Commit graph

648 commits

Author SHA1 Message Date
Gwenole Beauchesne
1d16669aca wayland: use a local event queue to avoid lock contention.
This improves performance when rendering several surfaces from within
the same process. e.g. a tee of vaapidecode'd buffers to vaapisink.
2013-01-28 19:06:59 +01:00
Gwenole Beauchesne
96d12f9eb1 wayland: fix thread-safe issues.
The Wayland API is not fully thread-safe and client applications shall
perform locking themselves on key functions. Besides, make sure to
release the lock if the _render() function fails.
2013-01-28 19:06:59 +01:00
Gwenole Beauchesne
1fb25b0853 wayland: really wait until the pending redraw completed.
Introduce gst_vaapi_window_wayland_sync() helper function to wait for
the completion of the redraw request. Use it in _render() function to
actually block until the previous draw request is completed.
2013-01-28 18:57:21 +01:00
Gwenole Beauchesne
23c6053b94 wayland: fix frame_redraw callback.
The redraw callback needs to be attached to the surface prior to the
commit. Otherwise, the callback notifies the next surface repaint,
which is not the desired behaviour. i.e. we want to be notified for
the surface we have just filled.

Another isse was the redraw_pending was reset before the actual completion
of the frame redraw callback function, thus causing concurrency issues.
e.g. the callback could have been called again, but with a NULL buffer.
2013-01-28 18:24:37 +01:00
Gwenole Beauchesne
087bf30c23 wayland: fix display sharing.
When the Wayland display is shared, we still have to create our own local
shell and compositor objects, since they are not propagated from the cache.
Likewise, we also need to determine the display size or vaapisink would
fail to account for the display aspect ratio, and will try to create a 0x0
window.
2013-01-28 18:24:37 +01:00
Gwenole Beauchesne
efd90114bf vc1: handle frames with multiple slices. 2013-01-23 17:02:45 +01:00
Gwenole Beauchesne
3eeff1786a vc1: handle CLOSED_ENTRY.
When CLOSED_ENTRY == 0, and if the B pictures that follow an entry-point
lack a reference anchor picture, these B pictures shall be discarded.

https://bugs.freedesktop.org/show_bug.cgi?id=59505
2013-01-23 11:19:32 +01:00
Gwenole Beauchesne
ee181d1b85 vc1: cope with latest codecparser changes.
Fix build with newer VC-1 codecparser where dqsbedge was renamed to
dqbedge, and now represents either DQSBEDGE or DQDBEDGE depending on
the actual value of DQPROFILE.
2013-01-23 10:25:52 +01:00
Gwenole Beauchesne
df411e435f vc1: fix size of encapsulated BDU.
Fix size of encapsulated BDUs since GstVC1BDU.size actually represents
the size of the BDU data, starting from offset, i.e. after any start
code is parsed.

This fixes a buffer overflow during the unescaping process.
2013-01-22 16:59:40 +01:00
Wind Yuan
a68c218c8e vc1: fix decoding of WMV3 videos in AVI format.
The AVI demuxer (avidemux) does not set a proper "format" attribute
to the generated caps. So, try to recover the video codec format from
the "wmvversion" property instead.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2013-01-22 16:10:07 +01:00
Gwenole Beauchesne
d69cb4389f vc1: review and report errors accordingly.
Use GST_ERROR() to report real errors instead of hiding them into
GST_DEBUG().
2013-01-22 16:10:07 +01:00
Gwenole Beauchesne
b134a97d42 vc1: don't create GstBuffers for all decoder units.
Don't create temporary GstBuffers for all decoder units, even if they
are lightweight "sub-buffers", since it is not really necessary to keep
the buffer data around.
2013-01-22 16:10:07 +01:00
Gwenole Beauchesne
ec3c4576c0 vc1: implement flush() hook.
Make it a simple DPB flush.
2013-01-22 16:09:45 +01:00
Gwenole Beauchesne
1513f52fc1 vc1: implement {start,end}_frame() hooks.
Implement GstVaapiDecoder.start_frame() and end_frame() semantics so
that to create new VA context earlier and submit VA pictures to the
HW for decoding as soon as possible. i.e. don't wait for the next
frame to start decoding the previous one.
2013-01-22 16:08:25 +01:00
Gwenole Beauchesne
823f16d7ce vc1: fix next POC for new sequence layers.
Fix next POC when a new sequence layer is reached. At this point, we
need to reset any previous reference picture, i.e. non B-frame.
2013-01-22 09:30:04 +01:00
Sreerenj Balachandran
472f3473e0 vc1: port to common GstVaapiDpb interface.
Use GstVaapiDpb interface instead of maintaining our own prev and next
picture pointers. While doing so, try to derive a sensible POC value.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2013-01-21 19:08:00 +01:00
Gwenole Beauchesne
142e13f637 vc1: fix decode_sequence_end() to return success, not EOS. 2013-01-21 16:35:58 +01:00
Gwenole Beauchesne
b24752da4f decoder: simplify gst_vaapi_decoder_get_surface().
Avoid extraenous branches, i.e. immediately return with success once we
have a decoded frame available.
2013-01-18 17:00:18 +01:00
Gwenole Beauchesne
a811a5de3d decoder: optimize and clean decode_step() up.
Avoid usage of goto. Simplify decode_step() process to first accumulate all
pending buffers into the GstAdapter, and then parse and decode units from
that input adapter. Stop the process once a frame is fully decoded or an
error occurred.
2013-01-18 16:56:15 +01:00
Gwenole Beauchesne
cd52fa315a display: move "vaapi" debug init to libgstvaapi_init_once(). 2013-01-18 14:49:15 +01:00
Gwenole Beauchesne
655f400e1c display: dump gstreamer-vaapi version for debugging purposes. 2013-01-18 14:49:15 +01:00
Gwenole Beauchesne
319f87f142 h264: implement GstVaapiDecoder::flush() as a DPB flush. 2013-01-17 18:40:36 +01:00
Gwenole Beauchesne
34162f5001 h264: handle end-of-stream NALU.
Handle <end-of-stream> NAL unit to actually flush any pending picture
from the DPB.
2013-01-17 18:40:36 +01:00
Gwenole Beauchesne
60acedb56e decoder: add GstVaapiDecoder::flush() hook. 2013-01-17 18:39:22 +01:00
Gwenole Beauchesne
e4e3a58bc0 decoder: fix check for end-of-stream in raw API mode.
Make sure to immediately return GST_VAAPI_DECODER_STATUS_END_OF_STREAM
if the end-of-stream was already reached at the previous iteration.
2013-01-17 18:39:22 +01:00
Gwenole Beauchesne
cc347cb41e decoder: make decode_step() return once the frame is decoded.
Make sure we always have a free surface left to use for decoding the
current frame. This means that decode_step() has to return once a frame
gets decoded. If the current adapter contains more buffers with valid
frames, they will get parsed and decoded on subsequent iterations.
2013-01-17 18:39:22 +01:00
Gwenole Beauchesne
059e032103 docs: expose new interfaces. 2013-01-14 12:58:41 +01:00
Gwenole Beauchesne
9372f14f94 dpb: cosmetics (clean-ups). 2013-01-14 10:58:49 +01:00
Gwenole Beauchesne
e44d8ee6e3 dpb: port to GstVaapiMiniObject. 2013-01-14 10:46:50 +01:00
Gwenole Beauchesne
6f4e01258d dpb: drop GstVaapiDpb2 interface, keep only one class.
Keep only one DPB interface and rename gst_vaapi_dpb2_get_references()
to gst_vaapi_dpb_get_neighbours() so that to retrieve pictures in DPB
around the specified picture POC.
2013-01-14 10:21:53 +01:00
Sreerenj Balachandran
eafdd771ae dpb: rename GstVaapiDpbMpeg2 to GstVaapiDpb2.
Move GstVaapiDpbMpeg2 API to a more generic version that could also be
useful to other decoders that require 2 reference pictures, e.g. VC-1.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2013-01-14 10:11:21 +01:00
Holger Kaelberer
082a56599b overlay: fix build without advanced GstVideoOverlayFormatFlags.
Check for global-alpha support in GstVideoOverlayComposition API.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2013-01-11 15:47:46 +01:00
Gwenole Beauchesne
7e1a8eabfe overlay: fix ordering of composition layers.
Make sure to maintain the association order of composition layers when
GstVideoOverlayRectangle objects are kept around (cached).
2013-01-11 15:32:17 +01:00
Holger Kaelberer
2ecb955626 overlay: fix support for global-alpha.
Fix support for global-alpha subpictures. The previous changes brought
the ability to check for GstVideoOverlayRectangle changes by comparing
the underlying pixel buffer pointers. If sequence number and pixel data
did not change, then this is an indication that only the global-alpha
value changed. Now, try to update the underlying VA subpicture global-alpha
value.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2013-01-11 15:29:12 +01:00
Gwenole Beauchesne
e6390d6e5f overlay: detect render-rect changes.
Don't re-upload VA subpicture if only the render rectangle changed.
Rather deassociate the subpicture and re-associate it with the new
render rectangle.
2013-01-11 15:29:12 +01:00
Gwenole Beauchesne
e876d9a581 overlay: fix check for pixels buffer change.
A GstVideoOverlayRectangle is created whenever the underlying pixels data
change. However, when global-alpha is supported, it is possible to re-use
the same GstVideoOverlayRectangle but with a change to the global-alpha
value. This process causes a change of sequence number, so we can no longer
check for that.

Still, if sequence numbers did not change, then there was no change in
global-alpha either. So, we need a way to compare the underlying GstBuffer
pointers. There is no API to retrieve the original pixels buffer from
a GstVideoOverlayRectangle. So, we use the following heuristics:

1. Use gst_video_overlay_rectangle_get_pixels_unscaled_argb() with the same
   format flags from which the GstVideoOverlayRectangle was created. This
   will work if there was no prior consumer of the GstVideoOverlayRectangle
   with alternate (non-"native") format flags.

2. In overlay_rectangle_has_changed_pixels(), we have to use the same
   gst_video_overlay_rectangle_get_pixels_unscaled_argb() function but
   with flags that match the subpicture. This is needed to cope with
   platforms that don't support global-alpha in HW, so the gst-video
   layer takes care of that and fixes this up with a possibly new
   GstBuffer, and hence pixels data (or) in-place by caching the current
   global-alpha value applied. So we have to determine the rectangle
   was previously used, based on what previous flags were used to
   retrieve the ARGB pixels buffer.
2013-01-11 15:29:12 +01:00
Gwenole Beauchesne
a14d259060 overlay: optimize cache at the GstVideoOverlayRectangle level.
We previously assumed that an overlay composition changed if the number
of overlay rectangles in there actually changed, or that the rectangle
was updated, and thus its seqnum was also updated.

Now, we can cope with cases where the GstVideoOverlayComposition grew
by one or a few more overlay rectangles, and the initial overlay rectangles
are kept as is.
2013-01-11 15:29:11 +01:00
Gwenole Beauchesne
ad6cdc0b52 overlay: simplify caching of GstVideoOverlayComposition objects.
Create the GPtrArray once in the _init() function and destroy it only
in the _finalize() function. Then use overlay_clear() to remove all
subpicture associations for intermediate updates, don't recreate the
GPtrArray.

Make GstVaapiOverlayRectangle a reference counted object. Also make
sure that overlay_rectangle_new() actually creates and associates the
VA subpicture.
2013-01-11 15:28:25 +01:00
Holger Kaelberer
2d2334afed overlay: add support for global-alpha.
Handle global-alpha from GstVideoOverlayComposition API. Likewise,
the same code path could also work for premultiplied-alpha but this
was not tested.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2013-01-11 14:23:08 +01:00
Holger Kaelberer
1cd4a8fc04 subpicture: add support for global-alpha.
Add the necessary helpers in GstVaapiDisplay to determine whether subpictures
with global alpha are supported or not. Also add accessors in GstVaapiSubpicture
to address this feature.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>
2013-01-11 14:23:06 +01:00
Gwenole Beauchesne
1307a5b4ef subpicture: add premultiplied-alpha and global-alpha feature flags.
Add premultiplied-alpha and global-alpha feature flags, along with converters
between VA-API and gstreamer-vaapi definitions. Another round of helpers is
also necessary for GstVideoOverlayComposition API.
2013-01-11 14:17:30 +01:00
Gwenole Beauchesne
fc1f9a64e4 display: allow image/subpicture formats with additional flags.
Introduce new GstVaapiFormatInfo to store the actual GstVaapiImageFormat
and any additional flags needed. Currently, all flags are set to zero.
2013-01-11 14:17:30 +01:00
Gwenole Beauchesne
591b637cfc decoder: fix mini object implementation on 64-bit systems.
Use GPOINTER_TO_SIZE() instead of GPOINTER_TO_UINT() while manipulating
pointers. The latter is meant to be 32-bit only, not uintptr_t like size.
Only a gsize can hold all bits of a pointer.

Thanks to Ouping Zhang for spotting this error.
2013-01-10 10:14:42 +01:00
Gwenole Beauchesne
0ecb3a627b mpeg2: optimize scan for the end of the frame.
Heuristic: if the second start-code is available, check whether that
one marks the start of a new frame because e.g. this is a sequence
or picture header. This doesn't save much, since we already cache the
results.
2013-01-09 16:05:39 +01:00
Gwenole Beauchesne
3bc6078f32 mpeg2: optimize scan for start codes.
Accelerate scan for start codes by skipping up to 3 bytes per iteration.
A start code prefix is defined by the following bytes: 00 00 01. Thus,
for any group of 3 bytes (xx yy zz), we have the following possible cases:

  1. If zz != 1, this cannot be a start code, then skip 3 bytes;
  2. If yy != 0, this cannot be a start code, then skip 2 bytes;
  3. If xx != 0 or zz != 1, this cannot be a start code, then skip 1 byte;
  4. xx == 00, yy == 00, zz == 1, we have match!

This algorithm requires to peek bytes from the adapter. This increases the
amount of bytes copied to a temporary buffer, but this process is much faster
than scanning for all the bytes and using shift/masks. So, overall, this is
a win.
2013-01-09 16:05:18 +01:00
Gwenole Beauchesne
0180ef635c mpeg2: drop useless gst_adapter_peek().
Drop useless gst_adapter_peek() since the returned buffer was not used
and this could incur superfluous memcpy().
2013-01-08 16:41:44 +01:00
Gwenole Beauchesne
4a69e395cd mpeg2: cosmetics: move parse_slice() down. 2013-01-07 16:12:42 +01:00
Gwenole Beauchesne
6dd8eab023 mpeg2: avoid too many allocations of parser info objects.
Move parsing back to decoding step, but keep functions separate for now.
This is needed for future optimizations that may introduce some meta data
for parsed info attached to codec frames.
2013-01-07 16:12:42 +01:00
Gwenole Beauchesne
8c403c2d98 decoder: decoder units are no longer dynamically allocated objects. 2013-01-07 14:11:39 +01:00
Gwenole Beauchesne
a486d1af66 decoder: optimize pre-allocation of decoder units.
Optimize pre-allocation of decoder units, thus avoiding un-necessary
memory reallocations. The heuristic used is that we could have around
one slice unit per macroblock line.
2013-01-07 14:11:39 +01:00