[PATCH RFC 00/11] drm/tilcdc: Atomic modeset support

[Daniel Vetter]

On Tue, May 10, 2016 at 12:29:23AM +0300, Jyri Sarha wrote:
> On 05/09/16 17:42, Daniel Vetter wrote:
> >> > It's not clear to me if a (primary) plane is required with atomic
> >> > universal planes and modesetting or not... I guess it is, when thinking
> >> > of how e.g. a framebuffer is configured to be shown on a screen when
> >> > using the atomic modesetting uapi.
> > You need a primary plane, but atomic doesn't require that it's enabled.
> > Which this simple display controller probably wont like, so seems like
> > this implementation of a primary plane is a bit too simple. You also need
> 
> So I do what I can, by checking in crtc check that the plane is part of
> the new state. What more should I do?g

that's not enough, since if you disable the primary plane it'll still be
part of the state, but disabled. You need to check that it actually is
enabled, and placed correctly (i.e. fullscreen). There's a helper for
that. But if you switch over to drm_simple_display_pipe that helper will
take care of things. At least in the final revision, current patch version
also lacks that check.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch