git.sesse.net Git - bcachefs-tools-debian/blob - libbcachefs/dirent.c

   1 // SPDX-License-Identifier: GPL-2.0
   2
   3 #include "bcachefs.h"
   4 #include "bkey_buf.h"
   5 #include "bkey_methods.h"
   6 #include "btree_update.h"
   7 #include "extents.h"
   8 #include "dirent.h"
   9 #include "fs.h"
  10 #include "keylist.h"
  11 #include "str_hash.h"
  12 #include "subvolume.h"
  13
  14 #include <linux/dcache.h>
  15
  16 static unsigned bch2_dirent_name_bytes(struct bkey_s_c_dirent d)
  17 {
  18         unsigned bkey_u64s = bkey_val_u64s(d.k);
  19         unsigned bkey_bytes = bkey_u64s * sizeof(u64);
  20         u64 last_u64 = ((u64*)d.v)[bkey_u64s - 1];
  21 #if CPU_BIG_ENDIAN
  22         unsigned trailing_nuls = last_u64 ? __builtin_ctzll(last_u64) / 8 : 64 / 8;
  23 #else
  24         unsigned trailing_nuls = last_u64 ? __builtin_clzll(last_u64) / 8 : 64 / 8;
  25 #endif
  26
  27         return bkey_bytes -
  28                 offsetof(struct bch_dirent, d_name) -
  29                 trailing_nuls;
  30 }
  31
  32 struct qstr bch2_dirent_get_name(struct bkey_s_c_dirent d)
  33 {
  34         return (struct qstr) QSTR_INIT(d.v->d_name, bch2_dirent_name_bytes(d));
  35 }
  36
  37 static u64 bch2_dirent_hash(const struct bch_hash_info *info,
  38                             const struct qstr *name)
  39 {
  40         struct bch_str_hash_ctx ctx;
  41
  42         bch2_str_hash_init(&ctx, info);
  43         bch2_str_hash_update(&ctx, info, name->name, name->len);
  44
  45         /* [0,2) reserved for dots */
  46         return max_t(u64, bch2_str_hash_end(&ctx, info), 2);
  47 }
  48
  49 static u64 dirent_hash_key(const struct bch_hash_info *info, const void *key)
  50 {
  51         return bch2_dirent_hash(info, key);
  52 }
  53
  54 static u64 dirent_hash_bkey(const struct bch_hash_info *info, struct bkey_s_c k)
  55 {
  56         struct bkey_s_c_dirent d = bkey_s_c_to_dirent(k);
  57         struct qstr name = bch2_dirent_get_name(d);
  58
  59         return bch2_dirent_hash(info, &name);
  60 }
  61
  62 static bool dirent_cmp_key(struct bkey_s_c _l, const void *_r)
  63 {
  64         struct bkey_s_c_dirent l = bkey_s_c_to_dirent(_l);
  65         const struct qstr l_name = bch2_dirent_get_name(l);
  66         const struct qstr *r_name = _r;
  67
  68         return l_name.len - r_name->len ?: memcmp(l_name.name, r_name->name, l_name.len);
  69 }
  70
  71 static bool dirent_cmp_bkey(struct bkey_s_c _l, struct bkey_s_c _r)
  72 {
  73         struct bkey_s_c_dirent l = bkey_s_c_to_dirent(_l);
  74         struct bkey_s_c_dirent r = bkey_s_c_to_dirent(_r);
  75         const struct qstr l_name = bch2_dirent_get_name(l);
  76         const struct qstr r_name = bch2_dirent_get_name(r);
  77
  78         return l_name.len - r_name.len ?: memcmp(l_name.name, r_name.name, l_name.len);
  79 }
  80
  81 static bool dirent_is_visible(subvol_inum inum, struct bkey_s_c k)
  82 {
  83         struct bkey_s_c_dirent d = bkey_s_c_to_dirent(k);
  84
  85         if (d.v->d_type == DT_SUBVOL)
  86                 return le32_to_cpu(d.v->d_parent_subvol) == inum.subvol;
  87         return true;
  88 }
  89
  90 const struct bch_hash_desc bch2_dirent_hash_desc = {
  91         .btree_id       = BTREE_ID_dirents,
  92         .key_type       = KEY_TYPE_dirent,
  93         .hash_key       = dirent_hash_key,
  94         .hash_bkey      = dirent_hash_bkey,
  95         .cmp_key        = dirent_cmp_key,
  96         .cmp_bkey       = dirent_cmp_bkey,
  97         .is_visible     = dirent_is_visible,
  98 };
  99
 100 int bch2_dirent_invalid(struct bch_fs *c, struct bkey_s_c k,
 101                         enum bkey_invalid_flags flags,
 102                         struct printbuf *err)
 103 {
 104         struct bkey_s_c_dirent d = bkey_s_c_to_dirent(k);
 105         struct qstr d_name = bch2_dirent_get_name(d);
 106         int ret = 0;
 107
 108         bkey_fsck_err_on(!d_name.len, c, err,
 109                          dirent_empty_name,
 110                          "empty name");
 111
 112         bkey_fsck_err_on(bkey_val_u64s(k.k) > dirent_val_u64s(d_name.len), c, err,
 113                          dirent_val_too_big,
 114                          "value too big (%zu > %u)",
 115                          bkey_val_u64s(k.k), dirent_val_u64s(d_name.len));
 116
 117         /*
 118          * Check new keys don't exceed the max length
 119          * (older keys may be larger.)
 120          */
 121         bkey_fsck_err_on((flags & BKEY_INVALID_COMMIT) && d_name.len > BCH_NAME_MAX, c, err,
 122                          dirent_name_too_long,
 123                          "dirent name too big (%u > %u)",
 124                          d_name.len, BCH_NAME_MAX);
 125
 126         bkey_fsck_err_on(d_name.len != strnlen(d_name.name, d_name.len), c, err,
 127                          dirent_name_embedded_nul,
 128                          "dirent has stray data after name's NUL");
 129
 130         bkey_fsck_err_on((d_name.len == 1 && !memcmp(d_name.name, ".", 1)) ||
 131                          (d_name.len == 2 && !memcmp(d_name.name, "..", 2)), c, err,
 132                          dirent_name_dot_or_dotdot,
 133                          "invalid name");
 134
 135         bkey_fsck_err_on(memchr(d_name.name, '/', d_name.len), c, err,
 136                          dirent_name_has_slash,
 137                          "name with /");
 138
 139         bkey_fsck_err_on(d.v->d_type != DT_SUBVOL &&
 140                          le64_to_cpu(d.v->d_inum) == d.k->p.inode, c, err,
 141                          dirent_to_itself,
 142                          "dirent points to own directory");
 143 fsck_err:
 144         return ret;
 145 }
 146
 147 void bch2_dirent_to_text(struct printbuf *out, struct bch_fs *c,
 148                          struct bkey_s_c k)
 149 {
 150         struct bkey_s_c_dirent d = bkey_s_c_to_dirent(k);
 151         struct qstr d_name = bch2_dirent_get_name(d);
 152
 153         prt_printf(out, "%.*s -> %llu type %s",
 154                d_name.len,
 155                d_name.name,
 156                d.v->d_type != DT_SUBVOL
 157                ? le64_to_cpu(d.v->d_inum)
 158                : le32_to_cpu(d.v->d_child_subvol),
 159                bch2_d_type_str(d.v->d_type));
 160 }
 161
 162 static struct bkey_i_dirent *dirent_create_key(struct btree_trans *trans,
 163                                 subvol_inum dir, u8 type,
 164                                 const struct qstr *name, u64 dst)
 165 {
 166         struct bkey_i_dirent *dirent;
 167         unsigned u64s = BKEY_U64s + dirent_val_u64s(name->len);
 168
 169         if (name->len > BCH_NAME_MAX)
 170                 return ERR_PTR(-ENAMETOOLONG);
 171
 172         BUG_ON(u64s > U8_MAX);
 173
 174         dirent = bch2_trans_kmalloc(trans, u64s * sizeof(u64));
 175         if (IS_ERR(dirent))
 176                 return dirent;
 177
 178         bkey_dirent_init(&dirent->k_i);
 179         dirent->k.u64s = u64s;
 180
 181         if (type != DT_SUBVOL) {
 182                 dirent->v.d_inum = cpu_to_le64(dst);
 183         } else {
 184                 dirent->v.d_parent_subvol = cpu_to_le32(dir.subvol);
 185                 dirent->v.d_child_subvol = cpu_to_le32(dst);
 186         }
 187
 188         dirent->v.d_type = type;
 189
 190         memcpy(dirent->v.d_name, name->name, name->len);
 191         memset(dirent->v.d_name + name->len, 0,
 192                bkey_val_bytes(&dirent->k) -
 193                offsetof(struct bch_dirent, d_name) -
 194                name->len);
 195
 196         EBUG_ON(bch2_dirent_name_bytes(dirent_i_to_s_c(dirent)) != name->len);
 197
 198         return dirent;
 199 }
 200
 201 int bch2_dirent_create(struct btree_trans *trans, subvol_inum dir,
 202                        const struct bch_hash_info *hash_info,
 203                        u8 type, const struct qstr *name, u64 dst_inum,
 204                        u64 *dir_offset, int flags)
 205 {
 206         struct bkey_i_dirent *dirent;
 207         int ret;
 208
 209         dirent = dirent_create_key(trans, dir, type, name, dst_inum);
 210         ret = PTR_ERR_OR_ZERO(dirent);
 211         if (ret)
 212                 return ret;
 213
 214         ret = bch2_hash_set(trans, bch2_dirent_hash_desc, hash_info,
 215                             dir, &dirent->k_i, flags);
 216         *dir_offset = dirent->k.p.offset;
 217
 218         return ret;
 219 }
 220
 221 static void dirent_copy_target(struct bkey_i_dirent *dst,
 222                                struct bkey_s_c_dirent src)
 223 {
 224         dst->v.d_inum = src.v->d_inum;
 225         dst->v.d_type = src.v->d_type;
 226 }
 227
 228 int bch2_dirent_read_target(struct btree_trans *trans, subvol_inum dir,
 229                             struct bkey_s_c_dirent d, subvol_inum *target)
 230 {
 231         struct bch_subvolume s;
 232         int ret = 0;
 233
 234         if (d.v->d_type == DT_SUBVOL &&
 235             le32_to_cpu(d.v->d_parent_subvol) != dir.subvol)
 236                 return 1;
 237
 238         if (likely(d.v->d_type != DT_SUBVOL)) {
 239                 target->subvol  = dir.subvol;
 240                 target->inum    = le64_to_cpu(d.v->d_inum);
 241         } else {
 242                 target->subvol  = le32_to_cpu(d.v->d_child_subvol);
 243
 244                 ret = bch2_subvolume_get(trans, target->subvol, true, BTREE_ITER_CACHED, &s);
 245
 246                 target->inum    = le64_to_cpu(s.inode);
 247         }
 248
 249         return ret;
 250 }
 251
 252 int bch2_dirent_rename(struct btree_trans *trans,
 253                 subvol_inum src_dir, struct bch_hash_info *src_hash,
 254                 subvol_inum dst_dir, struct bch_hash_info *dst_hash,
 255                 const struct qstr *src_name, subvol_inum *src_inum, u64 *src_offset,
 256                 const struct qstr *dst_name, subvol_inum *dst_inum, u64 *dst_offset,
 257                 enum bch_rename_mode mode)
 258 {
 259         struct btree_iter src_iter = { NULL };
 260         struct btree_iter dst_iter = { NULL };
 261         struct bkey_s_c old_src, old_dst = bkey_s_c_null;
 262         struct bkey_i_dirent *new_src = NULL, *new_dst = NULL;
 263         struct bpos dst_pos =
 264                 POS(dst_dir.inum, bch2_dirent_hash(dst_hash, dst_name));
 265         unsigned src_type = 0, dst_type = 0, src_update_flags = 0;
 266         int ret = 0;
 267
 268         if (src_dir.subvol != dst_dir.subvol)
 269                 return -EXDEV;
 270
 271         memset(src_inum, 0, sizeof(*src_inum));
 272         memset(dst_inum, 0, sizeof(*dst_inum));
 273
 274         /* Lookup src: */
 275         ret = bch2_hash_lookup(trans, &src_iter, bch2_dirent_hash_desc,
 276                                src_hash, src_dir, src_name,
 277                                BTREE_ITER_INTENT);
 278         if (ret)
 279                 goto out;
 280
 281         old_src = bch2_btree_iter_peek_slot(&src_iter);
 282         ret = bkey_err(old_src);
 283         if (ret)
 284                 goto out;
 285
 286         ret = bch2_dirent_read_target(trans, src_dir,
 287                         bkey_s_c_to_dirent(old_src), src_inum);
 288         if (ret)
 289                 goto out;
 290
 291         src_type = bkey_s_c_to_dirent(old_src).v->d_type;
 292
 293         if (src_type == DT_SUBVOL && mode == BCH_RENAME_EXCHANGE)
 294                 return -EOPNOTSUPP;
 295
 296
 297         /* Lookup dst: */
 298         if (mode == BCH_RENAME) {
 299                 /*
 300                  * Note that we're _not_ checking if the target already exists -
 301                  * we're relying on the VFS to do that check for us for
 302                  * correctness:
 303                  */
 304                 ret = bch2_hash_hole(trans, &dst_iter, bch2_dirent_hash_desc,
 305                                      dst_hash, dst_dir, dst_name);
 306                 if (ret)
 307                         goto out;
 308         } else {
 309                 ret = bch2_hash_lookup(trans, &dst_iter, bch2_dirent_hash_desc,
 310                                        dst_hash, dst_dir, dst_name,
 311                                        BTREE_ITER_INTENT);
 312                 if (ret)
 313                         goto out;
 314
 315                 old_dst = bch2_btree_iter_peek_slot(&dst_iter);
 316                 ret = bkey_err(old_dst);
 317                 if (ret)
 318                         goto out;
 319
 320                 ret = bch2_dirent_read_target(trans, dst_dir,
 321                                 bkey_s_c_to_dirent(old_dst), dst_inum);
 322                 if (ret)
 323                         goto out;
 324
 325                 dst_type = bkey_s_c_to_dirent(old_dst).v->d_type;
 326
 327                 if (dst_type == DT_SUBVOL)
 328                         return -EOPNOTSUPP;
 329         }
 330
 331         if (mode != BCH_RENAME_EXCHANGE)
 332                 *src_offset = dst_iter.pos.offset;
 333
 334         /* Create new dst key: */
 335         new_dst = dirent_create_key(trans, dst_dir, 0, dst_name, 0);
 336         ret = PTR_ERR_OR_ZERO(new_dst);
 337         if (ret)
 338                 goto out;
 339
 340         dirent_copy_target(new_dst, bkey_s_c_to_dirent(old_src));
 341         new_dst->k.p = dst_iter.pos;
 342
 343         /* Create new src key: */
 344         if (mode == BCH_RENAME_EXCHANGE) {
 345                 new_src = dirent_create_key(trans, src_dir, 0, src_name, 0);
 346                 ret = PTR_ERR_OR_ZERO(new_src);
 347                 if (ret)
 348                         goto out;
 349
 350                 dirent_copy_target(new_src, bkey_s_c_to_dirent(old_dst));
 351                 new_src->k.p = src_iter.pos;
 352         } else {
 353                 new_src = bch2_trans_kmalloc(trans, sizeof(struct bkey_i));
 354                 ret = PTR_ERR_OR_ZERO(new_src);
 355                 if (ret)
 356                         goto out;
 357
 358                 bkey_init(&new_src->k);
 359                 new_src->k.p = src_iter.pos;
 360
 361                 if (bkey_le(dst_pos, src_iter.pos) &&
 362                     bkey_lt(src_iter.pos, dst_iter.pos)) {
 363                         /*
 364                          * We have a hash collision for the new dst key,
 365                          * and new_src - the key we're deleting - is between
 366                          * new_dst's hashed slot and the slot we're going to be
 367                          * inserting it into - oops.  This will break the hash
 368                          * table if we don't deal with it:
 369                          */
 370                         if (mode == BCH_RENAME) {
 371                                 /*
 372                                  * If we're not overwriting, we can just insert
 373                                  * new_dst at the src position:
 374                                  */
 375                                 new_src = new_dst;
 376                                 new_src->k.p = src_iter.pos;
 377                                 goto out_set_src;
 378                         } else {
 379                                 /* If we're overwriting, we can't insert new_dst
 380                                  * at a different slot because it has to
 381                                  * overwrite old_dst - just make sure to use a
 382                                  * whiteout when deleting src:
 383                                  */
 384                                 new_src->k.type = KEY_TYPE_hash_whiteout;
 385                         }
 386                 } else {
 387                         /* Check if we need a whiteout to delete src: */
 388                         ret = bch2_hash_needs_whiteout(trans, bch2_dirent_hash_desc,
 389                                                        src_hash, &src_iter);
 390                         if (ret < 0)
 391                                 goto out;
 392
 393                         if (ret)
 394                                 new_src->k.type = KEY_TYPE_hash_whiteout;
 395                 }
 396         }
 397
 398         ret = bch2_trans_update(trans, &dst_iter, &new_dst->k_i, 0);
 399         if (ret)
 400                 goto out;
 401 out_set_src:
 402
 403         /*
 404          * If we're deleting a subvolume, we need to really delete the dirent,
 405          * not just emit a whiteout in the current snapshot:
 406          */
 407         if (src_type == DT_SUBVOL) {
 408                 bch2_btree_iter_set_snapshot(&src_iter, old_src.k->p.snapshot);
 409                 ret = bch2_btree_iter_traverse(&src_iter);
 410                 if (ret)
 411                         goto out;
 412
 413                 new_src->k.p = src_iter.pos;
 414                 src_update_flags |= BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE;
 415         }
 416
 417         ret = bch2_trans_update(trans, &src_iter, &new_src->k_i, src_update_flags);
 418         if (ret)
 419                 goto out;
 420
 421         if (mode == BCH_RENAME_EXCHANGE)
 422                 *src_offset = new_src->k.p.offset;
 423         *dst_offset = new_dst->k.p.offset;
 424 out:
 425         bch2_trans_iter_exit(trans, &src_iter);
 426         bch2_trans_iter_exit(trans, &dst_iter);
 427         return ret;
 428 }
 429
 430 int __bch2_dirent_lookup_trans(struct btree_trans *trans,
 431                                struct btree_iter *iter,
 432                                subvol_inum dir,
 433                                const struct bch_hash_info *hash_info,
 434                                const struct qstr *name, subvol_inum *inum,
 435                                unsigned flags)
 436 {
 437         struct bkey_s_c k;
 438         struct bkey_s_c_dirent d;
 439         u32 snapshot;
 440         int ret;
 441
 442         ret = bch2_subvolume_get_snapshot(trans, dir.subvol, &snapshot);
 443         if (ret)
 444                 return ret;
 445
 446         ret = bch2_hash_lookup(trans, iter, bch2_dirent_hash_desc,
 447                                hash_info, dir, name, flags);
 448         if (ret)
 449                 return ret;
 450
 451         k = bch2_btree_iter_peek_slot(iter);
 452         ret = bkey_err(k);
 453         if (ret)
 454                 goto err;
 455
 456         d = bkey_s_c_to_dirent(k);
 457
 458         ret = bch2_dirent_read_target(trans, dir, d, inum);
 459         if (ret > 0)
 460                 ret = -ENOENT;
 461 err:
 462         if (ret)
 463                 bch2_trans_iter_exit(trans, iter);
 464
 465         return ret;
 466 }
 467
 468 u64 bch2_dirent_lookup(struct bch_fs *c, subvol_inum dir,
 469                        const struct bch_hash_info *hash_info,
 470                        const struct qstr *name, subvol_inum *inum)
 471 {
 472         struct btree_trans *trans = bch2_trans_get(c);
 473         struct btree_iter iter;
 474         int ret;
 475 retry:
 476         bch2_trans_begin(trans);
 477
 478         ret = __bch2_dirent_lookup_trans(trans, &iter, dir, hash_info,
 479                                           name, inum, 0);
 480         if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
 481                 goto retry;
 482         if (!ret)
 483                 bch2_trans_iter_exit(trans, &iter);
 484         bch2_trans_put(trans);
 485         return ret;
 486 }
 487
 488 int bch2_empty_dir_trans(struct btree_trans *trans, subvol_inum dir)
 489 {
 490         struct btree_iter iter;
 491         struct bkey_s_c k;
 492         u32 snapshot;
 493         int ret;
 494
 495         ret = bch2_subvolume_get_snapshot(trans, dir.subvol, &snapshot);
 496         if (ret)
 497                 return ret;
 498
 499         for_each_btree_key_upto_norestart(trans, iter, BTREE_ID_dirents,
 500                            SPOS(dir.inum, 0, snapshot),
 501                            POS(dir.inum, U64_MAX), 0, k, ret)
 502                 if (k.k->type == KEY_TYPE_dirent) {
 503                         ret = -ENOTEMPTY;
 504                         break;
 505                 }
 506         bch2_trans_iter_exit(trans, &iter);
 507
 508         return ret;
 509 }
 510
 511 int bch2_readdir(struct bch_fs *c, subvol_inum inum, struct dir_context *ctx)
 512 {
 513         struct btree_trans *trans = bch2_trans_get(c);
 514         struct btree_iter iter;
 515         struct bkey_s_c k;
 516         struct bkey_s_c_dirent dirent;
 517         subvol_inum target;
 518         u32 snapshot;
 519         struct bkey_buf sk;
 520         struct qstr name;
 521         int ret;
 522
 523         bch2_bkey_buf_init(&sk);
 524 retry:
 525         bch2_trans_begin(trans);
 526
 527         ret = bch2_subvolume_get_snapshot(trans, inum.subvol, &snapshot);
 528         if (ret)
 529                 goto err;
 530
 531         for_each_btree_key_upto_norestart(trans, iter, BTREE_ID_dirents,
 532                            SPOS(inum.inum, ctx->pos, snapshot),
 533                            POS(inum.inum, U64_MAX), 0, k, ret) {
 534                 if (k.k->type != KEY_TYPE_dirent)
 535                         continue;
 536
 537                 dirent = bkey_s_c_to_dirent(k);
 538
 539                 ret = bch2_dirent_read_target(trans, inum, dirent, &target);
 540                 if (ret < 0)
 541                         break;
 542                 if (ret)
 543                         continue;
 544
 545                 /* dir_emit() can fault and block: */
 546                 bch2_bkey_buf_reassemble(&sk, c, k);
 547                 dirent = bkey_i_to_s_c_dirent(sk.k);
 548                 bch2_trans_unlock(trans);
 549
 550                 name = bch2_dirent_get_name(dirent);
 551
 552                 ctx->pos = dirent.k->p.offset;
 553                 if (!dir_emit(ctx, name.name,
 554                               name.len,
 555                               target.inum,
 556                               vfs_d_type(dirent.v->d_type)))
 557                         break;
 558                 ctx->pos = dirent.k->p.offset + 1;
 559
 560                 /*
 561                  * read_target looks up subvolumes, we can overflow paths if the
 562                  * directory has many subvolumes in it
 563                  */
 564                 ret = btree_trans_too_many_iters(trans);
 565                 if (ret)
 566                         break;
 567         }
 568         bch2_trans_iter_exit(trans, &iter);
 569 err:
 570         if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
 571                 goto retry;
 572
 573         bch2_trans_put(trans);
 574         bch2_bkey_buf_exit(&sk, c);
 575
 576         return ret;
 577 }