fs/btrfs/extent-tree.c

   1 /*
   2  * Copyright (C) 2007 Oracle.  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 #include <linux/sched.h>
  19 #include <linux/pagemap.h>
  20 #include <linux/writeback.h>
  21 #include <linux/blkdev.h>
  22 #include <linux/sort.h>
  23 #include <linux/rcupdate.h>
  24 #include "compat.h"
  25 #include "hash.h"
  26 #include "ctree.h"
  27 #include "disk-io.h"
  28 #include "print-tree.h"
  29 #include "transaction.h"
  30 #include "volumes.h"
  31 #include "locking.h"
  32 #include "free-space-cache.h"
  33
  34 static int update_reserved_extents(struct btrfs_root *root,
  35                                    u64 bytenr, u64 num, int reserve);
  36 static int update_block_group(struct btrfs_trans_handle *trans,
  37                               struct btrfs_root *root,
  38                               u64 bytenr, u64 num_bytes, int alloc,
  39                               int mark_free);
  40 static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
  41                                 struct btrfs_root *root,
  42                                 u64 bytenr, u64 num_bytes, u64 parent,
  43                                 u64 root_objectid, u64 owner_objectid,
  44                                 u64 owner_offset, int refs_to_drop,
  45                                 struct btrfs_delayed_extent_op *extra_op);
  46 static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op,
  47                                     struct extent_buffer *leaf,
  48                                     struct btrfs_extent_item *ei);
  49 static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
  50                                       struct btrfs_root *root,
  51                                       u64 parent, u64 root_objectid,
  52                                       u64 flags, u64 owner, u64 offset,
  53                                       struct btrfs_key *ins, int ref_mod);
  54 static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
  55                                      struct btrfs_root *root,
  56                                      u64 parent, u64 root_objectid,
  57                                      u64 flags, struct btrfs_disk_key *key,
  58                                      int level, struct btrfs_key *ins);
  59
  60 static int do_chunk_alloc(struct btrfs_trans_handle *trans,
  61                           struct btrfs_root *extent_root, u64 alloc_bytes,
  62                           u64 flags, int force);
  63
  64 static int block_group_bits(struct btrfs_block_group_cache *cache, u64 bits)
  65 {
  66         return (cache->flags & bits) == bits;
  67 }
  68
  69 /*
  70  * this adds the block group to the fs_info rb tree for the block group
  71  * cache
  72  */
  73 static int btrfs_add_block_group_cache(struct btrfs_fs_info *info,
  74                                 struct btrfs_block_group_cache *block_group)
  75 {
  76         struct rb_node **p;
  77         struct rb_node *parent = NULL;
  78         struct btrfs_block_group_cache *cache;
  79
  80         spin_lock(&info->block_group_cache_lock);
  81         p = &info->block_group_cache_tree.rb_node;
  82
  83         while (*p) {
  84                 parent = *p;
  85                 cache = rb_entry(parent, struct btrfs_block_group_cache,
  86                                  cache_node);
  87                 if (block_group->key.objectid < cache->key.objectid) {
  88                         p = &(*p)->rb_left;
  89                 } else if (block_group->key.objectid > cache->key.objectid) {
  90                         p = &(*p)->rb_right;
  91                 } else {
  92                         spin_unlock(&info->block_group_cache_lock);
  93                         return -EEXIST;
  94                 }
  95         }
  96
  97         rb_link_node(&block_group->cache_node, parent, p);
  98         rb_insert_color(&block_group->cache_node,
  99                         &info->block_group_cache_tree);
 100         spin_unlock(&info->block_group_cache_lock);
 101
 102         return 0;
 103 }
 104
 105 /*
 106  * This will return the block group at or after bytenr if contains is 0, else
 107  * it will return the block group that contains the bytenr
 108  */
 109 static struct btrfs_block_group_cache *
 110 block_group_cache_tree_search(struct btrfs_fs_info *info, u64 bytenr,
 111                               int contains)
 112 {
 113         struct btrfs_block_group_cache *cache, *ret = NULL;
 114         struct rb_node *n;
 115         u64 end, start;
 116
 117         spin_lock(&info->block_group_cache_lock);
 118         n = info->block_group_cache_tree.rb_node;
 119
 120         while (n) {
 121                 cache = rb_entry(n, struct btrfs_block_group_cache,
 122                                  cache_node);
 123                 end = cache->key.objectid + cache->key.offset - 1;
 124                 start = cache->key.objectid;
 125
 126                 if (bytenr < start) {
 127                         if (!contains && (!ret || start < ret->key.objectid))
 128                                 ret = cache;
 129                         n = n->rb_left;
 130                 } else if (bytenr > start) {
 131                         if (contains && bytenr <= end) {
 132                                 ret = cache;
 133                                 break;
 134                         }
 135                         n = n->rb_right;
 136                 } else {
 137                         ret = cache;
 138                         break;
 139                 }
 140         }
 141         if (ret)
 142                 atomic_inc(&ret->count);
 143         spin_unlock(&info->block_group_cache_lock);
 144
 145         return ret;
 146 }
 147
 148 /*
 149  * this is only called by cache_block_group, since we could have freed extents
 150  * we need to check the pinned_extents for any extents that can't be used yet
 151  * since their free space will be released as soon as the transaction commits.
 152  */
 153 static int add_new_free_space(struct btrfs_block_group_cache *block_group,
 154                               struct btrfs_fs_info *info, u64 start, u64 end)
 155 {
 156         u64 extent_start, extent_end, size;
 157         int ret;
 158
 159         while (start < end) {
 160                 ret = find_first_extent_bit(&info->pinned_extents, start,
 161                                             &extent_start, &extent_end,
 162                                             EXTENT_DIRTY);
 163                 if (ret)
 164                         break;
 165
 166                 if (extent_start == start) {
 167                         start = extent_end + 1;
 168                 } else if (extent_start > start && extent_start < end) {
 169                         size = extent_start - start;
 170                         ret = btrfs_add_free_space(block_group, start,
 171                                                    size);
 172                         BUG_ON(ret);
 173                         start = extent_end + 1;
 174                 } else {
 175                         break;
 176                 }
 177         }
 178
 179         if (start < end) {
 180                 size = end - start;
 181                 ret = btrfs_add_free_space(block_group, start, size);
 182                 BUG_ON(ret);
 183         }
 184
 185         return 0;
 186 }
 187
 188 static int remove_sb_from_cache(struct btrfs_root *root,
 189                                 struct btrfs_block_group_cache *cache)
 190 {
 191         u64 bytenr;
 192         u64 *logical;
 193         int stripe_len;
 194         int i, nr, ret;
 195
 196         for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) {
 197                 bytenr = btrfs_sb_offset(i);
 198                 ret = btrfs_rmap_block(&root->fs_info->mapping_tree,
 199                                        cache->key.objectid, bytenr, 0,
 200                                        &logical, &nr, &stripe_len);
 201                 BUG_ON(ret);
 202                 while (nr--) {
 203                         btrfs_remove_free_space(cache, logical[nr],
 204                                                 stripe_len);
 205                 }
 206                 kfree(logical);
 207         }
 208         return 0;
 209 }
 210
 211 static int cache_block_group(struct btrfs_root *root,
 212                              struct btrfs_block_group_cache *block_group)
 213 {
 214         struct btrfs_path *path;
 215         int ret = 0;
 216         struct btrfs_key key;
 217         struct extent_buffer *leaf;
 218         int slot;
 219         u64 last;
 220
 221         if (!block_group)
 222                 return 0;
 223
 224         root = root->fs_info->extent_root;
 225
 226         if (block_group->cached)
 227                 return 0;
 228
 229         path = btrfs_alloc_path();
 230         if (!path)
 231                 return -ENOMEM;
 232
 233         path->reada = 2;
 234         /*
 235          * we get into deadlocks with paths held by callers of this function.
 236          * since the alloc_mutex is protecting things right now, just
 237          * skip the locking here
 238          */
 239         path->skip_locking = 1;
 240         last = max_t(u64, block_group->key.objectid, BTRFS_SUPER_INFO_OFFSET);
 241         key.objectid = last;
 242         key.offset = 0;
 243         btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
 244         ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 245         if (ret < 0)
 246                 goto err;
 247
 248         while (1) {
 249                 leaf = path->nodes[0];
 250                 slot = path->slots[0];
 251                 if (slot >= btrfs_header_nritems(leaf)) {
 252                         ret = btrfs_next_leaf(root, path);
 253                         if (ret < 0)
 254                                 goto err;
 255                         if (ret == 0)
 256                                 continue;
 257                         else
 258                                 break;
 259                 }
 260                 btrfs_item_key_to_cpu(leaf, &key, slot);
 261                 if (key.objectid < block_group->key.objectid)
 262                         goto next;
 263
 264                 if (key.objectid >= block_group->key.objectid +
 265                     block_group->key.offset)
 266                         break;
 267
 268                 if (btrfs_key_type(&key) == BTRFS_EXTENT_ITEM_KEY) {
 269                         add_new_free_space(block_group, root->fs_info, last,
 270                                            key.objectid);
 271
 272                         last = key.objectid + key.offset;
 273                 }
 274 next:
 275                 path->slots[0]++;
 276         }
 277
 278         add_new_free_space(block_group, root->fs_info, last,
 279                            block_group->key.objectid +
 280                            block_group->key.offset);
 281
 282         block_group->cached = 1;
 283         remove_sb_from_cache(root, block_group);
 284         ret = 0;
 285 err:
 286         btrfs_free_path(path);
 287         return ret;
 288 }
 289
 290 /*
 291  * return the block group that starts at or after bytenr
 292  */
 293 static struct btrfs_block_group_cache *
 294 btrfs_lookup_first_block_group(struct btrfs_fs_info *info, u64 bytenr)
 295 {
 296         struct btrfs_block_group_cache *cache;
 297
 298         cache = block_group_cache_tree_search(info, bytenr, 0);
 299
 300         return cache;
 301 }
 302
 303 /*
 304  * return the block group that contains the given bytenr
 305  */
 306 struct btrfs_block_group_cache *btrfs_lookup_block_group(
 307                                                  struct btrfs_fs_info *info,
 308                                                  u64 bytenr)
 309 {
 310         struct btrfs_block_group_cache *cache;
 311
 312         cache = block_group_cache_tree_search(info, bytenr, 1);
 313
 314         return cache;
 315 }
 316
 317 void btrfs_put_block_group(struct btrfs_block_group_cache *cache)
 318 {
 319         if (atomic_dec_and_test(&cache->count))
 320                 kfree(cache);
 321 }
 322
 323 static struct btrfs_space_info *__find_space_info(struct btrfs_fs_info *info,
 324                                                   u64 flags)
 325 {
 326         struct list_head *head = &info->space_info;
 327         struct btrfs_space_info *found;
 328
 329         rcu_read_lock();
 330         list_for_each_entry_rcu(found, head, list) {
 331                 if (found->flags == flags) {
 332                         rcu_read_unlock();
 333                         return found;
 334                 }
 335         }
 336         rcu_read_unlock();
 337         return NULL;
 338 }
 339
 340 /*
 341  * after adding space to the filesystem, we need to clear the full flags
 342  * on all the space infos.
 343  */
 344 void btrfs_clear_space_info_full(struct btrfs_fs_info *info)
 345 {
 346         struct list_head *head = &info->space_info;
 347         struct btrfs_space_info *found;
 348
 349         rcu_read_lock();
 350         list_for_each_entry_rcu(found, head, list)
 351                 found->full = 0;
 352         rcu_read_unlock();
 353 }
 354
 355 static u64 div_factor(u64 num, int factor)
 356 {
 357         if (factor == 10)
 358                 return num;
 359         num *= factor;
 360         do_div(num, 10);
 361         return num;
 362 }
 363
 364 u64 btrfs_find_block_group(struct btrfs_root *root,
 365                            u64 search_start, u64 search_hint, int owner)
 366 {
 367         struct btrfs_block_group_cache *cache;
 368         u64 used;
 369         u64 last = max(search_hint, search_start);
 370         u64 group_start = 0;
 371         int full_search = 0;
 372         int factor = 9;
 373         int wrapped = 0;
 374 again:
 375         while (1) {
 376                 cache = btrfs_lookup_first_block_group(root->fs_info, last);
 377                 if (!cache)
 378                         break;
 379
 380                 spin_lock(&cache->lock);
 381                 last = cache->key.objectid + cache->key.offset;
 382                 used = btrfs_block_group_used(&cache->item);
 383
 384                 if ((full_search || !cache->ro) &&
 385                     block_group_bits(cache, BTRFS_BLOCK_GROUP_METADATA)) {
 386                         if (used + cache->pinned + cache->reserved <
 387                             div_factor(cache->key.offset, factor)) {
 388                                 group_start = cache->key.objectid;
 389                                 spin_unlock(&cache->lock);
 390                                 btrfs_put_block_group(cache);
 391                                 goto found;
 392                         }
 393                 }
 394                 spin_unlock(&cache->lock);
 395                 btrfs_put_block_group(cache);
 396                 cond_resched();
 397         }
 398         if (!wrapped) {
 399                 last = search_start;
 400                 wrapped = 1;
 401                 goto again;
 402         }
 403         if (!full_search && factor < 10) {
 404                 last = search_start;
 405                 full_search = 1;
 406                 factor = 10;
 407                 goto again;
 408         }
 409 found:
 410         return group_start;
 411 }
 412
 413 /* simple helper to search for an existing extent at a given offset */
 414 int btrfs_lookup_extent(struct btrfs_root *root, u64 start, u64 len)
 415 {
 416         int ret;
 417         struct btrfs_key key;
 418         struct btrfs_path *path;
 419
 420         path = btrfs_alloc_path();
 421         BUG_ON(!path);
 422         key.objectid = start;
 423         key.offset = len;
 424         btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
 425         ret = btrfs_search_slot(NULL, root->fs_info->extent_root, &key, path,
 426                                 0, 0);
 427         btrfs_free_path(path);
 428         return ret;
 429 }
 430
 431 /*
 432  * Back reference rules.  Back refs have three main goals:
 433  *
 434  * 1) differentiate between all holders of references to an extent so that
 435  *    when a reference is dropped we can make sure it was a valid reference
 436  *    before freeing the extent.
 437  *
 438  * 2) Provide enough information to quickly find the holders of an extent
 439  *    if we notice a given block is corrupted or bad.
 440  *
 441  * 3) Make it easy to migrate blocks for FS shrinking or storage pool
 442  *    maintenance.  This is actually the same as #2, but with a slightly
 443  *    different use case.
 444  *
 445  * There are two kinds of back refs. The implicit back refs is optimized
 446  * for pointers in non-shared tree blocks. For a given pointer in a block,
 447  * back refs of this kind provide information about the block's owner tree
 448  * and the pointer's key. These information allow us to find the block by
 449  * b-tree searching. The full back refs is for pointers in tree blocks not
 450  * referenced by their owner trees. The location of tree block is recorded
 451  * in the back refs. Actually the full back refs is generic, and can be
 452  * used in all cases the implicit back refs is used. The major shortcoming
 453  * of the full back refs is its overhead. Every time a tree block gets
 454  * COWed, we have to update back refs entry for all pointers in it.
 455  *
 456  * For a newly allocated tree block, we use implicit back refs for
 457  * pointers in it. This means most tree related operations only involve
 458  * implicit back refs. For a tree block created in old transaction, the
 459  * only way to drop a reference to it is COW it. So we can detect the
 460  * event that tree block loses its owner tree's reference and do the
 461  * back refs conversion.
 462  *
 463  * When a tree block is COW'd through a tree, there are four cases:
 464  *
 465  * The reference count of the block is one and the tree is the block's
 466  * owner tree. Nothing to do in this case.
 467  *
 468  * The reference count of the block is one and the tree is not the
 469  * block's owner tree. In this case, full back refs is used for pointers
 470  * in the block. Remove these full back refs, add implicit back refs for
 471  * every pointers in the new block.
 472  *
 473  * The reference count of the block is greater than one and the tree is
 474  * the block's owner tree. In this case, implicit back refs is used for
 475  * pointers in the block. Add full back refs for every pointers in the
 476  * block, increase lower level extents' reference counts. The original
 477  * implicit back refs are entailed to the new block.
 478  *
 479  * The reference count of the block is greater than one and the tree is
 480  * not the block's owner tree. Add implicit back refs for every pointer in
 481  * the new block, increase lower level extents' reference count.
 482  *
 483  * Back Reference Key composing:
 484  *
 485  * The key objectid corresponds to the first byte in the extent,
 486  * The key type is used to differentiate between types of back refs.
 487  * There are different meanings of the key offset for different types
 488  * of back refs.
 489  *
 490  * File extents can be referenced by:
 491  *
 492  * - multiple snapshots, subvolumes, or different generations in one subvol
 493  * - different files inside a single subvolume
 494  * - different offsets inside a file (bookend extents in file.c)
 495  *
 496  * The extent ref structure for the implicit back refs has fields for:
 497  *
 498  * - Objectid of the subvolume root
 499  * - objectid of the file holding the reference
 500  * - original offset in the file
 501  * - how many bookend extents
 502  *
 503  * The key offset for the implicit back refs is hash of the first
 504  * three fields.
 505  *
 506  * The extent ref structure for the full back refs has field for:
 507  *
 508  * - number of pointers in the tree leaf
 509  *
 510  * The key offset for the implicit back refs is the first byte of
 511  * the tree leaf
 512  *
 513  * When a file extent is allocated, The implicit back refs is used.
 514  * the fields are filled in:
 515  *
 516  *     (root_key.objectid, inode objectid, offset in file, 1)
 517  *
 518  * When a file extent is removed file truncation, we find the
 519  * corresponding implicit back refs and check the following fields:
 520  *
 521  *     (btrfs_header_owner(leaf), inode objectid, offset in file)
 522  *
 523  * Btree extents can be referenced by:
 524  *
 525  * - Different subvolumes
 526  *
 527  * Both the implicit back refs and the full back refs for tree blocks
 528  * only consist of key. The key offset for the implicit back refs is
 529  * objectid of block's owner tree. The key offset for the full back refs
 530  * is the first byte of parent block.
 531  *
 532  * When implicit back refs is used, information about the lowest key and
 533  * level of the tree block are required. These information are stored in
 534  * tree block info structure.
 535  */
 536
 537 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
 538 static int convert_extent_item_v0(struct btrfs_trans_handle *trans,
 539                                   struct btrfs_root *root,
 540                                   struct btrfs_path *path,
 541                                   u64 owner, u32 extra_size)
 542 {
 543         struct btrfs_extent_item *item;
 544         struct btrfs_extent_item_v0 *ei0;
 545         struct btrfs_extent_ref_v0 *ref0;
 546         struct btrfs_tree_block_info *bi;
 547         struct extent_buffer *leaf;
 548         struct btrfs_key key;
 549         struct btrfs_key found_key;
 550         u32 new_size = sizeof(*item);
 551         u64 refs;
 552         int ret;
 553
 554         leaf = path->nodes[0];
 555         BUG_ON(btrfs_item_size_nr(leaf, path->slots[0]) != sizeof(*ei0));
 556
 557         btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
 558         ei0 = btrfs_item_ptr(leaf, path->slots[0],
 559                              struct btrfs_extent_item_v0);
 560         refs = btrfs_extent_refs_v0(leaf, ei0);
 561
 562         if (owner == (u64)-1) {
 563                 while (1) {
 564                         if (path->slots[0] >= btrfs_header_nritems(leaf)) {
 565                                 ret = btrfs_next_leaf(root, path);
 566                                 if (ret < 0)
 567                                         return ret;
 568                                 BUG_ON(ret > 0);
 569                                 leaf = path->nodes[0];
 570                         }
 571                         btrfs_item_key_to_cpu(leaf, &found_key,
 572                                               path->slots[0]);
 573                         BUG_ON(key.objectid != found_key.objectid);
 574                         if (found_key.type != BTRFS_EXTENT_REF_V0_KEY) {
 575                                 path->slots[0]++;
 576                                 continue;
 577                         }
 578                         ref0 = btrfs_item_ptr(leaf, path->slots[0],
 579                                               struct btrfs_extent_ref_v0);
 580                         owner = btrfs_ref_objectid_v0(leaf, ref0);
 581                         break;
 582                 }
 583         }
 584         btrfs_release_path(root, path);
 585
 586         if (owner < BTRFS_FIRST_FREE_OBJECTID)
 587                 new_size += sizeof(*bi);
 588
 589         new_size -= sizeof(*ei0);
 590         ret = btrfs_search_slot(trans, root, &key, path,
 591                                 new_size + extra_size, 1);
 592         if (ret < 0)
 593                 return ret;
 594         BUG_ON(ret);
 595
 596         ret = btrfs_extend_item(trans, root, path, new_size);
 597         BUG_ON(ret);
 598
 599         leaf = path->nodes[0];
 600         item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
 601         btrfs_set_extent_refs(leaf, item, refs);
 602         /* FIXME: get real generation */
 603         btrfs_set_extent_generation(leaf, item, 0);
 604         if (owner < BTRFS_FIRST_FREE_OBJECTID) {
 605                 btrfs_set_extent_flags(leaf, item,
 606                                        BTRFS_EXTENT_FLAG_TREE_BLOCK |
 607                                        BTRFS_BLOCK_FLAG_FULL_BACKREF);
 608                 bi = (struct btrfs_tree_block_info *)(item + 1);
 609                 /* FIXME: get first key of the block */
 610                 memset_extent_buffer(leaf, 0, (unsigned long)bi, sizeof(*bi));
 611                 btrfs_set_tree_block_level(leaf, bi, (int)owner);
 612         } else {
 613                 btrfs_set_extent_flags(leaf, item, BTRFS_EXTENT_FLAG_DATA);
 614         }
 615         btrfs_mark_buffer_dirty(leaf);
 616         return 0;
 617 }
 618 #endif
 619
 620 static u64 hash_extent_data_ref(u64 root_objectid, u64 owner, u64 offset)
 621 {
 622         u32 high_crc = ~(u32)0;
 623         u32 low_crc = ~(u32)0;
 624         __le64 lenum;
 625
 626         lenum = cpu_to_le64(root_objectid);
 627         high_crc = crc32c(high_crc, &lenum, sizeof(lenum));
 628         lenum = cpu_to_le64(owner);
 629         low_crc = crc32c(low_crc, &lenum, sizeof(lenum));
 630         lenum = cpu_to_le64(offset);
 631         low_crc = crc32c(low_crc, &lenum, sizeof(lenum));
 632
 633         return ((u64)high_crc << 31) ^ (u64)low_crc;
 634 }
 635
 636 static u64 hash_extent_data_ref_item(struct extent_buffer *leaf,
 637                                      struct btrfs_extent_data_ref *ref)
 638 {
 639         return hash_extent_data_ref(btrfs_extent_data_ref_root(leaf, ref),
 640                                     btrfs_extent_data_ref_objectid(leaf, ref),
 641                                     btrfs_extent_data_ref_offset(leaf, ref));
 642 }
 643
 644 static int match_extent_data_ref(struct extent_buffer *leaf,
 645                                  struct btrfs_extent_data_ref *ref,
 646                                  u64 root_objectid, u64 owner, u64 offset)
 647 {
 648         if (btrfs_extent_data_ref_root(leaf, ref) != root_objectid ||
 649             btrfs_extent_data_ref_objectid(leaf, ref) != owner ||
 650             btrfs_extent_data_ref_offset(leaf, ref) != offset)
 651                 return 0;
 652         return 1;
 653 }
 654
 655 static noinline int lookup_extent_data_ref(struct btrfs_trans_handle *trans,
 656                                            struct btrfs_root *root,
 657                                            struct btrfs_path *path,
 658                                            u64 bytenr, u64 parent,
 659                                            u64 root_objectid,
 660                                            u64 owner, u64 offset)
 661 {
 662         struct btrfs_key key;
 663         struct btrfs_extent_data_ref *ref;
 664         struct extent_buffer *leaf;
 665         u32 nritems;
 666         int ret;
 667         int recow;
 668         int err = -ENOENT;
 669
 670         key.objectid = bytenr;
 671         if (parent) {
 672                 key.type = BTRFS_SHARED_DATA_REF_KEY;
 673                 key.offset = parent;
 674         } else {
 675                 key.type = BTRFS_EXTENT_DATA_REF_KEY;
 676                 key.offset = hash_extent_data_ref(root_objectid,
 677                                                   owner, offset);
 678         }
 679 again:
 680         recow = 0;
 681         ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 682         if (ret < 0) {
 683                 err = ret;
 684                 goto fail;
 685         }
 686
 687         if (parent) {
 688                 if (!ret)
 689                         return 0;
 690 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
 691                 key.type = BTRFS_EXTENT_REF_V0_KEY;
 692                 btrfs_release_path(root, path);
 693                 ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 694                 if (ret < 0) {
 695                         err = ret;
 696                         goto fail;
 697                 }
 698                 if (!ret)
 699                         return 0;
 700 #endif
 701                 goto fail;
 702         }
 703
 704         leaf = path->nodes[0];
 705         nritems = btrfs_header_nritems(leaf);
 706         while (1) {
 707                 if (path->slots[0] >= nritems) {
 708                         ret = btrfs_next_leaf(root, path);
 709                         if (ret < 0)
 710                                 err = ret;
 711                         if (ret)
 712                                 goto fail;
 713
 714                         leaf = path->nodes[0];
 715                         nritems = btrfs_header_nritems(leaf);
 716                         recow = 1;
 717                 }
 718
 719                 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
 720                 if (key.objectid != bytenr ||
 721                     key.type != BTRFS_EXTENT_DATA_REF_KEY)
 722                         goto fail;
 723
 724                 ref = btrfs_item_ptr(leaf, path->slots[0],
 725                                      struct btrfs_extent_data_ref);
 726
 727                 if (match_extent_data_ref(leaf, ref, root_objectid,
 728                                           owner, offset)) {
 729                         if (recow) {
 730                                 btrfs_release_path(root, path);
 731                                 goto again;
 732                         }
 733                         err = 0;
 734                         break;
 735                 }
 736                 path->slots[0]++;
 737         }
 738 fail:
 739         return err;
 740 }
 741
 742 static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans,
 743                                            struct btrfs_root *root,
 744                                            struct btrfs_path *path,
 745                                            u64 bytenr, u64 parent,
 746                                            u64 root_objectid, u64 owner,
 747                                            u64 offset, int refs_to_add)
 748 {
 749         struct btrfs_key key;
 750         struct extent_buffer *leaf;
 751         u32 size;
 752         u32 num_refs;
 753         int ret;
 754
 755         key.objectid = bytenr;
 756         if (parent) {
 757                 key.type = BTRFS_SHARED_DATA_REF_KEY;
 758                 key.offset = parent;
 759                 size = sizeof(struct btrfs_shared_data_ref);
 760         } else {
 761                 key.type = BTRFS_EXTENT_DATA_REF_KEY;
 762                 key.offset = hash_extent_data_ref(root_objectid,
 763                                                   owner, offset);
 764                 size = sizeof(struct btrfs_extent_data_ref);
 765         }
 766
 767         ret = btrfs_insert_empty_item(trans, root, path, &key, size);
 768         if (ret && ret != -EEXIST)
 769                 goto fail;
 770
 771         leaf = path->nodes[0];
 772         if (parent) {
 773                 struct btrfs_shared_data_ref *ref;
 774                 ref = btrfs_item_ptr(leaf, path->slots[0],
 775                                      struct btrfs_shared_data_ref);
 776                 if (ret == 0) {
 777                         btrfs_set_shared_data_ref_count(leaf, ref, refs_to_add);
 778                 } else {
 779                         num_refs = btrfs_shared_data_ref_count(leaf, ref);
 780                         num_refs += refs_to_add;
 781                         btrfs_set_shared_data_ref_count(leaf, ref, num_refs);
 782                 }
 783         } else {
 784                 struct btrfs_extent_data_ref *ref;
 785                 while (ret == -EEXIST) {
 786                         ref = btrfs_item_ptr(leaf, path->slots[0],
 787                                              struct btrfs_extent_data_ref);
 788                         if (match_extent_data_ref(leaf, ref, root_objectid,
 789                                                   owner, offset))
 790                                 break;
 791                         btrfs_release_path(root, path);
 792                         key.offset++;
 793                         ret = btrfs_insert_empty_item(trans, root, path, &key,
 794                                                       size);
 795                         if (ret && ret != -EEXIST)
 796                                 goto fail;
 797
 798                         leaf = path->nodes[0];
 799                 }
 800                 ref = btrfs_item_ptr(leaf, path->slots[0],
 801                                      struct btrfs_extent_data_ref);
 802                 if (ret == 0) {
 803                         btrfs_set_extent_data_ref_root(leaf, ref,
 804                                                        root_objectid);
 805                         btrfs_set_extent_data_ref_objectid(leaf, ref, owner);
 806                         btrfs_set_extent_data_ref_offset(leaf, ref, offset);
 807                         btrfs_set_extent_data_ref_count(leaf, ref, refs_to_add);
 808                 } else {
 809                         num_refs = btrfs_extent_data_ref_count(leaf, ref);
 810                         num_refs += refs_to_add;
 811                         btrfs_set_extent_data_ref_count(leaf, ref, num_refs);
 812                 }
 813         }
 814         btrfs_mark_buffer_dirty(leaf);
 815         ret = 0;
 816 fail:
 817         btrfs_release_path(root, path);
 818         return ret;
 819 }
 820
 821 static noinline int remove_extent_data_ref(struct btrfs_trans_handle *trans,
 822                                            struct btrfs_root *root,
 823                                            struct btrfs_path *path,
 824                                            int refs_to_drop)
 825 {
 826         struct btrfs_key key;
 827         struct btrfs_extent_data_ref *ref1 = NULL;
 828         struct btrfs_shared_data_ref *ref2 = NULL;
 829         struct extent_buffer *leaf;
 830         u32 num_refs = 0;
 831         int ret = 0;
 832
 833         leaf = path->nodes[0];
 834         btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
 835
 836         if (key.type == BTRFS_EXTENT_DATA_REF_KEY) {
 837                 ref1 = btrfs_item_ptr(leaf, path->slots[0],
 838                                       struct btrfs_extent_data_ref);
 839                 num_refs = btrfs_extent_data_ref_count(leaf, ref1);
 840         } else if (key.type == BTRFS_SHARED_DATA_REF_KEY) {
 841                 ref2 = btrfs_item_ptr(leaf, path->slots[0],
 842                                       struct btrfs_shared_data_ref);
 843                 num_refs = btrfs_shared_data_ref_count(leaf, ref2);
 844 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
 845         } else if (key.type == BTRFS_EXTENT_REF_V0_KEY) {
 846                 struct btrfs_extent_ref_v0 *ref0;
 847                 ref0 = btrfs_item_ptr(leaf, path->slots[0],
 848                                       struct btrfs_extent_ref_v0);
 849                 num_refs = btrfs_ref_count_v0(leaf, ref0);
 850 #endif
 851         } else {
 852                 BUG();
 853         }
 854
 855         BUG_ON(num_refs < refs_to_drop);
 856         num_refs -= refs_to_drop;
 857
 858         if (num_refs == 0) {
 859                 ret = btrfs_del_item(trans, root, path);
 860         } else {
 861                 if (key.type == BTRFS_EXTENT_DATA_REF_KEY)
 862                         btrfs_set_extent_data_ref_count(leaf, ref1, num_refs);
 863                 else if (key.type == BTRFS_SHARED_DATA_REF_KEY)
 864                         btrfs_set_shared_data_ref_count(leaf, ref2, num_refs);
 865 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
 866                 else {
 867                         struct btrfs_extent_ref_v0 *ref0;
 868                         ref0 = btrfs_item_ptr(leaf, path->slots[0],
 869                                         struct btrfs_extent_ref_v0);
 870                         btrfs_set_ref_count_v0(leaf, ref0, num_refs);
 871                 }
 872 #endif
 873                 btrfs_mark_buffer_dirty(leaf);
 874         }
 875         return ret;
 876 }
 877
 878 static noinline u32 extent_data_ref_count(struct btrfs_root *root,
 879                                           struct btrfs_path *path,
 880                                           struct btrfs_extent_inline_ref *iref)
 881 {
 882         struct btrfs_key key;
 883         struct extent_buffer *leaf;
 884         struct btrfs_extent_data_ref *ref1;
 885         struct btrfs_shared_data_ref *ref2;
 886         u32 num_refs = 0;
 887
 888         leaf = path->nodes[0];
 889         btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
 890         if (iref) {
 891                 if (btrfs_extent_inline_ref_type(leaf, iref) ==
 892                     BTRFS_EXTENT_DATA_REF_KEY) {
 893                         ref1 = (struct btrfs_extent_data_ref *)(&iref->offset);
 894                         num_refs = btrfs_extent_data_ref_count(leaf, ref1);
 895                 } else {
 896                         ref2 = (struct btrfs_shared_data_ref *)(iref + 1);
 897                         num_refs = btrfs_shared_data_ref_count(leaf, ref2);
 898                 }
 899         } else if (key.type == BTRFS_EXTENT_DATA_REF_KEY) {
 900                 ref1 = btrfs_item_ptr(leaf, path->slots[0],
 901                                       struct btrfs_extent_data_ref);
 902                 num_refs = btrfs_extent_data_ref_count(leaf, ref1);
 903         } else if (key.type == BTRFS_SHARED_DATA_REF_KEY) {
 904                 ref2 = btrfs_item_ptr(leaf, path->slots[0],
 905                                       struct btrfs_shared_data_ref);
 906                 num_refs = btrfs_shared_data_ref_count(leaf, ref2);
 907 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
 908         } else if (key.type == BTRFS_EXTENT_REF_V0_KEY) {
 909                 struct btrfs_extent_ref_v0 *ref0;
 910                 ref0 = btrfs_item_ptr(leaf, path->slots[0],
 911                                       struct btrfs_extent_ref_v0);
 912                 num_refs = btrfs_ref_count_v0(leaf, ref0);
 913 #endif
 914         } else {
 915                 WARN_ON(1);
 916         }
 917         return num_refs;
 918 }
 919
 920 static noinline int lookup_tree_block_ref(struct btrfs_trans_handle *trans,
 921                                           struct btrfs_root *root,
 922                                           struct btrfs_path *path,
 923                                           u64 bytenr, u64 parent,
 924                                           u64 root_objectid)
 925 {
 926         struct btrfs_key key;
 927         int ret;
 928
 929         key.objectid = bytenr;
 930         if (parent) {
 931                 key.type = BTRFS_SHARED_BLOCK_REF_KEY;
 932                 key.offset = parent;
 933         } else {
 934                 key.type = BTRFS_TREE_BLOCK_REF_KEY;
 935                 key.offset = root_objectid;
 936         }
 937
 938         ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 939         if (ret > 0)
 940                 ret = -ENOENT;
 941 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
 942         if (ret == -ENOENT && parent) {
 943                 btrfs_release_path(root, path);
 944                 key.type = BTRFS_EXTENT_REF_V0_KEY;
 945                 ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 946                 if (ret > 0)
 947                         ret = -ENOENT;
 948         }
 949 #endif
 950         return ret;
 951 }
 952
 953 static noinline int insert_tree_block_ref(struct btrfs_trans_handle *trans,
 954                                           struct btrfs_root *root,
 955                                           struct btrfs_path *path,
 956                                           u64 bytenr, u64 parent,
 957                                           u64 root_objectid)
 958 {
 959         struct btrfs_key key;
 960         int ret;
 961
 962         key.objectid = bytenr;
 963         if (parent) {
 964                 key.type = BTRFS_SHARED_BLOCK_REF_KEY;
 965                 key.offset = parent;
 966         } else {
 967                 key.type = BTRFS_TREE_BLOCK_REF_KEY;
 968                 key.offset = root_objectid;
 969         }
 970
 971         ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
 972         btrfs_release_path(root, path);
 973         return ret;
 974 }
 975
 976 static inline int extent_ref_type(u64 parent, u64 owner)
 977 {
 978         int type;
 979         if (owner < BTRFS_FIRST_FREE_OBJECTID) {
 980                 if (parent > 0)
 981                         type = BTRFS_SHARED_BLOCK_REF_KEY;
 982                 else
 983                         type = BTRFS_TREE_BLOCK_REF_KEY;
 984         } else {
 985                 if (parent > 0)
 986                         type = BTRFS_SHARED_DATA_REF_KEY;
 987                 else
 988                         type = BTRFS_EXTENT_DATA_REF_KEY;
 989         }
 990         return type;
 991 }
 992
 993 static int find_next_key(struct btrfs_path *path, int level,
 994                          struct btrfs_key *key)
 995
 996 {
 997         for (; level < BTRFS_MAX_LEVEL; level++) {
 998                 if (!path->nodes[level])
 999                         break;
1000                 if (path->slots[level] + 1 >=
1001                     btrfs_header_nritems(path->nodes[level]))
1002                         continue;
1003                 if (level == 0)
1004                         btrfs_item_key_to_cpu(path->nodes[level], key,
1005                                               path->slots[level] + 1);
1006                 else
1007                         btrfs_node_key_to_cpu(path->nodes[level], key,
1008                                               path->slots[level] + 1);
1009                 return 0;
1010         }
1011         return 1;
1012 }
1013
1014 /*
1015  * look for inline back ref. if back ref is found, *ref_ret is set
1016  * to the address of inline back ref, and 0 is returned.
1017  *
1018  * if back ref isn't found, *ref_ret is set to the address where it
1019  * should be inserted, and -ENOENT is returned.
1020  *
1021  * if insert is true and there are too many inline back refs, the path
1022  * points to the extent item, and -EAGAIN is returned.
1023  *
1024  * NOTE: inline back refs are ordered in the same way that back ref
1025  *       items in the tree are ordered.
1026  */
1027 static noinline_for_stack
1028 int lookup_inline_extent_backref(struct btrfs_trans_handle *trans,
1029                                  struct btrfs_root *root,
1030                                  struct btrfs_path *path,
1031                                  struct btrfs_extent_inline_ref **ref_ret,
1032                                  u64 bytenr, u64 num_bytes,
1033                                  u64 parent, u64 root_objectid,
1034                                  u64 owner, u64 offset, int insert)
1035 {
1036         struct btrfs_key key;
1037         struct extent_buffer *leaf;
1038         struct btrfs_extent_item *ei;
1039         struct btrfs_extent_inline_ref *iref;
1040         u64 flags;
1041         u64 item_size;
1042         unsigned long ptr;
1043         unsigned long end;
1044         int extra_size;
1045         int type;
1046         int want;
1047         int ret;
1048         int err = 0;
1049
1050         key.objectid = bytenr;
1051         key.type = BTRFS_EXTENT_ITEM_KEY;
1052         key.offset = num_bytes;
1053
1054         want = extent_ref_type(parent, owner);
1055         if (insert) {
1056                 extra_size = btrfs_extent_inline_ref_size(want);
1057                 path->keep_locks = 1;
1058         } else
1059                 extra_size = -1;
1060         ret = btrfs_search_slot(trans, root, &key, path, extra_size, 1);
1061         if (ret < 0) {
1062                 err = ret;
1063                 goto out;
1064         }
1065         BUG_ON(ret);
1066
1067         leaf = path->nodes[0];
1068         item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1069 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
1070         if (item_size < sizeof(*ei)) {
1071                 if (!insert) {
1072                         err = -ENOENT;
1073                         goto out;
1074                 }
1075                 ret = convert_extent_item_v0(trans, root, path, owner,
1076                                              extra_size);
1077                 if (ret < 0) {
1078                         err = ret;
1079                         goto out;
1080                 }
1081                 leaf = path->nodes[0];
1082                 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1083         }
1084 #endif
1085         BUG_ON(item_size < sizeof(*ei));
1086
1087         ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1088         flags = btrfs_extent_flags(leaf, ei);
1089
1090         ptr = (unsigned long)(ei + 1);
1091         end = (unsigned long)ei + item_size;
1092
1093         if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
1094                 ptr += sizeof(struct btrfs_tree_block_info);
1095                 BUG_ON(ptr > end);
1096         } else {
1097                 BUG_ON(!(flags & BTRFS_EXTENT_FLAG_DATA));
1098         }
1099
1100         err = -ENOENT;
1101         while (1) {
1102                 if (ptr >= end) {
1103                         WARN_ON(ptr > end);
1104                         break;
1105                 }
1106                 iref = (struct btrfs_extent_inline_ref *)ptr;
1107                 type = btrfs_extent_inline_ref_type(leaf, iref);
1108                 if (want < type)
1109                         break;
1110                 if (want > type) {
1111                         ptr += btrfs_extent_inline_ref_size(type);
1112                         continue;
1113                 }
1114
1115                 if (type == BTRFS_EXTENT_DATA_REF_KEY) {
1116                         struct btrfs_extent_data_ref *dref;
1117                         dref = (struct btrfs_extent_data_ref *)(&iref->offset);
1118                         if (match_extent_data_ref(leaf, dref, root_objectid,
1119                                                   owner, offset)) {
1120                                 err = 0;
1121                                 break;
1122                         }
1123                         if (hash_extent_data_ref_item(leaf, dref) <
1124                             hash_extent_data_ref(root_objectid, owner, offset))
1125                                 break;
1126                 } else {
1127                         u64 ref_offset;
1128                         ref_offset = btrfs_extent_inline_ref_offset(leaf, iref);
1129                         if (parent > 0) {
1130                                 if (parent == ref_offset) {
1131                                         err = 0;
1132                                         break;
1133                                 }
1134                                 if (ref_offset < parent)
1135                                         break;
1136                         } else {
1137                                 if (root_objectid == ref_offset) {
1138                                         err = 0;
1139                                         break;
1140                                 }
1141                                 if (ref_offset < root_objectid)
1142                                         break;
1143                         }
1144                 }
1145                 ptr += btrfs_extent_inline_ref_size(type);
1146         }
1147         if (err == -ENOENT && insert) {
1148                 if (item_size + extra_size >=
1149                     BTRFS_MAX_EXTENT_ITEM_SIZE(root)) {
1150                         err = -EAGAIN;
1151                         goto out;
1152                 }
1153                 /*
1154                  * To add new inline back ref, we have to make sure
1155                  * there is no corresponding back ref item.
1156                  * For simplicity, we just do not add new inline back
1157                  * ref if there is any kind of item for this block
1158                  */
1159                 if (find_next_key(path, 0, &key) == 0 &&
1160                     key.objectid == bytenr &&
1161                     key.type < BTRFS_BLOCK_GROUP_ITEM_KEY) {
1162                         err = -EAGAIN;
1163                         goto out;
1164                 }
1165         }
1166         *ref_ret = (struct btrfs_extent_inline_ref *)ptr;
1167 out:
1168         if (insert) {
1169                 path->keep_locks = 0;
1170                 btrfs_unlock_up_safe(path, 1);
1171         }
1172         return err;
1173 }
1174
1175 /*
1176  * helper to add new inline back ref
1177  */
1178 static noinline_for_stack
1179 int setup_inline_extent_backref(struct btrfs_trans_handle *trans,
1180                                 struct btrfs_root *root,
1181                                 struct btrfs_path *path,
1182                                 struct btrfs_extent_inline_ref *iref,
1183                                 u64 parent, u64 root_objectid,
1184                                 u64 owner, u64 offset, int refs_to_add,
1185                                 struct btrfs_delayed_extent_op *extent_op)
1186 {
1187         struct extent_buffer *leaf;
1188         struct btrfs_extent_item *ei;
1189         unsigned long ptr;
1190         unsigned long end;
1191         unsigned long item_offset;
1192         u64 refs;
1193         int size;
1194         int type;
1195         int ret;
1196
1197         leaf = path->nodes[0];
1198         ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1199         item_offset = (unsigned long)iref - (unsigned long)ei;
1200
1201         type = extent_ref_type(parent, owner);
1202         size = btrfs_extent_inline_ref_size(type);
1203
1204         ret = btrfs_extend_item(trans, root, path, size);
1205         BUG_ON(ret);
1206
1207         ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1208         refs = btrfs_extent_refs(leaf, ei);
1209         refs += refs_to_add;
1210         btrfs_set_extent_refs(leaf, ei, refs);
1211         if (extent_op)
1212                 __run_delayed_extent_op(extent_op, leaf, ei);
1213
1214         ptr = (unsigned long)ei + item_offset;
1215         end = (unsigned long)ei + btrfs_item_size_nr(leaf, path->slots[0]);
1216         if (ptr < end - size)
1217                 memmove_extent_buffer(leaf, ptr + size, ptr,
1218                                       end - size - ptr);
1219
1220         iref = (struct btrfs_extent_inline_ref *)ptr;
1221         btrfs_set_extent_inline_ref_type(leaf, iref, type);
1222         if (type == BTRFS_EXTENT_DATA_REF_KEY) {
1223                 struct btrfs_extent_data_ref *dref;
1224                 dref = (struct btrfs_extent_data_ref *)(&iref->offset);
1225                 btrfs_set_extent_data_ref_root(leaf, dref, root_objectid);
1226                 btrfs_set_extent_data_ref_objectid(leaf, dref, owner);
1227                 btrfs_set_extent_data_ref_offset(leaf, dref, offset);
1228                 btrfs_set_extent_data_ref_count(leaf, dref, refs_to_add);
1229         } else if (type == BTRFS_SHARED_DATA_REF_KEY) {
1230                 struct btrfs_shared_data_ref *sref;
1231                 sref = (struct btrfs_shared_data_ref *)(iref + 1);
1232                 btrfs_set_shared_data_ref_count(leaf, sref, refs_to_add);
1233                 btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
1234         } else if (type == BTRFS_SHARED_BLOCK_REF_KEY) {
1235                 btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
1236         } else {
1237                 btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid);
1238         }
1239         btrfs_mark_buffer_dirty(leaf);
1240         return 0;
1241 }
1242
1243 static int lookup_extent_backref(struct btrfs_trans_handle *trans,
1244                                  struct btrfs_root *root,
1245                                  struct btrfs_path *path,
1246                                  struct btrfs_extent_inline_ref **ref_ret,
1247                                  u64 bytenr, u64 num_bytes, u64 parent,
1248                                  u64 root_objectid, u64 owner, u64 offset)
1249 {
1250         int ret;
1251
1252         ret = lookup_inline_extent_backref(trans, root, path, ref_ret,
1253                                            bytenr, num_bytes, parent,
1254                                            root_objectid, owner, offset, 0);
1255         if (ret != -ENOENT)
1256                 return ret;
1257
1258         btrfs_release_path(root, path);
1259         *ref_ret = NULL;
1260
1261         if (owner < BTRFS_FIRST_FREE_OBJECTID) {
1262                 ret = lookup_tree_block_ref(trans, root, path, bytenr, parent,
1263                                             root_objectid);
1264         } else {
1265                 ret = lookup_extent_data_ref(trans, root, path, bytenr, parent,
1266                                              root_objectid, owner, offset);
1267         }
1268         return ret;
1269 }
1270
1271 /*
1272  * helper to update/remove inline back ref
1273  */
1274 static noinline_for_stack
1275 int update_inline_extent_backref(struct btrfs_trans_handle *trans,
1276                                  struct btrfs_root *root,
1277                                  struct btrfs_path *path,
1278                                  struct btrfs_extent_inline_ref *iref,
1279                                  int refs_to_mod,
1280                                  struct btrfs_delayed_extent_op *extent_op)
1281 {
1282         struct extent_buffer *leaf;
1283         struct btrfs_extent_item *ei;
1284         struct btrfs_extent_data_ref *dref = NULL;
1285         struct btrfs_shared_data_ref *sref = NULL;
1286         unsigned long ptr;
1287         unsigned long end;
1288         u32 item_size;
1289         int size;
1290         int type;
1291         int ret;
1292         u64 refs;
1293
1294         leaf = path->nodes[0];
1295         ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1296         refs = btrfs_extent_refs(leaf, ei);
1297         WARN_ON(refs_to_mod < 0 && refs + refs_to_mod <= 0);
1298         refs += refs_to_mod;
1299         btrfs_set_extent_refs(leaf, ei, refs);
1300         if (extent_op)
1301                 __run_delayed_extent_op(extent_op, leaf, ei);
1302
1303         type = btrfs_extent_inline_ref_type(leaf, iref);
1304
1305         if (type == BTRFS_EXTENT_DATA_REF_KEY) {
1306                 dref = (struct btrfs_extent_data_ref *)(&iref->offset);
1307                 refs = btrfs_extent_data_ref_count(leaf, dref);
1308         } else if (type == BTRFS_SHARED_DATA_REF_KEY) {
1309                 sref = (struct btrfs_shared_data_ref *)(iref + 1);
1310                 refs = btrfs_shared_data_ref_count(leaf, sref);
1311         } else {
1312                 refs = 1;
1313                 BUG_ON(refs_to_mod != -1);
1314         }
1315
1316         BUG_ON(refs_to_mod < 0 && refs < -refs_to_mod);
1317         refs += refs_to_mod;
1318
1319         if (refs > 0) {
1320                 if (type == BTRFS_EXTENT_DATA_REF_KEY)
1321                         btrfs_set_extent_data_ref_count(leaf, dref, refs);
1322                 else
1323                         btrfs_set_shared_data_ref_count(leaf, sref, refs);
1324         } else {
1325                 size =  btrfs_extent_inline_ref_size(type);
1326                 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1327                 ptr = (unsigned long)iref;
1328                 end = (unsigned long)ei + item_size;
1329                 if (ptr + size < end)
1330                         memmove_extent_buffer(leaf, ptr, ptr + size,
1331                                               end - ptr - size);
1332                 item_size -= size;
1333                 ret = btrfs_truncate_item(trans, root, path, item_size, 1);
1334                 BUG_ON(ret);
1335         }
1336         btrfs_mark_buffer_dirty(leaf);
1337         return 0;
1338 }
1339
1340 static noinline_for_stack
1341 int insert_inline_extent_backref(struct btrfs_trans_handle *trans,
1342                                  struct btrfs_root *root,
1343                                  struct btrfs_path *path,
1344                                  u64 bytenr, u64 num_bytes, u64 parent,
1345                                  u64 root_objectid, u64 owner,
1346                                  u64 offset, int refs_to_add,
1347                                  struct btrfs_delayed_extent_op *extent_op)
1348 {
1349         struct btrfs_extent_inline_ref *iref;
1350         int ret;
1351
1352         ret = lookup_inline_extent_backref(trans, root, path, &iref,
1353                                            bytenr, num_bytes, parent,
1354                                            root_objectid, owner, offset, 1);
1355         if (ret == 0) {
1356                 BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID);
1357                 ret = update_inline_extent_backref(trans, root, path, iref,
1358                                                    refs_to_add, extent_op);
1359         } else if (ret == -ENOENT) {
1360                 ret = setup_inline_extent_backref(trans, root, path, iref,
1361                                                   parent, root_objectid,
1362                                                   owner, offset, refs_to_add,
1363                                                   extent_op);
1364         }
1365         return ret;
1366 }
1367
1368 static int insert_extent_backref(struct btrfs_trans_handle *trans,
1369                                  struct btrfs_root *root,
1370                                  struct btrfs_path *path,
1371                                  u64 bytenr, u64 parent, u64 root_objectid,
1372                                  u64 owner, u64 offset, int refs_to_add)
1373 {
1374         int ret;
1375         if (owner < BTRFS_FIRST_FREE_OBJECTID) {
1376                 BUG_ON(refs_to_add != 1);
1377                 ret = insert_tree_block_ref(trans, root, path, bytenr,
1378                                             parent, root_objectid);
1379         } else {
1380                 ret = insert_extent_data_ref(trans, root, path, bytenr,
1381                                              parent, root_objectid,
1382                                              owner, offset, refs_to_add);
1383         }
1384         return ret;
1385 }
1386
1387 static int remove_extent_backref(struct btrfs_trans_handle *trans,
1388                                  struct btrfs_root *root,
1389                                  struct btrfs_path *path,
1390                                  struct btrfs_extent_inline_ref *iref,
1391                                  int refs_to_drop, int is_data)
1392 {
1393         int ret;
1394
1395         BUG_ON(!is_data && refs_to_drop != 1);
1396         if (iref) {
1397                 ret = update_inline_extent_backref(trans, root, path, iref,
1398                                                    -refs_to_drop, NULL);
1399         } else if (is_data) {
1400                 ret = remove_extent_data_ref(trans, root, path, refs_to_drop);
1401         } else {
1402                 ret = btrfs_del_item(trans, root, path);
1403         }
1404         return ret;
1405 }
1406
1407 #ifdef BIO_RW_DISCARD
1408 static void btrfs_issue_discard(struct block_device *bdev,
1409                                 u64 start, u64 len)
1410 {
1411         blkdev_issue_discard(bdev, start >> 9, len >> 9, GFP_KERNEL);
1412 }
1413 #endif
1414
1415 static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr,
1416                                 u64 num_bytes)
1417 {
1418 #ifdef BIO_RW_DISCARD
1419         int ret;
1420         u64 map_length = num_bytes;
1421         struct btrfs_multi_bio *multi = NULL;
1422
1423         /* Tell the block device(s) that the sectors can be discarded */
1424         ret = btrfs_map_block(&root->fs_info->mapping_tree, READ,
1425                               bytenr, &map_length, &multi, 0);
1426         if (!ret) {
1427                 struct btrfs_bio_stripe *stripe = multi->stripes;
1428                 int i;
1429
1430                 if (map_length > num_bytes)
1431                         map_length = num_bytes;
1432
1433                 for (i = 0; i < multi->num_stripes; i++, stripe++) {
1434                         btrfs_issue_discard(stripe->dev->bdev,
1435                                             stripe->physical,
1436                                             map_length);
1437                 }
1438                 kfree(multi);
1439         }
1440
1441         return ret;
1442 #else
1443         return 0;
1444 #endif
1445 }
1446
1447 int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
1448                          struct btrfs_root *root,
1449                          u64 bytenr, u64 num_bytes, u64 parent,
1450                          u64 root_objectid, u64 owner, u64 offset)
1451 {
1452         int ret;
1453         BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID &&
1454                root_objectid == BTRFS_TREE_LOG_OBJECTID);
1455
1456         if (owner < BTRFS_FIRST_FREE_OBJECTID) {
1457                 ret = btrfs_add_delayed_tree_ref(trans, bytenr, num_bytes,
1458                                         parent, root_objectid, (int)owner,
1459                                         BTRFS_ADD_DELAYED_REF, NULL);
1460         } else {
1461                 ret = btrfs_add_delayed_data_ref(trans, bytenr, num_bytes,
1462                                         parent, root_objectid, owner, offset,
1463                                         BTRFS_ADD_DELAYED_REF, NULL);
1464         }
1465         return ret;
1466 }
1467
1468 static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
1469                                   struct btrfs_root *root,
1470                                   u64 bytenr, u64 num_bytes,
1471                                   u64 parent, u64 root_objectid,
1472                                   u64 owner, u64 offset, int refs_to_add,
1473                                   struct btrfs_delayed_extent_op *extent_op)
1474 {
1475         struct btrfs_path *path;
1476         struct extent_buffer *leaf;
1477         struct btrfs_extent_item *item;
1478         u64 refs;
1479         int ret;
1480         int err = 0;
1481
1482         path = btrfs_alloc_path();
1483         if (!path)
1484                 return -ENOMEM;
1485
1486         path->reada = 1;
1487         path->leave_spinning = 1;
1488         /* this will setup the path even if it fails to insert the back ref */
1489         ret = insert_inline_extent_backref(trans, root->fs_info->extent_root,
1490                                            path, bytenr, num_bytes, parent,
1491                                            root_objectid, owner, offset,
1492                                            refs_to_add, extent_op);
1493         if (ret == 0)
1494                 goto out;
1495
1496         if (ret != -EAGAIN) {
1497                 err = ret;
1498                 goto out;
1499         }
1500
1501         leaf = path->nodes[0];
1502         item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1503         refs = btrfs_extent_refs(leaf, item);
1504         btrfs_set_extent_refs(leaf, item, refs + refs_to_add);
1505         if (extent_op)
1506                 __run_delayed_extent_op(extent_op, leaf, item);
1507
1508         btrfs_mark_buffer_dirty(leaf);
1509         btrfs_release_path(root->fs_info->extent_root, path);
1510
1511         path->reada = 1;
1512         path->leave_spinning = 1;
1513
1514         /* now insert the actual backref */
1515         ret = insert_extent_backref(trans, root->fs_info->extent_root,
1516                                     path, bytenr, parent, root_objectid,
1517                                     owner, offset, refs_to_add);
1518         BUG_ON(ret);
1519 out:
1520         btrfs_free_path(path);
1521         return err;
1522 }
1523
1524 static int run_delayed_data_ref(struct btrfs_trans_handle *trans,
1525                                 struct btrfs_root *root,
1526                                 struct btrfs_delayed_ref_node *node,
1527                                 struct btrfs_delayed_extent_op *extent_op,
1528                                 int insert_reserved)
1529 {
1530         int ret = 0;
1531         struct btrfs_delayed_data_ref *ref;
1532         struct btrfs_key ins;
1533         u64 parent = 0;
1534         u64 ref_root = 0;
1535         u64 flags = 0;
1536
1537         ins.objectid = node->bytenr;
1538         ins.offset = node->num_bytes;
1539         ins.type = BTRFS_EXTENT_ITEM_KEY;
1540
1541         ref = btrfs_delayed_node_to_data_ref(node);
1542         if (node->type == BTRFS_SHARED_DATA_REF_KEY)
1543                 parent = ref->parent;
1544         else
1545                 ref_root = ref->root;
1546
1547         if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) {
1548                 if (extent_op) {
1549                         BUG_ON(extent_op->update_key);
1550                         flags |= extent_op->flags_to_set;
1551                 }
1552                 ret = alloc_reserved_file_extent(trans, root,
1553                                                  parent, ref_root, flags,
1554                                                  ref->objectid, ref->offset,
1555                                                  &ins, node->ref_mod);
1556                 update_reserved_extents(root, ins.objectid, ins.offset, 0);
1557         } else if (node->action == BTRFS_ADD_DELAYED_REF) {
1558                 ret = __btrfs_inc_extent_ref(trans, root, node->bytenr,
1559                                              node->num_bytes, parent,
1560                                              ref_root, ref->objectid,
1561                                              ref->offset, node->ref_mod,
1562                                              extent_op);
1563         } else if (node->action == BTRFS_DROP_DELAYED_REF) {
1564                 ret = __btrfs_free_extent(trans, root, node->bytenr,
1565                                           node->num_bytes, parent,
1566                                           ref_root, ref->objectid,
1567                                           ref->offset, node->ref_mod,
1568                                           extent_op);
1569         } else {
1570                 BUG();
1571         }
1572         return ret;
1573 }
1574
1575 static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op,
1576                                     struct extent_buffer *leaf,
1577                                     struct btrfs_extent_item *ei)
1578 {
1579         u64 flags = btrfs_extent_flags(leaf, ei);
1580         if (extent_op->update_flags) {
1581                 flags |= extent_op->flags_to_set;
1582                 btrfs_set_extent_flags(leaf, ei, flags);
1583         }
1584
1585         if (extent_op->update_key) {
1586                 struct btrfs_tree_block_info *bi;
1587                 BUG_ON(!(flags & BTRFS_EXTENT_FLAG_TREE_BLOCK));
1588                 bi = (struct btrfs_tree_block_info *)(ei + 1);
1589                 btrfs_set_tree_block_key(leaf, bi, &extent_op->key);
1590         }
1591 }
1592
1593 static int run_delayed_extent_op(struct btrfs_trans_handle *trans,
1594                                  struct btrfs_root *root,
1595                                  struct btrfs_delayed_ref_node *node,
1596                                  struct btrfs_delayed_extent_op *extent_op)
1597 {
1598         struct btrfs_key key;
1599         struct btrfs_path *path;
1600         struct btrfs_extent_item *ei;
1601         struct extent_buffer *leaf;
1602         u32 item_size;
1603         int ret;
1604         int err = 0;
1605
1606         path = btrfs_alloc_path();
1607         if (!path)
1608                 return -ENOMEM;
1609
1610         key.objectid = node->bytenr;
1611         key.type = BTRFS_EXTENT_ITEM_KEY;
1612         key.offset = node->num_bytes;
1613
1614         path->reada = 1;
1615         path->leave_spinning = 1;
1616         ret = btrfs_search_slot(trans, root->fs_info->extent_root, &key,
1617                                 path, 0, 1);
1618         if (ret < 0) {
1619                 err = ret;
1620                 goto out;
1621         }
1622         if (ret > 0) {
1623                 err = -EIO;
1624                 goto out;
1625         }
1626
1627         leaf = path->nodes[0];
1628         item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1629 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
1630         if (item_size < sizeof(*ei)) {
1631                 ret = convert_extent_item_v0(trans, root->fs_info->extent_root,
1632                                              path, (u64)-1, 0);
1633                 if (ret < 0) {
1634                         err = ret;
1635                         goto out;
1636                 }
1637                 leaf = path->nodes[0];
1638                 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1639         }
1640 #endif
1641         BUG_ON(item_size < sizeof(*ei));
1642         ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1643         __run_delayed_extent_op(extent_op, leaf, ei);
1644
1645         btrfs_mark_buffer_dirty(leaf);
1646 out:
1647         btrfs_free_path(path);
1648         return err;
1649 }
1650
1651 static int run_delayed_tree_ref(struct btrfs_trans_handle *trans,
1652                                 struct btrfs_root *root,
1653                                 struct btrfs_delayed_ref_node *node,
1654                                 struct btrfs_delayed_extent_op *extent_op,
1655                                 int insert_reserved)
1656 {
1657         int ret = 0;
1658         struct btrfs_delayed_tree_ref *ref;
1659         struct btrfs_key ins;
1660         u64 parent = 0;
1661         u64 ref_root = 0;
1662
1663         ins.objectid = node->bytenr;
1664         ins.offset = node->num_bytes;
1665         ins.type = BTRFS_EXTENT_ITEM_KEY;
1666
1667         ref = btrfs_delayed_node_to_tree_ref(node);
1668         if (node->type == BTRFS_SHARED_BLOCK_REF_KEY)
1669                 parent = ref->parent;
1670         else
1671                 ref_root = ref->root;
1672
1673         BUG_ON(node->ref_mod != 1);
1674         if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) {
1675                 BUG_ON(!extent_op || !extent_op->update_flags ||
1676                        !extent_op->update_key);
1677                 ret = alloc_reserved_tree_block(trans, root,
1678                                                 parent, ref_root,
1679                                                 extent_op->flags_to_set,
1680                                                 &extent_op->key,
1681                                                 ref->level, &ins);
1682                 update_reserved_extents(root, ins.objectid, ins.offset, 0);
1683         } else if (node->action == BTRFS_ADD_DELAYED_REF) {
1684                 ret = __btrfs_inc_extent_ref(trans, root, node->bytenr,
1685                                              node->num_bytes, parent, ref_root,
1686                                              ref->level, 0, 1, extent_op);
1687         } else if (node->action == BTRFS_DROP_DELAYED_REF) {
1688                 ret = __btrfs_free_extent(trans, root, node->bytenr,
1689                                           node->num_bytes, parent, ref_root,
1690                                           ref->level, 0, 1, extent_op);
1691         } else {
1692                 BUG();
1693         }
1694         return ret;
1695 }
1696
1697
1698 /* helper function to actually process a single delayed ref entry */
1699 static int run_one_delayed_ref(struct btrfs_trans_handle *trans,
1700                                struct btrfs_root *root,
1701                                struct btrfs_delayed_ref_node *node,
1702                                struct btrfs_delayed_extent_op *extent_op,
1703                                int insert_reserved)
1704 {
1705         int ret;
1706         if (btrfs_delayed_ref_is_head(node)) {
1707                 struct btrfs_delayed_ref_head *head;
1708                 /*
1709                  * we've hit the end of the chain and we were supposed
1710                  * to insert this extent into the tree.  But, it got
1711                  * deleted before we ever needed to insert it, so all
1712                  * we have to do is clean up the accounting
1713                  */
1714                 BUG_ON(extent_op);
1715                 head = btrfs_delayed_node_to_head(node);
1716                 if (insert_reserved) {
1717                         if (head->is_data) {
1718                                 ret = btrfs_del_csums(trans, root,
1719                                                       node->bytenr,
1720                                                       node->num_bytes);
1721                                 BUG_ON(ret);
1722                         }
1723                         btrfs_update_pinned_extents(root, node->bytenr,
1724                                                     node->num_bytes, 1);
1725                         update_reserved_extents(root, node->bytenr,
1726                                                 node->num_bytes, 0);
1727                 }
1728                 mutex_unlock(&head->mutex);
1729                 return 0;
1730         }
1731
1732         if (node->type == BTRFS_TREE_BLOCK_REF_KEY ||
1733             node->type == BTRFS_SHARED_BLOCK_REF_KEY)
1734                 ret = run_delayed_tree_ref(trans, root, node, extent_op,
1735                                            insert_reserved);
1736         else if (node->type == BTRFS_EXTENT_DATA_REF_KEY ||
1737                  node->type == BTRFS_SHARED_DATA_REF_KEY)
1738                 ret = run_delayed_data_ref(trans, root, node, extent_op,
1739                                            insert_reserved);
1740         else
1741                 BUG();
1742         return ret;
1743 }
1744
1745 static noinline struct btrfs_delayed_ref_node *
1746 select_delayed_ref(struct btrfs_delayed_ref_head *head)
1747 {
1748         struct rb_node *node;
1749         struct btrfs_delayed_ref_node *ref;
1750         int action = BTRFS_ADD_DELAYED_REF;
1751 again:
1752         /*
1753          * select delayed ref of type BTRFS_ADD_DELAYED_REF first.
1754          * this prevents ref count from going down to zero when
1755          * there still are pending delayed ref.
1756          */
1757         node = rb_prev(&head->node.rb_node);
1758         while (1) {
1759                 if (!node)
1760                         break;
1761                 ref = rb_entry(node, struct btrfs_delayed_ref_node,
1762                                 rb_node);
1763                 if (ref->bytenr != head->node.bytenr)
1764                         break;
1765                 if (ref->action == action)
1766                         return ref;
1767                 node = rb_prev(node);
1768         }
1769         if (action == BTRFS_ADD_DELAYED_REF) {
1770                 action = BTRFS_DROP_DELAYED_REF;
1771                 goto again;
1772         }
1773         return NULL;
1774 }
1775
1776 static noinline int run_clustered_refs(struct btrfs_trans_handle *trans,
1777                                        struct btrfs_root *root,
1778                                        struct list_head *cluster)
1779 {
1780         struct btrfs_delayed_ref_root *delayed_refs;
1781         struct btrfs_delayed_ref_node *ref;
1782         struct btrfs_delayed_ref_head *locked_ref = NULL;
1783         struct btrfs_delayed_extent_op *extent_op;
1784         int ret;
1785         int count = 0;
1786         int must_insert_reserved = 0;
1787
1788         delayed_refs = &trans->transaction->delayed_refs;
1789         while (1) {
1790                 if (!locked_ref) {
1791                         /* pick a new head ref from the cluster list */
1792                         if (list_empty(cluster))
1793                                 break;
1794
1795                         locked_ref = list_entry(cluster->next,
1796                                      struct btrfs_delayed_ref_head, cluster);
1797
1798                         /* grab the lock that says we are going to process
1799                          * all the refs for this head */
1800                         ret = btrfs_delayed_ref_lock(trans, locked_ref);
1801
1802                         /*
1803                          * we may have dropped the spin lock to get the head
1804                          * mutex lock, and that might have given someone else
1805                          * time to free the head.  If that's true, it has been
1806                          * removed from our list and we can move on.
1807                          */
1808                         if (ret == -EAGAIN) {
1809                                 locked_ref = NULL;
1810                                 count++;
1811                                 continue;
1812                         }
1813                 }
1814
1815                 /*
1816                  * record the must insert reserved flag before we
1817                  * drop the spin lock.
1818                  */
1819                 must_insert_reserved = locked_ref->must_insert_reserved;
1820                 locked_ref->must_insert_reserved = 0;
1821
1822                 extent_op = locked_ref->extent_op;
1823                 locked_ref->extent_op = NULL;
1824
1825                 /*
1826                  * locked_ref is the head node, so we have to go one
1827                  * node back for any delayed ref updates
1828                  */
1829                 ref = select_delayed_ref(locked_ref);
1830                 if (!ref) {
1831                         /* All delayed refs have been processed, Go ahead
1832                          * and send the head node to run_one_delayed_ref,
1833                          * so that any accounting fixes can happen
1834                          */
1835                         ref = &locked_ref->node;
1836
1837                         if (extent_op && must_insert_reserved) {
1838                                 kfree(extent_op);
1839                                 extent_op = NULL;
1840                         }
1841
1842                         if (extent_op) {
1843                                 spin_unlock(&delayed_refs->lock);
1844
1845                                 ret = run_delayed_extent_op(trans, root,
1846                                                             ref, extent_op);
1847                                 BUG_ON(ret);
1848                                 kfree(extent_op);
1849
1850                                 cond_resched();
1851                                 spin_lock(&delayed_refs->lock);
1852                                 continue;
1853                         }
1854
1855                         list_del_init(&locked_ref->cluster);
1856                         locked_ref = NULL;
1857                 }
1858
1859                 ref->in_tree = 0;
1860                 rb_erase(&ref->rb_node, &delayed_refs->root);
1861                 delayed_refs->num_entries--;
1862
1863                 spin_unlock(&delayed_refs->lock);
1864
1865                 ret = run_one_delayed_ref(trans, root, ref, extent_op,
1866                                           must_insert_reserved);
1867                 BUG_ON(ret);
1868
1869                 btrfs_put_delayed_ref(ref);
1870                 kfree(extent_op);
1871                 count++;
1872
1873                 cond_resched();
1874                 spin_lock(&delayed_refs->lock);
1875         }
1876         return count;
1877 }
1878
1879 /*
1880  * this starts processing the delayed reference count updates and
1881  * extent insertions we have queued up so far.  count can be
1882  * 0, which means to process everything in the tree at the start
1883  * of the run (but not newly added entries), or it can be some target
1884  * number you'd like to process.
1885  */
1886 int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
1887                            struct btrfs_root *root, unsigned long count)
1888 {
1889         struct rb_node *node;
1890         struct btrfs_delayed_ref_root *delayed_refs;
1891         struct btrfs_delayed_ref_node *ref;
1892         struct list_head cluster;
1893         int ret;
1894         int run_all = count == (unsigned long)-1;
1895         int run_most = 0;
1896
1897         if (root == root->fs_info->extent_root)
1898                 root = root->fs_info->tree_root;
1899
1900         delayed_refs = &trans->transaction->delayed_refs;
1901         INIT_LIST_HEAD(&cluster);
1902 again:
1903         spin_lock(&delayed_refs->lock);
1904         if (count == 0) {
1905                 count = delayed_refs->num_entries * 2;
1906                 run_most = 1;
1907         }
1908         while (1) {
1909                 if (!(run_all || run_most) &&
1910                     delayed_refs->num_heads_ready < 64)
1911                         break;
1912
1913                 /*
1914                  * go find something we can process in the rbtree.  We start at
1915                  * the beginning of the tree, and then build a cluster
1916                  * of refs to process starting at the first one we are able to
1917                  * lock
1918                  */
1919                 ret = btrfs_find_ref_cluster(trans, &cluster,
1920                                              delayed_refs->run_delayed_start);
1921                 if (ret)
1922                         break;
1923
1924                 ret = run_clustered_refs(trans, root, &cluster);
1925                 BUG_ON(ret < 0);
1926
1927                 count -= min_t(unsigned long, ret, count);
1928
1929                 if (count == 0)
1930                         break;
1931         }
1932
1933         if (run_all) {
1934                 node = rb_first(&delayed_refs->root);
1935                 if (!node)
1936                         goto out;
1937                 count = (unsigned long)-1;
1938
1939                 while (node) {
1940                         ref = rb_entry(node, struct btrfs_delayed_ref_node,
1941                                        rb_node);
1942                         if (btrfs_delayed_ref_is_head(ref)) {
1943                                 struct btrfs_delayed_ref_head *head;
1944
1945                                 head = btrfs_delayed_node_to_head(ref);
1946                                 atomic_inc(&ref->refs);
1947
1948                                 spin_unlock(&delayed_refs->lock);
1949                                 mutex_lock(&head->mutex);
1950                                 mutex_unlock(&head->mutex);
1951
1952                                 btrfs_put_delayed_ref(ref);
1953                                 cond_resched();
1954                                 goto again;
1955                         }
1956                         node = rb_next(node);
1957                 }
1958                 spin_unlock(&delayed_refs->lock);
1959                 schedule_timeout(1);
1960                 goto again;
1961         }
1962 out:
1963         spin_unlock(&delayed_refs->lock);
1964         return 0;
1965 }
1966
1967 int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans,
1968                                 struct btrfs_root *root,
1969                                 u64 bytenr, u64 num_bytes, u64 flags,
1970                                 int is_data)
1971 {
1972         struct btrfs_delayed_extent_op *extent_op;
1973         int ret;
1974
1975         extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS);
1976         if (!extent_op)
1977                 return -ENOMEM;
1978
1979         extent_op->flags_to_set = flags;
1980         extent_op->update_flags = 1;
1981         extent_op->update_key = 0;
1982         extent_op->is_data = is_data ? 1 : 0;
1983
1984         ret = btrfs_add_delayed_extent_op(trans, bytenr, num_bytes, extent_op);
1985         if (ret)
1986                 kfree(extent_op);
1987         return ret;
1988 }
1989
1990 static noinline int check_delayed_ref(struct btrfs_trans_handle *trans,
1991                                       struct btrfs_root *root,
1992                                       struct btrfs_path *path,
1993                                       u64 objectid, u64 offset, u64 bytenr)
1994 {
1995         struct btrfs_delayed_ref_head *head;
1996         struct btrfs_delayed_ref_node *ref;
1997         struct btrfs_delayed_data_ref *data_ref;
1998         struct btrfs_delayed_ref_root *delayed_refs;
1999         struct rb_node *node;
2000         int ret = 0;
2001
2002         ret = -ENOENT;
2003         delayed_refs = &trans->transaction->delayed_refs;
2004         spin_lock(&delayed_refs->lock);
2005         head = btrfs_find_delayed_ref_head(trans, bytenr);
2006         if (!head)
2007                 goto out;
2008
2009         if (!mutex_trylock(&head->mutex)) {
2010                 atomic_inc(&head->node.refs);
2011                 spin_unlock(&delayed_refs->lock);
2012
2013                 btrfs_release_path(root->fs_info->extent_root, path);
2014
2015                 mutex_lock(&head->mutex);
2016                 mutex_unlock(&head->mutex);
2017                 btrfs_put_delayed_ref(&head->node);
2018                 return -EAGAIN;
2019         }
2020
2021         node = rb_prev(&head->node.rb_node);
2022         if (!node)
2023                 goto out_unlock;
2024
2025         ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
2026
2027         if (ref->bytenr != bytenr)
2028                 goto out_unlock;
2029
2030         ret = 1;
2031         if (ref->type != BTRFS_EXTENT_DATA_REF_KEY)
2032                 goto out_unlock;
2033
2034         data_ref = btrfs_delayed_node_to_data_ref(ref);
2035
2036         node = rb_prev(node);
2037         if (node) {
2038                 ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
2039                 if (ref->bytenr == bytenr)
2040                         goto out_unlock;
2041         }
2042
2043         if (data_ref->root != root->root_key.objectid ||
2044             data_ref->objectid != objectid || data_ref->offset != offset)
2045                 goto out_unlock;
2046
2047         ret = 0;
2048 out_unlock:
2049         mutex_unlock(&head->mutex);
2050 out:
2051         spin_unlock(&delayed_refs->lock);
2052         return ret;
2053 }
2054
2055 static noinline int check_committed_ref(struct btrfs_trans_handle *trans,
2056                                         struct btrfs_root *root,
2057                                         struct btrfs_path *path,
2058                                         u64 objectid, u64 offset, u64 bytenr)
2059 {
2060         struct btrfs_root *extent_root = root->fs_info->extent_root;
2061         struct extent_buffer *leaf;
2062         struct btrfs_extent_data_ref *ref;
2063         struct btrfs_extent_inline_ref *iref;
2064         struct btrfs_extent_item *ei;
2065         struct btrfs_key key;
2066         u32 item_size;
2067         int ret;
2068
2069         key.objectid = bytenr;
2070         key.offset = (u64)-1;
2071         key.type = BTRFS_EXTENT_ITEM_KEY;
2072
2073         ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
2074         if (ret < 0)
2075                 goto out;
2076         BUG_ON(ret == 0);
2077
2078         ret = -ENOENT;
2079         if (path->slots[0] == 0)
2080                 goto out;
2081
2082         path->slots[0]--;
2083         leaf = path->nodes[0];
2084         btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
2085
2086         if (key.objectid != bytenr || key.type != BTRFS_EXTENT_ITEM_KEY)
2087                 goto out;
2088
2089         ret = 1;
2090         item_size = btrfs_item_size_nr(leaf, path->slots[0]);
2091 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
2092         if (item_size < sizeof(*ei)) {
2093                 WARN_ON(item_size != sizeof(struct btrfs_extent_item_v0));
2094                 goto out;
2095         }
2096 #endif
2097         ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
2098
2099         if (item_size != sizeof(*ei) +
2100             btrfs_extent_inline_ref_size(BTRFS_EXTENT_DATA_REF_KEY))
2101                 goto out;
2102
2103         if (btrfs_extent_generation(leaf, ei) <=
2104             btrfs_root_last_snapshot(&root->root_item))
2105                 goto out;
2106
2107         iref = (struct btrfs_extent_inline_ref *)(ei + 1);
2108         if (btrfs_extent_inline_ref_type(leaf, iref) !=
2109             BTRFS_EXTENT_DATA_REF_KEY)
2110                 goto out;
2111
2112         ref = (struct btrfs_extent_data_ref *)(&iref->offset);
2113         if (btrfs_extent_refs(leaf, ei) !=
2114             btrfs_extent_data_ref_count(leaf, ref) ||
2115             btrfs_extent_data_ref_root(leaf, ref) !=
2116             root->root_key.objectid ||
2117             btrfs_extent_data_ref_objectid(leaf, ref) != objectid ||
2118             btrfs_extent_data_ref_offset(leaf, ref) != offset)
2119                 goto out;
2120
2121         ret = 0;
2122 out:
2123         return ret;
2124 }
2125
2126 int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans,
2127                           struct btrfs_root *root,
2128                           u64 objectid, u64 offset, u64 bytenr)
2129 {
2130         struct btrfs_path *path;
2131         int ret;
2132         int ret2;
2133
2134         path = btrfs_alloc_path();
2135         if (!path)
2136                 return -ENOENT;
2137
2138         do {
2139                 ret = check_committed_ref(trans, root, path, objectid,
2140                                           offset, bytenr);
2141                 if (ret && ret != -ENOENT)
2142                         goto out;
2143
2144                 ret2 = check_delayed_ref(trans, root, path, objectid,
2145                                          offset, bytenr);
2146         } while (ret2 == -EAGAIN);
2147
2148         if (ret2 && ret2 != -ENOENT) {
2149                 ret = ret2;
2150                 goto out;
2151         }
2152
2153         if (ret != -ENOENT || ret2 != -ENOENT)
2154                 ret = 0;
2155 out:
2156         btrfs_free_path(path);
2157         return ret;
2158 }
2159
2160 #if 0
2161 int btrfs_cache_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
2162                     struct extent_buffer *buf, u32 nr_extents)
2163 {
2164         struct btrfs_key key;
2165         struct btrfs_file_extent_item *fi;
2166         u64 root_gen;
2167         u32 nritems;
2168         int i;
2169         int level;
2170         int ret = 0;
2171         int shared = 0;
2172
2173         if (!root->ref_cows)
2174                 return 0;
2175
2176         if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) {
2177                 shared = 0;
2178                 root_gen = root->root_key.offset;
2179         } else {
2180                 shared = 1;
2181                 root_gen = trans->transid - 1;
2182         }
2183
2184         level = btrfs_header_level(buf);
2185         nritems = btrfs_header_nritems(buf);
2186
2187         if (level == 0) {
2188                 struct btrfs_leaf_ref *ref;
2189                 struct btrfs_extent_info *info;
2190
2191                 ref = btrfs_alloc_leaf_ref(root, nr_extents);
2192                 if (!ref) {
2193                         ret = -ENOMEM;
2194                         goto out;
2195                 }
2196
2197                 ref->root_gen = root_gen;
2198                 ref->bytenr = buf->start;
2199                 ref->owner = btrfs_header_owner(buf);
2200                 ref->generation = btrfs_header_generation(buf);
2201                 ref->nritems = nr_extents;
2202                 info = ref->extents;
2203
2204                 for (i = 0; nr_extents > 0 && i < nritems; i++) {
2205                         u64 disk_bytenr;
2206                         btrfs_item_key_to_cpu(buf, &key, i);
2207                         if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
2208                                 continue;
2209                         fi = btrfs_item_ptr(buf, i,
2210                                             struct btrfs_file_extent_item);
2211                         if (btrfs_file_extent_type(buf, fi) ==
2212                             BTRFS_FILE_EXTENT_INLINE)
2213                                 continue;
2214                         disk_bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
2215                         if (disk_bytenr == 0)
2216                                 continue;
2217
2218                         info->bytenr = disk_bytenr;
2219                         info->num_bytes =
2220                                 btrfs_file_extent_disk_num_bytes(buf, fi);
2221                         info->objectid = key.objectid;
2222                         info->offset = key.offset;
2223                         info++;
2224                 }
2225
2226                 ret = btrfs_add_leaf_ref(root, ref, shared);
2227                 if (ret == -EEXIST && shared) {
2228                         struct btrfs_leaf_ref *old;
2229                         old = btrfs_lookup_leaf_ref(root, ref->bytenr);
2230                         BUG_ON(!old);
2231                         btrfs_remove_leaf_ref(root, old);
2232                         btrfs_free_leaf_ref(root, old);
2233                         ret = btrfs_add_leaf_ref(root, ref, shared);
2234                 }
2235                 WARN_ON(ret);
2236                 btrfs_free_leaf_ref(root, ref);
2237         }
2238 out:
2239         return ret;
2240 }
2241
2242 /* when a block goes through cow, we update the reference counts of
2243  * everything that block points to.  The internal pointers of the block
2244  * can be in just about any order, and it is likely to have clusters of
2245  * things that are close together and clusters of things that are not.
2246  *
2247  * To help reduce the seeks that come with updating all of these reference
2248  * counts, sort them by byte number before actual updates are done.
2249  *
2250  * struct refsort is used to match byte number to slot in the btree block.
2251  * we sort based on the byte number and then use the slot to actually
2252  * find the item.
2253  *
2254  * struct refsort is smaller than strcut btrfs_item and smaller than
2255  * struct btrfs_key_ptr.  Since we're currently limited to the page size
2256  * for a btree block, there's no way for a kmalloc of refsorts for a
2257  * single node to be bigger than a page.
2258  */
2259 struct refsort {
2260         u64 bytenr;
2261         u32 slot;
2262 };
2263
2264 /*
2265  * for passing into sort()
2266  */
2267 static int refsort_cmp(const void *a_void, const void *b_void)
2268 {
2269         const struct refsort *a = a_void;
2270         const struct refsort *b = b_void;
2271
2272         if (a->bytenr < b->bytenr)
2273                 return -1;
2274         if (a->bytenr > b->bytenr)
2275                 return 1;
2276         return 0;
2277 }
2278 #endif
2279
2280 static int __btrfs_mod_ref(struct btrfs_trans_handle *trans,
2281                            struct btrfs_root *root,
2282                            struct extent_buffer *buf,
2283                            int full_backref, int inc)
2284 {
2285         u64 bytenr;
2286         u64 num_bytes;
2287         u64 parent;
2288         u64 ref_root;
2289         u32 nritems;
2290         struct btrfs_key key;
2291         struct btrfs_file_extent_item *fi;
2292         int i;
2293         int level;
2294         int ret = 0;
2295         int (*process_func)(struct btrfs_trans_handle *, struct btrfs_root *,
2296                             u64, u64, u64, u64, u64, u64);
2297
2298         ref_root = btrfs_header_owner(buf);
2299         nritems = btrfs_header_nritems(buf);
2300         level = btrfs_header_level(buf);
2301
2302         if (!root->ref_cows && level == 0)
2303                 return 0;
2304
2305         if (inc)
2306                 process_func = btrfs_inc_extent_ref;
2307         else
2308                 process_func = btrfs_free_extent;
2309
2310         if (full_backref)
2311                 parent = buf->start;
2312         else
2313                 parent = 0;
2314
2315         for (i = 0; i < nritems; i++) {
2316                 if (level == 0) {
2317                         btrfs_item_key_to_cpu(buf, &key, i);
2318                         if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
2319                                 continue;
2320                         fi = btrfs_item_ptr(buf, i,
2321                                             struct btrfs_file_extent_item);
2322                         if (btrfs_file_extent_type(buf, fi) ==
2323                             BTRFS_FILE_EXTENT_INLINE)
2324                                 continue;
2325                         bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
2326                         if (bytenr == 0)
2327                                 continue;
2328
2329                         num_bytes = btrfs_file_extent_disk_num_bytes(buf, fi);
2330                         key.offset -= btrfs_file_extent_offset(buf, fi);
2331                         ret = process_func(trans, root, bytenr, num_bytes,
2332                                            parent, ref_root, key.objectid,
2333                                            key.offset);
2334                         if (ret)
2335                                 goto fail;
2336                 } else {
2337                         bytenr = btrfs_node_blockptr(buf, i);
2338                         num_bytes = btrfs_level_size(root, level - 1);
2339                         ret = process_func(trans, root, bytenr, num_bytes,
2340                                            parent, ref_root, level - 1, 0);
2341                         if (ret)
2342                                 goto fail;
2343                 }
2344         }
2345         return 0;
2346 fail:
2347         BUG();
2348         return ret;
2349 }
2350
2351 int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
2352                   struct extent_buffer *buf, int full_backref)
2353 {
2354         return __btrfs_mod_ref(trans, root, buf, full_backref, 1);
2355 }
2356
2357 int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
2358                   struct extent_buffer *buf, int full_backref)
2359 {
2360         return __btrfs_mod_ref(trans, root, buf, full_backref, 0);
2361 }
2362
2363 static int write_one_cache_group(struct btrfs_trans_handle *trans,
2364                                  struct btrfs_root *root,
2365                                  struct btrfs_path *path,
2366                                  struct btrfs_block_group_cache *cache)
2367 {
2368         int ret;
2369         struct btrfs_root *extent_root = root->fs_info->extent_root;
2370         unsigned long bi;
2371         struct extent_buffer *leaf;
2372
2373         ret = btrfs_search_slot(trans, extent_root, &cache->key, path, 0, 1);
2374         if (ret < 0)
2375                 goto fail;
2376         BUG_ON(ret);
2377
2378         leaf = path->nodes[0];
2379         bi = btrfs_item_ptr_offset(leaf, path->slots[0]);
2380         write_extent_buffer(leaf, &cache->item, bi, sizeof(cache->item));
2381         btrfs_mark_buffer_dirty(leaf);
2382         btrfs_release_path(extent_root, path);
2383 fail:
2384         if (ret)
2385                 return ret;
2386         return 0;
2387
2388 }
2389
2390 int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans,
2391                                    struct btrfs_root *root)
2392 {
2393         struct btrfs_block_group_cache *cache, *entry;
2394         struct rb_node *n;
2395         int err = 0;
2396         int werr = 0;
2397         struct btrfs_path *path;
2398         u64 last = 0;
2399
2400         path = btrfs_alloc_path();
2401         if (!path)
2402                 return -ENOMEM;
2403
2404         while (1) {
2405                 cache = NULL;
2406                 spin_lock(&root->fs_info->block_group_cache_lock);
2407                 for (n = rb_first(&root->fs_info->block_group_cache_tree);
2408                      n; n = rb_next(n)) {
2409                         entry = rb_entry(n, struct btrfs_block_group_cache,
2410                                          cache_node);
2411                         if (entry->dirty) {
2412                                 cache = entry;
2413                                 break;
2414                         }
2415                 }
2416                 spin_unlock(&root->fs_info->block_group_cache_lock);
2417
2418                 if (!cache)
2419                         break;
2420
2421                 cache->dirty = 0;
2422                 last += cache->key.offset;
2423
2424                 err = write_one_cache_group(trans, root,
2425                                             path, cache);
2426                 /*
2427                  * if we fail to write the cache group, we want
2428                  * to keep it marked dirty in hopes that a later
2429                  * write will work
2430                  */
2431                 if (err) {
2432                         werr = err;
2433                         continue;
2434                 }
2435         }
2436         btrfs_free_path(path);
2437         return werr;
2438 }
2439
2440 int btrfs_extent_readonly(struct btrfs_root *root, u64 bytenr)
2441 {
2442         struct btrfs_block_group_cache *block_group;
2443         int readonly = 0;
2444
2445         block_group = btrfs_lookup_block_group(root->fs_info, bytenr);
2446         if (!block_group || block_group->ro)
2447                 readonly = 1;
2448         if (block_group)
2449                 btrfs_put_block_group(block_group);
2450         return readonly;
2451 }
2452
2453 static int update_space_info(struct btrfs_fs_info *info, u64 flags,
2454                              u64 total_bytes, u64 bytes_used,
2455                              struct btrfs_space_info **space_info)
2456 {
2457         struct btrfs_space_info *found;
2458
2459         found = __find_space_info(info, flags);
2460         if (found) {
2461                 spin_lock(&found->lock);
2462                 found->total_bytes += total_bytes;
2463                 found->bytes_used += bytes_used;
2464                 found->full = 0;
2465                 spin_unlock(&found->lock);
2466                 *space_info = found;
2467                 return 0;
2468         }
2469         found = kzalloc(sizeof(*found), GFP_NOFS);
2470         if (!found)
2471                 return -ENOMEM;
2472
2473         INIT_LIST_HEAD(&found->block_groups);
2474         init_rwsem(&found->groups_sem);
2475         spin_lock_init(&found->lock);
2476         found->flags = flags;
2477         found->total_bytes = total_bytes;
2478         found->bytes_used = bytes_used;
2479         found->bytes_pinned = 0;
2480         found->bytes_reserved = 0;
2481         found->bytes_readonly = 0;
2482         found->bytes_delalloc = 0;
2483         found->full = 0;
2484         found->force_alloc = 0;
2485         *space_info = found;
2486         list_add_rcu(&found->list, &info->space_info);
2487         return 0;
2488 }
2489
2490 static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
2491 {
2492         u64 extra_flags = flags & (BTRFS_BLOCK_GROUP_RAID0 |
2493                                    BTRFS_BLOCK_GROUP_RAID1 |
2494                                    BTRFS_BLOCK_GROUP_RAID10 |
2495                                    BTRFS_BLOCK_GROUP_DUP);
2496         if (extra_flags) {
2497                 if (flags & BTRFS_BLOCK_GROUP_DATA)
2498                         fs_info->avail_data_alloc_bits |= extra_flags;
2499                 if (flags & BTRFS_BLOCK_GROUP_METADATA)
2500                         fs_info->avail_metadata_alloc_bits |= extra_flags;
2501                 if (flags & BTRFS_BLOCK_GROUP_SYSTEM)
2502                         fs_info->avail_system_alloc_bits |= extra_flags;
2503         }
2504 }
2505
2506 static void set_block_group_readonly(struct btrfs_block_group_cache *cache)
2507 {
2508         spin_lock(&cache->space_info->lock);
2509         spin_lock(&cache->lock);
2510         if (!cache->ro) {
2511                 cache->space_info->bytes_readonly += cache->key.offset -
2512                                         btrfs_block_group_used(&cache->item);
2513                 cache->ro = 1;
2514         }
2515         spin_unlock(&cache->lock);
2516         spin_unlock(&cache->space_info->lock);
2517 }
2518
2519 u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags)
2520 {
2521         u64 num_devices = root->fs_info->fs_devices->rw_devices;
2522
2523         if (num_devices == 1)
2524                 flags &= ~(BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID0);
2525         if (num_devices < 4)
2526                 flags &= ~BTRFS_BLOCK_GROUP_RAID10;
2527
2528         if ((flags & BTRFS_BLOCK_GROUP_DUP) &&
2529             (flags & (BTRFS_BLOCK_GROUP_RAID1 |
2530                       BTRFS_BLOCK_GROUP_RAID10))) {
2531                 flags &= ~BTRFS_BLOCK_GROUP_DUP;
2532         }
2533
2534         if ((flags & BTRFS_BLOCK_GROUP_RAID1) &&
2535             (flags & BTRFS_BLOCK_GROUP_RAID10)) {
2536                 flags &= ~BTRFS_BLOCK_GROUP_RAID1;
2537         }
2538
2539         if ((flags & BTRFS_BLOCK_GROUP_RAID0) &&
2540             ((flags & BTRFS_BLOCK_GROUP_RAID1) |
2541              (flags & BTRFS_BLOCK_GROUP_RAID10) |
2542              (flags & BTRFS_BLOCK_GROUP_DUP)))
2543                 flags &= ~BTRFS_BLOCK_GROUP_RAID0;
2544         return flags;
2545 }
2546
2547 static u64 btrfs_get_alloc_profile(struct btrfs_root *root, u64 data)
2548 {
2549         struct btrfs_fs_info *info = root->fs_info;
2550         u64 alloc_profile;
2551
2552         if (data) {
2553                 alloc_profile = info->avail_data_alloc_bits &
2554                         info->data_alloc_profile;
2555                 data = BTRFS_BLOCK_GROUP_DATA | alloc_profile;
2556         } else if (root == root->fs_info->chunk_root) {
2557                 alloc_profile = info->avail_system_alloc_bits &
2558                         info->system_alloc_profile;
2559                 data = BTRFS_BLOCK_GROUP_SYSTEM | alloc_profile;
2560         } else {
2561                 alloc_profile = info->avail_metadata_alloc_bits &
2562                         info->metadata_alloc_profile;
2563                 data = BTRFS_BLOCK_GROUP_METADATA | alloc_profile;
2564         }
2565
2566         return btrfs_reduce_alloc_profile(root, data);
2567 }
2568
2569 void btrfs_set_inode_space_info(struct btrfs_root *root, struct inode *inode)
2570 {
2571         u64 alloc_target;
2572
2573         alloc_target = btrfs_get_alloc_profile(root, 1);
2574         BTRFS_I(inode)->space_info = __find_space_info(root->fs_info,
2575                                                        alloc_target);
2576 }
2577
2578 /*
2579  * for now this just makes sure we have at least 5% of our metadata space free
2580  * for use.
2581  */
2582 int btrfs_check_metadata_free_space(struct btrfs_root *root)
2583 {
2584         struct btrfs_fs_info *info = root->fs_info;
2585         struct btrfs_space_info *meta_sinfo;
2586         u64 alloc_target, thresh;
2587         int committed = 0, ret;
2588
2589         /* get the space info for where the metadata will live */
2590         alloc_target = btrfs_get_alloc_profile(root, 0);
2591         meta_sinfo = __find_space_info(info, alloc_target);
2592
2593 again:
2594         spin_lock(&meta_sinfo->lock);
2595         if (!meta_sinfo->full)
2596                 thresh = meta_sinfo->total_bytes * 80;
2597         else
2598                 thresh = meta_sinfo->total_bytes * 95;
2599
2600         do_div(thresh, 100);
2601
2602         if (meta_sinfo->bytes_used + meta_sinfo->bytes_reserved +
2603             meta_sinfo->bytes_pinned + meta_sinfo->bytes_readonly > thresh) {
2604                 struct btrfs_trans_handle *trans;
2605                 if (!meta_sinfo->full) {
2606                         meta_sinfo->force_alloc = 1;
2607                         spin_unlock(&meta_sinfo->lock);
2608
2609                         trans = btrfs_start_transaction(root, 1);
2610                         if (!trans)
2611                                 return -ENOMEM;
2612
2613                         ret = do_chunk_alloc(trans, root->fs_info->extent_root,
2614                                              2 * 1024 * 1024, alloc_target, 0);
2615                         btrfs_end_transaction(trans, root);
2616                         goto again;
2617                 }
2618                 spin_unlock(&meta_sinfo->lock);
2619
2620                 if (!committed) {
2621                         committed = 1;
2622                         trans = btrfs_join_transaction(root, 1);
2623                         if (!trans)
2624                                 return -ENOMEM;
2625                         ret = btrfs_commit_transaction(trans, root);
2626                         if (ret)
2627                                 return ret;
2628                         goto again;
2629                 }
2630                 return -ENOSPC;
2631         }
2632         spin_unlock(&meta_sinfo->lock);
2633
2634         return 0;
2635 }
2636
2637 /*
2638  * This will check the space that the inode allocates from to make sure we have
2639  * enough space for bytes.
2640  */
2641 int btrfs_check_data_free_space(struct btrfs_root *root, struct inode *inode,
2642                                 u64 bytes)
2643 {
2644         struct btrfs_space_info *data_sinfo;
2645         int ret = 0, committed = 0;
2646
2647         /* make sure bytes are sectorsize aligned */
2648         bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1);
2649
2650         data_sinfo = BTRFS_I(inode)->space_info;
2651 again:
2652         /* make sure we have enough space to handle the data first */
2653         spin_lock(&data_sinfo->lock);
2654         if (data_sinfo->total_bytes - data_sinfo->bytes_used -
2655             data_sinfo->bytes_delalloc - data_sinfo->bytes_reserved -
2656             data_sinfo->bytes_pinned - data_sinfo->bytes_readonly -
2657             data_sinfo->bytes_may_use < bytes) {
2658                 struct btrfs_trans_handle *trans;
2659
2660                 /*
2661                  * if we don't have enough free bytes in this space then we need
2662                  * to alloc a new chunk.
2663                  */
2664                 if (!data_sinfo->full) {
2665                         u64 alloc_target;
2666
2667                         data_sinfo->force_alloc = 1;
2668                         spin_unlock(&data_sinfo->lock);
2669
2670                         alloc_target = btrfs_get_alloc_profile(root, 1);
2671                         trans = btrfs_start_transaction(root, 1);
2672                         if (!trans)
2673                                 return -ENOMEM;
2674
2675                         ret = do_chunk_alloc(trans, root->fs_info->extent_root,
2676                                              bytes + 2 * 1024 * 1024,
2677                                              alloc_target, 0);
2678                         btrfs_end_transaction(trans, root);
2679                         if (ret)
2680                                 return ret;
2681                         goto again;
2682                 }
2683                 spin_unlock(&data_sinfo->lock);
2684
2685                 /* commit the current transaction and try again */
2686                 if (!committed) {
2687                         committed = 1;
2688                         trans = btrfs_join_transaction(root, 1);
2689                         if (!trans)
2690                                 return -ENOMEM;
2691                         ret = btrfs_commit_transaction(trans, root);
2692                         if (ret)
2693                                 return ret;
2694                         goto again;
2695                 }
2696
2697                 printk(KERN_ERR "no space left, need %llu, %llu delalloc bytes"
2698                        ", %llu bytes_used, %llu bytes_reserved, "
2699                        "%llu bytes_pinned, %llu bytes_readonly, %llu may use "
2700                        "%llu total\n", (unsigned long long)bytes,
2701                        (unsigned long long)data_sinfo->bytes_delalloc,
2702                        (unsigned long long)data_sinfo->bytes_used,
2703                        (unsigned long long)data_sinfo->bytes_reserved,
2704                        (unsigned long long)data_sinfo->bytes_pinned,
2705                        (unsigned long long)data_sinfo->bytes_readonly,
2706                        (unsigned long long)data_sinfo->bytes_may_use,
2707                        (unsigned long long)data_sinfo->total_bytes);
2708                 return -ENOSPC;
2709         }
2710         data_sinfo->bytes_may_use += bytes;
2711         BTRFS_I(inode)->reserved_bytes += bytes;
2712         spin_unlock(&data_sinfo->lock);
2713
2714         return btrfs_check_metadata_free_space(root);
2715 }
2716
2717 /*
2718  * if there was an error for whatever reason after calling
2719  * btrfs_check_data_free_space, call this so we can cleanup the counters.
2720  */
2721 void btrfs_free_reserved_data_space(struct btrfs_root *root,
2722                                     struct inode *inode, u64 bytes)
2723 {
2724         struct btrfs_space_info *data_sinfo;
2725
2726         /* make sure bytes are sectorsize aligned */
2727         bytes = (bytes + root->sectorsize - 1) & ~((u64)root->sectorsize - 1);
2728
2729         data_sinfo = BTRFS_I(inode)->space_info;
2730         spin_lock(&data_sinfo->lock);
2731         data_sinfo->bytes_may_use -= bytes;
2732         BTRFS_I(inode)->reserved_bytes -= bytes;
2733         spin_unlock(&data_sinfo->lock);
2734 }
2735
2736 /* called when we are adding a delalloc extent to the inode's io_tree */
2737 void btrfs_delalloc_reserve_space(struct btrfs_root *root, struct inode *inode,
2738                                   u64 bytes)
2739 {
2740         struct btrfs_space_info *data_sinfo;
2741
2742         /* get the space info for where this inode will be storing its data */
2743         data_sinfo = BTRFS_I(inode)->space_info;
2744
2745         /* make sure we have enough space to handle the data first */
2746         spin_lock(&data_sinfo->lock);
2747         data_sinfo->bytes_delalloc += bytes;
2748
2749         /*
2750          * we are adding a delalloc extent without calling
2751          * btrfs_check_data_free_space first.  This happens on a weird
2752          * writepage condition, but shouldn't hurt our accounting
2753          */
2754         if (unlikely(bytes > BTRFS_I(inode)->reserved_bytes)) {
2755                 data_sinfo->bytes_may_use -= BTRFS_I(inode)->reserved_bytes;
2756                 BTRFS_I(inode)->reserved_bytes = 0;
2757         } else {
2758                 data_sinfo->bytes_may_use -= bytes;
2759                 BTRFS_I(inode)->reserved_bytes -= bytes;
2760         }
2761
2762         spin_unlock(&data_sinfo->lock);
2763 }
2764
2765 /* called when we are clearing an delalloc extent from the inode's io_tree */
2766 void btrfs_delalloc_free_space(struct btrfs_root *root, struct inode *inode,
2767                               u64 bytes)
2768 {
2769         struct btrfs_space_info *info;
2770
2771         info = BTRFS_I(inode)->space_info;
2772
2773         spin_lock(&info->lock);
2774         info->bytes_delalloc -= bytes;
2775         spin_unlock(&info->lock);
2776 }
2777
2778 static void force_metadata_allocation(struct btrfs_fs_info *info)
2779 {
2780         struct list_head *head = &info->space_info;
2781         struct btrfs_space_info *found;
2782
2783         rcu_read_lock();
2784         list_for_each_entry_rcu(found, head, list) {
2785                 if (found->flags & BTRFS_BLOCK_GROUP_METADATA)
2786                         found->force_alloc = 1;
2787         }
2788         rcu_read_unlock();
2789 }
2790
2791 static int do_chunk_alloc(struct btrfs_trans_handle *trans,
2792                           struct btrfs_root *extent_root, u64 alloc_bytes,
2793                           u64 flags, int force)
2794 {
2795         struct btrfs_space_info *space_info;
2796         struct btrfs_fs_info *fs_info = extent_root->fs_info;
2797         u64 thresh;
2798         int ret = 0;
2799
2800         mutex_lock(&fs_info->chunk_mutex);
2801
2802         flags = btrfs_reduce_alloc_profile(extent_root, flags);
2803
2804         space_info = __find_space_info(extent_root->fs_info, flags);
2805         if (!space_info) {
2806                 ret = update_space_info(extent_root->fs_info, flags,
2807                                         0, 0, &space_info);
2808                 BUG_ON(ret);
2809         }
2810         BUG_ON(!space_info);
2811
2812         spin_lock(&space_info->lock);
2813         if (space_info->force_alloc) {
2814                 force = 1;
2815                 space_info->force_alloc = 0;
2816         }
2817         if (space_info->full) {
2818                 spin_unlock(&space_info->lock);
2819                 goto out;
2820         }
2821
2822         thresh = space_info->total_bytes - space_info->bytes_readonly;
2823         thresh = div_factor(thresh, 6);
2824         if (!force &&
2825            (space_info->bytes_used + space_info->bytes_pinned +
2826             space_info->bytes_reserved + alloc_bytes) < thresh) {
2827                 spin_unlock(&space_info->lock);
2828                 goto out;
2829         }
2830         spin_unlock(&space_info->lock);
2831
2832         /*
2833          * if we're doing a data chunk, go ahead and make sure that
2834          * we keep a reasonable number of metadata chunks allocated in the
2835          * FS as well.
2836          */
2837         if (flags & BTRFS_BLOCK_GROUP_DATA) {
2838                 fs_info->data_chunk_allocations++;
2839                 if (!(fs_info->data_chunk_allocations %
2840                       fs_info->metadata_ratio))
2841                         force_metadata_allocation(fs_info);
2842         }
2843
2844         ret = btrfs_alloc_chunk(trans, extent_root, flags);
2845         if (ret)
2846                 space_info->full = 1;
2847 out:
2848         mutex_unlock(&extent_root->fs_info->chunk_mutex);
2849         return ret;
2850 }
2851
2852 static int update_block_group(struct btrfs_trans_handle *trans,
2853                               struct btrfs_root *root,
2854                               u64 bytenr, u64 num_bytes, int alloc,
2855                               int mark_free)
2856 {
2857         struct btrfs_block_group_cache *cache;
2858         struct btrfs_fs_info *info = root->fs_info;
2859         u64 total = num_bytes;
2860         u64 old_val;
2861         u64 byte_in_group;
2862
2863         /* block accounting for super block */
2864         spin_lock(&info->delalloc_lock);
2865         old_val = btrfs_super_bytes_used(&info->super_copy);
2866         if (alloc)
2867                 old_val += num_bytes;
2868         else
2869                 old_val -= num_bytes;
2870         btrfs_set_super_bytes_used(&info->super_copy, old_val);
2871
2872         /* block accounting for root item */
2873         old_val = btrfs_root_used(&root->root_item);
2874         if (alloc)
2875                 old_val += num_bytes;
2876         else
2877                 old_val -= num_bytes;
2878         btrfs_set_root_used(&root->root_item, old_val);
2879         spin_unlock(&info->delalloc_lock);
2880
2881         while (total) {
2882                 cache = btrfs_lookup_block_group(info, bytenr);
2883                 if (!cache)
2884                         return -1;
2885                 byte_in_group = bytenr - cache->key.objectid;
2886                 WARN_ON(byte_in_group > cache->key.offset);
2887
2888                 spin_lock(&cache->space_info->lock);
2889                 spin_lock(&cache->lock);
2890                 cache->dirty = 1;
2891                 old_val = btrfs_block_group_used(&cache->item);
2892                 num_bytes = min(total, cache->key.offset - byte_in_group);
2893                 if (alloc) {
2894                         old_val += num_bytes;
2895                         cache->space_info->bytes_used += num_bytes;
2896                         if (cache->ro)
2897                                 cache->space_info->bytes_readonly -= num_bytes;
2898                         btrfs_set_block_group_used(&cache->item, old_val);
2899                         spin_unlock(&cache->lock);
2900                         spin_unlock(&cache->space_info->lock);
2901                 } else {
2902                         old_val -= num_bytes;
2903                         cache->space_info->bytes_used -= num_bytes;
2904                         if (cache->ro)
2905                                 cache->space_info->bytes_readonly += num_bytes;
2906                         btrfs_set_block_group_used(&cache->item, old_val);
2907                         spin_unlock(&cache->lock);
2908                         spin_unlock(&cache->space_info->lock);
2909                         if (mark_free) {
2910                                 int ret;
2911
2912                                 ret = btrfs_discard_extent(root, bytenr,
2913                                                            num_bytes);
2914                                 WARN_ON(ret);
2915
2916                                 ret = btrfs_add_free_space(cache, bytenr,
2917                                                            num_bytes);
2918                                 WARN_ON(ret);
2919                         }
2920                 }
2921                 btrfs_put_block_group(cache);
2922                 total -= num_bytes;
2923                 bytenr += num_bytes;
2924         }
2925         return 0;
2926 }
2927
2928 static u64 first_logical_byte(struct btrfs_root *root, u64 search_start)
2929 {
2930         struct btrfs_block_group_cache *cache;
2931         u64 bytenr;
2932
2933         cache = btrfs_lookup_first_block_group(root->fs_info, search_start);
2934         if (!cache)
2935                 return 0;
2936
2937         bytenr = cache->key.objectid;
2938         btrfs_put_block_group(cache);
2939
2940         return bytenr;
2941 }
2942
2943 int btrfs_update_pinned_extents(struct btrfs_root *root,
2944                                 u64 bytenr, u64 num, int pin)
2945 {
2946         u64 len;
2947         struct btrfs_block_group_cache *cache;
2948         struct btrfs_fs_info *fs_info = root->fs_info;
2949
2950         if (pin) {
2951                 set_extent_dirty(&fs_info->pinned_extents,
2952                                 bytenr, bytenr + num - 1, GFP_NOFS);
2953         } else {
2954                 clear_extent_dirty(&fs_info->pinned_extents,
2955                                 bytenr, bytenr + num - 1, GFP_NOFS);
2956         }
2957
2958         while (num > 0) {
2959                 cache = btrfs_lookup_block_group(fs_info, bytenr);
2960                 BUG_ON(!cache);
2961                 len = min(num, cache->key.offset -
2962                           (bytenr - cache->key.objectid));
2963                 if (pin) {
2964                         spin_lock(&cache->space_info->lock);
2965                         spin_lock(&cache->lock);
2966                         cache->pinned += len;
2967                         cache->space_info->bytes_pinned += len;
2968                         spin_unlock(&cache->lock);
2969                         spin_unlock(&cache->space_info->lock);
2970                         fs_info->total_pinned += len;
2971                 } else {
2972                         spin_lock(&cache->space_info->lock);
2973                         spin_lock(&cache->lock);
2974                         cache->pinned -= len;
2975                         cache->space_info->bytes_pinned -= len;
2976                         spin_unlock(&cache->lock);
2977                         spin_unlock(&cache->space_info->lock);
2978                         fs_info->total_pinned -= len;
2979                         if (cache->cached)
2980                                 btrfs_add_free_space(cache, bytenr, len);
2981                 }
2982                 btrfs_put_block_group(cache);
2983                 bytenr += len;
2984                 num -= len;
2985         }
2986         return 0;
2987 }
2988
2989 static int update_reserved_extents(struct btrfs_root *root,
2990                                    u64 bytenr, u64 num, int reserve)
2991 {
2992         u64 len;
2993         struct btrfs_block_group_cache *cache;
2994         struct btrfs_fs_info *fs_info = root->fs_info;
2995
2996         while (num > 0) {
2997                 cache = btrfs_lookup_block_group(fs_info, bytenr);
2998                 BUG_ON(!cache);
2999                 len = min(num, cache->key.offset -
3000                           (bytenr - cache->key.objectid));
3001
3002                 spin_lock(&cache->space_info->lock);
3003                 spin_lock(&cache->lock);
3004                 if (reserve) {
3005                         cache->reserved += len;
3006                         cache->space_info->bytes_reserved += len;
3007                 } else {
3008                         cache->reserved -= len;
3009                         cache->space_info->bytes_reserved -= len;
3010                 }
3011                 spin_unlock(&cache->lock);
3012                 spin_unlock(&cache->space_info->lock);
3013                 btrfs_put_block_group(cache);
3014                 bytenr += len;
3015                 num -= len;
3016         }
3017         return 0;
3018 }
3019
3020 int btrfs_copy_pinned(struct btrfs_root *root, struct extent_io_tree *copy)
3021 {
3022         u64 last = 0;
3023         u64 start;
3024         u64 end;
3025         struct extent_io_tree *pinned_extents = &root->fs_info->pinned_extents;
3026         int ret;
3027
3028         while (1) {
3029                 ret = find_first_extent_bit(pinned_extents, last,
3030                                             &start, &end, EXTENT_DIRTY);
3031                 if (ret)
3032                         break;
3033                 set_extent_dirty(copy, start, end, GFP_NOFS);
3034                 last = end + 1;
3035         }
3036         return 0;
3037 }
3038
3039 int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
3040                                struct btrfs_root *root,
3041                                struct extent_io_tree *unpin)
3042 {
3043         u64 start;
3044         u64 end;
3045         int ret;
3046
3047         while (1) {
3048                 ret = find_first_extent_bit(unpin, 0, &start, &end,
3049                                             EXTENT_DIRTY);
3050                 if (ret)
3051                         break;
3052
3053                 ret = btrfs_discard_extent(root, start, end + 1 - start);
3054
3055                 /* unlocks the pinned mutex */
3056                 btrfs_update_pinned_extents(root, start, end + 1 - start, 0);
3057                 clear_extent_dirty(unpin, start, end, GFP_NOFS);
3058
3059                 cond_resched();
3060         }
3061         return ret;
3062 }
3063
3064 static int pin_down_bytes(struct btrfs_trans_handle *trans,
3065                           struct btrfs_root *root,
3066                           struct btrfs_path *path,
3067                           u64 bytenr, u64 num_bytes, int is_data,
3068                           struct extent_buffer **must_clean)
3069 {
3070         int err = 0;
3071         struct extent_buffer *buf;
3072
3073         if (is_data)
3074                 goto pinit;
3075
3076         buf = btrfs_find_tree_block(root, bytenr, num_bytes);
3077         if (!buf)
3078                 goto pinit;
3079
3080         /* we can reuse a block if it hasn't been written
3081          * and it is from this transaction.  We can't
3082          * reuse anything from the tree log root because
3083          * it has tiny sub-transactions.
3084          */
3085         if (btrfs_buffer_uptodate(buf, 0) &&
3086             btrfs_try_tree_lock(buf)) {
3087                 u64 header_owner = btrfs_header_owner(buf);
3088                 u64 header_transid = btrfs_header_generation(buf);
3089                 if (header_owner != BTRFS_TREE_LOG_OBJECTID &&
3090                     header_transid == trans->transid &&
3091                     !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) {
3092                         *must_clean = buf;
3093                         return 1;
3094                 }
3095                 btrfs_tree_unlock(buf);
3096         }
3097         free_extent_buffer(buf);
3098 pinit:
3099         btrfs_set_path_blocking(path);
3100         /* unlocks the pinned mutex */
3101         btrfs_update_pinned_extents(root, bytenr, num_bytes, 1);
3102
3103         BUG_ON(err < 0);
3104         return 0;
3105 }
3106
3107
3108 static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
3109                                 struct btrfs_root *root,
3110                                 u64 bytenr, u64 num_bytes, u64 parent,
3111                                 u64 root_objectid, u64 owner_objectid,
3112                                 u64 owner_offset, int refs_to_drop,
3113                                 struct btrfs_delayed_extent_op *extent_op)
3114 {
3115         struct btrfs_key key;
3116         struct btrfs_path *path;
3117         struct btrfs_fs_info *info = root->fs_info;
3118         struct btrfs_root *extent_root = info->extent_root;
3119         struct extent_buffer *leaf;
3120         struct btrfs_extent_item *ei;
3121         struct btrfs_extent_inline_ref *iref;
3122         int ret;
3123         int is_data;
3124         int extent_slot = 0;
3125         int found_extent = 0;
3126         int num_to_del = 1;
3127         u32 item_size;
3128         u64 refs;
3129
3130         path = btrfs_alloc_path();
3131         if (!path)
3132                 return -ENOMEM;
3133
3134         path->reada = 1;
3135         path->leave_spinning = 1;
3136
3137         is_data = owner_objectid >= BTRFS_FIRST_FREE_OBJECTID;
3138         BUG_ON(!is_data && refs_to_drop != 1);
3139
3140         ret = lookup_extent_backref(trans, extent_root, path, &iref,
3141                                     bytenr, num_bytes, parent,
3142                                     root_objectid, owner_objectid,
3143                                     owner_offset);
3144         if (ret == 0) {
3145                 extent_slot = path->slots[0];
3146                 while (extent_slot >= 0) {
3147                         btrfs_item_key_to_cpu(path->nodes[0], &key,
3148                                               extent_slot);
3149                         if (key.objectid != bytenr)
3150                                 break;
3151                         if (key.type == BTRFS_EXTENT_ITEM_KEY &&
3152                             key.offset == num_bytes) {
3153                                 found_extent = 1;
3154                                 break;
3155                         }
3156                         if (path->slots[0] - extent_slot > 5)
3157                                 break;
3158                         extent_slot--;
3159                 }
3160 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
3161                 item_size = btrfs_item_size_nr(path->nodes[0], extent_slot);
3162                 if (found_extent && item_size < sizeof(*ei))
3163                         found_extent = 0;
3164 #endif
3165                 if (!found_extent) {
3166                         BUG_ON(iref);
3167                         ret = remove_extent_backref(trans, extent_root, path,
3168                                                     NULL, refs_to_drop,
3169                                                     is_data);
3170                         BUG_ON(ret);
3171                         btrfs_release_path(extent_root, path);
3172                         path->leave_spinning = 1;
3173
3174                         key.objectid = bytenr;
3175                         key.type = BTRFS_EXTENT_ITEM_KEY;
3176                         key.offset = num_bytes;
3177
3178                         ret = btrfs_search_slot(trans, extent_root,
3179                                                 &key, path, -1, 1);
3180                         if (ret) {
3181                                 printk(KERN_ERR "umm, got %d back from search"
3182                                        ", was looking for %llu\n", ret,
3183                                        (unsigned long long)bytenr);
3184                                 btrfs_print_leaf(extent_root, path->nodes[0]);
3185                         }
3186                         BUG_ON(ret);
3187                         extent_slot = path->slots[0];
3188                 }
3189         } else {
3190                 btrfs_print_leaf(extent_root, path->nodes[0]);
3191                 WARN_ON(1);
3192                 printk(KERN_ERR "btrfs unable to find ref byte nr %llu "
3193                        "parent %llu root %llu  owner %llu offset %llu\n",
3194                        (unsigned long long)bytenr,
3195                        (unsigned long long)parent,
3196                        (unsigned long long)root_objectid,
3197                        (unsigned long long)owner_objectid,
3198                        (unsigned long long)owner_offset);
3199         }
3200
3201         leaf = path->nodes[0];
3202         item_size = btrfs_item_size_nr(leaf, extent_slot);
3203 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
3204         if (item_size < sizeof(*ei)) {
3205                 BUG_ON(found_extent || extent_slot != path->slots[0]);
3206                 ret = convert_extent_item_v0(trans, extent_root, path,
3207                                              owner_objectid, 0);
3208                 BUG_ON(ret < 0);
3209
3210                 btrfs_release_path(extent_root, path);
3211                 path->leave_spinning = 1;
3212
3213                 key.objectid = bytenr;
3214                 key.type = BTRFS_EXTENT_ITEM_KEY;
3215                 key.offset = num_bytes;
3216
3217                 ret = btrfs_search_slot(trans, extent_root, &key, path,
3218                                         -1, 1);
3219                 if (ret) {
3220                         printk(KERN_ERR "umm, got %d back from search"
3221                                ", was looking for %llu\n", ret,
3222                                (unsigned long long)bytenr);
3223                         btrfs_print_leaf(extent_root, path->nodes[0]);
3224                 }
3225                 BUG_ON(ret);
3226                 extent_slot = path->slots[0];
3227                 leaf = path->nodes[0];
3228                 item_size = btrfs_item_size_nr(leaf, extent_slot);
3229         }
3230 #endif
3231         BUG_ON(item_size < sizeof(*ei));
3232         ei = btrfs_item_ptr(leaf, extent_slot,
3233                             struct btrfs_extent_item);
3234         if (owner_objectid < BTRFS_FIRST_FREE_OBJECTID) {
3235                 struct btrfs_tree_block_info *bi;
3236                 BUG_ON(item_size < sizeof(*ei) + sizeof(*bi));
3237                 bi = (struct btrfs_tree_block_info *)(ei + 1);
3238                 WARN_ON(owner_objectid != btrfs_tree_block_level(leaf, bi));
3239         }
3240
3241         refs = btrfs_extent_refs(leaf, ei);
3242         BUG_ON(refs < refs_to_drop);
3243         refs -= refs_to_drop;
3244
3245         if (refs > 0) {
3246                 if (extent_op)
3247                         __run_delayed_extent_op(extent_op, leaf, ei);
3248                 /*
3249                  * In the case of inline back ref, reference count will
3250                  * be updated by remove_extent_backref
3251                  */
3252                 if (iref) {
3253                         BUG_ON(!found_extent);
3254                 } else {
3255                         btrfs_set_extent_refs(leaf, ei, refs);
3256                         btrfs_mark_buffer_dirty(leaf);
3257                 }
3258                 if (found_extent) {
3259                         ret = remove_extent_backref(trans, extent_root, path,
3260                                                     iref, refs_to_drop,
3261                                                     is_data);
3262                         BUG_ON(ret);
3263                 }
3264         } else {
3265                 int mark_free = 0;
3266                 struct extent_buffer *must_clean = NULL;
3267
3268                 if (found_extent) {
3269                         BUG_ON(is_data && refs_to_drop !=
3270                                extent_data_ref_count(root, path, iref));
3271                         if (iref) {
3272                                 BUG_ON(path->slots[0] != extent_slot);
3273                         } else {
3274                                 BUG_ON(path->slots[0] != extent_slot + 1);
3275                                 path->slots[0] = extent_slot;
3276                                 num_to_del = 2;
3277                         }
3278                 }
3279
3280                 ret = pin_down_bytes(trans, root, path, bytenr,
3281                                      num_bytes, is_data, &must_clean);
3282                 if (ret > 0)
3283                         mark_free = 1;
3284                 BUG_ON(ret < 0);
3285                 /*
3286                  * it is going to be very rare for someone to be waiting
3287                  * on the block we're freeing.  del_items might need to
3288                  * schedule, so rather than get fancy, just force it
3289                  * to blocking here
3290                  */
3291                 if (must_clean)
3292                         btrfs_set_lock_blocking(must_clean);
3293
3294                 ret = btrfs_del_items(trans, extent_root, path, path->slots[0],
3295                                       num_to_del);
3296                 BUG_ON(ret);
3297                 btrfs_release_path(extent_root, path);
3298
3299                 if (must_clean) {
3300                         clean_tree_block(NULL, root, must_clean);
3301                         btrfs_tree_unlock(must_clean);
3302                         free_extent_buffer(must_clean);
3303                 }
3304
3305                 if (is_data) {
3306                         ret = btrfs_del_csums(trans, root, bytenr, num_bytes);
3307                         BUG_ON(ret);
3308                 } else {
3309                         invalidate_mapping_pages(info->btree_inode->i_mapping,
3310                              bytenr >> PAGE_CACHE_SHIFT,
3311                              (bytenr + num_bytes - 1) >> PAGE_CACHE_SHIFT);
3312                 }
3313
3314                 ret = update_block_group(trans, root, bytenr, num_bytes, 0,
3315                                          mark_free);
3316                 BUG_ON(ret);
3317         }
3318         btrfs_free_path(path);
3319         return ret;
3320 }
3321
3322 /*
3323  * when we free an extent, it is possible (and likely) that we free the last
3324  * delayed ref for that extent as well.  This searches the delayed ref tree for
3325  * a given extent, and if there are no other delayed refs to be processed, it
3326  * removes it from the tree.
3327  */
3328 static noinline int check_ref_cleanup(struct btrfs_trans_handle *trans,
3329                                       struct btrfs_root *root, u64 bytenr)
3330 {
3331         struct btrfs_delayed_ref_head *head;
3332         struct btrfs_delayed_ref_root *delayed_refs;
3333         struct btrfs_delayed_ref_node *ref;
3334         struct rb_node *node;
3335         int ret;
3336
3337         delayed_refs = &trans->transaction->delayed_refs;
3338         spin_lock(&delayed_refs->lock);
3339         head = btrfs_find_delayed_ref_head(trans, bytenr);
3340         if (!head)
3341                 goto out;
3342
3343         node = rb_prev(&head->node.rb_node);
3344         if (!node)
3345                 goto out;
3346
3347         ref = rb_entry(node, struct btrfs_delayed_ref_node, rb_node);
3348
3349         /* there are still entries for this ref, we can't drop it */
3350         if (ref->bytenr == bytenr)
3351                 goto out;
3352
3353         if (head->extent_op) {
3354                 if (!head->must_insert_reserved)
3355                         goto out;
3356                 kfree(head->extent_op);
3357                 head->extent_op = NULL;
3358         }
3359
3360         /*
3361          * waiting for the lock here would deadlock.  If someone else has it
3362          * locked they are already in the process of dropping it anyway
3363          */
3364         if (!mutex_trylock(&head->mutex))
3365                 goto out;
3366
3367         /*
3368          * at this point we have a head with no other entries.  Go
3369          * ahead and process it.
3370          */
3371         head->node.in_tree = 0;
3372         rb_erase(&head->node.rb_node, &delayed_refs->root);
3373
3374         delayed_refs->num_entries--;
3375
3376         /*
3377          * we don't take a ref on the node because we're removing it from the
3378          * tree, so we just steal the ref the tree was holding.
3379          */
3380         delayed_refs->num_heads--;
3381         if (list_empty(&head->cluster))
3382                 delayed_refs->num_heads_ready--;
3383
3384         list_del_init(&head->cluster);
3385         spin_unlock(&delayed_refs->lock);
3386
3387         ret = run_one_delayed_ref(trans, root->fs_info->tree_root,
3388                                   &head->node, head->extent_op,
3389                                   head->must_insert_reserved);
3390         BUG_ON(ret);
3391         btrfs_put_delayed_ref(&head->node);
3392         return 0;
3393 out:
3394         spin_unlock(&delayed_refs->lock);
3395         return 0;
3396 }
3397
3398 int btrfs_free_extent(struct btrfs_trans_handle *trans,
3399                       struct btrfs_root *root,
3400                       u64 bytenr, u64 num_bytes, u64 parent,
3401                       u64 root_objectid, u64 owner, u64 offset)
3402 {
3403         int ret;
3404
3405         /*
3406          * tree log blocks never actually go into the extent allocation
3407          * tree, just update pinning info and exit early.
3408          */
3409         if (root_objectid == BTRFS_TREE_LOG_OBJECTID) {
3410                 WARN_ON(owner >= BTRFS_FIRST_FREE_OBJECTID);
3411                 /* unlocks the pinned mutex */
3412                 btrfs_update_pinned_extents(root, bytenr, num_bytes, 1);
3413                 update_reserved_extents(root, bytenr, num_bytes, 0);
3414                 ret = 0;
3415         } else if (owner < BTRFS_FIRST_FREE_OBJECTID) {
3416                 ret = btrfs_add_delayed_tree_ref(trans, bytenr, num_bytes,
3417                                         parent, root_objectid, (int)owner,
3418                                         BTRFS_DROP_DELAYED_REF, NULL);
3419                 BUG_ON(ret);
3420                 ret = check_ref_cleanup(trans, root, bytenr);
3421                 BUG_ON(ret);
3422         } else {
3423                 ret = btrfs_add_delayed_data_ref(trans, bytenr, num_bytes,
3424                                         parent, root_objectid, owner,
3425                                         offset, BTRFS_DROP_DELAYED_REF, NULL);
3426                 BUG_ON(ret);
3427         }
3428         return ret;
3429 }
3430
3431 static u64 stripe_align(struct btrfs_root *root, u64 val)
3432 {
3433         u64 mask = ((u64)root->stripesize - 1);
3434         u64 ret = (val + mask) & ~mask;
3435         return ret;
3436 }
3437
3438 /*
3439  * walks the btree of allocated extents and find a hole of a given size.
3440  * The key ins is changed to record the hole:
3441  * ins->objectid == block start
3442  * ins->flags = BTRFS_EXTENT_ITEM_KEY
3443  * ins->offset == number of blocks
3444  * Any available blocks before search_start are skipped.
3445  */
3446 static noinline int find_free_extent(struct btrfs_trans_handle *trans,
3447                                      struct btrfs_root *orig_root,
3448                                      u64 num_bytes, u64 empty_size,
3449                                      u64 search_start, u64 search_end,
3450                                      u64 hint_byte, struct btrfs_key *ins,
3451                                      u64 exclude_start, u64 exclude_nr,
3452                                      int data)
3453 {
3454         int ret = 0;
3455         struct btrfs_root *root = orig_root->fs_info->extent_root;
3456         struct btrfs_free_cluster *last_ptr = NULL;
3457         struct btrfs_block_group_cache *block_group = NULL;
3458         int empty_cluster = 2 * 1024 * 1024;
3459         int allowed_chunk_alloc = 0;
3460         struct btrfs_space_info *space_info;
3461         int last_ptr_loop = 0;
3462         int loop = 0;
3463
3464         WARN_ON(num_bytes < root->sectorsize);
3465         btrfs_set_key_type(ins, BTRFS_EXTENT_ITEM_KEY);
3466         ins->objectid = 0;
3467         ins->offset = 0;
3468
3469         space_info = __find_space_info(root->fs_info, data);
3470
3471         if (orig_root->ref_cows || empty_size)
3472                 allowed_chunk_alloc = 1;
3473
3474         if (data & BTRFS_BLOCK_GROUP_METADATA) {
3475                 last_ptr = &root->fs_info->meta_alloc_cluster;
3476                 if (!btrfs_test_opt(root, SSD))
3477                         empty_cluster = 64 * 1024;
3478         }
3479
3480         if ((data & BTRFS_BLOCK_GROUP_DATA) && btrfs_test_opt(root, SSD)) {
3481                 last_ptr = &root->fs_info->data_alloc_cluster;
3482         }
3483
3484         if (last_ptr) {
3485                 spin_lock(&last_ptr->lock);
3486                 if (last_ptr->block_group)
3487                         hint_byte = last_ptr->window_start;
3488                 spin_unlock(&last_ptr->lock);
3489         }
3490
3491         search_start = max(search_start, first_logical_byte(root, 0));
3492         search_start = max(search_start, hint_byte);
3493
3494         if (!last_ptr) {
3495                 empty_cluster = 0;
3496                 loop = 1;
3497         }
3498
3499         if (search_start == hint_byte) {
3500                 block_group = btrfs_lookup_block_group(root->fs_info,
3501                                                        search_start);
3502                 if (block_group && block_group_bits(block_group, data)) {
3503                         down_read(&space_info->groups_sem);
3504                         if (list_empty(&block_group->list) ||
3505                             block_group->ro) {
3506                                 /*
3507                                  * someone is removing this block group,
3508                                  * we can't jump into the have_block_group
3509                                  * target because our list pointers are not
3510                                  * valid
3511                                  */
3512                                 btrfs_put_block_group(block_group);
3513                                 up_read(&space_info->groups_sem);
3514                         } else
3515                                 goto have_block_group;
3516                 } else if (block_group) {
3517                         btrfs_put_block_group(block_group);
3518                 }
3519         }
3520
3521 search:
3522         down_read(&space_info->groups_sem);
3523         list_for_each_entry(block_group, &space_info->block_groups, list) {
3524                 u64 offset;
3525
3526                 atomic_inc(&block_group->count);
3527                 search_start = block_group->key.objectid;
3528
3529 have_block_group:
3530                 if (unlikely(!block_group->cached)) {
3531                         mutex_lock(&block_group->cache_mutex);
3532                         ret = cache_block_group(root, block_group);
3533                         mutex_unlock(&block_group->cache_mutex);
3534                         if (ret) {
3535                                 btrfs_put_block_group(block_group);
3536                                 break;
3537                         }
3538                 }
3539
3540                 if (unlikely(block_group->ro))
3541                         goto loop;
3542
3543                 if (last_ptr) {
3544                         /*
3545                          * the refill lock keeps out other
3546                          * people trying to start a new cluster
3547                          */
3548                         spin_lock(&last_ptr->refill_lock);
3549                         if (last_ptr->block_group &&
3550                             (last_ptr->block_group->ro ||
3551                             !block_group_bits(last_ptr->block_group, data))) {
3552                                 offset = 0;
3553                                 goto refill_cluster;
3554                         }
3555
3556                         offset = btrfs_alloc_from_cluster(block_group, last_ptr,
3557                                                  num_bytes, search_start);
3558                         if (offset) {
3559                                 /* we have a block, we're done */
3560                                 spin_unlock(&last_ptr->refill_lock);
3561                                 goto checks;
3562                         }
3563
3564                         spin_lock(&last_ptr->lock);
3565                         /*
3566                          * whoops, this cluster doesn't actually point to
3567                          * this block group.  Get a ref on the block
3568                          * group is does point to and try again
3569                          */
3570                         if (!last_ptr_loop && last_ptr->block_group &&
3571                             last_ptr->block_group != block_group) {
3572
3573                                 btrfs_put_block_group(block_group);
3574                                 block_group = last_ptr->block_group;
3575                                 atomic_inc(&block_group->count);
3576                                 spin_unlock(&last_ptr->lock);
3577                                 spin_unlock(&last_ptr->refill_lock);
3578
3579                                 last_ptr_loop = 1;
3580                                 search_start = block_group->key.objectid;
3581                                 /*
3582                                  * we know this block group is properly
3583                                  * in the list because
3584                                  * btrfs_remove_block_group, drops the
3585                                  * cluster before it removes the block
3586                                  * group from the list
3587                                  */
3588                                 goto have_block_group;
3589                         }
3590                         spin_unlock(&last_ptr->lock);
3591 refill_cluster:
3592                         /*
3593                          * this cluster didn't work out, free it and
3594                          * start over
3595                          */
3596                         btrfs_return_cluster_to_free_space(NULL, last_ptr);
3597
3598                         last_ptr_loop = 0;
3599
3600                         /* allocate a cluster in this block group */
3601                         ret = btrfs_find_space_cluster(trans, root,
3602                                                block_group, last_ptr,
3603                                                offset, num_bytes,
3604                                                empty_cluster + empty_size);
3605                         if (ret == 0) {
3606                                 /*
3607                                  * now pull our allocation out of this
3608                                  * cluster
3609                                  */
3610                                 offset = btrfs_alloc_from_cluster(block_group,
3611                                                   last_ptr, num_bytes,
3612                                                   search_start);
3613                                 if (offset) {
3614                                         /* we found one, proceed */
3615                                         spin_unlock(&last_ptr->refill_lock);
3616                                         goto checks;
3617                                 }
3618                         }
3619                         /*
3620                          * at this point we either didn't find a cluster
3621                          * or we weren't able to allocate a block from our
3622                          * cluster.  Free the cluster we've been trying
3623                          * to use, and go to the next block group
3624                          */
3625                         if (loop < 2) {
3626                                 btrfs_return_cluster_to_free_space(NULL,
3627                                                                    last_ptr);
3628                                 spin_unlock(&last_ptr->refill_lock);
3629                                 goto loop;
3630                         }
3631                         spin_unlock(&last_ptr->refill_lock);
3632                 }
3633
3634                 offset = btrfs_find_space_for_alloc(block_group, search_start,
3635                                                     num_bytes, empty_size);
3636                 if (!offset)
3637                         goto loop;
3638 checks:
3639                 search_start = stripe_align(root, offset);
3640
3641                 /* move on to the next group */
3642                 if (search_start + num_bytes >= search_end) {
3643                         btrfs_add_free_space(block_group, offset, num_bytes);
3644                         goto loop;
3645                 }
3646
3647                 /* move on to the next group */
3648                 if (search_start + num_bytes >
3649                     block_group->key.objectid + block_group->key.offset) {
3650                         btrfs_add_free_space(block_group, offset, num_bytes);
3651                         goto loop;
3652                 }
3653
3654                 if (exclude_nr > 0 &&
3655                     (search_start + num_bytes > exclude_start &&
3656                      search_start < exclude_start + exclude_nr)) {
3657                         search_start = exclude_start + exclude_nr;
3658
3659                         btrfs_add_free_space(block_group, offset, num_bytes);
3660                         /*
3661                          * if search_start is still in this block group
3662                          * then we just re-search this block group
3663                          */
3664                         if (search_start >= block_group->key.objectid &&
3665                             search_start < (block_group->key.objectid +
3666                                             block_group->key.offset))
3667                                 goto have_block_group;
3668                         goto loop;
3669                 }
3670
3671                 ins->objectid = search_start;
3672                 ins->offset = num_bytes;
3673
3674                 if (offset < search_start)
3675                         btrfs_add_free_space(block_group, offset,
3676                                              search_start - offset);
3677                 BUG_ON(offset > search_start);
3678
3679                 /* we are all good, lets return */
3680                 break;
3681 loop:
3682                 btrfs_put_block_group(block_group);
3683         }
3684         up_read(&space_info->groups_sem);
3685
3686         /* loop == 0, try to find a clustered alloc in every block group
3687          * loop == 1, try again after forcing a chunk allocation
3688          * loop == 2, set empty_size and empty_cluster to 0 and try again
3689          */
3690         if (!ins->objectid && loop < 3 &&
3691             (empty_size || empty_cluster || allowed_chunk_alloc)) {
3692                 if (loop >= 2) {
3693                         empty_size = 0;
3694                         empty_cluster = 0;
3695                 }
3696
3697                 if (allowed_chunk_alloc) {
3698                         ret = do_chunk_alloc(trans, root, num_bytes +
3699                                              2 * 1024 * 1024, data, 1);
3700                         allowed_chunk_alloc = 0;
3701                 } else {
3702                         space_info->force_alloc = 1;
3703                 }
3704
3705                 if (loop < 3) {
3706                         loop++;
3707                         goto search;
3708                 }
3709                 ret = -ENOSPC;
3710         } else if (!ins->objectid) {
3711                 ret = -ENOSPC;
3712         }
3713
3714         /* we found what we needed */
3715         if (ins->objectid) {
3716                 if (!(data & BTRFS_BLOCK_GROUP_DATA))
3717                         trans->block_group = block_group->key.objectid;
3718
3719                 btrfs_put_block_group(block_group);
3720                 ret = 0;
3721         }
3722
3723         return ret;
3724 }
3725
3726 static void dump_space_info(struct btrfs_space_info *info, u64 bytes)
3727 {
3728         struct btrfs_block_group_cache *cache;
3729
3730         printk(KERN_INFO "space_info has %llu free, is %sfull\n",
3731                (unsigned long long)(info->total_bytes - info->bytes_used -
3732                                     info->bytes_pinned - info->bytes_reserved),
3733                (info->full) ? "" : "not ");
3734         printk(KERN_INFO "space_info total=%llu, pinned=%llu, delalloc=%llu,"
3735                " may_use=%llu, used=%llu\n",
3736                (unsigned long long)info->total_bytes,
3737                (unsigned long long)info->bytes_pinned,
3738                (unsigned long long)info->bytes_delalloc,
3739                (unsigned long long)info->bytes_may_use,
3740                (unsigned long long)info->bytes_used);
3741
3742         down_read(&info->groups_sem);
3743         list_for_each_entry(cache, &info->block_groups, list) {
3744                 spin_lock(&cache->lock);
3745                 printk(KERN_INFO "block group %llu has %llu bytes, %llu used "
3746                        "%llu pinned %llu reserved\n",
3747                        (unsigned long long)cache->key.objectid,
3748                        (unsigned long long)cache->key.offset,
3749                        (unsigned long long)btrfs_block_group_used(&cache->item),
3750                        (unsigned long long)cache->pinned,
3751                        (unsigned long long)cache->reserved);
3752                 btrfs_dump_free_space(cache, bytes);
3753                 spin_unlock(&cache->lock);
3754         }
3755         up_read(&info->groups_sem);
3756 }
3757
3758 static int __btrfs_reserve_extent(struct btrfs_trans_handle *trans,
3759                                   struct btrfs_root *root,
3760                                   u64 num_bytes, u64 min_alloc_size,
3761                                   u64 empty_size, u64 hint_byte,
3762                                   u64 search_end, struct btrfs_key *ins,
3763                                   u64 data)
3764 {
3765         int ret;
3766         u64 search_start = 0;
3767         struct btrfs_fs_info *info = root->fs_info;
3768
3769         data = btrfs_get_alloc_profile(root, data);
3770 again:
3771         /*
3772          * the only place that sets empty_size is btrfs_realloc_node, which
3773          * is not called recursively on allocations
3774          */
3775         if (empty_size || root->ref_cows) {
3776                 if (!(data & BTRFS_BLOCK_GROUP_METADATA)) {
3777                         ret = do_chunk_alloc(trans, root->fs_info->extent_root,
3778                                      2 * 1024 * 1024,
3779                                      BTRFS_BLOCK_GROUP_METADATA |
3780                                      (info->metadata_alloc_profile &
3781                                       info->avail_metadata_alloc_bits), 0);
3782                 }
3783                 ret = do_chunk_alloc(trans, root->fs_info->extent_root,
3784                                      num_bytes + 2 * 1024 * 1024, data, 0);
3785         }
3786
3787         WARN_ON(num_bytes < root->sectorsize);
3788         ret = find_free_extent(trans, root, num_bytes, empty_size,
3789                                search_start, search_end, hint_byte, ins,
3790                                trans->alloc_exclude_start,
3791                                trans->alloc_exclude_nr, data);
3792
3793         if (ret == -ENOSPC && num_bytes > min_alloc_size) {
3794                 num_bytes = num_bytes >> 1;
3795                 num_bytes = num_bytes & ~(root->sectorsize - 1);
3796                 num_bytes = max(num_bytes, min_alloc_size);
3797                 do_chunk_alloc(trans, root->fs_info->extent_root,
3798                                num_bytes, data, 1);
3799                 goto again;
3800         }
3801         if (ret) {
3802                 struct btrfs_space_info *sinfo;
3803
3804                 sinfo = __find_space_info(root->fs_info, data);
3805                 printk(KERN_ERR "btrfs allocation failed flags %llu, "
3806                        "wanted %llu\n", (unsigned long long)data,
3807                        (unsigned long long)num_bytes);
3808                 dump_space_info(sinfo, num_bytes);
3809                 BUG();
3810         }
3811
3812         return ret;
3813 }
3814
3815 int btrfs_free_reserved_extent(struct btrfs_root *root, u64 start, u64 len)
3816 {
3817         struct btrfs_block_group_cache *cache;
3818         int ret = 0;
3819
3820         cache = btrfs_lookup_block_group(root->fs_info, start);
3821         if (!cache) {
3822                 printk(KERN_ERR "Unable to find block group for %llu\n",
3823                        (unsigned long long)start);
3824                 return -ENOSPC;
3825         }
3826
3827         ret = btrfs_discard_extent(root, start, len);
3828
3829         btrfs_add_free_space(cache, start, len);
3830         btrfs_put_block_group(cache);
3831         update_reserved_extents(root, start, len, 0);
3832
3833         return ret;
3834 }
3835
3836 int btrfs_reserve_extent(struct btrfs_trans_handle *trans,
3837                                   struct btrfs_root *root,
3838                                   u64 num_bytes, u64 min_alloc_size,
3839                                   u64 empty_size, u64 hint_byte,
3840                                   u64 search_end, struct btrfs_key *ins,
3841                                   u64 data)
3842 {
3843         int ret;
3844         ret = __btrfs_reserve_extent(trans, root, num_bytes, min_alloc_size,
3845                                      empty_size, hint_byte, search_end, ins,
3846                                      data);
3847         update_reserved_extents(root, ins->objectid, ins->offset, 1);
3848         return ret;
3849 }
3850
3851 static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
3852                                       struct btrfs_root *root,
3853                                       u64 parent, u64 root_objectid,
3854                                       u64 flags, u64 owner, u64 offset,
3855                                       struct btrfs_key *ins, int ref_mod)
3856 {
3857         int ret;
3858         struct btrfs_fs_info *fs_info = root->fs_info;
3859         struct btrfs_extent_item *extent_item;
3860         struct btrfs_extent_inline_ref *iref;
3861         struct btrfs_path *path;
3862         struct extent_buffer *leaf;
3863         int type;
3864         u32 size;
3865
3866         if (parent > 0)
3867                 type = BTRFS_SHARED_DATA_REF_KEY;
3868         else
3869                 type = BTRFS_EXTENT_DATA_REF_KEY;
3870
3871         size = sizeof(*extent_item) + btrfs_extent_inline_ref_size(type);
3872
3873         path = btrfs_alloc_path();
3874         BUG_ON(!path);
3875
3876         path->leave_spinning = 1;
3877         ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path,
3878                                       ins, size);
3879         BUG_ON(ret);
3880
3881         leaf = path->nodes[0];
3882         extent_item = btrfs_item_ptr(leaf, path->slots[0],
3883                                      struct btrfs_extent_item);
3884         btrfs_set_extent_refs(leaf, extent_item, ref_mod);
3885         btrfs_set_extent_generation(leaf, extent_item, trans->transid);
3886         btrfs_set_extent_flags(leaf, extent_item,
3887                                flags | BTRFS_EXTENT_FLAG_DATA);
3888
3889         iref = (struct btrfs_extent_inline_ref *)(extent_item + 1);
3890         btrfs_set_extent_inline_ref_type(leaf, iref, type);
3891         if (parent > 0) {
3892                 struct btrfs_shared_data_ref *ref;
3893                 ref = (struct btrfs_shared_data_ref *)(iref + 1);
3894                 btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
3895                 btrfs_set_shared_data_ref_count(leaf, ref, ref_mod);
3896         } else {
3897                 struct btrfs_extent_data_ref *ref;
3898                 ref = (struct btrfs_extent_data_ref *)(&iref->offset);
3899                 btrfs_set_extent_data_ref_root(leaf, ref, root_objectid);
3900                 btrfs_set_extent_data_ref_objectid(leaf, ref, owner);
3901                 btrfs_set_extent_data_ref_offset(leaf, ref, offset);
3902                 btrfs_set_extent_data_ref_count(leaf, ref, ref_mod);
3903         }
3904
3905         btrfs_mark_buffer_dirty(path->nodes[0]);
3906         btrfs_free_path(path);
3907
3908         ret = update_block_group(trans, root, ins->objectid, ins->offset,
3909                                  1, 0);
3910         if (ret) {
3911                 printk(KERN_ERR "btrfs update block group failed for %llu "
3912                        "%llu\n", (unsigned long long)ins->objectid,
3913                        (unsigned long long)ins->offset);
3914                 BUG();
3915         }
3916         return ret;
3917 }
3918
3919 static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
3920                                      struct btrfs_root *root,
3921                                      u64 parent, u64 root_objectid,
3922                                      u64 flags, struct btrfs_disk_key *key,
3923                                      int level, struct btrfs_key *ins)
3924 {
3925         int ret;
3926         struct btrfs_fs_info *fs_info = root->fs_info;
3927         struct btrfs_extent_item *extent_item;
3928         struct btrfs_tree_block_info *block_info;
3929         struct btrfs_extent_inline_ref *iref;
3930         struct btrfs_path *path;
3931         struct extent_buffer *leaf;
3932         u32 size = sizeof(*extent_item) + sizeof(*block_info) + sizeof(*iref);
3933
3934         path = btrfs_alloc_path();
3935         BUG_ON(!path);
3936
3937         path->leave_spinning = 1;
3938         ret = btrfs_insert_empty_item(trans, fs_info->extent_root, path,
3939                                       ins, size);
3940         BUG_ON(ret);
3941
3942         leaf = path->nodes[0];
3943         extent_item = btrfs_item_ptr(leaf, path->slots[0],
3944                                      struct btrfs_extent_item);
3945         btrfs_set_extent_refs(leaf, extent_item, 1);
3946         btrfs_set_extent_generation(leaf, extent_item, trans->transid);
3947         btrfs_set_extent_flags(leaf, extent_item,
3948                                flags | BTRFS_EXTENT_FLAG_TREE_BLOCK);
3949         block_info = (struct btrfs_tree_block_info *)(extent_item + 1);
3950
3951         btrfs_set_tree_block_key(leaf, block_info, key);
3952         btrfs_set_tree_block_level(leaf, block_info, level);
3953
3954         iref = (struct btrfs_extent_inline_ref *)(block_info + 1);
3955         if (parent > 0) {
3956                 BUG_ON(!(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF));
3957                 btrfs_set_extent_inline_ref_type(leaf, iref,
3958                                                  BTRFS_SHARED_BLOCK_REF_KEY);
3959                 btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
3960         } else {
3961                 btrfs_set_extent_inline_ref_type(leaf, iref,
3962                                                  BTRFS_TREE_BLOCK_REF_KEY);
3963                 btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid);
3964         }
3965
3966         btrfs_mark_buffer_dirty(leaf);
3967         btrfs_free_path(path);
3968
3969         ret = update_block_group(trans, root, ins->objectid, ins->offset,
3970                                  1, 0);
3971         if (ret) {
3972                 printk(KERN_ERR "btrfs update block group failed for %llu "
3973                        "%llu\n", (unsigned long long)ins->objectid,
3974                        (unsigned long long)ins->offset);
3975                 BUG();
3976         }
3977         return ret;
3978 }
3979
3980 int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
3981                                      struct btrfs_root *root,
3982                                      u64 root_objectid, u64 owner,
3983                                      u64 offset, struct btrfs_key *ins)
3984 {
3985         int ret;
3986
3987         BUG_ON(root_objectid == BTRFS_TREE_LOG_OBJECTID);
3988
3989         ret = btrfs_add_delayed_data_ref(trans, ins->objectid, ins->offset,
3990                                          0, root_objectid, owner, offset,
3991                                          BTRFS_ADD_DELAYED_EXTENT, NULL);
3992         return ret;
3993 }
3994
3995 /*
3996  * this is used by the tree logging recovery code.  It records that
3997  * an extent has been allocated and makes sure to clear the free
3998  * space cache bits as well
3999  */
4000 int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans,
4001                                    struct btrfs_root *root,
4002                                    u64 root_objectid, u64 owner, u64 offset,
4003                                    struct btrfs_key *ins)
4004 {
4005         int ret;
4006         struct btrfs_block_group_cache *block_group;
4007
4008         block_group = btrfs_lookup_block_group(root->fs_info, ins->objectid);
4009         mutex_lock(&block_group->cache_mutex);
4010         cache_block_group(root, block_group);
4011         mutex_unlock(&block_group->cache_mutex);
4012
4013         ret = btrfs_remove_free_space(block_group, ins->objectid,
4014                                       ins->offset);
4015         BUG_ON(ret);
4016         btrfs_put_block_group(block_group);
4017         ret = alloc_reserved_file_extent(trans, root, 0, root_objectid,
4018                                          0, owner, offset, ins, 1);
4019         return ret;
4020 }
4021
4022 /*
4023  * finds a free extent and does all the dirty work required for allocation
4024  * returns the key for the extent through ins, and a tree buffer for
4025  * the first block of the extent through buf.
4026  *
4027  * returns 0 if everything worked, non-zero otherwise.
4028  */
4029 static int alloc_tree_block(struct btrfs_trans_handle *trans,
4030                             struct btrfs_root *root,
4031                             u64 num_bytes, u64 parent, u64 root_objectid,
4032                             struct btrfs_disk_key *key, int level,
4033                             u64 empty_size, u64 hint_byte, u64 search_end,
4034                             struct btrfs_key *ins)
4035 {
4036         int ret;
4037         u64 flags = 0;
4038
4039         ret = __btrfs_reserve_extent(trans, root, num_bytes, num_bytes,
4040                                      empty_size, hint_byte, search_end,
4041                                      ins, 0);
4042         BUG_ON(ret);
4043
4044         if (root_objectid == BTRFS_TREE_RELOC_OBJECTID) {
4045                 if (parent == 0)
4046                         parent = ins->objectid;
4047                 flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF;
4048         } else
4049                 BUG_ON(parent > 0);
4050
4051         update_reserved_extents(root, ins->objectid, ins->offset, 1);
4052         if (root_objectid != BTRFS_TREE_LOG_OBJECTID) {
4053                 struct btrfs_delayed_extent_op *extent_op;
4054                 extent_op = kmalloc(sizeof(*extent_op), GFP_NOFS);
4055                 BUG_ON(!extent_op);
4056                 if (key)
4057                         memcpy(&extent_op->key, key, sizeof(extent_op->key));
4058                 else
4059                         memset(&extent_op->key, 0, sizeof(extent_op->key));
4060                 extent_op->flags_to_set = flags;
4061                 extent_op->update_key = 1;
4062                 extent_op->update_flags = 1;
4063                 extent_op->is_data = 0;
4064
4065                 ret = btrfs_add_delayed_tree_ref(trans, ins->objectid,
4066                                         ins->offset, parent, root_objectid,
4067                                         level, BTRFS_ADD_DELAYED_EXTENT,
4068                                         extent_op);
4069                 BUG_ON(ret);
4070         }
4071         return ret;
4072 }
4073
4074 struct extent_buffer *btrfs_init_new_buffer(struct btrfs_trans_handle *trans,
4075                                             struct btrfs_root *root,
4076                                             u64 bytenr, u32 blocksize,
4077                                             int level)
4078 {
4079         struct extent_buffer *buf;
4080
4081         buf = btrfs_find_create_tree_block(root, bytenr, blocksize);
4082         if (!buf)
4083                 return ERR_PTR(-ENOMEM);
4084         btrfs_set_header_generation(buf, trans->transid);
4085         btrfs_set_buffer_lockdep_class(buf, level);
4086         btrfs_tree_lock(buf);
4087         clean_tree_block(trans, root, buf);
4088
4089         btrfs_set_lock_blocking(buf);
4090         btrfs_set_buffer_uptodate(buf);
4091
4092         if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) {
4093                 set_extent_dirty(&root->dirty_log_pages, buf->start,
4094                          buf->start + buf->len - 1, GFP_NOFS);
4095         } else {
4096                 set_extent_dirty(&trans->transaction->dirty_pages, buf->start,
4097                          buf->start + buf->len - 1, GFP_NOFS);
4098         }
4099         trans->blocks_used++;
4100         /* this returns a buffer locked for blocking */
4101         return buf;
4102 }
4103
4104 /*
4105  * helper function to allocate a block for a given tree
4106  * returns the tree buffer or NULL.
4107  */
4108 struct extent_buffer *btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
4109                                         struct btrfs_root *root, u32 blocksize,
4110                                         u64 parent, u64 root_objectid,
4111                                         struct btrfs_disk_key *key, int level,
4112                                         u64 hint, u64 empty_size)
4113 {
4114         struct btrfs_key ins;
4115         int ret;
4116         struct extent_buffer *buf;
4117
4118         ret = alloc_tree_block(trans, root, blocksize, parent, root_objectid,
4119                                key, level, empty_size, hint, (u64)-1, &ins);
4120         if (ret) {
4121                 BUG_ON(ret > 0);
4122                 return ERR_PTR(ret);
4123         }
4124
4125         buf = btrfs_init_new_buffer(trans, root, ins.objectid,
4126                                     blocksize, level);
4127         return buf;
4128 }
4129
4130 #if 0
4131 int btrfs_drop_leaf_ref(struct btrfs_trans_handle *trans,
4132                         struct btrfs_root *root, struct extent_buffer *leaf)
4133 {
4134         u64 disk_bytenr;
4135         u64 num_bytes;
4136         struct btrfs_key key;
4137         struct btrfs_file_extent_item *fi;
4138         u32 nritems;
4139         int i;
4140         int ret;
4141
4142         BUG_ON(!btrfs_is_leaf(leaf));
4143         nritems = btrfs_header_nritems(leaf);
4144
4145         for (i = 0; i < nritems; i++) {
4146                 cond_resched();
4147                 btrfs_item_key_to_cpu(leaf, &key, i);
4148
4149                 /* only extents have references, skip everything else */
4150                 if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
4151                         continue;
4152
4153                 fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item);
4154
4155                 /* inline extents live in the btree, they don't have refs */
4156                 if (btrfs_file_extent_type(leaf, fi) ==
4157                     BTRFS_FILE_EXTENT_INLINE)
4158                         continue;
4159
4160                 disk_bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
4161
4162                 /* holes don't have refs */
4163                 if (disk_bytenr == 0)
4164                         continue;
4165
4166                 num_bytes = btrfs_file_extent_disk_num_bytes(leaf, fi);
4167                 ret = btrfs_free_extent(trans, root, disk_bytenr, num_bytes,
4168                                         leaf->start, 0, key.objectid, 0);
4169                 BUG_ON(ret);
4170         }
4171         return 0;
4172 }
4173
4174 static noinline int cache_drop_leaf_ref(struct btrfs_trans_handle *trans,
4175                                         struct btrfs_root *root,
4176                                         struct btrfs_leaf_ref *ref)
4177 {
4178         int i;
4179         int ret;
4180         struct btrfs_extent_info *info;
4181         struct refsort *sorted;
4182
4183         if (ref->nritems == 0)
4184                 return 0;
4185
4186         sorted = kmalloc(sizeof(*sorted) * ref->nritems, GFP_NOFS);
4187         for (i = 0; i < ref->nritems; i++) {
4188                 sorted[i].bytenr = ref->extents[i].bytenr;
4189                 sorted[i].slot = i;
4190         }
4191         sort(sorted, ref->nritems, sizeof(struct refsort), refsort_cmp, NULL);
4192
4193         /*
4194          * the items in the ref were sorted when the ref was inserted
4195          * into the ref cache, so this is already in order
4196          */
4197         for (i = 0; i < ref->nritems; i++) {
4198                 info = ref->extents + sorted[i].slot;
4199                 ret = btrfs_free_extent(trans, root, info->bytenr,
4200                                           info->num_bytes, ref->bytenr,
4201                                           ref->owner, ref->generation,
4202                                           info->objectid, 0);
4203
4204                 atomic_inc(&root->fs_info->throttle_gen);
4205                 wake_up(&root->fs_info->transaction_throttle);
4206                 cond_resched();
4207
4208                 BUG_ON(ret);
4209                 info++;
4210         }
4211
4212         kfree(sorted);
4213         return 0;
4214 }
4215
4216
4217 static int drop_snap_lookup_refcount(struct btrfs_trans_handle *trans,
4218                                      struct btrfs_root *root, u64 start,
4219                                      u64 len, u32 *refs)
4220 {
4221         int ret;
4222
4223         ret = btrfs_lookup_extent_refs(trans, root, start, len, refs);
4224         BUG_ON(ret);
4225
4226 #if 0 /* some debugging code in case we see problems here */
4227         /* if the refs count is one, it won't get increased again.  But
4228          * if the ref count is > 1, someone may be decreasing it at
4229          * the same time we are.
4230          */
4231         if (*refs != 1) {
4232                 struct extent_buffer *eb = NULL;
4233                 eb = btrfs_find_create_tree_block(root, start, len);
4234                 if (eb)
4235                         btrfs_tree_lock(eb);
4236
4237                 mutex_lock(&root->fs_info->alloc_mutex);
4238                 ret = lookup_extent_ref(NULL, root, start, len, refs);
4239                 BUG_ON(ret);
4240                 mutex_unlock(&root->fs_info->alloc_mutex);
4241
4242                 if (eb) {
4243                         btrfs_tree_unlock(eb);
4244                         free_extent_buffer(eb);
4245                 }
4246                 if (*refs == 1) {
4247                         printk(KERN_ERR "btrfs block %llu went down to one "
4248                                "during drop_snap\n", (unsigned long long)start);
4249                 }
4250
4251         }
4252 #endif
4253
4254         cond_resched();
4255         return ret;
4256 }
4257
4258
4259 /*
4260  * this is used while deleting old snapshots, and it drops the refs
4261  * on a whole subtree starting from a level 1 node.
4262  *
4263  * The idea is to sort all the leaf pointers, and then drop the
4264  * ref on all the leaves in order.  Most of the time the leaves
4265  * will have ref cache entries, so no leaf IOs will be required to
4266  * find the extents they have references on.
4267  *
4268  * For each leaf, any references it has are also dropped in order
4269  *
4270  * This ends up dropping the references in something close to optimal
4271  * order for reading and modifying the extent allocation tree.
4272  */
4273 static noinline int drop_level_one_refs(struct btrfs_trans_handle *trans,
4274                                         struct btrfs_root *root,
4275                                         struct btrfs_path *path)
4276 {
4277         u64 bytenr;
4278         u64 root_owner;
4279         u64 root_gen;
4280         struct extent_buffer *eb = path->nodes[1];
4281         struct extent_buffer *leaf;
4282         struct btrfs_leaf_ref *ref;
4283         struct refsort *sorted = NULL;
4284         int nritems = btrfs_header_nritems(eb);
4285         int ret;
4286         int i;
4287         int refi = 0;
4288         int slot = path->slots[1];
4289         u32 blocksize = btrfs_level_size(root, 0);
4290         u32 refs;
4291
4292         if (nritems == 0)
4293                 goto out;
4294
4295         root_owner = btrfs_header_owner(eb);
4296         root_gen = btrfs_header_generation(eb);
4297         sorted = kmalloc(sizeof(*sorted) * nritems, GFP_NOFS);
4298
4299         /*
4300          * step one, sort all the leaf pointers so we don't scribble
4301          * randomly into the extent allocation tree
4302          */
4303         for (i = slot; i < nritems; i++) {
4304                 sorted[refi].bytenr = btrfs_node_blockptr(eb, i);
4305                 sorted[refi].slot = i;
4306                 refi++;
4307         }
4308
4309         /*
4310          * nritems won't be zero, but if we're picking up drop_snapshot
4311          * after a crash, slot might be > 0, so double check things
4312          * just in case.
4313          */
4314         if (refi == 0)
4315                 goto out;
4316
4317         sort(sorted, refi, sizeof(struct refsort), refsort_cmp, NULL);
4318
4319         /*
4320          * the first loop frees everything the leaves point to
4321          */
4322         for (i = 0; i < refi; i++) {
4323                 u64 ptr_gen;
4324
4325                 bytenr = sorted[i].bytenr;
4326
4327                 /*
4328                  * check the reference count on this leaf.  If it is > 1
4329                  * we just decrement it below and don't update any
4330                  * of the refs the leaf points to.
4331                  */
4332                 ret = drop_snap_lookup_refcount(trans, root, bytenr,
4333                                                 blocksize, &refs);
4334                 BUG_ON(ret);
4335                 if (refs != 1)
4336                         continue;
4337
4338                 ptr_gen = btrfs_node_ptr_generation(eb, sorted[i].slot);
4339
4340                 /*
4341                  * the leaf only had one reference, which means the
4342                  * only thing pointing to this leaf is the snapshot
4343                  * we're deleting.  It isn't possible for the reference
4344                  * count to increase again later
4345                  *
4346                  * The reference cache is checked for the leaf,
4347                  * and if found we'll be able to drop any refs held by
4348                  * the leaf without needing to read it in.
4349                  */
4350                 ref = btrfs_lookup_leaf_ref(root, bytenr);
4351                 if (ref && ref->generation != ptr_gen) {
4352                         btrfs_free_leaf_ref(root, ref);
4353                         ref = NULL;
4354                 }
4355                 if (ref) {
4356                         ret = cache_drop_leaf_ref(trans, root, ref);
4357                         BUG_ON(ret);
4358                         btrfs_remove_leaf_ref(root, ref);
4359                         btrfs_free_leaf_ref(root, ref);
4360                 } else {
4361                         /*
4362                          * the leaf wasn't in the reference cache, so
4363                          * we have to read it.
4364                          */
4365                         leaf = read_tree_block(root, bytenr, blocksize,
4366                                                ptr_gen);
4367                         ret = btrfs_drop_leaf_ref(trans, root, leaf);
4368                         BUG_ON(ret);
4369                         free_extent_buffer(leaf);
4370                 }
4371                 atomic_inc(&root->fs_info->throttle_gen);
4372                 wake_up(&root->fs_info->transaction_throttle);
4373                 cond_resched();
4374         }
4375
4376         /*
4377          * run through the loop again to free the refs on the leaves.
4378          * This is faster than doing it in the loop above because
4379          * the leaves are likely to be clustered together.  We end up
4380          * working in nice chunks on the extent allocation tree.
4381          */
4382         for (i = 0; i < refi; i++) {
4383                 bytenr = sorted[i].bytenr;
4384                 ret = btrfs_free_extent(trans, root, bytenr,
4385                                         blocksize, eb->start,
4386                                         root_owner, root_gen, 0, 1);
4387                 BUG_ON(ret);
4388
4389                 atomic_inc(&root->fs_info->throttle_gen);
4390                 wake_up(&root->fs_info->transaction_throttle);
4391                 cond_resched();
4392         }
4393 out:
4394         kfree(sorted);
4395
4396         /*
4397          * update the path to show we've processed the entire level 1
4398          * node.  This will get saved into the root's drop_snapshot_progress
4399          * field so these drops are not repeated again if this transaction
4400          * commits.
4401          */
4402         path->slots[1] = nritems;
4403         return 0;
4404 }
4405
4406 /*
4407  * helper function for drop_snapshot, this walks down the tree dropping ref
4408  * counts as it goes.
4409  */
4410 static noinline int walk_down_tree(struct btrfs_trans_handle *trans,
4411                                    struct btrfs_root *root,
4412                                    struct btrfs_path *path, int *level)
4413 {
4414         u64 root_owner;
4415         u64 root_gen;
4416         u64 bytenr;
4417         u64 ptr_gen;
4418         struct extent_buffer *next;
4419         struct extent_buffer *cur;
4420         struct extent_buffer *parent;
4421         u32 blocksize;
4422         int ret;
4423         u32 refs;
4424
4425         WARN_ON(*level < 0);
4426         WARN_ON(*level >= BTRFS_MAX_LEVEL);
4427         ret = drop_snap_lookup_refcount(trans, root, path->nodes[*level]->start,
4428                                 path->nodes[*level]->len, &refs);
4429         BUG_ON(ret);
4430         if (refs > 1)
4431                 goto out;
4432
4433         /*
4434          * walk down to the last node level and free all the leaves
4435          */
4436         while (*level >= 0) {
4437                 WARN_ON(*level < 0);
4438                 WARN_ON(*level >= BTRFS_MAX_LEVEL);
4439                 cur = path->nodes[*level];
4440
4441                 if (btrfs_header_level(cur) != *level)
4442                         WARN_ON(1);
4443
4444                 if (path->slots[*level] >=
4445                     btrfs_header_nritems(cur))
4446                         break;
4447
4448                 /* the new code goes down to level 1 and does all the
4449                  * leaves pointed to that node in bulk.  So, this check
4450                  * for level 0 will always be false.
4451                  *
4452                  * But, the disk format allows the drop_snapshot_progress
4453                  * field in the root to leave things in a state where
4454                  * a leaf will need cleaning up here.  If someone crashes
4455                  * with the old code and then boots with the new code,
4456                  * we might find a leaf here.
4457                  */
4458                 if (*level == 0) {
4459                         ret = btrfs_drop_leaf_ref(trans, root, cur);
4460                         BUG_ON(ret);
4461                         break;
4462                 }
4463
4464                 /*
4465                  * once we get to level one, process the whole node
4466                  * at once, including everything below it.
4467                  */
4468                 if (*level == 1) {
4469                         ret = drop_level_one_refs(trans, root, path);
4470                         BUG_ON(ret);
4471                         break;
4472                 }
4473
4474                 bytenr = btrfs_node_blockptr(cur, path->slots[*level]);
4475                 ptr_gen = btrfs_node_ptr_generation(cur, path->slots[*level]);
4476                 blocksize = btrfs_level_size(root, *level - 1);
4477
4478                 ret = drop_snap_lookup_refcount(trans, root, bytenr,
4479                                                 blocksize, &refs);
4480                 BUG_ON(ret);
4481
4482                 /*
4483                  * if there is more than one reference, we don't need
4484                  * to read that node to drop any references it has.  We
4485                  * just drop the ref we hold on that node and move on to the
4486                  * next slot in this level.
4487                  */
4488                 if (refs != 1) {
4489                         parent = path->nodes[*level];
4490                         root_owner = btrfs_header_owner(parent);
4491                         root_gen = btrfs_header_generation(parent);
4492                         path->slots[*level]++;
4493
4494                         ret = btrfs_free_extent(trans, root, bytenr,
4495                                                 blocksize, parent->start,
4496                                                 root_owner, root_gen,
4497                                                 *level - 1, 1);
4498                         BUG_ON(ret);
4499
4500                         atomic_inc(&root->fs_info->throttle_gen);
4501                         wake_up(&root->fs_info->transaction_throttle);
4502                         cond_resched();
4503
4504                         continue;
4505                 }
4506
4507                 /*
4508                  * we need to keep freeing things in the next level down.
4509                  * read the block and loop around to process it
4510                  */
4511                 next = read_tree_block(root, bytenr, blocksize, ptr_gen);
4512                 WARN_ON(*level <= 0);
4513                 if (path->nodes[*level-1])
4514                         free_extent_buffer(path->nodes[*level-1]);
4515                 path->nodes[*level-1] = next;
4516                 *level = btrfs_header_level(next);
4517                 path->slots[*level] = 0;
4518                 cond_resched();
4519         }
4520 out:
4521         WARN_ON(*level < 0);
4522         WARN_ON(*level >= BTRFS_MAX_LEVEL);
4523
4524         if (path->nodes[*level] == root->node) {
4525                 parent = path->nodes[*level];
4526                 bytenr = path->nodes[*level]->start;
4527         } else {
4528                 parent = path->nodes[*level + 1];
4529                 bytenr = btrfs_node_blockptr(parent, path->slots[*level + 1]);
4530         }
4531
4532         blocksize = btrfs_level_size(root, *level);
4533         root_owner = btrfs_header_owner(parent);
4534         root_gen = btrfs_header_generation(parent);
4535
4536         /*
4537          * cleanup and free the reference on the last node
4538          * we processed
4539          */
4540         ret = btrfs_free_extent(trans, root, bytenr, blocksize,
4541                                   parent->start, root_owner, root_gen,
4542                                   *level, 1);
4543         free_extent_buffer(path->nodes[*level]);
4544         path->nodes[*level] = NULL;
4545
4546         *level += 1;
4547         BUG_ON(ret);
4548
4549         cond_resched();
4550         return 0;
4551 }
4552 #endif
4553
4554 struct walk_control {
4555         u64 refs[BTRFS_MAX_LEVEL];
4556         u64 flags[BTRFS_MAX_LEVEL];
4557         struct btrfs_key update_progress;
4558         int stage;
4559         int level;
4560         int shared_level;
4561         int update_ref;
4562         int keep_locks;
4563 };
4564
4565 #define DROP_REFERENCE  1
4566 #define UPDATE_BACKREF  2
4567
4568 /*
4569  * hepler to process tree block while walking down the tree.
4570  *
4571  * when wc->stage == DROP_REFERENCE, this function checks
4572  * reference count of the block. if the block is shared and
4573  * we need update back refs for the subtree rooted at the
4574  * block, this function changes wc->stage to UPDATE_BACKREF
4575  *
4576  * when wc->stage == UPDATE_BACKREF, this function updates
4577  * back refs for pointers in the block.
4578  *
4579  * NOTE: return value 1 means we should stop walking down.
4580  */
4581 static noinline int walk_down_proc(struct btrfs_trans_handle *trans,
4582                                    struct btrfs_root *root,
4583                                    struct btrfs_path *path,
4584                                    struct walk_control *wc)
4585 {
4586         int level = wc->level;
4587         struct extent_buffer *eb = path->nodes[level];
4588         struct btrfs_key key;
4589         u64 flag = BTRFS_BLOCK_FLAG_FULL_BACKREF;
4590         int ret;
4591
4592         if (wc->stage == UPDATE_BACKREF &&
4593             btrfs_header_owner(eb) != root->root_key.objectid)
4594                 return 1;
4595
4596         /*
4597          * when reference count of tree block is 1, it won't increase
4598          * again. once full backref flag is set, we never clear it.
4599          */
4600         if ((wc->stage == DROP_REFERENCE && wc->refs[level] != 1) ||
4601             (wc->stage == UPDATE_BACKREF && !(wc->flags[level] & flag))) {
4602                 BUG_ON(!path->locks[level]);
4603                 ret = btrfs_lookup_extent_info(trans, root,
4604                                                eb->start, eb->len,
4605                                                &wc->refs[level],
4606                                                &wc->flags[level]);
4607                 BUG_ON(ret);
4608                 BUG_ON(wc->refs[level] == 0);
4609         }
4610
4611         if (wc->stage == DROP_REFERENCE &&
4612             wc->update_ref && wc->refs[level] > 1) {
4613                 BUG_ON(eb == root->node);
4614                 BUG_ON(path->slots[level] > 0);
4615                 if (level == 0)
4616                         btrfs_item_key_to_cpu(eb, &key, path->slots[level]);
4617                 else
4618                         btrfs_node_key_to_cpu(eb, &key, path->slots[level]);
4619                 if (btrfs_header_owner(eb) == root->root_key.objectid &&
4620                     btrfs_comp_cpu_keys(&key, &wc->update_progress) >= 0) {
4621                         wc->stage = UPDATE_BACKREF;
4622                         wc->shared_level = level;
4623                 }
4624         }
4625
4626         if (wc->stage == DROP_REFERENCE) {
4627                 if (wc->refs[level] > 1)
4628                         return 1;
4629
4630                 if (path->locks[level] && !wc->keep_locks) {
4631                         btrfs_tree_unlock(eb);
4632                         path->locks[level] = 0;
4633                 }
4634                 return 0;
4635         }
4636
4637         /* wc->stage == UPDATE_BACKREF */
4638         if (!(wc->flags[level] & flag)) {
4639                 BUG_ON(!path->locks[level]);
4640                 ret = btrfs_inc_ref(trans, root, eb, 1);
4641                 BUG_ON(ret);
4642                 ret = btrfs_dec_ref(trans, root, eb, 0);
4643                 BUG_ON(ret);
4644                 ret = btrfs_set_disk_extent_flags(trans, root, eb->start,
4645                                                   eb->len, flag, 0);
4646                 BUG_ON(ret);
4647                 wc->flags[level] |= flag;
4648         }
4649
4650         /*
4651          * the block is shared by multiple trees, so it's not good to
4652          * keep the tree lock
4653          */
4654         if (path->locks[level] && level > 0) {
4655                 btrfs_tree_unlock(eb);
4656                 path->locks[level] = 0;
4657         }
4658         return 0;
4659 }
4660
4661 /*
4662  * hepler to process tree block while walking up the tree.
4663  *
4664  * when wc->stage == DROP_REFERENCE, this function drops
4665  * reference count on the block.
4666  *
4667  * when wc->stage == UPDATE_BACKREF, this function changes
4668  * wc->stage back to DROP_REFERENCE if we changed wc->stage
4669  * to UPDATE_BACKREF previously while processing the block.
4670  *
4671  * NOTE: return value 1 means we should stop walking up.
4672  */
4673 static noinline int walk_up_proc(struct btrfs_trans_handle *trans,
4674                                  struct btrfs_root *root,
4675                                  struct btrfs_path *path,
4676                                  struct walk_control *wc)
4677 {
4678         int ret = 0;
4679         int level = wc->level;
4680         struct extent_buffer *eb = path->nodes[level];
4681         u64 parent = 0;
4682
4683         if (wc->stage == UPDATE_BACKREF) {
4684                 BUG_ON(wc->shared_level < level);
4685                 if (level < wc->shared_level)
4686                         goto out;
4687
4688                 BUG_ON(wc->refs[level] <= 1);
4689                 ret = find_next_key(path, level + 1, &wc->update_progress);
4690                 if (ret > 0)
4691                         wc->update_ref = 0;
4692
4693                 wc->stage = DROP_REFERENCE;
4694                 wc->shared_level = -1;
4695                 path->slots[level] = 0;
4696
4697                 /*
4698                  * check reference count again if the block isn't locked.
4699                  * we should start walking down the tree again if reference
4700                  * count is one.
4701                  */
4702                 if (!path->locks[level]) {
4703                         BUG_ON(level == 0);
4704                         btrfs_tree_lock(eb);
4705                         btrfs_set_lock_blocking(eb);
4706                         path->locks[level] = 1;
4707
4708                         ret = btrfs_lookup_extent_info(trans, root,
4709                                                        eb->start, eb->len,
4710                                                        &wc->refs[level],
4711                                                        &wc->flags[level]);
4712                         BUG_ON(ret);
4713                         BUG_ON(wc->refs[level] == 0);
4714                         if (wc->refs[level] == 1) {
4715                                 btrfs_tree_unlock(eb);
4716                                 path->locks[level] = 0;
4717                                 return 1;
4718                         }
4719                 } else {
4720                         BUG_ON(level != 0);
4721                 }
4722         }
4723
4724         /* wc->stage == DROP_REFERENCE */
4725         BUG_ON(wc->refs[level] > 1 && !path->locks[level]);
4726
4727         if (wc->refs[level] == 1) {
4728                 if (level == 0) {
4729                         if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
4730                                 ret = btrfs_dec_ref(trans, root, eb, 1);
4731                         else
4732                                 ret = btrfs_dec_ref(trans, root, eb, 0);
4733                         BUG_ON(ret);
4734                 }
4735                 /* make block locked assertion in clean_tree_block happy */
4736                 if (!path->locks[level] &&
4737                     btrfs_header_generation(eb) == trans->transid) {
4738                         btrfs_tree_lock(eb);
4739                         btrfs_set_lock_blocking(eb);
4740                         path->locks[level] = 1;
4741                 }
4742                 clean_tree_block(trans, root, eb);
4743         }
4744
4745         if (eb == root->node) {
4746                 if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
4747                         parent = eb->start;
4748                 else
4749                         BUG_ON(root->root_key.objectid !=
4750                                btrfs_header_owner(eb));
4751         } else {
4752                 if (wc->flags[level + 1] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
4753                         parent = path->nodes[level + 1]->start;
4754                 else
4755                         BUG_ON(root->root_key.objectid !=
4756                                btrfs_header_owner(path->nodes[level + 1]));
4757         }
4758
4759         ret = btrfs_free_extent(trans, root, eb->start, eb->len, parent,
4760                                 root->root_key.objectid, level, 0);
4761         BUG_ON(ret);
4762 out:
4763         wc->refs[level] = 0;
4764         wc->flags[level] = 0;
4765         return ret;
4766 }
4767
4768 static noinline int walk_down_tree(struct btrfs_trans_handle *trans,
4769                                    struct btrfs_root *root,
4770                                    struct btrfs_path *path,
4771                                    struct walk_control *wc)
4772 {
4773         struct extent_buffer *next;
4774         struct extent_buffer *cur;
4775         u64 bytenr;
4776         u64 ptr_gen;
4777         u32 blocksize;
4778         int level = wc->level;
4779         int ret;
4780
4781         while (level >= 0) {
4782                 cur = path->nodes[level];
4783                 BUG_ON(path->slots[level] >= btrfs_header_nritems(cur));
4784
4785                 ret = walk_down_proc(trans, root, path, wc);
4786                 if (ret > 0)
4787                         break;
4788
4789                 if (level == 0)
4790                         break;
4791
4792                 bytenr = btrfs_node_blockptr(cur, path->slots[level]);
4793                 blocksize = btrfs_level_size(root, level - 1);
4794                 ptr_gen = btrfs_node_ptr_generation(cur, path->slots[level]);
4795
4796                 next = read_tree_block(root, bytenr, blocksize, ptr_gen);
4797                 btrfs_tree_lock(next);
4798                 btrfs_set_lock_blocking(next);
4799
4800                 level--;
4801                 BUG_ON(level != btrfs_header_level(next));
4802                 path->nodes[level] = next;
4803                 path->slots[level] = 0;
4804                 path->locks[level] = 1;
4805                 wc->level = level;
4806         }
4807         return 0;
4808 }
4809
4810 static noinline int walk_up_tree(struct btrfs_trans_handle *trans,
4811                                  struct btrfs_root *root,
4812                                  struct btrfs_path *path,
4813                                  struct walk_control *wc, int max_level)
4814 {
4815         int level = wc->level;
4816         int ret;
4817
4818         path->slots[level] = btrfs_header_nritems(path->nodes[level]);
4819         while (level < max_level && path->nodes[level]) {
4820                 wc->level = level;
4821                 if (path->slots[level] + 1 <
4822                     btrfs_header_nritems(path->nodes[level])) {
4823                         path->slots[level]++;
4824                         return 0;
4825                 } else {
4826                         ret = walk_up_proc(trans, root, path, wc);
4827                         if (ret > 0)
4828                                 return 0;
4829
4830                         if (path->locks[level]) {
4831                                 btrfs_tree_unlock(path->nodes[level]);
4832                                 path->locks[level] = 0;
4833                         }
4834                         free_extent_buffer(path->nodes[level]);
4835                         path->nodes[level] = NULL;
4836                         level++;
4837                 }
4838         }
4839         return 1;
4840 }
4841
4842 /*
4843  * drop a subvolume tree.
4844  *
4845  * this function traverses the tree freeing any blocks that only
4846  * referenced by the tree.
4847  *
4848  * when a shared tree block is found. this function decreases its
4849  * reference count by one. if update_ref is true, this function
4850  * also make sure backrefs for the shared block and all lower level
4851  * blocks are properly updated.
4852  */
4853 int btrfs_drop_snapshot(struct btrfs_root *root, int update_ref)
4854 {
4855         struct btrfs_path *path;
4856         struct btrfs_trans_handle *trans;
4857         struct btrfs_root *tree_root = root->fs_info->tree_root;
4858         struct btrfs_root_item *root_item = &root->root_item;
4859         struct walk_control *wc;
4860         struct btrfs_key key;
4861         int err = 0;
4862         int ret;
4863         int level;
4864
4865         path = btrfs_alloc_path();
4866         BUG_ON(!path);
4867
4868         wc = kzalloc(sizeof(*wc), GFP_NOFS);
4869         BUG_ON(!wc);
4870
4871         trans = btrfs_start_transaction(tree_root, 1);
4872
4873         if (btrfs_disk_key_objectid(&root_item->drop_progress) == 0) {
4874                 level = btrfs_header_level(root->node);
4875                 path->nodes[level] = btrfs_lock_root_node(root);
4876                 btrfs_set_lock_blocking(path->nodes[level]);
4877                 path->slots[level] = 0;
4878                 path->locks[level] = 1;
4879                 memset(&wc->update_progress, 0,
4880                        sizeof(wc->update_progress));
4881         } else {
4882                 btrfs_disk_key_to_cpu(&key, &root_item->drop_progress);
4883                 memcpy(&wc->update_progress, &key,
4884                        sizeof(wc->update_progress));
4885
4886                 level = root_item->drop_level;
4887                 BUG_ON(level == 0);
4888                 path->lowest_level = level;
4889                 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
4890                 path->lowest_level = 0;
4891                 if (ret < 0) {
4892                         err = ret;
4893                         goto out;
4894                 }
4895                 btrfs_node_key_to_cpu(path->nodes[level], &key,
4896                                       path->slots[level]);
4897                 WARN_ON(memcmp(&key, &wc->update_progress, sizeof(key)));
4898
4899                 /*
4900                  * unlock our path, this is safe because only this
4901                  * function is allowed to delete this snapshot
4902                  */
4903                 btrfs_unlock_up_safe(path, 0);
4904
4905                 level = btrfs_header_level(root->node);
4906                 while (1) {
4907                         btrfs_tree_lock(path->nodes[level]);
4908                         btrfs_set_lock_blocking(path->nodes[level]);
4909
4910                         ret = btrfs_lookup_extent_info(trans, root,
4911                                                 path->nodes[level]->start,
4912                                                 path->nodes[level]->len,
4913                                                 &wc->refs[level],
4914                                                 &wc->flags[level]);
4915                         BUG_ON(ret);
4916                         BUG_ON(wc->refs[level] == 0);
4917
4918                         if (level == root_item->drop_level)
4919                                 break;
4920
4921                         btrfs_tree_unlock(path->nodes[level]);
4922                         WARN_ON(wc->refs[level] != 1);
4923                         level--;
4924                 }
4925         }
4926
4927         wc->level = level;
4928         wc->shared_level = -1;
4929         wc->stage = DROP_REFERENCE;
4930         wc->update_ref = update_ref;
4931         wc->keep_locks = 0;
4932
4933         while (1) {
4934                 ret = walk_down_tree(trans, root, path, wc);
4935                 if (ret < 0) {
4936                         err = ret;
4937                         break;
4938                 }
4939
4940                 ret = walk_up_tree(trans, root, path, wc, BTRFS_MAX_LEVEL);
4941                 if (ret < 0) {
4942                         err = ret;
4943                         break;
4944                 }
4945
4946                 if (ret > 0) {
4947                         BUG_ON(wc->stage != DROP_REFERENCE);
4948                         break;
4949                 }
4950
4951                 if (wc->stage == DROP_REFERENCE) {
4952                         level = wc->level;
4953                         btrfs_node_key(path->nodes[level],
4954                                        &root_item->drop_progress,
4955                                        path->slots[level]);
4956                         root_item->drop_level = level;
4957                 }
4958
4959                 BUG_ON(wc->level == 0);
4960                 if (trans->transaction->in_commit ||
4961                     trans->transaction->delayed_refs.flushing) {
4962                         ret = btrfs_update_root(trans, tree_root,
4963                                                 &root->root_key,
4964                                                 root_item);
4965                         BUG_ON(ret);
4966
4967                         btrfs_end_transaction(trans, tree_root);
4968                         trans = btrfs_start_transaction(tree_root, 1);
4969                 } else {
4970                         unsigned long update;
4971                         update = trans->delayed_ref_updates;
4972                         trans->delayed_ref_updates = 0;
4973                         if (update)
4974                                 btrfs_run_delayed_refs(trans, tree_root,
4975                                                        update);
4976                 }
4977         }
4978         btrfs_release_path(root, path);
4979         BUG_ON(err);
4980
4981         ret = btrfs_del_root(trans, tree_root, &root->root_key);
4982         BUG_ON(ret);
4983
4984         free_extent_buffer(root->node);
4985         free_extent_buffer(root->commit_root);
4986         kfree(root);
4987 out:
4988         btrfs_end_transaction(trans, tree_root);
4989         kfree(wc);
4990         btrfs_free_path(path);
4991         return err;
4992 }
4993
4994 /*
4995  * drop subtree rooted at tree block 'node'.
4996  *
4997  * NOTE: this function will unlock and release tree block 'node'
4998  */
4999 int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
5000                         struct btrfs_root *root,
5001                         struct extent_buffer *node,
5002                         struct extent_buffer *parent)
5003 {
5004         struct btrfs_path *path;
5005         struct walk_control *wc;
5006         int level;
5007         int parent_level;
5008         int ret = 0;
5009         int wret;
5010
5011         BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID);
5012
5013         path = btrfs_alloc_path();
5014         BUG_ON(!path);
5015
5016         wc = kzalloc(sizeof(*wc), GFP_NOFS);
5017         BUG_ON(!wc);
5018
5019         btrfs_assert_tree_locked(parent);
5020         parent_level = btrfs_header_level(parent);
5021         extent_buffer_get(parent);
5022         path->nodes[parent_level] = parent;
5023         path->slots[parent_level] = btrfs_header_nritems(parent);
5024
5025         btrfs_assert_tree_locked(node);
5026         level = btrfs_header_level(node);
5027         path->nodes[level] = node;
5028         path->slots[level] = 0;
5029         path->locks[level] = 1;
5030
5031         wc->refs[parent_level] = 1;
5032         wc->flags[parent_level] = BTRFS_BLOCK_FLAG_FULL_BACKREF;
5033         wc->level = level;
5034         wc->shared_level = -1;
5035         wc->stage = DROP_REFERENCE;
5036         wc->update_ref = 0;
5037         wc->keep_locks = 1;
5038
5039         while (1) {
5040                 wret = walk_down_tree(trans, root, path, wc);
5041                 if (wret < 0) {
5042                         ret = wret;
5043                         break;
5044                 }
5045
5046                 wret = walk_up_tree(trans, root, path, wc, parent_level);
5047                 if (wret < 0)
5048                         ret = wret;
5049                 if (wret != 0)
5050                         break;
5051         }
5052
5053         kfree(wc);
5054         btrfs_free_path(path);
5055         return ret;
5056 }
5057
5058 #if 0
5059 static unsigned long calc_ra(unsigned long start, unsigned long last,
5060                              unsigned long nr)
5061 {
5062         return min(last, start + nr - 1);
5063 }
5064
5065 static noinline int relocate_inode_pages(struct inode *inode, u64 start,
5066                                          u64 len)
5067 {
5068         u64 page_start;
5069         u64 page_end;
5070         unsigned long first_index;
5071         unsigned long last_index;
5072         unsigned long i;
5073         struct page *page;
5074         struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
5075         struct file_ra_state *ra;
5076         struct btrfs_ordered_extent *ordered;
5077         unsigned int total_read = 0;
5078         unsigned int total_dirty = 0;
5079         int ret = 0;
5080
5081         ra = kzalloc(sizeof(*ra), GFP_NOFS);
5082
5083         mutex_lock(&inode->i_mutex);
5084         first_index = start >> PAGE_CACHE_SHIFT;
5085         last_index = (start + len - 1) >> PAGE_CACHE_SHIFT;
5086
5087         /* make sure the dirty trick played by the caller work */
5088         ret = invalidate_inode_pages2_range(inode->i_mapping,
5089                                             first_index, last_index);
5090         if (ret)
5091                 goto out_unlock;
5092
5093         file_ra_state_init(ra, inode->i_mapping);
5094
5095         for (i = first_index ; i <= last_index; i++) {
5096                 if (total_read % ra->ra_pages == 0) {
5097                         btrfs_force_ra(inode->i_mapping, ra, NULL, i,
5098                                        calc_ra(i, last_index, ra->ra_pages));
5099                 }
5100                 total_read++;
5101 again:
5102                 if (((u64)i << PAGE_CACHE_SHIFT) > i_size_read(inode))
5103                         BUG_ON(1);
5104                 page = grab_cache_page(inode->i_mapping, i);
5105                 if (!page) {
5106                         ret = -ENOMEM;
5107                         goto out_unlock;
5108                 }
5109                 if (!PageUptodate(page)) {
5110                         btrfs_readpage(NULL, page);
5111                         lock_page(page);
5112                         if (!PageUptodate(page)) {
5113                                 unlock_page(page);
5114                                 page_cache_release(page);
5115                                 ret = -EIO;
5116                                 goto out_unlock;
5117                         }
5118                 }
5119                 wait_on_page_writeback(page);
5120
5121                 page_start = (u64)page->index << PAGE_CACHE_SHIFT;
5122                 page_end = page_start + PAGE_CACHE_SIZE - 1;
5123                 lock_extent(io_tree, page_start, page_end, GFP_NOFS);
5124
5125                 ordered = btrfs_lookup_ordered_extent(inode, page_start);
5126                 if (ordered) {
5127                         unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
5128                         unlock_page(page);
5129                         page_cache_release(page);
5130                         btrfs_start_ordered_extent(inode, ordered, 1);
5131                         btrfs_put_ordered_extent(ordered);
5132                         goto again;
5133                 }
5134                 set_page_extent_mapped(page);
5135
5136                 if (i == first_index)
5137                         set_extent_bits(io_tree, page_start, page_end,
5138                                         EXTENT_BOUNDARY, GFP_NOFS);
5139                 btrfs_set_extent_delalloc(inode, page_start, page_end);
5140
5141                 set_page_dirty(page);
5142                 total_dirty++;
5143
5144                 unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
5145                 unlock_page(page);
5146                 page_cache_release(page);
5147         }
5148
5149 out_unlock:
5150         kfree(ra);
5151         mutex_unlock(&inode->i_mutex);
5152         balance_dirty_pages_ratelimited_nr(inode->i_mapping, total_dirty);
5153         return ret;
5154 }
5155
5156 static noinline int relocate_data_extent(struct inode *reloc_inode,
5157                                          struct btrfs_key *extent_key,
5158                                          u64 offset)
5159 {
5160         struct btrfs_root *root = BTRFS_I(reloc_inode)->root;
5161         struct extent_map_tree *em_tree = &BTRFS_I(reloc_inode)->extent_tree;
5162         struct extent_map *em;
5163         u64 start = extent_key->objectid - offset;
5164         u64 end = start + extent_key->offset - 1;
5165
5166         em = alloc_extent_map(GFP_NOFS);
5167         BUG_ON(!em || IS_ERR(em));
5168
5169         em->start = start;
5170         em->len = extent_key->offset;
5171         em->block_len = extent_key->offset;
5172         em->block_start = extent_key->objectid;
5173         em->bdev = root->fs_info->fs_devices->latest_bdev;
5174         set_bit(EXTENT_FLAG_PINNED, &em->flags);
5175
5176         /* setup extent map to cheat btrfs_readpage */
5177         lock_extent(&BTRFS_I(reloc_inode)->io_tree, start, end, GFP_NOFS);
5178         while (1) {
5179                 int ret;
5180                 spin_lock(&em_tree->lock);
5181                 ret = add_extent_mapping(em_tree, em);
5182                 spin_unlock(&em_tree->lock);
5183                 if (ret != -EEXIST) {
5184                         free_extent_map(em);
5185                         break;
5186                 }
5187                 btrfs_drop_extent_cache(reloc_inode, start, end, 0);
5188         }
5189         unlock_extent(&BTRFS_I(reloc_inode)->io_tree, start, end, GFP_NOFS);
5190
5191         return relocate_inode_pages(reloc_inode, start, extent_key->offset);
5192 }
5193
5194 struct btrfs_ref_path {
5195         u64 extent_start;
5196         u64 nodes[BTRFS_MAX_LEVEL];
5197         u64 root_objectid;
5198         u64 root_generation;
5199         u64 owner_objectid;
5200         u32 num_refs;
5201         int lowest_level;
5202         int current_level;
5203         int shared_level;
5204
5205         struct btrfs_key node_keys[BTRFS_MAX_LEVEL];
5206         u64 new_nodes[BTRFS_MAX_LEVEL];
5207 };
5208
5209 struct disk_extent {
5210         u64 ram_bytes;
5211         u64 disk_bytenr;
5212         u64 disk_num_bytes;
5213         u64 offset;
5214         u64 num_bytes;
5215         u8 compression;
5216         u8 encryption;
5217         u16 other_encoding;
5218 };
5219
5220 static int is_cowonly_root(u64 root_objectid)
5221 {
5222         if (root_objectid == BTRFS_ROOT_TREE_OBJECTID ||
5223             root_objectid == BTRFS_EXTENT_TREE_OBJECTID ||
5224             root_objectid == BTRFS_CHUNK_TREE_OBJECTID ||
5225             root_objectid == BTRFS_DEV_TREE_OBJECTID ||
5226             root_objectid == BTRFS_TREE_LOG_OBJECTID ||
5227             root_objectid == BTRFS_CSUM_TREE_OBJECTID)
5228                 return 1;
5229         return 0;
5230 }
5231
5232 static noinline int __next_ref_path(struct btrfs_trans_handle *trans,
5233                                     struct btrfs_root *extent_root,
5234                                     struct btrfs_ref_path *ref_path,
5235                                     int first_time)
5236 {
5237         struct extent_buffer *leaf;
5238         struct btrfs_path *path;
5239         struct btrfs_extent_ref *ref;
5240         struct btrfs_key key;
5241         struct btrfs_key found_key;
5242         u64 bytenr;
5243         u32 nritems;
5244         int level;
5245         int ret = 1;
5246
5247         path = btrfs_alloc_path();
5248         if (!path)
5249                 return -ENOMEM;
5250
5251         if (first_time) {
5252                 ref_path->lowest_level = -1;
5253                 ref_path->current_level = -1;
5254                 ref_path->shared_level = -1;
5255                 goto walk_up;
5256         }
5257 walk_down:
5258         level = ref_path->current_level - 1;
5259         while (level >= -1) {
5260                 u64 parent;
5261                 if (level < ref_path->lowest_level)
5262                         break;
5263
5264                 if (level >= 0)
5265                         bytenr = ref_path->nodes[level];
5266                 else
5267                         bytenr = ref_path->extent_start;
5268                 BUG_ON(bytenr == 0);
5269
5270                 parent = ref_path->nodes[level + 1];
5271                 ref_path->nodes[level + 1] = 0;
5272                 ref_path->current_level = level;
5273                 BUG_ON(parent == 0);
5274
5275                 key.objectid = bytenr;
5276                 key.offset = parent + 1;
5277                 key.type = BTRFS_EXTENT_REF_KEY;
5278
5279                 ret = btrfs_search_slot(trans, extent_root, &key, path, 0, 0);
5280                 if (ret < 0)
5281                         goto out;
5282                 BUG_ON(ret == 0);
5283
5284                 leaf = path->nodes[0];
5285                 nritems = btrfs_header_nritems(leaf);
5286                 if (path->slots[0] >= nritems) {
5287                         ret = btrfs_next_leaf(extent_root, path);
5288                         if (ret < 0)
5289                                 goto out;
5290                         if (ret > 0)
5291                                 goto next;
5292                         leaf = path->nodes[0];
5293                 }
5294
5295                 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
5296                 if (found_key.objectid == bytenr &&
5297                     found_key.type == BTRFS_EXTENT_REF_KEY) {
5298                         if (level < ref_path->shared_level)
5299                                 ref_path->shared_level = level;
5300                         goto found;
5301                 }
5302 next:
5303                 level--;
5304                 btrfs_release_path(extent_root, path);
5305                 cond_resched();
5306         }
5307         /* reached lowest level */
5308         ret = 1;
5309         goto out;
5310 walk_up:
5311         level = ref_path->current_level;
5312         while (level < BTRFS_MAX_LEVEL - 1) {
5313                 u64 ref_objectid;
5314
5315                 if (level >= 0)
5316                         bytenr = ref_path->nodes[level];
5317                 else
5318                         bytenr = ref_path->extent_start;
5319
5320                 BUG_ON(bytenr == 0);
5321
5322                 key.objectid = bytenr;
5323                 key.offset = 0;
5324                 key.type = BTRFS_EXTENT_REF_KEY;
5325
5326                 ret = btrfs_search_slot(trans, extent_root, &key, path, 0, 0);
5327                 if (ret < 0)
5328                         goto out;
5329
5330                 leaf = path->nodes[0];
5331                 nritems = btrfs_header_nritems(leaf);
5332                 if (path->slots[0] >= nritems) {
5333                         ret = btrfs_next_leaf(extent_root, path);
5334                         if (ret < 0)
5335                                 goto out;
5336                         if (ret > 0) {
5337                                 /* the extent was freed by someone */
5338                                 if (ref_path->lowest_level == level)
5339                                         goto out;
5340                                 btrfs_release_path(extent_root, path);
5341                                 goto walk_down;
5342                         }
5343                         leaf = path->nodes[0];
5344                 }
5345
5346                 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
5347                 if (found_key.objectid != bytenr ||
5348                                 found_key.type != BTRFS_EXTENT_REF_KEY) {
5349                         /* the extent was freed by someone */
5350                         if (ref_path->lowest_level == level) {
5351                                 ret = 1;
5352                                 goto out;
5353                         }
5354                         btrfs_release_path(extent_root, path);
5355                         goto walk_down;
5356                 }
5357 found:
5358                 ref = btrfs_item_ptr(leaf, path->slots[0],
5359                                 struct btrfs_extent_ref);
5360                 ref_objectid = btrfs_ref_objectid(leaf, ref);
5361                 if (ref_objectid < BTRFS_FIRST_FREE_OBJECTID) {
5362                         if (first_time) {
5363                                 level = (int)ref_objectid;
5364                                 BUG_ON(level >= BTRFS_MAX_LEVEL);
5365                                 ref_path->lowest_level = level;
5366                                 ref_path->current_level = level;
5367                                 ref_path->nodes[level] = bytenr;
5368                         } else {
5369                                 WARN_ON(ref_objectid != level);
5370                         }
5371                 } else {
5372                         WARN_ON(level != -1);
5373                 }
5374                 first_time = 0;
5375
5376                 if (ref_path->lowest_level == level) {
5377                         ref_path->owner_objectid = ref_objectid;
5378                         ref_path->num_refs = btrfs_ref_num_refs(leaf, ref);
5379                 }
5380
5381                 /*
5382                  * the block is tree root or the block isn't in reference
5383                  * counted tree.
5384                  */
5385                 if (found_key.objectid == found_key.offset ||
5386                     is_cowonly_root(btrfs_ref_root(leaf, ref))) {
5387                         ref_path->root_objectid = btrfs_ref_root(leaf, ref);
5388                         ref_path->root_generation =
5389                                 btrfs_ref_generation(leaf, ref);
5390                         if (level < 0) {
5391                                 /* special reference from the tree log */
5392                                 ref_path->nodes[0] = found_key.offset;
5393                                 ref_path->current_level = 0;
5394                         }
5395                         ret = 0;
5396                         goto out;
5397                 }
5398
5399                 level++;
5400                 BUG_ON(ref_path->nodes[level] != 0);
5401                 ref_path->nodes[level] = found_key.offset;
5402                 ref_path->current_level = level;
5403
5404                 /*
5405                  * the reference was created in the running transaction,
5406                  * no need to continue walking up.
5407                  */
5408                 if (btrfs_ref_generation(leaf, ref) == trans->transid) {
5409                         ref_path->root_objectid = btrfs_ref_root(leaf, ref);
5410                         ref_path->root_generation =
5411                                 btrfs_ref_generation(leaf, ref);
5412                         ret = 0;
5413                         goto out;
5414                 }
5415
5416                 btrfs_release_path(extent_root, path);
5417                 cond_resched();
5418         }
5419         /* reached max tree level, but no tree root found. */
5420         BUG();
5421 out:
5422         btrfs_free_path(path);
5423         return ret;
5424 }
5425
5426 static int btrfs_first_ref_path(struct btrfs_trans_handle *trans,
5427                                 struct btrfs_root *extent_root,
5428                                 struct btrfs_ref_path *ref_path,
5429                                 u64 extent_start)
5430 {
5431         memset(ref_path, 0, sizeof(*ref_path));
5432         ref_path->extent_start = extent_start;
5433
5434         return __next_ref_path(trans, extent_root, ref_path, 1);
5435 }
5436
5437 static int btrfs_next_ref_path(struct btrfs_trans_handle *trans,
5438                                struct btrfs_root *extent_root,
5439                                struct btrfs_ref_path *ref_path)
5440 {
5441         return __next_ref_path(trans, extent_root, ref_path, 0);
5442 }
5443
5444 static noinline int get_new_locations(struct inode *reloc_inode,
5445                                       struct btrfs_key *extent_key,
5446                                       u64 offset, int no_fragment,
5447                                       struct disk_extent **extents,
5448                                       int *nr_extents)
5449 {
5450         struct btrfs_root *root = BTRFS_I(reloc_inode)->root;
5451         struct btrfs_path *path;
5452         struct btrfs_file_extent_item *fi;
5453         struct extent_buffer *leaf;
5454         struct disk_extent *exts = *extents;
5455         struct btrfs_key found_key;
5456         u64 cur_pos;
5457         u64 last_byte;
5458         u32 nritems;
5459         int nr = 0;
5460         int max = *nr_extents;
5461         int ret;
5462
5463         WARN_ON(!no_fragment && *extents);
5464         if (!exts) {
5465                 max = 1;
5466                 exts = kmalloc(sizeof(*exts) * max, GFP_NOFS);
5467                 if (!exts)
5468                         return -ENOMEM;
5469         }
5470
5471         path = btrfs_alloc_path();
5472         BUG_ON(!path);
5473
5474         cur_pos = extent_key->objectid - offset;
5475         last_byte = extent_key->objectid + extent_key->offset;
5476         ret = btrfs_lookup_file_extent(NULL, root, path, reloc_inode->i_ino,
5477                                        cur_pos, 0);
5478         if (ret < 0)
5479                 goto out;
5480         if (ret > 0) {
5481                 ret = -ENOENT;
5482                 goto out;
5483         }
5484
5485         while (1) {
5486                 leaf = path->nodes[0];
5487                 nritems = btrfs_header_nritems(leaf);
5488                 if (path->slots[0] >= nritems) {
5489                         ret = btrfs_next_leaf(root, path);
5490                         if (ret < 0)
5491                                 goto out;
5492                         if (ret > 0)
5493                                 break;
5494                         leaf = path->nodes[0];
5495                 }
5496
5497                 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
5498                 if (found_key.offset != cur_pos ||
5499                     found_key.type != BTRFS_EXTENT_DATA_KEY ||
5500                     found_key.objectid != reloc_inode->i_ino)
5501                         break;
5502
5503                 fi = btrfs_item_ptr(leaf, path->slots[0],
5504                                     struct btrfs_file_extent_item);
5505                 if (btrfs_file_extent_type(leaf, fi) !=
5506                     BTRFS_FILE_EXTENT_REG ||
5507                     btrfs_file_extent_disk_bytenr(leaf, fi) == 0)
5508                         break;
5509
5510                 if (nr == max) {
5511                         struct disk_extent *old = exts;
5512                         max *= 2;
5513                         exts = kzalloc(sizeof(*exts) * max, GFP_NOFS);
5514                         memcpy(exts, old, sizeof(*exts) * nr);
5515                         if (old != *extents)
5516                                 kfree(old);
5517                 }
5518
5519                 exts[nr].disk_bytenr =
5520                         btrfs_file_extent_disk_bytenr(leaf, fi);
5521                 exts[nr].disk_num_bytes =
5522                         btrfs_file_extent_disk_num_bytes(leaf, fi);
5523                 exts[nr].offset = btrfs_file_extent_offset(leaf, fi);
5524                 exts[nr].num_bytes = btrfs_file_extent_num_bytes(leaf, fi);
5525                 exts[nr].ram_bytes = btrfs_file_extent_ram_bytes(leaf, fi);
5526                 exts[nr].compression = btrfs_file_extent_compression(leaf, fi);
5527                 exts[nr].encryption = btrfs_file_extent_encryption(leaf, fi);
5528                 exts[nr].other_encoding = btrfs_file_extent_other_encoding(leaf,
5529                                                                            fi);
5530                 BUG_ON(exts[nr].offset > 0);
5531                 BUG_ON(exts[nr].compression || exts[nr].encryption);
5532                 BUG_ON(exts[nr].num_bytes != exts[nr].disk_num_bytes);
5533
5534                 cur_pos += exts[nr].num_bytes;
5535                 nr++;
5536
5537                 if (cur_pos + offset >= last_byte)
5538                         break;
5539
5540                 if (no_fragment) {
5541                         ret = 1;
5542                         goto out;
5543                 }
5544                 path->slots[0]++;
5545         }
5546
5547         BUG_ON(cur_pos + offset > last_byte);
5548         if (cur_pos + offset < last_byte) {
5549                 ret = -ENOENT;
5550                 goto out;
5551         }
5552         ret = 0;
5553 out:
5554         btrfs_free_path(path);
5555         if (ret) {
5556                 if (exts != *extents)
5557                         kfree(exts);
5558         } else {
5559                 *extents = exts;
5560                 *nr_extents = nr;
5561         }
5562         return ret;
5563 }
5564
5565 static noinline int replace_one_extent(struct btrfs_trans_handle *trans,
5566                                         struct btrfs_root *root,
5567                                         struct btrfs_path *path,
5568                                         struct btrfs_key *extent_key,
5569                                         struct btrfs_key *leaf_key,
5570                                         struct btrfs_ref_path *ref_path,
5571                                         struct disk_extent *new_extents,
5572                                         int nr_extents)
5573 {
5574         struct extent_buffer *leaf;
5575         struct btrfs_file_extent_item *fi;
5576         struct inode *inode = NULL;
5577         struct btrfs_key key;
5578         u64 lock_start = 0;
5579         u64 lock_end = 0;
5580         u64 num_bytes;
5581         u64 ext_offset;
5582         u64 search_end = (u64)-1;
5583         u32 nritems;
5584         int nr_scaned = 0;
5585         int extent_locked = 0;
5586         int extent_type;
5587         int ret;
5588
5589         memcpy(&key, leaf_key, sizeof(key));
5590         if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS) {
5591                 if (key.objectid < ref_path->owner_objectid ||
5592                     (key.objectid == ref_path->owner_objectid &&
5593                      key.type < BTRFS_EXTENT_DATA_KEY)) {
5594                         key.objectid = ref_path->owner_objectid;
5595                         key.type = BTRFS_EXTENT_DATA_KEY;
5596                         key.offset = 0;
5597                 }
5598         }
5599
5600         while (1) {
5601                 ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
5602                 if (ret < 0)
5603                         goto out;
5604
5605                 leaf = path->nodes[0];
5606                 nritems = btrfs_header_nritems(leaf);
5607 next:
5608                 if (extent_locked && ret > 0) {
5609                         /*
5610                          * the file extent item was modified by someone
5611                          * before the extent got locked.
5612                          */
5613                         unlock_extent(&BTRFS_I(inode)->io_tree, lock_start,
5614                                       lock_end, GFP_NOFS);
5615                         extent_locked = 0;
5616                 }
5617
5618                 if (path->slots[0] >= nritems) {
5619                         if (++nr_scaned > 2)
5620                                 break;
5621
5622                         BUG_ON(extent_locked);
5623                         ret = btrfs_next_leaf(root, path);
5624                         if (ret < 0)
5625                                 goto out;
5626                         if (ret > 0)
5627                                 break;
5628                         leaf = path->nodes[0];
5629                         nritems = btrfs_header_nritems(leaf);
5630                 }
5631
5632                 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
5633
5634                 if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS) {
5635                         if ((key.objectid > ref_path->owner_objectid) ||
5636                             (key.objectid == ref_path->owner_objectid &&
5637                              key.type > BTRFS_EXTENT_DATA_KEY) ||
5638                             key.offset >= search_end)
5639                                 break;
5640                 }
5641
5642                 if (inode && key.objectid != inode->i_ino) {
5643                         BUG_ON(extent_locked);
5644                         btrfs_release_path(root, path);
5645                         mutex_unlock(&inode->i_mutex);
5646                         iput(inode);
5647                         inode = NULL;
5648                         continue;
5649                 }
5650
5651                 if (key.type != BTRFS_EXTENT_DATA_KEY) {
5652                         path->slots[0]++;
5653                         ret = 1;
5654                         goto next;
5655                 }
5656                 fi = btrfs_item_ptr(leaf, path->slots[0],
5657                                     struct btrfs_file_extent_item);
5658                 extent_type = btrfs_file_extent_type(leaf, fi);
5659                 if ((extent_type != BTRFS_FILE_EXTENT_REG &&
5660                      extent_type != BTRFS_FILE_EXTENT_PREALLOC) ||
5661                     (btrfs_file_extent_disk_bytenr(leaf, fi) !=
5662                      extent_key->objectid)) {
5663                         path->slots[0]++;
5664                         ret = 1;
5665                         goto next;
5666                 }
5667
5668                 num_bytes = btrfs_file_extent_num_bytes(leaf, fi);
5669                 ext_offset = btrfs_file_extent_offset(leaf, fi);
5670
5671                 if (search_end == (u64)-1) {
5672                         search_end = key.offset - ext_offset +
5673                                 btrfs_file_extent_ram_bytes(leaf, fi);
5674                 }
5675
5676                 if (!extent_locked) {
5677                         lock_start = key.offset;
5678                         lock_end = lock_start + num_bytes - 1;
5679                 } else {
5680                         if (lock_start > key.offset ||
5681                             lock_end + 1 < key.offset + num_bytes) {
5682                                 unlock_extent(&BTRFS_I(inode)->io_tree,
5683                                               lock_start, lock_end, GFP_NOFS);
5684                                 extent_locked = 0;
5685                         }
5686                 }
5687
5688                 if (!inode) {
5689                         btrfs_release_path(root, path);
5690
5691                         inode = btrfs_iget_locked(root->fs_info->sb,
5692                                                   key.objectid, root);
5693                         if (inode->i_state & I_NEW) {
5694                                 BTRFS_I(inode)->root = root;
5695                                 BTRFS_I(inode)->location.objectid =
5696                                         key.objectid;
5697                                 BTRFS_I(inode)->location.type =
5698                                         BTRFS_INODE_ITEM_KEY;
5699                                 BTRFS_I(inode)->location.offset = 0;
5700                                 btrfs_read_locked_inode(inode);
5701                                 unlock_new_inode(inode);
5702                         }
5703                         /*
5704                          * some code call btrfs_commit_transaction while
5705                          * holding the i_mutex, so we can't use mutex_lock
5706                          * here.
5707                          */
5708                         if (is_bad_inode(inode) ||
5709                             !mutex_trylock(&inode->i_mutex)) {
5710                                 iput(inode);
5711                                 inode = NULL;
5712                                 key.offset = (u64)-1;
5713                                 goto skip;
5714                         }
5715                 }
5716
5717                 if (!extent_locked) {
5718                         struct btrfs_ordered_extent *ordered;
5719
5720                         btrfs_release_path(root, path);
5721
5722                         lock_extent(&BTRFS_I(inode)->io_tree, lock_start,
5723                                     lock_end, GFP_NOFS);
5724                         ordered = btrfs_lookup_first_ordered_extent(inode,
5725                                                                     lock_end);
5726                         if (ordered &&
5727                             ordered->file_offset <= lock_end &&
5728                             ordered->file_offset + ordered->len > lock_start) {
5729                                 unlock_extent(&BTRFS_I(inode)->io_tree,
5730                                               lock_start, lock_end, GFP_NOFS);
5731                                 btrfs_start_ordered_extent(inode, ordered, 1);
5732                                 btrfs_put_ordered_extent(ordered);
5733                                 key.offset += num_bytes;
5734                                 goto skip;
5735                         }
5736                         if (ordered)
5737                                 btrfs_put_ordered_extent(ordered);
5738
5739                         extent_locked = 1;
5740                         continue;
5741                 }
5742
5743                 if (nr_extents == 1) {
5744                         /* update extent pointer in place */
5745                         btrfs_set_file_extent_disk_bytenr(leaf, fi,
5746                                                 new_extents[0].disk_bytenr);
5747                         btrfs_set_file_extent_disk_num_bytes(leaf, fi,
5748                                                 new_extents[0].disk_num_bytes);
5749                         btrfs_mark_buffer_dirty(leaf);
5750
5751                         btrfs_drop_extent_cache(inode, key.offset,
5752                                                 key.offset + num_bytes - 1, 0);
5753
5754                         ret = btrfs_inc_extent_ref(trans, root,
5755                                                 new_extents[0].disk_bytenr,
5756                                                 new_extents[0].disk_num_bytes,
5757                                                 leaf->start,
5758                                                 root->root_key.objectid,
5759                                                 trans->transid,
5760                                                 key.objectid);
5761                         BUG_ON(ret);
5762
5763                         ret = btrfs_free_extent(trans, root,
5764                                                 extent_key->objectid,
5765                                                 extent_key->offset,
5766                                                 leaf->start,
5767                                                 btrfs_header_owner(leaf),
5768                                                 btrfs_header_generation(leaf),
5769                                                 key.objectid, 0);
5770                         BUG_ON(ret);
5771
5772                         btrfs_release_path(root, path);
5773                         key.offset += num_bytes;
5774                 } else {
5775                         BUG_ON(1);
5776 #if 0
5777                         u64 alloc_hint;
5778                         u64 extent_len;
5779                         int i;
5780                         /*
5781                          * drop old extent pointer at first, then insert the
5782                          * new pointers one bye one
5783                          */
5784                         btrfs_release_path(root, path);
5785                         ret = btrfs_drop_extents(trans, root, inode, key.offset,
5786                                                  key.offset + num_bytes,
5787                                                  key.offset, &alloc_hint);
5788                         BUG_ON(ret);
5789
5790                         for (i = 0; i < nr_extents; i++) {
5791                                 if (ext_offset >= new_extents[i].num_bytes) {
5792                                         ext_offset -= new_extents[i].num_bytes;
5793                                         continue;
5794                                 }
5795                                 extent_len = min(new_extents[i].num_bytes -
5796                                                  ext_offset, num_bytes);
5797
5798                                 ret = btrfs_insert_empty_item(trans, root,
5799                                                               path, &key,
5800                                                               sizeof(*fi));
5801                                 BUG_ON(ret);
5802
5803                                 leaf = path->nodes[0];
5804                                 fi = btrfs_item_ptr(leaf, path->slots[0],
5805                                                 struct btrfs_file_extent_item);
5806                                 btrfs_set_file_extent_generation(leaf, fi,
5807                                                         trans->transid);
5808                                 btrfs_set_file_extent_type(leaf, fi,
5809                                                         BTRFS_FILE_EXTENT_REG);
5810                                 btrfs_set_file_extent_disk_bytenr(leaf, fi,
5811                                                 new_extents[i].disk_bytenr);
5812                                 btrfs_set_file_extent_disk_num_bytes(leaf, fi,
5813                                                 new_extents[i].disk_num_bytes);
5814                                 btrfs_set_file_extent_ram_bytes(leaf, fi,
5815                                                 new_extents[i].ram_bytes);
5816
5817                                 btrfs_set_file_extent_compression(leaf, fi,
5818                                                 new_extents[i].compression);
5819                                 btrfs_set_file_extent_encryption(leaf, fi,
5820                                                 new_extents[i].encryption);
5821                                 btrfs_set_file_extent_other_encoding(leaf, fi,
5822                                                 new_extents[i].other_encoding);
5823
5824                                 btrfs_set_file_extent_num_bytes(leaf, fi,
5825                                                         extent_len);
5826                                 ext_offset += new_extents[i].offset;
5827                                 btrfs_set_file_extent_offset(leaf, fi,
5828                                                         ext_offset);
5829                                 btrfs_mark_buffer_dirty(leaf);
5830
5831                                 btrfs_drop_extent_cache(inode, key.offset,
5832                                                 key.offset + extent_len - 1, 0);
5833
5834                                 ret = btrfs_inc_extent_ref(trans, root,
5835                                                 new_extents[i].disk_bytenr,
5836                                                 new_extents[i].disk_num_bytes,
5837                                                 leaf->start,
5838                                                 root->root_key.objectid,
5839                                                 trans->transid, key.objectid);
5840                                 BUG_ON(ret);
5841                                 btrfs_release_path(root, path);
5842
5843                                 inode_add_bytes(inode, extent_len);
5844
5845                                 ext_offset = 0;
5846                                 num_bytes -= extent_len;
5847                                 key.offset += extent_len;
5848
5849                                 if (num_bytes == 0)
5850                                         break;
5851                         }
5852                         BUG_ON(i >= nr_extents);
5853 #endif
5854                 }
5855
5856                 if (extent_locked) {
5857                         unlock_extent(&BTRFS_I(inode)->io_tree, lock_start,
5858                                       lock_end, GFP_NOFS);
5859                         extent_locked = 0;
5860                 }
5861 skip:
5862                 if (ref_path->owner_objectid != BTRFS_MULTIPLE_OBJECTIDS &&
5863                     key.offset >= search_end)
5864                         break;
5865
5866                 cond_resched();
5867         }
5868         ret = 0;
5869 out:
5870         btrfs_release_path(root, path);
5871         if (inode) {
5872                 mutex_unlock(&inode->i_mutex);
5873                 if (extent_locked) {
5874                         unlock_extent(&BTRFS_I(inode)->io_tree, lock_start,
5875                                       lock_end, GFP_NOFS);
5876                 }
5877                 iput(inode);
5878         }
5879         return ret;
5880 }
5881
5882 int btrfs_reloc_tree_cache_ref(struct btrfs_trans_handle *trans,
5883                                struct btrfs_root *root,
5884                                struct extent_buffer *buf, u64 orig_start)
5885 {
5886         int level;
5887         int ret;
5888
5889         BUG_ON(btrfs_header_generation(buf) != trans->transid);
5890         BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID);
5891
5892         level = btrfs_header_level(buf);
5893         if (level == 0) {
5894                 struct btrfs_leaf_ref *ref;
5895                 struct btrfs_leaf_ref *orig_ref;
5896
5897                 orig_ref = btrfs_lookup_leaf_ref(root, orig_start);
5898                 if (!orig_ref)
5899                         return -ENOENT;
5900
5901                 ref = btrfs_alloc_leaf_ref(root, orig_ref->nritems);
5902                 if (!ref) {
5903                         btrfs_free_leaf_ref(root, orig_ref);
5904                         return -ENOMEM;
5905                 }
5906
5907                 ref->nritems = orig_ref->nritems;
5908                 memcpy(ref->extents, orig_ref->extents,
5909                         sizeof(ref->extents[0]) * ref->nritems);
5910
5911                 btrfs_free_leaf_ref(root, orig_ref);
5912
5913                 ref->root_gen = trans->transid;
5914                 ref->bytenr = buf->start;
5915                 ref->owner = btrfs_header_owner(buf);
5916                 ref->generation = btrfs_header_generation(buf);
5917
5918                 ret = btrfs_add_leaf_ref(root, ref, 0);
5919                 WARN_ON(ret);
5920                 btrfs_free_leaf_ref(root, ref);
5921         }
5922         return 0;
5923 }
5924
5925 static noinline int invalidate_extent_cache(struct btrfs_root *root,
5926                                         struct extent_buffer *leaf,
5927                                         struct btrfs_block_group_cache *group,
5928                                         struct btrfs_root *target_root)
5929 {
5930         struct btrfs_key key;
5931         struct inode *inode = NULL;
5932         struct btrfs_file_extent_item *fi;
5933         u64 num_bytes;
5934         u64 skip_objectid = 0;
5935         u32 nritems;
5936         u32 i;
5937
5938         nritems = btrfs_header_nritems(leaf);
5939         for (i = 0; i < nritems; i++) {
5940                 btrfs_item_key_to_cpu(leaf, &key, i);
5941                 if (key.objectid == skip_objectid ||
5942                     key.type != BTRFS_EXTENT_DATA_KEY)
5943                         continue;
5944                 fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item);
5945                 if (btrfs_file_extent_type(leaf, fi) ==
5946                     BTRFS_FILE_EXTENT_INLINE)
5947                         continue;
5948                 if (btrfs_file_extent_disk_bytenr(leaf, fi) == 0)
5949                         continue;
5950                 if (!inode || inode->i_ino != key.objectid) {
5951                         iput(inode);
5952                         inode = btrfs_ilookup(target_root->fs_info->sb,
5953                                               key.objectid, target_root, 1);
5954                 }
5955                 if (!inode) {
5956                         skip_objectid = key.objectid;
5957                         continue;
5958                 }
5959                 num_bytes = btrfs_file_extent_num_bytes(leaf, fi);
5960
5961                 lock_extent(&BTRFS_I(inode)->io_tree, key.offset,
5962                             key.offset + num_bytes - 1, GFP_NOFS);
5963                 btrfs_drop_extent_cache(inode, key.offset,
5964                                         key.offset + num_bytes - 1, 1);
5965                 unlock_extent(&BTRFS_I(inode)->io_tree, key.offset,
5966                               key.offset + num_bytes - 1, GFP_NOFS);
5967                 cond_resched();
5968         }
5969         iput(inode);
5970         return 0;
5971 }
5972
5973 static noinline int replace_extents_in_leaf(struct btrfs_trans_handle *trans,
5974                                         struct btrfs_root *root,
5975                                         struct extent_buffer *leaf,
5976                                         struct btrfs_block_group_cache *group,
5977                                         struct inode *reloc_inode)
5978 {
5979         struct btrfs_key key;
5980         struct btrfs_key extent_key;
5981         struct btrfs_file_extent_item *fi;
5982         struct btrfs_leaf_ref *ref;
5983         struct disk_extent *new_extent;
5984         u64 bytenr;
5985         u64 num_bytes;
5986         u32 nritems;
5987         u32 i;
5988         int ext_index;
5989         int nr_extent;
5990         int ret;
5991
5992         new_extent = kmalloc(sizeof(*new_extent), GFP_NOFS);
5993         BUG_ON(!new_extent);
5994
5995         ref = btrfs_lookup_leaf_ref(root, leaf->start);
5996         BUG_ON(!ref);
5997
5998         ext_index = -1;
5999         nritems = btrfs_header_nritems(leaf);
6000         for (i = 0; i < nritems; i++) {
6001                 btrfs_item_key_to_cpu(leaf, &key, i);
6002                 if (btrfs_key_type(&key) != BTRFS_EXTENT_DATA_KEY)
6003                         continue;
6004                 fi = btrfs_item_ptr(leaf, i, struct btrfs_file_extent_item);
6005                 if (btrfs_file_extent_type(leaf, fi) ==
6006                     BTRFS_FILE_EXTENT_INLINE)
6007                         continue;
6008                 bytenr = btrfs_file_extent_disk_bytenr(leaf, fi);
6009                 num_bytes = btrfs_file_extent_disk_num_bytes(leaf, fi);
6010                 if (bytenr == 0)
6011                         continue;
6012
6013                 ext_index++;
6014                 if (bytenr >= group->key.objectid + group->key.offset ||
6015                     bytenr + num_bytes <= group->key.objectid)
6016                         continue;
6017
6018                 extent_key.objectid = bytenr;
6019                 extent_key.offset = num_bytes;
6020                 extent_key.type = BTRFS_EXTENT_ITEM_KEY;
6021                 nr_extent = 1;
6022                 ret = get_new_locations(reloc_inode, &extent_key,
6023                                         group->key.objectid, 1,
6024                                         &new_extent, &nr_extent);
6025                 if (ret > 0)
6026                         continue;
6027                 BUG_ON(ret < 0);
6028
6029                 BUG_ON(ref->extents[ext_index].bytenr != bytenr);
6030                 BUG_ON(ref->extents[ext_index].num_bytes != num_bytes);
6031                 ref->extents[ext_index].bytenr = new_extent->disk_bytenr;
6032                 ref->extents[ext_index].num_bytes = new_extent->disk_num_bytes;
6033
6034                 btrfs_set_file_extent_disk_bytenr(leaf, fi,
6035                                                 new_extent->disk_bytenr);
6036                 btrfs_set_file_extent_disk_num_bytes(leaf, fi,
6037                                                 new_extent->disk_num_bytes);
6038                 btrfs_mark_buffer_dirty(leaf);
6039
6040                 ret = btrfs_inc_extent_ref(trans, root,
6041                                         new_extent->disk_bytenr,
6042                                         new_extent->disk_num_bytes,
6043                                         leaf->start,
6044                                         root->root_key.objectid,
6045                                         trans->transid, key.objectid);
6046                 BUG_ON(ret);
6047
6048                 ret = btrfs_free_extent(trans, root,
6049                                         bytenr, num_bytes, leaf->start,
6050                                         btrfs_header_owner(leaf),
6051                                         btrfs_header_generation(leaf),
6052                                         key.objectid, 0);
6053                 BUG_ON(ret);
6054                 cond_resched();
6055         }
6056         kfree(new_extent);
6057         BUG_ON(ext_index + 1 != ref->nritems);
6058         btrfs_free_leaf_ref(root, ref);
6059         return 0;
6060 }
6061
6062 int btrfs_free_reloc_root(struct btrfs_trans_handle *trans,
6063                           struct btrfs_root *root)
6064 {
6065         struct btrfs_root *reloc_root;
6066         int ret;
6067
6068         if (root->reloc_root) {
6069                 reloc_root = root->reloc_root;
6070                 root->reloc_root = NULL;
6071                 list_add(&reloc_root->dead_list,
6072                          &root->fs_info->dead_reloc_roots);
6073
6074                 btrfs_set_root_bytenr(&reloc_root->root_item,
6075                                       reloc_root->node->start);
6076                 btrfs_set_root_level(&root->root_item,
6077                                      btrfs_header_level(reloc_root->node));
6078                 memset(&reloc_root->root_item.drop_progress, 0,
6079                         sizeof(struct btrfs_disk_key));
6080                 reloc_root->root_item.drop_level = 0;
6081
6082                 ret = btrfs_update_root(trans, root->fs_info->tree_root,
6083                                         &reloc_root->root_key,
6084                                         &reloc_root->root_item);
6085                 BUG_ON(ret);
6086         }
6087         return 0;
6088 }
6089
6090 int btrfs_drop_dead_reloc_roots(struct btrfs_root *root)
6091 {
6092         struct btrfs_trans_handle *trans;
6093         struct btrfs_root *reloc_root;
6094         struct btrfs_root *prev_root = NULL;
6095         struct list_head dead_roots;
6096         int ret;
6097         unsigned long nr;
6098
6099         INIT_LIST_HEAD(&dead_roots);
6100         list_splice_init(&root->fs_info->dead_reloc_roots, &dead_roots);
6101
6102         while (!list_empty(&dead_roots)) {
6103                 reloc_root = list_entry(dead_roots.prev,
6104                                         struct btrfs_root, dead_list);
6105                 list_del_init(&reloc_root->dead_list);
6106
6107                 BUG_ON(reloc_root->commit_root != NULL);
6108                 while (1) {
6109                         trans = btrfs_join_transaction(root, 1);
6110                         BUG_ON(!trans);
6111
6112                         mutex_lock(&root->fs_info->drop_mutex);
6113                         ret = btrfs_drop_snapshot(trans, reloc_root);
6114                         if (ret != -EAGAIN)
6115                                 break;
6116                         mutex_unlock(&root->fs_info->drop_mutex);
6117
6118                         nr = trans->blocks_used;
6119                         ret = btrfs_end_transaction(trans, root);
6120                         BUG_ON(ret);
6121                         btrfs_btree_balance_dirty(root, nr);
6122                 }
6123
6124                 free_extent_buffer(reloc_root->node);
6125
6126                 ret = btrfs_del_root(trans, root->fs_info->tree_root,
6127                                      &reloc_root->root_key);
6128                 BUG_ON(ret);
6129                 mutex_unlock(&root->fs_info->drop_mutex);
6130
6131                 nr = trans->blocks_used;
6132                 ret = btrfs_end_transaction(trans, root);
6133                 BUG_ON(ret);
6134                 btrfs_btree_balance_dirty(root, nr);
6135
6136                 kfree(prev_root);
6137                 prev_root = reloc_root;
6138         }
6139         if (prev_root) {
6140                 btrfs_remove_leaf_refs(prev_root, (u64)-1, 0);
6141                 kfree(prev_root);
6142         }
6143         return 0;
6144 }
6145
6146 int btrfs_add_dead_reloc_root(struct btrfs_root *root)
6147 {
6148         list_add(&root->dead_list, &root->fs_info->dead_reloc_roots);
6149         return 0;
6150 }
6151
6152 int btrfs_cleanup_reloc_trees(struct btrfs_root *root)
6153 {
6154         struct btrfs_root *reloc_root;
6155         struct btrfs_trans_handle *trans;
6156         struct btrfs_key location;
6157         int found;
6158         int ret;
6159
6160         mutex_lock(&root->fs_info->tree_reloc_mutex);
6161         ret = btrfs_find_dead_roots(root, BTRFS_TREE_RELOC_OBJECTID, NULL);
6162         BUG_ON(ret);
6163         found = !list_empty(&root->fs_info->dead_reloc_roots);
6164         mutex_unlock(&root->fs_info->tree_reloc_mutex);
6165
6166         if (found) {
6167                 trans = btrfs_start_transaction(root, 1);
6168                 BUG_ON(!trans);
6169                 ret = btrfs_commit_transaction(trans, root);
6170                 BUG_ON(ret);
6171         }
6172
6173         location.objectid = BTRFS_DATA_RELOC_TREE_OBJECTID;
6174         location.offset = (u64)-1;
6175         location.type = BTRFS_ROOT_ITEM_KEY;
6176
6177         reloc_root = btrfs_read_fs_root_no_name(root->fs_info, &location);
6178         BUG_ON(!reloc_root);
6179         btrfs_orphan_cleanup(reloc_root);
6180         return 0;
6181 }
6182
6183 static noinline int init_reloc_tree(struct btrfs_trans_handle *trans,
6184                                     struct btrfs_root *root)
6185 {
6186         struct btrfs_root *reloc_root;
6187         struct extent_buffer *eb;
6188         struct btrfs_root_item *root_item;
6189         struct btrfs_key root_key;
6190         int ret;
6191
6192         BUG_ON(!root->ref_cows);
6193         if (root->reloc_root)
6194                 return 0;
6195
6196         root_item = kmalloc(sizeof(*root_item), GFP_NOFS);
6197         BUG_ON(!root_item);
6198
6199         ret = btrfs_copy_root(trans, root, root->commit_root,
6200                               &eb, BTRFS_TREE_RELOC_OBJECTID);
6201         BUG_ON(ret);
6202
6203         root_key.objectid = BTRFS_TREE_RELOC_OBJECTID;
6204         root_key.offset = root->root_key.objectid;
6205         root_key.type = BTRFS_ROOT_ITEM_KEY;
6206
6207         memcpy(root_item, &root->root_item, sizeof(root_item));
6208         btrfs_set_root_refs(root_item, 0);
6209         btrfs_set_root_bytenr(root_item, eb->start);
6210         btrfs_set_root_level(root_item, btrfs_header_level(eb));
6211         btrfs_set_root_generation(root_item, trans->transid);
6212
6213         btrfs_tree_unlock(eb);
6214         free_extent_buffer(eb);
6215
6216         ret = btrfs_insert_root(trans, root->fs_info->tree_root,
6217                                 &root_key, root_item);
6218         BUG_ON(ret);
6219         kfree(root_item);
6220
6221         reloc_root = btrfs_read_fs_root_no_radix(root->fs_info->tree_root,
6222                                                  &root_key);
6223         BUG_ON(!reloc_root);
6224         reloc_root->last_trans = trans->transid;
6225         reloc_root->commit_root = NULL;
6226         reloc_root->ref_tree = &root->fs_info->reloc_ref_tree;
6227
6228         root->reloc_root = reloc_root;
6229         return 0;
6230 }
6231
6232 /*
6233  * Core function of space balance.
6234  *
6235  * The idea is using reloc trees to relocate tree blocks in reference
6236  * counted roots. There is one reloc tree for each subvol, and all
6237  * reloc trees share same root key objectid. Reloc trees are snapshots
6238  * of the latest committed roots of subvols (root->commit_root).
6239  *
6240  * To relocate a tree block referenced by a subvol, there are two steps.
6241  * COW the block through subvol's reloc tree, then update block pointer
6242  * in the subvol to point to the new block. Since all reloc trees share
6243  * same root key objectid, doing special handing for tree blocks owned
6244  * by them is easy. Once a tree block has been COWed in one reloc tree,
6245  * we can use the resulting new block directly when the same block is
6246  * required to COW again through other reloc trees. By this way, relocated
6247  * tree blocks are shared between reloc trees, so they are also shared
6248  * between subvols.
6249  */
6250 static noinline int relocate_one_path(struct btrfs_trans_handle *trans,
6251                                       struct btrfs_root *root,
6252                                       struct btrfs_path *path,
6253                                       struct btrfs_key *first_key,
6254                                       struct btrfs_ref_path *ref_path,
6255                                       struct btrfs_block_group_cache *group,
6256                                       struct inode *reloc_inode)
6257 {
6258         struct btrfs_root *reloc_root;
6259         struct extent_buffer *eb = NULL;
6260         struct btrfs_key *keys;
6261         u64 *nodes;
6262         int level;
6263         int shared_level;
6264         int lowest_level = 0;
6265         int ret;
6266
6267         if (ref_path->owner_objectid < BTRFS_FIRST_FREE_OBJECTID)
6268                 lowest_level = ref_path->owner_objectid;
6269
6270         if (!root->ref_cows) {
6271                 path->lowest_level = lowest_level;
6272                 ret = btrfs_search_slot(trans, root, first_key, path, 0, 1);
6273                 BUG_ON(ret < 0);
6274                 path->lowest_level = 0;
6275                 btrfs_release_path(root, path);
6276                 return 0;
6277         }
6278
6279         mutex_lock(&root->fs_info->tree_reloc_mutex);
6280         ret = init_reloc_tree(trans, root);
6281         BUG_ON(ret);
6282         reloc_root = root->reloc_root;
6283
6284         shared_level = ref_path->shared_level;
6285         ref_path->shared_level = BTRFS_MAX_LEVEL - 1;
6286
6287         keys = ref_path->node_keys;
6288         nodes = ref_path->new_nodes;
6289         memset(&keys[shared_level + 1], 0,
6290                sizeof(*keys) * (BTRFS_MAX_LEVEL - shared_level - 1));
6291         memset(&nodes[shared_level + 1], 0,
6292                sizeof(*nodes) * (BTRFS_MAX_LEVEL - shared_level - 1));
6293
6294         if (nodes[lowest_level] == 0) {
6295                 path->lowest_level = lowest_level;
6296                 ret = btrfs_search_slot(trans, reloc_root, first_key, path,
6297                                         0, 1);
6298                 BUG_ON(ret);
6299                 for (level = lowest_level; level < BTRFS_MAX_LEVEL; level++) {
6300                         eb = path->nodes[level];
6301                         if (!eb || eb == reloc_root->node)
6302                                 break;
6303                         nodes[level] = eb->start;
6304                         if (level == 0)
6305                                 btrfs_item_key_to_cpu(eb, &keys[level], 0);
6306                         else
6307                                 btrfs_node_key_to_cpu(eb, &keys[level], 0);
6308                 }
6309                 if (nodes[0] &&
6310                     ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
6311                         eb = path->nodes[0];
6312                         ret = replace_extents_in_leaf(trans, reloc_root, eb,
6313                                                       group, reloc_inode);
6314                         BUG_ON(ret);
6315                 }
6316                 btrfs_release_path(reloc_root, path);
6317         } else {
6318                 ret = btrfs_merge_path(trans, reloc_root, keys, nodes,
6319                                        lowest_level);
6320                 BUG_ON(ret);
6321         }
6322
6323         /*
6324          * replace tree blocks in the fs tree with tree blocks in
6325          * the reloc tree.
6326          */
6327         ret = btrfs_merge_path(trans, root, keys, nodes, lowest_level);
6328         BUG_ON(ret < 0);
6329
6330         if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
6331                 ret = btrfs_search_slot(trans, reloc_root, first_key, path,
6332                                         0, 0);
6333                 BUG_ON(ret);
6334                 extent_buffer_get(path->nodes[0]);
6335                 eb = path->nodes[0];
6336                 btrfs_release_path(reloc_root, path);
6337                 ret = invalidate_extent_cache(reloc_root, eb, group, root);
6338                 BUG_ON(ret);
6339                 free_extent_buffer(eb);
6340         }
6341
6342         mutex_unlock(&root->fs_info->tree_reloc_mutex);
6343         path->lowest_level = 0;
6344         return 0;
6345 }
6346
6347 static noinline int relocate_tree_block(struct btrfs_trans_handle *trans,
6348                                         struct btrfs_root *root,
6349                                         struct btrfs_path *path,
6350                                         struct btrfs_key *first_key,
6351                                         struct btrfs_ref_path *ref_path)
6352 {
6353         int ret;
6354
6355         ret = relocate_one_path(trans, root, path, first_key,
6356                                 ref_path, NULL, NULL);
6357         BUG_ON(ret);
6358
6359         return 0;
6360 }
6361
6362 static noinline int del_extent_zero(struct btrfs_trans_handle *trans,
6363                                     struct btrfs_root *extent_root,
6364                                     struct btrfs_path *path,
6365                                     struct btrfs_key *extent_key)
6366 {
6367         int ret;
6368
6369         ret = btrfs_search_slot(trans, extent_root, extent_key, path, -1, 1);
6370         if (ret)
6371                 goto out;
6372         ret = btrfs_del_item(trans, extent_root, path);
6373 out:
6374         btrfs_release_path(extent_root, path);
6375         return ret;
6376 }
6377
6378 static noinline struct btrfs_root *read_ref_root(struct btrfs_fs_info *fs_info,
6379                                                 struct btrfs_ref_path *ref_path)
6380 {
6381         struct btrfs_key root_key;
6382
6383         root_key.objectid = ref_path->root_objectid;
6384         root_key.type = BTRFS_ROOT_ITEM_KEY;
6385         if (is_cowonly_root(ref_path->root_objectid))
6386                 root_key.offset = 0;
6387         else
6388                 root_key.offset = (u64)-1;
6389
6390         return btrfs_read_fs_root_no_name(fs_info, &root_key);
6391 }
6392
6393 static noinline int relocate_one_extent(struct btrfs_root *extent_root,
6394                                         struct btrfs_path *path,
6395                                         struct btrfs_key *extent_key,
6396                                         struct btrfs_block_group_cache *group,
6397                                         struct inode *reloc_inode, int pass)
6398 {
6399         struct btrfs_trans_handle *trans;
6400         struct btrfs_root *found_root;
6401         struct btrfs_ref_path *ref_path = NULL;
6402         struct disk_extent *new_extents = NULL;
6403         int nr_extents = 0;
6404         int loops;
6405         int ret;
6406         int level;
6407         struct btrfs_key first_key;
6408         u64 prev_block = 0;
6409
6410
6411         trans = btrfs_start_transaction(extent_root, 1);
6412         BUG_ON(!trans);
6413
6414         if (extent_key->objectid == 0) {
6415                 ret = del_extent_zero(trans, extent_root, path, extent_key);
6416                 goto out;
6417         }
6418
6419         ref_path = kmalloc(sizeof(*ref_path), GFP_NOFS);
6420         if (!ref_path) {
6421                 ret = -ENOMEM;
6422                 goto out;
6423         }
6424
6425         for (loops = 0; ; loops++) {
6426                 if (loops == 0) {
6427                         ret = btrfs_first_ref_path(trans, extent_root, ref_path,
6428                                                    extent_key->objectid);
6429                 } else {
6430                         ret = btrfs_next_ref_path(trans, extent_root, ref_path);
6431                 }
6432                 if (ret < 0)
6433                         goto out;
6434                 if (ret > 0)
6435                         break;
6436
6437                 if (ref_path->root_objectid == BTRFS_TREE_LOG_OBJECTID ||
6438                     ref_path->root_objectid == BTRFS_TREE_RELOC_OBJECTID)
6439                         continue;
6440
6441                 found_root = read_ref_root(extent_root->fs_info, ref_path);
6442                 BUG_ON(!found_root);
6443                 /*
6444                  * for reference counted tree, only process reference paths
6445                  * rooted at the latest committed root.
6446                  */
6447                 if (found_root->ref_cows &&
6448                     ref_path->root_generation != found_root->root_key.offset)
6449                         continue;
6450
6451                 if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
6452                         if (pass == 0) {
6453                                 /*
6454                                  * copy data extents to new locations
6455                                  */
6456                                 u64 group_start = group->key.objectid;
6457                                 ret = relocate_data_extent(reloc_inode,
6458                                                            extent_key,
6459                                                            group_start);
6460                                 if (ret < 0)
6461                                         goto out;
6462                                 break;
6463                         }
6464                         level = 0;
6465                 } else {
6466                         level = ref_path->owner_objectid;
6467                 }
6468
6469                 if (prev_block != ref_path->nodes[level]) {
6470                         struct extent_buffer *eb;
6471                         u64 block_start = ref_path->nodes[level];
6472                         u64 block_size = btrfs_level_size(found_root, level);
6473
6474                         eb = read_tree_block(found_root, block_start,
6475                                              block_size, 0);
6476                         btrfs_tree_lock(eb);
6477                         BUG_ON(level != btrfs_header_level(eb));
6478
6479                         if (level == 0)
6480                                 btrfs_item_key_to_cpu(eb, &first_key, 0);
6481                         else
6482                                 btrfs_node_key_to_cpu(eb, &first_key, 0);
6483
6484                         btrfs_tree_unlock(eb);
6485                         free_extent_buffer(eb);
6486                         prev_block = block_start;
6487                 }
6488
6489                 mutex_lock(&extent_root->fs_info->trans_mutex);
6490                 btrfs_record_root_in_trans(found_root);
6491                 mutex_unlock(&extent_root->fs_info->trans_mutex);
6492                 if (ref_path->owner_objectid >= BTRFS_FIRST_FREE_OBJECTID) {
6493                         /*
6494                          * try to update data extent references while
6495                          * keeping metadata shared between snapshots.
6496                          */
6497                         if (pass == 1) {
6498                                 ret = relocate_one_path(trans, found_root,
6499                                                 path, &first_key, ref_path,
6500                                                 group, reloc_inode);
6501                                 if (ret < 0)
6502                                         goto out;
6503                                 continue;
6504                         }
6505                         /*
6506                          * use fallback method to process the remaining
6507                          * references.
6508                          */
6509                         if (!new_extents) {
6510                                 u64 group_start = group->key.objectid;
6511                                 new_extents = kmalloc(sizeof(*new_extents),
6512                                                       GFP_NOFS);
6513                                 nr_extents = 1;
6514                                 ret = get_new_locations(reloc_inode,
6515                                                         extent_key,
6516                                                         group_start, 1,
6517                                                         &new_extents,
6518                                                         &nr_extents);
6519                                 if (ret)
6520                                         goto out;
6521                         }
6522                         ret = replace_one_extent(trans, found_root,
6523                                                 path, extent_key,
6524                                                 &first_key, ref_path,
6525                                                 new_extents, nr_extents);
6526                 } else {
6527                         ret = relocate_tree_block(trans, found_root, path,
6528                                                   &first_key, ref_path);
6529                 }
6530                 if (ret < 0)
6531                         goto out;
6532         }
6533         ret = 0;
6534 out:
6535         btrfs_end_transaction(trans, extent_root);
6536         kfree(new_extents);
6537         kfree(ref_path);
6538         return ret;
6539 }
6540 #endif
6541
6542 static u64 update_block_group_flags(struct btrfs_root *root, u64 flags)
6543 {
6544         u64 num_devices;
6545         u64 stripped = BTRFS_BLOCK_GROUP_RAID0 |
6546                 BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10;
6547
6548         num_devices = root->fs_info->fs_devices->rw_devices;
6549         if (num_devices == 1) {
6550                 stripped |= BTRFS_BLOCK_GROUP_DUP;
6551                 stripped = flags & ~stripped;
6552
6553                 /* turn raid0 into single device chunks */
6554                 if (flags & BTRFS_BLOCK_GROUP_RAID0)
6555                         return stripped;
6556
6557                 /* turn mirroring into duplication */
6558                 if (flags & (BTRFS_BLOCK_GROUP_RAID1 |
6559                              BTRFS_BLOCK_GROUP_RAID10))
6560                         return stripped | BTRFS_BLOCK_GROUP_DUP;
6561                 return flags;
6562         } else {
6563                 /* they already had raid on here, just return */
6564                 if (flags & stripped)
6565                         return flags;
6566
6567                 stripped |= BTRFS_BLOCK_GROUP_DUP;
6568                 stripped = flags & ~stripped;
6569
6570                 /* switch duplicated blocks with raid1 */
6571                 if (flags & BTRFS_BLOCK_GROUP_DUP)
6572                         return stripped | BTRFS_BLOCK_GROUP_RAID1;
6573
6574                 /* turn single device chunks into raid0 */
6575                 return stripped | BTRFS_BLOCK_GROUP_RAID0;
6576         }
6577         return flags;
6578 }
6579
6580 static int __alloc_chunk_for_shrink(struct btrfs_root *root,
6581                      struct btrfs_block_group_cache *shrink_block_group,
6582                      int force)
6583 {
6584         struct btrfs_trans_handle *trans;
6585         u64 new_alloc_flags;
6586         u64 calc;
6587
6588         spin_lock(&shrink_block_group->lock);
6589         if (btrfs_block_group_used(&shrink_block_group->item) +
6590             shrink_block_group->reserved > 0) {
6591                 spin_unlock(&shrink_block_group->lock);
6592
6593                 trans = btrfs_start_transaction(root, 1);
6594                 spin_lock(&shrink_block_group->lock);
6595
6596                 new_alloc_flags = update_block_group_flags(root,
6597                                                    shrink_block_group->flags);
6598                 if (new_alloc_flags != shrink_block_group->flags) {
6599                         calc =
6600                              btrfs_block_group_used(&shrink_block_group->item);
6601                 } else {
6602                         calc = shrink_block_group->key.offset;
6603                 }
6604                 spin_unlock(&shrink_block_group->lock);
6605
6606                 do_chunk_alloc(trans, root->fs_info->extent_root,
6607                                calc + 2 * 1024 * 1024, new_alloc_flags, force);
6608
6609                 btrfs_end_transaction(trans, root);
6610         } else
6611                 spin_unlock(&shrink_block_group->lock);
6612         return 0;
6613 }
6614
6615
6616 int btrfs_prepare_block_group_relocation(struct btrfs_root *root,
6617                                          struct btrfs_block_group_cache *group)
6618
6619 {
6620         __alloc_chunk_for_shrink(root, group, 1);
6621         set_block_group_readonly(group);
6622         return 0;
6623 }
6624
6625 #if 0
6626 static int __insert_orphan_inode(struct btrfs_trans_handle *trans,
6627                                  struct btrfs_root *root,
6628                                  u64 objectid, u64 size)
6629 {
6630         struct btrfs_path *path;
6631         struct btrfs_inode_item *item;
6632         struct extent_buffer *leaf;
6633         int ret;
6634
6635         path = btrfs_alloc_path();
6636         if (!path)
6637                 return -ENOMEM;
6638
6639         path->leave_spinning = 1;
6640         ret = btrfs_insert_empty_inode(trans, root, path, objectid);
6641         if (ret)
6642                 goto out;
6643
6644         leaf = path->nodes[0];
6645         item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_inode_item);
6646         memset_extent_buffer(leaf, 0, (unsigned long)item, sizeof(*item));
6647         btrfs_set_inode_generation(leaf, item, 1);
6648         btrfs_set_inode_size(leaf, item, size);
6649         btrfs_set_inode_mode(leaf, item, S_IFREG | 0600);
6650         btrfs_set_inode_flags(leaf, item, BTRFS_INODE_NOCOMPRESS);
6651         btrfs_mark_buffer_dirty(leaf);
6652         btrfs_release_path(root, path);
6653 out:
6654         btrfs_free_path(path);
6655         return ret;
6656 }
6657
6658 static noinline struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info,
6659                                         struct btrfs_block_group_cache *group)
6660 {
6661         struct inode *inode = NULL;
6662         struct btrfs_trans_handle *trans;
6663         struct btrfs_root *root;
6664         struct btrfs_key root_key;
6665         u64 objectid = BTRFS_FIRST_FREE_OBJECTID;
6666         int err = 0;
6667
6668         root_key.objectid = BTRFS_DATA_RELOC_TREE_OBJECTID;
6669         root_key.type = BTRFS_ROOT_ITEM_KEY;
6670         root_key.offset = (u64)-1;
6671         root = btrfs_read_fs_root_no_name(fs_info, &root_key);
6672         if (IS_ERR(root))
6673                 return ERR_CAST(root);
6674
6675         trans = btrfs_start_transaction(root, 1);
6676         BUG_ON(!trans);
6677
6678         err = btrfs_find_free_objectid(trans, root, objectid, &objectid);
6679         if (err)
6680                 goto out;
6681
6682         err = __insert_orphan_inode(trans, root, objectid, group->key.offset);
6683         BUG_ON(err);
6684
6685         err = btrfs_insert_file_extent(trans, root, objectid, 0, 0, 0,
6686                                        group->key.offset, 0, group->key.offset,
6687                                        0, 0, 0);
6688         BUG_ON(err);
6689
6690         inode = btrfs_iget_locked(root->fs_info->sb, objectid, root);
6691         if (inode->i_state & I_NEW) {
6692                 BTRFS_I(inode)->root = root;
6693                 BTRFS_I(inode)->location.objectid = objectid;
6694                 BTRFS_I(inode)->location.type = BTRFS_INODE_ITEM_KEY;
6695                 BTRFS_I(inode)->location.offset = 0;
6696                 btrfs_read_locked_inode(inode);
6697                 unlock_new_inode(inode);
6698                 BUG_ON(is_bad_inode(inode));
6699         } else {
6700                 BUG_ON(1);
6701         }
6702         BTRFS_I(inode)->index_cnt = group->key.objectid;
6703
6704         err = btrfs_orphan_add(trans, inode);
6705 out:
6706         btrfs_end_transaction(trans, root);
6707         if (err) {
6708                 if (inode)
6709                         iput(inode);
6710                 inode = ERR_PTR(err);
6711         }
6712         return inode;
6713 }
6714
6715 int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len)
6716 {
6717
6718         struct btrfs_ordered_sum *sums;
6719         struct btrfs_sector_sum *sector_sum;
6720         struct btrfs_ordered_extent *ordered;
6721         struct btrfs_root *root = BTRFS_I(inode)->root;
6722         struct list_head list;
6723         size_t offset;
6724         int ret;
6725         u64 disk_bytenr;
6726
6727         INIT_LIST_HEAD(&list);
6728
6729         ordered = btrfs_lookup_ordered_extent(inode, file_pos);
6730         BUG_ON(ordered->file_offset != file_pos || ordered->len != len);
6731
6732         disk_bytenr = file_pos + BTRFS_I(inode)->index_cnt;
6733         ret = btrfs_lookup_csums_range(root->fs_info->csum_root, disk_bytenr,
6734                                        disk_bytenr + len - 1, &list);
6735
6736         while (!list_empty(&list)) {
6737                 sums = list_entry(list.next, struct btrfs_ordered_sum, list);
6738                 list_del_init(&sums->list);
6739
6740                 sector_sum = sums->sums;
6741                 sums->bytenr = ordered->start;
6742
6743                 offset = 0;
6744                 while (offset < sums->len) {
6745                         sector_sum->bytenr += ordered->start - disk_bytenr;
6746                         sector_sum++;
6747                         offset += root->sectorsize;
6748                 }
6749
6750                 btrfs_add_ordered_sum(inode, ordered, sums);
6751         }
6752         btrfs_put_ordered_extent(ordered);
6753         return 0;
6754 }
6755
6756 int btrfs_relocate_block_group(struct btrfs_root *root, u64 group_start)
6757 {
6758         struct btrfs_trans_handle *trans;
6759         struct btrfs_path *path;
6760         struct btrfs_fs_info *info = root->fs_info;
6761         struct extent_buffer *leaf;
6762         struct inode *reloc_inode;
6763         struct btrfs_block_group_cache *block_group;
6764         struct btrfs_key key;
6765         u64 skipped;
6766         u64 cur_byte;
6767         u64 total_found;
6768         u32 nritems;
6769         int ret;
6770         int progress;
6771         int pass = 0;
6772
6773         root = root->fs_info->extent_root;
6774
6775         block_group = btrfs_lookup_block_group(info, group_start);
6776         BUG_ON(!block_group);
6777
6778         printk(KERN_INFO "btrfs relocating block group %llu flags %llu\n",
6779                (unsigned long long)block_group->key.objectid,
6780                (unsigned long long)block_group->flags);
6781
6782         path = btrfs_alloc_path();
6783         BUG_ON(!path);
6784
6785         reloc_inode = create_reloc_inode(info, block_group);
6786         BUG_ON(IS_ERR(reloc_inode));
6787
6788         __alloc_chunk_for_shrink(root, block_group, 1);
6789         set_block_group_readonly(block_group);
6790
6791         btrfs_start_delalloc_inodes(info->tree_root);
6792         btrfs_wait_ordered_extents(info->tree_root, 0);
6793 again:
6794         skipped = 0;
6795         total_found = 0;
6796         progress = 0;
6797         key.objectid = block_group->key.objectid;
6798         key.offset = 0;
6799         key.type = 0;
6800         cur_byte = key.objectid;
6801
6802         trans = btrfs_start_transaction(info->tree_root, 1);
6803         btrfs_commit_transaction(trans, info->tree_root);
6804
6805         mutex_lock(&root->fs_info->cleaner_mutex);
6806         btrfs_clean_old_snapshots(info->tree_root);
6807         btrfs_remove_leaf_refs(info->tree_root, (u64)-1, 1);
6808         mutex_unlock(&root->fs_info->cleaner_mutex);
6809
6810         trans = btrfs_start_transaction(info->tree_root, 1);
6811         btrfs_commit_transaction(trans, info->tree_root);
6812
6813         while (1) {
6814                 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
6815                 if (ret < 0)
6816                         goto out;
6817 next:
6818                 leaf = path->nodes[0];
6819                 nritems = btrfs_header_nritems(leaf);
6820                 if (path->slots[0] >= nritems) {
6821                         ret = btrfs_next_leaf(root, path);
6822                         if (ret < 0)
6823                                 goto out;
6824                         if (ret == 1) {
6825                                 ret = 0;
6826                                 break;
6827                         }
6828                         leaf = path->nodes[0];
6829                         nritems = btrfs_header_nritems(leaf);
6830                 }
6831
6832                 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
6833
6834                 if (key.objectid >= block_group->key.objectid +
6835                     block_group->key.offset)
6836                         break;
6837
6838                 if (progress && need_resched()) {
6839                         btrfs_release_path(root, path);
6840                         cond_resched();
6841                         progress = 0;
6842                         continue;
6843                 }
6844                 progress = 1;
6845
6846                 if (btrfs_key_type(&key) != BTRFS_EXTENT_ITEM_KEY ||
6847                     key.objectid + key.offset <= cur_byte) {
6848                         path->slots[0]++;
6849                         goto next;
6850                 }
6851
6852                 total_found++;
6853                 cur_byte = key.objectid + key.offset;
6854                 btrfs_release_path(root, path);
6855
6856                 __alloc_chunk_for_shrink(root, block_group, 0);
6857                 ret = relocate_one_extent(root, path, &key, block_group,
6858                                           reloc_inode, pass);
6859                 BUG_ON(ret < 0);
6860                 if (ret > 0)
6861                         skipped++;
6862
6863                 key.objectid = cur_byte;
6864                 key.type = 0;
6865                 key.offset = 0;
6866         }
6867
6868         btrfs_release_path(root, path);
6869
6870         if (pass == 0) {
6871                 btrfs_wait_ordered_range(reloc_inode, 0, (u64)-1);
6872                 invalidate_mapping_pages(reloc_inode->i_mapping, 0, -1);
6873         }
6874
6875         if (total_found > 0) {
6876                 printk(KERN_INFO "btrfs found %llu extents in pass %d\n",
6877                        (unsigned long long)total_found, pass);
6878                 pass++;
6879                 if (total_found == skipped && pass > 2) {
6880                         iput(reloc_inode);
6881                         reloc_inode = create_reloc_inode(info, block_group);
6882                         pass = 0;
6883                 }
6884                 goto again;
6885         }
6886
6887         /* delete reloc_inode */
6888         iput(reloc_inode);
6889
6890         /* unpin extents in this range */
6891         trans = btrfs_start_transaction(info->tree_root, 1);
6892         btrfs_commit_transaction(trans, info->tree_root);
6893
6894         spin_lock(&block_group->lock);
6895         WARN_ON(block_group->pinned > 0);
6896         WARN_ON(block_group->reserved > 0);
6897         WARN_ON(btrfs_block_group_used(&block_group->item) > 0);
6898         spin_unlock(&block_group->lock);
6899         btrfs_put_block_group(block_group);
6900         ret = 0;
6901 out:
6902         btrfs_free_path(path);
6903         return ret;
6904 }
6905 #endif
6906
6907 static int find_first_block_group(struct btrfs_root *root,
6908                 struct btrfs_path *path, struct btrfs_key *key)
6909 {
6910         int ret = 0;
6911         struct btrfs_key found_key;
6912         struct extent_buffer *leaf;
6913         int slot;
6914
6915         ret = btrfs_search_slot(NULL, root, key, path, 0, 0);
6916         if (ret < 0)
6917                 goto out;
6918
6919         while (1) {
6920                 slot = path->slots[0];
6921                 leaf = path->nodes[0];
6922                 if (slot >= btrfs_header_nritems(leaf)) {
6923                         ret = btrfs_next_leaf(root, path);
6924                         if (ret == 0)
6925                                 continue;
6926                         if (ret < 0)
6927                                 goto out;
6928                         break;
6929                 }
6930                 btrfs_item_key_to_cpu(leaf, &found_key, slot);
6931
6932                 if (found_key.objectid >= key->objectid &&
6933                     found_key.type == BTRFS_BLOCK_GROUP_ITEM_KEY) {
6934                         ret = 0;
6935                         goto out;
6936                 }
6937                 path->slots[0]++;
6938         }
6939         ret = -ENOENT;
6940 out:
6941         return ret;
6942 }
6943
6944 int btrfs_free_block_groups(struct btrfs_fs_info *info)
6945 {
6946         struct btrfs_block_group_cache *block_group;
6947         struct btrfs_space_info *space_info;
6948         struct rb_node *n;
6949
6950         spin_lock(&info->block_group_cache_lock);
6951         while ((n = rb_last(&info->block_group_cache_tree)) != NULL) {
6952                 block_group = rb_entry(n, struct btrfs_block_group_cache,
6953                                        cache_node);
6954                 rb_erase(&block_group->cache_node,
6955                          &info->block_group_cache_tree);
6956                 spin_unlock(&info->block_group_cache_lock);
6957
6958                 btrfs_remove_free_space_cache(block_group);
6959                 down_write(&block_group->space_info->groups_sem);
6960                 list_del(&block_group->list);
6961                 up_write(&block_group->space_info->groups_sem);
6962
6963                 WARN_ON(atomic_read(&block_group->count) != 1);
6964                 kfree(block_group);
6965
6966                 spin_lock(&info->block_group_cache_lock);
6967         }
6968         spin_unlock(&info->block_group_cache_lock);
6969
6970         /* now that all the block groups are freed, go through and
6971          * free all the space_info structs.  This is only called during
6972          * the final stages of unmount, and so we know nobody is
6973          * using them.  We call synchronize_rcu() once before we start,
6974          * just to be on the safe side.
6975          */
6976         synchronize_rcu();
6977
6978         while(!list_empty(&info->space_info)) {
6979                 space_info = list_entry(info->space_info.next,
6980                                         struct btrfs_space_info,
6981                                         list);
6982
6983                 list_del(&space_info->list);
6984                 kfree(space_info);
6985         }
6986         return 0;
6987 }
6988
6989 int btrfs_read_block_groups(struct btrfs_root *root)
6990 {
6991         struct btrfs_path *path;
6992         int ret;
6993         struct btrfs_block_group_cache *cache;
6994         struct btrfs_fs_info *info = root->fs_info;
6995         struct btrfs_space_info *space_info;
6996         struct btrfs_key key;
6997         struct btrfs_key found_key;
6998         struct extent_buffer *leaf;
6999
7000         root = info->extent_root;
7001         key.objectid = 0;
7002         key.offset = 0;
7003         btrfs_set_key_type(&key, BTRFS_BLOCK_GROUP_ITEM_KEY);
7004         path = btrfs_alloc_path();
7005         if (!path)
7006                 return -ENOMEM;
7007
7008         while (1) {
7009                 ret = find_first_block_group(root, path, &key);
7010                 if (ret > 0) {
7011                         ret = 0;
7012                         goto error;
7013                 }
7014                 if (ret != 0)
7015                         goto error;
7016
7017                 leaf = path->nodes[0];
7018                 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
7019                 cache = kzalloc(sizeof(*cache), GFP_NOFS);
7020                 if (!cache) {
7021                         ret = -ENOMEM;
7022                         break;
7023                 }
7024
7025                 atomic_set(&cache->count, 1);
7026                 spin_lock_init(&cache->lock);
7027                 spin_lock_init(&cache->tree_lock);
7028                 mutex_init(&cache->cache_mutex);
7029                 INIT_LIST_HEAD(&cache->list);
7030                 INIT_LIST_HEAD(&cache->cluster_list);
7031                 read_extent_buffer(leaf, &cache->item,
7032                                    btrfs_item_ptr_offset(leaf, path->slots[0]),
7033                                    sizeof(cache->item));
7034                 memcpy(&cache->key, &found_key, sizeof(found_key));
7035
7036                 key.objectid = found_key.objectid + found_key.offset;
7037                 btrfs_release_path(root, path);
7038                 cache->flags = btrfs_block_group_flags(&cache->item);
7039
7040                 ret = update_space_info(info, cache->flags, found_key.offset,
7041                                         btrfs_block_group_used(&cache->item),
7042                                         &space_info);
7043                 BUG_ON(ret);
7044                 cache->space_info = space_info;
7045                 down_write(&space_info->groups_sem);
7046                 list_add_tail(&cache->list, &space_info->block_groups);
7047                 up_write(&space_info->groups_sem);
7048
7049                 ret = btrfs_add_block_group_cache(root->fs_info, cache);
7050                 BUG_ON(ret);
7051
7052                 set_avail_alloc_bits(root->fs_info, cache->flags);
7053                 if (btrfs_chunk_readonly(root, cache->key.objectid))
7054                         set_block_group_readonly(cache);
7055         }
7056         ret = 0;
7057 error:
7058         btrfs_free_path(path);
7059         return ret;
7060 }
7061
7062 int btrfs_make_block_group(struct btrfs_trans_handle *trans,
7063                            struct btrfs_root *root, u64 bytes_used,
7064                            u64 type, u64 chunk_objectid, u64 chunk_offset,
7065                            u64 size)
7066 {
7067         int ret;
7068         struct btrfs_root *extent_root;
7069         struct btrfs_block_group_cache *cache;
7070
7071         extent_root = root->fs_info->extent_root;
7072
7073         root->fs_info->last_trans_log_full_commit = trans->transid;
7074
7075         cache = kzalloc(sizeof(*cache), GFP_NOFS);
7076         if (!cache)
7077                 return -ENOMEM;
7078
7079         cache->key.objectid = chunk_offset;
7080         cache->key.offset = size;
7081         cache->key.type = BTRFS_BLOCK_GROUP_ITEM_KEY;
7082         atomic_set(&cache->count, 1);
7083         spin_lock_init(&cache->lock);
7084         spin_lock_init(&cache->tree_lock);
7085         mutex_init(&cache->cache_mutex);
7086         INIT_LIST_HEAD(&cache->list);
7087         INIT_LIST_HEAD(&cache->cluster_list);
7088
7089         btrfs_set_block_group_used(&cache->item, bytes_used);
7090         btrfs_set_block_group_chunk_objectid(&cache->item, chunk_objectid);
7091         cache->flags = type;
7092         btrfs_set_block_group_flags(&cache->item, type);
7093
7094         ret = update_space_info(root->fs_info, cache->flags, size, bytes_used,
7095                                 &cache->space_info);
7096         BUG_ON(ret);
7097         down_write(&cache->space_info->groups_sem);
7098         list_add_tail(&cache->list, &cache->space_info->block_groups);
7099         up_write(&cache->space_info->groups_sem);
7100
7101         ret = btrfs_add_block_group_cache(root->fs_info, cache);
7102         BUG_ON(ret);
7103
7104         ret = btrfs_insert_item(trans, extent_root, &cache->key, &cache->item,
7105                                 sizeof(cache->item));
7106         BUG_ON(ret);
7107
7108         set_avail_alloc_bits(extent_root->fs_info, type);
7109
7110         return 0;
7111 }
7112
7113 int btrfs_remove_block_group(struct btrfs_trans_handle *trans,
7114                              struct btrfs_root *root, u64 group_start)
7115 {
7116         struct btrfs_path *path;
7117         struct btrfs_block_group_cache *block_group;
7118         struct btrfs_free_cluster *cluster;
7119         struct btrfs_key key;
7120         int ret;
7121
7122         root = root->fs_info->extent_root;
7123
7124         block_group = btrfs_lookup_block_group(root->fs_info, group_start);
7125         BUG_ON(!block_group);
7126         BUG_ON(!block_group->ro);
7127
7128         memcpy(&key, &block_group->key, sizeof(key));
7129
7130         /* make sure this block group isn't part of an allocation cluster */
7131         cluster = &root->fs_info->data_alloc_cluster;
7132         spin_lock(&cluster->refill_lock);
7133         btrfs_return_cluster_to_free_space(block_group, cluster);
7134         spin_unlock(&cluster->refill_lock);
7135
7136         /*
7137          * make sure this block group isn't part of a metadata
7138          * allocation cluster
7139          */
7140         cluster = &root->fs_info->meta_alloc_cluster;
7141         spin_lock(&cluster->refill_lock);
7142         btrfs_return_cluster_to_free_space(block_group, cluster);
7143         spin_unlock(&cluster->refill_lock);
7144
7145         path = btrfs_alloc_path();
7146         BUG_ON(!path);
7147
7148         spin_lock(&root->fs_info->block_group_cache_lock);
7149         rb_erase(&block_group->cache_node,
7150                  &root->fs_info->block_group_cache_tree);
7151         spin_unlock(&root->fs_info->block_group_cache_lock);
7152         btrfs_remove_free_space_cache(block_group);
7153         down_write(&block_group->space_info->groups_sem);
7154         /*
7155          * we must use list_del_init so people can check to see if they
7156          * are still on the list after taking the semaphore
7157          */
7158         list_del_init(&block_group->list);
7159         up_write(&block_group->space_info->groups_sem);
7160
7161         spin_lock(&block_group->space_info->lock);
7162         block_group->space_info->total_bytes -= block_group->key.offset;
7163         block_group->space_info->bytes_readonly -= block_group->key.offset;
7164         spin_unlock(&block_group->space_info->lock);
7165         block_group->space_info->full = 0;
7166
7167         btrfs_put_block_group(block_group);
7168         btrfs_put_block_group(block_group);
7169
7170         ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
7171         if (ret > 0)
7172                 ret = -EIO;
7173         if (ret < 0)
7174                 goto out;
7175
7176         ret = btrfs_del_item(trans, root, path);
7177 out:
7178         btrfs_free_path(path);
7179         return ret;
7180 }