1 /*
2  * linux/net/sunrpc/sched.c
3  *
4  * Scheduling for synchronous and asynchronous RPC requests.
5  *
6  * Copyright (C) 1996 Olaf Kirch, <okir@monad.swb.de>
7  *
8  * TCP NFS related read + write fixes
9  * (C) 1999 Dave Airlie, University of Limerick, Ireland <airlied@linux.ie>
10  */
11 
12 #include <linux/module.h>
13 
14 #include <linux/sched.h>
15 #include <linux/interrupt.h>
16 #include <linux/slab.h>
17 #include <linux/mempool.h>
18 #include <linux/smp.h>
19 #include <linux/spinlock.h>
20 #include <linux/mutex.h>
21 
22 #include <linux/sunrpc/clnt.h>
23 
24 #include "sunrpc.h"
25 
26 #ifdef RPC_DEBUG
27 #define RPCDBG_FACILITY		RPCDBG_SCHED
28 #endif
29 
30 /*
31  * RPC slabs and memory pools
32  */
33 #define RPC_BUFFER_MAXSIZE	(2048)
34 #define RPC_BUFFER_POOLSIZE	(8)
35 #define RPC_TASK_POOLSIZE	(8)
36 static struct kmem_cache	*rpc_task_slabp __read_mostly;
37 static struct kmem_cache	*rpc_buffer_slabp __read_mostly;
38 static mempool_t	*rpc_task_mempool __read_mostly;
39 static mempool_t	*rpc_buffer_mempool __read_mostly;
40 
41 static void			rpc_async_schedule(struct work_struct *);
42 static void			 rpc_release_task(struct rpc_task *task);
43 static void __rpc_queue_timer_fn(unsigned long ptr);
44 
45 /*
46  * RPC tasks sit here while waiting for conditions to improve.
47  */
48 static struct rpc_wait_queue delay_queue;
49 
50 /*
51  * rpciod-related stuff
52  */
53 struct workqueue_struct *rpciod_workqueue;
54 
55 /*
56  * Disable the timer for a given RPC task. Should be called with
57  * queue->lock and bh_disabled in order to avoid races within
58  * rpc_run_timer().
59  */
60 static void
__rpc_disable_timer(struct rpc_wait_queue * queue,struct rpc_task * task)61 __rpc_disable_timer(struct rpc_wait_queue *queue, struct rpc_task *task)
62 {
63 	if (task->tk_timeout == 0)
64 		return;
65 	dprintk("RPC: %5u disabling timer\n", task->tk_pid);
66 	task->tk_timeout = 0;
67 	list_del(&task->u.tk_wait.timer_list);
68 	if (list_empty(&queue->timer_list.list))
69 		del_timer(&queue->timer_list.timer);
70 }
71 
72 static void
rpc_set_queue_timer(struct rpc_wait_queue * queue,unsigned long expires)73 rpc_set_queue_timer(struct rpc_wait_queue *queue, unsigned long expires)
74 {
75 	queue->timer_list.expires = expires;
76 	mod_timer(&queue->timer_list.timer, expires);
77 }
78 
79 /*
80  * Set up a timer for the current task.
81  */
82 static void
__rpc_add_timer(struct rpc_wait_queue * queue,struct rpc_task * task)83 __rpc_add_timer(struct rpc_wait_queue *queue, struct rpc_task *task)
84 {
85 	if (!task->tk_timeout)
86 		return;
87 
88 	dprintk("RPC: %5u setting alarm for %lu ms\n",
89 			task->tk_pid, task->tk_timeout * 1000 / HZ);
90 
91 	task->u.tk_wait.expires = jiffies + task->tk_timeout;
92 	if (list_empty(&queue->timer_list.list) || time_before(task->u.tk_wait.expires, queue->timer_list.expires))
93 		rpc_set_queue_timer(queue, task->u.tk_wait.expires);
94 	list_add(&task->u.tk_wait.timer_list, &queue->timer_list.list);
95 }
96 
97 /*
98  * Add new request to a priority queue.
99  */
__rpc_add_wait_queue_priority(struct rpc_wait_queue * queue,struct rpc_task * task)100 static void __rpc_add_wait_queue_priority(struct rpc_wait_queue *queue, struct rpc_task *task)
101 {
102 	struct list_head *q;
103 	struct rpc_task *t;
104 
105 	INIT_LIST_HEAD(&task->u.tk_wait.links);
106 	q = &queue->tasks[task->tk_priority];
107 	if (unlikely(task->tk_priority > queue->maxpriority))
108 		q = &queue->tasks[queue->maxpriority];
109 	list_for_each_entry(t, q, u.tk_wait.list) {
110 		if (t->tk_owner == task->tk_owner) {
111 			list_add_tail(&task->u.tk_wait.list, &t->u.tk_wait.links);
112 			return;
113 		}
114 	}
115 	list_add_tail(&task->u.tk_wait.list, q);
116 }
117 
118 /*
119  * Add new request to wait queue.
120  *
121  * Swapper tasks always get inserted at the head of the queue.
122  * This should avoid many nasty memory deadlocks and hopefully
123  * improve overall performance.
124  * Everyone else gets appended to the queue to ensure proper FIFO behavior.
125  */
__rpc_add_wait_queue(struct rpc_wait_queue * queue,struct rpc_task * task)126 static void __rpc_add_wait_queue(struct rpc_wait_queue *queue, struct rpc_task *task)
127 {
128 	BUG_ON (RPC_IS_QUEUED(task));
129 
130 	if (RPC_IS_PRIORITY(queue))
131 		__rpc_add_wait_queue_priority(queue, task);
132 	else if (RPC_IS_SWAPPER(task))
133 		list_add(&task->u.tk_wait.list, &queue->tasks[0]);
134 	else
135 		list_add_tail(&task->u.tk_wait.list, &queue->tasks[0]);
136 	task->tk_waitqueue = queue;
137 	queue->qlen++;
138 	rpc_set_queued(task);
139 
140 	dprintk("RPC: %5u added to queue %p \"%s\"\n",
141 			task->tk_pid, queue, rpc_qname(queue));
142 }
143 
144 /*
145  * Remove request from a priority queue.
146  */
__rpc_remove_wait_queue_priority(struct rpc_task * task)147 static void __rpc_remove_wait_queue_priority(struct rpc_task *task)
148 {
149 	struct rpc_task *t;
150 
151 	if (!list_empty(&task->u.tk_wait.links)) {
152 		t = list_entry(task->u.tk_wait.links.next, struct rpc_task, u.tk_wait.list);
153 		list_move(&t->u.tk_wait.list, &task->u.tk_wait.list);
154 		list_splice_init(&task->u.tk_wait.links, &t->u.tk_wait.links);
155 	}
156 }
157 
158 /*
159  * Remove request from queue.
160  * Note: must be called with spin lock held.
161  */
__rpc_remove_wait_queue(struct rpc_wait_queue * queue,struct rpc_task * task)162 static void __rpc_remove_wait_queue(struct rpc_wait_queue *queue, struct rpc_task *task)
163 {
164 	__rpc_disable_timer(queue, task);
165 	if (RPC_IS_PRIORITY(queue))
166 		__rpc_remove_wait_queue_priority(task);
167 	list_del(&task->u.tk_wait.list);
168 	queue->qlen--;
169 	dprintk("RPC: %5u removed from queue %p \"%s\"\n",
170 			task->tk_pid, queue, rpc_qname(queue));
171 }
172 
rpc_set_waitqueue_priority(struct rpc_wait_queue * queue,int priority)173 static inline void rpc_set_waitqueue_priority(struct rpc_wait_queue *queue, int priority)
174 {
175 	queue->priority = priority;
176 	queue->count = 1 << (priority * 2);
177 }
178 
rpc_set_waitqueue_owner(struct rpc_wait_queue * queue,pid_t pid)179 static inline void rpc_set_waitqueue_owner(struct rpc_wait_queue *queue, pid_t pid)
180 {
181 	queue->owner = pid;
182 	queue->nr = RPC_BATCH_COUNT;
183 }
184 
rpc_reset_waitqueue_priority(struct rpc_wait_queue * queue)185 static inline void rpc_reset_waitqueue_priority(struct rpc_wait_queue *queue)
186 {
187 	rpc_set_waitqueue_priority(queue, queue->maxpriority);
188 	rpc_set_waitqueue_owner(queue, 0);
189 }
190 
__rpc_init_priority_wait_queue(struct rpc_wait_queue * queue,const char * qname,unsigned char nr_queues)191 static void __rpc_init_priority_wait_queue(struct rpc_wait_queue *queue, const char *qname, unsigned char nr_queues)
192 {
193 	int i;
194 
195 	spin_lock_init(&queue->lock);
196 	for (i = 0; i < ARRAY_SIZE(queue->tasks); i++)
197 		INIT_LIST_HEAD(&queue->tasks[i]);
198 	queue->maxpriority = nr_queues - 1;
199 	rpc_reset_waitqueue_priority(queue);
200 	queue->qlen = 0;
201 	setup_timer(&queue->timer_list.timer, __rpc_queue_timer_fn, (unsigned long)queue);
202 	INIT_LIST_HEAD(&queue->timer_list.list);
203 #ifdef RPC_DEBUG
204 	queue->name = qname;
205 #endif
206 }
207 
rpc_init_priority_wait_queue(struct rpc_wait_queue * queue,const char * qname)208 void rpc_init_priority_wait_queue(struct rpc_wait_queue *queue, const char *qname)
209 {
210 	__rpc_init_priority_wait_queue(queue, qname, RPC_NR_PRIORITY);
211 }
212 EXPORT_SYMBOL_GPL(rpc_init_priority_wait_queue);
213 
rpc_init_wait_queue(struct rpc_wait_queue * queue,const char * qname)214 void rpc_init_wait_queue(struct rpc_wait_queue *queue, const char *qname)
215 {
216 	__rpc_init_priority_wait_queue(queue, qname, 1);
217 }
218 EXPORT_SYMBOL_GPL(rpc_init_wait_queue);
219 
rpc_destroy_wait_queue(struct rpc_wait_queue * queue)220 void rpc_destroy_wait_queue(struct rpc_wait_queue *queue)
221 {
222 	del_timer_sync(&queue->timer_list.timer);
223 }
224 EXPORT_SYMBOL_GPL(rpc_destroy_wait_queue);
225 
rpc_wait_bit_killable(void * word)226 static int rpc_wait_bit_killable(void *word)
227 {
228 	if (fatal_signal_pending(current))
229 		return -ERESTARTSYS;
230 	schedule();
231 	return 0;
232 }
233 
234 #ifdef RPC_DEBUG
rpc_task_set_debuginfo(struct rpc_task * task)235 static void rpc_task_set_debuginfo(struct rpc_task *task)
236 {
237 	static atomic_t rpc_pid;
238 
239 	task->tk_pid = atomic_inc_return(&rpc_pid);
240 }
241 #else
rpc_task_set_debuginfo(struct rpc_task * task)242 static inline void rpc_task_set_debuginfo(struct rpc_task *task)
243 {
244 }
245 #endif
246 
rpc_set_active(struct rpc_task * task)247 static void rpc_set_active(struct rpc_task *task)
248 {
249 	rpc_task_set_debuginfo(task);
250 	set_bit(RPC_TASK_ACTIVE, &task->tk_runstate);
251 }
252 
253 /*
254  * Mark an RPC call as having completed by clearing the 'active' bit
255  * and then waking up all tasks that were sleeping.
256  */
rpc_complete_task(struct rpc_task * task)257 static int rpc_complete_task(struct rpc_task *task)
258 {
259 	void *m = &task->tk_runstate;
260 	wait_queue_head_t *wq = bit_waitqueue(m, RPC_TASK_ACTIVE);
261 	struct wait_bit_key k = __WAIT_BIT_KEY_INITIALIZER(m, RPC_TASK_ACTIVE);
262 	unsigned long flags;
263 	int ret;
264 
265 	spin_lock_irqsave(&wq->lock, flags);
266 	clear_bit(RPC_TASK_ACTIVE, &task->tk_runstate);
267 	ret = atomic_dec_and_test(&task->tk_count);
268 	if (waitqueue_active(wq))
269 		__wake_up_locked_key(wq, TASK_NORMAL, &k);
270 	spin_unlock_irqrestore(&wq->lock, flags);
271 	return ret;
272 }
273 
274 /*
275  * Allow callers to wait for completion of an RPC call
276  *
277  * Note the use of out_of_line_wait_on_bit() rather than wait_on_bit()
278  * to enforce taking of the wq->lock and hence avoid races with
279  * rpc_complete_task().
280  */
__rpc_wait_for_completion_task(struct rpc_task * task,int (* action)(void *))281 int __rpc_wait_for_completion_task(struct rpc_task *task, int (*action)(void *))
282 {
283 	if (action == NULL)
284 		action = rpc_wait_bit_killable;
285 	return out_of_line_wait_on_bit(&task->tk_runstate, RPC_TASK_ACTIVE,
286 			action, TASK_KILLABLE);
287 }
288 EXPORT_SYMBOL_GPL(__rpc_wait_for_completion_task);
289 
290 /*
291  * Make an RPC task runnable.
292  *
293  * Note: If the task is ASYNC, this must be called with
294  * the spinlock held to protect the wait queue operation.
295  */
rpc_make_runnable(struct rpc_task * task)296 static void rpc_make_runnable(struct rpc_task *task)
297 {
298 	rpc_clear_queued(task);
299 	if (rpc_test_and_set_running(task))
300 		return;
301 	if (RPC_IS_ASYNC(task)) {
302 		INIT_WORK(&task->u.tk_work, rpc_async_schedule);
303 		queue_work(rpciod_workqueue, &task->u.tk_work);
304 	} else
305 		wake_up_bit(&task->tk_runstate, RPC_TASK_QUEUED);
306 }
307 
308 /*
309  * Prepare for sleeping on a wait queue.
310  * By always appending tasks to the list we ensure FIFO behavior.
311  * NB: An RPC task will only receive interrupt-driven events as long
312  * as it's on a wait queue.
313  */
__rpc_sleep_on(struct rpc_wait_queue * q,struct rpc_task * task,rpc_action action)314 static void __rpc_sleep_on(struct rpc_wait_queue *q, struct rpc_task *task,
315 			rpc_action action)
316 {
317 	dprintk("RPC: %5u sleep_on(queue \"%s\" time %lu)\n",
318 			task->tk_pid, rpc_qname(q), jiffies);
319 
320 	__rpc_add_wait_queue(q, task);
321 
322 	BUG_ON(task->tk_callback != NULL);
323 	task->tk_callback = action;
324 	__rpc_add_timer(q, task);
325 }
326 
rpc_sleep_on(struct rpc_wait_queue * q,struct rpc_task * task,rpc_action action)327 void rpc_sleep_on(struct rpc_wait_queue *q, struct rpc_task *task,
328 				rpc_action action)
329 {
330 	/* We shouldn't ever put an inactive task to sleep */
331 	BUG_ON(!RPC_IS_ACTIVATED(task));
332 
333 	/*
334 	 * Protect the queue operations.
335 	 */
336 	spin_lock_bh(&q->lock);
337 	__rpc_sleep_on(q, task, action);
338 	spin_unlock_bh(&q->lock);
339 }
340 EXPORT_SYMBOL_GPL(rpc_sleep_on);
341 
342 /**
343  * __rpc_do_wake_up_task - wake up a single rpc_task
344  * @queue: wait queue
345  * @task: task to be woken up
346  *
347  * Caller must hold queue->lock, and have cleared the task queued flag.
348  */
__rpc_do_wake_up_task(struct rpc_wait_queue * queue,struct rpc_task * task)349 static void __rpc_do_wake_up_task(struct rpc_wait_queue *queue, struct rpc_task *task)
350 {
351 	dprintk("RPC: %5u __rpc_wake_up_task (now %lu)\n",
352 			task->tk_pid, jiffies);
353 
354 	/* Has the task been executed yet? If not, we cannot wake it up! */
355 	if (!RPC_IS_ACTIVATED(task)) {
356 		printk(KERN_ERR "RPC: Inactive task (%p) being woken up!\n", task);
357 		return;
358 	}
359 
360 	__rpc_remove_wait_queue(queue, task);
361 
362 	rpc_make_runnable(task);
363 
364 	dprintk("RPC:       __rpc_wake_up_task done\n");
365 }
366 
367 /*
368  * Wake up a queued task while the queue lock is being held
369  */
rpc_wake_up_task_queue_locked(struct rpc_wait_queue * queue,struct rpc_task * task)370 static void rpc_wake_up_task_queue_locked(struct rpc_wait_queue *queue, struct rpc_task *task)
371 {
372 	if (RPC_IS_QUEUED(task) && task->tk_waitqueue == queue)
373 		__rpc_do_wake_up_task(queue, task);
374 }
375 
376 /*
377  * Tests whether rpc queue is empty
378  */
rpc_queue_empty(struct rpc_wait_queue * queue)379 int rpc_queue_empty(struct rpc_wait_queue *queue)
380 {
381 	int res;
382 
383 	spin_lock_bh(&queue->lock);
384 	res = queue->qlen;
385 	spin_unlock_bh(&queue->lock);
386 	return res == 0;
387 }
388 EXPORT_SYMBOL_GPL(rpc_queue_empty);
389 
390 /*
391  * Wake up a task on a specific queue
392  */
rpc_wake_up_queued_task(struct rpc_wait_queue * queue,struct rpc_task * task)393 void rpc_wake_up_queued_task(struct rpc_wait_queue *queue, struct rpc_task *task)
394 {
395 	spin_lock_bh(&queue->lock);
396 	rpc_wake_up_task_queue_locked(queue, task);
397 	spin_unlock_bh(&queue->lock);
398 }
399 EXPORT_SYMBOL_GPL(rpc_wake_up_queued_task);
400 
401 /*
402  * Wake up the next task on a priority queue.
403  */
__rpc_wake_up_next_priority(struct rpc_wait_queue * queue)404 static struct rpc_task * __rpc_wake_up_next_priority(struct rpc_wait_queue *queue)
405 {
406 	struct list_head *q;
407 	struct rpc_task *task;
408 
409 	/*
410 	 * Service a batch of tasks from a single owner.
411 	 */
412 	q = &queue->tasks[queue->priority];
413 	if (!list_empty(q)) {
414 		task = list_entry(q->next, struct rpc_task, u.tk_wait.list);
415 		if (queue->owner == task->tk_owner) {
416 			if (--queue->nr)
417 				goto out;
418 			list_move_tail(&task->u.tk_wait.list, q);
419 		}
420 		/*
421 		 * Check if we need to switch queues.
422 		 */
423 		if (--queue->count)
424 			goto new_owner;
425 	}
426 
427 	/*
428 	 * Service the next queue.
429 	 */
430 	do {
431 		if (q == &queue->tasks[0])
432 			q = &queue->tasks[queue->maxpriority];
433 		else
434 			q = q - 1;
435 		if (!list_empty(q)) {
436 			task = list_entry(q->next, struct rpc_task, u.tk_wait.list);
437 			goto new_queue;
438 		}
439 	} while (q != &queue->tasks[queue->priority]);
440 
441 	rpc_reset_waitqueue_priority(queue);
442 	return NULL;
443 
444 new_queue:
445 	rpc_set_waitqueue_priority(queue, (unsigned int)(q - &queue->tasks[0]));
446 new_owner:
447 	rpc_set_waitqueue_owner(queue, task->tk_owner);
448 out:
449 	rpc_wake_up_task_queue_locked(queue, task);
450 	return task;
451 }
452 
453 /*
454  * Wake up the next task on the wait queue.
455  */
rpc_wake_up_next(struct rpc_wait_queue * queue)456 struct rpc_task * rpc_wake_up_next(struct rpc_wait_queue *queue)
457 {
458 	struct rpc_task	*task = NULL;
459 
460 	dprintk("RPC:       wake_up_next(%p \"%s\")\n",
461 			queue, rpc_qname(queue));
462 	spin_lock_bh(&queue->lock);
463 	if (RPC_IS_PRIORITY(queue))
464 		task = __rpc_wake_up_next_priority(queue);
465 	else {
466 		task_for_first(task, &queue->tasks[0])
467 			rpc_wake_up_task_queue_locked(queue, task);
468 	}
469 	spin_unlock_bh(&queue->lock);
470 
471 	return task;
472 }
473 EXPORT_SYMBOL_GPL(rpc_wake_up_next);
474 
475 /**
476  * rpc_wake_up - wake up all rpc_tasks
477  * @queue: rpc_wait_queue on which the tasks are sleeping
478  *
479  * Grabs queue->lock
480  */
rpc_wake_up(struct rpc_wait_queue * queue)481 void rpc_wake_up(struct rpc_wait_queue *queue)
482 {
483 	struct rpc_task *task, *next;
484 	struct list_head *head;
485 
486 	spin_lock_bh(&queue->lock);
487 	head = &queue->tasks[queue->maxpriority];
488 	for (;;) {
489 		list_for_each_entry_safe(task, next, head, u.tk_wait.list)
490 			rpc_wake_up_task_queue_locked(queue, task);
491 		if (head == &queue->tasks[0])
492 			break;
493 		head--;
494 	}
495 	spin_unlock_bh(&queue->lock);
496 }
497 EXPORT_SYMBOL_GPL(rpc_wake_up);
498 
499 /**
500  * rpc_wake_up_status - wake up all rpc_tasks and set their status value.
501  * @queue: rpc_wait_queue on which the tasks are sleeping
502  * @status: status value to set
503  *
504  * Grabs queue->lock
505  */
rpc_wake_up_status(struct rpc_wait_queue * queue,int status)506 void rpc_wake_up_status(struct rpc_wait_queue *queue, int status)
507 {
508 	struct rpc_task *task, *next;
509 	struct list_head *head;
510 
511 	spin_lock_bh(&queue->lock);
512 	head = &queue->tasks[queue->maxpriority];
513 	for (;;) {
514 		list_for_each_entry_safe(task, next, head, u.tk_wait.list) {
515 			task->tk_status = status;
516 			rpc_wake_up_task_queue_locked(queue, task);
517 		}
518 		if (head == &queue->tasks[0])
519 			break;
520 		head--;
521 	}
522 	spin_unlock_bh(&queue->lock);
523 }
524 EXPORT_SYMBOL_GPL(rpc_wake_up_status);
525 
__rpc_queue_timer_fn(unsigned long ptr)526 static void __rpc_queue_timer_fn(unsigned long ptr)
527 {
528 	struct rpc_wait_queue *queue = (struct rpc_wait_queue *)ptr;
529 	struct rpc_task *task, *n;
530 	unsigned long expires, now, timeo;
531 
532 	spin_lock(&queue->lock);
533 	expires = now = jiffies;
534 	list_for_each_entry_safe(task, n, &queue->timer_list.list, u.tk_wait.timer_list) {
535 		timeo = task->u.tk_wait.expires;
536 		if (time_after_eq(now, timeo)) {
537 			dprintk("RPC: %5u timeout\n", task->tk_pid);
538 			task->tk_status = -ETIMEDOUT;
539 			rpc_wake_up_task_queue_locked(queue, task);
540 			continue;
541 		}
542 		if (expires == now || time_after(expires, timeo))
543 			expires = timeo;
544 	}
545 	if (!list_empty(&queue->timer_list.list))
546 		rpc_set_queue_timer(queue, expires);
547 	spin_unlock(&queue->lock);
548 }
549 
__rpc_atrun(struct rpc_task * task)550 static void __rpc_atrun(struct rpc_task *task)
551 {
552 	task->tk_status = 0;
553 }
554 
555 /*
556  * Run a task at a later time
557  */
rpc_delay(struct rpc_task * task,unsigned long delay)558 void rpc_delay(struct rpc_task *task, unsigned long delay)
559 {
560 	task->tk_timeout = delay;
561 	rpc_sleep_on(&delay_queue, task, __rpc_atrun);
562 }
563 EXPORT_SYMBOL_GPL(rpc_delay);
564 
565 /*
566  * Helper to call task->tk_ops->rpc_call_prepare
567  */
rpc_prepare_task(struct rpc_task * task)568 void rpc_prepare_task(struct rpc_task *task)
569 {
570 	task->tk_ops->rpc_call_prepare(task, task->tk_calldata);
571 }
572 
573 /*
574  * Helper that calls task->tk_ops->rpc_call_done if it exists
575  */
rpc_exit_task(struct rpc_task * task)576 void rpc_exit_task(struct rpc_task *task)
577 {
578 	task->tk_action = NULL;
579 	if (task->tk_ops->rpc_call_done != NULL) {
580 		task->tk_ops->rpc_call_done(task, task->tk_calldata);
581 		if (task->tk_action != NULL) {
582 			WARN_ON(RPC_ASSASSINATED(task));
583 			/* Always release the RPC slot and buffer memory */
584 			xprt_release(task);
585 		}
586 	}
587 }
588 
rpc_exit(struct rpc_task * task,int status)589 void rpc_exit(struct rpc_task *task, int status)
590 {
591 	task->tk_status = status;
592 	task->tk_action = rpc_exit_task;
593 	if (RPC_IS_QUEUED(task))
594 		rpc_wake_up_queued_task(task->tk_waitqueue, task);
595 }
596 EXPORT_SYMBOL_GPL(rpc_exit);
597 
rpc_release_calldata(const struct rpc_call_ops * ops,void * calldata)598 void rpc_release_calldata(const struct rpc_call_ops *ops, void *calldata)
599 {
600 	if (ops->rpc_release != NULL)
601 		ops->rpc_release(calldata);
602 }
603 
604 /*
605  * This is the RPC `scheduler' (or rather, the finite state machine).
606  */
__rpc_execute(struct rpc_task * task)607 static void __rpc_execute(struct rpc_task *task)
608 {
609 	struct rpc_wait_queue *queue;
610 	int task_is_async = RPC_IS_ASYNC(task);
611 	int status = 0;
612 
613 	dprintk("RPC: %5u __rpc_execute flags=0x%x\n",
614 			task->tk_pid, task->tk_flags);
615 
616 	BUG_ON(RPC_IS_QUEUED(task));
617 
618 	for (;;) {
619 
620 		/*
621 		 * Execute any pending callback.
622 		 */
623 		if (task->tk_callback) {
624 			void (*save_callback)(struct rpc_task *);
625 
626 			/*
627 			 * We set tk_callback to NULL before calling it,
628 			 * in case it sets the tk_callback field itself:
629 			 */
630 			save_callback = task->tk_callback;
631 			task->tk_callback = NULL;
632 			save_callback(task);
633 		} else {
634 			/*
635 			 * Perform the next FSM step.
636 			 * tk_action may be NULL when the task has been killed
637 			 * by someone else.
638 			 */
639 			if (task->tk_action == NULL)
640 				break;
641 			task->tk_action(task);
642 		}
643 
644 		/*
645 		 * Lockless check for whether task is sleeping or not.
646 		 */
647 		if (!RPC_IS_QUEUED(task))
648 			continue;
649 		/*
650 		 * The queue->lock protects against races with
651 		 * rpc_make_runnable().
652 		 *
653 		 * Note that once we clear RPC_TASK_RUNNING on an asynchronous
654 		 * rpc_task, rpc_make_runnable() can assign it to a
655 		 * different workqueue. We therefore cannot assume that the
656 		 * rpc_task pointer may still be dereferenced.
657 		 */
658 		queue = task->tk_waitqueue;
659 		spin_lock_bh(&queue->lock);
660 		if (!RPC_IS_QUEUED(task)) {
661 			spin_unlock_bh(&queue->lock);
662 			continue;
663 		}
664 		rpc_clear_running(task);
665 		spin_unlock_bh(&queue->lock);
666 		if (task_is_async)
667 			return;
668 
669 		/* sync task: sleep here */
670 		dprintk("RPC: %5u sync task going to sleep\n", task->tk_pid);
671 		status = out_of_line_wait_on_bit(&task->tk_runstate,
672 				RPC_TASK_QUEUED, rpc_wait_bit_killable,
673 				TASK_KILLABLE);
674 		if (status == -ERESTARTSYS) {
675 			/*
676 			 * When a sync task receives a signal, it exits with
677 			 * -ERESTARTSYS. In order to catch any callbacks that
678 			 * clean up after sleeping on some queue, we don't
679 			 * break the loop here, but go around once more.
680 			 */
681 			dprintk("RPC: %5u got signal\n", task->tk_pid);
682 			task->tk_flags |= RPC_TASK_KILLED;
683 			rpc_exit(task, -ERESTARTSYS);
684 		}
685 		rpc_set_running(task);
686 		dprintk("RPC: %5u sync task resuming\n", task->tk_pid);
687 	}
688 
689 	dprintk("RPC: %5u return %d, status %d\n", task->tk_pid, status,
690 			task->tk_status);
691 	/* Release all resources associated with the task */
692 	rpc_release_task(task);
693 }
694 
695 /*
696  * User-visible entry point to the scheduler.
697  *
698  * This may be called recursively if e.g. an async NFS task updates
699  * the attributes and finds that dirty pages must be flushed.
700  * NOTE: Upon exit of this function the task is guaranteed to be
701  *	 released. In particular note that tk_release() will have
702  *	 been called, so your task memory may have been freed.
703  */
rpc_execute(struct rpc_task * task)704 void rpc_execute(struct rpc_task *task)
705 {
706 	rpc_set_active(task);
707 	rpc_make_runnable(task);
708 	if (!RPC_IS_ASYNC(task))
709 		__rpc_execute(task);
710 }
711 
rpc_async_schedule(struct work_struct * work)712 static void rpc_async_schedule(struct work_struct *work)
713 {
714 	__rpc_execute(container_of(work, struct rpc_task, u.tk_work));
715 }
716 
717 /**
718  * rpc_malloc - allocate an RPC buffer
719  * @task: RPC task that will use this buffer
720  * @size: requested byte size
721  *
722  * To prevent rpciod from hanging, this allocator never sleeps,
723  * returning NULL if the request cannot be serviced immediately.
724  * The caller can arrange to sleep in a way that is safe for rpciod.
725  *
726  * Most requests are 'small' (under 2KiB) and can be serviced from a
727  * mempool, ensuring that NFS reads and writes can always proceed,
728  * and that there is good locality of reference for these buffers.
729  *
730  * In order to avoid memory starvation triggering more writebacks of
731  * NFS requests, we avoid using GFP_KERNEL.
732  */
rpc_malloc(struct rpc_task * task,size_t size)733 void *rpc_malloc(struct rpc_task *task, size_t size)
734 {
735 	struct rpc_buffer *buf;
736 	gfp_t gfp = RPC_IS_SWAPPER(task) ? GFP_ATOMIC : GFP_NOWAIT;
737 
738 	size += sizeof(struct rpc_buffer);
739 	if (size <= RPC_BUFFER_MAXSIZE)
740 		buf = mempool_alloc(rpc_buffer_mempool, gfp);
741 	else
742 		buf = kmalloc(size, gfp);
743 
744 	if (!buf)
745 		return NULL;
746 
747 	buf->len = size;
748 	dprintk("RPC: %5u allocated buffer of size %zu at %p\n",
749 			task->tk_pid, size, buf);
750 	return &buf->data;
751 }
752 EXPORT_SYMBOL_GPL(rpc_malloc);
753 
754 /**
755  * rpc_free - free buffer allocated via rpc_malloc
756  * @buffer: buffer to free
757  *
758  */
rpc_free(void * buffer)759 void rpc_free(void *buffer)
760 {
761 	size_t size;
762 	struct rpc_buffer *buf;
763 
764 	if (!buffer)
765 		return;
766 
767 	buf = container_of(buffer, struct rpc_buffer, data);
768 	size = buf->len;
769 
770 	dprintk("RPC:       freeing buffer of size %zu at %p\n",
771 			size, buf);
772 
773 	if (size <= RPC_BUFFER_MAXSIZE)
774 		mempool_free(buf, rpc_buffer_mempool);
775 	else
776 		kfree(buf);
777 }
778 EXPORT_SYMBOL_GPL(rpc_free);
779 
780 /*
781  * Creation and deletion of RPC task structures
782  */
rpc_init_task(struct rpc_task * task,const struct rpc_task_setup * task_setup_data)783 static void rpc_init_task(struct rpc_task *task, const struct rpc_task_setup *task_setup_data)
784 {
785 	memset(task, 0, sizeof(*task));
786 	atomic_set(&task->tk_count, 1);
787 	task->tk_flags  = task_setup_data->flags;
788 	task->tk_ops = task_setup_data->callback_ops;
789 	task->tk_calldata = task_setup_data->callback_data;
790 	INIT_LIST_HEAD(&task->tk_task);
791 
792 	/* Initialize retry counters */
793 	task->tk_garb_retry = 2;
794 	task->tk_cred_retry = 2;
795 
796 	task->tk_priority = task_setup_data->priority - RPC_PRIORITY_LOW;
797 	task->tk_owner = current->tgid;
798 
799 	/* Initialize workqueue for async tasks */
800 	task->tk_workqueue = task_setup_data->workqueue;
801 
802 	if (task->tk_ops->rpc_call_prepare != NULL)
803 		task->tk_action = rpc_prepare_task;
804 
805 	/* starting timestamp */
806 	task->tk_start = ktime_get();
807 
808 	dprintk("RPC:       new task initialized, procpid %u\n",
809 				task_pid_nr(current));
810 }
811 
812 static struct rpc_task *
rpc_alloc_task(void)813 rpc_alloc_task(void)
814 {
815 	return (struct rpc_task *)mempool_alloc(rpc_task_mempool, GFP_NOFS);
816 }
817 
818 /*
819  * Create a new task for the specified client.
820  */
rpc_new_task(const struct rpc_task_setup * setup_data)821 struct rpc_task *rpc_new_task(const struct rpc_task_setup *setup_data)
822 {
823 	struct rpc_task	*task = setup_data->task;
824 	unsigned short flags = 0;
825 
826 	if (task == NULL) {
827 		task = rpc_alloc_task();
828 		if (task == NULL) {
829 			rpc_release_calldata(setup_data->callback_ops,
830 					setup_data->callback_data);
831 			return ERR_PTR(-ENOMEM);
832 		}
833 		flags = RPC_TASK_DYNAMIC;
834 	}
835 
836 	rpc_init_task(task, setup_data);
837 	task->tk_flags |= flags;
838 	dprintk("RPC:       allocated task %p\n", task);
839 	return task;
840 }
841 
rpc_free_task(struct rpc_task * task)842 static void rpc_free_task(struct rpc_task *task)
843 {
844 	const struct rpc_call_ops *tk_ops = task->tk_ops;
845 	void *calldata = task->tk_calldata;
846 
847 	if (task->tk_flags & RPC_TASK_DYNAMIC) {
848 		dprintk("RPC: %5u freeing task\n", task->tk_pid);
849 		mempool_free(task, rpc_task_mempool);
850 	}
851 	rpc_release_calldata(tk_ops, calldata);
852 }
853 
rpc_async_release(struct work_struct * work)854 static void rpc_async_release(struct work_struct *work)
855 {
856 	rpc_free_task(container_of(work, struct rpc_task, u.tk_work));
857 }
858 
rpc_release_resources_task(struct rpc_task * task)859 static void rpc_release_resources_task(struct rpc_task *task)
860 {
861 	if (task->tk_rqstp)
862 		xprt_release(task);
863 	if (task->tk_msg.rpc_cred) {
864 		put_rpccred(task->tk_msg.rpc_cred);
865 		task->tk_msg.rpc_cred = NULL;
866 	}
867 	rpc_task_release_client(task);
868 }
869 
rpc_final_put_task(struct rpc_task * task,struct workqueue_struct * q)870 static void rpc_final_put_task(struct rpc_task *task,
871 		struct workqueue_struct *q)
872 {
873 	if (q != NULL) {
874 		INIT_WORK(&task->u.tk_work, rpc_async_release);
875 		queue_work(q, &task->u.tk_work);
876 	} else
877 		rpc_free_task(task);
878 }
879 
rpc_do_put_task(struct rpc_task * task,struct workqueue_struct * q)880 static void rpc_do_put_task(struct rpc_task *task, struct workqueue_struct *q)
881 {
882 	if (atomic_dec_and_test(&task->tk_count)) {
883 		rpc_release_resources_task(task);
884 		rpc_final_put_task(task, q);
885 	}
886 }
887 
rpc_put_task(struct rpc_task * task)888 void rpc_put_task(struct rpc_task *task)
889 {
890 	rpc_do_put_task(task, NULL);
891 }
892 EXPORT_SYMBOL_GPL(rpc_put_task);
893 
rpc_put_task_async(struct rpc_task * task)894 void rpc_put_task_async(struct rpc_task *task)
895 {
896 	rpc_do_put_task(task, task->tk_workqueue);
897 }
898 EXPORT_SYMBOL_GPL(rpc_put_task_async);
899 
rpc_release_task(struct rpc_task * task)900 static void rpc_release_task(struct rpc_task *task)
901 {
902 	dprintk("RPC: %5u release task\n", task->tk_pid);
903 
904 	BUG_ON (RPC_IS_QUEUED(task));
905 
906 	rpc_release_resources_task(task);
907 
908 	/*
909 	 * Note: at this point we have been removed from rpc_clnt->cl_tasks,
910 	 * so it should be safe to use task->tk_count as a test for whether
911 	 * or not any other processes still hold references to our rpc_task.
912 	 */
913 	if (atomic_read(&task->tk_count) != 1 + !RPC_IS_ASYNC(task)) {
914 		/* Wake up anyone who may be waiting for task completion */
915 		if (!rpc_complete_task(task))
916 			return;
917 	} else {
918 		if (!atomic_dec_and_test(&task->tk_count))
919 			return;
920 	}
921 	rpc_final_put_task(task, task->tk_workqueue);
922 }
923 
rpciod_up(void)924 int rpciod_up(void)
925 {
926 	return try_module_get(THIS_MODULE) ? 0 : -EINVAL;
927 }
928 
rpciod_down(void)929 void rpciod_down(void)
930 {
931 	module_put(THIS_MODULE);
932 }
933 
934 /*
935  * Start up the rpciod workqueue.
936  */
rpciod_start(void)937 static int rpciod_start(void)
938 {
939 	struct workqueue_struct *wq;
940 
941 	/*
942 	 * Create the rpciod thread and wait for it to start.
943 	 */
944 	dprintk("RPC:       creating workqueue rpciod\n");
945 	wq = alloc_workqueue("rpciod", WQ_MEM_RECLAIM, 0);
946 	rpciod_workqueue = wq;
947 	return rpciod_workqueue != NULL;
948 }
949 
rpciod_stop(void)950 static void rpciod_stop(void)
951 {
952 	struct workqueue_struct *wq = NULL;
953 
954 	if (rpciod_workqueue == NULL)
955 		return;
956 	dprintk("RPC:       destroying workqueue rpciod\n");
957 
958 	wq = rpciod_workqueue;
959 	rpciod_workqueue = NULL;
960 	destroy_workqueue(wq);
961 }
962 
963 void
rpc_destroy_mempool(void)964 rpc_destroy_mempool(void)
965 {
966 	rpciod_stop();
967 	if (rpc_buffer_mempool)
968 		mempool_destroy(rpc_buffer_mempool);
969 	if (rpc_task_mempool)
970 		mempool_destroy(rpc_task_mempool);
971 	if (rpc_task_slabp)
972 		kmem_cache_destroy(rpc_task_slabp);
973 	if (rpc_buffer_slabp)
974 		kmem_cache_destroy(rpc_buffer_slabp);
975 	rpc_destroy_wait_queue(&delay_queue);
976 }
977 
978 int
rpc_init_mempool(void)979 rpc_init_mempool(void)
980 {
981 	/*
982 	 * The following is not strictly a mempool initialisation,
983 	 * but there is no harm in doing it here
984 	 */
985 	rpc_init_wait_queue(&delay_queue, "delayq");
986 	if (!rpciod_start())
987 		goto err_nomem;
988 
989 	rpc_task_slabp = kmem_cache_create("rpc_tasks",
990 					     sizeof(struct rpc_task),
991 					     0, SLAB_HWCACHE_ALIGN,
992 					     NULL);
993 	if (!rpc_task_slabp)
994 		goto err_nomem;
995 	rpc_buffer_slabp = kmem_cache_create("rpc_buffers",
996 					     RPC_BUFFER_MAXSIZE,
997 					     0, SLAB_HWCACHE_ALIGN,
998 					     NULL);
999 	if (!rpc_buffer_slabp)
1000 		goto err_nomem;
1001 	rpc_task_mempool = mempool_create_slab_pool(RPC_TASK_POOLSIZE,
1002 						    rpc_task_slabp);
1003 	if (!rpc_task_mempool)
1004 		goto err_nomem;
1005 	rpc_buffer_mempool = mempool_create_slab_pool(RPC_BUFFER_POOLSIZE,
1006 						      rpc_buffer_slabp);
1007 	if (!rpc_buffer_mempool)
1008 		goto err_nomem;
1009 	return 0;
1010 err_nomem:
1011 	rpc_destroy_mempool();
1012 	return -ENOMEM;
1013 }
1014