cache.c 41.9 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
/*
 * net/sunrpc/cache.c
 *
 * Generic code for various authentication-related caches
 * used by sunrpc clients and servers.
 *
 * Copyright (C) 2002 Neil Brown <neilb@cse.unsw.edu.au>
 *
 * Released under terms in GPL version 2.  See COPYING.
 *
 */

#include <linux/types.h>
#include <linux/fs.h>
#include <linux/file.h>
#include <linux/slab.h>
#include <linux/signal.h>
#include <linux/sched.h>
#include <linux/kmod.h>
#include <linux/list.h>
#include <linux/module.h>
#include <linux/ctype.h>
#include <asm/uaccess.h>
#include <linux/poll.h>
#include <linux/seq_file.h>
#include <linux/proc_fs.h>
#include <linux/net.h>
#include <linux/workqueue.h>
Arjan van de Ven's avatar
Arjan van de Ven committed
29
#include <linux/mutex.h>
30
#include <linux/pagemap.h>
31
#include <linux/smp_lock.h>
Linus Torvalds's avatar
Linus Torvalds committed
32 33 34 35
#include <asm/ioctls.h>
#include <linux/sunrpc/types.h>
#include <linux/sunrpc/cache.h>
#include <linux/sunrpc/stats.h>
36
#include <linux/sunrpc/rpc_pipe_fs.h>
Linus Torvalds's avatar
Linus Torvalds committed
37 38 39

#define	 RPCDBG_FACILITY RPCDBG_CACHE

40
static int cache_defer_req(struct cache_req *req, struct cache_head *item);
Linus Torvalds's avatar
Linus Torvalds committed
41 42
static void cache_revisit_request(struct cache_head *item);

43
static void cache_init(struct cache_head *h)
Linus Torvalds's avatar
Linus Torvalds committed
44
{
45
	time_t now = seconds_since_boot();
Linus Torvalds's avatar
Linus Torvalds committed
46 47
	h->next = NULL;
	h->flags = 0;
48
	kref_init(&h->ref);
Linus Torvalds's avatar
Linus Torvalds committed
49 50 51 52
	h->expiry_time = now + CACHE_NEW_EXPIRY;
	h->last_refresh = now;
}

53 54
static inline int cache_is_expired(struct cache_detail *detail, struct cache_head *h)
{
55
	return  (h->expiry_time < seconds_since_boot()) ||
56 57 58
		(detail->flush_time > h->last_refresh);
}

59 60 61 62
struct cache_head *sunrpc_cache_lookup(struct cache_detail *detail,
				       struct cache_head *key, int hash)
{
	struct cache_head **head,  **hp;
63
	struct cache_head *new = NULL, *freeme = NULL;
64 65 66 67 68 69 70 71

	head = &detail->hash_table[hash];

	read_lock(&detail->hash_lock);

	for (hp=head; *hp != NULL ; hp = &(*hp)->next) {
		struct cache_head *tmp = *hp;
		if (detail->match(tmp, key)) {
72 73 74
			if (cache_is_expired(detail, tmp))
				/* This entry is expired, we will discard it. */
				break;
75 76 77 78 79 80 81 82 83 84 85
			cache_get(tmp);
			read_unlock(&detail->hash_lock);
			return tmp;
		}
	}
	read_unlock(&detail->hash_lock);
	/* Didn't find anything, insert an empty entry */

	new = detail->alloc();
	if (!new)
		return NULL;
86 87 88 89
	/* must fully initialise 'new', else
	 * we might get lose if we need to
	 * cache_put it soon.
	 */
90
	cache_init(new);
91
	detail->init(new, key);
92 93 94 95 96 97 98

	write_lock(&detail->hash_lock);

	/* check if entry appeared while we slept */
	for (hp=head; *hp != NULL ; hp = &(*hp)->next) {
		struct cache_head *tmp = *hp;
		if (detail->match(tmp, key)) {
99 100 101 102 103 104 105
			if (cache_is_expired(detail, tmp)) {
				*hp = tmp->next;
				tmp->next = NULL;
				detail->entries --;
				freeme = tmp;
				break;
			}
106 107
			cache_get(tmp);
			write_unlock(&detail->hash_lock);
108
			cache_put(new, detail);
109 110 111 112 113 114 115 116 117
			return tmp;
		}
	}
	new->next = *head;
	*head = new;
	detail->entries++;
	cache_get(new);
	write_unlock(&detail->hash_lock);

118 119
	if (freeme)
		cache_put(freeme, detail);
120 121
	return new;
}
122
EXPORT_SYMBOL_GPL(sunrpc_cache_lookup);
123

124

125
static void cache_dequeue(struct cache_detail *detail, struct cache_head *ch);
126

127
static void cache_fresh_locked(struct cache_head *head, time_t expiry)
128 129
{
	head->expiry_time = expiry;
130
	head->last_refresh = seconds_since_boot();
131
	set_bit(CACHE_VALID, &head->flags);
132 133 134
}

static void cache_fresh_unlocked(struct cache_head *head,
135
				 struct cache_detail *detail)
136 137 138
{
	if (test_and_clear_bit(CACHE_PENDING, &head->flags)) {
		cache_revisit_request(head);
139
		cache_dequeue(detail, head);
140 141 142
	}
}

143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159
struct cache_head *sunrpc_cache_update(struct cache_detail *detail,
				       struct cache_head *new, struct cache_head *old, int hash)
{
	/* The 'old' entry is to be replaced by 'new'.
	 * If 'old' is not VALID, we update it directly,
	 * otherwise we need to replace it
	 */
	struct cache_head **head;
	struct cache_head *tmp;

	if (!test_bit(CACHE_VALID, &old->flags)) {
		write_lock(&detail->hash_lock);
		if (!test_bit(CACHE_VALID, &old->flags)) {
			if (test_bit(CACHE_NEGATIVE, &new->flags))
				set_bit(CACHE_NEGATIVE, &old->flags);
			else
				detail->update(old, new);
160
			cache_fresh_locked(old, new->expiry_time);
161
			write_unlock(&detail->hash_lock);
162
			cache_fresh_unlocked(old, detail);
163 164 165 166 167 168 169
			return old;
		}
		write_unlock(&detail->hash_lock);
	}
	/* We need to insert a new entry */
	tmp = detail->alloc();
	if (!tmp) {
170
		cache_put(old, detail);
171 172 173 174 175 176 177 178 179 180 181 182 183
		return NULL;
	}
	cache_init(tmp);
	detail->init(tmp, old);
	head = &detail->hash_table[hash];

	write_lock(&detail->hash_lock);
	if (test_bit(CACHE_NEGATIVE, &new->flags))
		set_bit(CACHE_NEGATIVE, &tmp->flags);
	else
		detail->update(tmp, new);
	tmp->next = *head;
	*head = tmp;
184
	detail->entries++;
185
	cache_get(tmp);
186
	cache_fresh_locked(tmp, new->expiry_time);
187
	cache_fresh_locked(old, 0);
188
	write_unlock(&detail->hash_lock);
189 190
	cache_fresh_unlocked(tmp, detail);
	cache_fresh_unlocked(old, detail);
191
	cache_put(old, detail);
192 193
	return tmp;
}
194
EXPORT_SYMBOL_GPL(sunrpc_cache_update);
Linus Torvalds's avatar
Linus Torvalds committed
195

196 197 198 199 200 201
static int cache_make_upcall(struct cache_detail *cd, struct cache_head *h)
{
	if (!cd->cache_upcall)
		return -EINVAL;
	return cd->cache_upcall(cd, h);
}
202 203 204

static inline int cache_is_valid(struct cache_detail *detail, struct cache_head *h)
{
205
	if (!test_bit(CACHE_VALID, &h->flags))
206 207 208 209 210 211 212 213 214
		return -EAGAIN;
	else {
		/* entry is valid */
		if (test_bit(CACHE_NEGATIVE, &h->flags))
			return -ENOENT;
		else
			return 0;
	}
}
215

Linus Torvalds's avatar
Linus Torvalds committed
216 217 218 219 220 221 222 223
/*
 * This is the generic cache management routine for all
 * the authentication caches.
 * It checks the currency of a cache item and will (later)
 * initiate an upcall to fill it if needed.
 *
 *
 * Returns 0 if the cache_head can be used, or cache_puts it and returns
224 225 226 227
 * -EAGAIN if upcall is pending and request has been queued
 * -ETIMEDOUT if upcall failed or request could not be queue or
 *           upcall completed but item is still invalid (implying that
 *           the cache item has been replaced with a newer one).
Linus Torvalds's avatar
Linus Torvalds committed
228 229 230 231 232 233 234 235 236
 * -ENOENT if cache entry was negative
 */
int cache_check(struct cache_detail *detail,
		    struct cache_head *h, struct cache_req *rqstp)
{
	int rv;
	long refresh_age, age;

	/* First decide return status as best we can */
237
	rv = cache_is_valid(detail, h);
Linus Torvalds's avatar
Linus Torvalds committed
238 239 240

	/* now see if we want to start an upcall */
	refresh_age = (h->expiry_time - h->last_refresh);
241
	age = seconds_since_boot() - h->last_refresh;
Linus Torvalds's avatar
Linus Torvalds committed
242 243 244 245 246

	if (rqstp == NULL) {
		if (rv == -EAGAIN)
			rv = -ENOENT;
	} else if (rv == -EAGAIN || age > refresh_age/2) {
247 248
		dprintk("RPC:       Want update, refage=%ld, age=%ld\n",
				refresh_age, age);
Linus Torvalds's avatar
Linus Torvalds committed
249 250 251 252
		if (!test_and_set_bit(CACHE_PENDING, &h->flags)) {
			switch (cache_make_upcall(detail, h)) {
			case -EINVAL:
				clear_bit(CACHE_PENDING, &h->flags);
253
				cache_revisit_request(h);
Linus Torvalds's avatar
Linus Torvalds committed
254 255
				if (rv == -EAGAIN) {
					set_bit(CACHE_NEGATIVE, &h->flags);
256
					cache_fresh_locked(h, seconds_since_boot()+CACHE_NEW_EXPIRY);
257
					cache_fresh_unlocked(h, detail);
Linus Torvalds's avatar
Linus Torvalds committed
258 259 260 261 262 263 264 265 266 267 268 269
					rv = -ENOENT;
				}
				break;

			case -EAGAIN:
				clear_bit(CACHE_PENDING, &h->flags);
				cache_revisit_request(h);
				break;
			}
		}
	}

270
	if (rv == -EAGAIN) {
271
		if (cache_defer_req(rqstp, h) < 0) {
272 273 274 275 276 277
			/* Request is not deferred */
			rv = cache_is_valid(detail, h);
			if (rv == -EAGAIN)
				rv = -ETIMEDOUT;
		}
	}
278
	if (rv)
279
		cache_put(h, detail);
Linus Torvalds's avatar
Linus Torvalds committed
280 281
	return rv;
}
282
EXPORT_SYMBOL_GPL(cache_check);
Linus Torvalds's avatar
Linus Torvalds committed
283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312

/*
 * caches need to be periodically cleaned.
 * For this we maintain a list of cache_detail and
 * a current pointer into that list and into the table
 * for that entry.
 *
 * Each time clean_cache is called it finds the next non-empty entry
 * in the current table and walks the list in that entry
 * looking for entries that can be removed.
 *
 * An entry gets removed if:
 * - The expiry is before current time
 * - The last_refresh time is before the flush_time for that cache
 *
 * later we might drop old entries with non-NEVER expiry if that table
 * is getting 'full' for some definition of 'full'
 *
 * The question of "how often to scan a table" is an interesting one
 * and is answered in part by the use of the "nextcheck" field in the
 * cache_detail.
 * When a scan of a table begins, the nextcheck field is set to a time
 * that is well into the future.
 * While scanning, if an expiry time is found that is earlier than the
 * current nextcheck time, nextcheck is set to that expiry time.
 * If the flush_time is ever set to a time earlier than the nextcheck
 * time, the nextcheck time is then set to that flush_time.
 *
 * A table is then only scanned if the current time is at least
 * the nextcheck time.
313
 *
Linus Torvalds's avatar
Linus Torvalds committed
314 315 316 317 318 319 320
 */

static LIST_HEAD(cache_list);
static DEFINE_SPINLOCK(cache_list_lock);
static struct cache_detail *current_detail;
static int current_index;

321
static void do_cache_clean(struct work_struct *work);
322
static struct delayed_work cache_cleaner;
Linus Torvalds's avatar
Linus Torvalds committed
323

324
static void sunrpc_init_cache_detail(struct cache_detail *cd)
325
{
Linus Torvalds's avatar
Linus Torvalds committed
326 327 328 329 330 331 332 333 334 335 336 337
	rwlock_init(&cd->hash_lock);
	INIT_LIST_HEAD(&cd->queue);
	spin_lock(&cache_list_lock);
	cd->nextcheck = 0;
	cd->entries = 0;
	atomic_set(&cd->readers, 0);
	cd->last_close = 0;
	cd->last_warn = -1;
	list_add(&cd->others, &cache_list);
	spin_unlock(&cache_list_lock);

	/* start the cleaning process */
338
	schedule_delayed_work(&cache_cleaner, 0);
Linus Torvalds's avatar
Linus Torvalds committed
339 340
}

341
static void sunrpc_destroy_cache_detail(struct cache_detail *cd)
Linus Torvalds's avatar
Linus Torvalds committed
342 343 344 345 346 347 348
{
	cache_purge(cd);
	spin_lock(&cache_list_lock);
	write_lock(&cd->hash_lock);
	if (cd->entries || atomic_read(&cd->inuse)) {
		write_unlock(&cd->hash_lock);
		spin_unlock(&cache_list_lock);
349
		goto out;
Linus Torvalds's avatar
Linus Torvalds committed
350 351 352 353 354 355 356 357
	}
	if (current_detail == cd)
		current_detail = NULL;
	list_del_init(&cd->others);
	write_unlock(&cd->hash_lock);
	spin_unlock(&cache_list_lock);
	if (list_empty(&cache_list)) {
		/* module must be being unloaded so its safe to kill the worker */
358
		cancel_delayed_work_sync(&cache_cleaner);
Linus Torvalds's avatar
Linus Torvalds committed
359
	}
360 361 362
	return;
out:
	printk(KERN_ERR "nfsd: failed to unregister %s cache\n", cd->name);
Linus Torvalds's avatar
Linus Torvalds committed
363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390
}

/* clean cache tries to find something to clean
 * and cleans it.
 * It returns 1 if it cleaned something,
 *            0 if it didn't find anything this time
 *           -1 if it fell off the end of the list.
 */
static int cache_clean(void)
{
	int rv = 0;
	struct list_head *next;

	spin_lock(&cache_list_lock);

	/* find a suitable table if we don't already have one */
	while (current_detail == NULL ||
	    current_index >= current_detail->hash_size) {
		if (current_detail)
			next = current_detail->others.next;
		else
			next = cache_list.next;
		if (next == &cache_list) {
			current_detail = NULL;
			spin_unlock(&cache_list_lock);
			return -1;
		}
		current_detail = list_entry(next, struct cache_detail, others);
391
		if (current_detail->nextcheck > seconds_since_boot())
Linus Torvalds's avatar
Linus Torvalds committed
392 393 394
			current_index = current_detail->hash_size;
		else {
			current_index = 0;
395
			current_detail->nextcheck = seconds_since_boot()+30*60;
Linus Torvalds's avatar
Linus Torvalds committed
396 397 398 399 400 401 402 403 404 405
		}
	}

	/* find a non-empty bucket in the table */
	while (current_detail &&
	       current_index < current_detail->hash_size &&
	       current_detail->hash_table[current_index] == NULL)
		current_index++;

	/* find a cleanable entry in the bucket and clean it, or set to next bucket */
406

Linus Torvalds's avatar
Linus Torvalds committed
407 408 409
	if (current_detail && current_index < current_detail->hash_size) {
		struct cache_head *ch, **cp;
		struct cache_detail *d;
410

Linus Torvalds's avatar
Linus Torvalds committed
411 412 413
		write_lock(&current_detail->hash_lock);

		/* Ok, now to clean this strand */
414

Linus Torvalds's avatar
Linus Torvalds committed
415
		cp = & current_detail->hash_table[current_index];
416
		for (ch = *cp ; ch ; cp = & ch->next, ch = *cp) {
Linus Torvalds's avatar
Linus Torvalds committed
417 418
			if (current_detail->nextcheck > ch->expiry_time)
				current_detail->nextcheck = ch->expiry_time+1;
419
			if (!cache_is_expired(current_detail, ch))
Linus Torvalds's avatar
Linus Torvalds committed
420 421 422 423 424 425
				continue;

			*cp = ch->next;
			ch->next = NULL;
			current_detail->entries--;
			rv = 1;
426
			break;
Linus Torvalds's avatar
Linus Torvalds committed
427
		}
428

Linus Torvalds's avatar
Linus Torvalds committed
429 430 431 432 433
		write_unlock(&current_detail->hash_lock);
		d = current_detail;
		if (!ch)
			current_index ++;
		spin_unlock(&cache_list_lock);
434
		if (ch) {
435 436
			if (test_and_clear_bit(CACHE_PENDING, &ch->flags))
				cache_dequeue(current_detail, ch);
437
			cache_revisit_request(ch);
438
			cache_put(ch, d);
439
		}
Linus Torvalds's avatar
Linus Torvalds committed
440 441 442 443 444 445 446 447 448
	} else
		spin_unlock(&cache_list_lock);

	return rv;
}

/*
 * We want to regularly clean the cache, so we need to schedule some work ...
 */
449
static void do_cache_clean(struct work_struct *work)
Linus Torvalds's avatar
Linus Torvalds committed
450 451 452
{
	int delay = 5;
	if (cache_clean() == -1)
453
		delay = round_jiffies_relative(30*HZ);
Linus Torvalds's avatar
Linus Torvalds committed
454 455 456 457 458 459 460 461 462

	if (list_empty(&cache_list))
		delay = 0;

	if (delay)
		schedule_delayed_work(&cache_cleaner, delay);
}


463
/*
Linus Torvalds's avatar
Linus Torvalds committed
464
 * Clean all caches promptly.  This just calls cache_clean
465
 * repeatedly until we are sure that every cache has had a chance to
Linus Torvalds's avatar
Linus Torvalds committed
466 467 468 469 470 471 472 473 474
 * be fully cleaned
 */
void cache_flush(void)
{
	while (cache_clean() != -1)
		cond_resched();
	while (cache_clean() != -1)
		cond_resched();
}
475
EXPORT_SYMBOL_GPL(cache_flush);
Linus Torvalds's avatar
Linus Torvalds committed
476 477 478 479

void cache_purge(struct cache_detail *detail)
{
	detail->flush_time = LONG_MAX;
480
	detail->nextcheck = seconds_since_boot();
Linus Torvalds's avatar
Linus Torvalds committed
481 482 483
	cache_flush();
	detail->flush_time = 1;
}
484
EXPORT_SYMBOL_GPL(cache_purge);
Linus Torvalds's avatar
Linus Torvalds committed
485 486 487 488 489 490 491 492 493 494


/*
 * Deferral and Revisiting of Requests.
 *
 * If a cache lookup finds a pending entry, we
 * need to defer the request and revisit it later.
 * All deferred requests are stored in a hash table,
 * indexed by "struct cache_head *".
 * As it may be wasteful to store a whole request
495
 * structure, we allow the request to provide a
Linus Torvalds's avatar
Linus Torvalds committed
496 497 498 499 500 501 502 503 504 505 506 507 508
 * deferred form, which must contain a
 * 'struct cache_deferred_req'
 * This cache_deferred_req contains a method to allow
 * it to be revisited when cache info is available
 */

#define	DFR_HASHSIZE	(PAGE_SIZE/sizeof(struct list_head))
#define	DFR_HASH(item)	((((long)item)>>4 ^ (((long)item)>>13)) % DFR_HASHSIZE)

#define	DFR_MAX	300	/* ??? */

static DEFINE_SPINLOCK(cache_defer_lock);
static LIST_HEAD(cache_defer_list);
509
static struct hlist_head cache_defer_hash[DFR_HASHSIZE];
Linus Torvalds's avatar
Linus Torvalds committed
510 511
static int cache_defer_cnt;

512 513 514
static void __unhash_deferred_req(struct cache_deferred_req *dreq)
{
	list_del_init(&dreq->recent);
515
	hlist_del_init(&dreq->hash);
516 517 518 519 520 521 522 523
	cache_defer_cnt--;
}

static void __hash_deferred_req(struct cache_deferred_req *dreq, struct cache_head *item)
{
	int hash = DFR_HASH(item);

	list_add(&dreq->recent, &cache_defer_list);
524
	hlist_add_head(&dreq->hash, &cache_defer_hash[hash]);
525 526
}

527
static int setup_deferral(struct cache_deferred_req *dreq, struct cache_head *item)
Linus Torvalds's avatar
Linus Torvalds committed
528
{
529
	struct cache_deferred_req *discard;
Linus Torvalds's avatar
Linus Torvalds committed
530 531 532 533 534

	dreq->item = item;

	spin_lock(&cache_defer_lock);

535
	__hash_deferred_req(dreq, item);
Linus Torvalds's avatar
Linus Torvalds committed
536 537

	/* it is in, now maybe clean up */
538
	discard = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
539
	if (++cache_defer_cnt > DFR_MAX) {
540 541
		discard = list_entry(cache_defer_list.prev,
				     struct cache_deferred_req, recent);
542
		__unhash_deferred_req(discard);
Linus Torvalds's avatar
Linus Torvalds committed
543 544 545
	}
	spin_unlock(&cache_defer_lock);

546
	if (discard)
Linus Torvalds's avatar
Linus Torvalds committed
547
		/* there was one too many */
548 549
		discard->revisit(discard, 1);

550
	if (!test_bit(CACHE_PENDING, &item->flags)) {
Linus Torvalds's avatar
Linus Torvalds committed
551 552
		/* must have just been validated... */
		cache_revisit_request(item);
553
		return -EAGAIN;
Linus Torvalds's avatar
Linus Torvalds committed
554
	}
555 556
	return 0;
}
557

558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588
struct thread_deferred_req {
	struct cache_deferred_req handle;
	struct completion completion;
};

static void cache_restart_thread(struct cache_deferred_req *dreq, int too_many)
{
	struct thread_deferred_req *dr =
		container_of(dreq, struct thread_deferred_req, handle);
	complete(&dr->completion);
}

static int cache_wait_req(struct cache_req *req, struct cache_head *item)
{
	struct thread_deferred_req sleeper;
	struct cache_deferred_req *dreq = &sleeper.handle;
	int ret;

	sleeper.completion = COMPLETION_INITIALIZER_ONSTACK(sleeper.completion);
	dreq->revisit = cache_restart_thread;

	ret = setup_deferral(dreq, item);
	if (ret)
		return ret;

	if (wait_for_completion_interruptible_timeout(
		    &sleeper.completion, req->thread_wait) <= 0) {
		/* The completion wasn't completed, so we need
		 * to clean up
		 */
		spin_lock(&cache_defer_lock);
589
		if (!hlist_unhashed(&sleeper.handle.hash)) {
590 591 592 593 594 595 596
			__unhash_deferred_req(&sleeper.handle);
			spin_unlock(&cache_defer_lock);
		} else {
			/* cache_revisit_request already removed
			 * this from the hash table, but hasn't
			 * called ->revisit yet.  It will very soon
			 * and we need to wait for it.
597
			 */
598 599
			spin_unlock(&cache_defer_lock);
			wait_for_completion(&sleeper.completion);
600
		}
601 602 603 604
	}
	if (test_bit(CACHE_PENDING, &item->flags)) {
		/* item is still pending, try request
		 * deferral
605
		 */
606
		return -ETIMEDOUT;
607
	}
608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636
	/* only return success if we actually deferred the
	 * request.  In this case we waited until it was
	 * answered so no deferral has happened - rather
	 * an answer already exists.
	 */
	return -EEXIST;
}

static int cache_defer_req(struct cache_req *req, struct cache_head *item)
{
	struct cache_deferred_req *dreq;
	int ret;

	if (cache_defer_cnt >= DFR_MAX) {
		/* too much in the cache, randomly drop this one,
		 * or continue and drop the oldest
		 */
		if (net_random()&1)
			return -ENOMEM;
	}
	if (req->thread_wait) {
		ret = cache_wait_req(req, item);
		if (ret != -ETIMEDOUT)
			return ret;
	}
	dreq = req->defer(req);
	if (dreq == NULL)
		return -ENOMEM;
	return setup_deferral(dreq, item);
Linus Torvalds's avatar
Linus Torvalds committed
637 638 639 640 641 642
}

static void cache_revisit_request(struct cache_head *item)
{
	struct cache_deferred_req *dreq;
	struct list_head pending;
643
	struct hlist_node *lp, *tmp;
Linus Torvalds's avatar
Linus Torvalds committed
644 645 646 647
	int hash = DFR_HASH(item);

	INIT_LIST_HEAD(&pending);
	spin_lock(&cache_defer_lock);
648

649 650 651 652
	hlist_for_each_entry_safe(dreq, lp, tmp, &cache_defer_hash[hash], hash)
		if (dreq->item == item) {
			__unhash_deferred_req(dreq);
			list_add(&dreq->recent, &pending);
Linus Torvalds's avatar
Linus Torvalds committed
653
		}
654

Linus Torvalds's avatar
Linus Torvalds committed
655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671
	spin_unlock(&cache_defer_lock);

	while (!list_empty(&pending)) {
		dreq = list_entry(pending.next, struct cache_deferred_req, recent);
		list_del_init(&dreq->recent);
		dreq->revisit(dreq, 0);
	}
}

void cache_clean_deferred(void *owner)
{
	struct cache_deferred_req *dreq, *tmp;
	struct list_head pending;


	INIT_LIST_HEAD(&pending);
	spin_lock(&cache_defer_lock);
672

Linus Torvalds's avatar
Linus Torvalds committed
673
	list_for_each_entry_safe(dreq, tmp, &cache_defer_list, recent) {
674
		if (dreq->owner == owner) {
675
			__unhash_deferred_req(dreq);
676 677
			list_add(&dreq->recent, &pending);
		}
Linus Torvalds's avatar
Linus Torvalds committed
678 679 680 681 682 683 684 685 686 687 688 689 690
	}
	spin_unlock(&cache_defer_lock);

	while (!list_empty(&pending)) {
		dreq = list_entry(pending.next, struct cache_deferred_req, recent);
		list_del_init(&dreq->recent);
		dreq->revisit(dreq, 1);
	}
}

/*
 * communicate with user-space
 *
J. Bruce Fields's avatar
J. Bruce Fields committed
691 692 693 694
 * We have a magic /proc file - /proc/sunrpc/<cachename>/channel.
 * On read, you get a full request, or block.
 * On write, an update request is processed.
 * Poll works if anything to read, and always allows write.
Linus Torvalds's avatar
Linus Torvalds committed
695
 *
696
 * Implemented by linked list of requests.  Each open file has
J. Bruce Fields's avatar
J. Bruce Fields committed
697
 * a ->private that also exists in this list.  New requests are added
Linus Torvalds's avatar
Linus Torvalds committed
698 699 700 701 702 703 704
 * to the end and may wakeup and preceding readers.
 * New readers are added to the head.  If, on read, an item is found with
 * CACHE_UPCALLING clear, we free it from the list.
 *
 */

static DEFINE_SPINLOCK(queue_lock);
Arjan van de Ven's avatar
Arjan van de Ven committed
705
static DEFINE_MUTEX(queue_io_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722

struct cache_queue {
	struct list_head	list;
	int			reader;	/* if 0, then request */
};
struct cache_request {
	struct cache_queue	q;
	struct cache_head	*item;
	char			* buf;
	int			len;
	int			readers;
};
struct cache_reader {
	struct cache_queue	q;
	int			offset;	/* if non-0, we have a refcnt on next request */
};

723 724
static ssize_t cache_read(struct file *filp, char __user *buf, size_t count,
			  loff_t *ppos, struct cache_detail *cd)
Linus Torvalds's avatar
Linus Torvalds committed
725 726 727
{
	struct cache_reader *rp = filp->private_data;
	struct cache_request *rq;
728
	struct inode *inode = filp->f_path.dentry->d_inode;
Linus Torvalds's avatar
Linus Torvalds committed
729 730 731 732 733
	int err;

	if (count == 0)
		return 0;

734
	mutex_lock(&inode->i_mutex); /* protect against multiple concurrent
Linus Torvalds's avatar
Linus Torvalds committed
735 736 737 738 739 740 741 742 743 744 745 746
			      * readers on this file */
 again:
	spin_lock(&queue_lock);
	/* need to find next request */
	while (rp->q.list.next != &cd->queue &&
	       list_entry(rp->q.list.next, struct cache_queue, list)
	       ->reader) {
		struct list_head *next = rp->q.list.next;
		list_move(&rp->q.list, next);
	}
	if (rp->q.list.next == &cd->queue) {
		spin_unlock(&queue_lock);
747
		mutex_unlock(&inode->i_mutex);
748
		BUG_ON(rp->offset);
Linus Torvalds's avatar
Linus Torvalds committed
749 750 751
		return 0;
	}
	rq = container_of(rp->q.list.next, struct cache_request, q.list);
752
	BUG_ON(rq->q.reader);
Linus Torvalds's avatar
Linus Torvalds committed
753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785
	if (rp->offset == 0)
		rq->readers++;
	spin_unlock(&queue_lock);

	if (rp->offset == 0 && !test_bit(CACHE_PENDING, &rq->item->flags)) {
		err = -EAGAIN;
		spin_lock(&queue_lock);
		list_move(&rp->q.list, &rq->q.list);
		spin_unlock(&queue_lock);
	} else {
		if (rp->offset + count > rq->len)
			count = rq->len - rp->offset;
		err = -EFAULT;
		if (copy_to_user(buf, rq->buf + rp->offset, count))
			goto out;
		rp->offset += count;
		if (rp->offset >= rq->len) {
			rp->offset = 0;
			spin_lock(&queue_lock);
			list_move(&rp->q.list, &rq->q.list);
			spin_unlock(&queue_lock);
		}
		err = 0;
	}
 out:
	if (rp->offset == 0) {
		/* need to release rq */
		spin_lock(&queue_lock);
		rq->readers--;
		if (rq->readers == 0 &&
		    !test_bit(CACHE_PENDING, &rq->item->flags)) {
			list_del(&rq->q.list);
			spin_unlock(&queue_lock);
786
			cache_put(rq->item, cd);
Linus Torvalds's avatar
Linus Torvalds committed
787 788 789 790 791 792 793
			kfree(rq->buf);
			kfree(rq);
		} else
			spin_unlock(&queue_lock);
	}
	if (err == -EAGAIN)
		goto again;
794
	mutex_unlock(&inode->i_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
795 796 797
	return err ? err :  count;
}

798 799 800 801
static ssize_t cache_do_downcall(char *kaddr, const char __user *buf,
				 size_t count, struct cache_detail *cd)
{
	ssize_t ret;
Linus Torvalds's avatar
Linus Torvalds committed
802

803 804 805 806 807 808 809 810 811 812 813
	if (copy_from_user(kaddr, buf, count))
		return -EFAULT;
	kaddr[count] = '\0';
	ret = cd->cache_parse(cd, kaddr, count);
	if (!ret)
		ret = count;
	return ret;
}

static ssize_t cache_slow_downcall(const char __user *buf,
				   size_t count, struct cache_detail *cd)
Linus Torvalds's avatar
Linus Torvalds committed
814
{
815 816
	static char write_buf[8192]; /* protected by queue_io_mutex */
	ssize_t ret = -EINVAL;
Linus Torvalds's avatar
Linus Torvalds committed
817 818

	if (count >= sizeof(write_buf))
819
		goto out;
Arjan van de Ven's avatar
Arjan van de Ven committed
820
	mutex_lock(&queue_io_mutex);
821 822 823 824 825
	ret = cache_do_downcall(write_buf, buf, count, cd);
	mutex_unlock(&queue_io_mutex);
out:
	return ret;
}
Linus Torvalds's avatar
Linus Torvalds committed
826

827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850
static ssize_t cache_downcall(struct address_space *mapping,
			      const char __user *buf,
			      size_t count, struct cache_detail *cd)
{
	struct page *page;
	char *kaddr;
	ssize_t ret = -ENOMEM;

	if (count >= PAGE_CACHE_SIZE)
		goto out_slow;

	page = find_or_create_page(mapping, 0, GFP_KERNEL);
	if (!page)
		goto out_slow;

	kaddr = kmap(page);
	ret = cache_do_downcall(kaddr, buf, count, cd);
	kunmap(page);
	unlock_page(page);
	page_cache_release(page);
	return ret;
out_slow:
	return cache_slow_downcall(buf, count, cd);
}
Linus Torvalds's avatar
Linus Torvalds committed
851

852 853 854
static ssize_t cache_write(struct file *filp, const char __user *buf,
			   size_t count, loff_t *ppos,
			   struct cache_detail *cd)
855 856 857 858 859 860 861 862 863 864 865 866 867
{
	struct address_space *mapping = filp->f_mapping;
	struct inode *inode = filp->f_path.dentry->d_inode;
	ssize_t ret = -EINVAL;

	if (!cd->cache_parse)
		goto out;

	mutex_lock(&inode->i_mutex);
	ret = cache_downcall(mapping, buf, count, cd);
	mutex_unlock(&inode->i_mutex);
out:
	return ret;
Linus Torvalds's avatar
Linus Torvalds committed
868 869 870 871
}

static DECLARE_WAIT_QUEUE_HEAD(queue_wait);

872 873
static unsigned int cache_poll(struct file *filp, poll_table *wait,
			       struct cache_detail *cd)
Linus Torvalds's avatar
Linus Torvalds committed
874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898
{
	unsigned int mask;
	struct cache_reader *rp = filp->private_data;
	struct cache_queue *cq;

	poll_wait(filp, &queue_wait, wait);

	/* alway allow write */
	mask = POLL_OUT | POLLWRNORM;

	if (!rp)
		return mask;

	spin_lock(&queue_lock);

	for (cq= &rp->q; &cq->list != &cd->queue;
	     cq = list_entry(cq->list.next, struct cache_queue, list))
		if (!cq->reader) {
			mask |= POLLIN | POLLRDNORM;
			break;
		}
	spin_unlock(&queue_lock);
	return mask;
}

899 900 901
static int cache_ioctl(struct inode *ino, struct file *filp,
		       unsigned int cmd, unsigned long arg,
		       struct cache_detail *cd)
Linus Torvalds's avatar
Linus Torvalds committed
902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927
{
	int len = 0;
	struct cache_reader *rp = filp->private_data;
	struct cache_queue *cq;

	if (cmd != FIONREAD || !rp)
		return -EINVAL;

	spin_lock(&queue_lock);

	/* only find the length remaining in current request,
	 * or the length of the next request
	 */
	for (cq= &rp->q; &cq->list != &cd->queue;
	     cq = list_entry(cq->list.next, struct cache_queue, list))
		if (!cq->reader) {
			struct cache_request *cr =
				container_of(cq, struct cache_request, q);
			len = cr->len - rp->offset;
			break;
		}
	spin_unlock(&queue_lock);

	return put_user(len, (int __user *)arg);
}

928 929
static int cache_open(struct inode *inode, struct file *filp,
		      struct cache_detail *cd)
Linus Torvalds's avatar
Linus Torvalds committed
930 931 932
{
	struct cache_reader *rp = NULL;

933 934
	if (!cd || !try_module_get(cd->owner))
		return -EACCES;
Linus Torvalds's avatar
Linus Torvalds committed
935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950
	nonseekable_open(inode, filp);
	if (filp->f_mode & FMODE_READ) {
		rp = kmalloc(sizeof(*rp), GFP_KERNEL);
		if (!rp)
			return -ENOMEM;
		rp->offset = 0;
		rp->q.reader = 1;
		atomic_inc(&cd->readers);
		spin_lock(&queue_lock);
		list_add(&rp->q.list, &cd->queue);
		spin_unlock(&queue_lock);
	}
	filp->private_data = rp;
	return 0;
}

951 952
static int cache_release(struct inode *inode, struct file *filp,
			 struct cache_detail *cd)
Linus Torvalds's avatar
Linus Torvalds committed
953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974
{
	struct cache_reader *rp = filp->private_data;

	if (rp) {
		spin_lock(&queue_lock);
		if (rp->offset) {
			struct cache_queue *cq;
			for (cq= &rp->q; &cq->list != &cd->queue;
			     cq = list_entry(cq->list.next, struct cache_queue, list))
				if (!cq->reader) {
					container_of(cq, struct cache_request, q)
						->readers--;
					break;
				}
			rp->offset = 0;
		}
		list_del(&rp->q.list);
		spin_unlock(&queue_lock);

		filp->private_data = NULL;
		kfree(rp);

975
		cd->last_close = seconds_since_boot();
Linus Torvalds's avatar
Linus Torvalds committed
976 977
		atomic_dec(&cd->readers);
	}
978
	module_put(cd->owner);
Linus Torvalds's avatar
Linus Torvalds committed
979 980 981 982 983
	return 0;
}



984
static void cache_dequeue(struct cache_detail *detail, struct cache_head *ch)
Linus Torvalds's avatar
Linus Torvalds committed
985 986 987 988 989 990 991 992 993
{
	struct cache_queue *cq;
	spin_lock(&queue_lock);
	list_for_each_entry(cq, &detail->queue, list)
		if (!cq->reader) {
			struct cache_request *cr = container_of(cq, struct cache_request, q);
			if (cr->item != ch)
				continue;
			if (cr->readers != 0)
994
				continue;
Linus Torvalds's avatar
Linus Torvalds committed
995 996
			list_del(&cr->q.list);
			spin_unlock(&queue_lock);
997
			cache_put(cr->item, detail);
Linus Torvalds's avatar
Linus Torvalds committed
998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047
			kfree(cr->buf);
			kfree(cr);
			return;
		}
	spin_unlock(&queue_lock);
}

/*
 * Support routines for text-based upcalls.
 * Fields are separated by spaces.
 * Fields are either mangled to quote space tab newline slosh with slosh
 * or a hexified with a leading \x
 * Record is terminated with newline.
 *
 */

void qword_add(char **bpp, int *lp, char *str)
{
	char *bp = *bpp;
	int len = *lp;
	char c;

	if (len < 0) return;

	while ((c=*str++) && len)
		switch(c) {
		case ' ':
		case '\t':
		case '\n':
		case '\\':
			if (len >= 4) {
				*bp++ = '\\';
				*bp++ = '0' + ((c & 0300)>>6);
				*bp++ = '0' + ((c & 0070)>>3);
				*bp++ = '0' + ((c & 0007)>>0);
			}
			len -= 4;
			break;
		default:
			*bp++ = c;
			len--;
		}
	if (c || len <1) len = -1;
	else {
		*bp++ = ' ';
		len--;
	}
	*bpp = bp;
	*lp = len;
}
1048
EXPORT_SYMBOL_GPL(qword_add);
Linus Torvalds's avatar
Linus Torvalds committed
1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076

void qword_addhex(char **bpp, int *lp, char *buf, int blen)
{
	char *bp = *bpp;
	int len = *lp;

	if (len < 0) return;

	if (len > 2) {
		*bp++ = '\\';
		*bp++ = 'x';
		len -= 2;
		while (blen && len >= 2) {
			unsigned char c = *buf++;
			*bp++ = '0' + ((c&0xf0)>>4) + (c>=0xa0)*('a'-'9'-1);
			*bp++ = '0' + (c&0x0f) + ((c&0x0f)>=0x0a)*('a'-'9'-1);
			len -= 2;
			blen--;
		}
	}
	if (blen || len<1) len = -1;
	else {
		*bp++ = ' ';
		len--;
	}
	*bpp = bp;
	*lp = len;
}
1077
EXPORT_SYMBOL_GPL(qword_addhex);
Linus Torvalds's avatar
Linus Torvalds committed
1078 1079 1080 1081 1082 1083

static void warn_no_listener(struct cache_detail *detail)
{
	if (detail->last_warn != detail->last_close) {
		detail->last_warn = detail->last_close;
		if (detail->warn_no_listener)
1084
			detail->warn_no_listener(detail, detail->last_close != 0);
Linus Torvalds's avatar
Linus Torvalds committed
1085 1086 1087
	}
}

1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104
static bool cache_listeners_exist(struct cache_detail *detail)
{
	if (atomic_read(&detail->readers))
		return true;
	if (detail->last_close == 0)
		/* This cache was never opened */
		return false;
	if (detail->last_close < seconds_since_boot() - 30)
		/*
		 * We allow for the possibility that someone might
		 * restart a userspace daemon without restarting the
		 * server; but after 30 seconds, we give up.
		 */
		 return false;
	return true;
}

Linus Torvalds's avatar
Linus Torvalds committed
1105
/*
1106 1107 1108
 * register an upcall request to user-space and queue it up for read() by the
 * upcall daemon.
 *
Linus Torvalds's avatar
Linus Torvalds committed
1109 1110
 * Each request is at most one page long.
 */
1111 1112 1113 1114 1115
int sunrpc_cache_pipe_upcall(struct cache_detail *detail, struct cache_head *h,
		void (*cache_request)(struct cache_detail *,
				      struct cache_head *,
				      char **,
				      int *))
Linus Torvalds's avatar
Linus Torvalds committed
1116 1117 1118 1119 1120 1121 1122
{

	char *buf;
	struct cache_request *crq;
	char *bp;
	int len;

1123 1124 1125
	if (!cache_listeners_exist(detail)) {
		warn_no_listener(detail);
		return -EINVAL;
Linus Torvalds's avatar
Linus Torvalds committed
1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139
	}

	buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
	if (!buf)
		return -EAGAIN;

	crq = kmalloc(sizeof (*crq), GFP_KERNEL);
	if (!crq) {
		kfree(buf);
		return -EAGAIN;
	}

	bp = buf; len = PAGE_SIZE;

1140
	cache_request(detail, h, &bp, &len);
Linus Torvalds's avatar
Linus Torvalds committed
1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157

	if (len < 0) {
		kfree(buf);
		kfree(crq);
		return -EAGAIN;
	}
	crq->q.reader = 0;
	crq->item = cache_get(h);
	crq->buf = buf;
	crq->len = PAGE_SIZE - len;
	crq->readers = 0;
	spin_lock(&queue_lock);
	list_add_tail(&crq->q.list, &detail->queue);
	spin_unlock(&queue_lock);
	wake_up(&queue_wait);
	return 0;
}
1158
EXPORT_SYMBOL_GPL(sunrpc_cache_pipe_upcall);
Linus Torvalds's avatar
Linus Torvalds committed
1159 1160 1161 1162 1163 1164 1165

/*
 * parse a message from user-space and pass it
 * to an appropriate cache
 * Messages are, like requests, separated into fields by
 * spaces and dequotes as \xHEXSTRING or embedded \nnn octal
 *
1166
 * Message is
Linus Torvalds's avatar
Linus Torvalds committed
1167 1168
 *   reply cachename expiry key ... content....
 *
1169
 * key and content are both parsed by cache
Linus Torvalds's avatar
Linus Torvalds committed
1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183
 */

#define isodigit(c) (isdigit(c) && c <= '7')
int qword_get(char **bpp, char *dest, int bufsize)
{
	/* return bytes copied, or -1 on error */
	char *bp = *bpp;
	int len = 0;

	while (*bp == ' ') bp++;

	if (bp[0] == '\\' && bp[1] == 'x') {
		/* HEX STRING */
		bp += 2;
1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196
		while (len < bufsize) {
			int h, l;

			h = hex_to_bin(bp[0]);
			if (h < 0)
				break;

			l = hex_to_bin(bp[1]);
			if (l < 0)
				break;

			*dest++ = (h << 4) | l;
			bp += 2;
Linus Torvalds's avatar
Linus Torvalds committed
1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225
			len++;
		}
	} else {
		/* text with \nnn octal quoting */
		while (*bp != ' ' && *bp != '\n' && *bp && len < bufsize-1) {
			if (*bp == '\\' &&
			    isodigit(bp[1]) && (bp[1] <= '3') &&
			    isodigit(bp[2]) &&
			    isodigit(bp[3])) {
				int byte = (*++bp -'0');
				bp++;
				byte = (byte << 3) | (*bp++ - '0');
				byte = (byte << 3) | (*bp++ - '0');
				*dest++ = byte;
				len++;
			} else {
				*dest++ = *bp++;
				len++;
			}
		}
	}

	if (*bp != ' ' && *bp != '\n' && *bp != '\0')
		return -1;
	while (*bp == ' ') bp++;
	*bpp = bp;
	*dest = '\0';
	return len;
}
1226
EXPORT_SYMBOL_GPL(qword_get);
Linus Torvalds's avatar
Linus Torvalds committed
1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240


/*
 * support /proc/sunrpc/cache/$CACHENAME/content
 * as a seqfile.
 * We call ->cache_show passing NULL for the item to
 * get a header, then pass each real item in the cache
 */

struct handle {
	struct cache_detail *cd;
};

static void *c_start(struct seq_file *m, loff_t *pos)
1241
	__acquires(cd->hash_lock)
Linus Torvalds's avatar
Linus Torvalds committed
1242 1243 1244 1245 1246
{
	loff_t n = *pos;
	unsigned hash, entry;
	struct cache_head *ch;
	struct cache_detail *cd = ((struct handle*)m->private)->cd;
1247

Linus Torvalds's avatar
Linus Torvalds committed
1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261

	read_lock(&cd->hash_lock);
	if (!n--)
		return SEQ_START_TOKEN;
	hash = n >> 32;
	entry = n & ((1LL<<32) - 1);

	for (ch=cd->hash_table[hash]; ch; ch=ch->next)
		if (!entry--)
			return ch;
	n &= ~((1LL<<32) - 1);
	do {
		hash++;
		n += 1LL<<32;
1262
	} while(hash < cd->hash_size &&
Linus Torvalds's avatar
Linus Torvalds committed
1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297
		cd->hash_table[hash]==NULL);
	if (hash >= cd->hash_size)
		return NULL;
	*pos = n+1;
	return cd->hash_table[hash];
}

static void *c_next(struct seq_file *m, void *p, loff_t *pos)
{
	struct cache_head *ch = p;
	int hash = (*pos >> 32);
	struct cache_detail *cd = ((struct handle*)m->private)->cd;

	if (p == SEQ_START_TOKEN)
		hash = 0;
	else if (ch->next == NULL) {
		hash++;
		*pos += 1LL<<32;
	} else {
		++*pos;
		return ch->next;
	}
	*pos &= ~((1LL<<32) - 1);
	while (hash < cd->hash_size &&
	       cd->hash_table[hash] == NULL) {
		hash++;
		*pos += 1LL<<32;
	}
	if (hash >= cd->hash_size)
		return NULL;
	++*pos;
	return cd->hash_table[hash];
}

static void c_stop(struct seq_file *m, void *p)
1298
	__releases(cd->hash_lock)
Linus Torvalds's avatar
Linus Torvalds committed
1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312
{
	struct cache_detail *cd = ((struct handle*)m->private)->cd;
	read_unlock(&cd->hash_lock);
}

static int c_show(struct seq_file *m, void *p)
{
	struct cache_head *cp = p;
	struct cache_detail *cd = ((struct handle*)m->private)->cd;

	if (p == SEQ_START_TOKEN)
		return cd->cache_show(m, cd, NULL);

	ifdebug(CACHE)
1313
		seq_printf(m, "# expiry=%ld refcnt=%d flags=%lx\n",
1314 1315
			   convert_to_wallclock(cp->expiry_time),
			   atomic_read(&cp->ref.refcount), cp->flags);
Linus Torvalds's avatar
Linus Torvalds committed
1316 1317 1318 1319 1320 1321 1322 1323 1324 1325
	cache_get(cp);
	if (cache_check(cd, cp, NULL))
		/* cache_check does a cache_put on failure */
		seq_printf(m, "# ");
	else
		cache_put(cp, cd);

	return cd->cache_show(m, cd, cp);
}

1326
static const struct seq_operations cache_content_op = {
Linus Torvalds's avatar
Linus Torvalds committed
1327 1328 1329 1330 1331 1332
	.start	= c_start,
	.next	= c_next,
	.stop	= c_stop,
	.show	= c_show,
};

1333 1334
static int content_open(struct inode *inode, struct file *file,
			struct cache_detail *cd)
Linus Torvalds's avatar
Linus Torvalds committed
1335 1336 1337
{
	struct handle *han;

1338 1339
	if (!cd || !try_module_get(cd->owner))
		return -EACCES;
1340
	han = __seq_open_private(file, &cache_content_op, sizeof(*han));
1341 1342
	if (han == NULL) {
		module_put(cd->owner);
Linus Torvalds's avatar
Linus Torvalds committed
1343
		return -ENOMEM;
1344
	}
Linus Torvalds's avatar
Linus Torvalds committed
1345 1346

	han->cd = cd;
1347
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
1348 1349
}

1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371
static int content_release(struct inode *inode, struct file *file,
		struct cache_detail *cd)
{
	int ret = seq_release_private(inode, file);
	module_put(cd->owner);
	return ret;
}

static int open_flush(struct inode *inode, struct file *file,
			struct cache_detail *cd)
{
	if (!cd || !try_module_get(cd->owner))
		return -EACCES;
	return nonseekable_open(inode, file);
}

static int release_flush(struct inode *inode, struct file *file,
			struct cache_detail *cd)
{
	module_put(cd->owner);
	return 0;
}
Linus Torvalds's avatar
Linus Torvalds committed
1372 1373

static ssize_t read_flush(struct file *file, char __user *buf,
1374 1375
			  size_t count, loff_t *ppos,
			  struct cache_detail *cd)
Linus Torvalds's avatar
Linus Torvalds committed
1376 1377 1378
{
	char tbuf[20];
	unsigned long p = *ppos;
1379
	size_t len;
Linus Torvalds's avatar
Linus Torvalds committed
1380

1381
	sprintf(tbuf, "%lu\n", convert_to_wallclock(cd->flush_time));
Linus Torvalds's avatar
Linus Torvalds committed
1382 1383 1384 1385
	len = strlen(tbuf);
	if (p >= len)
		return 0;
	len -= p;
1386 1387
	if (len > count)
		len = count;
Linus Torvalds's avatar
Linus Torvalds committed
1388
	if (copy_to_user(buf, (void*)(tbuf+p), len))
1389 1390
		return -EFAULT;
	*ppos += len;
Linus Torvalds's avatar
Linus Torvalds committed
1391 1392 1393
	return len;
}

1394 1395 1396
static ssize_t write_flush(struct file *file, const char __user *buf,
			   size_t count, loff_t *ppos,
			   struct cache_detail *cd)
Linus Torvalds's avatar
Linus Torvalds committed
1397 1398
{
	char tbuf[20];
1399 1400
	char *bp, *ep;

Linus Torvalds's avatar
Linus Torvalds committed
1401 1402 1403 1404 1405
	if (*ppos || count > sizeof(tbuf)-1)
		return -EINVAL;
	if (copy_from_user(tbuf, buf, count))
		return -EFAULT;
	tbuf[count] = 0;
1406
	simple_strtoul(tbuf, &ep, 0);
Linus Torvalds's avatar
Linus Torvalds committed
1407 1408 1409
	if (*ep && *ep != '\n')
		return -EINVAL;

1410 1411 1412
	bp = tbuf;
	cd->flush_time = get_expiry(&bp);
	cd->nextcheck = seconds_since_boot();
Linus Torvalds's avatar
Linus Torvalds committed
1413 1414 1415 1416 1417 1418
	cache_flush();

	*ppos += count;
	return count;
}

1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441
static ssize_t cache_read_procfs(struct file *filp, char __user *buf,
				 size_t count, loff_t *ppos)
{
	struct cache_detail *cd = PDE(filp->f_path.dentry->d_inode)->data;

	return cache_read(filp, buf, count, ppos, cd);
}

static ssize_t cache_write_procfs(struct file *filp, const char __user *buf,
				  size_t count, loff_t *ppos)
{
	struct cache_detail *cd = PDE(filp->f_path.dentry->d_inode)->data;

	return cache_write(filp, buf, count, ppos, cd);
}

static unsigned int cache_poll_procfs(struct file *filp, poll_table *wait)
{
	struct cache_detail *cd = PDE(filp->f_path.dentry->d_inode)->data;

	return cache_poll(filp, wait, cd);
}

1442 1443
static long cache_ioctl_procfs(struct file *filp,
			       unsigned int cmd, unsigned long arg)
1444
{
1445 1446
	long ret;
	struct inode *inode = filp->f_path.dentry->d_inode;
1447 1448
	struct cache_detail *cd = PDE(inode)->data;

1449 1450 1451 1452 1453
	lock_kernel();
	ret = cache_ioctl(inode, filp, cmd, arg, cd);
	unlock_kernel();

	return ret;
1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475
}

static int cache_open_procfs(struct inode *inode, struct file *filp)
{
	struct cache_detail *cd = PDE(inode)->data;

	return cache_open(inode, filp, cd);
}

static int cache_release_procfs(struct inode *inode, struct file *filp)
{
	struct cache_detail *cd = PDE(inode)->data;

	return cache_release(inode, filp, cd);
}

static const struct file_operations cache_file_operations_procfs = {
	.owner		= THIS_MODULE,
	.llseek		= no_llseek,
	.read		= cache_read_procfs,
	.write		= cache_write_procfs,
	.poll		= cache_poll_procfs,
1476
	.unlocked_ioctl	= cache_ioctl_procfs, /* for FIONREAD */
1477 1478
	.open		= cache_open_procfs,
	.release	= cache_release_procfs,
Linus Torvalds's avatar
Linus Torvalds committed
1479
};
1480 1481 1482 1483 1484 1485 1486 1487

static int content_open_procfs(struct inode *inode, struct file *filp)
{
	struct cache_detail *cd = PDE(inode)->data;

	return content_open(inode, filp, cd);
}

1488 1489 1490 1491 1492 1493 1494
static int content_release_procfs(struct inode *inode, struct file *filp)
{
	struct cache_detail *cd = PDE(inode)->data;

	return content_release(inode, filp, cd);
}

1495 1496 1497 1498
static const struct file_operations content_file_operations_procfs = {
	.open		= content_open_procfs,
	.read		= seq_read,
	.llseek		= seq_lseek,
1499
	.release	= content_release_procfs,
1500 1501
};

1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515
static int open_flush_procfs(struct inode *inode, struct file *filp)
{
	struct cache_detail *cd = PDE(inode)->data;

	return open_flush(inode, filp, cd);
}

static int release_flush_procfs(struct inode *inode, struct file *filp)
{
	struct cache_detail *cd = PDE(inode)->data;

	return release_flush(inode, filp, cd);
}

1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533
static ssize_t read_flush_procfs(struct file *filp, char __user *buf,
			    size_t count, loff_t *ppos)
{
	struct cache_detail *cd = PDE(filp->f_path.dentry->d_inode)->data;

	return read_flush(filp, buf, count, ppos, cd);
}

static ssize_t write_flush_procfs(struct file *filp,
				  const char __user *buf,
				  size_t count, loff_t *ppos)
{
	struct cache_detail *cd = PDE(filp->f_path.dentry->d_inode)->data;

	return write_flush(filp, buf, count, ppos, cd);
}

static const struct file_operations cache_flush_operations_procfs = {
1534
	.open		= open_flush_procfs,
1535 1536
	.read		= read_flush_procfs,
	.write		= write_flush_procfs,
1537
	.release	= release_flush_procfs,
Linus Torvalds's avatar
Linus Torvalds committed
1538
};
1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599

static void remove_cache_proc_entries(struct cache_detail *cd)
{
	if (cd->u.procfs.proc_ent == NULL)
		return;
	if (cd->u.procfs.flush_ent)
		remove_proc_entry("flush", cd->u.procfs.proc_ent);
	if (cd->u.procfs.channel_ent)
		remove_proc_entry("channel", cd->u.procfs.proc_ent);
	if (cd->u.procfs.content_ent)
		remove_proc_entry("content", cd->u.procfs.proc_ent);
	cd->u.procfs.proc_ent = NULL;
	remove_proc_entry(cd->name, proc_net_rpc);
}

#ifdef CONFIG_PROC_FS
static int create_cache_proc_entries(struct cache_detail *cd)
{
	struct proc_dir_entry *p;

	cd->u.procfs.proc_ent = proc_mkdir(cd->name, proc_net_rpc);
	if (cd->u.procfs.proc_ent == NULL)
		goto out_nomem;
	cd->u.procfs.channel_ent = NULL;
	cd->u.procfs.content_ent = NULL;

	p = proc_create_data("flush", S_IFREG|S_IRUSR|S_IWUSR,
			     cd->u.procfs.proc_ent,
			     &cache_flush_operations_procfs, cd);
	cd->u.procfs.flush_ent = p;
	if (p == NULL)
		goto out_nomem;

	if (cd->cache_upcall || cd->cache_parse) {
		p = proc_create_data("channel", S_IFREG|S_IRUSR|S_IWUSR,
				     cd->u.procfs.proc_ent,
				     &cache_file_operations_procfs, cd);
		cd->u.procfs.channel_ent = p;
		if (p == NULL)
			goto out_nomem;
	}
	if (cd->cache_show) {
		p = proc_create_data("content", S_IFREG|S_IRUSR|S_IWUSR,
				cd->u.procfs.proc_ent,
				&content_file_operations_procfs, cd);
		cd->u.procfs.content_ent = p;
		if (p == NULL)
			goto out_nomem;
	}
	return 0;
out_nomem:
	remove_cache_proc_entries(cd);
	return -ENOMEM;
}
#else /* CONFIG_PROC_FS */
static int create_cache_proc_entries(struct cache_detail *cd)
{
	return 0;
}
#endif

1600 1601 1602 1603 1604
void __init cache_initialize(void)
{
	INIT_DELAYED_WORK_DEFERRABLE(&cache_cleaner, do_cache_clean);
}

1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622
int cache_register(struct cache_detail *cd)
{
	int ret;

	sunrpc_init_cache_detail(cd);
	ret = create_cache_proc_entries(cd);
	if (ret)
		sunrpc_destroy_cache_detail(cd);
	return ret;
}
EXPORT_SYMBOL_GPL(cache_register);

void cache_unregister(struct cache_detail *cd)
{
	remove_cache_proc_entries(cd);
	sunrpc_destroy_cache_detail(cd);
}
EXPORT_SYMBOL_GPL(cache_unregister);
1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646

static ssize_t cache_read_pipefs(struct file *filp, char __user *buf,
				 size_t count, loff_t *ppos)
{
	struct cache_detail *cd = RPC_I(filp->f_path.dentry->d_inode)->private;

	return cache_read(filp, buf, count, ppos, cd);
}

static ssize_t cache_write_pipefs(struct file *filp, const char __user *buf,
				  size_t count, loff_t *ppos)
{
	struct cache_detail *cd = RPC_I(filp->f_path.dentry->d_inode)->private;

	return cache_write(filp, buf, count, ppos, cd);
}

static unsigned int cache_poll_pipefs(struct file *filp, poll_table *wait)
{
	struct cache_detail *cd = RPC_I(filp->f_path.dentry->d_inode)->private;

	return cache_poll(filp, wait, cd);
}

1647
static long cache_ioctl_pipefs(struct file *filp,
1648 1649
			      unsigned int cmd, unsigned long arg)
{
1650
	struct inode *inode = filp->f_dentry->d_inode;
1651
	struct cache_detail *cd = RPC_I(inode)->private;
1652
	long ret;
1653

1654 1655 1656 1657 1658
	lock_kernel();
	ret = cache_ioctl(inode, filp, cmd, arg, cd);
	unlock_kernel();

	return ret;
1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680
}

static int cache_open_pipefs(struct inode *inode, struct file *filp)
{
	struct cache_detail *cd = RPC_I(inode)->private;

	return cache_open(inode, filp, cd);
}

static int cache_release_pipefs(struct inode *inode, struct file *filp)
{
	struct cache_detail *cd = RPC_I(inode)->private;

	return cache_release(inode, filp, cd);
}

const struct file_operations cache_file_operations_pipefs = {
	.owner		= THIS_MODULE,
	.llseek		= no_llseek,
	.read		= cache_read_pipefs,
	.write		= cache_write_pipefs,
	.poll		= cache_poll_pipefs,
1681
	.unlocked_ioctl	= cache_ioctl_pipefs, /* for FIONREAD */
1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692
	.open		= cache_open_pipefs,
	.release	= cache_release_pipefs,
};

static int content_open_pipefs(struct inode *inode, struct file *filp)
{
	struct cache_detail *cd = RPC_I(inode)->private;

	return content_open(inode, filp, cd);
}

1693 1694 1695 1696 1697 1698 1699
static int content_release_pipefs(struct inode *inode, struct file *filp)
{
	struct cache_detail *cd = RPC_I(inode)->private;

	return content_release(inode, filp, cd);
}

1700 1701 1702 1703
const struct file_operations content_file_operations_pipefs = {
	.open		= content_open_pipefs,
	.read		= seq_read,
	.llseek		= seq_lseek,
1704
	.release	= content_release_pipefs,
1705 1706
};

1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720
static int open_flush_pipefs(struct inode *inode, struct file *filp)
{
	struct cache_detail *cd = RPC_I(inode)->private;

	return open_flush(inode, filp, cd);
}

static int release_flush_pipefs(struct inode *inode, struct file *filp)
{
	struct cache_detail *cd = RPC_I(inode)->private;

	return release_flush(inode, filp, cd);
}

1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738
static ssize_t read_flush_pipefs(struct file *filp, char __user *buf,
			    size_t count, loff_t *ppos)
{
	struct cache_detail *cd = RPC_I(filp->f_path.dentry->d_inode)->private;

	return read_flush(filp, buf, count, ppos, cd);
}

static ssize_t write_flush_pipefs(struct file *filp,
				  const char __user *buf,
				  size_t count, loff_t *ppos)
{
	struct cache_detail *cd = RPC_I(filp->f_path.dentry->d_inode)->private;

	return write_flush(filp, buf, count, ppos, cd);
}

const struct file_operations cache_flush_operations_pipefs = {
1739
	.open		= open_flush_pipefs,
1740 1741
	.read		= read_flush_pipefs,
	.write		= write_flush_pipefs,
1742
	.release	= release_flush_pipefs,
1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775
};

int sunrpc_cache_register_pipefs(struct dentry *parent,
				 const char *name, mode_t umode,
				 struct cache_detail *cd)
{
	struct qstr q;
	struct dentry *dir;
	int ret = 0;

	sunrpc_init_cache_detail(cd);
	q.name = name;
	q.len = strlen(name);
	q.hash = full_name_hash(q.name, q.len);
	dir = rpc_create_cache_dir(parent, &q, umode, cd);
	if (!IS_ERR(dir))
		cd->u.pipefs.dir = dir;
	else {
		sunrpc_destroy_cache_detail(cd);
		ret = PTR_ERR(dir);
	}
	return ret;
}
EXPORT_SYMBOL_GPL(sunrpc_cache_register_pipefs);

void sunrpc_cache_unregister_pipefs(struct cache_detail *cd)
{
	rpc_remove_cache_dir(cd->u.pipefs.dir);
	cd->u.pipefs.dir = NULL;
	sunrpc_destroy_cache_detail(cd);
}
EXPORT_SYMBOL_GPL(sunrpc_cache_unregister_pipefs);