Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <20190603112835.19661-1-hverkuil-cisco@xs4all.nl>
 <20190603112835.19661-2-hverkuil-cisco@xs4all.nl> <CAAFQd5Aa-PQEakeg3sC_EDYdKy15hHx09Qmk6Jik4COeBe3xVA@mail.gmail.com>
 <02da6340-3174-c03b-ffad-cc9a0a58afab@xs4all.nl>
In-Reply-To: <02da6340-3174-c03b-ffad-cc9a0a58afab@xs4all.nl>
From:   Tomasz Figa <tfiga@chromium.org>
Date:   Wed, 10 Jul 2019 17:23:52 +0900
Message-ID: <CAAFQd5A4+o9MZL8_TTdKOYa04O87GEi81PU2Kipa_Seeg98oMA@mail.gmail.com>
Subject: Re: [PATCHv4 1/2] media: docs-rst: Document memory-to-memory video
 decoder interface
To:     Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc:     Linux Media Mailing List <linux-media@vger.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Alexandre Courbot <acourbot@chromium.org>,
        Philipp Zabel <p.zabel@pengutronix.de>,
        Stanimir Varbanov <stanimir.varbanov@linaro.org>,
        Andrew-CT Chen <andrew-ct.chen@mediatek.com>,
        Tiffany Lin <tiffany.lin@mediatek.com>,
        Pawel Osciak <posciak@chromium.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, Jul 10, 2019 at 5:09 PM Hans Verkuil <hverkuil-cisco@xs4all.nl> wro=
te:
>
> On 7/3/19 6:58 AM, Tomasz Figa wrote:
> > Hi Hans,
> >
> > On Mon, Jun 3, 2019 at 8:28 PM Hans Verkuil <hverkuil-cisco@xs4all.nl> =
wrote:
> >>
> >> From: Tomasz Figa <tfiga@chromium.org>
> >>
> >> Due to complexity of the video decoding process, the V4L2 drivers of
> >> stateful decoder hardware require specific sequences of V4L2 API calls
> >> to be followed. These include capability enumeration, initialization,
> >> decoding, seek, pause, dynamic resolution change, drain and end of
> >> stream.
> >>
> >> Specifics of the above have been discussed during Media Workshops at
> >> LinuxCon Europe 2012 in Barcelona and then later Embedded Linux
> >> Conference Europe 2014 in D=C3=BCsseldorf. The de facto Codec API that
> >> originated at those events was later implemented by the drivers we alr=
eady
> >> have merged in mainline, such as s5p-mfc or coda.
> >>
> >> The only thing missing was the real specification included as a part o=
f
> >> Linux Media documentation. Fix it now and document the decoder part of
> >> the Codec API.
> >>
> >> Signed-off-by: Tomasz Figa <tfiga@chromium.org>
> >> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
> >> ---
> >>  Documentation/media/uapi/v4l/dev-decoder.rst  | 1084 ++++++++++++++++=
+
> >>  Documentation/media/uapi/v4l/dev-mem2mem.rst  |    8 +-
> >>  Documentation/media/uapi/v4l/pixfmt-v4l2.rst  |    5 +
> >>  Documentation/media/uapi/v4l/v4l2.rst         |   10 +-
> >>  .../media/uapi/v4l/vidioc-decoder-cmd.rst     |   41 +-
> >>  5 files changed, 1132 insertions(+), 16 deletions(-)
> >>  create mode 100644 Documentation/media/uapi/v4l/dev-decoder.rst
> >>
> >
> > Thanks a lot for helping with remaining changes.
> >
> > Just one thing inline our team member found recently.
> >
> > [snip]
> >> +Capture setup
> >> +=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> >> +
> > [snip]
> >> +4.  **Optional.** Set the ``CAPTURE`` format via :c:func:`VIDIOC_S_FM=
T` on the
> >> +    ``CAPTURE`` queue. The client may choose a different format than
> >> +    selected/suggested by the decoder in :c:func:`VIDIOC_G_FMT`.
> >> +
> >> +    * **Required fields:**
> >> +
> >> +      ``type``
> >> +          a ``V4L2_BUF_TYPE_*`` enum appropriate for ``CAPTURE``.
> >> +
> >> +      ``pixelformat``
> >> +          a raw pixel format.
> >
> > The client should be able to set the width and height as well. It's a
> > quite frequent case, especially in DMA-buf import mode, that the
> > buffers are actually bigger (e.g. more alignment) than what we could
> > get from the decoder by default. For sane hardware platforms it's
> > reasonable to expect that such bigger buffers could be handled as
> > well, as long as we update the width and height here.
>
> I've added this:
>
>      ``width``, ``height``
>          frame buffer resolution of the decoded stream; typically unchang=
ed from
>          what was returned with :c:func:`VIDIOC_G_FMT`, but it may be dif=
ferent
>          if the hardware supports composition and/or scaling.
>
> Is that what you were looking for?
>

Not sure if composition is a requirement here, but I guess it depends
on how we define composition. Most of the hardware today at least
support arbitrary strides (+/- some alignment), but still write the
pixels at (0,0)x(w,h).

In fact, there would be already some composition happening, even
without arbitrary strides, because G_FMT would return values aligned
in some way, but only the visible rectangle would contain meaningful
pixel data.

Best regards,
Tomasz