fs/bcachefs/fsck.c

   1 // SPDX-License-Identifier: GPL-2.0
   2
   3 #include "bcachefs.h"
   4 #include "bkey_buf.h"
   5 #include "btree_cache.h"
   6 #include "btree_update.h"
   7 #include "buckets.h"
   8 #include "darray.h"
   9 #include "dirent.h"
  10 #include "error.h"
  11 #include "fs-common.h"
  12 #include "fsck.h"
  13 #include "inode.h"
  14 #include "keylist.h"
  15 #include "recovery.h"
  16 #include "snapshot.h"
  17 #include "super.h"
  18 #include "xattr.h"
  19
  20 #include <linux/bsearch.h>
  21 #include <linux/dcache.h> /* struct qstr */
  22
  23 #define QSTR(n) { { { .len = strlen(n) } }, .name = n }
  24
  25 /*
  26  * XXX: this is handling transaction restarts without returning
  27  * -BCH_ERR_transaction_restart_nested, this is not how we do things anymore:
  28  */
  29 static s64 bch2_count_inode_sectors(struct btree_trans *trans, u64 inum,
  30                                     u32 snapshot)
  31 {
  32         struct btree_iter iter;
  33         struct bkey_s_c k;
  34         u64 sectors = 0;
  35         int ret;
  36
  37         for_each_btree_key_upto(trans, iter, BTREE_ID_extents,
  38                                 SPOS(inum, 0, snapshot),
  39                                 POS(inum, U64_MAX),
  40                                 0, k, ret)
  41                 if (bkey_extent_is_allocation(k.k))
  42                         sectors += k.k->size;
  43
  44         bch2_trans_iter_exit(trans, &iter);
  45
  46         return ret ?: sectors;
  47 }
  48
  49 static s64 bch2_count_subdirs(struct btree_trans *trans, u64 inum,
  50                                     u32 snapshot)
  51 {
  52         struct btree_iter iter;
  53         struct bkey_s_c k;
  54         struct bkey_s_c_dirent d;
  55         u64 subdirs = 0;
  56         int ret;
  57
  58         for_each_btree_key_upto(trans, iter, BTREE_ID_dirents,
  59                                 SPOS(inum, 0, snapshot),
  60                                 POS(inum, U64_MAX),
  61                                 0, k, ret) {
  62                 if (k.k->type != KEY_TYPE_dirent)
  63                         continue;
  64
  65                 d = bkey_s_c_to_dirent(k);
  66                 if (d.v->d_type == DT_DIR)
  67                         subdirs++;
  68         }
  69         bch2_trans_iter_exit(trans, &iter);
  70
  71         return ret ?: subdirs;
  72 }
  73
  74 static int __snapshot_lookup_subvol(struct btree_trans *trans, u32 snapshot,
  75                                     u32 *subvol)
  76 {
  77         struct bch_snapshot s;
  78         int ret = bch2_bkey_get_val_typed(trans, BTREE_ID_snapshots,
  79                                           POS(0, snapshot), 0,
  80                                           snapshot, &s);
  81         if (!ret)
  82                 *subvol = le32_to_cpu(s.subvol);
  83         else if (bch2_err_matches(ret, ENOENT))
  84                 bch_err(trans->c, "snapshot %u not found", snapshot);
  85         return ret;
  86
  87 }
  88
  89 static int __subvol_lookup(struct btree_trans *trans, u32 subvol,
  90                            u32 *snapshot, u64 *inum)
  91 {
  92         struct bch_subvolume s;
  93         int ret;
  94
  95         ret = bch2_subvolume_get(trans, subvol, false, 0, &s);
  96
  97         *snapshot = le32_to_cpu(s.snapshot);
  98         *inum = le64_to_cpu(s.inode);
  99         return ret;
 100 }
 101
 102 static int subvol_lookup(struct btree_trans *trans, u32 subvol,
 103                          u32 *snapshot, u64 *inum)
 104 {
 105         return lockrestart_do(trans, __subvol_lookup(trans, subvol, snapshot, inum));
 106 }
 107
 108 static int lookup_first_inode(struct btree_trans *trans, u64 inode_nr,
 109                               struct bch_inode_unpacked *inode)
 110 {
 111         struct btree_iter iter;
 112         struct bkey_s_c k;
 113         int ret;
 114
 115         bch2_trans_iter_init(trans, &iter, BTREE_ID_inodes,
 116                              POS(0, inode_nr),
 117                              BTREE_ITER_ALL_SNAPSHOTS);
 118         k = bch2_btree_iter_peek(&iter);
 119         ret = bkey_err(k);
 120         if (ret)
 121                 goto err;
 122
 123         if (!k.k || !bkey_eq(k.k->p, POS(0, inode_nr))) {
 124                 ret = -BCH_ERR_ENOENT_inode;
 125                 goto err;
 126         }
 127
 128         ret = bch2_inode_unpack(k, inode);
 129 err:
 130         bch_err_msg(trans->c, ret, "fetching inode %llu", inode_nr);
 131         bch2_trans_iter_exit(trans, &iter);
 132         return ret;
 133 }
 134
 135 static int __lookup_inode(struct btree_trans *trans, u64 inode_nr,
 136                           struct bch_inode_unpacked *inode,
 137                           u32 *snapshot)
 138 {
 139         struct btree_iter iter;
 140         struct bkey_s_c k;
 141         int ret;
 142
 143         k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_inodes,
 144                                SPOS(0, inode_nr, *snapshot), 0);
 145         ret = bkey_err(k);
 146         if (ret)
 147                 goto err;
 148
 149         ret = bkey_is_inode(k.k)
 150                 ? bch2_inode_unpack(k, inode)
 151                 : -BCH_ERR_ENOENT_inode;
 152         if (!ret)
 153                 *snapshot = iter.pos.snapshot;
 154 err:
 155         bch_err_msg(trans->c, ret, "fetching inode %llu:%u", inode_nr, *snapshot);
 156         bch2_trans_iter_exit(trans, &iter);
 157         return ret;
 158 }
 159
 160 static int lookup_inode(struct btree_trans *trans, u64 inode_nr,
 161                         struct bch_inode_unpacked *inode,
 162                         u32 *snapshot)
 163 {
 164         return lockrestart_do(trans, __lookup_inode(trans, inode_nr, inode, snapshot));
 165 }
 166
 167 static int __lookup_dirent(struct btree_trans *trans,
 168                            struct bch_hash_info hash_info,
 169                            subvol_inum dir, struct qstr *name,
 170                            u64 *target, unsigned *type)
 171 {
 172         struct btree_iter iter;
 173         struct bkey_s_c_dirent d;
 174         int ret;
 175
 176         ret = bch2_hash_lookup(trans, &iter, bch2_dirent_hash_desc,
 177                                &hash_info, dir, name, 0);
 178         if (ret)
 179                 return ret;
 180
 181         d = bkey_s_c_to_dirent(bch2_btree_iter_peek_slot(&iter));
 182         *target = le64_to_cpu(d.v->d_inum);
 183         *type = d.v->d_type;
 184         bch2_trans_iter_exit(trans, &iter);
 185         return 0;
 186 }
 187
 188 static int __write_inode(struct btree_trans *trans,
 189                          struct bch_inode_unpacked *inode,
 190                          u32 snapshot)
 191 {
 192         struct bkey_inode_buf *inode_p =
 193                 bch2_trans_kmalloc(trans, sizeof(*inode_p));
 194
 195         if (IS_ERR(inode_p))
 196                 return PTR_ERR(inode_p);
 197
 198         bch2_inode_pack(inode_p, inode);
 199         inode_p->inode.k.p.snapshot = snapshot;
 200
 201         return bch2_btree_insert_nonextent(trans, BTREE_ID_inodes,
 202                                 &inode_p->inode.k_i,
 203                                 BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
 204 }
 205
 206 static int fsck_write_inode(struct btree_trans *trans,
 207                             struct bch_inode_unpacked *inode,
 208                             u32 snapshot)
 209 {
 210         int ret = commit_do(trans, NULL, NULL,
 211                                   BTREE_INSERT_NOFAIL|
 212                                   BTREE_INSERT_LAZY_RW,
 213                                   __write_inode(trans, inode, snapshot));
 214         if (ret)
 215                 bch_err_fn(trans->c, ret);
 216         return ret;
 217 }
 218
 219 static int __remove_dirent(struct btree_trans *trans, struct bpos pos)
 220 {
 221         struct bch_fs *c = trans->c;
 222         struct btree_iter iter;
 223         struct bch_inode_unpacked dir_inode;
 224         struct bch_hash_info dir_hash_info;
 225         int ret;
 226
 227         ret = lookup_first_inode(trans, pos.inode, &dir_inode);
 228         if (ret)
 229                 goto err;
 230
 231         dir_hash_info = bch2_hash_info_init(c, &dir_inode);
 232
 233         bch2_trans_iter_init(trans, &iter, BTREE_ID_dirents, pos, BTREE_ITER_INTENT);
 234
 235         ret = bch2_hash_delete_at(trans, bch2_dirent_hash_desc,
 236                                   &dir_hash_info, &iter,
 237                                   BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
 238         bch2_trans_iter_exit(trans, &iter);
 239 err:
 240         bch_err_fn(c, ret);
 241         return ret;
 242 }
 243
 244 /* Get lost+found, create if it doesn't exist: */
 245 static int lookup_lostfound(struct btree_trans *trans, u32 subvol,
 246                             struct bch_inode_unpacked *lostfound)
 247 {
 248         struct bch_fs *c = trans->c;
 249         struct bch_inode_unpacked root;
 250         struct bch_hash_info root_hash_info;
 251         struct qstr lostfound_str = QSTR("lost+found");
 252         subvol_inum root_inum = { .subvol = subvol };
 253         u64 inum = 0;
 254         unsigned d_type = 0;
 255         u32 snapshot;
 256         int ret;
 257
 258         ret = __subvol_lookup(trans, subvol, &snapshot, &root_inum.inum);
 259         if (ret)
 260                 return ret;
 261
 262         ret = __lookup_inode(trans, root_inum.inum, &root, &snapshot);
 263         if (ret)
 264                 return ret;
 265
 266         root_hash_info = bch2_hash_info_init(c, &root);
 267
 268         ret = __lookup_dirent(trans, root_hash_info, root_inum,
 269                             &lostfound_str, &inum, &d_type);
 270         if (bch2_err_matches(ret, ENOENT)) {
 271                 bch_notice(c, "creating lost+found");
 272                 goto create_lostfound;
 273         }
 274
 275         bch_err_fn(c, ret);
 276         if (ret)
 277                 return ret;
 278
 279         if (d_type != DT_DIR) {
 280                 bch_err(c, "error looking up lost+found: not a directory");
 281                 return -BCH_ERR_ENOENT_not_directory;
 282         }
 283
 284         /*
 285          * The bch2_check_dirents pass has already run, dangling dirents
 286          * shouldn't exist here:
 287          */
 288         return __lookup_inode(trans, inum, lostfound, &snapshot);
 289
 290 create_lostfound:
 291         bch2_inode_init_early(c, lostfound);
 292
 293         ret = bch2_create_trans(trans, root_inum, &root,
 294                                 lostfound, &lostfound_str,
 295                                 0, 0, S_IFDIR|0700, 0, NULL, NULL,
 296                                 (subvol_inum) { }, 0);
 297         bch_err_msg(c, ret, "creating lost+found");
 298         return ret;
 299 }
 300
 301 static int __reattach_inode(struct btree_trans *trans,
 302                           struct bch_inode_unpacked *inode,
 303                           u32 inode_snapshot)
 304 {
 305         struct bch_hash_info dir_hash;
 306         struct bch_inode_unpacked lostfound;
 307         char name_buf[20];
 308         struct qstr name;
 309         u64 dir_offset = 0;
 310         u32 subvol;
 311         int ret;
 312
 313         ret = __snapshot_lookup_subvol(trans, inode_snapshot, &subvol);
 314         if (ret)
 315                 return ret;
 316
 317         ret = lookup_lostfound(trans, subvol, &lostfound);
 318         if (ret)
 319                 return ret;
 320
 321         if (S_ISDIR(inode->bi_mode)) {
 322                 lostfound.bi_nlink++;
 323
 324                 ret = __write_inode(trans, &lostfound, U32_MAX);
 325                 if (ret)
 326                         return ret;
 327         }
 328
 329         dir_hash = bch2_hash_info_init(trans->c, &lostfound);
 330
 331         snprintf(name_buf, sizeof(name_buf), "%llu", inode->bi_inum);
 332         name = (struct qstr) QSTR(name_buf);
 333
 334         ret = bch2_dirent_create(trans,
 335                                  (subvol_inum) {
 336                                         .subvol = subvol,
 337                                         .inum = lostfound.bi_inum,
 338                                  },
 339                                  &dir_hash,
 340                                  inode_d_type(inode),
 341                                  &name, inode->bi_inum, &dir_offset,
 342                                  BCH_HASH_SET_MUST_CREATE);
 343         if (ret)
 344                 return ret;
 345
 346         inode->bi_dir           = lostfound.bi_inum;
 347         inode->bi_dir_offset    = dir_offset;
 348
 349         return __write_inode(trans, inode, inode_snapshot);
 350 }
 351
 352 static int reattach_inode(struct btree_trans *trans,
 353                           struct bch_inode_unpacked *inode,
 354                           u32 inode_snapshot)
 355 {
 356         int ret = commit_do(trans, NULL, NULL,
 357                                   BTREE_INSERT_LAZY_RW|
 358                                   BTREE_INSERT_NOFAIL,
 359                         __reattach_inode(trans, inode, inode_snapshot));
 360         bch_err_msg(trans->c, ret, "reattaching inode %llu", inode->bi_inum);
 361         return ret;
 362 }
 363
 364 static int remove_backpointer(struct btree_trans *trans,
 365                               struct bch_inode_unpacked *inode)
 366 {
 367         struct btree_iter iter;
 368         struct bkey_s_c_dirent d;
 369         int ret;
 370
 371         d = bch2_bkey_get_iter_typed(trans, &iter, BTREE_ID_dirents,
 372                                      POS(inode->bi_dir, inode->bi_dir_offset), 0,
 373                                      dirent);
 374         ret =   bkey_err(d) ?:
 375                 __remove_dirent(trans, d.k->p);
 376         bch2_trans_iter_exit(trans, &iter);
 377         return ret;
 378 }
 379
 380 struct snapshots_seen_entry {
 381         u32                             id;
 382         u32                             equiv;
 383 };
 384
 385 struct snapshots_seen {
 386         struct bpos                     pos;
 387         DARRAY(struct snapshots_seen_entry) ids;
 388 };
 389
 390 static inline void snapshots_seen_exit(struct snapshots_seen *s)
 391 {
 392         darray_exit(&s->ids);
 393 }
 394
 395 static inline void snapshots_seen_init(struct snapshots_seen *s)
 396 {
 397         memset(s, 0, sizeof(*s));
 398 }
 399
 400 static int snapshots_seen_add_inorder(struct bch_fs *c, struct snapshots_seen *s, u32 id)
 401 {
 402         struct snapshots_seen_entry *i, n = {
 403                 .id     = id,
 404                 .equiv  = bch2_snapshot_equiv(c, id),
 405         };
 406         int ret = 0;
 407
 408         darray_for_each(s->ids, i) {
 409                 if (i->id == id)
 410                         return 0;
 411                 if (i->id > id)
 412                         break;
 413         }
 414
 415         ret = darray_insert_item(&s->ids, i - s->ids.data, n);
 416         if (ret)
 417                 bch_err(c, "error reallocating snapshots_seen table (size %zu)",
 418                         s->ids.size);
 419         return ret;
 420 }
 421
 422 static int snapshots_seen_update(struct bch_fs *c, struct snapshots_seen *s,
 423                                  enum btree_id btree_id, struct bpos pos)
 424 {
 425         struct snapshots_seen_entry *i, n = {
 426                 .id     = pos.snapshot,
 427                 .equiv  = bch2_snapshot_equiv(c, pos.snapshot),
 428         };
 429         int ret = 0;
 430
 431         if (!bkey_eq(s->pos, pos))
 432                 s->ids.nr = 0;
 433
 434         s->pos = pos;
 435         s->pos.snapshot = n.equiv;
 436
 437         darray_for_each(s->ids, i) {
 438                 if (i->id == n.id)
 439                         return 0;
 440
 441                 /*
 442                  * We currently don't rigorously track for snapshot cleanup
 443                  * needing to be run, so it shouldn't be a fsck error yet:
 444                  */
 445                 if (i->equiv == n.equiv) {
 446                         bch_err(c, "snapshot deletion did not finish:\n"
 447                                 "  duplicate keys in btree %s at %llu:%llu snapshots %u, %u (equiv %u)\n",
 448                                 bch2_btree_id_str(btree_id),
 449                                 pos.inode, pos.offset,
 450                                 i->id, n.id, n.equiv);
 451                         set_bit(BCH_FS_NEED_DELETE_DEAD_SNAPSHOTS, &c->flags);
 452                         return bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_delete_dead_snapshots);
 453                 }
 454         }
 455
 456         ret = darray_push(&s->ids, n);
 457         if (ret)
 458                 bch_err(c, "error reallocating snapshots_seen table (size %zu)",
 459                         s->ids.size);
 460         return ret;
 461 }
 462
 463 /**
 464  * key_visible_in_snapshot - returns true if @id is a descendent of @ancestor,
 465  * and @ancestor hasn't been overwritten in @seen
 466  *
 467  * @c:          filesystem handle
 468  * @seen:       list of snapshot ids already seen at current position
 469  * @id:         descendent snapshot id
 470  * @ancestor:   ancestor snapshot id
 471  *
 472  * Returns:     whether key in @ancestor snapshot is visible in @id snapshot
 473  */
 474 static bool key_visible_in_snapshot(struct bch_fs *c, struct snapshots_seen *seen,
 475                                     u32 id, u32 ancestor)
 476 {
 477         ssize_t i;
 478
 479         EBUG_ON(id > ancestor);
 480         EBUG_ON(!bch2_snapshot_is_equiv(c, id));
 481         EBUG_ON(!bch2_snapshot_is_equiv(c, ancestor));
 482
 483         /* @ancestor should be the snapshot most recently added to @seen */
 484         EBUG_ON(ancestor != seen->pos.snapshot);
 485         EBUG_ON(ancestor != seen->ids.data[seen->ids.nr - 1].equiv);
 486
 487         if (id == ancestor)
 488                 return true;
 489
 490         if (!bch2_snapshot_is_ancestor(c, id, ancestor))
 491                 return false;
 492
 493         /*
 494          * We know that @id is a descendant of @ancestor, we're checking if
 495          * we've seen a key that overwrote @ancestor - i.e. also a descendent of
 496          * @ascestor and with @id as a descendent.
 497          *
 498          * But we already know that we're scanning IDs between @id and @ancestor
 499          * numerically, since snapshot ID lists are kept sorted, so if we find
 500          * an id that's an ancestor of @id we're done:
 501          */
 502
 503         for (i = seen->ids.nr - 2;
 504              i >= 0 && seen->ids.data[i].equiv >= id;
 505              --i)
 506                 if (bch2_snapshot_is_ancestor(c, id, seen->ids.data[i].equiv))
 507                         return false;
 508
 509         return true;
 510 }
 511
 512 /**
 513  * ref_visible - given a key with snapshot id @src that points to a key with
 514  * snapshot id @dst, test whether there is some snapshot in which @dst is
 515  * visible.
 516  *
 517  * @c:          filesystem handle
 518  * @s:          list of snapshot IDs already seen at @src
 519  * @src:        snapshot ID of src key
 520  * @dst:        snapshot ID of dst key
 521  * Returns:     true if there is some snapshot in which @dst is visible
 522  *
 523  * Assumes we're visiting @src keys in natural key order
 524  */
 525 static bool ref_visible(struct bch_fs *c, struct snapshots_seen *s,
 526                         u32 src, u32 dst)
 527 {
 528         return dst <= src
 529                 ? key_visible_in_snapshot(c, s, dst, src)
 530                 : bch2_snapshot_is_ancestor(c, src, dst);
 531 }
 532
 533 static int ref_visible2(struct bch_fs *c,
 534                         u32 src, struct snapshots_seen *src_seen,
 535                         u32 dst, struct snapshots_seen *dst_seen)
 536 {
 537         src = bch2_snapshot_equiv(c, src);
 538         dst = bch2_snapshot_equiv(c, dst);
 539
 540         if (dst > src) {
 541                 swap(dst, src);
 542                 swap(dst_seen, src_seen);
 543         }
 544         return key_visible_in_snapshot(c, src_seen, dst, src);
 545 }
 546
 547 #define for_each_visible_inode(_c, _s, _w, _snapshot, _i)                               \
 548         for (_i = (_w)->inodes.data; _i < (_w)->inodes.data + (_w)->inodes.nr &&        \
 549              (_i)->snapshot <= (_snapshot); _i++)                                       \
 550                 if (key_visible_in_snapshot(_c, _s, _i->snapshot, _snapshot))
 551
 552 struct inode_walker_entry {
 553         struct bch_inode_unpacked inode;
 554         u32                     snapshot;
 555         bool                    seen_this_pos;
 556         u64                     count;
 557 };
 558
 559 struct inode_walker {
 560         bool                            first_this_inode;
 561         bool                            recalculate_sums;
 562         struct bpos                     last_pos;
 563
 564         DARRAY(struct inode_walker_entry) inodes;
 565 };
 566
 567 static void inode_walker_exit(struct inode_walker *w)
 568 {
 569         darray_exit(&w->inodes);
 570 }
 571
 572 static struct inode_walker inode_walker_init(void)
 573 {
 574         return (struct inode_walker) { 0, };
 575 }
 576
 577 static int add_inode(struct bch_fs *c, struct inode_walker *w,
 578                      struct bkey_s_c inode)
 579 {
 580         struct bch_inode_unpacked u;
 581
 582         BUG_ON(bch2_inode_unpack(inode, &u));
 583
 584         return darray_push(&w->inodes, ((struct inode_walker_entry) {
 585                 .inode          = u,
 586                 .snapshot       = bch2_snapshot_equiv(c, inode.k->p.snapshot),
 587         }));
 588 }
 589
 590 static int get_inodes_all_snapshots(struct btree_trans *trans,
 591                                     struct inode_walker *w, u64 inum)
 592 {
 593         struct bch_fs *c = trans->c;
 594         struct btree_iter iter;
 595         struct bkey_s_c k;
 596         u32 restart_count = trans->restart_count;
 597         int ret;
 598
 599         w->recalculate_sums = false;
 600         w->inodes.nr = 0;
 601
 602         for_each_btree_key(trans, iter, BTREE_ID_inodes, POS(0, inum),
 603                            BTREE_ITER_ALL_SNAPSHOTS, k, ret) {
 604                 if (k.k->p.offset != inum)
 605                         break;
 606
 607                 if (bkey_is_inode(k.k))
 608                         add_inode(c, w, k);
 609         }
 610         bch2_trans_iter_exit(trans, &iter);
 611
 612         if (ret)
 613                 return ret;
 614
 615         w->first_this_inode = true;
 616
 617         return trans_was_restarted(trans, restart_count);
 618 }
 619
 620 static struct inode_walker_entry *
 621 lookup_inode_for_snapshot(struct bch_fs *c, struct inode_walker *w,
 622                           u32 snapshot, bool is_whiteout)
 623 {
 624         struct inode_walker_entry *i;
 625
 626         snapshot = bch2_snapshot_equiv(c, snapshot);
 627
 628         darray_for_each(w->inodes, i)
 629                 if (bch2_snapshot_is_ancestor(c, snapshot, i->snapshot))
 630                         goto found;
 631
 632         return NULL;
 633 found:
 634         BUG_ON(snapshot > i->snapshot);
 635
 636         if (snapshot != i->snapshot && !is_whiteout) {
 637                 struct inode_walker_entry new = *i;
 638                 size_t pos;
 639                 int ret;
 640
 641                 new.snapshot = snapshot;
 642                 new.count = 0;
 643
 644                 bch_info(c, "have key for inode %llu:%u but have inode in ancestor snapshot %u",
 645                          w->last_pos.inode, snapshot, i->snapshot);
 646
 647                 while (i > w->inodes.data && i[-1].snapshot > snapshot)
 648                         --i;
 649
 650                 pos = i - w->inodes.data;
 651                 ret = darray_insert_item(&w->inodes, pos, new);
 652                 if (ret)
 653                         return ERR_PTR(ret);
 654
 655                 i = w->inodes.data + pos;
 656         }
 657
 658         return i;
 659 }
 660
 661 static struct inode_walker_entry *walk_inode(struct btree_trans *trans,
 662                                              struct inode_walker *w, struct bpos pos,
 663                                              bool is_whiteout)
 664 {
 665         if (w->last_pos.inode != pos.inode) {
 666                 int ret = get_inodes_all_snapshots(trans, w, pos.inode);
 667                 if (ret)
 668                         return ERR_PTR(ret);
 669         } else if (bkey_cmp(w->last_pos, pos)) {
 670                 struct inode_walker_entry *i;
 671
 672                 darray_for_each(w->inodes, i)
 673                         i->seen_this_pos = false;
 674
 675         }
 676
 677         w->last_pos = pos;
 678
 679         return lookup_inode_for_snapshot(trans->c, w, pos.snapshot, is_whiteout);
 680 }
 681
 682 static int __get_visible_inodes(struct btree_trans *trans,
 683                                 struct inode_walker *w,
 684                                 struct snapshots_seen *s,
 685                                 u64 inum)
 686 {
 687         struct bch_fs *c = trans->c;
 688         struct btree_iter iter;
 689         struct bkey_s_c k;
 690         int ret;
 691
 692         w->inodes.nr = 0;
 693
 694         for_each_btree_key_norestart(trans, iter, BTREE_ID_inodes, POS(0, inum),
 695                            BTREE_ITER_ALL_SNAPSHOTS, k, ret) {
 696                 u32 equiv = bch2_snapshot_equiv(c, k.k->p.snapshot);
 697
 698                 if (k.k->p.offset != inum)
 699                         break;
 700
 701                 if (!ref_visible(c, s, s->pos.snapshot, equiv))
 702                         continue;
 703
 704                 if (bkey_is_inode(k.k))
 705                         add_inode(c, w, k);
 706
 707                 if (equiv >= s->pos.snapshot)
 708                         break;
 709         }
 710         bch2_trans_iter_exit(trans, &iter);
 711
 712         return ret;
 713 }
 714
 715 static int check_key_has_snapshot(struct btree_trans *trans,
 716                                   struct btree_iter *iter,
 717                                   struct bkey_s_c k)
 718 {
 719         struct bch_fs *c = trans->c;
 720         struct printbuf buf = PRINTBUF;
 721         int ret = 0;
 722
 723         if (mustfix_fsck_err_on(!bch2_snapshot_equiv(c, k.k->p.snapshot), c,
 724                                 bkey_in_missing_snapshot,
 725                                 "key in missing snapshot: %s",
 726                                 (bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
 727                 ret = bch2_btree_delete_at(trans, iter,
 728                                             BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE) ?: 1;
 729 fsck_err:
 730         printbuf_exit(&buf);
 731         return ret;
 732 }
 733
 734 static int hash_redo_key(struct btree_trans *trans,
 735                          const struct bch_hash_desc desc,
 736                          struct bch_hash_info *hash_info,
 737                          struct btree_iter *k_iter, struct bkey_s_c k)
 738 {
 739         struct bkey_i *delete;
 740         struct bkey_i *tmp;
 741
 742         delete = bch2_trans_kmalloc(trans, sizeof(*delete));
 743         if (IS_ERR(delete))
 744                 return PTR_ERR(delete);
 745
 746         tmp = bch2_bkey_make_mut_noupdate(trans, k);
 747         if (IS_ERR(tmp))
 748                 return PTR_ERR(tmp);
 749
 750         bkey_init(&delete->k);
 751         delete->k.p = k_iter->pos;
 752         return  bch2_btree_iter_traverse(k_iter) ?:
 753                 bch2_trans_update(trans, k_iter, delete, 0) ?:
 754                 bch2_hash_set_snapshot(trans, desc, hash_info,
 755                                        (subvol_inum) { 0, k.k->p.inode },
 756                                        k.k->p.snapshot, tmp,
 757                                        BCH_HASH_SET_MUST_CREATE,
 758                                        BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE) ?:
 759                 bch2_trans_commit(trans, NULL, NULL,
 760                                   BTREE_INSERT_NOFAIL|
 761                                   BTREE_INSERT_LAZY_RW);
 762 }
 763
 764 static int hash_check_key(struct btree_trans *trans,
 765                           const struct bch_hash_desc desc,
 766                           struct bch_hash_info *hash_info,
 767                           struct btree_iter *k_iter, struct bkey_s_c hash_k)
 768 {
 769         struct bch_fs *c = trans->c;
 770         struct btree_iter iter = { NULL };
 771         struct printbuf buf = PRINTBUF;
 772         struct bkey_s_c k;
 773         u64 hash;
 774         int ret = 0;
 775
 776         if (hash_k.k->type != desc.key_type)
 777                 return 0;
 778
 779         hash = desc.hash_bkey(hash_info, hash_k);
 780
 781         if (likely(hash == hash_k.k->p.offset))
 782                 return 0;
 783
 784         if (hash_k.k->p.offset < hash)
 785                 goto bad_hash;
 786
 787         for_each_btree_key_norestart(trans, iter, desc.btree_id,
 788                                      SPOS(hash_k.k->p.inode, hash, hash_k.k->p.snapshot),
 789                                      BTREE_ITER_SLOTS, k, ret) {
 790                 if (bkey_eq(k.k->p, hash_k.k->p))
 791                         break;
 792
 793                 if (fsck_err_on(k.k->type == desc.key_type &&
 794                                 !desc.cmp_bkey(k, hash_k), c,
 795                                 hash_table_key_duplicate,
 796                                 "duplicate hash table keys:\n%s",
 797                                 (printbuf_reset(&buf),
 798                                  bch2_bkey_val_to_text(&buf, c, hash_k),
 799                                  buf.buf))) {
 800                         ret = bch2_hash_delete_at(trans, desc, hash_info, k_iter, 0) ?: 1;
 801                         break;
 802                 }
 803
 804                 if (bkey_deleted(k.k)) {
 805                         bch2_trans_iter_exit(trans, &iter);
 806                         goto bad_hash;
 807                 }
 808         }
 809 out:
 810         bch2_trans_iter_exit(trans, &iter);
 811         printbuf_exit(&buf);
 812         return ret;
 813 bad_hash:
 814         if (fsck_err(c, hash_table_key_wrong_offset,
 815                      "hash table key at wrong offset: btree %s inode %llu offset %llu, hashed to %llu\n%s",
 816                      bch2_btree_id_str(desc.btree_id), hash_k.k->p.inode, hash_k.k->p.offset, hash,
 817                      (printbuf_reset(&buf),
 818                       bch2_bkey_val_to_text(&buf, c, hash_k), buf.buf))) {
 819                 ret = hash_redo_key(trans, desc, hash_info, k_iter, hash_k);
 820                 bch_err_fn(c, ret);
 821                 if (ret)
 822                         return ret;
 823                 ret = -BCH_ERR_transaction_restart_nested;
 824         }
 825 fsck_err:
 826         goto out;
 827 }
 828
 829 static int check_inode(struct btree_trans *trans,
 830                        struct btree_iter *iter,
 831                        struct bkey_s_c k,
 832                        struct bch_inode_unpacked *prev,
 833                        struct snapshots_seen *s,
 834                        bool full)
 835 {
 836         struct bch_fs *c = trans->c;
 837         struct bch_inode_unpacked u;
 838         bool do_update = false;
 839         int ret;
 840
 841         ret = check_key_has_snapshot(trans, iter, k);
 842         if (ret < 0)
 843                 goto err;
 844         if (ret)
 845                 return 0;
 846
 847         ret = snapshots_seen_update(c, s, iter->btree_id, k.k->p);
 848         if (ret)
 849                 goto err;
 850
 851         if (!bkey_is_inode(k.k))
 852                 return 0;
 853
 854         BUG_ON(bch2_inode_unpack(k, &u));
 855
 856         if (!full &&
 857             !(u.bi_flags & (BCH_INODE_i_size_dirty|
 858                             BCH_INODE_i_sectors_dirty|
 859                             BCH_INODE_unlinked)))
 860                 return 0;
 861
 862         if (prev->bi_inum != u.bi_inum)
 863                 *prev = u;
 864
 865         if (fsck_err_on(prev->bi_hash_seed      != u.bi_hash_seed ||
 866                         inode_d_type(prev)      != inode_d_type(&u),
 867                         c, inode_snapshot_mismatch,
 868                         "inodes in different snapshots don't match")) {
 869                 bch_err(c, "repair not implemented yet");
 870                 return -EINVAL;
 871         }
 872
 873         if ((u.bi_flags & (BCH_INODE_i_size_dirty|BCH_INODE_unlinked)) &&
 874             bch2_key_has_snapshot_overwrites(trans, BTREE_ID_inodes, k.k->p)) {
 875                 struct bpos new_min_pos;
 876
 877                 ret = bch2_propagate_key_to_snapshot_leaves(trans, iter->btree_id, k, &new_min_pos);
 878                 if (ret)
 879                         goto err;
 880
 881                 u.bi_flags &= ~BCH_INODE_i_size_dirty|BCH_INODE_unlinked;
 882
 883                 ret = __write_inode(trans, &u, iter->pos.snapshot);
 884                 bch_err_msg(c, ret, "in fsck updating inode");
 885                 if (ret)
 886                         return ret;
 887
 888                 if (!bpos_eq(new_min_pos, POS_MIN))
 889                         bch2_btree_iter_set_pos(iter, bpos_predecessor(new_min_pos));
 890                 return 0;
 891         }
 892
 893         if (u.bi_flags & BCH_INODE_unlinked &&
 894             (!c->sb.clean ||
 895              fsck_err(c, inode_unlinked_but_clean,
 896                       "filesystem marked clean, but inode %llu unlinked",
 897                       u.bi_inum))) {
 898                 bch2_trans_unlock(trans);
 899                 bch2_fs_lazy_rw(c);
 900
 901                 ret = bch2_inode_rm_snapshot(trans, u.bi_inum, iter->pos.snapshot);
 902                 bch_err_msg(c, ret, "in fsck deleting inode");
 903                 return ret;
 904         }
 905
 906         if (u.bi_flags & BCH_INODE_i_size_dirty &&
 907             (!c->sb.clean ||
 908              fsck_err(c, inode_i_size_dirty_but_clean,
 909                       "filesystem marked clean, but inode %llu has i_size dirty",
 910                       u.bi_inum))) {
 911                 bch_verbose(c, "truncating inode %llu", u.bi_inum);
 912
 913                 bch2_trans_unlock(trans);
 914                 bch2_fs_lazy_rw(c);
 915
 916                 /*
 917                  * XXX: need to truncate partial blocks too here - or ideally
 918                  * just switch units to bytes and that issue goes away
 919                  */
 920                 ret = bch2_btree_delete_range_trans(trans, BTREE_ID_extents,
 921                                 SPOS(u.bi_inum, round_up(u.bi_size, block_bytes(c)) >> 9,
 922                                      iter->pos.snapshot),
 923                                 POS(u.bi_inum, U64_MAX),
 924                                 0, NULL);
 925                 bch_err_msg(c, ret, "in fsck truncating inode");
 926                 if (ret)
 927                         return ret;
 928
 929                 /*
 930                  * We truncated without our normal sector accounting hook, just
 931                  * make sure we recalculate it:
 932                  */
 933                 u.bi_flags |= BCH_INODE_i_sectors_dirty;
 934
 935                 u.bi_flags &= ~BCH_INODE_i_size_dirty;
 936                 do_update = true;
 937         }
 938
 939         if (u.bi_flags & BCH_INODE_i_sectors_dirty &&
 940             (!c->sb.clean ||
 941              fsck_err(c, inode_i_sectors_dirty_but_clean,
 942                       "filesystem marked clean, but inode %llu has i_sectors dirty",
 943                       u.bi_inum))) {
 944                 s64 sectors;
 945
 946                 bch_verbose(c, "recounting sectors for inode %llu",
 947                             u.bi_inum);
 948
 949                 sectors = bch2_count_inode_sectors(trans, u.bi_inum, iter->pos.snapshot);
 950                 if (sectors < 0) {
 951                         bch_err_msg(c, sectors, "in fsck recounting inode sectors");
 952                         return sectors;
 953                 }
 954
 955                 u.bi_sectors = sectors;
 956                 u.bi_flags &= ~BCH_INODE_i_sectors_dirty;
 957                 do_update = true;
 958         }
 959
 960         if (u.bi_flags & BCH_INODE_backptr_untrusted) {
 961                 u.bi_dir = 0;
 962                 u.bi_dir_offset = 0;
 963                 u.bi_flags &= ~BCH_INODE_backptr_untrusted;
 964                 do_update = true;
 965         }
 966
 967         if (do_update) {
 968                 ret = __write_inode(trans, &u, iter->pos.snapshot);
 969                 bch_err_msg(c, ret, "in fsck updating inode");
 970                 if (ret)
 971                         return ret;
 972         }
 973 err:
 974 fsck_err:
 975         bch_err_fn(c, ret);
 976         return ret;
 977 }
 978
 979 noinline_for_stack
 980 int bch2_check_inodes(struct bch_fs *c)
 981 {
 982         bool full = c->opts.fsck;
 983         struct btree_trans *trans = bch2_trans_get(c);
 984         struct btree_iter iter;
 985         struct bch_inode_unpacked prev = { 0 };
 986         struct snapshots_seen s;
 987         struct bkey_s_c k;
 988         int ret;
 989
 990         snapshots_seen_init(&s);
 991
 992         ret = for_each_btree_key_commit(trans, iter, BTREE_ID_inodes,
 993                         POS_MIN,
 994                         BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS, k,
 995                         NULL, NULL, BTREE_INSERT_LAZY_RW|BTREE_INSERT_NOFAIL,
 996                 check_inode(trans, &iter, k, &prev, &s, full));
 997
 998         snapshots_seen_exit(&s);
 999         bch2_trans_put(trans);
1000         bch_err_fn(c, ret);
1001         return ret;
1002 }
1003
1004 static struct bkey_s_c_dirent dirent_get_by_pos(struct btree_trans *trans,
1005                                                 struct btree_iter *iter,
1006                                                 struct bpos pos)
1007 {
1008         return bch2_bkey_get_iter_typed(trans, iter, BTREE_ID_dirents, pos, 0, dirent);
1009 }
1010
1011 static bool inode_points_to_dirent(struct bch_inode_unpacked *inode,
1012                                    struct bkey_s_c_dirent d)
1013 {
1014         return  inode->bi_dir           == d.k->p.inode &&
1015                 inode->bi_dir_offset    == d.k->p.offset;
1016 }
1017
1018 static bool dirent_points_to_inode(struct bkey_s_c_dirent d,
1019                                    struct bch_inode_unpacked *inode)
1020 {
1021         return d.v->d_type == DT_SUBVOL
1022                 ? le32_to_cpu(d.v->d_child_subvol)      == inode->bi_subvol
1023                 : le64_to_cpu(d.v->d_inum)              == inode->bi_inum;
1024 }
1025
1026 static int inode_backpointer_exists(struct btree_trans *trans,
1027                                     struct bch_inode_unpacked *inode,
1028                                     u32 snapshot)
1029 {
1030         struct btree_iter iter;
1031         struct bkey_s_c_dirent d;
1032         int ret;
1033
1034         d = dirent_get_by_pos(trans, &iter,
1035                         SPOS(inode->bi_dir, inode->bi_dir_offset, snapshot));
1036         ret = bkey_err(d);
1037         if (ret)
1038                 return bch2_err_matches(ret, ENOENT) ? 0 : ret;
1039
1040         ret = dirent_points_to_inode(d, inode);
1041         bch2_trans_iter_exit(trans, &iter);
1042         return ret;
1043 }
1044
1045 static int check_i_sectors(struct btree_trans *trans, struct inode_walker *w)
1046 {
1047         struct bch_fs *c = trans->c;
1048         struct inode_walker_entry *i;
1049         u32 restart_count = trans->restart_count;
1050         int ret = 0;
1051         s64 count2;
1052
1053         darray_for_each(w->inodes, i) {
1054                 if (i->inode.bi_sectors == i->count)
1055                         continue;
1056
1057                 count2 = bch2_count_inode_sectors(trans, w->last_pos.inode, i->snapshot);
1058
1059                 if (w->recalculate_sums)
1060                         i->count = count2;
1061
1062                 if (i->count != count2) {
1063                         bch_err(c, "fsck counted i_sectors wrong for inode %llu:%u: got %llu should be %llu",
1064                                 w->last_pos.inode, i->snapshot, i->count, count2);
1065                         return -BCH_ERR_internal_fsck_err;
1066                 }
1067
1068                 if (fsck_err_on(!(i->inode.bi_flags & BCH_INODE_i_sectors_dirty),
1069                                 c, inode_i_sectors_wrong,
1070                                 "inode %llu:%u has incorrect i_sectors: got %llu, should be %llu",
1071                                 w->last_pos.inode, i->snapshot,
1072                                 i->inode.bi_sectors, i->count)) {
1073                         i->inode.bi_sectors = i->count;
1074                         ret = fsck_write_inode(trans, &i->inode, i->snapshot);
1075                         if (ret)
1076                                 break;
1077                 }
1078         }
1079 fsck_err:
1080         bch_err_fn(c, ret);
1081         return ret ?: trans_was_restarted(trans, restart_count);
1082 }
1083
1084 struct extent_end {
1085         u32                     snapshot;
1086         u64                     offset;
1087         struct snapshots_seen   seen;
1088 };
1089
1090 struct extent_ends {
1091         struct bpos                     last_pos;
1092         DARRAY(struct extent_end)       e;
1093 };
1094
1095 static void extent_ends_reset(struct extent_ends *extent_ends)
1096 {
1097         struct extent_end *i;
1098
1099         darray_for_each(extent_ends->e, i)
1100                 snapshots_seen_exit(&i->seen);
1101
1102         extent_ends->e.nr = 0;
1103 }
1104
1105 static void extent_ends_exit(struct extent_ends *extent_ends)
1106 {
1107         extent_ends_reset(extent_ends);
1108         darray_exit(&extent_ends->e);
1109 }
1110
1111 static void extent_ends_init(struct extent_ends *extent_ends)
1112 {
1113         memset(extent_ends, 0, sizeof(*extent_ends));
1114 }
1115
1116 static int extent_ends_at(struct bch_fs *c,
1117                           struct extent_ends *extent_ends,
1118                           struct snapshots_seen *seen,
1119                           struct bkey_s_c k)
1120 {
1121         struct extent_end *i, n = (struct extent_end) {
1122                 .offset         = k.k->p.offset,
1123                 .snapshot       = k.k->p.snapshot,
1124                 .seen           = *seen,
1125         };
1126
1127         n.seen.ids.data = kmemdup(seen->ids.data,
1128                               sizeof(seen->ids.data[0]) * seen->ids.size,
1129                               GFP_KERNEL);
1130         if (!n.seen.ids.data)
1131                 return -BCH_ERR_ENOMEM_fsck_extent_ends_at;
1132
1133         darray_for_each(extent_ends->e, i) {
1134                 if (i->snapshot == k.k->p.snapshot) {
1135                         snapshots_seen_exit(&i->seen);
1136                         *i = n;
1137                         return 0;
1138                 }
1139
1140                 if (i->snapshot >= k.k->p.snapshot)
1141                         break;
1142         }
1143
1144         return darray_insert_item(&extent_ends->e, i - extent_ends->e.data, n);
1145 }
1146
1147 static int overlapping_extents_found(struct btree_trans *trans,
1148                                      enum btree_id btree,
1149                                      struct bpos pos1, struct snapshots_seen *pos1_seen,
1150                                      struct bkey pos2,
1151                                      bool *fixed,
1152                                      struct extent_end *extent_end)
1153 {
1154         struct bch_fs *c = trans->c;
1155         struct printbuf buf = PRINTBUF;
1156         struct btree_iter iter1, iter2 = { NULL };
1157         struct bkey_s_c k1, k2;
1158         int ret;
1159
1160         BUG_ON(bkey_le(pos1, bkey_start_pos(&pos2)));
1161
1162         bch2_trans_iter_init(trans, &iter1, btree, pos1,
1163                              BTREE_ITER_ALL_SNAPSHOTS|
1164                              BTREE_ITER_NOT_EXTENTS);
1165         k1 = bch2_btree_iter_peek_upto(&iter1, POS(pos1.inode, U64_MAX));
1166         ret = bkey_err(k1);
1167         if (ret)
1168                 goto err;
1169
1170         prt_str(&buf, "\n  ");
1171         bch2_bkey_val_to_text(&buf, c, k1);
1172
1173         if (!bpos_eq(pos1, k1.k->p)) {
1174                 prt_str(&buf, "\n  wanted\n  ");
1175                 bch2_bpos_to_text(&buf, pos1);
1176                 prt_str(&buf, "\n  ");
1177                 bch2_bkey_to_text(&buf, &pos2);
1178
1179                 bch_err(c, "%s: error finding first overlapping extent when repairing, got%s",
1180                         __func__, buf.buf);
1181                 ret = -BCH_ERR_internal_fsck_err;
1182                 goto err;
1183         }
1184
1185         bch2_trans_copy_iter(&iter2, &iter1);
1186
1187         while (1) {
1188                 bch2_btree_iter_advance(&iter2);
1189
1190                 k2 = bch2_btree_iter_peek_upto(&iter2, POS(pos1.inode, U64_MAX));
1191                 ret = bkey_err(k2);
1192                 if (ret)
1193                         goto err;
1194
1195                 if (bpos_ge(k2.k->p, pos2.p))
1196                         break;
1197         }
1198
1199         prt_str(&buf, "\n  ");
1200         bch2_bkey_val_to_text(&buf, c, k2);
1201
1202         if (bpos_gt(k2.k->p, pos2.p) ||
1203             pos2.size != k2.k->size) {
1204                 bch_err(c, "%s: error finding seconding overlapping extent when repairing%s",
1205                         __func__, buf.buf);
1206                 ret = -BCH_ERR_internal_fsck_err;
1207                 goto err;
1208         }
1209
1210         prt_printf(&buf, "\n  overwriting %s extent",
1211                    pos1.snapshot >= pos2.p.snapshot ? "first" : "second");
1212
1213         if (fsck_err(c, extent_overlapping,
1214                      "overlapping extents%s", buf.buf)) {
1215                 struct btree_iter *old_iter = &iter1;
1216                 struct disk_reservation res = { 0 };
1217
1218                 if (pos1.snapshot < pos2.p.snapshot) {
1219                         old_iter = &iter2;
1220                         swap(k1, k2);
1221                 }
1222
1223                 trans->extra_journal_res += bch2_bkey_sectors_compressed(k2);
1224
1225                 ret =   bch2_trans_update_extent_overwrite(trans, old_iter,
1226                                 BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE,
1227                                 k1, k2) ?:
1228                         bch2_trans_commit(trans, &res, NULL,
1229                                 BTREE_INSERT_LAZY_RW|BTREE_INSERT_NOFAIL);
1230                 bch2_disk_reservation_put(c, &res);
1231
1232                 if (ret)
1233                         goto err;
1234
1235                 *fixed = true;
1236
1237                 if (pos1.snapshot == pos2.p.snapshot) {
1238                         /*
1239                          * We overwrote the first extent, and did the overwrite
1240                          * in the same snapshot:
1241                          */
1242                         extent_end->offset = bkey_start_offset(&pos2);
1243                 } else if (pos1.snapshot > pos2.p.snapshot) {
1244                         /*
1245                          * We overwrote the first extent in pos2's snapshot:
1246                          */
1247                         ret = snapshots_seen_add_inorder(c, pos1_seen, pos2.p.snapshot);
1248                 } else {
1249                         /*
1250                          * We overwrote the second extent - restart
1251                          * check_extent() from the top:
1252                          */
1253                         ret = -BCH_ERR_transaction_restart_nested;
1254                 }
1255         }
1256 fsck_err:
1257 err:
1258         bch2_trans_iter_exit(trans, &iter2);
1259         bch2_trans_iter_exit(trans, &iter1);
1260         printbuf_exit(&buf);
1261         return ret;
1262 }
1263
1264 static int check_overlapping_extents(struct btree_trans *trans,
1265                               struct snapshots_seen *seen,
1266                               struct extent_ends *extent_ends,
1267                               struct bkey_s_c k,
1268                               u32 equiv,
1269                               struct btree_iter *iter,
1270                               bool *fixed)
1271 {
1272         struct bch_fs *c = trans->c;
1273         struct extent_end *i;
1274         int ret = 0;
1275
1276         /* transaction restart, running again */
1277         if (bpos_eq(extent_ends->last_pos, k.k->p))
1278                 return 0;
1279
1280         if (extent_ends->last_pos.inode != k.k->p.inode)
1281                 extent_ends_reset(extent_ends);
1282
1283         darray_for_each(extent_ends->e, i) {
1284                 if (i->offset <= bkey_start_offset(k.k))
1285                         continue;
1286
1287                 if (!ref_visible2(c,
1288                                   k.k->p.snapshot, seen,
1289                                   i->snapshot, &i->seen))
1290                         continue;
1291
1292                 ret = overlapping_extents_found(trans, iter->btree_id,
1293                                                 SPOS(iter->pos.inode,
1294                                                      i->offset,
1295                                                      i->snapshot),
1296                                                 &i->seen,
1297                                                 *k.k, fixed, i);
1298                 if (ret)
1299                         goto err;
1300         }
1301
1302         ret = extent_ends_at(c, extent_ends, seen, k);
1303         if (ret)
1304                 goto err;
1305
1306         extent_ends->last_pos = k.k->p;
1307 err:
1308         return ret;
1309 }
1310
1311 static int check_extent_overbig(struct btree_trans *trans, struct btree_iter *iter,
1312                                 struct bkey_s_c k)
1313 {
1314         struct bch_fs *c = trans->c;
1315         struct bkey_ptrs_c ptrs = bch2_bkey_ptrs_c(k);
1316         struct bch_extent_crc_unpacked crc;
1317         const union bch_extent_entry *i;
1318         unsigned encoded_extent_max_sectors = c->opts.encoded_extent_max >> 9;
1319
1320         bkey_for_each_crc(k.k, ptrs, crc, i)
1321                 if (crc_is_encoded(crc) &&
1322                     crc.uncompressed_size > encoded_extent_max_sectors) {
1323                         struct printbuf buf = PRINTBUF;
1324
1325                         bch2_bkey_val_to_text(&buf, c, k);
1326                         bch_err(c, "overbig encoded extent, please report this:\n  %s", buf.buf);
1327                         printbuf_exit(&buf);
1328                 }
1329
1330         return 0;
1331 }
1332
1333 static int check_extent(struct btree_trans *trans, struct btree_iter *iter,
1334                         struct bkey_s_c k,
1335                         struct inode_walker *inode,
1336                         struct snapshots_seen *s,
1337                         struct extent_ends *extent_ends)
1338 {
1339         struct bch_fs *c = trans->c;
1340         struct inode_walker_entry *i;
1341         struct printbuf buf = PRINTBUF;
1342         struct bpos equiv = k.k->p;
1343         int ret = 0;
1344
1345         equiv.snapshot = bch2_snapshot_equiv(c, k.k->p.snapshot);
1346
1347         ret = check_key_has_snapshot(trans, iter, k);
1348         if (ret) {
1349                 ret = ret < 0 ? ret : 0;
1350                 goto out;
1351         }
1352
1353         if (inode->last_pos.inode != k.k->p.inode) {
1354                 ret = check_i_sectors(trans, inode);
1355                 if (ret)
1356                         goto err;
1357         }
1358
1359         i = walk_inode(trans, inode, equiv, k.k->type == KEY_TYPE_whiteout);
1360         ret = PTR_ERR_OR_ZERO(i);
1361         if (ret)
1362                 goto err;
1363
1364         ret = snapshots_seen_update(c, s, iter->btree_id, k.k->p);
1365         if (ret)
1366                 goto err;
1367
1368         if (k.k->type != KEY_TYPE_whiteout) {
1369                 if (fsck_err_on(!i, c, extent_in_missing_inode,
1370                                 "extent in missing inode:\n  %s",
1371                                 (printbuf_reset(&buf),
1372                                  bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
1373                         goto delete;
1374
1375                 if (fsck_err_on(i &&
1376                                 !S_ISREG(i->inode.bi_mode) &&
1377                                 !S_ISLNK(i->inode.bi_mode),
1378                                 c, extent_in_non_reg_inode,
1379                                 "extent in non regular inode mode %o:\n  %s",
1380                                 i->inode.bi_mode,
1381                                 (printbuf_reset(&buf),
1382                                  bch2_bkey_val_to_text(&buf, c, k), buf.buf)))
1383                         goto delete;
1384
1385                 ret = check_overlapping_extents(trans, s, extent_ends, k,
1386                                                 equiv.snapshot, iter,
1387                                                 &inode->recalculate_sums);
1388                 if (ret)
1389                         goto err;
1390         }
1391
1392         /*
1393          * Check inodes in reverse order, from oldest snapshots to newest,
1394          * starting from the inode that matches this extent's snapshot. If we
1395          * didn't have one, iterate over all inodes:
1396          */
1397         if (!i)
1398                 i = inode->inodes.data + inode->inodes.nr - 1;
1399
1400         for (;
1401              inode->inodes.data && i >= inode->inodes.data;
1402              --i) {
1403                 if (i->snapshot > equiv.snapshot ||
1404                     !key_visible_in_snapshot(c, s, i->snapshot, equiv.snapshot))
1405                         continue;
1406
1407                 if (k.k->type != KEY_TYPE_whiteout) {
1408                         if (fsck_err_on(!(i->inode.bi_flags & BCH_INODE_i_size_dirty) &&
1409                                         k.k->p.offset > round_up(i->inode.bi_size, block_bytes(c)) >> 9 &&
1410                                         !bkey_extent_is_reservation(k),
1411                                         c, extent_past_end_of_inode,
1412                                         "extent type past end of inode %llu:%u, i_size %llu\n  %s",
1413                                         i->inode.bi_inum, i->snapshot, i->inode.bi_size,
1414                                         (bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
1415                                 struct btree_iter iter2;
1416
1417                                 bch2_trans_copy_iter(&iter2, iter);
1418                                 bch2_btree_iter_set_snapshot(&iter2, i->snapshot);
1419                                 ret =   bch2_btree_iter_traverse(&iter2) ?:
1420                                         bch2_btree_delete_at(trans, &iter2,
1421                                                 BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
1422                                 bch2_trans_iter_exit(trans, &iter2);
1423                                 if (ret)
1424                                         goto err;
1425
1426                                 iter->k.type = KEY_TYPE_whiteout;
1427                         }
1428
1429                         if (bkey_extent_is_allocation(k.k))
1430                                 i->count += k.k->size;
1431                 }
1432
1433                 i->seen_this_pos = true;
1434         }
1435 out:
1436 err:
1437 fsck_err:
1438         printbuf_exit(&buf);
1439         bch_err_fn(c, ret);
1440         return ret;
1441 delete:
1442         ret = bch2_btree_delete_at(trans, iter, BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
1443         goto out;
1444 }
1445
1446 /*
1447  * Walk extents: verify that extents have a corresponding S_ISREG inode, and
1448  * that i_size an i_sectors are consistent
1449  */
1450 int bch2_check_extents(struct bch_fs *c)
1451 {
1452         struct inode_walker w = inode_walker_init();
1453         struct snapshots_seen s;
1454         struct btree_trans *trans = bch2_trans_get(c);
1455         struct btree_iter iter;
1456         struct bkey_s_c k;
1457         struct extent_ends extent_ends;
1458         struct disk_reservation res = { 0 };
1459         int ret = 0;
1460
1461         snapshots_seen_init(&s);
1462         extent_ends_init(&extent_ends);
1463
1464         ret = for_each_btree_key_commit(trans, iter, BTREE_ID_extents,
1465                         POS(BCACHEFS_ROOT_INO, 0),
1466                         BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS, k,
1467                         &res, NULL,
1468                         BTREE_INSERT_LAZY_RW|BTREE_INSERT_NOFAIL, ({
1469                 bch2_disk_reservation_put(c, &res);
1470                 check_extent(trans, &iter, k, &w, &s, &extent_ends) ?:
1471                 check_extent_overbig(trans, &iter, k);
1472         })) ?:
1473         check_i_sectors(trans, &w);
1474
1475         bch2_disk_reservation_put(c, &res);
1476         extent_ends_exit(&extent_ends);
1477         inode_walker_exit(&w);
1478         snapshots_seen_exit(&s);
1479         bch2_trans_put(trans);
1480
1481         bch_err_fn(c, ret);
1482         return ret;
1483 }
1484
1485 int bch2_check_indirect_extents(struct bch_fs *c)
1486 {
1487         struct btree_trans *trans = bch2_trans_get(c);
1488         struct btree_iter iter;
1489         struct bkey_s_c k;
1490         struct disk_reservation res = { 0 };
1491         int ret = 0;
1492
1493         ret = for_each_btree_key_commit(trans, iter, BTREE_ID_reflink,
1494                         POS_MIN,
1495                         BTREE_ITER_PREFETCH, k,
1496                         &res, NULL,
1497                         BTREE_INSERT_LAZY_RW|BTREE_INSERT_NOFAIL, ({
1498                 bch2_disk_reservation_put(c, &res);
1499                 check_extent_overbig(trans, &iter, k);
1500         }));
1501
1502         bch2_disk_reservation_put(c, &res);
1503         bch2_trans_put(trans);
1504
1505         bch_err_fn(c, ret);
1506         return ret;
1507 }
1508
1509 static int check_subdir_count(struct btree_trans *trans, struct inode_walker *w)
1510 {
1511         struct bch_fs *c = trans->c;
1512         struct inode_walker_entry *i;
1513         u32 restart_count = trans->restart_count;
1514         int ret = 0;
1515         s64 count2;
1516
1517         darray_for_each(w->inodes, i) {
1518                 if (i->inode.bi_nlink == i->count)
1519                         continue;
1520
1521                 count2 = bch2_count_subdirs(trans, w->last_pos.inode, i->snapshot);
1522                 if (count2 < 0)
1523                         return count2;
1524
1525                 if (i->count != count2) {
1526                         bch_err(c, "fsck counted subdirectories wrong: got %llu should be %llu",
1527                                 i->count, count2);
1528                         i->count = count2;
1529                         if (i->inode.bi_nlink == i->count)
1530                                 continue;
1531                 }
1532
1533                 if (fsck_err_on(i->inode.bi_nlink != i->count,
1534                                 c, inode_dir_wrong_nlink,
1535                                 "directory %llu:%u with wrong i_nlink: got %u, should be %llu",
1536                                 w->last_pos.inode, i->snapshot, i->inode.bi_nlink, i->count)) {
1537                         i->inode.bi_nlink = i->count;
1538                         ret = fsck_write_inode(trans, &i->inode, i->snapshot);
1539                         if (ret)
1540                                 break;
1541                 }
1542         }
1543 fsck_err:
1544         bch_err_fn(c, ret);
1545         return ret ?: trans_was_restarted(trans, restart_count);
1546 }
1547
1548 static int check_dirent_target(struct btree_trans *trans,
1549                                struct btree_iter *iter,
1550                                struct bkey_s_c_dirent d,
1551                                struct bch_inode_unpacked *target,
1552                                u32 target_snapshot)
1553 {
1554         struct bch_fs *c = trans->c;
1555         struct bkey_i_dirent *n;
1556         bool backpointer_exists = true;
1557         struct printbuf buf = PRINTBUF;
1558         int ret = 0;
1559
1560         if (!target->bi_dir &&
1561             !target->bi_dir_offset) {
1562                 target->bi_dir          = d.k->p.inode;
1563                 target->bi_dir_offset   = d.k->p.offset;
1564
1565                 ret = __write_inode(trans, target, target_snapshot);
1566                 if (ret)
1567                         goto err;
1568         }
1569
1570         if (!inode_points_to_dirent(target, d)) {
1571                 ret = inode_backpointer_exists(trans, target, d.k->p.snapshot);
1572                 if (ret < 0)
1573                         goto err;
1574
1575                 backpointer_exists = ret;
1576                 ret = 0;
1577
1578                 if (fsck_err_on(S_ISDIR(target->bi_mode) && backpointer_exists,
1579                                 c, inode_dir_multiple_links,
1580                                 "directory %llu with multiple links",
1581                                 target->bi_inum)) {
1582                         ret = __remove_dirent(trans, d.k->p);
1583                         goto out;
1584                 }
1585
1586                 if (fsck_err_on(backpointer_exists && !target->bi_nlink,
1587                                 c, inode_multiple_links_but_nlink_0,
1588                                 "inode %llu type %s has multiple links but i_nlink 0",
1589                                 target->bi_inum, bch2_d_types[d.v->d_type])) {
1590                         target->bi_nlink++;
1591                         target->bi_flags &= ~BCH_INODE_unlinked;
1592
1593                         ret = __write_inode(trans, target, target_snapshot);
1594                         if (ret)
1595                                 goto err;
1596                 }
1597
1598                 if (fsck_err_on(!backpointer_exists,
1599                                 c, inode_wrong_backpointer,
1600                                 "inode %llu:%u has wrong backpointer:\n"
1601                                 "got       %llu:%llu\n"
1602                                 "should be %llu:%llu",
1603                                 target->bi_inum, target_snapshot,
1604                                 target->bi_dir,
1605                                 target->bi_dir_offset,
1606                                 d.k->p.inode,
1607                                 d.k->p.offset)) {
1608                         target->bi_dir          = d.k->p.inode;
1609                         target->bi_dir_offset   = d.k->p.offset;
1610
1611                         ret = __write_inode(trans, target, target_snapshot);
1612                         if (ret)
1613                                 goto err;
1614                 }
1615         }
1616
1617         if (fsck_err_on(d.v->d_type != inode_d_type(target),
1618                         c, dirent_d_type_wrong,
1619                         "incorrect d_type: got %s, should be %s:\n%s",
1620                         bch2_d_type_str(d.v->d_type),
1621                         bch2_d_type_str(inode_d_type(target)),
1622                         (printbuf_reset(&buf),
1623                          bch2_bkey_val_to_text(&buf, c, d.s_c), buf.buf))) {
1624                 n = bch2_trans_kmalloc(trans, bkey_bytes(d.k));
1625                 ret = PTR_ERR_OR_ZERO(n);
1626                 if (ret)
1627                         goto err;
1628
1629                 bkey_reassemble(&n->k_i, d.s_c);
1630                 n->v.d_type = inode_d_type(target);
1631
1632                 ret = bch2_trans_update(trans, iter, &n->k_i, 0);
1633                 if (ret)
1634                         goto err;
1635
1636                 d = dirent_i_to_s_c(n);
1637         }
1638
1639         if (d.v->d_type == DT_SUBVOL &&
1640             target->bi_parent_subvol != le32_to_cpu(d.v->d_parent_subvol) &&
1641             (c->sb.version < bcachefs_metadata_version_subvol_dirent ||
1642              fsck_err(c, dirent_d_parent_subvol_wrong,
1643                       "dirent has wrong d_parent_subvol field: got %u, should be %u",
1644                       le32_to_cpu(d.v->d_parent_subvol),
1645                       target->bi_parent_subvol))) {
1646                 n = bch2_trans_kmalloc(trans, bkey_bytes(d.k));
1647                 ret = PTR_ERR_OR_ZERO(n);
1648                 if (ret)
1649                         goto err;
1650
1651                 bkey_reassemble(&n->k_i, d.s_c);
1652                 n->v.d_parent_subvol = cpu_to_le32(target->bi_parent_subvol);
1653
1654                 ret = bch2_trans_update(trans, iter, &n->k_i, 0);
1655                 if (ret)
1656                         goto err;
1657
1658                 d = dirent_i_to_s_c(n);
1659         }
1660 out:
1661 err:
1662 fsck_err:
1663         printbuf_exit(&buf);
1664         bch_err_fn(c, ret);
1665         return ret;
1666 }
1667
1668 static int check_dirent(struct btree_trans *trans, struct btree_iter *iter,
1669                         struct bkey_s_c k,
1670                         struct bch_hash_info *hash_info,
1671                         struct inode_walker *dir,
1672                         struct inode_walker *target,
1673                         struct snapshots_seen *s)
1674 {
1675         struct bch_fs *c = trans->c;
1676         struct bkey_s_c_dirent d;
1677         struct inode_walker_entry *i;
1678         struct printbuf buf = PRINTBUF;
1679         struct bpos equiv;
1680         int ret = 0;
1681
1682         ret = check_key_has_snapshot(trans, iter, k);
1683         if (ret) {
1684                 ret = ret < 0 ? ret : 0;
1685                 goto out;
1686         }
1687
1688         equiv = k.k->p;
1689         equiv.snapshot = bch2_snapshot_equiv(c, k.k->p.snapshot);
1690
1691         ret = snapshots_seen_update(c, s, iter->btree_id, k.k->p);
1692         if (ret)
1693                 goto err;
1694
1695         if (k.k->type == KEY_TYPE_whiteout)
1696                 goto out;
1697
1698         if (dir->last_pos.inode != k.k->p.inode) {
1699                 ret = check_subdir_count(trans, dir);
1700                 if (ret)
1701                         goto err;
1702         }
1703
1704         BUG_ON(!iter->path->should_be_locked);
1705
1706         i = walk_inode(trans, dir, equiv, k.k->type == KEY_TYPE_whiteout);
1707         ret = PTR_ERR_OR_ZERO(i);
1708         if (ret < 0)
1709                 goto err;
1710
1711         if (dir->first_this_inode && dir->inodes.nr)
1712                 *hash_info = bch2_hash_info_init(c, &dir->inodes.data[0].inode);
1713         dir->first_this_inode = false;
1714
1715         if (fsck_err_on(!i, c, dirent_in_missing_dir_inode,
1716                         "dirent in nonexisting directory:\n%s",
1717                         (printbuf_reset(&buf),
1718                          bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
1719                 ret = bch2_btree_delete_at(trans, iter,
1720                                 BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE);
1721                 goto out;
1722         }
1723
1724         if (!i)
1725                 goto out;
1726
1727         if (fsck_err_on(!S_ISDIR(i->inode.bi_mode),
1728                         c, dirent_in_non_dir_inode,
1729                         "dirent in non directory inode type %s:\n%s",
1730                         bch2_d_type_str(inode_d_type(&i->inode)),
1731                         (printbuf_reset(&buf),
1732                          bch2_bkey_val_to_text(&buf, c, k), buf.buf))) {
1733                 ret = bch2_btree_delete_at(trans, iter, 0);
1734                 goto out;
1735         }
1736
1737         ret = hash_check_key(trans, bch2_dirent_hash_desc, hash_info, iter, k);
1738         if (ret < 0)
1739                 goto err;
1740         if (ret) {
1741                 /* dirent has been deleted */
1742                 ret = 0;
1743                 goto out;
1744         }
1745
1746         if (k.k->type != KEY_TYPE_dirent)
1747                 goto out;
1748
1749         d = bkey_s_c_to_dirent(k);
1750
1751         if (d.v->d_type == DT_SUBVOL) {
1752                 struct bch_inode_unpacked subvol_root;
1753                 u32 target_subvol = le32_to_cpu(d.v->d_child_subvol);
1754                 u32 target_snapshot;
1755                 u64 target_inum;
1756
1757                 ret = __subvol_lookup(trans, target_subvol,
1758                                       &target_snapshot, &target_inum);
1759                 if (ret && !bch2_err_matches(ret, ENOENT))
1760                         goto err;
1761
1762                 if (fsck_err_on(ret, c, dirent_to_missing_subvol,
1763                                 "dirent points to missing subvolume %u",
1764                                 le32_to_cpu(d.v->d_child_subvol))) {
1765                         ret = __remove_dirent(trans, d.k->p);
1766                         goto err;
1767                 }
1768
1769                 ret = __lookup_inode(trans, target_inum,
1770                                    &subvol_root, &target_snapshot);
1771                 if (ret && !bch2_err_matches(ret, ENOENT))
1772                         goto err;
1773
1774                 if (fsck_err_on(ret, c, subvol_to_missing_root,
1775                                 "subvolume %u points to missing subvolume root %llu",
1776                                 target_subvol,
1777                                 target_inum)) {
1778                         bch_err(c, "repair not implemented yet");
1779                         ret = -EINVAL;
1780                         goto err;
1781                 }
1782
1783                 if (fsck_err_on(subvol_root.bi_subvol != target_subvol,
1784                                 c, subvol_root_wrong_bi_subvol,
1785                                 "subvol root %llu has wrong bi_subvol field: got %u, should be %u",
1786                                 target_inum,
1787                                 subvol_root.bi_subvol, target_subvol)) {
1788                         subvol_root.bi_subvol = target_subvol;
1789                         ret = __write_inode(trans, &subvol_root, target_snapshot);
1790                         if (ret)
1791                                 goto err;
1792                 }
1793
1794                 ret = check_dirent_target(trans, iter, d, &subvol_root,
1795                                           target_snapshot);
1796                 if (ret)
1797                         goto err;
1798         } else {
1799                 ret = __get_visible_inodes(trans, target, s, le64_to_cpu(d.v->d_inum));
1800                 if (ret)
1801                         goto err;
1802
1803                 if (fsck_err_on(!target->inodes.nr,
1804                                 c, dirent_to_missing_inode,
1805                                 "dirent points to missing inode: (equiv %u)\n%s",
1806                                 equiv.snapshot,
1807                                 (printbuf_reset(&buf),
1808                                  bch2_bkey_val_to_text(&buf, c, k),
1809                                  buf.buf))) {
1810                         ret = __remove_dirent(trans, d.k->p);
1811                         if (ret)
1812                                 goto err;
1813                 }
1814
1815                 darray_for_each(target->inodes, i) {
1816                         ret = check_dirent_target(trans, iter, d,
1817                                                   &i->inode, i->snapshot);
1818                         if (ret)
1819                                 goto err;
1820                 }
1821         }
1822
1823         if (d.v->d_type == DT_DIR)
1824                 for_each_visible_inode(c, s, dir, equiv.snapshot, i)
1825                         i->count++;
1826
1827 out:
1828 err:
1829 fsck_err:
1830         printbuf_exit(&buf);
1831         bch_err_fn(c, ret);
1832         return ret;
1833 }
1834
1835 /*
1836  * Walk dirents: verify that they all have a corresponding S_ISDIR inode,
1837  * validate d_type
1838  */
1839 int bch2_check_dirents(struct bch_fs *c)
1840 {
1841         struct inode_walker dir = inode_walker_init();
1842         struct inode_walker target = inode_walker_init();
1843         struct snapshots_seen s;
1844         struct bch_hash_info hash_info;
1845         struct btree_trans *trans = bch2_trans_get(c);
1846         struct btree_iter iter;
1847         struct bkey_s_c k;
1848         int ret = 0;
1849
1850         snapshots_seen_init(&s);
1851
1852         ret = for_each_btree_key_commit(trans, iter, BTREE_ID_dirents,
1853                         POS(BCACHEFS_ROOT_INO, 0),
1854                         BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS,
1855                         k,
1856                         NULL, NULL,
1857                         BTREE_INSERT_LAZY_RW|BTREE_INSERT_NOFAIL,
1858                 check_dirent(trans, &iter, k, &hash_info, &dir, &target, &s));
1859
1860         bch2_trans_put(trans);
1861         snapshots_seen_exit(&s);
1862         inode_walker_exit(&dir);
1863         inode_walker_exit(&target);
1864         bch_err_fn(c, ret);
1865         return ret;
1866 }
1867
1868 static int check_xattr(struct btree_trans *trans, struct btree_iter *iter,
1869                        struct bkey_s_c k,
1870                        struct bch_hash_info *hash_info,
1871                        struct inode_walker *inode)
1872 {
1873         struct bch_fs *c = trans->c;
1874         struct inode_walker_entry *i;
1875         int ret;
1876
1877         ret = check_key_has_snapshot(trans, iter, k);
1878         if (ret)
1879                 return ret;
1880
1881         i = walk_inode(trans, inode, k.k->p, k.k->type == KEY_TYPE_whiteout);
1882         ret = PTR_ERR_OR_ZERO(i);
1883         if (ret)
1884                 return ret;
1885
1886         if (inode->first_this_inode && inode->inodes.nr)
1887                 *hash_info = bch2_hash_info_init(c, &inode->inodes.data[0].inode);
1888         inode->first_this_inode = false;
1889
1890         if (fsck_err_on(!i, c, xattr_in_missing_inode,
1891                         "xattr for missing inode %llu",
1892                         k.k->p.inode))
1893                 return bch2_btree_delete_at(trans, iter, 0);
1894
1895         if (!i)
1896                 return 0;
1897
1898         ret = hash_check_key(trans, bch2_xattr_hash_desc, hash_info, iter, k);
1899 fsck_err:
1900         bch_err_fn(c, ret);
1901         return ret;
1902 }
1903
1904 /*
1905  * Walk xattrs: verify that they all have a corresponding inode
1906  */
1907 int bch2_check_xattrs(struct bch_fs *c)
1908 {
1909         struct inode_walker inode = inode_walker_init();
1910         struct bch_hash_info hash_info;
1911         struct btree_iter iter;
1912         struct bkey_s_c k;
1913         int ret = 0;
1914
1915         ret = bch2_trans_run(c,
1916                 for_each_btree_key_commit(trans, iter, BTREE_ID_xattrs,
1917                         POS(BCACHEFS_ROOT_INO, 0),
1918                         BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS,
1919                         k,
1920                         NULL, NULL,
1921                         BTREE_INSERT_LAZY_RW|BTREE_INSERT_NOFAIL,
1922                 check_xattr(trans, &iter, k, &hash_info, &inode)));
1923         bch_err_fn(c, ret);
1924         return ret;
1925 }
1926
1927 static int check_root_trans(struct btree_trans *trans)
1928 {
1929         struct bch_fs *c = trans->c;
1930         struct bch_inode_unpacked root_inode;
1931         u32 snapshot;
1932         u64 inum;
1933         int ret;
1934
1935         ret = __subvol_lookup(trans, BCACHEFS_ROOT_SUBVOL, &snapshot, &inum);
1936         if (ret && !bch2_err_matches(ret, ENOENT))
1937                 return ret;
1938
1939         if (mustfix_fsck_err_on(ret, c, root_subvol_missing,
1940                                 "root subvol missing")) {
1941                 struct bkey_i_subvolume root_subvol;
1942
1943                 snapshot        = U32_MAX;
1944                 inum            = BCACHEFS_ROOT_INO;
1945
1946                 bkey_subvolume_init(&root_subvol.k_i);
1947                 root_subvol.k.p.offset = BCACHEFS_ROOT_SUBVOL;
1948                 root_subvol.v.flags     = 0;
1949                 root_subvol.v.snapshot  = cpu_to_le32(snapshot);
1950                 root_subvol.v.inode     = cpu_to_le64(inum);
1951                 ret = commit_do(trans, NULL, NULL,
1952                                       BTREE_INSERT_NOFAIL|
1953                                       BTREE_INSERT_LAZY_RW,
1954                         bch2_btree_insert_trans(trans, BTREE_ID_subvolumes,
1955                                             &root_subvol.k_i, 0));
1956                 bch_err_msg(c, ret, "writing root subvol");
1957                 if (ret)
1958                         goto err;
1959
1960         }
1961
1962         ret = __lookup_inode(trans, BCACHEFS_ROOT_INO, &root_inode, &snapshot);
1963         if (ret && !bch2_err_matches(ret, ENOENT))
1964                 return ret;
1965
1966         if (mustfix_fsck_err_on(ret, c, root_dir_missing,
1967                                 "root directory missing") ||
1968             mustfix_fsck_err_on(!S_ISDIR(root_inode.bi_mode),
1969                                 c, root_inode_not_dir,
1970                                 "root inode not a directory")) {
1971                 bch2_inode_init(c, &root_inode, 0, 0, S_IFDIR|0755,
1972                                 0, NULL);
1973                 root_inode.bi_inum = inum;
1974
1975                 ret = __write_inode(trans, &root_inode, snapshot);
1976                 bch_err_msg(c, ret, "writing root inode");
1977         }
1978 err:
1979 fsck_err:
1980         return ret;
1981 }
1982
1983 /* Get root directory, create if it doesn't exist: */
1984 int bch2_check_root(struct bch_fs *c)
1985 {
1986         int ret;
1987
1988         ret = bch2_trans_do(c, NULL, NULL,
1989                              BTREE_INSERT_NOFAIL|
1990                              BTREE_INSERT_LAZY_RW,
1991                 check_root_trans(trans));
1992         bch_err_fn(c, ret);
1993         return ret;
1994 }
1995
1996 struct pathbuf_entry {
1997         u64     inum;
1998         u32     snapshot;
1999 };
2000
2001 typedef DARRAY(struct pathbuf_entry) pathbuf;
2002
2003 static bool path_is_dup(pathbuf *p, u64 inum, u32 snapshot)
2004 {
2005         struct pathbuf_entry *i;
2006
2007         darray_for_each(*p, i)
2008                 if (i->inum     == inum &&
2009                     i->snapshot == snapshot)
2010                         return true;
2011
2012         return false;
2013 }
2014
2015 static int path_down(struct bch_fs *c, pathbuf *p,
2016                      u64 inum, u32 snapshot)
2017 {
2018         int ret = darray_push(p, ((struct pathbuf_entry) {
2019                 .inum           = inum,
2020                 .snapshot       = snapshot,
2021         }));
2022
2023         if (ret)
2024                 bch_err(c, "fsck: error allocating memory for pathbuf, size %zu",
2025                         p->size);
2026         return ret;
2027 }
2028
2029 /*
2030  * Check that a given inode is reachable from the root:
2031  *
2032  * XXX: we should also be verifying that inodes are in the right subvolumes
2033  */
2034 static int check_path(struct btree_trans *trans,
2035                       pathbuf *p,
2036                       struct bch_inode_unpacked *inode,
2037                       u32 snapshot)
2038 {
2039         struct bch_fs *c = trans->c;
2040         int ret = 0;
2041
2042         snapshot = bch2_snapshot_equiv(c, snapshot);
2043         p->nr = 0;
2044
2045         while (!(inode->bi_inum == BCACHEFS_ROOT_INO &&
2046                  inode->bi_subvol == BCACHEFS_ROOT_SUBVOL)) {
2047                 struct btree_iter dirent_iter;
2048                 struct bkey_s_c_dirent d;
2049                 u32 parent_snapshot = snapshot;
2050
2051                 if (inode->bi_subvol) {
2052                         u64 inum;
2053
2054                         ret = subvol_lookup(trans, inode->bi_parent_subvol,
2055                                             &parent_snapshot, &inum);
2056                         if (ret)
2057                                 break;
2058                 }
2059
2060                 ret = lockrestart_do(trans,
2061                         PTR_ERR_OR_ZERO((d = dirent_get_by_pos(trans, &dirent_iter,
2062                                           SPOS(inode->bi_dir, inode->bi_dir_offset,
2063                                                parent_snapshot))).k));
2064                 if (ret && !bch2_err_matches(ret, ENOENT))
2065                         break;
2066
2067                 if (!ret && !dirent_points_to_inode(d, inode)) {
2068                         bch2_trans_iter_exit(trans, &dirent_iter);
2069                         ret = -BCH_ERR_ENOENT_dirent_doesnt_match_inode;
2070                 }
2071
2072                 if (bch2_err_matches(ret, ENOENT)) {
2073                         if (fsck_err(c,  inode_unreachable,
2074                                      "unreachable inode %llu:%u, type %s nlink %u backptr %llu:%llu",
2075                                      inode->bi_inum, snapshot,
2076                                      bch2_d_type_str(inode_d_type(inode)),
2077                                      inode->bi_nlink,
2078                                      inode->bi_dir,
2079                                      inode->bi_dir_offset))
2080                                 ret = reattach_inode(trans, inode, snapshot);
2081                         break;
2082                 }
2083
2084                 bch2_trans_iter_exit(trans, &dirent_iter);
2085
2086                 if (!S_ISDIR(inode->bi_mode))
2087                         break;
2088
2089                 ret = path_down(c, p, inode->bi_inum, snapshot);
2090                 if (ret) {
2091                         bch_err(c, "memory allocation failure");
2092                         return ret;
2093                 }
2094
2095                 snapshot = parent_snapshot;
2096
2097                 ret = lookup_inode(trans, inode->bi_dir, inode, &snapshot);
2098                 if (ret) {
2099                         /* Should have been caught in dirents pass */
2100                         bch_err(c, "error looking up parent directory: %i", ret);
2101                         break;
2102                 }
2103
2104                 if (path_is_dup(p, inode->bi_inum, snapshot)) {
2105                         struct pathbuf_entry *i;
2106
2107                         /* XXX print path */
2108                         bch_err(c, "directory structure loop");
2109
2110                         darray_for_each(*p, i)
2111                                 pr_err("%llu:%u", i->inum, i->snapshot);
2112                         pr_err("%llu:%u", inode->bi_inum, snapshot);
2113
2114                         if (!fsck_err(c, dir_loop,
2115                                       "directory structure loop"))
2116                                 return 0;
2117
2118                         ret = commit_do(trans, NULL, NULL,
2119                                               BTREE_INSERT_NOFAIL|
2120                                               BTREE_INSERT_LAZY_RW,
2121                                         remove_backpointer(trans, inode));
2122                         if (ret) {
2123                                 bch_err(c, "error removing dirent: %i", ret);
2124                                 break;
2125                         }
2126
2127                         ret = reattach_inode(trans, inode, snapshot);
2128                 }
2129         }
2130 fsck_err:
2131         bch_err_fn(c, ret);
2132         return ret;
2133 }
2134
2135 /*
2136  * Check for unreachable inodes, as well as loops in the directory structure:
2137  * After bch2_check_dirents(), if an inode backpointer doesn't exist that means it's
2138  * unreachable:
2139  */
2140 int bch2_check_directory_structure(struct bch_fs *c)
2141 {
2142         struct btree_trans *trans = bch2_trans_get(c);
2143         struct btree_iter iter;
2144         struct bkey_s_c k;
2145         struct bch_inode_unpacked u;
2146         pathbuf path = { 0, };
2147         int ret;
2148
2149         for_each_btree_key(trans, iter, BTREE_ID_inodes, POS_MIN,
2150                            BTREE_ITER_INTENT|
2151                            BTREE_ITER_PREFETCH|
2152                            BTREE_ITER_ALL_SNAPSHOTS, k, ret) {
2153                 if (!bkey_is_inode(k.k))
2154                         continue;
2155
2156                 ret = bch2_inode_unpack(k, &u);
2157                 if (ret) {
2158                         /* Should have been caught earlier in fsck: */
2159                         bch_err(c, "error unpacking inode %llu: %i", k.k->p.offset, ret);
2160                         break;
2161                 }
2162
2163                 if (u.bi_flags & BCH_INODE_unlinked)
2164                         continue;
2165
2166                 ret = check_path(trans, &path, &u, iter.pos.snapshot);
2167                 if (ret)
2168                         break;
2169         }
2170         bch2_trans_iter_exit(trans, &iter);
2171         bch2_trans_put(trans);
2172         darray_exit(&path);
2173         bch_err_fn(c, ret);
2174         return ret;
2175 }
2176
2177 struct nlink_table {
2178         size_t          nr;
2179         size_t          size;
2180
2181         struct nlink {
2182                 u64     inum;
2183                 u32     snapshot;
2184                 u32     count;
2185         }               *d;
2186 };
2187
2188 static int add_nlink(struct bch_fs *c, struct nlink_table *t,
2189                      u64 inum, u32 snapshot)
2190 {
2191         if (t->nr == t->size) {
2192                 size_t new_size = max_t(size_t, 128UL, t->size * 2);
2193                 void *d = kvmalloc_array(new_size, sizeof(t->d[0]), GFP_KERNEL);
2194
2195                 if (!d) {
2196                         bch_err(c, "fsck: error allocating memory for nlink_table, size %zu",
2197                                 new_size);
2198                         return -BCH_ERR_ENOMEM_fsck_add_nlink;
2199                 }
2200
2201                 if (t->d)
2202                         memcpy(d, t->d, t->size * sizeof(t->d[0]));
2203                 kvfree(t->d);
2204
2205                 t->d = d;
2206                 t->size = new_size;
2207         }
2208
2209
2210         t->d[t->nr++] = (struct nlink) {
2211                 .inum           = inum,
2212                 .snapshot       = snapshot,
2213         };
2214
2215         return 0;
2216 }
2217
2218 static int nlink_cmp(const void *_l, const void *_r)
2219 {
2220         const struct nlink *l = _l;
2221         const struct nlink *r = _r;
2222
2223         return cmp_int(l->inum, r->inum) ?: cmp_int(l->snapshot, r->snapshot);
2224 }
2225
2226 static void inc_link(struct bch_fs *c, struct snapshots_seen *s,
2227                      struct nlink_table *links,
2228                      u64 range_start, u64 range_end, u64 inum, u32 snapshot)
2229 {
2230         struct nlink *link, key = {
2231                 .inum = inum, .snapshot = U32_MAX,
2232         };
2233
2234         if (inum < range_start || inum >= range_end)
2235                 return;
2236
2237         link = __inline_bsearch(&key, links->d, links->nr,
2238                                 sizeof(links->d[0]), nlink_cmp);
2239         if (!link)
2240                 return;
2241
2242         while (link > links->d && link[0].inum == link[-1].inum)
2243                 --link;
2244
2245         for (; link < links->d + links->nr && link->inum == inum; link++)
2246                 if (ref_visible(c, s, snapshot, link->snapshot)) {
2247                         link->count++;
2248                         if (link->snapshot >= snapshot)
2249                                 break;
2250                 }
2251 }
2252
2253 noinline_for_stack
2254 static int check_nlinks_find_hardlinks(struct bch_fs *c,
2255                                        struct nlink_table *t,
2256                                        u64 start, u64 *end)
2257 {
2258         struct btree_trans *trans = bch2_trans_get(c);
2259         struct btree_iter iter;
2260         struct bkey_s_c k;
2261         struct bch_inode_unpacked u;
2262         int ret = 0;
2263
2264         for_each_btree_key(trans, iter, BTREE_ID_inodes,
2265                            POS(0, start),
2266                            BTREE_ITER_INTENT|
2267                            BTREE_ITER_PREFETCH|
2268                            BTREE_ITER_ALL_SNAPSHOTS, k, ret) {
2269                 if (!bkey_is_inode(k.k))
2270                         continue;
2271
2272                 /* Should never fail, checked by bch2_inode_invalid: */
2273                 BUG_ON(bch2_inode_unpack(k, &u));
2274
2275                 /*
2276                  * Backpointer and directory structure checks are sufficient for
2277                  * directories, since they can't have hardlinks:
2278                  */
2279                 if (S_ISDIR(u.bi_mode))
2280                         continue;
2281
2282                 if (!u.bi_nlink)
2283                         continue;
2284
2285                 ret = add_nlink(c, t, k.k->p.offset, k.k->p.snapshot);
2286                 if (ret) {
2287                         *end = k.k->p.offset;
2288                         ret = 0;
2289                         break;
2290                 }
2291
2292         }
2293         bch2_trans_iter_exit(trans, &iter);
2294         bch2_trans_put(trans);
2295
2296         if (ret)
2297                 bch_err(c, "error in fsck: btree error %i while walking inodes", ret);
2298
2299         return ret;
2300 }
2301
2302 noinline_for_stack
2303 static int check_nlinks_walk_dirents(struct bch_fs *c, struct nlink_table *links,
2304                                      u64 range_start, u64 range_end)
2305 {
2306         struct btree_trans *trans = bch2_trans_get(c);
2307         struct snapshots_seen s;
2308         struct btree_iter iter;
2309         struct bkey_s_c k;
2310         struct bkey_s_c_dirent d;
2311         int ret;
2312
2313         snapshots_seen_init(&s);
2314
2315         for_each_btree_key(trans, iter, BTREE_ID_dirents, POS_MIN,
2316                            BTREE_ITER_INTENT|
2317                            BTREE_ITER_PREFETCH|
2318                            BTREE_ITER_ALL_SNAPSHOTS, k, ret) {
2319                 ret = snapshots_seen_update(c, &s, iter.btree_id, k.k->p);
2320                 if (ret)
2321                         break;
2322
2323                 switch (k.k->type) {
2324                 case KEY_TYPE_dirent:
2325                         d = bkey_s_c_to_dirent(k);
2326
2327                         if (d.v->d_type != DT_DIR &&
2328                             d.v->d_type != DT_SUBVOL)
2329                                 inc_link(c, &s, links, range_start, range_end,
2330                                          le64_to_cpu(d.v->d_inum),
2331                                          bch2_snapshot_equiv(c, d.k->p.snapshot));
2332                         break;
2333                 }
2334         }
2335         bch2_trans_iter_exit(trans, &iter);
2336
2337         if (ret)
2338                 bch_err(c, "error in fsck: btree error %i while walking dirents", ret);
2339
2340         bch2_trans_put(trans);
2341         snapshots_seen_exit(&s);
2342         return ret;
2343 }
2344
2345 static int check_nlinks_update_inode(struct btree_trans *trans, struct btree_iter *iter,
2346                                      struct bkey_s_c k,
2347                                      struct nlink_table *links,
2348                                      size_t *idx, u64 range_end)
2349 {
2350         struct bch_fs *c = trans->c;
2351         struct bch_inode_unpacked u;
2352         struct nlink *link = &links->d[*idx];
2353         int ret = 0;
2354
2355         if (k.k->p.offset >= range_end)
2356                 return 1;
2357
2358         if (!bkey_is_inode(k.k))
2359                 return 0;
2360
2361         BUG_ON(bch2_inode_unpack(k, &u));
2362
2363         if (S_ISDIR(u.bi_mode))
2364                 return 0;
2365
2366         if (!u.bi_nlink)
2367                 return 0;
2368
2369         while ((cmp_int(link->inum, k.k->p.offset) ?:
2370                 cmp_int(link->snapshot, k.k->p.snapshot)) < 0) {
2371                 BUG_ON(*idx == links->nr);
2372                 link = &links->d[++*idx];
2373         }
2374
2375         if (fsck_err_on(bch2_inode_nlink_get(&u) != link->count,
2376                         c, inode_wrong_nlink,
2377                         "inode %llu type %s has wrong i_nlink (%u, should be %u)",
2378                         u.bi_inum, bch2_d_types[mode_to_type(u.bi_mode)],
2379                         bch2_inode_nlink_get(&u), link->count)) {
2380                 bch2_inode_nlink_set(&u, link->count);
2381                 ret = __write_inode(trans, &u, k.k->p.snapshot);
2382         }
2383 fsck_err:
2384         return ret;
2385 }
2386
2387 noinline_for_stack
2388 static int check_nlinks_update_hardlinks(struct bch_fs *c,
2389                                struct nlink_table *links,
2390                                u64 range_start, u64 range_end)
2391 {
2392         struct btree_iter iter;
2393         struct bkey_s_c k;
2394         size_t idx = 0;
2395         int ret = 0;
2396
2397         ret = bch2_trans_run(c,
2398                 for_each_btree_key_commit(trans, iter, BTREE_ID_inodes,
2399                                 POS(0, range_start),
2400                                 BTREE_ITER_INTENT|BTREE_ITER_PREFETCH|BTREE_ITER_ALL_SNAPSHOTS, k,
2401                                 NULL, NULL, BTREE_INSERT_LAZY_RW|BTREE_INSERT_NOFAIL,
2402                         check_nlinks_update_inode(trans, &iter, k, links, &idx, range_end)));
2403         if (ret < 0) {
2404                 bch_err(c, "error in fsck: btree error %i while walking inodes", ret);
2405                 return ret;
2406         }
2407
2408         return 0;
2409 }
2410
2411 int bch2_check_nlinks(struct bch_fs *c)
2412 {
2413         struct nlink_table links = { 0 };
2414         u64 this_iter_range_start, next_iter_range_start = 0;
2415         int ret = 0;
2416
2417         do {
2418                 this_iter_range_start = next_iter_range_start;
2419                 next_iter_range_start = U64_MAX;
2420
2421                 ret = check_nlinks_find_hardlinks(c, &links,
2422                                                   this_iter_range_start,
2423                                                   &next_iter_range_start);
2424
2425                 ret = check_nlinks_walk_dirents(c, &links,
2426                                           this_iter_range_start,
2427                                           next_iter_range_start);
2428                 if (ret)
2429                         break;
2430
2431                 ret = check_nlinks_update_hardlinks(c, &links,
2432                                          this_iter_range_start,
2433                                          next_iter_range_start);
2434                 if (ret)
2435                         break;
2436
2437                 links.nr = 0;
2438         } while (next_iter_range_start != U64_MAX);
2439
2440         kvfree(links.d);
2441         bch_err_fn(c, ret);
2442         return ret;
2443 }
2444
2445 static int fix_reflink_p_key(struct btree_trans *trans, struct btree_iter *iter,
2446                              struct bkey_s_c k)
2447 {
2448         struct bkey_s_c_reflink_p p;
2449         struct bkey_i_reflink_p *u;
2450         int ret;
2451
2452         if (k.k->type != KEY_TYPE_reflink_p)
2453                 return 0;
2454
2455         p = bkey_s_c_to_reflink_p(k);
2456
2457         if (!p.v->front_pad && !p.v->back_pad)
2458                 return 0;
2459
2460         u = bch2_trans_kmalloc(trans, sizeof(*u));
2461         ret = PTR_ERR_OR_ZERO(u);
2462         if (ret)
2463                 return ret;
2464
2465         bkey_reassemble(&u->k_i, k);
2466         u->v.front_pad  = 0;
2467         u->v.back_pad   = 0;
2468
2469         return bch2_trans_update(trans, iter, &u->k_i, BTREE_TRIGGER_NORUN);
2470 }
2471
2472 int bch2_fix_reflink_p(struct bch_fs *c)
2473 {
2474         struct btree_iter iter;
2475         struct bkey_s_c k;
2476         int ret;
2477
2478         if (c->sb.version >= bcachefs_metadata_version_reflink_p_fix)
2479                 return 0;
2480
2481         ret = bch2_trans_run(c,
2482                 for_each_btree_key_commit(trans, iter,
2483                                 BTREE_ID_extents, POS_MIN,
2484                                 BTREE_ITER_INTENT|BTREE_ITER_PREFETCH|
2485                                 BTREE_ITER_ALL_SNAPSHOTS, k,
2486                                 NULL, NULL, BTREE_INSERT_NOFAIL|BTREE_INSERT_LAZY_RW,
2487                         fix_reflink_p_key(trans, &iter, k)));
2488         bch_err_fn(c, ret);
2489         return ret;
2490 }