GstTask going to pause

[Matteo Valdina]

Hi,

I'm facing an issue with a GstTask gone in a pause state sometimes (1 out
of 20 times).

I have a pipeline that is a v4lsrc -> tee -> queue -> vaapisink,
periodically I attach and detach a "queue -> vaaph24enc -> rtph264pay ->
rtpbin -> udpsink".

The problem happens when I attach the encoder branch to the tee element. It
does not happen regularly it took something like 1 each 20 attach/detach to
happen. Sometimes more sometimes less.

When I attach the encoder I don't do anything special like. I add all
elements to the pipeline, link together, request a new pad to the tee, link
to the newly added branch with the tee and sync with the parent state.

Logs is not very helpful. I only see that the GstTask went on pause.
```
Apr 24 07:54:55 linux vd[23663]: trying to link tee0:src_10 and q_enc:sink
Apr 24 07:54:55 linux vd[23663]: ts: 68:03:24.609568000 now
68:03:24.642953584 delay 0:00:00.033385584
Apr 24 07:54:55 linux vd[23663]: sync to 0:02:43.266665034 out ts
0:03:58.745488463
Apr 24 07:54:55 linux vd[23663]: lost frames detected: count = 1 - ts:
0:03:58.745488463
Apr 24 07:54:55 linux vd[23663]: Received event on flushing pad. Discarding
Apr 24 07:54:55 linux vd[23663]: pausing after gst_pad_push() = flushing
Apr 24 07:54:55 linux vd[23663]: Task going to paused
Apr 24 07:54:55 linux vd[23663]: linked tee0:src_10 and q_enc:sink,
successful
```

Any suggestion? Should I add the new branch in a probe IDLE?

Best
Matteo Valdina

-- 
“There are two ways of constructing a software design: One way is to make
it so simple that there are obviously no deficiencies, and the other way is
to make it so complicated that there are no obvious deficiencies. The first
method is far more difficult.”
- Tony Hoare
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/gstreamer-devel/attachments/20180424/5641d81c/attachment.html>