[Bug 650916] REGRESSION: ssrcdemux causing FLOW_NOT_LINKED

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=650916
  GStreamer | gst-plugins-good | git

--- Comment #2 from Håvard Graff (hgr) <havard.graff at tandberg.com> 2011-05-24 04:21:55 UTC ---
(In reply to comment #1)
> The problem the original patch wants to fix is that if an event is pushed
> through the pad while the signal is emitted, then it will deadlock.. Maybe we
> should hold the STREAM lock of the new pad while emitting the signal. 
> 

With the old code, AFAICT, if an event arrived while signal was emitted, the
event would be waiting for the PAD_LOCK. Now, unless the event was directly and
blockingly triggered by the emitted signal, I don't see that there should be a
deadlock?

> 
> While reading the code, I'm always worried that if in the chain functions the
> clear_ssrc function is called between the time when the lock is released and
> the time where the pad_push function is called.

Did not the original implementation hold the PAD_LOCK right through the push?

-- 
Configure bugmail: https://bugzilla.gnome.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
You are the assignee for the bug.