[PATCH v5 4/4] sched/fair: update util_est only on util_avg updates

[Patrick Bellasi]

The estimated utilization of a task is currently updated every time the
task is dequeued. However, to keep overheads under control, PELT signals
are effectively updated at maximum once every 1ms.

Thus, for really short running tasks, it can happen that their util_avg
value has not been updates since their last enqueue.  If such tasks are
also frequently running tasks (e.g. the kind of workload generated by
hackbench) it can also happen that their util_avg is updated only every
few activations.

This means that updating util_est at every dequeue potentially introduces
not necessary overheads and it's also conceptually wrong if the util_avg
signal has never been updated during a task activation.

Let's introduce a throttling mechanism on task's util_est updates
to sync them with util_avg updates. To make the solution memory
efficient, both in terms of space and load/store operations, we encode a
synchronization flag into the LSB of util_est.enqueued.
This makes util_est an even values only metric, which is still
considered good enough for its purpose.
The synchronization bit is (re)set by __update_load_avg_se() once the
PELT signal of a task has been updated during its last activation.

Such a throttling mechanism allows to keep under control util_est
overheads in the wakeup hot path, thus making it a suitable mechanism
which can be enabled also on high-intensity workload systems.
Thus, this now switches on by default the estimation utilization
scheduler feature.

Suggested-by: Chris Redpath <chris.redpath@xxxxxxx>
Signed-off-by: Patrick Bellasi <patrick.bellasi@xxxxxxx>

---
Changes in v5:
 - set SCHED_FEAT(UTIL_EST, true) as default (Peter)
---
 kernel/sched/fair.c     | 39 +++++++++++++++++++++++++++++++++++----
 kernel/sched/features.h |  2 +-
 2 files changed, 36 insertions(+), 5 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 8364771f7301..1bf9a86ebc39 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -3047,6 +3047,29 @@ static inline void cfs_rq_util_change(struct cfs_rq *cfs_rq)
 	}
 }
 
+/*
+ * When a task is dequeued, its estimated utilization should not be update if
+ * its util_avg has not been updated at least once.
+ * This flag is used to synchronize util_avg updates with util_est updates.
+ * We map this information into the LSB bit of the utilization saved at
+ * dequeue time (i.e. util_est.dequeued).
+ */
+#define UTIL_EST_NEED_UPDATE_FLAG 0x1
+
+static inline void cfs_se_util_change(struct sched_avg *avg)
+{
+	if (sched_feat(UTIL_EST)) {
+		struct util_est ue = READ_ONCE(avg->util_est);
+
+		if (!(ue.enqueued & UTIL_EST_NEED_UPDATE_FLAG))
+			return;
+
+		/* Reset flag to report util_avg has been updated */
+		ue.enqueued &= ~UTIL_EST_NEED_UPDATE_FLAG;
+		WRITE_ONCE(avg->util_est, ue);
+	}
+}
+
 #ifdef CONFIG_SMP
 /*
  * Approximate:
@@ -3308,6 +3331,7 @@ __update_load_avg_se(u64 now, int cpu, struct cfs_rq *cfs_rq, struct sched_entit
 				cfs_rq->curr == se)) {
 
 		___update_load_avg(&se->avg, se_weight(se), se_runnable(se));
+		cfs_se_util_change(&se->avg);
 		return 1;
 	}
 
@@ -5218,7 +5242,7 @@ static inline void util_est_enqueue(struct cfs_rq *cfs_rq,
 
 	/* Update root cfs_rq's estimated utilization */
 	enqueued  = READ_ONCE(cfs_rq->avg.util_est.enqueued);
-	enqueued += _task_util_est(p);
+	enqueued += (_task_util_est(p) | 0x1);
 	WRITE_ONCE(cfs_rq->avg.util_est.enqueued, enqueued);
 }
 
@@ -5310,7 +5334,7 @@ static inline void util_est_dequeue(struct cfs_rq *cfs_rq,
 	if (cfs_rq->nr_running) {
 		ue.enqueued  = READ_ONCE(cfs_rq->avg.util_est.enqueued);
 		ue.enqueued -= min_t(unsigned int, ue.enqueued,
-				     _task_util_est(p));
+				     (_task_util_est(p) | UTIL_EST_NEED_UPDATE_FLAG));
 	}
 	WRITE_ONCE(cfs_rq->avg.util_est.enqueued, ue.enqueued);
 
@@ -5321,12 +5345,19 @@ static inline void util_est_dequeue(struct cfs_rq *cfs_rq,
 	if (!task_sleep)
 		return;
 
+	/*
+	 * Skip update of task's estimated utilization if the PELT signal has
+	 * never been updated (at least once) since last enqueue time.
+	 */
+	ue = READ_ONCE(p->se.avg.util_est);
+	if (ue.enqueued & UTIL_EST_NEED_UPDATE_FLAG)
+		return;
+
 	/*
 	 * Skip update of task's estimated utilization when its EWMA is
 	 * already ~1% close to its last activation value.
 	 */
-	ue = READ_ONCE(p->se.avg.util_est);
-	ue.enqueued = task_util(p);
+	ue.enqueued = (task_util(p) | UTIL_EST_NEED_UPDATE_FLAG);
 	last_ewma_diff = ue.enqueued - ue.ewma;
 	if (within_margin(last_ewma_diff, (SCHED_CAPACITY_SCALE / 100)))
 		return;
diff --git a/kernel/sched/features.h b/kernel/sched/features.h
index c459a4b61544..85ae8488039c 100644
--- a/kernel/sched/features.h
+++ b/kernel/sched/features.h
@@ -89,4 +89,4 @@ SCHED_FEAT(WA_BIAS, true)
 /*
  * UtilEstimation. Use estimated CPU utilization.
  */
-SCHED_FEAT(UTIL_EST, false)
+SCHED_FEAT(UTIL_EST, true)
-- 
2.15.1