- 14 May, 2019 1 commit
-
-
Oleksandr Byelkin authored
-
- 13 May, 2019 5 commits
-
-
Elena Stepanova authored
-
Vladislav Vaintroub authored
-
Vladislav Vaintroub authored
String is not guaranteed to be null-terminated, thus push_warning_printf also needs to specify the length.
-
Vladislav Vaintroub authored
-
Marko Mäkelä authored
In the merge c51f85f8 the files row0trunc.cc and read0read.cc were not supposed to be changed in MariaDB 10.3. The row0trunc.cc code that was changed in MariaDB 10.2 commit 3db94d24 does not exist, because it was removed earlier in 10.3 commit 5a1868b5. When it comes to read0read.cc, as explained in the 10.2 commit e0271a7b the code had been moved to trx0sys.h in 10.3. In 10.3 commit 380069c2 appproriate mutex protection was added to checking the transaction system state. The assertion expression was properly updated in the earlier merge commit acf6f92a.
-
- 12 May, 2019 1 commit
-
-
Oleksandr Byelkin authored
-
- 10 May, 2019 9 commits
-
-
Monty authored
-
Marko Mäkelä authored
row_purge_upd_exist_or_extern_func(): Check for node->vcol_op_failed() after row_purge_remove_sec_if_poss(), like row_purge_del_mark() did. This avoids us dereferencing the node->table=NULL pointer. The test case, submitted by Elena Stepanova, is not deterministic and does not repeat the bug on 10.2. With the added loop, for me, it reliably crashes 10.3 without the fix. I was unable to create a deterministic test case for either 10.2 or 10.3. Reviewed by Thirunarayanan Balathandayuthapani
-
Marko Mäkelä authored
-
Thirunarayanan Balathandayuthapani authored
buf_dblwr_process(): Remove the useless warning that a copy of a page in the doublewrite buffer is corrupted. We already report an error if a corrupted page cannot be recovered from the doublewrite buffer. Note: In MariaDB 10.1, the original bug reported in MDEV-13893 could still be easily repeatable. In MariaDB 10.2.24, MDEV-12699 should have reduced the probability considerably.
-
Marko Mäkelä authored
fts_get_table_name(): Add the parameter bool dict_locked=false.
-
Marko Mäkelä authored
fts_get_table_name(): Output to a caller-allocated buffer. fts_get_table_name_prefix(): Use the lower-overhead allocation ut_malloc() instead of mem_alloc(). This is based on mysql/mysql-server@d1584b9f38ff0bcf609d181db35f74108e022168 in MySQL 5.7.4.
-
Marko Mäkelä authored
fts_table_t::parent: Remove the redundant field. Refer to table->name.m_name instead. fts_update_sync_doc_id(), fts_update_next_doc_id(): Remove the redundant parameter table_name. fts_get_table_name_prefix(): Access the dict_table_t::name. FIXME: Ensure that this access is always covered by dict_sys->mutex.
-
Marko Mäkelä authored
fts_state_t, fts_slot_t::state: Remove. Replaced by fts_slot_t::running and fts_slot_t::table_id as follows. FTS_STATE_SUSPENDED: Removed (unused). FTS_STATE_EMPTY: Removed. table_id=0 will denote empty slots. FTS_STATE_RUNNING: Equivalent to running=true. FTS_STATE_LOADED, FTS_STATE_DONE: Equivalent to running=false. fts_slot_t::table: Remove. Tables will be identified by table_id. After opening a table, we will check fil_table_accessible() before accessing the data. fts_optimize_new_table(), fts_optimize_del_table(), fts_optimize_how_many(), fts_is_sync_needed(): Remove the parameter tables, and use the static variable fts_slots (which was introduced in MariaDB 10.2) instead.
-
Eugene Kosov authored
No functional change. Call my_timer_init() only once and then reuse it from InnoDB and perfschema storage engines. This patch speeds up empty test for me like this: ./mtr -mem innodb.kevg,xtradb 1.21s user 0.84s system 34% cpu 5.999 total ./mtr -mem innodb.kevg,xtradb 1.12s user 0.60s system 31% cpu 5.385 total
-
- 09 May, 2019 7 commits
-
-
Daniel Bartholomew authored
-
Vladislav Vaintroub authored
The crash happens when writing into log file. The reason is likely that the call to WriteFile() was missing a valid parameter for lpNumberOfBytesWritten. This seems only to happen on ancient version of Windows. Since the fix to MDEV-16430 in 141bc58a, null pointer was passed instead of valid pointer. The fix is to provide a valid lpNumberOfBytesWritten parameter.
-
Marko Mäkelä authored
Avoid accessing the table cache while the ALTER TABLE statement is blocked by DEBUG_SYNC. Use explicit COMMIT for forcing the redo log flush (whose main purpose is to ensure that the incomplete state of the blocked ALTER TABLE statement is persisted).
-
Marko Mäkelä authored
Ensure that the 'auxiliary transactions' that are there for flushing the incomplete undo log of the to-be-recovered DDL transactions are actually making modifications. This is a backport of 2fe40a7a from MariaDB 10.4.
-
Rasmus Johansson authored
-
Sergey Vojtovich authored
-
Sergey Vojtovich authored
A sequel to 9180e866 and 149b7547. ALTER TABLE ... ADD FOREIGN KEY may crash if parent table is updated concurrently. Block FK parent table updates even earlier, before intermediate child table is created. Use proper charset info for my_casedn_str() and don't update original identifiers so that lower_cast_table_names == 2 is honoured.
-
- 08 May, 2019 9 commits
-
-
Jan Lindström authored
-
Igor Babaev authored
query with VALUES() A table value constructor can be used in all contexts where a select can be used. In particular an ORDER BY clause or a LIMIT clause or both of them can be attached to a table value constructor to produce a new query. Unfortunately execution of such queries was not supported. This patch fixes the problem.
-
Daniel Bartholomew authored
-
Marko Mäkelä authored
-
Andrea Kao authored
GitHub uses a library called Licensee to identify a project's license type. It shows this information in the status bar and via the API if it can unambiguously identify the license. This commit modifies a few of MariaDB's docs so that Licensee is able to recognize the repository's license type. It renames the README's "License" section to "Licensing" and renames COPYING.thirdparty to THIRDPARTY. These changes allow Licensee to bypass both files when it scans the repo for license files, which thus allows Licensee to successfully identify the license type of MariaDB as GPL 2.0. Signed-off-by: Andrea Kao <eirinikos@gmail.com>
-
Sergey Vojtovich authored
Causes valgrind warnings, harmless otherwise. Part of MDEV-14984 - regression in connect performance
-
Galina Shalygina authored
The bug occurs because Item_func_set_user var is allowed to be pushed into materialized derived table/view. To fix it excl_dep_on_table() as added to Item_func_set_user_var class to prevent pushdown.
-
Thirunarayanan Balathandayuthapani authored
Fix: ==== 1) Combined innodb_ft_result_cache_limit_32.test and innodb_ft_result_cache_limit_64.test test case in sys_vars suite. 2) Use word_size.inc for combinations of innodb_ft_result_cache_limit test case.
-
Marko Mäkelä authored
ReadView::copy_trx_ids(): Relax a debug check. It failed to account for TRX_STATE_PREPARED_RECOVERED, which was introduced in MDEV-15772. It was also reading trx->state twice and failed to tolerate TRX_STATE_COMMITTED_IN_MEMORY, which could be concurrently assigned in lock_trx_release_locks(), which is not holding trx_sys->mutex. This bug is specific to the MariaDB 10.2 series. The ReadView was introduced in MariaDB 10.2.2 by merging the code that had been introduced in MySQL 5.7.2. In MariaDB 10.3, ReadView::snapshot() would use the lock-free trx_sys.rw_trx_hash. MDEV-14638 moved the corresponding assertion to trx_sys_t::find(), where it was duly protected by trx->mutex, and later MDEV-14756 moved the check to rw_trx_hash_t::validate_element(). This check was correctly adjusted when MDEV-15772 was merged to 10.3.
-
- 07 May, 2019 8 commits
-
-
Elena Stepanova authored
-
Daniel Bartholomew authored
-
Sergei Golubchik authored
-
Sergei Golubchik authored
SHOW STATUS LIKE 'Open_files' was showing 18446744073709551615 my_file_opened used statistic_increment/statistic_decrement, so one-off errors were normal and expected. But they confused monitoring tools, so let's move my_file_opened to use atomics.
-
Sergei Golubchik authored
-
Sergei Golubchik authored
make live checksum to be returned in handler::info(), and slow table-scan checksum to be calculated in handler::checksum(). part of MDEV-16249 CHECKSUM TABLE for a spider table is not parallel and saves all data in memory in the spider head by default
-
Sergei Golubchik authored
regression after 3db6de33 in RPMs continue creating the $datadir outside of mysql_install_db. RPMs put the socket in there, so it cannot be chmod 0700.
-
Alexey Botchkov authored
Do not consider rows of Information_schema.files where LOGFILE_GROUP_NAME is NULL.
-