Commit graph

90 commits

Author SHA1 Message Date
Xu Guangxin
610e477565 msdkdec: hold a reference for the surfaces locked by msdk
previous code releases GstBuffer too earlier. so we will see

ERROR                default gstmsdkvideomemory.c:77:gst_msdk_video_allocator_get_surface: failed to get surface available
ERROR         msdkbufferpool gstmsdkbufferpool.c:270:gst_msdk_buffer_pool_alloc_buffer:<msdkbufferpool0> failed to create new MSDK memory

We need to hold GstBuffer reference for msdk if the surfaced locked by msdk.

step to reproduce.
1. ffmpeg -f lavfi -i testsrc=duration=10:size=320x240:rate=30 -pix_fmt yuv420p -c:v libx265 test.265
2. GST_GL_PLATFORM=egl  gst-launch-1.0 -v filesrc location=test.265  ! h265parse ! msdkh265dec  ! queue ! glimagesink

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1277>
2020-06-15 02:46:53 +00:00
Jordan Petridis
87fc038f67
gstmsdkdec: fix logical operation that misses parenthesis
in C, & is weaker than the ! operator and clang is giving the following
error about it.

```
../subprojects/gst-plugins-bad/sys/msdk/gstmsdkdec.c:731:7: error: logical not is only applied to the left hand side of this bitwise operator [-Werror,-Wlogical-not-parentheses]
  if (!gst_msdk_context_get_job_type (thiz->context) & GST_MSDK_JOB_DECODER) {
      ^                                              ~
../subprojects/gst-plugins-bad/sys/msdk/gstmsdkdec.c:731:7: note: add parentheses after the '!' to evaluate the bitwise operator first
  if (!gst_msdk_context_get_job_type (thiz->context) & GST_MSDK_JOB_DECODER) {
      ^
       (                                                                   )
../subprojects/gst-plugins-bad/sys/msdk/gstmsdkdec.c:731:7: note: add parentheses around left hand side expression to silence this warning
  if (!gst_msdk_context_get_job_type (thiz->context) & GST_MSDK_JOB_DECODER) {
      ^
      (                                             )
1 error generated.
```
2020-04-10 18:22:21 +03:00
Seungha Yang
5d160a8968 msdkdec: Fix GstMsdkContext leak 2020-01-29 15:35:21 +09:00
Seungha Yang
a8fff639c8 msdk: Clear reference counted object in dispose() method
Follow GObject's memory management model
2020-01-29 15:35:21 +09:00
Nirbheek Chauhan
1bcf44bb1f msdk: Fix increasing memory usage in dynamic pipelines
Our context is non-persistent, and we propagate it throughout the
pipeline. This means that if we try to reuse any gstmsdk element by
removing it from the pipeline and then re-adding it, we'll clone the
mfxSession and create a new gstmsdk context as a child of the old one
inside `gst_msdk_context_new_with_parent()`.

Normally this only allocates a few KB inside the driver, but on
Windows it seems to allocate tens of MBs which leads to linearly
increasing memory usage for each PLAYING->NULL->PLAYING state cycle
for the process. The contexts will only be freed when the pipeline
itself goes to `NULL`, which would defeat the purpose of dynamic
pipelines.

Essentially, we need to optimize the case in which the element is
removed from the pipeline and re-added and the same context is re-set
on it. To detect that case, we set the context on `old_context`, and
compare it to the new one when preparing the context. If they're the
same, we don't need to do anything.

Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/issues/946
2020-01-21 00:38:41 +00:00
Nirbheek Chauhan
6834a12112 msdk: Reorganize context preparation code
Split it out into a separate function with early exits to make the
flow clearer, and document what the function is doing clearly.
No functional changes.
2020-01-21 00:38:41 +00:00
Nirbheek Chauhan
c0d778c2c7 msdk: Use gst_clear_object()
`gst_object_replace()` is not supposed to be used for unreffing and
NULLing objects.
2020-01-21 00:38:41 +00:00
Josep Torra
ef37446460 msdkdec: align frame list using decoded timestamp
Before this change decoder used the oldest frame in the list to pair it
with the decoded surface. This only works when there's a perfect stream
like HEADERS,SYNCPOINT,DELTA...

When playing RTSP streams we can get imperfect streams like HEADERS,
DELTA,SYNCPOINT,DELTA... In this case decoder drops the frames
between HEADERS and SYNCPOINT which leads into using wrong PTS on
the output frames.

With this change we inject the input PTS into the bitstream and use it
to align the internal frame list with the actually decoded position.

Fixes playback with:
```
gst-launch-1.0 rtspsrc location=... latency=0 drop-on-latency=1 ! ...
```
2020-01-19 04:26:15 +00:00
Haihao Xiang
52abbeeefb msdkdec: support transform_meta virtual method
Allow the base class to copy GstVideoRegionOfInterestMeta data to the
output buffer when calling gst_video_decoder_finish_frame
2020-01-15 00:47:12 +00:00
Haihao Xiang
bdf9a76226 Revert "msdkdec: make sure to use video memory on Linux"
Commit a1584b6 caused big performance drop if the downstream element
is not a msdk element because it is very slow to read data from video
memory directly.

This reverts commit a1584b6f99.
2020-01-10 01:36:40 +00:00
Julien Isorce
4e6a1b9634 msdkdec: trigger renegotiation if video info changes
Useful when framerate changes. Previously it was only checking
for resolution change but renego should happen if any video
info changes.
2020-01-03 10:19:57 -08:00
Julien Isorce
a1584b6f99 msdkdec: make sure to use video memory on Linux
The block that sets use_video_memory flag is after the
the condition `if gst_msdk_context_prepare` but it
always returns false when there is no other msdk elements.
So the decoder ends up with use_video_memory as FALSE.
Note that msdkvpp always set use_video_memory as TRUE.

When use_video_memory is FALSE then the msdkdec allocates
the output frames with posix_memalign (see gstmsdksystemmemory.c).
The result is then copied back to the GstVideoPool's buffers
(or to the downstream pool's buffers if any).
When use_video_memory is TRUE then the msdkdec uses vaCreateSurfaces
to create vaapi surfaces for the hw decoder to decode into
(see gstmsdkvideomemory.c). The result is then copied to either
the internal GstVideoPool and to the downstream pool if any.
(vaDeriveImage/vaMapBuffer is used in order to read the surfaces)
2020-01-03 07:08:23 +00:00
Seungha Yang
b88af2e4fe msdkdec: Correct return value of GstVideoDecoder::flush()
Use boolean instead of GstFlowReturn as declared.
Note that since base class does not check return value of GstVideoDecoder::flush(),
this would not cause any change of behavior.
2020-01-02 12:22:07 +09:00
Haihao Xiang
a819f05851 msdkdec: free unlocked msdk surface before output buffer allocation
https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/merge_requests/924
is trying to use video memory for decoding on Linux, which reveals a
hidden bug in msdkdec.

For video memory, it is possible that a locked mfx surface is not used
indeed and it will be un-locked later in MSDK, so we have to check the
associated MSDK surface to find out and free un-used surfaces, otherwise
it is easy to exhaust all pre-allocated mfx surfaces and get errors below:

0:00:00.777324879 27290 0x564b65a510a0 ERROR                default
gstmsdkvideomemory.c:77:gst_msdk_video_allocator_get_surface: failed to
get surface available
0:00:00.777429079 27290 0x564b65a510a0 ERROR         msdkbufferpool
gstmsdkbufferpool.c:260:gst_msdk_buffer_pool_alloc_buffer:<msdkbufferpool0>
failed to create new MSDK memory

Note the sample code in MSDK does similar thing in
CBuffering::SyncFrameSurfaces()
2020-01-02 00:43:36 +00:00
Julien Isorce
93bc74284f msdkdec: call finalize on the parent class
Otherwise GstVideoDecoder is not finalized and
resources are leaked.

Somehow GST_TRACERS="leaks" GST_DEBUG="GST_TRACER:7" did not catch it.

Valgrind output:

==31645== 22,480 (1,400 direct, 21,080 indirect) bytes in 5 blocks are definitely lost in loss record 5,042 of 5,049
==31645==    at 0x4C2FB0F: malloc
==31645==    by 0x51D9E88: g_malloc
==31645==    by 0x51FA7B5: g_slice_alloc
==31645==    by 0x51FAC68: g_slice_alloc0
==31645==    by 0x58D9984: g_type_create_instance
==31645==    by 0x58BA344: g_object_new_with_properties
==31645==    by 0x58BADA0: g_object_new
==31645==    by 0x8ECA966: gst_video_decoder_init
==31645==    by 0x58D99E7: g_type_create_instance
==31645==    by 0x58BA344: g_object_new_with_properties
2019-12-13 15:36:26 -08:00
Stéphane Cerveau
6bc0e9527e remove various useless linefeed in logs 2019-12-11 10:51:29 +01:00
Julien Isorce
3bae199863 msdkdec: fix assertion 'frame->ref_count > 0' failed
Can happen if the oldest frame is the current frame
and if gst_video_decoder_finish_frame failed in which
case the current is unref and then drop instead of
just drop.
This patch also removes some assumptions, it was strange
to call unref and finish_frame in gst_msdkdec_finish_task.
In principle when owning a frame, the code should either
unref, or drop or finish.
2019-12-09 16:10:51 +00:00
Julien Isorce
2af672d00a msdkdec: log an error if leaking surfaces in finalize
The for loop in gst_msdkdec_handle_frame is error prone
about how it manages surfaces. Because sometimes it sets
the surface variable to NULL and sometimes it needs to free
it right away. So better to print an error if surfaces are
leaked to help with any change around the loop.
2019-11-25 17:14:20 +00:00
Julien Isorce
a7dc407041 msdkdec: fix surface leak in msdkdec_handle_frame
Can be reproduced with:
  videotestsrc ! x264enc key-int-max=$N ! \
  h264parse ! msdkh264dec ! fakesink sync=1

It happens with any gop size but the smaller is the distance N
between key frames, the quicker it is leaking.

Fixes #1023
2019-11-19 09:14:51 -08:00
Víctor Manuel Jáquez Leal
10d4c0c511 msdkdec: use decoder base class packetized flag
Instead of using a proxy of `is_packetized` flag this patch
replaces it with the accessor to that flag in decoder base class,
avoiding probable mismatches.
2019-10-29 14:58:54 +00:00
Haihao Xiang
cd883427e9 msdkdec: add parse callback for non-packetized input
commit 55c0d720 added the capability to handle non-packetized bitstream,
and there is a loop to handle multiple frames in a non-packetized buffer
in gst_msdkdec_handle_frame. However it is possible that a
non-packetized buffer still contains valid data but there is no long any
pending unfinished frame. Currently gst_video_decoder_decode_frame is
invoked to send a new frame with new input data, the situaltion is
repeated till an EOS is received. An application has to exit when
receiving an EOS, however there is still valid data in a
non-packetezied input buffer, hence some frames are dropped.

This fix adds a parse callback for non-packeteized input, a new frame
will be sent to the subclass as soon as the input buffer has valid data

This fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/issues/665
2019-10-29 14:58:54 +00:00
Haihao Xiang
71ff8236b5 msdkdec: set lockable flag
When postpone_free_surface is TRUE, the output buffer is not writable,
however the base decoder needs a writable buffer as output buffer,
otherwise it will make a copy of the output buffer. As the underlying
memory is always lockable, so we may set the LOCKABLE flag for this buffer
to avoid buffer copy in the base class.

The refcount of the output buffer is 1 when postpone_free_surface is
FALSE, so needn't set the LOCKABLE flag for this case.
2019-09-30 18:15:38 +00:00
Haihao Xiang
f840e304f3 msdkdec: postpone surface free for VC1
For a skipped frame in VC1, MSDK returns the mfx surface of the reference
frame, so we have to make sure the corresponding surface for the
reference frame is not freed. In this fix, we postpone surface free because
we don't know whether a surface is referenced

Before this fix, the error is like as below:

New clock: GstSystemClock
0:00:00.181793130 23098 0x55f8a9d622d0 ERROR                msdkdec
gstmsdkdec.c:622:gst_msdkdec_finish_task:<msdkvc1dec0> Couldn't find the
cached MSDK surface

Sample pipeline:
gst-launch-1.0 filesrc location=input_has_skipped_frame.wmv ! asfdemux !
vc1parse ! msdkvc1dec ! glimagesink
2019-09-21 18:05:42 +00:00
Haihao Xiang
c689c94458 msdkdec: release the surface if this surface is not in use
If the surface is not in use, we may release it even if GST_FLOW_OK is going
to be returned, which may avoid the issue of failing to get surface
available

This fixes the regression caused by commit c05acf4
2019-09-21 17:34:23 +00:00
Haihao Xiang
c05acf40a8 msdkdec: set pointer to NULL to avoid freeing an used surface
The surface should be cached in the surface list when GST_FLOW_OK is
going to be returned

This fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/issues/1051
2019-09-13 17:32:19 +00:00
Haihao Xiang
2d4b0a5649 msdkdec: unref the caps to avoid memory leaks 2019-09-13 16:54:21 +00:00
Seungha Yang
54c5ebbd53 msdk: Remove all DMABuf caps features on Windows
Add new macro for sink/src pad template to ensure no DMABuf caps
features are exposed on Windows. Some DMABuf caps features
were not handled by the commit 9ec62418c3
2019-09-10 13:29:11 +00:00
Seungha Yang
eb3a117548 msdkdec: Do not use video memory on Windows
Like msdkenc, do not use video memory by default on Windows.
2019-09-10 13:29:11 +00:00
Haihao Xiang
18728310c9 msdkdec: output the decoded frame immediately if decoded order is required
DecodedOrder was deprecated in msdk-2017 version, but some customers
still use this for low-latency streaming of non-b-frame encoded streams,
which needs to output the frame at once
2019-09-04 05:25:13 +00:00
Haihao Xiang
4841c2759d msdk: avoid reading data from freed memory
Both g_list_delete_link and g_list_remove remove an element and free it,
so l->next is invalid (catched by valgrind) after calling g_list_delete_link
or g_list_remove
2019-08-28 14:51:54 +00:00
Yan Wang
f92b5614db msdkdec: Fix buffer allocation based on frame.
gst_msdkdec_finish_task() may release all frames in
GstVideoDecoder object. In this case, allocate_output_buffer()
cannot get the oldest frame to allocate buffer.
So gst_msdkdec_handle_frame() should return GST_FLOW_OK for
letting gst_video_decoder_decode_frame() to send a new frame
for decoding.

Fixes #664.
Fixes #665.
2019-08-26 13:59:16 +00:00
Fuwei Tang
df59db988b msdkdec: fix decoding issue for interlaced streams
Currently h264parser produces a field or a frame for
alignment=au for interlaced streams, but the flag
MFX_BITSTREAM_COMPLETE_FRAME needs a complete frame
or complementary field pair of data, this results in
broken images being output.

Some patches have been sent out to fix h264parser,
but they are pending on some unfinished work. In
order to make gstreamer-msdk decoding work properly
for interlaced streams before h264parser is fixed,
this flag will be removed temporarily and will be
added back once h264parser if fixed.

Related to:
https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/merge_requests/399
https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/merge_requests/228
2019-08-08 07:55:41 +08:00
Aaron Boxer
7dbcddff61 msdkdec: no need to cache output state info 2019-08-07 11:37:28 +03:00
Haihao Xiang
ef81c9d6c4 msdkdec: Update frame info from video parameters
A 10bit stream may have different depth values for Luma and Chroma, and
MSDK requires the frame info must match the corresponding video
parameters
2019-08-06 09:20:21 +00:00
Aaron Boxer
40f7001bef msdkdec: no need to cache allocation_caps 2019-08-05 13:32:15 +00:00
Aaron Boxer
96dc150b8b msdkdec: improve spelling and grammar of comments 2019-07-24 12:30:31 +02:00
Haihao Xiang
20f61db2e5 msdkdec: make sure mfx frame width/height meets MSDK's requirement
It is possible that the output region size (e.g. 192x144) is different
from the coded picture size (e.g. 192x256). We may adjust the alignment
parameters so that the padding is respected in GstVideoInfo and use
GstVideoInfo to calculate mfx frame width and height

This fixes the error below when decoding a stream which has different
output region size and coded picture size

0:00:00.057726900 28634 0x55df6c3220a0 ERROR                msdkdec
gstmsdkdec.c:1065:gst_msdkdec_handle_frame:<msdkh265dec0>
DecodeFrameAsync failed (failed to allocate memory)

Sample pipeline:

gst-launch-1.0 filesrc location=output.h265 ! h265parse ! msdkh265dec !
glimagesink
2019-07-24 08:41:13 +08:00
Haihao Xiang
8145a2236d msdkdec: remove unneeded code
Before calling gst_msdkdec_create_buffer_pool, the alignment parameters
have been adjusted.
2019-07-24 08:17:29 +08:00
Seungha Yang
9ec62418c3 msdk: Do not expose DMA buffer caps feature on Windows
On Windows, DMA buffer is not supported. PadTemplate with actually
supported feature seems to more make sense.
2019-07-22 16:13:43 +00:00
U. Artie Eoff
70bfd34b41 msdkdec: add post_configure virtual method
A post_configure virtual method is added to allow
codec subclasses to adjust the initialized parameters
after MFXVideoDECODE_DecodeHeader is called from the
gstmsdkdec::gst_msdkdec_handle_frame function.

This is useful if codecs want to adjust the output
parameters based on the codec-specific decoding
options that are present in the mfxInfoMFX structure
after MFXVideoDECODE_DecodeHeader initializes them.
2019-06-29 17:40:06 +00:00
Peter Seiderer
670481bdeb msdkdec: fix error handling in case of unsupported hardware
Check the return value of gst_msdk_context_ensure_context and
abort in case of failure.

https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/issues/945
2019-04-10 15:54:31 +00:00
Haihao Xiang
8e4d7c3642 msdkdec: add postinit_decoder
The workaround for https://github.com/Intel-Media-SDK/MediaSDK/issues/1139
is required for vp8 only, so move this workaround to the corresponding
postinit_decoder function

The pipeline below works with this change

gst-launch-1.0 filesrc location=SA10104.vc1 ! \
'video/x-wmv,profile=(string)advanced',width=720,height=480,framerate=14/1 ! \
msdkvc1dec ! fakesink
2019-03-29 13:32:06 +08:00
Haihao Xiang
01051e7dfd msdkdec: don't set unknown picture struct to progressive for vc1
MFXVideoDECODE_DecodeHeader only parses the sequence layer for VC1, so
the structure is unknown for a stream with interlace flag set in the
sequence layer. If forcing the struct to progressive in this plugin,
MediaSDK will fail to decode such streams.
2019-03-29 13:13:07 +08:00
Haihao Xiang
28a1b0c418 msdkdec: avoid infinite loop
It is possible MFXVideoDECODE_DecodeFrameAsync returns MFX_ERR_INCOMPATIBLE_VIDEO_PARAM
and this error can't be recovered by retrying MFXVideoDECODE_DecodeFrameAsync
in some cases, so we need to limit the number of retries to avoid infinite loop.

This fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/issues/909
2019-03-25 09:10:38 +00:00
Haihao Xiang
4b8e3b215a msdkdec: fix for small resolution 2019-02-28 11:51:06 +00:00
Haihao Xiang
18d410b81a msdkdec: fix for resolution change
Returning MFX_ERR_INCOMPATIBLE_VIDEO_PARAM from
MFXVideoDECODE_DecodeFrameAsync means the allocated mfx surface is not
suitable for the current frame, we need a new mfx surface and try
MFXVideoDECODE_DecodeFrameAsync again.
2019-02-28 11:08:24 +00:00
Haihao Xiang
a751b33072 msdkdec: force the alignment of width/height to 16 for vp8/vp9
MSDK library requires 16 alignment for vp8/vp9, otherwise a pipeline
for vp8/vp9 decoding might fail.

example pipeline:
gst-launch-1.0 filesrc location=vp8_1280x720.webm ! matroskademux ! \
msdkvp8dec ! fakesink

0:00:00.150565444 10657 0x55c8484036d0 ERROR                msdkdec
gstmsdkdec.c:1056:gst_msdkdec_handle_frame:<msdkvp8dec0>
DecodeFrameAsync failed (invalid video parameters)
2019-02-27 15:56:29 +00:00
Haihao Xiang
6287016a74 msdkdec: release the occupied surface for MFX_WRN_DEVICE_BUSY
When MFXVideoDECODE_DecodeFrameAsync () returns MFX_WRN_DEVICE_BUSY with
an output surface, a new input surface is required when retrying
MFXVideoDECODE_DecodeFrameAsync ().

This fixes the out-of-surface issue mentioned in
https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/issues/890
2019-02-24 19:13:57 +00:00
Haihao Xiang
028338ff44 msdkdec: set decode_only for output only
MSDK may return MFX_ERR_MORE_DATA but without output surface

Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/issues/887
2019-02-12 11:39:04 +00:00
Haihao Xiang
6c953438f5 msdkdec: Release occupied surface for MFX_ERR_MORE_DATA
An output surface is returned but without sync point when when
MFXVideoDECODE_DecodeFrameAsync () returns MFX_ERR_MORE_DATA, this
surface should be released too, otherwise the surface is occupied
and it is easy to exhaust all pre-allocated mfx surfaces.

Example pipeline (input_vp8.webm contains lots of frame with show_frame
set to 0):

gst-launch-1.0 filesrc location=input_vp8.webm ! matroskademux !
msdkvp8dec ! msdkvpp ! fakesink

0:00:05.995959693 19866 0x563f30f14590 ERROR                default
gstmsdkvideomemory.c:77:gst_msdk_video_allocator_get_surface: failed to
get surface available
ERROR: from element
/GstPipeline:pipeline0/GstMatroskaDemux:matroskademux0: Internal data
stream error.
2019-02-03 13:44:56 +08:00