[asterisk-dev] [Code Review] [17949] Error that mutex 'dialog' has been free'd more than it's been locked in add_header_max_forwards

[Olle E Johansson]

> On 2010-11-03 16:35:22, Russell Bryant wrote:
> > It would be good to add a comment to add_header_max_forwards() that documents the assumption about the sip_pvt lock being held before the function is called.
> > 
> > Also, add some info to the "testing done" section of this review.  I would suggest starting with just making some basic calls.  From there, try running the Asterisk test suite to make sure the tests using SIP pass before and after the patch (or at least fail in the same way if something goes wrong ...).  Be aware, though, that the test suite will wipe out your configuration, so back up your configs before you try to run it.

I think we should revisit my original design. We could get the value of the channel variable earlier, when we set up the channel instead of reading the channel variable for each message. We only need to get it once from the dialplan, when we start the dial.


- Olle E


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/997/#review2880
-----------------------------------------------------------


On 2010-11-03 16:28:14, Brett Bryant wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/997/
> -----------------------------------------------------------
> 
> (Updated 2010-11-03 16:28:14)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Summary
> -------
> 
> Some functions passing the sip_pvt structure dialog to add_header_max_forwards were not following proper locking order, and this patch is intended to fix that.
> 
> 
> This addresses bug 17949.
>     https://issues.asterisk.org/view.php?id=17949
> 
> 
> Diffs
> -----
> 
>   /branches/1.8/channels/chan_sip.c 283596 
> 
> Diff: https://reviewboard.asterisk.org/r/997/diff
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Brett
> 
>