bug #57006 "Deadlock between HANDLER and FLUSH TABLES WITH READ
LOCK" and bug #54673 "It takes too long to get readlock for
'FLUSH TABLES WITH READ LOCK'".
The first bug manifested itself as a deadlock which occurred
when a connection, which had some table open through HANDLER
statement, tried to update some data through DML statement
while another connection tried to execute FLUSH TABLES WITH
READ LOCK concurrently.
What happened was that FTWRL in the second connection managed
to perform first step of GRL acquisition and thus blocked all
upcoming DML. After that it started to wait for table open
through HANDLER statement to be flushed. When the first connection
tried to execute DML it has started to wait for GRL/the second
connection creating deadlock.
The second bug manifested itself as starvation of FLUSH TABLES
WITH READ LOCK statements in cases when there was a constant
stream of concurrent DML statements (in two or more
connections).
This has happened because requests for protection against GRL
which were acquired by DML statements were ignoring presence of
pending GRL and thus the latter was starved.
This patch solves both these problems by re-implementing GRL
using metadata locks.
Similar to the old implementation acquisition of GRL in new
implementation is two-step. During the first step we block
all concurrent DML and DDL statements by acquiring global S
metadata lock (each DML and DDL statement acquires global IX
lock for its duration). During the second step we block commits
by acquiring global S lock in COMMIT namespace (commit code
acquires global IX lock in this namespace).
Note that unlike in old implementation acquisition of
protection against GRL in DML and DDL is semi-automatic.
We assume that any statement which should be blocked by GRL
will either open and acquires write-lock on tables or acquires
metadata locks on objects it is going to modify. For any such
statement global IX metadata lock is automatically acquired
for its duration.
The first problem is solved because waits for GRL become
visible to deadlock detector in metadata locking subsystem
and thus deadlocks like one in the first bug become impossible.
The second problem is solved because global S locks which
are used for GRL implementation are given preference over
IX locks which are acquired by concurrent DML (and we can
switch to fair scheduling in future if needed).
Important change:
FTWRL/GRL no longer blocks DML and DDL on temporary tables.
Before this patch behavior was not consistent in this respect:
in some cases DML/DDL statements on temporary tables were
blocked while in others they were not. Since the main use cases
for FTWRL are various forms of backups and temporary tables are
not preserved during backups we have opted for consistently
allowing DML/DDL on temporary tables during FTWRL/GRL.
Important change:
This patch changes thread state names which are used when
DML/DDL of FTWRL is waiting for global read lock. It is now
either "Waiting for global read lock" or "Waiting for commit
lock" depending on the stage on which FTWRL is.
Incompatible change:
To solve deadlock in events code which was exposed by this
patch we have to replace LOCK_event_metadata mutex with
metadata locks on events. As result we have to prohibit
DDL on events under LOCK TABLES.
This patch also adds extensive test coverage for interaction
of DML/DDL and FTWRL.
Performance of new and old global read lock implementations
in sysbench tests were compared. There were no significant
difference between new and old implementations.
- Fixed bug in pagecache_delete_internal() when deleting block that was flushed by another thread (fixed bug when block->next_used was unexpectedly null)
Fixed some using uninitialized memory warnings found by valgrind.
mysql-test/t/information_schema_all_engines-master.opt:
Added options to make slow test run faster
sql/sp.cc:
Fixed valgrind warning.
sql/sql_show.cc:
Fixed valgrind warning.
storage/maria/ma_bitmap.c:
Fixed wrong call parameter to pagecache_unlock_by_link() that caused assert crash in pagecache
storage/maria/ma_pagecache.c:
Fixed possible error in pagecache_wait_lock() when hash_link was not set. (We never hit this issue, but could be possible when two threads updates the same page.
Fixed bug in pagecache_delete_internal() when deleting block that was flushed by another thread (fixed bug when block->next_used was unexpectedly null)
Cleanup: moved pagecache_pthread_mutex_unlock() over comments and asserts to be just before pagecache_fwrite()
Avoid doing reallocs
Prealloc some strings / provide extension allocation size to some strings
This gave a 25 % speedup in some mysql-test-run tests.
mysys/safemalloc.c:
More DBUG_PRINT
sql/net_serv.cc:
Make all mallocs() look the similar. (just-for-safety fix)
sql/protocol.cc:
Ensure that communication packet buffer is allocated.
(It's freed by stored precedures and some DLL statements)
sql/sp.cc:
Fixed valgrind warning
sql/sql_select.cc:
Set extent allocation for buffer that has a lot of append() calls.
sql/sql_show.cc:
Fixed wrong usage of string buffer. Old code worked in test suite 'just-by-chance'
sql/sql_string.cc:
Call realloc_with_extra_if_needed() in append() functions.
sql/sql_string.h:
Added 'extra_alloc' member, to specify chunck size for realloc().
extra_alloc is addaptive to catch cases where preallocation of buffers is not done properly.
Simplified free() to allow compiler to optimize things better (and to keep things consistent).
Fixed shrink() to take into account the extra memory added to the Alloced_length in realloc(). This saves us a realloc() per query.
sql/sql_test.cc:
Set extent allocation for buffer that has a lot of append() calls.
sql/table.cc:
Set extent allocation for buffer that has a lot of append() calls.
This crash could happen if SHOW CREATE VIEW indirectly failed to open a
view due to failures to open underlying tables (or functions). Several
such errors were hidden and converted to ER_VIEW_INVALID warnings to
prevent exposing details of underlying tables for which the user have
no privileges.
However, with the changes introduced by the patch for Bug#52044,
failing to open a view will cause opened tables, views and functions
to be closed. Since the errors causing these failures were converted
to warnings, SHOW CREATE VIEW would try to continue. This made it
possible to try to access memory that had been freed, causing a crash.
This patch fixes the problem by not closing opened tables, views and
functions in these cases. This allows SHOW CREATE VIEW to continue
and also prevents it from accessing freed memory.
Test case added to lock_sync.test.
This crash could happen if SHOW CREATE VIEW indirectly failed to open a
view due to failures to open underlying tables (or functions). Several
such errors were hidden and converted to ER_VIEW_INVALID warnings to
prevent exposing details of underlying tables for which the user have
no privileges.
However, with the changes introduced by the patch for Bug#52044,
failing to open a view will cause opened tables, views and functions
to be closed. Since the errors causing these failures were converted
to warnings, SHOW CREATE VIEW would try to continue. This made it
possible to try to access memory that had been freed, causing a crash.
This patch fixes the problem by not closing opened tables, views and
functions in these cases. This allows SHOW CREATE VIEW to continue
and also prevents it from accessing freed memory.
Test case added to lock_sync.test.
sql_show.cc during rqg_info_schema test on Windows".
Ensure we do not access freed memory when filling
information_schema.views when one of the views
could not be properly opened.
mysql-test/r/information_schema.result:
Update results - a fix for Bug#56540.
mysql-test/t/information_schema.test:
Add a test case for Bug#56540
sql/sql_base.cc:
Push an error into the Diagnostics area
when we return an error.
This directs get_all_tables() to the execution
branch which doesn't involve 'process_table()'
when no table/view was opened.
sql/sql_show.cc:
Do not try to access underlying table fields
when opening of a view failed. The underlying
table is closed in that case, and accessing
its fields may lead to dereferencing a damaged
pointer.
sql_show.cc during rqg_info_schema test on Windows".
Ensure we do not access freed memory when filling
information_schema.views when one of the views
could not be properly opened.
Bug#54678: InnoDB, TRUNCATE, ALTER, I_S SELECT, crash or deadlock
- Incompatible change: truncate no longer resorts to a row by
row delete if the storage engine does not support the truncate
method. Consequently, the count of affected rows does not, in
any case, reflect the actual number of rows.
- Incompatible change: it is no longer possible to truncate a
table that participates as a parent in a foreign key constraint,
unless it is a self-referencing constraint (both parent and child
are in the same table). To work around this incompatible change
and still be able to truncate such tables, disable foreign checks
with SET foreign_key_checks=0 before truncate. Alternatively, if
foreign key checks are necessary, please use a DELETE statement
without a WHERE condition.
Problem description:
The problem was that for storage engines that do not support
truncate table via a external drop and recreate, such as InnoDB
which implements truncate via a internal drop and recreate, the
delete_all_rows method could be invoked with a shared metadata
lock, causing problems if the engine needed exclusive access
to some internal metadata. This problem originated with the
fact that there is no truncate specific handler method, which
ended up leading to a abuse of the delete_all_rows method that
is primarily used for delete operations without a condition.
Solution:
The solution is to introduce a truncate handler method that is
invoked when the engine does not support truncation via a table
drop and recreate. This method is invoked under a exclusive
metadata lock, so that there is only a single instance of the
table when the method is invoked.
Also, the method is not invoked and a error is thrown if
the table is a parent in a non-self-referencing foreign key
relationship. This was necessary to avoid inconsistency as
some integrity checks are bypassed. This is inline with the
fact that truncate is primarily a DDL operation that was
designed to quickly remove all data from a table.
mysql-test/suite/innodb/t/innodb-truncate.test:
Add test cases for truncate and foreign key checks.
Also test that InnoDB resets auto-increment on truncate.
mysql-test/suite/innodb/t/innodb.test:
FK is not necessary, test is related to auto-increment.
Update error number, truncate is no longer invoked if
table is parent in a FK relationship.
mysql-test/suite/innodb/t/innodb_mysql.test:
Update error number, truncate is no longer invoked if
table is parent in a FK relationship.
Use delete instead of truncate, test is used to check
the interaction of FKs, triggers and delete.
mysql-test/suite/parts/inc/partition_check.inc:
Fix typo.
mysql-test/suite/sys_vars/t/foreign_key_checks_func.test:
Update error number, truncate is no longer invoked if
table is parent in a FK relationship.
mysql-test/t/mdl_sync.test:
Modify test case to reflect and ensure that truncate takes
a exclusive metadata lock.
mysql-test/t/trigger-trans.test:
Update error number, truncate is no longer invoked if
table is parent in a FK relationship.
sql/ha_partition.cc:
Reorganize the various truncate methods. delete_all_rows is now
passed directly to the underlying engines, so as truncate. The
code responsible for truncating individual partitions is moved
to ha_partition::truncate_partition, which is invoked when a
ALTER TABLE t1 TRUNCATE PARTITION p statement is executed.
Since the partition truncate no longer can be invoked via
delete, the bitmap operations are not necessary anymore. The
explicit reset of the auto-increment value is also removed
as the underlying engines are now responsible for reseting
the value.
sql/handler.cc:
Wire up the handler truncate method.
sql/handler.h:
Introduce and document the truncate handler method. It assumes
certain use cases of delete_all_rows.
Add method to retrieve the list of foreign keys referencing a
table. Method is used to avoid truncating tables that are
parent in a foreign key relationship.
sql/share/errmsg-utf8.txt:
Add error message for truncate and FK.
sql/sql_lex.h:
Introduce a flag so that the partition engine can detect when
a partition is being truncated. Used to give a special error.
sql/sql_parse.cc:
Function mysql_truncate_table no longer exists.
sql/sql_partition_admin.cc:
Implement the TRUNCATE PARTITION statement.
sql/sql_truncate.cc:
Change the truncate table implementation to use the new truncate
handler method and to not rely on row-by-row delete anymore.
The truncate handler method is always invoked with a exclusive
metadata lock. Also, it is no longer possible to truncate a
table that is parent in some non-self-referencing foreign key.
storage/archive/ha_archive.cc:
Rename method as the description indicates that in the future
this could be a truncate operation.
storage/blackhole/ha_blackhole.cc:
Implement truncate as no operation for the blackhole engine in
order to remain compatible with older releases.
storage/federated/ha_federated.cc:
Introduce truncate method that invokes delete_all_rows.
This is required to support partition truncate as this
form of truncate does not implement the drop and recreate
protocol.
storage/heap/ha_heap.cc:
Introduce truncate method that invokes delete_all_rows.
This is required to support partition truncate as this
form of truncate does not implement the drop and recreate
protocol.
storage/ibmdb2i/ha_ibmdb2i.cc:
Introduce truncate method that invokes delete_all_rows.
This is required to support partition truncate as this
form of truncate does not implement the drop and recreate
protocol.
storage/innobase/handler/ha_innodb.cc:
Rename delete_all_rows to truncate. InnoDB now does truncate
under a exclusive metadata lock.
Introduce and reorganize methods used to retrieve the list
of foreign keys referenced by a or referencing a table.
storage/myisammrg/ha_myisammrg.cc:
Introduce truncate method that invokes delete_all_rows.
This is required in order to remain compatible with earlier
releases where truncate would resort to a row-by-row delete.
Bug#54678: InnoDB, TRUNCATE, ALTER, I_S SELECT, crash or deadlock
- Incompatible change: truncate no longer resorts to a row by
row delete if the storage engine does not support the truncate
method. Consequently, the count of affected rows does not, in
any case, reflect the actual number of rows.
- Incompatible change: it is no longer possible to truncate a
table that participates as a parent in a foreign key constraint,
unless it is a self-referencing constraint (both parent and child
are in the same table). To work around this incompatible change
and still be able to truncate such tables, disable foreign checks
with SET foreign_key_checks=0 before truncate. Alternatively, if
foreign key checks are necessary, please use a DELETE statement
without a WHERE condition.
Problem description:
The problem was that for storage engines that do not support
truncate table via a external drop and recreate, such as InnoDB
which implements truncate via a internal drop and recreate, the
delete_all_rows method could be invoked with a shared metadata
lock, causing problems if the engine needed exclusive access
to some internal metadata. This problem originated with the
fact that there is no truncate specific handler method, which
ended up leading to a abuse of the delete_all_rows method that
is primarily used for delete operations without a condition.
Solution:
The solution is to introduce a truncate handler method that is
invoked when the engine does not support truncation via a table
drop and recreate. This method is invoked under a exclusive
metadata lock, so that there is only a single instance of the
table when the method is invoked.
Also, the method is not invoked and a error is thrown if
the table is a parent in a non-self-referencing foreign key
relationship. This was necessary to avoid inconsistency as
some integrity checks are bypassed. This is inline with the
fact that truncate is primarily a DDL operation that was
designed to quickly remove all data from a table.
This patch implements I_S.PLUGINS.LOAD_OPTION column
as specified by WL#5496.
mysql-test/r/plugin_load_option.result:
A test case for WL#5496.
mysql-test/suite/funcs_1/r/is_columns_is.result:
Adjusted a test case according to WL#5496.
mysql-test/t/plugin_load_option.test:
A test case for WL#5496.
sql/sql_show.cc:
Added LOAD_OPTION column to I_S.PLUGINS.
Fall back to use ALTER TABLE for engines that doesn't support REPAIR when doing repair for upgrade.
Nicer output from mysql_upgrade and mysql_check
Updated all arrays that used NAME_LEN to use SAFE_NAME_LEN to ensure that we don't break things accidently as names can now have a #mysql50# prefix.
client/mysql_upgrade.c:
If we are using verbose, also run mysqlcheck in verbose mode.
client/mysqlcheck.c:
Add more information if running in verbose mode
Print 'Needs upgrade' instead of complex error if table needs to be upgraded
Don't write connect information if verbose is not 2 or above
mysql-test/r/drop.result:
Updated test and results as we now support full table names
mysql-test/r/grant.result:
Now you get a correct error message if using #mysql with paths
mysql-test/r/show_check.result:
Update results as table names can temporarly be bigger than NAME_LEN (during upgrade)
mysql-test/r/upgrade.result:
Test upgrade for long table names.
mysql-test/suite/funcs_1/r/is_tables_is.result:
Updated old test result (had note been updated in a while)
mysql-test/t/drop.test:
Updated test and results as we now support full table names
mysql-test/t/grant.test:
Now you get a correct error message if using #mysql with paths
mysql-test/t/upgrade.test:
Test upgrade for long table names.
sql/ha_partition.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/item.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/log_event.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/mysql_priv.h:
Added SAFE_NAME_LEN
sql/rpl_filter.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/sp.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/sp_head.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/sql_acl.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/sql_base.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/sql_connect.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/sql_parse.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/sql_prepare.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/sql_select.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/sql_show.cc:
NAME_LEN -> SAFE_NAME_LEN
Enlarge table names for SHOW TABLES to also include optional #mysql50#
sql/sql_table.cc:
Fall back to use ALTER TABLE for engines that doesn't support REPAIR when doing repair for upgrade.
sql/sql_trigger.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/sql_udf.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/sql_view.cc:
NAME_LEN -> SAFE_NAME_LEN
sql/table.cc:
Fixed check_table_name() to not count #mysql50# as part of name
If #mysql50# is part of the name, don't allow path characters in name.
- Changed to still use bcmp() in certain cases becasue
- Faster for short unaligneed strings than memcmp()
- Bettern when using valgrind
- Changed to use my_sprintf() instead of sprintf() to get higher portability for old systems
- Changed code to use MariaDB version of select->skip_record()
- Removed -%::SCCS/s.% from Makefile.am:s to remove automake warnings
Remove acquisition of LOCK_open around file system operations,
since such operations are now protected by metadata locks.
Rework table discovery algorithm to not require LOCK_open.
No new tests added since all MDL locking operations are covered
in lock.test and mdl_sync.test, and as long as these tests
pass despite the increased concurrency, consistency must be
unaffected.
mysql-test/t/disabled.def:
Disable NDB tests due to Bug#55799.
sql/datadict.cc:
No longer necessary to protect ha_create_table() with
LOCK_open. Serial execution is now ensured by metadata
locks.
sql/ha_ndbcluster.cc:
Do not manipulate with LOCK_open in cluster code.
sql/ha_ndbcluster_binlog.cc:
Do not manipulate with LOCK_open in cluster code.
sql/ha_ndbcluster_binlog.h:
Update function signature.
sql/handler.cc:
Implement ha_check_if_table_exists().
@todo: some engines provide ha_table_exists_in_engine()
handlerton call, for those we perhaps shouldn't
call ha_discover(), to be more efficient.
Since currently it's only NDB, postpone till
integration with NDB.
sql/handler.h:
Declare ha_check_if_table_exists() function.
sql/mdl.cc:
Remove an obsolete comment.
sql/sql_base.cc:
Update to a new signature of close_cached_tables():
from now on we always call it without LOCK_open.
Update comments.
Remove get_table_share_with_create(), we should
not attempt to create a table under LOCK_open.
Introduce get_table_share_with_discover() instead,
which would request a back off action if the table
exists in engine.
Remove acquisition of LOCK_open for
data dictionary operations, such as check_if_table_exists().
Do not use get_table_share_with_create/discover for views,
where it's not needed.
Make tdc_remove_table() optionally acquire LOCK_open
to simplify usage of this function.
Use the right mutex in the partitioning code when
manipulating with thd->open_tables.
sql/sql_base.h:
Update signatures of changes functions.
sql/sql_insert.cc:
Do not wrap quick_rm_table() with LOCK_open acquisition,
this is unnecessary.
sql/sql_parse.cc:
Update to the new calling convention of tdc_remove_table().
Update to the new signature of close_cached_tables().
Update comments.
sql/sql_rename.cc:
Update to the new calling convention of tdc_remove_table().
Remove acquisition of LOCK_open around filesystem
operations.
sql/sql_show.cc:
Remove get_trigger_table_impl().
Do not acquire LOCK_open for a dirty read of the trigger
file.
sql/sql_table.cc:
Do not acquire LOCK_open for filesystem operations.
sql/sql_trigger.cc:
Do not require LOCK_open for trigger file I/O.
sql/sql_truncate.cc:
Update to the new signature of tdc_remove_table().
sql/sql_view.cc:
Do not require LOCK_open for view I/O.
Use tdc_remove_table() to expel view share.
Update comments.
sql/sys_vars.cc:
Update to the new signature of close_cached_tables().
Remove acquisition of LOCK_open around file system operations,
since such operations are now protected by metadata locks.
Rework table discovery algorithm to not require LOCK_open.
No new tests added since all MDL locking operations are covered
in lock.test and mdl_sync.test, and as long as these tests
pass despite the increased concurrency, consistency must be
unaffected.
Changed to use longlong10_to_str() instead of longlong2str() when base is 10 or -10 as former is much faster than later
Changed my_vsnprintf() to use longlong2str instead of int2str() to get rid of warnings and to get support for long pointers even when long is 32 bit.
client/mysqltest.cc:
longlong2str() -> longlong10_to_str()
include/m_string.h:
Added extra argument to longlong2str() to make it have same prototype is int2str()
mysys/charset.c:
Fixed compiler warning
mysys/mf_soundex.c:
Fixed compiler warning
mysys/my_getopt.c:
longlong2str() -> longlong10_to_str()
sql/create_options.cc:
Fixed compiler warning
sql/item_strfunc.cc:
Added extra argument to longlong2str
sql/opt_range.cc:
longlong2str() -> longlong10_to_str()
sql/partition_info.cc:
longlong2str() -> longlong10_to_str()
sql/slave.cc:
longlong2str() -> longlong10_to_str()
sql/sql_bitmap.h:
Added extra argument to longlong2str
sql/sql_partition.cc:
Added extra argument to longlong2str
sql/sql_select.cc:
longlong2str() -> longlong10_to_str()
sql/sql_show.cc:
Added extra argument to longlong2str
storage/innodb_plugin/handler/ha_innodb.cc:
Update to new parameters for longlong2str()
storage/maria/ma_dbug.c:
longlong2str() -> longlong10_to_str()
storage/maria/maria_chk.c:
Added extra argument to longlong2str
storage/myisam/mi_dbug.c:
longlong2str() -> longlong10_to_str()
storage/myisam/myisamchk.c:
Added extra argument to longlong2str
storage/xtradb/handler/ha_innodb.cc:
Update to new parameters for longlong2str()
strings/longlong2str.c:
Added extra argument to longlong2str() to make it have same prototype is int2str()
strings/my_vsnprintf.c:
Changed my_vsnprintf() to use longlong2str instead of int2str() to get rid of warnings and to get support for long pointers even when long is 32 bit.
Added cast to get rid of compiler warnings
