Re: [RFC v3 02/19] media/v4l2-core: add new ioctl VIDIOC_G_DEF_EXT_CTRLS

[Hans Verkuil <hverkuil@xs4all.nl>]

On 07/20/2015 03:52 PM, Ricardo Ribalda Delgado wrote:
> Hello
> 
> I have no preference over the two implementations, but I see an issue
> with this suggestion.
> 
> 
> What happens to out out tree drivers, or drivers that don't support
> this functionality?
> 
> With the ioctl, the user receives a -ENOTTY. So he knows there is
> something wrong with the driver.
> 
> With this class, the driver might interpret this a simple G_VAL and
> return he current value with no way for the user to know what is going
> on.

Drivers that implement the current API correctly will return an error
if V4L2_CTRL_WHICH_DEF_VAL was specified. Such drivers will interpret
the value as a control class, and no control classes in that range exist.
See also class_check() in v4l2-ctrls.c.

The exception here is uvc which doesn't have this class check and it will
just return the current value :-(

I don't see a way around this, unfortunately.

Out-of-tree drivers that use the control framework are fine, and I don't
really care about drivers (out-of-tree or otherwise) that do not use the
control framework.

> Regarding the new implementation.... I can make some code next week,
> this week I am 120% busy :)

Wait until there is a decision first :-)

It's not a lot of work, I think.

Regards,

	Hans

> What do you think?
> --
> To unsubscribe from this list: send the line "unsubscribe linux-media" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>