sched/pelt: Relax the sync of load_sum with load_avg

mainline inclusion
from mainline-v5.17-rc2
commit 2d02fa8c
category: bugfix
bugzilla: https://gitee.com/openeuler/kernel/issues/I78WM8

Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?h=v6.4-rc3&id=2d02fa8cc21a93da35cfba462bf8ab87bf2db651

--------------------------------

Similarly to util_avg and util_sum, don't sync load_sum with the low
bound of load_avg but only ensure that load_sum stays in the correct range.
Signed-off-by: NVincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: NPeter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: NDietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: NSachin Sant <sachinp@linux.ibm.com>
Link: https://lkml.kernel.org/r/20220111134659.24961-5-vincent.guittot@linaro.orgSigned-off-by: NGuan Jing <guanjing6@huawei.com>

kernel/sched/fair.c

@@ -3084,6 +3084,9 @@ dequeue_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *se)
 {
 	sub_positive(&cfs_rq->avg.load_avg, se->avg.load_avg);
 	sub_positive(&cfs_rq->avg.load_sum, se_weight(se) * se->avg.load_sum);
+	/* See update_cfs_rq_load_avg() */
+	cfs_rq->avg.load_sum = max_t(u32, cfs_rq->avg.load_sum,
+					  cfs_rq->avg.load_avg * PELT_MIN_DIVIDER);
 }
 #else
 static inline void
@@ -3511,9 +3514,10 @@ update_tg_cfs_runnable(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cf
 static inline void
 update_tg_cfs_load(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cfs_rq *gcfs_rq)
 {
-	long delta, running_sum, runnable_sum = gcfs_rq->prop_runnable_sum;
+	long delta_avg, running_sum, runnable_sum = gcfs_rq->prop_runnable_sum;
 	unsigned long load_avg;
 	u64 load_sum = 0;
+	s64 delta_sum;
 	u32 divider;
 
 	if (!runnable_sum)
@@ -3540,7 +3544,7 @@ update_tg_cfs_load(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cfs_rq
 		 * assuming all tasks are equally runnable.
 		 */
 		if (scale_load_down(gcfs_rq->load.weight)) {
-			load_sum = div_s64(gcfs_rq->avg.load_sum,
+			load_sum = div_u64(gcfs_rq->avg.load_sum,
 				scale_load_down(gcfs_rq->load.weight));
 		}
 
@@ -3557,16 +3561,22 @@ update_tg_cfs_load(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cfs_rq
 	running_sum = se->avg.util_sum >> SCHED_CAPACITY_SHIFT;
 	runnable_sum = max(runnable_sum, running_sum);
 
-	load_sum = (s64)se_weight(se) * runnable_sum;
-	load_avg = div_s64(load_sum, divider);
+	load_sum = se_weight(se) * runnable_sum;
+	load_avg = div_u64(load_sum, divider);
+
+	delta_avg = load_avg - se->avg.load_avg;
+	if (!delta_avg)
+		return;
 
-	delta = load_avg - se->avg.load_avg;
+	delta_sum = load_sum - (s64)se_weight(se) * se->avg.load_sum;
 
 	se->avg.load_sum = runnable_sum;
 	se->avg.load_avg = load_avg;
-
-	add_positive(&cfs_rq->avg.load_avg, delta);
-	cfs_rq->avg.load_sum = cfs_rq->avg.load_avg * divider;
+	add_positive(&cfs_rq->avg.load_avg, delta_avg);
+	add_positive(&cfs_rq->avg.load_sum, delta_sum);
+	/* See update_cfs_rq_load_avg() */
+	cfs_rq->avg.load_sum = max_t(u32, cfs_rq->avg.load_sum,
+					  cfs_rq->avg.load_avg * PELT_MIN_DIVIDER);
 }
 
 static inline void add_tg_cfs_propagate(struct cfs_rq *cfs_rq, long runnable_sum)
@@ -3682,7 +3692,9 @@ update_cfs_rq_load_avg(u64 now, struct cfs_rq *cfs_rq)
 
 		r = removed_load;
 		sub_positive(&sa->load_avg, r);
-		sa->load_sum = sa->load_avg * divider;
+		sub_positive(&sa->load_sum, r * divider);
+		/* See sa->util_sum below */
+		sa->load_sum = max_t(u32, sa->load_sum, sa->load_avg * PELT_MIN_DIVIDER);
 
 		r = removed_util;
 		sub_positive(&sa->util_avg, r);