql-test/r/mix2_myisam.result:
Added warnings for duplicate key error
mysql-test/r/myisam_mrr.result:
Added warnings for duplicate key error
mysql-test/r/null_key.result:
Added warnings for duplicate key error
mysql-test/r/replace.result:
Update to new error codes
mysql-test/r/strict_autoinc_1myisam.result:
Update to new error codes
mysql-test/r/strict_autoinc_2innodb.result:
Update to new error codes
mysql-test/r/strict_autoinc_3heap.result:
Update to new error codes
mysql-test/r/trigger.result:
Added warnings for duplicate key error
mysql-test/r/xtradb_mrr.result:
Added warnings for duplicate key error
mysql-test/suite/binlog/r/binlog_innodb_row.result:
Updated result
mysql-test/suite/binlog/r/binlog_row_binlog.result:
Out of range data for auto-increment is not inserted anymore
mysql-test/suite/binlog/r/binlog_statement_insert_delayed.result:
Updated result
mysql-test/suite/binlog/r/binlog_stm_binlog.result:
Out of range data for auto-increment is not inserted anymore
mysql-test/suite/binlog/r/binlog_unsafe.result:
Updated result
mysql-test/suite/innodb/r/innodb-autoinc.result:
Update to new error codes
mysql-test/suite/innodb/r/innodb-lock.result:
Updated results
mysql-test/suite/innodb/r/innodb.result:
Updated results
mysql-test/suite/innodb/r/innodb_bug56947.result:
Updated results
mysql-test/suite/innodb/r/innodb_mysql.result:
Updated results
mysql-test/suite/innodb/t/innodb-autoinc.test:
Update to new error codes
mysql-test/suite/maria/maria3.result:
Updated result
mysql-test/suite/maria/mrr.result:
Updated result
mysql-test/suite/optimizer_unfixed_bugs/r/bug43617.result:
Updated result
mysql-test/suite/rpl/r/rpl_auto_increment.result:
Update to new error codes
mysql-test/suite/rpl/r/rpl_insert_delayed,stmt.rdiff:
Updated results
mysql-test/suite/rpl/r/rpl_loaddatalocal.result:
Updated results
mysql-test/t/auto_increment.test:
Update to new error codes
mysql-test/t/auto_increment_ranges.inc:
Test new behaviour
mysql-test/t/auto_increment_ranges_innodb.test:
Test new behaviour
mysql-test/t/auto_increment_ranges_myisam.test:
Test new behaviour
mysql-test/t/replace.test:
Update to new error codes
mysys/my_getopt.c:
Fixed bug when using --skip-log-bin followed by --log-bin, which did set log-bin to "0"
sql/handler.cc:
Ignore negative values for signed auto-increment columns
Always give an error if we get an overflow for an auto-increment-column (instead of inserting the max value)
Ensure that the row number is correct for the out-of-range-value error message.
******
Fixed wrong printing of column namn for "Out of range value" errors
Fixed that INSERT_ID is correctly replicated also for out-of-range autoincrement values
Fixed that print_keydup_error() can also be used to generate warnings
******
Return HA_ERR_AUTOINC_ERANGE (167) instead of ER_WARN_DATA_OUT_OF_RANGE for auto-increment overflow
sql/handler.h:
Allow INSERT IGNORE to continue also after out-of-range inserts.
Fixed that print_keydup_error() can also be used to generate warnings
sql/log_event.cc:
Added DBUG_PRINT
Fixed the ER_AUTOINC_READ_FAILED, ER_DUP_ENTRY and HA_ERR_AUTOINC_ERANGE are compared the same between master and slave.
This ensures that replication works between an old server to a new slave for auto-increment overflow errors.
sql/sql_insert.cc:
Add warnings for duplicate key errors when using INSERT IGNORE
sql/sql_state.c:
Added handler errors
sql/sql_table.cc:
Update call to print_keydup_error()
storage/innobase/handler/ha_innodb.cc:
Fixed increment handling of auto-increment columns to be consistent with rest of MariaDB.
storage/xtradb/handler/ha_innodb.cc:
Fixed increment handling of auto-increment columns to be consistent with rest of MariaDB.
sql/handler.cc:
SHOW INNODB STATUS sometimes returns 0 even if it has generated an error.
This code is here to catch it until InnoDB some day is fixed.
storage/innobase/handler/ha_innodb.cc:
Catch at least one of the possible errors from SHOW INNODB STATUS to provide a correct return code.
storage/xtradb/handler/ha_innodb.cc:
Catch at least one of the possible errors from SHOW INNODB STATUS to provide a correct return code.
support-files/my-huge.cnf.sh:
Fixed typo
sql/item_subselect.cc:
Added purecov info
sql/sql_select.cc:
Added cast
storage/innobase/handler/ha_innodb.cc:
Added cast
storage/xtradb/btr/btr0btr.c:
Added buf_block_get_frame_fast() to avoid compiler warning
storage/xtradb/handler/ha_innodb.cc:
Added cast
storage/xtradb/include/buf0buf.h:
Innodb has buf_block_get_frame(block) defined as (block)->frame.
Didn't want to do a big change to break xtradb as it may use block_get_frame() differently, so I mad this quick hack to patch one compiler warning.
- Better error messages
This fixes that one again can run the test systems with many threads without having to increase fs.aio-max-nr.
mysql-test/include/mtr_check.sql:
Ignore the INNODB_USE_NATIVE_AIO variable (may change during execution)
mysql-test/mysql-test-run.pl:
Ignore warnings for failure to setup AIO
storage/innobase/os/os0file.c:
Continue without AIO even if we can't allocate resources for AIO
storage/xtradb/os/os0file.c:
Continue without AIO even if we can't allocate resources for AIO
storage/xtradb/srv/srv0start.c:
Give an error message (instead of core dump) if AIO can't be initialized
