fs/btrfs/tree-mod-log.c

   1 // SPDX-License-Identifier: GPL-2.0
   2
   3 #include "tree-mod-log.h"
   4 #include "disk-io.h"
   5
   6 struct tree_mod_root {
   7         u64 logical;
   8         u8 level;
   9 };
  10
  11 struct tree_mod_elem {
  12         struct rb_node node;
  13         u64 logical;
  14         u64 seq;
  15         enum btrfs_mod_log_op op;
  16
  17         /*
  18          * This is used for BTRFS_MOD_LOG_KEY_* and BTRFS_MOD_LOG_MOVE_KEYS
  19          * operations.
  20          */
  21         int slot;
  22
  23         /* This is used for BTRFS_MOD_LOG_KEY* and BTRFS_MOD_LOG_ROOT_REPLACE. */
  24         u64 generation;
  25
  26         /* Those are used for op == BTRFS_MOD_LOG_KEY_{REPLACE,REMOVE}. */
  27         struct btrfs_disk_key key;
  28         u64 blockptr;
  29
  30         /* This is used for op == BTRFS_MOD_LOG_MOVE_KEYS. */
  31         struct {
  32                 int dst_slot;
  33                 int nr_items;
  34         } move;
  35
  36         /* This is used for op == BTRFS_MOD_LOG_ROOT_REPLACE. */
  37         struct tree_mod_root old_root;
  38 };
  39
  40 /*
  41  * Pull a new tree mod seq number for our operation.
  42  */
  43 static inline u64 btrfs_inc_tree_mod_seq(struct btrfs_fs_info *fs_info)
  44 {
  45         return atomic64_inc_return(&fs_info->tree_mod_seq);
  46 }
  47
  48 /*
  49  * This adds a new blocker to the tree mod log's blocker list if the @elem
  50  * passed does not already have a sequence number set. So when a caller expects
  51  * to record tree modifications, it should ensure to set elem->seq to zero
  52  * before calling btrfs_get_tree_mod_seq.
  53  * Returns a fresh, unused tree log modification sequence number, even if no new
  54  * blocker was added.
  55  */
  56 u64 btrfs_get_tree_mod_seq(struct btrfs_fs_info *fs_info,
  57                            struct btrfs_seq_list *elem)
  58 {
  59         write_lock(&fs_info->tree_mod_log_lock);
  60         if (!elem->seq) {
  61                 elem->seq = btrfs_inc_tree_mod_seq(fs_info);
  62                 list_add_tail(&elem->list, &fs_info->tree_mod_seq_list);
  63         }
  64         write_unlock(&fs_info->tree_mod_log_lock);
  65
  66         return elem->seq;
  67 }
  68
  69 void btrfs_put_tree_mod_seq(struct btrfs_fs_info *fs_info,
  70                             struct btrfs_seq_list *elem)
  71 {
  72         struct rb_root *tm_root;
  73         struct rb_node *node;
  74         struct rb_node *next;
  75         struct tree_mod_elem *tm;
  76         u64 min_seq = BTRFS_SEQ_LAST;
  77         u64 seq_putting = elem->seq;
  78
  79         if (!seq_putting)
  80                 return;
  81
  82         write_lock(&fs_info->tree_mod_log_lock);
  83         list_del(&elem->list);
  84         elem->seq = 0;
  85
  86         if (!list_empty(&fs_info->tree_mod_seq_list)) {
  87                 struct btrfs_seq_list *first;
  88
  89                 first = list_first_entry(&fs_info->tree_mod_seq_list,
  90                                          struct btrfs_seq_list, list);
  91                 if (seq_putting > first->seq) {
  92                         /*
  93                          * Blocker with lower sequence number exists, we cannot
  94                          * remove anything from the log.
  95                          */
  96                         write_unlock(&fs_info->tree_mod_log_lock);
  97                         return;
  98                 }
  99                 min_seq = first->seq;
 100         }
 101
 102         /*
 103          * Anything that's lower than the lowest existing (read: blocked)
 104          * sequence number can be removed from the tree.
 105          */
 106         tm_root = &fs_info->tree_mod_log;
 107         for (node = rb_first(tm_root); node; node = next) {
 108                 next = rb_next(node);
 109                 tm = rb_entry(node, struct tree_mod_elem, node);
 110                 if (tm->seq >= min_seq)
 111                         continue;
 112                 rb_erase(node, tm_root);
 113                 kfree(tm);
 114         }
 115         write_unlock(&fs_info->tree_mod_log_lock);
 116 }
 117
 118 /*
 119  * Key order of the log:
 120  *       node/leaf start address -> sequence
 121  *
 122  * The 'start address' is the logical address of the *new* root node for root
 123  * replace operations, or the logical address of the affected block for all
 124  * other operations.
 125  */
 126 static noinline int tree_mod_log_insert(struct btrfs_fs_info *fs_info,
 127                                         struct tree_mod_elem *tm)
 128 {
 129         struct rb_root *tm_root;
 130         struct rb_node **new;
 131         struct rb_node *parent = NULL;
 132         struct tree_mod_elem *cur;
 133
 134         lockdep_assert_held_write(&fs_info->tree_mod_log_lock);
 135
 136         tm->seq = btrfs_inc_tree_mod_seq(fs_info);
 137
 138         tm_root = &fs_info->tree_mod_log;
 139         new = &tm_root->rb_node;
 140         while (*new) {
 141                 cur = rb_entry(*new, struct tree_mod_elem, node);
 142                 parent = *new;
 143                 if (cur->logical < tm->logical)
 144                         new = &((*new)->rb_left);
 145                 else if (cur->logical > tm->logical)
 146                         new = &((*new)->rb_right);
 147                 else if (cur->seq < tm->seq)
 148                         new = &((*new)->rb_left);
 149                 else if (cur->seq > tm->seq)
 150                         new = &((*new)->rb_right);
 151                 else
 152                         return -EEXIST;
 153         }
 154
 155         rb_link_node(&tm->node, parent, new);
 156         rb_insert_color(&tm->node, tm_root);
 157         return 0;
 158 }
 159
 160 /*
 161  * Determines if logging can be omitted. Returns 1 if it can. Otherwise, it
 162  * returns zero with the tree_mod_log_lock acquired. The caller must hold
 163  * this until all tree mod log insertions are recorded in the rb tree and then
 164  * write unlock fs_info::tree_mod_log_lock.
 165  */
 166 static inline int tree_mod_dont_log(struct btrfs_fs_info *fs_info,
 167                                     struct extent_buffer *eb)
 168 {
 169         smp_mb();
 170         if (list_empty(&(fs_info)->tree_mod_seq_list))
 171                 return 1;
 172         if (eb && btrfs_header_level(eb) == 0)
 173                 return 1;
 174
 175         write_lock(&fs_info->tree_mod_log_lock);
 176         if (list_empty(&(fs_info)->tree_mod_seq_list)) {
 177                 write_unlock(&fs_info->tree_mod_log_lock);
 178                 return 1;
 179         }
 180
 181         return 0;
 182 }
 183
 184 /* Similar to tree_mod_dont_log, but doesn't acquire any locks. */
 185 static inline int tree_mod_need_log(const struct btrfs_fs_info *fs_info,
 186                                     struct extent_buffer *eb)
 187 {
 188         smp_mb();
 189         if (list_empty(&(fs_info)->tree_mod_seq_list))
 190                 return 0;
 191         if (eb && btrfs_header_level(eb) == 0)
 192                 return 0;
 193
 194         return 1;
 195 }
 196
 197 static struct tree_mod_elem *alloc_tree_mod_elem(struct extent_buffer *eb,
 198                                                  int slot,
 199                                                  enum btrfs_mod_log_op op,
 200                                                  gfp_t flags)
 201 {
 202         struct tree_mod_elem *tm;
 203
 204         tm = kzalloc(sizeof(*tm), flags);
 205         if (!tm)
 206                 return NULL;
 207
 208         tm->logical = eb->start;
 209         if (op != BTRFS_MOD_LOG_KEY_ADD) {
 210                 btrfs_node_key(eb, &tm->key, slot);
 211                 tm->blockptr = btrfs_node_blockptr(eb, slot);
 212         }
 213         tm->op = op;
 214         tm->slot = slot;
 215         tm->generation = btrfs_node_ptr_generation(eb, slot);
 216         RB_CLEAR_NODE(&tm->node);
 217
 218         return tm;
 219 }
 220
 221 int btrfs_tree_mod_log_insert_key(struct extent_buffer *eb, int slot,
 222                                   enum btrfs_mod_log_op op, gfp_t flags)
 223 {
 224         struct tree_mod_elem *tm;
 225         int ret;
 226
 227         if (!tree_mod_need_log(eb->fs_info, eb))
 228                 return 0;
 229
 230         tm = alloc_tree_mod_elem(eb, slot, op, flags);
 231         if (!tm)
 232                 return -ENOMEM;
 233
 234         if (tree_mod_dont_log(eb->fs_info, eb)) {
 235                 kfree(tm);
 236                 return 0;
 237         }
 238
 239         ret = tree_mod_log_insert(eb->fs_info, tm);
 240         write_unlock(&eb->fs_info->tree_mod_log_lock);
 241         if (ret)
 242                 kfree(tm);
 243
 244         return ret;
 245 }
 246
 247 int btrfs_tree_mod_log_insert_move(struct extent_buffer *eb,
 248                                    int dst_slot, int src_slot,
 249                                    int nr_items)
 250 {
 251         struct tree_mod_elem *tm = NULL;
 252         struct tree_mod_elem **tm_list = NULL;
 253         int ret = 0;
 254         int i;
 255         int locked = 0;
 256
 257         if (!tree_mod_need_log(eb->fs_info, eb))
 258                 return 0;
 259
 260         tm_list = kcalloc(nr_items, sizeof(struct tree_mod_elem *), GFP_NOFS);
 261         if (!tm_list)
 262                 return -ENOMEM;
 263
 264         tm = kzalloc(sizeof(*tm), GFP_NOFS);
 265         if (!tm) {
 266                 ret = -ENOMEM;
 267                 goto free_tms;
 268         }
 269
 270         tm->logical = eb->start;
 271         tm->slot = src_slot;
 272         tm->move.dst_slot = dst_slot;
 273         tm->move.nr_items = nr_items;
 274         tm->op = BTRFS_MOD_LOG_MOVE_KEYS;
 275
 276         for (i = 0; i + dst_slot < src_slot && i < nr_items; i++) {
 277                 tm_list[i] = alloc_tree_mod_elem(eb, i + dst_slot,
 278                                 BTRFS_MOD_LOG_KEY_REMOVE_WHILE_MOVING, GFP_NOFS);
 279                 if (!tm_list[i]) {
 280                         ret = -ENOMEM;
 281                         goto free_tms;
 282                 }
 283         }
 284
 285         if (tree_mod_dont_log(eb->fs_info, eb))
 286                 goto free_tms;
 287         locked = 1;
 288
 289         /*
 290          * When we override something during the move, we log these removals.
 291          * This can only happen when we move towards the beginning of the
 292          * buffer, i.e. dst_slot < src_slot.
 293          */
 294         for (i = 0; i + dst_slot < src_slot && i < nr_items; i++) {
 295                 ret = tree_mod_log_insert(eb->fs_info, tm_list[i]);
 296                 if (ret)
 297                         goto free_tms;
 298         }
 299
 300         ret = tree_mod_log_insert(eb->fs_info, tm);
 301         if (ret)
 302                 goto free_tms;
 303         write_unlock(&eb->fs_info->tree_mod_log_lock);
 304         kfree(tm_list);
 305
 306         return 0;
 307
 308 free_tms:
 309         for (i = 0; i < nr_items; i++) {
 310                 if (tm_list[i] && !RB_EMPTY_NODE(&tm_list[i]->node))
 311                         rb_erase(&tm_list[i]->node, &eb->fs_info->tree_mod_log);
 312                 kfree(tm_list[i]);
 313         }
 314         if (locked)
 315                 write_unlock(&eb->fs_info->tree_mod_log_lock);
 316         kfree(tm_list);
 317         kfree(tm);
 318
 319         return ret;
 320 }
 321
 322 static inline int tree_mod_log_free_eb(struct btrfs_fs_info *fs_info,
 323                                        struct tree_mod_elem **tm_list,
 324                                        int nritems)
 325 {
 326         int i, j;
 327         int ret;
 328
 329         for (i = nritems - 1; i >= 0; i--) {
 330                 ret = tree_mod_log_insert(fs_info, tm_list[i]);
 331                 if (ret) {
 332                         for (j = nritems - 1; j > i; j--)
 333                                 rb_erase(&tm_list[j]->node,
 334                                          &fs_info->tree_mod_log);
 335                         return ret;
 336                 }
 337         }
 338
 339         return 0;
 340 }
 341
 342 int btrfs_tree_mod_log_insert_root(struct extent_buffer *old_root,
 343                                    struct extent_buffer *new_root,
 344                                    int log_removal)
 345 {
 346         struct btrfs_fs_info *fs_info = old_root->fs_info;
 347         struct tree_mod_elem *tm = NULL;
 348         struct tree_mod_elem **tm_list = NULL;
 349         int nritems = 0;
 350         int ret = 0;
 351         int i;
 352
 353         if (!tree_mod_need_log(fs_info, NULL))
 354                 return 0;
 355
 356         if (log_removal && btrfs_header_level(old_root) > 0) {
 357                 nritems = btrfs_header_nritems(old_root);
 358                 tm_list = kcalloc(nritems, sizeof(struct tree_mod_elem *),
 359                                   GFP_NOFS);
 360                 if (!tm_list) {
 361                         ret = -ENOMEM;
 362                         goto free_tms;
 363                 }
 364                 for (i = 0; i < nritems; i++) {
 365                         tm_list[i] = alloc_tree_mod_elem(old_root, i,
 366                             BTRFS_MOD_LOG_KEY_REMOVE_WHILE_FREEING, GFP_NOFS);
 367                         if (!tm_list[i]) {
 368                                 ret = -ENOMEM;
 369                                 goto free_tms;
 370                         }
 371                 }
 372         }
 373
 374         tm = kzalloc(sizeof(*tm), GFP_NOFS);
 375         if (!tm) {
 376                 ret = -ENOMEM;
 377                 goto free_tms;
 378         }
 379
 380         tm->logical = new_root->start;
 381         tm->old_root.logical = old_root->start;
 382         tm->old_root.level = btrfs_header_level(old_root);
 383         tm->generation = btrfs_header_generation(old_root);
 384         tm->op = BTRFS_MOD_LOG_ROOT_REPLACE;
 385
 386         if (tree_mod_dont_log(fs_info, NULL))
 387                 goto free_tms;
 388
 389         if (tm_list)
 390                 ret = tree_mod_log_free_eb(fs_info, tm_list, nritems);
 391         if (!ret)
 392                 ret = tree_mod_log_insert(fs_info, tm);
 393
 394         write_unlock(&fs_info->tree_mod_log_lock);
 395         if (ret)
 396                 goto free_tms;
 397         kfree(tm_list);
 398
 399         return ret;
 400
 401 free_tms:
 402         if (tm_list) {
 403                 for (i = 0; i < nritems; i++)
 404                         kfree(tm_list[i]);
 405                 kfree(tm_list);
 406         }
 407         kfree(tm);
 408
 409         return ret;
 410 }
 411
 412 static struct tree_mod_elem *__tree_mod_log_search(struct btrfs_fs_info *fs_info,
 413                                                    u64 start, u64 min_seq,
 414                                                    int smallest)
 415 {
 416         struct rb_root *tm_root;
 417         struct rb_node *node;
 418         struct tree_mod_elem *cur = NULL;
 419         struct tree_mod_elem *found = NULL;
 420
 421         read_lock(&fs_info->tree_mod_log_lock);
 422         tm_root = &fs_info->tree_mod_log;
 423         node = tm_root->rb_node;
 424         while (node) {
 425                 cur = rb_entry(node, struct tree_mod_elem, node);
 426                 if (cur->logical < start) {
 427                         node = node->rb_left;
 428                 } else if (cur->logical > start) {
 429                         node = node->rb_right;
 430                 } else if (cur->seq < min_seq) {
 431                         node = node->rb_left;
 432                 } else if (!smallest) {
 433                         /* We want the node with the highest seq */
 434                         if (found)
 435                                 BUG_ON(found->seq > cur->seq);
 436                         found = cur;
 437                         node = node->rb_left;
 438                 } else if (cur->seq > min_seq) {
 439                         /* We want the node with the smallest seq */
 440                         if (found)
 441                                 BUG_ON(found->seq < cur->seq);
 442                         found = cur;
 443                         node = node->rb_right;
 444                 } else {
 445                         found = cur;
 446                         break;
 447                 }
 448         }
 449         read_unlock(&fs_info->tree_mod_log_lock);
 450
 451         return found;
 452 }
 453
 454 /*
 455  * This returns the element from the log with the smallest time sequence
 456  * value that's in the log (the oldest log item). Any element with a time
 457  * sequence lower than min_seq will be ignored.
 458  */
 459 static struct tree_mod_elem *tree_mod_log_search_oldest(struct btrfs_fs_info *fs_info,
 460                                                         u64 start, u64 min_seq)
 461 {
 462         return __tree_mod_log_search(fs_info, start, min_seq, 1);
 463 }
 464
 465 /*
 466  * This returns the element from the log with the largest time sequence
 467  * value that's in the log (the most recent log item). Any element with
 468  * a time sequence lower than min_seq will be ignored.
 469  */
 470 static struct tree_mod_elem *tree_mod_log_search(struct btrfs_fs_info *fs_info,
 471                                                  u64 start, u64 min_seq)
 472 {
 473         return __tree_mod_log_search(fs_info, start, min_seq, 0);
 474 }
 475
 476 int btrfs_tree_mod_log_eb_copy(struct extent_buffer *dst,
 477                                struct extent_buffer *src,
 478                                unsigned long dst_offset,
 479                                unsigned long src_offset,
 480                                int nr_items)
 481 {
 482         struct btrfs_fs_info *fs_info = dst->fs_info;
 483         int ret = 0;
 484         struct tree_mod_elem **tm_list = NULL;
 485         struct tree_mod_elem **tm_list_add, **tm_list_rem;
 486         int i;
 487         int locked = 0;
 488
 489         if (!tree_mod_need_log(fs_info, NULL))
 490                 return 0;
 491
 492         if (btrfs_header_level(dst) == 0 && btrfs_header_level(src) == 0)
 493                 return 0;
 494
 495         tm_list = kcalloc(nr_items * 2, sizeof(struct tree_mod_elem *),
 496                           GFP_NOFS);
 497         if (!tm_list)
 498                 return -ENOMEM;
 499
 500         tm_list_add = tm_list;
 501         tm_list_rem = tm_list + nr_items;
 502         for (i = 0; i < nr_items; i++) {
 503                 tm_list_rem[i] = alloc_tree_mod_elem(src, i + src_offset,
 504                     BTRFS_MOD_LOG_KEY_REMOVE, GFP_NOFS);
 505                 if (!tm_list_rem[i]) {
 506                         ret = -ENOMEM;
 507                         goto free_tms;
 508                 }
 509
 510                 tm_list_add[i] = alloc_tree_mod_elem(dst, i + dst_offset,
 511                                                 BTRFS_MOD_LOG_KEY_ADD, GFP_NOFS);
 512                 if (!tm_list_add[i]) {
 513                         ret = -ENOMEM;
 514                         goto free_tms;
 515                 }
 516         }
 517
 518         if (tree_mod_dont_log(fs_info, NULL))
 519                 goto free_tms;
 520         locked = 1;
 521
 522         for (i = 0; i < nr_items; i++) {
 523                 ret = tree_mod_log_insert(fs_info, tm_list_rem[i]);
 524                 if (ret)
 525                         goto free_tms;
 526                 ret = tree_mod_log_insert(fs_info, tm_list_add[i]);
 527                 if (ret)
 528                         goto free_tms;
 529         }
 530
 531         write_unlock(&fs_info->tree_mod_log_lock);
 532         kfree(tm_list);
 533
 534         return 0;
 535
 536 free_tms:
 537         for (i = 0; i < nr_items * 2; i++) {
 538                 if (tm_list[i] && !RB_EMPTY_NODE(&tm_list[i]->node))
 539                         rb_erase(&tm_list[i]->node, &fs_info->tree_mod_log);
 540                 kfree(tm_list[i]);
 541         }
 542         if (locked)
 543                 write_unlock(&fs_info->tree_mod_log_lock);
 544         kfree(tm_list);
 545
 546         return ret;
 547 }
 548
 549 int btrfs_tree_mod_log_free_eb(struct extent_buffer *eb)
 550 {
 551         struct tree_mod_elem **tm_list = NULL;
 552         int nritems = 0;
 553         int i;
 554         int ret = 0;
 555
 556         if (btrfs_header_level(eb) == 0)
 557                 return 0;
 558
 559         if (!tree_mod_need_log(eb->fs_info, NULL))
 560                 return 0;
 561
 562         nritems = btrfs_header_nritems(eb);
 563         tm_list = kcalloc(nritems, sizeof(struct tree_mod_elem *), GFP_NOFS);
 564         if (!tm_list)
 565                 return -ENOMEM;
 566
 567         for (i = 0; i < nritems; i++) {
 568                 tm_list[i] = alloc_tree_mod_elem(eb, i,
 569                     BTRFS_MOD_LOG_KEY_REMOVE_WHILE_FREEING, GFP_NOFS);
 570                 if (!tm_list[i]) {
 571                         ret = -ENOMEM;
 572                         goto free_tms;
 573                 }
 574         }
 575
 576         if (tree_mod_dont_log(eb->fs_info, eb))
 577                 goto free_tms;
 578
 579         ret = tree_mod_log_free_eb(eb->fs_info, tm_list, nritems);
 580         write_unlock(&eb->fs_info->tree_mod_log_lock);
 581         if (ret)
 582                 goto free_tms;
 583         kfree(tm_list);
 584
 585         return 0;
 586
 587 free_tms:
 588         for (i = 0; i < nritems; i++)
 589                 kfree(tm_list[i]);
 590         kfree(tm_list);
 591
 592         return ret;
 593 }
 594
 595 /*
 596  * Returns the logical address of the oldest predecessor of the given root.
 597  * Entries older than time_seq are ignored.
 598  */
 599 static struct tree_mod_elem *tree_mod_log_oldest_root(struct extent_buffer *eb_root,
 600                                                       u64 time_seq)
 601 {
 602         struct tree_mod_elem *tm;
 603         struct tree_mod_elem *found = NULL;
 604         u64 root_logical = eb_root->start;
 605         int looped = 0;
 606
 607         if (!time_seq)
 608                 return NULL;
 609
 610         /*
 611          * The very last operation that's logged for a root is the replacement
 612          * operation (if it is replaced at all). This has the logical address
 613          * of the *new* root, making it the very first operation that's logged
 614          * for this root.
 615          */
 616         while (1) {
 617                 tm = tree_mod_log_search_oldest(eb_root->fs_info, root_logical,
 618                                                 time_seq);
 619                 if (!looped && !tm)
 620                         return NULL;
 621                 /*
 622                  * If there are no tree operation for the oldest root, we simply
 623                  * return it. This should only happen if that (old) root is at
 624                  * level 0.
 625                  */
 626                 if (!tm)
 627                         break;
 628
 629                 /*
 630                  * If there's an operation that's not a root replacement, we
 631                  * found the oldest version of our root. Normally, we'll find a
 632                  * BTRFS_MOD_LOG_KEY_REMOVE_WHILE_FREEING operation here.
 633                  */
 634                 if (tm->op != BTRFS_MOD_LOG_ROOT_REPLACE)
 635                         break;
 636
 637                 found = tm;
 638                 root_logical = tm->old_root.logical;
 639                 looped = 1;
 640         }
 641
 642         /* If there's no old root to return, return what we found instead */
 643         if (!found)
 644                 found = tm;
 645
 646         return found;
 647 }
 648
 649
 650 /*
 651  * tm is a pointer to the first operation to rewind within eb. Then, all
 652  * previous operations will be rewound (until we reach something older than
 653  * time_seq).
 654  */
 655 static void tree_mod_log_rewind(struct btrfs_fs_info *fs_info,
 656                                 struct extent_buffer *eb,
 657                                 u64 time_seq,
 658                                 struct tree_mod_elem *first_tm)
 659 {
 660         u32 n;
 661         struct rb_node *next;
 662         struct tree_mod_elem *tm = first_tm;
 663         unsigned long o_dst;
 664         unsigned long o_src;
 665         unsigned long p_size = sizeof(struct btrfs_key_ptr);
 666
 667         n = btrfs_header_nritems(eb);
 668         read_lock(&fs_info->tree_mod_log_lock);
 669         while (tm && tm->seq >= time_seq) {
 670                 /*
 671                  * All the operations are recorded with the operator used for
 672                  * the modification. As we're going backwards, we do the
 673                  * opposite of each operation here.
 674                  */
 675                 switch (tm->op) {
 676                 case BTRFS_MOD_LOG_KEY_REMOVE_WHILE_FREEING:
 677                         BUG_ON(tm->slot < n);
 678                         fallthrough;
 679                 case BTRFS_MOD_LOG_KEY_REMOVE_WHILE_MOVING:
 680                 case BTRFS_MOD_LOG_KEY_REMOVE:
 681                         btrfs_set_node_key(eb, &tm->key, tm->slot);
 682                         btrfs_set_node_blockptr(eb, tm->slot, tm->blockptr);
 683                         btrfs_set_node_ptr_generation(eb, tm->slot,
 684                                                       tm->generation);
 685                         n++;
 686                         break;
 687                 case BTRFS_MOD_LOG_KEY_REPLACE:
 688                         BUG_ON(tm->slot >= n);
 689                         btrfs_set_node_key(eb, &tm->key, tm->slot);
 690                         btrfs_set_node_blockptr(eb, tm->slot, tm->blockptr);
 691                         btrfs_set_node_ptr_generation(eb, tm->slot,
 692                                                       tm->generation);
 693                         break;
 694                 case BTRFS_MOD_LOG_KEY_ADD:
 695                         /* if a move operation is needed it's in the log */
 696                         n--;
 697                         break;
 698                 case BTRFS_MOD_LOG_MOVE_KEYS:
 699                         o_dst = btrfs_node_key_ptr_offset(tm->slot);
 700                         o_src = btrfs_node_key_ptr_offset(tm->move.dst_slot);
 701                         memmove_extent_buffer(eb, o_dst, o_src,
 702                                               tm->move.nr_items * p_size);
 703                         break;
 704                 case BTRFS_MOD_LOG_ROOT_REPLACE:
 705                         /*
 706                          * This operation is special. For roots, this must be
 707                          * handled explicitly before rewinding.
 708                          * For non-roots, this operation may exist if the node
 709                          * was a root: root A -> child B; then A gets empty and
 710                          * B is promoted to the new root. In the mod log, we'll
 711                          * have a root-replace operation for B, a tree block
 712                          * that is no root. We simply ignore that operation.
 713                          */
 714                         break;
 715                 }
 716                 next = rb_next(&tm->node);
 717                 if (!next)
 718                         break;
 719                 tm = rb_entry(next, struct tree_mod_elem, node);
 720                 if (tm->logical != first_tm->logical)
 721                         break;
 722         }
 723         read_unlock(&fs_info->tree_mod_log_lock);
 724         btrfs_set_header_nritems(eb, n);
 725 }
 726
 727 /*
 728  * Called with eb read locked. If the buffer cannot be rewound, the same buffer
 729  * is returned. If rewind operations happen, a fresh buffer is returned. The
 730  * returned buffer is always read-locked. If the returned buffer is not the
 731  * input buffer, the lock on the input buffer is released and the input buffer
 732  * is freed (its refcount is decremented).
 733  */
 734 struct extent_buffer *btrfs_tree_mod_log_rewind(struct btrfs_fs_info *fs_info,
 735                                                 struct btrfs_path *path,
 736                                                 struct extent_buffer *eb,
 737                                                 u64 time_seq)
 738 {
 739         struct extent_buffer *eb_rewin;
 740         struct tree_mod_elem *tm;
 741
 742         if (!time_seq)
 743                 return eb;
 744
 745         if (btrfs_header_level(eb) == 0)
 746                 return eb;
 747
 748         tm = tree_mod_log_search(fs_info, eb->start, time_seq);
 749         if (!tm)
 750                 return eb;
 751
 752         if (tm->op == BTRFS_MOD_LOG_KEY_REMOVE_WHILE_FREEING) {
 753                 BUG_ON(tm->slot != 0);
 754                 eb_rewin = alloc_dummy_extent_buffer(fs_info, eb->start);
 755                 if (!eb_rewin) {
 756                         btrfs_tree_read_unlock(eb);
 757                         free_extent_buffer(eb);
 758                         return NULL;
 759                 }
 760                 btrfs_set_header_bytenr(eb_rewin, eb->start);
 761                 btrfs_set_header_backref_rev(eb_rewin,
 762                                              btrfs_header_backref_rev(eb));
 763                 btrfs_set_header_owner(eb_rewin, btrfs_header_owner(eb));
 764                 btrfs_set_header_level(eb_rewin, btrfs_header_level(eb));
 765         } else {
 766                 eb_rewin = btrfs_clone_extent_buffer(eb);
 767                 if (!eb_rewin) {
 768                         btrfs_tree_read_unlock(eb);
 769                         free_extent_buffer(eb);
 770                         return NULL;
 771                 }
 772         }
 773
 774         btrfs_tree_read_unlock(eb);
 775         free_extent_buffer(eb);
 776
 777         btrfs_set_buffer_lockdep_class(btrfs_header_owner(eb_rewin),
 778                                        eb_rewin, btrfs_header_level(eb_rewin));
 779         btrfs_tree_read_lock(eb_rewin);
 780         tree_mod_log_rewind(fs_info, eb_rewin, time_seq, tm);
 781         WARN_ON(btrfs_header_nritems(eb_rewin) >
 782                 BTRFS_NODEPTRS_PER_BLOCK(fs_info));
 783
 784         return eb_rewin;
 785 }
 786
 787 /*
 788  * Rewind the state of @root's root node to the given @time_seq value.
 789  * If there are no changes, the current root->root_node is returned. If anything
 790  * changed in between, there's a fresh buffer allocated on which the rewind
 791  * operations are done. In any case, the returned buffer is read locked.
 792  * Returns NULL on error (with no locks held).
 793  */
 794 struct extent_buffer *btrfs_get_old_root(struct btrfs_root *root, u64 time_seq)
 795 {
 796         struct btrfs_fs_info *fs_info = root->fs_info;
 797         struct tree_mod_elem *tm;
 798         struct extent_buffer *eb = NULL;
 799         struct extent_buffer *eb_root;
 800         u64 eb_root_owner = 0;
 801         struct extent_buffer *old;
 802         struct tree_mod_root *old_root = NULL;
 803         u64 old_generation = 0;
 804         u64 logical;
 805         int level;
 806
 807         eb_root = btrfs_read_lock_root_node(root);
 808         tm = tree_mod_log_oldest_root(eb_root, time_seq);
 809         if (!tm)
 810                 return eb_root;
 811
 812         if (tm->op == BTRFS_MOD_LOG_ROOT_REPLACE) {
 813                 old_root = &tm->old_root;
 814                 old_generation = tm->generation;
 815                 logical = old_root->logical;
 816                 level = old_root->level;
 817         } else {
 818                 logical = eb_root->start;
 819                 level = btrfs_header_level(eb_root);
 820         }
 821
 822         tm = tree_mod_log_search(fs_info, logical, time_seq);
 823         if (old_root && tm && tm->op != BTRFS_MOD_LOG_KEY_REMOVE_WHILE_FREEING) {
 824                 btrfs_tree_read_unlock(eb_root);
 825                 free_extent_buffer(eb_root);
 826                 old = read_tree_block(fs_info, logical, root->root_key.objectid,
 827                                       0, level, NULL);
 828                 if (WARN_ON(IS_ERR(old) || !extent_buffer_uptodate(old))) {
 829                         if (!IS_ERR(old))
 830                                 free_extent_buffer(old);
 831                         btrfs_warn(fs_info,
 832                                    "failed to read tree block %llu from get_old_root",
 833                                    logical);
 834                 } else {
 835                         btrfs_tree_read_lock(old);
 836                         eb = btrfs_clone_extent_buffer(old);
 837                         btrfs_tree_read_unlock(old);
 838                         free_extent_buffer(old);
 839                 }
 840         } else if (old_root) {
 841                 eb_root_owner = btrfs_header_owner(eb_root);
 842                 btrfs_tree_read_unlock(eb_root);
 843                 free_extent_buffer(eb_root);
 844                 eb = alloc_dummy_extent_buffer(fs_info, logical);
 845         } else {
 846                 eb = btrfs_clone_extent_buffer(eb_root);
 847                 btrfs_tree_read_unlock(eb_root);
 848                 free_extent_buffer(eb_root);
 849         }
 850
 851         if (!eb)
 852                 return NULL;
 853         if (old_root) {
 854                 btrfs_set_header_bytenr(eb, eb->start);
 855                 btrfs_set_header_backref_rev(eb, BTRFS_MIXED_BACKREF_REV);
 856                 btrfs_set_header_owner(eb, eb_root_owner);
 857                 btrfs_set_header_level(eb, old_root->level);
 858                 btrfs_set_header_generation(eb, old_generation);
 859         }
 860         btrfs_set_buffer_lockdep_class(btrfs_header_owner(eb), eb,
 861                                        btrfs_header_level(eb));
 862         btrfs_tree_read_lock(eb);
 863         if (tm)
 864                 tree_mod_log_rewind(fs_info, eb, time_seq, tm);
 865         else
 866                 WARN_ON(btrfs_header_level(eb) != 0);
 867         WARN_ON(btrfs_header_nritems(eb) > BTRFS_NODEPTRS_PER_BLOCK(fs_info));
 868
 869         return eb;
 870 }
 871
 872 int btrfs_old_root_level(struct btrfs_root *root, u64 time_seq)
 873 {
 874         struct tree_mod_elem *tm;
 875         int level;
 876         struct extent_buffer *eb_root = btrfs_root_node(root);
 877
 878         tm = tree_mod_log_oldest_root(eb_root, time_seq);
 879         if (tm && tm->op == BTRFS_MOD_LOG_ROOT_REPLACE)
 880                 level = tm->old_root.level;
 881         else
 882                 level = btrfs_header_level(eb_root);
 883
 884         free_extent_buffer(eb_root);
 885
 886         return level;
 887 }
 888