[Intel-gfx] [PATCH igt] igt/gem_exec_scheduler: HAS_SCHEDULER no longer means HAS_PREEMPTION

[Chris Wilson]

Quoting Joonas Lahtinen (2017-09-26 13:14:39)
> On Tue, 2017-09-26 at 10:32 +0100, Chris Wilson wrote:
> > Michal wants to limit machines that can do preemption, which means that
> > we no longer can assume that if we have a scheduler for execbuf, that
> > implies we have preemption.
> > 
> > v2: Try a capability mask instead
> > 
> > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > ---
> >  tests/gem_exec_schedule.c | 45 ++++++++++++++++++++++++++++-----------------
> >  1 file changed, 28 insertions(+), 17 deletions(-)
> > 
> > diff --git a/tests/gem_exec_schedule.c b/tests/gem_exec_schedule.c
> > index 0b1925f1..85c69703 100644
> > --- a/tests/gem_exec_schedule.c
> > +++ b/tests/gem_exec_schedule.c
> > @@ -33,6 +33,8 @@
> >  #include "igt_sysfs.h"
> >  
> >  #define LOCAL_PARAM_HAS_SCHEDULER 41
> > +#define   HAS_SCHEDULER (1u << 0)
> > +#define   HAS_PREEMPTION (1u << 2)
> 
> How about some BIT()? I think wehave it in IGT, at least I wrote
> patches for it.
> 
> Looks good to me, can you reference the latest Mesa patches in here and
> the kernel counterpart (reference the kernel counterpart here too).

Mesa doesn't care about this, no extension that I know declares
preemption. It is quite happy just to check whether or not it can change
the context priority, for the IMG_contxt_priority support.

The kennel patches depends upon having this inplace for CI to run the
tests on the series.
-Chris