[PATCH] drm/amdgpu/dc: Require primary plane to be enabled whenever the CRTC is

[Daniel Vetter]

On Tue, Sep 01, 2020 at 10:56:42AM +0200, Michel Dänzer wrote:
> On 2020-09-01 9:57 a.m., Daniel Vetter wrote:
> > On Tue, Aug 25, 2020 at 04:55:28PM +0200, Michel Dänzer wrote:
> >> On 2020-08-24 9:43 a.m., Pekka Paalanen wrote:
> >> 
> >>> Sounds like the helpers you refer to are inadequate for your case.
> >>> Can't you fix the helpers in the long run and land this patch as an
> >>> immediate fix?
> >> 
> >> I'd rather leave working on those helpers to KMS developers. :)
> > 
> > I thought the issue is the rmfb ioctl trickery, which has this assumption
> > fully backed in. The wiggle room in there for semantic changes is iirc
> > pretty much nil, it took us a pile of patches to just get to the current
> > state. So it's not helper code, it's core legacy ioctl code for atomic
> > drivers.
> 
> My bad. Should I respin with an amended commit log?

Yeah if it's not yet merged then I think would be good to clarify that
this assumption is baked into rmfb, and that together with the assumption
the the legacy cursor ioctls just work we have fallout.

If (and hey I've been on vacations for 2 weeks, so good chances I remebers
this all wrong) this is indeed what we discussed a few weeks ago.

Cheers, Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch