sched/fair: Add new cfs_rq.h_nr_runnable

With delayed dequeued feature, a sleeping sched_entity remains queued in
the rq until its lag has elapsed. As a result, it stays also visible
in the statistics that are used to balance the system and in particular
the field cfs.h_nr_queued when the sched_entity is associated to a task.

Create a new h_nr_runnable that tracks only queued and runnable tasks.

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>
Link: https://lore.kernel.org/r/20241202174606.4074512-5-vincent.guittot@linaro.org

diff --git a/kernel/sched/debug.c b/kernel/sched/debug.c
index 08d6c2b7caa393eeb57a92a10f34f1727f2fa3fe..fd711cc4d44c06eb96c7dca4f8743d56fe4653ea 100644 (file)
--- a/kernel/sched/debug.c
+++ b/kernel/sched/debug.c
@@ -844,6 +844,7 @@ void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq)
        spread = right_vruntime - left_vruntime;
        SEQ_printf(m, "  .%-30s: %Ld.%06ld\n", "spread", SPLIT_NS(spread));
        SEQ_printf(m, "  .%-30s: %d\n", "nr_running", cfs_rq->nr_running);
+       SEQ_printf(m, "  .%-30s: %d\n", "h_nr_runnable", cfs_rq->h_nr_runnable);
        SEQ_printf(m, "  .%-30s: %d\n", "h_nr_queued", cfs_rq->h_nr_queued);
        SEQ_printf(m, "  .%-30s: %d\n", "h_nr_delayed", cfs_rq->h_nr_delayed);
        SEQ_printf(m, "  .%-30s: %d\n", "idle_nr_running",

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index d6a9447e5e230a1e1b844c4cc717e2160c2910a8..ed01e72b2b7726db2f194b275666046368542db5 100644 (file)
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -5469,6 +5469,7 @@ static void set_delayed(struct sched_entity *se)
        for_each_sched_entity(se) {
                struct cfs_rq *cfs_rq = cfs_rq_of(se);
 
+               cfs_rq->h_nr_runnable--;
                cfs_rq->h_nr_delayed++;
                if (cfs_rq_throttled(cfs_rq))
                        break;
@@ -5481,6 +5482,7 @@ static void clear_delayed(struct sched_entity *se)
        for_each_sched_entity(se) {
                struct cfs_rq *cfs_rq = cfs_rq_of(se);
 
+               cfs_rq->h_nr_runnable++;
                cfs_rq->h_nr_delayed--;
                if (cfs_rq_throttled(cfs_rq))
                        break;
@@ -5930,7 +5932,7 @@ static bool throttle_cfs_rq(struct cfs_rq *cfs_rq)
        struct rq *rq = rq_of(cfs_rq);
        struct cfs_bandwidth *cfs_b = tg_cfs_bandwidth(cfs_rq->tg);
        struct sched_entity *se;
-       long queued_delta, idle_task_delta, delayed_delta, dequeue = 1;
+       long queued_delta, runnable_delta, idle_task_delta, delayed_delta, dequeue = 1;
        long rq_h_nr_queued = rq->cfs.h_nr_queued;
 
        raw_spin_lock(&cfs_b->lock);
@@ -5962,6 +5964,7 @@ static bool throttle_cfs_rq(struct cfs_rq *cfs_rq)
        rcu_read_unlock();
 
        queued_delta = cfs_rq->h_nr_queued;
+       runnable_delta = cfs_rq->h_nr_runnable;
        idle_task_delta = cfs_rq->idle_h_nr_running;
        delayed_delta = cfs_rq->h_nr_delayed;
        for_each_sched_entity(se) {
@@ -5986,6 +5989,7 @@ static bool throttle_cfs_rq(struct cfs_rq *cfs_rq)
                        idle_task_delta = cfs_rq->h_nr_queued;
 
                qcfs_rq->h_nr_queued -= queued_delta;
+               qcfs_rq->h_nr_runnable -= runnable_delta;
                qcfs_rq->idle_h_nr_running -= idle_task_delta;
                qcfs_rq->h_nr_delayed -= delayed_delta;
 
@@ -6009,6 +6013,7 @@ static bool throttle_cfs_rq(struct cfs_rq *cfs_rq)
                        idle_task_delta = cfs_rq->h_nr_queued;
 
                qcfs_rq->h_nr_queued -= queued_delta;
+               qcfs_rq->h_nr_runnable -= runnable_delta;
                qcfs_rq->idle_h_nr_running -= idle_task_delta;
                qcfs_rq->h_nr_delayed -= delayed_delta;
        }
@@ -6036,7 +6041,7 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
        struct rq *rq = rq_of(cfs_rq);
        struct cfs_bandwidth *cfs_b = tg_cfs_bandwidth(cfs_rq->tg);
        struct sched_entity *se;
-       long queued_delta, idle_task_delta, delayed_delta;
+       long queued_delta, runnable_delta, idle_task_delta, delayed_delta;
        long rq_h_nr_queued = rq->cfs.h_nr_queued;
 
        se = cfs_rq->tg->se[cpu_of(rq)];
@@ -6071,6 +6076,7 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
        }
 
        queued_delta = cfs_rq->h_nr_queued;
+       runnable_delta = cfs_rq->h_nr_runnable;
        idle_task_delta = cfs_rq->idle_h_nr_running;
        delayed_delta = cfs_rq->h_nr_delayed;
        for_each_sched_entity(se) {
@@ -6089,6 +6095,7 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
                        idle_task_delta = cfs_rq->h_nr_queued;
 
                qcfs_rq->h_nr_queued += queued_delta;
+               qcfs_rq->h_nr_runnable += runnable_delta;
                qcfs_rq->idle_h_nr_running += idle_task_delta;
                qcfs_rq->h_nr_delayed += delayed_delta;
 
@@ -6107,6 +6114,7 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
                        idle_task_delta = cfs_rq->h_nr_queued;
 
                qcfs_rq->h_nr_queued += queued_delta;
+               qcfs_rq->h_nr_runnable += runnable_delta;
                qcfs_rq->idle_h_nr_running += idle_task_delta;
                qcfs_rq->h_nr_delayed += delayed_delta;
 
@@ -7021,6 +7029,8 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
                enqueue_entity(cfs_rq, se, flags);
                slice = cfs_rq_min_slice(cfs_rq);
 
+               if (!h_nr_delayed)
+                       cfs_rq->h_nr_runnable++;
                cfs_rq->h_nr_queued++;
                cfs_rq->idle_h_nr_running += idle_h_nr_running;
                cfs_rq->h_nr_delayed += h_nr_delayed;
@@ -7045,6 +7055,8 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
                se->slice = slice;
                slice = cfs_rq_min_slice(cfs_rq);
 
+               if (!h_nr_delayed)
+                       cfs_rq->h_nr_runnable++;
                cfs_rq->h_nr_queued++;
                cfs_rq->idle_h_nr_running += idle_h_nr_running;
                cfs_rq->h_nr_delayed += h_nr_delayed;
@@ -7135,6 +7147,8 @@ static int dequeue_entities(struct rq *rq, struct sched_entity *se, int flags)
                        break;
                }
 
+               if (!h_nr_delayed)
+                       cfs_rq->h_nr_runnable -= h_nr_queued;
                cfs_rq->h_nr_queued -= h_nr_queued;
                cfs_rq->idle_h_nr_running -= idle_h_nr_running;
                cfs_rq->h_nr_delayed -= h_nr_delayed;
@@ -7174,6 +7188,8 @@ static int dequeue_entities(struct rq *rq, struct sched_entity *se, int flags)
                se->slice = slice;
                slice = cfs_rq_min_slice(cfs_rq);
 
+               if (!h_nr_delayed)
+                       cfs_rq->h_nr_runnable -= h_nr_queued;
                cfs_rq->h_nr_queued -= h_nr_queued;
                cfs_rq->idle_h_nr_running -= idle_h_nr_running;
                cfs_rq->h_nr_delayed -= h_nr_delayed;

diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index b011081aff9754f157c902741fb0a3ff1226db31..869d5d3521f20c3c150885b0bd861dfc23abb770 100644 (file)
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -647,6 +647,7 @@ struct cfs_rq {
        struct load_weight      load;
        unsigned int            nr_running;
        unsigned int            h_nr_queued;       /* SCHED_{NORMAL,BATCH,IDLE} */
+       unsigned int            h_nr_runnable;     /* SCHED_{NORMAL,BATCH,IDLE} */
        unsigned int            idle_nr_running;   /* SCHED_IDLE */
        unsigned int            idle_h_nr_running; /* SCHED_IDLE */
        unsigned int            h_nr_delayed;