Bug #46081 Valgrind warnings (JOIN code) in test rpl_trigger
Submitted: 9 Jul 2009 13:08 Modified: 4 Aug 2009 7:48
Reporter: Guilhem Bichot Email Updates:
Status: Duplicate Impact on me:
None 
Category:MySQL Server: Optimizer Severity:S3 (Non-critical)
Version:5.1-bzr OS:Linux (64)
Assigned to: Sergei Glukhov CPU Architecture:Any
Triage: Triaged: D2 (Serious)

[9 Jul 2009 13:08] Guilhem Bichot
Description:
Using 5.1 with latest revision:
joro@sun.com-20090708131116-kyz8iotbum8w9yic
BUILD/compile-pentium64-valgrind-max,
./mtr --mysqld=--binlog-format=mixed rpl_trigger --valgrind
shows:
==30171== Thread 10:
==30171== Source and destination overlap in memcpy(0x77630A8, 0x77630A8, 28)
==30171==    at 0x4C271DA: memcpy (mc_replace_strmem.c:402)
==30171==    by 0x66A6EC: JOIN::init(THD*, List<Item>&, unsigned long long, select_result*) (sql_select.h:469)
==30171==    by 0x66A83D: JOIN::JOIN(THD*, List<Item>&, unsigned long long, select_result*) (sql_select.h:418)
==30171==    by 0x7589F1: mysql_select(THD*, Item***, TABLE_LIST*, unsigned, List<Item>&, Item*, unsigned, st_order*, st_order*, Item*, st
_order*, unsigned long long, select_result*, st_select_lex_unit*, st_select_lex*) (sql_select.cc:2361)
==30171==    by 0x75E1E0: handle_select(THD*, st_lex*, select_result*, unsigned long) (sql_select.cc:268)
==30171==    by 0x6C827D: _ZL21execute_sqlcom_selectP3THDP10TABLE_LIST (sql_parse.cc:5012)
==30171==    by 0x6CA19D: mysql_execute_command(THD*) (sql_parse.cc:2207)
==30171==    by 0x895572: sp_instr_stmt::exec_core(THD*, unsigned*) (sp_head.cc:2907)
==30171==    by 0x8957AA: sp_lex_keeper::reset_lex_and_exec_core(THD*, unsigned*, bool, sp_instr*) (sp_head.cc:2735)
==30171==    by 0x89BD18: sp_instr_stmt::execute(THD*, unsigned*) (sp_head.cc:2849)
==30171==    by 0x897BED: sp_head::execute(THD*) (sp_head.cc:1253)
==30171==    by 0x8989BC: sp_head::execute_procedure(THD*, List<Item>*) (sp_head.cc:1983)
==30171==    by 0x6D0CFB: mysql_execute_command(THD*) (sql_parse.cc:4353)
==30171==    by 0x6D2EA7: mysql_parse(THD*, char const*, unsigned, char const**) (sql_parse.cc:5932)
==30171==    by 0x6D3C5E: dispatch_command(enum_server_command, THD*, char*, unsigned) (sql_parse.cc:1213)
==30171==    by 0x6D503C: do_command(THD*) (sql_parse.cc:854)
==30171== 
==30171== Source and destination overlap in memcpy(0x77630A8, 0x77630A8, 28)
==30171==    at 0x4C271DA: memcpy (mc_replace_strmem.c:402)
==30171==    by 0x712F8C: setup_wild(THD*, TABLE_LIST*, List<Item>&, List<Item>*, unsigned) (sql_base.cc:7378)
==30171==    by 0x757ABD: JOIN::prepare(Item***, TABLE_LIST*, unsigned, Item*, unsigned, st_order*, st_order*, Item*, st_order*, st_select_lex*, st_select_lex_unit*) (sql_select.cc:494)
==30171==    by 0x758AC9: mysql_select(THD*, Item***, TABLE_LIST*, unsigned, List<Item>&, Item*, unsigned, st_order*, st_order*, Item*, st_order*, unsigned long long, select_result*, st_select_lex_unit*, st_select_lex*) (sql_select.cc:2367)
==30171==    by 0x75E1E0: handle_select(THD*, st_lex*, select_result*, unsigned long) (sql_select.cc:268)
==30171==    by 0x6C827D: _ZL21execute_sqlcom_selectP3THDP10TABLE_LIST (sql_parse.cc:5012)
==30171==    by 0x6CA19D: mysql_execute_command(THD*) (sql_parse.cc:2207)
==30171==    by 0x895572: sp_instr_stmt::exec_core(THD*, unsigned*) (sp_head.cc:2907)
==30171==    by 0x8957AA: sp_lex_keeper::reset_lex_and_exec_core(THD*, unsigned*, bool, sp_instr*) (sp_head.cc:2735)
==30171==    by 0x89BD18: sp_instr_stmt::execute(THD*, unsigned*) (sp_head.cc:2849)
==30171==    by 0x897BED: sp_head::execute(THD*) (sp_head.cc:1253)
==30171==    by 0x8989BC: sp_head::execute_procedure(THD*, List<Item>*) (sp_head.cc:1983)
==30171==    by 0x6D0CFB: mysql_execute_command(THD*) (sql_parse.cc:4353)
==30171==    by 0x6D2EA7: mysql_parse(THD*, char const*, unsigned, char const**) (sql_parse.cc:5932)
==30171==    by 0x6D3C5E: dispatch_command(enum_server_command, THD*, char*, unsigned) (sql_parse.cc:1213)
==30171==    by 0x6D503C: do_command(THD*) (sql_parse.cc:854)

I'm using valgrind-3.3.1-Debian

How to repeat:
run the test.
However, I don't see this warning in pushbuild2 for 5.1, so maybe this is a problem of valgrind version, platform...
For example, my Valgrind also prints:
--30171-- WARNING: unhandled syscall: 142
--30171-- You may be able to write your own handler.
--30171-- Read the file README_MISSING_SYSCALL_OR_IOCTL.
--30171-- Nevertheless we consider this a bug.  Please report
--30171-- it at http://valgrind.org/support/bug_reports.html.
so maybe it has a bug.
[9 Jul 2009 13:13] Guilhem Bichot
With tree 5.1-performance-version, I don't have the "unhandled signal 142" but have the same JOIN warnings. So maybe my Valgrind is working.
[29 Jul 2009 21:03] Susanne Ebrecht
Verified as described
[4 Aug 2009 7:48] Sergei Glukhov
duplicated with bug#46003