linux-kernel - Re: [PATCH 2/2] media: docs-rst: Document memory-to-memory video encoder interface

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <7b8b56e7-1617-5de6-9fa9-a10897a8f2f1@xs4all.nl>
Date:   Tue, 16 Oct 2018 15:49:58 +0200
From:   Hans Verkuil <hverkuil@...all.nl>
To:     Tomasz Figa <tfiga@...omium.org>,
        Philipp Zabel <p.zabel@...gutronix.de>
Cc:     Linux Media Mailing List <linux-media@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Stanimir Varbanov <stanimir.varbanov@...aro.org>,
        Mauro Carvalho Chehab <mchehab@...nel.org>,
        Pawel Osciak <posciak@...omium.org>,
        Alexandre Courbot <acourbot@...omium.org>, kamil@...as.org,
        a.hajda@...sung.com, Kyungmin Park <kyungmin.park@...sung.com>,
        jtp.park@...sung.com,
        Tiffany Lin (林慧珊) 
        <tiffany.lin@...iatek.com>,
        Andrew-CT Chen (陳智迪) 
        <andrew-ct.chen@...iatek.com>, todor.tomov@...aro.org,
        nicolas@...fresne.ca,
        Paul Kocialkowski <paul.kocialkowski@...tlin.com>,
        Laurent Pinchart <laurent.pinchart@...asonboard.com>,
        dave.stevenson@...pberrypi.org,
        Ezequiel Garcia <ezequiel@...labora.com>
Subject: Re: [PATCH 2/2] media: docs-rst: Document memory-to-memory video
 encoder interface

On 10/16/18 09:36, Tomasz Figa wrote:
> On Tue, Aug 7, 2018 at 3:54 PM Tomasz Figa <tfiga@...omium.org> wrote:
>>>> +   * The driver must expose following selection targets on ``OUTPUT``:
>>>> +
>>>> +     ``V4L2_SEL_TGT_CROP_BOUNDS``
>>>> +         maximum crop bounds within the source buffer supported by the
>>>> +         encoder
>>>> +
>>>> +     ``V4L2_SEL_TGT_CROP_DEFAULT``
>>>> +         suggested cropping rectangle that covers the whole source picture
>>>> +
>>>> +     ``V4L2_SEL_TGT_CROP``
>>>> +         rectangle within the source buffer to be encoded into the
>>>> +         ``CAPTURE`` stream; defaults to ``V4L2_SEL_TGT_CROP_DEFAULT``
>>>> +
>>>> +     ``V4L2_SEL_TGT_COMPOSE_BOUNDS``
>>>> +         maximum rectangle within the coded resolution, which the cropped
>>>> +         source frame can be output into; always equal to (0, 0)x(width of
>>>> +         ``V4L2_SEL_TGT_CROP``, height of ``V4L2_SEL_TGT_CROP``), if the
>>>> +         hardware does not support compose/scaling

Re-reading this I would rewrite this a bit:

if the hardware does not support composition or scaling, then this is always
equal to (0, 0)x(width of ``V4L2_SEL_TGT_CROP``, height of ``V4L2_SEL_TGT_CROP``).

>>>> +
>>>> +     ``V4L2_SEL_TGT_COMPOSE_DEFAULT``
>>>> +         equal to ``V4L2_SEL_TGT_CROP``
>>>> +
>>>> +     ``V4L2_SEL_TGT_COMPOSE``
>>>> +         rectangle within the coded frame, which the cropped source frame
>>>> +         is to be output into; defaults to
>>>> +         ``V4L2_SEL_TGT_COMPOSE_DEFAULT``; read-only on hardware without
>>>> +         additional compose/scaling capabilities; resulting stream will
>>>> +         have this rectangle encoded as the visible rectangle in its
>>>> +         metadata
>>>> +
>>>> +     ``V4L2_SEL_TGT_COMPOSE_PADDED``
>>>> +         always equal to coded resolution of the stream, as selected by the
>>>> +         encoder based on source resolution and crop/compose rectangles
>>>
>>> Are there codec drivers that support composition? I can't remember seeing any.
>>>
>>
>> Hmm, I was convinced that MFC could scale and we just lacked support
>> in the driver, but I checked the documentation and it doesn't seem to
>> be able to do so. I guess we could drop the COMPOSE rectangles for
>> now, until we find any hardware that can do scaling or composing on
>> the fly.
>>
> 
> On the other hand, having them defined already wouldn't complicate
> existing drivers too much either, because they would just handle all
> of them in the same switch case, i.e.
> 
> case V4L2_SEL_TGT_COMPOSE_BOUNDS:
> case V4L2_SEL_TGT_COMPOSE_DEFAULT:
> case V4L2_SEL_TGT_COMPOSE:
> case V4L2_SEL_TGT_COMPOSE_PADDED:
>      return visible_rectangle;
> 
> That would need one change, though. We would define
> V4L2_SEL_TGT_COMPOSE_DEFAULT to be equal to (0, 0)x(width of
> V4L2_SEL_TGT_CROP - 1, height of ``V4L2_SEL_TGT_CROP - 1), which

" - 1"? Where does that come from?

Usually rectangles are specified as widthxheight@...t,top.

> makes more sense than current definition, since it would bypass any
> compose/scaling by default.

I have no problem with drivers optionally implementing these rectangles,
even if they don't do scaling or composition. The question is, should it
be required for decoders? If there is a good reason, then I'm OK with it.

Regards,

	Hans