[Bug 787311] vtenc: leaks frames when shut down before EOS

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=787311

Nicolas Dufresne (stormer) <nicolas at ndufresne.ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
                 CC|                            |nicolas at ndufresne.ca
         Resolution|FIXED                       |---

--- Comment #36 from Nicolas Dufresne (stormer) <nicolas at ndufresne.ca> ---
(In reply to Matthew Waters (ystreet00) from comment #35)
> commit e7cf4c058ddd466a797623b5ddf162c263d07059
> Author: Nicola Murino <nicola.murino at gmail.com>
> Date:   Wed Sep 27 16:41:51 2017 +0200
> 
>     audiodecoder: flush decoder in transition PAUSED->READY
>     
>     https://bugzilla.gnome.org/show_bug.cgi?id=787311

That patch is problematic in GstV4l2VideoDecoder. First, we where already
flushing, so it's a duplicated. But then, the actual flush implementation
brings back the decoder to it's active state, to be ready to receive future
buffers. That's is a terrible idea, since it is a lot of IOCTL to requeue all
the flushed buffers into capture queue. It's makes tracing of the driver
interaction quite horrible and the quick STREAMOFF/STREAMON sequence caused by
the fact that we'll shut it down again in _stop() is known to trigger race
condition in some buggy firmware.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.