[asterisk-dev] [Code Review] 3401: Sorcery: Give sorcery instances references to the underlying wizards.
Joshua Colp
reviewboard at asterisk.org
Wed Mar 26 13:31:43 CDT 2014
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/3401/#review11383
-----------------------------------------------------------
Ship it!
Ship It!
- Joshua Colp
On March 26, 2014, 6:20 p.m., Mark Michelson wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/3401/
> -----------------------------------------------------------
>
> (Updated March 26, 2014, 6:20 p.m.)
>
>
> Review request for Asterisk Developers.
>
>
> Repository: Asterisk
>
>
> Description
> -------
>
> This small change is intended to address a crash seen in the testsuite, specifically with the PJSIP qualify test.
>
> In that test, when Asterisk shuts down, sorcery kills its wizards container along with its sorcery instance container. After this occurs, a scheduled qualify is triggered, which attempts to use a freed wizard, therefore causing a crash. The reason for this is that sorcery instances have a bare pointer to a wizard (through an intermediary ast_sorcery_object_wizard struct). When the container of wizards is freed, the pointer on the sorcery instance is left dangling.
>
> This patch addresses the problem by keeping wizards alive as long as a sorcery instance has a reference to the wizard.
>
>
> Diffs
> -----
>
> /branches/12/main/sorcery.c 411188
>
> Diff: https://reviewboard.asterisk.org/r/3401/diff/
>
>
> Testing
> -------
>
> This change makes the PJSIP qualify test go from crashing 1 in 5 runs to running 120+ times with no crashes.
>
>
> Thanks,
>
> Mark Michelson
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-dev/attachments/20140326/7bc92a8a/attachment.html>
More information about the asterisk-dev
mailing list