[PATCH 12/12] exec-active-count

[Chris Wilson]

---
 drivers/gpu/drm/i915/i915_irq.c         |  6 ++++--
 drivers/gpu/drm/i915/intel_engine_cs.c  |  1 +
 drivers/gpu/drm/i915/intel_lrc.c        | 20 ++++++++++++++++----
 drivers/gpu/drm/i915/intel_ringbuffer.h |  6 ++++++
 4 files changed, 27 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index b1296a55c1e4..f8205841868b 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -1388,8 +1388,10 @@ gen8_cs_irq_handler(struct intel_engine_cs *engine, u32 iir, int test_shift)
 	bool tasklet = false;
 
 	if (iir & (GT_CONTEXT_SWITCH_INTERRUPT << test_shift)) {
-		__set_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted);
-		tasklet = true;
+		if (READ_ONCE(engine->execlists.active)) {
+			__set_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted);
+			tasklet = true;
+		}
 	}
 
 	if (iir & (GT_RENDER_USER_INTERRUPT << test_shift)) {
diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c
index 8d6ae8753d7b..2473c9cb205a 100644
--- a/drivers/gpu/drm/i915/intel_engine_cs.c
+++ b/drivers/gpu/drm/i915/intel_engine_cs.c
@@ -1688,6 +1688,7 @@ void intel_engine_dump(struct intel_engine_cs *engine, struct drm_printer *m)
 					   idx);
 			}
 		}
+		drm_printf(m, "\t\tELSP active %d\n", execlists->active);
 		rcu_read_unlock();
 	} else if (INTEL_GEN(dev_priv) > 6) {
 		drm_printf(m, "\tPP_DIR_BASE: 0x%08x\n",
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
index 7f45dd7dc3e5..14187d148a0a 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -482,15 +482,23 @@ static bool can_merge_ctx(const struct i915_gem_context *prev,
 	return true;
 }
 
-static void port_assign(struct execlist_port *port,
+static bool port_assign(struct execlist_port *port,
 			struct drm_i915_gem_request *rq)
 {
+	bool was_idle;
+
 	GEM_BUG_ON(rq == port_request(port));
 
-	if (port_isset(port))
+	if (port_isset(port)) {
 		i915_gem_request_put(port_request(port));
+		was_idle = false;
+	} else {
+		was_idle = true;
+	}
 
 	port_set(port, port_pack(i915_gem_request_get(rq), port_count(port)));
+
+	return was_idle;
 }
 
 static void inject_preempt_context(struct intel_engine_cs *engine)
@@ -657,7 +665,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 				GEM_BUG_ON(last->ctx == rq->ctx);
 
 				if (submit)
-					port_assign(port, last);
+					execlists->active += port_assign(port, last);
 				port++;
 
 				GEM_BUG_ON(port_isset(port));
@@ -679,7 +687,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 done:
 	execlists->first = rb;
 	if (submit)
-		port_assign(port, last);
+		execlists->active += port_assign(port, last);
 unlock:
 	spin_unlock_irq(&engine->timeline->lock);
 
@@ -700,6 +708,7 @@ execlist_cancel_port_requests(struct intel_engine_execlists *execlists)
 		i915_gem_request_put(rq);
 
 		memset(port, 0, sizeof(*port));
+		execlists->active--;
 		port++;
 	}
 }
@@ -863,6 +872,8 @@ static void intel_lrc_irq_handler(unsigned long data)
 
 				GEM_BUG_ON(!execlists->preempt);
 				execlists->preempt = false;
+				execlists->active--;
+				GEM_BUG_ON(execlists->active);
 				continue;
 			}
 
@@ -1461,6 +1472,7 @@ static int gen8_init_common_ring(struct intel_engine_cs *engine)
 	clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted);
 	execlists->csb_head = -1;
 	execlists->preempt = false;
+	execlists->active = 0;
 
 	/* After a GPU reset, we may have requests to replay */
 	if (!i915_modparams.enable_guc_submission && execlists->first)
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index 17186f067408..bf121fb7cf85 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -245,6 +245,11 @@ struct intel_engine_execlists {
 	 */
 	bool preempt;
 
+	/**
+	 * @active: activity counter
+	 */
+	unsigned int active;
+
 	/**
 	 * @port_mask: number of execlist ports - 1
 	 */
@@ -541,6 +546,7 @@ execlists_port_complete(struct intel_engine_execlists * const execlists,
 
 	memmove(port, port + 1, m * sizeof(struct execlist_port));
 	memset(port + m, 0, sizeof(struct execlist_port));
+	execlists->active--;
 }
 
 static inline unsigned int
-- 
2.15.0.rc1