platform/upstream/gstreamer.git
3 years agod3d11decoder: Don't print error log when no DPB texture is available
Seungha Yang [Thu, 10 Jun 2021 14:35:38 +0000 (23:35 +0900)]
d3d11decoder: Don't print error log when no DPB texture is available

... but we are flushing. The condition is quite expected situation
when pipeline is in the middle of seeking operation.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2320>

3 years agomsdkenc: add extbrc support in ext-coding-props property
Yinhang Liu [Sun, 23 May 2021 10:17:38 +0000 (18:17 +0800)]
msdkenc: add extbrc support in ext-coding-props property

The SDK can support external bitrate control [1], so add extbrc
to enable this feature.

[1] https://github.com/Intel-Media-SDK/MediaSDK/blob/master/doc/mediasdk-man.md#mfxextcodingoption2

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2139>

3 years agomsdkenc: add ext-coding-props for external coding options
Yinhang Liu [Sun, 23 May 2021 10:13:25 +0000 (18:13 +0800)]
msdkenc: add ext-coding-props for external coding options

This property supports passing multiple parameters using GstStructure.

Example usage:
ext-coding-props="props,key0=value0,key1=value1,..."

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2139>

3 years agocodecs: Fix the H265 poc out of order warning.
He Junyan [Sat, 5 Jun 2021 13:59:50 +0000 (21:59 +0800)]
codecs: Fix the H265 poc out of order warning.

We always get a warning such as:
h265decoder gsth265decoder.c:1432:gst_h265_decoder_do_output_picture: \
<vah265dec0> Outputting out of order 255 -> 0, likely a broken stream
in H265 decoder.

The problem is caused because we fail to reset the last_output_poc when
we get IDR and BLA. The incoming IDR and BLA frame already bump all the
frames in the DPB, but we forget to reset the last_output_poc, which
make the POC out of order and generate the warning all the time.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2294>

3 years agowasapi2sink: Fix ringbuffer object leak
Seungha Yang [Wed, 9 Jun 2021 16:09:44 +0000 (01:09 +0900)]
wasapi2sink: Fix ringbuffer object leak

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2315>

3 years agowasapi2ringbuffer: Implement GstAudioRingBuffer::pause()
Seungha Yang [Wed, 9 Jun 2021 15:24:24 +0000 (00:24 +0900)]
wasapi2ringbuffer: Implement GstAudioRingBuffer::pause()

WASAPI doesn't support PAUSE so it's not different from Stop().
When pipeline is in paused state, we don't need to waste CPU resource
for feeding silent buffers.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2315>

3 years agod3d11memory: Implement GstAllocator::mem_copy method
Seungha Yang [Sun, 6 Jun 2021 16:49:26 +0000 (01:49 +0900)]
d3d11memory: Implement GstAllocator::mem_copy method

There are a few places which require deep copy
(basesink on drain for example). Also this implementation can be
useful for future use case.
One probable future use case is that copying DPB texture to
another texture for in-place transform since our DPB texture is never
writable, and therefore copying is unavoidable.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2308>

3 years agowasapi2src: Add support for loopback recording
Seungha Yang [Tue, 8 Jun 2021 12:35:20 +0000 (21:35 +0900)]
wasapi2src: Add support for loopback recording

... and add various device error handling.

This loopback implementation is functionally identical to that of wasapisrc.
When it's enabled, wasapi2src will read data from render device instead of
capture device.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2311>

3 years agowasapi2: Rewrite plugin and implement audioringbuffer subclass
Seungha Yang [Mon, 10 May 2021 11:45:28 +0000 (20:45 +0900)]
wasapi2: Rewrite plugin and implement audioringbuffer subclass

... based on MediaFoundation work queue API.

By this commit, wasapi2 plugin will make use of pull mode scheduling
with audioringbuffer subclass.
There are several drawbacks of audiosrc/audiosink subclassing
(not audiobasesrc/audiobasesink) for WASAPI API, which are:
* audiosrc/audiosink classes try to set high priority to
  read/write thread via MMCSS (Multimedia Class Scheduler Service)
  but it's not allowed in case of UWP application.
  In order to use MMCSS in UWP, application should use MediaFoundation
  work queue indirectly.
  Since audiosrc/audiosink scheduling model is not compatible with
  MediaFoundation's work queue model, audioringbuffer subclassing
  is required.
* WASAPI capture device might report larger packet size than expected
  (i.e., larger frames we can read than expected frame size per period).
  Meanwhile, in any case, application should drain all packets at that moment.
  In order to handle the case, wasapi/wasapi2 plugins were making use of
  GstAdapter which is obviously sub-optimal because it requires additional
  memory allocation and copy.
  By implementing audioringbuffer subclassing, we can avoid such inefficiency.

In this commit, all the device read/write operations will be moved
to newly implemented wasapi2ringbuffer class and
existing wasapi2client class will take care of device enumeration
and activation parts only.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2306>

3 years agowasapi2: Use AUDCLNT_STREAMFLAGS_NOPERSIST flag
Seungha Yang [Sun, 6 Jun 2021 08:32:59 +0000 (17:32 +0900)]
wasapi2: Use AUDCLNT_STREAMFLAGS_NOPERSIST flag

... so that we can disable persistence of our mute/volume status

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2306>

3 years agowasapi2src: Fix doc typo
Seungha Yang [Sun, 6 Jun 2021 08:28:56 +0000 (17:28 +0900)]
wasapi2src: Fix doc typo

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2306>

3 years agowpe: Rename `undeserializable_type` to `not_deserializable_type`
Thibault Saunier [Wed, 26 May 2021 04:12:59 +0000 (00:12 -0400)]
wpe: Rename `undeserializable_type` to `not_deserializable_type`

Making it more readable.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2273>

3 years agowpe: Make forwarded messages layout more like GstBinForwaded messages
Thibault Saunier [Wed, 26 May 2021 03:58:27 +0000 (23:58 -0400)]
wpe: Make forwarded messages layout more like GstBinForwaded messages

Making it look more like how we do this kind of things in other places.

See: https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2252#note_927653
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2273>

3 years agowpe: Make wpesrc!video pad an always pad
Thibault Saunier [Fri, 21 May 2021 14:52:01 +0000 (10:52 -0400)]
wpe: Make wpesrc!video pad an always pad

There should always be a `video` pad no matter what.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2273>

3 years agowpe: Remove unused env var
Thibault Saunier [Fri, 21 May 2021 14:31:53 +0000 (10:31 -0400)]
wpe: Remove unused env var

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2273>

3 years agowpe: Fix atomic usage
Thibault Saunier [Fri, 21 May 2021 14:31:37 +0000 (10:31 -0400)]
wpe: Fix atomic usage

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2273>

3 years agowpe: Add a note able requiring tracing subsystem for message forwarding
Thibault Saunier [Fri, 21 May 2021 14:29:11 +0000 (10:29 -0400)]
wpe: Add a note able requiring tracing subsystem for message forwarding

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2273>

3 years agowpe: Fix check on whether MEMFD_CREATE is available
Thibault Saunier [Fri, 21 May 2021 14:18:21 +0000 (10:18 -0400)]
wpe: Fix check on whether MEMFD_CREATE is available

The ordering of the ifdef was wrong

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2273>

3 years agowpe: Plug a leak
Thibault Saunier [Fri, 21 May 2021 14:13:01 +0000 (10:13 -0400)]
wpe: Plug a leak

We were freeing after returning

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2273>

3 years agoRevert "wpe: Properly respect LIBGL_ALWAYS_SOFTWARE"
Thibault Saunier [Fri, 21 May 2021 13:54:33 +0000 (09:54 -0400)]
Revert "wpe: Properly respect LIBGL_ALWAYS_SOFTWARE"

This causes issues I didn't see:
     https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2252#note_927633

Let's just tell people to use capsfilter to force software rendering in
`wpesrc` for now.

The intent was to allow forcing it easily in playbin2 for the CI, but
we will do it some other way and see when time comes.

This reverts commit 9415106b029e5469ca28d882dc46ecc38786d4c9.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2273>

3 years agodebugutils: Introduce videocodectestsink
Nicolas Dufresne [Fri, 28 May 2021 19:18:53 +0000 (15:18 -0400)]
debugutils: Introduce videocodectestsink

This is a video specific sink used to test video CODEC conformance. This is similar
to a combination of filesink and testsink, but will skip over any type of
padding that GStreamer Video library introduces. This is needed in order to obtain the
correct checksum or raw yuv data.

This element currently support writing back non-padded raw I420 through the
location property and will calculate an MD5 and post it as an element message
of type conformance/checksum. More output format or checksum type could be
added in the future as needed.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2287>

3 years agovkinstance: Don't abort in case that system has no available vulkan device
Seungha Yang [Thu, 3 Jun 2021 16:44:47 +0000 (01:44 +0900)]
vkinstance: Don't abort in case that system has no available vulkan device

Specification doesn't have restriction that returned
pPhysicalDeviceCount value must be non-zero

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2304>

3 years agotsdemux: Clear all streams when rewinding
Edward Hervey [Thu, 3 Jun 2021 09:24:53 +0000 (11:24 +0200)]
tsdemux: Clear all streams when rewinding

This avoids sending out partial invalid data downstream which could cause
decoders (ex: `dvdlpmdec`) to error out.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2301>

3 years agod3d11: Add support for YV12 and NV21 formats
Seungha Yang [Fri, 28 May 2021 16:48:15 +0000 (01:48 +0900)]
d3d11: Add support for YV12 and NV21 formats

Handle UV swapped 4:2:0 8bits formats

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2290>

3 years agod3d11window_win32: Ensure closing internal HWND from window thread
Seungha Yang [Thu, 3 Jun 2021 09:28:26 +0000 (18:28 +0900)]
d3d11window_win32: Ensure closing internal HWND from window thread

Window handle must be closed from its own message thread

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2302>

3 years agodoc: Update cache after pixel format reorder
Nicolas Dufresne [Thu, 3 Jun 2021 14:31:39 +0000 (10:31 -0400)]
doc: Update cache after pixel format reorder

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2303>

3 years agodoc: add the msdk elements
Yinhang Liu [Thu, 3 Jun 2021 02:03:19 +0000 (10:03 +0800)]
doc: add the msdk elements

Supported elements:
msdkav1dec, msdkh264dec, msdkh264enc, msdkh265dec, msdkh265enc,
msdkmjpegdec, msdkmjpegenc, msdkmpeg2dec, msdkmpeg2enc, msdkvc1dec,
msdkvp8dec, msdkvp9dec, msdkvp9enc, msdkvpp.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2296>

3 years agowaylandsink: Fix for missing initial configure
Nicolas Dufresne [Wed, 2 Jun 2021 18:17:13 +0000 (14:17 -0400)]
waylandsink: Fix for missing initial configure

We were doing our initial "empty" commit on the subsurface instead of the
toplevel surface. As an incidence, we should not have received a configure
event ever, not just on mutter. This fixes the following warning when using
mutter compositor (aka gnome-shell):

  waylandsink wlwindow.c:304:gst_wl_window_new_toplevel: The compositor did not send configure event.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2299>

3 years agotest: camerabin: Fix buffer size calculation
Nicolas Dufresne [Wed, 2 Jun 2021 15:26:41 +0000 (11:26 -0400)]
test: camerabin: Fix buffer size calculation

We were assunming that GStreamer size for RGB (24bit packed) data was width x
height x 3, but GStreamer defaults to specific alignment. Use GstVideoInfo API
in order to obtain the buffer size.

This fixes failure seen when trying to merge: https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/998
which make us negoaite 1x1 instead of 16x16 in this test.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2297>

3 years agocodecs: Integrate H265 DPB full check into need_bump().
He Junyan [Mon, 31 May 2021 09:51:58 +0000 (17:51 +0800)]
codecs: Integrate H265 DPB full check into need_bump().

The current DPB check of H265 is not very correct. The current frame
is already in the DPB when we check whether the DPB is full.
For example, the DPB max size is 16 and we have 15 ref frames in the
DPB, so the gst_h265_dpb_delete_unused() cleans no one, and then plus
the current frame, the DPB is 16. This causes an error return, but in
fact, the stream is correct.
We now integrate the DPB full check into the need_bump() function.
We add the correct frame into to DPB and then check whether the picture
num is bigger than max_num_pics of DPB(which means there is no room for
the current picture). If true, we bump the DPB immediately.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2291>

3 years agoBack to development
Tim-Philipp Müller [Tue, 1 Jun 2021 14:28:57 +0000 (15:28 +0100)]
Back to development

3 years agoRelease 1.19.1
Tim-Philipp Müller [Mon, 31 May 2021 23:14:22 +0000 (00:14 +0100)]
Release 1.19.1

3 years agov4l2codecs: gstv4l2codecsvp8dec: implement a render delay
Daniel Almeida [Thu, 8 Apr 2021 13:11:52 +0000 (10:11 -0300)]
v4l2codecs: gstv4l2codecsvp8dec: implement a render delay

The v4l2 backend support delayed output for performance reasons.
It is then possible to use render delays to queue multiple requests
simultaneously, thus increasing performance.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2150>

3 years agocodecs: gstvp8decoder: add support for render delay
Daniel Almeida [Wed, 7 Apr 2021 21:24:27 +0000 (18:24 -0300)]
codecs: gstvp8decoder: add support for render delay

Some decoding APIs support delayed output for performance reasons.
One example would be to request decoding for multiple frames and
then query for the oldest frame in the output queue.

This also increases throughput for transcoding and improves seek
performance when supported by the underlying backend.

Introduce support in the vp8 base class, so that backends that
support render delays can actually implement it.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2150>

3 years agowebrtcbin: Stop transceivers update after first SDP error
Philippe Normand [Mon, 17 May 2021 09:49:41 +0000 (10:49 +0100)]
webrtcbin: Stop transceivers update after first SDP error

When invalid SDP is supplied, _update_transceiver_from_sdp_media() sets the
GError, so it is invalid to continue any further SDP processing, we have to exit
early when the first error is raised.

Fixes #1595

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2254>

3 years agod3d11: Suppress some warning logs
Seungha Yang [Fri, 28 May 2021 14:21:19 +0000 (23:21 +0900)]
d3d11: Suppress some warning logs

We uses gst_d3d11_device_new() for enumerating device which can
fail for some reason. Don't print warning log for the case.
And decoding capability check is the same case as well.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2286>

3 years agohlssink2: Initialize debug category to prevent an assert with `fatal-warnings`
Roman Sivriver [Fri, 28 May 2021 21:05:02 +0000 (17:05 -0400)]
hlssink2: Initialize debug category to prevent an assert with `fatal-warnings`

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2288>

3 years agod3d11compositor: Reuse converter on alpha update
Seungha Yang [Fri, 21 May 2021 11:02:53 +0000 (20:02 +0900)]
d3d11compositor: Reuse converter on alpha update

... instead of creating converter object

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2276>

3 years agod3d11converter: Introduce config to be extensible
Seungha Yang [Mon, 17 May 2021 16:24:29 +0000 (01:24 +0900)]
d3d11converter: Introduce config to be extensible

Add a config argument like that of GstVideoConverter so that
we can add more options without modifying existing methods

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2276>

3 years agod3d11shader: Don't hold state object in GstD3D11Quad
Seungha Yang [Fri, 21 May 2021 12:30:42 +0000 (21:30 +0900)]
d3d11shader: Don't hold state object in GstD3D11Quad

We might want to update state object

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2276>

3 years agov4lcodecs: Validate src formats
Nicolas Dufresne [Thu, 27 May 2021 20:22:42 +0000 (16:22 -0400)]
v4lcodecs: Validate src formats

This add src format validation, this avoid registering element for
drivers we don't support any of their src formats. This also special
case the AlphaDecodeBin wrapper, as we know that alphacombine element
only support I420 and NV12 for now.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2272>

3 years agov4l2codecs: add wrappers for alpha decode
Daniel Almeida [Sat, 22 May 2021 19:29:09 +0000 (16:29 -0300)]
v4l2codecs: add wrappers for alpha decode

codecalpha is a new plugin introduced to support VP8/VP9 alpha as
defined in the WebM and Matroska specifications. It splits the stream
into two streams, one for the alpha and one for the actual content,
then it decodes them separately with vpxdec and finally combine the
results as A420 or AV12 (i.e. YUV + an extra alpha plane).

The workflow above is setup by means of a bin, gstcodecalphabin.

This patch simulates the same workflow into the v4l2codecs namespace,
thus using the new v4l2 stateless decoders for hardware acceleration.

This is so we can register the new alpha decode elements only if the
hardware produces formats we support, i.e. I420 or NV12 for now.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2272>

3 years agocodecalpha: alphacombine: add support for NV12/AV12
Daniel Almeida [Wed, 19 May 2021 21:45:19 +0000 (18:45 -0300)]
codecalpha: alphacombine: add support for NV12/AV12

Alpha combine works by appending the GstMemory for the alpha channel
to the GstBuffer containing I420, thereby pushing A420 on its src pad.

Add support for the same workflow for NV12, thereby producing the
recently introduced AV12 format (NV12 + Alpha).

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2277>

3 years agointerlace: Don't set field-order field for progressive caps
Seungha Yang [Tue, 25 May 2021 11:21:34 +0000 (20:21 +0900)]
interlace: Don't set field-order field for progressive caps

That would cause negotiation issue

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2282>

3 years agointerlace: Drop framerate from query caps of sinkpad
Seungha Yang [Tue, 25 May 2021 10:47:28 +0000 (19:47 +0900)]
interlace: Drop framerate from query caps of sinkpad

Query caps should return caps which represent the element can accept,
not resulting format.

Fixing negotiation error with
gst-launch-1.0 videotestsrc ! video/x-raw,framerate=25/1 ! interlace field-pattern=0 ! fakesink

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2282>

3 years agod3d11compositor: Fix caps update handling
Seungha Yang [Wed, 26 May 2021 07:37:06 +0000 (16:37 +0900)]
d3d11compositor: Fix caps update handling

New caps is applied only when previous buffer is consumed if any.
So, the lastest given caps might not be corresponding to the current buffer
to be handled.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2284>

3 years agolibs: va: display: Handle auto clean up macros.
Víctor Manuel Jáquez Leal [Thu, 20 May 2021 11:47:11 +0000 (13:47 +0200)]
libs: va: display: Handle auto clean up macros.

Add G_DEFINE_AUTOPTR_CLEANUP_FUNC macro for display classes, so auto
pointers are possible to users.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2269>

3 years agova: h265dec: Add current picture into reference list for SCC.
He Junyan [Mon, 17 May 2021 11:06:34 +0000 (19:06 +0800)]
va: h265dec: Add current picture into reference list for SCC.

The current picture is not in the DPB, so we need to add it manually
to the reference list when SCC is enabled.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2255>

3 years agova: h265dec: Set Screen Content extension (SCC) for picture parameters.
He Junyan [Fri, 21 May 2021 15:47:14 +0000 (23:47 +0800)]
va: h265dec: Set Screen Content extension (SCC) for picture parameters.

We already declare the support of HEVC screen content extension profiles
in the profile mapping list, but we fail to generate the correct VA picture
parameters buffers. This may cause the GPU hang.

We need to fill the buffer of VAPictureParameterBufferHEVCExtension correctly.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2255>

3 years agova: h265dec: Use get_profile_from_sps() to recognize the profile.
He Junyan [Mon, 17 May 2021 09:47:07 +0000 (17:47 +0800)]
va: h265dec: Use get_profile_from_sps() to recognize the profile.

The function of gst_h265_get_profile_from_sps() is better than the
function gst_h265_profile_tier_level_get_profile() when we recognize
the profile of the stream, becaue it considers the compatibility.

It is also used by h265parse to recognize the profile. So it is
better to keep the same behaviour with the parse and other decoders.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2255>

3 years agova: h265dec: Set range extension for picture and slice parameters.
He Junyan [Fri, 21 May 2021 15:21:12 +0000 (23:21 +0800)]
va: h265dec: Set range extension for picture and slice parameters.

We already declare the support of HEVC range extension profiles in
the profile mapping list, but we fail to generate the correct VA
picture and slice parameters buffers. This may cause the GPU hang.

We need to fill the buffer of VAPictureParameterBufferHEVCExtension
and VASliceParameterBufferHEVCExtension correctly.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2255>

3 years agova: h265dec: Use picture and slide extension parameters.
Víctor Manuel Jáquez Leal [Mon, 24 May 2021 16:18:52 +0000 (18:18 +0200)]
va: h265dec: Use picture and slide extension parameters.

This is transitional commit to later implement extended and screen
profiles.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2255>

3 years agoUse gst_buffer_new_memdup()
Tim-Philipp Müller [Mon, 24 May 2021 17:34:25 +0000 (18:34 +0100)]
Use gst_buffer_new_memdup()

Update for function rename in core.

https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/827

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2281>

3 years agowpe: Bump wpebackend-fdo version requirement to 1.8
Philippe Normand [Sat, 22 May 2021 17:05:18 +0000 (18:05 +0100)]
wpe: Bump wpebackend-fdo version requirement to 1.8

Debian bullseye has this version already, and this allows us to get rid of many
ifdefs. The mouse scroll handling is actually functional now as well.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2278>

3 years agoUse new gst_buffer_new_copy()
Tim-Philipp Müller [Sun, 23 May 2021 15:10:53 +0000 (16:10 +0100)]
Use new gst_buffer_new_copy()

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2279>

3 years agodoc: update gst_plugins_cache.json
Daniel Almeida [Fri, 21 May 2021 18:18:21 +0000 (15:18 -0300)]
doc: update gst_plugins_cache.json

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1152>

3 years agodxgiscreencapsrc: renegotiate caps on resolution change
Jakub Adam [Fri, 14 May 2021 18:22:26 +0000 (20:22 +0200)]
dxgiscreencapsrc: renegotiate caps on resolution change

When desktop gets resized, recreate the textures and renegotiate the
source caps with the updated video dimensions.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2249>

3 years agova: h265dec: Set LastSliceOfPic for multi sliced frames.
He Junyan [Tue, 18 May 2021 06:09:01 +0000 (14:09 +0800)]
va: h265dec: Set LastSliceOfPic for multi sliced frames.

VA-API HEVC decoding needs to known which is the last slice of a
picture, but slices are processed sequencially, so we know the
last slice until all the slices are already pushed into the
VABuffer array.

In order to mark the last slice, they are pushed into the
VABuffer array with a delay of one slice: the first slice is
hold, and when the second slice come, the first one is pushed
while holding the second, and so on. Finally, at end_picture(),
the last slice is marked and pushed into the array.

Co-author: Victor Jaquez <vjaquez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2246>

3 years agod3d11desktopdupsrc: Add support for desktop size/rotation mode change
Seungha Yang [Thu, 20 May 2021 08:03:15 +0000 (17:03 +0900)]
d3d11desktopdupsrc: Add support for desktop size/rotation mode change

Re-negotiates with updated size on desktop size
(i.e., resolution, scaling factor), and rotation mode change

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2268>

3 years agompegtsmux: Fixup program array indices after stream removal
Jan Alexander Steffens (heftig) [Thu, 20 May 2021 08:09:57 +0000 (10:09 +0200)]
mpegtsmux: Fixup program array indices after stream removal

Each stream stores the `program_array_index` of its position in its
program's `streams` array. When we remove a stream from this array, we
need to correct the `program_array_index` of all streams that were
backshifted by the removal.

Also extract the removal into a new function and add some more safety
checks.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2266>

3 years agod3d11memory: Protect map and unmap with device lock
Seungha Yang [Thu, 20 May 2021 09:49:01 +0000 (18:49 +0900)]
d3d11memory: Protect map and unmap with device lock

We should lock memory object with gst_d3d11_device_lock() first
then GST_D3D11_MEMORY_LOCK() need to be used.

One observed deadlock case is that:
- Thread A takes d3d11 device lock
- At the same time, Thread B tries CPU map to d3d11memory which requires
  d3d11 device lock as well, but it's already taken by Thread A.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2267>

3 years agod3d11memory: Add trace log for debugging locking thread
Seungha Yang [Thu, 20 May 2021 09:38:17 +0000 (18:38 +0900)]
d3d11memory: Add trace log for debugging locking thread

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2267>

3 years agoaudiolatency: Drop incoming downstream stick events
Seungha Yang [Thu, 20 May 2021 06:39:39 +0000 (15:39 +0900)]
audiolatency: Drop incoming downstream stick events

stream-start, caps, and segment events will be pushed by internal
audiotestsrc element.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2265>

3 years agoaudiolatency: Use live mode audiotestsrc
Seungha Yang [Thu, 20 May 2021 06:28:13 +0000 (15:28 +0900)]
audiolatency: Use live mode audiotestsrc

Expected use case of audiolatency element is that mimic audio capture
device which is most likely live source. So audiolatency element
should use live mode as well.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2265>

3 years agowpe: Bump WPE dependency to 2.28
Thibault Saunier [Wed, 19 May 2021 22:48:29 +0000 (18:48 -0400)]
wpe: Bump WPE dependency to 2.28

The new audio feature depends on WPE 2.28 so we should just bump our
requirement to that.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2264>

3 years agod3d11compositor: Fix missing D3D11 prefix
Seungha Yang [Wed, 19 May 2021 15:51:08 +0000 (00:51 +0900)]
d3d11compositor: Fix missing D3D11 prefix

Fix typo, no functional change

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2262>

3 years agowpe: Update doc cache
Thibault Saunier [Tue, 18 May 2021 21:49:23 +0000 (17:49 -0400)]
wpe: Update doc cache

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2252>

3 years agowpe: Properly respect LIBGL_ALWAYS_SOFTWARE
Thibault Saunier [Thu, 22 Apr 2021 03:14:13 +0000 (23:14 -0400)]
wpe: Properly respect LIBGL_ALWAYS_SOFTWARE

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2252>

3 years agowpe: Relay messages from WPE internal pipelines
Thibault Saunier [Sun, 2 May 2021 01:48:23 +0000 (21:48 -0400)]
wpe: Relay messages from WPE internal pipelines

It is based on a tracer as it allows us to very easily get
every message that are posted on any bus inside the process.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2252>

3 years agowpe: Base wpe audio implementation on a web extension
Thibault Saunier [Tue, 20 Apr 2021 00:46:46 +0000 (20:46 -0400)]
wpe: Base wpe audio implementation on a web extension

This makes the implementation simpler and enable us to map
webviews and audio stream much more easily

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2252>

3 years agowpe: Enable WebAudio
Philippe Normand [Sun, 8 Dec 2019 13:16:38 +0000 (13:16 +0000)]
wpe: Enable WebAudio

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2252>

3 years agowpe: Implement audio support
Philippe Normand [Sun, 8 Dec 2019 11:49:20 +0000 (11:49 +0000)]
wpe: Implement audio support

The wpesrc bin now exposes "sometimes" audio src pads, one for every PCM audio
stream created by WPEWebKit.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2252>

3 years agowpe: Move wpesrc to wpevideosrc and add a wrapper bin `wpesrc`
Thibault Saunier [Wed, 10 Mar 2021 20:27:52 +0000 (17:27 -0300)]
wpe: Move wpesrc to wpevideosrc and add a wrapper bin `wpesrc`

Currently the bin contains a single element but we are going
to implement audio support and expose extra pads for audio

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2252>

3 years agosctp: Ensure pad is still a child of element before removal
Doug Nazar [Tue, 18 May 2021 04:43:23 +0000 (00:43 -0400)]
sctp: Ensure pad is still a child of element before removal

During pipeline shutdown there are several competing paths to remove
pads. Avoids tests failing due to:

Unexpected critical/warning: Padname '':sink_1 does not belong to element sctpenc1 when removing

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2256>

3 years agosctp: Fix race of pad removal during reset/stop
Doug Nazar [Mon, 17 May 2021 13:13:28 +0000 (09:13 -0400)]
sctp: Fix race of pad removal during reset/stop

Both reset & stop remove existing pads. Can result in warning from
gst_element_remove_pad().

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2256>

3 years agowebrtcbin: Fix race bringing up sctp data channel
Doug Nazar [Mon, 17 May 2021 13:11:54 +0000 (09:11 -0400)]
webrtcbin: Fix race bringing up sctp data channel

Notifying before pads are linked can cause the stream to fail to start.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2256>

3 years agowebrtcbin: advertise harder the rtcp-mux-only requirement
Matthew Waters [Thu, 13 May 2021 11:11:30 +0000 (21:11 +1000)]
webrtcbin: advertise harder the rtcp-mux-only requirement

And ignore rtcp ICE candidates

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2239>

3 years agowebrtcbin: update default jb latency docs
Sid Sethupathi [Fri, 14 May 2021 15:47:05 +0000 (10:47 -0500)]
webrtcbin: update default jb latency docs

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2242>

3 years agodtls: Let sender know when we are flushing
Doug Nazar [Tue, 18 May 2021 20:38:04 +0000 (16:38 -0400)]
dtls: Let sender know when we are flushing

Prevents endless loop during shutdown where we end up sending 0 bytes.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2229>

3 years agodtls: Add ability to set custom GstFlowReturn on callback error
Doug Nazar [Tue, 18 May 2021 20:31:47 +0000 (16:31 -0400)]
dtls: Add ability to set custom GstFlowReturn on callback error

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2229>

3 years agowebrtc: Remove reundundant context object in transportsendbin
Olivier Crête [Wed, 19 May 2021 00:26:38 +0000 (20:26 -0400)]
webrtc: Remove reundundant context object in transportsendbin

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2260>

3 years agowebrtc: Wait until ICE is connected to start DTLS handshake process
Olivier Crête [Wed, 19 May 2021 00:18:28 +0000 (20:18 -0400)]
webrtc: Wait until ICE is connected to start DTLS handshake process

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2260>

3 years agowebrtcbin: Remove pad probe on nicesink
Olivier Crête [Tue, 18 May 2021 22:29:16 +0000 (18:29 -0400)]
webrtcbin: Remove pad probe on nicesink

This pad probe is no longer necessary, libnice now drops
all buffers before the stream is connected. This pad problem
also caused deadlocks in some situations.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2260>

3 years agokate: Initialize debug categories
Olivier Crête [Tue, 18 May 2021 00:59:19 +0000 (20:59 -0400)]
kate: Initialize debug categories

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2258>

3 years agolibs: va: Documentation and annotations.
Víctor Manuel Jáquez Leal [Thu, 13 May 2021 08:27:49 +0000 (10:27 +0200)]
libs: va: Documentation and annotations.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2196>

3 years agoexamples: va: Update the VA examples because of the new va lib.
He Junyan [Fri, 7 May 2021 09:05:38 +0000 (17:05 +0800)]
examples: va: Update the VA examples because of the new va lib.

Because we introduce the new va lib, the va examples need to include
new header files and add more library linkage.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2196>

3 years agolibs: va: display_wrapper: Use gpointer for VADisplay.
Víctor Manuel Jáquez Leal [Thu, 13 May 2021 16:46:21 +0000 (18:46 +0200)]
libs: va: display_wrapper: Use gpointer for VADisplay.

In order to be coherent along all the implementation.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2196>

3 years agolibs: va: Move the VA common logic as a lib.
He Junyan [Thu, 6 May 2021 10:23:23 +0000 (18:23 +0800)]
libs: va: Move the VA common logic as a lib.

The VA acceleration now has more usages in linux-like platforms,
such as the MSDK. The different plugins based on the VA acceleration
need to share some common logic and types. We now move the display
related functions and types into a common va lib.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2196>

3 years agomsdk: add profile main-still-picture for hevc encoder
mkba [Mon, 17 May 2021 03:42:07 +0000 (11:42 +0800)]
msdk: add profile main-still-picture for hevc encoder

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2253>

3 years agointerlace: Fix too small buffer size error
Seungha Yang [Fri, 14 May 2021 15:39:57 +0000 (00:39 +0900)]
interlace: Fix too small buffer size error

Even though input/output resolutions are identical there, default
buffer size of progressive and interleaved formats could be different
because we are rounding up height of all plane of interlaced frame
to be multiple of two.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2244>

3 years agomsdk: use MFXJoinSession() to join the parent and child sessions
Haihao Xiang [Mon, 1 Mar 2021 04:09:43 +0000 (12:09 +0800)]
msdk: use MFXJoinSession() to join the parent and child sessions

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1503>

3 years agomsdk: use a new method to create mfx session when using oneVPL dispatcher
Haihao Xiang [Thu, 4 Feb 2021 07:27:13 +0000 (15:27 +0800)]
msdk: use a new method to create mfx session when using oneVPL dispatcher

In oneVPL, MFXLoad() and MFXCreateSession() are required to create a
workable mfx session[1]

[1] https://spec.oneapi.com/versions/latest/elements/oneVPL/source/programming_guide/VPL_prg_session.html#onevpl-dispatcher

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1503>

3 years agomsdk: allow user build this plugin against MFX version 2.2+ (oneVPL)
Haihao Xiang [Thu, 18 Feb 2021 05:38:25 +0000 (13:38 +0800)]
msdk: allow user build this plugin against MFX version 2.2+ (oneVPL)

Intel oneVPL SDK (oneVPL) is a successor to Intel Media SDK (MSDK)[1].
User may use -Dmfx_api=MSDK or -Dmfx_api=oneVPL to specify the required
SDK when building this plugin. If the SDK is not specified, meson will
try MSDK firstly, then oneVPL if MSDK is not available

Version 2.2+ is required in this patch because pkg-config file was not
provided officially before version 2.2

[1]https://spec.oneapi.com/versions/latest/elements/oneVPL/source/appendix/VPL_intel_media_sdk.html

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1503>

3 years agomsdkvp9dec: do not include mfxvp9.h
Haihao Xiang [Fri, 14 May 2021 03:56:49 +0000 (11:56 +0800)]
msdkvp9dec: do not include mfxvp9.h

The VP9 related definitions in mfxvp9.h are available under the
condition of 'MFX_VERSION >= MFX_VERSION_NEXT', which implies that these
definitions are never used in a public release.

This is in preparation for oneVPL support because mfxvp9.h was
removed from oneVPL

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1503>

3 years agomsdk: don't load user plugins for MFX version 2.0+
Haihao Xiang [Tue, 4 Aug 2020 04:53:35 +0000 (12:53 +0800)]
msdk: don't load user plugins for MFX version 2.0+

MFX version 2.0+ no longer supports user plugins, please refer to the
links for details

https://spec.oneapi.com/versions/latest/elements/oneVPL/source/appendix/VPL_intel_media_sdk.html#msdk-full-name-feature-removals
https://github.com/oneapi-src/oneVPL

This is in preparation for oneVPL support

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1503>

3 years agomsdk: exclude the audio code for MFX version 2.0+
Haihao Xiang [Tue, 4 Aug 2020 04:55:35 +0000 (12:55 +0800)]
msdk: exclude the audio code for MFX version 2.0+

MFX version 2.0+ no longer supports audio functions, please refer to the
links below for details

https://spec.oneapi.com/versions/latest/elements/oneVPL/source/appendix/VPL_intel_media_sdk.html#msdk-full-name-feature-removals
https://github.com/oneapi-src/oneVPL

This is in preparation for oneVPL support

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1503>

3 years agoalphacombine: Ignore all events coming from the alpha_pad
Nicolas Dufresne [Fri, 14 May 2021 18:08:17 +0000 (14:08 -0400)]
alphacombine: Ignore all events coming from the alpha_pad

As per usage of this element, everything from this pad is a
duplicate. Instead of implemented needless aggregation, simply
drop all events from this pad and let the one from the main stream
passthrough. Also stop proxying some queries from the alpha pad_too.

This fixes racy test failure:
- validate.file.playback.scrub_forward_seeking.opus_vp9-alpha_webm

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2247>

3 years agocodecalphademux: Do not set a GstFlowReturn from a boolean
Nicolas Dufresne [Fri, 14 May 2021 18:05:59 +0000 (14:05 -0400)]
codecalphademux: Do not set a GstFlowReturn from a boolean

This was a small overlook, gst_pad_send_event() returns a boolean,
so setting it into ret could confuse the flow combiner. Though,
it didn't bug, since both 0 and 1 are success (though 1 being
undefined).

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2247>

3 years agocodecalphademux: Remove eos flow return workaround
Nicolas Dufresne [Fri, 14 May 2021 18:04:00 +0000 (14:04 -0400)]
codecalphademux: Remove eos flow return workaround

It turns out that downstream returning OK after EOS is a bug in
multiqueue. As we moved to queue, we no longer have this issue.
Let's keep the code clean and just assuming that downstream will
keep returning EOS and allow convergence of flow.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2247>

3 years agoopenh264: Don't use GOnce for ABI check
Olivier Crête [Thu, 13 May 2021 19:18:34 +0000 (15:18 -0400)]
openh264: Don't use GOnce for ABI check

It turns out the value used for g_once_* APIs can't be
zero. And this is a very cheap check, so let's just do it every time.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2240>