The current code sets the prediction weight table by doing a memcpy of the
libva structure to the v4l2's structure.
However, for the offset and weight parameters, libva's structure uses
16-bits integer, while v4l2 uses 8-bits, which obviously doesn't work well
with memcpy.
Create a function to copy those arrays and matrices instead that follows
the algorithm defined in the H264 spec, and use it so that it works
properly.
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
The libva only provides the reference images needed to decode the current
picture, but not the full DPB. However, some codecs need that whole DPB in
order to decode a picture.
For example, the Allwinner hardware codec has an internal SRAM, with each
picture getting a slot in that SRAM, and during each decoding process, some
metadata will then be generated from that SRAM content to a separate
buffer. Therefore, each frames must be located at the same SRAM position
each time so that the metadata are then re-used properly.
However, since libva will only pass a few reference images, we can end up
in a situation where multiple, subsequent, frames will have the same
reference images set, but might all be used as reference later on and
cannot therefore be located at the same position.
And from a more theorical point of view, Linux expects a full blown DPB in
its H264 control.
In order to work around this, we can create a shadow of the DPB by simply
maintaining a list of 16 decoded images, each associated with their
VAPictureH264 and an age. This age is the last time we used that frame as
reference. When a new picture is decoded, either we assign it to a free
slot, or we reuse the slot from the frame that hasn't been used as a
reference for the longest time.
This is a much simpler approach than the one documented in the H264 spec,
but this shouldn't really be a problem since we don't handle the reference
frames ourselves, but just re-use the one from the libva, and taken from
the bitstream before. As such, frames that are not supposed to be used for
reference will not be anymore, their age will not increase, and therefore
after a while we will garbage-collect their slot to store a much newer
frame.
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Some functions setting the controls in the H264 code will need the context
in order to access the DPB. Make sure that we pass it as an argument.
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Some functions setting the controls will need the context in the future.
Make sure that we provide it as an argument.
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
The coding style has been a bit erratic. Enforce the linux kernel coding
style by reusing their .clang-format file, running clang-format on the
source, and ignoring the few shortcomings that clang-format has at the
moment (especially on aligning the define values).
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
The object_config structure embeds some cedrus.h defines, without including
its header, resulting in a compilation breakage if we don't include
sunxi_cedrus.h before config.h. Fix this.
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
This long structure name makes it quite difficult to fit within the 80
characters limit. Shorten it.
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
malloc returns a void pointer that can be casted to any pointer type
without any warning. Remove that cast.
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
The video_v4l2_format function is defined as inline in the header, but not
static. This will lead to a linker failure when you have multiple files
including the header.
Since it's not used anywhere but in video.c, move it there.
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Using the same words but not in the same order for both the type and the
variable name isn't particularly helpful, and prevents to stay within 80
characters. Shorten the name a bit.
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
The MPEG2 header doesn't need any header, since it's only using pointers.
Remove them, and add them to the mpeg2 file when needed.
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
The h264_fill_controls isn't used anywhere, but the h264_set_controls
function is. Since the one defined in the header is _fill_controls, this
leads to a warning at compile time. Fix it.
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
The BUFFER macro takes an implicit driver_data argument. In order to make
it obvious that we need it, let's put it as an explicit parameter.
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Using VAAPI as a video output (through vaPutSurface) is deprecated and
definitely not recommended for any use case. Since we're starting to
support non-X11 pipelines, remove X11 support altogether.
Signed-off-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
The libVA API expects rendering to be done after calling EndPicture.
Since SyncSurface calls are generally not issued in a way compatible
with dequeuing buffers (they might be called too early or too late),
always call SyncSurface from EndPicture.
Signed-off-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
Maxime Ripard