sched.c 225.7 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
/*
 *  kernel/sched.c
 *
 *  Kernel scheduler and related syscalls
 *
 *  Copyright (C) 1991-2002  Linus Torvalds
 *
 *  1996-12-23  Modified by Dave Grothe to fix bugs in semaphores and
 *		make semaphores SMP safe
 *  1998-11-19	Implemented schedule_timeout() and related stuff
 *		by Andrea Arcangeli
 *  2002-01-04	New ultra-scalable O(1) scheduler by Ingo Molnar:
 *		hybrid priority-list and round-robin design with
 *		an array-switch method of distributing timeslices
 *		and per-CPU runqueues.  Cleanups and useful suggestions
 *		by Davide Libenzi, preemptible kernel bits by Robert Love.
 *  2003-09-03	Interactivity tuning by Con Kolivas.
 *  2004-04-02	Scheduler domains code by Nick Piggin
I
Ingo Molnar 已提交
19 20 21 22 23 24
 *  2007-04-15  Work begun on replacing all interactivity tuning with a
 *              fair scheduling design by Con Kolivas.
 *  2007-05-05  Load balancing (smp-nice) and other improvements
 *              by Peter Williams
 *  2007-05-06  Interactivity improvements to CFS by Mike Galbraith
 *  2007-07-01  Group scheduling enhancements by Srivatsa Vaddagiri
25 26
 *  2007-11-29  RT balancing improvements by Steven Rostedt, Gregory Haskins,
 *              Thomas Gleixner, Mike Kravetz
L
Linus Torvalds 已提交
27 28 29 30 31 32
 */

#include <linux/mm.h>
#include <linux/module.h>
#include <linux/nmi.h>
#include <linux/init.h>
33
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
34 35 36 37
#include <linux/highmem.h>
#include <linux/smp_lock.h>
#include <asm/mmu_context.h>
#include <linux/interrupt.h>
38
#include <linux/capability.h>
L
Linus Torvalds 已提交
39 40
#include <linux/completion.h>
#include <linux/kernel_stat.h>
41
#include <linux/debug_locks.h>
42
#include <linux/perf_event.h>
L
Linus Torvalds 已提交
43 44 45
#include <linux/security.h>
#include <linux/notifier.h>
#include <linux/profile.h>
46
#include <linux/freezer.h>
47
#include <linux/vmalloc.h>
L
Linus Torvalds 已提交
48 49
#include <linux/blkdev.h>
#include <linux/delay.h>
50
#include <linux/pid_namespace.h>
L
Linus Torvalds 已提交
51 52 53 54 55 56 57
#include <linux/smp.h>
#include <linux/threads.h>
#include <linux/timer.h>
#include <linux/rcupdate.h>
#include <linux/cpu.h>
#include <linux/cpuset.h>
#include <linux/percpu.h>
58
#include <linux/proc_fs.h>
L
Linus Torvalds 已提交
59
#include <linux/seq_file.h>
60
#include <linux/stop_machine.h>
61
#include <linux/sysctl.h>
L
Linus Torvalds 已提交
62 63
#include <linux/syscalls.h>
#include <linux/times.h>
64
#include <linux/tsacct_kern.h>
65
#include <linux/kprobes.h>
66
#include <linux/delayacct.h>
67
#include <linux/unistd.h>
J
Jens Axboe 已提交
68
#include <linux/pagemap.h>
P
Peter Zijlstra 已提交
69
#include <linux/hrtimer.h>
R
Reynes Philippe 已提交
70
#include <linux/tick.h>
P
Peter Zijlstra 已提交
71 72
#include <linux/debugfs.h>
#include <linux/ctype.h>
73
#include <linux/ftrace.h>
74
#include <linux/slab.h>
L
Linus Torvalds 已提交
75

76
#include <asm/tlb.h>
77
#include <asm/irq_regs.h>
L
Linus Torvalds 已提交
78

79
#include "sched_cpupri.h"
T
Tejun Heo 已提交
80
#include "workqueue_sched.h"
81

82
#define CREATE_TRACE_POINTS
83
#include <trace/events/sched.h>
84

L
Linus Torvalds 已提交
85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103
/*
 * Convert user-nice values [ -20 ... 0 ... 19 ]
 * to static priority [ MAX_RT_PRIO..MAX_PRIO-1 ],
 * and back.
 */
#define NICE_TO_PRIO(nice)	(MAX_RT_PRIO + (nice) + 20)
#define PRIO_TO_NICE(prio)	((prio) - MAX_RT_PRIO - 20)
#define TASK_NICE(p)		PRIO_TO_NICE((p)->static_prio)

/*
 * 'User priority' is the nice value converted to something we
 * can work with better when scaling various scheduler parameters,
 * it's a [ 0 ... 39 ] range.
 */
#define USER_PRIO(p)		((p)-MAX_RT_PRIO)
#define TASK_USER_PRIO(p)	USER_PRIO((p)->static_prio)
#define MAX_USER_PRIO		(USER_PRIO(MAX_PRIO))

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

I
Ingo Molnar 已提交
108 109 110
#define NICE_0_LOAD		SCHED_LOAD_SCALE
#define NICE_0_SHIFT		SCHED_LOAD_SHIFT

L
Linus Torvalds 已提交
111 112 113
/*
 * These are the 'tuning knobs' of the scheduler:
 *
D
Dmitry Adamushko 已提交
114
 * default timeslice is 100 msecs (used only for SCHED_RR tasks).
L
Linus Torvalds 已提交
115 116 117
 * Timeslices get refilled after they expire.
 */
#define DEF_TIMESLICE		(100 * HZ / 1000)
118

119 120 121 122 123
/*
 * single value that denotes runtime == period, ie unlimited time.
 */
#define RUNTIME_INF	((u64)~0ULL)

124 125
static inline int rt_policy(int policy)
{
126
	if (unlikely(policy == SCHED_FIFO || policy == SCHED_RR))
127 128 129 130 131 132 133 134 135
		return 1;
	return 0;
}

static inline int task_has_rt_policy(struct task_struct *p)
{
	return rt_policy(p->policy);
}

L
Linus Torvalds 已提交
136
/*
I
Ingo Molnar 已提交
137
 * This is the priority-queue data structure of the RT scheduling class:
L
Linus Torvalds 已提交
138
 */
I
Ingo Molnar 已提交
139 140 141 142 143
struct rt_prio_array {
	DECLARE_BITMAP(bitmap, MAX_RT_PRIO+1); /* include 1 bit for delimiter */
	struct list_head queue[MAX_RT_PRIO];
};

144
struct rt_bandwidth {
I
Ingo Molnar 已提交
145
	/* nests inside the rq lock: */
146
	raw_spinlock_t		rt_runtime_lock;
I
Ingo Molnar 已提交
147 148 149
	ktime_t			rt_period;
	u64			rt_runtime;
	struct hrtimer		rt_period_timer;
150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182
};

static struct rt_bandwidth def_rt_bandwidth;

static int do_sched_rt_period_timer(struct rt_bandwidth *rt_b, int overrun);

static enum hrtimer_restart sched_rt_period_timer(struct hrtimer *timer)
{
	struct rt_bandwidth *rt_b =
		container_of(timer, struct rt_bandwidth, rt_period_timer);
	ktime_t now;
	int overrun;
	int idle = 0;

	for (;;) {
		now = hrtimer_cb_get_time(timer);
		overrun = hrtimer_forward(timer, now, rt_b->rt_period);

		if (!overrun)
			break;

		idle = do_sched_rt_period_timer(rt_b, overrun);
	}

	return idle ? HRTIMER_NORESTART : HRTIMER_RESTART;
}

static
void init_rt_bandwidth(struct rt_bandwidth *rt_b, u64 period, u64 runtime)
{
	rt_b->rt_period = ns_to_ktime(period);
	rt_b->rt_runtime = runtime;

183
	raw_spin_lock_init(&rt_b->rt_runtime_lock);
P
Peter Zijlstra 已提交
184

185 186 187 188 189
	hrtimer_init(&rt_b->rt_period_timer,
			CLOCK_MONOTONIC, HRTIMER_MODE_REL);
	rt_b->rt_period_timer.function = sched_rt_period_timer;
}

190 191 192
static inline int rt_bandwidth_enabled(void)
{
	return sysctl_sched_rt_runtime >= 0;
193 194 195 196 197 198
}

static void start_rt_bandwidth(struct rt_bandwidth *rt_b)
{
	ktime_t now;

199
	if (!rt_bandwidth_enabled() || rt_b->rt_runtime == RUNTIME_INF)
200 201 202 203 204
		return;

	if (hrtimer_active(&rt_b->rt_period_timer))
		return;

205
	raw_spin_lock(&rt_b->rt_runtime_lock);
206
	for (;;) {
207 208 209
		unsigned long delta;
		ktime_t soft, hard;

210 211 212 213 214
		if (hrtimer_active(&rt_b->rt_period_timer))
			break;

		now = hrtimer_cb_get_time(&rt_b->rt_period_timer);
		hrtimer_forward(&rt_b->rt_period_timer, now, rt_b->rt_period);
215 216 217 218 219

		soft = hrtimer_get_softexpires(&rt_b->rt_period_timer);
		hard = hrtimer_get_expires(&rt_b->rt_period_timer);
		delta = ktime_to_ns(ktime_sub(hard, soft));
		__hrtimer_start_range_ns(&rt_b->rt_period_timer, soft, delta,
220
				HRTIMER_MODE_ABS_PINNED, 0);
221
	}
222
	raw_spin_unlock(&rt_b->rt_runtime_lock);
223 224 225 226 227 228 229 230 231
}

#ifdef CONFIG_RT_GROUP_SCHED
static void destroy_rt_bandwidth(struct rt_bandwidth *rt_b)
{
	hrtimer_cancel(&rt_b->rt_period_timer);
}
#endif

232 233 234 235 236 237
/*
 * sched_domains_mutex serializes calls to arch_init_sched_domains,
 * detach_destroy_domains and partition_sched_domains.
 */
static DEFINE_MUTEX(sched_domains_mutex);

D
Dhaval Giani 已提交
238
#ifdef CONFIG_CGROUP_SCHED
S
Srivatsa Vaddagiri 已提交
239

240 241
#include <linux/cgroup.h>

S
Srivatsa Vaddagiri 已提交
242 243
struct cfs_rq;

P
Peter Zijlstra 已提交
244 245
static LIST_HEAD(task_groups);

S
Srivatsa Vaddagiri 已提交
246
/* task group related information */
247
struct task_group {
248
	struct cgroup_subsys_state css;
249

250
#ifdef CONFIG_FAIR_GROUP_SCHED
S
Srivatsa Vaddagiri 已提交
251 252 253 254 255
	/* 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;
256 257 258 259 260 261
#endif

#ifdef CONFIG_RT_GROUP_SCHED
	struct sched_rt_entity **rt_se;
	struct rt_rq **rt_rq;

262
	struct rt_bandwidth rt_bandwidth;
263
#endif
264

265
	struct rcu_head rcu;
P
Peter Zijlstra 已提交
266
	struct list_head list;
P
Peter Zijlstra 已提交
267 268 269 270

	struct task_group *parent;
	struct list_head siblings;
	struct list_head children;
S
Srivatsa Vaddagiri 已提交
271 272
};

273
#define root_task_group init_task_group
P
Peter Zijlstra 已提交
274

275
/* task_group_lock serializes add/remove of task groups and also changes to
276 277
 * a task group's cpu shares.
 */
278
static DEFINE_SPINLOCK(task_group_lock);
279

280 281
#ifdef CONFIG_FAIR_GROUP_SCHED

282 283 284 285 286 287 288
#ifdef CONFIG_SMP
static int root_task_group_empty(void)
{
	return list_empty(&root_task_group.children);
}
#endif

289 290
# define INIT_TASK_GROUP_LOAD	NICE_0_LOAD

291
/*
292 293 294 295
 * 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.
296 297 298
 * (The default weight is 1024 - so there's no practical
 *  limitation from this.)
 */
299
#define MIN_SHARES	2
300
#define MAX_SHARES	(1UL << 18)
301

302 303 304
static int init_task_group_load = INIT_TASK_GROUP_LOAD;
#endif

S
Srivatsa Vaddagiri 已提交
305
/* Default task group.
I
Ingo Molnar 已提交
306
 *	Every task in system belong to this group at bootup.
S
Srivatsa Vaddagiri 已提交
307
 */
308
struct task_group init_task_group;
S
Srivatsa Vaddagiri 已提交
309

D
Dhaval Giani 已提交
310
#endif	/* CONFIG_CGROUP_SCHED */
S
Srivatsa Vaddagiri 已提交
311

I
Ingo Molnar 已提交
312 313 314 315 316 317
/* CFS-related fields in a runqueue */
struct cfs_rq {
	struct load_weight load;
	unsigned long nr_running;

	u64 exec_clock;
I
Ingo Molnar 已提交
318
	u64 min_vruntime;
I
Ingo Molnar 已提交
319 320 321

	struct rb_root tasks_timeline;
	struct rb_node *rb_leftmost;
322 323 324 325 326 327

	struct list_head tasks;
	struct list_head *balance_iterator;

	/*
	 * 'curr' points to currently running entity on this cfs_rq.
I
Ingo Molnar 已提交
328 329
	 * It is set to NULL otherwise (i.e when none are currently running).
	 */
P
Peter Zijlstra 已提交
330
	struct sched_entity *curr, *next, *last;
P
Peter Zijlstra 已提交
331

P
Peter Zijlstra 已提交
332
	unsigned int nr_spread_over;
P
Peter Zijlstra 已提交
333

334
#ifdef CONFIG_FAIR_GROUP_SCHED
I
Ingo Molnar 已提交
335 336
	struct rq *rq;	/* cpu runqueue to which this cfs_rq is attached */

I
Ingo Molnar 已提交
337 338
	/*
	 * leaf cfs_rqs are those that hold tasks (lowest schedulable entity in
I
Ingo Molnar 已提交
339 340 341 342 343 344
	 * a hierarchy). Non-leaf lrqs hold other higher schedulable entities
	 * (like users, containers etc.)
	 *
	 * leaf_cfs_rq_list ties together list of leaf cfs_rq's in a cpu. This
	 * list is used during load balance.
	 */
I
Ingo Molnar 已提交
345 346
	struct list_head leaf_cfs_rq_list;
	struct task_group *tg;	/* group that "owns" this runqueue */
347 348 349

#ifdef CONFIG_SMP
	/*
350
	 * the part of load.weight contributed by tasks
351
	 */
352
	unsigned long task_weight;
353

354 355 356 357 358 359 360
	/*
	 *   h_load = weight * f(tg)
	 *
	 * Where f(tg) is the recursive weight fraction assigned to
	 * this group.
	 */
	unsigned long h_load;
361

362 363 364 365
	/*
	 * this cpu's part of tg->shares
	 */
	unsigned long shares;
366 367 368 369 370

	/*
	 * load.weight at the time we set shares
	 */
	unsigned long rq_weight;
371
#endif
I
Ingo Molnar 已提交
372 373
#endif
};
L
Linus Torvalds 已提交
374

I
Ingo Molnar 已提交
375 376 377
/* Real-Time classes' related field in a runqueue: */
struct rt_rq {
	struct rt_prio_array active;
378
	unsigned long rt_nr_running;
379
#if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED
380 381
	struct {
		int curr; /* highest queued rt task prio */
382
#ifdef CONFIG_SMP
383
		int next; /* next highest */
384
#endif
385
	} highest_prio;
P
Peter Zijlstra 已提交
386
#endif
P
Peter Zijlstra 已提交
387
#ifdef CONFIG_SMP
388
	unsigned long rt_nr_migratory;
389
	unsigned long rt_nr_total;
G
Gregory Haskins 已提交
390
	int overloaded;
391
	struct plist_head pushable_tasks;
P
Peter Zijlstra 已提交
392
#endif
P
Peter Zijlstra 已提交
393
	int rt_throttled;
P
Peter Zijlstra 已提交
394
	u64 rt_time;
P
Peter Zijlstra 已提交
395
	u64 rt_runtime;
I
Ingo Molnar 已提交
396
	/* Nests inside the rq lock: */
397
	raw_spinlock_t rt_runtime_lock;
P
Peter Zijlstra 已提交
398

399
#ifdef CONFIG_RT_GROUP_SCHED
P
Peter Zijlstra 已提交
400 401
	unsigned long rt_nr_boosted;

P
Peter Zijlstra 已提交
402 403 404 405
	struct rq *rq;
	struct list_head leaf_rt_rq_list;
	struct task_group *tg;
#endif
I
Ingo Molnar 已提交
406 407
};

G
Gregory Haskins 已提交
408 409 410 411
#ifdef CONFIG_SMP

/*
 * We add the notion of a root-domain which will be used to define per-domain
I
Ingo Molnar 已提交
412 413
 * variables. Each exclusive cpuset essentially defines an island domain by
 * fully partitioning the member cpus from any other cpuset. Whenever a new
G
Gregory Haskins 已提交
414 415 416 417 418 419
 * exclusive cpuset is created, we also create and attach a new root-domain
 * object.
 *
 */
struct root_domain {
	atomic_t refcount;
420 421
	cpumask_var_t span;
	cpumask_var_t online;
422

I
Ingo Molnar 已提交
423
	/*
424 425 426
	 * The "RT overload" flag: it gets set if a CPU has more than
	 * one runnable RT task.
	 */
427
	cpumask_var_t rto_mask;
I
Ingo Molnar 已提交
428
	atomic_t rto_count;
429
	struct cpupri cpupri;
G
Gregory Haskins 已提交
430 431
};

432 433 434 435
/*
 * By default the system creates a single root-domain with all cpus as
 * members (mimicking the global state we have today).
 */
G
Gregory Haskins 已提交
436 437
static struct root_domain def_root_domain;

438
#endif /* CONFIG_SMP */
G
Gregory Haskins 已提交
439

L
Linus Torvalds 已提交
440 441 442 443 444 445 446
/*
 * 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.
 */
447
struct rq {
448
	/* runqueue lock: */
449
	raw_spinlock_t lock;
L
Linus Torvalds 已提交
450 451 452 453 454 455

	/*
	 * nr_running and cpu_load should be in the same cacheline because
	 * remote CPUs use both these fields when doing load calculation.
	 */
	unsigned long nr_running;
I
Ingo Molnar 已提交
456 457
	#define CPU_LOAD_IDX_MAX 5
	unsigned long cpu_load[CPU_LOAD_IDX_MAX];
458
	unsigned long last_load_update_tick;
459
#ifdef CONFIG_NO_HZ
M
Mike Galbraith 已提交
460
	u64 nohz_stamp;
461
	unsigned char nohz_balance_kick;
462
#endif
463 464
	unsigned int skip_clock_update;

465 466
	/* capture load from *all* tasks on this cpu: */
	struct load_weight load;
I
Ingo Molnar 已提交
467 468 469 470
	unsigned long nr_load_updates;
	u64 nr_switches;

	struct cfs_rq cfs;
P
Peter Zijlstra 已提交
471 472
	struct rt_rq rt;

I
Ingo Molnar 已提交
473
#ifdef CONFIG_FAIR_GROUP_SCHED
474 475
	/* list of leaf cfs_rq on this cpu: */
	struct list_head leaf_cfs_rq_list;
476 477
#endif
#ifdef CONFIG_RT_GROUP_SCHED
P
Peter Zijlstra 已提交
478
	struct list_head leaf_rt_rq_list;
L
Linus Torvalds 已提交
479 480 481 482 483 484 485 486 487 488
#endif

	/*
	 * 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:
	 */
	unsigned long nr_uninterruptible;

489
	struct task_struct *curr, *idle, *stop;
490
	unsigned long next_balance;
L
Linus Torvalds 已提交
491
	struct mm_struct *prev_mm;
I
Ingo Molnar 已提交
492

493
	u64 clock;
494
	u64 clock_task;
I
Ingo Molnar 已提交
495

L
Linus Torvalds 已提交
496 497 498
	atomic_t nr_iowait;

#ifdef CONFIG_SMP
I
Ingo Molnar 已提交
499
	struct root_domain *rd;
L
Linus Torvalds 已提交
500 501
	struct sched_domain *sd;

502 503
	unsigned long cpu_power;

504
	unsigned char idle_at_tick;
L
Linus Torvalds 已提交
505
	/* For active balancing */
506
	int post_schedule;
L
Linus Torvalds 已提交
507 508
	int active_balance;
	int push_cpu;
509
	struct cpu_stop_work active_balance_work;
510 511
	/* cpu of this runqueue: */
	int cpu;
512
	int online;
L
Linus Torvalds 已提交
513

514
	unsigned long avg_load_per_task;
L
Linus Torvalds 已提交
515

516 517
	u64 rt_avg;
	u64 age_stamp;
M
Mike Galbraith 已提交
518 519
	u64 idle_stamp;
	u64 avg_idle;
L
Linus Torvalds 已提交
520 521
#endif

522 523 524 525
#ifdef CONFIG_IRQ_TIME_ACCOUNTING
	u64 prev_irq_time;
#endif

526 527 528 529
	/* calc_load related fields */
	unsigned long calc_load_update;
	long calc_load_active;

P
Peter Zijlstra 已提交
530
#ifdef CONFIG_SCHED_HRTICK
531 532 533 534
#ifdef CONFIG_SMP
	int hrtick_csd_pending;
	struct call_single_data hrtick_csd;
#endif
P
Peter Zijlstra 已提交
535 536 537
	struct hrtimer hrtick_timer;
#endif

L
Linus Torvalds 已提交
538 539 540
#ifdef CONFIG_SCHEDSTATS
	/* latency stats */
	struct sched_info rq_sched_info;
541 542
	unsigned long long rq_cpu_time;
	/* could above be rq->cfs_rq.exec_clock + rq->rt_rq.rt_runtime ? */
L
Linus Torvalds 已提交
543 544

	/* sys_sched_yield() stats */
545
	unsigned int yld_count;
L
Linus Torvalds 已提交
546 547

	/* schedule() stats */
548 549 550
	unsigned int sched_switch;
	unsigned int sched_count;
	unsigned int sched_goidle;
L
Linus Torvalds 已提交
551 552

	/* try_to_wake_up() stats */
553 554
	unsigned int ttwu_count;
	unsigned int ttwu_local;
I
Ingo Molnar 已提交
555 556

	/* BKL stats */
557
	unsigned int bkl_count;
L
Linus Torvalds 已提交
558 559 560
#endif
};

561
static DEFINE_PER_CPU_SHARED_ALIGNED(struct rq, runqueues);
L
Linus Torvalds 已提交
562

P
Peter Zijlstra 已提交
563 564
static inline
void check_preempt_curr(struct rq *rq, struct task_struct *p, int flags)
I
Ingo Molnar 已提交
565
{
P
Peter Zijlstra 已提交
566
	rq->curr->sched_class->check_preempt_curr(rq, p, flags);
567 568 569 570 571 572 573

	/*
	 * A queue event has occurred, and we're going to schedule.  In
	 * this case, we can save a useless back to back clock update.
	 */
	if (test_tsk_need_resched(p))
		rq->skip_clock_update = 1;
I
Ingo Molnar 已提交
574 575
}

576 577 578 579 580 581 582 583 584
static inline int cpu_of(struct rq *rq)
{
#ifdef CONFIG_SMP
	return rq->cpu;
#else
	return 0;
#endif
}

585
#define rcu_dereference_check_sched_domain(p) \
586 587 588 589
	rcu_dereference_check((p), \
			      rcu_read_lock_sched_held() || \
			      lockdep_is_held(&sched_domains_mutex))

N
Nick Piggin 已提交
590 591
/*
 * The domain tree (rq->sd) is protected by RCU's quiescent state transition.
592
 * See detach_destroy_domains: synchronize_sched for details.
N
Nick Piggin 已提交
593 594 595 596
 *
 * The domain tree of any CPU may only be accessed from within
 * preempt-disabled sections.
 */
597
#define for_each_domain(cpu, __sd) \
598
	for (__sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd); __sd; __sd = __sd->parent)
L
Linus Torvalds 已提交
599 600 601 602 603

#define cpu_rq(cpu)		(&per_cpu(runqueues, (cpu)))
#define this_rq()		(&__get_cpu_var(runqueues))
#define task_rq(p)		cpu_rq(task_cpu(p))
#define cpu_curr(cpu)		(cpu_rq(cpu)->curr)
604
#define raw_rq()		(&__raw_get_cpu_var(runqueues))
L
Linus Torvalds 已提交
605

606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648
#ifdef CONFIG_CGROUP_SCHED

/*
 * Return the group to which this tasks belongs.
 *
 * We use task_subsys_state_check() and extend the RCU verification
 * with lockdep_is_held(&task_rq(p)->lock) because cpu_cgroup_attach()
 * holds that lock for each task it moves into the cgroup. Therefore
 * by holding that lock, we pin the task to the current cgroup.
 */
static inline struct task_group *task_group(struct task_struct *p)
{
	struct cgroup_subsys_state *css;

	css = task_subsys_state_check(p, cpu_cgroup_subsys_id,
			lockdep_is_held(&task_rq(p)->lock));
	return container_of(css, struct task_group, css);
}

/* 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)
{
#ifdef CONFIG_FAIR_GROUP_SCHED
	p->se.cfs_rq = task_group(p)->cfs_rq[cpu];
	p->se.parent = task_group(p)->se[cpu];
#endif

#ifdef CONFIG_RT_GROUP_SCHED
	p->rt.rt_rq  = task_group(p)->rt_rq[cpu];
	p->rt.parent = task_group(p)->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 */

649
static u64 irq_time_cpu(int cpu);
650
static void sched_irq_time_avg_update(struct rq *rq, u64 irq_time);
651

I
Ingo Molnar 已提交
652
inline void update_rq_clock(struct rq *rq)
653
{
654 655 656 657 658 659 660 661
	if (!rq->skip_clock_update) {
		int cpu = cpu_of(rq);
		u64 irq_time;

		rq->clock = sched_clock_cpu(cpu);
		irq_time = irq_time_cpu(cpu);
		if (rq->clock - irq_time > rq->clock_task)
			rq->clock_task = rq->clock - irq_time;
662 663

		sched_irq_time_avg_update(rq, irq_time);
664
	}
665 666
}

I
Ingo Molnar 已提交
667 668 669 670 671 672 673 674 675
/*
 * Tunables that become constants when CONFIG_SCHED_DEBUG is off:
 */
#ifdef CONFIG_SCHED_DEBUG
# define const_debug __read_mostly
#else
# define const_debug static const
#endif

I
Ingo Molnar 已提交
676 677
/**
 * runqueue_is_locked
678
 * @cpu: the processor in question.
I
Ingo Molnar 已提交
679 680 681 682 683
 *
 * Returns true if the current cpu runqueue is locked.
 * This interface allows printk to be called with the runqueue lock
 * held and know whether or not it is OK to wake up the klogd.
 */
684
int runqueue_is_locked(int cpu)
I
Ingo Molnar 已提交
685
{
686
	return raw_spin_is_locked(&cpu_rq(cpu)->lock);
I
Ingo Molnar 已提交
687 688
}

I
Ingo Molnar 已提交
689 690 691
/*
 * Debugging: various feature bits
 */
P
Peter Zijlstra 已提交
692 693 694 695

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

I
Ingo Molnar 已提交
696
enum {
P
Peter Zijlstra 已提交
697
#include "sched_features.h"
I
Ingo Molnar 已提交
698 699
};

P
Peter Zijlstra 已提交
700 701 702 703 704
#undef SCHED_FEAT

#define SCHED_FEAT(name, enabled)	\
	(1UL << __SCHED_FEAT_##name) * enabled |

I
Ingo Molnar 已提交
705
const_debug unsigned int sysctl_sched_features =
P
Peter Zijlstra 已提交
706 707 708 709 710 711 712 713 714
#include "sched_features.h"
	0;

#undef SCHED_FEAT

#ifdef CONFIG_SCHED_DEBUG
#define SCHED_FEAT(name, enabled)	\
	#name ,

715
static __read_mostly char *sched_feat_names[] = {
P
Peter Zijlstra 已提交
716 717 718 719 720 721
#include "sched_features.h"
	NULL
};

#undef SCHED_FEAT

L
Li Zefan 已提交
722
static int sched_feat_show(struct seq_file *m, void *v)
P
Peter Zijlstra 已提交
723 724 725 726
{
	int i;

	for (i = 0; sched_feat_names[i]; i++) {
L
Li Zefan 已提交
727 728 729
		if (!(sysctl_sched_features & (1UL << i)))
			seq_puts(m, "NO_");
		seq_printf(m, "%s ", sched_feat_names[i]);
P
Peter Zijlstra 已提交
730
	}
L
Li Zefan 已提交
731
	seq_puts(m, "\n");
P
Peter Zijlstra 已提交
732

L
Li Zefan 已提交
733
	return 0;
P
Peter Zijlstra 已提交
734 735 736 737 738 739 740
}

static ssize_t
sched_feat_write(struct file *filp, const char __user *ubuf,
		size_t cnt, loff_t *ppos)
{
	char buf[64];
741
	char *cmp;
P
Peter Zijlstra 已提交
742 743 744 745 746 747 748 749 750 751
	int neg = 0;
	int i;

	if (cnt > 63)
		cnt = 63;

	if (copy_from_user(&buf, ubuf, cnt))
		return -EFAULT;

	buf[cnt] = 0;
752
	cmp = strstrip(buf);
P
Peter Zijlstra 已提交
753

I
Ingo Molnar 已提交
754
	if (strncmp(buf, "NO_", 3) == 0) {
P
Peter Zijlstra 已提交
755 756 757 758 759
		neg = 1;
		cmp += 3;
	}

	for (i = 0; sched_feat_names[i]; i++) {
760
		if (strcmp(cmp, sched_feat_names[i]) == 0) {
P
Peter Zijlstra 已提交
761 762 763 764 765 766 767 768 769 770 771
			if (neg)
				sysctl_sched_features &= ~(1UL << i);
			else
				sysctl_sched_features |= (1UL << i);
			break;
		}
	}

	if (!sched_feat_names[i])
		return -EINVAL;

772
	*ppos += cnt;
P
Peter Zijlstra 已提交
773 774 775 776

	return cnt;
}

L
Li Zefan 已提交
777 778 779 780 781
static int sched_feat_open(struct inode *inode, struct file *filp)
{
	return single_open(filp, sched_feat_show, NULL);
}

782
static const struct file_operations sched_feat_fops = {
L
Li Zefan 已提交
783 784 785 786 787
	.open		= sched_feat_open,
	.write		= sched_feat_write,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= single_release,
P
Peter Zijlstra 已提交
788 789 790 791 792 793 794 795 796 797 798 799 800 801
};

static __init int sched_init_debug(void)
{
	debugfs_create_file("sched_features", 0644, NULL, NULL,
			&sched_feat_fops);

	return 0;
}
late_initcall(sched_init_debug);

#endif

#define sched_feat(x) (sysctl_sched_features & (1UL << __SCHED_FEAT_##x))
I
Ingo Molnar 已提交
802

803 804 805 806 807 808
/*
 * Number of tasks to iterate in a single balance run.
 * Limited because this is done with IRQs disabled.
 */
const_debug unsigned int sysctl_sched_nr_migrate = 32;

P
Peter Zijlstra 已提交
809 810
/*
 * ratelimit for updating the group shares.
811
 * default: 0.25ms
P
Peter Zijlstra 已提交
812
 */
813
unsigned int sysctl_sched_shares_ratelimit = 250000;
814
unsigned int normalized_sysctl_sched_shares_ratelimit = 250000;
P
Peter Zijlstra 已提交
815

816 817 818 819 820 821 822
/*
 * Inject some fuzzyness into changing the per-cpu group shares
 * this avoids remote rq-locks at the expense of fairness.
 * default: 4
 */
unsigned int sysctl_sched_shares_thresh = 4;

823 824 825 826 827 828 829 830
/*
 * period over which we average the RT time consumption, measured
 * in ms.
 *
 * default: 1s
 */
const_debug unsigned int sysctl_sched_time_avg = MSEC_PER_SEC;

P
Peter Zijlstra 已提交
831
/*
P
Peter Zijlstra 已提交
832
 * period over which we measure -rt task cpu usage in us.
P
Peter Zijlstra 已提交
833 834
 * default: 1s
 */
P
Peter Zijlstra 已提交
835
unsigned int sysctl_sched_rt_period = 1000000;
P
Peter Zijlstra 已提交
836

837 838
static __read_mostly int scheduler_running;

P
Peter Zijlstra 已提交
839 840 841 842 843
/*
 * part of the period that we allow rt tasks to run in us.
 * default: 0.95s
 */
int sysctl_sched_rt_runtime = 950000;
P
Peter Zijlstra 已提交
844

845 846 847 848 849 850 851
static inline u64 global_rt_period(void)
{
	return (u64)sysctl_sched_rt_period * NSEC_PER_USEC;
}

static inline u64 global_rt_runtime(void)
{
852
	if (sysctl_sched_rt_runtime < 0)
853 854 855 856
		return RUNTIME_INF;

	return (u64)sysctl_sched_rt_runtime * NSEC_PER_USEC;
}
P
Peter Zijlstra 已提交
857

L
Linus Torvalds 已提交
858
#ifndef prepare_arch_switch
859 860 861 862 863 864
# define prepare_arch_switch(next)	do { } while (0)
#endif
#ifndef finish_arch_switch
# define finish_arch_switch(prev)	do { } while (0)
#endif

865 866 867 868 869
static inline int task_current(struct rq *rq, struct task_struct *p)
{
	return rq->curr == p;
}

870
#ifndef __ARCH_WANT_UNLOCKED_CTXSW
871
static inline int task_running(struct rq *rq, struct task_struct *p)
872
{
873
	return task_current(rq, p);
874 875
}

876
static inline void prepare_lock_switch(struct rq *rq, struct task_struct *next)
877 878 879
{
}

880
static inline void finish_lock_switch(struct rq *rq, struct task_struct *prev)
881
{
882 883 884 885
#ifdef CONFIG_DEBUG_SPINLOCK
	/* this is a valid case when another task releases the spinlock */
	rq->lock.owner = current;
#endif
886 887 888 889 890 891 892
	/*
	 * If we are tracking spinlock dependencies then we have to
	 * fix up the runqueue lock - which gets 'carried over' from
	 * prev into current:
	 */
	spin_acquire(&rq->lock.dep_map, 0, 0, _THIS_IP_);

893
	raw_spin_unlock_irq(&rq->lock);
894 895 896
}

#else /* __ARCH_WANT_UNLOCKED_CTXSW */
897
static inline int task_running(struct rq *rq, struct task_struct *p)
898 899 900 901
{
#ifdef CONFIG_SMP
	return p->oncpu;
#else
902
	return task_current(rq, p);
903 904 905
#endif
}

906
static inline void prepare_lock_switch(struct rq *rq, struct task_struct *next)
907 908 909 910 911 912 913 914 915 916
{
#ifdef CONFIG_SMP
	/*
	 * We can optimise this out completely for !SMP, because the
	 * SMP rebalancing from interrupt is the only thing that cares
	 * here.
	 */
	next->oncpu = 1;
#endif
#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
917
	raw_spin_unlock_irq(&rq->lock);
918
#else
919
	raw_spin_unlock(&rq->lock);
920 921 922
#endif
}

923
static inline void finish_lock_switch(struct rq *rq, struct task_struct *prev)
924 925 926 927 928 929 930 931 932 933 934 935
{
#ifdef CONFIG_SMP
	/*
	 * After ->oncpu is cleared, the task can be moved to a different CPU.
	 * We must ensure this doesn't happen until the switch is completely
	 * finished.
	 */
	smp_wmb();
	prev->oncpu = 0;
#endif
#ifndef __ARCH_WANT_INTERRUPTS_ON_CTXSW
	local_irq_enable();
L
Linus Torvalds 已提交
936
#endif
937 938
}
#endif /* __ARCH_WANT_UNLOCKED_CTXSW */
L
Linus Torvalds 已提交
939

940
/*
P
Peter Zijlstra 已提交
941 942
 * Check whether the task is waking, we use this to synchronize ->cpus_allowed
 * against ttwu().
943 944 945
 */
static inline int task_is_waking(struct task_struct *p)
{
946
	return unlikely(p->state == TASK_WAKING);
947 948
}

949 950 951 952
/*
 * __task_rq_lock - lock the runqueue a given task resides on.
 * Must be called interrupts disabled.
 */
953
static inline struct rq *__task_rq_lock(struct task_struct *p)
954 955
	__acquires(rq->lock)
{
956 957
	struct rq *rq;

958
	for (;;) {
959
		rq = task_rq(p);
960
		raw_spin_lock(&rq->lock);
P
Peter Zijlstra 已提交
961
		if (likely(rq == task_rq(p)))
962
			return rq;
963
		raw_spin_unlock(&rq->lock);
964 965 966
	}
}

L
Linus Torvalds 已提交
967 968
/*
 * task_rq_lock - lock the runqueue a given task resides on and disable
I
Ingo Molnar 已提交
969
 * interrupts. Note the ordering: we can safely lookup the task_rq without
L
Linus Torvalds 已提交
970 971
 * explicitly disabling preemption.
 */
972
static struct rq *task_rq_lock(struct task_struct *p, unsigned long *flags)
L
Linus Torvalds 已提交
973 974
	__acquires(rq->lock)
{
975
	struct rq *rq;
L
Linus Torvalds 已提交
976

977 978 979
	for (;;) {
		local_irq_save(*flags);
		rq = task_rq(p);
980
		raw_spin_lock(&rq->lock);
P
Peter Zijlstra 已提交
981
		if (likely(rq == task_rq(p)))
982
			return rq;
983
		raw_spin_unlock_irqrestore(&rq->lock, *flags);
L
Linus Torvalds 已提交
984 985 986
	}
}

A
Alexey Dobriyan 已提交
987
static void __task_rq_unlock(struct rq *rq)
988 989
	__releases(rq->lock)
{
990
	raw_spin_unlock(&rq->lock);
991 992
}

993
static inline void task_rq_unlock(struct rq *rq, unsigned long *flags)
L
Linus Torvalds 已提交
994 995
	__releases(rq->lock)
{
996
	raw_spin_unlock_irqrestore(&rq->lock, *flags);
L
Linus Torvalds 已提交
997 998 999
}

/*
1000
 * this_rq_lock - lock this runqueue and disable interrupts.
L
Linus Torvalds 已提交
1001
 */
A
Alexey Dobriyan 已提交
1002
static struct rq *this_rq_lock(void)
L
Linus Torvalds 已提交
1003 1004
	__acquires(rq->lock)
{
1005
	struct rq *rq;
L
Linus Torvalds 已提交
1006 1007 1008

	local_irq_disable();
	rq = this_rq();
1009
	raw_spin_lock(&rq->lock);
L
Linus Torvalds 已提交
1010 1011 1012 1013

	return rq;
}

P
Peter Zijlstra 已提交
1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034
#ifdef CONFIG_SCHED_HRTICK
/*
 * Use HR-timers to deliver accurate preemption points.
 *
 * Its all a bit involved since we cannot program an hrt while holding the
 * rq->lock. So what we do is store a state in in rq->hrtick_* and ask for a
 * reschedule event.
 *
 * When we get rescheduled we reprogram the hrtick_timer outside of the
 * rq->lock.
 */

/*
 * 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;
1035
	if (!cpu_active(cpu_of(rq)))
1036
		return 0;
P
Peter Zijlstra 已提交
1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055
	return hrtimer_is_hres_active(&rq->hrtick_timer);
}

static void hrtick_clear(struct rq *rq)
{
	if (hrtimer_active(&rq->hrtick_timer))
		hrtimer_cancel(&rq->hrtick_timer);
}

/*
 * High-resolution timer tick.
 * Runs from hardirq context with interrupts disabled.
 */
static enum hrtimer_restart hrtick(struct hrtimer *timer)
{
	struct rq *rq = container_of(timer, struct rq, hrtick_timer);

	WARN_ON_ONCE(cpu_of(rq) != smp_processor_id());

1056
	raw_spin_lock(&rq->lock);
1057
	update_rq_clock(rq);
P
Peter Zijlstra 已提交
1058
	rq->curr->sched_class->task_tick(rq, rq->curr, 1);
1059
	raw_spin_unlock(&rq->lock);
P
Peter Zijlstra 已提交
1060 1061 1062 1063

	return HRTIMER_NORESTART;
}

1064
#ifdef CONFIG_SMP
1065 1066 1067 1068
/*
 * called from hardirq (IPI) context
 */
static void __hrtick_start(void *arg)
1069
{
1070
	struct rq *rq = arg;
1071

1072
	raw_spin_lock(&rq->lock);
1073 1074
	hrtimer_restart(&rq->hrtick_timer);
	rq->hrtick_csd_pending = 0;
1075
	raw_spin_unlock(&rq->lock);
1076 1077
}

1078 1079 1080 1081 1082 1083
/*
 * Called to set the hrtick timer state.
 *
 * called with rq->lock held and irqs disabled
 */
static void hrtick_start(struct rq *rq, u64 delay)
1084
{
1085 1086
	struct hrtimer *timer = &rq->hrtick_timer;
	ktime_t time = ktime_add_ns(timer->base->get_time(), delay);
1087

1088
	hrtimer_set_expires(timer, time);
1089 1090 1091 1092

	if (rq == this_rq()) {
		hrtimer_restart(timer);
	} else if (!rq->hrtick_csd_pending) {
1093
		__smp_call_function_single(cpu_of(rq), &rq->hrtick_csd, 0);
1094 1095
		rq->hrtick_csd_pending = 1;
	}
1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109
}

static int
hotplug_hrtick(struct notifier_block *nfb, unsigned long action, void *hcpu)
{
	int cpu = (int)(long)hcpu;

	switch (action) {
	case CPU_UP_CANCELED:
	case CPU_UP_CANCELED_FROZEN:
	case CPU_DOWN_PREPARE:
	case CPU_DOWN_PREPARE_FROZEN:
	case CPU_DEAD:
	case CPU_DEAD_FROZEN:
1110
		hrtick_clear(cpu_rq(cpu));
1111 1112 1113 1114 1115 1116
		return NOTIFY_OK;
	}

	return NOTIFY_DONE;
}

1117
static __init void init_hrtick(void)
1118 1119 1120
{
	hotcpu_notifier(hotplug_hrtick, 0);
}
1121 1122 1123 1124 1125 1126 1127 1128
#else
/*
 * Called to set the hrtick timer state.
 *
 * called with rq->lock held and irqs disabled
 */
static void hrtick_start(struct rq *rq, u64 delay)
{
1129
	__hrtimer_start_range_ns(&rq->hrtick_timer, ns_to_ktime(delay), 0,
1130
			HRTIMER_MODE_REL_PINNED, 0);
1131
}
1132

A
Andrew Morton 已提交
1133
static inline void init_hrtick(void)
P
Peter Zijlstra 已提交
1134 1135
{
}
1136
#endif /* CONFIG_SMP */
P
Peter Zijlstra 已提交
1137

1138
static void init_rq_hrtick(struct rq *rq)
P
Peter Zijlstra 已提交
1139
{
1140 1141
#ifdef CONFIG_SMP
	rq->hrtick_csd_pending = 0;
P
Peter Zijlstra 已提交
1142

1143 1144 1145 1146
	rq->hrtick_csd.flags = 0;
	rq->hrtick_csd.func = __hrtick_start;
	rq->hrtick_csd.info = rq;
#endif
P
Peter Zijlstra 已提交
1147

1148 1149
	hrtimer_init(&rq->hrtick_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
	rq->hrtick_timer.function = hrtick;
P
Peter Zijlstra 已提交
1150
}
A
Andrew Morton 已提交
1151
#else	/* CONFIG_SCHED_HRTICK */
P
Peter Zijlstra 已提交
1152 1153 1154 1155 1156 1157 1158 1159
static inline void hrtick_clear(struct rq *rq)
{
}

static inline void init_rq_hrtick(struct rq *rq)
{
}

1160 1161 1162
static inline void init_hrtick(void)
{
}
A
Andrew Morton 已提交
1163
#endif	/* CONFIG_SCHED_HRTICK */
P
Peter Zijlstra 已提交
1164

I
Ingo Molnar 已提交
1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177
/*
 * resched_task - mark a task 'to be rescheduled now'.
 *
 * On UP this means the setting of the need_resched flag, on SMP it
 * might also involve a cross-CPU call to trigger the scheduler on
 * the target CPU.
 */
#ifdef CONFIG_SMP

#ifndef tsk_is_polling
#define tsk_is_polling(t) test_tsk_thread_flag(t, TIF_POLLING_NRFLAG)
#endif

1178
static void resched_task(struct task_struct *p)
I
Ingo Molnar 已提交
1179 1180 1181
{
	int cpu;

1182
	assert_raw_spin_locked(&task_rq(p)->lock);
I
Ingo Molnar 已提交
1183

1184
	if (test_tsk_need_resched(p))
I
Ingo Molnar 已提交
1185 1186
		return;

1187
	set_tsk_need_resched(p);
I
Ingo Molnar 已提交
1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203

	cpu = task_cpu(p);
	if (cpu == smp_processor_id())
		return;

	/* NEED_RESCHED must be visible before we test polling */
	smp_mb();
	if (!tsk_is_polling(p))
		smp_send_reschedule(cpu);
}

static void resched_cpu(int cpu)
{
	struct rq *rq = cpu_rq(cpu);
	unsigned long flags;

1204
	if (!raw_spin_trylock_irqsave(&rq->lock, flags))
I
Ingo Molnar 已提交
1205 1206
		return;
	resched_task(cpu_curr(cpu));
1207
	raw_spin_unlock_irqrestore(&rq->lock, flags);
I
Ingo Molnar 已提交
1208
}
1209 1210

#ifdef CONFIG_NO_HZ
1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231
/*
 * In the semi idle case, use the nearest busy cpu for migrating timers
 * from an idle cpu.  This is good for power-savings.
 *
 * We don't do similar optimization for completely idle system, as
 * selecting an idle cpu will add more delays to the timers than intended
 * (as that cpu's timer base may not be uptodate wrt jiffies etc).
 */
int get_nohz_timer_target(void)
{
	int cpu = smp_processor_id();
	int i;
	struct sched_domain *sd;

	for_each_domain(cpu, sd) {
		for_each_cpu(i, sched_domain_span(sd))
			if (!idle_cpu(i))
				return i;
	}
	return cpu;
}
1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263
/*
 * When add_timer_on() enqueues a timer into the timer wheel of an
 * idle CPU then this timer might expire before the next timer event
 * which is scheduled to wake up that CPU. In case of a completely
 * idle system the next event might even be infinite time into the
 * future. wake_up_idle_cpu() ensures that the CPU is woken up and
 * leaves the inner idle loop so the newly added timer is taken into
 * account when the CPU goes back to idle and evaluates the timer
 * wheel for the next timer event.
 */
void wake_up_idle_cpu(int cpu)
{
	struct rq *rq = cpu_rq(cpu);

	if (cpu == smp_processor_id())
		return;

	/*
	 * This is safe, as this function is called with the timer
	 * wheel base lock of (cpu) held. When the CPU is on the way
	 * to idle and has not yet set rq->curr to idle then it will
	 * be serialized on the timer wheel base lock and take the new
	 * timer into account automatically.
	 */
	if (rq->curr != rq->idle)
		return;

	/*
	 * We can set TIF_RESCHED on the idle task of the other CPU
	 * lockless. The worst case is that the other CPU runs the
	 * idle task through an additional NOOP schedule()
	 */
1264
	set_tsk_need_resched(rq->idle);
1265 1266 1267 1268 1269 1270

	/* NEED_RESCHED must be visible before we test polling */
	smp_mb();
	if (!tsk_is_polling(rq->idle))
		smp_send_reschedule(cpu);
}
M
Mike Galbraith 已提交
1271

1272
#endif /* CONFIG_NO_HZ */
1273

1274 1275 1276 1277 1278 1279 1280 1281 1282 1283
static u64 sched_avg_period(void)
{
	return (u64)sysctl_sched_time_avg * NSEC_PER_MSEC / 2;
}

static void sched_avg_update(struct rq *rq)
{
	s64 period = sched_avg_period();

	while ((s64)(rq->clock - rq->age_stamp) > period) {
1284 1285 1286 1287 1288 1289
		/*
		 * Inline assembly required to prevent the compiler
		 * optimising this loop into a divmod call.
		 * See __iter_div_u64_rem() for another example of this.
		 */
		asm("" : "+rm" (rq->age_stamp));
1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300
		rq->age_stamp += period;
		rq->rt_avg /= 2;
	}
}

static void sched_rt_avg_update(struct rq *rq, u64 rt_delta)
{
	rq->rt_avg += rt_delta;
	sched_avg_update(rq);
}

1301
#else /* !CONFIG_SMP */
1302
static void resched_task(struct task_struct *p)
I
Ingo Molnar 已提交
1303
{
1304
	assert_raw_spin_locked(&task_rq(p)->lock);
1305
	set_tsk_need_resched(p);
I
Ingo Molnar 已提交
1306
}
1307 1308 1309 1310

static void sched_rt_avg_update(struct rq *rq, u64 rt_delta)
{
}
1311 1312 1313 1314

static void sched_avg_update(struct rq *rq)
{
}
1315
#endif /* CONFIG_SMP */
I
Ingo Molnar 已提交
1316

1317 1318 1319 1320 1321 1322 1323 1324
#if BITS_PER_LONG == 32
# define WMULT_CONST	(~0UL)
#else
# define WMULT_CONST	(1UL << 32)
#endif

#define WMULT_SHIFT	32

I
Ingo Molnar 已提交
1325 1326 1327
/*
 * Shift right and round:
 */
I
Ingo Molnar 已提交
1328
#define SRR(x, y) (((x) + (1UL << ((y) - 1))) >> (y))
I
Ingo Molnar 已提交
1329

1330 1331 1332
/*
 * delta *= weight / lw
 */
1333
static unsigned long
1334 1335 1336 1337 1338
calc_delta_mine(unsigned long delta_exec, unsigned long weight,
		struct load_weight *lw)
{
	u64 tmp;

1339 1340 1341 1342 1343 1344 1345
	if (!lw->inv_weight) {
		if (BITS_PER_LONG > 32 && unlikely(lw->weight >= WMULT_CONST))
			lw->inv_weight = 1;
		else
			lw->inv_weight = 1 + (WMULT_CONST-lw->weight/2)
				/ (lw->weight+1);
	}
1346 1347 1348 1349 1350

	tmp = (u64)delta_exec * weight;
	/*
	 * Check whether we'd overflow the 64-bit multiplication:
	 */
I
Ingo Molnar 已提交
1351
	if (unlikely(tmp > WMULT_CONST))
I
Ingo Molnar 已提交
1352
		tmp = SRR(SRR(tmp, WMULT_SHIFT/2) * lw->inv_weight,
I
Ingo Molnar 已提交
1353 1354
			WMULT_SHIFT/2);
	else
I
Ingo Molnar 已提交
1355
		tmp = SRR(tmp * lw->inv_weight, WMULT_SHIFT);
1356

1357
	return (unsigned long)min(tmp, (u64)(unsigned long)LONG_MAX);
1358 1359
}

1360
static inline void update_load_add(struct load_weight *lw, unsigned long inc)
1361 1362
{
	lw->weight += inc;
I
Ingo Molnar 已提交
1363
	lw->inv_weight = 0;
1364 1365
}

1366
static inline void update_load_sub(struct load_weight *lw, unsigned long dec)
1367 1368
{
	lw->weight -= dec;
I
Ingo Molnar 已提交
1369
	lw->inv_weight = 0;
1370 1371
}

1372 1373 1374 1375
/*
 * 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
I
Ingo Molnar 已提交
1376
 * scheduling class and "nice" value. For SCHED_NORMAL tasks this is just a
1377 1378 1379 1380
 * scaled version of the new time slice allocation that they receive on time
 * slice expiry etc.
 */

P
Peter Zijlstra 已提交
1381 1382
#define WEIGHT_IDLEPRIO                3
#define WMULT_IDLEPRIO         1431655765
I
Ingo Molnar 已提交
1383 1384 1385 1386 1387 1388 1389 1390 1391

/*
 * Nice levels are multiplicative, with a gentle 10% change for every
 * nice level changed. I.e. when a CPU-bound task goes from nice 0 to
 * nice 1, it will get ~10% less CPU time than another CPU-bound task
 * that remained on nice 0.
 *
 * The "10% effect" is relative and cumulative: from _any_ nice level,
 * if you go up 1 level, it's -10% CPU usage, if you go down 1 level
1392 1393 1394
 * it's +10% CPU usage. (to achieve that we use a multiplier of 1.25.
 * If a task goes up by ~10% and another task goes down by ~10% then
 * the relative distance between them is ~25%.)
I
Ingo Molnar 已提交
1395 1396
 */
static const int prio_to_weight[40] = {
1397 1398 1399 1400 1401 1402 1403 1404
 /* -20 */     88761,     71755,     56483,     46273,     36291,
 /* -15 */     29154,     23254,     18705,     14949,     11916,
 /* -10 */      9548,      7620,      6100,      4904,      3906,
 /*  -5 */      3121,      2501,      1991,      1586,      1277,
 /*   0 */      1024,       820,       655,       526,       423,
 /*   5 */       335,       272,       215,       172,       137,
 /*  10 */       110,        87,        70,        56,        45,
 /*  15 */        36,        29,        23,        18,        15,
I
Ingo Molnar 已提交
1405 1406
};

1407 1408 1409 1410 1411 1412 1413
/*
 * Inverse (2^32/x) values of the prio_to_weight[] array, precalculated.
 *
 * In cases where the weight does not change often, we can use the
 * precalculated inverse to speed up arithmetics by turning divisions
 * into multiplications:
 */
I
Ingo Molnar 已提交
1414
static const u32 prio_to_wmult[40] = {
1415 1416 1417 1418 1419 1420 1421 1422
 /* -20 */     48388,     59856,     76040,     92818,    118348,
 /* -15 */    147320,    184698,    229616,    287308,    360437,
 /* -10 */    449829,    563644,    704093,    875809,   1099582,
 /*  -5 */   1376151,   1717300,   2157191,   2708050,   3363326,
 /*   0 */   4194304,   5237765,   6557202,   8165337,  10153587,
 /*   5 */  12820798,  15790321,  19976592,  24970740,  31350126,
 /*  10 */  39045157,  49367440,  61356676,  76695844,  95443717,
 /*  15 */ 119304647, 148102320, 186737708, 238609294, 286331153,
I
Ingo Molnar 已提交
1423
};
1424

1425 1426 1427 1428 1429 1430 1431 1432
/* Time spent by the tasks of the cpu accounting group executing in ... */
enum cpuacct_stat_index {
	CPUACCT_STAT_USER,	/* ... user mode */
	CPUACCT_STAT_SYSTEM,	/* ... kernel mode */

	CPUACCT_STAT_NSTATS,
};

1433 1434
#ifdef CONFIG_CGROUP_CPUACCT
static void cpuacct_charge(struct task_struct *tsk, u64 cputime);
1435 1436
static void cpuacct_update_stats(struct task_struct *tsk,
		enum cpuacct_stat_index idx, cputime_t val);
1437 1438
#else
static inline void cpuacct_charge(struct task_struct *tsk, u64 cputime) {}
1439 1440
static inline void cpuacct_update_stats(struct task_struct *tsk,
		enum cpuacct_stat_index idx, cputime_t val) {}
1441 1442
#endif

1443 1444 1445 1446 1447 1448 1449 1450 1451 1452
static inline void inc_cpu_load(struct rq *rq, unsigned long load)
{
	update_load_add(&rq->load, load);
}

static inline void dec_cpu_load(struct rq *rq, unsigned long load)
{
	update_load_sub(&rq->load, load);
}

I
Ingo Molnar 已提交
1453
#if (defined(CONFIG_SMP) && defined(CONFIG_FAIR_GROUP_SCHED)) || defined(CONFIG_RT_GROUP_SCHED)
P
Peter Zijlstra 已提交
1454
typedef int (*tg_visitor)(struct task_group *, void *);
1455 1456 1457 1458 1459

/*
 * Iterate the full tree, calling @down when first entering a node and @up when
 * leaving it for the final time.
 */
P
Peter Zijlstra 已提交
1460
static int walk_tg_tree(tg_visitor down, tg_visitor up, void *data)
1461 1462
{
	struct task_group *parent, *child;
P
Peter Zijlstra 已提交
1463
	int ret;
1464 1465 1466 1467

	rcu_read_lock();
	parent = &root_task_group;
down:
P
Peter Zijlstra 已提交
1468 1469 1470
	ret = (*down)(parent, data);
	if (ret)
		goto out_unlock;
1471 1472 1473 1474 1475 1476 1477
	list_for_each_entry_rcu(child, &parent->children, siblings) {
		parent = child;
		goto down;

up:
		continue;
	}
P
Peter Zijlstra 已提交
1478 1479 1480
	ret = (*up)(parent, data);
	if (ret)
		goto out_unlock;
1481 1482 1483 1484 1485

	child = parent;
	parent = parent->parent;
	if (parent)
		goto up;
P
Peter Zijlstra 已提交
1486
out_unlock:
1487
	rcu_read_unlock();
P
Peter Zijlstra 已提交
1488 1489

	return ret;
1490 1491
}

P
Peter Zijlstra 已提交
1492 1493 1494
static int tg_nop(struct task_group *tg, void *data)
{
	return 0;
1495
}
P
Peter Zijlstra 已提交
1496 1497 1498
#endif

#ifdef CONFIG_SMP
P
Peter Zijlstra 已提交
1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537
/* Used instead of source_load when we know the type == 0 */
static unsigned long weighted_cpuload(const int cpu)
{
	return cpu_rq(cpu)->load.weight;
}

/*
 * Return a low guess at the load of a migration-source cpu weighted
 * according to the scheduling class and "nice" value.
 *
 * We want to under-estimate the load of migration sources, to
 * balance conservatively.
 */
static unsigned long source_load(int cpu, int type)
{
	struct rq *rq = cpu_rq(cpu);
	unsigned long total = weighted_cpuload(cpu);

	if (type == 0 || !sched_feat(LB_BIAS))
		return total;

	return min(rq->cpu_load[type-1], total);
}

/*
 * Return a high guess at the load of a migration-target cpu weighted
 * according to the scheduling class and "nice" value.
 */
static unsigned long target_load(int cpu, int type)
{
	struct rq *rq = cpu_rq(cpu);
	unsigned long total = weighted_cpuload(cpu);

	if (type == 0 || !sched_feat(LB_BIAS))
		return total;

	return max(rq->cpu_load[type-1], total);
}

1538 1539
static unsigned long power_of(int cpu)
{
1540
	return cpu_rq(cpu)->cpu_power;
1541 1542
}

P
Peter Zijlstra 已提交
1543 1544 1545 1546 1547
static int task_hot(struct task_struct *p, u64 now, struct sched_domain *sd);

static unsigned long cpu_avg_load_per_task(int cpu)
{
	struct rq *rq = cpu_rq(cpu);
1548
	unsigned long nr_running = ACCESS_ONCE(rq->nr_running);
P
Peter Zijlstra 已提交
1549

1550 1551
	if (nr_running)
		rq->avg_load_per_task = rq->load.weight / nr_running;
1552 1553
	else
		rq->avg_load_per_task = 0;
P
Peter Zijlstra 已提交
1554 1555 1556 1557 1558

	return rq->avg_load_per_task;
}

#ifdef CONFIG_FAIR_GROUP_SCHED
1559

1560
static __read_mostly unsigned long __percpu *update_shares_data;
1561

1562 1563 1564 1565 1566
static void __set_se_shares(struct sched_entity *se, unsigned long shares);

/*
 * Calculate and set the cpu's group shares.
 */
1567 1568 1569
static void update_group_shares_cpu(struct task_group *tg, int cpu,
				    unsigned long sd_shares,
				    unsigned long sd_rq_weight,
1570
				    unsigned long *usd_rq_weight)
1571
{
1572
	unsigned long shares, rq_weight;
P
Peter Zijlstra 已提交
1573
	int boost = 0;
1574

1575
	rq_weight = usd_rq_weight[cpu];
P
Peter Zijlstra 已提交
1576 1577 1578 1579
	if (!rq_weight) {
		boost = 1;
		rq_weight = NICE_0_LOAD;
	}
1580

1581
	/*
P
Peter Zijlstra 已提交
1582 1583 1584
	 *             \Sum_j shares_j * rq_weight_i
	 * shares_i =  -----------------------------
	 *                  \Sum_j rq_weight_j
1585
	 */
1586
	shares = (sd_shares * rq_weight) / sd_rq_weight;
1587
	shares = clamp_t(unsigned long, shares, MIN_SHARES, MAX_SHARES);
1588

1589 1590 1591 1592
	if (abs(shares - tg->se[cpu]->load.weight) >
			sysctl_sched_shares_thresh) {
		struct rq *rq = cpu_rq(cpu);
		unsigned long flags;
1593

1594
		raw_spin_lock_irqsave(&rq->lock, flags);
1595
		tg->cfs_rq[cpu]->rq_weight = boost ? 0 : rq_weight;
P
Peter Zijlstra 已提交
1596
		tg->cfs_rq[cpu]->shares = boost ? 0 : shares;
1597
		__set_se_shares(tg->se[cpu], shares);
1598
		raw_spin_unlock_irqrestore(&rq->lock, flags);
1599
	}
1600
}
1601 1602

/*
1603 1604 1605
 * Re-compute the task group their per cpu shares over the given domain.
 * This needs to be done in a bottom-up fashion because the rq weight of a
 * parent group depends on the shares of its child groups.
1606
 */
P
Peter Zijlstra 已提交
1607
static int tg_shares_up(struct task_group *tg, void *data)
1608
{
1609
	unsigned long weight, rq_weight = 0, sum_weight = 0, shares = 0;
1610
	unsigned long *usd_rq_weight;
P
Peter Zijlstra 已提交
1611
	struct sched_domain *sd = data;
1612
	unsigned long flags;
1613
	int i;
1614

1615 1616 1617 1618
	if (!tg->se[0])
		return 0;

	local_irq_save(flags);
1619
	usd_rq_weight = per_cpu_ptr(update_shares_data, smp_processor_id());
1620

1621
	for_each_cpu(i, sched_domain_span(sd)) {
1622
		weight = tg->cfs_rq[i]->load.weight;
1623
		usd_rq_weight[i] = weight;
1624

1625
		rq_weight += weight;
1626 1627 1628 1629 1630 1631 1632 1633
		/*
		 * If there are currently no tasks on the cpu pretend there
		 * is one of average load so that when a new task gets to
		 * run here it will not get delayed by group starvation.
		 */
		if (!weight)
			weight = NICE_0_LOAD;

1634
		sum_weight += weight;
1635
		shares += tg->cfs_rq[i]->shares;
1636 1637
	}

1638 1639 1640
	if (!rq_weight)
		rq_weight = sum_weight;

1641 1642 1643 1644 1645
	if ((!shares && rq_weight) || shares > tg->shares)
		shares = tg->shares;

	if (!sd->parent || !(sd->parent->flags & SD_LOAD_BALANCE))
		shares = tg->shares;
1646

1647
	for_each_cpu(i, sched_domain_span(sd))
1648
		update_group_shares_cpu(tg, i, shares, rq_weight, usd_rq_weight);
1649 1650

	local_irq_restore(flags);
P
Peter Zijlstra 已提交
1651 1652

	return 0;
1653 1654 1655
}

/*
1656 1657 1658
 * Compute the cpu's hierarchical load factor for each task group.
 * This needs to be done in a top-down fashion because the load of a child
 * group is a fraction of its parents load.
1659
 */
P
Peter Zijlstra 已提交
1660
static int tg_load_down(struct task_group *tg, void *data)
1661
{
1662
	unsigned long load;
P
Peter Zijlstra 已提交
1663
	long cpu = (long)data;
1664

1665 1666 1667 1668 1669 1670 1671
	if (!tg->parent) {
		load = cpu_rq(cpu)->load.weight;
	} else {
		load = tg->parent->cfs_rq[cpu]->h_load;
		load *= tg->cfs_rq[cpu]->shares;
		load /= tg->parent->cfs_rq[cpu]->load.weight + 1;
	}
1672

1673
	tg->cfs_rq[cpu]->h_load = load;
1674

P
Peter Zijlstra 已提交
1675
	return 0;
1676 1677
}

1678
static void update_shares(struct sched_domain *sd)
1679
{
1680 1681 1682 1683 1684 1685
	s64 elapsed;
	u64 now;

	if (root_task_group_empty())
		return;

1686
	now = local_clock();
1687
	elapsed = now - sd->last_update;
P
Peter Zijlstra 已提交
1688 1689 1690

	if (elapsed >= (s64)(u64)sysctl_sched_shares_ratelimit) {
		sd->last_update = now;
P
Peter Zijlstra 已提交
1691
		walk_tg_tree(tg_nop, tg_shares_up, sd);
P
Peter Zijlstra 已提交
1692
	}
1693 1694
}

P
Peter Zijlstra 已提交
1695
static void update_h_load(long cpu)
1696
{
P
Peter Zijlstra 已提交
1697
	walk_tg_tree(tg_load_down, tg_nop, (void *)cpu);
1698 1699 1700 1701
}

#else

1702
static inline void update_shares(struct sched_domain *sd)
1703 1704 1705
{
}

1706 1707
#endif

1708 1709
#ifdef CONFIG_PREEMPT

1710 1711
static void double_rq_lock(struct rq *rq1, struct rq *rq2);

1712
/*
1713 1714 1715 1716 1717 1718
 * 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.
1719
 */
1720 1721 1722 1723 1724
static inline int _double_lock_balance(struct rq *this_rq, struct rq *busiest)
	__releases(this_rq->lock)
	__acquires(busiest->lock)
	__acquires(this_rq->lock)
{
1725
	raw_spin_unlock(&this_rq->lock);
1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739
	double_rq_lock(this_rq, busiest);

	return 1;
}

#else
/*
 * Unfair double_lock_balance: Optimizes throughput at the expense of
 * latency by eliminating extra atomic operations when the locks are
 * already in proper order on entry.  This favors lower cpu-ids and will
 * grant the double lock to lower cpus over higher ids under contention,
 * regardless of entry order into the function.
 */
static int _double_lock_balance(struct rq *this_rq, struct rq *busiest)
1740 1741 1742 1743 1744 1745
	__releases(this_rq->lock)
	__acquires(busiest->lock)
	__acquires(this_rq->lock)
{
	int ret = 0;

1746
	if (unlikely(!raw_spin_trylock(&busiest->lock))) {
1747
		if (busiest < this_rq) {
1748 1749 1750 1751
			raw_spin_unlock(&this_rq->lock);
			raw_spin_lock(&busiest->lock);
			raw_spin_lock_nested(&this_rq->lock,
					      SINGLE_DEPTH_NESTING);
1752 1753
			ret = 1;
		} else
1754 1755
			raw_spin_lock_nested(&busiest->lock,
					      SINGLE_DEPTH_NESTING);
1756 1757 1758 1759
	}
	return ret;
}

1760 1761 1762 1763 1764 1765 1766 1767 1768
#endif /* CONFIG_PREEMPT */

/*
 * double_lock_balance - lock the busiest runqueue, this_rq is locked already.
 */
static int double_lock_balance(struct rq *this_rq, struct rq *busiest)
{
	if (unlikely(!irqs_disabled())) {
		/* printk() doesn't work good under rq->lock */
1769
		raw_spin_unlock(&this_rq->lock);
1770 1771 1772 1773 1774 1775
		BUG_ON(1);
	}

	return _double_lock_balance(this_rq, busiest);
}

1776 1777 1778
static inline void double_unlock_balance(struct rq *this_rq, struct rq *busiest)
	__releases(busiest->lock)
{
1779
	raw_spin_unlock(&busiest->lock);
1780 1781
	lock_set_subclass(&this_rq->lock.dep_map, 0, _RET_IP_);
}
1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824

/*
 * 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 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 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);
}

1825 1826
#endif

V
Vegard Nossum 已提交
1827
#ifdef CONFIG_FAIR_GROUP_SCHED
I
Ingo Molnar 已提交
1828 1829
static void cfs_rq_set_shares(struct cfs_rq *cfs_rq, unsigned long shares)
{
V
Vegard Nossum 已提交
1830
#ifdef CONFIG_SMP
I
Ingo Molnar 已提交
1831 1832 1833
	cfs_rq->shares = shares;
#endif
}
V
Vegard Nossum 已提交
1834
#endif
1835

1836
static void calc_load_account_idle(struct rq *this_rq);
1837
static void update_sysctl(void);
1838
static int get_update_sysctl_factor(void);
1839
static void update_cpu_load(struct rq *this_rq);
1840

P
Peter Zijlstra 已提交
1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853
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();
	task_thread_info(p)->cpu = cpu;
#endif
}
1854

1855
static const struct sched_class rt_sched_class;
I
Ingo Molnar 已提交
1856

1857
#define sched_class_highest (&stop_sched_class)
1858 1859
#define for_each_class(class) \
   for (class = sched_class_highest; class; class = class->next)
I
Ingo Molnar 已提交
1860

1861 1862
#include "sched_stats.h"

1863
static void inc_nr_running(struct rq *rq)
1864 1865 1866 1867
{
	rq->nr_running++;
}

1868
static void dec_nr_running(struct rq *rq)
1869 1870 1871 1872
{
	rq->nr_running--;
}

1873 1874
static void set_load_weight(struct task_struct *p)
{
I
Ingo Molnar 已提交
1875 1876 1877 1878 1879 1880 1881 1882
	/*
	 * SCHED_IDLE tasks get minimal weight:
	 */
	if (p->policy == SCHED_IDLE) {
		p->se.load.weight = WEIGHT_IDLEPRIO;
		p->se.load.inv_weight = WMULT_IDLEPRIO;
		return;
	}
1883

I
Ingo Molnar 已提交
1884 1885
	p->se.load.weight = prio_to_weight[p->static_prio - MAX_RT_PRIO];
	p->se.load.inv_weight = prio_to_wmult[p->static_prio - MAX_RT_PRIO];
1886 1887
}

1888
static void enqueue_task(struct rq *rq, struct task_struct *p, int flags)
1889
{
1890
	update_rq_clock(rq);
I
Ingo Molnar 已提交
1891
	sched_info_queued(p);
1892
	p->sched_class->enqueue_task(rq, p, flags);
I
Ingo Molnar 已提交
1893
	p->se.on_rq = 1;
1894 1895
}

1896
static void dequeue_task(struct rq *rq, struct task_struct *p, int flags)
1897
{
1898
	update_rq_clock(rq);
1899
	sched_info_dequeued(p);
1900
	p->sched_class->dequeue_task(rq, p, flags);
I
Ingo Molnar 已提交
1901
	p->se.on_rq = 0;
1902 1903
}

1904 1905 1906
/*
 * activate_task - move a task to the runqueue.
 */
1907
static void activate_task(struct rq *rq, struct task_struct *p, int flags)
1908 1909 1910 1911
{
	if (task_contributes_to_load(p))
		rq->nr_uninterruptible--;

1912
	enqueue_task(rq, p, flags);
1913 1914 1915 1916 1917 1918
	inc_nr_running(rq);
}

/*
 * deactivate_task - remove a task from the runqueue.
 */
1919
static void deactivate_task(struct rq *rq, struct task_struct *p, int flags)
1920 1921 1922 1923
{
	if (task_contributes_to_load(p))
		rq->nr_uninterruptible++;

1924
	dequeue_task(rq, p, flags);
1925 1926 1927
	dec_nr_running(rq);
}

1928 1929
#ifdef CONFIG_IRQ_TIME_ACCOUNTING

1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941
/*
 * There are no locks covering percpu hardirq/softirq time.
 * They are only modified in account_system_vtime, on corresponding CPU
 * with interrupts disabled. So, writes are safe.
 * They are read and saved off onto struct rq in update_rq_clock().
 * This may result in other CPU reading this CPU's irq time and can
 * race with irq/account_system_vtime on this CPU. We would either get old
 * or new value (or semi updated value on 32 bit) with a side effect of
 * accounting a slice of irq time to wrong task when irq is in progress
 * while we read rq->clock. That is a worthy compromise in place of having
 * locks on each irq in account_system_time.
 */
1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957
static DEFINE_PER_CPU(u64, cpu_hardirq_time);
static DEFINE_PER_CPU(u64, cpu_softirq_time);

static DEFINE_PER_CPU(u64, irq_start_time);
static int sched_clock_irqtime;

void enable_sched_clock_irqtime(void)
{
	sched_clock_irqtime = 1;
}

void disable_sched_clock_irqtime(void)
{
	sched_clock_irqtime = 0;
}

1958 1959 1960 1961 1962 1963 1964 1965
static u64 irq_time_cpu(int cpu)
{
	if (!sched_clock_irqtime)
		return 0;

	return per_cpu(cpu_softirq_time, cpu) + per_cpu(cpu_hardirq_time, cpu);
}

1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994
void account_system_vtime(struct task_struct *curr)
{
	unsigned long flags;
	int cpu;
	u64 now, delta;

	if (!sched_clock_irqtime)
		return;

	local_irq_save(flags);

	now = sched_clock();
	cpu = smp_processor_id();
	delta = now - per_cpu(irq_start_time, cpu);
	per_cpu(irq_start_time, cpu) = now;
	/*
	 * We do not account for softirq time from ksoftirqd here.
	 * We want to continue accounting softirq time to ksoftirqd thread
	 * in that case, so as not to confuse scheduler with a special task
	 * that do not consume any time, but still wants to run.
	 */
	if (hardirq_count())
		per_cpu(cpu_hardirq_time, cpu) += delta;
	else if (in_serving_softirq() && !(curr->flags & PF_KSOFTIRQD))
		per_cpu(cpu_softirq_time, cpu) += delta;

	local_irq_restore(flags);
}

1995 1996 1997 1998 1999 2000 2001 2002 2003
static void sched_irq_time_avg_update(struct rq *rq, u64 curr_irq_time)
{
	if (sched_clock_irqtime && sched_feat(NONIRQ_POWER)) {
		u64 delta_irq = curr_irq_time - rq->prev_irq_time;
		rq->prev_irq_time = curr_irq_time;
		sched_rt_avg_update(rq, delta_irq);
	}
}

2004 2005 2006 2007 2008 2009 2010
#else

static u64 irq_time_cpu(int cpu)
{
	return 0;
}

2011 2012
static void sched_irq_time_avg_update(struct rq *rq, u64 curr_irq_time) { }

2013 2014
#endif

2015 2016 2017
#include "sched_idletask.c"
#include "sched_fair.c"
#include "sched_rt.c"
2018
#include "sched_stoptask.c"
2019 2020 2021 2022
#ifdef CONFIG_SCHED_DEBUG
# include "sched_debug.c"
#endif

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
void sched_set_stop_task(int cpu, struct task_struct *stop)
{
	struct sched_param param = { .sched_priority = MAX_RT_PRIO - 1 };
	struct task_struct *old_stop = cpu_rq(cpu)->stop;

	if (stop) {
		/*
		 * Make it appear like a SCHED_FIFO task, its something
		 * userspace knows about and won't get confused about.
		 *
		 * Also, it will make PI more or less work without too
		 * much confusion -- but then, stop work should not
		 * rely on PI working anyway.
		 */
		sched_setscheduler_nocheck(stop, SCHED_FIFO, &param);

		stop->sched_class = &stop_sched_class;
	}

	cpu_rq(cpu)->stop = stop;

	if (old_stop) {
		/*
		 * Reset it back to a normal scheduling class so that
		 * it can die in pieces.
		 */
		old_stop->sched_class = &rt_sched_class;
	}
}

2053
/*
I
Ingo Molnar 已提交
2054
 * __normal_prio - return the priority that is based on the static prio
2055 2056 2057
 */
static inline int __normal_prio(struct task_struct *p)
{
I
Ingo Molnar 已提交
2058
	return p->static_prio;
2059 2060
}

2061 2062 2063 2064 2065 2066 2067
/*
 * Calculate the expected normal priority: i.e. priority
 * without taking RT-inheritance into account. Might be
 * boosted by interactivity modifiers. Changes upon fork,
 * setprio syscalls, and whenever the interactivity
 * estimator recalculates.
 */
2068
static inline int normal_prio(struct task_struct *p)
2069 2070 2071
{
	int prio;

2072
	if (task_has_rt_policy(p))
2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085
		prio = MAX_RT_PRIO-1 - p->rt_priority;
	else
		prio = __normal_prio(p);
	return prio;
}

/*
 * Calculate the current priority, i.e. the priority
 * taken into account by the scheduler. This value might
 * be boosted by RT tasks, or might be boosted by
 * interactivity modifiers. Will be RT if the task got
 * RT-boosted. If not then it returns p->normal_prio.
 */
2086
static int effective_prio(struct task_struct *p)
2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098
{
	p->normal_prio = normal_prio(p);
	/*
	 * If we are RT tasks or we were boosted to RT priority,
	 * keep the priority unchanged. Otherwise, update priority
	 * to the normal priority:
	 */
	if (!rt_prio(p->prio))
		return p->normal_prio;
	return p->prio;
}

L
Linus Torvalds 已提交
2099 2100 2101 2102
/**
 * task_curr - is this task currently executing on a CPU?
 * @p: the task in question.
 */
2103
inline int task_curr(const struct task_struct *p)
L
Linus Torvalds 已提交
2104 2105 2106 2107
{
	return cpu_curr(task_cpu(p)) == p;
}

2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119
static inline void check_class_changed(struct rq *rq, struct task_struct *p,
				       const struct sched_class *prev_class,
				       int oldprio, int running)
{
	if (prev_class != p->sched_class) {
		if (prev_class->switched_from)
			prev_class->switched_from(rq, p, running);
		p->sched_class->switched_to(rq, p, running);
	} else
		p->sched_class->prio_changed(rq, p, oldprio, running);
}

L
Linus Torvalds 已提交
2120
#ifdef CONFIG_SMP
2121 2122 2123
/*
 * Is this task likely cache-hot:
 */
2124
static int
2125 2126 2127 2128
task_hot(struct task_struct *p, u64 now, struct sched_domain *sd)
{
	s64 delta;

P
Peter Zijlstra 已提交
2129 2130 2131
	if (p->sched_class != &fair_sched_class)
		return 0;

2132 2133 2134
	if (unlikely(p->policy == SCHED_IDLE))
		return 0;

2135 2136 2137
	/*
	 * Buddy candidates are cache hot:
	 */
2138
	if (sched_feat(CACHE_HOT_BUDDY) && this_rq()->nr_running &&
P
Peter Zijlstra 已提交
2139 2140
			(&p->se == cfs_rq_of(&p->se)->next ||
			 &p->se == cfs_rq_of(&p->se)->last))
2141 2142
		return 1;

2143 2144 2145 2146 2147
	if (sysctl_sched_migration_cost == -1)
		return 1;
	if (sysctl_sched_migration_cost == 0)
		return 0;

2148 2149 2150 2151 2152
	delta = now - p->se.exec_start;

	return delta < (s64)sysctl_sched_migration_cost;
}

I
Ingo Molnar 已提交
2153
void set_task_cpu(struct task_struct *p, unsigned int new_cpu)
I
Ingo Molnar 已提交
2154
{
2155 2156 2157 2158 2159
#ifdef CONFIG_SCHED_DEBUG
	/*
	 * We should never call set_task_cpu() on a blocked task,
	 * ttwu() will sort out the placement.
	 */
P
Peter Zijlstra 已提交
2160 2161
	WARN_ON_ONCE(p->state != TASK_RUNNING && p->state != TASK_WAKING &&
			!(task_thread_info(p)->preempt_count & PREEMPT_ACTIVE));
2162 2163
#endif

2164
	trace_sched_migrate_task(p, new_cpu);
2165

2166 2167 2168 2169
	if (task_cpu(p) != new_cpu) {
		p->se.nr_migrations++;
		perf_sw_event(PERF_COUNT_SW_CPU_MIGRATIONS, 1, 1, NULL, 0);
	}
I
Ingo Molnar 已提交
2170 2171

	__set_task_cpu(p, new_cpu);
I
Ingo Molnar 已提交
2172 2173
}

2174
struct migration_arg {
2175
	struct task_struct *task;
L
Linus Torvalds 已提交
2176
	int dest_cpu;
2177
};
L
Linus Torvalds 已提交
2178

2179 2180
static int migration_cpu_stop(void *data);

L
Linus Torvalds 已提交
2181 2182 2183 2184
/*
 * The task's runqueue lock must be held.
 * Returns true if you have to wait for migration thread.
 */
2185
static bool migrate_task(struct task_struct *p, int dest_cpu)
L
Linus Torvalds 已提交
2186
{
2187
	struct rq *rq = task_rq(p);
L
Linus Torvalds 已提交
2188 2189 2190

	/*
	 * If the task is not on a runqueue (and not running), then
2191
	 * the next wake-up will properly place the task.
L
Linus Torvalds 已提交
2192
	 */
2193
	return p->se.on_rq || task_running(rq, p);
L
Linus Torvalds 已提交
2194 2195 2196 2197 2198
}

/*
 * wait_task_inactive - wait for a thread to unschedule.
 *
R
Roland McGrath 已提交
2199 2200 2201 2202 2203 2204 2205
 * If @match_state is nonzero, it's the @p->state value just checked and
 * not expected to change.  If it changes, i.e. @p might have woken up,
 * then return zero.  When we succeed in waiting for @p to be off its CPU,
 * we return a positive number (its total switch count).  If a second call
 * a short while later returns the same number, the caller can be sure that
 * @p has remained unscheduled the whole time.
 *
L
Linus Torvalds 已提交
2206 2207 2208 2209 2210 2211
 * The caller must ensure that the task *will* unschedule sometime soon,
 * else this function might spin for a *long* time. This function can't
 * be called with interrupts off, or it may introduce deadlock with
 * smp_call_function() if an IPI is sent by the same process we are
 * waiting to become inactive.
 */
R
Roland McGrath 已提交
2212
unsigned long wait_task_inactive(struct task_struct *p, long match_state)
L
Linus Torvalds 已提交
2213 2214
{
	unsigned long flags;
I
Ingo Molnar 已提交
2215
	int running, on_rq;
R
Roland McGrath 已提交
2216
	unsigned long ncsw;
2217
	struct rq *rq;
L
Linus Torvalds 已提交
2218

2219 2220 2221 2222 2223 2224 2225 2226
	for (;;) {
		/*
		 * We do the initial early heuristics without holding
		 * any task-queue locks at all. We'll only try to get
		 * the runqueue lock when things look like they will
		 * work out!
		 */
		rq = task_rq(p);
2227

2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238
		/*
		 * If the task is actively running on another CPU
		 * still, just relax and busy-wait without holding
		 * any locks.
		 *
		 * NOTE! Since we don't hold any locks, it's not
		 * even sure that "rq" stays as the right runqueue!
		 * But we don't care, since "task_running()" will
		 * return false if the runqueue has changed and p
		 * is actually now running somewhere else!
		 */
R
Roland McGrath 已提交
2239 2240 2241
		while (task_running(rq, p)) {
			if (match_state && unlikely(p->state != match_state))
				return 0;
2242
			cpu_relax();
R
Roland McGrath 已提交
2243
		}
2244

2245 2246 2247 2248 2249 2250
		/*
		 * Ok, time to look more closely! We need the rq
		 * lock now, to be *sure*. If we're wrong, we'll
		 * just go back and repeat.
		 */
		rq = task_rq_lock(p, &flags);
2251
		trace_sched_wait_task(p);
2252 2253
		running = task_running(rq, p);
		on_rq = p->se.on_rq;
R
Roland McGrath 已提交
2254
		ncsw = 0;
2255
		if (!match_state || p->state == match_state)
2256
			ncsw = p->nvcsw | LONG_MIN; /* sets MSB */
2257
		task_rq_unlock(rq, &flags);
2258

R
Roland McGrath 已提交
2259 2260 2261 2262 2263 2264
		/*
		 * If it changed from the expected state, bail out now.
		 */
		if (unlikely(!ncsw))
			break;

2265 2266 2267 2268 2269 2270 2271 2272 2273 2274
		/*
		 * Was it really running after all now that we
		 * checked with the proper locks actually held?
		 *
		 * Oops. Go back and try again..
		 */
		if (unlikely(running)) {
			cpu_relax();
			continue;
		}
2275

2276 2277 2278 2279 2280
		/*
		 * It's not enough that it's not actively running,
		 * it must be off the runqueue _entirely_, and not
		 * preempted!
		 *
2281
		 * So if it was still runnable (but just not actively
2282 2283 2284 2285 2286 2287 2288
		 * running right now), it's preempted, and we should
		 * yield - it could be a while.
		 */
		if (unlikely(on_rq)) {
			schedule_timeout_uninterruptible(1);
			continue;
		}
2289

2290 2291 2292 2293 2294 2295 2296
		/*
		 * Ahh, all good. It wasn't running, and it wasn't
		 * runnable, which means that it will never become
		 * running in the future either. We're all done!
		 */
		break;
	}
R
Roland McGrath 已提交
2297 2298

	return ncsw;
L
Linus Torvalds 已提交
2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313
}

/***
 * kick_process - kick a running thread to enter/exit the kernel
 * @p: the to-be-kicked thread
 *
 * Cause a process which is running on another CPU to enter
 * kernel-mode, without any delay. (to get signals handled.)
 *
 * NOTE: this function doesnt have to take the runqueue lock,
 * because all it wants to ensure is that the remote task enters
 * the kernel. If the IPI races and the task has been migrated
 * to another CPU then no harm is done and the purpose has been
 * achieved as well.
 */
2314
void kick_process(struct task_struct *p)
L
Linus Torvalds 已提交
2315 2316 2317 2318 2319 2320 2321 2322 2323
{
	int cpu;

	preempt_disable();
	cpu = task_cpu(p);
	if ((cpu != smp_processor_id()) && task_curr(p))
		smp_send_reschedule(cpu);
	preempt_enable();
}
R
Rusty Russell 已提交
2324
EXPORT_SYMBOL_GPL(kick_process);
N
Nick Piggin 已提交
2325
#endif /* CONFIG_SMP */
L
Linus Torvalds 已提交
2326

T
Thomas Gleixner 已提交
2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347
/**
 * task_oncpu_function_call - call a function on the cpu on which a task runs
 * @p:		the task to evaluate
 * @func:	the function to be called
 * @info:	the function call argument
 *
 * Calls the function @func when the task is currently running. This might
 * be on the current CPU, which just calls the function directly
 */
void task_oncpu_function_call(struct task_struct *p,
			      void (*func) (void *info), void *info)
{
	int cpu;

	preempt_disable();
	cpu = task_cpu(p);
	if (task_curr(p))
		smp_call_function_single(cpu, func, info, 1);
	preempt_enable();
}

2348
#ifdef CONFIG_SMP
2349 2350 2351
/*
 * ->cpus_allowed is protected by either TASK_WAKING or rq->lock held.
 */
2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367
static int select_fallback_rq(int cpu, struct task_struct *p)
{
	int dest_cpu;
	const struct cpumask *nodemask = cpumask_of_node(cpu_to_node(cpu));

	/* Look for allowed, online CPU in same node. */
	for_each_cpu_and(dest_cpu, nodemask, cpu_active_mask)
		if (cpumask_test_cpu(dest_cpu, &p->cpus_allowed))
			return dest_cpu;

	/* Any allowed, online CPU? */
	dest_cpu = cpumask_any_and(&p->cpus_allowed, cpu_active_mask);
	if (dest_cpu < nr_cpu_ids)
		return dest_cpu;

	/* No more Mr. Nice Guy. */
2368
	if (unlikely(dest_cpu >= nr_cpu_ids)) {
2369
		dest_cpu = cpuset_cpus_allowed_fallback(p);
2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384
		/*
		 * Don't tell them about moving exiting tasks or
		 * kernel threads (both mm NULL), since they never
		 * leave kernel.
		 */
		if (p->mm && printk_ratelimit()) {
			printk(KERN_INFO "process %d (%s) no "
			       "longer affine to cpu%d\n",
			       task_pid_nr(p), p->comm, cpu);
		}
	}

	return dest_cpu;
}

2385
/*
2386
 * The caller (fork, wakeup) owns TASK_WAKING, ->cpus_allowed is stable.
2387
 */
2388
static inline
2389
int select_task_rq(struct rq *rq, struct task_struct *p, int sd_flags, int wake_flags)
2390
{
2391
	int cpu = p->sched_class->select_task_rq(rq, p, sd_flags, wake_flags);
2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403

	/*
	 * In order not to call set_task_cpu() on a blocking task we need
	 * to rely on ttwu() to place the task on a valid ->cpus_allowed
	 * cpu.
	 *
	 * Since this is common to all placement strategies, this lives here.
	 *
	 * [ this allows ->select_task() to simply return task_cpu(p) and
	 *   not worry about this generic constraint ]
	 */
	if (unlikely(!cpumask_test_cpu(cpu, &p->cpus_allowed) ||
P
Peter Zijlstra 已提交
2404
		     !cpu_online(cpu)))
2405
		cpu = select_fallback_rq(task_cpu(p), p);
2406 2407

	return cpu;
2408
}
2409 2410 2411 2412 2413 2414

static void update_avg(u64 *avg, u64 sample)
{
	s64 diff = sample - *avg;
	*avg += diff >> 3;
}
2415 2416
#endif

T
Tejun Heo 已提交
2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455
static inline void ttwu_activate(struct task_struct *p, struct rq *rq,
				 bool is_sync, bool is_migrate, bool is_local,
				 unsigned long en_flags)
{
	schedstat_inc(p, se.statistics.nr_wakeups);
	if (is_sync)
		schedstat_inc(p, se.statistics.nr_wakeups_sync);
	if (is_migrate)
		schedstat_inc(p, se.statistics.nr_wakeups_migrate);
	if (is_local)
		schedstat_inc(p, se.statistics.nr_wakeups_local);
	else
		schedstat_inc(p, se.statistics.nr_wakeups_remote);

	activate_task(rq, p, en_flags);
}

static inline void ttwu_post_activation(struct task_struct *p, struct rq *rq,
					int wake_flags, bool success)
{
	trace_sched_wakeup(p, success);
	check_preempt_curr(rq, p, wake_flags);

	p->state = TASK_RUNNING;
#ifdef CONFIG_SMP
	if (p->sched_class->task_woken)
		p->sched_class->task_woken(rq, p);

	if (unlikely(rq->idle_stamp)) {
		u64 delta = rq->clock - rq->idle_stamp;
		u64 max = 2*sysctl_sched_migration_cost;

		if (delta > max)
			rq->avg_idle = max;
		else
			update_avg(&rq->avg_idle, delta);
		rq->idle_stamp = 0;
	}
#endif
T
Tejun Heo 已提交
2456 2457 2458
	/* if a worker is waking up, notify workqueue */
	if ((p->flags & PF_WQ_WORKER) && success)
		wq_worker_waking_up(p, cpu_of(rq));
T
Tejun Heo 已提交
2459 2460 2461
}

/**
L
Linus Torvalds 已提交
2462
 * try_to_wake_up - wake up a thread
T
Tejun Heo 已提交
2463
 * @p: the thread to be awakened
L
Linus Torvalds 已提交
2464
 * @state: the mask of task states that can be woken
T
Tejun Heo 已提交
2465
 * @wake_flags: wake modifier flags (WF_*)
L
Linus Torvalds 已提交
2466 2467 2468 2469 2470 2471 2472
 *
 * Put it on the run-queue if it's not already there. The "current"
 * thread is always on the run-queue (except when the actual
 * re-schedule is in progress), and as such you're allowed to do
 * the simpler "current->state = TASK_RUNNING" to mark yourself
 * runnable without the overhead of this.
 *
T
Tejun Heo 已提交
2473 2474
 * Returns %true if @p was woken up, %false if it was already running
 * or @state didn't match @p's state.
L
Linus Torvalds 已提交
2475
 */
P
Peter Zijlstra 已提交
2476 2477
static int try_to_wake_up(struct task_struct *p, unsigned int state,
			  int wake_flags)
L
Linus Torvalds 已提交
2478
{
2479
	int cpu, orig_cpu, this_cpu, success = 0;
L
Linus Torvalds 已提交
2480
	unsigned long flags;
2481
	unsigned long en_flags = ENQUEUE_WAKEUP;
2482
	struct rq *rq;
L
Linus Torvalds 已提交
2483

P
Peter Zijlstra 已提交
2484
	this_cpu = get_cpu();
P
Peter Zijlstra 已提交
2485

2486
	smp_wmb();
2487
	rq = task_rq_lock(p, &flags);
P
Peter Zijlstra 已提交
2488
	if (!(p->state & state))
L
Linus Torvalds 已提交
2489 2490
		goto out;

I
Ingo Molnar 已提交
2491
	if (p->se.on_rq)
L
Linus Torvalds 已提交
2492 2493 2494
		goto out_running;

	cpu = task_cpu(p);
2495
	orig_cpu = cpu;
L
Linus Torvalds 已提交
2496 2497 2498 2499 2500

#ifdef CONFIG_SMP
	if (unlikely(task_running(rq, p)))
		goto out_activate;

P
Peter Zijlstra 已提交
2501 2502 2503
	/*
	 * In order to handle concurrent wakeups and release the rq->lock
	 * we put the task in TASK_WAKING state.
2504 2505
	 *
	 * First fix up the nr_uninterruptible count:
P
Peter Zijlstra 已提交
2506
	 */
2507 2508 2509 2510 2511 2512
	if (task_contributes_to_load(p)) {
		if (likely(cpu_online(orig_cpu)))
			rq->nr_uninterruptible--;
		else
			this_rq()->nr_uninterruptible--;
	}
P
Peter Zijlstra 已提交
2513
	p->state = TASK_WAKING;
2514

2515
	if (p->sched_class->task_waking) {
2516
		p->sched_class->task_waking(rq, p);
2517 2518
		en_flags |= ENQUEUE_WAKING;
	}
2519

2520 2521
	cpu = select_task_rq(rq, p, SD_BALANCE_WAKE, wake_flags);
	if (cpu != orig_cpu)
2522
		set_task_cpu(p, cpu);
2523
	__task_rq_unlock(rq);
P
Peter Zijlstra 已提交
2524

2525 2526
	rq = cpu_rq(cpu);
	raw_spin_lock(&rq->lock);
2527

2528 2529 2530 2531 2532 2533 2534
	/*
	 * We migrated the task without holding either rq->lock, however
	 * since the task is not on the task list itself, nobody else
	 * will try and migrate the task, hence the rq should match the
	 * cpu we just moved it to.
	 */
	WARN_ON(task_cpu(p) != cpu);
P
Peter Zijlstra 已提交
2535
	WARN_ON(p->state != TASK_WAKING);
L
Linus Torvalds 已提交
2536

2537 2538 2539 2540 2541 2542 2543
#ifdef CONFIG_SCHEDSTATS
	schedstat_inc(rq, ttwu_count);
	if (cpu == this_cpu)
		schedstat_inc(rq, ttwu_local);
	else {
		struct sched_domain *sd;
		for_each_domain(this_cpu, sd) {
2544
			if (cpumask_test_cpu(cpu, sched_domain_span(sd))) {
2545 2546 2547 2548 2549
				schedstat_inc(sd, ttwu_wake_remote);
				break;
			}
		}
	}
2550
#endif /* CONFIG_SCHEDSTATS */
2551

L
Linus Torvalds 已提交
2552 2553
out_activate:
#endif /* CONFIG_SMP */
T
Tejun Heo 已提交
2554 2555
	ttwu_activate(p, rq, wake_flags & WF_SYNC, orig_cpu != cpu,
		      cpu == this_cpu, en_flags);
L
Linus Torvalds 已提交
2556 2557
	success = 1;
out_running:
T
Tejun Heo 已提交
2558
	ttwu_post_activation(p, rq, wake_flags, success);
L
Linus Torvalds 已提交
2559 2560
out:
	task_rq_unlock(rq, &flags);
P
Peter Zijlstra 已提交
2561
	put_cpu();
L
Linus Torvalds 已提交
2562 2563 2564 2565

	return success;
}

T
Tejun Heo 已提交
2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596
/**
 * try_to_wake_up_local - try to wake up a local task with rq lock held
 * @p: the thread to be awakened
 *
 * Put @p on the run-queue if it's not alredy there.  The caller must
 * ensure that this_rq() is locked, @p is bound to this_rq() and not
 * the current task.  this_rq() stays locked over invocation.
 */
static void try_to_wake_up_local(struct task_struct *p)
{
	struct rq *rq = task_rq(p);
	bool success = false;

	BUG_ON(rq != this_rq());
	BUG_ON(p == current);
	lockdep_assert_held(&rq->lock);

	if (!(p->state & TASK_NORMAL))
		return;

	if (!p->se.on_rq) {
		if (likely(!task_running(rq, p))) {
			schedstat_inc(rq, ttwu_count);
			schedstat_inc(rq, ttwu_local);
		}
		ttwu_activate(p, rq, false, false, true, ENQUEUE_WAKEUP);
		success = true;
	}
	ttwu_post_activation(p, rq, 0, success);
}

2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607
/**
 * wake_up_process - Wake up a specific process
 * @p: The process to be woken up.
 *
 * Attempt to wake up the nominated process and move it to the set of runnable
 * processes.  Returns 1 if the process was woken up, 0 if it was already
 * running.
 *
 * It may be assumed that this function implies a write memory barrier before
 * changing the task state if and only if any tasks are woken up.
 */
2608
int wake_up_process(struct task_struct *p)
L
Linus Torvalds 已提交
2609
{
2610
	return try_to_wake_up(p, TASK_ALL, 0);
L
Linus Torvalds 已提交
2611 2612 2613
}
EXPORT_SYMBOL(wake_up_process);

2614
int wake_up_state(struct task_struct *p, unsigned int state)
L
Linus Torvalds 已提交
2615 2616 2617 2618 2619 2620 2621
{
	return try_to_wake_up(p, state, 0);
}

/*
 * Perform scheduler related setup for a newly forked process p.
 * p is forked by current.
I
Ingo Molnar 已提交
2622 2623 2624 2625 2626 2627 2628
 *
 * __sched_fork() is basic setup used by init_idle() too:
 */
static void __sched_fork(struct task_struct *p)
{
	p->se.exec_start		= 0;
	p->se.sum_exec_runtime		= 0;
2629
	p->se.prev_sum_exec_runtime	= 0;
2630
	p->se.nr_migrations		= 0;
I
Ingo Molnar 已提交
2631 2632

#ifdef CONFIG_SCHEDSTATS
2633
	memset(&p->se.statistics, 0, sizeof(p->se.statistics));
I
Ingo Molnar 已提交
2634
#endif
N
Nick Piggin 已提交
2635

P
Peter Zijlstra 已提交
2636
	INIT_LIST_HEAD(&p->rt.run_list);
I
Ingo Molnar 已提交
2637
	p->se.on_rq = 0;
2638
	INIT_LIST_HEAD(&p->se.group_node);
N
Nick Piggin 已提交
2639

2640 2641 2642
#ifdef CONFIG_PREEMPT_NOTIFIERS
	INIT_HLIST_HEAD(&p->preempt_notifiers);
#endif
I
Ingo Molnar 已提交
2643 2644 2645 2646 2647 2648 2649 2650 2651 2652
}

/*
 * fork()/clone()-time setup:
 */
void sched_fork(struct task_struct *p, int clone_flags)
{
	int cpu = get_cpu();

	__sched_fork(p);
2653
	/*
2654
	 * We mark the process as running here. This guarantees that
2655 2656 2657
	 * nobody will actually run it, and a signal or other external
	 * event cannot wake it up and insert it on the runqueue either.
	 */
2658
	p->state = TASK_RUNNING;
I
Ingo Molnar 已提交
2659

2660 2661 2662 2663
	/*
	 * Revert to default priority/policy on fork if requested.
	 */
	if (unlikely(p->sched_reset_on_fork)) {
2664
		if (p->policy == SCHED_FIFO || p->policy == SCHED_RR) {
2665
			p->policy = SCHED_NORMAL;
2666 2667
			p->normal_prio = p->static_prio;
		}
2668

2669 2670
		if (PRIO_TO_NICE(p->static_prio) < 0) {
			p->static_prio = NICE_TO_PRIO(0);
2671
			p->normal_prio = p->static_prio;
2672 2673 2674
			set_load_weight(p);
		}

2675 2676 2677 2678 2679 2680
		/*
		 * We don't need the reset flag anymore after the fork. It has
		 * fulfilled its duty:
		 */
		p->sched_reset_on_fork = 0;
	}
2681

2682 2683 2684 2685 2686
	/*
	 * Make sure we do not leak PI boosting priority to the child.
	 */
	p->prio = current->normal_prio;

H
Hiroshi Shimamoto 已提交
2687 2688
	if (!rt_prio(p->prio))
		p->sched_class = &fair_sched_class;
2689

P
Peter Zijlstra 已提交
2690 2691 2692
	if (p->sched_class->task_fork)
		p->sched_class->task_fork(p);

2693 2694 2695 2696 2697 2698 2699 2700
	/*
	 * The child is not yet in the pid-hash so no cgroup attach races,
	 * and the cgroup is pinned to this child due to cgroup_fork()
	 * is ran before sched_fork().
	 *
	 * Silence PROVE_RCU.
	 */
	rcu_read_lock();
2701
	set_task_cpu(p, cpu);
2702
	rcu_read_unlock();
2703

2704
#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
I
Ingo Molnar 已提交
2705
	if (likely(sched_info_on()))
2706
		memset(&p->sched_info, 0, sizeof(p->sched_info));
L
Linus Torvalds 已提交
2707
#endif
2708
#if defined(CONFIG_SMP) && defined(__ARCH_WANT_UNLOCKED_CTXSW)
2709 2710
	p->oncpu = 0;
#endif
L
Linus Torvalds 已提交
2711
#ifdef CONFIG_PREEMPT
2712
	/* Want to start with kernel preemption disabled. */
A
Al Viro 已提交
2713
	task_thread_info(p)->preempt_count = 1;
L
Linus Torvalds 已提交
2714
#endif
2715 2716
	plist_node_init(&p->pushable_tasks, MAX_PRIO);

N
Nick Piggin 已提交
2717
	put_cpu();
L
Linus Torvalds 已提交
2718 2719 2720 2721 2722 2723 2724 2725 2726
}

/*
 * wake_up_new_task - wake up a newly created task for the first time.
 *
 * This function will do some initial scheduler statistics housekeeping
 * that must be done for every newly created context, then puts the task
 * on the runqueue and wakes it.
 */
2727
void wake_up_new_task(struct task_struct *p, unsigned long clone_flags)
L
Linus Torvalds 已提交
2728 2729
{
	unsigned long flags;
I
Ingo Molnar 已提交
2730
	struct rq *rq;
2731
	int cpu __maybe_unused = get_cpu();
2732 2733

#ifdef CONFIG_SMP
2734 2735 2736
	rq = task_rq_lock(p, &flags);
	p->state = TASK_WAKING;

2737 2738 2739 2740 2741
	/*
	 * Fork balancing, do it here and not earlier because:
	 *  - cpus_allowed can change in the fork path
	 *  - any previously selected cpu might disappear through hotplug
	 *
2742 2743
	 * We set TASK_WAKING so that select_task_rq() can drop rq->lock
	 * without people poking at ->cpus_allowed.
2744
	 */
2745
	cpu = select_task_rq(rq, p, SD_BALANCE_FORK, 0);
2746
	set_task_cpu(p, cpu);
L
Linus Torvalds 已提交
2747

2748
	p->state = TASK_RUNNING;
2749 2750 2751 2752
	task_rq_unlock(rq, &flags);
#endif

	rq = task_rq_lock(p, &flags);
P
Peter Zijlstra 已提交
2753
	activate_task(rq, p, 0);
2754
	trace_sched_wakeup_new(p, 1);
P
Peter Zijlstra 已提交
2755
	check_preempt_curr(rq, p, WF_FORK);
2756
#ifdef CONFIG_SMP
2757 2758
	if (p->sched_class->task_woken)
		p->sched_class->task_woken(rq, p);
2759
#endif
I
Ingo Molnar 已提交
2760
	task_rq_unlock(rq, &flags);
2761
	put_cpu();
L
Linus Torvalds 已提交
2762 2763
}

2764 2765 2766
#ifdef CONFIG_PREEMPT_NOTIFIERS

/**
2767
 * preempt_notifier_register - tell me when current is being preempted & rescheduled
R
Randy Dunlap 已提交
2768
 * @notifier: notifier struct to register
2769 2770 2771 2772 2773 2774 2775 2776 2777
 */
void preempt_notifier_register(struct preempt_notifier *notifier)
{
	hlist_add_head(&notifier->link, &current->preempt_notifiers);
}
EXPORT_SYMBOL_GPL(preempt_notifier_register);

/**
 * preempt_notifier_unregister - no longer interested in preemption notifications
R
Randy Dunlap 已提交
2778
 * @notifier: notifier struct to unregister
2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807
 *
 * This is safe to call from within a preemption notifier.
 */
void preempt_notifier_unregister(struct preempt_notifier *notifier)
{
	hlist_del(&notifier->link);
}
EXPORT_SYMBOL_GPL(preempt_notifier_unregister);

static void fire_sched_in_preempt_notifiers(struct task_struct *curr)
{
	struct preempt_notifier *notifier;
	struct hlist_node *node;

	hlist_for_each_entry(notifier, node, &curr->preempt_notifiers, link)
		notifier->ops->sched_in(notifier, raw_smp_processor_id());
}

static void
fire_sched_out_preempt_notifiers(struct task_struct *curr,
				 struct task_struct *next)
{
	struct preempt_notifier *notifier;
	struct hlist_node *node;

	hlist_for_each_entry(notifier, node, &curr->preempt_notifiers, link)
		notifier->ops->sched_out(notifier, next);
}

2808
#else /* !CONFIG_PREEMPT_NOTIFIERS */
2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819

static void fire_sched_in_preempt_notifiers(struct task_struct *curr)
{
}

static void
fire_sched_out_preempt_notifiers(struct task_struct *curr,
				 struct task_struct *next)
{
}

2820
#endif /* CONFIG_PREEMPT_NOTIFIERS */
2821

2822 2823 2824
/**
 * prepare_task_switch - prepare to switch tasks
 * @rq: the runqueue preparing to switch
R
Randy Dunlap 已提交
2825
 * @prev: the current task that is being switched out
2826 2827 2828 2829 2830 2831 2832 2833 2834
 * @next: the task we are going to switch to.
 *
 * This is called with the rq lock held and interrupts off. It must
 * be paired with a subsequent finish_task_switch after the context
 * switch.
 *
 * prepare_task_switch sets up locking and calls architecture specific
 * hooks.
 */
2835 2836 2837
static inline void
prepare_task_switch(struct rq *rq, struct task_struct *prev,
		    struct task_struct *next)
2838
{
2839
	fire_sched_out_preempt_notifiers(prev, next);
2840 2841 2842 2843
	prepare_lock_switch(rq, next);
	prepare_arch_switch(next);
}

L
Linus Torvalds 已提交
2844 2845
/**
 * finish_task_switch - clean up after a task-switch
2846
 * @rq: runqueue associated with task-switch
L
Linus Torvalds 已提交
2847 2848
 * @prev: the thread we just switched away from.
 *
2849 2850 2851 2852
 * finish_task_switch must be called after the context switch, paired
 * with a prepare_task_switch call before the context switch.
 * finish_task_switch will reconcile locking set up by prepare_task_switch,
 * and do any other architecture-specific cleanup actions.
L
Linus Torvalds 已提交
2853 2854
 *
 * Note that we may have delayed dropping an mm in context_switch(). If
I
Ingo Molnar 已提交
2855
 * so, we finish that here outside of the runqueue lock. (Doing it
L
Linus Torvalds 已提交
2856 2857 2858
 * with the lock held can cause deadlocks; see schedule() for
 * details.)
 */
A
Alexey Dobriyan 已提交
2859
static void finish_task_switch(struct rq *rq, struct task_struct *prev)
L
Linus Torvalds 已提交
2860 2861 2862
	__releases(rq->lock)
{
	struct mm_struct *mm = rq->prev_mm;
O
Oleg Nesterov 已提交
2863
	long prev_state;
L
Linus Torvalds 已提交
2864 2865 2866 2867 2868

	rq->prev_mm = NULL;

	/*
	 * A task struct has one reference for the use as "current".
2869
	 * If a task dies, then it sets TASK_DEAD in tsk->state and calls
O
Oleg Nesterov 已提交
2870 2871
	 * schedule one last time. The schedule call will never return, and
	 * the scheduled task must drop that reference.
2872
	 * The test for TASK_DEAD must occur while the runqueue locks are
L
Linus Torvalds 已提交
2873 2874 2875 2876 2877
	 * still held, otherwise prev could be scheduled on another cpu, die
	 * there before we look at prev->state, and then the reference would
	 * be dropped twice.
	 *		Manfred Spraul <manfred@colorfullife.com>
	 */
O
Oleg Nesterov 已提交
2878
	prev_state = prev->state;
2879
	finish_arch_switch(prev);
2880 2881 2882
#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
	local_irq_disable();
#endif /* __ARCH_WANT_INTERRUPTS_ON_CTXSW */
2883
	perf_event_task_sched_in(current);
2884 2885 2886
#ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
	local_irq_enable();
#endif /* __ARCH_WANT_INTERRUPTS_ON_CTXSW */
2887
	finish_lock_switch(rq, prev);
S
Steven Rostedt 已提交
2888

2889
	fire_sched_in_preempt_notifiers(current);
L
Linus Torvalds 已提交
2890 2891
	if (mm)
		mmdrop(mm);
2892
	if (unlikely(prev_state == TASK_DEAD)) {
2893 2894 2895
		/*
		 * Remove function-return probe instances associated with this
		 * task and put them back on the free list.
I
Ingo Molnar 已提交
2896
		 */
2897
		kprobe_flush_task(prev);
L
Linus Torvalds 已提交
2898
		put_task_struct(prev);
2899
	}
L
Linus Torvalds 已提交
2900 2901
}

2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916
#ifdef CONFIG_SMP

/* assumes rq->lock is held */
static inline void pre_schedule(struct rq *rq, struct task_struct *prev)
{
	if (prev->sched_class->pre_schedule)
		prev->sched_class->pre_schedule(rq, prev);
}

/* rq->lock is NOT held, but preemption is disabled */
static inline void post_schedule(struct rq *rq)
{
	if (rq->post_schedule) {
		unsigned long flags;

2917
		raw_spin_lock_irqsave(&rq->lock, flags);
2918 2919
		if (rq->curr->sched_class->post_schedule)
			rq->curr->sched_class->post_schedule(rq);
2920
		raw_spin_unlock_irqrestore(&rq->lock, flags);
2921 2922 2923 2924 2925 2926

		rq->post_schedule = 0;
	}
}

#else
2927

2928 2929 2930 2931 2932 2933
static inline void pre_schedule(struct rq *rq, struct task_struct *p)
{
}

static inline void post_schedule(struct rq *rq)
{
L
Linus Torvalds 已提交
2934 2935
}

2936 2937
#endif

L
Linus Torvalds 已提交
2938 2939 2940 2941
/**
 * schedule_tail - first thing a freshly forked thread must call.
 * @prev: the thread we just switched away from.
 */
2942
asmlinkage void schedule_tail(struct task_struct *prev)
L
Linus Torvalds 已提交
2943 2944
	__releases(rq->lock)
{
2945 2946
	struct rq *rq = this_rq();

2947
	finish_task_switch(rq, prev);
2948

2949 2950 2951 2952 2953
	/*
	 * FIXME: do we need to worry about rq being invalidated by the
	 * task_switch?
	 */
	post_schedule(rq);
2954

2955 2956 2957 2958
#ifdef __ARCH_WANT_UNLOCKED_CTXSW
	/* In this case, finish_task_switch does not reenable preemption */
	preempt_enable();
#endif
L
Linus Torvalds 已提交
2959
	if (current->set_child_tid)
2960
		put_user(task_pid_vnr(current), current->set_child_tid);
L
Linus Torvalds 已提交
2961 2962 2963 2964 2965 2966
}

/*
 * context_switch - switch to the new MM and the new
 * thread's register state.
 */
I
Ingo Molnar 已提交
2967
static inline void
2968
context_switch(struct rq *rq, struct task_struct *prev,
2969
	       struct task_struct *next)
L
Linus Torvalds 已提交
2970
{
I
Ingo Molnar 已提交
2971
	struct mm_struct *mm, *oldmm;
L
Linus Torvalds 已提交
2972

2973
	prepare_task_switch(rq, prev, next);
2974
	trace_sched_switch(prev, next);
I
Ingo Molnar 已提交
2975 2976
	mm = next->mm;
	oldmm = prev->active_mm;
2977 2978 2979 2980 2981
	/*
	 * For paravirt, this is coupled with an exit in switch_to to
	 * combine the page table reload and the switch backend into
	 * one hypercall.
	 */
2982
	arch_start_context_switch(prev);
2983

2984
	if (!mm) {
L
Linus Torvalds 已提交
2985 2986 2987 2988 2989 2990
		next->active_mm = oldmm;
		atomic_inc(&oldmm->mm_count);
		enter_lazy_tlb(oldmm, next);
	} else
		switch_mm(oldmm, mm, next);

2991
	if (!prev->mm) {
L
Linus Torvalds 已提交
2992 2993 2994
		prev->active_mm = NULL;
		rq->prev_mm = oldmm;
	}
2995 2996 2997 2998 2999 3000 3001
	/*
	 * Since the runqueue lock will be released by the next
	 * task (which is an invalid locking op but in the case
	 * of the scheduler it's an obvious special-case), so we
	 * do an early lockdep release here:
	 */
#ifndef __ARCH_WANT_UNLOCKED_CTXSW
3002
	spin_release(&rq->lock.dep_map, 1, _THIS_IP_);
3003
#endif
L
Linus Torvalds 已提交
3004 3005 3006 3007

	/* Here we just switch the register state and the stack. */
	switch_to(prev, next, prev);

I
Ingo Molnar 已提交
3008 3009 3010 3011 3012 3013 3014
	barrier();
	/*
	 * this_rq must be evaluated again because prev may have moved
	 * CPUs since it called schedule(), thus the 'rq' on its stack
	 * frame will be invalid.
	 */
	finish_task_switch(this_rq(), prev);
L
Linus Torvalds 已提交
3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031
}

/*
 * nr_running, nr_uninterruptible and nr_context_switches:
 *
 * externally visible scheduler statistics: current number of runnable
 * threads, current number of uninterruptible-sleeping threads, total
 * number of context switches performed since bootup.
 */
unsigned long nr_running(void)
{
	unsigned long i, sum = 0;

	for_each_online_cpu(i)
		sum += cpu_rq(i)->nr_running;

	return sum;
3032
}
L
Linus Torvalds 已提交
3033 3034

unsigned long nr_uninterruptible(void)
3035
{
L
Linus Torvalds 已提交
3036
	unsigned long i, sum = 0;
3037

3038
	for_each_possible_cpu(i)
L
Linus Torvalds 已提交
3039
		sum += cpu_rq(i)->nr_uninterruptible;
3040 3041

	/*
L
Linus Torvalds 已提交
3042 3043
	 * Since we read the counters lockless, it might be slightly
	 * inaccurate. Do not allow it to go below zero though:
3044
	 */
L
Linus Torvalds 已提交
3045 3046
	if (unlikely((long)sum < 0))
		sum = 0;
3047

L
Linus Torvalds 已提交
3048
	return sum;
3049 3050
}

L
Linus Torvalds 已提交
3051
unsigned long long nr_context_switches(void)
3052
{
3053 3054
	int i;
	unsigned long long sum = 0;
3055

3056
	for_each_possible_cpu(i)
L
Linus Torvalds 已提交
3057
		sum += cpu_rq(i)->nr_switches;
3058

L
Linus Torvalds 已提交
3059 3060
	return sum;
}
3061

L
Linus Torvalds 已提交
3062 3063 3064
unsigned long nr_iowait(void)
{
	unsigned long i, sum = 0;
3065

3066
	for_each_possible_cpu(i)
L
Linus Torvalds 已提交
3067
		sum += atomic_read(&cpu_rq(i)->nr_iowait);
3068

L
Linus Torvalds 已提交
3069 3070
	return sum;
}
3071

3072
unsigned long nr_iowait_cpu(int cpu)
3073
{
3074
	struct rq *this = cpu_rq(cpu);
3075 3076
	return atomic_read(&this->nr_iowait);
}
3077

3078 3079 3080 3081 3082
unsigned long this_cpu_load(void)
{
	struct rq *this = this_rq();
	return this->cpu_load[0];
}
3083

3084

3085 3086 3087 3088 3089
/* Variables and functions for calc_load */
static atomic_long_t calc_load_tasks;
static unsigned long calc_load_update;
unsigned long avenrun[3];
EXPORT_SYMBOL(avenrun);
3090

3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145
static long calc_load_fold_active(struct rq *this_rq)
{
	long nr_active, delta = 0;

	nr_active = this_rq->nr_running;
	nr_active += (long) this_rq->nr_uninterruptible;

	if (nr_active != this_rq->calc_load_active) {
		delta = nr_active - this_rq->calc_load_active;
		this_rq->calc_load_active = nr_active;
	}

	return delta;
}

#ifdef CONFIG_NO_HZ
/*
 * For NO_HZ we delay the active fold to the next LOAD_FREQ update.
 *
 * When making the ILB scale, we should try to pull this in as well.
 */
static atomic_long_t calc_load_tasks_idle;

static void calc_load_account_idle(struct rq *this_rq)
{
	long delta;

	delta = calc_load_fold_active(this_rq);
	if (delta)
		atomic_long_add(delta, &calc_load_tasks_idle);
}

static long calc_load_fold_idle(void)
{
	long delta = 0;

	/*
	 * Its got a race, we don't care...
	 */
	if (atomic_long_read(&calc_load_tasks_idle))
		delta = atomic_long_xchg(&calc_load_tasks_idle, 0);

	return delta;
}
#else
static void calc_load_account_idle(struct rq *this_rq)
{
}

static inline long calc_load_fold_idle(void)
{
	return 0;
}
#endif

3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158
/**
 * get_avenrun - get the load average array
 * @loads:	pointer to dest load array
 * @offset:	offset to add
 * @shift:	shift count to shift the result left
 *
 * These values are estimates at best, so no need for locking.
 */
void get_avenrun(unsigned long *loads, unsigned long offset, int shift)
{
	loads[0] = (avenrun[0] + offset) << shift;
	loads[1] = (avenrun[1] + offset) << shift;
	loads[2] = (avenrun[2] + offset) << shift;
3159 3160
}

3161 3162
static unsigned long
calc_load(unsigned long load, unsigned long exp, unsigned long active)
3163
{
3164 3165 3166 3167
	load *= exp;
	load += active * (FIXED_1 - exp);
	return load >> FSHIFT;
}
3168 3169

/*
3170 3171
 * calc_load - update the avenrun load estimates 10 ticks after the
 * CPUs have updated calc_load_tasks.
3172
 */
3173
void calc_global_load(void)
3174
{
3175 3176
	unsigned long upd = calc_load_update + 10;
	long active;
L
Linus Torvalds 已提交
3177

3178 3179
	if (time_before(jiffies, upd))
		return;
L
Linus Torvalds 已提交
3180

3181 3182
	active = atomic_long_read(&calc_load_tasks);
	active = active > 0 ? active * FIXED_1 : 0;
L
Linus Torvalds 已提交
3183

3184 3185 3186
	avenrun[0] = calc_load(avenrun[0], EXP_1, active);
	avenrun[1] = calc_load(avenrun[1], EXP_5, active);
	avenrun[2] = calc_load(avenrun[2], EXP_15, active);
I
Ingo Molnar 已提交
3187

3188 3189
	calc_load_update += LOAD_FREQ;
}
L
Linus Torvalds 已提交
3190

3191
/*
3192 3193
 * Called from update_cpu_load() to periodically update this CPU's
 * active count.
3194 3195 3196
 */
static void calc_load_account_active(struct rq *this_rq)
{
3197
	long delta;
3198

3199 3200
	if (time_before(jiffies, this_rq->calc_load_update))
		return;
3201

3202 3203 3204
	delta  = calc_load_fold_active(this_rq);
	delta += calc_load_fold_idle();
	if (delta)
3205
		atomic_long_add(delta, &calc_load_tasks);
3206 3207

	this_rq->calc_load_update += LOAD_FREQ;
3208 3209
}

3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276
/*
 * The exact cpuload at various idx values, calculated at every tick would be
 * load = (2^idx - 1) / 2^idx * load + 1 / 2^idx * cur_load
 *
 * If a cpu misses updates for n-1 ticks (as it was idle) and update gets called
 * on nth tick when cpu may be busy, then we have:
 * load = ((2^idx - 1) / 2^idx)^(n-1) * load
 * load = (2^idx - 1) / 2^idx) * load + 1 / 2^idx * cur_load
 *
 * decay_load_missed() below does efficient calculation of
 * load = ((2^idx - 1) / 2^idx)^(n-1) * load
 * avoiding 0..n-1 loop doing load = ((2^idx - 1) / 2^idx) * load
 *
 * The calculation is approximated on a 128 point scale.
 * degrade_zero_ticks is the number of ticks after which load at any
 * particular idx is approximated to be zero.
 * degrade_factor is a precomputed table, a row for each load idx.
 * Each column corresponds to degradation factor for a power of two ticks,
 * based on 128 point scale.
 * Example:
 * row 2, col 3 (=12) says that the degradation at load idx 2 after
 * 8 ticks is 12/128 (which is an approximation of exact factor 3^8/4^8).
 *
 * With this power of 2 load factors, we can degrade the load n times
 * by looking at 1 bits in n and doing as many mult/shift instead of
 * n mult/shifts needed by the exact degradation.
 */
#define DEGRADE_SHIFT		7
static const unsigned char
		degrade_zero_ticks[CPU_LOAD_IDX_MAX] = {0, 8, 32, 64, 128};
static const unsigned char
		degrade_factor[CPU_LOAD_IDX_MAX][DEGRADE_SHIFT + 1] = {
					{0, 0, 0, 0, 0, 0, 0, 0},
					{64, 32, 8, 0, 0, 0, 0, 0},
					{96, 72, 40, 12, 1, 0, 0},
					{112, 98, 75, 43, 15, 1, 0},
					{120, 112, 98, 76, 45, 16, 2} };

/*
 * Update cpu_load for any missed ticks, due to tickless idle. The backlog
 * would be when CPU is idle and so we just decay the old load without
 * adding any new load.
 */
static unsigned long
decay_load_missed(unsigned long load, unsigned long missed_updates, int idx)
{
	int j = 0;

	if (!missed_updates)
		return load;

	if (missed_updates >= degrade_zero_ticks[idx])
		return 0;

	if (idx == 1)
		return load >> missed_updates;

	while (missed_updates) {
		if (missed_updates % 2)
			load = (load * degrade_factor[idx][j]) >> DEGRADE_SHIFT;

		missed_updates >>= 1;
		j++;
	}
	return load;
}

3277
/*
I
Ingo Molnar 已提交
3278
 * Update rq->cpu_load[] statistics. This function is usually called every
3279 3280
 * scheduler tick (TICK_NSEC). With tickless idle this will not be called
 * every tick. We fix it up based on jiffies.
3281
 */
I
Ingo Molnar 已提交
3282
static void update_cpu_load(struct rq *this_rq)
3283
{
3284
	unsigned long this_load = this_rq->load.weight;
3285 3286
	unsigned long curr_jiffies = jiffies;
	unsigned long pending_updates;
I
Ingo Molnar 已提交
3287
	int i, scale;
3288

I
Ingo Molnar 已提交
3289
	this_rq->nr_load_updates++;
3290

3291 3292 3293 3294 3295 3296 3297
	/* Avoid repeated calls on same jiffy, when moving in and out of idle */
	if (curr_jiffies == this_rq->last_load_update_tick)
		return;

	pending_updates = curr_jiffies - this_rq->last_load_update_tick;
	this_rq->last_load_update_tick = curr_jiffies;

I
Ingo Molnar 已提交
3298
	/* Update our load: */
3299 3300
	this_rq->cpu_load[0] = this_load; /* Fasttrack for idx 0 */
	for (i = 1, scale = 2; i < CPU_LOAD_IDX_MAX; i++, scale += scale) {
I
Ingo Molnar 已提交
3301
		unsigned long old_load, new_load;
3302

I
Ingo Molnar 已提交
3303
		/* scale is effectively 1 << i now, and >> i divides by scale */
3304

I
Ingo Molnar 已提交
3305
		old_load = this_rq->cpu_load[i];
3306
		old_load = decay_load_missed(old_load, pending_updates - 1, i);
I
Ingo Molnar 已提交
3307
		new_load = this_load;
I
Ingo Molnar 已提交
3308 3309 3310 3311 3312 3313
		/*
		 * Round up the averaging division if load is increasing. This
		 * prevents us from getting stuck on 9 if the load is 10, for
		 * example.
		 */
		if (new_load > old_load)
3314 3315 3316
			new_load += scale - 1;

		this_rq->cpu_load[i] = (old_load * (scale - 1) + new_load) >> i;
I
Ingo Molnar 已提交
3317
	}
3318 3319

	sched_avg_update(this_rq);
3320 3321 3322 3323 3324
}

static void update_cpu_load_active(struct rq *this_rq)
{
	update_cpu_load(this_rq);
3325

3326
	calc_load_account_active(this_rq);
3327 3328
}

I
Ingo Molnar 已提交
3329
#ifdef CONFIG_SMP
3330

3331
/*
P
Peter Zijlstra 已提交
3332 3333
 * sched_exec - execve() is a valuable balancing opportunity, because at
 * this point the task has the smallest effective memory and cache footprint.
3334
 */
P
Peter Zijlstra 已提交
3335
void sched_exec(void)
3336
{
P
Peter Zijlstra 已提交
3337
	struct task_struct *p = current;
L
Linus Torvalds 已提交
3338
	unsigned long flags;
3339
	struct rq *rq;
3340
	int dest_cpu;
3341

L
Linus Torvalds 已提交
3342
	rq = task_rq_lock(p, &flags);
3343 3344 3345
	dest_cpu = p->sched_class->select_task_rq(rq, p, SD_BALANCE_EXEC, 0);
	if (dest_cpu == smp_processor_id())
		goto unlock;
P
Peter Zijlstra 已提交
3346

3347
	/*
P
Peter Zijlstra 已提交
3348
	 * select_task_rq() can race against ->cpus_allowed
3349
	 */
3350
	if (cpumask_test_cpu(dest_cpu, &p->cpus_allowed) &&
3351 3352
	    likely(cpu_active(dest_cpu)) && migrate_task(p, dest_cpu)) {
		struct migration_arg arg = { p, dest_cpu };
3353

L
Linus Torvalds 已提交
3354
		task_rq_unlock(rq, &flags);
3355
		stop_one_cpu(cpu_of(rq), migration_cpu_stop, &arg);
L
Linus Torvalds 已提交
3356 3357
		return;
	}
3358
unlock:
L
Linus Torvalds 已提交
3359 3360
	task_rq_unlock(rq, &flags);
}
I
Ingo Molnar 已提交
3361

L
Linus Torvalds 已提交
3362 3363 3364 3365 3366 3367 3368
#endif

DEFINE_PER_CPU(struct kernel_stat, kstat);

EXPORT_PER_CPU_SYMBOL(kstat);

/*
3369
 * Return any ns on the sched_clock that have not yet been accounted in
3370
 * @p in case that task is currently running.
3371 3372
 *
 * Called with task_rq_lock() held on @rq.
L
Linus Torvalds 已提交
3373
 */
3374 3375 3376 3377 3378 3379
static u64 do_task_delta_exec(struct task_struct *p, struct rq *rq)
{
	u64 ns = 0;

	if (task_current(rq, p)) {
		update_rq_clock(rq);
3380
		ns = rq->clock_task - p->se.exec_start;
3381 3382 3383 3384 3385 3386 3387
		if ((s64)ns < 0)
			ns = 0;
	}

	return ns;
}

3388
unsigned long long task_delta_exec(struct task_struct *p)
L
Linus Torvalds 已提交
3389 3390
{
	unsigned long flags;
3391
	struct rq *rq;
3392
	u64 ns = 0;
3393

3394
	rq = task_rq_lock(p, &flags);
3395 3396
	ns = do_task_delta_exec(p, rq);
	task_rq_unlock(rq, &flags);
3397

3398 3399
	return ns;
}
3400

3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417
/*
 * Return accounted runtime for the task.
 * In case the task is currently running, return the runtime plus current's
 * pending runtime that have not been accounted yet.
 */
unsigned long long task_sched_runtime(struct task_struct *p)
{
	unsigned long flags;
	struct rq *rq;
	u64 ns = 0;

	rq = task_rq_lock(p, &flags);
	ns = p->se.sum_exec_runtime + do_task_delta_exec(p, rq);
	task_rq_unlock(rq, &flags);

	return ns;
}
3418

3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437
/*
 * Return sum_exec_runtime for the thread group.
 * In case the task is currently running, return the sum plus current's
 * pending runtime that have not been accounted yet.
 *
 * Note that the thread group might have other running tasks as well,
 * so the return value not includes other pending runtime that other
 * running tasks might have.
 */
unsigned long long thread_group_sched_runtime(struct task_struct *p)
{
	struct task_cputime totals;
	unsigned long flags;
	struct rq *rq;
	u64 ns;

	rq = task_rq_lock(p, &flags);
	thread_group_cputime(p, &totals);
	ns = totals.sum_exec_runtime + do_task_delta_exec(p, rq);
3438
	task_rq_unlock(rq, &flags);
3439

L
Linus Torvalds 已提交
3440 3441 3442 3443 3444 3445 3446
	return ns;
}

/*
 * Account user cpu time to a process.
 * @p: the process that the cpu time gets accounted to
 * @cputime: the cpu time spent in user space since the last update
3447
 * @cputime_scaled: cputime scaled by cpu frequency
L
Linus Torvalds 已提交
3448
 */
3449 3450
void account_user_time(struct task_struct *p, cputime_t cputime,
		       cputime_t cputime_scaled)
L
Linus Torvalds 已提交
3451 3452 3453 3454
{
	struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat;
	cputime64_t tmp;

3455
	/* Add user time to process. */
L
Linus Torvalds 已提交
3456
	p->utime = cputime_add(p->utime, cputime);
3457
	p->utimescaled = cputime_add(p->utimescaled, cputime_scaled);
3458
	account_group_user_time(p, cputime);
L
Linus Torvalds 已提交
3459 3460 3461 3462 3463 3464 3465

	/* Add user time to cpustat. */
	tmp = cputime_to_cputime64(cputime);
	if (TASK_NICE(p) > 0)
		cpustat->nice = cputime64_add(cpustat->nice, tmp);
	else
		cpustat->user = cputime64_add(cpustat->user, tmp);
3466 3467

	cpuacct_update_stats(p, CPUACCT_STAT_USER, cputime);
3468 3469
	/* Account for user time used */
	acct_update_integrals(p);
L
Linus Torvalds 已提交
3470 3471
}

3472 3473 3474 3475
/*
 * Account guest cpu time to a process.
 * @p: the process that the cpu time gets accounted to
 * @cputime: the cpu time spent in virtual machine since the last update
3476
 * @cputime_scaled: cputime scaled by cpu frequency
3477
 */
3478 3479
static void account_guest_time(struct task_struct *p, cputime_t cputime,
			       cputime_t cputime_scaled)
3480 3481 3482 3483 3484 3485
{
	cputime64_t tmp;
	struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat;

	tmp = cputime_to_cputime64(cputime);

3486
	/* Add guest time to process. */
3487
	p->utime = cputime_add(p->utime, cputime);
3488
	p->utimescaled = cputime_add(p->utimescaled, cputime_scaled);
3489
	account_group_user_time(p, cputime);
3490 3491
	p->gtime = cputime_add(p->gtime, cputime);

3492
	/* Add guest time to cpustat. */
3493 3494 3495 3496 3497 3498 3499
	if (TASK_NICE(p) > 0) {
		cpustat->nice = cputime64_add(cpustat->nice, tmp);
		cpustat->guest_nice = cputime64_add(cpustat->guest_nice, tmp);
	} else {
		cpustat->user = cputime64_add(cpustat->user, tmp);
		cpustat->guest = cputime64_add(cpustat->guest, tmp);
	}
3500 3501
}

L
Linus Torvalds 已提交
3502 3503 3504 3505 3506
/*
 * Account system cpu time to a process.
 * @p: the process that the cpu time gets accounted to
 * @hardirq_offset: the offset to subtract from hardirq_count()
 * @cputime: the cpu time spent in kernel space since the last update
3507
 * @cputime_scaled: cputime scaled by cpu frequency
L
Linus Torvalds 已提交
3508 3509
 */
void account_system_time(struct task_struct *p, int hardirq_offset,
3510
			 cputime_t cputime, cputime_t cputime_scaled)
L
Linus Torvalds 已提交
3511 3512 3513 3514
{
	struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat;
	cputime64_t tmp;

3515
	if ((p->flags & PF_VCPU) && (irq_count() - hardirq_offset == 0)) {
3516
		account_guest_time(p, cputime, cputime_scaled);
3517 3518
		return;
	}
3519

3520
	/* Add system time to process. */
L
Linus Torvalds 已提交
3521
	p->stime = cputime_add(p->stime, cputime);
3522
	p->stimescaled = cputime_add(p->stimescaled, cputime_scaled);
3523
	account_group_system_time(p, cputime);
L
Linus Torvalds 已提交
3524 3525 3526 3527 3528

	/* Add system time to cpustat. */
	tmp = cputime_to_cputime64(cputime);
	if (hardirq_count() - hardirq_offset)
		cpustat->irq = cputime64_add(cpustat->irq, tmp);
3529
	else if (in_serving_softirq())
L
Linus Torvalds 已提交
3530 3531
		cpustat->softirq = cputime64_add(cpustat->softirq, tmp);
	else
3532 3533
		cpustat->system = cputime64_add(cpustat->system, tmp);

3534 3535
	cpuacct_update_stats(p, CPUACCT_STAT_SYSTEM, cputime);

L
Linus Torvalds 已提交
3536 3537 3538 3539
	/* Account for system time used */
	acct_update_integrals(p);
}

3540
/*
L
Linus Torvalds 已提交
3541 3542
 * Account for involuntary wait time.
 * @steal: the cpu time spent in involuntary wait
3543
 */
3544
void account_steal_time(cputime_t cputime)
3545
{
3546 3547 3548 3549
	struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat;
	cputime64_t cputime64 = cputime_to_cputime64(cputime);

	cpustat->steal = cputime64_add(cpustat->steal, cputime64);
3550 3551
}

L
Linus Torvalds 已提交
3552
/*
3553 3554
 * Account for idle time.
 * @cputime: the cpu time spent in idle wait
L
Linus Torvalds 已提交
3555
 */
3556
void account_idle_time(cputime_t cputime)
L
Linus Torvalds 已提交
3557 3558
{
	struct cpu_usage_stat *cpustat = &kstat_this_cpu.cpustat;
3559
	cputime64_t cputime64 = cputime_to_cputime64(cputime);
3560
	struct rq *rq = this_rq();
L
Linus Torvalds 已提交
3561

3562 3563 3564 3565
	if (atomic_read(&rq->nr_iowait) > 0)
		cpustat->iowait = cputime64_add(cpustat->iowait, cputime64);
	else
		cpustat->idle = cputime64_add(cpustat->idle, cputime64);
L
Linus Torvalds 已提交
3566 3567
}

3568 3569 3570 3571 3572 3573 3574 3575 3576
#ifndef CONFIG_VIRT_CPU_ACCOUNTING

/*
 * Account a single tick of cpu time.
 * @p: the process that the cpu time gets accounted to
 * @user_tick: indicates if the tick is a user or a system tick
 */
void account_process_tick(struct task_struct *p, int user_tick)
{
3577
	cputime_t one_jiffy_scaled = cputime_to_scaled(cputime_one_jiffy);
3578 3579 3580
	struct rq *rq = this_rq();

	if (user_tick)
3581
		account_user_time(p, cputime_one_jiffy, one_jiffy_scaled);
3582
	else if ((p != rq->idle) || (irq_count() != HARDIRQ_OFFSET))
3583
		account_system_time(p, HARDIRQ_OFFSET, cputime_one_jiffy,
3584 3585
				    one_jiffy_scaled);
	else
3586
		account_idle_time(cputime_one_jiffy);
3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605
}

/*
 * Account multiple ticks of steal time.
 * @p: the process from which the cpu time has been stolen
 * @ticks: number of stolen ticks
 */
void account_steal_ticks(unsigned long ticks)
{
	account_steal_time(jiffies_to_cputime(ticks));
}

/*
 * Account multiple ticks of idle time.
 * @ticks: number of stolen ticks
 */
void account_idle_ticks(unsigned long ticks)
{
	account_idle_time(jiffies_to_cputime(ticks));
L
Linus Torvalds 已提交
3606 3607
}

3608 3609
#endif

3610 3611 3612 3613
/*
 * Use precise platform statistics if available:
 */
#ifdef CONFIG_VIRT_CPU_ACCOUNTING
3614
void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
3615
{
3616 3617
	*ut = p->utime;
	*st = p->stime;
3618 3619
}

3620
void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
3621
{
3622 3623 3624 3625 3626 3627
	struct task_cputime cputime;

	thread_group_cputime(p, &cputime);

	*ut = cputime.utime;
	*st = cputime.stime;
3628 3629
}
#else
3630 3631

#ifndef nsecs_to_cputime
3632
# define nsecs_to_cputime(__nsecs)	nsecs_to_jiffies(__nsecs)
3633 3634
#endif

3635
void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
3636
{
3637
	cputime_t rtime, utime = p->utime, total = cputime_add(utime, p->stime);
3638 3639 3640 3641

	/*
	 * Use CFS's precise accounting:
	 */
3642
	rtime = nsecs_to_cputime(p->se.sum_exec_runtime);
3643 3644

	if (total) {
3645
		u64 temp = rtime;
3646

3647
		temp *= utime;
3648
		do_div(temp, total);
3649 3650 3651
		utime = (cputime_t)temp;
	} else
		utime = rtime;
3652

3653 3654 3655
	/*
	 * Compare with previous values, to keep monotonicity:
	 */
3656
	p->prev_utime = max(p->prev_utime, utime);
3657
	p->prev_stime = max(p->prev_stime, cputime_sub(rtime, p->prev_utime));
3658

3659 3660
	*ut = p->prev_utime;
	*st = p->prev_stime;
3661 3662
}

3663 3664 3665 3666
/*
 * Must be called with siglock held.
 */
void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
3667
{
3668 3669 3670
	struct signal_struct *sig = p->signal;
	struct task_cputime cputime;
	cputime_t rtime, utime, total;
3671

3672
	thread_group_cputime(p, &cputime);
3673

3674 3675
	total = cputime_add(cputime.utime, cputime.stime);
	rtime = nsecs_to_cputime(cputime.sum_exec_runtime);
3676

3677
	if (total) {
3678
		u64 temp = rtime;
3679

3680
		temp *= cputime.utime;
3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691
		do_div(temp, total);
		utime = (cputime_t)temp;
	} else
		utime = rtime;

	sig->prev_utime = max(sig->prev_utime, utime);
	sig->prev_stime = max(sig->prev_stime,
			      cputime_sub(rtime, sig->prev_utime));

	*ut = sig->prev_utime;
	*st = sig->prev_stime;
3692 3693 3694
}
#endif

3695 3696 3697 3698 3699 3700 3701 3702 3703 3704 3705
/*
 * This function gets called by the timer code, with HZ frequency.
 * We call it with interrupts disabled.
 *
 * It also gets called by the fork code, when changing the parent's
 * timeslices.
 */
void scheduler_tick(void)
{
	int cpu = smp_processor_id();
	struct rq *rq = cpu_rq(cpu);
I
Ingo Molnar 已提交
3706
	struct task_struct *curr = rq->curr;
3707 3708

	sched_clock_tick();
I
Ingo Molnar 已提交
3709

3710
	raw_spin_lock(&rq->lock);
3711
	update_rq_clock(rq);
3712
	update_cpu_load_active(rq);
P
Peter Zijlstra 已提交
3713
	curr->sched_class->task_tick(rq, curr, 0);
3714
	raw_spin_unlock(&rq->lock);
3715

3716
	perf_event_task_tick(curr);
3717

3718
#ifdef CONFIG_SMP
I
Ingo Molnar 已提交
3719 3720
	rq->idle_at_tick = idle_cpu(cpu);
	trigger_load_balance(rq, cpu);
3721
#endif
L
Linus Torvalds 已提交
3722 3723
}

3724
notrace unsigned long get_parent_ip(unsigned long addr)
3725 3726 3727 3728 3729 3730 3731 3732
{
	if (in_lock_functions(addr)) {
		addr = CALLER_ADDR2;
		if (in_lock_functions(addr))
			addr = CALLER_ADDR3;
	}
	return addr;
}
L
Linus Torvalds 已提交
3733

3734 3735 3736
#if defined(CONFIG_PREEMPT) && (defined(CONFIG_DEBUG_PREEMPT) || \
				defined(CONFIG_PREEMPT_TRACER))

3737
void __kprobes add_preempt_count(int val)
L
Linus Torvalds 已提交
3738
{
3739
#ifdef CONFIG_DEBUG_PREEMPT
L
Linus Torvalds 已提交
3740 3741 3742
	/*
	 * Underflow?
	 */
3743 3744
	if (DEBUG_LOCKS_WARN_ON((preempt_count() < 0)))
		return;
3745
#endif
L
Linus Torvalds 已提交
3746
	preempt_count() += val;
3747
#ifdef CONFIG_DEBUG_PREEMPT
L
Linus Torvalds 已提交
3748 3749 3750
	/*
	 * Spinlock count overflowing soon?
	 */
3751 3752
	DEBUG_LOCKS_WARN_ON((preempt_count() & PREEMPT_MASK) >=
				PREEMPT_MASK - 10);
3753 3754 3755
#endif
	if (preempt_count() == val)
		trace_preempt_off(CALLER_ADDR0, get_parent_ip(CALLER_ADDR1));
L
Linus Torvalds 已提交
3756 3757 3758
}
EXPORT_SYMBOL(add_preempt_count);

3759
void __kprobes sub_preempt_count(int val)
L
Linus Torvalds 已提交
3760
{
3761
#ifdef CONFIG_DEBUG_PREEMPT
L
Linus Torvalds 已提交
3762 3763 3764
	/*
	 * Underflow?
	 */
3765
	if (DEBUG_LOCKS_WARN_ON(val > preempt_count()))
3766
		return;
L
Linus Torvalds 已提交
3767 3768 3769
	/*
	 * Is the spinlock portion underflowing?
	 */
3770 3771 3772
	if (DEBUG_LOCKS_WARN_ON((val < PREEMPT_MASK) &&
			!(preempt_count() & PREEMPT_MASK)))
		return;
3773
#endif
3774

3775 3776
	if (preempt_count() == val)
		trace_preempt_on(CALLER_ADDR0, get_parent_ip(CALLER_ADDR1));
L
Linus Torvalds 已提交
3777 3778 3779 3780 3781 3782 3783
	preempt_count() -= val;
}
EXPORT_SYMBOL(sub_preempt_count);

#endif

/*
I
Ingo Molnar 已提交
3784
 * Print scheduling while atomic bug:
L
Linus Torvalds 已提交
3785
 */
I
Ingo Molnar 已提交
3786
static noinline void __schedule_bug(struct task_struct *prev)
L
Linus Torvalds 已提交
3787
{
3788 3789
	struct pt_regs *regs = get_irq_regs();

P
Peter Zijlstra 已提交
3790 3791
	printk(KERN_ERR "BUG: scheduling while atomic: %s/%d/0x%08x\n",
		prev->comm, prev->pid, preempt_count());
3792

I
Ingo Molnar 已提交
3793
	debug_show_held_locks(prev);
3794
	print_modules();
I
Ingo Molnar 已提交
3795 3796
	if (irqs_disabled())
		print_irqtrace_events(prev);
3797 3798 3799 3800 3801

	if (regs)
		show_regs(regs);
	else
		dump_stack();
I
Ingo Molnar 已提交
3802
}
L
Linus Torvalds 已提交
3803

I
Ingo Molnar 已提交
3804 3805 3806 3807 3808
/*
 * Various schedule()-time debugging checks and statistics:
 */
static inline void schedule_debug(struct task_struct *prev)
{
L
Linus Torvalds 已提交
3809
	/*
I
Ingo Molnar 已提交
3810
	 * Test if we are atomic. Since do_exit() needs to call into
L
Linus Torvalds 已提交
3811 3812 3813
	 * schedule() atomically, we ignore that path for now.
	 * Otherwise, whine if we are scheduling when we should not be.
	 */
3814
	if (unlikely(in_atomic_preempt_off() && !prev->exit_state))
I
Ingo Molnar 已提交
3815 3816
		__schedule_bug(prev);

L
Linus Torvalds 已提交
3817 3818
	profile_hit(SCHED_PROFILING, __builtin_return_address(0));

3819
	schedstat_inc(this_rq(), sched_count);
I
Ingo Molnar 已提交
3820 3821
#ifdef CONFIG_SCHEDSTATS
	if (unlikely(prev->lock_depth >= 0)) {
3822 3823
		schedstat_inc(this_rq(), bkl_count);
		schedstat_inc(prev, sched_info.bkl_count);
I
Ingo Molnar 已提交
3824 3825
	}
#endif
I
Ingo Molnar 已提交
3826 3827
}

P
Peter Zijlstra 已提交
3828
static void put_prev_task(struct rq *rq, struct task_struct *prev)
M
Mike Galbraith 已提交
3829
{
3830 3831 3832
	if (prev->se.on_rq)
		update_rq_clock(rq);
	rq->skip_clock_update = 0;
P
Peter Zijlstra 已提交
3833
	prev->sched_class->put_prev_task(rq, prev);
M
Mike Galbraith 已提交
3834 3835
}

I
Ingo Molnar 已提交
3836 3837 3838 3839
/*
 * Pick up the highest-prio task:
 */
static inline struct task_struct *
3840
pick_next_task(struct rq *rq)
I
Ingo Molnar 已提交
3841
{
3842
	const struct sched_class *class;
I
Ingo Molnar 已提交
3843
	struct task_struct *p;
L
Linus Torvalds 已提交
3844 3845

	/*
I
Ingo Molnar 已提交
3846 3847
	 * Optimization: we know that if all tasks are in
	 * the fair class we can call that function directly:
L
Linus Torvalds 已提交
3848
	 */
I
Ingo Molnar 已提交
3849
	if (likely(rq->nr_running == rq->cfs.nr_running)) {
3850
		p = fair_sched_class.pick_next_task(rq);
I
Ingo Molnar 已提交
3851 3852
		if (likely(p))
			return p;
L
Linus Torvalds 已提交
3853 3854
	}

3855
	for_each_class(class) {
3856
		p = class->pick_next_task(rq);
I
Ingo Molnar 已提交
3857 3858 3859
		if (p)
			return p;
	}
3860 3861

	BUG(); /* the idle class will always have a runnable task */
I
Ingo Molnar 已提交
3862
}
L
Linus Torvalds 已提交
3863

I
Ingo Molnar 已提交
3864 3865 3866
/*
 * schedule() is the main scheduler function.
 */
3867
asmlinkage void __sched schedule(void)
I
Ingo Molnar 已提交
3868 3869
{
	struct task_struct *prev, *next;
3870
	unsigned long *switch_count;
I
Ingo Molnar 已提交
3871
	struct rq *rq;
3872
	int cpu;
I
Ingo Molnar 已提交
3873

3874 3875
need_resched:
	preempt_disable();
I
Ingo Molnar 已提交
3876 3877
	cpu = smp_processor_id();
	rq = cpu_rq(cpu);
3878
	rcu_note_context_switch(cpu);
I
Ingo Molnar 已提交
3879 3880 3881 3882 3883 3884
	prev = rq->curr;

	release_kernel_lock(prev);
need_resched_nonpreemptible:

	schedule_debug(prev);
L
Linus Torvalds 已提交
3885

3886
	if (sched_feat(HRTICK))
M
Mike Galbraith 已提交
3887
		hrtick_clear(rq);
P
Peter Zijlstra 已提交
3888

3889
	raw_spin_lock_irq(&rq->lock);
3890
	clear_tsk_need_resched(prev);
L
Linus Torvalds 已提交
3891

3892
	switch_count = &prev->nivcsw;
L
Linus Torvalds 已提交
3893
	if (prev->state && !(preempt_count() & PREEMPT_ACTIVE)) {
T
Tejun Heo 已提交
3894
		if (unlikely(signal_pending_state(prev->state, prev))) {
L
Linus Torvalds 已提交
3895
			prev->state = TASK_RUNNING;
T
Tejun Heo 已提交
3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909
		} else {
			/*
			 * If a worker is going to sleep, notify and
			 * ask workqueue whether it wants to wake up a
			 * task to maintain concurrency.  If so, wake
			 * up the task.
			 */
			if (prev->flags & PF_WQ_WORKER) {
				struct task_struct *to_wakeup;

				to_wakeup = wq_worker_sleeping(prev, cpu);
				if (to_wakeup)
					try_to_wake_up_local(to_wakeup);
			}
3910
			deactivate_task(rq, prev, DEQUEUE_SLEEP);
T
Tejun Heo 已提交
3911
		}
I
Ingo Molnar 已提交
3912
		switch_count = &prev->nvcsw;
L
Linus Torvalds 已提交
3913 3914
	}

3915
	pre_schedule(rq, prev);
3916

I
Ingo Molnar 已提交
3917
	if (unlikely(!rq->nr_running))
L
Linus Torvalds 已提交
3918 3919
		idle_balance(cpu, rq);

M
Mike Galbraith 已提交
3920
	put_prev_task(rq, prev);
3921
	next = pick_next_task(rq);
L
Linus Torvalds 已提交
3922 3923

	if (likely(prev != next)) {
3924
		sched_info_switch(prev, next);
3925
		perf_event_task_sched_out(prev, next);
3926

L
Linus Torvalds 已提交
3927 3928 3929 3930
		rq->nr_switches++;
		rq->curr = next;
		++*switch_count;

I
Ingo Molnar 已提交
3931
		context_switch(rq, prev, next); /* unlocks the rq */
P
Peter Zijlstra 已提交
3932
		/*
3933 3934 3935 3936
		 * The context switch have flipped the stack from under us
		 * and restored the local variables which were saved when
		 * this task called schedule() in the past. prev == current
		 * is still correct, but it can be moved to another cpu/rq.
P
Peter Zijlstra 已提交
3937 3938 3939
		 */
		cpu = smp_processor_id();
		rq = cpu_rq(cpu);
L
Linus Torvalds 已提交
3940
	} else
3941
		raw_spin_unlock_irq(&rq->lock);
L
Linus Torvalds 已提交
3942

3943
	post_schedule(rq);
L
Linus Torvalds 已提交
3944

3945
	if (unlikely(reacquire_kernel_lock(prev)))
L
Linus Torvalds 已提交
3946
		goto need_resched_nonpreemptible;
P
Peter Zijlstra 已提交
3947

L
Linus Torvalds 已提交
3948
	preempt_enable_no_resched();
3949
	if (need_resched())
L
Linus Torvalds 已提交
3950 3951 3952 3953
		goto need_resched;
}
EXPORT_SYMBOL(schedule);

3954
#ifdef CONFIG_MUTEX_SPIN_ON_OWNER
3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973
/*
 * Look out! "owner" is an entirely speculative pointer
 * access and not reliable.
 */
int mutex_spin_on_owner(struct mutex *lock, struct thread_info *owner)
{
	unsigned int cpu;
	struct rq *rq;

	if (!sched_feat(OWNER_SPIN))
		return 0;

#ifdef CONFIG_DEBUG_PAGEALLOC
	/*
	 * Need to access the cpu field knowing that
	 * DEBUG_PAGEALLOC could have unmapped it if
	 * the mutex owner just released it and exited.
	 */
	if (probe_kernel_address(&owner->cpu, cpu))
3974
		return 0;
3975 3976 3977 3978 3979 3980 3981 3982 3983
#else
	cpu = owner->cpu;
#endif

	/*
	 * Even if the access succeeded (likely case),
	 * the cpu field may no longer be valid.
	 */
	if (cpu >= nr_cpumask_bits)
3984
		return 0;
3985 3986 3987 3988 3989 3990

	/*
	 * We need to validate that we can do a
	 * get_cpu() and that we have the percpu area.
	 */
	if (!cpu_online(cpu))
3991
		return 0;
3992 3993 3994 3995 3996 3997 3998

	rq = cpu_rq(cpu);

	for (;;) {
		/*
		 * Owner changed, break to re-assess state.
		 */
3999 4000 4001 4002 4003 4004 4005 4006
		if (lock->owner != owner) {
			/*
			 * If the lock has switched to a different owner,
			 * we likely have heavy contention. Return 0 to quit
			 * optimistic spinning and not contend further:
			 */
			if (lock->owner)
				return 0;
4007
			break;
4008
		}
4009 4010 4011 4012 4013 4014 4015 4016 4017

		/*
		 * Is that owner really running on that cpu?
		 */
		if (task_thread_info(rq->curr) != owner || need_resched())
			return 0;

		cpu_relax();
	}
4018

4019 4020 4021 4022
	return 1;
}
#endif

L
Linus Torvalds 已提交
4023 4024
#ifdef CONFIG_PREEMPT
/*
4025
 * this is the entry point to schedule() from in-kernel preemption
I
Ingo Molnar 已提交
4026
 * off of preempt_enable. Kernel preemptions off return from interrupt
L
Linus Torvalds 已提交
4027 4028
 * occur there and call schedule directly.
 */
4029
asmlinkage void __sched notrace preempt_schedule(void)
L
Linus Torvalds 已提交
4030 4031
{
	struct thread_info *ti = current_thread_info();
4032

L
Linus Torvalds 已提交
4033 4034
	/*
	 * If there is a non-zero preempt_count or interrupts are disabled,
I
Ingo Molnar 已提交
4035
	 * we do not want to preempt the current task. Just return..
L
Linus Torvalds 已提交
4036
	 */
N
Nick Piggin 已提交
4037
	if (likely(ti->preempt_count || irqs_disabled()))
L
Linus Torvalds 已提交
4038 4039
		return;

4040
	do {
4041
		add_preempt_count_notrace(PREEMPT_ACTIVE);
4042
		schedule();
4043
		sub_preempt_count_notrace(PREEMPT_ACTIVE);
L
Linus Torvalds 已提交
4044

4045 4046 4047 4048 4049
		/*
		 * Check again in case we missed a preemption opportunity
		 * between schedule and now.
		 */
		barrier();
4050
	} while (need_resched());
L
Linus Torvalds 已提交
4051 4052 4053 4054
}
EXPORT_SYMBOL(preempt_schedule);

/*
4055
 * this is the entry point to schedule() from kernel preemption
L
Linus Torvalds 已提交
4056 4057 4058 4059 4060 4061 4062
 * off of irq context.
 * Note, that this is called and return with irqs disabled. This will
 * protect us against recursive calling from irq.
 */
asmlinkage void __sched preempt_schedule_irq(void)
{
	struct thread_info *ti = current_thread_info();
4063

4064
	/* Catch callers which need to be fixed */
L
Linus Torvalds 已提交
4065 4066
	BUG_ON(ti->preempt_count || !irqs_disabled());

4067 4068 4069 4070 4071 4072
	do {
		add_preempt_count(PREEMPT_ACTIVE);
		local_irq_enable();
		schedule();
		local_irq_disable();
		sub_preempt_count(PREEMPT_ACTIVE);
L
Linus Torvalds 已提交
4073

4074 4075 4076 4077 4078
		/*
		 * Check again in case we missed a preemption opportunity
		 * between schedule and now.
		 */
		barrier();
4079
	} while (need_resched());
L
Linus Torvalds 已提交
4080 4081 4082 4083
}

#endif /* CONFIG_PREEMPT */

P
Peter Zijlstra 已提交
4084
int default_wake_function(wait_queue_t *curr, unsigned mode, int wake_flags,
I
Ingo Molnar 已提交
4085
			  void *key)
L
Linus Torvalds 已提交
4086
{
P
Peter Zijlstra 已提交
4087
	return try_to_wake_up(curr->private, mode, wake_flags);
L
Linus Torvalds 已提交
4088 4089 4090 4091
}
EXPORT_SYMBOL(default_wake_function);

/*
I
Ingo Molnar 已提交
4092 4093
 * The core wakeup function. Non-exclusive wakeups (nr_exclusive == 0) just
 * wake everything up. If it's an exclusive wakeup (nr_exclusive == small +ve
L
Linus Torvalds 已提交
4094 4095 4096
 * number) then we wake all the non-exclusive tasks and one exclusive task.
 *
 * There are circumstances in which we can try to wake a task which has already
I
Ingo Molnar 已提交
4097
 * started to run but is not in state TASK_RUNNING. try_to_wake_up() returns
L
Linus Torvalds 已提交
4098 4099
 * zero in this (rare) case, and we handle it by continuing to scan the queue.
 */
4100
static void __wake_up_common(wait_queue_head_t *q, unsigned int mode,
P
Peter Zijlstra 已提交
4101
			int nr_exclusive, int wake_flags, void *key)
L
Linus Torvalds 已提交
4102
{
4103
	wait_queue_t *curr, *next;
L
Linus Torvalds 已提交
4104

4105
	list_for_each_entry_safe(curr, next, &q->task_list, task_list) {
4106 4107
		unsigned flags = curr->flags;

P
Peter Zijlstra 已提交
4108
		if (curr->func(curr, mode, wake_flags, key) &&
4109
				(flags & WQ_FLAG_EXCLUSIVE) && !--nr_exclusive)
L
Linus Torvalds 已提交
4110 4111 4112 4113 4114 4115 4116 4117 4118
			break;
	}
}

/**
 * __wake_up - wake up threads blocked on a waitqueue.
 * @q: the waitqueue
 * @mode: which threads
 * @nr_exclusive: how many wake-one or wake-many threads to wake up
4119
 * @key: is directly passed to the wakeup function
4120 4121 4122
 *
 * It may be assumed that this function implies a write memory barrier before
 * changing the task state if and only if any tasks are woken up.
L
Linus Torvalds 已提交
4123
 */
4124
void __wake_up(wait_queue_head_t *q, unsigned int mode,
I
Ingo Molnar 已提交
4125
			int nr_exclusive, void *key)
L
Linus Torvalds 已提交
4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137
{
	unsigned long flags;

	spin_lock_irqsave(&q->lock, flags);
	__wake_up_common(q, mode, nr_exclusive, 0, key);
	spin_unlock_irqrestore(&q->lock, flags);
}
EXPORT_SYMBOL(__wake_up);

/*
 * Same as __wake_up but called with the spinlock in wait_queue_head_t held.
 */
4138
void __wake_up_locked(wait_queue_head_t *q, unsigned int mode)
L
Linus Torvalds 已提交
4139 4140 4141
{
	__wake_up_common(q, mode, 1, 0, NULL);
}
4142
EXPORT_SYMBOL_GPL(__wake_up_locked);
L
Linus Torvalds 已提交
4143

4144 4145 4146 4147 4148
void __wake_up_locked_key(wait_queue_head_t *q, unsigned int mode, void *key)
{
	__wake_up_common(q, mode, 1, 0, key);
}

L
Linus Torvalds 已提交
4149
/**
4150
 * __wake_up_sync_key - wake up threads blocked on a waitqueue.
L
Linus Torvalds 已提交
4151 4152 4153
 * @q: the waitqueue
 * @mode: which threads
 * @nr_exclusive: how many wake-one or wake-many threads to wake up
4154
 * @key: opaque value to be passed to wakeup targets
L
Linus Torvalds 已提交
4155 4156 4157 4158 4159 4160 4161
 *
 * The sync wakeup differs that the waker knows that it will schedule
 * away soon, so while the target thread will be woken up, it will not
 * be migrated to another CPU - ie. the two threads are 'synchronized'
 * with each other. This can prevent needless bouncing between CPUs.
 *
 * On UP it can prevent extra preemption.
4162 4163 4164
 *
 * It may be assumed that this function implies a write memory barrier before
 * changing the task state if and only if any tasks are woken up.
L
Linus Torvalds 已提交
4165
 */
4166 4167
void __wake_up_sync_key(wait_queue_head_t *q, unsigned int mode,
			int nr_exclusive, void *key)
L
Linus Torvalds 已提交
4168 4169
{
	unsigned long flags;
P
Peter Zijlstra 已提交
4170
	int wake_flags = WF_SYNC;
L
Linus Torvalds 已提交
4171 4172 4173 4174 4175

	if (unlikely(!q))
		return;

	if (unlikely(!nr_exclusive))
P
Peter Zijlstra 已提交
4176
		wake_flags = 0;
L
Linus Torvalds 已提交
4177 4178

	spin_lock_irqsave(&q->lock, flags);
P
Peter Zijlstra 已提交
4179
	__wake_up_common(q, mode, nr_exclusive, wake_flags, key);
L
Linus Torvalds 已提交
4180 4181
	spin_unlock_irqrestore(&q->lock, flags);
}
4182 4183 4184 4185 4186 4187 4188 4189 4190
EXPORT_SYMBOL_GPL(__wake_up_sync_key);

/*
 * __wake_up_sync - see __wake_up_sync_key()
 */
void __wake_up_sync(wait_queue_head_t *q, unsigned int mode, int nr_exclusive)
{
	__wake_up_sync_key(q, mode, nr_exclusive, NULL);
}
L
Linus Torvalds 已提交
4191 4192
EXPORT_SYMBOL_GPL(__wake_up_sync);	/* For internal use only */

4193 4194 4195 4196 4197 4198 4199 4200
/**
 * complete: - signals a single thread waiting on this completion
 * @x:  holds the state of this particular completion
 *
 * This will wake up a single thread waiting on this completion. Threads will be
 * awakened in the same order in which they were queued.
 *
 * See also complete_all(), wait_for_completion() and related routines.
4201 4202 4203
 *
 * It may be assumed that this function implies a write memory barrier before
 * changing the task state if and only if any tasks are woken up.
4204
 */
4205
void complete(struct completion *x)
L
Linus Torvalds 已提交
4206 4207 4208 4209 4210
{
	unsigned long flags;

	spin_lock_irqsave(&x->wait.lock, flags);
	x->done++;
4211
	__wake_up_common(&x->wait, TASK_NORMAL, 1, 0, NULL);
L
Linus Torvalds 已提交
4212 4213 4214 4215
	spin_unlock_irqrestore(&x->wait.lock, flags);
}
EXPORT_SYMBOL(complete);

4216 4217 4218 4219 4220
/**
 * complete_all: - signals all threads waiting on this completion
 * @x:  holds the state of this particular completion
 *
 * This will wake up all threads waiting on this particular completion event.
4221 4222 4223
 *
 * It may be assumed that this function implies a write memory barrier before
 * changing the task state if and only if any tasks are woken up.
4224
 */
4225
void complete_all(struct completion *x)
L
Linus Torvalds 已提交
4226 4227 4228 4229 4230
{
	unsigned long flags;

	spin_lock_irqsave(&x->wait.lock, flags);
	x->done += UINT_MAX/2;
4231
	__wake_up_common(&x->wait, TASK_NORMAL, 0, 0, NULL);
L
Linus Torvalds 已提交
4232 4233 4234 4235
	spin_unlock_irqrestore(&x->wait.lock, flags);
}
EXPORT_SYMBOL(complete_all);

4236 4237
static inline long __sched
do_wait_for_common(struct completion *x, long timeout, int state)
L
Linus Torvalds 已提交
4238 4239 4240 4241
{
	if (!x->done) {
		DECLARE_WAITQUEUE(wait, current);

C
Changli Gao 已提交
4242
		__add_wait_queue_tail_exclusive(&x->wait, &wait);
L
Linus Torvalds 已提交
4243
		do {
4244
			if (signal_pending_state(state, current)) {
4245 4246
				timeout = -ERESTARTSYS;
				break;
4247 4248
			}
			__set_current_state(state);
L
Linus Torvalds 已提交
4249 4250 4251
			spin_unlock_irq(&x->wait.lock);
			timeout = schedule_timeout(timeout);
			spin_lock_irq(&x->wait.lock);
4252
		} while (!x->done && timeout);
L
Linus Torvalds 已提交
4253
		__remove_wait_queue(&x->wait, &wait);
4254 4255
		if (!x->done)
			return timeout;
L
Linus Torvalds 已提交
4256 4257
	}
	x->done--;
4258
	return timeout ?: 1;
L
Linus Torvalds 已提交
4259 4260
}

4261 4262
static long __sched
wait_for_common(struct completion *x, long timeout, int state)
L
Linus Torvalds 已提交
4263 4264 4265 4266
{
	might_sleep();

	spin_lock_irq(&x->wait.lock);
4267
	timeout = do_wait_for_common(x, timeout, state);
L
Linus Torvalds 已提交
4268
	spin_unlock_irq(&x->wait.lock);
4269 4270
	return timeout;
}
L
Linus Torvalds 已提交
4271

4272 4273 4274 4275 4276 4277 4278 4279 4280 4281
/**
 * wait_for_completion: - waits for completion of a task
 * @x:  holds the state of this particular completion
 *
 * This waits to be signaled for completion of a specific task. It is NOT
 * interruptible and there is no timeout.
 *
 * See also similar routines (i.e. wait_for_completion_timeout()) with timeout
 * and interrupt capability. Also see complete().
 */
4282
void __sched wait_for_completion(struct completion *x)
4283 4284
{
	wait_for_common(x, MAX_SCHEDULE_TIMEOUT, TASK_UNINTERRUPTIBLE);
L
Linus Torvalds 已提交
4285
}
4286
EXPORT_SYMBOL(wait_for_completion);
L
Linus Torvalds 已提交
4287

4288 4289 4290 4291 4292 4293 4294 4295 4296
/**
 * wait_for_completion_timeout: - waits for completion of a task (w/timeout)
 * @x:  holds the state of this particular completion
 * @timeout:  timeout value in jiffies
 *
 * This waits for either a completion of a specific task to be signaled or for a
 * specified timeout to expire. The timeout is in jiffies. It is not
 * interruptible.
 */
4297
unsigned long __sched
4298
wait_for_completion_timeout(struct completion *x, unsigned long timeout)
L
Linus Torvalds 已提交
4299
{
4300
	return wait_for_common(x, timeout, TASK_UNINTERRUPTIBLE);
L
Linus Torvalds 已提交
4301
}
4302
EXPORT_SYMBOL(wait_for_completion_timeout);
L
Linus Torvalds 已提交
4303

4304 4305 4306 4307 4308 4309 4310
/**
 * wait_for_completion_interruptible: - waits for completion of a task (w/intr)
 * @x:  holds the state of this particular completion
 *
 * This waits for completion of a specific task to be signaled. It is
 * interruptible.
 */
4311
int __sched wait_for_completion_interruptible(struct completion *x)
I
Ingo Molnar 已提交
4312
{
4313 4314 4315 4316
	long t = wait_for_common(x, MAX_SCHEDULE_TIMEOUT, TASK_INTERRUPTIBLE);
	if (t == -ERESTARTSYS)
		return t;
	return 0;
I
Ingo Molnar 已提交
4317
}
4318
EXPORT_SYMBOL(wait_for_completion_interruptible);
L
Linus Torvalds 已提交
4319

4320 4321 4322 4323 4324 4325 4326 4327
/**
 * wait_for_completion_interruptible_timeout: - waits for completion (w/(to,intr))
 * @x:  holds the state of this particular completion
 * @timeout:  timeout value in jiffies
 *
 * This waits for either a completion of a specific task to be signaled or for a
 * specified timeout to expire. It is interruptible. The timeout is in jiffies.
 */
4328
unsigned long __sched
4329 4330
wait_for_completion_interruptible_timeout(struct completion *x,
					  unsigned long timeout)
I
Ingo Molnar 已提交
4331
{
4332
	return wait_for_common(x, timeout, TASK_INTERRUPTIBLE);
I
Ingo Molnar 已提交
4333
}
4334
EXPORT_SYMBOL(wait_for_completion_interruptible_timeout);
L
Linus Torvalds 已提交
4335

4336 4337 4338 4339 4340 4341 4342
/**
 * wait_for_completion_killable: - waits for completion of a task (killable)
 * @x:  holds the state of this particular completion
 *
 * This waits to be signaled for completion of a specific task. It can be
 * interrupted by a kill signal.
 */
M
Matthew Wilcox 已提交
4343 4344 4345 4346 4347 4348 4349 4350 4351
int __sched wait_for_completion_killable(struct completion *x)
{
	long t = wait_for_common(x, MAX_SCHEDULE_TIMEOUT, TASK_KILLABLE);
	if (t == -ERESTARTSYS)
		return t;
	return 0;
}
EXPORT_SYMBOL(wait_for_completion_killable);

4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368
/**
 * wait_for_completion_killable_timeout: - waits for completion of a task (w/(to,killable))
 * @x:  holds the state of this particular completion
 * @timeout:  timeout value in jiffies
 *
 * This waits for either a completion of a specific task to be
 * signaled or for a specified timeout to expire. It can be
 * interrupted by a kill signal. The timeout is in jiffies.
 */
unsigned long __sched
wait_for_completion_killable_timeout(struct completion *x,
				     unsigned long timeout)
{
	return wait_for_common(x, timeout, TASK_KILLABLE);
}
EXPORT_SYMBOL(wait_for_completion_killable_timeout);

4369 4370 4371 4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382
/**
 *	try_wait_for_completion - try to decrement a completion without blocking
 *	@x:	completion structure
 *
 *	Returns: 0 if a decrement cannot be done without blocking
 *		 1 if a decrement succeeded.
 *
 *	If a completion is being used as a counting completion,
 *	attempt to decrement the counter without blocking. This
 *	enables us to avoid waiting if the resource the completion
 *	is protecting is not available.
 */
bool try_wait_for_completion(struct completion *x)
{
4383
	unsigned long flags;
4384 4385
	int ret = 1;

4386
	spin_lock_irqsave(&x->wait.lock, flags);
4387 4388 4389 4390
	if (!x->done)
		ret = 0;
	else
		x->done--;
4391
	spin_unlock_irqrestore(&x->wait.lock, flags);
4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405
	return ret;
}
EXPORT_SYMBOL(try_wait_for_completion);

/**
 *	completion_done - Test to see if a completion has any waiters
 *	@x:	completion structure
 *
 *	Returns: 0 if there are waiters (wait_for_completion() in progress)
 *		 1 if there are no waiters.
 *
 */
bool completion_done(struct completion *x)
{
4406
	unsigned long flags;
4407 4408
	int ret = 1;

4409
	spin_lock_irqsave(&x->wait.lock, flags);
4410 4411
	if (!x->done)
		ret = 0;
4412
	spin_unlock_irqrestore(&x->wait.lock, flags);
4413 4414 4415 4416
	return ret;
}
EXPORT_SYMBOL(completion_done);

4417 4418
static long __sched
sleep_on_common(wait_queue_head_t *q, int state, long timeout)
L
Linus Torvalds 已提交
4419
{
I
Ingo Molnar 已提交
4420 4421 4422 4423
	unsigned long flags;
	wait_queue_t wait;

	init_waitqueue_entry(&wait, current);
L
Linus Torvalds 已提交
4424

4425
	__set_current_state(state);
L
Linus Torvalds 已提交
4426

4427 4428 4429 4430 4431 4432 4433 4434 4435 4436 4437 4438 4439 4440
	spin_lock_irqsave(&q->lock, flags);
	__add_wait_queue(q, &wait);
	spin_unlock(&q->lock);
	timeout = schedule_timeout(timeout);
	spin_lock_irq(&q->lock);
	__remove_wait_queue(q, &wait);
	spin_unlock_irqrestore(&q->lock, flags);

	return timeout;
}

void __sched interruptible_sleep_on(wait_queue_head_t *q)
{
	sleep_on_common(q, TASK_INTERRUPTIBLE, MAX_SCHEDULE_TIMEOUT);
L
Linus Torvalds 已提交
4441 4442 4443
}
EXPORT_SYMBOL(interruptible_sleep_on);

I
Ingo Molnar 已提交
4444
long __sched
I
Ingo Molnar 已提交
4445
interruptible_sleep_on_timeout(wait_queue_head_t *q, long timeout)
L
Linus Torvalds 已提交
4446
{
4447
	return sleep_on_common(q, TASK_INTERRUPTIBLE, timeout);
L
Linus Torvalds 已提交
4448 4449 4450
}
EXPORT_SYMBOL(interruptible_sleep_on_timeout);

I
Ingo Molnar 已提交
4451
void __sched sleep_on(wait_queue_head_t *q)
L
Linus Torvalds 已提交
4452
{
4453
	sleep_on_common(q, TASK_UNINTERRUPTIBLE, MAX_SCHEDULE_TIMEOUT);
L
Linus Torvalds 已提交
4454 4455 4456
}
EXPORT_SYMBOL(sleep_on);

I
Ingo Molnar 已提交
4457
long __sched sleep_on_timeout(wait_queue_head_t *q, long timeout)
L
Linus Torvalds 已提交
4458
{
4459
	return sleep_on_common(q, TASK_UNINTERRUPTIBLE, timeout);
L
Linus Torvalds 已提交
4460 4461 4462
}
EXPORT_SYMBOL(sleep_on_timeout);

4463 4464 4465 4466 4467 4468 4469 4470 4471 4472 4473 4474
#ifdef CONFIG_RT_MUTEXES

/*
 * rt_mutex_setprio - set the current priority of a task
 * @p: task
 * @prio: prio value (kernel-internal form)
 *
 * This function changes the 'effective' priority of a task. It does
 * not touch ->normal_prio like __setscheduler().
 *
 * Used by the rt_mutex code to implement priority inheritance logic.
 */
4475
void rt_mutex_setprio(struct task_struct *p, int prio)
4476 4477
{
	unsigned long flags;
4478
	int oldprio, on_rq, running;
4479
	struct rq *rq;
4480
	const struct sched_class *prev_class;
4481 4482 4483 4484 4485

	BUG_ON(prio < 0 || prio > MAX_PRIO);

	rq = task_rq_lock(p, &flags);

4486
	trace_sched_pi_setprio(p, prio);
4487
	oldprio = p->prio;
4488
	prev_class = p->sched_class;
I
Ingo Molnar 已提交
4489
	on_rq = p->se.on_rq;
4490
	running = task_current(rq, p);
4491
	if (on_rq)
4492
		dequeue_task(rq, p, 0);
4493 4494
	if (running)
		p->sched_class->put_prev_task(rq, p);
I
Ingo Molnar 已提交
4495 4496 4497 4498 4499 4500

	if (rt_prio(prio))
		p->sched_class = &rt_sched_class;
	else
		p->sched_class = &fair_sched_class;

4501 4502
	p->prio = prio;

4503 4504
	if (running)
		p->sched_class->set_curr_task(rq);
I
Ingo Molnar 已提交
4505
	if (on_rq) {
4506
		enqueue_task(rq, p, oldprio < prio ? ENQUEUE_HEAD : 0);
4507 4508

		check_class_changed(rq, p, prev_class, oldprio, running);
4509 4510 4511 4512 4513 4514
	}
	task_rq_unlock(rq, &flags);
}

#endif

4515
void set_user_nice(struct task_struct *p, long nice)
L
Linus Torvalds 已提交
4516
{
I
Ingo Molnar 已提交
4517
	int old_prio, delta, on_rq;
L
Linus Torvalds 已提交
4518
	unsigned long flags;
4519
	struct rq *rq;
L
Linus Torvalds 已提交
4520 4521 4522 4523 4524 4525 4526 4527 4528 4529 4530 4531

	if (TASK_NICE(p) == nice || nice < -20 || nice > 19)
		return;
	/*
	 * We have to be careful, if called from sys_setpriority(),
	 * the task might be in the middle of scheduling on another CPU.
	 */
	rq = task_rq_lock(p, &flags);
	/*
	 * The RT priorities are set via sched_setscheduler(), but we still
	 * allow the 'normal' nice value to be set - but as expected
	 * it wont have any effect on scheduling until the task is
I
Ingo Molnar 已提交
4532
	 * SCHED_FIFO/SCHED_RR:
L
Linus Torvalds 已提交
4533
	 */
4534
	if (task_has_rt_policy(p)) {
L
Linus Torvalds 已提交
4535 4536 4537
		p->static_prio = NICE_TO_PRIO(nice);
		goto out_unlock;
	}
I
Ingo Molnar 已提交
4538
	on_rq = p->se.on_rq;
4539
	if (on_rq)
4540
		dequeue_task(rq, p, 0);
L
Linus Torvalds 已提交
4541 4542

	p->static_prio = NICE_TO_PRIO(nice);
4543
	set_load_weight(p);
4544 4545 4546
	old_prio = p->prio;
	p->prio = effective_prio(p);
	delta = p->prio - old_prio;
L
Linus Torvalds 已提交
4547

I
Ingo Molnar 已提交
4548
	if (on_rq) {
4549
		enqueue_task(rq, p, 0);
L
Linus Torvalds 已提交
4550
		/*
4551 4552
		 * If the task increased its priority or is running and
		 * lowered its priority, then reschedule its CPU:
L
Linus Torvalds 已提交
4553
		 */
4554
		if (delta < 0 || (delta > 0 && task_running(rq, p)))
L
Linus Torvalds 已提交
4555 4556 4557 4558 4559 4560 4561
			resched_task(rq->curr);
	}
out_unlock:
	task_rq_unlock(rq, &flags);
}
EXPORT_SYMBOL(set_user_nice);

M
Matt Mackall 已提交
4562 4563 4564 4565 4566
/*
 * can_nice - check if a task can reduce its nice value
 * @p: task
 * @nice: nice value
 */
4567
int can_nice(const struct task_struct *p, const int nice)
M
Matt Mackall 已提交
4568
{
4569 4570
	/* convert nice value [19,-20] to rlimit style value [1,40] */
	int nice_rlim = 20 - nice;
4571

4572
	return (nice_rlim <= task_rlimit(p, RLIMIT_NICE) ||
M
Matt Mackall 已提交
4573 4574 4575
		capable(CAP_SYS_NICE));
}

L
Linus Torvalds 已提交
4576 4577 4578 4579 4580 4581 4582 4583 4584
#ifdef __ARCH_WANT_SYS_NICE

/*
 * sys_nice - change the priority of the current process.
 * @increment: priority increment
 *
 * sys_setpriority is a more generic, but much slower function that
 * does similar things.
 */
4585
SYSCALL_DEFINE1(nice, int, increment)
L
Linus Torvalds 已提交
4586
{
4587
	long nice, retval;
L
Linus Torvalds 已提交
4588 4589 4590 4591 4592 4593

	/*
	 * Setpriority might change our priority at the same moment.
	 * We don't have to worry. Conceptually one call occurs first
	 * and we have a single winner.
	 */
M
Matt Mackall 已提交
4594 4595
	if (increment < -40)
		increment = -40;
L
Linus Torvalds 已提交
4596 4597 4598
	if (increment > 40)
		increment = 40;

4599
	nice = TASK_NICE(current) + increment;
L
Linus Torvalds 已提交
4600 4601 4602 4603 4604
	if (nice < -20)
		nice = -20;
	if (nice > 19)
		nice = 19;

M
Matt Mackall 已提交
4605 4606 4607
	if (increment < 0 && !can_nice(current, nice))
		return -EPERM;

L
Linus Torvalds 已提交
4608 4609 4610 4611 4612 4613 4614 4615 4616 4617 4618 4619 4620 4621 4622 4623 4624 4625
	retval = security_task_setnice(current, nice);
	if (retval)
		return retval;

	set_user_nice(current, nice);
	return 0;
}

#endif

/**
 * task_prio - return the priority value of a given task.
 * @p: the task in question.
 *
 * This is the priority value as seen by users in /proc.
 * RT tasks are offset by -200. Normal tasks are centered
 * around 0, value goes from -16 to +15.
 */
4626
int task_prio(const struct task_struct *p)
L
Linus Torvalds 已提交
4627 4628 4629 4630 4631 4632 4633 4634
{
	return p->prio - MAX_RT_PRIO;
}

/**
 * task_nice - return the nice value of a given task.
 * @p: the task in question.
 */
4635
int task_nice(const struct task_struct *p)
L
Linus Torvalds 已提交
4636 4637 4638
{
	return TASK_NICE(p);
}
P
Pavel Roskin 已提交
4639
EXPORT_SYMBOL(task_nice);
L
Linus Torvalds 已提交
4640 4641 4642 4643 4644 4645 4646 4647 4648 4649 4650 4651 4652 4653

/**
 * idle_cpu - is a given cpu idle currently?
 * @cpu: the processor in question.
 */
int idle_cpu(int cpu)
{
	return cpu_curr(cpu) == cpu_rq(cpu)->idle;
}

/**
 * idle_task - return the idle task for a given cpu.
 * @cpu: the processor in question.
 */
4654
struct task_struct *idle_task(int cpu)
L
Linus Torvalds 已提交
4655 4656 4657 4658 4659 4660 4661 4662
{
	return cpu_rq(cpu)->idle;
}

/**
 * find_process_by_pid - find a process with a matching PID value.
 * @pid: the pid in question.
 */
A
Alexey Dobriyan 已提交
4663
static struct task_struct *find_process_by_pid(pid_t pid)
L
Linus Torvalds 已提交
4664
{
4665
	return pid ? find_task_by_vpid(pid) : current;
L
Linus Torvalds 已提交
4666 4667 4668
}

/* Actually do priority change: must hold rq lock. */
I
Ingo Molnar 已提交
4669 4670
static void
__setscheduler(struct rq *rq, struct task_struct *p, int policy, int prio)
L
Linus Torvalds 已提交
4671
{
I
Ingo Molnar 已提交
4672
	BUG_ON(p->se.on_rq);
4673

L
Linus Torvalds 已提交
4674 4675
	p->policy = policy;
	p->rt_priority = prio;
4676 4677 4678
	p->normal_prio = normal_prio(p);
	/* we are holding p->pi_lock already */
	p->prio = rt_mutex_getprio(p);
4679 4680 4681 4682
	if (rt_prio(p->prio))
		p->sched_class = &rt_sched_class;
	else
		p->sched_class = &fair_sched_class;
4683
	set_load_weight(p);
L
Linus Torvalds 已提交
4684 4685
}

4686 4687 4688 4689 4690 4691 4692 4693 4694 4695 4696 4697 4698 4699 4700 4701
/*
 * check the target process has a UID that matches the current process's
 */
static bool check_same_owner(struct task_struct *p)
{
	const struct cred *cred = current_cred(), *pcred;
	bool match;

	rcu_read_lock();
	pcred = __task_cred(p);
	match = (cred->euid == pcred->euid ||
		 cred->euid == pcred->uid);
	rcu_read_unlock();
	return match;
}

4702 4703
static int __sched_setscheduler(struct task_struct *p, int policy,
				struct sched_param *param, bool user)
L
Linus Torvalds 已提交
4704
{
4705
	int retval, oldprio, oldpolicy = -1, on_rq, running;
L
Linus Torvalds 已提交
4706
	unsigned long flags;
4707
	const struct sched_class *prev_class;
4708
	struct rq *rq;
4709
	int reset_on_fork;
L
Linus Torvalds 已提交
4710

4711 4712
	/* may grab non-irq protected spin_locks */
	BUG_ON(in_interrupt());
L
Linus Torvalds 已提交
4713 4714
recheck:
	/* double check policy once rq lock held */
4715 4716
	if (policy < 0) {
		reset_on_fork = p->sched_reset_on_fork;
L
Linus Torvalds 已提交
4717
		policy = oldpolicy = p->policy;
4718 4719 4720 4721 4722 4723 4724 4725 4726 4727
	} else {
		reset_on_fork = !!(policy & SCHED_RESET_ON_FORK);
		policy &= ~SCHED_RESET_ON_FORK;

		if (policy != SCHED_FIFO && policy != SCHED_RR &&
				policy != SCHED_NORMAL && policy != SCHED_BATCH &&
				policy != SCHED_IDLE)
			return -EINVAL;
	}

L
Linus Torvalds 已提交
4728 4729
	/*
	 * Valid priorities for SCHED_FIFO and SCHED_RR are
I
Ingo Molnar 已提交
4730 4731
	 * 1..MAX_USER_RT_PRIO-1, valid priority for SCHED_NORMAL,
	 * SCHED_BATCH and SCHED_IDLE is 0.
L
Linus Torvalds 已提交
4732 4733
	 */
	if (param->sched_priority < 0 ||
I
Ingo Molnar 已提交
4734
	    (p->mm && param->sched_priority > MAX_USER_RT_PRIO-1) ||
4735
	    (!p->mm && param->sched_priority > MAX_RT_PRIO-1))
L
Linus Torvalds 已提交
4736
		return -EINVAL;
4737
	if (rt_policy(policy) != (param->sched_priority != 0))
L
Linus Torvalds 已提交
4738 4739
		return -EINVAL;

4740 4741 4742
	/*
	 * Allow unprivileged RT tasks to decrease priority:
	 */
4743
	if (user && !capable(CAP_SYS_NICE)) {
4744
		if (rt_policy(policy)) {
4745 4746
			unsigned long rlim_rtprio =
					task_rlimit(p, RLIMIT_RTPRIO);
4747 4748 4749 4750 4751 4752 4753 4754 4755 4756

			/* can't set/change the rt policy */
			if (policy != p->policy && !rlim_rtprio)
				return -EPERM;

			/* can't increase priority */
			if (param->sched_priority > p->rt_priority &&
			    param->sched_priority > rlim_rtprio)
				return -EPERM;
		}
I
Ingo Molnar 已提交
4757 4758 4759 4760 4761 4762
		/*
		 * Like positive nice levels, dont allow tasks to
		 * move out of SCHED_IDLE either:
		 */
		if (p->policy == SCHED_IDLE && policy != SCHED_IDLE)
			return -EPERM;
4763

4764
		/* can't change other user's priorities */
4765
		if (!check_same_owner(p))
4766
			return -EPERM;
4767 4768 4769 4770

		/* Normal users shall not reset the sched_reset_on_fork flag */
		if (p->sched_reset_on_fork && !reset_on_fork)
			return -EPERM;
4771
	}
L
Linus Torvalds 已提交
4772

4773 4774 4775 4776 4777 4778
	if (user) {
		retval = security_task_setscheduler(p, policy, param);
		if (retval)
			return retval;
	}

4779 4780 4781 4782
	/*
	 * make sure no PI-waiters arrive (or leave) while we are
	 * changing the priority of the task:
	 */
4783
	raw_spin_lock_irqsave(&p->pi_lock, flags);
L
Linus Torvalds 已提交
4784 4785 4786 4787
	/*
	 * To be able to change p->policy safely, the apropriate
	 * runqueue lock must be held.
	 */
4788
	rq = __task_rq_lock(p);
4789

4790 4791 4792 4793 4794 4795 4796 4797 4798
	/*
	 * Changing the policy of the stop threads its a very bad idea
	 */
	if (p == rq->stop) {
		__task_rq_unlock(rq);
		raw_spin_unlock_irqrestore(&p->pi_lock, flags);
		return -EINVAL;
	}

4799 4800 4801 4802 4803 4804 4805 4806 4807 4808 4809 4810 4811 4812 4813
#ifdef CONFIG_RT_GROUP_SCHED
	if (user) {
		/*
		 * Do not allow realtime tasks into groups that have no runtime
		 * assigned.
		 */
		if (rt_bandwidth_enabled() && rt_policy(policy) &&
				task_group(p)->rt_bandwidth.rt_runtime == 0) {
			__task_rq_unlock(rq);
			raw_spin_unlock_irqrestore(&p->pi_lock, flags);
			return -EPERM;
		}
	}
#endif

L
Linus Torvalds 已提交
4814 4815 4816
	/* recheck policy now with rq lock held */
	if (unlikely(oldpolicy != -1 && oldpolicy != p->policy)) {
		policy = oldpolicy = -1;
4817
		__task_rq_unlock(rq);
4818
		raw_spin_unlock_irqrestore(&p->pi_lock, flags);
L
Linus Torvalds 已提交
4819 4820
		goto recheck;
	}
I
Ingo Molnar 已提交
4821
	on_rq = p->se.on_rq;
4822
	running = task_current(rq, p);
4823
	if (on_rq)
4824
		deactivate_task(rq, p, 0);
4825 4826
	if (running)
		p->sched_class->put_prev_task(rq, p);
4827

4828 4829
	p->sched_reset_on_fork = reset_on_fork;

L
Linus Torvalds 已提交
4830
	oldprio = p->prio;
4831
	prev_class = p->sched_class;
I
Ingo Molnar 已提交
4832
	__setscheduler(rq, p, policy, param->sched_priority);
4833

4834 4835
	if (running)
		p->sched_class->set_curr_task(rq);
I
Ingo Molnar 已提交
4836 4837
	if (on_rq) {
		activate_task(rq, p, 0);
4838 4839

		check_class_changed(rq, p, prev_class, oldprio, running);
L
Linus Torvalds 已提交
4840
	}
4841
	__task_rq_unlock(rq);
4842
	raw_spin_unlock_irqrestore(&p->pi_lock, flags);
4843

4844 4845
	rt_mutex_adjust_pi(p);

L
Linus Torvalds 已提交
4846 4847
	return 0;
}
4848 4849 4850 4851 4852 4853 4854 4855 4856 4857 4858 4859 4860 4861

/**
 * sched_setscheduler - change the scheduling policy and/or RT priority of a thread.
 * @p: the task in question.
 * @policy: new policy.
 * @param: structure containing the new RT priority.
 *
 * NOTE that the task may be already dead.
 */
int sched_setscheduler(struct task_struct *p, int policy,
		       struct sched_param *param)
{
	return __sched_setscheduler(p, policy, param, true);
}
L
Linus Torvalds 已提交
4862 4863
EXPORT_SYMBOL_GPL(sched_setscheduler);

4864 4865 4866 4867 4868 4869 4870 4871 4872 4873 4874 4875 4876 4877 4878 4879 4880
/**
 * sched_setscheduler_nocheck - change the scheduling policy and/or RT priority of a thread from kernelspace.
 * @p: the task in question.
 * @policy: new policy.
 * @param: structure containing the new RT priority.
 *
 * Just like sched_setscheduler, only don't bother checking if the
 * current context has permission.  For example, this is needed in
 * stop_machine(): we create temporary high priority worker threads,
 * but our caller might not have that capability.
 */
int sched_setscheduler_nocheck(struct task_struct *p, int policy,
			       struct sched_param *param)
{
	return __sched_setscheduler(p, policy, param, false);
}

I
Ingo Molnar 已提交
4881 4882
static int
do_sched_setscheduler(pid_t pid, int policy, struct sched_param __user *param)
L
Linus Torvalds 已提交
4883 4884 4885
{
	struct sched_param lparam;
	struct task_struct *p;
4886
	int retval;
L
Linus Torvalds 已提交
4887 4888 4889 4890 4891

	if (!param || pid < 0)
		return -EINVAL;
	if (copy_from_user(&lparam, param, sizeof(struct sched_param)))
		return -EFAULT;
4892 4893 4894

	rcu_read_lock();
	retval = -ESRCH;
L
Linus Torvalds 已提交
4895
	p = find_process_by_pid(pid);
4896 4897 4898
	if (p != NULL)
		retval = sched_setscheduler(p, policy, &lparam);
	rcu_read_unlock();
4899

L
Linus Torvalds 已提交
4900 4901 4902 4903 4904 4905 4906 4907 4908
	return retval;
}

/**
 * sys_sched_setscheduler - set/change the scheduler policy and RT priority
 * @pid: the pid in question.
 * @policy: new policy.
 * @param: structure containing the new RT priority.
 */
4909 4910
SYSCALL_DEFINE3(sched_setscheduler, pid_t, pid, int, policy,
		struct sched_param __user *, param)
L
Linus Torvalds 已提交
4911
{
4912 4913 4914 4915
	/* negative values for policy are not valid */
	if (policy < 0)
		return -EINVAL;

L
Linus Torvalds 已提交
4916 4917 4918 4919 4920 4921 4922 4923
	return do_sched_setscheduler(pid, policy, param);
}

/**
 * sys_sched_setparam - set/change the RT priority of a thread
 * @pid: the pid in question.
 * @param: structure containing the new RT priority.
 */
4924
SYSCALL_DEFINE2(sched_setparam, pid_t, pid, struct sched_param __user *, param)
L
Linus Torvalds 已提交
4925 4926 4927 4928 4929 4930 4931 4932
{
	return do_sched_setscheduler(pid, -1, param);
}

/**
 * sys_sched_getscheduler - get the policy (scheduling class) of a thread
 * @pid: the pid in question.
 */
4933
SYSCALL_DEFINE1(sched_getscheduler, pid_t, pid)
L
Linus Torvalds 已提交
4934
{
4935
	struct task_struct *p;
4936
	int retval;
L
Linus Torvalds 已提交
4937 4938

	if (pid < 0)
4939
		return -EINVAL;
L
Linus Torvalds 已提交
4940 4941

	retval = -ESRCH;
4942
	rcu_read_lock();
L
Linus Torvalds 已提交
4943 4944 4945 4946
	p = find_process_by_pid(pid);
	if (p) {
		retval = security_task_getscheduler(p);
		if (!retval)
4947 4948
			retval = p->policy
				| (p->sched_reset_on_fork ? SCHED_RESET_ON_FORK : 0);
L
Linus Torvalds 已提交
4949
	}
4950
	rcu_read_unlock();
L
Linus Torvalds 已提交
4951 4952 4953 4954
	return retval;
}

/**
4955
 * sys_sched_getparam - get the RT priority of a thread
L
Linus Torvalds 已提交
4956 4957 4958
 * @pid: the pid in question.
 * @param: structure containing the RT priority.
 */
4959
SYSCALL_DEFINE2(sched_getparam, pid_t, pid, struct sched_param __user *, param)
L
Linus Torvalds 已提交
4960 4961
{
	struct sched_param lp;
4962
	struct task_struct *p;
4963
	int retval;
L
Linus Torvalds 已提交
4964 4965

	if (!param || pid < 0)
4966
		return -EINVAL;
L
Linus Torvalds 已提交
4967

4968
	rcu_read_lock();
L
Linus Torvalds 已提交
4969 4970 4971 4972 4973 4974 4975 4976 4977 4978
	p = find_process_by_pid(pid);
	retval = -ESRCH;
	if (!p)
		goto out_unlock;

	retval = security_task_getscheduler(p);
	if (retval)
		goto out_unlock;

	lp.sched_priority = p->rt_priority;
4979
	rcu_read_unlock();
L
Linus Torvalds 已提交
4980 4981 4982 4983 4984 4985 4986 4987 4988

	/*
	 * This one might sleep, we cannot do it with a spinlock held ...
	 */
	retval = copy_to_user(param, &lp, sizeof(*param)) ? -EFAULT : 0;

	return retval;

out_unlock:
4989
	rcu_read_unlock();
L
Linus Torvalds 已提交
4990 4991 4992
	return retval;
}

4993
long sched_setaffinity(pid_t pid, const struct cpumask *in_mask)
L
Linus Torvalds 已提交
4994
{
4995
	cpumask_var_t cpus_allowed, new_mask;
4996 4997
	struct task_struct *p;
	int retval;
L
Linus Torvalds 已提交
4998

4999
	get_online_cpus();
5000
	rcu_read_lock();
L
Linus Torvalds 已提交
5001 5002 5003

	p = find_process_by_pid(pid);
	if (!p) {
5004
		rcu_read_unlock();
5005
		put_online_cpus();
L
Linus Torvalds 已提交
5006 5007 5008
		return -ESRCH;
	}

5009
	/* Prevent p going away */
L
Linus Torvalds 已提交
5010
	get_task_struct(p);
5011
	rcu_read_unlock();
L
Linus Torvalds 已提交
5012

5013 5014 5015 5016 5017 5018 5019 5020
	if (!alloc_cpumask_var(&cpus_allowed, GFP_KERNEL)) {
		retval = -ENOMEM;
		goto out_put_task;
	}
	if (!alloc_cpumask_var(&new_mask, GFP_KERNEL)) {
		retval = -ENOMEM;
		goto out_free_cpus_allowed;
	}
L
Linus Torvalds 已提交
5021
	retval = -EPERM;
5022
	if (!check_same_owner(p) && !capable(CAP_SYS_NICE))
L
Linus Torvalds 已提交
5023 5024
		goto out_unlock;

5025 5026 5027 5028
	retval = security_task_setscheduler(p, 0, NULL);
	if (retval)
		goto out_unlock;

5029 5030
	cpuset_cpus_allowed(p, cpus_allowed);
	cpumask_and(new_mask, in_mask, cpus_allowed);
P
Peter Zijlstra 已提交
5031
again:
5032
	retval = set_cpus_allowed_ptr(p, new_mask);
L
Linus Torvalds 已提交
5033

P
Paul Menage 已提交
5034
	if (!retval) {
5035 5036
		cpuset_cpus_allowed(p, cpus_allowed);
		if (!cpumask_subset(new_mask, cpus_allowed)) {
P
Paul Menage 已提交
5037 5038 5039 5040 5041
			/*
			 * We must have raced with a concurrent cpuset
			 * update. Just reset the cpus_allowed to the
			 * cpuset's cpus_allowed
			 */
5042
			cpumask_copy(new_mask, cpus_allowed);
P
Paul Menage 已提交
5043 5044 5045
			goto again;
		}
	}
L
Linus Torvalds 已提交
5046
out_unlock:
5047 5048 5049 5050
	free_cpumask_var(new_mask);
out_free_cpus_allowed:
	free_cpumask_var(cpus_allowed);
out_put_task:
L
Linus Torvalds 已提交
5051
	put_task_struct(p);
5052
	put_online_cpus();
L
Linus Torvalds 已提交
5053 5054 5055 5056
	return retval;
}

static int get_user_cpu_mask(unsigned long __user *user_mask_ptr, unsigned len,
5057
			     struct cpumask *new_mask)
L
Linus Torvalds 已提交
5058
{
5059 5060 5061 5062 5063
	if (len < cpumask_size())
		cpumask_clear(new_mask);
	else if (len > cpumask_size())
		len = cpumask_size();

L
Linus Torvalds 已提交
5064 5065 5066 5067 5068 5069 5070 5071 5072
	return copy_from_user(new_mask, user_mask_ptr, len) ? -EFAULT : 0;
}

/**
 * sys_sched_setaffinity - set the cpu affinity of a process
 * @pid: pid of the process
 * @len: length in bytes of the bitmask pointed to by user_mask_ptr
 * @user_mask_ptr: user-space pointer to the new cpu mask
 */
5073 5074
SYSCALL_DEFINE3(sched_setaffinity, pid_t, pid, unsigned int, len,
		unsigned long __user *, user_mask_ptr)
L
Linus Torvalds 已提交
5075
{
5076
	cpumask_var_t new_mask;
L
Linus Torvalds 已提交
5077 5078
	int retval;

5079 5080
	if (!alloc_cpumask_var(&new_mask, GFP_KERNEL))
		return -ENOMEM;
L
Linus Torvalds 已提交
5081

5082 5083 5084 5085 5086
	retval = get_user_cpu_mask(user_mask_ptr, len, new_mask);
	if (retval == 0)
		retval = sched_setaffinity(pid, new_mask);
	free_cpumask_var(new_mask);
	return retval;
L
Linus Torvalds 已提交
5087 5088
}

5089
long sched_getaffinity(pid_t pid, struct cpumask *mask)
L
Linus Torvalds 已提交
5090
{
5091
	struct task_struct *p;
5092 5093
	unsigned long flags;
	struct rq *rq;
L
Linus Torvalds 已提交
5094 5095
	int retval;

5096
	get_online_cpus();
5097
	rcu_read_lock();
L
Linus Torvalds 已提交
5098 5099 5100 5101 5102 5103

	retval = -ESRCH;
	p = find_process_by_pid(pid);
	if (!p)
		goto out_unlock;

5104 5105 5106 5107
	retval = security_task_getscheduler(p);
	if (retval)
		goto out_unlock;

5108
	rq = task_rq_lock(p, &flags);
5109
	cpumask_and(mask, &p->cpus_allowed, cpu_online_mask);
5110
	task_rq_unlock(rq, &flags);
L
Linus Torvalds 已提交
5111 5112

out_unlock:
5113
	rcu_read_unlock();
5114
	put_online_cpus();
L
Linus Torvalds 已提交
5115

5116
	return retval;
L
Linus Torvalds 已提交
5117 5118 5119 5120 5121 5122 5123 5124
}

/**
 * sys_sched_getaffinity - get the cpu affinity of a process
 * @pid: pid of the process
 * @len: length in bytes of the bitmask pointed to by user_mask_ptr
 * @user_mask_ptr: user-space pointer to hold the current cpu mask
 */
5125 5126
SYSCALL_DEFINE3(sched_getaffinity, pid_t, pid, unsigned int, len,
		unsigned long __user *, user_mask_ptr)
L
Linus Torvalds 已提交
5127 5128
{
	int ret;
5129
	cpumask_var_t mask;
L
Linus Torvalds 已提交
5130

A
Anton Blanchard 已提交
5131
	if ((len * BITS_PER_BYTE) < nr_cpu_ids)
5132 5133
		return -EINVAL;
	if (len & (sizeof(unsigned long)-1))
L
Linus Torvalds 已提交
5134 5135
		return -EINVAL;

5136 5137
	if (!alloc_cpumask_var(&mask, GFP_KERNEL))
		return -ENOMEM;
L
Linus Torvalds 已提交
5138

5139 5140
	ret = sched_getaffinity(pid, mask);
	if (ret == 0) {
5141
		size_t retlen = min_t(size_t, len, cpumask_size());
5142 5143

		if (copy_to_user(user_mask_ptr, mask, retlen))
5144 5145
			ret = -EFAULT;
		else
5146
			ret = retlen;
5147 5148
	}
	free_cpumask_var(mask);
L
Linus Torvalds 已提交
5149

5150
	return ret;
L
Linus Torvalds 已提交
5151 5152 5153 5154 5155
}

/**
 * sys_sched_yield - yield the current processor to other threads.
 *
I
Ingo Molnar 已提交
5156 5157
 * This function yields the current CPU to other tasks. If there are no
 * other threads running on this CPU then this function will return.
L
Linus Torvalds 已提交
5158
 */
5159
SYSCALL_DEFINE0(sched_yield)
L
Linus Torvalds 已提交
5160
{
5161
	struct rq *rq = this_rq_lock();
L
Linus Torvalds 已提交
5162

5163
	schedstat_inc(rq, yld_count);
5164
	current->sched_class->yield_task(rq);
L
Linus Torvalds 已提交
5165 5166 5167 5168 5169 5170

	/*
	 * Since we are going to call schedule() anyway, there's
	 * no need to preempt or enable interrupts:
	 */
	__release(rq->lock);
5171
	spin_release(&rq->lock.dep_map, 1, _THIS_IP_);
5172
	do_raw_spin_unlock(&rq->lock);
L
Linus Torvalds 已提交
5173 5174 5175 5176 5177 5178 5179
	preempt_enable_no_resched();

	schedule();

	return 0;
}

P
Peter Zijlstra 已提交
5180 5181 5182 5183 5184
static inline int should_resched(void)
{
	return need_resched() && !(preempt_count() & PREEMPT_ACTIVE);
}

A
Andrew Morton 已提交
5185
static void __cond_resched(void)
L
Linus Torvalds 已提交
5186
{
5187 5188 5189
	add_preempt_count(PREEMPT_ACTIVE);
	schedule();
	sub_preempt_count(PREEMPT_ACTIVE);
L
Linus Torvalds 已提交
5190 5191
}

5192
int __sched _cond_resched(void)
L
Linus Torvalds 已提交
5193
{
P
Peter Zijlstra 已提交
5194
	if (should_resched()) {
L
Linus Torvalds 已提交
5195 5196 5197 5198 5199
		__cond_resched();
		return 1;
	}
	return 0;
}
5200
EXPORT_SYMBOL(_cond_resched);
L
Linus Torvalds 已提交
5201 5202

/*
5203
 * __cond_resched_lock() - if a reschedule is pending, drop the given lock,
L
Linus Torvalds 已提交
5204 5205
 * call schedule, and on return reacquire the lock.
 *
I
Ingo Molnar 已提交
5206
 * This works OK both with and without CONFIG_PREEMPT. We do strange low-level
L
Linus Torvalds 已提交
5207 5208 5209
 * operations here to prevent schedule() from being called twice (once via
 * spin_unlock(), once by hand).
 */
5210
int __cond_resched_lock(spinlock_t *lock)
L
Linus Torvalds 已提交
5211
{
P
Peter Zijlstra 已提交
5212
	int resched = should_resched();
J
Jan Kara 已提交
5213 5214
	int ret = 0;

5215 5216
	lockdep_assert_held(lock);

N
Nick Piggin 已提交
5217
	if (spin_needbreak(lock) || resched) {
L
Linus Torvalds 已提交
5218
		spin_unlock(lock);
P
Peter Zijlstra 已提交
5219
		if (resched)
N
Nick Piggin 已提交
5220 5221 5222
			__cond_resched();
		else
			cpu_relax();
J
Jan Kara 已提交
5223
		ret = 1;
L
Linus Torvalds 已提交
5224 5225
		spin_lock(lock);
	}
J
Jan Kara 已提交
5226
	return ret;
L
Linus Torvalds 已提交
5227
}
5228
EXPORT_SYMBOL(__cond_resched_lock);
L
Linus Torvalds 已提交
5229

5230
int __sched __cond_resched_softirq(void)
L
Linus Torvalds 已提交
5231 5232 5233
{
	BUG_ON(!in_softirq());

P
Peter Zijlstra 已提交
5234
	if (should_resched()) {
5235
		local_bh_enable();
L
Linus Torvalds 已提交
5236 5237 5238 5239 5240 5241
		__cond_resched();
		local_bh_disable();
		return 1;
	}
	return 0;
}
5242
EXPORT_SYMBOL(__cond_resched_softirq);
L
Linus Torvalds 已提交
5243 5244 5245 5246

/**
 * yield - yield the current processor to other threads.
 *
5247
 * This is a shortcut for kernel-space yielding - it marks the
L
Linus Torvalds 已提交
5248 5249 5250 5251 5252 5253 5254 5255 5256 5257
 * thread runnable and calls sys_sched_yield().
 */
void __sched yield(void)
{
	set_current_state(TASK_RUNNING);
	sys_sched_yield();
}
EXPORT_SYMBOL(yield);

/*
I
Ingo Molnar 已提交
5258
 * This task is about to go to sleep on IO. Increment rq->nr_iowait so
L
Linus Torvalds 已提交
5259 5260 5261 5262
 * that process accounting knows that this is a task in IO wait state.
 */
void __sched io_schedule(void)
{
5263
	struct rq *rq = raw_rq();
L
Linus Torvalds 已提交
5264

5265
	delayacct_blkio_start();
L
Linus Torvalds 已提交
5266
	atomic_inc(&rq->nr_iowait);
5267
	current->in_iowait = 1;
L
Linus Torvalds 已提交
5268
	schedule();
5269
	current->in_iowait = 0;
L
Linus Torvalds 已提交
5270
	atomic_dec(&rq->nr_iowait);
5271
	delayacct_blkio_end();
L
Linus Torvalds 已提交
5272 5273 5274 5275 5276
}
EXPORT_SYMBOL(io_schedule);

long __sched io_schedule_timeout(long timeout)
{
5277
	struct rq *rq = raw_rq();
L
Linus Torvalds 已提交
5278 5279
	long ret;

5280
	delayacct_blkio_start();
L
Linus Torvalds 已提交
5281
	atomic_inc(&rq->nr_iowait);
5282
	current->in_iowait = 1;
L
Linus Torvalds 已提交
5283
	ret = schedule_timeout(timeout);
5284
	current->in_iowait = 0;
L
Linus Torvalds 已提交
5285
	atomic_dec(&rq->nr_iowait);
5286
	delayacct_blkio_end();
L
Linus Torvalds 已提交
5287 5288 5289 5290 5291 5292 5293 5294 5295 5296
	return ret;
}

/**
 * sys_sched_get_priority_max - return maximum RT priority.
 * @policy: scheduling class.
 *
 * this syscall returns the maximum rt_priority that can be used
 * by a given scheduling class.
 */
5297
SYSCALL_DEFINE1(sched_get_priority_max, int, policy)
L
Linus Torvalds 已提交
5298 5299 5300 5301 5302 5303 5304 5305 5306
{
	int ret = -EINVAL;

	switch (policy) {
	case SCHED_FIFO:
	case SCHED_RR:
		ret = MAX_USER_RT_PRIO-1;
		break;
	case SCHED_NORMAL:
5307
	case SCHED_BATCH:
I
Ingo Molnar 已提交
5308
	case SCHED_IDLE:
L
Linus Torvalds 已提交
5309 5310 5311 5312 5313 5314 5315 5316 5317 5318 5319 5320 5321
		ret = 0;
		break;
	}
	return ret;
}

/**
 * sys_sched_get_priority_min - return minimum RT priority.
 * @policy: scheduling class.
 *
 * this syscall returns the minimum rt_priority that can be used
 * by a given scheduling class.
 */
5322
SYSCALL_DEFINE1(sched_get_priority_min, int, policy)
L
Linus Torvalds 已提交
5323 5324 5325 5326 5327 5328 5329 5330 5331
{
	int ret = -EINVAL;

	switch (policy) {
	case SCHED_FIFO:
	case SCHED_RR:
		ret = 1;
		break;
	case SCHED_NORMAL:
5332
	case SCHED_BATCH:
I
Ingo Molnar 已提交
5333
	case SCHED_IDLE:
L
Linus Torvalds 已提交
5334 5335 5336 5337 5338 5339 5340 5341 5342 5343 5344 5345 5346
		ret = 0;
	}
	return ret;
}

/**
 * sys_sched_rr_get_interval - return the default timeslice of a process.
 * @pid: pid of the process.
 * @interval: userspace pointer to the timeslice value.
 *
 * this syscall writes the default timeslice value of a given process
 * into the user-space timespec buffer. A value of '0' means infinity.
 */
5347
SYSCALL_DEFINE2(sched_rr_get_interval, pid_t, pid,
5348
		struct timespec __user *, interval)
L
Linus Torvalds 已提交
5349
{
5350
	struct task_struct *p;
D
Dmitry Adamushko 已提交
5351
	unsigned int time_slice;
5352 5353
	unsigned long flags;
	struct rq *rq;
5354
	int retval;
L
Linus Torvalds 已提交
5355 5356 5357
	struct timespec t;

	if (pid < 0)
5358
		return -EINVAL;
L
Linus Torvalds 已提交
5359 5360

	retval = -ESRCH;
5361
	rcu_read_lock();
L
Linus Torvalds 已提交
5362 5363 5364 5365 5366 5367 5368 5369
	p = find_process_by_pid(pid);
	if (!p)
		goto out_unlock;

	retval = security_task_getscheduler(p);
	if (retval)
		goto out_unlock;

5370 5371 5372
	rq = task_rq_lock(p, &flags);
	time_slice = p->sched_class->get_rr_interval(rq, p);
	task_rq_unlock(rq, &flags);
D
Dmitry Adamushko 已提交
5373

5374
	rcu_read_unlock();
D
Dmitry Adamushko 已提交
5375
	jiffies_to_timespec(time_slice, &t);
L
Linus Torvalds 已提交
5376 5377
	retval = copy_to_user(interval, &t, sizeof(t)) ? -EFAULT : 0;
	return retval;
5378

L
Linus Torvalds 已提交
5379
out_unlock:
5380
	rcu_read_unlock();
L
Linus Torvalds 已提交
5381 5382 5383
	return retval;
}

5384
static const char stat_nam[] = TASK_STATE_TO_CHAR_STR;
5385

5386
void sched_show_task(struct task_struct *p)
L
Linus Torvalds 已提交
5387 5388
{
	unsigned long free = 0;
5389
	unsigned state;
L
Linus Torvalds 已提交
5390 5391

	state = p->state ? __ffs(p->state) + 1 : 0;
P
Peter Zijlstra 已提交
5392
	printk(KERN_INFO "%-13.13s %c", p->comm,
5393
		state < sizeof(stat_nam) - 1 ? stat_nam[state] : '?');
5394
#if BITS_PER_LONG == 32
L
Linus Torvalds 已提交
5395
	if (state == TASK_RUNNING)
P
Peter Zijlstra 已提交
5396
		printk(KERN_CONT " running  ");
L
Linus Torvalds 已提交
5397
	else
P
Peter Zijlstra 已提交
5398
		printk(KERN_CONT " %08lx ", thread_saved_pc(p));
L
Linus Torvalds 已提交
5399 5400
#else
	if (state == TASK_RUNNING)
P
Peter Zijlstra 已提交
5401
		printk(KERN_CONT "  running task    ");
L
Linus Torvalds 已提交
5402
	else
P
Peter Zijlstra 已提交
5403
		printk(KERN_CONT " %016lx ", thread_saved_pc(p));
L
Linus Torvalds 已提交
5404 5405
#endif
#ifdef CONFIG_DEBUG_STACK_USAGE
5406
	free = stack_not_used(p);
L
Linus Torvalds 已提交
5407
#endif
P
Peter Zijlstra 已提交
5408
	printk(KERN_CONT "%5lu %5d %6d 0x%08lx\n", free,
5409 5410
		task_pid_nr(p), task_pid_nr(p->real_parent),
		(unsigned long)task_thread_info(p)->flags);
L
Linus Torvalds 已提交
5411

5412
	show_stack(p, NULL);
L
Linus Torvalds 已提交
5413 5414
}

I
Ingo Molnar 已提交
5415
void show_state_filter(unsigned long state_filter)
L
Linus Torvalds 已提交
5416
{
5417
	struct task_struct *g, *p;
L
Linus Torvalds 已提交
5418

5419
#if BITS_PER_LONG == 32
P
Peter Zijlstra 已提交
5420 5421
	printk(KERN_INFO
		"  task                PC stack   pid father\n");
L
Linus Torvalds 已提交
5422
#else
P
Peter Zijlstra 已提交
5423 5424
	printk(KERN_INFO
		"  task                        PC stack   pid father\n");
L
Linus Torvalds 已提交
5425 5426 5427 5428 5429 5430 5431 5432
#endif
	read_lock(&tasklist_lock);
	do_each_thread(g, p) {
		/*
		 * reset the NMI-timeout, listing all files on a slow
		 * console might take alot of time:
		 */
		touch_nmi_watchdog();
I
Ingo Molnar 已提交
5433
		if (!state_filter || (p->state & state_filter))
5434
			sched_show_task(p);
L
Linus Torvalds 已提交
5435 5436
	} while_each_thread(g, p);

5437 5438
	touch_all_softlockup_watchdogs();

I
Ingo Molnar 已提交
5439 5440 5441
#ifdef CONFIG_SCHED_DEBUG
	sysrq_sched_debug_show();
#endif
L
Linus Torvalds 已提交
5442
	read_unlock(&tasklist_lock);
I
Ingo Molnar 已提交
5443 5444 5445
	/*
	 * Only show locks if all tasks are dumped:
	 */
5446
	if (!state_filter)
I
Ingo Molnar 已提交
5447
		debug_show_all_locks();
L
Linus Torvalds 已提交
5448 5449
}

I
Ingo Molnar 已提交
5450 5451
void __cpuinit init_idle_bootup_task(struct task_struct *idle)
{
I
Ingo Molnar 已提交
5452
	idle->sched_class = &idle_sched_class;
I
Ingo Molnar 已提交
5453 5454
}

5455 5456 5457 5458 5459 5460 5461 5462
/**
 * init_idle - set up an idle thread for a given CPU
 * @idle: task in question
 * @cpu: cpu the idle task belongs to
 *
 * NOTE: this function does not set the idle thread's NEED_RESCHED
 * flag, to make booting more robust.
 */
5463
void __cpuinit init_idle(struct task_struct *idle, int cpu)
L
Linus Torvalds 已提交
5464
{
5465
	struct rq *rq = cpu_rq(cpu);
L
Linus Torvalds 已提交
5466 5467
	unsigned long flags;

5468
	raw_spin_lock_irqsave(&rq->lock, flags);
5469

I
Ingo Molnar 已提交
5470
	__sched_fork(idle);
5471
	idle->state = TASK_RUNNING;
I
Ingo Molnar 已提交
5472 5473
	idle->se.exec_start = sched_clock();

5474
	cpumask_copy(&idle->cpus_allowed, cpumask_of(cpu));
I
Ingo Molnar 已提交
5475
	__set_task_cpu(idle, cpu);
L
Linus Torvalds 已提交
5476 5477

	rq->curr = rq->idle = idle;
5478 5479 5480
#if defined(CONFIG_SMP) && defined(__ARCH_WANT_UNLOCKED_CTXSW)
	idle->oncpu = 1;
#endif
5481
	raw_spin_unlock_irqrestore(&rq->lock, flags);
L
Linus Torvalds 已提交
5482 5483

	/* Set the preempt count _outside_ the spinlocks! */
5484 5485 5486
#if defined(CONFIG_PREEMPT)
	task_thread_info(idle)->preempt_count = (idle->lock_depth >= 0);
#else
A
Al Viro 已提交
5487
	task_thread_info(idle)->preempt_count = 0;
5488
#endif
I
Ingo Molnar 已提交
5489 5490 5491 5492
	/*
	 * The idle tasks have their own, simple scheduling class:
	 */
	idle->sched_class = &idle_sched_class;
5493
	ftrace_graph_init_task(idle);
L
Linus Torvalds 已提交
5494 5495 5496 5497 5498 5499 5500
}

/*
 * In a system that switches off the HZ timer nohz_cpu_mask
 * indicates which cpus entered this state. This is used
 * in the rcu update to wait only for active cpus. For system
 * which do not switch off the HZ timer nohz_cpu_mask should
5501
 * always be CPU_BITS_NONE.
L
Linus Torvalds 已提交
5502
 */
5503
cpumask_var_t nohz_cpu_mask;
L
Linus Torvalds 已提交
5504

I
Ingo Molnar 已提交
5505 5506 5507 5508 5509 5510 5511 5512 5513
/*
 * Increase the granularity value when there are more CPUs,
 * because with more CPUs the 'effective latency' as visible
 * to users decreases. But the relationship is not linear,
 * so pick a second-best guess by going with the log2 of the
 * number of CPUs.
 *
 * This idea comes from the SD scheduler of Con Kolivas:
 */
5514
static int get_update_sysctl_factor(void)
I
Ingo Molnar 已提交
5515
{
5516
	unsigned int cpus = min_t(int, num_online_cpus(), 8);
5517 5518 5519 5520 5521 5522 5523 5524 5525 5526 5527 5528 5529 5530
	unsigned int factor;

	switch (sysctl_sched_tunable_scaling) {
	case SCHED_TUNABLESCALING_NONE:
		factor = 1;
		break;
	case SCHED_TUNABLESCALING_LINEAR:
		factor = cpus;
		break;
	case SCHED_TUNABLESCALING_LOG:
	default:
		factor = 1 + ilog2(cpus);
		break;
	}
I
Ingo Molnar 已提交
5531

5532 5533
	return factor;
}
I
Ingo Molnar 已提交
5534

5535 5536 5537
static void update_sysctl(void)
{
	unsigned int factor = get_update_sysctl_factor();
I
Ingo Molnar 已提交
5538

5539 5540 5541 5542 5543 5544 5545 5546
#define SET_SYSCTL(name) \
	(sysctl_##name = (factor) * normalized_sysctl_##name)
	SET_SYSCTL(sched_min_granularity);
	SET_SYSCTL(sched_latency);
	SET_SYSCTL(sched_wakeup_granularity);
	SET_SYSCTL(sched_shares_ratelimit);
#undef SET_SYSCTL
}
5547

5548 5549 5550
static inline void sched_init_granularity(void)
{
	update_sysctl();
I
Ingo Molnar 已提交
5551 5552
}

L
Linus Torvalds 已提交
5553 5554 5555 5556
#ifdef CONFIG_SMP
/*
 * This is how migration works:
 *
5557 5558 5559 5560 5561 5562
 * 1) we invoke migration_cpu_stop() on the target CPU using
 *    stop_one_cpu().
 * 2) stopper starts to run (implicitly forcing the migrated thread
 *    off the CPU)
 * 3) it checks whether the migrated task is still in the wrong runqueue.
 * 4) if it's in the wrong runqueue then the migration thread removes
L
Linus Torvalds 已提交
5563
 *    it and puts it into the right queue.
5564 5565
 * 5) stopper completes and stop_one_cpu() returns and the migration
 *    is done.
L
Linus Torvalds 已提交
5566 5567 5568 5569 5570 5571 5572 5573
 */

/*
 * Change a given task's CPU affinity. Migrate the thread to a
 * proper CPU and schedule it away if the CPU it's executing on
 * is removed from the allowed bitmask.
 *
 * NOTE: the caller must have a valid reference to the task, the
I
Ingo Molnar 已提交
5574
 * task must not exit() & deallocate itself prematurely. The
L
Linus Torvalds 已提交
5575 5576
 * call is not atomic; no spinlocks may be held.
 */
5577
int set_cpus_allowed_ptr(struct task_struct *p, const struct cpumask *new_mask)
L
Linus Torvalds 已提交
5578 5579
{
	unsigned long flags;
5580
	struct rq *rq;
5581
	unsigned int dest_cpu;
5582
	int ret = 0;
L
Linus Torvalds 已提交
5583

P
Peter Zijlstra 已提交
5584 5585 5586 5587 5588 5589 5590
	/*
	 * Serialize against TASK_WAKING so that ttwu() and wunt() can
	 * drop the rq->lock and still rely on ->cpus_allowed.
	 */
again:
	while (task_is_waking(p))
		cpu_relax();
L
Linus Torvalds 已提交
5591
	rq = task_rq_lock(p, &flags);
P
Peter Zijlstra 已提交
5592 5593 5594 5595
	if (task_is_waking(p)) {
		task_rq_unlock(rq, &flags);
		goto again;
	}
5596

5597
	if (!cpumask_intersects(new_mask, cpu_active_mask)) {
L
Linus Torvalds 已提交
5598 5599 5600 5601
		ret = -EINVAL;
		goto out;
	}

5602
	if (unlikely((p->flags & PF_THREAD_BOUND) && p != current &&
5603
		     !cpumask_equal(&p->cpus_allowed, new_mask))) {
5604 5605 5606 5607
		ret = -EINVAL;
		goto out;
	}

5608
	if (p->sched_class->set_cpus_allowed)
5609
		p->sched_class->set_cpus_allowed(p, new_mask);
5610
	else {
5611 5612
		cpumask_copy(&p->cpus_allowed, new_mask);
		p->rt.nr_cpus_allowed = cpumask_weight(new_mask);
5613 5614
	}

L
Linus Torvalds 已提交
5615
	/* Can the task run on the task's current CPU? If so, we're done */
5616
	if (cpumask_test_cpu(task_cpu(p), new_mask))
L
Linus Torvalds 已提交
5617 5618
		goto out;

5619 5620 5621
	dest_cpu = cpumask_any_and(cpu_active_mask, new_mask);
	if (migrate_task(p, dest_cpu)) {
		struct migration_arg arg = { p, dest_cpu };
L
Linus Torvalds 已提交
5622 5623
		/* Need help from migration thread: drop lock and wait. */
		task_rq_unlock(rq, &flags);
5624
		stop_one_cpu(cpu_of(rq), migration_cpu_stop, &arg);
L
Linus Torvalds 已提交
5625 5626 5627 5628 5629
		tlb_migrate_finish(p->mm);
		return 0;
	}
out:
	task_rq_unlock(rq, &flags);
5630

L
Linus Torvalds 已提交
5631 5632
	return ret;
}
5633
EXPORT_SYMBOL_GPL(set_cpus_allowed_ptr);
L
Linus Torvalds 已提交
5634 5635

/*
I
Ingo Molnar 已提交
5636
 * Move (not current) task off this cpu, onto dest cpu. We're doing
L
Linus Torvalds 已提交
5637 5638 5639 5640 5641 5642
 * this because either it can't run here any more (set_cpus_allowed()
 * away from this CPU, or CPU going down), or because we're
 * attempting to rebalance this task on exec (sched_exec).
 *
 * So we race with normal scheduler movements, but that's OK, as long
 * as the task is no longer on this CPU.
5643 5644
 *
 * Returns non-zero if task was successfully migrated.
L
Linus Torvalds 已提交
5645
 */
5646
static int __migrate_task(struct task_struct *p, int src_cpu, int dest_cpu)
L
Linus Torvalds 已提交
5647
{
5648
	struct rq *rq_dest, *rq_src;
5649
	int ret = 0;
L
Linus Torvalds 已提交
5650

5651
	if (unlikely(!cpu_active(dest_cpu)))
5652
		return ret;
L
Linus Torvalds 已提交
5653 5654 5655 5656 5657 5658 5659

	rq_src = cpu_rq(src_cpu);
	rq_dest = cpu_rq(dest_cpu);

	double_rq_lock(rq_src, rq_dest);
	/* Already moved. */
	if (task_cpu(p) != src_cpu)
L
Linus Torvalds 已提交
5660
		goto done;
L
Linus Torvalds 已提交
5661
	/* Affinity changed (again). */
5662
	if (!cpumask_test_cpu(dest_cpu, &p->cpus_allowed))
L
Linus Torvalds 已提交
5663
		goto fail;
L
Linus Torvalds 已提交
5664

5665 5666 5667 5668 5669
	/*
	 * If we're not on a rq, the next wake-up will ensure we're
	 * placed properly.
	 */
	if (p->se.on_rq) {
5670
		deactivate_task(rq_src, p, 0);
5671
		set_task_cpu(p, dest_cpu);
I
Ingo Molnar 已提交
5672
		activate_task(rq_dest, p, 0);
5673
		check_preempt_curr(rq_dest, p, 0);
L
Linus Torvalds 已提交
5674
	}
L
Linus Torvalds 已提交
5675
done:
5676
	ret = 1;
L
Linus Torvalds 已提交
5677
fail:
L
Linus Torvalds 已提交
5678
	double_rq_unlock(rq_src, rq_dest);
5679
	return ret;
L
Linus Torvalds 已提交
5680 5681 5682
}

/*
5683 5684 5685
 * migration_cpu_stop - this will be executed by a highprio stopper thread
 * and performs thread migration by bumping thread off CPU then
 * 'pushing' onto another runqueue.
L
Linus Torvalds 已提交
5686
 */
5687
static int migration_cpu_stop(void *data)
L
Linus Torvalds 已提交
5688
{
5689
	struct migration_arg *arg = data;
5690

5691 5692 5693 5694
	/*
	 * The original target cpu might have gone down and we might
	 * be on another cpu but it doesn't matter.
	 */
5695
	local_irq_disable();
5696
	__migrate_task(arg->task, raw_smp_processor_id(), arg->dest_cpu);
5697
	local_irq_enable();
L
Linus Torvalds 已提交
5698
	return 0;
5699 5700
}

L
Linus Torvalds 已提交
5701
#ifdef CONFIG_HOTPLUG_CPU
5702
/*
5703
 * Figure out where task on dead CPU should go, use force if necessary.
5704
 */
5705
void move_task_off_dead_cpu(int dead_cpu, struct task_struct *p)
L
Linus Torvalds 已提交
5706
{
5707 5708 5709
	struct rq *rq = cpu_rq(dead_cpu);
	int needs_cpu, uninitialized_var(dest_cpu);
	unsigned long flags;
5710

5711
	local_irq_save(flags);
5712

5713 5714 5715 5716 5717
	raw_spin_lock(&rq->lock);
	needs_cpu = (task_cpu(p) == dead_cpu) && (p->state != TASK_WAKING);
	if (needs_cpu)
		dest_cpu = select_fallback_rq(dead_cpu, p);
	raw_spin_unlock(&rq->lock);
5718 5719 5720 5721
	/*
	 * It can only fail if we race with set_cpus_allowed(),
	 * in the racer should migrate the task anyway.
	 */
5722
	if (needs_cpu)
5723
		__migrate_task(p, dead_cpu, dest_cpu);
5724
	local_irq_restore(flags);
L
Linus Torvalds 已提交
5725 5726 5727 5728 5729 5730 5731 5732 5733
}

/*
 * While a dead CPU has no uninterruptible tasks queued at this point,
 * it might still have a nonzero ->nr_uninterruptible counter, because
 * for performance reasons the counter is not stricly tracking tasks to
 * their home CPUs. So we just add the counter to another CPU's counter,
 * to keep the global sum constant after CPU-down:
 */
5734
static void migrate_nr_uninterruptible(struct rq *rq_src)
L
Linus Torvalds 已提交
5735
{
5736
	struct rq *rq_dest = cpu_rq(cpumask_any(cpu_active_mask));
L
Linus Torvalds 已提交
5737 5738 5739 5740 5741 5742 5743 5744 5745 5746 5747 5748 5749
	unsigned long flags;

	local_irq_save(flags);
	double_rq_lock(rq_src, rq_dest);
	rq_dest->nr_uninterruptible += rq_src->nr_uninterruptible;
	rq_src->nr_uninterruptible = 0;
	double_rq_unlock(rq_src, rq_dest);
	local_irq_restore(flags);
}

/* Run through task list and migrate tasks from the dead cpu. */
static void migrate_live_tasks(int src_cpu)
{
5750
	struct task_struct *p, *t;
L
Linus Torvalds 已提交
5751

5752
	read_lock(&tasklist_lock);
L
Linus Torvalds 已提交
5753

5754 5755
	do_each_thread(t, p) {
		if (p == current)
L
Linus Torvalds 已提交
5756 5757
			continue;

5758 5759 5760
		if (task_cpu(p) == src_cpu)
			move_task_off_dead_cpu(src_cpu, p);
	} while_each_thread(t, p);
L
Linus Torvalds 已提交
5761

5762
	read_unlock(&tasklist_lock);
L
Linus Torvalds 已提交
5763 5764
}

I
Ingo Molnar 已提交
5765 5766
/*
 * Schedules idle task to be the next runnable task on current CPU.
5767 5768
 * It does so by boosting its priority to highest possible.
 * Used by CPU offline code.
L
Linus Torvalds 已提交
5769 5770 5771
 */
void sched_idle_next(void)
{
5772
	int this_cpu = smp_processor_id();
5773
	struct rq *rq = cpu_rq(this_cpu);
L
Linus Torvalds 已提交
5774 5775 5776 5777
	struct task_struct *p = rq->idle;
	unsigned long flags;

	/* cpu has to be offline */
5778
	BUG_ON(cpu_online(this_cpu));
L
Linus Torvalds 已提交
5779

5780 5781 5782
	/*
	 * Strictly not necessary since rest of the CPUs are stopped by now
	 * and interrupts disabled on the current cpu.
L
Linus Torvalds 已提交
5783
	 */
5784
	raw_spin_lock_irqsave(&rq->lock, flags);
L
Linus Torvalds 已提交
5785

I
Ingo Molnar 已提交
5786
	__setscheduler(rq, p, SCHED_FIFO, MAX_RT_PRIO-1);
5787

5788
	activate_task(rq, p, 0);
L
Linus Torvalds 已提交
5789

5790
	raw_spin_unlock_irqrestore(&rq->lock, flags);
L
Linus Torvalds 已提交
5791 5792
}

5793 5794
/*
 * Ensures that the idle task is using init_mm right before its cpu goes
L
Linus Torvalds 已提交
5795 5796 5797 5798 5799 5800 5801 5802 5803 5804 5805 5806 5807
 * offline.
 */
void idle_task_exit(void)
{
	struct mm_struct *mm = current->active_mm;

	BUG_ON(cpu_online(smp_processor_id()));

	if (mm != &init_mm)
		switch_mm(mm, &init_mm, current);
	mmdrop(mm);
}

5808
/* called under rq->lock with disabled interrupts */
5809
static void migrate_dead(unsigned int dead_cpu, struct task_struct *p)
L
Linus Torvalds 已提交
5810
{
5811
	struct rq *rq = cpu_rq(dead_cpu);
L
Linus Torvalds 已提交
5812 5813

	/* Must be exiting, otherwise would be on tasklist. */
E
Eugene Teo 已提交
5814
	BUG_ON(!p->exit_state);
L
Linus Torvalds 已提交
5815 5816

	/* Cannot have done final schedule yet: would have vanished. */
5817
	BUG_ON(p->state == TASK_DEAD);
L
Linus Torvalds 已提交
5818

5819
	get_task_struct(p);
L
Linus Torvalds 已提交
5820 5821 5822

	/*
	 * Drop lock around migration; if someone else moves it,
I
Ingo Molnar 已提交
5823
	 * that's OK. No task can be added to this CPU, so iteration is
L
Linus Torvalds 已提交
5824 5825
	 * fine.
	 */
5826
	raw_spin_unlock_irq(&rq->lock);
5827
	move_task_off_dead_cpu(dead_cpu, p);
5828
	raw_spin_lock_irq(&rq->lock);
L
Linus Torvalds 已提交
5829

5830
	put_task_struct(p);
L
Linus Torvalds 已提交
5831 5832 5833 5834 5835
}

/* release_task() removes task from tasklist, so we won't find dead tasks. */
static void migrate_dead_tasks(unsigned int dead_cpu)
{
5836
	struct rq *rq = cpu_rq(dead_cpu);
I
Ingo Molnar 已提交
5837
	struct task_struct *next;
5838

I
Ingo Molnar 已提交
5839 5840 5841
	for ( ; ; ) {
		if (!rq->nr_running)
			break;
5842
		next = pick_next_task(rq);
I
Ingo Molnar 已提交
5843 5844
		if (!next)
			break;
D
Dmitry Adamushko 已提交
5845
		next->sched_class->put_prev_task(rq, next);
I
Ingo Molnar 已提交
5846
		migrate_dead(dead_cpu, next);
5847

L
Linus Torvalds 已提交
5848 5849
	}
}
5850 5851 5852 5853 5854 5855 5856

/*
 * remove the tasks which were accounted by rq from calc_load_tasks.
 */
static void calc_global_load_remove(struct rq *rq)
{
	atomic_long_sub(rq->calc_load_active, &calc_load_tasks);
5857
	rq->calc_load_active = 0;
5858
}
L
Linus Torvalds 已提交
5859 5860
#endif /* CONFIG_HOTPLUG_CPU */

5861 5862 5863
#if defined(CONFIG_SCHED_DEBUG) && defined(CONFIG_SYSCTL)

static struct ctl_table sd_ctl_dir[] = {
5864 5865
	{
		.procname	= "sched_domain",
5866
		.mode		= 0555,
5867
	},
5868
	{}
5869 5870 5871
};

static struct ctl_table sd_ctl_root[] = {
5872 5873
	{
		.procname	= "kernel",
5874
		.mode		= 0555,
5875 5876
		.child		= sd_ctl_dir,
	},
5877
	{}
5878 5879 5880 5881 5882
};

static struct ctl_table *sd_alloc_ctl_entry(int n)
{
	struct ctl_table *entry =
5883
		kcalloc(n, sizeof(struct ctl_table), GFP_KERNEL);
5884 5885 5886 5887

	return entry;
}

5888 5889
static void sd_free_ctl_entry(struct ctl_table **tablep)
{
5890
	struct ctl_table *entry;
5891

5892 5893 5894
	/*
	 * In the intermediate directories, both the child directory and
	 * procname are dynamically allocated and could fail but the mode
I
Ingo Molnar 已提交
5895
	 * will always be set. In the lowest directory the names are
5896 5897 5898
	 * static strings and all have proc handlers.
	 */
	for (entry = *tablep; entry->mode; entry++) {
5899 5900
		if (entry->child)
			sd_free_ctl_entry(&entry->child);
5901 5902 5903
		if (entry->proc_handler == NULL)
			kfree(entry->procname);
	}
5904 5905 5906 5907 5908

	kfree(*tablep);
	*tablep = NULL;
}

5909
static void
5910
set_table_entry(struct ctl_table *entry,
5911 5912 5913 5914 5915 5916 5917 5918 5919 5920 5921 5922 5923
		const char *procname, void *data, int maxlen,
		mode_t mode, proc_handler *proc_handler)
{
	entry->procname = procname;
	entry->data = data;
	entry->maxlen = maxlen;
	entry->mode = mode;
	entry->proc_handler = proc_handler;
}

static struct ctl_table *
sd_alloc_ctl_domain_table(struct sched_domain *sd)
{
5924
	struct ctl_table *table = sd_alloc_ctl_entry(13);
5925

5926 5927 5928
	if (table == NULL)
		return NULL;

5929
	set_table_entry(&table[0], "min_interval", &sd->min_interval,
5930
		sizeof(long), 0644, proc_doulongvec_minmax);
5931
	set_table_entry(&table[1], "max_interval", &sd->max_interval,
5932
		sizeof(long), 0644, proc_doulongvec_minmax);
5933
	set_table_entry(&table[2], "busy_idx", &sd->busy_idx,
5934
		sizeof(int), 0644, proc_dointvec_minmax);
5935
	set_table_entry(&table[3], "idle_idx", &sd->idle_idx,
5936
		sizeof(int), 0644, proc_dointvec_minmax);
5937
	set_table_entry(&table[4], "newidle_idx", &sd->newidle_idx,
5938
		sizeof(int), 0644, proc_dointvec_minmax);
5939
	set_table_entry(&table[5], "wake_idx", &sd->wake_idx,
5940
		sizeof(int), 0644, proc_dointvec_minmax);
5941
	set_table_entry(&table[6], "forkexec_idx", &sd->forkexec_idx,
5942
		sizeof(int), 0644, proc_dointvec_minmax);
5943
	set_table_entry(&table[7], "busy_factor", &sd->busy_factor,
5944
		sizeof(int), 0644, proc_dointvec_minmax);
5945
	set_table_entry(&table[8], "imbalance_pct", &sd->imbalance_pct,
5946
		sizeof(int), 0644, proc_dointvec_minmax);
5947
	set_table_entry(&table[9], "cache_nice_tries",
5948 5949
		&sd->cache_nice_tries,
		sizeof(int), 0644, proc_dointvec_minmax);
5950
	set_table_entry(&table[10], "flags", &sd->flags,
5951
		sizeof(int), 0644, proc_dointvec_minmax);
5952 5953 5954
	set_table_entry(&table[11], "name", sd->name,
		CORENAME_MAX_SIZE, 0444, proc_dostring);
	/* &table[12] is terminator */
5955 5956 5957 5958

	return table;
}

5959
static ctl_table *sd_alloc_ctl_cpu_table(int cpu)
5960 5961 5962 5963 5964 5965 5966 5967 5968
{
	struct ctl_table *entry, *table;
	struct sched_domain *sd;
	int domain_num = 0, i;
	char buf[32];

	for_each_domain(cpu, sd)
		domain_num++;
	entry = table = sd_alloc_ctl_entry(domain_num + 1);
5969 5970
	if (table == NULL)
		return NULL;
5971 5972 5973 5974 5975

	i = 0;
	for_each_domain(cpu, sd) {
		snprintf(buf, 32, "domain%d", i);
		entry->procname = kstrdup(buf, GFP_KERNEL);
5976
		entry->mode = 0555;
5977 5978 5979 5980 5981 5982 5983 5984
		entry->child = sd_alloc_ctl_domain_table(sd);
		entry++;
		i++;
	}
	return table;
}

static struct ctl_table_header *sd_sysctl_header;
5985
static void register_sched_domain_sysctl(void)
5986
{
5987
	int i, cpu_num = num_possible_cpus();
5988 5989 5990
	struct ctl_table *entry = sd_alloc_ctl_entry(cpu_num + 1);
	char buf[32];

5991 5992 5993
	WARN_ON(sd_ctl_dir[0].child);
	sd_ctl_dir[0].child = entry;

5994 5995 5996
	if (entry == NULL)
		return;

5997
	for_each_possible_cpu(i) {
5998 5999
		snprintf(buf, 32, "cpu%d", i);
		entry->procname = kstrdup(buf, GFP_KERNEL);
6000
		entry->mode = 0555;
6001
		entry->child = sd_alloc_ctl_cpu_table(i);
6002
		entry++;
6003
	}
6004 6005

	WARN_ON(sd_sysctl_header);
6006 6007
	sd_sysctl_header = register_sysctl_table(sd_ctl_root);
}
6008

6009
/* may be called multiple times per register */
6010 6011
static void unregister_sched_domain_sysctl(void)
{
6012 6013
	if (sd_sysctl_header)
		unregister_sysctl_table(sd_sysctl_header);
6014
	sd_sysctl_header = NULL;
6015 6016
	if (sd_ctl_dir[0].child)
		sd_free_ctl_entry(&sd_ctl_dir[0].child);
6017
}
6018
#else
6019 6020 6021 6022
static void register_sched_domain_sysctl(void)
{
}
static void unregister_sched_domain_sysctl(void)
6023 6024 6025 6026
{
}
#endif

6027 6028 6029 6030 6031
static void set_rq_online(struct rq *rq)
{
	if (!rq->online) {
		const struct sched_class *class;

6032
		cpumask_set_cpu(rq->cpu, rq->rd->online);
6033 6034 6035 6036 6037 6038 6039 6040 6041 6042 6043 6044 6045 6046 6047 6048 6049 6050 6051
		rq->online = 1;

		for_each_class(class) {
			if (class->rq_online)
				class->rq_online(rq);
		}
	}
}

static void set_rq_offline(struct rq *rq)
{
	if (rq->online) {
		const struct sched_class *class;

		for_each_class(class) {
			if (class->rq_offline)
				class->rq_offline(rq);
		}

6052
		cpumask_clear_cpu(rq->cpu, rq->rd->online);
6053 6054 6055 6056
		rq->online = 0;
	}
}

L
Linus Torvalds 已提交
6057 6058 6059 6060
/*
 * migration_call - callback that gets triggered when a CPU is added.
 * Here we can start up the necessary migration thread for the new CPU.
 */
6061 6062
static int __cpuinit
migration_call(struct notifier_block *nfb, unsigned long action, void *hcpu)
L
Linus Torvalds 已提交
6063
{
6064
	int cpu = (long)hcpu;
L
Linus Torvalds 已提交
6065
	unsigned long flags;
6066
	struct rq *rq = cpu_rq(cpu);
L
Linus Torvalds 已提交
6067 6068

	switch (action) {
6069

L
Linus Torvalds 已提交
6070
	case CPU_UP_PREPARE:
6071
	case CPU_UP_PREPARE_FROZEN:
6072
		rq->calc_load_update = calc_load_update;
L
Linus Torvalds 已提交
6073
		break;
6074

L
Linus Torvalds 已提交
6075
	case CPU_ONLINE:
6076
	case CPU_ONLINE_FROZEN:
6077
		/* Update our root-domain */
6078
		raw_spin_lock_irqsave(&rq->lock, flags);
6079
		if (rq->rd) {
6080
			BUG_ON(!cpumask_test_cpu(cpu, rq->rd->span));
6081 6082

			set_rq_online(rq);
6083
		}
6084
		raw_spin_unlock_irqrestore(&rq->lock, flags);
L
Linus Torvalds 已提交
6085
		break;
6086

L
Linus Torvalds 已提交
6087 6088
#ifdef CONFIG_HOTPLUG_CPU
	case CPU_DEAD:
6089
	case CPU_DEAD_FROZEN:
L
Linus Torvalds 已提交
6090 6091
		migrate_live_tasks(cpu);
		/* Idle task back to normal (off runqueue, low prio) */
6092
		raw_spin_lock_irq(&rq->lock);
6093
		deactivate_task(rq, rq->idle, 0);
I
Ingo Molnar 已提交
6094 6095
		__setscheduler(rq, rq->idle, SCHED_NORMAL, 0);
		rq->idle->sched_class = &idle_sched_class;
L
Linus Torvalds 已提交
6096
		migrate_dead_tasks(cpu);
6097
		raw_spin_unlock_irq(&rq->lock);
L
Linus Torvalds 已提交
6098 6099
		migrate_nr_uninterruptible(rq);
		BUG_ON(rq->nr_running != 0);
6100
		calc_global_load_remove(rq);
L
Linus Torvalds 已提交
6101
		break;
G
Gregory Haskins 已提交
6102

6103 6104
	case CPU_DYING:
	case CPU_DYING_FROZEN:
G
Gregory Haskins 已提交
6105
		/* Update our root-domain */
6106
		raw_spin_lock_irqsave(&rq->lock, flags);
G
Gregory Haskins 已提交
6107
		if (rq->rd) {
6108
			BUG_ON(!cpumask_test_cpu(cpu, rq->rd->span));
6109
			set_rq_offline(rq);
G
Gregory Haskins 已提交
6110
		}
6111
		raw_spin_unlock_irqrestore(&rq->lock, flags);
G
Gregory Haskins 已提交
6112
		break;
L
Linus Torvalds 已提交
6113 6114 6115 6116 6117
#endif
	}
	return NOTIFY_OK;
}

6118 6119 6120
/*
 * Register at high priority so that task migration (migrate_all_tasks)
 * happens before everything else.  This has to be lower priority than
6121
 * the notifier in the perf_event subsystem, though.
L
Linus Torvalds 已提交
6122
 */
6123
static struct notifier_block __cpuinitdata migration_notifier = {
L
Linus Torvalds 已提交
6124
	.notifier_call = migration_call,
6125
	.priority = CPU_PRI_MIGRATION,
L
Linus Torvalds 已提交
6126 6127
};

6128 6129 6130 6131 6132 6133 6134 6135 6136 6137 6138 6139 6140 6141 6142 6143 6144 6145 6146 6147 6148 6149 6150 6151 6152
static int __cpuinit sched_cpu_active(struct notifier_block *nfb,
				      unsigned long action, void *hcpu)
{
	switch (action & ~CPU_TASKS_FROZEN) {
	case CPU_ONLINE:
	case CPU_DOWN_FAILED:
		set_cpu_active((long)hcpu, true);
		return NOTIFY_OK;
	default:
		return NOTIFY_DONE;
	}
}

static int __cpuinit sched_cpu_inactive(struct notifier_block *nfb,
					unsigned long action, void *hcpu)
{
	switch (action & ~CPU_TASKS_FROZEN) {
	case CPU_DOWN_PREPARE:
		set_cpu_active((long)hcpu, false);
		return NOTIFY_OK;
	default:
		return NOTIFY_DONE;
	}
}

6153
static int __init migration_init(void)
L
Linus Torvalds 已提交
6154 6155
{
	void *cpu = (void *)(long)smp_processor_id();
6156
	int err;
6157

6158
	/* Initialize migration for the boot CPU */
6159 6160
	err = migration_call(&migration_notifier, CPU_UP_PREPARE, cpu);
	BUG_ON(err == NOTIFY_BAD);
L
Linus Torvalds 已提交
6161 6162
	migration_call(&migration_notifier, CPU_ONLINE, cpu);
	register_cpu_notifier(&migration_notifier);
6163

6164 6165 6166 6167
	/* Register cpu active notifiers */
	cpu_notifier(sched_cpu_active, CPU_PRI_SCHED_ACTIVE);
	cpu_notifier(sched_cpu_inactive, CPU_PRI_SCHED_INACTIVE);

6168
	return 0;
L
Linus Torvalds 已提交
6169
}
6170
early_initcall(migration_init);
L
Linus Torvalds 已提交
6171 6172 6173
#endif

#ifdef CONFIG_SMP
6174

6175
#ifdef CONFIG_SCHED_DEBUG
I
Ingo Molnar 已提交
6176

6177 6178 6179 6180 6181 6182 6183 6184 6185 6186
static __read_mostly int sched_domain_debug_enabled;

static int __init sched_domain_debug_setup(char *str)
{
	sched_domain_debug_enabled = 1;

	return 0;
}
early_param("sched_debug", sched_domain_debug_setup);

6187
static int sched_domain_debug_one(struct sched_domain *sd, int cpu, int level,
6188
				  struct cpumask *groupmask)
L
Linus Torvalds 已提交
6189
{
I
Ingo Molnar 已提交
6190
	struct sched_group *group = sd->groups;
6191
	char str[256];
L
Linus Torvalds 已提交
6192

R
Rusty Russell 已提交
6193
	cpulist_scnprintf(str, sizeof(str), sched_domain_span(sd));
6194
	cpumask_clear(groupmask);
I
Ingo Molnar 已提交
6195 6196 6197 6198

	printk(KERN_DEBUG "%*s domain %d: ", level, "", level);

	if (!(sd->flags & SD_LOAD_BALANCE)) {
P
Peter Zijlstra 已提交
6199
		printk("does not load-balance\n");
I
Ingo Molnar 已提交
6200
		if (sd->parent)
P
Peter Zijlstra 已提交
6201 6202
			printk(KERN_ERR "ERROR: !SD_LOAD_BALANCE domain"
					" has parent");
I
Ingo Molnar 已提交
6203
		return -1;
N
Nick Piggin 已提交
6204 6205
	}

P
Peter Zijlstra 已提交
6206
	printk(KERN_CONT "span %s level %s\n", str, sd->name);
I
Ingo Molnar 已提交
6207

6208
	if (!cpumask_test_cpu(cpu, sched_domain_span(sd))) {
P
Peter Zijlstra 已提交
6209 6210
		printk(KERN_ERR "ERROR: domain->span does not contain "
				"CPU%d\n", cpu);
I
Ingo Molnar 已提交
6211
	}
6212
	if (!cpumask_test_cpu(cpu, sched_group_cpus(group))) {
P
Peter Zijlstra 已提交
6213 6214
		printk(KERN_ERR "ERROR: domain->groups does not contain"
				" CPU%d\n", cpu);
I
Ingo Molnar 已提交
6215
	}
L
Linus Torvalds 已提交
6216

I
Ingo Molnar 已提交
6217
	printk(KERN_DEBUG "%*s groups:", level + 1, "");
L
Linus Torvalds 已提交
6218
	do {
I
Ingo Molnar 已提交
6219
		if (!group) {
P
Peter Zijlstra 已提交
6220 6221
			printk("\n");
			printk(KERN_ERR "ERROR: group is NULL\n");
L
Linus Torvalds 已提交
6222 6223 6224
			break;
		}

6225
		if (!group->cpu_power) {
P
Peter Zijlstra 已提交
6226 6227 6228
			printk(KERN_CONT "\n");
			printk(KERN_ERR "ERROR: domain->cpu_power not "
					"set\n");
I
Ingo Molnar 已提交
6229 6230
			break;
		}
L
Linus Torvalds 已提交
6231

6232
		if (!cpumask_weight(sched_group_cpus(group))) {
P
Peter Zijlstra 已提交
6233 6234
			printk(KERN_CONT "\n");
			printk(KERN_ERR "ERROR: empty group\n");
I
Ingo Molnar 已提交
6235 6236
			break;
		}
L
Linus Torvalds 已提交
6237

6238
		if (cpumask_intersects(groupmask, sched_group_cpus(group))) {
P
Peter Zijlstra 已提交
6239 6240
			printk(KERN_CONT "\n");
			printk(KERN_ERR "ERROR: repeated CPUs\n");
I
Ingo Molnar 已提交
6241 6242
			break;
		}
L
Linus Torvalds 已提交
6243

6244
		cpumask_or(groupmask, groupmask, sched_group_cpus(group));
L
Linus Torvalds 已提交
6245

R
Rusty Russell 已提交
6246
		cpulist_scnprintf(str, sizeof(str), sched_group_cpus(group));
6247

P
Peter Zijlstra 已提交
6248
		printk(KERN_CONT " %s", str);
6249
		if (group->cpu_power != SCHED_LOAD_SCALE) {
P
Peter Zijlstra 已提交
6250 6251
			printk(KERN_CONT " (cpu_power = %d)",
				group->cpu_power);
6252
		}
L
Linus Torvalds 已提交
6253

I
Ingo Molnar 已提交
6254 6255
		group = group->next;
	} while (group != sd->groups);
P
Peter Zijlstra 已提交
6256
	printk(KERN_CONT "\n");
L
Linus Torvalds 已提交
6257

6258
	if (!cpumask_equal(sched_domain_span(sd), groupmask))
P
Peter Zijlstra 已提交
6259
		printk(KERN_ERR "ERROR: groups don't span domain->span\n");
L
Linus Torvalds 已提交
6260

6261 6262
	if (sd->parent &&
	    !cpumask_subset(groupmask, sched_domain_span(sd->parent)))
P
Peter Zijlstra 已提交
6263 6264
		printk(KERN_ERR "ERROR: parent span is not a superset "
			"of domain->span\n");
I
Ingo Molnar 已提交
6265 6266
	return 0;
}
L
Linus Torvalds 已提交
6267

I
Ingo Molnar 已提交
6268 6269
static void sched_domain_debug(struct sched_domain *sd, int cpu)
{
6270
	cpumask_var_t groupmask;
I
Ingo Molnar 已提交
6271
	int level = 0;
L
Linus Torvalds 已提交
6272

6273 6274 6275
	if (!sched_domain_debug_enabled)
		return;

I
Ingo Molnar 已提交
6276 6277 6278 6279
	if (!sd) {
		printk(KERN_DEBUG "CPU%d attaching NULL sched-domain.\n", cpu);
		return;
	}
L
Linus Torvalds 已提交
6280

I
Ingo Molnar 已提交
6281 6282
	printk(KERN_DEBUG "CPU%d attaching sched-domain:\n", cpu);

6283
	if (!alloc_cpumask_var(&groupmask, GFP_KERNEL)) {
6284 6285 6286 6287
		printk(KERN_DEBUG "Cannot load-balance (out of memory)\n");
		return;
	}

I
Ingo Molnar 已提交
6288
	for (;;) {
6289
		if (sched_domain_debug_one(sd, cpu, level, groupmask))
I
Ingo Molnar 已提交
6290
			break;
L
Linus Torvalds 已提交
6291 6292
		level++;
		sd = sd->parent;
6293
		if (!sd)
I
Ingo Molnar 已提交
6294 6295
			break;
	}
6296
	free_cpumask_var(groupmask);
L
Linus Torvalds 已提交
6297
}
6298
#else /* !CONFIG_SCHED_DEBUG */
6299
# define sched_domain_debug(sd, cpu) do { } while (0)
6300
#endif /* CONFIG_SCHED_DEBUG */
L
Linus Torvalds 已提交
6301

6302
static int sd_degenerate(struct sched_domain *sd)
6303
{
6304
	if (cpumask_weight(sched_domain_span(sd)) == 1)
6305 6306 6307 6308 6309 6310
		return 1;

	/* Following flags need at least 2 groups */
	if (sd->flags & (SD_LOAD_BALANCE |
			 SD_BALANCE_NEWIDLE |
			 SD_BALANCE_FORK |
6311 6312 6313
			 SD_BALANCE_EXEC |
			 SD_SHARE_CPUPOWER |
			 SD_SHARE_PKG_RESOURCES)) {
6314 6315 6316 6317 6318
		if (sd->groups != sd->groups->next)
			return 0;
	}

	/* Following flags don't use groups */
6319
	if (sd->flags & (SD_WAKE_AFFINE))
6320 6321 6322 6323 6324
		return 0;

	return 1;
}

6325 6326
static int
sd_parent_degenerate(struct sched_domain *sd, struct sched_domain *parent)
6327 6328 6329 6330 6331 6332
{
	unsigned long cflags = sd->flags, pflags = parent->flags;

	if (sd_degenerate(parent))
		return 1;

6333
	if (!cpumask_equal(sched_domain_span(sd), sched_domain_span(parent)))
6334 6335 6336 6337 6338 6339 6340
		return 0;

	/* Flags needing groups don't count if only 1 group in parent */
	if (parent->groups == parent->groups->next) {
		pflags &= ~(SD_LOAD_BALANCE |
				SD_BALANCE_NEWIDLE |
				SD_BALANCE_FORK |
6341 6342 6343
				SD_BALANCE_EXEC |
				SD_SHARE_CPUPOWER |
				SD_SHARE_PKG_RESOURCES);
6344 6345
		if (nr_node_ids == 1)
			pflags &= ~SD_SERIALIZE;
6346 6347 6348 6349 6350 6351 6352
	}
	if (~cflags & pflags)
		return 0;

	return 1;
}

6353 6354
static void free_rootdomain(struct root_domain *rd)
{
6355 6356
	synchronize_sched();

6357 6358
	cpupri_cleanup(&rd->cpupri);

6359 6360 6361 6362 6363 6364
	free_cpumask_var(rd->rto_mask);
	free_cpumask_var(rd->online);
	free_cpumask_var(rd->span);
	kfree(rd);
}

G
Gregory Haskins 已提交
6365 6366
static void rq_attach_root(struct rq *rq, struct root_domain *rd)
{
I
Ingo Molnar 已提交
6367
	struct root_domain *old_rd = NULL;
G
Gregory Haskins 已提交
6368 6369
	unsigned long flags;

6370
	raw_spin_lock_irqsave(&rq->lock, flags);
G
Gregory Haskins 已提交
6371 6372

	if (rq->rd) {
I
Ingo Molnar 已提交
6373
		old_rd = rq->rd;
G
Gregory Haskins 已提交
6374

6375
		if (cpumask_test_cpu(rq->cpu, old_rd->online))
6376
			set_rq_offline(rq);
G
Gregory Haskins 已提交
6377

6378
		cpumask_clear_cpu(rq->cpu, old_rd->span);
6379

I
Ingo Molnar 已提交
6380 6381 6382 6383 6384 6385 6386
		/*
		 * If we dont want to free the old_rt yet then
		 * set old_rd to NULL to skip the freeing later
		 * in this function:
		 */
		if (!atomic_dec_and_test(&old_rd->refcount))
			old_rd = NULL;
G
Gregory Haskins 已提交
6387 6388 6389 6390 6391
	}

	atomic_inc(&rd->refcount);
	rq->rd = rd;

6392
	cpumask_set_cpu(rq->cpu, rd->span);
6393
	if (cpumask_test_cpu(rq->cpu, cpu_active_mask))
6394
		set_rq_online(rq);
G
Gregory Haskins 已提交
6395

6396
	raw_spin_unlock_irqrestore(&rq->lock, flags);
I
Ingo Molnar 已提交
6397 6398 6399

	if (old_rd)
		free_rootdomain(old_rd);
G
Gregory Haskins 已提交
6400 6401
}

6402
static int init_rootdomain(struct root_domain *rd)
G
Gregory Haskins 已提交
6403 6404 6405
{
	memset(rd, 0, sizeof(*rd));

6406
	if (!alloc_cpumask_var(&rd->span, GFP_KERNEL))
6407
		goto out;
6408
	if (!alloc_cpumask_var(&rd->online, GFP_KERNEL))
6409
		goto free_span;
6410
	if (!alloc_cpumask_var(&rd->rto_mask, GFP_KERNEL))
6411
		goto free_online;
6412

6413
	if (cpupri_init(&rd->cpupri) != 0)
6414
		goto free_rto_mask;
6415
	return 0;
6416

6417 6418
free_rto_mask:
	free_cpumask_var(rd->rto_mask);
6419 6420 6421 6422
free_online:
	free_cpumask_var(rd->online);
free_span:
	free_cpumask_var(rd->span);
6423
out:
6424
	return -ENOMEM;
G
Gregory Haskins 已提交
6425 6426 6427 6428
}

static void init_defrootdomain(void)
{
6429
	init_rootdomain(&def_root_domain);
6430

G
Gregory Haskins 已提交
6431 6432 6433
	atomic_set(&def_root_domain.refcount, 1);
}

6434
static struct root_domain *alloc_rootdomain(void)
G
Gregory Haskins 已提交
6435 6436 6437 6438 6439 6440 6441
{
	struct root_domain *rd;

	rd = kmalloc(sizeof(*rd), GFP_KERNEL);
	if (!rd)
		return NULL;

6442
	if (init_rootdomain(rd) != 0) {
6443 6444 6445
		kfree(rd);
		return NULL;
	}
G
Gregory Haskins 已提交
6446 6447 6448 6449

	return rd;
}

L
Linus Torvalds 已提交
6450
/*
I
Ingo Molnar 已提交
6451
 * Attach the domain 'sd' to 'cpu' as its base domain. Callers must
L
Linus Torvalds 已提交
6452 6453
 * hold the hotplug lock.
 */
I
Ingo Molnar 已提交
6454 6455
static void
cpu_attach_domain(struct sched_domain *sd, struct root_domain *rd, int cpu)
L
Linus Torvalds 已提交
6456
{
6457
	struct rq *rq = cpu_rq(cpu);
6458 6459
	struct sched_domain *tmp;

6460 6461 6462
	for (tmp = sd; tmp; tmp = tmp->parent)
		tmp->span_weight = cpumask_weight(sched_domain_span(tmp));

6463
	/* Remove the sched domains which do not contribute to scheduling. */
6464
	for (tmp = sd; tmp; ) {
6465 6466 6467
		struct sched_domain *parent = tmp->parent;
		if (!parent)
			break;
6468

6469
		if (sd_parent_degenerate(tmp, parent)) {
6470
			tmp->parent = parent->parent;
6471 6472
			if (parent->parent)
				parent->parent->child = tmp;
6473 6474
		} else
			tmp = tmp->parent;
6475 6476
	}

6477
	if (sd && sd_degenerate(sd)) {
6478
		sd = sd->parent;
6479 6480 6481
		if (sd)
			sd->child = NULL;
	}
L
Linus Torvalds 已提交
6482 6483 6484

	sched_domain_debug(sd, cpu);

G
Gregory Haskins 已提交
6485
	rq_attach_root(rq, rd);
N
Nick Piggin 已提交
6486
	rcu_assign_pointer(rq->sd, sd);
L
Linus Torvalds 已提交
6487 6488 6489
}

/* cpus with isolated domains */
6490
static cpumask_var_t cpu_isolated_map;
L
Linus Torvalds 已提交
6491 6492 6493 6494

/* Setup the mask of cpus configured for isolated domains */
static int __init isolated_cpu_setup(char *str)
{
R
Rusty Russell 已提交
6495
	alloc_bootmem_cpumask_var(&cpu_isolated_map);
R
Rusty Russell 已提交
6496
	cpulist_parse(str, cpu_isolated_map);
L
Linus Torvalds 已提交
6497 6498 6499
	return 1;
}

I
Ingo Molnar 已提交
6500
__setup("isolcpus=", isolated_cpu_setup);
L
Linus Torvalds 已提交
6501 6502

/*
6503 6504
 * init_sched_build_groups takes the cpumask we wish to span, and a pointer
 * to a function which identifies what group(along with sched group) a CPU
6505 6506
 * belongs to. The return value of group_fn must be a >= 0 and < nr_cpu_ids
 * (due to the fact that we keep track of groups covered with a struct cpumask).
L
Linus Torvalds 已提交
6507 6508 6509 6510 6511
 *
 * init_sched_build_groups will build a circular linked list of the groups
 * covered by the given span, and will set each group's ->cpumask correctly,
 * and ->cpu_power to 0.
 */
6512
static void
6513 6514 6515
init_sched_build_groups(const struct cpumask *span,
			const struct cpumask *cpu_map,
			int (*group_fn)(int cpu, const struct cpumask *cpu_map,
6516
					struct sched_group **sg,
6517 6518
					struct cpumask *tmpmask),
			struct cpumask *covered, struct cpumask *tmpmask)
L
Linus Torvalds 已提交
6519 6520 6521 6522
{
	struct sched_group *first = NULL, *last = NULL;
	int i;

6523
	cpumask_clear(covered);
6524

6525
	for_each_cpu(i, span) {
6526
		struct sched_group *sg;
6527
		int group = group_fn(i, cpu_map, &sg, tmpmask);
L
Linus Torvalds 已提交
6528 6529
		int j;

6530
		if (cpumask_test_cpu(i, covered))
L
Linus Torvalds 已提交
6531 6532
			continue;

6533
		cpumask_clear(sched_group_cpus(sg));
6534
		sg->cpu_power = 0;
L
Linus Torvalds 已提交
6535

6536
		for_each_cpu(j, span) {
6537
			if (group_fn(j, cpu_map, NULL, tmpmask) != group)
L
Linus Torvalds 已提交
6538 6539
				continue;

6540
			cpumask_set_cpu(j, covered);
6541
			cpumask_set_cpu(j, sched_group_cpus(sg));
L
Linus Torvalds 已提交
6542 6543 6544 6545 6546 6547 6548 6549 6550 6551
		}
		if (!first)
			first = sg;
		if (last)
			last->next = sg;
		last = sg;
	}
	last->next = first;
}

6552
#define SD_NODES_PER_DOMAIN 16
L
Linus Torvalds 已提交
6553

6554
#ifdef CONFIG_NUMA
6555

6556 6557 6558 6559 6560
/**
 * find_next_best_node - find the next node to include in a sched_domain
 * @node: node whose sched_domain we're building
 * @used_nodes: nodes already in the sched_domain
 *
I
Ingo Molnar 已提交
6561
 * Find the next node to include in a given scheduling domain. Simply
6562 6563 6564 6565
 * finds the closest node not already in the @used_nodes map.
 *
 * Should use nodemask_t.
 */
6566
static int find_next_best_node(int node, nodemask_t *used_nodes)
6567 6568 6569 6570 6571
{
	int i, n, val, min_val, best_node = 0;

	min_val = INT_MAX;

6572
	for (i = 0; i < nr_node_ids; i++) {
6573
		/* Start at @node */
6574
		n = (node + i) % nr_node_ids;
6575 6576 6577 6578 6579

		if (!nr_cpus_node(n))
			continue;

		/* Skip already used nodes */
6580
		if (node_isset(n, *used_nodes))
6581 6582 6583 6584 6585 6586 6587 6588 6589 6590 6591
			continue;

		/* Simple min distance search */
		val = node_distance(node, n);

		if (val < min_val) {
			min_val = val;
			best_node = n;
		}
	}

6592
	node_set(best_node, *used_nodes);
6593 6594 6595 6596 6597 6598
	return best_node;
}

/**
 * sched_domain_node_span - get a cpumask for a node's sched_domain
 * @node: node whose cpumask we're constructing
6599
 * @span: resulting cpumask
6600
 *
I
Ingo Molnar 已提交
6601
 * Given a node, construct a good cpumask for its sched_domain to span. It
6602 6603 6604
 * should be one that prevents unnecessary balancing, but also spreads tasks
 * out optimally.
 */
6605
static void sched_domain_node_span(int node, struct cpumask *span)
6606
{
6607
	nodemask_t used_nodes;
6608
	int i;
6609

6610
	cpumask_clear(span);
6611
	nodes_clear(used_nodes);
6612

6613
	cpumask_or(span, span, cpumask_of_node(node));
6614
	node_set(node, used_nodes);
6615 6616

	for (i = 1; i < SD_NODES_PER_DOMAIN; i++) {
6617
		int next_node = find_next_best_node(node, &used_nodes);
6618

6619
		cpumask_or(span, span, cpumask_of_node(next_node));
6620 6621
	}
}
6622
#endif /* CONFIG_NUMA */
6623

6624
int sched_smt_power_savings = 0, sched_mc_power_savings = 0;
6625

6626 6627
/*
 * The cpus mask in sched_group and sched_domain hangs off the end.
6628 6629 6630
 *
 * ( See the the comments in include/linux/sched.h:struct sched_group
 *   and struct sched_domain. )
6631 6632 6633 6634 6635 6636 6637 6638 6639 6640 6641
 */
struct static_sched_group {
	struct sched_group sg;
	DECLARE_BITMAP(cpus, CONFIG_NR_CPUS);
};

struct static_sched_domain {
	struct sched_domain sd;
	DECLARE_BITMAP(span, CONFIG_NR_CPUS);
};

6642 6643 6644 6645 6646 6647 6648 6649 6650 6651
struct s_data {
#ifdef CONFIG_NUMA
	int			sd_allnodes;
	cpumask_var_t		domainspan;
	cpumask_var_t		covered;
	cpumask_var_t		notcovered;
#endif
	cpumask_var_t		nodemask;
	cpumask_var_t		this_sibling_map;
	cpumask_var_t		this_core_map;
6652
	cpumask_var_t		this_book_map;
6653 6654 6655 6656 6657 6658
	cpumask_var_t		send_covered;
	cpumask_var_t		tmpmask;
	struct sched_group	**sched_group_nodes;
	struct root_domain	*rd;
};

6659 6660 6661 6662 6663
enum s_alloc {
	sa_sched_groups = 0,
	sa_rootdomain,
	sa_tmpmask,
	sa_send_covered,
6664
	sa_this_book_map,
6665 6666 6667 6668 6669 6670 6671 6672 6673 6674 6675 6676
	sa_this_core_map,
	sa_this_sibling_map,
	sa_nodemask,
	sa_sched_group_nodes,
#ifdef CONFIG_NUMA
	sa_notcovered,
	sa_covered,
	sa_domainspan,
#endif
	sa_none,
};

6677
/*
6678
 * SMT sched-domains:
6679
 */
L
Linus Torvalds 已提交
6680
#ifdef CONFIG_SCHED_SMT
6681
static DEFINE_PER_CPU(struct static_sched_domain, cpu_domains);
6682
static DEFINE_PER_CPU(struct static_sched_group, sched_groups);
6683

I
Ingo Molnar 已提交
6684
static int
6685 6686
cpu_to_cpu_group(int cpu, const struct cpumask *cpu_map,
		 struct sched_group **sg, struct cpumask *unused)
L
Linus Torvalds 已提交
6687
{
6688
	if (sg)
6689
		*sg = &per_cpu(sched_groups, cpu).sg;
L
Linus Torvalds 已提交
6690 6691
	return cpu;
}
6692
#endif /* CONFIG_SCHED_SMT */
L
Linus Torvalds 已提交
6693

6694 6695 6696
/*
 * multi-core sched-domains:
 */
6697
#ifdef CONFIG_SCHED_MC
6698 6699
static DEFINE_PER_CPU(struct static_sched_domain, core_domains);
static DEFINE_PER_CPU(struct static_sched_group, sched_group_core);
6700

I
Ingo Molnar 已提交
6701
static int
6702 6703
cpu_to_core_group(int cpu, const struct cpumask *cpu_map,
		  struct sched_group **sg, struct cpumask *mask)
6704
{
6705
	int group;
6706
#ifdef CONFIG_SCHED_SMT
6707
	cpumask_and(mask, topology_thread_cpumask(cpu), cpu_map);
6708
	group = cpumask_first(mask);
6709 6710 6711
#else
	group = cpu;
#endif
6712
	if (sg)
6713
		*sg = &per_cpu(sched_group_core, group).sg;
6714
	return group;
6715
}
6716
#endif /* CONFIG_SCHED_MC */
6717

6718 6719 6720 6721 6722 6723 6724 6725 6726 6727 6728 6729 6730 6731 6732 6733 6734 6735 6736 6737 6738 6739 6740 6741 6742
/*
 * book sched-domains:
 */
#ifdef CONFIG_SCHED_BOOK
static DEFINE_PER_CPU(struct static_sched_domain, book_domains);
static DEFINE_PER_CPU(struct static_sched_group, sched_group_book);

static int
cpu_to_book_group(int cpu, const struct cpumask *cpu_map,
		  struct sched_group **sg, struct cpumask *mask)
{
	int group = cpu;
#ifdef CONFIG_SCHED_MC
	cpumask_and(mask, cpu_coregroup_mask(cpu), cpu_map);
	group = cpumask_first(mask);
#elif defined(CONFIG_SCHED_SMT)
	cpumask_and(mask, topology_thread_cpumask(cpu), cpu_map);
	group = cpumask_first(mask);
#endif
	if (sg)
		*sg = &per_cpu(sched_group_book, group).sg;
	return group;
}
#endif /* CONFIG_SCHED_BOOK */

6743 6744
static DEFINE_PER_CPU(struct static_sched_domain, phys_domains);
static DEFINE_PER_CPU(struct static_sched_group, sched_group_phys);
6745

I
Ingo Molnar 已提交
6746
static int
6747 6748
cpu_to_phys_group(int cpu, const struct cpumask *cpu_map,
		  struct sched_group **sg, struct cpumask *mask)
L
Linus Torvalds 已提交
6749
{
6750
	int group;
6751 6752 6753 6754
#ifdef CONFIG_SCHED_BOOK
	cpumask_and(mask, cpu_book_mask(cpu), cpu_map);
	group = cpumask_first(mask);
#elif defined(CONFIG_SCHED_MC)
6755
	cpumask_and(mask, cpu_coregroup_mask(cpu), cpu_map);
6756
	group = cpumask_first(mask);
6757
#elif defined(CONFIG_SCHED_SMT)
6758
	cpumask_and(mask, topology_thread_cpumask(cpu), cpu_map);
6759
	group = cpumask_first(mask);
L
Linus Torvalds 已提交
6760
#else
6761
	group = cpu;
L
Linus Torvalds 已提交
6762
#endif
6763
	if (sg)
6764
		*sg = &per_cpu(sched_group_phys, group).sg;
6765
	return group;
L
Linus Torvalds 已提交
6766 6767 6768 6769
}

#ifdef CONFIG_NUMA
/*
6770 6771 6772
 * The init_sched_build_groups can't handle what we want to do with node
 * groups, so roll our own. Now each node has its own list of groups which
 * gets dynamically allocated.
L
Linus Torvalds 已提交
6773
 */
6774
static DEFINE_PER_CPU(struct static_sched_domain, node_domains);
6775
static struct sched_group ***sched_group_nodes_bycpu;
L
Linus Torvalds 已提交
6776

6777
static DEFINE_PER_CPU(struct static_sched_domain, allnodes_domains);
6778
static DEFINE_PER_CPU(struct static_sched_group, sched_group_allnodes);
6779

6780 6781 6782
static int cpu_to_allnodes_group(int cpu, const struct cpumask *cpu_map,
				 struct sched_group **sg,
				 struct cpumask *nodemask)
6783
{
6784 6785
	int group;

6786
	cpumask_and(nodemask, cpumask_of_node(cpu_to_node(cpu)), cpu_map);
6787
	group = cpumask_first(nodemask);
6788 6789

	if (sg)
6790
		*sg = &per_cpu(sched_group_allnodes, group).sg;
6791
	return group;
L
Linus Torvalds 已提交
6792
}
6793

6794 6795 6796 6797 6798 6799 6800
static void init_numa_sched_groups_power(struct sched_group *group_head)
{
	struct sched_group *sg = group_head;
	int j;

	if (!sg)
		return;
6801
	do {
6802
		for_each_cpu(j, sched_group_cpus(sg)) {
6803
			struct sched_domain *sd;
6804

6805
			sd = &per_cpu(phys_domains, j).sd;
6806
			if (j != group_first_cpu(sd->groups)) {
6807 6808 6809 6810 6811 6812
				/*
				 * Only add "power" once for each
				 * physical package.
				 */
				continue;
			}
6813

6814
			sg->cpu_power += sd->groups->cpu_power;
6815 6816 6817
		}
		sg = sg->next;
	} while (sg != group_head);
6818
}
6819 6820 6821 6822 6823 6824 6825 6826 6827 6828 6829 6830 6831 6832 6833 6834 6835 6836 6837 6838 6839

static int build_numa_sched_groups(struct s_data *d,
				   const struct cpumask *cpu_map, int num)
{
	struct sched_domain *sd;
	struct sched_group *sg, *prev;
	int n, j;

	cpumask_clear(d->covered);
	cpumask_and(d->nodemask, cpumask_of_node(num), cpu_map);
	if (cpumask_empty(d->nodemask)) {
		d->sched_group_nodes[num] = NULL;
		goto out;
	}

	sched_domain_node_span(num, d->domainspan);
	cpumask_and(d->domainspan, d->domainspan, cpu_map);

	sg = kmalloc_node(sizeof(struct sched_group) + cpumask_size(),
			  GFP_KERNEL, num);
	if (!sg) {
P
Peter Zijlstra 已提交
6840 6841
		printk(KERN_WARNING "Can not alloc domain group for node %d\n",
		       num);
6842 6843 6844 6845 6846 6847 6848 6849 6850
		return -ENOMEM;
	}
	d->sched_group_nodes[num] = sg;

	for_each_cpu(j, d->nodemask) {
		sd = &per_cpu(node_domains, j).sd;
		sd->groups = sg;
	}

6851
	sg->cpu_power = 0;
6852 6853 6854 6855 6856 6857 6858 6859 6860 6861 6862 6863 6864 6865 6866 6867 6868 6869
	cpumask_copy(sched_group_cpus(sg), d->nodemask);
	sg->next = sg;
	cpumask_or(d->covered, d->covered, d->nodemask);

	prev = sg;
	for (j = 0; j < nr_node_ids; j++) {
		n = (num + j) % nr_node_ids;
		cpumask_complement(d->notcovered, d->covered);
		cpumask_and(d->tmpmask, d->notcovered, cpu_map);
		cpumask_and(d->tmpmask, d->tmpmask, d->domainspan);
		if (cpumask_empty(d->tmpmask))
			break;
		cpumask_and(d->tmpmask, d->tmpmask, cpumask_of_node(n));
		if (cpumask_empty(d->tmpmask))
			continue;
		sg = kmalloc_node(sizeof(struct sched_group) + cpumask_size(),
				  GFP_KERNEL, num);
		if (!sg) {
P
Peter Zijlstra 已提交
6870 6871
			printk(KERN_WARNING
			       "Can not alloc domain group for node %d\n", j);
6872 6873
			return -ENOMEM;
		}
6874
		sg->cpu_power = 0;
6875 6876 6877 6878 6879 6880 6881 6882 6883
		cpumask_copy(sched_group_cpus(sg), d->tmpmask);
		sg->next = prev->next;
		cpumask_or(d->covered, d->covered, d->tmpmask);
		prev->next = sg;
		prev = sg;
	}
out:
	return 0;
}
6884
#endif /* CONFIG_NUMA */
L
Linus Torvalds 已提交
6885

6886
#ifdef CONFIG_NUMA
6887
/* Free memory allocated for various sched_group structures */
6888 6889
static void free_sched_groups(const struct cpumask *cpu_map,
			      struct cpumask *nodemask)
6890
{
6891
	int cpu, i;
6892

6893
	for_each_cpu(cpu, cpu_map) {
6894 6895 6896 6897 6898 6899
		struct sched_group **sched_group_nodes
			= sched_group_nodes_bycpu[cpu];

		if (!sched_group_nodes)
			continue;

6900
		for (i = 0; i < nr_node_ids; i++) {
6901 6902
			struct sched_group *oldsg, *sg = sched_group_nodes[i];

6903
			cpumask_and(nodemask, cpumask_of_node(i), cpu_map);
6904
			if (cpumask_empty(nodemask))
6905 6906 6907 6908 6909 6910 6911 6912 6913 6914 6915 6916 6917 6918 6919 6920
				continue;

			if (sg == NULL)
				continue;
			sg = sg->next;
next_sg:
			oldsg = sg;
			sg = sg->next;
			kfree(oldsg);
			if (oldsg != sched_group_nodes[i])
				goto next_sg;
		}
		kfree(sched_group_nodes);
		sched_group_nodes_bycpu[cpu] = NULL;
	}
}
6921
#else /* !CONFIG_NUMA */
6922 6923
static void free_sched_groups(const struct cpumask *cpu_map,
			      struct cpumask *nodemask)
6924 6925
{
}
6926
#endif /* CONFIG_NUMA */
6927

6928 6929 6930 6931 6932 6933 6934 6935 6936 6937 6938 6939 6940 6941
/*
 * Initialize sched groups cpu_power.
 *
 * cpu_power indicates the capacity of sched group, which is used while
 * distributing the load between different sched groups in a sched domain.
 * Typically cpu_power for all the groups in a sched domain will be same unless
 * there are asymmetries in the topology. If there are asymmetries, group
 * having more cpu_power will pickup more load compared to the group having
 * less cpu_power.
 */
static void init_sched_groups_power(int cpu, struct sched_domain *sd)
{
	struct sched_domain *child;
	struct sched_group *group;
6942 6943
	long power;
	int weight;
6944 6945 6946

	WARN_ON(!sd || !sd->groups);

6947
	if (cpu != group_first_cpu(sd->groups))
6948 6949 6950 6951
		return;

	child = sd->child;

6952
	sd->groups->cpu_power = 0;
6953

6954 6955 6956 6957 6958
	if (!child) {
		power = SCHED_LOAD_SCALE;
		weight = cpumask_weight(sched_domain_span(sd));
		/*
		 * SMT siblings share the power of a single core.
P
Peter Zijlstra 已提交
6959 6960 6961
		 * Usually multiple threads get a better yield out of
		 * that one core than a single thread would have,
		 * reflect that in sd->smt_gain.
6962
		 */
P
Peter Zijlstra 已提交
6963 6964
		if ((sd->flags & SD_SHARE_CPUPOWER) && weight > 1) {
			power *= sd->smt_gain;
6965
			power /= weight;
P
Peter Zijlstra 已提交
6966 6967
			power >>= SCHED_LOAD_SHIFT;
		}
6968
		sd->groups->cpu_power += power;
6969 6970 6971 6972
		return;
	}

	/*
6973
	 * Add cpu_power of each child group to this groups cpu_power.
6974 6975 6976
	 */
	group = child->groups;
	do {
6977
		sd->groups->cpu_power += group->cpu_power;
6978 6979 6980 6981
		group = group->next;
	} while (group != child->groups);
}

6982 6983 6984 6985 6986
/*
 * Initializers for schedule domains
 * Non-inlined to reduce accumulated stack pressure in build_sched_domains()
 */

6987 6988 6989 6990 6991 6992
#ifdef CONFIG_SCHED_DEBUG
# define SD_INIT_NAME(sd, type)		sd->name = #type
#else
# define SD_INIT_NAME(sd, type)		do { } while (0)
#endif

6993
#define	SD_INIT(sd, type)	sd_init_##type(sd)
6994

6995 6996 6997 6998 6999
#define SD_INIT_FUNC(type)	\
static noinline void sd_init_##type(struct sched_domain *sd)	\
{								\
	memset(sd, 0, sizeof(*sd));				\
	*sd = SD_##type##_INIT;					\
7000
	sd->level = SD_LV_##type;				\
7001
	SD_INIT_NAME(sd, type);					\
7002 7003 7004 7005 7006 7007 7008 7009 7010 7011 7012 7013 7014
}

SD_INIT_FUNC(CPU)
#ifdef CONFIG_NUMA
 SD_INIT_FUNC(ALLNODES)
 SD_INIT_FUNC(NODE)
#endif
#ifdef CONFIG_SCHED_SMT
 SD_INIT_FUNC(SIBLING)
#endif
#ifdef CONFIG_SCHED_MC
 SD_INIT_FUNC(MC)
#endif
7015 7016 7017
#ifdef CONFIG_SCHED_BOOK
 SD_INIT_FUNC(BOOK)
#endif
7018

7019 7020 7021 7022
static int default_relax_domain_level = -1;

static int __init setup_relax_domain_level(char *str)
{
7023 7024 7025 7026 7027 7028
	unsigned long val;

	val = simple_strtoul(str, NULL, 0);
	if (val < SD_LV_MAX)
		default_relax_domain_level = val;

7029 7030 7031 7032 7033 7034 7035 7036 7037 7038 7039 7040 7041 7042 7043 7044 7045 7046
	return 1;
}
__setup("relax_domain_level=", setup_relax_domain_level);

static void set_domain_attribute(struct sched_domain *sd,
				 struct sched_domain_attr *attr)
{
	int request;

	if (!attr || attr->relax_domain_level < 0) {
		if (default_relax_domain_level < 0)
			return;
		else
			request = default_relax_domain_level;
	} else
		request = attr->relax_domain_level;
	if (request < sd->level) {
		/* turn off idle balance on this domain */
7047
		sd->flags &= ~(SD_BALANCE_WAKE|SD_BALANCE_NEWIDLE);
7048 7049
	} else {
		/* turn on idle balance on this domain */
7050
		sd->flags |= (SD_BALANCE_WAKE|SD_BALANCE_NEWIDLE);
7051 7052 7053
	}
}

7054 7055 7056 7057 7058 7059 7060 7061 7062 7063 7064 7065 7066
static void __free_domain_allocs(struct s_data *d, enum s_alloc what,
				 const struct cpumask *cpu_map)
{
	switch (what) {
	case sa_sched_groups:
		free_sched_groups(cpu_map, d->tmpmask); /* fall through */
		d->sched_group_nodes = NULL;
	case sa_rootdomain:
		free_rootdomain(d->rd); /* fall through */
	case sa_tmpmask:
		free_cpumask_var(d->tmpmask); /* fall through */
	case sa_send_covered:
		free_cpumask_var(d->send_covered); /* fall through */
7067 7068
	case sa_this_book_map:
		free_cpumask_var(d->this_book_map); /* fall through */
7069 7070 7071 7072 7073 7074 7075
	case sa_this_core_map:
		free_cpumask_var(d->this_core_map); /* fall through */
	case sa_this_sibling_map:
		free_cpumask_var(d->this_sibling_map); /* fall through */
	case sa_nodemask:
		free_cpumask_var(d->nodemask); /* fall through */
	case sa_sched_group_nodes:
7076
#ifdef CONFIG_NUMA
7077 7078 7079 7080 7081 7082 7083
		kfree(d->sched_group_nodes); /* fall through */
	case sa_notcovered:
		free_cpumask_var(d->notcovered); /* fall through */
	case sa_covered:
		free_cpumask_var(d->covered); /* fall through */
	case sa_domainspan:
		free_cpumask_var(d->domainspan); /* fall through */
7084
#endif
7085 7086 7087 7088
	case sa_none:
		break;
	}
}
7089

7090 7091 7092
static enum s_alloc __visit_domain_allocation_hell(struct s_data *d,
						   const struct cpumask *cpu_map)
{
7093
#ifdef CONFIG_NUMA
7094 7095 7096 7097 7098 7099 7100 7101 7102 7103
	if (!alloc_cpumask_var(&d->domainspan, GFP_KERNEL))
		return sa_none;
	if (!alloc_cpumask_var(&d->covered, GFP_KERNEL))
		return sa_domainspan;
	if (!alloc_cpumask_var(&d->notcovered, GFP_KERNEL))
		return sa_covered;
	/* Allocate the per-node list of sched groups */
	d->sched_group_nodes = kcalloc(nr_node_ids,
				      sizeof(struct sched_group *), GFP_KERNEL);
	if (!d->sched_group_nodes) {
P
Peter Zijlstra 已提交
7104
		printk(KERN_WARNING "Can not alloc sched group node list\n");
7105
		return sa_notcovered;
7106
	}
7107
	sched_group_nodes_bycpu[cpumask_first(cpu_map)] = d->sched_group_nodes;
7108
#endif
7109 7110 7111 7112 7113 7114
	if (!alloc_cpumask_var(&d->nodemask, GFP_KERNEL))
		return sa_sched_group_nodes;
	if (!alloc_cpumask_var(&d->this_sibling_map, GFP_KERNEL))
		return sa_nodemask;
	if (!alloc_cpumask_var(&d->this_core_map, GFP_KERNEL))
		return sa_this_sibling_map;
7115
	if (!alloc_cpumask_var(&d->this_book_map, GFP_KERNEL))
7116
		return sa_this_core_map;
7117 7118
	if (!alloc_cpumask_var(&d->send_covered, GFP_KERNEL))
		return sa_this_book_map;
7119 7120 7121 7122
	if (!alloc_cpumask_var(&d->tmpmask, GFP_KERNEL))
		return sa_send_covered;
	d->rd = alloc_rootdomain();
	if (!d->rd) {
P
Peter Zijlstra 已提交
7123
		printk(KERN_WARNING "Cannot alloc root domain\n");
7124
		return sa_tmpmask;
G
Gregory Haskins 已提交
7125
	}
7126 7127
	return sa_rootdomain;
}
G
Gregory Haskins 已提交
7128

7129 7130 7131 7132
static struct sched_domain *__build_numa_sched_domains(struct s_data *d,
	const struct cpumask *cpu_map, struct sched_domain_attr *attr, int i)
{
	struct sched_domain *sd = NULL;
7133
#ifdef CONFIG_NUMA
7134
	struct sched_domain *parent;
L
Linus Torvalds 已提交
7135

7136 7137 7138 7139 7140
	d->sd_allnodes = 0;
	if (cpumask_weight(cpu_map) >
	    SD_NODES_PER_DOMAIN * cpumask_weight(d->nodemask)) {
		sd = &per_cpu(allnodes_domains, i).sd;
		SD_INIT(sd, ALLNODES);
7141
		set_domain_attribute(sd, attr);
7142 7143 7144 7145 7146 7147 7148 7149 7150 7151 7152 7153 7154 7155
		cpumask_copy(sched_domain_span(sd), cpu_map);
		cpu_to_allnodes_group(i, cpu_map, &sd->groups, d->tmpmask);
		d->sd_allnodes = 1;
	}
	parent = sd;

	sd = &per_cpu(node_domains, i).sd;
	SD_INIT(sd, NODE);
	set_domain_attribute(sd, attr);
	sched_domain_node_span(cpu_to_node(i), sched_domain_span(sd));
	sd->parent = parent;
	if (parent)
		parent->child = sd;
	cpumask_and(sched_domain_span(sd), sched_domain_span(sd), cpu_map);
L
Linus Torvalds 已提交
7156
#endif
7157 7158
	return sd;
}
L
Linus Torvalds 已提交
7159

7160 7161 7162 7163 7164 7165 7166 7167 7168 7169 7170 7171 7172 7173 7174
static struct sched_domain *__build_cpu_sched_domain(struct s_data *d,
	const struct cpumask *cpu_map, struct sched_domain_attr *attr,
	struct sched_domain *parent, int i)
{
	struct sched_domain *sd;
	sd = &per_cpu(phys_domains, i).sd;
	SD_INIT(sd, CPU);
	set_domain_attribute(sd, attr);
	cpumask_copy(sched_domain_span(sd), d->nodemask);
	sd->parent = parent;
	if (parent)
		parent->child = sd;
	cpu_to_phys_group(i, cpu_map, &sd->groups, d->tmpmask);
	return sd;
}
L
Linus Torvalds 已提交
7175

7176 7177 7178 7179 7180 7181 7182 7183 7184 7185 7186 7187 7188 7189 7190 7191 7192
static struct sched_domain *__build_book_sched_domain(struct s_data *d,
	const struct cpumask *cpu_map, struct sched_domain_attr *attr,
	struct sched_domain *parent, int i)
{
	struct sched_domain *sd = parent;
#ifdef CONFIG_SCHED_BOOK
	sd = &per_cpu(book_domains, i).sd;
	SD_INIT(sd, BOOK);
	set_domain_attribute(sd, attr);
	cpumask_and(sched_domain_span(sd), cpu_map, cpu_book_mask(i));
	sd->parent = parent;
	parent->child = sd;
	cpu_to_book_group(i, cpu_map, &sd->groups, d->tmpmask);
#endif
	return sd;
}

7193 7194 7195 7196 7197
static struct sched_domain *__build_mc_sched_domain(struct s_data *d,
	const struct cpumask *cpu_map, struct sched_domain_attr *attr,
	struct sched_domain *parent, int i)
{
	struct sched_domain *sd = parent;
7198
#ifdef CONFIG_SCHED_MC
7199 7200 7201 7202 7203 7204 7205
	sd = &per_cpu(core_domains, i).sd;
	SD_INIT(sd, MC);
	set_domain_attribute(sd, attr);
	cpumask_and(sched_domain_span(sd), cpu_map, cpu_coregroup_mask(i));
	sd->parent = parent;
	parent->child = sd;
	cpu_to_core_group(i, cpu_map, &sd->groups, d->tmpmask);
7206
#endif
7207 7208
	return sd;
}
7209

7210 7211 7212 7213 7214
static struct sched_domain *__build_smt_sched_domain(struct s_data *d,
	const struct cpumask *cpu_map, struct sched_domain_attr *attr,
	struct sched_domain *parent, int i)
{
	struct sched_domain *sd = parent;
L
Linus Torvalds 已提交
7215
#ifdef CONFIG_SCHED_SMT
7216 7217 7218 7219 7220 7221 7222
	sd = &per_cpu(cpu_domains, i).sd;
	SD_INIT(sd, SIBLING);
	set_domain_attribute(sd, attr);
	cpumask_and(sched_domain_span(sd), cpu_map, topology_thread_cpumask(i));
	sd->parent = parent;
	parent->child = sd;
	cpu_to_cpu_group(i, cpu_map, &sd->groups, d->tmpmask);
L
Linus Torvalds 已提交
7223
#endif
7224 7225
	return sd;
}
L
Linus Torvalds 已提交
7226

7227 7228 7229 7230
static void build_sched_groups(struct s_data *d, enum sched_domain_level l,
			       const struct cpumask *cpu_map, int cpu)
{
	switch (l) {
L
Linus Torvalds 已提交
7231
#ifdef CONFIG_SCHED_SMT
7232 7233 7234 7235 7236 7237 7238 7239
	case SD_LV_SIBLING: /* set up CPU (sibling) groups */
		cpumask_and(d->this_sibling_map, cpu_map,
			    topology_thread_cpumask(cpu));
		if (cpu == cpumask_first(d->this_sibling_map))
			init_sched_build_groups(d->this_sibling_map, cpu_map,
						&cpu_to_cpu_group,
						d->send_covered, d->tmpmask);
		break;
L
Linus Torvalds 已提交
7240
#endif
7241
#ifdef CONFIG_SCHED_MC
7242 7243 7244 7245 7246 7247 7248
	case SD_LV_MC: /* set up multi-core groups */
		cpumask_and(d->this_core_map, cpu_map, cpu_coregroup_mask(cpu));
		if (cpu == cpumask_first(d->this_core_map))
			init_sched_build_groups(d->this_core_map, cpu_map,
						&cpu_to_core_group,
						d->send_covered, d->tmpmask);
		break;
7249 7250 7251 7252 7253 7254 7255 7256 7257
#endif
#ifdef CONFIG_SCHED_BOOK
	case SD_LV_BOOK: /* set up book groups */
		cpumask_and(d->this_book_map, cpu_map, cpu_book_mask(cpu));
		if (cpu == cpumask_first(d->this_book_map))
			init_sched_build_groups(d->this_book_map, cpu_map,
						&cpu_to_book_group,
						d->send_covered, d->tmpmask);
		break;
7258
#endif
7259 7260 7261 7262 7263 7264 7265
	case SD_LV_CPU: /* set up physical groups */
		cpumask_and(d->nodemask, cpumask_of_node(cpu), cpu_map);
		if (!cpumask_empty(d->nodemask))
			init_sched_build_groups(d->nodemask, cpu_map,
						&cpu_to_phys_group,
						d->send_covered, d->tmpmask);
		break;
L
Linus Torvalds 已提交
7266
#ifdef CONFIG_NUMA
7267 7268 7269 7270 7271
	case SD_LV_ALLNODES:
		init_sched_build_groups(cpu_map, cpu_map, &cpu_to_allnodes_group,
					d->send_covered, d->tmpmask);
		break;
#endif
7272 7273
	default:
		break;
7274
	}
7275
}
7276

7277 7278 7279 7280 7281 7282 7283 7284 7285
/*
 * Build sched domains for a given set of cpus and attach the sched domains
 * to the individual cpus
 */
static int __build_sched_domains(const struct cpumask *cpu_map,
				 struct sched_domain_attr *attr)
{
	enum s_alloc alloc_state = sa_none;
	struct s_data d;
7286
	struct sched_domain *sd;
7287
	int i;
7288
#ifdef CONFIG_NUMA
7289
	d.sd_allnodes = 0;
7290
#endif
7291

7292 7293 7294 7295
	alloc_state = __visit_domain_allocation_hell(&d, cpu_map);
	if (alloc_state != sa_rootdomain)
		goto error;
	alloc_state = sa_sched_groups;
7296

L
Linus Torvalds 已提交
7297
	/*
7298
	 * Set up domains for cpus specified by the cpu_map.
L
Linus Torvalds 已提交
7299
	 */
7300
	for_each_cpu(i, cpu_map) {
7301 7302
		cpumask_and(d.nodemask, cpumask_of_node(cpu_to_node(i)),
			    cpu_map);
I
Ingo Molnar 已提交
7303

7304
		sd = __build_numa_sched_domains(&d, cpu_map, attr, i);
7305
		sd = __build_cpu_sched_domain(&d, cpu_map, attr, sd, i);
7306
		sd = __build_book_sched_domain(&d, cpu_map, attr, sd, i);
7307
		sd = __build_mc_sched_domain(&d, cpu_map, attr, sd, i);
7308
		sd = __build_smt_sched_domain(&d, cpu_map, attr, sd, i);
L
Linus Torvalds 已提交
7309
	}
7310

7311
	for_each_cpu(i, cpu_map) {
7312
		build_sched_groups(&d, SD_LV_SIBLING, cpu_map, i);
7313
		build_sched_groups(&d, SD_LV_BOOK, cpu_map, i);
7314
		build_sched_groups(&d, SD_LV_MC, cpu_map, i);
L
Linus Torvalds 已提交
7315
	}
7316

L
Linus Torvalds 已提交
7317
	/* Set up physical groups */
7318 7319
	for (i = 0; i < nr_node_ids; i++)
		build_sched_groups(&d, SD_LV_CPU, cpu_map, i);
7320

L
Linus Torvalds 已提交
7321 7322
#ifdef CONFIG_NUMA
	/* Set up node groups */
7323 7324
	if (d.sd_allnodes)
		build_sched_groups(&d, SD_LV_ALLNODES, cpu_map, 0);
7325

7326 7327
	for (i = 0; i < nr_node_ids; i++)
		if (build_numa_sched_groups(&d, cpu_map, i))
7328
			goto error;
L
Linus Torvalds 已提交
7329 7330 7331
#endif

	/* Calculate CPU power for physical packages and nodes */
7332
#ifdef CONFIG_SCHED_SMT
7333
	for_each_cpu(i, cpu_map) {
7334
		sd = &per_cpu(cpu_domains, i).sd;
7335
		init_sched_groups_power(i, sd);
7336
	}
L
Linus Torvalds 已提交
7337
#endif
7338
#ifdef CONFIG_SCHED_MC
7339
	for_each_cpu(i, cpu_map) {
7340
		sd = &per_cpu(core_domains, i).sd;
7341
		init_sched_groups_power(i, sd);
7342 7343
	}
#endif
7344 7345 7346 7347 7348 7349
#ifdef CONFIG_SCHED_BOOK
	for_each_cpu(i, cpu_map) {
		sd = &per_cpu(book_domains, i).sd;
		init_sched_groups_power(i, sd);
	}
#endif
7350

7351
	for_each_cpu(i, cpu_map) {
7352
		sd = &per_cpu(phys_domains, i).sd;
7353
		init_sched_groups_power(i, sd);
L
Linus Torvalds 已提交
7354 7355
	}

7356
#ifdef CONFIG_NUMA
7357
	for (i = 0; i < nr_node_ids; i++)
7358
		init_numa_sched_groups_power(d.sched_group_nodes[i]);
7359

7360
	if (d.sd_allnodes) {
7361
		struct sched_group *sg;
7362

7363
		cpu_to_allnodes_group(cpumask_first(cpu_map), cpu_map, &sg,
7364
								d.tmpmask);
7365 7366
		init_numa_sched_groups_power(sg);
	}
7367 7368
#endif

L
Linus Torvalds 已提交
7369
	/* Attach the domains */
7370
	for_each_cpu(i, cpu_map) {
L
Linus Torvalds 已提交
7371
#ifdef CONFIG_SCHED_SMT
7372
		sd = &per_cpu(cpu_domains, i).sd;
7373
#elif defined(CONFIG_SCHED_MC)
7374
		sd = &per_cpu(core_domains, i).sd;
7375 7376
#elif defined(CONFIG_SCHED_BOOK)
		sd = &per_cpu(book_domains, i).sd;
L
Linus Torvalds 已提交
7377
#else
7378
		sd = &per_cpu(phys_domains, i).sd;
L
Linus Torvalds 已提交
7379
#endif
7380
		cpu_attach_domain(sd, d.rd, i);
L
Linus Torvalds 已提交
7381
	}
7382

7383 7384 7385
	d.sched_group_nodes = NULL; /* don't free this we still need it */
	__free_domain_allocs(&d, sa_tmpmask, cpu_map);
	return 0;
7386 7387

error:
7388 7389
	__free_domain_allocs(&d, alloc_state, cpu_map);
	return -ENOMEM;
L
Linus Torvalds 已提交
7390
}
P
Paul Jackson 已提交
7391

7392
static int build_sched_domains(const struct cpumask *cpu_map)
7393 7394 7395 7396
{
	return __build_sched_domains(cpu_map, NULL);
}

7397
static cpumask_var_t *doms_cur;	/* current sched domains */
P
Paul Jackson 已提交
7398
static int ndoms_cur;		/* number of sched domains in 'doms_cur' */
I
Ingo Molnar 已提交
7399 7400
static struct sched_domain_attr *dattr_cur;
				/* attribues of custom domains in 'doms_cur' */
P
Paul Jackson 已提交
7401 7402 7403

/*
 * Special case: If a kmalloc of a doms_cur partition (array of
7404 7405
 * cpumask) fails, then fallback to a single sched domain,
 * as determined by the single cpumask fallback_doms.
P
Paul Jackson 已提交
7406
 */
7407
static cpumask_var_t fallback_doms;
P
Paul Jackson 已提交
7408

7409 7410 7411 7412 7413 7414
/*
 * arch_update_cpu_topology lets virtualized architectures update the
 * cpu core maps. It is supposed to return 1 if the topology changed
 * or 0 if it stayed the same.
 */
int __attribute__((weak)) arch_update_cpu_topology(void)
7415
{
7416
	return 0;
7417 7418
}

7419 7420 7421 7422 7423 7424 7425 7426 7427 7428 7429 7430 7431 7432 7433 7434 7435 7436 7437 7438 7439 7440 7441 7442 7443
cpumask_var_t *alloc_sched_domains(unsigned int ndoms)
{
	int i;
	cpumask_var_t *doms;

	doms = kmalloc(sizeof(*doms) * ndoms, GFP_KERNEL);
	if (!doms)
		return NULL;
	for (i = 0; i < ndoms; i++) {
		if (!alloc_cpumask_var(&doms[i], GFP_KERNEL)) {
			free_sched_domains(doms, i);
			return NULL;
		}
	}
	return doms;
}

void free_sched_domains(cpumask_var_t doms[], unsigned int ndoms)
{
	unsigned int i;
	for (i = 0; i < ndoms; i++)
		free_cpumask_var(doms[i]);
	kfree(doms);
}

7444
/*
I
Ingo Molnar 已提交
7445
 * Set up scheduler domains and groups. Callers must hold the hotplug lock.
P
Paul Jackson 已提交
7446 7447
 * For now this just excludes isolated cpus, but could be used to
 * exclude other special cases in the future.
7448
 */
7449
static int arch_init_sched_domains(const struct cpumask *cpu_map)
7450
{
7451 7452
	int err;

7453
	arch_update_cpu_topology();
P
Paul Jackson 已提交
7454
	ndoms_cur = 1;
7455
	doms_cur = alloc_sched_domains(ndoms_cur);
P
Paul Jackson 已提交
7456
	if (!doms_cur)
7457 7458
		doms_cur = &fallback_doms;
	cpumask_andnot(doms_cur[0], cpu_map, cpu_isolated_map);
7459
	dattr_cur = NULL;
7460
	err = build_sched_domains(doms_cur[0]);
7461
	register_sched_domain_sysctl();
7462 7463

	return err;
7464 7465
}

7466 7467
static void arch_destroy_sched_domains(const struct cpumask *cpu_map,
				       struct cpumask *tmpmask)
L
Linus Torvalds 已提交
7468
{
7469
	free_sched_groups(cpu_map, tmpmask);
7470
}
L
Linus Torvalds 已提交
7471

7472 7473 7474 7475
/*
 * Detach sched domains from a group of cpus specified in cpu_map
 * These cpus will now be attached to the NULL domain
 */
7476
static void detach_destroy_domains(const struct cpumask *cpu_map)
7477
{
7478 7479
	/* Save because hotplug lock held. */
	static DECLARE_BITMAP(tmpmask, CONFIG_NR_CPUS);
7480 7481
	int i;

7482
	for_each_cpu(i, cpu_map)
G
Gregory Haskins 已提交
7483
		cpu_attach_domain(NULL, &def_root_domain, i);
7484
	synchronize_sched();
7485
	arch_destroy_sched_domains(cpu_map, to_cpumask(tmpmask));
7486 7487
}

7488 7489 7490 7491 7492 7493 7494 7495 7496 7497 7498 7499 7500 7501 7502 7503
/* handle null as "default" */
static int dattrs_equal(struct sched_domain_attr *cur, int idx_cur,
			struct sched_domain_attr *new, int idx_new)
{
	struct sched_domain_attr tmp;

	/* fast path */
	if (!new && !cur)
		return 1;

	tmp = SD_ATTR_INIT;
	return !memcmp(cur ? (cur + idx_cur) : &tmp,
			new ? (new + idx_new) : &tmp,
			sizeof(struct sched_domain_attr));
}

P
Paul Jackson 已提交
7504 7505
/*
 * Partition sched domains as specified by the 'ndoms_new'
I
Ingo Molnar 已提交
7506
 * cpumasks in the array doms_new[] of cpumasks. This compares
P
Paul Jackson 已提交
7507 7508 7509
 * doms_new[] to the current sched domain partitioning, doms_cur[].
 * It destroys each deleted domain and builds each new domain.
 *
7510
 * 'doms_new' is an array of cpumask_var_t's of length 'ndoms_new'.
I
Ingo Molnar 已提交
7511 7512 7513
 * The masks don't intersect (don't overlap.) We should setup one
 * sched domain for each mask. CPUs not in any of the cpumasks will
 * not be load balanced. If the same cpumask appears both in the
P
Paul Jackson 已提交
7514 7515 7516
 * current 'doms_cur' domains and in the new 'doms_new', we can leave
 * it as it is.
 *
7517 7518 7519 7520 7521 7522
 * The passed in 'doms_new' should be allocated using
 * alloc_sched_domains.  This routine takes ownership of it and will
 * free_sched_domains it when done with it. If the caller failed the
 * alloc call, then it can pass in doms_new == NULL && ndoms_new == 1,
 * and partition_sched_domains() will fallback to the single partition
 * 'fallback_doms', it also forces the domains to be rebuilt.
P
Paul Jackson 已提交
7523
 *
7524
 * If doms_new == NULL it will be replaced with cpu_online_mask.
7525 7526
 * ndoms_new == 0 is a special case for destroying existing domains,
 * and it will not create the default domain.
7527
 *
P
Paul Jackson 已提交
7528 7529
 * Call with hotplug lock held
 */
7530
void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
7531
			     struct sched_domain_attr *dattr_new)
P
Paul Jackson 已提交
7532
{
7533
	int i, j, n;
7534
	int new_topology;
P
Paul Jackson 已提交
7535

7536
	mutex_lock(&sched_domains_mutex);
7537

7538 7539 7540
	/* always unregister in case we don't destroy any domains */
	unregister_sched_domain_sysctl();

7541 7542 7543
	/* Let architecture update cpu core mappings. */
	new_topology = arch_update_cpu_topology();

7544
	n = doms_new ? ndoms_new : 0;
P
Paul Jackson 已提交
7545 7546 7547

	/* Destroy deleted domains */
	for (i = 0; i < ndoms_cur; i++) {
7548
		for (j = 0; j < n && !new_topology; j++) {
7549
			if (cpumask_equal(doms_cur[i], doms_new[j])
7550
			    && dattrs_equal(dattr_cur, i, dattr_new, j))
P
Paul Jackson 已提交
7551 7552 7553
				goto match1;
		}
		/* no match - a current sched domain not in new doms_new[] */
7554
		detach_destroy_domains(doms_cur[i]);
P
Paul Jackson 已提交
7555 7556 7557 7558
match1:
		;
	}

7559 7560
	if (doms_new == NULL) {
		ndoms_cur = 0;
7561
		doms_new = &fallback_doms;
7562
		cpumask_andnot(doms_new[0], cpu_active_mask, cpu_isolated_map);
7563
		WARN_ON_ONCE(dattr_new);
7564 7565
	}

P
Paul Jackson 已提交
7566 7567
	/* Build new domains */
	for (i = 0; i < ndoms_new; i++) {
7568
		for (j = 0; j < ndoms_cur && !new_topology; j++) {
7569
			if (cpumask_equal(doms_new[i], doms_cur[j])
7570
			    && dattrs_equal(dattr_new, i, dattr_cur, j))
P
Paul Jackson 已提交
7571 7572 7573
				goto match2;
		}
		/* no match - add a new doms_new */
7574
		__build_sched_domains(doms_new[i],
7575
					dattr_new ? dattr_new + i : NULL);
P
Paul Jackson 已提交
7576 7577 7578 7579 7580
match2:
		;
	}

	/* Remember the new sched domains */
7581 7582
	if (doms_cur != &fallback_doms)
		free_sched_domains(doms_cur, ndoms_cur);
7583
	kfree(dattr_cur);	/* kfree(NULL) is safe */
P
Paul Jackson 已提交
7584
	doms_cur = doms_new;
7585
	dattr_cur = dattr_new;
P
Paul Jackson 已提交
7586
	ndoms_cur = ndoms_new;
7587 7588

	register_sched_domain_sysctl();
7589

7590
	mutex_unlock(&sched_domains_mutex);
P
Paul Jackson 已提交
7591 7592
}

7593
#if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT)
7594
static void arch_reinit_sched_domains(void)
7595
{
7596
	get_online_cpus();
7597 7598 7599 7600

	/* Destroy domains first to force the rebuild */
	partition_sched_domains(0, NULL, NULL);

7601
	rebuild_sched_domains();
7602
	put_online_cpus();
7603 7604 7605 7606
}

static ssize_t sched_power_savings_store(const char *buf, size_t count, int smt)
{
7607
	unsigned int level = 0;
7608

7609 7610 7611 7612 7613 7614 7615 7616 7617 7618 7619
	if (sscanf(buf, "%u", &level) != 1)
		return -EINVAL;

	/*
	 * level is always be positive so don't check for
	 * level < POWERSAVINGS_BALANCE_NONE which is 0
	 * What happens on 0 or 1 byte write,
	 * need to check for count as well?
	 */

	if (level >= MAX_POWERSAVINGS_BALANCE_LEVELS)
7620 7621 7622
		return -EINVAL;

	if (smt)
7623
		sched_smt_power_savings = level;
7624
	else
7625
		sched_mc_power_savings = level;
7626

7627
	arch_reinit_sched_domains();
7628

7629
	return count;
7630 7631 7632
}

#ifdef CONFIG_SCHED_MC
7633
static ssize_t sched_mc_power_savings_show(struct sysdev_class *class,
7634
					   struct sysdev_class_attribute *attr,
7635
					   char *page)
7636 7637 7638
{
	return sprintf(page, "%u\n", sched_mc_power_savings);
}
7639
static ssize_t sched_mc_power_savings_store(struct sysdev_class *class,
7640
					    struct sysdev_class_attribute *attr,
7641
					    const char *buf, size_t count)
7642 7643 7644
{
	return sched_power_savings_store(buf, count, 0);
}
7645 7646 7647
static SYSDEV_CLASS_ATTR(sched_mc_power_savings, 0644,
			 sched_mc_power_savings_show,
			 sched_mc_power_savings_store);
7648 7649 7650
#endif

#ifdef CONFIG_SCHED_SMT
7651
static ssize_t sched_smt_power_savings_show(struct sysdev_class *dev,
7652
					    struct sysdev_class_attribute *attr,
7653
					    char *page)
7654 7655 7656
{
	return sprintf(page, "%u\n", sched_smt_power_savings);
}
7657
static ssize_t sched_smt_power_savings_store(struct sysdev_class *dev,
7658
					     struct sysdev_class_attribute *attr,
7659
					     const char *buf, size_t count)
7660 7661 7662
{
	return sched_power_savings_store(buf, count, 1);
}
7663 7664
static SYSDEV_CLASS_ATTR(sched_smt_power_savings, 0644,
		   sched_smt_power_savings_show,
A
Adrian Bunk 已提交
7665 7666 7667
		   sched_smt_power_savings_store);
#endif

7668
int __init sched_create_sysfs_power_savings_entries(struct sysdev_class *cls)
A
Adrian Bunk 已提交
7669 7670 7671 7672 7673 7674 7675 7676 7677 7678 7679 7680 7681 7682 7683
{
	int err = 0;

#ifdef CONFIG_SCHED_SMT
	if (smt_capable())
		err = sysfs_create_file(&cls->kset.kobj,
					&attr_sched_smt_power_savings.attr);
#endif
#ifdef CONFIG_SCHED_MC
	if (!err && mc_capable())
		err = sysfs_create_file(&cls->kset.kobj,
					&attr_sched_mc_power_savings.attr);
#endif
	return err;
}
7684
#endif /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */
7685

L
Linus Torvalds 已提交
7686
/*
7687 7688 7689
 * Update cpusets according to cpu_active mask.  If cpusets are
 * disabled, cpuset_update_active_cpus() becomes a simple wrapper
 * around partition_sched_domains().
L
Linus Torvalds 已提交
7690
 */
7691 7692
static int cpuset_cpu_active(struct notifier_block *nfb, unsigned long action,
			     void *hcpu)
7693
{
7694
	switch (action & ~CPU_TASKS_FROZEN) {
7695
	case CPU_ONLINE:
7696
	case CPU_DOWN_FAILED:
7697
		cpuset_update_active_cpus();
7698
		return NOTIFY_OK;
7699 7700 7701 7702
	default:
		return NOTIFY_DONE;
	}
}
7703

7704 7705
static int cpuset_cpu_inactive(struct notifier_block *nfb, unsigned long action,
			       void *hcpu)
7706 7707 7708 7709 7710
{
	switch (action & ~CPU_TASKS_FROZEN) {
	case CPU_DOWN_PREPARE:
		cpuset_update_active_cpus();
		return NOTIFY_OK;
7711 7712 7713 7714 7715 7716 7717
	default:
		return NOTIFY_DONE;
	}
}

static int update_runtime(struct notifier_block *nfb,
				unsigned long action, void *hcpu)
L
Linus Torvalds 已提交
7718
{
P
Peter Zijlstra 已提交
7719 7720
	int cpu = (int)(long)hcpu;

L
Linus Torvalds 已提交
7721 7722
	switch (action) {
	case CPU_DOWN_PREPARE:
7723
	case CPU_DOWN_PREPARE_FROZEN:
P
Peter Zijlstra 已提交
7724
		disable_runtime(cpu_rq(cpu));
L
Linus Torvalds 已提交
7725 7726 7727
		return NOTIFY_OK;

	case CPU_DOWN_FAILED:
7728
	case CPU_DOWN_FAILED_FROZEN:
L
Linus Torvalds 已提交
7729
	case CPU_ONLINE:
7730
	case CPU_ONLINE_FROZEN:
P
Peter Zijlstra 已提交
7731
		enable_runtime(cpu_rq(cpu));
7732 7733
		return NOTIFY_OK;

L
Linus Torvalds 已提交
7734 7735 7736 7737 7738 7739 7740
	default:
		return NOTIFY_DONE;
	}
}

void __init sched_init_smp(void)
{
7741 7742 7743
	cpumask_var_t non_isolated_cpus;

	alloc_cpumask_var(&non_isolated_cpus, GFP_KERNEL);
7744
	alloc_cpumask_var(&fallback_doms, GFP_KERNEL);
7745

7746 7747 7748 7749 7750
#if defined(CONFIG_NUMA)
	sched_group_nodes_bycpu = kzalloc(nr_cpu_ids * sizeof(void **),
								GFP_KERNEL);
	BUG_ON(sched_group_nodes_bycpu == NULL);
#endif
7751
	get_online_cpus();
7752
	mutex_lock(&sched_domains_mutex);
7753
	arch_init_sched_domains(cpu_active_mask);
7754 7755 7756
	cpumask_andnot(non_isolated_cpus, cpu_possible_mask, cpu_isolated_map);
	if (cpumask_empty(non_isolated_cpus))
		cpumask_set_cpu(smp_processor_id(), non_isolated_cpus);
7757
	mutex_unlock(&sched_domains_mutex);
7758
	put_online_cpus();
7759

7760 7761
	hotcpu_notifier(cpuset_cpu_active, CPU_PRI_CPUSET_ACTIVE);
	hotcpu_notifier(cpuset_cpu_inactive, CPU_PRI_CPUSET_INACTIVE);
7762 7763 7764 7765

	/* RT runtime code needs to handle some hotplug events */
	hotcpu_notifier(update_runtime, 0);

7766
	init_hrtick();
7767 7768

	/* Move init over to a non-isolated CPU */
7769
	if (set_cpus_allowed_ptr(current, non_isolated_cpus) < 0)
7770
		BUG();
I
Ingo Molnar 已提交
7771
	sched_init_granularity();
7772
	free_cpumask_var(non_isolated_cpus);
7773

7774
	init_sched_rt_class();
L
Linus Torvalds 已提交
7775 7776 7777 7778
}
#else
void __init sched_init_smp(void)
{
I
Ingo Molnar 已提交
7779
	sched_init_granularity();
L
Linus Torvalds 已提交
7780 7781 7782
}
#endif /* CONFIG_SMP */

7783 7784
const_debug unsigned int sysctl_timer_migration = 1;

L
Linus Torvalds 已提交
7785 7786 7787 7788 7789 7790 7791
int in_sched_functions(unsigned long addr)
{
	return in_lock_functions(addr) ||
		(addr >= (unsigned long)__sched_text_start
		&& addr < (unsigned long)__sched_text_end);
}

A
Alexey Dobriyan 已提交
7792
static void init_cfs_rq(struct cfs_rq *cfs_rq, struct rq *rq)
I
Ingo Molnar 已提交
7793 7794
{
	cfs_rq->tasks_timeline = RB_ROOT;
7795
	INIT_LIST_HEAD(&cfs_rq->tasks);
I
Ingo Molnar 已提交
7796 7797 7798
#ifdef CONFIG_FAIR_GROUP_SCHED
	cfs_rq->rq = rq;
#endif
P
Peter Zijlstra 已提交
7799
	cfs_rq->min_vruntime = (u64)(-(1LL << 20));
I
Ingo Molnar 已提交
7800 7801
}

P
Peter Zijlstra 已提交
7802 7803 7804 7805 7806 7807 7808 7809 7810 7811 7812 7813 7814
static void init_rt_rq(struct rt_rq *rt_rq, struct rq *rq)
{
	struct rt_prio_array *array;
	int i;

	array = &rt_rq->active;
	for (i = 0; i < MAX_RT_PRIO; i++) {
		INIT_LIST_HEAD(array->queue + i);
		__clear_bit(i, array->bitmap);
	}
	/* delimiter for bitsearch: */
	__set_bit(MAX_RT_PRIO, array->bitmap);

7815
#if defined CONFIG_SMP || defined CONFIG_RT_GROUP_SCHED
7816
	rt_rq->highest_prio.curr = MAX_RT_PRIO;
7817
#ifdef CONFIG_SMP
7818
	rt_rq->highest_prio.next = MAX_RT_PRIO;
P
Peter Zijlstra 已提交
7819 7820
#endif
#endif
P
Peter Zijlstra 已提交
7821 7822 7823
#ifdef CONFIG_SMP
	rt_rq->rt_nr_migratory = 0;
	rt_rq->overloaded = 0;
7824
	plist_head_init_raw(&rt_rq->pushable_tasks, &rq->lock);
P
Peter Zijlstra 已提交
7825 7826 7827 7828
#endif

	rt_rq->rt_time = 0;
	rt_rq->rt_throttled = 0;
P
Peter Zijlstra 已提交
7829
	rt_rq->rt_runtime = 0;
7830
	raw_spin_lock_init(&rt_rq->rt_runtime_lock);
P
Peter Zijlstra 已提交
7831

7832
#ifdef CONFIG_RT_GROUP_SCHED
P
Peter Zijlstra 已提交
7833
	rt_rq->rt_nr_boosted = 0;
P
Peter Zijlstra 已提交
7834 7835
	rt_rq->rq = rq;
#endif
P
Peter Zijlstra 已提交
7836 7837
}

P
Peter Zijlstra 已提交
7838
#ifdef CONFIG_FAIR_GROUP_SCHED
7839 7840 7841
static void init_tg_cfs_entry(struct task_group *tg, struct cfs_rq *cfs_rq,
				struct sched_entity *se, int cpu, int add,
				struct sched_entity *parent)
P
Peter Zijlstra 已提交
7842
{
7843
	struct rq *rq = cpu_rq(cpu);
P
Peter Zijlstra 已提交
7844 7845 7846 7847 7848 7849 7850
	tg->cfs_rq[cpu] = cfs_rq;
	init_cfs_rq(cfs_rq, rq);
	cfs_rq->tg = tg;
	if (add)
		list_add(&cfs_rq->leaf_cfs_rq_list, &rq->leaf_cfs_rq_list);

	tg->se[cpu] = se;
D
Dhaval Giani 已提交
7851 7852 7853 7854
	/* se could be NULL for init_task_group */
	if (!se)
		return;

7855 7856 7857 7858 7859
	if (!parent)
		se->cfs_rq = &rq->cfs;
	else
		se->cfs_rq = parent->my_q;

P
Peter Zijlstra 已提交
7860 7861
	se->my_q = cfs_rq;
	se->load.weight = tg->shares;
7862
	se->load.inv_weight = 0;
7863
	se->parent = parent;
P
Peter Zijlstra 已提交
7864
}
7865
#endif
P
Peter Zijlstra 已提交
7866

7867
#ifdef CONFIG_RT_GROUP_SCHED
7868 7869 7870
static void init_tg_rt_entry(struct task_group *tg, struct rt_rq *rt_rq,
		struct sched_rt_entity *rt_se, int cpu, int add,
		struct sched_rt_entity *parent)
P
Peter Zijlstra 已提交
7871
{
7872 7873
	struct rq *rq = cpu_rq(cpu);

P
Peter Zijlstra 已提交
7874 7875 7876
	tg->rt_rq[cpu] = rt_rq;
	init_rt_rq(rt_rq, rq);
	rt_rq->tg = tg;
P
Peter Zijlstra 已提交
7877
	rt_rq->rt_runtime = tg->rt_bandwidth.rt_runtime;
P
Peter Zijlstra 已提交
7878 7879 7880 7881
	if (add)
		list_add(&rt_rq->leaf_rt_rq_list, &rq->leaf_rt_rq_list);

	tg->rt_se[cpu] = rt_se;
D
Dhaval Giani 已提交
7882 7883 7884
	if (!rt_se)
		return;

7885 7886 7887 7888 7889
	if (!parent)
		rt_se->rt_rq = &rq->rt;
	else
		rt_se->rt_rq = parent->my_q;

P
Peter Zijlstra 已提交
7890
	rt_se->my_q = rt_rq;
7891
	rt_se->parent = parent;
P
Peter Zijlstra 已提交
7892 7893 7894 7895
	INIT_LIST_HEAD(&rt_se->run_list);
}
#endif

L
Linus Torvalds 已提交
7896 7897
void __init sched_init(void)
{
I
Ingo Molnar 已提交
7898
	int i, j;
7899 7900 7901 7902 7903 7904 7905
	unsigned long alloc_size = 0, ptr;

#ifdef CONFIG_FAIR_GROUP_SCHED
	alloc_size += 2 * nr_cpu_ids * sizeof(void **);
#endif
#ifdef CONFIG_RT_GROUP_SCHED
	alloc_size += 2 * nr_cpu_ids * sizeof(void **);
7906
#endif
7907
#ifdef CONFIG_CPUMASK_OFFSTACK
7908
	alloc_size += num_possible_cpus() * cpumask_size();
7909 7910
#endif
	if (alloc_size) {
7911
		ptr = (unsigned long)kzalloc(alloc_size, GFP_NOWAIT);
7912 7913 7914 7915 7916 7917 7918

#ifdef CONFIG_FAIR_GROUP_SCHED
		init_task_group.se = (struct sched_entity **)ptr;
		ptr += nr_cpu_ids * sizeof(void **);

		init_task_group.cfs_rq = (struct cfs_rq **)ptr;
		ptr += nr_cpu_ids * sizeof(void **);
7919

7920
#endif /* CONFIG_FAIR_GROUP_SCHED */
7921 7922 7923 7924 7925
#ifdef CONFIG_RT_GROUP_SCHED
		init_task_group.rt_se = (struct sched_rt_entity **)ptr;
		ptr += nr_cpu_ids * sizeof(void **);

		init_task_group.rt_rq = (struct rt_rq **)ptr;
7926 7927
		ptr += nr_cpu_ids * sizeof(void **);

7928
#endif /* CONFIG_RT_GROUP_SCHED */
7929 7930 7931 7932 7933 7934
#ifdef CONFIG_CPUMASK_OFFSTACK
		for_each_possible_cpu(i) {
			per_cpu(load_balance_tmpmask, i) = (void *)ptr;
			ptr += cpumask_size();
		}
#endif /* CONFIG_CPUMASK_OFFSTACK */
7935
	}
I
Ingo Molnar 已提交
7936

G
Gregory Haskins 已提交
7937 7938 7939 7940
#ifdef CONFIG_SMP
	init_defrootdomain();
#endif

7941 7942 7943 7944 7945 7946
	init_rt_bandwidth(&def_rt_bandwidth,
			global_rt_period(), global_rt_runtime());

#ifdef CONFIG_RT_GROUP_SCHED
	init_rt_bandwidth(&init_task_group.rt_bandwidth,
			global_rt_period(), global_rt_runtime());
7947
#endif /* CONFIG_RT_GROUP_SCHED */
7948

D
Dhaval Giani 已提交
7949
#ifdef CONFIG_CGROUP_SCHED
P
Peter Zijlstra 已提交
7950
	list_add(&init_task_group.list, &task_groups);
P
Peter Zijlstra 已提交
7951 7952
	INIT_LIST_HEAD(&init_task_group.children);

D
Dhaval Giani 已提交
7953
#endif /* CONFIG_CGROUP_SCHED */
P
Peter Zijlstra 已提交
7954

7955 7956 7957 7958
#if defined CONFIG_FAIR_GROUP_SCHED && defined CONFIG_SMP
	update_shares_data = __alloc_percpu(nr_cpu_ids * sizeof(unsigned long),
					    __alignof__(unsigned long));
#endif
7959
	for_each_possible_cpu(i) {
7960
		struct rq *rq;
L
Linus Torvalds 已提交
7961 7962

		rq = cpu_rq(i);
7963
		raw_spin_lock_init(&rq->lock);
N
Nick Piggin 已提交
7964
		rq->nr_running = 0;
7965 7966
		rq->calc_load_active = 0;
		rq->calc_load_update = jiffies + LOAD_FREQ;
I
Ingo Molnar 已提交
7967
		init_cfs_rq(&rq->cfs, rq);
P
Peter Zijlstra 已提交
7968
		init_rt_rq(&rq->rt, rq);
I
Ingo Molnar 已提交
7969
#ifdef CONFIG_FAIR_GROUP_SCHED
7970
		init_task_group.shares = init_task_group_load;
P
Peter Zijlstra 已提交
7971
		INIT_LIST_HEAD(&rq->leaf_cfs_rq_list);
D
Dhaval Giani 已提交
7972 7973 7974 7975 7976 7977 7978 7979 7980 7981 7982 7983 7984 7985 7986
#ifdef CONFIG_CGROUP_SCHED
		/*
		 * How much cpu bandwidth does init_task_group get?
		 *
		 * In case of task-groups formed thr' the cgroup filesystem, it
		 * gets 100% of the cpu resources in the system. This overall
		 * system cpu resource is divided among the tasks of
		 * init_task_group and its child task-groups in a fair manner,
		 * based on each entity's (task or task-group's) weight
		 * (se->load.weight).
		 *
		 * In other words, if init_task_group has 10 tasks of weight
		 * 1024) and two child groups A0 and A1 (of weight 1024 each),
		 * then A0's share of the cpu resource is:
		 *
7987
		 *	A0's bandwidth = 1024 / (10*1024 + 1024 + 1024) = 8.33%
D
Dhaval Giani 已提交
7988 7989 7990 7991
		 *
		 * We achieve this by letting init_task_group's tasks sit
		 * directly in rq->cfs (i.e init_task_group->se[] = NULL).
		 */
7992
		init_tg_cfs_entry(&init_task_group, &rq->cfs, NULL, i, 1, NULL);
7993
#endif
D
Dhaval Giani 已提交
7994 7995 7996
#endif /* CONFIG_FAIR_GROUP_SCHED */

		rq->rt.rt_runtime = def_rt_bandwidth.rt_runtime;
7997
#ifdef CONFIG_RT_GROUP_SCHED
P
Peter Zijlstra 已提交
7998
		INIT_LIST_HEAD(&rq->leaf_rt_rq_list);
D
Dhaval Giani 已提交
7999
#ifdef CONFIG_CGROUP_SCHED
8000
		init_tg_rt_entry(&init_task_group, &rq->rt, NULL, i, 1, NULL);
D
Dhaval Giani 已提交
8001
#endif
I
Ingo Molnar 已提交
8002
#endif
L
Linus Torvalds 已提交
8003

I
Ingo Molnar 已提交
8004 8005
		for (j = 0; j < CPU_LOAD_IDX_MAX; j++)
			rq->cpu_load[j] = 0;
8006 8007 8008

		rq->last_load_update_tick = jiffies;

L
Linus Torvalds 已提交
8009
#ifdef CONFIG_SMP
N
Nick Piggin 已提交
8010
		rq->sd = NULL;
G
Gregory Haskins 已提交
8011
		rq->rd = NULL;
8012
		rq->cpu_power = SCHED_LOAD_SCALE;
8013
		rq->post_schedule = 0;
L
Linus Torvalds 已提交
8014
		rq->active_balance = 0;
I
Ingo Molnar 已提交
8015
		rq->next_balance = jiffies;
L
Linus Torvalds 已提交
8016
		rq->push_cpu = 0;
8017
		rq->cpu = i;
8018
		rq->online = 0;
8019 8020
		rq->idle_stamp = 0;
		rq->avg_idle = 2*sysctl_sched_migration_cost;
8021
		rq_attach_root(rq, &def_root_domain);
8022 8023 8024 8025
#ifdef CONFIG_NO_HZ
		rq->nohz_balance_kick = 0;
		init_sched_softirq_csd(&per_cpu(remote_sched_softirq_cb, i));
#endif
L
Linus Torvalds 已提交
8026
#endif
P
Peter Zijlstra 已提交
8027
		init_rq_hrtick(rq);
L
Linus Torvalds 已提交
8028 8029 8030
		atomic_set(&rq->nr_iowait, 0);
	}

8031
	set_load_weight(&init_task);
8032

8033 8034 8035 8036
#ifdef CONFIG_PREEMPT_NOTIFIERS
	INIT_HLIST_HEAD(&init_task.preempt_notifiers);
#endif

8037
#ifdef CONFIG_SMP
8038
	open_softirq(SCHED_SOFTIRQ, run_rebalance_domains);
8039 8040
#endif

8041
#ifdef CONFIG_RT_MUTEXES
8042
	plist_head_init_raw(&init_task.pi_waiters, &init_task.pi_lock);
8043 8044
#endif

L
Linus Torvalds 已提交
8045 8046 8047 8048 8049 8050 8051 8052 8053 8054 8055 8056 8057
	/*
	 * The boot idle thread does lazy MMU switching as well:
	 */
	atomic_inc(&init_mm.mm_count);
	enter_lazy_tlb(&init_mm, current);

	/*
	 * Make us the idle thread. Technically, schedule() should not be
	 * called from this thread, however somewhere below it might be,
	 * but because we are the idle thread, we just pick up running again
	 * when this runqueue becomes "idle".
	 */
	init_idle(current, smp_processor_id());
8058 8059 8060

	calc_load_update = jiffies + LOAD_FREQ;

I
Ingo Molnar 已提交
8061 8062 8063 8064
	/*
	 * During early bootup we pretend to be a normal task:
	 */
	current->sched_class = &fair_sched_class;
8065

8066
	/* Allocate the nohz_cpu_mask if CONFIG_CPUMASK_OFFSTACK */
8067
	zalloc_cpumask_var(&nohz_cpu_mask, GFP_NOWAIT);
8068
#ifdef CONFIG_SMP
8069
#ifdef CONFIG_NO_HZ
8070 8071 8072 8073 8074
	zalloc_cpumask_var(&nohz.idle_cpus_mask, GFP_NOWAIT);
	alloc_cpumask_var(&nohz.grp_idle_mask, GFP_NOWAIT);
	atomic_set(&nohz.load_balancer, nr_cpu_ids);
	atomic_set(&nohz.first_pick_cpu, nr_cpu_ids);
	atomic_set(&nohz.second_pick_cpu, nr_cpu_ids);
8075
#endif
R
Rusty Russell 已提交
8076 8077 8078
	/* May be allocated at isolcpus cmdline parse time */
	if (cpu_isolated_map == NULL)
		zalloc_cpumask_var(&cpu_isolated_map, GFP_NOWAIT);
8079
#endif /* SMP */
8080

8081
	perf_event_init();
8082

8083
	scheduler_running = 1;
L
Linus Torvalds 已提交
8084 8085 8086
}

#ifdef CONFIG_DEBUG_SPINLOCK_SLEEP
8087 8088
static inline int preempt_count_equals(int preempt_offset)
{
8089
	int nested = (preempt_count() & ~PREEMPT_ACTIVE) + rcu_preempt_depth();
8090 8091 8092 8093

	return (nested == PREEMPT_INATOMIC_BASE + preempt_offset);
}

8094
void __might_sleep(const char *file, int line, int preempt_offset)
L
Linus Torvalds 已提交
8095
{
8096
#ifdef in_atomic
L
Linus Torvalds 已提交
8097 8098
	static unsigned long prev_jiffy;	/* ratelimiting */

8099 8100
	if ((preempt_count_equals(preempt_offset) && !irqs_disabled()) ||
	    system_state != SYSTEM_RUNNING || oops_in_progress)
I
Ingo Molnar 已提交
8101 8102 8103 8104 8105
		return;
	if (time_before(jiffies, prev_jiffy + HZ) && prev_jiffy)
		return;
	prev_jiffy = jiffies;

P
Peter Zijlstra 已提交
8106 8107 8108 8109 8110 8111 8112
	printk(KERN_ERR
		"BUG: sleeping function called from invalid context at %s:%d\n",
			file, line);
	printk(KERN_ERR
		"in_atomic(): %d, irqs_disabled(): %d, pid: %d, name: %s\n",
			in_atomic(), irqs_disabled(),
			current->pid, current->comm);
I
Ingo Molnar 已提交
8113 8114 8115 8116 8117

	debug_show_held_locks(current);
	if (irqs_disabled())
		print_irqtrace_events(current);
	dump_stack();
L
Linus Torvalds 已提交
8118 8119 8120 8121 8122 8123
#endif
}
EXPORT_SYMBOL(__might_sleep);
#endif

#ifdef CONFIG_MAGIC_SYSRQ
8124 8125 8126
static void normalize_task(struct rq *rq, struct task_struct *p)
{
	int on_rq;
8127

8128 8129 8130 8131 8132 8133 8134 8135 8136 8137
	on_rq = p->se.on_rq;
	if (on_rq)
		deactivate_task(rq, p, 0);
	__setscheduler(rq, p, SCHED_NORMAL, 0);
	if (on_rq) {
		activate_task(rq, p, 0);
		resched_task(rq->curr);
	}
}

L
Linus Torvalds 已提交
8138 8139
void normalize_rt_tasks(void)
{
8140
	struct task_struct *g, *p;
L
Linus Torvalds 已提交
8141
	unsigned long flags;
8142
	struct rq *rq;
L
Linus Torvalds 已提交
8143

8144
	read_lock_irqsave(&tasklist_lock, flags);
8145
	do_each_thread(g, p) {
8146 8147 8148 8149 8150 8151
		/*
		 * Only normalize user tasks:
		 */
		if (!p->mm)
			continue;

I
Ingo Molnar 已提交
8152 8153
		p->se.exec_start		= 0;
#ifdef CONFIG_SCHEDSTATS
8154 8155 8156
		p->se.statistics.wait_start	= 0;
		p->se.statistics.sleep_start	= 0;
		p->se.statistics.block_start	= 0;
I
Ingo Molnar 已提交
8157
#endif
I
Ingo Molnar 已提交
8158 8159 8160 8161 8162 8163 8164 8165

		if (!rt_task(p)) {
			/*
			 * Renice negative nice level userspace
			 * tasks back to 0:
			 */
			if (TASK_NICE(p) < 0 && p->mm)
				set_user_nice(p, 0);
L
Linus Torvalds 已提交
8166
			continue;
I
Ingo Molnar 已提交
8167
		}
L
Linus Torvalds 已提交
8168

8169
		raw_spin_lock(&p->pi_lock);
8170
		rq = __task_rq_lock(p);
L
Linus Torvalds 已提交
8171

8172
		normalize_task(rq, p);
8173

8174
		__task_rq_unlock(rq);
8175
		raw_spin_unlock(&p->pi_lock);
8176 8177
	} while_each_thread(g, p);

8178
	read_unlock_irqrestore(&tasklist_lock, flags);
L
Linus Torvalds 已提交
8179 8180 8181
}

#endif /* CONFIG_MAGIC_SYSRQ */
8182

8183
#if defined(CONFIG_IA64) || defined(CONFIG_KGDB_KDB)
8184
/*
8185
 * These functions are only useful for the IA64 MCA handling, or kdb.
8186 8187 8188 8189 8190 8191 8192 8193 8194 8195 8196 8197 8198 8199
 *
 * They can only be called when the whole system has been
 * stopped - every CPU needs to be quiescent, and no scheduling
 * activity can take place. Using them for anything else would
 * be a serious bug, and as a result, they aren't even visible
 * under any other configuration.
 */

/**
 * curr_task - return the current task for a given cpu.
 * @cpu: the processor in question.
 *
 * ONLY VALID WHEN THE WHOLE SYSTEM IS STOPPED!
 */
8200
struct task_struct *curr_task(int cpu)
8201 8202 8203 8204
{
	return cpu_curr(cpu);
}

8205 8206 8207
#endif /* defined(CONFIG_IA64) || defined(CONFIG_KGDB_KDB) */

#ifdef CONFIG_IA64
8208 8209 8210 8211 8212 8213
/**
 * set_curr_task - set the current task for a given cpu.
 * @cpu: the processor in question.
 * @p: the task pointer to set.
 *
 * Description: This function must only be used when non-maskable interrupts
I
Ingo Molnar 已提交
8214 8215
 * are serviced on a separate stack. It allows the architecture to switch the
 * notion of the current task on a cpu in a non-blocking manner. This function
8216 8217 8218 8219 8220 8221 8222
 * must be called with all CPU's synchronized, and interrupts disabled, the
 * and caller must save the original value of the current task (see
 * curr_task() above) and restore that value before reenabling interrupts and
 * re-starting the system.
 *
 * ONLY VALID WHEN THE WHOLE SYSTEM IS STOPPED!
 */
8223
void set_curr_task(int cpu, struct task_struct *p)
8224 8225 8226 8227 8228
{
	cpu_curr(cpu) = p;
}

#endif
S
Srivatsa Vaddagiri 已提交
8229

8230 8231
#ifdef CONFIG_FAIR_GROUP_SCHED
static void free_fair_sched_group(struct task_group *tg)
P
Peter Zijlstra 已提交
8232 8233 8234 8235 8236 8237 8238 8239 8240 8241 8242 8243 8244 8245
{
	int i;

	for_each_possible_cpu(i) {
		if (tg->cfs_rq)
			kfree(tg->cfs_rq[i]);
		if (tg->se)
			kfree(tg->se[i]);
	}

	kfree(tg->cfs_rq);
	kfree(tg->se);
}

8246 8247
static
int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent)
S
Srivatsa Vaddagiri 已提交
8248 8249
{
	struct cfs_rq *cfs_rq;
8250
	struct sched_entity *se;
8251
	struct rq *rq;
S
Srivatsa Vaddagiri 已提交
8252 8253
	int i;

8254
	tg->cfs_rq = kzalloc(sizeof(cfs_rq) * nr_cpu_ids, GFP_KERNEL);
S
Srivatsa Vaddagiri 已提交
8255 8256
	if (!tg->cfs_rq)
		goto err;
8257
	tg->se = kzalloc(sizeof(se) * nr_cpu_ids, GFP_KERNEL);
S
Srivatsa Vaddagiri 已提交
8258 8259
	if (!tg->se)
		goto err;
8260 8261

	tg->shares = NICE_0_LOAD;
S
Srivatsa Vaddagiri 已提交
8262 8263

	for_each_possible_cpu(i) {
8264
		rq = cpu_rq(i);
S
Srivatsa Vaddagiri 已提交
8265

8266 8267
		cfs_rq = kzalloc_node(sizeof(struct cfs_rq),
				      GFP_KERNEL, cpu_to_node(i));
S
Srivatsa Vaddagiri 已提交
8268 8269 8270
		if (!cfs_rq)
			goto err;

8271 8272
		se = kzalloc_node(sizeof(struct sched_entity),
				  GFP_KERNEL, cpu_to_node(i));
S
Srivatsa Vaddagiri 已提交
8273
		if (!se)
8274
			goto err_free_rq;
S
Srivatsa Vaddagiri 已提交
8275

8276
		init_tg_cfs_entry(tg, cfs_rq, se, i, 0, parent->se[i]);
8277 8278 8279 8280
	}

	return 1;

P
Peter Zijlstra 已提交
8281
err_free_rq:
8282
	kfree(cfs_rq);
P
Peter Zijlstra 已提交
8283
err:
8284 8285 8286 8287 8288 8289 8290 8291 8292 8293 8294 8295 8296
	return 0;
}

static inline void register_fair_sched_group(struct task_group *tg, int cpu)
{
	list_add_rcu(&tg->cfs_rq[cpu]->leaf_cfs_rq_list,
			&cpu_rq(cpu)->leaf_cfs_rq_list);
}

static inline void unregister_fair_sched_group(struct task_group *tg, int cpu)
{
	list_del_rcu(&tg->cfs_rq[cpu]->leaf_cfs_rq_list);
}
8297
#else /* !CONFG_FAIR_GROUP_SCHED */
8298 8299 8300 8301
static inline void free_fair_sched_group(struct task_group *tg)
{
}

8302 8303
static inline
int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent)
8304 8305 8306 8307 8308 8309 8310 8311 8312 8313 8314
{
	return 1;
}

static inline void register_fair_sched_group(struct task_group *tg, int cpu)
{
}

static inline void unregister_fair_sched_group(struct task_group *tg, int cpu)
{
}
8315
#endif /* CONFIG_FAIR_GROUP_SCHED */
8316 8317

#ifdef CONFIG_RT_GROUP_SCHED
8318 8319 8320 8321
static void free_rt_sched_group(struct task_group *tg)
{
	int i;

8322 8323
	destroy_rt_bandwidth(&tg->rt_bandwidth);

8324 8325 8326 8327 8328 8329 8330 8331 8332 8333 8334
	for_each_possible_cpu(i) {
		if (tg->rt_rq)
			kfree(tg->rt_rq[i]);
		if (tg->rt_se)
			kfree(tg->rt_se[i]);
	}

	kfree(tg->rt_rq);
	kfree(tg->rt_se);
}

8335 8336
static
int alloc_rt_sched_group(struct task_group *tg, struct task_group *parent)
8337 8338
{
	struct rt_rq *rt_rq;
8339
	struct sched_rt_entity *rt_se;
8340 8341 8342
	struct rq *rq;
	int i;

8343
	tg->rt_rq = kzalloc(sizeof(rt_rq) * nr_cpu_ids, GFP_KERNEL);
8344 8345
	if (!tg->rt_rq)
		goto err;
8346
	tg->rt_se = kzalloc(sizeof(rt_se) * nr_cpu_ids, GFP_KERNEL);
8347 8348 8349
	if (!tg->rt_se)
		goto err;

8350 8351
	init_rt_bandwidth(&tg->rt_bandwidth,
			ktime_to_ns(def_rt_bandwidth.rt_period), 0);
8352 8353 8354 8355

	for_each_possible_cpu(i) {
		rq = cpu_rq(i);

8356 8357
		rt_rq = kzalloc_node(sizeof(struct rt_rq),
				     GFP_KERNEL, cpu_to_node(i));
P
Peter Zijlstra 已提交
8358 8359
		if (!rt_rq)
			goto err;
S
Srivatsa Vaddagiri 已提交
8360

8361 8362
		rt_se = kzalloc_node(sizeof(struct sched_rt_entity),
				     GFP_KERNEL, cpu_to_node(i));
P
Peter Zijlstra 已提交
8363
		if (!rt_se)
8364
			goto err_free_rq;
S
Srivatsa Vaddagiri 已提交
8365

8366
		init_tg_rt_entry(tg, rt_rq, rt_se, i, 0, parent->rt_se[i]);
S
Srivatsa Vaddagiri 已提交
8367 8368
	}

8369 8370
	return 1;

P
Peter Zijlstra 已提交
8371
err_free_rq:
8372
	kfree(rt_rq);
P
Peter Zijlstra 已提交
8373
err:
8374 8375 8376 8377 8378 8379 8380 8381 8382 8383 8384 8385 8386
	return 0;
}

static inline void register_rt_sched_group(struct task_group *tg, int cpu)
{
	list_add_rcu(&tg->rt_rq[cpu]->leaf_rt_rq_list,
			&cpu_rq(cpu)->leaf_rt_rq_list);
}

static inline void unregister_rt_sched_group(struct task_group *tg, int cpu)
{
	list_del_rcu(&tg->rt_rq[cpu]->leaf_rt_rq_list);
}
8387
#else /* !CONFIG_RT_GROUP_SCHED */
8388 8389 8390 8391
static inline void free_rt_sched_group(struct task_group *tg)
{
}

8392 8393
static inline
int alloc_rt_sched_group(struct task_group *tg, struct task_group *parent)
8394 8395 8396 8397 8398 8399 8400 8401 8402 8403 8404
{
	return 1;
}

static inline void register_rt_sched_group(struct task_group *tg, int cpu)
{
}

static inline void unregister_rt_sched_group(struct task_group *tg, int cpu)
{
}
8405
#endif /* CONFIG_RT_GROUP_SCHED */
8406

D
Dhaval Giani 已提交
8407
#ifdef CONFIG_CGROUP_SCHED
8408 8409 8410 8411 8412 8413 8414 8415
static void free_sched_group(struct task_group *tg)
{
	free_fair_sched_group(tg);
	free_rt_sched_group(tg);
	kfree(tg);
}

/* allocate runqueue etc for a new task group */
8416
struct task_group *sched_create_group(struct task_group *parent)
8417 8418 8419 8420 8421 8422 8423 8424 8425
{
	struct task_group *tg;
	unsigned long flags;
	int i;

	tg = kzalloc(sizeof(*tg), GFP_KERNEL);
	if (!tg)
		return ERR_PTR(-ENOMEM);

8426
	if (!alloc_fair_sched_group(tg, parent))
8427 8428
		goto err;

8429
	if (!alloc_rt_sched_group(tg, parent))
8430 8431
		goto err;

8432
	spin_lock_irqsave(&task_group_lock, flags);
8433
	for_each_possible_cpu(i) {
8434 8435
		register_fair_sched_group(tg, i);
		register_rt_sched_group(tg, i);
8436
	}
P
Peter Zijlstra 已提交
8437
	list_add_rcu(&tg->list, &task_groups);
P
Peter Zijlstra 已提交
8438 8439 8440 8441 8442

	WARN_ON(!parent); /* root should already exist */

	tg->parent = parent;
	INIT_LIST_HEAD(&tg->children);
8443
	list_add_rcu(&tg->siblings, &parent->children);
8444
	spin_unlock_irqrestore(&task_group_lock, flags);
S
Srivatsa Vaddagiri 已提交
8445

8446
	return tg;
S
Srivatsa Vaddagiri 已提交
8447 8448

err:
P
Peter Zijlstra 已提交
8449
	free_sched_group(tg);
S
Srivatsa Vaddagiri 已提交
8450 8451 8452
	return ERR_PTR(-ENOMEM);
}

8453
/* rcu callback to free various structures associated with a task group */
P
Peter Zijlstra 已提交
8454
static void free_sched_group_rcu(struct rcu_head *rhp)
S
Srivatsa Vaddagiri 已提交
8455 8456
{
	/* now it should be safe to free those cfs_rqs */
P
Peter Zijlstra 已提交
8457
	free_sched_group(container_of(rhp, struct task_group, rcu));
S
Srivatsa Vaddagiri 已提交
8458 8459
}

8460
/* Destroy runqueue etc associated with a task group */
8461
void sched_destroy_group(struct task_group *tg)
S
Srivatsa Vaddagiri 已提交
8462
{
8463
	unsigned long flags;
8464
	int i;
S
Srivatsa Vaddagiri 已提交
8465

8466
	spin_lock_irqsave(&task_group_lock, flags);
8467
	for_each_possible_cpu(i) {
8468 8469
		unregister_fair_sched_group(tg, i);
		unregister_rt_sched_group(tg, i);
8470
	}
P
Peter Zijlstra 已提交
8471
	list_del_rcu(&tg->list);
P
Peter Zijlstra 已提交
8472
	list_del_rcu(&tg->siblings);
8473
	spin_unlock_irqrestore(&task_group_lock, flags);
8474 8475

	/* wait for possible concurrent references to cfs_rqs complete */
P
Peter Zijlstra 已提交
8476
	call_rcu(&tg->rcu, free_sched_group_rcu);
S
Srivatsa Vaddagiri 已提交
8477 8478
}

8479
/* change task's runqueue when it moves between groups.
I
Ingo Molnar 已提交
8480 8481 8482
 *	The caller of this function should have put the task in its new group
 *	by now. This function just updates tsk->se.cfs_rq and tsk->se.parent to
 *	reflect its new group.
8483 8484
 */
void sched_move_task(struct task_struct *tsk)
S
Srivatsa Vaddagiri 已提交
8485 8486 8487 8488 8489 8490 8491
{
	int on_rq, running;
	unsigned long flags;
	struct rq *rq;

	rq = task_rq_lock(tsk, &flags);

8492
	running = task_current(rq, tsk);
S
Srivatsa Vaddagiri 已提交
8493 8494
	on_rq = tsk->se.on_rq;

8495
	if (on_rq)
S
Srivatsa Vaddagiri 已提交
8496
		dequeue_task(rq, tsk, 0);
8497 8498
	if (unlikely(running))
		tsk->sched_class->put_prev_task(rq, tsk);
S
Srivatsa Vaddagiri 已提交
8499

P
Peter Zijlstra 已提交
8500
	set_task_rq(tsk, task_cpu(tsk));
S
Srivatsa Vaddagiri 已提交
8501

P
Peter Zijlstra 已提交
8502 8503
#ifdef CONFIG_FAIR_GROUP_SCHED
	if (tsk->sched_class->moved_group)
8504
		tsk->sched_class->moved_group(tsk, on_rq);
P
Peter Zijlstra 已提交
8505 8506
#endif

8507 8508 8509
	if (unlikely(running))
		tsk->sched_class->set_curr_task(rq);
	if (on_rq)
8510
		enqueue_task(rq, tsk, 0);
S
Srivatsa Vaddagiri 已提交
8511 8512 8513

	task_rq_unlock(rq, &flags);
}
D
Dhaval Giani 已提交
8514
#endif /* CONFIG_CGROUP_SCHED */
S
Srivatsa Vaddagiri 已提交
8515

8516
#ifdef CONFIG_FAIR_GROUP_SCHED
8517
static void __set_se_shares(struct sched_entity *se, unsigned long shares)
S
Srivatsa Vaddagiri 已提交
8518 8519 8520 8521 8522
{
	struct cfs_rq *cfs_rq = se->cfs_rq;
	int on_rq;

	on_rq = se->on_rq;
8523
	if (on_rq)
S
Srivatsa Vaddagiri 已提交
8524 8525 8526
		dequeue_entity(cfs_rq, se, 0);

	se->load.weight = shares;
8527
	se->load.inv_weight = 0;
S
Srivatsa Vaddagiri 已提交
8528

8529
	if (on_rq)
S
Srivatsa Vaddagiri 已提交
8530
		enqueue_entity(cfs_rq, se, 0);
8531
}
8532

8533 8534 8535 8536 8537 8538
static void set_se_shares(struct sched_entity *se, unsigned long shares)
{
	struct cfs_rq *cfs_rq = se->cfs_rq;
	struct rq *rq = cfs_rq->rq;
	unsigned long flags;

8539
	raw_spin_lock_irqsave(&rq->lock, flags);
8540
	__set_se_shares(se, shares);
8541
	raw_spin_unlock_irqrestore(&rq->lock, flags);
S
Srivatsa Vaddagiri 已提交
8542 8543
}

8544 8545
static DEFINE_MUTEX(shares_mutex);

8546
int sched_group_set_shares(struct task_group *tg, unsigned long shares)
S
Srivatsa Vaddagiri 已提交
8547 8548
{
	int i;
8549
	unsigned long flags;
8550

8551 8552 8553 8554 8555 8556
	/*
	 * We can't change the weight of the root cgroup.
	 */
	if (!tg->se[0])
		return -EINVAL;

8557 8558
	if (shares < MIN_SHARES)
		shares = MIN_SHARES;
8559 8560
	else if (shares > MAX_SHARES)
		shares = MAX_SHARES;
8561

8562
	mutex_lock(&shares_mutex);
8563
	if (tg->shares == shares)
8564
		goto done;
S
Srivatsa Vaddagiri 已提交
8565

8566
	spin_lock_irqsave(&task_group_lock, flags);
8567 8568
	for_each_possible_cpu(i)
		unregister_fair_sched_group(tg, i);
P
Peter Zijlstra 已提交
8569
	list_del_rcu(&tg->siblings);
8570
	spin_unlock_irqrestore(&task_group_lock, flags);
8571 8572 8573 8574 8575 8576 8577 8578

	/* wait for any ongoing reference to this group to finish */
	synchronize_sched();

	/*
	 * Now we are free to modify the group's share on each cpu
	 * w/o tripping rebalance_share or load_balance_fair.
	 */
8579
	tg->shares = shares;
8580 8581 8582 8583 8584
	for_each_possible_cpu(i) {
		/*
		 * force a rebalance
		 */
		cfs_rq_set_shares(tg->cfs_rq[i], 0);
8585
		set_se_shares(tg->se[i], shares);
8586
	}
S
Srivatsa Vaddagiri 已提交
8587

8588 8589 8590 8591
	/*
	 * Enable load balance activity on this group, by inserting it back on
	 * each cpu's rq->leaf_cfs_rq_list.
	 */
8592
	spin_lock_irqsave(&task_group_lock, flags);
8593 8594
	for_each_possible_cpu(i)
		register_fair_sched_group(tg, i);
P
Peter Zijlstra 已提交
8595
	list_add_rcu(&tg->siblings, &tg->parent->children);
8596
	spin_unlock_irqrestore(&task_group_lock, flags);
8597
done:
8598
	mutex_unlock(&shares_mutex);
8599
	return 0;
S
Srivatsa Vaddagiri 已提交
8600 8601
}

8602 8603 8604 8605
unsigned long sched_group_shares(struct task_group *tg)
{
	return tg->shares;
}
8606
#endif
8607

8608
#ifdef CONFIG_RT_GROUP_SCHED
P
Peter Zijlstra 已提交
8609
/*
P
Peter Zijlstra 已提交
8610
 * Ensure that the real time constraints are schedulable.
P
Peter Zijlstra 已提交
8611
 */
P
Peter Zijlstra 已提交
8612 8613 8614 8615 8616
static DEFINE_MUTEX(rt_constraints_mutex);

static unsigned long to_ratio(u64 period, u64 runtime)
{
	if (runtime == RUNTIME_INF)
P
Peter Zijlstra 已提交
8617
		return 1ULL << 20;
P
Peter Zijlstra 已提交
8618

P
Peter Zijlstra 已提交
8619
	return div64_u64(runtime << 20, period);
P
Peter Zijlstra 已提交
8620 8621
}

P
Peter Zijlstra 已提交
8622 8623
/* Must be called with tasklist_lock held */
static inline int tg_has_rt_tasks(struct task_group *tg)
8624
{
P
Peter Zijlstra 已提交
8625
	struct task_struct *g, *p;
8626

P
Peter Zijlstra 已提交
8627 8628 8629 8630
	do_each_thread(g, p) {
		if (rt_task(p) && rt_rq_of_se(&p->rt)->tg == tg)
			return 1;
	} while_each_thread(g, p);
8631

P
Peter Zijlstra 已提交
8632 8633
	return 0;
}
8634

P
Peter Zijlstra 已提交
8635 8636 8637 8638 8639
struct rt_schedulable_data {
	struct task_group *tg;
	u64 rt_period;
	u64 rt_runtime;
};
8640

P
Peter Zijlstra 已提交
8641 8642 8643 8644 8645 8646
static int tg_schedulable(struct task_group *tg, void *data)
{
	struct rt_schedulable_data *d = data;
	struct task_group *child;
	unsigned long total, sum = 0;
	u64 period, runtime;
8647

P
Peter Zijlstra 已提交
8648 8649
	period = ktime_to_ns(tg->rt_bandwidth.rt_period);
	runtime = tg->rt_bandwidth.rt_runtime;
8650

P
Peter Zijlstra 已提交
8651 8652 8653
	if (tg == d->tg) {
		period = d->rt_period;
		runtime = d->rt_runtime;
8654 8655
	}

8656 8657 8658 8659 8660
	/*
	 * Cannot have more runtime than the period.
	 */
	if (runtime > period && runtime != RUNTIME_INF)
		return -EINVAL;
P
Peter Zijlstra 已提交
8661

8662 8663 8664
	/*
	 * Ensure we don't starve existing RT tasks.
	 */
P
Peter Zijlstra 已提交
8665 8666
	if (rt_bandwidth_enabled() && !runtime && tg_has_rt_tasks(tg))
		return -EBUSY;
P
Peter Zijlstra 已提交
8667

P
Peter Zijlstra 已提交
8668
	total = to_ratio(period, runtime);
P
Peter Zijlstra 已提交
8669

8670 8671 8672 8673 8674
	/*
	 * Nobody can have more than the global setting allows.
	 */
	if (total > to_ratio(global_rt_period(), global_rt_runtime()))
		return -EINVAL;
P
Peter Zijlstra 已提交
8675

8676 8677 8678
	/*
	 * The sum of our children's runtime should not exceed our own.
	 */
P
Peter Zijlstra 已提交
8679 8680 8681
	list_for_each_entry_rcu(child, &tg->children, siblings) {
		period = ktime_to_ns(child->rt_bandwidth.rt_period);
		runtime = child->rt_bandwidth.rt_runtime;
P
Peter Zijlstra 已提交
8682

P
Peter Zijlstra 已提交
8683 8684 8685 8686
		if (child == d->tg) {
			period = d->rt_period;
			runtime = d->rt_runtime;
		}
P
Peter Zijlstra 已提交
8687

P
Peter Zijlstra 已提交
8688
		sum += to_ratio(period, runtime);
P
Peter Zijlstra 已提交
8689
	}
P
Peter Zijlstra 已提交
8690

P
Peter Zijlstra 已提交
8691 8692 8693 8694
	if (sum > total)
		return -EINVAL;

	return 0;
P
Peter Zijlstra 已提交
8695 8696
}

P
Peter Zijlstra 已提交
8697
static int __rt_schedulable(struct task_group *tg, u64 period, u64 runtime)
8698
{
P
Peter Zijlstra 已提交
8699 8700 8701 8702 8703 8704 8705
	struct rt_schedulable_data data = {
		.tg = tg,
		.rt_period = period,
		.rt_runtime = runtime,
	};

	return walk_tg_tree(tg_schedulable, tg_nop, &data);
8706 8707
}

8708 8709
static int tg_set_bandwidth(struct task_group *tg,
		u64 rt_period, u64 rt_runtime)
P
Peter Zijlstra 已提交
8710
{
P
Peter Zijlstra 已提交
8711
	int i, err = 0;
P
Peter Zijlstra 已提交
8712 8713

	mutex_lock(&rt_constraints_mutex);
8714
	read_lock(&tasklist_lock);
P
Peter Zijlstra 已提交
8715 8716
	err = __rt_schedulable(tg, rt_period, rt_runtime);
	if (err)
P
Peter Zijlstra 已提交
8717
		goto unlock;
P
Peter Zijlstra 已提交
8718

8719
	raw_spin_lock_irq(&tg->rt_bandwidth.rt_runtime_lock);
8720 8721
	tg->rt_bandwidth.rt_period = ns_to_ktime(rt_period);
	tg->rt_bandwidth.rt_runtime = rt_runtime;
P
Peter Zijlstra 已提交
8722 8723 8724 8725

	for_each_possible_cpu(i) {
		struct rt_rq *rt_rq = tg->rt_rq[i];

8726
		raw_spin_lock(&rt_rq->rt_runtime_lock);
P
Peter Zijlstra 已提交
8727
		rt_rq->rt_runtime = rt_runtime;
8728
		raw_spin_unlock(&rt_rq->rt_runtime_lock);
P
Peter Zijlstra 已提交
8729
	}
8730
	raw_spin_unlock_irq(&tg->rt_bandwidth.rt_runtime_lock);
P
Peter Zijlstra 已提交
8731
unlock:
8732
	read_unlock(&tasklist_lock);
P
Peter Zijlstra 已提交
8733 8734 8735
	mutex_unlock(&rt_constraints_mutex);

	return err;
P
Peter Zijlstra 已提交
8736 8737
}

8738 8739 8740 8741 8742 8743 8744 8745 8746 8747 8748 8749
int sched_group_set_rt_runtime(struct task_group *tg, long rt_runtime_us)
{
	u64 rt_runtime, rt_period;

	rt_period = ktime_to_ns(tg->rt_bandwidth.rt_period);
	rt_runtime = (u64)rt_runtime_us * NSEC_PER_USEC;
	if (rt_runtime_us < 0)
		rt_runtime = RUNTIME_INF;

	return tg_set_bandwidth(tg, rt_period, rt_runtime);
}

P
Peter Zijlstra 已提交
8750 8751 8752 8753
long sched_group_rt_runtime(struct task_group *tg)
{
	u64 rt_runtime_us;

8754
	if (tg->rt_bandwidth.rt_runtime == RUNTIME_INF)
P
Peter Zijlstra 已提交
8755 8756
		return -1;

8757
	rt_runtime_us = tg->rt_bandwidth.rt_runtime;
P
Peter Zijlstra 已提交
8758 8759 8760
	do_div(rt_runtime_us, NSEC_PER_USEC);
	return rt_runtime_us;
}
8761 8762 8763 8764 8765 8766 8767 8768

int sched_group_set_rt_period(struct task_group *tg, long rt_period_us)
{
	u64 rt_runtime, rt_period;

	rt_period = (u64)rt_period_us * NSEC_PER_USEC;
	rt_runtime = tg->rt_bandwidth.rt_runtime;

8769 8770 8771
	if (rt_period == 0)
		return -EINVAL;

8772 8773 8774 8775 8776 8777 8778 8779 8780 8781 8782 8783 8784 8785
	return tg_set_bandwidth(tg, rt_period, rt_runtime);
}

long sched_group_rt_period(struct task_group *tg)
{
	u64 rt_period_us;

	rt_period_us = ktime_to_ns(tg->rt_bandwidth.rt_period);
	do_div(rt_period_us, NSEC_PER_USEC);
	return rt_period_us;
}

static int sched_rt_global_constraints(void)
{
8786
	u64 runtime, period;
8787 8788
	int ret = 0;

8789 8790 8791
	if (sysctl_sched_rt_period <= 0)
		return -EINVAL;

8792 8793 8794 8795 8796 8797 8798 8799
	runtime = global_rt_runtime();
	period = global_rt_period();

	/*
	 * Sanity check on the sysctl variables.
	 */
	if (runtime > period && runtime != RUNTIME_INF)
		return -EINVAL;
8800

8801
	mutex_lock(&rt_constraints_mutex);
P
Peter Zijlstra 已提交
8802
	read_lock(&tasklist_lock);
8803
	ret = __rt_schedulable(NULL, 0, 0);
P
Peter Zijlstra 已提交
8804
	read_unlock(&tasklist_lock);
8805 8806 8807 8808
	mutex_unlock(&rt_constraints_mutex);

	return ret;
}
8809 8810 8811 8812 8813 8814 8815 8816 8817 8818

int sched_rt_can_attach(struct task_group *tg, struct task_struct *tsk)
{
	/* Don't accept realtime tasks when there is no way for them to run */
	if (rt_task(tsk) && tg->rt_bandwidth.rt_runtime == 0)
		return 0;

	return 1;
}

8819
#else /* !CONFIG_RT_GROUP_SCHED */
8820 8821
static int sched_rt_global_constraints(void)
{
P
Peter Zijlstra 已提交
8822 8823 8824
	unsigned long flags;
	int i;

8825 8826 8827
	if (sysctl_sched_rt_period <= 0)
		return -EINVAL;

8828 8829 8830 8831 8832 8833 8834
	/*
	 * There's always some RT tasks in the root group
	 * -- migration, kstopmachine etc..
	 */
	if (sysctl_sched_rt_runtime == 0)
		return -EBUSY;

8835
	raw_spin_lock_irqsave(&def_rt_bandwidth.rt_runtime_lock, flags);
P
Peter Zijlstra 已提交
8836 8837 8838
	for_each_possible_cpu(i) {
		struct rt_rq *rt_rq = &cpu_rq(i)->rt;

8839
		raw_spin_lock(&rt_rq->rt_runtime_lock);
P
Peter Zijlstra 已提交
8840
		rt_rq->rt_runtime = global_rt_runtime();
8841
		raw_spin_unlock(&rt_rq->rt_runtime_lock);
P
Peter Zijlstra 已提交
8842
	}
8843
	raw_spin_unlock_irqrestore(&def_rt_bandwidth.rt_runtime_lock, flags);
P
Peter Zijlstra 已提交
8844

8845 8846
	return 0;
}
8847
#endif /* CONFIG_RT_GROUP_SCHED */
8848 8849

int sched_rt_handler(struct ctl_table *table, int write,
8850
		void __user *buffer, size_t *lenp,
8851 8852 8853 8854 8855 8856 8857 8858 8859 8860
		loff_t *ppos)
{
	int ret;
	int old_period, old_runtime;
	static DEFINE_MUTEX(mutex);

	mutex_lock(&mutex);
	old_period = sysctl_sched_rt_period;
	old_runtime = sysctl_sched_rt_runtime;

8861
	ret = proc_dointvec(table, write, buffer, lenp, ppos);
8862 8863 8864 8865 8866 8867 8868 8869 8870 8871 8872 8873 8874 8875 8876 8877

	if (!ret && write) {
		ret = sched_rt_global_constraints();
		if (ret) {
			sysctl_sched_rt_period = old_period;
			sysctl_sched_rt_runtime = old_runtime;
		} else {
			def_rt_bandwidth.rt_runtime = global_rt_runtime();
			def_rt_bandwidth.rt_period =
				ns_to_ktime(global_rt_period());
		}
	}
	mutex_unlock(&mutex);

	return ret;
}
8878

8879
#ifdef CONFIG_CGROUP_SCHED
8880 8881

/* return corresponding task_group object of a cgroup */
8882
static inline struct task_group *cgroup_tg(struct cgroup *cgrp)
8883
{
8884 8885
	return container_of(cgroup_subsys_state(cgrp, cpu_cgroup_subsys_id),
			    struct task_group, css);
8886 8887 8888
}

static struct cgroup_subsys_state *
8889
cpu_cgroup_create(struct cgroup_subsys *ss, struct cgroup *cgrp)
8890
{
8891
	struct task_group *tg, *parent;
8892

8893
	if (!cgrp->parent) {
8894 8895 8896 8897
		/* This is early initialization for the top cgroup */
		return &init_task_group.css;
	}

8898 8899
	parent = cgroup_tg(cgrp->parent);
	tg = sched_create_group(parent);
8900 8901 8902 8903 8904 8905
	if (IS_ERR(tg))
		return ERR_PTR(-ENOMEM);

	return &tg->css;
}

I
Ingo Molnar 已提交
8906 8907
static void
cpu_cgroup_destroy(struct cgroup_subsys *ss, struct cgroup *cgrp)
8908
{
8909
	struct task_group *tg = cgroup_tg(cgrp);
8910 8911 8912 8913

	sched_destroy_group(tg);
}

I
Ingo Molnar 已提交
8914
static int
8915
cpu_cgroup_can_attach_task(struct cgroup *cgrp, struct task_struct *tsk)
8916
{
8917
#ifdef CONFIG_RT_GROUP_SCHED
8918
	if (!sched_rt_can_attach(cgroup_tg(cgrp), tsk))
8919 8920
		return -EINVAL;
#else
8921 8922 8923
	/* We don't support RT-tasks being in separate groups */
	if (tsk->sched_class != &fair_sched_class)
		return -EINVAL;
8924
#endif
8925 8926
	return 0;
}
8927

8928 8929 8930 8931 8932 8933 8934 8935 8936 8937 8938 8939 8940 8941 8942 8943 8944 8945 8946
static int
cpu_cgroup_can_attach(struct cgroup_subsys *ss, struct cgroup *cgrp,
		      struct task_struct *tsk, bool threadgroup)
{
	int retval = cpu_cgroup_can_attach_task(cgrp, tsk);
	if (retval)
		return retval;
	if (threadgroup) {
		struct task_struct *c;
		rcu_read_lock();
		list_for_each_entry_rcu(c, &tsk->thread_group, thread_group) {
			retval = cpu_cgroup_can_attach_task(cgrp, c);
			if (retval) {
				rcu_read_unlock();
				return retval;
			}
		}
		rcu_read_unlock();
	}
8947 8948 8949 8950
	return 0;
}

static void
8951
cpu_cgroup_attach(struct cgroup_subsys *ss, struct cgroup *cgrp,
8952 8953
		  struct cgroup *old_cont, struct task_struct *tsk,
		  bool threadgroup)
8954 8955
{
	sched_move_task(tsk);
8956 8957 8958 8959 8960 8961 8962 8963
	if (threadgroup) {
		struct task_struct *c;
		rcu_read_lock();
		list_for_each_entry_rcu(c, &tsk->thread_group, thread_group) {
			sched_move_task(c);
		}
		rcu_read_unlock();
	}
8964 8965
}

8966
#ifdef CONFIG_FAIR_GROUP_SCHED
8967
static int cpu_shares_write_u64(struct cgroup *cgrp, struct cftype *cftype,
8968
				u64 shareval)
8969
{
8970
	return sched_group_set_shares(cgroup_tg(cgrp), shareval);
8971 8972
}

8973
static u64 cpu_shares_read_u64(struct cgroup *cgrp, struct cftype *cft)
8974
{
8975
	struct task_group *tg = cgroup_tg(cgrp);
8976 8977 8978

	return (u64) tg->shares;
}
8979
#endif /* CONFIG_FAIR_GROUP_SCHED */
8980

8981
#ifdef CONFIG_RT_GROUP_SCHED
M
Mirco Tischler 已提交
8982
static int cpu_rt_runtime_write(struct cgroup *cgrp, struct cftype *cft,
8983
				s64 val)
P
Peter Zijlstra 已提交
8984
{
8985
	return sched_group_set_rt_runtime(cgroup_tg(cgrp), val);
P
Peter Zijlstra 已提交
8986 8987
}

8988
static s64 cpu_rt_runtime_read(struct cgroup *cgrp, struct cftype *cft)
P
Peter Zijlstra 已提交
8989
{
8990
	return sched_group_rt_runtime(cgroup_tg(cgrp));
P
Peter Zijlstra 已提交
8991
}
8992 8993 8994 8995 8996 8997 8998 8999 9000 9001 9002

static int cpu_rt_period_write_uint(struct cgroup *cgrp, struct cftype *cftype,
		u64 rt_period_us)
{
	return sched_group_set_rt_period(cgroup_tg(cgrp), rt_period_us);
}

static u64 cpu_rt_period_read_uint(struct cgroup *cgrp, struct cftype *cft)
{
	return sched_group_rt_period(cgroup_tg(cgrp));
}
9003
#endif /* CONFIG_RT_GROUP_SCHED */
P
Peter Zijlstra 已提交
9004

9005
static struct cftype cpu_files[] = {
9006
#ifdef CONFIG_FAIR_GROUP_SCHED
9007 9008
	{
		.name = "shares",
9009 9010
		.read_u64 = cpu_shares_read_u64,
		.write_u64 = cpu_shares_write_u64,
9011
	},
9012 9013
#endif
#ifdef CONFIG_RT_GROUP_SCHED
P
Peter Zijlstra 已提交
9014
	{
P
Peter Zijlstra 已提交
9015
		.name = "rt_runtime_us",
9016 9017
		.read_s64 = cpu_rt_runtime_read,
		.write_s64 = cpu_rt_runtime_write,
P
Peter Zijlstra 已提交
9018
	},
9019 9020
	{
		.name = "rt_period_us",
9021 9022
		.read_u64 = cpu_rt_period_read_uint,
		.write_u64 = cpu_rt_period_write_uint,
9023
	},
9024
#endif
9025 9026 9027 9028
};

static int cpu_cgroup_populate(struct cgroup_subsys *ss, struct cgroup *cont)
{
9029
	return cgroup_add_files(cont, ss, cpu_files, ARRAY_SIZE(cpu_files));
9030 9031 9032
}

struct cgroup_subsys cpu_cgroup_subsys = {
I
Ingo Molnar 已提交
9033 9034 9035 9036 9037 9038 9039
	.name		= "cpu",
	.create		= cpu_cgroup_create,
	.destroy	= cpu_cgroup_destroy,
	.can_attach	= cpu_cgroup_can_attach,
	.attach		= cpu_cgroup_attach,
	.populate	= cpu_cgroup_populate,
	.subsys_id	= cpu_cgroup_subsys_id,
9040 9041 9042
	.early_init	= 1,
};

9043
#endif	/* CONFIG_CGROUP_SCHED */
9044 9045 9046 9047 9048 9049 9050 9051 9052 9053

#ifdef CONFIG_CGROUP_CPUACCT

/*
 * CPU accounting code for task groups.
 *
 * Based on the work by Paul Menage (menage@google.com) and Balbir Singh
 * (balbir@in.ibm.com).
 */

9054
/* track cpu usage of a group of tasks and its child groups */
9055 9056 9057
struct cpuacct {
	struct cgroup_subsys_state css;
	/* cpuusage holds pointer to a u64-type object on every cpu */
9058
	u64 __percpu *cpuusage;
9059
	struct percpu_counter cpustat[CPUACCT_STAT_NSTATS];
9060
	struct cpuacct *parent;
9061 9062 9063 9064 9065
};

struct cgroup_subsys cpuacct_subsys;

/* return cpu accounting group corresponding to this container */
9066
static inline struct cpuacct *cgroup_ca(struct cgroup *cgrp)
9067
{
9068
	return container_of(cgroup_subsys_state(cgrp, cpuacct_subsys_id),
9069 9070 9071 9072 9073 9074 9075 9076 9077 9078 9079 9080
			    struct cpuacct, css);
}

/* return cpu accounting group to which this task belongs */
static inline struct cpuacct *task_ca(struct task_struct *tsk)
{
	return container_of(task_subsys_state(tsk, cpuacct_subsys_id),
			    struct cpuacct, css);
}

/* create a new cpu accounting group */
static struct cgroup_subsys_state *cpuacct_create(
9081
	struct cgroup_subsys *ss, struct cgroup *cgrp)
9082 9083
{
	struct cpuacct *ca = kzalloc(sizeof(*ca), GFP_KERNEL);
9084
	int i;
9085 9086

	if (!ca)
9087
		goto out;
9088 9089

	ca->cpuusage = alloc_percpu(u64);
9090 9091 9092 9093 9094 9095
	if (!ca->cpuusage)
		goto out_free_ca;

	for (i = 0; i < CPUACCT_STAT_NSTATS; i++)
		if (percpu_counter_init(&ca->cpustat[i], 0))
			goto out_free_counters;
9096

9097 9098 9099
	if (cgrp->parent)
		ca->parent = cgroup_ca(cgrp->parent);

9100
	return &ca->css;
9101 9102 9103 9104 9105 9106 9107 9108 9109

out_free_counters:
	while (--i >= 0)
		percpu_counter_destroy(&ca->cpustat[i]);
	free_percpu(ca->cpuusage);
out_free_ca:
	kfree(ca);
out:
	return ERR_PTR(-ENOMEM);
9110 9111 9112
}

/* destroy an existing cpu accounting group */
I
Ingo Molnar 已提交
9113
static void
9114
cpuacct_destroy(struct cgroup_subsys *ss, struct cgroup *cgrp)
9115
{
9116
	struct cpuacct *ca = cgroup_ca(cgrp);
9117
	int i;
9118

9119 9120
	for (i = 0; i < CPUACCT_STAT_NSTATS; i++)
		percpu_counter_destroy(&ca->cpustat[i]);
9121 9122 9123 9124
	free_percpu(ca->cpuusage);
	kfree(ca);
}

9125 9126
static u64 cpuacct_cpuusage_read(struct cpuacct *ca, int cpu)
{
9127
	u64 *cpuusage = per_cpu_ptr(ca->cpuusage, cpu);
9128 9129 9130 9131 9132 9133
	u64 data;

#ifndef CONFIG_64BIT
	/*
	 * Take rq->lock to make 64-bit read safe on 32-bit platforms.
	 */
9134
	raw_spin_lock_irq(&cpu_rq(cpu)->lock);
9135
	data = *cpuusage;
9136
	raw_spin_unlock_irq(&cpu_rq(cpu)->lock);
9137 9138 9139 9140 9141 9142 9143 9144 9145
#else
	data = *cpuusage;
#endif

	return data;
}

static void cpuacct_cpuusage_write(struct cpuacct *ca, int cpu, u64 val)
{
9146
	u64 *cpuusage = per_cpu_ptr(ca->cpuusage, cpu);
9147 9148 9149 9150 9151

#ifndef CONFIG_64BIT
	/*
	 * Take rq->lock to make 64-bit write safe on 32-bit platforms.
	 */
9152
	raw_spin_lock_irq(&cpu_rq(cpu)->lock);
9153
	*cpuusage = val;
9154
	raw_spin_unlock_irq(&cpu_rq(cpu)->lock);
9155 9156 9157 9158 9159
#else
	*cpuusage = val;
#endif
}

9160
/* return total cpu usage (in nanoseconds) of a group */
9161
static u64 cpuusage_read(struct cgroup *cgrp, struct cftype *cft)
9162
{
9163
	struct cpuacct *ca = cgroup_ca(cgrp);
9164 9165 9166
	u64 totalcpuusage = 0;
	int i;

9167 9168
	for_each_present_cpu(i)
		totalcpuusage += cpuacct_cpuusage_read(ca, i);
9169 9170 9171 9172

	return totalcpuusage;
}

9173 9174 9175 9176 9177 9178 9179 9180 9181 9182 9183 9184
static int cpuusage_write(struct cgroup *cgrp, struct cftype *cftype,
								u64 reset)
{
	struct cpuacct *ca = cgroup_ca(cgrp);
	int err = 0;
	int i;

	if (reset) {
		err = -EINVAL;
		goto out;
	}

9185 9186
	for_each_present_cpu(i)
		cpuacct_cpuusage_write(ca, i, 0);
9187 9188 9189 9190 9191

out:
	return err;
}

9192 9193 9194 9195 9196 9197 9198 9199 9200 9201 9202 9203 9204 9205 9206
static int cpuacct_percpu_seq_read(struct cgroup *cgroup, struct cftype *cft,
				   struct seq_file *m)
{
	struct cpuacct *ca = cgroup_ca(cgroup);
	u64 percpu;
	int i;

	for_each_present_cpu(i) {
		percpu = cpuacct_cpuusage_read(ca, i);
		seq_printf(m, "%llu ", (unsigned long long) percpu);
	}
	seq_printf(m, "\n");
	return 0;
}

9207 9208 9209 9210 9211 9212 9213 9214 9215 9216 9217 9218 9219 9220 9221 9222 9223 9224 9225
static const char *cpuacct_stat_desc[] = {
	[CPUACCT_STAT_USER] = "user",
	[CPUACCT_STAT_SYSTEM] = "system",
};

static int cpuacct_stats_show(struct cgroup *cgrp, struct cftype *cft,
		struct cgroup_map_cb *cb)
{
	struct cpuacct *ca = cgroup_ca(cgrp);
	int i;

	for (i = 0; i < CPUACCT_STAT_NSTATS; i++) {
		s64 val = percpu_counter_read(&ca->cpustat[i]);
		val = cputime64_to_clock_t(val);
		cb->fill(cb, cpuacct_stat_desc[i], val);
	}
	return 0;
}

9226 9227 9228
static struct cftype files[] = {
	{
		.name = "usage",
9229 9230
		.read_u64 = cpuusage_read,
		.write_u64 = cpuusage_write,
9231
	},
9232 9233 9234 9235
	{
		.name = "usage_percpu",
		.read_seq_string = cpuacct_percpu_seq_read,
	},
9236 9237 9238 9239
	{
		.name = "stat",
		.read_map = cpuacct_stats_show,
	},
9240 9241
};

9242
static int cpuacct_populate(struct cgroup_subsys *ss, struct cgroup *cgrp)
9243
{
9244
	return cgroup_add_files(cgrp, ss, files, ARRAY_SIZE(files));
9245 9246 9247 9248 9249 9250 9251 9252 9253 9254
}

/*
 * charge this task's execution time to its accounting group.
 *
 * called with rq->lock held.
 */
static void cpuacct_charge(struct task_struct *tsk, u64 cputime)
{
	struct cpuacct *ca;
9255
	int cpu;
9256

L
Li Zefan 已提交
9257
	if (unlikely(!cpuacct_subsys.active))
9258 9259
		return;

9260
	cpu = task_cpu(tsk);
9261 9262 9263

	rcu_read_lock();

9264 9265
	ca = task_ca(tsk);

9266
	for (; ca; ca = ca->parent) {
9267
		u64 *cpuusage = per_cpu_ptr(ca->cpuusage, cpu);
9268 9269
		*cpuusage += cputime;
	}
9270 9271

	rcu_read_unlock();
9272 9273
}

9274 9275 9276 9277 9278 9279 9280 9281 9282 9283 9284 9285 9286 9287 9288 9289 9290
/*
 * When CONFIG_VIRT_CPU_ACCOUNTING is enabled one jiffy can be very large
 * in cputime_t units. As a result, cpuacct_update_stats calls
 * percpu_counter_add with values large enough to always overflow the
 * per cpu batch limit causing bad SMP scalability.
 *
 * To fix this we scale percpu_counter_batch by cputime_one_jiffy so we
 * batch the same amount of time with CONFIG_VIRT_CPU_ACCOUNTING disabled
 * and enabled. We cap it at INT_MAX which is the largest allowed batch value.
 */
#ifdef CONFIG_SMP
#define CPUACCT_BATCH	\
	min_t(long, percpu_counter_batch * cputime_one_jiffy, INT_MAX)
#else
#define CPUACCT_BATCH	0
#endif

9291 9292 9293 9294 9295 9296 9297
/*
 * Charge the system/user time to the task's accounting group.
 */
static void cpuacct_update_stats(struct task_struct *tsk,
		enum cpuacct_stat_index idx, cputime_t val)
{
	struct cpuacct *ca;
9298
	int batch = CPUACCT_BATCH;
9299 9300 9301 9302 9303 9304 9305 9306

	if (unlikely(!cpuacct_subsys.active))
		return;

	rcu_read_lock();
	ca = task_ca(tsk);

	do {
9307
		__percpu_counter_add(&ca->cpustat[idx], val, batch);
9308 9309 9310 9311 9312
		ca = ca->parent;
	} while (ca);
	rcu_read_unlock();
}

9313 9314 9315 9316 9317 9318 9319 9320
struct cgroup_subsys cpuacct_subsys = {
	.name = "cpuacct",
	.create = cpuacct_create,
	.destroy = cpuacct_destroy,
	.populate = cpuacct_populate,
	.subsys_id = cpuacct_subsys_id,
};
#endif	/* CONFIG_CGROUP_CPUACCT */
9321 9322 9323 9324 9325

#ifndef CONFIG_SMP

void synchronize_sched_expedited(void)
{
9326
	barrier();
9327 9328 9329 9330 9331
}
EXPORT_SYMBOL_GPL(synchronize_sched_expedited);

#else /* #ifndef CONFIG_SMP */

9332
static atomic_t synchronize_sched_expedited_count = ATOMIC_INIT(0);
9333

9334
static int synchronize_sched_expedited_cpu_stop(void *data)
9335
{
9336 9337 9338 9339 9340 9341 9342 9343 9344 9345 9346
	/*
	 * There must be a full memory barrier on each affected CPU
	 * between the time that try_stop_cpus() is called and the
	 * time that it returns.
	 *
	 * In the current initial implementation of cpu_stop, the
	 * above condition is already met when the control reaches
	 * this point and the following smp_mb() is not strictly
	 * necessary.  Do smp_mb() anyway for documentation and
	 * robustness against future implementation changes.
	 */
9347
	smp_mb(); /* See above comment block. */
9348
	return 0;
9349 9350 9351 9352 9353 9354 9355 9356 9357 9358 9359 9360 9361 9362
}

/*
 * Wait for an rcu-sched grace period to elapse, but use "big hammer"
 * approach to force grace period to end quickly.  This consumes
 * significant time on all CPUs, and is thus not recommended for
 * any sort of common-case code.
 *
 * Note that it is illegal to call this function while holding any
 * lock that is acquired by a CPU-hotplug notifier.  Failing to
 * observe this restriction will result in deadlock.
 */
void synchronize_sched_expedited(void)
{
9363
	int snap, trycount = 0;
9364 9365

	smp_mb();  /* ensure prior mod happens before capturing snap. */
9366
	snap = atomic_read(&synchronize_sched_expedited_count) + 1;
9367
	get_online_cpus();
9368 9369
	while (try_stop_cpus(cpu_online_mask,
			     synchronize_sched_expedited_cpu_stop,
9370
			     NULL) == -EAGAIN) {
9371 9372 9373 9374 9375 9376 9377
		put_online_cpus();
		if (trycount++ < 10)
			udelay(trycount * num_online_cpus());
		else {
			synchronize_sched();
			return;
		}
9378
		if (atomic_read(&synchronize_sched_expedited_count) - snap > 0) {
9379 9380 9381 9382 9383
			smp_mb(); /* ensure test happens before caller kfree */
			return;
		}
		get_online_cpus();
	}
9384
	atomic_inc(&synchronize_sched_expedited_count);
9385
	smp_mb__after_atomic_inc(); /* ensure post-GP actions seen after GP. */
9386 9387 9388 9389 9390
	put_online_cpus();
}
EXPORT_SYMBOL_GPL(synchronize_sched_expedited);

#endif /* #else #ifndef CONFIG_SMP */