[PATCH v4] sched/fair: Sync se's load_avg with cfs_rq in reweight_task

[Chuyi Zhou]

In reweight_task(), there are two situations:

1. The task was on_rq, then the task's load_avg is accurate because in
__sched_setscheduler()/set_user_nice(), we would dequeue the on_rq tasks
before doing reweight. The task's load_avg would be synchronized with
cfs_rq through update_load_avg() in dequeue_task().

2. The task is sleeping, its load_avg might not have been updated for some
time, which can result in inaccurate dequeue_load_avg() in
reweight_entity().

This patch solves this by using sync_entity_load_avg() to synchronize the
load_avg of se with cfs_rq before dequeue_load_avg() in reweight_entity().
For tasks were on_rq, since we already update load_avg to accurate values
in dequeue_task(), this change will not have other effects due to the short
time interval between the two updates.

Suggested-by: Dietmar Eggemann <dietmar.eggemann@xxxxxxx>
Signed-off-by: Chuyi Zhou <zhouchuyi@xxxxxxxxxxxxx>
Reviewed-by: Chengming Zhou <chengming.zhou@xxxxxxxxx>
Reviewed-by: Vishal Chourasia <vishalc@xxxxxxxxxxxxx>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@xxxxxxx>

---
Changes in v4:
- Fix the 'if else' code style issue.(Dietmar)
- Add a description of __sched_setscheduler()/set_user_nice() in the commit
  log.(Dietmar)
- Add comment before calling sync_entity_load_avg().(Qais)
Changes in v3:
- use sync_entity_load_avg() rather than update_load_avg() to sync the
sleeping task with its cfs_rq suggested by Dietmar.
- Link t0 v2: https://lore.kernel.org/lkml/20240720051248.59608-1-zhouchuyi@xxxxxxxxxxxxx/
Changes in v2:
- change the description in commit log.
- use update_load_avg() in reweight_task() rather than in reweight_entity
suggested by chengming.
- Link to v1: https://lore.kernel.org/lkml/20240716150840.23061-1-zhouchuyi@xxxxxxxxxxxxx/
---
 kernel/sched/fair.c | 46 +++++++++++++++++++++++++++------------------
 1 file changed, 28 insertions(+), 18 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 9057584ec06d..1e3c7c582541 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -3669,11 +3669,32 @@ dequeue_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *se)
 	cfs_rq->avg.load_sum = max_t(u32, cfs_rq->avg.load_sum,
 					  cfs_rq->avg.load_avg * PELT_MIN_DIVIDER);
 }
+
+static inline u64 cfs_rq_last_update_time(struct cfs_rq *cfs_rq)
+{
+	return u64_u32_load_copy(cfs_rq->avg.last_update_time,
+				 cfs_rq->last_update_time_copy);
+}
+
+/*
+ * Synchronize entity load avg of dequeued entity without locking
+ * the previous rq.
+ */
+static void sync_entity_load_avg(struct sched_entity *se)
+{
+	struct cfs_rq *cfs_rq = cfs_rq_of(se);
+	u64 last_update_time;
+
+	last_update_time = cfs_rq_last_update_time(cfs_rq);
+	__update_load_avg_blocked_se(last_update_time, se);
+}
+
 #else
 static inline void
 enqueue_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *se) { }
 static inline void
 dequeue_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *se) { }
+static void sync_entity_load_avg(struct sched_entity *se) { }
 #endif
 
 static void reweight_eevdf(struct sched_entity *se, u64 avruntime,
@@ -3795,7 +3816,14 @@ static void reweight_entity(struct cfs_rq *cfs_rq, struct sched_entity *se,
 		if (!curr)
 			__dequeue_entity(cfs_rq, se);
 		update_load_sub(&cfs_rq->load, se->load.weight);
+	} else if (entity_is_task(se)) {
+		/*
+		 * If the task is sleeping, we need to synchronize entity load avg
+		 * before dequeue_load_avg().
+		 */
+		sync_entity_load_avg(se);
 	}
+
 	dequeue_load_avg(cfs_rq, se);
 
 	if (se->on_rq) {
@@ -4034,11 +4062,6 @@ static inline bool load_avg_is_decayed(struct sched_avg *sa)
 	return true;
 }
 
-static inline u64 cfs_rq_last_update_time(struct cfs_rq *cfs_rq)
-{
-	return u64_u32_load_copy(cfs_rq->avg.last_update_time,
-				 cfs_rq->last_update_time_copy);
-}
 #ifdef CONFIG_FAIR_GROUP_SCHED
 /*
  * Because list_add_leaf_cfs_rq always places a child cfs_rq on the list
@@ -4773,19 +4796,6 @@ static inline void update_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *s
 	}
 }
 
-/*
- * Synchronize entity load avg of dequeued entity without locking
- * the previous rq.
- */
-static void sync_entity_load_avg(struct sched_entity *se)
-{
-	struct cfs_rq *cfs_rq = cfs_rq_of(se);
-	u64 last_update_time;
-
-	last_update_time = cfs_rq_last_update_time(cfs_rq);
-	__update_load_avg_blocked_se(last_update_time, se);
-}
-
 /*
  * Task first catches up with cfs_rq, and then subtract
  * itself from the cfs_rq (task must be off the queue now).
-- 
2.20.1