sched.h 57.6 KB
Newer Older
1
/* SPDX-License-Identifier: GPL-2.0 */
2 3 4
/*
 * Scheduler internal types and methods:
 */
5
#include <linux/sched.h>
6

7
#include <linux/sched/autogroup.h>
8
#include <linux/sched/clock.h>
9
#include <linux/sched/coredump.h>
10
#include <linux/sched/cpufreq.h>
11 12
#include <linux/sched/cputime.h>
#include <linux/sched/deadline.h>
13
#include <linux/sched/debug.h>
14
#include <linux/sched/hotplug.h>
15 16 17 18 19 20 21 22 23 24 25 26 27
#include <linux/sched/idle.h>
#include <linux/sched/init.h>
#include <linux/sched/isolation.h>
#include <linux/sched/jobctl.h>
#include <linux/sched/loadavg.h>
#include <linux/sched/mm.h>
#include <linux/sched/nohz.h>
#include <linux/sched/numa_balancing.h>
#include <linux/sched/prio.h>
#include <linux/sched/rt.h>
#include <linux/sched/signal.h>
#include <linux/sched/stat.h>
#include <linux/sched/sysctl.h>
28
#include <linux/sched/task.h>
29
#include <linux/sched/task_stack.h>
30 31 32 33 34 35
#include <linux/sched/topology.h>
#include <linux/sched/user.h>
#include <linux/sched/wake_q.h>
#include <linux/sched/xacct.h>

#include <uapi/linux/sched/types.h>
36

37
#include <linux/binfmts.h>
38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
#include <linux/blkdev.h>
#include <linux/compat.h>
#include <linux/context_tracking.h>
#include <linux/cpufreq.h>
#include <linux/cpuidle.h>
#include <linux/cpuset.h>
#include <linux/ctype.h>
#include <linux/debugfs.h>
#include <linux/delayacct.h>
#include <linux/init_task.h>
#include <linux/kprobes.h>
#include <linux/kthread.h>
#include <linux/membarrier.h>
#include <linux/migrate.h>
#include <linux/mmu_context.h>
#include <linux/nmi.h>
#include <linux/proc_fs.h>
#include <linux/prefetch.h>
#include <linux/profile.h>
#include <linux/rcupdate_wait.h>
#include <linux/security.h>
#include <linux/stackprotector.h>
60
#include <linux/stop_machine.h>
61 62 63 64 65 66 67
#include <linux/suspend.h>
#include <linux/swait.h>
#include <linux/syscalls.h>
#include <linux/task_work.h>
#include <linux/tsacct_kern.h>

#include <asm/tlb.h>
68

69
#ifdef CONFIG_PARAVIRT
70
# include <asm/paravirt.h>
71 72
#endif

73
#include "cpupri.h"
74
#include "cpudeadline.h"
75

76
#ifdef CONFIG_SCHED_DEBUG
77
# define SCHED_WARN_ON(x)	WARN_ONCE(x, #x)
78
#else
79
# define SCHED_WARN_ON(x)	({ (void)(x), 0; })
80 81
#endif

82
struct rq;
83
struct cpuidle_state;
84

85 86
/* task_struct::on_rq states: */
#define TASK_ON_RQ_QUEUED	1
87
#define TASK_ON_RQ_MIGRATING	2
88

89 90
extern __read_mostly int scheduler_running;

91 92 93
extern unsigned long calc_load_update;
extern atomic_long_t calc_load_tasks;

94
extern void calc_global_load_tick(struct rq *this_rq);
95
extern long calc_load_fold_active(struct rq *this_rq, long adjust);
96 97

#ifdef CONFIG_SMP
98
extern void cpu_load_update_active(struct rq *this_rq);
99
#else
100
static inline void cpu_load_update_active(struct rq *this_rq) { }
101
#endif
102

103 104 105 106 107
/*
 * Helpers for converting nanosecond timing to jiffy resolution
 */
#define NS_TO_JIFFIES(TIME)	((unsigned long)(TIME) / (NSEC_PER_SEC / HZ))

108 109 110 111 112 113 114 115
/*
 * Increase resolution of nice-level calculations for 64-bit architectures.
 * The extra resolution improves shares distribution and load balancing of
 * low-weight task groups (eg. nice +19 on an autogroup), deeper taskgroup
 * hierarchies, especially on larger systems. This is not a user-visible change
 * and does not change the user-interface for setting shares/weights.
 *
 * We increase resolution only if we have enough bits to allow this increased
116 117
 * resolution (i.e. 64-bit). The costs for increasing resolution when 32-bit
 * are pretty high and the returns do not justify the increased costs.
118
 *
119 120
 * Really only required when CONFIG_FAIR_GROUP_SCHED=y is also set, but to
 * increase coverage and consistency always enable it on 64-bit platforms.
121
 */
122
#ifdef CONFIG_64BIT
123
# define NICE_0_LOAD_SHIFT	(SCHED_FIXEDPOINT_SHIFT + SCHED_FIXEDPOINT_SHIFT)
124 125
# define scale_load(w)		((w) << SCHED_FIXEDPOINT_SHIFT)
# define scale_load_down(w)	((w) >> SCHED_FIXEDPOINT_SHIFT)
126
#else
127
# define NICE_0_LOAD_SHIFT	(SCHED_FIXEDPOINT_SHIFT)
128 129 130 131
# define scale_load(w)		(w)
# define scale_load_down(w)	(w)
#endif

132
/*
133 134 135 136 137 138 139
 * Task weight (visible to users) and its load (invisible to users) have
 * independent resolution, but they should be well calibrated. We use
 * scale_load() and scale_load_down(w) to convert between them. The
 * following must be true:
 *
 *  scale_load(sched_prio_to_weight[USER_PRIO(NICE_TO_PRIO(0))]) == NICE_0_LOAD
 *
140
 */
141
#define NICE_0_LOAD		(1L << NICE_0_LOAD_SHIFT)
142

143 144 145 146 147
/*
 * Single value that decides SCHED_DEADLINE internal math precision.
 * 10 -> just above 1us
 * 9  -> just above 0.5us
 */
148
#define DL_SCALE		10
149 150

/*
151
 * Single value that denotes runtime == period, ie unlimited time.
152
 */
153
#define RUNTIME_INF		((u64)~0ULL)
154

155 156 157 158
static inline int idle_policy(int policy)
{
	return policy == SCHED_IDLE;
}
159 160 161 162 163
static inline int fair_policy(int policy)
{
	return policy == SCHED_NORMAL || policy == SCHED_BATCH;
}

164 165
static inline int rt_policy(int policy)
{
166
	return policy == SCHED_FIFO || policy == SCHED_RR;
167 168
}

169 170 171 172
static inline int dl_policy(int policy)
{
	return policy == SCHED_DEADLINE;
}
173 174 175 176 177
static inline bool valid_policy(int policy)
{
	return idle_policy(policy) || fair_policy(policy) ||
		rt_policy(policy) || dl_policy(policy);
}
178

179 180 181 182 183
static inline int task_has_rt_policy(struct task_struct *p)
{
	return rt_policy(p->policy);
}

184 185 186 187 188
static inline int task_has_dl_policy(struct task_struct *p)
{
	return dl_policy(p->policy);
}

189 190
#define cap_scale(v, s) ((v)*(s) >> SCHED_CAPACITY_SHIFT)

191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213
/*
 * !! For sched_setattr_nocheck() (kernel) only !!
 *
 * This is actually gross. :(
 *
 * It is used to make schedutil kworker(s) higher priority than SCHED_DEADLINE
 * tasks, but still be able to sleep. We need this on platforms that cannot
 * atomically change clock frequency. Remove once fast switching will be
 * available on such platforms.
 *
 * SUGOV stands for SchedUtil GOVernor.
 */
#define SCHED_FLAG_SUGOV	0x10000000

static inline bool dl_entity_is_special(struct sched_dl_entity *dl_se)
{
#ifdef CONFIG_CPU_FREQ_GOV_SCHEDUTIL
	return unlikely(dl_se->flags & SCHED_FLAG_SUGOV);
#else
	return false;
#endif
}

214 215 216
/*
 * Tells if entity @a should preempt entity @b.
 */
217 218
static inline bool
dl_entity_preempt(struct sched_dl_entity *a, struct sched_dl_entity *b)
219
{
220 221
	return dl_entity_is_special(a) ||
	       dl_time_before(a->deadline, b->deadline);
222 223
}

224 225 226 227 228 229 230 231 232 233 234 235 236 237
/*
 * This is the priority-queue data structure of the RT scheduling class:
 */
struct rt_prio_array {
	DECLARE_BITMAP(bitmap, MAX_RT_PRIO+1); /* include 1 bit for delimiter */
	struct list_head queue[MAX_RT_PRIO];
};

struct rt_bandwidth {
	/* nests inside the rq lock: */
	raw_spinlock_t		rt_runtime_lock;
	ktime_t			rt_period;
	u64			rt_runtime;
	struct hrtimer		rt_period_timer;
P
Peter Zijlstra 已提交
238
	unsigned int		rt_period_active;
239
};
240 241 242

void __dl_clear_params(struct task_struct *p);

243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267
/*
 * To keep the bandwidth of -deadline tasks and groups under control
 * we need some place where:
 *  - store the maximum -deadline bandwidth of the system (the group);
 *  - cache the fraction of that bandwidth that is currently allocated.
 *
 * This is all done in the data structure below. It is similar to the
 * one used for RT-throttling (rt_bandwidth), with the main difference
 * that, since here we are only interested in admission control, we
 * do not decrease any runtime while the group "executes", neither we
 * need a timer to replenish it.
 *
 * With respect to SMP, the bandwidth is given on a per-CPU basis,
 * meaning that:
 *  - dl_bw (< 100%) is the bandwidth of the system (group) on each CPU;
 *  - dl_total_bw array contains, in the i-eth element, the currently
 *    allocated bandwidth on the i-eth CPU.
 * Moreover, groups consume bandwidth on each CPU, while tasks only
 * consume bandwidth on the CPU they're running on.
 * Finally, dl_total_bw_cpu is used to cache the index of dl_total_bw
 * that will be shown the next time the proc or cgroup controls will
 * be red. It on its turn can be changed by writing on its own
 * control.
 */
struct dl_bandwidth {
268 269 270
	raw_spinlock_t		dl_runtime_lock;
	u64			dl_runtime;
	u64			dl_period;
271 272 273 274
};

static inline int dl_bandwidth_enabled(void)
{
275
	return sysctl_sched_rt_runtime >= 0;
276 277 278
}

struct dl_bw {
279 280 281
	raw_spinlock_t		lock;
	u64			bw;
	u64			total_bw;
282 283
};

284 285
static inline void __dl_update(struct dl_bw *dl_b, s64 bw);

286
static inline
287
void __dl_sub(struct dl_bw *dl_b, u64 tsk_bw, int cpus)
288 289
{
	dl_b->total_bw -= tsk_bw;
290
	__dl_update(dl_b, (s32)tsk_bw / cpus);
291 292 293
}

static inline
294
void __dl_add(struct dl_bw *dl_b, u64 tsk_bw, int cpus)
295 296
{
	dl_b->total_bw += tsk_bw;
297
	__dl_update(dl_b, -((s32)tsk_bw / cpus));
298 299 300 301 302 303 304 305 306
}

static inline
bool __dl_overflow(struct dl_bw *dl_b, int cpus, u64 old_bw, u64 new_bw)
{
	return dl_b->bw != -1 &&
	       dl_b->bw * cpus < dl_b->total_bw - old_bw + new_bw;
}

307
extern void dl_change_utilization(struct task_struct *p, u64 new_bw);
308
extern void init_dl_bw(struct dl_bw *dl_b);
309
extern int  sched_dl_global_validate(void);
310
extern void sched_dl_do_global(void);
311
extern int  sched_dl_overflow(struct task_struct *p, int policy, const struct sched_attr *attr);
312 313 314 315
extern void __setparam_dl(struct task_struct *p, const struct sched_attr *attr);
extern void __getparam_dl(struct task_struct *p, struct sched_attr *attr);
extern bool __checkparam_dl(const struct sched_attr *attr);
extern bool dl_param_changed(struct task_struct *p, const struct sched_attr *attr);
316 317
extern int  dl_task_can_attach(struct task_struct *p, const struct cpumask *cs_cpus_allowed);
extern int  dl_cpuset_cpumask_can_shrink(const struct cpumask *cur, const struct cpumask *trial);
318
extern bool dl_cpu_busy(unsigned int cpu);
319 320 321 322 323 324 325 326

#ifdef CONFIG_CGROUP_SCHED

#include <linux/cgroup.h>

struct cfs_rq;
struct rt_rq;

327
extern struct list_head task_groups;
328 329 330

struct cfs_bandwidth {
#ifdef CONFIG_CFS_BANDWIDTH
331 332 333 334 335 336
	raw_spinlock_t		lock;
	ktime_t			period;
	u64			quota;
	u64			runtime;
	s64			hierarchical_quota;
	u64			runtime_expires;
337
	int			expires_seq;
338

339 340
	short			idle;
	short			period_active;
341 342 343 344 345 346 347 348
	struct hrtimer		period_timer;
	struct hrtimer		slack_timer;
	struct list_head	throttled_cfs_rq;

	/* Statistics: */
	int			nr_periods;
	int			nr_throttled;
	u64			throttled_time;
349 350 351
#endif
};

352
/* Task group related information */
353 354 355 356
struct task_group {
	struct cgroup_subsys_state css;

#ifdef CONFIG_FAIR_GROUP_SCHED
357 358 359 360 361
	/* schedulable entities of this group on each CPU */
	struct sched_entity	**se;
	/* runqueue "owned" by this group on each CPU */
	struct cfs_rq		**cfs_rq;
	unsigned long		shares;
362

363
#ifdef	CONFIG_SMP
364 365 366 367 368
	/*
	 * load_avg can be heavily contended at clock tick time, so put
	 * it in its own cacheline separated from the fields above which
	 * will also be accessed at each tick.
	 */
369
	atomic_long_t		load_avg ____cacheline_aligned;
370
#endif
371
#endif
372 373

#ifdef CONFIG_RT_GROUP_SCHED
374 375
	struct sched_rt_entity	**rt_se;
	struct rt_rq		**rt_rq;
376

377
	struct rt_bandwidth	rt_bandwidth;
378 379
#endif

380 381
	struct rcu_head		rcu;
	struct list_head	list;
382

383 384 385
	struct task_group	*parent;
	struct list_head	siblings;
	struct list_head	children;
386 387

#ifdef CONFIG_SCHED_AUTOGROUP
388
	struct autogroup	*autogroup;
389 390
#endif

391
	struct cfs_bandwidth	cfs_bandwidth;
392 393 394 395 396 397 398 399 400 401 402 403 404
};

#ifdef CONFIG_FAIR_GROUP_SCHED
#define ROOT_TASK_GROUP_LOAD	NICE_0_LOAD

/*
 * A weight of 0 or 1 can cause arithmetics problems.
 * A weight of a cfs_rq is the sum of weights of which entities
 * are queued on this cfs_rq, so a weight of a entity should not be
 * too large, so as the shares value of a task group.
 * (The default weight is 1024 - so there's no practical
 *  limitation from this.)
 */
405 406
#define MIN_SHARES		(1UL <<  1)
#define MAX_SHARES		(1UL << 18)
407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428
#endif

typedef int (*tg_visitor)(struct task_group *, void *);

extern int walk_tg_tree_from(struct task_group *from,
			     tg_visitor down, tg_visitor up, void *data);

/*
 * Iterate the full tree, calling @down when first entering a node and @up when
 * leaving it for the final time.
 *
 * Caller must hold rcu_lock or sufficient equivalent.
 */
static inline int walk_tg_tree(tg_visitor down, tg_visitor up, void *data)
{
	return walk_tg_tree_from(&root_task_group, down, up, data);
}

extern int tg_nop(struct task_group *tg, void *data);

extern void free_fair_sched_group(struct task_group *tg);
extern int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent);
429
extern void online_fair_sched_group(struct task_group *tg);
430
extern void unregister_fair_sched_group(struct task_group *tg);
431 432 433 434 435 436
extern void init_tg_cfs_entry(struct task_group *tg, struct cfs_rq *cfs_rq,
			struct sched_entity *se, int cpu,
			struct sched_entity *parent);
extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b);

extern void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b);
P
Peter Zijlstra 已提交
437
extern void start_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
438 439 440 441 442 443 444
extern void unthrottle_cfs_rq(struct cfs_rq *cfs_rq);

extern void free_rt_sched_group(struct task_group *tg);
extern int alloc_rt_sched_group(struct task_group *tg, struct task_group *parent);
extern void init_tg_rt_entry(struct task_group *tg, struct rt_rq *rt_rq,
		struct sched_rt_entity *rt_se, int cpu,
		struct sched_rt_entity *parent);
445 446 447 448 449
extern int sched_group_set_rt_runtime(struct task_group *tg, long rt_runtime_us);
extern int sched_group_set_rt_period(struct task_group *tg, u64 rt_period_us);
extern long sched_group_rt_runtime(struct task_group *tg);
extern long sched_group_rt_period(struct task_group *tg);
extern int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk);
450

451 452 453 454 455 456 457 458 459 460
extern struct task_group *sched_create_group(struct task_group *parent);
extern void sched_online_group(struct task_group *tg,
			       struct task_group *parent);
extern void sched_destroy_group(struct task_group *tg);
extern void sched_offline_group(struct task_group *tg);

extern void sched_move_task(struct task_struct *tsk);

#ifdef CONFIG_FAIR_GROUP_SCHED
extern int sched_group_set_shares(struct task_group *tg, unsigned long shares);
461 462 463 464 465 466 467 468 469

#ifdef CONFIG_SMP
extern void set_task_rq_fair(struct sched_entity *se,
			     struct cfs_rq *prev, struct cfs_rq *next);
#else /* !CONFIG_SMP */
static inline void set_task_rq_fair(struct sched_entity *se,
			     struct cfs_rq *prev, struct cfs_rq *next) { }
#endif /* CONFIG_SMP */
#endif /* CONFIG_FAIR_GROUP_SCHED */
470

471 472 473 474 475 476 477 478
#else /* CONFIG_CGROUP_SCHED */

struct cfs_bandwidth { };

#endif	/* CONFIG_CGROUP_SCHED */

/* CFS-related fields in a runqueue */
struct cfs_rq {
479 480 481 482
	struct load_weight	load;
	unsigned long		runnable_weight;
	unsigned int		nr_running;
	unsigned int		h_nr_running;
483

484 485
	u64			exec_clock;
	u64			min_vruntime;
486
#ifndef CONFIG_64BIT
487
	u64			min_vruntime_copy;
488 489
#endif

490
	struct rb_root_cached	tasks_timeline;
491 492 493 494 495

	/*
	 * 'curr' points to currently running entity on this cfs_rq.
	 * It is set to NULL otherwise (i.e when none are currently running).
	 */
496 497 498 499
	struct sched_entity	*curr;
	struct sched_entity	*next;
	struct sched_entity	*last;
	struct sched_entity	*skip;
500 501

#ifdef	CONFIG_SCHED_DEBUG
502
	unsigned int		nr_spread_over;
503 504
#endif

505 506
#ifdef CONFIG_SMP
	/*
507
	 * CFS load tracking
508
	 */
509
	struct sched_avg	avg;
510
#ifndef CONFIG_64BIT
511
	u64			load_last_update_time_copy;
512
#endif
513 514 515 516 517
	struct {
		raw_spinlock_t	lock ____cacheline_aligned;
		int		nr;
		unsigned long	load_avg;
		unsigned long	util_avg;
518
		unsigned long	runnable_sum;
519
	} removed;
520

521
#ifdef CONFIG_FAIR_GROUP_SCHED
522 523 524
	unsigned long		tg_load_avg_contrib;
	long			propagate;
	long			prop_runnable_sum;
525

526 527 528 529 530 531
	/*
	 *   h_load = weight * f(tg)
	 *
	 * Where f(tg) is the recursive weight fraction assigned to
	 * this group.
	 */
532 533 534
	unsigned long		h_load;
	u64			last_h_load_update;
	struct sched_entity	*h_load_next;
535
#endif /* CONFIG_FAIR_GROUP_SCHED */
536 537
#endif /* CONFIG_SMP */

538
#ifdef CONFIG_FAIR_GROUP_SCHED
539
	struct rq		*rq;	/* CPU runqueue to which this cfs_rq is attached */
540 541 542 543 544 545

	/*
	 * leaf cfs_rqs are those that hold tasks (lowest schedulable entity in
	 * a hierarchy). Non-leaf lrqs hold other higher schedulable entities
	 * (like users, containers etc.)
	 *
546 547
	 * leaf_cfs_rq_list ties together list of leaf cfs_rq's in a CPU.
	 * This list is used during load balance.
548
	 */
549 550 551
	int			on_list;
	struct list_head	leaf_cfs_rq_list;
	struct task_group	*tg;	/* group that "owns" this runqueue */
552 553

#ifdef CONFIG_CFS_BANDWIDTH
554
	int			runtime_enabled;
555
	int			expires_seq;
556 557 558 559 560 561 562 563 564
	u64			runtime_expires;
	s64			runtime_remaining;

	u64			throttled_clock;
	u64			throttled_clock_task;
	u64			throttled_clock_task_time;
	int			throttled;
	int			throttle_count;
	struct list_head	throttled_list;
565 566 567 568 569 570 571 572 573
#endif /* CONFIG_CFS_BANDWIDTH */
#endif /* CONFIG_FAIR_GROUP_SCHED */
};

static inline int rt_bandwidth_enabled(void)
{
	return sysctl_sched_rt_runtime >= 0;
}

574
/* RT IPI pull logic requires IRQ_WORK */
575
#if defined(CONFIG_IRQ_WORK) && defined(CONFIG_SMP)
576 577 578
# define HAVE_RT_PUSH_IPI
#endif

579 580
/* Real-Time classes' related field in a runqueue: */
struct rt_rq {
581 582 583
	struct rt_prio_array	active;
	unsigned int		rt_nr_running;
	unsigned int		rr_nr_running;
584 585
#if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED
	struct {
586
		int		curr; /* highest queued rt task prio */
587
#ifdef CONFIG_SMP
588
		int		next; /* next highest */
589 590 591 592
#endif
	} highest_prio;
#endif
#ifdef CONFIG_SMP
593 594 595 596
	unsigned long		rt_nr_migratory;
	unsigned long		rt_nr_total;
	int			overloaded;
	struct plist_head	pushable_tasks;
597

598
#endif /* CONFIG_SMP */
599
	int			rt_queued;
600

601 602 603
	int			rt_throttled;
	u64			rt_time;
	u64			rt_runtime;
604
	/* Nests inside the rq lock: */
605
	raw_spinlock_t		rt_runtime_lock;
606 607

#ifdef CONFIG_RT_GROUP_SCHED
608
	unsigned long		rt_nr_boosted;
609

610 611
	struct rq		*rq;
	struct task_group	*tg;
612 613 614
#endif
};

615 616 617 618 619
static inline bool rt_rq_is_runnable(struct rt_rq *rt_rq)
{
	return rt_rq->rt_queued && rt_rq->rt_nr_running;
}

620 621 622
/* Deadline class' related fields in a runqueue */
struct dl_rq {
	/* runqueue is an rbtree, ordered by deadline */
623
	struct rb_root_cached	root;
624

625
	unsigned long		dl_nr_running;
626 627 628 629 630 631 632 633 634

#ifdef CONFIG_SMP
	/*
	 * Deadline values of the currently executing and the
	 * earliest ready task on this rq. Caching these facilitates
	 * the decision wether or not a ready but not running task
	 * should migrate somewhere else.
	 */
	struct {
635 636
		u64		curr;
		u64		next;
637 638
	} earliest_dl;

639 640
	unsigned long		dl_nr_migratory;
	int			overloaded;
641 642 643 644 645 646

	/*
	 * Tasks on this rq that can be pushed away. They are kept in
	 * an rb-tree, ordered by tasks' deadlines, with caching
	 * of the leftmost (earliest deadline) element.
	 */
647
	struct rb_root_cached	pushable_dl_tasks_root;
648
#else
649
	struct dl_bw		dl_bw;
650
#endif
651 652 653 654 655
	/*
	 * "Active utilization" for this runqueue: increased when a
	 * task wakes up (becomes TASK_RUNNING) and decreased when a
	 * task blocks
	 */
656
	u64			running_bw;
657

658 659 660 661 662 663 664 665 666
	/*
	 * Utilization of the tasks "assigned" to this runqueue (including
	 * the tasks that are in runqueue and the tasks that executed on this
	 * CPU and blocked). Increased when a task moves to this runqueue, and
	 * decreased when the task moves away (migrates, changes scheduling
	 * policy, or terminates).
	 * This is needed to compute the "inactive utilization" for the
	 * runqueue (inactive utilization = this_bw - running_bw).
	 */
667 668
	u64			this_bw;
	u64			extra_bw;
669

670 671 672 673
	/*
	 * Inverse of the fraction of CPU utilization that can be reclaimed
	 * by the GRUB algorithm.
	 */
674
	u64			bw_ratio;
675 676
};

677 678 679 680 681 682 683
#ifdef CONFIG_FAIR_GROUP_SCHED
/* An entity is a task if it doesn't "own" a runqueue */
#define entity_is_task(se)	(!se->my_q)
#else
#define entity_is_task(se)	1
#endif

684
#ifdef CONFIG_SMP
685 686 687 688 689 690 691 692 693 694 695 696
/*
 * XXX we want to get rid of these helpers and use the full load resolution.
 */
static inline long se_weight(struct sched_entity *se)
{
	return scale_load_down(se->load.weight);
}

static inline long se_runnable(struct sched_entity *se)
{
	return scale_load_down(se->runnable_weight);
}
697

T
Tim Chen 已提交
698 699 700 701 702
static inline bool sched_asym_prefer(int a, int b)
{
	return arch_asym_cpu_priority(a) > arch_asym_cpu_priority(b);
}

703 704 705
/*
 * We add the notion of a root-domain which will be used to define per-domain
 * variables. Each exclusive cpuset essentially defines an island domain by
706
 * fully partitioning the member CPUs from any other cpuset. Whenever a new
707 708 709 710 711
 * exclusive cpuset is created, we also create and attach a new root-domain
 * object.
 *
 */
struct root_domain {
712 713 714 715 716
	atomic_t		refcount;
	atomic_t		rto_count;
	struct rcu_head		rcu;
	cpumask_var_t		span;
	cpumask_var_t		online;
717

718
	/* Indicate more than one runnable task for any CPU */
719
	bool			overload;
720

721 722 723 724
	/*
	 * The bit corresponding to a CPU gets set here if such CPU has more
	 * than one runnable -deadline task (as it is below for RT tasks).
	 */
725 726 727 728
	cpumask_var_t		dlo_mask;
	atomic_t		dlo_count;
	struct dl_bw		dl_bw;
	struct cpudl		cpudl;
729

730 731 732 733
#ifdef HAVE_RT_PUSH_IPI
	/*
	 * For IPI pull requests, loop across the rto_mask.
	 */
734 735
	struct irq_work		rto_push_work;
	raw_spinlock_t		rto_lock;
736
	/* These are only updated and read within rto_lock */
737 738
	int			rto_loop;
	int			rto_cpu;
739
	/* These atomics are updated outside of a lock */
740 741
	atomic_t		rto_loop_next;
	atomic_t		rto_loop_start;
742
#endif
743 744 745 746
	/*
	 * The "RT overload" flag: it gets set if a CPU has more than
	 * one runnable RT task.
	 */
747 748
	cpumask_var_t		rto_mask;
	struct cpupri		cpupri;
749

750
	unsigned long		max_cpu_capacity;
751 752 753
};

extern struct root_domain def_root_domain;
754 755 756
extern struct mutex sched_domains_mutex;

extern void init_defrootdomain(void);
P
Peter Zijlstra 已提交
757
extern int sched_init_domains(const struct cpumask *cpu_map);
758
extern void rq_attach_root(struct rq *rq, struct root_domain *rd);
759 760
extern void sched_get_rd(struct root_domain *rd);
extern void sched_put_rd(struct root_domain *rd);
761

762 763 764
#ifdef HAVE_RT_PUSH_IPI
extern void rto_push_irq_work_func(struct irq_work *work);
#endif
765 766 767 768 769 770 771 772 773 774 775
#endif /* CONFIG_SMP */

/*
 * This is the main, per-CPU runqueue data structure.
 *
 * Locking rule: those places that want to lock multiple runqueues
 * (such as the load balancing or the thread migration code), lock
 * acquire operations must be ordered by ascending &runqueue.
 */
struct rq {
	/* runqueue lock: */
776
	raw_spinlock_t		lock;
777 778 779 780 781

	/*
	 * nr_running and cpu_load should be in the same cacheline because
	 * remote CPUs use both these fields when doing load calculation.
	 */
782
	unsigned int		nr_running;
783
#ifdef CONFIG_NUMA_BALANCING
784 785
	unsigned int		nr_numa_running;
	unsigned int		nr_preferred_running;
786
#endif
787
	#define CPU_LOAD_IDX_MAX 5
788
	unsigned long		cpu_load[CPU_LOAD_IDX_MAX];
789
#ifdef CONFIG_NO_HZ_COMMON
790
#ifdef CONFIG_SMP
791
	unsigned long		last_load_update_tick;
792
	unsigned long		last_blocked_load_update_tick;
793
	unsigned int		has_blocked_load;
794
#endif /* CONFIG_SMP */
795
	unsigned int		nohz_tick_stopped;
796
	atomic_t nohz_flags;
797
#endif /* CONFIG_NO_HZ_COMMON */
798

799 800 801 802
	/* capture load from *all* tasks on this CPU: */
	struct load_weight	load;
	unsigned long		nr_load_updates;
	u64			nr_switches;
803

804 805 806
	struct cfs_rq		cfs;
	struct rt_rq		rt;
	struct dl_rq		dl;
807 808

#ifdef CONFIG_FAIR_GROUP_SCHED
809 810 811
	/* list of leaf cfs_rq on this CPU: */
	struct list_head	leaf_cfs_rq_list;
	struct list_head	*tmp_alone_branch;
812 813
#endif /* CONFIG_FAIR_GROUP_SCHED */

814 815 816 817 818 819
	/*
	 * This is part of a global counter where only the total sum
	 * over all CPUs matters. A task can increase this counter on
	 * one CPU and if it got migrated afterwards it may decrease
	 * it on another CPU. Always updated under the runqueue lock:
	 */
820
	unsigned long		nr_uninterruptible;
821

822 823 824 825 826
	struct task_struct	*curr;
	struct task_struct	*idle;
	struct task_struct	*stop;
	unsigned long		next_balance;
	struct mm_struct	*prev_mm;
827

828 829 830
	unsigned int		clock_update_flags;
	u64			clock;
	u64			clock_task;
831

832
	atomic_t		nr_iowait;
833 834

#ifdef CONFIG_SMP
835 836 837 838 839
	struct root_domain	*rd;
	struct sched_domain	*sd;

	unsigned long		cpu_capacity;
	unsigned long		cpu_capacity_orig;
840

841
	struct callback_head	*balance_callback;
842

843
	unsigned char		idle_balance;
844

845
	/* For active balancing */
846 847 848 849 850 851 852
	int			active_balance;
	int			push_cpu;
	struct cpu_stop_work	active_balance_work;

	/* CPU of this runqueue: */
	int			cpu;
	int			online;
853

854 855
	struct list_head cfs_tasks;

856 857
	u64			rt_avg;
	u64			age_stamp;
858
	struct sched_avg	avg_rt;
859
	struct sched_avg	avg_dl;
860 861
	u64			idle_stamp;
	u64			avg_idle;
862 863

	/* This is used to determine avg_idle's max value */
864
	u64			max_idle_balance_cost;
865 866 867
#endif

#ifdef CONFIG_IRQ_TIME_ACCOUNTING
868
	u64			prev_irq_time;
869 870
#endif
#ifdef CONFIG_PARAVIRT
871
	u64			prev_steal_time;
872 873
#endif
#ifdef CONFIG_PARAVIRT_TIME_ACCOUNTING
874
	u64			prev_steal_time_rq;
875 876 877
#endif

	/* calc_load related fields */
878 879
	unsigned long		calc_load_update;
	long			calc_load_active;
880 881 882

#ifdef CONFIG_SCHED_HRTICK
#ifdef CONFIG_SMP
883 884
	int			hrtick_csd_pending;
	call_single_data_t	hrtick_csd;
885
#endif
886
	struct hrtimer		hrtick_timer;
887 888 889 890
#endif

#ifdef CONFIG_SCHEDSTATS
	/* latency stats */
891 892
	struct sched_info	rq_sched_info;
	unsigned long long	rq_cpu_time;
893 894 895
	/* could above be rq->cfs_rq.exec_clock + rq->rt_rq.rt_runtime ? */

	/* sys_sched_yield() stats */
896
	unsigned int		yld_count;
897 898

	/* schedule() stats */
899 900
	unsigned int		sched_count;
	unsigned int		sched_goidle;
901 902

	/* try_to_wake_up() stats */
903 904
	unsigned int		ttwu_count;
	unsigned int		ttwu_local;
905 906 907
#endif

#ifdef CONFIG_SMP
908
	struct llist_head	wake_list;
909
#endif
910 911 912

#ifdef CONFIG_CPU_IDLE
	/* Must be inspected within a rcu lock section */
913
	struct cpuidle_state	*idle_state;
914
#endif
915 916 917 918 919 920 921 922 923 924 925
};

static inline int cpu_of(struct rq *rq)
{
#ifdef CONFIG_SMP
	return rq->cpu;
#else
	return 0;
#endif
}

P
Peter Zijlstra 已提交
926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942

#ifdef CONFIG_SCHED_SMT

extern struct static_key_false sched_smt_present;

extern void __update_idle_core(struct rq *rq);

static inline void update_idle_core(struct rq *rq)
{
	if (static_branch_unlikely(&sched_smt_present))
		__update_idle_core(rq);
}

#else
static inline void update_idle_core(struct rq *rq) { }
#endif

943
DECLARE_PER_CPU_SHARED_ALIGNED(struct rq, runqueues);
944

945
#define cpu_rq(cpu)		(&per_cpu(runqueues, (cpu)))
946
#define this_rq()		this_cpu_ptr(&runqueues)
947 948
#define task_rq(p)		cpu_rq(task_cpu(p))
#define cpu_curr(cpu)		(cpu_rq(cpu)->curr)
949
#define raw_rq()		raw_cpu_ptr(&runqueues)
950

951 952
static inline u64 __rq_clock_broken(struct rq *rq)
{
953
	return READ_ONCE(rq->clock);
954 955
}

956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978
/*
 * rq::clock_update_flags bits
 *
 * %RQCF_REQ_SKIP - will request skipping of clock update on the next
 *  call to __schedule(). This is an optimisation to avoid
 *  neighbouring rq clock updates.
 *
 * %RQCF_ACT_SKIP - is set from inside of __schedule() when skipping is
 *  in effect and calls to update_rq_clock() are being ignored.
 *
 * %RQCF_UPDATED - is a debug flag that indicates whether a call has been
 *  made to update_rq_clock() since the last time rq::lock was pinned.
 *
 * If inside of __schedule(), clock_update_flags will have been
 * shifted left (a left shift is a cheap operation for the fast path
 * to promote %RQCF_REQ_SKIP to %RQCF_ACT_SKIP), so you must use,
 *
 *	if (rq-clock_update_flags >= RQCF_UPDATED)
 *
 * to check if %RQCF_UPADTED is set. It'll never be shifted more than
 * one position though, because the next rq_unpin_lock() will shift it
 * back.
 */
979 980 981
#define RQCF_REQ_SKIP		0x01
#define RQCF_ACT_SKIP		0x02
#define RQCF_UPDATED		0x04
982 983 984 985 986 987 988 989 990 991

static inline void assert_clock_updated(struct rq *rq)
{
	/*
	 * The only reason for not seeing a clock update since the
	 * last rq_pin_lock() is if we're currently skipping updates.
	 */
	SCHED_WARN_ON(rq->clock_update_flags < RQCF_ACT_SKIP);
}

992 993
static inline u64 rq_clock(struct rq *rq)
{
994
	lockdep_assert_held(&rq->lock);
995 996
	assert_clock_updated(rq);

997 998 999 1000 1001
	return rq->clock;
}

static inline u64 rq_clock_task(struct rq *rq)
{
1002
	lockdep_assert_held(&rq->lock);
1003 1004
	assert_clock_updated(rq);

1005 1006 1007
	return rq->clock_task;
}

1008
static inline void rq_clock_skip_update(struct rq *rq)
1009 1010
{
	lockdep_assert_held(&rq->lock);
1011 1012 1013 1014
	rq->clock_update_flags |= RQCF_REQ_SKIP;
}

/*
D
Davidlohr Bueso 已提交
1015
 * See rt task throttling, which is the only time a skip
1016 1017 1018 1019 1020 1021
 * request is cancelled.
 */
static inline void rq_clock_cancel_skipupdate(struct rq *rq)
{
	lockdep_assert_held(&rq->lock);
	rq->clock_update_flags &= ~RQCF_REQ_SKIP;
1022 1023
}

1024 1025 1026
struct rq_flags {
	unsigned long flags;
	struct pin_cookie cookie;
1027 1028 1029 1030 1031 1032 1033 1034
#ifdef CONFIG_SCHED_DEBUG
	/*
	 * A copy of (rq::clock_update_flags & RQCF_UPDATED) for the
	 * current pin context is stashed here in case it needs to be
	 * restored in rq_repin_lock().
	 */
	unsigned int clock_update_flags;
#endif
1035 1036 1037 1038 1039
};

static inline void rq_pin_lock(struct rq *rq, struct rq_flags *rf)
{
	rf->cookie = lockdep_pin_lock(&rq->lock);
1040 1041 1042 1043 1044

#ifdef CONFIG_SCHED_DEBUG
	rq->clock_update_flags &= (RQCF_REQ_SKIP|RQCF_ACT_SKIP);
	rf->clock_update_flags = 0;
#endif
1045 1046 1047 1048
}

static inline void rq_unpin_lock(struct rq *rq, struct rq_flags *rf)
{
1049 1050 1051 1052 1053
#ifdef CONFIG_SCHED_DEBUG
	if (rq->clock_update_flags > RQCF_ACT_SKIP)
		rf->clock_update_flags = RQCF_UPDATED;
#endif

1054 1055 1056 1057 1058 1059
	lockdep_unpin_lock(&rq->lock, rf->cookie);
}

static inline void rq_repin_lock(struct rq *rq, struct rq_flags *rf)
{
	lockdep_repin_lock(&rq->lock, rf->cookie);
1060 1061 1062 1063 1064 1065 1066

#ifdef CONFIG_SCHED_DEBUG
	/*
	 * Restore the value we stashed in @rf for this pin context.
	 */
	rq->clock_update_flags |= rf->clock_update_flags;
#endif
1067 1068
}

1069
#ifdef CONFIG_NUMA
1070 1071 1072 1073 1074 1075
enum numa_topology_type {
	NUMA_DIRECT,
	NUMA_GLUELESS_MESH,
	NUMA_BACKPLANE,
};
extern enum numa_topology_type sched_numa_topology_type;
1076 1077 1078 1079
extern int sched_max_numa_distance;
extern bool find_numa_distance(int distance);
#endif

1080 1081 1082 1083 1084 1085 1086 1087 1088 1089
#ifdef CONFIG_NUMA
extern void sched_init_numa(void);
extern void sched_domains_numa_masks_set(unsigned int cpu);
extern void sched_domains_numa_masks_clear(unsigned int cpu);
#else
static inline void sched_init_numa(void) { }
static inline void sched_domains_numa_masks_set(unsigned int cpu) { }
static inline void sched_domains_numa_masks_clear(unsigned int cpu) { }
#endif

1090
#ifdef CONFIG_NUMA_BALANCING
1091 1092 1093 1094 1095 1096 1097
/* The regions in numa_faults array from task_struct */
enum numa_faults_stats {
	NUMA_MEM = 0,
	NUMA_CPU,
	NUMA_MEMBUF,
	NUMA_CPUBUF
};
1098
extern void sched_setnuma(struct task_struct *p, int node);
1099
extern int migrate_task_to(struct task_struct *p, int cpu);
1100
extern int migrate_swap(struct task_struct *, struct task_struct *);
1101 1102 1103 1104 1105 1106
extern void init_numa_balancing(unsigned long clone_flags, struct task_struct *p);
#else
static inline void
init_numa_balancing(unsigned long clone_flags, struct task_struct *p)
{
}
1107 1108
#endif /* CONFIG_NUMA_BALANCING */

1109 1110
#ifdef CONFIG_SMP

1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125
static inline void
queue_balance_callback(struct rq *rq,
		       struct callback_head *head,
		       void (*func)(struct rq *rq))
{
	lockdep_assert_held(&rq->lock);

	if (unlikely(head->next))
		return;

	head->func = (void (*)(struct callback_head *))func;
	head->next = rq->balance_callback;
	rq->balance_callback = head;
}

1126 1127
extern void sched_ttwu_pending(void);

1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139
#define rcu_dereference_check_sched_domain(p) \
	rcu_dereference_check((p), \
			      lockdep_is_held(&sched_domains_mutex))

/*
 * The domain tree (rq->sd) is protected by RCU's quiescent state transition.
 * See detach_destroy_domains: synchronize_sched for details.
 *
 * The domain tree of any CPU may only be accessed from within
 * preempt-disabled sections.
 */
#define for_each_domain(cpu, __sd) \
1140 1141
	for (__sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd); \
			__sd; __sd = __sd->parent)
1142

1143 1144
#define for_each_lower_domain(sd) for (; sd; sd = sd->child)

1145 1146
/**
 * highest_flag_domain - Return highest sched_domain containing flag.
1147
 * @cpu:	The CPU whose highest level of sched domain is to
1148 1149
 *		be returned.
 * @flag:	The flag to check for the highest sched_domain
1150
 *		for the given CPU.
1151
 *
1152
 * Returns the highest sched_domain of a CPU which contains the given flag.
1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166
 */
static inline struct sched_domain *highest_flag_domain(int cpu, int flag)
{
	struct sched_domain *sd, *hsd = NULL;

	for_each_domain(cpu, sd) {
		if (!(sd->flags & flag))
			break;
		hsd = sd;
	}

	return hsd;
}

1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178
static inline struct sched_domain *lowest_flag_domain(int cpu, int flag)
{
	struct sched_domain *sd;

	for_each_domain(cpu, sd) {
		if (sd->flags & flag)
			break;
	}

	return sd;
}

1179
DECLARE_PER_CPU(struct sched_domain *, sd_llc);
1180
DECLARE_PER_CPU(int, sd_llc_size);
1181
DECLARE_PER_CPU(int, sd_llc_id);
1182
DECLARE_PER_CPU(struct sched_domain_shared *, sd_llc_shared);
1183
DECLARE_PER_CPU(struct sched_domain *, sd_numa);
1184
DECLARE_PER_CPU(struct sched_domain *, sd_asym);
1185

1186
struct sched_group_capacity {
1187
	atomic_t		ref;
1188
	/*
1189
	 * CPU capacity of this group, SCHED_CAPACITY_SCALE being max capacity
1190
	 * for a single CPU.
1191
	 */
1192 1193 1194 1195
	unsigned long		capacity;
	unsigned long		min_capacity;		/* Min per-CPU capacity in group */
	unsigned long		next_update;
	int			imbalance;		/* XXX unrelated to capacity but shared group state */
1196

1197
#ifdef CONFIG_SCHED_DEBUG
1198
	int			id;
1199 1200
#endif

1201
	unsigned long		cpumask[0];		/* Balance mask */
1202 1203 1204
};

struct sched_group {
1205 1206
	struct sched_group	*next;			/* Must be a circular list */
	atomic_t		ref;
1207

1208
	unsigned int		group_weight;
1209
	struct sched_group_capacity *sgc;
1210
	int			asym_prefer_cpu;	/* CPU of highest priority in group */
1211 1212 1213 1214 1215 1216 1217 1218

	/*
	 * The CPUs this group covers.
	 *
	 * NOTE: this field is variable length. (Allocated dynamically
	 * by attaching extra space to the end of the structure,
	 * depending on how many CPUs the kernel has booted up with)
	 */
1219
	unsigned long		cpumask[0];
1220 1221
};

1222
static inline struct cpumask *sched_group_span(struct sched_group *sg)
1223 1224 1225 1226 1227
{
	return to_cpumask(sg->cpumask);
}

/*
1228
 * See build_balance_mask().
1229
 */
1230
static inline struct cpumask *group_balance_mask(struct sched_group *sg)
1231
{
1232
	return to_cpumask(sg->sgc->cpumask);
1233 1234 1235
}

/**
1236 1237
 * group_first_cpu - Returns the first CPU in the cpumask of a sched_group.
 * @group: The group whose first CPU is to be returned.
1238 1239 1240
 */
static inline unsigned int group_first_cpu(struct sched_group *group)
{
1241
	return cpumask_first(sched_group_span(group));
1242 1243
}

P
Peter Zijlstra 已提交
1244 1245
extern int group_balance_cpu(struct sched_group *sg);

1246 1247
#if defined(CONFIG_SCHED_DEBUG) && defined(CONFIG_SYSCTL)
void register_sched_domain_sysctl(void);
1248
void dirty_sched_domain_sysctl(int cpu);
1249 1250 1251 1252 1253
void unregister_sched_domain_sysctl(void);
#else
static inline void register_sched_domain_sysctl(void)
{
}
1254 1255 1256
static inline void dirty_sched_domain_sysctl(int cpu)
{
}
1257 1258 1259 1260 1261
static inline void unregister_sched_domain_sysctl(void)
{
}
#endif

1262 1263 1264 1265
#else

static inline void sched_ttwu_pending(void) { }

1266
#endif /* CONFIG_SMP */
1267

1268
#include "stats.h"
1269
#include "autogroup.h"
1270 1271 1272 1273 1274 1275

#ifdef CONFIG_CGROUP_SCHED

/*
 * Return the group to which this tasks belongs.
 *
1276 1277 1278
 * We cannot use task_css() and friends because the cgroup subsystem
 * changes that value before the cgroup_subsys::attach() method is called,
 * therefore we cannot pin it and might observe the wrong value.
P
Peter Zijlstra 已提交
1279 1280 1281 1282 1283 1284
 *
 * The same is true for autogroup's p->signal->autogroup->tg, the autogroup
 * core changes this before calling sched_move_task().
 *
 * Instead we use a 'copy' which is updated from sched_move_task() while
 * holding both task_struct::pi_lock and rq::lock.
1285 1286 1287
 */
static inline struct task_group *task_group(struct task_struct *p)
{
P
Peter Zijlstra 已提交
1288
	return p->sched_task_group;
1289 1290 1291 1292 1293 1294 1295 1296 1297 1298
}

/* Change a task's cfs_rq and parent entity if it moves across CPUs/groups */
static inline void set_task_rq(struct task_struct *p, unsigned int cpu)
{
#if defined(CONFIG_FAIR_GROUP_SCHED) || defined(CONFIG_RT_GROUP_SCHED)
	struct task_group *tg = task_group(p);
#endif

#ifdef CONFIG_FAIR_GROUP_SCHED
1299
	set_task_rq_fair(&p->se, p->se.cfs_rq, tg->cfs_rq[cpu]);
1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329
	p->se.cfs_rq = tg->cfs_rq[cpu];
	p->se.parent = tg->se[cpu];
#endif

#ifdef CONFIG_RT_GROUP_SCHED
	p->rt.rt_rq  = tg->rt_rq[cpu];
	p->rt.parent = tg->rt_se[cpu];
#endif
}

#else /* CONFIG_CGROUP_SCHED */

static inline void set_task_rq(struct task_struct *p, unsigned int cpu) { }
static inline struct task_group *task_group(struct task_struct *p)
{
	return NULL;
}

#endif /* CONFIG_CGROUP_SCHED */

static inline void __set_task_cpu(struct task_struct *p, unsigned int cpu)
{
	set_task_rq(p, cpu);
#ifdef CONFIG_SMP
	/*
	 * After ->cpu is set up to a new value, task_rq_lock(p, ...) can be
	 * successfuly executed on another CPU. We must ensure that updates of
	 * per-task data have been completed by this moment.
	 */
	smp_wmb();
1330 1331 1332
#ifdef CONFIG_THREAD_INFO_IN_TASK
	p->cpu = cpu;
#else
1333
	task_thread_info(p)->cpu = cpu;
1334
#endif
1335
	p->wake_cpu = cpu;
1336 1337 1338 1339 1340 1341 1342
#endif
}

/*
 * Tunables that become constants when CONFIG_SCHED_DEBUG is off:
 */
#ifdef CONFIG_SCHED_DEBUG
1343
# include <linux/static_key.h>
1344 1345 1346 1347 1348 1349 1350 1351 1352
# define const_debug __read_mostly
#else
# define const_debug const
#endif

#define SCHED_FEAT(name, enabled)	\
	__SCHED_FEAT_##name ,

enum {
1353
#include "features.h"
1354
	__SCHED_FEAT_NR,
1355 1356 1357 1358
};

#undef SCHED_FEAT

1359
#if defined(CONFIG_SCHED_DEBUG) && defined(HAVE_JUMP_LABEL)
1360 1361 1362 1363 1364 1365 1366

/*
 * To support run-time toggling of sched features, all the translation units
 * (but core.c) reference the sysctl_sched_features defined in core.c.
 */
extern const_debug unsigned int sysctl_sched_features;

1367
#define SCHED_FEAT(name, enabled)					\
1368
static __always_inline bool static_branch_##name(struct static_key *key) \
1369
{									\
1370
	return static_key_##enabled(key);				\
1371 1372 1373 1374 1375
}

#include "features.h"
#undef SCHED_FEAT

1376
extern struct static_key sched_feat_keys[__SCHED_FEAT_NR];
1377
#define sched_feat(x) (static_branch_##x(&sched_feat_keys[__SCHED_FEAT_##x]))
1378

1379
#else /* !(SCHED_DEBUG && HAVE_JUMP_LABEL) */
1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392

/*
 * Each translation unit has its own copy of sysctl_sched_features to allow
 * constants propagation at compile time and compiler optimization based on
 * features default.
 */
#define SCHED_FEAT(name, enabled)	\
	(1UL << __SCHED_FEAT_##name) * enabled |
static const_debug __maybe_unused unsigned int sysctl_sched_features =
#include "features.h"
	0;
#undef SCHED_FEAT

1393
#define sched_feat(x) (sysctl_sched_features & (1UL << __SCHED_FEAT_##x))
1394

1395
#endif /* SCHED_DEBUG && HAVE_JUMP_LABEL */
1396

1397
extern struct static_key_false sched_numa_balancing;
1398
extern struct static_key_false sched_schedstats;
1399

1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426
static inline u64 global_rt_period(void)
{
	return (u64)sysctl_sched_rt_period * NSEC_PER_USEC;
}

static inline u64 global_rt_runtime(void)
{
	if (sysctl_sched_rt_runtime < 0)
		return RUNTIME_INF;

	return (u64)sysctl_sched_rt_runtime * NSEC_PER_USEC;
}

static inline int task_current(struct rq *rq, struct task_struct *p)
{
	return rq->curr == p;
}

static inline int task_running(struct rq *rq, struct task_struct *p)
{
#ifdef CONFIG_SMP
	return p->on_cpu;
#else
	return task_current(rq, p);
#endif
}

1427 1428 1429 1430
static inline int task_on_rq_queued(struct task_struct *p)
{
	return p->on_rq == TASK_ON_RQ_QUEUED;
}
1431

1432 1433 1434 1435 1436
static inline int task_on_rq_migrating(struct task_struct *p)
{
	return p->on_rq == TASK_ON_RQ_MIGRATING;
}

1437 1438 1439
/*
 * wake flags
 */
1440 1441 1442
#define WF_SYNC			0x01		/* Waker goes to sleep after wakeup */
#define WF_FORK			0x02		/* Child wakeup after fork */
#define WF_MIGRATED		0x4		/* Internal use, task got migrated */
1443

1444 1445 1446 1447 1448 1449 1450 1451 1452
/*
 * To aid in avoiding the subversion of "niceness" due to uneven distribution
 * of tasks with abnormal "nice" values across CPUs the contribution that
 * each task makes to its run queue's load is weighted according to its
 * scheduling class and "nice" value. For SCHED_NORMAL tasks this is just a
 * scaled version of the new time slice allocation that they receive on time
 * slice expiry etc.
 */

1453 1454
#define WEIGHT_IDLEPRIO		3
#define WMULT_IDLEPRIO		1431655765
1455

1456 1457
extern const int		sched_prio_to_weight[40];
extern const u32		sched_prio_to_wmult[40];
1458

1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473
/*
 * {de,en}queue flags:
 *
 * DEQUEUE_SLEEP  - task is no longer runnable
 * ENQUEUE_WAKEUP - task just became runnable
 *
 * SAVE/RESTORE - an otherwise spurious dequeue/enqueue, done to ensure tasks
 *                are in a known state which allows modification. Such pairs
 *                should preserve as much state as possible.
 *
 * MOVE - paired with SAVE/RESTORE, explicitly does not preserve the location
 *        in the runqueue.
 *
 * ENQUEUE_HEAD      - place at front of runqueue (tail if not specified)
 * ENQUEUE_REPLENISH - CBS (replenish runtime and postpone deadline)
1474
 * ENQUEUE_MIGRATED  - the task was migrated during wakeup
1475 1476 1477 1478
 *
 */

#define DEQUEUE_SLEEP		0x01
1479 1480 1481
#define DEQUEUE_SAVE		0x02 /* Matches ENQUEUE_RESTORE */
#define DEQUEUE_MOVE		0x04 /* Matches ENQUEUE_MOVE */
#define DEQUEUE_NOCLOCK		0x08 /* Matches ENQUEUE_NOCLOCK */
1482

1483
#define ENQUEUE_WAKEUP		0x01
1484 1485
#define ENQUEUE_RESTORE		0x02
#define ENQUEUE_MOVE		0x04
1486
#define ENQUEUE_NOCLOCK		0x08
1487

1488 1489
#define ENQUEUE_HEAD		0x10
#define ENQUEUE_REPLENISH	0x20
1490
#ifdef CONFIG_SMP
1491
#define ENQUEUE_MIGRATED	0x40
1492
#else
1493
#define ENQUEUE_MIGRATED	0x00
1494 1495
#endif

1496 1497
#define RETRY_TASK		((void *)-1UL)

1498 1499 1500 1501 1502
struct sched_class {
	const struct sched_class *next;

	void (*enqueue_task) (struct rq *rq, struct task_struct *p, int flags);
	void (*dequeue_task) (struct rq *rq, struct task_struct *p, int flags);
1503 1504
	void (*yield_task)   (struct rq *rq);
	bool (*yield_to_task)(struct rq *rq, struct task_struct *p, bool preempt);
1505

1506
	void (*check_preempt_curr)(struct rq *rq, struct task_struct *p, int flags);
1507

1508 1509 1510 1511
	/*
	 * It is the responsibility of the pick_next_task() method that will
	 * return the next task to call put_prev_task() on the @prev task or
	 * something equivalent.
1512 1513 1514
	 *
	 * May return RETRY_TASK when it finds a higher prio class has runnable
	 * tasks.
1515
	 */
1516 1517 1518 1519
	struct task_struct * (*pick_next_task)(struct rq *rq,
					       struct task_struct *prev,
					       struct rq_flags *rf);
	void (*put_prev_task)(struct rq *rq, struct task_struct *p);
1520 1521

#ifdef CONFIG_SMP
1522
	int  (*select_task_rq)(struct task_struct *p, int task_cpu, int sd_flag, int flags);
1523
	void (*migrate_task_rq)(struct task_struct *p);
1524

1525
	void (*task_woken)(struct rq *this_rq, struct task_struct *task);
1526 1527 1528 1529 1530 1531 1532 1533

	void (*set_cpus_allowed)(struct task_struct *p,
				 const struct cpumask *newmask);

	void (*rq_online)(struct rq *rq);
	void (*rq_offline)(struct rq *rq);
#endif

1534 1535 1536 1537
	void (*set_curr_task)(struct rq *rq);
	void (*task_tick)(struct rq *rq, struct task_struct *p, int queued);
	void (*task_fork)(struct task_struct *p);
	void (*task_dead)(struct task_struct *p);
1538

1539 1540 1541 1542 1543
	/*
	 * The switched_from() call is allowed to drop rq->lock, therefore we
	 * cannot assume the switched_from/switched_to pair is serliazed by
	 * rq->lock. They are however serialized by p->pi_lock.
	 */
1544 1545
	void (*switched_from)(struct rq *this_rq, struct task_struct *task);
	void (*switched_to)  (struct rq *this_rq, struct task_struct *task);
1546
	void (*prio_changed) (struct rq *this_rq, struct task_struct *task,
1547
			      int oldprio);
1548

1549 1550
	unsigned int (*get_rr_interval)(struct rq *rq,
					struct task_struct *task);
1551

1552
	void (*update_curr)(struct rq *rq);
1553

1554 1555
#define TASK_SET_GROUP		0
#define TASK_MOVE_GROUP		1
1556

1557
#ifdef CONFIG_FAIR_GROUP_SCHED
1558
	void (*task_change_group)(struct task_struct *p, int type);
1559 1560
#endif
};
1561

1562 1563 1564 1565 1566
static inline void put_prev_task(struct rq *rq, struct task_struct *prev)
{
	prev->sched_class->put_prev_task(rq, prev);
}

1567 1568 1569 1570 1571
static inline void set_curr_task(struct rq *rq, struct task_struct *curr)
{
	curr->sched_class->set_curr_task(rq);
}

1572
#ifdef CONFIG_SMP
1573
#define sched_class_highest (&stop_sched_class)
1574 1575 1576
#else
#define sched_class_highest (&dl_sched_class)
#endif
1577 1578 1579 1580
#define for_each_class(class) \
   for (class = sched_class_highest; class; class = class->next)

extern const struct sched_class stop_sched_class;
1581
extern const struct sched_class dl_sched_class;
1582 1583 1584 1585 1586 1587 1588
extern const struct sched_class rt_sched_class;
extern const struct sched_class fair_sched_class;
extern const struct sched_class idle_sched_class;


#ifdef CONFIG_SMP

1589
extern void update_group_capacity(struct sched_domain *sd, int cpu);
1590

1591
extern void trigger_load_balance(struct rq *rq);
1592

1593 1594
extern void set_cpus_allowed_common(struct task_struct *p, const struct cpumask *new_mask);

1595 1596
#endif

1597 1598 1599 1600 1601 1602 1603 1604 1605
#ifdef CONFIG_CPU_IDLE
static inline void idle_set_state(struct rq *rq,
				  struct cpuidle_state *idle_state)
{
	rq->idle_state = idle_state;
}

static inline struct cpuidle_state *idle_get_state(struct rq *rq)
{
1606
	SCHED_WARN_ON(!rcu_read_lock_held());
1607

1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621
	return rq->idle_state;
}
#else
static inline void idle_set_state(struct rq *rq,
				  struct cpuidle_state *idle_state)
{
}

static inline struct cpuidle_state *idle_get_state(struct rq *rq)
{
	return NULL;
}
#endif

1622 1623
extern void schedule_idle(void);

1624 1625 1626
extern void sysrq_sched_debug_show(void);
extern void sched_init_granularity(void);
extern void update_max_interval(void);
1627 1628

extern void init_sched_dl_class(void);
1629 1630 1631
extern void init_sched_rt_class(void);
extern void init_sched_fair_class(void);

1632 1633
extern void reweight_task(struct task_struct *p, int prio);

1634
extern void resched_curr(struct rq *rq);
1635 1636 1637 1638 1639
extern void resched_cpu(int cpu);

extern struct rt_bandwidth def_rt_bandwidth;
extern void init_rt_bandwidth(struct rt_bandwidth *rt_b, u64 period, u64 runtime);

1640 1641
extern struct dl_bandwidth def_dl_bandwidth;
extern void init_dl_bandwidth(struct dl_bandwidth *dl_b, u64 period, u64 runtime);
1642
extern void init_dl_task_timer(struct sched_dl_entity *dl_se);
1643
extern void init_dl_inactive_task_timer(struct sched_dl_entity *dl_se);
1644
extern void init_dl_rq_bw_ratio(struct dl_rq *dl_rq);
1645

1646 1647 1648
#define BW_SHIFT		20
#define BW_UNIT			(1 << BW_SHIFT)
#define RATIO_SHIFT		8
1649 1650
unsigned long to_ratio(u64 period, u64 runtime);

1651
extern void init_entity_runnable_average(struct sched_entity *se);
1652
extern void post_init_entity_util_avg(struct sched_entity *se);
1653

1654 1655
#ifdef CONFIG_NO_HZ_FULL
extern bool sched_can_stop_tick(struct rq *rq);
1656
extern int __init sched_tick_offload_init(void);
1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680

/*
 * Tick may be needed by tasks in the runqueue depending on their policy and
 * requirements. If tick is needed, lets send the target an IPI to kick it out of
 * nohz mode if necessary.
 */
static inline void sched_update_tick_dependency(struct rq *rq)
{
	int cpu;

	if (!tick_nohz_full_enabled())
		return;

	cpu = cpu_of(rq);

	if (!tick_nohz_full_cpu(cpu))
		return;

	if (sched_can_stop_tick(rq))
		tick_nohz_dep_clear_cpu(cpu, TICK_DEP_BIT_SCHED);
	else
		tick_nohz_dep_set_cpu(cpu, TICK_DEP_BIT_SCHED);
}
#else
1681
static inline int sched_tick_offload_init(void) { return 0; }
1682 1683 1684
static inline void sched_update_tick_dependency(struct rq *rq) { }
#endif

1685
static inline void add_nr_running(struct rq *rq, unsigned count)
1686
{
1687 1688 1689
	unsigned prev_nr = rq->nr_running;

	rq->nr_running = prev_nr + count;
1690

1691
	if (prev_nr < 2 && rq->nr_running >= 2) {
1692 1693 1694 1695 1696
#ifdef CONFIG_SMP
		if (!rq->rd->overload)
			rq->rd->overload = true;
#endif
	}
1697 1698

	sched_update_tick_dependency(rq);
1699 1700
}

1701
static inline void sub_nr_running(struct rq *rq, unsigned count)
1702
{
1703
	rq->nr_running -= count;
1704 1705
	/* Check if we still need preemption */
	sched_update_tick_dependency(rq);
1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741
}

extern void update_rq_clock(struct rq *rq);

extern void activate_task(struct rq *rq, struct task_struct *p, int flags);
extern void deactivate_task(struct rq *rq, struct task_struct *p, int flags);

extern void check_preempt_curr(struct rq *rq, struct task_struct *p, int flags);

extern const_debug unsigned int sysctl_sched_time_avg;
extern const_debug unsigned int sysctl_sched_nr_migrate;
extern const_debug unsigned int sysctl_sched_migration_cost;

static inline u64 sched_avg_period(void)
{
	return (u64)sysctl_sched_time_avg * NSEC_PER_MSEC / 2;
}

#ifdef CONFIG_SCHED_HRTICK

/*
 * Use hrtick when:
 *  - enabled by features
 *  - hrtimer is actually high res
 */
static inline int hrtick_enabled(struct rq *rq)
{
	if (!sched_feat(HRTICK))
		return 0;
	if (!cpu_active(cpu_of(rq)))
		return 0;
	return hrtimer_is_hres_active(&rq->hrtick_timer);
}

void hrtick_start(struct rq *rq, u64 delay);

1742 1743 1744 1745 1746 1747 1748
#else

static inline int hrtick_enabled(struct rq *rq)
{
	return 0;
}

1749 1750
#endif /* CONFIG_SCHED_HRTICK */

1751 1752
#ifndef arch_scale_freq_capacity
static __always_inline
1753
unsigned long arch_scale_freq_capacity(int cpu)
1754 1755 1756 1757
{
	return SCHED_CAPACITY_SCALE;
}
#endif
1758

1759 1760 1761
#ifdef CONFIG_SMP
extern void sched_avg_update(struct rq *rq);

1762 1763 1764 1765
#ifndef arch_scale_cpu_capacity
static __always_inline
unsigned long arch_scale_cpu_capacity(struct sched_domain *sd, int cpu)
{
1766
	if (sd && (sd->flags & SD_SHARE_CPUCAPACITY) && (sd->span_weight > 1))
1767 1768 1769 1770 1771 1772
		return sd->smt_gain / sd->span_weight;

	return SCHED_CAPACITY_SCALE;
}
#endif

1773 1774
static inline void sched_rt_avg_update(struct rq *rq, u64 rt_delta)
{
1775
	rq->rt_avg += rt_delta * arch_scale_freq_capacity(cpu_of(rq));
1776 1777 1778
	sched_avg_update(rq);
}
#else
1779 1780 1781 1782 1783 1784 1785
#ifndef arch_scale_cpu_capacity
static __always_inline
unsigned long arch_scale_cpu_capacity(void __always_unused *sd, int cpu)
{
	return SCHED_CAPACITY_SCALE;
}
#endif
1786 1787 1788 1789
static inline void sched_rt_avg_update(struct rq *rq, u64 rt_delta) { }
static inline void sched_avg_update(struct rq *rq) { }
#endif

1790
struct rq *__task_rq_lock(struct task_struct *p, struct rq_flags *rf)
1791
	__acquires(rq->lock);
1792

1793
struct rq *task_rq_lock(struct task_struct *p, struct rq_flags *rf)
1794
	__acquires(p->pi_lock)
1795
	__acquires(rq->lock);
1796

1797
static inline void __task_rq_unlock(struct rq *rq, struct rq_flags *rf)
1798 1799
	__releases(rq->lock)
{
1800
	rq_unpin_lock(rq, rf);
1801 1802 1803 1804
	raw_spin_unlock(&rq->lock);
}

static inline void
1805
task_rq_unlock(struct rq *rq, struct task_struct *p, struct rq_flags *rf)
1806 1807 1808
	__releases(rq->lock)
	__releases(p->pi_lock)
{
1809
	rq_unpin_lock(rq, rf);
1810
	raw_spin_unlock(&rq->lock);
1811
	raw_spin_unlock_irqrestore(&p->pi_lock, rf->flags);
1812 1813
}

1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869
static inline void
rq_lock_irqsave(struct rq *rq, struct rq_flags *rf)
	__acquires(rq->lock)
{
	raw_spin_lock_irqsave(&rq->lock, rf->flags);
	rq_pin_lock(rq, rf);
}

static inline void
rq_lock_irq(struct rq *rq, struct rq_flags *rf)
	__acquires(rq->lock)
{
	raw_spin_lock_irq(&rq->lock);
	rq_pin_lock(rq, rf);
}

static inline void
rq_lock(struct rq *rq, struct rq_flags *rf)
	__acquires(rq->lock)
{
	raw_spin_lock(&rq->lock);
	rq_pin_lock(rq, rf);
}

static inline void
rq_relock(struct rq *rq, struct rq_flags *rf)
	__acquires(rq->lock)
{
	raw_spin_lock(&rq->lock);
	rq_repin_lock(rq, rf);
}

static inline void
rq_unlock_irqrestore(struct rq *rq, struct rq_flags *rf)
	__releases(rq->lock)
{
	rq_unpin_lock(rq, rf);
	raw_spin_unlock_irqrestore(&rq->lock, rf->flags);
}

static inline void
rq_unlock_irq(struct rq *rq, struct rq_flags *rf)
	__releases(rq->lock)
{
	rq_unpin_lock(rq, rf);
	raw_spin_unlock_irq(&rq->lock);
}

static inline void
rq_unlock(struct rq *rq, struct rq_flags *rf)
	__releases(rq->lock)
{
	rq_unpin_lock(rq, rf);
	raw_spin_unlock(&rq->lock);
}

1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897
#ifdef CONFIG_SMP
#ifdef CONFIG_PREEMPT

static inline void double_rq_lock(struct rq *rq1, struct rq *rq2);

/*
 * fair double_lock_balance: Safely acquires both rq->locks in a fair
 * way at the expense of forcing extra atomic operations in all
 * invocations.  This assures that the double_lock is acquired using the
 * same underlying policy as the spinlock_t on this architecture, which
 * reduces latency compared to the unfair variant below.  However, it
 * also adds more overhead and therefore may reduce throughput.
 */
static inline int _double_lock_balance(struct rq *this_rq, struct rq *busiest)
	__releases(this_rq->lock)
	__acquires(busiest->lock)
	__acquires(this_rq->lock)
{
	raw_spin_unlock(&this_rq->lock);
	double_rq_lock(this_rq, busiest);

	return 1;
}

#else
/*
 * Unfair double_lock_balance: Optimizes throughput at the expense of
 * latency by eliminating extra atomic operations when the locks are
1898 1899
 * already in proper order on entry.  This favors lower CPU-ids and will
 * grant the double lock to lower CPUs over higher ids under contention,
1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930
 * regardless of entry order into the function.
 */
static inline int _double_lock_balance(struct rq *this_rq, struct rq *busiest)
	__releases(this_rq->lock)
	__acquires(busiest->lock)
	__acquires(this_rq->lock)
{
	int ret = 0;

	if (unlikely(!raw_spin_trylock(&busiest->lock))) {
		if (busiest < this_rq) {
			raw_spin_unlock(&this_rq->lock);
			raw_spin_lock(&busiest->lock);
			raw_spin_lock_nested(&this_rq->lock,
					      SINGLE_DEPTH_NESTING);
			ret = 1;
		} else
			raw_spin_lock_nested(&busiest->lock,
					      SINGLE_DEPTH_NESTING);
	}
	return ret;
}

#endif /* CONFIG_PREEMPT */

/*
 * double_lock_balance - lock the busiest runqueue, this_rq is locked already.
 */
static inline int double_lock_balance(struct rq *this_rq, struct rq *busiest)
{
	if (unlikely(!irqs_disabled())) {
1931
		/* printk() doesn't work well under rq->lock */
1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945
		raw_spin_unlock(&this_rq->lock);
		BUG_ON(1);
	}

	return _double_lock_balance(this_rq, busiest);
}

static inline void double_unlock_balance(struct rq *this_rq, struct rq *busiest)
	__releases(busiest->lock)
{
	raw_spin_unlock(&busiest->lock);
	lock_set_subclass(&this_rq->lock.dep_map, 0, _RET_IP_);
}

1946 1947 1948 1949 1950 1951 1952 1953 1954
static inline void double_lock(spinlock_t *l1, spinlock_t *l2)
{
	if (l1 > l2)
		swap(l1, l2);

	spin_lock(l1);
	spin_lock_nested(l2, SINGLE_DEPTH_NESTING);
}

1955 1956 1957 1958 1959 1960 1961 1962 1963
static inline void double_lock_irq(spinlock_t *l1, spinlock_t *l2)
{
	if (l1 > l2)
		swap(l1, l2);

	spin_lock_irq(l1);
	spin_lock_nested(l2, SINGLE_DEPTH_NESTING);
}

1964 1965 1966 1967 1968 1969 1970 1971 1972
static inline void double_raw_lock(raw_spinlock_t *l1, raw_spinlock_t *l2)
{
	if (l1 > l2)
		swap(l1, l2);

	raw_spin_lock(l1);
	raw_spin_lock_nested(l2, SINGLE_DEPTH_NESTING);
}

1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
/*
 * double_rq_lock - safely lock two runqueues
 *
 * Note this does not disable interrupts like task_rq_lock,
 * you need to do so manually before calling.
 */
static inline void double_rq_lock(struct rq *rq1, struct rq *rq2)
	__acquires(rq1->lock)
	__acquires(rq2->lock)
{
	BUG_ON(!irqs_disabled());
	if (rq1 == rq2) {
		raw_spin_lock(&rq1->lock);
		__acquire(rq2->lock);	/* Fake it out ;) */
	} else {
		if (rq1 < rq2) {
			raw_spin_lock(&rq1->lock);
			raw_spin_lock_nested(&rq2->lock, SINGLE_DEPTH_NESTING);
		} else {
			raw_spin_lock(&rq2->lock);
			raw_spin_lock_nested(&rq1->lock, SINGLE_DEPTH_NESTING);
		}
	}
}

/*
 * double_rq_unlock - safely unlock two runqueues
 *
 * Note this does not restore interrupts like task_rq_unlock,
 * you need to do so manually after calling.
 */
static inline void double_rq_unlock(struct rq *rq1, struct rq *rq2)
	__releases(rq1->lock)
	__releases(rq2->lock)
{
	raw_spin_unlock(&rq1->lock);
	if (rq1 != rq2)
		raw_spin_unlock(&rq2->lock);
	else
		__release(rq2->lock);
}

2015 2016 2017 2018
extern void set_rq_online (struct rq *rq);
extern void set_rq_offline(struct rq *rq);
extern bool sched_smp_initialized;

2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055
#else /* CONFIG_SMP */

/*
 * double_rq_lock - safely lock two runqueues
 *
 * Note this does not disable interrupts like task_rq_lock,
 * you need to do so manually before calling.
 */
static inline void double_rq_lock(struct rq *rq1, struct rq *rq2)
	__acquires(rq1->lock)
	__acquires(rq2->lock)
{
	BUG_ON(!irqs_disabled());
	BUG_ON(rq1 != rq2);
	raw_spin_lock(&rq1->lock);
	__acquire(rq2->lock);	/* Fake it out ;) */
}

/*
 * double_rq_unlock - safely unlock two runqueues
 *
 * Note this does not restore interrupts like task_rq_unlock,
 * you need to do so manually after calling.
 */
static inline void double_rq_unlock(struct rq *rq1, struct rq *rq2)
	__releases(rq1->lock)
	__releases(rq2->lock)
{
	BUG_ON(rq1 != rq2);
	raw_spin_unlock(&rq1->lock);
	__release(rq2->lock);
}

#endif

extern struct sched_entity *__pick_first_entity(struct cfs_rq *cfs_rq);
extern struct sched_entity *__pick_last_entity(struct cfs_rq *cfs_rq);
2056 2057

#ifdef	CONFIG_SCHED_DEBUG
2058 2059
extern bool sched_debug_enabled;

2060 2061
extern void print_cfs_stats(struct seq_file *m, int cpu);
extern void print_rt_stats(struct seq_file *m, int cpu);
2062
extern void print_dl_stats(struct seq_file *m, int cpu);
2063 2064 2065
extern void print_cfs_rq(struct seq_file *m, int cpu, struct cfs_rq *cfs_rq);
extern void print_rt_rq(struct seq_file *m, int cpu, struct rt_rq *rt_rq);
extern void print_dl_rq(struct seq_file *m, int cpu, struct dl_rq *dl_rq);
2066 2067 2068 2069 2070 2071 2072 2073
#ifdef CONFIG_NUMA_BALANCING
extern void
show_numa_stats(struct task_struct *p, struct seq_file *m);
extern void
print_numa_stats(struct seq_file *m, int node, unsigned long tsf,
	unsigned long tpf, unsigned long gsf, unsigned long gpf);
#endif /* CONFIG_NUMA_BALANCING */
#endif /* CONFIG_SCHED_DEBUG */
2074 2075

extern void init_cfs_rq(struct cfs_rq *cfs_rq);
2076 2077
extern void init_rt_rq(struct rt_rq *rt_rq);
extern void init_dl_rq(struct dl_rq *dl_rq);
2078

2079 2080
extern void cfs_bandwidth_usage_inc(void);
extern void cfs_bandwidth_usage_dec(void);
2081

2082
#ifdef CONFIG_NO_HZ_COMMON
2083 2084
#define NOHZ_BALANCE_KICK_BIT	0
#define NOHZ_STATS_KICK_BIT	1
2085 2086

#define NOHZ_BALANCE_KICK	BIT(NOHZ_BALANCE_KICK_BIT)
P
Peter Zijlstra 已提交
2087 2088 2089
#define NOHZ_STATS_KICK		BIT(NOHZ_STATS_KICK_BIT)

#define NOHZ_KICK_MASK	(NOHZ_BALANCE_KICK | NOHZ_STATS_KICK)
2090 2091

#define nohz_flags(cpu)	(&cpu_rq(cpu)->nohz_flags)
2092

2093
extern void nohz_balance_exit_idle(struct rq *rq);
2094
#else
2095
static inline void nohz_balance_exit_idle(struct rq *rq) { }
2096
#endif
2097

2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124

#ifdef CONFIG_SMP
static inline
void __dl_update(struct dl_bw *dl_b, s64 bw)
{
	struct root_domain *rd = container_of(dl_b, struct root_domain, dl_bw);
	int i;

	RCU_LOCKDEP_WARN(!rcu_read_lock_sched_held(),
			 "sched RCU must be held");
	for_each_cpu_and(i, rd->span, cpu_active_mask) {
		struct rq *rq = cpu_rq(i);

		rq->dl.extra_bw += bw;
	}
}
#else
static inline
void __dl_update(struct dl_bw *dl_b, s64 bw)
{
	struct dl_rq *dl = container_of(dl_b, struct dl_rq, dl_bw);

	dl->extra_bw += bw;
}
#endif


2125
#ifdef CONFIG_IRQ_TIME_ACCOUNTING
2126
struct irqtime {
2127
	u64			total;
2128
	u64			tick_delta;
2129 2130 2131
	u64			irq_start_time;
	struct u64_stats_sync	sync;
};
2132

2133
DECLARE_PER_CPU(struct irqtime, cpu_irqtime);
2134

2135 2136 2137 2138 2139
/*
 * Returns the irqtime minus the softirq time computed by ksoftirqd.
 * Otherwise ksoftirqd's sum_exec_runtime is substracted its own runtime
 * and never move forward.
 */
2140 2141
static inline u64 irq_time_read(int cpu)
{
2142 2143 2144
	struct irqtime *irqtime = &per_cpu(cpu_irqtime, cpu);
	unsigned int seq;
	u64 total;
2145 2146

	do {
2147
		seq = __u64_stats_fetch_begin(&irqtime->sync);
2148
		total = irqtime->total;
2149
	} while (__u64_stats_fetch_retry(&irqtime->sync, seq));
2150

2151
	return total;
2152 2153
}
#endif /* CONFIG_IRQ_TIME_ACCOUNTING */
2154 2155 2156 2157 2158 2159

#ifdef CONFIG_CPU_FREQ
DECLARE_PER_CPU(struct update_util_data *, cpufreq_update_util_data);

/**
 * cpufreq_update_util - Take a note about CPU utilization changes.
2160
 * @rq: Runqueue to carry out the update for.
2161
 * @flags: Update reason flags.
2162
 *
2163 2164
 * This function is called by the scheduler on the CPU whose utilization is
 * being updated.
2165 2166 2167 2168 2169 2170
 *
 * It can only be called from RCU-sched read-side critical sections.
 *
 * The way cpufreq is currently arranged requires it to evaluate the CPU
 * performance state (frequency/voltage) on a regular basis to prevent it from
 * being stuck in a completely inadequate performance level for too long.
2171 2172 2173
 * That is not guaranteed to happen if the updates are only triggered from CFS
 * and DL, though, because they may not be coming in if only RT tasks are
 * active all the time (or there are RT tasks only).
2174
 *
2175 2176
 * As a workaround for that issue, this function is called periodically by the
 * RT sched class to trigger extra cpufreq updates to prevent it from stalling,
2177
 * but that really is a band-aid.  Going forward it should be replaced with
2178
 * solutions targeted more specifically at RT tasks.
2179
 */
2180
static inline void cpufreq_update_util(struct rq *rq, unsigned int flags)
2181
{
2182 2183
	struct update_util_data *data;

2184 2185
	data = rcu_dereference_sched(*per_cpu_ptr(&cpufreq_update_util_data,
						  cpu_of(rq)));
2186
	if (data)
2187 2188
		data->func(data, rq_clock(rq), flags);
}
2189
#else
2190
static inline void cpufreq_update_util(struct rq *rq, unsigned int flags) {}
2191
#endif /* CONFIG_CPU_FREQ */
2192

2193
#ifdef arch_scale_freq_capacity
2194 2195 2196 2197 2198
# ifndef arch_scale_freq_invariant
#  define arch_scale_freq_invariant()	true
# endif
#else
# define arch_scale_freq_invariant()	false
2199
#endif
2200

2201
#ifdef CONFIG_CPU_FREQ_GOV_SCHEDUTIL
2202 2203 2204 2205 2206 2207 2208
static inline unsigned long cpu_util_dl(struct rq *rq)
{
	return (rq->dl.running_bw * SCHED_CAPACITY_SCALE) >> BW_SHIFT;
}

static inline unsigned long cpu_util_cfs(struct rq *rq)
{
2209 2210 2211 2212 2213 2214 2215 2216
	unsigned long util = READ_ONCE(rq->cfs.avg.util_avg);

	if (sched_feat(UTIL_EST)) {
		util = max_t(unsigned long, util,
			     READ_ONCE(rq->cfs.avg.util_est.enqueued));
	}

	return util;
2217
}
2218 2219 2220 2221 2222

static inline unsigned long cpu_util_rt(struct rq *rq)
{
	return rq->avg_rt.util_avg;
}
2223
#endif