fs/xfs/libxfs/xfs_refcount_btree.c

   1 // SPDX-License-Identifier: GPL-2.0+
   2 /*
   3  * Copyright (C) 2016 Oracle.  All Rights Reserved.
   4  * Author: Darrick J. Wong <darrick.wong@oracle.com>
   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_mount.h"
  13 #include "xfs_btree.h"
  14 #include "xfs_btree_staging.h"
  15 #include "xfs_refcount_btree.h"
  16 #include "xfs_refcount.h"
  17 #include "xfs_alloc.h"
  18 #include "xfs_error.h"
  19 #include "xfs_health.h"
  20 #include "xfs_trace.h"
  21 #include "xfs_trans.h"
  22 #include "xfs_bit.h"
  23 #include "xfs_rmap.h"
  24 #include "xfs_ag.h"
  25
  26 static struct kmem_cache        *xfs_refcountbt_cur_cache;
  27
  28 static struct xfs_btree_cur *
  29 xfs_refcountbt_dup_cursor(
  30         struct xfs_btree_cur    *cur)
  31 {
  32         return xfs_refcountbt_init_cursor(cur->bc_mp, cur->bc_tp,
  33                         cur->bc_ag.agbp, cur->bc_ag.pag);
  34 }
  35
  36 STATIC void
  37 xfs_refcountbt_set_root(
  38         struct xfs_btree_cur            *cur,
  39         const union xfs_btree_ptr       *ptr,
  40         int                             inc)
  41 {
  42         struct xfs_buf          *agbp = cur->bc_ag.agbp;
  43         struct xfs_agf          *agf = agbp->b_addr;
  44         struct xfs_perag        *pag = agbp->b_pag;
  45
  46         ASSERT(ptr->s != 0);
  47
  48         agf->agf_refcount_root = ptr->s;
  49         be32_add_cpu(&agf->agf_refcount_level, inc);
  50         pag->pagf_refcount_level += inc;
  51
  52         xfs_alloc_log_agf(cur->bc_tp, agbp,
  53                         XFS_AGF_REFCOUNT_ROOT | XFS_AGF_REFCOUNT_LEVEL);
  54 }
  55
  56 STATIC int
  57 xfs_refcountbt_alloc_block(
  58         struct xfs_btree_cur            *cur,
  59         const union xfs_btree_ptr       *start,
  60         union xfs_btree_ptr             *new,
  61         int                             *stat)
  62 {
  63         struct xfs_buf          *agbp = cur->bc_ag.agbp;
  64         struct xfs_agf          *agf = agbp->b_addr;
  65         struct xfs_alloc_arg    args;           /* block allocation args */
  66         int                     error;          /* error return value */
  67
  68         memset(&args, 0, sizeof(args));
  69         args.tp = cur->bc_tp;
  70         args.mp = cur->bc_mp;
  71         args.pag = cur->bc_ag.pag;
  72         args.oinfo = XFS_RMAP_OINFO_REFC;
  73         args.minlen = args.maxlen = args.prod = 1;
  74         args.resv = XFS_AG_RESV_METADATA;
  75
  76         error = xfs_alloc_vextent_near_bno(&args,
  77                         XFS_AGB_TO_FSB(args.mp, args.pag->pag_agno,
  78                                         xfs_refc_block(args.mp)));
  79         if (error)
  80                 goto out_error;
  81         if (args.fsbno == NULLFSBLOCK) {
  82                 *stat = 0;
  83                 return 0;
  84         }
  85         ASSERT(args.agno == cur->bc_ag.pag->pag_agno);
  86         ASSERT(args.len == 1);
  87
  88         new->s = cpu_to_be32(args.agbno);
  89         be32_add_cpu(&agf->agf_refcount_blocks, 1);
  90         xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_REFCOUNT_BLOCKS);
  91
  92         *stat = 1;
  93         return 0;
  94
  95 out_error:
  96         return error;
  97 }
  98
  99 STATIC int
 100 xfs_refcountbt_free_block(
 101         struct xfs_btree_cur    *cur,
 102         struct xfs_buf          *bp)
 103 {
 104         struct xfs_mount        *mp = cur->bc_mp;
 105         struct xfs_buf          *agbp = cur->bc_ag.agbp;
 106         struct xfs_agf          *agf = agbp->b_addr;
 107         xfs_fsblock_t           fsbno = XFS_DADDR_TO_FSB(mp, xfs_buf_daddr(bp));
 108
 109         be32_add_cpu(&agf->agf_refcount_blocks, -1);
 110         xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_REFCOUNT_BLOCKS);
 111         return xfs_free_extent_later(cur->bc_tp, fsbno, 1,
 112                         &XFS_RMAP_OINFO_REFC, XFS_AG_RESV_METADATA, false);
 113 }
 114
 115 STATIC int
 116 xfs_refcountbt_get_minrecs(
 117         struct xfs_btree_cur    *cur,
 118         int                     level)
 119 {
 120         return cur->bc_mp->m_refc_mnr[level != 0];
 121 }
 122
 123 STATIC int
 124 xfs_refcountbt_get_maxrecs(
 125         struct xfs_btree_cur    *cur,
 126         int                     level)
 127 {
 128         return cur->bc_mp->m_refc_mxr[level != 0];
 129 }
 130
 131 STATIC void
 132 xfs_refcountbt_init_key_from_rec(
 133         union xfs_btree_key             *key,
 134         const union xfs_btree_rec       *rec)
 135 {
 136         key->refc.rc_startblock = rec->refc.rc_startblock;
 137 }
 138
 139 STATIC void
 140 xfs_refcountbt_init_high_key_from_rec(
 141         union xfs_btree_key             *key,
 142         const union xfs_btree_rec       *rec)
 143 {
 144         __u32                           x;
 145
 146         x = be32_to_cpu(rec->refc.rc_startblock);
 147         x += be32_to_cpu(rec->refc.rc_blockcount) - 1;
 148         key->refc.rc_startblock = cpu_to_be32(x);
 149 }
 150
 151 STATIC void
 152 xfs_refcountbt_init_rec_from_cur(
 153         struct xfs_btree_cur    *cur,
 154         union xfs_btree_rec     *rec)
 155 {
 156         const struct xfs_refcount_irec *irec = &cur->bc_rec.rc;
 157         uint32_t                start;
 158
 159         start = xfs_refcount_encode_startblock(irec->rc_startblock,
 160                         irec->rc_domain);
 161         rec->refc.rc_startblock = cpu_to_be32(start);
 162         rec->refc.rc_blockcount = cpu_to_be32(cur->bc_rec.rc.rc_blockcount);
 163         rec->refc.rc_refcount = cpu_to_be32(cur->bc_rec.rc.rc_refcount);
 164 }
 165
 166 STATIC void
 167 xfs_refcountbt_init_ptr_from_cur(
 168         struct xfs_btree_cur    *cur,
 169         union xfs_btree_ptr     *ptr)
 170 {
 171         struct xfs_agf          *agf = cur->bc_ag.agbp->b_addr;
 172
 173         ASSERT(cur->bc_ag.pag->pag_agno == be32_to_cpu(agf->agf_seqno));
 174
 175         ptr->s = agf->agf_refcount_root;
 176 }
 177
 178 STATIC int64_t
 179 xfs_refcountbt_key_diff(
 180         struct xfs_btree_cur            *cur,
 181         const union xfs_btree_key       *key)
 182 {
 183         const struct xfs_refcount_key   *kp = &key->refc;
 184         const struct xfs_refcount_irec  *irec = &cur->bc_rec.rc;
 185         uint32_t                        start;
 186
 187         start = xfs_refcount_encode_startblock(irec->rc_startblock,
 188                         irec->rc_domain);
 189         return (int64_t)be32_to_cpu(kp->rc_startblock) - start;
 190 }
 191
 192 STATIC int64_t
 193 xfs_refcountbt_diff_two_keys(
 194         struct xfs_btree_cur            *cur,
 195         const union xfs_btree_key       *k1,
 196         const union xfs_btree_key       *k2,
 197         const union xfs_btree_key       *mask)
 198 {
 199         ASSERT(!mask || mask->refc.rc_startblock);
 200
 201         return (int64_t)be32_to_cpu(k1->refc.rc_startblock) -
 202                         be32_to_cpu(k2->refc.rc_startblock);
 203 }
 204
 205 STATIC xfs_failaddr_t
 206 xfs_refcountbt_verify(
 207         struct xfs_buf          *bp)
 208 {
 209         struct xfs_mount        *mp = bp->b_mount;
 210         struct xfs_btree_block  *block = XFS_BUF_TO_BLOCK(bp);
 211         struct xfs_perag        *pag = bp->b_pag;
 212         xfs_failaddr_t          fa;
 213         unsigned int            level;
 214
 215         if (!xfs_verify_magic(bp, block->bb_magic))
 216                 return __this_address;
 217
 218         if (!xfs_has_reflink(mp))
 219                 return __this_address;
 220         fa = xfs_btree_agblock_v5hdr_verify(bp);
 221         if (fa)
 222                 return fa;
 223
 224         level = be16_to_cpu(block->bb_level);
 225         if (pag && xfs_perag_initialised_agf(pag)) {
 226                 unsigned int    maxlevel = pag->pagf_refcount_level;
 227
 228 #ifdef CONFIG_XFS_ONLINE_REPAIR
 229                 /*
 230                  * Online repair could be rewriting the refcount btree, so
 231                  * we'll validate against the larger of either tree while this
 232                  * is going on.
 233                  */
 234                 maxlevel = max_t(unsigned int, maxlevel,
 235                                 pag->pagf_repair_refcount_level);
 236 #endif
 237                 if (level >= maxlevel)
 238                         return __this_address;
 239         } else if (level >= mp->m_refc_maxlevels)
 240                 return __this_address;
 241
 242         return xfs_btree_agblock_verify(bp, mp->m_refc_mxr[level != 0]);
 243 }
 244
 245 STATIC void
 246 xfs_refcountbt_read_verify(
 247         struct xfs_buf  *bp)
 248 {
 249         xfs_failaddr_t  fa;
 250
 251         if (!xfs_btree_agblock_verify_crc(bp))
 252                 xfs_verifier_error(bp, -EFSBADCRC, __this_address);
 253         else {
 254                 fa = xfs_refcountbt_verify(bp);
 255                 if (fa)
 256                         xfs_verifier_error(bp, -EFSCORRUPTED, fa);
 257         }
 258
 259         if (bp->b_error)
 260                 trace_xfs_btree_corrupt(bp, _RET_IP_);
 261 }
 262
 263 STATIC void
 264 xfs_refcountbt_write_verify(
 265         struct xfs_buf  *bp)
 266 {
 267         xfs_failaddr_t  fa;
 268
 269         fa = xfs_refcountbt_verify(bp);
 270         if (fa) {
 271                 trace_xfs_btree_corrupt(bp, _RET_IP_);
 272                 xfs_verifier_error(bp, -EFSCORRUPTED, fa);
 273                 return;
 274         }
 275         xfs_btree_agblock_calc_crc(bp);
 276
 277 }
 278
 279 const struct xfs_buf_ops xfs_refcountbt_buf_ops = {
 280         .name                   = "xfs_refcountbt",
 281         .magic                  = { 0, cpu_to_be32(XFS_REFC_CRC_MAGIC) },
 282         .verify_read            = xfs_refcountbt_read_verify,
 283         .verify_write           = xfs_refcountbt_write_verify,
 284         .verify_struct          = xfs_refcountbt_verify,
 285 };
 286
 287 STATIC int
 288 xfs_refcountbt_keys_inorder(
 289         struct xfs_btree_cur            *cur,
 290         const union xfs_btree_key       *k1,
 291         const union xfs_btree_key       *k2)
 292 {
 293         return be32_to_cpu(k1->refc.rc_startblock) <
 294                be32_to_cpu(k2->refc.rc_startblock);
 295 }
 296
 297 STATIC int
 298 xfs_refcountbt_recs_inorder(
 299         struct xfs_btree_cur            *cur,
 300         const union xfs_btree_rec       *r1,
 301         const union xfs_btree_rec       *r2)
 302 {
 303         return  be32_to_cpu(r1->refc.rc_startblock) +
 304                 be32_to_cpu(r1->refc.rc_blockcount) <=
 305                 be32_to_cpu(r2->refc.rc_startblock);
 306 }
 307
 308 STATIC enum xbtree_key_contig
 309 xfs_refcountbt_keys_contiguous(
 310         struct xfs_btree_cur            *cur,
 311         const union xfs_btree_key       *key1,
 312         const union xfs_btree_key       *key2,
 313         const union xfs_btree_key       *mask)
 314 {
 315         ASSERT(!mask || mask->refc.rc_startblock);
 316
 317         return xbtree_key_contig(be32_to_cpu(key1->refc.rc_startblock),
 318                                  be32_to_cpu(key2->refc.rc_startblock));
 319 }
 320
 321 const struct xfs_btree_ops xfs_refcountbt_ops = {
 322         .name                   = "refcount",
 323         .type                   = XFS_BTREE_TYPE_AG,
 324
 325         .rec_len                = sizeof(struct xfs_refcount_rec),
 326         .key_len                = sizeof(struct xfs_refcount_key),
 327         .ptr_len                = XFS_BTREE_SHORT_PTR_LEN,
 328
 329         .lru_refs               = XFS_REFC_BTREE_REF,
 330         .statoff                = XFS_STATS_CALC_INDEX(xs_refcbt_2),
 331         .sick_mask              = XFS_SICK_AG_REFCNTBT,
 332
 333         .dup_cursor             = xfs_refcountbt_dup_cursor,
 334         .set_root               = xfs_refcountbt_set_root,
 335         .alloc_block            = xfs_refcountbt_alloc_block,
 336         .free_block             = xfs_refcountbt_free_block,
 337         .get_minrecs            = xfs_refcountbt_get_minrecs,
 338         .get_maxrecs            = xfs_refcountbt_get_maxrecs,
 339         .init_key_from_rec      = xfs_refcountbt_init_key_from_rec,
 340         .init_high_key_from_rec = xfs_refcountbt_init_high_key_from_rec,
 341         .init_rec_from_cur      = xfs_refcountbt_init_rec_from_cur,
 342         .init_ptr_from_cur      = xfs_refcountbt_init_ptr_from_cur,
 343         .key_diff               = xfs_refcountbt_key_diff,
 344         .buf_ops                = &xfs_refcountbt_buf_ops,
 345         .diff_two_keys          = xfs_refcountbt_diff_two_keys,
 346         .keys_inorder           = xfs_refcountbt_keys_inorder,
 347         .recs_inorder           = xfs_refcountbt_recs_inorder,
 348         .keys_contiguous        = xfs_refcountbt_keys_contiguous,
 349 };
 350
 351 /*
 352  * Create a new refcount btree cursor.
 353  *
 354  * For staging cursors tp and agbp are NULL.
 355  */
 356 struct xfs_btree_cur *
 357 xfs_refcountbt_init_cursor(
 358         struct xfs_mount        *mp,
 359         struct xfs_trans        *tp,
 360         struct xfs_buf          *agbp,
 361         struct xfs_perag        *pag)
 362 {
 363         struct xfs_btree_cur    *cur;
 364
 365         ASSERT(pag->pag_agno < mp->m_sb.sb_agcount);
 366
 367         cur = xfs_btree_alloc_cursor(mp, tp, &xfs_refcountbt_ops,
 368                         mp->m_refc_maxlevels, xfs_refcountbt_cur_cache);
 369         cur->bc_ag.pag = xfs_perag_hold(pag);
 370         cur->bc_refc.nr_ops = 0;
 371         cur->bc_refc.shape_changes = 0;
 372         cur->bc_ag.agbp = agbp;
 373         if (agbp) {
 374                 struct xfs_agf          *agf = agbp->b_addr;
 375
 376                 cur->bc_nlevels = be32_to_cpu(agf->agf_refcount_level);
 377         }
 378         return cur;
 379 }
 380
 381 /*
 382  * Swap in the new btree root.  Once we pass this point the newly rebuilt btree
 383  * is in place and we have to kill off all the old btree blocks.
 384  */
 385 void
 386 xfs_refcountbt_commit_staged_btree(
 387         struct xfs_btree_cur    *cur,
 388         struct xfs_trans        *tp,
 389         struct xfs_buf          *agbp)
 390 {
 391         struct xfs_agf          *agf = agbp->b_addr;
 392         struct xbtree_afakeroot *afake = cur->bc_ag.afake;
 393
 394         ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
 395
 396         agf->agf_refcount_root = cpu_to_be32(afake->af_root);
 397         agf->agf_refcount_level = cpu_to_be32(afake->af_levels);
 398         agf->agf_refcount_blocks = cpu_to_be32(afake->af_blocks);
 399         xfs_alloc_log_agf(tp, agbp, XFS_AGF_REFCOUNT_BLOCKS |
 400                                     XFS_AGF_REFCOUNT_ROOT |
 401                                     XFS_AGF_REFCOUNT_LEVEL);
 402         xfs_btree_commit_afakeroot(cur, tp, agbp);
 403 }
 404
 405 /* Calculate number of records in a refcount btree block. */
 406 static inline unsigned int
 407 xfs_refcountbt_block_maxrecs(
 408         unsigned int            blocklen,
 409         bool                    leaf)
 410 {
 411         if (leaf)
 412                 return blocklen / sizeof(struct xfs_refcount_rec);
 413         return blocklen / (sizeof(struct xfs_refcount_key) +
 414                            sizeof(xfs_refcount_ptr_t));
 415 }
 416
 417 /*
 418  * Calculate the number of records in a refcount btree block.
 419  */
 420 int
 421 xfs_refcountbt_maxrecs(
 422         int                     blocklen,
 423         bool                    leaf)
 424 {
 425         blocklen -= XFS_REFCOUNT_BLOCK_LEN;
 426         return xfs_refcountbt_block_maxrecs(blocklen, leaf);
 427 }
 428
 429 /* Compute the max possible height of the maximally sized refcount btree. */
 430 unsigned int
 431 xfs_refcountbt_maxlevels_ondisk(void)
 432 {
 433         unsigned int            minrecs[2];
 434         unsigned int            blocklen;
 435
 436         blocklen = XFS_MIN_CRC_BLOCKSIZE - XFS_BTREE_SBLOCK_CRC_LEN;
 437
 438         minrecs[0] = xfs_refcountbt_block_maxrecs(blocklen, true) / 2;
 439         minrecs[1] = xfs_refcountbt_block_maxrecs(blocklen, false) / 2;
 440
 441         return xfs_btree_compute_maxlevels(minrecs, XFS_MAX_CRC_AG_BLOCKS);
 442 }
 443
 444 /* Compute the maximum height of a refcount btree. */
 445 void
 446 xfs_refcountbt_compute_maxlevels(
 447         struct xfs_mount                *mp)
 448 {
 449         if (!xfs_has_reflink(mp)) {
 450                 mp->m_refc_maxlevels = 0;
 451                 return;
 452         }
 453
 454         mp->m_refc_maxlevels = xfs_btree_compute_maxlevels(
 455                         mp->m_refc_mnr, mp->m_sb.sb_agblocks);
 456         ASSERT(mp->m_refc_maxlevels <= xfs_refcountbt_maxlevels_ondisk());
 457 }
 458
 459 /* Calculate the refcount btree size for some records. */
 460 xfs_extlen_t
 461 xfs_refcountbt_calc_size(
 462         struct xfs_mount        *mp,
 463         unsigned long long      len)
 464 {
 465         return xfs_btree_calc_size(mp->m_refc_mnr, len);
 466 }
 467
 468 /*
 469  * Calculate the maximum refcount btree size.
 470  */
 471 xfs_extlen_t
 472 xfs_refcountbt_max_size(
 473         struct xfs_mount        *mp,
 474         xfs_agblock_t           agblocks)
 475 {
 476         /* Bail out if we're uninitialized, which can happen in mkfs. */
 477         if (mp->m_refc_mxr[0] == 0)
 478                 return 0;
 479
 480         return xfs_refcountbt_calc_size(mp, agblocks);
 481 }
 482
 483 /*
 484  * Figure out how many blocks to reserve and how many are used by this btree.
 485  */
 486 int
 487 xfs_refcountbt_calc_reserves(
 488         struct xfs_mount        *mp,
 489         struct xfs_trans        *tp,
 490         struct xfs_perag        *pag,
 491         xfs_extlen_t            *ask,
 492         xfs_extlen_t            *used)
 493 {
 494         struct xfs_buf          *agbp;
 495         struct xfs_agf          *agf;
 496         xfs_agblock_t           agblocks;
 497         xfs_extlen_t            tree_len;
 498         int                     error;
 499
 500         if (!xfs_has_reflink(mp))
 501                 return 0;
 502
 503         error = xfs_alloc_read_agf(pag, tp, 0, &agbp);
 504         if (error)
 505                 return error;
 506
 507         agf = agbp->b_addr;
 508         agblocks = be32_to_cpu(agf->agf_length);
 509         tree_len = be32_to_cpu(agf->agf_refcount_blocks);
 510         xfs_trans_brelse(tp, agbp);
 511
 512         /*
 513          * The log is permanently allocated, so the space it occupies will
 514          * never be available for the kinds of things that would require btree
 515          * expansion.  We therefore can pretend the space isn't there.
 516          */
 517         if (xfs_ag_contains_log(mp, pag->pag_agno))
 518                 agblocks -= mp->m_sb.sb_logblocks;
 519
 520         *ask += xfs_refcountbt_max_size(mp, agblocks);
 521         *used += tree_len;
 522
 523         return error;
 524 }
 525
 526 int __init
 527 xfs_refcountbt_init_cur_cache(void)
 528 {
 529         xfs_refcountbt_cur_cache = kmem_cache_create("xfs_refcbt_cur",
 530                         xfs_btree_cur_sizeof(xfs_refcountbt_maxlevels_ondisk()),
 531                         0, 0, NULL);
 532
 533         if (!xfs_refcountbt_cur_cache)
 534                 return -ENOMEM;
 535         return 0;
 536 }
 537
 538 void
 539 xfs_refcountbt_destroy_cur_cache(void)
 540 {
 541         kmem_cache_destroy(xfs_refcountbt_cur_cache);
 542         xfs_refcountbt_cur_cache = NULL;
 543 }