ufs_dirhash.c revision 298433
1/*- 2 * Copyright (c) 2001, 2002 Ian Dowse. All rights reserved. 3 * 4 * Redistribution and use in source and binary forms, with or without 5 * modification, are permitted provided that the following conditions 6 * are met: 7 * 1. Redistributions of source code must retain the above copyright 8 * notice, this list of conditions and the following disclaimer. 9 * 2. Redistributions in binary form must reproduce the above copyright 10 * notice, this list of conditions and the following disclaimer in the 11 * documentation and/or other materials provided with the distribution. 12 * 13 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 14 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 15 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 16 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 17 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 18 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 19 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 20 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 21 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 22 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 23 * SUCH DAMAGE. 24 */ 25 26/* 27 * This implements a hash-based lookup scheme for UFS directories. 28 */ 29 30#include <sys/cdefs.h> 31__FBSDID("$FreeBSD: head/sys/ufs/ufs/ufs_dirhash.c 298433 2016-04-21 19:57:40Z pfg $"); 32 33#include "opt_ufs.h" 34 35#ifdef UFS_DIRHASH 36 37#include <sys/param.h> 38#include <sys/systm.h> 39#include <sys/kernel.h> 40#include <sys/lock.h> 41#include <sys/mutex.h> 42#include <sys/malloc.h> 43#include <sys/fnv_hash.h> 44#include <sys/proc.h> 45#include <sys/bio.h> 46#include <sys/buf.h> 47#include <sys/vnode.h> 48#include <sys/mount.h> 49#include <sys/refcount.h> 50#include <sys/sysctl.h> 51#include <sys/sx.h> 52#include <sys/eventhandler.h> 53#include <sys/time.h> 54#include <vm/uma.h> 55 56#include <ufs/ufs/quota.h> 57#include <ufs/ufs/inode.h> 58#include <ufs/ufs/dir.h> 59#include <ufs/ufs/dirhash.h> 60#include <ufs/ufs/extattr.h> 61#include <ufs/ufs/ufsmount.h> 62#include <ufs/ufs/ufs_extern.h> 63 64#define WRAPINCR(val, limit) (((val) + 1 == (limit)) ? 0 : ((val) + 1)) 65#define WRAPDECR(val, limit) (((val) == 0) ? ((limit) - 1) : ((val) - 1)) 66#define OFSFMT(vp) ((vp)->v_mount->mnt_maxsymlinklen <= 0) 67#define BLKFREE2IDX(n) ((n) > DH_NFSTATS ? DH_NFSTATS : (n)) 68 69static MALLOC_DEFINE(M_DIRHASH, "ufs_dirhash", "UFS directory hash tables"); 70 71static int ufs_mindirhashsize = DIRBLKSIZ * 5; 72SYSCTL_INT(_vfs_ufs, OID_AUTO, dirhash_minsize, CTLFLAG_RW, 73 &ufs_mindirhashsize, 74 0, "minimum directory size in bytes for which to use hashed lookup"); 75static int ufs_dirhashmaxmem = 2 * 1024 * 1024; /* NOTE: initial value. It is 76 tuned in ufsdirhash_init() */ 77SYSCTL_INT(_vfs_ufs, OID_AUTO, dirhash_maxmem, CTLFLAG_RW, &ufs_dirhashmaxmem, 78 0, "maximum allowed dirhash memory usage"); 79static int ufs_dirhashmem; 80SYSCTL_INT(_vfs_ufs, OID_AUTO, dirhash_mem, CTLFLAG_RD, &ufs_dirhashmem, 81 0, "current dirhash memory usage"); 82static int ufs_dirhashcheck = 0; 83SYSCTL_INT(_vfs_ufs, OID_AUTO, dirhash_docheck, CTLFLAG_RW, &ufs_dirhashcheck, 84 0, "enable extra sanity tests"); 85static int ufs_dirhashlowmemcount = 0; 86SYSCTL_INT(_vfs_ufs, OID_AUTO, dirhash_lowmemcount, CTLFLAG_RD, 87 &ufs_dirhashlowmemcount, 0, "number of times low memory hook called"); 88static int ufs_dirhashreclaimpercent = 10; 89static int ufsdirhash_set_reclaimpercent(SYSCTL_HANDLER_ARGS); 90SYSCTL_PROC(_vfs_ufs, OID_AUTO, dirhash_reclaimpercent, 91 CTLTYPE_INT | CTLFLAG_RW, 0, 0, ufsdirhash_set_reclaimpercent, "I", 92 "set percentage of dirhash cache to be removed in low VM events"); 93 94 95static int ufsdirhash_hash(struct dirhash *dh, char *name, int namelen); 96static void ufsdirhash_adjfree(struct dirhash *dh, doff_t offset, int diff); 97static void ufsdirhash_delslot(struct dirhash *dh, int slot); 98static int ufsdirhash_findslot(struct dirhash *dh, char *name, int namelen, 99 doff_t offset); 100static doff_t ufsdirhash_getprev(struct direct *dp, doff_t offset); 101static int ufsdirhash_recycle(int wanted); 102static void ufsdirhash_lowmem(void); 103static void ufsdirhash_free_locked(struct inode *ip); 104 105static uma_zone_t ufsdirhash_zone; 106 107#define DIRHASHLIST_LOCK() mtx_lock(&ufsdirhash_mtx) 108#define DIRHASHLIST_UNLOCK() mtx_unlock(&ufsdirhash_mtx) 109#define DIRHASH_BLKALLOC_WAITOK() uma_zalloc(ufsdirhash_zone, M_WAITOK) 110#define DIRHASH_BLKFREE(ptr) uma_zfree(ufsdirhash_zone, (ptr)) 111#define DIRHASH_ASSERT_LOCKED(dh) \ 112 sx_assert(&(dh)->dh_lock, SA_LOCKED) 113 114/* Dirhash list; recently-used entries are near the tail. */ 115static TAILQ_HEAD(, dirhash) ufsdirhash_list; 116 117/* Protects: ufsdirhash_list, `dh_list' field, ufs_dirhashmem. */ 118static struct mtx ufsdirhash_mtx; 119 120/* 121 * Locking: 122 * 123 * The relationship between inode and dirhash is protected either by an 124 * exclusive vnode lock or the vnode interlock where a shared vnode lock 125 * may be used. The dirhash_mtx is acquired after the dirhash lock. To 126 * handle teardown races, code wishing to lock the dirhash for an inode 127 * when using a shared vnode lock must obtain a private reference on the 128 * dirhash while holding the vnode interlock. They can drop it once they 129 * have obtained the dirhash lock and verified that the dirhash wasn't 130 * recycled while they waited for the dirhash lock. 131 * 132 * ufsdirhash_build() acquires a shared lock on the dirhash when it is 133 * successful. This lock is released after a call to ufsdirhash_lookup(). 134 * 135 * Functions requiring exclusive access use ufsdirhash_acquire() which may 136 * free a dirhash structure that was recycled by ufsdirhash_recycle(). 137 * 138 * The dirhash lock may be held across io operations. 139 * 140 * WITNESS reports a lock order reversal between the "bufwait" lock 141 * and the "dirhash" lock. However, this specific reversal will not 142 * cause a deadlock. To get a deadlock, one would have to lock a 143 * buffer followed by the dirhash while a second thread locked a 144 * buffer while holding the dirhash lock. The second order can happen 145 * under a shared or exclusive vnode lock for the associated directory 146 * in lookup(). The first order, however, can only happen under an 147 * exclusive vnode lock (e.g. unlink(), rename(), etc.). Thus, for 148 * a thread to be doing a "bufwait" -> "dirhash" order, it has to hold 149 * an exclusive vnode lock. That exclusive vnode lock will prevent 150 * any other threads from doing a "dirhash" -> "bufwait" order. 151 */ 152 153static void 154ufsdirhash_hold(struct dirhash *dh) 155{ 156 157 refcount_acquire(&dh->dh_refcount); 158} 159 160static void 161ufsdirhash_drop(struct dirhash *dh) 162{ 163 164 if (refcount_release(&dh->dh_refcount)) { 165 sx_destroy(&dh->dh_lock); 166 free(dh, M_DIRHASH); 167 } 168} 169 170/* 171 * Release the lock on a dirhash. 172 */ 173static void 174ufsdirhash_release(struct dirhash *dh) 175{ 176 177 sx_unlock(&dh->dh_lock); 178} 179 180/* 181 * Either acquire an existing hash locked shared or create a new hash and 182 * return it exclusively locked. May return NULL if the allocation fails. 183 * 184 * The vnode interlock is used to protect the i_dirhash pointer from 185 * simultaneous access while only a shared vnode lock is held. 186 */ 187static struct dirhash * 188ufsdirhash_create(struct inode *ip) 189{ 190 struct dirhash *ndh; 191 struct dirhash *dh; 192 struct vnode *vp; 193 194 ndh = dh = NULL; 195 vp = ip->i_vnode; 196 for (;;) { 197 /* Racy check for i_dirhash to prefetch a dirhash structure. */ 198 if (ip->i_dirhash == NULL && ndh == NULL) { 199 ndh = malloc(sizeof *dh, M_DIRHASH, 200 M_NOWAIT | M_ZERO); 201 if (ndh == NULL) 202 return (NULL); 203 refcount_init(&ndh->dh_refcount, 1); 204 205 /* 206 * The DUPOK is to prevent warnings from the 207 * sx_slock() a few lines down which is safe 208 * since the duplicate lock in that case is 209 * the one for this dirhash we are creating 210 * now which has no external references until 211 * after this function returns. 212 */ 213 sx_init_flags(&ndh->dh_lock, "dirhash", SX_DUPOK); 214 sx_xlock(&ndh->dh_lock); 215 } 216 /* 217 * Check i_dirhash. If it's NULL just try to use a 218 * preallocated structure. If none exists loop and try again. 219 */ 220 VI_LOCK(vp); 221 dh = ip->i_dirhash; 222 if (dh == NULL) { 223 ip->i_dirhash = ndh; 224 VI_UNLOCK(vp); 225 if (ndh == NULL) 226 continue; 227 return (ndh); 228 } 229 ufsdirhash_hold(dh); 230 VI_UNLOCK(vp); 231 232 /* Acquire a shared lock on existing hashes. */ 233 sx_slock(&dh->dh_lock); 234 235 /* The hash could've been recycled while we were waiting. */ 236 VI_LOCK(vp); 237 if (ip->i_dirhash != dh) { 238 VI_UNLOCK(vp); 239 ufsdirhash_release(dh); 240 ufsdirhash_drop(dh); 241 continue; 242 } 243 VI_UNLOCK(vp); 244 ufsdirhash_drop(dh); 245 246 /* If the hash is still valid we've succeeded. */ 247 if (dh->dh_hash != NULL) 248 break; 249 /* 250 * If the hash is NULL it has been recycled. Try to upgrade 251 * so we can recreate it. If we fail the upgrade, drop our 252 * lock and try again. 253 */ 254 if (sx_try_upgrade(&dh->dh_lock)) 255 break; 256 sx_sunlock(&dh->dh_lock); 257 } 258 /* Free the preallocated structure if it was not necessary. */ 259 if (ndh) { 260 ufsdirhash_release(ndh); 261 ufsdirhash_drop(ndh); 262 } 263 return (dh); 264} 265 266/* 267 * Acquire an exclusive lock on an existing hash. Requires an exclusive 268 * vnode lock to protect the i_dirhash pointer. hashes that have been 269 * recycled are reclaimed here and NULL is returned. 270 */ 271static struct dirhash * 272ufsdirhash_acquire(struct inode *ip) 273{ 274 struct dirhash *dh; 275 276 ASSERT_VOP_ELOCKED(ip->i_vnode, __FUNCTION__); 277 278 dh = ip->i_dirhash; 279 if (dh == NULL) 280 return (NULL); 281 sx_xlock(&dh->dh_lock); 282 if (dh->dh_hash != NULL) 283 return (dh); 284 ufsdirhash_free_locked(ip); 285 return (NULL); 286} 287 288/* 289 * Acquire exclusively and free the hash pointed to by ip. Works with a 290 * shared or exclusive vnode lock. 291 */ 292void 293ufsdirhash_free(struct inode *ip) 294{ 295 struct dirhash *dh; 296 struct vnode *vp; 297 298 vp = ip->i_vnode; 299 for (;;) { 300 /* Grab a reference on this inode's dirhash if it has one. */ 301 VI_LOCK(vp); 302 dh = ip->i_dirhash; 303 if (dh == NULL) { 304 VI_UNLOCK(vp); 305 return; 306 } 307 ufsdirhash_hold(dh); 308 VI_UNLOCK(vp); 309 310 /* Exclusively lock the dirhash. */ 311 sx_xlock(&dh->dh_lock); 312 313 /* If this dirhash still belongs to this inode, then free it. */ 314 VI_LOCK(vp); 315 if (ip->i_dirhash == dh) { 316 VI_UNLOCK(vp); 317 ufsdirhash_drop(dh); 318 break; 319 } 320 VI_UNLOCK(vp); 321 322 /* 323 * This inode's dirhash has changed while we were 324 * waiting for the dirhash lock, so try again. 325 */ 326 ufsdirhash_release(dh); 327 ufsdirhash_drop(dh); 328 } 329 ufsdirhash_free_locked(ip); 330} 331 332/* 333 * Attempt to build up a hash table for the directory contents in 334 * inode 'ip'. Returns 0 on success, or -1 of the operation failed. 335 */ 336int 337ufsdirhash_build(struct inode *ip) 338{ 339 struct dirhash *dh; 340 struct buf *bp = NULL; 341 struct direct *ep; 342 struct vnode *vp; 343 doff_t bmask, pos; 344 int dirblocks, i, j, memreqd, nblocks, narrays, nslots, slot; 345 346 /* Take care of a decreased sysctl value. */ 347 while (ufs_dirhashmem > ufs_dirhashmaxmem) { 348 if (ufsdirhash_recycle(0) != 0) 349 return (-1); 350 /* Recycled enough memory, so unlock the list. */ 351 DIRHASHLIST_UNLOCK(); 352 } 353 354 /* Check if we can/should use dirhash. */ 355 if (ip->i_size < ufs_mindirhashsize || OFSFMT(ip->i_vnode) || 356 ip->i_effnlink == 0) { 357 if (ip->i_dirhash) 358 ufsdirhash_free(ip); 359 return (-1); 360 } 361 dh = ufsdirhash_create(ip); 362 if (dh == NULL) 363 return (-1); 364 if (dh->dh_hash != NULL) 365 return (0); 366 367 vp = ip->i_vnode; 368 /* Allocate 50% more entries than this dir size could ever need. */ 369 KASSERT(ip->i_size >= DIRBLKSIZ, ("ufsdirhash_build size")); 370 nslots = ip->i_size / DIRECTSIZ(1); 371 nslots = (nslots * 3 + 1) / 2; 372 narrays = howmany(nslots, DH_NBLKOFF); 373 nslots = narrays * DH_NBLKOFF; 374 dirblocks = howmany(ip->i_size, DIRBLKSIZ); 375 nblocks = (dirblocks * 3 + 1) / 2; 376 memreqd = sizeof(*dh) + narrays * sizeof(*dh->dh_hash) + 377 narrays * DH_NBLKOFF * sizeof(**dh->dh_hash) + 378 nblocks * sizeof(*dh->dh_blkfree); 379 DIRHASHLIST_LOCK(); 380 if (memreqd + ufs_dirhashmem > ufs_dirhashmaxmem) { 381 DIRHASHLIST_UNLOCK(); 382 if (memreqd > ufs_dirhashmaxmem / 2) 383 goto fail; 384 /* Try to free some space. */ 385 if (ufsdirhash_recycle(memreqd) != 0) 386 goto fail; 387 /* Enough was freed, and list has been locked. */ 388 } 389 ufs_dirhashmem += memreqd; 390 DIRHASHLIST_UNLOCK(); 391 392 /* Initialise the hash table and block statistics. */ 393 dh->dh_memreq = memreqd; 394 dh->dh_narrays = narrays; 395 dh->dh_hlen = nslots; 396 dh->dh_nblk = nblocks; 397 dh->dh_dirblks = dirblocks; 398 for (i = 0; i < DH_NFSTATS; i++) 399 dh->dh_firstfree[i] = -1; 400 dh->dh_firstfree[DH_NFSTATS] = 0; 401 dh->dh_hused = 0; 402 dh->dh_seqoff = -1; 403 dh->dh_score = DH_SCOREINIT; 404 dh->dh_lastused = time_second; 405 406 /* 407 * Use non-blocking mallocs so that we will revert to a linear 408 * lookup on failure rather than potentially blocking forever. 409 */ 410 dh->dh_hash = malloc(narrays * sizeof(dh->dh_hash[0]), 411 M_DIRHASH, M_NOWAIT | M_ZERO); 412 if (dh->dh_hash == NULL) 413 goto fail; 414 dh->dh_blkfree = malloc(nblocks * sizeof(dh->dh_blkfree[0]), 415 M_DIRHASH, M_NOWAIT); 416 if (dh->dh_blkfree == NULL) 417 goto fail; 418 for (i = 0; i < narrays; i++) { 419 if ((dh->dh_hash[i] = DIRHASH_BLKALLOC_WAITOK()) == NULL) 420 goto fail; 421 for (j = 0; j < DH_NBLKOFF; j++) 422 dh->dh_hash[i][j] = DIRHASH_EMPTY; 423 } 424 for (i = 0; i < dirblocks; i++) 425 dh->dh_blkfree[i] = DIRBLKSIZ / DIRALIGN; 426 bmask = vp->v_mount->mnt_stat.f_iosize - 1; 427 pos = 0; 428 while (pos < ip->i_size) { 429 /* If necessary, get the next directory block. */ 430 if ((pos & bmask) == 0) { 431 if (bp != NULL) 432 brelse(bp); 433 if (UFS_BLKATOFF(vp, (off_t)pos, NULL, &bp) != 0) 434 goto fail; 435 } 436 437 /* Add this entry to the hash. */ 438 ep = (struct direct *)((char *)bp->b_data + (pos & bmask)); 439 if (ep->d_reclen == 0 || ep->d_reclen > 440 DIRBLKSIZ - (pos & (DIRBLKSIZ - 1))) { 441 /* Corrupted directory. */ 442 brelse(bp); 443 goto fail; 444 } 445 if (ep->d_ino != 0) { 446 /* Add the entry (simplified ufsdirhash_add). */ 447 slot = ufsdirhash_hash(dh, ep->d_name, ep->d_namlen); 448 while (DH_ENTRY(dh, slot) != DIRHASH_EMPTY) 449 slot = WRAPINCR(slot, dh->dh_hlen); 450 dh->dh_hused++; 451 DH_ENTRY(dh, slot) = pos; 452 ufsdirhash_adjfree(dh, pos, -DIRSIZ(0, ep)); 453 } 454 pos += ep->d_reclen; 455 } 456 457 if (bp != NULL) 458 brelse(bp); 459 DIRHASHLIST_LOCK(); 460 TAILQ_INSERT_TAIL(&ufsdirhash_list, dh, dh_list); 461 dh->dh_onlist = 1; 462 DIRHASHLIST_UNLOCK(); 463 sx_downgrade(&dh->dh_lock); 464 return (0); 465 466fail: 467 ufsdirhash_free_locked(ip); 468 return (-1); 469} 470 471/* 472 * Free any hash table associated with inode 'ip'. 473 */ 474static void 475ufsdirhash_free_locked(struct inode *ip) 476{ 477 struct dirhash *dh; 478 struct vnode *vp; 479 int i; 480 481 DIRHASH_ASSERT_LOCKED(ip->i_dirhash); 482 483 /* 484 * Clear the pointer in the inode to prevent new threads from 485 * finding the dead structure. 486 */ 487 vp = ip->i_vnode; 488 VI_LOCK(vp); 489 dh = ip->i_dirhash; 490 ip->i_dirhash = NULL; 491 VI_UNLOCK(vp); 492 493 /* 494 * Remove the hash from the list since we are going to free its 495 * memory. 496 */ 497 DIRHASHLIST_LOCK(); 498 if (dh->dh_onlist) 499 TAILQ_REMOVE(&ufsdirhash_list, dh, dh_list); 500 ufs_dirhashmem -= dh->dh_memreq; 501 DIRHASHLIST_UNLOCK(); 502 503 /* 504 * At this point, any waiters for the lock should hold their 505 * own reference on the dirhash structure. They will drop 506 * that reference once they grab the vnode interlock and see 507 * that ip->i_dirhash is NULL. 508 */ 509 sx_xunlock(&dh->dh_lock); 510 511 /* 512 * Handle partially recycled as well as fully constructed hashes. 513 */ 514 if (dh->dh_hash != NULL) { 515 for (i = 0; i < dh->dh_narrays; i++) 516 if (dh->dh_hash[i] != NULL) 517 DIRHASH_BLKFREE(dh->dh_hash[i]); 518 free(dh->dh_hash, M_DIRHASH); 519 if (dh->dh_blkfree != NULL) 520 free(dh->dh_blkfree, M_DIRHASH); 521 } 522 523 /* 524 * Drop the inode's reference to the data structure. 525 */ 526 ufsdirhash_drop(dh); 527} 528 529/* 530 * Find the offset of the specified name within the given inode. 531 * Returns 0 on success, ENOENT if the entry does not exist, or 532 * EJUSTRETURN if the caller should revert to a linear search. 533 * 534 * If successful, the directory offset is stored in *offp, and a 535 * pointer to a struct buf containing the entry is stored in *bpp. If 536 * prevoffp is non-NULL, the offset of the previous entry within 537 * the DIRBLKSIZ-sized block is stored in *prevoffp (if the entry 538 * is the first in a block, the start of the block is used). 539 * 540 * Must be called with the hash locked. Returns with the hash unlocked. 541 */ 542int 543ufsdirhash_lookup(struct inode *ip, char *name, int namelen, doff_t *offp, 544 struct buf **bpp, doff_t *prevoffp) 545{ 546 struct dirhash *dh, *dh_next; 547 struct direct *dp; 548 struct vnode *vp; 549 struct buf *bp; 550 doff_t blkoff, bmask, offset, prevoff, seqoff; 551 int i, slot; 552 int error; 553 554 dh = ip->i_dirhash; 555 KASSERT(dh != NULL && dh->dh_hash != NULL, 556 ("ufsdirhash_lookup: Invalid dirhash %p\n", dh)); 557 DIRHASH_ASSERT_LOCKED(dh); 558 /* 559 * Move this dirhash towards the end of the list if it has a 560 * score higher than the next entry, and acquire the dh_lock. 561 */ 562 DIRHASHLIST_LOCK(); 563 if (TAILQ_NEXT(dh, dh_list) != NULL) { 564 /* 565 * If the new score will be greater than that of the next 566 * entry, then move this entry past it. With both mutexes 567 * held, dh_next won't go away, but its dh_score could 568 * change; that's not important since it is just a hint. 569 */ 570 if ((dh_next = TAILQ_NEXT(dh, dh_list)) != NULL && 571 dh->dh_score >= dh_next->dh_score) { 572 KASSERT(dh->dh_onlist, ("dirhash: not on list")); 573 TAILQ_REMOVE(&ufsdirhash_list, dh, dh_list); 574 TAILQ_INSERT_AFTER(&ufsdirhash_list, dh_next, dh, 575 dh_list); 576 } 577 } 578 /* Update the score. */ 579 if (dh->dh_score < DH_SCOREMAX) 580 dh->dh_score++; 581 582 /* Update last used time. */ 583 dh->dh_lastused = time_second; 584 DIRHASHLIST_UNLOCK(); 585 586 vp = ip->i_vnode; 587 bmask = vp->v_mount->mnt_stat.f_iosize - 1; 588 blkoff = -1; 589 bp = NULL; 590 seqoff = dh->dh_seqoff; 591restart: 592 slot = ufsdirhash_hash(dh, name, namelen); 593 594 if (seqoff != -1) { 595 /* 596 * Sequential access optimisation. seqoff contains the 597 * offset of the directory entry immediately following 598 * the last entry that was looked up. Check if this offset 599 * appears in the hash chain for the name we are looking for. 600 */ 601 for (i = slot; (offset = DH_ENTRY(dh, i)) != DIRHASH_EMPTY; 602 i = WRAPINCR(i, dh->dh_hlen)) 603 if (offset == seqoff) 604 break; 605 if (offset == seqoff) { 606 /* 607 * We found an entry with the expected offset. This 608 * is probably the entry we want, but if not, the 609 * code below will retry. 610 */ 611 slot = i; 612 } else 613 seqoff = -1; 614 } 615 616 for (; (offset = DH_ENTRY(dh, slot)) != DIRHASH_EMPTY; 617 slot = WRAPINCR(slot, dh->dh_hlen)) { 618 if (offset == DIRHASH_DEL) 619 continue; 620 if (offset < 0 || offset >= ip->i_size) 621 panic("ufsdirhash_lookup: bad offset in hash array"); 622 if ((offset & ~bmask) != blkoff) { 623 if (bp != NULL) 624 brelse(bp); 625 blkoff = offset & ~bmask; 626 if (UFS_BLKATOFF(vp, (off_t)blkoff, NULL, &bp) != 0) { 627 error = EJUSTRETURN; 628 goto fail; 629 } 630 } 631 KASSERT(bp != NULL, ("no buffer allocated")); 632 dp = (struct direct *)(bp->b_data + (offset & bmask)); 633 if (dp->d_reclen == 0 || dp->d_reclen > 634 DIRBLKSIZ - (offset & (DIRBLKSIZ - 1))) { 635 /* Corrupted directory. */ 636 error = EJUSTRETURN; 637 goto fail; 638 } 639 if (dp->d_namlen == namelen && 640 bcmp(dp->d_name, name, namelen) == 0) { 641 /* Found. Get the prev offset if needed. */ 642 if (prevoffp != NULL) { 643 if (offset & (DIRBLKSIZ - 1)) { 644 prevoff = ufsdirhash_getprev(dp, 645 offset); 646 if (prevoff == -1) { 647 error = EJUSTRETURN; 648 goto fail; 649 } 650 } else 651 prevoff = offset; 652 *prevoffp = prevoff; 653 } 654 655 /* Update offset. */ 656 dh->dh_seqoff = offset + DIRSIZ(0, dp); 657 *bpp = bp; 658 *offp = offset; 659 ufsdirhash_release(dh); 660 return (0); 661 } 662 663 /* 664 * When the name doesn't match in the sequential 665 * optimization case, go back and search normally. 666 */ 667 if (seqoff != -1) { 668 seqoff = -1; 669 goto restart; 670 } 671 } 672 error = ENOENT; 673fail: 674 ufsdirhash_release(dh); 675 if (bp != NULL) 676 brelse(bp); 677 return (error); 678} 679 680/* 681 * Find a directory block with room for 'slotneeded' bytes. Returns 682 * the offset of the directory entry that begins the free space. 683 * This will either be the offset of an existing entry that has free 684 * space at the end, or the offset of an entry with d_ino == 0 at 685 * the start of a DIRBLKSIZ block. 686 * 687 * To use the space, the caller may need to compact existing entries in 688 * the directory. The total number of bytes in all of the entries involved 689 * in the compaction is stored in *slotsize. In other words, all of 690 * the entries that must be compacted are exactly contained in the 691 * region beginning at the returned offset and spanning *slotsize bytes. 692 * 693 * Returns -1 if no space was found, indicating that the directory 694 * must be extended. 695 */ 696doff_t 697ufsdirhash_findfree(struct inode *ip, int slotneeded, int *slotsize) 698{ 699 struct direct *dp; 700 struct dirhash *dh; 701 struct buf *bp; 702 doff_t pos, slotstart; 703 int dirblock, error, freebytes, i; 704 705 dh = ip->i_dirhash; 706 KASSERT(dh != NULL && dh->dh_hash != NULL, 707 ("ufsdirhash_findfree: Invalid dirhash %p\n", dh)); 708 DIRHASH_ASSERT_LOCKED(dh); 709 710 /* Find a directory block with the desired free space. */ 711 dirblock = -1; 712 for (i = howmany(slotneeded, DIRALIGN); i <= DH_NFSTATS; i++) 713 if ((dirblock = dh->dh_firstfree[i]) != -1) 714 break; 715 if (dirblock == -1) 716 return (-1); 717 718 KASSERT(dirblock < dh->dh_nblk && 719 dh->dh_blkfree[dirblock] >= howmany(slotneeded, DIRALIGN), 720 ("ufsdirhash_findfree: bad stats")); 721 pos = dirblock * DIRBLKSIZ; 722 error = UFS_BLKATOFF(ip->i_vnode, (off_t)pos, (char **)&dp, &bp); 723 if (error) 724 return (-1); 725 726 /* Find the first entry with free space. */ 727 for (i = 0; i < DIRBLKSIZ; ) { 728 if (dp->d_reclen == 0) { 729 brelse(bp); 730 return (-1); 731 } 732 if (dp->d_ino == 0 || dp->d_reclen > DIRSIZ(0, dp)) 733 break; 734 i += dp->d_reclen; 735 dp = (struct direct *)((char *)dp + dp->d_reclen); 736 } 737 if (i > DIRBLKSIZ) { 738 brelse(bp); 739 return (-1); 740 } 741 slotstart = pos + i; 742 743 /* Find the range of entries needed to get enough space */ 744 freebytes = 0; 745 while (i < DIRBLKSIZ && freebytes < slotneeded) { 746 freebytes += dp->d_reclen; 747 if (dp->d_ino != 0) 748 freebytes -= DIRSIZ(0, dp); 749 if (dp->d_reclen == 0) { 750 brelse(bp); 751 return (-1); 752 } 753 i += dp->d_reclen; 754 dp = (struct direct *)((char *)dp + dp->d_reclen); 755 } 756 if (i > DIRBLKSIZ) { 757 brelse(bp); 758 return (-1); 759 } 760 if (freebytes < slotneeded) 761 panic("ufsdirhash_findfree: free mismatch"); 762 brelse(bp); 763 *slotsize = pos + i - slotstart; 764 return (slotstart); 765} 766 767/* 768 * Return the start of the unused space at the end of a directory, or 769 * -1 if there are no trailing unused blocks. 770 */ 771doff_t 772ufsdirhash_enduseful(struct inode *ip) 773{ 774 775 struct dirhash *dh; 776 int i; 777 778 dh = ip->i_dirhash; 779 DIRHASH_ASSERT_LOCKED(dh); 780 KASSERT(dh != NULL && dh->dh_hash != NULL, 781 ("ufsdirhash_enduseful: Invalid dirhash %p\n", dh)); 782 783 if (dh->dh_blkfree[dh->dh_dirblks - 1] != DIRBLKSIZ / DIRALIGN) 784 return (-1); 785 786 for (i = dh->dh_dirblks - 1; i >= 0; i--) 787 if (dh->dh_blkfree[i] != DIRBLKSIZ / DIRALIGN) 788 break; 789 790 return ((doff_t)(i + 1) * DIRBLKSIZ); 791} 792 793/* 794 * Insert information into the hash about a new directory entry. dirp 795 * points to a struct direct containing the entry, and offset specifies 796 * the offset of this entry. 797 */ 798void 799ufsdirhash_add(struct inode *ip, struct direct *dirp, doff_t offset) 800{ 801 struct dirhash *dh; 802 int slot; 803 804 if ((dh = ufsdirhash_acquire(ip)) == NULL) 805 return; 806 807 KASSERT(offset < dh->dh_dirblks * DIRBLKSIZ, 808 ("ufsdirhash_add: bad offset")); 809 /* 810 * Normal hash usage is < 66%. If the usage gets too high then 811 * remove the hash entirely and let it be rebuilt later. 812 */ 813 if (dh->dh_hused >= (dh->dh_hlen * 3) / 4) { 814 ufsdirhash_free_locked(ip); 815 return; 816 } 817 818 /* Find a free hash slot (empty or deleted), and add the entry. */ 819 slot = ufsdirhash_hash(dh, dirp->d_name, dirp->d_namlen); 820 while (DH_ENTRY(dh, slot) >= 0) 821 slot = WRAPINCR(slot, dh->dh_hlen); 822 if (DH_ENTRY(dh, slot) == DIRHASH_EMPTY) 823 dh->dh_hused++; 824 DH_ENTRY(dh, slot) = offset; 825 826 /* Update last used time. */ 827 dh->dh_lastused = time_second; 828 829 /* Update the per-block summary info. */ 830 ufsdirhash_adjfree(dh, offset, -DIRSIZ(0, dirp)); 831 ufsdirhash_release(dh); 832} 833 834/* 835 * Remove the specified directory entry from the hash. The entry to remove 836 * is defined by the name in `dirp', which must exist at the specified 837 * `offset' within the directory. 838 */ 839void 840ufsdirhash_remove(struct inode *ip, struct direct *dirp, doff_t offset) 841{ 842 struct dirhash *dh; 843 int slot; 844 845 if ((dh = ufsdirhash_acquire(ip)) == NULL) 846 return; 847 848 KASSERT(offset < dh->dh_dirblks * DIRBLKSIZ, 849 ("ufsdirhash_remove: bad offset")); 850 /* Find the entry */ 851 slot = ufsdirhash_findslot(dh, dirp->d_name, dirp->d_namlen, offset); 852 853 /* Remove the hash entry. */ 854 ufsdirhash_delslot(dh, slot); 855 856 /* Update the per-block summary info. */ 857 ufsdirhash_adjfree(dh, offset, DIRSIZ(0, dirp)); 858 ufsdirhash_release(dh); 859} 860 861/* 862 * Change the offset associated with a directory entry in the hash. Used 863 * when compacting directory blocks. 864 */ 865void 866ufsdirhash_move(struct inode *ip, struct direct *dirp, doff_t oldoff, 867 doff_t newoff) 868{ 869 struct dirhash *dh; 870 int slot; 871 872 if ((dh = ufsdirhash_acquire(ip)) == NULL) 873 return; 874 875 KASSERT(oldoff < dh->dh_dirblks * DIRBLKSIZ && 876 newoff < dh->dh_dirblks * DIRBLKSIZ, 877 ("ufsdirhash_move: bad offset")); 878 /* Find the entry, and update the offset. */ 879 slot = ufsdirhash_findslot(dh, dirp->d_name, dirp->d_namlen, oldoff); 880 DH_ENTRY(dh, slot) = newoff; 881 ufsdirhash_release(dh); 882} 883 884/* 885 * Inform dirhash that the directory has grown by one block that 886 * begins at offset (i.e. the new length is offset + DIRBLKSIZ). 887 */ 888void 889ufsdirhash_newblk(struct inode *ip, doff_t offset) 890{ 891 struct dirhash *dh; 892 int block; 893 894 if ((dh = ufsdirhash_acquire(ip)) == NULL) 895 return; 896 897 KASSERT(offset == dh->dh_dirblks * DIRBLKSIZ, 898 ("ufsdirhash_newblk: bad offset")); 899 block = offset / DIRBLKSIZ; 900 if (block >= dh->dh_nblk) { 901 /* Out of space; must rebuild. */ 902 ufsdirhash_free_locked(ip); 903 return; 904 } 905 dh->dh_dirblks = block + 1; 906 907 /* Account for the new free block. */ 908 dh->dh_blkfree[block] = DIRBLKSIZ / DIRALIGN; 909 if (dh->dh_firstfree[DH_NFSTATS] == -1) 910 dh->dh_firstfree[DH_NFSTATS] = block; 911 ufsdirhash_release(dh); 912} 913 914/* 915 * Inform dirhash that the directory is being truncated. 916 */ 917void 918ufsdirhash_dirtrunc(struct inode *ip, doff_t offset) 919{ 920 struct dirhash *dh; 921 int block, i; 922 923 if ((dh = ufsdirhash_acquire(ip)) == NULL) 924 return; 925 926 KASSERT(offset <= dh->dh_dirblks * DIRBLKSIZ, 927 ("ufsdirhash_dirtrunc: bad offset")); 928 block = howmany(offset, DIRBLKSIZ); 929 /* 930 * If the directory shrinks to less than 1/8 of dh_nblk blocks 931 * (about 20% of its original size due to the 50% extra added in 932 * ufsdirhash_build) then free it, and let the caller rebuild 933 * if necessary. 934 */ 935 if (block < dh->dh_nblk / 8 && dh->dh_narrays > 1) { 936 ufsdirhash_free_locked(ip); 937 return; 938 } 939 940 /* 941 * Remove any `first free' information pertaining to the 942 * truncated blocks. All blocks we're removing should be 943 * completely unused. 944 */ 945 if (dh->dh_firstfree[DH_NFSTATS] >= block) 946 dh->dh_firstfree[DH_NFSTATS] = -1; 947 for (i = block; i < dh->dh_dirblks; i++) 948 if (dh->dh_blkfree[i] != DIRBLKSIZ / DIRALIGN) 949 panic("ufsdirhash_dirtrunc: blocks in use"); 950 for (i = 0; i < DH_NFSTATS; i++) 951 if (dh->dh_firstfree[i] >= block) 952 panic("ufsdirhash_dirtrunc: first free corrupt"); 953 dh->dh_dirblks = block; 954 ufsdirhash_release(dh); 955} 956 957/* 958 * Debugging function to check that the dirhash information about 959 * a directory block matches its actual contents. Panics if a mismatch 960 * is detected. 961 * 962 * On entry, `buf' should point to the start of an in-core 963 * DIRBLKSIZ-sized directory block, and `offset' should contain the 964 * offset from the start of the directory of that block. 965 */ 966void 967ufsdirhash_checkblock(struct inode *ip, char *buf, doff_t offset) 968{ 969 struct dirhash *dh; 970 struct direct *dp; 971 int block, ffslot, i, nfree; 972 973 if (!ufs_dirhashcheck) 974 return; 975 if ((dh = ufsdirhash_acquire(ip)) == NULL) 976 return; 977 978 block = offset / DIRBLKSIZ; 979 if ((offset & (DIRBLKSIZ - 1)) != 0 || block >= dh->dh_dirblks) 980 panic("ufsdirhash_checkblock: bad offset"); 981 982 nfree = 0; 983 for (i = 0; i < DIRBLKSIZ; i += dp->d_reclen) { 984 dp = (struct direct *)(buf + i); 985 if (dp->d_reclen == 0 || i + dp->d_reclen > DIRBLKSIZ) 986 panic("ufsdirhash_checkblock: bad dir"); 987 988 if (dp->d_ino == 0) { 989#if 0 990 /* 991 * XXX entries with d_ino == 0 should only occur 992 * at the start of a DIRBLKSIZ block. However the 993 * ufs code is tolerant of such entries at other 994 * offsets, and fsck does not fix them. 995 */ 996 if (i != 0) 997 panic("ufsdirhash_checkblock: bad dir inode"); 998#endif 999 nfree += dp->d_reclen; 1000 continue; 1001 } 1002 1003 /* Check that the entry exists (will panic if it doesn't). */ 1004 ufsdirhash_findslot(dh, dp->d_name, dp->d_namlen, offset + i); 1005 1006 nfree += dp->d_reclen - DIRSIZ(0, dp); 1007 } 1008 if (i != DIRBLKSIZ) 1009 panic("ufsdirhash_checkblock: bad dir end"); 1010 1011 if (dh->dh_blkfree[block] * DIRALIGN != nfree) 1012 panic("ufsdirhash_checkblock: bad free count"); 1013 1014 ffslot = BLKFREE2IDX(nfree / DIRALIGN); 1015 for (i = 0; i <= DH_NFSTATS; i++) 1016 if (dh->dh_firstfree[i] == block && i != ffslot) 1017 panic("ufsdirhash_checkblock: bad first-free"); 1018 if (dh->dh_firstfree[ffslot] == -1) 1019 panic("ufsdirhash_checkblock: missing first-free entry"); 1020 ufsdirhash_release(dh); 1021} 1022 1023/* 1024 * Hash the specified filename into a dirhash slot. 1025 */ 1026static int 1027ufsdirhash_hash(struct dirhash *dh, char *name, int namelen) 1028{ 1029 u_int32_t hash; 1030 1031 /* 1032 * We hash the name and then some other bit of data that is 1033 * invariant over the dirhash's lifetime. Otherwise names 1034 * differing only in the last byte are placed close to one 1035 * another in the table, which is bad for linear probing. 1036 */ 1037 hash = fnv_32_buf(name, namelen, FNV1_32_INIT); 1038 hash = fnv_32_buf(&dh, sizeof(dh), hash); 1039 return (hash % dh->dh_hlen); 1040} 1041 1042/* 1043 * Adjust the number of free bytes in the block containing `offset' 1044 * by the value specified by `diff'. 1045 * 1046 * The caller must ensure we have exclusive access to `dh'; normally 1047 * that means that dh_lock should be held, but this is also called 1048 * from ufsdirhash_build() where exclusive access can be assumed. 1049 */ 1050static void 1051ufsdirhash_adjfree(struct dirhash *dh, doff_t offset, int diff) 1052{ 1053 int block, i, nfidx, ofidx; 1054 1055 /* Update the per-block summary info. */ 1056 block = offset / DIRBLKSIZ; 1057 KASSERT(block < dh->dh_nblk && block < dh->dh_dirblks, 1058 ("dirhash bad offset")); 1059 ofidx = BLKFREE2IDX(dh->dh_blkfree[block]); 1060 dh->dh_blkfree[block] = (int)dh->dh_blkfree[block] + (diff / DIRALIGN); 1061 nfidx = BLKFREE2IDX(dh->dh_blkfree[block]); 1062 1063 /* Update the `first free' list if necessary. */ 1064 if (ofidx != nfidx) { 1065 /* If removing, scan forward for the next block. */ 1066 if (dh->dh_firstfree[ofidx] == block) { 1067 for (i = block + 1; i < dh->dh_dirblks; i++) 1068 if (BLKFREE2IDX(dh->dh_blkfree[i]) == ofidx) 1069 break; 1070 dh->dh_firstfree[ofidx] = (i < dh->dh_dirblks) ? i : -1; 1071 } 1072 1073 /* Make this the new `first free' if necessary */ 1074 if (dh->dh_firstfree[nfidx] > block || 1075 dh->dh_firstfree[nfidx] == -1) 1076 dh->dh_firstfree[nfidx] = block; 1077 } 1078} 1079 1080/* 1081 * Find the specified name which should have the specified offset. 1082 * Returns a slot number, and panics on failure. 1083 * 1084 * `dh' must be locked on entry and remains so on return. 1085 */ 1086static int 1087ufsdirhash_findslot(struct dirhash *dh, char *name, int namelen, doff_t offset) 1088{ 1089 int slot; 1090 1091 DIRHASH_ASSERT_LOCKED(dh); 1092 1093 /* Find the entry. */ 1094 KASSERT(dh->dh_hused < dh->dh_hlen, ("dirhash find full")); 1095 slot = ufsdirhash_hash(dh, name, namelen); 1096 while (DH_ENTRY(dh, slot) != offset && 1097 DH_ENTRY(dh, slot) != DIRHASH_EMPTY) 1098 slot = WRAPINCR(slot, dh->dh_hlen); 1099 if (DH_ENTRY(dh, slot) != offset) 1100 panic("ufsdirhash_findslot: '%.*s' not found", namelen, name); 1101 1102 return (slot); 1103} 1104 1105/* 1106 * Remove the entry corresponding to the specified slot from the hash array. 1107 * 1108 * `dh' must be locked on entry and remains so on return. 1109 */ 1110static void 1111ufsdirhash_delslot(struct dirhash *dh, int slot) 1112{ 1113 int i; 1114 1115 DIRHASH_ASSERT_LOCKED(dh); 1116 1117 /* Mark the entry as deleted. */ 1118 DH_ENTRY(dh, slot) = DIRHASH_DEL; 1119 1120 /* If this is the end of a chain of DIRHASH_DEL slots, remove them. */ 1121 for (i = slot; DH_ENTRY(dh, i) == DIRHASH_DEL; ) 1122 i = WRAPINCR(i, dh->dh_hlen); 1123 if (DH_ENTRY(dh, i) == DIRHASH_EMPTY) { 1124 i = WRAPDECR(i, dh->dh_hlen); 1125 while (DH_ENTRY(dh, i) == DIRHASH_DEL) { 1126 DH_ENTRY(dh, i) = DIRHASH_EMPTY; 1127 dh->dh_hused--; 1128 i = WRAPDECR(i, dh->dh_hlen); 1129 } 1130 KASSERT(dh->dh_hused >= 0, ("ufsdirhash_delslot neg hlen")); 1131 } 1132} 1133 1134/* 1135 * Given a directory entry and its offset, find the offset of the 1136 * previous entry in the same DIRBLKSIZ-sized block. Returns an 1137 * offset, or -1 if there is no previous entry in the block or some 1138 * other problem occurred. 1139 */ 1140static doff_t 1141ufsdirhash_getprev(struct direct *dirp, doff_t offset) 1142{ 1143 struct direct *dp; 1144 char *blkbuf; 1145 doff_t blkoff, prevoff; 1146 int entrypos, i; 1147 1148 blkoff = rounddown2(offset, DIRBLKSIZ); /* offset of start of block */ 1149 entrypos = offset & (DIRBLKSIZ - 1); /* entry relative to block */ 1150 blkbuf = (char *)dirp - entrypos; 1151 prevoff = blkoff; 1152 1153 /* If `offset' is the start of a block, there is no previous entry. */ 1154 if (entrypos == 0) 1155 return (-1); 1156 1157 /* Scan from the start of the block until we get to the entry. */ 1158 for (i = 0; i < entrypos; i += dp->d_reclen) { 1159 dp = (struct direct *)(blkbuf + i); 1160 if (dp->d_reclen == 0 || i + dp->d_reclen > entrypos) 1161 return (-1); /* Corrupted directory. */ 1162 prevoff = blkoff + i; 1163 } 1164 return (prevoff); 1165} 1166 1167/* 1168 * Delete the given dirhash and reclaim its memory. Assumes that 1169 * ufsdirhash_list is locked, and leaves it locked. Also assumes 1170 * that dh is locked. Returns the amount of memory freed. 1171 */ 1172static int 1173ufsdirhash_destroy(struct dirhash *dh) 1174{ 1175 doff_t **hash; 1176 u_int8_t *blkfree; 1177 int i, mem, narrays; 1178 1179 KASSERT(dh->dh_hash != NULL, ("dirhash: NULL hash on list")); 1180 1181 /* Remove it from the list and detach its memory. */ 1182 TAILQ_REMOVE(&ufsdirhash_list, dh, dh_list); 1183 dh->dh_onlist = 0; 1184 hash = dh->dh_hash; 1185 dh->dh_hash = NULL; 1186 blkfree = dh->dh_blkfree; 1187 dh->dh_blkfree = NULL; 1188 narrays = dh->dh_narrays; 1189 mem = dh->dh_memreq; 1190 dh->dh_memreq = 0; 1191 1192 /* Unlock dirhash and free the detached memory. */ 1193 ufsdirhash_release(dh); 1194 for (i = 0; i < narrays; i++) 1195 DIRHASH_BLKFREE(hash[i]); 1196 free(hash, M_DIRHASH); 1197 free(blkfree, M_DIRHASH); 1198 1199 /* Account for the returned memory. */ 1200 ufs_dirhashmem -= mem; 1201 1202 return (mem); 1203} 1204 1205/* 1206 * Try to free up `wanted' bytes by stealing memory from existing 1207 * dirhashes. Returns zero with list locked if successful. 1208 */ 1209static int 1210ufsdirhash_recycle(int wanted) 1211{ 1212 struct dirhash *dh; 1213 1214 DIRHASHLIST_LOCK(); 1215 dh = TAILQ_FIRST(&ufsdirhash_list); 1216 while (wanted + ufs_dirhashmem > ufs_dirhashmaxmem) { 1217 /* Decrement the score; only recycle if it becomes zero. */ 1218 if (dh == NULL || --dh->dh_score > 0) { 1219 DIRHASHLIST_UNLOCK(); 1220 return (-1); 1221 } 1222 /* 1223 * If we can't lock it it's in use and we don't want to 1224 * recycle it anyway. 1225 */ 1226 if (!sx_try_xlock(&dh->dh_lock)) { 1227 dh = TAILQ_NEXT(dh, dh_list); 1228 continue; 1229 } 1230 1231 ufsdirhash_destroy(dh); 1232 1233 /* Repeat if necessary. */ 1234 dh = TAILQ_FIRST(&ufsdirhash_list); 1235 } 1236 /* Success; return with list locked. */ 1237 return (0); 1238} 1239 1240/* 1241 * Callback that frees some dirhashes when the system is low on virtual memory. 1242 */ 1243static void 1244ufsdirhash_lowmem() 1245{ 1246 struct dirhash *dh, *dh_temp; 1247 int memfreed, memwanted; 1248 1249 ufs_dirhashlowmemcount++; 1250 memfreed = 0; 1251 memwanted = ufs_dirhashmem * ufs_dirhashreclaimpercent / 100; 1252 1253 DIRHASHLIST_LOCK(); 1254 1255 /* 1256 * Reclaim up to memwanted from the oldest dirhashes. This will allow 1257 * us to make some progress when the system is running out of memory 1258 * without compromising the dinamicity of maximum age. If the situation 1259 * does not improve lowmem will be eventually retriggered and free some 1260 * other entry in the cache. The entries on the head of the list should 1261 * be the oldest. If during list traversal we can't get a lock on the 1262 * dirhash, it will be skipped. 1263 */ 1264 TAILQ_FOREACH_SAFE(dh, &ufsdirhash_list, dh_list, dh_temp) { 1265 if (sx_try_xlock(&dh->dh_lock)) 1266 memfreed += ufsdirhash_destroy(dh); 1267 if (memfreed >= memwanted) 1268 break; 1269 } 1270 DIRHASHLIST_UNLOCK(); 1271} 1272 1273static int 1274ufsdirhash_set_reclaimpercent(SYSCTL_HANDLER_ARGS) 1275{ 1276 int error, v; 1277 1278 v = ufs_dirhashreclaimpercent; 1279 error = sysctl_handle_int(oidp, &v, v, req); 1280 if (error) 1281 return (error); 1282 if (req->newptr == NULL) 1283 return (error); 1284 if (v == ufs_dirhashreclaimpercent) 1285 return (0); 1286 1287 /* Refuse invalid percentages */ 1288 if (v < 0 || v > 100) 1289 return (EINVAL); 1290 ufs_dirhashreclaimpercent = v; 1291 return (0); 1292} 1293 1294void 1295ufsdirhash_init() 1296{ 1297 ufs_dirhashmaxmem = lmax(roundup(hibufspace / 64, PAGE_SIZE), 1298 2 * 1024 * 1024); 1299 1300 ufsdirhash_zone = uma_zcreate("DIRHASH", DH_NBLKOFF * sizeof(doff_t), 1301 NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, 0); 1302 mtx_init(&ufsdirhash_mtx, "dirhash list", NULL, MTX_DEF); 1303 TAILQ_INIT(&ufsdirhash_list); 1304 1305 /* Register a callback function to handle low memory signals */ 1306 EVENTHANDLER_REGISTER(vm_lowmem, ufsdirhash_lowmem, NULL, 1307 EVENTHANDLER_PRI_FIRST); 1308} 1309 1310void 1311ufsdirhash_uninit() 1312{ 1313 KASSERT(TAILQ_EMPTY(&ufsdirhash_list), ("ufsdirhash_uninit")); 1314 uma_zdestroy(ufsdirhash_zone); 1315 mtx_destroy(&ufsdirhash_mtx); 1316} 1317 1318#endif /* UFS_DIRHASH */ 1319