Memory controller: OOM handling
[linux-2.6.git] / kernel / sched_rt.c
1 /*
2  * Real-Time Scheduling Class (mapped to the SCHED_FIFO and SCHED_RR
3  * policies)
4  */
5
6 #ifdef CONFIG_SMP
7
8 static inline int rt_overloaded(struct rq *rq)
9 {
10         return atomic_read(&rq->rd->rto_count);
11 }
12
13 static inline void rt_set_overload(struct rq *rq)
14 {
15         cpu_set(rq->cpu, rq->rd->rto_mask);
16         /*
17          * Make sure the mask is visible before we set
18          * the overload count. That is checked to determine
19          * if we should look at the mask. It would be a shame
20          * if we looked at the mask, but the mask was not
21          * updated yet.
22          */
23         wmb();
24         atomic_inc(&rq->rd->rto_count);
25 }
26
27 static inline void rt_clear_overload(struct rq *rq)
28 {
29         /* the order here really doesn't matter */
30         atomic_dec(&rq->rd->rto_count);
31         cpu_clear(rq->cpu, rq->rd->rto_mask);
32 }
33
34 static void update_rt_migration(struct rq *rq)
35 {
36         if (rq->rt.rt_nr_migratory && (rq->rt.rt_nr_running > 1)) {
37                 if (!rq->rt.overloaded) {
38                         rt_set_overload(rq);
39                         rq->rt.overloaded = 1;
40                 }
41         } else if (rq->rt.overloaded) {
42                 rt_clear_overload(rq);
43                 rq->rt.overloaded = 0;
44         }
45 }
46 #endif /* CONFIG_SMP */
47
48 static inline struct task_struct *rt_task_of(struct sched_rt_entity *rt_se)
49 {
50         return container_of(rt_se, struct task_struct, rt);
51 }
52
53 static inline int on_rt_rq(struct sched_rt_entity *rt_se)
54 {
55         return !list_empty(&rt_se->run_list);
56 }
57
58 #ifdef CONFIG_FAIR_GROUP_SCHED
59
60 static inline unsigned int sched_rt_ratio(struct rt_rq *rt_rq)
61 {
62         if (!rt_rq->tg)
63                 return SCHED_RT_FRAC;
64
65         return rt_rq->tg->rt_ratio;
66 }
67
68 #define for_each_leaf_rt_rq(rt_rq, rq) \
69         list_for_each_entry(rt_rq, &rq->leaf_rt_rq_list, leaf_rt_rq_list)
70
71 static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq)
72 {
73         return rt_rq->rq;
74 }
75
76 static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se)
77 {
78         return rt_se->rt_rq;
79 }
80
81 #define for_each_sched_rt_entity(rt_se) \
82         for (; rt_se; rt_se = rt_se->parent)
83
84 static inline struct rt_rq *group_rt_rq(struct sched_rt_entity *rt_se)
85 {
86         return rt_se->my_q;
87 }
88
89 static void enqueue_rt_entity(struct sched_rt_entity *rt_se);
90 static void dequeue_rt_entity(struct sched_rt_entity *rt_se);
91
92 static void sched_rt_ratio_enqueue(struct rt_rq *rt_rq)
93 {
94         struct sched_rt_entity *rt_se = rt_rq->rt_se;
95
96         if (rt_se && !on_rt_rq(rt_se) && rt_rq->rt_nr_running) {
97                 struct task_struct *curr = rq_of_rt_rq(rt_rq)->curr;
98
99                 enqueue_rt_entity(rt_se);
100                 if (rt_rq->highest_prio < curr->prio)
101                         resched_task(curr);
102         }
103 }
104
105 static void sched_rt_ratio_dequeue(struct rt_rq *rt_rq)
106 {
107         struct sched_rt_entity *rt_se = rt_rq->rt_se;
108
109         if (rt_se && on_rt_rq(rt_se))
110                 dequeue_rt_entity(rt_se);
111 }
112
113 #else
114
115 static inline unsigned int sched_rt_ratio(struct rt_rq *rt_rq)
116 {
117         return sysctl_sched_rt_ratio;
118 }
119
120 #define for_each_leaf_rt_rq(rt_rq, rq) \
121         for (rt_rq = &rq->rt; rt_rq; rt_rq = NULL)
122
123 static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq)
124 {
125         return container_of(rt_rq, struct rq, rt);
126 }
127
128 static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se)
129 {
130         struct task_struct *p = rt_task_of(rt_se);
131         struct rq *rq = task_rq(p);
132
133         return &rq->rt;
134 }
135
136 #define for_each_sched_rt_entity(rt_se) \
137         for (; rt_se; rt_se = NULL)
138
139 static inline struct rt_rq *group_rt_rq(struct sched_rt_entity *rt_se)
140 {
141         return NULL;
142 }
143
144 static inline void sched_rt_ratio_enqueue(struct rt_rq *rt_rq)
145 {
146 }
147
148 static inline void sched_rt_ratio_dequeue(struct rt_rq *rt_rq)
149 {
150 }
151
152 #endif
153
154 static inline int rt_se_prio(struct sched_rt_entity *rt_se)
155 {
156 #ifdef CONFIG_FAIR_GROUP_SCHED
157         struct rt_rq *rt_rq = group_rt_rq(rt_se);
158
159         if (rt_rq)
160                 return rt_rq->highest_prio;
161 #endif
162
163         return rt_task_of(rt_se)->prio;
164 }
165
166 static int sched_rt_ratio_exceeded(struct rt_rq *rt_rq)
167 {
168         unsigned int rt_ratio = sched_rt_ratio(rt_rq);
169         u64 period, ratio;
170
171         if (rt_ratio == SCHED_RT_FRAC)
172                 return 0;
173
174         if (rt_rq->rt_throttled)
175                 return 1;
176
177         period = (u64)sysctl_sched_rt_period * NSEC_PER_MSEC;
178         ratio = (period * rt_ratio) >> SCHED_RT_FRAC_SHIFT;
179
180         if (rt_rq->rt_time > ratio) {
181                 struct rq *rq = rq_of_rt_rq(rt_rq);
182
183                 rq->rt_throttled = 1;
184                 rt_rq->rt_throttled = 1;
185
186                 sched_rt_ratio_dequeue(rt_rq);
187                 return 1;
188         }
189
190         return 0;
191 }
192
193 static void update_sched_rt_period(struct rq *rq)
194 {
195         struct rt_rq *rt_rq;
196         u64 period;
197
198         while (rq->clock > rq->rt_period_expire) {
199                 period = (u64)sysctl_sched_rt_period * NSEC_PER_MSEC;
200                 rq->rt_period_expire += period;
201
202                 for_each_leaf_rt_rq(rt_rq, rq) {
203                         unsigned long rt_ratio = sched_rt_ratio(rt_rq);
204                         u64 ratio = (period * rt_ratio) >> SCHED_RT_FRAC_SHIFT;
205
206                         rt_rq->rt_time -= min(rt_rq->rt_time, ratio);
207                         if (rt_rq->rt_throttled) {
208                                 rt_rq->rt_throttled = 0;
209                                 sched_rt_ratio_enqueue(rt_rq);
210                         }
211                 }
212
213                 rq->rt_throttled = 0;
214         }
215 }
216
217 /*
218  * Update the current task's runtime statistics. Skip current tasks that
219  * are not in our scheduling class.
220  */
221 static void update_curr_rt(struct rq *rq)
222 {
223         struct task_struct *curr = rq->curr;
224         struct sched_rt_entity *rt_se = &curr->rt;
225         struct rt_rq *rt_rq = rt_rq_of_se(rt_se);
226         u64 delta_exec;
227
228         if (!task_has_rt_policy(curr))
229                 return;
230
231         delta_exec = rq->clock - curr->se.exec_start;
232         if (unlikely((s64)delta_exec < 0))
233                 delta_exec = 0;
234
235         schedstat_set(curr->se.exec_max, max(curr->se.exec_max, delta_exec));
236
237         curr->se.sum_exec_runtime += delta_exec;
238         curr->se.exec_start = rq->clock;
239         cpuacct_charge(curr, delta_exec);
240
241         rt_rq->rt_time += delta_exec;
242         /*
243          * might make it a tad more accurate:
244          *
245          * update_sched_rt_period(rq);
246          */
247         if (sched_rt_ratio_exceeded(rt_rq))
248                 resched_task(curr);
249 }
250
251 static inline
252 void inc_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
253 {
254         WARN_ON(!rt_prio(rt_se_prio(rt_se)));
255         rt_rq->rt_nr_running++;
256 #if defined CONFIG_SMP || defined CONFIG_FAIR_GROUP_SCHED
257         if (rt_se_prio(rt_se) < rt_rq->highest_prio)
258                 rt_rq->highest_prio = rt_se_prio(rt_se);
259 #endif
260 #ifdef CONFIG_SMP
261         if (rt_se->nr_cpus_allowed > 1) {
262                 struct rq *rq = rq_of_rt_rq(rt_rq);
263                 rq->rt.rt_nr_migratory++;
264         }
265
266         update_rt_migration(rq_of_rt_rq(rt_rq));
267 #endif
268 }
269
270 static inline
271 void dec_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
272 {
273         WARN_ON(!rt_prio(rt_se_prio(rt_se)));
274         WARN_ON(!rt_rq->rt_nr_running);
275         rt_rq->rt_nr_running--;
276 #if defined CONFIG_SMP || defined CONFIG_FAIR_GROUP_SCHED
277         if (rt_rq->rt_nr_running) {
278                 struct rt_prio_array *array;
279
280                 WARN_ON(rt_se_prio(rt_se) < rt_rq->highest_prio);
281                 if (rt_se_prio(rt_se) == rt_rq->highest_prio) {
282                         /* recalculate */
283                         array = &rt_rq->active;
284                         rt_rq->highest_prio =
285                                 sched_find_first_bit(array->bitmap);
286                 } /* otherwise leave rq->highest prio alone */
287         } else
288                 rt_rq->highest_prio = MAX_RT_PRIO;
289 #endif
290 #ifdef CONFIG_SMP
291         if (rt_se->nr_cpus_allowed > 1) {
292                 struct rq *rq = rq_of_rt_rq(rt_rq);
293                 rq->rt.rt_nr_migratory--;
294         }
295
296         update_rt_migration(rq_of_rt_rq(rt_rq));
297 #endif /* CONFIG_SMP */
298 }
299
300 static void enqueue_rt_entity(struct sched_rt_entity *rt_se)
301 {
302         struct rt_rq *rt_rq = rt_rq_of_se(rt_se);
303         struct rt_prio_array *array = &rt_rq->active;
304         struct rt_rq *group_rq = group_rt_rq(rt_se);
305
306         if (group_rq && group_rq->rt_throttled)
307                 return;
308
309         list_add_tail(&rt_se->run_list, array->queue + rt_se_prio(rt_se));
310         __set_bit(rt_se_prio(rt_se), array->bitmap);
311
312         inc_rt_tasks(rt_se, rt_rq);
313 }
314
315 static void dequeue_rt_entity(struct sched_rt_entity *rt_se)
316 {
317         struct rt_rq *rt_rq = rt_rq_of_se(rt_se);
318         struct rt_prio_array *array = &rt_rq->active;
319
320         list_del_init(&rt_se->run_list);
321         if (list_empty(array->queue + rt_se_prio(rt_se)))
322                 __clear_bit(rt_se_prio(rt_se), array->bitmap);
323
324         dec_rt_tasks(rt_se, rt_rq);
325 }
326
327 /*
328  * Because the prio of an upper entry depends on the lower
329  * entries, we must remove entries top - down.
330  *
331  * XXX: O(1/2 h^2) because we can only walk up, not down the chain.
332  *      doesn't matter much for now, as h=2 for GROUP_SCHED.
333  */
334 static void dequeue_rt_stack(struct task_struct *p)
335 {
336         struct sched_rt_entity *rt_se, *top_se;
337
338         /*
339          * dequeue all, top - down.
340          */
341         do {
342                 rt_se = &p->rt;
343                 top_se = NULL;
344                 for_each_sched_rt_entity(rt_se) {
345                         if (on_rt_rq(rt_se))
346                                 top_se = rt_se;
347                 }
348                 if (top_se)
349                         dequeue_rt_entity(top_se);
350         } while (top_se);
351 }
352
353 /*
354  * Adding/removing a task to/from a priority array:
355  */
356 static void enqueue_task_rt(struct rq *rq, struct task_struct *p, int wakeup)
357 {
358         struct sched_rt_entity *rt_se = &p->rt;
359
360         if (wakeup)
361                 rt_se->timeout = 0;
362
363         dequeue_rt_stack(p);
364
365         /*
366          * enqueue everybody, bottom - up.
367          */
368         for_each_sched_rt_entity(rt_se)
369                 enqueue_rt_entity(rt_se);
370
371         inc_cpu_load(rq, p->se.load.weight);
372 }
373
374 static void dequeue_task_rt(struct rq *rq, struct task_struct *p, int sleep)
375 {
376         struct sched_rt_entity *rt_se = &p->rt;
377         struct rt_rq *rt_rq;
378
379         update_curr_rt(rq);
380
381         dequeue_rt_stack(p);
382
383         /*
384          * re-enqueue all non-empty rt_rq entities.
385          */
386         for_each_sched_rt_entity(rt_se) {
387                 rt_rq = group_rt_rq(rt_se);
388                 if (rt_rq && rt_rq->rt_nr_running)
389                         enqueue_rt_entity(rt_se);
390         }
391
392         dec_cpu_load(rq, p->se.load.weight);
393 }
394
395 /*
396  * Put task to the end of the run list without the overhead of dequeue
397  * followed by enqueue.
398  */
399 static
400 void requeue_rt_entity(struct rt_rq *rt_rq, struct sched_rt_entity *rt_se)
401 {
402         struct rt_prio_array *array = &rt_rq->active;
403
404         list_move_tail(&rt_se->run_list, array->queue + rt_se_prio(rt_se));
405 }
406
407 static void requeue_task_rt(struct rq *rq, struct task_struct *p)
408 {
409         struct sched_rt_entity *rt_se = &p->rt;
410         struct rt_rq *rt_rq;
411
412         for_each_sched_rt_entity(rt_se) {
413                 rt_rq = rt_rq_of_se(rt_se);
414                 requeue_rt_entity(rt_rq, rt_se);
415         }
416 }
417
418 static void yield_task_rt(struct rq *rq)
419 {
420         requeue_task_rt(rq, rq->curr);
421 }
422
423 #ifdef CONFIG_SMP
424 static int find_lowest_rq(struct task_struct *task);
425
426 static int select_task_rq_rt(struct task_struct *p, int sync)
427 {
428         struct rq *rq = task_rq(p);
429
430         /*
431          * If the current task is an RT task, then
432          * try to see if we can wake this RT task up on another
433          * runqueue. Otherwise simply start this RT task
434          * on its current runqueue.
435          *
436          * We want to avoid overloading runqueues. Even if
437          * the RT task is of higher priority than the current RT task.
438          * RT tasks behave differently than other tasks. If
439          * one gets preempted, we try to push it off to another queue.
440          * So trying to keep a preempting RT task on the same
441          * cache hot CPU will force the running RT task to
442          * a cold CPU. So we waste all the cache for the lower
443          * RT task in hopes of saving some of a RT task
444          * that is just being woken and probably will have
445          * cold cache anyway.
446          */
447         if (unlikely(rt_task(rq->curr)) &&
448             (p->rt.nr_cpus_allowed > 1)) {
449                 int cpu = find_lowest_rq(p);
450
451                 return (cpu == -1) ? task_cpu(p) : cpu;
452         }
453
454         /*
455          * Otherwise, just let it ride on the affined RQ and the
456          * post-schedule router will push the preempted task away
457          */
458         return task_cpu(p);
459 }
460 #endif /* CONFIG_SMP */
461
462 /*
463  * Preempt the current task with a newly woken task if needed:
464  */
465 static void check_preempt_curr_rt(struct rq *rq, struct task_struct *p)
466 {
467         if (p->prio < rq->curr->prio)
468                 resched_task(rq->curr);
469 }
470
471 static struct sched_rt_entity *pick_next_rt_entity(struct rq *rq,
472                                                    struct rt_rq *rt_rq)
473 {
474         struct rt_prio_array *array = &rt_rq->active;
475         struct sched_rt_entity *next = NULL;
476         struct list_head *queue;
477         int idx;
478
479         idx = sched_find_first_bit(array->bitmap);
480         BUG_ON(idx >= MAX_RT_PRIO);
481
482         queue = array->queue + idx;
483         next = list_entry(queue->next, struct sched_rt_entity, run_list);
484
485         return next;
486 }
487
488 static struct task_struct *pick_next_task_rt(struct rq *rq)
489 {
490         struct sched_rt_entity *rt_se;
491         struct task_struct *p;
492         struct rt_rq *rt_rq;
493
494         rt_rq = &rq->rt;
495
496         if (unlikely(!rt_rq->rt_nr_running))
497                 return NULL;
498
499         if (sched_rt_ratio_exceeded(rt_rq))
500                 return NULL;
501
502         do {
503                 rt_se = pick_next_rt_entity(rq, rt_rq);
504                 BUG_ON(!rt_se);
505                 rt_rq = group_rt_rq(rt_se);
506         } while (rt_rq);
507
508         p = rt_task_of(rt_se);
509         p->se.exec_start = rq->clock;
510         return p;
511 }
512
513 static void put_prev_task_rt(struct rq *rq, struct task_struct *p)
514 {
515         update_curr_rt(rq);
516         p->se.exec_start = 0;
517 }
518
519 #ifdef CONFIG_SMP
520
521 /* Only try algorithms three times */
522 #define RT_MAX_TRIES 3
523
524 static int double_lock_balance(struct rq *this_rq, struct rq *busiest);
525 static void deactivate_task(struct rq *rq, struct task_struct *p, int sleep);
526
527 static int pick_rt_task(struct rq *rq, struct task_struct *p, int cpu)
528 {
529         if (!task_running(rq, p) &&
530             (cpu < 0 || cpu_isset(cpu, p->cpus_allowed)) &&
531             (p->rt.nr_cpus_allowed > 1))
532                 return 1;
533         return 0;
534 }
535
536 /* Return the second highest RT task, NULL otherwise */
537 static struct task_struct *pick_next_highest_task_rt(struct rq *rq, int cpu)
538 {
539         struct task_struct *next = NULL;
540         struct sched_rt_entity *rt_se;
541         struct rt_prio_array *array;
542         struct rt_rq *rt_rq;
543         int idx;
544
545         for_each_leaf_rt_rq(rt_rq, rq) {
546                 array = &rt_rq->active;
547                 idx = sched_find_first_bit(array->bitmap);
548  next_idx:
549                 if (idx >= MAX_RT_PRIO)
550                         continue;
551                 if (next && next->prio < idx)
552                         continue;
553                 list_for_each_entry(rt_se, array->queue + idx, run_list) {
554                         struct task_struct *p = rt_task_of(rt_se);
555                         if (pick_rt_task(rq, p, cpu)) {
556                                 next = p;
557                                 break;
558                         }
559                 }
560                 if (!next) {
561                         idx = find_next_bit(array->bitmap, MAX_RT_PRIO, idx+1);
562                         goto next_idx;
563                 }
564         }
565
566         return next;
567 }
568
569 static DEFINE_PER_CPU(cpumask_t, local_cpu_mask);
570
571 static int find_lowest_cpus(struct task_struct *task, cpumask_t *lowest_mask)
572 {
573         int       lowest_prio = -1;
574         int       lowest_cpu  = -1;
575         int       count       = 0;
576         int       cpu;
577
578         cpus_and(*lowest_mask, task_rq(task)->rd->online, task->cpus_allowed);
579
580         /*
581          * Scan each rq for the lowest prio.
582          */
583         for_each_cpu_mask(cpu, *lowest_mask) {
584                 struct rq *rq = cpu_rq(cpu);
585
586                 /* We look for lowest RT prio or non-rt CPU */
587                 if (rq->rt.highest_prio >= MAX_RT_PRIO) {
588                         /*
589                          * if we already found a low RT queue
590                          * and now we found this non-rt queue
591                          * clear the mask and set our bit.
592                          * Otherwise just return the queue as is
593                          * and the count==1 will cause the algorithm
594                          * to use the first bit found.
595                          */
596                         if (lowest_cpu != -1) {
597                                 cpus_clear(*lowest_mask);
598                                 cpu_set(rq->cpu, *lowest_mask);
599                         }
600                         return 1;
601                 }
602
603                 /* no locking for now */
604                 if ((rq->rt.highest_prio > task->prio)
605                     && (rq->rt.highest_prio >= lowest_prio)) {
606                         if (rq->rt.highest_prio > lowest_prio) {
607                                 /* new low - clear old data */
608                                 lowest_prio = rq->rt.highest_prio;
609                                 lowest_cpu = cpu;
610                                 count = 0;
611                         }
612                         count++;
613                 } else
614                         cpu_clear(cpu, *lowest_mask);
615         }
616
617         /*
618          * Clear out all the set bits that represent
619          * runqueues that were of higher prio than
620          * the lowest_prio.
621          */
622         if (lowest_cpu > 0) {
623                 /*
624                  * Perhaps we could add another cpumask op to
625                  * zero out bits. Like cpu_zero_bits(cpumask, nrbits);
626                  * Then that could be optimized to use memset and such.
627                  */
628                 for_each_cpu_mask(cpu, *lowest_mask) {
629                         if (cpu >= lowest_cpu)
630                                 break;
631                         cpu_clear(cpu, *lowest_mask);
632                 }
633         }
634
635         return count;
636 }
637
638 static inline int pick_optimal_cpu(int this_cpu, cpumask_t *mask)
639 {
640         int first;
641
642         /* "this_cpu" is cheaper to preempt than a remote processor */
643         if ((this_cpu != -1) && cpu_isset(this_cpu, *mask))
644                 return this_cpu;
645
646         first = first_cpu(*mask);
647         if (first != NR_CPUS)
648                 return first;
649
650         return -1;
651 }
652
653 static int find_lowest_rq(struct task_struct *task)
654 {
655         struct sched_domain *sd;
656         cpumask_t *lowest_mask = &__get_cpu_var(local_cpu_mask);
657         int this_cpu = smp_processor_id();
658         int cpu      = task_cpu(task);
659         int count    = find_lowest_cpus(task, lowest_mask);
660
661         if (!count)
662                 return -1; /* No targets found */
663
664         /*
665          * There is no sense in performing an optimal search if only one
666          * target is found.
667          */
668         if (count == 1)
669                 return first_cpu(*lowest_mask);
670
671         /*
672          * At this point we have built a mask of cpus representing the
673          * lowest priority tasks in the system.  Now we want to elect
674          * the best one based on our affinity and topology.
675          *
676          * We prioritize the last cpu that the task executed on since
677          * it is most likely cache-hot in that location.
678          */
679         if (cpu_isset(cpu, *lowest_mask))
680                 return cpu;
681
682         /*
683          * Otherwise, we consult the sched_domains span maps to figure
684          * out which cpu is logically closest to our hot cache data.
685          */
686         if (this_cpu == cpu)
687                 this_cpu = -1; /* Skip this_cpu opt if the same */
688
689         for_each_domain(cpu, sd) {
690                 if (sd->flags & SD_WAKE_AFFINE) {
691                         cpumask_t domain_mask;
692                         int       best_cpu;
693
694                         cpus_and(domain_mask, sd->span, *lowest_mask);
695
696                         best_cpu = pick_optimal_cpu(this_cpu,
697                                                     &domain_mask);
698                         if (best_cpu != -1)
699                                 return best_cpu;
700                 }
701         }
702
703         /*
704          * And finally, if there were no matches within the domains
705          * just give the caller *something* to work with from the compatible
706          * locations.
707          */
708         return pick_optimal_cpu(this_cpu, lowest_mask);
709 }
710
711 /* Will lock the rq it finds */
712 static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq)
713 {
714         struct rq *lowest_rq = NULL;
715         int tries;
716         int cpu;
717
718         for (tries = 0; tries < RT_MAX_TRIES; tries++) {
719                 cpu = find_lowest_rq(task);
720
721                 if ((cpu == -1) || (cpu == rq->cpu))
722                         break;
723
724                 lowest_rq = cpu_rq(cpu);
725
726                 /* if the prio of this runqueue changed, try again */
727                 if (double_lock_balance(rq, lowest_rq)) {
728                         /*
729                          * We had to unlock the run queue. In
730                          * the mean time, task could have
731                          * migrated already or had its affinity changed.
732                          * Also make sure that it wasn't scheduled on its rq.
733                          */
734                         if (unlikely(task_rq(task) != rq ||
735                                      !cpu_isset(lowest_rq->cpu,
736                                                 task->cpus_allowed) ||
737                                      task_running(rq, task) ||
738                                      !task->se.on_rq)) {
739
740                                 spin_unlock(&lowest_rq->lock);
741                                 lowest_rq = NULL;
742                                 break;
743                         }
744                 }
745
746                 /* If this rq is still suitable use it. */
747                 if (lowest_rq->rt.highest_prio > task->prio)
748                         break;
749
750                 /* try again */
751                 spin_unlock(&lowest_rq->lock);
752                 lowest_rq = NULL;
753         }
754
755         return lowest_rq;
756 }
757
758 /*
759  * If the current CPU has more than one RT task, see if the non
760  * running task can migrate over to a CPU that is running a task
761  * of lesser priority.
762  */
763 static int push_rt_task(struct rq *rq)
764 {
765         struct task_struct *next_task;
766         struct rq *lowest_rq;
767         int ret = 0;
768         int paranoid = RT_MAX_TRIES;
769
770         if (!rq->rt.overloaded)
771                 return 0;
772
773         next_task = pick_next_highest_task_rt(rq, -1);
774         if (!next_task)
775                 return 0;
776
777  retry:
778         if (unlikely(next_task == rq->curr)) {
779                 WARN_ON(1);
780                 return 0;
781         }
782
783         /*
784          * It's possible that the next_task slipped in of
785          * higher priority than current. If that's the case
786          * just reschedule current.
787          */
788         if (unlikely(next_task->prio < rq->curr->prio)) {
789                 resched_task(rq->curr);
790                 return 0;
791         }
792
793         /* We might release rq lock */
794         get_task_struct(next_task);
795
796         /* find_lock_lowest_rq locks the rq if found */
797         lowest_rq = find_lock_lowest_rq(next_task, rq);
798         if (!lowest_rq) {
799                 struct task_struct *task;
800                 /*
801                  * find lock_lowest_rq releases rq->lock
802                  * so it is possible that next_task has changed.
803                  * If it has, then try again.
804                  */
805                 task = pick_next_highest_task_rt(rq, -1);
806                 if (unlikely(task != next_task) && task && paranoid--) {
807                         put_task_struct(next_task);
808                         next_task = task;
809                         goto retry;
810                 }
811                 goto out;
812         }
813
814         deactivate_task(rq, next_task, 0);
815         set_task_cpu(next_task, lowest_rq->cpu);
816         activate_task(lowest_rq, next_task, 0);
817
818         resched_task(lowest_rq->curr);
819
820         spin_unlock(&lowest_rq->lock);
821
822         ret = 1;
823 out:
824         put_task_struct(next_task);
825
826         return ret;
827 }
828
829 /*
830  * TODO: Currently we just use the second highest prio task on
831  *       the queue, and stop when it can't migrate (or there's
832  *       no more RT tasks).  There may be a case where a lower
833  *       priority RT task has a different affinity than the
834  *       higher RT task. In this case the lower RT task could
835  *       possibly be able to migrate where as the higher priority
836  *       RT task could not.  We currently ignore this issue.
837  *       Enhancements are welcome!
838  */
839 static void push_rt_tasks(struct rq *rq)
840 {
841         /* push_rt_task will return true if it moved an RT */
842         while (push_rt_task(rq))
843                 ;
844 }
845
846 static int pull_rt_task(struct rq *this_rq)
847 {
848         int this_cpu = this_rq->cpu, ret = 0, cpu;
849         struct task_struct *p, *next;
850         struct rq *src_rq;
851
852         if (likely(!rt_overloaded(this_rq)))
853                 return 0;
854
855         next = pick_next_task_rt(this_rq);
856
857         for_each_cpu_mask(cpu, this_rq->rd->rto_mask) {
858                 if (this_cpu == cpu)
859                         continue;
860
861                 src_rq = cpu_rq(cpu);
862                 /*
863                  * We can potentially drop this_rq's lock in
864                  * double_lock_balance, and another CPU could
865                  * steal our next task - hence we must cause
866                  * the caller to recalculate the next task
867                  * in that case:
868                  */
869                 if (double_lock_balance(this_rq, src_rq)) {
870                         struct task_struct *old_next = next;
871
872                         next = pick_next_task_rt(this_rq);
873                         if (next != old_next)
874                                 ret = 1;
875                 }
876
877                 /*
878                  * Are there still pullable RT tasks?
879                  */
880                 if (src_rq->rt.rt_nr_running <= 1)
881                         goto skip;
882
883                 p = pick_next_highest_task_rt(src_rq, this_cpu);
884
885                 /*
886                  * Do we have an RT task that preempts
887                  * the to-be-scheduled task?
888                  */
889                 if (p && (!next || (p->prio < next->prio))) {
890                         WARN_ON(p == src_rq->curr);
891                         WARN_ON(!p->se.on_rq);
892
893                         /*
894                          * There's a chance that p is higher in priority
895                          * than what's currently running on its cpu.
896                          * This is just that p is wakeing up and hasn't
897                          * had a chance to schedule. We only pull
898                          * p if it is lower in priority than the
899                          * current task on the run queue or
900                          * this_rq next task is lower in prio than
901                          * the current task on that rq.
902                          */
903                         if (p->prio < src_rq->curr->prio ||
904                             (next && next->prio < src_rq->curr->prio))
905                                 goto skip;
906
907                         ret = 1;
908
909                         deactivate_task(src_rq, p, 0);
910                         set_task_cpu(p, this_cpu);
911                         activate_task(this_rq, p, 0);
912                         /*
913                          * We continue with the search, just in
914                          * case there's an even higher prio task
915                          * in another runqueue. (low likelyhood
916                          * but possible)
917                          *
918                          * Update next so that we won't pick a task
919                          * on another cpu with a priority lower (or equal)
920                          * than the one we just picked.
921                          */
922                         next = p;
923
924                 }
925  skip:
926                 spin_unlock(&src_rq->lock);
927         }
928
929         return ret;
930 }
931
932 static void pre_schedule_rt(struct rq *rq, struct task_struct *prev)
933 {
934         /* Try to pull RT tasks here if we lower this rq's prio */
935         if (unlikely(rt_task(prev)) && rq->rt.highest_prio > prev->prio)
936                 pull_rt_task(rq);
937 }
938
939 static void post_schedule_rt(struct rq *rq)
940 {
941         /*
942          * If we have more than one rt_task queued, then
943          * see if we can push the other rt_tasks off to other CPUS.
944          * Note we may release the rq lock, and since
945          * the lock was owned by prev, we need to release it
946          * first via finish_lock_switch and then reaquire it here.
947          */
948         if (unlikely(rq->rt.overloaded)) {
949                 spin_lock_irq(&rq->lock);
950                 push_rt_tasks(rq);
951                 spin_unlock_irq(&rq->lock);
952         }
953 }
954
955
956 static void task_wake_up_rt(struct rq *rq, struct task_struct *p)
957 {
958         if (!task_running(rq, p) &&
959             (p->prio >= rq->rt.highest_prio) &&
960             rq->rt.overloaded)
961                 push_rt_tasks(rq);
962 }
963
964 static unsigned long
965 load_balance_rt(struct rq *this_rq, int this_cpu, struct rq *busiest,
966                 unsigned long max_load_move,
967                 struct sched_domain *sd, enum cpu_idle_type idle,
968                 int *all_pinned, int *this_best_prio)
969 {
970         /* don't touch RT tasks */
971         return 0;
972 }
973
974 static int
975 move_one_task_rt(struct rq *this_rq, int this_cpu, struct rq *busiest,
976                  struct sched_domain *sd, enum cpu_idle_type idle)
977 {
978         /* don't touch RT tasks */
979         return 0;
980 }
981
982 static void set_cpus_allowed_rt(struct task_struct *p, cpumask_t *new_mask)
983 {
984         int weight = cpus_weight(*new_mask);
985
986         BUG_ON(!rt_task(p));
987
988         /*
989          * Update the migration status of the RQ if we have an RT task
990          * which is running AND changing its weight value.
991          */
992         if (p->se.on_rq && (weight != p->rt.nr_cpus_allowed)) {
993                 struct rq *rq = task_rq(p);
994
995                 if ((p->rt.nr_cpus_allowed <= 1) && (weight > 1)) {
996                         rq->rt.rt_nr_migratory++;
997                 } else if ((p->rt.nr_cpus_allowed > 1) && (weight <= 1)) {
998                         BUG_ON(!rq->rt.rt_nr_migratory);
999                         rq->rt.rt_nr_migratory--;
1000                 }
1001
1002                 update_rt_migration(rq);
1003         }
1004
1005         p->cpus_allowed    = *new_mask;
1006         p->rt.nr_cpus_allowed = weight;
1007 }
1008
1009 /* Assumes rq->lock is held */
1010 static void join_domain_rt(struct rq *rq)
1011 {
1012         if (rq->rt.overloaded)
1013                 rt_set_overload(rq);
1014 }
1015
1016 /* Assumes rq->lock is held */
1017 static void leave_domain_rt(struct rq *rq)
1018 {
1019         if (rq->rt.overloaded)
1020                 rt_clear_overload(rq);
1021 }
1022
1023 /*
1024  * When switch from the rt queue, we bring ourselves to a position
1025  * that we might want to pull RT tasks from other runqueues.
1026  */
1027 static void switched_from_rt(struct rq *rq, struct task_struct *p,
1028                            int running)
1029 {
1030         /*
1031          * If there are other RT tasks then we will reschedule
1032          * and the scheduling of the other RT tasks will handle
1033          * the balancing. But if we are the last RT task
1034          * we may need to handle the pulling of RT tasks
1035          * now.
1036          */
1037         if (!rq->rt.rt_nr_running)
1038                 pull_rt_task(rq);
1039 }
1040 #endif /* CONFIG_SMP */
1041
1042 /*
1043  * When switching a task to RT, we may overload the runqueue
1044  * with RT tasks. In this case we try to push them off to
1045  * other runqueues.
1046  */
1047 static void switched_to_rt(struct rq *rq, struct task_struct *p,
1048                            int running)
1049 {
1050         int check_resched = 1;
1051
1052         /*
1053          * If we are already running, then there's nothing
1054          * that needs to be done. But if we are not running
1055          * we may need to preempt the current running task.
1056          * If that current running task is also an RT task
1057          * then see if we can move to another run queue.
1058          */
1059         if (!running) {
1060 #ifdef CONFIG_SMP
1061                 if (rq->rt.overloaded && push_rt_task(rq) &&
1062                     /* Don't resched if we changed runqueues */
1063                     rq != task_rq(p))
1064                         check_resched = 0;
1065 #endif /* CONFIG_SMP */
1066                 if (check_resched && p->prio < rq->curr->prio)
1067                         resched_task(rq->curr);
1068         }
1069 }
1070
1071 /*
1072  * Priority of the task has changed. This may cause
1073  * us to initiate a push or pull.
1074  */
1075 static void prio_changed_rt(struct rq *rq, struct task_struct *p,
1076                             int oldprio, int running)
1077 {
1078         if (running) {
1079 #ifdef CONFIG_SMP
1080                 /*
1081                  * If our priority decreases while running, we
1082                  * may need to pull tasks to this runqueue.
1083                  */
1084                 if (oldprio < p->prio)
1085                         pull_rt_task(rq);
1086                 /*
1087                  * If there's a higher priority task waiting to run
1088                  * then reschedule.
1089                  */
1090                 if (p->prio > rq->rt.highest_prio)
1091                         resched_task(p);
1092 #else
1093                 /* For UP simply resched on drop of prio */
1094                 if (oldprio < p->prio)
1095                         resched_task(p);
1096 #endif /* CONFIG_SMP */
1097         } else {
1098                 /*
1099                  * This task is not running, but if it is
1100                  * greater than the current running task
1101                  * then reschedule.
1102                  */
1103                 if (p->prio < rq->curr->prio)
1104                         resched_task(rq->curr);
1105         }
1106 }
1107
1108 static void watchdog(struct rq *rq, struct task_struct *p)
1109 {
1110         unsigned long soft, hard;
1111
1112         if (!p->signal)
1113                 return;
1114
1115         soft = p->signal->rlim[RLIMIT_RTTIME].rlim_cur;
1116         hard = p->signal->rlim[RLIMIT_RTTIME].rlim_max;
1117
1118         if (soft != RLIM_INFINITY) {
1119                 unsigned long next;
1120
1121                 p->rt.timeout++;
1122                 next = DIV_ROUND_UP(min(soft, hard), USEC_PER_SEC/HZ);
1123                 if (p->rt.timeout > next)
1124                         p->it_sched_expires = p->se.sum_exec_runtime;
1125         }
1126 }
1127
1128 static void task_tick_rt(struct rq *rq, struct task_struct *p, int queued)
1129 {
1130         update_curr_rt(rq);
1131
1132         watchdog(rq, p);
1133
1134         /*
1135          * RR tasks need a special form of timeslice management.
1136          * FIFO tasks have no timeslices.
1137          */
1138         if (p->policy != SCHED_RR)
1139                 return;
1140
1141         if (--p->rt.time_slice)
1142                 return;
1143
1144         p->rt.time_slice = DEF_TIMESLICE;
1145
1146         /*
1147          * Requeue to the end of queue if we are not the only element
1148          * on the queue:
1149          */
1150         if (p->rt.run_list.prev != p->rt.run_list.next) {
1151                 requeue_task_rt(rq, p);
1152                 set_tsk_need_resched(p);
1153         }
1154 }
1155
1156 static void set_curr_task_rt(struct rq *rq)
1157 {
1158         struct task_struct *p = rq->curr;
1159
1160         p->se.exec_start = rq->clock;
1161 }
1162
1163 const struct sched_class rt_sched_class = {
1164         .next                   = &fair_sched_class,
1165         .enqueue_task           = enqueue_task_rt,
1166         .dequeue_task           = dequeue_task_rt,
1167         .yield_task             = yield_task_rt,
1168 #ifdef CONFIG_SMP
1169         .select_task_rq         = select_task_rq_rt,
1170 #endif /* CONFIG_SMP */
1171
1172         .check_preempt_curr     = check_preempt_curr_rt,
1173
1174         .pick_next_task         = pick_next_task_rt,
1175         .put_prev_task          = put_prev_task_rt,
1176
1177 #ifdef CONFIG_SMP
1178         .load_balance           = load_balance_rt,
1179         .move_one_task          = move_one_task_rt,
1180         .set_cpus_allowed       = set_cpus_allowed_rt,
1181         .join_domain            = join_domain_rt,
1182         .leave_domain           = leave_domain_rt,
1183         .pre_schedule           = pre_schedule_rt,
1184         .post_schedule          = post_schedule_rt,
1185         .task_wake_up           = task_wake_up_rt,
1186         .switched_from          = switched_from_rt,
1187 #endif
1188
1189         .set_curr_task          = set_curr_task_rt,
1190         .task_tick              = task_tick_rt,
1191
1192         .prio_changed           = prio_changed_rt,
1193         .switched_to            = switched_to_rt,
1194 };