Message-ID: <c941a8b9-6ed8-ff54-6114-f31af37d25a1@xs4all.nl>
Date:   Mon, 28 Sep 2020 16:01:00 +0200
From:   Hans Verkuil <hverkuil@...all.nl>
To:     Laurent Pinchart <laurent.pinchart@...asonboard.com>
Cc:     Corentin Labbe <clabbe@...libre.com>, gregkh@...uxfoundation.org,
        mchehab@...nel.org, devel@...verdev.osuosl.org,
        linux-kernel@...r.kernel.org, linux-media@...r.kernel.org
Subject: Re: [PATCH RFT/RFC v2 00/47] staging: media: bring back zoran driver

On 28/09/2020 15:53, Laurent Pinchart wrote:
> Hi Hans,
> 
> On Mon, Sep 28, 2020 at 03:45:03PM +0200, Hans Verkuil wrote:
>> On 25/09/2020 20:30, Corentin Labbe wrote:
>>> Hello
>>>
>>> The zoran driver was removed in 5.3
>>> The main reason of the removing was lack of motivation to convert it to
>>> VB2
>>> Since I need it, I worked on bringing it back.
>>>
>>> So the plan to achieve it was:
>>> - clean up the coding style.
>>> - convert old usage/API
>>> - clean unused code
>>> - convert to VB2
>>>
>>> I have tried to split a bit the VB2 patch (by adding/removing code in
>>> another patch), but the result is unfortunately still a big patch.
>>>
>>> The result of this serie is a working zoran driver.
>>> Furthermore it is now compliant to v4l-compliance.
>>>
>>> In the process some old (useless) feature (fbuf, overlay) was removed
>>> for simplifying maintenance.
>>>
>>> The zoran hardware support MJPEG decompression, but this feature is
>>> temporarily disabled by this serie.
>>> But basically, this feature was unusable, since the only tool which
>>> permitted to use it was a mplayer option.
>>> But this mplayer option no longer compile (probably since a long time)
>>> and is such a hack (a copy of some private ffmpeg structure) that it is
>>> unfixable.
>>> Happily, when I started to work on zoran, a patch was posted on ffmpeg
>>> ML which permit it to output non-raw video stream (and so MJPEG for
>>> zoran case).
>>> But the zoran hw does not like some part of JPEG header it receives, so
>>> a filter need to be written.
>>> Anyway, this disabling is not a regression, since this feature was not
>>> usable since a long time.
>>>
>>> Since the driver was in staging, I targeted staging, but probably the
>>> driver is in a sufficient good shape to target media and bypass staging.
>>>
>>> This driver is tested on a DC10+ (on x86). Tests on different hardware
>>> are welcome.
>>>
>>> I would like to thanks all people that helped me to achieve this (mostly
>>> #v4l)
>>>
>>> Regards
>>>
>>> PS: this serie is resent a bit soon since linux-media didnt get some patch
>>> and cover letter
>>
>> Thank you very much for your hard work!
>>
>> I've just posted a PR for this driver since it is in good enough shape to
>> resurrect in staging.
>>
>> This means that starting with 5.10 this driver will once again be available.
>>
>> There are some things that I would like to see fixed before moving it out
>> of staging:
>>
>> 1) Check the driver with checkpatch --strict: I noticed various warnings
>> that would be good to fix. Let's have this cleaned up before moving it
>> out of staging.
>>
>> 2) I would like to see the output support fixed.
> 
> As lots of time has elapsed since the very first driver for this card
> was merged, and the (Linux) world has changed quite a bit since then,
> would it make sense to expose this feature through DRM/KMS instead ?

No, this is really a video output only, not something you would use as a
desktop output. I'm sure you can probably shoehorn it into DRM/KMS, but
it would not be a logical fit for this type of hardware. And besides, it
makes no sense to put a lot of effort in that for such old hardware.

Regards,

	Hans

Powered by blists - more mailing lists