Re: [RFC PATCH 2/2] media/uapi/v4l: clarify cropcap/crop/selection behavior

[Sylwester Nawrocki <snawrocki@kernel.org>]

On 06/16/2017 03:08 PM, Sakari Ailus wrote:
> On Mon, May 08, 2017 at 04:35:06PM +0200, Hans Verkuil wrote:
>> From: Hans Verkuil <hans.verkuil@cisco.com>
>>
>> Unfortunately the use of 'type' was inconsistent for multiplanar
>> buffer types. Starting with 4.12 both the normal and _MPLANE variants
>> are allowed, thus making it possible to write sensible code.
>>
>> Yes, we messed up :-(
> 
> Things worse than this have happened. :-)
> 
> I don't think in general I would write about driver bugs that have already
> been fixed in developer documentation. That said, I'm not sure how otherwise
> developers would learn about this, but OTOH has it been reported to us as a
> bug?
> 
> Marek, Sylwester: any idea how widely the drivers in question are in use? If
> there's a real chance of hitting this, then it does make sense to mention it
> in the documentation.

I'm not sure how widely are used those drivers, I think we should just assume 
they are deployed and whatever we do should be backwards compatible. I don't 
think it is much helpful for Exynos to add notes like this in the documentation, 
so far I didn't receive any related bug report.

And even though now there is some confusion because drivers see "regular" 
buffer type while user space uses _MPLANE I like the $subject patch set
as it makes the API clearer from user perspective.

--
Regards,
Sylwester