1 // SPDX-License-Identifier: GPL-2.0 2 /* 3 * Copyright (c) 2000-2001,2005 Silicon Graphics, Inc. 4 * All Rights Reserved. 5 */ 6 #include "xfs.h" 7 #include "xfs_fs.h" 8 #include "xfs_shared.h" 9 #include "xfs_format.h" 10 #include "xfs_log_format.h" 11 #include "xfs_trans_resv.h" 12 #include "xfs_bit.h" 13 #include "xfs_mount.h" 14 #include "xfs_inode.h" 15 #include "xfs_btree.h" 16 #include "xfs_ialloc.h" 17 #include "xfs_ialloc_btree.h" 18 #include "xfs_alloc.h" 19 #include "xfs_error.h" 20 #include "xfs_trace.h" 21 #include "xfs_cksum.h" 22 #include "xfs_trans.h" 23 #include "xfs_rmap.h" 24 25 26 STATIC int 27 xfs_inobt_get_minrecs( 28 struct xfs_btree_cur *cur, 29 int level) 30 { 31 return cur->bc_mp->m_inobt_mnr[level != 0]; 32 } 33 34 STATIC struct xfs_btree_cur * 35 xfs_inobt_dup_cursor( 36 struct xfs_btree_cur *cur) 37 { 38 return xfs_inobt_init_cursor(cur->bc_mp, cur->bc_tp, 39 cur->bc_private.a.agbp, cur->bc_private.a.agno, 40 cur->bc_btnum); 41 } 42 43 STATIC void 44 xfs_inobt_set_root( 45 struct xfs_btree_cur *cur, 46 union xfs_btree_ptr *nptr, 47 int inc) /* level change */ 48 { 49 struct xfs_buf *agbp = cur->bc_private.a.agbp; 50 struct xfs_agi *agi = XFS_BUF_TO_AGI(agbp); 51 52 agi->agi_root = nptr->s; 53 be32_add_cpu(&agi->agi_level, inc); 54 xfs_ialloc_log_agi(cur->bc_tp, agbp, XFS_AGI_ROOT | XFS_AGI_LEVEL); 55 } 56 57 STATIC void 58 xfs_finobt_set_root( 59 struct xfs_btree_cur *cur, 60 union xfs_btree_ptr *nptr, 61 int inc) /* level change */ 62 { 63 struct xfs_buf *agbp = cur->bc_private.a.agbp; 64 struct xfs_agi *agi = XFS_BUF_TO_AGI(agbp); 65 66 agi->agi_free_root = nptr->s; 67 be32_add_cpu(&agi->agi_free_level, inc); 68 xfs_ialloc_log_agi(cur->bc_tp, agbp, 69 XFS_AGI_FREE_ROOT | XFS_AGI_FREE_LEVEL); 70 } 71 72 STATIC int 73 __xfs_inobt_alloc_block( 74 struct xfs_btree_cur *cur, 75 union xfs_btree_ptr *start, 76 union xfs_btree_ptr *new, 77 int *stat, 78 enum xfs_ag_resv_type resv) 79 { 80 xfs_alloc_arg_t args; /* block allocation args */ 81 int error; /* error return value */ 82 xfs_agblock_t sbno = be32_to_cpu(start->s); 83 84 memset(&args, 0, sizeof(args)); 85 args.tp = cur->bc_tp; 86 args.mp = cur->bc_mp; 87 xfs_rmap_ag_owner(&args.oinfo, XFS_RMAP_OWN_INOBT); 88 args.fsbno = XFS_AGB_TO_FSB(args.mp, cur->bc_private.a.agno, sbno); 89 args.minlen = 1; 90 args.maxlen = 1; 91 args.prod = 1; 92 args.type = XFS_ALLOCTYPE_NEAR_BNO; 93 args.resv = resv; 94 95 error = xfs_alloc_vextent(&args); 96 if (error) 97 return error; 98 99 if (args.fsbno == NULLFSBLOCK) { 100 *stat = 0; 101 return 0; 102 } 103 ASSERT(args.len == 1); 104 105 new->s = cpu_to_be32(XFS_FSB_TO_AGBNO(args.mp, args.fsbno)); 106 *stat = 1; 107 return 0; 108 } 109 110 STATIC int 111 xfs_inobt_alloc_block( 112 struct xfs_btree_cur *cur, 113 union xfs_btree_ptr *start, 114 union xfs_btree_ptr *new, 115 int *stat) 116 { 117 return __xfs_inobt_alloc_block(cur, start, new, stat, XFS_AG_RESV_NONE); 118 } 119 120 STATIC int 121 xfs_finobt_alloc_block( 122 struct xfs_btree_cur *cur, 123 union xfs_btree_ptr *start, 124 union xfs_btree_ptr *new, 125 int *stat) 126 { 127 if (cur->bc_mp->m_inotbt_nores) 128 return xfs_inobt_alloc_block(cur, start, new, stat); 129 return __xfs_inobt_alloc_block(cur, start, new, stat, 130 XFS_AG_RESV_METADATA); 131 } 132 133 STATIC int 134 __xfs_inobt_free_block( 135 struct xfs_btree_cur *cur, 136 struct xfs_buf *bp, 137 enum xfs_ag_resv_type resv) 138 { 139 struct xfs_owner_info oinfo; 140 141 xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_INOBT); 142 return xfs_free_extent(cur->bc_tp, 143 XFS_DADDR_TO_FSB(cur->bc_mp, XFS_BUF_ADDR(bp)), 1, 144 &oinfo, resv); 145 } 146 147 STATIC int 148 xfs_inobt_free_block( 149 struct xfs_btree_cur *cur, 150 struct xfs_buf *bp) 151 { 152 return __xfs_inobt_free_block(cur, bp, XFS_AG_RESV_NONE); 153 } 154 155 STATIC int 156 xfs_finobt_free_block( 157 struct xfs_btree_cur *cur, 158 struct xfs_buf *bp) 159 { 160 if (cur->bc_mp->m_inotbt_nores) 161 return xfs_inobt_free_block(cur, bp); 162 return __xfs_inobt_free_block(cur, bp, XFS_AG_RESV_METADATA); 163 } 164 165 STATIC int 166 xfs_inobt_get_maxrecs( 167 struct xfs_btree_cur *cur, 168 int level) 169 { 170 return cur->bc_mp->m_inobt_mxr[level != 0]; 171 } 172 173 STATIC void 174 xfs_inobt_init_key_from_rec( 175 union xfs_btree_key *key, 176 union xfs_btree_rec *rec) 177 { 178 key->inobt.ir_startino = rec->inobt.ir_startino; 179 } 180 181 STATIC void 182 xfs_inobt_init_high_key_from_rec( 183 union xfs_btree_key *key, 184 union xfs_btree_rec *rec) 185 { 186 __u32 x; 187 188 x = be32_to_cpu(rec->inobt.ir_startino); 189 x += XFS_INODES_PER_CHUNK - 1; 190 key->inobt.ir_startino = cpu_to_be32(x); 191 } 192 193 STATIC void 194 xfs_inobt_init_rec_from_cur( 195 struct xfs_btree_cur *cur, 196 union xfs_btree_rec *rec) 197 { 198 rec->inobt.ir_startino = cpu_to_be32(cur->bc_rec.i.ir_startino); 199 if (xfs_sb_version_hassparseinodes(&cur->bc_mp->m_sb)) { 200 rec->inobt.ir_u.sp.ir_holemask = 201 cpu_to_be16(cur->bc_rec.i.ir_holemask); 202 rec->inobt.ir_u.sp.ir_count = cur->bc_rec.i.ir_count; 203 rec->inobt.ir_u.sp.ir_freecount = cur->bc_rec.i.ir_freecount; 204 } else { 205 /* ir_holemask/ir_count not supported on-disk */ 206 rec->inobt.ir_u.f.ir_freecount = 207 cpu_to_be32(cur->bc_rec.i.ir_freecount); 208 } 209 rec->inobt.ir_free = cpu_to_be64(cur->bc_rec.i.ir_free); 210 } 211 212 /* 213 * initial value of ptr for lookup 214 */ 215 STATIC void 216 xfs_inobt_init_ptr_from_cur( 217 struct xfs_btree_cur *cur, 218 union xfs_btree_ptr *ptr) 219 { 220 struct xfs_agi *agi = XFS_BUF_TO_AGI(cur->bc_private.a.agbp); 221 222 ASSERT(cur->bc_private.a.agno == be32_to_cpu(agi->agi_seqno)); 223 224 ptr->s = agi->agi_root; 225 } 226 227 STATIC void 228 xfs_finobt_init_ptr_from_cur( 229 struct xfs_btree_cur *cur, 230 union xfs_btree_ptr *ptr) 231 { 232 struct xfs_agi *agi = XFS_BUF_TO_AGI(cur->bc_private.a.agbp); 233 234 ASSERT(cur->bc_private.a.agno == be32_to_cpu(agi->agi_seqno)); 235 ptr->s = agi->agi_free_root; 236 } 237 238 STATIC int64_t 239 xfs_inobt_key_diff( 240 struct xfs_btree_cur *cur, 241 union xfs_btree_key *key) 242 { 243 return (int64_t)be32_to_cpu(key->inobt.ir_startino) - 244 cur->bc_rec.i.ir_startino; 245 } 246 247 STATIC int64_t 248 xfs_inobt_diff_two_keys( 249 struct xfs_btree_cur *cur, 250 union xfs_btree_key *k1, 251 union xfs_btree_key *k2) 252 { 253 return (int64_t)be32_to_cpu(k1->inobt.ir_startino) - 254 be32_to_cpu(k2->inobt.ir_startino); 255 } 256 257 static xfs_failaddr_t 258 xfs_inobt_verify( 259 struct xfs_buf *bp) 260 { 261 struct xfs_mount *mp = bp->b_target->bt_mount; 262 struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp); 263 xfs_failaddr_t fa; 264 unsigned int level; 265 266 /* 267 * During growfs operations, we can't verify the exact owner as the 268 * perag is not fully initialised and hence not attached to the buffer. 269 * 270 * Similarly, during log recovery we will have a perag structure 271 * attached, but the agi information will not yet have been initialised 272 * from the on disk AGI. We don't currently use any of this information, 273 * but beware of the landmine (i.e. need to check pag->pagi_init) if we 274 * ever do. 275 */ 276 switch (block->bb_magic) { 277 case cpu_to_be32(XFS_IBT_CRC_MAGIC): 278 case cpu_to_be32(XFS_FIBT_CRC_MAGIC): 279 fa = xfs_btree_sblock_v5hdr_verify(bp); 280 if (fa) 281 return fa; 282 /* fall through */ 283 case cpu_to_be32(XFS_IBT_MAGIC): 284 case cpu_to_be32(XFS_FIBT_MAGIC): 285 break; 286 default: 287 return __this_address; 288 } 289 290 /* level verification */ 291 level = be16_to_cpu(block->bb_level); 292 if (level >= mp->m_in_maxlevels) 293 return __this_address; 294 295 return xfs_btree_sblock_verify(bp, mp->m_inobt_mxr[level != 0]); 296 } 297 298 static void 299 xfs_inobt_read_verify( 300 struct xfs_buf *bp) 301 { 302 xfs_failaddr_t fa; 303 304 if (!xfs_btree_sblock_verify_crc(bp)) 305 xfs_verifier_error(bp, -EFSBADCRC, __this_address); 306 else { 307 fa = xfs_inobt_verify(bp); 308 if (fa) 309 xfs_verifier_error(bp, -EFSCORRUPTED, fa); 310 } 311 312 if (bp->b_error) 313 trace_xfs_btree_corrupt(bp, _RET_IP_); 314 } 315 316 static void 317 xfs_inobt_write_verify( 318 struct xfs_buf *bp) 319 { 320 xfs_failaddr_t fa; 321 322 fa = xfs_inobt_verify(bp); 323 if (fa) { 324 trace_xfs_btree_corrupt(bp, _RET_IP_); 325 xfs_verifier_error(bp, -EFSCORRUPTED, fa); 326 return; 327 } 328 xfs_btree_sblock_calc_crc(bp); 329 330 } 331 332 const struct xfs_buf_ops xfs_inobt_buf_ops = { 333 .name = "xfs_inobt", 334 .verify_read = xfs_inobt_read_verify, 335 .verify_write = xfs_inobt_write_verify, 336 .verify_struct = xfs_inobt_verify, 337 }; 338 339 STATIC int 340 xfs_inobt_keys_inorder( 341 struct xfs_btree_cur *cur, 342 union xfs_btree_key *k1, 343 union xfs_btree_key *k2) 344 { 345 return be32_to_cpu(k1->inobt.ir_startino) < 346 be32_to_cpu(k2->inobt.ir_startino); 347 } 348 349 STATIC int 350 xfs_inobt_recs_inorder( 351 struct xfs_btree_cur *cur, 352 union xfs_btree_rec *r1, 353 union xfs_btree_rec *r2) 354 { 355 return be32_to_cpu(r1->inobt.ir_startino) + XFS_INODES_PER_CHUNK <= 356 be32_to_cpu(r2->inobt.ir_startino); 357 } 358 359 static const struct xfs_btree_ops xfs_inobt_ops = { 360 .rec_len = sizeof(xfs_inobt_rec_t), 361 .key_len = sizeof(xfs_inobt_key_t), 362 363 .dup_cursor = xfs_inobt_dup_cursor, 364 .set_root = xfs_inobt_set_root, 365 .alloc_block = xfs_inobt_alloc_block, 366 .free_block = xfs_inobt_free_block, 367 .get_minrecs = xfs_inobt_get_minrecs, 368 .get_maxrecs = xfs_inobt_get_maxrecs, 369 .init_key_from_rec = xfs_inobt_init_key_from_rec, 370 .init_high_key_from_rec = xfs_inobt_init_high_key_from_rec, 371 .init_rec_from_cur = xfs_inobt_init_rec_from_cur, 372 .init_ptr_from_cur = xfs_inobt_init_ptr_from_cur, 373 .key_diff = xfs_inobt_key_diff, 374 .buf_ops = &xfs_inobt_buf_ops, 375 .diff_two_keys = xfs_inobt_diff_two_keys, 376 .keys_inorder = xfs_inobt_keys_inorder, 377 .recs_inorder = xfs_inobt_recs_inorder, 378 }; 379 380 static const struct xfs_btree_ops xfs_finobt_ops = { 381 .rec_len = sizeof(xfs_inobt_rec_t), 382 .key_len = sizeof(xfs_inobt_key_t), 383 384 .dup_cursor = xfs_inobt_dup_cursor, 385 .set_root = xfs_finobt_set_root, 386 .alloc_block = xfs_finobt_alloc_block, 387 .free_block = xfs_finobt_free_block, 388 .get_minrecs = xfs_inobt_get_minrecs, 389 .get_maxrecs = xfs_inobt_get_maxrecs, 390 .init_key_from_rec = xfs_inobt_init_key_from_rec, 391 .init_high_key_from_rec = xfs_inobt_init_high_key_from_rec, 392 .init_rec_from_cur = xfs_inobt_init_rec_from_cur, 393 .init_ptr_from_cur = xfs_finobt_init_ptr_from_cur, 394 .key_diff = xfs_inobt_key_diff, 395 .buf_ops = &xfs_inobt_buf_ops, 396 .diff_two_keys = xfs_inobt_diff_two_keys, 397 .keys_inorder = xfs_inobt_keys_inorder, 398 .recs_inorder = xfs_inobt_recs_inorder, 399 }; 400 401 /* 402 * Allocate a new inode btree cursor. 403 */ 404 struct xfs_btree_cur * /* new inode btree cursor */ 405 xfs_inobt_init_cursor( 406 struct xfs_mount *mp, /* file system mount point */ 407 struct xfs_trans *tp, /* transaction pointer */ 408 struct xfs_buf *agbp, /* buffer for agi structure */ 409 xfs_agnumber_t agno, /* allocation group number */ 410 xfs_btnum_t btnum) /* ialloc or free ino btree */ 411 { 412 struct xfs_agi *agi = XFS_BUF_TO_AGI(agbp); 413 struct xfs_btree_cur *cur; 414 415 cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS); 416 417 cur->bc_tp = tp; 418 cur->bc_mp = mp; 419 cur->bc_btnum = btnum; 420 if (btnum == XFS_BTNUM_INO) { 421 cur->bc_nlevels = be32_to_cpu(agi->agi_level); 422 cur->bc_ops = &xfs_inobt_ops; 423 cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_ibt_2); 424 } else { 425 cur->bc_nlevels = be32_to_cpu(agi->agi_free_level); 426 cur->bc_ops = &xfs_finobt_ops; 427 cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_fibt_2); 428 } 429 430 cur->bc_blocklog = mp->m_sb.sb_blocklog; 431 432 if (xfs_sb_version_hascrc(&mp->m_sb)) 433 cur->bc_flags |= XFS_BTREE_CRC_BLOCKS; 434 435 cur->bc_private.a.agbp = agbp; 436 cur->bc_private.a.agno = agno; 437 438 return cur; 439 } 440 441 /* 442 * Calculate number of records in an inobt btree block. 443 */ 444 int 445 xfs_inobt_maxrecs( 446 struct xfs_mount *mp, 447 int blocklen, 448 int leaf) 449 { 450 blocklen -= XFS_INOBT_BLOCK_LEN(mp); 451 452 if (leaf) 453 return blocklen / sizeof(xfs_inobt_rec_t); 454 return blocklen / (sizeof(xfs_inobt_key_t) + sizeof(xfs_inobt_ptr_t)); 455 } 456 457 /* 458 * Convert the inode record holemask to an inode allocation bitmap. The inode 459 * allocation bitmap is inode granularity and specifies whether an inode is 460 * physically allocated on disk (not whether the inode is considered allocated 461 * or free by the fs). 462 * 463 * A bit value of 1 means the inode is allocated, a value of 0 means it is free. 464 */ 465 uint64_t 466 xfs_inobt_irec_to_allocmask( 467 struct xfs_inobt_rec_incore *rec) 468 { 469 uint64_t bitmap = 0; 470 uint64_t inodespbit; 471 int nextbit; 472 uint allocbitmap; 473 474 /* 475 * The holemask has 16-bits for a 64 inode record. Therefore each 476 * holemask bit represents multiple inodes. Create a mask of bits to set 477 * in the allocmask for each holemask bit. 478 */ 479 inodespbit = (1 << XFS_INODES_PER_HOLEMASK_BIT) - 1; 480 481 /* 482 * Allocated inodes are represented by 0 bits in holemask. Invert the 0 483 * bits to 1 and convert to a uint so we can use xfs_next_bit(). Mask 484 * anything beyond the 16 holemask bits since this casts to a larger 485 * type. 486 */ 487 allocbitmap = ~rec->ir_holemask & ((1 << XFS_INOBT_HOLEMASK_BITS) - 1); 488 489 /* 490 * allocbitmap is the inverted holemask so every set bit represents 491 * allocated inodes. To expand from 16-bit holemask granularity to 492 * 64-bit (e.g., bit-per-inode), set inodespbit bits in the target 493 * bitmap for every holemask bit. 494 */ 495 nextbit = xfs_next_bit(&allocbitmap, 1, 0); 496 while (nextbit != -1) { 497 ASSERT(nextbit < (sizeof(rec->ir_holemask) * NBBY)); 498 499 bitmap |= (inodespbit << 500 (nextbit * XFS_INODES_PER_HOLEMASK_BIT)); 501 502 nextbit = xfs_next_bit(&allocbitmap, 1, nextbit + 1); 503 } 504 505 return bitmap; 506 } 507 508 #if defined(DEBUG) || defined(XFS_WARN) 509 /* 510 * Verify that an in-core inode record has a valid inode count. 511 */ 512 int 513 xfs_inobt_rec_check_count( 514 struct xfs_mount *mp, 515 struct xfs_inobt_rec_incore *rec) 516 { 517 int inocount = 0; 518 int nextbit = 0; 519 uint64_t allocbmap; 520 int wordsz; 521 522 wordsz = sizeof(allocbmap) / sizeof(unsigned int); 523 allocbmap = xfs_inobt_irec_to_allocmask(rec); 524 525 nextbit = xfs_next_bit((uint *) &allocbmap, wordsz, nextbit); 526 while (nextbit != -1) { 527 inocount++; 528 nextbit = xfs_next_bit((uint *) &allocbmap, wordsz, 529 nextbit + 1); 530 } 531 532 if (inocount != rec->ir_count) 533 return -EFSCORRUPTED; 534 535 return 0; 536 } 537 #endif /* DEBUG */ 538 539 static xfs_extlen_t 540 xfs_inobt_max_size( 541 struct xfs_mount *mp) 542 { 543 /* Bail out if we're uninitialized, which can happen in mkfs. */ 544 if (mp->m_inobt_mxr[0] == 0) 545 return 0; 546 547 return xfs_btree_calc_size(mp->m_inobt_mnr, 548 (uint64_t)mp->m_sb.sb_agblocks * mp->m_sb.sb_inopblock / 549 XFS_INODES_PER_CHUNK); 550 } 551 552 static int 553 xfs_inobt_count_blocks( 554 struct xfs_mount *mp, 555 struct xfs_trans *tp, 556 xfs_agnumber_t agno, 557 xfs_btnum_t btnum, 558 xfs_extlen_t *tree_blocks) 559 { 560 struct xfs_buf *agbp; 561 struct xfs_btree_cur *cur; 562 int error; 563 564 error = xfs_ialloc_read_agi(mp, tp, agno, &agbp); 565 if (error) 566 return error; 567 568 cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, btnum); 569 error = xfs_btree_count_blocks(cur, tree_blocks); 570 xfs_btree_del_cursor(cur, error); 571 xfs_trans_brelse(tp, agbp); 572 573 return error; 574 } 575 576 /* 577 * Figure out how many blocks to reserve and how many are used by this btree. 578 */ 579 int 580 xfs_finobt_calc_reserves( 581 struct xfs_mount *mp, 582 struct xfs_trans *tp, 583 xfs_agnumber_t agno, 584 xfs_extlen_t *ask, 585 xfs_extlen_t *used) 586 { 587 xfs_extlen_t tree_len = 0; 588 int error; 589 590 if (!xfs_sb_version_hasfinobt(&mp->m_sb)) 591 return 0; 592 593 error = xfs_inobt_count_blocks(mp, tp, agno, XFS_BTNUM_FINO, &tree_len); 594 if (error) 595 return error; 596 597 *ask += xfs_inobt_max_size(mp); 598 *used += tree_len; 599 return 0; 600 } 601 602 /* Calculate the inobt btree size for some records. */ 603 xfs_extlen_t 604 xfs_iallocbt_calc_size( 605 struct xfs_mount *mp, 606 unsigned long long len) 607 { 608 return xfs_btree_calc_size(mp->m_inobt_mnr, len); 609 } 610