blob: ade8a7e99cd3049f1e4f5e5284146be446a9e67a [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001/*
2 * net/sunrpc/cache.c
3 *
4 * Generic code for various authentication-related caches
5 * used by sunrpc clients and servers.
6 *
7 * Copyright (C) 2002 Neil Brown <neilb@cse.unsw.edu.au>
8 *
9 * Released under terms in GPL version 2. See COPYING.
10 *
11 */
12
13#include <linux/types.h>
14#include <linux/fs.h>
15#include <linux/file.h>
16#include <linux/slab.h>
17#include <linux/signal.h>
18#include <linux/sched.h>
19#include <linux/kmod.h>
20#include <linux/list.h>
21#include <linux/module.h>
22#include <linux/ctype.h>
23#include <asm/uaccess.h>
24#include <linux/poll.h>
25#include <linux/seq_file.h>
26#include <linux/proc_fs.h>
27#include <linux/net.h>
28#include <linux/workqueue.h>
Arjan van de Ven4a3e2f72006-03-20 22:33:17 -080029#include <linux/mutex.h>
Trond Myklebustda770052009-08-09 15:14:28 -040030#include <linux/pagemap.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070031#include <asm/ioctls.h>
32#include <linux/sunrpc/types.h>
33#include <linux/sunrpc/cache.h>
34#include <linux/sunrpc/stats.h>
Trond Myklebust8854e822009-08-09 15:14:30 -040035#include <linux/sunrpc/rpc_pipe_fs.h>
Linus Torvalds1da177e2005-04-16 15:20:36 -070036
37#define RPCDBG_FACILITY RPCDBG_CACHE
38
J.Bruce Fieldse0bb89e2006-12-13 00:35:25 -080039static int cache_defer_req(struct cache_req *req, struct cache_head *item);
Linus Torvalds1da177e2005-04-16 15:20:36 -070040static void cache_revisit_request(struct cache_head *item);
41
Adrian Bunk74cae612006-03-27 01:15:10 -080042static void cache_init(struct cache_head *h)
Linus Torvalds1da177e2005-04-16 15:20:36 -070043{
44 time_t now = get_seconds();
45 h->next = NULL;
46 h->flags = 0;
NeilBrownbaab9352006-03-27 01:15:09 -080047 kref_init(&h->ref);
Linus Torvalds1da177e2005-04-16 15:20:36 -070048 h->expiry_time = now + CACHE_NEW_EXPIRY;
49 h->last_refresh = now;
50}
51
NeilBrown15a5f6b2006-03-27 01:15:02 -080052struct cache_head *sunrpc_cache_lookup(struct cache_detail *detail,
53 struct cache_head *key, int hash)
54{
55 struct cache_head **head, **hp;
56 struct cache_head *new = NULL;
57
58 head = &detail->hash_table[hash];
59
60 read_lock(&detail->hash_lock);
61
62 for (hp=head; *hp != NULL ; hp = &(*hp)->next) {
63 struct cache_head *tmp = *hp;
64 if (detail->match(tmp, key)) {
65 cache_get(tmp);
66 read_unlock(&detail->hash_lock);
67 return tmp;
68 }
69 }
70 read_unlock(&detail->hash_lock);
71 /* Didn't find anything, insert an empty entry */
72
73 new = detail->alloc();
74 if (!new)
75 return NULL;
Neil Brown2f349312006-08-05 12:14:29 -070076 /* must fully initialise 'new', else
77 * we might get lose if we need to
78 * cache_put it soon.
79 */
NeilBrown15a5f6b2006-03-27 01:15:02 -080080 cache_init(new);
Neil Brown2f349312006-08-05 12:14:29 -070081 detail->init(new, key);
NeilBrown15a5f6b2006-03-27 01:15:02 -080082
83 write_lock(&detail->hash_lock);
84
85 /* check if entry appeared while we slept */
86 for (hp=head; *hp != NULL ; hp = &(*hp)->next) {
87 struct cache_head *tmp = *hp;
88 if (detail->match(tmp, key)) {
89 cache_get(tmp);
90 write_unlock(&detail->hash_lock);
NeilBrownbaab9352006-03-27 01:15:09 -080091 cache_put(new, detail);
NeilBrown15a5f6b2006-03-27 01:15:02 -080092 return tmp;
93 }
94 }
NeilBrown15a5f6b2006-03-27 01:15:02 -080095 new->next = *head;
96 *head = new;
97 detail->entries++;
98 cache_get(new);
99 write_unlock(&detail->hash_lock);
100
101 return new;
102}
Trond Myklebust24c37672008-12-23 16:30:12 -0500103EXPORT_SYMBOL_GPL(sunrpc_cache_lookup);
NeilBrown15a5f6b2006-03-27 01:15:02 -0800104
NeilBrownebd0cb12006-03-27 01:15:08 -0800105
NeilBrownf866a812009-08-04 15:22:38 +1000106static void cache_dequeue(struct cache_detail *detail, struct cache_head *ch);
NeilBrownebd0cb12006-03-27 01:15:08 -0800107
108static int cache_fresh_locked(struct cache_head *head, time_t expiry)
109{
110 head->expiry_time = expiry;
111 head->last_refresh = get_seconds();
112 return !test_and_set_bit(CACHE_VALID, &head->flags);
113}
114
115static void cache_fresh_unlocked(struct cache_head *head,
116 struct cache_detail *detail, int new)
117{
118 if (new)
119 cache_revisit_request(head);
120 if (test_and_clear_bit(CACHE_PENDING, &head->flags)) {
121 cache_revisit_request(head);
NeilBrownf866a812009-08-04 15:22:38 +1000122 cache_dequeue(detail, head);
NeilBrownebd0cb12006-03-27 01:15:08 -0800123 }
124}
125
NeilBrown15a5f6b2006-03-27 01:15:02 -0800126struct cache_head *sunrpc_cache_update(struct cache_detail *detail,
127 struct cache_head *new, struct cache_head *old, int hash)
128{
129 /* The 'old' entry is to be replaced by 'new'.
130 * If 'old' is not VALID, we update it directly,
131 * otherwise we need to replace it
132 */
133 struct cache_head **head;
134 struct cache_head *tmp;
NeilBrownebd0cb12006-03-27 01:15:08 -0800135 int is_new;
NeilBrown15a5f6b2006-03-27 01:15:02 -0800136
137 if (!test_bit(CACHE_VALID, &old->flags)) {
138 write_lock(&detail->hash_lock);
139 if (!test_bit(CACHE_VALID, &old->flags)) {
140 if (test_bit(CACHE_NEGATIVE, &new->flags))
141 set_bit(CACHE_NEGATIVE, &old->flags);
142 else
143 detail->update(old, new);
NeilBrownebd0cb12006-03-27 01:15:08 -0800144 is_new = cache_fresh_locked(old, new->expiry_time);
NeilBrown15a5f6b2006-03-27 01:15:02 -0800145 write_unlock(&detail->hash_lock);
NeilBrownebd0cb12006-03-27 01:15:08 -0800146 cache_fresh_unlocked(old, detail, is_new);
NeilBrown15a5f6b2006-03-27 01:15:02 -0800147 return old;
148 }
149 write_unlock(&detail->hash_lock);
150 }
151 /* We need to insert a new entry */
152 tmp = detail->alloc();
153 if (!tmp) {
NeilBrownbaab9352006-03-27 01:15:09 -0800154 cache_put(old, detail);
NeilBrown15a5f6b2006-03-27 01:15:02 -0800155 return NULL;
156 }
157 cache_init(tmp);
158 detail->init(tmp, old);
159 head = &detail->hash_table[hash];
160
161 write_lock(&detail->hash_lock);
162 if (test_bit(CACHE_NEGATIVE, &new->flags))
163 set_bit(CACHE_NEGATIVE, &tmp->flags);
164 else
165 detail->update(tmp, new);
166 tmp->next = *head;
167 *head = tmp;
NeilBrownf2d39582006-05-22 22:35:25 -0700168 detail->entries++;
NeilBrown15a5f6b2006-03-27 01:15:02 -0800169 cache_get(tmp);
NeilBrownebd0cb12006-03-27 01:15:08 -0800170 is_new = cache_fresh_locked(tmp, new->expiry_time);
171 cache_fresh_locked(old, 0);
NeilBrown15a5f6b2006-03-27 01:15:02 -0800172 write_unlock(&detail->hash_lock);
NeilBrownebd0cb12006-03-27 01:15:08 -0800173 cache_fresh_unlocked(tmp, detail, is_new);
174 cache_fresh_unlocked(old, detail, 0);
NeilBrownbaab9352006-03-27 01:15:09 -0800175 cache_put(old, detail);
NeilBrown15a5f6b2006-03-27 01:15:02 -0800176 return tmp;
177}
Trond Myklebust24c37672008-12-23 16:30:12 -0500178EXPORT_SYMBOL_GPL(sunrpc_cache_update);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700179
Trond Myklebustbc74b4f2009-08-09 15:14:29 -0400180static int cache_make_upcall(struct cache_detail *cd, struct cache_head *h)
181{
182 if (!cd->cache_upcall)
183 return -EINVAL;
184 return cd->cache_upcall(cd, h);
185}
NeilBrown989a19b2009-08-04 15:22:38 +1000186
187static inline int cache_is_valid(struct cache_detail *detail, struct cache_head *h)
188{
189 if (!test_bit(CACHE_VALID, &h->flags) ||
190 h->expiry_time < get_seconds())
191 return -EAGAIN;
192 else if (detail->flush_time > h->last_refresh)
193 return -EAGAIN;
194 else {
195 /* entry is valid */
196 if (test_bit(CACHE_NEGATIVE, &h->flags))
197 return -ENOENT;
198 else
199 return 0;
200 }
201}
J. Bruce Fieldse9dc1222009-08-21 11:27:29 -0400202
Linus Torvalds1da177e2005-04-16 15:20:36 -0700203/*
204 * This is the generic cache management routine for all
205 * the authentication caches.
206 * It checks the currency of a cache item and will (later)
207 * initiate an upcall to fill it if needed.
208 *
209 *
210 * Returns 0 if the cache_head can be used, or cache_puts it and returns
NeilBrown989a19b2009-08-04 15:22:38 +1000211 * -EAGAIN if upcall is pending and request has been queued
212 * -ETIMEDOUT if upcall failed or request could not be queue or
213 * upcall completed but item is still invalid (implying that
214 * the cache item has been replaced with a newer one).
Linus Torvalds1da177e2005-04-16 15:20:36 -0700215 * -ENOENT if cache entry was negative
216 */
217int cache_check(struct cache_detail *detail,
218 struct cache_head *h, struct cache_req *rqstp)
219{
220 int rv;
221 long refresh_age, age;
222
223 /* First decide return status as best we can */
NeilBrown989a19b2009-08-04 15:22:38 +1000224 rv = cache_is_valid(detail, h);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700225
226 /* now see if we want to start an upcall */
227 refresh_age = (h->expiry_time - h->last_refresh);
228 age = get_seconds() - h->last_refresh;
229
230 if (rqstp == NULL) {
231 if (rv == -EAGAIN)
232 rv = -ENOENT;
233 } else if (rv == -EAGAIN || age > refresh_age/2) {
Chuck Lever46121cf2007-01-31 12:14:08 -0500234 dprintk("RPC: Want update, refage=%ld, age=%ld\n",
235 refresh_age, age);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700236 if (!test_and_set_bit(CACHE_PENDING, &h->flags)) {
237 switch (cache_make_upcall(detail, h)) {
238 case -EINVAL:
239 clear_bit(CACHE_PENDING, &h->flags);
NeilBrown5c4d2632009-08-04 15:22:38 +1000240 cache_revisit_request(h);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700241 if (rv == -EAGAIN) {
242 set_bit(CACHE_NEGATIVE, &h->flags);
NeilBrownebd0cb12006-03-27 01:15:08 -0800243 cache_fresh_unlocked(h, detail,
244 cache_fresh_locked(h, get_seconds()+CACHE_NEW_EXPIRY));
Linus Torvalds1da177e2005-04-16 15:20:36 -0700245 rv = -ENOENT;
246 }
247 break;
248
249 case -EAGAIN:
250 clear_bit(CACHE_PENDING, &h->flags);
251 cache_revisit_request(h);
252 break;
253 }
254 }
255 }
256
NeilBrown989a19b2009-08-04 15:22:38 +1000257 if (rv == -EAGAIN) {
258 if (cache_defer_req(rqstp, h) == 0) {
259 /* Request is not deferred */
260 rv = cache_is_valid(detail, h);
261 if (rv == -EAGAIN)
262 rv = -ETIMEDOUT;
263 }
264 }
NeilBrown4013ede2006-03-27 01:15:07 -0800265 if (rv)
NeilBrownbaab9352006-03-27 01:15:09 -0800266 cache_put(h, detail);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700267 return rv;
268}
Trond Myklebust24c37672008-12-23 16:30:12 -0500269EXPORT_SYMBOL_GPL(cache_check);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700270
Linus Torvalds1da177e2005-04-16 15:20:36 -0700271/*
272 * caches need to be periodically cleaned.
273 * For this we maintain a list of cache_detail and
274 * a current pointer into that list and into the table
275 * for that entry.
276 *
277 * Each time clean_cache is called it finds the next non-empty entry
278 * in the current table and walks the list in that entry
279 * looking for entries that can be removed.
280 *
281 * An entry gets removed if:
282 * - The expiry is before current time
283 * - The last_refresh time is before the flush_time for that cache
284 *
285 * later we might drop old entries with non-NEVER expiry if that table
286 * is getting 'full' for some definition of 'full'
287 *
288 * The question of "how often to scan a table" is an interesting one
289 * and is answered in part by the use of the "nextcheck" field in the
290 * cache_detail.
291 * When a scan of a table begins, the nextcheck field is set to a time
292 * that is well into the future.
293 * While scanning, if an expiry time is found that is earlier than the
294 * current nextcheck time, nextcheck is set to that expiry time.
295 * If the flush_time is ever set to a time earlier than the nextcheck
296 * time, the nextcheck time is then set to that flush_time.
297 *
298 * A table is then only scanned if the current time is at least
299 * the nextcheck time.
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -0800300 *
Linus Torvalds1da177e2005-04-16 15:20:36 -0700301 */
302
303static LIST_HEAD(cache_list);
304static DEFINE_SPINLOCK(cache_list_lock);
305static struct cache_detail *current_detail;
306static int current_index;
307
David Howells65f27f32006-11-22 14:55:48 +0000308static void do_cache_clean(struct work_struct *work);
309static DECLARE_DELAYED_WORK(cache_cleaner, do_cache_clean);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700310
Trond Myklebust5b7a1b92009-08-09 15:14:27 -0400311static void sunrpc_init_cache_detail(struct cache_detail *cd)
J. Bruce Fieldsffe93862007-11-12 17:04:29 -0500312{
Linus Torvalds1da177e2005-04-16 15:20:36 -0700313 rwlock_init(&cd->hash_lock);
314 INIT_LIST_HEAD(&cd->queue);
315 spin_lock(&cache_list_lock);
316 cd->nextcheck = 0;
317 cd->entries = 0;
318 atomic_set(&cd->readers, 0);
319 cd->last_close = 0;
320 cd->last_warn = -1;
321 list_add(&cd->others, &cache_list);
322 spin_unlock(&cache_list_lock);
323
324 /* start the cleaning process */
David Howells52bad642006-11-22 14:54:01 +0000325 schedule_delayed_work(&cache_cleaner, 0);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700326}
327
Trond Myklebust5b7a1b92009-08-09 15:14:27 -0400328static void sunrpc_destroy_cache_detail(struct cache_detail *cd)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700329{
330 cache_purge(cd);
331 spin_lock(&cache_list_lock);
332 write_lock(&cd->hash_lock);
333 if (cd->entries || atomic_read(&cd->inuse)) {
334 write_unlock(&cd->hash_lock);
335 spin_unlock(&cache_list_lock);
J. Bruce Fieldsdf95a9d2007-11-08 16:09:59 -0500336 goto out;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700337 }
338 if (current_detail == cd)
339 current_detail = NULL;
340 list_del_init(&cd->others);
341 write_unlock(&cd->hash_lock);
342 spin_unlock(&cache_list_lock);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700343 if (list_empty(&cache_list)) {
344 /* module must be being unloaded so its safe to kill the worker */
Trond Myklebust4011cd92007-08-07 15:33:01 -0400345 cancel_delayed_work_sync(&cache_cleaner);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700346 }
J. Bruce Fieldsdf95a9d2007-11-08 16:09:59 -0500347 return;
348out:
349 printk(KERN_ERR "nfsd: failed to unregister %s cache\n", cd->name);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700350}
351
352/* clean cache tries to find something to clean
353 * and cleans it.
354 * It returns 1 if it cleaned something,
355 * 0 if it didn't find anything this time
356 * -1 if it fell off the end of the list.
357 */
358static int cache_clean(void)
359{
360 int rv = 0;
361 struct list_head *next;
362
363 spin_lock(&cache_list_lock);
364
365 /* find a suitable table if we don't already have one */
366 while (current_detail == NULL ||
367 current_index >= current_detail->hash_size) {
368 if (current_detail)
369 next = current_detail->others.next;
370 else
371 next = cache_list.next;
372 if (next == &cache_list) {
373 current_detail = NULL;
374 spin_unlock(&cache_list_lock);
375 return -1;
376 }
377 current_detail = list_entry(next, struct cache_detail, others);
378 if (current_detail->nextcheck > get_seconds())
379 current_index = current_detail->hash_size;
380 else {
381 current_index = 0;
382 current_detail->nextcheck = get_seconds()+30*60;
383 }
384 }
385
386 /* find a non-empty bucket in the table */
387 while (current_detail &&
388 current_index < current_detail->hash_size &&
389 current_detail->hash_table[current_index] == NULL)
390 current_index++;
391
392 /* find a cleanable entry in the bucket and clean it, or set to next bucket */
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -0800393
Linus Torvalds1da177e2005-04-16 15:20:36 -0700394 if (current_detail && current_index < current_detail->hash_size) {
395 struct cache_head *ch, **cp;
396 struct cache_detail *d;
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -0800397
Linus Torvalds1da177e2005-04-16 15:20:36 -0700398 write_lock(&current_detail->hash_lock);
399
400 /* Ok, now to clean this strand */
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -0800401
Linus Torvalds1da177e2005-04-16 15:20:36 -0700402 cp = & current_detail->hash_table[current_index];
403 ch = *cp;
404 for (; ch; cp= & ch->next, ch= *cp) {
405 if (current_detail->nextcheck > ch->expiry_time)
406 current_detail->nextcheck = ch->expiry_time+1;
407 if (ch->expiry_time >= get_seconds()
408 && ch->last_refresh >= current_detail->flush_time
409 )
410 continue;
411 if (test_and_clear_bit(CACHE_PENDING, &ch->flags))
NeilBrownf866a812009-08-04 15:22:38 +1000412 cache_dequeue(current_detail, ch);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700413
NeilBrownbaab9352006-03-27 01:15:09 -0800414 if (atomic_read(&ch->ref.refcount) == 1)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700415 break;
416 }
417 if (ch) {
418 *cp = ch->next;
419 ch->next = NULL;
420 current_detail->entries--;
421 rv = 1;
422 }
423 write_unlock(&current_detail->hash_lock);
424 d = current_detail;
425 if (!ch)
426 current_index ++;
427 spin_unlock(&cache_list_lock);
NeilBrown5c4d2632009-08-04 15:22:38 +1000428 if (ch) {
429 cache_revisit_request(ch);
NeilBrownbaab9352006-03-27 01:15:09 -0800430 cache_put(ch, d);
NeilBrown5c4d2632009-08-04 15:22:38 +1000431 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700432 } else
433 spin_unlock(&cache_list_lock);
434
435 return rv;
436}
437
438/*
439 * We want to regularly clean the cache, so we need to schedule some work ...
440 */
David Howells65f27f32006-11-22 14:55:48 +0000441static void do_cache_clean(struct work_struct *work)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700442{
443 int delay = 5;
444 if (cache_clean() == -1)
Anton Blanchard6aad89c2009-06-10 12:52:21 -0700445 delay = round_jiffies_relative(30*HZ);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700446
447 if (list_empty(&cache_list))
448 delay = 0;
449
450 if (delay)
451 schedule_delayed_work(&cache_cleaner, delay);
452}
453
454
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -0800455/*
Linus Torvalds1da177e2005-04-16 15:20:36 -0700456 * Clean all caches promptly. This just calls cache_clean
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -0800457 * repeatedly until we are sure that every cache has had a chance to
Linus Torvalds1da177e2005-04-16 15:20:36 -0700458 * be fully cleaned
459 */
460void cache_flush(void)
461{
462 while (cache_clean() != -1)
463 cond_resched();
464 while (cache_clean() != -1)
465 cond_resched();
466}
Trond Myklebust24c37672008-12-23 16:30:12 -0500467EXPORT_SYMBOL_GPL(cache_flush);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700468
469void cache_purge(struct cache_detail *detail)
470{
471 detail->flush_time = LONG_MAX;
472 detail->nextcheck = get_seconds();
473 cache_flush();
474 detail->flush_time = 1;
475}
Trond Myklebust24c37672008-12-23 16:30:12 -0500476EXPORT_SYMBOL_GPL(cache_purge);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700477
478
479/*
480 * Deferral and Revisiting of Requests.
481 *
482 * If a cache lookup finds a pending entry, we
483 * need to defer the request and revisit it later.
484 * All deferred requests are stored in a hash table,
485 * indexed by "struct cache_head *".
486 * As it may be wasteful to store a whole request
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -0800487 * structure, we allow the request to provide a
Linus Torvalds1da177e2005-04-16 15:20:36 -0700488 * deferred form, which must contain a
489 * 'struct cache_deferred_req'
490 * This cache_deferred_req contains a method to allow
491 * it to be revisited when cache info is available
492 */
493
494#define DFR_HASHSIZE (PAGE_SIZE/sizeof(struct list_head))
495#define DFR_HASH(item) ((((long)item)>>4 ^ (((long)item)>>13)) % DFR_HASHSIZE)
496
497#define DFR_MAX 300 /* ??? */
498
499static DEFINE_SPINLOCK(cache_defer_lock);
500static LIST_HEAD(cache_defer_list);
501static struct list_head cache_defer_hash[DFR_HASHSIZE];
502static int cache_defer_cnt;
503
J.Bruce Fieldse0bb89e2006-12-13 00:35:25 -0800504static int cache_defer_req(struct cache_req *req, struct cache_head *item)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700505{
506 struct cache_deferred_req *dreq;
507 int hash = DFR_HASH(item);
508
J.Bruce Fields01f3bd12006-12-13 00:35:26 -0800509 if (cache_defer_cnt >= DFR_MAX) {
510 /* too much in the cache, randomly drop this one,
511 * or continue and drop the oldest below
512 */
513 if (net_random()&1)
NeilBrown989a19b2009-08-04 15:22:38 +1000514 return 0;
J.Bruce Fields01f3bd12006-12-13 00:35:26 -0800515 }
Linus Torvalds1da177e2005-04-16 15:20:36 -0700516 dreq = req->defer(req);
517 if (dreq == NULL)
NeilBrown989a19b2009-08-04 15:22:38 +1000518 return 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700519
520 dreq->item = item;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700521
522 spin_lock(&cache_defer_lock);
523
524 list_add(&dreq->recent, &cache_defer_list);
525
526 if (cache_defer_hash[hash].next == NULL)
527 INIT_LIST_HEAD(&cache_defer_hash[hash]);
528 list_add(&dreq->hash, &cache_defer_hash[hash]);
529
530 /* it is in, now maybe clean up */
531 dreq = NULL;
532 if (++cache_defer_cnt > DFR_MAX) {
J.Bruce Fields01f3bd12006-12-13 00:35:26 -0800533 dreq = list_entry(cache_defer_list.prev,
534 struct cache_deferred_req, recent);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700535 list_del(&dreq->recent);
536 list_del(&dreq->hash);
537 cache_defer_cnt--;
538 }
539 spin_unlock(&cache_defer_lock);
540
541 if (dreq) {
542 /* there was one too many */
543 dreq->revisit(dreq, 1);
544 }
NeilBrown4013ede2006-03-27 01:15:07 -0800545 if (!test_bit(CACHE_PENDING, &item->flags)) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700546 /* must have just been validated... */
547 cache_revisit_request(item);
NeilBrown989a19b2009-08-04 15:22:38 +1000548 return 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700549 }
NeilBrown989a19b2009-08-04 15:22:38 +1000550 return 1;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700551}
552
553static void cache_revisit_request(struct cache_head *item)
554{
555 struct cache_deferred_req *dreq;
556 struct list_head pending;
557
558 struct list_head *lp;
559 int hash = DFR_HASH(item);
560
561 INIT_LIST_HEAD(&pending);
562 spin_lock(&cache_defer_lock);
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -0800563
Linus Torvalds1da177e2005-04-16 15:20:36 -0700564 lp = cache_defer_hash[hash].next;
565 if (lp) {
566 while (lp != &cache_defer_hash[hash]) {
567 dreq = list_entry(lp, struct cache_deferred_req, hash);
568 lp = lp->next;
569 if (dreq->item == item) {
570 list_del(&dreq->hash);
571 list_move(&dreq->recent, &pending);
572 cache_defer_cnt--;
573 }
574 }
575 }
576 spin_unlock(&cache_defer_lock);
577
578 while (!list_empty(&pending)) {
579 dreq = list_entry(pending.next, struct cache_deferred_req, recent);
580 list_del_init(&dreq->recent);
581 dreq->revisit(dreq, 0);
582 }
583}
584
585void cache_clean_deferred(void *owner)
586{
587 struct cache_deferred_req *dreq, *tmp;
588 struct list_head pending;
589
590
591 INIT_LIST_HEAD(&pending);
592 spin_lock(&cache_defer_lock);
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -0800593
Linus Torvalds1da177e2005-04-16 15:20:36 -0700594 list_for_each_entry_safe(dreq, tmp, &cache_defer_list, recent) {
595 if (dreq->owner == owner) {
596 list_del(&dreq->hash);
597 list_move(&dreq->recent, &pending);
598 cache_defer_cnt--;
599 }
600 }
601 spin_unlock(&cache_defer_lock);
602
603 while (!list_empty(&pending)) {
604 dreq = list_entry(pending.next, struct cache_deferred_req, recent);
605 list_del_init(&dreq->recent);
606 dreq->revisit(dreq, 1);
607 }
608}
609
610/*
611 * communicate with user-space
612 *
J. Bruce Fieldsa490c682007-11-06 14:15:19 -0500613 * We have a magic /proc file - /proc/sunrpc/<cachename>/channel.
614 * On read, you get a full request, or block.
615 * On write, an update request is processed.
616 * Poll works if anything to read, and always allows write.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700617 *
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -0800618 * Implemented by linked list of requests. Each open file has
J. Bruce Fieldsa490c682007-11-06 14:15:19 -0500619 * a ->private that also exists in this list. New requests are added
Linus Torvalds1da177e2005-04-16 15:20:36 -0700620 * to the end and may wakeup and preceding readers.
621 * New readers are added to the head. If, on read, an item is found with
622 * CACHE_UPCALLING clear, we free it from the list.
623 *
624 */
625
626static DEFINE_SPINLOCK(queue_lock);
Arjan van de Ven4a3e2f72006-03-20 22:33:17 -0800627static DEFINE_MUTEX(queue_io_mutex);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700628
629struct cache_queue {
630 struct list_head list;
631 int reader; /* if 0, then request */
632};
633struct cache_request {
634 struct cache_queue q;
635 struct cache_head *item;
636 char * buf;
637 int len;
638 int readers;
639};
640struct cache_reader {
641 struct cache_queue q;
642 int offset; /* if non-0, we have a refcnt on next request */
643};
644
Trond Myklebust173912a2009-08-09 15:14:29 -0400645static ssize_t cache_read(struct file *filp, char __user *buf, size_t count,
646 loff_t *ppos, struct cache_detail *cd)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700647{
648 struct cache_reader *rp = filp->private_data;
649 struct cache_request *rq;
Trond Myklebustda770052009-08-09 15:14:28 -0400650 struct inode *inode = filp->f_path.dentry->d_inode;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700651 int err;
652
653 if (count == 0)
654 return 0;
655
Trond Myklebustda770052009-08-09 15:14:28 -0400656 mutex_lock(&inode->i_mutex); /* protect against multiple concurrent
Linus Torvalds1da177e2005-04-16 15:20:36 -0700657 * readers on this file */
658 again:
659 spin_lock(&queue_lock);
660 /* need to find next request */
661 while (rp->q.list.next != &cd->queue &&
662 list_entry(rp->q.list.next, struct cache_queue, list)
663 ->reader) {
664 struct list_head *next = rp->q.list.next;
665 list_move(&rp->q.list, next);
666 }
667 if (rp->q.list.next == &cd->queue) {
668 spin_unlock(&queue_lock);
Trond Myklebustda770052009-08-09 15:14:28 -0400669 mutex_unlock(&inode->i_mutex);
Kris Katterjohn09a62662006-01-08 22:24:28 -0800670 BUG_ON(rp->offset);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700671 return 0;
672 }
673 rq = container_of(rp->q.list.next, struct cache_request, q.list);
Kris Katterjohn09a62662006-01-08 22:24:28 -0800674 BUG_ON(rq->q.reader);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700675 if (rp->offset == 0)
676 rq->readers++;
677 spin_unlock(&queue_lock);
678
679 if (rp->offset == 0 && !test_bit(CACHE_PENDING, &rq->item->flags)) {
680 err = -EAGAIN;
681 spin_lock(&queue_lock);
682 list_move(&rp->q.list, &rq->q.list);
683 spin_unlock(&queue_lock);
684 } else {
685 if (rp->offset + count > rq->len)
686 count = rq->len - rp->offset;
687 err = -EFAULT;
688 if (copy_to_user(buf, rq->buf + rp->offset, count))
689 goto out;
690 rp->offset += count;
691 if (rp->offset >= rq->len) {
692 rp->offset = 0;
693 spin_lock(&queue_lock);
694 list_move(&rp->q.list, &rq->q.list);
695 spin_unlock(&queue_lock);
696 }
697 err = 0;
698 }
699 out:
700 if (rp->offset == 0) {
701 /* need to release rq */
702 spin_lock(&queue_lock);
703 rq->readers--;
704 if (rq->readers == 0 &&
705 !test_bit(CACHE_PENDING, &rq->item->flags)) {
706 list_del(&rq->q.list);
707 spin_unlock(&queue_lock);
NeilBrownbaab9352006-03-27 01:15:09 -0800708 cache_put(rq->item, cd);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700709 kfree(rq->buf);
710 kfree(rq);
711 } else
712 spin_unlock(&queue_lock);
713 }
714 if (err == -EAGAIN)
715 goto again;
Trond Myklebustda770052009-08-09 15:14:28 -0400716 mutex_unlock(&inode->i_mutex);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700717 return err ? err : count;
718}
719
Trond Myklebustda770052009-08-09 15:14:28 -0400720static ssize_t cache_do_downcall(char *kaddr, const char __user *buf,
721 size_t count, struct cache_detail *cd)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700722{
Trond Myklebustda770052009-08-09 15:14:28 -0400723 ssize_t ret;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700724
Trond Myklebustda770052009-08-09 15:14:28 -0400725 if (copy_from_user(kaddr, buf, count))
Linus Torvalds1da177e2005-04-16 15:20:36 -0700726 return -EFAULT;
Trond Myklebustda770052009-08-09 15:14:28 -0400727 kaddr[count] = '\0';
728 ret = cd->cache_parse(cd, kaddr, count);
729 if (!ret)
730 ret = count;
731 return ret;
732}
Linus Torvalds1da177e2005-04-16 15:20:36 -0700733
Trond Myklebustda770052009-08-09 15:14:28 -0400734static ssize_t cache_slow_downcall(const char __user *buf,
735 size_t count, struct cache_detail *cd)
736{
737 static char write_buf[8192]; /* protected by queue_io_mutex */
738 ssize_t ret = -EINVAL;
739
740 if (count >= sizeof(write_buf))
741 goto out;
742 mutex_lock(&queue_io_mutex);
743 ret = cache_do_downcall(write_buf, buf, count, cd);
Arjan van de Ven4a3e2f72006-03-20 22:33:17 -0800744 mutex_unlock(&queue_io_mutex);
Trond Myklebustda770052009-08-09 15:14:28 -0400745out:
746 return ret;
747}
748
749static ssize_t cache_downcall(struct address_space *mapping,
750 const char __user *buf,
751 size_t count, struct cache_detail *cd)
752{
753 struct page *page;
754 char *kaddr;
755 ssize_t ret = -ENOMEM;
756
757 if (count >= PAGE_CACHE_SIZE)
758 goto out_slow;
759
760 page = find_or_create_page(mapping, 0, GFP_KERNEL);
761 if (!page)
762 goto out_slow;
763
764 kaddr = kmap(page);
765 ret = cache_do_downcall(kaddr, buf, count, cd);
766 kunmap(page);
767 unlock_page(page);
768 page_cache_release(page);
769 return ret;
770out_slow:
771 return cache_slow_downcall(buf, count, cd);
772}
Linus Torvalds1da177e2005-04-16 15:20:36 -0700773
Trond Myklebust173912a2009-08-09 15:14:29 -0400774static ssize_t cache_write(struct file *filp, const char __user *buf,
775 size_t count, loff_t *ppos,
776 struct cache_detail *cd)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700777{
Trond Myklebustda770052009-08-09 15:14:28 -0400778 struct address_space *mapping = filp->f_mapping;
779 struct inode *inode = filp->f_path.dentry->d_inode;
Trond Myklebustda770052009-08-09 15:14:28 -0400780 ssize_t ret = -EINVAL;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700781
Trond Myklebustda770052009-08-09 15:14:28 -0400782 if (!cd->cache_parse)
783 goto out;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700784
Trond Myklebustda770052009-08-09 15:14:28 -0400785 mutex_lock(&inode->i_mutex);
786 ret = cache_downcall(mapping, buf, count, cd);
787 mutex_unlock(&inode->i_mutex);
788out:
789 return ret;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700790}
791
792static DECLARE_WAIT_QUEUE_HEAD(queue_wait);
793
Trond Myklebust173912a2009-08-09 15:14:29 -0400794static unsigned int cache_poll(struct file *filp, poll_table *wait,
795 struct cache_detail *cd)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700796{
797 unsigned int mask;
798 struct cache_reader *rp = filp->private_data;
799 struct cache_queue *cq;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700800
801 poll_wait(filp, &queue_wait, wait);
802
803 /* alway allow write */
804 mask = POLL_OUT | POLLWRNORM;
805
806 if (!rp)
807 return mask;
808
809 spin_lock(&queue_lock);
810
811 for (cq= &rp->q; &cq->list != &cd->queue;
812 cq = list_entry(cq->list.next, struct cache_queue, list))
813 if (!cq->reader) {
814 mask |= POLLIN | POLLRDNORM;
815 break;
816 }
817 spin_unlock(&queue_lock);
818 return mask;
819}
820
Trond Myklebust173912a2009-08-09 15:14:29 -0400821static int cache_ioctl(struct inode *ino, struct file *filp,
822 unsigned int cmd, unsigned long arg,
823 struct cache_detail *cd)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700824{
825 int len = 0;
826 struct cache_reader *rp = filp->private_data;
827 struct cache_queue *cq;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700828
829 if (cmd != FIONREAD || !rp)
830 return -EINVAL;
831
832 spin_lock(&queue_lock);
833
834 /* only find the length remaining in current request,
835 * or the length of the next request
836 */
837 for (cq= &rp->q; &cq->list != &cd->queue;
838 cq = list_entry(cq->list.next, struct cache_queue, list))
839 if (!cq->reader) {
840 struct cache_request *cr =
841 container_of(cq, struct cache_request, q);
842 len = cr->len - rp->offset;
843 break;
844 }
845 spin_unlock(&queue_lock);
846
847 return put_user(len, (int __user *)arg);
848}
849
Trond Myklebust173912a2009-08-09 15:14:29 -0400850static int cache_open(struct inode *inode, struct file *filp,
851 struct cache_detail *cd)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700852{
853 struct cache_reader *rp = NULL;
854
Trond Myklebustf7e86ab2009-08-19 18:13:00 -0400855 if (!cd || !try_module_get(cd->owner))
856 return -EACCES;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700857 nonseekable_open(inode, filp);
858 if (filp->f_mode & FMODE_READ) {
Linus Torvalds1da177e2005-04-16 15:20:36 -0700859 rp = kmalloc(sizeof(*rp), GFP_KERNEL);
860 if (!rp)
861 return -ENOMEM;
862 rp->offset = 0;
863 rp->q.reader = 1;
864 atomic_inc(&cd->readers);
865 spin_lock(&queue_lock);
866 list_add(&rp->q.list, &cd->queue);
867 spin_unlock(&queue_lock);
868 }
869 filp->private_data = rp;
870 return 0;
871}
872
Trond Myklebust173912a2009-08-09 15:14:29 -0400873static int cache_release(struct inode *inode, struct file *filp,
874 struct cache_detail *cd)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700875{
876 struct cache_reader *rp = filp->private_data;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700877
878 if (rp) {
879 spin_lock(&queue_lock);
880 if (rp->offset) {
881 struct cache_queue *cq;
882 for (cq= &rp->q; &cq->list != &cd->queue;
883 cq = list_entry(cq->list.next, struct cache_queue, list))
884 if (!cq->reader) {
885 container_of(cq, struct cache_request, q)
886 ->readers--;
887 break;
888 }
889 rp->offset = 0;
890 }
891 list_del(&rp->q.list);
892 spin_unlock(&queue_lock);
893
894 filp->private_data = NULL;
895 kfree(rp);
896
897 cd->last_close = get_seconds();
898 atomic_dec(&cd->readers);
899 }
Trond Myklebustf7e86ab2009-08-19 18:13:00 -0400900 module_put(cd->owner);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700901 return 0;
902}
903
904
905
NeilBrownf866a812009-08-04 15:22:38 +1000906static void cache_dequeue(struct cache_detail *detail, struct cache_head *ch)
Linus Torvalds1da177e2005-04-16 15:20:36 -0700907{
908 struct cache_queue *cq;
909 spin_lock(&queue_lock);
910 list_for_each_entry(cq, &detail->queue, list)
911 if (!cq->reader) {
912 struct cache_request *cr = container_of(cq, struct cache_request, q);
913 if (cr->item != ch)
914 continue;
915 if (cr->readers != 0)
NeilBrown4013ede2006-03-27 01:15:07 -0800916 continue;
Linus Torvalds1da177e2005-04-16 15:20:36 -0700917 list_del(&cr->q.list);
918 spin_unlock(&queue_lock);
NeilBrownbaab9352006-03-27 01:15:09 -0800919 cache_put(cr->item, detail);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700920 kfree(cr->buf);
921 kfree(cr);
922 return;
923 }
924 spin_unlock(&queue_lock);
925}
926
927/*
928 * Support routines for text-based upcalls.
929 * Fields are separated by spaces.
930 * Fields are either mangled to quote space tab newline slosh with slosh
931 * or a hexified with a leading \x
932 * Record is terminated with newline.
933 *
934 */
935
936void qword_add(char **bpp, int *lp, char *str)
937{
938 char *bp = *bpp;
939 int len = *lp;
940 char c;
941
942 if (len < 0) return;
943
944 while ((c=*str++) && len)
945 switch(c) {
946 case ' ':
947 case '\t':
948 case '\n':
949 case '\\':
950 if (len >= 4) {
951 *bp++ = '\\';
952 *bp++ = '0' + ((c & 0300)>>6);
953 *bp++ = '0' + ((c & 0070)>>3);
954 *bp++ = '0' + ((c & 0007)>>0);
955 }
956 len -= 4;
957 break;
958 default:
959 *bp++ = c;
960 len--;
961 }
962 if (c || len <1) len = -1;
963 else {
964 *bp++ = ' ';
965 len--;
966 }
967 *bpp = bp;
968 *lp = len;
969}
Trond Myklebust24c37672008-12-23 16:30:12 -0500970EXPORT_SYMBOL_GPL(qword_add);
Linus Torvalds1da177e2005-04-16 15:20:36 -0700971
972void qword_addhex(char **bpp, int *lp, char *buf, int blen)
973{
974 char *bp = *bpp;
975 int len = *lp;
976
977 if (len < 0) return;
978
979 if (len > 2) {
980 *bp++ = '\\';
981 *bp++ = 'x';
982 len -= 2;
983 while (blen && len >= 2) {
984 unsigned char c = *buf++;
985 *bp++ = '0' + ((c&0xf0)>>4) + (c>=0xa0)*('a'-'9'-1);
986 *bp++ = '0' + (c&0x0f) + ((c&0x0f)>=0x0a)*('a'-'9'-1);
987 len -= 2;
988 blen--;
989 }
990 }
991 if (blen || len<1) len = -1;
992 else {
993 *bp++ = ' ';
994 len--;
995 }
996 *bpp = bp;
997 *lp = len;
998}
Trond Myklebust24c37672008-12-23 16:30:12 -0500999EXPORT_SYMBOL_GPL(qword_addhex);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001000
1001static void warn_no_listener(struct cache_detail *detail)
1002{
1003 if (detail->last_warn != detail->last_close) {
1004 detail->last_warn = detail->last_close;
1005 if (detail->warn_no_listener)
Trond Myklebust2da8ca22009-08-09 15:14:26 -04001006 detail->warn_no_listener(detail, detail->last_close != 0);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001007 }
1008}
1009
1010/*
Trond Myklebustbc74b4f2009-08-09 15:14:29 -04001011 * register an upcall request to user-space and queue it up for read() by the
1012 * upcall daemon.
1013 *
Linus Torvalds1da177e2005-04-16 15:20:36 -07001014 * Each request is at most one page long.
1015 */
Trond Myklebustbc74b4f2009-08-09 15:14:29 -04001016int sunrpc_cache_pipe_upcall(struct cache_detail *detail, struct cache_head *h,
1017 void (*cache_request)(struct cache_detail *,
1018 struct cache_head *,
1019 char **,
1020 int *))
Linus Torvalds1da177e2005-04-16 15:20:36 -07001021{
1022
1023 char *buf;
1024 struct cache_request *crq;
1025 char *bp;
1026 int len;
1027
Linus Torvalds1da177e2005-04-16 15:20:36 -07001028 if (atomic_read(&detail->readers) == 0 &&
1029 detail->last_close < get_seconds() - 30) {
1030 warn_no_listener(detail);
1031 return -EINVAL;
1032 }
1033
1034 buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
1035 if (!buf)
1036 return -EAGAIN;
1037
1038 crq = kmalloc(sizeof (*crq), GFP_KERNEL);
1039 if (!crq) {
1040 kfree(buf);
1041 return -EAGAIN;
1042 }
1043
1044 bp = buf; len = PAGE_SIZE;
1045
Trond Myklebustbc74b4f2009-08-09 15:14:29 -04001046 cache_request(detail, h, &bp, &len);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001047
1048 if (len < 0) {
1049 kfree(buf);
1050 kfree(crq);
1051 return -EAGAIN;
1052 }
1053 crq->q.reader = 0;
1054 crq->item = cache_get(h);
1055 crq->buf = buf;
1056 crq->len = PAGE_SIZE - len;
1057 crq->readers = 0;
1058 spin_lock(&queue_lock);
1059 list_add_tail(&crq->q.list, &detail->queue);
1060 spin_unlock(&queue_lock);
1061 wake_up(&queue_wait);
1062 return 0;
1063}
Trond Myklebustbc74b4f2009-08-09 15:14:29 -04001064EXPORT_SYMBOL_GPL(sunrpc_cache_pipe_upcall);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001065
1066/*
1067 * parse a message from user-space and pass it
1068 * to an appropriate cache
1069 * Messages are, like requests, separated into fields by
1070 * spaces and dequotes as \xHEXSTRING or embedded \nnn octal
1071 *
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -08001072 * Message is
Linus Torvalds1da177e2005-04-16 15:20:36 -07001073 * reply cachename expiry key ... content....
1074 *
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -08001075 * key and content are both parsed by cache
Linus Torvalds1da177e2005-04-16 15:20:36 -07001076 */
1077
1078#define isodigit(c) (isdigit(c) && c <= '7')
1079int qword_get(char **bpp, char *dest, int bufsize)
1080{
1081 /* return bytes copied, or -1 on error */
1082 char *bp = *bpp;
1083 int len = 0;
1084
1085 while (*bp == ' ') bp++;
1086
1087 if (bp[0] == '\\' && bp[1] == 'x') {
1088 /* HEX STRING */
1089 bp += 2;
1090 while (isxdigit(bp[0]) && isxdigit(bp[1]) && len < bufsize) {
1091 int byte = isdigit(*bp) ? *bp-'0' : toupper(*bp)-'A'+10;
1092 bp++;
1093 byte <<= 4;
1094 byte |= isdigit(*bp) ? *bp-'0' : toupper(*bp)-'A'+10;
1095 *dest++ = byte;
1096 bp++;
1097 len++;
1098 }
1099 } else {
1100 /* text with \nnn octal quoting */
1101 while (*bp != ' ' && *bp != '\n' && *bp && len < bufsize-1) {
1102 if (*bp == '\\' &&
1103 isodigit(bp[1]) && (bp[1] <= '3') &&
1104 isodigit(bp[2]) &&
1105 isodigit(bp[3])) {
1106 int byte = (*++bp -'0');
1107 bp++;
1108 byte = (byte << 3) | (*bp++ - '0');
1109 byte = (byte << 3) | (*bp++ - '0');
1110 *dest++ = byte;
1111 len++;
1112 } else {
1113 *dest++ = *bp++;
1114 len++;
1115 }
1116 }
1117 }
1118
1119 if (*bp != ' ' && *bp != '\n' && *bp != '\0')
1120 return -1;
1121 while (*bp == ' ') bp++;
1122 *bpp = bp;
1123 *dest = '\0';
1124 return len;
1125}
Trond Myklebust24c37672008-12-23 16:30:12 -05001126EXPORT_SYMBOL_GPL(qword_get);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001127
1128
1129/*
1130 * support /proc/sunrpc/cache/$CACHENAME/content
1131 * as a seqfile.
1132 * We call ->cache_show passing NULL for the item to
1133 * get a header, then pass each real item in the cache
1134 */
1135
1136struct handle {
1137 struct cache_detail *cd;
1138};
1139
1140static void *c_start(struct seq_file *m, loff_t *pos)
Eric Dumazet9a429c42008-01-01 21:58:02 -08001141 __acquires(cd->hash_lock)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001142{
1143 loff_t n = *pos;
1144 unsigned hash, entry;
1145 struct cache_head *ch;
1146 struct cache_detail *cd = ((struct handle*)m->private)->cd;
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -08001147
Linus Torvalds1da177e2005-04-16 15:20:36 -07001148
1149 read_lock(&cd->hash_lock);
1150 if (!n--)
1151 return SEQ_START_TOKEN;
1152 hash = n >> 32;
1153 entry = n & ((1LL<<32) - 1);
1154
1155 for (ch=cd->hash_table[hash]; ch; ch=ch->next)
1156 if (!entry--)
1157 return ch;
1158 n &= ~((1LL<<32) - 1);
1159 do {
1160 hash++;
1161 n += 1LL<<32;
YOSHIFUJI Hideakicca51722007-02-09 15:38:13 -08001162 } while(hash < cd->hash_size &&
Linus Torvalds1da177e2005-04-16 15:20:36 -07001163 cd->hash_table[hash]==NULL);
1164 if (hash >= cd->hash_size)
1165 return NULL;
1166 *pos = n+1;
1167 return cd->hash_table[hash];
1168}
1169
1170static void *c_next(struct seq_file *m, void *p, loff_t *pos)
1171{
1172 struct cache_head *ch = p;
1173 int hash = (*pos >> 32);
1174 struct cache_detail *cd = ((struct handle*)m->private)->cd;
1175
1176 if (p == SEQ_START_TOKEN)
1177 hash = 0;
1178 else if (ch->next == NULL) {
1179 hash++;
1180 *pos += 1LL<<32;
1181 } else {
1182 ++*pos;
1183 return ch->next;
1184 }
1185 *pos &= ~((1LL<<32) - 1);
1186 while (hash < cd->hash_size &&
1187 cd->hash_table[hash] == NULL) {
1188 hash++;
1189 *pos += 1LL<<32;
1190 }
1191 if (hash >= cd->hash_size)
1192 return NULL;
1193 ++*pos;
1194 return cd->hash_table[hash];
1195}
1196
1197static void c_stop(struct seq_file *m, void *p)
Eric Dumazet9a429c42008-01-01 21:58:02 -08001198 __releases(cd->hash_lock)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001199{
1200 struct cache_detail *cd = ((struct handle*)m->private)->cd;
1201 read_unlock(&cd->hash_lock);
1202}
1203
1204static int c_show(struct seq_file *m, void *p)
1205{
1206 struct cache_head *cp = p;
1207 struct cache_detail *cd = ((struct handle*)m->private)->cd;
1208
1209 if (p == SEQ_START_TOKEN)
1210 return cd->cache_show(m, cd, NULL);
1211
1212 ifdebug(CACHE)
NeilBrown4013ede2006-03-27 01:15:07 -08001213 seq_printf(m, "# expiry=%ld refcnt=%d flags=%lx\n",
NeilBrownbaab9352006-03-27 01:15:09 -08001214 cp->expiry_time, atomic_read(&cp->ref.refcount), cp->flags);
Linus Torvalds1da177e2005-04-16 15:20:36 -07001215 cache_get(cp);
1216 if (cache_check(cd, cp, NULL))
1217 /* cache_check does a cache_put on failure */
1218 seq_printf(m, "# ");
1219 else
1220 cache_put(cp, cd);
1221
1222 return cd->cache_show(m, cd, cp);
1223}
1224
Philippe De Muyter56b3d972007-07-10 23:07:31 -07001225static const struct seq_operations cache_content_op = {
Linus Torvalds1da177e2005-04-16 15:20:36 -07001226 .start = c_start,
1227 .next = c_next,
1228 .stop = c_stop,
1229 .show = c_show,
1230};
1231
Trond Myklebust173912a2009-08-09 15:14:29 -04001232static int content_open(struct inode *inode, struct file *file,
1233 struct cache_detail *cd)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001234{
Linus Torvalds1da177e2005-04-16 15:20:36 -07001235 struct handle *han;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001236
Trond Myklebustf7e86ab2009-08-19 18:13:00 -04001237 if (!cd || !try_module_get(cd->owner))
1238 return -EACCES;
Pavel Emelyanovec931032007-10-10 02:31:07 -07001239 han = __seq_open_private(file, &cache_content_op, sizeof(*han));
Linus Torvalds1da177e2005-04-16 15:20:36 -07001240 if (han == NULL)
1241 return -ENOMEM;
1242
1243 han->cd = cd;
Pavel Emelyanovec931032007-10-10 02:31:07 -07001244 return 0;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001245}
Linus Torvalds1da177e2005-04-16 15:20:36 -07001246
Trond Myklebustf7e86ab2009-08-19 18:13:00 -04001247static int content_release(struct inode *inode, struct file *file,
1248 struct cache_detail *cd)
1249{
1250 int ret = seq_release_private(inode, file);
1251 module_put(cd->owner);
1252 return ret;
1253}
1254
1255static int open_flush(struct inode *inode, struct file *file,
1256 struct cache_detail *cd)
1257{
1258 if (!cd || !try_module_get(cd->owner))
1259 return -EACCES;
1260 return nonseekable_open(inode, file);
1261}
1262
1263static int release_flush(struct inode *inode, struct file *file,
1264 struct cache_detail *cd)
1265{
1266 module_put(cd->owner);
1267 return 0;
1268}
Linus Torvalds1da177e2005-04-16 15:20:36 -07001269
1270static ssize_t read_flush(struct file *file, char __user *buf,
Trond Myklebust173912a2009-08-09 15:14:29 -04001271 size_t count, loff_t *ppos,
1272 struct cache_detail *cd)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001273{
Linus Torvalds1da177e2005-04-16 15:20:36 -07001274 char tbuf[20];
1275 unsigned long p = *ppos;
Chuck Lever01b29692007-10-26 13:31:20 -04001276 size_t len;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001277
1278 sprintf(tbuf, "%lu\n", cd->flush_time);
1279 len = strlen(tbuf);
1280 if (p >= len)
1281 return 0;
1282 len -= p;
Chuck Lever01b29692007-10-26 13:31:20 -04001283 if (len > count)
1284 len = count;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001285 if (copy_to_user(buf, (void*)(tbuf+p), len))
Chuck Lever01b29692007-10-26 13:31:20 -04001286 return -EFAULT;
1287 *ppos += len;
Linus Torvalds1da177e2005-04-16 15:20:36 -07001288 return len;
1289}
1290
Trond Myklebust173912a2009-08-09 15:14:29 -04001291static ssize_t write_flush(struct file *file, const char __user *buf,
1292 size_t count, loff_t *ppos,
1293 struct cache_detail *cd)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001294{
Linus Torvalds1da177e2005-04-16 15:20:36 -07001295 char tbuf[20];
1296 char *ep;
1297 long flushtime;
1298 if (*ppos || count > sizeof(tbuf)-1)
1299 return -EINVAL;
1300 if (copy_from_user(tbuf, buf, count))
1301 return -EFAULT;
1302 tbuf[count] = 0;
1303 flushtime = simple_strtoul(tbuf, &ep, 0);
1304 if (*ep && *ep != '\n')
1305 return -EINVAL;
1306
1307 cd->flush_time = flushtime;
1308 cd->nextcheck = get_seconds();
1309 cache_flush();
1310
1311 *ppos += count;
1312 return count;
1313}
1314
Trond Myklebust173912a2009-08-09 15:14:29 -04001315static ssize_t cache_read_procfs(struct file *filp, char __user *buf,
1316 size_t count, loff_t *ppos)
1317{
1318 struct cache_detail *cd = PDE(filp->f_path.dentry->d_inode)->data;
1319
1320 return cache_read(filp, buf, count, ppos, cd);
1321}
1322
1323static ssize_t cache_write_procfs(struct file *filp, const char __user *buf,
1324 size_t count, loff_t *ppos)
1325{
1326 struct cache_detail *cd = PDE(filp->f_path.dentry->d_inode)->data;
1327
1328 return cache_write(filp, buf, count, ppos, cd);
1329}
1330
1331static unsigned int cache_poll_procfs(struct file *filp, poll_table *wait)
1332{
1333 struct cache_detail *cd = PDE(filp->f_path.dentry->d_inode)->data;
1334
1335 return cache_poll(filp, wait, cd);
1336}
1337
1338static int cache_ioctl_procfs(struct inode *inode, struct file *filp,
1339 unsigned int cmd, unsigned long arg)
1340{
1341 struct cache_detail *cd = PDE(inode)->data;
1342
1343 return cache_ioctl(inode, filp, cmd, arg, cd);
1344}
1345
1346static int cache_open_procfs(struct inode *inode, struct file *filp)
1347{
1348 struct cache_detail *cd = PDE(inode)->data;
1349
1350 return cache_open(inode, filp, cd);
1351}
1352
1353static int cache_release_procfs(struct inode *inode, struct file *filp)
1354{
1355 struct cache_detail *cd = PDE(inode)->data;
1356
1357 return cache_release(inode, filp, cd);
1358}
1359
1360static const struct file_operations cache_file_operations_procfs = {
1361 .owner = THIS_MODULE,
1362 .llseek = no_llseek,
1363 .read = cache_read_procfs,
1364 .write = cache_write_procfs,
1365 .poll = cache_poll_procfs,
1366 .ioctl = cache_ioctl_procfs, /* for FIONREAD */
1367 .open = cache_open_procfs,
1368 .release = cache_release_procfs,
Linus Torvalds1da177e2005-04-16 15:20:36 -07001369};
Trond Myklebust173912a2009-08-09 15:14:29 -04001370
1371static int content_open_procfs(struct inode *inode, struct file *filp)
1372{
1373 struct cache_detail *cd = PDE(inode)->data;
1374
1375 return content_open(inode, filp, cd);
1376}
1377
Trond Myklebustf7e86ab2009-08-19 18:13:00 -04001378static int content_release_procfs(struct inode *inode, struct file *filp)
1379{
1380 struct cache_detail *cd = PDE(inode)->data;
1381
1382 return content_release(inode, filp, cd);
1383}
1384
Trond Myklebust173912a2009-08-09 15:14:29 -04001385static const struct file_operations content_file_operations_procfs = {
1386 .open = content_open_procfs,
1387 .read = seq_read,
1388 .llseek = seq_lseek,
Trond Myklebustf7e86ab2009-08-19 18:13:00 -04001389 .release = content_release_procfs,
Trond Myklebust173912a2009-08-09 15:14:29 -04001390};
1391
Trond Myklebustf7e86ab2009-08-19 18:13:00 -04001392static int open_flush_procfs(struct inode *inode, struct file *filp)
1393{
1394 struct cache_detail *cd = PDE(inode)->data;
1395
1396 return open_flush(inode, filp, cd);
1397}
1398
1399static int release_flush_procfs(struct inode *inode, struct file *filp)
1400{
1401 struct cache_detail *cd = PDE(inode)->data;
1402
1403 return release_flush(inode, filp, cd);
1404}
1405
Trond Myklebust173912a2009-08-09 15:14:29 -04001406static ssize_t read_flush_procfs(struct file *filp, char __user *buf,
1407 size_t count, loff_t *ppos)
1408{
1409 struct cache_detail *cd = PDE(filp->f_path.dentry->d_inode)->data;
1410
1411 return read_flush(filp, buf, count, ppos, cd);
1412}
1413
1414static ssize_t write_flush_procfs(struct file *filp,
1415 const char __user *buf,
1416 size_t count, loff_t *ppos)
1417{
1418 struct cache_detail *cd = PDE(filp->f_path.dentry->d_inode)->data;
1419
1420 return write_flush(filp, buf, count, ppos, cd);
1421}
1422
1423static const struct file_operations cache_flush_operations_procfs = {
Trond Myklebustf7e86ab2009-08-19 18:13:00 -04001424 .open = open_flush_procfs,
Trond Myklebust173912a2009-08-09 15:14:29 -04001425 .read = read_flush_procfs,
1426 .write = write_flush_procfs,
Trond Myklebustf7e86ab2009-08-19 18:13:00 -04001427 .release = release_flush_procfs,
Trond Myklebust173912a2009-08-09 15:14:29 -04001428};
1429
1430static void remove_cache_proc_entries(struct cache_detail *cd)
1431{
1432 if (cd->u.procfs.proc_ent == NULL)
1433 return;
1434 if (cd->u.procfs.flush_ent)
1435 remove_proc_entry("flush", cd->u.procfs.proc_ent);
1436 if (cd->u.procfs.channel_ent)
1437 remove_proc_entry("channel", cd->u.procfs.proc_ent);
1438 if (cd->u.procfs.content_ent)
1439 remove_proc_entry("content", cd->u.procfs.proc_ent);
1440 cd->u.procfs.proc_ent = NULL;
1441 remove_proc_entry(cd->name, proc_net_rpc);
1442}
1443
1444#ifdef CONFIG_PROC_FS
1445static int create_cache_proc_entries(struct cache_detail *cd)
1446{
1447 struct proc_dir_entry *p;
1448
1449 cd->u.procfs.proc_ent = proc_mkdir(cd->name, proc_net_rpc);
1450 if (cd->u.procfs.proc_ent == NULL)
1451 goto out_nomem;
1452 cd->u.procfs.channel_ent = NULL;
1453 cd->u.procfs.content_ent = NULL;
1454
1455 p = proc_create_data("flush", S_IFREG|S_IRUSR|S_IWUSR,
1456 cd->u.procfs.proc_ent,
1457 &cache_flush_operations_procfs, cd);
1458 cd->u.procfs.flush_ent = p;
1459 if (p == NULL)
1460 goto out_nomem;
1461
1462 if (cd->cache_upcall || cd->cache_parse) {
1463 p = proc_create_data("channel", S_IFREG|S_IRUSR|S_IWUSR,
1464 cd->u.procfs.proc_ent,
1465 &cache_file_operations_procfs, cd);
1466 cd->u.procfs.channel_ent = p;
1467 if (p == NULL)
1468 goto out_nomem;
1469 }
1470 if (cd->cache_show) {
1471 p = proc_create_data("content", S_IFREG|S_IRUSR|S_IWUSR,
1472 cd->u.procfs.proc_ent,
1473 &content_file_operations_procfs, cd);
1474 cd->u.procfs.content_ent = p;
1475 if (p == NULL)
1476 goto out_nomem;
1477 }
1478 return 0;
1479out_nomem:
1480 remove_cache_proc_entries(cd);
1481 return -ENOMEM;
1482}
1483#else /* CONFIG_PROC_FS */
1484static int create_cache_proc_entries(struct cache_detail *cd)
1485{
1486 return 0;
1487}
1488#endif
1489
1490int cache_register(struct cache_detail *cd)
1491{
1492 int ret;
1493
1494 sunrpc_init_cache_detail(cd);
1495 ret = create_cache_proc_entries(cd);
1496 if (ret)
1497 sunrpc_destroy_cache_detail(cd);
1498 return ret;
1499}
1500EXPORT_SYMBOL_GPL(cache_register);
1501
1502void cache_unregister(struct cache_detail *cd)
1503{
1504 remove_cache_proc_entries(cd);
1505 sunrpc_destroy_cache_detail(cd);
1506}
1507EXPORT_SYMBOL_GPL(cache_unregister);
Trond Myklebust8854e822009-08-09 15:14:30 -04001508
1509static ssize_t cache_read_pipefs(struct file *filp, char __user *buf,
1510 size_t count, loff_t *ppos)
1511{
1512 struct cache_detail *cd = RPC_I(filp->f_path.dentry->d_inode)->private;
1513
1514 return cache_read(filp, buf, count, ppos, cd);
1515}
1516
1517static ssize_t cache_write_pipefs(struct file *filp, const char __user *buf,
1518 size_t count, loff_t *ppos)
1519{
1520 struct cache_detail *cd = RPC_I(filp->f_path.dentry->d_inode)->private;
1521
1522 return cache_write(filp, buf, count, ppos, cd);
1523}
1524
1525static unsigned int cache_poll_pipefs(struct file *filp, poll_table *wait)
1526{
1527 struct cache_detail *cd = RPC_I(filp->f_path.dentry->d_inode)->private;
1528
1529 return cache_poll(filp, wait, cd);
1530}
1531
1532static int cache_ioctl_pipefs(struct inode *inode, struct file *filp,
1533 unsigned int cmd, unsigned long arg)
1534{
1535 struct cache_detail *cd = RPC_I(inode)->private;
1536
1537 return cache_ioctl(inode, filp, cmd, arg, cd);
1538}
1539
1540static int cache_open_pipefs(struct inode *inode, struct file *filp)
1541{
1542 struct cache_detail *cd = RPC_I(inode)->private;
1543
1544 return cache_open(inode, filp, cd);
1545}
1546
1547static int cache_release_pipefs(struct inode *inode, struct file *filp)
1548{
1549 struct cache_detail *cd = RPC_I(inode)->private;
1550
1551 return cache_release(inode, filp, cd);
1552}
1553
1554const struct file_operations cache_file_operations_pipefs = {
1555 .owner = THIS_MODULE,
1556 .llseek = no_llseek,
1557 .read = cache_read_pipefs,
1558 .write = cache_write_pipefs,
1559 .poll = cache_poll_pipefs,
1560 .ioctl = cache_ioctl_pipefs, /* for FIONREAD */
1561 .open = cache_open_pipefs,
1562 .release = cache_release_pipefs,
1563};
1564
1565static int content_open_pipefs(struct inode *inode, struct file *filp)
1566{
1567 struct cache_detail *cd = RPC_I(inode)->private;
1568
1569 return content_open(inode, filp, cd);
1570}
1571
Trond Myklebustf7e86ab2009-08-19 18:13:00 -04001572static int content_release_pipefs(struct inode *inode, struct file *filp)
1573{
1574 struct cache_detail *cd = RPC_I(inode)->private;
1575
1576 return content_release(inode, filp, cd);
1577}
1578
Trond Myklebust8854e822009-08-09 15:14:30 -04001579const struct file_operations content_file_operations_pipefs = {
1580 .open = content_open_pipefs,
1581 .read = seq_read,
1582 .llseek = seq_lseek,
Trond Myklebustf7e86ab2009-08-19 18:13:00 -04001583 .release = content_release_pipefs,
Trond Myklebust8854e822009-08-09 15:14:30 -04001584};
1585
Trond Myklebustf7e86ab2009-08-19 18:13:00 -04001586static int open_flush_pipefs(struct inode *inode, struct file *filp)
1587{
1588 struct cache_detail *cd = RPC_I(inode)->private;
1589
1590 return open_flush(inode, filp, cd);
1591}
1592
1593static int release_flush_pipefs(struct inode *inode, struct file *filp)
1594{
1595 struct cache_detail *cd = RPC_I(inode)->private;
1596
1597 return release_flush(inode, filp, cd);
1598}
1599
Trond Myklebust8854e822009-08-09 15:14:30 -04001600static ssize_t read_flush_pipefs(struct file *filp, char __user *buf,
1601 size_t count, loff_t *ppos)
1602{
1603 struct cache_detail *cd = RPC_I(filp->f_path.dentry->d_inode)->private;
1604
1605 return read_flush(filp, buf, count, ppos, cd);
1606}
1607
1608static ssize_t write_flush_pipefs(struct file *filp,
1609 const char __user *buf,
1610 size_t count, loff_t *ppos)
1611{
1612 struct cache_detail *cd = RPC_I(filp->f_path.dentry->d_inode)->private;
1613
1614 return write_flush(filp, buf, count, ppos, cd);
1615}
1616
1617const struct file_operations cache_flush_operations_pipefs = {
Trond Myklebustf7e86ab2009-08-19 18:13:00 -04001618 .open = open_flush_pipefs,
Trond Myklebust8854e822009-08-09 15:14:30 -04001619 .read = read_flush_pipefs,
1620 .write = write_flush_pipefs,
Trond Myklebustf7e86ab2009-08-19 18:13:00 -04001621 .release = release_flush_pipefs,
Trond Myklebust8854e822009-08-09 15:14:30 -04001622};
1623
1624int sunrpc_cache_register_pipefs(struct dentry *parent,
1625 const char *name, mode_t umode,
1626 struct cache_detail *cd)
1627{
1628 struct qstr q;
1629 struct dentry *dir;
1630 int ret = 0;
1631
1632 sunrpc_init_cache_detail(cd);
1633 q.name = name;
1634 q.len = strlen(name);
1635 q.hash = full_name_hash(q.name, q.len);
1636 dir = rpc_create_cache_dir(parent, &q, umode, cd);
1637 if (!IS_ERR(dir))
1638 cd->u.pipefs.dir = dir;
1639 else {
1640 sunrpc_destroy_cache_detail(cd);
1641 ret = PTR_ERR(dir);
1642 }
1643 return ret;
1644}
1645EXPORT_SYMBOL_GPL(sunrpc_cache_register_pipefs);
1646
1647void sunrpc_cache_unregister_pipefs(struct cache_detail *cd)
1648{
1649 rpc_remove_cache_dir(cd->u.pipefs.dir);
1650 cd->u.pipefs.dir = NULL;
1651 sunrpc_destroy_cache_detail(cd);
1652}
1653EXPORT_SYMBOL_GPL(sunrpc_cache_unregister_pipefs);
1654