[PATCH 10/81] drm/i915: Delay queuing hangcheck to wait-request

[Chris Wilson]

We can forgo queuing the hangcheck from the start of every request to
until we wait upon a request. This reduces the overhead of every
request, but may increase the latency of detecting a hang. Howeever, if
nothing every waits upon a hang, did it ever hang? It also improves the
robustness of the wait-request by ensuring that the hangchecker is
indeed running before we sleep indefinitely (and thereby ensuring that
we never actually sleep forever waiting for a dead GPU).

v2: Also queue the hangcheck from retire work in case the GPU become
stuck when no one is watching.

Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/i915_drv.h |  2 +-
 drivers/gpu/drm/i915/i915_gem.c | 13 ++++++++-----
 drivers/gpu/drm/i915/i915_irq.c | 16 +++++++---------
 3 files changed, 16 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 6349f9bf7c78..0d059c811d6e 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -2810,7 +2810,7 @@ void intel_hpd_cancel_work(struct drm_i915_private *dev_priv);
 bool intel_hpd_pin_to_port(enum hpd_pin pin, enum port *port);
 
 /* i915_irq.c */
-void i915_queue_hangcheck(struct drm_device *dev);
+void i915_queue_hangcheck(struct drm_i915_private *dev_priv);
 __printf(3, 4)
 void i915_handle_error(struct drm_device *dev, u32 engine_mask,
 		       const char *fmt, ...);
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index cf50e3386de4..3f890f317ec3 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1308,6 +1308,9 @@ int __i915_wait_request(struct drm_i915_gem_request *req,
 			break;
 		}
 
+		/* Ensure that even if the GPU hangs, we get woken up. */
+		i915_queue_hangcheck(dev_priv);
+
 		timer.function = NULL;
 		if (timeout || missed_irq(dev_priv, engine)) {
 			unsigned long expire;
@@ -2652,8 +2655,6 @@ void __i915_add_request(struct drm_i915_gem_request *request,
 	/* Not allowed to fail! */
 	WARN(ret, "emit|add_request failed: %d!\n", ret);
 
-	i915_queue_hangcheck(engine->dev);
-
 	queue_delayed_work(dev_priv->wq,
 			   &dev_priv->mm.retire_work,
 			   round_jiffies_up_relative(HZ));
@@ -3025,8 +3026,8 @@ i915_gem_retire_requests(struct drm_device *dev)
 
 	if (idle)
 		mod_delayed_work(dev_priv->wq,
-				   &dev_priv->mm.idle_work,
-				   msecs_to_jiffies(100));
+				 &dev_priv->mm.idle_work,
+				 msecs_to_jiffies(100));
 
 	return idle;
 }
@@ -3045,9 +3046,11 @@ i915_gem_retire_work_handler(struct work_struct *work)
 		idle = i915_gem_retire_requests(dev);
 		mutex_unlock(&dev->struct_mutex);
 	}
-	if (!idle)
+	if (!idle) {
+		i915_queue_hangcheck(dev_priv);
 		queue_delayed_work(dev_priv->wq, &dev_priv->mm.retire_work,
 				   round_jiffies_up_relative(HZ));
+	}
 }
 
 static void
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index d4b1867b06d4..e37be627cef7 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -3169,10 +3169,10 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 	intel_uncore_arm_unclaimed_mmio_detection(dev_priv);
 
 	for_each_engine_id(engine, dev_priv, id) {
+		bool busy = waitqueue_active(&engine->irq_queue);
 		u64 acthd;
 		u32 seqno;
 		unsigned user_interrupts;
-		bool busy = true;
 
 		semaphore_clear_deadlocks(dev_priv);
 
@@ -3195,12 +3195,11 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 		if (engine->hangcheck.seqno == seqno) {
 			if (ring_idle(engine, seqno)) {
 				engine->hangcheck.action = HANGCHECK_IDLE;
-				if (waitqueue_active(&engine->irq_queue)) {
+				if (busy) {
 					/* Safeguard against driver failure */
 					user_interrupts = kick_waiters(engine);
 					engine->hangcheck.score += BUSY;
-				} else
-					busy = false;
+				}
 			} else {
 				/* We always increment the hangcheck score
 				 * if the ring is busy and still processing
@@ -3274,18 +3273,17 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 		goto out;
 	}
 
+	/* Reset timer in case GPU hangs without another request being added */
 	if (busy_count)
-		/* Reset timer case chip hangs without another request
-		 * being added */
-		i915_queue_hangcheck(dev);
+		i915_queue_hangcheck(dev_priv);
 
 out:
 	ENABLE_RPM_WAKEREF_ASSERTS(dev_priv);
 }
 
-void i915_queue_hangcheck(struct drm_device *dev)
+void i915_queue_hangcheck(struct drm_i915_private *dev_priv)
 {
-	struct i915_gpu_error *e = &to_i915(dev)->gpu_error;
+	struct i915_gpu_error *e = &dev_priv->gpu_error;
 
 	if (!i915.enable_hangcheck)
 		return;
-- 
2.8.0.rc3