The problem is that EGLNativeWindowSurface and wl_egl_surface are the
same object underneath, so we must recreate both together. As an
optimization, the EGLNativeWindowSurface wrapper is only re-created
if the window_handle changed.
On Mesa, this would cause crash, which will be fixed by:
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11979
And will lead to proper errors in the future or on other GL stack. This
issue was encounter using a permanent GstGLDisplay after cycling one of
multiple independent pipelines through NULL state.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1230>
Currently the decoder selection is very naive: The type with the highest
rank that matches the current caps is used. This works well for software
decoders. The exact supported caps are always known and the static caps are
defined accordingly.
With hardware decoders, e.g. vaapi, the situation is different. The decoder
may reject the caps later during a caps query. At that point, a new decoder
is created. However, the same type is chosen an after several tries,
decodebin fails.
To avoid this, do the caps query while adding the decoder and try again
with other decoder types if the query fails:
1. create the decoder from the next matching type
2. add and link the decoder
3. change the decoder state to READY
4. do the caps query
if it fails then remove the decoder again and go back to 1.
5. expose the source pad
6. sync the decoder state with the parent.
This way, the decoder is already part of the pipeline when the state change
to READY happens. So context handling should work as before.
Exposing the source pad after the query was successful is important:
Otherwise the thread from the decoder source pad may block in a blocked pad
downstream in the playsink waiting for other pads to be ready.
The thread now blocks trying to set the state back to NULL while holding
the SELECTION_LOCK. Other streams may block on the SELECTION_LOCK and the
playsink never unblocks the pad. The result is a deadlock.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1201>
When the window is called, we properly destroy all surfaces, which effectively
will unmap that surface and should make it disapear on screen, but we also
destroy the wl_source, a GSource that is resposibble of dispatching and executing
messages to/from the Wayland server.
As a side effect, the server never gets the message and the surfaces are
"leaked" on the server. We fix this using wl_display_flush() before destroying
the wl_source.
Fixes#815
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1226>
Previously one of the branches did not check for the property value. To
avoid this in the future, check inside the QoS calculation function
instead.
As a side effect this now always prints the debug messages into the logs
when samples are dropped, which is useful information even without the
QoS messages.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1224>
If a buffer is dropped during resyncing on a discont because either its
end offset is already before the current output offset of the
aggregator or because it fully overlaps with the part of the current
output buffer that was already filled, then don't just assume that the
next buffer is going to start at exactly the expected offset. It might
still require some more dropping of samples.
This caused the input to be mixed with an offset to its actual position
in the output stream, causing additional latency and wrong
synchronization between the different input streams.
Instead consider each buffer after a discont as a discont until the
aggregator actually resynced and starts mixing samples from the input
again.
Also update the start output offset of a new input buffer if samples
have to be dropped at the beginning. Otherwise it might be mixed too
early into the output and overwrite part of the output buffer that
already took samples from this input into account.
Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/issues/912
which is a regression introduced by https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1180/
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1224>
A video decoder can now receive subframes and start decoding
instead of waiting for the full frame to be complete.
Subframe support will reduce latency as described in the
video encoder base class.
A unit test illustrating this API is available in
tests/check/libs/videodecoder.c.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/454>
Implement 8-bit values of SMPTE RP 2019-1:2014. The bar widths and
heights are the result of fractions as integers. The remainders of
widths are distributed in a way that they match the values in Table
C.1 (a) in the specification.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1063>
GstRTPHeaderExtension::write can map the RTP buffer for reading. If that
happens on a buffer that is already mapped WRITE-only by the payloader,
the payloader's mapping gets invalidated (GstRTPBuffer::map will point
to a different instance of GstMemory).
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1173>
The finish() virtual function documentation state that "Sub-classes can refuse
to decode new data after." Though, it is very common to issue a non-flushing
seek after that event in gapless playback uses case. This fixes potential
stalls with code using segment seeks, by using drain() virtual funciton
instead.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1206>
There may be two or more threads involved here however the important
interaction is the use of ogg->seeK_event_drop_till value that was only
set in the push-mode seek-event thread and could race with upstream
sending e.g. and EOS (or data).
Scenario is this:
1. oggdemux performs a seek to near the end of the file to try and find
the duration. ogg->push_state is set to PUSH_DURATION.
2. Seek is picked up by the dedicated seek event thread and sets
ogg->seek_event_drop_till to the seek event's seqnum.
3. Most operations are blocked or dropped waiting on the duration to
be determined and processing continues until a duration is found.
4. Two branching options for how this ultimately plays out
4a. The source is too fast and we receive an EOS event which is dropped
because ogg->push_state == PUSH_DURATION. In this case everything
works.
4b. We hit our 'almost at the end' check in
gst_ogg_pad_handle_push_mode_state() and attempt to seek back to the
beginning (or to a user-provided seek). This seek is marshalled to
the seek event thread without setting ogg->seek_event_drop_till but
with change ogg->push_state = PUSH_PLAYING. If an EOS event or
e.g. buffers arrive from upstream before the seek event thread has
picked up the seek event, then the EOS/data is processed as if it
came as a result of the seek event. This is the case that fails.
The fix is two-fold:
1. Preemptively set ogg->seek_event_drop_till when setting the seek
event so that data and other events can be dropped correctly.
2. In addition to dropping and EOS events while ogg->push_state ==
PUSH_DURATION, also drop any EOS events that are received before the
seek event has been processed by also tracking the seqnum of the seek.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1196>
below commit change the window resize thread and cause viv-fb backend
hang, need move resize code after window->open is called. Otherwise,
the resize message will send to a thread that not start running and
window resize call will waiting forever.
Commit: b887db1efe
glwindow: fix racy resize updates
Take locks around resize handling and marshall all resizes to the
windowing thread by default.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1195>
