Re: [Intel-gfx] [PATCH v2] drm/atomic: Handle vblank events in atomic ioctl correctly.

2015-06-15 Thread Maarten Lankhorst
Op 15-06-15 om 08:22 schreef Daniel Vetter: > On Tue, Jun 02, 2015 at 09:00:46AM +0200, Maarten Lankhorst wrote: >> All users of async updates seem to clear clear crtc_state->event >> correctly, so move destroying vblank event to crtc_destroy_state. >> >> This is better than manually free'ing it in

Re: [Intel-gfx] [PATCH v2] drm/atomic: Handle vblank events in atomic ioctl correctly.

2015-06-14 Thread Daniel Vetter
On Tue, Jun 02, 2015 at 09:00:46AM +0200, Maarten Lankhorst wrote: > All users of async updates seem to clear clear crtc_state->event > correctly, so move destroying vblank event to crtc_destroy_state. > > This is better than manually free'ing it in the atomic ioctl, since > this code seems to do

[Intel-gfx] [PATCH v2] drm/atomic: Handle vblank events in atomic ioctl correctly.

2015-06-02 Thread Maarten Lankhorst
All users of async updates seem to clear clear crtc_state->event correctly, so move destroying vblank event to crtc_destroy_state. This is better than manually free'ing it in the atomic ioctl, since this code seems to do it wrong. While we're at it handle -EDEADLK from atomic_commit correctly, be