[Intel-gfx] [PATCH 15/26] drm/arc: Implement nonblocking commit correctly

[Daniel Vetter]

On Mon, May 30, 2016 at 11:36:06AM +0200, Maarten Lankhorst wrote:
> Op 30-05-16 om 11:24 schreef Daniel Vetter:
> > On Mon, May 30, 2016 at 10:15:14AM +0200, Maarten Lankhorst wrote:
> >> Op 29-05-16 om 20:35 schreef Daniel Vetter:
> >>> Committing with block it is not.
> >>>
> >>> Thanks to the fixed up vblank event handling we can just use the
> >>> helper support for nonblocking commits now.
> >>>
> >>> Cc: Carlos Palminha <palminha at synopsys.com>
> >>> Cc: Alexey Brodkin <abrodkin at synopsys.com>
> >>> Cc: linux-snps-arc at lists.infradead.org
> >>> Signed-off-by: Daniel Vetter <daniel.vetter at intel.com>
> >> This breaks page flip on arc, it uses drm_atomic_helper_page_flip.
> > How exactly does this break page_flip? Note that the commit to implement
> > generic nonblocking commit in drm_atomic_helper_commit is _before_ this
> > patch in the series.
> > -Daniel
> 
> This is patch 15/16, I found this in patch 21/26:
> 
> @@ -1147,8 +1206,11 @@ int drm_atomic_helper_commit(struct drm_device *dev,
>  {
>  	int ret;
>  
> -	if (nonblock)
> -		return -EBUSY;

Oh right, I mixed things up. Most of the prep work is for drm event
handling. We need to do that _before_ the non-blocking commit lands for
the drivers that just directly reuse drm_atomic_helper_commit, since the
nonblocking helpers will time out when drm events don't work.

But this one indeed should only happen after that commit. I'll reorder.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch