[Intel-gfx] [PATCH v2] drm/i915/execlists: Inhibit context save/restore for the fake preempt context
Chris Wilson
chris at chris-wilson.co.uk
Wed Jan 24 20:46:34 UTC 2018
Quoting Chris Wilson (2018-01-24 09:44:45)
> Quoting Michel Thierry (2018-01-24 01:24:25)
> > On 1/23/2018 1:04 PM, Chris Wilson wrote:
> > > We only use the preempt context to inject an idle point into execlists.
> > > We never need to reference its logical state, so tell the GPU never to
> > > load it or save it.
> > >
> > > v2: BIT(2) for save-inhibit.
> > >
> > > N.B. Daniele mentioned this bit mbz for ICL, and has been moved into the
> > > submission process rather than the context image.
> > >
> > > Suggested-by: Daniele Ceraolo Spurio <daniele.ceraolospurio at intel.com>
> > > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > > Cc: Michal Winiarski <michal.winiarski at intel.com>
> > > Cc: Michel Thierry <michel.thierry at intel.com>
> > > Cc: Michal Wajdeczko <michal.wajdeczko at intel.com>
> > > Cc: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
> > > Cc: Mika Kuoppala <mika.kuoppala at intel.com>
> > > Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio at intel.com>
> > > ---
> > > drivers/gpu/drm/i915/intel_lrc.c | 4 ++++
> > > drivers/gpu/drm/i915/intel_lrc.h | 1 +
> > > 2 files changed, 5 insertions(+)
> > >
> > > diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
> > > index 22d471a4228d..c28f267a8417 100644
> > > --- a/drivers/gpu/drm/i915/intel_lrc.c
> > > +++ b/drivers/gpu/drm/i915/intel_lrc.c
> > > @@ -2285,6 +2285,10 @@ populate_lr_context(struct i915_gem_context *ctx,
> > > if (!engine->default_state)
> > > regs[CTX_CONTEXT_CONTROL + 1] |=
> > > _MASKED_BIT_ENABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT);
> > > + if (ctx->hw_id == PREEMPT_ID)
> > > + regs[CTX_CONTEXT_CONTROL + 1] |=
> > > + _MASKED_BIT_ENABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT |
> > > + CTX_CTRL_ENGINE_CTX_SAVE_INHIBIT);
> > >
> >
> > This shouldn't break anything and ICL is not merged yet (plus things may
> > still change) so if you want to merge this,
>
> I think it's the right thing conceptually to do.
Uhoh, seeing GPU hangs on Broxton when thrashing preemption.
Is the bit still valid?
-Chris
More information about the Intel-gfx
mailing list