[PATCH v2] sched/rt: fix call to cpufreq_update_util

[Vincent Guittot]

With commit 8f111bc357aa ("cpufreq/schedutil: Rewrite CPUFREQ_RT support")
schedutil governor uses rq->rt.rt_nr_running to detect whether a RT task is
currently running on the CPU and to set frequency to max if necessary.

cpufreq_update_util() is called in enqueue/dequeue_top_rt_rq() but
rq->rt.rt_nr_running has not been updated yet when dequeue_top_rt_rq() is
called so schedutil still considers that a RT task is running when the
last task is dequeued. The update of rq->rt.rt_nr_running happens later
in dequeue_rt_stack().

In fact, we can take advantage of the sequence that the dequeue then
re-enqueue rt entities when a rt task is enqueued or dequeued;
As a result enqueue_top_rt_rq() is always called when a task is
enqueued or dequeued and also when groups are throttled or unthrottled.
The only place that not use enqueue_top_rt_rq() is when root rt_rq is
throttled.

Fixes: 8f111bc357aa ('cpufreq/schedutil: Rewrite CPUFREQ_RT support')
Cc: <stable@xxxxxxxxxxxxxxx> # v4.16+
Signed-off-by: Vincent Guittot <vincent.guittot@xxxxxxxxxx>
---
 kernel/sched/cpufreq_schedutil.c |  2 +-
 kernel/sched/rt.c                | 16 ++++++++++------
 kernel/sched/sched.h             |  5 +++++
 3 files changed, 16 insertions(+), 7 deletions(-)

diff --git a/kernel/sched/cpufreq_schedutil.c b/kernel/sched/cpufreq_schedutil.c
index 3cde464..c907fde 100644
--- a/kernel/sched/cpufreq_schedutil.c
+++ b/kernel/sched/cpufreq_schedutil.c
@@ -192,7 +192,7 @@ static unsigned long sugov_aggregate_util(struct sugov_cpu *sg_cpu)
 {
 	struct rq *rq = cpu_rq(sg_cpu->cpu);
 
-	if (rq->rt.rt_nr_running)
+	if (rt_rq_is_runnable(&rq->rt))
 		return sg_cpu->max;
 
 	/*
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index 47556b0..5725670 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -508,8 +508,11 @@ static void sched_rt_rq_dequeue(struct rt_rq *rt_rq)
 
 	rt_se = rt_rq->tg->rt_se[cpu];
 
-	if (!rt_se)
+	if (!rt_se) {
 		dequeue_top_rt_rq(rt_rq);
+		/* Kick cpufreq (see the comment in kernel/sched/sched.h). */
+		cpufreq_update_util(rq_of_rt_rq(rt_rq), 0);
+	}
 	else if (on_rt_rq(rt_se))
 		dequeue_rt_entity(rt_se, 0);
 }
@@ -1001,8 +1004,6 @@ dequeue_top_rt_rq(struct rt_rq *rt_rq)
 	sub_nr_running(rq, rt_rq->rt_nr_running);
 	rt_rq->rt_queued = 0;
 
-	/* Kick cpufreq (see the comment in kernel/sched/sched.h). */
-	cpufreq_update_util(rq, 0);
 }
 
 static void
@@ -1014,11 +1015,14 @@ enqueue_top_rt_rq(struct rt_rq *rt_rq)
 
 	if (rt_rq->rt_queued)
 		return;
-	if (rt_rq_throttled(rt_rq) || !rt_rq->rt_nr_running)
+
+	if (rt_rq_throttled(rt_rq))
 		return;
 
-	add_nr_running(rq, rt_rq->rt_nr_running);
-	rt_rq->rt_queued = 1;
+	if (rt_rq->rt_nr_running) {
+		add_nr_running(rq, rt_rq->rt_nr_running);
+		rt_rq->rt_queued = 1;
+	}
 
 	/* Kick cpufreq (see the comment in kernel/sched/sched.h). */
 	cpufreq_update_util(rq, 0);
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 6601baf..27ddec3 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -609,6 +609,11 @@ struct rt_rq {
 #endif
 };
 
+static inline bool rt_rq_is_runnable(struct rt_rq *rt_rq)
+{
+	return rt_rq->rt_queued && rt_rq->rt_nr_running;
+}
+
 /* Deadline class' related fields in a runqueue */
 struct dl_rq {
 	/* runqueue is an rbtree, ordered by deadline */
-- 
2.7.4