[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <7869d6b4-6b6e-4fb6-95d9-bbe497caeceb@foss.st.com>
Date: Thu, 12 Sep 2024 14:32:41 +0200
From: Hugues FRUCHET <hugues.fruchet@...s.st.com>
To: Nicolas Dufresne <nicolas.dufresne@...labora.com>,
Mauro Carvalho Chehab
<mchehab@...nel.org>,
Ezequiel Garcia <ezequiel@...guardiasur.com.ar>,
Philipp Zabel <p.zabel@...gutronix.de>,
Hans Verkuil
<hverkuil-cisco@...all.nl>,
Fritz Koenig <frkoenig@...omium.org>,
Sebastian
Fricke <sebastian.fricke@...labora.com>,
Daniel Almeida
<daniel.almeida@...labora.com>,
Andrzej Pietrasiewicz
<andrzej.p@...labora.com>,
Benjamin Gaignard
<benjamin.gaignard@...labora.com>,
<linux-media@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
<linux-rockchip@...ts.infradead.org>,
<linux-stm32@...md-mailman.stormreply.com>
Subject: Re: [PATCH 1/2] media: uapi: add WebP VP8 frame flag
Hi Nicolas,
Thanks for reviewing.
On 9/11/24 19:12, Nicolas Dufresne wrote:
> Hi Hugues,
>
> Le mercredi 11 septembre 2024 à 15:50 +0200, Hugues Fruchet a écrit :
>> Add a flag indicating that VP8 bitstream is a WebP picture.
>
> Sounds like there should be some code changes in GStreamer that you haven't
> disclosed. Mind sharing how this new uAPI is used ? I would also expect this
> commit message to give more insight on what is special about WebP that makes
> this flag required.
GStreamer changes here:
https://gitlab.freedesktop.org/gstreamer/gstreamer/-/commit/138ecfac54ce85b273a26ff6f0fefe3998f8d436?merge_request_iid=7505
Verisilicon datasheet is not explicit on why WebP must be signaled to
hardware but WebP decoding fails if not.
Seems to me that such a simple addition on an already existing flag is
something acceptable and preferable to the development of a new complete
uAPI for WebP decoding.
>
> I would also need some more API or documentation that explain how we can
> differentiate a upstream decoder that is capable of WebP decoding from one that
> does not. I wonder if it would not have been better to define a new format ?
> That being said, I haven't looked at all in the specification and only rely on
> your cover letter and patch series.
>
> Nicolas
>
>>
>> Signed-off-by: Hugues Fruchet <hugues.fruchet@...s.st.com>
>> ---
>> .../userspace-api/media/v4l/ext-ctrls-codec-stateless.rst | 3 +++
>> include/uapi/linux/v4l2-controls.h | 1 +
>> 2 files changed, 4 insertions(+)
>>
>> diff --git a/Documentation/userspace-api/media/v4l/ext-ctrls-codec-stateless.rst b/Documentation/userspace-api/media/v4l/ext-ctrls-codec-stateless.rst
>> index 0da635691fdc..bb08aacddc9c 100644
>> --- a/Documentation/userspace-api/media/v4l/ext-ctrls-codec-stateless.rst
>> +++ b/Documentation/userspace-api/media/v4l/ext-ctrls-codec-stateless.rst
>> @@ -1062,6 +1062,9 @@ FWHT Flags
>> * - ``V4L2_VP8_FRAME_FLAG_SIGN_BIAS_ALT``
>> - 0x20
>> - Sign of motion vectors when the alt frame is referenced.
>> + * - ``V4L2_VP8_FRAME_FLAG_WEBP``
>> + - 0x40
>> + - Indicates that this frame is a WebP picture.
>>
>> .. c:type:: v4l2_vp8_entropy_coder_state
>>
>> diff --git a/include/uapi/linux/v4l2-controls.h b/include/uapi/linux/v4l2-controls.h
>> index 974fd254e573..e41b62f2cb2b 100644
>> --- a/include/uapi/linux/v4l2-controls.h
>> +++ b/include/uapi/linux/v4l2-controls.h
>> @@ -1897,6 +1897,7 @@ struct v4l2_vp8_entropy_coder_state {
>> #define V4L2_VP8_FRAME_FLAG_MB_NO_SKIP_COEFF 0x08
>> #define V4L2_VP8_FRAME_FLAG_SIGN_BIAS_GOLDEN 0x10
>> #define V4L2_VP8_FRAME_FLAG_SIGN_BIAS_ALT 0x20
>> +#define V4L2_VP8_FRAME_FLAG_WEBP 0x40
>>
>> #define V4L2_VP8_FRAME_IS_KEY_FRAME(hdr) \
>> (!!((hdr)->flags & V4L2_VP8_FRAME_FLAG_KEY_FRAME))
>
BR,
Hugues.
Powered by blists - more mailing lists