Re: [PATCHv2 0/2] Stateful Encoding: final bits

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Hans,

On Wed, May 20, 2020 at 12:02 PM Hans Verkuil <hverkuil-cisco@xxxxxxxxx> wrote:
>
> This series adds the encoder spec and updates the VIDIOC_G/S_PARM
> documentation.
>
> This is a follow-up of the original "Stateful Encoding: final bits"
> series (1).
>
> The patches in that series that add support for V4L2_CID_MPEG_VIDEO_ENC_FRAME_RATE
> and V4L2_BUF_FLAG_TOO_SMALL have been dropped (the first is not necessary
> and the second can be skipped for now, see the irc discussion with
> Nicolas [3]).
>

I've reviewed both patches and had just one minor comment for each.
Otherwise they look good to me. Since I authored major parts of the
text, I'm not sure if my Reviewed-by or Acked-by would be worth
anything, but feel free to add them after fixing those minor issues.
:)

Best regards,
Tomasz

> The encoder spec has been updated since [2] with the following
> changes:
>
> - Document the optional VIDIOC_ENUM_FRAMEINTERVALS ioctl.
>
> - Document how to use VIDIOC_S_PARM:
>
>   1) calling S_PARM for the OUTPUT queue sets both the raw frame interval
>      (this is a hint only) and the coded frame interval.
>
>   2) calling S_PARM for the CAPTURE queue sets the coded frame interval
>      only. This is optional and can be used for off-line encoding. In
>      that case the OUTPUT frame interval can be used by the driver to
>      schedule multiple encoders.
>
>   Ideally S_PARM for the OUTPUT queue would just provide a hint, but
>   existing encoder drivers all use S_PARM for the OUTPUT queue to
>   define the coded frame interval, and that can't be changed.
>
> - Added a note that if a capture buffer is too small it will be
>   returned with V4L2_BUF_FLAG_ERROR and that more work has to be
>   done to properly support this corner case.
>
> - Clarify in the 'Encoding' section that there are more reasons
>   why 'a buffer queued to OUTPUT may result in more than one buffer
>   produced on CAPTURE'.
>
> I think that with these changes this stateful encoder spec is ready
> to be merged.
>
> Regards,
>
>         Hans
>
> [1] https://lore.kernel.org/linux-media/20191119113457.57833-6-hverkuil-cisco@xxxxxxxxx/T/
> [2] https://www.mail-archive.com/linux-media@xxxxxxxxxxxxxxx/msg149211.html
> [3] https://linuxtv.org/irc/irclogger_log/v4l?date=2020-05-19,Tue
>
> Hans Verkuil (1):
>   vidioc-g-parm.rst: update the VIDIOC_G/S_PARM documentation
>
> Tomasz Figa (1):
>   media: docs-rst: Document memory-to-memory video encoder interface
>
>  .../userspace-api/media/v4l/dev-encoder.rst   | 727 ++++++++++++++++++
>  .../userspace-api/media/v4l/dev-mem2mem.rst   |   1 +
>  .../userspace-api/media/v4l/pixfmt-v4l2.rst   |   5 +
>  .../userspace-api/media/v4l/v4l2.rst          |   2 +
>  .../media/v4l/vidioc-encoder-cmd.rst          |  51 +-
>  .../userspace-api/media/v4l/vidioc-g-parm.rst |  51 +-
>  6 files changed, 798 insertions(+), 39 deletions(-)
>  create mode 100644 Documentation/userspace-api/media/v4l/dev-encoder.rst
>
> --
> 2.25.1
>



[Index of Archives]     [Linux Input]     [Video for Linux]     [Gstreamer Embedded]     [Mplayer Users]     [Linux USB Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Yosemite Backpacking]

  Powered by Linux