2 * Real-Time Scheduling Class (mapped to the SCHED_FIFO and SCHED_RR
8 #include <linux/slab.h>
10 static int do_sched_rt_period_timer(struct rt_bandwidth
*rt_b
, int overrun
);
12 struct rt_bandwidth def_rt_bandwidth
;
14 static enum hrtimer_restart
sched_rt_period_timer(struct hrtimer
*timer
)
16 struct rt_bandwidth
*rt_b
=
17 container_of(timer
, struct rt_bandwidth
, rt_period_timer
);
23 now
= hrtimer_cb_get_time(timer
);
24 overrun
= hrtimer_forward(timer
, now
, rt_b
->rt_period
);
29 idle
= do_sched_rt_period_timer(rt_b
, overrun
);
32 return idle
? HRTIMER_NORESTART
: HRTIMER_RESTART
;
35 void init_rt_bandwidth(struct rt_bandwidth
*rt_b
, u64 period
, u64 runtime
)
37 rt_b
->rt_period
= ns_to_ktime(period
);
38 rt_b
->rt_runtime
= runtime
;
40 raw_spin_lock_init(&rt_b
->rt_runtime_lock
);
42 hrtimer_init(&rt_b
->rt_period_timer
,
43 CLOCK_MONOTONIC
, HRTIMER_MODE_REL
);
44 rt_b
->rt_period_timer
.function
= sched_rt_period_timer
;
47 static void start_rt_bandwidth(struct rt_bandwidth
*rt_b
)
49 if (!rt_bandwidth_enabled() || rt_b
->rt_runtime
== RUNTIME_INF
)
52 if (hrtimer_active(&rt_b
->rt_period_timer
))
55 raw_spin_lock(&rt_b
->rt_runtime_lock
);
56 start_bandwidth_timer(&rt_b
->rt_period_timer
, rt_b
->rt_period
);
57 raw_spin_unlock(&rt_b
->rt_runtime_lock
);
60 void init_rt_rq(struct rt_rq
*rt_rq
, struct rq
*rq
)
62 struct rt_prio_array
*array
;
65 array
= &rt_rq
->active
;
66 for (i
= 0; i
< MAX_RT_PRIO
; i
++) {
67 INIT_LIST_HEAD(array
->queue
+ i
);
68 __clear_bit(i
, array
->bitmap
);
70 /* delimiter for bitsearch: */
71 __set_bit(MAX_RT_PRIO
, array
->bitmap
);
73 #if defined CONFIG_SMP
74 rt_rq
->highest_prio
.curr
= MAX_RT_PRIO
;
75 rt_rq
->highest_prio
.next
= MAX_RT_PRIO
;
76 rt_rq
->rt_nr_migratory
= 0;
77 rt_rq
->overloaded
= 0;
78 plist_head_init(&rt_rq
->pushable_tasks
);
82 rt_rq
->rt_throttled
= 0;
83 rt_rq
->rt_runtime
= 0;
84 raw_spin_lock_init(&rt_rq
->rt_runtime_lock
);
87 #ifdef CONFIG_RT_GROUP_SCHED
88 static void destroy_rt_bandwidth(struct rt_bandwidth
*rt_b
)
90 hrtimer_cancel(&rt_b
->rt_period_timer
);
93 #define rt_entity_is_task(rt_se) (!(rt_se)->my_q)
95 static inline struct task_struct
*rt_task_of(struct sched_rt_entity
*rt_se
)
97 #ifdef CONFIG_SCHED_DEBUG
98 WARN_ON_ONCE(!rt_entity_is_task(rt_se
));
100 return container_of(rt_se
, struct task_struct
, rt
);
103 static inline struct rq
*rq_of_rt_rq(struct rt_rq
*rt_rq
)
108 static inline struct rt_rq
*rt_rq_of_se(struct sched_rt_entity
*rt_se
)
113 void free_rt_sched_group(struct task_group
*tg
)
118 destroy_rt_bandwidth(&tg
->rt_bandwidth
);
120 for_each_possible_cpu(i
) {
131 void init_tg_rt_entry(struct task_group
*tg
, struct rt_rq
*rt_rq
,
132 struct sched_rt_entity
*rt_se
, int cpu
,
133 struct sched_rt_entity
*parent
)
135 struct rq
*rq
= cpu_rq(cpu
);
137 rt_rq
->highest_prio
.curr
= MAX_RT_PRIO
;
138 rt_rq
->rt_nr_boosted
= 0;
142 tg
->rt_rq
[cpu
] = rt_rq
;
143 tg
->rt_se
[cpu
] = rt_se
;
149 rt_se
->rt_rq
= &rq
->rt
;
151 rt_se
->rt_rq
= parent
->my_q
;
154 rt_se
->parent
= parent
;
155 INIT_LIST_HEAD(&rt_se
->run_list
);
158 int alloc_rt_sched_group(struct task_group
*tg
, struct task_group
*parent
)
161 struct sched_rt_entity
*rt_se
;
164 tg
->rt_rq
= kzalloc(sizeof(rt_rq
) * nr_cpu_ids
, GFP_KERNEL
);
167 tg
->rt_se
= kzalloc(sizeof(rt_se
) * nr_cpu_ids
, GFP_KERNEL
);
171 init_rt_bandwidth(&tg
->rt_bandwidth
,
172 ktime_to_ns(def_rt_bandwidth
.rt_period
), 0);
174 for_each_possible_cpu(i
) {
175 rt_rq
= kzalloc_node(sizeof(struct rt_rq
),
176 GFP_KERNEL
, cpu_to_node(i
));
180 rt_se
= kzalloc_node(sizeof(struct sched_rt_entity
),
181 GFP_KERNEL
, cpu_to_node(i
));
185 init_rt_rq(rt_rq
, cpu_rq(i
));
186 rt_rq
->rt_runtime
= tg
->rt_bandwidth
.rt_runtime
;
187 init_tg_rt_entry(tg
, rt_rq
, rt_se
, i
, parent
->rt_se
[i
]);
198 #else /* CONFIG_RT_GROUP_SCHED */
200 #define rt_entity_is_task(rt_se) (1)
202 static inline struct task_struct
*rt_task_of(struct sched_rt_entity
*rt_se
)
204 return container_of(rt_se
, struct task_struct
, rt
);
207 static inline struct rq
*rq_of_rt_rq(struct rt_rq
*rt_rq
)
209 return container_of(rt_rq
, struct rq
, rt
);
212 static inline struct rt_rq
*rt_rq_of_se(struct sched_rt_entity
*rt_se
)
214 struct task_struct
*p
= rt_task_of(rt_se
);
215 struct rq
*rq
= task_rq(p
);
220 void free_rt_sched_group(struct task_group
*tg
) { }
222 int alloc_rt_sched_group(struct task_group
*tg
, struct task_group
*parent
)
226 #endif /* CONFIG_RT_GROUP_SCHED */
230 static inline int rt_overloaded(struct rq
*rq
)
232 return atomic_read(&rq
->rd
->rto_count
);
235 static inline void rt_set_overload(struct rq
*rq
)
240 cpumask_set_cpu(rq
->cpu
, rq
->rd
->rto_mask
);
242 * Make sure the mask is visible before we set
243 * the overload count. That is checked to determine
244 * if we should look at the mask. It would be a shame
245 * if we looked at the mask, but the mask was not
249 atomic_inc(&rq
->rd
->rto_count
);
252 static inline void rt_clear_overload(struct rq
*rq
)
257 /* the order here really doesn't matter */
258 atomic_dec(&rq
->rd
->rto_count
);
259 cpumask_clear_cpu(rq
->cpu
, rq
->rd
->rto_mask
);
262 static void update_rt_migration(struct rt_rq
*rt_rq
)
264 if (rt_rq
->rt_nr_migratory
&& rt_rq
->rt_nr_total
> 1) {
265 if (!rt_rq
->overloaded
) {
266 rt_set_overload(rq_of_rt_rq(rt_rq
));
267 rt_rq
->overloaded
= 1;
269 } else if (rt_rq
->overloaded
) {
270 rt_clear_overload(rq_of_rt_rq(rt_rq
));
271 rt_rq
->overloaded
= 0;
275 static void inc_rt_migration(struct sched_rt_entity
*rt_se
, struct rt_rq
*rt_rq
)
277 if (!rt_entity_is_task(rt_se
))
280 rt_rq
= &rq_of_rt_rq(rt_rq
)->rt
;
282 rt_rq
->rt_nr_total
++;
283 if (rt_se
->nr_cpus_allowed
> 1)
284 rt_rq
->rt_nr_migratory
++;
286 update_rt_migration(rt_rq
);
289 static void dec_rt_migration(struct sched_rt_entity
*rt_se
, struct rt_rq
*rt_rq
)
291 if (!rt_entity_is_task(rt_se
))
294 rt_rq
= &rq_of_rt_rq(rt_rq
)->rt
;
296 rt_rq
->rt_nr_total
--;
297 if (rt_se
->nr_cpus_allowed
> 1)
298 rt_rq
->rt_nr_migratory
--;
300 update_rt_migration(rt_rq
);
303 static inline int has_pushable_tasks(struct rq
*rq
)
305 return !plist_head_empty(&rq
->rt
.pushable_tasks
);
308 static void enqueue_pushable_task(struct rq
*rq
, struct task_struct
*p
)
310 plist_del(&p
->pushable_tasks
, &rq
->rt
.pushable_tasks
);
311 plist_node_init(&p
->pushable_tasks
, p
->prio
);
312 plist_add(&p
->pushable_tasks
, &rq
->rt
.pushable_tasks
);
314 /* Update the highest prio pushable task */
315 if (p
->prio
< rq
->rt
.highest_prio
.next
)
316 rq
->rt
.highest_prio
.next
= p
->prio
;
319 static void dequeue_pushable_task(struct rq
*rq
, struct task_struct
*p
)
321 plist_del(&p
->pushable_tasks
, &rq
->rt
.pushable_tasks
);
323 /* Update the new highest prio pushable task */
324 if (has_pushable_tasks(rq
)) {
325 p
= plist_first_entry(&rq
->rt
.pushable_tasks
,
326 struct task_struct
, pushable_tasks
);
327 rq
->rt
.highest_prio
.next
= p
->prio
;
329 rq
->rt
.highest_prio
.next
= MAX_RT_PRIO
;
334 static inline void enqueue_pushable_task(struct rq
*rq
, struct task_struct
*p
)
338 static inline void dequeue_pushable_task(struct rq
*rq
, struct task_struct
*p
)
343 void inc_rt_migration(struct sched_rt_entity
*rt_se
, struct rt_rq
*rt_rq
)
348 void dec_rt_migration(struct sched_rt_entity
*rt_se
, struct rt_rq
*rt_rq
)
352 #endif /* CONFIG_SMP */
354 static inline int on_rt_rq(struct sched_rt_entity
*rt_se
)
356 return !list_empty(&rt_se
->run_list
);
359 #ifdef CONFIG_RT_GROUP_SCHED
361 static inline u64
sched_rt_runtime(struct rt_rq
*rt_rq
)
366 return rt_rq
->rt_runtime
;
369 static inline u64
sched_rt_period(struct rt_rq
*rt_rq
)
371 return ktime_to_ns(rt_rq
->tg
->rt_bandwidth
.rt_period
);
374 typedef struct task_group
*rt_rq_iter_t
;
376 static inline struct task_group
*next_task_group(struct task_group
*tg
)
379 tg
= list_entry_rcu(tg
->list
.next
,
380 typeof(struct task_group
), list
);
381 } while (&tg
->list
!= &task_groups
&& task_group_is_autogroup(tg
));
383 if (&tg
->list
== &task_groups
)
389 #define for_each_rt_rq(rt_rq, iter, rq) \
390 for (iter = container_of(&task_groups, typeof(*iter), list); \
391 (iter = next_task_group(iter)) && \
392 (rt_rq = iter->rt_rq[cpu_of(rq)]);)
394 static inline void list_add_leaf_rt_rq(struct rt_rq
*rt_rq
)
396 list_add_rcu(&rt_rq
->leaf_rt_rq_list
,
397 &rq_of_rt_rq(rt_rq
)->leaf_rt_rq_list
);
400 static inline void list_del_leaf_rt_rq(struct rt_rq
*rt_rq
)
402 list_del_rcu(&rt_rq
->leaf_rt_rq_list
);
405 #define for_each_leaf_rt_rq(rt_rq, rq) \
406 list_for_each_entry_rcu(rt_rq, &rq->leaf_rt_rq_list, leaf_rt_rq_list)
408 #define for_each_sched_rt_entity(rt_se) \
409 for (; rt_se; rt_se = rt_se->parent)
411 static inline struct rt_rq
*group_rt_rq(struct sched_rt_entity
*rt_se
)
416 static void enqueue_rt_entity(struct sched_rt_entity
*rt_se
, bool head
);
417 static void dequeue_rt_entity(struct sched_rt_entity
*rt_se
);
419 static void sched_rt_rq_enqueue(struct rt_rq
*rt_rq
)
421 struct task_struct
*curr
= rq_of_rt_rq(rt_rq
)->curr
;
422 struct sched_rt_entity
*rt_se
;
424 int cpu
= cpu_of(rq_of_rt_rq(rt_rq
));
426 rt_se
= rt_rq
->tg
->rt_se
[cpu
];
428 if (rt_rq
->rt_nr_running
) {
429 if (rt_se
&& !on_rt_rq(rt_se
))
430 enqueue_rt_entity(rt_se
, false);
431 if (rt_rq
->highest_prio
.curr
< curr
->prio
)
436 static void sched_rt_rq_dequeue(struct rt_rq
*rt_rq
)
438 struct sched_rt_entity
*rt_se
;
439 int cpu
= cpu_of(rq_of_rt_rq(rt_rq
));
441 rt_se
= rt_rq
->tg
->rt_se
[cpu
];
443 if (rt_se
&& on_rt_rq(rt_se
))
444 dequeue_rt_entity(rt_se
);
447 static inline int rt_rq_throttled(struct rt_rq
*rt_rq
)
449 return rt_rq
->rt_throttled
&& !rt_rq
->rt_nr_boosted
;
452 static int rt_se_boosted(struct sched_rt_entity
*rt_se
)
454 struct rt_rq
*rt_rq
= group_rt_rq(rt_se
);
455 struct task_struct
*p
;
458 return !!rt_rq
->rt_nr_boosted
;
460 p
= rt_task_of(rt_se
);
461 return p
->prio
!= p
->normal_prio
;
465 static inline const struct cpumask
*sched_rt_period_mask(void)
467 return cpu_rq(smp_processor_id())->rd
->span
;
470 static inline const struct cpumask
*sched_rt_period_mask(void)
472 return cpu_online_mask
;
477 struct rt_rq
*sched_rt_period_rt_rq(struct rt_bandwidth
*rt_b
, int cpu
)
479 return container_of(rt_b
, struct task_group
, rt_bandwidth
)->rt_rq
[cpu
];
482 static inline struct rt_bandwidth
*sched_rt_bandwidth(struct rt_rq
*rt_rq
)
484 return &rt_rq
->tg
->rt_bandwidth
;
487 #else /* !CONFIG_RT_GROUP_SCHED */
489 static inline u64
sched_rt_runtime(struct rt_rq
*rt_rq
)
491 return rt_rq
->rt_runtime
;
494 static inline u64
sched_rt_period(struct rt_rq
*rt_rq
)
496 return ktime_to_ns(def_rt_bandwidth
.rt_period
);
499 typedef struct rt_rq
*rt_rq_iter_t
;
501 #define for_each_rt_rq(rt_rq, iter, rq) \
502 for ((void) iter, rt_rq = &rq->rt; rt_rq; rt_rq = NULL)
504 static inline void list_add_leaf_rt_rq(struct rt_rq
*rt_rq
)
508 static inline void list_del_leaf_rt_rq(struct rt_rq
*rt_rq
)
512 #define for_each_leaf_rt_rq(rt_rq, rq) \
513 for (rt_rq = &rq->rt; rt_rq; rt_rq = NULL)
515 #define for_each_sched_rt_entity(rt_se) \
516 for (; rt_se; rt_se = NULL)
518 static inline struct rt_rq
*group_rt_rq(struct sched_rt_entity
*rt_se
)
523 static inline void sched_rt_rq_enqueue(struct rt_rq
*rt_rq
)
525 if (rt_rq
->rt_nr_running
)
526 resched_task(rq_of_rt_rq(rt_rq
)->curr
);
529 static inline void sched_rt_rq_dequeue(struct rt_rq
*rt_rq
)
533 static inline int rt_rq_throttled(struct rt_rq
*rt_rq
)
535 return rt_rq
->rt_throttled
;
538 static inline const struct cpumask
*sched_rt_period_mask(void)
540 return cpu_online_mask
;
544 struct rt_rq
*sched_rt_period_rt_rq(struct rt_bandwidth
*rt_b
, int cpu
)
546 return &cpu_rq(cpu
)->rt
;
549 static inline struct rt_bandwidth
*sched_rt_bandwidth(struct rt_rq
*rt_rq
)
551 return &def_rt_bandwidth
;
554 #endif /* CONFIG_RT_GROUP_SCHED */
558 * We ran out of runtime, see if we can borrow some from our neighbours.
560 static int do_balance_runtime(struct rt_rq
*rt_rq
)
562 struct rt_bandwidth
*rt_b
= sched_rt_bandwidth(rt_rq
);
563 struct root_domain
*rd
= cpu_rq(smp_processor_id())->rd
;
564 int i
, weight
, more
= 0;
567 weight
= cpumask_weight(rd
->span
);
569 raw_spin_lock(&rt_b
->rt_runtime_lock
);
570 rt_period
= ktime_to_ns(rt_b
->rt_period
);
571 for_each_cpu(i
, rd
->span
) {
572 struct rt_rq
*iter
= sched_rt_period_rt_rq(rt_b
, i
);
578 raw_spin_lock(&iter
->rt_runtime_lock
);
580 * Either all rqs have inf runtime and there's nothing to steal
581 * or __disable_runtime() below sets a specific rq to inf to
582 * indicate its been disabled and disalow stealing.
584 if (iter
->rt_runtime
== RUNTIME_INF
)
588 * From runqueues with spare time, take 1/n part of their
589 * spare time, but no more than our period.
591 diff
= iter
->rt_runtime
- iter
->rt_time
;
593 diff
= div_u64((u64
)diff
, weight
);
594 if (rt_rq
->rt_runtime
+ diff
> rt_period
)
595 diff
= rt_period
- rt_rq
->rt_runtime
;
596 iter
->rt_runtime
-= diff
;
597 rt_rq
->rt_runtime
+= diff
;
599 if (rt_rq
->rt_runtime
== rt_period
) {
600 raw_spin_unlock(&iter
->rt_runtime_lock
);
605 raw_spin_unlock(&iter
->rt_runtime_lock
);
607 raw_spin_unlock(&rt_b
->rt_runtime_lock
);
613 * Ensure this RQ takes back all the runtime it lend to its neighbours.
615 static void __disable_runtime(struct rq
*rq
)
617 struct root_domain
*rd
= rq
->rd
;
621 if (unlikely(!scheduler_running
))
624 for_each_rt_rq(rt_rq
, iter
, rq
) {
625 struct rt_bandwidth
*rt_b
= sched_rt_bandwidth(rt_rq
);
629 raw_spin_lock(&rt_b
->rt_runtime_lock
);
630 raw_spin_lock(&rt_rq
->rt_runtime_lock
);
632 * Either we're all inf and nobody needs to borrow, or we're
633 * already disabled and thus have nothing to do, or we have
634 * exactly the right amount of runtime to take out.
636 if (rt_rq
->rt_runtime
== RUNTIME_INF
||
637 rt_rq
->rt_runtime
== rt_b
->rt_runtime
)
639 raw_spin_unlock(&rt_rq
->rt_runtime_lock
);
642 * Calculate the difference between what we started out with
643 * and what we current have, that's the amount of runtime
644 * we lend and now have to reclaim.
646 want
= rt_b
->rt_runtime
- rt_rq
->rt_runtime
;
649 * Greedy reclaim, take back as much as we can.
651 for_each_cpu(i
, rd
->span
) {
652 struct rt_rq
*iter
= sched_rt_period_rt_rq(rt_b
, i
);
656 * Can't reclaim from ourselves or disabled runqueues.
658 if (iter
== rt_rq
|| iter
->rt_runtime
== RUNTIME_INF
)
661 raw_spin_lock(&iter
->rt_runtime_lock
);
663 diff
= min_t(s64
, iter
->rt_runtime
, want
);
664 iter
->rt_runtime
-= diff
;
667 iter
->rt_runtime
-= want
;
670 raw_spin_unlock(&iter
->rt_runtime_lock
);
676 raw_spin_lock(&rt_rq
->rt_runtime_lock
);
678 * We cannot be left wanting - that would mean some runtime
679 * leaked out of the system.
684 * Disable all the borrow logic by pretending we have inf
685 * runtime - in which case borrowing doesn't make sense.
687 rt_rq
->rt_runtime
= RUNTIME_INF
;
688 raw_spin_unlock(&rt_rq
->rt_runtime_lock
);
689 raw_spin_unlock(&rt_b
->rt_runtime_lock
);
693 static void disable_runtime(struct rq
*rq
)
697 raw_spin_lock_irqsave(&rq
->lock
, flags
);
698 __disable_runtime(rq
);
699 raw_spin_unlock_irqrestore(&rq
->lock
, flags
);
702 static void __enable_runtime(struct rq
*rq
)
707 if (unlikely(!scheduler_running
))
711 * Reset each runqueue's bandwidth settings
713 for_each_rt_rq(rt_rq
, iter
, rq
) {
714 struct rt_bandwidth
*rt_b
= sched_rt_bandwidth(rt_rq
);
716 raw_spin_lock(&rt_b
->rt_runtime_lock
);
717 raw_spin_lock(&rt_rq
->rt_runtime_lock
);
718 rt_rq
->rt_runtime
= rt_b
->rt_runtime
;
720 rt_rq
->rt_throttled
= 0;
721 raw_spin_unlock(&rt_rq
->rt_runtime_lock
);
722 raw_spin_unlock(&rt_b
->rt_runtime_lock
);
726 static void enable_runtime(struct rq
*rq
)
730 raw_spin_lock_irqsave(&rq
->lock
, flags
);
731 __enable_runtime(rq
);
732 raw_spin_unlock_irqrestore(&rq
->lock
, flags
);
735 int update_runtime(struct notifier_block
*nfb
, unsigned long action
, void *hcpu
)
737 int cpu
= (int)(long)hcpu
;
740 case CPU_DOWN_PREPARE
:
741 case CPU_DOWN_PREPARE_FROZEN
:
742 disable_runtime(cpu_rq(cpu
));
745 case CPU_DOWN_FAILED
:
746 case CPU_DOWN_FAILED_FROZEN
:
748 case CPU_ONLINE_FROZEN
:
749 enable_runtime(cpu_rq(cpu
));
757 static int balance_runtime(struct rt_rq
*rt_rq
)
761 if (!sched_feat(RT_RUNTIME_SHARE
))
764 if (rt_rq
->rt_time
> rt_rq
->rt_runtime
) {
765 raw_spin_unlock(&rt_rq
->rt_runtime_lock
);
766 more
= do_balance_runtime(rt_rq
);
767 raw_spin_lock(&rt_rq
->rt_runtime_lock
);
772 #else /* !CONFIG_SMP */
773 static inline int balance_runtime(struct rt_rq
*rt_rq
)
777 #endif /* CONFIG_SMP */
779 static int do_sched_rt_period_timer(struct rt_bandwidth
*rt_b
, int overrun
)
782 const struct cpumask
*span
;
784 if (!rt_bandwidth_enabled() || rt_b
->rt_runtime
== RUNTIME_INF
)
787 span
= sched_rt_period_mask();
788 for_each_cpu(i
, span
) {
790 struct rt_rq
*rt_rq
= sched_rt_period_rt_rq(rt_b
, i
);
791 struct rq
*rq
= rq_of_rt_rq(rt_rq
);
793 raw_spin_lock(&rq
->lock
);
794 if (rt_rq
->rt_time
) {
797 raw_spin_lock(&rt_rq
->rt_runtime_lock
);
798 if (rt_rq
->rt_throttled
)
799 balance_runtime(rt_rq
);
800 runtime
= rt_rq
->rt_runtime
;
801 rt_rq
->rt_time
-= min(rt_rq
->rt_time
, overrun
*runtime
);
802 if (rt_rq
->rt_throttled
&& rt_rq
->rt_time
< runtime
) {
803 rt_rq
->rt_throttled
= 0;
807 * Force a clock update if the CPU was idle,
808 * lest wakeup -> unthrottle time accumulate.
810 if (rt_rq
->rt_nr_running
&& rq
->curr
== rq
->idle
)
811 rq
->skip_clock_update
= -1;
813 if (rt_rq
->rt_time
|| rt_rq
->rt_nr_running
)
815 raw_spin_unlock(&rt_rq
->rt_runtime_lock
);
816 } else if (rt_rq
->rt_nr_running
) {
818 if (!rt_rq_throttled(rt_rq
))
823 sched_rt_rq_enqueue(rt_rq
);
824 raw_spin_unlock(&rq
->lock
);
830 static inline int rt_se_prio(struct sched_rt_entity
*rt_se
)
832 #ifdef CONFIG_RT_GROUP_SCHED
833 struct rt_rq
*rt_rq
= group_rt_rq(rt_se
);
836 return rt_rq
->highest_prio
.curr
;
839 return rt_task_of(rt_se
)->prio
;
842 static int sched_rt_runtime_exceeded(struct rt_rq
*rt_rq
)
844 u64 runtime
= sched_rt_runtime(rt_rq
);
846 if (rt_rq
->rt_throttled
)
847 return rt_rq_throttled(rt_rq
);
849 if (runtime
>= sched_rt_period(rt_rq
))
852 balance_runtime(rt_rq
);
853 runtime
= sched_rt_runtime(rt_rq
);
854 if (runtime
== RUNTIME_INF
)
857 if (rt_rq
->rt_time
> runtime
) {
858 rt_rq
->rt_throttled
= 1;
859 printk_once(KERN_WARNING
"sched: RT throttling activated\n");
860 if (rt_rq_throttled(rt_rq
)) {
861 sched_rt_rq_dequeue(rt_rq
);
870 * Update the current task's runtime statistics. Skip current tasks that
871 * are not in our scheduling class.
873 static void update_curr_rt(struct rq
*rq
)
875 struct task_struct
*curr
= rq
->curr
;
876 struct sched_rt_entity
*rt_se
= &curr
->rt
;
877 struct rt_rq
*rt_rq
= rt_rq_of_se(rt_se
);
880 if (curr
->sched_class
!= &rt_sched_class
)
883 delta_exec
= rq
->clock_task
- curr
->se
.exec_start
;
884 if (unlikely((s64
)delta_exec
< 0))
887 schedstat_set(curr
->se
.statistics
.exec_max
, max(curr
->se
.statistics
.exec_max
, delta_exec
));
889 curr
->se
.sum_exec_runtime
+= delta_exec
;
890 account_group_exec_runtime(curr
, delta_exec
);
892 curr
->se
.exec_start
= rq
->clock_task
;
893 cpuacct_charge(curr
, delta_exec
);
895 sched_rt_avg_update(rq
, delta_exec
);
897 if (!rt_bandwidth_enabled())
900 for_each_sched_rt_entity(rt_se
) {
901 rt_rq
= rt_rq_of_se(rt_se
);
903 if (sched_rt_runtime(rt_rq
) != RUNTIME_INF
) {
904 raw_spin_lock(&rt_rq
->rt_runtime_lock
);
905 rt_rq
->rt_time
+= delta_exec
;
906 if (sched_rt_runtime_exceeded(rt_rq
))
908 raw_spin_unlock(&rt_rq
->rt_runtime_lock
);
913 #if defined CONFIG_SMP
916 inc_rt_prio_smp(struct rt_rq
*rt_rq
, int prio
, int prev_prio
)
918 struct rq
*rq
= rq_of_rt_rq(rt_rq
);
920 if (rq
->online
&& prio
< prev_prio
)
921 cpupri_set(&rq
->rd
->cpupri
, rq
->cpu
, prio
);
925 dec_rt_prio_smp(struct rt_rq
*rt_rq
, int prio
, int prev_prio
)
927 struct rq
*rq
= rq_of_rt_rq(rt_rq
);
929 if (rq
->online
&& rt_rq
->highest_prio
.curr
!= prev_prio
)
930 cpupri_set(&rq
->rd
->cpupri
, rq
->cpu
, rt_rq
->highest_prio
.curr
);
933 #else /* CONFIG_SMP */
936 void inc_rt_prio_smp(struct rt_rq
*rt_rq
, int prio
, int prev_prio
) {}
938 void dec_rt_prio_smp(struct rt_rq
*rt_rq
, int prio
, int prev_prio
) {}
940 #endif /* CONFIG_SMP */
942 #if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED
944 inc_rt_prio(struct rt_rq
*rt_rq
, int prio
)
946 int prev_prio
= rt_rq
->highest_prio
.curr
;
948 if (prio
< prev_prio
)
949 rt_rq
->highest_prio
.curr
= prio
;
951 inc_rt_prio_smp(rt_rq
, prio
, prev_prio
);
955 dec_rt_prio(struct rt_rq
*rt_rq
, int prio
)
957 int prev_prio
= rt_rq
->highest_prio
.curr
;
959 if (rt_rq
->rt_nr_running
) {
961 WARN_ON(prio
< prev_prio
);
964 * This may have been our highest task, and therefore
965 * we may have some recomputation to do
967 if (prio
== prev_prio
) {
968 struct rt_prio_array
*array
= &rt_rq
->active
;
970 rt_rq
->highest_prio
.curr
=
971 sched_find_first_bit(array
->bitmap
);
975 rt_rq
->highest_prio
.curr
= MAX_RT_PRIO
;
977 dec_rt_prio_smp(rt_rq
, prio
, prev_prio
);
982 static inline void inc_rt_prio(struct rt_rq
*rt_rq
, int prio
) {}
983 static inline void dec_rt_prio(struct rt_rq
*rt_rq
, int prio
) {}
985 #endif /* CONFIG_SMP || CONFIG_RT_GROUP_SCHED */
987 #ifdef CONFIG_RT_GROUP_SCHED
990 inc_rt_group(struct sched_rt_entity
*rt_se
, struct rt_rq
*rt_rq
)
992 if (rt_se_boosted(rt_se
))
993 rt_rq
->rt_nr_boosted
++;
996 start_rt_bandwidth(&rt_rq
->tg
->rt_bandwidth
);
1000 dec_rt_group(struct sched_rt_entity
*rt_se
, struct rt_rq
*rt_rq
)
1002 if (rt_se_boosted(rt_se
))
1003 rt_rq
->rt_nr_boosted
--;
1005 WARN_ON(!rt_rq
->rt_nr_running
&& rt_rq
->rt_nr_boosted
);
1008 #else /* CONFIG_RT_GROUP_SCHED */
1011 inc_rt_group(struct sched_rt_entity
*rt_se
, struct rt_rq
*rt_rq
)
1013 start_rt_bandwidth(&def_rt_bandwidth
);
1017 void dec_rt_group(struct sched_rt_entity
*rt_se
, struct rt_rq
*rt_rq
) {}
1019 #endif /* CONFIG_RT_GROUP_SCHED */
1022 void inc_rt_tasks(struct sched_rt_entity
*rt_se
, struct rt_rq
*rt_rq
)
1024 int prio
= rt_se_prio(rt_se
);
1026 WARN_ON(!rt_prio(prio
));
1027 rt_rq
->rt_nr_running
++;
1029 inc_rt_prio(rt_rq
, prio
);
1030 inc_rt_migration(rt_se
, rt_rq
);
1031 inc_rt_group(rt_se
, rt_rq
);
1035 void dec_rt_tasks(struct sched_rt_entity
*rt_se
, struct rt_rq
*rt_rq
)
1037 WARN_ON(!rt_prio(rt_se_prio(rt_se
)));
1038 WARN_ON(!rt_rq
->rt_nr_running
);
1039 rt_rq
->rt_nr_running
--;
1041 dec_rt_prio(rt_rq
, rt_se_prio(rt_se
));
1042 dec_rt_migration(rt_se
, rt_rq
);
1043 dec_rt_group(rt_se
, rt_rq
);
1046 static void __enqueue_rt_entity(struct sched_rt_entity
*rt_se
, bool head
)
1048 struct rt_rq
*rt_rq
= rt_rq_of_se(rt_se
);
1049 struct rt_prio_array
*array
= &rt_rq
->active
;
1050 struct rt_rq
*group_rq
= group_rt_rq(rt_se
);
1051 struct list_head
*queue
= array
->queue
+ rt_se_prio(rt_se
);
1054 * Don't enqueue the group if its throttled, or when empty.
1055 * The latter is a consequence of the former when a child group
1056 * get throttled and the current group doesn't have any other
1059 if (group_rq
&& (rt_rq_throttled(group_rq
) || !group_rq
->rt_nr_running
))
1062 if (!rt_rq
->rt_nr_running
)
1063 list_add_leaf_rt_rq(rt_rq
);
1066 list_add(&rt_se
->run_list
, queue
);
1068 list_add_tail(&rt_se
->run_list
, queue
);
1069 __set_bit(rt_se_prio(rt_se
), array
->bitmap
);
1071 inc_rt_tasks(rt_se
, rt_rq
);
1074 static void __dequeue_rt_entity(struct sched_rt_entity
*rt_se
)
1076 struct rt_rq
*rt_rq
= rt_rq_of_se(rt_se
);
1077 struct rt_prio_array
*array
= &rt_rq
->active
;
1079 list_del_init(&rt_se
->run_list
);
1080 if (list_empty(array
->queue
+ rt_se_prio(rt_se
)))
1081 __clear_bit(rt_se_prio(rt_se
), array
->bitmap
);
1083 dec_rt_tasks(rt_se
, rt_rq
);
1084 if (!rt_rq
->rt_nr_running
)
1085 list_del_leaf_rt_rq(rt_rq
);
1089 * Because the prio of an upper entry depends on the lower
1090 * entries, we must remove entries top - down.
1092 static void dequeue_rt_stack(struct sched_rt_entity
*rt_se
)
1094 struct sched_rt_entity
*back
= NULL
;
1096 for_each_sched_rt_entity(rt_se
) {
1101 for (rt_se
= back
; rt_se
; rt_se
= rt_se
->back
) {
1102 if (on_rt_rq(rt_se
))
1103 __dequeue_rt_entity(rt_se
);
1107 static void enqueue_rt_entity(struct sched_rt_entity
*rt_se
, bool head
)
1109 dequeue_rt_stack(rt_se
);
1110 for_each_sched_rt_entity(rt_se
)
1111 __enqueue_rt_entity(rt_se
, head
);
1114 static void dequeue_rt_entity(struct sched_rt_entity
*rt_se
)
1116 dequeue_rt_stack(rt_se
);
1118 for_each_sched_rt_entity(rt_se
) {
1119 struct rt_rq
*rt_rq
= group_rt_rq(rt_se
);
1121 if (rt_rq
&& rt_rq
->rt_nr_running
)
1122 __enqueue_rt_entity(rt_se
, false);
1127 * Adding/removing a task to/from a priority array:
1130 enqueue_task_rt(struct rq
*rq
, struct task_struct
*p
, int flags
)
1132 struct sched_rt_entity
*rt_se
= &p
->rt
;
1134 if (flags
& ENQUEUE_WAKEUP
)
1137 enqueue_rt_entity(rt_se
, flags
& ENQUEUE_HEAD
);
1139 if (!task_current(rq
, p
) && p
->rt
.nr_cpus_allowed
> 1)
1140 enqueue_pushable_task(rq
, p
);
1145 static void dequeue_task_rt(struct rq
*rq
, struct task_struct
*p
, int flags
)
1147 struct sched_rt_entity
*rt_se
= &p
->rt
;
1150 dequeue_rt_entity(rt_se
);
1152 dequeue_pushable_task(rq
, p
);
1158 * Put task to the head or the end of the run list without the overhead of
1159 * dequeue followed by enqueue.
1162 requeue_rt_entity(struct rt_rq
*rt_rq
, struct sched_rt_entity
*rt_se
, int head
)
1164 if (on_rt_rq(rt_se
)) {
1165 struct rt_prio_array
*array
= &rt_rq
->active
;
1166 struct list_head
*queue
= array
->queue
+ rt_se_prio(rt_se
);
1169 list_move(&rt_se
->run_list
, queue
);
1171 list_move_tail(&rt_se
->run_list
, queue
);
1175 static void requeue_task_rt(struct rq
*rq
, struct task_struct
*p
, int head
)
1177 struct sched_rt_entity
*rt_se
= &p
->rt
;
1178 struct rt_rq
*rt_rq
;
1180 for_each_sched_rt_entity(rt_se
) {
1181 rt_rq
= rt_rq_of_se(rt_se
);
1182 requeue_rt_entity(rt_rq
, rt_se
, head
);
1186 static void yield_task_rt(struct rq
*rq
)
1188 requeue_task_rt(rq
, rq
->curr
, 0);
1192 static int find_lowest_rq(struct task_struct
*task
);
1195 select_task_rq_rt(struct task_struct
*p
, int sd_flag
, int flags
)
1197 struct task_struct
*curr
;
1203 if (p
->rt
.nr_cpus_allowed
== 1)
1206 /* For anything but wake ups, just return the task_cpu */
1207 if (sd_flag
!= SD_BALANCE_WAKE
&& sd_flag
!= SD_BALANCE_FORK
)
1213 curr
= ACCESS_ONCE(rq
->curr
); /* unlocked access */
1216 * If the current task on @p's runqueue is an RT task, then
1217 * try to see if we can wake this RT task up on another
1218 * runqueue. Otherwise simply start this RT task
1219 * on its current runqueue.
1221 * We want to avoid overloading runqueues. If the woken
1222 * task is a higher priority, then it will stay on this CPU
1223 * and the lower prio task should be moved to another CPU.
1224 * Even though this will probably make the lower prio task
1225 * lose its cache, we do not want to bounce a higher task
1226 * around just because it gave up its CPU, perhaps for a
1229 * For equal prio tasks, we just let the scheduler sort it out.
1231 * Otherwise, just let it ride on the affined RQ and the
1232 * post-schedule router will push the preempted task away
1234 * This test is optimistic, if we get it wrong the load-balancer
1235 * will have to sort it out.
1237 if (curr
&& unlikely(rt_task(curr
)) &&
1238 (curr
->rt
.nr_cpus_allowed
< 2 ||
1239 curr
->prio
<= p
->prio
) &&
1240 (p
->rt
.nr_cpus_allowed
> 1)) {
1241 int target
= find_lowest_rq(p
);
1252 static void check_preempt_equal_prio(struct rq
*rq
, struct task_struct
*p
)
1254 if (rq
->curr
->rt
.nr_cpus_allowed
== 1)
1257 if (p
->rt
.nr_cpus_allowed
!= 1
1258 && cpupri_find(&rq
->rd
->cpupri
, p
, NULL
))
1261 if (!cpupri_find(&rq
->rd
->cpupri
, rq
->curr
, NULL
))
1265 * There appears to be other cpus that can accept
1266 * current and none to run 'p', so lets reschedule
1267 * to try and push current away:
1269 requeue_task_rt(rq
, p
, 1);
1270 resched_task(rq
->curr
);
1273 #endif /* CONFIG_SMP */
1276 * Preempt the current task with a newly woken task if needed:
1278 static void check_preempt_curr_rt(struct rq
*rq
, struct task_struct
*p
, int flags
)
1280 if (p
->prio
< rq
->curr
->prio
) {
1281 resched_task(rq
->curr
);
1289 * - the newly woken task is of equal priority to the current task
1290 * - the newly woken task is non-migratable while current is migratable
1291 * - current will be preempted on the next reschedule
1293 * we should check to see if current can readily move to a different
1294 * cpu. If so, we will reschedule to allow the push logic to try
1295 * to move current somewhere else, making room for our non-migratable
1298 if (p
->prio
== rq
->curr
->prio
&& !test_tsk_need_resched(rq
->curr
))
1299 check_preempt_equal_prio(rq
, p
);
1303 static struct sched_rt_entity
*pick_next_rt_entity(struct rq
*rq
,
1304 struct rt_rq
*rt_rq
)
1306 struct rt_prio_array
*array
= &rt_rq
->active
;
1307 struct sched_rt_entity
*next
= NULL
;
1308 struct list_head
*queue
;
1311 idx
= sched_find_first_bit(array
->bitmap
);
1312 BUG_ON(idx
>= MAX_RT_PRIO
);
1314 queue
= array
->queue
+ idx
;
1315 next
= list_entry(queue
->next
, struct sched_rt_entity
, run_list
);
1320 static struct task_struct
*_pick_next_task_rt(struct rq
*rq
)
1322 struct sched_rt_entity
*rt_se
;
1323 struct task_struct
*p
;
1324 struct rt_rq
*rt_rq
;
1328 if (!rt_rq
->rt_nr_running
)
1331 if (rt_rq_throttled(rt_rq
))
1335 rt_se
= pick_next_rt_entity(rq
, rt_rq
);
1337 rt_rq
= group_rt_rq(rt_se
);
1340 p
= rt_task_of(rt_se
);
1341 p
->se
.exec_start
= rq
->clock_task
;
1346 static struct task_struct
*pick_next_task_rt(struct rq
*rq
)
1348 struct task_struct
*p
= _pick_next_task_rt(rq
);
1350 /* The running task is never eligible for pushing */
1352 dequeue_pushable_task(rq
, p
);
1356 * We detect this state here so that we can avoid taking the RQ
1357 * lock again later if there is no need to push
1359 rq
->post_schedule
= has_pushable_tasks(rq
);
1365 static void put_prev_task_rt(struct rq
*rq
, struct task_struct
*p
)
1370 * The previous task needs to be made eligible for pushing
1371 * if it is still active
1373 if (on_rt_rq(&p
->rt
) && p
->rt
.nr_cpus_allowed
> 1)
1374 enqueue_pushable_task(rq
, p
);
1379 /* Only try algorithms three times */
1380 #define RT_MAX_TRIES 3
1382 static int pick_rt_task(struct rq
*rq
, struct task_struct
*p
, int cpu
)
1384 if (!task_running(rq
, p
) &&
1385 (cpu
< 0 || cpumask_test_cpu(cpu
, tsk_cpus_allowed(p
))) &&
1386 (p
->rt
.nr_cpus_allowed
> 1))
1391 /* Return the second highest RT task, NULL otherwise */
1392 static struct task_struct
*pick_next_highest_task_rt(struct rq
*rq
, int cpu
)
1394 struct task_struct
*next
= NULL
;
1395 struct sched_rt_entity
*rt_se
;
1396 struct rt_prio_array
*array
;
1397 struct rt_rq
*rt_rq
;
1400 for_each_leaf_rt_rq(rt_rq
, rq
) {
1401 array
= &rt_rq
->active
;
1402 idx
= sched_find_first_bit(array
->bitmap
);
1404 if (idx
>= MAX_RT_PRIO
)
1406 if (next
&& next
->prio
< idx
)
1408 list_for_each_entry(rt_se
, array
->queue
+ idx
, run_list
) {
1409 struct task_struct
*p
;
1411 if (!rt_entity_is_task(rt_se
))
1414 p
= rt_task_of(rt_se
);
1415 if (pick_rt_task(rq
, p
, cpu
)) {
1421 idx
= find_next_bit(array
->bitmap
, MAX_RT_PRIO
, idx
+1);
1429 static DEFINE_PER_CPU(cpumask_var_t
, local_cpu_mask
);
1431 static int find_lowest_rq(struct task_struct
*task
)
1433 struct sched_domain
*sd
;
1434 struct cpumask
*lowest_mask
= __get_cpu_var(local_cpu_mask
);
1435 int this_cpu
= smp_processor_id();
1436 int cpu
= task_cpu(task
);
1438 /* Make sure the mask is initialized first */
1439 if (unlikely(!lowest_mask
))
1442 if (task
->rt
.nr_cpus_allowed
== 1)
1443 return -1; /* No other targets possible */
1445 if (!cpupri_find(&task_rq(task
)->rd
->cpupri
, task
, lowest_mask
))
1446 return -1; /* No targets found */
1449 * At this point we have built a mask of cpus representing the
1450 * lowest priority tasks in the system. Now we want to elect
1451 * the best one based on our affinity and topology.
1453 * We prioritize the last cpu that the task executed on since
1454 * it is most likely cache-hot in that location.
1456 if (cpumask_test_cpu(cpu
, lowest_mask
))
1460 * Otherwise, we consult the sched_domains span maps to figure
1461 * out which cpu is logically closest to our hot cache data.
1463 if (!cpumask_test_cpu(this_cpu
, lowest_mask
))
1464 this_cpu
= -1; /* Skip this_cpu opt if not among lowest */
1467 for_each_domain(cpu
, sd
) {
1468 if (sd
->flags
& SD_WAKE_AFFINE
) {
1472 * "this_cpu" is cheaper to preempt than a
1475 if (this_cpu
!= -1 &&
1476 cpumask_test_cpu(this_cpu
, sched_domain_span(sd
))) {
1481 best_cpu
= cpumask_first_and(lowest_mask
,
1482 sched_domain_span(sd
));
1483 if (best_cpu
< nr_cpu_ids
) {
1492 * And finally, if there were no matches within the domains
1493 * just give the caller *something* to work with from the compatible
1499 cpu
= cpumask_any(lowest_mask
);
1500 if (cpu
< nr_cpu_ids
)
1505 /* Will lock the rq it finds */
1506 static struct rq
*find_lock_lowest_rq(struct task_struct
*task
, struct rq
*rq
)
1508 struct rq
*lowest_rq
= NULL
;
1512 for (tries
= 0; tries
< RT_MAX_TRIES
; tries
++) {
1513 cpu
= find_lowest_rq(task
);
1515 if ((cpu
== -1) || (cpu
== rq
->cpu
))
1518 lowest_rq
= cpu_rq(cpu
);
1520 /* if the prio of this runqueue changed, try again */
1521 if (double_lock_balance(rq
, lowest_rq
)) {
1523 * We had to unlock the run queue. In
1524 * the mean time, task could have
1525 * migrated already or had its affinity changed.
1526 * Also make sure that it wasn't scheduled on its rq.
1528 if (unlikely(task_rq(task
) != rq
||
1529 !cpumask_test_cpu(lowest_rq
->cpu
,
1530 tsk_cpus_allowed(task
)) ||
1531 task_running(rq
, task
) ||
1534 raw_spin_unlock(&lowest_rq
->lock
);
1540 /* If this rq is still suitable use it. */
1541 if (lowest_rq
->rt
.highest_prio
.curr
> task
->prio
)
1545 double_unlock_balance(rq
, lowest_rq
);
1552 static struct task_struct
*pick_next_pushable_task(struct rq
*rq
)
1554 struct task_struct
*p
;
1556 if (!has_pushable_tasks(rq
))
1559 p
= plist_first_entry(&rq
->rt
.pushable_tasks
,
1560 struct task_struct
, pushable_tasks
);
1562 BUG_ON(rq
->cpu
!= task_cpu(p
));
1563 BUG_ON(task_current(rq
, p
));
1564 BUG_ON(p
->rt
.nr_cpus_allowed
<= 1);
1567 BUG_ON(!rt_task(p
));
1573 * If the current CPU has more than one RT task, see if the non
1574 * running task can migrate over to a CPU that is running a task
1575 * of lesser priority.
1577 static int push_rt_task(struct rq
*rq
)
1579 struct task_struct
*next_task
;
1580 struct rq
*lowest_rq
;
1583 if (!rq
->rt
.overloaded
)
1586 next_task
= pick_next_pushable_task(rq
);
1591 if (unlikely(next_task
== rq
->curr
)) {
1597 * It's possible that the next_task slipped in of
1598 * higher priority than current. If that's the case
1599 * just reschedule current.
1601 if (unlikely(next_task
->prio
< rq
->curr
->prio
)) {
1602 resched_task(rq
->curr
);
1606 /* We might release rq lock */
1607 get_task_struct(next_task
);
1609 /* find_lock_lowest_rq locks the rq if found */
1610 lowest_rq
= find_lock_lowest_rq(next_task
, rq
);
1612 struct task_struct
*task
;
1614 * find_lock_lowest_rq releases rq->lock
1615 * so it is possible that next_task has migrated.
1617 * We need to make sure that the task is still on the same
1618 * run-queue and is also still the next task eligible for
1621 task
= pick_next_pushable_task(rq
);
1622 if (task_cpu(next_task
) == rq
->cpu
&& task
== next_task
) {
1624 * The task hasn't migrated, and is still the next
1625 * eligible task, but we failed to find a run-queue
1626 * to push it to. Do not retry in this case, since
1627 * other cpus will pull from us when ready.
1633 /* No more tasks, just exit */
1637 * Something has shifted, try again.
1639 put_task_struct(next_task
);
1644 deactivate_task(rq
, next_task
, 0);
1645 set_task_cpu(next_task
, lowest_rq
->cpu
);
1646 activate_task(lowest_rq
, next_task
, 0);
1649 resched_task(lowest_rq
->curr
);
1651 double_unlock_balance(rq
, lowest_rq
);
1654 put_task_struct(next_task
);
1659 static void push_rt_tasks(struct rq
*rq
)
1661 /* push_rt_task will return true if it moved an RT */
1662 while (push_rt_task(rq
))
1666 static int pull_rt_task(struct rq
*this_rq
)
1668 int this_cpu
= this_rq
->cpu
, ret
= 0, cpu
;
1669 struct task_struct
*p
;
1672 if (likely(!rt_overloaded(this_rq
)))
1675 for_each_cpu(cpu
, this_rq
->rd
->rto_mask
) {
1676 if (this_cpu
== cpu
)
1679 src_rq
= cpu_rq(cpu
);
1682 * Don't bother taking the src_rq->lock if the next highest
1683 * task is known to be lower-priority than our current task.
1684 * This may look racy, but if this value is about to go
1685 * logically higher, the src_rq will push this task away.
1686 * And if its going logically lower, we do not care
1688 if (src_rq
->rt
.highest_prio
.next
>=
1689 this_rq
->rt
.highest_prio
.curr
)
1693 * We can potentially drop this_rq's lock in
1694 * double_lock_balance, and another CPU could
1697 double_lock_balance(this_rq
, src_rq
);
1700 * Are there still pullable RT tasks?
1702 if (src_rq
->rt
.rt_nr_running
<= 1)
1705 p
= pick_next_highest_task_rt(src_rq
, this_cpu
);
1708 * Do we have an RT task that preempts
1709 * the to-be-scheduled task?
1711 if (p
&& (p
->prio
< this_rq
->rt
.highest_prio
.curr
)) {
1712 WARN_ON(p
== src_rq
->curr
);
1716 * There's a chance that p is higher in priority
1717 * than what's currently running on its cpu.
1718 * This is just that p is wakeing up and hasn't
1719 * had a chance to schedule. We only pull
1720 * p if it is lower in priority than the
1721 * current task on the run queue
1723 if (p
->prio
< src_rq
->curr
->prio
)
1728 deactivate_task(src_rq
, p
, 0);
1729 set_task_cpu(p
, this_cpu
);
1730 activate_task(this_rq
, p
, 0);
1732 * We continue with the search, just in
1733 * case there's an even higher prio task
1734 * in another runqueue. (low likelihood
1739 double_unlock_balance(this_rq
, src_rq
);
1745 static void pre_schedule_rt(struct rq
*rq
, struct task_struct
*prev
)
1747 /* Try to pull RT tasks here if we lower this rq's prio */
1748 if (rq
->rt
.highest_prio
.curr
> prev
->prio
)
1752 static void post_schedule_rt(struct rq
*rq
)
1758 * If we are not running and we are not going to reschedule soon, we should
1759 * try to push tasks away now
1761 static void task_woken_rt(struct rq
*rq
, struct task_struct
*p
)
1763 if (!task_running(rq
, p
) &&
1764 !test_tsk_need_resched(rq
->curr
) &&
1765 has_pushable_tasks(rq
) &&
1766 p
->rt
.nr_cpus_allowed
> 1 &&
1767 rt_task(rq
->curr
) &&
1768 (rq
->curr
->rt
.nr_cpus_allowed
< 2 ||
1769 rq
->curr
->prio
<= p
->prio
))
1773 static void set_cpus_allowed_rt(struct task_struct
*p
,
1774 const struct cpumask
*new_mask
)
1776 int weight
= cpumask_weight(new_mask
);
1778 BUG_ON(!rt_task(p
));
1781 * Update the migration status of the RQ if we have an RT task
1782 * which is running AND changing its weight value.
1784 if (p
->on_rq
&& (weight
!= p
->rt
.nr_cpus_allowed
)) {
1785 struct rq
*rq
= task_rq(p
);
1787 if (!task_current(rq
, p
)) {
1789 * Make sure we dequeue this task from the pushable list
1790 * before going further. It will either remain off of
1791 * the list because we are no longer pushable, or it
1794 if (p
->rt
.nr_cpus_allowed
> 1)
1795 dequeue_pushable_task(rq
, p
);
1798 * Requeue if our weight is changing and still > 1
1801 enqueue_pushable_task(rq
, p
);
1805 if ((p
->rt
.nr_cpus_allowed
<= 1) && (weight
> 1)) {
1806 rq
->rt
.rt_nr_migratory
++;
1807 } else if ((p
->rt
.nr_cpus_allowed
> 1) && (weight
<= 1)) {
1808 BUG_ON(!rq
->rt
.rt_nr_migratory
);
1809 rq
->rt
.rt_nr_migratory
--;
1812 update_rt_migration(&rq
->rt
);
1816 /* Assumes rq->lock is held */
1817 static void rq_online_rt(struct rq
*rq
)
1819 if (rq
->rt
.overloaded
)
1820 rt_set_overload(rq
);
1822 __enable_runtime(rq
);
1824 cpupri_set(&rq
->rd
->cpupri
, rq
->cpu
, rq
->rt
.highest_prio
.curr
);
1827 /* Assumes rq->lock is held */
1828 static void rq_offline_rt(struct rq
*rq
)
1830 if (rq
->rt
.overloaded
)
1831 rt_clear_overload(rq
);
1833 __disable_runtime(rq
);
1835 cpupri_set(&rq
->rd
->cpupri
, rq
->cpu
, CPUPRI_INVALID
);
1839 * When switch from the rt queue, we bring ourselves to a position
1840 * that we might want to pull RT tasks from other runqueues.
1842 static void switched_from_rt(struct rq
*rq
, struct task_struct
*p
)
1845 * If there are other RT tasks then we will reschedule
1846 * and the scheduling of the other RT tasks will handle
1847 * the balancing. But if we are the last RT task
1848 * we may need to handle the pulling of RT tasks
1851 if (p
->on_rq
&& !rq
->rt
.rt_nr_running
)
1855 void init_sched_rt_class(void)
1859 for_each_possible_cpu(i
) {
1860 zalloc_cpumask_var_node(&per_cpu(local_cpu_mask
, i
),
1861 GFP_KERNEL
, cpu_to_node(i
));
1864 #endif /* CONFIG_SMP */
1867 * When switching a task to RT, we may overload the runqueue
1868 * with RT tasks. In this case we try to push them off to
1871 static void switched_to_rt(struct rq
*rq
, struct task_struct
*p
)
1873 int check_resched
= 1;
1876 * If we are already running, then there's nothing
1877 * that needs to be done. But if we are not running
1878 * we may need to preempt the current running task.
1879 * If that current running task is also an RT task
1880 * then see if we can move to another run queue.
1882 if (p
->on_rq
&& rq
->curr
!= p
) {
1884 if (rq
->rt
.overloaded
&& push_rt_task(rq
) &&
1885 /* Don't resched if we changed runqueues */
1888 #endif /* CONFIG_SMP */
1889 if (check_resched
&& p
->prio
< rq
->curr
->prio
)
1890 resched_task(rq
->curr
);
1895 * Priority of the task has changed. This may cause
1896 * us to initiate a push or pull.
1899 prio_changed_rt(struct rq
*rq
, struct task_struct
*p
, int oldprio
)
1904 if (rq
->curr
== p
) {
1907 * If our priority decreases while running, we
1908 * may need to pull tasks to this runqueue.
1910 if (oldprio
< p
->prio
)
1913 * If there's a higher priority task waiting to run
1914 * then reschedule. Note, the above pull_rt_task
1915 * can release the rq lock and p could migrate.
1916 * Only reschedule if p is still on the same runqueue.
1918 if (p
->prio
> rq
->rt
.highest_prio
.curr
&& rq
->curr
== p
)
1921 /* For UP simply resched on drop of prio */
1922 if (oldprio
< p
->prio
)
1924 #endif /* CONFIG_SMP */
1927 * This task is not running, but if it is
1928 * greater than the current running task
1931 if (p
->prio
< rq
->curr
->prio
)
1932 resched_task(rq
->curr
);
1936 static void watchdog(struct rq
*rq
, struct task_struct
*p
)
1938 unsigned long soft
, hard
;
1940 /* max may change after cur was read, this will be fixed next tick */
1941 soft
= task_rlimit(p
, RLIMIT_RTTIME
);
1942 hard
= task_rlimit_max(p
, RLIMIT_RTTIME
);
1944 if (soft
!= RLIM_INFINITY
) {
1948 next
= DIV_ROUND_UP(min(soft
, hard
), USEC_PER_SEC
/HZ
);
1949 if (p
->rt
.timeout
> next
)
1950 p
->cputime_expires
.sched_exp
= p
->se
.sum_exec_runtime
;
1954 static void task_tick_rt(struct rq
*rq
, struct task_struct
*p
, int queued
)
1961 * RR tasks need a special form of timeslice management.
1962 * FIFO tasks have no timeslices.
1964 if (p
->policy
!= SCHED_RR
)
1967 if (--p
->rt
.time_slice
)
1970 p
->rt
.time_slice
= DEF_TIMESLICE
;
1973 * Requeue to the end of queue if we are not the only element
1976 if (p
->rt
.run_list
.prev
!= p
->rt
.run_list
.next
) {
1977 requeue_task_rt(rq
, p
, 0);
1978 set_tsk_need_resched(p
);
1982 static void set_curr_task_rt(struct rq
*rq
)
1984 struct task_struct
*p
= rq
->curr
;
1986 p
->se
.exec_start
= rq
->clock_task
;
1988 /* The running task is never eligible for pushing */
1989 dequeue_pushable_task(rq
, p
);
1992 static unsigned int get_rr_interval_rt(struct rq
*rq
, struct task_struct
*task
)
1995 * Time slice is 0 for SCHED_FIFO tasks
1997 if (task
->policy
== SCHED_RR
)
1998 return DEF_TIMESLICE
;
2003 const struct sched_class rt_sched_class
= {
2004 .next
= &fair_sched_class
,
2005 .enqueue_task
= enqueue_task_rt
,
2006 .dequeue_task
= dequeue_task_rt
,
2007 .yield_task
= yield_task_rt
,
2009 .check_preempt_curr
= check_preempt_curr_rt
,
2011 .pick_next_task
= pick_next_task_rt
,
2012 .put_prev_task
= put_prev_task_rt
,
2015 .select_task_rq
= select_task_rq_rt
,
2017 .set_cpus_allowed
= set_cpus_allowed_rt
,
2018 .rq_online
= rq_online_rt
,
2019 .rq_offline
= rq_offline_rt
,
2020 .pre_schedule
= pre_schedule_rt
,
2021 .post_schedule
= post_schedule_rt
,
2022 .task_woken
= task_woken_rt
,
2023 .switched_from
= switched_from_rt
,
2026 .set_curr_task
= set_curr_task_rt
,
2027 .task_tick
= task_tick_rt
,
2029 .get_rr_interval
= get_rr_interval_rt
,
2031 .prio_changed
= prio_changed_rt
,
2032 .switched_to
= switched_to_rt
,
2035 #ifdef CONFIG_SCHED_DEBUG
2036 extern void print_rt_rq(struct seq_file
*m
, int cpu
, struct rt_rq
*rt_rq
);
2038 void print_rt_stats(struct seq_file
*m
, int cpu
)
2041 struct rt_rq
*rt_rq
;
2044 for_each_rt_rq(rt_rq
, iter
, cpu_rq(cpu
))
2045 print_rt_rq(m
, cpu
, rt_rq
);
2048 #endif /* CONFIG_SCHED_DEBUG */