[PATCH 2/2] modesetting: Detect whether damage tracking is needed

[Jason Ekstrand]

On Dec 19, 2014 9:48 PM, "Keith Packard" <keithp at keithp.com> wrote:
>
> Jason Ekstrand <jason at jlekstrand.net> writes:
>
>
> >> +    if (err != -EINVAL && err != -ENOSYS) {
> >
> > I'm not terribly familiar with the ioctls here, but why are we ignoring
> > EINVAL?  The previous patch made it sound like ENOSYS was the "I don't
> > support this" error and EINVAL was a genuine error.
>
> We're treating EINVAL and ENOSYS the same; an indication that the driver
> doesn't actually care about this information.
>
> I looked at the kernel source, and it returns -ENOSYS if there isn't a
> driver hook for receiving the dirty rects; perhaps there's some notion
> that -EINVAL would be another possible return value? Mostly, I didn't
> want to break a potentially working case, just to make it treat -ENOSYS
> exactly as it treated -EINVAL.

Seems perfectly reasonable assuming EINVAL isn't supposed to be fatal which
it doesn't sound like it is.

> > Apart from this my limited X knowledge says it looks perfectly
> > reasonable.
>
> You stand at the beginning on the path to X knowledge. In time, perhaps
> you will gain wisdom enough to review Peter's patches in the input
> subsystem. I also hope to become that wise someday.

I don't.  I've seen what understanding the input subsystem has done to some
people.

> > Reviewed-by: Jason Ekstrand <jason.ekstrand at intel.com>
>
> Thanks!
>
> --
> keith.packard at intel.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.x.org/archives/xorg-devel/attachments/20141219/b86ebbf8/attachment.html>