aboutsummaryrefslogtreecommitdiffstats
path: root/fs/fscache
diff options
context:
space:
mode:
authorDavid Howells <dhowells@redhat.com>2009-04-03 11:42:39 -0400
committerDavid Howells <dhowells@redhat.com>2009-04-03 11:42:39 -0400
commit952efe7b7840e1c726ae88222245e4efe6bd88f3 (patch)
treefed43da6709707d0eb51b8695b6b382a221e99bd /fs/fscache
parentccc4fc3d11e91477036d1f82bfa2d442f6ce77f0 (diff)
FS-Cache: Add and document asynchronous operation handling
Add and document asynchronous operation handling for use by FS-Cache's data storage and retrieval routines. The following documentation is added to: Documentation/filesystems/caching/operations.txt ================================ ASYNCHRONOUS OPERATIONS HANDLING ================================ ======== OVERVIEW ======== FS-Cache has an asynchronous operations handling facility that it uses for its data storage and retrieval routines. Its operations are represented by fscache_operation structs, though these are usually embedded into some other structure. This facility is available to and expected to be be used by the cache backends, and FS-Cache will create operations and pass them off to the appropriate cache backend for completion. To make use of this facility, <linux/fscache-cache.h> should be #included. =============================== OPERATION RECORD INITIALISATION =============================== An operation is recorded in an fscache_operation struct: struct fscache_operation { union { struct work_struct fast_work; struct slow_work slow_work; }; unsigned long flags; fscache_operation_processor_t processor; ... }; Someone wanting to issue an operation should allocate something with this struct embedded in it. They should initialise it by calling: void fscache_operation_init(struct fscache_operation *op, fscache_operation_release_t release); with the operation to be initialised and the release function to use. The op->flags parameter should be set to indicate the CPU time provision and the exclusivity (see the Parameters section). The op->fast_work, op->slow_work and op->processor flags should be set as appropriate for the CPU time provision (see the Parameters section). FSCACHE_OP_WAITING may be set in op->flags prior to each submission of the operation and waited for afterwards. ========== PARAMETERS ========== There are a number of parameters that can be set in the operation record's flag parameter. There are three options for the provision of CPU time in these operations: (1) The operation may be done synchronously (FSCACHE_OP_MYTHREAD). A thread may decide it wants to handle an operation itself without deferring it to another thread. This is, for example, used in read operations for calling readpages() on the backing filesystem in CacheFiles. Although readpages() does an asynchronous data fetch, the determination of whether pages exist is done synchronously - and the netfs does not proceed until this has been determined. If this option is to be used, FSCACHE_OP_WAITING must be set in op->flags before submitting the operation, and the operating thread must wait for it to be cleared before proceeding: wait_on_bit(&op->flags, FSCACHE_OP_WAITING, fscache_wait_bit, TASK_UNINTERRUPTIBLE); (2) The operation may be fast asynchronous (FSCACHE_OP_FAST), in which case it will be given to keventd to process. Such an operation is not permitted to sleep on I/O. This is, for example, used by CacheFiles to copy data from a backing fs page to a netfs page after the backing fs has read the page in. If this option is used, op->fast_work and op->processor must be initialised before submitting the operation: INIT_WORK(&op->fast_work, do_some_work); (3) The operation may be slow asynchronous (FSCACHE_OP_SLOW), in which case it will be given to the slow work facility to process. Such an operation is permitted to sleep on I/O. This is, for example, used by FS-Cache to handle background writes of pages that have just been fetched from a remote server. If this option is used, op->slow_work and op->processor must be initialised before submitting the operation: fscache_operation_init_slow(op, processor) Furthermore, operations may be one of two types: (1) Exclusive (FSCACHE_OP_EXCLUSIVE). Operations of this type may not run in conjunction with any other operation on the object being operated upon. An example of this is the attribute change operation, in which the file being written to may need truncation. (2) Shareable. Operations of this type may be running simultaneously. It's up to the operation implementation to prevent interference between other operations running at the same time. ========= PROCEDURE ========= Operations are used through the following procedure: (1) The submitting thread must allocate the operation and initialise it itself. Normally this would be part of a more specific structure with the generic op embedded within. (2) The submitting thread must then submit the operation for processing using one of the following two functions: int fscache_submit_op(struct fscache_object *object, struct fscache_operation *op); int fscache_submit_exclusive_op(struct fscache_object *object, struct fscache_operation *op); The first function should be used to submit non-exclusive ops and the second to submit exclusive ones. The caller must still set the FSCACHE_OP_EXCLUSIVE flag. If successful, both functions will assign the operation to the specified object and return 0. -ENOBUFS will be returned if the object specified is permanently unavailable. The operation manager will defer operations on an object that is still undergoing lookup or creation. The operation will also be deferred if an operation of conflicting exclusivity is in progress on the object. If the operation is asynchronous, the manager will retain a reference to it, so the caller should put their reference to it by passing it to: void fscache_put_operation(struct fscache_operation *op); (3) If the submitting thread wants to do the work itself, and has marked the operation with FSCACHE_OP_MYTHREAD, then it should monitor FSCACHE_OP_WAITING as described above and check the state of the object if necessary (the object might have died whilst the thread was waiting). When it has finished doing its processing, it should call fscache_put_operation() on it. (4) The operation holds an effective lock upon the object, preventing other exclusive ops conflicting until it is released. The operation can be enqueued for further immediate asynchronous processing by adjusting the CPU time provisioning option if necessary, eg: op->flags &= ~FSCACHE_OP_TYPE; op->flags |= ~FSCACHE_OP_FAST; and calling: void fscache_enqueue_operation(struct fscache_operation *op) This can be used to allow other things to have use of the worker thread pools. ===================== ASYNCHRONOUS CALLBACK ===================== When used in asynchronous mode, the worker thread pool will invoke the processor method with a pointer to the operation. This should then get at the container struct by using container_of(): static void fscache_write_op(struct fscache_operation *_op) { struct fscache_storage *op = container_of(_op, struct fscache_storage, op); ... } The caller holds a reference on the operation, and will invoke fscache_put_operation() when the processor function returns. The processor function is at liberty to call fscache_enqueue_operation() or to take extra references. Signed-off-by: David Howells <dhowells@redhat.com> Acked-by: Steve Dickson <steved@redhat.com> Acked-by: Trond Myklebust <Trond.Myklebust@netapp.com> Acked-by: Al Viro <viro@zeniv.linux.org.uk> Tested-by: Daire Byrne <Daire.Byrne@framestore.com>
Diffstat (limited to 'fs/fscache')
-rw-r--r--fs/fscache/Makefile3
-rw-r--r--fs/fscache/cache.c2
-rw-r--r--fs/fscache/internal.h8
-rw-r--r--fs/fscache/operation.c459
4 files changed, 469 insertions, 3 deletions
diff --git a/fs/fscache/Makefile b/fs/fscache/Makefile
index 4420ac6ea10d..6f82da2aa9d1 100644
--- a/fs/fscache/Makefile
+++ b/fs/fscache/Makefile
@@ -8,7 +8,8 @@ fscache-y := \
8 fsdef.o \ 8 fsdef.o \
9 main.o \ 9 main.o \
10 netfs.o \ 10 netfs.o \
11 object.o 11 object.o \
12 operation.o
12 13
13fscache-$(CONFIG_PROC_FS) += proc.o 14fscache-$(CONFIG_PROC_FS) += proc.o
14fscache-$(CONFIG_FSCACHE_STATS) += stats.o 15fscache-$(CONFIG_FSCACHE_STATS) += stats.o
diff --git a/fs/fscache/cache.c b/fs/fscache/cache.c
index 355172f785fb..e21985bbb1fb 100644
--- a/fs/fscache/cache.c
+++ b/fs/fscache/cache.c
@@ -194,7 +194,7 @@ void fscache_init_cache(struct fscache_cache *cache,
194 vsnprintf(cache->identifier, sizeof(cache->identifier), idfmt, va); 194 vsnprintf(cache->identifier, sizeof(cache->identifier), idfmt, va);
195 va_end(va); 195 va_end(va);
196 196
197 INIT_WORK(&cache->op_gc, NULL); 197 INIT_WORK(&cache->op_gc, fscache_operation_gc);
198 INIT_LIST_HEAD(&cache->link); 198 INIT_LIST_HEAD(&cache->link);
199 INIT_LIST_HEAD(&cache->object_list); 199 INIT_LIST_HEAD(&cache->object_list);
200 INIT_LIST_HEAD(&cache->op_gc_list); 200 INIT_LIST_HEAD(&cache->op_gc_list);
diff --git a/fs/fscache/internal.h b/fs/fscache/internal.h
index 529f4de328c2..014a830c8b37 100644
--- a/fs/fscache/internal.h
+++ b/fs/fscache/internal.h
@@ -95,7 +95,13 @@ extern void fscache_enqueue_object(struct fscache_object *);
95/* 95/*
96 * fsc-operation.c 96 * fsc-operation.c
97 */ 97 */
98#define fscache_start_operations(obj) BUG() 98extern int fscache_submit_exclusive_op(struct fscache_object *,
99 struct fscache_operation *);
100extern int fscache_submit_op(struct fscache_object *,
101 struct fscache_operation *);
102extern void fscache_abort_object(struct fscache_object *);
103extern void fscache_start_operations(struct fscache_object *);
104extern void fscache_operation_gc(struct work_struct *);
99 105
100/* 106/*
101 * fsc-proc.c 107 * fsc-proc.c
diff --git a/fs/fscache/operation.c b/fs/fscache/operation.c
new file mode 100644
index 000000000000..e7f8d53b8b6b
--- /dev/null
+++ b/fs/fscache/operation.c
@@ -0,0 +1,459 @@
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 "internal.h"
17
18atomic_t fscache_op_debug_id;
19EXPORT_SYMBOL(fscache_op_debug_id);
20
21/**
22 * fscache_enqueue_operation - Enqueue an operation for processing
23 * @op: The operation to enqueue
24 *
25 * Enqueue an operation for processing by the FS-Cache thread pool.
26 *
27 * This will get its own ref on the object.
28 */
29void fscache_enqueue_operation(struct fscache_operation *op)
30{
31 _enter("{OBJ%x OP%x,%u}",
32 op->object->debug_id, op->debug_id, atomic_read(&op->usage));
33
34 ASSERT(op->processor != NULL);
35 ASSERTCMP(op->object->state, >=, FSCACHE_OBJECT_AVAILABLE);
36 ASSERTCMP(atomic_read(&op->usage), >, 0);
37
38 if (list_empty(&op->pend_link)) {
39 switch (op->flags & FSCACHE_OP_TYPE) {
40 case FSCACHE_OP_FAST:
41 _debug("queue fast");
42 atomic_inc(&op->usage);
43 if (!schedule_work(&op->fast_work))
44 fscache_put_operation(op);
45 break;
46 case FSCACHE_OP_SLOW:
47 _debug("queue slow");
48 slow_work_enqueue(&op->slow_work);
49 break;
50 case FSCACHE_OP_MYTHREAD:
51 _debug("queue for caller's attention");
52 break;
53 default:
54 printk(KERN_ERR "FS-Cache: Unexpected op type %lx",
55 op->flags);
56 BUG();
57 break;
58 }
59 fscache_stat(&fscache_n_op_enqueue);
60 }
61}
62EXPORT_SYMBOL(fscache_enqueue_operation);
63
64/*
65 * start an op running
66 */
67static void fscache_run_op(struct fscache_object *object,
68 struct fscache_operation *op)
69{
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 spin_lock(&object->lock);
91 ASSERTCMP(object->n_ops, >=, object->n_in_progress);
92 ASSERTCMP(object->n_ops, >=, object->n_exclusive);
93
94 ret = -ENOBUFS;
95 if (fscache_object_is_active(object)) {
96 op->object = object;
97 object->n_ops++;
98 object->n_exclusive++; /* reads and writes must wait */
99
100 if (object->n_ops > 0) {
101 atomic_inc(&op->usage);
102 list_add_tail(&op->pend_link, &object->pending_ops);
103 fscache_stat(&fscache_n_op_pend);
104 } else if (!list_empty(&object->pending_ops)) {
105 atomic_inc(&op->usage);
106 list_add_tail(&op->pend_link, &object->pending_ops);
107 fscache_stat(&fscache_n_op_pend);
108 fscache_start_operations(object);
109 } else {
110 ASSERTCMP(object->n_in_progress, ==, 0);
111 fscache_run_op(object, op);
112 }
113
114 /* need to issue a new write op after this */
115 clear_bit(FSCACHE_OBJECT_PENDING_WRITE, &object->flags);
116 ret = 0;
117 } else if (object->state == FSCACHE_OBJECT_CREATING) {
118 op->object = object;
119 object->n_ops++;
120 object->n_exclusive++; /* reads and writes must wait */
121 atomic_inc(&op->usage);
122 list_add_tail(&op->pend_link, &object->pending_ops);
123 fscache_stat(&fscache_n_op_pend);
124 ret = 0;
125 } else {
126 /* not allowed to submit ops in any other state */
127 BUG();
128 }
129
130 spin_unlock(&object->lock);
131 return ret;
132}
133
134/*
135 * report an unexpected submission
136 */
137static void fscache_report_unexpected_submission(struct fscache_object *object,
138 struct fscache_operation *op,
139 unsigned long ostate)
140{
141 static bool once_only;
142 struct fscache_operation *p;
143 unsigned n;
144
145 if (once_only)
146 return;
147 once_only = true;
148
149 kdebug("unexpected submission OP%x [OBJ%x %s]",
150 op->debug_id, object->debug_id,
151 fscache_object_states[object->state]);
152 kdebug("objstate=%s [%s]",
153 fscache_object_states[object->state],
154 fscache_object_states[ostate]);
155 kdebug("objflags=%lx", object->flags);
156 kdebug("objevent=%lx [%lx]", object->events, object->event_mask);
157 kdebug("ops=%u inp=%u exc=%u",
158 object->n_ops, object->n_in_progress, object->n_exclusive);
159
160 if (!list_empty(&object->pending_ops)) {
161 n = 0;
162 list_for_each_entry(p, &object->pending_ops, pend_link) {
163 ASSERTCMP(p->object, ==, object);
164 kdebug("%p %p", op->processor, op->release);
165 n++;
166 }
167
168 kdebug("n=%u", n);
169 }
170
171 dump_stack();
172}
173
174/*
175 * submit an operation for an object
176 * - objects may be submitted only in the following states:
177 * - during object creation (write ops may be submitted)
178 * - whilst the object is active
179 * - after an I/O error incurred in one of the two above states (op rejected)
180 * - this gets any extra refs it needs on an op
181 */
182int fscache_submit_op(struct fscache_object *object,
183 struct fscache_operation *op)
184{
185 unsigned long ostate;
186 int ret;
187
188 _enter("{OBJ%x OP%x},{%u}",
189 object->debug_id, op->debug_id, atomic_read(&op->usage));
190
191 ASSERTCMP(atomic_read(&op->usage), >, 0);
192
193 spin_lock(&object->lock);
194 ASSERTCMP(object->n_ops, >=, object->n_in_progress);
195 ASSERTCMP(object->n_ops, >=, object->n_exclusive);
196
197 ostate = object->state;
198 smp_rmb();
199
200 if (fscache_object_is_active(object)) {
201 op->object = object;
202 object->n_ops++;
203
204 if (object->n_exclusive > 0) {
205 atomic_inc(&op->usage);
206 list_add_tail(&op->pend_link, &object->pending_ops);
207 fscache_stat(&fscache_n_op_pend);
208 } else if (!list_empty(&object->pending_ops)) {
209 atomic_inc(&op->usage);
210 list_add_tail(&op->pend_link, &object->pending_ops);
211 fscache_stat(&fscache_n_op_pend);
212 fscache_start_operations(object);
213 } else {
214 ASSERTCMP(object->n_exclusive, ==, 0);
215 fscache_run_op(object, op);
216 }
217 ret = 0;
218 } else if (object->state == FSCACHE_OBJECT_CREATING) {
219 op->object = object;
220 object->n_ops++;
221 atomic_inc(&op->usage);
222 list_add_tail(&op->pend_link, &object->pending_ops);
223 fscache_stat(&fscache_n_op_pend);
224 ret = 0;
225 } else if (!test_bit(FSCACHE_IOERROR, &object->cache->flags)) {
226 fscache_report_unexpected_submission(object, op, ostate);
227 ASSERT(!fscache_object_is_active(object));
228 ret = -ENOBUFS;
229 } else {
230 ret = -ENOBUFS;
231 }
232
233 spin_unlock(&object->lock);
234 return ret;
235}
236
237/*
238 * queue an object for withdrawal on error, aborting all following asynchronous
239 * operations
240 */
241void fscache_abort_object(struct fscache_object *object)
242{
243 _enter("{OBJ%x}", object->debug_id);
244
245 fscache_raise_event(object, FSCACHE_OBJECT_EV_ERROR);
246}
247
248/*
249 * jump start the operation processing on an object
250 * - caller must hold object->lock
251 */
252void fscache_start_operations(struct fscache_object *object)
253{
254 struct fscache_operation *op;
255 bool stop = false;
256
257 while (!list_empty(&object->pending_ops) && !stop) {
258 op = list_entry(object->pending_ops.next,
259 struct fscache_operation, pend_link);
260
261 if (test_bit(FSCACHE_OP_EXCLUSIVE, &op->flags)) {
262 if (object->n_in_progress > 0)
263 break;
264 stop = true;
265 }
266 list_del_init(&op->pend_link);
267 object->n_in_progress++;
268
269 if (test_and_clear_bit(FSCACHE_OP_WAITING, &op->flags))
270 wake_up_bit(&op->flags, FSCACHE_OP_WAITING);
271 if (op->processor)
272 fscache_enqueue_operation(op);
273
274 /* the pending queue was holding a ref on the object */
275 fscache_put_operation(op);
276 }
277
278 ASSERTCMP(object->n_in_progress, <=, object->n_ops);
279
280 _debug("woke %d ops on OBJ%x",
281 object->n_in_progress, object->debug_id);
282}
283
284/*
285 * release an operation
286 * - queues pending ops if this is the last in-progress op
287 */
288void fscache_put_operation(struct fscache_operation *op)
289{
290 struct fscache_object *object;
291 struct fscache_cache *cache;
292
293 _enter("{OBJ%x OP%x,%d}",
294 op->object->debug_id, op->debug_id, atomic_read(&op->usage));
295
296 ASSERTCMP(atomic_read(&op->usage), >, 0);
297
298 if (!atomic_dec_and_test(&op->usage))
299 return;
300
301 _debug("PUT OP");
302 if (test_and_set_bit(FSCACHE_OP_DEAD, &op->flags))
303 BUG();
304
305 fscache_stat(&fscache_n_op_release);
306
307 if (op->release) {
308 op->release(op);
309 op->release = NULL;
310 }
311
312 object = op->object;
313
314 /* now... we may get called with the object spinlock held, so we
315 * complete the cleanup here only if we can immediately acquire the
316 * lock, and defer it otherwise */
317 if (!spin_trylock(&object->lock)) {
318 _debug("defer put");
319 fscache_stat(&fscache_n_op_deferred_release);
320
321 cache = object->cache;
322 spin_lock(&cache->op_gc_list_lock);
323 list_add_tail(&op->pend_link, &cache->op_gc_list);
324 spin_unlock(&cache->op_gc_list_lock);
325 schedule_work(&cache->op_gc);
326 _leave(" [defer]");
327 return;
328 }
329
330 if (test_bit(FSCACHE_OP_EXCLUSIVE, &op->flags)) {
331 ASSERTCMP(object->n_exclusive, >, 0);
332 object->n_exclusive--;
333 }
334
335 ASSERTCMP(object->n_in_progress, >, 0);
336 object->n_in_progress--;
337 if (object->n_in_progress == 0)
338 fscache_start_operations(object);
339
340 ASSERTCMP(object->n_ops, >, 0);
341 object->n_ops--;
342 if (object->n_ops == 0)
343 fscache_raise_event(object, FSCACHE_OBJECT_EV_CLEARED);
344
345 spin_unlock(&object->lock);
346
347 kfree(op);
348 _leave(" [done]");
349}
350EXPORT_SYMBOL(fscache_put_operation);
351
352/*
353 * garbage collect operations that have had their release deferred
354 */
355void fscache_operation_gc(struct work_struct *work)
356{
357 struct fscache_operation *op;
358 struct fscache_object *object;
359 struct fscache_cache *cache =
360 container_of(work, struct fscache_cache, op_gc);
361 int count = 0;
362
363 _enter("");
364
365 do {
366 spin_lock(&cache->op_gc_list_lock);
367 if (list_empty(&cache->op_gc_list)) {
368 spin_unlock(&cache->op_gc_list_lock);
369 break;
370 }
371
372 op = list_entry(cache->op_gc_list.next,
373 struct fscache_operation, pend_link);
374 list_del(&op->pend_link);
375 spin_unlock(&cache->op_gc_list_lock);
376
377 object = op->object;
378
379 _debug("GC DEFERRED REL OBJ%x OP%x",
380 object->debug_id, op->debug_id);
381 fscache_stat(&fscache_n_op_gc);
382
383 ASSERTCMP(atomic_read(&op->usage), ==, 0);
384
385 spin_lock(&object->lock);
386 if (test_bit(FSCACHE_OP_EXCLUSIVE, &op->flags)) {
387 ASSERTCMP(object->n_exclusive, >, 0);
388 object->n_exclusive--;
389 }
390
391 ASSERTCMP(object->n_in_progress, >, 0);
392 object->n_in_progress--;
393 if (object->n_in_progress == 0)
394 fscache_start_operations(object);
395
396 ASSERTCMP(object->n_ops, >, 0);
397 object->n_ops--;
398 if (object->n_ops == 0)
399 fscache_raise_event(object, FSCACHE_OBJECT_EV_CLEARED);
400
401 spin_unlock(&object->lock);
402
403 } while (count++ < 20);
404
405 if (!list_empty(&cache->op_gc_list))
406 schedule_work(&cache->op_gc);
407
408 _leave("");
409}
410
411/*
412 * allow the slow work item processor to get a ref on an operation
413 */
414static int fscache_op_get_ref(struct slow_work *work)
415{
416 struct fscache_operation *op =
417 container_of(work, struct fscache_operation, slow_work);
418
419 atomic_inc(&op->usage);
420 return 0;
421}
422
423/*
424 * allow the slow work item processor to discard a ref on an operation
425 */
426static void fscache_op_put_ref(struct slow_work *work)
427{
428 struct fscache_operation *op =
429 container_of(work, struct fscache_operation, slow_work);
430
431 fscache_put_operation(op);
432}
433
434/*
435 * execute an operation using the slow thread pool to provide processing context
436 * - the caller holds a ref to this object, so we don't need to hold one
437 */
438static void fscache_op_execute(struct slow_work *work)
439{
440 struct fscache_operation *op =
441 container_of(work, struct fscache_operation, slow_work);
442 unsigned long start;
443
444 _enter("{OBJ%x OP%x,%d}",
445 op->object->debug_id, op->debug_id, atomic_read(&op->usage));
446
447 ASSERT(op->processor != NULL);
448 start = jiffies;
449 op->processor(op);
450 fscache_hist(fscache_ops_histogram, start);
451
452 _leave("");
453}
454
455const struct slow_work_ops fscache_op_slow_work_ops = {
456 .get_ref = fscache_op_get_ref,
457 .put_ref = fscache_op_put_ref,
458 .execute = fscache_op_execute,
459};