fs/btrfs/backref.c

   1 /*
   2  * Copyright (C) 2011 STRATO.  All rights reserved.
   3  *
   4  * This program is free software; you can redistribute it and/or
   5  * modify it under the terms of the GNU General Public
   6  * License v2 as published by the Free Software Foundation.
   7  *
   8  * This program is distributed in the hope that it will be useful,
   9  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  11  * General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU General Public
  14  * License along with this program; if not, write to the
  15  * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  16  * Boston, MA 021110-1307, USA.
  17  */
  18
  19 #include <linux/vmalloc.h>
  20 #include "ctree.h"
  21 #include "disk-io.h"
  22 #include "backref.h"
  23 #include "ulist.h"
  24 #include "transaction.h"
  25 #include "delayed-ref.h"
  26 #include "locking.h"
  27
  28 struct extent_inode_elem {
  29         u64 inum;
  30         u64 offset;
  31         struct extent_inode_elem *next;
  32 };
  33
  34 static int check_extent_in_eb(struct btrfs_key *key, struct extent_buffer *eb,
  35                                 struct btrfs_file_extent_item *fi,
  36                                 u64 extent_item_pos,
  37                                 struct extent_inode_elem **eie)
  38 {
  39         u64 offset = 0;
  40         struct extent_inode_elem *e;
  41
  42         if (!btrfs_file_extent_compression(eb, fi) &&
  43             !btrfs_file_extent_encryption(eb, fi) &&
  44             !btrfs_file_extent_other_encoding(eb, fi)) {
  45                 u64 data_offset;
  46                 u64 data_len;
  47
  48                 data_offset = btrfs_file_extent_offset(eb, fi);
  49                 data_len = btrfs_file_extent_num_bytes(eb, fi);
  50
  51                 if (extent_item_pos < data_offset ||
  52                     extent_item_pos >= data_offset + data_len)
  53                         return 1;
  54                 offset = extent_item_pos - data_offset;
  55         }
  56
  57         e = kmalloc(sizeof(*e), GFP_NOFS);
  58         if (!e)
  59                 return -ENOMEM;
  60
  61         e->next = *eie;
  62         e->inum = key->objectid;
  63         e->offset = key->offset + offset;
  64         *eie = e;
  65
  66         return 0;
  67 }
  68
  69 static int find_extent_in_eb(struct extent_buffer *eb, u64 wanted_disk_byte,
  70                                 u64 extent_item_pos,
  71                                 struct extent_inode_elem **eie)
  72 {
  73         u64 disk_byte;
  74         struct btrfs_key key;
  75         struct btrfs_file_extent_item *fi;
  76         int slot;
  77         int nritems;
  78         int extent_type;
  79         int ret;
  80
  81         /*
  82          * from the shared data ref, we only have the leaf but we need
  83          * the key. thus, we must look into all items and see that we
  84          * find one (some) with a reference to our extent item.
  85          */
  86         nritems = btrfs_header_nritems(eb);
  87         for (slot = 0; slot < nritems; ++slot) {
  88                 btrfs_item_key_to_cpu(eb, &key, slot);
  89                 if (key.type != BTRFS_EXTENT_DATA_KEY)
  90                         continue;
  91                 fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
  92                 extent_type = btrfs_file_extent_type(eb, fi);
  93                 if (extent_type == BTRFS_FILE_EXTENT_INLINE)
  94                         continue;
  95                 /* don't skip BTRFS_FILE_EXTENT_PREALLOC, we can handle that */
  96                 disk_byte = btrfs_file_extent_disk_bytenr(eb, fi);
  97                 if (disk_byte != wanted_disk_byte)
  98                         continue;
  99
 100                 ret = check_extent_in_eb(&key, eb, fi, extent_item_pos, eie);
 101                 if (ret < 0)
 102                         return ret;
 103         }
 104
 105         return 0;
 106 }
 107
 108 /*
 109  * this structure records all encountered refs on the way up to the root
 110  */
 111 struct __prelim_ref {
 112         struct list_head list;
 113         u64 root_id;
 114         struct btrfs_key key_for_search;
 115         int level;
 116         int count;
 117         struct extent_inode_elem *inode_list;
 118         u64 parent;
 119         u64 wanted_disk_byte;
 120 };
 121
 122 /*
 123  * the rules for all callers of this function are:
 124  * - obtaining the parent is the goal
 125  * - if you add a key, you must know that it is a correct key
 126  * - if you cannot add the parent or a correct key, then we will look into the
 127  *   block later to set a correct key
 128  *
 129  * delayed refs
 130  * ============
 131  *        backref type | shared | indirect | shared | indirect
 132  * information         |   tree |     tree |   data |     data
 133  * --------------------+--------+----------+--------+----------
 134  *      parent logical |    y   |     -    |    -   |     -
 135  *      key to resolve |    -   |     y    |    y   |     y
 136  *  tree block logical |    -   |     -    |    -   |     -
 137  *  root for resolving |    y   |     y    |    y   |     y
 138  *
 139  * - column 1:       we've the parent -> done
 140  * - column 2, 3, 4: we use the key to find the parent
 141  *
 142  * on disk refs (inline or keyed)
 143  * ==============================
 144  *        backref type | shared | indirect | shared | indirect
 145  * information         |   tree |     tree |   data |     data
 146  * --------------------+--------+----------+--------+----------
 147  *      parent logical |    y   |     -    |    y   |     -
 148  *      key to resolve |    -   |     -    |    -   |     y
 149  *  tree block logical |    y   |     y    |    y   |     y
 150  *  root for resolving |    -   |     y    |    y   |     y
 151  *
 152  * - column 1, 3: we've the parent -> done
 153  * - column 2:    we take the first key from the block to find the parent
 154  *                (see __add_missing_keys)
 155  * - column 4:    we use the key to find the parent
 156  *
 157  * additional information that's available but not required to find the parent
 158  * block might help in merging entries to gain some speed.
 159  */
 160
 161 static int __add_prelim_ref(struct list_head *head, u64 root_id,
 162                             struct btrfs_key *key, int level,
 163                             u64 parent, u64 wanted_disk_byte, int count)
 164 {
 165         struct __prelim_ref *ref;
 166
 167         /* in case we're adding delayed refs, we're holding the refs spinlock */
 168         ref = kmalloc(sizeof(*ref), GFP_ATOMIC);
 169         if (!ref)
 170                 return -ENOMEM;
 171
 172         ref->root_id = root_id;
 173         if (key)
 174                 ref->key_for_search = *key;
 175         else
 176                 memset(&ref->key_for_search, 0, sizeof(ref->key_for_search));
 177
 178         ref->inode_list = NULL;
 179         ref->level = level;
 180         ref->count = count;
 181         ref->parent = parent;
 182         ref->wanted_disk_byte = wanted_disk_byte;
 183         list_add_tail(&ref->list, head);
 184
 185         return 0;
 186 }
 187
 188 static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path,
 189                                 struct ulist *parents, int level,
 190                                 struct btrfs_key *key_for_search, u64 time_seq,
 191                                 u64 wanted_disk_byte,
 192                                 const u64 *extent_item_pos)
 193 {
 194         int ret = 0;
 195         int slot;
 196         struct extent_buffer *eb;
 197         struct btrfs_key key;
 198         struct btrfs_file_extent_item *fi;
 199         struct extent_inode_elem *eie = NULL, *old = NULL;
 200         u64 disk_byte;
 201
 202         if (level != 0) {
 203                 eb = path->nodes[level];
 204                 ret = ulist_add(parents, eb->start, 0, GFP_NOFS);
 205                 if (ret < 0)
 206                         return ret;
 207                 return 0;
 208         }
 209
 210         /*
 211          * We normally enter this function with the path already pointing to
 212          * the first item to check. But sometimes, we may enter it with
 213          * slot==nritems. In that case, go to the next leaf before we continue.
 214          */
 215         if (path->slots[0] >= btrfs_header_nritems(path->nodes[0]))
 216                 ret = btrfs_next_old_leaf(root, path, time_seq);
 217
 218         while (!ret) {
 219                 eb = path->nodes[0];
 220                 slot = path->slots[0];
 221
 222                 btrfs_item_key_to_cpu(eb, &key, slot);
 223
 224                 if (key.objectid != key_for_search->objectid ||
 225                     key.type != BTRFS_EXTENT_DATA_KEY)
 226                         break;
 227
 228                 fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
 229                 disk_byte = btrfs_file_extent_disk_bytenr(eb, fi);
 230
 231                 if (disk_byte == wanted_disk_byte) {
 232                         eie = NULL;
 233                         old = NULL;
 234                         if (extent_item_pos) {
 235                                 ret = check_extent_in_eb(&key, eb, fi,
 236                                                 *extent_item_pos,
 237                                                 &eie);
 238                                 if (ret < 0)
 239                                         break;
 240                         }
 241                         if (ret > 0)
 242                                 goto next;
 243                         ret = ulist_add_merge(parents, eb->start,
 244                                               (uintptr_t)eie,
 245                                               (u64 *)&old, GFP_NOFS);
 246                         if (ret < 0)
 247                                 break;
 248                         if (!ret && extent_item_pos) {
 249                                 while (old->next)
 250                                         old = old->next;
 251                                 old->next = eie;
 252                         }
 253                 }
 254 next:
 255                 ret = btrfs_next_old_item(root, path, time_seq);
 256         }
 257
 258         if (ret > 0)
 259                 ret = 0;
 260         return ret;
 261 }
 262
 263 /*
 264  * resolve an indirect backref in the form (root_id, key, level)
 265  * to a logical address
 266  */
 267 static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info,
 268                                   struct btrfs_path *path, u64 time_seq,
 269                                   struct __prelim_ref *ref,
 270                                   struct ulist *parents,
 271                                   const u64 *extent_item_pos)
 272 {
 273         struct btrfs_root *root;
 274         struct btrfs_key root_key;
 275         struct extent_buffer *eb;
 276         int ret = 0;
 277         int root_level;
 278         int level = ref->level;
 279
 280         root_key.objectid = ref->root_id;
 281         root_key.type = BTRFS_ROOT_ITEM_KEY;
 282         root_key.offset = (u64)-1;
 283         root = btrfs_read_fs_root_no_name(fs_info, &root_key);
 284         if (IS_ERR(root)) {
 285                 ret = PTR_ERR(root);
 286                 goto out;
 287         }
 288
 289         root_level = btrfs_old_root_level(root, time_seq);
 290
 291         if (root_level + 1 == level)
 292                 goto out;
 293
 294         path->lowest_level = level;
 295         ret = btrfs_search_old_slot(root, &ref->key_for_search, path, time_seq);
 296         pr_debug("search slot in root %llu (level %d, ref count %d) returned "
 297                  "%d for key (%llu %u %llu)\n",
 298                  (unsigned long long)ref->root_id, level, ref->count, ret,
 299                  (unsigned long long)ref->key_for_search.objectid,
 300                  ref->key_for_search.type,
 301                  (unsigned long long)ref->key_for_search.offset);
 302         if (ret < 0)
 303                 goto out;
 304
 305         eb = path->nodes[level];
 306         while (!eb) {
 307                 if (!level) {
 308                         WARN_ON(1);
 309                         ret = 1;
 310                         goto out;
 311                 }
 312                 level--;
 313                 eb = path->nodes[level];
 314         }
 315
 316         ret = add_all_parents(root, path, parents, level, &ref->key_for_search,
 317                                 time_seq, ref->wanted_disk_byte,
 318                                 extent_item_pos);
 319 out:
 320         path->lowest_level = 0;
 321         btrfs_release_path(path);
 322         return ret;
 323 }
 324
 325 /*
 326  * resolve all indirect backrefs from the list
 327  */
 328 static int __resolve_indirect_refs(struct btrfs_fs_info *fs_info,
 329                                    struct btrfs_path *path, u64 time_seq,
 330                                    struct list_head *head,
 331                                    const u64 *extent_item_pos)
 332 {
 333         int err;
 334         int ret = 0;
 335         struct __prelim_ref *ref;
 336         struct __prelim_ref *ref_safe;
 337         struct __prelim_ref *new_ref;
 338         struct ulist *parents;
 339         struct ulist_node *node;
 340         struct ulist_iterator uiter;
 341
 342         parents = ulist_alloc(GFP_NOFS);
 343         if (!parents)
 344                 return -ENOMEM;
 345
 346         /*
 347          * _safe allows us to insert directly after the current item without
 348          * iterating over the newly inserted items.
 349          * we're also allowed to re-assign ref during iteration.
 350          */
 351         list_for_each_entry_safe(ref, ref_safe, head, list) {
 352                 if (ref->parent)        /* already direct */
 353                         continue;
 354                 if (ref->count == 0)
 355                         continue;
 356                 err = __resolve_indirect_ref(fs_info, path, time_seq, ref,
 357                                              parents, extent_item_pos);
 358                 if (err == -ENOMEM)
 359                         goto out;
 360                 if (err)
 361                         continue;
 362
 363                 /* we put the first parent into the ref at hand */
 364                 ULIST_ITER_INIT(&uiter);
 365                 node = ulist_next(parents, &uiter);
 366                 ref->parent = node ? node->val : 0;
 367                 ref->inode_list = node ?
 368                         (struct extent_inode_elem *)(uintptr_t)node->aux : 0;
 369
 370                 /* additional parents require new refs being added here */
 371                 while ((node = ulist_next(parents, &uiter))) {
 372                         new_ref = kmalloc(sizeof(*new_ref), GFP_NOFS);
 373                         if (!new_ref) {
 374                                 ret = -ENOMEM;
 375                                 goto out;
 376                         }
 377                         memcpy(new_ref, ref, sizeof(*ref));
 378                         new_ref->parent = node->val;
 379                         new_ref->inode_list = (struct extent_inode_elem *)
 380                                                         (uintptr_t)node->aux;
 381                         list_add(&new_ref->list, &ref->list);
 382                 }
 383                 ulist_reinit(parents);
 384         }
 385 out:
 386         ulist_free(parents);
 387         return ret;
 388 }
 389
 390 static inline int ref_for_same_block(struct __prelim_ref *ref1,
 391                                      struct __prelim_ref *ref2)
 392 {
 393         if (ref1->level != ref2->level)
 394                 return 0;
 395         if (ref1->root_id != ref2->root_id)
 396                 return 0;
 397         if (ref1->key_for_search.type != ref2->key_for_search.type)
 398                 return 0;
 399         if (ref1->key_for_search.objectid != ref2->key_for_search.objectid)
 400                 return 0;
 401         if (ref1->key_for_search.offset != ref2->key_for_search.offset)
 402                 return 0;
 403         if (ref1->parent != ref2->parent)
 404                 return 0;
 405
 406         return 1;
 407 }
 408
 409 /*
 410  * read tree blocks and add keys where required.
 411  */
 412 static int __add_missing_keys(struct btrfs_fs_info *fs_info,
 413                               struct list_head *head)
 414 {
 415         struct list_head *pos;
 416         struct extent_buffer *eb;
 417
 418         list_for_each(pos, head) {
 419                 struct __prelim_ref *ref;
 420                 ref = list_entry(pos, struct __prelim_ref, list);
 421
 422                 if (ref->parent)
 423                         continue;
 424                 if (ref->key_for_search.type)
 425                         continue;
 426                 BUG_ON(!ref->wanted_disk_byte);
 427                 eb = read_tree_block(fs_info->tree_root, ref->wanted_disk_byte,
 428                                      fs_info->tree_root->leafsize, 0);
 429                 if (!eb || !extent_buffer_uptodate(eb)) {
 430                         free_extent_buffer(eb);
 431                         return -EIO;
 432                 }
 433                 btrfs_tree_read_lock(eb);
 434                 if (btrfs_header_level(eb) == 0)
 435                         btrfs_item_key_to_cpu(eb, &ref->key_for_search, 0);
 436                 else
 437                         btrfs_node_key_to_cpu(eb, &ref->key_for_search, 0);
 438                 btrfs_tree_read_unlock(eb);
 439                 free_extent_buffer(eb);
 440         }
 441         return 0;
 442 }
 443
 444 /*
 445  * merge two lists of backrefs and adjust counts accordingly
 446  *
 447  * mode = 1: merge identical keys, if key is set
 448  *    FIXME: if we add more keys in __add_prelim_ref, we can merge more here.
 449  *           additionally, we could even add a key range for the blocks we
 450  *           looked into to merge even more (-> replace unresolved refs by those
 451  *           having a parent).
 452  * mode = 2: merge identical parents
 453  */
 454 static void __merge_refs(struct list_head *head, int mode)
 455 {
 456         struct list_head *pos1;
 457
 458         list_for_each(pos1, head) {
 459                 struct list_head *n2;
 460                 struct list_head *pos2;
 461                 struct __prelim_ref *ref1;
 462
 463                 ref1 = list_entry(pos1, struct __prelim_ref, list);
 464
 465                 for (pos2 = pos1->next, n2 = pos2->next; pos2 != head;
 466                      pos2 = n2, n2 = pos2->next) {
 467                         struct __prelim_ref *ref2;
 468                         struct __prelim_ref *xchg;
 469                         struct extent_inode_elem *eie;
 470
 471                         ref2 = list_entry(pos2, struct __prelim_ref, list);
 472
 473                         if (mode == 1) {
 474                                 if (!ref_for_same_block(ref1, ref2))
 475                                         continue;
 476                                 if (!ref1->parent && ref2->parent) {
 477                                         xchg = ref1;
 478                                         ref1 = ref2;
 479                                         ref2 = xchg;
 480                                 }
 481                         } else {
 482                                 if (ref1->parent != ref2->parent)
 483                                         continue;
 484                         }
 485
 486                         eie = ref1->inode_list;
 487                         while (eie && eie->next)
 488                                 eie = eie->next;
 489                         if (eie)
 490                                 eie->next = ref2->inode_list;
 491                         else
 492                                 ref1->inode_list = ref2->inode_list;
 493                         ref1->count += ref2->count;
 494
 495                         list_del(&ref2->list);
 496                         kfree(ref2);
 497                 }
 498
 499         }
 500 }
 501
 502 /*
 503  * add all currently queued delayed refs from this head whose seq nr is
 504  * smaller or equal that seq to the list
 505  */
 506 static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq,
 507                               struct list_head *prefs)
 508 {
 509         struct btrfs_delayed_extent_op *extent_op = head->extent_op;
 510         struct rb_node *n = &head->node.rb_node;
 511         struct btrfs_key key;
 512         struct btrfs_key op_key = {0};
 513         int sgn;
 514         int ret = 0;
 515
 516         if (extent_op && extent_op->update_key)
 517                 btrfs_disk_key_to_cpu(&op_key, &extent_op->key);
 518
 519         while ((n = rb_prev(n))) {
 520                 struct btrfs_delayed_ref_node *node;
 521                 node = rb_entry(n, struct btrfs_delayed_ref_node,
 522                                 rb_node);
 523                 if (node->bytenr != head->node.bytenr)
 524                         break;
 525                 WARN_ON(node->is_head);
 526
 527                 if (node->seq > seq)
 528                         continue;
 529
 530                 switch (node->action) {
 531                 case BTRFS_ADD_DELAYED_EXTENT:
 532                 case BTRFS_UPDATE_DELAYED_HEAD:
 533                         WARN_ON(1);
 534                         continue;
 535                 case BTRFS_ADD_DELAYED_REF:
 536                         sgn = 1;
 537                         break;
 538                 case BTRFS_DROP_DELAYED_REF:
 539                         sgn = -1;
 540                         break;
 541                 default:
 542                         BUG_ON(1);
 543                 }
 544                 switch (node->type) {
 545                 case BTRFS_TREE_BLOCK_REF_KEY: {
 546                         struct btrfs_delayed_tree_ref *ref;
 547
 548                         ref = btrfs_delayed_node_to_tree_ref(node);
 549                         ret = __add_prelim_ref(prefs, ref->root, &op_key,
 550                                                ref->level + 1, 0, node->bytenr,
 551                                                node->ref_mod * sgn);
 552                         break;
 553                 }
 554                 case BTRFS_SHARED_BLOCK_REF_KEY: {
 555                         struct btrfs_delayed_tree_ref *ref;
 556
 557                         ref = btrfs_delayed_node_to_tree_ref(node);
 558                         ret = __add_prelim_ref(prefs, ref->root, NULL,
 559                                                ref->level + 1, ref->parent,
 560                                                node->bytenr,
 561                                                node->ref_mod * sgn);
 562                         break;
 563                 }
 564                 case BTRFS_EXTENT_DATA_REF_KEY: {
 565                         struct btrfs_delayed_data_ref *ref;
 566                         ref = btrfs_delayed_node_to_data_ref(node);
 567
 568                         key.objectid = ref->objectid;
 569                         key.type = BTRFS_EXTENT_DATA_KEY;
 570                         key.offset = ref->offset;
 571                         ret = __add_prelim_ref(prefs, ref->root, &key, 0, 0,
 572                                                node->bytenr,
 573                                                node->ref_mod * sgn);
 574                         break;
 575                 }
 576                 case BTRFS_SHARED_DATA_REF_KEY: {
 577                         struct btrfs_delayed_data_ref *ref;
 578
 579                         ref = btrfs_delayed_node_to_data_ref(node);
 580
 581                         key.objectid = ref->objectid;
 582                         key.type = BTRFS_EXTENT_DATA_KEY;
 583                         key.offset = ref->offset;
 584                         ret = __add_prelim_ref(prefs, ref->root, &key, 0,
 585                                                ref->parent, node->bytenr,
 586                                                node->ref_mod * sgn);
 587                         break;
 588                 }
 589                 default:
 590                         WARN_ON(1);
 591                 }
 592                 if (ret)
 593                         return ret;
 594         }
 595
 596         return 0;
 597 }
 598
 599 /*
 600  * add all inline backrefs for bytenr to the list
 601  */
 602 static int __add_inline_refs(struct btrfs_fs_info *fs_info,
 603                              struct btrfs_path *path, u64 bytenr,
 604                              int *info_level, struct list_head *prefs)
 605 {
 606         int ret = 0;
 607         int slot;
 608         struct extent_buffer *leaf;
 609         struct btrfs_key key;
 610         struct btrfs_key found_key;
 611         unsigned long ptr;
 612         unsigned long end;
 613         struct btrfs_extent_item *ei;
 614         u64 flags;
 615         u64 item_size;
 616
 617         /*
 618          * enumerate all inline refs
 619          */
 620         leaf = path->nodes[0];
 621         slot = path->slots[0];
 622
 623         item_size = btrfs_item_size_nr(leaf, slot);
 624         BUG_ON(item_size < sizeof(*ei));
 625
 626         ei = btrfs_item_ptr(leaf, slot, struct btrfs_extent_item);
 627         flags = btrfs_extent_flags(leaf, ei);
 628         btrfs_item_key_to_cpu(leaf, &found_key, slot);
 629
 630         ptr = (unsigned long)(ei + 1);
 631         end = (unsigned long)ei + item_size;
 632
 633         if (found_key.type == BTRFS_EXTENT_ITEM_KEY &&
 634             flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
 635                 struct btrfs_tree_block_info *info;
 636
 637                 info = (struct btrfs_tree_block_info *)ptr;
 638                 *info_level = btrfs_tree_block_level(leaf, info);
 639                 ptr += sizeof(struct btrfs_tree_block_info);
 640                 BUG_ON(ptr > end);
 641         } else if (found_key.type == BTRFS_METADATA_ITEM_KEY) {
 642                 *info_level = found_key.offset;
 643         } else {
 644                 BUG_ON(!(flags & BTRFS_EXTENT_FLAG_DATA));
 645         }
 646
 647         while (ptr < end) {
 648                 struct btrfs_extent_inline_ref *iref;
 649                 u64 offset;
 650                 int type;
 651
 652                 iref = (struct btrfs_extent_inline_ref *)ptr;
 653                 type = btrfs_extent_inline_ref_type(leaf, iref);
 654                 offset = btrfs_extent_inline_ref_offset(leaf, iref);
 655
 656                 switch (type) {
 657                 case BTRFS_SHARED_BLOCK_REF_KEY:
 658                         ret = __add_prelim_ref(prefs, 0, NULL,
 659                                                 *info_level + 1, offset,
 660                                                 bytenr, 1);
 661                         break;
 662                 case BTRFS_SHARED_DATA_REF_KEY: {
 663                         struct btrfs_shared_data_ref *sdref;
 664                         int count;
 665
 666                         sdref = (struct btrfs_shared_data_ref *)(iref + 1);
 667                         count = btrfs_shared_data_ref_count(leaf, sdref);
 668                         ret = __add_prelim_ref(prefs, 0, NULL, 0, offset,
 669                                                bytenr, count);
 670                         break;
 671                 }
 672                 case BTRFS_TREE_BLOCK_REF_KEY:
 673                         ret = __add_prelim_ref(prefs, offset, NULL,
 674                                                *info_level + 1, 0,
 675                                                bytenr, 1);
 676                         break;
 677                 case BTRFS_EXTENT_DATA_REF_KEY: {
 678                         struct btrfs_extent_data_ref *dref;
 679                         int count;
 680                         u64 root;
 681
 682                         dref = (struct btrfs_extent_data_ref *)(&iref->offset);
 683                         count = btrfs_extent_data_ref_count(leaf, dref);
 684                         key.objectid = btrfs_extent_data_ref_objectid(leaf,
 685                                                                       dref);
 686                         key.type = BTRFS_EXTENT_DATA_KEY;
 687                         key.offset = btrfs_extent_data_ref_offset(leaf, dref);
 688                         root = btrfs_extent_data_ref_root(leaf, dref);
 689                         ret = __add_prelim_ref(prefs, root, &key, 0, 0,
 690                                                bytenr, count);
 691                         break;
 692                 }
 693                 default:
 694                         WARN_ON(1);
 695                 }
 696                 if (ret)
 697                         return ret;
 698                 ptr += btrfs_extent_inline_ref_size(type);
 699         }
 700
 701         return 0;
 702 }
 703
 704 /*
 705  * add all non-inline backrefs for bytenr to the list
 706  */
 707 static int __add_keyed_refs(struct btrfs_fs_info *fs_info,
 708                             struct btrfs_path *path, u64 bytenr,
 709                             int info_level, struct list_head *prefs)
 710 {
 711         struct btrfs_root *extent_root = fs_info->extent_root;
 712         int ret;
 713         int slot;
 714         struct extent_buffer *leaf;
 715         struct btrfs_key key;
 716
 717         while (1) {
 718                 ret = btrfs_next_item(extent_root, path);
 719                 if (ret < 0)
 720                         break;
 721                 if (ret) {
 722                         ret = 0;
 723                         break;
 724                 }
 725
 726                 slot = path->slots[0];
 727                 leaf = path->nodes[0];
 728                 btrfs_item_key_to_cpu(leaf, &key, slot);
 729
 730                 if (key.objectid != bytenr)
 731                         break;
 732                 if (key.type < BTRFS_TREE_BLOCK_REF_KEY)
 733                         continue;
 734                 if (key.type > BTRFS_SHARED_DATA_REF_KEY)
 735                         break;
 736
 737                 switch (key.type) {
 738                 case BTRFS_SHARED_BLOCK_REF_KEY:
 739                         ret = __add_prelim_ref(prefs, 0, NULL,
 740                                                 info_level + 1, key.offset,
 741                                                 bytenr, 1);
 742                         break;
 743                 case BTRFS_SHARED_DATA_REF_KEY: {
 744                         struct btrfs_shared_data_ref *sdref;
 745                         int count;
 746
 747                         sdref = btrfs_item_ptr(leaf, slot,
 748                                               struct btrfs_shared_data_ref);
 749                         count = btrfs_shared_data_ref_count(leaf, sdref);
 750                         ret = __add_prelim_ref(prefs, 0, NULL, 0, key.offset,
 751                                                 bytenr, count);
 752                         break;
 753                 }
 754                 case BTRFS_TREE_BLOCK_REF_KEY:
 755                         ret = __add_prelim_ref(prefs, key.offset, NULL,
 756                                                info_level + 1, 0,
 757                                                bytenr, 1);
 758                         break;
 759                 case BTRFS_EXTENT_DATA_REF_KEY: {
 760                         struct btrfs_extent_data_ref *dref;
 761                         int count;
 762                         u64 root;
 763
 764                         dref = btrfs_item_ptr(leaf, slot,
 765                                               struct btrfs_extent_data_ref);
 766                         count = btrfs_extent_data_ref_count(leaf, dref);
 767                         key.objectid = btrfs_extent_data_ref_objectid(leaf,
 768                                                                       dref);
 769                         key.type = BTRFS_EXTENT_DATA_KEY;
 770                         key.offset = btrfs_extent_data_ref_offset(leaf, dref);
 771                         root = btrfs_extent_data_ref_root(leaf, dref);
 772                         ret = __add_prelim_ref(prefs, root, &key, 0, 0,
 773                                                bytenr, count);
 774                         break;
 775                 }
 776                 default:
 777                         WARN_ON(1);
 778                 }
 779                 if (ret)
 780                         return ret;
 781
 782         }
 783
 784         return ret;
 785 }
 786
 787 /*
 788  * this adds all existing backrefs (inline backrefs, backrefs and delayed
 789  * refs) for the given bytenr to the refs list, merges duplicates and resolves
 790  * indirect refs to their parent bytenr.
 791  * When roots are found, they're added to the roots list
 792  *
 793  * FIXME some caching might speed things up
 794  */
 795 static int find_parent_nodes(struct btrfs_trans_handle *trans,
 796                              struct btrfs_fs_info *fs_info, u64 bytenr,
 797                              u64 time_seq, struct ulist *refs,
 798                              struct ulist *roots, const u64 *extent_item_pos)
 799 {
 800         struct btrfs_key key;
 801         struct btrfs_path *path;
 802         struct btrfs_delayed_ref_root *delayed_refs = NULL;
 803         struct btrfs_delayed_ref_head *head;
 804         int info_level = 0;
 805         int ret;
 806         struct list_head prefs_delayed;
 807         struct list_head prefs;
 808         struct __prelim_ref *ref;
 809
 810         INIT_LIST_HEAD(&prefs);
 811         INIT_LIST_HEAD(&prefs_delayed);
 812
 813         key.objectid = bytenr;
 814         key.offset = (u64)-1;
 815         if (btrfs_fs_incompat(fs_info, SKINNY_METADATA))
 816                 key.type = BTRFS_METADATA_ITEM_KEY;
 817         else
 818                 key.type = BTRFS_EXTENT_ITEM_KEY;
 819
 820         path = btrfs_alloc_path();
 821         if (!path)
 822                 return -ENOMEM;
 823         if (!trans)
 824                 path->search_commit_root = 1;
 825
 826         /*
 827          * grab both a lock on the path and a lock on the delayed ref head.
 828          * We need both to get a consistent picture of how the refs look
 829          * at a specified point in time
 830          */
 831 again:
 832         head = NULL;
 833
 834         ret = btrfs_search_slot(trans, fs_info->extent_root, &key, path, 0, 0);
 835         if (ret < 0)
 836                 goto out;
 837         BUG_ON(ret == 0);
 838
 839         if (trans) {
 840                 /*
 841                  * look if there are updates for this ref queued and lock the
 842                  * head
 843                  */
 844                 delayed_refs = &trans->transaction->delayed_refs;
 845                 spin_lock(&delayed_refs->lock);
 846                 head = btrfs_find_delayed_ref_head(trans, bytenr);
 847                 if (head) {
 848                         if (!mutex_trylock(&head->mutex)) {
 849                                 atomic_inc(&head->node.refs);
 850                                 spin_unlock(&delayed_refs->lock);
 851
 852                                 btrfs_release_path(path);
 853
 854                                 /*
 855                                  * Mutex was contended, block until it's
 856                                  * released and try again
 857                                  */
 858                                 mutex_lock(&head->mutex);
 859                                 mutex_unlock(&head->mutex);
 860                                 btrfs_put_delayed_ref(&head->node);
 861                                 goto again;
 862                         }
 863                         ret = __add_delayed_refs(head, time_seq,
 864                                                  &prefs_delayed);
 865                         mutex_unlock(&head->mutex);
 866                         if (ret) {
 867                                 spin_unlock(&delayed_refs->lock);
 868                                 goto out;
 869                         }
 870                 }
 871                 spin_unlock(&delayed_refs->lock);
 872         }
 873
 874         if (path->slots[0]) {
 875                 struct extent_buffer *leaf;
 876                 int slot;
 877
 878                 path->slots[0]--;
 879                 leaf = path->nodes[0];
 880                 slot = path->slots[0];
 881                 btrfs_item_key_to_cpu(leaf, &key, slot);
 882                 if (key.objectid == bytenr &&
 883                     (key.type == BTRFS_EXTENT_ITEM_KEY ||
 884                      key.type == BTRFS_METADATA_ITEM_KEY)) {
 885                         ret = __add_inline_refs(fs_info, path, bytenr,
 886                                                 &info_level, &prefs);
 887                         if (ret)
 888                                 goto out;
 889                         ret = __add_keyed_refs(fs_info, path, bytenr,
 890                                                info_level, &prefs);
 891                         if (ret)
 892                                 goto out;
 893                 }
 894         }
 895         btrfs_release_path(path);
 896
 897         list_splice_init(&prefs_delayed, &prefs);
 898
 899         ret = __add_missing_keys(fs_info, &prefs);
 900         if (ret)
 901                 goto out;
 902
 903         __merge_refs(&prefs, 1);
 904
 905         ret = __resolve_indirect_refs(fs_info, path, time_seq, &prefs,
 906                                       extent_item_pos);
 907         if (ret)
 908                 goto out;
 909
 910         __merge_refs(&prefs, 2);
 911
 912         while (!list_empty(&prefs)) {
 913                 ref = list_first_entry(&prefs, struct __prelim_ref, list);
 914                 list_del(&ref->list);
 915                 WARN_ON(ref->count < 0);
 916                 if (ref->count && ref->root_id && ref->parent == 0) {
 917                         /* no parent == root of tree */
 918                         ret = ulist_add(roots, ref->root_id, 0, GFP_NOFS);
 919                         if (ret < 0)
 920                                 goto out;
 921                 }
 922                 if (ref->count && ref->parent) {
 923                         struct extent_inode_elem *eie = NULL;
 924                         if (extent_item_pos && !ref->inode_list) {
 925                                 u32 bsz;
 926                                 struct extent_buffer *eb;
 927                                 bsz = btrfs_level_size(fs_info->extent_root,
 928                                                         info_level);
 929                                 eb = read_tree_block(fs_info->extent_root,
 930                                                            ref->parent, bsz, 0);
 931                                 if (!eb || !extent_buffer_uptodate(eb)) {
 932                                         free_extent_buffer(eb);
 933                                         ret = -EIO;
 934                                         goto out;
 935                                 }
 936                                 ret = find_extent_in_eb(eb, bytenr,
 937                                                         *extent_item_pos, &eie);
 938                                 ref->inode_list = eie;
 939                                 free_extent_buffer(eb);
 940                         }
 941                         ret = ulist_add_merge(refs, ref->parent,
 942                                               (uintptr_t)ref->inode_list,
 943                                               (u64 *)&eie, GFP_NOFS);
 944                         if (ret < 0)
 945                                 goto out;
 946                         if (!ret && extent_item_pos) {
 947                                 /*
 948                                  * we've recorded that parent, so we must extend
 949                                  * its inode list here
 950                                  */
 951                                 BUG_ON(!eie);
 952                                 while (eie->next)
 953                                         eie = eie->next;
 954                                 eie->next = ref->inode_list;
 955                         }
 956                 }
 957                 kfree(ref);
 958         }
 959
 960 out:
 961         btrfs_free_path(path);
 962         while (!list_empty(&prefs)) {
 963                 ref = list_first_entry(&prefs, struct __prelim_ref, list);
 964                 list_del(&ref->list);
 965                 kfree(ref);
 966         }
 967         while (!list_empty(&prefs_delayed)) {
 968                 ref = list_first_entry(&prefs_delayed, struct __prelim_ref,
 969                                        list);
 970                 list_del(&ref->list);
 971                 kfree(ref);
 972         }
 973
 974         return ret;
 975 }
 976
 977 static void free_leaf_list(struct ulist *blocks)
 978 {
 979         struct ulist_node *node = NULL;
 980         struct extent_inode_elem *eie;
 981         struct extent_inode_elem *eie_next;
 982         struct ulist_iterator uiter;
 983
 984         ULIST_ITER_INIT(&uiter);
 985         while ((node = ulist_next(blocks, &uiter))) {
 986                 if (!node->aux)
 987                         continue;
 988                 eie = (struct extent_inode_elem *)(uintptr_t)node->aux;
 989                 for (; eie; eie = eie_next) {
 990                         eie_next = eie->next;
 991                         kfree(eie);
 992                 }
 993                 node->aux = 0;
 994         }
 995
 996         ulist_free(blocks);
 997 }
 998
 999 /*
1000  * Finds all leafs with a reference to the specified combination of bytenr and
1001  * offset. key_list_head will point to a list of corresponding keys (caller must
1002  * free each list element). The leafs will be stored in the leafs ulist, which
1003  * must be freed with ulist_free.
1004  *
1005  * returns 0 on success, <0 on error
1006  */
1007 static int btrfs_find_all_leafs(struct btrfs_trans_handle *trans,
1008                                 struct btrfs_fs_info *fs_info, u64 bytenr,
1009                                 u64 time_seq, struct ulist **leafs,
1010                                 const u64 *extent_item_pos)
1011 {
1012         struct ulist *tmp;
1013         int ret;
1014
1015         tmp = ulist_alloc(GFP_NOFS);
1016         if (!tmp)
1017                 return -ENOMEM;
1018         *leafs = ulist_alloc(GFP_NOFS);
1019         if (!*leafs) {
1020                 ulist_free(tmp);
1021                 return -ENOMEM;
1022         }
1023
1024         ret = find_parent_nodes(trans, fs_info, bytenr,
1025                                 time_seq, *leafs, tmp, extent_item_pos);
1026         ulist_free(tmp);
1027
1028         if (ret < 0 && ret != -ENOENT) {
1029                 free_leaf_list(*leafs);
1030                 return ret;
1031         }
1032
1033         return 0;
1034 }
1035
1036 /*
1037  * walk all backrefs for a given extent to find all roots that reference this
1038  * extent. Walking a backref means finding all extents that reference this
1039  * extent and in turn walk the backrefs of those, too. Naturally this is a
1040  * recursive process, but here it is implemented in an iterative fashion: We
1041  * find all referencing extents for the extent in question and put them on a
1042  * list. In turn, we find all referencing extents for those, further appending
1043  * to the list. The way we iterate the list allows adding more elements after
1044  * the current while iterating. The process stops when we reach the end of the
1045  * list. Found roots are added to the roots list.
1046  *
1047  * returns 0 on success, < 0 on error.
1048  */
1049 int btrfs_find_all_roots(struct btrfs_trans_handle *trans,
1050                                 struct btrfs_fs_info *fs_info, u64 bytenr,
1051                                 u64 time_seq, struct ulist **roots)
1052 {
1053         struct ulist *tmp;
1054         struct ulist_node *node = NULL;
1055         struct ulist_iterator uiter;
1056         int ret;
1057
1058         tmp = ulist_alloc(GFP_NOFS);
1059         if (!tmp)
1060                 return -ENOMEM;
1061         *roots = ulist_alloc(GFP_NOFS);
1062         if (!*roots) {
1063                 ulist_free(tmp);
1064                 return -ENOMEM;
1065         }
1066
1067         ULIST_ITER_INIT(&uiter);
1068         while (1) {
1069                 ret = find_parent_nodes(trans, fs_info, bytenr,
1070                                         time_seq, tmp, *roots, NULL);
1071                 if (ret < 0 && ret != -ENOENT) {
1072                         ulist_free(tmp);
1073                         ulist_free(*roots);
1074                         return ret;
1075                 }
1076                 node = ulist_next(tmp, &uiter);
1077                 if (!node)
1078                         break;
1079                 bytenr = node->val;
1080         }
1081
1082         ulist_free(tmp);
1083         return 0;
1084 }
1085
1086
1087 static int __inode_info(u64 inum, u64 ioff, u8 key_type,
1088                         struct btrfs_root *fs_root, struct btrfs_path *path,
1089                         struct btrfs_key *found_key)
1090 {
1091         int ret;
1092         struct btrfs_key key;
1093         struct extent_buffer *eb;
1094
1095         key.type = key_type;
1096         key.objectid = inum;
1097         key.offset = ioff;
1098
1099         ret = btrfs_search_slot(NULL, fs_root, &key, path, 0, 0);
1100         if (ret < 0)
1101                 return ret;
1102
1103         eb = path->nodes[0];
1104         if (ret && path->slots[0] >= btrfs_header_nritems(eb)) {
1105                 ret = btrfs_next_leaf(fs_root, path);
1106                 if (ret)
1107                         return ret;
1108                 eb = path->nodes[0];
1109         }
1110
1111         btrfs_item_key_to_cpu(eb, found_key, path->slots[0]);
1112         if (found_key->type != key.type || found_key->objectid != key.objectid)
1113                 return 1;
1114
1115         return 0;
1116 }
1117
1118 /*
1119  * this makes the path point to (inum INODE_ITEM ioff)
1120  */
1121 int inode_item_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
1122                         struct btrfs_path *path)
1123 {
1124         struct btrfs_key key;
1125         return __inode_info(inum, ioff, BTRFS_INODE_ITEM_KEY, fs_root, path,
1126                                 &key);
1127 }
1128
1129 static int inode_ref_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
1130                                 struct btrfs_path *path,
1131                                 struct btrfs_key *found_key)
1132 {
1133         return __inode_info(inum, ioff, BTRFS_INODE_REF_KEY, fs_root, path,
1134                                 found_key);
1135 }
1136
1137 int btrfs_find_one_extref(struct btrfs_root *root, u64 inode_objectid,
1138                           u64 start_off, struct btrfs_path *path,
1139                           struct btrfs_inode_extref **ret_extref,
1140                           u64 *found_off)
1141 {
1142         int ret, slot;
1143         struct btrfs_key key;
1144         struct btrfs_key found_key;
1145         struct btrfs_inode_extref *extref;
1146         struct extent_buffer *leaf;
1147         unsigned long ptr;
1148
1149         key.objectid = inode_objectid;
1150         btrfs_set_key_type(&key, BTRFS_INODE_EXTREF_KEY);
1151         key.offset = start_off;
1152
1153         ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
1154         if (ret < 0)
1155                 return ret;
1156
1157         while (1) {
1158                 leaf = path->nodes[0];
1159                 slot = path->slots[0];
1160                 if (slot >= btrfs_header_nritems(leaf)) {
1161                         /*
1162                          * If the item at offset is not found,
1163                          * btrfs_search_slot will point us to the slot
1164                          * where it should be inserted. In our case
1165                          * that will be the slot directly before the
1166                          * next INODE_REF_KEY_V2 item. In the case
1167                          * that we're pointing to the last slot in a
1168                          * leaf, we must move one leaf over.
1169                          */
1170                         ret = btrfs_next_leaf(root, path);
1171                         if (ret) {
1172                                 if (ret >= 1)
1173                                         ret = -ENOENT;
1174                                 break;
1175                         }
1176                         continue;
1177                 }
1178
1179                 btrfs_item_key_to_cpu(leaf, &found_key, slot);
1180
1181                 /*
1182                  * Check that we're still looking at an extended ref key for
1183                  * this particular objectid. If we have different
1184                  * objectid or type then there are no more to be found
1185                  * in the tree and we can exit.
1186                  */
1187                 ret = -ENOENT;
1188                 if (found_key.objectid != inode_objectid)
1189                         break;
1190                 if (btrfs_key_type(&found_key) != BTRFS_INODE_EXTREF_KEY)
1191                         break;
1192
1193                 ret = 0;
1194                 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
1195                 extref = (struct btrfs_inode_extref *)ptr;
1196                 *ret_extref = extref;
1197                 if (found_off)
1198                         *found_off = found_key.offset;
1199                 break;
1200         }
1201
1202         return ret;
1203 }
1204
1205 /*
1206  * this iterates to turn a name (from iref/extref) into a full filesystem path.
1207  * Elements of the path are separated by '/' and the path is guaranteed to be
1208  * 0-terminated. the path is only given within the current file system.
1209  * Therefore, it never starts with a '/'. the caller is responsible to provide
1210  * "size" bytes in "dest". the dest buffer will be filled backwards. finally,
1211  * the start point of the resulting string is returned. this pointer is within
1212  * dest, normally.
1213  * in case the path buffer would overflow, the pointer is decremented further
1214  * as if output was written to the buffer, though no more output is actually
1215  * generated. that way, the caller can determine how much space would be
1216  * required for the path to fit into the buffer. in that case, the returned
1217  * value will be smaller than dest. callers must check this!
1218  */
1219 char *btrfs_ref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
1220                         u32 name_len, unsigned long name_off,
1221                         struct extent_buffer *eb_in, u64 parent,
1222                         char *dest, u32 size)
1223 {
1224         int slot;
1225         u64 next_inum;
1226         int ret;
1227         s64 bytes_left = ((s64)size) - 1;
1228         struct extent_buffer *eb = eb_in;
1229         struct btrfs_key found_key;
1230         int leave_spinning = path->leave_spinning;
1231         struct btrfs_inode_ref *iref;
1232
1233         if (bytes_left >= 0)
1234                 dest[bytes_left] = '\0';
1235
1236         path->leave_spinning = 1;
1237         while (1) {
1238                 bytes_left -= name_len;
1239                 if (bytes_left >= 0)
1240                         read_extent_buffer(eb, dest + bytes_left,
1241                                            name_off, name_len);
1242                 if (eb != eb_in) {
1243                         btrfs_tree_read_unlock_blocking(eb);
1244                         free_extent_buffer(eb);
1245                 }
1246                 ret = inode_ref_info(parent, 0, fs_root, path, &found_key);
1247                 if (ret > 0)
1248                         ret = -ENOENT;
1249                 if (ret)
1250                         break;
1251
1252                 next_inum = found_key.offset;
1253
1254                 /* regular exit ahead */
1255                 if (parent == next_inum)
1256                         break;
1257
1258                 slot = path->slots[0];
1259                 eb = path->nodes[0];
1260                 /* make sure we can use eb after releasing the path */
1261                 if (eb != eb_in) {
1262                         atomic_inc(&eb->refs);
1263                         btrfs_tree_read_lock(eb);
1264                         btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1265                 }
1266                 btrfs_release_path(path);
1267                 iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
1268
1269                 name_len = btrfs_inode_ref_name_len(eb, iref);
1270                 name_off = (unsigned long)(iref + 1);
1271
1272                 parent = next_inum;
1273                 --bytes_left;
1274                 if (bytes_left >= 0)
1275                         dest[bytes_left] = '/';
1276         }
1277
1278         btrfs_release_path(path);
1279         path->leave_spinning = leave_spinning;
1280
1281         if (ret)
1282                 return ERR_PTR(ret);
1283
1284         return dest + bytes_left;
1285 }
1286
1287 /*
1288  * this makes the path point to (logical EXTENT_ITEM *)
1289  * returns BTRFS_EXTENT_FLAG_DATA for data, BTRFS_EXTENT_FLAG_TREE_BLOCK for
1290  * tree blocks and <0 on error.
1291  */
1292 int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical,
1293                         struct btrfs_path *path, struct btrfs_key *found_key,
1294                         u64 *flags_ret)
1295 {
1296         int ret;
1297         u64 flags;
1298         u64 size = 0;
1299         u32 item_size;
1300         struct extent_buffer *eb;
1301         struct btrfs_extent_item *ei;
1302         struct btrfs_key key;
1303
1304         if (btrfs_fs_incompat(fs_info, SKINNY_METADATA))
1305                 key.type = BTRFS_METADATA_ITEM_KEY;
1306         else
1307                 key.type = BTRFS_EXTENT_ITEM_KEY;
1308         key.objectid = logical;
1309         key.offset = (u64)-1;
1310
1311         ret = btrfs_search_slot(NULL, fs_info->extent_root, &key, path, 0, 0);
1312         if (ret < 0)
1313                 return ret;
1314         ret = btrfs_previous_item(fs_info->extent_root, path,
1315                                         0, BTRFS_EXTENT_ITEM_KEY);
1316         if (ret < 0)
1317                 return ret;
1318
1319         btrfs_item_key_to_cpu(path->nodes[0], found_key, path->slots[0]);
1320         if (found_key->type == BTRFS_METADATA_ITEM_KEY)
1321                 size = fs_info->extent_root->leafsize;
1322         else if (found_key->type == BTRFS_EXTENT_ITEM_KEY)
1323                 size = found_key->offset;
1324
1325         if ((found_key->type != BTRFS_EXTENT_ITEM_KEY &&
1326              found_key->type != BTRFS_METADATA_ITEM_KEY) ||
1327             found_key->objectid > logical ||
1328             found_key->objectid + size <= logical) {
1329                 pr_debug("logical %llu is not within any extent\n",
1330                          (unsigned long long)logical);
1331                 return -ENOENT;
1332         }
1333
1334         eb = path->nodes[0];
1335         item_size = btrfs_item_size_nr(eb, path->slots[0]);
1336         BUG_ON(item_size < sizeof(*ei));
1337
1338         ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item);
1339         flags = btrfs_extent_flags(eb, ei);
1340
1341         pr_debug("logical %llu is at position %llu within the extent (%llu "
1342                  "EXTENT_ITEM %llu) flags %#llx size %u\n",
1343                  (unsigned long long)logical,
1344                  (unsigned long long)(logical - found_key->objectid),
1345                  (unsigned long long)found_key->objectid,
1346                  (unsigned long long)found_key->offset,
1347                  (unsigned long long)flags, item_size);
1348
1349         WARN_ON(!flags_ret);
1350         if (flags_ret) {
1351                 if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)
1352                         *flags_ret = BTRFS_EXTENT_FLAG_TREE_BLOCK;
1353                 else if (flags & BTRFS_EXTENT_FLAG_DATA)
1354                         *flags_ret = BTRFS_EXTENT_FLAG_DATA;
1355                 else
1356                         BUG_ON(1);
1357                 return 0;
1358         }
1359
1360         return -EIO;
1361 }
1362
1363 /*
1364  * helper function to iterate extent inline refs. ptr must point to a 0 value
1365  * for the first call and may be modified. it is used to track state.
1366  * if more refs exist, 0 is returned and the next call to
1367  * __get_extent_inline_ref must pass the modified ptr parameter to get the
1368  * next ref. after the last ref was processed, 1 is returned.
1369  * returns <0 on error
1370  */
1371 static int __get_extent_inline_ref(unsigned long *ptr, struct extent_buffer *eb,
1372                                 struct btrfs_extent_item *ei, u32 item_size,
1373                                 struct btrfs_extent_inline_ref **out_eiref,
1374                                 int *out_type)
1375 {
1376         unsigned long end;
1377         u64 flags;
1378         struct btrfs_tree_block_info *info;
1379
1380         if (!*ptr) {
1381                 /* first call */
1382                 flags = btrfs_extent_flags(eb, ei);
1383                 if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
1384                         info = (struct btrfs_tree_block_info *)(ei + 1);
1385                         *out_eiref =
1386                                 (struct btrfs_extent_inline_ref *)(info + 1);
1387                 } else {
1388                         *out_eiref = (struct btrfs_extent_inline_ref *)(ei + 1);
1389                 }
1390                 *ptr = (unsigned long)*out_eiref;
1391                 if ((void *)*ptr >= (void *)ei + item_size)
1392                         return -ENOENT;
1393         }
1394
1395         end = (unsigned long)ei + item_size;
1396         *out_eiref = (struct btrfs_extent_inline_ref *)*ptr;
1397         *out_type = btrfs_extent_inline_ref_type(eb, *out_eiref);
1398
1399         *ptr += btrfs_extent_inline_ref_size(*out_type);
1400         WARN_ON(*ptr > end);
1401         if (*ptr == end)
1402                 return 1; /* last */
1403
1404         return 0;
1405 }
1406
1407 /*
1408  * reads the tree block backref for an extent. tree level and root are returned
1409  * through out_level and out_root. ptr must point to a 0 value for the first
1410  * call and may be modified (see __get_extent_inline_ref comment).
1411  * returns 0 if data was provided, 1 if there was no more data to provide or
1412  * <0 on error.
1413  */
1414 int tree_backref_for_extent(unsigned long *ptr, struct extent_buffer *eb,
1415                                 struct btrfs_extent_item *ei, u32 item_size,
1416                                 u64 *out_root, u8 *out_level)
1417 {
1418         int ret;
1419         int type;
1420         struct btrfs_tree_block_info *info;
1421         struct btrfs_extent_inline_ref *eiref;
1422
1423         if (*ptr == (unsigned long)-1)
1424                 return 1;
1425
1426         while (1) {
1427                 ret = __get_extent_inline_ref(ptr, eb, ei, item_size,
1428                                                 &eiref, &type);
1429                 if (ret < 0)
1430                         return ret;
1431
1432                 if (type == BTRFS_TREE_BLOCK_REF_KEY ||
1433                     type == BTRFS_SHARED_BLOCK_REF_KEY)
1434                         break;
1435
1436                 if (ret == 1)
1437                         return 1;
1438         }
1439
1440         /* we can treat both ref types equally here */
1441         info = (struct btrfs_tree_block_info *)(ei + 1);
1442         *out_root = btrfs_extent_inline_ref_offset(eb, eiref);
1443         *out_level = btrfs_tree_block_level(eb, info);
1444
1445         if (ret == 1)
1446                 *ptr = (unsigned long)-1;
1447
1448         return 0;
1449 }
1450
1451 static int iterate_leaf_refs(struct extent_inode_elem *inode_list,
1452                                 u64 root, u64 extent_item_objectid,
1453                                 iterate_extent_inodes_t *iterate, void *ctx)
1454 {
1455         struct extent_inode_elem *eie;
1456         int ret = 0;
1457
1458         for (eie = inode_list; eie; eie = eie->next) {
1459                 pr_debug("ref for %llu resolved, key (%llu EXTEND_DATA %llu), "
1460                          "root %llu\n", extent_item_objectid,
1461                          eie->inum, eie->offset, root);
1462                 ret = iterate(eie->inum, eie->offset, root, ctx);
1463                 if (ret) {
1464                         pr_debug("stopping iteration for %llu due to ret=%d\n",
1465                                  extent_item_objectid, ret);
1466                         break;
1467                 }
1468         }
1469
1470         return ret;
1471 }
1472
1473 /*
1474  * calls iterate() for every inode that references the extent identified by
1475  * the given parameters.
1476  * when the iterator function returns a non-zero value, iteration stops.
1477  */
1478 int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
1479                                 u64 extent_item_objectid, u64 extent_item_pos,
1480                                 int search_commit_root,
1481                                 iterate_extent_inodes_t *iterate, void *ctx)
1482 {
1483         int ret;
1484         struct btrfs_trans_handle *trans = NULL;
1485         struct ulist *refs = NULL;
1486         struct ulist *roots = NULL;
1487         struct ulist_node *ref_node = NULL;
1488         struct ulist_node *root_node = NULL;
1489         struct seq_list tree_mod_seq_elem = {};
1490         struct ulist_iterator ref_uiter;
1491         struct ulist_iterator root_uiter;
1492
1493         pr_debug("resolving all inodes for extent %llu\n",
1494                         extent_item_objectid);
1495
1496         if (!search_commit_root) {
1497                 trans = btrfs_join_transaction(fs_info->extent_root);
1498                 if (IS_ERR(trans))
1499                         return PTR_ERR(trans);
1500                 btrfs_get_tree_mod_seq(fs_info, &tree_mod_seq_elem);
1501         }
1502
1503         ret = btrfs_find_all_leafs(trans, fs_info, extent_item_objectid,
1504                                    tree_mod_seq_elem.seq, &refs,
1505                                    &extent_item_pos);
1506         if (ret)
1507                 goto out;
1508
1509         ULIST_ITER_INIT(&ref_uiter);
1510         while (!ret && (ref_node = ulist_next(refs, &ref_uiter))) {
1511                 ret = btrfs_find_all_roots(trans, fs_info, ref_node->val,
1512                                            tree_mod_seq_elem.seq, &roots);
1513                 if (ret)
1514                         break;
1515                 ULIST_ITER_INIT(&root_uiter);
1516                 while (!ret && (root_node = ulist_next(roots, &root_uiter))) {
1517                         pr_debug("root %llu references leaf %llu, data list "
1518                                  "%#llx\n", root_node->val, ref_node->val,
1519                                  (long long)ref_node->aux);
1520                         ret = iterate_leaf_refs((struct extent_inode_elem *)
1521                                                 (uintptr_t)ref_node->aux,
1522                                                 root_node->val,
1523                                                 extent_item_objectid,
1524                                                 iterate, ctx);
1525                 }
1526                 ulist_free(roots);
1527         }
1528
1529         free_leaf_list(refs);
1530 out:
1531         if (!search_commit_root) {
1532                 btrfs_put_tree_mod_seq(fs_info, &tree_mod_seq_elem);
1533                 btrfs_end_transaction(trans, fs_info->extent_root);
1534         }
1535
1536         return ret;
1537 }
1538
1539 int iterate_inodes_from_logical(u64 logical, struct btrfs_fs_info *fs_info,
1540                                 struct btrfs_path *path,
1541                                 iterate_extent_inodes_t *iterate, void *ctx)
1542 {
1543         int ret;
1544         u64 extent_item_pos;
1545         u64 flags = 0;
1546         struct btrfs_key found_key;
1547         int search_commit_root = path->search_commit_root;
1548
1549         ret = extent_from_logical(fs_info, logical, path, &found_key, &flags);
1550         btrfs_release_path(path);
1551         if (ret < 0)
1552                 return ret;
1553         if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)
1554                 return -EINVAL;
1555
1556         extent_item_pos = logical - found_key.objectid;
1557         ret = iterate_extent_inodes(fs_info, found_key.objectid,
1558                                         extent_item_pos, search_commit_root,
1559                                         iterate, ctx);
1560
1561         return ret;
1562 }
1563
1564 typedef int (iterate_irefs_t)(u64 parent, u32 name_len, unsigned long name_off,
1565                               struct extent_buffer *eb, void *ctx);
1566
1567 static int iterate_inode_refs(u64 inum, struct btrfs_root *fs_root,
1568                               struct btrfs_path *path,
1569                               iterate_irefs_t *iterate, void *ctx)
1570 {
1571         int ret = 0;
1572         int slot;
1573         u32 cur;
1574         u32 len;
1575         u32 name_len;
1576         u64 parent = 0;
1577         int found = 0;
1578         struct extent_buffer *eb;
1579         struct btrfs_item *item;
1580         struct btrfs_inode_ref *iref;
1581         struct btrfs_key found_key;
1582
1583         while (!ret) {
1584                 path->leave_spinning = 1;
1585                 ret = inode_ref_info(inum, parent ? parent+1 : 0, fs_root, path,
1586                                      &found_key);
1587                 if (ret < 0)
1588                         break;
1589                 if (ret) {
1590                         ret = found ? 0 : -ENOENT;
1591                         break;
1592                 }
1593                 ++found;
1594
1595                 parent = found_key.offset;
1596                 slot = path->slots[0];
1597                 eb = path->nodes[0];
1598                 /* make sure we can use eb after releasing the path */
1599                 atomic_inc(&eb->refs);
1600                 btrfs_tree_read_lock(eb);
1601                 btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1602                 btrfs_release_path(path);
1603
1604                 item = btrfs_item_nr(eb, slot);
1605                 iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
1606
1607                 for (cur = 0; cur < btrfs_item_size(eb, item); cur += len) {
1608                         name_len = btrfs_inode_ref_name_len(eb, iref);
1609                         /* path must be released before calling iterate()! */
1610                         pr_debug("following ref at offset %u for inode %llu in "
1611                                  "tree %llu\n", cur,
1612                                  (unsigned long long)found_key.objectid,
1613                                  (unsigned long long)fs_root->objectid);
1614                         ret = iterate(parent, name_len,
1615                                       (unsigned long)(iref + 1), eb, ctx);
1616                         if (ret)
1617                                 break;
1618                         len = sizeof(*iref) + name_len;
1619                         iref = (struct btrfs_inode_ref *)((char *)iref + len);
1620                 }
1621                 btrfs_tree_read_unlock_blocking(eb);
1622                 free_extent_buffer(eb);
1623         }
1624
1625         btrfs_release_path(path);
1626
1627         return ret;
1628 }
1629
1630 static int iterate_inode_extrefs(u64 inum, struct btrfs_root *fs_root,
1631                                  struct btrfs_path *path,
1632                                  iterate_irefs_t *iterate, void *ctx)
1633 {
1634         int ret;
1635         int slot;
1636         u64 offset = 0;
1637         u64 parent;
1638         int found = 0;
1639         struct extent_buffer *eb;
1640         struct btrfs_inode_extref *extref;
1641         struct extent_buffer *leaf;
1642         u32 item_size;
1643         u32 cur_offset;
1644         unsigned long ptr;
1645
1646         while (1) {
1647                 ret = btrfs_find_one_extref(fs_root, inum, offset, path, &extref,
1648                                             &offset);
1649                 if (ret < 0)
1650                         break;
1651                 if (ret) {
1652                         ret = found ? 0 : -ENOENT;
1653                         break;
1654                 }
1655                 ++found;
1656
1657                 slot = path->slots[0];
1658                 eb = path->nodes[0];
1659                 /* make sure we can use eb after releasing the path */
1660                 atomic_inc(&eb->refs);
1661
1662                 btrfs_tree_read_lock(eb);
1663                 btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1664                 btrfs_release_path(path);
1665
1666                 leaf = path->nodes[0];
1667                 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1668                 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
1669                 cur_offset = 0;
1670
1671                 while (cur_offset < item_size) {
1672                         u32 name_len;
1673
1674                         extref = (struct btrfs_inode_extref *)(ptr + cur_offset);
1675                         parent = btrfs_inode_extref_parent(eb, extref);
1676                         name_len = btrfs_inode_extref_name_len(eb, extref);
1677                         ret = iterate(parent, name_len,
1678                                       (unsigned long)&extref->name, eb, ctx);
1679                         if (ret)
1680                                 break;
1681
1682                         cur_offset += btrfs_inode_extref_name_len(leaf, extref);
1683                         cur_offset += sizeof(*extref);
1684                 }
1685                 btrfs_tree_read_unlock_blocking(eb);
1686                 free_extent_buffer(eb);
1687
1688                 offset++;
1689         }
1690
1691         btrfs_release_path(path);
1692
1693         return ret;
1694 }
1695
1696 static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
1697                          struct btrfs_path *path, iterate_irefs_t *iterate,
1698                          void *ctx)
1699 {
1700         int ret;
1701         int found_refs = 0;
1702
1703         ret = iterate_inode_refs(inum, fs_root, path, iterate, ctx);
1704         if (!ret)
1705                 ++found_refs;
1706         else if (ret != -ENOENT)
1707                 return ret;
1708
1709         ret = iterate_inode_extrefs(inum, fs_root, path, iterate, ctx);
1710         if (ret == -ENOENT && found_refs)
1711                 return 0;
1712
1713         return ret;
1714 }
1715
1716 /*
1717  * returns 0 if the path could be dumped (probably truncated)
1718  * returns <0 in case of an error
1719  */
1720 static int inode_to_path(u64 inum, u32 name_len, unsigned long name_off,
1721                          struct extent_buffer *eb, void *ctx)
1722 {
1723         struct inode_fs_paths *ipath = ctx;
1724         char *fspath;
1725         char *fspath_min;
1726         int i = ipath->fspath->elem_cnt;
1727         const int s_ptr = sizeof(char *);
1728         u32 bytes_left;
1729
1730         bytes_left = ipath->fspath->bytes_left > s_ptr ?
1731                                         ipath->fspath->bytes_left - s_ptr : 0;
1732
1733         fspath_min = (char *)ipath->fspath->val + (i + 1) * s_ptr;
1734         fspath = btrfs_ref_to_path(ipath->fs_root, ipath->btrfs_path, name_len,
1735                                    name_off, eb, inum, fspath_min, bytes_left);
1736         if (IS_ERR(fspath))
1737                 return PTR_ERR(fspath);
1738
1739         if (fspath > fspath_min) {
1740                 ipath->fspath->val[i] = (u64)(unsigned long)fspath;
1741                 ++ipath->fspath->elem_cnt;
1742                 ipath->fspath->bytes_left = fspath - fspath_min;
1743         } else {
1744                 ++ipath->fspath->elem_missed;
1745                 ipath->fspath->bytes_missing += fspath_min - fspath;
1746                 ipath->fspath->bytes_left = 0;
1747         }
1748
1749         return 0;
1750 }
1751
1752 /*
1753  * this dumps all file system paths to the inode into the ipath struct, provided
1754  * is has been created large enough. each path is zero-terminated and accessed
1755  * from ipath->fspath->val[i].
1756  * when it returns, there are ipath->fspath->elem_cnt number of paths available
1757  * in ipath->fspath->val[]. when the allocated space wasn't sufficient, the
1758  * number of missed paths in recored in ipath->fspath->elem_missed, otherwise,
1759  * it's zero. ipath->fspath->bytes_missing holds the number of bytes that would
1760  * have been needed to return all paths.
1761  */
1762 int paths_from_inode(u64 inum, struct inode_fs_paths *ipath)
1763 {
1764         return iterate_irefs(inum, ipath->fs_root, ipath->btrfs_path,
1765                              inode_to_path, ipath);
1766 }
1767
1768 struct btrfs_data_container *init_data_container(u32 total_bytes)
1769 {
1770         struct btrfs_data_container *data;
1771         size_t alloc_bytes;
1772
1773         alloc_bytes = max_t(size_t, total_bytes, sizeof(*data));
1774         data = vmalloc(alloc_bytes);
1775         if (!data)
1776                 return ERR_PTR(-ENOMEM);
1777
1778         if (total_bytes >= sizeof(*data)) {
1779                 data->bytes_left = total_bytes - sizeof(*data);
1780                 data->bytes_missing = 0;
1781         } else {
1782                 data->bytes_missing = sizeof(*data) - total_bytes;
1783                 data->bytes_left = 0;
1784         }
1785
1786         data->elem_cnt = 0;
1787         data->elem_missed = 0;
1788
1789         return data;
1790 }
1791
1792 /*
1793  * allocates space to return multiple file system paths for an inode.
1794  * total_bytes to allocate are passed, note that space usable for actual path
1795  * information will be total_bytes - sizeof(struct inode_fs_paths).
1796  * the returned pointer must be freed with free_ipath() in the end.
1797  */
1798 struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root,
1799                                         struct btrfs_path *path)
1800 {
1801         struct inode_fs_paths *ifp;
1802         struct btrfs_data_container *fspath;
1803
1804         fspath = init_data_container(total_bytes);
1805         if (IS_ERR(fspath))
1806                 return (void *)fspath;
1807
1808         ifp = kmalloc(sizeof(*ifp), GFP_NOFS);
1809         if (!ifp) {
1810                 kfree(fspath);
1811                 return ERR_PTR(-ENOMEM);
1812         }
1813
1814         ifp->btrfs_path = path;
1815         ifp->fspath = fspath;
1816         ifp->fs_root = fs_root;
1817
1818         return ifp;
1819 }
1820
1821 void free_ipath(struct inode_fs_paths *ipath)
1822 {
1823         if (!ipath)
1824                 return;
1825         vfree(ipath->fspath);
1826         kfree(ipath);
1827 }