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_RT_GROUP_SCHED
   7
   8 #define rt_entity_is_task(rt_se) (!(rt_se)->my_q)
   9
  10 static inline struct task_struct *rt_task_of(struct sched_rt_entity *rt_se)
  11 {
  12 #ifdef CONFIG_SCHED_DEBUG
  13         WARN_ON_ONCE(!rt_entity_is_task(rt_se));
  14 #endif
  15         return container_of(rt_se, struct task_struct, rt);
  16 }
  17
  18 static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq)
  19 {
  20         return rt_rq->rq;
  21 }
  22
  23 static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se)
  24 {
  25         return rt_se->rt_rq;
  26 }
  27
  28 #else /* CONFIG_RT_GROUP_SCHED */
  29
  30 #define rt_entity_is_task(rt_se) (1)
  31
  32 static inline struct task_struct *rt_task_of(struct sched_rt_entity *rt_se)
  33 {
  34         return container_of(rt_se, struct task_struct, rt);
  35 }
  36
  37 static inline struct rq *rq_of_rt_rq(struct rt_rq *rt_rq)
  38 {
  39         return container_of(rt_rq, struct rq, rt);
  40 }
  41
  42 static inline struct rt_rq *rt_rq_of_se(struct sched_rt_entity *rt_se)
  43 {
  44         struct task_struct *p = rt_task_of(rt_se);
  45         struct rq *rq = task_rq(p);
  46
  47         return &rq->rt;
  48 }
  49
  50 #endif /* CONFIG_RT_GROUP_SCHED */
  51
  52 #ifdef CONFIG_SMP
  53
  54 static inline int rt_overloaded(struct rq *rq)
  55 {
  56         return atomic_read(&rq->rd->rto_count);
  57 }
  58
  59 static inline void rt_set_overload(struct rq *rq)
  60 {
  61         if (!rq->online)
  62                 return;
  63
  64         cpumask_set_cpu(rq->cpu, rq->rd->rto_mask);
  65         /*
  66          * Make sure the mask is visible before we set
  67          * the overload count. That is checked to determine
  68          * if we should look at the mask. It would be a shame
  69          * if we looked at the mask, but the mask was not
  70          * updated yet.
  71          */
  72         wmb();
  73         atomic_inc(&rq->rd->rto_count);
  74 }
  75
  76 static inline void rt_clear_overload(struct rq *rq)
  77 {
  78         if (!rq->online)
  79                 return;
  80
  81         /* the order here really doesn't matter */
  82         atomic_dec(&rq->rd->rto_count);
  83         cpumask_clear_cpu(rq->cpu, rq->rd->rto_mask);
  84 }
  85
  86 static void update_rt_migration(struct rt_rq *rt_rq)
  87 {
  88         if (rt_rq->rt_nr_migratory && rt_rq->rt_nr_total > 1) {
  89                 if (!rt_rq->overloaded) {
  90                         rt_set_overload(rq_of_rt_rq(rt_rq));
  91                         rt_rq->overloaded = 1;
  92                 }
  93         } else if (rt_rq->overloaded) {
  94                 rt_clear_overload(rq_of_rt_rq(rt_rq));
  95                 rt_rq->overloaded = 0;
  96         }
  97 }
  98
  99 static void inc_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
 100 {
 101         if (!rt_entity_is_task(rt_se))
 102                 return;
 103
 104         rt_rq = &rq_of_rt_rq(rt_rq)->rt;
 105
 106         rt_rq->rt_nr_total++;
 107         if (rt_se->nr_cpus_allowed > 1)
 108                 rt_rq->rt_nr_migratory++;
 109
 110         update_rt_migration(rt_rq);
 111 }
 112
 113 static void dec_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
 114 {
 115         if (!rt_entity_is_task(rt_se))
 116                 return;
 117
 118         rt_rq = &rq_of_rt_rq(rt_rq)->rt;
 119
 120         rt_rq->rt_nr_total--;
 121         if (rt_se->nr_cpus_allowed > 1)
 122                 rt_rq->rt_nr_migratory--;
 123
 124         update_rt_migration(rt_rq);
 125 }
 126
 127 static inline int has_pushable_tasks(struct rq *rq)
 128 {
 129         return !plist_head_empty(&rq->rt.pushable_tasks);
 130 }
 131
 132 static void enqueue_pushable_task(struct rq *rq, struct task_struct *p)
 133 {
 134         plist_del(&p->pushable_tasks, &rq->rt.pushable_tasks);
 135         plist_node_init(&p->pushable_tasks, p->prio);
 136         plist_add(&p->pushable_tasks, &rq->rt.pushable_tasks);
 137
 138         /* Update the highest prio pushable task */
 139         if (p->prio < rq->rt.highest_prio.next)
 140                 rq->rt.highest_prio.next = p->prio;
 141 }
 142
 143 static void dequeue_pushable_task(struct rq *rq, struct task_struct *p)
 144 {
 145         plist_del(&p->pushable_tasks, &rq->rt.pushable_tasks);
 146
 147         /* Update the new highest prio pushable task */
 148         if (has_pushable_tasks(rq)) {
 149                 p = plist_first_entry(&rq->rt.pushable_tasks,
 150                                       struct task_struct, pushable_tasks);
 151                 rq->rt.highest_prio.next = p->prio;
 152         } else
 153                 rq->rt.highest_prio.next = MAX_RT_PRIO;
 154 }
 155
 156 #else
 157
 158 static inline void enqueue_pushable_task(struct rq *rq, struct task_struct *p)
 159 {
 160 }
 161
 162 static inline void dequeue_pushable_task(struct rq *rq, struct task_struct *p)
 163 {
 164 }
 165
 166 static inline
 167 void inc_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
 168 {
 169 }
 170
 171 static inline
 172 void dec_rt_migration(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
 173 {
 174 }
 175
 176 #endif /* CONFIG_SMP */
 177
 178 static inline int on_rt_rq(struct sched_rt_entity *rt_se)
 179 {
 180         return !list_empty(&rt_se->run_list);
 181 }
 182
 183 #ifdef CONFIG_RT_GROUP_SCHED
 184
 185 static inline u64 sched_rt_runtime(struct rt_rq *rt_rq)
 186 {
 187         if (!rt_rq->tg)
 188                 return RUNTIME_INF;
 189
 190         return rt_rq->rt_runtime;
 191 }
 192
 193 static inline u64 sched_rt_period(struct rt_rq *rt_rq)
 194 {
 195         return ktime_to_ns(rt_rq->tg->rt_bandwidth.rt_period);
 196 }
 197
 198 typedef struct task_group *rt_rq_iter_t;
 199
 200 static inline struct task_group *next_task_group(struct task_group *tg)
 201 {
 202         do {
 203                 tg = list_entry_rcu(tg->list.next,
 204                         typeof(struct task_group), list);
 205         } while (&tg->list != &task_groups && task_group_is_autogroup(tg));
 206
 207         if (&tg->list == &task_groups)
 208                 tg = NULL;
 209
 210         return tg;
 211 }
 212
 213 #define for_each_rt_rq(rt_rq, iter, rq)                                 \
 214         for (iter = container_of(&task_groups, typeof(*iter), list);    \
 215                 (iter = next_task_group(iter)) &&                       \
 216                 (rt_rq = iter->rt_rq[cpu_of(rq)]);)
 217
 218 static inline void list_add_leaf_rt_rq(struct rt_rq *rt_rq)
 219 {
 220         list_add_rcu(&rt_rq->leaf_rt_rq_list,
 221                         &rq_of_rt_rq(rt_rq)->leaf_rt_rq_list);
 222 }
 223
 224 static inline void list_del_leaf_rt_rq(struct rt_rq *rt_rq)
 225 {
 226         list_del_rcu(&rt_rq->leaf_rt_rq_list);
 227 }
 228
 229 #define for_each_leaf_rt_rq(rt_rq, rq) \
 230         list_for_each_entry_rcu(rt_rq, &rq->leaf_rt_rq_list, leaf_rt_rq_list)
 231
 232 #define for_each_sched_rt_entity(rt_se) \
 233         for (; rt_se; rt_se = rt_se->parent)
 234
 235 static inline struct rt_rq *group_rt_rq(struct sched_rt_entity *rt_se)
 236 {
 237         return rt_se->my_q;
 238 }
 239
 240 static void enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head);
 241 static void dequeue_rt_entity(struct sched_rt_entity *rt_se);
 242
 243 static void sched_rt_rq_enqueue(struct rt_rq *rt_rq)
 244 {
 245         struct task_struct *curr = rq_of_rt_rq(rt_rq)->curr;
 246         struct sched_rt_entity *rt_se;
 247
 248         int cpu = cpu_of(rq_of_rt_rq(rt_rq));
 249
 250         rt_se = rt_rq->tg->rt_se[cpu];
 251
 252         if (rt_rq->rt_nr_running) {
 253                 if (rt_se && !on_rt_rq(rt_se))
 254                         enqueue_rt_entity(rt_se, false);
 255                 if (rt_rq->highest_prio.curr < curr->prio)
 256                         resched_task(curr);
 257         }
 258 }
 259
 260 static void sched_rt_rq_dequeue(struct rt_rq *rt_rq)
 261 {
 262         struct sched_rt_entity *rt_se;
 263         int cpu = cpu_of(rq_of_rt_rq(rt_rq));
 264
 265         rt_se = rt_rq->tg->rt_se[cpu];
 266
 267         if (rt_se && on_rt_rq(rt_se))
 268                 dequeue_rt_entity(rt_se);
 269 }
 270
 271 static inline int rt_rq_throttled(struct rt_rq *rt_rq)
 272 {
 273         return rt_rq->rt_throttled && !rt_rq->rt_nr_boosted;
 274 }
 275
 276 static int rt_se_boosted(struct sched_rt_entity *rt_se)
 277 {
 278         struct rt_rq *rt_rq = group_rt_rq(rt_se);
 279         struct task_struct *p;
 280
 281         if (rt_rq)
 282                 return !!rt_rq->rt_nr_boosted;
 283
 284         p = rt_task_of(rt_se);
 285         return p->prio != p->normal_prio;
 286 }
 287
 288 #ifdef CONFIG_SMP
 289 static inline const struct cpumask *sched_rt_period_mask(void)
 290 {
 291         return cpu_rq(smp_processor_id())->rd->span;
 292 }
 293 #else
 294 static inline const struct cpumask *sched_rt_period_mask(void)
 295 {
 296         return cpu_online_mask;
 297 }
 298 #endif
 299
 300 static inline
 301 struct rt_rq *sched_rt_period_rt_rq(struct rt_bandwidth *rt_b, int cpu)
 302 {
 303         return container_of(rt_b, struct task_group, rt_bandwidth)->rt_rq[cpu];
 304 }
 305
 306 static inline struct rt_bandwidth *sched_rt_bandwidth(struct rt_rq *rt_rq)
 307 {
 308         return &rt_rq->tg->rt_bandwidth;
 309 }
 310
 311 #else /* !CONFIG_RT_GROUP_SCHED */
 312
 313 static inline u64 sched_rt_runtime(struct rt_rq *rt_rq)
 314 {
 315         return rt_rq->rt_runtime;
 316 }
 317
 318 static inline u64 sched_rt_period(struct rt_rq *rt_rq)
 319 {
 320         return ktime_to_ns(def_rt_bandwidth.rt_period);
 321 }
 322
 323 typedef struct rt_rq *rt_rq_iter_t;
 324
 325 #define for_each_rt_rq(rt_rq, iter, rq) \
 326         for ((void) iter, rt_rq = &rq->rt; rt_rq; rt_rq = NULL)
 327
 328 static inline void list_add_leaf_rt_rq(struct rt_rq *rt_rq)
 329 {
 330 }
 331
 332 static inline void list_del_leaf_rt_rq(struct rt_rq *rt_rq)
 333 {
 334 }
 335
 336 #define for_each_leaf_rt_rq(rt_rq, rq) \
 337         for (rt_rq = &rq->rt; rt_rq; rt_rq = NULL)
 338
 339 #define for_each_sched_rt_entity(rt_se) \
 340         for (; rt_se; rt_se = NULL)
 341
 342 static inline struct rt_rq *group_rt_rq(struct sched_rt_entity *rt_se)
 343 {
 344         return NULL;
 345 }
 346
 347 static inline void sched_rt_rq_enqueue(struct rt_rq *rt_rq)
 348 {
 349         if (rt_rq->rt_nr_running)
 350                 resched_task(rq_of_rt_rq(rt_rq)->curr);
 351 }
 352
 353 static inline void sched_rt_rq_dequeue(struct rt_rq *rt_rq)
 354 {
 355 }
 356
 357 static inline int rt_rq_throttled(struct rt_rq *rt_rq)
 358 {
 359         return rt_rq->rt_throttled;
 360 }
 361
 362 static inline const struct cpumask *sched_rt_period_mask(void)
 363 {
 364         return cpu_online_mask;
 365 }
 366
 367 static inline
 368 struct rt_rq *sched_rt_period_rt_rq(struct rt_bandwidth *rt_b, int cpu)
 369 {
 370         return &cpu_rq(cpu)->rt;
 371 }
 372
 373 static inline struct rt_bandwidth *sched_rt_bandwidth(struct rt_rq *rt_rq)
 374 {
 375         return &def_rt_bandwidth;
 376 }
 377
 378 #endif /* CONFIG_RT_GROUP_SCHED */
 379
 380 #ifdef CONFIG_SMP
 381 /*
 382  * We ran out of runtime, see if we can borrow some from our neighbours.
 383  */
 384 static int do_balance_runtime(struct rt_rq *rt_rq)
 385 {
 386         struct rt_bandwidth *rt_b = sched_rt_bandwidth(rt_rq);
 387         struct root_domain *rd = cpu_rq(smp_processor_id())->rd;
 388         int i, weight, more = 0;
 389         u64 rt_period;
 390
 391         weight = cpumask_weight(rd->span);
 392
 393         raw_spin_lock(&rt_b->rt_runtime_lock);
 394         rt_period = ktime_to_ns(rt_b->rt_period);
 395         for_each_cpu(i, rd->span) {
 396                 struct rt_rq *iter = sched_rt_period_rt_rq(rt_b, i);
 397                 s64 diff;
 398
 399                 if (iter == rt_rq)
 400                         continue;
 401
 402                 raw_spin_lock(&iter->rt_runtime_lock);
 403                 /*
 404                  * Either all rqs have inf runtime and there's nothing to steal
 405                  * or __disable_runtime() below sets a specific rq to inf to
 406                  * indicate its been disabled and disalow stealing.
 407                  */
 408                 if (iter->rt_runtime == RUNTIME_INF)
 409                         goto next;
 410
 411                 /*
 412                  * From runqueues with spare time, take 1/n part of their
 413                  * spare time, but no more than our period.
 414                  */
 415                 diff = iter->rt_runtime - iter->rt_time;
 416                 if (diff > 0) {
 417                         diff = div_u64((u64)diff, weight);
 418                         if (rt_rq->rt_runtime + diff > rt_period)
 419                                 diff = rt_period - rt_rq->rt_runtime;
 420                         iter->rt_runtime -= diff;
 421                         rt_rq->rt_runtime += diff;
 422                         more = 1;
 423                         if (rt_rq->rt_runtime == rt_period) {
 424                                 raw_spin_unlock(&iter->rt_runtime_lock);
 425                                 break;
 426                         }
 427                 }
 428 next:
 429                 raw_spin_unlock(&iter->rt_runtime_lock);
 430         }
 431         raw_spin_unlock(&rt_b->rt_runtime_lock);
 432
 433         return more;
 434 }
 435
 436 /*
 437  * Ensure this RQ takes back all the runtime it lend to its neighbours.
 438  */
 439 static void __disable_runtime(struct rq *rq)
 440 {
 441         struct root_domain *rd = rq->rd;
 442         rt_rq_iter_t iter;
 443         struct rt_rq *rt_rq;
 444
 445         if (unlikely(!scheduler_running))
 446                 return;
 447
 448         for_each_rt_rq(rt_rq, iter, rq) {
 449                 struct rt_bandwidth *rt_b = sched_rt_bandwidth(rt_rq);
 450                 s64 want;
 451                 int i;
 452
 453                 raw_spin_lock(&rt_b->rt_runtime_lock);
 454                 raw_spin_lock(&rt_rq->rt_runtime_lock);
 455                 /*
 456                  * Either we're all inf and nobody needs to borrow, or we're
 457                  * already disabled and thus have nothing to do, or we have
 458                  * exactly the right amount of runtime to take out.
 459                  */
 460                 if (rt_rq->rt_runtime == RUNTIME_INF ||
 461                                 rt_rq->rt_runtime == rt_b->rt_runtime)
 462                         goto balanced;
 463                 raw_spin_unlock(&rt_rq->rt_runtime_lock);
 464
 465                 /*
 466                  * Calculate the difference between what we started out with
 467                  * and what we current have, that's the amount of runtime
 468                  * we lend and now have to reclaim.
 469                  */
 470                 want = rt_b->rt_runtime - rt_rq->rt_runtime;
 471
 472                 /*
 473                  * Greedy reclaim, take back as much as we can.
 474                  */
 475                 for_each_cpu(i, rd->span) {
 476                         struct rt_rq *iter = sched_rt_period_rt_rq(rt_b, i);
 477                         s64 diff;
 478
 479                         /*
 480                          * Can't reclaim from ourselves or disabled runqueues.
 481                          */
 482                         if (iter == rt_rq || iter->rt_runtime == RUNTIME_INF)
 483                                 continue;
 484
 485                         raw_spin_lock(&iter->rt_runtime_lock);
 486                         if (want > 0) {
 487                                 diff = min_t(s64, iter->rt_runtime, want);
 488                                 iter->rt_runtime -= diff;
 489                                 want -= diff;
 490                         } else {
 491                                 iter->rt_runtime -= want;
 492                                 want -= want;
 493                         }
 494                         raw_spin_unlock(&iter->rt_runtime_lock);
 495
 496                         if (!want)
 497                                 break;
 498                 }
 499
 500                 raw_spin_lock(&rt_rq->rt_runtime_lock);
 501                 /*
 502                  * We cannot be left wanting - that would mean some runtime
 503                  * leaked out of the system.
 504                  */
 505                 BUG_ON(want);
 506 balanced:
 507                 /*
 508                  * Disable all the borrow logic by pretending we have inf
 509                  * runtime - in which case borrowing doesn't make sense.
 510                  */
 511                 rt_rq->rt_runtime = RUNTIME_INF;
 512                 raw_spin_unlock(&rt_rq->rt_runtime_lock);
 513                 raw_spin_unlock(&rt_b->rt_runtime_lock);
 514         }
 515 }
 516
 517 static void disable_runtime(struct rq *rq)
 518 {
 519         unsigned long flags;
 520
 521         raw_spin_lock_irqsave(&rq->lock, flags);
 522         __disable_runtime(rq);
 523         raw_spin_unlock_irqrestore(&rq->lock, flags);
 524 }
 525
 526 static void __enable_runtime(struct rq *rq)
 527 {
 528         rt_rq_iter_t iter;
 529         struct rt_rq *rt_rq;
 530
 531         if (unlikely(!scheduler_running))
 532                 return;
 533
 534         /*
 535          * Reset each runqueue's bandwidth settings
 536          */
 537         for_each_rt_rq(rt_rq, iter, rq) {
 538                 struct rt_bandwidth *rt_b = sched_rt_bandwidth(rt_rq);
 539
 540                 raw_spin_lock(&rt_b->rt_runtime_lock);
 541                 raw_spin_lock(&rt_rq->rt_runtime_lock);
 542                 rt_rq->rt_runtime = rt_b->rt_runtime;
 543                 rt_rq->rt_time = 0;
 544                 rt_rq->rt_throttled = 0;
 545                 raw_spin_unlock(&rt_rq->rt_runtime_lock);
 546                 raw_spin_unlock(&rt_b->rt_runtime_lock);
 547         }
 548 }
 549
 550 static void enable_runtime(struct rq *rq)
 551 {
 552         unsigned long flags;
 553
 554         raw_spin_lock_irqsave(&rq->lock, flags);
 555         __enable_runtime(rq);
 556         raw_spin_unlock_irqrestore(&rq->lock, flags);
 557 }
 558
 559 static int balance_runtime(struct rt_rq *rt_rq)
 560 {
 561         int more = 0;
 562
 563         if (!sched_feat(RT_RUNTIME_SHARE))
 564                 return more;
 565
 566         if (rt_rq->rt_time > rt_rq->rt_runtime) {
 567                 raw_spin_unlock(&rt_rq->rt_runtime_lock);
 568                 more = do_balance_runtime(rt_rq);
 569                 raw_spin_lock(&rt_rq->rt_runtime_lock);
 570         }
 571
 572         return more;
 573 }
 574 #else /* !CONFIG_SMP */
 575 static inline int balance_runtime(struct rt_rq *rt_rq)
 576 {
 577         return 0;
 578 }
 579 #endif /* CONFIG_SMP */
 580
 581 static int do_sched_rt_period_timer(struct rt_bandwidth *rt_b, int overrun)
 582 {
 583         int i, idle = 1;
 584         const struct cpumask *span;
 585
 586         if (!rt_bandwidth_enabled() || rt_b->rt_runtime == RUNTIME_INF)
 587                 return 1;
 588
 589         span = sched_rt_period_mask();
 590         for_each_cpu(i, span) {
 591                 int enqueue = 0;
 592                 struct rt_rq *rt_rq = sched_rt_period_rt_rq(rt_b, i);
 593                 struct rq *rq = rq_of_rt_rq(rt_rq);
 594
 595                 raw_spin_lock(&rq->lock);
 596                 if (rt_rq->rt_time) {
 597                         u64 runtime;
 598
 599                         raw_spin_lock(&rt_rq->rt_runtime_lock);
 600                         if (rt_rq->rt_throttled)
 601                                 balance_runtime(rt_rq);
 602                         runtime = rt_rq->rt_runtime;
 603                         rt_rq->rt_time -= min(rt_rq->rt_time, overrun*runtime);
 604                         if (rt_rq->rt_throttled && rt_rq->rt_time < runtime) {
 605                                 rt_rq->rt_throttled = 0;
 606                                 enqueue = 1;
 607
 608                                 /*
 609                                  * Force a clock update if the CPU was idle,
 610                                  * lest wakeup -> unthrottle time accumulate.
 611                                  */
 612                                 if (rt_rq->rt_nr_running && rq->curr == rq->idle)
 613                                         rq->skip_clock_update = -1;
 614                         }
 615                         if (rt_rq->rt_time || rt_rq->rt_nr_running)
 616                                 idle = 0;
 617                         raw_spin_unlock(&rt_rq->rt_runtime_lock);
 618                 } else if (rt_rq->rt_nr_running) {
 619                         idle = 0;
 620                         if (!rt_rq_throttled(rt_rq))
 621                                 enqueue = 1;
 622                 }
 623
 624                 if (enqueue)
 625                         sched_rt_rq_enqueue(rt_rq);
 626                 raw_spin_unlock(&rq->lock);
 627         }
 628
 629         return idle;
 630 }
 631
 632 static inline int rt_se_prio(struct sched_rt_entity *rt_se)
 633 {
 634 #ifdef CONFIG_RT_GROUP_SCHED
 635         struct rt_rq *rt_rq = group_rt_rq(rt_se);
 636
 637         if (rt_rq)
 638                 return rt_rq->highest_prio.curr;
 639 #endif
 640
 641         return rt_task_of(rt_se)->prio;
 642 }
 643
 644 static int sched_rt_runtime_exceeded(struct rt_rq *rt_rq)
 645 {
 646         u64 runtime = sched_rt_runtime(rt_rq);
 647
 648         if (rt_rq->rt_throttled)
 649                 return rt_rq_throttled(rt_rq);
 650
 651         if (sched_rt_runtime(rt_rq) >= sched_rt_period(rt_rq))
 652                 return 0;
 653
 654         balance_runtime(rt_rq);
 655         runtime = sched_rt_runtime(rt_rq);
 656         if (runtime == RUNTIME_INF)
 657                 return 0;
 658
 659         if (rt_rq->rt_time > runtime) {
 660                 rt_rq->rt_throttled = 1;
 661                 printk_once(KERN_WARNING "sched: RT throttling activated\n");
 662                 if (rt_rq_throttled(rt_rq)) {
 663                         sched_rt_rq_dequeue(rt_rq);
 664                         return 1;
 665                 }
 666         }
 667
 668         return 0;
 669 }
 670
 671 /*
 672  * Update the current task's runtime statistics. Skip current tasks that
 673  * are not in our scheduling class.
 674  */
 675 static void update_curr_rt(struct rq *rq)
 676 {
 677         struct task_struct *curr = rq->curr;
 678         struct sched_rt_entity *rt_se = &curr->rt;
 679         struct rt_rq *rt_rq = rt_rq_of_se(rt_se);
 680         u64 delta_exec;
 681
 682         if (curr->sched_class != &rt_sched_class)
 683                 return;
 684
 685         delta_exec = rq->clock_task - curr->se.exec_start;
 686         if (unlikely((s64)delta_exec < 0))
 687                 delta_exec = 0;
 688
 689         schedstat_set(curr->se.statistics.exec_max, max(curr->se.statistics.exec_max, delta_exec));
 690
 691         curr->se.sum_exec_runtime += delta_exec;
 692         account_group_exec_runtime(curr, delta_exec);
 693
 694         curr->se.exec_start = rq->clock_task;
 695         cpuacct_charge(curr, delta_exec);
 696
 697         sched_rt_avg_update(rq, delta_exec);
 698
 699         if (!rt_bandwidth_enabled())
 700                 return;
 701
 702         for_each_sched_rt_entity(rt_se) {
 703                 rt_rq = rt_rq_of_se(rt_se);
 704
 705                 if (sched_rt_runtime(rt_rq) != RUNTIME_INF) {
 706                         raw_spin_lock(&rt_rq->rt_runtime_lock);
 707                         rt_rq->rt_time += delta_exec;
 708                         if (sched_rt_runtime_exceeded(rt_rq))
 709                                 resched_task(curr);
 710                         raw_spin_unlock(&rt_rq->rt_runtime_lock);
 711                 }
 712         }
 713 }
 714
 715 #if defined CONFIG_SMP
 716
 717 static void
 718 inc_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio)
 719 {
 720         struct rq *rq = rq_of_rt_rq(rt_rq);
 721
 722         if (rq->online && prio < prev_prio)
 723                 cpupri_set(&rq->rd->cpupri, rq->cpu, prio);
 724 }
 725
 726 static void
 727 dec_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio)
 728 {
 729         struct rq *rq = rq_of_rt_rq(rt_rq);
 730
 731         if (rq->online && rt_rq->highest_prio.curr != prev_prio)
 732                 cpupri_set(&rq->rd->cpupri, rq->cpu, rt_rq->highest_prio.curr);
 733 }
 734
 735 #else /* CONFIG_SMP */
 736
 737 static inline
 738 void inc_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio) {}
 739 static inline
 740 void dec_rt_prio_smp(struct rt_rq *rt_rq, int prio, int prev_prio) {}
 741
 742 #endif /* CONFIG_SMP */
 743
 744 #if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED
 745 static void
 746 inc_rt_prio(struct rt_rq *rt_rq, int prio)
 747 {
 748         int prev_prio = rt_rq->highest_prio.curr;
 749
 750         if (prio < prev_prio)
 751                 rt_rq->highest_prio.curr = prio;
 752
 753         inc_rt_prio_smp(rt_rq, prio, prev_prio);
 754 }
 755
 756 static void
 757 dec_rt_prio(struct rt_rq *rt_rq, int prio)
 758 {
 759         int prev_prio = rt_rq->highest_prio.curr;
 760
 761         if (rt_rq->rt_nr_running) {
 762
 763                 WARN_ON(prio < prev_prio);
 764
 765                 /*
 766                  * This may have been our highest task, and therefore
 767                  * we may have some recomputation to do
 768                  */
 769                 if (prio == prev_prio) {
 770                         struct rt_prio_array *array = &rt_rq->active;
 771
 772                         rt_rq->highest_prio.curr =
 773                                 sched_find_first_bit(array->bitmap);
 774                 }
 775
 776         } else
 777                 rt_rq->highest_prio.curr = MAX_RT_PRIO;
 778
 779         dec_rt_prio_smp(rt_rq, prio, prev_prio);
 780 }
 781
 782 #else
 783
 784 static inline void inc_rt_prio(struct rt_rq *rt_rq, int prio) {}
 785 static inline void dec_rt_prio(struct rt_rq *rt_rq, int prio) {}
 786
 787 #endif /* CONFIG_SMP || CONFIG_RT_GROUP_SCHED */
 788
 789 #ifdef CONFIG_RT_GROUP_SCHED
 790
 791 static void
 792 inc_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
 793 {
 794         if (rt_se_boosted(rt_se))
 795                 rt_rq->rt_nr_boosted++;
 796
 797         if (rt_rq->tg)
 798                 start_rt_bandwidth(&rt_rq->tg->rt_bandwidth);
 799 }
 800
 801 static void
 802 dec_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
 803 {
 804         if (rt_se_boosted(rt_se))
 805                 rt_rq->rt_nr_boosted--;
 806
 807         WARN_ON(!rt_rq->rt_nr_running && rt_rq->rt_nr_boosted);
 808 }
 809
 810 #else /* CONFIG_RT_GROUP_SCHED */
 811
 812 static void
 813 inc_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
 814 {
 815         start_rt_bandwidth(&def_rt_bandwidth);
 816 }
 817
 818 static inline
 819 void dec_rt_group(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq) {}
 820
 821 #endif /* CONFIG_RT_GROUP_SCHED */
 822
 823 static inline
 824 void inc_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
 825 {
 826         int prio = rt_se_prio(rt_se);
 827
 828         WARN_ON(!rt_prio(prio));
 829         rt_rq->rt_nr_running++;
 830
 831         inc_rt_prio(rt_rq, prio);
 832         inc_rt_migration(rt_se, rt_rq);
 833         inc_rt_group(rt_se, rt_rq);
 834 }
 835
 836 static inline
 837 void dec_rt_tasks(struct sched_rt_entity *rt_se, struct rt_rq *rt_rq)
 838 {
 839         WARN_ON(!rt_prio(rt_se_prio(rt_se)));
 840         WARN_ON(!rt_rq->rt_nr_running);
 841         rt_rq->rt_nr_running--;
 842
 843         dec_rt_prio(rt_rq, rt_se_prio(rt_se));
 844         dec_rt_migration(rt_se, rt_rq);
 845         dec_rt_group(rt_se, rt_rq);
 846 }
 847
 848 static void __enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head)
 849 {
 850         struct rt_rq *rt_rq = rt_rq_of_se(rt_se);
 851         struct rt_prio_array *array = &rt_rq->active;
 852         struct rt_rq *group_rq = group_rt_rq(rt_se);
 853         struct list_head *queue = array->queue + rt_se_prio(rt_se);
 854
 855         /*
 856          * Don't enqueue the group if its throttled, or when empty.
 857          * The latter is a consequence of the former when a child group
 858          * get throttled and the current group doesn't have any other
 859          * active members.
 860          */
 861         if (group_rq && (rt_rq_throttled(group_rq) || !group_rq->rt_nr_running))
 862                 return;
 863
 864         if (!rt_rq->rt_nr_running)
 865                 list_add_leaf_rt_rq(rt_rq);
 866
 867         if (head)
 868                 list_add(&rt_se->run_list, queue);
 869         else
 870                 list_add_tail(&rt_se->run_list, queue);
 871         __set_bit(rt_se_prio(rt_se), array->bitmap);
 872
 873         inc_rt_tasks(rt_se, rt_rq);
 874 }
 875
 876 static void __dequeue_rt_entity(struct sched_rt_entity *rt_se)
 877 {
 878         struct rt_rq *rt_rq = rt_rq_of_se(rt_se);
 879         struct rt_prio_array *array = &rt_rq->active;
 880
 881         list_del_init(&rt_se->run_list);
 882         if (list_empty(array->queue + rt_se_prio(rt_se)))
 883                 __clear_bit(rt_se_prio(rt_se), array->bitmap);
 884
 885         dec_rt_tasks(rt_se, rt_rq);
 886         if (!rt_rq->rt_nr_running)
 887                 list_del_leaf_rt_rq(rt_rq);
 888 }
 889
 890 /*
 891  * Because the prio of an upper entry depends on the lower
 892  * entries, we must remove entries top - down.
 893  */
 894 static void dequeue_rt_stack(struct sched_rt_entity *rt_se)
 895 {
 896         struct sched_rt_entity *back = NULL;
 897
 898         for_each_sched_rt_entity(rt_se) {
 899                 rt_se->back = back;
 900                 back = rt_se;
 901         }
 902
 903         for (rt_se = back; rt_se; rt_se = rt_se->back) {
 904                 if (on_rt_rq(rt_se))
 905                         __dequeue_rt_entity(rt_se);
 906         }
 907 }
 908
 909 static void enqueue_rt_entity(struct sched_rt_entity *rt_se, bool head)
 910 {
 911         dequeue_rt_stack(rt_se);
 912         for_each_sched_rt_entity(rt_se)
 913                 __enqueue_rt_entity(rt_se, head);
 914 }
 915
 916 static void dequeue_rt_entity(struct sched_rt_entity *rt_se)
 917 {
 918         dequeue_rt_stack(rt_se);
 919
 920         for_each_sched_rt_entity(rt_se) {
 921                 struct rt_rq *rt_rq = group_rt_rq(rt_se);
 922
 923                 if (rt_rq && rt_rq->rt_nr_running)
 924                         __enqueue_rt_entity(rt_se, false);
 925         }
 926 }
 927
 928 /*
 929  * Adding/removing a task to/from a priority array:
 930  */
 931 static void
 932 enqueue_task_rt(struct rq *rq, struct task_struct *p, int flags)
 933 {
 934         struct sched_rt_entity *rt_se = &p->rt;
 935
 936         if (flags & ENQUEUE_WAKEUP)
 937                 rt_se->timeout = 0;
 938
 939         enqueue_rt_entity(rt_se, flags & ENQUEUE_HEAD);
 940
 941         if (!task_current(rq, p) && p->rt.nr_cpus_allowed > 1)
 942                 enqueue_pushable_task(rq, p);
 943
 944         inc_nr_running(rq);
 945 }
 946
 947 static void dequeue_task_rt(struct rq *rq, struct task_struct *p, int flags)
 948 {
 949         struct sched_rt_entity *rt_se = &p->rt;
 950
 951         update_curr_rt(rq);
 952         dequeue_rt_entity(rt_se);
 953
 954         dequeue_pushable_task(rq, p);
 955
 956         dec_nr_running(rq);
 957 }
 958
 959 /*
 960  * Put task to the end of the run list without the overhead of dequeue
 961  * followed by enqueue.
 962  */
 963 static void
 964 requeue_rt_entity(struct rt_rq *rt_rq, struct sched_rt_entity *rt_se, int head)
 965 {
 966         if (on_rt_rq(rt_se)) {
 967                 struct rt_prio_array *array = &rt_rq->active;
 968                 struct list_head *queue = array->queue + rt_se_prio(rt_se);
 969
 970                 if (head)
 971                         list_move(&rt_se->run_list, queue);
 972                 else
 973                         list_move_tail(&rt_se->run_list, queue);
 974         }
 975 }
 976
 977 static void requeue_task_rt(struct rq *rq, struct task_struct *p, int head)
 978 {
 979         struct sched_rt_entity *rt_se = &p->rt;
 980         struct rt_rq *rt_rq;
 981
 982         for_each_sched_rt_entity(rt_se) {
 983                 rt_rq = rt_rq_of_se(rt_se);
 984                 requeue_rt_entity(rt_rq, rt_se, head);
 985         }
 986 }
 987
 988 static void yield_task_rt(struct rq *rq)
 989 {
 990         requeue_task_rt(rq, rq->curr, 0);
 991 }
 992
 993 #ifdef CONFIG_SMP
 994 static int find_lowest_rq(struct task_struct *task);
 995
 996 static int
 997 select_task_rq_rt(struct task_struct *p, int sd_flag, int flags)
 998 {
 999         struct task_struct *curr;
1000         struct rq *rq;
1001         int cpu;
1002
1003         cpu = task_cpu(p);
1004
1005         /* For anything but wake ups, just return the task_cpu */
1006         if (sd_flag != SD_BALANCE_WAKE && sd_flag != SD_BALANCE_FORK)
1007                 goto out;
1008
1009         rq = cpu_rq(cpu);
1010
1011         rcu_read_lock();
1012         curr = ACCESS_ONCE(rq->curr); /* unlocked access */
1013
1014         /*
1015          * If the current task on @p's runqueue is an RT task, then
1016          * try to see if we can wake this RT task up on another
1017          * runqueue. Otherwise simply start this RT task
1018          * on its current runqueue.
1019          *
1020          * We want to avoid overloading runqueues. If the woken
1021          * task is a higher priority, then it will stay on this CPU
1022          * and the lower prio task should be moved to another CPU.
1023          * Even though this will probably make the lower prio task
1024          * lose its cache, we do not want to bounce a higher task
1025          * around just because it gave up its CPU, perhaps for a
1026          * lock?
1027          *
1028          * For equal prio tasks, we just let the scheduler sort it out.
1029          *
1030          * Otherwise, just let it ride on the affined RQ and the
1031          * post-schedule router will push the preempted task away
1032          *
1033          * This test is optimistic, if we get it wrong the load-balancer
1034          * will have to sort it out.
1035          */
1036         if (curr && unlikely(rt_task(curr)) &&
1037             (curr->rt.nr_cpus_allowed < 2 ||
1038              curr->prio <= p->prio) &&
1039             (p->rt.nr_cpus_allowed > 1)) {
1040                 int target = find_lowest_rq(p);
1041
1042                 if (target != -1)
1043                         cpu = target;
1044         }
1045         rcu_read_unlock();
1046
1047 out:
1048         return cpu;
1049 }
1050
1051 static void check_preempt_equal_prio(struct rq *rq, struct task_struct *p)
1052 {
1053         if (rq->curr->rt.nr_cpus_allowed == 1)
1054                 return;
1055
1056         if (p->rt.nr_cpus_allowed != 1
1057             && cpupri_find(&rq->rd->cpupri, p, NULL))
1058                 return;
1059
1060         if (!cpupri_find(&rq->rd->cpupri, rq->curr, NULL))
1061                 return;
1062
1063         /*
1064          * There appears to be other cpus that can accept
1065          * current and none to run 'p', so lets reschedule
1066          * to try and push current away:
1067          */
1068         requeue_task_rt(rq, p, 1);
1069         resched_task(rq->curr);
1070 }
1071
1072 #endif /* CONFIG_SMP */
1073
1074 /*
1075  * Preempt the current task with a newly woken task if needed:
1076  */
1077 static void check_preempt_curr_rt(struct rq *rq, struct task_struct *p, int flags)
1078 {
1079         if (p->prio < rq->curr->prio) {
1080                 resched_task(rq->curr);
1081                 return;
1082         }
1083
1084 #ifdef CONFIG_SMP
1085         /*
1086          * If:
1087          *
1088          * - the newly woken task is of equal priority to the current task
1089          * - the newly woken task is non-migratable while current is migratable
1090          * - current will be preempted on the next reschedule
1091          *
1092          * we should check to see if current can readily move to a different
1093          * cpu.  If so, we will reschedule to allow the push logic to try
1094          * to move current somewhere else, making room for our non-migratable
1095          * task.
1096          */
1097         if (p->prio == rq->curr->prio && !test_tsk_need_resched(rq->curr))
1098                 check_preempt_equal_prio(rq, p);
1099 #endif
1100 }
1101
1102 static struct sched_rt_entity *pick_next_rt_entity(struct rq *rq,
1103                                                    struct rt_rq *rt_rq)
1104 {
1105         struct rt_prio_array *array = &rt_rq->active;
1106         struct sched_rt_entity *next = NULL;
1107         struct list_head *queue;
1108         int idx;
1109
1110         idx = sched_find_first_bit(array->bitmap);
1111         BUG_ON(idx >= MAX_RT_PRIO);
1112
1113         queue = array->queue + idx;
1114         next = list_entry(queue->next, struct sched_rt_entity, run_list);
1115
1116         return next;
1117 }
1118
1119 static struct task_struct *_pick_next_task_rt(struct rq *rq)
1120 {
1121         struct sched_rt_entity *rt_se;
1122         struct task_struct *p;
1123         struct rt_rq *rt_rq;
1124
1125         rt_rq = &rq->rt;
1126
1127         if (!rt_rq->rt_nr_running)
1128                 return NULL;
1129
1130         if (rt_rq_throttled(rt_rq))
1131                 return NULL;
1132
1133         do {
1134                 rt_se = pick_next_rt_entity(rq, rt_rq);
1135                 BUG_ON(!rt_se);
1136                 rt_rq = group_rt_rq(rt_se);
1137         } while (rt_rq);
1138
1139         p = rt_task_of(rt_se);
1140         p->se.exec_start = rq->clock_task;
1141
1142         return p;
1143 }
1144
1145 static struct task_struct *pick_next_task_rt(struct rq *rq)
1146 {
1147         struct task_struct *p = _pick_next_task_rt(rq);
1148
1149         /* The running task is never eligible for pushing */
1150         if (p)
1151                 dequeue_pushable_task(rq, p);
1152
1153 #ifdef CONFIG_SMP
1154         /*
1155          * We detect this state here so that we can avoid taking the RQ
1156          * lock again later if there is no need to push
1157          */
1158         rq->post_schedule = has_pushable_tasks(rq);
1159 #endif
1160
1161         return p;
1162 }
1163
1164 static void put_prev_task_rt(struct rq *rq, struct task_struct *p)
1165 {
1166         update_curr_rt(rq);
1167
1168         /*
1169          * The previous task needs to be made eligible for pushing
1170          * if it is still active
1171          */
1172         if (on_rt_rq(&p->rt) && p->rt.nr_cpus_allowed > 1)
1173                 enqueue_pushable_task(rq, p);
1174 }
1175
1176 #ifdef CONFIG_SMP
1177
1178 /* Only try algorithms three times */
1179 #define RT_MAX_TRIES 3
1180
1181 static void deactivate_task(struct rq *rq, struct task_struct *p, int sleep);
1182
1183 static int pick_rt_task(struct rq *rq, struct task_struct *p, int cpu)
1184 {
1185         if (!task_running(rq, p) &&
1186             (cpu < 0 || cpumask_test_cpu(cpu, tsk_cpus_allowed(p))) &&
1187             (p->rt.nr_cpus_allowed > 1))
1188                 return 1;
1189         return 0;
1190 }
1191
1192 /* Return the second highest RT task, NULL otherwise */
1193 static struct task_struct *pick_next_highest_task_rt(struct rq *rq, int cpu)
1194 {
1195         struct task_struct *next = NULL;
1196         struct sched_rt_entity *rt_se;
1197         struct rt_prio_array *array;
1198         struct rt_rq *rt_rq;
1199         int idx;
1200
1201         for_each_leaf_rt_rq(rt_rq, rq) {
1202                 array = &rt_rq->active;
1203                 idx = sched_find_first_bit(array->bitmap);
1204 next_idx:
1205                 if (idx >= MAX_RT_PRIO)
1206                         continue;
1207                 if (next && next->prio < idx)
1208                         continue;
1209                 list_for_each_entry(rt_se, array->queue + idx, run_list) {
1210                         struct task_struct *p;
1211
1212                         if (!rt_entity_is_task(rt_se))
1213                                 continue;
1214
1215                         p = rt_task_of(rt_se);
1216                         if (pick_rt_task(rq, p, cpu)) {
1217                                 next = p;
1218                                 break;
1219                         }
1220                 }
1221                 if (!next) {
1222                         idx = find_next_bit(array->bitmap, MAX_RT_PRIO, idx+1);
1223                         goto next_idx;
1224                 }
1225         }
1226
1227         return next;
1228 }
1229
1230 static DEFINE_PER_CPU(cpumask_var_t, local_cpu_mask);
1231
1232 static int find_lowest_rq(struct task_struct *task)
1233 {
1234         struct sched_domain *sd;
1235         struct cpumask *lowest_mask = __get_cpu_var(local_cpu_mask);
1236         int this_cpu = smp_processor_id();
1237         int cpu      = task_cpu(task);
1238
1239         /* Make sure the mask is initialized first */
1240         if (unlikely(!lowest_mask))
1241                 return -1;
1242
1243         if (task->rt.nr_cpus_allowed == 1)
1244                 return -1; /* No other targets possible */
1245
1246         if (!cpupri_find(&task_rq(task)->rd->cpupri, task, lowest_mask))
1247                 return -1; /* No targets found */
1248
1249         /*
1250          * At this point we have built a mask of cpus representing the
1251          * lowest priority tasks in the system.  Now we want to elect
1252          * the best one based on our affinity and topology.
1253          *
1254          * We prioritize the last cpu that the task executed on since
1255          * it is most likely cache-hot in that location.
1256          */
1257         if (cpumask_test_cpu(cpu, lowest_mask))
1258                 return cpu;
1259
1260         /*
1261          * Otherwise, we consult the sched_domains span maps to figure
1262          * out which cpu is logically closest to our hot cache data.
1263          */
1264         if (!cpumask_test_cpu(this_cpu, lowest_mask))
1265                 this_cpu = -1; /* Skip this_cpu opt if not among lowest */
1266
1267         rcu_read_lock();
1268         for_each_domain(cpu, sd) {
1269                 if (sd->flags & SD_WAKE_AFFINE) {
1270                         int best_cpu;
1271
1272                         /*
1273                          * "this_cpu" is cheaper to preempt than a
1274                          * remote processor.
1275                          */
1276                         if (this_cpu != -1 &&
1277                             cpumask_test_cpu(this_cpu, sched_domain_span(sd))) {
1278                                 rcu_read_unlock();
1279                                 return this_cpu;
1280                         }
1281
1282                         best_cpu = cpumask_first_and(lowest_mask,
1283                                                      sched_domain_span(sd));
1284                         if (best_cpu < nr_cpu_ids) {
1285                                 rcu_read_unlock();
1286                                 return best_cpu;
1287                         }
1288                 }
1289         }
1290         rcu_read_unlock();
1291
1292         /*
1293          * And finally, if there were no matches within the domains
1294          * just give the caller *something* to work with from the compatible
1295          * locations.
1296          */
1297         if (this_cpu != -1)
1298                 return this_cpu;
1299
1300         cpu = cpumask_any(lowest_mask);
1301         if (cpu < nr_cpu_ids)
1302                 return cpu;
1303         return -1;
1304 }
1305
1306 /* Will lock the rq it finds */
1307 static struct rq *find_lock_lowest_rq(struct task_struct *task, struct rq *rq)
1308 {
1309         struct rq *lowest_rq = NULL;
1310         int tries;
1311         int cpu;
1312
1313         for (tries = 0; tries < RT_MAX_TRIES; tries++) {
1314                 cpu = find_lowest_rq(task);
1315
1316                 if ((cpu == -1) || (cpu == rq->cpu))
1317                         break;
1318
1319                 lowest_rq = cpu_rq(cpu);
1320
1321                 /* if the prio of this runqueue changed, try again */
1322                 if (double_lock_balance(rq, lowest_rq)) {
1323                         /*
1324                          * We had to unlock the run queue. In
1325                          * the mean time, task could have
1326                          * migrated already or had its affinity changed.
1327                          * Also make sure that it wasn't scheduled on its rq.
1328                          */
1329                         if (unlikely(task_rq(task) != rq ||
1330                                      !cpumask_test_cpu(lowest_rq->cpu,
1331                                                        tsk_cpus_allowed(task)) ||
1332                                      task_running(rq, task) ||
1333                                      !task->on_rq)) {
1334
1335                                 raw_spin_unlock(&lowest_rq->lock);
1336                                 lowest_rq = NULL;
1337                                 break;
1338                         }
1339                 }
1340
1341                 /* If this rq is still suitable use it. */
1342                 if (lowest_rq->rt.highest_prio.curr > task->prio)
1343                         break;
1344
1345                 /* try again */
1346                 double_unlock_balance(rq, lowest_rq);
1347                 lowest_rq = NULL;
1348         }
1349
1350         return lowest_rq;
1351 }
1352
1353 static struct task_struct *pick_next_pushable_task(struct rq *rq)
1354 {
1355         struct task_struct *p;
1356
1357         if (!has_pushable_tasks(rq))
1358                 return NULL;
1359
1360         p = plist_first_entry(&rq->rt.pushable_tasks,
1361                               struct task_struct, pushable_tasks);
1362
1363         BUG_ON(rq->cpu != task_cpu(p));
1364         BUG_ON(task_current(rq, p));
1365         BUG_ON(p->rt.nr_cpus_allowed <= 1);
1366
1367         BUG_ON(!p->on_rq);
1368         BUG_ON(!rt_task(p));
1369
1370         return p;
1371 }
1372
1373 /*
1374  * If the current CPU has more than one RT task, see if the non
1375  * running task can migrate over to a CPU that is running a task
1376  * of lesser priority.
1377  */
1378 static int push_rt_task(struct rq *rq)
1379 {
1380         struct task_struct *next_task;
1381         struct rq *lowest_rq;
1382         int ret = 0;
1383
1384         if (!rq->rt.overloaded)
1385                 return 0;
1386
1387         next_task = pick_next_pushable_task(rq);
1388         if (!next_task)
1389                 return 0;
1390
1391 retry:
1392         if (unlikely(next_task == rq->curr)) {
1393                 WARN_ON(1);
1394                 return 0;
1395         }
1396
1397         /*
1398          * It's possible that the next_task slipped in of
1399          * higher priority than current. If that's the case
1400          * just reschedule current.
1401          */
1402         if (unlikely(next_task->prio < rq->curr->prio)) {
1403                 resched_task(rq->curr);
1404                 return 0;
1405         }
1406
1407         /* We might release rq lock */
1408         get_task_struct(next_task);
1409
1410         /* find_lock_lowest_rq locks the rq if found */
1411         lowest_rq = find_lock_lowest_rq(next_task, rq);
1412         if (!lowest_rq) {
1413                 struct task_struct *task;
1414                 /*
1415                  * find_lock_lowest_rq releases rq->lock
1416                  * so it is possible that next_task has migrated.
1417                  *
1418                  * We need to make sure that the task is still on the same
1419                  * run-queue and is also still the next task eligible for
1420                  * pushing.
1421                  */
1422                 task = pick_next_pushable_task(rq);
1423                 if (task_cpu(next_task) == rq->cpu && task == next_task) {
1424                         /*
1425                          * The task hasn't migrated, and is still the next
1426                          * eligible task, but we failed to find a run-queue
1427                          * to push it to.  Do not retry in this case, since
1428                          * other cpus will pull from us when ready.
1429                          */
1430                         goto out;
1431                 }
1432
1433                 if (!task)
1434                         /* No more tasks, just exit */
1435                         goto out;
1436
1437                 /*
1438                  * Something has shifted, try again.
1439                  */
1440                 put_task_struct(next_task);
1441                 next_task = task;
1442                 goto retry;
1443         }
1444
1445         deactivate_task(rq, next_task, 0);
1446         set_task_cpu(next_task, lowest_rq->cpu);
1447         activate_task(lowest_rq, next_task, 0);
1448         ret = 1;
1449
1450         resched_task(lowest_rq->curr);
1451
1452         double_unlock_balance(rq, lowest_rq);
1453
1454 out:
1455         put_task_struct(next_task);
1456
1457         return ret;
1458 }
1459
1460 static void push_rt_tasks(struct rq *rq)
1461 {
1462         /* push_rt_task will return true if it moved an RT */
1463         while (push_rt_task(rq))
1464                 ;
1465 }
1466
1467 static int pull_rt_task(struct rq *this_rq)
1468 {
1469         int this_cpu = this_rq->cpu, ret = 0, cpu;
1470         struct task_struct *p;
1471         struct rq *src_rq;
1472
1473         if (likely(!rt_overloaded(this_rq)))
1474                 return 0;
1475
1476         for_each_cpu(cpu, this_rq->rd->rto_mask) {
1477                 if (this_cpu == cpu)
1478                         continue;
1479
1480                 src_rq = cpu_rq(cpu);
1481
1482                 /*
1483                  * Don't bother taking the src_rq->lock if the next highest
1484                  * task is known to be lower-priority than our current task.
1485                  * This may look racy, but if this value is about to go
1486                  * logically higher, the src_rq will push this task away.
1487                  * And if its going logically lower, we do not care
1488                  */
1489                 if (src_rq->rt.highest_prio.next >=
1490                     this_rq->rt.highest_prio.curr)
1491                         continue;
1492
1493                 /*
1494                  * We can potentially drop this_rq's lock in
1495                  * double_lock_balance, and another CPU could
1496                  * alter this_rq
1497                  */
1498                 double_lock_balance(this_rq, src_rq);
1499
1500                 /*
1501                  * Are there still pullable RT tasks?
1502                  */
1503                 if (src_rq->rt.rt_nr_running <= 1)
1504                         goto skip;
1505
1506                 p = pick_next_highest_task_rt(src_rq, this_cpu);
1507
1508                 /*
1509                  * Do we have an RT task that preempts
1510                  * the to-be-scheduled task?
1511                  */
1512                 if (p && (p->prio < this_rq->rt.highest_prio.curr)) {
1513                         WARN_ON(p == src_rq->curr);
1514                         WARN_ON(!p->on_rq);
1515
1516                         /*
1517                          * There's a chance that p is higher in priority
1518                          * than what's currently running on its cpu.
1519                          * This is just that p is wakeing up and hasn't
1520                          * had a chance to schedule. We only pull
1521                          * p if it is lower in priority than the
1522                          * current task on the run queue
1523                          */
1524                         if (p->prio < src_rq->curr->prio)
1525                                 goto skip;
1526
1527                         ret = 1;
1528
1529                         deactivate_task(src_rq, p, 0);
1530                         set_task_cpu(p, this_cpu);
1531                         activate_task(this_rq, p, 0);
1532                         /*
1533                          * We continue with the search, just in
1534                          * case there's an even higher prio task
1535                          * in another runqueue. (low likelihood
1536                          * but possible)
1537                          */
1538                 }
1539 skip:
1540                 double_unlock_balance(this_rq, src_rq);
1541         }
1542
1543         return ret;
1544 }
1545
1546 static void pre_schedule_rt(struct rq *rq, struct task_struct *prev)
1547 {
1548         /* Try to pull RT tasks here if we lower this rq's prio */
1549         if (rq->rt.highest_prio.curr > prev->prio)
1550                 pull_rt_task(rq);
1551 }
1552
1553 static void post_schedule_rt(struct rq *rq)
1554 {
1555         push_rt_tasks(rq);
1556 }
1557
1558 /*
1559  * If we are not running and we are not going to reschedule soon, we should
1560  * try to push tasks away now
1561  */
1562 static void task_woken_rt(struct rq *rq, struct task_struct *p)
1563 {
1564         if (!task_running(rq, p) &&
1565             !test_tsk_need_resched(rq->curr) &&
1566             has_pushable_tasks(rq) &&
1567             p->rt.nr_cpus_allowed > 1 &&
1568             rt_task(rq->curr) &&
1569             (rq->curr->rt.nr_cpus_allowed < 2 ||
1570              rq->curr->prio <= p->prio))
1571                 push_rt_tasks(rq);
1572 }
1573
1574 static void set_cpus_allowed_rt(struct task_struct *p,
1575                                 const struct cpumask *new_mask)
1576 {
1577         int weight = cpumask_weight(new_mask);
1578
1579         BUG_ON(!rt_task(p));
1580
1581         /*
1582          * Update the migration status of the RQ if we have an RT task
1583          * which is running AND changing its weight value.
1584          */
1585         if (p->on_rq && (weight != p->rt.nr_cpus_allowed)) {
1586                 struct rq *rq = task_rq(p);
1587
1588                 if (!task_current(rq, p)) {
1589                         /*
1590                          * Make sure we dequeue this task from the pushable list
1591                          * before going further.  It will either remain off of
1592                          * the list because we are no longer pushable, or it
1593                          * will be requeued.
1594                          */
1595                         if (p->rt.nr_cpus_allowed > 1)
1596                                 dequeue_pushable_task(rq, p);
1597
1598                         /*
1599                          * Requeue if our weight is changing and still > 1
1600                          */
1601                         if (weight > 1)
1602                                 enqueue_pushable_task(rq, p);
1603
1604                 }
1605
1606                 if ((p->rt.nr_cpus_allowed <= 1) && (weight > 1)) {
1607                         rq->rt.rt_nr_migratory++;
1608                 } else if ((p->rt.nr_cpus_allowed > 1) && (weight <= 1)) {
1609                         BUG_ON(!rq->rt.rt_nr_migratory);
1610                         rq->rt.rt_nr_migratory--;
1611                 }
1612
1613                 update_rt_migration(&rq->rt);
1614         }
1615 }
1616
1617 /* Assumes rq->lock is held */
1618 static void rq_online_rt(struct rq *rq)
1619 {
1620         if (rq->rt.overloaded)
1621                 rt_set_overload(rq);
1622
1623         __enable_runtime(rq);
1624
1625         cpupri_set(&rq->rd->cpupri, rq->cpu, rq->rt.highest_prio.curr);
1626 }
1627
1628 /* Assumes rq->lock is held */
1629 static void rq_offline_rt(struct rq *rq)
1630 {
1631         if (rq->rt.overloaded)
1632                 rt_clear_overload(rq);
1633
1634         __disable_runtime(rq);
1635
1636         cpupri_set(&rq->rd->cpupri, rq->cpu, CPUPRI_INVALID);
1637 }
1638
1639 /*
1640  * When switch from the rt queue, we bring ourselves to a position
1641  * that we might want to pull RT tasks from other runqueues.
1642  */
1643 static void switched_from_rt(struct rq *rq, struct task_struct *p)
1644 {
1645         /*
1646          * If there are other RT tasks then we will reschedule
1647          * and the scheduling of the other RT tasks will handle
1648          * the balancing. But if we are the last RT task
1649          * we may need to handle the pulling of RT tasks
1650          * now.
1651          */
1652         if (p->on_rq && !rq->rt.rt_nr_running)
1653                 pull_rt_task(rq);
1654 }
1655
1656 static inline void init_sched_rt_class(void)
1657 {
1658         unsigned int i;
1659
1660         for_each_possible_cpu(i)
1661                 zalloc_cpumask_var_node(&per_cpu(local_cpu_mask, i),
1662                                         GFP_KERNEL, cpu_to_node(i));
1663 }
1664 #endif /* CONFIG_SMP */
1665
1666 /*
1667  * When switching a task to RT, we may overload the runqueue
1668  * with RT tasks. In this case we try to push them off to
1669  * other runqueues.
1670  */
1671 static void switched_to_rt(struct rq *rq, struct task_struct *p)
1672 {
1673         int check_resched = 1;
1674
1675         /*
1676          * If we are already running, then there's nothing
1677          * that needs to be done. But if we are not running
1678          * we may need to preempt the current running task.
1679          * If that current running task is also an RT task
1680          * then see if we can move to another run queue.
1681          */
1682         if (p->on_rq && rq->curr != p) {
1683 #ifdef CONFIG_SMP
1684                 if (rq->rt.overloaded && push_rt_task(rq) &&
1685                     /* Don't resched if we changed runqueues */
1686                     rq != task_rq(p))
1687                         check_resched = 0;
1688 #endif /* CONFIG_SMP */
1689                 if (check_resched && p->prio < rq->curr->prio)
1690                         resched_task(rq->curr);
1691         }
1692 }
1693
1694 /*
1695  * Priority of the task has changed. This may cause
1696  * us to initiate a push or pull.
1697  */
1698 static void
1699 prio_changed_rt(struct rq *rq, struct task_struct *p, int oldprio)
1700 {
1701         if (!p->on_rq)
1702                 return;
1703
1704         if (rq->curr == p) {
1705 #ifdef CONFIG_SMP
1706                 /*
1707                  * If our priority decreases while running, we
1708                  * may need to pull tasks to this runqueue.
1709                  */
1710                 if (oldprio < p->prio)
1711                         pull_rt_task(rq);
1712                 /*
1713                  * If there's a higher priority task waiting to run
1714                  * then reschedule. Note, the above pull_rt_task
1715                  * can release the rq lock and p could migrate.
1716                  * Only reschedule if p is still on the same runqueue.
1717                  */
1718                 if (p->prio > rq->rt.highest_prio.curr && rq->curr == p)
1719                         resched_task(p);
1720 #else
1721                 /* For UP simply resched on drop of prio */
1722                 if (oldprio < p->prio)
1723                         resched_task(p);
1724 #endif /* CONFIG_SMP */
1725         } else {
1726                 /*
1727                  * This task is not running, but if it is
1728                  * greater than the current running task
1729                  * then reschedule.
1730                  */
1731                 if (p->prio < rq->curr->prio)
1732                         resched_task(rq->curr);
1733         }
1734 }
1735
1736 static void watchdog(struct rq *rq, struct task_struct *p)
1737 {
1738         unsigned long soft, hard;
1739
1740         /* max may change after cur was read, this will be fixed next tick */
1741         soft = task_rlimit(p, RLIMIT_RTTIME);
1742         hard = task_rlimit_max(p, RLIMIT_RTTIME);
1743
1744         if (soft != RLIM_INFINITY) {
1745                 unsigned long next;
1746
1747                 p->rt.timeout++;
1748                 next = DIV_ROUND_UP(min(soft, hard), USEC_PER_SEC/HZ);
1749                 if (p->rt.timeout > next)
1750                         p->cputime_expires.sched_exp = p->se.sum_exec_runtime;
1751         }
1752 }
1753
1754 static void task_tick_rt(struct rq *rq, struct task_struct *p, int queued)
1755 {
1756         update_curr_rt(rq);
1757
1758         watchdog(rq, p);
1759
1760         /*
1761          * RR tasks need a special form of timeslice management.
1762          * FIFO tasks have no timeslices.
1763          */
1764         if (p->policy != SCHED_RR)
1765                 return;
1766
1767         if (--p->rt.time_slice)
1768                 return;
1769
1770         p->rt.time_slice = DEF_TIMESLICE;
1771
1772         /*
1773          * Requeue to the end of queue if we are not the only element
1774          * on the queue:
1775          */
1776         if (p->rt.run_list.prev != p->rt.run_list.next) {
1777                 requeue_task_rt(rq, p, 0);
1778                 set_tsk_need_resched(p);
1779         }
1780 }
1781
1782 static void set_curr_task_rt(struct rq *rq)
1783 {
1784         struct task_struct *p = rq->curr;
1785
1786         p->se.exec_start = rq->clock_task;
1787
1788         /* The running task is never eligible for pushing */
1789         dequeue_pushable_task(rq, p);
1790 }
1791
1792 static unsigned int get_rr_interval_rt(struct rq *rq, struct task_struct *task)
1793 {
1794         /*
1795          * Time slice is 0 for SCHED_FIFO tasks
1796          */
1797         if (task->policy == SCHED_RR)
1798                 return DEF_TIMESLICE;
1799         else
1800                 return 0;
1801 }
1802
1803 static const struct sched_class rt_sched_class = {
1804         .next                   = &fair_sched_class,
1805         .enqueue_task           = enqueue_task_rt,
1806         .dequeue_task           = dequeue_task_rt,
1807         .yield_task             = yield_task_rt,
1808
1809         .check_preempt_curr     = check_preempt_curr_rt,
1810
1811         .pick_next_task         = pick_next_task_rt,
1812         .put_prev_task          = put_prev_task_rt,
1813
1814 #ifdef CONFIG_SMP
1815         .select_task_rq         = select_task_rq_rt,
1816
1817         .set_cpus_allowed       = set_cpus_allowed_rt,
1818         .rq_online              = rq_online_rt,
1819         .rq_offline             = rq_offline_rt,
1820         .pre_schedule           = pre_schedule_rt,
1821         .post_schedule          = post_schedule_rt,
1822         .task_woken             = task_woken_rt,
1823         .switched_from          = switched_from_rt,
1824 #endif
1825
1826         .set_curr_task          = set_curr_task_rt,
1827         .task_tick              = task_tick_rt,
1828
1829         .get_rr_interval        = get_rr_interval_rt,
1830
1831         .prio_changed           = prio_changed_rt,
1832         .switched_to            = switched_to_rt,
1833 };
1834
1835 #ifdef CONFIG_SCHED_DEBUG
1836 extern void print_rt_rq(struct seq_file *m, int cpu, struct rt_rq *rt_rq);
1837
1838 static void print_rt_stats(struct seq_file *m, int cpu)
1839 {
1840         rt_rq_iter_t iter;
1841         struct rt_rq *rt_rq;
1842
1843         rcu_read_lock();
1844         for_each_rt_rq(rt_rq, iter, cpu_rq(cpu))
1845                 print_rt_rq(m, cpu, rt_rq);
1846         rcu_read_unlock();
1847 }
1848 #endif /* CONFIG_SCHED_DEBUG */