Recording SSRC issues

[Pete Kelly]

Hi, I was wondering if you would be able to cast your eyes over a change that I have made (and tested) to the stream slot logic part of the recorder.

The issue I have been seeing is a situation where there are a lot of media re-establishments via reinvite etc, where the codec PT is changed but the codec remains the same, and the SSRC remains the same!

So a new slot was being created, but then errors were appearing in the logs 

"[S tag-0-media-1-component-1-RTP-id-2] [0xefcbee67] [core] Failed to add frame to mixer: received samples for old re-used input channel"

and ultimately the SSRC was not recorded.

I have made a change to use the SSRC as the key for the mixer input_ref and it has resolved the issue. This is the change-set, I realise it may not be suitable to be pushed into the main project but I would be interested to know your comments on the suitability of what I have done and whether in your experience it could expose some issues:

https://github.com/petekelly/rtpengine/commit/f3cfb0cc45f674db65f508f8ffe23440dc48794e

[Richard Fuchs]

This actually seems perfectly reasonable to me. At least I can't immediately think of any scenario where this would lead to a regression. 😁

Feel free to open a PR.

Cheers

[Pete Kelly]

I've tested this for a while and it seems fine. Opened PR https://github.com/sipwise/rtpengine/pull/2009 

However it's against an old branch! If you're OK with the changes I can re-open against master? (although it looks like types.h has a few changes now...)