drm/i915: Enable irq to trace batch buffer completion.

If we trigger a tracepoint for batch buffer submission, it is a reasonable
assumption that we wish to also trace the batch buffer completion. So in
order to capture the completion events, we need to enable irqs... However,
we cannot rely on the completion event to disable the irq later, so we
defer the irq disable to the retire request.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

5 files changed, 21 insertions, 1 deletions

diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 45d507ebd3f..92aeb918e0c 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1468,6 +1468,7 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	spin_lock_init(&dev_priv->user_irq_lock);
 	spin_lock_init(&dev_priv->error_lock);
 	dev_priv->user_irq_refcount = 0;
+	dev_priv->trace_irq_seqno = 0;
 
 	ret = drm_vblank_init(dev, I915_NUM_PIPE);
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index b24b2d145b7..6035d3dae85 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -202,6 +202,7 @@ typedef struct drm_i915_private {
 	spinlock_t user_irq_lock;
 	/** Refcount for i915_user_irq_get() versus i915_user_irq_put(). */
 	int user_irq_refcount;
+	u32 trace_irq_seqno;
 	/** Cached value of IMR to avoid reads in updating the bitfield */
 	u32 irq_mask_reg;
 	u32 pipestat[2];
@@ -665,6 +666,7 @@ extern int i915_irq_emit(struct drm_device *dev, void *data,
 extern int i915_irq_wait(struct drm_device *dev, void *data,
 			 struct drm_file *file_priv);
 void i915_user_irq_get(struct drm_device *dev);
+void i915_trace_irq_get(struct drm_device *dev, u32 seqno);
 void i915_user_irq_put(struct drm_device *dev);
 extern void i915_enable_interrupt (struct drm_device *dev);
 
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index b5f9df230d0..abfc27b0c2e 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1770,7 +1770,7 @@ i915_gem_retire_requests(struct drm_device *dev)
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	uint32_t seqno;
 
-	if (!dev_priv->hw_status_page)
+	if (!dev_priv->hw_status_page || list_empty(&dev_priv->mm.request_list))
 		return;
 
 	seqno = i915_get_gem_seqno(dev);
@@ -1794,6 +1794,12 @@ i915_gem_retire_requests(struct drm_device *dev)
 		} else
 			break;
 	}
+
+	if (unlikely (dev_priv->trace_irq_seqno &&
+		      i915_seqno_passed(dev_priv->trace_irq_seqno, seqno))) {
+		i915_user_irq_put(dev);
+		dev_priv->trace_irq_seqno = 0;
+	}
 }
 
 void
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 4dfeec7cdd4..c3ceffa46ea 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -725,6 +725,16 @@ void i915_user_irq_put(struct drm_device *dev)
 	spin_unlock_irqrestore(&dev_priv->user_irq_lock, irqflags);
 }
 
+void i915_trace_irq_get(struct drm_device *dev, u32 seqno)
+{
+	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
+
+	if (dev_priv->trace_irq_seqno == 0)
+		i915_user_irq_get(dev);
+
+	dev_priv->trace_irq_seqno = seqno;
+}
+
 static int i915_wait_irq(struct drm_device * dev, int irq_nr)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h
index 908b3c4d8cf..01840d9bc38 100644
--- a/drivers/gpu/drm/i915/i915_trace.h
+++ b/drivers/gpu/drm/i915/i915_trace.h
@@ -165,6 +165,7 @@ TRACE_EVENT(i915_gem_request_submit,
 	    TP_fast_assign(
 			   __entry->dev = dev->primary->index;
 			   __entry->seqno = seqno;
+			   i915_trace_irq_get(dev, seqno);
 			   ),
 
 	    TP_printk("dev=%u, seqno=%u", __entry->dev, __entry->seqno)