[Devel] [ openser-Bugs-1721057 ] failed pkg_malloc() causes real memory to leak.

Kobi Eshun kobi at sightspeed.com
Thu May 24 00:20:43 CEST 2007


Hi,

Sorry about the delayed response. Here is the output from a valgrind  
run under the known-bad conditions (i.e., with the bad query against  
a bloated watchers table):

==26483== 26,340 (1,276 direct, 25,064 indirect) bytes in 11 blocks  
are definitely lost in loss record 26 of 43
==26483==    at 0x4018524: malloc (in /usr/lib/valgrind/x86-linux/ 
vgpreload_memcheck.so)
==26483==    by 0x434A1BD: my_malloc (in /usr/lib/mysql/ 
libmysqlclient.so.15.0.0)
==26483==    by 0x436FDEF: mysql_store_result (in /usr/lib/mysql/ 
libmysqlclient.so.15.0.0)
==26483==    by 0x42F292D: db_mysql_store_result (in /etc/openser/ 
togo/mysql.so)
==26483==    by 0x4671C4D: msg_watchers_clean (in /etc/openser/togo/ 
presence.so)
==26483==    by 0x80985EF: timer_ticker (timer.c:226)
==26483==    by 0x8098459: run_timer (timer.c:301)
==26483==    by 0x8068F23: main_loop (main.c:702)
==26483==    by 0x806A78D: main (main.c:1399)
==26483==
==26483==
==26483== 4,555,628 bytes in 31 blocks are possibly lost in loss  
record 42 of 43
==26483==    at 0x4018524: malloc (in /usr/lib/valgrind/x86-linux/ 
vgpreload_memcheck.so)
==26483==    by 0x434A1BD: my_malloc (in /usr/lib/mysql/ 
libmysqlclient.so.15.0.0)
==26483==    by 0x436FDEF: mysql_store_result (in /usr/lib/mysql/ 
libmysqlclient.so.15.0.0)
==26483==    by 0x42F292D: db_mysql_store_result (in /etc/openser/ 
togo/mysql.so)
==26483==    by 0x4671C4D: msg_watchers_clean (in /etc/openser/togo/ 
presence.so)
==26483==    by 0x80985EF: timer_ticker (timer.c:226)
==26483==    by 0x8098459: run_timer (timer.c:301)
==26483==    by 0x8068F23: main_loop (main.c:702)
==26483==    by 0x806A78D: main (main.c:1399)

Hope this helps. Now that I'm back in the code, I'll take a look as  
well. Cheers,
--
kobi


On May 21, 2007, at 8:11 AM, SourceForge.net wrote:

> Hi Ekobi,
>
> which presence table is about? I'm warking with Anca on this issue,  
> but we
> need fitst to locate it. If the conventional mem is leaking, it  
> means it is
> a mysql driver error - either the free_result() is not called,  
> either there
> is a bug in the driver/library.
>
> Thanks and regards,
> Bogdan

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://openser.org/pipermail/devel/attachments/20070523/8be15c3b/attachment.html


More information about the Devel mailing list