1 /* 2 * CDDL HEADER START 3 * 4 * The contents of this file are subject to the terms of the 5 * Common Development and Distribution License (the "License"). 6 * You may not use this file except in compliance with the License. 7 * 8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE 9 * or http://www.opensolaris.org/os/licensing. 10 * See the License for the specific language governing permissions 11 * and limitations under the License. 12 * 13 * When distributing Covered Code, include this CDDL HEADER in each 14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE. 15 * If applicable, add the following below this CDDL HEADER, with the 16 * fields enclosed by brackets "[]" replaced with your own identifying 17 * information: Portions Copyright [yyyy] [name of copyright owner] 18 * 19 * CDDL HEADER END 20 */ 21 /* 22 * Copyright 2009 Sun Microsystems, Inc. All rights reserved. 23 * Use is subject to license terms. 24 */ 25 26 /* 27 * This file contains the top half of the zfs directory structure 28 * implementation. The bottom half is in zap_leaf.c. 29 * 30 * The zdir is an extendable hash data structure. There is a table of 31 * pointers to buckets (zap_t->zd_data->zd_leafs). The buckets are 32 * each a constant size and hold a variable number of directory entries. 33 * The buckets (aka "leaf nodes") are implemented in zap_leaf.c. 34 * 35 * The pointer table holds a power of 2 number of pointers. 36 * (1<<zap_t->zd_data->zd_phys->zd_prefix_len). The bucket pointed to 37 * by the pointer at index i in the table holds entries whose hash value 38 * has a zd_prefix_len - bit prefix 39 */ 40 41 #include <sys/spa.h> 42 #include <sys/dmu.h> 43 #include <sys/zfs_context.h> 44 #include <sys/zfs_znode.h> 45 #include <sys/fs/zfs.h> 46 #include <sys/zap.h> 47 #include <sys/refcount.h> 48 #include <sys/zap_impl.h> 49 #include <sys/zap_leaf.h> 50 51 int fzap_default_block_shift = 14; /* 16k blocksize */ 52 53 static void zap_leaf_pageout(dmu_buf_t *db, void *vl); 54 static uint64_t zap_allocate_blocks(zap_t *zap, int nblocks); 55 56 57 void 58 fzap_byteswap(void *vbuf, size_t size) 59 { 60 uint64_t block_type; 61 62 block_type = *(uint64_t *)vbuf; 63 64 if (block_type == ZBT_LEAF || block_type == BSWAP_64(ZBT_LEAF)) 65 zap_leaf_byteswap(vbuf, size); 66 else { 67 /* it's a ptrtbl block */ 68 byteswap_uint64_array(vbuf, size); 69 } 70 } 71 72 void 73 fzap_upgrade(zap_t *zap, dmu_tx_t *tx) 74 { 75 dmu_buf_t *db; 76 zap_leaf_t *l; 77 int i; 78 zap_phys_t *zp; 79 80 ASSERT(RW_WRITE_HELD(&zap->zap_rwlock)); 81 zap->zap_ismicro = FALSE; 82 83 (void) dmu_buf_update_user(zap->zap_dbuf, zap, zap, 84 &zap->zap_f.zap_phys, zap_evict); 85 86 mutex_init(&zap->zap_f.zap_num_entries_mtx, 0, 0, 0); 87 zap->zap_f.zap_block_shift = highbit(zap->zap_dbuf->db_size) - 1; 88 89 zp = zap->zap_f.zap_phys; 90 /* 91 * explicitly zero it since it might be coming from an 92 * initialized microzap 93 */ 94 bzero(zap->zap_dbuf->db_data, zap->zap_dbuf->db_size); 95 zp->zap_block_type = ZBT_HEADER; 96 zp->zap_magic = ZAP_MAGIC; 97 98 zp->zap_ptrtbl.zt_shift = ZAP_EMBEDDED_PTRTBL_SHIFT(zap); 99 100 zp->zap_freeblk = 2; /* block 1 will be the first leaf */ 101 zp->zap_num_leafs = 1; 102 zp->zap_num_entries = 0; 103 zp->zap_salt = zap->zap_salt; 104 zp->zap_normflags = zap->zap_normflags; 105 106 /* block 1 will be the first leaf */ 107 for (i = 0; i < (1<<zp->zap_ptrtbl.zt_shift); i++) 108 ZAP_EMBEDDED_PTRTBL_ENT(zap, i) = 1; 109 110 /* 111 * set up block 1 - the first leaf 112 */ 113 VERIFY(0 == dmu_buf_hold(zap->zap_objset, zap->zap_object, 114 1<<FZAP_BLOCK_SHIFT(zap), FTAG, &db)); 115 dmu_buf_will_dirty(db, tx); 116 117 l = kmem_zalloc(sizeof (zap_leaf_t), KM_SLEEP); 118 l->l_dbuf = db; 119 l->l_phys = db->db_data; 120 121 zap_leaf_init(l, zp->zap_normflags != 0); 122 123 kmem_free(l, sizeof (zap_leaf_t)); 124 dmu_buf_rele(db, FTAG); 125 } 126 127 static int 128 zap_tryupgradedir(zap_t *zap, dmu_tx_t *tx) 129 { 130 if (RW_WRITE_HELD(&zap->zap_rwlock)) 131 return (1); 132 if (rw_tryupgrade(&zap->zap_rwlock)) { 133 dmu_buf_will_dirty(zap->zap_dbuf, tx); 134 return (1); 135 } 136 return (0); 137 } 138 139 /* 140 * Generic routines for dealing with the pointer & cookie tables. 141 */ 142 143 static int 144 zap_table_grow(zap_t *zap, zap_table_phys_t *tbl, 145 void (*transfer_func)(const uint64_t *src, uint64_t *dst, int n), 146 dmu_tx_t *tx) 147 { 148 uint64_t b, newblk; 149 dmu_buf_t *db_old, *db_new; 150 int err; 151 int bs = FZAP_BLOCK_SHIFT(zap); 152 int hepb = 1<<(bs-4); 153 /* hepb = half the number of entries in a block */ 154 155 ASSERT(RW_WRITE_HELD(&zap->zap_rwlock)); 156 ASSERT(tbl->zt_blk != 0); 157 ASSERT(tbl->zt_numblks > 0); 158 159 if (tbl->zt_nextblk != 0) { 160 newblk = tbl->zt_nextblk; 161 } else { 162 newblk = zap_allocate_blocks(zap, tbl->zt_numblks * 2); 163 tbl->zt_nextblk = newblk; 164 ASSERT3U(tbl->zt_blks_copied, ==, 0); 165 dmu_prefetch(zap->zap_objset, zap->zap_object, 166 tbl->zt_blk << bs, tbl->zt_numblks << bs); 167 } 168 169 /* 170 * Copy the ptrtbl from the old to new location. 171 */ 172 173 b = tbl->zt_blks_copied; 174 err = dmu_buf_hold(zap->zap_objset, zap->zap_object, 175 (tbl->zt_blk + b) << bs, FTAG, &db_old); 176 if (err) 177 return (err); 178 179 /* first half of entries in old[b] go to new[2*b+0] */ 180 VERIFY(0 == dmu_buf_hold(zap->zap_objset, zap->zap_object, 181 (newblk + 2*b+0) << bs, FTAG, &db_new)); 182 dmu_buf_will_dirty(db_new, tx); 183 transfer_func(db_old->db_data, db_new->db_data, hepb); 184 dmu_buf_rele(db_new, FTAG); 185 186 /* second half of entries in old[b] go to new[2*b+1] */ 187 VERIFY(0 == dmu_buf_hold(zap->zap_objset, zap->zap_object, 188 (newblk + 2*b+1) << bs, FTAG, &db_new)); 189 dmu_buf_will_dirty(db_new, tx); 190 transfer_func((uint64_t *)db_old->db_data + hepb, 191 db_new->db_data, hepb); 192 dmu_buf_rele(db_new, FTAG); 193 194 dmu_buf_rele(db_old, FTAG); 195 196 tbl->zt_blks_copied++; 197 198 dprintf("copied block %llu of %llu\n", 199 tbl->zt_blks_copied, tbl->zt_numblks); 200 201 if (tbl->zt_blks_copied == tbl->zt_numblks) { 202 (void) dmu_free_range(zap->zap_objset, zap->zap_object, 203 tbl->zt_blk << bs, tbl->zt_numblks << bs, tx); 204 205 tbl->zt_blk = newblk; 206 tbl->zt_numblks *= 2; 207 tbl->zt_shift++; 208 tbl->zt_nextblk = 0; 209 tbl->zt_blks_copied = 0; 210 211 dprintf("finished; numblocks now %llu (%lluk entries)\n", 212 tbl->zt_numblks, 1<<(tbl->zt_shift-10)); 213 } 214 215 return (0); 216 } 217 218 static int 219 zap_table_store(zap_t *zap, zap_table_phys_t *tbl, uint64_t idx, uint64_t val, 220 dmu_tx_t *tx) 221 { 222 int err; 223 uint64_t blk, off; 224 int bs = FZAP_BLOCK_SHIFT(zap); 225 dmu_buf_t *db; 226 227 ASSERT(RW_LOCK_HELD(&zap->zap_rwlock)); 228 ASSERT(tbl->zt_blk != 0); 229 230 dprintf("storing %llx at index %llx\n", val, idx); 231 232 blk = idx >> (bs-3); 233 off = idx & ((1<<(bs-3))-1); 234 235 err = dmu_buf_hold(zap->zap_objset, zap->zap_object, 236 (tbl->zt_blk + blk) << bs, FTAG, &db); 237 if (err) 238 return (err); 239 dmu_buf_will_dirty(db, tx); 240 241 if (tbl->zt_nextblk != 0) { 242 uint64_t idx2 = idx * 2; 243 uint64_t blk2 = idx2 >> (bs-3); 244 uint64_t off2 = idx2 & ((1<<(bs-3))-1); 245 dmu_buf_t *db2; 246 247 err = dmu_buf_hold(zap->zap_objset, zap->zap_object, 248 (tbl->zt_nextblk + blk2) << bs, FTAG, &db2); 249 if (err) { 250 dmu_buf_rele(db, FTAG); 251 return (err); 252 } 253 dmu_buf_will_dirty(db2, tx); 254 ((uint64_t *)db2->db_data)[off2] = val; 255 ((uint64_t *)db2->db_data)[off2+1] = val; 256 dmu_buf_rele(db2, FTAG); 257 } 258 259 ((uint64_t *)db->db_data)[off] = val; 260 dmu_buf_rele(db, FTAG); 261 262 return (0); 263 } 264 265 static int 266 zap_table_load(zap_t *zap, zap_table_phys_t *tbl, uint64_t idx, uint64_t *valp) 267 { 268 uint64_t blk, off; 269 int err; 270 dmu_buf_t *db; 271 int bs = FZAP_BLOCK_SHIFT(zap); 272 273 ASSERT(RW_LOCK_HELD(&zap->zap_rwlock)); 274 275 blk = idx >> (bs-3); 276 off = idx & ((1<<(bs-3))-1); 277 278 err = dmu_buf_hold(zap->zap_objset, zap->zap_object, 279 (tbl->zt_blk + blk) << bs, FTAG, &db); 280 if (err) 281 return (err); 282 *valp = ((uint64_t *)db->db_data)[off]; 283 dmu_buf_rele(db, FTAG); 284 285 if (tbl->zt_nextblk != 0) { 286 /* 287 * read the nextblk for the sake of i/o error checking, 288 * so that zap_table_load() will catch errors for 289 * zap_table_store. 290 */ 291 blk = (idx*2) >> (bs-3); 292 293 err = dmu_buf_hold(zap->zap_objset, zap->zap_object, 294 (tbl->zt_nextblk + blk) << bs, FTAG, &db); 295 dmu_buf_rele(db, FTAG); 296 } 297 return (err); 298 } 299 300 /* 301 * Routines for growing the ptrtbl. 302 */ 303 304 static void 305 zap_ptrtbl_transfer(const uint64_t *src, uint64_t *dst, int n) 306 { 307 int i; 308 for (i = 0; i < n; i++) { 309 uint64_t lb = src[i]; 310 dst[2*i+0] = lb; 311 dst[2*i+1] = lb; 312 } 313 } 314 315 static int 316 zap_grow_ptrtbl(zap_t *zap, dmu_tx_t *tx) 317 { 318 /* In case things go horribly wrong. */ 319 if (zap->zap_f.zap_phys->zap_ptrtbl.zt_shift >= ZAP_HASHBITS-2) 320 return (ENOSPC); 321 322 if (zap->zap_f.zap_phys->zap_ptrtbl.zt_numblks == 0) { 323 /* 324 * We are outgrowing the "embedded" ptrtbl (the one 325 * stored in the header block). Give it its own entire 326 * block, which will double the size of the ptrtbl. 327 */ 328 uint64_t newblk; 329 dmu_buf_t *db_new; 330 int err; 331 332 ASSERT3U(zap->zap_f.zap_phys->zap_ptrtbl.zt_shift, ==, 333 ZAP_EMBEDDED_PTRTBL_SHIFT(zap)); 334 ASSERT3U(zap->zap_f.zap_phys->zap_ptrtbl.zt_blk, ==, 0); 335 336 newblk = zap_allocate_blocks(zap, 1); 337 err = dmu_buf_hold(zap->zap_objset, zap->zap_object, 338 newblk << FZAP_BLOCK_SHIFT(zap), FTAG, &db_new); 339 if (err) 340 return (err); 341 dmu_buf_will_dirty(db_new, tx); 342 zap_ptrtbl_transfer(&ZAP_EMBEDDED_PTRTBL_ENT(zap, 0), 343 db_new->db_data, 1 << ZAP_EMBEDDED_PTRTBL_SHIFT(zap)); 344 dmu_buf_rele(db_new, FTAG); 345 346 zap->zap_f.zap_phys->zap_ptrtbl.zt_blk = newblk; 347 zap->zap_f.zap_phys->zap_ptrtbl.zt_numblks = 1; 348 zap->zap_f.zap_phys->zap_ptrtbl.zt_shift++; 349 350 ASSERT3U(1ULL << zap->zap_f.zap_phys->zap_ptrtbl.zt_shift, ==, 351 zap->zap_f.zap_phys->zap_ptrtbl.zt_numblks << 352 (FZAP_BLOCK_SHIFT(zap)-3)); 353 354 return (0); 355 } else { 356 return (zap_table_grow(zap, &zap->zap_f.zap_phys->zap_ptrtbl, 357 zap_ptrtbl_transfer, tx)); 358 } 359 } 360 361 static void 362 zap_increment_num_entries(zap_t *zap, int delta, dmu_tx_t *tx) 363 { 364 dmu_buf_will_dirty(zap->zap_dbuf, tx); 365 mutex_enter(&zap->zap_f.zap_num_entries_mtx); 366 ASSERT(delta > 0 || zap->zap_f.zap_phys->zap_num_entries >= -delta); 367 zap->zap_f.zap_phys->zap_num_entries += delta; 368 mutex_exit(&zap->zap_f.zap_num_entries_mtx); 369 } 370 371 static uint64_t 372 zap_allocate_blocks(zap_t *zap, int nblocks) 373 { 374 uint64_t newblk; 375 ASSERT(RW_WRITE_HELD(&zap->zap_rwlock)); 376 newblk = zap->zap_f.zap_phys->zap_freeblk; 377 zap->zap_f.zap_phys->zap_freeblk += nblocks; 378 return (newblk); 379 } 380 381 static zap_leaf_t * 382 zap_create_leaf(zap_t *zap, dmu_tx_t *tx) 383 { 384 void *winner; 385 zap_leaf_t *l = kmem_alloc(sizeof (zap_leaf_t), KM_SLEEP); 386 387 ASSERT(RW_WRITE_HELD(&zap->zap_rwlock)); 388 389 rw_init(&l->l_rwlock, 0, 0, 0); 390 rw_enter(&l->l_rwlock, RW_WRITER); 391 l->l_blkid = zap_allocate_blocks(zap, 1); 392 l->l_dbuf = NULL; 393 l->l_phys = NULL; 394 395 VERIFY(0 == dmu_buf_hold(zap->zap_objset, zap->zap_object, 396 l->l_blkid << FZAP_BLOCK_SHIFT(zap), NULL, &l->l_dbuf)); 397 winner = dmu_buf_set_user(l->l_dbuf, l, &l->l_phys, zap_leaf_pageout); 398 ASSERT(winner == NULL); 399 dmu_buf_will_dirty(l->l_dbuf, tx); 400 401 zap_leaf_init(l, zap->zap_normflags != 0); 402 403 zap->zap_f.zap_phys->zap_num_leafs++; 404 405 return (l); 406 } 407 408 int 409 fzap_count(zap_t *zap, uint64_t *count) 410 { 411 ASSERT(!zap->zap_ismicro); 412 mutex_enter(&zap->zap_f.zap_num_entries_mtx); /* unnecessary */ 413 *count = zap->zap_f.zap_phys->zap_num_entries; 414 mutex_exit(&zap->zap_f.zap_num_entries_mtx); 415 return (0); 416 } 417 418 /* 419 * Routines for obtaining zap_leaf_t's 420 */ 421 422 void 423 zap_put_leaf(zap_leaf_t *l) 424 { 425 rw_exit(&l->l_rwlock); 426 dmu_buf_rele(l->l_dbuf, NULL); 427 } 428 429 _NOTE(ARGSUSED(0)) 430 static void 431 zap_leaf_pageout(dmu_buf_t *db, void *vl) 432 { 433 zap_leaf_t *l = vl; 434 435 rw_destroy(&l->l_rwlock); 436 kmem_free(l, sizeof (zap_leaf_t)); 437 } 438 439 static zap_leaf_t * 440 zap_open_leaf(uint64_t blkid, dmu_buf_t *db) 441 { 442 zap_leaf_t *l, *winner; 443 444 ASSERT(blkid != 0); 445 446 l = kmem_alloc(sizeof (zap_leaf_t), KM_SLEEP); 447 rw_init(&l->l_rwlock, 0, 0, 0); 448 rw_enter(&l->l_rwlock, RW_WRITER); 449 l->l_blkid = blkid; 450 l->l_bs = highbit(db->db_size)-1; 451 l->l_dbuf = db; 452 l->l_phys = NULL; 453 454 winner = dmu_buf_set_user(db, l, &l->l_phys, zap_leaf_pageout); 455 456 rw_exit(&l->l_rwlock); 457 if (winner != NULL) { 458 /* someone else set it first */ 459 zap_leaf_pageout(NULL, l); 460 l = winner; 461 } 462 463 /* 464 * lhr_pad was previously used for the next leaf in the leaf 465 * chain. There should be no chained leafs (as we have removed 466 * support for them). 467 */ 468 ASSERT3U(l->l_phys->l_hdr.lh_pad1, ==, 0); 469 470 /* 471 * There should be more hash entries than there can be 472 * chunks to put in the hash table 473 */ 474 ASSERT3U(ZAP_LEAF_HASH_NUMENTRIES(l), >, ZAP_LEAF_NUMCHUNKS(l) / 3); 475 476 /* The chunks should begin at the end of the hash table */ 477 ASSERT3P(&ZAP_LEAF_CHUNK(l, 0), ==, 478 &l->l_phys->l_hash[ZAP_LEAF_HASH_NUMENTRIES(l)]); 479 480 /* The chunks should end at the end of the block */ 481 ASSERT3U((uintptr_t)&ZAP_LEAF_CHUNK(l, ZAP_LEAF_NUMCHUNKS(l)) - 482 (uintptr_t)l->l_phys, ==, l->l_dbuf->db_size); 483 484 return (l); 485 } 486 487 static int 488 zap_get_leaf_byblk(zap_t *zap, uint64_t blkid, dmu_tx_t *tx, krw_t lt, 489 zap_leaf_t **lp) 490 { 491 dmu_buf_t *db; 492 zap_leaf_t *l; 493 int bs = FZAP_BLOCK_SHIFT(zap); 494 int err; 495 496 ASSERT(RW_LOCK_HELD(&zap->zap_rwlock)); 497 498 err = dmu_buf_hold(zap->zap_objset, zap->zap_object, 499 blkid << bs, NULL, &db); 500 if (err) 501 return (err); 502 503 ASSERT3U(db->db_object, ==, zap->zap_object); 504 ASSERT3U(db->db_offset, ==, blkid << bs); 505 ASSERT3U(db->db_size, ==, 1 << bs); 506 ASSERT(blkid != 0); 507 508 l = dmu_buf_get_user(db); 509 510 if (l == NULL) 511 l = zap_open_leaf(blkid, db); 512 513 rw_enter(&l->l_rwlock, lt); 514 /* 515 * Must lock before dirtying, otherwise l->l_phys could change, 516 * causing ASSERT below to fail. 517 */ 518 if (lt == RW_WRITER) 519 dmu_buf_will_dirty(db, tx); 520 ASSERT3U(l->l_blkid, ==, blkid); 521 ASSERT3P(l->l_dbuf, ==, db); 522 ASSERT3P(l->l_phys, ==, l->l_dbuf->db_data); 523 ASSERT3U(l->l_phys->l_hdr.lh_block_type, ==, ZBT_LEAF); 524 ASSERT3U(l->l_phys->l_hdr.lh_magic, ==, ZAP_LEAF_MAGIC); 525 526 *lp = l; 527 return (0); 528 } 529 530 static int 531 zap_idx_to_blk(zap_t *zap, uint64_t idx, uint64_t *valp) 532 { 533 ASSERT(RW_LOCK_HELD(&zap->zap_rwlock)); 534 535 if (zap->zap_f.zap_phys->zap_ptrtbl.zt_numblks == 0) { 536 ASSERT3U(idx, <, 537 (1ULL << zap->zap_f.zap_phys->zap_ptrtbl.zt_shift)); 538 *valp = ZAP_EMBEDDED_PTRTBL_ENT(zap, idx); 539 return (0); 540 } else { 541 return (zap_table_load(zap, &zap->zap_f.zap_phys->zap_ptrtbl, 542 idx, valp)); 543 } 544 } 545 546 static int 547 zap_set_idx_to_blk(zap_t *zap, uint64_t idx, uint64_t blk, dmu_tx_t *tx) 548 { 549 ASSERT(tx != NULL); 550 ASSERT(RW_WRITE_HELD(&zap->zap_rwlock)); 551 552 if (zap->zap_f.zap_phys->zap_ptrtbl.zt_blk == 0) { 553 ZAP_EMBEDDED_PTRTBL_ENT(zap, idx) = blk; 554 return (0); 555 } else { 556 return (zap_table_store(zap, &zap->zap_f.zap_phys->zap_ptrtbl, 557 idx, blk, tx)); 558 } 559 } 560 561 static int 562 zap_deref_leaf(zap_t *zap, uint64_t h, dmu_tx_t *tx, krw_t lt, zap_leaf_t **lp) 563 { 564 uint64_t idx, blk; 565 int err; 566 567 ASSERT(zap->zap_dbuf == NULL || 568 zap->zap_f.zap_phys == zap->zap_dbuf->db_data); 569 ASSERT3U(zap->zap_f.zap_phys->zap_magic, ==, ZAP_MAGIC); 570 idx = ZAP_HASH_IDX(h, zap->zap_f.zap_phys->zap_ptrtbl.zt_shift); 571 err = zap_idx_to_blk(zap, idx, &blk); 572 if (err != 0) 573 return (err); 574 err = zap_get_leaf_byblk(zap, blk, tx, lt, lp); 575 576 ASSERT(err || ZAP_HASH_IDX(h, (*lp)->l_phys->l_hdr.lh_prefix_len) == 577 (*lp)->l_phys->l_hdr.lh_prefix); 578 return (err); 579 } 580 581 static int 582 zap_expand_leaf(zap_name_t *zn, zap_leaf_t *l, dmu_tx_t *tx, zap_leaf_t **lp) 583 { 584 zap_t *zap = zn->zn_zap; 585 uint64_t hash = zn->zn_hash; 586 zap_leaf_t *nl; 587 int prefix_diff, i, err; 588 uint64_t sibling; 589 int old_prefix_len = l->l_phys->l_hdr.lh_prefix_len; 590 591 ASSERT3U(old_prefix_len, <=, zap->zap_f.zap_phys->zap_ptrtbl.zt_shift); 592 ASSERT(RW_LOCK_HELD(&zap->zap_rwlock)); 593 594 ASSERT3U(ZAP_HASH_IDX(hash, old_prefix_len), ==, 595 l->l_phys->l_hdr.lh_prefix); 596 597 if (zap_tryupgradedir(zap, tx) == 0 || 598 old_prefix_len == zap->zap_f.zap_phys->zap_ptrtbl.zt_shift) { 599 /* We failed to upgrade, or need to grow the pointer table */ 600 objset_t *os = zap->zap_objset; 601 uint64_t object = zap->zap_object; 602 603 zap_put_leaf(l); 604 zap_unlockdir(zap); 605 err = zap_lockdir(os, object, tx, RW_WRITER, 606 FALSE, FALSE, &zn->zn_zap); 607 zap = zn->zn_zap; 608 if (err) 609 return (err); 610 ASSERT(!zap->zap_ismicro); 611 612 while (old_prefix_len == 613 zap->zap_f.zap_phys->zap_ptrtbl.zt_shift) { 614 err = zap_grow_ptrtbl(zap, tx); 615 if (err) 616 return (err); 617 } 618 619 err = zap_deref_leaf(zap, hash, tx, RW_WRITER, &l); 620 if (err) 621 return (err); 622 623 if (l->l_phys->l_hdr.lh_prefix_len != old_prefix_len) { 624 /* it split while our locks were down */ 625 *lp = l; 626 return (0); 627 } 628 } 629 ASSERT(RW_WRITE_HELD(&zap->zap_rwlock)); 630 ASSERT3U(old_prefix_len, <, zap->zap_f.zap_phys->zap_ptrtbl.zt_shift); 631 ASSERT3U(ZAP_HASH_IDX(hash, old_prefix_len), ==, 632 l->l_phys->l_hdr.lh_prefix); 633 634 prefix_diff = zap->zap_f.zap_phys->zap_ptrtbl.zt_shift - 635 (old_prefix_len + 1); 636 sibling = (ZAP_HASH_IDX(hash, old_prefix_len + 1) | 1) << prefix_diff; 637 638 /* check for i/o errors before doing zap_leaf_split */ 639 for (i = 0; i < (1ULL<<prefix_diff); i++) { 640 uint64_t blk; 641 err = zap_idx_to_blk(zap, sibling+i, &blk); 642 if (err) 643 return (err); 644 ASSERT3U(blk, ==, l->l_blkid); 645 } 646 647 nl = zap_create_leaf(zap, tx); 648 zap_leaf_split(l, nl, zap->zap_normflags != 0); 649 650 /* set sibling pointers */ 651 for (i = 0; i < (1ULL<<prefix_diff); i++) { 652 err = zap_set_idx_to_blk(zap, sibling+i, nl->l_blkid, tx); 653 ASSERT3U(err, ==, 0); /* we checked for i/o errors above */ 654 } 655 656 if (hash & (1ULL << (64 - l->l_phys->l_hdr.lh_prefix_len))) { 657 /* we want the sibling */ 658 zap_put_leaf(l); 659 *lp = nl; 660 } else { 661 zap_put_leaf(nl); 662 *lp = l; 663 } 664 665 return (0); 666 } 667 668 static void 669 zap_put_leaf_maybe_grow_ptrtbl(zap_name_t *zn, zap_leaf_t *l, dmu_tx_t *tx) 670 { 671 zap_t *zap = zn->zn_zap; 672 int shift = zap->zap_f.zap_phys->zap_ptrtbl.zt_shift; 673 int leaffull = (l->l_phys->l_hdr.lh_prefix_len == shift && 674 l->l_phys->l_hdr.lh_nfree < ZAP_LEAF_LOW_WATER); 675 676 zap_put_leaf(l); 677 678 if (leaffull || zap->zap_f.zap_phys->zap_ptrtbl.zt_nextblk) { 679 int err; 680 681 /* 682 * We are in the middle of growing the pointer table, or 683 * this leaf will soon make us grow it. 684 */ 685 if (zap_tryupgradedir(zap, tx) == 0) { 686 objset_t *os = zap->zap_objset; 687 uint64_t zapobj = zap->zap_object; 688 689 zap_unlockdir(zap); 690 err = zap_lockdir(os, zapobj, tx, 691 RW_WRITER, FALSE, FALSE, &zn->zn_zap); 692 zap = zn->zn_zap; 693 if (err) 694 return; 695 } 696 697 /* could have finished growing while our locks were down */ 698 if (zap->zap_f.zap_phys->zap_ptrtbl.zt_shift == shift) 699 (void) zap_grow_ptrtbl(zap, tx); 700 } 701 } 702 703 704 static int 705 fzap_checksize(const char *name, uint64_t integer_size, uint64_t num_integers) 706 { 707 if (name && strlen(name) > ZAP_MAXNAMELEN) 708 return (E2BIG); 709 710 /* Only integer sizes supported by C */ 711 switch (integer_size) { 712 case 1: 713 case 2: 714 case 4: 715 case 8: 716 break; 717 default: 718 return (EINVAL); 719 } 720 721 if (integer_size * num_integers > ZAP_MAXVALUELEN) 722 return (E2BIG); 723 724 return (0); 725 } 726 727 /* 728 * Routines for manipulating attributes. 729 */ 730 int 731 fzap_lookup(zap_name_t *zn, 732 uint64_t integer_size, uint64_t num_integers, void *buf, 733 char *realname, int rn_len, boolean_t *ncp) 734 { 735 zap_leaf_t *l; 736 int err; 737 zap_entry_handle_t zeh; 738 739 err = fzap_checksize(zn->zn_name_orij, integer_size, num_integers); 740 if (err != 0) 741 return (err); 742 743 err = zap_deref_leaf(zn->zn_zap, zn->zn_hash, NULL, RW_READER, &l); 744 if (err != 0) 745 return (err); 746 err = zap_leaf_lookup(l, zn, &zeh); 747 if (err == 0) { 748 err = zap_entry_read(&zeh, integer_size, num_integers, buf); 749 (void) zap_entry_read_name(&zeh, rn_len, realname); 750 if (ncp) { 751 *ncp = zap_entry_normalization_conflict(&zeh, 752 zn, NULL, zn->zn_zap); 753 } 754 } 755 756 zap_put_leaf(l); 757 return (err); 758 } 759 760 int 761 fzap_add_cd(zap_name_t *zn, 762 uint64_t integer_size, uint64_t num_integers, 763 const void *val, uint32_t cd, dmu_tx_t *tx) 764 { 765 zap_leaf_t *l; 766 int err; 767 zap_entry_handle_t zeh; 768 zap_t *zap = zn->zn_zap; 769 770 ASSERT(RW_LOCK_HELD(&zap->zap_rwlock)); 771 ASSERT(!zap->zap_ismicro); 772 ASSERT(fzap_checksize(zn->zn_name_orij, 773 integer_size, num_integers) == 0); 774 775 err = zap_deref_leaf(zap, zn->zn_hash, tx, RW_WRITER, &l); 776 if (err != 0) 777 return (err); 778 retry: 779 err = zap_leaf_lookup(l, zn, &zeh); 780 if (err == 0) { 781 err = EEXIST; 782 goto out; 783 } 784 if (err != ENOENT) 785 goto out; 786 787 err = zap_entry_create(l, zn->zn_name_orij, zn->zn_hash, cd, 788 integer_size, num_integers, val, &zeh); 789 790 if (err == 0) { 791 zap_increment_num_entries(zap, 1, tx); 792 } else if (err == EAGAIN) { 793 err = zap_expand_leaf(zn, l, tx, &l); 794 zap = zn->zn_zap; /* zap_expand_leaf() may change zap */ 795 if (err == 0) 796 goto retry; 797 } 798 799 out: 800 if (zap != NULL) 801 zap_put_leaf_maybe_grow_ptrtbl(zn, l, tx); 802 return (err); 803 } 804 805 int 806 fzap_add(zap_name_t *zn, 807 uint64_t integer_size, uint64_t num_integers, 808 const void *val, dmu_tx_t *tx) 809 { 810 int err = fzap_checksize(zn->zn_name_orij, integer_size, num_integers); 811 if (err != 0) 812 return (err); 813 814 return (fzap_add_cd(zn, integer_size, num_integers, 815 val, ZAP_MAXCD, tx)); 816 } 817 818 int 819 fzap_update(zap_name_t *zn, 820 int integer_size, uint64_t num_integers, const void *val, dmu_tx_t *tx) 821 { 822 zap_leaf_t *l; 823 int err, create; 824 zap_entry_handle_t zeh; 825 zap_t *zap = zn->zn_zap; 826 827 ASSERT(RW_LOCK_HELD(&zap->zap_rwlock)); 828 err = fzap_checksize(zn->zn_name_orij, integer_size, num_integers); 829 if (err != 0) 830 return (err); 831 832 err = zap_deref_leaf(zap, zn->zn_hash, tx, RW_WRITER, &l); 833 if (err != 0) 834 return (err); 835 retry: 836 err = zap_leaf_lookup(l, zn, &zeh); 837 create = (err == ENOENT); 838 ASSERT(err == 0 || err == ENOENT); 839 840 if (create) { 841 err = zap_entry_create(l, zn->zn_name_orij, zn->zn_hash, 842 ZAP_MAXCD, integer_size, num_integers, val, &zeh); 843 if (err == 0) 844 zap_increment_num_entries(zap, 1, tx); 845 } else { 846 err = zap_entry_update(&zeh, integer_size, num_integers, val); 847 } 848 849 if (err == EAGAIN) { 850 err = zap_expand_leaf(zn, l, tx, &l); 851 zap = zn->zn_zap; /* zap_expand_leaf() may change zap */ 852 if (err == 0) 853 goto retry; 854 } 855 856 if (zap != NULL) 857 zap_put_leaf_maybe_grow_ptrtbl(zn, l, tx); 858 return (err); 859 } 860 861 int 862 fzap_length(zap_name_t *zn, 863 uint64_t *integer_size, uint64_t *num_integers) 864 { 865 zap_leaf_t *l; 866 int err; 867 zap_entry_handle_t zeh; 868 869 err = zap_deref_leaf(zn->zn_zap, zn->zn_hash, NULL, RW_READER, &l); 870 if (err != 0) 871 return (err); 872 err = zap_leaf_lookup(l, zn, &zeh); 873 if (err != 0) 874 goto out; 875 876 if (integer_size) 877 *integer_size = zeh.zeh_integer_size; 878 if (num_integers) 879 *num_integers = zeh.zeh_num_integers; 880 out: 881 zap_put_leaf(l); 882 return (err); 883 } 884 885 int 886 fzap_remove(zap_name_t *zn, dmu_tx_t *tx) 887 { 888 zap_leaf_t *l; 889 int err; 890 zap_entry_handle_t zeh; 891 892 err = zap_deref_leaf(zn->zn_zap, zn->zn_hash, tx, RW_WRITER, &l); 893 if (err != 0) 894 return (err); 895 err = zap_leaf_lookup(l, zn, &zeh); 896 if (err == 0) { 897 zap_entry_remove(&zeh); 898 zap_increment_num_entries(zn->zn_zap, -1, tx); 899 } 900 zap_put_leaf(l); 901 return (err); 902 } 903 904 /* 905 * Helper functions for consumers. 906 */ 907 908 int 909 zap_value_search(objset_t *os, uint64_t zapobj, uint64_t value, uint64_t mask, 910 char *name) 911 { 912 zap_cursor_t zc; 913 zap_attribute_t *za; 914 int err; 915 916 if (mask == 0) 917 mask = -1ULL; 918 919 za = kmem_alloc(sizeof (zap_attribute_t), KM_SLEEP); 920 for (zap_cursor_init(&zc, os, zapobj); 921 (err = zap_cursor_retrieve(&zc, za)) == 0; 922 zap_cursor_advance(&zc)) { 923 if ((za->za_first_integer & mask) == (value & mask)) { 924 (void) strcpy(name, za->za_name); 925 break; 926 } 927 } 928 zap_cursor_fini(&zc); 929 kmem_free(za, sizeof (zap_attribute_t)); 930 return (err); 931 } 932 933 int 934 zap_join(objset_t *os, uint64_t fromobj, uint64_t intoobj, dmu_tx_t *tx) 935 { 936 zap_cursor_t zc; 937 zap_attribute_t za; 938 int err; 939 940 for (zap_cursor_init(&zc, os, fromobj); 941 zap_cursor_retrieve(&zc, &za) == 0; 942 (void) zap_cursor_advance(&zc)) { 943 if (za.za_integer_length != 8 || za.za_num_integers != 1) 944 return (EINVAL); 945 err = zap_add(os, intoobj, za.za_name, 946 8, 1, &za.za_first_integer, tx); 947 if (err) 948 return (err); 949 } 950 zap_cursor_fini(&zc); 951 return (0); 952 } 953 954 int 955 zap_add_int(objset_t *os, uint64_t obj, uint64_t value, dmu_tx_t *tx) 956 { 957 char name[20]; 958 959 (void) snprintf(name, sizeof (name), "%llx", (longlong_t)value); 960 return (zap_add(os, obj, name, 8, 1, &value, tx)); 961 } 962 963 int 964 zap_remove_int(objset_t *os, uint64_t obj, uint64_t value, dmu_tx_t *tx) 965 { 966 char name[20]; 967 968 (void) snprintf(name, sizeof (name), "%llx", (longlong_t)value); 969 return (zap_remove(os, obj, name, tx)); 970 } 971 972 int 973 zap_lookup_int(objset_t *os, uint64_t obj, uint64_t value) 974 { 975 char name[20]; 976 977 (void) snprintf(name, sizeof (name), "%llx", (longlong_t)value); 978 return (zap_lookup(os, obj, name, 8, 1, &value)); 979 } 980 981 int 982 zap_increment_int(objset_t *os, uint64_t obj, uint64_t key, int64_t delta, 983 dmu_tx_t *tx) 984 { 985 char name[20]; 986 uint64_t value = 0; 987 int err; 988 989 if (delta == 0) 990 return (0); 991 992 (void) snprintf(name, sizeof (name), "%llx", (longlong_t)key); 993 err = zap_lookup(os, obj, name, 8, 1, &value); 994 if (err != 0 && err != ENOENT) 995 return (err); 996 value += delta; 997 if (value == 0) 998 err = zap_remove(os, obj, name, tx); 999 else 1000 err = zap_update(os, obj, name, 8, 1, &value, tx); 1001 return (err); 1002 } 1003 1004 1005 /* 1006 * Routines for iterating over the attributes. 1007 */ 1008 1009 int 1010 fzap_cursor_retrieve(zap_t *zap, zap_cursor_t *zc, zap_attribute_t *za) 1011 { 1012 int err = ENOENT; 1013 zap_entry_handle_t zeh; 1014 zap_leaf_t *l; 1015 1016 /* retrieve the next entry at or after zc_hash/zc_cd */ 1017 /* if no entry, return ENOENT */ 1018 1019 if (zc->zc_leaf && 1020 (ZAP_HASH_IDX(zc->zc_hash, 1021 zc->zc_leaf->l_phys->l_hdr.lh_prefix_len) != 1022 zc->zc_leaf->l_phys->l_hdr.lh_prefix)) { 1023 rw_enter(&zc->zc_leaf->l_rwlock, RW_READER); 1024 zap_put_leaf(zc->zc_leaf); 1025 zc->zc_leaf = NULL; 1026 } 1027 1028 again: 1029 if (zc->zc_leaf == NULL) { 1030 err = zap_deref_leaf(zap, zc->zc_hash, NULL, RW_READER, 1031 &zc->zc_leaf); 1032 if (err != 0) 1033 return (err); 1034 } else { 1035 rw_enter(&zc->zc_leaf->l_rwlock, RW_READER); 1036 } 1037 l = zc->zc_leaf; 1038 1039 err = zap_leaf_lookup_closest(l, zc->zc_hash, zc->zc_cd, &zeh); 1040 1041 if (err == ENOENT) { 1042 uint64_t nocare = 1043 (1ULL << (64 - l->l_phys->l_hdr.lh_prefix_len)) - 1; 1044 zc->zc_hash = (zc->zc_hash & ~nocare) + nocare + 1; 1045 zc->zc_cd = 0; 1046 if (l->l_phys->l_hdr.lh_prefix_len == 0 || zc->zc_hash == 0) { 1047 zc->zc_hash = -1ULL; 1048 } else { 1049 zap_put_leaf(zc->zc_leaf); 1050 zc->zc_leaf = NULL; 1051 goto again; 1052 } 1053 } 1054 1055 if (err == 0) { 1056 zc->zc_hash = zeh.zeh_hash; 1057 zc->zc_cd = zeh.zeh_cd; 1058 za->za_integer_length = zeh.zeh_integer_size; 1059 za->za_num_integers = zeh.zeh_num_integers; 1060 if (zeh.zeh_num_integers == 0) { 1061 za->za_first_integer = 0; 1062 } else { 1063 err = zap_entry_read(&zeh, 8, 1, &za->za_first_integer); 1064 ASSERT(err == 0 || err == EOVERFLOW); 1065 } 1066 err = zap_entry_read_name(&zeh, 1067 sizeof (za->za_name), za->za_name); 1068 ASSERT(err == 0); 1069 1070 za->za_normalization_conflict = 1071 zap_entry_normalization_conflict(&zeh, 1072 NULL, za->za_name, zap); 1073 } 1074 rw_exit(&zc->zc_leaf->l_rwlock); 1075 return (err); 1076 } 1077 1078 1079 static void 1080 zap_stats_ptrtbl(zap_t *zap, uint64_t *tbl, int len, zap_stats_t *zs) 1081 { 1082 int i, err; 1083 uint64_t lastblk = 0; 1084 1085 /* 1086 * NB: if a leaf has more pointers than an entire ptrtbl block 1087 * can hold, then it'll be accounted for more than once, since 1088 * we won't have lastblk. 1089 */ 1090 for (i = 0; i < len; i++) { 1091 zap_leaf_t *l; 1092 1093 if (tbl[i] == lastblk) 1094 continue; 1095 lastblk = tbl[i]; 1096 1097 err = zap_get_leaf_byblk(zap, tbl[i], NULL, RW_READER, &l); 1098 if (err == 0) { 1099 zap_leaf_stats(zap, l, zs); 1100 zap_put_leaf(l); 1101 } 1102 } 1103 } 1104 1105 void 1106 fzap_get_stats(zap_t *zap, zap_stats_t *zs) 1107 { 1108 int bs = FZAP_BLOCK_SHIFT(zap); 1109 zs->zs_blocksize = 1ULL << bs; 1110 1111 /* 1112 * Set zap_phys_t fields 1113 */ 1114 zs->zs_num_leafs = zap->zap_f.zap_phys->zap_num_leafs; 1115 zs->zs_num_entries = zap->zap_f.zap_phys->zap_num_entries; 1116 zs->zs_num_blocks = zap->zap_f.zap_phys->zap_freeblk; 1117 zs->zs_block_type = zap->zap_f.zap_phys->zap_block_type; 1118 zs->zs_magic = zap->zap_f.zap_phys->zap_magic; 1119 zs->zs_salt = zap->zap_f.zap_phys->zap_salt; 1120 1121 /* 1122 * Set zap_ptrtbl fields 1123 */ 1124 zs->zs_ptrtbl_len = 1ULL << zap->zap_f.zap_phys->zap_ptrtbl.zt_shift; 1125 zs->zs_ptrtbl_nextblk = zap->zap_f.zap_phys->zap_ptrtbl.zt_nextblk; 1126 zs->zs_ptrtbl_blks_copied = 1127 zap->zap_f.zap_phys->zap_ptrtbl.zt_blks_copied; 1128 zs->zs_ptrtbl_zt_blk = zap->zap_f.zap_phys->zap_ptrtbl.zt_blk; 1129 zs->zs_ptrtbl_zt_numblks = zap->zap_f.zap_phys->zap_ptrtbl.zt_numblks; 1130 zs->zs_ptrtbl_zt_shift = zap->zap_f.zap_phys->zap_ptrtbl.zt_shift; 1131 1132 if (zap->zap_f.zap_phys->zap_ptrtbl.zt_numblks == 0) { 1133 /* the ptrtbl is entirely in the header block. */ 1134 zap_stats_ptrtbl(zap, &ZAP_EMBEDDED_PTRTBL_ENT(zap, 0), 1135 1 << ZAP_EMBEDDED_PTRTBL_SHIFT(zap), zs); 1136 } else { 1137 int b; 1138 1139 dmu_prefetch(zap->zap_objset, zap->zap_object, 1140 zap->zap_f.zap_phys->zap_ptrtbl.zt_blk << bs, 1141 zap->zap_f.zap_phys->zap_ptrtbl.zt_numblks << bs); 1142 1143 for (b = 0; b < zap->zap_f.zap_phys->zap_ptrtbl.zt_numblks; 1144 b++) { 1145 dmu_buf_t *db; 1146 int err; 1147 1148 err = dmu_buf_hold(zap->zap_objset, zap->zap_object, 1149 (zap->zap_f.zap_phys->zap_ptrtbl.zt_blk + b) << bs, 1150 FTAG, &db); 1151 if (err == 0) { 1152 zap_stats_ptrtbl(zap, db->db_data, 1153 1<<(bs-3), zs); 1154 dmu_buf_rele(db, FTAG); 1155 } 1156 } 1157 } 1158 } 1159 1160 int 1161 fzap_count_write(zap_name_t *zn, int add, uint64_t *towrite, 1162 uint64_t *tooverwrite) 1163 { 1164 zap_t *zap = zn->zn_zap; 1165 zap_leaf_t *l; 1166 int err; 1167 1168 /* 1169 * Account for the header block of the fatzap. 1170 */ 1171 if (!add && dmu_buf_freeable(zap->zap_dbuf)) { 1172 *tooverwrite += zap->zap_dbuf->db_size; 1173 } else { 1174 *towrite += zap->zap_dbuf->db_size; 1175 } 1176 1177 /* 1178 * Account for the pointer table blocks. 1179 * If we are adding we need to account for the following cases : 1180 * - If the pointer table is embedded, this operation could force an 1181 * external pointer table. 1182 * - If this already has an external pointer table this operation 1183 * could extend the table. 1184 */ 1185 if (add) { 1186 if (zap->zap_f.zap_phys->zap_ptrtbl.zt_blk == 0) 1187 *towrite += zap->zap_dbuf->db_size; 1188 else 1189 *towrite += (zap->zap_dbuf->db_size * 3); 1190 } 1191 1192 /* 1193 * Now, check if the block containing leaf is freeable 1194 * and account accordingly. 1195 */ 1196 err = zap_deref_leaf(zap, zn->zn_hash, NULL, RW_READER, &l); 1197 if (err != 0) { 1198 return (err); 1199 } 1200 1201 if (!add && dmu_buf_freeable(l->l_dbuf)) { 1202 *tooverwrite += l->l_dbuf->db_size; 1203 } else { 1204 /* 1205 * If this an add operation, the leaf block could split. 1206 * Hence, we need to account for an additional leaf block. 1207 */ 1208 *towrite += (add ? 2 : 1) * l->l_dbuf->db_size; 1209 } 1210 1211 zap_put_leaf(l); 1212 return (0); 1213 } 1214