We try to find an idle CPU to run the next task, but in case we don't
find an idle CPU it is better to pick a CPU which will run the task the
soonest, for performance reason.

A CPU which isn't idle but has only SCHED_IDLE activity queued on it
should be a good target based on this criteria as any normal fair task
will most likely preempt the currently running SCHED_IDLE task
immediately. In fact, choosing a SCHED_IDLE CPU over a fully idle one
shall give better results as it should be able to run the task sooner
than an idle CPU (which requires to be woken up from an idle state).

This patch updates both fast and slow paths with this optimization.

Signed-off-by: Viresh Kumar <viresh.ku...@linaro.org>
---
 kernel/sched/fair.c | 43 +++++++++++++++++++++++++++++++++----------
 1 file changed, 33 insertions(+), 10 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 1277adc3e7ed..2e0527fd468c 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -5376,6 +5376,15 @@ static struct {
 
 #endif /* CONFIG_NO_HZ_COMMON */
 
+/* CPU only has SCHED_IDLE tasks enqueued */
+static int sched_idle_cpu(int cpu)
+{
+       struct rq *rq = cpu_rq(cpu);
+
+       return unlikely(rq->nr_running == rq->cfs.idle_h_nr_running &&
+                       rq->nr_running);
+}
+
 static unsigned long cpu_runnable_load(struct rq *rq)
 {
        return cfs_rq_runnable_load_avg(&rq->cfs);
@@ -5698,7 +5707,7 @@ find_idlest_group_cpu(struct sched_group *group, struct 
task_struct *p, int this
        unsigned int min_exit_latency = UINT_MAX;
        u64 latest_idle_timestamp = 0;
        int least_loaded_cpu = this_cpu;
-       int shallowest_idle_cpu = -1;
+       int shallowest_idle_cpu = -1, si_cpu = -1;
        int i;
 
        /* Check if we have any choice: */
@@ -5729,7 +5738,12 @@ find_idlest_group_cpu(struct sched_group *group, struct 
task_struct *p, int this
                                latest_idle_timestamp = rq->idle_stamp;
                                shallowest_idle_cpu = i;
                        }
-               } else if (shallowest_idle_cpu == -1) {
+               } else if (shallowest_idle_cpu == -1 && si_cpu == -1) {
+                       if (sched_idle_cpu(i)) {
+                               si_cpu = i;
+                               continue;
+                       }
+
                        load = cpu_runnable_load(cpu_rq(i));
                        if (load < min_load) {
                                min_load = load;
@@ -5738,7 +5752,11 @@ find_idlest_group_cpu(struct sched_group *group, struct 
task_struct *p, int this
                }
        }
 
-       return shallowest_idle_cpu != -1 ? shallowest_idle_cpu : 
least_loaded_cpu;
+       if (shallowest_idle_cpu != -1)
+               return shallowest_idle_cpu;
+       if (si_cpu != -1)
+               return si_cpu;
+       return least_loaded_cpu;
 }
 
 static inline int find_idlest_cpu(struct sched_domain *sd, struct task_struct 
*p,
@@ -5891,7 +5909,7 @@ static int select_idle_core(struct task_struct *p, struct 
sched_domain *sd, int
  */
 static int select_idle_smt(struct task_struct *p, int target)
 {
-       int cpu;
+       int cpu, si_cpu = -1;
 
        if (!static_branch_likely(&sched_smt_present))
                return -1;
@@ -5901,9 +5919,11 @@ static int select_idle_smt(struct task_struct *p, int 
target)
                        continue;
                if (available_idle_cpu(cpu))
                        return cpu;
+               if (si_cpu == -1 && sched_idle_cpu(cpu))
+                       si_cpu = cpu;
        }
 
-       return -1;
+       return si_cpu;
 }
 
 #else /* CONFIG_SCHED_SMT */
@@ -5931,7 +5951,7 @@ static int select_idle_cpu(struct task_struct *p, struct 
sched_domain *sd, int t
        u64 avg_cost, avg_idle;
        u64 time, cost;
        s64 delta;
-       int cpu, nr = INT_MAX;
+       int cpu, nr = INT_MAX, si_cpu = -1;
        int this = smp_processor_id();
 
        this_sd = rcu_dereference(*this_cpu_ptr(&sd_llc));
@@ -5960,11 +5980,13 @@ static int select_idle_cpu(struct task_struct *p, 
struct sched_domain *sd, int t
 
        for_each_cpu_wrap(cpu, sched_domain_span(sd), target) {
                if (!--nr)
-                       return -1;
+                       return si_cpu;
                if (!cpumask_test_cpu(cpu, p->cpus_ptr))
                        continue;
                if (available_idle_cpu(cpu))
                        break;
+               if (si_cpu == -1 && sched_idle_cpu(cpu))
+                       si_cpu = cpu;
        }
 
        time = cpu_clock(this) - time;
@@ -5983,13 +6005,14 @@ static int select_idle_sibling(struct task_struct *p, 
int prev, int target)
        struct sched_domain *sd;
        int i, recent_used_cpu;
 
-       if (available_idle_cpu(target))
+       if (available_idle_cpu(target) || sched_idle_cpu(target))
                return target;
 
        /*
         * If the previous CPU is cache affine and idle, don't be stupid:
         */
-       if (prev != target && cpus_share_cache(prev, target) && 
available_idle_cpu(prev))
+       if (prev != target && cpus_share_cache(prev, target) &&
+           (available_idle_cpu(prev) || sched_idle_cpu(prev)))
                return prev;
 
        /* Check a recently used CPU as a potential idle candidate: */
@@ -5997,7 +6020,7 @@ static int select_idle_sibling(struct task_struct *p, int 
prev, int target)
        if (recent_used_cpu != prev &&
            recent_used_cpu != target &&
            cpus_share_cache(recent_used_cpu, target) &&
-           available_idle_cpu(recent_used_cpu) &&
+           (available_idle_cpu(recent_used_cpu) || 
sched_idle_cpu(recent_used_cpu)) &&
            cpumask_test_cpu(p->recent_used_cpu, p->cpus_ptr)) {
                /*
                 * Replace recent_used_cpu with prev as it is a potential
-- 
2.21.0.rc0.269.g1a574e7a288b

Reply via email to