5 files changed, 553 insertions, 193 deletions

diff --git a/kernel/sched.c b/kernel/sched.c
index 8ee437a5ec1d..fcc3483e9955 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -467,11 +467,17 @@ struct rt_rq {
         struct rt_prio_array active;
         unsigned long rt_nr_running;
 #if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED
-        int highest_prio; /* highest queued rt task prio */
+        struct {
+                int curr; /* highest queued rt task prio */
+#ifdef CONFIG_SMP
+                int next; /* next highest */
+#endif
+        } highest_prio;
 #endif
 #ifdef CONFIG_SMP
         unsigned long rt_nr_migratory;
         int overloaded;
+        struct plist_head pushable_tasks;
 #endif
         int rt_throttled;
         u64 rt_time;
@@ -1610,21 +1616,42 @@ static inline void update_shares_locked(struct rq *rq, struct sched_domain *sd)
 
 #endif
 
+#ifdef CONFIG_PREEMPT
+
 /*
- * double_lock_balance - lock the busiest runqueue, this_rq is locked already.
+ * fair double_lock_balance: Safely acquires both rq->locks in a fair
+ * way at the expense of forcing extra atomic operations in all
+ * invocations.  This assures that the double_lock is acquired using the
+ * same underlying policy as the spinlock_t on this architecture, which
+ * reduces latency compared to the unfair variant below.  However, it
+ * also adds more overhead and therefore may reduce throughput.
  */
-static int double_lock_balance(struct rq *this_rq, struct rq *busiest)
+static inline int _double_lock_balance(struct rq *this_rq, struct rq *busiest)
+        __releases(this_rq->lock)
+        __acquires(busiest->lock)
+        __acquires(this_rq->lock)
+{
+        spin_unlock(&this_rq->lock);
+        double_rq_lock(this_rq, busiest);
+
+        return 1;
+}
+
+#else
+/*
+ * Unfair double_lock_balance: Optimizes throughput at the expense of
+ * latency by eliminating extra atomic operations when the locks are
+ * already in proper order on entry.  This favors lower cpu-ids and will
+ * grant the double lock to lower cpus over higher ids under contention,
+ * regardless of entry order into the function.
+ */
+static int _double_lock_balance(struct rq *this_rq, struct rq *busiest)
         __releases(this_rq->lock)
         __acquires(busiest->lock)
         __acquires(this_rq->lock)
 {
         int ret = 0;
 
-        if (unlikely(!irqs_disabled())) {
-                /* printk() doesn't work good under rq->lock */
-                spin_unlock(&this_rq->lock);
-                BUG_ON(1);
-        }
         if (unlikely(!spin_trylock(&busiest->lock))) {
                 if (busiest < this_rq) {
                         spin_unlock(&this_rq->lock);
@@ -1637,6 +1664,22 @@ static int double_lock_balance(struct rq *this_rq, struct rq *busiest)
         return ret;
 }
 
+#endif /* CONFIG_PREEMPT */
+
+/*
+ * double_lock_balance - lock the busiest runqueue, this_rq is locked already.
+ */
+static int double_lock_balance(struct rq *this_rq, struct rq *busiest)
+{
+        if (unlikely(!irqs_disabled())) {
+                /* printk() doesn't work good under rq->lock */
+                spin_unlock(&this_rq->lock);
+                BUG_ON(1);
+        }
+
+        return _double_lock_balance(this_rq, busiest);
+}
+
 static inline void double_unlock_balance(struct rq *this_rq, struct rq *busiest)
         __releases(busiest->lock)
 {
@@ -1705,6 +1748,9 @@ static void update_avg(u64 *avg, u64 sample)
 
 static void enqueue_task(struct rq *rq, struct task_struct *p, int wakeup)
 {
+        if (wakeup)
+                p->se.start_runtime = p->se.sum_exec_runtime;
+
         sched_info_queued(p);
         p->sched_class->enqueue_task(rq, p, wakeup);
         p->se.on_rq = 1;
@@ -1712,10 +1758,15 @@ static void enqueue_task(struct rq *rq, struct task_struct *p, int wakeup)
 
 static void dequeue_task(struct rq *rq, struct task_struct *p, int sleep)
 {
-        if (sleep && p->se.last_wakeup) {
-                update_avg(&p->se.avg_overlap,
-                           p->se.sum_exec_runtime - p->se.last_wakeup);
-                p->se.last_wakeup = 0;
+        if (sleep) {
+                if (p->se.last_wakeup) {
+                        update_avg(&p->se.avg_overlap,
+                                p->se.sum_exec_runtime - p->se.last_wakeup);
+                        p->se.last_wakeup = 0;
+                } else {
+                        update_avg(&p->se.avg_wakeup,
+                                sysctl_sched_wakeup_granularity);
+                }
         }
 
         sched_info_dequeued(p);
@@ -2355,6 +2406,22 @@ out_activate:
         activate_task(rq, p, 1);
         success = 1;
 
+        /*
+         * Only attribute actual wakeups done by this task.
+         */
+        if (!in_interrupt()) {
+                struct sched_entity *se = &current->se;
+                u64 sample = se->sum_exec_runtime;
+
+                if (se->last_wakeup)
+                        sample -= se->last_wakeup;
+                else
+                        sample -= se->start_runtime;
+                update_avg(&se->avg_wakeup, sample);
+
+                se->last_wakeup = se->sum_exec_runtime;
+        }
+
 out_running:
         trace_sched_wakeup(rq, p, success);
         check_preempt_curr(rq, p, sync);
@@ -2365,8 +2432,6 @@ out_running:
                 p->sched_class->task_wake_up(rq, p);
 #endif
 out:
-        current->se.last_wakeup = current->se.sum_exec_runtime;
-
         task_rq_unlock(rq, &flags);
 
         return success;
@@ -2396,6 +2461,8 @@ static void __sched_fork(struct task_struct *p)
         p->se.prev_sum_exec_runtime     = 0;
         p->se.last_wakeup               = 0;
         p->se.avg_overlap               = 0;
+        p->se.start_runtime             = 0;
+        p->se.avg_wakeup                = sysctl_sched_wakeup_granularity;
 
 #ifdef CONFIG_SCHEDSTATS
         p->se.wait_start                = 0;
@@ -2458,6 +2525,8 @@ void sched_fork(struct task_struct *p, int clone_flags)
         /* Want to start with kernel preemption disabled. */
         task_thread_info(p)->preempt_count = 1;
 #endif
+        plist_node_init(&p->pushable_tasks, MAX_PRIO);
+
         put_cpu();
 }
 
@@ -2598,6 +2667,12 @@ static void finish_task_switch(struct rq *rq, struct task_struct *prev)
 {
         struct mm_struct *mm = rq->prev_mm;
         long prev_state;
+#ifdef CONFIG_SMP
+        int post_schedule = 0;
+
+        if (current->sched_class->needs_post_schedule)
+                post_schedule = current->sched_class->needs_post_schedule(rq);
+#endif
 
         rq->prev_mm = NULL;
 
@@ -2616,7 +2691,7 @@ static void finish_task_switch(struct rq *rq, struct task_struct *prev)
         finish_arch_switch(prev);
         finish_lock_switch(rq, prev);
 #ifdef CONFIG_SMP
-        if (current->sched_class->post_schedule)
+        if (post_schedule)
                 current->sched_class->post_schedule(rq);
 #endif
 
@@ -2997,6 +3072,16 @@ next:
         pulled++;
         rem_load_move -= p->se.load.weight;
 
+#ifdef CONFIG_PREEMPT
+        /*
+         * NEWIDLE balancing is a source of latency, so preemptible kernels
+         * will stop after the first task is pulled to minimize the critical
+         * section.
+         */
+        if (idle == CPU_NEWLY_IDLE)
+                goto out;
+#endif
+
         /*
          * We only want to steal up to the prescribed amount of weighted load.
          */
@@ -3043,9 +3128,15 @@ static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest,
                                 sd, idle, all_pinned, &this_best_prio);
                 class = class->next;
 
+#ifdef CONFIG_PREEMPT
+                /*
+                 * NEWIDLE balancing is a source of latency, so preemptible
+                 * kernels will stop after the first task is pulled to minimize
+                 * the critical section.
+                 */
                 if (idle == CPU_NEWLY_IDLE && this_rq->nr_running)
                         break;
-
+#endif
         } while (class && max_load_move > total_load_moved);
 
         return total_load_moved > 0;
@@ -3890,19 +3981,24 @@ int select_nohz_load_balancer(int stop_tick)
         int cpu = smp_processor_id();
 
         if (stop_tick) {
-                cpumask_set_cpu(cpu, nohz.cpu_mask);
                 cpu_rq(cpu)->in_nohz_recently = 1;
 
-                /*
-                 * If we are going offline and still the leader, give up!
-                 */
-                if (!cpu_active(cpu) &&
-                    atomic_read(&nohz.load_balancer) == cpu) {
+                if (!cpu_active(cpu)) {
+                        if (atomic_read(&nohz.load_balancer) != cpu)
+                                return 0;
+
+                        /*
+                         * If we are going offline and still the leader,
+                         * give up!
+                         */
                         if (atomic_cmpxchg(&nohz.load_balancer, cpu, -1) != cpu)
                                 BUG();
+
                         return 0;
                 }
 
+                cpumask_set_cpu(cpu, nohz.cpu_mask);
+
                 /* time for ilb owner also to sleep */
                 if (cpumask_weight(nohz.cpu_mask) == num_online_cpus()) {
                         if (atomic_read(&nohz.load_balancer) == cpu)
@@ -8214,11 +8310,15 @@ static void init_rt_rq(struct rt_rq *rt_rq, struct rq *rq)
         __set_bit(MAX_RT_PRIO, array->bitmap);
 
 #if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED
-        rt_rq->highest_prio = MAX_RT_PRIO;
+        rt_rq->highest_prio.curr = MAX_RT_PRIO;
+#ifdef CONFIG_SMP
+        rt_rq->highest_prio.next = MAX_RT_PRIO;
+#endif
 #endif
 #ifdef CONFIG_SMP
         rt_rq->rt_nr_migratory = 0;
         rt_rq->overloaded = 0;
+        plist_head_init(&rq->rt.pushable_tasks, &rq->lock);
 #endif
 
         rt_rq->rt_time = 0;
diff --git a/kernel/sched_debug.c b/kernel/sched_debug.c
index 16eeba4e4169..2b1260f0e800 100644
--- a/kernel/sched_debug.c
+++ b/kernel/sched_debug.c
@@ -397,6 +397,7 @@ void proc_sched_show_task(struct task_struct *p, struct seq_file *m)
         PN(se.vruntime);
         PN(se.sum_exec_runtime);
         PN(se.avg_overlap);
+        PN(se.avg_wakeup);
 
         nr_switches = p->nvcsw + p->nivcsw;
 
diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c
index a7e50ba185ac..bc1563e7a248 100644
--- a/kernel/sched_fair.c
+++ b/kernel/sched_fair.c
@@ -1309,16 +1309,63 @@ out:
 }
 #endif /* CONFIG_SMP */
 
-static unsigned long wakeup_gran(struct sched_entity *se)
+/*
+ * Adaptive granularity
+ *
+ * se->avg_wakeup gives the average time a task runs until it does a wakeup,
+ * with the limit of wakeup_gran -- when it never does a wakeup.
+ *
+ * So the smaller avg_wakeup is the faster we want this task to preempt,
+ * but we don't want to treat the preemptee unfairly and therefore allow it
+ * to run for at least the amount of time we'd like to run.
+ *
+ * NOTE: we use 2*avg_wakeup to increase the probability of actually doing one
+ *
+ * NOTE: we use *nr_running to scale with load, this nicely matches the
+ *       degrading latency on load.
+ */
+static unsigned long
+adaptive_gran(struct sched_entity *curr, struct sched_entity *se)
+{
+        u64 this_run = curr->sum_exec_runtime - curr->prev_sum_exec_runtime;
+        u64 expected_wakeup = 2*se->avg_wakeup * cfs_rq_of(se)->nr_running;
+        u64 gran = 0;
+
+        if (this_run < expected_wakeup)
+                gran = expected_wakeup - this_run;
+
+        return min_t(s64, gran, sysctl_sched_wakeup_granularity);
+}
+
+static unsigned long
+wakeup_gran(struct sched_entity *curr, struct sched_entity *se)
 {
         unsigned long gran = sysctl_sched_wakeup_granularity;
 
+        if (cfs_rq_of(curr)->curr && sched_feat(ADAPTIVE_GRAN))
+                gran = adaptive_gran(curr, se);
+
         /*
-         * More easily preempt - nice tasks, while not making it harder for
-         * + nice tasks.
+         * Since its curr running now, convert the gran from real-time
+         * to virtual-time in his units.
          */
-        if (!sched_feat(ASYM_GRAN) || se->load.weight > NICE_0_LOAD)
-                gran = calc_delta_fair(sysctl_sched_wakeup_granularity, se);
+        if (sched_feat(ASYM_GRAN)) {
+                /*
+                 * By using 'se' instead of 'curr' we penalize light tasks, so
+                 * they get preempted easier. That is, if 'se' < 'curr' then
+                 * the resulting gran will be larger, therefore penalizing the
+                 * lighter, if otoh 'se' > 'curr' then the resulting gran will
+                 * be smaller, again penalizing the lighter task.
+                 *
+                 * This is especially important for buddies when the leftmost
+                 * task is higher priority than the buddy.
+                 */
+                if (unlikely(se->load.weight != NICE_0_LOAD))
+                        gran = calc_delta_fair(gran, se);
+        } else {
+                if (unlikely(curr->load.weight != NICE_0_LOAD))
+                        gran = calc_delta_fair(gran, curr);
+        }
 
         return gran;
 }
@@ -1345,7 +1392,7 @@ wakeup_preempt_entity(struct sched_entity *curr, struct sched_entity *se)
         if (vdiff <= 0)
                 return -1;
 
-        gran = wakeup_gran(curr);
+        gran = wakeup_gran(curr, se);
         if (vdiff > gran)
                 return 1;
 
diff --git a/kernel/sched_features.h b/kernel/sched_features.h
index da5d93b5d2c6..76f61756e677 100644
--- a/kernel/sched_features.h
+++ b/kernel/sched_features.h
@@ -1,5 +1,6 @@
 SCHED_FEAT(NEW_FAIR_SLEEPERS, 1)
-SCHED_FEAT(NORMALIZED_SLEEPER, 1)
+SCHED_FEAT(NORMALIZED_SLEEPER, 0)
+SCHED_FEAT(ADAPTIVE_GRAN, 1)
 SCHED_FEAT(WAKEUP_PREEMPT, 1)
 SCHED_FEAT(START_DEBIT, 1)
 SCHED_FEAT(AFFINE_WAKEUPS, 1)
diff --git a/kernel/sched_rt.c b/kernel/sched_rt.c
index bac1061cea2f..c79dc7844012 100644
--- a/kernel/sched_rt.c
+++ b/kernel/sched_rt.c
@@ -3,6 +3,40 @@
  * policies)
  */
 
+static inline struct task_struct *rt_task_of(struct sched_rt_entity *rt_se)
+{
+        return container_of(rt_se, struct task_struct, rt);
+}
+
+#ifdef CONFIG_RT_GROUP_SCHED
+
+static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq)
+{
+        return rt_rq->rq;
+}
+
+static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se)
+{
+        return rt_se->rt_rq;
+}
+
+#else /* CONFIG_RT_GROUP_SCHED */
+
+static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq)
+{
+        return container_of(rt_rq, struct rq, rt);
+}
+
+static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se)
+{
+        struct task_struct *p = rt_task_of(rt_se);
+        struct rq *rq = task_rq(p);
+
+        return &rq->rt;
+}
+
+#endif /* CONFIG_RT_GROUP_SCHED */
+
 #ifdef CONFIG_SMP
 
 static inline int rt_overloaded(struct rq *rq)
@@ -37,25 +71,69 @@ static inline void rt_clear_overload(struct rq *rq)
         cpumask_clear_cpu(rq->cpu, rq->rd->rto_mask);
 }
 
-static void update_rt_migration(struct rq *rq)
+static void update_rt_migration(struct rt_rq *rt_rq)
 {
-        if (rq->rt.rt_nr_migratory && (rq->rt.rt_nr_running > 1)) {
-                if (!rq->rt.overloaded) {
-                        rt_set_overload(rq);
-                        rq->rt.overloaded = 1;
+        if (rt_rq->rt_nr_migratory && (rt_rq->rt_nr_running > 1)) {
+                if (!rt_rq->overloaded) {
+                        rt_set_overload(rq_of_rt_rq(rt_rq));
+                        rt_rq->overloaded = 1;
                 }
-        } else if (rq->rt.overloaded) {
-                rt_clear_overload(rq);
-                rq->rt.overloaded = 0;
+        } else if (rt_rq->overloaded) {
+                rt_clear_overload(rq_of_rt_rq(rt_rq));
+                rt_rq->overloaded = 0;
         }
 }
-#endif /* CONFIG_SMP */
 
-static inline struct task_struct *rt_task_of(struct sched_rt_entity *rt_se)
+static void inc_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
+{
+        if (rt_se->nr_cpus_allowed > 1)
+                rt_rq->rt_nr_migratory++;
+
+        update_rt_migration(rt_rq);
+}
+
+static void dec_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
+{
+        if (rt_se->nr_cpus_allowed > 1)
+                rt_rq->rt_nr_migratory--;
+
+        update_rt_migration(rt_rq);
+}
+
+static void enqueue_pushable_task(struct rq *rq, struct task_struct *p)
+{
+        plist_del(&p->pushable_tasks, &rq->rt.pushable_tasks);
+        plist_node_init(&p->pushable_tasks, p->prio);
+        plist_add(&p->pushable_tasks, &rq->rt.pushable_tasks);
+}
+
+static void dequeue_pushable_task(struct rq *rq, struct task_struct *p)
+{
+        plist_del(&p->pushable_tasks, &rq->rt.pushable_tasks);
+}
+
+#else
+
+static inline void enqueue_pushable_task(struct rq *rq, struct task_struct *p)
 {
-        return container_of(rt_se, struct task_struct, rt);
 }
 
+static inline void dequeue_pushable_task(struct rq *rq, struct task_struct *p)
+{
+}
+
+static inline
+void inc_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
+{
+}
+
+static inline
+void dec_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
+{
+}
+
+#endif /* CONFIG_SMP */
+
 static inline int on_rt_rq(struct sched_rt_entity *rt_se)
 {
         return !list_empty(&rt_se->run_list);
@@ -79,16 +157,6 @@ static inline u64 sched_rt_period(struct rt_rq *rt_rq)
 #define for_each_leaf_rt_rq(rt_rq, rq) \
         list_for_each_entry_rcu(rt_rq, &rq->leaf_rt_rq_list, leaf_rt_rq_list)
 
-static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq)
-{
-        return rt_rq->rq;
-}
-
-static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se)
-{
-        return rt_se->rt_rq;
-}
-
 #define for_each_sched_rt_entity(rt_se) \
         for (; rt_se; rt_se = rt_se->parent)
 
@@ -108,7 +176,7 @@ static void sched_rt_rq_enqueue(struct rt_rq *rt_rq)
         if (rt_rq->rt_nr_running) {
                 if (rt_se && !on_rt_rq(rt_se))
                         enqueue_rt_entity(rt_se);
-                if (rt_rq->highest_prio < curr->prio)
+                if (rt_rq->highest_prio.curr < curr->prio)
                         resched_task(curr);
         }
 }
@@ -176,19 +244,6 @@ static inline u64 sched_rt_period(struct rt_rq *rt_rq)
 #define for_each_leaf_rt_rq(rt_rq, rq) \
         for (rt_rq = &rq->rt; rt_rq; rt_rq = NULL)
 
-static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq)
-{
-        return container_of(rt_rq, struct rq, rt);
-}
-
-static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se)
-{
-        struct task_struct *p = rt_task_of(rt_se);
-        struct rq *rq = task_rq(p);
-
-        return &rq->rt;
-}
-
 #define for_each_sched_rt_entity(rt_se) \
         for (; rt_se; rt_se = NULL)
 
@@ -473,7 +528,7 @@ static inline int rt_se_prio(struct sched_rt_entity *rt_se)
         struct rt_rq *rt_rq = group_rt_rq(rt_se);
 
         if (rt_rq)
-                return rt_rq->highest_prio;
+                return rt_rq->highest_prio.curr;
 #endif
 
         return rt_task_of(rt_se)->prio;
@@ -547,91 +602,174 @@ static void update_curr_rt(struct rq *rq)
         }
 }
 
-static inline
-void inc_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
+#if defined CONFIG_SMP
+
+static struct task_struct *pick_next_highest_task_rt(struct rq *rq, int cpu);
+
+static inline int next_prio(struct rq *rq)
 {
-        WARN_ON(!rt_prio(rt_se_prio(rt_se)));
-        rt_rq->rt_nr_running++;
-#if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED
-        if (rt_se_prio(rt_se) < rt_rq->highest_prio) {
-#ifdef CONFIG_SMP
-                struct rq *rq = rq_of_rt_rq(rt_rq);
-#endif
+        struct task_struct *next = pick_next_highest_task_rt(rq, rq->cpu);
+
+        if (next && rt_prio(next->prio))
+                return next->prio;
+        else
+                return MAX_RT_PRIO;
+}
+
+static void
+inc_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio)
+{
+        struct rq *rq = rq_of_rt_rq(rt_rq);
+
+        if (prio < prev_prio) {
+
+                /*
+                 * If the new task is higher in priority than anything on the
+                 * run-queue, we know that the previous high becomes our
+                 * next-highest.
+                 */
+                rt_rq->highest_prio.next = prev_prio;
 
-                rt_rq->highest_prio = rt_se_prio(rt_se);
-#ifdef CONFIG_SMP
                 if (rq->online)
-                        cpupri_set(&rq->rd->cpupri, rq->cpu,
-                                   rt_se_prio(rt_se));
-#endif
-        }
-#endif
-#ifdef CONFIG_SMP
-        if (rt_se->nr_cpus_allowed > 1) {
-                struct rq *rq = rq_of_rt_rq(rt_rq);
+                        cpupri_set(&rq->rd->cpupri, rq->cpu, prio);
 
-                rq->rt.rt_nr_migratory++;
-        }
+        } else if (prio == rt_rq->highest_prio.curr)
+                /*
+                 * If the next task is equal in priority to the highest on
+                 * the run-queue, then we implicitly know that the next highest
+                 * task cannot be any lower than current
+                 */
+                rt_rq->highest_prio.next = prio;
+        else if (prio < rt_rq->highest_prio.next)
+                /*
+                 * Otherwise, we need to recompute next-highest
+                 */
+                rt_rq->highest_prio.next = next_prio(rq);
+}
 
-        update_rt_migration(rq_of_rt_rq(rt_rq));
-#endif
-#ifdef CONFIG_RT_GROUP_SCHED
-        if (rt_se_boosted(rt_se))
-                rt_rq->rt_nr_boosted++;
+static void
+dec_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio)
+{
+        struct rq *rq = rq_of_rt_rq(rt_rq);
 
-        if (rt_rq->tg)
-                start_rt_bandwidth(&rt_rq->tg->rt_bandwidth);
-#else
-        start_rt_bandwidth(&def_rt_bandwidth);
-#endif
+        if (rt_rq->rt_nr_running && (prio <= rt_rq->highest_prio.next))
+                rt_rq->highest_prio.next = next_prio(rq);
+
+        if (rq->online && rt_rq->highest_prio.curr != prev_prio)
+                cpupri_set(&rq->rd->cpupri, rq->cpu, rt_rq->highest_prio.curr);
 }
 
+#else /* CONFIG_SMP */
+
 static inline
-void dec_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
-{
-#ifdef CONFIG_SMP
-        int highest_prio = rt_rq->highest_prio;
-#endif
+void inc_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio) {}
+static inline
+void dec_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio) {}
+
+#endif /* CONFIG_SMP */
 
-        WARN_ON(!rt_prio(rt_se_prio(rt_se)));
-        WARN_ON(!rt_rq->rt_nr_running);
-        rt_rq->rt_nr_running--;
 #if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED
+static void
+inc_rt_prio(struct rt_rq *rt_rq, int prio)
+{
+        int prev_prio = rt_rq->highest_prio.curr;
+
+        if (prio < prev_prio)
+                rt_rq->highest_prio.curr = prio;
+
+        inc_rt_prio_smp(rt_rq, prio, prev_prio);
+}
+
+static void
+dec_rt_prio(struct rt_rq *rt_rq, int prio)
+{
+        int prev_prio = rt_rq->highest_prio.curr;
+
         if (rt_rq->rt_nr_running) {
-                struct rt_prio_array *array;
 
-                WARN_ON(rt_se_prio(rt_se) < rt_rq->highest_prio);
-                if (rt_se_prio(rt_se) == rt_rq->highest_prio) {
-                        /* recalculate */
-                        array = &rt_rq->active;
-                        rt_rq->highest_prio =
+                WARN_ON(prio < prev_prio);
+
+                /*
+                 * This may have been our highest task, and therefore
+                 * we may have some recomputation to do
+                 */
+                if (prio == prev_prio) {
+                        struct rt_prio_array *array = &rt_rq->active;
+
+                        rt_rq->highest_prio.curr =
                                 sched_find_first_bit(array->bitmap);
-                } /* otherwise leave rq->highest prio alone */
+                }
+
         } else
-                rt_rq->highest_prio = MAX_RT_PRIO;
-#endif
-#ifdef CONFIG_SMP
-        if (rt_se->nr_cpus_allowed > 1) {
-                struct rq *rq = rq_of_rt_rq(rt_rq);
-                rq->rt.rt_nr_migratory--;
-        }
+                rt_rq->highest_prio.curr = MAX_RT_PRIO;
 
-        if (rt_rq->highest_prio != highest_prio) {
-                struct rq *rq = rq_of_rt_rq(rt_rq);
+        dec_rt_prio_smp(rt_rq, prio, prev_prio);
+}
 
-                if (rq->online)
-                        cpupri_set(&rq->rd->cpupri, rq->cpu,
-                                   rt_rq->highest_prio);
-        }
+#else
+
+static inline void inc_rt_prio(struct rt_rq *rt_rq, int prio) {}
+static inline void dec_rt_prio(struct rt_rq *rt_rq, int prio) {}
+
+#endif /* CONFIG_SMP || CONFIG_RT_GROUP_SCHED */
 
-        update_rt_migration(rq_of_rt_rq(rt_rq));
-#endif /* CONFIG_SMP */
 #ifdef CONFIG_RT_GROUP_SCHED
+
+static void
+inc_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
+{
+        if (rt_se_boosted(rt_se))
+                rt_rq->rt_nr_boosted++;
+
+        if (rt_rq->tg)
+                start_rt_bandwidth(&rt_rq->tg->rt_bandwidth);
+}
+
+static void
+dec_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
+{
         if (rt_se_boosted(rt_se))
                 rt_rq->rt_nr_boosted--;
 
         WARN_ON(!rt_rq->rt_nr_running && rt_rq->rt_nr_boosted);
-#endif
+}
+
+#else /* CONFIG_RT_GROUP_SCHED */
+
+static void
+inc_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
+{
+        start_rt_bandwidth(&def_rt_bandwidth);
+}
+
+static inline
+void dec_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) {}
+
+#endif /* CONFIG_RT_GROUP_SCHED */
+
+static inline
+void inc_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
+{
+        int prio = rt_se_prio(rt_se);
+
+        WARN_ON(!rt_prio(prio));
+        rt_rq->rt_nr_running++;
+
+        inc_rt_prio(rt_rq, prio);
+        inc_rt_migration(rt_se, rt_rq);
+        inc_rt_group(rt_se, rt_rq);
+}
+
+static inline
+void dec_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
+{
+        WARN_ON(!rt_prio(rt_se_prio(rt_se)));
+        WARN_ON(!rt_rq->rt_nr_running);
+        rt_rq->rt_nr_running--;
+
+        dec_rt_prio(rt_rq, rt_se_prio(rt_se));
+        dec_rt_migration(rt_se, rt_rq);
+        dec_rt_group(rt_se, rt_rq);
 }
 
 static void __enqueue_rt_entity(struct sched_rt_entity *rt_se)
@@ -718,6 +856,9 @@ static void enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup)
 
         enqueue_rt_entity(rt_se);
 
+        if (!task_current(rq, p) && p->rt.nr_cpus_allowed > 1)
+                enqueue_pushable_task(rq, p);
+
         inc_cpu_load(rq, p->se.load.weight);
 }
 
@@ -728,6 +869,8 @@ static void dequeue_task_rt(struct rq *rq, struct task_struct *p, int sleep)
         update_curr_rt(rq);
         dequeue_rt_entity(rt_se);
 
+        dequeue_pushable_task(rq, p);
+
         dec_cpu_load(rq, p->se.load.weight);
 }
 
@@ -878,7 +1021,7 @@ static struct sched_rt_entity *pick_next_rt_entity(struct rq *rq,
         return next;
 }
 
-static struct task_struct *pick_next_task_rt(struct rq *rq)
+static struct task_struct *_pick_next_task_rt(struct rq *rq)
 {
         struct sched_rt_entity *rt_se;
         struct task_struct *p;
@@ -900,6 +1043,18 @@ static struct task_struct *pick_next_task_rt(struct rq *rq)
 
         p = rt_task_of(rt_se);
         p->se.exec_start = rq->clock;
+
+        return p;
+}
+
+static struct task_struct *pick_next_task_rt(struct rq *rq)
+{
+        struct task_struct *p = _pick_next_task_rt(rq);
+
+        /* The running task is never eligible for pushing */
+        if (p)
+                dequeue_pushable_task(rq, p);
+
         return p;
 }
 
@@ -907,6 +1062,13 @@ static void put_prev_task_rt(struct rq *rq, struct task_struct *p)
 {
         update_curr_rt(rq);
         p->se.exec_start = 0;
+
+        /*
+         * The previous task needs to be made eligible for pushing
+         * if it is still active
+         */
+        if (p->se.on_rq && p->rt.nr_cpus_allowed > 1)
+                enqueue_pushable_task(rq, p);
 }
 
 #ifdef CONFIG_SMP
@@ -1072,7 +1234,7 @@ static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq)
                 }
 
                 /* If this rq is still suitable use it. */
-                if (lowest_rq->rt.highest_prio > task->prio)
+                if (lowest_rq->rt.highest_prio.curr > task->prio)
                         break;
 
                 /* try again */
@@ -1083,6 +1245,31 @@ static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq)
         return lowest_rq;
 }
 
+static inline int has_pushable_tasks(struct rq *rq)
+{
+        return !plist_head_empty(&rq->rt.pushable_tasks);
+}
+
+static struct task_struct *pick_next_pushable_task(struct rq *rq)
+{
+        struct task_struct *p;
+
+        if (!has_pushable_tasks(rq))
+                return NULL;
+
+        p = plist_first_entry(&rq->rt.pushable_tasks,
+                              struct task_struct, pushable_tasks);
+
+        BUG_ON(rq->cpu != task_cpu(p));
+        BUG_ON(task_current(rq, p));
+        BUG_ON(p->rt.nr_cpus_allowed <= 1);
+
+        BUG_ON(!p->se.on_rq);
+        BUG_ON(!rt_task(p));
+
+        return p;
+}
+
 /*
  * If the current CPU has more than one RT task, see if the non
  * running task can migrate over to a CPU that is running a task
@@ -1092,13 +1279,11 @@ static int push_rt_task(struct rq *rq)
 {
         struct task_struct *next_task;
         struct rq *lowest_rq;
-        int ret = 0;
-        int paranoid = RT_MAX_TRIES;
 
         if (!rq->rt.overloaded)
                 return 0;
 
-        next_task = pick_next_highest_task_rt(rq, -1);
+        next_task = pick_next_pushable_task(rq);
         if (!next_task)
                 return 0;
 
@@ -1127,16 +1312,34 @@ static int push_rt_task(struct rq *rq)
                 struct task_struct *task;
                 /*
                  * find lock_lowest_rq releases rq->lock
-                 * so it is possible that next_task has changed.
-                 * If it has, then try again.
+                 * so it is possible that next_task has migrated.
+                 *
+                 * We need to make sure that the task is still on the same
+                 * run-queue and is also still the next task eligible for
+                 * pushing.
                  */
-                task = pick_next_highest_task_rt(rq, -1);
-                if (unlikely(task != next_task) && task && paranoid--) {
-                        put_task_struct(next_task);
-                        next_task = task;
-                        goto retry;
+                task = pick_next_pushable_task(rq);
+                if (task_cpu(next_task) == rq->cpu && task == next_task) {
+                        /*
+                         * If we get here, the task hasnt moved at all, but
+                         * it has failed to push.  We will not try again,
+                         * since the other cpus will pull from us when they
+                         * are ready.
+                         */
+                        dequeue_pushable_task(rq, next_task);
+                        goto out;
                 }
-                goto out;
+
+                if (!task)
+                        /* No more tasks, just exit */
+                        goto out;
+
+                /*
+                 * Something has shifted, try again.
+                 */
+                put_task_struct(next_task);
+                next_task = task;
+                goto retry;
         }
 
         deactivate_task(rq, next_task, 0);
@@ -1147,23 +1350,12 @@ static int push_rt_task(struct rq *rq)
 
         double_unlock_balance(rq, lowest_rq);
 
-        ret = 1;
 out:
         put_task_struct(next_task);
 
-        return ret;
+        return 1;
 }
 
-/*
- * TODO: Currently we just use the second highest prio task on
- *       the queue, and stop when it can't migrate (or there's
- *       no more RT tasks).  There may be a case where a lower
- *       priority RT task has a different affinity than the
- *       higher RT task. In this case the lower RT task could
- *       possibly be able to migrate where as the higher priority
- *       RT task could not.  We currently ignore this issue.
- *       Enhancements are welcome!
- */
 static void push_rt_tasks(struct rq *rq)
 {
         /* push_rt_task will return true if it moved an RT */
@@ -1174,33 +1366,35 @@ static void push_rt_tasks(struct rq *rq)
 static int pull_rt_task(struct rq *this_rq)
 {
         int this_cpu = this_rq->cpu, ret = 0, cpu;
-        struct task_struct *p, *next;
+        struct task_struct *p;
         struct rq *src_rq;
 
         if (likely(!rt_overloaded(this_rq)))
                 return 0;
 
-        next = pick_next_task_rt(this_rq);
-
         for_each_cpu(cpu, this_rq->rd->rto_mask) {
                 if (this_cpu == cpu)
                         continue;
 
                 src_rq = cpu_rq(cpu);
+
+                /*
+                 * Don't bother taking the src_rq->lock if the next highest
+                 * task is known to be lower-priority than our current task.
+                 * This may look racy, but if this value is about to go
+                 * logically higher, the src_rq will push this task away.
+                 * And if its going logically lower, we do not care
+                 */
+                if (src_rq->rt.highest_prio.next >=
+                    this_rq->rt.highest_prio.curr)
+                        continue;
+
                 /*
                  * We can potentially drop this_rq's lock in
                  * double_lock_balance, and another CPU could
-                 * steal our next task - hence we must cause
-                 * the caller to recalculate the next task
-                 * in that case:
+                 * alter this_rq
                  */
-                if (double_lock_balance(this_rq, src_rq)) {
-                        struct task_struct *old_next = next;
-
-                        next = pick_next_task_rt(this_rq);
-                        if (next != old_next)
-                                ret = 1;
-                }
+                double_lock_balance(this_rq, src_rq);
 
                 /*
                  * Are there still pullable RT tasks?
@@ -1214,7 +1408,7 @@ static int pull_rt_task(struct rq *this_rq)
                  * Do we have an RT task that preempts
                  * the to-be-scheduled task?
                  */
-                if (p && (!next || (p->prio < next->prio))) {
+                if (p && (p->prio < this_rq->rt.highest_prio.curr)) {
                         WARN_ON(p == src_rq->curr);
                         WARN_ON(!p->se.on_rq);
 
@@ -1224,12 +1418,9 @@ static int pull_rt_task(struct rq *this_rq)
                          * This is just that p is wakeing up and hasn't
                          * had a chance to schedule. We only pull
                          * p if it is lower in priority than the
-                         * current task on the run queue or
-                         * this_rq next task is lower in prio than
-                         * the current task on that rq.
+                         * current task on the run queue
                          */
-                        if (p->prio < src_rq->curr->prio ||
-                            (next && next->prio < src_rq->curr->prio))
+                        if (p->prio < src_rq->curr->prio)
                                 goto skip;
 
                         ret = 1;
@@ -1242,13 +1433,7 @@ static int pull_rt_task(struct rq *this_rq)
                          * case there's an even higher prio task
                          * in another runqueue. (low likelyhood
                          * but possible)
-                         *
-                         * Update next so that we won't pick a task
-                         * on another cpu with a priority lower (or equal)
-                         * than the one we just picked.
                          */
-                        next = p;
-
                 }
  skip:
                 double_unlock_balance(this_rq, src_rq);
@@ -1260,24 +1445,27 @@ static int pull_rt_task(struct rq *this_rq)
 static void pre_schedule_rt(struct rq *rq, struct task_struct *prev)
 {
         /* Try to pull RT tasks here if we lower this rq's prio */
-        if (unlikely(rt_task(prev)) && rq->rt.highest_prio > prev->prio)
+        if (unlikely(rt_task(prev)) && rq->rt.highest_prio.curr > prev->prio)
                 pull_rt_task(rq);
 }
 
+/*
+ * assumes rq->lock is held
+ */
+static int needs_post_schedule_rt(struct rq *rq)
+{
+        return has_pushable_tasks(rq);
+}
+
 static void post_schedule_rt(struct rq *rq)
 {
         /*
-         * If we have more than one rt_task queued, then
-         * see if we can push the other rt_tasks off to other CPUS.
-         * Note we may release the rq lock, and since
-         * the lock was owned by prev, we need to release it
-         * first via finish_lock_switch and then reaquire it here.
+         * This is only called if needs_post_schedule_rt() indicates that
+         * we need to push tasks away
          */
-        if (unlikely(rq->rt.overloaded)) {
-                spin_lock_irq(&rq->lock);
-                push_rt_tasks(rq);
-                spin_unlock_irq(&rq->lock);
-        }
+        spin_lock_irq(&rq->lock);
+        push_rt_tasks(rq);
+        spin_unlock_irq(&rq->lock);
 }
 
 /*
@@ -1288,7 +1476,8 @@ static void task_wake_up_rt(struct rq *rq, struct task_struct *p)
 {
         if (!task_running(rq, p) &&
             !test_tsk_need_resched(rq->curr) &&
-            rq->rt.overloaded)
+            has_pushable_tasks(rq) &&
+            p->rt.nr_cpus_allowed > 1)
                 push_rt_tasks(rq);
 }
 
@@ -1324,6 +1513,24 @@ static void set_cpus_allowed_rt(struct task_struct *p,
         if (p->se.on_rq && (weight != p->rt.nr_cpus_allowed)) {
                 struct rq *rq = task_rq(p);
 
+                if (!task_current(rq, p)) {
+                        /*
+                         * Make sure we dequeue this task from the pushable list
+                         * before going further.  It will either remain off of
+                         * the list because we are no longer pushable, or it
+                         * will be requeued.
+                         */
+                        if (p->rt.nr_cpus_allowed > 1)
+                                dequeue_pushable_task(rq, p);
+
+                        /*
+                         * Requeue if our weight is changing and still > 1
+                         */
+                        if (weight > 1)
+                                enqueue_pushable_task(rq, p);
+
+                }
+
                 if ((p->rt.nr_cpus_allowed <= 1) && (weight > 1)) {
                         rq->rt.rt_nr_migratory++;
                 } else if ((p->rt.nr_cpus_allowed > 1) && (weight <= 1)) {
@@ -1331,7 +1538,7 @@ static void set_cpus_allowed_rt(struct task_struct *p,
                         rq->rt.rt_nr_migratory--;
                 }
 
-                update_rt_migration(rq);
+                update_rt_migration(&rq->rt);
         }
 
         cpumask_copy(&p->cpus_allowed, new_mask);
@@ -1346,7 +1553,7 @@ static void rq_online_rt(struct rq *rq)
 
         __enable_runtime(rq);
 
-        cpupri_set(&rq->rd->cpupri, rq->cpu, rq->rt.highest_prio);
+        cpupri_set(&rq->rd->cpupri, rq->cpu, rq->rt.highest_prio.curr);
 }
 
 /* Assumes rq->lock is held */
@@ -1438,7 +1645,7 @@ static void prio_changed_rt(struct rq *rq, struct task_struct *p,
                  * can release the rq lock and p could migrate.
                  * Only reschedule if p is still on the same runqueue.
                  */
-                if (p->prio > rq->rt.highest_prio && rq->curr == p)
+                if (p->prio > rq->rt.highest_prio.curr && rq->curr == p)
                         resched_task(p);
 #else
                 /* For UP simply resched on drop of prio */
@@ -1509,6 +1716,9 @@ static void set_curr_task_rt(struct rq *rq)
         struct task_struct *p = rq->curr;
 
         p->se.exec_start = rq->clock;
+
+        /* The running task is never eligible for pushing */
+        dequeue_pushable_task(rq, p);
 }
 
 static const struct sched_class rt_sched_class = {
@@ -1531,6 +1741,7 @@ static const struct sched_class rt_sched_class = {
         .rq_online              = rq_online_rt,
         .rq_offline             = rq_offline_rt,
         .pre_schedule           = pre_schedule_rt,
+        .needs_post_schedule    = needs_post_schedule_rt,
         .post_schedule          = post_schedule_rt,
         .task_wake_up           = task_wake_up_rt,
         .switched_from          = switched_from_rt,