corrupts a MERGE table
Bug 26867 - LOCK TABLES + REPAIR + merge table result in
            memory/cpu hogging
Bug 26377 - Deadlock with MERGE and FLUSH TABLE
Bug 25038 - Waiting TRUNCATE
Bug 25700 - merge base tables get corrupted by
            optimize/analyze/repair table
Bug 30275 - Merge tables: flush tables or unlock tables
            causes server to crash
Bug 19627 - temporary merge table locking
Bug 27660 - Falcon: merge table possible
Bug 30273 - merge tables: Can't lock file (errno: 155)

The problems were:

Bug 26379 - Combination of FLUSH TABLE and REPAIR TABLE
                corrupts a MERGE table

  1. A thread trying to lock a MERGE table performs busy waiting while
     REPAIR TABLE or a similar table administration task is ongoing on
     one or more of its MyISAM tables.
  
  2. A thread trying to lock a MERGE table performs busy waiting until all
     threads that did REPAIR TABLE or similar table administration tasks
     on one or more of its MyISAM tables in LOCK TABLES segments do UNLOCK
     TABLES. The difference against problem #1 is that the busy waiting
     takes place *after* the administration task. It is terminated by
     UNLOCK TABLES only.
  
  3. Two FLUSH TABLES within a LOCK TABLES segment can invalidate the
     lock. This does *not* require a MERGE table. The first FLUSH TABLES
     can be replaced by any statement that requires other threads to
     reopen the table. In 5.0 and 5.1 a single FLUSH TABLES can provoke
     the problem.

Bug 26867 - LOCK TABLES + REPAIR + merge table result in
            memory/cpu hogging

  Trying DML on a MERGE table, which has a child locked and
  repaired by another thread, made an infinite loop in the server.

Bug 26377 - Deadlock with MERGE and FLUSH TABLE

  Locking a MERGE table and its children in parent-child order
  and flushing the child deadlocked the server.

Bug 25038 - Waiting TRUNCATE

  Truncating a MERGE child, while the MERGE table was in use,
  let the truncate fail instead of waiting for the table to
  become free.

Bug 25700 - merge base tables get corrupted by
            optimize/analyze/repair table

  Repairing a child of an open MERGE table corrupted the child.
  It was necessary to FLUSH the child first.

Bug 30275 - Merge tables: flush tables or unlock tables
            causes server to crash

  Flushing and optimizing locked MERGE children crashed the server.

Bug 19627 - temporary merge table locking

  Use of a temporary MERGE table with non-temporary children
  could corrupt the children.

  Temporary tables are never locked. So we do now prohibit
  non-temporary chidlren of a temporary MERGE table.

Bug 27660 - Falcon: merge table possible

  It was possible to create a MERGE table with non-MyISAM children.

Bug 30273 - merge tables: Can't lock file (errno: 155)

  This was a Windows-only bug. Table administration statements
  sometimes failed with "Can't lock file (errno: 155)".

These bugs are fixed by a new implementation of MERGE table open.

When opening a MERGE table in open_tables() we do now add the
child tables to the list of tables to be opened by open_tables()
(the "query_list"). The children are not opened in the handler at
this stage.

After opening the parent, open_tables() opens each child from the
now extended query_list. When the last child is opened, we remove
the children from the query_list again and attach the children to
the parent. This behaves similar to the old open. However it does
not open the MyISAM tables directly, but grabs them from the already
open children.

When closing a MERGE table in close_thread_table() we detach the
children only. Closing of the children is done implicitly because
they are in thd->open_tables.

For more detail see the comment at the top of ha_myisammrg.cc.

Changed from open_ltable() to open_and_lock_tables() in all places
that can be relevant for MERGE tables. The latter can handle tables
added to the list on the fly. When open_ltable() was used in a loop
over a list of tables, the list must be temporarily terminated
after every table for open_and_lock_tables().
table_list->required_type is set to FRMTYPE_TABLE to avoid open of
special tables. Handling of derived tables is suppressed.
These details are handled by the new function
open_n_lock_single_table(), which has nearly the same signature as
open_ltable() and can replace it in most cases.

In reopen_tables() some of the tables open by a thread can be
closed and reopened. When a MERGE child is affected, the parent
must be closed and reopened too. Closing of the parent is forced
before the first child is closed. Reopen happens in the order of
thd->open_tables. MERGE parents do not attach their children
automatically at open. This is done after all tables are reopened.
So all children are open when attaching them.

Special lock handling like mysql_lock_abort() or mysql_lock_remove()
needs to be suppressed for MERGE children or forwarded to the parent.
This depends on the situation. In loops over all open tables one
suppresses child lock handling. When a single table is touched,
forwarding is done.

Behavioral changes:
===================

This patch changes the behavior of temporary MERGE tables.
Temporary MERGE must have temporary children.
The old behavior was wrong. A temporary table is not locked. Hence
even non-temporary children were not locked. See
Bug 19627 - temporary merge table locking.

You cannot change the union list of a non-temporary MERGE table
when LOCK TABLES is in effect. The following does *not* work:
CREATE TABLE m1 ... ENGINE=MRG_MYISAM ...;
LOCK TABLES t1 WRITE, t2 WRITE, m1 WRITE;
ALTER TABLE m1 ... UNION=(t1,t2) ...;
However, you can do this with a temporary MERGE table.

You cannot create a MERGE table with CREATE ... SELECT, neither
as a temporary MERGE table, nor as a non-temporary MERGE table.
CREATE TABLE m1 ... ENGINE=MRG_MYISAM ... SELECT ...;
Gives error message: table is not BASE TABLE.


include/my_base.h:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added HA_EXTRA_ATTACH_CHILDREN and HA_EXTRA_DETACH_CHILDREN.
include/myisammrg.h:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added element 'children_attached' to MYRG_INFO.
  Added declarations for myrg_parent_open(),
  myrg_attach_children() and myrg_detach_children()
  for the new MERGE table open approach.
mysql-test/extra/binlog_tests/blackhole.test:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Preliminarily added new error message with a comment.
mysql-test/r/create.result:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Fixed test result.
mysql-test/r/delayed.result:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Moved test result from here to merge.result.
mysql-test/r/merge.result:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Fixed/added test result.
mysql-test/r/myisam.result:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Moved test result for bug 8306 from here to merge.result.
mysql-test/suite/binlog/r/binlog_stm_blackhole.result:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Fixed test result.
mysql-test/t/create.test:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Fixed error number.
mysql-test/t/delayed.test:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Moved test from here to merge.test.
mysql-test/t/merge.test:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Fixed test for new temporary MERGE table behavior.
  Exchanged error numbers by symbolic codes.
  Added tests. Included are tests for bugs
  8306 (moved from myisam.test), 26379, 19627, 25038, 25700, 26377,
  26867, 27660, 30275, and 30273.
  Fixed changes resulting from disabled CREATE...SELECT.
  Integrated tests moved from delayed.test and myisam.test to here.
mysql-test/t/myisam.test:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Moved test for bug 8306 from here to merge.test.
mysys/thr_lock.c:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added code to let the owner of a high priority lock (TL_WRITE_ONLY)
  to bypass its own lock.
sql/ha_ndbcluster_binlog.cc:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added 'thd' argument to init_tmp_table_share().
sql/handler.cc:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added 'thd' argument to init_tmp_table_share().
sql/mysql_priv.h:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Removed declaration of check_merge_table_access(). It is now static
  in sql_parse.cc.
  Added declaration for fix_merge_after_open().
  Renamed open_and_lock_tables() to open_and_lock_tables_derived()
  with additional parameter 'derived'.
  Added inline functions simple_open_n_lock_tables() and
  open_and_lock_tables(), which call open_and_lock_tables_derived()
  and add the argument for 'derived'.
  Added new function open_n_lock_single_table(), which can be used
  as an replacement for open_ltable() in most situations. Internally
  it calls simple_open_n_lock_tables() so hat it is appropriate for
  MERGE tables.
  Added 'thd' argument to init_tmp_table_share().
sql/slave.cc:
  ug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added comment.
sql/sql_base.cc:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  
  Defined new functions add_merge_table_list(),
  attach_merge_children(), detach_merge_children(), and
  fix_merge_after_open() for the new MERGE table open approach.
  
  Added calls of the new functions to
  close_handle_and_leave_table_as_lock(), close_thread_tables(),
  close_thread_table(), unlink_open_table(), reopen_name_locked_table(),
  reopen_table(), drop_locked_tables(), close_temporary_table(),
  and open_tables() respectively.
  
  Prevented special lock handling of merge children (like
  mysql_lock_remove, mysql_lock_merge or mysql_lock_abort)
  at many places. Some of these calls are forwarded to the
  parent table instead.
  
  Added code to set thd->some_tables_deleted for every thread that has
  a table open that we are flushing.
  Added code for MERGE tables to unlink_open_table().
  Added MERGE children to the list of unusable tables in open_table().
  Added MERGE table handling to reopen_table().
  Added lock handling and closing of a parent before the children
  in close_data_files_and_morph_locks().
  Added code for re-attaching children in reopen_tables().
  Added MYSQL_LOCK_NOTIFY_IF_NEED_REOPEN to the locking flags and
  error reporting after mysql_lock_tables() in reopen_tables().
  Added lock handling and closing of a parent before the children
  in close_old_data_files().
  Added lock handling and detaching in drop_locked_tables().
  Added code for removing the children list from the statement list
  to prepare for a repetition in open_tables().
  Added new function open_n_lock_single_table(), which can be used
  as an replacement for open_ltable() in most situations. Internally
  it calls simple_open_n_lock_tables() so hat it is appropriate for
  MERGE tables.
  Disabled use of open_ltable() for MERGE tables.
  Removed function simple_open_n_lock_tables(). It is now inline
  declared in mysql_priv.h.
  Renamed open_and_lock_tables() to open_and_lock_tables_derived()
  with additional parameter 'derived'. open_and_lock_tables() is now
  inline declared in mysql_priv.h.
  Added a check for end-of-list in two loops in lock_tables().
  Added 'thd' argument to init_tmp_table_share().
sql/sql_insert.cc:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Changed from open_ltable() to open_n_lock_single_table() in
  handle_delayed_insert().
  Reestablished LEX settings after lex initialization.
  Added 'thd' argument to init_tmp_table_share().
sql/sql_parse.cc:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Made check_merge_table_access() a static function.
  Disabled use of CREATE...SELECT for MERGE tables.
sql/sql_partition.cc:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Fixed comment typo.
sql/sql_select.cc:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added 'thd' argument to init_tmp_table_share().
sql/sql_table.cc:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Optimized use of mysql_ha_flush() in mysql_rm_table_part2().
  Disabled the use of MERGE tables with prepare_for_restore() and
  prepare_for_repair().
  Changed from open_ltable() to open_n_lock_single_table() in
  mysql_alter_table() and mysql_checksum_table().
  Disabled change of child list under LOCK TABLES.
  Initialized table_list->table in mysql_recreate_table().
sql/sql_trigger.cc:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added code for allowing CREATE TRIGGER under LOCK TABLE, to be able
  to test it with MERGE tables.
sql/table.cc:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added 'thd' argument to init_tmp_table_share().
  Setting table_map_id from query_id in init_tmp_table_share().
  Added member function TABLE::is_children_attached().
sql/table.h:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added access method get_table_def_version() to TABLE_SHARE.
  Added elements for MERGE tables to TABLE and TABLE_LIST.
storage/myisam/ha_myisam.cc:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added an unrelated comment to the function comment of table2myisam().
storage/myisam/ha_myisam.h:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added new member function MI_INFO::file_ptr().
storage/myisammrg/ha_myisammrg.cc:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added callback functions to support parent open and children attach
  of MERGE tables.
  Changed ha_myisammrg::open() to initialize storage engine structures
  and create a list of child tables only. Child tables are not opened.
  Added ha_myisammrg::attach_children(), which does now the main part
  of MERGE open.
  Added ha_myisammrg::detach_children().
  Added calls to ::attach_children() and ::detach_children() to
  ::extra() on HA_EXTRA_ATTACH_CHILDREN and HA_EXTRA_DETACH_CHILDREN
  respectively.
  Added a check for matching TEMPORARY type for children against
  parent.
  Added a check for table def version.
  Added support for thd->open_options to attach_children().
  Changed child path name generation for temporary tables so that
  it does nothing special for temporary tables.
storage/myisammrg/ha_myisammrg.h:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added elements to class ha_myisammrg to support the new
  open approach.
  Changed empty destructor definition to a declaration.
  Implemented in ha_myisammrg.cc.
  Added declaration for methods attach_children() and
  detach_children().
  Added definition for method table_ptr() for use with
  callback functions.
storage/myisammrg/myrg_close.c:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Added a check to avoid closing of MyISAM tables when the
  child tables are not attached.
  Added freeing of rec_per_key_part when the child tables
  are not attached.
storage/myisammrg/myrg_extra.c:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  Some ::extra() functions and ::reset() can be called when
  children are detached.
storage/myisammrg/myrg_open.c:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE
              corrupts a MERGE table
  
  Kept old myrg_open() for MERGE use independent from MySQL.
  Removed an always true condition in myrg_open().
  Set children_attached for independent MERGE use in myrg_open().
  
  Added myrg_parent_open(), myrg_attach_children(), and
  myrg_detach_children() for the new MERGE table open approach.
mysql-test/r/merge-big.result:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE corrupts a MERGE table
  New test result
mysql-test/t/merge-big.test:
  Bug#26379 - Combination of FLUSH TABLE and REPAIR TABLE corrupts a MERGE table
  New test case