---
 kernel/sched/bfs.c |   49 +++++++++++++++++++------------------------------
 1 file changed, 19 insertions(+), 30 deletions(-)

Index: linux-4.1-bfs/kernel/sched/bfs.c
===================================================================
--- linux-4.1-bfs.orig/kernel/sched/bfs.c	2015-08-02 17:34:13.340388825 +1000
+++ linux-4.1-bfs/kernel/sched/bfs.c	2015-08-02 18:21:33.857686300 +1000
@@ -134,7 +134,7 @@
 
 void print_scheduler_version(void)
 {
-	printk(KERN_INFO "BFS CPU scheduler v0.462 by Con Kolivas.\n");
+	printk(KERN_INFO "BFS CPU scheduler v0.463 by Con Kolivas.\n");
 }
 
 /*
@@ -3361,18 +3361,16 @@ static void __sched __schedule(void)
 {
 	struct task_struct *prev, *next, *idle;
 	unsigned long *switch_count;
-	bool deactivate;
+	bool deactivate = false;
 	struct rq *rq;
 	int cpu;
 
-need_resched:
 	preempt_disable();
 	cpu = smp_processor_id();
 	rq = cpu_rq(cpu);
 	rcu_note_context_switch();
 	prev = rq->curr;
 
-	deactivate = false;
 	schedule_debug(prev);
 
 	/*
@@ -3413,17 +3411,6 @@ need_resched:
 		switch_count = &prev->nvcsw;
 	}
 
-	/*
-	 * If we are going to sleep and we have plugged IO queued, make
-	 * sure to submit it to avoid deadlocks. This usually clears before
-	 * grabbing the lock but still may rarely happen here. */
-	if (unlikely(deactivate && blk_needs_flush_plug(prev))) {
-		grq_unlock_irq();
-		preempt_enable_no_resched();
-		blk_schedule_flush_plug(prev);
-		goto need_resched;
-	}
-
 	update_clocks(rq);
 	update_cpu_clock_switch(rq, prev);
 	if (rq->clock - rq->last_tick > HALF_JIFFY_NS)
@@ -3912,8 +3899,8 @@ static inline struct task_struct *find_p
 }
 
 /* Actually do priority change: must hold grq lock. */
-static void
-__setscheduler(struct task_struct *p, struct rq *rq, int policy, int prio)
+static void __setscheduler(struct task_struct *p, struct rq *rq, int policy,
+			   int prio, bool keep_boost)
 {
 	int oldrtprio, oldprio;
 
@@ -3922,8 +3909,14 @@ __setscheduler(struct task_struct *p, st
 	p->rt_priority = prio;
 	p->normal_prio = normal_prio(p);
 	oldprio = p->prio;
-	/* we are holding p->pi_lock already */
-	p->prio = rt_mutex_getprio(p);
+	/*
+	 * Keep a potential priority boosting if called from
+	 * sched_setscheduler().
+	 */
+	if (keep_boost)
+		p->prio = rt_mutex_get_effective_prio(p, p->normal_prio);
+	else
+		p->prio = p->normal_prio;
 	if (task_running(p)) {
 		reset_rq_task(rq, p);
 		/* Resched only if we might now be preempted */
@@ -4104,7 +4097,7 @@ recheck:
 	queued = task_queued(p);
 	if (queued)
 		dequeue_task(p);
-	__setscheduler(p, rq, policy, param->sched_priority);
+	__setscheduler(p, rq, policy, param->sched_priority, true);
 	if (queued) {
 		enqueue_task(p, rq);
 		try_preempt(p, rq);
@@ -4825,22 +4818,18 @@ EXPORT_SYMBOL_GPL(yield_to);
 
 long __sched io_schedule_timeout(long timeout)
 {
-	struct task_struct *curr = current;
-	int old_iowait = curr->in_iowait;
+	int old_iowait = current->in_iowait;
 	struct rq *rq;
 	long ret;
 
-	curr->in_iowait = 1;
-	if (old_iowait)
-		blk_schedule_flush_plug(curr);
-	else
-		blk_flush_plug(curr);
+	current->in_iowait = 1;
+	blk_schedule_flush_plug(current);
 
 	delayacct_blkio_start();
 	rq = raw_rq();
 	atomic_inc(&rq->nr_iowait);
 	ret = schedule_timeout(timeout);
-	curr->in_iowait = old_iowait;
+	current->in_iowait = old_iowait;
 	atomic_dec(&rq->nr_iowait);
 	delayacct_blkio_end();
 
@@ -6153,6 +6142,7 @@ sd_init(struct sched_domain_topology_lev
 	 */
 
 	if (sd->flags & SD_SHARE_CPUCAPACITY) {
+		sd->flags |= SD_PREFER_SIBLING;
 		sd->imbalance_pct = 110;
 		sd->smt_gain = 1178; /* ~15% */
 
@@ -6809,7 +6799,6 @@ static int cpuset_cpu_active(struct noti
 		 */
 
 	case CPU_ONLINE:
-	case CPU_DOWN_FAILED:
 		cpuset_update_active_cpus(true);
 		break;
 	default:
@@ -7166,7 +7155,7 @@ void normalize_rt_tasks(void)
 		queued = task_queued(p);
 		if (queued)
 			dequeue_task(p);
-		__setscheduler(p, rq, SCHED_NORMAL, 0);
+		__setscheduler(p, rq, SCHED_NORMAL, 0, false);
 		if (queued) {
 			enqueue_task(p, rq);
 			try_preempt(p, rq);