X-Git-Url: http://git.rot13.org/?a=blobdiff_plain;f=kernel%2Fsched_fair.c;h=de13a6f5b9777e2ed899fec5c23b61714b39ab15;hb=b9fa3df33f9166daf81bfa8253d339f5a7726122;hp=827a0636525a36e35ead96313235a54ba143021e;hpb=87fefa381ef27f46c1182622ea01eb9504cd2e24;p=powerpc.git

diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c
index 827a063652..de13a6f5b9 100644
--- a/kernel/sched_fair.c
+++ b/kernel/sched_fair.c
@@ -46,7 +46,7 @@ const_debug unsigned int sysctl_sched_child_runs_first = 1;
  * Minimal preemption granularity for CPU-bound tasks:
  * (default: 2 msec, units: nanoseconds)
  */
-unsigned int sysctl_sched_min_granularity __read_mostly = 2000000ULL;
+const_debug unsigned int sysctl_sched_nr_latency = 20;
 
 /*
  * sys_sched_yield() compat mode
@@ -76,8 +76,6 @@ const_debug unsigned int sysctl_sched_batch_wakeup_granularity = 25000000UL;
  */
 const_debug unsigned int sysctl_sched_wakeup_granularity = 2000000UL;
 
-unsigned int sysctl_sched_runtime_limit __read_mostly;
-
 extern struct sched_class fair_sched_class;
 
 /**************************************************************
@@ -119,21 +117,21 @@ static inline struct task_struct *task_of(struct sched_entity *se)
 static inline u64
 max_vruntime(u64 min_vruntime, u64 vruntime)
 {
-	if ((vruntime > min_vruntime) ||
-	    (min_vruntime > (1ULL << 61) && vruntime < (1ULL << 50)))
+	s64 delta = (s64)(vruntime - min_vruntime);
+	if (delta > 0)
 		min_vruntime = vruntime;
 
 	return min_vruntime;
 }
 
-static inline void
-set_leftmost(struct cfs_rq *cfs_rq, struct rb_node *leftmost)
+static inline u64
+min_vruntime(u64 min_vruntime, u64 vruntime)
 {
-	struct sched_entity *se;
+	s64 delta = (s64)(vruntime - min_vruntime);
+	if (delta < 0)
+		min_vruntime = vruntime;
 
-	cfs_rq->rb_leftmost = leftmost;
-	if (leftmost)
-		se = rb_entry(leftmost, struct sched_entity, run_node);
+	return min_vruntime;
 }
 
 static inline s64
@@ -177,7 +175,7 @@ __enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se)
 	 * used):
 	 */
 	if (leftmost)
-		set_leftmost(cfs_rq, &se->run_node);
+		cfs_rq->rb_leftmost = &se->run_node;
 
 	rb_link_node(&se->run_node, parent, link);
 	rb_insert_color(&se->run_node, &cfs_rq->tasks_timeline);
@@ -187,7 +185,7 @@ static void
 __dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se)
 {
 	if (cfs_rq->rb_leftmost == &se->run_node)
-		set_leftmost(cfs_rq, rb_next(&se->run_node));
+		cfs_rq->rb_leftmost = rb_next(&se->run_node);
 
 	rb_erase(&se->run_node, &cfs_rq->tasks_timeline);
 }
@@ -224,8 +222,7 @@ static inline struct sched_entity *__pick_last_entity(struct cfs_rq *cfs_rq)
 static u64 __sched_period(unsigned long nr_running)
 {
 	u64 period = sysctl_sched_latency;
-	unsigned long nr_latency =
-		sysctl_sched_latency / sysctl_sched_min_granularity;
+	unsigned long nr_latency = sysctl_sched_nr_latency;
 
 	if (unlikely(nr_running > nr_latency)) {
 		period *= nr_running;
@@ -245,6 +242,19 @@ static u64 sched_slice(struct cfs_rq *cfs_rq, struct sched_entity *se)
 	return period;
 }
 
+static u64 __sched_vslice(unsigned long nr_running)
+{
+	unsigned long period = sysctl_sched_latency;
+	unsigned long nr_latency = sysctl_sched_nr_latency;
+
+	if (unlikely(nr_running > nr_latency))
+		nr_running = nr_latency;
+
+	period /= nr_running;
+
+	return (u64)period;
+}
+
 /*
  * Update the current task's runtime statistics. Skip current tasks that
  * are not in our scheduling class.
@@ -254,7 +264,7 @@ __update_curr(struct cfs_rq *cfs_rq, struct sched_entity *curr,
 	      unsigned long delta_exec)
 {
 	unsigned long delta_exec_weighted;
-	u64 next_vruntime, min_vruntime;
+	u64 vruntime;
 
 	schedstat_set(curr->exec_max, max((u64)delta_exec, curr->exec_max));
 
@@ -272,19 +282,13 @@ __update_curr(struct cfs_rq *cfs_rq, struct sched_entity *curr,
 	 * value tracking the leftmost vruntime in the tree.
 	 */
 	if (first_fair(cfs_rq)) {
-		next_vruntime = __pick_next_entity(cfs_rq)->vruntime;
-
-		/* min_vruntime() := !max_vruntime() */
-		min_vruntime = max_vruntime(curr->vruntime, next_vruntime);
-		if (min_vruntime == next_vruntime)
-			min_vruntime = curr->vruntime;
-		else
-			min_vruntime = next_vruntime;
+		vruntime = min_vruntime(curr->vruntime,
+				__pick_next_entity(cfs_rq)->vruntime);
 	} else
-		min_vruntime = curr->vruntime;
+		vruntime = curr->vruntime;
 
 	cfs_rq->min_vruntime =
-		max_vruntime(cfs_rq->min_vruntime, min_vruntime);
+		max_vruntime(cfs_rq->min_vruntime, vruntime);
 }
 
 static void update_curr(struct cfs_rq *cfs_rq)
@@ -440,35 +444,47 @@ static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se)
 #endif
 }
 
+static void check_spread(struct cfs_rq *cfs_rq, struct sched_entity *se)
+{
+#ifdef CONFIG_SCHED_DEBUG
+	s64 d = se->vruntime - cfs_rq->min_vruntime;
+
+	if (d < 0)
+		d = -d;
+
+	if (d > 3*sysctl_sched_latency)
+		schedstat_inc(cfs_rq, nr_spread_over);
+#endif
+}
+
 static void
 place_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int initial)
 {
-	u64 min_runtime, latency;
+	u64 vruntime;
 
-	min_runtime = cfs_rq->min_vruntime;
+	vruntime = cfs_rq->min_vruntime;
 
 	if (sched_feat(USE_TREE_AVG)) {
 		struct sched_entity *last = __pick_last_entity(cfs_rq);
 		if (last) {
-			min_runtime = __pick_next_entity(cfs_rq)->vruntime;
-			min_runtime += last->vruntime;
-			min_runtime >>= 1;
+			vruntime += last->vruntime;
+			vruntime >>= 1;
 		}
-	} else if (sched_feat(APPROX_AVG))
-		min_runtime += sysctl_sched_latency/2;
+	} else if (sched_feat(APPROX_AVG) && cfs_rq->nr_running)
+		vruntime += __sched_vslice(cfs_rq->nr_running)/2;
 
 	if (initial && sched_feat(START_DEBIT))
-		min_runtime += sched_slice(cfs_rq, se);
-
-	if (!initial && sched_feat(NEW_FAIR_SLEEPERS)) {
-		latency = sysctl_sched_latency;
-		if (min_runtime > latency)
-			min_runtime -= latency;
-		else
-			min_runtime = 0;
+		vruntime += __sched_vslice(cfs_rq->nr_running + 1);
+
+	if (!initial) {
+		if (sched_feat(NEW_FAIR_SLEEPERS))
+			vruntime -= sysctl_sched_latency;
+
+		vruntime = max_t(s64, vruntime, se->vruntime);
 	}
 
-	se->vruntime = max(se->vruntime, min_runtime);
+	se->vruntime = vruntime;
+
 }
 
 static void
@@ -485,6 +501,7 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int wakeup)
 	}
 
 	update_stats_enqueue(cfs_rq, se);
+	check_spread(cfs_rq, se);
 	if (se != cfs_rq->curr)
 		__enqueue_entity(cfs_rq, se);
 	account_entity_enqueue(cfs_rq, se);
@@ -494,8 +511,8 @@ static void
 dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int sleep)
 {
 	update_stats_dequeue(cfs_rq, se);
-#ifdef CONFIG_SCHEDSTATS
 	if (sleep) {
+#ifdef CONFIG_SCHEDSTATS
 		if (entity_is_task(se)) {
 			struct task_struct *tsk = task_of(se);
 
@@ -504,8 +521,9 @@ dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int sleep)
 			if (tsk->state & TASK_UNINTERRUPTIBLE)
 				se->block_start = rq_of(cfs_rq)->clock;
 		}
-	}
 #endif
+	}
+
 	if (se != cfs_rq->curr)
 		__dequeue_entity(cfs_rq, se);
 	account_entity_dequeue(cfs_rq, se);
@@ -525,15 +543,20 @@ check_preempt_tick(struct cfs_rq *cfs_rq, struct sched_entity *curr)
 		resched_task(rq_of(cfs_rq)->curr);
 }
 
-static inline void
+static void
 set_next_entity(struct cfs_rq *cfs_rq, struct sched_entity *se)
 {
-	/*
-	 * Any task has to be enqueued before it get to execute on
-	 * a CPU. So account for the time it spent waiting on the
-	 * runqueue.
-	 */
-	update_stats_wait_end(cfs_rq, se);
+	/* 'current' is not kept within the tree. */
+	if (se->on_rq) {
+		/*
+		 * Any task has to be enqueued before it get to execute on
+		 * a CPU. So account for the time it spent waiting on the
+		 * runqueue.
+		 */
+		update_stats_wait_end(cfs_rq, se);
+		__dequeue_entity(cfs_rq, se);
+	}
+
 	update_stats_curr_start(cfs_rq, se);
 	cfs_rq->curr = se;
 #ifdef CONFIG_SCHEDSTATS
@@ -554,10 +577,6 @@ static struct sched_entity *pick_next_entity(struct cfs_rq *cfs_rq)
 {
 	struct sched_entity *se = __pick_next_entity(cfs_rq);
 
-	/* 'current' is not kept within the tree. */
-	if (se)
-		__dequeue_entity(cfs_rq, se);
-
 	set_next_entity(cfs_rq, se);
 
 	return se;
@@ -574,6 +593,7 @@ static void put_prev_entity(struct cfs_rq *cfs_rq, struct sched_entity *prev)
 
 	update_stats_curr_end(cfs_rq, prev);
 
+	check_spread(cfs_rq, prev);
 	if (prev->on_rq) {
 		update_stats_wait_start(cfs_rq, prev);
 		/* Put 'current' back into the tree. */
@@ -632,15 +652,21 @@ static inline struct cfs_rq *cpu_cfs_rq(struct cfs_rq *cfs_rq, int this_cpu)
 #define for_each_leaf_cfs_rq(rq, cfs_rq) \
 	list_for_each_entry(cfs_rq, &rq->leaf_cfs_rq_list, leaf_cfs_rq_list)
 
-/* Do the two (enqueued) tasks belong to the same group ? */
-static inline int is_same_group(struct task_struct *curr, struct task_struct *p)
+/* Do the two (enqueued) entities belong to the same group ? */
+static inline int
+is_same_group(struct sched_entity *se, struct sched_entity *pse)
 {
-	if (curr->se.cfs_rq == p->se.cfs_rq)
+	if (se->cfs_rq == pse->cfs_rq)
 		return 1;
 
 	return 0;
 }
 
+static inline struct sched_entity *parent_entity(struct sched_entity *se)
+{
+	return se->parent;
+}
+
 #else	/* CONFIG_FAIR_GROUP_SCHED */
 
 #define for_each_sched_entity(se) \
@@ -673,11 +699,17 @@ static inline struct cfs_rq *cpu_cfs_rq(struct cfs_rq *cfs_rq, int this_cpu)
 #define for_each_leaf_cfs_rq(rq, cfs_rq) \
 		for (cfs_rq = &rq->cfs; cfs_rq; cfs_rq = NULL)
 
-static inline int is_same_group(struct task_struct *curr, struct task_struct *p)
+static inline int
+is_same_group(struct sched_entity *se, struct sched_entity *pse)
 {
 	return 1;
 }
 
+static inline struct sched_entity *parent_entity(struct sched_entity *se)
+{
+	return NULL;
+}
+
 #endif	/* CONFIG_FAIR_GROUP_SCHED */
 
 /*
@@ -695,6 +727,7 @@ static void enqueue_task_fair(struct rq *rq, struct task_struct *p, int wakeup)
 			break;
 		cfs_rq = cfs_rq_of(se);
 		enqueue_entity(cfs_rq, se, wakeup);
+		wakeup = 1;
 	}
 }
 
@@ -714,6 +747,7 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int sleep)
 		/* Don't dequeue parent if it has other entities besides us */
 		if (cfs_rq->load.weight)
 			break;
+		sleep = 1;
 	}
 }
 
@@ -722,12 +756,10 @@ static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int sleep)
  *
  * If compat_yield is turned on then we requeue to the end of the tree.
  */
-static void yield_task_fair(struct rq *rq, struct task_struct *p)
+static void yield_task_fair(struct rq *rq)
 {
-	struct cfs_rq *cfs_rq = task_cfs_rq(p);
-	struct rb_node **link = &cfs_rq->tasks_timeline.rb_node;
-	struct sched_entity *rightmost, *se = &p->se;
-	struct rb_node *parent;
+	struct cfs_rq *cfs_rq = task_cfs_rq(rq->curr);
+	struct sched_entity *rightmost, *se = &rq->curr->se;
 
 	/*
 	 * Are we the only task in the tree?
@@ -741,39 +773,26 @@ static void yield_task_fair(struct rq *rq, struct task_struct *p)
 		 * Dequeue and enqueue the task to update its
 		 * position within the tree:
 		 */
-		dequeue_entity(cfs_rq, &p->se, 0);
-		enqueue_entity(cfs_rq, &p->se, 0);
+		update_curr(cfs_rq);
 
 		return;
 	}
 	/*
 	 * Find the rightmost entry in the rbtree:
 	 */
-	do {
-		parent = *link;
-		link = &parent->rb_right;
-	} while (*link);
-
-	rightmost = rb_entry(parent, struct sched_entity, run_node);
+	rightmost = __pick_last_entity(cfs_rq);
 	/*
 	 * Already in the rightmost position?
 	 */
-	if (unlikely(rightmost == se))
+	if (unlikely(rightmost->vruntime < se->vruntime))
 		return;
 
 	/*
 	 * Minimally necessary key value to be last in the tree:
+	 * Upon rescheduling, sched_class::put_prev_task() will place
+	 * 'current' within the tree based on its new key value.
 	 */
 	se->vruntime = rightmost->vruntime + 1;
-
-	if (cfs_rq->rb_leftmost == &se->run_node)
-		cfs_rq->rb_leftmost = rb_next(&se->run_node);
-	/*
-	 * Relink the task to the rightmost position:
-	 */
-	rb_erase(&se->run_node, &cfs_rq->tasks_timeline);
-	rb_link_node(&se->run_node, parent, link);
-	rb_insert_color(&se->run_node, &cfs_rq->tasks_timeline);
 }
 
 /*
@@ -783,6 +802,8 @@ 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;
 
 	if (unlikely(rt_prio(p->prio))) {
 		update_rq_clock(rq);
@@ -790,12 +811,16 @@ static void check_preempt_wakeup(struct rq *rq, struct task_struct *p)
 		resched_task(curr);
 		return;
 	}
-	if (is_same_group(curr, p)) {
-		s64 delta = curr->se.vruntime - p->se.vruntime;
 
-		if (delta > (s64)sysctl_sched_wakeup_granularity)
-			resched_task(curr);
+	while (!is_same_group(se, pse)) {
+		se = parent_entity(se);
+		pse = parent_entity(pse);
 	}
+
+	delta = se->vruntime - pse->vruntime;
+
+	if (delta > (s64)sysctl_sched_wakeup_granularity)
+		resched_task(curr);
 }
 
 static struct task_struct *pick_next_task_fair(struct rq *rq)
@@ -876,7 +901,10 @@ static int cfs_rq_best_prio(struct cfs_rq *cfs_rq)
 	if (!cfs_rq->nr_running)
 		return MAX_PRIO;
 
-	curr = __pick_next_entity(cfs_rq);
+	curr = cfs_rq->curr;
+	if (!curr)
+		curr = __pick_next_entity(cfs_rq);
+
 	p = task_of(curr);
 
 	return p->prio;
@@ -973,19 +1001,20 @@ static void task_new_fair(struct rq *rq, struct task_struct *p)
 	if (sysctl_sched_child_runs_first &&
 			curr->vruntime < se->vruntime) {
 		/*
- 		 * Upon rescheduling, sched_class::put_prev_task() will place
- 		 * 'current' within the tree based on its new key value.
- 		 */
+		 * Upon rescheduling, sched_class::put_prev_task() will place
+		 * 'current' within the tree based on its new key value.
+		 */
 		swap(curr->vruntime, se->vruntime);
 	}
 
 	update_stats_enqueue(cfs_rq, se);
+	check_spread(cfs_rq, se);
+	check_spread(cfs_rq, curr);
 	__enqueue_entity(cfs_rq, se);
 	account_entity_enqueue(cfs_rq, se);
 	resched_task(rq->curr);
 }
 
-#ifdef CONFIG_FAIR_GROUP_SCHED
 /* Account for a task changing its policy or group.
  *
  * This routine is mostly called to set cfs_rq->curr field when a task
@@ -998,15 +1027,6 @@ static void set_curr_task_fair(struct rq *rq)
 	for_each_sched_entity(se)
 		set_next_entity(cfs_rq_of(se), se);
 }
-#else
-static void set_curr_task_fair(struct rq *rq)
-{
-	struct sched_entity *se = &rq->curr->se;
-	struct cfs_rq *cfs_rq = cfs_rq_of(se);
-
-	cfs_rq->curr = se;
-}
-#endif
 
 /*
  * All the scheduling class methods:
@@ -1033,6 +1053,9 @@ static void print_cfs_stats(struct seq_file *m, int cpu)
 {
 	struct cfs_rq *cfs_rq;
 
+#ifdef CONFIG_FAIR_GROUP_SCHED
+	print_cfs_rq(m, cpu, &cpu_rq(cpu)->cfs);
+#endif
 	for_each_leaf_cfs_rq(cpu_rq(cpu), cfs_rq)
 		print_cfs_rq(m, cpu, cfs_rq);
 }