X-Git-Url: https://err.no/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=kernel%2Fsched_fair.c;h=da7c061e72062eacd6afee8a9870e9044ad0e2c5;hb=b5677d848cbb94220ac2cfd36d93bcdbe49c3280;hp=c495dcf7031b2ec62200f68bd3fff8f6a87df832;hpb=b2be5e96dc0b5a179cf4cb98e65cfb605752ca26;p=linux-2.6

diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c
index c495dcf703..da7c061e72 100644
--- a/kernel/sched_fair.c
+++ b/kernel/sched_fair.c
@@ -22,7 +22,7 @@
 
 /*
  * Targeted preemption latency for CPU-bound tasks:
- * (default: 20ms, units: nanoseconds)
+ * (default: 20ms * (1 + ilog(ncpus)), units: nanoseconds)
  *
  * NOTE: this latency value is not the same as the concept of
  * 'timeslice length' - timeslices in CFS are of variable length
@@ -32,18 +32,18 @@
  * (to see the precise effective timeslice length of your workload,
  *  run vmstat and monitor the context-switches (cs) field)
  */
-const_debug unsigned int sysctl_sched_latency = 20000000ULL;
+unsigned int sysctl_sched_latency = 20000000ULL;
 
 /*
  * Minimal preemption granularity for CPU-bound tasks:
- * (default: 1 msec, units: nanoseconds)
+ * (default: 4 msec * (1 + ilog(ncpus)), units: nanoseconds)
  */
-const_debug unsigned int sysctl_sched_min_granularity = 1000000ULL;
+unsigned int sysctl_sched_min_granularity = 4000000ULL;
 
 /*
  * is kept at sysctl_sched_latency / sysctl_sched_min_granularity
  */
-const_debug unsigned int sched_nr_latency = 20;
+static unsigned int sched_nr_latency = 5;
 
 /*
  * After fork, child runs first. (default) If set to 0 then
@@ -61,23 +61,23 @@ unsigned int __read_mostly sysctl_sched_compat_yield;
 
 /*
  * SCHED_BATCH wake-up granularity.
- * (default: 10 msec, units: nanoseconds)
+ * (default: 10 msec * (1 + ilog(ncpus)), units: nanoseconds)
  *
  * This option delays the preemption effects of decoupled workloads
  * and reduces their over-scheduling. Synchronous workloads will still
  * have immediate wakeup/sleep latencies.
  */
-const_debug unsigned int sysctl_sched_batch_wakeup_granularity = 10000000UL;
+unsigned int sysctl_sched_batch_wakeup_granularity = 10000000UL;
 
 /*
  * SCHED_OTHER wake-up granularity.
- * (default: 10 msec, units: nanoseconds)
+ * (default: 10 msec * (1 + ilog(ncpus)), units: nanoseconds)
  *
  * This option delays the preemption effects of decoupled workloads
  * and reduces their over-scheduling. Synchronous workloads will still
  * have immediate wakeup/sleep latencies.
  */
-const_debug unsigned int sysctl_sched_wakeup_granularity = 10000000UL;
+unsigned int sysctl_sched_wakeup_granularity = 10000000UL;
 
 const_debug unsigned int sysctl_sched_migration_cost = 500000UL;
 
@@ -351,6 +351,12 @@ static void update_curr(struct cfs_rq *cfs_rq)
 
 	__update_curr(cfs_rq, curr, delta_exec);
 	curr->exec_start = now;
+
+	if (entity_is_task(curr)) {
+		struct task_struct *curtask = task_of(curr);
+
+		cpuacct_charge(curtask, delta_exec);
+	}
 }
 
 static inline void
@@ -505,8 +511,7 @@ place_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int initial)
 
 	if (!initial) {
 		/* sleeps upto a single latency don't count. */
-		if (sched_feat(NEW_FAIR_SLEEPERS) && entity_is_task(se) &&
-				task_of(se)->policy != SCHED_BATCH)
+		if (sched_feat(NEW_FAIR_SLEEPERS) && entity_is_task(se))
 			vruntime -= sysctl_sched_latency;
 
 		/* ensure we never gain time by being placed backwards. */
@@ -546,7 +551,6 @@ dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int sleep)
 
 	update_stats_dequeue(cfs_rq, se);
 	if (sleep) {
-		se->peer_preempt = 0;
 #ifdef CONFIG_SCHEDSTATS
 		if (entity_is_task(se)) {
 			struct task_struct *tsk = task_of(se);
@@ -574,10 +578,8 @@ check_preempt_tick(struct cfs_rq *cfs_rq, struct sched_entity *curr)
 
 	ideal_runtime = sched_slice(cfs_rq, curr);
 	delta_exec = curr->sum_exec_runtime - curr->prev_sum_exec_runtime;
-	if (delta_exec > ideal_runtime ||
-			(sched_feat(PREEMPT_RESTRICT) && curr->peer_preempt))
+	if (delta_exec > ideal_runtime)
 		resched_task(rq_of(cfs_rq)->curr);
-	curr->peer_preempt = 0;
 }
 
 static void
@@ -796,8 +798,9 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int sleep)
  */
 static void yield_task_fair(struct rq *rq)
 {
-	struct cfs_rq *cfs_rq = task_cfs_rq(rq->curr);
-	struct sched_entity *rightmost, *se = &rq->curr->se;
+	struct task_struct *curr = rq->curr;
+	struct cfs_rq *cfs_rq = task_cfs_rq(curr);
+	struct sched_entity *rightmost, *se = &curr->se;
 
 	/*
 	 * Are we the only task in the tree?
@@ -805,7 +808,7 @@ static void yield_task_fair(struct rq *rq)
 	if (unlikely(cfs_rq->nr_running == 1))
 		return;
 
-	if (likely(!sysctl_sched_compat_yield)) {
+	if (likely(!sysctl_sched_compat_yield) && curr->policy != SCHED_BATCH) {
 		__update_rq_clock(rq);
 		/*
 		 * Update run-time statistics of the 'current'.
@@ -840,7 +843,7 @@ static void check_preempt_wakeup(struct rq *rq, struct task_struct *p)
 	struct task_struct *curr = rq->curr;
 	struct cfs_rq *cfs_rq = task_cfs_rq(curr);
 	struct sched_entity *se = &curr->se, *pse = &p->se;
-	s64 delta, gran;
+	unsigned long gran;
 
 	if (unlikely(rt_prio(p->prio))) {
 		update_rq_clock(rq);
@@ -855,24 +858,20 @@ static void check_preempt_wakeup(struct rq *rq, struct task_struct *p)
 	if (unlikely(p->policy == SCHED_BATCH))
 		return;
 
-	if (sched_feat(WAKEUP_PREEMPT)) {
-		while (!is_same_group(se, pse)) {
-			se = parent_entity(se);
-			pse = parent_entity(pse);
-		}
+	if (!sched_feat(WAKEUP_PREEMPT))
+		return;
 
-		delta = se->vruntime - pse->vruntime;
-		gran = sysctl_sched_wakeup_granularity;
-		if (unlikely(se->load.weight != NICE_0_LOAD))
-			gran = calc_delta_fair(gran, &se->load);
+	while (!is_same_group(se, pse)) {
+		se = parent_entity(se);
+		pse = parent_entity(pse);
+	}
 
-		if (delta > gran) {
-			int now = !sched_feat(PREEMPT_RESTRICT);
+	gran = sysctl_sched_wakeup_granularity;
+	if (unlikely(se->load.weight != NICE_0_LOAD))
+		gran = calc_delta_fair(gran, &se->load);
 
-			if (now || p->prio < curr->prio || !se->peer_preempt++)
-				resched_task(curr);
-		}
-	}
+	if (pse->vruntime + gran < se->vruntime)
+		resched_task(curr);
 }
 
 static struct task_struct *pick_next_task_fair(struct rq *rq)
@@ -1074,8 +1073,9 @@ static void task_new_fair(struct rq *rq, struct task_struct *p)
 	update_curr(cfs_rq);
 	place_entity(cfs_rq, se, 1);
 
+	/* 'curr' will be NULL if the child belongs to a different group */
 	if (sysctl_sched_child_runs_first && this_cpu == task_cpu(p) &&
-			curr->vruntime < se->vruntime) {
+			curr && curr->vruntime < se->vruntime) {
 		/*
 		 * Upon rescheduling, sched_class::put_prev_task() will place
 		 * 'current' within the tree based on its new key value.
@@ -1083,7 +1083,6 @@ static void task_new_fair(struct rq *rq, struct task_struct *p)
 		swap(curr->vruntime, se->vruntime);
 	}
 
-	se->peer_preempt = 0;
 	enqueue_task_fair(rq, p, 0);
 	resched_task(rq->curr);
 }