Quoting Michal Wajdeczko (2019-05-17 17:54:53) > On Fri, 17 May 2019 18:27:44 +0200, Chris Wilson > <chris@xxxxxxxxxxxxxxxxxx> wrote: > > > Quoting Michal Wajdeczko (2019-05-17 17:22:27) > >> When we reset engines using ALL_ENGINES mask, we will do > >> full GPU reset and GuC will be also affected. Let GuC be > >> prepared for upcoming reset. > >> > >> Signed-off-by: Michal Wajdeczko <michal.wajdeczko@xxxxxxxxx> > >> Cc: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> > >> Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@xxxxxxxxx> > >> --- > >> drivers/gpu/drm/i915/gt/intel_reset.c | 4 ++++ > >> 1 file changed, 4 insertions(+) > >> > >> diff --git a/drivers/gpu/drm/i915/gt/intel_reset.c > >> b/drivers/gpu/drm/i915/gt/intel_reset.c > >> index 464369bc55ad..ca6e40b6b4e2 100644 > >> --- a/drivers/gpu/drm/i915/gt/intel_reset.c > >> +++ b/drivers/gpu/drm/i915/gt/intel_reset.c > >> @@ -564,6 +564,10 @@ static int gen8_reset_engines(struct > >> drm_i915_private *i915, > >> */ > >> } > >> > >> + /* We are about to do full GPU reset, don't forget about GuC */ > >> + if (engine_mask == ALL_ENGINES) > >> + intel_uc_reset_prepare(i915); > > > > Eh, this is done in reset_prepare already. The only other path to call > > intel_gpu_reset() directly is along sanitization, which should also have > > already sanitized the guc as well. No? > > There is igt_atomic_reset selftest which does not call reset_prepare. > And since we lost GuC in gen6_reset_engines due to GEN6_GRDOM_FULL, > our later graceful goodbye with GuC was not working. > > This is hidden with current GuC fw, but with new ICL fw with CT is was > visible as: Imagine we fix the selftest, is there a GEM_BUG_ON(intel_uc_active()) we could put here to enforce sanitization first? Does such a assert want to be here or up a level in intel_gpu_reset()? -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
