[PATCH] drm: rcar-du: Create immutable zpos property for primary planes

Daniel Vetter daniel at ffwll.ch
Thu Apr 2 11:13:37 UTC 2020


On Thu, Apr 2, 2020 at 12:57 PM Daniel Stone <daniel at fooishbar.org> wrote:
>
> On Thu, 2 Apr 2020 at 11:40, Laurent Pinchart
> <laurent.pinchart+renesas at ideasonboard.com> wrote:
> > The R-Car DU driver creates a zpos property, ranging from 1 to 7, for
> > all the overlay planes, but leaves the primary plane without a zpos
> > property. The DRM/KMS API doesn't clearly specify if this is acceptable,
> > of it the property is mandatory for all planes when exposed for some of
> > the planes. Nonetheless, weston v8.0 has been reported to have trouble
> > handling this situation.
>
> Yeah. It didn't even occur to me/us that someone would do that, to be
> honest. We need to have zpos information for all planes that we're
> using in order for zpos to be at all meaningful, and we can't exactly
> avoid using the primary plane. Without knowing the primary plane's
> zpos, we can't know if the overlays are actually overlays or in fact
> underlays.

Maybe we need to clarify docs that if you expose zpos, then you should
expose it on all planes (opting for immutable zpos where it can't be
adjusted)? Care to type up that quick doc patch please?
-Daniel

>
> > The DRM core offers support for immutable zpos properties. Creating an
> > immutable zpos property set to 0 for the primary planes seems to be a
> > good way forward, as it shouldn't introduce any regression, and can fix
> > the issue. Do so.
>
> Perfect. We support immutable properties entirely well, we just need
> to know about them.
>
> > Reported-by: Kuninori Morimoto <kuninori.morimoto.gx at renesas.com>
> > Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas at ideasonboard.com>
>
> Reviewed-by: Daniel Stone <daniels at collabora.com>
>
> Cheers,
> Daniel
> _______________________________________________
> dri-devel mailing list
> dri-devel at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel



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


More information about the dri-devel mailing list