fs/btrfs/extent-tree.c

   1 // SPDX-License-Identifier: GPL-2.0
   2 /*
   3  * Copyright (C) 2007 Oracle.  All rights reserved.
   4  */
   5
   6 #include <linux/sched.h>
   7 #include <linux/sched/signal.h>
   8 #include <linux/pagemap.h>
   9 #include <linux/writeback.h>
  10 #include <linux/blkdev.h>
  11 #include <linux/sort.h>
  12 #include <linux/rcupdate.h>
  13 #include <linux/kthread.h>
  14 #include <linux/slab.h>
  15 #include <linux/ratelimit.h>
  16 #include <linux/percpu_counter.h>
  17 #include <linux/lockdep.h>
  18 #include <linux/crc32c.h>
  19 #include "misc.h"
  20 #include "tree-log.h"
  21 #include "disk-io.h"
  22 #include "print-tree.h"
  23 #include "volumes.h"
  24 #include "raid56.h"
  25 #include "locking.h"
  26 #include "free-space-cache.h"
  27 #include "free-space-tree.h"
  28 #include "sysfs.h"
  29 #include "qgroup.h"
  30 #include "ref-verify.h"
  31 #include "space-info.h"
  32 #include "block-rsv.h"
  33 #include "delalloc-space.h"
  34 #include "block-group.h"
  35 #include "discard.h"
  36 #include "rcu-string.h"
  37 #include "zoned.h"
  38 #include "dev-replace.h"
  39
  40 #undef SCRAMBLE_DELAYED_REFS
  41
  42
  43 static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
  44                                struct btrfs_delayed_ref_node *node, u64 parent,
  45                                u64 root_objectid, u64 owner_objectid,
  46                                u64 owner_offset, int refs_to_drop,
  47                                struct btrfs_delayed_extent_op *extra_op);
  48 static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op,
  49                                     struct extent_buffer *leaf,
  50                                     struct btrfs_extent_item *ei);
  51 static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
  52                                       u64 parent, u64 root_objectid,
  53                                       u64 flags, u64 owner, u64 offset,
  54                                       struct btrfs_key *ins, int ref_mod);
  55 static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
  56                                      struct btrfs_delayed_ref_node *node,
  57                                      struct btrfs_delayed_extent_op *extent_op);
  58 static int find_next_key(struct btrfs_path *path, int level,
  59                          struct btrfs_key *key);
  60
  61 static int block_group_bits(struct btrfs_block_group *cache, u64 bits)
  62 {
  63         return (cache->flags & bits) == bits;
  64 }
  65
  66 int btrfs_add_excluded_extent(struct btrfs_fs_info *fs_info,
  67                               u64 start, u64 num_bytes)
  68 {
  69         u64 end = start + num_bytes - 1;
  70         set_extent_bits(&fs_info->excluded_extents, start, end,
  71                         EXTENT_UPTODATE);
  72         return 0;
  73 }
  74
  75 void btrfs_free_excluded_extents(struct btrfs_block_group *cache)
  76 {
  77         struct btrfs_fs_info *fs_info = cache->fs_info;
  78         u64 start, end;
  79
  80         start = cache->start;
  81         end = start + cache->length - 1;
  82
  83         clear_extent_bits(&fs_info->excluded_extents, start, end,
  84                           EXTENT_UPTODATE);
  85 }
  86
  87 /* simple helper to search for an existing data extent at a given offset */
  88 int btrfs_lookup_data_extent(struct btrfs_fs_info *fs_info, u64 start, u64 len)
  89 {
  90         struct btrfs_root *root = btrfs_extent_root(fs_info, start);
  91         int ret;
  92         struct btrfs_key key;
  93         struct btrfs_path *path;
  94
  95         path = btrfs_alloc_path();
  96         if (!path)
  97                 return -ENOMEM;
  98
  99         key.objectid = start;
 100         key.offset = len;
 101         key.type = BTRFS_EXTENT_ITEM_KEY;
 102         ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
 103         btrfs_free_path(path);
 104         return ret;
 105 }
 106
 107 /*
 108  * helper function to lookup reference count and flags of a tree block.
 109  *
 110  * the head node for delayed ref is used to store the sum of all the
 111  * reference count modifications queued up in the rbtree. the head
 112  * node may also store the extent flags to set. This way you can check
 113  * to see what the reference count and extent flags would be if all of
 114  * the delayed refs are not processed.
 115  */
 116 int btrfs_lookup_extent_info(struct btrfs_trans_handle *trans,
 117                              struct btrfs_fs_info *fs_info, u64 bytenr,
 118                              u64 offset, int metadata, u64 *refs, u64 *flags)
 119 {
 120         struct btrfs_root *extent_root;
 121         struct btrfs_delayed_ref_head *head;
 122         struct btrfs_delayed_ref_root *delayed_refs;
 123         struct btrfs_path *path;
 124         struct btrfs_extent_item *ei;
 125         struct extent_buffer *leaf;
 126         struct btrfs_key key;
 127         u32 item_size;
 128         u64 num_refs;
 129         u64 extent_flags;
 130         int ret;
 131
 132         /*
 133          * If we don't have skinny metadata, don't bother doing anything
 134          * different
 135          */
 136         if (metadata && !btrfs_fs_incompat(fs_info, SKINNY_METADATA)) {
 137                 offset = fs_info->nodesize;
 138                 metadata = 0;
 139         }
 140
 141         path = btrfs_alloc_path();
 142         if (!path)
 143                 return -ENOMEM;
 144
 145         if (!trans) {
 146                 path->skip_locking = 1;
 147                 path->search_commit_root = 1;
 148         }
 149
 150 search_again:
 151         key.objectid = bytenr;
 152         key.offset = offset;
 153         if (metadata)
 154                 key.type = BTRFS_METADATA_ITEM_KEY;
 155         else
 156                 key.type = BTRFS_EXTENT_ITEM_KEY;
 157
 158         extent_root = btrfs_extent_root(fs_info, bytenr);
 159         ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
 160         if (ret < 0)
 161                 goto out_free;
 162
 163         if (ret > 0 && metadata && key.type == BTRFS_METADATA_ITEM_KEY) {
 164                 if (path->slots[0]) {
 165                         path->slots[0]--;
 166                         btrfs_item_key_to_cpu(path->nodes[0], &key,
 167                                               path->slots[0]);
 168                         if (key.objectid == bytenr &&
 169                             key.type == BTRFS_EXTENT_ITEM_KEY &&
 170                             key.offset == fs_info->nodesize)
 171                                 ret = 0;
 172                 }
 173         }
 174
 175         if (ret == 0) {
 176                 leaf = path->nodes[0];
 177                 item_size = btrfs_item_size(leaf, path->slots[0]);
 178                 if (item_size >= sizeof(*ei)) {
 179                         ei = btrfs_item_ptr(leaf, path->slots[0],
 180                                             struct btrfs_extent_item);
 181                         num_refs = btrfs_extent_refs(leaf, ei);
 182                         extent_flags = btrfs_extent_flags(leaf, ei);
 183                 } else {
 184                         ret = -EINVAL;
 185                         btrfs_print_v0_err(fs_info);
 186                         if (trans)
 187                                 btrfs_abort_transaction(trans, ret);
 188                         else
 189                                 btrfs_handle_fs_error(fs_info, ret, NULL);
 190
 191                         goto out_free;
 192                 }
 193
 194                 BUG_ON(num_refs == 0);
 195         } else {
 196                 num_refs = 0;
 197                 extent_flags = 0;
 198                 ret = 0;
 199         }
 200
 201         if (!trans)
 202                 goto out;
 203
 204         delayed_refs = &trans->transaction->delayed_refs;
 205         spin_lock(&delayed_refs->lock);
 206         head = btrfs_find_delayed_ref_head(delayed_refs, bytenr);
 207         if (head) {
 208                 if (!mutex_trylock(&head->mutex)) {
 209                         refcount_inc(&head->refs);
 210                         spin_unlock(&delayed_refs->lock);
 211
 212                         btrfs_release_path(path);
 213
 214                         /*
 215                          * Mutex was contended, block until it's released and try
 216                          * again
 217                          */
 218                         mutex_lock(&head->mutex);
 219                         mutex_unlock(&head->mutex);
 220                         btrfs_put_delayed_ref_head(head);
 221                         goto search_again;
 222                 }
 223                 spin_lock(&head->lock);
 224                 if (head->extent_op && head->extent_op->update_flags)
 225                         extent_flags |= head->extent_op->flags_to_set;
 226                 else
 227                         BUG_ON(num_refs == 0);
 228
 229                 num_refs += head->ref_mod;
 230                 spin_unlock(&head->lock);
 231                 mutex_unlock(&head->mutex);
 232         }
 233         spin_unlock(&delayed_refs->lock);
 234 out:
 235         WARN_ON(num_refs == 0);
 236         if (refs)
 237                 *refs = num_refs;
 238         if (flags)
 239                 *flags = extent_flags;
 240 out_free:
 241         btrfs_free_path(path);
 242         return ret;
 243 }
 244
 245 /*
 246  * Back reference rules.  Back refs have three main goals:
 247  *
 248  * 1) differentiate between all holders of references to an extent so that
 249  *    when a reference is dropped we can make sure it was a valid reference
 250  *    before freeing the extent.
 251  *
 252  * 2) Provide enough information to quickly find the holders of an extent
 253  *    if we notice a given block is corrupted or bad.
 254  *
 255  * 3) Make it easy to migrate blocks for FS shrinking or storage pool
 256  *    maintenance.  This is actually the same as #2, but with a slightly
 257  *    different use case.
 258  *
 259  * There are two kinds of back refs. The implicit back refs is optimized
 260  * for pointers in non-shared tree blocks. For a given pointer in a block,
 261  * back refs of this kind provide information about the block's owner tree
 262  * and the pointer's key. These information allow us to find the block by
 263  * b-tree searching. The full back refs is for pointers in tree blocks not
 264  * referenced by their owner trees. The location of tree block is recorded
 265  * in the back refs. Actually the full back refs is generic, and can be
 266  * used in all cases the implicit back refs is used. The major shortcoming
 267  * of the full back refs is its overhead. Every time a tree block gets
 268  * COWed, we have to update back refs entry for all pointers in it.
 269  *
 270  * For a newly allocated tree block, we use implicit back refs for
 271  * pointers in it. This means most tree related operations only involve
 272  * implicit back refs. For a tree block created in old transaction, the
 273  * only way to drop a reference to it is COW it. So we can detect the
 274  * event that tree block loses its owner tree's reference and do the
 275  * back refs conversion.
 276  *
 277  * When a tree block is COWed through a tree, there are four cases:
 278  *
 279  * The reference count of the block is one and the tree is the block's
 280  * owner tree. Nothing to do in this case.
 281  *
 282  * The reference count of the block is one and the tree is not the
 283  * block's owner tree. In this case, full back refs is used for pointers
 284  * in the block. Remove these full back refs, add implicit back refs for
 285  * every pointers in the new block.
 286  *
 287  * The reference count of the block is greater than one and the tree is
 288  * the block's owner tree. In this case, implicit back refs is used for
 289  * pointers in the block. Add full back refs for every pointers in the
 290  * block, increase lower level extents' reference counts. The original
 291  * implicit back refs are entailed to the new block.
 292  *
 293  * The reference count of the block is greater than one and the tree is
 294  * not the block's owner tree. Add implicit back refs for every pointer in
 295  * the new block, increase lower level extents' reference count.
 296  *
 297  * Back Reference Key composing:
 298  *
 299  * The key objectid corresponds to the first byte in the extent,
 300  * The key type is used to differentiate between types of back refs.
 301  * There are different meanings of the key offset for different types
 302  * of back refs.
 303  *
 304  * File extents can be referenced by:
 305  *
 306  * - multiple snapshots, subvolumes, or different generations in one subvol
 307  * - different files inside a single subvolume
 308  * - different offsets inside a file (bookend extents in file.c)
 309  *
 310  * The extent ref structure for the implicit back refs has fields for:
 311  *
 312  * - Objectid of the subvolume root
 313  * - objectid of the file holding the reference
 314  * - original offset in the file
 315  * - how many bookend extents
 316  *
 317  * The key offset for the implicit back refs is hash of the first
 318  * three fields.
 319  *
 320  * The extent ref structure for the full back refs has field for:
 321  *
 322  * - number of pointers in the tree leaf
 323  *
 324  * The key offset for the implicit back refs is the first byte of
 325  * the tree leaf
 326  *
 327  * When a file extent is allocated, The implicit back refs is used.
 328  * the fields are filled in:
 329  *
 330  *     (root_key.objectid, inode objectid, offset in file, 1)
 331  *
 332  * When a file extent is removed file truncation, we find the
 333  * corresponding implicit back refs and check the following fields:
 334  *
 335  *     (btrfs_header_owner(leaf), inode objectid, offset in file)
 336  *
 337  * Btree extents can be referenced by:
 338  *
 339  * - Different subvolumes
 340  *
 341  * Both the implicit back refs and the full back refs for tree blocks
 342  * only consist of key. The key offset for the implicit back refs is
 343  * objectid of block's owner tree. The key offset for the full back refs
 344  * is the first byte of parent block.
 345  *
 346  * When implicit back refs is used, information about the lowest key and
 347  * level of the tree block are required. These information are stored in
 348  * tree block info structure.
 349  */
 350
 351 /*
 352  * is_data == BTRFS_REF_TYPE_BLOCK, tree block type is required,
 353  * is_data == BTRFS_REF_TYPE_DATA, data type is requiried,
 354  * is_data == BTRFS_REF_TYPE_ANY, either type is OK.
 355  */
 356 int btrfs_get_extent_inline_ref_type(const struct extent_buffer *eb,
 357                                      struct btrfs_extent_inline_ref *iref,
 358                                      enum btrfs_inline_ref_type is_data)
 359 {
 360         int type = btrfs_extent_inline_ref_type(eb, iref);
 361         u64 offset = btrfs_extent_inline_ref_offset(eb, iref);
 362
 363         if (type == BTRFS_TREE_BLOCK_REF_KEY ||
 364             type == BTRFS_SHARED_BLOCK_REF_KEY ||
 365             type == BTRFS_SHARED_DATA_REF_KEY ||
 366             type == BTRFS_EXTENT_DATA_REF_KEY) {
 367                 if (is_data == BTRFS_REF_TYPE_BLOCK) {
 368                         if (type == BTRFS_TREE_BLOCK_REF_KEY)
 369                                 return type;
 370                         if (type == BTRFS_SHARED_BLOCK_REF_KEY) {
 371                                 ASSERT(eb->fs_info);
 372                                 /*
 373                                  * Every shared one has parent tree block,
 374                                  * which must be aligned to sector size.
 375                                  */
 376                                 if (offset &&
 377                                     IS_ALIGNED(offset, eb->fs_info->sectorsize))
 378                                         return type;
 379                         }
 380                 } else if (is_data == BTRFS_REF_TYPE_DATA) {
 381                         if (type == BTRFS_EXTENT_DATA_REF_KEY)
 382                                 return type;
 383                         if (type == BTRFS_SHARED_DATA_REF_KEY) {
 384                                 ASSERT(eb->fs_info);
 385                                 /*
 386                                  * Every shared one has parent tree block,
 387                                  * which must be aligned to sector size.
 388                                  */
 389                                 if (offset &&
 390                                     IS_ALIGNED(offset, eb->fs_info->sectorsize))
 391                                         return type;
 392                         }
 393                 } else {
 394                         ASSERT(is_data == BTRFS_REF_TYPE_ANY);
 395                         return type;
 396                 }
 397         }
 398
 399         btrfs_print_leaf((struct extent_buffer *)eb);
 400         btrfs_err(eb->fs_info,
 401                   "eb %llu iref 0x%lx invalid extent inline ref type %d",
 402                   eb->start, (unsigned long)iref, type);
 403         WARN_ON(1);
 404
 405         return BTRFS_REF_TYPE_INVALID;
 406 }
 407
 408 u64 hash_extent_data_ref(u64 root_objectid, u64 owner, u64 offset)
 409 {
 410         u32 high_crc = ~(u32)0;
 411         u32 low_crc = ~(u32)0;
 412         __le64 lenum;
 413
 414         lenum = cpu_to_le64(root_objectid);
 415         high_crc = btrfs_crc32c(high_crc, &lenum, sizeof(lenum));
 416         lenum = cpu_to_le64(owner);
 417         low_crc = btrfs_crc32c(low_crc, &lenum, sizeof(lenum));
 418         lenum = cpu_to_le64(offset);
 419         low_crc = btrfs_crc32c(low_crc, &lenum, sizeof(lenum));
 420
 421         return ((u64)high_crc << 31) ^ (u64)low_crc;
 422 }
 423
 424 static u64 hash_extent_data_ref_item(struct extent_buffer *leaf,
 425                                      struct btrfs_extent_data_ref *ref)
 426 {
 427         return hash_extent_data_ref(btrfs_extent_data_ref_root(leaf, ref),
 428                                     btrfs_extent_data_ref_objectid(leaf, ref),
 429                                     btrfs_extent_data_ref_offset(leaf, ref));
 430 }
 431
 432 static int match_extent_data_ref(struct extent_buffer *leaf,
 433                                  struct btrfs_extent_data_ref *ref,
 434                                  u64 root_objectid, u64 owner, u64 offset)
 435 {
 436         if (btrfs_extent_data_ref_root(leaf, ref) != root_objectid ||
 437             btrfs_extent_data_ref_objectid(leaf, ref) != owner ||
 438             btrfs_extent_data_ref_offset(leaf, ref) != offset)
 439                 return 0;
 440         return 1;
 441 }
 442
 443 static noinline int lookup_extent_data_ref(struct btrfs_trans_handle *trans,
 444                                            struct btrfs_path *path,
 445                                            u64 bytenr, u64 parent,
 446                                            u64 root_objectid,
 447                                            u64 owner, u64 offset)
 448 {
 449         struct btrfs_root *root = btrfs_extent_root(trans->fs_info, bytenr);
 450         struct btrfs_key key;
 451         struct btrfs_extent_data_ref *ref;
 452         struct extent_buffer *leaf;
 453         u32 nritems;
 454         int ret;
 455         int recow;
 456         int err = -ENOENT;
 457
 458         key.objectid = bytenr;
 459         if (parent) {
 460                 key.type = BTRFS_SHARED_DATA_REF_KEY;
 461                 key.offset = parent;
 462         } else {
 463                 key.type = BTRFS_EXTENT_DATA_REF_KEY;
 464                 key.offset = hash_extent_data_ref(root_objectid,
 465                                                   owner, offset);
 466         }
 467 again:
 468         recow = 0;
 469         ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 470         if (ret < 0) {
 471                 err = ret;
 472                 goto fail;
 473         }
 474
 475         if (parent) {
 476                 if (!ret)
 477                         return 0;
 478                 goto fail;
 479         }
 480
 481         leaf = path->nodes[0];
 482         nritems = btrfs_header_nritems(leaf);
 483         while (1) {
 484                 if (path->slots[0] >= nritems) {
 485                         ret = btrfs_next_leaf(root, path);
 486                         if (ret < 0)
 487                                 err = ret;
 488                         if (ret)
 489                                 goto fail;
 490
 491                         leaf = path->nodes[0];
 492                         nritems = btrfs_header_nritems(leaf);
 493                         recow = 1;
 494                 }
 495
 496                 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
 497                 if (key.objectid != bytenr ||
 498                     key.type != BTRFS_EXTENT_DATA_REF_KEY)
 499                         goto fail;
 500
 501                 ref = btrfs_item_ptr(leaf, path->slots[0],
 502                                      struct btrfs_extent_data_ref);
 503
 504                 if (match_extent_data_ref(leaf, ref, root_objectid,
 505                                           owner, offset)) {
 506                         if (recow) {
 507                                 btrfs_release_path(path);
 508                                 goto again;
 509                         }
 510                         err = 0;
 511                         break;
 512                 }
 513                 path->slots[0]++;
 514         }
 515 fail:
 516         return err;
 517 }
 518
 519 static noinline int insert_extent_data_ref(struct btrfs_trans_handle *trans,
 520                                            struct btrfs_path *path,
 521                                            u64 bytenr, u64 parent,
 522                                            u64 root_objectid, u64 owner,
 523                                            u64 offset, int refs_to_add)
 524 {
 525         struct btrfs_root *root = btrfs_extent_root(trans->fs_info, bytenr);
 526         struct btrfs_key key;
 527         struct extent_buffer *leaf;
 528         u32 size;
 529         u32 num_refs;
 530         int ret;
 531
 532         key.objectid = bytenr;
 533         if (parent) {
 534                 key.type = BTRFS_SHARED_DATA_REF_KEY;
 535                 key.offset = parent;
 536                 size = sizeof(struct btrfs_shared_data_ref);
 537         } else {
 538                 key.type = BTRFS_EXTENT_DATA_REF_KEY;
 539                 key.offset = hash_extent_data_ref(root_objectid,
 540                                                   owner, offset);
 541                 size = sizeof(struct btrfs_extent_data_ref);
 542         }
 543
 544         ret = btrfs_insert_empty_item(trans, root, path, &key, size);
 545         if (ret && ret != -EEXIST)
 546                 goto fail;
 547
 548         leaf = path->nodes[0];
 549         if (parent) {
 550                 struct btrfs_shared_data_ref *ref;
 551                 ref = btrfs_item_ptr(leaf, path->slots[0],
 552                                      struct btrfs_shared_data_ref);
 553                 if (ret == 0) {
 554                         btrfs_set_shared_data_ref_count(leaf, ref, refs_to_add);
 555                 } else {
 556                         num_refs = btrfs_shared_data_ref_count(leaf, ref);
 557                         num_refs += refs_to_add;
 558                         btrfs_set_shared_data_ref_count(leaf, ref, num_refs);
 559                 }
 560         } else {
 561                 struct btrfs_extent_data_ref *ref;
 562                 while (ret == -EEXIST) {
 563                         ref = btrfs_item_ptr(leaf, path->slots[0],
 564                                              struct btrfs_extent_data_ref);
 565                         if (match_extent_data_ref(leaf, ref, root_objectid,
 566                                                   owner, offset))
 567                                 break;
 568                         btrfs_release_path(path);
 569                         key.offset++;
 570                         ret = btrfs_insert_empty_item(trans, root, path, &key,
 571                                                       size);
 572                         if (ret && ret != -EEXIST)
 573                                 goto fail;
 574
 575                         leaf = path->nodes[0];
 576                 }
 577                 ref = btrfs_item_ptr(leaf, path->slots[0],
 578                                      struct btrfs_extent_data_ref);
 579                 if (ret == 0) {
 580                         btrfs_set_extent_data_ref_root(leaf, ref,
 581                                                        root_objectid);
 582                         btrfs_set_extent_data_ref_objectid(leaf, ref, owner);
 583                         btrfs_set_extent_data_ref_offset(leaf, ref, offset);
 584                         btrfs_set_extent_data_ref_count(leaf, ref, refs_to_add);
 585                 } else {
 586                         num_refs = btrfs_extent_data_ref_count(leaf, ref);
 587                         num_refs += refs_to_add;
 588                         btrfs_set_extent_data_ref_count(leaf, ref, num_refs);
 589                 }
 590         }
 591         btrfs_mark_buffer_dirty(leaf);
 592         ret = 0;
 593 fail:
 594         btrfs_release_path(path);
 595         return ret;
 596 }
 597
 598 static noinline int remove_extent_data_ref(struct btrfs_trans_handle *trans,
 599                                            struct btrfs_root *root,
 600                                            struct btrfs_path *path,
 601                                            int refs_to_drop)
 602 {
 603         struct btrfs_key key;
 604         struct btrfs_extent_data_ref *ref1 = NULL;
 605         struct btrfs_shared_data_ref *ref2 = NULL;
 606         struct extent_buffer *leaf;
 607         u32 num_refs = 0;
 608         int ret = 0;
 609
 610         leaf = path->nodes[0];
 611         btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
 612
 613         if (key.type == BTRFS_EXTENT_DATA_REF_KEY) {
 614                 ref1 = btrfs_item_ptr(leaf, path->slots[0],
 615                                       struct btrfs_extent_data_ref);
 616                 num_refs = btrfs_extent_data_ref_count(leaf, ref1);
 617         } else if (key.type == BTRFS_SHARED_DATA_REF_KEY) {
 618                 ref2 = btrfs_item_ptr(leaf, path->slots[0],
 619                                       struct btrfs_shared_data_ref);
 620                 num_refs = btrfs_shared_data_ref_count(leaf, ref2);
 621         } else if (unlikely(key.type == BTRFS_EXTENT_REF_V0_KEY)) {
 622                 btrfs_print_v0_err(trans->fs_info);
 623                 btrfs_abort_transaction(trans, -EINVAL);
 624                 return -EINVAL;
 625         } else {
 626                 BUG();
 627         }
 628
 629         BUG_ON(num_refs < refs_to_drop);
 630         num_refs -= refs_to_drop;
 631
 632         if (num_refs == 0) {
 633                 ret = btrfs_del_item(trans, root, path);
 634         } else {
 635                 if (key.type == BTRFS_EXTENT_DATA_REF_KEY)
 636                         btrfs_set_extent_data_ref_count(leaf, ref1, num_refs);
 637                 else if (key.type == BTRFS_SHARED_DATA_REF_KEY)
 638                         btrfs_set_shared_data_ref_count(leaf, ref2, num_refs);
 639                 btrfs_mark_buffer_dirty(leaf);
 640         }
 641         return ret;
 642 }
 643
 644 static noinline u32 extent_data_ref_count(struct btrfs_path *path,
 645                                           struct btrfs_extent_inline_ref *iref)
 646 {
 647         struct btrfs_key key;
 648         struct extent_buffer *leaf;
 649         struct btrfs_extent_data_ref *ref1;
 650         struct btrfs_shared_data_ref *ref2;
 651         u32 num_refs = 0;
 652         int type;
 653
 654         leaf = path->nodes[0];
 655         btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
 656
 657         BUG_ON(key.type == BTRFS_EXTENT_REF_V0_KEY);
 658         if (iref) {
 659                 /*
 660                  * If type is invalid, we should have bailed out earlier than
 661                  * this call.
 662                  */
 663                 type = btrfs_get_extent_inline_ref_type(leaf, iref, BTRFS_REF_TYPE_DATA);
 664                 ASSERT(type != BTRFS_REF_TYPE_INVALID);
 665                 if (type == BTRFS_EXTENT_DATA_REF_KEY) {
 666                         ref1 = (struct btrfs_extent_data_ref *)(&iref->offset);
 667                         num_refs = btrfs_extent_data_ref_count(leaf, ref1);
 668                 } else {
 669                         ref2 = (struct btrfs_shared_data_ref *)(iref + 1);
 670                         num_refs = btrfs_shared_data_ref_count(leaf, ref2);
 671                 }
 672         } else if (key.type == BTRFS_EXTENT_DATA_REF_KEY) {
 673                 ref1 = btrfs_item_ptr(leaf, path->slots[0],
 674                                       struct btrfs_extent_data_ref);
 675                 num_refs = btrfs_extent_data_ref_count(leaf, ref1);
 676         } else if (key.type == BTRFS_SHARED_DATA_REF_KEY) {
 677                 ref2 = btrfs_item_ptr(leaf, path->slots[0],
 678                                       struct btrfs_shared_data_ref);
 679                 num_refs = btrfs_shared_data_ref_count(leaf, ref2);
 680         } else {
 681                 WARN_ON(1);
 682         }
 683         return num_refs;
 684 }
 685
 686 static noinline int lookup_tree_block_ref(struct btrfs_trans_handle *trans,
 687                                           struct btrfs_path *path,
 688                                           u64 bytenr, u64 parent,
 689                                           u64 root_objectid)
 690 {
 691         struct btrfs_root *root = btrfs_extent_root(trans->fs_info, bytenr);
 692         struct btrfs_key key;
 693         int ret;
 694
 695         key.objectid = bytenr;
 696         if (parent) {
 697                 key.type = BTRFS_SHARED_BLOCK_REF_KEY;
 698                 key.offset = parent;
 699         } else {
 700                 key.type = BTRFS_TREE_BLOCK_REF_KEY;
 701                 key.offset = root_objectid;
 702         }
 703
 704         ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
 705         if (ret > 0)
 706                 ret = -ENOENT;
 707         return ret;
 708 }
 709
 710 static noinline int insert_tree_block_ref(struct btrfs_trans_handle *trans,
 711                                           struct btrfs_path *path,
 712                                           u64 bytenr, u64 parent,
 713                                           u64 root_objectid)
 714 {
 715         struct btrfs_root *root = btrfs_extent_root(trans->fs_info, bytenr);
 716         struct btrfs_key key;
 717         int ret;
 718
 719         key.objectid = bytenr;
 720         if (parent) {
 721                 key.type = BTRFS_SHARED_BLOCK_REF_KEY;
 722                 key.offset = parent;
 723         } else {
 724                 key.type = BTRFS_TREE_BLOCK_REF_KEY;
 725                 key.offset = root_objectid;
 726         }
 727
 728         ret = btrfs_insert_empty_item(trans, root, path, &key, 0);
 729         btrfs_release_path(path);
 730         return ret;
 731 }
 732
 733 static inline int extent_ref_type(u64 parent, u64 owner)
 734 {
 735         int type;
 736         if (owner < BTRFS_FIRST_FREE_OBJECTID) {
 737                 if (parent > 0)
 738                         type = BTRFS_SHARED_BLOCK_REF_KEY;
 739                 else
 740                         type = BTRFS_TREE_BLOCK_REF_KEY;
 741         } else {
 742                 if (parent > 0)
 743                         type = BTRFS_SHARED_DATA_REF_KEY;
 744                 else
 745                         type = BTRFS_EXTENT_DATA_REF_KEY;
 746         }
 747         return type;
 748 }
 749
 750 static int find_next_key(struct btrfs_path *path, int level,
 751                          struct btrfs_key *key)
 752
 753 {
 754         for (; level < BTRFS_MAX_LEVEL; level++) {
 755                 if (!path->nodes[level])
 756                         break;
 757                 if (path->slots[level] + 1 >=
 758                     btrfs_header_nritems(path->nodes[level]))
 759                         continue;
 760                 if (level == 0)
 761                         btrfs_item_key_to_cpu(path->nodes[level], key,
 762                                               path->slots[level] + 1);
 763                 else
 764                         btrfs_node_key_to_cpu(path->nodes[level], key,
 765                                               path->slots[level] + 1);
 766                 return 0;
 767         }
 768         return 1;
 769 }
 770
 771 /*
 772  * look for inline back ref. if back ref is found, *ref_ret is set
 773  * to the address of inline back ref, and 0 is returned.
 774  *
 775  * if back ref isn't found, *ref_ret is set to the address where it
 776  * should be inserted, and -ENOENT is returned.
 777  *
 778  * if insert is true and there are too many inline back refs, the path
 779  * points to the extent item, and -EAGAIN is returned.
 780  *
 781  * NOTE: inline back refs are ordered in the same way that back ref
 782  *       items in the tree are ordered.
 783  */
 784 static noinline_for_stack
 785 int lookup_inline_extent_backref(struct btrfs_trans_handle *trans,
 786                                  struct btrfs_path *path,
 787                                  struct btrfs_extent_inline_ref **ref_ret,
 788                                  u64 bytenr, u64 num_bytes,
 789                                  u64 parent, u64 root_objectid,
 790                                  u64 owner, u64 offset, int insert)
 791 {
 792         struct btrfs_fs_info *fs_info = trans->fs_info;
 793         struct btrfs_root *root = btrfs_extent_root(fs_info, bytenr);
 794         struct btrfs_key key;
 795         struct extent_buffer *leaf;
 796         struct btrfs_extent_item *ei;
 797         struct btrfs_extent_inline_ref *iref;
 798         u64 flags;
 799         u64 item_size;
 800         unsigned long ptr;
 801         unsigned long end;
 802         int extra_size;
 803         int type;
 804         int want;
 805         int ret;
 806         int err = 0;
 807         bool skinny_metadata = btrfs_fs_incompat(fs_info, SKINNY_METADATA);
 808         int needed;
 809
 810         key.objectid = bytenr;
 811         key.type = BTRFS_EXTENT_ITEM_KEY;
 812         key.offset = num_bytes;
 813
 814         want = extent_ref_type(parent, owner);
 815         if (insert) {
 816                 extra_size = btrfs_extent_inline_ref_size(want);
 817                 path->search_for_extension = 1;
 818                 path->keep_locks = 1;
 819         } else
 820                 extra_size = -1;
 821
 822         /*
 823          * Owner is our level, so we can just add one to get the level for the
 824          * block we are interested in.
 825          */
 826         if (skinny_metadata && owner < BTRFS_FIRST_FREE_OBJECTID) {
 827                 key.type = BTRFS_METADATA_ITEM_KEY;
 828                 key.offset = owner;
 829         }
 830
 831 again:
 832         ret = btrfs_search_slot(trans, root, &key, path, extra_size, 1);
 833         if (ret < 0) {
 834                 err = ret;
 835                 goto out;
 836         }
 837
 838         /*
 839          * We may be a newly converted file system which still has the old fat
 840          * extent entries for metadata, so try and see if we have one of those.
 841          */
 842         if (ret > 0 && skinny_metadata) {
 843                 skinny_metadata = false;
 844                 if (path->slots[0]) {
 845                         path->slots[0]--;
 846                         btrfs_item_key_to_cpu(path->nodes[0], &key,
 847                                               path->slots[0]);
 848                         if (key.objectid == bytenr &&
 849                             key.type == BTRFS_EXTENT_ITEM_KEY &&
 850                             key.offset == num_bytes)
 851                                 ret = 0;
 852                 }
 853                 if (ret) {
 854                         key.objectid = bytenr;
 855                         key.type = BTRFS_EXTENT_ITEM_KEY;
 856                         key.offset = num_bytes;
 857                         btrfs_release_path(path);
 858                         goto again;
 859                 }
 860         }
 861
 862         if (ret && !insert) {
 863                 err = -ENOENT;
 864                 goto out;
 865         } else if (WARN_ON(ret)) {
 866                 err = -EIO;
 867                 goto out;
 868         }
 869
 870         leaf = path->nodes[0];
 871         item_size = btrfs_item_size(leaf, path->slots[0]);
 872         if (unlikely(item_size < sizeof(*ei))) {
 873                 err = -EINVAL;
 874                 btrfs_print_v0_err(fs_info);
 875                 btrfs_abort_transaction(trans, err);
 876                 goto out;
 877         }
 878
 879         ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
 880         flags = btrfs_extent_flags(leaf, ei);
 881
 882         ptr = (unsigned long)(ei + 1);
 883         end = (unsigned long)ei + item_size;
 884
 885         if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK && !skinny_metadata) {
 886                 ptr += sizeof(struct btrfs_tree_block_info);
 887                 BUG_ON(ptr > end);
 888         }
 889
 890         if (owner >= BTRFS_FIRST_FREE_OBJECTID)
 891                 needed = BTRFS_REF_TYPE_DATA;
 892         else
 893                 needed = BTRFS_REF_TYPE_BLOCK;
 894
 895         err = -ENOENT;
 896         while (1) {
 897                 if (ptr >= end) {
 898                         if (ptr > end) {
 899                                 err = -EUCLEAN;
 900                                 btrfs_print_leaf(path->nodes[0]);
 901                                 btrfs_crit(fs_info,
 902 "overrun extent record at slot %d while looking for inline extent for root %llu owner %llu offset %llu parent %llu",
 903                                         path->slots[0], root_objectid, owner, offset, parent);
 904                         }
 905                         break;
 906                 }
 907                 iref = (struct btrfs_extent_inline_ref *)ptr;
 908                 type = btrfs_get_extent_inline_ref_type(leaf, iref, needed);
 909                 if (type == BTRFS_REF_TYPE_INVALID) {
 910                         err = -EUCLEAN;
 911                         goto out;
 912                 }
 913
 914                 if (want < type)
 915                         break;
 916                 if (want > type) {
 917                         ptr += btrfs_extent_inline_ref_size(type);
 918                         continue;
 919                 }
 920
 921                 if (type == BTRFS_EXTENT_DATA_REF_KEY) {
 922                         struct btrfs_extent_data_ref *dref;
 923                         dref = (struct btrfs_extent_data_ref *)(&iref->offset);
 924                         if (match_extent_data_ref(leaf, dref, root_objectid,
 925                                                   owner, offset)) {
 926                                 err = 0;
 927                                 break;
 928                         }
 929                         if (hash_extent_data_ref_item(leaf, dref) <
 930                             hash_extent_data_ref(root_objectid, owner, offset))
 931                                 break;
 932                 } else {
 933                         u64 ref_offset;
 934                         ref_offset = btrfs_extent_inline_ref_offset(leaf, iref);
 935                         if (parent > 0) {
 936                                 if (parent == ref_offset) {
 937                                         err = 0;
 938                                         break;
 939                                 }
 940                                 if (ref_offset < parent)
 941                                         break;
 942                         } else {
 943                                 if (root_objectid == ref_offset) {
 944                                         err = 0;
 945                                         break;
 946                                 }
 947                                 if (ref_offset < root_objectid)
 948                                         break;
 949                         }
 950                 }
 951                 ptr += btrfs_extent_inline_ref_size(type);
 952         }
 953         if (err == -ENOENT && insert) {
 954                 if (item_size + extra_size >=
 955                     BTRFS_MAX_EXTENT_ITEM_SIZE(root)) {
 956                         err = -EAGAIN;
 957                         goto out;
 958                 }
 959                 /*
 960                  * To add new inline back ref, we have to make sure
 961                  * there is no corresponding back ref item.
 962                  * For simplicity, we just do not add new inline back
 963                  * ref if there is any kind of item for this block
 964                  */
 965                 if (find_next_key(path, 0, &key) == 0 &&
 966                     key.objectid == bytenr &&
 967                     key.type < BTRFS_BLOCK_GROUP_ITEM_KEY) {
 968                         err = -EAGAIN;
 969                         goto out;
 970                 }
 971         }
 972         *ref_ret = (struct btrfs_extent_inline_ref *)ptr;
 973 out:
 974         if (insert) {
 975                 path->keep_locks = 0;
 976                 path->search_for_extension = 0;
 977                 btrfs_unlock_up_safe(path, 1);
 978         }
 979         return err;
 980 }
 981
 982 /*
 983  * helper to add new inline back ref
 984  */
 985 static noinline_for_stack
 986 void setup_inline_extent_backref(struct btrfs_fs_info *fs_info,
 987                                  struct btrfs_path *path,
 988                                  struct btrfs_extent_inline_ref *iref,
 989                                  u64 parent, u64 root_objectid,
 990                                  u64 owner, u64 offset, int refs_to_add,
 991                                  struct btrfs_delayed_extent_op *extent_op)
 992 {
 993         struct extent_buffer *leaf;
 994         struct btrfs_extent_item *ei;
 995         unsigned long ptr;
 996         unsigned long end;
 997         unsigned long item_offset;
 998         u64 refs;
 999         int size;
1000         int type;
1001
1002         leaf = path->nodes[0];
1003         ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1004         item_offset = (unsigned long)iref - (unsigned long)ei;
1005
1006         type = extent_ref_type(parent, owner);
1007         size = btrfs_extent_inline_ref_size(type);
1008
1009         btrfs_extend_item(path, size);
1010
1011         ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1012         refs = btrfs_extent_refs(leaf, ei);
1013         refs += refs_to_add;
1014         btrfs_set_extent_refs(leaf, ei, refs);
1015         if (extent_op)
1016                 __run_delayed_extent_op(extent_op, leaf, ei);
1017
1018         ptr = (unsigned long)ei + item_offset;
1019         end = (unsigned long)ei + btrfs_item_size(leaf, path->slots[0]);
1020         if (ptr < end - size)
1021                 memmove_extent_buffer(leaf, ptr + size, ptr,
1022                                       end - size - ptr);
1023
1024         iref = (struct btrfs_extent_inline_ref *)ptr;
1025         btrfs_set_extent_inline_ref_type(leaf, iref, type);
1026         if (type == BTRFS_EXTENT_DATA_REF_KEY) {
1027                 struct btrfs_extent_data_ref *dref;
1028                 dref = (struct btrfs_extent_data_ref *)(&iref->offset);
1029                 btrfs_set_extent_data_ref_root(leaf, dref, root_objectid);
1030                 btrfs_set_extent_data_ref_objectid(leaf, dref, owner);
1031                 btrfs_set_extent_data_ref_offset(leaf, dref, offset);
1032                 btrfs_set_extent_data_ref_count(leaf, dref, refs_to_add);
1033         } else if (type == BTRFS_SHARED_DATA_REF_KEY) {
1034                 struct btrfs_shared_data_ref *sref;
1035                 sref = (struct btrfs_shared_data_ref *)(iref + 1);
1036                 btrfs_set_shared_data_ref_count(leaf, sref, refs_to_add);
1037                 btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
1038         } else if (type == BTRFS_SHARED_BLOCK_REF_KEY) {
1039                 btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
1040         } else {
1041                 btrfs_set_extent_inline_ref_offset(leaf, iref, root_objectid);
1042         }
1043         btrfs_mark_buffer_dirty(leaf);
1044 }
1045
1046 static int lookup_extent_backref(struct btrfs_trans_handle *trans,
1047                                  struct btrfs_path *path,
1048                                  struct btrfs_extent_inline_ref **ref_ret,
1049                                  u64 bytenr, u64 num_bytes, u64 parent,
1050                                  u64 root_objectid, u64 owner, u64 offset)
1051 {
1052         int ret;
1053
1054         ret = lookup_inline_extent_backref(trans, path, ref_ret, bytenr,
1055                                            num_bytes, parent, root_objectid,
1056                                            owner, offset, 0);
1057         if (ret != -ENOENT)
1058                 return ret;
1059
1060         btrfs_release_path(path);
1061         *ref_ret = NULL;
1062
1063         if (owner < BTRFS_FIRST_FREE_OBJECTID) {
1064                 ret = lookup_tree_block_ref(trans, path, bytenr, parent,
1065                                             root_objectid);
1066         } else {
1067                 ret = lookup_extent_data_ref(trans, path, bytenr, parent,
1068                                              root_objectid, owner, offset);
1069         }
1070         return ret;
1071 }
1072
1073 /*
1074  * helper to update/remove inline back ref
1075  */
1076 static noinline_for_stack
1077 void update_inline_extent_backref(struct btrfs_path *path,
1078                                   struct btrfs_extent_inline_ref *iref,
1079                                   int refs_to_mod,
1080                                   struct btrfs_delayed_extent_op *extent_op)
1081 {
1082         struct extent_buffer *leaf = path->nodes[0];
1083         struct btrfs_extent_item *ei;
1084         struct btrfs_extent_data_ref *dref = NULL;
1085         struct btrfs_shared_data_ref *sref = NULL;
1086         unsigned long ptr;
1087         unsigned long end;
1088         u32 item_size;
1089         int size;
1090         int type;
1091         u64 refs;
1092
1093         ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1094         refs = btrfs_extent_refs(leaf, ei);
1095         WARN_ON(refs_to_mod < 0 && refs + refs_to_mod <= 0);
1096         refs += refs_to_mod;
1097         btrfs_set_extent_refs(leaf, ei, refs);
1098         if (extent_op)
1099                 __run_delayed_extent_op(extent_op, leaf, ei);
1100
1101         /*
1102          * If type is invalid, we should have bailed out after
1103          * lookup_inline_extent_backref().
1104          */
1105         type = btrfs_get_extent_inline_ref_type(leaf, iref, BTRFS_REF_TYPE_ANY);
1106         ASSERT(type != BTRFS_REF_TYPE_INVALID);
1107
1108         if (type == BTRFS_EXTENT_DATA_REF_KEY) {
1109                 dref = (struct btrfs_extent_data_ref *)(&iref->offset);
1110                 refs = btrfs_extent_data_ref_count(leaf, dref);
1111         } else if (type == BTRFS_SHARED_DATA_REF_KEY) {
1112                 sref = (struct btrfs_shared_data_ref *)(iref + 1);
1113                 refs = btrfs_shared_data_ref_count(leaf, sref);
1114         } else {
1115                 refs = 1;
1116                 BUG_ON(refs_to_mod != -1);
1117         }
1118
1119         BUG_ON(refs_to_mod < 0 && refs < -refs_to_mod);
1120         refs += refs_to_mod;
1121
1122         if (refs > 0) {
1123                 if (type == BTRFS_EXTENT_DATA_REF_KEY)
1124                         btrfs_set_extent_data_ref_count(leaf, dref, refs);
1125                 else
1126                         btrfs_set_shared_data_ref_count(leaf, sref, refs);
1127         } else {
1128                 size =  btrfs_extent_inline_ref_size(type);
1129                 item_size = btrfs_item_size(leaf, path->slots[0]);
1130                 ptr = (unsigned long)iref;
1131                 end = (unsigned long)ei + item_size;
1132                 if (ptr + size < end)
1133                         memmove_extent_buffer(leaf, ptr, ptr + size,
1134                                               end - ptr - size);
1135                 item_size -= size;
1136                 btrfs_truncate_item(path, item_size, 1);
1137         }
1138         btrfs_mark_buffer_dirty(leaf);
1139 }
1140
1141 static noinline_for_stack
1142 int insert_inline_extent_backref(struct btrfs_trans_handle *trans,
1143                                  struct btrfs_path *path,
1144                                  u64 bytenr, u64 num_bytes, u64 parent,
1145                                  u64 root_objectid, u64 owner,
1146                                  u64 offset, int refs_to_add,
1147                                  struct btrfs_delayed_extent_op *extent_op)
1148 {
1149         struct btrfs_extent_inline_ref *iref;
1150         int ret;
1151
1152         ret = lookup_inline_extent_backref(trans, path, &iref, bytenr,
1153                                            num_bytes, parent, root_objectid,
1154                                            owner, offset, 1);
1155         if (ret == 0) {
1156                 /*
1157                  * We're adding refs to a tree block we already own, this
1158                  * should not happen at all.
1159                  */
1160                 if (owner < BTRFS_FIRST_FREE_OBJECTID) {
1161                         btrfs_crit(trans->fs_info,
1162 "adding refs to an existing tree ref, bytenr %llu num_bytes %llu root_objectid %llu",
1163                                    bytenr, num_bytes, root_objectid);
1164                         if (IS_ENABLED(CONFIG_BTRFS_DEBUG)) {
1165                                 WARN_ON(1);
1166                                 btrfs_crit(trans->fs_info,
1167                         "path->slots[0]=%d path->nodes[0]:", path->slots[0]);
1168                                 btrfs_print_leaf(path->nodes[0]);
1169                         }
1170                         return -EUCLEAN;
1171                 }
1172                 update_inline_extent_backref(path, iref, refs_to_add, extent_op);
1173         } else if (ret == -ENOENT) {
1174                 setup_inline_extent_backref(trans->fs_info, path, iref, parent,
1175                                             root_objectid, owner, offset,
1176                                             refs_to_add, extent_op);
1177                 ret = 0;
1178         }
1179         return ret;
1180 }
1181
1182 static int remove_extent_backref(struct btrfs_trans_handle *trans,
1183                                  struct btrfs_root *root,
1184                                  struct btrfs_path *path,
1185                                  struct btrfs_extent_inline_ref *iref,
1186                                  int refs_to_drop, int is_data)
1187 {
1188         int ret = 0;
1189
1190         BUG_ON(!is_data && refs_to_drop != 1);
1191         if (iref)
1192                 update_inline_extent_backref(path, iref, -refs_to_drop, NULL);
1193         else if (is_data)
1194                 ret = remove_extent_data_ref(trans, root, path, refs_to_drop);
1195         else
1196                 ret = btrfs_del_item(trans, root, path);
1197         return ret;
1198 }
1199
1200 static int btrfs_issue_discard(struct block_device *bdev, u64 start, u64 len,
1201                                u64 *discarded_bytes)
1202 {
1203         int j, ret = 0;
1204         u64 bytes_left, end;
1205         u64 aligned_start = ALIGN(start, 1 << 9);
1206
1207         if (WARN_ON(start != aligned_start)) {
1208                 len -= aligned_start - start;
1209                 len = round_down(len, 1 << 9);
1210                 start = aligned_start;
1211         }
1212
1213         *discarded_bytes = 0;
1214
1215         if (!len)
1216                 return 0;
1217
1218         end = start + len;
1219         bytes_left = len;
1220
1221         /* Skip any superblocks on this device. */
1222         for (j = 0; j < BTRFS_SUPER_MIRROR_MAX; j++) {
1223                 u64 sb_start = btrfs_sb_offset(j);
1224                 u64 sb_end = sb_start + BTRFS_SUPER_INFO_SIZE;
1225                 u64 size = sb_start - start;
1226
1227                 if (!in_range(sb_start, start, bytes_left) &&
1228                     !in_range(sb_end, start, bytes_left) &&
1229                     !in_range(start, sb_start, BTRFS_SUPER_INFO_SIZE))
1230                         continue;
1231
1232                 /*
1233                  * Superblock spans beginning of range.  Adjust start and
1234                  * try again.
1235                  */
1236                 if (sb_start <= start) {
1237                         start += sb_end - start;
1238                         if (start > end) {
1239                                 bytes_left = 0;
1240                                 break;
1241                         }
1242                         bytes_left = end - start;
1243                         continue;
1244                 }
1245
1246                 if (size) {
1247                         ret = blkdev_issue_discard(bdev, start >> 9, size >> 9,
1248                                                    GFP_NOFS);
1249                         if (!ret)
1250                                 *discarded_bytes += size;
1251                         else if (ret != -EOPNOTSUPP)
1252                                 return ret;
1253                 }
1254
1255                 start = sb_end;
1256                 if (start > end) {
1257                         bytes_left = 0;
1258                         break;
1259                 }
1260                 bytes_left = end - start;
1261         }
1262
1263         if (bytes_left) {
1264                 ret = blkdev_issue_discard(bdev, start >> 9, bytes_left >> 9,
1265                                            GFP_NOFS);
1266                 if (!ret)
1267                         *discarded_bytes += bytes_left;
1268         }
1269         return ret;
1270 }
1271
1272 static int do_discard_extent(struct btrfs_io_stripe *stripe, u64 *bytes)
1273 {
1274         struct btrfs_device *dev = stripe->dev;
1275         struct btrfs_fs_info *fs_info = dev->fs_info;
1276         struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
1277         u64 phys = stripe->physical;
1278         u64 len = stripe->length;
1279         u64 discarded = 0;
1280         int ret = 0;
1281
1282         /* Zone reset on a zoned filesystem */
1283         if (btrfs_can_zone_reset(dev, phys, len)) {
1284                 u64 src_disc;
1285
1286                 ret = btrfs_reset_device_zone(dev, phys, len, &discarded);
1287                 if (ret)
1288                         goto out;
1289
1290                 if (!btrfs_dev_replace_is_ongoing(dev_replace) ||
1291                     dev != dev_replace->srcdev)
1292                         goto out;
1293
1294                 src_disc = discarded;
1295
1296                 /* Send to replace target as well */
1297                 ret = btrfs_reset_device_zone(dev_replace->tgtdev, phys, len,
1298                                               &discarded);
1299                 discarded += src_disc;
1300         } else if (bdev_max_discard_sectors(stripe->dev->bdev)) {
1301                 ret = btrfs_issue_discard(dev->bdev, phys, len, &discarded);
1302         } else {
1303                 ret = 0;
1304                 *bytes = 0;
1305         }
1306
1307 out:
1308         *bytes = discarded;
1309         return ret;
1310 }
1311
1312 int btrfs_discard_extent(struct btrfs_fs_info *fs_info, u64 bytenr,
1313                          u64 num_bytes, u64 *actual_bytes)
1314 {
1315         int ret = 0;
1316         u64 discarded_bytes = 0;
1317         u64 end = bytenr + num_bytes;
1318         u64 cur = bytenr;
1319         struct btrfs_io_context *bioc = NULL;
1320
1321         /*
1322          * Avoid races with device replace and make sure our bioc has devices
1323          * associated to its stripes that don't go away while we are discarding.
1324          */
1325         btrfs_bio_counter_inc_blocked(fs_info);
1326         while (cur < end) {
1327                 struct btrfs_io_stripe *stripe;
1328                 int i;
1329
1330                 num_bytes = end - cur;
1331                 /* Tell the block device(s) that the sectors can be discarded */
1332                 ret = btrfs_map_block(fs_info, BTRFS_MAP_DISCARD, cur,
1333                                       &num_bytes, &bioc, 0);
1334                 /*
1335                  * Error can be -ENOMEM, -ENOENT (no such chunk mapping) or
1336                  * -EOPNOTSUPP. For any such error, @num_bytes is not updated,
1337                  * thus we can't continue anyway.
1338                  */
1339                 if (ret < 0)
1340                         goto out;
1341
1342                 stripe = bioc->stripes;
1343                 for (i = 0; i < bioc->num_stripes; i++, stripe++) {
1344                         u64 bytes;
1345                         struct btrfs_device *device = stripe->dev;
1346
1347                         if (!device->bdev) {
1348                                 ASSERT(btrfs_test_opt(fs_info, DEGRADED));
1349                                 continue;
1350                         }
1351
1352                         if (!test_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state))
1353                                 continue;
1354
1355                         ret = do_discard_extent(stripe, &bytes);
1356                         if (!ret) {
1357                                 discarded_bytes += bytes;
1358                         } else if (ret != -EOPNOTSUPP) {
1359                                 /*
1360                                  * Logic errors or -ENOMEM, or -EIO, but
1361                                  * unlikely to happen.
1362                                  *
1363                                  * And since there are two loops, explicitly
1364                                  * go to out to avoid confusion.
1365                                  */
1366                                 btrfs_put_bioc(bioc);
1367                                 goto out;
1368                         }
1369
1370                         /*
1371                          * Just in case we get back EOPNOTSUPP for some reason,
1372                          * just ignore the return value so we don't screw up
1373                          * people calling discard_extent.
1374                          */
1375                         ret = 0;
1376                 }
1377                 btrfs_put_bioc(bioc);
1378                 cur += num_bytes;
1379         }
1380 out:
1381         btrfs_bio_counter_dec(fs_info);
1382
1383         if (actual_bytes)
1384                 *actual_bytes = discarded_bytes;
1385
1386
1387         if (ret == -EOPNOTSUPP)
1388                 ret = 0;
1389         return ret;
1390 }
1391
1392 /* Can return -ENOMEM */
1393 int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
1394                          struct btrfs_ref *generic_ref)
1395 {
1396         struct btrfs_fs_info *fs_info = trans->fs_info;
1397         int ret;
1398
1399         ASSERT(generic_ref->type != BTRFS_REF_NOT_SET &&
1400                generic_ref->action);
1401         BUG_ON(generic_ref->type == BTRFS_REF_METADATA &&
1402                generic_ref->tree_ref.owning_root == BTRFS_TREE_LOG_OBJECTID);
1403
1404         if (generic_ref->type == BTRFS_REF_METADATA)
1405                 ret = btrfs_add_delayed_tree_ref(trans, generic_ref, NULL);
1406         else
1407                 ret = btrfs_add_delayed_data_ref(trans, generic_ref, 0);
1408
1409         btrfs_ref_tree_mod(fs_info, generic_ref);
1410
1411         return ret;
1412 }
1413
1414 /*
1415  * __btrfs_inc_extent_ref - insert backreference for a given extent
1416  *
1417  * The counterpart is in __btrfs_free_extent(), with examples and more details
1418  * how it works.
1419  *
1420  * @trans:          Handle of transaction
1421  *
1422  * @node:           The delayed ref node used to get the bytenr/length for
1423  *                  extent whose references are incremented.
1424  *
1425  * @parent:         If this is a shared extent (BTRFS_SHARED_DATA_REF_KEY/
1426  *                  BTRFS_SHARED_BLOCK_REF_KEY) then it holds the logical
1427  *                  bytenr of the parent block. Since new extents are always
1428  *                  created with indirect references, this will only be the case
1429  *                  when relocating a shared extent. In that case, root_objectid
1430  *                  will be BTRFS_TREE_RELOC_OBJECTID. Otherwise, parent must
1431  *                  be 0
1432  *
1433  * @root_objectid:  The id of the root where this modification has originated,
1434  *                  this can be either one of the well-known metadata trees or
1435  *                  the subvolume id which references this extent.
1436  *
1437  * @owner:          For data extents it is the inode number of the owning file.
1438  *                  For metadata extents this parameter holds the level in the
1439  *                  tree of the extent.
1440  *
1441  * @offset:         For metadata extents the offset is ignored and is currently
1442  *                  always passed as 0. For data extents it is the fileoffset
1443  *                  this extent belongs to.
1444  *
1445  * @refs_to_add     Number of references to add
1446  *
1447  * @extent_op       Pointer to a structure, holding information necessary when
1448  *                  updating a tree block's flags
1449  *
1450  */
1451 static int __btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
1452                                   struct btrfs_delayed_ref_node *node,
1453                                   u64 parent, u64 root_objectid,
1454                                   u64 owner, u64 offset, int refs_to_add,
1455                                   struct btrfs_delayed_extent_op *extent_op)
1456 {
1457         struct btrfs_path *path;
1458         struct extent_buffer *leaf;
1459         struct btrfs_extent_item *item;
1460         struct btrfs_key key;
1461         u64 bytenr = node->bytenr;
1462         u64 num_bytes = node->num_bytes;
1463         u64 refs;
1464         int ret;
1465
1466         path = btrfs_alloc_path();
1467         if (!path)
1468                 return -ENOMEM;
1469
1470         /* this will setup the path even if it fails to insert the back ref */
1471         ret = insert_inline_extent_backref(trans, path, bytenr, num_bytes,
1472                                            parent, root_objectid, owner,
1473                                            offset, refs_to_add, extent_op);
1474         if ((ret < 0 && ret != -EAGAIN) || !ret)
1475                 goto out;
1476
1477         /*
1478          * Ok we had -EAGAIN which means we didn't have space to insert and
1479          * inline extent ref, so just update the reference count and add a
1480          * normal backref.
1481          */
1482         leaf = path->nodes[0];
1483         btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
1484         item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1485         refs = btrfs_extent_refs(leaf, item);
1486         btrfs_set_extent_refs(leaf, item, refs + refs_to_add);
1487         if (extent_op)
1488                 __run_delayed_extent_op(extent_op, leaf, item);
1489
1490         btrfs_mark_buffer_dirty(leaf);
1491         btrfs_release_path(path);
1492
1493         /* now insert the actual backref */
1494         if (owner < BTRFS_FIRST_FREE_OBJECTID) {
1495                 BUG_ON(refs_to_add != 1);
1496                 ret = insert_tree_block_ref(trans, path, bytenr, parent,
1497                                             root_objectid);
1498         } else {
1499                 ret = insert_extent_data_ref(trans, path, bytenr, parent,
1500                                              root_objectid, owner, offset,
1501                                              refs_to_add);
1502         }
1503         if (ret)
1504                 btrfs_abort_transaction(trans, ret);
1505 out:
1506         btrfs_free_path(path);
1507         return ret;
1508 }
1509
1510 static int run_delayed_data_ref(struct btrfs_trans_handle *trans,
1511                                 struct btrfs_delayed_ref_node *node,
1512                                 struct btrfs_delayed_extent_op *extent_op,
1513                                 int insert_reserved)
1514 {
1515         int ret = 0;
1516         struct btrfs_delayed_data_ref *ref;
1517         struct btrfs_key ins;
1518         u64 parent = 0;
1519         u64 ref_root = 0;
1520         u64 flags = 0;
1521
1522         ins.objectid = node->bytenr;
1523         ins.offset = node->num_bytes;
1524         ins.type = BTRFS_EXTENT_ITEM_KEY;
1525
1526         ref = btrfs_delayed_node_to_data_ref(node);
1527         trace_run_delayed_data_ref(trans->fs_info, node, ref, node->action);
1528
1529         if (node->type == BTRFS_SHARED_DATA_REF_KEY)
1530                 parent = ref->parent;
1531         ref_root = ref->root;
1532
1533         if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) {
1534                 if (extent_op)
1535                         flags |= extent_op->flags_to_set;
1536                 ret = alloc_reserved_file_extent(trans, parent, ref_root,
1537                                                  flags, ref->objectid,
1538                                                  ref->offset, &ins,
1539                                                  node->ref_mod);
1540         } else if (node->action == BTRFS_ADD_DELAYED_REF) {
1541                 ret = __btrfs_inc_extent_ref(trans, node, parent, ref_root,
1542                                              ref->objectid, ref->offset,
1543                                              node->ref_mod, extent_op);
1544         } else if (node->action == BTRFS_DROP_DELAYED_REF) {
1545                 ret = __btrfs_free_extent(trans, node, parent,
1546                                           ref_root, ref->objectid,
1547                                           ref->offset, node->ref_mod,
1548                                           extent_op);
1549         } else {
1550                 BUG();
1551         }
1552         return ret;
1553 }
1554
1555 static void __run_delayed_extent_op(struct btrfs_delayed_extent_op *extent_op,
1556                                     struct extent_buffer *leaf,
1557                                     struct btrfs_extent_item *ei)
1558 {
1559         u64 flags = btrfs_extent_flags(leaf, ei);
1560         if (extent_op->update_flags) {
1561                 flags |= extent_op->flags_to_set;
1562                 btrfs_set_extent_flags(leaf, ei, flags);
1563         }
1564
1565         if (extent_op->update_key) {
1566                 struct btrfs_tree_block_info *bi;
1567                 BUG_ON(!(flags & BTRFS_EXTENT_FLAG_TREE_BLOCK));
1568                 bi = (struct btrfs_tree_block_info *)(ei + 1);
1569                 btrfs_set_tree_block_key(leaf, bi, &extent_op->key);
1570         }
1571 }
1572
1573 static int run_delayed_extent_op(struct btrfs_trans_handle *trans,
1574                                  struct btrfs_delayed_ref_head *head,
1575                                  struct btrfs_delayed_extent_op *extent_op)
1576 {
1577         struct btrfs_fs_info *fs_info = trans->fs_info;
1578         struct btrfs_root *root;
1579         struct btrfs_key key;
1580         struct btrfs_path *path;
1581         struct btrfs_extent_item *ei;
1582         struct extent_buffer *leaf;
1583         u32 item_size;
1584         int ret;
1585         int err = 0;
1586         int metadata = 1;
1587
1588         if (TRANS_ABORTED(trans))
1589                 return 0;
1590
1591         if (!btrfs_fs_incompat(fs_info, SKINNY_METADATA))
1592                 metadata = 0;
1593
1594         path = btrfs_alloc_path();
1595         if (!path)
1596                 return -ENOMEM;
1597
1598         key.objectid = head->bytenr;
1599
1600         if (metadata) {
1601                 key.type = BTRFS_METADATA_ITEM_KEY;
1602                 key.offset = extent_op->level;
1603         } else {
1604                 key.type = BTRFS_EXTENT_ITEM_KEY;
1605                 key.offset = head->num_bytes;
1606         }
1607
1608         root = btrfs_extent_root(fs_info, key.objectid);
1609 again:
1610         ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
1611         if (ret < 0) {
1612                 err = ret;
1613                 goto out;
1614         }
1615         if (ret > 0) {
1616                 if (metadata) {
1617                         if (path->slots[0] > 0) {
1618                                 path->slots[0]--;
1619                                 btrfs_item_key_to_cpu(path->nodes[0], &key,
1620                                                       path->slots[0]);
1621                                 if (key.objectid == head->bytenr &&
1622                                     key.type == BTRFS_EXTENT_ITEM_KEY &&
1623                                     key.offset == head->num_bytes)
1624                                         ret = 0;
1625                         }
1626                         if (ret > 0) {
1627                                 btrfs_release_path(path);
1628                                 metadata = 0;
1629
1630                                 key.objectid = head->bytenr;
1631                                 key.offset = head->num_bytes;
1632                                 key.type = BTRFS_EXTENT_ITEM_KEY;
1633                                 goto again;
1634                         }
1635                 } else {
1636                         err = -EIO;
1637                         goto out;
1638                 }
1639         }
1640
1641         leaf = path->nodes[0];
1642         item_size = btrfs_item_size(leaf, path->slots[0]);
1643
1644         if (unlikely(item_size < sizeof(*ei))) {
1645                 err = -EINVAL;
1646                 btrfs_print_v0_err(fs_info);
1647                 btrfs_abort_transaction(trans, err);
1648                 goto out;
1649         }
1650
1651         ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
1652         __run_delayed_extent_op(extent_op, leaf, ei);
1653
1654         btrfs_mark_buffer_dirty(leaf);
1655 out:
1656         btrfs_free_path(path);
1657         return err;
1658 }
1659
1660 static int run_delayed_tree_ref(struct btrfs_trans_handle *trans,
1661                                 struct btrfs_delayed_ref_node *node,
1662                                 struct btrfs_delayed_extent_op *extent_op,
1663                                 int insert_reserved)
1664 {
1665         int ret = 0;
1666         struct btrfs_delayed_tree_ref *ref;
1667         u64 parent = 0;
1668         u64 ref_root = 0;
1669
1670         ref = btrfs_delayed_node_to_tree_ref(node);
1671         trace_run_delayed_tree_ref(trans->fs_info, node, ref, node->action);
1672
1673         if (node->type == BTRFS_SHARED_BLOCK_REF_KEY)
1674                 parent = ref->parent;
1675         ref_root = ref->root;
1676
1677         if (node->ref_mod != 1) {
1678                 btrfs_err(trans->fs_info,
1679         "btree block(%llu) has %d references rather than 1: action %d ref_root %llu parent %llu",
1680                           node->bytenr, node->ref_mod, node->action, ref_root,
1681                           parent);
1682                 return -EIO;
1683         }
1684         if (node->action == BTRFS_ADD_DELAYED_REF && insert_reserved) {
1685                 BUG_ON(!extent_op || !extent_op->update_flags);
1686                 ret = alloc_reserved_tree_block(trans, node, extent_op);
1687         } else if (node->action == BTRFS_ADD_DELAYED_REF) {
1688                 ret = __btrfs_inc_extent_ref(trans, node, parent, ref_root,
1689                                              ref->level, 0, 1, extent_op);
1690         } else if (node->action == BTRFS_DROP_DELAYED_REF) {
1691                 ret = __btrfs_free_extent(trans, node, parent, ref_root,
1692                                           ref->level, 0, 1, extent_op);
1693         } else {
1694                 BUG();
1695         }
1696         return ret;
1697 }
1698
1699 /* helper function to actually process a single delayed ref entry */
1700 static int run_one_delayed_ref(struct btrfs_trans_handle *trans,
1701                                struct btrfs_delayed_ref_node *node,
1702                                struct btrfs_delayed_extent_op *extent_op,
1703                                int insert_reserved)
1704 {
1705         int ret = 0;
1706
1707         if (TRANS_ABORTED(trans)) {
1708                 if (insert_reserved)
1709                         btrfs_pin_extent(trans, node->bytenr, node->num_bytes, 1);
1710                 return 0;
1711         }
1712
1713         if (node->type == BTRFS_TREE_BLOCK_REF_KEY ||
1714             node->type == BTRFS_SHARED_BLOCK_REF_KEY)
1715                 ret = run_delayed_tree_ref(trans, node, extent_op,
1716                                            insert_reserved);
1717         else if (node->type == BTRFS_EXTENT_DATA_REF_KEY ||
1718                  node->type == BTRFS_SHARED_DATA_REF_KEY)
1719                 ret = run_delayed_data_ref(trans, node, extent_op,
1720                                            insert_reserved);
1721         else
1722                 BUG();
1723         if (ret && insert_reserved)
1724                 btrfs_pin_extent(trans, node->bytenr, node->num_bytes, 1);
1725         return ret;
1726 }
1727
1728 static inline struct btrfs_delayed_ref_node *
1729 select_delayed_ref(struct btrfs_delayed_ref_head *head)
1730 {
1731         struct btrfs_delayed_ref_node *ref;
1732
1733         if (RB_EMPTY_ROOT(&head->ref_tree.rb_root))
1734                 return NULL;
1735
1736         /*
1737          * Select a delayed ref of type BTRFS_ADD_DELAYED_REF first.
1738          * This is to prevent a ref count from going down to zero, which deletes
1739          * the extent item from the extent tree, when there still are references
1740          * to add, which would fail because they would not find the extent item.
1741          */
1742         if (!list_empty(&head->ref_add_list))
1743                 return list_first_entry(&head->ref_add_list,
1744                                 struct btrfs_delayed_ref_node, add_list);
1745
1746         ref = rb_entry(rb_first_cached(&head->ref_tree),
1747                        struct btrfs_delayed_ref_node, ref_node);
1748         ASSERT(list_empty(&ref->add_list));
1749         return ref;
1750 }
1751
1752 static void unselect_delayed_ref_head(struct btrfs_delayed_ref_root *delayed_refs,
1753                                       struct btrfs_delayed_ref_head *head)
1754 {
1755         spin_lock(&delayed_refs->lock);
1756         head->processing = 0;
1757         delayed_refs->num_heads_ready++;
1758         spin_unlock(&delayed_refs->lock);
1759         btrfs_delayed_ref_unlock(head);
1760 }
1761
1762 static struct btrfs_delayed_extent_op *cleanup_extent_op(
1763                                 struct btrfs_delayed_ref_head *head)
1764 {
1765         struct btrfs_delayed_extent_op *extent_op = head->extent_op;
1766
1767         if (!extent_op)
1768                 return NULL;
1769
1770         if (head->must_insert_reserved) {
1771                 head->extent_op = NULL;
1772                 btrfs_free_delayed_extent_op(extent_op);
1773                 return NULL;
1774         }
1775         return extent_op;
1776 }
1777
1778 static int run_and_cleanup_extent_op(struct btrfs_trans_handle *trans,
1779                                      struct btrfs_delayed_ref_head *head)
1780 {
1781         struct btrfs_delayed_extent_op *extent_op;
1782         int ret;
1783
1784         extent_op = cleanup_extent_op(head);
1785         if (!extent_op)
1786                 return 0;
1787         head->extent_op = NULL;
1788         spin_unlock(&head->lock);
1789         ret = run_delayed_extent_op(trans, head, extent_op);
1790         btrfs_free_delayed_extent_op(extent_op);
1791         return ret ? ret : 1;
1792 }
1793
1794 void btrfs_cleanup_ref_head_accounting(struct btrfs_fs_info *fs_info,
1795                                   struct btrfs_delayed_ref_root *delayed_refs,
1796                                   struct btrfs_delayed_ref_head *head)
1797 {
1798         int nr_items = 1;       /* Dropping this ref head update. */
1799
1800         /*
1801          * We had csum deletions accounted for in our delayed refs rsv, we need
1802          * to drop the csum leaves for this update from our delayed_refs_rsv.
1803          */
1804         if (head->total_ref_mod < 0 && head->is_data) {
1805                 spin_lock(&delayed_refs->lock);
1806                 delayed_refs->pending_csums -= head->num_bytes;
1807                 spin_unlock(&delayed_refs->lock);
1808                 nr_items += btrfs_csum_bytes_to_leaves(fs_info, head->num_bytes);
1809         }
1810
1811         btrfs_delayed_refs_rsv_release(fs_info, nr_items);
1812 }
1813
1814 static int cleanup_ref_head(struct btrfs_trans_handle *trans,
1815                             struct btrfs_delayed_ref_head *head)
1816 {
1817
1818         struct btrfs_fs_info *fs_info = trans->fs_info;
1819         struct btrfs_delayed_ref_root *delayed_refs;
1820         int ret;
1821
1822         delayed_refs = &trans->transaction->delayed_refs;
1823
1824         ret = run_and_cleanup_extent_op(trans, head);
1825         if (ret < 0) {
1826                 unselect_delayed_ref_head(delayed_refs, head);
1827                 btrfs_debug(fs_info, "run_delayed_extent_op returned %d", ret);
1828                 return ret;
1829         } else if (ret) {
1830                 return ret;
1831         }
1832
1833         /*
1834          * Need to drop our head ref lock and re-acquire the delayed ref lock
1835          * and then re-check to make sure nobody got added.
1836          */
1837         spin_unlock(&head->lock);
1838         spin_lock(&delayed_refs->lock);
1839         spin_lock(&head->lock);
1840         if (!RB_EMPTY_ROOT(&head->ref_tree.rb_root) || head->extent_op) {
1841                 spin_unlock(&head->lock);
1842                 spin_unlock(&delayed_refs->lock);
1843                 return 1;
1844         }
1845         btrfs_delete_ref_head(delayed_refs, head);
1846         spin_unlock(&head->lock);
1847         spin_unlock(&delayed_refs->lock);
1848
1849         if (head->must_insert_reserved) {
1850                 btrfs_pin_extent(trans, head->bytenr, head->num_bytes, 1);
1851                 if (head->is_data) {
1852                         struct btrfs_root *csum_root;
1853
1854                         csum_root = btrfs_csum_root(fs_info, head->bytenr);
1855                         ret = btrfs_del_csums(trans, csum_root, head->bytenr,
1856                                               head->num_bytes);
1857                 }
1858         }
1859
1860         btrfs_cleanup_ref_head_accounting(fs_info, delayed_refs, head);
1861
1862         trace_run_delayed_ref_head(fs_info, head, 0);
1863         btrfs_delayed_ref_unlock(head);
1864         btrfs_put_delayed_ref_head(head);
1865         return ret;
1866 }
1867
1868 static struct btrfs_delayed_ref_head *btrfs_obtain_ref_head(
1869                                         struct btrfs_trans_handle *trans)
1870 {
1871         struct btrfs_delayed_ref_root *delayed_refs =
1872                 &trans->transaction->delayed_refs;
1873         struct btrfs_delayed_ref_head *head = NULL;
1874         int ret;
1875
1876         spin_lock(&delayed_refs->lock);
1877         head = btrfs_select_ref_head(delayed_refs);
1878         if (!head) {
1879                 spin_unlock(&delayed_refs->lock);
1880                 return head;
1881         }
1882
1883         /*
1884          * Grab the lock that says we are going to process all the refs for
1885          * this head
1886          */
1887         ret = btrfs_delayed_ref_lock(delayed_refs, head);
1888         spin_unlock(&delayed_refs->lock);
1889
1890         /*
1891          * We may have dropped the spin lock to get the head mutex lock, and
1892          * that might have given someone else time to free the head.  If that's
1893          * true, it has been removed from our list and we can move on.
1894          */
1895         if (ret == -EAGAIN)
1896                 head = ERR_PTR(-EAGAIN);
1897
1898         return head;
1899 }
1900
1901 static int btrfs_run_delayed_refs_for_head(struct btrfs_trans_handle *trans,
1902                                     struct btrfs_delayed_ref_head *locked_ref,
1903                                     unsigned long *run_refs)
1904 {
1905         struct btrfs_fs_info *fs_info = trans->fs_info;
1906         struct btrfs_delayed_ref_root *delayed_refs;
1907         struct btrfs_delayed_extent_op *extent_op;
1908         struct btrfs_delayed_ref_node *ref;
1909         int must_insert_reserved = 0;
1910         int ret;
1911
1912         delayed_refs = &trans->transaction->delayed_refs;
1913
1914         lockdep_assert_held(&locked_ref->mutex);
1915         lockdep_assert_held(&locked_ref->lock);
1916
1917         while ((ref = select_delayed_ref(locked_ref))) {
1918                 if (ref->seq &&
1919                     btrfs_check_delayed_seq(fs_info, ref->seq)) {
1920                         spin_unlock(&locked_ref->lock);
1921                         unselect_delayed_ref_head(delayed_refs, locked_ref);
1922                         return -EAGAIN;
1923                 }
1924
1925                 (*run_refs)++;
1926                 ref->in_tree = 0;
1927                 rb_erase_cached(&ref->ref_node, &locked_ref->ref_tree);
1928                 RB_CLEAR_NODE(&ref->ref_node);
1929                 if (!list_empty(&ref->add_list))
1930                         list_del(&ref->add_list);
1931                 /*
1932                  * When we play the delayed ref, also correct the ref_mod on
1933                  * head
1934                  */
1935                 switch (ref->action) {
1936                 case BTRFS_ADD_DELAYED_REF:
1937                 case BTRFS_ADD_DELAYED_EXTENT:
1938                         locked_ref->ref_mod -= ref->ref_mod;
1939                         break;
1940                 case BTRFS_DROP_DELAYED_REF:
1941                         locked_ref->ref_mod += ref->ref_mod;
1942                         break;
1943                 default:
1944                         WARN_ON(1);
1945                 }
1946                 atomic_dec(&delayed_refs->num_entries);
1947
1948                 /*
1949                  * Record the must_insert_reserved flag before we drop the
1950                  * spin lock.
1951                  */
1952                 must_insert_reserved = locked_ref->must_insert_reserved;
1953                 locked_ref->must_insert_reserved = 0;
1954
1955                 extent_op = locked_ref->extent_op;
1956                 locked_ref->extent_op = NULL;
1957                 spin_unlock(&locked_ref->lock);
1958
1959                 ret = run_one_delayed_ref(trans, ref, extent_op,
1960                                           must_insert_reserved);
1961
1962                 btrfs_free_delayed_extent_op(extent_op);
1963                 if (ret) {
1964                         unselect_delayed_ref_head(delayed_refs, locked_ref);
1965                         btrfs_put_delayed_ref(ref);
1966                         btrfs_debug(fs_info, "run_one_delayed_ref returned %d",
1967                                     ret);
1968                         return ret;
1969                 }
1970
1971                 btrfs_put_delayed_ref(ref);
1972                 cond_resched();
1973
1974                 spin_lock(&locked_ref->lock);
1975                 btrfs_merge_delayed_refs(trans, delayed_refs, locked_ref);
1976         }
1977
1978         return 0;
1979 }
1980
1981 /*
1982  * Returns 0 on success or if called with an already aborted transaction.
1983  * Returns -ENOMEM or -EIO on failure and will abort the transaction.
1984  */
1985 static noinline int __btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
1986                                              unsigned long nr)
1987 {
1988         struct btrfs_fs_info *fs_info = trans->fs_info;
1989         struct btrfs_delayed_ref_root *delayed_refs;
1990         struct btrfs_delayed_ref_head *locked_ref = NULL;
1991         ktime_t start = ktime_get();
1992         int ret;
1993         unsigned long count = 0;
1994         unsigned long actual_count = 0;
1995
1996         delayed_refs = &trans->transaction->delayed_refs;
1997         do {
1998                 if (!locked_ref) {
1999                         locked_ref = btrfs_obtain_ref_head(trans);
2000                         if (IS_ERR_OR_NULL(locked_ref)) {
2001                                 if (PTR_ERR(locked_ref) == -EAGAIN) {
2002                                         continue;
2003                                 } else {
2004                                         break;
2005                                 }
2006                         }
2007                         count++;
2008                 }
2009                 /*
2010                  * We need to try and merge add/drops of the same ref since we
2011                  * can run into issues with relocate dropping the implicit ref
2012                  * and then it being added back again before the drop can
2013                  * finish.  If we merged anything we need to re-loop so we can
2014                  * get a good ref.
2015                  * Or we can get node references of the same type that weren't
2016                  * merged when created due to bumps in the tree mod seq, and
2017                  * we need to merge them to prevent adding an inline extent
2018                  * backref before dropping it (triggering a BUG_ON at
2019                  * insert_inline_extent_backref()).
2020                  */
2021                 spin_lock(&locked_ref->lock);
2022                 btrfs_merge_delayed_refs(trans, delayed_refs, locked_ref);
2023
2024                 ret = btrfs_run_delayed_refs_for_head(trans, locked_ref,
2025                                                       &actual_count);
2026                 if (ret < 0 && ret != -EAGAIN) {
2027                         /*
2028                          * Error, btrfs_run_delayed_refs_for_head already
2029                          * unlocked everything so just bail out
2030                          */
2031                         return ret;
2032                 } else if (!ret) {
2033                         /*
2034                          * Success, perform the usual cleanup of a processed
2035                          * head
2036                          */
2037                         ret = cleanup_ref_head(trans, locked_ref);
2038                         if (ret > 0 ) {
2039                                 /* We dropped our lock, we need to loop. */
2040                                 ret = 0;
2041                                 continue;
2042                         } else if (ret) {
2043                                 return ret;
2044                         }
2045                 }
2046
2047                 /*
2048                  * Either success case or btrfs_run_delayed_refs_for_head
2049                  * returned -EAGAIN, meaning we need to select another head
2050                  */
2051
2052                 locked_ref = NULL;
2053                 cond_resched();
2054         } while ((nr != -1 && count < nr) || locked_ref);
2055
2056         /*
2057          * We don't want to include ref heads since we can have empty ref heads
2058          * and those will drastically skew our runtime down since we just do
2059          * accounting, no actual extent tree updates.
2060          */
2061         if (actual_count > 0) {
2062                 u64 runtime = ktime_to_ns(ktime_sub(ktime_get(), start));
2063                 u64 avg;
2064
2065                 /*
2066                  * We weigh the current average higher than our current runtime
2067                  * to avoid large swings in the average.
2068                  */
2069                 spin_lock(&delayed_refs->lock);
2070                 avg = fs_info->avg_delayed_ref_runtime * 3 + runtime;
2071                 fs_info->avg_delayed_ref_runtime = avg >> 2;    /* div by 4 */
2072                 spin_unlock(&delayed_refs->lock);
2073         }
2074         return 0;
2075 }
2076
2077 #ifdef SCRAMBLE_DELAYED_REFS
2078 /*
2079  * Normally delayed refs get processed in ascending bytenr order. This
2080  * correlates in most cases to the order added. To expose dependencies on this
2081  * order, we start to process the tree in the middle instead of the beginning
2082  */
2083 static u64 find_middle(struct rb_root *root)
2084 {
2085         struct rb_node *n = root->rb_node;
2086         struct btrfs_delayed_ref_node *entry;
2087         int alt = 1;
2088         u64 middle;
2089         u64 first = 0, last = 0;
2090
2091         n = rb_first(root);
2092         if (n) {
2093                 entry = rb_entry(n, struct btrfs_delayed_ref_node, rb_node);
2094                 first = entry->bytenr;
2095         }
2096         n = rb_last(root);
2097         if (n) {
2098                 entry = rb_entry(n, struct btrfs_delayed_ref_node, rb_node);
2099                 last = entry->bytenr;
2100         }
2101         n = root->rb_node;
2102
2103         while (n) {
2104                 entry = rb_entry(n, struct btrfs_delayed_ref_node, rb_node);
2105                 WARN_ON(!entry->in_tree);
2106
2107                 middle = entry->bytenr;
2108
2109                 if (alt)
2110                         n = n->rb_left;
2111                 else
2112                         n = n->rb_right;
2113
2114                 alt = 1 - alt;
2115         }
2116         return middle;
2117 }
2118 #endif
2119
2120 /*
2121  * this starts processing the delayed reference count updates and
2122  * extent insertions we have queued up so far.  count can be
2123  * 0, which means to process everything in the tree at the start
2124  * of the run (but not newly added entries), or it can be some target
2125  * number you'd like to process.
2126  *
2127  * Returns 0 on success or if called with an aborted transaction
2128  * Returns <0 on error and aborts the transaction
2129  */
2130 int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans,
2131                            unsigned long count)
2132 {
2133         struct btrfs_fs_info *fs_info = trans->fs_info;
2134         struct rb_node *node;
2135         struct btrfs_delayed_ref_root *delayed_refs;
2136         struct btrfs_delayed_ref_head *head;
2137         int ret;
2138         int run_all = count == (unsigned long)-1;
2139
2140         /* We'll clean this up in btrfs_cleanup_transaction */
2141         if (TRANS_ABORTED(trans))
2142                 return 0;
2143
2144         if (test_bit(BTRFS_FS_CREATING_FREE_SPACE_TREE, &fs_info->flags))
2145                 return 0;
2146
2147         delayed_refs = &trans->transaction->delayed_refs;
2148         if (count == 0)
2149                 count = delayed_refs->num_heads_ready;
2150
2151 again:
2152 #ifdef SCRAMBLE_DELAYED_REFS
2153         delayed_refs->run_delayed_start = find_middle(&delayed_refs->root);
2154 #endif
2155         ret = __btrfs_run_delayed_refs(trans, count);
2156         if (ret < 0) {
2157                 btrfs_abort_transaction(trans, ret);
2158                 return ret;
2159         }
2160
2161         if (run_all) {
2162                 btrfs_create_pending_block_groups(trans);
2163
2164                 spin_lock(&delayed_refs->lock);
2165                 node = rb_first_cached(&delayed_refs->href_root);
2166                 if (!node) {
2167                         spin_unlock(&delayed_refs->lock);
2168                         goto out;
2169                 }
2170                 head = rb_entry(node, struct btrfs_delayed_ref_head,
2171                                 href_node);
2172                 refcount_inc(&head->refs);
2173                 spin_unlock(&delayed_refs->lock);
2174
2175                 /* Mutex was contended, block until it's released and retry. */
2176                 mutex_lock(&head->mutex);
2177                 mutex_unlock(&head->mutex);
2178
2179                 btrfs_put_delayed_ref_head(head);
2180                 cond_resched();
2181                 goto again;
2182         }
2183 out:
2184         return 0;
2185 }
2186
2187 int btrfs_set_disk_extent_flags(struct btrfs_trans_handle *trans,
2188                                 struct extent_buffer *eb, u64 flags,
2189                                 int level)
2190 {
2191         struct btrfs_delayed_extent_op *extent_op;
2192         int ret;
2193
2194         extent_op = btrfs_alloc_delayed_extent_op();
2195         if (!extent_op)
2196                 return -ENOMEM;
2197
2198         extent_op->flags_to_set = flags;
2199         extent_op->update_flags = true;
2200         extent_op->update_key = false;
2201         extent_op->level = level;
2202
2203         ret = btrfs_add_delayed_extent_op(trans, eb->start, eb->len, extent_op);
2204         if (ret)
2205                 btrfs_free_delayed_extent_op(extent_op);
2206         return ret;
2207 }
2208
2209 static noinline int check_delayed_ref(struct btrfs_root *root,
2210                                       struct btrfs_path *path,
2211                                       u64 objectid, u64 offset, u64 bytenr)
2212 {
2213         struct btrfs_delayed_ref_head *head;
2214         struct btrfs_delayed_ref_node *ref;
2215         struct btrfs_delayed_data_ref *data_ref;
2216         struct btrfs_delayed_ref_root *delayed_refs;
2217         struct btrfs_transaction *cur_trans;
2218         struct rb_node *node;
2219         int ret = 0;
2220
2221         spin_lock(&root->fs_info->trans_lock);
2222         cur_trans = root->fs_info->running_transaction;
2223         if (cur_trans)
2224                 refcount_inc(&cur_trans->use_count);
2225         spin_unlock(&root->fs_info->trans_lock);
2226         if (!cur_trans)
2227                 return 0;
2228
2229         delayed_refs = &cur_trans->delayed_refs;
2230         spin_lock(&delayed_refs->lock);
2231         head = btrfs_find_delayed_ref_head(delayed_refs, bytenr);
2232         if (!head) {
2233                 spin_unlock(&delayed_refs->lock);
2234                 btrfs_put_transaction(cur_trans);
2235                 return 0;
2236         }
2237
2238         if (!mutex_trylock(&head->mutex)) {
2239                 refcount_inc(&head->refs);
2240                 spin_unlock(&delayed_refs->lock);
2241
2242                 btrfs_release_path(path);
2243
2244                 /*
2245                  * Mutex was contended, block until it's released and let
2246                  * caller try again
2247                  */
2248                 mutex_lock(&head->mutex);
2249                 mutex_unlock(&head->mutex);
2250                 btrfs_put_delayed_ref_head(head);
2251                 btrfs_put_transaction(cur_trans);
2252                 return -EAGAIN;
2253         }
2254         spin_unlock(&delayed_refs->lock);
2255
2256         spin_lock(&head->lock);
2257         /*
2258          * XXX: We should replace this with a proper search function in the
2259          * future.
2260          */
2261         for (node = rb_first_cached(&head->ref_tree); node;
2262              node = rb_next(node)) {
2263                 ref = rb_entry(node, struct btrfs_delayed_ref_node, ref_node);
2264                 /* If it's a shared ref we know a cross reference exists */
2265                 if (ref->type != BTRFS_EXTENT_DATA_REF_KEY) {
2266                         ret = 1;
2267                         break;
2268                 }
2269
2270                 data_ref = btrfs_delayed_node_to_data_ref(ref);
2271
2272                 /*
2273                  * If our ref doesn't match the one we're currently looking at
2274                  * then we have a cross reference.
2275                  */
2276                 if (data_ref->root != root->root_key.objectid ||
2277                     data_ref->objectid != objectid ||
2278                     data_ref->offset != offset) {
2279                         ret = 1;
2280                         break;
2281                 }
2282         }
2283         spin_unlock(&head->lock);
2284         mutex_unlock(&head->mutex);
2285         btrfs_put_transaction(cur_trans);
2286         return ret;
2287 }
2288
2289 static noinline int check_committed_ref(struct btrfs_root *root,
2290                                         struct btrfs_path *path,
2291                                         u64 objectid, u64 offset, u64 bytenr,
2292                                         bool strict)
2293 {
2294         struct btrfs_fs_info *fs_info = root->fs_info;
2295         struct btrfs_root *extent_root = btrfs_extent_root(fs_info, bytenr);
2296         struct extent_buffer *leaf;
2297         struct btrfs_extent_data_ref *ref;
2298         struct btrfs_extent_inline_ref *iref;
2299         struct btrfs_extent_item *ei;
2300         struct btrfs_key key;
2301         u32 item_size;
2302         int type;
2303         int ret;
2304
2305         key.objectid = bytenr;
2306         key.offset = (u64)-1;
2307         key.type = BTRFS_EXTENT_ITEM_KEY;
2308
2309         ret = btrfs_search_slot(NULL, extent_root, &key, path, 0, 0);
2310         if (ret < 0)
2311                 goto out;
2312         BUG_ON(ret == 0); /* Corruption */
2313
2314         ret = -ENOENT;
2315         if (path->slots[0] == 0)
2316                 goto out;
2317
2318         path->slots[0]--;
2319         leaf = path->nodes[0];
2320         btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
2321
2322         if (key.objectid != bytenr || key.type != BTRFS_EXTENT_ITEM_KEY)
2323                 goto out;
2324
2325         ret = 1;
2326         item_size = btrfs_item_size(leaf, path->slots[0]);
2327         ei = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_extent_item);
2328
2329         /* If extent item has more than 1 inline ref then it's shared */
2330         if (item_size != sizeof(*ei) +
2331             btrfs_extent_inline_ref_size(BTRFS_EXTENT_DATA_REF_KEY))
2332                 goto out;
2333
2334         /*
2335          * If extent created before last snapshot => it's shared unless the
2336          * snapshot has been deleted. Use the heuristic if strict is false.
2337          */
2338         if (!strict &&
2339             (btrfs_extent_generation(leaf, ei) <=
2340              btrfs_root_last_snapshot(&root->root_item)))
2341                 goto out;
2342
2343         iref = (struct btrfs_extent_inline_ref *)(ei + 1);
2344
2345         /* If this extent has SHARED_DATA_REF then it's shared */
2346         type = btrfs_get_extent_inline_ref_type(leaf, iref, BTRFS_REF_TYPE_DATA);
2347         if (type != BTRFS_EXTENT_DATA_REF_KEY)
2348                 goto out;
2349
2350         ref = (struct btrfs_extent_data_ref *)(&iref->offset);
2351         if (btrfs_extent_refs(leaf, ei) !=
2352             btrfs_extent_data_ref_count(leaf, ref) ||
2353             btrfs_extent_data_ref_root(leaf, ref) !=
2354             root->root_key.objectid ||
2355             btrfs_extent_data_ref_objectid(leaf, ref) != objectid ||
2356             btrfs_extent_data_ref_offset(leaf, ref) != offset)
2357                 goto out;
2358
2359         ret = 0;
2360 out:
2361         return ret;
2362 }
2363
2364 int btrfs_cross_ref_exist(struct btrfs_root *root, u64 objectid, u64 offset,
2365                           u64 bytenr, bool strict, struct btrfs_path *path)
2366 {
2367         int ret;
2368
2369         do {
2370                 ret = check_committed_ref(root, path, objectid,
2371                                           offset, bytenr, strict);
2372                 if (ret && ret != -ENOENT)
2373                         goto out;
2374
2375                 ret = check_delayed_ref(root, path, objectid, offset, bytenr);
2376         } while (ret == -EAGAIN);
2377
2378 out:
2379         btrfs_release_path(path);
2380         if (btrfs_is_data_reloc_root(root))
2381                 WARN_ON(ret > 0);
2382         return ret;
2383 }
2384
2385 static int __btrfs_mod_ref(struct btrfs_trans_handle *trans,
2386                            struct btrfs_root *root,
2387                            struct extent_buffer *buf,
2388                            int full_backref, int inc)
2389 {
2390         struct btrfs_fs_info *fs_info = root->fs_info;
2391         u64 bytenr;
2392         u64 num_bytes;
2393         u64 parent;
2394         u64 ref_root;
2395         u32 nritems;
2396         struct btrfs_key key;
2397         struct btrfs_file_extent_item *fi;
2398         struct btrfs_ref generic_ref = { 0 };
2399         bool for_reloc = btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC);
2400         int i;
2401         int action;
2402         int level;
2403         int ret = 0;
2404
2405         if (btrfs_is_testing(fs_info))
2406                 return 0;
2407
2408         ref_root = btrfs_header_owner(buf);
2409         nritems = btrfs_header_nritems(buf);
2410         level = btrfs_header_level(buf);
2411
2412         if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state) && level == 0)
2413                 return 0;
2414
2415         if (full_backref)
2416                 parent = buf->start;
2417         else
2418                 parent = 0;
2419         if (inc)
2420                 action = BTRFS_ADD_DELAYED_REF;
2421         else
2422                 action = BTRFS_DROP_DELAYED_REF;
2423
2424         for (i = 0; i < nritems; i++) {
2425                 if (level == 0) {
2426                         btrfs_item_key_to_cpu(buf, &key, i);
2427                         if (key.type != BTRFS_EXTENT_DATA_KEY)
2428                                 continue;
2429                         fi = btrfs_item_ptr(buf, i,
2430                                             struct btrfs_file_extent_item);
2431                         if (btrfs_file_extent_type(buf, fi) ==
2432                             BTRFS_FILE_EXTENT_INLINE)
2433                                 continue;
2434                         bytenr = btrfs_file_extent_disk_bytenr(buf, fi);
2435                         if (bytenr == 0)
2436                                 continue;
2437
2438                         num_bytes = btrfs_file_extent_disk_num_bytes(buf, fi);
2439                         key.offset -= btrfs_file_extent_offset(buf, fi);
2440                         btrfs_init_generic_ref(&generic_ref, action, bytenr,
2441                                                num_bytes, parent);
2442                         btrfs_init_data_ref(&generic_ref, ref_root, key.objectid,
2443                                             key.offset, root->root_key.objectid,
2444                                             for_reloc);
2445                         if (inc)
2446                                 ret = btrfs_inc_extent_ref(trans, &generic_ref);
2447                         else
2448                                 ret = btrfs_free_extent(trans, &generic_ref);
2449                         if (ret)
2450                                 goto fail;
2451                 } else {
2452                         bytenr = btrfs_node_blockptr(buf, i);
2453                         num_bytes = fs_info->nodesize;
2454                         btrfs_init_generic_ref(&generic_ref, action, bytenr,
2455                                                num_bytes, parent);
2456                         btrfs_init_tree_ref(&generic_ref, level - 1, ref_root,
2457                                             root->root_key.objectid, for_reloc);
2458                         if (inc)
2459                                 ret = btrfs_inc_extent_ref(trans, &generic_ref);
2460                         else
2461                                 ret = btrfs_free_extent(trans, &generic_ref);
2462                         if (ret)
2463                                 goto fail;
2464                 }
2465         }
2466         return 0;
2467 fail:
2468         return ret;
2469 }
2470
2471 int btrfs_inc_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
2472                   struct extent_buffer *buf, int full_backref)
2473 {
2474         return __btrfs_mod_ref(trans, root, buf, full_backref, 1);
2475 }
2476
2477 int btrfs_dec_ref(struct btrfs_trans_handle *trans, struct btrfs_root *root,
2478                   struct extent_buffer *buf, int full_backref)
2479 {
2480         return __btrfs_mod_ref(trans, root, buf, full_backref, 0);
2481 }
2482
2483 static u64 get_alloc_profile_by_root(struct btrfs_root *root, int data)
2484 {
2485         struct btrfs_fs_info *fs_info = root->fs_info;
2486         u64 flags;
2487         u64 ret;
2488
2489         if (data)
2490                 flags = BTRFS_BLOCK_GROUP_DATA;
2491         else if (root == fs_info->chunk_root)
2492                 flags = BTRFS_BLOCK_GROUP_SYSTEM;
2493         else
2494                 flags = BTRFS_BLOCK_GROUP_METADATA;
2495
2496         ret = btrfs_get_alloc_profile(fs_info, flags);
2497         return ret;
2498 }
2499
2500 static u64 first_logical_byte(struct btrfs_fs_info *fs_info)
2501 {
2502         struct rb_node *leftmost;
2503         u64 bytenr = 0;
2504
2505         read_lock(&fs_info->block_group_cache_lock);
2506         /* Get the block group with the lowest logical start address. */
2507         leftmost = rb_first_cached(&fs_info->block_group_cache_tree);
2508         if (leftmost) {
2509                 struct btrfs_block_group *bg;
2510
2511                 bg = rb_entry(leftmost, struct btrfs_block_group, cache_node);
2512                 bytenr = bg->start;
2513         }
2514         read_unlock(&fs_info->block_group_cache_lock);
2515
2516         return bytenr;
2517 }
2518
2519 static int pin_down_extent(struct btrfs_trans_handle *trans,
2520                            struct btrfs_block_group *cache,
2521                            u64 bytenr, u64 num_bytes, int reserved)
2522 {
2523         struct btrfs_fs_info *fs_info = cache->fs_info;
2524
2525         spin_lock(&cache->space_info->lock);
2526         spin_lock(&cache->lock);
2527         cache->pinned += num_bytes;
2528         btrfs_space_info_update_bytes_pinned(fs_info, cache->space_info,
2529                                              num_bytes);
2530         if (reserved) {
2531                 cache->reserved -= num_bytes;
2532                 cache->space_info->bytes_reserved -= num_bytes;
2533         }
2534         spin_unlock(&cache->lock);
2535         spin_unlock(&cache->space_info->lock);
2536
2537         set_extent_dirty(&trans->transaction->pinned_extents, bytenr,
2538                          bytenr + num_bytes - 1, GFP_NOFS | __GFP_NOFAIL);
2539         return 0;
2540 }
2541
2542 int btrfs_pin_extent(struct btrfs_trans_handle *trans,
2543                      u64 bytenr, u64 num_bytes, int reserved)
2544 {
2545         struct btrfs_block_group *cache;
2546
2547         cache = btrfs_lookup_block_group(trans->fs_info, bytenr);
2548         BUG_ON(!cache); /* Logic error */
2549
2550         pin_down_extent(trans, cache, bytenr, num_bytes, reserved);
2551
2552         btrfs_put_block_group(cache);
2553         return 0;
2554 }
2555
2556 /*
2557  * this function must be called within transaction
2558  */
2559 int btrfs_pin_extent_for_log_replay(struct btrfs_trans_handle *trans,
2560                                     u64 bytenr, u64 num_bytes)
2561 {
2562         struct btrfs_block_group *cache;
2563         int ret;
2564
2565         cache = btrfs_lookup_block_group(trans->fs_info, bytenr);
2566         if (!cache)
2567                 return -EINVAL;
2568
2569         /*
2570          * pull in the free space cache (if any) so that our pin
2571          * removes the free space from the cache.  We have load_only set
2572          * to one because the slow code to read in the free extents does check
2573          * the pinned extents.
2574          */
2575         btrfs_cache_block_group(cache, 1);
2576         /*
2577          * Make sure we wait until the cache is completely built in case it is
2578          * missing or is invalid and therefore needs to be rebuilt.
2579          */
2580         ret = btrfs_wait_block_group_cache_done(cache);
2581         if (ret)
2582                 goto out;
2583
2584         pin_down_extent(trans, cache, bytenr, num_bytes, 0);
2585
2586         /* remove us from the free space cache (if we're there at all) */
2587         ret = btrfs_remove_free_space(cache, bytenr, num_bytes);
2588 out:
2589         btrfs_put_block_group(cache);
2590         return ret;
2591 }
2592
2593 static int __exclude_logged_extent(struct btrfs_fs_info *fs_info,
2594                                    u64 start, u64 num_bytes)
2595 {
2596         int ret;
2597         struct btrfs_block_group *block_group;
2598
2599         block_group = btrfs_lookup_block_group(fs_info, start);
2600         if (!block_group)
2601                 return -EINVAL;
2602
2603         btrfs_cache_block_group(block_group, 1);
2604         /*
2605          * Make sure we wait until the cache is completely built in case it is
2606          * missing or is invalid and therefore needs to be rebuilt.
2607          */
2608         ret = btrfs_wait_block_group_cache_done(block_group);
2609         if (ret)
2610                 goto out;
2611
2612         ret = btrfs_remove_free_space(block_group, start, num_bytes);
2613 out:
2614         btrfs_put_block_group(block_group);
2615         return ret;
2616 }
2617
2618 int btrfs_exclude_logged_extents(struct extent_buffer *eb)
2619 {
2620         struct btrfs_fs_info *fs_info = eb->fs_info;
2621         struct btrfs_file_extent_item *item;
2622         struct btrfs_key key;
2623         int found_type;
2624         int i;
2625         int ret = 0;
2626
2627         if (!btrfs_fs_incompat(fs_info, MIXED_GROUPS))
2628                 return 0;
2629
2630         for (i = 0; i < btrfs_header_nritems(eb); i++) {
2631                 btrfs_item_key_to_cpu(eb, &key, i);
2632                 if (key.type != BTRFS_EXTENT_DATA_KEY)
2633                         continue;
2634                 item = btrfs_item_ptr(eb, i, struct btrfs_file_extent_item);
2635                 found_type = btrfs_file_extent_type(eb, item);
2636                 if (found_type == BTRFS_FILE_EXTENT_INLINE)
2637                         continue;
2638                 if (btrfs_file_extent_disk_bytenr(eb, item) == 0)
2639                         continue;
2640                 key.objectid = btrfs_file_extent_disk_bytenr(eb, item);
2641                 key.offset = btrfs_file_extent_disk_num_bytes(eb, item);
2642                 ret = __exclude_logged_extent(fs_info, key.objectid, key.offset);
2643                 if (ret)
2644                         break;
2645         }
2646
2647         return ret;
2648 }
2649
2650 static void
2651 btrfs_inc_block_group_reservations(struct btrfs_block_group *bg)
2652 {
2653         atomic_inc(&bg->reservations);
2654 }
2655
2656 /*
2657  * Returns the free cluster for the given space info and sets empty_cluster to
2658  * what it should be based on the mount options.
2659  */
2660 static struct btrfs_free_cluster *
2661 fetch_cluster_info(struct btrfs_fs_info *fs_info,
2662                    struct btrfs_space_info *space_info, u64 *empty_cluster)
2663 {
2664         struct btrfs_free_cluster *ret = NULL;
2665
2666         *empty_cluster = 0;
2667         if (btrfs_mixed_space_info(space_info))
2668                 return ret;
2669
2670         if (space_info->flags & BTRFS_BLOCK_GROUP_METADATA) {
2671                 ret = &fs_info->meta_alloc_cluster;
2672                 if (btrfs_test_opt(fs_info, SSD))
2673                         *empty_cluster = SZ_2M;
2674                 else
2675                         *empty_cluster = SZ_64K;
2676         } else if ((space_info->flags & BTRFS_BLOCK_GROUP_DATA) &&
2677                    btrfs_test_opt(fs_info, SSD_SPREAD)) {
2678                 *empty_cluster = SZ_2M;
2679                 ret = &fs_info->data_alloc_cluster;
2680         }
2681
2682         return ret;
2683 }
2684
2685 static int unpin_extent_range(struct btrfs_fs_info *fs_info,
2686                               u64 start, u64 end,
2687                               const bool return_free_space)
2688 {
2689         struct btrfs_block_group *cache = NULL;
2690         struct btrfs_space_info *space_info;
2691         struct btrfs_block_rsv *global_rsv = &fs_info->global_block_rsv;
2692         struct btrfs_free_cluster *cluster = NULL;
2693         u64 len;
2694         u64 total_unpinned = 0;
2695         u64 empty_cluster = 0;
2696         bool readonly;
2697
2698         while (start <= end) {
2699                 readonly = false;
2700                 if (!cache ||
2701                     start >= cache->start + cache->length) {
2702                         if (cache)
2703                                 btrfs_put_block_group(cache);
2704                         total_unpinned = 0;
2705                         cache = btrfs_lookup_block_group(fs_info, start);
2706                         BUG_ON(!cache); /* Logic error */
2707
2708                         cluster = fetch_cluster_info(fs_info,
2709                                                      cache->space_info,
2710                                                      &empty_cluster);
2711                         empty_cluster <<= 1;
2712                 }
2713
2714                 len = cache->start + cache->length - start;
2715                 len = min(len, end + 1 - start);
2716
2717                 down_read(&fs_info->commit_root_sem);
2718                 if (start < cache->last_byte_to_unpin && return_free_space) {
2719                         u64 add_len = min(len, cache->last_byte_to_unpin - start);
2720
2721                         btrfs_add_free_space(cache, start, add_len);
2722                 }
2723                 up_read(&fs_info->commit_root_sem);
2724
2725                 start += len;
2726                 total_unpinned += len;
2727                 space_info = cache->space_info;
2728
2729                 /*
2730                  * If this space cluster has been marked as fragmented and we've
2731                  * unpinned enough in this block group to potentially allow a
2732                  * cluster to be created inside of it go ahead and clear the
2733                  * fragmented check.
2734                  */
2735                 if (cluster && cluster->fragmented &&
2736                     total_unpinned > empty_cluster) {
2737                         spin_lock(&cluster->lock);
2738                         cluster->fragmented = 0;
2739                         spin_unlock(&cluster->lock);
2740                 }
2741
2742                 spin_lock(&space_info->lock);
2743                 spin_lock(&cache->lock);
2744                 cache->pinned -= len;
2745                 btrfs_space_info_update_bytes_pinned(fs_info, space_info, -len);
2746                 space_info->max_extent_size = 0;
2747                 if (cache->ro) {
2748                         space_info->bytes_readonly += len;
2749                         readonly = true;
2750                 } else if (btrfs_is_zoned(fs_info)) {
2751                         /* Need reset before reusing in a zoned block group */
2752                         space_info->bytes_zone_unusable += len;
2753                         readonly = true;
2754                 }
2755                 spin_unlock(&cache->lock);
2756                 if (!readonly && return_free_space &&
2757                     global_rsv->space_info == space_info) {
2758                         spin_lock(&global_rsv->lock);
2759                         if (!global_rsv->full) {
2760                                 u64 to_add = min(len, global_rsv->size -
2761                                                       global_rsv->reserved);
2762
2763                                 global_rsv->reserved += to_add;
2764                                 btrfs_space_info_update_bytes_may_use(fs_info,
2765                                                 space_info, to_add);
2766                                 if (global_rsv->reserved >= global_rsv->size)
2767                                         global_rsv->full = 1;
2768                                 len -= to_add;
2769                         }
2770                         spin_unlock(&global_rsv->lock);
2771                 }
2772                 /* Add to any tickets we may have */
2773                 if (!readonly && return_free_space && len)
2774                         btrfs_try_granting_tickets(fs_info, space_info);
2775                 spin_unlock(&space_info->lock);
2776         }
2777
2778         if (cache)
2779                 btrfs_put_block_group(cache);
2780         return 0;
2781 }
2782
2783 int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans)
2784 {
2785         struct btrfs_fs_info *fs_info = trans->fs_info;
2786         struct btrfs_block_group *block_group, *tmp;
2787         struct list_head *deleted_bgs;
2788         struct extent_io_tree *unpin;
2789         u64 start;
2790         u64 end;
2791         int ret;
2792
2793         unpin = &trans->transaction->pinned_extents;
2794
2795         while (!TRANS_ABORTED(trans)) {
2796                 struct extent_state *cached_state = NULL;
2797
2798                 mutex_lock(&fs_info->unused_bg_unpin_mutex);
2799                 ret = find_first_extent_bit(unpin, 0, &start, &end,
2800                                             EXTENT_DIRTY, &cached_state);
2801                 if (ret) {
2802                         mutex_unlock(&fs_info->unused_bg_unpin_mutex);
2803                         break;
2804                 }
2805
2806                 if (btrfs_test_opt(fs_info, DISCARD_SYNC))
2807                         ret = btrfs_discard_extent(fs_info, start,
2808                                                    end + 1 - start, NULL);
2809
2810                 clear_extent_dirty(unpin, start, end, &cached_state);
2811                 unpin_extent_range(fs_info, start, end, true);
2812                 mutex_unlock(&fs_info->unused_bg_unpin_mutex);
2813                 free_extent_state(cached_state);
2814                 cond_resched();
2815         }
2816
2817         if (btrfs_test_opt(fs_info, DISCARD_ASYNC)) {
2818                 btrfs_discard_calc_delay(&fs_info->discard_ctl);
2819                 btrfs_discard_schedule_work(&fs_info->discard_ctl, true);
2820         }
2821
2822         /*
2823          * Transaction is finished.  We don't need the lock anymore.  We
2824          * do need to clean up the block groups in case of a transaction
2825          * abort.
2826          */
2827         deleted_bgs = &trans->transaction->deleted_bgs;
2828         list_for_each_entry_safe(block_group, tmp, deleted_bgs, bg_list) {
2829                 u64 trimmed = 0;
2830
2831                 ret = -EROFS;
2832                 if (!TRANS_ABORTED(trans))
2833                         ret = btrfs_discard_extent(fs_info,
2834                                                    block_group->start,
2835                                                    block_group->length,
2836                                                    &trimmed);
2837
2838                 list_del_init(&block_group->bg_list);
2839                 btrfs_unfreeze_block_group(block_group);
2840                 btrfs_put_block_group(block_group);
2841
2842                 if (ret) {
2843                         const char *errstr = btrfs_decode_error(ret);
2844                         btrfs_warn(fs_info,
2845                            "discard failed while removing blockgroup: errno=%d %s",
2846                                    ret, errstr);
2847                 }
2848         }
2849
2850         return 0;
2851 }
2852
2853 static int do_free_extent_accounting(struct btrfs_trans_handle *trans,
2854                                      u64 bytenr, u64 num_bytes, bool is_data)
2855 {
2856         int ret;
2857
2858         if (is_data) {
2859                 struct btrfs_root *csum_root;
2860
2861                 csum_root = btrfs_csum_root(trans->fs_info, bytenr);
2862                 ret = btrfs_del_csums(trans, csum_root, bytenr, num_bytes);
2863                 if (ret) {
2864                         btrfs_abort_transaction(trans, ret);
2865                         return ret;
2866                 }
2867         }
2868
2869         ret = add_to_free_space_tree(trans, bytenr, num_bytes);
2870         if (ret) {
2871                 btrfs_abort_transaction(trans, ret);
2872                 return ret;
2873         }
2874
2875         ret = btrfs_update_block_group(trans, bytenr, num_bytes, false);
2876         if (ret)
2877                 btrfs_abort_transaction(trans, ret);
2878
2879         return ret;
2880 }
2881
2882 /*
2883  * Drop one or more refs of @node.
2884  *
2885  * 1. Locate the extent refs.
2886  *    It's either inline in EXTENT/METADATA_ITEM or in keyed SHARED_* item.
2887  *    Locate it, then reduce the refs number or remove the ref line completely.
2888  *
2889  * 2. Update the refs count in EXTENT/METADATA_ITEM
2890  *
2891  * Inline backref case:
2892  *
2893  * in extent tree we have:
2894  *
2895  *      item 0 key (13631488 EXTENT_ITEM 1048576) itemoff 16201 itemsize 82
2896  *              refs 2 gen 6 flags DATA
2897  *              extent data backref root FS_TREE objectid 258 offset 0 count 1
2898  *              extent data backref root FS_TREE objectid 257 offset 0 count 1
2899  *
2900  * This function gets called with:
2901  *
2902  *    node->bytenr = 13631488
2903  *    node->num_bytes = 1048576
2904  *    root_objectid = FS_TREE
2905  *    owner_objectid = 257
2906  *    owner_offset = 0
2907  *    refs_to_drop = 1
2908  *
2909  * Then we should get some like:
2910  *
2911  *      item 0 key (13631488 EXTENT_ITEM 1048576) itemoff 16201 itemsize 82
2912  *              refs 1 gen 6 flags DATA
2913  *              extent data backref root FS_TREE objectid 258 offset 0 count 1
2914  *
2915  * Keyed backref case:
2916  *
2917  * in extent tree we have:
2918  *
2919  *      item 0 key (13631488 EXTENT_ITEM 1048576) itemoff 3971 itemsize 24
2920  *              refs 754 gen 6 flags DATA
2921  *      [...]
2922  *      item 2 key (13631488 EXTENT_DATA_REF <HASH>) itemoff 3915 itemsize 28
2923  *              extent data backref root FS_TREE objectid 866 offset 0 count 1
2924  *
2925  * This function get called with:
2926  *
2927  *    node->bytenr = 13631488
2928  *    node->num_bytes = 1048576
2929  *    root_objectid = FS_TREE
2930  *    owner_objectid = 866
2931  *    owner_offset = 0
2932  *    refs_to_drop = 1
2933  *
2934  * Then we should get some like:
2935  *
2936  *      item 0 key (13631488 EXTENT_ITEM 1048576) itemoff 3971 itemsize 24
2937  *              refs 753 gen 6 flags DATA
2938  *
2939  * And that (13631488 EXTENT_DATA_REF <HASH>) gets removed.
2940  */
2941 static int __btrfs_free_extent(struct btrfs_trans_handle *trans,
2942                                struct btrfs_delayed_ref_node *node, u64 parent,
2943                                u64 root_objectid, u64 owner_objectid,
2944                                u64 owner_offset, int refs_to_drop,
2945                                struct btrfs_delayed_extent_op *extent_op)
2946 {
2947         struct btrfs_fs_info *info = trans->fs_info;
2948         struct btrfs_key key;
2949         struct btrfs_path *path;
2950         struct btrfs_root *extent_root;
2951         struct extent_buffer *leaf;
2952         struct btrfs_extent_item *ei;
2953         struct btrfs_extent_inline_ref *iref;
2954         int ret;
2955         int is_data;
2956         int extent_slot = 0;
2957         int found_extent = 0;
2958         int num_to_del = 1;
2959         u32 item_size;
2960         u64 refs;
2961         u64 bytenr = node->bytenr;
2962         u64 num_bytes = node->num_bytes;
2963         bool skinny_metadata = btrfs_fs_incompat(info, SKINNY_METADATA);
2964
2965         extent_root = btrfs_extent_root(info, bytenr);
2966         ASSERT(extent_root);
2967
2968         path = btrfs_alloc_path();
2969         if (!path)
2970                 return -ENOMEM;
2971
2972         is_data = owner_objectid >= BTRFS_FIRST_FREE_OBJECTID;
2973
2974         if (!is_data && refs_to_drop != 1) {
2975                 btrfs_crit(info,
2976 "invalid refs_to_drop, dropping more than 1 refs for tree block %llu refs_to_drop %u",
2977                            node->bytenr, refs_to_drop);
2978                 ret = -EINVAL;
2979                 btrfs_abort_transaction(trans, ret);
2980                 goto out;
2981         }
2982
2983         if (is_data)
2984                 skinny_metadata = false;
2985
2986         ret = lookup_extent_backref(trans, path, &iref, bytenr, num_bytes,
2987                                     parent, root_objectid, owner_objectid,
2988                                     owner_offset);
2989         if (ret == 0) {
2990                 /*
2991                  * Either the inline backref or the SHARED_DATA_REF/
2992                  * SHARED_BLOCK_REF is found
2993                  *
2994                  * Here is a quick path to locate EXTENT/METADATA_ITEM.
2995                  * It's possible the EXTENT/METADATA_ITEM is near current slot.
2996                  */
2997                 extent_slot = path->slots[0];
2998                 while (extent_slot >= 0) {
2999                         btrfs_item_key_to_cpu(path->nodes[0], &key,
3000                                               extent_slot);
3001                         if (key.objectid != bytenr)
3002                                 break;
3003                         if (key.type == BTRFS_EXTENT_ITEM_KEY &&
3004                             key.offset == num_bytes) {
3005                                 found_extent = 1;
3006                                 break;
3007                         }
3008                         if (key.type == BTRFS_METADATA_ITEM_KEY &&
3009                             key.offset == owner_objectid) {
3010                                 found_extent = 1;
3011                                 break;
3012                         }
3013
3014                         /* Quick path didn't find the EXTEMT/METADATA_ITEM */
3015                         if (path->slots[0] - extent_slot > 5)
3016                                 break;
3017                         extent_slot--;
3018                 }
3019
3020                 if (!found_extent) {
3021                         if (iref) {
3022                                 btrfs_crit(info,
3023 "invalid iref, no EXTENT/METADATA_ITEM found but has inline extent ref");
3024                                 btrfs_abort_transaction(trans, -EUCLEAN);
3025                                 goto err_dump;
3026                         }
3027                         /* Must be SHARED_* item, remove the backref first */
3028                         ret = remove_extent_backref(trans, extent_root, path,
3029                                                     NULL, refs_to_drop, is_data);
3030                         if (ret) {
3031                                 btrfs_abort_transaction(trans, ret);
3032                                 goto out;
3033                         }
3034                         btrfs_release_path(path);
3035
3036                         /* Slow path to locate EXTENT/METADATA_ITEM */
3037                         key.objectid = bytenr;
3038                         key.type = BTRFS_EXTENT_ITEM_KEY;
3039                         key.offset = num_bytes;
3040
3041                         if (!is_data && skinny_metadata) {
3042                                 key.type = BTRFS_METADATA_ITEM_KEY;
3043                                 key.offset = owner_objectid;
3044                         }
3045
3046                         ret = btrfs_search_slot(trans, extent_root,
3047                                                 &key, path, -1, 1);
3048                         if (ret > 0 && skinny_metadata && path->slots[0]) {
3049                                 /*
3050                                  * Couldn't find our skinny metadata item,
3051                                  * see if we have ye olde extent item.
3052                                  */
3053                                 path->slots[0]--;
3054                                 btrfs_item_key_to_cpu(path->nodes[0], &key,
3055                                                       path->slots[0]);
3056                                 if (key.objectid == bytenr &&
3057                                     key.type == BTRFS_EXTENT_ITEM_KEY &&
3058                                     key.offset == num_bytes)
3059                                         ret = 0;
3060                         }
3061
3062                         if (ret > 0 && skinny_metadata) {
3063                                 skinny_metadata = false;
3064                                 key.objectid = bytenr;
3065                                 key.type = BTRFS_EXTENT_ITEM_KEY;
3066                                 key.offset = num_bytes;
3067                                 btrfs_release_path(path);
3068                                 ret = btrfs_search_slot(trans, extent_root,
3069                                                         &key, path, -1, 1);
3070                         }
3071
3072                         if (ret) {
3073                                 btrfs_err(info,
3074                                           "umm, got %d back from search, was looking for %llu",
3075                                           ret, bytenr);
3076                                 if (ret > 0)
3077                                         btrfs_print_leaf(path->nodes[0]);
3078                         }
3079                         if (ret < 0) {
3080                                 btrfs_abort_transaction(trans, ret);
3081                                 goto out;
3082                         }
3083                         extent_slot = path->slots[0];
3084                 }
3085         } else if (WARN_ON(ret == -ENOENT)) {
3086                 btrfs_print_leaf(path->nodes[0]);
3087                 btrfs_err(info,
3088                         "unable to find ref byte nr %llu parent %llu root %llu  owner %llu offset %llu",
3089                         bytenr, parent, root_objectid, owner_objectid,
3090                         owner_offset);
3091                 btrfs_abort_transaction(trans, ret);
3092                 goto out;
3093         } else {
3094                 btrfs_abort_transaction(trans, ret);
3095                 goto out;
3096         }
3097
3098         leaf = path->nodes[0];
3099         item_size = btrfs_item_size(leaf, extent_slot);
3100         if (unlikely(item_size < sizeof(*ei))) {
3101                 ret = -EINVAL;
3102                 btrfs_print_v0_err(info);
3103                 btrfs_abort_transaction(trans, ret);
3104                 goto out;
3105         }
3106         ei = btrfs_item_ptr(leaf, extent_slot,
3107                             struct btrfs_extent_item);
3108         if (owner_objectid < BTRFS_FIRST_FREE_OBJECTID &&
3109             key.type == BTRFS_EXTENT_ITEM_KEY) {
3110                 struct btrfs_tree_block_info *bi;
3111                 if (item_size < sizeof(*ei) + sizeof(*bi)) {
3112                         btrfs_crit(info,
3113 "invalid extent item size for key (%llu, %u, %llu) owner %llu, has %u expect >= %zu",
3114                                    key.objectid, key.type, key.offset,
3115                                    owner_objectid, item_size,
3116                                    sizeof(*ei) + sizeof(*bi));
3117                         btrfs_abort_transaction(trans, -EUCLEAN);
3118                         goto err_dump;
3119                 }
3120                 bi = (struct btrfs_tree_block_info *)(ei + 1);
3121                 WARN_ON(owner_objectid != btrfs_tree_block_level(leaf, bi));
3122         }
3123
3124         refs = btrfs_extent_refs(leaf, ei);
3125         if (refs < refs_to_drop) {
3126                 btrfs_crit(info,
3127                 "trying to drop %d refs but we only have %llu for bytenr %llu",
3128                           refs_to_drop, refs, bytenr);
3129                 btrfs_abort_transaction(trans, -EUCLEAN);
3130                 goto err_dump;
3131         }
3132         refs -= refs_to_drop;
3133
3134         if (refs > 0) {
3135                 if (extent_op)
3136                         __run_delayed_extent_op(extent_op, leaf, ei);
3137                 /*
3138                  * In the case of inline back ref, reference count will
3139                  * be updated by remove_extent_backref
3140                  */
3141                 if (iref) {
3142                         if (!found_extent) {
3143                                 btrfs_crit(info,
3144 "invalid iref, got inlined extent ref but no EXTENT/METADATA_ITEM found");
3145                                 btrfs_abort_transaction(trans, -EUCLEAN);
3146                                 goto err_dump;
3147                         }
3148                 } else {
3149                         btrfs_set_extent_refs(leaf, ei, refs);
3150                         btrfs_mark_buffer_dirty(leaf);
3151                 }
3152                 if (found_extent) {
3153                         ret = remove_extent_backref(trans, extent_root, path,
3154                                                     iref, refs_to_drop, is_data);
3155                         if (ret) {
3156                                 btrfs_abort_transaction(trans, ret);
3157                                 goto out;
3158                         }
3159                 }
3160         } else {
3161                 /* In this branch refs == 1 */
3162                 if (found_extent) {
3163                         if (is_data && refs_to_drop !=
3164                             extent_data_ref_count(path, iref)) {
3165                                 btrfs_crit(info,
3166                 "invalid refs_to_drop, current refs %u refs_to_drop %u",
3167                                            extent_data_ref_count(path, iref),
3168                                            refs_to_drop);
3169                                 btrfs_abort_transaction(trans, -EUCLEAN);
3170                                 goto err_dump;
3171                         }
3172                         if (iref) {
3173                                 if (path->slots[0] != extent_slot) {
3174                                         btrfs_crit(info,
3175 "invalid iref, extent item key (%llu %u %llu) doesn't have wanted iref",
3176                                                    key.objectid, key.type,
3177                                                    key.offset);
3178                                         btrfs_abort_transaction(trans, -EUCLEAN);
3179                                         goto err_dump;
3180                                 }
3181                         } else {
3182                                 /*
3183                                  * No inline ref, we must be at SHARED_* item,
3184                                  * And it's single ref, it must be:
3185                                  * |    extent_slot       ||extent_slot + 1|
3186                                  * [ EXTENT/METADATA_ITEM ][ SHARED_* ITEM ]
3187                                  */
3188                                 if (path->slots[0] != extent_slot + 1) {
3189                                         btrfs_crit(info,
3190         "invalid SHARED_* item, previous item is not EXTENT/METADATA_ITEM");
3191                                         btrfs_abort_transaction(trans, -EUCLEAN);
3192                                         goto err_dump;
3193                                 }
3194                                 path->slots[0] = extent_slot;
3195                                 num_to_del = 2;
3196                         }
3197                 }
3198
3199                 ret = btrfs_del_items(trans, extent_root, path, path->slots[0],
3200                                       num_to_del);
3201                 if (ret) {
3202                         btrfs_abort_transaction(trans, ret);
3203                         goto out;
3204                 }
3205                 btrfs_release_path(path);
3206
3207                 ret = do_free_extent_accounting(trans, bytenr, num_bytes, is_data);
3208         }
3209         btrfs_release_path(path);
3210
3211 out:
3212         btrfs_free_path(path);
3213         return ret;
3214 err_dump:
3215         /*
3216          * Leaf dump can take up a lot of log buffer, so we only do full leaf
3217          * dump for debug build.
3218          */
3219         if (IS_ENABLED(CONFIG_BTRFS_DEBUG)) {
3220                 btrfs_crit(info, "path->slots[0]=%d extent_slot=%d",
3221                            path->slots[0], extent_slot);
3222                 btrfs_print_leaf(path->nodes[0]);
3223         }
3224
3225         btrfs_free_path(path);
3226         return -EUCLEAN;
3227 }
3228
3229 /*
3230  * when we free an block, it is possible (and likely) that we free the last
3231  * delayed ref for that extent as well.  This searches the delayed ref tree for
3232  * a given extent, and if there are no other delayed refs to be processed, it
3233  * removes it from the tree.
3234  */
3235 static noinline int check_ref_cleanup(struct btrfs_trans_handle *trans,
3236                                       u64 bytenr)
3237 {
3238         struct btrfs_delayed_ref_head *head;
3239         struct btrfs_delayed_ref_root *delayed_refs;
3240         int ret = 0;
3241
3242         delayed_refs = &trans->transaction->delayed_refs;
3243         spin_lock(&delayed_refs->lock);
3244         head = btrfs_find_delayed_ref_head(delayed_refs, bytenr);
3245         if (!head)
3246                 goto out_delayed_unlock;
3247
3248         spin_lock(&head->lock);
3249         if (!RB_EMPTY_ROOT(&head->ref_tree.rb_root))
3250                 goto out;
3251
3252         if (cleanup_extent_op(head) != NULL)
3253                 goto out;
3254
3255         /*
3256          * waiting for the lock here would deadlock.  If someone else has it
3257          * locked they are already in the process of dropping it anyway
3258          */
3259         if (!mutex_trylock(&head->mutex))
3260                 goto out;
3261
3262         btrfs_delete_ref_head(delayed_refs, head);
3263         head->processing = 0;
3264
3265         spin_unlock(&head->lock);
3266         spin_unlock(&delayed_refs->lock);
3267
3268         BUG_ON(head->extent_op);
3269         if (head->must_insert_reserved)
3270                 ret = 1;
3271
3272         btrfs_cleanup_ref_head_accounting(trans->fs_info, delayed_refs, head);
3273         mutex_unlock(&head->mutex);
3274         btrfs_put_delayed_ref_head(head);
3275         return ret;
3276 out:
3277         spin_unlock(&head->lock);
3278
3279 out_delayed_unlock:
3280         spin_unlock(&delayed_refs->lock);
3281         return 0;
3282 }
3283
3284 void btrfs_free_tree_block(struct btrfs_trans_handle *trans,
3285                            u64 root_id,
3286                            struct extent_buffer *buf,
3287                            u64 parent, int last_ref)
3288 {
3289         struct btrfs_fs_info *fs_info = trans->fs_info;
3290         struct btrfs_ref generic_ref = { 0 };
3291         int ret;
3292
3293         btrfs_init_generic_ref(&generic_ref, BTRFS_DROP_DELAYED_REF,
3294                                buf->start, buf->len, parent);
3295         btrfs_init_tree_ref(&generic_ref, btrfs_header_level(buf),
3296                             root_id, 0, false);
3297
3298         if (root_id != BTRFS_TREE_LOG_OBJECTID) {
3299                 btrfs_ref_tree_mod(fs_info, &generic_ref);
3300                 ret = btrfs_add_delayed_tree_ref(trans, &generic_ref, NULL);
3301                 BUG_ON(ret); /* -ENOMEM */
3302         }
3303
3304         if (last_ref && btrfs_header_generation(buf) == trans->transid) {
3305                 struct btrfs_block_group *cache;
3306                 bool must_pin = false;
3307
3308                 if (root_id != BTRFS_TREE_LOG_OBJECTID) {
3309                         ret = check_ref_cleanup(trans, buf->start);
3310                         if (!ret) {
3311                                 btrfs_redirty_list_add(trans->transaction, buf);
3312                                 goto out;
3313                         }
3314                 }
3315
3316                 cache = btrfs_lookup_block_group(fs_info, buf->start);
3317
3318                 if (btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN)) {
3319                         pin_down_extent(trans, cache, buf->start, buf->len, 1);
3320                         btrfs_put_block_group(cache);
3321                         goto out;
3322                 }
3323
3324                 /*
3325                  * If this is a leaf and there are tree mod log users, we may
3326                  * have recorded mod log operations that point to this leaf.
3327                  * So we must make sure no one reuses this leaf's extent before
3328                  * mod log operations are applied to a node, otherwise after
3329                  * rewinding a node using the mod log operations we get an
3330                  * inconsistent btree, as the leaf's extent may now be used as
3331                  * a node or leaf for another different btree.
3332                  * We are safe from races here because at this point no other
3333                  * node or root points to this extent buffer, so if after this
3334                  * check a new tree mod log user joins, it will not be able to
3335                  * find a node pointing to this leaf and record operations that
3336                  * point to this leaf.
3337                  */
3338                 if (btrfs_header_level(buf) == 0 &&
3339                     test_bit(BTRFS_FS_TREE_MOD_LOG_USERS, &fs_info->flags))
3340                         must_pin = true;
3341
3342                 if (must_pin || btrfs_is_zoned(fs_info)) {
3343                         btrfs_redirty_list_add(trans->transaction, buf);
3344                         pin_down_extent(trans, cache, buf->start, buf->len, 1);
3345                         btrfs_put_block_group(cache);
3346                         goto out;
3347                 }
3348
3349                 WARN_ON(test_bit(EXTENT_BUFFER_DIRTY, &buf->bflags));
3350
3351                 btrfs_add_free_space(cache, buf->start, buf->len);
3352                 btrfs_free_reserved_bytes(cache, buf->len, 0);
3353                 btrfs_put_block_group(cache);
3354                 trace_btrfs_reserved_extent_free(fs_info, buf->start, buf->len);
3355         }
3356 out:
3357         if (last_ref) {
3358                 /*
3359                  * Deleting the buffer, clear the corrupt flag since it doesn't
3360                  * matter anymore.
3361                  */
3362                 clear_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags);
3363         }
3364 }
3365
3366 /* Can return -ENOMEM */
3367 int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_ref *ref)
3368 {
3369         struct btrfs_fs_info *fs_info = trans->fs_info;
3370         int ret;
3371
3372         if (btrfs_is_testing(fs_info))
3373                 return 0;
3374
3375         /*
3376          * tree log blocks never actually go into the extent allocation
3377          * tree, just update pinning info and exit early.
3378          */
3379         if ((ref->type == BTRFS_REF_METADATA &&
3380              ref->tree_ref.owning_root == BTRFS_TREE_LOG_OBJECTID) ||
3381             (ref->type == BTRFS_REF_DATA &&
3382              ref->data_ref.owning_root == BTRFS_TREE_LOG_OBJECTID)) {
3383                 /* unlocks the pinned mutex */
3384                 btrfs_pin_extent(trans, ref->bytenr, ref->len, 1);
3385                 ret = 0;
3386         } else if (ref->type == BTRFS_REF_METADATA) {
3387                 ret = btrfs_add_delayed_tree_ref(trans, ref, NULL);
3388         } else {
3389                 ret = btrfs_add_delayed_data_ref(trans, ref, 0);
3390         }
3391
3392         if (!((ref->type == BTRFS_REF_METADATA &&
3393                ref->tree_ref.owning_root == BTRFS_TREE_LOG_OBJECTID) ||
3394               (ref->type == BTRFS_REF_DATA &&
3395                ref->data_ref.owning_root == BTRFS_TREE_LOG_OBJECTID)))
3396                 btrfs_ref_tree_mod(fs_info, ref);
3397
3398         return ret;
3399 }
3400
3401 enum btrfs_loop_type {
3402         LOOP_CACHING_NOWAIT,
3403         LOOP_CACHING_WAIT,
3404         LOOP_ALLOC_CHUNK,
3405         LOOP_NO_EMPTY_SIZE,
3406 };
3407
3408 static inline void
3409 btrfs_lock_block_group(struct btrfs_block_group *cache,
3410                        int delalloc)
3411 {
3412         if (delalloc)
3413                 down_read(&cache->data_rwsem);
3414 }
3415
3416 static inline void btrfs_grab_block_group(struct btrfs_block_group *cache,
3417                        int delalloc)
3418 {
3419         btrfs_get_block_group(cache);
3420         if (delalloc)
3421                 down_read(&cache->data_rwsem);
3422 }
3423
3424 static struct btrfs_block_group *btrfs_lock_cluster(
3425                    struct btrfs_block_group *block_group,
3426                    struct btrfs_free_cluster *cluster,
3427                    int delalloc)
3428         __acquires(&cluster->refill_lock)
3429 {
3430         struct btrfs_block_group *used_bg = NULL;
3431
3432         spin_lock(&cluster->refill_lock);
3433         while (1) {
3434                 used_bg = cluster->block_group;
3435                 if (!used_bg)
3436                         return NULL;
3437
3438                 if (used_bg == block_group)
3439                         return used_bg;
3440
3441                 btrfs_get_block_group(used_bg);
3442
3443                 if (!delalloc)
3444                         return used_bg;
3445
3446                 if (down_read_trylock(&used_bg->data_rwsem))
3447                         return used_bg;
3448
3449                 spin_unlock(&cluster->refill_lock);
3450
3451                 /* We should only have one-level nested. */
3452                 down_read_nested(&used_bg->data_rwsem, SINGLE_DEPTH_NESTING);
3453
3454                 spin_lock(&cluster->refill_lock);
3455                 if (used_bg == cluster->block_group)
3456                         return used_bg;
3457
3458                 up_read(&used_bg->data_rwsem);
3459                 btrfs_put_block_group(used_bg);
3460         }
3461 }
3462
3463 static inline void
3464 btrfs_release_block_group(struct btrfs_block_group *cache,
3465                          int delalloc)
3466 {
3467         if (delalloc)
3468                 up_read(&cache->data_rwsem);
3469         btrfs_put_block_group(cache);
3470 }
3471
3472 enum btrfs_extent_allocation_policy {
3473         BTRFS_EXTENT_ALLOC_CLUSTERED,
3474         BTRFS_EXTENT_ALLOC_ZONED,
3475 };
3476
3477 /*
3478  * Structure used internally for find_free_extent() function.  Wraps needed
3479  * parameters.
3480  */
3481 struct find_free_extent_ctl {
3482         /* Basic allocation info */
3483         u64 ram_bytes;
3484         u64 num_bytes;
3485         u64 min_alloc_size;
3486         u64 empty_size;
3487         u64 flags;
3488         int delalloc;
3489
3490         /* Where to start the search inside the bg */
3491         u64 search_start;
3492
3493         /* For clustered allocation */
3494         u64 empty_cluster;
3495         struct btrfs_free_cluster *last_ptr;
3496         bool use_cluster;
3497
3498         bool have_caching_bg;
3499         bool orig_have_caching_bg;
3500
3501         /* Allocation is called for tree-log */
3502         bool for_treelog;
3503
3504         /* Allocation is called for data relocation */
3505         bool for_data_reloc;
3506
3507         /* RAID index, converted from flags */
3508         int index;
3509
3510         /*
3511          * Current loop number, check find_free_extent_update_loop() for details
3512          */
3513         int loop;
3514
3515         /*
3516          * Whether we're refilling a cluster, if true we need to re-search
3517          * current block group but don't try to refill the cluster again.
3518          */
3519         bool retry_clustered;
3520
3521         /*
3522          * Whether we're updating free space cache, if true we need to re-search
3523          * current block group but don't try updating free space cache again.
3524          */
3525         bool retry_unclustered;
3526
3527         /* If current block group is cached */
3528         int cached;
3529
3530         /* Max contiguous hole found */
3531         u64 max_extent_size;
3532
3533         /* Total free space from free space cache, not always contiguous */
3534         u64 total_free_space;
3535
3536         /* Found result */
3537         u64 found_offset;
3538
3539         /* Hint where to start looking for an empty space */
3540         u64 hint_byte;
3541
3542         /* Allocation policy */
3543         enum btrfs_extent_allocation_policy policy;
3544 };
3545
3546
3547 /*
3548  * Helper function for find_free_extent().
3549  *
3550  * Return -ENOENT to inform caller that we need fallback to unclustered mode.
3551  * Return -EAGAIN to inform caller that we need to re-search this block group
3552  * Return >0 to inform caller that we find nothing
3553  * Return 0 means we have found a location and set ffe_ctl->found_offset.
3554  */
3555 static int find_free_extent_clustered(struct btrfs_block_group *bg,
3556                                       struct find_free_extent_ctl *ffe_ctl,
3557                                       struct btrfs_block_group **cluster_bg_ret)
3558 {
3559         struct btrfs_block_group *cluster_bg;
3560         struct btrfs_free_cluster *last_ptr = ffe_ctl->last_ptr;
3561         u64 aligned_cluster;
3562         u64 offset;
3563         int ret;
3564
3565         cluster_bg = btrfs_lock_cluster(bg, last_ptr, ffe_ctl->delalloc);
3566         if (!cluster_bg)
3567                 goto refill_cluster;
3568         if (cluster_bg != bg && (cluster_bg->ro ||
3569             !block_group_bits(cluster_bg, ffe_ctl->flags)))
3570                 goto release_cluster;
3571
3572         offset = btrfs_alloc_from_cluster(cluster_bg, last_ptr,
3573                         ffe_ctl->num_bytes, cluster_bg->start,
3574                         &ffe_ctl->max_extent_size);
3575         if (offset) {
3576                 /* We have a block, we're done */
3577                 spin_unlock(&last_ptr->refill_lock);
3578                 trace_btrfs_reserve_extent_cluster(cluster_bg,
3579                                 ffe_ctl->search_start, ffe_ctl->num_bytes);
3580                 *cluster_bg_ret = cluster_bg;
3581                 ffe_ctl->found_offset = offset;
3582                 return 0;
3583         }
3584         WARN_ON(last_ptr->block_group != cluster_bg);
3585
3586 release_cluster:
3587         /*
3588          * If we are on LOOP_NO_EMPTY_SIZE, we can't set up a new clusters, so
3589          * lets just skip it and let the allocator find whatever block it can
3590          * find. If we reach this point, we will have tried the cluster
3591          * allocator plenty of times and not have found anything, so we are
3592          * likely way too fragmented for the clustering stuff to find anything.
3593          *
3594          * However, if the cluster is taken from the current block group,
3595          * release the cluster first, so that we stand a better chance of
3596          * succeeding in the unclustered allocation.
3597          */
3598         if (ffe_ctl->loop >= LOOP_NO_EMPTY_SIZE && cluster_bg != bg) {
3599                 spin_unlock(&last_ptr->refill_lock);
3600                 btrfs_release_block_group(cluster_bg, ffe_ctl->delalloc);
3601                 return -ENOENT;
3602         }
3603
3604         /* This cluster didn't work out, free it and start over */
3605         btrfs_return_cluster_to_free_space(NULL, last_ptr);
3606
3607         if (cluster_bg != bg)
3608                 btrfs_release_block_group(cluster_bg, ffe_ctl->delalloc);
3609
3610 refill_cluster:
3611         if (ffe_ctl->loop >= LOOP_NO_EMPTY_SIZE) {
3612                 spin_unlock(&last_ptr->refill_lock);
3613                 return -ENOENT;
3614         }
3615
3616         aligned_cluster = max_t(u64,
3617                         ffe_ctl->empty_cluster + ffe_ctl->empty_size,
3618                         bg->full_stripe_len);
3619         ret = btrfs_find_space_cluster(bg, last_ptr, ffe_ctl->search_start,
3620                         ffe_ctl->num_bytes, aligned_cluster);
3621         if (ret == 0) {
3622                 /* Now pull our allocation out of this cluster */
3623                 offset = btrfs_alloc_from_cluster(bg, last_ptr,
3624                                 ffe_ctl->num_bytes, ffe_ctl->search_start,
3625                                 &ffe_ctl->max_extent_size);
3626                 if (offset) {
3627                         /* We found one, proceed */
3628                         spin_unlock(&last_ptr->refill_lock);
3629                         trace_btrfs_reserve_extent_cluster(bg,
3630                                         ffe_ctl->search_start,
3631                                         ffe_ctl->num_bytes);
3632                         ffe_ctl->found_offset = offset;
3633                         return 0;
3634                 }
3635         } else if (!ffe_ctl->cached && ffe_ctl->loop > LOOP_CACHING_NOWAIT &&
3636                    !ffe_ctl->retry_clustered) {
3637                 spin_unlock(&last_ptr->refill_lock);
3638
3639                 ffe_ctl->retry_clustered = true;
3640                 btrfs_wait_block_group_cache_progress(bg, ffe_ctl->num_bytes +
3641                                 ffe_ctl->empty_cluster + ffe_ctl->empty_size);
3642                 return -EAGAIN;
3643         }
3644         /*
3645          * At this point we either didn't find a cluster or we weren't able to
3646          * allocate a block from our cluster.  Free the cluster we've been
3647          * trying to use, and go to the next block group.
3648          */
3649         btrfs_return_cluster_to_free_space(NULL, last_ptr);
3650         spin_unlock(&last_ptr->refill_lock);
3651         return 1;
3652 }
3653
3654 /*
3655  * Return >0 to inform caller that we find nothing
3656  * Return 0 when we found an free extent and set ffe_ctrl->found_offset
3657  * Return -EAGAIN to inform caller that we need to re-search this block group
3658  */
3659 static int find_free_extent_unclustered(struct btrfs_block_group *bg,
3660                                         struct find_free_extent_ctl *ffe_ctl)
3661 {
3662         struct btrfs_free_cluster *last_ptr = ffe_ctl->last_ptr;
3663         u64 offset;
3664
3665         /*
3666          * We are doing an unclustered allocation, set the fragmented flag so
3667          * we don't bother trying to setup a cluster again until we get more
3668          * space.
3669          */
3670         if (unlikely(last_ptr)) {
3671                 spin_lock(&last_ptr->lock);
3672                 last_ptr->fragmented = 1;
3673                 spin_unlock(&last_ptr->lock);
3674         }
3675         if (ffe_ctl->cached) {
3676                 struct btrfs_free_space_ctl *free_space_ctl;
3677
3678                 free_space_ctl = bg->free_space_ctl;
3679                 spin_lock(&free_space_ctl->tree_lock);
3680                 if (free_space_ctl->free_space <
3681                     ffe_ctl->num_bytes + ffe_ctl->empty_cluster +
3682                     ffe_ctl->empty_size) {
3683                         ffe_ctl->total_free_space = max_t(u64,
3684                                         ffe_ctl->total_free_space,
3685                                         free_space_ctl->free_space);
3686                         spin_unlock(&free_space_ctl->tree_lock);
3687                         return 1;
3688                 }
3689                 spin_unlock(&free_space_ctl->tree_lock);
3690         }
3691
3692         offset = btrfs_find_space_for_alloc(bg, ffe_ctl->search_start,
3693                         ffe_ctl->num_bytes, ffe_ctl->empty_size,
3694                         &ffe_ctl->max_extent_size);
3695
3696         /*
3697          * If we didn't find a chunk, and we haven't failed on this block group
3698          * before, and this block group is in the middle of caching and we are
3699          * ok with waiting, then go ahead and wait for progress to be made, and
3700          * set @retry_unclustered to true.
3701          *
3702          * If @retry_unclustered is true then we've already waited on this
3703          * block group once and should move on to the next block group.
3704          */
3705         if (!offset && !ffe_ctl->retry_unclustered && !ffe_ctl->cached &&
3706             ffe_ctl->loop > LOOP_CACHING_NOWAIT) {
3707                 btrfs_wait_block_group_cache_progress(bg, ffe_ctl->num_bytes +
3708                                                       ffe_ctl->empty_size);
3709                 ffe_ctl->retry_unclustered = true;
3710                 return -EAGAIN;
3711         } else if (!offset) {
3712                 return 1;
3713         }
3714         ffe_ctl->found_offset = offset;
3715         return 0;
3716 }
3717
3718 static int do_allocation_clustered(struct btrfs_block_group *block_group,
3719                                    struct find_free_extent_ctl *ffe_ctl,
3720                                    struct btrfs_block_group **bg_ret)
3721 {
3722         int ret;
3723
3724         /* We want to try and use the cluster allocator, so lets look there */
3725         if (ffe_ctl->last_ptr && ffe_ctl->use_cluster) {
3726                 ret = find_free_extent_clustered(block_group, ffe_ctl, bg_ret);
3727                 if (ret >= 0 || ret == -EAGAIN)
3728                         return ret;
3729                 /* ret == -ENOENT case falls through */
3730         }
3731
3732         return find_free_extent_unclustered(block_group, ffe_ctl);
3733 }
3734
3735 /*
3736  * Tree-log block group locking
3737  * ============================
3738  *
3739  * fs_info::treelog_bg_lock protects the fs_info::treelog_bg which
3740  * indicates the starting address of a block group, which is reserved only
3741  * for tree-log metadata.
3742  *
3743  * Lock nesting
3744  * ============
3745  *
3746  * space_info::lock
3747  *   block_group::lock
3748  *     fs_info::treelog_bg_lock
3749  */
3750
3751 /*
3752  * Simple allocator for sequential-only block group. It only allows sequential
3753  * allocation. No need to play with trees. This function also reserves the
3754  * bytes as in btrfs_add_reserved_bytes.
3755  */
3756 static int do_allocation_zoned(struct btrfs_block_group *block_group,
3757                                struct find_free_extent_ctl *ffe_ctl,
3758                                struct btrfs_block_group **bg_ret)
3759 {
3760         struct btrfs_fs_info *fs_info = block_group->fs_info;
3761         struct btrfs_space_info *space_info = block_group->space_info;
3762         struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
3763         u64 start = block_group->start;
3764         u64 num_bytes = ffe_ctl->num_bytes;
3765         u64 avail;
3766         u64 bytenr = block_group->start;
3767         u64 log_bytenr;
3768         u64 data_reloc_bytenr;
3769         int ret = 0;
3770         bool skip = false;
3771
3772         ASSERT(btrfs_is_zoned(block_group->fs_info));
3773
3774         /*
3775          * Do not allow non-tree-log blocks in the dedicated tree-log block
3776          * group, and vice versa.
3777          */
3778         spin_lock(&fs_info->treelog_bg_lock);
3779         log_bytenr = fs_info->treelog_bg;
3780         if (log_bytenr && ((ffe_ctl->for_treelog && bytenr != log_bytenr) ||
3781                            (!ffe_ctl->for_treelog && bytenr == log_bytenr)))
3782                 skip = true;
3783         spin_unlock(&fs_info->treelog_bg_lock);
3784         if (skip)
3785                 return 1;
3786
3787         /*
3788          * Do not allow non-relocation blocks in the dedicated relocation block
3789          * group, and vice versa.
3790          */
3791         spin_lock(&fs_info->relocation_bg_lock);
3792         data_reloc_bytenr = fs_info->data_reloc_bg;
3793         if (data_reloc_bytenr &&
3794             ((ffe_ctl->for_data_reloc && bytenr != data_reloc_bytenr) ||
3795              (!ffe_ctl->for_data_reloc && bytenr == data_reloc_bytenr)))
3796                 skip = true;
3797         spin_unlock(&fs_info->relocation_bg_lock);
3798         if (skip)
3799                 return 1;
3800
3801         /* Check RO and no space case before trying to activate it */
3802         spin_lock(&block_group->lock);
3803         if (block_group->ro || btrfs_zoned_bg_is_full(block_group)) {
3804                 ret = 1;
3805                 /*
3806                  * May need to clear fs_info->{treelog,data_reloc}_bg.
3807                  * Return the error after taking the locks.
3808                  */
3809         }
3810         spin_unlock(&block_group->lock);
3811
3812         if (!ret && !btrfs_zone_activate(block_group)) {
3813                 ret = 1;
3814                 /*
3815                  * May need to clear fs_info->{treelog,data_reloc}_bg.
3816                  * Return the error after taking the locks.
3817                  */
3818         }
3819
3820         spin_lock(&space_info->lock);
3821         spin_lock(&block_group->lock);
3822         spin_lock(&fs_info->treelog_bg_lock);
3823         spin_lock(&fs_info->relocation_bg_lock);
3824
3825         if (ret)
3826                 goto out;
3827
3828         ASSERT(!ffe_ctl->for_treelog ||
3829                block_group->start == fs_info->treelog_bg ||
3830                fs_info->treelog_bg == 0);
3831         ASSERT(!ffe_ctl->for_data_reloc ||
3832                block_group->start == fs_info->data_reloc_bg ||
3833                fs_info->data_reloc_bg == 0);
3834
3835         if (block_group->ro || block_group->zoned_data_reloc_ongoing) {
3836                 ret = 1;
3837                 goto out;
3838         }
3839
3840         /*
3841          * Do not allow currently using block group to be tree-log dedicated
3842          * block group.
3843          */
3844         if (ffe_ctl->for_treelog && !fs_info->treelog_bg &&
3845             (block_group->used || block_group->reserved)) {
3846                 ret = 1;
3847                 goto out;
3848         }
3849
3850         /*
3851          * Do not allow currently used block group to be the data relocation
3852          * dedicated block group.
3853          */
3854         if (ffe_ctl->for_data_reloc && !fs_info->data_reloc_bg &&
3855             (block_group->used || block_group->reserved)) {
3856                 ret = 1;
3857                 goto out;
3858         }
3859
3860         WARN_ON_ONCE(block_group->alloc_offset > block_group->zone_capacity);
3861         avail = block_group->zone_capacity - block_group->alloc_offset;
3862         if (avail < num_bytes) {
3863                 if (ffe_ctl->max_extent_size < avail) {
3864                         /*
3865                          * With sequential allocator, free space is always
3866                          * contiguous
3867                          */
3868                         ffe_ctl->max_extent_size = avail;
3869                         ffe_ctl->total_free_space = avail;
3870                 }
3871                 ret = 1;
3872                 goto out;
3873         }
3874
3875         if (ffe_ctl->for_treelog && !fs_info->treelog_bg)
3876                 fs_info->treelog_bg = block_group->start;
3877
3878         if (ffe_ctl->for_data_reloc && !fs_info->data_reloc_bg)
3879                 fs_info->data_reloc_bg = block_group->start;
3880
3881         ffe_ctl->found_offset = start + block_group->alloc_offset;
3882         block_group->alloc_offset += num_bytes;
3883         spin_lock(&ctl->tree_lock);
3884         ctl->free_space -= num_bytes;
3885         spin_unlock(&ctl->tree_lock);
3886
3887         /*
3888          * We do not check if found_offset is aligned to stripesize. The
3889          * address is anyway rewritten when using zone append writing.
3890          */
3891
3892         ffe_ctl->search_start = ffe_ctl->found_offset;
3893
3894 out:
3895         if (ret && ffe_ctl->for_treelog)
3896                 fs_info->treelog_bg = 0;
3897         if (ret && ffe_ctl->for_data_reloc &&
3898             fs_info->data_reloc_bg == block_group->start) {
3899                 /*
3900                  * Do not allow further allocations from this block group.
3901                  * Compared to increasing the ->ro, setting the
3902                  * ->zoned_data_reloc_ongoing flag still allows nocow
3903                  *  writers to come in. See btrfs_inc_nocow_writers().
3904                  *
3905                  * We need to disable an allocation to avoid an allocation of
3906                  * regular (non-relocation data) extent. With mix of relocation
3907                  * extents and regular extents, we can dispatch WRITE commands
3908                  * (for relocation extents) and ZONE APPEND commands (for
3909                  * regular extents) at the same time to the same zone, which
3910                  * easily break the write pointer.
3911                  */
3912                 block_group->zoned_data_reloc_ongoing = 1;
3913                 fs_info->data_reloc_bg = 0;
3914         }
3915         spin_unlock(&fs_info->relocation_bg_lock);
3916         spin_unlock(&fs_info->treelog_bg_lock);
3917         spin_unlock(&block_group->lock);
3918         spin_unlock(&space_info->lock);
3919         return ret;
3920 }
3921
3922 static int do_allocation(struct btrfs_block_group *block_group,
3923                          struct find_free_extent_ctl *ffe_ctl,
3924                          struct btrfs_block_group **bg_ret)
3925 {
3926         switch (ffe_ctl->policy) {
3927         case BTRFS_EXTENT_ALLOC_CLUSTERED:
3928                 return do_allocation_clustered(block_group, ffe_ctl, bg_ret);
3929         case BTRFS_EXTENT_ALLOC_ZONED:
3930                 return do_allocation_zoned(block_group, ffe_ctl, bg_ret);
3931         default:
3932                 BUG();
3933         }
3934 }
3935
3936 static void release_block_group(struct btrfs_block_group *block_group,
3937                                 struct find_free_extent_ctl *ffe_ctl,
3938                                 int delalloc)
3939 {
3940         switch (ffe_ctl->policy) {
3941         case BTRFS_EXTENT_ALLOC_CLUSTERED:
3942                 ffe_ctl->retry_clustered = false;
3943                 ffe_ctl->retry_unclustered = false;
3944                 break;
3945         case BTRFS_EXTENT_ALLOC_ZONED:
3946                 /* Nothing to do */
3947                 break;
3948         default:
3949                 BUG();
3950         }
3951
3952         BUG_ON(btrfs_bg_flags_to_raid_index(block_group->flags) !=
3953                ffe_ctl->index);
3954         btrfs_release_block_group(block_group, delalloc);
3955 }
3956
3957 static void found_extent_clustered(struct find_free_extent_ctl *ffe_ctl,
3958                                    struct btrfs_key *ins)
3959 {
3960         struct btrfs_free_cluster *last_ptr = ffe_ctl->last_ptr;
3961
3962         if (!ffe_ctl->use_cluster && last_ptr) {
3963                 spin_lock(&last_ptr->lock);
3964                 last_ptr->window_start = ins->objectid;
3965                 spin_unlock(&last_ptr->lock);
3966         }
3967 }
3968
3969 static void found_extent(struct find_free_extent_ctl *ffe_ctl,
3970                          struct btrfs_key *ins)
3971 {
3972         switch (ffe_ctl->policy) {
3973         case BTRFS_EXTENT_ALLOC_CLUSTERED:
3974                 found_extent_clustered(ffe_ctl, ins);
3975                 break;
3976         case BTRFS_EXTENT_ALLOC_ZONED:
3977                 /* Nothing to do */
3978                 break;
3979         default:
3980                 BUG();
3981         }
3982 }
3983
3984 static bool can_allocate_chunk(struct btrfs_fs_info *fs_info,
3985                                struct find_free_extent_ctl *ffe_ctl)
3986 {
3987         switch (ffe_ctl->policy) {
3988         case BTRFS_EXTENT_ALLOC_CLUSTERED:
3989                 return true;
3990         case BTRFS_EXTENT_ALLOC_ZONED:
3991                 /*
3992                  * If we have enough free space left in an already
3993                  * active block group and we can't activate any other
3994                  * zone now, do not allow allocating a new chunk and
3995                  * let find_free_extent() retry with a smaller size.
3996                  */
3997                 if (ffe_ctl->max_extent_size >= ffe_ctl->min_alloc_size &&
3998                     !btrfs_can_activate_zone(fs_info->fs_devices, ffe_ctl->flags))
3999                         return false;
4000                 return true;
4001         default:
4002                 BUG();
4003         }
4004 }
4005
4006 static int chunk_allocation_failed(struct find_free_extent_ctl *ffe_ctl)
4007 {
4008         switch (ffe_ctl->policy) {
4009         case BTRFS_EXTENT_ALLOC_CLUSTERED:
4010                 /*
4011                  * If we can't allocate a new chunk we've already looped through
4012                  * at least once, move on to the NO_EMPTY_SIZE case.
4013                  */
4014                 ffe_ctl->loop = LOOP_NO_EMPTY_SIZE;
4015                 return 0;
4016         case BTRFS_EXTENT_ALLOC_ZONED:
4017                 /* Give up here */
4018                 return -ENOSPC;
4019         default:
4020                 BUG();
4021         }
4022 }
4023
4024 /*
4025  * Return >0 means caller needs to re-search for free extent
4026  * Return 0 means we have the needed free extent.
4027  * Return <0 means we failed to locate any free extent.
4028  */
4029 static int find_free_extent_update_loop(struct btrfs_fs_info *fs_info,
4030                                         struct btrfs_key *ins,
4031                                         struct find_free_extent_ctl *ffe_ctl,
4032                                         bool full_search)
4033 {
4034         struct btrfs_root *root = fs_info->chunk_root;
4035         int ret;
4036
4037         if ((ffe_ctl->loop == LOOP_CACHING_NOWAIT) &&
4038             ffe_ctl->have_caching_bg && !ffe_ctl->orig_have_caching_bg)
4039                 ffe_ctl->orig_have_caching_bg = true;
4040
4041         if (ins->objectid) {
4042                 found_extent(ffe_ctl, ins);
4043                 return 0;
4044         }
4045
4046         if (ffe_ctl->loop >= LOOP_CACHING_WAIT && ffe_ctl->have_caching_bg)
4047                 return 1;
4048
4049         ffe_ctl->index++;
4050         if (ffe_ctl->index < BTRFS_NR_RAID_TYPES)
4051                 return 1;
4052
4053         /*
4054          * LOOP_CACHING_NOWAIT, search partially cached block groups, kicking
4055          *                      caching kthreads as we move along
4056          * LOOP_CACHING_WAIT, search everything, and wait if our bg is caching
4057          * LOOP_ALLOC_CHUNK, force a chunk allocation and try again
4058          * LOOP_NO_EMPTY_SIZE, set empty_size and empty_cluster to 0 and try
4059          *                     again
4060          */
4061         if (ffe_ctl->loop < LOOP_NO_EMPTY_SIZE) {
4062                 ffe_ctl->index = 0;
4063                 if (ffe_ctl->loop == LOOP_CACHING_NOWAIT) {
4064                         /*
4065                          * We want to skip the LOOP_CACHING_WAIT step if we
4066                          * don't have any uncached bgs and we've already done a
4067                          * full search through.
4068                          */
4069                         if (ffe_ctl->orig_have_caching_bg || !full_search)
4070                                 ffe_ctl->loop = LOOP_CACHING_WAIT;
4071                         else
4072                                 ffe_ctl->loop = LOOP_ALLOC_CHUNK;
4073                 } else {
4074                         ffe_ctl->loop++;
4075                 }
4076
4077                 if (ffe_ctl->loop == LOOP_ALLOC_CHUNK) {
4078                         struct btrfs_trans_handle *trans;
4079                         int exist = 0;
4080
4081                         /*Check if allocation policy allows to create a new chunk */
4082                         if (!can_allocate_chunk(fs_info, ffe_ctl))
4083                                 return -ENOSPC;
4084
4085                         trans = current->journal_info;
4086                         if (trans)
4087                                 exist = 1;
4088                         else
4089                                 trans = btrfs_join_transaction(root);
4090
4091                         if (IS_ERR(trans)) {
4092                                 ret = PTR_ERR(trans);
4093                                 return ret;
4094                         }
4095
4096                         ret = btrfs_chunk_alloc(trans, ffe_ctl->flags,
4097                                                 CHUNK_ALLOC_FORCE_FOR_EXTENT);
4098
4099                         /* Do not bail out on ENOSPC since we can do more. */
4100                         if (ret == -ENOSPC)
4101                                 ret = chunk_allocation_failed(ffe_ctl);
4102                         else if (ret < 0)
4103                                 btrfs_abort_transaction(trans, ret);
4104                         else
4105                                 ret = 0;
4106                         if (!exist)
4107                                 btrfs_end_transaction(trans);
4108                         if (ret)
4109                                 return ret;
4110                 }
4111
4112                 if (ffe_ctl->loop == LOOP_NO_EMPTY_SIZE) {
4113                         if (ffe_ctl->policy != BTRFS_EXTENT_ALLOC_CLUSTERED)
4114                                 return -ENOSPC;
4115
4116                         /*
4117                          * Don't loop again if we already have no empty_size and
4118                          * no empty_cluster.
4119                          */
4120                         if (ffe_ctl->empty_size == 0 &&
4121                             ffe_ctl->empty_cluster == 0)
4122                                 return -ENOSPC;
4123                         ffe_ctl->empty_size = 0;
4124                         ffe_ctl->empty_cluster = 0;
4125                 }
4126                 return 1;
4127         }
4128         return -ENOSPC;
4129 }
4130
4131 static int prepare_allocation_clustered(struct btrfs_fs_info *fs_info,
4132                                         struct find_free_extent_ctl *ffe_ctl,
4133                                         struct btrfs_space_info *space_info,
4134                                         struct btrfs_key *ins)
4135 {
4136         /*
4137          * If our free space is heavily fragmented we may not be able to make
4138          * big contiguous allocations, so instead of doing the expensive search
4139          * for free space, simply return ENOSPC with our max_extent_size so we
4140          * can go ahead and search for a more manageable chunk.
4141          *
4142          * If our max_extent_size is large enough for our allocation simply
4143          * disable clustering since we will likely not be able to find enough
4144          * space to create a cluster and induce latency trying.
4145          */
4146         if (space_info->max_extent_size) {
4147                 spin_lock(&space_info->lock);
4148                 if (space_info->max_extent_size &&
4149                     ffe_ctl->num_bytes > space_info->max_extent_size) {
4150                         ins->offset = space_info->max_extent_size;
4151                         spin_unlock(&space_info->lock);
4152                         return -ENOSPC;
4153                 } else if (space_info->max_extent_size) {
4154                         ffe_ctl->use_cluster = false;
4155                 }
4156                 spin_unlock(&space_info->lock);
4157         }
4158
4159         ffe_ctl->last_ptr = fetch_cluster_info(fs_info, space_info,
4160                                                &ffe_ctl->empty_cluster);
4161         if (ffe_ctl->last_ptr) {
4162                 struct btrfs_free_cluster *last_ptr = ffe_ctl->last_ptr;
4163
4164                 spin_lock(&last_ptr->lock);
4165                 if (last_ptr->block_group)
4166                         ffe_ctl->hint_byte = last_ptr->window_start;
4167                 if (last_ptr->fragmented) {
4168                         /*
4169                          * We still set window_start so we can keep track of the
4170                          * last place we found an allocation to try and save
4171                          * some time.
4172                          */
4173                         ffe_ctl->hint_byte = last_ptr->window_start;
4174                         ffe_ctl->use_cluster = false;
4175                 }
4176                 spin_unlock(&last_ptr->lock);
4177         }
4178
4179         return 0;
4180 }
4181
4182 static int prepare_allocation(struct btrfs_fs_info *fs_info,
4183                               struct find_free_extent_ctl *ffe_ctl,
4184                               struct btrfs_space_info *space_info,
4185                               struct btrfs_key *ins)
4186 {
4187         switch (ffe_ctl->policy) {
4188         case BTRFS_EXTENT_ALLOC_CLUSTERED:
4189                 return prepare_allocation_clustered(fs_info, ffe_ctl,
4190                                                     space_info, ins);
4191         case BTRFS_EXTENT_ALLOC_ZONED:
4192                 if (ffe_ctl->for_treelog) {
4193                         spin_lock(&fs_info->treelog_bg_lock);
4194                         if (fs_info->treelog_bg)
4195                                 ffe_ctl->hint_byte = fs_info->treelog_bg;
4196                         spin_unlock(&fs_info->treelog_bg_lock);
4197                 }
4198                 if (ffe_ctl->for_data_reloc) {
4199                         spin_lock(&fs_info->relocation_bg_lock);
4200                         if (fs_info->data_reloc_bg)
4201                                 ffe_ctl->hint_byte = fs_info->data_reloc_bg;
4202                         spin_unlock(&fs_info->relocation_bg_lock);
4203                 }
4204                 return 0;
4205         default:
4206                 BUG();
4207         }
4208 }
4209
4210 /*
4211  * walks the btree of allocated extents and find a hole of a given size.
4212  * The key ins is changed to record the hole:
4213  * ins->objectid == start position
4214  * ins->flags = BTRFS_EXTENT_ITEM_KEY
4215  * ins->offset == the size of the hole.
4216  * Any available blocks before search_start are skipped.
4217  *
4218  * If there is no suitable free space, we will record the max size of
4219  * the free space extent currently.
4220  *
4221  * The overall logic and call chain:
4222  *
4223  * find_free_extent()
4224  * |- Iterate through all block groups
4225  * |  |- Get a valid block group
4226  * |  |- Try to do clustered allocation in that block group
4227  * |  |- Try to do unclustered allocation in that block group
4228  * |  |- Check if the result is valid
4229  * |  |  |- If valid, then exit
4230  * |  |- Jump to next block group
4231  * |
4232  * |- Push harder to find free extents
4233  *    |- If not found, re-iterate all block groups
4234  */
4235 static noinline int find_free_extent(struct btrfs_root *root,
4236                                      struct btrfs_key *ins,
4237                                      struct find_free_extent_ctl *ffe_ctl)
4238 {
4239         struct btrfs_fs_info *fs_info = root->fs_info;
4240         int ret = 0;
4241         int cache_block_group_error = 0;
4242         struct btrfs_block_group *block_group = NULL;
4243         struct btrfs_space_info *space_info;
4244         bool full_search = false;
4245
4246         WARN_ON(ffe_ctl->num_bytes < fs_info->sectorsize);
4247
4248         ffe_ctl->search_start = 0;
4249         /* For clustered allocation */
4250         ffe_ctl->empty_cluster = 0;
4251         ffe_ctl->last_ptr = NULL;
4252         ffe_ctl->use_cluster = true;
4253         ffe_ctl->have_caching_bg = false;
4254         ffe_ctl->orig_have_caching_bg = false;
4255         ffe_ctl->index = btrfs_bg_flags_to_raid_index(ffe_ctl->flags);
4256         ffe_ctl->loop = 0;
4257         /* For clustered allocation */
4258         ffe_ctl->retry_clustered = false;
4259         ffe_ctl->retry_unclustered = false;
4260         ffe_ctl->cached = 0;
4261         ffe_ctl->max_extent_size = 0;
4262         ffe_ctl->total_free_space = 0;
4263         ffe_ctl->found_offset = 0;
4264         ffe_ctl->policy = BTRFS_EXTENT_ALLOC_CLUSTERED;
4265
4266         if (btrfs_is_zoned(fs_info))
4267                 ffe_ctl->policy = BTRFS_EXTENT_ALLOC_ZONED;
4268
4269         ins->type = BTRFS_EXTENT_ITEM_KEY;
4270         ins->objectid = 0;
4271         ins->offset = 0;
4272
4273         trace_find_free_extent(root, ffe_ctl->num_bytes, ffe_ctl->empty_size,
4274                                ffe_ctl->flags);
4275
4276         space_info = btrfs_find_space_info(fs_info, ffe_ctl->flags);
4277         if (!space_info) {
4278                 btrfs_err(fs_info, "No space info for %llu", ffe_ctl->flags);
4279                 return -ENOSPC;
4280         }
4281
4282         ret = prepare_allocation(fs_info, ffe_ctl, space_info, ins);
4283         if (ret < 0)
4284                 return ret;
4285
4286         ffe_ctl->search_start = max(ffe_ctl->search_start,
4287                                     first_logical_byte(fs_info));
4288         ffe_ctl->search_start = max(ffe_ctl->search_start, ffe_ctl->hint_byte);
4289         if (ffe_ctl->search_start == ffe_ctl->hint_byte) {
4290                 block_group = btrfs_lookup_block_group(fs_info,
4291                                                        ffe_ctl->search_start);
4292                 /*
4293                  * we don't want to use the block group if it doesn't match our
4294                  * allocation bits, or if its not cached.
4295                  *
4296                  * However if we are re-searching with an ideal block group
4297                  * picked out then we don't care that the block group is cached.
4298                  */
4299                 if (block_group && block_group_bits(block_group, ffe_ctl->flags) &&
4300                     block_group->cached != BTRFS_CACHE_NO) {
4301                         down_read(&space_info->groups_sem);
4302                         if (list_empty(&block_group->list) ||
4303                             block_group->ro) {
4304                                 /*
4305                                  * someone is removing this block group,
4306                                  * we can't jump into the have_block_group
4307                                  * target because our list pointers are not
4308                                  * valid
4309                                  */
4310                                 btrfs_put_block_group(block_group);
4311                                 up_read(&space_info->groups_sem);
4312                         } else {
4313                                 ffe_ctl->index = btrfs_bg_flags_to_raid_index(
4314                                                         block_group->flags);
4315                                 btrfs_lock_block_group(block_group,
4316                                                        ffe_ctl->delalloc);
4317                                 goto have_block_group;
4318                         }
4319                 } else if (block_group) {
4320                         btrfs_put_block_group(block_group);
4321                 }
4322         }
4323 search:
4324         ffe_ctl->have_caching_bg = false;
4325         if (ffe_ctl->index == btrfs_bg_flags_to_raid_index(ffe_ctl->flags) ||
4326             ffe_ctl->index == 0)
4327                 full_search = true;
4328         down_read(&space_info->groups_sem);
4329         list_for_each_entry(block_group,
4330                             &space_info->block_groups[ffe_ctl->index], list) {
4331                 struct btrfs_block_group *bg_ret;
4332
4333                 /* If the block group is read-only, we can skip it entirely. */
4334                 if (unlikely(block_group->ro)) {
4335                         if (ffe_ctl->for_treelog)
4336                                 btrfs_clear_treelog_bg(block_group);
4337                         if (ffe_ctl->for_data_reloc)
4338                                 btrfs_clear_data_reloc_bg(block_group);
4339                         continue;
4340                 }
4341
4342                 btrfs_grab_block_group(block_group, ffe_ctl->delalloc);
4343                 ffe_ctl->search_start = block_group->start;
4344
4345                 /*
4346                  * this can happen if we end up cycling through all the
4347                  * raid types, but we want to make sure we only allocate
4348                  * for the proper type.
4349                  */
4350                 if (!block_group_bits(block_group, ffe_ctl->flags)) {
4351                         u64 extra = BTRFS_BLOCK_GROUP_DUP |
4352                                 BTRFS_BLOCK_GROUP_RAID1_MASK |
4353                                 BTRFS_BLOCK_GROUP_RAID56_MASK |
4354                                 BTRFS_BLOCK_GROUP_RAID10;
4355
4356                         /*
4357                          * if they asked for extra copies and this block group
4358                          * doesn't provide them, bail.  This does allow us to
4359                          * fill raid0 from raid1.
4360                          */
4361                         if ((ffe_ctl->flags & extra) && !(block_group->flags & extra))
4362                                 goto loop;
4363
4364                         /*
4365                          * This block group has different flags than we want.
4366                          * It's possible that we have MIXED_GROUP flag but no
4367                          * block group is mixed.  Just skip such block group.
4368                          */
4369                         btrfs_release_block_group(block_group, ffe_ctl->delalloc);
4370                         continue;
4371                 }
4372
4373 have_block_group:
4374                 ffe_ctl->cached = btrfs_block_group_done(block_group);
4375                 if (unlikely(!ffe_ctl->cached)) {
4376                         ffe_ctl->have_caching_bg = true;
4377                         ret = btrfs_cache_block_group(block_group, 0);
4378
4379                         /*
4380                          * If we get ENOMEM here or something else we want to
4381                          * try other block groups, because it may not be fatal.
4382                          * However if we can't find anything else we need to
4383                          * save our return here so that we return the actual
4384                          * error that caused problems, not ENOSPC.
4385                          */
4386                         if (ret < 0) {
4387                                 if (!cache_block_group_error)
4388                                         cache_block_group_error = ret;
4389                                 ret = 0;
4390                                 goto loop;
4391                         }
4392                         ret = 0;
4393                 }
4394
4395                 if (unlikely(block_group->cached == BTRFS_CACHE_ERROR))
4396                         goto loop;
4397
4398                 bg_ret = NULL;
4399                 ret = do_allocation(block_group, ffe_ctl, &bg_ret);
4400                 if (ret == 0) {
4401                         if (bg_ret && bg_ret != block_group) {
4402                                 btrfs_release_block_group(block_group,
4403                                                           ffe_ctl->delalloc);
4404                                 block_group = bg_ret;
4405                         }
4406                 } else if (ret == -EAGAIN) {
4407                         goto have_block_group;
4408                 } else if (ret > 0) {
4409                         goto loop;
4410                 }
4411
4412                 /* Checks */
4413                 ffe_ctl->search_start = round_up(ffe_ctl->found_offset,
4414                                                  fs_info->stripesize);
4415
4416                 /* move on to the next group */
4417                 if (ffe_ctl->search_start + ffe_ctl->num_bytes >
4418                     block_group->start + block_group->length) {
4419                         btrfs_add_free_space_unused(block_group,
4420                                             ffe_ctl->found_offset,
4421                                             ffe_ctl->num_bytes);
4422                         goto loop;
4423                 }
4424
4425                 if (ffe_ctl->found_offset < ffe_ctl->search_start)
4426                         btrfs_add_free_space_unused(block_group,
4427                                         ffe_ctl->found_offset,
4428                                         ffe_ctl->search_start - ffe_ctl->found_offset);
4429
4430                 ret = btrfs_add_reserved_bytes(block_group, ffe_ctl->ram_bytes,
4431                                                ffe_ctl->num_bytes,
4432                                                ffe_ctl->delalloc);
4433                 if (ret == -EAGAIN) {
4434                         btrfs_add_free_space_unused(block_group,
4435                                         ffe_ctl->found_offset,
4436                                         ffe_ctl->num_bytes);
4437                         goto loop;
4438                 }
4439                 btrfs_inc_block_group_reservations(block_group);
4440
4441                 /* we are all good, lets return */
4442                 ins->objectid = ffe_ctl->search_start;
4443                 ins->offset = ffe_ctl->num_bytes;
4444
4445                 trace_btrfs_reserve_extent(block_group, ffe_ctl->search_start,
4446                                            ffe_ctl->num_bytes);
4447                 btrfs_release_block_group(block_group, ffe_ctl->delalloc);
4448                 break;
4449 loop:
4450                 release_block_group(block_group, ffe_ctl, ffe_ctl->delalloc);
4451                 cond_resched();
4452         }
4453         up_read(&space_info->groups_sem);
4454
4455         ret = find_free_extent_update_loop(fs_info, ins, ffe_ctl, full_search);
4456         if (ret > 0)
4457                 goto search;
4458
4459         if (ret == -ENOSPC && !cache_block_group_error) {
4460                 /*
4461                  * Use ffe_ctl->total_free_space as fallback if we can't find
4462                  * any contiguous hole.
4463                  */
4464                 if (!ffe_ctl->max_extent_size)
4465                         ffe_ctl->max_extent_size = ffe_ctl->total_free_space;
4466                 spin_lock(&space_info->lock);
4467                 space_info->max_extent_size = ffe_ctl->max_extent_size;
4468                 spin_unlock(&space_info->lock);
4469                 ins->offset = ffe_ctl->max_extent_size;
4470         } else if (ret == -ENOSPC) {
4471                 ret = cache_block_group_error;
4472         }
4473         return ret;
4474 }
4475
4476 /*
4477  * btrfs_reserve_extent - entry point to the extent allocator. Tries to find a
4478  *                        hole that is at least as big as @num_bytes.
4479  *
4480  * @root           -    The root that will contain this extent
4481  *
4482  * @ram_bytes      -    The amount of space in ram that @num_bytes take. This
4483  *                      is used for accounting purposes. This value differs
4484  *                      from @num_bytes only in the case of compressed extents.
4485  *
4486  * @num_bytes      -    Number of bytes to allocate on-disk.
4487  *
4488  * @min_alloc_size -    Indicates the minimum amount of space that the
4489  *                      allocator should try to satisfy. In some cases
4490  *                      @num_bytes may be larger than what is required and if
4491  *                      the filesystem is fragmented then allocation fails.
4492  *                      However, the presence of @min_alloc_size gives a
4493  *                      chance to try and satisfy the smaller allocation.
4494  *
4495  * @empty_size     -    A hint that you plan on doing more COW. This is the
4496  *                      size in bytes the allocator should try to find free
4497  *                      next to the block it returns.  This is just a hint and
4498  *                      may be ignored by the allocator.
4499  *
4500  * @hint_byte      -    Hint to the allocator to start searching above the byte
4501  *                      address passed. It might be ignored.
4502  *
4503  * @ins            -    This key is modified to record the found hole. It will
4504  *                      have the following values:
4505  *                      ins->objectid == start position
4506  *                      ins->flags = BTRFS_EXTENT_ITEM_KEY
4507  *                      ins->offset == the size of the hole.
4508  *
4509  * @is_data        -    Boolean flag indicating whether an extent is
4510  *                      allocated for data (true) or metadata (false)
4511  *
4512  * @delalloc       -    Boolean flag indicating whether this allocation is for
4513  *                      delalloc or not. If 'true' data_rwsem of block groups
4514  *                      is going to be acquired.
4515  *
4516  *
4517  * Returns 0 when an allocation succeeded or < 0 when an error occurred. In
4518  * case -ENOSPC is returned then @ins->offset will contain the size of the
4519  * largest available hole the allocator managed to find.
4520  */
4521 int btrfs_reserve_extent(struct btrfs_root *root, u64 ram_bytes,
4522                          u64 num_bytes, u64 min_alloc_size,
4523                          u64 empty_size, u64 hint_byte,
4524                          struct btrfs_key *ins, int is_data, int delalloc)
4525 {
4526         struct btrfs_fs_info *fs_info = root->fs_info;
4527         struct find_free_extent_ctl ffe_ctl = {};
4528         bool final_tried = num_bytes == min_alloc_size;
4529         u64 flags;
4530         int ret;
4531         bool for_treelog = (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID);
4532         bool for_data_reloc = (btrfs_is_data_reloc_root(root) && is_data);
4533
4534         flags = get_alloc_profile_by_root(root, is_data);
4535 again:
4536         WARN_ON(num_bytes < fs_info->sectorsize);
4537
4538         ffe_ctl.ram_bytes = ram_bytes;
4539         ffe_ctl.num_bytes = num_bytes;
4540         ffe_ctl.min_alloc_size = min_alloc_size;
4541         ffe_ctl.empty_size = empty_size;
4542         ffe_ctl.flags = flags;
4543         ffe_ctl.delalloc = delalloc;
4544         ffe_ctl.hint_byte = hint_byte;
4545         ffe_ctl.for_treelog = for_treelog;
4546         ffe_ctl.for_data_reloc = for_data_reloc;
4547
4548         ret = find_free_extent(root, ins, &ffe_ctl);
4549         if (!ret && !is_data) {
4550                 btrfs_dec_block_group_reservations(fs_info, ins->objectid);
4551         } else if (ret == -ENOSPC) {
4552                 if (!final_tried && ins->offset) {
4553                         num_bytes = min(num_bytes >> 1, ins->offset);
4554                         num_bytes = round_down(num_bytes,
4555                                                fs_info->sectorsize);
4556                         num_bytes = max(num_bytes, min_alloc_size);
4557                         ram_bytes = num_bytes;
4558                         if (num_bytes == min_alloc_size)
4559                                 final_tried = true;
4560                         goto again;
4561                 } else if (btrfs_test_opt(fs_info, ENOSPC_DEBUG)) {
4562                         struct btrfs_space_info *sinfo;
4563
4564                         sinfo = btrfs_find_space_info(fs_info, flags);
4565                         btrfs_err(fs_info,
4566         "allocation failed flags %llu, wanted %llu tree-log %d, relocation: %d",
4567                                   flags, num_bytes, for_treelog, for_data_reloc);
4568                         if (sinfo)
4569                                 btrfs_dump_space_info(fs_info, sinfo,
4570                                                       num_bytes, 1);
4571                 }
4572         }
4573
4574         return ret;
4575 }
4576
4577 int btrfs_free_reserved_extent(struct btrfs_fs_info *fs_info,
4578                                u64 start, u64 len, int delalloc)
4579 {
4580         struct btrfs_block_group *cache;
4581
4582         cache = btrfs_lookup_block_group(fs_info, start);
4583         if (!cache) {
4584                 btrfs_err(fs_info, "Unable to find block group for %llu",
4585                           start);
4586                 return -ENOSPC;
4587         }
4588
4589         btrfs_add_free_space(cache, start, len);
4590         btrfs_free_reserved_bytes(cache, len, delalloc);
4591         trace_btrfs_reserved_extent_free(fs_info, start, len);
4592
4593         btrfs_put_block_group(cache);
4594         return 0;
4595 }
4596
4597 int btrfs_pin_reserved_extent(struct btrfs_trans_handle *trans, u64 start,
4598                               u64 len)
4599 {
4600         struct btrfs_block_group *cache;
4601         int ret = 0;
4602
4603         cache = btrfs_lookup_block_group(trans->fs_info, start);
4604         if (!cache) {
4605                 btrfs_err(trans->fs_info, "unable to find block group for %llu",
4606                           start);
4607                 return -ENOSPC;
4608         }
4609
4610         ret = pin_down_extent(trans, cache, start, len, 1);
4611         btrfs_put_block_group(cache);
4612         return ret;
4613 }
4614
4615 static int alloc_reserved_extent(struct btrfs_trans_handle *trans, u64 bytenr,
4616                                  u64 num_bytes)
4617 {
4618         struct btrfs_fs_info *fs_info = trans->fs_info;
4619         int ret;
4620
4621         ret = remove_from_free_space_tree(trans, bytenr, num_bytes);
4622         if (ret)
4623                 return ret;
4624
4625         ret = btrfs_update_block_group(trans, bytenr, num_bytes, true);
4626         if (ret) {
4627                 ASSERT(!ret);
4628                 btrfs_err(fs_info, "update block group failed for %llu %llu",
4629                           bytenr, num_bytes);
4630                 return ret;
4631         }
4632
4633         trace_btrfs_reserved_extent_alloc(fs_info, bytenr, num_bytes);
4634         return 0;
4635 }
4636
4637 static int alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
4638                                       u64 parent, u64 root_objectid,
4639                                       u64 flags, u64 owner, u64 offset,
4640                                       struct btrfs_key *ins, int ref_mod)
4641 {
4642         struct btrfs_fs_info *fs_info = trans->fs_info;
4643         struct btrfs_root *extent_root;
4644         int ret;
4645         struct btrfs_extent_item *extent_item;
4646         struct btrfs_extent_inline_ref *iref;
4647         struct btrfs_path *path;
4648         struct extent_buffer *leaf;
4649         int type;
4650         u32 size;
4651
4652         if (parent > 0)
4653                 type = BTRFS_SHARED_DATA_REF_KEY;
4654         else
4655                 type = BTRFS_EXTENT_DATA_REF_KEY;
4656
4657         size = sizeof(*extent_item) + btrfs_extent_inline_ref_size(type);
4658
4659         path = btrfs_alloc_path();
4660         if (!path)
4661                 return -ENOMEM;
4662
4663         extent_root = btrfs_extent_root(fs_info, ins->objectid);
4664         ret = btrfs_insert_empty_item(trans, extent_root, path, ins, size);
4665         if (ret) {
4666                 btrfs_free_path(path);
4667                 return ret;
4668         }
4669
4670         leaf = path->nodes[0];
4671         extent_item = btrfs_item_ptr(leaf, path->slots[0],
4672                                      struct btrfs_extent_item);
4673         btrfs_set_extent_refs(leaf, extent_item, ref_mod);
4674         btrfs_set_extent_generation(leaf, extent_item, trans->transid);
4675         btrfs_set_extent_flags(leaf, extent_item,
4676                                flags | BTRFS_EXTENT_FLAG_DATA);
4677
4678         iref = (struct btrfs_extent_inline_ref *)(extent_item + 1);
4679         btrfs_set_extent_inline_ref_type(leaf, iref, type);
4680         if (parent > 0) {
4681                 struct btrfs_shared_data_ref *ref;
4682                 ref = (struct btrfs_shared_data_ref *)(iref + 1);
4683                 btrfs_set_extent_inline_ref_offset(leaf, iref, parent);
4684                 btrfs_set_shared_data_ref_count(leaf, ref, ref_mod);
4685         } else {
4686                 struct btrfs_extent_data_ref *ref;
4687                 ref = (struct btrfs_extent_data_ref *)(&iref->offset);
4688                 btrfs_set_extent_data_ref_root(leaf, ref, root_objectid);
4689                 btrfs_set_extent_data_ref_objectid(leaf, ref, owner);
4690                 btrfs_set_extent_data_ref_offset(leaf, ref, offset);
4691                 btrfs_set_extent_data_ref_count(leaf, ref, ref_mod);
4692         }
4693
4694         btrfs_mark_buffer_dirty(path->nodes[0]);
4695         btrfs_free_path(path);
4696
4697         return alloc_reserved_extent(trans, ins->objectid, ins->offset);
4698 }
4699
4700 static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,
4701                                      struct btrfs_delayed_ref_node *node,
4702                                      struct btrfs_delayed_extent_op *extent_op)
4703 {
4704         struct btrfs_fs_info *fs_info = trans->fs_info;
4705         struct btrfs_root *extent_root;
4706         int ret;
4707         struct btrfs_extent_item *extent_item;
4708         struct btrfs_key extent_key;
4709         struct btrfs_tree_block_info *block_info;
4710         struct btrfs_extent_inline_ref *iref;
4711         struct btrfs_path *path;
4712         struct extent_buffer *leaf;
4713         struct btrfs_delayed_tree_ref *ref;
4714         u32 size = sizeof(*extent_item) + sizeof(*iref);
4715         u64 flags = extent_op->flags_to_set;
4716         bool skinny_metadata = btrfs_fs_incompat(fs_info, SKINNY_METADATA);
4717
4718         ref = btrfs_delayed_node_to_tree_ref(node);
4719
4720         extent_key.objectid = node->bytenr;
4721         if (skinny_metadata) {
4722                 extent_key.offset = ref->level;
4723                 extent_key.type = BTRFS_METADATA_ITEM_KEY;
4724         } else {
4725                 extent_key.offset = node->num_bytes;
4726                 extent_key.type = BTRFS_EXTENT_ITEM_KEY;
4727                 size += sizeof(*block_info);
4728         }
4729
4730         path = btrfs_alloc_path();
4731         if (!path)
4732                 return -ENOMEM;
4733
4734         extent_root = btrfs_extent_root(fs_info, extent_key.objectid);
4735         ret = btrfs_insert_empty_item(trans, extent_root, path, &extent_key,
4736                                       size);
4737         if (ret) {
4738                 btrfs_free_path(path);
4739                 return ret;
4740         }
4741
4742         leaf = path->nodes[0];
4743         extent_item = btrfs_item_ptr(leaf, path->slots[0],
4744                                      struct btrfs_extent_item);
4745         btrfs_set_extent_refs(leaf, extent_item, 1);
4746         btrfs_set_extent_generation(leaf, extent_item, trans->transid);
4747         btrfs_set_extent_flags(leaf, extent_item,
4748                                flags | BTRFS_EXTENT_FLAG_TREE_BLOCK);
4749
4750         if (skinny_metadata) {
4751                 iref = (struct btrfs_extent_inline_ref *)(extent_item + 1);
4752         } else {
4753                 block_info = (struct btrfs_tree_block_info *)(extent_item + 1);
4754                 btrfs_set_tree_block_key(leaf, block_info, &extent_op->key);
4755                 btrfs_set_tree_block_level(leaf, block_info, ref->level);
4756                 iref = (struct btrfs_extent_inline_ref *)(block_info + 1);
4757         }
4758
4759         if (node->type == BTRFS_SHARED_BLOCK_REF_KEY) {
4760                 btrfs_set_extent_inline_ref_type(leaf, iref,
4761                                                  BTRFS_SHARED_BLOCK_REF_KEY);
4762                 btrfs_set_extent_inline_ref_offset(leaf, iref, ref->parent);
4763         } else {
4764                 btrfs_set_extent_inline_ref_type(leaf, iref,
4765                                                  BTRFS_TREE_BLOCK_REF_KEY);
4766                 btrfs_set_extent_inline_ref_offset(leaf, iref, ref->root);
4767         }
4768
4769         btrfs_mark_buffer_dirty(leaf);
4770         btrfs_free_path(path);
4771
4772         return alloc_reserved_extent(trans, node->bytenr, fs_info->nodesize);
4773 }
4774
4775 int btrfs_alloc_reserved_file_extent(struct btrfs_trans_handle *trans,
4776                                      struct btrfs_root *root, u64 owner,
4777                                      u64 offset, u64 ram_bytes,
4778                                      struct btrfs_key *ins)
4779 {
4780         struct btrfs_ref generic_ref = { 0 };
4781
4782         BUG_ON(root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID);
4783
4784         btrfs_init_generic_ref(&generic_ref, BTRFS_ADD_DELAYED_EXTENT,
4785                                ins->objectid, ins->offset, 0);
4786         btrfs_init_data_ref(&generic_ref, root->root_key.objectid, owner,
4787                             offset, 0, false);
4788         btrfs_ref_tree_mod(root->fs_info, &generic_ref);
4789
4790         return btrfs_add_delayed_data_ref(trans, &generic_ref, ram_bytes);
4791 }
4792
4793 /*
4794  * this is used by the tree logging recovery code.  It records that
4795  * an extent has been allocated and makes sure to clear the free
4796  * space cache bits as well
4797  */
4798 int btrfs_alloc_logged_file_extent(struct btrfs_trans_handle *trans,
4799                                    u64 root_objectid, u64 owner, u64 offset,
4800                                    struct btrfs_key *ins)
4801 {
4802         struct btrfs_fs_info *fs_info = trans->fs_info;
4803         int ret;
4804         struct btrfs_block_group *block_group;
4805         struct btrfs_space_info *space_info;
4806
4807         /*
4808          * Mixed block groups will exclude before processing the log so we only
4809          * need to do the exclude dance if this fs isn't mixed.
4810          */
4811         if (!btrfs_fs_incompat(fs_info, MIXED_GROUPS)) {
4812                 ret = __exclude_logged_extent(fs_info, ins->objectid,
4813                                               ins->offset);
4814                 if (ret)
4815                         return ret;
4816         }
4817
4818         block_group = btrfs_lookup_block_group(fs_info, ins->objectid);
4819         if (!block_group)
4820                 return -EINVAL;
4821
4822         space_info = block_group->space_info;
4823         spin_lock(&space_info->lock);
4824         spin_lock(&block_group->lock);
4825         space_info->bytes_reserved += ins->offset;
4826         block_group->reserved += ins->offset;
4827         spin_unlock(&block_group->lock);
4828         spin_unlock(&space_info->lock);
4829
4830         ret = alloc_reserved_file_extent(trans, 0, root_objectid, 0, owner,
4831                                          offset, ins, 1);
4832         if (ret)
4833                 btrfs_pin_extent(trans, ins->objectid, ins->offset, 1);
4834         btrfs_put_block_group(block_group);
4835         return ret;
4836 }
4837
4838 static struct extent_buffer *
4839 btrfs_init_new_buffer(struct btrfs_trans_handle *trans, struct btrfs_root *root,
4840                       u64 bytenr, int level, u64 owner,
4841                       enum btrfs_lock_nesting nest)
4842 {
4843         struct btrfs_fs_info *fs_info = root->fs_info;
4844         struct extent_buffer *buf;
4845
4846         buf = btrfs_find_create_tree_block(fs_info, bytenr, owner, level);
4847         if (IS_ERR(buf))
4848                 return buf;
4849
4850         /*
4851          * Extra safety check in case the extent tree is corrupted and extent
4852          * allocator chooses to use a tree block which is already used and
4853          * locked.
4854          */
4855         if (buf->lock_owner == current->pid) {
4856                 btrfs_err_rl(fs_info,
4857 "tree block %llu owner %llu already locked by pid=%d, extent tree corruption detected",
4858                         buf->start, btrfs_header_owner(buf), current->pid);
4859                 free_extent_buffer(buf);
4860                 return ERR_PTR(-EUCLEAN);
4861         }
4862
4863         /*
4864          * This needs to stay, because we could allocate a freed block from an
4865          * old tree into a new tree, so we need to make sure this new block is
4866          * set to the appropriate level and owner.
4867          */
4868         btrfs_set_buffer_lockdep_class(owner, buf, level);
4869         __btrfs_tree_lock(buf, nest);
4870         btrfs_clean_tree_block(buf);
4871         clear_bit(EXTENT_BUFFER_STALE, &buf->bflags);
4872         clear_bit(EXTENT_BUFFER_NO_CHECK, &buf->bflags);
4873
4874         set_extent_buffer_uptodate(buf);
4875
4876         memzero_extent_buffer(buf, 0, sizeof(struct btrfs_header));
4877         btrfs_set_header_level(buf, level);
4878         btrfs_set_header_bytenr(buf, buf->start);
4879         btrfs_set_header_generation(buf, trans->transid);
4880         btrfs_set_header_backref_rev(buf, BTRFS_MIXED_BACKREF_REV);
4881         btrfs_set_header_owner(buf, owner);
4882         write_extent_buffer_fsid(buf, fs_info->fs_devices->metadata_uuid);
4883         write_extent_buffer_chunk_tree_uuid(buf, fs_info->chunk_tree_uuid);
4884         if (root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID) {
4885                 buf->log_index = root->log_transid % 2;
4886                 /*
4887                  * we allow two log transactions at a time, use different
4888                  * EXTENT bit to differentiate dirty pages.
4889                  */
4890                 if (buf->log_index == 0)
4891                         set_extent_dirty(&root->dirty_log_pages, buf->start,
4892                                         buf->start + buf->len - 1, GFP_NOFS);
4893                 else
4894                         set_extent_new(&root->dirty_log_pages, buf->start,
4895                                         buf->start + buf->len - 1);
4896         } else {
4897                 buf->log_index = -1;
4898                 set_extent_dirty(&trans->transaction->dirty_pages, buf->start,
4899                          buf->start + buf->len - 1, GFP_NOFS);
4900         }
4901         /* this returns a buffer locked for blocking */
4902         return buf;
4903 }
4904
4905 /*
4906  * finds a free extent and does all the dirty work required for allocation
4907  * returns the tree buffer or an ERR_PTR on error.
4908  */
4909 struct extent_buffer *btrfs_alloc_tree_block(struct btrfs_trans_handle *trans,
4910                                              struct btrfs_root *root,
4911                                              u64 parent, u64 root_objectid,
4912                                              const struct btrfs_disk_key *key,
4913                                              int level, u64 hint,
4914                                              u64 empty_size,
4915                                              enum btrfs_lock_nesting nest)
4916 {
4917         struct btrfs_fs_info *fs_info = root->fs_info;
4918         struct btrfs_key ins;
4919         struct btrfs_block_rsv *block_rsv;
4920         struct extent_buffer *buf;
4921         struct btrfs_delayed_extent_op *extent_op;
4922         struct btrfs_ref generic_ref = { 0 };
4923         u64 flags = 0;
4924         int ret;
4925         u32 blocksize = fs_info->nodesize;
4926         bool skinny_metadata = btrfs_fs_incompat(fs_info, SKINNY_METADATA);
4927
4928 #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
4929         if (btrfs_is_testing(fs_info)) {
4930                 buf = btrfs_init_new_buffer(trans, root, root->alloc_bytenr,
4931                                             level, root_objectid, nest);
4932                 if (!IS_ERR(buf))
4933                         root->alloc_bytenr += blocksize;
4934                 return buf;
4935         }
4936 #endif
4937
4938         block_rsv = btrfs_use_block_rsv(trans, root, blocksize);
4939         if (IS_ERR(block_rsv))
4940                 return ERR_CAST(block_rsv);
4941
4942         ret = btrfs_reserve_extent(root, blocksize, blocksize, blocksize,
4943                                    empty_size, hint, &ins, 0, 0);
4944         if (ret)
4945                 goto out_unuse;
4946
4947         buf = btrfs_init_new_buffer(trans, root, ins.objectid, level,
4948                                     root_objectid, nest);
4949         if (IS_ERR(buf)) {
4950                 ret = PTR_ERR(buf);
4951                 goto out_free_reserved;
4952         }
4953
4954         if (root_objectid == BTRFS_TREE_RELOC_OBJECTID) {
4955                 if (parent == 0)
4956                         parent = ins.objectid;
4957                 flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF;
4958         } else
4959                 BUG_ON(parent > 0);
4960
4961         if (root_objectid != BTRFS_TREE_LOG_OBJECTID) {
4962                 extent_op = btrfs_alloc_delayed_extent_op();
4963                 if (!extent_op) {
4964                         ret = -ENOMEM;
4965                         goto out_free_buf;
4966                 }
4967                 if (key)
4968                         memcpy(&extent_op->key, key, sizeof(extent_op->key));
4969                 else
4970                         memset(&extent_op->key, 0, sizeof(extent_op->key));
4971                 extent_op->flags_to_set = flags;
4972                 extent_op->update_key = skinny_metadata ? false : true;
4973                 extent_op->update_flags = true;
4974                 extent_op->level = level;
4975
4976                 btrfs_init_generic_ref(&generic_ref, BTRFS_ADD_DELAYED_EXTENT,
4977                                        ins.objectid, ins.offset, parent);
4978                 btrfs_init_tree_ref(&generic_ref, level, root_objectid,
4979                                     root->root_key.objectid, false);
4980                 btrfs_ref_tree_mod(fs_info, &generic_ref);
4981                 ret = btrfs_add_delayed_tree_ref(trans, &generic_ref, extent_op);
4982                 if (ret)
4983                         goto out_free_delayed;
4984         }
4985         return buf;
4986
4987 out_free_delayed:
4988         btrfs_free_delayed_extent_op(extent_op);
4989 out_free_buf:
4990         btrfs_tree_unlock(buf);
4991         free_extent_buffer(buf);
4992 out_free_reserved:
4993         btrfs_free_reserved_extent(fs_info, ins.objectid, ins.offset, 0);
4994 out_unuse:
4995         btrfs_unuse_block_rsv(fs_info, block_rsv, blocksize);
4996         return ERR_PTR(ret);
4997 }
4998
4999 struct walk_control {
5000         u64 refs[BTRFS_MAX_LEVEL];
5001         u64 flags[BTRFS_MAX_LEVEL];
5002         struct btrfs_key update_progress;
5003         struct btrfs_key drop_progress;
5004         int drop_level;
5005         int stage;
5006         int level;
5007         int shared_level;
5008         int update_ref;
5009         int keep_locks;
5010         int reada_slot;
5011         int reada_count;
5012         int restarted;
5013 };
5014
5015 #define DROP_REFERENCE  1
5016 #define UPDATE_BACKREF  2
5017
5018 static noinline void reada_walk_down(struct btrfs_trans_handle *trans,
5019                                      struct btrfs_root *root,
5020                                      struct walk_control *wc,
5021                                      struct btrfs_path *path)
5022 {
5023         struct btrfs_fs_info *fs_info = root->fs_info;
5024         u64 bytenr;
5025         u64 generation;
5026         u64 refs;
5027         u64 flags;
5028         u32 nritems;
5029         struct btrfs_key key;
5030         struct extent_buffer *eb;
5031         int ret;
5032         int slot;
5033         int nread = 0;
5034
5035         if (path->slots[wc->level] < wc->reada_slot) {
5036                 wc->reada_count = wc->reada_count * 2 / 3;
5037                 wc->reada_count = max(wc->reada_count, 2);
5038         } else {
5039                 wc->reada_count = wc->reada_count * 3 / 2;
5040                 wc->reada_count = min_t(int, wc->reada_count,
5041                                         BTRFS_NODEPTRS_PER_BLOCK(fs_info));
5042         }
5043
5044         eb = path->nodes[wc->level];
5045         nritems = btrfs_header_nritems(eb);
5046
5047         for (slot = path->slots[wc->level]; slot < nritems; slot++) {
5048                 if (nread >= wc->reada_count)
5049                         break;
5050
5051                 cond_resched();
5052                 bytenr = btrfs_node_blockptr(eb, slot);
5053                 generation = btrfs_node_ptr_generation(eb, slot);
5054
5055                 if (slot == path->slots[wc->level])
5056                         goto reada;
5057
5058                 if (wc->stage == UPDATE_BACKREF &&
5059                     generation <= root->root_key.offset)
5060                         continue;
5061
5062                 /* We don't lock the tree block, it's OK to be racy here */
5063                 ret = btrfs_lookup_extent_info(trans, fs_info, bytenr,
5064                                                wc->level - 1, 1, &refs,
5065                                                &flags);
5066                 /* We don't care about errors in readahead. */
5067                 if (ret < 0)
5068                         continue;
5069                 BUG_ON(refs == 0);
5070
5071                 if (wc->stage == DROP_REFERENCE) {
5072                         if (refs == 1)
5073                                 goto reada;
5074
5075                         if (wc->level == 1 &&
5076                             (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF))
5077                                 continue;
5078                         if (!wc->update_ref ||
5079                             generation <= root->root_key.offset)
5080                                 continue;
5081                         btrfs_node_key_to_cpu(eb, &key, slot);
5082                         ret = btrfs_comp_cpu_keys(&key,
5083                                                   &wc->update_progress);
5084                         if (ret < 0)
5085                                 continue;
5086                 } else {
5087                         if (wc->level == 1 &&
5088                             (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF))
5089                                 continue;
5090                 }
5091 reada:
5092                 btrfs_readahead_node_child(eb, slot);
5093                 nread++;
5094         }
5095         wc->reada_slot = slot;
5096 }
5097
5098 /*
5099  * helper to process tree block while walking down the tree.
5100  *
5101  * when wc->stage == UPDATE_BACKREF, this function updates
5102  * back refs for pointers in the block.
5103  *
5104  * NOTE: return value 1 means we should stop walking down.
5105  */
5106 static noinline int walk_down_proc(struct btrfs_trans_handle *trans,
5107                                    struct btrfs_root *root,
5108                                    struct btrfs_path *path,
5109                                    struct walk_control *wc, int lookup_info)
5110 {
5111         struct btrfs_fs_info *fs_info = root->fs_info;
5112         int level = wc->level;
5113         struct extent_buffer *eb = path->nodes[level];
5114         u64 flag = BTRFS_BLOCK_FLAG_FULL_BACKREF;
5115         int ret;
5116
5117         if (wc->stage == UPDATE_BACKREF &&
5118             btrfs_header_owner(eb) != root->root_key.objectid)
5119                 return 1;
5120
5121         /*
5122          * when reference count of tree block is 1, it won't increase
5123          * again. once full backref flag is set, we never clear it.
5124          */
5125         if (lookup_info &&
5126             ((wc->stage == DROP_REFERENCE && wc->refs[level] != 1) ||
5127              (wc->stage == UPDATE_BACKREF && !(wc->flags[level] & flag)))) {
5128                 BUG_ON(!path->locks[level]);
5129                 ret = btrfs_lookup_extent_info(trans, fs_info,
5130                                                eb->start, level, 1,
5131                                                &wc->refs[level],
5132                                                &wc->flags[level]);
5133                 BUG_ON(ret == -ENOMEM);
5134                 if (ret)
5135                         return ret;
5136                 BUG_ON(wc->refs[level] == 0);
5137         }
5138
5139         if (wc->stage == DROP_REFERENCE) {
5140                 if (wc->refs[level] > 1)
5141                         return 1;
5142
5143                 if (path->locks[level] && !wc->keep_locks) {
5144                         btrfs_tree_unlock_rw(eb, path->locks[level]);
5145                         path->locks[level] = 0;
5146                 }
5147                 return 0;
5148         }
5149
5150         /* wc->stage == UPDATE_BACKREF */
5151         if (!(wc->flags[level] & flag)) {
5152                 BUG_ON(!path->locks[level]);
5153                 ret = btrfs_inc_ref(trans, root, eb, 1);
5154                 BUG_ON(ret); /* -ENOMEM */
5155                 ret = btrfs_dec_ref(trans, root, eb, 0);
5156                 BUG_ON(ret); /* -ENOMEM */
5157                 ret = btrfs_set_disk_extent_flags(trans, eb, flag,
5158                                                   btrfs_header_level(eb));
5159                 BUG_ON(ret); /* -ENOMEM */
5160                 wc->flags[level] |= flag;
5161         }
5162
5163         /*
5164          * the block is shared by multiple trees, so it's not good to
5165          * keep the tree lock
5166          */
5167         if (path->locks[level] && level > 0) {
5168                 btrfs_tree_unlock_rw(eb, path->locks[level]);
5169                 path->locks[level] = 0;
5170         }
5171         return 0;
5172 }
5173
5174 /*
5175  * This is used to verify a ref exists for this root to deal with a bug where we
5176  * would have a drop_progress key that hadn't been updated properly.
5177  */
5178 static int check_ref_exists(struct btrfs_trans_handle *trans,
5179                             struct btrfs_root *root, u64 bytenr, u64 parent,
5180                             int level)
5181 {
5182         struct btrfs_path *path;
5183         struct btrfs_extent_inline_ref *iref;
5184         int ret;
5185
5186         path = btrfs_alloc_path();
5187         if (!path)
5188                 return -ENOMEM;
5189
5190         ret = lookup_extent_backref(trans, path, &iref, bytenr,
5191                                     root->fs_info->nodesize, parent,
5192                                     root->root_key.objectid, level, 0);
5193         btrfs_free_path(path);
5194         if (ret == -ENOENT)
5195                 return 0;
5196         if (ret < 0)
5197                 return ret;
5198         return 1;
5199 }
5200
5201 /*
5202  * helper to process tree block pointer.
5203  *
5204  * when wc->stage == DROP_REFERENCE, this function checks
5205  * reference count of the block pointed to. if the block
5206  * is shared and we need update back refs for the subtree
5207  * rooted at the block, this function changes wc->stage to
5208  * UPDATE_BACKREF. if the block is shared and there is no
5209  * need to update back, this function drops the reference
5210  * to the block.
5211  *
5212  * NOTE: return value 1 means we should stop walking down.
5213  */
5214 static noinline int do_walk_down(struct btrfs_trans_handle *trans,
5215                                  struct btrfs_root *root,
5216                                  struct btrfs_path *path,
5217                                  struct walk_control *wc, int *lookup_info)
5218 {
5219         struct btrfs_fs_info *fs_info = root->fs_info;
5220         u64 bytenr;
5221         u64 generation;
5222         u64 parent;
5223         struct btrfs_key key;
5224         struct btrfs_key first_key;
5225         struct btrfs_ref ref = { 0 };
5226         struct extent_buffer *next;
5227         int level = wc->level;
5228         int reada = 0;
5229         int ret = 0;
5230         bool need_account = false;
5231
5232         generation = btrfs_node_ptr_generation(path->nodes[level],
5233                                                path->slots[level]);
5234         /*
5235          * if the lower level block was created before the snapshot
5236          * was created, we know there is no need to update back refs
5237          * for the subtree
5238          */
5239         if (wc->stage == UPDATE_BACKREF &&
5240             generation <= root->root_key.offset) {
5241                 *lookup_info = 1;
5242                 return 1;
5243         }
5244
5245         bytenr = btrfs_node_blockptr(path->nodes[level], path->slots[level]);
5246         btrfs_node_key_to_cpu(path->nodes[level], &first_key,
5247                               path->slots[level]);
5248
5249         next = find_extent_buffer(fs_info, bytenr);
5250         if (!next) {
5251                 next = btrfs_find_create_tree_block(fs_info, bytenr,
5252                                 root->root_key.objectid, level - 1);
5253                 if (IS_ERR(next))
5254                         return PTR_ERR(next);
5255                 reada = 1;
5256         }
5257         btrfs_tree_lock(next);
5258
5259         ret = btrfs_lookup_extent_info(trans, fs_info, bytenr, level - 1, 1,
5260                                        &wc->refs[level - 1],
5261                                        &wc->flags[level - 1]);
5262         if (ret < 0)
5263                 goto out_unlock;
5264
5265         if (unlikely(wc->refs[level - 1] == 0)) {
5266                 btrfs_err(fs_info, "Missing references.");
5267                 ret = -EIO;
5268                 goto out_unlock;
5269         }
5270         *lookup_info = 0;
5271
5272         if (wc->stage == DROP_REFERENCE) {
5273                 if (wc->refs[level - 1] > 1) {
5274                         need_account = true;
5275                         if (level == 1 &&
5276                             (wc->flags[0] & BTRFS_BLOCK_FLAG_FULL_BACKREF))
5277                                 goto skip;
5278
5279                         if (!wc->update_ref ||
5280                             generation <= root->root_key.offset)
5281                                 goto skip;
5282
5283                         btrfs_node_key_to_cpu(path->nodes[level], &key,
5284                                               path->slots[level]);
5285                         ret = btrfs_comp_cpu_keys(&key, &wc->update_progress);
5286                         if (ret < 0)
5287                                 goto skip;
5288
5289                         wc->stage = UPDATE_BACKREF;
5290                         wc->shared_level = level - 1;
5291                 }
5292         } else {
5293                 if (level == 1 &&
5294                     (wc->flags[0] & BTRFS_BLOCK_FLAG_FULL_BACKREF))
5295                         goto skip;
5296         }
5297
5298         if (!btrfs_buffer_uptodate(next, generation, 0)) {
5299                 btrfs_tree_unlock(next);
5300                 free_extent_buffer(next);
5301                 next = NULL;
5302                 *lookup_info = 1;
5303         }
5304
5305         if (!next) {
5306                 if (reada && level == 1)
5307                         reada_walk_down(trans, root, wc, path);
5308                 next = read_tree_block(fs_info, bytenr, root->root_key.objectid,
5309                                        generation, level - 1, &first_key);
5310                 if (IS_ERR(next)) {
5311                         return PTR_ERR(next);
5312                 } else if (!extent_buffer_uptodate(next)) {
5313                         free_extent_buffer(next);
5314                         return -EIO;
5315                 }
5316                 btrfs_tree_lock(next);
5317         }
5318
5319         level--;
5320         ASSERT(level == btrfs_header_level(next));
5321         if (level != btrfs_header_level(next)) {
5322                 btrfs_err(root->fs_info, "mismatched level");
5323                 ret = -EIO;
5324                 goto out_unlock;
5325         }
5326         path->nodes[level] = next;
5327         path->slots[level] = 0;
5328         path->locks[level] = BTRFS_WRITE_LOCK;
5329         wc->level = level;
5330         if (wc->level == 1)
5331                 wc->reada_slot = 0;
5332         return 0;
5333 skip:
5334         wc->refs[level - 1] = 0;
5335         wc->flags[level - 1] = 0;
5336         if (wc->stage == DROP_REFERENCE) {
5337                 if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF) {
5338                         parent = path->nodes[level]->start;
5339                 } else {
5340                         ASSERT(root->root_key.objectid ==
5341                                btrfs_header_owner(path->nodes[level]));
5342                         if (root->root_key.objectid !=
5343                             btrfs_header_owner(path->nodes[level])) {
5344                                 btrfs_err(root->fs_info,
5345                                                 "mismatched block owner");
5346                                 ret = -EIO;
5347                                 goto out_unlock;
5348                         }
5349                         parent = 0;
5350                 }
5351
5352                 /*
5353                  * If we had a drop_progress we need to verify the refs are set
5354                  * as expected.  If we find our ref then we know that from here
5355                  * on out everything should be correct, and we can clear the
5356                  * ->restarted flag.
5357                  */
5358                 if (wc->restarted) {
5359                         ret = check_ref_exists(trans, root, bytenr, parent,
5360                                                level - 1);
5361                         if (ret < 0)
5362                                 goto out_unlock;
5363                         if (ret == 0)
5364                                 goto no_delete;
5365                         ret = 0;
5366                         wc->restarted = 0;
5367                 }
5368
5369                 /*
5370                  * Reloc tree doesn't contribute to qgroup numbers, and we have
5371                  * already accounted them at merge time (replace_path),
5372                  * thus we could skip expensive subtree trace here.
5373                  */
5374                 if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID &&
5375                     need_account) {
5376                         ret = btrfs_qgroup_trace_subtree(trans, next,
5377                                                          generation, level - 1);
5378                         if (ret) {
5379                                 btrfs_err_rl(fs_info,
5380                                              "Error %d accounting shared subtree. Quota is out of sync, rescan required.",
5381                                              ret);
5382                         }
5383                 }
5384
5385                 /*
5386                  * We need to update the next key in our walk control so we can
5387                  * update the drop_progress key accordingly.  We don't care if
5388                  * find_next_key doesn't find a key because that means we're at
5389                  * the end and are going to clean up now.
5390                  */
5391                 wc->drop_level = level;
5392                 find_next_key(path, level, &wc->drop_progress);
5393
5394                 btrfs_init_generic_ref(&ref, BTRFS_DROP_DELAYED_REF, bytenr,
5395                                        fs_info->nodesize, parent);
5396                 btrfs_init_tree_ref(&ref, level - 1, root->root_key.objectid,
5397                                     0, false);
5398                 ret = btrfs_free_extent(trans, &ref);
5399                 if (ret)
5400                         goto out_unlock;
5401         }
5402 no_delete:
5403         *lookup_info = 1;
5404         ret = 1;
5405
5406 out_unlock:
5407         btrfs_tree_unlock(next);
5408         free_extent_buffer(next);
5409
5410         return ret;
5411 }
5412
5413 /*
5414  * helper to process tree block while walking up the tree.
5415  *
5416  * when wc->stage == DROP_REFERENCE, this function drops
5417  * reference count on the block.
5418  *
5419  * when wc->stage == UPDATE_BACKREF, this function changes
5420  * wc->stage back to DROP_REFERENCE if we changed wc->stage
5421  * to UPDATE_BACKREF previously while processing the block.
5422  *
5423  * NOTE: return value 1 means we should stop walking up.
5424  */
5425 static noinline int walk_up_proc(struct btrfs_trans_handle *trans,
5426                                  struct btrfs_root *root,
5427                                  struct btrfs_path *path,
5428                                  struct walk_control *wc)
5429 {
5430         struct btrfs_fs_info *fs_info = root->fs_info;
5431         int ret;
5432         int level = wc->level;
5433         struct extent_buffer *eb = path->nodes[level];
5434         u64 parent = 0;
5435
5436         if (wc->stage == UPDATE_BACKREF) {
5437                 BUG_ON(wc->shared_level < level);
5438                 if (level < wc->shared_level)
5439                         goto out;
5440
5441                 ret = find_next_key(path, level + 1, &wc->update_progress);
5442                 if (ret > 0)
5443                         wc->update_ref = 0;
5444
5445                 wc->stage = DROP_REFERENCE;
5446                 wc->shared_level = -1;
5447                 path->slots[level] = 0;
5448
5449                 /*
5450                  * check reference count again if the block isn't locked.
5451                  * we should start walking down the tree again if reference
5452                  * count is one.
5453                  */
5454                 if (!path->locks[level]) {
5455                         BUG_ON(level == 0);
5456                         btrfs_tree_lock(eb);
5457                         path->locks[level] = BTRFS_WRITE_LOCK;
5458
5459                         ret = btrfs_lookup_extent_info(trans, fs_info,
5460                                                        eb->start, level, 1,
5461                                                        &wc->refs[level],
5462                                                        &wc->flags[level]);
5463                         if (ret < 0) {
5464                                 btrfs_tree_unlock_rw(eb, path->locks[level]);
5465                                 path->locks[level] = 0;
5466                                 return ret;
5467                         }
5468                         BUG_ON(wc->refs[level] == 0);
5469                         if (wc->refs[level] == 1) {
5470                                 btrfs_tree_unlock_rw(eb, path->locks[level]);
5471                                 path->locks[level] = 0;
5472                                 return 1;
5473                         }
5474                 }
5475         }
5476
5477         /* wc->stage == DROP_REFERENCE */
5478         BUG_ON(wc->refs[level] > 1 && !path->locks[level]);
5479
5480         if (wc->refs[level] == 1) {
5481                 if (level == 0) {
5482                         if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
5483                                 ret = btrfs_dec_ref(trans, root, eb, 1);
5484                         else
5485                                 ret = btrfs_dec_ref(trans, root, eb, 0);
5486                         BUG_ON(ret); /* -ENOMEM */
5487                         if (is_fstree(root->root_key.objectid)) {
5488                                 ret = btrfs_qgroup_trace_leaf_items(trans, eb);
5489                                 if (ret) {
5490                                         btrfs_err_rl(fs_info,
5491         "error %d accounting leaf items, quota is out of sync, rescan required",
5492                                              ret);
5493                                 }
5494                         }
5495                 }
5496                 /* make block locked assertion in btrfs_clean_tree_block happy */
5497                 if (!path->locks[level] &&
5498                     btrfs_header_generation(eb) == trans->transid) {
5499                         btrfs_tree_lock(eb);
5500                         path->locks[level] = BTRFS_WRITE_LOCK;
5501                 }
5502                 btrfs_clean_tree_block(eb);
5503         }
5504
5505         if (eb == root->node) {
5506                 if (wc->flags[level] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
5507                         parent = eb->start;
5508                 else if (root->root_key.objectid != btrfs_header_owner(eb))
5509                         goto owner_mismatch;
5510         } else {
5511                 if (wc->flags[level + 1] & BTRFS_BLOCK_FLAG_FULL_BACKREF)
5512                         parent = path->nodes[level + 1]->start;
5513                 else if (root->root_key.objectid !=
5514                          btrfs_header_owner(path->nodes[level + 1]))
5515                         goto owner_mismatch;
5516         }
5517
5518         btrfs_free_tree_block(trans, btrfs_root_id(root), eb, parent,
5519                               wc->refs[level] == 1);
5520 out:
5521         wc->refs[level] = 0;
5522         wc->flags[level] = 0;
5523         return 0;
5524
5525 owner_mismatch:
5526         btrfs_err_rl(fs_info, "unexpected tree owner, have %llu expect %llu",
5527                      btrfs_header_owner(eb), root->root_key.objectid);
5528         return -EUCLEAN;
5529 }
5530
5531 static noinline int walk_down_tree(struct btrfs_trans_handle *trans,
5532                                    struct btrfs_root *root,
5533                                    struct btrfs_path *path,
5534                                    struct walk_control *wc)
5535 {
5536         int level = wc->level;
5537         int lookup_info = 1;
5538         int ret;
5539
5540         while (level >= 0) {
5541                 ret = walk_down_proc(trans, root, path, wc, lookup_info);
5542                 if (ret > 0)
5543                         break;
5544
5545                 if (level == 0)
5546                         break;
5547
5548                 if (path->slots[level] >=
5549                     btrfs_header_nritems(path->nodes[level]))
5550                         break;
5551
5552                 ret = do_walk_down(trans, root, path, wc, &lookup_info);
5553                 if (ret > 0) {
5554                         path->slots[level]++;
5555                         continue;
5556                 } else if (ret < 0)
5557                         return ret;
5558                 level = wc->level;
5559         }
5560         return 0;
5561 }
5562
5563 static noinline int walk_up_tree(struct btrfs_trans_handle *trans,
5564                                  struct btrfs_root *root,
5565                                  struct btrfs_path *path,
5566                                  struct walk_control *wc, int max_level)
5567 {
5568         int level = wc->level;
5569         int ret;
5570
5571         path->slots[level] = btrfs_header_nritems(path->nodes[level]);
5572         while (level < max_level && path->nodes[level]) {
5573                 wc->level = level;
5574                 if (path->slots[level] + 1 <
5575                     btrfs_header_nritems(path->nodes[level])) {
5576                         path->slots[level]++;
5577                         return 0;
5578                 } else {
5579                         ret = walk_up_proc(trans, root, path, wc);
5580                         if (ret > 0)
5581                                 return 0;
5582                         if (ret < 0)
5583                                 return ret;
5584
5585                         if (path->locks[level]) {
5586                                 btrfs_tree_unlock_rw(path->nodes[level],
5587                                                      path->locks[level]);
5588                                 path->locks[level] = 0;
5589                         }
5590                         free_extent_buffer(path->nodes[level]);
5591                         path->nodes[level] = NULL;
5592                         level++;
5593                 }
5594         }
5595         return 1;
5596 }
5597
5598 /*
5599  * drop a subvolume tree.
5600  *
5601  * this function traverses the tree freeing any blocks that only
5602  * referenced by the tree.
5603  *
5604  * when a shared tree block is found. this function decreases its
5605  * reference count by one. if update_ref is true, this function
5606  * also make sure backrefs for the shared block and all lower level
5607  * blocks are properly updated.
5608  *
5609  * If called with for_reloc == 0, may exit early with -EAGAIN
5610  */
5611 int btrfs_drop_snapshot(struct btrfs_root *root, int update_ref, int for_reloc)
5612 {
5613         struct btrfs_fs_info *fs_info = root->fs_info;
5614         struct btrfs_path *path;
5615         struct btrfs_trans_handle *trans;
5616         struct btrfs_root *tree_root = fs_info->tree_root;
5617         struct btrfs_root_item *root_item = &root->root_item;
5618         struct walk_control *wc;
5619         struct btrfs_key key;
5620         int err = 0;
5621         int ret;
5622         int level;
5623         bool root_dropped = false;
5624         bool unfinished_drop = false;
5625
5626         btrfs_debug(fs_info, "Drop subvolume %llu", root->root_key.objectid);
5627
5628         path = btrfs_alloc_path();
5629         if (!path) {
5630                 err = -ENOMEM;
5631                 goto out;
5632         }
5633
5634         wc = kzalloc(sizeof(*wc), GFP_NOFS);
5635         if (!wc) {
5636                 btrfs_free_path(path);
5637                 err = -ENOMEM;
5638                 goto out;
5639         }
5640
5641         /*
5642          * Use join to avoid potential EINTR from transaction start. See
5643          * wait_reserve_ticket and the whole reservation callchain.
5644          */
5645         if (for_reloc)
5646                 trans = btrfs_join_transaction(tree_root);
5647         else
5648                 trans = btrfs_start_transaction(tree_root, 0);
5649         if (IS_ERR(trans)) {
5650                 err = PTR_ERR(trans);
5651                 goto out_free;
5652         }
5653
5654         err = btrfs_run_delayed_items(trans);
5655         if (err)
5656                 goto out_end_trans;
5657
5658         /*
5659          * This will help us catch people modifying the fs tree while we're
5660          * dropping it.  It is unsafe to mess with the fs tree while it's being
5661          * dropped as we unlock the root node and parent nodes as we walk down
5662          * the tree, assuming nothing will change.  If something does change
5663          * then we'll have stale information and drop references to blocks we've
5664          * already dropped.
5665          */
5666         set_bit(BTRFS_ROOT_DELETING, &root->state);
5667         unfinished_drop = test_bit(BTRFS_ROOT_UNFINISHED_DROP, &root->state);
5668
5669         if (btrfs_disk_key_objectid(&root_item->drop_progress) == 0) {
5670                 level = btrfs_header_level(root->node);
5671                 path->nodes[level] = btrfs_lock_root_node(root);
5672                 path->slots[level] = 0;
5673                 path->locks[level] = BTRFS_WRITE_LOCK;
5674                 memset(&wc->update_progress, 0,
5675                        sizeof(wc->update_progress));
5676         } else {
5677                 btrfs_disk_key_to_cpu(&key, &root_item->drop_progress);
5678                 memcpy(&wc->update_progress, &key,
5679                        sizeof(wc->update_progress));
5680
5681                 level = btrfs_root_drop_level(root_item);
5682                 BUG_ON(level == 0);
5683                 path->lowest_level = level;
5684                 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
5685                 path->lowest_level = 0;
5686                 if (ret < 0) {
5687                         err = ret;
5688                         goto out_end_trans;
5689                 }
5690                 WARN_ON(ret > 0);
5691
5692                 /*
5693                  * unlock our path, this is safe because only this
5694                  * function is allowed to delete this snapshot
5695                  */
5696                 btrfs_unlock_up_safe(path, 0);
5697
5698                 level = btrfs_header_level(root->node);
5699                 while (1) {
5700                         btrfs_tree_lock(path->nodes[level]);
5701                         path->locks[level] = BTRFS_WRITE_LOCK;
5702
5703                         ret = btrfs_lookup_extent_info(trans, fs_info,
5704                                                 path->nodes[level]->start,
5705                                                 level, 1, &wc->refs[level],
5706                                                 &wc->flags[level]);
5707                         if (ret < 0) {
5708                                 err = ret;
5709                                 goto out_end_trans;
5710                         }
5711                         BUG_ON(wc->refs[level] == 0);
5712
5713                         if (level == btrfs_root_drop_level(root_item))
5714                                 break;
5715
5716                         btrfs_tree_unlock(path->nodes[level]);
5717                         path->locks[level] = 0;
5718                         WARN_ON(wc->refs[level] != 1);
5719                         level--;
5720                 }
5721         }
5722
5723         wc->restarted = test_bit(BTRFS_ROOT_DEAD_TREE, &root->state);
5724         wc->level = level;
5725         wc->shared_level = -1;
5726         wc->stage = DROP_REFERENCE;
5727         wc->update_ref = update_ref;
5728         wc->keep_locks = 0;
5729         wc->reada_count = BTRFS_NODEPTRS_PER_BLOCK(fs_info);
5730
5731         while (1) {
5732
5733                 ret = walk_down_tree(trans, root, path, wc);
5734                 if (ret < 0) {
5735                         err = ret;
5736                         break;
5737                 }
5738
5739                 ret = walk_up_tree(trans, root, path, wc, BTRFS_MAX_LEVEL);
5740                 if (ret < 0) {
5741                         err = ret;
5742                         break;
5743                 }
5744
5745                 if (ret > 0) {
5746                         BUG_ON(wc->stage != DROP_REFERENCE);
5747                         break;
5748                 }
5749
5750                 if (wc->stage == DROP_REFERENCE) {
5751                         wc->drop_level = wc->level;
5752                         btrfs_node_key_to_cpu(path->nodes[wc->drop_level],
5753                                               &wc->drop_progress,
5754                                               path->slots[wc->drop_level]);
5755                 }
5756                 btrfs_cpu_key_to_disk(&root_item->drop_progress,
5757                                       &wc->drop_progress);
5758                 btrfs_set_root_drop_level(root_item, wc->drop_level);
5759
5760                 BUG_ON(wc->level == 0);
5761                 if (btrfs_should_end_transaction(trans) ||
5762                     (!for_reloc && btrfs_need_cleaner_sleep(fs_info))) {
5763                         ret = btrfs_update_root(trans, tree_root,
5764                                                 &root->root_key,
5765                                                 root_item);
5766                         if (ret) {
5767                                 btrfs_abort_transaction(trans, ret);
5768                                 err = ret;
5769                                 goto out_end_trans;
5770                         }
5771
5772                         btrfs_end_transaction_throttle(trans);
5773                         if (!for_reloc && btrfs_need_cleaner_sleep(fs_info)) {
5774                                 btrfs_debug(fs_info,
5775                                             "drop snapshot early exit");
5776                                 err = -EAGAIN;
5777                                 goto out_free;
5778                         }
5779
5780                        /*
5781                         * Use join to avoid potential EINTR from transaction
5782                         * start. See wait_reserve_ticket and the whole
5783                         * reservation callchain.
5784                         */
5785                         if (for_reloc)
5786                                 trans = btrfs_join_transaction(tree_root);
5787                         else
5788                                 trans = btrfs_start_transaction(tree_root, 0);
5789                         if (IS_ERR(trans)) {
5790                                 err = PTR_ERR(trans);
5791                                 goto out_free;
5792                         }
5793                 }
5794         }
5795         btrfs_release_path(path);
5796         if (err)
5797                 goto out_end_trans;
5798
5799         ret = btrfs_del_root(trans, &root->root_key);
5800         if (ret) {
5801                 btrfs_abort_transaction(trans, ret);
5802                 err = ret;
5803                 goto out_end_trans;
5804         }
5805
5806         if (root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID) {
5807                 ret = btrfs_find_root(tree_root, &root->root_key, path,
5808                                       NULL, NULL);
5809                 if (ret < 0) {
5810                         btrfs_abort_transaction(trans, ret);
5811                         err = ret;
5812                         goto out_end_trans;
5813                 } else if (ret > 0) {
5814                         /* if we fail to delete the orphan item this time
5815                          * around, it'll get picked up the next time.
5816                          *
5817                          * The most common failure here is just -ENOENT.
5818                          */
5819                         btrfs_del_orphan_item(trans, tree_root,
5820                                               root->root_key.objectid);
5821                 }
5822         }
5823
5824         /*
5825          * This subvolume is going to be completely dropped, and won't be
5826          * recorded as dirty roots, thus pertrans meta rsv will not be freed at
5827          * commit transaction time.  So free it here manually.
5828          */
5829         btrfs_qgroup_convert_reserved_meta(root, INT_MAX);
5830         btrfs_qgroup_free_meta_all_pertrans(root);
5831
5832         if (test_bit(BTRFS_ROOT_REGISTERED, &root->state))
5833                 btrfs_add_dropped_root(trans, root);
5834         else
5835                 btrfs_put_root(root);
5836         root_dropped = true;
5837 out_end_trans:
5838         btrfs_end_transaction_throttle(trans);
5839 out_free:
5840         kfree(wc);
5841         btrfs_free_path(path);
5842 out:
5843         /*
5844          * We were an unfinished drop root, check to see if there are any
5845          * pending, and if not clear and wake up any waiters.
5846          */
5847         if (!err && unfinished_drop)
5848                 btrfs_maybe_wake_unfinished_drop(fs_info);
5849
5850         /*
5851          * So if we need to stop dropping the snapshot for whatever reason we
5852          * need to make sure to add it back to the dead root list so that we
5853          * keep trying to do the work later.  This also cleans up roots if we
5854          * don't have it in the radix (like when we recover after a power fail
5855          * or unmount) so we don't leak memory.
5856          */
5857         if (!for_reloc && !root_dropped)
5858                 btrfs_add_dead_root(root);
5859         return err;
5860 }
5861
5862 /*
5863  * drop subtree rooted at tree block 'node'.
5864  *
5865  * NOTE: this function will unlock and release tree block 'node'
5866  * only used by relocation code
5867  */
5868 int btrfs_drop_subtree(struct btrfs_trans_handle *trans,
5869                         struct btrfs_root *root,
5870                         struct extent_buffer *node,
5871                         struct extent_buffer *parent)
5872 {
5873         struct btrfs_fs_info *fs_info = root->fs_info;
5874         struct btrfs_path *path;
5875         struct walk_control *wc;
5876         int level;
5877         int parent_level;
5878         int ret = 0;
5879         int wret;
5880
5881         BUG_ON(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID);
5882
5883         path = btrfs_alloc_path();
5884         if (!path)
5885                 return -ENOMEM;
5886
5887         wc = kzalloc(sizeof(*wc), GFP_NOFS);
5888         if (!wc) {
5889                 btrfs_free_path(path);
5890                 return -ENOMEM;
5891         }
5892
5893         btrfs_assert_tree_write_locked(parent);
5894         parent_level = btrfs_header_level(parent);
5895         atomic_inc(&parent->refs);
5896         path->nodes[parent_level] = parent;
5897         path->slots[parent_level] = btrfs_header_nritems(parent);
5898
5899         btrfs_assert_tree_write_locked(node);
5900         level = btrfs_header_level(node);
5901         path->nodes[level] = node;
5902         path->slots[level] = 0;
5903         path->locks[level] = BTRFS_WRITE_LOCK;
5904
5905         wc->refs[parent_level] = 1;
5906         wc->flags[parent_level] = BTRFS_BLOCK_FLAG_FULL_BACKREF;
5907         wc->level = level;
5908         wc->shared_level = -1;
5909         wc->stage = DROP_REFERENCE;
5910         wc->update_ref = 0;
5911         wc->keep_locks = 1;
5912         wc->reada_count = BTRFS_NODEPTRS_PER_BLOCK(fs_info);
5913
5914         while (1) {
5915                 wret = walk_down_tree(trans, root, path, wc);
5916                 if (wret < 0) {
5917                         ret = wret;
5918                         break;
5919                 }
5920
5921                 wret = walk_up_tree(trans, root, path, wc, parent_level);
5922                 if (wret < 0)
5923                         ret = wret;
5924                 if (wret != 0)
5925                         break;
5926         }
5927
5928         kfree(wc);
5929         btrfs_free_path(path);
5930         return ret;
5931 }
5932
5933 /*
5934  * helper to account the unused space of all the readonly block group in the
5935  * space_info. takes mirrors into account.
5936  */
5937 u64 btrfs_account_ro_block_groups_free_space(struct btrfs_space_info *sinfo)
5938 {
5939         struct btrfs_block_group *block_group;
5940         u64 free_bytes = 0;
5941         int factor;
5942
5943         /* It's df, we don't care if it's racy */
5944         if (list_empty(&sinfo->ro_bgs))
5945                 return 0;
5946
5947         spin_lock(&sinfo->lock);
5948         list_for_each_entry(block_group, &sinfo->ro_bgs, ro_list) {
5949                 spin_lock(&block_group->lock);
5950
5951                 if (!block_group->ro) {
5952                         spin_unlock(&block_group->lock);
5953                         continue;
5954                 }
5955
5956                 factor = btrfs_bg_type_to_factor(block_group->flags);
5957                 free_bytes += (block_group->length -
5958                                block_group->used) * factor;
5959
5960                 spin_unlock(&block_group->lock);
5961         }
5962         spin_unlock(&sinfo->lock);
5963
5964         return free_bytes;
5965 }
5966
5967 int btrfs_error_unpin_extent_range(struct btrfs_fs_info *fs_info,
5968                                    u64 start, u64 end)
5969 {
5970         return unpin_extent_range(fs_info, start, end, false);
5971 }
5972
5973 /*
5974  * It used to be that old block groups would be left around forever.
5975  * Iterating over them would be enough to trim unused space.  Since we
5976  * now automatically remove them, we also need to iterate over unallocated
5977  * space.
5978  *
5979  * We don't want a transaction for this since the discard may take a
5980  * substantial amount of time.  We don't require that a transaction be
5981  * running, but we do need to take a running transaction into account
5982  * to ensure that we're not discarding chunks that were released or
5983  * allocated in the current transaction.
5984  *
5985  * Holding the chunks lock will prevent other threads from allocating
5986  * or releasing chunks, but it won't prevent a running transaction
5987  * from committing and releasing the memory that the pending chunks
5988  * list head uses.  For that, we need to take a reference to the
5989  * transaction and hold the commit root sem.  We only need to hold
5990  * it while performing the free space search since we have already
5991  * held back allocations.
5992  */
5993 static int btrfs_trim_free_extents(struct btrfs_device *device, u64 *trimmed)
5994 {
5995         u64 start = SZ_1M, len = 0, end = 0;
5996         int ret;
5997
5998         *trimmed = 0;
5999
6000         /* Discard not supported = nothing to do. */
6001         if (!bdev_max_discard_sectors(device->bdev))
6002                 return 0;
6003
6004         /* Not writable = nothing to do. */
6005         if (!test_bit(BTRFS_DEV_STATE_WRITEABLE, &device->dev_state))
6006                 return 0;
6007
6008         /* No free space = nothing to do. */
6009         if (device->total_bytes <= device->bytes_used)
6010                 return 0;
6011
6012         ret = 0;
6013
6014         while (1) {
6015                 struct btrfs_fs_info *fs_info = device->fs_info;
6016                 u64 bytes;
6017
6018                 ret = mutex_lock_interruptible(&fs_info->chunk_mutex);
6019                 if (ret)
6020                         break;
6021
6022                 find_first_clear_extent_bit(&device->alloc_state, start,
6023                                             &start, &end,
6024                                             CHUNK_TRIMMED | CHUNK_ALLOCATED);
6025
6026                 /* Check if there are any CHUNK_* bits left */
6027                 if (start > device->total_bytes) {
6028                         WARN_ON(IS_ENABLED(CONFIG_BTRFS_DEBUG));
6029                         btrfs_warn_in_rcu(fs_info,
6030 "ignoring attempt to trim beyond device size: offset %llu length %llu device %s device size %llu",
6031                                           start, end - start + 1,
6032                                           rcu_str_deref(device->name),
6033                                           device->total_bytes);
6034                         mutex_unlock(&fs_info->chunk_mutex);
6035                         ret = 0;
6036                         break;
6037                 }
6038
6039                 /* Ensure we skip the reserved area in the first 1M */
6040                 start = max_t(u64, start, SZ_1M);
6041
6042                 /*
6043                  * If find_first_clear_extent_bit find a range that spans the
6044                  * end of the device it will set end to -1, in this case it's up
6045                  * to the caller to trim the value to the size of the device.
6046                  */
6047                 end = min(end, device->total_bytes - 1);
6048
6049                 len = end - start + 1;
6050
6051                 /* We didn't find any extents */
6052                 if (!len) {
6053                         mutex_unlock(&fs_info->chunk_mutex);
6054                         ret = 0;
6055                         break;
6056                 }
6057
6058                 ret = btrfs_issue_discard(device->bdev, start, len,
6059                                           &bytes);
6060                 if (!ret)
6061                         set_extent_bits(&device->alloc_state, start,
6062                                         start + bytes - 1,
6063                                         CHUNK_TRIMMED);
6064                 mutex_unlock(&fs_info->chunk_mutex);
6065
6066                 if (ret)
6067                         break;
6068
6069                 start += len;
6070                 *trimmed += bytes;
6071
6072                 if (fatal_signal_pending(current)) {
6073                         ret = -ERESTARTSYS;
6074                         break;
6075                 }
6076
6077                 cond_resched();
6078         }
6079
6080         return ret;
6081 }
6082
6083 /*
6084  * Trim the whole filesystem by:
6085  * 1) trimming the free space in each block group
6086  * 2) trimming the unallocated space on each device
6087  *
6088  * This will also continue trimming even if a block group or device encounters
6089  * an error.  The return value will be the last error, or 0 if nothing bad
6090  * happens.
6091  */
6092 int btrfs_trim_fs(struct btrfs_fs_info *fs_info, struct fstrim_range *range)
6093 {
6094         struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
6095         struct btrfs_block_group *cache = NULL;
6096         struct btrfs_device *device;
6097         u64 group_trimmed;
6098         u64 range_end = U64_MAX;
6099         u64 start;
6100         u64 end;
6101         u64 trimmed = 0;
6102         u64 bg_failed = 0;
6103         u64 dev_failed = 0;
6104         int bg_ret = 0;
6105         int dev_ret = 0;
6106         int ret = 0;
6107
6108         if (range->start == U64_MAX)
6109                 return -EINVAL;
6110
6111         /*
6112          * Check range overflow if range->len is set.
6113          * The default range->len is U64_MAX.
6114          */
6115         if (range->len != U64_MAX &&
6116             check_add_overflow(range->start, range->len, &range_end))
6117                 return -EINVAL;
6118
6119         cache = btrfs_lookup_first_block_group(fs_info, range->start);
6120         for (; cache; cache = btrfs_next_block_group(cache)) {
6121                 if (cache->start >= range_end) {
6122                         btrfs_put_block_group(cache);
6123                         break;
6124                 }
6125
6126                 start = max(range->start, cache->start);
6127                 end = min(range_end, cache->start + cache->length);
6128
6129                 if (end - start >= range->minlen) {
6130                         if (!btrfs_block_group_done(cache)) {
6131                                 ret = btrfs_cache_block_group(cache, 0);
6132                                 if (ret) {
6133                                         bg_failed++;
6134                                         bg_ret = ret;
6135                                         continue;
6136                                 }
6137                                 ret = btrfs_wait_block_group_cache_done(cache);
6138                                 if (ret) {
6139                                         bg_failed++;
6140                                         bg_ret = ret;
6141                                         continue;
6142                                 }
6143                         }
6144                         ret = btrfs_trim_block_group(cache,
6145                                                      &group_trimmed,
6146                                                      start,
6147                                                      end,
6148                                                      range->minlen);
6149
6150                         trimmed += group_trimmed;
6151                         if (ret) {
6152                                 bg_failed++;
6153                                 bg_ret = ret;
6154                                 continue;
6155                         }
6156                 }
6157         }
6158
6159         if (bg_failed)
6160                 btrfs_warn(fs_info,
6161                         "failed to trim %llu block group(s), last error %d",
6162                         bg_failed, bg_ret);
6163
6164         mutex_lock(&fs_devices->device_list_mutex);
6165         list_for_each_entry(device, &fs_devices->devices, dev_list) {
6166                 if (test_bit(BTRFS_DEV_STATE_MISSING, &device->dev_state))
6167                         continue;
6168
6169                 ret = btrfs_trim_free_extents(device, &group_trimmed);
6170                 if (ret) {
6171                         dev_failed++;
6172                         dev_ret = ret;
6173                         break;
6174                 }
6175
6176                 trimmed += group_trimmed;
6177         }
6178         mutex_unlock(&fs_devices->device_list_mutex);
6179
6180         if (dev_failed)
6181                 btrfs_warn(fs_info,
6182                         "failed to trim %llu device(s), last error %d",
6183                         dev_failed, dev_ret);
6184         range->len = trimmed;
6185         if (bg_ret)
6186                 return bg_ret;
6187         return dev_ret;
6188 }