1 // SPDX-License-Identifier: GPL-2.0
3 * Basic worker thread pool for io_uring
5 * Copyright (C) 2019 Jens Axboe
8 #include <linux/kernel.h>
9 #include <linux/init.h>
10 #include <linux/errno.h>
11 #include <linux/sched/signal.h>
13 #include <linux/mmu_context.h>
14 #include <linux/sched/mm.h>
15 #include <linux/percpu.h>
16 #include <linux/slab.h>
17 #include <linux/kthread.h>
18 #include <linux/rculist_nulls.h>
22 #define WORKER_IDLE_TIMEOUT (5 * HZ)
25 IO_WORKER_F_UP = 1, /* up and active */
26 IO_WORKER_F_RUNNING = 2, /* account as running */
27 IO_WORKER_F_FREE = 4, /* worker on free list */
28 IO_WORKER_F_EXITING = 8, /* worker exiting */
29 IO_WORKER_F_FIXED = 16, /* static idle worker */
30 IO_WORKER_F_BOUND = 32, /* is doing bounded work */
34 IO_WQ_BIT_EXIT = 0, /* wq exiting */
35 IO_WQ_BIT_CANCEL = 1, /* cancel work on list */
36 IO_WQ_BIT_ERROR = 2, /* error on setup */
40 IO_WQE_FLAG_STALLED = 1, /* stalled on hash */
44 * One for each thread in a wqe pool
49 struct hlist_nulls_node nulls_node;
50 struct list_head all_list;
51 struct task_struct *task;
54 struct io_wq_work *cur_work;
59 const struct cred *cur_creds;
60 const struct cred *saved_creds;
61 struct files_struct *restore_files;
64 #if BITS_PER_LONG == 64
65 #define IO_WQ_HASH_ORDER 6
67 #define IO_WQ_HASH_ORDER 5
82 * Per-node worker thread pool
87 struct io_wq_work_list work_list;
88 unsigned long hash_map;
90 } ____cacheline_aligned_in_smp;
93 struct io_wqe_acct acct[2];
95 struct hlist_nulls_head free_list;
96 struct list_head all_list;
105 struct io_wqe **wqes;
108 get_work_fn *get_work;
109 put_work_fn *put_work;
111 struct task_struct *manager;
112 struct user_struct *user;
114 struct completion done;
119 static bool io_worker_get(struct io_worker *worker)
121 return refcount_inc_not_zero(&worker->ref);
124 static void io_worker_release(struct io_worker *worker)
126 if (refcount_dec_and_test(&worker->ref))
127 wake_up_process(worker->task);
131 * Note: drops the wqe->lock if returning true! The caller must re-acquire
132 * the lock in that case. Some callers need to restart handling if this
133 * happens, so we can't just re-acquire the lock on behalf of the caller.
135 static bool __io_worker_unuse(struct io_wqe *wqe, struct io_worker *worker)
137 bool dropped_lock = false;
139 if (worker->saved_creds) {
140 revert_creds(worker->saved_creds);
141 worker->cur_creds = worker->saved_creds = NULL;
144 if (current->files != worker->restore_files) {
145 __acquire(&wqe->lock);
146 spin_unlock_irq(&wqe->lock);
150 current->files = worker->restore_files;
151 task_unlock(current);
155 * If we have an active mm, we need to drop the wq lock before unusing
156 * it. If we do, return true and let the caller retry the idle loop.
160 __acquire(&wqe->lock);
161 spin_unlock_irq(&wqe->lock);
164 __set_current_state(TASK_RUNNING);
166 unuse_mm(worker->mm);
174 static inline struct io_wqe_acct *io_work_get_acct(struct io_wqe *wqe,
175 struct io_wq_work *work)
177 if (work->flags & IO_WQ_WORK_UNBOUND)
178 return &wqe->acct[IO_WQ_ACCT_UNBOUND];
180 return &wqe->acct[IO_WQ_ACCT_BOUND];
183 static inline struct io_wqe_acct *io_wqe_get_acct(struct io_wqe *wqe,
184 struct io_worker *worker)
186 if (worker->flags & IO_WORKER_F_BOUND)
187 return &wqe->acct[IO_WQ_ACCT_BOUND];
189 return &wqe->acct[IO_WQ_ACCT_UNBOUND];
192 static void io_worker_exit(struct io_worker *worker)
194 struct io_wqe *wqe = worker->wqe;
195 struct io_wqe_acct *acct = io_wqe_get_acct(wqe, worker);
199 * If we're not at zero, someone else is holding a brief reference
200 * to the worker. Wait for that to go away.
202 set_current_state(TASK_INTERRUPTIBLE);
203 if (!refcount_dec_and_test(&worker->ref))
205 __set_current_state(TASK_RUNNING);
208 current->flags &= ~PF_IO_WORKER;
209 if (worker->flags & IO_WORKER_F_RUNNING)
210 atomic_dec(&acct->nr_running);
211 if (!(worker->flags & IO_WORKER_F_BOUND))
212 atomic_dec(&wqe->wq->user->processes);
216 spin_lock_irq(&wqe->lock);
217 hlist_nulls_del_rcu(&worker->nulls_node);
218 list_del_rcu(&worker->all_list);
219 if (__io_worker_unuse(wqe, worker)) {
220 __release(&wqe->lock);
221 spin_lock_irq(&wqe->lock);
224 nr_workers = wqe->acct[IO_WQ_ACCT_BOUND].nr_workers +
225 wqe->acct[IO_WQ_ACCT_UNBOUND].nr_workers;
226 spin_unlock_irq(&wqe->lock);
228 /* all workers gone, wq exit can proceed */
229 if (!nr_workers && refcount_dec_and_test(&wqe->wq->refs))
230 complete(&wqe->wq->done);
232 kfree_rcu(worker, rcu);
235 static inline bool io_wqe_run_queue(struct io_wqe *wqe)
236 __must_hold(wqe->lock)
238 if (!wq_list_empty(&wqe->work_list) &&
239 !(wqe->flags & IO_WQE_FLAG_STALLED))
245 * Check head of free list for an available worker. If one isn't available,
246 * caller must wake up the wq manager to create one.
248 static bool io_wqe_activate_free_worker(struct io_wqe *wqe)
251 struct hlist_nulls_node *n;
252 struct io_worker *worker;
254 n = rcu_dereference(hlist_nulls_first_rcu(&wqe->free_list));
258 worker = hlist_nulls_entry(n, struct io_worker, nulls_node);
259 if (io_worker_get(worker)) {
260 wake_up_process(worker->task);
261 io_worker_release(worker);
269 * We need a worker. If we find a free one, we're good. If not, and we're
270 * below the max number of workers, wake up the manager to create one.
272 static void io_wqe_wake_worker(struct io_wqe *wqe, struct io_wqe_acct *acct)
277 * Most likely an attempt to queue unbounded work on an io_wq that
278 * wasn't setup with any unbounded workers.
280 WARN_ON_ONCE(!acct->max_workers);
283 ret = io_wqe_activate_free_worker(wqe);
286 if (!ret && acct->nr_workers < acct->max_workers)
287 wake_up_process(wqe->wq->manager);
290 static void io_wqe_inc_running(struct io_wqe *wqe, struct io_worker *worker)
292 struct io_wqe_acct *acct = io_wqe_get_acct(wqe, worker);
294 atomic_inc(&acct->nr_running);
297 static void io_wqe_dec_running(struct io_wqe *wqe, struct io_worker *worker)
298 __must_hold(wqe->lock)
300 struct io_wqe_acct *acct = io_wqe_get_acct(wqe, worker);
302 if (atomic_dec_and_test(&acct->nr_running) && io_wqe_run_queue(wqe))
303 io_wqe_wake_worker(wqe, acct);
306 static void io_worker_start(struct io_wqe *wqe, struct io_worker *worker)
308 allow_kernel_signal(SIGINT);
310 current->flags |= PF_IO_WORKER;
312 worker->flags |= (IO_WORKER_F_UP | IO_WORKER_F_RUNNING);
313 worker->restore_files = current->files;
314 io_wqe_inc_running(wqe, worker);
318 * Worker will start processing some work. Move it to the busy list, if
319 * it's currently on the freelist
321 static void __io_worker_busy(struct io_wqe *wqe, struct io_worker *worker,
322 struct io_wq_work *work)
323 __must_hold(wqe->lock)
325 bool worker_bound, work_bound;
327 if (worker->flags & IO_WORKER_F_FREE) {
328 worker->flags &= ~IO_WORKER_F_FREE;
329 hlist_nulls_del_init_rcu(&worker->nulls_node);
333 * If worker is moving from bound to unbound (or vice versa), then
334 * ensure we update the running accounting.
336 worker_bound = (worker->flags & IO_WORKER_F_BOUND) != 0;
337 work_bound = (work->flags & IO_WQ_WORK_UNBOUND) == 0;
338 if (worker_bound != work_bound) {
339 io_wqe_dec_running(wqe, worker);
341 worker->flags |= IO_WORKER_F_BOUND;
342 wqe->acct[IO_WQ_ACCT_UNBOUND].nr_workers--;
343 wqe->acct[IO_WQ_ACCT_BOUND].nr_workers++;
344 atomic_dec(&wqe->wq->user->processes);
346 worker->flags &= ~IO_WORKER_F_BOUND;
347 wqe->acct[IO_WQ_ACCT_UNBOUND].nr_workers++;
348 wqe->acct[IO_WQ_ACCT_BOUND].nr_workers--;
349 atomic_inc(&wqe->wq->user->processes);
351 io_wqe_inc_running(wqe, worker);
356 * No work, worker going to sleep. Move to freelist, and unuse mm if we
357 * have one attached. Dropping the mm may potentially sleep, so we drop
358 * the lock in that case and return success. Since the caller has to
359 * retry the loop in that case (we changed task state), we don't regrab
360 * the lock if we return success.
362 static bool __io_worker_idle(struct io_wqe *wqe, struct io_worker *worker)
363 __must_hold(wqe->lock)
365 if (!(worker->flags & IO_WORKER_F_FREE)) {
366 worker->flags |= IO_WORKER_F_FREE;
367 hlist_nulls_add_head_rcu(&worker->nulls_node, &wqe->free_list);
370 return __io_worker_unuse(wqe, worker);
373 static struct io_wq_work *io_get_next_work(struct io_wqe *wqe, unsigned *hash)
374 __must_hold(wqe->lock)
376 struct io_wq_work_node *node, *prev;
377 struct io_wq_work *work;
379 wq_list_for_each(node, prev, &wqe->work_list) {
380 work = container_of(node, struct io_wq_work, list);
382 /* not hashed, can run anytime */
383 if (!(work->flags & IO_WQ_WORK_HASHED)) {
384 wq_node_del(&wqe->work_list, node, prev);
388 /* hashed, can run if not already running */
389 *hash = work->flags >> IO_WQ_HASH_SHIFT;
390 if (!(wqe->hash_map & BIT_ULL(*hash))) {
391 wqe->hash_map |= BIT_ULL(*hash);
392 wq_node_del(&wqe->work_list, node, prev);
400 static void io_wq_switch_mm(struct io_worker *worker, struct io_wq_work *work)
403 unuse_mm(worker->mm);
411 if (mmget_not_zero(work->mm)) {
415 worker->mm = work->mm;
416 /* hang on to this mm */
421 /* failed grabbing mm, ensure work gets cancelled */
422 work->flags |= IO_WQ_WORK_CANCEL;
425 static void io_wq_switch_creds(struct io_worker *worker,
426 struct io_wq_work *work)
428 const struct cred *old_creds = override_creds(work->creds);
430 worker->cur_creds = work->creds;
431 if (worker->saved_creds)
432 put_cred(old_creds); /* creds set by previous switch */
434 worker->saved_creds = old_creds;
437 static void io_worker_handle_work(struct io_worker *worker)
438 __releases(wqe->lock)
440 struct io_wq_work *work, *old_work = NULL, *put_work = NULL;
441 struct io_wqe *wqe = worker->wqe;
442 struct io_wq *wq = wqe->wq;
448 * If we got some work, mark us as busy. If we didn't, but
449 * the list isn't empty, it means we stalled on hashed work.
450 * Mark us stalled so we don't keep looking for work when we
451 * can't make progress, any work completion or insertion will
452 * clear the stalled flag.
454 work = io_get_next_work(wqe, &hash);
456 __io_worker_busy(wqe, worker, work);
457 else if (!wq_list_empty(&wqe->work_list))
458 wqe->flags |= IO_WQE_FLAG_STALLED;
460 spin_unlock_irq(&wqe->lock);
461 if (put_work && wq->put_work)
462 wq->put_work(old_work);
466 /* flush any pending signals before assigning new work */
467 if (signal_pending(current))
468 flush_signals(current);
472 spin_lock_irq(&worker->lock);
473 worker->cur_work = work;
474 spin_unlock_irq(&worker->lock);
476 if (work->flags & IO_WQ_WORK_CB)
479 if ((work->flags & IO_WQ_WORK_NEEDS_FILES) &&
480 current->files != work->files) {
482 current->files = work->files;
483 task_unlock(current);
485 if (work->mm != worker->mm)
486 io_wq_switch_mm(worker, work);
487 if (worker->cur_creds != work->creds)
488 io_wq_switch_creds(worker, work);
490 * OK to set IO_WQ_WORK_CANCEL even for uncancellable work,
491 * the worker function will do the right thing.
493 if (test_bit(IO_WQ_BIT_CANCEL, &wq->state))
494 work->flags |= IO_WQ_WORK_CANCEL;
496 work->flags |= IO_WQ_WORK_HAS_MM;
498 if (wq->get_work && !(work->flags & IO_WQ_WORK_INTERNAL)) {
506 spin_lock_irq(&worker->lock);
507 worker->cur_work = NULL;
508 spin_unlock_irq(&worker->lock);
510 spin_lock_irq(&wqe->lock);
513 wqe->hash_map &= ~BIT_ULL(hash);
514 wqe->flags &= ~IO_WQE_FLAG_STALLED;
516 if (work && work != old_work) {
517 spin_unlock_irq(&wqe->lock);
519 if (put_work && wq->put_work) {
520 wq->put_work(put_work);
524 /* dependent work not hashed */
531 static inline void io_worker_spin_for_work(struct io_wqe *wqe)
536 if (io_wqe_run_queue(wqe))
544 static int io_wqe_worker(void *data)
546 struct io_worker *worker = data;
547 struct io_wqe *wqe = worker->wqe;
548 struct io_wq *wq = wqe->wq;
551 io_worker_start(wqe, worker);
554 while (!test_bit(IO_WQ_BIT_EXIT, &wq->state)) {
555 set_current_state(TASK_INTERRUPTIBLE);
558 io_worker_spin_for_work(wqe);
559 spin_lock_irq(&wqe->lock);
560 if (io_wqe_run_queue(wqe)) {
561 __set_current_state(TASK_RUNNING);
562 io_worker_handle_work(worker);
567 /* drops the lock on success, retry */
568 if (__io_worker_idle(wqe, worker)) {
569 __release(&wqe->lock);
572 spin_unlock_irq(&wqe->lock);
573 if (signal_pending(current))
574 flush_signals(current);
575 if (schedule_timeout(WORKER_IDLE_TIMEOUT))
577 /* timed out, exit unless we're the fixed worker */
578 if (test_bit(IO_WQ_BIT_EXIT, &wq->state) ||
579 !(worker->flags & IO_WORKER_F_FIXED))
583 if (test_bit(IO_WQ_BIT_EXIT, &wq->state)) {
584 spin_lock_irq(&wqe->lock);
585 if (!wq_list_empty(&wqe->work_list))
586 io_worker_handle_work(worker);
588 spin_unlock_irq(&wqe->lock);
591 io_worker_exit(worker);
596 * Called when a worker is scheduled in. Mark us as currently running.
598 void io_wq_worker_running(struct task_struct *tsk)
600 struct io_worker *worker = kthread_data(tsk);
601 struct io_wqe *wqe = worker->wqe;
603 if (!(worker->flags & IO_WORKER_F_UP))
605 if (worker->flags & IO_WORKER_F_RUNNING)
607 worker->flags |= IO_WORKER_F_RUNNING;
608 io_wqe_inc_running(wqe, worker);
612 * Called when worker is going to sleep. If there are no workers currently
613 * running and we have work pending, wake up a free one or have the manager
616 void io_wq_worker_sleeping(struct task_struct *tsk)
618 struct io_worker *worker = kthread_data(tsk);
619 struct io_wqe *wqe = worker->wqe;
621 if (!(worker->flags & IO_WORKER_F_UP))
623 if (!(worker->flags & IO_WORKER_F_RUNNING))
626 worker->flags &= ~IO_WORKER_F_RUNNING;
628 spin_lock_irq(&wqe->lock);
629 io_wqe_dec_running(wqe, worker);
630 spin_unlock_irq(&wqe->lock);
633 static bool create_io_worker(struct io_wq *wq, struct io_wqe *wqe, int index)
635 struct io_wqe_acct *acct =&wqe->acct[index];
636 struct io_worker *worker;
638 worker = kzalloc_node(sizeof(*worker), GFP_KERNEL, wqe->node);
642 refcount_set(&worker->ref, 1);
643 worker->nulls_node.pprev = NULL;
645 spin_lock_init(&worker->lock);
647 worker->task = kthread_create_on_node(io_wqe_worker, worker, wqe->node,
648 "io_wqe_worker-%d/%d", index, wqe->node);
649 if (IS_ERR(worker->task)) {
654 spin_lock_irq(&wqe->lock);
655 hlist_nulls_add_head_rcu(&worker->nulls_node, &wqe->free_list);
656 list_add_tail_rcu(&worker->all_list, &wqe->all_list);
657 worker->flags |= IO_WORKER_F_FREE;
658 if (index == IO_WQ_ACCT_BOUND)
659 worker->flags |= IO_WORKER_F_BOUND;
660 if (!acct->nr_workers && (worker->flags & IO_WORKER_F_BOUND))
661 worker->flags |= IO_WORKER_F_FIXED;
663 spin_unlock_irq(&wqe->lock);
665 if (index == IO_WQ_ACCT_UNBOUND)
666 atomic_inc(&wq->user->processes);
668 wake_up_process(worker->task);
672 static inline bool io_wqe_need_worker(struct io_wqe *wqe, int index)
673 __must_hold(wqe->lock)
675 struct io_wqe_acct *acct = &wqe->acct[index];
677 /* if we have available workers or no work, no need */
678 if (!hlist_nulls_empty(&wqe->free_list) || !io_wqe_run_queue(wqe))
680 return acct->nr_workers < acct->max_workers;
684 * Manager thread. Tasked with creating new workers, if we need them.
686 static int io_wq_manager(void *data)
688 struct io_wq *wq = data;
689 int workers_to_create = num_possible_nodes();
692 /* create fixed workers */
693 refcount_set(&wq->refs, workers_to_create);
694 for_each_node(node) {
695 if (!create_io_worker(wq, wq->wqes[node], IO_WQ_ACCT_BOUND))
702 while (!kthread_should_stop()) {
703 for_each_node(node) {
704 struct io_wqe *wqe = wq->wqes[node];
705 bool fork_worker[2] = { false, false };
707 spin_lock_irq(&wqe->lock);
708 if (io_wqe_need_worker(wqe, IO_WQ_ACCT_BOUND))
709 fork_worker[IO_WQ_ACCT_BOUND] = true;
710 if (io_wqe_need_worker(wqe, IO_WQ_ACCT_UNBOUND))
711 fork_worker[IO_WQ_ACCT_UNBOUND] = true;
712 spin_unlock_irq(&wqe->lock);
713 if (fork_worker[IO_WQ_ACCT_BOUND])
714 create_io_worker(wq, wqe, IO_WQ_ACCT_BOUND);
715 if (fork_worker[IO_WQ_ACCT_UNBOUND])
716 create_io_worker(wq, wqe, IO_WQ_ACCT_UNBOUND);
718 set_current_state(TASK_INTERRUPTIBLE);
719 schedule_timeout(HZ);
724 set_bit(IO_WQ_BIT_ERROR, &wq->state);
725 set_bit(IO_WQ_BIT_EXIT, &wq->state);
726 if (refcount_sub_and_test(workers_to_create, &wq->refs))
731 static bool io_wq_can_queue(struct io_wqe *wqe, struct io_wqe_acct *acct,
732 struct io_wq_work *work)
736 if (!(work->flags & IO_WQ_WORK_UNBOUND))
738 if (atomic_read(&acct->nr_running))
742 free_worker = !hlist_nulls_empty(&wqe->free_list);
747 if (atomic_read(&wqe->wq->user->processes) >= acct->max_workers &&
748 !(capable(CAP_SYS_RESOURCE) || capable(CAP_SYS_ADMIN)))
754 static void io_wqe_enqueue(struct io_wqe *wqe, struct io_wq_work *work)
756 struct io_wqe_acct *acct = io_work_get_acct(wqe, work);
761 * Do early check to see if we need a new unbound worker, and if we do,
762 * if we're allowed to do so. This isn't 100% accurate as there's a
763 * gap between this check and incrementing the value, but that's OK.
764 * It's close enough to not be an issue, fork() has the same delay.
766 if (unlikely(!io_wq_can_queue(wqe, acct, work))) {
767 work->flags |= IO_WQ_WORK_CANCEL;
772 work_flags = work->flags;
773 spin_lock_irqsave(&wqe->lock, flags);
774 wq_list_add_tail(&work->list, &wqe->work_list);
775 wqe->flags &= ~IO_WQE_FLAG_STALLED;
776 spin_unlock_irqrestore(&wqe->lock, flags);
778 if ((work_flags & IO_WQ_WORK_CONCURRENT) ||
779 !atomic_read(&acct->nr_running))
780 io_wqe_wake_worker(wqe, acct);
783 void io_wq_enqueue(struct io_wq *wq, struct io_wq_work *work)
785 struct io_wqe *wqe = wq->wqes[numa_node_id()];
787 io_wqe_enqueue(wqe, work);
791 * Enqueue work, hashed by some key. Work items that hash to the same value
792 * will not be done in parallel. Used to limit concurrent writes, generally
795 void io_wq_enqueue_hashed(struct io_wq *wq, struct io_wq_work *work, void *val)
797 struct io_wqe *wqe = wq->wqes[numa_node_id()];
801 bit = hash_ptr(val, IO_WQ_HASH_ORDER);
802 work->flags |= (IO_WQ_WORK_HASHED | (bit << IO_WQ_HASH_SHIFT));
803 io_wqe_enqueue(wqe, work);
806 static bool io_wqe_worker_send_sig(struct io_worker *worker, void *data)
808 send_sig(SIGINT, worker->task, 1);
813 * Iterate the passed in list and call the specific function for each
814 * worker that isn't exiting
816 static bool io_wq_for_each_worker(struct io_wqe *wqe,
817 bool (*func)(struct io_worker *, void *),
820 struct io_worker *worker;
823 list_for_each_entry_rcu(worker, &wqe->all_list, all_list) {
824 if (io_worker_get(worker)) {
825 ret = func(worker, data);
826 io_worker_release(worker);
835 void io_wq_cancel_all(struct io_wq *wq)
839 set_bit(IO_WQ_BIT_CANCEL, &wq->state);
842 for_each_node(node) {
843 struct io_wqe *wqe = wq->wqes[node];
845 io_wq_for_each_worker(wqe, io_wqe_worker_send_sig, NULL);
850 struct io_cb_cancel_data {
852 work_cancel_fn *cancel;
856 static bool io_work_cancel(struct io_worker *worker, void *cancel_data)
858 struct io_cb_cancel_data *data = cancel_data;
863 * Hold the lock to avoid ->cur_work going out of scope, caller
864 * may dereference the passed in work.
866 spin_lock_irqsave(&worker->lock, flags);
867 if (worker->cur_work &&
868 !(worker->cur_work->flags & IO_WQ_WORK_NO_CANCEL) &&
869 data->cancel(worker->cur_work, data->caller_data)) {
870 send_sig(SIGINT, worker->task, 1);
873 spin_unlock_irqrestore(&worker->lock, flags);
878 static enum io_wq_cancel io_wqe_cancel_cb_work(struct io_wqe *wqe,
879 work_cancel_fn *cancel,
882 struct io_cb_cancel_data data = {
885 .caller_data = cancel_data,
887 struct io_wq_work_node *node, *prev;
888 struct io_wq_work *work;
892 spin_lock_irqsave(&wqe->lock, flags);
893 wq_list_for_each(node, prev, &wqe->work_list) {
894 work = container_of(node, struct io_wq_work, list);
896 if (cancel(work, cancel_data)) {
897 wq_node_del(&wqe->work_list, node, prev);
902 spin_unlock_irqrestore(&wqe->lock, flags);
905 work->flags |= IO_WQ_WORK_CANCEL;
907 return IO_WQ_CANCEL_OK;
911 found = io_wq_for_each_worker(wqe, io_work_cancel, &data);
913 return found ? IO_WQ_CANCEL_RUNNING : IO_WQ_CANCEL_NOTFOUND;
916 enum io_wq_cancel io_wq_cancel_cb(struct io_wq *wq, work_cancel_fn *cancel,
919 enum io_wq_cancel ret = IO_WQ_CANCEL_NOTFOUND;
922 for_each_node(node) {
923 struct io_wqe *wqe = wq->wqes[node];
925 ret = io_wqe_cancel_cb_work(wqe, cancel, data);
926 if (ret != IO_WQ_CANCEL_NOTFOUND)
933 static bool io_wq_worker_cancel(struct io_worker *worker, void *data)
935 struct io_wq_work *work = data;
939 if (worker->cur_work != work)
942 spin_lock_irqsave(&worker->lock, flags);
943 if (worker->cur_work == work &&
944 !(worker->cur_work->flags & IO_WQ_WORK_NO_CANCEL)) {
945 send_sig(SIGINT, worker->task, 1);
948 spin_unlock_irqrestore(&worker->lock, flags);
953 static enum io_wq_cancel io_wqe_cancel_work(struct io_wqe *wqe,
954 struct io_wq_work *cwork)
956 struct io_wq_work_node *node, *prev;
957 struct io_wq_work *work;
961 cwork->flags |= IO_WQ_WORK_CANCEL;
964 * First check pending list, if we're lucky we can just remove it
965 * from there. CANCEL_OK means that the work is returned as-new,
966 * no completion will be posted for it.
968 spin_lock_irqsave(&wqe->lock, flags);
969 wq_list_for_each(node, prev, &wqe->work_list) {
970 work = container_of(node, struct io_wq_work, list);
973 wq_node_del(&wqe->work_list, node, prev);
978 spin_unlock_irqrestore(&wqe->lock, flags);
981 work->flags |= IO_WQ_WORK_CANCEL;
983 return IO_WQ_CANCEL_OK;
987 * Now check if a free (going busy) or busy worker has the work
988 * currently running. If we find it there, we'll return CANCEL_RUNNING
989 * as an indication that we attempt to signal cancellation. The
990 * completion will run normally in this case.
993 found = io_wq_for_each_worker(wqe, io_wq_worker_cancel, cwork);
995 return found ? IO_WQ_CANCEL_RUNNING : IO_WQ_CANCEL_NOTFOUND;
998 enum io_wq_cancel io_wq_cancel_work(struct io_wq *wq, struct io_wq_work *cwork)
1000 enum io_wq_cancel ret = IO_WQ_CANCEL_NOTFOUND;
1003 for_each_node(node) {
1004 struct io_wqe *wqe = wq->wqes[node];
1006 ret = io_wqe_cancel_work(wqe, cwork);
1007 if (ret != IO_WQ_CANCEL_NOTFOUND)
1014 struct io_wq_flush_data {
1015 struct io_wq_work work;
1016 struct completion done;
1019 static void io_wq_flush_func(struct io_wq_work **workptr)
1021 struct io_wq_work *work = *workptr;
1022 struct io_wq_flush_data *data;
1024 data = container_of(work, struct io_wq_flush_data, work);
1025 complete(&data->done);
1029 * Doesn't wait for previously queued work to finish. When this completes,
1030 * it just means that previously queued work was started.
1032 void io_wq_flush(struct io_wq *wq)
1034 struct io_wq_flush_data data;
1037 for_each_node(node) {
1038 struct io_wqe *wqe = wq->wqes[node];
1040 init_completion(&data.done);
1041 INIT_IO_WORK(&data.work, io_wq_flush_func);
1042 data.work.flags |= IO_WQ_WORK_INTERNAL;
1043 io_wqe_enqueue(wqe, &data.work);
1044 wait_for_completion(&data.done);
1048 struct io_wq *io_wq_create(unsigned bounded, struct io_wq_data *data)
1050 int ret = -ENOMEM, node;
1053 wq = kzalloc(sizeof(*wq), GFP_KERNEL);
1055 return ERR_PTR(-ENOMEM);
1057 wq->wqes = kcalloc(nr_node_ids, sizeof(struct io_wqe *), GFP_KERNEL);
1060 return ERR_PTR(-ENOMEM);
1063 wq->get_work = data->get_work;
1064 wq->put_work = data->put_work;
1066 /* caller must already hold a reference to this */
1067 wq->user = data->user;
1069 for_each_node(node) {
1072 wqe = kzalloc_node(sizeof(struct io_wqe), GFP_KERNEL, node);
1075 wq->wqes[node] = wqe;
1077 wqe->acct[IO_WQ_ACCT_BOUND].max_workers = bounded;
1078 atomic_set(&wqe->acct[IO_WQ_ACCT_BOUND].nr_running, 0);
1080 wqe->acct[IO_WQ_ACCT_UNBOUND].max_workers =
1081 task_rlimit(current, RLIMIT_NPROC);
1083 atomic_set(&wqe->acct[IO_WQ_ACCT_UNBOUND].nr_running, 0);
1086 spin_lock_init(&wqe->lock);
1087 INIT_WQ_LIST(&wqe->work_list);
1088 INIT_HLIST_NULLS_HEAD(&wqe->free_list, 0);
1089 INIT_LIST_HEAD(&wqe->all_list);
1092 init_completion(&wq->done);
1094 wq->manager = kthread_create(io_wq_manager, wq, "io_wq_manager");
1095 if (!IS_ERR(wq->manager)) {
1096 wake_up_process(wq->manager);
1097 wait_for_completion(&wq->done);
1098 if (test_bit(IO_WQ_BIT_ERROR, &wq->state)) {
1102 refcount_set(&wq->use_refs, 1);
1103 reinit_completion(&wq->done);
1107 ret = PTR_ERR(wq->manager);
1108 complete(&wq->done);
1111 kfree(wq->wqes[node]);
1114 return ERR_PTR(ret);
1117 static bool io_wq_worker_wake(struct io_worker *worker, void *data)
1119 wake_up_process(worker->task);
1123 static void __io_wq_destroy(struct io_wq *wq)
1127 set_bit(IO_WQ_BIT_EXIT, &wq->state);
1129 kthread_stop(wq->manager);
1133 io_wq_for_each_worker(wq->wqes[node], io_wq_worker_wake, NULL);
1136 wait_for_completion(&wq->done);
1139 kfree(wq->wqes[node]);
1144 void io_wq_destroy(struct io_wq *wq)
1146 if (refcount_dec_and_test(&wq->use_refs))
1147 __io_wq_destroy(wq);