sched: Avoid SMT siblings in select_idle_sibling() if possible

Avoid select_idle_sibling() from picking a sibling thread if there's
an idle core that shares cache.

This fixes SMT balancing in the increasingly common case where there's
a shared cache core available to balance to.

Tested-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1321350377.1421.55.camel@twins
Signed-off-by: Ingo Molnar <mingo@elte.hu>

1 files changed, 28 insertions, 14 deletions

diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c
index 7e51b5bb27cc..ba0e1f49a22f 100644
--- a/kernel/sched_fair.c
+++ b/kernel/sched_fair.c
@@ -2326,7 +2326,8 @@ static int select_idle_sibling(struct task_struct *p, int target)
         int cpu = smp_processor_id();
         int prev_cpu = task_cpu(p);
         struct sched_domain *sd;
-        int i;
+        struct sched_group *sg;
+        int i, smt = 0;
 
         /*
          * If the task is going to be woken-up on this cpu and if it is
@@ -2346,25 +2347,38 @@ static int select_idle_sibling(struct task_struct *p, int target)
          * Otherwise, iterate the domains and find an elegible idle cpu.
          */
         rcu_read_lock();
+again:
         for_each_domain(target, sd) {
-                if (!(sd->flags & SD_SHARE_PKG_RESOURCES))
-                        break;
+                if (!smt && (sd->flags & SD_SHARE_CPUPOWER))
+                        continue;
 
-                for_each_cpu_and(i, sched_domain_span(sd), tsk_cpus_allowed(p)) {
-                        if (idle_cpu(i)) {
-                                target = i;
-                                break;
+                if (!(sd->flags & SD_SHARE_PKG_RESOURCES)) {
+                        if (!smt) {
+                                smt = 1;
+                                goto again;
                         }
+                        break;
                 }
 
-                /*
-                 * Lets stop looking for an idle sibling when we reached
-                 * the domain that spans the current cpu and prev_cpu.
-                 */
-                if (cpumask_test_cpu(cpu, sched_domain_span(sd)) &&
-                    cpumask_test_cpu(prev_cpu, sched_domain_span(sd)))
-                        break;
+                sg = sd->groups;
+                do {
+                        if (!cpumask_intersects(sched_group_cpus(sg),
+                                                tsk_cpus_allowed(p)))
+                                goto next;
+
+                        for_each_cpu(i, sched_group_cpus(sg)) {
+                                if (!idle_cpu(i))
+                                        goto next;
+                        }
+
+                        target = cpumask_first_and(sched_group_cpus(sg),
+                                        tsk_cpus_allowed(p));
+                        goto done;
+next:
+                        sg = sg->next;
+                } while (sg != sd->groups);
         }
+done:
         rcu_read_unlock();
 
         return target;