[Intel-gfx] [PATCH 06/18] tests/drv_suspend: mark sysfs tests as basic

Jesse Barnes jbarnes at virtuousgeek.org
Fri Aug 14 09:10:18 PDT 2015


On 08/14/2015 09:01 AM, Daniel Vetter wrote:
> On Fri, Aug 14, 2015 at 08:29:40AM -0700, Jesse Barnes wrote:
>> On 08/14/2015 05:29 AM, Daniel Vetter wrote:
>>> On Thu, Aug 13, 2015 at 01:31:29PM -0700, Jesse Barnes wrote:
>>>> debugfs may not be mounted, but sysfs should always be restored after
>>>> suspend or hibernate.
>>>>
>>>> Signed-off-by: Jesse Barnes <jbarnes at virtuousgeek.org>
>>>
>>> We already have a suspend/resume testcase in kms_pipc_crc_basic. Do we
>>> have enough budget for this one?
>>
>> Yeah I thought about getting rid of the suspend/resume ones in pipe_crc
>> (marking them as non-basic), since I really just want the one set of
>> tests.  Any preference?
> 
> Imo crc is such a basic (hah!) validation tool that we really should try
> to run all the corner cases. Otherwise we have ugly situations where some
> tests randomyl fail, depending upon when they've been run after system
> suspend/resume or not. And with all our troubles we need to reboot a few
> times for a full run, so this is likely.
> 
> This actually happened (that's why we have these tests) and resulted in
> lots and lots of wtf until resolved. So yeah I think for crc we really
> want them all, even if it's a bit expensive, since if we don't there's a
> good chance for lots of flaky test results. And that's the prime problem
> we have with igt.

Ok sounds good, I'll drop the drv_suspend test then.

Jesse



More information about the Intel-gfx mailing list