Commit graph

184259 commits

Author SHA1 Message Date
Dmitry Shulga
259e5243fa MDEV-24860: Incorrect behaviour of SET STATEMENT in case it is executed as a prepared statement
Running statements with SET STATEMENT FOR clause is handled incorrectly in
case the whole statement is executed in prepared statement mode.
For example, running of the following statement
  SET STATEMENT sql_mode = 'NO_ENGINE_SUBSTITUTION' FOR CREATE TABLE t1 AS SELECT CONCAT('abc') AS c1;
results in different definition of the table t1 depending on whether
the statement is executed as a prepared or as a regular statement.

In first case the column c1 is defined as
  `c1` varchar(3) DEFAULT NULL
in the last case the column c1 is defined as
  `c1` varchar(3) NOT NULL

Different definition for the column c1 arise due to the fact that
a value of the data memeber Item_func_concat::maybe_null depends on
whether strict mode is on or off. Below is definition of the method
fix_fields() of the class Item_str_func that is base class for the
class Item_func_concat that is created on parsing the
SET STATEMENT FOR clause.

bool Item_str_func::fix_fields(THD *thd, Item **ref)
{
  bool res= Item_func::fix_fields(thd, ref);
  /*
    In Item_str_func::check_well_formed_result() we may set null_value
    flag on the same condition as in test() below.
  */
  maybe_null= maybe_null || thd->is_strict_mode();
  return res;
}

Although the clause SET STATEMENT sql_mode = 'NO_ENGINE_SUBSTITUTION' FOR
is parsed on PREPARE phase during processing of the prepared statement,
real setting of the sql_mode system variable is done on EXECUTION phase.
On the other hand, the method Item_str_func::fix_fields is called on PREPARE
phase. In result, thd->is_strict_mode() returns true during calling the method
Item_str_func::fix_fields(), the data member maybe_null is assigned the value
true and column c1 is defined as DEFAULT NULL.

To fix the issue the system variables listed in the SET STATEMENT FOR clause
are set at the beginning of handling the PREPARE phase just right before
calling  the function check_prepared_statement() and their original values
restored immediate after return from this function.

Additionally, to avoid code duplication the source code used in the function
mysql_execute_command for setting variables, specified by SET STATEMENT
clause, were extracted to the standalone functions
run_set_statement_if_requested(). This new function is called from
the function  mysql_execute_command() and the method
Prepared_statement::prepare().
2021-02-25 14:36:09 +07:00
Varun Gupta
0a95c922c8 Fixed the innodb_ext_key test by adding replace_column 2021-02-25 12:42:01 +05:30
Daniel Black
577c970c52 MDEV-24728: Debian include client caching_sha2_password plugin
Backport of 4bc31a904f

Include client libraries for auth caching_sha2_password and
sha256_password in the libmariadb3 client library package.
2021-02-25 12:12:04 +11:00
Daniel Black
1635686b50 MDEV-23510: arm64 lf_hash alignment of pointers
volatile != atomic.

volatile has no memory barrier schemantics, its for mmaped IO
so lets allow some optimizer gains and stop pretending it helps
with memory atomicity.

The MDEV lists a SEGV an assumption is made that an address was
partially read. As C packs structs strictly in order and on arm64 the
cache line size is 128 bits. A pointer (link - 64 bits), followed
by a hashnr (uint32 - 32 bits), leaves the following key (uchar *
64 bits), neither naturally aligned to any pointer and worse, split
across a cache line which is the processors view of an atomic
reservation of memory.

lf_dynarray_lvalue is assumed to return a 64 bit aligned address.

As a solution move the 32bit hashnr to the end so we don't get the
*key pointer split across two cache lines.

Tested by: Krunal Bauskar
Reviewer: Marko Mäkelä
2021-02-25 09:13:10 +11:00
Vicențiu Ciorbaru
9e259d58c2 Remove race condition during make dist
Introduced by 85828b8f22

This is running 2 git processes in parallel, which, if unlucky can cause
either of them to fail with "File already exists" error.
2021-02-23 17:26:08 +02:00
Thirunarayanan Balathandayuthapani
787c47586e MDEV-24913 Assertion !recv_no_log_write in log_write_up_to()
- The commit 5fd3c7471e3e0673b50d309567c9747d36f09412(MDEV-24709)
resets the recv_no_ibuf_operations in
recv_recovery_from_checkpoint_start(), but InnoDB fails to reset
the variable recv_no_log_write() during that time and that leads
to the assert failure.
2021-02-23 13:56:40 +05:30
Sergei Golubchik
3c021485c9 fix binlog_xa_recover test
1. wait for the binlog thread to reach the certain state, don't use
   a debug_sync that's incorrectly placed to detect the state
2. no need to do a (non-deterministic) `show binlog events` to verify
   what is guaranteed by the directly preceding line
2021-02-22 19:43:08 +01:00
Sergei Golubchik
bb98c6bf44 cleanup: renames, no need to create a new .inc file
if it's the whole content of a test anyway.
2021-02-22 19:43:08 +01:00
Sergei Golubchik
7fe351aba4 mtr fixes for old (5.10.1) perl 2021-02-22 19:43:08 +01:00
Sergei Golubchik
77c23c62ae support for mtr --valgdb
add a new "debugger" to mtr, that runs the executable
under valgrind in gdb. valgrind pid is auto-detected,
but the delay (sleep) and vgdb path are hard-coded for now
2021-02-22 19:43:08 +01:00
Sergei Golubchik
feacc0aaf2 unify mtr handling of debuggers
"debugger" is anything that wraps execution of a target
binary (mysqld or mysqltest). Currently the list includes:
gdb, ddd, dbx, lldb, valgrind, strace, ktrace, rr,
devenv, windbg, vsjitdebugger.

for every debugger xxx, mtr will recognize four options:
--xxx, --boot-xxx, --manual-xxx, --client-xxx.
They all support an optional "=string" argument. String
being a semicolon-separated list of commands (e.g. for gdb)
or one (not semicolon-separated) command line of options
(e.g. for valgrind). Or both (e.g. --gdb='-quiet -nh;info files'

In embedded both --xxx and --client-xxx work.

Functionality changed/removed:
* --rr-args is gone
* --rr-dir is gone
* --manual-debug is gone
* --debugger={devenv|vc|windbg|vc_express|vsjitdebugger} is gone
* --strace-option is gone
* --stracer={strace|ktrace} is gone
* --valgrind only enables it for the server, not for everything
* --valgrind-all is gone
* --valgrind-mysqltest is gone
* --valgrind-mysqld is gone
* --valgrind-options is gone
* --valgrind-option is gone
* --valgrind-path is gone
* --callgrind is gone
* one cannot combine --valgrind --gdb anymore
* valgrind report doesn't add a fake test line to the output
* vc and vcexpress on windows are no longer supported
2021-02-22 19:43:08 +01:00
Sergei Golubchik
3b0b4e614c cleanup: remove dead code in mtr 2021-02-22 19:27:12 +01:00
Sergei Golubchik
c4f0133444 cleanup: stat tables
don't allocate Column_statistics_collected objects that won't
be used.

minor style fixes (StringBuffer<>, etc)
2021-02-22 19:27:12 +01:00
Sergei Golubchik
06a791aa12 MDEV-23753: SIGSEGV in Column_stat::store_stat_fields
only collect persistent stats for columns explicitly listed
by the user in the  ANALYZE TABLE PERSISTENT FOR COLUMNS (...)
clause. The engine can extend table->read_set as much as
it wants, it should not affect the collected statistics.

Test case from the 3b94309a6c applies - it used to crash,
because ha_partition extended table->read_set after the loop that
initialized some objects based on bits in the read_set but before the
loop that used these objects based on bits in the read_set.
2021-02-22 19:27:12 +01:00
Sergei Golubchik
caad32ca92 Revert "MDEV-23753: SIGSEGV in Column_stat::store_stat_fields"
This reverts the commit 3b94309a6c but keeps the test

Because the fix is a hack that isn't supposed to do anything,
and relies on a side-effect of rnd_init inside ha_partition.

A different fix is coming up.
2021-02-22 19:27:12 +01:00
Sergei Golubchik
a638f1577a Merge branch 'bb-10.2-release' into 10.2 2021-02-22 18:43:03 +01:00
Daniel Bartholomew
6aa909745d
bump the VERSION 2021-02-22 10:04:25 -05:00
Thirunarayanan Balathandayuthapani
d7fc4f5236 MDEV-24863 AHI entries mismatch with the index while reloading the evicted tables.
- This is caused by commit ad6171b91c
(MDEV-22456). InnoDB reloads the evicted table again from dictionary.
In that case, AHI entries and current index object mismatches
happens. When index object mismatches then InnoDB should drop
the page hash AHI entries for the block. In
btr_search_drop_page_hash_index(), InnoDB should take exclusive
lock on the AHI latch if index is already freed to avoid the
freed memory access during buf_pool_resize()
2021-02-22 18:08:10 +05:30
Jan Lindström
374f4c3f8c Suppress warning on galera_ssl_upgrade test. 2021-02-22 13:57:28 +02:00
Sergei Golubchik
ce3a2a688d make @@wsrep_provider and @@wsrep_notify_cmd read-only
this should simplify run-time cluster management
2021-02-18 19:03:01 +01:00
Sergei Golubchik
cbbbdb9c3b don't allocate 64K on the stack 2021-02-18 12:16:11 +01:00
Jan Lindström
5ecaf52d42 MDEV-24873 : galera.galera_as_slave_ctas MTR failed: Assertion `(&(&LOCK_thd_data)->m_mutex)->count > 0 && pthread_equal(pthread_self(), (&(&LOCK_thd_data)->m_mutex)->thread)' failed in sql_class.cc on THD::awake(killed_state)
Problem was that thd::awake assumes now that you hold THD::LOCK_thd_data
so we need to keep it when we call wsrep_thd_awake from
wsrep_abort_transaction.
2021-02-17 10:28:37 +02:00
Jan Lindström
45e33e05e2 MDEV-24872 : galera.galera_insert_multi MTR failed: crash with SIGABRT
Problem was that we tried to lock THD::LOCK_thd_data after we have
acquired lock_sys mutex. This is against mutex ordering rules.
2021-02-17 10:28:37 +02:00
Jan Lindström
4d300ab1a8 MDEV-24867 : wsrep.variables MTR failed: Result length mismatch
Stabilize test case.
2021-02-17 10:28:37 +02:00
Marko Mäkelä
067465cd2f MDEV-15641 fixup: Make the test faster
Let us avoid the excessive allocation of explicit record locks
(a work-around of MDEV-24813) so that the test will execute
much faster under AddressSanitizer, MemorySanitizer, Valgrind.
2021-02-16 12:07:48 +02:00
Varun Gupta
3544643f09 MDEV-23291: SUM column from a derived table returns invalid values
The issue here was the read_set bitmap was not set for a field which
was used as a reference in an inner select.
We need to make sure that if we are in an inner select and we have
references from outer select then we update the table bitmaps for
such references.

Introduced a function in the class Item_subselect that would
update bitmaps of table for the references within a
subquery that are defined in outer selects.
2021-02-16 11:53:13 +05:30
Varun Gupta
7e9a6b7f09 MDEV-24779: main.subselect fails in buildbot with --ps-protocol
Follow-up fix to commit 26f5033(MDEV-23449)
The GROUP BY clause inside IN/ALL/ANY subquery is removed
when there is no aggregate function or HAVING clause in the subquery.

When the GROUP BY clause is removed, a subquery can also be removed
if it part of the GROUP BY clause. This is done inside the function
remove_redundant_subquery_clauses. Here we walk over the GROUP BY list
and remove a subselect from its unit via the callback function
eliminate_subselect_processor.

The issue here was that when the query was being re-executed it was trying
to reinitialize the select that was removed as stated above.
This is not required, so the fix would be to remove select_lex
both from tree lex structure and the global list of nodes so that
we don't do the reinitialization again.
2021-02-16 11:53:13 +05:30
Varun Gupta
a461e4d306 MDEV-19474: Histogram statistics are used even with optimizer_use_condition_selectivity=3
The issue here was histogram statistics were being used even when
the level of optimizer_use_condition_selectivity doesn't allow
usage of statistics from histogram.

The histogram statistics are read for a table only when
optimizer_use_condition_selectivity > 3. But the TABLE structure can be
stored in the internal table cache and be reused for the next query.
So in this case the histogram statistics will be available for the next query.

The fix would be to make sure to use the histogram statistics only when
optimizer_use_condition_selectivity > 3.
2021-02-16 11:53:13 +05:30
Marko Mäkelä
e926964cb8 Remove useless test innodb.innodb_bug60049
The test innodb.innodb_bug60049 used to check that the record
(ID,NAME)=(12,'SYS_FOREIGN_COLS') is the last record in the
secondary index of the system table SYS_TABLES.
But, ever since commit 2336558423
or mysql/mysql-server@082d59670f
that record no longer is the last one in the table!

The more recent test innodb.purge_secondary covers the purge
functionality much better.
2021-02-15 18:12:31 +02:00
Alice Sherepa
6f14230643 MDEV-11862 rpl.rpl_semi_sync_event_after_sync 'fails if there is no semisync plugin 2021-02-15 13:41:44 +01:00
Marko Mäkelä
da3211e487 MDEV-24763 fixup: Use deterministic ORDER BY 2021-02-12 14:03:25 +02:00
Marko Mäkelä
6f3f191cfa MDEV-24763 ALTER TABLE fails to rename a column in SYS_FIELDS
innobase_rename_column_try(): When renaming SYS_FIELDS records
for secondary indexes, try to use both formats of SYS_FIELDS.POS
as keys, in case the PRIMARY KEY includes a column prefix.

Without this fix, an ALTER TABLE that renames a column followed
by a server restart (or LRU eviction of the table definition
from dict_sys) would make the table inaccessible.
2021-02-12 09:48:36 +02:00
Marko Mäkelä
028ba10d0b MDEV-18468 fixup: Make test case robust w.r.t. deferred DROP TABLE 2021-02-12 09:41:15 +02:00
Julius Goryavsky
95003eab45 MDEV-19950: Galera test failure on galera_ssl_upgrade
The test requires adaptation for MariaDB, which is done
in this patch. In addition, this patch includes a fix for
the SST script startup code that adds escaping for special
characters on the command line (in case they are contained
in the arguments to mysqld). The fix does not require
separate tests, as the required tests are already part
of the mtr suite for Galera.
2021-02-11 13:44:01 +01:00
Jan Lindström
362dcf9e01 Update Galera disabled.def 2021-02-11 08:02:24 +02:00
Alexander Barkov
afc5bac49d MDEV-24790 CAST('0e1111111111' AS DECIMAL(38,0)) returns a wrong result 2021-02-08 16:19:45 +04:00
Marko Mäkelä
739abf5195 Make innodb_gis.rtree_purge run faster
A locking SELECT from an InnoDB table is very slow especially
in debug builds. Replacing some INSERT...SELECT should not reduce
the test coverage, because the test will still do DELETE
(which will acquire explicit record locks).
2021-02-07 14:31:48 +02:00
Olivier Bertrand
e7d9c1d498 Fix connect engine ppc64 fail 2021-02-05 18:34:06 +01:00
Olivier Bertrand
dc31627c2d Fix of connect engine crashes 2021-02-05 11:25:24 +01:00
Oleksandr Byelkin
c04ae0d365 Fix of crashes of connect engine.
Use size_t everywhere and remove suspicious expression.
2021-02-05 10:56:05 +01:00
Oleksandr Byelkin
87bf594bc5 Fix of random crashes of connect engine (probably depend on addresses used)- 2021-02-05 10:55:53 +01:00
Oleksandr Byelkin
716b0b44f8 Fix compiler warnings of the new connect engine. 2021-02-05 10:55:40 +01:00
Sujatha
eef4c5d378 MDEV-22741: *SAN: ERROR: AddressSanitizer: use-after-poison on address in instrings/strmake.c:36 from change_master (on optimized builds)
Problem:
========
CHANGE MASTER TO MASTER_USER='root', MASTER_SSL=0, MASTER_SSL_CA='',
  MASTER_SSL_CERT='', MASTER_SSL_KEY='', MASTER_SSL_CRL='',
  MASTER_SSL_CRLPATH='';

CHANGE MASTER TO MASTER_USER='root', MASTER_PASSWORD='', MASTER_SSL=0;

use-after-poison is reported for lex_mi->ssl_crl

File: sql_repl.cc

if (lex_mi->ssl_crl)
  strmake_buf(mi->ssl_crl, lex_mi->ssl_crl);

Analysis:
========
At the end of CHANGE MASTER statement execution, the LEX_MASTER_INFO
parameters are reset so that the next query will have a clean state. But
'ssl_crl' and 'ssl_crl_path' members of LEX_MASTER_INFO object are not
cleared during 'LEX_MASTER_INFO::reset'. Hence when a new CHANGE MASTER
statement is executed, the stale value of lex_mi->ssl_crl is used, so ASAN
reports use-after-poison.

Fix:
===
Clear 'ssl_crl' and 'ssl_crl_path' as part of 'reset'.
2021-02-03 12:18:29 +05:30
Oleksandr Byelkin
b3cecb7bfc Revert "Fix of warnings on aarch64 like:"
Fixed by the author in other way (char -> short)

This reverts commit 496f7090a8.
2021-02-02 10:37:54 +01:00
Oleksandr Byelkin
7f7e66147d Merge remote-tracking branch 'connect/10.2' into 10.2 2021-02-02 10:36:46 +01:00
Vladislav Vaintroub
6ede84f477 MDEV-24762 - HeidiSQL 11.2 2021-02-02 10:28:31 +01:00
Olivier Bertrand
571294f954 Fix failed test bson and xml 2021-02-02 00:03:07 +01:00
Varun Gupta
ceb3976191 Updating test results in rocksdb test suite after MDEV-11172 is fixed 2021-02-01 17:59:38 +05:30
Elena Stepanova
ff5186fd2b List of unstable tests for 10.2.37 release
Test code modifications and new failures from buildbot only
registered for the main suite. The rest was updated partially,
based on the status of existing JIRA items
2021-02-01 08:51:22 +01:00
Varun Gupta
26f5033555 MDEV-23449: alias do not exist and a query do not report an error
For an IN/ANY/ALL subquery without an aggregate function and HAVING clause,
the GROUP BY clause is removed.
Due to the GROUP BY list being removed, the invalid reference in the GROUP BY
clause was never resolved.
Remove the GROUP BY list only when the all the items in the GROUP BY list
are resolved.
Also removing the GROUP BY list later would not affect the extension that allows
using non-aggregated field in an aggregate function (when ONLY_FULL_GROUP_BY
is not set) because the GROUP BY list is removed only when their is
NO aggregate function in IN/ALL/ANY subquery.
2021-01-31 19:55:07 +05:30