[Intel-gfx] [PATCH 3/3] drm/i915: Improve long running OCL w/a for GuC submission

[John.C.Harrison at Intel.com]

From: John Harrison <John.C.Harrison at Intel.com>

A workaround was added to the driver to allow OpenCL workloads to run
'forever' by disabling pre-emption on the RCS engine for Gen12.
It is not totally unbound as the heartbeat will kick in eventually
and cause a reset of the hung engine.

However, this does not work well in GuC submission mode. In GuC mode,
the pre-emption timeout is how GuC detects hung contexts and triggers
a per engine reset. Thus, disabling the timeout means also losing all
per engine reset ability. A full GT reset will still occur when the
heartbeat finally expires, but that is a much more destructive and
undesirable mechanism.

The purpose of the workaround is actually to give OpenCL tasks longer
to reach a pre-emption point after a pre-emption request has been
issued. This is necessary because Gen12 does not support mid-thread
pre-emption and OpenCL can have long running threads.

So, rather than disabling the timeout completely, just set it to a
'long' value.

CC: Michal Mrozek <michal.mrozek at intel.com>
Signed-off-by: John Harrison <John.C.Harrison at Intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio at intel.com>
---
 drivers/gpu/drm/i915/gt/intel_engine_cs.c | 22 +++++++++++++++++++---
 1 file changed, 19 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/i915/gt/intel_engine_cs.c b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
index 2a1e9f36e6f5..64249301a227 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine_cs.c
+++ b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
@@ -385,9 +385,25 @@ static int intel_engine_setup(struct intel_gt *gt, enum intel_engine_id id,
 	engine->props.timeslice_duration_ms =
 		CONFIG_DRM_I915_TIMESLICE_DURATION;
 
-	/* Override to uninterruptible for OpenCL workloads. */
-	if (GRAPHICS_VER(i915) == 12 && engine->class == RENDER_CLASS)
-		engine->props.preempt_timeout_ms = 0;
+	/*
+	 * Mid-thread pre-emption is not available in Gen12. Unfortunately,
+	 * some OpenCL workloads run quite long threads. That means they get
+	 * reset due to not pre-empting in a timely manner. So, bump the
+	 * pre-emption timeout value to be much higher for compute engines.
+	 * Using three times the heartbeat period seems long enough for a
+	 * reasonable task to reach a pre-emption point but not so long as to
+	 * allow genuine hangs to go unresolved.
+	 */
+	if (GRAPHICS_VER(i915) == 12 && engine->class == RENDER_CLASS) {
+		unsigned long triple_beat = engine->props.heartbeat_interval_ms * 3;
+
+		if (triple_beat > engine->props.preempt_timeout_ms) {
+			drm_info(&gt->i915->drm, "Bumping pre-emption timeout from %ld to %ld on %s to allow slow compute pre-emption\n",
+				 engine->props.preempt_timeout_ms, triple_beat, engine->name);
+
+			engine->props.preempt_timeout_ms = triple_beat;
+		}
+	}
 
 	/* Cap timeouts to prevent overflow inside GuC */
 	if (intel_guc_submission_is_wanted(&gt->uc.guc)) {
-- 
2.25.1