MDEV-23356 InnoDB: Failing assertion: field->col->mtype == type, crash or ASAN failures in row_sel_convert_mysql_key_to_innobase, InnoDB indexes are inconsistent after INDEX changes

innobase_rename_indexes_cache(): fix corruption of index cache. Index ids
help distinguish indexes when their names clash.

innobase_rename_indexes_cache(): fix corruption of index statistics table.
Use unique temporary names to avoid names clashing.

Reviewed by: Marko Mäkelä

Recent gcc/clang versions failed to compile the existing code.

Updating a later upstream SDK version was simple and required
only implementing a flush method. This was left blank as
there was no strong requirement to keep the error log
atomic or durable.

Reviewed-by: wlad@mariadb.com

The upstream SDK version added a flush method which was simple
to complete.

protect global toku_global_status_rows with a mutex

v_indexes was wrongly copied from the source table.

We must make the test depend on a debug version of Galera.

In commit 7eda5561 we removed a
valid debug assertion.

AddressSanitizer would trip when writing undo log for an INSERT
operation that follows the problematic ALTER TABLE because the
v_indexes would refer to an index object that was freed during
the execution of ALTER TABLE.

The operation to remove NOT NULL attribute from a column should
lead into any affected indexes to be dropped and re-created.
But, the SQL layer set the ha_alter_info->handler_flags to
HA_INPLACE_ADD_UNIQUE_INDEX_NO_WRITE|ALTER_COLUMN_NULLABLE
(that is, InnoDB was asked to add an index and change the column,
but not to drop the index that depended on the column).

Let us restore the debug assertion that catches this problem
outside AddressSanitizer, and 'defuse' the offending ALTER TABLE
statement in the test until something has been fixed in the SQL layer.

MDEV-23935: Fix warnings generated during compilation of plugin/auth_pam/testing/pam_mariadb_mtr.c on MacOS

During build server 10.4 on MacOS warnings like the following one
are generated on compiling the file plugin/auth_pam/testing/pam_mariadb_mtr.c

server-10.4/plugin/auth_pam/testing/pam_mariadb_mtr.c:25:22: error:
initializing 'char *' with an expression of type 'const char [23]'
discards qualifiers [-Werror,-Wincompatible-pointer-types-discards-qualifiers]
{ PAM_TEXT_INFO, "Challenge input first." }

The reason of the warnings is that the data member pam_message::msg is declared
as 'char *' on MacOS but initializer of the data member is 'const char *'.

To eliminate warnings the compiler option
  -Wno-incompatible-pointer-types-discards-qualifiers
has been added to the compiler flags used for compiling the file
pam_mariadb_mtr.c.

MDEV-20945: BACKUP UNLOCK + FTWRL assertion failure | SIGSEGV in I_P_List
from MDL_context::release_lock on INSERT w/ BACKUP LOCK (on optimized
builds) | Assertion `ticket->m_duration == MDL_EXPLICIT' failed

BACKUP LOCK behavior is modified so it won't be used wrong:
- BACKUP LOCK should commit any active transactions.
- BACKUP LOCK should not be allowed in stored procedures.
- When BACKUP LOCK is active, don't allow any DDL's for that connection.
- FTWRL is forbidden on the same connection while BACKUP LOCK is active.

Reviewed-by: monty@mariadb.com

Some functions on ha_partition call functions on all partitions, but handler->reset() is only called that pruned by m_partitions_to_reset. So Spider didn't clear pointer on unpruned partitions, if the unpruned partitions are used by next query, Spider reference the pointer that is already freed.

Change xarecover_handlerton so that transaction with WSREP prefixed
xids are rolled back when Galera is disabled.
Reviewd-by: Jan Lindström <jan.lindstrom@mariadb.com>

Compiler warnings like one listed below are generated during server build on MacOS:
In file included from server-10.2-MDEV-23564/mysys_ssl/openssl.c:33:
In file included from /usr/local/include/openssl/evp.h:16:
In file included from /usr/local/include/openssl/bio.h:20:
/usr/local/include/openssl/crypto.h:206:10: warning: 'CRYPTO_cleanup_all_ex_data' macro redefined [-Wmacro-redefined]
           ^
  /mariadb/server-10.2-MDEV-23564/include/ssl_compat.h:46:9: note: previous definition is here
          ^

In case MariaDB serer is build with -DCMAKE_BUILD_TYPE=Debug it results in
build error.

The reason of compiler warnings is that header file <ssl_compat.h>
included before the openssl system header files. File ssl_compat.h
contains some macros with the same names as SSL API functions declared
in the openssl system header files. It resulted in duplicate
symbols that produces compiler warnings.

To fix the issue the header file ssl_compat.h should be included
after a line where openssl system header is included.

It was only from CMake-3.14.0 that CMAKE_REQUIRED_LINK_OPTIONS
was used in CHECK_CXX_SOURCE_COMPILES. Without this, it could be
the case (as was on OSX) that a flag was never checked in
CHECK_CXX_SOURCE_COMPILES, the CHECK successfully passed, but
failed at link time.

As such we use CMAKE_REQUIRED_LIBRARIES to include the flags to check
as its compatible enough with the cmake versions for non-Windows
compilers/linkers.

Tested on x86_64 with:
* 3.11.4
* 3.17.4

Corrects: 7473e184

In the future:
* cmake >=3.14.0 can use CMAKE_REQUIRED_LINK_OPTIONS
* cmake >=3.18.0 can use CHECK_LINKER_FLAG (with policy CMP0057 NEW)
(e.g: commit c7ac2deff9a2c965887dcc67cbf2a3a7c3e0123d)

CMAKE_REQUIRED_LIBRARIES suggested by serg@mariadb.com

Reviewed-by: anel@mariadb.org

The only applicable InnoDB change to MariaDB that was made
between MySQL 5.6.49 and MySQL 5.6.50 is MDEV-23999.

fts_query_t::nested_sub_exp: Keep track of nested
fts_ast_visit_sub_exp() calls.

fts_ast_visit_sub_exp(): Return DB_OUT_OF_MEMORY if the
maximum recursion depth is exceeded.

This is motivated by a change in MySQL 5.6.50:
mysql/mysql-server@e2a46b4834a0030e5380d45789772cb6beab6e92
Bug #29929684 USING MANY NESTED ARGUMENTS WITH BOOLEAN FTS CAN LEAD
TO TERMINATE SERVER

In row_undo_ins_remove_clust_rec() and similar places,
an assertion !node->trx->dict_operation_lock_mode could fail,
because an online ALTER is not allowed to run at the same time
while DDL operations on the table are being rolled back.

This race condition would be fixed by always acquiring an InnoDB
table lock in ha_innobase::prepare_inplace_alter_table() or
prepare_inplace_alter_table_dict(), or by ensuring that recovered
transactions are protected by MDL that would block concurrent DDL
until the rollback has been completed.

This reverts commit 15093639
and commit 22c4a751.

The function innodb_show_mutex_status() is the only ultimate caller of
LatchCounter::iterate() via MutexMonitor::iterate(). Because the call
is not protected by LatchCounter::m_mutex, any mutex_create() or
mutex_free() that is invoked concurrently during the execution, bad
things such as a crash could happen.

The most likely way for this to happen is buffer pool resizing,
which could cause buf_block_t::mutex (which existed before MDEV-15053)
to be created or freed. We could also register InnoDB mutexes in
TrxFactory::init() if trx_pools needs to grow.

The view INFORMATION_SCHEMA.INNODB_MUTEXES is not affected, because it
only displays information about rw-locks, not mutexes.

This commit intentionally touches also MutexMonitor::iterate()
and the only code that interfaces with LatchCounter::iterate()
to make it clearer for future readers that the scattered code
that is obfuscated by templates belongs together.

This is based on
mysql/mysql-server@273a93396f49c7e0a8b07b260128d9a990c2b154

- Based on patch: d6a983351c5a454bd0cb113852f
- Update combination example for 10.2 (commit 2a3fe45d added change
for 10.3+)
```
==============================================================================

TEST                                      RESULT   TIME (ms) or COMMENT
--------------------------------------------------------------------------

worker[1] Using MTR_BUILD_THREAD 300, with reserved ports 16000..16019
rpl.rpl_invoked_features 'innodb,mix'    [ pass ]   1677
rpl.rpl_invoked_features 'innodb,row'    [ pass ]   3516
rpl.rpl_invoked_features 'innodb,stmt'   [ pass ]   1609
--------------------------------------------------------------------------
```
- `gdb` option will be added during the merge

- Updated combination example to use `innodb_plugin`
```
$ ./mysql-test/mtr rpl.rpl_invoked_features # no rpl.rpl_invoked_features,mix,xtradb_plugin

worker[1] Using MTR_BUILD_THREAD 300, with reserved ports 16000..16019
rpl.rpl_invoked_features 'innodb_plugin,mix' [ pass ]    904
rpl.rpl_invoked_features 'mix,xtradb'    [ pass ]   1707
rpl.rpl_invoked_features 'innodb_plugin,row' [ pass ]    927
rpl.rpl_invoked_features 'row,xtradb'    [ pass ]    828
rpl.rpl_invoked_features 'innodb_plugin,stmt' [ pass ]    855
rpl.rpl_invoked_features 'stmt,xtradb'   [ pass ]    952
--------------------------------------------------------------------------
```
- `gdb` option can take optional gdb arguments (good documentation in
KB)

Some functions on ha_partition call functions on all partitions, but handler->reset() is only called that pruned by m_partitions_to_reset. So Spider didn't clear pointer on unpruned partitions, if the unpruned partitions are used by next query, Spider reference the pointer that is already freed.

To fix this, it is necessary to add an option to exclude the
database with the name "lost+found" from processing (the database
name will be checked by the check_if_skip_database_by_path() or
by the check_if_skip_database() function, and as a result
"lost+found" will be skipped).

In addition, it is necessary to slightly modify the verification
logic in the check_if_skip_database() function.

Also added a new test galera_sst_mariabackup_lost_found.test

fix an error with locked taböes

dict_v_idx_t node was shared between two dict_v_col_t objects because
of wrong object copy. Replace memory plain copy with copy constructor.

Tha patch also removes n_v_indexes property and improves "page full"
judgements for trx_undo_log_v_idx().

MDEV-23968 CREATE TEMPORARY TABLE .. LIKE (system versioned table) returns error if unique index is defined in the table

- Remove row_start/row_end from keys in fix_create_like();

- Disable manual adding of implicit row_start/row_end to indexes on
  CREATE TABLE. INVISIBLE_SYSTEM fields are unoperable by user;

- Fix memory leak on allocation of Key_part_spec.

PARTITION clause in SELECT means query is non-versioned (see
WITH_PARTITION_STORAGE_ENGINE in vers_setup_conds()).

vers_setup_conds() expands such query to SYSTEM_TIME_ALL which is then
added to VIEW specification. When VIEW is queried both clauses
PARTITION and FOR SYSTEM_TIME ALL lead to ER_VERS_QUERY_IN_PARTITION
(same place WITH_PARTITION_STORAGE_ENGINE).

Fix removes FOR SYSTEM_TIME ALL from VIEW by accessing original
SYSTEM_TIME clause: the one specified in parser. As a side-effect
EXPLAIN SELECT displays SYSTEM_TIME specified in SELECT which is
user-friendly.

For join to work correctly versioning condition must be added to table
on_expr. Without that JOIN_CACHE gets expression (1)

  trigcond(xtitle.row_end = TIMESTAMP'2038-01-19 06:14:07.999999') and
  trigcond(xtitle.elementId = x.`id` and xtitle.pkey = 'title')

instead of (2)

  trigcond(xtitle.elementId = x.`id` and xtitle.pkey = 'title')

for join_null_complements(). It is NULL-row of xtitle for
complementing the join and the above comparisons of course FALSE, but
trigcond (Item_func_trig_cond) makes them TRUE via its trig_var
property which is bound to some boolean properties of JOIN_TAB.

Expression (2) evaluated to TRUE because its trig_var is bound to
first_inner_tab->not_null_compl. The expression (1) does not evaluate
correctly because row_end comparison's trig_var is bound to
first_inner->found earlier. As a result JOIN_CACHE::check_match()
skipped the row for join_null_complements().

When we add versioning condition to table's on_expr the optimizer in
make_join_select() distributes conditions differently. tmp_cond
inherits on_expr value and in Good case it is full expression

xgender.elementId = x.`id` and xgender.pkey = 'gender' and
xgender.row_end = TIMESTAMP'2038-01-19 06:14:07.999999'

while in Bad case it is only

xgender.elementId = x.`id` and xgender.pkey = 'gender'.

Later in Good row_end condition is optimized out and we get one
trigcond in form of (2).

- This issue is caused by commit a4948daf.
Purge doesn't free the externally stored page associated with the
last record of the root page. In that case, purge thread does empty
the root page and leads to more orphaned blob page in the tablespace.
Purge thread should free the blob even for the last record of the
root page.

Reviewed-by: Marko Mäkelä

Analysis: When we reach the maximum limit to examine rows killed_state is set
as ABORT. But this isn't an actual error and we still return TRUE. This
eventually sets error as UNKNOWN ERROR.
Fix: Check if need to stop execution by checking the killed state. If we have
to abort it, return false because this isn't an actual error.

is_bulk_op())' fails on UPDATE on a partitioned table with subquery
(MySQL:71630)

Analysis and fix: Error is not checked. So correct error state is not returned.
Fix: Check for error and return the error state.

Diagnostics_area::set_error_status

Analysis: When strict mode is enabled, all warnings are converted to errors
including those which do not occur because of bad data.
Fix: Query should not be aborted when we have warning because limit to
examine rows was reached because it doesn't happen due to bad data.
So thd->abort_on_warning should be false.

MDEV-13318 introduced a condition to Mariabackup that can cause it to
hang if the server goes idle after writing a log block that has no
payload after the 12-byte header. Normal recovery in log0recv.cc would
allow blocks with exactly 12 bytes of length, and only reject blocks
where the length field is shorter than that.

This issue is caused by the commit af40a2b4.
In btr_search_update_hash_on_insert(), btr_search_sys->hash_tables
is being accessed without taking proper ahi latch. During buffer pool
resizing, btr_get_search_table() is being accessed and it leads to
segmentation fault.

Reviewed-by: Marko Mäkelä

Problem:
=======
 InnoDB allows virtual index to be referenced index in foreign key
relations. While dropping the virtual column, Inplace alter does
allow the table to be closed and open it using table name to
update dict_table_t::v_cols. While loading the table, it doesn't
allow any error to be ignored. InnoDB can't find the referenced
virtual index and fails to load the table during Inplace alter.

Solution:
=========
  During inplace alter, ignore the foreign key error while loading
the table.

Reviewed-by: Marko Mäkelä

Made cleanup of DROP (udf) FUNCTION procedure and also check of mysql.func (not only loaded udf).

MDEV-16676 Using malloc-lib=jemalloc in MariaDB 10.2 causes non-critical error about missing mysql_config on startup

- This commit is based on patch 84fe9720a4d2483ff67b6a and suggestion of
Jean Weisbuch to use `ldconfig -p`
- Format of ldconfig output:
"libjemalloc.so.1 (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/libjemalloc.so.1"

```
$ ./scripts/mysqld_safe
201013 13:36:50 mysqld_safe Adding '/usr/lib/x86_64-linux-gnu/libjemalloc.so.1' to LD_PRELOAD for mysqld
201013 13:36:50 mysqld_safe Logging to '/home/anel/builds/data-10.5/mysqld_safe_anel.err'.
201013 13:36:50 mysqld_safe Starting mysqld daemon with databases from /home/anel/builds/data-10.5
```
Reviewed by: Daniel Black, Faustin Lammler

Instead of pointlessly waiting for a page flush to occur, take
the matter into our own hands and request an explicit flush.
Also, test with the minimum necessary amount of data (0 or 1 rows)
so that both page encryption and decryption will be exercised.

This also fixes some issues with
MDEV-23730 s3.replication_partition 'innodb,mix' segv

The problem was that mysql_change_partitions() closes all handler files
in case of error, which was not properly reflected in
fast_alter_partition_table(). This caused handle_alter_part_error() to
try to close already closed tables, which caused the crash.

Fixed fast_alter_partion_table() to reflect when tables are opened.
I also fixed that ha_partition::change_partitions() resets m_new_file in
case of errors.
Either of the above changes fixes the issue, but both are needed to ensure
that the code works as expected.