1/* SPDX-License-Identifier: GPL-2.0 */ 2/* 3 * Copyright (c) 2000-2005 Silicon Graphics, Inc. 4 * Copyright (c) 2018 Red Hat, Inc. 5 * All rights reserved. 6 */ 7 8#include "xfs.h" 9#include "xfs_fs.h" 10#include "xfs_shared.h" 11#include "xfs_format.h" 12#include "xfs_trans_resv.h" 13#include "xfs_bit.h" 14#include "xfs_sb.h" 15#include "xfs_mount.h" 16#include "xfs_btree.h" 17#include "xfs_alloc_btree.h" 18#include "xfs_rmap_btree.h" 19#include "xfs_alloc.h" 20#include "xfs_ialloc.h" 21#include "xfs_rmap.h" 22#include "xfs_ag.h" 23#include "xfs_ag_resv.h" 24#include "xfs_health.h" 25 26static int 27xfs_get_aghdr_buf( 28 struct xfs_mount *mp, 29 xfs_daddr_t blkno, 30 size_t numblks, 31 struct xfs_buf **bpp, 32 const struct xfs_buf_ops *ops) 33{ 34 struct xfs_buf *bp; 35 int error; 36 37 error = xfs_buf_get_uncached(mp->m_ddev_targp, numblks, 0, &bp); 38 if (error) 39 return error; 40 41 xfs_buf_zero(bp, 0, BBTOB(bp->b_length)); 42 bp->b_bn = blkno; 43 bp->b_maps[0].bm_bn = blkno; 44 bp->b_ops = ops; 45 46 *bpp = bp; 47 return 0; 48} 49 50static inline bool is_log_ag(struct xfs_mount *mp, struct aghdr_init_data *id) 51{ 52 return mp->m_sb.sb_logstart > 0 && 53 id->agno == XFS_FSB_TO_AGNO(mp, mp->m_sb.sb_logstart); 54} 55 56/* 57 * Generic btree root block init function 58 */ 59static void 60xfs_btroot_init( 61 struct xfs_mount *mp, 62 struct xfs_buf *bp, 63 struct aghdr_init_data *id) 64{ 65 xfs_btree_init_block(mp, bp, id->type, 0, 0, id->agno); 66} 67 68/* Finish initializing a free space btree. */ 69static void 70xfs_freesp_init_recs( 71 struct xfs_mount *mp, 72 struct xfs_buf *bp, 73 struct aghdr_init_data *id) 74{ 75 struct xfs_alloc_rec *arec; 76 struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp); 77 78 arec = XFS_ALLOC_REC_ADDR(mp, XFS_BUF_TO_BLOCK(bp), 1); 79 arec->ar_startblock = cpu_to_be32(mp->m_ag_prealloc_blocks); 80 81 if (is_log_ag(mp, id)) { 82 struct xfs_alloc_rec *nrec; 83 xfs_agblock_t start = XFS_FSB_TO_AGBNO(mp, 84 mp->m_sb.sb_logstart); 85 86 ASSERT(start >= mp->m_ag_prealloc_blocks); 87 if (start != mp->m_ag_prealloc_blocks) { 88 /* 89 * Modify first record to pad stripe align of log 90 */ 91 arec->ar_blockcount = cpu_to_be32(start - 92 mp->m_ag_prealloc_blocks); 93 nrec = arec + 1; 94 95 /* 96 * Insert second record at start of internal log 97 * which then gets trimmed. 98 */ 99 nrec->ar_startblock = cpu_to_be32( 100 be32_to_cpu(arec->ar_startblock) + 101 be32_to_cpu(arec->ar_blockcount)); 102 arec = nrec; 103 be16_add_cpu(&block->bb_numrecs, 1); 104 } 105 /* 106 * Change record start to after the internal log 107 */ 108 be32_add_cpu(&arec->ar_startblock, mp->m_sb.sb_logblocks); 109 } 110 111 /* 112 * Calculate the record block count and check for the case where 113 * the log might have consumed all available space in the AG. If 114 * so, reset the record count to 0 to avoid exposure of an invalid 115 * record start block. 116 */ 117 arec->ar_blockcount = cpu_to_be32(id->agsize - 118 be32_to_cpu(arec->ar_startblock)); 119 if (!arec->ar_blockcount) 120 block->bb_numrecs = 0; 121} 122 123/* 124 * Alloc btree root block init functions 125 */ 126static void 127xfs_bnoroot_init( 128 struct xfs_mount *mp, 129 struct xfs_buf *bp, 130 struct aghdr_init_data *id) 131{ 132 xfs_btree_init_block(mp, bp, XFS_BTNUM_BNO, 0, 1, id->agno); 133 xfs_freesp_init_recs(mp, bp, id); 134} 135 136static void 137xfs_cntroot_init( 138 struct xfs_mount *mp, 139 struct xfs_buf *bp, 140 struct aghdr_init_data *id) 141{ 142 xfs_btree_init_block(mp, bp, XFS_BTNUM_CNT, 0, 1, id->agno); 143 xfs_freesp_init_recs(mp, bp, id); 144} 145 146/* 147 * Reverse map root block init 148 */ 149static void 150xfs_rmaproot_init( 151 struct xfs_mount *mp, 152 struct xfs_buf *bp, 153 struct aghdr_init_data *id) 154{ 155 struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp); 156 struct xfs_rmap_rec *rrec; 157 158 xfs_btree_init_block(mp, bp, XFS_BTNUM_RMAP, 0, 4, id->agno); 159 160 /* 161 * mark the AG header regions as static metadata The BNO 162 * btree block is the first block after the headers, so 163 * it's location defines the size of region the static 164 * metadata consumes. 165 * 166 * Note: unlike mkfs, we never have to account for log 167 * space when growing the data regions 168 */ 169 rrec = XFS_RMAP_REC_ADDR(block, 1); 170 rrec->rm_startblock = 0; 171 rrec->rm_blockcount = cpu_to_be32(XFS_BNO_BLOCK(mp)); 172 rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_FS); 173 rrec->rm_offset = 0; 174 175 /* account freespace btree root blocks */ 176 rrec = XFS_RMAP_REC_ADDR(block, 2); 177 rrec->rm_startblock = cpu_to_be32(XFS_BNO_BLOCK(mp)); 178 rrec->rm_blockcount = cpu_to_be32(2); 179 rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_AG); 180 rrec->rm_offset = 0; 181 182 /* account inode btree root blocks */ 183 rrec = XFS_RMAP_REC_ADDR(block, 3); 184 rrec->rm_startblock = cpu_to_be32(XFS_IBT_BLOCK(mp)); 185 rrec->rm_blockcount = cpu_to_be32(XFS_RMAP_BLOCK(mp) - 186 XFS_IBT_BLOCK(mp)); 187 rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_INOBT); 188 rrec->rm_offset = 0; 189 190 /* account for rmap btree root */ 191 rrec = XFS_RMAP_REC_ADDR(block, 4); 192 rrec->rm_startblock = cpu_to_be32(XFS_RMAP_BLOCK(mp)); 193 rrec->rm_blockcount = cpu_to_be32(1); 194 rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_AG); 195 rrec->rm_offset = 0; 196 197 /* account for refc btree root */ 198 if (xfs_sb_version_hasreflink(&mp->m_sb)) { 199 rrec = XFS_RMAP_REC_ADDR(block, 5); 200 rrec->rm_startblock = cpu_to_be32(xfs_refc_block(mp)); 201 rrec->rm_blockcount = cpu_to_be32(1); 202 rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_REFC); 203 rrec->rm_offset = 0; 204 be16_add_cpu(&block->bb_numrecs, 1); 205 } 206 207 /* account for the log space */ 208 if (is_log_ag(mp, id)) { 209 rrec = XFS_RMAP_REC_ADDR(block, 210 be16_to_cpu(block->bb_numrecs) + 1); 211 rrec->rm_startblock = cpu_to_be32( 212 XFS_FSB_TO_AGBNO(mp, mp->m_sb.sb_logstart)); 213 rrec->rm_blockcount = cpu_to_be32(mp->m_sb.sb_logblocks); 214 rrec->rm_owner = cpu_to_be64(XFS_RMAP_OWN_LOG); 215 rrec->rm_offset = 0; 216 be16_add_cpu(&block->bb_numrecs, 1); 217 } 218} 219 220/* 221 * Initialise new secondary superblocks with the pre-grow geometry, but mark 222 * them as "in progress" so we know they haven't yet been activated. This will 223 * get cleared when the update with the new geometry information is done after 224 * changes to the primary are committed. This isn't strictly necessary, but we 225 * get it for free with the delayed buffer write lists and it means we can tell 226 * if a grow operation didn't complete properly after the fact. 227 */ 228static void 229xfs_sbblock_init( 230 struct xfs_mount *mp, 231 struct xfs_buf *bp, 232 struct aghdr_init_data *id) 233{ 234 struct xfs_dsb *dsb = bp->b_addr; 235 236 xfs_sb_to_disk(dsb, &mp->m_sb); 237 dsb->sb_inprogress = 1; 238} 239 240static void 241xfs_agfblock_init( 242 struct xfs_mount *mp, 243 struct xfs_buf *bp, 244 struct aghdr_init_data *id) 245{ 246 struct xfs_agf *agf = bp->b_addr; 247 xfs_extlen_t tmpsize; 248 249 agf->agf_magicnum = cpu_to_be32(XFS_AGF_MAGIC); 250 agf->agf_versionnum = cpu_to_be32(XFS_AGF_VERSION); 251 agf->agf_seqno = cpu_to_be32(id->agno); 252 agf->agf_length = cpu_to_be32(id->agsize); 253 agf->agf_roots[XFS_BTNUM_BNOi] = cpu_to_be32(XFS_BNO_BLOCK(mp)); 254 agf->agf_roots[XFS_BTNUM_CNTi] = cpu_to_be32(XFS_CNT_BLOCK(mp)); 255 agf->agf_levels[XFS_BTNUM_BNOi] = cpu_to_be32(1); 256 agf->agf_levels[XFS_BTNUM_CNTi] = cpu_to_be32(1); 257 if (xfs_sb_version_hasrmapbt(&mp->m_sb)) { 258 agf->agf_roots[XFS_BTNUM_RMAPi] = 259 cpu_to_be32(XFS_RMAP_BLOCK(mp)); 260 agf->agf_levels[XFS_BTNUM_RMAPi] = cpu_to_be32(1); 261 agf->agf_rmap_blocks = cpu_to_be32(1); 262 } 263 264 agf->agf_flfirst = cpu_to_be32(1); 265 agf->agf_fllast = 0; 266 agf->agf_flcount = 0; 267 tmpsize = id->agsize - mp->m_ag_prealloc_blocks; 268 agf->agf_freeblks = cpu_to_be32(tmpsize); 269 agf->agf_longest = cpu_to_be32(tmpsize); 270 if (xfs_sb_version_hascrc(&mp->m_sb)) 271 uuid_copy(&agf->agf_uuid, &mp->m_sb.sb_meta_uuid); 272 if (xfs_sb_version_hasreflink(&mp->m_sb)) { 273 agf->agf_refcount_root = cpu_to_be32( 274 xfs_refc_block(mp)); 275 agf->agf_refcount_level = cpu_to_be32(1); 276 agf->agf_refcount_blocks = cpu_to_be32(1); 277 } 278 279 if (is_log_ag(mp, id)) { 280 int64_t logblocks = mp->m_sb.sb_logblocks; 281 282 be32_add_cpu(&agf->agf_freeblks, -logblocks); 283 agf->agf_longest = cpu_to_be32(id->agsize - 284 XFS_FSB_TO_AGBNO(mp, mp->m_sb.sb_logstart) - logblocks); 285 } 286} 287 288static void 289xfs_agflblock_init( 290 struct xfs_mount *mp, 291 struct xfs_buf *bp, 292 struct aghdr_init_data *id) 293{ 294 struct xfs_agfl *agfl = XFS_BUF_TO_AGFL(bp); 295 __be32 *agfl_bno; 296 int bucket; 297 298 if (xfs_sb_version_hascrc(&mp->m_sb)) { 299 agfl->agfl_magicnum = cpu_to_be32(XFS_AGFL_MAGIC); 300 agfl->agfl_seqno = cpu_to_be32(id->agno); 301 uuid_copy(&agfl->agfl_uuid, &mp->m_sb.sb_meta_uuid); 302 } 303 304 agfl_bno = xfs_buf_to_agfl_bno(bp); 305 for (bucket = 0; bucket < xfs_agfl_size(mp); bucket++) 306 agfl_bno[bucket] = cpu_to_be32(NULLAGBLOCK); 307} 308 309static void 310xfs_agiblock_init( 311 struct xfs_mount *mp, 312 struct xfs_buf *bp, 313 struct aghdr_init_data *id) 314{ 315 struct xfs_agi *agi = bp->b_addr; 316 int bucket; 317 318 agi->agi_magicnum = cpu_to_be32(XFS_AGI_MAGIC); 319 agi->agi_versionnum = cpu_to_be32(XFS_AGI_VERSION); 320 agi->agi_seqno = cpu_to_be32(id->agno); 321 agi->agi_length = cpu_to_be32(id->agsize); 322 agi->agi_count = 0; 323 agi->agi_root = cpu_to_be32(XFS_IBT_BLOCK(mp)); 324 agi->agi_level = cpu_to_be32(1); 325 agi->agi_freecount = 0; 326 agi->agi_newino = cpu_to_be32(NULLAGINO); 327 agi->agi_dirino = cpu_to_be32(NULLAGINO); 328 if (xfs_sb_version_hascrc(&mp->m_sb)) 329 uuid_copy(&agi->agi_uuid, &mp->m_sb.sb_meta_uuid); 330 if (xfs_sb_version_hasfinobt(&mp->m_sb)) { 331 agi->agi_free_root = cpu_to_be32(XFS_FIBT_BLOCK(mp)); 332 agi->agi_free_level = cpu_to_be32(1); 333 } 334 for (bucket = 0; bucket < XFS_AGI_UNLINKED_BUCKETS; bucket++) 335 agi->agi_unlinked[bucket] = cpu_to_be32(NULLAGINO); 336 if (xfs_sb_version_hasinobtcounts(&mp->m_sb)) { 337 agi->agi_iblocks = cpu_to_be32(1); 338 if (xfs_sb_version_hasfinobt(&mp->m_sb)) 339 agi->agi_fblocks = cpu_to_be32(1); 340 } 341} 342 343typedef void (*aghdr_init_work_f)(struct xfs_mount *mp, struct xfs_buf *bp, 344 struct aghdr_init_data *id); 345static int 346xfs_ag_init_hdr( 347 struct xfs_mount *mp, 348 struct aghdr_init_data *id, 349 aghdr_init_work_f work, 350 const struct xfs_buf_ops *ops) 351{ 352 struct xfs_buf *bp; 353 int error; 354 355 error = xfs_get_aghdr_buf(mp, id->daddr, id->numblks, &bp, ops); 356 if (error) 357 return error; 358 359 (*work)(mp, bp, id); 360 361 xfs_buf_delwri_queue(bp, &id->buffer_list); 362 xfs_buf_relse(bp); 363 return 0; 364} 365 366struct xfs_aghdr_grow_data { 367 xfs_daddr_t daddr; 368 size_t numblks; 369 const struct xfs_buf_ops *ops; 370 aghdr_init_work_f work; 371 xfs_btnum_t type; 372 bool need_init; 373}; 374 375/* 376 * Prepare new AG headers to be written to disk. We use uncached buffers here, 377 * as it is assumed these new AG headers are currently beyond the currently 378 * valid filesystem address space. Using cached buffers would trip over EOFS 379 * corruption detection alogrithms in the buffer cache lookup routines. 380 * 381 * This is a non-transactional function, but the prepared buffers are added to a 382 * delayed write buffer list supplied by the caller so they can submit them to 383 * disk and wait on them as required. 384 */ 385int 386xfs_ag_init_headers( 387 struct xfs_mount *mp, 388 struct aghdr_init_data *id) 389 390{ 391 struct xfs_aghdr_grow_data aghdr_data[] = { 392 { /* SB */ 393 .daddr = XFS_AG_DADDR(mp, id->agno, XFS_SB_DADDR), 394 .numblks = XFS_FSS_TO_BB(mp, 1), 395 .ops = &xfs_sb_buf_ops, 396 .work = &xfs_sbblock_init, 397 .need_init = true 398 }, 399 { /* AGF */ 400 .daddr = XFS_AG_DADDR(mp, id->agno, XFS_AGF_DADDR(mp)), 401 .numblks = XFS_FSS_TO_BB(mp, 1), 402 .ops = &xfs_agf_buf_ops, 403 .work = &xfs_agfblock_init, 404 .need_init = true 405 }, 406 { /* AGFL */ 407 .daddr = XFS_AG_DADDR(mp, id->agno, XFS_AGFL_DADDR(mp)), 408 .numblks = XFS_FSS_TO_BB(mp, 1), 409 .ops = &xfs_agfl_buf_ops, 410 .work = &xfs_agflblock_init, 411 .need_init = true 412 }, 413 { /* AGI */ 414 .daddr = XFS_AG_DADDR(mp, id->agno, XFS_AGI_DADDR(mp)), 415 .numblks = XFS_FSS_TO_BB(mp, 1), 416 .ops = &xfs_agi_buf_ops, 417 .work = &xfs_agiblock_init, 418 .need_init = true 419 }, 420 { /* BNO root block */ 421 .daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_BNO_BLOCK(mp)), 422 .numblks = BTOBB(mp->m_sb.sb_blocksize), 423 .ops = &xfs_bnobt_buf_ops, 424 .work = &xfs_bnoroot_init, 425 .need_init = true 426 }, 427 { /* CNT root block */ 428 .daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_CNT_BLOCK(mp)), 429 .numblks = BTOBB(mp->m_sb.sb_blocksize), 430 .ops = &xfs_cntbt_buf_ops, 431 .work = &xfs_cntroot_init, 432 .need_init = true 433 }, 434 { /* INO root block */ 435 .daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_IBT_BLOCK(mp)), 436 .numblks = BTOBB(mp->m_sb.sb_blocksize), 437 .ops = &xfs_inobt_buf_ops, 438 .work = &xfs_btroot_init, 439 .type = XFS_BTNUM_INO, 440 .need_init = true 441 }, 442 { /* FINO root block */ 443 .daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_FIBT_BLOCK(mp)), 444 .numblks = BTOBB(mp->m_sb.sb_blocksize), 445 .ops = &xfs_finobt_buf_ops, 446 .work = &xfs_btroot_init, 447 .type = XFS_BTNUM_FINO, 448 .need_init = xfs_sb_version_hasfinobt(&mp->m_sb) 449 }, 450 { /* RMAP root block */ 451 .daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_RMAP_BLOCK(mp)), 452 .numblks = BTOBB(mp->m_sb.sb_blocksize), 453 .ops = &xfs_rmapbt_buf_ops, 454 .work = &xfs_rmaproot_init, 455 .need_init = xfs_sb_version_hasrmapbt(&mp->m_sb) 456 }, 457 { /* REFC root block */ 458 .daddr = XFS_AGB_TO_DADDR(mp, id->agno, xfs_refc_block(mp)), 459 .numblks = BTOBB(mp->m_sb.sb_blocksize), 460 .ops = &xfs_refcountbt_buf_ops, 461 .work = &xfs_btroot_init, 462 .type = XFS_BTNUM_REFC, 463 .need_init = xfs_sb_version_hasreflink(&mp->m_sb) 464 }, 465 { /* NULL terminating block */ 466 .daddr = XFS_BUF_DADDR_NULL, 467 } 468 }; 469 struct xfs_aghdr_grow_data *dp; 470 int error = 0; 471 472 /* Account for AG free space in new AG */ 473 id->nfree += id->agsize - mp->m_ag_prealloc_blocks; 474 for (dp = &aghdr_data[0]; dp->daddr != XFS_BUF_DADDR_NULL; dp++) { 475 if (!dp->need_init) 476 continue; 477 478 id->daddr = dp->daddr; 479 id->numblks = dp->numblks; 480 id->type = dp->type; 481 error = xfs_ag_init_hdr(mp, id, dp->work, dp->ops); 482 if (error) 483 break; 484 } 485 return error; 486} 487 488/* 489 * Extent the AG indicated by the @id by the length passed in 490 */ 491int 492xfs_ag_extend_space( 493 struct xfs_mount *mp, 494 struct xfs_trans *tp, 495 struct aghdr_init_data *id, 496 xfs_extlen_t len) 497{ 498 struct xfs_buf *bp; 499 struct xfs_agi *agi; 500 struct xfs_agf *agf; 501 int error; 502 503 /* 504 * Change the agi length. 505 */ 506 error = xfs_ialloc_read_agi(mp, tp, id->agno, &bp); 507 if (error) 508 return error; 509 510 agi = bp->b_addr; 511 be32_add_cpu(&agi->agi_length, len); 512 ASSERT(id->agno == mp->m_sb.sb_agcount - 1 || 513 be32_to_cpu(agi->agi_length) == mp->m_sb.sb_agblocks); 514 xfs_ialloc_log_agi(tp, bp, XFS_AGI_LENGTH); 515 516 /* 517 * Change agf length. 518 */ 519 error = xfs_alloc_read_agf(mp, tp, id->agno, 0, &bp); 520 if (error) 521 return error; 522 523 agf = bp->b_addr; 524 be32_add_cpu(&agf->agf_length, len); 525 ASSERT(agf->agf_length == agi->agi_length); 526 xfs_alloc_log_agf(tp, bp, XFS_AGF_LENGTH); 527 528 /* 529 * Free the new space. 530 * 531 * XFS_RMAP_OINFO_SKIP_UPDATE is used here to tell the rmap btree that 532 * this doesn't actually exist in the rmap btree. 533 */ 534 error = xfs_rmap_free(tp, bp, id->agno, 535 be32_to_cpu(agf->agf_length) - len, 536 len, &XFS_RMAP_OINFO_SKIP_UPDATE); 537 if (error) 538 return error; 539 540 return xfs_free_extent(tp, XFS_AGB_TO_FSB(mp, id->agno, 541 be32_to_cpu(agf->agf_length) - len), 542 len, &XFS_RMAP_OINFO_SKIP_UPDATE, 543 XFS_AG_RESV_NONE); 544} 545 546/* Retrieve AG geometry. */ 547int 548xfs_ag_get_geometry( 549 struct xfs_mount *mp, 550 xfs_agnumber_t agno, 551 struct xfs_ag_geometry *ageo) 552{ 553 struct xfs_buf *agi_bp; 554 struct xfs_buf *agf_bp; 555 struct xfs_agi *agi; 556 struct xfs_agf *agf; 557 struct xfs_perag *pag; 558 unsigned int freeblks; 559 int error; 560 561 if (agno >= mp->m_sb.sb_agcount) 562 return -EINVAL; 563 564 /* Lock the AG headers. */ 565 error = xfs_ialloc_read_agi(mp, NULL, agno, &agi_bp); 566 if (error) 567 return error; 568 error = xfs_alloc_read_agf(mp, NULL, agno, 0, &agf_bp); 569 if (error) 570 goto out_agi; 571 572 pag = agi_bp->b_pag; 573 574 /* Fill out form. */ 575 memset(ageo, 0, sizeof(*ageo)); 576 ageo->ag_number = agno; 577 578 agi = agi_bp->b_addr; 579 ageo->ag_icount = be32_to_cpu(agi->agi_count); 580 ageo->ag_ifree = be32_to_cpu(agi->agi_freecount); 581 582 agf = agf_bp->b_addr; 583 ageo->ag_length = be32_to_cpu(agf->agf_length); 584 freeblks = pag->pagf_freeblks + 585 pag->pagf_flcount + 586 pag->pagf_btreeblks - 587 xfs_ag_resv_needed(pag, XFS_AG_RESV_NONE); 588 ageo->ag_freeblks = freeblks; 589 xfs_ag_geom_health(pag, ageo); 590 591 /* Release resources. */ 592 xfs_buf_relse(agf_bp); 593out_agi: 594 xfs_buf_relse(agi_bp); 595 return error; 596} 597