MySQL Bugs: #59022: Valgrind warning in write

Bug #59022	Valgrind warning in write_delayed() after InnoDB merge
Submitted:	18 Dec 2010 6:47	Modified:	6 Jan 2011 1:19
Reporter:	Alexander Nozdrin	Email Updates:
Status:	Closed	Impact on me:	None
Category:	MySQL Server: General	Severity:	S1 (Critical)
Version:		OS:	Any
Assigned to:	Alexander Nozdrin	CPU Architecture:	Any
Tags:	pb2, test failure

Description:
Tree: mysql-trunk-bugfixing
Test: binlog.binlog_max_extension

==6019== 4 bytes in 1 blocks are definitely lost in loss record 2 of 237
==6019==    at 0x4C20E1C: malloc (vg_replace_malloc.c:195)
==6019==    by 0x90417C: my_malloc (my_malloc.c:38)
==6019==    by 0x5C3FB0: write_delayed(THD*, TABLE*, enum_duplicates, st_mysql_lex_string, bool, bool) (sql_insert.cc:2424)
==6019==    by 0x5C65E6: mysql_insert(THD*, TABLE_LIST*, List<Item>&, List<List<Item> >&, List<Item>&, List<Item>&, enum_duplicates, bool) (sql_insert.cc:923)
==6019==    by 0x5DA265: mysql_execute_command(THD*) (sql_parse.cc:2845)
==6019==    by 0x5DF440: mysql_parse(THD*, char*, unsigned int, Parser_state*) (sql_parse.cc:5550)
==6019==    by 0x5E0A89: dispatch_command(enum_server_command, THD*, char*, unsigned int) (sql_parse.cc:1078)
==6019==    by 0x5E1FD2: do_command(THD*) (sql_parse.cc:815)
==6019==    by 0x6B3795: do_handle_one_connection(THD*) (sql_connect.cc:748)
==6019==    by 0x6B388E: handle_one_connection (sql_connect.cc:684)
==6019==    by 0x92D973: pfs_spawn_thread (pfs.cc:1360)
==6019==    by 0x4E2B73C: start_thread (in /lib64/libpthread-2.5.so)
==6019==    by 0x59DBD1C: clone (in /lib64/libc-2.5.so)

How to repeat:
Check out PB.

It seems, this begun in trunk-innodb-pb2 and came to trunk-bugfixing
with the recent merge. Assigning to Vasil to start with.

Please push a fix into trunk-bugfixing directly!

NOTE: I can not reproduce the warning by running that single test case.
Neither locally (Ubuntu 10.10, valgrind 3.6.0), nor on tyr29 (PB host).

However, it is perfectly reproducible in PB.

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/127482

3465 Alexander Nozdrin	2010-12-22
      A patch for Bug#59022 (Valgrind warning in write_delayed() after InnoDB merge).
      
      Free allocated bitmap.

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/127504

3465 Alexander Nozdrin	2010-12-22
      A patch for Bug#59022 (Valgrind warning in write_delayed() after InnoDB merge).
      
      Free allocated bitmap.

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/127506

3465 Alexander Nozdrin	2010-12-22
      A patch for Bug#59022 (Valgrind warning in write_delayed() after InnoDB merge).
      
      Fix memory leak introduced by WL#5092 'RBR: Options for writing
      partial or full row images in RBR events'.
      
      Free memory which is allocated for column bitmap associated with
      each row to be processed by delayed insert handler thread.

Pushed to trunk-bugfixing.

Pushed into mysql-trunk 5.6.1 (revid:alexander.nozdrin@oracle.com-20101222212842-y0t3ibtd32wd9qaw) (version source revid:alexander.nozdrin@oracle.com-20101222212842-y0t3ibtd32wd9qaw) (merge vers: 5.6.1) (pib:24)

Bug does not appear in any released version. No changelog entry needed.