~ [ source navigation ] ~ [ diff markup ] ~ [ identifier search ] ~

TOMOYO Linux Cross Reference
Linux/fs/btrfs/ctree.c

Version: ~ [ linux-5.9 ] ~ [ linux-5.8.14 ] ~ [ linux-5.7.19 ] ~ [ linux-5.6.19 ] ~ [ linux-5.5.19 ] ~ [ linux-5.4.70 ] ~ [ linux-5.3.18 ] ~ [ linux-5.2.21 ] ~ [ linux-5.1.21 ] ~ [ linux-5.0.21 ] ~ [ linux-4.20.17 ] ~ [ linux-4.19.150 ] ~ [ linux-4.18.20 ] ~ [ linux-4.17.19 ] ~ [ linux-4.16.18 ] ~ [ linux-4.15.18 ] ~ [ linux-4.14.200 ] ~ [ linux-4.13.16 ] ~ [ linux-4.12.14 ] ~ [ linux-4.11.12 ] ~ [ linux-4.10.17 ] ~ [ linux-4.9.238 ] ~ [ linux-4.8.17 ] ~ [ linux-4.7.10 ] ~ [ linux-4.6.7 ] ~ [ linux-4.5.7 ] ~ [ linux-4.4.238 ] ~ [ linux-4.3.6 ] ~ [ linux-4.2.8 ] ~ [ linux-4.1.52 ] ~ [ linux-4.0.9 ] ~ [ linux-3.19.8 ] ~ [ linux-3.18.140 ] ~ [ linux-3.17.8 ] ~ [ linux-3.16.85 ] ~ [ linux-3.15.10 ] ~ [ linux-3.14.79 ] ~ [ linux-3.13.11 ] ~ [ linux-3.12.74 ] ~ [ linux-3.11.10 ] ~ [ linux-3.10.108 ] ~ [ linux-2.6.32.71 ] ~ [ linux-2.6.0 ] ~ [ linux-2.4.37.11 ] ~ [ unix-v6-master ] ~ [ ccs-tools-1.8.5 ] ~ [ policy-sample ] ~
Architecture: ~ [ i386 ] ~ [ alpha ] ~ [ m68k ] ~ [ mips ] ~ [ ppc ] ~ [ sparc ] ~ [ sparc64 ] ~

  1 /*
  2  * Copyright (C) 2007,2008 Oracle.  All rights reserved.
  3  *
  4  * This program is free software; you can redistribute it and/or
  5  * modify it under the terms of the GNU General Public
  6  * License v2 as published by the Free Software Foundation.
  7  *
  8  * This program is distributed in the hope that it will be useful,
  9  * but WITHOUT ANY WARRANTY; without even the implied warranty of
 10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 11  * General Public License for more details.
 12  *
 13  * You should have received a copy of the GNU General Public
 14  * License along with this program; if not, write to the
 15  * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 16  * Boston, MA 021110-1307, USA.
 17  */
 18 
 19 #include <linux/sched.h>
 20 #include <linux/slab.h>
 21 #include <linux/rbtree.h>
 22 #include <linux/mm.h>
 23 #include "ctree.h"
 24 #include "disk-io.h"
 25 #include "transaction.h"
 26 #include "print-tree.h"
 27 #include "locking.h"
 28 
 29 static int split_node(struct btrfs_trans_handle *trans, struct btrfs_root
 30                       *root, struct btrfs_path *path, int level);
 31 static int split_leaf(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 32                       const struct btrfs_key *ins_key, struct btrfs_path *path,
 33                       int data_size, int extend);
 34 static int push_node_left(struct btrfs_trans_handle *trans,
 35                           struct btrfs_fs_info *fs_info,
 36                           struct extent_buffer *dst,
 37                           struct extent_buffer *src, int empty);
 38 static int balance_node_right(struct btrfs_trans_handle *trans,
 39                               struct btrfs_fs_info *fs_info,
 40                               struct extent_buffer *dst_buf,
 41                               struct extent_buffer *src_buf);
 42 static void del_ptr(struct btrfs_root *root, struct btrfs_path *path,
 43                     int level, int slot);
 44 static int tree_mod_log_free_eb(struct btrfs_fs_info *fs_info,
 45                                  struct extent_buffer *eb);
 46 
 47 struct btrfs_path *btrfs_alloc_path(void)
 48 {
 49         return kmem_cache_zalloc(btrfs_path_cachep, GFP_NOFS);
 50 }
 51 
 52 /*
 53  * set all locked nodes in the path to blocking locks.  This should
 54  * be done before scheduling
 55  */
 56 noinline void btrfs_set_path_blocking(struct btrfs_path *p)
 57 {
 58         int i;
 59         for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
 60                 if (!p->nodes[i] || !p->locks[i])
 61                         continue;
 62                 btrfs_set_lock_blocking_rw(p->nodes[i], p->locks[i]);
 63                 if (p->locks[i] == BTRFS_READ_LOCK)
 64                         p->locks[i] = BTRFS_READ_LOCK_BLOCKING;
 65                 else if (p->locks[i] == BTRFS_WRITE_LOCK)
 66                         p->locks[i] = BTRFS_WRITE_LOCK_BLOCKING;
 67         }
 68 }
 69 
 70 /*
 71  * reset all the locked nodes in the patch to spinning locks.
 72  *
 73  * held is used to keep lockdep happy, when lockdep is enabled
 74  * we set held to a blocking lock before we go around and
 75  * retake all the spinlocks in the path.  You can safely use NULL
 76  * for held
 77  */
 78 noinline void btrfs_clear_path_blocking(struct btrfs_path *p,
 79                                         struct extent_buffer *held, int held_rw)
 80 {
 81         int i;
 82 
 83         if (held) {
 84                 btrfs_set_lock_blocking_rw(held, held_rw);
 85                 if (held_rw == BTRFS_WRITE_LOCK)
 86                         held_rw = BTRFS_WRITE_LOCK_BLOCKING;
 87                 else if (held_rw == BTRFS_READ_LOCK)
 88                         held_rw = BTRFS_READ_LOCK_BLOCKING;
 89         }
 90         btrfs_set_path_blocking(p);
 91 
 92         for (i = BTRFS_MAX_LEVEL - 1; i >= 0; i--) {
 93                 if (p->nodes[i] && p->locks[i]) {
 94                         btrfs_clear_lock_blocking_rw(p->nodes[i], p->locks[i]);
 95                         if (p->locks[i] == BTRFS_WRITE_LOCK_BLOCKING)
 96                                 p->locks[i] = BTRFS_WRITE_LOCK;
 97                         else if (p->locks[i] == BTRFS_READ_LOCK_BLOCKING)
 98                                 p->locks[i] = BTRFS_READ_LOCK;
 99                 }
100         }
101 
102         if (held)
103                 btrfs_clear_lock_blocking_rw(held, held_rw);
104 }
105 
106 /* this also releases the path */
107 void btrfs_free_path(struct btrfs_path *p)
108 {
109         if (!p)
110                 return;
111         btrfs_release_path(p);
112         kmem_cache_free(btrfs_path_cachep, p);
113 }
114 
115 /*
116  * path release drops references on the extent buffers in the path
117  * and it drops any locks held by this path
118  *
119  * It is safe to call this on paths that no locks or extent buffers held.
120  */
121 noinline void btrfs_release_path(struct btrfs_path *p)
122 {
123         int i;
124 
125         for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
126                 p->slots[i] = 0;
127                 if (!p->nodes[i])
128                         continue;
129                 if (p->locks[i]) {
130                         btrfs_tree_unlock_rw(p->nodes[i], p->locks[i]);
131                         p->locks[i] = 0;
132                 }
133                 free_extent_buffer(p->nodes[i]);
134                 p->nodes[i] = NULL;
135         }
136 }
137 
138 /*
139  * safely gets a reference on the root node of a tree.  A lock
140  * is not taken, so a concurrent writer may put a different node
141  * at the root of the tree.  See btrfs_lock_root_node for the
142  * looping required.
143  *
144  * The extent buffer returned by this has a reference taken, so
145  * it won't disappear.  It may stop being the root of the tree
146  * at any time because there are no locks held.
147  */
148 struct extent_buffer *btrfs_root_node(struct btrfs_root *root)
149 {
150         struct extent_buffer *eb;
151 
152         while (1) {
153                 rcu_read_lock();
154                 eb = rcu_dereference(root->node);
155 
156                 /*
157                  * RCU really hurts here, we could free up the root node because
158                  * it was COWed but we may not get the new root node yet so do
159                  * the inc_not_zero dance and if it doesn't work then
160                  * synchronize_rcu and try again.
161                  */
162                 if (atomic_inc_not_zero(&eb->refs)) {
163                         rcu_read_unlock();
164                         break;
165                 }
166                 rcu_read_unlock();
167                 synchronize_rcu();
168         }
169         return eb;
170 }
171 
172 /* loop around taking references on and locking the root node of the
173  * tree until you end up with a lock on the root.  A locked buffer
174  * is returned, with a reference held.
175  */
176 struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root)
177 {
178         struct extent_buffer *eb;
179 
180         while (1) {
181                 eb = btrfs_root_node(root);
182                 btrfs_tree_lock(eb);
183                 if (eb == root->node)
184                         break;
185                 btrfs_tree_unlock(eb);
186                 free_extent_buffer(eb);
187         }
188         return eb;
189 }
190 
191 /* loop around taking references on and locking the root node of the
192  * tree until you end up with a lock on the root.  A locked buffer
193  * is returned, with a reference held.
194  */
195 struct extent_buffer *btrfs_read_lock_root_node(struct btrfs_root *root)
196 {
197         struct extent_buffer *eb;
198 
199         while (1) {
200                 eb = btrfs_root_node(root);
201                 btrfs_tree_read_lock(eb);
202                 if (eb == root->node)
203                         break;
204                 btrfs_tree_read_unlock(eb);
205                 free_extent_buffer(eb);
206         }
207         return eb;
208 }
209 
210 /* cowonly root (everything not a reference counted cow subvolume), just get
211  * put onto a simple dirty list.  transaction.c walks this to make sure they
212  * get properly updated on disk.
213  */
214 static void add_root_to_dirty_list(struct btrfs_root *root)
215 {
216         struct btrfs_fs_info *fs_info = root->fs_info;
217 
218         if (test_bit(BTRFS_ROOT_DIRTY, &root->state) ||
219             !test_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state))
220                 return;
221 
222         spin_lock(&fs_info->trans_lock);
223         if (!test_and_set_bit(BTRFS_ROOT_DIRTY, &root->state)) {
224                 /* Want the extent tree to be the last on the list */
225                 if (root->objectid == BTRFS_EXTENT_TREE_OBJECTID)
226                         list_move_tail(&root->dirty_list,
227                                        &fs_info->dirty_cowonly_roots);
228                 else
229                         list_move(&root->dirty_list,
230                                   &fs_info->dirty_cowonly_roots);
231         }
232         spin_unlock(&fs_info->trans_lock);
233 }
234 
235 /*
236  * used by snapshot creation to make a copy of a root for a tree with
237  * a given objectid.  The buffer with the new root node is returned in
238  * cow_ret, and this func returns zero on success or a negative error code.
239  */
240 int btrfs_copy_root(struct btrfs_trans_handle *trans,
241                       struct btrfs_root *root,
242                       struct extent_buffer *buf,
243                       struct extent_buffer **cow_ret, u64 new_root_objectid)
244 {
245         struct btrfs_fs_info *fs_info = root->fs_info;
246         struct extent_buffer *cow;
247         int ret = 0;
248         int level;
249         struct btrfs_disk_key disk_key;
250 
251         WARN_ON(test_bit(BTRFS_ROOT_REF_COWS, &root->state) &&
252                 trans->transid != fs_info->running_transaction->transid);
253         WARN_ON(test_bit(BTRFS_ROOT_REF_COWS, &root->state) &&
254                 trans->transid != root->last_trans);
255 
256         level = btrfs_header_level(buf);
257         if (level == 0)
258                 btrfs_item_key(buf, &disk_key, 0);
259         else
260                 btrfs_node_key(buf, &disk_key, 0);
261 
262         cow = btrfs_alloc_tree_block(trans, root, 0, new_root_objectid,
263                         &disk_key, level, buf->start, 0);
264         if (IS_ERR(cow))
265                 return PTR_ERR(cow);
266 
267         copy_extent_buffer_full(cow, buf);
268         btrfs_set_header_bytenr(cow, cow->start);
269         btrfs_set_header_generation(cow, trans->transid);
270         btrfs_set_header_backref_rev(cow, BTRFS_MIXED_BACKREF_REV);
271         btrfs_clear_header_flag(cow, BTRFS_HEADER_FLAG_WRITTEN |
272                                      BTRFS_HEADER_FLAG_RELOC);
273         if (new_root_objectid == BTRFS_TREE_RELOC_OBJECTID)
274                 btrfs_set_header_flag(cow, BTRFS_HEADER_FLAG_RELOC);
275         else
276                 btrfs_set_header_owner(cow, new_root_objectid);
277 
278         write_extent_buffer_fsid(cow, fs_info->fsid);
279 
280         WARN_ON(btrfs_header_generation(buf) > trans->transid);
281         if (new_root_objectid == BTRFS_TREE_RELOC_OBJECTID)
282                 ret = btrfs_inc_ref(trans, root, cow, 1);
283         else
284                 ret = btrfs_inc_ref(trans, root, cow, 0);
285 
286         if (ret)
287                 return ret;
288 
289         btrfs_mark_buffer_dirty(cow);
290         *cow_ret = cow;
291         return 0;
292 }
293 
294 enum mod_log_op {
295         MOD_LOG_KEY_REPLACE,
296         MOD_LOG_KEY_ADD,
297         MOD_LOG_KEY_REMOVE,
298         MOD_LOG_KEY_REMOVE_WHILE_FREEING,
299         MOD_LOG_KEY_REMOVE_WHILE_MOVING,
300         MOD_LOG_MOVE_KEYS,
301         MOD_LOG_ROOT_REPLACE,
302 };
303 
304 struct tree_mod_move {
305         int dst_slot;
306         int nr_items;
307 };
308 
309 struct tree_mod_root {
310         u64 logical;
311         u8 level;
312 };
313 
314 struct tree_mod_elem {
315         struct rb_node node;
316         u64 logical;
317         u64 seq;
318         enum mod_log_op op;
319 
320         /* this is used for MOD_LOG_KEY_* and MOD_LOG_MOVE_KEYS operations */
321         int slot;
322 
323         /* this is used for MOD_LOG_KEY* and MOD_LOG_ROOT_REPLACE */
324         u64 generation;
325 
326         /* those are used for op == MOD_LOG_KEY_{REPLACE,REMOVE} */
327         struct btrfs_disk_key key;
328         u64 blockptr;
329 
330         /* this is used for op == MOD_LOG_MOVE_KEYS */
331         struct tree_mod_move move;
332 
333         /* this is used for op == MOD_LOG_ROOT_REPLACE */
334         struct tree_mod_root old_root;
335 };
336 
337 static inline void tree_mod_log_read_lock(struct btrfs_fs_info *fs_info)
338 {
339         read_lock(&fs_info->tree_mod_log_lock);
340 }
341 
342 static inline void tree_mod_log_read_unlock(struct btrfs_fs_info *fs_info)
343 {
344         read_unlock(&fs_info->tree_mod_log_lock);
345 }
346 
347 static inline void tree_mod_log_write_lock(struct btrfs_fs_info *fs_info)
348 {
349         write_lock(&fs_info->tree_mod_log_lock);
350 }
351 
352 static inline void tree_mod_log_write_unlock(struct btrfs_fs_info *fs_info)
353 {
354         write_unlock(&fs_info->tree_mod_log_lock);
355 }
356 
357 /*
358  * Pull a new tree mod seq number for our operation.
359  */
360 static inline u64 btrfs_inc_tree_mod_seq(struct btrfs_fs_info *fs_info)
361 {
362         return atomic64_inc_return(&fs_info->tree_mod_seq);
363 }
364 
365 /*
366  * This adds a new blocker to the tree mod log's blocker list if the @elem
367  * passed does not already have a sequence number set. So when a caller expects
368  * to record tree modifications, it should ensure to set elem->seq to zero
369  * before calling btrfs_get_tree_mod_seq.
370  * Returns a fresh, unused tree log modification sequence number, even if no new
371  * blocker was added.
372  */
373 u64 btrfs_get_tree_mod_seq(struct btrfs_fs_info *fs_info,
374                            struct seq_list *elem)
375 {
376         tree_mod_log_write_lock(fs_info);
377         spin_lock(&fs_info->tree_mod_seq_lock);
378         if (!elem->seq) {
379                 elem->seq = btrfs_inc_tree_mod_seq(fs_info);
380                 list_add_tail(&elem->list, &fs_info->tree_mod_seq_list);
381         }
382         spin_unlock(&fs_info->tree_mod_seq_lock);
383         tree_mod_log_write_unlock(fs_info);
384 
385         return elem->seq;
386 }
387 
388 void btrfs_put_tree_mod_seq(struct btrfs_fs_info *fs_info,
389                             struct seq_list *elem)
390 {
391         struct rb_root *tm_root;
392         struct rb_node *node;
393         struct rb_node *next;
394         struct seq_list *cur_elem;
395         struct tree_mod_elem *tm;
396         u64 min_seq = (u64)-1;
397         u64 seq_putting = elem->seq;
398 
399         if (!seq_putting)
400                 return;
401 
402         spin_lock(&fs_info->tree_mod_seq_lock);
403         list_del(&elem->list);
404         elem->seq = 0;
405 
406         list_for_each_entry(cur_elem, &fs_info->tree_mod_seq_list, list) {
407                 if (cur_elem->seq < min_seq) {
408                         if (seq_putting > cur_elem->seq) {
409                                 /*
410                                  * blocker with lower sequence number exists, we
411                                  * cannot remove anything from the log
412                                  */
413                                 spin_unlock(&fs_info->tree_mod_seq_lock);
414                                 return;
415                         }
416                         min_seq = cur_elem->seq;
417                 }
418         }
419         spin_unlock(&fs_info->tree_mod_seq_lock);
420 
421         /*
422          * anything that's lower than the lowest existing (read: blocked)
423          * sequence number can be removed from the tree.
424          */
425         tree_mod_log_write_lock(fs_info);
426         tm_root = &fs_info->tree_mod_log;
427         for (node = rb_first(tm_root); node; node = next) {
428                 next = rb_next(node);
429                 tm = rb_entry(node, struct tree_mod_elem, node);
430                 if (tm->seq > min_seq)
431                         continue;
432                 rb_erase(node, tm_root);
433                 kfree(tm);
434         }
435         tree_mod_log_write_unlock(fs_info);
436 }
437 
438 /*
439  * key order of the log:
440  *       node/leaf start address -> sequence
441  *
442  * The 'start address' is the logical address of the *new* root node
443  * for root replace operations, or the logical address of the affected
444  * block for all other operations.
445  *
446  * Note: must be called with write lock (tree_mod_log_write_lock).
447  */
448 static noinline int
449 __tree_mod_log_insert(struct btrfs_fs_info *fs_info, struct tree_mod_elem *tm)
450 {
451         struct rb_root *tm_root;
452         struct rb_node **new;
453         struct rb_node *parent = NULL;
454         struct tree_mod_elem *cur;
455 
456         tm->seq = btrfs_inc_tree_mod_seq(fs_info);
457 
458         tm_root = &fs_info->tree_mod_log;
459         new = &tm_root->rb_node;
460         while (*new) {
461                 cur = rb_entry(*new, struct tree_mod_elem, node);
462                 parent = *new;
463                 if (cur->logical < tm->logical)
464                         new = &((*new)->rb_left);
465                 else if (cur->logical > tm->logical)
466                         new = &((*new)->rb_right);
467                 else if (cur->seq < tm->seq)
468                         new = &((*new)->rb_left);
469                 else if (cur->seq > tm->seq)
470                         new = &((*new)->rb_right);
471                 else
472                         return -EEXIST;
473         }
474 
475         rb_link_node(&tm->node, parent, new);
476         rb_insert_color(&tm->node, tm_root);
477         return 0;
478 }
479 
480 /*
481  * Determines if logging can be omitted. Returns 1 if it can. Otherwise, it
482  * returns zero with the tree_mod_log_lock acquired. The caller must hold
483  * this until all tree mod log insertions are recorded in the rb tree and then
484  * call tree_mod_log_write_unlock() to release.
485  */
486 static inline int tree_mod_dont_log(struct btrfs_fs_info *fs_info,
487                                     struct extent_buffer *eb) {
488         smp_mb();
489         if (list_empty(&(fs_info)->tree_mod_seq_list))
490                 return 1;
491         if (eb && btrfs_header_level(eb) == 0)
492                 return 1;
493 
494         tree_mod_log_write_lock(fs_info);
495         if (list_empty(&(fs_info)->tree_mod_seq_list)) {
496                 tree_mod_log_write_unlock(fs_info);
497                 return 1;
498         }
499 
500         return 0;
501 }
502 
503 /* Similar to tree_mod_dont_log, but doesn't acquire any locks. */
504 static inline int tree_mod_need_log(const struct btrfs_fs_info *fs_info,
505                                     struct extent_buffer *eb)
506 {
507         smp_mb();
508         if (list_empty(&(fs_info)->tree_mod_seq_list))
509                 return 0;
510         if (eb && btrfs_header_level(eb) == 0)
511                 return 0;
512 
513         return 1;
514 }
515 
516 static struct tree_mod_elem *
517 alloc_tree_mod_elem(struct extent_buffer *eb, int slot,
518                     enum mod_log_op op, gfp_t flags)
519 {
520         struct tree_mod_elem *tm;
521 
522         tm = kzalloc(sizeof(*tm), flags);
523         if (!tm)
524                 return NULL;
525 
526         tm->logical = eb->start;
527         if (op != MOD_LOG_KEY_ADD) {
528                 btrfs_node_key(eb, &tm->key, slot);
529                 tm->blockptr = btrfs_node_blockptr(eb, slot);
530         }
531         tm->op = op;
532         tm->slot = slot;
533         tm->generation = btrfs_node_ptr_generation(eb, slot);
534         RB_CLEAR_NODE(&tm->node);
535 
536         return tm;
537 }
538 
539 static noinline int
540 tree_mod_log_insert_key(struct btrfs_fs_info *fs_info,
541                         struct extent_buffer *eb, int slot,
542                         enum mod_log_op op, gfp_t flags)
543 {
544         struct tree_mod_elem *tm;
545         int ret;
546 
547         if (!tree_mod_need_log(fs_info, eb))
548                 return 0;
549 
550         tm = alloc_tree_mod_elem(eb, slot, op, flags);
551         if (!tm)
552                 return -ENOMEM;
553 
554         if (tree_mod_dont_log(fs_info, eb)) {
555                 kfree(tm);
556                 return 0;
557         }
558 
559         ret = __tree_mod_log_insert(fs_info, tm);
560         tree_mod_log_write_unlock(fs_info);
561         if (ret)
562                 kfree(tm);
563 
564         return ret;
565 }
566 
567 static noinline int
568 tree_mod_log_insert_move(struct btrfs_fs_info *fs_info,
569                          struct extent_buffer *eb, int dst_slot, int src_slot,
570                          int nr_items)
571 {
572         struct tree_mod_elem *tm = NULL;
573         struct tree_mod_elem **tm_list = NULL;
574         int ret = 0;
575         int i;
576         int locked = 0;
577 
578         if (!tree_mod_need_log(fs_info, eb))
579                 return 0;
580 
581         tm_list = kcalloc(nr_items, sizeof(struct tree_mod_elem *), GFP_NOFS);
582         if (!tm_list)
583                 return -ENOMEM;
584 
585         tm = kzalloc(sizeof(*tm), GFP_NOFS);
586         if (!tm) {
587                 ret = -ENOMEM;
588                 goto free_tms;
589         }
590 
591         tm->logical = eb->start;
592         tm->slot = src_slot;
593         tm->move.dst_slot = dst_slot;
594         tm->move.nr_items = nr_items;
595         tm->op = MOD_LOG_MOVE_KEYS;
596 
597         for (i = 0; i + dst_slot < src_slot && i < nr_items; i++) {
598                 tm_list[i] = alloc_tree_mod_elem(eb, i + dst_slot,
599                     MOD_LOG_KEY_REMOVE_WHILE_MOVING, GFP_NOFS);
600                 if (!tm_list[i]) {
601                         ret = -ENOMEM;
602                         goto free_tms;
603                 }
604         }
605 
606         if (tree_mod_dont_log(fs_info, eb))
607                 goto free_tms;
608         locked = 1;
609 
610         /*
611          * When we override something during the move, we log these removals.
612          * This can only happen when we move towards the beginning of the
613          * buffer, i.e. dst_slot < src_slot.
614          */
615         for (i = 0; i + dst_slot < src_slot && i < nr_items; i++) {
616                 ret = __tree_mod_log_insert(fs_info, tm_list[i]);
617                 if (ret)
618                         goto free_tms;
619         }
620 
621         ret = __tree_mod_log_insert(fs_info, tm);
622         if (ret)
623                 goto free_tms;
624         tree_mod_log_write_unlock(fs_info);
625         kfree(tm_list);
626 
627         return 0;
628 free_tms:
629         for (i = 0; i < nr_items; i++) {
630                 if (tm_list[i] && !RB_EMPTY_NODE(&tm_list[i]->node))
631                         rb_erase(&tm_list[i]->node, &fs_info->tree_mod_log);
632                 kfree(tm_list[i]);
633         }
634         if (locked)
635                 tree_mod_log_write_unlock(fs_info);
636         kfree(tm_list);
637         kfree(tm);
638 
639         return ret;
640 }
641 
642 static inline int
643 __tree_mod_log_free_eb(struct btrfs_fs_info *fs_info,
644                        struct tree_mod_elem **tm_list,
645                        int nritems)
646 {
647         int i, j;
648         int ret;
649 
650         for (i = nritems - 1; i >= 0; i--) {
651                 ret = __tree_mod_log_insert(fs_info, tm_list[i]);
652                 if (ret) {
653                         for (j = nritems - 1; j > i; j--)
654                                 rb_erase(&tm_list[j]->node,
655                                          &fs_info->tree_mod_log);
656                         return ret;
657                 }
658         }
659 
660         return 0;
661 }
662 
663 static noinline int
664 tree_mod_log_insert_root(struct btrfs_fs_info *fs_info,
665                          struct extent_buffer *old_root,
666                          struct extent_buffer *new_root,
667                          int log_removal)
668 {
669         struct tree_mod_elem *tm = NULL;
670         struct tree_mod_elem **tm_list = NULL;
671         int nritems = 0;
672         int ret = 0;
673         int i;
674 
675         if (!tree_mod_need_log(fs_info, NULL))
676                 return 0;
677 
678         if (log_removal && btrfs_header_level(old_root) > 0) {
679                 nritems = btrfs_header_nritems(old_root);
680                 tm_list = kcalloc(nritems, sizeof(struct tree_mod_elem *),
681                                   GFP_NOFS);
682                 if (!tm_list) {
683                         ret = -ENOMEM;
684                         goto free_tms;
685                 }
686                 for (i = 0; i < nritems; i++) {
687                         tm_list[i] = alloc_tree_mod_elem(old_root, i,
688                             MOD_LOG_KEY_REMOVE_WHILE_FREEING, GFP_NOFS);
689                         if (!tm_list[i]) {
690                                 ret = -ENOMEM;
691                                 goto free_tms;
692                         }
693                 }
694         }
695 
696         tm = kzalloc(sizeof(*tm), GFP_NOFS);
697         if (!tm) {
698                 ret = -ENOMEM;
699                 goto free_tms;
700         }
701 
702         tm->logical = new_root->start;
703         tm->old_root.logical = old_root->start;
704         tm->old_root.level = btrfs_header_level(old_root);
705         tm->generation = btrfs_header_generation(old_root);
706         tm->op = MOD_LOG_ROOT_REPLACE;
707 
708         if (tree_mod_dont_log(fs_info, NULL))
709                 goto free_tms;
710 
711         if (tm_list)
712                 ret = __tree_mod_log_free_eb(fs_info, tm_list, nritems);
713         if (!ret)
714                 ret = __tree_mod_log_insert(fs_info, tm);
715 
716         tree_mod_log_write_unlock(fs_info);
717         if (ret)
718                 goto free_tms;
719         kfree(tm_list);
720 
721         return ret;
722 
723 free_tms:
724         if (tm_list) {
725                 for (i = 0; i < nritems; i++)
726                         kfree(tm_list[i]);
727                 kfree(tm_list);
728         }
729         kfree(tm);
730 
731         return ret;
732 }
733 
734 static struct tree_mod_elem *
735 __tree_mod_log_search(struct btrfs_fs_info *fs_info, u64 start, u64 min_seq,
736                       int smallest)
737 {
738         struct rb_root *tm_root;
739         struct rb_node *node;
740         struct tree_mod_elem *cur = NULL;
741         struct tree_mod_elem *found = NULL;
742 
743         tree_mod_log_read_lock(fs_info);
744         tm_root = &fs_info->tree_mod_log;
745         node = tm_root->rb_node;
746         while (node) {
747                 cur = rb_entry(node, struct tree_mod_elem, node);
748                 if (cur->logical < start) {
749                         node = node->rb_left;
750                 } else if (cur->logical > start) {
751                         node = node->rb_right;
752                 } else if (cur->seq < min_seq) {
753                         node = node->rb_left;
754                 } else if (!smallest) {
755                         /* we want the node with the highest seq */
756                         if (found)
757                                 BUG_ON(found->seq > cur->seq);
758                         found = cur;
759                         node = node->rb_left;
760                 } else if (cur->seq > min_seq) {
761                         /* we want the node with the smallest seq */
762                         if (found)
763                                 BUG_ON(found->seq < cur->seq);
764                         found = cur;
765                         node = node->rb_right;
766                 } else {
767                         found = cur;
768                         break;
769                 }
770         }
771         tree_mod_log_read_unlock(fs_info);
772 
773         return found;
774 }
775 
776 /*
777  * this returns the element from the log with the smallest time sequence
778  * value that's in the log (the oldest log item). any element with a time
779  * sequence lower than min_seq will be ignored.
780  */
781 static struct tree_mod_elem *
782 tree_mod_log_search_oldest(struct btrfs_fs_info *fs_info, u64 start,
783                            u64 min_seq)
784 {
785         return __tree_mod_log_search(fs_info, start, min_seq, 1);
786 }
787 
788 /*
789  * this returns the element from the log with the largest time sequence
790  * value that's in the log (the most recent log item). any element with
791  * a time sequence lower than min_seq will be ignored.
792  */
793 static struct tree_mod_elem *
794 tree_mod_log_search(struct btrfs_fs_info *fs_info, u64 start, u64 min_seq)
795 {
796         return __tree_mod_log_search(fs_info, start, min_seq, 0);
797 }
798 
799 static noinline int
800 tree_mod_log_eb_copy(struct btrfs_fs_info *fs_info, struct extent_buffer *dst,
801                      struct extent_buffer *src, unsigned long dst_offset,
802                      unsigned long src_offset, int nr_items)
803 {
804         int ret = 0;
805         struct tree_mod_elem **tm_list = NULL;
806         struct tree_mod_elem **tm_list_add, **tm_list_rem;
807         int i;
808         int locked = 0;
809 
810         if (!tree_mod_need_log(fs_info, NULL))
811                 return 0;
812 
813         if (btrfs_header_level(dst) == 0 && btrfs_header_level(src) == 0)
814                 return 0;
815 
816         tm_list = kcalloc(nr_items * 2, sizeof(struct tree_mod_elem *),
817                           GFP_NOFS);
818         if (!tm_list)
819                 return -ENOMEM;
820 
821         tm_list_add = tm_list;
822         tm_list_rem = tm_list + nr_items;
823         for (i = 0; i < nr_items; i++) {
824                 tm_list_rem[i] = alloc_tree_mod_elem(src, i + src_offset,
825                     MOD_LOG_KEY_REMOVE, GFP_NOFS);
826                 if (!tm_list_rem[i]) {
827                         ret = -ENOMEM;
828                         goto free_tms;
829                 }
830 
831                 tm_list_add[i] = alloc_tree_mod_elem(dst, i + dst_offset,
832                     MOD_LOG_KEY_ADD, GFP_NOFS);
833                 if (!tm_list_add[i]) {
834                         ret = -ENOMEM;
835                         goto free_tms;
836                 }
837         }
838 
839         if (tree_mod_dont_log(fs_info, NULL))
840                 goto free_tms;
841         locked = 1;
842 
843         for (i = 0; i < nr_items; i++) {
844                 ret = __tree_mod_log_insert(fs_info, tm_list_rem[i]);
845                 if (ret)
846                         goto free_tms;
847                 ret = __tree_mod_log_insert(fs_info, tm_list_add[i]);
848                 if (ret)
849                         goto free_tms;
850         }
851 
852         tree_mod_log_write_unlock(fs_info);
853         kfree(tm_list);
854 
855         return 0;
856 
857 free_tms:
858         for (i = 0; i < nr_items * 2; i++) {
859                 if (tm_list[i] && !RB_EMPTY_NODE(&tm_list[i]->node))
860                         rb_erase(&tm_list[i]->node, &fs_info->tree_mod_log);
861                 kfree(tm_list[i]);
862         }
863         if (locked)
864                 tree_mod_log_write_unlock(fs_info);
865         kfree(tm_list);
866 
867         return ret;
868 }
869 
870 static inline void
871 tree_mod_log_eb_move(struct btrfs_fs_info *fs_info, struct extent_buffer *dst,
872                      int dst_offset, int src_offset, int nr_items)
873 {
874         int ret;
875         ret = tree_mod_log_insert_move(fs_info, dst, dst_offset, src_offset,
876                                        nr_items);
877         BUG_ON(ret < 0);
878 }
879 
880 static noinline void
881 tree_mod_log_set_node_key(struct btrfs_fs_info *fs_info,
882                           struct extent_buffer *eb, int slot, int atomic)
883 {
884         int ret;
885 
886         ret = tree_mod_log_insert_key(fs_info, eb, slot,
887                                         MOD_LOG_KEY_REPLACE,
888                                         atomic ? GFP_ATOMIC : GFP_NOFS);
889         BUG_ON(ret < 0);
890 }
891 
892 static noinline int
893 tree_mod_log_free_eb(struct btrfs_fs_info *fs_info, struct extent_buffer *eb)
894 {
895         struct tree_mod_elem **tm_list = NULL;
896         int nritems = 0;
897         int i;
898         int ret = 0;
899 
900         if (btrfs_header_level(eb) == 0)
901                 return 0;
902 
903         if (!tree_mod_need_log(fs_info, NULL))
904                 return 0;
905 
906         nritems = btrfs_header_nritems(eb);
907         tm_list = kcalloc(nritems, sizeof(struct tree_mod_elem *), GFP_NOFS);
908         if (!tm_list)
909                 return -ENOMEM;
910 
911         for (i = 0; i < nritems; i++) {
912                 tm_list[i] = alloc_tree_mod_elem(eb, i,
913                     MOD_LOG_KEY_REMOVE_WHILE_FREEING, GFP_NOFS);
914                 if (!tm_list[i]) {
915                         ret = -ENOMEM;
916                         goto free_tms;
917                 }
918         }
919 
920         if (tree_mod_dont_log(fs_info, eb))
921                 goto free_tms;
922 
923         ret = __tree_mod_log_free_eb(fs_info, tm_list, nritems);
924         tree_mod_log_write_unlock(fs_info);
925         if (ret)
926                 goto free_tms;
927         kfree(tm_list);
928 
929         return 0;
930 
931 free_tms:
932         for (i = 0; i < nritems; i++)
933                 kfree(tm_list[i]);
934         kfree(tm_list);
935 
936         return ret;
937 }
938 
939 static noinline void
940 tree_mod_log_set_root_pointer(struct btrfs_root *root,
941                               struct extent_buffer *new_root_node,
942                               int log_removal)
943 {
944         int ret;
945         ret = tree_mod_log_insert_root(root->fs_info, root->node,
946                                        new_root_node, log_removal);
947         BUG_ON(ret < 0);
948 }
949 
950 /*
951  * check if the tree block can be shared by multiple trees
952  */
953 int btrfs_block_can_be_shared(struct btrfs_root *root,
954                               struct extent_buffer *buf)
955 {
956         /*
957          * Tree blocks not in reference counted trees and tree roots
958          * are never shared. If a block was allocated after the last
959          * snapshot and the block was not allocated by tree relocation,
960          * we know the block is not shared.
961          */
962         if (test_bit(BTRFS_ROOT_REF_COWS, &root->state) &&
963             buf != root->node && buf != root->commit_root &&
964             (btrfs_header_generation(buf) <=
965              btrfs_root_last_snapshot(&root->root_item) ||
966              btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC)))
967                 return 1;
968 #ifdef BTRFS_COMPAT_EXTENT_TREE_V0
969         if (test_bit(BTRFS_ROOT_REF_COWS, &root->state) &&
970             btrfs_header_backref_rev(buf) < BTRFS_MIXED_BACKREF_REV)
971                 return 1;
972 #endif
973         return 0;
974 }
975 
976 static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans,
977                                        struct btrfs_root *root,
978                                        struct extent_buffer *buf,
979                                        struct extent_buffer *cow,
980                                        int *last_ref)
981 {
982         struct btrfs_fs_info *fs_info = root->fs_info;
983         u64 refs;
984         u64 owner;
985         u64 flags;
986         u64 new_flags = 0;
987         int ret;
988 
989         /*
990          * Backrefs update rules:
991          *
992          * Always use full backrefs for extent pointers in tree block
993          * allocated by tree relocation.
994          *
995          * If a shared tree block is no longer referenced by its owner
996          * tree (btrfs_header_owner(buf) == root->root_key.objectid),
997          * use full backrefs for extent pointers in tree block.
998          *
999          * If a tree block is been relocating
1000          * (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID),
1001          * use full backrefs for extent pointers in tree block.
1002          * The reason for this is some operations (such as drop tree)
1003          * are only allowed for blocks use full backrefs.
1004          */
1005 
1006         if (btrfs_block_can_be_shared(root, buf)) {
1007                 ret = btrfs_lookup_extent_info(trans, fs_info, buf->start,
1008                                                btrfs_header_level(buf), 1,
1009                                                &refs, &flags);
1010                 if (ret)
1011                         return ret;
1012                 if (refs == 0) {
1013                         ret = -EROFS;
1014                         btrfs_handle_fs_error(fs_info, ret, NULL);
1015                         return ret;
1016                 }
1017         } else {
1018                 refs = 1;
1019                 if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID ||
1020                     btrfs_header_backref_rev(buf) < BTRFS_MIXED_BACKREF_REV)
1021                         flags = BTRFS_BLOCK_FLAG_FULL_BACKREF;
1022                 else
1023                         flags = 0;
1024         }
1025 
1026         owner = btrfs_header_owner(buf);
1027         BUG_ON(owner == BTRFS_TREE_RELOC_OBJECTID &&
1028                !(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF));
1029 
1030         if (refs > 1) {
1031                 if ((owner == root->root_key.objectid ||
1032                      root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) &&
1033                     !(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)) {
1034                         ret = btrfs_inc_ref(trans, root, buf, 1);
1035                         if (ret)
1036                                 return ret;
1037 
1038                         if (root->root_key.objectid ==
1039                             BTRFS_TREE_RELOC_OBJECTID) {
1040                                 ret = btrfs_dec_ref(trans, root, buf, 0);
1041                                 if (ret)
1042                                         return ret;
1043                                 ret = btrfs_inc_ref(trans, root, cow, 1);
1044                                 if (ret)
1045                                         return ret;
1046                         }
1047                         new_flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF;
1048                 } else {
1049 
1050                         if (root->root_key.objectid ==
1051                             BTRFS_TREE_RELOC_OBJECTID)
1052                                 ret = btrfs_inc_ref(trans, root, cow, 1);
1053                         else
1054                                 ret = btrfs_inc_ref(trans, root, cow, 0);
1055                         if (ret)
1056                                 return ret;
1057                 }
1058                 if (new_flags != 0) {
1059                         int level = btrfs_header_level(buf);
1060 
1061                         ret = btrfs_set_disk_extent_flags(trans, fs_info,
1062                                                           buf->start,
1063                                                           buf->len,
1064                                                           new_flags, level, 0);
1065                         if (ret)
1066                                 return ret;
1067                 }
1068         } else {
1069                 if (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF) {
1070                         if (root->root_key.objectid ==
1071                             BTRFS_TREE_RELOC_OBJECTID)
1072                                 ret = btrfs_inc_ref(trans, root, cow, 1);
1073                         else
1074                                 ret = btrfs_inc_ref(trans, root, cow, 0);
1075                         if (ret)
1076                                 return ret;
1077                         ret = btrfs_dec_ref(trans, root, buf, 1);
1078                         if (ret)
1079                                 return ret;
1080                 }
1081                 clean_tree_block(fs_info, buf);
1082                 *last_ref = 1;
1083         }
1084         return 0;
1085 }
1086 
1087 /*
1088  * does the dirty work in cow of a single block.  The parent block (if
1089  * supplied) is updated to point to the new cow copy.  The new buffer is marked
1090  * dirty and returned locked.  If you modify the block it needs to be marked
1091  * dirty again.
1092  *
1093  * search_start -- an allocation hint for the new block
1094  *
1095  * empty_size -- a hint that you plan on doing more cow.  This is the size in
1096  * bytes the allocator should try to find free next to the block it returns.
1097  * This is just a hint and may be ignored by the allocator.
1098  */
1099 static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans,
1100                              struct btrfs_root *root,
1101                              struct extent_buffer *buf,
1102                              struct extent_buffer *parent, int parent_slot,
1103                              struct extent_buffer **cow_ret,
1104                              u64 search_start, u64 empty_size)
1105 {
1106         struct btrfs_fs_info *fs_info = root->fs_info;
1107         struct btrfs_disk_key disk_key;
1108         struct extent_buffer *cow;
1109         int level, ret;
1110         int last_ref = 0;
1111         int unlock_orig = 0;
1112         u64 parent_start = 0;
1113 
1114         if (*cow_ret == buf)
1115                 unlock_orig = 1;
1116 
1117         btrfs_assert_tree_locked(buf);
1118 
1119         WARN_ON(test_bit(BTRFS_ROOT_REF_COWS, &root->state) &&
1120                 trans->transid != fs_info->running_transaction->transid);
1121         WARN_ON(test_bit(BTRFS_ROOT_REF_COWS, &root->state) &&
1122                 trans->transid != root->last_trans);
1123 
1124         level = btrfs_header_level(buf);
1125 
1126         if (level == 0)
1127                 btrfs_item_key(buf, &disk_key, 0);
1128         else
1129                 btrfs_node_key(buf, &disk_key, 0);
1130 
1131         if ((root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) && parent)
1132                 parent_start = parent->start;
1133 
1134         cow = btrfs_alloc_tree_block(trans, root, parent_start,
1135                         root->root_key.objectid, &disk_key, level,
1136                         search_start, empty_size);
1137         if (IS_ERR(cow))
1138                 return PTR_ERR(cow);
1139 
1140         /* cow is set to blocking by btrfs_init_new_buffer */
1141 
1142         copy_extent_buffer_full(cow, buf);
1143         btrfs_set_header_bytenr(cow, cow->start);
1144         btrfs_set_header_generation(cow, trans->transid);
1145         btrfs_set_header_backref_rev(cow, BTRFS_MIXED_BACKREF_REV);
1146         btrfs_clear_header_flag(cow, BTRFS_HEADER_FLAG_WRITTEN |
1147                                      BTRFS_HEADER_FLAG_RELOC);
1148         if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)
1149                 btrfs_set_header_flag(cow, BTRFS_HEADER_FLAG_RELOC);
1150         else
1151                 btrfs_set_header_owner(cow, root->root_key.objectid);
1152 
1153         write_extent_buffer_fsid(cow, fs_info->fsid);
1154 
1155         ret = update_ref_for_cow(trans, root, buf, cow, &last_ref);
1156         if (ret) {
1157                 btrfs_abort_transaction(trans, ret);
1158                 return ret;
1159         }
1160 
1161         if (test_bit(BTRFS_ROOT_REF_COWS, &root->state)) {
1162                 ret = btrfs_reloc_cow_block(trans, root, buf, cow);
1163                 if (ret) {
1164                         btrfs_abort_transaction(trans, ret);
1165                         return ret;
1166                 }
1167         }
1168 
1169         if (buf == root->node) {
1170                 WARN_ON(parent && parent != buf);
1171                 if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID ||
1172                     btrfs_header_backref_rev(buf) < BTRFS_MIXED_BACKREF_REV)
1173                         parent_start = buf->start;
1174 
1175                 extent_buffer_get(cow);
1176                 tree_mod_log_set_root_pointer(root, cow, 1);
1177                 rcu_assign_pointer(root->node, cow);
1178 
1179                 btrfs_free_tree_block(trans, root, buf, parent_start,
1180                                       last_ref);
1181                 free_extent_buffer(buf);
1182                 add_root_to_dirty_list(root);
1183         } else {
1184                 WARN_ON(trans->transid != btrfs_header_generation(parent));
1185                 tree_mod_log_insert_key(fs_info, parent, parent_slot,
1186                                         MOD_LOG_KEY_REPLACE, GFP_NOFS);
1187                 btrfs_set_node_blockptr(parent, parent_slot,
1188                                         cow->start);
1189                 btrfs_set_node_ptr_generation(parent, parent_slot,
1190                                               trans->transid);
1191                 btrfs_mark_buffer_dirty(parent);
1192                 if (last_ref) {
1193                         ret = tree_mod_log_free_eb(fs_info, buf);
1194                         if (ret) {
1195                                 btrfs_abort_transaction(trans, ret);
1196                                 return ret;
1197                         }
1198                 }
1199                 btrfs_free_tree_block(trans, root, buf, parent_start,
1200                                       last_ref);
1201         }
1202         if (unlock_orig)
1203                 btrfs_tree_unlock(buf);
1204         free_extent_buffer_stale(buf);
1205         btrfs_mark_buffer_dirty(cow);
1206         *cow_ret = cow;
1207         return 0;
1208 }
1209 
1210 /*
1211  * returns the logical address of the oldest predecessor of the given root.
1212  * entries older than time_seq are ignored.
1213  */
1214 static struct tree_mod_elem *
1215 __tree_mod_log_oldest_root(struct btrfs_fs_info *fs_info,
1216                            struct extent_buffer *eb_root, u64 time_seq)
1217 {
1218         struct tree_mod_elem *tm;
1219         struct tree_mod_elem *found = NULL;
1220         u64 root_logical = eb_root->start;
1221         int looped = 0;
1222 
1223         if (!time_seq)
1224                 return NULL;
1225 
1226         /*
1227          * the very last operation that's logged for a root is the
1228          * replacement operation (if it is replaced at all). this has
1229          * the logical address of the *new* root, making it the very
1230          * first operation that's logged for this root.
1231          */
1232         while (1) {
1233                 tm = tree_mod_log_search_oldest(fs_info, root_logical,
1234                                                 time_seq);
1235                 if (!looped && !tm)
1236                         return NULL;
1237                 /*
1238                  * if there are no tree operation for the oldest root, we simply
1239                  * return it. this should only happen if that (old) root is at
1240                  * level 0.
1241                  */
1242                 if (!tm)
1243                         break;
1244 
1245                 /*
1246                  * if there's an operation that's not a root replacement, we
1247                  * found the oldest version of our root. normally, we'll find a
1248                  * MOD_LOG_KEY_REMOVE_WHILE_FREEING operation here.
1249                  */
1250                 if (tm->op != MOD_LOG_ROOT_REPLACE)
1251                         break;
1252 
1253                 found = tm;
1254                 root_logical = tm->old_root.logical;
1255                 looped = 1;
1256         }
1257 
1258         /* if there's no old root to return, return what we found instead */
1259         if (!found)
1260                 found = tm;
1261 
1262         return found;
1263 }
1264 
1265 /*
1266  * tm is a pointer to the first operation to rewind within eb. then, all
1267  * previous operations will be rewound (until we reach something older than
1268  * time_seq).
1269  */
1270 static void
1271 __tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb,
1272                       u64 time_seq, struct tree_mod_elem *first_tm)
1273 {
1274         u32 n;
1275         struct rb_node *next;
1276         struct tree_mod_elem *tm = first_tm;
1277         unsigned long o_dst;
1278         unsigned long o_src;
1279         unsigned long p_size = sizeof(struct btrfs_key_ptr);
1280 
1281         n = btrfs_header_nritems(eb);
1282         tree_mod_log_read_lock(fs_info);
1283         while (tm && tm->seq >= time_seq) {
1284                 /*
1285                  * all the operations are recorded with the operator used for
1286                  * the modification. as we're going backwards, we do the
1287                  * opposite of each operation here.
1288                  */
1289                 switch (tm->op) {
1290                 case MOD_LOG_KEY_REMOVE_WHILE_FREEING:
1291                         BUG_ON(tm->slot < n);
1292                         /* Fallthrough */
1293                 case MOD_LOG_KEY_REMOVE_WHILE_MOVING:
1294                 case MOD_LOG_KEY_REMOVE:
1295                         btrfs_set_node_key(eb, &tm->key, tm->slot);
1296                         btrfs_set_node_blockptr(eb, tm->slot, tm->blockptr);
1297                         btrfs_set_node_ptr_generation(eb, tm->slot,
1298                                                       tm->generation);
1299                         n++;
1300                         break;
1301                 case MOD_LOG_KEY_REPLACE:
1302                         BUG_ON(tm->slot >= n);
1303                         btrfs_set_node_key(eb, &tm->key, tm->slot);
1304                         btrfs_set_node_blockptr(eb, tm->slot, tm->blockptr);
1305                         btrfs_set_node_ptr_generation(eb, tm->slot,
1306                                                       tm->generation);
1307                         break;
1308                 case MOD_LOG_KEY_ADD:
1309                         /* if a move operation is needed it's in the log */
1310                         n--;
1311                         break;
1312                 case MOD_LOG_MOVE_KEYS:
1313                         o_dst = btrfs_node_key_ptr_offset(tm->slot);
1314                         o_src = btrfs_node_key_ptr_offset(tm->move.dst_slot);
1315                         memmove_extent_buffer(eb, o_dst, o_src,
1316                                               tm->move.nr_items * p_size);
1317                         break;
1318                 case MOD_LOG_ROOT_REPLACE:
1319                         /*
1320                          * this operation is special. for roots, this must be
1321                          * handled explicitly before rewinding.
1322                          * for non-roots, this operation may exist if the node
1323                          * was a root: root A -> child B; then A gets empty and
1324                          * B is promoted to the new root. in the mod log, we'll
1325                          * have a root-replace operation for B, a tree block
1326                          * that is no root. we simply ignore that operation.
1327                          */
1328                         break;
1329                 }
1330                 next = rb_next(&tm->node);
1331                 if (!next)
1332                         break;
1333                 tm = rb_entry(next, struct tree_mod_elem, node);
1334                 if (tm->logical != first_tm->logical)
1335                         break;
1336         }
1337         tree_mod_log_read_unlock(fs_info);
1338         btrfs_set_header_nritems(eb, n);
1339 }
1340 
1341 /*
1342  * Called with eb read locked. If the buffer cannot be rewound, the same buffer
1343  * is returned. If rewind operations happen, a fresh buffer is returned. The
1344  * returned buffer is always read-locked. If the returned buffer is not the
1345  * input buffer, the lock on the input buffer is released and the input buffer
1346  * is freed (its refcount is decremented).
1347  */
1348 static struct extent_buffer *
1349 tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct btrfs_path *path,
1350                     struct extent_buffer *eb, u64 time_seq)
1351 {
1352         struct extent_buffer *eb_rewin;
1353         struct tree_mod_elem *tm;
1354 
1355         if (!time_seq)
1356                 return eb;
1357 
1358         if (btrfs_header_level(eb) == 0)
1359                 return eb;
1360 
1361         tm = tree_mod_log_search(fs_info, eb->start, time_seq);
1362         if (!tm)
1363                 return eb;
1364 
1365         btrfs_set_path_blocking(path);
1366         btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1367 
1368         if (tm->op == MOD_LOG_KEY_REMOVE_WHILE_FREEING) {
1369                 BUG_ON(tm->slot != 0);
1370                 eb_rewin = alloc_dummy_extent_buffer(fs_info, eb->start);
1371                 if (!eb_rewin) {
1372                         btrfs_tree_read_unlock_blocking(eb);
1373                         free_extent_buffer(eb);
1374                         return NULL;
1375                 }
1376                 btrfs_set_header_bytenr(eb_rewin, eb->start);
1377                 btrfs_set_header_backref_rev(eb_rewin,
1378                                              btrfs_header_backref_rev(eb));
1379                 btrfs_set_header_owner(eb_rewin, btrfs_header_owner(eb));
1380                 btrfs_set_header_level(eb_rewin, btrfs_header_level(eb));
1381         } else {
1382                 eb_rewin = btrfs_clone_extent_buffer(eb);
1383                 if (!eb_rewin) {
1384                         btrfs_tree_read_unlock_blocking(eb);
1385                         free_extent_buffer(eb);
1386                         return NULL;
1387                 }
1388         }
1389 
1390         btrfs_clear_path_blocking(path, NULL, BTRFS_READ_LOCK);
1391         btrfs_tree_read_unlock_blocking(eb);
1392         free_extent_buffer(eb);
1393 
1394         extent_buffer_get(eb_rewin);
1395         btrfs_tree_read_lock(eb_rewin);
1396         __tree_mod_log_rewind(fs_info, eb_rewin, time_seq, tm);
1397         WARN_ON(btrfs_header_nritems(eb_rewin) >
1398                 BTRFS_NODEPTRS_PER_BLOCK(fs_info));
1399 
1400         return eb_rewin;
1401 }
1402 
1403 /*
1404  * get_old_root() rewinds the state of @root's root node to the given @time_seq
1405  * value. If there are no changes, the current root->root_node is returned. If
1406  * anything changed in between, there's a fresh buffer allocated on which the
1407  * rewind operations are done. In any case, the returned buffer is read locked.
1408  * Returns NULL on error (with no locks held).
1409  */
1410 static inline struct extent_buffer *
1411 get_old_root(struct btrfs_root *root, u64 time_seq)
1412 {
1413         struct btrfs_fs_info *fs_info = root->fs_info;
1414         struct tree_mod_elem *tm;
1415         struct extent_buffer *eb = NULL;
1416         struct extent_buffer *eb_root;
1417         struct extent_buffer *old;
1418         struct tree_mod_root *old_root = NULL;
1419         u64 old_generation = 0;
1420         u64 logical;
1421 
1422         eb_root = btrfs_read_lock_root_node(root);
1423         tm = __tree_mod_log_oldest_root(fs_info, eb_root, time_seq);
1424         if (!tm)
1425                 return eb_root;
1426 
1427         if (tm->op == MOD_LOG_ROOT_REPLACE) {
1428                 old_root = &tm->old_root;
1429                 old_generation = tm->generation;
1430                 logical = old_root->logical;
1431         } else {
1432                 logical = eb_root->start;
1433         }
1434 
1435         tm = tree_mod_log_search(fs_info, logical, time_seq);
1436         if (old_root && tm && tm->op != MOD_LOG_KEY_REMOVE_WHILE_FREEING) {
1437                 btrfs_tree_read_unlock(eb_root);
1438                 free_extent_buffer(eb_root);
1439                 old = read_tree_block(fs_info, logical, 0);
1440                 if (WARN_ON(IS_ERR(old) || !extent_buffer_uptodate(old))) {
1441                         if (!IS_ERR(old))
1442                                 free_extent_buffer(old);
1443                         btrfs_warn(fs_info,
1444                                    "failed to read tree block %llu from get_old_root",
1445                                    logical);
1446                 } else {
1447                         eb = btrfs_clone_extent_buffer(old);
1448                         free_extent_buffer(old);
1449                 }
1450         } else if (old_root) {
1451                 btrfs_tree_read_unlock(eb_root);
1452                 free_extent_buffer(eb_root);
1453                 eb = alloc_dummy_extent_buffer(fs_info, logical);
1454         } else {
1455                 btrfs_set_lock_blocking_rw(eb_root, BTRFS_READ_LOCK);
1456                 eb = btrfs_clone_extent_buffer(eb_root);
1457                 btrfs_tree_read_unlock_blocking(eb_root);
1458                 free_extent_buffer(eb_root);
1459         }
1460 
1461         if (!eb)
1462                 return NULL;
1463         extent_buffer_get(eb);
1464         btrfs_tree_read_lock(eb);
1465         if (old_root) {
1466                 btrfs_set_header_bytenr(eb, eb->start);
1467                 btrfs_set_header_backref_rev(eb, BTRFS_MIXED_BACKREF_REV);
1468                 btrfs_set_header_owner(eb, btrfs_header_owner(eb_root));
1469                 btrfs_set_header_level(eb, old_root->level);
1470                 btrfs_set_header_generation(eb, old_generation);
1471         }
1472         if (tm)
1473                 __tree_mod_log_rewind(fs_info, eb, time_seq, tm);
1474         else
1475                 WARN_ON(btrfs_header_level(eb) != 0);
1476         WARN_ON(btrfs_header_nritems(eb) > BTRFS_NODEPTRS_PER_BLOCK(fs_info));
1477 
1478         return eb;
1479 }
1480 
1481 int btrfs_old_root_level(struct btrfs_root *root, u64 time_seq)
1482 {
1483         struct tree_mod_elem *tm;
1484         int level;
1485         struct extent_buffer *eb_root = btrfs_root_node(root);
1486 
1487         tm = __tree_mod_log_oldest_root(root->fs_info, eb_root, time_seq);
1488         if (tm && tm->op == MOD_LOG_ROOT_REPLACE) {
1489                 level = tm->old_root.level;
1490         } else {
1491                 level = btrfs_header_level(eb_root);
1492         }
1493         free_extent_buffer(eb_root);
1494 
1495         return level;
1496 }
1497 
1498 static inline int should_cow_block(struct btrfs_trans_handle *trans,
1499                                    struct btrfs_root *root,
1500                                    struct extent_buffer *buf)
1501 {
1502         if (btrfs_is_testing(root->fs_info))
1503                 return 0;
1504 
1505         /* ensure we can see the force_cow */
1506         smp_rmb();
1507 
1508         /*
1509          * We do not need to cow a block if
1510          * 1) this block is not created or changed in this transaction;
1511          * 2) this block does not belong to TREE_RELOC tree;
1512          * 3) the root is not forced COW.
1513          *
1514          * What is forced COW:
1515          *    when we create snapshot during committing the transaction,
1516          *    after we've finished coping src root, we must COW the shared
1517          *    block to ensure the metadata consistency.
1518          */
1519         if (btrfs_header_generation(buf) == trans->transid &&
1520             !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN) &&
1521             !(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID &&
1522               btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC)) &&
1523             !test_bit(BTRFS_ROOT_FORCE_COW, &root->state))
1524                 return 0;
1525         return 1;
1526 }
1527 
1528 /*
1529  * cows a single block, see __btrfs_cow_block for the real work.
1530  * This version of it has extra checks so that a block isn't COWed more than
1531  * once per transaction, as long as it hasn't been written yet
1532  */
1533 noinline int btrfs_cow_block(struct btrfs_trans_handle *trans,
1534                     struct btrfs_root *root, struct extent_buffer *buf,
1535                     struct extent_buffer *parent, int parent_slot,
1536                     struct extent_buffer **cow_ret)
1537 {
1538         struct btrfs_fs_info *fs_info = root->fs_info;
1539         u64 search_start;
1540         int ret;
1541 
1542         if (trans->transaction != fs_info->running_transaction)
1543                 WARN(1, KERN_CRIT "trans %llu running %llu\n",
1544                        trans->transid,
1545                        fs_info->running_transaction->transid);
1546 
1547         if (trans->transid != fs_info->generation)
1548                 WARN(1, KERN_CRIT "trans %llu running %llu\n",
1549                        trans->transid, fs_info->generation);
1550 
1551         if (!should_cow_block(trans, root, buf)) {
1552                 trans->dirty = true;
1553                 *cow_ret = buf;
1554                 return 0;
1555         }
1556 
1557         search_start = buf->start & ~((u64)SZ_1G - 1);
1558 
1559         if (parent)
1560                 btrfs_set_lock_blocking(parent);
1561         btrfs_set_lock_blocking(buf);
1562 
1563         ret = __btrfs_cow_block(trans, root, buf, parent,
1564                                  parent_slot, cow_ret, search_start, 0);
1565 
1566         trace_btrfs_cow_block(root, buf, *cow_ret);
1567 
1568         return ret;
1569 }
1570 
1571 /*
1572  * helper function for defrag to decide if two blocks pointed to by a
1573  * node are actually close by
1574  */
1575 static int close_blocks(u64 blocknr, u64 other, u32 blocksize)
1576 {
1577         if (blocknr < other && other - (blocknr + blocksize) < 32768)
1578                 return 1;
1579         if (blocknr > other && blocknr - (other + blocksize) < 32768)
1580                 return 1;
1581         return 0;
1582 }
1583 
1584 /*
1585  * compare two keys in a memcmp fashion
1586  */
1587 static int comp_keys(const struct btrfs_disk_key *disk,
1588                      const struct btrfs_key *k2)
1589 {
1590         struct btrfs_key k1;
1591 
1592         btrfs_disk_key_to_cpu(&k1, disk);
1593 
1594         return btrfs_comp_cpu_keys(&k1, k2);
1595 }
1596 
1597 /*
1598  * same as comp_keys only with two btrfs_key's
1599  */
1600 int btrfs_comp_cpu_keys(const struct btrfs_key *k1, const struct btrfs_key *k2)
1601 {
1602         if (k1->objectid > k2->objectid)
1603                 return 1;
1604         if (k1->objectid < k2->objectid)
1605                 return -1;
1606         if (k1->type > k2->type)
1607                 return 1;
1608         if (k1->type < k2->type)
1609                 return -1;
1610         if (k1->offset > k2->offset)
1611                 return 1;
1612         if (k1->offset < k2->offset)
1613                 return -1;
1614         return 0;
1615 }
1616 
1617 /*
1618  * this is used by the defrag code to go through all the
1619  * leaves pointed to by a node and reallocate them so that
1620  * disk order is close to key order
1621  */
1622 int btrfs_realloc_node(struct btrfs_trans_handle *trans,
1623                        struct btrfs_root *root, struct extent_buffer *parent,
1624                        int start_slot, u64 *last_ret,
1625                        struct btrfs_key *progress)
1626 {
1627         struct btrfs_fs_info *fs_info = root->fs_info;
1628         struct extent_buffer *cur;
1629         u64 blocknr;
1630         u64 gen;
1631         u64 search_start = *last_ret;
1632         u64 last_block = 0;
1633         u64 other;
1634         u32 parent_nritems;
1635         int end_slot;
1636         int i;
1637         int err = 0;
1638         int parent_level;
1639         int uptodate;
1640         u32 blocksize;
1641         int progress_passed = 0;
1642         struct btrfs_disk_key disk_key;
1643 
1644         parent_level = btrfs_header_level(parent);
1645 
1646         WARN_ON(trans->transaction != fs_info->running_transaction);
1647         WARN_ON(trans->transid != fs_info->generation);
1648 
1649         parent_nritems = btrfs_header_nritems(parent);
1650         blocksize = fs_info->nodesize;
1651         end_slot = parent_nritems - 1;
1652 
1653         if (parent_nritems <= 1)
1654                 return 0;
1655 
1656         btrfs_set_lock_blocking(parent);
1657 
1658         for (i = start_slot; i <= end_slot; i++) {
1659                 int close = 1;
1660 
1661                 btrfs_node_key(parent, &disk_key, i);
1662                 if (!progress_passed && comp_keys(&disk_key, progress) < 0)
1663                         continue;
1664 
1665                 progress_passed = 1;
1666                 blocknr = btrfs_node_blockptr(parent, i);
1667                 gen = btrfs_node_ptr_generation(parent, i);
1668                 if (last_block == 0)
1669                         last_block = blocknr;
1670 
1671                 if (i > 0) {
1672                         other = btrfs_node_blockptr(parent, i - 1);
1673                         close = close_blocks(blocknr, other, blocksize);
1674                 }
1675                 if (!close && i < end_slot) {
1676                         other = btrfs_node_blockptr(parent, i + 1);
1677                         close = close_blocks(blocknr, other, blocksize);
1678                 }
1679                 if (close) {
1680                         last_block = blocknr;
1681                         continue;
1682                 }
1683 
1684                 cur = find_extent_buffer(fs_info, blocknr);
1685                 if (cur)
1686                         uptodate = btrfs_buffer_uptodate(cur, gen, 0);
1687                 else
1688                         uptodate = 0;
1689                 if (!cur || !uptodate) {
1690                         if (!cur) {
1691                                 cur = read_tree_block(fs_info, blocknr, gen);
1692                                 if (IS_ERR(cur)) {
1693                                         return PTR_ERR(cur);
1694                                 } else if (!extent_buffer_uptodate(cur)) {
1695                                         free_extent_buffer(cur);
1696                                         return -EIO;
1697                                 }
1698                         } else if (!uptodate) {
1699                                 err = btrfs_read_buffer(cur, gen);
1700                                 if (err) {
1701                                         free_extent_buffer(cur);
1702                                         return err;
1703                                 }
1704                         }
1705                 }
1706                 if (search_start == 0)
1707                         search_start = last_block;
1708 
1709                 btrfs_tree_lock(cur);
1710                 btrfs_set_lock_blocking(cur);
1711                 err = __btrfs_cow_block(trans, root, cur, parent, i,
1712                                         &cur, search_start,
1713                                         min(16 * blocksize,
1714                                             (end_slot - i) * blocksize));
1715                 if (err) {
1716                         btrfs_tree_unlock(cur);
1717                         free_extent_buffer(cur);
1718                         break;
1719                 }
1720                 search_start = cur->start;
1721                 last_block = cur->start;
1722                 *last_ret = search_start;
1723                 btrfs_tree_unlock(cur);
1724                 free_extent_buffer(cur);
1725         }
1726         return err;
1727 }
1728 
1729 /*
1730  * search for key in the extent_buffer.  The items start at offset p,
1731  * and they are item_size apart.  There are 'max' items in p.
1732  *
1733  * the slot in the array is returned via slot, and it points to
1734  * the place where you would insert key if it is not found in
1735  * the array.
1736  *
1737  * slot may point to max if the key is bigger than all of the keys
1738  */
1739 static noinline int generic_bin_search(struct extent_buffer *eb,
1740                                        unsigned long p, int item_size,
1741                                        const struct btrfs_key *key,
1742                                        int max, int *slot)
1743 {
1744         int low = 0;
1745         int high = max;
1746         int mid;
1747         int ret;
1748         struct btrfs_disk_key *tmp = NULL;
1749         struct btrfs_disk_key unaligned;
1750         unsigned long offset;
1751         char *kaddr = NULL;
1752         unsigned long map_start = 0;
1753         unsigned long map_len = 0;
1754         int err;
1755 
1756         if (low > high) {
1757                 btrfs_err(eb->fs_info,
1758                  "%s: low (%d) > high (%d) eb %llu owner %llu level %d",
1759                           __func__, low, high, eb->start,
1760                           btrfs_header_owner(eb), btrfs_header_level(eb));
1761                 return -EINVAL;
1762         }
1763 
1764         while (low < high) {
1765                 mid = (low + high) / 2;
1766                 offset = p + mid * item_size;
1767 
1768                 if (!kaddr || offset < map_start ||
1769                     (offset + sizeof(struct btrfs_disk_key)) >
1770                     map_start + map_len) {
1771 
1772                         err = map_private_extent_buffer(eb, offset,
1773                                                 sizeof(struct btrfs_disk_key),
1774                                                 &kaddr, &map_start, &map_len);
1775 
1776                         if (!err) {
1777                                 tmp = (struct btrfs_disk_key *)(kaddr + offset -
1778                                                         map_start);
1779                         } else if (err == 1) {
1780                                 read_extent_buffer(eb, &unaligned,
1781                                                    offset, sizeof(unaligned));
1782                                 tmp = &unaligned;
1783                         } else {
1784                                 return err;
1785                         }
1786 
1787                 } else {
1788                         tmp = (struct btrfs_disk_key *)(kaddr + offset -
1789                                                         map_start);
1790                 }
1791                 ret = comp_keys(tmp, key);
1792 
1793                 if (ret < 0)
1794                         low = mid + 1;
1795                 else if (ret > 0)
1796                         high = mid;
1797                 else {
1798                         *slot = mid;
1799                         return 0;
1800                 }
1801         }
1802         *slot = low;
1803         return 1;
1804 }
1805 
1806 /*
1807  * simple bin_search frontend that does the right thing for
1808  * leaves vs nodes
1809  */
1810 static int bin_search(struct extent_buffer *eb, const struct btrfs_key *key,
1811                       int level, int *slot)
1812 {
1813         if (level == 0)
1814                 return generic_bin_search(eb,
1815                                           offsetof(struct btrfs_leaf, items),
1816                                           sizeof(struct btrfs_item),
1817                                           key, btrfs_header_nritems(eb),
1818                                           slot);
1819         else
1820                 return generic_bin_search(eb,
1821                                           offsetof(struct btrfs_node, ptrs),
1822                                           sizeof(struct btrfs_key_ptr),
1823                                           key, btrfs_header_nritems(eb),
1824                                           slot);
1825 }
1826 
1827 int btrfs_bin_search(struct extent_buffer *eb, const struct btrfs_key *key,
1828                      int level, int *slot)
1829 {
1830         return bin_search(eb, key, level, slot);
1831 }
1832 
1833 static void root_add_used(struct btrfs_root *root, u32 size)
1834 {
1835         spin_lock(&root->accounting_lock);
1836         btrfs_set_root_used(&root->root_item,
1837                             btrfs_root_used(&root->root_item) + size);
1838         spin_unlock(&root->accounting_lock);
1839 }
1840 
1841 static void root_sub_used(struct btrfs_root *root, u32 size)
1842 {
1843         spin_lock(&root->accounting_lock);
1844         btrfs_set_root_used(&root->root_item,
1845                             btrfs_root_used(&root->root_item) - size);
1846         spin_unlock(&root->accounting_lock);
1847 }
1848 
1849 /* given a node and slot number, this reads the blocks it points to.  The
1850  * extent buffer is returned with a reference taken (but unlocked).
1851  */
1852 static noinline struct extent_buffer *
1853 read_node_slot(struct btrfs_fs_info *fs_info, struct extent_buffer *parent,
1854                int slot)
1855 {
1856         int level = btrfs_header_level(parent);
1857         struct extent_buffer *eb;
1858 
1859         if (slot < 0 || slot >= btrfs_header_nritems(parent))
1860                 return ERR_PTR(-ENOENT);
1861 
1862         BUG_ON(level == 0);
1863 
1864         eb = read_tree_block(fs_info, btrfs_node_blockptr(parent, slot),
1865                              btrfs_node_ptr_generation(parent, slot));
1866         if (!IS_ERR(eb) && !extent_buffer_uptodate(eb)) {
1867                 free_extent_buffer(eb);
1868                 eb = ERR_PTR(-EIO);
1869         }
1870 
1871         return eb;
1872 }
1873 
1874 /*
1875  * node level balancing, used to make sure nodes are in proper order for
1876  * item deletion.  We balance from the top down, so we have to make sure
1877  * that a deletion won't leave an node completely empty later on.
1878  */
1879 static noinline int balance_level(struct btrfs_trans_handle *trans,
1880                          struct btrfs_root *root,
1881                          struct btrfs_path *path, int level)
1882 {
1883         struct btrfs_fs_info *fs_info = root->fs_info;
1884         struct extent_buffer *right = NULL;
1885         struct extent_buffer *mid;
1886         struct extent_buffer *left = NULL;
1887         struct extent_buffer *parent = NULL;
1888         int ret = 0;
1889         int wret;
1890         int pslot;
1891         int orig_slot = path->slots[level];
1892         u64 orig_ptr;
1893 
1894         if (level == 0)
1895                 return 0;
1896 
1897         mid = path->nodes[level];
1898 
1899         WARN_ON(path->locks[level] != BTRFS_WRITE_LOCK &&
1900                 path->locks[level] != BTRFS_WRITE_LOCK_BLOCKING);
1901         WARN_ON(btrfs_header_generation(mid) != trans->transid);
1902 
1903         orig_ptr = btrfs_node_blockptr(mid, orig_slot);
1904 
1905         if (level < BTRFS_MAX_LEVEL - 1) {
1906                 parent = path->nodes[level + 1];
1907                 pslot = path->slots[level + 1];
1908         }
1909 
1910         /*
1911          * deal with the case where there is only one pointer in the root
1912          * by promoting the node below to a root
1913          */
1914         if (!parent) {
1915                 struct extent_buffer *child;
1916 
1917                 if (btrfs_header_nritems(mid) != 1)
1918                         return 0;
1919 
1920                 /* promote the child to a root */
1921                 child = read_node_slot(fs_info, mid, 0);
1922                 if (IS_ERR(child)) {
1923                         ret = PTR_ERR(child);
1924                         btrfs_handle_fs_error(fs_info, ret, NULL);
1925                         goto enospc;
1926                 }
1927 
1928                 btrfs_tree_lock(child);
1929                 btrfs_set_lock_blocking(child);
1930                 ret = btrfs_cow_block(trans, root, child, mid, 0, &child);
1931                 if (ret) {
1932                         btrfs_tree_unlock(child);
1933                         free_extent_buffer(child);
1934                         goto enospc;
1935                 }
1936 
1937                 tree_mod_log_set_root_pointer(root, child, 1);
1938                 rcu_assign_pointer(root->node, child);
1939 
1940                 add_root_to_dirty_list(root);
1941                 btrfs_tree_unlock(child);
1942 
1943                 path->locks[level] = 0;
1944                 path->nodes[level] = NULL;
1945                 clean_tree_block(fs_info, mid);
1946                 btrfs_tree_unlock(mid);
1947                 /* once for the path */
1948                 free_extent_buffer(mid);
1949 
1950                 root_sub_used(root, mid->len);
1951                 btrfs_free_tree_block(trans, root, mid, 0, 1);
1952                 /* once for the root ptr */
1953                 free_extent_buffer_stale(mid);
1954                 return 0;
1955         }
1956         if (btrfs_header_nritems(mid) >
1957             BTRFS_NODEPTRS_PER_BLOCK(fs_info) / 4)
1958                 return 0;
1959 
1960         left = read_node_slot(fs_info, parent, pslot - 1);
1961         if (IS_ERR(left))
1962                 left = NULL;
1963 
1964         if (left) {
1965                 btrfs_tree_lock(left);
1966                 btrfs_set_lock_blocking(left);
1967                 wret = btrfs_cow_block(trans, root, left,
1968                                        parent, pslot - 1, &left);
1969                 if (wret) {
1970                         ret = wret;
1971                         goto enospc;
1972                 }
1973         }
1974 
1975         right = read_node_slot(fs_info, parent, pslot + 1);
1976         if (IS_ERR(right))
1977                 right = NULL;
1978 
1979         if (right) {
1980                 btrfs_tree_lock(right);
1981                 btrfs_set_lock_blocking(right);
1982                 wret = btrfs_cow_block(trans, root, right,
1983                                        parent, pslot + 1, &right);
1984                 if (wret) {
1985                         ret = wret;
1986                         goto enospc;
1987                 }
1988         }
1989 
1990         /* first, try to make some room in the middle buffer */
1991         if (left) {
1992                 orig_slot += btrfs_header_nritems(left);
1993                 wret = push_node_left(trans, fs_info, left, mid, 1);
1994                 if (wret < 0)
1995                         ret = wret;
1996         }
1997 
1998         /*
1999          * then try to empty the right most buffer into the middle
2000          */
2001         if (right) {
2002                 wret = push_node_left(trans, fs_info, mid, right, 1);
2003                 if (wret < 0 && wret != -ENOSPC)
2004                         ret = wret;
2005                 if (btrfs_header_nritems(right) == 0) {
2006                         clean_tree_block(fs_info, right);
2007                         btrfs_tree_unlock(right);
2008                         del_ptr(root, path, level + 1, pslot + 1);
2009                         root_sub_used(root, right->len);
2010                         btrfs_free_tree_block(trans, root, right, 0, 1);
2011                         free_extent_buffer_stale(right);
2012                         right = NULL;
2013                 } else {
2014                         struct btrfs_disk_key right_key;
2015                         btrfs_node_key(right, &right_key, 0);
2016                         tree_mod_log_set_node_key(fs_info, parent,
2017                                                   pslot + 1, 0);
2018                         btrfs_set_node_key(parent, &right_key, pslot + 1);
2019                         btrfs_mark_buffer_dirty(parent);
2020                 }
2021         }
2022         if (btrfs_header_nritems(mid) == 1) {
2023                 /*
2024                  * we're not allowed to leave a node with one item in the
2025                  * tree during a delete.  A deletion from lower in the tree
2026                  * could try to delete the only pointer in this node.
2027                  * So, pull some keys from the left.
2028                  * There has to be a left pointer at this point because
2029                  * otherwise we would have pulled some pointers from the
2030                  * right
2031                  */
2032                 if (!left) {
2033                         ret = -EROFS;
2034                         btrfs_handle_fs_error(fs_info, ret, NULL);
2035                         goto enospc;
2036                 }
2037                 wret = balance_node_right(trans, fs_info, mid, left);
2038                 if (wret < 0) {
2039                         ret = wret;
2040                         goto enospc;
2041                 }
2042                 if (wret == 1) {
2043                         wret = push_node_left(trans, fs_info, left, mid, 1);
2044                         if (wret < 0)
2045                                 ret = wret;
2046                 }
2047                 BUG_ON(wret == 1);
2048         }
2049         if (btrfs_header_nritems(mid) == 0) {
2050                 clean_tree_block(fs_info, mid);
2051                 btrfs_tree_unlock(mid);
2052                 del_ptr(root, path, level + 1, pslot);
2053                 root_sub_used(root, mid->len);
2054                 btrfs_free_tree_block(trans, root, mid, 0, 1);
2055                 free_extent_buffer_stale(mid);
2056                 mid = NULL;
2057         } else {
2058                 /* update the parent key to reflect our changes */
2059                 struct btrfs_disk_key mid_key;
2060                 btrfs_node_key(mid, &mid_key, 0);
2061                 tree_mod_log_set_node_key(fs_info, parent, pslot, 0);
2062                 btrfs_set_node_key(parent, &mid_key, pslot);
2063                 btrfs_mark_buffer_dirty(parent);
2064         }
2065 
2066         /* update the path */
2067         if (left) {
2068                 if (btrfs_header_nritems(left) > orig_slot) {
2069                         extent_buffer_get(left);
2070                         /* left was locked after cow */
2071                         path->nodes[level] = left;
2072                         path->slots[level + 1] -= 1;
2073                         path->slots[level] = orig_slot;
2074                         if (mid) {
2075                                 btrfs_tree_unlock(mid);
2076                                 free_extent_buffer(mid);
2077                         }
2078                 } else {
2079                         orig_slot -= btrfs_header_nritems(left);
2080                         path->slots[level] = orig_slot;
2081                 }
2082         }
2083         /* double check we haven't messed things up */
2084         if (orig_ptr !=
2085             btrfs_node_blockptr(path->nodes[level], path->slots[level]))
2086                 BUG();
2087 enospc:
2088         if (right) {
2089                 btrfs_tree_unlock(right);
2090                 free_extent_buffer(right);
2091         }
2092         if (left) {
2093                 if (path->nodes[level] != left)
2094                         btrfs_tree_unlock(left);
2095                 free_extent_buffer(left);
2096         }
2097         return ret;
2098 }
2099 
2100 /* Node balancing for insertion.  Here we only split or push nodes around
2101  * when they are completely full.  This is also done top down, so we
2102  * have to be pessimistic.
2103  */
2104 static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans,
2105                                           struct btrfs_root *root,
2106                                           struct btrfs_path *path, int level)
2107 {
2108         struct btrfs_fs_info *fs_info = root->fs_info;
2109         struct extent_buffer *right = NULL;
2110         struct extent_buffer *mid;
2111         struct extent_buffer *left = NULL;
2112         struct extent_buffer *parent = NULL;
2113         int ret = 0;
2114         int wret;
2115         int pslot;
2116         int orig_slot = path->slots[level];
2117 
2118         if (level == 0)
2119                 return 1;
2120 
2121         mid = path->nodes[level];
2122         WARN_ON(btrfs_header_generation(mid) != trans->transid);
2123 
2124         if (level < BTRFS_MAX_LEVEL - 1) {
2125                 parent = path->nodes[level + 1];
2126                 pslot = path->slots[level + 1];
2127         }
2128 
2129         if (!parent)
2130                 return 1;
2131 
2132         left = read_node_slot(fs_info, parent, pslot - 1);
2133         if (IS_ERR(left))
2134                 left = NULL;
2135 
2136         /* first, try to make some room in the middle buffer */
2137         if (left) {
2138                 u32 left_nr;
2139 
2140                 btrfs_tree_lock(left);
2141                 btrfs_set_lock_blocking(left);
2142 
2143                 left_nr = btrfs_header_nritems(left);
2144                 if (left_nr >= BTRFS_NODEPTRS_PER_BLOCK(fs_info) - 1) {
2145                         wret = 1;
2146                 } else {
2147                         ret = btrfs_cow_block(trans, root, left, parent,
2148                                               pslot - 1, &left);
2149                         if (ret)
2150                                 wret = 1;
2151                         else {
2152                                 wret = push_node_left(trans, fs_info,
2153                                                       left, mid, 0);
2154                         }
2155                 }
2156                 if (wret < 0)
2157                         ret = wret;
2158                 if (wret == 0) {
2159                         struct btrfs_disk_key disk_key;
2160                         orig_slot += left_nr;
2161                         btrfs_node_key(mid, &disk_key, 0);
2162                         tree_mod_log_set_node_key(fs_info, parent, pslot, 0);
2163                         btrfs_set_node_key(parent, &disk_key, pslot);
2164                         btrfs_mark_buffer_dirty(parent);
2165                         if (btrfs_header_nritems(left) > orig_slot) {
2166                                 path->nodes[level] = left;
2167                                 path->slots[level + 1] -= 1;
2168                                 path->slots[level] = orig_slot;
2169                                 btrfs_tree_unlock(mid);
2170                                 free_extent_buffer(mid);
2171                         } else {
2172                                 orig_slot -=
2173                                         btrfs_header_nritems(left);
2174                                 path->slots[level] = orig_slot;
2175                                 btrfs_tree_unlock(left);
2176                                 free_extent_buffer(left);
2177                         }
2178                         return 0;
2179                 }
2180                 btrfs_tree_unlock(left);
2181                 free_extent_buffer(left);
2182         }
2183         right = read_node_slot(fs_info, parent, pslot + 1);
2184         if (IS_ERR(right))
2185                 right = NULL;
2186 
2187         /*
2188          * then try to empty the right most buffer into the middle
2189          */
2190         if (right) {
2191                 u32 right_nr;
2192 
2193                 btrfs_tree_lock(right);
2194                 btrfs_set_lock_blocking(right);
2195 
2196                 right_nr = btrfs_header_nritems(right);
2197                 if (right_nr >= BTRFS_NODEPTRS_PER_BLOCK(fs_info) - 1) {
2198                         wret = 1;
2199                 } else {
2200                         ret = btrfs_cow_block(trans, root, right,
2201                                               parent, pslot + 1,
2202                                               &right);
2203                         if (ret)
2204                                 wret = 1;
2205                         else {
2206                                 wret = balance_node_right(trans, fs_info,
2207                                                           right, mid);
2208                         }
2209                 }
2210                 if (wret < 0)
2211                         ret = wret;
2212                 if (wret == 0) {
2213                         struct btrfs_disk_key disk_key;
2214 
2215                         btrfs_node_key(right, &disk_key, 0);
2216                         tree_mod_log_set_node_key(fs_info, parent,
2217                                                   pslot + 1, 0);
2218                         btrfs_set_node_key(parent, &disk_key, pslot + 1);
2219                         btrfs_mark_buffer_dirty(parent);
2220 
2221                         if (btrfs_header_nritems(mid) <= orig_slot) {
2222                                 path->nodes[level] = right;
2223                                 path->slots[level + 1] += 1;
2224                                 path->slots[level] = orig_slot -
2225                                         btrfs_header_nritems(mid);
2226                                 btrfs_tree_unlock(mid);
2227                                 free_extent_buffer(mid);
2228                         } else {
2229                                 btrfs_tree_unlock(right);
2230                                 free_extent_buffer(right);
2231                         }
2232                         return 0;
2233                 }
2234                 btrfs_tree_unlock(right);
2235                 free_extent_buffer(right);
2236         }
2237         return 1;
2238 }
2239 
2240 /*
2241  * readahead one full node of leaves, finding things that are close
2242  * to the block in 'slot', and triggering ra on them.
2243  */
2244 static void reada_for_search(struct btrfs_fs_info *fs_info,
2245                              struct btrfs_path *path,
2246                              int level, int slot, u64 objectid)
2247 {
2248         struct extent_buffer *node;
2249         struct btrfs_disk_key disk_key;
2250         u32 nritems;
2251         u64 search;
2252         u64 target;
2253         u64 nread = 0;
2254         struct extent_buffer *eb;
2255         u32 nr;
2256         u32 blocksize;
2257         u32 nscan = 0;
2258 
2259         if (level != 1)
2260                 return;
2261 
2262         if (!path->nodes[level])
2263                 return;
2264 
2265         node = path->nodes[level];
2266 
2267         search = btrfs_node_blockptr(node, slot);
2268         blocksize = fs_info->nodesize;
2269         eb = find_extent_buffer(fs_info, search);
2270         if (eb) {
2271                 free_extent_buffer(eb);
2272                 return;
2273         }
2274 
2275         target = search;
2276 
2277         nritems = btrfs_header_nritems(node);
2278         nr = slot;
2279 
2280         while (1) {
2281                 if (path->reada == READA_BACK) {
2282                         if (nr == 0)
2283                                 break;
2284                         nr--;
2285                 } else if (path->reada == READA_FORWARD) {
2286                         nr++;
2287                         if (nr >= nritems)
2288                                 break;
2289                 }
2290                 if (path->reada == READA_BACK && objectid) {
2291                         btrfs_node_key(node, &disk_key, nr);
2292                         if (btrfs_disk_key_objectid(&disk_key) != objectid)
2293                                 break;
2294                 }
2295                 search = btrfs_node_blockptr(node, nr);
2296                 if ((search <= target && target - search <= 65536) ||
2297                     (search > target && search - target <= 65536)) {
2298                         readahead_tree_block(fs_info, search);
2299                         nread += blocksize;
2300                 }
2301                 nscan++;
2302                 if ((nread > 65536 || nscan > 32))
2303                         break;
2304         }
2305 }
2306 
2307 static noinline void reada_for_balance(struct btrfs_fs_info *fs_info,
2308                                        struct btrfs_path *path, int level)
2309 {
2310         int slot;
2311         int nritems;
2312         struct extent_buffer *parent;
2313         struct extent_buffer *eb;
2314         u64 gen;
2315         u64 block1 = 0;
2316         u64 block2 = 0;
2317 
2318         parent = path->nodes[level + 1];
2319         if (!parent)
2320                 return;
2321 
2322         nritems = btrfs_header_nritems(parent);
2323         slot = path->slots[level + 1];
2324 
2325         if (slot > 0) {
2326                 block1 = btrfs_node_blockptr(parent, slot - 1);
2327                 gen = btrfs_node_ptr_generation(parent, slot - 1);
2328                 eb = find_extent_buffer(fs_info, block1);
2329                 /*
2330                  * if we get -eagain from btrfs_buffer_uptodate, we
2331                  * don't want to return eagain here.  That will loop
2332                  * forever
2333                  */
2334                 if (eb && btrfs_buffer_uptodate(eb, gen, 1) != 0)
2335                         block1 = 0;
2336                 free_extent_buffer(eb);
2337         }
2338         if (slot + 1 < nritems) {
2339                 block2 = btrfs_node_blockptr(parent, slot + 1);
2340                 gen = btrfs_node_ptr_generation(parent, slot + 1);
2341                 eb = find_extent_buffer(fs_info, block2);
2342                 if (eb && btrfs_buffer_uptodate(eb, gen, 1) != 0)
2343                         block2 = 0;
2344                 free_extent_buffer(eb);
2345         }
2346 
2347         if (block1)
2348                 readahead_tree_block(fs_info, block1);
2349         if (block2)
2350                 readahead_tree_block(fs_info, block2);
2351 }
2352 
2353 
2354 /*
2355  * when we walk down the tree, it is usually safe to unlock the higher layers
2356  * in the tree.  The exceptions are when our path goes through slot 0, because
2357  * operations on the tree might require changing key pointers higher up in the
2358  * tree.
2359  *
2360  * callers might also have set path->keep_locks, which tells this code to keep
2361  * the lock if the path points to the last slot in the block.  This is part of
2362  * walking through the tree, and selecting the next slot in the higher block.
2363  *
2364  * lowest_unlock sets the lowest level in the tree we're allowed to unlock.  so
2365  * if lowest_unlock is 1, level 0 won't be unlocked
2366  */
2367 static noinline void unlock_up(struct btrfs_path *path, int level,
2368                                int lowest_unlock, int min_write_lock_level,
2369                                int *write_lock_level)
2370 {
2371         int i;
2372         int skip_level = level;
2373         int no_skips = 0;
2374         struct extent_buffer *t;
2375 
2376         for (i = level; i < BTRFS_MAX_LEVEL; i++) {
2377                 if (!path->nodes[i])
2378                         break;
2379                 if (!path->locks[i])
2380                         break;
2381                 if (!no_skips && path->slots[i] == 0) {
2382                         skip_level = i + 1;
2383                         continue;
2384                 }
2385                 if (!no_skips && path->keep_locks) {
2386                         u32 nritems;
2387                         t = path->nodes[i];
2388                         nritems = btrfs_header_nritems(t);
2389                         if (nritems < 1 || path->slots[i] >= nritems - 1) {
2390                                 skip_level = i + 1;
2391                                 continue;
2392                         }
2393                 }
2394                 if (skip_level < i && i >= lowest_unlock)
2395                         no_skips = 1;
2396 
2397                 t = path->nodes[i];
2398                 if (i >= lowest_unlock && i > skip_level && path->locks[i]) {
2399                         btrfs_tree_unlock_rw(t, path->locks[i]);
2400                         path->locks[i] = 0;
2401                         if (write_lock_level &&
2402                             i > min_write_lock_level &&
2403                             i <= *write_lock_level) {
2404                                 *write_lock_level = i - 1;
2405                         }
2406                 }
2407         }
2408 }
2409 
2410 /*
2411  * This releases any locks held in the path starting at level and
2412  * going all the way up to the root.
2413  *
2414  * btrfs_search_slot will keep the lock held on higher nodes in a few
2415  * corner cases, such as COW of the block at slot zero in the node.  This
2416  * ignores those rules, and it should only be called when there are no
2417  * more updates to be done higher up in the tree.
2418  */
2419 noinline void btrfs_unlock_up_safe(struct btrfs_path *path, int level)
2420 {
2421         int i;
2422 
2423         if (path->keep_locks)
2424                 return;
2425 
2426         for (i = level; i < BTRFS_MAX_LEVEL; i++) {
2427                 if (!path->nodes[i])
2428                         continue;
2429                 if (!path->locks[i])
2430                         continue;
2431                 btrfs_tree_unlock_rw(path->nodes[i], path->locks[i]);
2432                 path->locks[i] = 0;
2433         }
2434 }
2435 
2436 /*
2437  * helper function for btrfs_search_slot.  The goal is to find a block
2438  * in cache without setting the path to blocking.  If we find the block
2439  * we return zero and the path is unchanged.
2440  *
2441  * If we can't find the block, we set the path blocking and do some
2442  * reada.  -EAGAIN is returned and the search must be repeated.
2443  */
2444 static int
2445 read_block_for_search(struct btrfs_root *root, struct btrfs_path *p,
2446                       struct extent_buffer **eb_ret, int level, int slot,
2447                       const struct btrfs_key *key)
2448 {
2449         struct btrfs_fs_info *fs_info = root->fs_info;
2450         u64 blocknr;
2451         u64 gen;
2452         struct extent_buffer *b = *eb_ret;
2453         struct extent_buffer *tmp;
2454         int ret;
2455 
2456         blocknr = btrfs_node_blockptr(b, slot);
2457         gen = btrfs_node_ptr_generation(b, slot);
2458 
2459         tmp = find_extent_buffer(fs_info, blocknr);
2460         if (tmp) {
2461                 /* first we do an atomic uptodate check */
2462                 if (btrfs_buffer_uptodate(tmp, gen, 1) > 0) {
2463                         *eb_ret = tmp;
2464                         return 0;
2465                 }
2466 
2467                 /* the pages were up to date, but we failed
2468                  * the generation number check.  Do a full
2469                  * read for the generation number that is correct.
2470                  * We must do this without dropping locks so
2471                  * we can trust our generation number
2472                  */
2473                 btrfs_set_path_blocking(p);
2474 
2475                 /* now we're allowed to do a blocking uptodate check */
2476                 ret = btrfs_read_buffer(tmp, gen);
2477                 if (!ret) {
2478                         *eb_ret = tmp;
2479                         return 0;
2480                 }
2481                 free_extent_buffer(tmp);
2482                 btrfs_release_path(p);
2483                 return -EIO;
2484         }
2485 
2486         /*
2487          * reduce lock contention at high levels
2488          * of the btree by dropping locks before
2489          * we read.  Don't release the lock on the current
2490          * level because we need to walk this node to figure
2491          * out which blocks to read.
2492          */
2493         btrfs_unlock_up_safe(p, level + 1);
2494         btrfs_set_path_blocking(p);
2495 
2496         free_extent_buffer(tmp);
2497         if (p->reada != READA_NONE)
2498                 reada_for_search(fs_info, p, level, slot, key->objectid);
2499 
2500         btrfs_release_path(p);
2501 
2502         ret = -EAGAIN;
2503         tmp = read_tree_block(fs_info, blocknr, 0);
2504         if (!IS_ERR(tmp)) {
2505                 /*
2506                  * If the read above didn't mark this buffer up to date,
2507                  * it will never end up being up to date.  Set ret to EIO now
2508                  * and give up so that our caller doesn't loop forever
2509                  * on our EAGAINs.
2510                  */
2511                 if (!btrfs_buffer_uptodate(tmp, 0, 0))
2512                         ret = -EIO;
2513                 free_extent_buffer(tmp);
2514         } else {
2515                 ret = PTR_ERR(tmp);
2516         }
2517         return ret;
2518 }
2519 
2520 /*
2521  * helper function for btrfs_search_slot.  This does all of the checks
2522  * for node-level blocks and does any balancing required based on
2523  * the ins_len.
2524  *
2525  * If no extra work was required, zero is returned.  If we had to
2526  * drop the path, -EAGAIN is returned and btrfs_search_slot must
2527  * start over
2528  */
2529 static int
2530 setup_nodes_for_search(struct btrfs_trans_handle *trans,
2531                        struct btrfs_root *root, struct btrfs_path *p,
2532                        struct extent_buffer *b, int level, int ins_len,
2533                        int *write_lock_level)
2534 {
2535         struct btrfs_fs_info *fs_info = root->fs_info;
2536         int ret;
2537 
2538         if ((p->search_for_split || ins_len > 0) && btrfs_header_nritems(b) >=
2539             BTRFS_NODEPTRS_PER_BLOCK(fs_info) - 3) {
2540                 int sret;
2541 
2542                 if (*write_lock_level < level + 1) {
2543                         *write_lock_level = level + 1;
2544                         btrfs_release_path(p);
2545                         goto again;
2546                 }
2547 
2548                 btrfs_set_path_blocking(p);
2549                 reada_for_balance(fs_info, p, level);
2550                 sret = split_node(trans, root, p, level);
2551                 btrfs_clear_path_blocking(p, NULL, 0);
2552 
2553                 BUG_ON(sret > 0);
2554                 if (sret) {
2555                         ret = sret;
2556                         goto done;
2557                 }
2558                 b = p->nodes[level];
2559         } else if (ins_len < 0 && btrfs_header_nritems(b) <
2560                    BTRFS_NODEPTRS_PER_BLOCK(fs_info) / 2) {
2561                 int sret;
2562 
2563                 if (*write_lock_level < level + 1) {
2564                         *write_lock_level = level + 1;
2565                         btrfs_release_path(p);
2566                         goto again;
2567                 }
2568 
2569                 btrfs_set_path_blocking(p);
2570                 reada_for_balance(fs_info, p, level);
2571                 sret = balance_level(trans, root, p, level);
2572                 btrfs_clear_path_blocking(p, NULL, 0);
2573 
2574                 if (sret) {
2575                         ret = sret;
2576                         goto done;
2577                 }
2578                 b = p->nodes[level];
2579                 if (!b) {
2580                         btrfs_release_path(p);
2581                         goto again;
2582                 }
2583                 BUG_ON(btrfs_header_nritems(b) == 1);
2584         }
2585         return 0;
2586 
2587 again:
2588         ret = -EAGAIN;
2589 done:
2590         return ret;
2591 }
2592 
2593 static void key_search_validate(struct extent_buffer *b,
2594                                 const struct btrfs_key *key,
2595                                 int level)
2596 {
2597 #ifdef CONFIG_BTRFS_ASSERT
2598         struct btrfs_disk_key disk_key;
2599 
2600         btrfs_cpu_key_to_disk(&disk_key, key);
2601 
2602         if (level == 0)
2603                 ASSERT(!memcmp_extent_buffer(b, &disk_key,
2604                     offsetof(struct btrfs_leaf, items[0].key),
2605                     sizeof(disk_key)));
2606         else
2607                 ASSERT(!memcmp_extent_buffer(b, &disk_key,
2608                     offsetof(struct btrfs_node, ptrs[0].key),
2609                     sizeof(disk_key)));
2610 #endif
2611 }
2612 
2613 static int key_search(struct extent_buffer *b, const struct btrfs_key *key,
2614                       int level, int *prev_cmp, int *slot)
2615 {
2616         if (*prev_cmp != 0) {
2617                 *prev_cmp = bin_search(b, key, level, slot);
2618                 return *prev_cmp;
2619         }
2620 
2621         key_search_validate(b, key, level);
2622         *slot = 0;
2623 
2624         return 0;
2625 }
2626 
2627 int btrfs_find_item(struct btrfs_root *fs_root, struct btrfs_path *path,
2628                 u64 iobjectid, u64 ioff, u8 key_type,
2629                 struct btrfs_key *found_key)
2630 {
2631         int ret;
2632         struct btrfs_key key;
2633         struct extent_buffer *eb;
2634 
2635         ASSERT(path);
2636         ASSERT(found_key);
2637 
2638         key.type = key_type;
2639         key.objectid = iobjectid;
2640         key.offset = ioff;
2641 
2642         ret = btrfs_search_slot(NULL, fs_root, &key, path, 0, 0);
2643         if (ret < 0)
2644                 return ret;
2645 
2646         eb = path->nodes[0];
2647         if (ret && path->slots[0] >= btrfs_header_nritems(eb)) {
2648                 ret = btrfs_next_leaf(fs_root, path);
2649                 if (ret)
2650                         return ret;
2651                 eb = path->nodes[0];
2652         }
2653 
2654         btrfs_item_key_to_cpu(eb, found_key, path->slots[0]);
2655         if (found_key->type != key.type ||
2656                         found_key->objectid != key.objectid)
2657                 return 1;
2658 
2659         return 0;
2660 }
2661 
2662 /*
2663  * look for key in the tree.  path is filled in with nodes along the way
2664  * if key is found, we return zero and you can find the item in the leaf
2665  * level of the path (level 0)
2666  *
2667  * If the key isn't found, the path points to the slot where it should
2668  * be inserted, and 1 is returned.  If there are other errors during the
2669  * search a negative error number is returned.
2670  *
2671  * if ins_len > 0, nodes and leaves will be split as we walk down the
2672  * tree.  if ins_len < 0, nodes will be merged as we walk down the tree (if
2673  * possible)
2674  */
2675 int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root *root,
2676                       const struct btrfs_key *key, struct btrfs_path *p,
2677                       int ins_len, int cow)
2678 {
2679         struct btrfs_fs_info *fs_info = root->fs_info;
2680         struct extent_buffer *b;
2681         int slot;
2682         int ret;
2683         int err;
2684         int level;
2685         int lowest_unlock = 1;
2686         int root_lock;
2687         /* everything at write_lock_level or lower must be write locked */
2688         int write_lock_level = 0;
2689         u8 lowest_level = 0;
2690         int min_write_lock_level;
2691         int prev_cmp;
2692 
2693         lowest_level = p->lowest_level;
2694         WARN_ON(lowest_level && ins_len > 0);
2695         WARN_ON(p->nodes[0] != NULL);
2696         BUG_ON(!cow && ins_len);
2697 
2698         if (ins_len < 0) {
2699                 lowest_unlock = 2;
2700 
2701                 /* when we are removing items, we might have to go up to level
2702                  * two as we update tree pointers  Make sure we keep write
2703                  * for those levels as well
2704                  */
2705                 write_lock_level = 2;
2706         } else if (ins_len > 0) {
2707                 /*
2708                  * for inserting items, make sure we have a write lock on
2709                  * level 1 so we can update keys
2710                  */
2711                 write_lock_level = 1;
2712         }
2713 
2714         if (!cow)
2715                 write_lock_level = -1;
2716 
2717         if (cow && (p->keep_locks || p->lowest_level))
2718                 write_lock_level = BTRFS_MAX_LEVEL;
2719 
2720         min_write_lock_level = write_lock_level;
2721 
2722 again:
2723         prev_cmp = -1;
2724         /*
2725          * we try very hard to do read locks on the root
2726          */
2727         root_lock = BTRFS_READ_LOCK;
2728         level = 0;
2729         if (p->search_commit_root) {
2730                 /*
2731                  * the commit roots are read only
2732                  * so we always do read locks
2733                  */
2734                 if (p->need_commit_sem)
2735                         down_read(&fs_info->commit_root_sem);
2736                 b = root->commit_root;
2737                 extent_buffer_get(b);
2738                 level = btrfs_header_level(b);
2739                 if (p->need_commit_sem)
2740                         up_read(&fs_info->commit_root_sem);
2741                 if (!p->skip_locking)
2742                         btrfs_tree_read_lock(b);
2743         } else {
2744                 if (p->skip_locking) {
2745                         b = btrfs_root_node(root);
2746                         level = btrfs_header_level(b);
2747                 } else {
2748                         /* we don't know the level of the root node
2749                          * until we actually have it read locked
2750                          */
2751                         b = btrfs_read_lock_root_node(root);
2752                         level = btrfs_header_level(b);
2753                         if (level <= write_lock_level) {
2754                                 /* whoops, must trade for write lock */
2755                                 btrfs_tree_read_unlock(b);
2756                                 free_extent_buffer(b);
2757                                 b = btrfs_lock_root_node(root);
2758                                 root_lock = BTRFS_WRITE_LOCK;
2759 
2760                                 /* the level might have changed, check again */
2761                                 level = btrfs_header_level(b);
2762                         }
2763                 }
2764         }
2765         p->nodes[level] = b;
2766         if (!p->skip_locking)
2767                 p->locks[level] = root_lock;
2768 
2769         while (b) {
2770                 level = btrfs_header_level(b);
2771 
2772                 /*
2773                  * setup the path here so we can release it under lock
2774                  * contention with the cow code
2775                  */
2776                 if (cow) {
2777                         /*
2778                          * if we don't really need to cow this block
2779                          * then we don't want to set the path blocking,
2780                          * so we test it here
2781                          */
2782                         if (!should_cow_block(trans, root, b)) {
2783                                 trans->dirty = true;
2784                                 goto cow_done;
2785                         }
2786 
2787                         /*
2788                          * must have write locks on this node and the
2789                          * parent
2790                          */
2791                         if (level > write_lock_level ||
2792                             (level + 1 > write_lock_level &&
2793                             level + 1 < BTRFS_MAX_LEVEL &&
2794                             p->nodes[level + 1])) {
2795                                 write_lock_level = level + 1;
2796                                 btrfs_release_path(p);
2797                                 goto again;
2798                         }
2799 
2800                         btrfs_set_path_blocking(p);
2801                         err = btrfs_cow_block(trans, root, b,
2802                                               p->nodes[level + 1],
2803                                               p->slots[level + 1], &b);
2804                         if (err) {
2805                                 ret = err;
2806                                 goto done;
2807                         }
2808                 }
2809 cow_done:
2810                 p->nodes[level] = b;
2811                 btrfs_clear_path_blocking(p, NULL, 0);
2812 
2813                 /*
2814                  * we have a lock on b and as long as we aren't changing
2815                  * the tree, there is no way to for the items in b to change.
2816                  * It is safe to drop the lock on our parent before we
2817                  * go through the expensive btree search on b.
2818                  *
2819                  * If we're inserting or deleting (ins_len != 0), then we might
2820                  * be changing slot zero, which may require changing the parent.
2821                  * So, we can't drop the lock until after we know which slot
2822                  * we're operating on.
2823                  */
2824                 if (!ins_len && !p->keep_locks) {
2825                         int u = level + 1;
2826 
2827                         if (u < BTRFS_MAX_LEVEL && p->locks[u]) {
2828                                 btrfs_tree_unlock_rw(p->nodes[u], p->locks[u]);
2829                                 p->locks[u] = 0;
2830                         }
2831                 }
2832 
2833                 ret = key_search(b, key, level, &prev_cmp, &slot);
2834                 if (ret < 0)
2835                         goto done;
2836 
2837                 if (level != 0) {
2838                         int dec = 0;
2839                         if (ret && slot > 0) {
2840                                 dec = 1;
2841                                 slot -= 1;
2842                         }
2843                         p->slots[level] = slot;
2844                         err = setup_nodes_for_search(trans, root, p, b, level,
2845                                              ins_len, &write_lock_level);
2846                         if (err == -EAGAIN)
2847                                 goto again;
2848                         if (err) {
2849                                 ret = err;
2850                                 goto done;
2851                         }
2852                         b = p->nodes[level];
2853                         slot = p->slots[level];
2854 
2855                         /*
2856                          * slot 0 is special, if we change the key
2857                          * we have to update the parent pointer
2858                          * which means we must have a write lock
2859                          * on the parent
2860                          */
2861                         if (slot == 0 && ins_len &&
2862                             write_lock_level < level + 1) {
2863                                 write_lock_level = level + 1;
2864                                 btrfs_release_path(p);
2865                                 goto again;
2866                         }
2867 
2868                         unlock_up(p, level, lowest_unlock,
2869                                   min_write_lock_level, &write_lock_level);
2870 
2871                         if (level == lowest_level) {
2872                                 if (dec)
2873                                         p->slots[level]++;
2874                                 goto done;
2875                         }
2876 
2877                         err = read_block_for_search(root, p, &b, level,
2878                                                     slot, key);
2879                         if (err == -EAGAIN)
2880                                 goto again;
2881                         if (err) {
2882                                 ret = err;
2883                                 goto done;
2884                         }
2885 
2886                         if (!p->skip_locking) {
2887                                 level = btrfs_header_level(b);
2888                                 if (level <= write_lock_level) {
2889                                         err = btrfs_try_tree_write_lock(b);
2890                                         if (!err) {
2891                                                 btrfs_set_path_blocking(p);
2892                                                 btrfs_tree_lock(b);
2893                                                 btrfs_clear_path_blocking(p, b,
2894                                                                   BTRFS_WRITE_LOCK);
2895                                         }
2896                                         p->locks[level] = BTRFS_WRITE_LOCK;
2897                                 } else {
2898                                         err = btrfs_tree_read_lock_atomic(b);
2899                                         if (!err) {
2900                                                 btrfs_set_path_blocking(p);
2901                                                 btrfs_tree_read_lock(b);
2902                                                 btrfs_clear_path_blocking(p, b,
2903                                                                   BTRFS_READ_LOCK);
2904                                         }
2905                                         p->locks[level] = BTRFS_READ_LOCK;
2906                                 }
2907                                 p->nodes[level] = b;
2908                         }
2909                 } else {
2910                         p->slots[level] = slot;
2911                         if (ins_len > 0 &&
2912                             btrfs_leaf_free_space(fs_info, b) < ins_len) {
2913                                 if (write_lock_level < 1) {
2914                                         write_lock_level = 1;
2915                                         btrfs_release_path(p);
2916                                         goto again;
2917                                 }
2918 
2919                                 btrfs_set_path_blocking(p);
2920                                 err = split_leaf(trans, root, key,
2921                                                  p, ins_len, ret == 0);
2922                                 btrfs_clear_path_blocking(p, NULL, 0);
2923 
2924                                 BUG_ON(err > 0);
2925                                 if (err) {
2926                                         ret = err;
2927                                         goto done;
2928                                 }
2929                         }
2930                         if (!p->search_for_split)
2931                                 unlock_up(p, level, lowest_unlock,
2932                                           min_write_lock_level, &write_lock_level);
2933                         goto done;
2934                 }
2935         }
2936         ret = 1;
2937 done:
2938         /*
2939          * we don't really know what they plan on doing with the path
2940          * from here on, so for now just mark it as blocking
2941          */
2942         if (!p->leave_spinning)
2943                 btrfs_set_path_blocking(p);
2944         if (ret < 0 && !p->skip_release_on_error)
2945                 btrfs_release_path(p);
2946         return ret;
2947 }
2948 
2949 /*
2950  * Like btrfs_search_slot, this looks for a key in the given tree. It uses the
2951  * current state of the tree together with the operations recorded in the tree
2952  * modification log to search for the key in a previous version of this tree, as
2953  * denoted by the time_seq parameter.
2954  *
2955  * Naturally, there is no support for insert, delete or cow operations.
2956  *
2957  * The resulting path and return value will be set up as if we called
2958  * btrfs_search_slot at that point in time with ins_len and cow both set to 0.
2959  */
2960 int btrfs_search_old_slot(struct btrfs_root *root, const struct btrfs_key *key,
2961                           struct btrfs_path *p, u64 time_seq)
2962 {
2963         struct btrfs_fs_info *fs_info = root->fs_info;
2964         struct extent_buffer *b;
2965         int slot;
2966         int ret;
2967         int err;
2968         int level;
2969         int lowest_unlock = 1;
2970         u8 lowest_level = 0;
2971         int prev_cmp = -1;
2972 
2973         lowest_level = p->lowest_level;
2974         WARN_ON(p->nodes[0] != NULL);
2975 
2976         if (p->search_commit_root) {
2977                 BUG_ON(time_seq);
2978                 return btrfs_search_slot(NULL, root, key, p, 0, 0);
2979         }
2980 
2981 again:
2982         b = get_old_root(root, time_seq);
2983         level = btrfs_header_level(b);
2984         p->locks[level] = BTRFS_READ_LOCK;
2985 
2986         while (b) {
2987                 level = btrfs_header_level(b);
2988                 p->nodes[level] = b;
2989                 btrfs_clear_path_blocking(p, NULL, 0);
2990 
2991                 /*
2992                  * we have a lock on b and as long as we aren't changing
2993                  * the tree, there is no way to for the items in b to change.
2994                  * It is safe to drop the lock on our parent before we
2995                  * go through the expensive btree search on b.
2996                  */
2997                 btrfs_unlock_up_safe(p, level + 1);
2998 
2999                 /*
3000                  * Since we can unwind ebs we want to do a real search every
3001                  * time.
3002                  */
3003                 prev_cmp = -1;
3004                 ret = key_search(b, key, level, &prev_cmp, &slot);
3005 
3006                 if (level != 0) {
3007                         int dec = 0;
3008                         if (ret && slot > 0) {
3009                                 dec = 1;
3010                                 slot -= 1;
3011                         }
3012                         p->slots[level] = slot;
3013                         unlock_up(p, level, lowest_unlock, 0, NULL);
3014 
3015                         if (level == lowest_level) {
3016                                 if (dec)
3017                                         p->slots[level]++;
3018                                 goto done;
3019                         }
3020 
3021                         err = read_block_for_search(root, p, &b, level,
3022                                                     slot, key);
3023                         if (err == -EAGAIN)
3024                                 goto again;
3025                         if (err) {
3026                                 ret = err;
3027                                 goto done;
3028                         }
3029 
3030                         level = btrfs_header_level(b);
3031                         err = btrfs_tree_read_lock_atomic(b);
3032                         if (!err) {
3033                                 btrfs_set_path_blocking(p);
3034                                 btrfs_tree_read_lock(b);
3035                                 btrfs_clear_path_blocking(p, b,
3036                                                           BTRFS_READ_LOCK);
3037                         }
3038                         b = tree_mod_log_rewind(fs_info, p, b, time_seq);
3039                         if (!b) {
3040                                 ret = -ENOMEM;
3041                                 goto done;
3042                         }
3043                         p->locks[level] = BTRFS_READ_LOCK;
3044                         p->nodes[level] = b;
3045                 } else {
3046                         p->slots[level] = slot;
3047                         unlock_up(p, level, lowest_unlock, 0, NULL);
3048                         goto done;
3049                 }
3050         }
3051         ret = 1;
3052 done:
3053         if (!p->leave_spinning)
3054                 btrfs_set_path_blocking(p);
3055         if (ret < 0)
3056                 btrfs_release_path(p);
3057 
3058         return ret;
3059 }
3060 
3061 /*
3062  * helper to use instead of search slot if no exact match is needed but
3063  * instead the next or previous item should be returned.
3064  * When find_higher is true, the next higher item is returned, the next lower
3065  * otherwise.
3066  * When return_any and find_higher are both true, and no higher item is found,
3067  * return the next lower instead.
3068  * When return_any is true and find_higher is false, and no lower item is found,
3069  * return the next higher instead.
3070  * It returns 0 if any item is found, 1 if none is found (tree empty), and
3071  * < 0 on error
3072  */
3073 int btrfs_search_slot_for_read(struct btrfs_root *root,
3074                                const struct btrfs_key *key,
3075                                struct btrfs_path *p, int find_higher,
3076                                int return_any)
3077 {
3078         int ret;
3079         struct extent_buffer *leaf;
3080 
3081 again:
3082         ret = btrfs_search_slot(NULL, root, key, p, 0, 0);
3083         if (ret <= 0)
3084                 return ret;
3085         /*
3086          * a return value of 1 means the path is at the position where the
3087          * item should be inserted. Normally this is the next bigger item,
3088          * but in case the previous item is the last in a leaf, path points
3089          * to the first free slot in the previous leaf, i.e. at an invalid
3090          * item.
3091          */
3092         leaf = p->nodes[0];
3093 
3094         if (find_higher) {
3095                 if (p->slots[0] >= btrfs_header_nritems(leaf)) {
3096                         ret = btrfs_next_leaf(root, p);
3097                         if (ret <= 0)
3098                                 return ret;
3099                         if (!return_any)
3100                                 return 1;
3101                         /*
3102                          * no higher item found, return the next
3103                          * lower instead
3104                          */
3105                         return_any = 0;
3106                         find_higher = 0;
3107                         btrfs_release_path(p);
3108                         goto again;
3109                 }
3110         } else {
3111                 if (p->slots[0] == 0) {
3112                         ret = btrfs_prev_leaf(root, p);
3113                         if (ret < 0)
3114                                 return ret;
3115                         if (!ret) {
3116                                 leaf = p->nodes[0];
3117                                 if (p->slots[0] == btrfs_header_nritems(leaf))
3118                                         p->slots[0]--;
3119                                 return 0;
3120                         }
3121                         if (!return_any)
3122                                 return 1;
3123                         /*
3124                          * no lower item found, return the next
3125                          * higher instead
3126                          */
3127                         return_any = 0;
3128                         find_higher = 1;
3129                         btrfs_release_path(p);
3130                         goto again;
3131                 } else {
3132                         --p->slots[0];
3133                 }
3134         }
3135         return 0;
3136 }
3137 
3138 /*
3139  * adjust the pointers going up the tree, starting at level
3140  * making sure the right key of each node is points to 'key'.
3141  * This is used after shifting pointers to the left, so it stops
3142  * fixing up pointers when a given leaf/node is not in slot 0 of the
3143  * higher levels
3144  *
3145  */
3146 static void fixup_low_keys(struct btrfs_fs_info *fs_info,
3147                            struct btrfs_path *path,
3148                            struct btrfs_disk_key *key, int level)
3149 {
3150         int i;
3151         struct extent_buffer *t;
3152 
3153         for (i = level; i < BTRFS_MAX_LEVEL; i++) {
3154                 int tslot = path->slots[i];
3155                 if (!path->nodes[i])
3156                         break;
3157                 t = path->nodes[i];
3158                 tree_mod_log_set_node_key(fs_info, t, tslot, 1);
3159                 btrfs_set_node_key(t, key, tslot);
3160                 btrfs_mark_buffer_dirty(path->nodes[i]);
3161                 if (tslot != 0)
3162                         break;
3163         }
3164 }
3165 
3166 /*
3167  * update item key.
3168  *
3169  * This function isn't completely safe. It's the caller's responsibility
3170  * that the new key won't break the order
3171  */
3172 void btrfs_set_item_key_safe(struct btrfs_fs_info *fs_info,
3173                              struct btrfs_path *path,
3174                              const struct btrfs_key *new_key)
3175 {
3176         struct btrfs_disk_key disk_key;
3177         struct extent_buffer *eb;
3178         int slot;
3179 
3180         eb = path->nodes[0];
3181         slot = path->slots[0];
3182         if (slot > 0) {
3183                 btrfs_item_key(eb, &disk_key, slot - 1);
3184                 BUG_ON(comp_keys(&disk_key, new_key) >= 0);
3185         }
3186         if (slot < btrfs_header_nritems(eb) - 1) {
3187                 btrfs_item_key(eb, &disk_key, slot + 1);
3188                 BUG_ON(comp_keys(&disk_key, new_key) <= 0);
3189         }
3190 
3191         btrfs_cpu_key_to_disk(&disk_key, new_key);
3192         btrfs_set_item_key(eb, &disk_key, slot);
3193         btrfs_mark_buffer_dirty(eb);
3194         if (slot == 0)
3195                 fixup_low_keys(fs_info, path, &disk_key, 1);
3196 }
3197 
3198 /*
3199  * try to push data from one node into the next node left in the
3200  * tree.
3201  *
3202  * returns 0 if some ptrs were pushed left, < 0 if there was some horrible
3203  * error, and > 0 if there was no room in the left hand block.
3204  */
3205 static int push_node_left(struct btrfs_trans_handle *trans,
3206                           struct btrfs_fs_info *fs_info,
3207                           struct extent_buffer *dst,
3208                           struct extent_buffer *src, int empty)
3209 {
3210         int push_items = 0;
3211         int src_nritems;
3212         int dst_nritems;
3213         int ret = 0;
3214 
3215         src_nritems = btrfs_header_nritems(src);
3216         dst_nritems = btrfs_header_nritems(dst);
3217         push_items = BTRFS_NODEPTRS_PER_BLOCK(fs_info) - dst_nritems;
3218         WARN_ON(btrfs_header_generation(src) != trans->transid);
3219         WARN_ON(btrfs_header_generation(dst) != trans->transid);
3220 
3221         if (!empty && src_nritems <= 8)
3222                 return 1;
3223 
3224         if (push_items <= 0)
3225                 return 1;
3226 
3227         if (empty) {
3228                 push_items = min(src_nritems, push_items);
3229                 if (push_items < src_nritems) {
3230                         /* leave at least 8 pointers in the node if
3231                          * we aren't going to empty it
3232                          */
3233                         if (src_nritems - push_items < 8) {
3234                                 if (push_items <= 8)
3235                                         return 1;
3236                                 push_items -= 8;
3237                         }
3238                 }
3239         } else
3240                 push_items = min(src_nritems - 8, push_items);
3241 
3242         ret = tree_mod_log_eb_copy(fs_info, dst, src, dst_nritems, 0,
3243                                    push_items);
3244         if (ret) {
3245                 btrfs_abort_transaction(trans, ret);
3246                 return ret;
3247         }
3248         copy_extent_buffer(dst, src,
3249                            btrfs_node_key_ptr_offset(dst_nritems),
3250                            btrfs_node_key_ptr_offset(0),
3251                            push_items * sizeof(struct btrfs_key_ptr));
3252 
3253         if (push_items < src_nritems) {
3254                 /*
3255                  * don't call tree_mod_log_eb_move here, key removal was already
3256                  * fully logged by tree_mod_log_eb_copy above.
3257                  */
3258                 memmove_extent_buffer(src, btrfs_node_key_ptr_offset(0),
3259                                       btrfs_node_key_ptr_offset(push_items),
3260                                       (src_nritems - push_items) *
3261                                       sizeof(struct btrfs_key_ptr));
3262         }
3263         btrfs_set_header_nritems(src, src_nritems - push_items);
3264         btrfs_set_header_nritems(dst, dst_nritems + push_items);
3265         btrfs_mark_buffer_dirty(src);
3266         btrfs_mark_buffer_dirty(dst);
3267 
3268         return ret;
3269 }
3270 
3271 /*
3272  * try to push data from one node into the next node right in the
3273  * tree.
3274  *
3275  * returns 0 if some ptrs were pushed, < 0 if there was some horrible
3276  * error, and > 0 if there was no room in the right hand block.
3277  *
3278  * this will  only push up to 1/2 the contents of the left node over
3279  */
3280 static int balance_node_right(struct btrfs_trans_handle *trans,
3281                               struct btrfs_fs_info *fs_info,
3282                               struct extent_buffer *dst,
3283                               struct extent_buffer *src)
3284 {
3285         int push_items = 0;
3286         int max_push;
3287         int src_nritems;
3288         int dst_nritems;
3289         int ret = 0;
3290 
3291         WARN_ON(btrfs_header_generation(src) != trans->transid);
3292         WARN_ON(btrfs_header_generation(dst) != trans->transid);
3293 
3294         src_nritems = btrfs_header_nritems(src);
3295         dst_nritems = btrfs_header_nritems(dst);
3296         push_items = BTRFS_NODEPTRS_PER_BLOCK(fs_info) - dst_nritems;
3297         if (push_items <= 0)
3298                 return 1;
3299 
3300         if (src_nritems < 4)
3301                 return 1;
3302 
3303         max_push = src_nritems / 2 + 1;
3304         /* don't try to empty the node */
3305         if (max_push >= src_nritems)
3306                 return 1;
3307 
3308         if (max_push < push_items)
3309                 push_items = max_push;
3310 
3311         tree_mod_log_eb_move(fs_info, dst, push_items, 0, dst_nritems);
3312         memmove_extent_buffer(dst, btrfs_node_key_ptr_offset(push_items),
3313                                       btrfs_node_key_ptr_offset(0),
3314                                       (dst_nritems) *
3315                                       sizeof(struct btrfs_key_ptr));
3316 
3317         ret = tree_mod_log_eb_copy(fs_info, dst, src, 0,
3318                                    src_nritems - push_items, push_items);
3319         if (ret) {
3320                 btrfs_abort_transaction(trans, ret);
3321                 return ret;
3322         }
3323         copy_extent_buffer(dst, src,
3324                            btrfs_node_key_ptr_offset(0),
3325                            btrfs_node_key_ptr_offset(src_nritems - push_items),
3326                            push_items * sizeof(struct btrfs_key_ptr));
3327 
3328         btrfs_set_header_nritems(src, src_nritems - push_items);
3329         btrfs_set_header_nritems(dst, dst_nritems + push_items);
3330 
3331         btrfs_mark_buffer_dirty(src);
3332         btrfs_mark_buffer_dirty(dst);
3333 
3334         return ret;
3335 }
3336 
3337 /*
3338  * helper function to insert a new root level in the tree.
3339  * A new node is allocated, and a single item is inserted to
3340  * point to the existing root
3341  *
3342  * returns zero on success or < 0 on failure.
3343  */
3344 static noinline int insert_new_root(struct btrfs_trans_handle *trans,
3345                            struct btrfs_root *root,
3346                            struct btrfs_path *path, int level)
3347 {
3348         struct btrfs_fs_info *fs_info = root->fs_info;
3349         u64 lower_gen;
3350         struct extent_buffer *lower;
3351         struct extent_buffer *c;
3352         struct extent_buffer *old;
3353         struct btrfs_disk_key lower_key;
3354 
3355         BUG_ON(path->nodes[level]);
3356         BUG_ON(path->nodes[level-1] != root->node);
3357 
3358         lower = path->nodes[level-1];
3359         if (level == 1)
3360                 btrfs_item_key(lower, &lower_key, 0);
3361         else
3362                 btrfs_node_key(lower, &lower_key, 0);
3363 
3364         c = btrfs_alloc_tree_block(trans, root, 0, root->root_key.objectid,
3365                                    &lower_key, level, root->node->start, 0);
3366         if (IS_ERR(c))
3367                 return PTR_ERR(c);
3368 
3369         root_add_used(root, fs_info->nodesize);
3370 
3371         memzero_extent_buffer(c, 0, sizeof(struct btrfs_header));
3372         btrfs_set_header_nritems(c, 1);
3373         btrfs_set_header_level(c, level);
3374         btrfs_set_header_bytenr(c, c->start);
3375         btrfs_set_header_generation(c, trans->transid);
3376         btrfs_set_header_backref_rev(c, BTRFS_MIXED_BACKREF_REV);
3377         btrfs_set_header_owner(c, root->root_key.objectid);
3378 
3379         write_extent_buffer_fsid(c, fs_info->fsid);
3380         write_extent_buffer_chunk_tree_uuid(c, fs_info->chunk_tree_uuid);
3381 
3382         btrfs_set_node_key(c, &lower_key, 0);
3383         btrfs_set_node_blockptr(c, 0, lower->start);
3384         lower_gen = btrfs_header_generation(lower);
3385         WARN_ON(lower_gen != trans->transid);
3386 
3387         btrfs_set_node_ptr_generation(c, 0, lower_gen);
3388 
3389         btrfs_mark_buffer_dirty(c);
3390 
3391         old = root->node;
3392         tree_mod_log_set_root_pointer(root, c, 0);
3393         rcu_assign_pointer(root->node, c);
3394 
3395         /* the super has an extra ref to root->node */
3396         free_extent_buffer(old);
3397 
3398         add_root_to_dirty_list(root);
3399         extent_buffer_get(c);
3400         path->nodes[level] = c;
3401         path->locks[level] = BTRFS_WRITE_LOCK_BLOCKING;
3402         path->slots[level] = 0;
3403         return 0;
3404 }
3405 
3406 /*
3407  * worker function to insert a single pointer in a node.
3408  * the node should have enough room for the pointer already
3409  *
3410  * slot and level indicate where you want the key to go, and
3411  * blocknr is the block the key points to.
3412  */
3413 static void insert_ptr(struct btrfs_trans_handle *trans,
3414                        struct btrfs_fs_info *fs_info, struct btrfs_path *path,
3415                        struct btrfs_disk_key *key, u64 bytenr,
3416                        int slot, int level)
3417 {
3418         struct extent_buffer *lower;
3419         int nritems;
3420         int ret;
3421 
3422         BUG_ON(!path->nodes[level]);
3423         btrfs_assert_tree_locked(path->nodes[level]);
3424         lower = path->nodes[level];
3425         nritems = btrfs_header_nritems(lower);
3426         BUG_ON(slot > nritems);
3427         BUG_ON(nritems == BTRFS_NODEPTRS_PER_BLOCK(fs_info));
3428         if (slot != nritems) {
3429                 if (level)
3430                         tree_mod_log_eb_move(fs_info, lower, slot + 1,
3431                                              slot, nritems - slot);
3432                 memmove_extent_buffer(lower,
3433                               btrfs_node_key_ptr_offset(slot + 1),
3434                               btrfs_node_key_ptr_offset(slot),
3435                               (nritems - slot) * sizeof(struct btrfs_key_ptr));
3436         }
3437         if (level) {
3438                 ret = tree_mod_log_insert_key(fs_info, lower, slot,
3439                                               MOD_LOG_KEY_ADD, GFP_NOFS);
3440                 BUG_ON(ret < 0);
3441         }
3442         btrfs_set_node_key(lower, key, slot);
3443         btrfs_set_node_blockptr(lower, slot, bytenr);
3444         WARN_ON(trans->transid == 0);
3445         btrfs_set_node_ptr_generation(lower, slot, trans->transid);
3446         btrfs_set_header_nritems(lower, nritems + 1);
3447         btrfs_mark_buffer_dirty(lower);
3448 }
3449 
3450 /*
3451  * split the node at the specified level in path in two.
3452  * The path is corrected to point to the appropriate node after the split
3453  *
3454  * Before splitting this tries to make some room in the node by pushing
3455  * left and right, if either one works, it returns right away.
3456  *
3457  * returns 0 on success and < 0 on failure
3458  */
3459 static noinline int split_node(struct btrfs_trans_handle *trans,
3460                                struct btrfs_root *root,
3461                                struct btrfs_path *path, int level)
3462 {
3463         struct btrfs_fs_info *fs_info = root->fs_info;
3464         struct extent_buffer *c;
3465         struct extent_buffer *split;
3466         struct btrfs_disk_key disk_key;
3467         int mid;
3468         int ret;
3469         u32 c_nritems;
3470 
3471         c = path->nodes[level];
3472         WARN_ON(btrfs_header_generation(c) != trans->transid);
3473         if (c == root->node) {
3474                 /*
3475                  * trying to split the root, lets make a new one
3476                  *
3477                  * tree mod log: We don't log_removal old root in
3478                  * insert_new_root, because that root buffer will be kept as a
3479                  * normal node. We are going to log removal of half of the
3480                  * elements below with tree_mod_log_eb_copy. We're holding a
3481                  * tree lock on the buffer, which is why we cannot race with
3482                  * other tree_mod_log users.
3483                  */
3484                 ret = insert_new_root(trans, root, path, level + 1);
3485                 if (ret)
3486                         return ret;
3487         } else {
3488                 ret = push_nodes_for_insert(trans, root, path, level);
3489                 c = path->nodes[level];
3490                 if (!ret && btrfs_header_nritems(c) <
3491                     BTRFS_NODEPTRS_PER_BLOCK(fs_info) - 3)
3492                         return 0;
3493                 if (ret < 0)
3494                         return ret;
3495         }
3496 
3497         c_nritems = btrfs_header_nritems(c);
3498         mid = (c_nritems + 1) / 2;
3499         btrfs_node_key(c, &disk_key, mid);
3500 
3501         split = btrfs_alloc_tree_block(trans, root, 0, root->root_key.objectid,
3502                         &disk_key, level, c->start, 0);
3503         if (IS_ERR(split))
3504                 return PTR_ERR(split);
3505 
3506         root_add_used(root, fs_info->nodesize);
3507 
3508         memzero_extent_buffer(split, 0, sizeof(struct btrfs_header));
3509         btrfs_set_header_level(split, btrfs_header_level(c));
3510         btrfs_set_header_bytenr(split, split->start);
3511         btrfs_set_header_generation(split, trans->transid);
3512         btrfs_set_header_backref_rev(split, BTRFS_MIXED_BACKREF_REV);
3513         btrfs_set_header_owner(split, root->root_key.objectid);
3514         write_extent_buffer_fsid(split, fs_info->fsid);
3515         write_extent_buffer_chunk_tree_uuid(split, fs_info->chunk_tree_uuid);
3516 
3517         ret = tree_mod_log_eb_copy(fs_info, split, c, 0, mid, c_nritems - mid);
3518         if (ret) {
3519                 btrfs_abort_transaction(trans, ret);
3520                 return ret;
3521         }
3522         copy_extent_buffer(split, c,
3523                            btrfs_node_key_ptr_offset(0),
3524                            btrfs_node_key_ptr_offset(mid),
3525                            (c_nritems - mid) * sizeof(struct btrfs_key_ptr));
3526         btrfs_set_header_nritems(split, c_nritems - mid);
3527         btrfs_set_header_nritems(c, mid);
3528         ret = 0;
3529 
3530         btrfs_mark_buffer_dirty(c);
3531         btrfs_mark_buffer_dirty(split);
3532 
3533         insert_ptr(trans, fs_info, path, &disk_key, split->start,
3534                    path->slots[level + 1] + 1, level + 1);
3535 
3536         if (path->slots[level] >= mid) {
3537                 path->slots[level] -= mid;
3538                 btrfs_tree_unlock(c);
3539                 free_extent_buffer(c);
3540                 path->nodes[level] = split;
3541                 path->slots[level + 1] += 1;
3542         } else {
3543                 btrfs_tree_unlock(split);
3544                 free_extent_buffer(split);
3545         }
3546         return ret;
3547 }
3548 
3549 /*
3550  * how many bytes are required to store the items in a leaf.  start
3551  * and nr indicate which items in the leaf to check.  This totals up the
3552  * space used both by the item structs and the item data
3553  */
3554 static int leaf_space_used(struct extent_buffer *l, int start, int nr)
3555 {
3556         struct btrfs_item *start_item;
3557         struct btrfs_item *end_item;
3558         struct btrfs_map_token token;
3559         int data_len;
3560         int nritems = btrfs_header_nritems(l);
3561         int end = min(nritems, start + nr) - 1;
3562 
3563         if (!nr)
3564                 return 0;
3565         btrfs_init_map_token(&token);
3566         start_item = btrfs_item_nr(start);
3567         end_item = btrfs_item_nr(end);
3568         data_len = btrfs_token_item_offset(l, start_item, &token) +
3569                 btrfs_token_item_size(l, start_item, &token);
3570         data_len = data_len - btrfs_token_item_offset(l, end_item, &token);
3571         data_len += sizeof(struct btrfs_item) * nr;
3572         WARN_ON(data_len < 0);
3573         return data_len;
3574 }
3575 
3576 /*
3577  * The space between the end of the leaf items and
3578  * the start of the leaf data.  IOW, how much room
3579  * the leaf has left for both items and data
3580  */
3581 noinline int btrfs_leaf_free_space(struct btrfs_fs_info *fs_info,
3582                                    struct extent_buffer *leaf)
3583 {
3584         int nritems = btrfs_header_nritems(leaf);
3585         int ret;
3586 
3587         ret = BTRFS_LEAF_DATA_SIZE(fs_info) - leaf_space_used(leaf, 0, nritems);
3588         if (ret < 0) {
3589                 btrfs_crit(fs_info,
3590                            "leaf free space ret %d, leaf data size %lu, used %d nritems %d",
3591                            ret,
3592                            (unsigned long) BTRFS_LEAF_DATA_SIZE(fs_info),
3593                            leaf_space_used(leaf, 0, nritems), nritems);
3594         }
3595         return ret;
3596 }
3597 
3598 /*
3599  * min slot controls the lowest index we're willing to push to the
3600  * right.  We'll push up to and including min_slot, but no lower
3601  */
3602 static noinline int __push_leaf_right(struct btrfs_fs_info *fs_info,
3603                                       struct btrfs_path *path,
3604                                       int data_size, int empty,
3605                                       struct extent_buffer *right,
3606                                       int free_space, u32 left_nritems,
3607                                       u32 min_slot)
3608 {
3609         struct extent_buffer *left = path->nodes[0];
3610         struct extent_buffer *upper = path->nodes[1];
3611         struct btrfs_map_token token;
3612         struct btrfs_disk_key disk_key;
3613         int slot;
3614         u32 i;
3615         int push_space = 0;
3616         int push_items = 0;
3617         struct btrfs_item *item;
3618         u32 nr;
3619         u32 right_nritems;
3620         u32 data_end;
3621         u32 this_item_size;
3622 
3623         btrfs_init_map_token(&token);
3624 
3625         if (empty)
3626                 nr = 0;
3627         else
3628                 nr = max_t(u32, 1, min_slot);
3629 
3630         if (path->slots[0] >= left_nritems)
3631                 push_space += data_size;
3632 
3633         slot = path->slots[1];
3634         i = left_nritems - 1;
3635         while (i >= nr) {
3636                 item = btrfs_item_nr(i);
3637 
3638                 if (!empty && push_items > 0) {
3639                         if (path->slots[0] > i)
3640                                 break;
3641                         if (path->slots[0] == i) {
3642                                 int space = btrfs_leaf_free_space(fs_info, left);
3643                                 if (space + push_space * 2 > free_space)
3644                                         break;
3645                         }
3646                 }
3647 
3648                 if (path->slots[0] == i)
3649                         push_space += data_size;
3650 
3651                 this_item_size = btrfs_item_size(left, item);
3652                 if (this_item_size + sizeof(*item) + push_space > free_space)
3653                         break;
3654 
3655                 push_items++;
3656                 push_space += this_item_size + sizeof(*item);
3657                 if (i == 0)
3658                         break;
3659                 i--;
3660         }
3661 
3662         if (push_items == 0)
3663                 goto out_unlock;
3664 
3665         WARN_ON(!empty && push_items == left_nritems);
3666 
3667         /* push left to right */
3668         right_nritems = btrfs_header_nritems(right);
3669 
3670         push_space = btrfs_item_end_nr(left, left_nritems - push_items);
3671         push_space -= leaf_data_end(fs_info, left);
3672 
3673         /* make room in the right data area */
3674         data_end = leaf_data_end(fs_info, right);
3675         memmove_extent_buffer(right,
3676                               BTRFS_LEAF_DATA_OFFSET + data_end - push_space,
3677                               BTRFS_LEAF_DATA_OFFSET + data_end,
3678                               BTRFS_LEAF_DATA_SIZE(fs_info) - data_end);
3679 
3680         /* copy from the left data area */
3681         copy_extent_buffer(right, left, BTRFS_LEAF_DATA_OFFSET +
3682                      BTRFS_LEAF_DATA_SIZE(fs_info) - push_space,
3683                      BTRFS_LEAF_DATA_OFFSET + leaf_data_end(fs_info, left),
3684                      push_space);
3685 
3686         memmove_extent_buffer(right, btrfs_item_nr_offset(push_items),
3687                               btrfs_item_nr_offset(0),
3688                               right_nritems * sizeof(struct btrfs_item));
3689 
3690         /* copy the items from left to right */
3691         copy_extent_buffer(right, left, btrfs_item_nr_offset(0),
3692                    btrfs_item_nr_offset(left_nritems - push_items),
3693                    push_items * sizeof(struct btrfs_item));
3694 
3695         /* update the item pointers */
3696         right_nritems += push_items;
3697         btrfs_set_header_nritems(right, right_nritems);
3698         push_space = BTRFS_LEAF_DATA_SIZE(fs_info);
3699         for (i = 0; i < right_nritems; i++) {
3700                 item = btrfs_item_nr(i);
3701                 push_space -= btrfs_token_item_size(right, item, &token);
3702                 btrfs_set_token_item_offset(right, item, push_space, &token);
3703         }
3704 
3705         left_nritems -= push_items;
3706         btrfs_set_header_nritems(left, left_nritems);
3707 
3708         if (left_nritems)
3709                 btrfs_mark_buffer_dirty(left);
3710         else
3711                 clean_tree_block(fs_info, left);
3712 
3713         btrfs_mark_buffer_dirty(right);
3714 
3715         btrfs_item_key(right, &disk_key, 0);
3716         btrfs_set_node_key(upper, &disk_key, slot + 1);
3717         btrfs_mark_buffer_dirty(upper);
3718 
3719         /* then fixup the leaf pointer in the path */
3720         if (path->slots[0] >= left_nritems) {
3721                 path->slots[0] -= left_nritems;
3722                 if (btrfs_header_nritems(path->nodes[0]) == 0)
3723                         clean_tree_block(fs_info, path->nodes[0]);
3724                 btrfs_tree_unlock(path->nodes[0]);
3725                 free_extent_buffer(path->nodes[0]);
3726                 path->nodes[0] = right;
3727                 path->slots[1] += 1;
3728         } else {
3729                 btrfs_tree_unlock(right);
3730                 free_extent_buffer(right);
3731         }
3732         return 0;
3733 
3734 out_unlock:
3735         btrfs_tree_unlock(right);
3736         free_extent_buffer(right);
3737         return 1;
3738 }
3739 
3740 /*
3741  * push some data in the path leaf to the right, trying to free up at
3742  * least data_size bytes.  returns zero if the push worked, nonzero otherwise
3743  *
3744  * returns 1 if the push failed because the other node didn't have enough
3745  * room, 0 if everything worked out and < 0 if there were major errors.
3746  *
3747  * this will push starting from min_slot to the end of the leaf.  It won't
3748  * push any slot lower than min_slot
3749  */
3750 static int push_leaf_right(struct btrfs_trans_handle *trans, struct btrfs_root
3751                            *root, struct btrfs_path *path,
3752                            int min_data_size, int data_size,
3753                            int empty, u32 min_slot)
3754 {
3755         struct btrfs_fs_info *fs_info = root->fs_info;
3756         struct extent_buffer *left = path->nodes[0];
3757         struct extent_buffer *right;
3758         struct extent_buffer *upper;
3759         int slot;
3760         int free_space;
3761         u32 left_nritems;
3762         int ret;
3763 
3764         if (!path->nodes[1])
3765                 return 1;
3766 
3767         slot = path->slots[1];
3768         upper = path->nodes[1];
3769         if (slot >= btrfs_header_nritems(upper) - 1)
3770                 return 1;
3771 
3772         btrfs_assert_tree_locked(path->nodes[1]);
3773 
3774         right = read_node_slot(fs_info, upper, slot + 1);
3775         /*
3776          * slot + 1 is not valid or we fail to read the right node,
3777          * no big deal, just return.
3778          */
3779         if (IS_ERR(right))
3780                 return 1;
3781 
3782         btrfs_tree_lock(right);
3783         btrfs_set_lock_blocking(right);
3784 
3785         free_space = btrfs_leaf_free_space(fs_info, right);
3786         if (free_space < data_size)
3787                 goto out_unlock;
3788 
3789         /* cow and double check */
3790         ret = btrfs_cow_block(trans, root, right, upper,
3791                               slot + 1, &right);
3792         if (ret)
3793                 goto out_unlock;
3794 
3795         free_space = btrfs_leaf_free_space(fs_info, right);
3796         if (free_space < data_size)
3797                 goto out_unlock;
3798 
3799         left_nritems = btrfs_header_nritems(left);
3800         if (left_nritems == 0)
3801                 goto out_unlock;
3802 
3803         if (path->slots[0] == left_nritems && !empty) {
3804                 /* Key greater than all keys in the leaf, right neighbor has
3805                  * enough room for it and we're not emptying our leaf to delete
3806                  * it, therefore use right neighbor to insert the new item and
3807                  * no need to touch/dirty our left leaft. */
3808                 btrfs_tree_unlock(left);
3809                 free_extent_buffer(left);
3810                 path->nodes[0] = right;
3811                 path->slots[0] = 0;
3812                 path->slots[1]++;
3813                 return 0;
3814         }
3815 
3816         return __push_leaf_right(fs_info, path, min_data_size, empty,
3817                                 right, free_space, left_nritems, min_slot);
3818 out_unlock:
3819         btrfs_tree_unlock(right);
3820         free_extent_buffer(right);
3821         return 1;
3822 }
3823 
3824 /*
3825  * push some data in the path leaf to the left, trying to free up at
3826  * least data_size bytes.  returns zero if the push worked, nonzero otherwise
3827  *
3828  * max_slot can put a limit on how far into the leaf we'll push items.  The
3829  * item at 'max_slot' won't be touched.  Use (u32)-1 to make us do all the
3830  * items
3831  */
3832 static noinline int __push_leaf_left(struct btrfs_fs_info *fs_info,
3833                                      struct btrfs_path *path, int data_size,
3834                                      int empty, struct extent_buffer *left,
3835                                      int free_space, u32 right_nritems,
3836                                      u32 max_slot)
3837 {
3838         struct btrfs_disk_key disk_key;
3839         struct extent_buffer *right = path->nodes[0];
3840         int i;
3841         int push_space = 0;
3842         int push_items = 0;
3843         struct btrfs_item *item;
3844         u32 old_left_nritems;
3845         u32 nr;
3846         int ret = 0;
3847         u32 this_item_size;
3848         u32 old_left_item_size;
3849         struct btrfs_map_token token;
3850 
3851         btrfs_init_map_token(&token);
3852 
3853         if (empty)
3854                 nr = min(right_nritems, max_slot);
3855         else
3856                 nr = min(right_nritems - 1, max_slot);
3857 
3858         for (i = 0; i < nr; i++) {
3859                 item = btrfs_item_nr(i);
3860 
3861                 if (!empty && push_items > 0) {
3862                         if (path->slots[0] < i)
3863                                 break;
3864                         if (path->slots[0] == i) {
3865                                 int space = btrfs_leaf_free_space(fs_info, right);
3866                                 if (space + push_space * 2 > free_space)
3867                                         break;
3868                         }
3869                 }
3870 
3871                 if (path->slots[0] == i)
3872                         push_space += data_size;
3873 
3874                 this_item_size = btrfs_item_size(right, item);
3875                 if (this_item_size + sizeof(*item) + push_space > free_space)
3876                         break;
3877 
3878                 push_items++;
3879                 push_space += this_item_size + sizeof(*item);
3880         }
3881 
3882         if (push_items == 0) {
3883                 ret = 1;
3884                 goto out;
3885         }
3886         WARN_ON(!empty && push_items == btrfs_header_nritems(right));
3887 
3888         /* push data from right to left */
3889         copy_extent_buffer(left, right,
3890                            btrfs_item_nr_offset(btrfs_header_nritems(left)),
3891                            btrfs_item_nr_offset(0),
3892                            push_items * sizeof(struct btrfs_item));
3893 
3894         push_space = BTRFS_LEAF_DATA_SIZE(fs_info) -
3895                      btrfs_item_offset_nr(right, push_items - 1);
3896 
3897         copy_extent_buffer(left, right, BTRFS_LEAF_DATA_OFFSET +
3898                      leaf_data_end(fs_info, left) - push_space,
3899                      BTRFS_LEAF_DATA_OFFSET +
3900                      btrfs_item_offset_nr(right, push_items - 1),
3901                      push_space);
3902         old_left_nritems = btrfs_header_nritems(left);
3903         BUG_ON(old_left_nritems <= 0);
3904 
3905         old_left_item_size = btrfs_item_offset_nr(left, old_left_nritems - 1);
3906         for (i = old_left_nritems; i < old_left_nritems + push_items; i++) {
3907                 u32 ioff;
3908 
3909                 item = btrfs_item_nr(i);
3910 
3911                 ioff = btrfs_token_item_offset(left, item, &token);
3912                 btrfs_set_token_item_offset(left, item,
3913                       ioff - (BTRFS_LEAF_DATA_SIZE(fs_info) - old_left_item_size),
3914                       &token);
3915         }
3916         btrfs_set_header_nritems(left, old_left_nritems + push_items);
3917 
3918         /* fixup right node */
3919         if (push_items > right_nritems)
3920                 WARN(1, KERN_CRIT "push items %d nr %u\n", push_items,
3921                        right_nritems);
3922 
3923         if (push_items < right_nritems) {
3924                 push_space = btrfs_item_offset_nr(right, push_items - 1) -
3925                                                   leaf_data_end(fs_info, right);
3926                 memmove_extent_buffer(right, BTRFS_LEAF_DATA_OFFSET +
3927                                       BTRFS_LEAF_DATA_SIZE(fs_info) - push_space,
3928                                       BTRFS_LEAF_DATA_OFFSET +
3929                                       leaf_data_end(fs_info, right), push_space);
3930 
3931                 memmove_extent_buffer(right, btrfs_item_nr_offset(0),
3932                               btrfs_item_nr_offset(push_items),
3933                              (btrfs_header_nritems(right) - push_items) *
3934                              sizeof(struct btrfs_item));
3935         }
3936         right_nritems -= push_items;
3937         btrfs_set_header_nritems(right, right_nritems);
3938         push_space = BTRFS_LEAF_DATA_SIZE(fs_info);
3939         for (i = 0; i < right_nritems; i++) {
3940                 item = btrfs_item_nr(i);
3941 
3942                 push_space = push_space - btrfs_token_item_size(right,
3943                                                                 item, &token);
3944                 btrfs_set_token_item_offset(right, item, push_space, &token);
3945         }
3946 
3947         btrfs_mark_buffer_dirty(left);
3948         if (right_nritems)
3949                 btrfs_mark_buffer_dirty(right);
3950         else
3951                 clean_tree_block(fs_info, right);
3952 
3953         btrfs_item_key(right, &disk_key, 0);
3954         fixup_low_keys(fs_info, path, &disk_key, 1);
3955 
3956         /* then fixup the leaf pointer in the path */
3957         if (path->slots[0] < push_items) {
3958                 path->slots[0] += old_left_nritems;
3959                 btrfs_tree_unlock(path->nodes[0]);
3960                 free_extent_buffer(path->nodes[0]);
3961                 path->nodes[0] = left;
3962                 path->slots[1] -= 1;
3963         } else {
3964                 btrfs_tree_unlock(left);
3965                 free_extent_buffer(left);
3966                 path->slots[0] -= push_items;
3967         }
3968         BUG_ON(path->slots[0] < 0);
3969         return ret;
3970 out:
3971         btrfs_tree_unlock(left);
3972         free_extent_buffer(left);
3973         return ret;
3974 }
3975 
3976 /*
3977  * push some data in the path leaf to the left, trying to free up at
3978  * least data_size bytes.  returns zero if the push worked, nonzero otherwise
3979  *
3980  * max_slot can put a limit on how far into the leaf we'll push items.  The
3981  * item at 'max_slot' won't be touched.  Use (u32)-1 to make us push all the
3982  * items
3983  */
3984 static int push_leaf_left(struct btrfs_trans_handle *trans, struct btrfs_root
3985                           *root, struct btrfs_path *path, int min_data_size,
3986                           int data_size, int empty, u32 max_slot)
3987 {
3988         struct btrfs_fs_info *fs_info = root->fs_info;
3989         struct extent_buffer *right = path->nodes[0];
3990         struct extent_buffer *left;
3991         int slot;
3992         int free_space;
3993         u32 right_nritems;
3994         int ret = 0;
3995 
3996         slot = path->slots[1];
3997         if (slot == 0)
3998                 return 1;
3999         if (!path->nodes[1])
4000                 return 1;
4001 
4002         right_nritems = btrfs_header_nritems(right);
4003         if (right_nritems == 0)
4004                 return 1;
4005 
4006         btrfs_assert_tree_locked(path->nodes[1]);
4007 
4008         left = read_node_slot(fs_info, path->nodes[1], slot - 1);
4009         /*
4010          * slot - 1 is not valid or we fail to read the left node,
4011          * no big deal, just return.
4012          */
4013         if (IS_ERR(left))
4014                 return 1;
4015 
4016         btrfs_tree_lock(left);
4017         btrfs_set_lock_blocking(left);
4018 
4019         free_space = btrfs_leaf_free_space(fs_info, left);
4020         if (free_space < data_size) {
4021                 ret = 1;
4022                 goto out;
4023         }
4024 
4025         /* cow and double check */
4026         ret = btrfs_cow_block(trans, root, left,
4027                               path->nodes[1], slot - 1, &left);
4028         if (ret) {
4029                 /* we hit -ENOSPC, but it isn't fatal here */
4030                 if (ret == -ENOSPC)
4031                         ret = 1;
4032                 goto out;
4033         }
4034 
4035         free_space = btrfs_leaf_free_space(fs_info, left);
4036         if (free_space < data_size) {
4037                 ret = 1;
4038                 goto out;
4039         }
4040 
4041         return __push_leaf_left(fs_info, path, min_data_size,
4042                                empty, left, free_space, right_nritems,
4043                                max_slot);
4044 out:
4045         btrfs_tree_unlock(left);
4046         free_extent_buffer(left);
4047         return ret;
4048 }
4049 
4050 /*
4051  * split the path's leaf in two, making sure there is at least data_size
4052  * available for the resulting leaf level of the path.
4053  */
4054 static noinline void copy_for_split(struct btrfs_trans_handle *trans,
4055                                     struct btrfs_fs_info *fs_info,
4056                                     struct btrfs_path *path,
4057                                     struct extent_buffer *l,
4058                                     struct extent_buffer *right,
4059                                     int slot, int mid, int nritems)
4060 {
4061         int data_copy_size;
4062         int rt_data_off;
4063         int i;
4064         struct btrfs_disk_key disk_key;
4065         struct btrfs_map_token token;
4066 
4067         btrfs_init_map_token(&token);
4068 
4069         nritems = nritems - mid;
4070         btrfs_set_header_nritems(right, nritems);
4071         data_copy_size = btrfs_item_end_nr(l, mid) - leaf_data_end(fs_info, l);
4072 
4073         copy_extent_buffer(right, l, btrfs_item_nr_offset(0),
4074                            btrfs_item_nr_offset(mid),
4075                            nritems * sizeof(struct btrfs_item));
4076 
4077         copy_extent_buffer(right, l,
4078                      BTRFS_LEAF_DATA_OFFSET + BTRFS_LEAF_DATA_SIZE(fs_info) -
4079                      data_copy_size, BTRFS_LEAF_DATA_OFFSET +
4080                      leaf_data_end(fs_info, l), data_copy_size);
4081 
4082         rt_data_off = BTRFS_LEAF_DATA_SIZE(fs_info) - btrfs_item_end_nr(l, mid);
4083 
4084         for (i = 0; i < nritems; i++) {
4085                 struct btrfs_item *item = btrfs_item_nr(i);
4086                 u32 ioff;
4087 
4088                 ioff = btrfs_token_item_offset(right, item, &token);
4089                 btrfs_set_token_item_offset(right, item,
4090                                             ioff + rt_data_off, &token);
4091         }
4092 
4093         btrfs_set_header_nritems(l, mid);
4094         btrfs_item_key(right, &disk_key, 0);
4095         insert_ptr(trans, fs_info, path, &disk_key, right->start,
4096                    path->slots[1] + 1, 1);
4097 
4098         btrfs_mark_buffer_dirty(right);
4099         btrfs_mark_buffer_dirty(l);
4100         BUG_ON(path->slots[0] != slot);
4101 
4102         if (mid <= slot) {
4103                 btrfs_tree_unlock(path->nodes[0]);
4104                 free_extent_buffer(path->nodes[0]);
4105                 path->nodes[0] = right;
4106                 path->slots[0] -= mid;
4107                 path->slots[1] += 1;
4108         } else {
4109                 btrfs_tree_unlock(right);
4110                 free_extent_buffer(right);
4111         }
4112 
4113         BUG_ON(path->slots[0] < 0);
4114 }
4115 
4116 /*
4117  * double splits happen when we need to insert a big item in the middle
4118  * of a leaf.  A double split can leave us with 3 mostly empty leaves:
4119  * leaf: [ slots 0 - N] [ our target ] [ N + 1 - total in leaf ]
4120  *          A                 B                 C
4121  *
4122  * We avoid this by trying to push the items on either side of our target
4123  * into the adjacent leaves.  If all goes well we can avoid the double split
4124  * completely.
4125  */
4126 static noinline int push_for_double_split(struct btrfs_trans_handle *trans,
4127                                           struct btrfs_root *root,
4128                                           struct btrfs_path *path,
4129                                           int data_size)
4130 {
4131         struct btrfs_fs_info *fs_info = root->fs_info;
4132         int ret;
4133         int progress = 0;
4134         int slot;
4135         u32 nritems;
4136         int space_needed = data_size;
4137 
4138         slot = path->slots[0];
4139         if (slot < btrfs_header_nritems(path->nodes[0]))
4140                 space_needed -= btrfs_leaf_free_space(fs_info, path->nodes[0]);
4141 
4142         /*
4143          * try to push all the items after our slot into the
4144          * right leaf
4145          */
4146         ret = push_leaf_right(trans, root, path, 1, space_needed, 0, slot);
4147         if (ret < 0)
4148                 return ret;
4149 
4150         if (ret == 0)
4151                 progress++;
4152 
4153         nritems = btrfs_header_nritems(path->nodes[0]);
4154         /*
4155          * our goal is to get our slot at the start or end of a leaf.  If
4156          * we've done so we're done
4157          */
4158         if (path->slots[0] == 0 || path->slots[0] == nritems)
4159                 return 0;
4160 
4161         if (btrfs_leaf_free_space(fs_info, path->nodes[0]) >= data_size)
4162                 return 0;
4163 
4164         /* try to push all the items before our slot into the next leaf */
4165         slot = path->slots[0];
4166         space_needed = data_size;
4167         if (slot > 0)
4168                 space_needed -= btrfs_leaf_free_space(fs_info, path->nodes[0]);
4169         ret = push_leaf_left(trans, root, path, 1, space_needed, 0, slot);
4170         if (ret < 0)
4171                 return ret;
4172 
4173         if (ret == 0)
4174                 progress++;
4175 
4176         if (progress)
4177                 return 0;
4178         return 1;
4179 }
4180 
4181 /*
4182  * split the path's leaf in two, making sure there is at least data_size
4183  * available for the resulting leaf level of the path.
4184  *
4185  * returns 0 if all went well and < 0 on failure.
4186  */
4187 static noinline int split_leaf(struct btrfs_trans_handle *trans,
4188                                struct btrfs_root *root,
4189                                const struct btrfs_key *ins_key,
4190                                struct btrfs_path *path, int data_size,
4191                                int extend)
4192 {
4193         struct btrfs_disk_key disk_key;
4194         struct extent_buffer *l;
4195         u32 nritems;
4196         int mid;
4197         int slot;
4198         struct extent_buffer *right;
4199         struct btrfs_fs_info *fs_info = root->fs_info;
4200         int ret = 0;
4201         int wret;
4202         int split;
4203         int num_doubles = 0;
4204         int tried_avoid_double = 0;
4205 
4206         l = path->nodes[0];
4207         slot = path->slots[0];
4208         if (extend && data_size + btrfs_item_size_nr(l, slot) +
4209             sizeof(struct btrfs_item) > BTRFS_LEAF_DATA_SIZE(fs_info))
4210                 return -EOVERFLOW;
4211 
4212         /* first try to make some room by pushing left and right */
4213         if (data_size && path->nodes[1]) {
4214                 int space_needed = data_size;
4215 
4216                 if (slot < btrfs_header_nritems(l))
4217                         space_needed -= btrfs_leaf_free_space(fs_info, l);
4218 
4219                 wret = push_leaf_right(trans, root, path, space_needed,
4220                                        space_needed, 0, 0);
4221                 if (wret < 0)
4222                         return wret;
4223                 if (wret) {
4224                         space_needed = data_size;
4225                         if (slot > 0)
4226                                 space_needed -= btrfs_leaf_free_space(fs_info,
4227                                                                       l);
4228                         wret = push_leaf_left(trans, root, path, space_needed,
4229                                               space_needed, 0, (u32)-1);
4230                         if (wret < 0)
4231                                 return wret;
4232                 }
4233                 l = path->nodes[0];
4234 
4235                 /* did the pushes work? */
4236                 if (btrfs_leaf_free_space(fs_info, l) >= data_size)
4237                         return 0;
4238         }
4239 
4240         if (!path->nodes[1]) {
4241                 ret = insert_new_root(trans, root, path, 1);
4242                 if (ret)
4243                         return ret;
4244         }
4245 again:
4246         split = 1;
4247         l = path->nodes[0];
4248         slot = path->slots[0];
4249         nritems = btrfs_header_nritems(l);
4250         mid = (nritems + 1) / 2;
4251 
4252         if (mid <= slot) {
4253                 if (nritems == 1 ||
4254                     leaf_space_used(l, mid, nritems - mid) + data_size >
4255                         BTRFS_LEAF_DATA_SIZE(fs_info)) {
4256                         if (slot >= nritems) {
4257                                 split = 0;
4258                         } else {
4259                                 mid = slot;
4260                                 if (mid != nritems &&
4261                                     leaf_space_used(l, mid, nritems - mid) +
4262                                     data_size > BTRFS_LEAF_DATA_SIZE(fs_info)) {
4263                                         if (data_size && !tried_avoid_double)
4264                                                 goto push_for_double;
4265                                         split = 2;
4266                                 }
4267                         }
4268                 }
4269         } else {
4270                 if (leaf_space_used(l, 0, mid) + data_size >
4271                         BTRFS_LEAF_DATA_SIZE(fs_info)) {
4272                         if (!extend && data_size && slot == 0) {
4273                                 split = 0;
4274                         } else if ((extend || !data_size) && slot == 0) {
4275                                 mid = 1;
4276                         } else {
4277                                 mid = slot;
4278                                 if (mid != nritems &&
4279                                     leaf_space_used(l, mid, nritems - mid) +
4280                                     data_size > BTRFS_LEAF_DATA_SIZE(fs_info)) {
4281                                         if (data_size && !tried_avoid_double)
4282                                                 goto push_for_double;
4283                                         split = 2;
4284                                 }
4285                         }
4286                 }
4287         }
4288 
4289         if (split == 0)
4290                 btrfs_cpu_key_to_disk(&disk_key, ins_key);
4291         else
4292                 btrfs_item_key(l, &disk_key, mid);
4293 
4294         right = btrfs_alloc_tree_block(trans, root, 0, root->root_key.objectid,
4295                         &disk_key, 0, l->start, 0);
4296         if (IS_ERR(right))
4297                 return PTR_ERR(right);
4298 
4299         root_add_used(root, fs_info->nodesize);
4300 
4301         memzero_extent_buffer(right, 0, sizeof(struct btrfs_header));
4302         btrfs_set_header_bytenr(right, right->start);
4303         btrfs_set_header_generation(right, trans->transid);
4304         btrfs_set_header_backref_rev(right, BTRFS_MIXED_BACKREF_REV);
4305         btrfs_set_header_owner(right, root->root_key.objectid);
4306         btrfs_set_header_level(right, 0);
4307         write_extent_buffer_fsid(right, fs_info->fsid);
4308         write_extent_buffer_chunk_tree_uuid(right, fs_info->chunk_tree_uuid);
4309 
4310         if (split == 0) {
4311                 if (mid <= slot) {
4312                         btrfs_set_header_nritems(right, 0);
4313                         insert_ptr(trans, fs_info, path, &disk_key,
4314                                    right->start, path->slots[1] + 1, 1);
4315                         btrfs_tree_unlock(path->nodes[0]);
4316                         free_extent_buffer(path->nodes[0]);
4317                         path->nodes[0] = right;
4318                         path->slots[0] = 0;
4319                         path->slots[1] += 1;
4320                 } else {
4321                         btrfs_set_header_nritems(right, 0);
4322                         insert_ptr(trans, fs_info, path, &disk_key,
4323                                    right->start, path->slots[1], 1);
4324                         btrfs_tree_unlock(path->nodes[0]);
4325                         free_extent_buffer(path->nodes[0]);
4326                         path->nodes[0] = right;
4327                         path->slots[0] = 0;
4328                         if (path->slots[1] == 0)
4329                                 fixup_low_keys(fs_info, path, &disk_key, 1);
4330                 }
4331                 /*
4332                  * We create a new leaf 'right' for the required ins_len and
4333                  * we'll do btrfs_mark_buffer_dirty() on this leaf after copying
4334                  * the content of ins_len to 'right'.
4335                  */
4336                 return ret;
4337         }
4338 
4339         copy_for_split(trans, fs_info, path, l, right, slot, mid, nritems);
4340 
4341         if (split == 2) {
4342                 BUG_ON(num_doubles != 0);
4343                 num_doubles++;
4344                 goto again;
4345         }
4346 
4347         return 0;
4348 
4349 push_for_double:
4350         push_for_double_split(trans, root, path, data_size);
4351         tried_avoid_double = 1;
4352         if (btrfs_leaf_free_space(fs_info, path->nodes[0]) >= data_size)
4353                 return 0;
4354         goto again;
4355 }
4356 
4357 static noinline int setup_leaf_for_split(struct btrfs_trans_handle *trans,
4358                                          struct btrfs_root *root,
4359                                          struct btrfs_path *path, int ins_len)
4360 {
4361         struct btrfs_fs_info *fs_info = root->fs_info;
4362         struct btrfs_key key;
4363         struct extent_buffer *leaf;
4364         struct btrfs_file_extent_item *fi;
4365         u64 extent_len = 0;
4366         u32 item_size;
4367         int ret;
4368 
4369         leaf = path->nodes[0];
4370         btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
4371 
4372         BUG_ON(key.type != BTRFS_EXTENT_DATA_KEY &&
4373                key.type != BTRFS_EXTENT_CSUM_KEY);
4374 
4375         if (btrfs_leaf_free_space(fs_info, leaf) >= ins_len)
4376                 return 0;
4377 
4378         item_size = btrfs_item_size_nr(leaf, path->slots[0]);
4379         if (key.type == BTRFS_EXTENT_DATA_KEY) {
4380                 fi = btrfs_item_ptr(leaf, path->slots[0],
4381                                     struct btrfs_file_extent_item);
4382                 extent_len = btrfs_file_extent_num_bytes(leaf, fi);
4383         }
4384         btrfs_release_path(path);
4385 
4386         path->keep_locks = 1;
4387         path->search_for_split = 1;
4388         ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
4389         path->search_for_split = 0;
4390         if (ret > 0)
4391                 ret = -EAGAIN;
4392         if (ret < 0)
4393                 goto err;
4394 
4395         ret = -EAGAIN;
4396         leaf = path->nodes[0];
4397         /* if our item isn't there, return now */
4398         if (item_size != btrfs_item_size_nr(leaf, path->slots[0]))
4399                 goto err;
4400 
4401         /* the leaf has  changed, it now has room.  return now */
4402         if (btrfs_leaf_free_space(fs_info, path->nodes[0]) >= ins_len)
4403                 goto err;
4404 
4405         if (key.type == BTRFS_EXTENT_DATA_KEY) {
4406                 fi = btrfs_item_ptr(leaf, path->slots[0],
4407                                     struct btrfs_file_extent_item);
4408                 if (extent_len != btrfs_file_extent_num_bytes(leaf, fi))
4409                         goto err;
4410         }
4411 
4412         btrfs_set_path_blocking(path);
4413         ret = split_leaf(trans, root, &key, path, ins_len, 1);
4414         if (ret)
4415                 goto err;
4416 
4417         path->keep_locks = 0;
4418         btrfs_unlock_up_safe(path, 1);
4419         return 0;
4420 err:
4421         path->keep_locks = 0;
4422         return ret;
4423 }
4424 
4425 static noinline int split_item(struct btrfs_fs_info *fs_info,
4426                                struct btrfs_path *path,
4427                                const struct btrfs_key *new_key,
4428                                unsigned long split_offset)
4429 {
4430         struct extent_buffer *leaf;
4431         struct btrfs_item *item;
4432         struct btrfs_item *new_item;
4433         int slot;
4434         char *buf;
4435         u32 nritems;
4436         u32 item_size;
4437         u32 orig_offset;
4438         struct btrfs_disk_key disk_key;
4439 
4440         leaf = path->nodes[0];
4441         BUG_ON(btrfs_leaf_free_space(fs_info, leaf) < sizeof(struct btrfs_item));
4442 
4443         btrfs_set_path_blocking(path);
4444 
4445         item = btrfs_item_nr(path->slots[0]);
4446         orig_offset = btrfs_item_offset(leaf, item);
4447         item_size = btrfs_item_size(leaf, item);
4448 
4449         buf = kmalloc(item_size, GFP_NOFS);
4450         if (!buf)
4451                 return -ENOMEM;
4452 
4453         read_extent_buffer(leaf, buf, btrfs_item_ptr_offset(leaf,
4454                             path->slots[0]), item_size);
4455 
4456         slot = path->slots[0] + 1;
4457         nritems = btrfs_header_nritems(leaf);
4458         if (slot != nritems) {
4459                 /* shift the items */
4460                 memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + 1),
4461                                 btrfs_item_nr_offset(slot),
4462                                 (nritems - slot) * sizeof(struct btrfs_item));
4463         }
4464 
4465         btrfs_cpu_key_to_disk(&disk_key, new_key);
4466         btrfs_set_item_key(leaf, &disk_key, slot);
4467 
4468         new_item = btrfs_item_nr(slot);
4469 
4470         btrfs_set_item_offset(leaf, new_item, orig_offset);
4471         btrfs_set_item_size(leaf, new_item, item_size - split_offset);
4472 
4473         btrfs_set_item_offset(leaf, item,
4474                               orig_offset + item_size - split_offset);
4475         btrfs_set_item_size(leaf, item, split_offset);
4476 
4477         btrfs_set_header_nritems(leaf, nritems + 1);
4478 
4479         /* write the data for the start of the original item */
4480         write_extent_buffer(leaf, buf,
4481                             btrfs_item_ptr_offset(leaf, path->slots[0]),
4482                             split_offset);
4483 
4484         /* write the data for the new item */
4485         write_extent_buffer(leaf, buf + split_offset,
4486                             btrfs_item_ptr_offset(leaf, slot),
4487                             item_size - split_offset);
4488         btrfs_mark_buffer_dirty(leaf);
4489 
4490         BUG_ON(btrfs_leaf_free_space(fs_info, leaf) < 0);
4491         kfree(buf);
4492         return 0;
4493 }
4494 
4495 /*
4496  * This function splits a single item into two items,
4497  * giving 'new_key' to the new item and splitting the
4498  * old one at split_offset (from the start of the item).
4499  *
4500  * The path may be released by this operation.  After
4501  * the split, the path is pointing to the old item.  The
4502  * new item is going to be in the same node as the old one.
4503  *
4504  * Note, the item being split must be smaller enough to live alone on
4505  * a tree block with room for one extra struct btrfs_item
4506  *
4507  * This allows us to split the item in place, keeping a lock on the
4508  * leaf the entire time.
4509  */
4510 int btrfs_split_item(struct btrfs_trans_handle *trans,
4511                      struct btrfs_root *root,
4512                      struct btrfs_path *path,
4513                      const struct btrfs_key *new_key,
4514                      unsigned long split_offset)
4515 {
4516         int ret;
4517         ret = setup_leaf_for_split(trans, root, path,
4518                                    sizeof(struct btrfs_item));
4519         if (ret)
4520                 return ret;
4521 
4522         ret = split_item(root->fs_info, path, new_key, split_offset);
4523         return ret;
4524 }
4525 
4526 /*
4527  * This function duplicate a item, giving 'new_key' to the new item.
4528  * It guarantees both items live in the same tree leaf and the new item
4529  * is contiguous with the original item.
4530  *
4531  * This allows us to split file extent in place, keeping a lock on the
4532  * leaf the entire time.
4533  */
4534 int btrfs_duplicate_item(struct btrfs_trans_handle *trans,
4535                          struct btrfs_root *root,
4536                          struct btrfs_path *path,
4537                          const struct btrfs_key *new_key)
4538 {
4539         struct extent_buffer *leaf;
4540         int ret;
4541         u32 item_size;
4542 
4543         leaf = path->nodes[0];
4544         item_size = btrfs_item_size_nr(leaf, path->slots[0]);
4545         ret = setup_leaf_for_split(trans, root, path,
4546                                    item_size + sizeof(struct btrfs_item));
4547         if (ret)
4548                 return ret;
4549 
4550         path->slots[0]++;
4551         setup_items_for_insert(root, path, new_key, &item_size,
4552                                item_size, item_size +
4553                                sizeof(struct btrfs_item), 1);
4554         leaf = path->nodes[0];
4555         memcpy_extent_buffer(leaf,
4556                              btrfs_item_ptr_offset(leaf, path->slots[0]),
4557                              btrfs_item_ptr_offset(leaf, path->slots[0] - 1),
4558                              item_size);
4559         return 0;
4560 }
4561 
4562 /*
4563  * make the item pointed to by the path smaller.  new_size indicates
4564  * how small to make it, and from_end tells us if we just chop bytes
4565  * off the end of the item or if we shift the item to chop bytes off
4566  * the front.
4567  */
4568 void btrfs_truncate_item(struct btrfs_fs_info *fs_info,
4569                          struct btrfs_path *path, u32 new_size, int from_end)
4570 {
4571         int slot;
4572         struct extent_buffer *leaf;
4573         struct btrfs_item *item;
4574         u32 nritems;
4575         unsigned int data_end;
4576         unsigned int old_data_start;
4577         unsigned int old_size;
4578         unsigned int size_diff;
4579         int i;
4580         struct btrfs_map_token token;
4581 
4582         btrfs_init_map_token(&token);
4583 
4584         leaf = path->nodes[0];
4585         slot = path->slots[0];
4586 
4587         old_size = btrfs_item_size_nr(leaf, slot);
4588         if (old_size == new_size)
4589                 return;
4590 
4591         nritems = btrfs_header_nritems(leaf);
4592         data_end = leaf_data_end(fs_info, leaf);
4593 
4594         old_data_start = btrfs_item_offset_nr(leaf, slot);
4595 
4596         size_diff = old_size - new_size;
4597 
4598         BUG_ON(slot < 0);
4599         BUG_ON(slot >= nritems);
4600 
4601         /*
4602          * item0..itemN ... dataN.offset..dataN.size .. data0.size
4603          */
4604         /* first correct the data pointers */
4605         for (i = slot; i < nritems; i++) {
4606                 u32 ioff;
4607                 item = btrfs_item_nr(i);
4608 
4609                 ioff = btrfs_token_item_offset(leaf, item, &token);
4610                 btrfs_set_token_item_offset(leaf, item,
4611                                             ioff + size_diff, &token);
4612         }
4613 
4614         /* shift the data */
4615         if (from_end) {
4616                 memmove_extent_buffer(leaf, BTRFS_LEAF_DATA_OFFSET +
4617                               data_end + size_diff, BTRFS_LEAF_DATA_OFFSET +
4618                               data_end, old_data_start + new_size - data_end);
4619         } else {
4620                 struct btrfs_disk_key disk_key;
4621                 u64 offset;
4622 
4623                 btrfs_item_key(leaf, &disk_key, slot);
4624 
4625                 if (btrfs_disk_key_type(&disk_key) == BTRFS_EXTENT_DATA_KEY) {
4626                         unsigned long ptr;
4627                         struct btrfs_file_extent_item *fi;
4628 
4629                         fi = btrfs_item_ptr(leaf, slot,
4630                                             struct btrfs_file_extent_item);
4631                         fi = (struct btrfs_file_extent_item *)(
4632                              (unsigned long)fi - size_diff);
4633 
4634                         if (btrfs_file_extent_type(leaf, fi) ==
4635                             BTRFS_FILE_EXTENT_INLINE) {
4636                                 ptr = btrfs_item_ptr_offset(leaf, slot);
4637                                 memmove_extent_buffer(leaf, ptr,
4638                                       (unsigned long)fi,
4639                                       BTRFS_FILE_EXTENT_INLINE_DATA_START);
4640                         }
4641                 }
4642 
4643                 memmove_extent_buffer(leaf, BTRFS_LEAF_DATA_OFFSET +
4644                               data_end + size_diff, BTRFS_LEAF_DATA_OFFSET +
4645                               data_end, old_data_start - data_end);
4646 
4647                 offset = btrfs_disk_key_offset(&disk_key);
4648                 btrfs_set_disk_key_offset(&disk_key, offset + size_diff);
4649                 btrfs_set_item_key(leaf, &disk_key, slot);
4650                 if (slot == 0)
4651                         fixup_low_keys(fs_info, path, &disk_key, 1);
4652         }
4653 
4654         item = btrfs_item_nr(slot);
4655         btrfs_set_item_size(leaf, item, new_size);
4656         btrfs_mark_buffer_dirty(leaf);
4657 
4658         if (btrfs_leaf_free_space(fs_info, leaf) < 0) {
4659                 btrfs_print_leaf(leaf);
4660                 BUG();
4661         }
4662 }
4663 
4664 /*
4665  * make the item pointed to by the path bigger, data_size is the added size.
4666  */
4667 void btrfs_extend_item(struct btrfs_fs_info *fs_info, struct btrfs_path *path,
4668                        u32 data_size)
4669 {
4670         int slot;
4671         struct extent_buffer *leaf;
4672         struct btrfs_item *item;
4673         u32 nritems;
4674         unsigned int data_end;
4675         unsigned int old_data;
4676         unsigned int old_size;
4677         int i;
4678         struct btrfs_map_token token;
4679 
4680         btrfs_init_map_token(&token);
4681 
4682         leaf = path->nodes[0];
4683 
4684         nritems = btrfs_header_nritems(leaf);
4685         data_end = leaf_data_end(fs_info, leaf);
4686 
4687         if (btrfs_leaf_free_space(fs_info, leaf) < data_size) {
4688                 btrfs_print_leaf(leaf);
4689                 BUG();
4690         }
4691         slot = path->slots[0];
4692         old_data = btrfs_item_end_nr(leaf, slot);
4693 
4694         BUG_ON(slot < 0);
4695         if (slot >= nritems) {
4696                 btrfs_print_leaf(leaf);
4697                 btrfs_crit(fs_info, "slot %d too large, nritems %d",
4698                            slot, nritems);
4699                 BUG_ON(1);
4700         }
4701 
4702         /*
4703          * item0..itemN ... dataN.offset..dataN.size .. data0.size
4704          */
4705         /* first correct the data pointers */
4706         for (i = slot; i < nritems; i++) {
4707                 u32 ioff;
4708                 item = btrfs_item_nr(i);
4709 
4710                 ioff = btrfs_token_item_offset(leaf, item, &token);
4711                 btrfs_set_token_item_offset(leaf, item,
4712                                             ioff - data_size, &token);
4713         }
4714 
4715         /* shift the data */
4716         memmove_extent_buffer(leaf, BTRFS_LEAF_DATA_OFFSET +
4717                       data_end - data_size, BTRFS_LEAF_DATA_OFFSET +
4718                       data_end, old_data - data_end);
4719 
4720         data_end = old_data;
4721         old_size = btrfs_item_size_nr(leaf, slot);
4722         item = btrfs_item_nr(slot);
4723         btrfs_set_item_size(leaf, item, old_size + data_size);
4724         btrfs_mark_buffer_dirty(leaf);
4725 
4726         if (btrfs_leaf_free_space(fs_info, leaf) < 0) {
4727                 btrfs_print_leaf(leaf);
4728                 BUG();
4729         }
4730 }
4731 
4732 /*
4733  * this is a helper for btrfs_insert_empty_items, the main goal here is
4734  * to save stack depth by doing the bulk of the work in a function
4735  * that doesn't call btrfs_search_slot
4736  */
4737 void setup_items_for_insert(struct btrfs_root *root, struct btrfs_path *path,
4738                             const struct btrfs_key *cpu_key, u32 *data_size,
4739                             u32 total_data, u32 total_size, int nr)
4740 {
4741         struct btrfs_fs_info *fs_info = root->fs_info;
4742         struct btrfs_item *item;
4743         int i;
4744         u32 nritems;
4745         unsigned int data_end;
4746         struct btrfs_disk_key disk_key;
4747         struct extent_buffer *leaf;
4748         int slot;
4749         struct btrfs_map_token token;
4750 
4751         if (path->slots[0] == 0) {
4752                 btrfs_cpu_key_to_disk(&disk_key, cpu_key);
4753                 fixup_low_keys(fs_info, path, &disk_key, 1);
4754         }
4755         btrfs_unlock_up_safe(path, 1);
4756 
4757         btrfs_init_map_token(&token);
4758 
4759         leaf = path->nodes[0];
4760         slot = path->slots[0];
4761 
4762         nritems = btrfs_header_nritems(leaf);
4763         data_end = leaf_data_end(fs_info, leaf);
4764 
4765         if (btrfs_leaf_free_space(fs_info, leaf) < total_size) {
4766                 btrfs_print_leaf(leaf);
4767                 btrfs_crit(fs_info, "not enough freespace need %u have %d",
4768                            total_size, btrfs_leaf_free_space(fs_info, leaf));
4769                 BUG();
4770         }
4771 
4772         if (slot != nritems) {
4773                 unsigned int old_data = btrfs_item_end_nr(leaf, slot);
4774 
4775                 if (old_data < data_end) {
4776                         btrfs_print_leaf(leaf);
4777                         btrfs_crit(fs_info, "slot %d old_data %d data_end %d",
4778                                    slot, old_data, data_end);
4779                         BUG_ON(1);
4780                 }
4781                 /*
4782                  * item0..itemN ... dataN.offset..dataN.size .. data0.size
4783                  */
4784                 /* first correct the data pointers */
4785                 for (i = slot; i < nritems; i++) {
4786                         u32 ioff;
4787 
4788                         item = btrfs_item_nr(i);
4789                         ioff = btrfs_token_item_offset(leaf, item, &token);
4790                         btrfs_set_token_item_offset(leaf, item,
4791                                                     ioff - total_data, &token);
4792                 }
4793                 /* shift the items */
4794                 memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr),
4795                               btrfs_item_nr_offset(slot),
4796                               (nritems - slot) * sizeof(struct btrfs_item));
4797 
4798                 /* shift the data */
4799                 memmove_extent_buffer(leaf, BTRFS_LEAF_DATA_OFFSET +
4800                               data_end - total_data, BTRFS_LEAF_DATA_OFFSET +
4801                               data_end, old_data - data_end);
4802                 data_end = old_data;
4803         }
4804 
4805         /* setup the item for the new data */
4806         for (i = 0; i < nr; i++) {
4807                 btrfs_cpu_key_to_disk(&disk_key, cpu_key + i);
4808                 btrfs_set_item_key(leaf, &disk_key, slot + i);
4809                 item = btrfs_item_nr(slot + i);
4810                 btrfs_set_token_item_offset(leaf, item,
4811                                             data_end - data_size[i], &token);
4812                 data_end -= data_size[i];
4813                 btrfs_set_token_item_size(leaf, item, data_size[i], &token);
4814         }
4815 
4816         btrfs_set_header_nritems(leaf, nritems + nr);
4817         btrfs_mark_buffer_dirty(leaf);
4818 
4819         if (btrfs_leaf_free_space(fs_info, leaf) < 0) {
4820                 btrfs_print_leaf(leaf);
4821                 BUG();
4822         }
4823 }
4824 
4825 /*
4826  * Given a key and some data, insert items into the tree.
4827  * This does all the path init required, making room in the tree if needed.
4828  */
4829 int btrfs_insert_empty_items(struct btrfs_trans_handle *trans,
4830                             struct btrfs_root *root,
4831                             struct btrfs_path *path,
4832                             const struct btrfs_key *cpu_key, u32 *data_size,
4833                             int nr)
4834 {
4835         int ret = 0;
4836         int slot;
4837         int i;
4838         u32 total_size = 0;
4839         u32 total_data = 0;
4840 
4841         for (i = 0; i < nr; i++)
4842                 total_data += data_size[i];
4843 
4844         total_size = total_data + (nr * sizeof(struct btrfs_item));
4845         ret = btrfs_search_slot(trans, root, cpu_key, path, total_size, 1);
4846         if (ret == 0)
4847                 return -EEXIST;
4848         if (ret < 0)
4849                 return ret;
4850 
4851         slot = path->slots[0];
4852         BUG_ON(slot < 0);
4853 
4854         setup_items_for_insert(root, path, cpu_key, data_size,
4855                                total_data, total_size, nr);
4856         return 0;
4857 }
4858 
4859 /*
4860  * Given a key and some data, insert an item into the tree.
4861  * This does all the path init required, making room in the tree if needed.
4862  */
4863 int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root *root,
4864                       const struct btrfs_key *cpu_key, void *data,
4865                       u32 data_size)
4866 {
4867         int ret = 0;
4868         struct btrfs_path *path;
4869         struct extent_buffer *leaf;
4870         unsigned long ptr;
4871 
4872         path = btrfs_alloc_path();
4873         if (!path)
4874                 return -ENOMEM;
4875         ret = btrfs_insert_empty_item(trans, root, path, cpu_key, data_size);
4876         if (!ret) {
4877                 leaf = path->nodes[0];
4878                 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
4879                 write_extent_buffer(leaf, data, ptr, data_size);
4880                 btrfs_mark_buffer_dirty(leaf);
4881         }
4882         btrfs_free_path(path);
4883         return ret;
4884 }
4885 
4886 /*
4887  * delete the pointer from a given node.
4888  *
4889  * the tree should have been previously balanced so the deletion does not
4890  * empty a node.
4891  */
4892 static void del_ptr(struct btrfs_root *root, struct btrfs_path *path,
4893                     int level, int slot)
4894 {
4895         struct btrfs_fs_info *fs_info = root->fs_info;
4896         struct extent_buffer *parent = path->nodes[level];
4897         u32 nritems;
4898         int ret;
4899 
4900         nritems = btrfs_header_nritems(parent);
4901         if (slot != nritems - 1) {
4902                 if (level)
4903                         tree_mod_log_eb_move(fs_info, parent, slot,
4904                                              slot + 1, nritems - slot - 1);
4905                 memmove_extent_buffer(parent,
4906                               btrfs_node_key_ptr_offset(slot),
4907                               btrfs_node_key_ptr_offset(slot + 1),
4908                               sizeof(struct btrfs_key_ptr) *
4909                               (nritems - slot - 1));
4910         } else if (level) {
4911                 ret = tree_mod_log_insert_key(fs_info, parent, slot,
4912                                               MOD_LOG_KEY_REMOVE, GFP_NOFS);
4913                 BUG_ON(ret < 0);
4914         }
4915 
4916         nritems--;
4917         btrfs_set_header_nritems(parent, nritems);
4918         if (nritems == 0 && parent == root->node) {
4919                 BUG_ON(btrfs_header_level(root->node) != 1);
4920                 /* just turn the root into a leaf and break */
4921                 btrfs_set_header_level(root->node, 0);
4922         } else if (slot == 0) {
4923                 struct btrfs_disk_key disk_key;
4924 
4925                 btrfs_node_key(parent, &disk_key, 0);
4926                 fixup_low_keys(fs_info, path, &disk_key, level + 1);
4927         }
4928         btrfs_mark_buffer_dirty(parent);
4929 }
4930 
4931 /*
4932  * a helper function to delete the leaf pointed to by path->slots[1] and
4933  * path->nodes[1].
4934  *
4935  * This deletes the pointer in path->nodes[1] and frees the leaf
4936  * block extent.  zero is returned if it all worked out, < 0 otherwise.
4937  *
4938  * The path must have already been setup for deleting the leaf, including
4939  * all the proper balancing.  path->nodes[1] must be locked.
4940  */
4941 static noinline void btrfs_del_leaf(struct btrfs_trans_handle *trans,
4942                                     struct btrfs_root *root,
4943                                     struct btrfs_path *path,
4944                                     struct extent_buffer *leaf)
4945 {
4946         WARN_ON(btrfs_header_generation(leaf) != trans->transid);
4947         del_ptr(root, path, 1, path->slots[1]);
4948 
4949         /*
4950          * btrfs_free_extent is expensive, we want to make sure we
4951          * aren't holding any locks when we call it
4952          */
4953         btrfs_unlock_up_safe(path, 0);
4954 
4955         root_sub_used(root, leaf->len);
4956 
4957         extent_buffer_get(leaf);
4958         btrfs_free_tree_block(trans, root, leaf, 0, 1);
4959         free_extent_buffer_stale(leaf);
4960 }
4961 /*
4962  * delete the item at the leaf level in path.  If that empties
4963  * the leaf, remove it from the tree
4964  */
4965 int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root,
4966                     struct btrfs_path *path, int slot, int nr)
4967 {
4968         struct btrfs_fs_info *fs_info = root->fs_info;
4969         struct extent_buffer *leaf;
4970         struct btrfs_item *item;
4971         u32 last_off;
4972         u32 dsize = 0;
4973         int ret = 0;
4974         int wret;
4975         int i;
4976         u32 nritems;
4977         struct btrfs_map_token token;
4978 
4979         btrfs_init_map_token(&token);
4980 
4981         leaf = path->nodes[0];
4982         last_off = btrfs_item_offset_nr(leaf, slot + nr - 1);
4983 
4984         for (i = 0; i < nr; i++)
4985                 dsize += btrfs_item_size_nr(leaf, slot + i);
4986 
4987         nritems = btrfs_header_nritems(leaf);
4988 
4989         if (slot + nr != nritems) {
4990                 int data_end = leaf_data_end(fs_info, leaf);
4991 
4992                 memmove_extent_buffer(leaf, BTRFS_LEAF_DATA_OFFSET +
4993                               data_end + dsize,
4994                               BTRFS_LEAF_DATA_OFFSET + data_end,
4995                               last_off - data_end);
4996 
4997                 for (i = slot + nr; i < nritems; i++) {
4998                         u32 ioff;
4999 
5000                         item = btrfs_item_nr(i);
5001                         ioff = btrfs_token_item_offset(leaf, item, &token);
5002                         btrfs_set_token_item_offset(leaf, item,
5003                                                     ioff + dsize, &token);
5004                 }
5005 
5006                 memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot),
5007                               btrfs_item_nr_offset(slot + nr),
5008                               sizeof(struct btrfs_item) *
5009                               (nritems - slot - nr));
5010         }
5011         btrfs_set_header_nritems(leaf, nritems - nr);
5012         nritems -= nr;
5013 
5014         /* delete the leaf if we've emptied it */
5015         if (nritems == 0) {
5016                 if (leaf == root->node) {
5017                         btrfs_set_header_level(leaf, 0);
5018                 } else {
5019                         btrfs_set_path_blocking(path);
5020                         clean_tree_block(fs_info, leaf);
5021                         btrfs_del_leaf(trans, root, path, leaf);
5022                 }
5023         } else {
5024                 int used = leaf_space_used(leaf, 0, nritems);
5025                 if (slot == 0) {
5026                         struct btrfs_disk_key disk_key;
5027 
5028                         btrfs_item_key(leaf, &disk_key, 0);
5029                         fixup_low_keys(fs_info, path, &disk_key, 1);
5030                 }
5031 
5032                 /* delete the leaf if it is mostly empty */
5033                 if (used < BTRFS_LEAF_DATA_SIZE(fs_info) / 3) {
5034                         /* push_leaf_left fixes the path.
5035                          * make sure the path still points to our leaf
5036                          * for possible call to del_ptr below
5037                          */
5038                         slot = path->slots[1];
5039                         extent_buffer_get(leaf);
5040 
5041                         btrfs_set_path_blocking(path);
5042                         wret = push_leaf_left(trans, root, path, 1, 1,
5043                                               1, (u32)-1);
5044                         if (wret < 0 && wret != -ENOSPC)
5045                                 ret = wret;
5046 
5047                         if (path->nodes[0] == leaf &&
5048                             btrfs_header_nritems(leaf)) {
5049                                 wret = push_leaf_right(trans, root, path, 1,
5050                                                        1, 1, 0);
5051                                 if (wret < 0 && wret != -ENOSPC)
5052                                         ret = wret;
5053                         }
5054 
5055                         if (btrfs_header_nritems(leaf) == 0) {
5056                                 path->slots[1] = slot;
5057                                 btrfs_del_leaf(trans, root, path, leaf);
5058                                 free_extent_buffer(leaf);
5059                                 ret = 0;
5060                         } else {
5061                                 /* if we're still in the path, make sure
5062                                  * we're dirty.  Otherwise, one of the
5063                                  * push_leaf functions must have already
5064                                  * dirtied this buffer
5065                                  */
5066                                 if (path->nodes[0] == leaf)
5067                                         btrfs_mark_buffer_dirty(leaf);
5068                                 free_extent_buffer(leaf);
5069                         }
5070                 } else {
5071                         btrfs_mark_buffer_dirty(leaf);
5072                 }
5073         }
5074         return ret;
5075 }
5076 
5077 /*
5078  * search the tree again to find a leaf with lesser keys
5079  * returns 0 if it found something or 1 if there are no lesser leaves.
5080  * returns < 0 on io errors.
5081  *
5082  * This may release the path, and so you may lose any locks held at the
5083  * time you call it.
5084  */
5085 int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path)
5086 {
5087         struct btrfs_key key;
5088         struct btrfs_disk_key found_key;
5089         int ret;
5090 
5091         btrfs_item_key_to_cpu(path->nodes[0], &key, 0);
5092 
5093         if (key.offset > 0) {
5094                 key.offset--;
5095         } else if (key.type > 0) {
5096                 key.type--;
5097                 key.offset = (u64)-1;
5098         } else if (key.objectid > 0) {
5099                 key.objectid--;
5100                 key.type = (u8)-1;
5101                 key.offset = (u64)-1;
5102         } else {
5103                 return 1;
5104         }
5105 
5106         btrfs_release_path(path);
5107         ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
5108         if (ret < 0)
5109                 return ret;
5110         btrfs_item_key(path->nodes[0], &found_key, 0);
5111         ret = comp_keys(&found_key, &key);
5112         /*
5113          * We might have had an item with the previous key in the tree right
5114          * before we released our path. And after we released our path, that
5115          * item might have been pushed to the first slot (0) of the leaf we
5116          * were holding due to a tree balance. Alternatively, an item with the
5117          * previous key can exist as the only element of a leaf (big fat item).
5118          * Therefore account for these 2 cases, so that our callers (like
5119          * btrfs_previous_item) don't miss an existing item with a key matching
5120          * the previous key we computed above.
5121          */
5122         if (ret <= 0)
5123                 return 0;
5124         return 1;
5125 }
5126 
5127 /*
5128  * A helper function to walk down the tree starting at min_key, and looking
5129  * for nodes or leaves that are have a minimum transaction id.
5130  * This is used by the btree defrag code, and tree logging
5131  *
5132  * This does not cow, but it does stuff the starting key it finds back
5133  * into min_key, so you can call btrfs_search_slot with cow=1 on the
5134  * key and get a writable path.
5135  *
5136  * This does lock as it descends, and path->keep_locks should be set
5137  * to 1 by the caller.
5138  *
5139  * This honors path->lowest_level to prevent descent past a given level
5140  * of the tree.
5141  *
5142  * min_trans indicates the oldest transaction that you are interested
5143  * in walking through.  Any nodes or leaves older than min_trans are
5144  * skipped over (without reading them).
5145  *
5146  * returns zero if something useful was found, < 0 on error and 1 if there
5147  * was nothing in the tree that matched the search criteria.
5148  */
5149 int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key,
5150                          struct btrfs_path *path,
5151                          u64 min_trans)
5152 {
5153         struct btrfs_fs_info *fs_info = root->fs_info;
5154         struct extent_buffer *cur;
5155         struct btrfs_key found_key;
5156         int slot;
5157         int sret;
5158         u32 nritems;
5159         int level;
5160         int ret = 1;
5161         int keep_locks = path->keep_locks;
5162 
5163         path->keep_locks = 1;
5164 again:
5165         cur = btrfs_read_lock_root_node(root);
5166         level = btrfs_header_level(cur);
5167         WARN_ON(path->nodes[level]);
5168         path->nodes[level] = cur;
5169         path->locks[level] = BTRFS_READ_LOCK;
5170 
5171         if (btrfs_header_generation(cur) < min_trans) {
5172                 ret = 1;
5173                 goto out;
5174         }
5175         while (1) {
5176                 nritems = btrfs_header_nritems(cur);
5177                 level = btrfs_header_level(cur);
5178                 sret = bin_search(cur, min_key, level, &slot);
5179 
5180                 /* at the lowest level, we're done, setup the path and exit */
5181                 if (level == path->lowest_level) {
5182                         if (slot >= nritems)
5183                                 goto find_next_key;
5184                         ret = 0;
5185                         path->slots[level] = slot;
5186                         btrfs_item_key_to_cpu(cur, &found_key, slot);
5187                         goto out;
5188                 }
5189                 if (sret && slot > 0)
5190                         slot--;
5191                 /*
5192                  * check this node pointer against the min_trans parameters.
5193                  * If it is too old, old, skip to the next one.
5194                  */
5195                 while (slot < nritems) {
5196                         u64 gen;
5197 
5198                         gen = btrfs_node_ptr_generation(cur, slot);
5199                         if (gen < min_trans) {
5200                                 slot++;
5201                                 continue;
5202                         }
5203                         break;
5204                 }
5205 find_next_key:
5206                 /*
5207                  * we didn't find a candidate key in this node, walk forward
5208                  * and find another one
5209                  */
5210                 if (slot >= nritems) {
5211                         path->slots[level] = slot;
5212                         btrfs_set_path_blocking(path);
5213                         sret = btrfs_find_next_key(root, path, min_key, level,
5214                                                   min_trans);
5215                         if (sret == 0) {
5216                                 btrfs_release_path(path);
5217                                 goto again;
5218                         } else {
5219                                 goto out;
5220                         }
5221                 }
5222                 /* save our key for returning back */
5223                 btrfs_node_key_to_cpu(cur, &found_key, slot);
5224                 path->slots[level] = slot;
5225                 if (level == path->lowest_level) {
5226                         ret = 0;
5227                         goto out;
5228                 }
5229                 btrfs_set_path_blocking(path);
5230                 cur = read_node_slot(fs_info, cur, slot);
5231                 if (IS_ERR(cur)) {
5232                         ret = PTR_ERR(cur);
5233                         goto out;
5234                 }
5235 
5236                 btrfs_tree_read_lock(cur);
5237 
5238                 path->locks[level - 1] = BTRFS_READ_LOCK;
5239                 path->nodes[level - 1] = cur;
5240                 unlock_up(path, level, 1, 0, NULL);
5241                 btrfs_clear_path_blocking(path, NULL, 0);
5242         }
5243 out:
5244         path->keep_locks = keep_locks;
5245         if (ret == 0) {
5246                 btrfs_unlock_up_safe(path, path->lowest_level + 1);
5247                 btrfs_set_path_blocking(path);
5248                 memcpy(min_key, &found_key, sizeof(found_key));
5249         }
5250         return ret;
5251 }
5252 
5253 static int tree_move_down(struct btrfs_fs_info *fs_info,
5254                            struct btrfs_path *path,
5255                            int *level)
5256 {
5257         struct extent_buffer *eb;
5258 
5259         BUG_ON(*level == 0);
5260         eb = read_node_slot(fs_info, path->nodes[*level], path->slots[*level]);
5261         if (IS_ERR(eb))
5262                 return PTR_ERR(eb);
5263 
5264         path->nodes[*level - 1] = eb;
5265         path->slots[*level - 1] = 0;
5266         (*level)--;
5267         return 0;
5268 }
5269 
5270 static int tree_move_next_or_upnext(struct btrfs_path *path,
5271                                     int *level, int root_level)
5272 {
5273         int ret = 0;
5274         int nritems;
5275         nritems = btrfs_header_nritems(path->nodes[*level]);
5276 
5277         path->slots[*level]++;
5278 
5279         while (path->slots[*level] >= nritems) {
5280                 if (*level == root_level)
5281                         return -1;
5282 
5283                 /* move upnext */
5284                 path->slots[*level] = 0;
5285                 free_extent_buffer(path->nodes[*level]);
5286                 path->nodes[*level] = NULL;
5287                 (*level)++;
5288                 path->slots[*level]++;
5289 
5290                 nritems = btrfs_header_nritems(path->nodes[*level]);
5291                 ret = 1;
5292         }
5293         return ret;
5294 }
5295 
5296 /*
5297  * Returns 1 if it had to move up and next. 0 is returned if it moved only next
5298  * or down.
5299  */
5300 static int tree_advance(struct btrfs_fs_info *fs_info,
5301                         struct btrfs_path *path,
5302                         int *level, int root_level,
5303                         int allow_down,
5304                         struct btrfs_key *key)
5305 {
5306         int ret;
5307 
5308         if (*level == 0 || !allow_down) {
5309                 ret = tree_move_next_or_upnext(path, level, root_level);
5310         } else {
5311                 ret = tree_move_down(fs_info, path, level);
5312         }
5313         if (ret >= 0) {
5314                 if (*level == 0)
5315                         btrfs_item_key_to_cpu(path->nodes[*level], key,
5316                                         path->slots[*level]);
5317                 else
5318                         btrfs_node_key_to_cpu(path->nodes[*level], key,
5319                                         path->slots[*level]);
5320         }
5321         return ret;
5322 }
5323 
5324 static int tree_compare_item(struct btrfs_path *left_path,
5325                              struct btrfs_path *right_path,
5326                              char *tmp_buf)
5327 {
5328         int cmp;
5329         int len1, len2;
5330         unsigned long off1, off2;
5331 
5332         len1 = btrfs_item_size_nr(left_path->nodes[0], left_path->slots[0]);
5333         len2 = btrfs_item_size_nr(right_path->nodes[0], right_path->slots[0]);
5334         if (len1 != len2)
5335                 return 1;
5336 
5337         off1 = btrfs_item_ptr_offset(left_path->nodes[0], left_path->slots[0]);
5338         off2 = btrfs_item_ptr_offset(right_path->nodes[0],
5339                                 right_path->slots[0]);
5340 
5341         read_extent_buffer(left_path->nodes[0], tmp_buf, off1, len1);
5342 
5343         cmp = memcmp_extent_buffer(right_path->nodes[0], tmp_buf, off2, len1);
5344         if (cmp)
5345                 return 1;
5346         return 0;
5347 }
5348 
5349 #define ADVANCE 1
5350 #define ADVANCE_ONLY_NEXT -1
5351 
5352 /*
5353  * This function compares two trees and calls the provided callback for
5354  * every changed/new/deleted item it finds.
5355  * If shared tree blocks are encountered, whole subtrees are skipped, making
5356  * the compare pretty fast on snapshotted subvolumes.
5357  *
5358  * This currently works on commit roots only. As commit roots are read only,
5359  * we don't do any locking. The commit roots are protected with transactions.
5360  * Transactions are ended and rejoined when a commit is tried in between.
5361  *
5362  * This function checks for modifications done to the trees while comparing.
5363  * If it detects a change, it aborts immediately.
5364  */
5365 int btrfs_compare_trees(struct btrfs_root *left_root,
5366                         struct btrfs_root *right_root,
5367                         btrfs_changed_cb_t changed_cb, void *ctx)
5368 {
5369         struct btrfs_fs_info *fs_info = left_root->fs_info;
5370         int ret;
5371         int cmp;
5372         struct btrfs_path *left_path = NULL;
5373         struct btrfs_path *right_path = NULL;
5374         struct btrfs_key left_key;
5375         struct btrfs_key right_key;
5376         char *tmp_buf = NULL;
5377         int left_root_level;
5378         int right_root_level;
5379         int left_level;
5380         int right_level;
5381         int left_end_reached;
5382         int right_end_reached;
5383         int advance_left;
5384         int advance_right;
5385         u64 left_blockptr;
5386         u64 right_blockptr;
5387         u64 left_gen;
5388         u64 right_gen;
5389 
5390         left_path = btrfs_alloc_path();
5391         if (!left_path) {
5392                 ret = -ENOMEM;
5393                 goto out;
5394         }
5395         right_path = btrfs_alloc_path();
5396         if (!right_path) {
5397                 ret = -ENOMEM;
5398                 goto out;
5399         }
5400 
5401         tmp_buf = kvmalloc(fs_info->nodesize, GFP_KERNEL);
5402         if (!tmp_buf) {
5403                 ret = -ENOMEM;
5404                 goto out;
5405         }
5406 
5407         left_path->search_commit_root = 1;
5408         left_path->skip_locking = 1;
5409         right_path->search_commit_root = 1;
5410         right_path->skip_locking = 1;
5411 
5412         /*
5413          * Strategy: Go to the first items of both trees. Then do
5414          *
5415          * If both trees are at level 0
5416          *   Compare keys of current items
5417          *     If left < right treat left item as new, advance left tree
5418          *       and repeat
5419          *     If left > right treat right item as deleted, advance right tree
5420          *       and repeat
5421          *     If left == right do deep compare of items, treat as changed if
5422          *       needed, advance both trees and repeat
5423          * If both trees are at the same level but not at level 0
5424          *   Compare keys of current nodes/leafs
5425          *     If left < right advance left tree and repeat
5426          *     If left > right advance right tree and repeat
5427          *     If left == right compare blockptrs of the next nodes/leafs
5428          *       If they match advance both trees but stay at the same level
5429          *         and repeat
5430          *       If they don't match advance both trees while allowing to go
5431          *         deeper and repeat
5432          * If tree levels are different
5433          *   Advance the tree that needs it and repeat
5434          *
5435          * Advancing a tree means:
5436          *   If we are at level 0, try to go to the next slot. If that's not
5437          *   possible, go one level up and repeat. Stop when we found a level
5438          *   where we could go to the next slot. We may at this point be on a
5439          *   node or a leaf.
5440          *
5441          *   If we are not at level 0 and not on shared tree blocks, go one
5442          *   level deeper.
5443          *
5444          *   If we are not at level 0 and on shared tree blocks, go one slot to
5445          *   the right if possible or go up and right.
5446          */
5447 
5448         down_read(&fs_info->commit_root_sem);
5449         left_level = btrfs_header_level(left_root->commit_root);
5450         left_root_level = left_level;
5451         left_path->nodes[left_level] = left_root->commit_root;
5452         extent_buffer_get(left_path->nodes[left_level]);
5453 
5454         right_level = btrfs_header_level(right_root->commit_root);
5455         right_root_level = right_level;
5456         right_path->nodes[right_level] = right_root->commit_root;
5457         extent_buffer_get(right_path->nodes[right_level]);
5458         up_read(&fs_info->commit_root_sem);
5459 
5460         if (left_level == 0)
5461                 btrfs_item_key_to_cpu(left_path->nodes[left_level],
5462                                 &left_key, left_path->slots[left_level]);
5463         else
5464                 btrfs_node_key_to_cpu(left_path->nodes[left_level],
5465                                 &left_key, left_path->slots[left_level]);
5466         if (right_level == 0)
5467                 btrfs_item_key_to_cpu(right_path->nodes[right_level],
5468                                 &right_key, right_path->slots[right_level]);
5469         else
5470                 btrfs_node_key_to_cpu(right_path->nodes[right_level],
5471                                 &right_key, right_path->slots[right_level]);
5472 
5473         left_end_reached = right_end_reached = 0;
5474         advance_left = advance_right = 0;
5475 
5476         while (1) {
5477                 if (advance_left && !left_end_reached) {
5478                         ret = tree_advance(fs_info, left_path, &left_level,
5479                                         left_root_level,
5480                                         advance_left != ADVANCE_ONLY_NEXT,
5481                                         &left_key);
5482                         if (ret == -1)
5483                                 left_end_reached = ADVANCE;
5484                         else if (ret < 0)
5485                                 goto out;
5486                         advance_left = 0;
5487                 }
5488                 if (advance_right && !right_end_reached) {
5489                         ret = tree_advance(fs_info, right_path, &right_level,
5490                                         right_root_level,
5491                                         advance_right != ADVANCE_ONLY_NEXT,
5492                                         &right_key);
5493                         if (ret == -1)
5494                                 right_end_reached = ADVANCE;
5495                         else if (ret < 0)
5496                                 goto out;
5497                         advance_right = 0;
5498                 }
5499 
5500                 if (left_end_reached && right_end_reached) {
5501                         ret = 0;
5502                         goto out;
5503                 } else if (left_end_reached) {
5504                         if (right_level == 0) {
5505                                 ret = changed_cb(left_path, right_path,
5506                                                 &right_key,
5507                                                 BTRFS_COMPARE_TREE_DELETED,
5508                                                 ctx);
5509                                 if (ret < 0)
5510                                         goto out;
5511                         }
5512                         advance_right = ADVANCE;
5513                         continue;
5514                 } else if (right_end_reached) {
5515                         if (left_level == 0) {
5516                                 ret = changed_cb(left_path, right_path,
5517                                                 &left_key,
5518                                                 BTRFS_COMPARE_TREE_NEW,
5519                                                 ctx);
5520                                 if (ret < 0)
5521                                         goto out;
5522                         }
5523                         advance_left = ADVANCE;
5524                         continue;
5525                 }
5526 
5527                 if (left_level == 0 && right_level == 0) {
5528                         cmp = btrfs_comp_cpu_keys(&left_key, &right_key);
5529                         if (cmp < 0) {
5530                                 ret = changed_cb(left_path, right_path,
5531                                                 &left_key,
5532                                                 BTRFS_COMPARE_TREE_NEW,
5533                                                 ctx);
5534                                 if (ret < 0)
5535                                         goto out;
5536                                 advance_left = ADVANCE;
5537                         } else if (cmp > 0) {
5538                                 ret = changed_cb(left_path, right_path,
5539                                                 &right_key,
5540                                                 BTRFS_COMPARE_TREE_DELETED,
5541                                                 ctx);
5542                                 if (ret < 0)
5543                                         goto out;
5544                                 advance_right = ADVANCE;
5545                         } else {
5546                                 enum btrfs_compare_tree_result result;
5547 
5548                                 WARN_ON(!extent_buffer_uptodate(left_path->nodes[0]));
5549                                 ret = tree_compare_item(left_path, right_path,
5550                                                         tmp_buf);
5551                                 if (ret)
5552                                         result = BTRFS_COMPARE_TREE_CHANGED;
5553                                 else
5554                                         result = BTRFS_COMPARE_TREE_SAME;
5555                                 ret = changed_cb(left_path, right_path,
5556                                                  &left_key, result, ctx);
5557                                 if (ret < 0)
5558                                         goto out;
5559                                 advance_left = ADVANCE;
5560                                 advance_right = ADVANCE;
5561                         }
5562                 } else if (left_level == right_level) {
5563                         cmp = btrfs_comp_cpu_keys(&left_key, &right_key);
5564                         if (cmp < 0) {
5565                                 advance_left = ADVANCE;
5566                         } else if (cmp > 0) {
5567                                 advance_right = ADVANCE;
5568                         } else {
5569                                 left_blockptr = btrfs_node_blockptr(
5570                                                 left_path->nodes[left_level],
5571                                                 left_path->slots[left_level]);
5572                                 right_blockptr = btrfs_node_blockptr(
5573                                                 right_path->nodes[right_level],
5574                                                 right_path->slots[right_level]);
5575                                 left_gen = btrfs_node_ptr_generation(
5576                                                 left_path->nodes[left_level],
5577                                                 left_path->slots[left_level]);
5578                                 right_gen = btrfs_node_ptr_generation(
5579                                                 right_path->nodes[right_level],
5580                                                 right_path->slots[right_level]);
5581                                 if (left_blockptr == right_blockptr &&
5582                                     left_gen == right_gen) {
5583                                         /*
5584                                          * As we're on a shared block, don't
5585                                          * allow to go deeper.
5586                                          */
5587                                         advance_left = ADVANCE_ONLY_NEXT;
5588                                         advance_right = ADVANCE_ONLY_NEXT;
5589                                 } else {
5590                                         advance_left = ADVANCE;
5591                                         advance_right = ADVANCE;
5592                                 }
5593                         }
5594                 } else if (left_level < right_level) {
5595                         advance_right = ADVANCE;
5596                 } else {
5597                         advance_left = ADVANCE;
5598                 }
5599         }
5600 
5601 out:
5602         btrfs_free_path(left_path);
5603         btrfs_free_path(right_path);
5604         kvfree(tmp_buf);
5605         return ret;
5606 }
5607 
5608 /*
5609  * this is similar to btrfs_next_leaf, but does not try to preserve
5610  * and fixup the path.  It looks for and returns the next key in the
5611  * tree based on the current path and the min_trans parameters.
5612  *
5613  * 0 is returned if another key is found, < 0 if there are any errors
5614  * and 1 is returned if there are no higher keys in the tree
5615  *
5616  * path->keep_locks should be set to 1 on the search made before
5617  * calling this function.
5618  */
5619 int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path,
5620                         struct btrfs_key *key, int level, u64 min_trans)
5621 {
5622         int slot;
5623         struct extent_buffer *c;
5624 
5625         WARN_ON(!path->keep_locks);
5626         while (level<