[Home]

Summary:ASTERISK-14283: Segmentation fault on Re-INVITE when t38pt_udptl=no
Reporter:Bastian Schern (schern)Labels:
Date Opened:2009-06-08 13:10:46Date Closed:2009-07-30 18:21:48
Priority:CriticalRegression?No
Status:Closed/CompleteComponents:Channels/chan_sip/T.38
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:( 0) gdb_2009-06-08_01.txt.gz
Description:Asterisk 1.6.1.1 crashes with a segmentation fault if T.38 is tried to use but t38pt_udptl=no.

****** STEPS TO REPRODUCE ******

ATA <- t38pt_udptl=yes -> Asterisk <- t38pt_udptl=no -> SIP-Carrier

OUTPUT:
~~~~~~~
[Jun  8 17:54:53]     -- SIP/gw2-00954190 is making progress passing it to SIP/base-1-00924250
[Jun  8 17:54:55]     -- SIP/gw2-00954190 is ringing
[Jun  8 17:55:01]     -- SIP/gw2-00954190 answered SIP/base-1-00924250

Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1076689248 (LWP 22540)]
0x00000000005415f5 in ast_udptl_get_us (udptl=0x0, us=0x402c1310) at udptl.c:876
876             memcpy(us, &udptl->us, sizeof(udptl->us));

****** ADDITIONAL INFORMATION ******

I created a gdb backtrace with the following steps (gdb_2009-06-08_01.txt.gz):
out-gw2:~# gdb --args asterisk -cvvvvvg
set logging on
bt
bt full
info thread
thread apply all bt
quit
Comments:By: Bastian Schern (schern) 2009-06-09 12:27:56

This bug does not appear in Asterisk 1.6.0.6.

By: Kevin P. Fleming (kpfleming) 2009-07-24 17:17:44

Please re-test with the latest 1.6.1 SVN branch; some important T.38 fixes were added this week. Thanks!

By: Bastian Schern (schern) 2009-07-27 16:01:14

I will test again during end of the week and report results back.

By: Bastian Schern (schern) 2009-07-30 16:06:14

Today I found the time to test again with the Asterisk SVN-branch-1.6.1-r209593M.
It works well and crashes not anymore. I think the bug is fixed.

By: snuffy (snuffy) 2009-07-30 18:21:47

Closed at reporters request