[asterisk-dev] [Code Review] 3753: media_formats: Prevent crash during masquerade

[rmudgett]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/3753/#review12614
-----------------------------------------------------------

Ship it!



/team/group/media_formats-reviewed-trunk/main/channel.c
<https://reviewboard.asterisk.org/r/3753/#comment22899>

    This may have been wrong anyway for the surviving channel.  After a masquerade, the surviving channel could be using the preferred codec instead of what was negotiated.
    
    The guts of SIP/100 (clonechan) masquerades into SIP/200 (original) for call pickup.
    SIP/100 (codec list: gsm, alaw) and was using alaw
    After the pickup masquerade it would be using gsm.


- rmudgett


On July 11, 2014, 9:53 p.m., Matt Jordan wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/3753/
> -----------------------------------------------------------
> 
> (Updated July 11, 2014, 9:53 p.m.)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Repository: Asterisk
> 
> 
> Description
> -------
> 
> When a masquerade occurs, a newly created channel replaces an existing channel and steals its private data structure. This includes swapping the formats and capabilities.
> 
> Since the newly created channel only contains ast_format_none, this causes an assert to fire when the masqueraded channel is destroyed.
> 
> Removing the assert isn't a good idea. However, there's also no real need to do the accessing that fires the asserts either: the masqueraded channel will be destroyed, the references will be cleaned up appropriately, and life will go on. free_translation is also called in a channel destructor, and again, things will be cleaned up appropriately without going through the accessors that have the asserts.
> 
> 
> Diffs
> -----
> 
>   /team/group/media_formats-reviewed-trunk/main/channel.c 418435 
> 
> Diff: https://reviewboard.asterisk.org/r/3753/diff/
> 
> 
> Testing
> -------
> 
> The crashing test (tests/apps/dial/dial_dtmf_hangup_cancel) now passes
> 
> 
> Thanks,
> 
> Matt Jordan
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-dev/attachments/20140714/8b53ebc9/attachment-0001.html>