[Intel-gfx] [PATCH 06/15] 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).

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 | 4 ++--
 drivers/gpu/drm/i915/i915_irq.c | 9 ++++-----
 3 files changed, 7 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index c24c23d8a0c0..e1980212ba37 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -2711,7 +2711,7 @@ 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, bool wedged,
 		       const char *fmt, ...);
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 646d189e23a1..f33c35c6130f 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1302,6 +1302,8 @@ int __i915_wait_request(struct drm_i915_gem_request *req,
 			break;
 		}
 
+		i915_queue_hangcheck(dev_priv);
+
 		timer.function = NULL;
 		if (timeout || missed_irq(dev_priv, ring)) {
 			unsigned long expire;
@@ -2579,8 +2581,6 @@ void __i915_add_request(struct drm_i915_gem_request *request,
 
 	trace_i915_gem_request_add(request);
 
-	i915_queue_hangcheck(ring->dev);
-
 	queue_delayed_work(dev_priv->wq,
 			   &dev_priv->mm.retire_work,
 			   round_jiffies_up_relative(HZ));
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 3b2a8fbe0392..28c0329f8281 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -3066,15 +3066,14 @@ static void i915_hangcheck_elapsed(struct work_struct *work)
 	if (rings_hung)
 		return i915_handle_error(dev, true, "Ring hung");
 
+	/* 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);
 }
 
-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;
-- 
2.6.2