sched/pelt: Relax the sync of runnable_sum with runnable_avg

Similarly to util_avg and util_sum, don't sync runnable_sum with the low
bound of runnable_avg but only ensure that runnable_sum stays in the
correct range.

Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Sachin Sant <sachinp@linux.ibm.com>
Link: https://lkml.kernel.org/r/20220111134659.24961-4-vincent.guittot@linaro.org

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index ad2809cc57c78a5b6ffdb242ff5ab28eb6aafeed..0e87e1916504462d0c32da765e45629f133e9aee 100644 (file)
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -3483,11 +3483,11 @@ update_tg_cfs_util(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cfs_rq
 static inline void
 update_tg_cfs_runnable(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cfs_rq *gcfs_rq)
 {
-       long delta = gcfs_rq->avg.runnable_avg - se->avg.runnable_avg;
-       u32 divider;
+       long delta_sum, delta_avg = gcfs_rq->avg.runnable_avg - se->avg.runnable_avg;
+       u32 new_sum, divider;
 
        /* Nothing to update */
-       if (!delta)
+       if (!delta_avg)
                return;
 
        /*
@@ -3498,11 +3498,16 @@ update_tg_cfs_runnable(struct cfs_rq *cfs_rq, struct sched_entity *se, struct cf
 
        /* Set new sched_entity's runnable */
        se->avg.runnable_avg = gcfs_rq->avg.runnable_avg;
-       se->avg.runnable_sum = se->avg.runnable_avg * divider;
+       new_sum = se->avg.runnable_avg * divider;
+       delta_sum = (long)new_sum - (long)se->avg.runnable_sum;
+       se->avg.runnable_sum = new_sum;
 
        /* Update parent cfs_rq runnable */
-       add_positive(&cfs_rq->avg.runnable_avg, delta);
-       cfs_rq->avg.runnable_sum = cfs_rq->avg.runnable_avg * divider;
+       add_positive(&cfs_rq->avg.runnable_avg, delta_avg);
+       add_positive(&cfs_rq->avg.runnable_sum, delta_sum);
+       /* See update_cfs_rq_load_avg() */
+       cfs_rq->avg.runnable_sum = max_t(u32, cfs_rq->avg.runnable_sum,
+                                             cfs_rq->avg.runnable_avg * PELT_MIN_DIVIDER);
 }
 
 static inline void
@@ -3702,7 +3707,10 @@ update_cfs_rq_load_avg(u64 now, struct cfs_rq *cfs_rq)
 
                r = removed_runnable;
                sub_positive(&sa->runnable_avg, r);
-               sa->runnable_sum = sa->runnable_avg * divider;
+               sub_positive(&sa->runnable_sum, r * divider);
+               /* See sa->util_sum above */
+               sa->runnable_sum = max_t(u32, sa->runnable_sum,
+                                             sa->runnable_avg * PELT_MIN_DIVIDER);
 
                /*
                 * removed_runnable is the unweighted version of removed_load so we
@@ -3789,12 +3797,6 @@ static void attach_entity_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *s
  */
 static void detach_entity_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *se)
 {
-       /*
-        * cfs_rq->avg.period_contrib can be used for both cfs_rq and se.
-        * See ___update_load_avg() for details.
-        */
-       u32 divider = get_pelt_divider(&cfs_rq->avg);
-
        dequeue_load_avg(cfs_rq, se);
        sub_positive(&cfs_rq->avg.util_avg, se->avg.util_avg);
        sub_positive(&cfs_rq->avg.util_sum, se->avg.util_sum);
@@ -3803,7 +3805,10 @@ static void detach_entity_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *s
                                          cfs_rq->avg.util_avg * PELT_MIN_DIVIDER);
 
        sub_positive(&cfs_rq->avg.runnable_avg, se->avg.runnable_avg);
-       cfs_rq->avg.runnable_sum = cfs_rq->avg.runnable_avg * divider;
+       sub_positive(&cfs_rq->avg.runnable_sum, se->avg.runnable_sum);
+       /* See update_cfs_rq_load_avg() */
+       cfs_rq->avg.runnable_sum = max_t(u32, cfs_rq->avg.runnable_sum,
+                                             cfs_rq->avg.runnable_avg * PELT_MIN_DIVIDER);
 
        add_tg_cfs_propagate(cfs_rq, -se->avg.load_sum);