workqueue.c 102.2 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
T
Tejun Heo 已提交
2
 * kernel/workqueue.c - generic async execution with shared worker pool
L
Linus Torvalds 已提交
3
 *
T
Tejun Heo 已提交
4
 * Copyright (C) 2002		Ingo Molnar
L
Linus Torvalds 已提交
5
 *
T
Tejun Heo 已提交
6 7 8 9 10
 *   Derived from the taskqueue/keventd code by:
 *     David Woodhouse <dwmw2@infradead.org>
 *     Andrew Morton
 *     Kai Petzke <wpp@marie.physik.tu-berlin.de>
 *     Theodore Ts'o <tytso@mit.edu>
L
Linus Torvalds 已提交
11
 *
T
Tejun Heo 已提交
12
 * Made to use alloc_percpu by Christoph Lameter.
L
Linus Torvalds 已提交
13
 *
T
Tejun Heo 已提交
14 15
 * Copyright (C) 2010		SUSE Linux Products GmbH
 * Copyright (C) 2010		Tejun Heo <tj@kernel.org>
16
 *
T
Tejun Heo 已提交
17 18 19 20 21 22 23
 * This is the generic async execution mechanism.  Work items as are
 * executed in process context.  The worker pool is shared and
 * automatically managed.  There is one worker pool for each CPU and
 * one extra for works which are better served by workers which are
 * not bound to any specific CPU.
 *
 * Please read Documentation/workqueue.txt for details.
L
Linus Torvalds 已提交
24 25 26 27 28 29 30 31 32 33 34 35 36
 */

#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/init.h>
#include <linux/signal.h>
#include <linux/completion.h>
#include <linux/workqueue.h>
#include <linux/slab.h>
#include <linux/cpu.h>
#include <linux/notifier.h>
#include <linux/kthread.h>
37
#include <linux/hardirq.h>
38
#include <linux/mempolicy.h>
39
#include <linux/freezer.h>
40 41
#include <linux/kallsyms.h>
#include <linux/debug_locks.h>
42
#include <linux/lockdep.h>
T
Tejun Heo 已提交
43
#include <linux/idr.h>
44 45

#include "workqueue_sched.h"
L
Linus Torvalds 已提交
46

T
Tejun Heo 已提交
47
enum {
48
	/* global_cwq flags */
49 50 51
	GCWQ_MANAGE_WORKERS	= 1 << 0,	/* need to manage workers */
	GCWQ_MANAGING_WORKERS	= 1 << 1,	/* managing workers */
	GCWQ_DISASSOCIATED	= 1 << 2,	/* cpu can't serve workers */
52
	GCWQ_FREEZING		= 1 << 3,	/* freeze in progress */
53
	GCWQ_HIGHPRI_PENDING	= 1 << 4,	/* highpri works on queue */
54

T
Tejun Heo 已提交
55 56 57 58
	/* worker flags */
	WORKER_STARTED		= 1 << 0,	/* started */
	WORKER_DIE		= 1 << 1,	/* die die die */
	WORKER_IDLE		= 1 << 2,	/* is idle */
59
	WORKER_PREP		= 1 << 3,	/* preparing to run works */
60
	WORKER_ROGUE		= 1 << 4,	/* not bound to any cpu */
61
	WORKER_REBIND		= 1 << 5,	/* mom is home, come back */
62
	WORKER_CPU_INTENSIVE	= 1 << 6,	/* cpu intensive */
63
	WORKER_UNBOUND		= 1 << 7,	/* worker is unbound */
64

65
	WORKER_NOT_RUNNING	= WORKER_PREP | WORKER_ROGUE | WORKER_REBIND |
66
				  WORKER_CPU_INTENSIVE | WORKER_UNBOUND,
67 68 69 70 71 72 73

	/* gcwq->trustee_state */
	TRUSTEE_START		= 0,		/* start */
	TRUSTEE_IN_CHARGE	= 1,		/* trustee in charge of gcwq */
	TRUSTEE_BUTCHER		= 2,		/* butcher workers */
	TRUSTEE_RELEASE		= 3,		/* release workers */
	TRUSTEE_DONE		= 4,		/* trustee is done */
T
Tejun Heo 已提交
74 75 76 77

	BUSY_WORKER_HASH_ORDER	= 6,		/* 64 pointers */
	BUSY_WORKER_HASH_SIZE	= 1 << BUSY_WORKER_HASH_ORDER,
	BUSY_WORKER_HASH_MASK	= BUSY_WORKER_HASH_SIZE - 1,
78

79 80 81 82 83 84
	MAX_IDLE_WORKERS_RATIO	= 4,		/* 1/4 of busy can be idle */
	IDLE_WORKER_TIMEOUT	= 300 * HZ,	/* keep idle ones for 5 mins */

	MAYDAY_INITIAL_TIMEOUT	= HZ / 100,	/* call for help after 10ms */
	MAYDAY_INTERVAL		= HZ / 10,	/* and then every 100ms */
	CREATE_COOLDOWN		= HZ,		/* time to breath after fail */
85
	TRUSTEE_COOLDOWN	= HZ / 10,	/* for trustee draining */
86 87 88 89 90 91

	/*
	 * Rescue workers are used only on emergencies and shared by
	 * all cpus.  Give -20.
	 */
	RESCUER_NICE_LEVEL	= -20,
T
Tejun Heo 已提交
92
};
L
Linus Torvalds 已提交
93 94

/*
T
Tejun Heo 已提交
95 96
 * Structure fields follow one of the following exclusion rules.
 *
97 98
 * I: Modifiable by initialization/destruction paths and read-only for
 *    everyone else.
T
Tejun Heo 已提交
99
 *
100 101 102
 * P: Preemption protected.  Disabling preemption is enough and should
 *    only be modified and accessed from the local cpu.
 *
103
 * L: gcwq->lock protected.  Access with gcwq->lock held.
T
Tejun Heo 已提交
104
 *
105 106 107
 * X: During normal operation, modification requires gcwq->lock and
 *    should be done only from local cpu.  Either disabling preemption
 *    on local cpu or grabbing gcwq->lock is enough for read access.
108
 *    If GCWQ_DISASSOCIATED is set, it's identical to L.
109
 *
110 111
 * F: wq->flush_mutex protected.
 *
T
Tejun Heo 已提交
112
 * W: workqueue_lock protected.
L
Linus Torvalds 已提交
113 114
 */

115
struct global_cwq;
L
Linus Torvalds 已提交
116

117 118 119 120
/*
 * The poor guys doing the actual heavy lifting.  All on-duty workers
 * are either serving the manager role, on idle list or on busy hash.
 */
T
Tejun Heo 已提交
121
struct worker {
T
Tejun Heo 已提交
122 123 124 125 126
	/* on idle list while idle, on busy hash table while busy */
	union {
		struct list_head	entry;	/* L: while idle */
		struct hlist_node	hentry;	/* L: while busy */
	};
L
Linus Torvalds 已提交
127

T
Tejun Heo 已提交
128
	struct work_struct	*current_work;	/* L: work being processed */
129
	struct cpu_workqueue_struct *current_cwq; /* L: current_work's cwq */
130
	struct list_head	scheduled;	/* L: scheduled works */
T
Tejun Heo 已提交
131
	struct task_struct	*task;		/* I: worker task */
132
	struct global_cwq	*gcwq;		/* I: the associated gcwq */
133 134 135
	/* 64 bytes boundary on 64bit, 32 on 32bit */
	unsigned long		last_active;	/* L: last active timestamp */
	unsigned int		flags;		/* X: flags */
T
Tejun Heo 已提交
136
	int			id;		/* I: worker id */
137
	struct work_struct	rebind_work;	/* L: rebind worker to cpu */
T
Tejun Heo 已提交
138 139
};

140
/*
141 142 143
 * Global per-cpu workqueue.  There's one and only one for each cpu
 * and all works are queued and processed here regardless of their
 * target workqueues.
144 145 146
 */
struct global_cwq {
	spinlock_t		lock;		/* the gcwq lock */
147
	struct list_head	worklist;	/* L: list of pending works */
148
	unsigned int		cpu;		/* I: the associated cpu */
149
	unsigned int		flags;		/* L: GCWQ_* flags */
T
Tejun Heo 已提交
150 151 152 153 154

	int			nr_workers;	/* L: total number of workers */
	int			nr_idle;	/* L: currently idle ones */

	/* workers are chained either in the idle_list or busy_hash */
155
	struct list_head	idle_list;	/* X: list of idle workers */
T
Tejun Heo 已提交
156 157 158
	struct hlist_head	busy_hash[BUSY_WORKER_HASH_SIZE];
						/* L: hash of busy workers */

159 160 161
	struct timer_list	idle_timer;	/* L: worker idle timeout */
	struct timer_list	mayday_timer;	/* L: SOS timer for dworkers */

162
	struct ida		worker_ida;	/* L: for worker IDs */
163 164 165 166

	struct task_struct	*trustee;	/* L: for gcwq shutdown */
	unsigned int		trustee_state;	/* L: trustee state */
	wait_queue_head_t	trustee_wait;	/* trustee wait */
167
	struct worker		*first_idle;	/* L: first idle worker */
168 169
} ____cacheline_aligned_in_smp;

L
Linus Torvalds 已提交
170
/*
171
 * The per-CPU workqueue.  The lower WORK_STRUCT_FLAG_BITS of
T
Tejun Heo 已提交
172 173
 * work_struct->data are used for flags and thus cwqs need to be
 * aligned at two's power of the number of flag bits.
L
Linus Torvalds 已提交
174 175
 */
struct cpu_workqueue_struct {
176
	struct global_cwq	*gcwq;		/* I: the associated gcwq */
T
Tejun Heo 已提交
177
	struct workqueue_struct *wq;		/* I: the owning workqueue */
178 179 180 181
	int			work_color;	/* L: current color */
	int			flush_color;	/* L: flushing color */
	int			nr_in_flight[WORK_NR_COLORS];
						/* L: nr of in_flight works */
182
	int			nr_active;	/* L: nr of active works */
183
	int			max_active;	/* L: max active works */
184
	struct list_head	delayed_works;	/* L: delayed works */
T
Tejun Heo 已提交
185
};
L
Linus Torvalds 已提交
186

187 188 189 190 191 192 193 194 195
/*
 * Structure used to wait for workqueue flush.
 */
struct wq_flusher {
	struct list_head	list;		/* F: list of flushers */
	int			flush_color;	/* F: flush color waiting for */
	struct completion	done;		/* flush completion */
};

196 197 198 199 200 201 202 203 204 205
/*
 * All cpumasks are assumed to be always set on UP and thus can't be
 * used to determine whether there's something to be done.
 */
#ifdef CONFIG_SMP
typedef cpumask_var_t mayday_mask_t;
#define mayday_test_and_set_cpu(cpu, mask)	\
	cpumask_test_and_set_cpu((cpu), (mask))
#define mayday_clear_cpu(cpu, mask)		cpumask_clear_cpu((cpu), (mask))
#define for_each_mayday_cpu(cpu, mask)		for_each_cpu((cpu), (mask))
206
#define alloc_mayday_mask(maskp, gfp)		zalloc_cpumask_var((maskp), (gfp))
207 208 209 210 211 212 213 214 215
#define free_mayday_mask(mask)			free_cpumask_var((mask))
#else
typedef unsigned long mayday_mask_t;
#define mayday_test_and_set_cpu(cpu, mask)	test_and_set_bit(0, &(mask))
#define mayday_clear_cpu(cpu, mask)		clear_bit(0, &(mask))
#define for_each_mayday_cpu(cpu, mask)		if ((cpu) = 0, (mask))
#define alloc_mayday_mask(maskp, gfp)		true
#define free_mayday_mask(mask)			do { } while (0)
#endif
L
Linus Torvalds 已提交
216 217 218 219 220 221

/*
 * The externally visible workqueue abstraction is an array of
 * per-CPU workqueues:
 */
struct workqueue_struct {
222
	unsigned int		flags;		/* I: WQ_* flags */
223 224 225 226 227
	union {
		struct cpu_workqueue_struct __percpu	*pcpu;
		struct cpu_workqueue_struct		*single;
		unsigned long				v;
	} cpu_wq;				/* I: cwq's */
T
Tejun Heo 已提交
228
	struct list_head	list;		/* W: list of all workqueues */
229 230 231 232 233 234 235 236 237

	struct mutex		flush_mutex;	/* protects wq flushing */
	int			work_color;	/* F: current work color */
	int			flush_color;	/* F: current flush color */
	atomic_t		nr_cwqs_to_flush; /* flush in progress */
	struct wq_flusher	*first_flusher;	/* F: first flusher */
	struct list_head	flusher_queue;	/* F: flush waiters */
	struct list_head	flusher_overflow; /* F: flush overflow list */

238
	mayday_mask_t		mayday_mask;	/* cpus requesting rescue */
239 240
	struct worker		*rescuer;	/* I: rescue worker */

241
	int			saved_max_active; /* W: saved cwq max_active */
T
Tejun Heo 已提交
242
	const char		*name;		/* I: workqueue name */
243
#ifdef CONFIG_LOCKDEP
T
Tejun Heo 已提交
244
	struct lockdep_map	lockdep_map;
245
#endif
L
Linus Torvalds 已提交
246 247
};

248 249 250
struct workqueue_struct *system_wq __read_mostly;
struct workqueue_struct *system_long_wq __read_mostly;
struct workqueue_struct *system_nrt_wq __read_mostly;
251
struct workqueue_struct *system_unbound_wq __read_mostly;
252 253 254
EXPORT_SYMBOL_GPL(system_wq);
EXPORT_SYMBOL_GPL(system_long_wq);
EXPORT_SYMBOL_GPL(system_nrt_wq);
255
EXPORT_SYMBOL_GPL(system_unbound_wq);
256

257 258 259
#define CREATE_TRACE_POINTS
#include <trace/events/workqueue.h>

260 261 262 263
#define for_each_busy_worker(worker, i, pos, gcwq)			\
	for (i = 0; i < BUSY_WORKER_HASH_SIZE; i++)			\
		hlist_for_each_entry(worker, pos, &gcwq->busy_hash[i], hentry)

264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284
static inline int __next_gcwq_cpu(int cpu, const struct cpumask *mask,
				  unsigned int sw)
{
	if (cpu < nr_cpu_ids) {
		if (sw & 1) {
			cpu = cpumask_next(cpu, mask);
			if (cpu < nr_cpu_ids)
				return cpu;
		}
		if (sw & 2)
			return WORK_CPU_UNBOUND;
	}
	return WORK_CPU_NONE;
}

static inline int __next_wq_cpu(int cpu, const struct cpumask *mask,
				struct workqueue_struct *wq)
{
	return __next_gcwq_cpu(cpu, mask, !(wq->flags & WQ_UNBOUND) ? 1 : 2);
}

285 286 287 288 289 290 291 292 293 294 295 296 297
/*
 * CPU iterators
 *
 * An extra gcwq is defined for an invalid cpu number
 * (WORK_CPU_UNBOUND) to host workqueues which are not bound to any
 * specific CPU.  The following iterators are similar to
 * for_each_*_cpu() iterators but also considers the unbound gcwq.
 *
 * for_each_gcwq_cpu()		: possible CPUs + WORK_CPU_UNBOUND
 * for_each_online_gcwq_cpu()	: online CPUs + WORK_CPU_UNBOUND
 * for_each_cwq_cpu()		: possible CPUs for bound workqueues,
 *				  WORK_CPU_UNBOUND for unbound workqueues
 */
298 299 300 301 302 303 304 305 306 307 308 309 310 311 312
#define for_each_gcwq_cpu(cpu)						\
	for ((cpu) = __next_gcwq_cpu(-1, cpu_possible_mask, 3);		\
	     (cpu) < WORK_CPU_NONE;					\
	     (cpu) = __next_gcwq_cpu((cpu), cpu_possible_mask, 3))

#define for_each_online_gcwq_cpu(cpu)					\
	for ((cpu) = __next_gcwq_cpu(-1, cpu_online_mask, 3);		\
	     (cpu) < WORK_CPU_NONE;					\
	     (cpu) = __next_gcwq_cpu((cpu), cpu_online_mask, 3))

#define for_each_cwq_cpu(cpu, wq)					\
	for ((cpu) = __next_wq_cpu(-1, cpu_possible_mask, (wq));	\
	     (cpu) < WORK_CPU_NONE;					\
	     (cpu) = __next_wq_cpu((cpu), cpu_possible_mask, (wq)))

313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351
#ifdef CONFIG_DEBUG_OBJECTS_WORK

static struct debug_obj_descr work_debug_descr;

/*
 * fixup_init is called when:
 * - an active object is initialized
 */
static int work_fixup_init(void *addr, enum debug_obj_state state)
{
	struct work_struct *work = addr;

	switch (state) {
	case ODEBUG_STATE_ACTIVE:
		cancel_work_sync(work);
		debug_object_init(work, &work_debug_descr);
		return 1;
	default:
		return 0;
	}
}

/*
 * fixup_activate is called when:
 * - an active object is activated
 * - an unknown object is activated (might be a statically initialized object)
 */
static int work_fixup_activate(void *addr, enum debug_obj_state state)
{
	struct work_struct *work = addr;

	switch (state) {

	case ODEBUG_STATE_NOTAVAILABLE:
		/*
		 * This is not really a fixup. The work struct was
		 * statically initialized. We just make sure that it
		 * is tracked in the object tracker.
		 */
352
		if (test_bit(WORK_STRUCT_STATIC_BIT, work_data_bits(work))) {
353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422
			debug_object_init(work, &work_debug_descr);
			debug_object_activate(work, &work_debug_descr);
			return 0;
		}
		WARN_ON_ONCE(1);
		return 0;

	case ODEBUG_STATE_ACTIVE:
		WARN_ON(1);

	default:
		return 0;
	}
}

/*
 * fixup_free is called when:
 * - an active object is freed
 */
static int work_fixup_free(void *addr, enum debug_obj_state state)
{
	struct work_struct *work = addr;

	switch (state) {
	case ODEBUG_STATE_ACTIVE:
		cancel_work_sync(work);
		debug_object_free(work, &work_debug_descr);
		return 1;
	default:
		return 0;
	}
}

static struct debug_obj_descr work_debug_descr = {
	.name		= "work_struct",
	.fixup_init	= work_fixup_init,
	.fixup_activate	= work_fixup_activate,
	.fixup_free	= work_fixup_free,
};

static inline void debug_work_activate(struct work_struct *work)
{
	debug_object_activate(work, &work_debug_descr);
}

static inline void debug_work_deactivate(struct work_struct *work)
{
	debug_object_deactivate(work, &work_debug_descr);
}

void __init_work(struct work_struct *work, int onstack)
{
	if (onstack)
		debug_object_init_on_stack(work, &work_debug_descr);
	else
		debug_object_init(work, &work_debug_descr);
}
EXPORT_SYMBOL_GPL(__init_work);

void destroy_work_on_stack(struct work_struct *work)
{
	debug_object_free(work, &work_debug_descr);
}
EXPORT_SYMBOL_GPL(destroy_work_on_stack);

#else
static inline void debug_work_activate(struct work_struct *work) { }
static inline void debug_work_deactivate(struct work_struct *work) { }
#endif

423 424
/* Serializes the accesses to the list of workqueues. */
static DEFINE_SPINLOCK(workqueue_lock);
L
Linus Torvalds 已提交
425
static LIST_HEAD(workqueues);
426
static bool workqueue_freezing;		/* W: have wqs started freezing? */
T
Tejun Heo 已提交
427

428 429 430 431 432
/*
 * The almighty global cpu workqueues.  nr_running is the only field
 * which is expected to be used frequently by other cpus via
 * try_to_wake_up().  Put it in a separate cacheline.
 */
433
static DEFINE_PER_CPU(struct global_cwq, global_cwq);
434
static DEFINE_PER_CPU_SHARED_ALIGNED(atomic_t, gcwq_nr_running);
435

436 437 438 439 440 441 442 443
/*
 * Global cpu workqueue and nr_running counter for unbound gcwq.  The
 * gcwq is always online, has GCWQ_DISASSOCIATED set, and all its
 * workers have WORKER_UNBOUND set.
 */
static struct global_cwq unbound_global_cwq;
static atomic_t unbound_gcwq_nr_running = ATOMIC_INIT(0);	/* always 0 */

T
Tejun Heo 已提交
444
static int worker_thread(void *__worker);
L
Linus Torvalds 已提交
445

446 447
static struct global_cwq *get_gcwq(unsigned int cpu)
{
448 449 450 451
	if (cpu != WORK_CPU_UNBOUND)
		return &per_cpu(global_cwq, cpu);
	else
		return &unbound_global_cwq;
452 453
}

454 455
static atomic_t *get_gcwq_nr_running(unsigned int cpu)
{
456 457 458 459
	if (cpu != WORK_CPU_UNBOUND)
		return &per_cpu(gcwq_nr_running, cpu);
	else
		return &unbound_gcwq_nr_running;
460 461
}

T
Tejun Heo 已提交
462 463
static struct cpu_workqueue_struct *get_cwq(unsigned int cpu,
					    struct workqueue_struct *wq)
464
{
465 466 467 468
	if (!(wq->flags & WQ_UNBOUND)) {
		if (likely(cpu < nr_cpu_ids)) {
#ifdef CONFIG_SMP
			return per_cpu_ptr(wq->cpu_wq.pcpu, cpu);
469
#else
470
			return wq->cpu_wq.single;
471
#endif
472 473 474 475
		}
	} else if (likely(cpu == WORK_CPU_UNBOUND))
		return wq->cpu_wq.single;
	return NULL;
476 477
}

478 479 480 481 482 483 484 485 486 487 488 489 490 491 492
static unsigned int work_color_to_flags(int color)
{
	return color << WORK_STRUCT_COLOR_SHIFT;
}

static int get_work_color(struct work_struct *work)
{
	return (*work_data_bits(work) >> WORK_STRUCT_COLOR_SHIFT) &
		((1 << WORK_STRUCT_COLOR_BITS) - 1);
}

static int work_next_color(int color)
{
	return (color + 1) % WORK_NR_COLORS;
}
L
Linus Torvalds 已提交
493

494
/*
495 496 497
 * A work's data points to the cwq with WORK_STRUCT_CWQ set while the
 * work is on queue.  Once execution starts, WORK_STRUCT_CWQ is
 * cleared and the work data contains the cpu number it was last on.
498 499 500 501 502 503 504 505 506
 *
 * set_work_{cwq|cpu}() and clear_work_data() can be used to set the
 * cwq, cpu or clear work->data.  These functions should only be
 * called while the work is owned - ie. while the PENDING bit is set.
 *
 * get_work_[g]cwq() can be used to obtain the gcwq or cwq
 * corresponding to a work.  gcwq is available once the work has been
 * queued anywhere after initialization.  cwq is available only from
 * queueing until execution starts.
507
 */
508 509
static inline void set_work_data(struct work_struct *work, unsigned long data,
				 unsigned long flags)
510
{
511
	BUG_ON(!work_pending(work));
512 513
	atomic_long_set(&work->data, data | flags | work_static(work));
}
514

515 516 517 518 519
static void set_work_cwq(struct work_struct *work,
			 struct cpu_workqueue_struct *cwq,
			 unsigned long extra_flags)
{
	set_work_data(work, (unsigned long)cwq,
520
		      WORK_STRUCT_PENDING | WORK_STRUCT_CWQ | extra_flags);
521 522
}

523 524 525 526
static void set_work_cpu(struct work_struct *work, unsigned int cpu)
{
	set_work_data(work, cpu << WORK_STRUCT_FLAG_BITS, WORK_STRUCT_PENDING);
}
527

528
static void clear_work_data(struct work_struct *work)
L
Linus Torvalds 已提交
529
{
530
	set_work_data(work, WORK_STRUCT_NO_CPU, 0);
L
Linus Torvalds 已提交
531 532
}

533
static struct cpu_workqueue_struct *get_work_cwq(struct work_struct *work)
534
{
535
	unsigned long data = atomic_long_read(&work->data);
536

537 538 539 540
	if (data & WORK_STRUCT_CWQ)
		return (void *)(data & WORK_STRUCT_WQ_DATA_MASK);
	else
		return NULL;
541 542
}

543
static struct global_cwq *get_work_gcwq(struct work_struct *work)
544
{
545
	unsigned long data = atomic_long_read(&work->data);
546 547
	unsigned int cpu;

548 549 550
	if (data & WORK_STRUCT_CWQ)
		return ((struct cpu_workqueue_struct *)
			(data & WORK_STRUCT_WQ_DATA_MASK))->gcwq;
551 552

	cpu = data >> WORK_STRUCT_FLAG_BITS;
553
	if (cpu == WORK_CPU_NONE)
554 555
		return NULL;

556
	BUG_ON(cpu >= nr_cpu_ids && cpu != WORK_CPU_UNBOUND);
557
	return get_gcwq(cpu);
558 559
}

560 561 562 563 564 565
/*
 * Policy functions.  These define the policies on how the global
 * worker pool is managed.  Unless noted otherwise, these functions
 * assume that they're being called with gcwq->lock held.
 */

566
static bool __need_more_worker(struct global_cwq *gcwq)
567
{
568 569
	return !atomic_read(get_gcwq_nr_running(gcwq->cpu)) ||
		gcwq->flags & GCWQ_HIGHPRI_PENDING;
570 571
}

572
/*
573 574
 * Need to wake up a worker?  Called from anything but currently
 * running workers.
575
 */
576
static bool need_more_worker(struct global_cwq *gcwq)
577
{
578
	return !list_empty(&gcwq->worklist) && __need_more_worker(gcwq);
579
}
580

581 582 583 584 585 586 587 588 589 590 591
/* Can I start working?  Called from busy but !running workers. */
static bool may_start_working(struct global_cwq *gcwq)
{
	return gcwq->nr_idle;
}

/* Do I need to keep working?  Called from currently running workers. */
static bool keep_working(struct global_cwq *gcwq)
{
	atomic_t *nr_running = get_gcwq_nr_running(gcwq->cpu);

592 593 594
	return !list_empty(&gcwq->worklist) &&
		(atomic_read(nr_running) <= 1 ||
		 gcwq->flags & GCWQ_HIGHPRI_PENDING);
595 596 597 598 599 600 601
}

/* Do we need a new worker?  Called from manager. */
static bool need_to_create_worker(struct global_cwq *gcwq)
{
	return need_more_worker(gcwq) && !may_start_working(gcwq);
}
602

603 604 605 606 607 608 609 610 611 612 613 614 615 616
/* Do I need to be the manager? */
static bool need_to_manage_workers(struct global_cwq *gcwq)
{
	return need_to_create_worker(gcwq) || gcwq->flags & GCWQ_MANAGE_WORKERS;
}

/* Do we have too many workers and should some go away? */
static bool too_many_workers(struct global_cwq *gcwq)
{
	bool managing = gcwq->flags & GCWQ_MANAGING_WORKERS;
	int nr_idle = gcwq->nr_idle + managing; /* manager is considered idle */
	int nr_busy = gcwq->nr_workers - nr_idle;

	return nr_idle > 2 && (nr_idle - 2) * MAX_IDLE_WORKERS_RATIO >= nr_busy;
617 618
}

619
/*
620 621 622
 * Wake up functions.
 */

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
/* Return the first worker.  Safe with preemption disabled */
static struct worker *first_worker(struct global_cwq *gcwq)
{
	if (unlikely(list_empty(&gcwq->idle_list)))
		return NULL;

	return list_first_entry(&gcwq->idle_list, struct worker, entry);
}

/**
 * wake_up_worker - wake up an idle worker
 * @gcwq: gcwq to wake worker for
 *
 * Wake up the first idle worker of @gcwq.
 *
 * CONTEXT:
 * spin_lock_irq(gcwq->lock).
 */
static void wake_up_worker(struct global_cwq *gcwq)
{
	struct worker *worker = first_worker(gcwq);

	if (likely(worker))
		wake_up_process(worker->task);
}

649
/**
650 651 652 653 654 655 656 657 658 659 660 661 662 663
 * wq_worker_waking_up - a worker is waking up
 * @task: task waking up
 * @cpu: CPU @task is waking up to
 *
 * This function is called during try_to_wake_up() when a worker is
 * being awoken.
 *
 * CONTEXT:
 * spin_lock_irq(rq->lock)
 */
void wq_worker_waking_up(struct task_struct *task, unsigned int cpu)
{
	struct worker *worker = kthread_data(task);

664
	if (!(worker->flags & WORKER_NOT_RUNNING))
665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689
		atomic_inc(get_gcwq_nr_running(cpu));
}

/**
 * wq_worker_sleeping - a worker is going to sleep
 * @task: task going to sleep
 * @cpu: CPU in question, must be the current CPU number
 *
 * This function is called during schedule() when a busy worker is
 * going to sleep.  Worker on the same cpu can be woken up by
 * returning pointer to its task.
 *
 * CONTEXT:
 * spin_lock_irq(rq->lock)
 *
 * RETURNS:
 * Worker task on @cpu to wake up, %NULL if none.
 */
struct task_struct *wq_worker_sleeping(struct task_struct *task,
				       unsigned int cpu)
{
	struct worker *worker = kthread_data(task), *to_wakeup = NULL;
	struct global_cwq *gcwq = get_gcwq(cpu);
	atomic_t *nr_running = get_gcwq_nr_running(cpu);

690
	if (worker->flags & WORKER_NOT_RUNNING)
691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713
		return NULL;

	/* this can only happen on the local cpu */
	BUG_ON(cpu != raw_smp_processor_id());

	/*
	 * The counterpart of the following dec_and_test, implied mb,
	 * worklist not empty test sequence is in insert_work().
	 * Please read comment there.
	 *
	 * NOT_RUNNING is clear.  This means that trustee is not in
	 * charge and we're running on the local cpu w/ rq lock held
	 * and preemption disabled, which in turn means that none else
	 * could be manipulating idle_list, so dereferencing idle_list
	 * without gcwq lock is safe.
	 */
	if (atomic_dec_and_test(nr_running) && !list_empty(&gcwq->worklist))
		to_wakeup = first_worker(gcwq);
	return to_wakeup ? to_wakeup->task : NULL;
}

/**
 * worker_set_flags - set worker flags and adjust nr_running accordingly
714
 * @worker: self
715 716 717
 * @flags: flags to set
 * @wakeup: wakeup an idle worker if necessary
 *
718 719 720
 * Set @flags in @worker->flags and adjust nr_running accordingly.  If
 * nr_running becomes zero and @wakeup is %true, an idle worker is
 * woken up.
721
 *
722 723
 * CONTEXT:
 * spin_lock_irq(gcwq->lock)
724 725 726 727
 */
static inline void worker_set_flags(struct worker *worker, unsigned int flags,
				    bool wakeup)
{
728 729
	struct global_cwq *gcwq = worker->gcwq;

730 731
	WARN_ON_ONCE(worker->task != current);

732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748
	/*
	 * If transitioning into NOT_RUNNING, adjust nr_running and
	 * wake up an idle worker as necessary if requested by
	 * @wakeup.
	 */
	if ((flags & WORKER_NOT_RUNNING) &&
	    !(worker->flags & WORKER_NOT_RUNNING)) {
		atomic_t *nr_running = get_gcwq_nr_running(gcwq->cpu);

		if (wakeup) {
			if (atomic_dec_and_test(nr_running) &&
			    !list_empty(&gcwq->worklist))
				wake_up_worker(gcwq);
		} else
			atomic_dec(nr_running);
	}

749 750 751 752
	worker->flags |= flags;
}

/**
753
 * worker_clr_flags - clear worker flags and adjust nr_running accordingly
754
 * @worker: self
755 756
 * @flags: flags to clear
 *
757
 * Clear @flags in @worker->flags and adjust nr_running accordingly.
758
 *
759 760
 * CONTEXT:
 * spin_lock_irq(gcwq->lock)
761 762 763
 */
static inline void worker_clr_flags(struct worker *worker, unsigned int flags)
{
764 765 766
	struct global_cwq *gcwq = worker->gcwq;
	unsigned int oflags = worker->flags;

767 768
	WARN_ON_ONCE(worker->task != current);

769
	worker->flags &= ~flags;
770

771 772 773 774 775
	/*
	 * If transitioning out of NOT_RUNNING, increment nr_running.  Note
	 * that the nested NOT_RUNNING is not a noop.  NOT_RUNNING is mask
	 * of multiple flags, not a single flag.
	 */
776 777 778
	if ((flags & WORKER_NOT_RUNNING) && (oflags & WORKER_NOT_RUNNING))
		if (!(worker->flags & WORKER_NOT_RUNNING))
			atomic_inc(get_gcwq_nr_running(gcwq->cpu));
779 780
}

T
Tejun Heo 已提交
781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807
/**
 * busy_worker_head - return the busy hash head for a work
 * @gcwq: gcwq of interest
 * @work: work to be hashed
 *
 * Return hash head of @gcwq for @work.
 *
 * CONTEXT:
 * spin_lock_irq(gcwq->lock).
 *
 * RETURNS:
 * Pointer to the hash head.
 */
static struct hlist_head *busy_worker_head(struct global_cwq *gcwq,
					   struct work_struct *work)
{
	const int base_shift = ilog2(sizeof(struct work_struct));
	unsigned long v = (unsigned long)work;

	/* simple shift and fold hash, do we need something better? */
	v >>= base_shift;
	v += v >> BUSY_WORKER_HASH_ORDER;
	v &= BUSY_WORKER_HASH_MASK;

	return &gcwq->busy_hash[v];
}

808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852
/**
 * __find_worker_executing_work - find worker which is executing a work
 * @gcwq: gcwq of interest
 * @bwh: hash head as returned by busy_worker_head()
 * @work: work to find worker for
 *
 * Find a worker which is executing @work on @gcwq.  @bwh should be
 * the hash head obtained by calling busy_worker_head() with the same
 * work.
 *
 * CONTEXT:
 * spin_lock_irq(gcwq->lock).
 *
 * RETURNS:
 * Pointer to worker which is executing @work if found, NULL
 * otherwise.
 */
static struct worker *__find_worker_executing_work(struct global_cwq *gcwq,
						   struct hlist_head *bwh,
						   struct work_struct *work)
{
	struct worker *worker;
	struct hlist_node *tmp;

	hlist_for_each_entry(worker, tmp, bwh, hentry)
		if (worker->current_work == work)
			return worker;
	return NULL;
}

/**
 * find_worker_executing_work - find worker which is executing a work
 * @gcwq: gcwq of interest
 * @work: work to find worker for
 *
 * Find a worker which is executing @work on @gcwq.  This function is
 * identical to __find_worker_executing_work() except that this
 * function calculates @bwh itself.
 *
 * CONTEXT:
 * spin_lock_irq(gcwq->lock).
 *
 * RETURNS:
 * Pointer to worker which is executing @work if found, NULL
 * otherwise.
853
 */
854 855
static struct worker *find_worker_executing_work(struct global_cwq *gcwq,
						 struct work_struct *work)
856
{
857 858
	return __find_worker_executing_work(gcwq, busy_worker_head(gcwq, work),
					    work);
859 860
}

861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880
/**
 * gcwq_determine_ins_pos - find insertion position
 * @gcwq: gcwq of interest
 * @cwq: cwq a work is being queued for
 *
 * A work for @cwq is about to be queued on @gcwq, determine insertion
 * position for the work.  If @cwq is for HIGHPRI wq, the work is
 * queued at the head of the queue but in FIFO order with respect to
 * other HIGHPRI works; otherwise, at the end of the queue.  This
 * function also sets GCWQ_HIGHPRI_PENDING flag to hint @gcwq that
 * there are HIGHPRI works pending.
 *
 * CONTEXT:
 * spin_lock_irq(gcwq->lock).
 *
 * RETURNS:
 * Pointer to inserstion position.
 */
static inline struct list_head *gcwq_determine_ins_pos(struct global_cwq *gcwq,
					       struct cpu_workqueue_struct *cwq)
881
{
882 883 884 885 886 887 888 889 890 891 892 893 894 895
	struct work_struct *twork;

	if (likely(!(cwq->wq->flags & WQ_HIGHPRI)))
		return &gcwq->worklist;

	list_for_each_entry(twork, &gcwq->worklist, entry) {
		struct cpu_workqueue_struct *tcwq = get_work_cwq(twork);

		if (!(tcwq->wq->flags & WQ_HIGHPRI))
			break;
	}

	gcwq->flags |= GCWQ_HIGHPRI_PENDING;
	return &twork->entry;
896 897
}

T
Tejun Heo 已提交
898
/**
899
 * insert_work - insert a work into gcwq
T
Tejun Heo 已提交
900 901 902 903 904
 * @cwq: cwq @work belongs to
 * @work: work to insert
 * @head: insertion point
 * @extra_flags: extra WORK_STRUCT_* flags to set
 *
905 906
 * Insert @work which belongs to @cwq into @gcwq after @head.
 * @extra_flags is or'd to work_struct flags.
T
Tejun Heo 已提交
907 908
 *
 * CONTEXT:
909
 * spin_lock_irq(gcwq->lock).
T
Tejun Heo 已提交
910
 */
O
Oleg Nesterov 已提交
911
static void insert_work(struct cpu_workqueue_struct *cwq,
T
Tejun Heo 已提交
912 913
			struct work_struct *work, struct list_head *head,
			unsigned int extra_flags)
O
Oleg Nesterov 已提交
914
{
915 916
	struct global_cwq *gcwq = cwq->gcwq;

T
Tejun Heo 已提交
917
	/* we own @work, set data and link */
918
	set_work_cwq(work, cwq, extra_flags);
919

920 921 922 923 924
	/*
	 * Ensure that we get the right work->data if we see the
	 * result of list_add() below, see try_to_grab_pending().
	 */
	smp_wmb();
T
Tejun Heo 已提交
925

926
	list_add_tail(&work->entry, head);
927 928 929 930 931 932 933 934

	/*
	 * Ensure either worker_sched_deactivated() sees the above
	 * list_add_tail() or we see zero nr_running to avoid workers
	 * lying around lazily while there are works to be processed.
	 */
	smp_mb();

935
	if (__need_more_worker(gcwq))
936
		wake_up_worker(gcwq);
O
Oleg Nesterov 已提交
937 938
}

939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970
/*
 * Test whether @work is being queued from another work executing on the
 * same workqueue.  This is rather expensive and should only be used from
 * cold paths.
 */
static bool is_chained_work(struct workqueue_struct *wq)
{
	unsigned long flags;
	unsigned int cpu;

	for_each_gcwq_cpu(cpu) {
		struct global_cwq *gcwq = get_gcwq(cpu);
		struct worker *worker;
		struct hlist_node *pos;
		int i;

		spin_lock_irqsave(&gcwq->lock, flags);
		for_each_busy_worker(worker, i, pos, gcwq) {
			if (worker->task != current)
				continue;
			spin_unlock_irqrestore(&gcwq->lock, flags);
			/*
			 * I'm @worker, no locking necessary.  See if @work
			 * is headed to the same workqueue.
			 */
			return worker->current_cwq->wq == wq;
		}
		spin_unlock_irqrestore(&gcwq->lock, flags);
	}
	return false;
}

T
Tejun Heo 已提交
971
static void __queue_work(unsigned int cpu, struct workqueue_struct *wq,
L
Linus Torvalds 已提交
972 973
			 struct work_struct *work)
{
974 975
	struct global_cwq *gcwq;
	struct cpu_workqueue_struct *cwq;
976
	struct list_head *worklist;
977
	unsigned int work_flags;
L
Linus Torvalds 已提交
978 979
	unsigned long flags;

980
	debug_work_activate(work);
981

982 983 984
	/* if dying, only works from the same workqueue are allowed */
	if (unlikely(wq->flags & WQ_DYING) &&
	    WARN_ON_ONCE(!is_chained_work(wq)))
985 986
		return;

987 988
	/* determine gcwq to use */
	if (!(wq->flags & WQ_UNBOUND)) {
989 990
		struct global_cwq *last_gcwq;

991 992 993
		if (unlikely(cpu == WORK_CPU_UNBOUND))
			cpu = raw_smp_processor_id();

994 995 996 997 998 999
		/*
		 * It's multi cpu.  If @wq is non-reentrant and @work
		 * was previously on a different cpu, it might still
		 * be running there, in which case the work needs to
		 * be queued on that cpu to guarantee non-reentrance.
		 */
1000
		gcwq = get_gcwq(cpu);
1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017
		if (wq->flags & WQ_NON_REENTRANT &&
		    (last_gcwq = get_work_gcwq(work)) && last_gcwq != gcwq) {
			struct worker *worker;

			spin_lock_irqsave(&last_gcwq->lock, flags);

			worker = find_worker_executing_work(last_gcwq, work);

			if (worker && worker->current_cwq->wq == wq)
				gcwq = last_gcwq;
			else {
				/* meh... not running there, queue here */
				spin_unlock_irqrestore(&last_gcwq->lock, flags);
				spin_lock_irqsave(&gcwq->lock, flags);
			}
		} else
			spin_lock_irqsave(&gcwq->lock, flags);
1018 1019 1020
	} else {
		gcwq = get_gcwq(WORK_CPU_UNBOUND);
		spin_lock_irqsave(&gcwq->lock, flags);
1021 1022 1023 1024
	}

	/* gcwq determined, get cwq and queue */
	cwq = get_cwq(gcwq->cpu, wq);
1025
	trace_workqueue_queue_work(cpu, cwq, work);
1026

T
Tejun Heo 已提交
1027
	BUG_ON(!list_empty(&work->entry));
1028

1029
	cwq->nr_in_flight[cwq->work_color]++;
1030
	work_flags = work_color_to_flags(cwq->work_color);
1031 1032

	if (likely(cwq->nr_active < cwq->max_active)) {
1033
		trace_workqueue_activate_work(work);
1034
		cwq->nr_active++;
1035
		worklist = gcwq_determine_ins_pos(gcwq, cwq);
1036 1037
	} else {
		work_flags |= WORK_STRUCT_DELAYED;
1038
		worklist = &cwq->delayed_works;
1039
	}
1040

1041
	insert_work(cwq, work, worklist, work_flags);
1042

1043
	spin_unlock_irqrestore(&gcwq->lock, flags);
L
Linus Torvalds 已提交
1044 1045
}

1046 1047 1048 1049 1050
/**
 * queue_work - queue work on a workqueue
 * @wq: workqueue to use
 * @work: work to queue
 *
A
Alan Stern 已提交
1051
 * Returns 0 if @work was already on a queue, non-zero otherwise.
L
Linus Torvalds 已提交
1052
 *
1053 1054
 * We queue the work to the CPU on which it was submitted, but if the CPU dies
 * it can be processed by another CPU.
L
Linus Torvalds 已提交
1055
 */
1056
int queue_work(struct workqueue_struct *wq, struct work_struct *work)
L
Linus Torvalds 已提交
1057
{
1058 1059 1060 1061 1062
	int ret;

	ret = queue_work_on(get_cpu(), wq, work);
	put_cpu();

L
Linus Torvalds 已提交
1063 1064
	return ret;
}
1065
EXPORT_SYMBOL_GPL(queue_work);
L
Linus Torvalds 已提交
1066

1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082
/**
 * queue_work_on - queue work on specific cpu
 * @cpu: CPU number to execute work on
 * @wq: workqueue to use
 * @work: work to queue
 *
 * Returns 0 if @work was already on a queue, non-zero otherwise.
 *
 * We queue the work to a specific CPU, the caller must ensure it
 * can't go away.
 */
int
queue_work_on(int cpu, struct workqueue_struct *wq, struct work_struct *work)
{
	int ret = 0;

1083
	if (!test_and_set_bit(WORK_STRUCT_PENDING_BIT, work_data_bits(work))) {
T
Tejun Heo 已提交
1084
		__queue_work(cpu, wq, work);
1085 1086 1087 1088 1089 1090
		ret = 1;
	}
	return ret;
}
EXPORT_SYMBOL_GPL(queue_work_on);

1091
static void delayed_work_timer_fn(unsigned long __data)
L
Linus Torvalds 已提交
1092
{
1093
	struct delayed_work *dwork = (struct delayed_work *)__data;
1094
	struct cpu_workqueue_struct *cwq = get_work_cwq(&dwork->work);
L
Linus Torvalds 已提交
1095

T
Tejun Heo 已提交
1096
	__queue_work(smp_processor_id(), cwq->wq, &dwork->work);
L
Linus Torvalds 已提交
1097 1098
}

1099 1100 1101
/**
 * queue_delayed_work - queue work on a workqueue after delay
 * @wq: workqueue to use
1102
 * @dwork: delayable work to queue
1103 1104
 * @delay: number of jiffies to wait before queueing
 *
A
Alan Stern 已提交
1105
 * Returns 0 if @work was already on a queue, non-zero otherwise.
1106
 */
1107
int queue_delayed_work(struct workqueue_struct *wq,
1108
			struct delayed_work *dwork, unsigned long delay)
L
Linus Torvalds 已提交
1109
{
1110
	if (delay == 0)
1111
		return queue_work(wq, &dwork->work);
L
Linus Torvalds 已提交
1112

1113
	return queue_delayed_work_on(-1, wq, dwork, delay);
L
Linus Torvalds 已提交
1114
}
1115
EXPORT_SYMBOL_GPL(queue_delayed_work);
L
Linus Torvalds 已提交
1116

1117 1118 1119 1120
/**
 * queue_delayed_work_on - queue work on specific CPU after delay
 * @cpu: CPU number to execute work on
 * @wq: workqueue to use
1121
 * @dwork: work to queue
1122 1123
 * @delay: number of jiffies to wait before queueing
 *
A
Alan Stern 已提交
1124
 * Returns 0 if @work was already on a queue, non-zero otherwise.
1125
 */
1126
int queue_delayed_work_on(int cpu, struct workqueue_struct *wq,
1127
			struct delayed_work *dwork, unsigned long delay)
1128 1129
{
	int ret = 0;
1130 1131
	struct timer_list *timer = &dwork->timer;
	struct work_struct *work = &dwork->work;
1132

1133
	if (!test_and_set_bit(WORK_STRUCT_PENDING_BIT, work_data_bits(work))) {
1134
		unsigned int lcpu;
1135

1136 1137 1138
		BUG_ON(timer_pending(timer));
		BUG_ON(!list_empty(&work->entry));

1139 1140
		timer_stats_timer_set_start_info(&dwork->timer);

1141 1142 1143 1144 1145
		/*
		 * This stores cwq for the moment, for the timer_fn.
		 * Note that the work's gcwq is preserved to allow
		 * reentrance detection for delayed works.
		 */
1146 1147 1148 1149 1150 1151 1152 1153 1154 1155
		if (!(wq->flags & WQ_UNBOUND)) {
			struct global_cwq *gcwq = get_work_gcwq(work);

			if (gcwq && gcwq->cpu != WORK_CPU_UNBOUND)
				lcpu = gcwq->cpu;
			else
				lcpu = raw_smp_processor_id();
		} else
			lcpu = WORK_CPU_UNBOUND;

1156
		set_work_cwq(work, get_cwq(lcpu, wq), 0);
1157

1158
		timer->expires = jiffies + delay;
1159
		timer->data = (unsigned long)dwork;
1160
		timer->function = delayed_work_timer_fn;
1161 1162 1163 1164 1165

		if (unlikely(cpu >= 0))
			add_timer_on(timer, cpu);
		else
			add_timer(timer);
1166 1167 1168 1169
		ret = 1;
	}
	return ret;
}
1170
EXPORT_SYMBOL_GPL(queue_delayed_work_on);
L
Linus Torvalds 已提交
1171

T
Tejun Heo 已提交
1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182
/**
 * worker_enter_idle - enter idle state
 * @worker: worker which is entering idle state
 *
 * @worker is entering idle state.  Update stats and idle timer if
 * necessary.
 *
 * LOCKING:
 * spin_lock_irq(gcwq->lock).
 */
static void worker_enter_idle(struct worker *worker)
L
Linus Torvalds 已提交
1183
{
T
Tejun Heo 已提交
1184 1185 1186 1187 1188 1189
	struct global_cwq *gcwq = worker->gcwq;

	BUG_ON(worker->flags & WORKER_IDLE);
	BUG_ON(!list_empty(&worker->entry) &&
	       (worker->hentry.next || worker->hentry.pprev));

1190 1191
	/* can't use worker_set_flags(), also called from start_worker() */
	worker->flags |= WORKER_IDLE;
T
Tejun Heo 已提交
1192
	gcwq->nr_idle++;
1193
	worker->last_active = jiffies;
T
Tejun Heo 已提交
1194 1195 1196

	/* idle_list is LIFO */
	list_add(&worker->entry, &gcwq->idle_list);
1197

1198 1199 1200 1201 1202
	if (likely(!(worker->flags & WORKER_ROGUE))) {
		if (too_many_workers(gcwq) && !timer_pending(&gcwq->idle_timer))
			mod_timer(&gcwq->idle_timer,
				  jiffies + IDLE_WORKER_TIMEOUT);
	} else
1203
		wake_up_all(&gcwq->trustee_wait);
1204 1205 1206 1207

	/* sanity check nr_running */
	WARN_ON_ONCE(gcwq->nr_workers == gcwq->nr_idle &&
		     atomic_read(get_gcwq_nr_running(gcwq->cpu)));
T
Tejun Heo 已提交
1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223
}

/**
 * worker_leave_idle - leave idle state
 * @worker: worker which is leaving idle state
 *
 * @worker is leaving idle state.  Update stats.
 *
 * LOCKING:
 * spin_lock_irq(gcwq->lock).
 */
static void worker_leave_idle(struct worker *worker)
{
	struct global_cwq *gcwq = worker->gcwq;

	BUG_ON(!(worker->flags & WORKER_IDLE));
1224
	worker_clr_flags(worker, WORKER_IDLE);
T
Tejun Heo 已提交
1225 1226 1227 1228
	gcwq->nr_idle--;
	list_del_init(&worker->entry);
}

1229 1230 1231 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
/**
 * worker_maybe_bind_and_lock - bind worker to its cpu if possible and lock gcwq
 * @worker: self
 *
 * Works which are scheduled while the cpu is online must at least be
 * scheduled to a worker which is bound to the cpu so that if they are
 * flushed from cpu callbacks while cpu is going down, they are
 * guaranteed to execute on the cpu.
 *
 * This function is to be used by rogue workers and rescuers to bind
 * themselves to the target cpu and may race with cpu going down or
 * coming online.  kthread_bind() can't be used because it may put the
 * worker to already dead cpu and set_cpus_allowed_ptr() can't be used
 * verbatim as it's best effort and blocking and gcwq may be
 * [dis]associated in the meantime.
 *
 * This function tries set_cpus_allowed() and locks gcwq and verifies
 * the binding against GCWQ_DISASSOCIATED which is set during
 * CPU_DYING and cleared during CPU_ONLINE, so if the worker enters
 * idle state or fetches works without dropping lock, it can guarantee
 * the scheduling requirement described in the first paragraph.
 *
 * CONTEXT:
 * Might sleep.  Called without any lock but returns with gcwq->lock
 * held.
 *
 * RETURNS:
 * %true if the associated gcwq is online (@worker is successfully
 * bound), %false if offline.
 */
static bool worker_maybe_bind_and_lock(struct worker *worker)
1260
__acquires(&gcwq->lock)
1261 1262 1263 1264 1265
{
	struct global_cwq *gcwq = worker->gcwq;
	struct task_struct *task = worker->task;

	while (true) {
1266
		/*
1267 1268 1269 1270
		 * The following call may fail, succeed or succeed
		 * without actually migrating the task to the cpu if
		 * it races with cpu hotunplug operation.  Verify
		 * against GCWQ_DISASSOCIATED.
1271
		 */
1272 1273
		if (!(gcwq->flags & GCWQ_DISASSOCIATED))
			set_cpus_allowed_ptr(task, get_cpu_mask(gcwq->cpu));
1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305

		spin_lock_irq(&gcwq->lock);
		if (gcwq->flags & GCWQ_DISASSOCIATED)
			return false;
		if (task_cpu(task) == gcwq->cpu &&
		    cpumask_equal(&current->cpus_allowed,
				  get_cpu_mask(gcwq->cpu)))
			return true;
		spin_unlock_irq(&gcwq->lock);

		/* CPU has come up inbetween, retry migration */
		cpu_relax();
	}
}

/*
 * Function for worker->rebind_work used to rebind rogue busy workers
 * to the associated cpu which is coming back online.  This is
 * scheduled by cpu up but can race with other cpu hotplug operations
 * and may be executed twice without intervening cpu down.
 */
static void worker_rebind_fn(struct work_struct *work)
{
	struct worker *worker = container_of(work, struct worker, rebind_work);
	struct global_cwq *gcwq = worker->gcwq;

	if (worker_maybe_bind_and_lock(worker))
		worker_clr_flags(worker, WORKER_REBIND);

	spin_unlock_irq(&gcwq->lock);
}

T
Tejun Heo 已提交
1306 1307 1308 1309 1310
static struct worker *alloc_worker(void)
{
	struct worker *worker;

	worker = kzalloc(sizeof(*worker), GFP_KERNEL);
T
Tejun Heo 已提交
1311 1312
	if (worker) {
		INIT_LIST_HEAD(&worker->entry);
1313
		INIT_LIST_HEAD(&worker->scheduled);
1314 1315 1316
		INIT_WORK(&worker->rebind_work, worker_rebind_fn);
		/* on creation a worker is in !idle && prep state */
		worker->flags = WORKER_PREP;
T
Tejun Heo 已提交
1317
	}
T
Tejun Heo 已提交
1318 1319 1320 1321 1322
	return worker;
}

/**
 * create_worker - create a new workqueue worker
1323
 * @gcwq: gcwq the new worker will belong to
T
Tejun Heo 已提交
1324 1325
 * @bind: whether to set affinity to @cpu or not
 *
1326
 * Create a new worker which is bound to @gcwq.  The returned worker
T
Tejun Heo 已提交
1327 1328 1329 1330 1331 1332 1333 1334 1335
 * can be started by calling start_worker() or destroyed using
 * destroy_worker().
 *
 * CONTEXT:
 * Might sleep.  Does GFP_KERNEL allocations.
 *
 * RETURNS:
 * Pointer to the newly created worker.
 */
1336
static struct worker *create_worker(struct global_cwq *gcwq, bool bind)
T
Tejun Heo 已提交
1337
{
1338
	bool on_unbound_cpu = gcwq->cpu == WORK_CPU_UNBOUND;
T
Tejun Heo 已提交
1339
	struct worker *worker = NULL;
1340
	int id = -1;
T
Tejun Heo 已提交
1341

1342 1343 1344 1345
	spin_lock_irq(&gcwq->lock);
	while (ida_get_new(&gcwq->worker_ida, &id)) {
		spin_unlock_irq(&gcwq->lock);
		if (!ida_pre_get(&gcwq->worker_ida, GFP_KERNEL))
T
Tejun Heo 已提交
1346
			goto fail;
1347
		spin_lock_irq(&gcwq->lock);
T
Tejun Heo 已提交
1348
	}
1349
	spin_unlock_irq(&gcwq->lock);
T
Tejun Heo 已提交
1350 1351 1352 1353 1354

	worker = alloc_worker();
	if (!worker)
		goto fail;

1355
	worker->gcwq = gcwq;
T
Tejun Heo 已提交
1356 1357
	worker->id = id;

1358 1359 1360 1361 1362 1363
	if (!on_unbound_cpu)
		worker->task = kthread_create(worker_thread, worker,
					      "kworker/%u:%d", gcwq->cpu, id);
	else
		worker->task = kthread_create(worker_thread, worker,
					      "kworker/u:%d", id);
T
Tejun Heo 已提交
1364 1365 1366
	if (IS_ERR(worker->task))
		goto fail;

1367 1368 1369 1370 1371
	/*
	 * A rogue worker will become a regular one if CPU comes
	 * online later on.  Make sure every worker has
	 * PF_THREAD_BOUND set.
	 */
1372
	if (bind && !on_unbound_cpu)
1373
		kthread_bind(worker->task, gcwq->cpu);
1374
	else {
1375
		worker->task->flags |= PF_THREAD_BOUND;
1376 1377 1378
		if (on_unbound_cpu)
			worker->flags |= WORKER_UNBOUND;
	}
T
Tejun Heo 已提交
1379 1380 1381 1382

	return worker;
fail:
	if (id >= 0) {
1383 1384 1385
		spin_lock_irq(&gcwq->lock);
		ida_remove(&gcwq->worker_ida, id);
		spin_unlock_irq(&gcwq->lock);
T
Tejun Heo 已提交
1386 1387 1388 1389 1390 1391 1392 1393 1394
	}
	kfree(worker);
	return NULL;
}

/**
 * start_worker - start a newly created worker
 * @worker: worker to start
 *
T
Tejun Heo 已提交
1395
 * Make the gcwq aware of @worker and start it.
T
Tejun Heo 已提交
1396 1397
 *
 * CONTEXT:
1398
 * spin_lock_irq(gcwq->lock).
T
Tejun Heo 已提交
1399 1400 1401
 */
static void start_worker(struct worker *worker)
{
1402
	worker->flags |= WORKER_STARTED;
T
Tejun Heo 已提交
1403 1404
	worker->gcwq->nr_workers++;
	worker_enter_idle(worker);
T
Tejun Heo 已提交
1405 1406 1407 1408 1409 1410 1411
	wake_up_process(worker->task);
}

/**
 * destroy_worker - destroy a workqueue worker
 * @worker: worker to be destroyed
 *
T
Tejun Heo 已提交
1412 1413 1414 1415
 * Destroy @worker and adjust @gcwq stats accordingly.
 *
 * CONTEXT:
 * spin_lock_irq(gcwq->lock) which is released and regrabbed.
T
Tejun Heo 已提交
1416 1417 1418
 */
static void destroy_worker(struct worker *worker)
{
1419
	struct global_cwq *gcwq = worker->gcwq;
T
Tejun Heo 已提交
1420 1421 1422 1423
	int id = worker->id;

	/* sanity check frenzy */
	BUG_ON(worker->current_work);
1424
	BUG_ON(!list_empty(&worker->scheduled));
T
Tejun Heo 已提交
1425

T
Tejun Heo 已提交
1426 1427 1428 1429 1430 1431
	if (worker->flags & WORKER_STARTED)
		gcwq->nr_workers--;
	if (worker->flags & WORKER_IDLE)
		gcwq->nr_idle--;

	list_del_init(&worker->entry);
1432
	worker->flags |= WORKER_DIE;
T
Tejun Heo 已提交
1433 1434 1435

	spin_unlock_irq(&gcwq->lock);

T
Tejun Heo 已提交
1436 1437 1438
	kthread_stop(worker->task);
	kfree(worker);

1439 1440
	spin_lock_irq(&gcwq->lock);
	ida_remove(&gcwq->worker_ida, id);
T
Tejun Heo 已提交
1441 1442
}

1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462
static void idle_worker_timeout(unsigned long __gcwq)
{
	struct global_cwq *gcwq = (void *)__gcwq;

	spin_lock_irq(&gcwq->lock);

	if (too_many_workers(gcwq)) {
		struct worker *worker;
		unsigned long expires;

		/* idle_list is kept in LIFO order, check the last one */
		worker = list_entry(gcwq->idle_list.prev, struct worker, entry);
		expires = worker->last_active + IDLE_WORKER_TIMEOUT;

		if (time_before(jiffies, expires))
			mod_timer(&gcwq->idle_timer, expires);
		else {
			/* it's been idle for too long, wake up manager */
			gcwq->flags |= GCWQ_MANAGE_WORKERS;
			wake_up_worker(gcwq);
1463
		}
1464 1465 1466 1467
	}

	spin_unlock_irq(&gcwq->lock);
}
1468

1469 1470 1471 1472
static bool send_mayday(struct work_struct *work)
{
	struct cpu_workqueue_struct *cwq = get_work_cwq(work);
	struct workqueue_struct *wq = cwq->wq;
1473
	unsigned int cpu;
1474 1475 1476 1477 1478

	if (!(wq->flags & WQ_RESCUER))
		return false;

	/* mayday mayday mayday */
1479 1480 1481 1482
	cpu = cwq->gcwq->cpu;
	/* WORK_CPU_UNBOUND can't be set in cpumask, use cpu 0 instead */
	if (cpu == WORK_CPU_UNBOUND)
		cpu = 0;
1483
	if (!mayday_test_and_set_cpu(cpu, wq->mayday_mask))
1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503
		wake_up_process(wq->rescuer->task);
	return true;
}

static void gcwq_mayday_timeout(unsigned long __gcwq)
{
	struct global_cwq *gcwq = (void *)__gcwq;
	struct work_struct *work;

	spin_lock_irq(&gcwq->lock);

	if (need_to_create_worker(gcwq)) {
		/*
		 * We've been trying to create a new worker but
		 * haven't been successful.  We might be hitting an
		 * allocation deadlock.  Send distress signals to
		 * rescuers.
		 */
		list_for_each_entry(work, &gcwq->worklist, entry)
			send_mayday(work);
L
Linus Torvalds 已提交
1504
	}
1505 1506 1507 1508

	spin_unlock_irq(&gcwq->lock);

	mod_timer(&gcwq->mayday_timer, jiffies + MAYDAY_INTERVAL);
L
Linus Torvalds 已提交
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
/**
 * maybe_create_worker - create a new worker if necessary
 * @gcwq: gcwq to create a new worker for
 *
 * Create a new worker for @gcwq if necessary.  @gcwq is guaranteed to
 * have at least one idle worker on return from this function.  If
 * creating a new worker takes longer than MAYDAY_INTERVAL, mayday is
 * sent to all rescuers with works scheduled on @gcwq to resolve
 * possible allocation deadlock.
 *
 * On return, need_to_create_worker() is guaranteed to be false and
 * may_start_working() true.
 *
 * LOCKING:
 * spin_lock_irq(gcwq->lock) which may be released and regrabbed
 * multiple times.  Does GFP_KERNEL allocations.  Called only from
 * manager.
 *
 * RETURNS:
 * false if no action was taken and gcwq->lock stayed locked, true
 * otherwise.
 */
static bool maybe_create_worker(struct global_cwq *gcwq)
1534 1535
__releases(&gcwq->lock)
__acquires(&gcwq->lock)
L
Linus Torvalds 已提交
1536
{
1537 1538 1539
	if (!need_to_create_worker(gcwq))
		return false;
restart:
1540 1541
	spin_unlock_irq(&gcwq->lock);

1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558
	/* if we don't make progress in MAYDAY_INITIAL_TIMEOUT, call for help */
	mod_timer(&gcwq->mayday_timer, jiffies + MAYDAY_INITIAL_TIMEOUT);

	while (true) {
		struct worker *worker;

		worker = create_worker(gcwq, true);
		if (worker) {
			del_timer_sync(&gcwq->mayday_timer);
			spin_lock_irq(&gcwq->lock);
			start_worker(worker);
			BUG_ON(need_to_create_worker(gcwq));
			return true;
		}

		if (!need_to_create_worker(gcwq))
			break;
L
Linus Torvalds 已提交
1559

1560 1561
		__set_current_state(TASK_INTERRUPTIBLE);
		schedule_timeout(CREATE_COOLDOWN);
1562

1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591
		if (!need_to_create_worker(gcwq))
			break;
	}

	del_timer_sync(&gcwq->mayday_timer);
	spin_lock_irq(&gcwq->lock);
	if (need_to_create_worker(gcwq))
		goto restart;
	return true;
}

/**
 * maybe_destroy_worker - destroy workers which have been idle for a while
 * @gcwq: gcwq to destroy workers for
 *
 * Destroy @gcwq workers which have been idle for longer than
 * IDLE_WORKER_TIMEOUT.
 *
 * LOCKING:
 * spin_lock_irq(gcwq->lock) which may be released and regrabbed
 * multiple times.  Called only from manager.
 *
 * RETURNS:
 * false if no action was taken and gcwq->lock stayed locked, true
 * otherwise.
 */
static bool maybe_destroy_workers(struct global_cwq *gcwq)
{
	bool ret = false;
L
Linus Torvalds 已提交
1592

1593 1594 1595
	while (too_many_workers(gcwq)) {
		struct worker *worker;
		unsigned long expires;
1596

1597 1598
		worker = list_entry(gcwq->idle_list.prev, struct worker, entry);
		expires = worker->last_active + IDLE_WORKER_TIMEOUT;
1599

1600 1601
		if (time_before(jiffies, expires)) {
			mod_timer(&gcwq->idle_timer, expires);
1602
			break;
1603
		}
L
Linus Torvalds 已提交
1604

1605 1606
		destroy_worker(worker);
		ret = true;
L
Linus Torvalds 已提交
1607
	}
1608

1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661
	return ret;
}

/**
 * manage_workers - manage worker pool
 * @worker: self
 *
 * Assume the manager role and manage gcwq worker pool @worker belongs
 * to.  At any given time, there can be only zero or one manager per
 * gcwq.  The exclusion is handled automatically by this function.
 *
 * The caller can safely start processing works on false return.  On
 * true return, it's guaranteed that need_to_create_worker() is false
 * and may_start_working() is true.
 *
 * CONTEXT:
 * spin_lock_irq(gcwq->lock) which may be released and regrabbed
 * multiple times.  Does GFP_KERNEL allocations.
 *
 * RETURNS:
 * false if no action was taken and gcwq->lock stayed locked, true if
 * some action was taken.
 */
static bool manage_workers(struct worker *worker)
{
	struct global_cwq *gcwq = worker->gcwq;
	bool ret = false;

	if (gcwq->flags & GCWQ_MANAGING_WORKERS)
		return ret;

	gcwq->flags &= ~GCWQ_MANAGE_WORKERS;
	gcwq->flags |= GCWQ_MANAGING_WORKERS;

	/*
	 * Destroy and then create so that may_start_working() is true
	 * on return.
	 */
	ret |= maybe_destroy_workers(gcwq);
	ret |= maybe_create_worker(gcwq);

	gcwq->flags &= ~GCWQ_MANAGING_WORKERS;

	/*
	 * The trustee might be waiting to take over the manager
	 * position, tell it we're done.
	 */
	if (unlikely(gcwq->trustee))
		wake_up_all(&gcwq->trustee_wait);

	return ret;
}

1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676
/**
 * move_linked_works - move linked works to a list
 * @work: start of series of works to be scheduled
 * @head: target list to append @work to
 * @nextp: out paramter for nested worklist walking
 *
 * Schedule linked works starting from @work to @head.  Work series to
 * be scheduled starts at @work and includes any consecutive work with
 * WORK_STRUCT_LINKED set in its predecessor.
 *
 * If @nextp is not NULL, it's updated to point to the next work of
 * the last scheduled work.  This allows move_linked_works() to be
 * nested inside outer list_for_each_entry_safe().
 *
 * CONTEXT:
1677
 * spin_lock_irq(gcwq->lock).
1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702
 */
static void move_linked_works(struct work_struct *work, struct list_head *head,
			      struct work_struct **nextp)
{
	struct work_struct *n;

	/*
	 * Linked worklist will always end before the end of the list,
	 * use NULL for list head.
	 */
	list_for_each_entry_safe_from(work, n, NULL, entry) {
		list_move_tail(&work->entry, head);
		if (!(*work_data_bits(work) & WORK_STRUCT_LINKED))
			break;
	}

	/*
	 * If we're already inside safe list traversal and have moved
	 * multiple works to the scheduled queue, the next position
	 * needs to be updated.
	 */
	if (nextp)
		*nextp = n;
}

1703 1704 1705 1706
static void cwq_activate_first_delayed(struct cpu_workqueue_struct *cwq)
{
	struct work_struct *work = list_first_entry(&cwq->delayed_works,
						    struct work_struct, entry);
1707
	struct list_head *pos = gcwq_determine_ins_pos(cwq->gcwq, cwq);
1708

1709
	trace_workqueue_activate_work(work);
1710
	move_linked_works(work, pos, NULL);
1711
	__clear_bit(WORK_STRUCT_DELAYED_BIT, work_data_bits(work));
1712 1713 1714
	cwq->nr_active++;
}

1715 1716 1717 1718
/**
 * cwq_dec_nr_in_flight - decrement cwq's nr_in_flight
 * @cwq: cwq of interest
 * @color: color of work which left the queue
1719
 * @delayed: for a delayed work
1720 1721 1722 1723 1724
 *
 * A work either has completed or is removed from pending queue,
 * decrement nr_in_flight of its cwq and handle workqueue flushing.
 *
 * CONTEXT:
1725
 * spin_lock_irq(gcwq->lock).
1726
 */
1727 1728
static void cwq_dec_nr_in_flight(struct cpu_workqueue_struct *cwq, int color,
				 bool delayed)
1729 1730 1731 1732 1733 1734
{
	/* ignore uncolored works */
	if (color == WORK_NO_COLOR)
		return;

	cwq->nr_in_flight[color]--;
1735

1736 1737 1738 1739 1740 1741 1742
	if (!delayed) {
		cwq->nr_active--;
		if (!list_empty(&cwq->delayed_works)) {
			/* one down, submit a delayed one */
			if (cwq->nr_active < cwq->max_active)
				cwq_activate_first_delayed(cwq);
		}
1743
	}
1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763

	/* is flush in progress and are we at the flushing tip? */
	if (likely(cwq->flush_color != color))
		return;

	/* are there still in-flight works? */
	if (cwq->nr_in_flight[color])
		return;

	/* this cwq is done, clear flush_color */
	cwq->flush_color = -1;

	/*
	 * If this was the last cwq, wake up the first flusher.  It
	 * will handle the rest.
	 */
	if (atomic_dec_and_test(&cwq->wq->nr_cwqs_to_flush))
		complete(&cwq->wq->first_flusher->done);
}

1764 1765
/**
 * process_one_work - process single work
T
Tejun Heo 已提交
1766
 * @worker: self
1767 1768 1769 1770 1771 1772 1773 1774 1775
 * @work: work to process
 *
 * Process @work.  This function contains all the logics necessary to
 * process a single work including synchronization against and
 * interaction with other workers on the same cpu, queueing and
 * flushing.  As long as context requirement is met, any worker can
 * call this function to process a work.
 *
 * CONTEXT:
1776
 * spin_lock_irq(gcwq->lock) which is released and regrabbed.
1777
 */
T
Tejun Heo 已提交
1778
static void process_one_work(struct worker *worker, struct work_struct *work)
1779 1780
__releases(&gcwq->lock)
__acquires(&gcwq->lock)
1781
{
1782
	struct cpu_workqueue_struct *cwq = get_work_cwq(work);
1783
	struct global_cwq *gcwq = cwq->gcwq;
T
Tejun Heo 已提交
1784
	struct hlist_head *bwh = busy_worker_head(gcwq, work);
1785
	bool cpu_intensive = cwq->wq->flags & WQ_CPU_INTENSIVE;
1786
	work_func_t f = work->func;
1787
	int work_color;
1788
	struct worker *collision;
1789 1790 1791 1792 1793 1794 1795 1796 1797 1798
#ifdef CONFIG_LOCKDEP
	/*
	 * It is permissible to free the struct work_struct from
	 * inside the function that is called from it, this we need to
	 * take into account for lockdep too.  To avoid bogus "held
	 * lock freed" warnings as well as problems when looking into
	 * work->lockdep_map, make a copy and use that here.
	 */
	struct lockdep_map lockdep_map = work->lockdep_map;
#endif
1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810
	/*
	 * A single work shouldn't be executed concurrently by
	 * multiple workers on a single cpu.  Check whether anyone is
	 * already processing the work.  If so, defer the work to the
	 * currently executing one.
	 */
	collision = __find_worker_executing_work(gcwq, bwh, work);
	if (unlikely(collision)) {
		move_linked_works(work, &collision->scheduled, NULL);
		return;
	}

1811 1812
	/* claim and process */
	debug_work_deactivate(work);
T
Tejun Heo 已提交
1813
	hlist_add_head(&worker->hentry, bwh);
T
Tejun Heo 已提交
1814
	worker->current_work = work;
1815
	worker->current_cwq = cwq;
1816
	work_color = get_work_color(work);
1817 1818 1819

	/* record the current cpu number in the work data and dequeue */
	set_work_cpu(work, gcwq->cpu);
1820 1821
	list_del_init(&work->entry);

1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836
	/*
	 * If HIGHPRI_PENDING, check the next work, and, if HIGHPRI,
	 * wake up another worker; otherwise, clear HIGHPRI_PENDING.
	 */
	if (unlikely(gcwq->flags & GCWQ_HIGHPRI_PENDING)) {
		struct work_struct *nwork = list_first_entry(&gcwq->worklist,
						struct work_struct, entry);

		if (!list_empty(&gcwq->worklist) &&
		    get_work_cwq(nwork)->wq->flags & WQ_HIGHPRI)
			wake_up_worker(gcwq);
		else
			gcwq->flags &= ~GCWQ_HIGHPRI_PENDING;
	}

1837 1838 1839 1840 1841 1842 1843
	/*
	 * CPU intensive works don't participate in concurrency
	 * management.  They're the scheduler's responsibility.
	 */
	if (unlikely(cpu_intensive))
		worker_set_flags(worker, WORKER_CPU_INTENSIVE, true);

1844
	spin_unlock_irq(&gcwq->lock);
1845 1846

	work_clear_pending(work);
1847
	lock_map_acquire_read(&cwq->wq->lockdep_map);
1848
	lock_map_acquire(&lockdep_map);
1849
	trace_workqueue_execute_start(work);
1850
	f(work);
1851 1852 1853 1854 1855
	/*
	 * While we must be careful to not use "work" after this, the trace
	 * point will only record its address.
	 */
	trace_workqueue_execute_end(work);
1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868
	lock_map_release(&lockdep_map);
	lock_map_release(&cwq->wq->lockdep_map);

	if (unlikely(in_atomic() || lockdep_depth(current) > 0)) {
		printk(KERN_ERR "BUG: workqueue leaked lock or atomic: "
		       "%s/0x%08x/%d\n",
		       current->comm, preempt_count(), task_pid_nr(current));
		printk(KERN_ERR "    last function: ");
		print_symbol("%s\n", (unsigned long)f);
		debug_show_held_locks(current);
		dump_stack();
	}

1869
	spin_lock_irq(&gcwq->lock);
1870

1871 1872 1873 1874
	/* clear cpu intensive status */
	if (unlikely(cpu_intensive))
		worker_clr_flags(worker, WORKER_CPU_INTENSIVE);

1875
	/* we're done with it, release */
T
Tejun Heo 已提交
1876
	hlist_del_init(&worker->hentry);
T
Tejun Heo 已提交
1877
	worker->current_work = NULL;
1878
	worker->current_cwq = NULL;
1879
	cwq_dec_nr_in_flight(cwq, work_color, false);
1880 1881
}

1882 1883 1884 1885 1886 1887 1888 1889 1890
/**
 * process_scheduled_works - process scheduled works
 * @worker: self
 *
 * Process all scheduled works.  Please note that the scheduled list
 * may change while processing a work, so this function repeatedly
 * fetches a work from the top and executes it.
 *
 * CONTEXT:
1891
 * spin_lock_irq(gcwq->lock) which may be released and regrabbed
1892 1893 1894
 * multiple times.
 */
static void process_scheduled_works(struct worker *worker)
L
Linus Torvalds 已提交
1895
{
1896 1897
	while (!list_empty(&worker->scheduled)) {
		struct work_struct *work = list_first_entry(&worker->scheduled,
L
Linus Torvalds 已提交
1898
						struct work_struct, entry);
T
Tejun Heo 已提交
1899
		process_one_work(worker, work);
L
Linus Torvalds 已提交
1900 1901 1902
	}
}

T
Tejun Heo 已提交
1903 1904
/**
 * worker_thread - the worker thread function
T
Tejun Heo 已提交
1905
 * @__worker: self
T
Tejun Heo 已提交
1906
 *
1907 1908 1909 1910 1911
 * The gcwq worker thread function.  There's a single dynamic pool of
 * these per each cpu.  These workers process all works regardless of
 * their specific target workqueue.  The only exception is works which
 * belong to workqueues with a rescuer which will be explained in
 * rescuer_thread().
T
Tejun Heo 已提交
1912
 */
T
Tejun Heo 已提交
1913
static int worker_thread(void *__worker)
L
Linus Torvalds 已提交
1914
{
T
Tejun Heo 已提交
1915
	struct worker *worker = __worker;
1916
	struct global_cwq *gcwq = worker->gcwq;
L
Linus Torvalds 已提交
1917

1918 1919
	/* tell the scheduler that this is a workqueue worker */
	worker->task->flags |= PF_WQ_WORKER;
T
Tejun Heo 已提交
1920 1921
woke_up:
	spin_lock_irq(&gcwq->lock);
L
Linus Torvalds 已提交
1922

T
Tejun Heo 已提交
1923 1924 1925
	/* DIE can be set only while we're idle, checking here is enough */
	if (worker->flags & WORKER_DIE) {
		spin_unlock_irq(&gcwq->lock);
1926
		worker->task->flags &= ~PF_WQ_WORKER;
T
Tejun Heo 已提交
1927 1928
		return 0;
	}
1929

T
Tejun Heo 已提交
1930
	worker_leave_idle(worker);
1931
recheck:
1932 1933 1934 1935 1936 1937 1938 1939
	/* no more worker necessary? */
	if (!need_more_worker(gcwq))
		goto sleep;

	/* do we need to manage? */
	if (unlikely(!may_start_working(gcwq)) && manage_workers(worker))
		goto recheck;

T
Tejun Heo 已提交
1940 1941 1942 1943 1944 1945 1946
	/*
	 * ->scheduled list can only be filled while a worker is
	 * preparing to process a work or actually processing it.
	 * Make sure nobody diddled with it while I was sleeping.
	 */
	BUG_ON(!list_empty(&worker->scheduled));

1947 1948 1949 1950 1951 1952 1953 1954
	/*
	 * When control reaches this point, we're guaranteed to have
	 * at least one idle worker or that someone else has already
	 * assumed the manager role.
	 */
	worker_clr_flags(worker, WORKER_PREP);

	do {
T
Tejun Heo 已提交
1955
		struct work_struct *work =
1956
			list_first_entry(&gcwq->worklist,
T
Tejun Heo 已提交
1957 1958 1959 1960 1961 1962
					 struct work_struct, entry);

		if (likely(!(*work_data_bits(work) & WORK_STRUCT_LINKED))) {
			/* optimization path, not strictly necessary */
			process_one_work(worker, work);
			if (unlikely(!list_empty(&worker->scheduled)))
1963
				process_scheduled_works(worker);
T
Tejun Heo 已提交
1964 1965 1966
		} else {
			move_linked_works(work, &worker->scheduled, NULL);
			process_scheduled_works(worker);
1967
		}
1968 1969 1970
	} while (keep_working(gcwq));

	worker_set_flags(worker, WORKER_PREP, false);
1971
sleep:
1972 1973
	if (unlikely(need_to_manage_workers(gcwq)) && manage_workers(worker))
		goto recheck;
1974

T
Tejun Heo 已提交
1975
	/*
1976 1977 1978 1979 1980
	 * gcwq->lock is held and there's no work to process and no
	 * need to manage, sleep.  Workers are woken up only while
	 * holding gcwq->lock or from local cpu, so setting the
	 * current state before releasing gcwq->lock is enough to
	 * prevent losing any event.
T
Tejun Heo 已提交
1981 1982 1983 1984 1985 1986
	 */
	worker_enter_idle(worker);
	__set_current_state(TASK_INTERRUPTIBLE);
	spin_unlock_irq(&gcwq->lock);
	schedule();
	goto woke_up;
L
Linus Torvalds 已提交
1987 1988
}

1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
/**
 * rescuer_thread - the rescuer thread function
 * @__wq: the associated workqueue
 *
 * Workqueue rescuer thread function.  There's one rescuer for each
 * workqueue which has WQ_RESCUER set.
 *
 * Regular work processing on a gcwq may block trying to create a new
 * worker which uses GFP_KERNEL allocation which has slight chance of
 * developing into deadlock if some works currently on the same queue
 * need to be processed to satisfy the GFP_KERNEL allocation.  This is
 * the problem rescuer solves.
 *
 * When such condition is possible, the gcwq summons rescuers of all
 * workqueues which have works queued on the gcwq and let them process
 * those works so that forward progress can be guaranteed.
 *
 * This should happen rarely.
 */
static int rescuer_thread(void *__wq)
{
	struct workqueue_struct *wq = __wq;
	struct worker *rescuer = wq->rescuer;
	struct list_head *scheduled = &rescuer->scheduled;
2013
	bool is_unbound = wq->flags & WQ_UNBOUND;
2014 2015 2016 2017 2018 2019 2020 2021 2022
	unsigned int cpu;

	set_user_nice(current, RESCUER_NICE_LEVEL);
repeat:
	set_current_state(TASK_INTERRUPTIBLE);

	if (kthread_should_stop())
		return 0;

2023 2024 2025 2026
	/*
	 * See whether any cpu is asking for help.  Unbounded
	 * workqueues use cpu 0 in mayday_mask for CPU_UNBOUND.
	 */
2027
	for_each_mayday_cpu(cpu, wq->mayday_mask) {
2028 2029
		unsigned int tcpu = is_unbound ? WORK_CPU_UNBOUND : cpu;
		struct cpu_workqueue_struct *cwq = get_cwq(tcpu, wq);
2030 2031 2032 2033
		struct global_cwq *gcwq = cwq->gcwq;
		struct work_struct *work, *n;

		__set_current_state(TASK_RUNNING);
2034
		mayday_clear_cpu(cpu, wq->mayday_mask);
2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054

		/* migrate to the target cpu if possible */
		rescuer->gcwq = gcwq;
		worker_maybe_bind_and_lock(rescuer);

		/*
		 * Slurp in all works issued via this workqueue and
		 * process'em.
		 */
		BUG_ON(!list_empty(&rescuer->scheduled));
		list_for_each_entry_safe(work, n, &gcwq->worklist, entry)
			if (get_work_cwq(work) == cwq)
				move_linked_works(work, scheduled, &n);

		process_scheduled_works(rescuer);
		spin_unlock_irq(&gcwq->lock);
	}

	schedule();
	goto repeat;
L
Linus Torvalds 已提交
2055 2056
}

O
Oleg Nesterov 已提交
2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067
struct wq_barrier {
	struct work_struct	work;
	struct completion	done;
};

static void wq_barrier_func(struct work_struct *work)
{
	struct wq_barrier *barr = container_of(work, struct wq_barrier, work);
	complete(&barr->done);
}

T
Tejun Heo 已提交
2068 2069 2070 2071
/**
 * insert_wq_barrier - insert a barrier work
 * @cwq: cwq to insert barrier into
 * @barr: wq_barrier to insert
2072 2073
 * @target: target work to attach @barr to
 * @worker: worker currently executing @target, NULL if @target is not executing
T
Tejun Heo 已提交
2074
 *
2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087
 * @barr is linked to @target such that @barr is completed only after
 * @target finishes execution.  Please note that the ordering
 * guarantee is observed only with respect to @target and on the local
 * cpu.
 *
 * Currently, a queued barrier can't be canceled.  This is because
 * try_to_grab_pending() can't determine whether the work to be
 * grabbed is at the head of the queue and thus can't clear LINKED
 * flag of the previous work while there must be a valid next work
 * after a work with LINKED flag set.
 *
 * Note that when @worker is non-NULL, @target may be modified
 * underneath us, so we can't reliably determine cwq from @target.
T
Tejun Heo 已提交
2088 2089
 *
 * CONTEXT:
2090
 * spin_lock_irq(gcwq->lock).
T
Tejun Heo 已提交
2091
 */
2092
static void insert_wq_barrier(struct cpu_workqueue_struct *cwq,
2093 2094
			      struct wq_barrier *barr,
			      struct work_struct *target, struct worker *worker)
O
Oleg Nesterov 已提交
2095
{
2096 2097 2098
	struct list_head *head;
	unsigned int linked = 0;

2099
	/*
2100
	 * debugobject calls are safe here even with gcwq->lock locked
2101 2102 2103 2104
	 * as we know for sure that this will not trigger any of the
	 * checks and call back into the fixup functions where we
	 * might deadlock.
	 */
A
Andrew Morton 已提交
2105
	INIT_WORK_ONSTACK(&barr->work, wq_barrier_func);
2106
	__set_bit(WORK_STRUCT_PENDING_BIT, work_data_bits(&barr->work));
O
Oleg Nesterov 已提交
2107
	init_completion(&barr->done);
2108

2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123
	/*
	 * If @target is currently being executed, schedule the
	 * barrier to the worker; otherwise, put it after @target.
	 */
	if (worker)
		head = worker->scheduled.next;
	else {
		unsigned long *bits = work_data_bits(target);

		head = target->entry.next;
		/* there can already be other linked works, inherit and set */
		linked = *bits & WORK_STRUCT_LINKED;
		__set_bit(WORK_STRUCT_LINKED_BIT, bits);
	}

2124
	debug_work_activate(&barr->work);
2125 2126
	insert_work(cwq, &barr->work, head,
		    work_color_to_flags(WORK_NO_COLOR) | linked);
O
Oleg Nesterov 已提交
2127 2128
}

2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161
/**
 * flush_workqueue_prep_cwqs - prepare cwqs for workqueue flushing
 * @wq: workqueue being flushed
 * @flush_color: new flush color, < 0 for no-op
 * @work_color: new work color, < 0 for no-op
 *
 * Prepare cwqs for workqueue flushing.
 *
 * If @flush_color is non-negative, flush_color on all cwqs should be
 * -1.  If no cwq has in-flight commands at the specified color, all
 * cwq->flush_color's stay at -1 and %false is returned.  If any cwq
 * has in flight commands, its cwq->flush_color is set to
 * @flush_color, @wq->nr_cwqs_to_flush is updated accordingly, cwq
 * wakeup logic is armed and %true is returned.
 *
 * The caller should have initialized @wq->first_flusher prior to
 * calling this function with non-negative @flush_color.  If
 * @flush_color is negative, no flush color update is done and %false
 * is returned.
 *
 * If @work_color is non-negative, all cwqs should have the same
 * work_color which is previous to @work_color and all will be
 * advanced to @work_color.
 *
 * CONTEXT:
 * mutex_lock(wq->flush_mutex).
 *
 * RETURNS:
 * %true if @flush_color >= 0 and there's something to flush.  %false
 * otherwise.
 */
static bool flush_workqueue_prep_cwqs(struct workqueue_struct *wq,
				      int flush_color, int work_color)
L
Linus Torvalds 已提交
2162
{
2163 2164
	bool wait = false;
	unsigned int cpu;
L
Linus Torvalds 已提交
2165

2166 2167 2168
	if (flush_color >= 0) {
		BUG_ON(atomic_read(&wq->nr_cwqs_to_flush));
		atomic_set(&wq->nr_cwqs_to_flush, 1);
L
Linus Torvalds 已提交
2169
	}
2170

2171
	for_each_cwq_cpu(cpu, wq) {
2172
		struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq);
2173
		struct global_cwq *gcwq = cwq->gcwq;
O
Oleg Nesterov 已提交
2174

2175
		spin_lock_irq(&gcwq->lock);
2176

2177 2178
		if (flush_color >= 0) {
			BUG_ON(cwq->flush_color != -1);
O
Oleg Nesterov 已提交
2179

2180 2181 2182 2183 2184 2185
			if (cwq->nr_in_flight[flush_color]) {
				cwq->flush_color = flush_color;
				atomic_inc(&wq->nr_cwqs_to_flush);
				wait = true;
			}
		}
L
Linus Torvalds 已提交
2186

2187 2188 2189 2190
		if (work_color >= 0) {
			BUG_ON(work_color != work_next_color(cwq->work_color));
			cwq->work_color = work_color;
		}
L
Linus Torvalds 已提交
2191

2192
		spin_unlock_irq(&gcwq->lock);
L
Linus Torvalds 已提交
2193
	}
2194

2195 2196
	if (flush_color >= 0 && atomic_dec_and_test(&wq->nr_cwqs_to_flush))
		complete(&wq->first_flusher->done);
2197

2198
	return wait;
L
Linus Torvalds 已提交
2199 2200
}

2201
/**
L
Linus Torvalds 已提交
2202
 * flush_workqueue - ensure that any scheduled work has run to completion.
2203
 * @wq: workqueue to flush
L
Linus Torvalds 已提交
2204 2205 2206 2207
 *
 * Forces execution of the workqueue and blocks until its completion.
 * This is typically used in driver shutdown handlers.
 *
O
Oleg Nesterov 已提交
2208 2209
 * We sleep until all works which were queued on entry have been handled,
 * but we are not livelocked by new incoming ones.
L
Linus Torvalds 已提交
2210
 */
2211
void flush_workqueue(struct workqueue_struct *wq)
L
Linus Torvalds 已提交
2212
{
2213 2214 2215 2216 2217 2218
	struct wq_flusher this_flusher = {
		.list = LIST_HEAD_INIT(this_flusher.list),
		.flush_color = -1,
		.done = COMPLETION_INITIALIZER_ONSTACK(this_flusher.done),
	};
	int next_color;
L
Linus Torvalds 已提交
2219

2220 2221
	lock_map_acquire(&wq->lockdep_map);
	lock_map_release(&wq->lockdep_map);
2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282

	mutex_lock(&wq->flush_mutex);

	/*
	 * Start-to-wait phase
	 */
	next_color = work_next_color(wq->work_color);

	if (next_color != wq->flush_color) {
		/*
		 * Color space is not full.  The current work_color
		 * becomes our flush_color and work_color is advanced
		 * by one.
		 */
		BUG_ON(!list_empty(&wq->flusher_overflow));
		this_flusher.flush_color = wq->work_color;
		wq->work_color = next_color;

		if (!wq->first_flusher) {
			/* no flush in progress, become the first flusher */
			BUG_ON(wq->flush_color != this_flusher.flush_color);

			wq->first_flusher = &this_flusher;

			if (!flush_workqueue_prep_cwqs(wq, wq->flush_color,
						       wq->work_color)) {
				/* nothing to flush, done */
				wq->flush_color = next_color;
				wq->first_flusher = NULL;
				goto out_unlock;
			}
		} else {
			/* wait in queue */
			BUG_ON(wq->flush_color == this_flusher.flush_color);
			list_add_tail(&this_flusher.list, &wq->flusher_queue);
			flush_workqueue_prep_cwqs(wq, -1, wq->work_color);
		}
	} else {
		/*
		 * Oops, color space is full, wait on overflow queue.
		 * The next flush completion will assign us
		 * flush_color and transfer to flusher_queue.
		 */
		list_add_tail(&this_flusher.list, &wq->flusher_overflow);
	}

	mutex_unlock(&wq->flush_mutex);

	wait_for_completion(&this_flusher.done);

	/*
	 * Wake-up-and-cascade phase
	 *
	 * First flushers are responsible for cascading flushes and
	 * handling overflow.  Non-first flushers can simply return.
	 */
	if (wq->first_flusher != &this_flusher)
		return;

	mutex_lock(&wq->flush_mutex);

2283 2284 2285 2286
	/* we might have raced, check again with mutex held */
	if (wq->first_flusher != &this_flusher)
		goto out_unlock;

2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353
	wq->first_flusher = NULL;

	BUG_ON(!list_empty(&this_flusher.list));
	BUG_ON(wq->flush_color != this_flusher.flush_color);

	while (true) {
		struct wq_flusher *next, *tmp;

		/* complete all the flushers sharing the current flush color */
		list_for_each_entry_safe(next, tmp, &wq->flusher_queue, list) {
			if (next->flush_color != wq->flush_color)
				break;
			list_del_init(&next->list);
			complete(&next->done);
		}

		BUG_ON(!list_empty(&wq->flusher_overflow) &&
		       wq->flush_color != work_next_color(wq->work_color));

		/* this flush_color is finished, advance by one */
		wq->flush_color = work_next_color(wq->flush_color);

		/* one color has been freed, handle overflow queue */
		if (!list_empty(&wq->flusher_overflow)) {
			/*
			 * Assign the same color to all overflowed
			 * flushers, advance work_color and append to
			 * flusher_queue.  This is the start-to-wait
			 * phase for these overflowed flushers.
			 */
			list_for_each_entry(tmp, &wq->flusher_overflow, list)
				tmp->flush_color = wq->work_color;

			wq->work_color = work_next_color(wq->work_color);

			list_splice_tail_init(&wq->flusher_overflow,
					      &wq->flusher_queue);
			flush_workqueue_prep_cwqs(wq, -1, wq->work_color);
		}

		if (list_empty(&wq->flusher_queue)) {
			BUG_ON(wq->flush_color != wq->work_color);
			break;
		}

		/*
		 * Need to flush more colors.  Make the next flusher
		 * the new first flusher and arm cwqs.
		 */
		BUG_ON(wq->flush_color == wq->work_color);
		BUG_ON(wq->flush_color != next->flush_color);

		list_del_init(&next->list);
		wq->first_flusher = next;

		if (flush_workqueue_prep_cwqs(wq, wq->flush_color, -1))
			break;

		/*
		 * Meh... this color is already done, clear first
		 * flusher and repeat cascading.
		 */
		wq->first_flusher = NULL;
	}

out_unlock:
	mutex_unlock(&wq->flush_mutex);
L
Linus Torvalds 已提交
2354
}
2355
EXPORT_SYMBOL_GPL(flush_workqueue);
L
Linus Torvalds 已提交
2356

2357 2358
static bool start_flush_work(struct work_struct *work, struct wq_barrier *barr,
			     bool wait_executing)
2359
{
2360
	struct worker *worker = NULL;
2361
	struct global_cwq *gcwq;
2362 2363 2364
	struct cpu_workqueue_struct *cwq;

	might_sleep();
2365 2366
	gcwq = get_work_gcwq(work);
	if (!gcwq)
2367
		return false;
2368

2369
	spin_lock_irq(&gcwq->lock);
2370 2371 2372
	if (!list_empty(&work->entry)) {
		/*
		 * See the comment near try_to_grab_pending()->smp_rmb().
2373 2374
		 * If it was re-queued to a different gcwq under us, we
		 * are not going to wait.
2375 2376
		 */
		smp_rmb();
2377 2378
		cwq = get_work_cwq(work);
		if (unlikely(!cwq || gcwq != cwq->gcwq))
T
Tejun Heo 已提交
2379
			goto already_gone;
2380
	} else if (wait_executing) {
2381
		worker = find_worker_executing_work(gcwq, work);
2382
		if (!worker)
T
Tejun Heo 已提交
2383
			goto already_gone;
2384
		cwq = worker->current_cwq;
2385 2386
	} else
		goto already_gone;
2387

2388
	insert_wq_barrier(cwq, barr, work, worker);
2389
	spin_unlock_irq(&gcwq->lock);
2390

2391 2392 2393 2394 2395 2396 2397 2398 2399 2400
	/*
	 * If @max_active is 1 or rescuer is in use, flushing another work
	 * item on the same workqueue may lead to deadlock.  Make sure the
	 * flusher is not running on the same workqueue by verifying write
	 * access.
	 */
	if (cwq->wq->saved_max_active == 1 || cwq->wq->flags & WQ_RESCUER)
		lock_map_acquire(&cwq->wq->lockdep_map);
	else
		lock_map_acquire_read(&cwq->wq->lockdep_map);
2401
	lock_map_release(&cwq->wq->lockdep_map);
2402

2403
	return true;
T
Tejun Heo 已提交
2404
already_gone:
2405
	spin_unlock_irq(&gcwq->lock);
2406
	return false;
2407
}
2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437

/**
 * flush_work - wait for a work to finish executing the last queueing instance
 * @work: the work to flush
 *
 * Wait until @work has finished execution.  This function considers
 * only the last queueing instance of @work.  If @work has been
 * enqueued across different CPUs on a non-reentrant workqueue or on
 * multiple workqueues, @work might still be executing on return on
 * some of the CPUs from earlier queueing.
 *
 * If @work was queued only on a non-reentrant, ordered or unbound
 * workqueue, @work is guaranteed to be idle on return if it hasn't
 * been requeued since flush started.
 *
 * RETURNS:
 * %true if flush_work() waited for the work to finish execution,
 * %false if it was already idle.
 */
bool flush_work(struct work_struct *work)
{
	struct wq_barrier barr;

	if (start_flush_work(work, &barr, true)) {
		wait_for_completion(&barr.done);
		destroy_work_on_stack(&barr.work);
		return true;
	} else
		return false;
}
2438 2439
EXPORT_SYMBOL_GPL(flush_work);

2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475
static bool wait_on_cpu_work(struct global_cwq *gcwq, struct work_struct *work)
{
	struct wq_barrier barr;
	struct worker *worker;

	spin_lock_irq(&gcwq->lock);

	worker = find_worker_executing_work(gcwq, work);
	if (unlikely(worker))
		insert_wq_barrier(worker->current_cwq, &barr, work, worker);

	spin_unlock_irq(&gcwq->lock);

	if (unlikely(worker)) {
		wait_for_completion(&barr.done);
		destroy_work_on_stack(&barr.work);
		return true;
	} else
		return false;
}

static bool wait_on_work(struct work_struct *work)
{
	bool ret = false;
	int cpu;

	might_sleep();

	lock_map_acquire(&work->lockdep_map);
	lock_map_release(&work->lockdep_map);

	for_each_gcwq_cpu(cpu)
		ret |= wait_on_cpu_work(get_gcwq(cpu), work);
	return ret;
}

2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510
/**
 * flush_work_sync - wait until a work has finished execution
 * @work: the work to flush
 *
 * Wait until @work has finished execution.  On return, it's
 * guaranteed that all queueing instances of @work which happened
 * before this function is called are finished.  In other words, if
 * @work hasn't been requeued since this function was called, @work is
 * guaranteed to be idle on return.
 *
 * RETURNS:
 * %true if flush_work_sync() waited for the work to finish execution,
 * %false if it was already idle.
 */
bool flush_work_sync(struct work_struct *work)
{
	struct wq_barrier barr;
	bool pending, waited;

	/* we'll wait for executions separately, queue barr only if pending */
	pending = start_flush_work(work, &barr, false);

	/* wait for executions to finish */
	waited = wait_on_work(work);

	/* wait for the pending one */
	if (pending) {
		wait_for_completion(&barr.done);
		destroy_work_on_stack(&barr.work);
	}

	return pending || waited;
}
EXPORT_SYMBOL_GPL(flush_work_sync);

2511
/*
2512
 * Upon a successful return (>= 0), the caller "owns" WORK_STRUCT_PENDING bit,
2513 2514 2515 2516
 * so this work can't be re-armed in any way.
 */
static int try_to_grab_pending(struct work_struct *work)
{
2517
	struct global_cwq *gcwq;
2518
	int ret = -1;
2519

2520
	if (!test_and_set_bit(WORK_STRUCT_PENDING_BIT, work_data_bits(work)))
2521
		return 0;
2522 2523 2524 2525 2526

	/*
	 * The queueing is in progress, or it is already queued. Try to
	 * steal it from ->worklist without clearing WORK_STRUCT_PENDING.
	 */
2527 2528
	gcwq = get_work_gcwq(work);
	if (!gcwq)
2529 2530
		return ret;

2531
	spin_lock_irq(&gcwq->lock);
2532 2533
	if (!list_empty(&work->entry)) {
		/*
2534
		 * This work is queued, but perhaps we locked the wrong gcwq.
2535 2536 2537 2538
		 * In that case we must see the new value after rmb(), see
		 * insert_work()->wmb().
		 */
		smp_rmb();
2539
		if (gcwq == get_work_gcwq(work)) {
2540
			debug_work_deactivate(work);
2541
			list_del_init(&work->entry);
2542
			cwq_dec_nr_in_flight(get_work_cwq(work),
2543 2544
				get_work_color(work),
				*work_data_bits(work) & WORK_STRUCT_DELAYED);
2545 2546 2547
			ret = 1;
		}
	}
2548
	spin_unlock_irq(&gcwq->lock);
2549 2550 2551 2552

	return ret;
}

2553
static bool __cancel_work_timer(struct work_struct *work,
2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564
				struct timer_list* timer)
{
	int ret;

	do {
		ret = (timer && likely(del_timer(timer)));
		if (!ret)
			ret = try_to_grab_pending(work);
		wait_on_work(work);
	} while (unlikely(ret < 0));

2565
	clear_work_data(work);
2566 2567 2568
	return ret;
}

2569
/**
2570 2571
 * cancel_work_sync - cancel a work and wait for it to finish
 * @work: the work to cancel
2572
 *
2573 2574 2575 2576
 * Cancel @work and wait for its execution to finish.  This function
 * can be used even if the work re-queues itself or migrates to
 * another workqueue.  On return from this function, @work is
 * guaranteed to be not pending or executing on any CPU.
2577
 *
2578 2579
 * cancel_work_sync(&delayed_work->work) must not be used for
 * delayed_work's.  Use cancel_delayed_work_sync() instead.
2580
 *
2581
 * The caller must ensure that the workqueue on which @work was last
2582
 * queued can't be destroyed before this function returns.
2583 2584 2585
 *
 * RETURNS:
 * %true if @work was pending, %false otherwise.
2586
 */
2587
bool cancel_work_sync(struct work_struct *work)
2588
{
2589
	return __cancel_work_timer(work, NULL);
O
Oleg Nesterov 已提交
2590
}
2591
EXPORT_SYMBOL_GPL(cancel_work_sync);
O
Oleg Nesterov 已提交
2592

2593
/**
2594 2595
 * flush_delayed_work - wait for a dwork to finish executing the last queueing
 * @dwork: the delayed work to flush
2596
 *
2597 2598 2599
 * Delayed timer is cancelled and the pending work is queued for
 * immediate execution.  Like flush_work(), this function only
 * considers the last queueing instance of @dwork.
2600
 *
2601 2602 2603
 * RETURNS:
 * %true if flush_work() waited for the work to finish execution,
 * %false if it was already idle.
2604
 */
2605 2606 2607 2608 2609 2610 2611 2612 2613
bool flush_delayed_work(struct delayed_work *dwork)
{
	if (del_timer_sync(&dwork->timer))
		__queue_work(raw_smp_processor_id(),
			     get_work_cwq(&dwork->work)->wq, &dwork->work);
	return flush_work(&dwork->work);
}
EXPORT_SYMBOL(flush_delayed_work);

2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634
/**
 * flush_delayed_work_sync - wait for a dwork to finish
 * @dwork: the delayed work to flush
 *
 * Delayed timer is cancelled and the pending work is queued for
 * execution immediately.  Other than timer handling, its behavior
 * is identical to flush_work_sync().
 *
 * RETURNS:
 * %true if flush_work_sync() waited for the work to finish execution,
 * %false if it was already idle.
 */
bool flush_delayed_work_sync(struct delayed_work *dwork)
{
	if (del_timer_sync(&dwork->timer))
		__queue_work(raw_smp_processor_id(),
			     get_work_cwq(&dwork->work)->wq, &dwork->work);
	return flush_work_sync(&dwork->work);
}
EXPORT_SYMBOL(flush_delayed_work_sync);

2635 2636 2637 2638 2639 2640 2641 2642 2643 2644
/**
 * cancel_delayed_work_sync - cancel a delayed work and wait for it to finish
 * @dwork: the delayed work cancel
 *
 * This is cancel_work_sync() for delayed works.
 *
 * RETURNS:
 * %true if @dwork was pending, %false otherwise.
 */
bool cancel_delayed_work_sync(struct delayed_work *dwork)
2645
{
2646
	return __cancel_work_timer(&dwork->work, &dwork->timer);
2647
}
2648
EXPORT_SYMBOL(cancel_delayed_work_sync);
L
Linus Torvalds 已提交
2649

2650 2651 2652 2653
/**
 * schedule_work - put work task in global workqueue
 * @work: job to be done
 *
2654 2655 2656 2657 2658 2659
 * Returns zero if @work was already on the kernel-global workqueue and
 * non-zero otherwise.
 *
 * This puts a job in the kernel-global workqueue if it was not already
 * queued and leaves it in the same position on the kernel-global
 * workqueue otherwise.
2660
 */
2661
int schedule_work(struct work_struct *work)
L
Linus Torvalds 已提交
2662
{
2663
	return queue_work(system_wq, work);
L
Linus Torvalds 已提交
2664
}
2665
EXPORT_SYMBOL(schedule_work);
L
Linus Torvalds 已提交
2666

2667 2668 2669 2670 2671 2672 2673 2674 2675
/*
 * schedule_work_on - put work task on a specific cpu
 * @cpu: cpu to put the work task on
 * @work: job to be done
 *
 * This puts a job on a specific cpu
 */
int schedule_work_on(int cpu, struct work_struct *work)
{
2676
	return queue_work_on(cpu, system_wq, work);
2677 2678 2679
}
EXPORT_SYMBOL(schedule_work_on);

2680 2681
/**
 * schedule_delayed_work - put work task in global workqueue after delay
2682 2683
 * @dwork: job to be done
 * @delay: number of jiffies to wait or 0 for immediate execution
2684 2685 2686 2687
 *
 * After waiting for a given time this puts a job in the kernel-global
 * workqueue.
 */
2688
int schedule_delayed_work(struct delayed_work *dwork,
2689
					unsigned long delay)
L
Linus Torvalds 已提交
2690
{
2691
	return queue_delayed_work(system_wq, dwork, delay);
L
Linus Torvalds 已提交
2692
}
2693
EXPORT_SYMBOL(schedule_delayed_work);
L
Linus Torvalds 已提交
2694

2695 2696 2697
/**
 * schedule_delayed_work_on - queue work in global workqueue on CPU after delay
 * @cpu: cpu to use
2698
 * @dwork: job to be done
2699 2700 2701 2702 2703
 * @delay: number of jiffies to wait
 *
 * After waiting for a given time this puts a job in the kernel-global
 * workqueue on the specified CPU.
 */
L
Linus Torvalds 已提交
2704
int schedule_delayed_work_on(int cpu,
2705
			struct delayed_work *dwork, unsigned long delay)
L
Linus Torvalds 已提交
2706
{
2707
	return queue_delayed_work_on(cpu, system_wq, dwork, delay);
L
Linus Torvalds 已提交
2708
}
2709
EXPORT_SYMBOL(schedule_delayed_work_on);
L
Linus Torvalds 已提交
2710

2711
/**
2712
 * schedule_on_each_cpu - execute a function synchronously on each online CPU
2713 2714
 * @func: the function to call
 *
2715 2716
 * schedule_on_each_cpu() executes @func on each online CPU using the
 * system workqueue and blocks until all CPUs have completed.
2717
 * schedule_on_each_cpu() is very slow.
2718 2719 2720
 *
 * RETURNS:
 * 0 on success, -errno on failure.
2721
 */
2722
int schedule_on_each_cpu(work_func_t func)
2723 2724
{
	int cpu;
2725
	struct work_struct __percpu *works;
2726

2727 2728
	works = alloc_percpu(struct work_struct);
	if (!works)
2729
		return -ENOMEM;
2730

2731 2732
	get_online_cpus();

2733
	for_each_online_cpu(cpu) {
2734 2735 2736
		struct work_struct *work = per_cpu_ptr(works, cpu);

		INIT_WORK(work, func);
2737
		schedule_work_on(cpu, work);
2738
	}
2739 2740 2741 2742

	for_each_online_cpu(cpu)
		flush_work(per_cpu_ptr(works, cpu));

2743
	put_online_cpus();
2744
	free_percpu(works);
2745 2746 2747
	return 0;
}

2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771
/**
 * flush_scheduled_work - ensure that any scheduled work has run to completion.
 *
 * Forces execution of the kernel-global workqueue and blocks until its
 * completion.
 *
 * Think twice before calling this function!  It's very easy to get into
 * trouble if you don't take great care.  Either of the following situations
 * will lead to deadlock:
 *
 *	One of the work items currently on the workqueue needs to acquire
 *	a lock held by your code or its caller.
 *
 *	Your code is running in the context of a work routine.
 *
 * They will be detected by lockdep when they occur, but the first might not
 * occur very often.  It depends on what work items are on the workqueue and
 * what locks they need, which you have no control over.
 *
 * In most situations flushing the entire workqueue is overkill; you merely
 * need to know that a particular work item isn't queued and isn't running.
 * In such cases you should use cancel_delayed_work_sync() or
 * cancel_work_sync() instead.
 */
L
Linus Torvalds 已提交
2772 2773
void flush_scheduled_work(void)
{
2774
	flush_workqueue(system_wq);
L
Linus Torvalds 已提交
2775
}
2776
EXPORT_SYMBOL(flush_scheduled_work);
L
Linus Torvalds 已提交
2777

2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789
/**
 * execute_in_process_context - reliably execute the routine with user context
 * @fn:		the function to execute
 * @ew:		guaranteed storage for the execute work structure (must
 *		be available when the work executes)
 *
 * Executes the function immediately if process context is available,
 * otherwise schedules the function for delayed execution.
 *
 * Returns:	0 - function was executed
 *		1 - function was scheduled for execution
 */
2790
int execute_in_process_context(work_func_t fn, struct execute_work *ew)
2791 2792
{
	if (!in_interrupt()) {
2793
		fn(&ew->work);
2794 2795 2796
		return 0;
	}

2797
	INIT_WORK(&ew->work, fn);
2798 2799 2800 2801 2802 2803
	schedule_work(&ew->work);

	return 1;
}
EXPORT_SYMBOL_GPL(execute_in_process_context);

L
Linus Torvalds 已提交
2804 2805
int keventd_up(void)
{
2806
	return system_wq != NULL;
L
Linus Torvalds 已提交
2807 2808
}

2809
static int alloc_cwqs(struct workqueue_struct *wq)
T
Tejun Heo 已提交
2810
{
2811
	/*
T
Tejun Heo 已提交
2812 2813 2814
	 * cwqs are forced aligned according to WORK_STRUCT_FLAG_BITS.
	 * Make sure that the alignment isn't lower than that of
	 * unsigned long long.
2815
	 */
T
Tejun Heo 已提交
2816 2817 2818
	const size_t size = sizeof(struct cpu_workqueue_struct);
	const size_t align = max_t(size_t, 1 << WORK_STRUCT_FLAG_BITS,
				   __alignof__(unsigned long long));
2819 2820 2821 2822 2823
#ifdef CONFIG_SMP
	bool percpu = !(wq->flags & WQ_UNBOUND);
#else
	bool percpu = false;
#endif
2824

2825
	if (percpu)
2826
		wq->cpu_wq.pcpu = __alloc_percpu(size, align);
2827
	else {
2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839
		void *ptr;

		/*
		 * Allocate enough room to align cwq and put an extra
		 * pointer at the end pointing back to the originally
		 * allocated pointer which will be used for free.
		 */
		ptr = kzalloc(size + align + sizeof(void *), GFP_KERNEL);
		if (ptr) {
			wq->cpu_wq.single = PTR_ALIGN(ptr, align);
			*(void **)(wq->cpu_wq.single + 1) = ptr;
		}
2840
	}
2841

2842 2843 2844
	/* just in case, make sure it's actually aligned
	 * - this is affected by PERCPU() alignment in vmlinux.lds.S
	 */
2845 2846
	BUG_ON(!IS_ALIGNED(wq->cpu_wq.v, align));
	return wq->cpu_wq.v ? 0 : -ENOMEM;
T
Tejun Heo 已提交
2847 2848
}

2849
static void free_cwqs(struct workqueue_struct *wq)
T
Tejun Heo 已提交
2850
{
2851 2852 2853 2854 2855 2856 2857
#ifdef CONFIG_SMP
	bool percpu = !(wq->flags & WQ_UNBOUND);
#else
	bool percpu = false;
#endif

	if (percpu)
2858 2859 2860
		free_percpu(wq->cpu_wq.pcpu);
	else if (wq->cpu_wq.single) {
		/* the pointer to free is stored right after the cwq */
2861
		kfree(*(void **)(wq->cpu_wq.single + 1));
2862
	}
T
Tejun Heo 已提交
2863 2864
}

2865 2866
static int wq_clamp_max_active(int max_active, unsigned int flags,
			       const char *name)
2867
{
2868 2869 2870
	int lim = flags & WQ_UNBOUND ? WQ_UNBOUND_MAX_ACTIVE : WQ_MAX_ACTIVE;

	if (max_active < 1 || max_active > lim)
2871 2872
		printk(KERN_WARNING "workqueue: max_active %d requested for %s "
		       "is out of range, clamping between %d and %d\n",
2873
		       max_active, name, 1, lim);
2874

2875
	return clamp_val(max_active, 1, lim);
2876 2877
}

2878 2879 2880 2881 2882
struct workqueue_struct *__alloc_workqueue_key(const char *name,
					       unsigned int flags,
					       int max_active,
					       struct lock_class_key *key,
					       const char *lock_name)
L
Linus Torvalds 已提交
2883 2884
{
	struct workqueue_struct *wq;
T
Tejun Heo 已提交
2885
	unsigned int cpu;
L
Linus Torvalds 已提交
2886

2887 2888 2889 2890 2891 2892 2893
	/*
	 * Workqueues which may be used during memory reclaim should
	 * have a rescuer to guarantee forward progress.
	 */
	if (flags & WQ_MEM_RECLAIM)
		flags |= WQ_RESCUER;

2894 2895 2896 2897 2898 2899 2900
	/*
	 * Unbound workqueues aren't concurrency managed and should be
	 * dispatched to workers immediately.
	 */
	if (flags & WQ_UNBOUND)
		flags |= WQ_HIGHPRI;

2901
	max_active = max_active ?: WQ_DFL_ACTIVE;
2902
	max_active = wq_clamp_max_active(max_active, flags, name);
2903

2904 2905
	wq = kzalloc(sizeof(*wq), GFP_KERNEL);
	if (!wq)
T
Tejun Heo 已提交
2906
		goto err;
2907

2908
	wq->flags = flags;
2909
	wq->saved_max_active = max_active;
2910 2911 2912 2913
	mutex_init(&wq->flush_mutex);
	atomic_set(&wq->nr_cwqs_to_flush, 0);
	INIT_LIST_HEAD(&wq->flusher_queue);
	INIT_LIST_HEAD(&wq->flusher_overflow);
2914

2915
	wq->name = name;
2916
	lockdep_init_map(&wq->lockdep_map, lock_name, key, 0);
2917
	INIT_LIST_HEAD(&wq->list);
2918

2919 2920 2921
	if (alloc_cwqs(wq) < 0)
		goto err;

2922
	for_each_cwq_cpu(cpu, wq) {
T
Tejun Heo 已提交
2923
		struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq);
2924
		struct global_cwq *gcwq = get_gcwq(cpu);
T
Tejun Heo 已提交
2925

T
Tejun Heo 已提交
2926
		BUG_ON((unsigned long)cwq & WORK_STRUCT_FLAG_MASK);
2927
		cwq->gcwq = gcwq;
T
Tejun Heo 已提交
2928
		cwq->wq = wq;
2929
		cwq->flush_color = -1;
2930 2931
		cwq->max_active = max_active;
		INIT_LIST_HEAD(&cwq->delayed_works);
2932
	}
T
Tejun Heo 已提交
2933

2934 2935 2936
	if (flags & WQ_RESCUER) {
		struct worker *rescuer;

2937
		if (!alloc_mayday_mask(&wq->mayday_mask, GFP_KERNEL))
2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949
			goto err;

		wq->rescuer = rescuer = alloc_worker();
		if (!rescuer)
			goto err;

		rescuer->task = kthread_create(rescuer_thread, wq, "%s", name);
		if (IS_ERR(rescuer->task))
			goto err;

		rescuer->task->flags |= PF_THREAD_BOUND;
		wake_up_process(rescuer->task);
2950 2951
	}

2952 2953 2954 2955 2956
	/*
	 * workqueue_lock protects global freeze state and workqueues
	 * list.  Grab it, set max_active accordingly and add the new
	 * workqueue to workqueues list.
	 */
T
Tejun Heo 已提交
2957
	spin_lock(&workqueue_lock);
2958 2959

	if (workqueue_freezing && wq->flags & WQ_FREEZEABLE)
2960
		for_each_cwq_cpu(cpu, wq)
2961 2962
			get_cwq(cpu, wq)->max_active = 0;

T
Tejun Heo 已提交
2963
	list_add(&wq->list, &workqueues);
2964

T
Tejun Heo 已提交
2965 2966
	spin_unlock(&workqueue_lock);

2967
	return wq;
T
Tejun Heo 已提交
2968 2969
err:
	if (wq) {
2970
		free_cwqs(wq);
2971
		free_mayday_mask(wq->mayday_mask);
2972
		kfree(wq->rescuer);
T
Tejun Heo 已提交
2973 2974 2975
		kfree(wq);
	}
	return NULL;
2976
}
2977
EXPORT_SYMBOL_GPL(__alloc_workqueue_key);
L
Linus Torvalds 已提交
2978

2979 2980 2981 2982 2983 2984 2985 2986
/**
 * destroy_workqueue - safely terminate a workqueue
 * @wq: target workqueue
 *
 * Safely destroy a workqueue. All work currently pending will be done first.
 */
void destroy_workqueue(struct workqueue_struct *wq)
{
2987
	unsigned int flush_cnt = 0;
T
Tejun Heo 已提交
2988
	unsigned int cpu;
2989

2990 2991 2992 2993 2994 2995 2996 2997
	/*
	 * Mark @wq dying and drain all pending works.  Once WQ_DYING is
	 * set, only chain queueing is allowed.  IOW, only currently
	 * pending or running work items on @wq can queue further work
	 * items on it.  @wq is flushed repeatedly until it becomes empty.
	 * The number of flushing is detemined by the depth of chaining and
	 * should be relatively short.  Whine if it takes too long.
	 */
2998
	wq->flags |= WQ_DYING;
2999
reflush:
3000 3001
	flush_workqueue(wq);

3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015
	for_each_cwq_cpu(cpu, wq) {
		struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq);

		if (!cwq->nr_active && list_empty(&cwq->delayed_works))
			continue;

		if (++flush_cnt == 10 ||
		    (flush_cnt % 100 == 0 && flush_cnt <= 1000))
			printk(KERN_WARNING "workqueue %s: flush on "
			       "destruction isn't complete after %u tries\n",
			       wq->name, flush_cnt);
		goto reflush;
	}

3016 3017 3018 3019
	/*
	 * wq list is used to freeze wq, remove from list after
	 * flushing is complete in case freeze races us.
	 */
3020
	spin_lock(&workqueue_lock);
3021
	list_del(&wq->list);
3022
	spin_unlock(&workqueue_lock);
3023

3024
	/* sanity check */
3025
	for_each_cwq_cpu(cpu, wq) {
3026 3027 3028 3029 3030
		struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq);
		int i;

		for (i = 0; i < WORK_NR_COLORS; i++)
			BUG_ON(cwq->nr_in_flight[i]);
3031 3032
		BUG_ON(cwq->nr_active);
		BUG_ON(!list_empty(&cwq->delayed_works));
3033
	}
3034

3035 3036
	if (wq->flags & WQ_RESCUER) {
		kthread_stop(wq->rescuer->task);
3037
		free_mayday_mask(wq->mayday_mask);
3038
		kfree(wq->rescuer);
3039 3040
	}

3041
	free_cwqs(wq);
3042 3043 3044 3045
	kfree(wq);
}
EXPORT_SYMBOL_GPL(destroy_workqueue);

3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059
/**
 * workqueue_set_max_active - adjust max_active of a workqueue
 * @wq: target workqueue
 * @max_active: new max_active value.
 *
 * Set max_active of @wq to @max_active.
 *
 * CONTEXT:
 * Don't call from IRQ context.
 */
void workqueue_set_max_active(struct workqueue_struct *wq, int max_active)
{
	unsigned int cpu;

3060
	max_active = wq_clamp_max_active(max_active, wq->flags, wq->name);
3061 3062 3063 3064 3065

	spin_lock(&workqueue_lock);

	wq->saved_max_active = max_active;

3066
	for_each_cwq_cpu(cpu, wq) {
3067 3068 3069 3070 3071 3072 3073
		struct global_cwq *gcwq = get_gcwq(cpu);

		spin_lock_irq(&gcwq->lock);

		if (!(wq->flags & WQ_FREEZEABLE) ||
		    !(gcwq->flags & GCWQ_FREEZING))
			get_cwq(gcwq->cpu, wq)->max_active = max_active;
3074

3075
		spin_unlock_irq(&gcwq->lock);
3076
	}
3077

3078
	spin_unlock(&workqueue_lock);
3079
}
3080
EXPORT_SYMBOL_GPL(workqueue_set_max_active);
3081

3082
/**
3083 3084 3085
 * workqueue_congested - test whether a workqueue is congested
 * @cpu: CPU in question
 * @wq: target workqueue
3086
 *
3087 3088 3089
 * Test whether @wq's cpu workqueue for @cpu is congested.  There is
 * no synchronization around this function and the test result is
 * unreliable and only useful as advisory hints or for debugging.
3090
 *
3091 3092
 * RETURNS:
 * %true if congested, %false otherwise.
3093
 */
3094
bool workqueue_congested(unsigned int cpu, struct workqueue_struct *wq)
L
Linus Torvalds 已提交
3095
{
3096 3097 3098
	struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq);

	return !list_empty(&cwq->delayed_works);
L
Linus Torvalds 已提交
3099
}
3100
EXPORT_SYMBOL_GPL(workqueue_congested);
L
Linus Torvalds 已提交
3101

3102
/**
3103 3104
 * work_cpu - return the last known associated cpu for @work
 * @work: the work of interest
3105
 *
3106
 * RETURNS:
3107
 * CPU number if @work was ever queued.  WORK_CPU_NONE otherwise.
3108
 */
3109
unsigned int work_cpu(struct work_struct *work)
3110
{
3111
	struct global_cwq *gcwq = get_work_gcwq(work);
3112

3113
	return gcwq ? gcwq->cpu : WORK_CPU_NONE;
3114
}
3115
EXPORT_SYMBOL_GPL(work_cpu);
3116

3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130
/**
 * work_busy - test whether a work is currently pending or running
 * @work: the work to be tested
 *
 * Test whether @work is currently pending or running.  There is no
 * synchronization around this function and the test result is
 * unreliable and only useful as advisory hints or for debugging.
 * Especially for reentrant wqs, the pending state might hide the
 * running state.
 *
 * RETURNS:
 * OR'd bitmask of WORK_BUSY_* bits.
 */
unsigned int work_busy(struct work_struct *work)
L
Linus Torvalds 已提交
3131
{
3132 3133 3134
	struct global_cwq *gcwq = get_work_gcwq(work);
	unsigned long flags;
	unsigned int ret = 0;
L
Linus Torvalds 已提交
3135

3136 3137
	if (!gcwq)
		return false;
L
Linus Torvalds 已提交
3138

3139
	spin_lock_irqsave(&gcwq->lock, flags);
L
Linus Torvalds 已提交
3140

3141 3142 3143 3144
	if (work_pending(work))
		ret |= WORK_BUSY_PENDING;
	if (find_worker_executing_work(gcwq, work))
		ret |= WORK_BUSY_RUNNING;
L
Linus Torvalds 已提交
3145

3146
	spin_unlock_irqrestore(&gcwq->lock, flags);
L
Linus Torvalds 已提交
3147

3148
	return ret;
L
Linus Torvalds 已提交
3149
}
3150
EXPORT_SYMBOL_GPL(work_busy);
L
Linus Torvalds 已提交
3151

3152 3153 3154
/*
 * CPU hotplug.
 *
3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166
 * There are two challenges in supporting CPU hotplug.  Firstly, there
 * are a lot of assumptions on strong associations among work, cwq and
 * gcwq which make migrating pending and scheduled works very
 * difficult to implement without impacting hot paths.  Secondly,
 * gcwqs serve mix of short, long and very long running works making
 * blocked draining impractical.
 *
 * This is solved by allowing a gcwq to be detached from CPU, running
 * it with unbound (rogue) workers and allowing it to be reattached
 * later if the cpu comes back online.  A separate thread is created
 * to govern a gcwq in such state and is called the trustee of the
 * gcwq.
3167 3168 3169 3170 3171 3172 3173
 *
 * Trustee states and their descriptions.
 *
 * START	Command state used on startup.  On CPU_DOWN_PREPARE, a
 *		new trustee is started with this state.
 *
 * IN_CHARGE	Once started, trustee will enter this state after
3174 3175 3176 3177 3178 3179
 *		assuming the manager role and making all existing
 *		workers rogue.  DOWN_PREPARE waits for trustee to
 *		enter this state.  After reaching IN_CHARGE, trustee
 *		tries to execute the pending worklist until it's empty
 *		and the state is set to BUTCHER, or the state is set
 *		to RELEASE.
3180 3181 3182 3183 3184 3185 3186 3187 3188 3189
 *
 * BUTCHER	Command state which is set by the cpu callback after
 *		the cpu has went down.  Once this state is set trustee
 *		knows that there will be no new works on the worklist
 *		and once the worklist is empty it can proceed to
 *		killing idle workers.
 *
 * RELEASE	Command state which is set by the cpu callback if the
 *		cpu down has been canceled or it has come online
 *		again.  After recognizing this state, trustee stops
3190 3191 3192
 *		trying to drain or butcher and clears ROGUE, rebinds
 *		all remaining workers back to the cpu and releases
 *		manager role.
3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203
 *
 * DONE		Trustee will enter this state after BUTCHER or RELEASE
 *		is complete.
 *
 *          trustee                 CPU                draining
 *         took over                down               complete
 * START -----------> IN_CHARGE -----------> BUTCHER -----------> DONE
 *                        |                     |                  ^
 *                        | CPU is back online  v   return workers |
 *                         ----------------> RELEASE --------------
 */
L
Linus Torvalds 已提交
3204

3205 3206 3207 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
/**
 * trustee_wait_event_timeout - timed event wait for trustee
 * @cond: condition to wait for
 * @timeout: timeout in jiffies
 *
 * wait_event_timeout() for trustee to use.  Handles locking and
 * checks for RELEASE request.
 *
 * CONTEXT:
 * spin_lock_irq(gcwq->lock) which may be released and regrabbed
 * multiple times.  To be used by trustee.
 *
 * RETURNS:
 * Positive indicating left time if @cond is satisfied, 0 if timed
 * out, -1 if canceled.
 */
#define trustee_wait_event_timeout(cond, timeout) ({			\
	long __ret = (timeout);						\
	while (!((cond) || (gcwq->trustee_state == TRUSTEE_RELEASE)) &&	\
	       __ret) {							\
		spin_unlock_irq(&gcwq->lock);				\
		__wait_event_timeout(gcwq->trustee_wait, (cond) ||	\
			(gcwq->trustee_state == TRUSTEE_RELEASE),	\
			__ret);						\
		spin_lock_irq(&gcwq->lock);				\
	}								\
	gcwq->trustee_state == TRUSTEE_RELEASE ? -1 : (__ret);		\
})
3233

3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252
/**
 * trustee_wait_event - event wait for trustee
 * @cond: condition to wait for
 *
 * wait_event() for trustee to use.  Automatically handles locking and
 * checks for CANCEL request.
 *
 * CONTEXT:
 * spin_lock_irq(gcwq->lock) which may be released and regrabbed
 * multiple times.  To be used by trustee.
 *
 * RETURNS:
 * 0 if @cond is satisfied, -1 if canceled.
 */
#define trustee_wait_event(cond) ({					\
	long __ret1;							\
	__ret1 = trustee_wait_event_timeout(cond, MAX_SCHEDULE_TIMEOUT);\
	__ret1 < 0 ? -1 : 0;						\
})
L
Linus Torvalds 已提交
3253

3254
static int __cpuinit trustee_thread(void *__gcwq)
3255
{
3256 3257
	struct global_cwq *gcwq = __gcwq;
	struct worker *worker;
3258
	struct work_struct *work;
3259
	struct hlist_node *pos;
3260
	long rc;
3261
	int i;
3262

3263 3264 3265
	BUG_ON(gcwq->cpu != smp_processor_id());

	spin_lock_irq(&gcwq->lock);
3266
	/*
3267 3268 3269
	 * Claim the manager position and make all workers rogue.
	 * Trustee must be bound to the target cpu and can't be
	 * cancelled.
3270
	 */
3271
	BUG_ON(gcwq->cpu != smp_processor_id());
3272 3273
	rc = trustee_wait_event(!(gcwq->flags & GCWQ_MANAGING_WORKERS));
	BUG_ON(rc < 0);
3274

3275
	gcwq->flags |= GCWQ_MANAGING_WORKERS;
3276

3277
	list_for_each_entry(worker, &gcwq->idle_list, entry)
3278
		worker->flags |= WORKER_ROGUE;
3279

3280
	for_each_busy_worker(worker, i, pos, gcwq)
3281
		worker->flags |= WORKER_ROGUE;
3282

3283 3284 3285 3286 3287 3288 3289 3290 3291
	/*
	 * Call schedule() so that we cross rq->lock and thus can
	 * guarantee sched callbacks see the rogue flag.  This is
	 * necessary as scheduler callbacks may be invoked from other
	 * cpus.
	 */
	spin_unlock_irq(&gcwq->lock);
	schedule();
	spin_lock_irq(&gcwq->lock);
3292

3293
	/*
3294 3295 3296 3297
	 * Sched callbacks are disabled now.  Zap nr_running.  After
	 * this, nr_running stays zero and need_more_worker() and
	 * keep_working() are always true as long as the worklist is
	 * not empty.
3298
	 */
3299
	atomic_set(get_gcwq_nr_running(gcwq->cpu), 0);
L
Linus Torvalds 已提交
3300

3301 3302 3303
	spin_unlock_irq(&gcwq->lock);
	del_timer_sync(&gcwq->idle_timer);
	spin_lock_irq(&gcwq->lock);
3304

3305 3306 3307 3308 3309 3310 3311 3312
	/*
	 * We're now in charge.  Notify and proceed to drain.  We need
	 * to keep the gcwq running during the whole CPU down
	 * procedure as other cpu hotunplug callbacks may need to
	 * flush currently running tasks.
	 */
	gcwq->trustee_state = TRUSTEE_IN_CHARGE;
	wake_up_all(&gcwq->trustee_wait);
3313

3314 3315 3316
	/*
	 * The original cpu is in the process of dying and may go away
	 * anytime now.  When that happens, we and all workers would
3317 3318 3319 3320 3321 3322
	 * be migrated to other cpus.  Try draining any left work.  We
	 * want to get it over with ASAP - spam rescuers, wake up as
	 * many idlers as necessary and create new ones till the
	 * worklist is empty.  Note that if the gcwq is frozen, there
	 * may be frozen works in freezeable cwqs.  Don't declare
	 * completion while frozen.
3323 3324 3325 3326
	 */
	while (gcwq->nr_workers != gcwq->nr_idle ||
	       gcwq->flags & GCWQ_FREEZING ||
	       gcwq->trustee_state == TRUSTEE_IN_CHARGE) {
3327 3328 3329 3330 3331 3332
		int nr_works = 0;

		list_for_each_entry(work, &gcwq->worklist, entry) {
			send_mayday(work);
			nr_works++;
		}
3333

3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344
		list_for_each_entry(worker, &gcwq->idle_list, entry) {
			if (!nr_works--)
				break;
			wake_up_process(worker->task);
		}

		if (need_to_create_worker(gcwq)) {
			spin_unlock_irq(&gcwq->lock);
			worker = create_worker(gcwq, false);
			spin_lock_irq(&gcwq->lock);
			if (worker) {
3345
				worker->flags |= WORKER_ROGUE;
3346 3347
				start_worker(worker);
			}
L
Linus Torvalds 已提交
3348
		}
3349

3350 3351 3352
		/* give a breather */
		if (trustee_wait_event_timeout(false, TRUSTEE_COOLDOWN) < 0)
			break;
3353
	}
L
Linus Torvalds 已提交
3354

3355
	/*
3356 3357 3358
	 * Either all works have been scheduled and cpu is down, or
	 * cpu down has already been canceled.  Wait for and butcher
	 * all workers till we're canceled.
3359
	 */
3360 3361 3362 3363 3364 3365
	do {
		rc = trustee_wait_event(!list_empty(&gcwq->idle_list));
		while (!list_empty(&gcwq->idle_list))
			destroy_worker(list_first_entry(&gcwq->idle_list,
							struct worker, entry));
	} while (gcwq->nr_workers && rc >= 0);
3366

3367
	/*
3368 3369 3370 3371 3372
	 * At this point, either draining has completed and no worker
	 * is left, or cpu down has been canceled or the cpu is being
	 * brought back up.  There shouldn't be any idle one left.
	 * Tell the remaining busy ones to rebind once it finishes the
	 * currently scheduled works by scheduling the rebind_work.
3373
	 */
3374 3375 3376 3377 3378 3379 3380 3381 3382 3383
	WARN_ON(!list_empty(&gcwq->idle_list));

	for_each_busy_worker(worker, i, pos, gcwq) {
		struct work_struct *rebind_work = &worker->rebind_work;

		/*
		 * Rebind_work may race with future cpu hotplug
		 * operations.  Use a separate flag to mark that
		 * rebinding is scheduled.
		 */
3384 3385
		worker->flags |= WORKER_REBIND;
		worker->flags &= ~WORKER_ROGUE;
3386 3387 3388 3389 3390 3391 3392

		/* queue rebind_work, wq doesn't matter, use the default one */
		if (test_and_set_bit(WORK_STRUCT_PENDING_BIT,
				     work_data_bits(rebind_work)))
			continue;

		debug_work_activate(rebind_work);
3393
		insert_work(get_cwq(gcwq->cpu, system_wq), rebind_work,
3394 3395 3396 3397 3398 3399 3400
			    worker->scheduled.next,
			    work_color_to_flags(WORK_NO_COLOR));
	}

	/* relinquish manager role */
	gcwq->flags &= ~GCWQ_MANAGING_WORKERS;

3401 3402 3403 3404 3405 3406
	/* notify completion */
	gcwq->trustee = NULL;
	gcwq->trustee_state = TRUSTEE_DONE;
	wake_up_all(&gcwq->trustee_wait);
	spin_unlock_irq(&gcwq->lock);
	return 0;
3407 3408 3409
}

/**
3410 3411 3412
 * wait_trustee_state - wait for trustee to enter the specified state
 * @gcwq: gcwq the trustee of interest belongs to
 * @state: target state to wait for
3413
 *
3414 3415 3416 3417 3418
 * Wait for the trustee to reach @state.  DONE is already matched.
 *
 * CONTEXT:
 * spin_lock_irq(gcwq->lock) which may be released and regrabbed
 * multiple times.  To be used by cpu_callback.
3419
 */
3420
static void __cpuinit wait_trustee_state(struct global_cwq *gcwq, int state)
3421 3422
__releases(&gcwq->lock)
__acquires(&gcwq->lock)
3423
{
3424 3425 3426 3427 3428 3429 3430 3431
	if (!(gcwq->trustee_state == state ||
	      gcwq->trustee_state == TRUSTEE_DONE)) {
		spin_unlock_irq(&gcwq->lock);
		__wait_event(gcwq->trustee_wait,
			     gcwq->trustee_state == state ||
			     gcwq->trustee_state == TRUSTEE_DONE);
		spin_lock_irq(&gcwq->lock);
	}
3432 3433 3434 3435 3436 3437 3438
}

static int __devinit workqueue_cpu_callback(struct notifier_block *nfb,
						unsigned long action,
						void *hcpu)
{
	unsigned int cpu = (unsigned long)hcpu;
3439 3440
	struct global_cwq *gcwq = get_gcwq(cpu);
	struct task_struct *new_trustee = NULL;
3441
	struct worker *uninitialized_var(new_worker);
3442
	unsigned long flags;
3443

3444 3445
	action &= ~CPU_TASKS_FROZEN;

3446
	switch (action) {
3447 3448 3449 3450 3451 3452
	case CPU_DOWN_PREPARE:
		new_trustee = kthread_create(trustee_thread, gcwq,
					     "workqueue_trustee/%d\n", cpu);
		if (IS_ERR(new_trustee))
			return notifier_from_errno(PTR_ERR(new_trustee));
		kthread_bind(new_trustee, cpu);
3453
		/* fall through */
3454
	case CPU_UP_PREPARE:
3455 3456 3457 3458 3459 3460
		BUG_ON(gcwq->first_idle);
		new_worker = create_worker(gcwq, false);
		if (!new_worker) {
			if (new_trustee)
				kthread_stop(new_trustee);
			return NOTIFY_BAD;
3461
		}
L
Linus Torvalds 已提交
3462 3463
	}

3464 3465
	/* some are called w/ irq disabled, don't disturb irq status */
	spin_lock_irqsave(&gcwq->lock, flags);
3466

3467
	switch (action) {
3468 3469 3470 3471 3472 3473 3474
	case CPU_DOWN_PREPARE:
		/* initialize trustee and tell it to acquire the gcwq */
		BUG_ON(gcwq->trustee || gcwq->trustee_state != TRUSTEE_DONE);
		gcwq->trustee = new_trustee;
		gcwq->trustee_state = TRUSTEE_START;
		wake_up_process(gcwq->trustee);
		wait_trustee_state(gcwq, TRUSTEE_IN_CHARGE);
3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488
		/* fall through */
	case CPU_UP_PREPARE:
		BUG_ON(gcwq->first_idle);
		gcwq->first_idle = new_worker;
		break;

	case CPU_DYING:
		/*
		 * Before this, the trustee and all workers except for
		 * the ones which are still executing works from
		 * before the last CPU down must be on the cpu.  After
		 * this, they'll all be diasporas.
		 */
		gcwq->flags |= GCWQ_DISASSOCIATED;
3489 3490
		break;

3491
	case CPU_POST_DEAD:
3492
		gcwq->trustee_state = TRUSTEE_BUTCHER;
3493 3494 3495 3496
		/* fall through */
	case CPU_UP_CANCELED:
		destroy_worker(gcwq->first_idle);
		gcwq->first_idle = NULL;
3497 3498 3499 3500
		break;

	case CPU_DOWN_FAILED:
	case CPU_ONLINE:
3501
		gcwq->flags &= ~GCWQ_DISASSOCIATED;
3502 3503 3504 3505
		if (gcwq->trustee_state != TRUSTEE_DONE) {
			gcwq->trustee_state = TRUSTEE_RELEASE;
			wake_up_process(gcwq->trustee);
			wait_trustee_state(gcwq, TRUSTEE_DONE);
3506
		}
3507

3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518
		/*
		 * Trustee is done and there might be no worker left.
		 * Put the first_idle in and request a real manager to
		 * take a look.
		 */
		spin_unlock_irq(&gcwq->lock);
		kthread_bind(gcwq->first_idle->task, cpu);
		spin_lock_irq(&gcwq->lock);
		gcwq->flags |= GCWQ_MANAGE_WORKERS;
		start_worker(gcwq->first_idle);
		gcwq->first_idle = NULL;
3519
		break;
3520 3521
	}

3522 3523
	spin_unlock_irqrestore(&gcwq->lock, flags);

T
Tejun Heo 已提交
3524
	return notifier_from_errno(0);
L
Linus Torvalds 已提交
3525 3526
}

3527
#ifdef CONFIG_SMP
3528

3529
struct work_for_cpu {
3530
	struct completion completion;
3531 3532 3533 3534 3535
	long (*fn)(void *);
	void *arg;
	long ret;
};

3536
static int do_work_for_cpu(void *_wfc)
3537
{
3538
	struct work_for_cpu *wfc = _wfc;
3539
	wfc->ret = wfc->fn(wfc->arg);
3540 3541
	complete(&wfc->completion);
	return 0;
3542 3543 3544 3545 3546 3547 3548 3549
}

/**
 * work_on_cpu - run a function in user context on a particular cpu
 * @cpu: the cpu to run on
 * @fn: the function to run
 * @arg: the function arg
 *
3550 3551
 * This will return the value @fn returns.
 * It is up to the caller to ensure that the cpu doesn't go offline.
3552
 * The caller must not hold any locks which would prevent @fn from completing.
3553 3554 3555
 */
long work_on_cpu(unsigned int cpu, long (*fn)(void *), void *arg)
{
3556 3557 3558 3559 3560 3561 3562 3563 3564 3565 3566 3567 3568
	struct task_struct *sub_thread;
	struct work_for_cpu wfc = {
		.completion = COMPLETION_INITIALIZER_ONSTACK(wfc.completion),
		.fn = fn,
		.arg = arg,
	};

	sub_thread = kthread_create(do_work_for_cpu, &wfc, "work_for_cpu");
	if (IS_ERR(sub_thread))
		return PTR_ERR(sub_thread);
	kthread_bind(sub_thread, cpu);
	wake_up_process(sub_thread);
	wait_for_completion(&wfc.completion);
3569 3570 3571 3572 3573
	return wfc.ret;
}
EXPORT_SYMBOL_GPL(work_on_cpu);
#endif /* CONFIG_SMP */

3574 3575 3576 3577 3578 3579 3580
#ifdef CONFIG_FREEZER

/**
 * freeze_workqueues_begin - begin freezing workqueues
 *
 * Start freezing workqueues.  After this function returns, all
 * freezeable workqueues will queue new works to their frozen_works
3581
 * list instead of gcwq->worklist.
3582 3583
 *
 * CONTEXT:
3584
 * Grabs and releases workqueue_lock and gcwq->lock's.
3585 3586 3587 3588 3589 3590 3591 3592 3593 3594
 */
void freeze_workqueues_begin(void)
{
	unsigned int cpu;

	spin_lock(&workqueue_lock);

	BUG_ON(workqueue_freezing);
	workqueue_freezing = true;

3595
	for_each_gcwq_cpu(cpu) {
3596
		struct global_cwq *gcwq = get_gcwq(cpu);
3597
		struct workqueue_struct *wq;
3598 3599 3600

		spin_lock_irq(&gcwq->lock);

3601 3602 3603
		BUG_ON(gcwq->flags & GCWQ_FREEZING);
		gcwq->flags |= GCWQ_FREEZING;

3604 3605 3606
		list_for_each_entry(wq, &workqueues, list) {
			struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq);

3607
			if (cwq && wq->flags & WQ_FREEZEABLE)
3608 3609
				cwq->max_active = 0;
		}
3610 3611

		spin_unlock_irq(&gcwq->lock);
3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638
	}

	spin_unlock(&workqueue_lock);
}

/**
 * freeze_workqueues_busy - are freezeable workqueues still busy?
 *
 * Check whether freezing is complete.  This function must be called
 * between freeze_workqueues_begin() and thaw_workqueues().
 *
 * CONTEXT:
 * Grabs and releases workqueue_lock.
 *
 * RETURNS:
 * %true if some freezeable workqueues are still busy.  %false if
 * freezing is complete.
 */
bool freeze_workqueues_busy(void)
{
	unsigned int cpu;
	bool busy = false;

	spin_lock(&workqueue_lock);

	BUG_ON(!workqueue_freezing);

3639
	for_each_gcwq_cpu(cpu) {
3640
		struct workqueue_struct *wq;
3641 3642 3643 3644 3645 3646 3647
		/*
		 * nr_active is monotonically decreasing.  It's safe
		 * to peek without lock.
		 */
		list_for_each_entry(wq, &workqueues, list) {
			struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq);

3648
			if (!cwq || !(wq->flags & WQ_FREEZEABLE))
3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662 3663 3664 3665 3666
				continue;

			BUG_ON(cwq->nr_active < 0);
			if (cwq->nr_active) {
				busy = true;
				goto out_unlock;
			}
		}
	}
out_unlock:
	spin_unlock(&workqueue_lock);
	return busy;
}

/**
 * thaw_workqueues - thaw workqueues
 *
 * Thaw workqueues.  Normal queueing is restored and all collected
3667
 * frozen works are transferred to their respective gcwq worklists.
3668 3669
 *
 * CONTEXT:
3670
 * Grabs and releases workqueue_lock and gcwq->lock's.
3671 3672 3673 3674 3675 3676 3677 3678 3679 3680
 */
void thaw_workqueues(void)
{
	unsigned int cpu;

	spin_lock(&workqueue_lock);

	if (!workqueue_freezing)
		goto out_unlock;

3681
	for_each_gcwq_cpu(cpu) {
3682
		struct global_cwq *gcwq = get_gcwq(cpu);
3683
		struct workqueue_struct *wq;
3684 3685 3686

		spin_lock_irq(&gcwq->lock);

3687 3688 3689
		BUG_ON(!(gcwq->flags & GCWQ_FREEZING));
		gcwq->flags &= ~GCWQ_FREEZING;

3690 3691 3692
		list_for_each_entry(wq, &workqueues, list) {
			struct cpu_workqueue_struct *cwq = get_cwq(cpu, wq);

3693
			if (!cwq || !(wq->flags & WQ_FREEZEABLE))
3694 3695 3696 3697 3698 3699 3700 3701 3702
				continue;

			/* restore max_active and repopulate worklist */
			cwq->max_active = wq->saved_max_active;

			while (!list_empty(&cwq->delayed_works) &&
			       cwq->nr_active < cwq->max_active)
				cwq_activate_first_delayed(cwq);
		}
3703

3704 3705
		wake_up_worker(gcwq);

3706
		spin_unlock_irq(&gcwq->lock);
3707 3708 3709 3710 3711 3712 3713 3714
	}

	workqueue_freezing = false;
out_unlock:
	spin_unlock(&workqueue_lock);
}
#endif /* CONFIG_FREEZER */

3715
static int __init init_workqueues(void)
L
Linus Torvalds 已提交
3716
{
T
Tejun Heo 已提交
3717
	unsigned int cpu;
T
Tejun Heo 已提交
3718
	int i;
T
Tejun Heo 已提交
3719

3720
	cpu_notifier(workqueue_cpu_callback, CPU_PRI_WORKQUEUE);
3721 3722

	/* initialize gcwqs */
3723
	for_each_gcwq_cpu(cpu) {
3724 3725 3726
		struct global_cwq *gcwq = get_gcwq(cpu);

		spin_lock_init(&gcwq->lock);
3727
		INIT_LIST_HEAD(&gcwq->worklist);
3728
		gcwq->cpu = cpu;
3729
		gcwq->flags |= GCWQ_DISASSOCIATED;
3730

T
Tejun Heo 已提交
3731 3732 3733 3734
		INIT_LIST_HEAD(&gcwq->idle_list);
		for (i = 0; i < BUSY_WORKER_HASH_SIZE; i++)
			INIT_HLIST_HEAD(&gcwq->busy_hash[i]);

3735 3736 3737
		init_timer_deferrable(&gcwq->idle_timer);
		gcwq->idle_timer.function = idle_worker_timeout;
		gcwq->idle_timer.data = (unsigned long)gcwq;
3738

3739 3740 3741
		setup_timer(&gcwq->mayday_timer, gcwq_mayday_timeout,
			    (unsigned long)gcwq);

3742
		ida_init(&gcwq->worker_ida);
3743 3744 3745

		gcwq->trustee_state = TRUSTEE_DONE;
		init_waitqueue_head(&gcwq->trustee_wait);
3746 3747
	}

3748
	/* create the initial worker */
3749
	for_each_online_gcwq_cpu(cpu) {
3750 3751 3752
		struct global_cwq *gcwq = get_gcwq(cpu);
		struct worker *worker;

3753 3754
		if (cpu != WORK_CPU_UNBOUND)
			gcwq->flags &= ~GCWQ_DISASSOCIATED;
3755 3756 3757 3758 3759 3760 3761
		worker = create_worker(gcwq, true);
		BUG_ON(!worker);
		spin_lock_irq(&gcwq->lock);
		start_worker(worker);
		spin_unlock_irq(&gcwq->lock);
	}

3762 3763 3764
	system_wq = alloc_workqueue("events", 0, 0);
	system_long_wq = alloc_workqueue("events_long", 0, 0);
	system_nrt_wq = alloc_workqueue("events_nrt", WQ_NON_REENTRANT, 0);
3765 3766
	system_unbound_wq = alloc_workqueue("events_unbound", WQ_UNBOUND,
					    WQ_UNBOUND_MAX_ACTIVE);
3767 3768
	BUG_ON(!system_wq || !system_long_wq || !system_nrt_wq ||
	       !system_unbound_wq);
3769
	return 0;
L
Linus Torvalds 已提交
3770
}
3771
early_initcall(init_workqueues);