sched: Add a few SYNC hint knobs to play with

Currently we use overlap to weaken the SYNC hint, but allow it to
set the hint as well.

 echo NO_SYNC_WAKEUP > /debug/sched_features
 echo SYNC_MORE > /debug/sched_features

preserves pipe-test behaviour without using the WF_SYNC hint.

Worth playing with on more workloads...

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c
index 6766959c7f449b4a3d6c0aab4deeb7b7dc396f53..280892e9d85e3543564a1397cd51ad82158f7545 100644 (file)
--- a/kernel/sched_fair.c
+++ b/kernel/sched_fair.c
@@ -1165,9 +1165,17 @@ static int wake_affine(struct sched_domain *sd, struct task_struct *p, int sync)
        load      = source_load(prev_cpu, idx);
        this_load = target_load(this_cpu, idx);
 
-       if (sync && (curr->se.avg_overlap > sysctl_sched_migration_cost ||
-                       p->se.avg_overlap > sysctl_sched_migration_cost))
-               sync = 0;
+       if (sync) {
+              if (sched_feat(SYNC_LESS) &&
+                  (curr->se.avg_overlap > sysctl_sched_migration_cost ||
+                   p->se.avg_overlap > sysctl_sched_migration_cost))
+                      sync = 0;
+       } else {
+               if (sched_feat(SYNC_MORE) &&
+                   (curr->se.avg_overlap < sysctl_sched_migration_cost &&
+                    p->se.avg_overlap < sysctl_sched_migration_cost))
+                       sync = 1;
+       }
 
        /*
         * If sync wakeup then subtract the (maximum possible)

diff --git a/kernel/sched_features.h b/kernel/sched_features.h
index 294e10edd3c8bc67e135cce272c772e9ca67379e..70115c69c7a96da02c742bd9a6d9d751fccc6e0d 100644 (file)
--- a/kernel/sched_features.h
+++ b/kernel/sched_features.h
@@ -62,6 +62,16 @@ SCHED_FEAT(SYNC_WAKEUPS, 1)
  */
 SCHED_FEAT(AFFINE_WAKEUPS, 1)
 
+/*
+ * Weaken SYNC hint based on overlap
+ */
+SCHED_FEAT(SYNC_LESS, 1)
+
+/*
+ * Add SYNC hint based on overlap
+ */
+SCHED_FEAT(SYNC_MORE, 0)
+
 /*
  * Prefer to schedule the task we woke last (assuming it failed
  * wakeup-preemption), since its likely going to consume data we