[Intel-gfx] [PATCH v2 i-g-t] kms_atomic: Measure speed of some plane ioctls

[Chris Wilson]

On Tue, Apr 28, 2015 at 10:18:53AM +0100, Tvrtko Ursulin wrote:
> On 04/27/2015 10:20 PM, Chris Wilson wrote:
> >Same comment for commit = COMMIT_LEGACY here, mainly now for consistency
> >with before.
> 
> Here we want to be able to choose between commit types if we want to
> keep two subtests.

I just meant to have the assignment closer to where we choose between
LEGACY/UNIVERSAL.

> >Looks like magic to me. Is this enough variation to flush out future
> >bugs?
> 
> On which part does this apply?

The igt/kms interface is like magic and seems to me hides a lot of
details, so I am not really sure what the test does.

The second part was just wondering if coverage is wide enough to catch
bugs like panning being synchronous, or cursor images, or cursor size,
or sprite position/size, etc.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre