Re: [Intel-gfx] [PATCH] drm/i915/guc: Cancel GuC engine busyness worker synchronously
On Fri, 26 Aug 2022 17:21:35 -0700, Umesh Nerlige Ramappa wrote: > > The worker is canceled in gt_park path, but earlier it was assumed that > gt_park path cannot sleep and the cancel is asynchronous. This caused a > race with suspend flow where the worker runs after suspend and causes an > unclaimed register access warning. Cancel the worker synchronously since > the gt_park is indeed allowed to sleep. Indeed, __gt_park already calls cancel_work_sync and synchronize_irq which can sleep: Reviewed-by: Ashutosh Dixit > v2: Fix author name and sign-off mismatch > > Signed-off-by: Umesh Nerlige Ramappa > Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/4419 > Fixes: 77cdd054dd2c ("drm/i915/pmu: Connect engine busyness stats from GuC to > pmu") > --- > drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c | 7 ++- > 1 file changed, 6 insertions(+), 1 deletion(-) > > 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 0d56b615bf78..e6275380b253 100644 > --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c > +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c > @@ -1438,7 +1438,12 @@ void intel_guc_busyness_park(struct intel_gt *gt) > if (!guc_submission_initialized(guc)) > return; > > - cancel_delayed_work(>timestamp.work); > + /* > + * There is a race with suspend flow where the worker runs after suspend > + * and causes an unclaimed register access warning. Cancel the worker > + * synchronously here. > + */ > + cancel_delayed_work_sync(>timestamp.work); > > /* >* Before parking, we should sample engine busyness stats if we need to. > -- > 2.25.1 >
[Intel-gfx] [PATCH] drm/i915/guc: Cancel GuC engine busyness worker synchronously
The worker is canceled in gt_park path, but earlier it was assumed that gt_park path cannot sleep and the cancel is asynchronous. This caused a race with suspend flow where the worker runs after suspend and causes an unclaimed register access warning. Cancel the worker synchronously since the gt_park is indeed allowed to sleep. v2: Fix author name and sign-off mismatch Signed-off-by: Umesh Nerlige Ramappa Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/4419 Fixes: 77cdd054dd2c ("drm/i915/pmu: Connect engine busyness stats from GuC to pmu") --- drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c | 7 ++- 1 file changed, 6 insertions(+), 1 deletion(-) 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 0d56b615bf78..e6275380b253 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c @@ -1438,7 +1438,12 @@ void intel_guc_busyness_park(struct intel_gt *gt) if (!guc_submission_initialized(guc)) return; - cancel_delayed_work(>timestamp.work); + /* +* There is a race with suspend flow where the worker runs after suspend +* and causes an unclaimed register access warning. Cancel the worker +* synchronously here. +*/ + cancel_delayed_work_sync(>timestamp.work); /* * Before parking, we should sample engine busyness stats if we need to. -- 2.25.1
[Intel-gfx] [PATCH] drm/i915/guc: Cancel GuC engine busyness worker synchronously
The worker is canceled in gt_park path, but earlier it was assumed that gt_park path cannot sleep and the cancel is asynchronous. This caused a race with suspend flow where the worker runs after suspend and causes an unclaimed register access warning. Cancel the worker synchronously since the gt_park is indeed allowed to sleep. Signed-off-by: Umesh Nerlige Ramappa --- drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c | 7 ++- 1 file changed, 6 insertions(+), 1 deletion(-) 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 0d56b615bf78..e6275380b253 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c @@ -1438,7 +1438,12 @@ void intel_guc_busyness_park(struct intel_gt *gt) if (!guc_submission_initialized(guc)) return; - cancel_delayed_work(>timestamp.work); + /* +* There is a race with suspend flow where the worker runs after suspend +* and causes an unclaimed register access warning. Cancel the worker +* synchronously here. +*/ + cancel_delayed_work_sync(>timestamp.work); /* * Before parking, we should sample engine busyness stats if we need to. -- 2.25.1
Re: [Intel-gfx] [PATCH] drm/i915/guc: Cancel GuC engine busyness worker synchronously
On 7/26/2022 13:51, Nerlige Ramappa, Umesh wrote: The worker is canceled in gt_park path, but earlier it was assumed that gt_park path cannot sleep and the cancel is asynchronous. This caused a race with suspend flow where the worker runs after suspend and causes an unclaimed register access warning. Cancel the worker synchronously since the gt_park is indeed allowed to sleep. Signed-off-by: Umesh Nerlige Ramappa Fixes: 77cdd054dd2c ("drm/i915/pmu: Connect engine busyness stats from GuC to pmu") Reviewed-by: John Harrison --- drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c | 7 ++- 1 file changed, 6 insertions(+), 1 deletion(-) 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 76916aed897a..0b7a5ecb640a 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c @@ -1438,7 +1438,12 @@ void intel_guc_busyness_park(struct intel_gt *gt) if (!guc_submission_initialized(guc)) return; - cancel_delayed_work(>timestamp.work); + /* +* There is a race with suspend flow where the worker runs after suspend +* and causes an unclaimed register access warning. Cancel the worker +* synchronously here. +*/ + cancel_delayed_work_sync(>timestamp.work); /* * Before parking, we should sample engine busyness stats if we need to.
[Intel-gfx] [PATCH] drm/i915/guc: Cancel GuC engine busyness worker synchronously
The worker is canceled in gt_park path, but earlier it was assumed that gt_park path cannot sleep and the cancel is asynchronous. This caused a race with suspend flow where the worker runs after suspend and causes an unclaimed register access warning. Cancel the worker synchronously since the gt_park is indeed allowed to sleep. Signed-off-by: Umesh Nerlige Ramappa Fixes: 77cdd054dd2c ("drm/i915/pmu: Connect engine busyness stats from GuC to pmu") --- drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c | 7 ++- 1 file changed, 6 insertions(+), 1 deletion(-) 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 76916aed897a..0b7a5ecb640a 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c @@ -1438,7 +1438,12 @@ void intel_guc_busyness_park(struct intel_gt *gt) if (!guc_submission_initialized(guc)) return; - cancel_delayed_work(>timestamp.work); + /* +* There is a race with suspend flow where the worker runs after suspend +* and causes an unclaimed register access warning. Cancel the worker +* synchronously here. +*/ + cancel_delayed_work_sync(>timestamp.work); /* * Before parking, we should sample engine busyness stats if we need to. -- 2.36.1