Previously multiple threads were allowed to load histograms concurrently.
There were no known problems caused by this. But given amount of data
races in this code, it'd happen sooner or later.
To avoid scalability bottleneck, histograms loading is protected by
per-TABLE_SHARE atomic variable.
Whenever histograms were loaded by preceding statement (hot-path), a
scalable load-acquire check is performed.
Whenever histograms have to be loaded anew, mutual exclusion for loaders
is established by atomic variable. If histograms are being loaded
concurrently, statement waits until load is completed.
- Table_statistics::total_hist_size moved to TABLE_STATISTICS_CB: only
meaningful within TABLE_SHARE (not used for collected stats).
- TABLE_STATISTICS_CB::histograms_can_be_read and
TABLE_STATISTICS_CB::histograms_are_read are replaced with a tri state
atomic variable.
- Simplified away alloc_histograms_for_table_share().
Note: there's still likely a data race if a thread attempts accessing
histograms data after it failed to load it (because of concurrent load).
It was there previously and goes out of the scope of this effort. One way
of fixing it could be reviving TABLE::histograms_are_read and adding
appropriate checks whenever it is needed.
Part of MDEV-19061 - table_share used for reading statistical tables is
not protected
Previously multiple threads were allowed to load statistics concurrently.
There were no known problems caused by this. But given amount of data
races in this code, it'd happen sooner or later.
To avoid scalability bottleneck, statistics loading is protected by
per-TABLE_SHARE atomic variable.
Whenever statistics were loaded by preceding statement (hot-path), a
scalable load-acquire check is performed.
Whenever statistics have to be loaded anew, mutual exclusion for loaders
is established by atomic variable. If statistics are being loaded
concurrently, statement waits until load is completed.
TABLE_STATISTICS_CB::stats_can_be_read and
TABLE_STATISTICS_CB::stats_is_read are replaced with a tri state atomic
variable.
Part of MDEV-19061 - table_share used for reading statistical tables is
not protected
Removed redundant loops, integrated logics into the caller instead.
Unified condition in read_statistics_for_tables(), less
"table_share != NULL" checks, no more potential "table_share == NULL"
dereferencing.
Part of MDEV-19061 - table_share used for reading statistical tables is
not protected
A crash was observed where dict_acquire_mdl_shared<trylock=false>
would invoke memcpy() with an apparently uninitialized tbl_len.
dict_table_t::parse_name(): Remove an unnecessary tbl_len--
operation. (This should be mostly non-functional cleanup.)
dict_acquire_mdl_shared(): If the second dict_table_t::parse_name()
returns false, terminate the loop just like we would do on the
first invocation.
In Item_nodeset_func_predicate::val_nodeset, args[1] is not necessarily
an Item_func descendant. It can be Item_bool.
Removing a wrong cast. It was not really needed anyway.
commit d09aec7a15 (MDEV-19940)
caused a regression. We made wait_lock_get_heap_no() return
uint16_t instead of ulint, and we mostly replaced the previous
magic value ULINT_UNDEFINED with 0. But, we failed to adjust
some assertions. Furthermore, 0 is a valid although rare value
for record locks. (Record locks can be temporarily stored on
page infimum in some operations that involve multiple leaf pages.)
Let us use 0xFFFF as the magic value. Valid heap numbers
are limited to less than 9362 = innodb_page_size/(5+1+1)
when using a minimal 1-byte PRIMARY KEY and a
secondary index on a NULL or '' column.
when checking for free port, use the same logic (IPv6 socket address
/ dual socket), like the server would.
Previous solution for testing whether port is free was trying to bind
IPv4 socket on INADDR_ANY.
This not work now on some reason, that attempt succeeds, even if there is
an existing IPv6-dual socket listening on 0.0.0.0:3306
In the merge 9e6e43551f we replaced
direct use of std::atomic with a wrapper class, so that
dict_index_t::lock will support the default assignment operator.
As part of that change, one occurrence of std::memory_order_release
was accidentally replaced with std::memory_order_relaxed.
Thanks to Sergey Vojtovich for noticing this.
Respect system fields in NO_ZERO_DATE mode.
This is the subject for refactoring in MDEV-19597
Conflict resolution from 7d5223310789f967106d86ce193ef31b315ecff0
The constructor of Lex_ident_sys returns LEX_CSTRING(NULL,0) if character set
conversion goes wrong, and raises the "wrong character string" error in
the diagnostics area.
The code in sql_yacc.yy did not check Lex_ident_sys::ptr against NULL,
so the execution entered functions that did not expect NULL (and crashed).
Fixing the code to do MYSQL_YYABORT if Lex_ident_sys::ptr is NULL
after constructing.
UPDATE gets access to history records because versioning conditions
are not set for VIEW. This leads to endless loop of inserting history
records when clustered index is rebuilt and ha_rnd_next() returns
newly inserted history record.
Return back original behavior of failing on write-locked table in
historical query.
35b679b9 assumed that SELECT_LEX::lock_type influences anything, but
actually at this point table is already locked. Original bug report
was tempesta-tech/mariadb#102
System versioning assertion fix. Since DROP SYSTEM VERSIONING does not
change list of dropped keys we should handle a special case.
Caused by MDEV-19751. This fix deprecates MDEV-17091.
- `SET DEFAULT ROLE xxx [FOR yyy]` should say:
"User yyy has not been granted a role xxx" if:
- The current user (not the user `yyy` in the FOR clause) can see the
role xxx. It can see the role if:
* role exists in `mysql.roles_mappings` (traverse the graph),
* If the current user has read access on `mysql.user` table - in
that case, it can see all roles, granted or not.
- Otherwise it should be "Invalid role specification".
In other words, it should not be possible to use `SET DEFAULT ROLE` to discover whether a specific role exist or not.
commit 3a37644a29 added a non-POD
member buf_page_info_t::id, and thus GCC 7 or later would complain
about a memset() call. Let my_malloc fill the memory for us.
InnoDB mutex monitor is accessing mutexes of poisoned (cached) trx
objects. Unpoison ReadView::m_mutex similarly to trx_t::mutex.
This is regression after MDEV-22593.
otherwise it reaches "0 tests left" state and then waits for
a few minutes for all workers to complete their tests.
show failures. account for retries.
This reverts commit 6f1f911497.
because it doesn't do anything now (the server doesn't check
my_disable_leak_check) and it never did anything before
(because without `extern` it simply created a local instance of
my_disable_leak_check, did not affect server's my_disable_leak_check).
cannot use the current THD::mem_root, because it can be temporarily
reassigned to something with a very different life time
(e.g. to TABLE::mem_root or range optimizer mem_root).
data_file_length == 0 in mi_repair() is normal for REPAIR ... USE_FRM.
But in-file links (for blocks and deleted chain) must be compared with
the real file length to avoid spurious "link points outside datafile"
warnings and arbitrary block skipping.
- Remove extra ',' and quotes
- Remove extra newline and remove double newlines
- Added options --lsn-redo-end and --lsn-undo-end to aria_read_log
- Allow one to give the aria_read_log lsn aruments as number,0xhexnumber,
the same way as lsn's are written by aria_read_log
- Don't write full pages to redo log with EXTRA_DEBUG as this takes up
a lot of disk and there has not been a need for this extra loggging for
a long time. Instead one should use EXTRA_ARIA_DEBUG instead.
