[PATCH 2/3] drm: rcar-du: Reject planes located fully off-screen

[Daniel Vetter]

On Thu, Sep 28, 2017 at 11:16 AM, Laurent Pinchart
<laurent.pinchart at ideasonboard.com> wrote:
> Hi Daniel,
>
> On Thursday, 28 September 2017 10:10:34 EEST Daniel Vetter wrote:
>> On Wed, Sep 27, 2017 at 07:58:42PM +0100, Kieran Bingham wrote:
>> > Hi Laurent,
>> >
>> > Thankyou for the patch,
>> >
>> > This looks good, and passes the tests.
>> >
>> > On 16/08/17 00:03, Laurent Pinchart wrote:
>> > > There is no point in accepting fully off-screen planes as they won't be
>> > > displayed. Reject them in the atomic check.
>> > >
>> > > Signed-off-by: Laurent Pinchart
>> > > <laurent.pinchart+renesas at ideasonboard.com>
>> >
>> > Reviewed-by: Kieran Bingham <kieran.bingham+renesas at ideasonboard.com>
>>
>> Just a note, this is different from what i915 does. Not sure you want to
>> intentionally be incompatible with i915 and hence defacto make this part
>> of atomic undefined (or break some apps, depending upon how this pans
>> out). I think X loves to put the cursor offscreen at least.
>>
>> I think we even have igts to test this properly.
>
> If you update the documentation I'll submit a v2 that allows fully off-screen
> planes :-)

Quoting them:

"The destination rectangle can lie outside of the visible area of the
current mode of the CRTC."

This shit is better than I thought!

https://dri.freedesktop.org/docs/drm/gpu/drm-kms.html#plane-composition-properties

Now can I raise you one and challenge you to run the kms tests in igt
and see how non-compliant your driver is :-)

Cheers, Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch