Add linux-next specific files for 20110716
[linux-2.6/next.git] / fs / btrfs / ctree.c
blob2e667868e0d2b8b75649572d10602ac56679a0d0
1 /*
2 * Copyright (C) 2007,2008 Oracle. All rights reserved.
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU General Public
6 * License v2 as published by the Free Software Foundation.
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * General Public License for more details.
13 * You should have received a copy of the GNU General Public
14 * License along with this program; if not, write to the
15 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
16 * Boston, MA 021110-1307, USA.
19 #include <linux/sched.h>
20 #include <linux/slab.h>
21 #include "ctree.h"
22 #include "disk-io.h"
23 #include "transaction.h"
24 #include "print-tree.h"
25 #include "locking.h"
27 static int split_node(struct btrfs_trans_handle *trans, struct btrfs_root
28 *root, struct btrfs_path *path, int level);
29 static int split_leaf(struct btrfs_trans_handle *trans, struct btrfs_root
30 *root, struct btrfs_key *ins_key,
31 struct btrfs_path *path, int data_size, int extend);
32 static int push_node_left(struct btrfs_trans_handle *trans,
33 struct btrfs_root *root, struct extent_buffer *dst,
34 struct extent_buffer *src, int empty);
35 static int balance_node_right(struct btrfs_trans_handle *trans,
36 struct btrfs_root *root,
37 struct extent_buffer *dst_buf,
38 struct extent_buffer *src_buf);
39 static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
40 struct btrfs_path *path, int level, int slot);
42 struct btrfs_path *btrfs_alloc_path(void)
44 struct btrfs_path *path;
45 path = kmem_cache_zalloc(btrfs_path_cachep, GFP_NOFS);
46 return path;
50 * set all locked nodes in the path to blocking locks. This should
51 * be done before scheduling
53 noinline void btrfs_set_path_blocking(struct btrfs_path *p)
55 int i;
56 for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
57 if (p->nodes[i] && p->locks[i])
58 btrfs_set_lock_blocking(p->nodes[i]);
63 * reset all the locked nodes in the patch to spinning locks.
65 * held is used to keep lockdep happy, when lockdep is enabled
66 * we set held to a blocking lock before we go around and
67 * retake all the spinlocks in the path. You can safely use NULL
68 * for held
70 noinline void btrfs_clear_path_blocking(struct btrfs_path *p,
71 struct extent_buffer *held)
73 int i;
75 #ifdef CONFIG_DEBUG_LOCK_ALLOC
76 /* lockdep really cares that we take all of these spinlocks
77 * in the right order. If any of the locks in the path are not
78 * currently blocking, it is going to complain. So, make really
79 * really sure by forcing the path to blocking before we clear
80 * the path blocking.
82 if (held)
83 btrfs_set_lock_blocking(held);
84 btrfs_set_path_blocking(p);
85 #endif
87 for (i = BTRFS_MAX_LEVEL - 1; i >= 0; i--) {
88 if (p->nodes[i] && p->locks[i])
89 btrfs_clear_lock_blocking(p->nodes[i]);
92 #ifdef CONFIG_DEBUG_LOCK_ALLOC
93 if (held)
94 btrfs_clear_lock_blocking(held);
95 #endif
98 /* this also releases the path */
99 void btrfs_free_path(struct btrfs_path *p)
101 if (!p)
102 return;
103 btrfs_release_path(p);
104 kmem_cache_free(btrfs_path_cachep, p);
108 * path release drops references on the extent buffers in the path
109 * and it drops any locks held by this path
111 * It is safe to call this on paths that no locks or extent buffers held.
113 noinline void btrfs_release_path(struct btrfs_path *p)
115 int i;
117 for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
118 p->slots[i] = 0;
119 if (!p->nodes[i])
120 continue;
121 if (p->locks[i]) {
122 btrfs_tree_unlock(p->nodes[i]);
123 p->locks[i] = 0;
125 free_extent_buffer(p->nodes[i]);
126 p->nodes[i] = NULL;
131 * safely gets a reference on the root node of a tree. A lock
132 * is not taken, so a concurrent writer may put a different node
133 * at the root of the tree. See btrfs_lock_root_node for the
134 * looping required.
136 * The extent buffer returned by this has a reference taken, so
137 * it won't disappear. It may stop being the root of the tree
138 * at any time because there are no locks held.
140 struct extent_buffer *btrfs_root_node(struct btrfs_root *root)
142 struct extent_buffer *eb;
144 rcu_read_lock();
145 eb = rcu_dereference(root->node);
146 extent_buffer_get(eb);
147 rcu_read_unlock();
148 return eb;
151 /* loop around taking references on and locking the root node of the
152 * tree until you end up with a lock on the root. A locked buffer
153 * is returned, with a reference held.
155 struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root)
157 struct extent_buffer *eb;
159 while (1) {
160 eb = btrfs_root_node(root);
161 btrfs_tree_lock(eb);
162 if (eb == root->node)
163 break;
164 btrfs_tree_unlock(eb);
165 free_extent_buffer(eb);
167 return eb;
170 /* cowonly root (everything not a reference counted cow subvolume), just get
171 * put onto a simple dirty list. transaction.c walks this to make sure they
172 * get properly updated on disk.
174 static void add_root_to_dirty_list(struct btrfs_root *root)
176 if (root->track_dirty && list_empty(&root->dirty_list)) {
177 list_add(&root->dirty_list,
178 &root->fs_info->dirty_cowonly_roots);
183 * used by snapshot creation to make a copy of a root for a tree with
184 * a given objectid. The buffer with the new root node is returned in
185 * cow_ret, and this func returns zero on success or a negative error code.
187 int btrfs_copy_root(struct btrfs_trans_handle *trans,
188 struct btrfs_root *root,
189 struct extent_buffer *buf,
190 struct extent_buffer **cow_ret, u64 new_root_objectid)
192 struct extent_buffer *cow;
193 int ret = 0;
194 int level;
195 struct btrfs_disk_key disk_key;
197 WARN_ON(root->ref_cows && trans->transid !=
198 root->fs_info->running_transaction->transid);
199 WARN_ON(root->ref_cows && trans->transid != root->last_trans);
201 level = btrfs_header_level(buf);
202 if (level == 0)
203 btrfs_item_key(buf, &disk_key, 0);
204 else
205 btrfs_node_key(buf, &disk_key, 0);
207 cow = btrfs_alloc_free_block(trans, root, buf->len, 0,
208 new_root_objectid, &disk_key, level,
209 buf->start, 0);
210 if (IS_ERR(cow))
211 return PTR_ERR(cow);
213 copy_extent_buffer(cow, buf, 0, 0, cow->len);
214 btrfs_set_header_bytenr(cow, cow->start);
215 btrfs_set_header_generation(cow, trans->transid);
216 btrfs_set_header_backref_rev(cow, BTRFS_MIXED_BACKREF_REV);
217 btrfs_clear_header_flag(cow, BTRFS_HEADER_FLAG_WRITTEN |
218 BTRFS_HEADER_FLAG_RELOC);
219 if (new_root_objectid == BTRFS_TREE_RELOC_OBJECTID)
220 btrfs_set_header_flag(cow, BTRFS_HEADER_FLAG_RELOC);
221 else
222 btrfs_set_header_owner(cow, new_root_objectid);
224 write_extent_buffer(cow, root->fs_info->fsid,
225 (unsigned long)btrfs_header_fsid(cow),
226 BTRFS_FSID_SIZE);
228 WARN_ON(btrfs_header_generation(buf) > trans->transid);
229 if (new_root_objectid == BTRFS_TREE_RELOC_OBJECTID)
230 ret = btrfs_inc_ref(trans, root, cow, 1);
231 else
232 ret = btrfs_inc_ref(trans, root, cow, 0);
234 if (ret)
235 return ret;
237 btrfs_mark_buffer_dirty(cow);
238 *cow_ret = cow;
239 return 0;
243 * check if the tree block can be shared by multiple trees
245 int btrfs_block_can_be_shared(struct btrfs_root *root,
246 struct extent_buffer *buf)
249 * Tree blocks not in refernece counted trees and tree roots
250 * are never shared. If a block was allocated after the last
251 * snapshot and the block was not allocated by tree relocation,
252 * we know the block is not shared.
254 if (root->ref_cows &&
255 buf != root->node && buf != root->commit_root &&
256 (btrfs_header_generation(buf) <=
257 btrfs_root_last_snapshot(&root->root_item) ||
258 btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC)))
259 return 1;
260 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
261 if (root->ref_cows &&
262 btrfs_header_backref_rev(buf) < BTRFS_MIXED_BACKREF_REV)
263 return 1;
264 #endif
265 return 0;
268 static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
269 struct btrfs_root *root,
270 struct extent_buffer *buf,
271 struct extent_buffer *cow,
272 int *last_ref)
274 u64 refs;
275 u64 owner;
276 u64 flags;
277 u64 new_flags = 0;
278 int ret;
281 * Backrefs update rules:
283 * Always use full backrefs for extent pointers in tree block
284 * allocated by tree relocation.
286 * If a shared tree block is no longer referenced by its owner
287 * tree (btrfs_header_owner(buf) == root->root_key.objectid),
288 * use full backrefs for extent pointers in tree block.
290 * If a tree block is been relocating
291 * (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID),
292 * use full backrefs for extent pointers in tree block.
293 * The reason for this is some operations (such as drop tree)
294 * are only allowed for blocks use full backrefs.
297 if (btrfs_block_can_be_shared(root, buf)) {
298 ret = btrfs_lookup_extent_info(trans, root, buf->start,
299 buf->len, &refs, &flags);
300 BUG_ON(ret);
301 BUG_ON(refs == 0);
302 } else {
303 refs = 1;
304 if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID ||
305 btrfs_header_backref_rev(buf) < BTRFS_MIXED_BACKREF_REV)
306 flags = BTRFS_BLOCK_FLAG_FULL_BACKREF;
307 else
308 flags = 0;
311 owner = btrfs_header_owner(buf);
312 BUG_ON(owner == BTRFS_TREE_RELOC_OBJECTID &&
313 !(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF));
315 if (refs > 1) {
316 if ((owner == root->root_key.objectid ||
317 root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) &&
318 !(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)) {
319 ret = btrfs_inc_ref(trans, root, buf, 1);
320 BUG_ON(ret);
322 if (root->root_key.objectid ==
323 BTRFS_TREE_RELOC_OBJECTID) {
324 ret = btrfs_dec_ref(trans, root, buf, 0);
325 BUG_ON(ret);
326 ret = btrfs_inc_ref(trans, root, cow, 1);
327 BUG_ON(ret);
329 new_flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF;
330 } else {
332 if (root->root_key.objectid ==
333 BTRFS_TREE_RELOC_OBJECTID)
334 ret = btrfs_inc_ref(trans, root, cow, 1);
335 else
336 ret = btrfs_inc_ref(trans, root, cow, 0);
337 BUG_ON(ret);
339 if (new_flags != 0) {
340 ret = btrfs_set_disk_extent_flags(trans, root,
341 buf->start,
342 buf->len,
343 new_flags, 0);
344 BUG_ON(ret);
346 } else {
347 if (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF) {
348 if (root->root_key.objectid ==
349 BTRFS_TREE_RELOC_OBJECTID)
350 ret = btrfs_inc_ref(trans, root, cow, 1);
351 else
352 ret = btrfs_inc_ref(trans, root, cow, 0);
353 BUG_ON(ret);
354 ret = btrfs_dec_ref(trans, root, buf, 1);
355 BUG_ON(ret);
357 clean_tree_block(trans, root, buf);
358 *last_ref = 1;
360 return 0;
364 * does the dirty work in cow of a single block. The parent block (if
365 * supplied) is updated to point to the new cow copy. The new buffer is marked
366 * dirty and returned locked. If you modify the block it needs to be marked
367 * dirty again.
369 * search_start -- an allocation hint for the new block
371 * empty_size -- a hint that you plan on doing more cow. This is the size in
372 * bytes the allocator should try to find free next to the block it returns.
373 * This is just a hint and may be ignored by the allocator.
375 static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
376 struct btrfs_root *root,
377 struct extent_buffer *buf,
378 struct extent_buffer *parent, int parent_slot,
379 struct extent_buffer **cow_ret,
380 u64 search_start, u64 empty_size)
382 struct btrfs_disk_key disk_key;
383 struct extent_buffer *cow;
384 int level;
385 int last_ref = 0;
386 int unlock_orig = 0;
387 u64 parent_start;
389 if (*cow_ret == buf)
390 unlock_orig = 1;
392 btrfs_assert_tree_locked(buf);
394 WARN_ON(root->ref_cows && trans->transid !=
395 root->fs_info->running_transaction->transid);
396 WARN_ON(root->ref_cows && trans->transid != root->last_trans);
398 level = btrfs_header_level(buf);
400 if (level == 0)
401 btrfs_item_key(buf, &disk_key, 0);
402 else
403 btrfs_node_key(buf, &disk_key, 0);
405 if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) {
406 if (parent)
407 parent_start = parent->start;
408 else
409 parent_start = 0;
410 } else
411 parent_start = 0;
413 cow = btrfs_alloc_free_block(trans, root, buf->len, parent_start,
414 root->root_key.objectid, &disk_key,
415 level, search_start, empty_size);
416 if (IS_ERR(cow))
417 return PTR_ERR(cow);
419 /* cow is set to blocking by btrfs_init_new_buffer */
421 copy_extent_buffer(cow, buf, 0, 0, cow->len);
422 btrfs_set_header_bytenr(cow, cow->start);
423 btrfs_set_header_generation(cow, trans->transid);
424 btrfs_set_header_backref_rev(cow, BTRFS_MIXED_BACKREF_REV);
425 btrfs_clear_header_flag(cow, BTRFS_HEADER_FLAG_WRITTEN |
426 BTRFS_HEADER_FLAG_RELOC);
427 if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)
428 btrfs_set_header_flag(cow, BTRFS_HEADER_FLAG_RELOC);
429 else
430 btrfs_set_header_owner(cow, root->root_key.objectid);
432 write_extent_buffer(cow, root->fs_info->fsid,
433 (unsigned long)btrfs_header_fsid(cow),
434 BTRFS_FSID_SIZE);
436 update_ref_for_cow(trans, root, buf, cow, &last_ref);
438 if (root->ref_cows)
439 btrfs_reloc_cow_block(trans, root, buf, cow);
441 if (buf == root->node) {
442 WARN_ON(parent && parent != buf);
443 if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID ||
444 btrfs_header_backref_rev(buf) < BTRFS_MIXED_BACKREF_REV)
445 parent_start = buf->start;
446 else
447 parent_start = 0;
449 extent_buffer_get(cow);
450 rcu_assign_pointer(root->node, cow);
452 btrfs_free_tree_block(trans, root, buf, parent_start,
453 last_ref);
454 free_extent_buffer(buf);
455 add_root_to_dirty_list(root);
456 } else {
457 if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)
458 parent_start = parent->start;
459 else
460 parent_start = 0;
462 WARN_ON(trans->transid != btrfs_header_generation(parent));
463 btrfs_set_node_blockptr(parent, parent_slot,
464 cow->start);
465 btrfs_set_node_ptr_generation(parent, parent_slot,
466 trans->transid);
467 btrfs_mark_buffer_dirty(parent);
468 btrfs_free_tree_block(trans, root, buf, parent_start,
469 last_ref);
471 if (unlock_orig)
472 btrfs_tree_unlock(buf);
473 free_extent_buffer(buf);
474 btrfs_mark_buffer_dirty(cow);
475 *cow_ret = cow;
476 return 0;
479 static inline int should_cow_block(struct btrfs_trans_handle *trans,
480 struct btrfs_root *root,
481 struct extent_buffer *buf)
483 if (btrfs_header_generation(buf) == trans->transid &&
484 !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN) &&
485 !(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID &&
486 btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC)))
487 return 0;
488 return 1;
492 * cows a single block, see __btrfs_cow_block for the real work.
493 * This version of it has extra checks so that a block isn't cow'd more than
494 * once per transaction, as long as it hasn't been written yet
496 noinline int btrfs_cow_block(struct btrfs_trans_handle *trans,
497 struct btrfs_root *root, struct extent_buffer *buf,
498 struct extent_buffer *parent, int parent_slot,
499 struct extent_buffer **cow_ret)
501 u64 search_start;
502 int ret;
504 if (trans->transaction != root->fs_info->running_transaction) {
505 printk(KERN_CRIT "trans %llu running %llu\n",
506 (unsigned long long)trans->transid,
507 (unsigned long long)
508 root->fs_info->running_transaction->transid);
509 WARN_ON(1);
511 if (trans->transid != root->fs_info->generation) {
512 printk(KERN_CRIT "trans %llu running %llu\n",
513 (unsigned long long)trans->transid,
514 (unsigned long long)root->fs_info->generation);
515 WARN_ON(1);
518 if (!should_cow_block(trans, root, buf)) {
519 *cow_ret = buf;
520 return 0;
523 search_start = buf->start & ~((u64)(1024 * 1024 * 1024) - 1);
525 if (parent)
526 btrfs_set_lock_blocking(parent);
527 btrfs_set_lock_blocking(buf);
529 ret = __btrfs_cow_block(trans, root, buf, parent,
530 parent_slot, cow_ret, search_start, 0);
532 trace_btrfs_cow_block(root, buf, *cow_ret);
534 return ret;
538 * helper function for defrag to decide if two blocks pointed to by a
539 * node are actually close by
541 static int close_blocks(u64 blocknr, u64 other, u32 blocksize)
543 if (blocknr < other && other - (blocknr + blocksize) < 32768)
544 return 1;
545 if (blocknr > other && blocknr - (other + blocksize) < 32768)
546 return 1;
547 return 0;
551 * compare two keys in a memcmp fashion
553 static int comp_keys(struct btrfs_disk_key *disk, struct btrfs_key *k2)
555 struct btrfs_key k1;
557 btrfs_disk_key_to_cpu(&k1, disk);
559 return btrfs_comp_cpu_keys(&k1, k2);
563 * same as comp_keys only with two btrfs_key's
565 int btrfs_comp_cpu_keys(struct btrfs_key *k1, struct btrfs_key *k2)
567 if (k1->objectid > k2->objectid)
568 return 1;
569 if (k1->objectid < k2->objectid)
570 return -1;
571 if (k1->type > k2->type)
572 return 1;
573 if (k1->type < k2->type)
574 return -1;
575 if (k1->offset > k2->offset)
576 return 1;
577 if (k1->offset < k2->offset)
578 return -1;
579 return 0;
583 * this is used by the defrag code to go through all the
584 * leaves pointed to by a node and reallocate them so that
585 * disk order is close to key order
587 int btrfs_realloc_node(struct btrfs_trans_handle *trans,
588 struct btrfs_root *root, struct extent_buffer *parent,
589 int start_slot, int cache_only, u64 *last_ret,
590 struct btrfs_key *progress)
592 struct extent_buffer *cur;
593 u64 blocknr;
594 u64 gen;
595 u64 search_start = *last_ret;
596 u64 last_block = 0;
597 u64 other;
598 u32 parent_nritems;
599 int end_slot;
600 int i;
601 int err = 0;
602 int parent_level;
603 int uptodate;
604 u32 blocksize;
605 int progress_passed = 0;
606 struct btrfs_disk_key disk_key;
608 parent_level = btrfs_header_level(parent);
609 if (cache_only && parent_level != 1)
610 return 0;
612 if (trans->transaction != root->fs_info->running_transaction)
613 WARN_ON(1);
614 if (trans->transid != root->fs_info->generation)
615 WARN_ON(1);
617 parent_nritems = btrfs_header_nritems(parent);
618 blocksize = btrfs_level_size(root, parent_level - 1);
619 end_slot = parent_nritems;
621 if (parent_nritems == 1)
622 return 0;
624 btrfs_set_lock_blocking(parent);
626 for (i = start_slot; i < end_slot; i++) {
627 int close = 1;
629 if (!parent->map_token) {
630 map_extent_buffer(parent,
631 btrfs_node_key_ptr_offset(i),
632 sizeof(struct btrfs_key_ptr),
633 &parent->map_token, &parent->kaddr,
634 &parent->map_start, &parent->map_len,
635 KM_USER1);
637 btrfs_node_key(parent, &disk_key, i);
638 if (!progress_passed && comp_keys(&disk_key, progress) < 0)
639 continue;
641 progress_passed = 1;
642 blocknr = btrfs_node_blockptr(parent, i);
643 gen = btrfs_node_ptr_generation(parent, i);
644 if (last_block == 0)
645 last_block = blocknr;
647 if (i > 0) {
648 other = btrfs_node_blockptr(parent, i - 1);
649 close = close_blocks(blocknr, other, blocksize);
651 if (!close && i < end_slot - 2) {
652 other = btrfs_node_blockptr(parent, i + 1);
653 close = close_blocks(blocknr, other, blocksize);
655 if (close) {
656 last_block = blocknr;
657 continue;
659 if (parent->map_token) {
660 unmap_extent_buffer(parent, parent->map_token,
661 KM_USER1);
662 parent->map_token = NULL;
665 cur = btrfs_find_tree_block(root, blocknr, blocksize);
666 if (cur)
667 uptodate = btrfs_buffer_uptodate(cur, gen);
668 else
669 uptodate = 0;
670 if (!cur || !uptodate) {
671 if (cache_only) {
672 free_extent_buffer(cur);
673 continue;
675 if (!cur) {
676 cur = read_tree_block(root, blocknr,
677 blocksize, gen);
678 if (!cur)
679 return -EIO;
680 } else if (!uptodate) {
681 btrfs_read_buffer(cur, gen);
684 if (search_start == 0)
685 search_start = last_block;
687 btrfs_tree_lock(cur);
688 btrfs_set_lock_blocking(cur);
689 err = __btrfs_cow_block(trans, root, cur, parent, i,
690 &cur, search_start,
691 min(16 * blocksize,
692 (end_slot - i) * blocksize));
693 if (err) {
694 btrfs_tree_unlock(cur);
695 free_extent_buffer(cur);
696 break;
698 search_start = cur->start;
699 last_block = cur->start;
700 *last_ret = search_start;
701 btrfs_tree_unlock(cur);
702 free_extent_buffer(cur);
704 if (parent->map_token) {
705 unmap_extent_buffer(parent, parent->map_token,
706 KM_USER1);
707 parent->map_token = NULL;
709 return err;
713 * The leaf data grows from end-to-front in the node.
714 * this returns the address of the start of the last item,
715 * which is the stop of the leaf data stack
717 static inline unsigned int leaf_data_end(struct btrfs_root *root,
718 struct extent_buffer *leaf)
720 u32 nr = btrfs_header_nritems(leaf);
721 if (nr == 0)
722 return BTRFS_LEAF_DATA_SIZE(root);
723 return btrfs_item_offset_nr(leaf, nr - 1);
728 * search for key in the extent_buffer. The items start at offset p,
729 * and they are item_size apart. There are 'max' items in p.
731 * the slot in the array is returned via slot, and it points to
732 * the place where you would insert key if it is not found in
733 * the array.
735 * slot may point to max if the key is bigger than all of the keys
737 static noinline int generic_bin_search(struct extent_buffer *eb,
738 unsigned long p,
739 int item_size, struct btrfs_key *key,
740 int max, int *slot)
742 int low = 0;
743 int high = max;
744 int mid;
745 int ret;
746 struct btrfs_disk_key *tmp = NULL;
747 struct btrfs_disk_key unaligned;
748 unsigned long offset;
749 char *map_token = NULL;
750 char *kaddr = NULL;
751 unsigned long map_start = 0;
752 unsigned long map_len = 0;
753 int err;
755 while (low < high) {
756 mid = (low + high) / 2;
757 offset = p + mid * item_size;
759 if (!map_token || offset < map_start ||
760 (offset + sizeof(struct btrfs_disk_key)) >
761 map_start + map_len) {
762 if (map_token) {
763 unmap_extent_buffer(eb, map_token, KM_USER0);
764 map_token = NULL;
767 err = map_private_extent_buffer(eb, offset,
768 sizeof(struct btrfs_disk_key),
769 &map_token, &kaddr,
770 &map_start, &map_len, KM_USER0);
772 if (!err) {
773 tmp = (struct btrfs_disk_key *)(kaddr + offset -
774 map_start);
775 } else {
776 read_extent_buffer(eb, &unaligned,
777 offset, sizeof(unaligned));
778 tmp = &unaligned;
781 } else {
782 tmp = (struct btrfs_disk_key *)(kaddr + offset -
783 map_start);
785 ret = comp_keys(tmp, key);
787 if (ret < 0)
788 low = mid + 1;
789 else if (ret > 0)
790 high = mid;
791 else {
792 *slot = mid;
793 if (map_token)
794 unmap_extent_buffer(eb, map_token, KM_USER0);
795 return 0;
798 *slot = low;
799 if (map_token)
800 unmap_extent_buffer(eb, map_token, KM_USER0);
801 return 1;
805 * simple bin_search frontend that does the right thing for
806 * leaves vs nodes
808 static int bin_search(struct extent_buffer *eb, struct btrfs_key *key,
809 int level, int *slot)
811 if (level == 0) {
812 return generic_bin_search(eb,
813 offsetof(struct btrfs_leaf, items),
814 sizeof(struct btrfs_item),
815 key, btrfs_header_nritems(eb),
816 slot);
817 } else {
818 return generic_bin_search(eb,
819 offsetof(struct btrfs_node, ptrs),
820 sizeof(struct btrfs_key_ptr),
821 key, btrfs_header_nritems(eb),
822 slot);
824 return -1;
827 int btrfs_bin_search(struct extent_buffer *eb, struct btrfs_key *key,
828 int level, int *slot)
830 return bin_search(eb, key, level, slot);
833 static void root_add_used(struct btrfs_root *root, u32 size)
835 spin_lock(&root->accounting_lock);
836 btrfs_set_root_used(&root->root_item,
837 btrfs_root_used(&root->root_item) + size);
838 spin_unlock(&root->accounting_lock);
841 static void root_sub_used(struct btrfs_root *root, u32 size)
843 spin_lock(&root->accounting_lock);
844 btrfs_set_root_used(&root->root_item,
845 btrfs_root_used(&root->root_item) - size);
846 spin_unlock(&root->accounting_lock);
849 /* given a node and slot number, this reads the blocks it points to. The
850 * extent buffer is returned with a reference taken (but unlocked).
851 * NULL is returned on error.
853 static noinline struct extent_buffer *read_node_slot(struct btrfs_root *root,
854 struct extent_buffer *parent, int slot)
856 int level = btrfs_header_level(parent);
857 if (slot < 0)
858 return NULL;
859 if (slot >= btrfs_header_nritems(parent))
860 return NULL;
862 BUG_ON(level == 0);
864 return read_tree_block(root, btrfs_node_blockptr(parent, slot),
865 btrfs_level_size(root, level - 1),
866 btrfs_node_ptr_generation(parent, slot));
870 * node level balancing, used to make sure nodes are in proper order for
871 * item deletion. We balance from the top down, so we have to make sure
872 * that a deletion won't leave an node completely empty later on.
874 static noinline int balance_level(struct btrfs_trans_handle *trans,
875 struct btrfs_root *root,
876 struct btrfs_path *path, int level)
878 struct extent_buffer *right = NULL;
879 struct extent_buffer *mid;
880 struct extent_buffer *left = NULL;
881 struct extent_buffer *parent = NULL;
882 int ret = 0;
883 int wret;
884 int pslot;
885 int orig_slot = path->slots[level];
886 u64 orig_ptr;
888 if (level == 0)
889 return 0;
891 mid = path->nodes[level];
893 WARN_ON(!path->locks[level]);
894 WARN_ON(btrfs_header_generation(mid) != trans->transid);
896 orig_ptr = btrfs_node_blockptr(mid, orig_slot);
898 if (level < BTRFS_MAX_LEVEL - 1)
899 parent = path->nodes[level + 1];
900 pslot = path->slots[level + 1];
903 * deal with the case where there is only one pointer in the root
904 * by promoting the node below to a root
906 if (!parent) {
907 struct extent_buffer *child;
909 if (btrfs_header_nritems(mid) != 1)
910 return 0;
912 /* promote the child to a root */
913 child = read_node_slot(root, mid, 0);
914 BUG_ON(!child);
915 btrfs_tree_lock(child);
916 btrfs_set_lock_blocking(child);
917 ret = btrfs_cow_block(trans, root, child, mid, 0, &child);
918 if (ret) {
919 btrfs_tree_unlock(child);
920 free_extent_buffer(child);
921 goto enospc;
924 rcu_assign_pointer(root->node, child);
926 add_root_to_dirty_list(root);
927 btrfs_tree_unlock(child);
929 path->locks[level] = 0;
930 path->nodes[level] = NULL;
931 clean_tree_block(trans, root, mid);
932 btrfs_tree_unlock(mid);
933 /* once for the path */
934 free_extent_buffer(mid);
936 root_sub_used(root, mid->len);
937 btrfs_free_tree_block(trans, root, mid, 0, 1);
938 /* once for the root ptr */
939 free_extent_buffer(mid);
940 return 0;
942 if (btrfs_header_nritems(mid) >
943 BTRFS_NODEPTRS_PER_BLOCK(root) / 4)
944 return 0;
946 btrfs_header_nritems(mid);
948 left = read_node_slot(root, parent, pslot - 1);
949 if (left) {
950 btrfs_tree_lock(left);
951 btrfs_set_lock_blocking(left);
952 wret = btrfs_cow_block(trans, root, left,
953 parent, pslot - 1, &left);
954 if (wret) {
955 ret = wret;
956 goto enospc;
959 right = read_node_slot(root, parent, pslot + 1);
960 if (right) {
961 btrfs_tree_lock(right);
962 btrfs_set_lock_blocking(right);
963 wret = btrfs_cow_block(trans, root, right,
964 parent, pslot + 1, &right);
965 if (wret) {
966 ret = wret;
967 goto enospc;
971 /* first, try to make some room in the middle buffer */
972 if (left) {
973 orig_slot += btrfs_header_nritems(left);
974 wret = push_node_left(trans, root, left, mid, 1);
975 if (wret < 0)
976 ret = wret;
977 btrfs_header_nritems(mid);
981 * then try to empty the right most buffer into the middle
983 if (right) {
984 wret = push_node_left(trans, root, mid, right, 1);
985 if (wret < 0 && wret != -ENOSPC)
986 ret = wret;
987 if (btrfs_header_nritems(right) == 0) {
988 clean_tree_block(trans, root, right);
989 btrfs_tree_unlock(right);
990 wret = del_ptr(trans, root, path, level + 1, pslot +
992 if (wret)
993 ret = wret;
994 root_sub_used(root, right->len);
995 btrfs_free_tree_block(trans, root, right, 0, 1);
996 free_extent_buffer(right);
997 right = NULL;
998 } else {
999 struct btrfs_disk_key right_key;
1000 btrfs_node_key(right, &right_key, 0);
1001 btrfs_set_node_key(parent, &right_key, pslot + 1);
1002 btrfs_mark_buffer_dirty(parent);
1005 if (btrfs_header_nritems(mid) == 1) {
1007 * we're not allowed to leave a node with one item in the
1008 * tree during a delete. A deletion from lower in the tree
1009 * could try to delete the only pointer in this node.
1010 * So, pull some keys from the left.
1011 * There has to be a left pointer at this point because
1012 * otherwise we would have pulled some pointers from the
1013 * right
1015 BUG_ON(!left);
1016 wret = balance_node_right(trans, root, mid, left);
1017 if (wret < 0) {
1018 ret = wret;
1019 goto enospc;
1021 if (wret == 1) {
1022 wret = push_node_left(trans, root, left, mid, 1);
1023 if (wret < 0)
1024 ret = wret;
1026 BUG_ON(wret == 1);
1028 if (btrfs_header_nritems(mid) == 0) {
1029 clean_tree_block(trans, root, mid);
1030 btrfs_tree_unlock(mid);
1031 wret = del_ptr(trans, root, path, level + 1, pslot);
1032 if (wret)
1033 ret = wret;
1034 root_sub_used(root, mid->len);
1035 btrfs_free_tree_block(trans, root, mid, 0, 1);
1036 free_extent_buffer(mid);
1037 mid = NULL;
1038 } else {
1039 /* update the parent key to reflect our changes */
1040 struct btrfs_disk_key mid_key;
1041 btrfs_node_key(mid, &mid_key, 0);
1042 btrfs_set_node_key(parent, &mid_key, pslot);
1043 btrfs_mark_buffer_dirty(parent);
1046 /* update the path */
1047 if (left) {
1048 if (btrfs_header_nritems(left) > orig_slot) {
1049 extent_buffer_get(left);
1050 /* left was locked after cow */
1051 path->nodes[level] = left;
1052 path->slots[level + 1] -= 1;
1053 path->slots[level] = orig_slot;
1054 if (mid) {
1055 btrfs_tree_unlock(mid);
1056 free_extent_buffer(mid);
1058 } else {
1059 orig_slot -= btrfs_header_nritems(left);
1060 path->slots[level] = orig_slot;
1063 /* double check we haven't messed things up */
1064 if (orig_ptr !=
1065 btrfs_node_blockptr(path->nodes[level], path->slots[level]))
1066 BUG();
1067 enospc:
1068 if (right) {
1069 btrfs_tree_unlock(right);
1070 free_extent_buffer(right);
1072 if (left) {
1073 if (path->nodes[level] != left)
1074 btrfs_tree_unlock(left);
1075 free_extent_buffer(left);
1077 return ret;
1080 /* Node balancing for insertion. Here we only split or push nodes around
1081 * when they are completely full. This is also done top down, so we
1082 * have to be pessimistic.
1084 static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans,
1085 struct btrfs_root *root,
1086 struct btrfs_path *path, int level)
1088 struct extent_buffer *right = NULL;
1089 struct extent_buffer *mid;
1090 struct extent_buffer *left = NULL;
1091 struct extent_buffer *parent = NULL;
1092 int ret = 0;
1093 int wret;
1094 int pslot;
1095 int orig_slot = path->slots[level];
1097 if (level == 0)
1098 return 1;
1100 mid = path->nodes[level];
1101 WARN_ON(btrfs_header_generation(mid) != trans->transid);
1103 if (level < BTRFS_MAX_LEVEL - 1)
1104 parent = path->nodes[level + 1];
1105 pslot = path->slots[level + 1];
1107 if (!parent)
1108 return 1;
1110 left = read_node_slot(root, parent, pslot - 1);
1112 /* first, try to make some room in the middle buffer */
1113 if (left) {
1114 u32 left_nr;
1116 btrfs_tree_lock(left);
1117 btrfs_set_lock_blocking(left);
1119 left_nr = btrfs_header_nritems(left);
1120 if (left_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
1121 wret = 1;
1122 } else {
1123 ret = btrfs_cow_block(trans, root, left, parent,
1124 pslot - 1, &left);
1125 if (ret)
1126 wret = 1;
1127 else {
1128 wret = push_node_left(trans, root,
1129 left, mid, 0);
1132 if (wret < 0)
1133 ret = wret;
1134 if (wret == 0) {
1135 struct btrfs_disk_key disk_key;
1136 orig_slot += left_nr;
1137 btrfs_node_key(mid, &disk_key, 0);
1138 btrfs_set_node_key(parent, &disk_key, pslot);
1139 btrfs_mark_buffer_dirty(parent);
1140 if (btrfs_header_nritems(left) > orig_slot) {
1141 path->nodes[level] = left;
1142 path->slots[level + 1] -= 1;
1143 path->slots[level] = orig_slot;
1144 btrfs_tree_unlock(mid);
1145 free_extent_buffer(mid);
1146 } else {
1147 orig_slot -=
1148 btrfs_header_nritems(left);
1149 path->slots[level] = orig_slot;
1150 btrfs_tree_unlock(left);
1151 free_extent_buffer(left);
1153 return 0;
1155 btrfs_tree_unlock(left);
1156 free_extent_buffer(left);
1158 right = read_node_slot(root, parent, pslot + 1);
1161 * then try to empty the right most buffer into the middle
1163 if (right) {
1164 u32 right_nr;
1166 btrfs_tree_lock(right);
1167 btrfs_set_lock_blocking(right);
1169 right_nr = btrfs_header_nritems(right);
1170 if (right_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) {
1171 wret = 1;
1172 } else {
1173 ret = btrfs_cow_block(trans, root, right,
1174 parent, pslot + 1,
1175 &right);
1176 if (ret)
1177 wret = 1;
1178 else {
1179 wret = balance_node_right(trans, root,
1180 right, mid);
1183 if (wret < 0)
1184 ret = wret;
1185 if (wret == 0) {
1186 struct btrfs_disk_key disk_key;
1188 btrfs_node_key(right, &disk_key, 0);
1189 btrfs_set_node_key(parent, &disk_key, pslot + 1);
1190 btrfs_mark_buffer_dirty(parent);
1192 if (btrfs_header_nritems(mid) <= orig_slot) {
1193 path->nodes[level] = right;
1194 path->slots[level + 1] += 1;
1195 path->slots[level] = orig_slot -
1196 btrfs_header_nritems(mid);
1197 btrfs_tree_unlock(mid);
1198 free_extent_buffer(mid);
1199 } else {
1200 btrfs_tree_unlock(right);
1201 free_extent_buffer(right);
1203 return 0;
1205 btrfs_tree_unlock(right);
1206 free_extent_buffer(right);
1208 return 1;
1212 * readahead one full node of leaves, finding things that are close
1213 * to the block in 'slot', and triggering ra on them.
1215 static void reada_for_search(struct btrfs_root *root,
1216 struct btrfs_path *path,
1217 int level, int slot, u64 objectid)
1219 struct extent_buffer *node;
1220 struct btrfs_disk_key disk_key;
1221 u32 nritems;
1222 u64 search;
1223 u64 target;
1224 u64 nread = 0;
1225 u64 gen;
1226 int direction = path->reada;
1227 struct extent_buffer *eb;
1228 u32 nr;
1229 u32 blocksize;
1230 u32 nscan = 0;
1231 bool map = true;
1233 if (level != 1)
1234 return;
1236 if (!path->nodes[level])
1237 return;
1239 node = path->nodes[level];
1241 search = btrfs_node_blockptr(node, slot);
1242 blocksize = btrfs_level_size(root, level - 1);
1243 eb = btrfs_find_tree_block(root, search, blocksize);
1244 if (eb) {
1245 free_extent_buffer(eb);
1246 return;
1249 target = search;
1251 nritems = btrfs_header_nritems(node);
1252 nr = slot;
1253 if (node->map_token || path->skip_locking)
1254 map = false;
1256 while (1) {
1257 if (map && !node->map_token) {
1258 unsigned long offset = btrfs_node_key_ptr_offset(nr);
1259 map_private_extent_buffer(node, offset,
1260 sizeof(struct btrfs_key_ptr),
1261 &node->map_token,
1262 &node->kaddr,
1263 &node->map_start,
1264 &node->map_len, KM_USER1);
1266 if (direction < 0) {
1267 if (nr == 0)
1268 break;
1269 nr--;
1270 } else if (direction > 0) {
1271 nr++;
1272 if (nr >= nritems)
1273 break;
1275 if (path->reada < 0 && objectid) {
1276 btrfs_node_key(node, &disk_key, nr);
1277 if (btrfs_disk_key_objectid(&disk_key) != objectid)
1278 break;
1280 search = btrfs_node_blockptr(node, nr);
1281 if ((search <= target && target - search <= 65536) ||
1282 (search > target && search - target <= 65536)) {
1283 gen = btrfs_node_ptr_generation(node, nr);
1284 if (map && node->map_token) {
1285 unmap_extent_buffer(node, node->map_token,
1286 KM_USER1);
1287 node->map_token = NULL;
1289 readahead_tree_block(root, search, blocksize, gen);
1290 nread += blocksize;
1292 nscan++;
1293 if ((nread > 65536 || nscan > 32))
1294 break;
1296 if (map && node->map_token) {
1297 unmap_extent_buffer(node, node->map_token, KM_USER1);
1298 node->map_token = NULL;
1303 * returns -EAGAIN if it had to drop the path, or zero if everything was in
1304 * cache
1306 static noinline int reada_for_balance(struct btrfs_root *root,
1307 struct btrfs_path *path, int level)
1309 int slot;
1310 int nritems;
1311 struct extent_buffer *parent;
1312 struct extent_buffer *eb;
1313 u64 gen;
1314 u64 block1 = 0;
1315 u64 block2 = 0;
1316 int ret = 0;
1317 int blocksize;
1319 parent = path->nodes[level + 1];
1320 if (!parent)
1321 return 0;
1323 nritems = btrfs_header_nritems(parent);
1324 slot = path->slots[level + 1];
1325 blocksize = btrfs_level_size(root, level);
1327 if (slot > 0) {
1328 block1 = btrfs_node_blockptr(parent, slot - 1);
1329 gen = btrfs_node_ptr_generation(parent, slot - 1);
1330 eb = btrfs_find_tree_block(root, block1, blocksize);
1331 if (eb && btrfs_buffer_uptodate(eb, gen))
1332 block1 = 0;
1333 free_extent_buffer(eb);
1335 if (slot + 1 < nritems) {
1336 block2 = btrfs_node_blockptr(parent, slot + 1);
1337 gen = btrfs_node_ptr_generation(parent, slot + 1);
1338 eb = btrfs_find_tree_block(root, block2, blocksize);
1339 if (eb && btrfs_buffer_uptodate(eb, gen))
1340 block2 = 0;
1341 free_extent_buffer(eb);
1343 if (block1 || block2) {
1344 ret = -EAGAIN;
1346 /* release the whole path */
1347 btrfs_release_path(path);
1349 /* read the blocks */
1350 if (block1)
1351 readahead_tree_block(root, block1, blocksize, 0);
1352 if (block2)
1353 readahead_tree_block(root, block2, blocksize, 0);
1355 if (block1) {
1356 eb = read_tree_block(root, block1, blocksize, 0);
1357 free_extent_buffer(eb);
1359 if (block2) {
1360 eb = read_tree_block(root, block2, blocksize, 0);
1361 free_extent_buffer(eb);
1364 return ret;
1369 * when we walk down the tree, it is usually safe to unlock the higher layers
1370 * in the tree. The exceptions are when our path goes through slot 0, because
1371 * operations on the tree might require changing key pointers higher up in the
1372 * tree.
1374 * callers might also have set path->keep_locks, which tells this code to keep
1375 * the lock if the path points to the last slot in the block. This is part of
1376 * walking through the tree, and selecting the next slot in the higher block.
1378 * lowest_unlock sets the lowest level in the tree we're allowed to unlock. so
1379 * if lowest_unlock is 1, level 0 won't be unlocked
1381 static noinline void unlock_up(struct btrfs_path *path, int level,
1382 int lowest_unlock)
1384 int i;
1385 int skip_level = level;
1386 int no_skips = 0;
1387 struct extent_buffer *t;
1389 for (i = level; i < BTRFS_MAX_LEVEL; i++) {
1390 if (!path->nodes[i])
1391 break;
1392 if (!path->locks[i])
1393 break;
1394 if (!no_skips && path->slots[i] == 0) {
1395 skip_level = i + 1;
1396 continue;
1398 if (!no_skips && path->keep_locks) {
1399 u32 nritems;
1400 t = path->nodes[i];
1401 nritems = btrfs_header_nritems(t);
1402 if (nritems < 1 || path->slots[i] >= nritems - 1) {
1403 skip_level = i + 1;
1404 continue;
1407 if (skip_level < i && i >= lowest_unlock)
1408 no_skips = 1;
1410 t = path->nodes[i];
1411 if (i >= lowest_unlock && i > skip_level && path->locks[i]) {
1412 btrfs_tree_unlock(t);
1413 path->locks[i] = 0;
1419 * This releases any locks held in the path starting at level and
1420 * going all the way up to the root.
1422 * btrfs_search_slot will keep the lock held on higher nodes in a few
1423 * corner cases, such as COW of the block at slot zero in the node. This
1424 * ignores those rules, and it should only be called when there are no
1425 * more updates to be done higher up in the tree.
1427 noinline void btrfs_unlock_up_safe(struct btrfs_path *path, int level)
1429 int i;
1431 if (path->keep_locks)
1432 return;
1434 for (i = level; i < BTRFS_MAX_LEVEL; i++) {
1435 if (!path->nodes[i])
1436 continue;
1437 if (!path->locks[i])
1438 continue;
1439 btrfs_tree_unlock(path->nodes[i]);
1440 path->locks[i] = 0;
1445 * helper function for btrfs_search_slot. The goal is to find a block
1446 * in cache without setting the path to blocking. If we find the block
1447 * we return zero and the path is unchanged.
1449 * If we can't find the block, we set the path blocking and do some
1450 * reada. -EAGAIN is returned and the search must be repeated.
1452 static int
1453 read_block_for_search(struct btrfs_trans_handle *trans,
1454 struct btrfs_root *root, struct btrfs_path *p,
1455 struct extent_buffer **eb_ret, int level, int slot,
1456 struct btrfs_key *key)
1458 u64 blocknr;
1459 u64 gen;
1460 u32 blocksize;
1461 struct extent_buffer *b = *eb_ret;
1462 struct extent_buffer *tmp;
1463 int ret;
1465 blocknr = btrfs_node_blockptr(b, slot);
1466 gen = btrfs_node_ptr_generation(b, slot);
1467 blocksize = btrfs_level_size(root, level - 1);
1469 tmp = btrfs_find_tree_block(root, blocknr, blocksize);
1470 if (tmp) {
1471 if (btrfs_buffer_uptodate(tmp, 0)) {
1472 if (btrfs_buffer_uptodate(tmp, gen)) {
1474 * we found an up to date block without
1475 * sleeping, return
1476 * right away
1478 *eb_ret = tmp;
1479 return 0;
1481 /* the pages were up to date, but we failed
1482 * the generation number check. Do a full
1483 * read for the generation number that is correct.
1484 * We must do this without dropping locks so
1485 * we can trust our generation number
1487 free_extent_buffer(tmp);
1488 tmp = read_tree_block(root, blocknr, blocksize, gen);
1489 if (tmp && btrfs_buffer_uptodate(tmp, gen)) {
1490 *eb_ret = tmp;
1491 return 0;
1493 free_extent_buffer(tmp);
1494 btrfs_release_path(p);
1495 return -EIO;
1500 * reduce lock contention at high levels
1501 * of the btree by dropping locks before
1502 * we read. Don't release the lock on the current
1503 * level because we need to walk this node to figure
1504 * out which blocks to read.
1506 btrfs_unlock_up_safe(p, level + 1);
1507 btrfs_set_path_blocking(p);
1509 free_extent_buffer(tmp);
1510 if (p->reada)
1511 reada_for_search(root, p, level, slot, key->objectid);
1513 btrfs_release_path(p);
1515 ret = -EAGAIN;
1516 tmp = read_tree_block(root, blocknr, blocksize, 0);
1517 if (tmp) {
1519 * If the read above didn't mark this buffer up to date,
1520 * it will never end up being up to date. Set ret to EIO now
1521 * and give up so that our caller doesn't loop forever
1522 * on our EAGAINs.
1524 if (!btrfs_buffer_uptodate(tmp, 0))
1525 ret = -EIO;
1526 free_extent_buffer(tmp);
1528 return ret;
1532 * helper function for btrfs_search_slot. This does all of the checks
1533 * for node-level blocks and does any balancing required based on
1534 * the ins_len.
1536 * If no extra work was required, zero is returned. If we had to
1537 * drop the path, -EAGAIN is returned and btrfs_search_slot must
1538 * start over
1540 static int
1541 setup_nodes_for_search(struct btrfs_trans_handle *trans,
1542 struct btrfs_root *root, struct btrfs_path *p,
1543 struct extent_buffer *b, int level, int ins_len)
1545 int ret;
1546 if ((p->search_for_split || ins_len > 0) && btrfs_header_nritems(b) >=
1547 BTRFS_NODEPTRS_PER_BLOCK(root) - 3) {
1548 int sret;
1550 sret = reada_for_balance(root, p, level);
1551 if (sret)
1552 goto again;
1554 btrfs_set_path_blocking(p);
1555 sret = split_node(trans, root, p, level);
1556 btrfs_clear_path_blocking(p, NULL);
1558 BUG_ON(sret > 0);
1559 if (sret) {
1560 ret = sret;
1561 goto done;
1563 b = p->nodes[level];
1564 } else if (ins_len < 0 && btrfs_header_nritems(b) <
1565 BTRFS_NODEPTRS_PER_BLOCK(root) / 2) {
1566 int sret;
1568 sret = reada_for_balance(root, p, level);
1569 if (sret)
1570 goto again;
1572 btrfs_set_path_blocking(p);
1573 sret = balance_level(trans, root, p, level);
1574 btrfs_clear_path_blocking(p, NULL);
1576 if (sret) {
1577 ret = sret;
1578 goto done;
1580 b = p->nodes[level];
1581 if (!b) {
1582 btrfs_release_path(p);
1583 goto again;
1585 BUG_ON(btrfs_header_nritems(b) == 1);
1587 return 0;
1589 again:
1590 ret = -EAGAIN;
1591 done:
1592 return ret;
1596 * look for key in the tree. path is filled in with nodes along the way
1597 * if key is found, we return zero and you can find the item in the leaf
1598 * level of the path (level 0)
1600 * If the key isn't found, the path points to the slot where it should
1601 * be inserted, and 1 is returned. If there are other errors during the
1602 * search a negative error number is returned.
1604 * if ins_len > 0, nodes and leaves will be split as we walk down the
1605 * tree. if ins_len < 0, nodes will be merged as we walk down the tree (if
1606 * possible)
1608 int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root
1609 *root, struct btrfs_key *key, struct btrfs_path *p, int
1610 ins_len, int cow)
1612 struct extent_buffer *b;
1613 int slot;
1614 int ret;
1615 int err;
1616 int level;
1617 int lowest_unlock = 1;
1618 u8 lowest_level = 0;
1620 lowest_level = p->lowest_level;
1621 WARN_ON(lowest_level && ins_len > 0);
1622 WARN_ON(p->nodes[0] != NULL);
1624 if (ins_len < 0)
1625 lowest_unlock = 2;
1627 again:
1628 if (p->search_commit_root) {
1629 b = root->commit_root;
1630 extent_buffer_get(b);
1631 if (!p->skip_locking)
1632 btrfs_tree_lock(b);
1633 } else {
1634 if (p->skip_locking)
1635 b = btrfs_root_node(root);
1636 else
1637 b = btrfs_lock_root_node(root);
1640 while (b) {
1641 level = btrfs_header_level(b);
1644 * setup the path here so we can release it under lock
1645 * contention with the cow code
1647 p->nodes[level] = b;
1648 if (!p->skip_locking)
1649 p->locks[level] = 1;
1651 if (cow) {
1653 * if we don't really need to cow this block
1654 * then we don't want to set the path blocking,
1655 * so we test it here
1657 if (!should_cow_block(trans, root, b))
1658 goto cow_done;
1660 btrfs_set_path_blocking(p);
1662 err = btrfs_cow_block(trans, root, b,
1663 p->nodes[level + 1],
1664 p->slots[level + 1], &b);
1665 if (err) {
1666 ret = err;
1667 goto done;
1670 cow_done:
1671 BUG_ON(!cow && ins_len);
1673 p->nodes[level] = b;
1674 if (!p->skip_locking)
1675 p->locks[level] = 1;
1677 btrfs_clear_path_blocking(p, NULL);
1680 * we have a lock on b and as long as we aren't changing
1681 * the tree, there is no way to for the items in b to change.
1682 * It is safe to drop the lock on our parent before we
1683 * go through the expensive btree search on b.
1685 * If cow is true, then we might be changing slot zero,
1686 * which may require changing the parent. So, we can't
1687 * drop the lock until after we know which slot we're
1688 * operating on.
1690 if (!cow)
1691 btrfs_unlock_up_safe(p, level + 1);
1693 ret = bin_search(b, key, level, &slot);
1695 if (level != 0) {
1696 int dec = 0;
1697 if (ret && slot > 0) {
1698 dec = 1;
1699 slot -= 1;
1701 p->slots[level] = slot;
1702 err = setup_nodes_for_search(trans, root, p, b, level,
1703 ins_len);
1704 if (err == -EAGAIN)
1705 goto again;
1706 if (err) {
1707 ret = err;
1708 goto done;
1710 b = p->nodes[level];
1711 slot = p->slots[level];
1713 unlock_up(p, level, lowest_unlock);
1715 if (level == lowest_level) {
1716 if (dec)
1717 p->slots[level]++;
1718 goto done;
1721 err = read_block_for_search(trans, root, p,
1722 &b, level, slot, key);
1723 if (err == -EAGAIN)
1724 goto again;
1725 if (err) {
1726 ret = err;
1727 goto done;
1730 if (!p->skip_locking) {
1731 btrfs_clear_path_blocking(p, NULL);
1732 err = btrfs_try_spin_lock(b);
1734 if (!err) {
1735 btrfs_set_path_blocking(p);
1736 btrfs_tree_lock(b);
1737 btrfs_clear_path_blocking(p, b);
1740 } else {
1741 p->slots[level] = slot;
1742 if (ins_len > 0 &&
1743 btrfs_leaf_free_space(root, b) < ins_len) {
1744 btrfs_set_path_blocking(p);
1745 err = split_leaf(trans, root, key,
1746 p, ins_len, ret == 0);
1747 btrfs_clear_path_blocking(p, NULL);
1749 BUG_ON(err > 0);
1750 if (err) {
1751 ret = err;
1752 goto done;
1755 if (!p->search_for_split)
1756 unlock_up(p, level, lowest_unlock);
1757 goto done;
1760 ret = 1;
1761 done:
1763 * we don't really know what they plan on doing with the path
1764 * from here on, so for now just mark it as blocking
1766 if (!p->leave_spinning)
1767 btrfs_set_path_blocking(p);
1768 if (ret < 0)
1769 btrfs_release_path(p);
1770 return ret;
1774 * adjust the pointers going up the tree, starting at level
1775 * making sure the right key of each node is points to 'key'.
1776 * This is used after shifting pointers to the left, so it stops
1777 * fixing up pointers when a given leaf/node is not in slot 0 of the
1778 * higher levels
1780 * If this fails to write a tree block, it returns -1, but continues
1781 * fixing up the blocks in ram so the tree is consistent.
1783 static int fixup_low_keys(struct btrfs_trans_handle *trans,
1784 struct btrfs_root *root, struct btrfs_path *path,
1785 struct btrfs_disk_key *key, int level)
1787 int i;
1788 int ret = 0;
1789 struct extent_buffer *t;
1791 for (i = level; i < BTRFS_MAX_LEVEL; i++) {
1792 int tslot = path->slots[i];
1793 if (!path->nodes[i])
1794 break;
1795 t = path->nodes[i];
1796 btrfs_set_node_key(t, key, tslot);
1797 btrfs_mark_buffer_dirty(path->nodes[i]);
1798 if (tslot != 0)
1799 break;
1801 return ret;
1805 * update item key.
1807 * This function isn't completely safe. It's the caller's responsibility
1808 * that the new key won't break the order
1810 int btrfs_set_item_key_safe(struct btrfs_trans_handle *trans,
1811 struct btrfs_root *root, struct btrfs_path *path,
1812 struct btrfs_key *new_key)
1814 struct btrfs_disk_key disk_key;
1815 struct extent_buffer *eb;
1816 int slot;
1818 eb = path->nodes[0];
1819 slot = path->slots[0];
1820 if (slot > 0) {
1821 btrfs_item_key(eb, &disk_key, slot - 1);
1822 if (comp_keys(&disk_key, new_key) >= 0)
1823 return -1;
1825 if (slot < btrfs_header_nritems(eb) - 1) {
1826 btrfs_item_key(eb, &disk_key, slot + 1);
1827 if (comp_keys(&disk_key, new_key) <= 0)
1828 return -1;
1831 btrfs_cpu_key_to_disk(&disk_key, new_key);
1832 btrfs_set_item_key(eb, &disk_key, slot);
1833 btrfs_mark_buffer_dirty(eb);
1834 if (slot == 0)
1835 fixup_low_keys(trans, root, path, &disk_key, 1);
1836 return 0;
1840 * try to push data from one node into the next node left in the
1841 * tree.
1843 * returns 0 if some ptrs were pushed left, < 0 if there was some horrible
1844 * error, and > 0 if there was no room in the left hand block.
1846 static int push_node_left(struct btrfs_trans_handle *trans,
1847 struct btrfs_root *root, struct extent_buffer *dst,
1848 struct extent_buffer *src, int empty)
1850 int push_items = 0;
1851 int src_nritems;
1852 int dst_nritems;
1853 int ret = 0;
1855 src_nritems = btrfs_header_nritems(src);
1856 dst_nritems = btrfs_header_nritems(dst);
1857 push_items = BTRFS_NODEPTRS_PER_BLOCK(root) - dst_nritems;
1858 WARN_ON(btrfs_header_generation(src) != trans->transid);
1859 WARN_ON(btrfs_header_generation(dst) != trans->transid);
1861 if (!empty && src_nritems <= 8)
1862 return 1;
1864 if (push_items <= 0)
1865 return 1;
1867 if (empty) {
1868 push_items = min(src_nritems, push_items);
1869 if (push_items < src_nritems) {
1870 /* leave at least 8 pointers in the node if
1871 * we aren't going to empty it
1873 if (src_nritems - push_items < 8) {
1874 if (push_items <= 8)
1875 return 1;
1876 push_items -= 8;
1879 } else
1880 push_items = min(src_nritems - 8, push_items);
1882 copy_extent_buffer(dst, src,
1883 btrfs_node_key_ptr_offset(dst_nritems),
1884 btrfs_node_key_ptr_offset(0),
1885 push_items * sizeof(struct btrfs_key_ptr));
1887 if (push_items < src_nritems) {
1888 memmove_extent_buffer(src, btrfs_node_key_ptr_offset(0),
1889 btrfs_node_key_ptr_offset(push_items),
1890 (src_nritems - push_items) *
1891 sizeof(struct btrfs_key_ptr));
1893 btrfs_set_header_nritems(src, src_nritems - push_items);
1894 btrfs_set_header_nritems(dst, dst_nritems + push_items);
1895 btrfs_mark_buffer_dirty(src);
1896 btrfs_mark_buffer_dirty(dst);
1898 return ret;
1902 * try to push data from one node into the next node right in the
1903 * tree.
1905 * returns 0 if some ptrs were pushed, < 0 if there was some horrible
1906 * error, and > 0 if there was no room in the right hand block.
1908 * this will only push up to 1/2 the contents of the left node over
1910 static int balance_node_right(struct btrfs_trans_handle *trans,
1911 struct btrfs_root *root,
1912 struct extent_buffer *dst,
1913 struct extent_buffer *src)
1915 int push_items = 0;
1916 int max_push;
1917 int src_nritems;
1918 int dst_nritems;
1919 int ret = 0;
1921 WARN_ON(btrfs_header_generation(src) != trans->transid);
1922 WARN_ON(btrfs_header_generation(dst) != trans->transid);
1924 src_nritems = btrfs_header_nritems(src);
1925 dst_nritems = btrfs_header_nritems(dst);
1926 push_items = BTRFS_NODEPTRS_PER_BLOCK(root) - dst_nritems;
1927 if (push_items <= 0)
1928 return 1;
1930 if (src_nritems < 4)
1931 return 1;
1933 max_push = src_nritems / 2 + 1;
1934 /* don't try to empty the node */
1935 if (max_push >= src_nritems)
1936 return 1;
1938 if (max_push < push_items)
1939 push_items = max_push;
1941 memmove_extent_buffer(dst, btrfs_node_key_ptr_offset(push_items),
1942 btrfs_node_key_ptr_offset(0),
1943 (dst_nritems) *
1944 sizeof(struct btrfs_key_ptr));
1946 copy_extent_buffer(dst, src,
1947 btrfs_node_key_ptr_offset(0),
1948 btrfs_node_key_ptr_offset(src_nritems - push_items),
1949 push_items * sizeof(struct btrfs_key_ptr));
1951 btrfs_set_header_nritems(src, src_nritems - push_items);
1952 btrfs_set_header_nritems(dst, dst_nritems + push_items);
1954 btrfs_mark_buffer_dirty(src);
1955 btrfs_mark_buffer_dirty(dst);
1957 return ret;
1961 * helper function to insert a new root level in the tree.
1962 * A new node is allocated, and a single item is inserted to
1963 * point to the existing root
1965 * returns zero on success or < 0 on failure.
1967 static noinline int insert_new_root(struct btrfs_trans_handle *trans,
1968 struct btrfs_root *root,
1969 struct btrfs_path *path, int level)
1971 u64 lower_gen;
1972 struct extent_buffer *lower;
1973 struct extent_buffer *c;
1974 struct extent_buffer *old;
1975 struct btrfs_disk_key lower_key;
1977 BUG_ON(path->nodes[level]);
1978 BUG_ON(path->nodes[level-1] != root->node);
1980 lower = path->nodes[level-1];
1981 if (level == 1)
1982 btrfs_item_key(lower, &lower_key, 0);
1983 else
1984 btrfs_node_key(lower, &lower_key, 0);
1986 c = btrfs_alloc_free_block(trans, root, root->nodesize, 0,
1987 root->root_key.objectid, &lower_key,
1988 level, root->node->start, 0);
1989 if (IS_ERR(c))
1990 return PTR_ERR(c);
1992 root_add_used(root, root->nodesize);
1994 memset_extent_buffer(c, 0, 0, sizeof(struct btrfs_header));
1995 btrfs_set_header_nritems(c, 1);
1996 btrfs_set_header_level(c, level);
1997 btrfs_set_header_bytenr(c, c->start);
1998 btrfs_set_header_generation(c, trans->transid);
1999 btrfs_set_header_backref_rev(c, BTRFS_MIXED_BACKREF_REV);
2000 btrfs_set_header_owner(c, root->root_key.objectid);
2002 write_extent_buffer(c, root->fs_info->fsid,
2003 (unsigned long)btrfs_header_fsid(c),
2004 BTRFS_FSID_SIZE);
2006 write_extent_buffer(c, root->fs_info->chunk_tree_uuid,
2007 (unsigned long)btrfs_header_chunk_tree_uuid(c),
2008 BTRFS_UUID_SIZE);
2010 btrfs_set_node_key(c, &lower_key, 0);
2011 btrfs_set_node_blockptr(c, 0, lower->start);
2012 lower_gen = btrfs_header_generation(lower);
2013 WARN_ON(lower_gen != trans->transid);
2015 btrfs_set_node_ptr_generation(c, 0, lower_gen);
2017 btrfs_mark_buffer_dirty(c);
2019 old = root->node;
2020 rcu_assign_pointer(root->node, c);
2022 /* the super has an extra ref to root->node */
2023 free_extent_buffer(old);
2025 add_root_to_dirty_list(root);
2026 extent_buffer_get(c);
2027 path->nodes[level] = c;
2028 path->locks[level] = 1;
2029 path->slots[level] = 0;
2030 return 0;
2034 * worker function to insert a single pointer in a node.
2035 * the node should have enough room for the pointer already
2037 * slot and level indicate where you want the key to go, and
2038 * blocknr is the block the key points to.
2040 * returns zero on success and < 0 on any error
2042 static int insert_ptr(struct btrfs_trans_handle *trans, struct btrfs_root
2043 *root, struct btrfs_path *path, struct btrfs_disk_key
2044 *key, u64 bytenr, int slot, int level)
2046 struct extent_buffer *lower;
2047 int nritems;
2049 BUG_ON(!path->nodes[level]);
2050 btrfs_assert_tree_locked(path->nodes[level]);
2051 lower = path->nodes[level];
2052 nritems = btrfs_header_nritems(lower);
2053 BUG_ON(slot > nritems);
2054 if (nritems == BTRFS_NODEPTRS_PER_BLOCK(root))
2055 BUG();
2056 if (slot != nritems) {
2057 memmove_extent_buffer(lower,
2058 btrfs_node_key_ptr_offset(slot + 1),
2059 btrfs_node_key_ptr_offset(slot),
2060 (nritems - slot) * sizeof(struct btrfs_key_ptr));
2062 btrfs_set_node_key(lower, key, slot);
2063 btrfs_set_node_blockptr(lower, slot, bytenr);
2064 WARN_ON(trans->transid == 0);
2065 btrfs_set_node_ptr_generation(lower, slot, trans->transid);
2066 btrfs_set_header_nritems(lower, nritems + 1);
2067 btrfs_mark_buffer_dirty(lower);
2068 return 0;
2072 * split the node at the specified level in path in two.
2073 * The path is corrected to point to the appropriate node after the split
2075 * Before splitting this tries to make some room in the node by pushing
2076 * left and right, if either one works, it returns right away.
2078 * returns 0 on success and < 0 on failure
2080 static noinline int split_node(struct btrfs_trans_handle *trans,
2081 struct btrfs_root *root,
2082 struct btrfs_path *path, int level)
2084 struct extent_buffer *c;
2085 struct extent_buffer *split;
2086 struct btrfs_disk_key disk_key;
2087 int mid;
2088 int ret;
2089 int wret;
2090 u32 c_nritems;
2092 c = path->nodes[level];
2093 WARN_ON(btrfs_header_generation(c) != trans->transid);
2094 if (c == root->node) {
2095 /* trying to split the root, lets make a new one */
2096 ret = insert_new_root(trans, root, path, level + 1);
2097 if (ret)
2098 return ret;
2099 } else {
2100 ret = push_nodes_for_insert(trans, root, path, level);
2101 c = path->nodes[level];
2102 if (!ret && btrfs_header_nritems(c) <
2103 BTRFS_NODEPTRS_PER_BLOCK(root) - 3)
2104 return 0;
2105 if (ret < 0)
2106 return ret;
2109 c_nritems = btrfs_header_nritems(c);
2110 mid = (c_nritems + 1) / 2;
2111 btrfs_node_key(c, &disk_key, mid);
2113 split = btrfs_alloc_free_block(trans, root, root->nodesize, 0,
2114 root->root_key.objectid,
2115 &disk_key, level, c->start, 0);
2116 if (IS_ERR(split))
2117 return PTR_ERR(split);
2119 root_add_used(root, root->nodesize);
2121 memset_extent_buffer(split, 0, 0, sizeof(struct btrfs_header));
2122 btrfs_set_header_level(split, btrfs_header_level(c));
2123 btrfs_set_header_bytenr(split, split->start);
2124 btrfs_set_header_generation(split, trans->transid);
2125 btrfs_set_header_backref_rev(split, BTRFS_MIXED_BACKREF_REV);
2126 btrfs_set_header_owner(split, root->root_key.objectid);
2127 write_extent_buffer(split, root->fs_info->fsid,
2128 (unsigned long)btrfs_header_fsid(split),
2129 BTRFS_FSID_SIZE);
2130 write_extent_buffer(split, root->fs_info->chunk_tree_uuid,
2131 (unsigned long)btrfs_header_chunk_tree_uuid(split),
2132 BTRFS_UUID_SIZE);
2135 copy_extent_buffer(split, c,
2136 btrfs_node_key_ptr_offset(0),
2137 btrfs_node_key_ptr_offset(mid),
2138 (c_nritems - mid) * sizeof(struct btrfs_key_ptr));
2139 btrfs_set_header_nritems(split, c_nritems - mid);
2140 btrfs_set_header_nritems(c, mid);
2141 ret = 0;
2143 btrfs_mark_buffer_dirty(c);
2144 btrfs_mark_buffer_dirty(split);
2146 wret = insert_ptr(trans, root, path, &disk_key, split->start,
2147 path->slots[level + 1] + 1,
2148 level + 1);
2149 if (wret)
2150 ret = wret;
2152 if (path->slots[level] >= mid) {
2153 path->slots[level] -= mid;
2154 btrfs_tree_unlock(c);
2155 free_extent_buffer(c);
2156 path->nodes[level] = split;
2157 path->slots[level + 1] += 1;
2158 } else {
2159 btrfs_tree_unlock(split);
2160 free_extent_buffer(split);
2162 return ret;
2166 * how many bytes are required to store the items in a leaf. start
2167 * and nr indicate which items in the leaf to check. This totals up the
2168 * space used both by the item structs and the item data
2170 static int leaf_space_used(struct extent_buffer *l, int start, int nr)
2172 int data_len;
2173 int nritems = btrfs_header_nritems(l);
2174 int end = min(nritems, start + nr) - 1;
2176 if (!nr)
2177 return 0;
2178 data_len = btrfs_item_end_nr(l, start);
2179 data_len = data_len - btrfs_item_offset_nr(l, end);
2180 data_len += sizeof(struct btrfs_item) * nr;
2181 WARN_ON(data_len < 0);
2182 return data_len;
2186 * The space between the end of the leaf items and
2187 * the start of the leaf data. IOW, how much room
2188 * the leaf has left for both items and data
2190 noinline int btrfs_leaf_free_space(struct btrfs_root *root,
2191 struct extent_buffer *leaf)
2193 int nritems = btrfs_header_nritems(leaf);
2194 int ret;
2195 ret = BTRFS_LEAF_DATA_SIZE(root) - leaf_space_used(leaf, 0, nritems);
2196 if (ret < 0) {
2197 printk(KERN_CRIT "leaf free space ret %d, leaf data size %lu, "
2198 "used %d nritems %d\n",
2199 ret, (unsigned long) BTRFS_LEAF_DATA_SIZE(root),
2200 leaf_space_used(leaf, 0, nritems), nritems);
2202 return ret;
2206 * min slot controls the lowest index we're willing to push to the
2207 * right. We'll push up to and including min_slot, but no lower
2209 static noinline int __push_leaf_right(struct btrfs_trans_handle *trans,
2210 struct btrfs_root *root,
2211 struct btrfs_path *path,
2212 int data_size, int empty,
2213 struct extent_buffer *right,
2214 int free_space, u32 left_nritems,
2215 u32 min_slot)
2217 struct extent_buffer *left = path->nodes[0];
2218 struct extent_buffer *upper = path->nodes[1];
2219 struct btrfs_disk_key disk_key;
2220 int slot;
2221 u32 i;
2222 int push_space = 0;
2223 int push_items = 0;
2224 struct btrfs_item *item;
2225 u32 nr;
2226 u32 right_nritems;
2227 u32 data_end;
2228 u32 this_item_size;
2230 if (empty)
2231 nr = 0;
2232 else
2233 nr = max_t(u32, 1, min_slot);
2235 if (path->slots[0] >= left_nritems)
2236 push_space += data_size;
2238 slot = path->slots[1];
2239 i = left_nritems - 1;
2240 while (i >= nr) {
2241 item = btrfs_item_nr(left, i);
2243 if (!empty && push_items > 0) {
2244 if (path->slots[0] > i)
2245 break;
2246 if (path->slots[0] == i) {
2247 int space = btrfs_leaf_free_space(root, left);
2248 if (space + push_space * 2 > free_space)
2249 break;
2253 if (path->slots[0] == i)
2254 push_space += data_size;
2256 if (!left->map_token) {
2257 map_extent_buffer(left, (unsigned long)item,
2258 sizeof(struct btrfs_item),
2259 &left->map_token, &left->kaddr,
2260 &left->map_start, &left->map_len,
2261 KM_USER1);
2264 this_item_size = btrfs_item_size(left, item);
2265 if (this_item_size + sizeof(*item) + push_space > free_space)
2266 break;
2268 push_items++;
2269 push_space += this_item_size + sizeof(*item);
2270 if (i == 0)
2271 break;
2272 i--;
2274 if (left->map_token) {
2275 unmap_extent_buffer(left, left->map_token, KM_USER1);
2276 left->map_token = NULL;
2279 if (push_items == 0)
2280 goto out_unlock;
2282 if (!empty && push_items == left_nritems)
2283 WARN_ON(1);
2285 /* push left to right */
2286 right_nritems = btrfs_header_nritems(right);
2288 push_space = btrfs_item_end_nr(left, left_nritems - push_items);
2289 push_space -= leaf_data_end(root, left);
2291 /* make room in the right data area */
2292 data_end = leaf_data_end(root, right);
2293 memmove_extent_buffer(right,
2294 btrfs_leaf_data(right) + data_end - push_space,
2295 btrfs_leaf_data(right) + data_end,
2296 BTRFS_LEAF_DATA_SIZE(root) - data_end);
2298 /* copy from the left data area */
2299 copy_extent_buffer(right, left, btrfs_leaf_data(right) +
2300 BTRFS_LEAF_DATA_SIZE(root) - push_space,
2301 btrfs_leaf_data(left) + leaf_data_end(root, left),
2302 push_space);
2304 memmove_extent_buffer(right, btrfs_item_nr_offset(push_items),
2305 btrfs_item_nr_offset(0),
2306 right_nritems * sizeof(struct btrfs_item));
2308 /* copy the items from left to right */
2309 copy_extent_buffer(right, left, btrfs_item_nr_offset(0),
2310 btrfs_item_nr_offset(left_nritems - push_items),
2311 push_items * sizeof(struct btrfs_item));
2313 /* update the item pointers */
2314 right_nritems += push_items;
2315 btrfs_set_header_nritems(right, right_nritems);
2316 push_space = BTRFS_LEAF_DATA_SIZE(root);
2317 for (i = 0; i < right_nritems; i++) {
2318 item = btrfs_item_nr(right, i);
2319 if (!right->map_token) {
2320 map_extent_buffer(right, (unsigned long)item,
2321 sizeof(struct btrfs_item),
2322 &right->map_token, &right->kaddr,
2323 &right->map_start, &right->map_len,
2324 KM_USER1);
2326 push_space -= btrfs_item_size(right, item);
2327 btrfs_set_item_offset(right, item, push_space);
2330 if (right->map_token) {
2331 unmap_extent_buffer(right, right->map_token, KM_USER1);
2332 right->map_token = NULL;
2334 left_nritems -= push_items;
2335 btrfs_set_header_nritems(left, left_nritems);
2337 if (left_nritems)
2338 btrfs_mark_buffer_dirty(left);
2339 else
2340 clean_tree_block(trans, root, left);
2342 btrfs_mark_buffer_dirty(right);
2344 btrfs_item_key(right, &disk_key, 0);
2345 btrfs_set_node_key(upper, &disk_key, slot + 1);
2346 btrfs_mark_buffer_dirty(upper);
2348 /* then fixup the leaf pointer in the path */
2349 if (path->slots[0] >= left_nritems) {
2350 path->slots[0] -= left_nritems;
2351 if (btrfs_header_nritems(path->nodes[0]) == 0)
2352 clean_tree_block(trans, root, path->nodes[0]);
2353 btrfs_tree_unlock(path->nodes[0]);
2354 free_extent_buffer(path->nodes[0]);
2355 path->nodes[0] = right;
2356 path->slots[1] += 1;
2357 } else {
2358 btrfs_tree_unlock(right);
2359 free_extent_buffer(right);
2361 return 0;
2363 out_unlock:
2364 btrfs_tree_unlock(right);
2365 free_extent_buffer(right);
2366 return 1;
2370 * push some data in the path leaf to the right, trying to free up at
2371 * least data_size bytes. returns zero if the push worked, nonzero otherwise
2373 * returns 1 if the push failed because the other node didn't have enough
2374 * room, 0 if everything worked out and < 0 if there were major errors.
2376 * this will push starting from min_slot to the end of the leaf. It won't
2377 * push any slot lower than min_slot
2379 static int push_leaf_right(struct btrfs_trans_handle *trans, struct btrfs_root
2380 *root, struct btrfs_path *path,
2381 int min_data_size, int data_size,
2382 int empty, u32 min_slot)
2384 struct extent_buffer *left = path->nodes[0];
2385 struct extent_buffer *right;
2386 struct extent_buffer *upper;
2387 int slot;
2388 int free_space;
2389 u32 left_nritems;
2390 int ret;
2392 if (!path->nodes[1])
2393 return 1;
2395 slot = path->slots[1];
2396 upper = path->nodes[1];
2397 if (slot >= btrfs_header_nritems(upper) - 1)
2398 return 1;
2400 btrfs_assert_tree_locked(path->nodes[1]);
2402 right = read_node_slot(root, upper, slot + 1);
2403 if (right == NULL)
2404 return 1;
2406 btrfs_tree_lock(right);
2407 btrfs_set_lock_blocking(right);
2409 free_space = btrfs_leaf_free_space(root, right);
2410 if (free_space < data_size)
2411 goto out_unlock;
2413 /* cow and double check */
2414 ret = btrfs_cow_block(trans, root, right, upper,
2415 slot + 1, &right);
2416 if (ret)
2417 goto out_unlock;
2419 free_space = btrfs_leaf_free_space(root, right);
2420 if (free_space < data_size)
2421 goto out_unlock;
2423 left_nritems = btrfs_header_nritems(left);
2424 if (left_nritems == 0)
2425 goto out_unlock;
2427 return __push_leaf_right(trans, root, path, min_data_size, empty,
2428 right, free_space, left_nritems, min_slot);
2429 out_unlock:
2430 btrfs_tree_unlock(right);
2431 free_extent_buffer(right);
2432 return 1;
2436 * push some data in the path leaf to the left, trying to free up at
2437 * least data_size bytes. returns zero if the push worked, nonzero otherwise
2439 * max_slot can put a limit on how far into the leaf we'll push items. The
2440 * item at 'max_slot' won't be touched. Use (u32)-1 to make us do all the
2441 * items
2443 static noinline int __push_leaf_left(struct btrfs_trans_handle *trans,
2444 struct btrfs_root *root,
2445 struct btrfs_path *path, int data_size,
2446 int empty, struct extent_buffer *left,
2447 int free_space, u32 right_nritems,
2448 u32 max_slot)
2450 struct btrfs_disk_key disk_key;
2451 struct extent_buffer *right = path->nodes[0];
2452 int i;
2453 int push_space = 0;
2454 int push_items = 0;
2455 struct btrfs_item *item;
2456 u32 old_left_nritems;
2457 u32 nr;
2458 int ret = 0;
2459 int wret;
2460 u32 this_item_size;
2461 u32 old_left_item_size;
2463 if (empty)
2464 nr = min(right_nritems, max_slot);
2465 else
2466 nr = min(right_nritems - 1, max_slot);
2468 for (i = 0; i < nr; i++) {
2469 item = btrfs_item_nr(right, i);
2470 if (!right->map_token) {
2471 map_extent_buffer(right, (unsigned long)item,
2472 sizeof(struct btrfs_item),
2473 &right->map_token, &right->kaddr,
2474 &right->map_start, &right->map_len,
2475 KM_USER1);
2478 if (!empty && push_items > 0) {
2479 if (path->slots[0] < i)
2480 break;
2481 if (path->slots[0] == i) {
2482 int space = btrfs_leaf_free_space(root, right);
2483 if (space + push_space * 2 > free_space)
2484 break;
2488 if (path->slots[0] == i)
2489 push_space += data_size;
2491 this_item_size = btrfs_item_size(right, item);
2492 if (this_item_size + sizeof(*item) + push_space > free_space)
2493 break;
2495 push_items++;
2496 push_space += this_item_size + sizeof(*item);
2499 if (right->map_token) {
2500 unmap_extent_buffer(right, right->map_token, KM_USER1);
2501 right->map_token = NULL;
2504 if (push_items == 0) {
2505 ret = 1;
2506 goto out;
2508 if (!empty && push_items == btrfs_header_nritems(right))
2509 WARN_ON(1);
2511 /* push data from right to left */
2512 copy_extent_buffer(left, right,
2513 btrfs_item_nr_offset(btrfs_header_nritems(left)),
2514 btrfs_item_nr_offset(0),
2515 push_items * sizeof(struct btrfs_item));
2517 push_space = BTRFS_LEAF_DATA_SIZE(root) -
2518 btrfs_item_offset_nr(right, push_items - 1);
2520 copy_extent_buffer(left, right, btrfs_leaf_data(left) +
2521 leaf_data_end(root, left) - push_space,
2522 btrfs_leaf_data(right) +
2523 btrfs_item_offset_nr(right, push_items - 1),
2524 push_space);
2525 old_left_nritems = btrfs_header_nritems(left);
2526 BUG_ON(old_left_nritems <= 0);
2528 old_left_item_size = btrfs_item_offset_nr(left, old_left_nritems - 1);
2529 for (i = old_left_nritems; i < old_left_nritems + push_items; i++) {
2530 u32 ioff;
2532 item = btrfs_item_nr(left, i);
2533 if (!left->map_token) {
2534 map_extent_buffer(left, (unsigned long)item,
2535 sizeof(struct btrfs_item),
2536 &left->map_token, &left->kaddr,
2537 &left->map_start, &left->map_len,
2538 KM_USER1);
2541 ioff = btrfs_item_offset(left, item);
2542 btrfs_set_item_offset(left, item,
2543 ioff - (BTRFS_LEAF_DATA_SIZE(root) - old_left_item_size));
2545 btrfs_set_header_nritems(left, old_left_nritems + push_items);
2546 if (left->map_token) {
2547 unmap_extent_buffer(left, left->map_token, KM_USER1);
2548 left->map_token = NULL;
2551 /* fixup right node */
2552 if (push_items > right_nritems) {
2553 printk(KERN_CRIT "push items %d nr %u\n", push_items,
2554 right_nritems);
2555 WARN_ON(1);
2558 if (push_items < right_nritems) {
2559 push_space = btrfs_item_offset_nr(right, push_items - 1) -
2560 leaf_data_end(root, right);
2561 memmove_extent_buffer(right, btrfs_leaf_data(right) +
2562 BTRFS_LEAF_DATA_SIZE(root) - push_space,
2563 btrfs_leaf_data(right) +
2564 leaf_data_end(root, right), push_space);
2566 memmove_extent_buffer(right, btrfs_item_nr_offset(0),
2567 btrfs_item_nr_offset(push_items),
2568 (btrfs_header_nritems(right) - push_items) *
2569 sizeof(struct btrfs_item));
2571 right_nritems -= push_items;
2572 btrfs_set_header_nritems(right, right_nritems);
2573 push_space = BTRFS_LEAF_DATA_SIZE(root);
2574 for (i = 0; i < right_nritems; i++) {
2575 item = btrfs_item_nr(right, i);
2577 if (!right->map_token) {
2578 map_extent_buffer(right, (unsigned long)item,
2579 sizeof(struct btrfs_item),
2580 &right->map_token, &right->kaddr,
2581 &right->map_start, &right->map_len,
2582 KM_USER1);
2585 push_space = push_space - btrfs_item_size(right, item);
2586 btrfs_set_item_offset(right, item, push_space);
2588 if (right->map_token) {
2589 unmap_extent_buffer(right, right->map_token, KM_USER1);
2590 right->map_token = NULL;
2593 btrfs_mark_buffer_dirty(left);
2594 if (right_nritems)
2595 btrfs_mark_buffer_dirty(right);
2596 else
2597 clean_tree_block(trans, root, right);
2599 btrfs_item_key(right, &disk_key, 0);
2600 wret = fixup_low_keys(trans, root, path, &disk_key, 1);
2601 if (wret)
2602 ret = wret;
2604 /* then fixup the leaf pointer in the path */
2605 if (path->slots[0] < push_items) {
2606 path->slots[0] += old_left_nritems;
2607 btrfs_tree_unlock(path->nodes[0]);
2608 free_extent_buffer(path->nodes[0]);
2609 path->nodes[0] = left;
2610 path->slots[1] -= 1;
2611 } else {
2612 btrfs_tree_unlock(left);
2613 free_extent_buffer(left);
2614 path->slots[0] -= push_items;
2616 BUG_ON(path->slots[0] < 0);
2617 return ret;
2618 out:
2619 btrfs_tree_unlock(left);
2620 free_extent_buffer(left);
2621 return ret;
2625 * push some data in the path leaf to the left, trying to free up at
2626 * least data_size bytes. returns zero if the push worked, nonzero otherwise
2628 * max_slot can put a limit on how far into the leaf we'll push items. The
2629 * item at 'max_slot' won't be touched. Use (u32)-1 to make us push all the
2630 * items
2632 static int push_leaf_left(struct btrfs_trans_handle *trans, struct btrfs_root
2633 *root, struct btrfs_path *path, int min_data_size,
2634 int data_size, int empty, u32 max_slot)
2636 struct extent_buffer *right = path->nodes[0];
2637 struct extent_buffer *left;
2638 int slot;
2639 int free_space;
2640 u32 right_nritems;
2641 int ret = 0;
2643 slot = path->slots[1];
2644 if (slot == 0)
2645 return 1;
2646 if (!path->nodes[1])
2647 return 1;
2649 right_nritems = btrfs_header_nritems(right);
2650 if (right_nritems == 0)
2651 return 1;
2653 btrfs_assert_tree_locked(path->nodes[1]);
2655 left = read_node_slot(root, path->nodes[1], slot - 1);
2656 if (left == NULL)
2657 return 1;
2659 btrfs_tree_lock(left);
2660 btrfs_set_lock_blocking(left);
2662 free_space = btrfs_leaf_free_space(root, left);
2663 if (free_space < data_size) {
2664 ret = 1;
2665 goto out;
2668 /* cow and double check */
2669 ret = btrfs_cow_block(trans, root, left,
2670 path->nodes[1], slot - 1, &left);
2671 if (ret) {
2672 /* we hit -ENOSPC, but it isn't fatal here */
2673 ret = 1;
2674 goto out;
2677 free_space = btrfs_leaf_free_space(root, left);
2678 if (free_space < data_size) {
2679 ret = 1;
2680 goto out;
2683 return __push_leaf_left(trans, root, path, min_data_size,
2684 empty, left, free_space, right_nritems,
2685 max_slot);
2686 out:
2687 btrfs_tree_unlock(left);
2688 free_extent_buffer(left);
2689 return ret;
2693 * split the path's leaf in two, making sure there is at least data_size
2694 * available for the resulting leaf level of the path.
2696 * returns 0 if all went well and < 0 on failure.
2698 static noinline int copy_for_split(struct btrfs_trans_handle *trans,
2699 struct btrfs_root *root,
2700 struct btrfs_path *path,
2701 struct extent_buffer *l,
2702 struct extent_buffer *right,
2703 int slot, int mid, int nritems)
2705 int data_copy_size;
2706 int rt_data_off;
2707 int i;
2708 int ret = 0;
2709 int wret;
2710 struct btrfs_disk_key disk_key;
2712 nritems = nritems - mid;
2713 btrfs_set_header_nritems(right, nritems);
2714 data_copy_size = btrfs_item_end_nr(l, mid) - leaf_data_end(root, l);
2716 copy_extent_buffer(right, l, btrfs_item_nr_offset(0),
2717 btrfs_item_nr_offset(mid),
2718 nritems * sizeof(struct btrfs_item));
2720 copy_extent_buffer(right, l,
2721 btrfs_leaf_data(right) + BTRFS_LEAF_DATA_SIZE(root) -
2722 data_copy_size, btrfs_leaf_data(l) +
2723 leaf_data_end(root, l), data_copy_size);
2725 rt_data_off = BTRFS_LEAF_DATA_SIZE(root) -
2726 btrfs_item_end_nr(l, mid);
2728 for (i = 0; i < nritems; i++) {
2729 struct btrfs_item *item = btrfs_item_nr(right, i);
2730 u32 ioff;
2732 if (!right->map_token) {
2733 map_extent_buffer(right, (unsigned long)item,
2734 sizeof(struct btrfs_item),
2735 &right->map_token, &right->kaddr,
2736 &right->map_start, &right->map_len,
2737 KM_USER1);
2740 ioff = btrfs_item_offset(right, item);
2741 btrfs_set_item_offset(right, item, ioff + rt_data_off);
2744 if (right->map_token) {
2745 unmap_extent_buffer(right, right->map_token, KM_USER1);
2746 right->map_token = NULL;
2749 btrfs_set_header_nritems(l, mid);
2750 ret = 0;
2751 btrfs_item_key(right, &disk_key, 0);
2752 wret = insert_ptr(trans, root, path, &disk_key, right->start,
2753 path->slots[1] + 1, 1);
2754 if (wret)
2755 ret = wret;
2757 btrfs_mark_buffer_dirty(right);
2758 btrfs_mark_buffer_dirty(l);
2759 BUG_ON(path->slots[0] != slot);
2761 if (mid <= slot) {
2762 btrfs_tree_unlock(path->nodes[0]);
2763 free_extent_buffer(path->nodes[0]);
2764 path->nodes[0] = right;
2765 path->slots[0] -= mid;
2766 path->slots[1] += 1;
2767 } else {
2768 btrfs_tree_unlock(right);
2769 free_extent_buffer(right);
2772 BUG_ON(path->slots[0] < 0);
2774 return ret;
2778 * double splits happen when we need to insert a big item in the middle
2779 * of a leaf. A double split can leave us with 3 mostly empty leaves:
2780 * leaf: [ slots 0 - N] [ our target ] [ N + 1 - total in leaf ]
2781 * A B C
2783 * We avoid this by trying to push the items on either side of our target
2784 * into the adjacent leaves. If all goes well we can avoid the double split
2785 * completely.
2787 static noinline int push_for_double_split(struct btrfs_trans_handle *trans,
2788 struct btrfs_root *root,
2789 struct btrfs_path *path,
2790 int data_size)
2792 int ret;
2793 int progress = 0;
2794 int slot;
2795 u32 nritems;
2797 slot = path->slots[0];
2800 * try to push all the items after our slot into the
2801 * right leaf
2803 ret = push_leaf_right(trans, root, path, 1, data_size, 0, slot);
2804 if (ret < 0)
2805 return ret;
2807 if (ret == 0)
2808 progress++;
2810 nritems = btrfs_header_nritems(path->nodes[0]);
2812 * our goal is to get our slot at the start or end of a leaf. If
2813 * we've done so we're done
2815 if (path->slots[0] == 0 || path->slots[0] == nritems)
2816 return 0;
2818 if (btrfs_leaf_free_space(root, path->nodes[0]) >= data_size)
2819 return 0;
2821 /* try to push all the items before our slot into the next leaf */
2822 slot = path->slots[0];
2823 ret = push_leaf_left(trans, root, path, 1, data_size, 0, slot);
2824 if (ret < 0)
2825 return ret;
2827 if (ret == 0)
2828 progress++;
2830 if (progress)
2831 return 0;
2832 return 1;
2836 * split the path's leaf in two, making sure there is at least data_size
2837 * available for the resulting leaf level of the path.
2839 * returns 0 if all went well and < 0 on failure.
2841 static noinline int split_leaf(struct btrfs_trans_handle *trans,
2842 struct btrfs_root *root,
2843 struct btrfs_key *ins_key,
2844 struct btrfs_path *path, int data_size,
2845 int extend)
2847 struct btrfs_disk_key disk_key;
2848 struct extent_buffer *l;
2849 u32 nritems;
2850 int mid;
2851 int slot;
2852 struct extent_buffer *right;
2853 int ret = 0;
2854 int wret;
2855 int split;
2856 int num_doubles = 0;
2857 int tried_avoid_double = 0;
2859 l = path->nodes[0];
2860 slot = path->slots[0];
2861 if (extend && data_size + btrfs_item_size_nr(l, slot) +
2862 sizeof(struct btrfs_item) > BTRFS_LEAF_DATA_SIZE(root))
2863 return -EOVERFLOW;
2865 /* first try to make some room by pushing left and right */
2866 if (data_size) {
2867 wret = push_leaf_right(trans, root, path, data_size,
2868 data_size, 0, 0);
2869 if (wret < 0)
2870 return wret;
2871 if (wret) {
2872 wret = push_leaf_left(trans, root, path, data_size,
2873 data_size, 0, (u32)-1);
2874 if (wret < 0)
2875 return wret;
2877 l = path->nodes[0];
2879 /* did the pushes work? */
2880 if (btrfs_leaf_free_space(root, l) >= data_size)
2881 return 0;
2884 if (!path->nodes[1]) {
2885 ret = insert_new_root(trans, root, path, 1);
2886 if (ret)
2887 return ret;
2889 again:
2890 split = 1;
2891 l = path->nodes[0];
2892 slot = path->slots[0];
2893 nritems = btrfs_header_nritems(l);
2894 mid = (nritems + 1) / 2;
2896 if (mid <= slot) {
2897 if (nritems == 1 ||
2898 leaf_space_used(l, mid, nritems - mid) + data_size >
2899 BTRFS_LEAF_DATA_SIZE(root)) {
2900 if (slot >= nritems) {
2901 split = 0;
2902 } else {
2903 mid = slot;
2904 if (mid != nritems &&
2905 leaf_space_used(l, mid, nritems - mid) +
2906 data_size > BTRFS_LEAF_DATA_SIZE(root)) {
2907 if (data_size && !tried_avoid_double)
2908 goto push_for_double;
2909 split = 2;
2913 } else {
2914 if (leaf_space_used(l, 0, mid) + data_size >
2915 BTRFS_LEAF_DATA_SIZE(root)) {
2916 if (!extend && data_size && slot == 0) {
2917 split = 0;
2918 } else if ((extend || !data_size) && slot == 0) {
2919 mid = 1;
2920 } else {
2921 mid = slot;
2922 if (mid != nritems &&
2923 leaf_space_used(l, mid, nritems - mid) +
2924 data_size > BTRFS_LEAF_DATA_SIZE(root)) {
2925 if (data_size && !tried_avoid_double)
2926 goto push_for_double;
2927 split = 2 ;
2933 if (split == 0)
2934 btrfs_cpu_key_to_disk(&disk_key, ins_key);
2935 else
2936 btrfs_item_key(l, &disk_key, mid);
2938 right = btrfs_alloc_free_block(trans, root, root->leafsize, 0,
2939 root->root_key.objectid,
2940 &disk_key, 0, l->start, 0);
2941 if (IS_ERR(right))
2942 return PTR_ERR(right);
2944 root_add_used(root, root->leafsize);
2946 memset_extent_buffer(right, 0, 0, sizeof(struct btrfs_header));
2947 btrfs_set_header_bytenr(right, right->start);
2948 btrfs_set_header_generation(right, trans->transid);
2949 btrfs_set_header_backref_rev(right, BTRFS_MIXED_BACKREF_REV);
2950 btrfs_set_header_owner(right, root->root_key.objectid);
2951 btrfs_set_header_level(right, 0);
2952 write_extent_buffer(right, root->fs_info->fsid,
2953 (unsigned long)btrfs_header_fsid(right),
2954 BTRFS_FSID_SIZE);
2956 write_extent_buffer(right, root->fs_info->chunk_tree_uuid,
2957 (unsigned long)btrfs_header_chunk_tree_uuid(right),
2958 BTRFS_UUID_SIZE);
2960 if (split == 0) {
2961 if (mid <= slot) {
2962 btrfs_set_header_nritems(right, 0);
2963 wret = insert_ptr(trans, root, path,
2964 &disk_key, right->start,
2965 path->slots[1] + 1, 1);
2966 if (wret)
2967 ret = wret;
2969 btrfs_tree_unlock(path->nodes[0]);
2970 free_extent_buffer(path->nodes[0]);
2971 path->nodes[0] = right;
2972 path->slots[0] = 0;
2973 path->slots[1] += 1;
2974 } else {
2975 btrfs_set_header_nritems(right, 0);
2976 wret = insert_ptr(trans, root, path,
2977 &disk_key,
2978 right->start,
2979 path->slots[1], 1);
2980 if (wret)
2981 ret = wret;
2982 btrfs_tree_unlock(path->nodes[0]);
2983 free_extent_buffer(path->nodes[0]);
2984 path->nodes[0] = right;
2985 path->slots[0] = 0;
2986 if (path->slots[1] == 0) {
2987 wret = fixup_low_keys(trans, root,
2988 path, &disk_key, 1);
2989 if (wret)
2990 ret = wret;
2993 btrfs_mark_buffer_dirty(right);
2994 return ret;
2997 ret = copy_for_split(trans, root, path, l, right, slot, mid, nritems);
2998 BUG_ON(ret);
3000 if (split == 2) {
3001 BUG_ON(num_doubles != 0);
3002 num_doubles++;
3003 goto again;
3006 return ret;
3008 push_for_double:
3009 push_for_double_split(trans, root, path, data_size);
3010 tried_avoid_double = 1;
3011 if (btrfs_leaf_free_space(root, path->nodes[0]) >= data_size)
3012 return 0;
3013 goto again;
3016 static noinline int setup_leaf_for_split(struct btrfs_trans_handle *trans,
3017 struct btrfs_root *root,
3018 struct btrfs_path *path, int ins_len)
3020 struct btrfs_key key;
3021 struct extent_buffer *leaf;
3022 struct btrfs_file_extent_item *fi;
3023 u64 extent_len = 0;
3024 u32 item_size;
3025 int ret;
3027 leaf = path->nodes[0];
3028 btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
3030 BUG_ON(key.type != BTRFS_EXTENT_DATA_KEY &&
3031 key.type != BTRFS_EXTENT_CSUM_KEY);
3033 if (btrfs_leaf_free_space(root, leaf) >= ins_len)
3034 return 0;
3036 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
3037 if (key.type == BTRFS_EXTENT_DATA_KEY) {
3038 fi = btrfs_item_ptr(leaf, path->slots[0],
3039 struct btrfs_file_extent_item);
3040 extent_len = btrfs_file_extent_num_bytes(leaf, fi);
3042 btrfs_release_path(path);
3044 path->keep_locks = 1;
3045 path->search_for_split = 1;
3046 ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
3047 path->search_for_split = 0;
3048 if (ret < 0)
3049 goto err;
3051 ret = -EAGAIN;
3052 leaf = path->nodes[0];
3053 /* if our item isn't there or got smaller, return now */
3054 if (ret > 0 || item_size != btrfs_item_size_nr(leaf, path->slots[0]))
3055 goto err;
3057 /* the leaf has changed, it now has room. return now */
3058 if (btrfs_leaf_free_space(root, path->nodes[0]) >= ins_len)
3059 goto err;
3061 if (key.type == BTRFS_EXTENT_DATA_KEY) {
3062 fi = btrfs_item_ptr(leaf, path->slots[0],
3063 struct btrfs_file_extent_item);
3064 if (extent_len != btrfs_file_extent_num_bytes(leaf, fi))
3065 goto err;
3068 btrfs_set_path_blocking(path);
3069 ret = split_leaf(trans, root, &key, path, ins_len, 1);
3070 if (ret)
3071 goto err;
3073 path->keep_locks = 0;
3074 btrfs_unlock_up_safe(path, 1);
3075 return 0;
3076 err:
3077 path->keep_locks = 0;
3078 return ret;
3081 static noinline int split_item(struct btrfs_trans_handle *trans,
3082 struct btrfs_root *root,
3083 struct btrfs_path *path,
3084 struct btrfs_key *new_key,
3085 unsigned long split_offset)
3087 struct extent_buffer *leaf;
3088 struct btrfs_item *item;
3089 struct btrfs_item *new_item;
3090 int slot;
3091 char *buf;
3092 u32 nritems;
3093 u32 item_size;
3094 u32 orig_offset;
3095 struct btrfs_disk_key disk_key;
3097 leaf = path->nodes[0];
3098 BUG_ON(btrfs_leaf_free_space(root, leaf) < sizeof(struct btrfs_item));
3100 btrfs_set_path_blocking(path);
3102 item = btrfs_item_nr(leaf, path->slots[0]);
3103 orig_offset = btrfs_item_offset(leaf, item);
3104 item_size = btrfs_item_size(leaf, item);
3106 buf = kmalloc(item_size, GFP_NOFS);
3107 if (!buf)
3108 return -ENOMEM;
3110 read_extent_buffer(leaf, buf, btrfs_item_ptr_offset(leaf,
3111 path->slots[0]), item_size);
3113 slot = path->slots[0] + 1;
3114 nritems = btrfs_header_nritems(leaf);
3115 if (slot != nritems) {
3116 /* shift the items */
3117 memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + 1),
3118 btrfs_item_nr_offset(slot),
3119 (nritems - slot) * sizeof(struct btrfs_item));
3122 btrfs_cpu_key_to_disk(&disk_key, new_key);
3123 btrfs_set_item_key(leaf, &disk_key, slot);
3125 new_item = btrfs_item_nr(leaf, slot);
3127 btrfs_set_item_offset(leaf, new_item, orig_offset);
3128 btrfs_set_item_size(leaf, new_item, item_size - split_offset);
3130 btrfs_set_item_offset(leaf, item,
3131 orig_offset + item_size - split_offset);
3132 btrfs_set_item_size(leaf, item, split_offset);
3134 btrfs_set_header_nritems(leaf, nritems + 1);
3136 /* write the data for the start of the original item */
3137 write_extent_buffer(leaf, buf,
3138 btrfs_item_ptr_offset(leaf, path->slots[0]),
3139 split_offset);
3141 /* write the data for the new item */
3142 write_extent_buffer(leaf, buf + split_offset,
3143 btrfs_item_ptr_offset(leaf, slot),
3144 item_size - split_offset);
3145 btrfs_mark_buffer_dirty(leaf);
3147 BUG_ON(btrfs_leaf_free_space(root, leaf) < 0);
3148 kfree(buf);
3149 return 0;
3153 * This function splits a single item into two items,
3154 * giving 'new_key' to the new item and splitting the
3155 * old one at split_offset (from the start of the item).
3157 * The path may be released by this operation. After
3158 * the split, the path is pointing to the old item. The
3159 * new item is going to be in the same node as the old one.
3161 * Note, the item being split must be smaller enough to live alone on
3162 * a tree block with room for one extra struct btrfs_item
3164 * This allows us to split the item in place, keeping a lock on the
3165 * leaf the entire time.
3167 int btrfs_split_item(struct btrfs_trans_handle *trans,
3168 struct btrfs_root *root,
3169 struct btrfs_path *path,
3170 struct btrfs_key *new_key,
3171 unsigned long split_offset)
3173 int ret;
3174 ret = setup_leaf_for_split(trans, root, path,
3175 sizeof(struct btrfs_item));
3176 if (ret)
3177 return ret;
3179 ret = split_item(trans, root, path, new_key, split_offset);
3180 return ret;
3184 * This function duplicate a item, giving 'new_key' to the new item.
3185 * It guarantees both items live in the same tree leaf and the new item
3186 * is contiguous with the original item.
3188 * This allows us to split file extent in place, keeping a lock on the
3189 * leaf the entire time.
3191 int btrfs_duplicate_item(struct btrfs_trans_handle *trans,
3192 struct btrfs_root *root,
3193 struct btrfs_path *path,
3194 struct btrfs_key *new_key)
3196 struct extent_buffer *leaf;
3197 int ret;
3198 u32 item_size;
3200 leaf = path->nodes[0];
3201 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
3202 ret = setup_leaf_for_split(trans, root, path,
3203 item_size + sizeof(struct btrfs_item));
3204 if (ret)
3205 return ret;
3207 path->slots[0]++;
3208 ret = setup_items_for_insert(trans, root, path, new_key, &item_size,
3209 item_size, item_size +
3210 sizeof(struct btrfs_item), 1);
3211 BUG_ON(ret);
3213 leaf = path->nodes[0];
3214 memcpy_extent_buffer(leaf,
3215 btrfs_item_ptr_offset(leaf, path->slots[0]),
3216 btrfs_item_ptr_offset(leaf, path->slots[0] - 1),
3217 item_size);
3218 return 0;
3222 * make the item pointed to by the path smaller. new_size indicates
3223 * how small to make it, and from_end tells us if we just chop bytes
3224 * off the end of the item or if we shift the item to chop bytes off
3225 * the front.
3227 int btrfs_truncate_item(struct btrfs_trans_handle *trans,
3228 struct btrfs_root *root,
3229 struct btrfs_path *path,
3230 u32 new_size, int from_end)
3232 int slot;
3233 struct extent_buffer *leaf;
3234 struct btrfs_item *item;
3235 u32 nritems;
3236 unsigned int data_end;
3237 unsigned int old_data_start;
3238 unsigned int old_size;
3239 unsigned int size_diff;
3240 int i;
3242 leaf = path->nodes[0];
3243 slot = path->slots[0];
3245 old_size = btrfs_item_size_nr(leaf, slot);
3246 if (old_size == new_size)
3247 return 0;
3249 nritems = btrfs_header_nritems(leaf);
3250 data_end = leaf_data_end(root, leaf);
3252 old_data_start = btrfs_item_offset_nr(leaf, slot);
3254 size_diff = old_size - new_size;
3256 BUG_ON(slot < 0);
3257 BUG_ON(slot >= nritems);
3260 * item0..itemN ... dataN.offset..dataN.size .. data0.size
3262 /* first correct the data pointers */
3263 for (i = slot; i < nritems; i++) {
3264 u32 ioff;
3265 item = btrfs_item_nr(leaf, i);
3267 if (!leaf->map_token) {
3268 map_extent_buffer(leaf, (unsigned long)item,
3269 sizeof(struct btrfs_item),
3270 &leaf->map_token, &leaf->kaddr,
3271 &leaf->map_start, &leaf->map_len,
3272 KM_USER1);
3275 ioff = btrfs_item_offset(leaf, item);
3276 btrfs_set_item_offset(leaf, item, ioff + size_diff);
3279 if (leaf->map_token) {
3280 unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
3281 leaf->map_token = NULL;
3284 /* shift the data */
3285 if (from_end) {
3286 memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
3287 data_end + size_diff, btrfs_leaf_data(leaf) +
3288 data_end, old_data_start + new_size - data_end);
3289 } else {
3290 struct btrfs_disk_key disk_key;
3291 u64 offset;
3293 btrfs_item_key(leaf, &disk_key, slot);
3295 if (btrfs_disk_key_type(&disk_key) == BTRFS_EXTENT_DATA_KEY) {
3296 unsigned long ptr;
3297 struct btrfs_file_extent_item *fi;
3299 fi = btrfs_item_ptr(leaf, slot,
3300 struct btrfs_file_extent_item);
3301 fi = (struct btrfs_file_extent_item *)(
3302 (unsigned long)fi - size_diff);
3304 if (btrfs_file_extent_type(leaf, fi) ==
3305 BTRFS_FILE_EXTENT_INLINE) {
3306 ptr = btrfs_item_ptr_offset(leaf, slot);
3307 memmove_extent_buffer(leaf, ptr,
3308 (unsigned long)fi,
3309 offsetof(struct btrfs_file_extent_item,
3310 disk_bytenr));
3314 memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
3315 data_end + size_diff, btrfs_leaf_data(leaf) +
3316 data_end, old_data_start - data_end);
3318 offset = btrfs_disk_key_offset(&disk_key);
3319 btrfs_set_disk_key_offset(&disk_key, offset + size_diff);
3320 btrfs_set_item_key(leaf, &disk_key, slot);
3321 if (slot == 0)
3322 fixup_low_keys(trans, root, path, &disk_key, 1);
3325 item = btrfs_item_nr(leaf, slot);
3326 btrfs_set_item_size(leaf, item, new_size);
3327 btrfs_mark_buffer_dirty(leaf);
3329 if (btrfs_leaf_free_space(root, leaf) < 0) {
3330 btrfs_print_leaf(root, leaf);
3331 BUG();
3333 return 0;
3337 * make the item pointed to by the path bigger, data_size is the new size.
3339 int btrfs_extend_item(struct btrfs_trans_handle *trans,
3340 struct btrfs_root *root, struct btrfs_path *path,
3341 u32 data_size)
3343 int slot;
3344 struct extent_buffer *leaf;
3345 struct btrfs_item *item;
3346 u32 nritems;
3347 unsigned int data_end;
3348 unsigned int old_data;
3349 unsigned int old_size;
3350 int i;
3352 leaf = path->nodes[0];
3354 nritems = btrfs_header_nritems(leaf);
3355 data_end = leaf_data_end(root, leaf);
3357 if (btrfs_leaf_free_space(root, leaf) < data_size) {
3358 btrfs_print_leaf(root, leaf);
3359 BUG();
3361 slot = path->slots[0];
3362 old_data = btrfs_item_end_nr(leaf, slot);
3364 BUG_ON(slot < 0);
3365 if (slot >= nritems) {
3366 btrfs_print_leaf(root, leaf);
3367 printk(KERN_CRIT "slot %d too large, nritems %d\n",
3368 slot, nritems);
3369 BUG_ON(1);
3373 * item0..itemN ... dataN.offset..dataN.size .. data0.size
3375 /* first correct the data pointers */
3376 for (i = slot; i < nritems; i++) {
3377 u32 ioff;
3378 item = btrfs_item_nr(leaf, i);
3380 if (!leaf->map_token) {
3381 map_extent_buffer(leaf, (unsigned long)item,
3382 sizeof(struct btrfs_item),
3383 &leaf->map_token, &leaf->kaddr,
3384 &leaf->map_start, &leaf->map_len,
3385 KM_USER1);
3387 ioff = btrfs_item_offset(leaf, item);
3388 btrfs_set_item_offset(leaf, item, ioff - data_size);
3391 if (leaf->map_token) {
3392 unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
3393 leaf->map_token = NULL;
3396 /* shift the data */
3397 memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
3398 data_end - data_size, btrfs_leaf_data(leaf) +
3399 data_end, old_data - data_end);
3401 data_end = old_data;
3402 old_size = btrfs_item_size_nr(leaf, slot);
3403 item = btrfs_item_nr(leaf, slot);
3404 btrfs_set_item_size(leaf, item, old_size + data_size);
3405 btrfs_mark_buffer_dirty(leaf);
3407 if (btrfs_leaf_free_space(root, leaf) < 0) {
3408 btrfs_print_leaf(root, leaf);
3409 BUG();
3411 return 0;
3415 * Given a key and some data, insert items into the tree.
3416 * This does all the path init required, making room in the tree if needed.
3417 * Returns the number of keys that were inserted.
3419 int btrfs_insert_some_items(struct btrfs_trans_handle *trans,
3420 struct btrfs_root *root,
3421 struct btrfs_path *path,
3422 struct btrfs_key *cpu_key, u32 *data_size,
3423 int nr)
3425 struct extent_buffer *leaf;
3426 struct btrfs_item *item;
3427 int ret = 0;
3428 int slot;
3429 int i;
3430 u32 nritems;
3431 u32 total_data = 0;
3432 u32 total_size = 0;
3433 unsigned int data_end;
3434 struct btrfs_disk_key disk_key;
3435 struct btrfs_key found_key;
3437 for (i = 0; i < nr; i++) {
3438 if (total_size + data_size[i] + sizeof(struct btrfs_item) >
3439 BTRFS_LEAF_DATA_SIZE(root)) {
3440 break;
3441 nr = i;
3443 total_data += data_size[i];
3444 total_size += data_size[i] + sizeof(struct btrfs_item);
3446 BUG_ON(nr == 0);
3448 ret = btrfs_search_slot(trans, root, cpu_key, path, total_size, 1);
3449 if (ret == 0)
3450 return -EEXIST;
3451 if (ret < 0)
3452 goto out;
3454 leaf = path->nodes[0];
3456 nritems = btrfs_header_nritems(leaf);
3457 data_end = leaf_data_end(root, leaf);
3459 if (btrfs_leaf_free_space(root, leaf) < total_size) {
3460 for (i = nr; i >= 0; i--) {
3461 total_data -= data_size[i];
3462 total_size -= data_size[i] + sizeof(struct btrfs_item);
3463 if (total_size < btrfs_leaf_free_space(root, leaf))
3464 break;
3466 nr = i;
3469 slot = path->slots[0];
3470 BUG_ON(slot < 0);
3472 if (slot != nritems) {
3473 unsigned int old_data = btrfs_item_end_nr(leaf, slot);
3475 item = btrfs_item_nr(leaf, slot);
3476 btrfs_item_key_to_cpu(leaf, &found_key, slot);
3478 /* figure out how many keys we can insert in here */
3479 total_data = data_size[0];
3480 for (i = 1; i < nr; i++) {
3481 if (btrfs_comp_cpu_keys(&found_key, cpu_key + i) <= 0)
3482 break;
3483 total_data += data_size[i];
3485 nr = i;
3487 if (old_data < data_end) {
3488 btrfs_print_leaf(root, leaf);
3489 printk(KERN_CRIT "slot %d old_data %d data_end %d\n",
3490 slot, old_data, data_end);
3491 BUG_ON(1);
3494 * item0..itemN ... dataN.offset..dataN.size .. data0.size
3496 /* first correct the data pointers */
3497 WARN_ON(leaf->map_token);
3498 for (i = slot; i < nritems; i++) {
3499 u32 ioff;
3501 item = btrfs_item_nr(leaf, i);
3502 if (!leaf->map_token) {
3503 map_extent_buffer(leaf, (unsigned long)item,
3504 sizeof(struct btrfs_item),
3505 &leaf->map_token, &leaf->kaddr,
3506 &leaf->map_start, &leaf->map_len,
3507 KM_USER1);
3510 ioff = btrfs_item_offset(leaf, item);
3511 btrfs_set_item_offset(leaf, item, ioff - total_data);
3513 if (leaf->map_token) {
3514 unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
3515 leaf->map_token = NULL;
3518 /* shift the items */
3519 memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr),
3520 btrfs_item_nr_offset(slot),
3521 (nritems - slot) * sizeof(struct btrfs_item));
3523 /* shift the data */
3524 memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
3525 data_end - total_data, btrfs_leaf_data(leaf) +
3526 data_end, old_data - data_end);
3527 data_end = old_data;
3528 } else {
3530 * this sucks but it has to be done, if we are inserting at
3531 * the end of the leaf only insert 1 of the items, since we
3532 * have no way of knowing whats on the next leaf and we'd have
3533 * to drop our current locks to figure it out
3535 nr = 1;
3538 /* setup the item for the new data */
3539 for (i = 0; i < nr; i++) {
3540 btrfs_cpu_key_to_disk(&disk_key, cpu_key + i);
3541 btrfs_set_item_key(leaf, &disk_key, slot + i);
3542 item = btrfs_item_nr(leaf, slot + i);
3543 btrfs_set_item_offset(leaf, item, data_end - data_size[i]);
3544 data_end -= data_size[i];
3545 btrfs_set_item_size(leaf, item, data_size[i]);
3547 btrfs_set_header_nritems(leaf, nritems + nr);
3548 btrfs_mark_buffer_dirty(leaf);
3550 ret = 0;
3551 if (slot == 0) {
3552 btrfs_cpu_key_to_disk(&disk_key, cpu_key);
3553 ret = fixup_low_keys(trans, root, path, &disk_key, 1);
3556 if (btrfs_leaf_free_space(root, leaf) < 0) {
3557 btrfs_print_leaf(root, leaf);
3558 BUG();
3560 out:
3561 if (!ret)
3562 ret = nr;
3563 return ret;
3567 * this is a helper for btrfs_insert_empty_items, the main goal here is
3568 * to save stack depth by doing the bulk of the work in a function
3569 * that doesn't call btrfs_search_slot
3571 int setup_items_for_insert(struct btrfs_trans_handle *trans,
3572 struct btrfs_root *root, struct btrfs_path *path,
3573 struct btrfs_key *cpu_key, u32 *data_size,
3574 u32 total_data, u32 total_size, int nr)
3576 struct btrfs_item *item;
3577 int i;
3578 u32 nritems;
3579 unsigned int data_end;
3580 struct btrfs_disk_key disk_key;
3581 int ret;
3582 struct extent_buffer *leaf;
3583 int slot;
3585 leaf = path->nodes[0];
3586 slot = path->slots[0];
3588 nritems = btrfs_header_nritems(leaf);
3589 data_end = leaf_data_end(root, leaf);
3591 if (btrfs_leaf_free_space(root, leaf) < total_size) {
3592 btrfs_print_leaf(root, leaf);
3593 printk(KERN_CRIT "not enough freespace need %u have %d\n",
3594 total_size, btrfs_leaf_free_space(root, leaf));
3595 BUG();
3598 if (slot != nritems) {
3599 unsigned int old_data = btrfs_item_end_nr(leaf, slot);
3601 if (old_data < data_end) {
3602 btrfs_print_leaf(root, leaf);
3603 printk(KERN_CRIT "slot %d old_data %d data_end %d\n",
3604 slot, old_data, data_end);
3605 BUG_ON(1);
3608 * item0..itemN ... dataN.offset..dataN.size .. data0.size
3610 /* first correct the data pointers */
3611 WARN_ON(leaf->map_token);
3612 for (i = slot; i < nritems; i++) {
3613 u32 ioff;
3615 item = btrfs_item_nr(leaf, i);
3616 if (!leaf->map_token) {
3617 map_extent_buffer(leaf, (unsigned long)item,
3618 sizeof(struct btrfs_item),
3619 &leaf->map_token, &leaf->kaddr,
3620 &leaf->map_start, &leaf->map_len,
3621 KM_USER1);
3624 ioff = btrfs_item_offset(leaf, item);
3625 btrfs_set_item_offset(leaf, item, ioff - total_data);
3627 if (leaf->map_token) {
3628 unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
3629 leaf->map_token = NULL;
3632 /* shift the items */
3633 memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr),
3634 btrfs_item_nr_offset(slot),
3635 (nritems - slot) * sizeof(struct btrfs_item));
3637 /* shift the data */
3638 memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
3639 data_end - total_data, btrfs_leaf_data(leaf) +
3640 data_end, old_data - data_end);
3641 data_end = old_data;
3644 /* setup the item for the new data */
3645 for (i = 0; i < nr; i++) {
3646 btrfs_cpu_key_to_disk(&disk_key, cpu_key + i);
3647 btrfs_set_item_key(leaf, &disk_key, slot + i);
3648 item = btrfs_item_nr(leaf, slot + i);
3649 btrfs_set_item_offset(leaf, item, data_end - data_size[i]);
3650 data_end -= data_size[i];
3651 btrfs_set_item_size(leaf, item, data_size[i]);
3654 btrfs_set_header_nritems(leaf, nritems + nr);
3656 ret = 0;
3657 if (slot == 0) {
3658 btrfs_cpu_key_to_disk(&disk_key, cpu_key);
3659 ret = fixup_low_keys(trans, root, path, &disk_key, 1);
3661 btrfs_unlock_up_safe(path, 1);
3662 btrfs_mark_buffer_dirty(leaf);
3664 if (btrfs_leaf_free_space(root, leaf) < 0) {
3665 btrfs_print_leaf(root, leaf);
3666 BUG();
3668 return ret;
3672 * Given a key and some data, insert items into the tree.
3673 * This does all the path init required, making room in the tree if needed.
3675 int btrfs_insert_empty_items(struct btrfs_trans_handle *trans,
3676 struct btrfs_root *root,
3677 struct btrfs_path *path,
3678 struct btrfs_key *cpu_key, u32 *data_size,
3679 int nr)
3681 int ret = 0;
3682 int slot;
3683 int i;
3684 u32 total_size = 0;
3685 u32 total_data = 0;
3687 for (i = 0; i < nr; i++)
3688 total_data += data_size[i];
3690 total_size = total_data + (nr * sizeof(struct btrfs_item));
3691 ret = btrfs_search_slot(trans, root, cpu_key, path, total_size, 1);
3692 if (ret == 0)
3693 return -EEXIST;
3694 if (ret < 0)
3695 goto out;
3697 slot = path->slots[0];
3698 BUG_ON(slot < 0);
3700 ret = setup_items_for_insert(trans, root, path, cpu_key, data_size,
3701 total_data, total_size, nr);
3703 out:
3704 return ret;
3708 * Given a key and some data, insert an item into the tree.
3709 * This does all the path init required, making room in the tree if needed.
3711 int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root
3712 *root, struct btrfs_key *cpu_key, void *data, u32
3713 data_size)
3715 int ret = 0;
3716 struct btrfs_path *path;
3717 struct extent_buffer *leaf;
3718 unsigned long ptr;
3720 path = btrfs_alloc_path();
3721 if (!path)
3722 return -ENOMEM;
3723 ret = btrfs_insert_empty_item(trans, root, path, cpu_key, data_size);
3724 if (!ret) {
3725 leaf = path->nodes[0];
3726 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
3727 write_extent_buffer(leaf, data, ptr, data_size);
3728 btrfs_mark_buffer_dirty(leaf);
3730 btrfs_free_path(path);
3731 return ret;
3735 * delete the pointer from a given node.
3737 * the tree should have been previously balanced so the deletion does not
3738 * empty a node.
3740 static int del_ptr(struct btrfs_trans_handle *trans, struct btrfs_root *root,
3741 struct btrfs_path *path, int level, int slot)
3743 struct extent_buffer *parent = path->nodes[level];
3744 u32 nritems;
3745 int ret = 0;
3746 int wret;
3748 nritems = btrfs_header_nritems(parent);
3749 if (slot != nritems - 1) {
3750 memmove_extent_buffer(parent,
3751 btrfs_node_key_ptr_offset(slot),
3752 btrfs_node_key_ptr_offset(slot + 1),
3753 sizeof(struct btrfs_key_ptr) *
3754 (nritems - slot - 1));
3756 nritems--;
3757 btrfs_set_header_nritems(parent, nritems);
3758 if (nritems == 0 && parent == root->node) {
3759 BUG_ON(btrfs_header_level(root->node) != 1);
3760 /* just turn the root into a leaf and break */
3761 btrfs_set_header_level(root->node, 0);
3762 } else if (slot == 0) {
3763 struct btrfs_disk_key disk_key;
3765 btrfs_node_key(parent, &disk_key, 0);
3766 wret = fixup_low_keys(trans, root, path, &disk_key, level + 1);
3767 if (wret)
3768 ret = wret;
3770 btrfs_mark_buffer_dirty(parent);
3771 return ret;
3775 * a helper function to delete the leaf pointed to by path->slots[1] and
3776 * path->nodes[1].
3778 * This deletes the pointer in path->nodes[1] and frees the leaf
3779 * block extent. zero is returned if it all worked out, < 0 otherwise.
3781 * The path must have already been setup for deleting the leaf, including
3782 * all the proper balancing. path->nodes[1] must be locked.
3784 static noinline int btrfs_del_leaf(struct btrfs_trans_handle *trans,
3785 struct btrfs_root *root,
3786 struct btrfs_path *path,
3787 struct extent_buffer *leaf)
3789 int ret;
3791 WARN_ON(btrfs_header_generation(leaf) != trans->transid);
3792 ret = del_ptr(trans, root, path, 1, path->slots[1]);
3793 if (ret)
3794 return ret;
3797 * btrfs_free_extent is expensive, we want to make sure we
3798 * aren't holding any locks when we call it
3800 btrfs_unlock_up_safe(path, 0);
3802 root_sub_used(root, leaf->len);
3804 btrfs_free_tree_block(trans, root, leaf, 0, 1);
3805 return 0;
3808 * delete the item at the leaf level in path. If that empties
3809 * the leaf, remove it from the tree
3811 int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
3812 struct btrfs_path *path, int slot, int nr)
3814 struct extent_buffer *leaf;
3815 struct btrfs_item *item;
3816 int last_off;
3817 int dsize = 0;
3818 int ret = 0;
3819 int wret;
3820 int i;
3821 u32 nritems;
3823 leaf = path->nodes[0];
3824 last_off = btrfs_item_offset_nr(leaf, slot + nr - 1);
3826 for (i = 0; i < nr; i++)
3827 dsize += btrfs_item_size_nr(leaf, slot + i);
3829 nritems = btrfs_header_nritems(leaf);
3831 if (slot + nr != nritems) {
3832 int data_end = leaf_data_end(root, leaf);
3834 memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) +
3835 data_end + dsize,
3836 btrfs_leaf_data(leaf) + data_end,
3837 last_off - data_end);
3839 for (i = slot + nr; i < nritems; i++) {
3840 u32 ioff;
3842 item = btrfs_item_nr(leaf, i);
3843 if (!leaf->map_token) {
3844 map_extent_buffer(leaf, (unsigned long)item,
3845 sizeof(struct btrfs_item),
3846 &leaf->map_token, &leaf->kaddr,
3847 &leaf->map_start, &leaf->map_len,
3848 KM_USER1);
3850 ioff = btrfs_item_offset(leaf, item);
3851 btrfs_set_item_offset(leaf, item, ioff + dsize);
3854 if (leaf->map_token) {
3855 unmap_extent_buffer(leaf, leaf->map_token, KM_USER1);
3856 leaf->map_token = NULL;
3859 memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot),
3860 btrfs_item_nr_offset(slot + nr),
3861 sizeof(struct btrfs_item) *
3862 (nritems - slot - nr));
3864 btrfs_set_header_nritems(leaf, nritems - nr);
3865 nritems -= nr;
3867 /* delete the leaf if we've emptied it */
3868 if (nritems == 0) {
3869 if (leaf == root->node) {
3870 btrfs_set_header_level(leaf, 0);
3871 } else {
3872 btrfs_set_path_blocking(path);
3873 clean_tree_block(trans, root, leaf);
3874 ret = btrfs_del_leaf(trans, root, path, leaf);
3875 BUG_ON(ret);
3877 } else {
3878 int used = leaf_space_used(leaf, 0, nritems);
3879 if (slot == 0) {
3880 struct btrfs_disk_key disk_key;
3882 btrfs_item_key(leaf, &disk_key, 0);
3883 wret = fixup_low_keys(trans, root, path,
3884 &disk_key, 1);
3885 if (wret)
3886 ret = wret;
3889 /* delete the leaf if it is mostly empty */
3890 if (used < BTRFS_LEAF_DATA_SIZE(root) / 3) {
3891 /* push_leaf_left fixes the path.
3892 * make sure the path still points to our leaf
3893 * for possible call to del_ptr below
3895 slot = path->slots[1];
3896 extent_buffer_get(leaf);
3898 btrfs_set_path_blocking(path);
3899 wret = push_leaf_left(trans, root, path, 1, 1,
3900 1, (u32)-1);
3901 if (wret < 0 && wret != -ENOSPC)
3902 ret = wret;
3904 if (path->nodes[0] == leaf &&
3905 btrfs_header_nritems(leaf)) {
3906 wret = push_leaf_right(trans, root, path, 1,
3907 1, 1, 0);
3908 if (wret < 0 && wret != -ENOSPC)
3909 ret = wret;
3912 if (btrfs_header_nritems(leaf) == 0) {
3913 path->slots[1] = slot;
3914 ret = btrfs_del_leaf(trans, root, path, leaf);
3915 BUG_ON(ret);
3916 free_extent_buffer(leaf);
3917 } else {
3918 /* if we're still in the path, make sure
3919 * we're dirty. Otherwise, one of the
3920 * push_leaf functions must have already
3921 * dirtied this buffer
3923 if (path->nodes[0] == leaf)
3924 btrfs_mark_buffer_dirty(leaf);
3925 free_extent_buffer(leaf);
3927 } else {
3928 btrfs_mark_buffer_dirty(leaf);
3931 return ret;
3935 * search the tree again to find a leaf with lesser keys
3936 * returns 0 if it found something or 1 if there are no lesser leaves.
3937 * returns < 0 on io errors.
3939 * This may release the path, and so you may lose any locks held at the
3940 * time you call it.
3942 int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path)
3944 struct btrfs_key key;
3945 struct btrfs_disk_key found_key;
3946 int ret;
3948 btrfs_item_key_to_cpu(path->nodes[0], &key, 0);
3950 if (key.offset > 0)
3951 key.offset--;
3952 else if (key.type > 0)
3953 key.type--;
3954 else if (key.objectid > 0)
3955 key.objectid--;
3956 else
3957 return 1;
3959 btrfs_release_path(path);
3960 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
3961 if (ret < 0)
3962 return ret;
3963 btrfs_item_key(path->nodes[0], &found_key, 0);
3964 ret = comp_keys(&found_key, &key);
3965 if (ret < 0)
3966 return 0;
3967 return 1;
3971 * A helper function to walk down the tree starting at min_key, and looking
3972 * for nodes or leaves that are either in cache or have a minimum
3973 * transaction id. This is used by the btree defrag code, and tree logging
3975 * This does not cow, but it does stuff the starting key it finds back
3976 * into min_key, so you can call btrfs_search_slot with cow=1 on the
3977 * key and get a writable path.
3979 * This does lock as it descends, and path->keep_locks should be set
3980 * to 1 by the caller.
3982 * This honors path->lowest_level to prevent descent past a given level
3983 * of the tree.
3985 * min_trans indicates the oldest transaction that you are interested
3986 * in walking through. Any nodes or leaves older than min_trans are
3987 * skipped over (without reading them).
3989 * returns zero if something useful was found, < 0 on error and 1 if there
3990 * was nothing in the tree that matched the search criteria.
3992 int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key,
3993 struct btrfs_key *max_key,
3994 struct btrfs_path *path, int cache_only,
3995 u64 min_trans)
3997 struct extent_buffer *cur;
3998 struct btrfs_key found_key;
3999 int slot;
4000 int sret;
4001 u32 nritems;
4002 int level;
4003 int ret = 1;
4005 WARN_ON(!path->keep_locks);
4006 again:
4007 cur = btrfs_lock_root_node(root);
4008 level = btrfs_header_level(cur);
4009 WARN_ON(path->nodes[level]);
4010 path->nodes[level] = cur;
4011 path->locks[level] = 1;
4013 if (btrfs_header_generation(cur) < min_trans) {
4014 ret = 1;
4015 goto out;
4017 while (1) {
4018 nritems = btrfs_header_nritems(cur);
4019 level = btrfs_header_level(cur);
4020 sret = bin_search(cur, min_key, level, &slot);
4022 /* at the lowest level, we're done, setup the path and exit */
4023 if (level == path->lowest_level) {
4024 if (slot >= nritems)
4025 goto find_next_key;
4026 ret = 0;
4027 path->slots[level] = slot;
4028 btrfs_item_key_to_cpu(cur, &found_key, slot);
4029 goto out;
4031 if (sret && slot > 0)
4032 slot--;
4034 * check this node pointer against the cache_only and
4035 * min_trans parameters. If it isn't in cache or is too
4036 * old, skip to the next one.
4038 while (slot < nritems) {
4039 u64 blockptr;
4040 u64 gen;
4041 struct extent_buffer *tmp;
4042 struct btrfs_disk_key disk_key;
4044 blockptr = btrfs_node_blockptr(cur, slot);
4045 gen = btrfs_node_ptr_generation(cur, slot);
4046 if (gen < min_trans) {
4047 slot++;
4048 continue;
4050 if (!cache_only)
4051 break;
4053 if (max_key) {
4054 btrfs_node_key(cur, &disk_key, slot);
4055 if (comp_keys(&disk_key, max_key) >= 0) {
4056 ret = 1;
4057 goto out;
4061 tmp = btrfs_find_tree_block(root, blockptr,
4062 btrfs_level_size(root, level - 1));
4064 if (tmp && btrfs_buffer_uptodate(tmp, gen)) {
4065 free_extent_buffer(tmp);
4066 break;
4068 if (tmp)
4069 free_extent_buffer(tmp);
4070 slot++;
4072 find_next_key:
4074 * we didn't find a candidate key in this node, walk forward
4075 * and find another one
4077 if (slot >= nritems) {
4078 path->slots[level] = slot;
4079 btrfs_set_path_blocking(path);
4080 sret = btrfs_find_next_key(root, path, min_key, level,
4081 cache_only, min_trans);
4082 if (sret == 0) {
4083 btrfs_release_path(path);
4084 goto again;
4085 } else {
4086 goto out;
4089 /* save our key for returning back */
4090 btrfs_node_key_to_cpu(cur, &found_key, slot);
4091 path->slots[level] = slot;
4092 if (level == path->lowest_level) {
4093 ret = 0;
4094 unlock_up(path, level, 1);
4095 goto out;
4097 btrfs_set_path_blocking(path);
4098 cur = read_node_slot(root, cur, slot);
4099 BUG_ON(!cur);
4101 btrfs_tree_lock(cur);
4103 path->locks[level - 1] = 1;
4104 path->nodes[level - 1] = cur;
4105 unlock_up(path, level, 1);
4106 btrfs_clear_path_blocking(path, NULL);
4108 out:
4109 if (ret == 0)
4110 memcpy(min_key, &found_key, sizeof(found_key));
4111 btrfs_set_path_blocking(path);
4112 return ret;
4116 * this is similar to btrfs_next_leaf, but does not try to preserve
4117 * and fixup the path. It looks for and returns the next key in the
4118 * tree based on the current path and the cache_only and min_trans
4119 * parameters.
4121 * 0 is returned if another key is found, < 0 if there are any errors
4122 * and 1 is returned if there are no higher keys in the tree
4124 * path->keep_locks should be set to 1 on the search made before
4125 * calling this function.
4127 int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path,
4128 struct btrfs_key *key, int level,
4129 int cache_only, u64 min_trans)
4131 int slot;
4132 struct extent_buffer *c;
4134 WARN_ON(!path->keep_locks);
4135 while (level < BTRFS_MAX_LEVEL) {
4136 if (!path->nodes[level])
4137 return 1;
4139 slot = path->slots[level] + 1;
4140 c = path->nodes[level];
4141 next:
4142 if (slot >= btrfs_header_nritems(c)) {
4143 int ret;
4144 int orig_lowest;
4145 struct btrfs_key cur_key;
4146 if (level + 1 >= BTRFS_MAX_LEVEL ||
4147 !path->nodes[level + 1])
4148 return 1;
4150 if (path->locks[level + 1]) {
4151 level++;
4152 continue;
4155 slot = btrfs_header_nritems(c) - 1;
4156 if (level == 0)
4157 btrfs_item_key_to_cpu(c, &cur_key, slot);
4158 else
4159 btrfs_node_key_to_cpu(c, &cur_key, slot);
4161 orig_lowest = path->lowest_level;
4162 btrfs_release_path(path);
4163 path->lowest_level = level;
4164 ret = btrfs_search_slot(NULL, root, &cur_key, path,
4165 0, 0);
4166 path->lowest_level = orig_lowest;
4167 if (ret < 0)
4168 return ret;
4170 c = path->nodes[level];
4171 slot = path->slots[level];
4172 if (ret == 0)
4173 slot++;
4174 goto next;
4177 if (level == 0)
4178 btrfs_item_key_to_cpu(c, key, slot);
4179 else {
4180 u64 blockptr = btrfs_node_blockptr(c, slot);
4181 u64 gen = btrfs_node_ptr_generation(c, slot);
4183 if (cache_only) {
4184 struct extent_buffer *cur;
4185 cur = btrfs_find_tree_block(root, blockptr,
4186 btrfs_level_size(root, level - 1));
4187 if (!cur || !btrfs_buffer_uptodate(cur, gen)) {
4188 slot++;
4189 if (cur)
4190 free_extent_buffer(cur);
4191 goto next;
4193 free_extent_buffer(cur);
4195 if (gen < min_trans) {
4196 slot++;
4197 goto next;
4199 btrfs_node_key_to_cpu(c, key, slot);
4201 return 0;
4203 return 1;
4207 * search the tree again to find a leaf with greater keys
4208 * returns 0 if it found something or 1 if there are no greater leaves.
4209 * returns < 0 on io errors.
4211 int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path)
4213 int slot;
4214 int level;
4215 struct extent_buffer *c;
4216 struct extent_buffer *next;
4217 struct btrfs_key key;
4218 u32 nritems;
4219 int ret;
4220 int old_spinning = path->leave_spinning;
4221 int force_blocking = 0;
4223 nritems = btrfs_header_nritems(path->nodes[0]);
4224 if (nritems == 0)
4225 return 1;
4228 * we take the blocks in an order that upsets lockdep. Using
4229 * blocking mode is the only way around it.
4231 #ifdef CONFIG_DEBUG_LOCK_ALLOC
4232 force_blocking = 1;
4233 #endif
4235 btrfs_item_key_to_cpu(path->nodes[0], &key, nritems - 1);
4236 again:
4237 level = 1;
4238 next = NULL;
4239 btrfs_release_path(path);
4241 path->keep_locks = 1;
4243 if (!force_blocking)
4244 path->leave_spinning = 1;
4246 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
4247 path->keep_locks = 0;
4249 if (ret < 0)
4250 return ret;
4252 nritems = btrfs_header_nritems(path->nodes[0]);
4254 * by releasing the path above we dropped all our locks. A balance
4255 * could have added more items next to the key that used to be
4256 * at the very end of the block. So, check again here and
4257 * advance the path if there are now more items available.
4259 if (nritems > 0 && path->slots[0] < nritems - 1) {
4260 if (ret == 0)
4261 path->slots[0]++;
4262 ret = 0;
4263 goto done;
4266 while (level < BTRFS_MAX_LEVEL) {
4267 if (!path->nodes[level]) {
4268 ret = 1;
4269 goto done;
4272 slot = path->slots[level] + 1;
4273 c = path->nodes[level];
4274 if (slot >= btrfs_header_nritems(c)) {
4275 level++;
4276 if (level == BTRFS_MAX_LEVEL) {
4277 ret = 1;
4278 goto done;
4280 continue;
4283 if (next) {
4284 btrfs_tree_unlock(next);
4285 free_extent_buffer(next);
4288 next = c;
4289 ret = read_block_for_search(NULL, root, path, &next, level,
4290 slot, &key);
4291 if (ret == -EAGAIN)
4292 goto again;
4294 if (ret < 0) {
4295 btrfs_release_path(path);
4296 goto done;
4299 if (!path->skip_locking) {
4300 ret = btrfs_try_spin_lock(next);
4301 if (!ret) {
4302 btrfs_set_path_blocking(path);
4303 btrfs_tree_lock(next);
4304 if (!force_blocking)
4305 btrfs_clear_path_blocking(path, next);
4307 if (force_blocking)
4308 btrfs_set_lock_blocking(next);
4310 break;
4312 path->slots[level] = slot;
4313 while (1) {
4314 level--;
4315 c = path->nodes[level];
4316 if (path->locks[level])
4317 btrfs_tree_unlock(c);
4319 free_extent_buffer(c);
4320 path->nodes[level] = next;
4321 path->slots[level] = 0;
4322 if (!path->skip_locking)
4323 path->locks[level] = 1;
4325 if (!level)
4326 break;
4328 ret = read_block_for_search(NULL, root, path, &next, level,
4329 0, &key);
4330 if (ret == -EAGAIN)
4331 goto again;
4333 if (ret < 0) {
4334 btrfs_release_path(path);
4335 goto done;
4338 if (!path->skip_locking) {
4339 btrfs_assert_tree_locked(path->nodes[level]);
4340 ret = btrfs_try_spin_lock(next);
4341 if (!ret) {
4342 btrfs_set_path_blocking(path);
4343 btrfs_tree_lock(next);
4344 if (!force_blocking)
4345 btrfs_clear_path_blocking(path, next);
4347 if (force_blocking)
4348 btrfs_set_lock_blocking(next);
4351 ret = 0;
4352 done:
4353 unlock_up(path, 0, 1);
4354 path->leave_spinning = old_spinning;
4355 if (!old_spinning)
4356 btrfs_set_path_blocking(path);
4358 return ret;
4362 * this uses btrfs_prev_leaf to walk backwards in the tree, and keeps
4363 * searching until it gets past min_objectid or finds an item of 'type'
4365 * returns 0 if something is found, 1 if nothing was found and < 0 on error
4367 int btrfs_previous_item(struct btrfs_root *root,
4368 struct btrfs_path *path, u64 min_objectid,
4369 int type)
4371 struct btrfs_key found_key;
4372 struct extent_buffer *leaf;
4373 u32 nritems;
4374 int ret;
4376 while (1) {
4377 if (path->slots[0] == 0) {
4378 btrfs_set_path_blocking(path);
4379 ret = btrfs_prev_leaf(root, path);
4380 if (ret != 0)
4381 return ret;
4382 } else {
4383 path->slots[0]--;
4385 leaf = path->nodes[0];
4386 nritems = btrfs_header_nritems(leaf);
4387 if (nritems == 0)
4388 return 1;
4389 if (path->slots[0] == nritems)
4390 path->slots[0]--;
4392 btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
4393 if (found_key.objectid < min_objectid)
4394 break;
4395 if (found_key.type == type)
4396 return 0;
4397 if (found_key.objectid == min_objectid &&
4398 found_key.type < type)
4399 break;
4401 return 1;