On Thu, Jan 23, 2025 at 11:38:39AM -0800, Umesh Nerlige Ramappa wrote: > When running igt@gem_exec_balancer@individual for multiple iterations, > it is seen that the delta busyness returned by PMU is 0. The issue stems > from a combination of 2 implementation specific details: > > 1) gt_park is throttling __update_guc_busyness_stats() so that it does > not hog PCI bandwidth for some use cases. (Ref: 59bcdb564b3ba) > > 2) busyness implementation always returns monotonically increasing > counters. (Ref: cf907f6d29421) > > If an application queried an engine while it was active, > engine->stats.guc.running is set to true. Following that, if all PM > wakeref's are released, then gt is parked. At this time the throttling > of __update_guc_busyness_stats() may result in a missed update to the > running state of the engine (due to (1) above). This means subsequent > calls to guc_engine_busyness() will think that the engine is still > running and they will keep updating the cached counter (stats->total). > This results in an inflated cached counter. > > Later when the application runs a workload and queries for busyness, we > return the cached value since it is larger than the actual value (due to > (2) above) > > All subsequent queries will return the same large (inflated) value, so > the application sees a delta busyness of zero. > > Fix the issue by resetting the running state of engines each time > intel_guc_busyness_park() is called. > > v2: (Rodrigo) > - Use the correct tag in commit message > - Drop the redundant wakeref check in guc_engine_busyness() and update > commit message Thank you > > Closes: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/13366 > Fixes: cf907f6d2942 ("i915/guc: Ensure busyness counter increases motonically") > Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@xxxxxxxxx> > --- > .../gpu/drm/i915/gt/uc/intel_guc_submission.c | 16 ++++++++++++++++ > 1 file changed, 16 insertions(+) > > diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c > index 3b1333a24a89..a33b67b83dc1 100644 > --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c > +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c > @@ -1469,6 +1469,19 @@ static void __reset_guc_busyness_stats(struct intel_guc *guc) > spin_unlock_irqrestore(&guc->timestamp.lock, flags); > } > > +static void __update_guc_busyness_running_state(struct intel_guc *guc) I hate those '__', but I forgot to spot in the previous email and I know I know... there's a lot of those in this file already :/ I was thinking of another name for this function as well since it is only stopping the running state, but I'm bad with naming... let's move on and close the real issue Reviewed-by: Rodrigo Vivi <rodrigo.vivi@xxxxxxxxx> > +{ > + struct intel_gt *gt = guc_to_gt(guc); > + struct intel_engine_cs *engine; > + enum intel_engine_id id; > + unsigned long flags; > + > + spin_lock_irqsave(&guc->timestamp.lock, flags); > + for_each_engine(engine, gt, id) > + engine->stats.guc.running = false; > + spin_unlock_irqrestore(&guc->timestamp.lock, flags); > +} > + > static void __update_guc_busyness_stats(struct intel_guc *guc) > { > struct intel_gt *gt = guc_to_gt(guc); > @@ -1619,6 +1632,9 @@ void intel_guc_busyness_park(struct intel_gt *gt) > if (!guc_submission_initialized(guc)) > return; > > + /* Assume no engines are running and set running state to false */ > + __update_guc_busyness_running_state(guc); > + > /* > * There is a race with suspend flow where the worker runs after suspend > * and causes an unclaimed register access warning. Cancel the worker > -- > 2.38.1 >
