Commit graph

199084 commits

Author SHA1 Message Date
Daniele Sciascia
e821c9fa7c MDEV-35281 SR transaction crashes with innodb_snapshot_isolation
Ignore snapshot isolation conflict during fragment removal, before
streaming transaction commits. This happens when a streaming
transaction creates a read view that precedes the INSERTion of
fragments into the streaming_log table. Fragments are INSERTed
using a different transaction. These fragment are then removed
as part of COMMIT of the streaming transaction. This fragment
removal operation could fail when the fragments were not part
the transaction's read view, thus violating snapshot isolation.
2024-11-29 08:06:32 +01:00
Alexander Barkov
fdb6db6b47 MDEV-29462 ASAN: heap-use-after-free in Binary_string::copy on DO CONVERT
Item_char_typecast::val_str_generic() uses Item::str_value as a buffer.
Item::val_str_ascii() also used Item::str_value as a buffer.
As a result, str_value tried to copy to itself.

Fixing val_str_ascii() to use a local buffer instead of str_value.
2024-11-28 16:34:32 +04:00
Jan Lindström
f39217da0c MDEV-35473 : Sporadic failures in the galera_3nodes.galera_evs_suspect_timeout mtr test
Remove unnecessary sleep and replace it with proper wait_conditions.

Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
2024-11-28 01:02:35 +01:00
Jan Lindström
e7e06b3cb7 MDEV-35481 : ER_LOCK_DEADLOCK instead of ER_NO_SUCH_TABLE in galera_var_ignore_apply_errors
Add wait_condition before select and set wsrep_sync_wait=0 to avoid
possible deadlock.

Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
2024-11-28 01:02:35 +01:00
Jan Lindström
ff45fdac29 MDEV-35440 : Protocol error warning in the galera_wsrep_schema_detached test
Make sure that node_1 remains in primary view by increasing it's
weight. Add suppression on expected warnings because we kill
node_2.

Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
2024-11-28 01:02:35 +01:00
Julius Goryavsky
8bc254dd62 MDEV-26516: WSREP: Record locking is disabled in this thread, but the table being modified
We periodically observe assertion failures in the mtr tests,
specifically in the /storage/innobase/row/row0ins.cc file,
following a WSREP error. The error message is: 'WSREP: record
locking is disabled in this thread, but the table being modified
is not mysql/wsrep_streaming_log: mysql/innodb_table_stats.'"
This issue seems to occur because, upon opening the table,
innodb_stats_auto_recalc may trigger, which Galera does not
anticipate. This commit should fix this bug.
2024-11-28 01:02:35 +01:00
Oleksandr Byelkin
0ea19c12fd MDEV-35507 ed25519 authentication plugin create user statement trigger plain text password in audit log
Mask also all cases of "password(PWD" in CREATE/ALTER USER and GRANT.
(minimal fix)
2024-11-27 19:45:33 +01:00
Ivan Prisyazhnyy
f39a61505f MDEV-33075 [backport/2f5174e556] use more robust self-pipe to wake up poll() in break_connect_loop()
Backport of 2f5174e556:
MDEV-33075 Resolve server shutdown issues on macOS, Solaris, and FreeBSD.

This commit addresses multiple server shutdown problems observed on macOS,
Solaris, and FreeBSD:

1. Corrected a non-portable assumption where socket shutdown was expected
to wake up poll() with listening sockets in the main thread.

Use more robust self-pipe to wake up poll() by writing to the pipe's write
end.

Signed-off-by: Ivan Prisyazhnyy <john.koepi@gmail.com>
2024-11-27 14:59:50 +02:00
Ivan Prisyazhnyy
c4cadb768f MDEV-33075 [backport/2f5174e556] fix rnd crash on macOS from pthread_kill(signal_handler)
Backport of 2f5174e556:
MDEV-33075 Resolve server shutdown issues on macOS, Solaris, and FreeBSD.

This commit addresses multiple server shutdown problems observed on macOS,
Solaris, and FreeBSD:

2. Fixed a random crash on macOS from pthread_kill(signal_handler)
when the signal_handler was detached and the thread had already exited.

Use more robust `kill(getpid(), SIGTERM)` to wake up the signal handler
thread.

Additionally, the shutdown code underwent light refactoring
for better readability and maintainability:

- Modified `break_connect_loop()` to no longer wait for the main thread,
  aligning behavior with Windows (since 10.4).
2024-11-27 14:59:50 +02:00
Ivan Prisyazhnyy
8214707699 MDEV-33075 [backport/2f5174e556] fix signal handler thread exit on abort
Backport of 2f5174e556:
MDEV-33075 Resolve server shutdown issues on macOS, Solaris, and FreeBSD.

This commit addresses multiple server shutdown problems observed on macOS,
Solaris, and FreeBSD:

3. Made sure, that signal handler thread always exits once `abort_loop` is
set, and also calls `my_thread_end()` and clears `signal_thread_in_use`
when exiting.

This fixes warning "1 thread did not exit"  by `my_global_thread_end()`
seen on FreeBSD/macOS when the process is terminated via signal.

Additionally, the shutdown code underwent light refactoring
for better readability and maintainability:

- Removed dead code related to the unused `USE_ONE_SIGNAL_HAND`
  preprocessor constant.

Signed-off-by: Ivan Prisyazhnyy <john.koepi@gmail.com>
2024-11-27 14:59:50 +02:00
Ivan Prisyazhnyy
490274e850 MDEV-33075 [backport/2f5174e556] eliminated support for #ifndef HAVE_POLL
Backport of 2f5174e556:
MDEV-33075 Resolve server shutdown issues on macOS, Solaris, and FreeBSD.

Eliminated support for `#ifndef HAVE_POLL` in `handle_connection_sockets`
This code is also dead, since 10.4

Signed-off-by: Ivan Prisyazhnyy <john.koepi@gmail.com>
2024-11-27 14:59:50 +02:00
Jan Lindström
f5aed74573 MDEV-35486 : MDEV-33997 test failed
Problem was that at wsrep_to_isolation_end saved_lock_wait_timeout
variable was set to thd->variables.lock_wait_timeout when RSU
is used and variable value was 0 leading sporadic lock wait timeout
errors. Fixed by removing incorrect variable set.

Signed-off-by: Julius Goryavsky <julius.goryavsky@mariadb.com>
2024-11-27 13:00:08 +01:00
Thirunarayanan Balathandayuthapani
9ba18d1aa0 MDEV-35394 Innochecksum misinterprets freed pages
- Innochecksum misinterprets the freed pages as active one.
This leads the user to think there are too many valid
pages exist.

- To avoid this confusion, innochecksum introduced one
more option --skip-freed-pages and -r to avoid the freed
pages while dumping or printing the summary of the tablespace.

- Innochecksum can safely assume the page is freed if
the respective extent doesn't belong to a segment and marked as
freed in XDES_BITMAP in extent descriptor page.

- Innochecksum shouldn't assume that zero-filled page as extent
descriptor page.

Reviewed-by: Marko Mäkelä
2024-11-27 13:00:51 +05:30
Yuchen Pei
5be859d52c
MDEV-30649 Adding a spider testcase showing copying from a remote to a local table
Also deleted some trailing whitespace in mdev_30191.test.
2024-11-27 10:25:14 +11:00
Yuchen Pei
a8cc40d9a4
MDEV-35064 Reduce the default spider connect retry counts to 2
The existing default value 1000 is too big and could result in
"hanging" when failing to connect a remote server. Three tries in
total is a more sensible default.
2024-11-27 10:25:14 +11:00
Monty
142851f120 Update my_print_defaults to accept --mariadbd as an option
--mariadbd and --mysqld are now synonymes for my_print_defaults

Other things
- Removed safemalloc warnings when using an unknown argument to
  my_print_defaults
2024-11-26 14:23:31 +02:00
Monty
f09020b3bb Fixed bug in subselect3.inc (not notable) 2024-11-26 14:23:31 +02:00
Alexander Barkov
225c17d35c MDEV-34090 Client allows to set character set to utf32 and crashes on the next command
Disallowing character sets with mbminlen>1 in the client.
2024-11-26 13:57:55 +04:00
Alexander Barkov
425d2521ec MDEV-33472 Assertion `0' failed in Item_row::illegal_method_call on CREATE EVENT
Do not accept rows as event parameters.
2024-11-25 18:29:13 +04:00
Alexander Barkov
20eba06d9b MDEV-35489 Assertion `!ldate->neg' or unexpected result upon extracting unit from invalid value
The patch for MDEV-23687 easlier fixed this problem. Adding tests only.
2024-11-25 16:13:16 +04:00
Marko Mäkelä
2255be0395 MDEV-35472 Server crash in ha_storage_put_memlim upon reading from INNODB_LOCKS
ha_storage_put_memlim(): Initialize node->next in order to avoid a
crash on a subsequent invocation, due to dereferencing an uninitialized
pointer.

This fixes a regression that had been introduced in
commit ccb6cd8053 (MDEV-35189).

Reviewed by: Debarun Banerjee
2024-11-25 10:31:57 +02:00
ParadoxV5
ec58fce3da MDEV-35478 Correction for table->space_id in dict_load_tablespace() was mistakenly applied on an earlier branch
It is `ulint` on 10.6 and `uint32_t` on 10.11+, but I included its
format specifier change in 10.6 (MDEV-35430, merged ) rather
than 10.11. This commit reverts that change so 10.11 can reapply it.
2024-11-25 18:25:02 +11:00
Alexander Barkov
2e404c9850 MDEV-21029 Incorrect result for expression with the <=> operator and IS NULL
Item_func_equal erroneously derived is_null() from the parent class.
Overriding it to return false because <=> can never return NULL.
2024-11-25 10:37:02 +04:00
Brandon Nesterenko
773cb726a8 MDEV-34348: MariaDB is violating clang-16 -Wcast-function-type-strict
This HEAD commit just makes cast-function-type-strict fatal.
It will stay at the HEAD, and prior commits contain the
actual fixes, organized by the part of the code that the fix
targets. Future changes will be force pushed to have their
fixes come before this HEAD.

This branch cherry-picks patches from daniel@mariadb.org
(in bb-10.5-MDEV-34508-ubsan-errors) which initially
incorporate the flag into the build, as well as fix some of
its findings.
2024-11-23 08:14:23 -07:00
Daniel Black
6456e437f2 MDEV-34348: Add cast-function-type-strict to warnings 2024-11-23 08:14:23 -07:00
Daniel Black
971a0ba23c MDEV-34408: Facilitate the addition of warnings into the build system
Create a MY_WARNING_FLAGS_NON_FATAL for testing warnings

Add -Weffc++ as no-error espect, disabling from rocksdb due
to excessive errors that will be corrected later.
2024-11-23 08:14:23 -07:00
Brandon Nesterenko
78d7bb1d27 MDEV-34348: Miscellaneous fixes
Partial commit of the greater MDEV-34348 scope.
MDEV-34348: MariaDB is violating clang-16 -Wcast-function-type-strict

Various additional fixes, each too small to put into
their own commit.

Reviewed By:
============
Marko Mäkelä <marko.makela@mariadb.com>
2024-11-23 08:14:23 -07:00
Brandon Nesterenko
3c785499da MDEV-34348: Fix casts relating to tree_walk_action
Partial commit of the greater MDEV-34348 scope.
MDEV-34348: MariaDB is violating clang-16 -Wcast-function-type-strict

Reviewed By:
============
Marko Mäkelä <marko.makela@mariadb.com>
2024-11-23 08:14:23 -07:00
Brandon Nesterenko
5432fa802b MDEV-34348: Fix casts in sql_acl
Partial commit of the greater MDEV-34348 scope.
MDEV-34348: MariaDB is violating clang-16 -Wcast-function-type-strict

Reviewed By:
============
Marko Mäkelä <marko.makela@mariadb.com>
2024-11-23 08:14:23 -07:00
Brandon Nesterenko
7a8eb26bda MDEV-34348: Fix casting related to plugins
Partial commit of the greater MDEV-34348 scope.
MDEV-34348: MariaDB is violating clang-16 -Wcast-function-type-strict

Reviewed By:
============
Marko Mäkelä <marko.makela@mariadb.com>
2024-11-23 08:14:23 -07:00
Brandon Nesterenko
840fe316d4 MDEV-34348: my_hash_get_key fixes
Partial commit of the greater MDEV-34348 scope.
MDEV-34348: MariaDB is violating clang-16 -Wcast-function-type-strict

Change the type of my_hash_get_key to:
 1) Return const
 2) Change the context parameter to be const void*

Also fix casting in hash adjacent areas.

Reviewed By:
============
Marko Mäkelä <marko.makela@mariadb.com>
2024-11-23 08:14:22 -07:00
Brandon Nesterenko
dbfee9fc2b MDEV-34348: Consolidate cmp function declarations
Partial commit of the greater MDEV-34348 scope.
MDEV-34348: MariaDB is violating clang-16 -Wcast-function-type-strict

The functions queue_compare, qsort2_cmp, and qsort_cmp2
all had similar interfaces, and were used interchangable
and unsafely cast to one another.

This patch consolidates the functions all into the
qsort_cmp2 interface.

Reviewed By:
============
Marko Mäkelä <marko.makela@mariadb.com>
2024-11-23 08:14:22 -07:00
Alexander Barkov
3997d28f48 MDEV-16698 ASAN: heap-use-after-free in field_longstr::uncompress
This bug was earlier fixed by MDEV-16699. Adding tests only.
2024-11-22 15:46:36 +04:00
Alexander Barkov
95df7ea33a MDEV-31881 ASAN: unknown-crash in check_ulonglong (sql/sql_analyse.cc) on SELECT ... FROM ... PROCEDURE ANALYSE()
Fixing a wrong condition which made the code read 1 byte behind the buffer.
2024-11-22 14:54:08 +04:00
Alexander Barkov
39f1f30f68 MDEV-23687 Assertion `is_valid_value_slow()' failed in Datetime::Datetime upon EXTRACT under mode ZERO_DATE_TIME_CAST
Item_{date|datetime}_typecase::get_date() erroneously passed the
TIME_INTERVAL_DAY flag from the caller to args[0] which made
CAST('100000:00:00' AS DATETIME) parse '100000:00:00' as TIME
rather that DATETIME.
Suppressing this flag.
2024-11-22 12:15:03 +04:00
Marko Mäkelä
26597b91b3 MDEV-35413 InnoDB: Cannot load compressed BLOB
A race condition was observed between two buf_page_get_zip() for a page.
One of them had proceeded to buf_read_page(), allocating and x-latching
a buf_block_t that initially comprises only an uncompressed page frame.
While that thread was waiting inside buf_block_alloc(), another thread
would try to access the same page. Without acquiring a page latch, it
would wrongly conclude that there is corruption because no compressed
page frame exists for the block.

buf_page_get_zip(): Simplify the logic and correct the documentation.
Always acquire a shared latch to prevent any race condition with a
concurrent read operation. No longer increment a buffer-fix; the latch
is sufficient for preventing page relocation or eviction.

buf_read_page(): Add the parameter bool unzip=true. In buf_page_get_zip()
there is no need to allocate an uncompressed page frame for reading a
compressed BLOB page. We only need that for other ROW_FORMAT=COMPRESSED
pages, or for writing compressed BLOB pages.

btr_copy_zblob_prefix(): Remove the message "Cannot load compressed BLOB"
because buf_page_get_zip() will already have reported a more specific
error whenever it returns nullptr.

row_merge_buf_add(): Do not crash on BLOB corruption, but return an
error instead. (In debug builds, an assertion will fail if this
corruption is noticed.)

Reviewed by: Debarun Banerjee
2024-11-22 08:33:03 +02:00
Brandon Nesterenko
a06d81ff3f MDEV-35477: rpl_semi_sync_no_missed_ack_after_add_slave fails after MDEV-35109
MTR test rpl_semi_sync_no_missed_ack_after_add_slave fails on
buildbot after the preparatory commit for MDEV-35109 (5290fa043b)
which changed a sleep to a debug_sync point. The problem is that the
debug_sync point would time-out on a slave while waiting to enter
the logic to send an ACK reply. More specifically, where the test
config is a primary with two replicas, and the test waits on one of
the replicas to start sending an ACK, if the other replica was able
to receive the event and respond with an ACK before the binlog dump
thread of the timing-out server would prepare to send event, it
wouldn't set the SEMI_SYNC_NEED_ACK flag, and the replica wouldn't
even try to respond with an ACK.

Fix is to use debug_sync for both replicas such that both replicas
are held before sending their ack, so one can’t temporarily disable
semi-sync for the other before it receives the transaction.
2024-11-21 11:30:25 -07:00
ParadoxV5
cf2d49ddcf Extract some of fixes to 10.5.x
That PR uncovered countless issues on `my_snprintf` uses.
This commit backports a squashed subset of their fixes.
2024-11-21 22:43:56 +11:00
Daniel Black
b414eca98d Correct cursor protocol tests main.{loaddata,grant_plugin} & innodb_fts.fulltext
Cursor protocol cannot handle select... into.

Disable this on loaddata.

For the grant_plugin/innodb_fts.fulltext changed
the tests to use a temporary table rather than a
user variable.
2024-11-21 21:43:44 +11:00
Daniel Black
2c89fe7ea6 main.stack view protocol - correct test result
View protocol requires expressions include a simple alias.
2024-11-21 21:43:44 +11:00
Marko Mäkelä
895cd553a3 MDEV-32175: Reduce page_align(), page_offset() calls
When srv_page_size and innodb_page_size were introduced,
the functions page_align() and page_offset() got more expensive.
Let us try to replace such calls with simpler pointer arithmetics
with respect to the buffer page frame.

page_rec_get_next_non_del_marked(): Add a page frame as a parameter,
and template<bool comp>.

page_rec_next_get(): A more efficient variant of page_rec_get_next(),
with template<bool comp> and const page_t* parameters.

lock_get_heap_no(): Replaces page_rec_get_heap_no() outside debug checks.

fseg_free_step(), fseg_free_step_not_header(): Take the header block
as a parameter.

Reviewed by: Vladislav Lesin
2024-11-21 11:01:30 +02:00
Marko Mäkelä
df3855a471 MDEV-35247: ut_hash_ulint() is a waste
ut_hash_ulint(): Remove. The exclusive OR before a modulus operation
does not serve any useful purpose; it is only obfuscating code and
wasting some CPU cycles.

Reviewed by: Debarun Banerjee
2024-11-21 08:59:31 +02:00
Marko Mäkelä
a9b0a1c5d0 MDEV-35247: ut_fold_ull() is a waste
ut_fold_ull(): For SIZEOF_SIZE_T < 8, we simulate universal hashing
(Carter and Wegman, 1977) by pretending that SIZE_T_MAX + 1
is a prime. In other words, we implement a Rabin–Karp rolling
hash algorithm similar to java.lang.String.hashCode().
This is used for representing 64-bit dict_index_t::id or
dict_table_t::id in the native word size.

For SIZEOF_SIZE_T >= 8, we just use an identity mapping.

Reviewed by: Debarun Banerjee
2024-11-21 08:59:17 +02:00
Marko Mäkelä
3c312d247c MDEV-35190 HASH_SEARCH duplicates effort before HASH_INSERT or HASH_DELETE
The HASH_ macros are unnecessarily obfuscating the logic,
so we had better replace them.

hash_cell_t::search(): Implement most of the HASH_DELETE logic,
for a subsequent insert or remove().

hash_cell_t::remove(): Remove an element.

hash_cell_t::find(): Implement the HASH_SEARCH logic.

xb_filter_hash_free(): Avoid any hash table lookup;
just traverse the hash bucket chains and free each element.

xb_register_filter_entry(): Search databases_hash only once.

rm_if_not_found(): Make use of find_filter_in_hashtable().

dict_sys_t::acquire_temporary_table(), dict_sys_t::find_table():
Define non-inline to avoid unnecessary code duplication.

dict_sys_t::add(dict_table_t *table), dict_table_rename_in_cache():
Look for duplicate while finding the insert position.

dict_table_change_id_in_cache(): Merged to the only caller
row_discard_tablespace().

hash_insert(): Helper function of dict_sys_t::resize().

fil_space_t::create(): Look for a duplicate (and crash if found)
when searching for the insert position.

lock_rec_discard(): Take the hash array cell as a parameter
to avoid a duplicated lookup.

lock_rec_free_all_from_discard_page(): Remove a parameter.

Reviewed by: Debarun Banerjee
2024-11-21 08:59:02 +02:00
Vlad Lesin
bcbeef6772 MDEV-35457 Remove btr_cur_t::path_arr
After MDEV-21136 fix, the btr_cur_t::path_arr field stayed declared, but
not used, wasting space in each btr_cur_t and btr_pcur_t. Remove it.
2024-11-20 17:43:04 +03:00
Monty
32962ea253 Do not read aria bitmap page for internal temporary tables
Instead create the bitmap page from scratch
2024-11-20 10:01:20 +02:00
Monty
93fb364cd9 Removed not used ha_drop_table()
This was done after changing call in sql_select.cc from
ha_drop_table() to drop_table(), like in 11.5
2024-11-20 09:59:43 +02:00
Monty
69be363daa Fixed that internal temporary Aria tables are not flushed to disk
This bug was caused by
MDEV-17070 Table corruption or Assertion `table->file->stats.records > 0
2024-11-19 20:27:51 +02:00
Monty
0de9e40f4b Added status variable "stack_usable" to be able to check stack usage 2024-11-19 19:02:45 +02:00
Alexander Barkov
ae0cbfe934 MDEV-28001 greatest/least with bigint unsigned maxium has unexpected results compared to 0
LEAST() and GREATEST() erroneously calcucalted the result as signed
for BIGINT UNSIGNED arguments.

Adding a new method for unsigned arguments:
  Item_func_min_max::val_uint_native()
2024-11-19 14:26:39 +04:00