[gstreamer-bugs] [Bug 601587] MiniObject race condition

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=601587
  GStreamer | gstreamer (core) | git

--- Comment #4 from Ole André Vadla Ravnås <ole.andre.ravnas at tandberg.com> 2009-11-11 18:14:18 UTC ---
I should probably add that reproducing this in a real-world application is
either impossible or very hard. It may only happen if the buffer alloc and
dealloc happen in different threads. So for example:
"videotestsrc ! xvimagesink" is safe
whereas:
"videotestsrc ! queue ! xvimagesink" is not

The reason we discovered this is that I implemented a buffer pool in
ClutterGstVideoSink and things were pretty mellow for a while -- until a queue
was later introduced before the sink. I guess our use-case is an acid-test
though, as there's three threads involved in our case;
1) Upstream buffer producer (H264 decoder)
2) queue's srcpad task
3) Clutter's GMainLoop

We've been testing with the attached fix and so far the "mysterious" crashes
that we were seeing seem to have gone away.

-- 
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.