mariadb

Tree: 014e9105ea

Author	SHA1	Message	Date
Monty	014e9105ea	Return >= 1 from matching_candidates_in_table if records > 0.0 Having rows >= 1.0 helps ensure that when we calculate total rows of joins the number of resulting rows will not be less after the join. Changes in test cases: - Join order change for some tables with few records - 'Filtered' is much higher for tables with few rows, as 1 row is a high procent of a table with few rows.	4 years ago
Monty	4235c65e98	Update matching_candidates_in_table() to treat all conditions similar Fixed also that the 'with_found_constraint parameter' to matching_candidates_in_table() is as documented: It is now true only if there is a reference to a previous table in the WHERE condition for the current examined table (as it was originally documented) Changes in test results: - Filtered was 25% smaller for some queries (expected). - Some join order changed (probably because the tables had very few rows). - Some more table scans, probably because there would be fewer returned rows. - Some tests exposes a bug that if there is more filtered rows, then the cost for table scan will be higher. This will be fixed in a later commit.	4 years ago
Monty	5a33de8d04	Fix calculation of selectivity calculate_cond_selectivity_for_table() is largely rewritten: - Process keys in the order of rows found, smaller ranges first. If two ranges has equal number of rows, use the one with more key parts. This helps us to mark more used fields to not be used for further selectivity calculations. See cmp_quick_ranges(). - Ignore keys with fields that where used by previous keys - Don't use rec_per_key[] to calculate selectivity for smaller secondary key parts. This does not work as rec_per_key[] value is calculated in the context of the previous key parts, not for the key part itself. The one exception is if the previous key parts is a constant. Other things: - Ensure that select->cond_selectivity is always between 0 and 1. - Ensure that select->opt_range_condition_rows is never updated to a higher value. It is initially set to the number of rows in table. - We know store in table->opt_range_condition_rows the lowest number of rows that any row-read-method has found so far. Before it was only done for QUICK_SELECT_I::QS_TYPE_ROR_UNION and QUICK_SELECT_I::QS_TYPE_INDEX_MERGE. Now it is done for a lot more methods. See calculate_cond_selectivity_for_table() for details. - Calculate and use selectivity for the first key part of a multiple key part if the first key part is a constant. WHERE key1_part1=5 and key2_part1=5. IF key1 is used, then we can still use selectivity for key2 Changes in test results: - 'filtered' is slighly changed, usually to something slightly smaller. - A few cases where for group by queries the table order changed. This was because the number of resulting rows from a group by query with MIN/MAX is now set to be smaller. - A few index was changed as we know prefer index with more key parts if the number of resulting rows is the same.	4 years ago
Monty	046514ba87	Fixed bug in SQL_SELECT_LIMIT We where comparing costs when we should be comparing number of rows that will be examined	4 years ago
Monty	8c64f1e846	Simple optimization to speed up some handler functions when checking killed - Avoid checking for has_transactions if killed flag is not checked - Simplify code (Have checked with gcc -O3 that there is improvements) - Added handler::fast_increment_statstics() to be used when a handler functions wants to increase two statistics for one row access. - Made check_limit_rows_examened() inline (even if it didn't make any difference for gcc 7.5.0), still the right thing to do	4 years ago
Monty	ef40093018	Adjusted Range_rowid_filter_cost_info lookup cost slightly. If the array size would be 1, the cost would be 0 which is wrong. Fixed by adding a small (0.001) base value to the lookup cost. This causes not changes in any result files.	4 years ago
Monty	98d2c0a311	Change class variable names in rowid_filter to longer, more clear names No code logic changes was done a -> gain b -> cost_of_building_range_filter a_adj -> gain_adj r -> row_combinations Other things: - Optimized the layout of class Range_rowid_filter_cost_info. One effect was that I moved key_no to the private section to get better alignment and had to introduce a get_key_no() function. - Indentation changes in rowid_filter.cc to avoid long rows.	4 years ago
Monty	13e5a5816a	Updated convert-debug-for-diff	4 years ago
Monty	c3f07a6ef2	Fixed crashing when using DBUG_PUSH_EMPTY DBUG_PUSH_EMPTY is used by thr_mutex.cc. If there are 4G of DBUG_PUSH_EMPTY calls, then DBUG_POP_EMPTY will cause a crash when DBUGCloseFile() will try to free an object that was never allocated.	4 years ago
Monty	c1e3f5ffa8	Optimizer code cleanups, no logic changes - Updated comments - Added some extra DEBUG - Indentation changes and break long lines - Trivial code changes like: - Combining 2 statements in one - Reorder DBUG lines - Use a variable to store a pointer that is used multiple times - Moved declaration of variables to start of loop/function - Removed dead or commented code - Removed wrong DBUG_EXECUTE code in best_extension_by_limited_search()	4 years ago
Monty	e71a7b212c	Limit calculated rows to the number of rows in the table The result file changes are mainly that number of rows is one smaller for some queries with DISTINCT or GROUP BY	4 years ago
Monty	4c755518d3	Ensure that test_quick_select doesn't return more rows than in the table Other changes: - In test_quick_select(), assume that if table->used_stats_records is 0 then the table has 0 rows. - Fixed prepare_simple_select() to populate table->used_stat_records - Enusre that set_statistics_for_tables() doesn't cause used_stats_records to be 0 when using stat_tables. - To get blackhole to work with replication, set stats.records to 2 so that test_quick_select() doesn't assume the table is empty.	4 years ago
Monty	9e8fa07f38	MDEV-14907 FEDERATEDX doesn't respect DISTINCT This is a minor cleanup of the original commit	4 years ago
Monty	dd458e5291	Improve comments in the optimizer	4 years ago
Monty	429fbccafa	Make Rocksdb error message independent from engine messages in my_base.h - Rocksdb unique error messages move to a range starting at 300 - Use same error code for Rocksdb error messages that exists in my_base.h	4 years ago
Marko Mäkelä	2d8e38bc94	Merge 10.6 into 10.7	4 years ago
Marko Mäkelä	4e1ca38838	MDEV-26781 InnoDB hangs when using SUX_LOCK_GENERIC The Shared/Update/Exclusive locks that were introduced in commit `03ca6495df` (MDEV-24142) did not work correctly when a futex-like system call interface was not available. On all tested implementations (IBM AIX as well as FreeBSD and GNU/Linux with the futex interface artificially disabled), the old implementation would cause hangs in some SPATIAL INDEX tests (innodb_gis suite). On FreeBSD, a hang was also observed in an encryption test. We will simply emulate the futex system calls with a single mutex and two condition variables, one for each wait queue. The condition variables basically shadow the lock words and are used as wait queues, just like the futex system calls would be. The storage overhead of ssux_lock_impl will be increased by 32 bits when using SUX_LOCK_GENERIC. Compared to the futex-based implementation, the SUX_LOCK_GENERIC implementation has an overhead of sizeof(pthread_mutex_t)+2*sizeof(pthread_cond_t). rw_lock: Remove all SUX_LOCK_GENERIC extensions. pthread_mutex_wrapper: A simple wrapper of pthread_mutex that implements srw_spin_mutex and srw_mutex for SUX_LOCK_GENERIC. srw_mutex_impl: Define this also for SUX_LOCK_GENERIC, but in that case add the fields mutex, cond. ssux_lock_impl: Define for SUX_LOCK_GENERIC with a minimal difference: adding readers_cond.	4 years ago
Marko Mäkelä	ff99413804	MDEV-25975: Merge 10.5 into 10.6	4 years ago
Marko Mäkelä	9d94c60f2b	Merge 10.5 into 10.6	4 years ago
Marko Mäkelä	fd6a464ae5	MDEV-13005 after-merge fixup	4 years ago
Marko Mäkelä	5d8dcfd86c	MDEV-25975: Merge 10.4 into 10.5	4 years ago
Marko Mäkelä	cbdf62ae90	MDEV-25975 merge fixup	4 years ago
Marko Mäkelä	cacb61b6be	Merge 10.4 into 10.5	4 years ago
Marko Mäkelä	d172df9913	MDEV-25975: Merge 10.3 into 10.4	4 years ago
Marko Mäkelä	f089f8d95e	MDEV-23328 fixup: sign mismatch in format strings kill_one_thread(): Fix integer sign mismatch in some format strings. Some of this was introduced in commit `5c230b21bf`	4 years ago
Marko Mäkelä	d6d66c6e90	Merge 10.3 into 10.4	4 years ago
sjaakola	d6758efbe1	MDEV-23328 Server hang due to Galera lock conflict resolution Cherry-pick the sql_kill and sql_user_kill from `ef2dbb8dbc` Changed ER_CANNOT_USER to ER_KILL_DENIED_ERROR to match other kill denied user messages. Cherry-pick by Daniel Black. Reviewed-by: Jan Lindström <jan.lindstrom@mariadb.com>	4 years ago
Faustin Lammler	aed87f1e0a	Change MDBF mirror URL As as discussed with @dbart we decided to move https://deb.mariadb.org/repo to https://deb.mariadb.org so that a request on https://deb.mariadb.org will end up directly in https://mirror.mariadb.org/repo. The same behavior was configured for rpms (https://rpm.mariadb.org --> https://mirror.mariadb.org/yum) this also makes the old yum.mariadb.org service transparently compatible with this new service (after DNS change to point to rpm.mariadb.org). See also: https://jira.mariadb.org/browse/MDBF-297	4 years ago
Marko Mäkelä	e9735a8185	MDEV-25975 innodb_disallow_writes causes shutdown to hang We will remove the parameter innodb_disallow_writes because it is badly designed and implemented. The parameter was never allowed at startup. It was only internally used by Galera snapshot transfer. If a user executed SET GLOBAL innodb_disallow_writes=ON; the server could hang even on subsequent read operations. During Galera snapshot transfer, we will block writes to implement an rsync friendly snapshot, as follows: sst_flush_tables() will acquire a global lock by executing FLUSH TABLES WITH READ LOCK, which will block any writes at the high level. sst_disable_innodb_writes(), invoked via ha_disable_internal_writes(true), will suspend or disable InnoDB background tasks or threads that could initiate writes. As part of this, log_make_checkpoint() will be invoked to ensure that anything in the InnoDB buf_pool.flush_list will be written to the data files. This has the nice side effect that the Galera joiner will avoid crash recovery. The changes to sql/wsrep.cc and to the tests are based on a prototype that was developed by Jan Lindström. Reviewed by: Jan Lindström	4 years ago
Marko Mäkelä	7c584d8270	Merge 10.2 into 10.3	4 years ago
Daniel Black	b249abde57	Merge branch '10.6' into 10.7 Closes #2082	4 years ago
Tuukka Pasanen	b4f3969a6c	MDEV-28121: Add new compression plugins as autopkg dependencies Add compression plugins: * bzip2 * lz4 * lzma * lzo * snappy To Autopkg's basic smoke-test dependencies. Also make sure that MariaDB service is reloaded when smoke-test is executed.	4 years ago
Dmitry Shulga	f6b09a7ce5	MDEV-21173: Assertion `m_thd == __null' failed in sp_head::~sp_head Some SQL statements that involves subqueries or stored routines could fail since execution of subqueries or stored routines is not supported for theses statements. Unfortunately, parsing error could result in abnormal termination by firing the following assert DBUG_ASSERT(m_thd == NULL); in a destructor of the class sp_head. The reason of the assert firing is that the method sp_head::restore_thd_mem_root() is not called on semantic action code to clean up resources allocated during parsing. This happens since the macros YYABORT is called instead of MYSQL_YYABORT by semantic action code for some grammar rules. So, to fix the bug YYABORT was just replaced with MYSQL_YYABORT.	4 years ago
Sergei Golubchik	2d2c3da8ec	MDEV-27673 Warning after "select progress from information_schema.processlist" after moving fields in optimize_schema_tables_memory_usage() store default values into their new, moved, locations.	4 years ago
Sergei Golubchik	cf8d30efd2	Revert "MDEV-28131 Unexpected warning while selecting from information_schema.processlist" This reverts commit `0812d0de8d`. But keeps the test case.	4 years ago
Sergei Golubchik	d7fd76456e	MDEV-19525 fix the test for embedded followup for `58cd2a8ded`	4 years ago
Marko Mäkelä	8d9c2561cd	Merge 10.5 into 10.6	4 years ago
Otto Kekäläinen	8c919330a7	Deb: Sync Salsa-CI from Debian MariaDB 10.6 As MariaDB 10.5 has been removed from Debian Sid and MariaDB 10.6 has entered it, the Salsa-CI testing needs to adapt. To achieve this, essentially sync most the the salsa-ci.yml contents from https://salsa.debian.org/mariadb-team/mariadb-server/-/tree/debian/latest This includes removing Stretch builds, as Stretch does not support uring nor pmem libraries, which MariaDB 10.6 depends on. Also add a couple Lintian overrides to make Salsa-CI pass. NOTE TO MERGERS: This commit is made on 10.6 branch and can be merged to all later branches (10.7, 10.8, 10.9..) for now, but later somebody needs to go in and update all the testing stages to do the upgrade testing correctly for 10.6->10.7->10.8->10.9 etc.	4 years ago
Thirunarayanan Balathandayuthapani	db655e1310	MDEV-28237 Assertion `0' failed in row_upd_sec_index_entry on DELETE - InnoDB bulk insert operation fails to rollback when it detect DB_DUPLICATE_KEY error. It leads to orphaned records in primary indexes. Consecutive update/delete operation assumes that record should exist in secondary index and it leads to failure.	4 years ago
Thirunarayanan Balathandayuthapani	b678f8811b	MDEV-28138 MariaDB Assertion Failed in mtr_buf_t::has_space - After MDEV-24621, InnoDB does buffer the insert bulk operation for all indexes expect spatial one. But it leads to search the primary key lookup and it leads to failure. So InnoDB should avoid bulk insert when table has spatial index involved.	4 years ago
Monty	c4ebb2bd04	Fixed that mysql_upgrade doesn't give errors about mariadb.sys The reason for this fix was that when I tried to run mysql_upgrade at home to update an old 10.5 installation, mysql_upgrade failed with warnings about mariadb.sys user not existing. If the server was started with --skip-grants, there would be no warnings from mysql_upgrade, but in some cases running mysql_upgrade again could produce new warnings. The reason for the warnings was that any access of the mysql.user view will produce a warning if the mariadb.sys user does not exists. Fixed with the following changes: - Disable warnings about mariadb.sys user not existing - Don't overwrite old mariadb.sys entries in tables_priv and global_priv - Ensure that tables_priv has an entry for mariadb.sys if the user exists. This fixes an issue that tables_priv would not be updated if there was a failure directly after global_priv was updated.	4 years ago
Monty	09c7f78c2e	Fixed double free issue in events Server crashed during shutdown with: "corrupted double-linked list" when running mysql_upgrade multiple times against the server. Reason was that db_repostitory could be freed twice.	4 years ago
Julius Goryavsky	daed558b2c	MDEV-28204: The tr utility does not work as expected on rsync SST This commit contains a fix to use modern syntax for selecting character classes in the tr utility options. Also one of the tests for SST via rsync (galera_sst_rysnc2) is made more reliable (to avoid rare failures during automatic testing).	4 years ago
Vlad Lesin	6a3545dd1e	MDEV-26322 Last binlog file and position are "empty" in mariabackup --prepare output The issue is caused by `59a0236da4` commit. The initial intention of the commit was to speed up "mariabackup --prepare". The call stack of binlog position reading is the following: ▾ trx_rseg_mem_restore ▾ trx_rseg_array_init ▾ trx_lists_init_at_db_start ▸ srv_start Both trx_lists_init_at_db_start() and trx_rseg_mem_restore() contain special cases for srv_operation == SRV_OPERATION_RESTORE condition, and on this condition only rseg headers are read to parse binlog position. Performance impact is not so big. The solution is to revert `59a0236da4`.	4 years ago
Daniel Black	7f5a3cd253	MDEV-28231: innodb: format string warning on aix UINT64PFx (ib_id_t) ib_id_t is a uint64. On AIX this isn't a long long unsigned and to prevent the compile warnings and potential wrong type, the UINT64PFx defination is corrected. As INT64PF is unused (last use, xtradb in 10.2), it is removed to remove the confusion that INT64PF and UINT64PFx would be different types otherwise.	4 years ago
Dmitry Shulga	cd56b40f6d	MDEV-28129: MariaDB UAF issue at lex_end_nops(LEX*) This bug report is not about ASAN Use After Free issue. This bug is about missed calling of the method LEX::cleanup_lex_after_parse_error that should happen on parse error. Aforementioned method calls sphead::restore_thd_mem_root to clean up resources acquired on processing a stored routine. Particularly, the method sp_head::restore_tht_mem_root is called to restore an original mem root and reset LEX::sphead into nullptr. The method LEX::cleanup_lex_after_parse_error is invoked by the macros MYSQL_YYABORT. Unfortunately, some rules of grammar for handling user variables in SQL use YYABORT instead of MYSQL_YYABORT to handle parser errors. As a consequence, in case a statement with setting of a user variable is called inside a stored routine, it results in assert failure in sp_head destructor. To fix the issue the macros YYABORT should be replaced by MYSQL_YYABORT in those grammar rules that handle assignment of user variables.	4 years ago
Alexander Barkov	0ffaf19c53	Adding a "const" qualifier to arguments of create_func(), create_native() etc The "const" qualifier was obviously forgotten. This change will also simpily fixing of MDEV-27744.	4 years ago
Alexander Barkov	d271fbd392	MDEV-28224 error: cannot initialize return object of type 'bool' with an rvalue of type 'nullptr_t' Fixing a typo in the fix for MDEV-19804, wrong return value in a bool function: < return NULL; > return true; The problem was found because it did not compile on some platforms. Strangley, it did not have visible problems on other platforms, which did not fail to compile, although "return NULL" should compile to "return false" rather than "return true".	4 years ago
Daniel Black	75b9014fed	MDEV-26136: Correct AIX/macOS cast warning (my_time.h) tv_usec is a (suseconds_t) so we cast to it. Prevents the AIX(gcc-10) warning: include/my_time.h: In function 'void my_timeval_trunc(timeval*, uint)': include/my_time.h:249:65: warning: conversion from 'long int' to 'suseconds_t' {aka 'int'} may change value [-Wconversion] 249 \| tv->tv_usec-= my_time_fraction_remainder(tv->tv_usec, decimals); \| macOS is: conversion from 'long int' to '__darwin_suseconds_t' {aka 'int'} may change value On Windows suseconds_t isn't defined so we use the existing long return type of my_time_fraction_remainder. Reviewed by Marko Mäkelä Closes: #2079	4 years ago
Sergei Golubchik	d48774e0e0	MDEV-27354 Change maturity of plugins	4 years ago

1 2 3 4 5 ...

195178 Commits (014e9105ea394f218c58fc2dd90fc3e0837e7d26) All Branches Search

195178 Commits (014e9105ea394f218c58fc2dd90fc3e0837e7d26)

All Branches