[asterisk-bugs] [Asterisk 0018735]: memory leak with IAX in 1.6.2

[Asterisk Bug Tracker]

A NOTE has been added to this issue. 
====================================================================== 
https://issues.asterisk.org/view.php?id=18735 
====================================================================== 
Reported By:                tzafrir
Assigned To:                
====================================================================== 
Project:                    Asterisk
Issue ID:                   18735
Category:                   Channels/chan_iax2
Reproducibility:            sometimes
Severity:                   minor
Priority:                   normal
Status:                     feedback
Asterisk Version:           SVN 
JIRA:                        
Regression:                 No 
Reviewboard Link:            
SVN Branch (only for SVN checkouts, not tarball releases):  trunk 
SVN Revision (number only!): 306010 
Request Review:              
====================================================================== 
Date Submitted:             2011-02-02 12:43 CST
Last Modified:              2011-02-14 13:35 CST
====================================================================== 
Summary:                    memory leak with IAX in 1.6.2
Description: 
I'm trying to track a certain memory leak: the sizes (rss, size, vsize)
keep increasing, whereas the memory size shown by 'memory show summary' and
'astobj2 show allocations' don't inflate.

When I stop the remote Asterisk server, the memory inflation stops. Thus
it must be triggered by the registration requests.

I managed to reproduce this on several Centos5 systems, but not on my
laptop (x86_64, Debian Squeeze). It has been around for a number of months,
at least.
====================================================================== 

---------------------------------------------------------------------- 
 (0131939) tzafrir (manager) - 2011-02-14 13:35
 https://issues.asterisk.org/view.php?id=18735#c131939 
---------------------------------------------------------------------- 
I ran valgrind twice. Once I stopped it quickly. The second time I stopped
it after an hour (or more, not sure).

valgrind was run with:  -v --leak-check=full

The main difference between the logs (besides obvious ones, such as
process IDs)

+==18036== 167,264 bytes in 1,101 blocks are possibly lost in loss record
139 of 143
+==18036==    at 0x4021E22: calloc (vg_replace_malloc.c:397)
+==18036==    by 0x4010BAB: _dl_allocate_tls (in /lib/ld-2.7.so)
+==18036==    by 0x430C672: pthread_create@@GLIBC_2.1 (in
/lib/i686/cmov/libpthread-2.7.so)
+==18036==    by 0x81AC655: ast_pthread_create_stack (utils.c:1064)
+==18036==    by 0x6C8593E: find_idle_thread (chan_iax2.c:1437)
+==18036==    by 0x6C9CE0B: socket_read (chan_iax2.c:9609)
+==18036==    by 0x8122BF0: ast_io_wait (io.c:288)
+==18036==    by 0x6C91CDA: network_thread (chan_iax2.c:12193)
+==18036==    by 0x81AA72A: dummy_start (utils.c:1010)
+==18036==    by 0x430D4BF: start_thread (in
/lib/i686/cmov/libpthread-2.7.so)
+==18036==    by 0x516184D: clone (in /lib/i686/cmov/libc-2.7.so)
+==18036==
+==18036== 

Issue History 
Date Modified    Username       Field                    Change               
====================================================================== 
2011-02-14 13:35 tzafrir        Note Added: 0131939                          
======================================================================