Quoting Matthew Auld (2019-07-03 10:48:28)
> On 03/07/2019 10:17, Chris Wilson wrote:
> > We frequently, and not frequently enough, remember to flush residual
> > openations and objects at the end of a live subtest. The purpose is to
>
> operations
>
> > cleanup after every subtest, leaving a clean slate for the next subtest,
> > and perform early detection of leaky state.
> >
> > Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> > Cc: Tvrtko Ursulin <tvrtko.ursulin@xxxxxxxxx>
> Reviewed-by: Matthew Auld <matthew.auld@xxxxxxxxx>
>
> A variant with separate arguments for data and i915? Meh.
My first attempt was to use
struct i915_live_data {
struct drm_i915_private *i915;
};
but for this pass everything is basically using i915 as the data
parameter, so went with that for a smaller patch.
It may also be interesting to a do a i915_mock_subtests() that creates
and destroy a mock_gem_device around each subtest. I am sure we will
start using setup/teardown more creatively in future :)
Also remind me to pay attention to kunit and see if we can have a smooth
transition over to a central framework.
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
