Re: [PATCH v4 1/2] media: uapi: Add H264 low-level decoder API compound controls.

From: Ezequiel Garcia
Date: Tue Mar 05 2019 - 14:54:21 EST


On Tue, 2019-03-05 at 12:16 +0100, Maxime Ripard wrote:
> On Fri, Feb 22, 2019 at 04:46:17PM +0900, Tomasz Figa wrote:
> > Hi Maxime,
> >
> > On Wed, Feb 20, 2019 at 11:17 PM Maxime Ripard
> > <maxime.ripard@xxxxxxxxxxx> wrote:
> > > From: Pawel Osciak <posciak@xxxxxxxxxxxx>
> > >
> > > Stateless video codecs will require both the H264 metadata and slices in
> > > order to be able to decode frames.
> > >
> > > This introduces the definitions for a new pixel format for H264 slices that
> > > have been parsed, as well as the structures used to pass the metadata from
> > > the userspace to the kernel.
> > >
> > > Co-Developped-by: Maxime Ripard <maxime.ripard@xxxxxxxxxxx>
> > > Signed-off-by: Pawel Osciak <posciak@xxxxxxxxxxxx>
> > > Signed-off-by: Guenter Roeck <groeck@xxxxxxxxxxxx>
> > > Signed-off-by: Maxime Ripard <maxime.ripard@xxxxxxxxxxx>
> >
> > Thanks for the patch. Some comments inline.
> >
> > [snip]
> > > +``V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS (struct)``
> > > + Specifies the slice parameters (as extracted from the bitstream)
> > > + for the associated H264 slice data. This includes the necessary
> > > + parameters for configuring a stateless hardware decoding pipeline
> > > + for H264. The bitstream parameters are defined according to
> > > + :ref:`h264`. Unless there's a specific comment, refer to the
> > > + specification for the documentation of these fields, section 7.4.3
> > > + "Slice Header Semantics".
> >
> > Note that this is expected to be an array, with entries for all the
> > slices included in the bitstream buffer.
> >
> > > +
> > > + .. note::
> > > +
> > > + This compound control is not yet part of the public kernel API and
> > > + it is expected to change.
> > > +
> > > +.. c:type:: v4l2_ctrl_h264_slice_param
> > > +
> > > +.. cssclass:: longtable
> > > +
> > > +.. flat-table:: struct v4l2_ctrl_h264_slice_param
> > > + :header-rows: 0
> > > + :stub-columns: 0
> > > + :widths: 1 1 2
> > > +
> > > + * - __u32
> > > + - ``size``
> > > + -
> > > + * - __u32
> > > + - ``header_bit_size``
> > > + -
> > > + * - __u16
> > > + - ``first_mb_in_slice``
> > > + -
> > > + * - __u8
> > > + - ``slice_type``
> > > + -
> > > + * - __u8
> > > + - ``pic_parameter_set_id``
> > > + -
> > > + * - __u8
> > > + - ``colour_plane_id``
> > > + -
> > > + * - __u8
> > > + - ``redundant_pic_cnt``
> > > + -
> > > + * - __u16
> > > + - ``frame_num``
> > > + -
> > > + * - __u16
> > > + - ``idr_pic_id``
> > > + -
> > > + * - __u16
> > > + - ``pic_order_cnt_lsb``
> > > + -
> > > + * - __s32
> > > + - ``delta_pic_order_cnt_bottom``
> > > + -
> > > + * - __s32
> > > + - ``delta_pic_order_cnt0``
> > > + -
> > > + * - __s32
> > > + - ``delta_pic_order_cnt1``
> > > + -
> > > + * - struct :c:type:`v4l2_h264_pred_weight_table`
> > > + - ``pred_weight_table``
> > > + -
> > > + * - __u32
> > > + - ``dec_ref_pic_marking_bit_size``
> > > + -
> > > + * - __u32
> > > + - ``pic_order_cnt_bit_size``
> > > + -
> > > + * - __u8
> > > + - ``cabac_init_idc``
> > > + -
> > > + * - __s8
> > > + - ``slice_qp_delta``
> > > + -
> > > + * - __s8
> > > + - ``slice_qs_delta``
> > > + -
> > > + * - __u8
> > > + - ``disable_deblocking_filter_idc``
> > > + -
> > > + * - __s8
> > > + - ``slice_alpha_c0_offset_div2``
> > > + -
> > > + * - __s8
> > > + - ``slice_beta_offset_div2``
> > > + -
> > > + * - __u8
> > > + - ``num_ref_idx_l0_active_minus1``
> > > + -
> > > + * - __u8
> > > + - ``num_ref_idx_l1_active_minus1``
> > > + -
> > > + * - __u32
> > > + - ``slice_group_change_cycle``
> > > + -
> > > + * - __u8
> > > + - ``ref_pic_list0[32]``
> > > + -
> > > + * - __u8
> > > + - ``ref_pic_list1[32]``
> > > + -
> >
> > Should we explicitly document that these are the lists after applying
> > the per-slice modifications, as opposed to the original order from
> > v4l2_ctrl_h264_decode_param?
> >
> > [snip]
> > > + * .. _V4L2-PIX-FMT-H264-SLICE:
> > > +
> > > + - ``V4L2_PIX_FMT_H264_SLICE``
> > > + - 'S264'
> > > + - H264 parsed slice data, as extracted from the H264 bitstream.
> > > + This format is adapted for stateless video decoders that
> > > + implement an H264 pipeline (using the :ref:`codec` and
> > > + :ref:`media-request-api`). Metadata associated with the frame
> > > + to decode are required to be passed through the
> > > + ``V4L2_CID_MPEG_VIDEO_H264_SPS``,
> > > + ``V4L2_CID_MPEG_VIDEO_H264_PPS``,
> > > + ``V4L2_CID_MPEG_VIDEO_H264_SLICE_PARAMS`` and
> > > + ``V4L2_CID_MPEG_VIDEO_H264_DECODE_PARAMS`` controls and
> > > + scaling matrices can optionally be specified through the
> > > + ``V4L2_CID_MPEG_VIDEO_H264_SCALING_MATRIX`` control. See the
> > > + :ref:`associated Codec Control IDs <v4l2-mpeg-h264>`.
> > > + Exactly one output and one capture buffer must be provided for
> > > + use with this pixel format. The output buffer must contain the
> > > + appropriate number of macroblocks to decode a full
> > > + corresponding frame to the matching capture buffer.
> >
> > What does it mean that a control can be optionally specified? A
> > control always has a value, so how do we decide that it was specified
> > or not? Should we have another control (or flag) that selects whether
> > to use the control? How is it better than just having the control
> > initialized with the default scaling matrix and always using it?
>
> Ok, I'll change it.
>
> > > diff --git a/include/uapi/linux/videodev2.h b/include/uapi/linux/videodev2.h
> > > index 9a920f071ff9..6443ae53597f 100644
> > > --- a/include/uapi/linux/videodev2.h
> > > +++ b/include/uapi/linux/videodev2.h
> > > @@ -653,6 +653,7 @@ struct v4l2_pix_format {
> > > #define V4L2_PIX_FMT_H264 v4l2_fourcc('H', '2', '6', '4') /* H264 with start codes */
> > > #define V4L2_PIX_FMT_H264_NO_SC v4l2_fourcc('A', 'V', 'C', '1') /* H264 without start codes */
> > > #define V4L2_PIX_FMT_H264_MVC v4l2_fourcc('M', '2', '6', '4') /* H264 MVC */
> > > +#define V4L2_PIX_FMT_H264_SLICE v4l2_fourcc('S', '2', '6', '4') /* H264 parsed slices */
> >
> > Are we okay with adding here already, without going through staging first?
>
> This is what we did for MPEG-2 already (the format is public but the
> controls are not), so I'm not sure this is causing any issue.
>

As pointed out by Nicolas on IRC, the V4L2_PIX_FMT_H264_SLICE_RAW and V4L2_PIX_FMT_H264_SLICE_ANNEX_B
should describe pretty well the pixel format.

I believe it's acceptable for them to go public.

Thanks!
Ezequiel