1/* FS-Cache worker operation management routines
2 *
3 * Copyright (C) 2008 Red Hat, Inc. All Rights Reserved.
4 * Written by David Howells (dhowells@redhat.com)
5 *
6 * This program is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU General Public License
8 * as published by the Free Software Foundation; either version
9 * 2 of the License, or (at your option) any later version.
10 *
11 * See Documentation/filesystems/caching/operations.txt
12 */
13
14#define FSCACHE_DEBUG_LEVEL OPERATION
15#include <linux/module.h>
16#include <linux/seq_file.h>
17#include <linux/slab.h>
18#include "internal.h"
19
20atomic_t fscache_op_debug_id;
21EXPORT_SYMBOL(fscache_op_debug_id);
22
23/**
24 * fscache_enqueue_operation - Enqueue an operation for processing
25 * @op: The operation to enqueue
26 *
27 * Enqueue an operation for processing by the FS-Cache thread pool.
28 *
29 * This will get its own ref on the object.
30 */
31void fscache_enqueue_operation(struct fscache_operation *op)
32{
33	_enter("{OBJ%x OP%x,%u}",
34	       op->object->debug_id, op->debug_id, atomic_read(&op->usage));
35
36	ASSERT(list_empty(&op->pend_link));
37	ASSERT(op->processor != NULL);
38	ASSERT(fscache_object_is_available(op->object));
39	ASSERTCMP(atomic_read(&op->usage), >, 0);
40	ASSERTCMP(op->state, ==, FSCACHE_OP_ST_IN_PROGRESS);
41
42	fscache_stat(&fscache_n_op_enqueue);
43	switch (op->flags & FSCACHE_OP_TYPE) {
44	case FSCACHE_OP_ASYNC:
45		_debug("queue async");
46		atomic_inc(&op->usage);
47		if (!queue_work(fscache_op_wq, &op->work))
48			fscache_put_operation(op);
49		break;
50	case FSCACHE_OP_MYTHREAD:
51		_debug("queue for caller's attention");
52		break;
53	default:
54		pr_err("Unexpected op type %lx", op->flags);
55		BUG();
56		break;
57	}
58}
59EXPORT_SYMBOL(fscache_enqueue_operation);
60
61/*
62 * start an op running
63 */
64static void fscache_run_op(struct fscache_object *object,
65			   struct fscache_operation *op)
66{
67	ASSERTCMP(op->state, ==, FSCACHE_OP_ST_PENDING);
68
69	op->state = FSCACHE_OP_ST_IN_PROGRESS;
70	object->n_in_progress++;
71	if (test_and_clear_bit(FSCACHE_OP_WAITING, &op->flags))
72		wake_up_bit(&op->flags, FSCACHE_OP_WAITING);
73	if (op->processor)
74		fscache_enqueue_operation(op);
75	fscache_stat(&fscache_n_op_run);
76}
77
78/*
79 * submit an exclusive operation for an object
80 * - other ops are excluded from running simultaneously with this one
81 * - this gets any extra refs it needs on an op
82 */
83int fscache_submit_exclusive_op(struct fscache_object *object,
84				struct fscache_operation *op)
85{
86	int ret;
87
88	_enter("{OBJ%x OP%x},", object->debug_id, op->debug_id);
89
90	ASSERTCMP(op->state, ==, FSCACHE_OP_ST_INITIALISED);
91	ASSERTCMP(atomic_read(&op->usage), >, 0);
92
93	spin_lock(&object->lock);
94	ASSERTCMP(object->n_ops, >=, object->n_in_progress);
95	ASSERTCMP(object->n_ops, >=, object->n_exclusive);
96	ASSERT(list_empty(&op->pend_link));
97
98	op->state = FSCACHE_OP_ST_PENDING;
99	if (fscache_object_is_active(object)) {
100		op->object = object;
101		object->n_ops++;
102		object->n_exclusive++;	/* reads and writes must wait */
103
104		if (object->n_in_progress > 0) {
105			atomic_inc(&op->usage);
106			list_add_tail(&op->pend_link, &object->pending_ops);
107			fscache_stat(&fscache_n_op_pend);
108		} else if (!list_empty(&object->pending_ops)) {
109			atomic_inc(&op->usage);
110			list_add_tail(&op->pend_link, &object->pending_ops);
111			fscache_stat(&fscache_n_op_pend);
112			fscache_start_operations(object);
113		} else {
114			ASSERTCMP(object->n_in_progress, ==, 0);
115			fscache_run_op(object, op);
116		}
117
118		/* need to issue a new write op after this */
119		clear_bit(FSCACHE_OBJECT_PENDING_WRITE, &object->flags);
120		ret = 0;
121	} else if (test_bit(FSCACHE_OBJECT_IS_LOOKED_UP, &object->flags)) {
122		op->object = object;
123		object->n_ops++;
124		object->n_exclusive++;	/* reads and writes must wait */
125		atomic_inc(&op->usage);
126		list_add_tail(&op->pend_link, &object->pending_ops);
127		fscache_stat(&fscache_n_op_pend);
128		ret = 0;
129	} else {
130		/* If we're in any other state, there must have been an I/O
131		 * error of some nature.
132		 */
133		ASSERT(test_bit(FSCACHE_IOERROR, &object->cache->flags));
134		ret = -EIO;
135	}
136
137	spin_unlock(&object->lock);
138	return ret;
139}
140
141/*
142 * report an unexpected submission
143 */
144static void fscache_report_unexpected_submission(struct fscache_object *object,
145						 struct fscache_operation *op,
146						 const struct fscache_state *ostate)
147{
148	static bool once_only;
149	struct fscache_operation *p;
150	unsigned n;
151
152	if (once_only)
153		return;
154	once_only = true;
155
156	kdebug("unexpected submission OP%x [OBJ%x %s]",
157	       op->debug_id, object->debug_id, object->state->name);
158	kdebug("objstate=%s [%s]", object->state->name, ostate->name);
159	kdebug("objflags=%lx", object->flags);
160	kdebug("objevent=%lx [%lx]", object->events, object->event_mask);
161	kdebug("ops=%u inp=%u exc=%u",
162	       object->n_ops, object->n_in_progress, object->n_exclusive);
163
164	if (!list_empty(&object->pending_ops)) {
165		n = 0;
166		list_for_each_entry(p, &object->pending_ops, pend_link) {
167			ASSERTCMP(p->object, ==, object);
168			kdebug("%p %p", op->processor, op->release);
169			n++;
170		}
171
172		kdebug("n=%u", n);
173	}
174
175	dump_stack();
176}
177
178/*
179 * submit an operation for an object
180 * - objects may be submitted only in the following states:
181 *   - during object creation (write ops may be submitted)
182 *   - whilst the object is active
183 *   - after an I/O error incurred in one of the two above states (op rejected)
184 * - this gets any extra refs it needs on an op
185 */
186int fscache_submit_op(struct fscache_object *object,
187		      struct fscache_operation *op)
188{
189	const struct fscache_state *ostate;
190	int ret;
191
192	_enter("{OBJ%x OP%x},{%u}",
193	       object->debug_id, op->debug_id, atomic_read(&op->usage));
194
195	ASSERTCMP(op->state, ==, FSCACHE_OP_ST_INITIALISED);
196	ASSERTCMP(atomic_read(&op->usage), >, 0);
197
198	spin_lock(&object->lock);
199	ASSERTCMP(object->n_ops, >=, object->n_in_progress);
200	ASSERTCMP(object->n_ops, >=, object->n_exclusive);
201	ASSERT(list_empty(&op->pend_link));
202
203	ostate = object->state;
204	smp_rmb();
205
206	op->state = FSCACHE_OP_ST_PENDING;
207	if (fscache_object_is_active(object)) {
208		op->object = object;
209		object->n_ops++;
210
211		if (object->n_exclusive > 0) {
212			atomic_inc(&op->usage);
213			list_add_tail(&op->pend_link, &object->pending_ops);
214			fscache_stat(&fscache_n_op_pend);
215		} else if (!list_empty(&object->pending_ops)) {
216			atomic_inc(&op->usage);
217			list_add_tail(&op->pend_link, &object->pending_ops);
218			fscache_stat(&fscache_n_op_pend);
219			fscache_start_operations(object);
220		} else {
221			ASSERTCMP(object->n_exclusive, ==, 0);
222			fscache_run_op(object, op);
223		}
224		ret = 0;
225	} else if (test_bit(FSCACHE_OBJECT_IS_LOOKED_UP, &object->flags)) {
226		op->object = object;
227		object->n_ops++;
228		atomic_inc(&op->usage);
229		list_add_tail(&op->pend_link, &object->pending_ops);
230		fscache_stat(&fscache_n_op_pend);
231		ret = 0;
232	} else if (fscache_object_is_dying(object)) {
233		fscache_stat(&fscache_n_op_rejected);
234		op->state = FSCACHE_OP_ST_CANCELLED;
235		ret = -ENOBUFS;
236	} else if (!test_bit(FSCACHE_IOERROR, &object->cache->flags)) {
237		fscache_report_unexpected_submission(object, op, ostate);
238		ASSERT(!fscache_object_is_active(object));
239		op->state = FSCACHE_OP_ST_CANCELLED;
240		ret = -ENOBUFS;
241	} else {
242		op->state = FSCACHE_OP_ST_CANCELLED;
243		ret = -ENOBUFS;
244	}
245
246	spin_unlock(&object->lock);
247	return ret;
248}
249
250/*
251 * queue an object for withdrawal on error, aborting all following asynchronous
252 * operations
253 */
254void fscache_abort_object(struct fscache_object *object)
255{
256	_enter("{OBJ%x}", object->debug_id);
257
258	fscache_raise_event(object, FSCACHE_OBJECT_EV_ERROR);
259}
260
261/*
262 * Jump start the operation processing on an object.  The caller must hold
263 * object->lock.
264 */
265void fscache_start_operations(struct fscache_object *object)
266{
267	struct fscache_operation *op;
268	bool stop = false;
269
270	while (!list_empty(&object->pending_ops) && !stop) {
271		op = list_entry(object->pending_ops.next,
272				struct fscache_operation, pend_link);
273
274		if (test_bit(FSCACHE_OP_EXCLUSIVE, &op->flags)) {
275			if (object->n_in_progress > 0)
276				break;
277			stop = true;
278		}
279		list_del_init(&op->pend_link);
280		fscache_run_op(object, op);
281
282		/* the pending queue was holding a ref on the object */
283		fscache_put_operation(op);
284	}
285
286	ASSERTCMP(object->n_in_progress, <=, object->n_ops);
287
288	_debug("woke %d ops on OBJ%x",
289	       object->n_in_progress, object->debug_id);
290}
291
292/*
293 * cancel an operation that's pending on an object
294 */
295int fscache_cancel_op(struct fscache_operation *op,
296		      void (*do_cancel)(struct fscache_operation *))
297{
298	struct fscache_object *object = op->object;
299	int ret;
300
301	_enter("OBJ%x OP%x}", op->object->debug_id, op->debug_id);
302
303	ASSERTCMP(op->state, >=, FSCACHE_OP_ST_PENDING);
304	ASSERTCMP(op->state, !=, FSCACHE_OP_ST_CANCELLED);
305	ASSERTCMP(atomic_read(&op->usage), >, 0);
306
307	spin_lock(&object->lock);
308
309	ret = -EBUSY;
310	if (op->state == FSCACHE_OP_ST_PENDING) {
311		ASSERT(!list_empty(&op->pend_link));
312		fscache_stat(&fscache_n_op_cancelled);
313		list_del_init(&op->pend_link);
314		if (do_cancel)
315			do_cancel(op);
316		op->state = FSCACHE_OP_ST_CANCELLED;
317		if (test_bit(FSCACHE_OP_EXCLUSIVE, &op->flags))
318			object->n_exclusive--;
319		if (test_and_clear_bit(FSCACHE_OP_WAITING, &op->flags))
320			wake_up_bit(&op->flags, FSCACHE_OP_WAITING);
321		fscache_put_operation(op);
322		ret = 0;
323	}
324
325	spin_unlock(&object->lock);
326	_leave(" = %d", ret);
327	return ret;
328}
329
330/*
331 * Cancel all pending operations on an object
332 */
333void fscache_cancel_all_ops(struct fscache_object *object)
334{
335	struct fscache_operation *op;
336
337	_enter("OBJ%x", object->debug_id);
338
339	spin_lock(&object->lock);
340
341	while (!list_empty(&object->pending_ops)) {
342		op = list_entry(object->pending_ops.next,
343				struct fscache_operation, pend_link);
344		fscache_stat(&fscache_n_op_cancelled);
345		list_del_init(&op->pend_link);
346
347		ASSERTCMP(op->state, ==, FSCACHE_OP_ST_PENDING);
348		op->state = FSCACHE_OP_ST_CANCELLED;
349
350		if (test_bit(FSCACHE_OP_EXCLUSIVE, &op->flags))
351			object->n_exclusive--;
352		if (test_and_clear_bit(FSCACHE_OP_WAITING, &op->flags))
353			wake_up_bit(&op->flags, FSCACHE_OP_WAITING);
354		fscache_put_operation(op);
355		cond_resched_lock(&object->lock);
356	}
357
358	spin_unlock(&object->lock);
359	_leave("");
360}
361
362/*
363 * Record the completion or cancellation of an in-progress operation.
364 */
365void fscache_op_complete(struct fscache_operation *op, bool cancelled)
366{
367	struct fscache_object *object = op->object;
368
369	_enter("OBJ%x", object->debug_id);
370
371	ASSERTCMP(op->state, ==, FSCACHE_OP_ST_IN_PROGRESS);
372	ASSERTCMP(object->n_in_progress, >, 0);
373	ASSERTIFCMP(test_bit(FSCACHE_OP_EXCLUSIVE, &op->flags),
374		    object->n_exclusive, >, 0);
375	ASSERTIFCMP(test_bit(FSCACHE_OP_EXCLUSIVE, &op->flags),
376		    object->n_in_progress, ==, 1);
377
378	spin_lock(&object->lock);
379
380	op->state = cancelled ?
381		FSCACHE_OP_ST_CANCELLED : FSCACHE_OP_ST_COMPLETE;
382
383	if (test_bit(FSCACHE_OP_EXCLUSIVE, &op->flags))
384		object->n_exclusive--;
385	object->n_in_progress--;
386	if (object->n_in_progress == 0)
387		fscache_start_operations(object);
388
389	spin_unlock(&object->lock);
390	_leave("");
391}
392EXPORT_SYMBOL(fscache_op_complete);
393
394/*
395 * release an operation
396 * - queues pending ops if this is the last in-progress op
397 */
398void fscache_put_operation(struct fscache_operation *op)
399{
400	struct fscache_object *object;
401	struct fscache_cache *cache;
402
403	_enter("{OBJ%x OP%x,%d}",
404	       op->object->debug_id, op->debug_id, atomic_read(&op->usage));
405
406	ASSERTCMP(atomic_read(&op->usage), >, 0);
407
408	if (!atomic_dec_and_test(&op->usage))
409		return;
410
411	_debug("PUT OP");
412	ASSERTIFCMP(op->state != FSCACHE_OP_ST_COMPLETE,
413		    op->state, ==, FSCACHE_OP_ST_CANCELLED);
414	op->state = FSCACHE_OP_ST_DEAD;
415
416	fscache_stat(&fscache_n_op_release);
417
418	if (op->release) {
419		op->release(op);
420		op->release = NULL;
421	}
422
423	object = op->object;
424
425	if (test_bit(FSCACHE_OP_DEC_READ_CNT, &op->flags))
426		atomic_dec(&object->n_reads);
427	if (test_bit(FSCACHE_OP_UNUSE_COOKIE, &op->flags))
428		fscache_unuse_cookie(object);
429
430	/* now... we may get called with the object spinlock held, so we
431	 * complete the cleanup here only if we can immediately acquire the
432	 * lock, and defer it otherwise */
433	if (!spin_trylock(&object->lock)) {
434		_debug("defer put");
435		fscache_stat(&fscache_n_op_deferred_release);
436
437		cache = object->cache;
438		spin_lock(&cache->op_gc_list_lock);
439		list_add_tail(&op->pend_link, &cache->op_gc_list);
440		spin_unlock(&cache->op_gc_list_lock);
441		schedule_work(&cache->op_gc);
442		_leave(" [defer]");
443		return;
444	}
445
446	ASSERTCMP(object->n_ops, >, 0);
447	object->n_ops--;
448	if (object->n_ops == 0)
449		fscache_raise_event(object, FSCACHE_OBJECT_EV_CLEARED);
450
451	spin_unlock(&object->lock);
452
453	kfree(op);
454	_leave(" [done]");
455}
456EXPORT_SYMBOL(fscache_put_operation);
457
458/*
459 * garbage collect operations that have had their release deferred
460 */
461void fscache_operation_gc(struct work_struct *work)
462{
463	struct fscache_operation *op;
464	struct fscache_object *object;
465	struct fscache_cache *cache =
466		container_of(work, struct fscache_cache, op_gc);
467	int count = 0;
468
469	_enter("");
470
471	do {
472		spin_lock(&cache->op_gc_list_lock);
473		if (list_empty(&cache->op_gc_list)) {
474			spin_unlock(&cache->op_gc_list_lock);
475			break;
476		}
477
478		op = list_entry(cache->op_gc_list.next,
479				struct fscache_operation, pend_link);
480		list_del(&op->pend_link);
481		spin_unlock(&cache->op_gc_list_lock);
482
483		object = op->object;
484		spin_lock(&object->lock);
485
486		_debug("GC DEFERRED REL OBJ%x OP%x",
487		       object->debug_id, op->debug_id);
488		fscache_stat(&fscache_n_op_gc);
489
490		ASSERTCMP(atomic_read(&op->usage), ==, 0);
491		ASSERTCMP(op->state, ==, FSCACHE_OP_ST_DEAD);
492
493		ASSERTCMP(object->n_ops, >, 0);
494		object->n_ops--;
495		if (object->n_ops == 0)
496			fscache_raise_event(object, FSCACHE_OBJECT_EV_CLEARED);
497
498		spin_unlock(&object->lock);
499		kfree(op);
500
501	} while (count++ < 20);
502
503	if (!list_empty(&cache->op_gc_list))
504		schedule_work(&cache->op_gc);
505
506	_leave("");
507}
508
509/*
510 * execute an operation using fs_op_wq to provide processing context -
511 * the caller holds a ref to this object, so we don't need to hold one
512 */
513void fscache_op_work_func(struct work_struct *work)
514{
515	struct fscache_operation *op =
516		container_of(work, struct fscache_operation, work);
517	unsigned long start;
518
519	_enter("{OBJ%x OP%x,%d}",
520	       op->object->debug_id, op->debug_id, atomic_read(&op->usage));
521
522	ASSERT(op->processor != NULL);
523	start = jiffies;
524	op->processor(op);
525	fscache_hist(fscache_ops_histogram, start);
526	fscache_put_operation(op);
527
528	_leave("");
529}
530