reimplement flush_workqueue()
[linux-2.6-block.git] / kernel / workqueue.c
CommitLineData
1da177e4
LT
1/*
2 * linux/kernel/workqueue.c
3 *
4 * Generic mechanism for defining kernel helper threads for running
5 * arbitrary tasks in process context.
6 *
7 * Started by Ingo Molnar, Copyright (C) 2002
8 *
9 * Derived from the taskqueue/keventd code by:
10 *
11 * David Woodhouse <dwmw2@infradead.org>
12 * Andrew Morton <andrewm@uow.edu.au>
13 * Kai Petzke <wpp@marie.physik.tu-berlin.de>
14 * Theodore Ts'o <tytso@mit.edu>
89ada679
CL
15 *
16 * Made to use alloc_percpu by Christoph Lameter <clameter@sgi.com>.
1da177e4
LT
17 */
18
19#include <linux/module.h>
20#include <linux/kernel.h>
21#include <linux/sched.h>
22#include <linux/init.h>
23#include <linux/signal.h>
24#include <linux/completion.h>
25#include <linux/workqueue.h>
26#include <linux/slab.h>
27#include <linux/cpu.h>
28#include <linux/notifier.h>
29#include <linux/kthread.h>
1fa44eca 30#include <linux/hardirq.h>
46934023 31#include <linux/mempolicy.h>
341a5958 32#include <linux/freezer.h>
d5abe669
PZ
33#include <linux/kallsyms.h>
34#include <linux/debug_locks.h>
1da177e4
LT
35
36/*
f756d5e2
NL
37 * The per-CPU workqueue (if single thread, we always use the first
38 * possible cpu).
1da177e4
LT
39 */
40struct cpu_workqueue_struct {
41
42 spinlock_t lock;
43
1da177e4
LT
44 struct list_head worklist;
45 wait_queue_head_t more_work;
1da177e4
LT
46
47 struct workqueue_struct *wq;
36c8b586 48 struct task_struct *thread;
1da177e4
LT
49
50 int run_depth; /* Detect run_workqueue() recursion depth */
341a5958
RW
51
52 int freezeable; /* Freeze the thread during suspend */
1da177e4
LT
53} ____cacheline_aligned;
54
55/*
56 * The externally visible workqueue abstraction is an array of
57 * per-CPU workqueues:
58 */
59struct workqueue_struct {
89ada679 60 struct cpu_workqueue_struct *cpu_wq;
1da177e4
LT
61 const char *name;
62 struct list_head list; /* Empty if single thread */
63};
64
65/* All the per-cpu workqueues on the system, for hotplug cpu to add/remove
66 threads to each one as cpus come/go. */
9b41ea72 67static DEFINE_MUTEX(workqueue_mutex);
1da177e4
LT
68static LIST_HEAD(workqueues);
69
f756d5e2
NL
70static int singlethread_cpu;
71
1da177e4
LT
72/* If it's single threaded, it isn't in the list of workqueues. */
73static inline int is_single_threaded(struct workqueue_struct *wq)
74{
75 return list_empty(&wq->list);
76}
77
4594bf15
DH
78/*
79 * Set the workqueue on which a work item is to be run
80 * - Must *only* be called if the pending flag is set
81 */
365970a1
DH
82static inline void set_wq_data(struct work_struct *work, void *wq)
83{
4594bf15
DH
84 unsigned long new;
85
86 BUG_ON(!work_pending(work));
365970a1 87
365970a1 88 new = (unsigned long) wq | (1UL << WORK_STRUCT_PENDING);
a08727ba
LT
89 new |= WORK_STRUCT_FLAG_MASK & *work_data_bits(work);
90 atomic_long_set(&work->data, new);
365970a1
DH
91}
92
93static inline void *get_wq_data(struct work_struct *work)
94{
a08727ba 95 return (void *) (atomic_long_read(&work->data) & WORK_STRUCT_WQ_DATA_MASK);
365970a1
DH
96}
97
68380b58
LT
98static int __run_work(struct cpu_workqueue_struct *cwq, struct work_struct *work)
99{
100 int ret = 0;
101 unsigned long flags;
102
103 spin_lock_irqsave(&cwq->lock, flags);
104 /*
105 * We need to re-validate the work info after we've gotten
106 * the cpu_workqueue lock. We can run the work now iff:
107 *
108 * - the wq_data still matches the cpu_workqueue_struct
109 * - AND the work is still marked pending
110 * - AND the work is still on a list (which will be this
111 * workqueue_struct list)
112 *
113 * All these conditions are important, because we
114 * need to protect against the work being run right
115 * now on another CPU (all but the last one might be
116 * true if it's currently running and has not been
117 * released yet, for example).
118 */
119 if (get_wq_data(work) == cwq
120 && work_pending(work)
121 && !list_empty(&work->entry)) {
122 work_func_t f = work->func;
123 list_del_init(&work->entry);
124 spin_unlock_irqrestore(&cwq->lock, flags);
125
a08727ba 126 if (!test_bit(WORK_STRUCT_NOAUTOREL, work_data_bits(work)))
68380b58
LT
127 work_release(work);
128 f(work);
129
130 spin_lock_irqsave(&cwq->lock, flags);
68380b58
LT
131 ret = 1;
132 }
133 spin_unlock_irqrestore(&cwq->lock, flags);
134 return ret;
135}
136
137/**
138 * run_scheduled_work - run scheduled work synchronously
139 * @work: work to run
140 *
141 * This checks if the work was pending, and runs it
142 * synchronously if so. It returns a boolean to indicate
143 * whether it had any scheduled work to run or not.
144 *
145 * NOTE! This _only_ works for normal work_structs. You
146 * CANNOT use this for delayed work, because the wq data
147 * for delayed work will not point properly to the per-
148 * CPU workqueue struct, but will change!
149 */
150int fastcall run_scheduled_work(struct work_struct *work)
151{
152 for (;;) {
153 struct cpu_workqueue_struct *cwq;
154
155 if (!work_pending(work))
156 return 0;
157 if (list_empty(&work->entry))
158 return 0;
159 /* NOTE! This depends intimately on __queue_work! */
160 cwq = get_wq_data(work);
161 if (!cwq)
162 return 0;
163 if (__run_work(cwq, work))
164 return 1;
165 }
166}
167EXPORT_SYMBOL(run_scheduled_work);
168
1da177e4
LT
169/* Preempt must be disabled. */
170static void __queue_work(struct cpu_workqueue_struct *cwq,
171 struct work_struct *work)
172{
173 unsigned long flags;
174
175 spin_lock_irqsave(&cwq->lock, flags);
365970a1 176 set_wq_data(work, cwq);
1da177e4 177 list_add_tail(&work->entry, &cwq->worklist);
1da177e4
LT
178 wake_up(&cwq->more_work);
179 spin_unlock_irqrestore(&cwq->lock, flags);
180}
181
0fcb78c2
REB
182/**
183 * queue_work - queue work on a workqueue
184 * @wq: workqueue to use
185 * @work: work to queue
186 *
057647fc 187 * Returns 0 if @work was already on a queue, non-zero otherwise.
1da177e4
LT
188 *
189 * We queue the work to the CPU it was submitted, but there is no
190 * guarantee that it will be processed by that CPU.
191 */
192int fastcall queue_work(struct workqueue_struct *wq, struct work_struct *work)
193{
194 int ret = 0, cpu = get_cpu();
195
a08727ba 196 if (!test_and_set_bit(WORK_STRUCT_PENDING, work_data_bits(work))) {
1da177e4 197 if (unlikely(is_single_threaded(wq)))
f756d5e2 198 cpu = singlethread_cpu;
1da177e4 199 BUG_ON(!list_empty(&work->entry));
89ada679 200 __queue_work(per_cpu_ptr(wq->cpu_wq, cpu), work);
1da177e4
LT
201 ret = 1;
202 }
203 put_cpu();
204 return ret;
205}
ae90dd5d 206EXPORT_SYMBOL_GPL(queue_work);
1da177e4 207
82f67cd9 208void delayed_work_timer_fn(unsigned long __data)
1da177e4 209{
52bad64d 210 struct delayed_work *dwork = (struct delayed_work *)__data;
365970a1 211 struct workqueue_struct *wq = get_wq_data(&dwork->work);
1da177e4
LT
212 int cpu = smp_processor_id();
213
214 if (unlikely(is_single_threaded(wq)))
f756d5e2 215 cpu = singlethread_cpu;
1da177e4 216
52bad64d 217 __queue_work(per_cpu_ptr(wq->cpu_wq, cpu), &dwork->work);
1da177e4
LT
218}
219
0fcb78c2
REB
220/**
221 * queue_delayed_work - queue work on a workqueue after delay
222 * @wq: workqueue to use
af9997e4 223 * @dwork: delayable work to queue
0fcb78c2
REB
224 * @delay: number of jiffies to wait before queueing
225 *
057647fc 226 * Returns 0 if @work was already on a queue, non-zero otherwise.
0fcb78c2 227 */
1da177e4 228int fastcall queue_delayed_work(struct workqueue_struct *wq,
52bad64d 229 struct delayed_work *dwork, unsigned long delay)
1da177e4
LT
230{
231 int ret = 0;
52bad64d
DH
232 struct timer_list *timer = &dwork->timer;
233 struct work_struct *work = &dwork->work;
234
82f67cd9 235 timer_stats_timer_set_start_info(timer);
52bad64d
DH
236 if (delay == 0)
237 return queue_work(wq, work);
1da177e4 238
a08727ba 239 if (!test_and_set_bit(WORK_STRUCT_PENDING, work_data_bits(work))) {
1da177e4
LT
240 BUG_ON(timer_pending(timer));
241 BUG_ON(!list_empty(&work->entry));
242
243 /* This stores wq for the moment, for the timer_fn */
365970a1 244 set_wq_data(work, wq);
1da177e4 245 timer->expires = jiffies + delay;
52bad64d 246 timer->data = (unsigned long)dwork;
1da177e4
LT
247 timer->function = delayed_work_timer_fn;
248 add_timer(timer);
249 ret = 1;
250 }
251 return ret;
252}
ae90dd5d 253EXPORT_SYMBOL_GPL(queue_delayed_work);
1da177e4 254
0fcb78c2
REB
255/**
256 * queue_delayed_work_on - queue work on specific CPU after delay
257 * @cpu: CPU number to execute work on
258 * @wq: workqueue to use
af9997e4 259 * @dwork: work to queue
0fcb78c2
REB
260 * @delay: number of jiffies to wait before queueing
261 *
057647fc 262 * Returns 0 if @work was already on a queue, non-zero otherwise.
0fcb78c2 263 */
7a6bc1cd 264int queue_delayed_work_on(int cpu, struct workqueue_struct *wq,
52bad64d 265 struct delayed_work *dwork, unsigned long delay)
7a6bc1cd
VP
266{
267 int ret = 0;
52bad64d
DH
268 struct timer_list *timer = &dwork->timer;
269 struct work_struct *work = &dwork->work;
7a6bc1cd 270
a08727ba 271 if (!test_and_set_bit(WORK_STRUCT_PENDING, work_data_bits(work))) {
7a6bc1cd
VP
272 BUG_ON(timer_pending(timer));
273 BUG_ON(!list_empty(&work->entry));
274
275 /* This stores wq for the moment, for the timer_fn */
365970a1 276 set_wq_data(work, wq);
7a6bc1cd 277 timer->expires = jiffies + delay;
52bad64d 278 timer->data = (unsigned long)dwork;
7a6bc1cd
VP
279 timer->function = delayed_work_timer_fn;
280 add_timer_on(timer, cpu);
281 ret = 1;
282 }
283 return ret;
284}
ae90dd5d 285EXPORT_SYMBOL_GPL(queue_delayed_work_on);
1da177e4 286
858119e1 287static void run_workqueue(struct cpu_workqueue_struct *cwq)
1da177e4
LT
288{
289 unsigned long flags;
290
291 /*
292 * Keep taking off work from the queue until
293 * done.
294 */
295 spin_lock_irqsave(&cwq->lock, flags);
296 cwq->run_depth++;
297 if (cwq->run_depth > 3) {
298 /* morton gets to eat his hat */
299 printk("%s: recursion depth exceeded: %d\n",
300 __FUNCTION__, cwq->run_depth);
301 dump_stack();
302 }
303 while (!list_empty(&cwq->worklist)) {
304 struct work_struct *work = list_entry(cwq->worklist.next,
305 struct work_struct, entry);
6bb49e59 306 work_func_t f = work->func;
1da177e4
LT
307
308 list_del_init(cwq->worklist.next);
309 spin_unlock_irqrestore(&cwq->lock, flags);
310
365970a1 311 BUG_ON(get_wq_data(work) != cwq);
a08727ba 312 if (!test_bit(WORK_STRUCT_NOAUTOREL, work_data_bits(work)))
65f27f38
DH
313 work_release(work);
314 f(work);
1da177e4 315
d5abe669
PZ
316 if (unlikely(in_atomic() || lockdep_depth(current) > 0)) {
317 printk(KERN_ERR "BUG: workqueue leaked lock or atomic: "
318 "%s/0x%08x/%d\n",
319 current->comm, preempt_count(),
320 current->pid);
321 printk(KERN_ERR " last function: ");
322 print_symbol("%s\n", (unsigned long)f);
323 debug_show_held_locks(current);
324 dump_stack();
325 }
326
1da177e4 327 spin_lock_irqsave(&cwq->lock, flags);
1da177e4
LT
328 }
329 cwq->run_depth--;
330 spin_unlock_irqrestore(&cwq->lock, flags);
331}
332
333static int worker_thread(void *__cwq)
334{
335 struct cpu_workqueue_struct *cwq = __cwq;
336 DECLARE_WAITQUEUE(wait, current);
337 struct k_sigaction sa;
338 sigset_t blocked;
339
341a5958
RW
340 if (!cwq->freezeable)
341 current->flags |= PF_NOFREEZE;
1da177e4
LT
342
343 set_user_nice(current, -5);
344
345 /* Block and flush all signals */
346 sigfillset(&blocked);
347 sigprocmask(SIG_BLOCK, &blocked, NULL);
348 flush_signals(current);
349
46934023
CL
350 /*
351 * We inherited MPOL_INTERLEAVE from the booting kernel.
352 * Set MPOL_DEFAULT to insure node local allocations.
353 */
354 numa_default_policy();
355
1da177e4
LT
356 /* SIG_IGN makes children autoreap: see do_notify_parent(). */
357 sa.sa.sa_handler = SIG_IGN;
358 sa.sa.sa_flags = 0;
359 siginitset(&sa.sa.sa_mask, sigmask(SIGCHLD));
360 do_sigaction(SIGCHLD, &sa, (struct k_sigaction *)0);
361
362 set_current_state(TASK_INTERRUPTIBLE);
363 while (!kthread_should_stop()) {
341a5958
RW
364 if (cwq->freezeable)
365 try_to_freeze();
366
1da177e4
LT
367 add_wait_queue(&cwq->more_work, &wait);
368 if (list_empty(&cwq->worklist))
369 schedule();
370 else
371 __set_current_state(TASK_RUNNING);
372 remove_wait_queue(&cwq->more_work, &wait);
373
374 if (!list_empty(&cwq->worklist))
375 run_workqueue(cwq);
376 set_current_state(TASK_INTERRUPTIBLE);
377 }
378 __set_current_state(TASK_RUNNING);
379 return 0;
380}
381
fc2e4d70
ON
382struct wq_barrier {
383 struct work_struct work;
384 struct completion done;
385};
386
387static void wq_barrier_func(struct work_struct *work)
388{
389 struct wq_barrier *barr = container_of(work, struct wq_barrier, work);
390 complete(&barr->done);
391}
392
393static inline void init_wq_barrier(struct wq_barrier *barr)
394{
395 INIT_WORK(&barr->work, wq_barrier_func);
396 __set_bit(WORK_STRUCT_PENDING, work_data_bits(&barr->work));
397
398 init_completion(&barr->done);
399}
400
1da177e4
LT
401static void flush_cpu_workqueue(struct cpu_workqueue_struct *cwq)
402{
403 if (cwq->thread == current) {
404 /*
405 * Probably keventd trying to flush its own queue. So simply run
406 * it by hand rather than deadlocking.
407 */
fc2e4d70 408 mutex_unlock(&workqueue_mutex);
1da177e4 409 run_workqueue(cwq);
fc2e4d70 410 mutex_lock(&workqueue_mutex);
1da177e4 411 } else {
fc2e4d70 412 struct wq_barrier barr;
1da177e4 413
fc2e4d70
ON
414 init_wq_barrier(&barr);
415 __queue_work(cwq, &barr.work);
1da177e4 416
fc2e4d70
ON
417 mutex_unlock(&workqueue_mutex);
418 wait_for_completion(&barr.done);
419 mutex_lock(&workqueue_mutex);
1da177e4
LT
420 }
421}
422
0fcb78c2 423/**
1da177e4 424 * flush_workqueue - ensure that any scheduled work has run to completion.
0fcb78c2 425 * @wq: workqueue to flush
1da177e4
LT
426 *
427 * Forces execution of the workqueue and blocks until its completion.
428 * This is typically used in driver shutdown handlers.
429 *
fc2e4d70
ON
430 * We sleep until all works which were queued on entry have been handled,
431 * but we are not livelocked by new incoming ones.
1da177e4
LT
432 *
433 * This function used to run the workqueues itself. Now we just wait for the
434 * helper threads to do it.
435 */
436void fastcall flush_workqueue(struct workqueue_struct *wq)
437{
fc2e4d70 438 mutex_lock(&workqueue_mutex);
1da177e4 439 if (is_single_threaded(wq)) {
bce61dd4 440 /* Always use first cpu's area. */
f756d5e2 441 flush_cpu_workqueue(per_cpu_ptr(wq->cpu_wq, singlethread_cpu));
1da177e4
LT
442 } else {
443 int cpu;
444
1da177e4 445 for_each_online_cpu(cpu)
89ada679 446 flush_cpu_workqueue(per_cpu_ptr(wq->cpu_wq, cpu));
1da177e4 447 }
fc2e4d70 448 mutex_unlock(&workqueue_mutex);
1da177e4 449}
ae90dd5d 450EXPORT_SYMBOL_GPL(flush_workqueue);
1da177e4
LT
451
452static struct task_struct *create_workqueue_thread(struct workqueue_struct *wq,
341a5958 453 int cpu, int freezeable)
1da177e4 454{
89ada679 455 struct cpu_workqueue_struct *cwq = per_cpu_ptr(wq->cpu_wq, cpu);
1da177e4
LT
456 struct task_struct *p;
457
458 spin_lock_init(&cwq->lock);
459 cwq->wq = wq;
460 cwq->thread = NULL;
341a5958 461 cwq->freezeable = freezeable;
1da177e4
LT
462 INIT_LIST_HEAD(&cwq->worklist);
463 init_waitqueue_head(&cwq->more_work);
1da177e4
LT
464
465 if (is_single_threaded(wq))
466 p = kthread_create(worker_thread, cwq, "%s", wq->name);
467 else
468 p = kthread_create(worker_thread, cwq, "%s/%d", wq->name, cpu);
469 if (IS_ERR(p))
470 return NULL;
471 cwq->thread = p;
472 return p;
473}
474
475struct workqueue_struct *__create_workqueue(const char *name,
341a5958 476 int singlethread, int freezeable)
1da177e4
LT
477{
478 int cpu, destroy = 0;
479 struct workqueue_struct *wq;
480 struct task_struct *p;
481
dd392710 482 wq = kzalloc(sizeof(*wq), GFP_KERNEL);
1da177e4
LT
483 if (!wq)
484 return NULL;
1da177e4 485
89ada679 486 wq->cpu_wq = alloc_percpu(struct cpu_workqueue_struct);
676121fc
BC
487 if (!wq->cpu_wq) {
488 kfree(wq);
489 return NULL;
490 }
491
1da177e4 492 wq->name = name;
9b41ea72 493 mutex_lock(&workqueue_mutex);
1da177e4
LT
494 if (singlethread) {
495 INIT_LIST_HEAD(&wq->list);
341a5958 496 p = create_workqueue_thread(wq, singlethread_cpu, freezeable);
1da177e4
LT
497 if (!p)
498 destroy = 1;
499 else
500 wake_up_process(p);
501 } else {
1da177e4 502 list_add(&wq->list, &workqueues);
1da177e4 503 for_each_online_cpu(cpu) {
341a5958 504 p = create_workqueue_thread(wq, cpu, freezeable);
1da177e4
LT
505 if (p) {
506 kthread_bind(p, cpu);
507 wake_up_process(p);
508 } else
509 destroy = 1;
510 }
511 }
9b41ea72 512 mutex_unlock(&workqueue_mutex);
1da177e4
LT
513
514 /*
515 * Was there any error during startup? If yes then clean up:
516 */
517 if (destroy) {
518 destroy_workqueue(wq);
519 wq = NULL;
520 }
521 return wq;
522}
ae90dd5d 523EXPORT_SYMBOL_GPL(__create_workqueue);
1da177e4
LT
524
525static void cleanup_workqueue_thread(struct workqueue_struct *wq, int cpu)
526{
527 struct cpu_workqueue_struct *cwq;
528 unsigned long flags;
529 struct task_struct *p;
530
89ada679 531 cwq = per_cpu_ptr(wq->cpu_wq, cpu);
1da177e4
LT
532 spin_lock_irqsave(&cwq->lock, flags);
533 p = cwq->thread;
534 cwq->thread = NULL;
535 spin_unlock_irqrestore(&cwq->lock, flags);
536 if (p)
537 kthread_stop(p);
538}
539
0fcb78c2
REB
540/**
541 * destroy_workqueue - safely terminate a workqueue
542 * @wq: target workqueue
543 *
544 * Safely destroy a workqueue. All work currently pending will be done first.
545 */
1da177e4
LT
546void destroy_workqueue(struct workqueue_struct *wq)
547{
548 int cpu;
549
550 flush_workqueue(wq);
551
552 /* We don't need the distraction of CPUs appearing and vanishing. */
9b41ea72 553 mutex_lock(&workqueue_mutex);
1da177e4 554 if (is_single_threaded(wq))
f756d5e2 555 cleanup_workqueue_thread(wq, singlethread_cpu);
1da177e4
LT
556 else {
557 for_each_online_cpu(cpu)
558 cleanup_workqueue_thread(wq, cpu);
1da177e4 559 list_del(&wq->list);
1da177e4 560 }
9b41ea72 561 mutex_unlock(&workqueue_mutex);
89ada679 562 free_percpu(wq->cpu_wq);
1da177e4
LT
563 kfree(wq);
564}
ae90dd5d 565EXPORT_SYMBOL_GPL(destroy_workqueue);
1da177e4
LT
566
567static struct workqueue_struct *keventd_wq;
568
0fcb78c2
REB
569/**
570 * schedule_work - put work task in global workqueue
571 * @work: job to be done
572 *
573 * This puts a job in the kernel-global workqueue.
574 */
1da177e4
LT
575int fastcall schedule_work(struct work_struct *work)
576{
577 return queue_work(keventd_wq, work);
578}
ae90dd5d 579EXPORT_SYMBOL(schedule_work);
1da177e4 580
0fcb78c2
REB
581/**
582 * schedule_delayed_work - put work task in global workqueue after delay
52bad64d
DH
583 * @dwork: job to be done
584 * @delay: number of jiffies to wait or 0 for immediate execution
0fcb78c2
REB
585 *
586 * After waiting for a given time this puts a job in the kernel-global
587 * workqueue.
588 */
82f67cd9
IM
589int fastcall schedule_delayed_work(struct delayed_work *dwork,
590 unsigned long delay)
1da177e4 591{
82f67cd9 592 timer_stats_timer_set_start_info(&dwork->timer);
52bad64d 593 return queue_delayed_work(keventd_wq, dwork, delay);
1da177e4 594}
ae90dd5d 595EXPORT_SYMBOL(schedule_delayed_work);
1da177e4 596
0fcb78c2
REB
597/**
598 * schedule_delayed_work_on - queue work in global workqueue on CPU after delay
599 * @cpu: cpu to use
52bad64d 600 * @dwork: job to be done
0fcb78c2
REB
601 * @delay: number of jiffies to wait
602 *
603 * After waiting for a given time this puts a job in the kernel-global
604 * workqueue on the specified CPU.
605 */
1da177e4 606int schedule_delayed_work_on(int cpu,
52bad64d 607 struct delayed_work *dwork, unsigned long delay)
1da177e4 608{
52bad64d 609 return queue_delayed_work_on(cpu, keventd_wq, dwork, delay);
1da177e4 610}
ae90dd5d 611EXPORT_SYMBOL(schedule_delayed_work_on);
1da177e4 612
b6136773
AM
613/**
614 * schedule_on_each_cpu - call a function on each online CPU from keventd
615 * @func: the function to call
b6136773
AM
616 *
617 * Returns zero on success.
618 * Returns -ve errno on failure.
619 *
620 * Appears to be racy against CPU hotplug.
621 *
622 * schedule_on_each_cpu() is very slow.
623 */
65f27f38 624int schedule_on_each_cpu(work_func_t func)
15316ba8
CL
625{
626 int cpu;
b6136773 627 struct work_struct *works;
15316ba8 628
b6136773
AM
629 works = alloc_percpu(struct work_struct);
630 if (!works)
15316ba8 631 return -ENOMEM;
b6136773 632
e18f3ffb 633 preempt_disable(); /* CPU hotplug */
15316ba8 634 for_each_online_cpu(cpu) {
9bfb1839
IM
635 struct work_struct *work = per_cpu_ptr(works, cpu);
636
637 INIT_WORK(work, func);
638 set_bit(WORK_STRUCT_PENDING, work_data_bits(work));
639 __queue_work(per_cpu_ptr(keventd_wq->cpu_wq, cpu), work);
15316ba8 640 }
e18f3ffb 641 preempt_enable();
15316ba8 642 flush_workqueue(keventd_wq);
b6136773 643 free_percpu(works);
15316ba8
CL
644 return 0;
645}
646
1da177e4
LT
647void flush_scheduled_work(void)
648{
649 flush_workqueue(keventd_wq);
650}
ae90dd5d 651EXPORT_SYMBOL(flush_scheduled_work);
1da177e4
LT
652
653/**
72fd4a35 654 * cancel_rearming_delayed_workqueue - reliably kill off a delayed work whose handler rearms the delayed work.
1da177e4 655 * @wq: the controlling workqueue structure
52bad64d 656 * @dwork: the delayed work struct
1da177e4 657 */
81ddef77 658void cancel_rearming_delayed_workqueue(struct workqueue_struct *wq,
52bad64d 659 struct delayed_work *dwork)
1da177e4 660{
52bad64d 661 while (!cancel_delayed_work(dwork))
1da177e4
LT
662 flush_workqueue(wq);
663}
81ddef77 664EXPORT_SYMBOL(cancel_rearming_delayed_workqueue);
1da177e4
LT
665
666/**
72fd4a35 667 * cancel_rearming_delayed_work - reliably kill off a delayed keventd work whose handler rearms the delayed work.
52bad64d 668 * @dwork: the delayed work struct
1da177e4 669 */
52bad64d 670void cancel_rearming_delayed_work(struct delayed_work *dwork)
1da177e4 671{
52bad64d 672 cancel_rearming_delayed_workqueue(keventd_wq, dwork);
1da177e4
LT
673}
674EXPORT_SYMBOL(cancel_rearming_delayed_work);
675
1fa44eca
JB
676/**
677 * execute_in_process_context - reliably execute the routine with user context
678 * @fn: the function to execute
1fa44eca
JB
679 * @ew: guaranteed storage for the execute work structure (must
680 * be available when the work executes)
681 *
682 * Executes the function immediately if process context is available,
683 * otherwise schedules the function for delayed execution.
684 *
685 * Returns: 0 - function was executed
686 * 1 - function was scheduled for execution
687 */
65f27f38 688int execute_in_process_context(work_func_t fn, struct execute_work *ew)
1fa44eca
JB
689{
690 if (!in_interrupt()) {
65f27f38 691 fn(&ew->work);
1fa44eca
JB
692 return 0;
693 }
694
65f27f38 695 INIT_WORK(&ew->work, fn);
1fa44eca
JB
696 schedule_work(&ew->work);
697
698 return 1;
699}
700EXPORT_SYMBOL_GPL(execute_in_process_context);
701
1da177e4
LT
702int keventd_up(void)
703{
704 return keventd_wq != NULL;
705}
706
707int current_is_keventd(void)
708{
709 struct cpu_workqueue_struct *cwq;
710 int cpu = smp_processor_id(); /* preempt-safe: keventd is per-cpu */
711 int ret = 0;
712
713 BUG_ON(!keventd_wq);
714
89ada679 715 cwq = per_cpu_ptr(keventd_wq->cpu_wq, cpu);
1da177e4
LT
716 if (current == cwq->thread)
717 ret = 1;
718
719 return ret;
720
721}
722
1da177e4
LT
723/* Take the work from this (downed) CPU. */
724static void take_over_work(struct workqueue_struct *wq, unsigned int cpu)
725{
89ada679 726 struct cpu_workqueue_struct *cwq = per_cpu_ptr(wq->cpu_wq, cpu);
626ab0e6 727 struct list_head list;
1da177e4
LT
728 struct work_struct *work;
729
730 spin_lock_irq(&cwq->lock);
626ab0e6 731 list_replace_init(&cwq->worklist, &list);
1da177e4
LT
732
733 while (!list_empty(&list)) {
734 printk("Taking work for %s\n", wq->name);
735 work = list_entry(list.next,struct work_struct,entry);
736 list_del(&work->entry);
89ada679 737 __queue_work(per_cpu_ptr(wq->cpu_wq, smp_processor_id()), work);
1da177e4
LT
738 }
739 spin_unlock_irq(&cwq->lock);
740}
741
742/* We're holding the cpucontrol mutex here */
9c7b216d 743static int __devinit workqueue_cpu_callback(struct notifier_block *nfb,
1da177e4
LT
744 unsigned long action,
745 void *hcpu)
746{
747 unsigned int hotcpu = (unsigned long)hcpu;
748 struct workqueue_struct *wq;
749
750 switch (action) {
751 case CPU_UP_PREPARE:
9b41ea72 752 mutex_lock(&workqueue_mutex);
1da177e4
LT
753 /* Create a new workqueue thread for it. */
754 list_for_each_entry(wq, &workqueues, list) {
341a5958 755 if (!create_workqueue_thread(wq, hotcpu, 0)) {
1da177e4
LT
756 printk("workqueue for %i failed\n", hotcpu);
757 return NOTIFY_BAD;
758 }
759 }
760 break;
761
762 case CPU_ONLINE:
763 /* Kick off worker threads. */
764 list_for_each_entry(wq, &workqueues, list) {
89ada679
CL
765 struct cpu_workqueue_struct *cwq;
766
767 cwq = per_cpu_ptr(wq->cpu_wq, hotcpu);
768 kthread_bind(cwq->thread, hotcpu);
769 wake_up_process(cwq->thread);
1da177e4 770 }
9b41ea72 771 mutex_unlock(&workqueue_mutex);
1da177e4
LT
772 break;
773
774 case CPU_UP_CANCELED:
775 list_for_each_entry(wq, &workqueues, list) {
fc75cdfa
HC
776 if (!per_cpu_ptr(wq->cpu_wq, hotcpu)->thread)
777 continue;
1da177e4 778 /* Unbind so it can run. */
89ada679 779 kthread_bind(per_cpu_ptr(wq->cpu_wq, hotcpu)->thread,
a4c4af7c 780 any_online_cpu(cpu_online_map));
1da177e4
LT
781 cleanup_workqueue_thread(wq, hotcpu);
782 }
9b41ea72
AM
783 mutex_unlock(&workqueue_mutex);
784 break;
785
786 case CPU_DOWN_PREPARE:
787 mutex_lock(&workqueue_mutex);
788 break;
789
790 case CPU_DOWN_FAILED:
791 mutex_unlock(&workqueue_mutex);
1da177e4
LT
792 break;
793
794 case CPU_DEAD:
795 list_for_each_entry(wq, &workqueues, list)
796 cleanup_workqueue_thread(wq, hotcpu);
797 list_for_each_entry(wq, &workqueues, list)
798 take_over_work(wq, hotcpu);
9b41ea72 799 mutex_unlock(&workqueue_mutex);
1da177e4
LT
800 break;
801 }
802
803 return NOTIFY_OK;
804}
1da177e4
LT
805
806void init_workqueues(void)
807{
f756d5e2 808 singlethread_cpu = first_cpu(cpu_possible_map);
1da177e4
LT
809 hotcpu_notifier(workqueue_cpu_callback, 0);
810 keventd_wq = create_workqueue("events");
811 BUG_ON(!keventd_wq);
812}
813