[PATCH 70/73] active-request

[Chris Wilson]

---
 drivers/gpu/drm/i915/gt/intel_engine.h        | 19 ----------
 drivers/gpu/drm/i915/gt/intel_engine_cs.c     | 37 +++----------------
 .../drm/i915/gt/intel_execlists_submission.c  | 24 ++++++++++++
 .../gpu/drm/i915/gt/intel_ring_scheduler.c    | 24 ++++++++++++
 .../gpu/drm/i915/gt/intel_ring_submission.c   | 21 +++++++++++
 drivers/gpu/drm/i915/i915_scheduler_types.h   |  2 +
 6 files changed, 76 insertions(+), 51 deletions(-)

diff --git a/drivers/gpu/drm/i915/gt/intel_engine.h b/drivers/gpu/drm/i915/gt/intel_engine.h
index fa257a305143..ce24ac6f58cb 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine.h
+++ b/drivers/gpu/drm/i915/gt/intel_engine.h
@@ -103,25 +103,6 @@ execlists_num_ports(const struct intel_engine_execlists * const execlists)
 	return execlists->port_mask + 1;
 }
 
-static inline struct i915_request *
-execlists_active(const struct intel_engine_execlists *execlists)
-{
-	struct i915_request * const *cur, * const *old, *active;
-
-	cur = READ_ONCE(execlists->active);
-	smp_rmb(); /* pairs with overwrite protection in process_csb() */
-	do {
-		old = cur;
-
-		active = READ_ONCE(*cur);
-		cur = READ_ONCE(execlists->active);
-
-		smp_rmb(); /* and complete the seqlock retry */
-	} while (unlikely(cur != old));
-
-	return active;
-}
-
 static inline u32
 intel_read_status_page(const struct intel_engine_cs *engine, int reg)
 {
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_cs.c b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
index e25772d8b0cd..c21243106b0f 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine_cs.c
+++ b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
@@ -1734,10 +1734,8 @@ ktime_t intel_engine_get_busy_time(struct intel_engine_cs *engine, ktime_t *now)
 }
 
 struct i915_request *
-intel_engine_find_active_request(struct intel_engine_cs *engine)
+i915_sched_engine_active_request(struct i915_sched_engine *se)
 {
-	struct i915_request *request, *active = NULL;
-
 	/*
 	 * We are called by the error capture, reset and to dump engine
 	 * state at random points in time. In particular, note that neither is
@@ -1749,36 +1747,11 @@ intel_engine_find_active_request(struct intel_engine_cs *engine)
 	 * At all other times, we must assume the GPU is still running, but
 	 * we only care about the snapshot of this moment.
 	 */
-	lockdep_assert_held(&engine->active.lock);
+	lockdep_assert_held(&se->lock);
+	if (!se->active_request)
+		return NULL;
 
-	rcu_read_lock();
-	request = execlists_active(&engine->execlists);
-	if (request) {
-		struct intel_timeline *tl = request->context->timeline;
-
-		list_for_each_entry_from_reverse(request, &tl->requests, link) {
-			if (__i915_request_is_complete(request))
-				break;
-
-			active = request;
-		}
-	}
-	rcu_read_unlock();
-	if (active)
-		return active;
-
-	list_for_each_entry(request, &engine->active.requests, sched.link) {
-		if (i915_request_completed(request))
-			continue;
-
-		if (!i915_request_started(request))
-			break;
-
-		active = request;
-		break;
-	}
-
-	return active;
+	return se->active_request(se);
 }
 
 ktime_t intel_runtime_stats_get_busy_time(const struct intel_runtime_stats *st)
diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
index 2387cde2426c..8ef6dd158423 100644
--- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
+++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
@@ -2907,6 +2907,29 @@ static void rcs_submission_override(struct intel_engine_cs *engine)
 	}
 }
 
+static struct i915_request *
+execlists_active_request(struct i915_sched_engine *se)
+{
+	struct intel_engine_cs *engine = container_of(se, typeof(*se), active);
+	struct i915_request *rq, *active = NULL;
+
+	rcu_read_lock();
+	rq = *READ_ONCE(engine->execlists.active);
+	if (rq) {
+		struct intel_timeline *tl = rq->context->timeline;
+
+		list_for_each_entry_from_reverse(rq, &tl->requests, link) {
+			if (__i915_request_is_complete(rq))
+				break;
+
+			active = rq;
+		}
+	}
+	rcu_read_unlock();
+
+	return active;
+}
+
 int intel_execlists_submission_setup(struct intel_engine_cs *engine)
 {
 	struct intel_engine_execlists * const el = &engine->execlists;
@@ -2918,6 +2941,7 @@ int intel_execlists_submission_setup(struct intel_engine_cs *engine)
 		     execlists_submission_tasklet, (unsigned long)engine);
 	timer_setup(&el->timer, execlists_timeslice, 0);
 	timer_setup(&el->preempt, execlists_preempt, 0);
+	engine->active.active_request = execlists_active_request;
 
 	logical_ring_default_vfuncs(engine);
 	logical_ring_default_irqs(engine);
diff --git a/drivers/gpu/drm/i915/gt/intel_ring_scheduler.c b/drivers/gpu/drm/i915/gt/intel_ring_scheduler.c
index 5080f106f913..21793a502866 100644
--- a/drivers/gpu/drm/i915/gt/intel_ring_scheduler.c
+++ b/drivers/gpu/drm/i915/gt/intel_ring_scheduler.c
@@ -1144,6 +1144,29 @@ static int gen7_ctx_switch_bb_init(struct intel_engine_cs *engine)
 	return err;
 }
 
+static struct i915_request *
+active_request(struct i915_sched_engine *se)
+{
+	struct intel_engine_cs *engine = container_of(se, typeof(*se), active);
+	struct i915_request *rq, *active = NULL;
+
+	rcu_read_lock();
+	rq = *READ_ONCE(engine->execlists.active);
+	if (rq) {
+		struct intel_timeline *tl = rq->context->timeline;
+
+		list_for_each_entry_from_reverse(rq, &tl->requests, link) {
+			if (__i915_request_is_complete(rq))
+				break;
+
+			active = rq;
+		}
+	}
+	rcu_read_unlock();
+
+	return active;
+}
+
 int intel_ring_scheduler_setup(struct intel_engine_cs *engine)
 {
 	struct intel_ring *ring;
@@ -1153,6 +1176,7 @@ int intel_ring_scheduler_setup(struct intel_engine_cs *engine)
 
 	tasklet_init(&engine->active.tasklet,
 		     submission_tasklet, (unsigned long)engine);
+	engine->active.active_request = active_request;
 
 	setup_common(engine);
 
diff --git a/drivers/gpu/drm/i915/gt/intel_ring_submission.c b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
index 3875e006668c..6a3a8bf8be60 100644
--- a/drivers/gpu/drm/i915/gt/intel_ring_submission.c
+++ b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
@@ -1043,6 +1043,25 @@ static void setup_irq(struct intel_engine_cs *engine)
 	}
 }
 
+static struct i915_request *
+legacy_active_request(struct i915_sched_engine *se)
+{
+	struct intel_engine_cs *engine = container_of(se, typeof(*se), active);
+	struct i915_request *rq;
+
+	list_for_each_entry(rq, &engine->active.requests, sched.link) {
+		if (__i915_request_is_complete(rq))
+			continue;
+
+		if (!__i915_request_has_started(rq))
+			break;
+
+		return rq;
+	}
+
+	return NULL;
+}
+
 static void setup_common(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *i915 = engine->i915;
@@ -1050,6 +1069,8 @@ static void setup_common(struct intel_engine_cs *engine)
 	/* gen8+ are only supported with execlists */
 	GEM_BUG_ON(INTEL_GEN(i915) >= 8);
 
+	engine->active.active_request = legacy_active_request;
+
 	setup_irq(engine);
 
 	engine->resume = intel_ring_submission_resume;
diff --git a/drivers/gpu/drm/i915/i915_scheduler_types.h b/drivers/gpu/drm/i915/i915_scheduler_types.h
index b8c484511185..a65354975074 100644
--- a/drivers/gpu/drm/i915/i915_scheduler_types.h
+++ b/drivers/gpu/drm/i915/i915_scheduler_types.h
@@ -130,6 +130,8 @@ struct i915_sched_engine {
 	 * @tasklet: softirq tasklet for bottom half
 	 */
 	struct tasklet_struct tasklet;
+
+	struct i915_request *(*active_request)(struct i915_sched_engine *se);
 };
 
 struct i915_dependency {
-- 
2.20.1