[Intel-gfx] [PATCH 1/3] drm/i915: Use pagecache write to prepopulate shmemfs from pwrite-ioctl

[Petri Latvala]

On Tue, Mar 07, 2017 at 10:22:09AM +0000, Chris Wilson wrote:
> > igt at gem_madvise@dontneed-before-exec: pass -> {'fail', 'incomplete'}
> That test is expected to fail, that it ever passed is a fluke.

That subtest should be removed from IGT then?


> incomplete there should be a failure in the runner? That should quite
> happily spot the purged object in execbuf.

The test executed right before was igt at drv_missed_irq and the used
versions of IGT and kernel did not have the recent fixes related to
that test. Quite possible that the incomplete result was due to that
one. 'incomplete' here means that the test didn't finish and the
machine was killed a watchdog. 'fail' also being in the set is because
the test order was different in that run.


-- 
Petri Latvala