kernel/fs/btrfs/delayed-inode.c

   1 /*
   2  * Copyright (C) 2011 Fujitsu.  All rights reserved.
   3  * Written by Miao Xie <miaox@cn.fujitsu.com>
   4  *
   5  * This program is free software; you can redistribute it and/or
   6  * modify it under the terms of the GNU General Public
   7  * License v2 as published by the Free Software Foundation.
   8  *
   9  * This program is distributed in the hope that it will be useful,
  10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  12  * General Public License for more details.
  13  *
  14  * You should have received a copy of the GNU General Public
  15  * License along with this program; if not, write to the
  16  * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  17  * Boston, MA 021110-1307, USA.
  18  */
  19
  20 #include <linux/slab.h>
  21 #include "delayed-inode.h"
  22 #include "disk-io.h"
  23 #include "transaction.h"
  24 #include "ctree.h"
  25
  26 #define BTRFS_DELAYED_WRITEBACK         512
  27 #define BTRFS_DELAYED_BACKGROUND        128
  28 #define BTRFS_DELAYED_BATCH             16
  29
  30 static struct kmem_cache *delayed_node_cache;
  31
  32 int __init btrfs_delayed_inode_init(void)
  33 {
  34         delayed_node_cache = kmem_cache_create("btrfs_delayed_node",
  35                                         sizeof(struct btrfs_delayed_node),
  36                                         0,
  37                                         SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD,
  38                                         NULL);
  39         if (!delayed_node_cache)
  40                 return -ENOMEM;
  41         return 0;
  42 }
  43
  44 void btrfs_delayed_inode_exit(void)
  45 {
  46         if (delayed_node_cache)
  47                 kmem_cache_destroy(delayed_node_cache);
  48 }
  49
  50 static inline void btrfs_init_delayed_node(
  51                                 struct btrfs_delayed_node *delayed_node,
  52                                 struct btrfs_root *root, u64 inode_id)
  53 {
  54         delayed_node->root = root;
  55         delayed_node->inode_id = inode_id;
  56         atomic_set(&delayed_node->refs, 0);
  57         delayed_node->count = 0;
  58         delayed_node->flags = 0;
  59         delayed_node->ins_root = RB_ROOT;
  60         delayed_node->del_root = RB_ROOT;
  61         mutex_init(&delayed_node->mutex);
  62         delayed_node->index_cnt = 0;
  63         INIT_LIST_HEAD(&delayed_node->n_list);
  64         INIT_LIST_HEAD(&delayed_node->p_list);
  65         delayed_node->bytes_reserved = 0;
  66         memset(&delayed_node->inode_item, 0, sizeof(delayed_node->inode_item));
  67 }
  68
  69 static inline int btrfs_is_continuous_delayed_item(
  70                                         struct btrfs_delayed_item *item1,
  71                                         struct btrfs_delayed_item *item2)
  72 {
  73         if (item1->key.type == BTRFS_DIR_INDEX_KEY &&
  74             item1->key.objectid == item2->key.objectid &&
  75             item1->key.type == item2->key.type &&
  76             item1->key.offset + 1 == item2->key.offset)
  77                 return 1;
  78         return 0;
  79 }
  80
  81 static inline struct btrfs_delayed_root *btrfs_get_delayed_root(
  82                                                         struct btrfs_root *root)
  83 {
  84         return root->fs_info->delayed_root;
  85 }
  86
  87 static struct btrfs_delayed_node *btrfs_get_delayed_node(struct inode *inode)
  88 {
  89         struct btrfs_inode *btrfs_inode = BTRFS_I(inode);
  90         struct btrfs_root *root = btrfs_inode->root;
  91         u64 ino = btrfs_ino(inode);
  92         struct btrfs_delayed_node *node;
  93
  94         node = ACCESS_ONCE(btrfs_inode->delayed_node);
  95         if (node) {
  96                 atomic_inc(&node->refs);
  97                 return node;
  98         }
  99
 100         spin_lock(&root->inode_lock);
 101         node = radix_tree_lookup(&root->delayed_nodes_tree, ino);
 102         if (node) {
 103                 if (btrfs_inode->delayed_node) {
 104                         atomic_inc(&node->refs);        /* can be accessed */
 105                         BUG_ON(btrfs_inode->delayed_node != node);
 106                         spin_unlock(&root->inode_lock);
 107                         return node;
 108                 }
 109                 btrfs_inode->delayed_node = node;
 110                 /* can be accessed and cached in the inode */
 111                 atomic_add(2, &node->refs);
 112                 spin_unlock(&root->inode_lock);
 113                 return node;
 114         }
 115         spin_unlock(&root->inode_lock);
 116
 117         return NULL;
 118 }
 119
 120 /* Will return either the node or PTR_ERR(-ENOMEM) */
 121 static struct btrfs_delayed_node *btrfs_get_or_create_delayed_node(
 122                                                         struct inode *inode)
 123 {
 124         struct btrfs_delayed_node *node;
 125         struct btrfs_inode *btrfs_inode = BTRFS_I(inode);
 126         struct btrfs_root *root = btrfs_inode->root;
 127         u64 ino = btrfs_ino(inode);
 128         int ret;
 129
 130 again:
 131         node = btrfs_get_delayed_node(inode);
 132         if (node)
 133                 return node;
 134
 135         node = kmem_cache_alloc(delayed_node_cache, GFP_NOFS);
 136         if (!node)
 137                 return ERR_PTR(-ENOMEM);
 138         btrfs_init_delayed_node(node, root, ino);
 139
 140         /* cached in the btrfs inode and can be accessed */
 141         atomic_add(2, &node->refs);
 142
 143         ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
 144         if (ret) {
 145                 kmem_cache_free(delayed_node_cache, node);
 146                 return ERR_PTR(ret);
 147         }
 148
 149         spin_lock(&root->inode_lock);
 150         ret = radix_tree_insert(&root->delayed_nodes_tree, ino, node);
 151         if (ret == -EEXIST) {
 152                 spin_unlock(&root->inode_lock);
 153                 kmem_cache_free(delayed_node_cache, node);
 154                 radix_tree_preload_end();
 155                 goto again;
 156         }
 157         btrfs_inode->delayed_node = node;
 158         spin_unlock(&root->inode_lock);
 159         radix_tree_preload_end();
 160
 161         return node;
 162 }
 163
 164 /*
 165  * Call it when holding delayed_node->mutex
 166  *
 167  * If mod = 1, add this node into the prepared list.
 168  */
 169 static void btrfs_queue_delayed_node(struct btrfs_delayed_root *root,
 170                                      struct btrfs_delayed_node *node,
 171                                      int mod)
 172 {
 173         spin_lock(&root->lock);
 174         if (test_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags)) {
 175                 if (!list_empty(&node->p_list))
 176                         list_move_tail(&node->p_list, &root->prepare_list);
 177                 else if (mod)
 178                         list_add_tail(&node->p_list, &root->prepare_list);
 179         } else {
 180                 list_add_tail(&node->n_list, &root->node_list);
 181                 list_add_tail(&node->p_list, &root->prepare_list);
 182                 atomic_inc(&node->refs);        /* inserted into list */
 183                 root->nodes++;
 184                 set_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags);
 185         }
 186         spin_unlock(&root->lock);
 187 }
 188
 189 /* Call it when holding delayed_node->mutex */
 190 static void btrfs_dequeue_delayed_node(struct btrfs_delayed_root *root,
 191                                        struct btrfs_delayed_node *node)
 192 {
 193         spin_lock(&root->lock);
 194         if (test_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags)) {
 195                 root->nodes--;
 196                 atomic_dec(&node->refs);        /* not in the list */
 197                 list_del_init(&node->n_list);
 198                 if (!list_empty(&node->p_list))
 199                         list_del_init(&node->p_list);
 200                 clear_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags);
 201         }
 202         spin_unlock(&root->lock);
 203 }
 204
 205 static struct btrfs_delayed_node *btrfs_first_delayed_node(
 206                         struct btrfs_delayed_root *delayed_root)
 207 {
 208         struct list_head *p;
 209         struct btrfs_delayed_node *node = NULL;
 210
 211         spin_lock(&delayed_root->lock);
 212         if (list_empty(&delayed_root->node_list))
 213                 goto out;
 214
 215         p = delayed_root->node_list.next;
 216         node = list_entry(p, struct btrfs_delayed_node, n_list);
 217         atomic_inc(&node->refs);
 218 out:
 219         spin_unlock(&delayed_root->lock);
 220
 221         return node;
 222 }
 223
 224 static struct btrfs_delayed_node *btrfs_next_delayed_node(
 225                                                 struct btrfs_delayed_node *node)
 226 {
 227         struct btrfs_delayed_root *delayed_root;
 228         struct list_head *p;
 229         struct btrfs_delayed_node *next = NULL;
 230
 231         delayed_root = node->root->fs_info->delayed_root;
 232         spin_lock(&delayed_root->lock);
 233         if (!test_bit(BTRFS_DELAYED_NODE_IN_LIST, &node->flags)) {
 234                 /* not in the list */
 235                 if (list_empty(&delayed_root->node_list))
 236                         goto out;
 237                 p = delayed_root->node_list.next;
 238         } else if (list_is_last(&node->n_list, &delayed_root->node_list))
 239                 goto out;
 240         else
 241                 p = node->n_list.next;
 242
 243         next = list_entry(p, struct btrfs_delayed_node, n_list);
 244         atomic_inc(&next->refs);
 245 out:
 246         spin_unlock(&delayed_root->lock);
 247
 248         return next;
 249 }
 250
 251 static void __btrfs_release_delayed_node(
 252                                 struct btrfs_delayed_node *delayed_node,
 253                                 int mod)
 254 {
 255         struct btrfs_delayed_root *delayed_root;
 256
 257         if (!delayed_node)
 258                 return;
 259
 260         delayed_root = delayed_node->root->fs_info->delayed_root;
 261
 262         mutex_lock(&delayed_node->mutex);
 263         if (delayed_node->count)
 264                 btrfs_queue_delayed_node(delayed_root, delayed_node, mod);
 265         else
 266                 btrfs_dequeue_delayed_node(delayed_root, delayed_node);
 267         mutex_unlock(&delayed_node->mutex);
 268
 269         if (atomic_dec_and_test(&delayed_node->refs)) {
 270                 bool free = false;
 271                 struct btrfs_root *root = delayed_node->root;
 272                 spin_lock(&root->inode_lock);
 273                 if (atomic_read(&delayed_node->refs) == 0) {
 274                         radix_tree_delete(&root->delayed_nodes_tree,
 275                                           delayed_node->inode_id);
 276                         free = true;
 277                 }
 278                 spin_unlock(&root->inode_lock);
 279                 if (free)
 280                         kmem_cache_free(delayed_node_cache, delayed_node);
 281         }
 282 }
 283
 284 static inline void btrfs_release_delayed_node(struct btrfs_delayed_node *node)
 285 {
 286         __btrfs_release_delayed_node(node, 0);
 287 }
 288
 289 static struct btrfs_delayed_node *btrfs_first_prepared_delayed_node(
 290                                         struct btrfs_delayed_root *delayed_root)
 291 {
 292         struct list_head *p;
 293         struct btrfs_delayed_node *node = NULL;
 294
 295         spin_lock(&delayed_root->lock);
 296         if (list_empty(&delayed_root->prepare_list))
 297                 goto out;
 298
 299         p = delayed_root->prepare_list.next;
 300         list_del_init(p);
 301         node = list_entry(p, struct btrfs_delayed_node, p_list);
 302         atomic_inc(&node->refs);
 303 out:
 304         spin_unlock(&delayed_root->lock);
 305
 306         return node;
 307 }
 308
 309 static inline void btrfs_release_prepared_delayed_node(
 310                                         struct btrfs_delayed_node *node)
 311 {
 312         __btrfs_release_delayed_node(node, 1);
 313 }
 314
 315 static struct btrfs_delayed_item *btrfs_alloc_delayed_item(u32 data_len)
 316 {
 317         struct btrfs_delayed_item *item;
 318         item = kmalloc(sizeof(*item) + data_len, GFP_NOFS);
 319         if (item) {
 320                 item->data_len = data_len;
 321                 item->ins_or_del = 0;
 322                 item->bytes_reserved = 0;
 323                 item->delayed_node = NULL;
 324                 atomic_set(&item->refs, 1);
 325         }
 326         return item;
 327 }
 328
 329 /*
 330  * __btrfs_lookup_delayed_item - look up the delayed item by key
 331  * @delayed_node: pointer to the delayed node
 332  * @key:          the key to look up
 333  * @prev:         used to store the prev item if the right item isn't found
 334  * @next:         used to store the next item if the right item isn't found
 335  *
 336  * Note: if we don't find the right item, we will return the prev item and
 337  * the next item.
 338  */
 339 static struct btrfs_delayed_item *__btrfs_lookup_delayed_item(
 340                                 struct rb_root *root,
 341                                 struct btrfs_key *key,
 342                                 struct btrfs_delayed_item **prev,
 343                                 struct btrfs_delayed_item **next)
 344 {
 345         struct rb_node *node, *prev_node = NULL;
 346         struct btrfs_delayed_item *delayed_item = NULL;
 347         int ret = 0;
 348
 349         node = root->rb_node;
 350
 351         while (node) {
 352                 delayed_item = rb_entry(node, struct btrfs_delayed_item,
 353                                         rb_node);
 354                 prev_node = node;
 355                 ret = btrfs_comp_cpu_keys(&delayed_item->key, key);
 356                 if (ret < 0)
 357                         node = node->rb_right;
 358                 else if (ret > 0)
 359                         node = node->rb_left;
 360                 else
 361                         return delayed_item;
 362         }
 363
 364         if (prev) {
 365                 if (!prev_node)
 366                         *prev = NULL;
 367                 else if (ret < 0)
 368                         *prev = delayed_item;
 369                 else if ((node = rb_prev(prev_node)) != NULL) {
 370                         *prev = rb_entry(node, struct btrfs_delayed_item,
 371                                          rb_node);
 372                 } else
 373                         *prev = NULL;
 374         }
 375
 376         if (next) {
 377                 if (!prev_node)
 378                         *next = NULL;
 379                 else if (ret > 0)
 380                         *next = delayed_item;
 381                 else if ((node = rb_next(prev_node)) != NULL) {
 382                         *next = rb_entry(node, struct btrfs_delayed_item,
 383                                          rb_node);
 384                 } else
 385                         *next = NULL;
 386         }
 387         return NULL;
 388 }
 389
 390 static struct btrfs_delayed_item *__btrfs_lookup_delayed_insertion_item(
 391                                         struct btrfs_delayed_node *delayed_node,
 392                                         struct btrfs_key *key)
 393 {
 394         struct btrfs_delayed_item *item;
 395
 396         item = __btrfs_lookup_delayed_item(&delayed_node->ins_root, key,
 397                                            NULL, NULL);
 398         return item;
 399 }
 400
 401 static int __btrfs_add_delayed_item(struct btrfs_delayed_node *delayed_node,
 402                                     struct btrfs_delayed_item *ins,
 403                                     int action)
 404 {
 405         struct rb_node **p, *node;
 406         struct rb_node *parent_node = NULL;
 407         struct rb_root *root;
 408         struct btrfs_delayed_item *item;
 409         int cmp;
 410
 411         if (action == BTRFS_DELAYED_INSERTION_ITEM)
 412                 root = &delayed_node->ins_root;
 413         else if (action == BTRFS_DELAYED_DELETION_ITEM)
 414                 root = &delayed_node->del_root;
 415         else
 416                 BUG();
 417         p = &root->rb_node;
 418         node = &ins->rb_node;
 419
 420         while (*p) {
 421                 parent_node = *p;
 422                 item = rb_entry(parent_node, struct btrfs_delayed_item,
 423                                  rb_node);
 424
 425                 cmp = btrfs_comp_cpu_keys(&item->key, &ins->key);
 426                 if (cmp < 0)
 427                         p = &(*p)->rb_right;
 428                 else if (cmp > 0)
 429                         p = &(*p)->rb_left;
 430                 else
 431                         return -EEXIST;
 432         }
 433
 434         rb_link_node(node, parent_node, p);
 435         rb_insert_color(node, root);
 436         ins->delayed_node = delayed_node;
 437         ins->ins_or_del = action;
 438
 439         if (ins->key.type == BTRFS_DIR_INDEX_KEY &&
 440             action == BTRFS_DELAYED_INSERTION_ITEM &&
 441             ins->key.offset >= delayed_node->index_cnt)
 442                         delayed_node->index_cnt = ins->key.offset + 1;
 443
 444         delayed_node->count++;
 445         atomic_inc(&delayed_node->root->fs_info->delayed_root->items);
 446         return 0;
 447 }
 448
 449 static int __btrfs_add_delayed_insertion_item(struct btrfs_delayed_node *node,
 450                                               struct btrfs_delayed_item *item)
 451 {
 452         return __btrfs_add_delayed_item(node, item,
 453                                         BTRFS_DELAYED_INSERTION_ITEM);
 454 }
 455
 456 static int __btrfs_add_delayed_deletion_item(struct btrfs_delayed_node *node,
 457                                              struct btrfs_delayed_item *item)
 458 {
 459         return __btrfs_add_delayed_item(node, item,
 460                                         BTRFS_DELAYED_DELETION_ITEM);
 461 }
 462
 463 static void finish_one_item(struct btrfs_delayed_root *delayed_root)
 464 {
 465         int seq = atomic_inc_return(&delayed_root->items_seq);
 466
 467         /*
 468          * atomic_dec_return implies a barrier for waitqueue_active
 469          */
 470         if ((atomic_dec_return(&delayed_root->items) <
 471             BTRFS_DELAYED_BACKGROUND || seq % BTRFS_DELAYED_BATCH == 0) &&
 472             waitqueue_active(&delayed_root->wait))
 473                 wake_up(&delayed_root->wait);
 474 }
 475
 476 static void __btrfs_remove_delayed_item(struct btrfs_delayed_item *delayed_item)
 477 {
 478         struct rb_root *root;
 479         struct btrfs_delayed_root *delayed_root;
 480
 481         delayed_root = delayed_item->delayed_node->root->fs_info->delayed_root;
 482
 483         BUG_ON(!delayed_root);
 484         BUG_ON(delayed_item->ins_or_del != BTRFS_DELAYED_DELETION_ITEM &&
 485                delayed_item->ins_or_del != BTRFS_DELAYED_INSERTION_ITEM);
 486
 487         if (delayed_item->ins_or_del == BTRFS_DELAYED_INSERTION_ITEM)
 488                 root = &delayed_item->delayed_node->ins_root;
 489         else
 490                 root = &delayed_item->delayed_node->del_root;
 491
 492         rb_erase(&delayed_item->rb_node, root);
 493         delayed_item->delayed_node->count--;
 494
 495         finish_one_item(delayed_root);
 496 }
 497
 498 static void btrfs_release_delayed_item(struct btrfs_delayed_item *item)
 499 {
 500         if (item) {
 501                 __btrfs_remove_delayed_item(item);
 502                 if (atomic_dec_and_test(&item->refs))
 503                         kfree(item);
 504         }
 505 }
 506
 507 static struct btrfs_delayed_item *__btrfs_first_delayed_insertion_item(
 508                                         struct btrfs_delayed_node *delayed_node)
 509 {
 510         struct rb_node *p;
 511         struct btrfs_delayed_item *item = NULL;
 512
 513         p = rb_first(&delayed_node->ins_root);
 514         if (p)
 515                 item = rb_entry(p, struct btrfs_delayed_item, rb_node);
 516
 517         return item;
 518 }
 519
 520 static struct btrfs_delayed_item *__btrfs_first_delayed_deletion_item(
 521                                         struct btrfs_delayed_node *delayed_node)
 522 {
 523         struct rb_node *p;
 524         struct btrfs_delayed_item *item = NULL;
 525
 526         p = rb_first(&delayed_node->del_root);
 527         if (p)
 528                 item = rb_entry(p, struct btrfs_delayed_item, rb_node);
 529
 530         return item;
 531 }
 532
 533 static struct btrfs_delayed_item *__btrfs_next_delayed_item(
 534                                                 struct btrfs_delayed_item *item)
 535 {
 536         struct rb_node *p;
 537         struct btrfs_delayed_item *next = NULL;
 538
 539         p = rb_next(&item->rb_node);
 540         if (p)
 541                 next = rb_entry(p, struct btrfs_delayed_item, rb_node);
 542
 543         return next;
 544 }
 545
 546 static int btrfs_delayed_item_reserve_metadata(struct btrfs_trans_handle *trans,
 547                                                struct btrfs_root *root,
 548                                                struct btrfs_delayed_item *item)
 549 {
 550         struct btrfs_block_rsv *src_rsv;
 551         struct btrfs_block_rsv *dst_rsv;
 552         u64 num_bytes;
 553         int ret;
 554
 555         if (!trans->bytes_reserved)
 556                 return 0;
 557
 558         src_rsv = trans->block_rsv;
 559         dst_rsv = &root->fs_info->delayed_block_rsv;
 560
 561         num_bytes = btrfs_calc_trans_metadata_size(root, 1);
 562         ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes);
 563         if (!ret) {
 564                 trace_btrfs_space_reservation(root->fs_info, "delayed_item",
 565                                               item->key.objectid,
 566                                               num_bytes, 1);
 567                 item->bytes_reserved = num_bytes;
 568         }
 569
 570         return ret;
 571 }
 572
 573 static void btrfs_delayed_item_release_metadata(struct btrfs_root *root,
 574                                                 struct btrfs_delayed_item *item)
 575 {
 576         struct btrfs_block_rsv *rsv;
 577
 578         if (!item->bytes_reserved)
 579                 return;
 580
 581         rsv = &root->fs_info->delayed_block_rsv;
 582         trace_btrfs_space_reservation(root->fs_info, "delayed_item",
 583                                       item->key.objectid, item->bytes_reserved,
 584                                       0);
 585         btrfs_block_rsv_release(root, rsv,
 586                                 item->bytes_reserved);
 587 }
 588
 589 static int btrfs_delayed_inode_reserve_metadata(
 590                                         struct btrfs_trans_handle *trans,
 591                                         struct btrfs_root *root,
 592                                         struct inode *inode,
 593                                         struct btrfs_delayed_node *node)
 594 {
 595         struct btrfs_block_rsv *src_rsv;
 596         struct btrfs_block_rsv *dst_rsv;
 597         u64 num_bytes;
 598         int ret;
 599         bool release = false;
 600
 601         src_rsv = trans->block_rsv;
 602         dst_rsv = &root->fs_info->delayed_block_rsv;
 603
 604         num_bytes = btrfs_calc_trans_metadata_size(root, 1);
 605
 606         /*
 607          * btrfs_dirty_inode will update the inode under btrfs_join_transaction
 608          * which doesn't reserve space for speed.  This is a problem since we
 609          * still need to reserve space for this update, so try to reserve the
 610          * space.
 611          *
 612          * Now if src_rsv == delalloc_block_rsv we'll let it just steal since
 613          * we're accounted for.
 614          */
 615         if (!src_rsv || (!trans->bytes_reserved &&
 616                          src_rsv->type != BTRFS_BLOCK_RSV_DELALLOC)) {
 617                 ret = btrfs_block_rsv_add(root, dst_rsv, num_bytes,
 618                                           BTRFS_RESERVE_NO_FLUSH);
 619                 /*
 620                  * Since we're under a transaction reserve_metadata_bytes could
 621                  * try to commit the transaction which will make it return
 622                  * EAGAIN to make us stop the transaction we have, so return
 623                  * ENOSPC instead so that btrfs_dirty_inode knows what to do.
 624                  */
 625                 if (ret == -EAGAIN)
 626                         ret = -ENOSPC;
 627                 if (!ret) {
 628                         node->bytes_reserved = num_bytes;
 629                         trace_btrfs_space_reservation(root->fs_info,
 630                                                       "delayed_inode",
 631                                                       btrfs_ino(inode),
 632                                                       num_bytes, 1);
 633                 }
 634                 return ret;
 635         } else if (src_rsv->type == BTRFS_BLOCK_RSV_DELALLOC) {
 636                 spin_lock(&BTRFS_I(inode)->lock);
 637                 if (test_and_clear_bit(BTRFS_INODE_DELALLOC_META_RESERVED,
 638                                        &BTRFS_I(inode)->runtime_flags)) {
 639                         spin_unlock(&BTRFS_I(inode)->lock);
 640                         release = true;
 641                         goto migrate;
 642                 }
 643                 spin_unlock(&BTRFS_I(inode)->lock);
 644
 645                 /* Ok we didn't have space pre-reserved.  This shouldn't happen
 646                  * too often but it can happen if we do delalloc to an existing
 647                  * inode which gets dirtied because of the time update, and then
 648                  * isn't touched again until after the transaction commits and
 649                  * then we try to write out the data.  First try to be nice and
 650                  * reserve something strictly for us.  If not be a pain and try
 651                  * to steal from the delalloc block rsv.
 652                  */
 653                 ret = btrfs_block_rsv_add(root, dst_rsv, num_bytes,
 654                                           BTRFS_RESERVE_NO_FLUSH);
 655                 if (!ret)
 656                         goto out;
 657
 658                 ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes);
 659                 if (!WARN_ON(ret))
 660                         goto out;
 661
 662                 /*
 663                  * Ok this is a problem, let's just steal from the global rsv
 664                  * since this really shouldn't happen that often.
 665                  */
 666                 ret = btrfs_block_rsv_migrate(&root->fs_info->global_block_rsv,
 667                                               dst_rsv, num_bytes);
 668                 goto out;
 669         }
 670
 671 migrate:
 672         ret = btrfs_block_rsv_migrate(src_rsv, dst_rsv, num_bytes);
 673
 674 out:
 675         /*
 676          * Migrate only takes a reservation, it doesn't touch the size of the
 677          * block_rsv.  This is to simplify people who don't normally have things
 678          * migrated from their block rsv.  If they go to release their
 679          * reservation, that will decrease the size as well, so if migrate
 680          * reduced size we'd end up with a negative size.  But for the
 681          * delalloc_meta_reserved stuff we will only know to drop 1 reservation,
 682          * but we could in fact do this reserve/migrate dance several times
 683          * between the time we did the original reservation and we'd clean it
 684          * up.  So to take care of this, release the space for the meta
 685          * reservation here.  I think it may be time for a documentation page on
 686          * how block rsvs. work.
 687          */
 688         if (!ret) {
 689                 trace_btrfs_space_reservation(root->fs_info, "delayed_inode",
 690                                               btrfs_ino(inode), num_bytes, 1);
 691                 node->bytes_reserved = num_bytes;
 692         }
 693
 694         if (release) {
 695                 trace_btrfs_space_reservation(root->fs_info, "delalloc",
 696                                               btrfs_ino(inode), num_bytes, 0);
 697                 btrfs_block_rsv_release(root, src_rsv, num_bytes);
 698         }
 699
 700         return ret;
 701 }
 702
 703 static void btrfs_delayed_inode_release_metadata(struct btrfs_root *root,
 704                                                 struct btrfs_delayed_node *node)
 705 {
 706         struct btrfs_block_rsv *rsv;
 707
 708         if (!node->bytes_reserved)
 709                 return;
 710
 711         rsv = &root->fs_info->delayed_block_rsv;
 712         trace_btrfs_space_reservation(root->fs_info, "delayed_inode",
 713                                       node->inode_id, node->bytes_reserved, 0);
 714         btrfs_block_rsv_release(root, rsv,
 715                                 node->bytes_reserved);
 716         node->bytes_reserved = 0;
 717 }
 718
 719 /*
 720  * This helper will insert some continuous items into the same leaf according
 721  * to the free space of the leaf.
 722  */
 723 static int btrfs_batch_insert_items(struct btrfs_root *root,
 724                                     struct btrfs_path *path,
 725                                     struct btrfs_delayed_item *item)
 726 {
 727         struct btrfs_delayed_item *curr, *next;
 728         int free_space;
 729         int total_data_size = 0, total_size = 0;
 730         struct extent_buffer *leaf;
 731         char *data_ptr;
 732         struct btrfs_key *keys;
 733         u32 *data_size;
 734         struct list_head head;
 735         int slot;
 736         int nitems;
 737         int i;
 738         int ret = 0;
 739
 740         BUG_ON(!path->nodes[0]);
 741
 742         leaf = path->nodes[0];
 743         free_space = btrfs_leaf_free_space(root, leaf);
 744         INIT_LIST_HEAD(&head);
 745
 746         next = item;
 747         nitems = 0;
 748
 749         /*
 750          * count the number of the continuous items that we can insert in batch
 751          */
 752         while (total_size + next->data_len + sizeof(struct btrfs_item) <=
 753                free_space) {
 754                 total_data_size += next->data_len;
 755                 total_size += next->data_len + sizeof(struct btrfs_item);
 756                 list_add_tail(&next->tree_list, &head);
 757                 nitems++;
 758
 759                 curr = next;
 760                 next = __btrfs_next_delayed_item(curr);
 761                 if (!next)
 762                         break;
 763
 764                 if (!btrfs_is_continuous_delayed_item(curr, next))
 765                         break;
 766         }
 767
 768         if (!nitems) {
 769                 ret = 0;
 770                 goto out;
 771         }
 772
 773         /*
 774          * we need allocate some memory space, but it might cause the task
 775          * to sleep, so we set all locked nodes in the path to blocking locks
 776          * first.
 777          */
 778         btrfs_set_path_blocking(path);
 779
 780         keys = kmalloc_array(nitems, sizeof(struct btrfs_key), GFP_NOFS);
 781         if (!keys) {
 782                 ret = -ENOMEM;
 783                 goto out;
 784         }
 785
 786         data_size = kmalloc_array(nitems, sizeof(u32), GFP_NOFS);
 787         if (!data_size) {
 788                 ret = -ENOMEM;
 789                 goto error;
 790         }
 791
 792         /* get keys of all the delayed items */
 793         i = 0;
 794         list_for_each_entry(next, &head, tree_list) {
 795                 keys[i] = next->key;
 796                 data_size[i] = next->data_len;
 797                 i++;
 798         }
 799
 800         /* reset all the locked nodes in the patch to spinning locks. */
 801         btrfs_clear_path_blocking(path, NULL, 0);
 802
 803         /* insert the keys of the items */
 804         setup_items_for_insert(root, path, keys, data_size,
 805                                total_data_size, total_size, nitems);
 806
 807         /* insert the dir index items */
 808         slot = path->slots[0];
 809         list_for_each_entry_safe(curr, next, &head, tree_list) {
 810                 data_ptr = btrfs_item_ptr(leaf, slot, char);
 811                 write_extent_buffer(leaf, &curr->data,
 812                                     (unsigned long)data_ptr,
 813                                     curr->data_len);
 814                 slot++;
 815
 816                 btrfs_delayed_item_release_metadata(root, curr);
 817
 818                 list_del(&curr->tree_list);
 819                 btrfs_release_delayed_item(curr);
 820         }
 821
 822 error:
 823         kfree(data_size);
 824         kfree(keys);
 825 out:
 826         return ret;
 827 }
 828
 829 /*
 830  * This helper can just do simple insertion that needn't extend item for new
 831  * data, such as directory name index insertion, inode insertion.
 832  */
 833 static int btrfs_insert_delayed_item(struct btrfs_trans_handle *trans,
 834                                      struct btrfs_root *root,
 835                                      struct btrfs_path *path,
 836                                      struct btrfs_delayed_item *delayed_item)
 837 {
 838         struct extent_buffer *leaf;
 839         char *ptr;
 840         int ret;
 841
 842         ret = btrfs_insert_empty_item(trans, root, path, &delayed_item->key,
 843                                       delayed_item->data_len);
 844         if (ret < 0 && ret != -EEXIST)
 845                 return ret;
 846
 847         leaf = path->nodes[0];
 848
 849         ptr = btrfs_item_ptr(leaf, path->slots[0], char);
 850
 851         write_extent_buffer(leaf, delayed_item->data, (unsigned long)ptr,
 852                             delayed_item->data_len);
 853         btrfs_mark_buffer_dirty(leaf);
 854
 855         btrfs_delayed_item_release_metadata(root, delayed_item);
 856         return 0;
 857 }
 858
 859 /*
 860  * we insert an item first, then if there are some continuous items, we try
 861  * to insert those items into the same leaf.
 862  */
 863 static int btrfs_insert_delayed_items(struct btrfs_trans_handle *trans,
 864                                       struct btrfs_path *path,
 865                                       struct btrfs_root *root,
 866                                       struct btrfs_delayed_node *node)
 867 {
 868         struct btrfs_delayed_item *curr, *prev;
 869         int ret = 0;
 870
 871 do_again:
 872         mutex_lock(&node->mutex);
 873         curr = __btrfs_first_delayed_insertion_item(node);
 874         if (!curr)
 875                 goto insert_end;
 876
 877         ret = btrfs_insert_delayed_item(trans, root, path, curr);
 878         if (ret < 0) {
 879                 btrfs_release_path(path);
 880                 goto insert_end;
 881         }
 882
 883         prev = curr;
 884         curr = __btrfs_next_delayed_item(prev);
 885         if (curr && btrfs_is_continuous_delayed_item(prev, curr)) {
 886                 /* insert the continuous items into the same leaf */
 887                 path->slots[0]++;
 888                 btrfs_batch_insert_items(root, path, curr);
 889         }
 890         btrfs_release_delayed_item(prev);
 891         btrfs_mark_buffer_dirty(path->nodes[0]);
 892
 893         btrfs_release_path(path);
 894         mutex_unlock(&node->mutex);
 895         goto do_again;
 896
 897 insert_end:
 898         mutex_unlock(&node->mutex);
 899         return ret;
 900 }
 901
 902 static int btrfs_batch_delete_items(struct btrfs_trans_handle *trans,
 903                                     struct btrfs_root *root,
 904                                     struct btrfs_path *path,
 905                                     struct btrfs_delayed_item *item)
 906 {
 907         struct btrfs_delayed_item *curr, *next;
 908         struct extent_buffer *leaf;
 909         struct btrfs_key key;
 910         struct list_head head;
 911         int nitems, i, last_item;
 912         int ret = 0;
 913
 914         BUG_ON(!path->nodes[0]);
 915
 916         leaf = path->nodes[0];
 917
 918         i = path->slots[0];
 919         last_item = btrfs_header_nritems(leaf) - 1;
 920         if (i > last_item)
 921                 return -ENOENT; /* FIXME: Is errno suitable? */
 922
 923         next = item;
 924         INIT_LIST_HEAD(&head);
 925         btrfs_item_key_to_cpu(leaf, &key, i);
 926         nitems = 0;
 927         /*
 928          * count the number of the dir index items that we can delete in batch
 929          */
 930         while (btrfs_comp_cpu_keys(&next->key, &key) == 0) {
 931                 list_add_tail(&next->tree_list, &head);
 932                 nitems++;
 933
 934                 curr = next;
 935                 next = __btrfs_next_delayed_item(curr);
 936                 if (!next)
 937                         break;
 938
 939                 if (!btrfs_is_continuous_delayed_item(curr, next))
 940                         break;
 941
 942                 i++;
 943                 if (i > last_item)
 944                         break;
 945                 btrfs_item_key_to_cpu(leaf, &key, i);
 946         }
 947
 948         if (!nitems)
 949                 return 0;
 950
 951         ret = btrfs_del_items(trans, root, path, path->slots[0], nitems);
 952         if (ret)
 953                 goto out;
 954
 955         list_for_each_entry_safe(curr, next, &head, tree_list) {
 956                 btrfs_delayed_item_release_metadata(root, curr);
 957                 list_del(&curr->tree_list);
 958                 btrfs_release_delayed_item(curr);
 959         }
 960
 961 out:
 962         return ret;
 963 }
 964
 965 static int btrfs_delete_delayed_items(struct btrfs_trans_handle *trans,
 966                                       struct btrfs_path *path,
 967                                       struct btrfs_root *root,
 968                                       struct btrfs_delayed_node *node)
 969 {
 970         struct btrfs_delayed_item *curr, *prev;
 971         int ret = 0;
 972
 973 do_again:
 974         mutex_lock(&node->mutex);
 975         curr = __btrfs_first_delayed_deletion_item(node);
 976         if (!curr)
 977                 goto delete_fail;
 978
 979         ret = btrfs_search_slot(trans, root, &curr->key, path, -1, 1);
 980         if (ret < 0)
 981                 goto delete_fail;
 982         else if (ret > 0) {
 983                 /*
 984                  * can't find the item which the node points to, so this node
 985                  * is invalid, just drop it.
 986                  */
 987                 prev = curr;
 988                 curr = __btrfs_next_delayed_item(prev);
 989                 btrfs_release_delayed_item(prev);
 990                 ret = 0;
 991                 btrfs_release_path(path);
 992                 if (curr) {
 993                         mutex_unlock(&node->mutex);
 994                         goto do_again;
 995                 } else
 996                         goto delete_fail;
 997         }
 998
 999         btrfs_batch_delete_items(trans, root, path, curr);
1000         btrfs_release_path(path);
1001         mutex_unlock(&node->mutex);
1002         goto do_again;
1003
1004 delete_fail:
1005         btrfs_release_path(path);
1006         mutex_unlock(&node->mutex);
1007         return ret;
1008 }
1009
1010 static void btrfs_release_delayed_inode(struct btrfs_delayed_node *delayed_node)
1011 {
1012         struct btrfs_delayed_root *delayed_root;
1013
1014         if (delayed_node &&
1015             test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) {
1016                 BUG_ON(!delayed_node->root);
1017                 clear_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags);
1018                 delayed_node->count--;
1019
1020                 delayed_root = delayed_node->root->fs_info->delayed_root;
1021                 finish_one_item(delayed_root);
1022         }
1023 }
1024
1025 static void btrfs_release_delayed_iref(struct btrfs_delayed_node *delayed_node)
1026 {
1027         struct btrfs_delayed_root *delayed_root;
1028
1029         ASSERT(delayed_node->root);
1030         clear_bit(BTRFS_DELAYED_NODE_DEL_IREF, &delayed_node->flags);
1031         delayed_node->count--;
1032
1033         delayed_root = delayed_node->root->fs_info->delayed_root;
1034         finish_one_item(delayed_root);
1035 }
1036
1037 static int __btrfs_update_delayed_inode(struct btrfs_trans_handle *trans,
1038                                         struct btrfs_root *root,
1039                                         struct btrfs_path *path,
1040                                         struct btrfs_delayed_node *node)
1041 {
1042         struct btrfs_key key;
1043         struct btrfs_inode_item *inode_item;
1044         struct extent_buffer *leaf;
1045         int mod;
1046         int ret;
1047
1048         key.objectid = node->inode_id;
1049         key.type = BTRFS_INODE_ITEM_KEY;
1050         key.offset = 0;
1051
1052         if (test_bit(BTRFS_DELAYED_NODE_DEL_IREF, &node->flags))
1053                 mod = -1;
1054         else
1055                 mod = 1;
1056
1057         ret = btrfs_lookup_inode(trans, root, path, &key, mod);
1058         if (ret > 0) {
1059                 btrfs_release_path(path);
1060                 return -ENOENT;
1061         } else if (ret < 0) {
1062                 return ret;
1063         }
1064
1065         leaf = path->nodes[0];
1066         inode_item = btrfs_item_ptr(leaf, path->slots[0],
1067                                     struct btrfs_inode_item);
1068         write_extent_buffer(leaf, &node->inode_item, (unsigned long)inode_item,
1069                             sizeof(struct btrfs_inode_item));
1070         btrfs_mark_buffer_dirty(leaf);
1071
1072         if (!test_bit(BTRFS_DELAYED_NODE_DEL_IREF, &node->flags))
1073                 goto no_iref;
1074
1075         path->slots[0]++;
1076         if (path->slots[0] >= btrfs_header_nritems(leaf))
1077                 goto search;
1078 again:
1079         btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
1080         if (key.objectid != node->inode_id)
1081                 goto out;
1082
1083         if (key.type != BTRFS_INODE_REF_KEY &&
1084             key.type != BTRFS_INODE_EXTREF_KEY)
1085                 goto out;
1086
1087         /*
1088          * Delayed iref deletion is for the inode who has only one link,
1089          * so there is only one iref. The case that several irefs are
1090          * in the same item doesn't exist.
1091          */
1092         btrfs_del_item(trans, root, path);
1093 out:
1094         btrfs_release_delayed_iref(node);
1095 no_iref:
1096         btrfs_release_path(path);
1097 err_out:
1098         btrfs_delayed_inode_release_metadata(root, node);
1099         btrfs_release_delayed_inode(node);
1100
1101         return ret;
1102
1103 search:
1104         btrfs_release_path(path);
1105
1106         key.type = BTRFS_INODE_EXTREF_KEY;
1107         key.offset = -1;
1108         ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
1109         if (ret < 0)
1110                 goto err_out;
1111         ASSERT(ret);
1112
1113         ret = 0;
1114         leaf = path->nodes[0];
1115         path->slots[0]--;
1116         goto again;
1117 }
1118
1119 static inline int btrfs_update_delayed_inode(struct btrfs_trans_handle *trans,
1120                                              struct btrfs_root *root,
1121                                              struct btrfs_path *path,
1122                                              struct btrfs_delayed_node *node)
1123 {
1124         int ret;
1125
1126         mutex_lock(&node->mutex);
1127         if (!test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &node->flags)) {
1128                 mutex_unlock(&node->mutex);
1129                 return 0;
1130         }
1131
1132         ret = __btrfs_update_delayed_inode(trans, root, path, node);
1133         mutex_unlock(&node->mutex);
1134         return ret;
1135 }
1136
1137 static inline int
1138 __btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
1139                                    struct btrfs_path *path,
1140                                    struct btrfs_delayed_node *node)
1141 {
1142         int ret;
1143
1144         ret = btrfs_insert_delayed_items(trans, path, node->root, node);
1145         if (ret)
1146                 return ret;
1147
1148         ret = btrfs_delete_delayed_items(trans, path, node->root, node);
1149         if (ret)
1150                 return ret;
1151
1152         ret = btrfs_update_delayed_inode(trans, node->root, path, node);
1153         return ret;
1154 }
1155
1156 /*
1157  * Called when committing the transaction.
1158  * Returns 0 on success.
1159  * Returns < 0 on error and returns with an aborted transaction with any
1160  * outstanding delayed items cleaned up.
1161  */
1162 static int __btrfs_run_delayed_items(struct btrfs_trans_handle *trans,
1163                                      struct btrfs_root *root, int nr)
1164 {
1165         struct btrfs_delayed_root *delayed_root;
1166         struct btrfs_delayed_node *curr_node, *prev_node;
1167         struct btrfs_path *path;
1168         struct btrfs_block_rsv *block_rsv;
1169         int ret = 0;
1170         bool count = (nr > 0);
1171
1172         if (trans->aborted)
1173                 return -EIO;
1174
1175         path = btrfs_alloc_path();
1176         if (!path)
1177                 return -ENOMEM;
1178         path->leave_spinning = 1;
1179
1180         block_rsv = trans->block_rsv;
1181         trans->block_rsv = &root->fs_info->delayed_block_rsv;
1182
1183         delayed_root = btrfs_get_delayed_root(root);
1184
1185         curr_node = btrfs_first_delayed_node(delayed_root);
1186         while (curr_node && (!count || (count && nr--))) {
1187                 ret = __btrfs_commit_inode_delayed_items(trans, path,
1188                                                          curr_node);
1189                 if (ret) {
1190                         btrfs_release_delayed_node(curr_node);
1191                         curr_node = NULL;
1192                         btrfs_abort_transaction(trans, root, ret);
1193                         break;
1194                 }
1195
1196                 prev_node = curr_node;
1197                 curr_node = btrfs_next_delayed_node(curr_node);
1198                 btrfs_release_delayed_node(prev_node);
1199         }
1200
1201         if (curr_node)
1202                 btrfs_release_delayed_node(curr_node);
1203         btrfs_free_path(path);
1204         trans->block_rsv = block_rsv;
1205
1206         return ret;
1207 }
1208
1209 int btrfs_run_delayed_items(struct btrfs_trans_handle *trans,
1210                             struct btrfs_root *root)
1211 {
1212         return __btrfs_run_delayed_items(trans, root, -1);
1213 }
1214
1215 int btrfs_run_delayed_items_nr(struct btrfs_trans_handle *trans,
1216                                struct btrfs_root *root, int nr)
1217 {
1218         return __btrfs_run_delayed_items(trans, root, nr);
1219 }
1220
1221 int btrfs_commit_inode_delayed_items(struct btrfs_trans_handle *trans,
1222                                      struct inode *inode)
1223 {
1224         struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode);
1225         struct btrfs_path *path;
1226         struct btrfs_block_rsv *block_rsv;
1227         int ret;
1228
1229         if (!delayed_node)
1230                 return 0;
1231
1232         mutex_lock(&delayed_node->mutex);
1233         if (!delayed_node->count) {
1234                 mutex_unlock(&delayed_node->mutex);
1235                 btrfs_release_delayed_node(delayed_node);
1236                 return 0;
1237         }
1238         mutex_unlock(&delayed_node->mutex);
1239
1240         path = btrfs_alloc_path();
1241         if (!path) {
1242                 btrfs_release_delayed_node(delayed_node);
1243                 return -ENOMEM;
1244         }
1245         path->leave_spinning = 1;
1246
1247         block_rsv = trans->block_rsv;
1248         trans->block_rsv = &delayed_node->root->fs_info->delayed_block_rsv;
1249
1250         ret = __btrfs_commit_inode_delayed_items(trans, path, delayed_node);
1251
1252         btrfs_release_delayed_node(delayed_node);
1253         btrfs_free_path(path);
1254         trans->block_rsv = block_rsv;
1255
1256         return ret;
1257 }
1258
1259 int btrfs_commit_inode_delayed_inode(struct inode *inode)
1260 {
1261         struct btrfs_trans_handle *trans;
1262         struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode);
1263         struct btrfs_path *path;
1264         struct btrfs_block_rsv *block_rsv;
1265         int ret;
1266
1267         if (!delayed_node)
1268                 return 0;
1269
1270         mutex_lock(&delayed_node->mutex);
1271         if (!test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) {
1272                 mutex_unlock(&delayed_node->mutex);
1273                 btrfs_release_delayed_node(delayed_node);
1274                 return 0;
1275         }
1276         mutex_unlock(&delayed_node->mutex);
1277
1278         trans = btrfs_join_transaction(delayed_node->root);
1279         if (IS_ERR(trans)) {
1280                 ret = PTR_ERR(trans);
1281                 goto out;
1282         }
1283
1284         path = btrfs_alloc_path();
1285         if (!path) {
1286                 ret = -ENOMEM;
1287                 goto trans_out;
1288         }
1289         path->leave_spinning = 1;
1290
1291         block_rsv = trans->block_rsv;
1292         trans->block_rsv = &delayed_node->root->fs_info->delayed_block_rsv;
1293
1294         mutex_lock(&delayed_node->mutex);
1295         if (test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags))
1296                 ret = __btrfs_update_delayed_inode(trans, delayed_node->root,
1297                                                    path, delayed_node);
1298         else
1299                 ret = 0;
1300         mutex_unlock(&delayed_node->mutex);
1301
1302         btrfs_free_path(path);
1303         trans->block_rsv = block_rsv;
1304 trans_out:
1305         btrfs_end_transaction(trans, delayed_node->root);
1306         btrfs_btree_balance_dirty(delayed_node->root);
1307 out:
1308         btrfs_release_delayed_node(delayed_node);
1309
1310         return ret;
1311 }
1312
1313 void btrfs_remove_delayed_node(struct inode *inode)
1314 {
1315         struct btrfs_delayed_node *delayed_node;
1316
1317         delayed_node = ACCESS_ONCE(BTRFS_I(inode)->delayed_node);
1318         if (!delayed_node)
1319                 return;
1320
1321         BTRFS_I(inode)->delayed_node = NULL;
1322         btrfs_release_delayed_node(delayed_node);
1323 }
1324
1325 struct btrfs_async_delayed_work {
1326         struct btrfs_delayed_root *delayed_root;
1327         int nr;
1328         struct btrfs_work work;
1329 };
1330
1331 static void btrfs_async_run_delayed_root(struct btrfs_work *work)
1332 {
1333         struct btrfs_async_delayed_work *async_work;
1334         struct btrfs_delayed_root *delayed_root;
1335         struct btrfs_trans_handle *trans;
1336         struct btrfs_path *path;
1337         struct btrfs_delayed_node *delayed_node = NULL;
1338         struct btrfs_root *root;
1339         struct btrfs_block_rsv *block_rsv;
1340         int total_done = 0;
1341
1342         async_work = container_of(work, struct btrfs_async_delayed_work, work);
1343         delayed_root = async_work->delayed_root;
1344
1345         path = btrfs_alloc_path();
1346         if (!path)
1347                 goto out;
1348
1349 again:
1350         if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND / 2)
1351                 goto free_path;
1352
1353         delayed_node = btrfs_first_prepared_delayed_node(delayed_root);
1354         if (!delayed_node)
1355                 goto free_path;
1356
1357         path->leave_spinning = 1;
1358         root = delayed_node->root;
1359
1360         trans = btrfs_join_transaction(root);
1361         if (IS_ERR(trans))
1362                 goto release_path;
1363
1364         block_rsv = trans->block_rsv;
1365         trans->block_rsv = &root->fs_info->delayed_block_rsv;
1366
1367         __btrfs_commit_inode_delayed_items(trans, path, delayed_node);
1368
1369         trans->block_rsv = block_rsv;
1370         btrfs_end_transaction(trans, root);
1371         btrfs_btree_balance_dirty_nodelay(root);
1372
1373 release_path:
1374         btrfs_release_path(path);
1375         total_done++;
1376
1377         btrfs_release_prepared_delayed_node(delayed_node);
1378         if (async_work->nr == 0 || total_done < async_work->nr)
1379                 goto again;
1380
1381 free_path:
1382         btrfs_free_path(path);
1383 out:
1384         wake_up(&delayed_root->wait);
1385         kfree(async_work);
1386 }
1387
1388
1389 static int btrfs_wq_run_delayed_node(struct btrfs_delayed_root *delayed_root,
1390                                      struct btrfs_fs_info *fs_info, int nr)
1391 {
1392         struct btrfs_async_delayed_work *async_work;
1393
1394         if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND)
1395                 return 0;
1396
1397         async_work = kmalloc(sizeof(*async_work), GFP_NOFS);
1398         if (!async_work)
1399                 return -ENOMEM;
1400
1401         async_work->delayed_root = delayed_root;
1402         btrfs_init_work(&async_work->work, btrfs_delayed_meta_helper,
1403                         btrfs_async_run_delayed_root, NULL, NULL);
1404         async_work->nr = nr;
1405
1406         btrfs_queue_work(fs_info->delayed_workers, &async_work->work);
1407         return 0;
1408 }
1409
1410 void btrfs_assert_delayed_root_empty(struct btrfs_root *root)
1411 {
1412         struct btrfs_delayed_root *delayed_root;
1413         delayed_root = btrfs_get_delayed_root(root);
1414         WARN_ON(btrfs_first_delayed_node(delayed_root));
1415 }
1416
1417 static int could_end_wait(struct btrfs_delayed_root *delayed_root, int seq)
1418 {
1419         int val = atomic_read(&delayed_root->items_seq);
1420
1421         if (val < seq || val >= seq + BTRFS_DELAYED_BATCH)
1422                 return 1;
1423
1424         if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND)
1425                 return 1;
1426
1427         return 0;
1428 }
1429
1430 void btrfs_balance_delayed_items(struct btrfs_root *root)
1431 {
1432         struct btrfs_delayed_root *delayed_root;
1433         struct btrfs_fs_info *fs_info = root->fs_info;
1434
1435         delayed_root = btrfs_get_delayed_root(root);
1436
1437         if (atomic_read(&delayed_root->items) < BTRFS_DELAYED_BACKGROUND)
1438                 return;
1439
1440         if (atomic_read(&delayed_root->items) >= BTRFS_DELAYED_WRITEBACK) {
1441                 int seq;
1442                 int ret;
1443
1444                 seq = atomic_read(&delayed_root->items_seq);
1445
1446                 ret = btrfs_wq_run_delayed_node(delayed_root, fs_info, 0);
1447                 if (ret)
1448                         return;
1449
1450                 wait_event_interruptible(delayed_root->wait,
1451                                          could_end_wait(delayed_root, seq));
1452                 return;
1453         }
1454
1455         btrfs_wq_run_delayed_node(delayed_root, fs_info, BTRFS_DELAYED_BATCH);
1456 }
1457
1458 /* Will return 0 or -ENOMEM */
1459 int btrfs_insert_delayed_dir_index(struct btrfs_trans_handle *trans,
1460                                    struct btrfs_root *root, const char *name,
1461                                    int name_len, struct inode *dir,
1462                                    struct btrfs_disk_key *disk_key, u8 type,
1463                                    u64 index)
1464 {
1465         struct btrfs_delayed_node *delayed_node;
1466         struct btrfs_delayed_item *delayed_item;
1467         struct btrfs_dir_item *dir_item;
1468         int ret;
1469
1470         delayed_node = btrfs_get_or_create_delayed_node(dir);
1471         if (IS_ERR(delayed_node))
1472                 return PTR_ERR(delayed_node);
1473
1474         delayed_item = btrfs_alloc_delayed_item(sizeof(*dir_item) + name_len);
1475         if (!delayed_item) {
1476                 ret = -ENOMEM;
1477                 goto release_node;
1478         }
1479
1480         delayed_item->key.objectid = btrfs_ino(dir);
1481         delayed_item->key.type = BTRFS_DIR_INDEX_KEY;
1482         delayed_item->key.offset = index;
1483
1484         dir_item = (struct btrfs_dir_item *)delayed_item->data;
1485         dir_item->location = *disk_key;
1486         btrfs_set_stack_dir_transid(dir_item, trans->transid);
1487         btrfs_set_stack_dir_data_len(dir_item, 0);
1488         btrfs_set_stack_dir_name_len(dir_item, name_len);
1489         btrfs_set_stack_dir_type(dir_item, type);
1490         memcpy((char *)(dir_item + 1), name, name_len);
1491
1492         ret = btrfs_delayed_item_reserve_metadata(trans, root, delayed_item);
1493         /*
1494          * we have reserved enough space when we start a new transaction,
1495          * so reserving metadata failure is impossible
1496          */
1497         BUG_ON(ret);
1498
1499
1500         mutex_lock(&delayed_node->mutex);
1501         ret = __btrfs_add_delayed_insertion_item(delayed_node, delayed_item);
1502         if (unlikely(ret)) {
1503                 btrfs_err(root->fs_info, "err add delayed dir index item(name: %.*s) "
1504                                 "into the insertion tree of the delayed node"
1505                                 "(root id: %llu, inode id: %llu, errno: %d)",
1506                                 name_len, name, delayed_node->root->objectid,
1507                                 delayed_node->inode_id, ret);
1508                 BUG();
1509         }
1510         mutex_unlock(&delayed_node->mutex);
1511
1512 release_node:
1513         btrfs_release_delayed_node(delayed_node);
1514         return ret;
1515 }
1516
1517 static int btrfs_delete_delayed_insertion_item(struct btrfs_root *root,
1518                                                struct btrfs_delayed_node *node,
1519                                                struct btrfs_key *key)
1520 {
1521         struct btrfs_delayed_item *item;
1522
1523         mutex_lock(&node->mutex);
1524         item = __btrfs_lookup_delayed_insertion_item(node, key);
1525         if (!item) {
1526                 mutex_unlock(&node->mutex);
1527                 return 1;
1528         }
1529
1530         btrfs_delayed_item_release_metadata(root, item);
1531         btrfs_release_delayed_item(item);
1532         mutex_unlock(&node->mutex);
1533         return 0;
1534 }
1535
1536 int btrfs_delete_delayed_dir_index(struct btrfs_trans_handle *trans,
1537                                    struct btrfs_root *root, struct inode *dir,
1538                                    u64 index)
1539 {
1540         struct btrfs_delayed_node *node;
1541         struct btrfs_delayed_item *item;
1542         struct btrfs_key item_key;
1543         int ret;
1544
1545         node = btrfs_get_or_create_delayed_node(dir);
1546         if (IS_ERR(node))
1547                 return PTR_ERR(node);
1548
1549         item_key.objectid = btrfs_ino(dir);
1550         item_key.type = BTRFS_DIR_INDEX_KEY;
1551         item_key.offset = index;
1552
1553         ret = btrfs_delete_delayed_insertion_item(root, node, &item_key);
1554         if (!ret)
1555                 goto end;
1556
1557         item = btrfs_alloc_delayed_item(0);
1558         if (!item) {
1559                 ret = -ENOMEM;
1560                 goto end;
1561         }
1562
1563         item->key = item_key;
1564
1565         ret = btrfs_delayed_item_reserve_metadata(trans, root, item);
1566         /*
1567          * we have reserved enough space when we start a new transaction,
1568          * so reserving metadata failure is impossible.
1569          */
1570         BUG_ON(ret);
1571
1572         mutex_lock(&node->mutex);
1573         ret = __btrfs_add_delayed_deletion_item(node, item);
1574         if (unlikely(ret)) {
1575                 btrfs_err(root->fs_info, "err add delayed dir index item(index: %llu) "
1576                                 "into the deletion tree of the delayed node"
1577                                 "(root id: %llu, inode id: %llu, errno: %d)",
1578                                 index, node->root->objectid, node->inode_id,
1579                                 ret);
1580                 BUG();
1581         }
1582         mutex_unlock(&node->mutex);
1583 end:
1584         btrfs_release_delayed_node(node);
1585         return ret;
1586 }
1587
1588 int btrfs_inode_delayed_dir_index_count(struct inode *inode)
1589 {
1590         struct btrfs_delayed_node *delayed_node = btrfs_get_delayed_node(inode);
1591
1592         if (!delayed_node)
1593                 return -ENOENT;
1594
1595         /*
1596          * Since we have held i_mutex of this directory, it is impossible that
1597          * a new directory index is added into the delayed node and index_cnt
1598          * is updated now. So we needn't lock the delayed node.
1599          */
1600         if (!delayed_node->index_cnt) {
1601                 btrfs_release_delayed_node(delayed_node);
1602                 return -EINVAL;
1603         }
1604
1605         BTRFS_I(inode)->index_cnt = delayed_node->index_cnt;
1606         btrfs_release_delayed_node(delayed_node);
1607         return 0;
1608 }
1609
1610 void btrfs_get_delayed_items(struct inode *inode, struct list_head *ins_list,
1611                              struct list_head *del_list)
1612 {
1613         struct btrfs_delayed_node *delayed_node;
1614         struct btrfs_delayed_item *item;
1615
1616         delayed_node = btrfs_get_delayed_node(inode);
1617         if (!delayed_node)
1618                 return;
1619
1620         mutex_lock(&delayed_node->mutex);
1621         item = __btrfs_first_delayed_insertion_item(delayed_node);
1622         while (item) {
1623                 atomic_inc(&item->refs);
1624                 list_add_tail(&item->readdir_list, ins_list);
1625                 item = __btrfs_next_delayed_item(item);
1626         }
1627
1628         item = __btrfs_first_delayed_deletion_item(delayed_node);
1629         while (item) {
1630                 atomic_inc(&item->refs);
1631                 list_add_tail(&item->readdir_list, del_list);
1632                 item = __btrfs_next_delayed_item(item);
1633         }
1634         mutex_unlock(&delayed_node->mutex);
1635         /*
1636          * This delayed node is still cached in the btrfs inode, so refs
1637          * must be > 1 now, and we needn't check it is going to be freed
1638          * or not.
1639          *
1640          * Besides that, this function is used to read dir, we do not
1641          * insert/delete delayed items in this period. So we also needn't
1642          * requeue or dequeue this delayed node.
1643          */
1644         atomic_dec(&delayed_node->refs);
1645 }
1646
1647 void btrfs_put_delayed_items(struct list_head *ins_list,
1648                              struct list_head *del_list)
1649 {
1650         struct btrfs_delayed_item *curr, *next;
1651
1652         list_for_each_entry_safe(curr, next, ins_list, readdir_list) {
1653                 list_del(&curr->readdir_list);
1654                 if (atomic_dec_and_test(&curr->refs))
1655                         kfree(curr);
1656         }
1657
1658         list_for_each_entry_safe(curr, next, del_list, readdir_list) {
1659                 list_del(&curr->readdir_list);
1660                 if (atomic_dec_and_test(&curr->refs))
1661                         kfree(curr);
1662         }
1663 }
1664
1665 int btrfs_should_delete_dir_index(struct list_head *del_list,
1666                                   u64 index)
1667 {
1668         struct btrfs_delayed_item *curr, *next;
1669         int ret;
1670
1671         if (list_empty(del_list))
1672                 return 0;
1673
1674         list_for_each_entry_safe(curr, next, del_list, readdir_list) {
1675                 if (curr->key.offset > index)
1676                         break;
1677
1678                 list_del(&curr->readdir_list);
1679                 ret = (curr->key.offset == index);
1680
1681                 if (atomic_dec_and_test(&curr->refs))
1682                         kfree(curr);
1683
1684                 if (ret)
1685                         return 1;
1686                 else
1687                         continue;
1688         }
1689         return 0;
1690 }
1691
1692 /*
1693  * btrfs_readdir_delayed_dir_index - read dir info stored in the delayed tree
1694  *
1695  */
1696 int btrfs_readdir_delayed_dir_index(struct dir_context *ctx,
1697                                     struct list_head *ins_list, bool *emitted)
1698 {
1699         struct btrfs_dir_item *di;
1700         struct btrfs_delayed_item *curr, *next;
1701         struct btrfs_key location;
1702         char *name;
1703         int name_len;
1704         int over = 0;
1705         unsigned char d_type;
1706
1707         if (list_empty(ins_list))
1708                 return 0;
1709
1710         /*
1711          * Changing the data of the delayed item is impossible. So
1712          * we needn't lock them. And we have held i_mutex of the
1713          * directory, nobody can delete any directory indexes now.
1714          */
1715         list_for_each_entry_safe(curr, next, ins_list, readdir_list) {
1716                 list_del(&curr->readdir_list);
1717
1718                 if (curr->key.offset < ctx->pos) {
1719                         if (atomic_dec_and_test(&curr->refs))
1720                                 kfree(curr);
1721                         continue;
1722                 }
1723
1724                 ctx->pos = curr->key.offset;
1725
1726                 di = (struct btrfs_dir_item *)curr->data;
1727                 name = (char *)(di + 1);
1728                 name_len = btrfs_stack_dir_name_len(di);
1729
1730                 d_type = btrfs_filetype_table[di->type];
1731                 btrfs_disk_key_to_cpu(&location, &di->location);
1732
1733                 over = !dir_emit(ctx, name, name_len,
1734                                location.objectid, d_type);
1735
1736                 if (atomic_dec_and_test(&curr->refs))
1737                         kfree(curr);
1738
1739                 if (over)
1740                         return 1;
1741                 *emitted = true;
1742         }
1743         return 0;
1744 }
1745
1746 static void fill_stack_inode_item(struct btrfs_trans_handle *trans,
1747                                   struct btrfs_inode_item *inode_item,
1748                                   struct inode *inode)
1749 {
1750         btrfs_set_stack_inode_uid(inode_item, i_uid_read(inode));
1751         btrfs_set_stack_inode_gid(inode_item, i_gid_read(inode));
1752         btrfs_set_stack_inode_size(inode_item, BTRFS_I(inode)->disk_i_size);
1753         btrfs_set_stack_inode_mode(inode_item, inode->i_mode);
1754         btrfs_set_stack_inode_nlink(inode_item, inode->i_nlink);
1755         btrfs_set_stack_inode_nbytes(inode_item, inode_get_bytes(inode));
1756         btrfs_set_stack_inode_generation(inode_item,
1757                                          BTRFS_I(inode)->generation);
1758         btrfs_set_stack_inode_sequence(inode_item, inode->i_version);
1759         btrfs_set_stack_inode_transid(inode_item, trans->transid);
1760         btrfs_set_stack_inode_rdev(inode_item, inode->i_rdev);
1761         btrfs_set_stack_inode_flags(inode_item, BTRFS_I(inode)->flags);
1762         btrfs_set_stack_inode_block_group(inode_item, 0);
1763
1764         btrfs_set_stack_timespec_sec(&inode_item->atime,
1765                                      inode->i_atime.tv_sec);
1766         btrfs_set_stack_timespec_nsec(&inode_item->atime,
1767                                       inode->i_atime.tv_nsec);
1768
1769         btrfs_set_stack_timespec_sec(&inode_item->mtime,
1770                                      inode->i_mtime.tv_sec);
1771         btrfs_set_stack_timespec_nsec(&inode_item->mtime,
1772                                       inode->i_mtime.tv_nsec);
1773
1774         btrfs_set_stack_timespec_sec(&inode_item->ctime,
1775                                      inode->i_ctime.tv_sec);
1776         btrfs_set_stack_timespec_nsec(&inode_item->ctime,
1777                                       inode->i_ctime.tv_nsec);
1778
1779         btrfs_set_stack_timespec_sec(&inode_item->otime,
1780                                      BTRFS_I(inode)->i_otime.tv_sec);
1781         btrfs_set_stack_timespec_nsec(&inode_item->otime,
1782                                      BTRFS_I(inode)->i_otime.tv_nsec);
1783 }
1784
1785 int btrfs_fill_inode(struct inode *inode, u32 *rdev)
1786 {
1787         struct btrfs_delayed_node *delayed_node;
1788         struct btrfs_inode_item *inode_item;
1789
1790         delayed_node = btrfs_get_delayed_node(inode);
1791         if (!delayed_node)
1792                 return -ENOENT;
1793
1794         mutex_lock(&delayed_node->mutex);
1795         if (!test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) {
1796                 mutex_unlock(&delayed_node->mutex);
1797                 btrfs_release_delayed_node(delayed_node);
1798                 return -ENOENT;
1799         }
1800
1801         inode_item = &delayed_node->inode_item;
1802
1803         i_uid_write(inode, btrfs_stack_inode_uid(inode_item));
1804         i_gid_write(inode, btrfs_stack_inode_gid(inode_item));
1805         btrfs_i_size_write(inode, btrfs_stack_inode_size(inode_item));
1806         inode->i_mode = btrfs_stack_inode_mode(inode_item);
1807         set_nlink(inode, btrfs_stack_inode_nlink(inode_item));
1808         inode_set_bytes(inode, btrfs_stack_inode_nbytes(inode_item));
1809         BTRFS_I(inode)->generation = btrfs_stack_inode_generation(inode_item);
1810         BTRFS_I(inode)->last_trans = btrfs_stack_inode_transid(inode_item);
1811
1812         inode->i_version = btrfs_stack_inode_sequence(inode_item);
1813         inode->i_rdev = 0;
1814         *rdev = btrfs_stack_inode_rdev(inode_item);
1815         BTRFS_I(inode)->flags = btrfs_stack_inode_flags(inode_item);
1816
1817         inode->i_atime.tv_sec = btrfs_stack_timespec_sec(&inode_item->atime);
1818         inode->i_atime.tv_nsec = btrfs_stack_timespec_nsec(&inode_item->atime);
1819
1820         inode->i_mtime.tv_sec = btrfs_stack_timespec_sec(&inode_item->mtime);
1821         inode->i_mtime.tv_nsec = btrfs_stack_timespec_nsec(&inode_item->mtime);
1822
1823         inode->i_ctime.tv_sec = btrfs_stack_timespec_sec(&inode_item->ctime);
1824         inode->i_ctime.tv_nsec = btrfs_stack_timespec_nsec(&inode_item->ctime);
1825
1826         BTRFS_I(inode)->i_otime.tv_sec =
1827                 btrfs_stack_timespec_sec(&inode_item->otime);
1828         BTRFS_I(inode)->i_otime.tv_nsec =
1829                 btrfs_stack_timespec_nsec(&inode_item->otime);
1830
1831         inode->i_generation = BTRFS_I(inode)->generation;
1832         BTRFS_I(inode)->index_cnt = (u64)-1;
1833
1834         mutex_unlock(&delayed_node->mutex);
1835         btrfs_release_delayed_node(delayed_node);
1836         return 0;
1837 }
1838
1839 int btrfs_delayed_update_inode(struct btrfs_trans_handle *trans,
1840                                struct btrfs_root *root, struct inode *inode)
1841 {
1842         struct btrfs_delayed_node *delayed_node;
1843         int ret = 0;
1844
1845         delayed_node = btrfs_get_or_create_delayed_node(inode);
1846         if (IS_ERR(delayed_node))
1847                 return PTR_ERR(delayed_node);
1848
1849         mutex_lock(&delayed_node->mutex);
1850         if (test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) {
1851                 fill_stack_inode_item(trans, &delayed_node->inode_item, inode);
1852                 goto release_node;
1853         }
1854
1855         ret = btrfs_delayed_inode_reserve_metadata(trans, root, inode,
1856                                                    delayed_node);
1857         if (ret)
1858                 goto release_node;
1859
1860         fill_stack_inode_item(trans, &delayed_node->inode_item, inode);
1861         set_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags);
1862         delayed_node->count++;
1863         atomic_inc(&root->fs_info->delayed_root->items);
1864 release_node:
1865         mutex_unlock(&delayed_node->mutex);
1866         btrfs_release_delayed_node(delayed_node);
1867         return ret;
1868 }
1869
1870 int btrfs_delayed_delete_inode_ref(struct inode *inode)
1871 {
1872         struct btrfs_delayed_node *delayed_node;
1873
1874         /*
1875          * we don't do delayed inode updates during log recovery because it
1876          * leads to enospc problems.  This means we also can't do
1877          * delayed inode refs
1878          */
1879         if (BTRFS_I(inode)->root->fs_info->log_root_recovering)
1880                 return -EAGAIN;
1881
1882         delayed_node = btrfs_get_or_create_delayed_node(inode);
1883         if (IS_ERR(delayed_node))
1884                 return PTR_ERR(delayed_node);
1885
1886         /*
1887          * We don't reserve space for inode ref deletion is because:
1888          * - We ONLY do async inode ref deletion for the inode who has only
1889          *   one link(i_nlink == 1), it means there is only one inode ref.
1890          *   And in most case, the inode ref and the inode item are in the
1891          *   same leaf, and we will deal with them at the same time.
1892          *   Since we are sure we will reserve the space for the inode item,
1893          *   it is unnecessary to reserve space for inode ref deletion.
1894          * - If the inode ref and the inode item are not in the same leaf,
1895          *   We also needn't worry about enospc problem, because we reserve
1896          *   much more space for the inode update than it needs.
1897          * - At the worst, we can steal some space from the global reservation.
1898          *   It is very rare.
1899          */
1900         mutex_lock(&delayed_node->mutex);
1901         if (test_bit(BTRFS_DELAYED_NODE_DEL_IREF, &delayed_node->flags))
1902                 goto release_node;
1903
1904         set_bit(BTRFS_DELAYED_NODE_DEL_IREF, &delayed_node->flags);
1905         delayed_node->count++;
1906         atomic_inc(&BTRFS_I(inode)->root->fs_info->delayed_root->items);
1907 release_node:
1908         mutex_unlock(&delayed_node->mutex);
1909         btrfs_release_delayed_node(delayed_node);
1910         return 0;
1911 }
1912
1913 static void __btrfs_kill_delayed_node(struct btrfs_delayed_node *delayed_node)
1914 {
1915         struct btrfs_root *root = delayed_node->root;
1916         struct btrfs_delayed_item *curr_item, *prev_item;
1917
1918         mutex_lock(&delayed_node->mutex);
1919         curr_item = __btrfs_first_delayed_insertion_item(delayed_node);
1920         while (curr_item) {
1921                 btrfs_delayed_item_release_metadata(root, curr_item);
1922                 prev_item = curr_item;
1923                 curr_item = __btrfs_next_delayed_item(prev_item);
1924                 btrfs_release_delayed_item(prev_item);
1925         }
1926
1927         curr_item = __btrfs_first_delayed_deletion_item(delayed_node);
1928         while (curr_item) {
1929                 btrfs_delayed_item_release_metadata(root, curr_item);
1930                 prev_item = curr_item;
1931                 curr_item = __btrfs_next_delayed_item(prev_item);
1932                 btrfs_release_delayed_item(prev_item);
1933         }
1934
1935         if (test_bit(BTRFS_DELAYED_NODE_DEL_IREF, &delayed_node->flags))
1936                 btrfs_release_delayed_iref(delayed_node);
1937
1938         if (test_bit(BTRFS_DELAYED_NODE_INODE_DIRTY, &delayed_node->flags)) {
1939                 btrfs_delayed_inode_release_metadata(root, delayed_node);
1940                 btrfs_release_delayed_inode(delayed_node);
1941         }
1942         mutex_unlock(&delayed_node->mutex);
1943 }
1944
1945 void btrfs_kill_delayed_inode_items(struct inode *inode)
1946 {
1947         struct btrfs_delayed_node *delayed_node;
1948
1949         delayed_node = btrfs_get_delayed_node(inode);
1950         if (!delayed_node)
1951                 return;
1952
1953         __btrfs_kill_delayed_node(delayed_node);
1954         btrfs_release_delayed_node(delayed_node);
1955 }
1956
1957 void btrfs_kill_all_delayed_nodes(struct btrfs_root *root)
1958 {
1959         u64 inode_id = 0;
1960         struct btrfs_delayed_node *delayed_nodes[8];
1961         int i, n;
1962
1963         while (1) {
1964                 spin_lock(&root->inode_lock);
1965                 n = radix_tree_gang_lookup(&root->delayed_nodes_tree,
1966                                            (void **)delayed_nodes, inode_id,
1967                                            ARRAY_SIZE(delayed_nodes));
1968                 if (!n) {
1969                         spin_unlock(&root->inode_lock);
1970                         break;
1971                 }
1972
1973                 inode_id = delayed_nodes[n - 1]->inode_id + 1;
1974
1975                 for (i = 0; i < n; i++)
1976                         atomic_inc(&delayed_nodes[i]->refs);
1977                 spin_unlock(&root->inode_lock);
1978
1979                 for (i = 0; i < n; i++) {
1980                         __btrfs_kill_delayed_node(delayed_nodes[i]);
1981                         btrfs_release_delayed_node(delayed_nodes[i]);
1982                 }
1983         }
1984 }
1985
1986 void btrfs_destroy_delayed_inodes(struct btrfs_root *root)
1987 {
1988         struct btrfs_delayed_root *delayed_root;
1989         struct btrfs_delayed_node *curr_node, *prev_node;
1990
1991         delayed_root = btrfs_get_delayed_root(root);
1992
1993         curr_node = btrfs_first_delayed_node(delayed_root);
1994         while (curr_node) {
1995                 __btrfs_kill_delayed_node(curr_node);
1996
1997                 prev_node = curr_node;
1998                 curr_node = btrfs_next_delayed_node(curr_node);
1999                 btrfs_release_delayed_node(prev_node);
2000         }
2001 }
2002