drm/i915/bdw: Help out the ctx switch interrupt handler

If we receive a storm of requests for the same context (see gem_storedw_loop_*)
we might end up iterating over too many elements in interrupt time, looking for
contexts to squash together. Instead, share the burden by giving more
intelligence to the queue function. At most, the interrupt will iterate over
three elements.

Signed-off-by: Oscar Mateo <oscar.mateo@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
[danvet: Checkpatch.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drivers/gpu/drm/i915/intel_lrc.c

@@ -392,10 +392,10 @@ static int execlists_context_queue(struct intel_engine_cs *ring,
 				   struct intel_context *to,
 				   u32 tail)
 {
-	struct intel_ctx_submit_request *req = NULL;
+	struct intel_ctx_submit_request *req = NULL, *cursor;
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 	unsigned long flags;
-	bool was_empty;
+	int num_elements = 0;
 
 	req = kzalloc(sizeof(*req), GFP_KERNEL);
 	if (req == NULL)
@@ -410,9 +410,27 @@ static int execlists_context_queue(struct intel_engine_cs *ring,
 
 	spin_lock_irqsave(&ring->execlist_lock, flags);
 
-	was_empty = list_empty(&ring->execlist_queue);
+	list_for_each_entry(cursor, &ring->execlist_queue, execlist_link)
+		if (++num_elements > 2)
+			break;
+
+	if (num_elements > 2) {
+		struct intel_ctx_submit_request *tail_req;
+
+		tail_req = list_last_entry(&ring->execlist_queue,
+					   struct intel_ctx_submit_request,
+					   execlist_link);
+
+		if (to == tail_req->ctx) {
+			WARN(tail_req->elsp_submitted != 0,
+			     "More than 2 already-submitted reqs queued\n");
+			list_del(&tail_req->execlist_link);
+			queue_work(dev_priv->wq, &tail_req->work);
+		}
+	}
+
 	list_add_tail(&req->execlist_link, &ring->execlist_queue);
-	if (was_empty)
+	if (num_elements == 0)
 		execlists_context_unqueue(ring);
 
 	spin_unlock_irqrestore(&ring->execlist_lock, flags);