[PATCH i-g-t] tests/intel/gem_exec_schedule: Update for MTL RCS/CCS w/a

Dixit, Ashutosh ashutosh.dixit at intel.com
Tue Jan 9 21:58:05 UTC 2024


On Tue, 09 Jan 2024 11:18:49 -0800, John Harrison wrote:
>
> On 1/9/2024 09:51, Dixit, Ashutosh wrote:
> > On Tue, 09 Jan 2024 09:37:47 -0800, John Harrison wrote:
> >> On 1/8/2024 17:55, Dixit, Ashutosh wrote:
> >>> On Mon, 08 Jan 2024 16:12:00 -0800, John.C.Harrison at Intel.com wrote:
> >>> Hi John,
> >>>
> >>>> From: johnharr <johnharr at invalid-email.com>
> >>>>
> >>>> The RCS/CCS w/a means that if a spinners is sent to both RCS and CCS
> >>>> engines then the system is dead and requires a reset. Even
> >>>> pre-emptible spinners cannot be pre-empted due to the nature of the
> >>>> w/a. So do not get into that situation.
> >>>>
> >>>> Signed-off-by: John Harrison <John.C.Harrison at Intel.com>
> >>>> ---
> >>>>    tests/intel/gem_exec_schedule.c | 14 ++++++++++++++
> >>>>    1 file changed, 14 insertions(+)
> >>>>
> >>>> diff --git a/tests/intel/gem_exec_schedule.c b/tests/intel/gem_exec_schedule.c
> >>>> index 183ee1a214bb..1b086733ad50 100644
> >>>> --- a/tests/intel/gem_exec_schedule.c
> >>>> +++ b/tests/intel/gem_exec_schedule.c
> >>>> @@ -443,6 +443,14 @@
> >>>>
> >>>>    IGT_TEST_DESCRIPTION("Check that we can control the order of execution");
> >>>>
> >>>> +/*
> >>>> + * On some platforms, there are hardware w/a's that are not compatible
> >>>> + * with these tests.
> >>> This is too general. I would basically add the commit description here in
> >>> the comment. Also, not sure if we add the WA HSD number in IGT's?
> >> Yeah, wasn't sure how much detail we should include here and was wanting to
> >> keep it general. I can add the text from the email description.
> > I think better to add some detail so that people will cleanly know the
> > reason why tests are skipped.
> How about:
>
> /*
>  * The RCS/CCS workaround on MTL means that if a spinners is sent to both RCS
>  * and CCS engines concurrently then the system is dead and requires a reset of
>  * both engines to recover. Even pre-emptible spinners cannot be pre-empted due
>  * to the nature of the w/a. The requests must run to completion or be reset.
>  * And in the case of an certain IGT spinners, there is no completion until
>  * after the pre-emption. The result being that the tests get killed by the
>  * heartbeat and fail.
>  * So avoid that situation when running on MTL by skipping the compute engines.
>  */

Yes LGTM but I don't the exact issue and how to best describe it. So you
are the final arbiter :)

Reviewed-by: Ashutosh Dixit <ashutosh.dixit at intel.com>


More information about the igt-dev mailing list