summaryrefslogtreecommitdiffstats
path: root/fs/nfsd/nfscache.c
blob: 6040da8830ffa629adf878458f05ebaca174e3fe (plain)
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
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
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
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
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
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
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
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
/*
 * Request reply cache. This is currently a global cache, but this may
 * change in the future and be a per-client cache.
 *
 * This code is heavily inspired by the 44BSD implementation, although
 * it does things a bit differently.
 *
 * Copyright (C) 1995, 1996 Olaf Kirch <okir@monad.swb.de>
 */

#include <linux/slab.h>
#include <linux/sunrpc/addr.h>
#include <linux/highmem.h>
#include <linux/log2.h>
#include <linux/hash.h>
#include <net/checksum.h>

#include "nfsd.h"
#include "cache.h"

#define NFSDDBG_FACILITY	NFSDDBG_REPCACHE

/*
 * We use this value to determine the number of hash buckets from the max
 * cache size, the idea being that when the cache is at its maximum number
 * of entries, then this should be the average number of entries per bucket.
 */
#define TARGET_BUCKET_SIZE	64

static struct hlist_head *	cache_hash;
static struct list_head 	lru_head;
static struct kmem_cache	*drc_slab;

/* max number of entries allowed in the cache */
static unsigned int		max_drc_entries;

/* number of significant bits in the hash value */
static unsigned int		maskbits;

/*
 * Stats and other tracking of on the duplicate reply cache. All of these and
 * the "rc" fields in nfsdstats are protected by the cache_lock
 */

/* total number of entries */
static unsigned int		num_drc_entries;

/* cache misses due only to checksum comparison failures */
static unsigned int		payload_misses;

/* amount of memory (in bytes) currently consumed by the DRC */
static unsigned int		drc_mem_usage;

/* longest hash chain seen */
static unsigned int		longest_chain;

/* size of cache when we saw the longest hash chain */
static unsigned int		longest_chain_cachesize;

static int	nfsd_cache_append(struct svc_rqst *rqstp, struct kvec *vec);
static void	cache_cleaner_func(struct work_struct *unused);
static unsigned long nfsd_reply_cache_count(struct shrinker *shrink,
					    struct shrink_control *sc);
static unsigned long nfsd_reply_cache_scan(struct shrinker *shrink,
					   struct shrink_control *sc);

static struct shrinker nfsd_reply_cache_shrinker = {
	.scan_objects = nfsd_reply_cache_scan,
	.count_objects = nfsd_reply_cache_count,
	.seeks	= 1,
};

/*
 * locking for the reply cache:
 * A cache entry is "single use" if c_state == RC_INPROG
 * Otherwise, it when accessing _prev or _next, the lock must be held.
 */
static DEFINE_SPINLOCK(cache_lock);
static DECLARE_DELAYED_WORK(cache_cleaner, cache_cleaner_func);

/*
 * Put a cap on the size of the DRC based on the amount of available
 * low memory in the machine.
 *
 *  64MB:    8192
 * 128MB:   11585
 * 256MB:   16384
 * 512MB:   23170
 *   1GB:   32768
 *   2GB:   46340
 *   4GB:   65536
 *   8GB:   92681
 *  16GB:  131072
 *
 * ...with a hard cap of 256k entries. In the worst case, each entry will be
 * ~1k, so the above numbers should give a rough max of the amount of memory
 * used in k.
 */
static unsigned int
nfsd_cache_size_limit(void)
{
	unsigned int limit;
	unsigned long low_pages = totalram_pages - totalhigh_pages;

	limit = (16 * int_sqrt(low_pages)) << (PAGE_SHIFT-10);
	return min_t(unsigned int, limit, 256*1024);
}

/*
 * Compute the number of hash buckets we need. Divide the max cachesize by
 * the "target" max bucket size, and round up to next power of two.
 */
static unsigned int
nfsd_hashsize(unsigned int limit)
{
	return roundup_pow_of_two(limit / TARGET_BUCKET_SIZE);
}

static struct svc_cacherep *
nfsd_reply_cache_alloc(void)
{
	struct svc_cacherep	*rp;

	rp = kmem_cache_alloc(drc_slab, GFP_KERNEL);
	if (rp) {
		rp->c_state = RC_UNUSED;
		rp->c_type = RC_NOCACHE;
		INIT_LIST_HEAD(&rp->c_lru);
		INIT_HLIST_NODE(&rp->c_hash);
	}
	return rp;
}

static void
nfsd_reply_cache_free_locked(struct svc_cacherep *rp)
{
	if (rp->c_type == RC_REPLBUFF && rp->c_replvec.iov_base) {
		drc_mem_usage -= rp->c_replvec.iov_len;
		kfree(rp->c_replvec.iov_base);
	}
	if (!hlist_unhashed(&rp->c_hash))
		hlist_del(&rp->c_hash);
	list_del(&rp->c_lru);
	--num_drc_entries;
	drc_mem_usage -= sizeof(*rp);
	kmem_cache_free(drc_slab, rp);
}

static void
nfsd_reply_cache_free(struct svc_cacherep *rp)
{
	spin_lock(&cache_lock);
	nfsd_reply_cache_free_locked(rp);
	spin_unlock(&cache_lock);
}

int nfsd_reply_cache_init(void)
{
	unsigned int hashsize;

	INIT_LIST_HEAD(&lru_head);
	max_drc_entries = nfsd_cache_size_limit();
	num_drc_entries = 0;
	hashsize = nfsd_hashsize(max_drc_entries);
	maskbits = ilog2(hashsize);

	register_shrinker(&nfsd_reply_cache_shrinker);
	drc_slab = kmem_cache_create("nfsd_drc", sizeof(struct svc_cacherep),
					0, 0, NULL);
	if (!drc_slab)
		goto out_nomem;

	cache_hash = kcalloc(hashsize, sizeof(struct hlist_head), GFP_KERNEL);
	if (!cache_hash)
		goto out_nomem;

	return 0;
out_nomem:
	printk(KERN_ERR "nfsd: failed to allocate reply cache\n");
	nfsd_reply_cache_shutdown();
	return -ENOMEM;
}

void nfsd_reply_cache_shutdown(void)
{
	struct svc_cacherep	*rp;

	unregister_shrinker(&nfsd_reply_cache_shrinker);
	cancel_delayed_work_sync(&cache_cleaner);

	while (!list_empty(&lru_head)) {
		rp = list_entry(lru_head.next, struct svc_cacherep, c_lru);
		nfsd_reply_cache_free_locked(rp);
	}

	kfree (cache_hash);
	cache_hash = NULL;

	if (drc_slab) {
		kmem_cache_destroy(drc_slab);
		drc_slab = NULL;
	}
}

/*
 * Move cache entry to end of LRU list, and queue the cleaner to run if it's
 * not already scheduled.
 */
static void
lru_put_end(struct svc_cacherep *rp)
{
	rp->c_timestamp = jiffies;
	list_move_tail(&rp->c_lru, &lru_head);
	schedule_delayed_work(&cache_cleaner, RC_EXPIRE);
}

/*
 * Move a cache entry from one hash list to another
 */
static void
hash_refile(struct svc_cacherep *rp)
{
	hlist_del_init(&rp->c_hash);
	hlist_add_head(&rp->c_hash, cache_hash + hash_32(rp->c_xid, maskbits));
}

/*
 * Walk the LRU list and prune off entries that are older than RC_EXPIRE.
 * Also prune the oldest ones when the total exceeds the max number of entries.
 */
static long
prune_cache_entries(void)
{
	struct svc_cacherep *rp, *tmp;
	long freed = 0;

	list_for_each_entry_safe(rp, tmp, &lru_head, c_lru) {
		/*
		 * Don't free entries attached to calls that are still
		 * in-progress, but do keep scanning the list.
		 */
		if (rp->c_state == RC_INPROG)
			continue;
		if (num_drc_entries <= max_drc_entries &&
		    time_before(jiffies, rp->c_timestamp + RC_EXPIRE))
			break;
		nfsd_reply_cache_free_locked(rp);
		freed++;
	}

	/*
	 * Conditionally rearm the job. If we cleaned out the list, then
	 * cancel any pending run (since there won't be any work to do).
	 * Otherwise, we rearm the job or modify the existing one to run in
	 * RC_EXPIRE since we just ran the pruner.
	 */
	if (list_empty(&lru_head))
		cancel_delayed_work(&cache_cleaner);
	else
		mod_delayed_work(system_wq, &cache_cleaner, RC_EXPIRE);
	return freed;
}

static void
cache_cleaner_func(struct work_struct *unused)
{
	spin_lock(&cache_lock);
	prune_cache_entries();
	spin_unlock(&cache_lock);
}

static unsigned long
nfsd_reply_cache_count(struct shrinker *shrink, struct shrink_control *sc)
{
	unsigned long num;

	spin_lock(&cache_lock);
	num = num_drc_entries;
	spin_unlock(&cache_lock);

	return num;
}

static unsigned long
nfsd_reply_cache_scan(struct shrinker *shrink, struct shrink_control *sc)
{
	unsigned long freed;

	spin_lock(&cache_lock);
	freed = prune_cache_entries();
	spin_unlock(&cache_lock);
	return freed;
}
/*
 * Walk an xdr_buf and get a CRC for at most the first RC_CSUMLEN bytes
 */
static __wsum
nfsd_cache_csum(struct svc_rqst *rqstp)
{
	int idx;
	unsigned int base;
	__wsum csum;
	struct xdr_buf *buf = &rqstp->rq_arg;
	const unsigned char *p = buf->head[0].iov_base;
	size_t csum_len = min_t(size_t, buf->head[0].iov_len + buf->page_len,
				RC_CSUMLEN);
	size_t len = min(buf->head[0].iov_len, csum_len);

	/* rq_arg.head first */
	csum = csum_partial(p, len, 0);
	csum_len -= len;

	/* Continue into page array */
	idx = buf->page_base / PAGE_SIZE;
	base = buf->page_base & ~PAGE_MASK;
	while (csum_len) {
		p = page_address(buf->pages[idx]) + base;
		len = min_t(size_t, PAGE_SIZE - base, csum_len);
		csum = csum_partial(p, len, csum);
		csum_len -= len;
		base = 0;
		++idx;
	}
	return csum;
}

static bool
nfsd_cache_match(struct svc_rqst *rqstp, __wsum csum, struct svc_cacherep *rp)
{
	/* Check RPC header info first */
	if (rqstp->rq_xid != rp->c_xid || rqstp->rq_proc != rp->c_proc ||
	    rqstp->rq_prot != rp->c_prot || rqstp->rq_vers != rp->c_vers ||
	    rqstp->rq_arg.len != rp->c_len ||
	    !rpc_cmp_addr(svc_addr(rqstp), (struct sockaddr *)&rp->c_addr) ||
	    rpc_get_port(svc_addr(rqstp)) != rpc_get_port((struct sockaddr *)&rp->c_addr))
		return false;

	/* compare checksum of NFS data */
	if (csum != rp->c_csum) {
		++payload_misses;
		return false;
	}

	return true;
}

/*
 * Search the request hash for an entry that matches the given rqstp.
 * Must be called with cache_lock held. Returns the found entry or
 * NULL on failure.
 */
static struct svc_cacherep *
nfsd_cache_search(struct svc_rqst *rqstp, __wsum csum)
{
	struct svc_cacherep	*rp, *ret = NULL;
	struct hlist_head 	*rh;
	unsigned int		entries = 0;

	rh = &cache_hash[hash_32(rqstp->rq_xid, maskbits)];
	hlist_for_each_entry(rp, rh, c_hash) {
		++entries;
		if (nfsd_cache_match(rqstp, csum, rp)) {
			ret = rp;
			break;
		}
	}

	/* tally hash chain length stats */
	if (entries > longest_chain) {
		longest_chain = entries;
		longest_chain_cachesize = num_drc_entries;
	} else if (entries == longest_chain) {
		/* prefer to keep the smallest cachesize possible here */
		longest_chain_cachesize = min(longest_chain_cachesize,
						num_drc_entries);
	}

	return ret;
}

/*
 * Try to find an entry matching the current call in the cache. When none
 * is found, we try to grab the oldest expired entry off the LRU list. If
 * a suitable one isn't there, then drop the cache_lock and allocate a
 * new one, then search again in case one got inserted while this thread
 * didn't hold the lock.
 */
int
nfsd_cache_lookup(struct svc_rqst *rqstp)
{
	struct svc_cacherep	*rp, *found;
	__be32			xid = rqstp->rq_xid;
	u32			proto =  rqstp->rq_prot,
				vers = rqstp->rq_vers,
				proc = rqstp->rq_proc;
	__wsum			csum;
	unsigned long		age;
	int type = rqstp->rq_cachetype;
	int rtn = RC_DOIT;

	rqstp->rq_cacherep = NULL;
	if (type == RC_NOCACHE) {
		nfsdstats.rcnocache++;
		return rtn;
	}

	csum = nfsd_cache_csum(rqstp);

	/*
	 * Since the common case is a cache miss followed by an insert,
	 * preallocate an entry.
	 */
	rp = nfsd_reply_cache_alloc();
	spin_lock(&cache_lock);
	if (likely(rp)) {
		++num_drc_entries;
		drc_mem_usage += sizeof(*rp);
	}

	/* go ahead and prune the cache */
	prune_cache_entries();

	found = nfsd_cache_search(rqstp, csum);
	if (found) {
		if (likely(rp))
			nfsd_reply_cache_free_locked(rp);
		rp = found;
		goto found_entry;
	}

	if (!rp) {
		dprintk("nfsd: unable to allocate DRC entry!\n");
		goto out;
	}

	nfsdstats.rcmisses++;
	rqstp->rq_cacherep = rp;
	rp->c_state = RC_INPROG;
	rp->c_xid = xid;
	rp->c_proc = proc;
	rpc_copy_addr((struct sockaddr *)&rp->c_addr, svc_addr(rqstp));
	rpc_set_port((struct sockaddr *)&rp->c_addr, rpc_get_port(svc_addr(rqstp)));
	rp->c_prot = proto;
	rp->c_vers = vers;
	rp->c_len = rqstp->rq_arg.len;
	rp->c_csum = csum;

	hash_refile(rp);
	lru_put_end(rp);

	/* release any buffer */
	if (rp->c_type == RC_REPLBUFF) {
		drc_mem_usage -= rp->c_replvec.iov_len;
		kfree(rp->c_replvec.iov_base);
		rp->c_replvec.iov_base = NULL;
	}
	rp->c_type = RC_NOCACHE;
 out:
	spin_unlock(&cache_lock);
	return rtn;

found_entry:
	nfsdstats.rchits++;
	/* We found a matching entry which is either in progress or done. */
	age = jiffies - rp->c_timestamp;
	lru_put_end(rp);

	rtn = RC_DROPIT;
	/* Request being processed or excessive rexmits */
	if (rp->c_state == RC_INPROG || age < RC_DELAY)
		goto out;

	/* From the hall of fame of impractical attacks:
	 * Is this a user who tries to snoop on the cache? */
	rtn = RC_DOIT;
	if (!rqstp->rq_secure && rp->c_secure)
		goto out;

	/* Compose RPC reply header */
	switch (rp->c_type) {
	case RC_NOCACHE:
		break;
	case RC_REPLSTAT:
		svc_putu32(&rqstp->rq_res.head[0], rp->c_replstat);
		rtn = RC_REPLY;
		break;
	case RC_REPLBUFF:
		if (!nfsd_cache_append(rqstp, &rp->c_replvec))
			goto out;	/* should not happen */
		rtn = RC_REPLY;
		break;
	default:
		printk(KERN_WARNING "nfsd: bad repcache type %d\n", rp->c_type);
		nfsd_reply_cache_free_locked(rp);
	}

	goto out;
}

/*
 * Update a cache entry. This is called from nfsd_dispatch when
 * the procedure has been executed and the complete reply is in
 * rqstp->rq_res.
 *
 * We're copying around data here rather than swapping buffers because
 * the toplevel loop requires max-sized buffers, which would be a waste
 * of memory for a cache with a max reply size of 100 bytes (diropokres).
 *
 * If we should start to use different types of cache entries tailored
 * specifically for attrstat and fh's, we may save even more space.
 *
 * Also note that a cachetype of RC_NOCACHE can legally be passed when
 * nfsd failed to encode a reply that otherwise would have been cached.
 * In this case, nfsd_cache_update is called with statp == NULL.
 */
void
nfsd_cache_update(struct svc_rqst *rqstp, int cachetype, __be32 *statp)
{
	struct svc_cacherep *rp = rqstp->rq_cacherep;
	struct kvec	*resv = &rqstp->rq_res.head[0], *cachv;
	int		len;
	size_t		bufsize = 0;

	if (!rp)
		return;

	len = resv->iov_len - ((char*)statp - (char*)resv->iov_base);
	len >>= 2;

	/* Don't cache excessive amounts of data and XDR failures */
	if (!statp || len > (256 >> 2)) {
		nfsd_reply_cache_free(rp);
		return;
	}

	switch (cachetype) {
	case RC_REPLSTAT:
		if (len != 1)
			printk("nfsd: RC_REPLSTAT/reply len %d!\n",len);
		rp->c_replstat = *statp;
		break;
	case RC_REPLBUFF:
		cachv = &rp->c_replvec;
		bufsize = len << 2;
		cachv->iov_base = kmalloc(bufsize, GFP_KERNEL);
		if (!cachv->iov_base) {
			nfsd_reply_cache_free(rp);
			return;
		}
		cachv->iov_len = bufsize;
		memcpy(cachv->iov_base, statp, bufsize);
		break;
	case RC_NOCACHE:
		nfsd_reply_cache_free(rp);
		return;
	}
	spin_lock(&cache_lock);
	drc_mem_usage += bufsize;
	lru_put_end(rp);
	rp->c_secure = rqstp->rq_secure;
	rp->c_type = cachetype;
	rp->c_state = RC_DONE;
	spin_unlock(&cache_lock);
	return;
}

/*
 * Copy cached reply to current reply buffer. Should always fit.
 * FIXME as reply is in a page, we should just attach the page, and
 * keep a refcount....
 */
static int
nfsd_cache_append(struct svc_rqst *rqstp, struct kvec *data)
{
	struct kvec	*vec = &rqstp->rq_res.head[0];

	if (vec->iov_len + data->iov_len > PAGE_SIZE) {
		printk(KERN_WARNING "nfsd: cached reply too large (%Zd).\n",
				data->iov_len);
		return 0;
	}
	memcpy((char*)vec->iov_base + vec->iov_len, data->iov_base, data->iov_len);
	vec->iov_len += data->iov_len;
	return 1;
}

/*
 * Note that fields may be added, removed or reordered in the future. Programs
 * scraping this file for info should test the labels to ensure they're
 * getting the correct field.
 */
static int nfsd_reply_cache_stats_show(struct seq_file *m, void *v)
{
	spin_lock(&cache_lock);
	seq_printf(m, "max entries:           %u\n", max_drc_entries);
	seq_printf(m, "num entries:           %u\n", num_drc_entries);
	seq_printf(m, "hash buckets:          %u\n", 1 << maskbits);
	seq_printf(m, "mem usage:             %u\n", drc_mem_usage);
	seq_printf(m, "cache hits:            %u\n", nfsdstats.rchits);
	seq_printf(m, "cache misses:          %u\n", nfsdstats.rcmisses);
	seq_printf(m, "not cached:            %u\n", nfsdstats.rcnocache);
	seq_printf(m, "payload misses:        %u\n", payload_misses);
	seq_printf(m, "longest chain len:     %u\n", longest_chain);
	seq_printf(m, "cachesize at longest:  %u\n", longest_chain_cachesize);
	spin_unlock(&cache_lock);
	return 0;
}

int nfsd_reply_cache_stats_open(struct inode *inode, struct file *file)
{
	return single_open(file, nfsd_reply_cache_stats_show, NULL);
}