[v2] drm/i915: Shrink search list for active timelines

Commit Message

Chris Wilson May 15, 2018, 2:31 p.m. UTC

When switching to the kernel context, we force the switch to occur after
all currently active requests (so that we know the GPU won't switch
immediately away and the kernel context remains current as we work). To
do so we have to inspect all the timelines and add a fence from the
active work to queue our switch afterwards. We can use the tracked set
of active rings to shrink our search for active timelines.

v2: Use a local to shrink the list_for_each_entry()

Signed-off-by: Chris Wilson <[e-mail]>
Cc: Tvrtko Ursulin <[e-mail]>
Reviewed-by: Tvrtko Ursulin <[e-mail]>
---
 drivers/gpu/drm/i915/i915_gem_context.c | 25 ++++++++++++++-----------
 1 file changed, 14 insertions(+), 11 deletions(-)

Comments

Chris Wilson May 16, 2018, 6:32 a.m. UTC | #1

Quoting Chris Wilson (2018-05-15 15:31:49)
> When switching to the kernel context, we force the switch to occur after
> all currently active requests (so that we know the GPU won't switch
> immediately away and the kernel context remains current as we work). To
> do so we have to inspect all the timelines and add a fence from the
> active work to queue our switch afterwards. We can use the tracked set
> of active rings to shrink our search for active timelines.
> 
> v2: Use a local to shrink the list_for_each_entry()
> 
> Signed-off-by: Chris Wilson <[e-mail]>
> Cc: Tvrtko Ursulin <[e-mail]>
> Reviewed-by: Tvrtko Ursulin <[e-mail]>

And pushed, thanks kindly for the review.
-Chris

Patch

diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
index 33f8a4b3c981..4bf18b5c6f1d 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/i915_gem_context.c
@@ -596,41 +596,44 @@  last_request_on_engine(struct i915_timeline *timeline,
 
 static bool engine_has_idle_kernel_context(struct intel_engine_cs *engine)
 {
-	struct i915_timeline *timeline;
+	struct list_head * const active_rings = &engine->i915->gt.active_rings;
+	struct intel_ring *ring;
 
-	list_for_each_entry(timeline, &engine->i915->gt.timelines, link) {
-		if (last_request_on_engine(timeline, engine))
+	lockdep_assert_held(&engine->i915->drm.struct_mutex);
+
+	list_for_each_entry(ring, active_rings, active_link) {
+		if (last_request_on_engine(ring->timeline, engine))
 			return false;
 	}
 
 	return intel_engine_has_kernel_context(engine);
 }
 
-int i915_gem_switch_to_kernel_context(struct drm_i915_private *dev_priv)
+int i915_gem_switch_to_kernel_context(struct drm_i915_private *i915)
 {
 	struct intel_engine_cs *engine;
-	struct i915_timeline *timeline;
 	enum intel_engine_id id;
 
-	lockdep_assert_held(&dev_priv->drm.struct_mutex);
+	lockdep_assert_held(&i915->drm.struct_mutex);
 
-	i915_retire_requests(dev_priv);
+	i915_retire_requests(i915);
 
-	for_each_engine(engine, dev_priv, id) {
+	for_each_engine(engine, i915, id) {
+		struct intel_ring *ring;
 		struct i915_request *rq;
 
 		if (engine_has_idle_kernel_context(engine))
 			continue;
 
-		rq = i915_request_alloc(engine, dev_priv->kernel_context);
+		rq = i915_request_alloc(engine, i915->kernel_context);
 		if (IS_ERR(rq))
 			return PTR_ERR(rq);
 
 		/* Queue this switch after all other activity */
-		list_for_each_entry(timeline, &dev_priv->gt.timelines, link) {
+		list_for_each_entry(ring, &i915->gt.active_rings, active_link) {
 			struct i915_request *prev;
 
-			prev = last_request_on_engine(timeline, engine);
+			prev = last_request_on_engine(ring->timeline, engine);
 			if (prev)
 				i915_sw_fence_await_sw_fence_gfp(&rq->submit,
 								 &prev->submit,