[RFC PATCH 14/14] sched: implement usage tracking

[Paul Turner <pjt@google.com>]

With the frame-work for runnable tracking now fully in place.  Per-entity usage
tracking is a simple and low-overhead addition.

Signed-off-by: Paul Turner <pjt@google.com>
Signed-off-by: Ben Segall <bsegall@google.com>
---
 include/linux/sched.h |    1 +
 kernel/sched/debug.c  |    3 +++
 kernel/sched/fair.c   |   29 +++++++++++++++++++++++------
 kernel/sched/sched.h  |    4 ++--
 4 files changed, 29 insertions(+), 8 deletions(-)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 09b8c45..209185f 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1159,6 +1159,7 @@ struct load_weight {
 struct sched_avg {
 	u64 runnable_avg_sum, runnable_avg_period;
 	u64 last_runnable_update, decay_count;
+	u32 usage_avg_sum;
 	unsigned long load_avg_contrib;
 
 	int contributes_blocked_load;
diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index 0911ec6..4d39069 100644
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -93,6 +93,7 @@ static void print_cfs_group_stats(struct seq_file *m, int cpu, struct task_group
 	P(se->load.weight);
 	P(se->avg.runnable_avg_sum);
 	P(se->avg.runnable_avg_period);
+	P(se->avg.usage_avg_sum);
 	P(se->avg.load_avg_contrib);
 	P(se->avg.decay_count);
 #undef PN
@@ -228,6 +229,8 @@ void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
 			cfs_rq->tg_runnable_contrib);
 	SEQ_printf(m, "  .%-30s: %d\n", "tg->runnable_avg",
 			atomic_read(&cfs_rq->tg->runnable_avg));
+	SEQ_printf(m, "  .%-30s: %d\n", "tg->usage_avg",
+			atomic_read(&cfs_rq->tg->usage_avg));
 #endif
 
 	print_cfs_group_stats(m, cpu, cfs_rq->tg);
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index ad524bb..222c2c9 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -997,7 +997,8 @@ static u32 __compute_runnable_contrib(int n)
  */
 static __always_inline int __update_entity_runnable_avg(u64 now,
 							struct sched_avg *sa,
-							int runnable)
+							int runnable,
+							int running)
 {
 	u64 delta;
 	u32 periods, runnable_contrib;
@@ -1026,6 +1027,8 @@ static __always_inline int __update_entity_runnable_avg(u64 now,
 		delta_w = 1024 - delta_w;
 		if (runnable)
 			sa->runnable_avg_sum += delta_w;
+		if (running)
+			sa->usage_avg_sum += delta_w;
 		sa->runnable_avg_period += delta_w;
 
 		delta -= delta_w;
@@ -1036,15 +1039,20 @@ static __always_inline int __update_entity_runnable_avg(u64 now,
 						  periods + 1);
 		sa->runnable_avg_period = decay_load(sa->runnable_avg_period,
 						  periods + 1);
+		sa->usage_avg_sum = decay_load(sa->usage_avg_sum, periods + 1);
 
 		runnable_contrib = __compute_runnable_contrib(periods);
 		if (runnable)
 			sa->runnable_avg_sum += runnable_contrib;
+		if (running)
+			sa->usage_avg_sum += runnable_contrib;
 		sa->runnable_avg_period += runnable_contrib;
 	}
 
 	if (runnable)
 		sa->runnable_avg_sum += delta;
+	if (running)
+		sa->usage_avg_sum += delta;
 	sa->runnable_avg_period += delta;
 
 	return decayed;
@@ -1081,14 +1089,21 @@ static inline void __update_tg_runnable_avg(struct sched_avg *sa,
 						  struct cfs_rq *cfs_rq)
 {
 	struct task_group *tg = cfs_rq->tg;
-	long contrib;
+	long contrib, usage_contrib;
 
 	contrib = (sa->runnable_avg_sum << 12) / (sa->runnable_avg_period + 1);
 	contrib -= cfs_rq->tg_runnable_contrib;
 
-	if (abs(contrib) > cfs_rq->tg_runnable_contrib/64) {
+	usage_contrib = (sa->usage_avg_sum << 12) / (sa->runnable_avg_period + 1);
+	usage_contrib -= cfs_rq->tg_usage_contrib;
+
+	if ((abs(contrib) > cfs_rq->tg_runnable_contrib/64) ||
+	    (abs(usage_contrib) > cfs_rq->tg_usage_contrib/64)) {
 		atomic_add(contrib, &tg->runnable_avg);
 		cfs_rq->tg_runnable_contrib += contrib;
+
+		atomic_add(usage_contrib, &tg->usage_avg);
+		cfs_rq->tg_usage_contrib += usage_contrib;
 	}
 }
 
@@ -1164,8 +1179,8 @@ static inline void update_entity_load_avg(struct sched_entity *se,
 	struct cfs_rq *cfs_rq = cfs_rq_of(se);
 	long contrib_delta;
 
-	if(!__update_entity_runnable_avg(rq_of(cfs_rq)->clock_task, &se->avg,
-					se->on_rq))
+	if(!__update_entity_runnable_avg(cfs_rq_clock_task(cfs_rq), &se->avg,
+					 se->on_rq, cfs_rq->curr == se))
 		return;
 
 	contrib_delta = __update_entity_load_avg_contrib(se);
@@ -1210,7 +1225,8 @@ static void update_cfs_rq_blocked_load(struct cfs_rq *cfs_rq, int force_update)
 
 static inline void update_rq_runnable_avg(struct rq *rq, int runnable)
 {
-	__update_entity_runnable_avg(rq->clock_task, &rq->avg, runnable);
+	__update_entity_runnable_avg(rq->clock_task, &rq->avg, runnable,
+				     runnable);
 	__update_tg_runnable_avg(&rq->avg, &rq->cfs);
 }
 
@@ -1590,6 +1606,7 @@ set_next_entity(struct cfs_rq *cfs_rq, struct sched_entity *se)
 		 */
 		update_stats_wait_end(cfs_rq, se);
 		__dequeue_entity(cfs_rq, se);
+		update_entity_load_avg(se, 1);
 	}
 
 	update_stats_curr_start(cfs_rq, se);
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 77f3297..9602a47 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -117,7 +117,7 @@ struct task_group {
 
 	atomic_t load_weight;
 	atomic64_t load_avg;
-	atomic_t runnable_avg;
+	atomic_t runnable_avg, usage_avg;
 #endif
 
 #ifdef CONFIG_RT_GROUP_SCHED
@@ -260,7 +260,7 @@ struct cfs_rq {
 	 */
 	unsigned long h_load;
 
-	u32 tg_runnable_contrib;
+	u32 tg_runnable_contrib, tg_usage_contrib;
 	u64 runnable_load_avg, blocked_load_avg;
 	u64 tg_load_contrib;
 	atomic64_t decay_counter, removed_load;