Op 25-07-17 om 10:27 schreef Daniel Vetter:
On Wed, Jul 19, 2017 at 04:39:17PM +0200, Maarten Lankhorst wrote:
Use the new iterator macro and look for crtc_state->active instead of enable, only crtc_state->enable implies that vblanks will happen.
s/enable/active/, since enable only means logically enabled (aka resources reserved). With that my r-b holds. -Daniel
Ok thanks. I've pushed patch 1, 4, 5 and 6.
I'll wait a bit longer for the conflict in patch 3, and for a better solution on async in patch 2.
It seems the nouveau patches caused a bit of a conflict in nv50_disp_atomic_commit_tail because of better event handling compared to drm-misc, I've fixed up drm-tip.