sched: virtual time buddy preemption

Since we moved wakeup preemption back to virtual time, it makes sense to move
the buddy stuff back as well. The purpose of the buddy scheduling is to allow
a quickly scheduling pair of tasks to run away from the group as far as a
regular busy task would be allowed under wakeup preemption.

This has the advantage that the pair can ping-pong for a while, enjoying
cache-hotness. Without buddy scheduling other tasks would interleave destroying
the cache.

Also, it saves a word in cfs_rq.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

diff --git a/kernel/sched.c b/kernel/sched.c
index 12bc367d92413705e044539af6489c311a4e9413..e8819bc6f462c18761c11861b4808fd28223b431 100644 (file)
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -386,7 +386,6 @@ struct cfs_rq {
 
        u64 exec_clock;
        u64 min_vruntime;
-       u64 pair_start;
 
        struct rb_root tasks_timeline;
        struct rb_node *rb_leftmost;

diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c
index 7af17e04a6db86cc6b753e0d5db68b13e4acc240..ce514afd78ff7998338b6661ae22475e808fa3fa 100644 (file)
--- a/kernel/sched_fair.c
+++ b/kernel/sched_fair.c
@@ -791,16 +791,14 @@ set_next_entity(struct cfs_rq *cfs_rq, struct sched_entity *se)
        se->prev_sum_exec_runtime = se->sum_exec_runtime;
 }
 
+static int
+wakeup_preempt_entity(struct sched_entity *curr, struct sched_entity *se);
+
 static struct sched_entity *
 pick_next(struct cfs_rq *cfs_rq, struct sched_entity *se)
 {
-       struct rq *rq = rq_of(cfs_rq);
-       u64 pair_slice = rq->clock - cfs_rq->pair_start;
-
-       if (!cfs_rq->next || pair_slice > sysctl_sched_min_granularity) {
-               cfs_rq->pair_start = rq->clock;
+       if (!cfs_rq->next || wakeup_preempt_entity(cfs_rq->next, se) == 1)
                return se;
-       }
 
        return cfs_rq->next;
 }