Home | History | Annotate | Line # | Download | only in kern
vfs_dirhash.c revision 1.1.8.1
      1 /* $NetBSD: vfs_dirhash.c,v 1.1.8.1 2009/01/19 13:19:40 skrll Exp $ */
      2 
      3 /*
      4  * Copyright (c) 2008 Reinoud Zandijk
      5  * All rights reserved.
      6  *
      7  * Redistribution and use in source and binary forms, with or without
      8  * modification, are permitted provided that the following conditions
      9  * are met:
     10  * 1. Redistributions of source code must retain the above copyright
     11  *    notice, this list of conditions and the following disclaimer.
     12  * 2. Redistributions in binary form must reproduce the above copyright
     13  *    notice, this list of conditions and the following disclaimer in the
     14  *    documentation and/or other materials provided with the distribution.
     15  *
     16  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
     17  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
     18  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
     19  * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
     20  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
     21  * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
     22  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
     23  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
     24  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
     25  * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
     26  *
     27  */
     28 
     29 
     30 #include <sys/cdefs.h>
     31 __KERNEL_RCSID(0, "$NetBSD: vfs_dirhash.c,v 1.1.8.1 2009/01/19 13:19:40 skrll Exp $");
     32 
     33 /* CLEAN UP! */
     34 #include <sys/param.h>
     35 #include <sys/kernel.h>
     36 #include <sys/buf.h>
     37 #include <sys/dirent.h>
     38 #include <sys/hash.h>
     39 #include <sys/mutex.h>
     40 #include <sys/pool.h>
     41 #include <sys/queue.h>
     42 #include <sys/vnode.h>
     43 #include <sys/sysctl.h>
     44 
     45 #include <sys/dirhash.h>
     46 
     47 #if 1
     48 #	define DPRINTF(a) ;
     49 #else
     50 #	define DPRINTF(a) printf a;
     51 #endif
     52 
     53 /*
     54  * The locking protocol of the dirhash structures is fairly simple:
     55  *
     56  * The global dirhash_queue is protected by the dirhashmutex. This lock is
     57  * internal only and is FS/mountpoint/vnode independent. On exit of the
     58  * exported functions this mutex is not helt.
     59  *
     60  * The dirhash structure is considered part of the vnode/inode/udf_node
     61  * structure and will thus use the lock that protects that vnode/inode.
     62  *
     63  * The dirhash entries are considered part of the dirhash structure and thus
     64  * are on the same lock.
     65  */
     66 
     67 static struct sysctllog *sysctl_log;
     68 static struct pool dirhash_pool;
     69 static struct pool dirhash_entry_pool;
     70 
     71 static kmutex_t dirhashmutex;
     72 static uint32_t maxdirhashsize = DIRHASH_SIZE;
     73 static uint32_t dirhashsize    = 0;
     74 static TAILQ_HEAD(_dirhash, dirhash) dirhash_queue;
     75 
     76 
     77 void
     78 dirhash_init(void)
     79 {
     80 	const struct sysctlnode *rnode, *cnode;
     81 	size_t sz;
     82 	uint32_t max_entries;
     83 
     84 	/* initialise dirhash queue */
     85 	TAILQ_INIT(&dirhash_queue);
     86 
     87 	/* init dirhash pools */
     88 	sz = sizeof(struct dirhash);
     89 	pool_init(&dirhash_pool, sz, 0, 0, 0,
     90 		"dirhpl", NULL, IPL_NONE);
     91 
     92 	sz = sizeof(struct dirhash_entry);
     93 	pool_init(&dirhash_entry_pool, sz, 0, 0, 0,
     94 		"dirhepl", NULL, IPL_NONE);
     95 
     96 	mutex_init(&dirhashmutex, MUTEX_DEFAULT, IPL_NONE);
     97 	max_entries = maxdirhashsize / sz;
     98 	pool_sethiwat(&dirhash_entry_pool, max_entries);
     99 	dirhashsize = 0;
    100 
    101 	/* create sysctl knobs and dials */
    102 	sysctl_log = NULL;
    103 	sysctl_createv(&sysctl_log, 0, NULL, &rnode,
    104 		       CTLFLAG_PERMANENT,
    105 		       CTLTYPE_NODE, "dirhash", NULL,
    106 		       NULL, 0, NULL, 0,
    107 		       CTL_VFS, VFS_GENERIC, CTL_CREATE, CTL_EOL);
    108 	sysctl_createv(&sysctl_log, 0, &rnode, &cnode,
    109 		       CTLFLAG_PERMANENT,
    110 		       CTLTYPE_INT, "memused",
    111 		       SYSCTL_DESCR("current dirhash memory usage"),
    112 		       NULL, 0, &dirhashsize, 0,
    113 		       CTL_CREATE, CTL_EOL);
    114 	sysctl_createv(&sysctl_log, 0, &rnode, &cnode,
    115 		       CTLFLAG_PERMANENT | CTLFLAG_READWRITE,
    116 		       CTLTYPE_INT, "maxmem",
    117 		       SYSCTL_DESCR("maximum dirhash memory usage"),
    118 		       NULL, 0, &maxdirhashsize, 0,
    119 		       CTL_CREATE, CTL_EOL);
    120 }
    121 
    122 
    123 #if 0
    124 void
    125 dirhash_finish(void)
    126 {
    127 	pool_destroy(&dirhash_pool);
    128 	pool_destroy(&dirhash_entry_pool);
    129 
    130 	mutex_destroy(&dirhashmutex);
    131 
    132 	/* sysctl_teardown(&sysctl_log); */
    133 }
    134 #endif
    135 
    136 
    137 /*
    138  * generic dirhash implementation
    139  */
    140 
    141 void
    142 dirhash_purge_entries(struct dirhash *dirh)
    143 {
    144 	struct dirhash_entry *dirh_e;
    145 	uint32_t hashline;
    146 
    147 	if (dirh == NULL)
    148 		return;
    149 
    150 	if (dirh->size == 0)
    151 		return;
    152 
    153 	for (hashline = 0; hashline < DIRHASH_HASHSIZE; hashline++) {
    154 		dirh_e = LIST_FIRST(&dirh->entries[hashline]);
    155 		while (dirh_e) {
    156 			LIST_REMOVE(dirh_e, next);
    157 			pool_put(&dirhash_entry_pool, dirh_e);
    158 			dirh_e = LIST_FIRST(&dirh->entries[hashline]);
    159 		}
    160 	}
    161 	dirh_e = LIST_FIRST(&dirh->free_entries);
    162 
    163 	while (dirh_e) {
    164 		LIST_REMOVE(dirh_e, next);
    165 		pool_put(&dirhash_entry_pool, dirh_e);
    166 		dirh_e = LIST_FIRST(&dirh->entries[hashline]);
    167 	}
    168 
    169 	dirh->flags &= ~DIRH_COMPLETE;
    170 	dirh->flags |=  DIRH_PURGED;
    171 
    172 	dirhashsize -= dirh->size;
    173 	dirh->size = 0;
    174 }
    175 
    176 
    177 void
    178 dirhash_purge(struct dirhash **dirhp)
    179 {
    180 	struct dirhash *dirh = *dirhp;
    181 
    182 	if (dirh == NULL)
    183 		return;
    184 
    185 	/* purge its entries */
    186 	dirhash_purge_entries(dirh);
    187 
    188 	/* recycle */
    189 	mutex_enter(&dirhashmutex);
    190 	TAILQ_REMOVE(&dirhash_queue, dirh, next);
    191 	mutex_exit(&dirhashmutex);
    192 
    193 	pool_put(&dirhash_pool, dirh);
    194 	*dirhp = NULL;
    195 }
    196 
    197 
    198 void
    199 dirhash_get(struct dirhash **dirhp)
    200 {
    201 	struct dirhash *dirh;
    202 	uint32_t hashline;
    203 
    204 	/* if no dirhash was given, allocate one */
    205 	dirh = *dirhp;
    206 	if (dirh == NULL) {
    207 		dirh = pool_get(&dirhash_pool, PR_WAITOK);
    208 		memset(dirh, 0, sizeof(struct dirhash));
    209 		for (hashline = 0; hashline < DIRHASH_HASHSIZE; hashline++) {
    210 			LIST_INIT(&dirh->entries[hashline]);
    211 		}
    212 	}
    213 
    214 	/* implement LRU on the dirhash queue */
    215 	mutex_enter(&dirhashmutex);
    216 	if (*dirhp) {
    217 		/* remove from queue to be requeued */
    218 		TAILQ_REMOVE(&dirhash_queue, dirh, next);
    219 	}
    220 	dirh->refcnt++;
    221 	TAILQ_INSERT_HEAD(&dirhash_queue, dirh, next);
    222 	mutex_exit(&dirhashmutex);
    223 
    224 	*dirhp = dirh;
    225 }
    226 
    227 
    228 void
    229 dirhash_put(struct dirhash *dirh)
    230 {
    231 
    232 	mutex_enter(&dirhashmutex);
    233 	dirh->refcnt--;
    234 	mutex_exit(&dirhashmutex);
    235 }
    236 
    237 
    238 void
    239 dirhash_enter(struct dirhash *dirh,
    240 	struct dirent *dirent, uint64_t offset, uint32_t entry_size, int new)
    241 {
    242 	struct dirhash *del_dirh, *prev_dirh;
    243 	struct dirhash_entry *dirh_e;
    244 	uint32_t hashvalue, hashline;
    245 	int entrysize;
    246 
    247 	/* make sure we have a dirhash to work on */
    248 	KASSERT(dirh);
    249 	KASSERT(dirh->refcnt > 0);
    250 
    251 	/* are we trying to re-enter an entry? */
    252 	if (!new && (dirh->flags & DIRH_COMPLETE))
    253 		return;
    254 
    255 	/* calculate our hash */
    256 	hashvalue = hash32_strn(dirent->d_name, dirent->d_namlen, HASH32_STR_INIT);
    257 	hashline  = hashvalue & DIRHASH_HASHMASK;
    258 
    259 	/* lookup and insert entry if not there yet */
    260 	LIST_FOREACH(dirh_e, &dirh->entries[hashline], next) {
    261 		/* check for hash collision */
    262 		if (dirh_e->hashvalue != hashvalue)
    263 			continue;
    264 		if (dirh_e->offset != offset)
    265 			continue;
    266 		/* got it already */
    267 		KASSERT(dirh_e->d_namlen == dirent->d_namlen);
    268 		KASSERT(dirh_e->entry_size == entry_size);
    269 		return;
    270 	}
    271 
    272 	DPRINTF(("dirhash enter %"PRIu64", %d, %d for `%*.*s`\n",
    273 		offset, entry_size, dirent->d_namlen,
    274 		dirent->d_namlen, dirent->d_namlen, dirent->d_name));
    275 
    276 	/* check if entry is in free space list */
    277 	LIST_FOREACH(dirh_e, &dirh->free_entries, next) {
    278 		if (dirh_e->offset == offset) {
    279 			DPRINTF(("\tremoving free entry\n"));
    280 			LIST_REMOVE(dirh_e, next);
    281 			pool_put(&dirhash_entry_pool, dirh_e);
    282 			break;
    283 		}
    284 	}
    285 
    286 	/* ensure we are not passing the dirhash limit */
    287 	entrysize = sizeof(struct dirhash_entry);
    288 	if (dirhashsize + entrysize > maxdirhashsize) {
    289 		/* we walk the dirhash_queue, so need to lock it */
    290 		mutex_enter(&dirhashmutex);
    291 		del_dirh = TAILQ_LAST(&dirhash_queue, _dirhash);
    292 		KASSERT(del_dirh);
    293 		while (dirhashsize + entrysize > maxdirhashsize) {
    294 			/* no use trying to delete myself */
    295 			if (del_dirh == dirh)
    296 				break;
    297 			prev_dirh = TAILQ_PREV(del_dirh, _dirhash, next);
    298 			if (del_dirh->refcnt == 0)
    299 				dirhash_purge_entries(del_dirh);
    300 			del_dirh = prev_dirh;
    301 		}
    302 		mutex_exit(&dirhashmutex);
    303 	}
    304 
    305 	/* add to the hashline */
    306 	dirh_e = pool_get(&dirhash_entry_pool, PR_WAITOK);
    307 	memset(dirh_e, 0, sizeof(struct dirhash_entry));
    308 
    309 	dirh_e->hashvalue = hashvalue;
    310 	dirh_e->offset    = offset;
    311 	dirh_e->d_namlen  = dirent->d_namlen;
    312 	dirh_e->entry_size  = entry_size;
    313 
    314 	dirh->size  += sizeof(struct dirhash_entry);
    315 	dirhashsize += sizeof(struct dirhash_entry);
    316 	LIST_INSERT_HEAD(&dirh->entries[hashline], dirh_e, next);
    317 }
    318 
    319 
    320 void
    321 dirhash_enter_freed(struct dirhash *dirh, uint64_t offset,
    322 	uint32_t entry_size)
    323 {
    324 	struct dirhash_entry *dirh_e;
    325 
    326 	/* make sure we have a dirhash to work on */
    327 	KASSERT(dirh);
    328 	KASSERT(dirh->refcnt > 0);
    329 
    330 	/* check for double entry of free space */
    331 	LIST_FOREACH(dirh_e, &dirh->free_entries, next) {
    332 		KASSERT(dirh_e->offset != offset);
    333 	}
    334 
    335 	DPRINTF(("dirhash enter FREED %"PRIu64", %d\n",
    336 		offset, entry_size));
    337 	dirh_e = pool_get(&dirhash_entry_pool, PR_WAITOK);
    338 	memset(dirh_e, 0, sizeof(struct dirhash_entry));
    339 
    340 	dirh_e->hashvalue = 0;		/* not relevant */
    341 	dirh_e->offset    = offset;
    342 	dirh_e->d_namlen  = 0;		/* not relevant */
    343 	dirh_e->entry_size  = entry_size;
    344 
    345 	/* XXX it might be preferable to append them at the tail */
    346 	LIST_INSERT_HEAD(&dirh->free_entries, dirh_e, next);
    347 	dirh->size  += sizeof(struct dirhash_entry);
    348 	dirhashsize += sizeof(struct dirhash_entry);
    349 }
    350 
    351 
    352 void
    353 dirhash_remove(struct dirhash *dirh, struct dirent *dirent,
    354 	uint64_t offset, uint32_t entry_size)
    355 {
    356 	struct dirhash_entry *dirh_e;
    357 	uint32_t hashvalue, hashline;
    358 
    359 	DPRINTF(("dirhash remove %"PRIu64", %d for `%*.*s`\n",
    360 		offset, entry_size,
    361 		dirent->d_namlen, dirent->d_namlen, dirent->d_name));
    362 
    363 	/* make sure we have a dirhash to work on */
    364 	KASSERT(dirh);
    365 	KASSERT(dirh->refcnt > 0);
    366 
    367 	/* calculate our hash */
    368 	hashvalue = hash32_strn(dirent->d_name, dirent->d_namlen, HASH32_STR_INIT);
    369 	hashline  = hashvalue & DIRHASH_HASHMASK;
    370 
    371 	/* lookup entry */
    372 	LIST_FOREACH(dirh_e, &dirh->entries[hashline], next) {
    373 		/* check for hash collision */
    374 		if (dirh_e->hashvalue != hashvalue)
    375 			continue;
    376 		if (dirh_e->offset != offset)
    377 			continue;
    378 
    379 		/* got it! */
    380 		KASSERT(dirh_e->d_namlen == dirent->d_namlen);
    381 		KASSERT(dirh_e->entry_size == entry_size);
    382 		LIST_REMOVE(dirh_e, next);
    383 		dirh->size -= sizeof(struct dirhash_entry);
    384 		dirhashsize -= sizeof(struct dirhash_entry);
    385 
    386 		dirhash_enter_freed(dirh, offset, entry_size);
    387 		return;
    388 	}
    389 
    390 	/* not found! */
    391 	panic("dirhash_remove couldn't find entry in hash table\n");
    392 }
    393 
    394 
    395 /*
    396  * BUGALERT: don't use result longer than needed, never past the node lock.
    397  * Call with NULL *result initially and it will return nonzero if again.
    398  */
    399 int
    400 dirhash_lookup(struct dirhash *dirh, const char *d_name, int d_namlen,
    401 	struct dirhash_entry **result)
    402 {
    403 	struct dirhash_entry *dirh_e;
    404 	uint32_t hashvalue, hashline;
    405 
    406 	/* make sure we have a dirhash to work on */
    407 	KASSERT(dirh);
    408 	KASSERT(dirh->refcnt > 0);
    409 
    410 	/* start where we were */
    411 	if (*result) {
    412 		dirh_e = *result;
    413 
    414 		/* retrieve information to avoid recalculation and advance */
    415 		hashvalue = dirh_e->hashvalue;
    416 		dirh_e = LIST_NEXT(*result, next);
    417 	} else {
    418 		/* calculate our hash and lookup all entries in hashline */
    419 		hashvalue = hash32_strn(d_name, d_namlen, HASH32_STR_INIT);
    420 		hashline  = hashvalue & DIRHASH_HASHMASK;
    421 		dirh_e = LIST_FIRST(&dirh->entries[hashline]);
    422 	}
    423 
    424 	for (; dirh_e; dirh_e = LIST_NEXT(dirh_e, next)) {
    425 		/* check for hash collision */
    426 		if (dirh_e->hashvalue != hashvalue)
    427 			continue;
    428 		if (dirh_e->d_namlen != d_namlen)
    429 			continue;
    430 		/* might have an entry in the cache */
    431 		*result = dirh_e;
    432 		return 1;
    433 	}
    434 
    435 	*result = NULL;
    436 	return 0;
    437 }
    438 
    439 
    440 /*
    441  * BUGALERT: don't use result longer than needed, never past the node lock.
    442  * Call with NULL *result initially and it will return nonzero if again.
    443  */
    444 
    445 int
    446 dirhash_lookup_freed(struct dirhash *dirh, uint32_t min_entrysize,
    447 	struct dirhash_entry **result)
    448 {
    449 	struct dirhash_entry *dirh_e;
    450 
    451 	/* make sure we have a dirhash to work on */
    452 	KASSERT(dirh);
    453 	KASSERT(dirh->refcnt > 0);
    454 
    455 	/* start where we were */
    456 	if (*result) {
    457 		dirh_e = LIST_NEXT(*result, next);
    458 	} else {
    459 		/* lookup all entries that match */
    460 		dirh_e = LIST_FIRST(&dirh->free_entries);
    461 	}
    462 
    463 	for (; dirh_e; dirh_e = LIST_NEXT(dirh_e, next)) {
    464 		/* check for minimum size */
    465 		if (dirh_e->entry_size < min_entrysize)
    466 			continue;
    467 		/* might be a candidate */
    468 		*result = dirh_e;
    469 		return 1;
    470 	}
    471 
    472 	*result = NULL;
    473 	return 0;
    474 }
    475