sched: make the scheduler converge to the ideal latency

de-HZ-ification of the granularity defaults unearthed a pre-existing
property of CFS: while it correctly converges to the granularity goal,
it does not prevent run-time fluctuations in the range of
[-gran ... 0 ... +gran].

With the increase of the granularity due to the removal of HZ
dependencies, this becomes visible in chew-max output (with 5 tasks
running):

 out:  28 . 27. 32 | flu:  0 .  0 | ran:    9 .   13 | per:   37 .   40
 out:  27 . 27. 32 | flu:  0 .  0 | ran:   17 .   13 | per:   44 .   40
 out:  27 . 27. 32 | flu:  0 .  0 | ran:    9 .   13 | per:   36 .   40
 out:  29 . 27. 32 | flu:  2 .  0 | ran:   17 .   13 | per:   46 .   40
 out:  28 . 27. 32 | flu:  0 .  0 | ran:    9 .   13 | per:   37 .   40
 out:  29 . 27. 32 | flu:  0 .  0 | ran:   18 .   13 | per:   47 .   40
 out:  28 . 27. 32 | flu:  0 .  0 | ran:    9 .   13 | per:   37 .   40

average slice is the ideal 13 msecs and the period is picture-perfect 40
msecs. But the 'ran' field fluctuates around 13.33 msecs and there's no
mechanism in CFS to keep that from happening: it's a perfectly valid
solution that CFS finds.

to fix this we add a granularity/preemption rule that knows about
the "target latency", which makes tasks that run longer than the ideal
latency run a bit less. The simplest approach is to simply decrease the
preemption granularity when a task overruns its ideal latency. For this
we have to track how much the task executed since its last preemption.

( this adds a new field to task_struct, but we can eliminate that
  overhead in 2.6.24 by putting all the scheduler timestamps into an
  anonymous union. )

with this change in place, chew-max output is fluctuation-less all
around:

 out:  28 . 27. 39 | flu:  0 .  2 | ran:   13 .   13 | per:   41 .   40
 out:  28 . 27. 39 | flu:  0 .  2 | ran:   13 .   13 | per:   41 .   40
 out:  28 . 27. 39 | flu:  0 .  2 | ran:   13 .   13 | per:   41 .   40
 out:  28 . 27. 39 | flu:  0 .  2 | ran:   13 .   13 | per:   41 .   40
 out:  28 . 27. 39 | flu:  0 .  1 | ran:   13 .   13 | per:   41 .   40
 out:  28 . 27. 39 | flu:  0 .  1 | ran:   13 .   13 | per:   41 .   40

this patch has no impact on any fastpath or on any globally observable
scheduling property. (unless you have sharp enough eyes to see
millisecond-level ruckles in glxgears smoothness :-)

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

3 files changed, 24 insertions, 4 deletions

diff --git a/include/linux/sched.h b/include/linux/sched.h
index bd6a0320a77..f4e324ed2e4 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -904,6 +904,7 @@ struct sched_entity {
 
 	u64			exec_start;
 	u64			sum_exec_runtime;
+	u64			prev_sum_exec_runtime;
 	u64			wait_start_fair;
 	u64			sleep_start_fair;
 
diff --git a/kernel/sched.c b/kernel/sched.c
index 9fe473a190d..b533d6db78a 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -1587,6 +1587,7 @@ static void __sched_fork(struct task_struct *p)
 	p->se.wait_start_fair		= 0;
 	p->se.exec_start		= 0;
 	p->se.sum_exec_runtime		= 0;
+	p->se.prev_sum_exec_runtime	= 0;
 	p->se.delta_exec		= 0;
 	p->se.delta_fair_run		= 0;
 	p->se.delta_fair_sleep		= 0;
diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c
index 9f53d49f3aa..721fe774487 100644
--- a/kernel/sched_fair.c
+++ b/kernel/sched_fair.c
@@ -668,7 +668,7 @@ dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int sleep)
 /*
  * Preempt the current task with a newly woken task if needed:
  */
-static void
+static int
 __check_preempt_curr_fair(struct cfs_rq *cfs_rq, struct sched_entity *se,
 			  struct sched_entity *curr, unsigned long granularity)
 {
@@ -679,8 +679,11 @@ __check_preempt_curr_fair(struct cfs_rq *cfs_rq, struct sched_entity *se,
 	 * preempt the current task unless the best task has
 	 * a larger than sched_granularity fairness advantage:
 	 */
-	if (__delta > niced_granularity(curr, granularity))
+	if (__delta > niced_granularity(curr, granularity)) {
 		resched_task(rq_of(cfs_rq)->curr);
+		return 1;
+	}
+	return 0;
 }
 
 static inline void
@@ -725,6 +728,7 @@ static void put_prev_entity(struct cfs_rq *cfs_rq, struct sched_entity *prev)
 
 static void entity_tick(struct cfs_rq *cfs_rq, struct sched_entity *curr)
 {
+	unsigned long gran, ideal_runtime, delta_exec;
 	struct sched_entity *next;
 
 	/*
@@ -741,8 +745,22 @@ static void entity_tick(struct cfs_rq *cfs_rq, struct sched_entity *curr)
 	if (next == curr)
 		return;
 
-	__check_preempt_curr_fair(cfs_rq, next, curr,
-			sched_granularity(cfs_rq));
+	gran = sched_granularity(cfs_rq);
+	ideal_runtime = niced_granularity(curr,
+		max(sysctl_sched_latency / cfs_rq->nr_running,
+		    (unsigned long)sysctl_sched_min_granularity));
+	/*
+	 * If we executed more than what the latency constraint suggests,
+	 * reduce the rescheduling granularity. This way the total latency
+	 * of how much a task is not scheduled converges to
+	 * sysctl_sched_latency:
+	 */
+	delta_exec = curr->sum_exec_runtime - curr->prev_sum_exec_runtime;
+	if (delta_exec > ideal_runtime)
+		gran = 0;
+
+	if (__check_preempt_curr_fair(cfs_rq, next, curr, gran))
+		curr->prev_sum_exec_runtime = curr->sum_exec_runtime;
 }
 
 /**************************************************