[asterisk-dev] [Code Review] Chan_sip: Voice frame dropped for every early media audio call

Olle E Johansson reviewboard at asterisk.org
Tue Apr 19 03:18:27 CDT 2011



> On 2011-04-18 12:50:26, Matthew Nicholson wrote:
> > /branches/1.4/channels/chan_sip.c, lines 4064-4067
> > <https://reviewboard.asterisk.org/r/1186/diff/1/?file=16181#file16181line4064>
> >
> >     I think this does need to stay within the lock.

Thanks for quick feedback. I'll make that happen.


- Olle E


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviewboard.asterisk.org/r/1186/#review3389
-----------------------------------------------------------


On 2011-04-18 05:55:30, Olle E Johansson wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviewboard.asterisk.org/r/1186/
> -----------------------------------------------------------
> 
> (Updated 2011-04-18 05:55:30)
> 
> 
> Review request for Asterisk Developers.
> 
> 
> Summary
> -------
> 
> The code for checking T.38 in sip_write accidentally drops one frame in situations where an audio frame forces early media . If you compare with the video code below this part of the code, the frame is not dropped even though we add an 183 message. It's not a big issue, but nevertheless, frames are frames and should be treated with love and care.
> 
> Moved the T.38 check out of the lock - maybe that's wrong?
> 
> 
> This addresses bug 19312.
>     https://issues.asterisk.org/view.php?id=19312
> 
> 
> Diffs
> -----
> 
>   /branches/1.4/channels/chan_sip.c 313187 
> 
> Diff: https://reviewboard.asterisk.org/r/1186/diff
> 
> 
> Testing
> -------
> 
> Can't test with T.38 - only with audio. It works.
> 
> 
> Thanks,
> 
> Olle E
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-dev/attachments/20110419/08f70647/attachment.htm>


More information about the asterisk-dev mailing list