[01/33] drm/i915/gt: Harden the heartbeat against a stuck driver
diff mbox series

Message ID 20200701084053.6086-1-chris@chris-wilson.co.uk
State New
Headers show
Series
  • [01/33] drm/i915/gt: Harden the heartbeat against a stuck driver
Related show

Commit Message

Chris Wilson July 1, 2020, 8:40 a.m. UTC
If the driver get stuck holding the kernel timeline, we cannot issue a
heartbeat and so fail to discover that the driver is indeed stuck and do
not issue a GPU reset (which would hopefully unstick the driver!).
Switch to using a trylock so that we can query if the heartbeat's
timelin mutex is locked elsewhere, and then use the timer to probe if it
remains stuck at the same spot for consecutive heartbeats, indicating
that the mutex has not been released and the engine has not progressed.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c | 14 ++++++++++++--
 drivers/gpu/drm/i915/gt/intel_engine_types.h     |  1 +
 2 files changed, 13 insertions(+), 2 deletions(-)

Comments

Mika Kuoppala July 2, 2020, 9:39 a.m. UTC | #1
Chris Wilson <chris@chris-wilson.co.uk> writes:

> If the driver get stuck holding the kernel timeline, we cannot issue a
> heartbeat and so fail to discover that the driver is indeed stuck and do
> not issue a GPU reset (which would hopefully unstick the driver!).
> Switch to using a trylock so that we can query if the heartbeat's
> timelin mutex is locked elsewhere, and then use the timer to probe if it

timeline

> remains stuck at the same spot for consecutive heartbeats, indicating
> that the mutex has not been released and the engine has not progressed.
>
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> ---
>  drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c | 14 ++++++++++++--
>  drivers/gpu/drm/i915/gt/intel_engine_types.h     |  1 +
>  2 files changed, 13 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
> index 8db7e93abde5..1663ab5c68a5 100644
> --- a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
> +++ b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
> @@ -65,6 +65,7 @@ static void heartbeat(struct work_struct *wrk)
>  		container_of(wrk, typeof(*engine), heartbeat.work.work);
>  	struct intel_context *ce = engine->kernel_context;
>  	struct i915_request *rq;
> +	unsigned long serial;
>  
>  	/* Just in case everything has gone horribly wrong, give it a kick */
>  	intel_engine_flush_submission(engine);
> @@ -122,10 +123,19 @@ static void heartbeat(struct work_struct *wrk)
>  		goto out;
>  	}
>  
> -	if (engine->wakeref_serial == engine->serial)
> +	serial = READ_ONCE(engine->serial);
> +	if (engine->wakeref_serial == serial)
>  		goto out;
>  
> -	mutex_lock(&ce->timeline->mutex);
> +	if (!mutex_trylock(&ce->timeline->mutex)) {
> +		/* Unable to lock the kernel timeline, is the engine stuck? */
> +		if (xchg(&engine->heartbeat.blocked, serial) == serial)
> +			intel_gt_handle_error(engine->gt, engine->mask,
> +					      I915_ERROR_CAPTURE,
> +					      "stopped heartbeat on %s",
> +					      engine->name);
> +		goto out;
> +	}

This should do the trick. I was worried on the submit signal (fence) block being
empty above in this function that what happens if we never manage to
submit.

But this should cover that case also.

Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

>  
>  	intel_context_enter(ce);
>  	rq = __i915_request_create(ce, GFP_NOWAIT | __GFP_NOWARN);
> diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h
> index 073c3769e8cc..490af81bd6f3 100644
> --- a/drivers/gpu/drm/i915/gt/intel_engine_types.h
> +++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h
> @@ -348,6 +348,7 @@ struct intel_engine_cs {
>  	struct {
>  		struct delayed_work work;
>  		struct i915_request *systole;
> +		unsigned long blocked;
>  	} heartbeat;
>  
>  	unsigned long serial;
> -- 
> 2.20.1
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Patch
diff mbox series

diff --git a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
index 8db7e93abde5..1663ab5c68a5 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
+++ b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
@@ -65,6 +65,7 @@  static void heartbeat(struct work_struct *wrk)
 		container_of(wrk, typeof(*engine), heartbeat.work.work);
 	struct intel_context *ce = engine->kernel_context;
 	struct i915_request *rq;
+	unsigned long serial;
 
 	/* Just in case everything has gone horribly wrong, give it a kick */
 	intel_engine_flush_submission(engine);
@@ -122,10 +123,19 @@  static void heartbeat(struct work_struct *wrk)
 		goto out;
 	}
 
-	if (engine->wakeref_serial == engine->serial)
+	serial = READ_ONCE(engine->serial);
+	if (engine->wakeref_serial == serial)
 		goto out;
 
-	mutex_lock(&ce->timeline->mutex);
+	if (!mutex_trylock(&ce->timeline->mutex)) {
+		/* Unable to lock the kernel timeline, is the engine stuck? */
+		if (xchg(&engine->heartbeat.blocked, serial) == serial)
+			intel_gt_handle_error(engine->gt, engine->mask,
+					      I915_ERROR_CAPTURE,
+					      "stopped heartbeat on %s",
+					      engine->name);
+		goto out;
+	}
 
 	intel_context_enter(ce);
 	rq = __i915_request_create(ce, GFP_NOWAIT | __GFP_NOWARN);
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_types.h b/drivers/gpu/drm/i915/gt/intel_engine_types.h
index 073c3769e8cc..490af81bd6f3 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_engine_types.h
@@ -348,6 +348,7 @@  struct intel_engine_cs {
 	struct {
 		struct delayed_work work;
 		struct i915_request *systole;
+		unsigned long blocked;
 	} heartbeat;
 
 	unsigned long serial;