[asterisk-bugs] [JIRA] (ASTERISK-26973) bridge: Crash when freeing frame and snooping

[Kevin Harwell (JIRA)]

     [ https://issues.asterisk.org/jira/browse/ASTERISK-26973?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Kevin Harwell updated ASTERISK-26973:
-------------------------------------

    Attachment: ASTERISK-26973.diff

Thanks for the information and testing!

Unfortunately I am still unable to duplicate the crash myself. However, I hope now that I have tracked down the offending patch - commit # for Asterisk 13:
{noformat}
430f6e5388b6e18b31e1a08582112838f6cc7557
{noformat}
It's related to issue ASTERISK-26419

I've attached a patch, ASTERISK-26973.diff. If you could apply the patch and run your test to see if it fixes your crash it would be most helpful. If the attached patch does not fix it then next I suggest removing the changes (https://gerrit.asterisk.org/#/c/4024/) made by the above commit and see if that fixes it.

If neither of those work then it's something else. If that is the case then, if you can, try running Asterisk with valgrind and post the results.

> bridge: Crash when freeing frame and snooping
> ---------------------------------------------
>
>                 Key: ASTERISK-26973
>                 URL: https://issues.asterisk.org/jira/browse/ASTERISK-26973
>             Project: Asterisk
>          Issue Type: Bug
>      Security Level: None
>          Components: Bridges/bridge_simple, Resources/res_stasis_snoop
>    Affects Versions: 13.15.0
>         Environment: Linux 2.6.32-642.11.1.el6.x86_64 #1 SMP Fri Nov 18 19:25:05 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
>            Reporter: Michel R. Vaillancourt
>            Assignee: Kevin Harwell
>         Attachments: ARI.snoop.code, ASTERISK-26973.diff, backtrace-and-debug.tgz, core.22813.backtrace.txt, core.asterisk.15474.backtrace.txt, Snoop Crash Process.txt
>
>
> {noformat}
> [Apr 26 00:11:24]     -- Channel Snoop/[REDACTED]-00000002 joined 'simple_bridge' stasis-bridge <br6308>
> [Apr 26 00:11:28]   == Setting global variable 'SIPDOMAIN' to '[REDACTED]'
> [Apr 26 00:11:28]     -- Executing [[REDACTED]@inbound:1] Verbose("PJSIP/[REDACTED]-0000395a", "2, Call from "[REDACTED]" <[REDACTED]> to [REDACTED]") in new stack
> [Apr 26 00:11:28]   ==  Call from "[REDACTED]" <[REDACTED]> to [REDACTED]
> [Apr 26 00:11:28]     -- Executing [[REDACTED]@inbound:2] Stasis("PJSIP/[REDACTED]-0000395a", "[REDACTED],external") in new stack
> [Apr 26 00:11:30]     -- <PJSIP/[REDACTED]-00003894> Playing 'beep.slin16' (language 'en')
> [Apr 26 00:11:30]     -- Channel PJSIP/[REDACTED]-0000395a joined 'simple_bridge' stasis-bridge <[REDACTED]>
> [Apr 26 00:11:30]        > 0x7fbc741cd8a0 -- Probation passed - setting RTP source address to [REDACTED]:17564
> asterisk2*CLI>
> Disconnected from Asterisk server
> [Apr 26 00:11:31] Asterisk cleanly ending (0).
> [Apr 26 00:11:31] Executing last minute cleanups
> Asterisk ending (0).
> {noformat}
> We upgraded one dev PBX and one production PBX to 13.15.0 and have discovered that suddenly a function that worked just fine before now causes Asterisk to SegFault.  Note that this ARI code function has been stable / safe for over a year;  the behavior began as soon as we upgraded.  
> Our DEV and PRD machines still running prior versions of Asterisk using the same ARI code do not exhibit this behavior.



--
This message was sent by Atlassian JIRA
(v6.2#6252)