Commit | Line | Data |
---|---|---|
6cbd5570 | 1 | /* |
d352ac68 | 2 | * Copyright (C) 2007,2008 Oracle. All rights reserved. |
6cbd5570 CM |
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 | ||
a6b6e75e | 19 | #include <linux/sched.h> |
5a0e3ad6 | 20 | #include <linux/slab.h> |
bd989ba3 | 21 | #include <linux/rbtree.h> |
eb60ceac CM |
22 | #include "ctree.h" |
23 | #include "disk-io.h" | |
7f5c1516 | 24 | #include "transaction.h" |
5f39d397 | 25 | #include "print-tree.h" |
925baedd | 26 | #include "locking.h" |
9a8dd150 | 27 | |
e089f05c CM |
28 | static int split_node(struct btrfs_trans_handle *trans, struct btrfs_root |
29 | *root, struct btrfs_path *path, int level); | |
30 | static int split_leaf(struct btrfs_trans_handle *trans, struct btrfs_root | |
d4dbff95 | 31 | *root, struct btrfs_key *ins_key, |
cc0c5538 | 32 | struct btrfs_path *path, int data_size, int extend); |
5f39d397 CM |
33 | static int push_node_left(struct btrfs_trans_handle *trans, |
34 | struct btrfs_root *root, struct extent_buffer *dst, | |
971a1f66 | 35 | struct extent_buffer *src, int empty); |
5f39d397 CM |
36 | static int balance_node_right(struct btrfs_trans_handle *trans, |
37 | struct btrfs_root *root, | |
38 | struct extent_buffer *dst_buf, | |
39 | struct extent_buffer *src_buf); | |
afe5fea7 TI |
40 | static void del_ptr(struct btrfs_root *root, struct btrfs_path *path, |
41 | int level, int slot); | |
f230475e JS |
42 | static void tree_mod_log_free_eb(struct btrfs_fs_info *fs_info, |
43 | struct extent_buffer *eb); | |
44 | struct extent_buffer *read_old_tree_block(struct btrfs_root *root, u64 bytenr, | |
45 | u32 blocksize, u64 parent_transid, | |
46 | u64 time_seq); | |
47 | struct extent_buffer *btrfs_find_old_tree_block(struct btrfs_root *root, | |
48 | u64 bytenr, u32 blocksize, | |
49 | u64 time_seq); | |
d97e63b6 | 50 | |
df24a2b9 | 51 | struct btrfs_path *btrfs_alloc_path(void) |
2c90e5d6 | 52 | { |
df24a2b9 | 53 | struct btrfs_path *path; |
e00f7308 | 54 | path = kmem_cache_zalloc(btrfs_path_cachep, GFP_NOFS); |
df24a2b9 | 55 | return path; |
2c90e5d6 CM |
56 | } |
57 | ||
b4ce94de CM |
58 | /* |
59 | * set all locked nodes in the path to blocking locks. This should | |
60 | * be done before scheduling | |
61 | */ | |
62 | noinline void btrfs_set_path_blocking(struct btrfs_path *p) | |
63 | { | |
64 | int i; | |
65 | for (i = 0; i < BTRFS_MAX_LEVEL; i++) { | |
bd681513 CM |
66 | if (!p->nodes[i] || !p->locks[i]) |
67 | continue; | |
68 | btrfs_set_lock_blocking_rw(p->nodes[i], p->locks[i]); | |
69 | if (p->locks[i] == BTRFS_READ_LOCK) | |
70 | p->locks[i] = BTRFS_READ_LOCK_BLOCKING; | |
71 | else if (p->locks[i] == BTRFS_WRITE_LOCK) | |
72 | p->locks[i] = BTRFS_WRITE_LOCK_BLOCKING; | |
b4ce94de CM |
73 | } |
74 | } | |
75 | ||
76 | /* | |
77 | * reset all the locked nodes in the patch to spinning locks. | |
4008c04a CM |
78 | * |
79 | * held is used to keep lockdep happy, when lockdep is enabled | |
80 | * we set held to a blocking lock before we go around and | |
81 | * retake all the spinlocks in the path. You can safely use NULL | |
82 | * for held | |
b4ce94de | 83 | */ |
4008c04a | 84 | noinline void btrfs_clear_path_blocking(struct btrfs_path *p, |
bd681513 | 85 | struct extent_buffer *held, int held_rw) |
b4ce94de CM |
86 | { |
87 | int i; | |
4008c04a CM |
88 | |
89 | #ifdef CONFIG_DEBUG_LOCK_ALLOC | |
90 | /* lockdep really cares that we take all of these spinlocks | |
91 | * in the right order. If any of the locks in the path are not | |
92 | * currently blocking, it is going to complain. So, make really | |
93 | * really sure by forcing the path to blocking before we clear | |
94 | * the path blocking. | |
95 | */ | |
bd681513 CM |
96 | if (held) { |
97 | btrfs_set_lock_blocking_rw(held, held_rw); | |
98 | if (held_rw == BTRFS_WRITE_LOCK) | |
99 | held_rw = BTRFS_WRITE_LOCK_BLOCKING; | |
100 | else if (held_rw == BTRFS_READ_LOCK) | |
101 | held_rw = BTRFS_READ_LOCK_BLOCKING; | |
102 | } | |
4008c04a CM |
103 | btrfs_set_path_blocking(p); |
104 | #endif | |
105 | ||
106 | for (i = BTRFS_MAX_LEVEL - 1; i >= 0; i--) { | |
bd681513 CM |
107 | if (p->nodes[i] && p->locks[i]) { |
108 | btrfs_clear_lock_blocking_rw(p->nodes[i], p->locks[i]); | |
109 | if (p->locks[i] == BTRFS_WRITE_LOCK_BLOCKING) | |
110 | p->locks[i] = BTRFS_WRITE_LOCK; | |
111 | else if (p->locks[i] == BTRFS_READ_LOCK_BLOCKING) | |
112 | p->locks[i] = BTRFS_READ_LOCK; | |
113 | } | |
b4ce94de | 114 | } |
4008c04a CM |
115 | |
116 | #ifdef CONFIG_DEBUG_LOCK_ALLOC | |
117 | if (held) | |
bd681513 | 118 | btrfs_clear_lock_blocking_rw(held, held_rw); |
4008c04a | 119 | #endif |
b4ce94de CM |
120 | } |
121 | ||
d352ac68 | 122 | /* this also releases the path */ |
df24a2b9 | 123 | void btrfs_free_path(struct btrfs_path *p) |
be0e5c09 | 124 | { |
ff175d57 JJ |
125 | if (!p) |
126 | return; | |
b3b4aa74 | 127 | btrfs_release_path(p); |
df24a2b9 | 128 | kmem_cache_free(btrfs_path_cachep, p); |
be0e5c09 CM |
129 | } |
130 | ||
d352ac68 CM |
131 | /* |
132 | * path release drops references on the extent buffers in the path | |
133 | * and it drops any locks held by this path | |
134 | * | |
135 | * It is safe to call this on paths that no locks or extent buffers held. | |
136 | */ | |
b3b4aa74 | 137 | noinline void btrfs_release_path(struct btrfs_path *p) |
eb60ceac CM |
138 | { |
139 | int i; | |
a2135011 | 140 | |
234b63a0 | 141 | for (i = 0; i < BTRFS_MAX_LEVEL; i++) { |
3f157a2f | 142 | p->slots[i] = 0; |
eb60ceac | 143 | if (!p->nodes[i]) |
925baedd CM |
144 | continue; |
145 | if (p->locks[i]) { | |
bd681513 | 146 | btrfs_tree_unlock_rw(p->nodes[i], p->locks[i]); |
925baedd CM |
147 | p->locks[i] = 0; |
148 | } | |
5f39d397 | 149 | free_extent_buffer(p->nodes[i]); |
3f157a2f | 150 | p->nodes[i] = NULL; |
eb60ceac CM |
151 | } |
152 | } | |
153 | ||
d352ac68 CM |
154 | /* |
155 | * safely gets a reference on the root node of a tree. A lock | |
156 | * is not taken, so a concurrent writer may put a different node | |
157 | * at the root of the tree. See btrfs_lock_root_node for the | |
158 | * looping required. | |
159 | * | |
160 | * The extent buffer returned by this has a reference taken, so | |
161 | * it won't disappear. It may stop being the root of the tree | |
162 | * at any time because there are no locks held. | |
163 | */ | |
925baedd CM |
164 | struct extent_buffer *btrfs_root_node(struct btrfs_root *root) |
165 | { | |
166 | struct extent_buffer *eb; | |
240f62c8 | 167 | |
3083ee2e JB |
168 | while (1) { |
169 | rcu_read_lock(); | |
170 | eb = rcu_dereference(root->node); | |
171 | ||
172 | /* | |
173 | * RCU really hurts here, we could free up the root node because | |
174 | * it was cow'ed but we may not get the new root node yet so do | |
175 | * the inc_not_zero dance and if it doesn't work then | |
176 | * synchronize_rcu and try again. | |
177 | */ | |
178 | if (atomic_inc_not_zero(&eb->refs)) { | |
179 | rcu_read_unlock(); | |
180 | break; | |
181 | } | |
182 | rcu_read_unlock(); | |
183 | synchronize_rcu(); | |
184 | } | |
925baedd CM |
185 | return eb; |
186 | } | |
187 | ||
d352ac68 CM |
188 | /* loop around taking references on and locking the root node of the |
189 | * tree until you end up with a lock on the root. A locked buffer | |
190 | * is returned, with a reference held. | |
191 | */ | |
925baedd CM |
192 | struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root) |
193 | { | |
194 | struct extent_buffer *eb; | |
195 | ||
d397712b | 196 | while (1) { |
925baedd CM |
197 | eb = btrfs_root_node(root); |
198 | btrfs_tree_lock(eb); | |
240f62c8 | 199 | if (eb == root->node) |
925baedd | 200 | break; |
925baedd CM |
201 | btrfs_tree_unlock(eb); |
202 | free_extent_buffer(eb); | |
203 | } | |
204 | return eb; | |
205 | } | |
206 | ||
bd681513 CM |
207 | /* loop around taking references on and locking the root node of the |
208 | * tree until you end up with a lock on the root. A locked buffer | |
209 | * is returned, with a reference held. | |
210 | */ | |
211 | struct extent_buffer *btrfs_read_lock_root_node(struct btrfs_root *root) | |
212 | { | |
213 | struct extent_buffer *eb; | |
214 | ||
215 | while (1) { | |
216 | eb = btrfs_root_node(root); | |
217 | btrfs_tree_read_lock(eb); | |
218 | if (eb == root->node) | |
219 | break; | |
220 | btrfs_tree_read_unlock(eb); | |
221 | free_extent_buffer(eb); | |
222 | } | |
223 | return eb; | |
224 | } | |
225 | ||
d352ac68 CM |
226 | /* cowonly root (everything not a reference counted cow subvolume), just get |
227 | * put onto a simple dirty list. transaction.c walks this to make sure they | |
228 | * get properly updated on disk. | |
229 | */ | |
0b86a832 CM |
230 | static void add_root_to_dirty_list(struct btrfs_root *root) |
231 | { | |
e5846fc6 | 232 | spin_lock(&root->fs_info->trans_lock); |
0b86a832 CM |
233 | if (root->track_dirty && list_empty(&root->dirty_list)) { |
234 | list_add(&root->dirty_list, | |
235 | &root->fs_info->dirty_cowonly_roots); | |
236 | } | |
e5846fc6 | 237 | spin_unlock(&root->fs_info->trans_lock); |
0b86a832 CM |
238 | } |
239 | ||
d352ac68 CM |
240 | /* |
241 | * used by snapshot creation to make a copy of a root for a tree with | |
242 | * a given objectid. The buffer with the new root node is returned in | |
243 | * cow_ret, and this func returns zero on success or a negative error code. | |
244 | */ | |
be20aa9d CM |
245 | int btrfs_copy_root(struct btrfs_trans_handle *trans, |
246 | struct btrfs_root *root, | |
247 | struct extent_buffer *buf, | |
248 | struct extent_buffer **cow_ret, u64 new_root_objectid) | |
249 | { | |
250 | struct extent_buffer *cow; | |
be20aa9d CM |
251 | int ret = 0; |
252 | int level; | |
5d4f98a2 | 253 | struct btrfs_disk_key disk_key; |
be20aa9d CM |
254 | |
255 | WARN_ON(root->ref_cows && trans->transid != | |
256 | root->fs_info->running_transaction->transid); | |
257 | WARN_ON(root->ref_cows && trans->transid != root->last_trans); | |
258 | ||
259 | level = btrfs_header_level(buf); | |
5d4f98a2 YZ |
260 | if (level == 0) |
261 | btrfs_item_key(buf, &disk_key, 0); | |
262 | else | |
263 | btrfs_node_key(buf, &disk_key, 0); | |
31840ae1 | 264 | |
5d4f98a2 YZ |
265 | cow = btrfs_alloc_free_block(trans, root, buf->len, 0, |
266 | new_root_objectid, &disk_key, level, | |
5581a51a | 267 | buf->start, 0); |
5d4f98a2 | 268 | if (IS_ERR(cow)) |
be20aa9d CM |
269 | return PTR_ERR(cow); |
270 | ||
271 | copy_extent_buffer(cow, buf, 0, 0, cow->len); | |
272 | btrfs_set_header_bytenr(cow, cow->start); | |
273 | btrfs_set_header_generation(cow, trans->transid); | |
5d4f98a2 YZ |
274 | btrfs_set_header_backref_rev(cow, BTRFS_MIXED_BACKREF_REV); |
275 | btrfs_clear_header_flag(cow, BTRFS_HEADER_FLAG_WRITTEN | | |
276 | BTRFS_HEADER_FLAG_RELOC); | |
277 | if (new_root_objectid == BTRFS_TREE_RELOC_OBJECTID) | |
278 | btrfs_set_header_flag(cow, BTRFS_HEADER_FLAG_RELOC); | |
279 | else | |
280 | btrfs_set_header_owner(cow, new_root_objectid); | |
be20aa9d | 281 | |
2b82032c YZ |
282 | write_extent_buffer(cow, root->fs_info->fsid, |
283 | (unsigned long)btrfs_header_fsid(cow), | |
284 | BTRFS_FSID_SIZE); | |
285 | ||
be20aa9d | 286 | WARN_ON(btrfs_header_generation(buf) > trans->transid); |
5d4f98a2 | 287 | if (new_root_objectid == BTRFS_TREE_RELOC_OBJECTID) |
66d7e7f0 | 288 | ret = btrfs_inc_ref(trans, root, cow, 1, 1); |
5d4f98a2 | 289 | else |
66d7e7f0 | 290 | ret = btrfs_inc_ref(trans, root, cow, 0, 1); |
4aec2b52 | 291 | |
be20aa9d CM |
292 | if (ret) |
293 | return ret; | |
294 | ||
295 | btrfs_mark_buffer_dirty(cow); | |
296 | *cow_ret = cow; | |
297 | return 0; | |
298 | } | |
299 | ||
bd989ba3 JS |
300 | enum mod_log_op { |
301 | MOD_LOG_KEY_REPLACE, | |
302 | MOD_LOG_KEY_ADD, | |
303 | MOD_LOG_KEY_REMOVE, | |
304 | MOD_LOG_KEY_REMOVE_WHILE_FREEING, | |
305 | MOD_LOG_KEY_REMOVE_WHILE_MOVING, | |
306 | MOD_LOG_MOVE_KEYS, | |
307 | MOD_LOG_ROOT_REPLACE, | |
308 | }; | |
309 | ||
310 | struct tree_mod_move { | |
311 | int dst_slot; | |
312 | int nr_items; | |
313 | }; | |
314 | ||
315 | struct tree_mod_root { | |
316 | u64 logical; | |
317 | u8 level; | |
318 | }; | |
319 | ||
320 | struct tree_mod_elem { | |
321 | struct rb_node node; | |
322 | u64 index; /* shifted logical */ | |
097b8a7c | 323 | u64 seq; |
bd989ba3 JS |
324 | enum mod_log_op op; |
325 | ||
326 | /* this is used for MOD_LOG_KEY_* and MOD_LOG_MOVE_KEYS operations */ | |
327 | int slot; | |
328 | ||
329 | /* this is used for MOD_LOG_KEY* and MOD_LOG_ROOT_REPLACE */ | |
330 | u64 generation; | |
331 | ||
332 | /* those are used for op == MOD_LOG_KEY_{REPLACE,REMOVE} */ | |
333 | struct btrfs_disk_key key; | |
334 | u64 blockptr; | |
335 | ||
336 | /* this is used for op == MOD_LOG_MOVE_KEYS */ | |
337 | struct tree_mod_move move; | |
338 | ||
339 | /* this is used for op == MOD_LOG_ROOT_REPLACE */ | |
340 | struct tree_mod_root old_root; | |
341 | }; | |
342 | ||
097b8a7c | 343 | static inline void tree_mod_log_read_lock(struct btrfs_fs_info *fs_info) |
bd989ba3 | 344 | { |
097b8a7c | 345 | read_lock(&fs_info->tree_mod_log_lock); |
bd989ba3 JS |
346 | } |
347 | ||
097b8a7c JS |
348 | static inline void tree_mod_log_read_unlock(struct btrfs_fs_info *fs_info) |
349 | { | |
350 | read_unlock(&fs_info->tree_mod_log_lock); | |
351 | } | |
352 | ||
353 | static inline void tree_mod_log_write_lock(struct btrfs_fs_info *fs_info) | |
354 | { | |
355 | write_lock(&fs_info->tree_mod_log_lock); | |
356 | } | |
357 | ||
358 | static inline void tree_mod_log_write_unlock(struct btrfs_fs_info *fs_info) | |
359 | { | |
360 | write_unlock(&fs_info->tree_mod_log_lock); | |
361 | } | |
362 | ||
fc36ed7e JS |
363 | /* |
364 | * Increment the upper half of tree_mod_seq, set lower half zero. | |
365 | * | |
366 | * Must be called with fs_info->tree_mod_seq_lock held. | |
367 | */ | |
368 | static inline u64 btrfs_inc_tree_mod_seq_major(struct btrfs_fs_info *fs_info) | |
369 | { | |
370 | u64 seq = atomic64_read(&fs_info->tree_mod_seq); | |
371 | seq &= 0xffffffff00000000ull; | |
372 | seq += 1ull << 32; | |
373 | atomic64_set(&fs_info->tree_mod_seq, seq); | |
374 | return seq; | |
375 | } | |
376 | ||
377 | /* | |
378 | * Increment the lower half of tree_mod_seq. | |
379 | * | |
380 | * Must be called with fs_info->tree_mod_seq_lock held. The way major numbers | |
381 | * are generated should not technically require a spin lock here. (Rationale: | |
382 | * incrementing the minor while incrementing the major seq number is between its | |
383 | * atomic64_read and atomic64_set calls doesn't duplicate sequence numbers, it | |
384 | * just returns a unique sequence number as usual.) We have decided to leave | |
385 | * that requirement in here and rethink it once we notice it really imposes a | |
386 | * problem on some workload. | |
387 | */ | |
388 | static inline u64 btrfs_inc_tree_mod_seq_minor(struct btrfs_fs_info *fs_info) | |
389 | { | |
390 | return atomic64_inc_return(&fs_info->tree_mod_seq); | |
391 | } | |
392 | ||
393 | /* | |
394 | * return the last minor in the previous major tree_mod_seq number | |
395 | */ | |
396 | u64 btrfs_tree_mod_seq_prev(u64 seq) | |
397 | { | |
398 | return (seq & 0xffffffff00000000ull) - 1ull; | |
399 | } | |
400 | ||
097b8a7c JS |
401 | /* |
402 | * This adds a new blocker to the tree mod log's blocker list if the @elem | |
403 | * passed does not already have a sequence number set. So when a caller expects | |
404 | * to record tree modifications, it should ensure to set elem->seq to zero | |
405 | * before calling btrfs_get_tree_mod_seq. | |
406 | * Returns a fresh, unused tree log modification sequence number, even if no new | |
407 | * blocker was added. | |
408 | */ | |
409 | u64 btrfs_get_tree_mod_seq(struct btrfs_fs_info *fs_info, | |
410 | struct seq_list *elem) | |
bd989ba3 | 411 | { |
097b8a7c JS |
412 | u64 seq; |
413 | ||
414 | tree_mod_log_write_lock(fs_info); | |
bd989ba3 | 415 | spin_lock(&fs_info->tree_mod_seq_lock); |
097b8a7c | 416 | if (!elem->seq) { |
fc36ed7e | 417 | elem->seq = btrfs_inc_tree_mod_seq_major(fs_info); |
097b8a7c JS |
418 | list_add_tail(&elem->list, &fs_info->tree_mod_seq_list); |
419 | } | |
fc36ed7e | 420 | seq = btrfs_inc_tree_mod_seq_minor(fs_info); |
bd989ba3 | 421 | spin_unlock(&fs_info->tree_mod_seq_lock); |
097b8a7c JS |
422 | tree_mod_log_write_unlock(fs_info); |
423 | ||
424 | return seq; | |
bd989ba3 JS |
425 | } |
426 | ||
427 | void btrfs_put_tree_mod_seq(struct btrfs_fs_info *fs_info, | |
428 | struct seq_list *elem) | |
429 | { | |
430 | struct rb_root *tm_root; | |
431 | struct rb_node *node; | |
432 | struct rb_node *next; | |
433 | struct seq_list *cur_elem; | |
434 | struct tree_mod_elem *tm; | |
435 | u64 min_seq = (u64)-1; | |
436 | u64 seq_putting = elem->seq; | |
437 | ||
438 | if (!seq_putting) | |
439 | return; | |
440 | ||
bd989ba3 JS |
441 | spin_lock(&fs_info->tree_mod_seq_lock); |
442 | list_del(&elem->list); | |
097b8a7c | 443 | elem->seq = 0; |
bd989ba3 JS |
444 | |
445 | list_for_each_entry(cur_elem, &fs_info->tree_mod_seq_list, list) { | |
097b8a7c | 446 | if (cur_elem->seq < min_seq) { |
bd989ba3 JS |
447 | if (seq_putting > cur_elem->seq) { |
448 | /* | |
449 | * blocker with lower sequence number exists, we | |
450 | * cannot remove anything from the log | |
451 | */ | |
097b8a7c JS |
452 | spin_unlock(&fs_info->tree_mod_seq_lock); |
453 | return; | |
bd989ba3 JS |
454 | } |
455 | min_seq = cur_elem->seq; | |
456 | } | |
457 | } | |
097b8a7c JS |
458 | spin_unlock(&fs_info->tree_mod_seq_lock); |
459 | ||
bd989ba3 JS |
460 | /* |
461 | * anything that's lower than the lowest existing (read: blocked) | |
462 | * sequence number can be removed from the tree. | |
463 | */ | |
097b8a7c | 464 | tree_mod_log_write_lock(fs_info); |
bd989ba3 JS |
465 | tm_root = &fs_info->tree_mod_log; |
466 | for (node = rb_first(tm_root); node; node = next) { | |
467 | next = rb_next(node); | |
468 | tm = container_of(node, struct tree_mod_elem, node); | |
097b8a7c | 469 | if (tm->seq > min_seq) |
bd989ba3 JS |
470 | continue; |
471 | rb_erase(node, tm_root); | |
bd989ba3 JS |
472 | kfree(tm); |
473 | } | |
097b8a7c | 474 | tree_mod_log_write_unlock(fs_info); |
bd989ba3 JS |
475 | } |
476 | ||
477 | /* | |
478 | * key order of the log: | |
479 | * index -> sequence | |
480 | * | |
481 | * the index is the shifted logical of the *new* root node for root replace | |
482 | * operations, or the shifted logical of the affected block for all other | |
483 | * operations. | |
484 | */ | |
485 | static noinline int | |
486 | __tree_mod_log_insert(struct btrfs_fs_info *fs_info, struct tree_mod_elem *tm) | |
487 | { | |
488 | struct rb_root *tm_root; | |
489 | struct rb_node **new; | |
490 | struct rb_node *parent = NULL; | |
491 | struct tree_mod_elem *cur; | |
bd989ba3 | 492 | |
097b8a7c | 493 | BUG_ON(!tm || !tm->seq); |
bd989ba3 | 494 | |
bd989ba3 JS |
495 | tm_root = &fs_info->tree_mod_log; |
496 | new = &tm_root->rb_node; | |
497 | while (*new) { | |
498 | cur = container_of(*new, struct tree_mod_elem, node); | |
499 | parent = *new; | |
500 | if (cur->index < tm->index) | |
501 | new = &((*new)->rb_left); | |
502 | else if (cur->index > tm->index) | |
503 | new = &((*new)->rb_right); | |
097b8a7c | 504 | else if (cur->seq < tm->seq) |
bd989ba3 | 505 | new = &((*new)->rb_left); |
097b8a7c | 506 | else if (cur->seq > tm->seq) |
bd989ba3 JS |
507 | new = &((*new)->rb_right); |
508 | else { | |
509 | kfree(tm); | |
097b8a7c | 510 | return -EEXIST; |
bd989ba3 JS |
511 | } |
512 | } | |
513 | ||
514 | rb_link_node(&tm->node, parent, new); | |
515 | rb_insert_color(&tm->node, tm_root); | |
097b8a7c | 516 | return 0; |
bd989ba3 JS |
517 | } |
518 | ||
097b8a7c JS |
519 | /* |
520 | * Determines if logging can be omitted. Returns 1 if it can. Otherwise, it | |
521 | * returns zero with the tree_mod_log_lock acquired. The caller must hold | |
522 | * this until all tree mod log insertions are recorded in the rb tree and then | |
523 | * call tree_mod_log_write_unlock() to release. | |
524 | */ | |
e9b7fd4d JS |
525 | static inline int tree_mod_dont_log(struct btrfs_fs_info *fs_info, |
526 | struct extent_buffer *eb) { | |
527 | smp_mb(); | |
528 | if (list_empty(&(fs_info)->tree_mod_seq_list)) | |
529 | return 1; | |
097b8a7c JS |
530 | if (eb && btrfs_header_level(eb) == 0) |
531 | return 1; | |
532 | ||
533 | tree_mod_log_write_lock(fs_info); | |
534 | if (list_empty(&fs_info->tree_mod_seq_list)) { | |
535 | /* | |
536 | * someone emptied the list while we were waiting for the lock. | |
537 | * we must not add to the list when no blocker exists. | |
538 | */ | |
539 | tree_mod_log_write_unlock(fs_info); | |
e9b7fd4d | 540 | return 1; |
097b8a7c JS |
541 | } |
542 | ||
e9b7fd4d JS |
543 | return 0; |
544 | } | |
545 | ||
3310c36e | 546 | /* |
097b8a7c | 547 | * This allocates memory and gets a tree modification sequence number. |
3310c36e | 548 | * |
097b8a7c JS |
549 | * Returns <0 on error. |
550 | * Returns >0 (the added sequence number) on success. | |
3310c36e | 551 | */ |
926dd8a6 JS |
552 | static inline int tree_mod_alloc(struct btrfs_fs_info *fs_info, gfp_t flags, |
553 | struct tree_mod_elem **tm_ret) | |
bd989ba3 JS |
554 | { |
555 | struct tree_mod_elem *tm; | |
bd989ba3 | 556 | |
097b8a7c JS |
557 | /* |
558 | * once we switch from spin locks to something different, we should | |
559 | * honor the flags parameter here. | |
560 | */ | |
561 | tm = *tm_ret = kzalloc(sizeof(*tm), GFP_ATOMIC); | |
bd989ba3 JS |
562 | if (!tm) |
563 | return -ENOMEM; | |
564 | ||
fc36ed7e JS |
565 | spin_lock(&fs_info->tree_mod_seq_lock); |
566 | tm->seq = btrfs_inc_tree_mod_seq_minor(fs_info); | |
567 | spin_unlock(&fs_info->tree_mod_seq_lock); | |
568 | ||
097b8a7c | 569 | return tm->seq; |
bd989ba3 JS |
570 | } |
571 | ||
097b8a7c JS |
572 | static inline int |
573 | __tree_mod_log_insert_key(struct btrfs_fs_info *fs_info, | |
574 | struct extent_buffer *eb, int slot, | |
575 | enum mod_log_op op, gfp_t flags) | |
bd989ba3 | 576 | { |
bd989ba3 | 577 | int ret; |
097b8a7c | 578 | struct tree_mod_elem *tm; |
bd989ba3 JS |
579 | |
580 | ret = tree_mod_alloc(fs_info, flags, &tm); | |
097b8a7c | 581 | if (ret < 0) |
bd989ba3 JS |
582 | return ret; |
583 | ||
584 | tm->index = eb->start >> PAGE_CACHE_SHIFT; | |
585 | if (op != MOD_LOG_KEY_ADD) { | |
586 | btrfs_node_key(eb, &tm->key, slot); | |
587 | tm->blockptr = btrfs_node_blockptr(eb, slot); | |
588 | } | |
589 | tm->op = op; | |
590 | tm->slot = slot; | |
591 | tm->generation = btrfs_node_ptr_generation(eb, slot); | |
592 | ||
097b8a7c JS |
593 | return __tree_mod_log_insert(fs_info, tm); |
594 | } | |
595 | ||
596 | static noinline int | |
597 | tree_mod_log_insert_key_mask(struct btrfs_fs_info *fs_info, | |
598 | struct extent_buffer *eb, int slot, | |
599 | enum mod_log_op op, gfp_t flags) | |
600 | { | |
601 | int ret; | |
602 | ||
603 | if (tree_mod_dont_log(fs_info, eb)) | |
604 | return 0; | |
605 | ||
606 | ret = __tree_mod_log_insert_key(fs_info, eb, slot, op, flags); | |
607 | ||
608 | tree_mod_log_write_unlock(fs_info); | |
3310c36e | 609 | return ret; |
bd989ba3 JS |
610 | } |
611 | ||
612 | static noinline int | |
613 | tree_mod_log_insert_key(struct btrfs_fs_info *fs_info, struct extent_buffer *eb, | |
614 | int slot, enum mod_log_op op) | |
615 | { | |
616 | return tree_mod_log_insert_key_mask(fs_info, eb, slot, op, GFP_NOFS); | |
617 | } | |
618 | ||
097b8a7c JS |
619 | static noinline int |
620 | tree_mod_log_insert_key_locked(struct btrfs_fs_info *fs_info, | |
621 | struct extent_buffer *eb, int slot, | |
622 | enum mod_log_op op) | |
623 | { | |
624 | return __tree_mod_log_insert_key(fs_info, eb, slot, op, GFP_NOFS); | |
625 | } | |
626 | ||
bd989ba3 JS |
627 | static noinline int |
628 | tree_mod_log_insert_move(struct btrfs_fs_info *fs_info, | |
629 | struct extent_buffer *eb, int dst_slot, int src_slot, | |
630 | int nr_items, gfp_t flags) | |
631 | { | |
632 | struct tree_mod_elem *tm; | |
633 | int ret; | |
634 | int i; | |
635 | ||
f395694c JS |
636 | if (tree_mod_dont_log(fs_info, eb)) |
637 | return 0; | |
bd989ba3 | 638 | |
01763a2e JS |
639 | /* |
640 | * When we override something during the move, we log these removals. | |
641 | * This can only happen when we move towards the beginning of the | |
642 | * buffer, i.e. dst_slot < src_slot. | |
643 | */ | |
bd989ba3 | 644 | for (i = 0; i + dst_slot < src_slot && i < nr_items; i++) { |
097b8a7c | 645 | ret = tree_mod_log_insert_key_locked(fs_info, eb, i + dst_slot, |
bd989ba3 JS |
646 | MOD_LOG_KEY_REMOVE_WHILE_MOVING); |
647 | BUG_ON(ret < 0); | |
648 | } | |
649 | ||
f395694c | 650 | ret = tree_mod_alloc(fs_info, flags, &tm); |
097b8a7c JS |
651 | if (ret < 0) |
652 | goto out; | |
f395694c | 653 | |
bd989ba3 JS |
654 | tm->index = eb->start >> PAGE_CACHE_SHIFT; |
655 | tm->slot = src_slot; | |
656 | tm->move.dst_slot = dst_slot; | |
657 | tm->move.nr_items = nr_items; | |
658 | tm->op = MOD_LOG_MOVE_KEYS; | |
659 | ||
3310c36e | 660 | ret = __tree_mod_log_insert(fs_info, tm); |
097b8a7c JS |
661 | out: |
662 | tree_mod_log_write_unlock(fs_info); | |
3310c36e | 663 | return ret; |
bd989ba3 JS |
664 | } |
665 | ||
097b8a7c JS |
666 | static inline void |
667 | __tree_mod_log_free_eb(struct btrfs_fs_info *fs_info, struct extent_buffer *eb) | |
668 | { | |
669 | int i; | |
670 | u32 nritems; | |
671 | int ret; | |
672 | ||
b12a3b1e CM |
673 | if (btrfs_header_level(eb) == 0) |
674 | return; | |
675 | ||
097b8a7c JS |
676 | nritems = btrfs_header_nritems(eb); |
677 | for (i = nritems - 1; i >= 0; i--) { | |
678 | ret = tree_mod_log_insert_key_locked(fs_info, eb, i, | |
679 | MOD_LOG_KEY_REMOVE_WHILE_FREEING); | |
680 | BUG_ON(ret < 0); | |
681 | } | |
682 | } | |
683 | ||
bd989ba3 JS |
684 | static noinline int |
685 | tree_mod_log_insert_root(struct btrfs_fs_info *fs_info, | |
686 | struct extent_buffer *old_root, | |
90f8d62e JS |
687 | struct extent_buffer *new_root, gfp_t flags, |
688 | int log_removal) | |
bd989ba3 JS |
689 | { |
690 | struct tree_mod_elem *tm; | |
691 | int ret; | |
692 | ||
097b8a7c JS |
693 | if (tree_mod_dont_log(fs_info, NULL)) |
694 | return 0; | |
695 | ||
90f8d62e JS |
696 | if (log_removal) |
697 | __tree_mod_log_free_eb(fs_info, old_root); | |
d9abbf1c | 698 | |
bd989ba3 | 699 | ret = tree_mod_alloc(fs_info, flags, &tm); |
097b8a7c JS |
700 | if (ret < 0) |
701 | goto out; | |
bd989ba3 JS |
702 | |
703 | tm->index = new_root->start >> PAGE_CACHE_SHIFT; | |
704 | tm->old_root.logical = old_root->start; | |
705 | tm->old_root.level = btrfs_header_level(old_root); | |
706 | tm->generation = btrfs_header_generation(old_root); | |
707 | tm->op = MOD_LOG_ROOT_REPLACE; | |
708 | ||
3310c36e | 709 | ret = __tree_mod_log_insert(fs_info, tm); |
097b8a7c JS |
710 | out: |
711 | tree_mod_log_write_unlock(fs_info); | |
3310c36e | 712 | return ret; |
bd989ba3 JS |
713 | } |
714 | ||
715 | static struct tree_mod_elem * | |
716 | __tree_mod_log_search(struct btrfs_fs_info *fs_info, u64 start, u64 min_seq, | |
717 | int smallest) | |
718 | { | |
719 | struct rb_root *tm_root; | |
720 | struct rb_node *node; | |
721 | struct tree_mod_elem *cur = NULL; | |
722 | struct tree_mod_elem *found = NULL; | |
723 | u64 index = start >> PAGE_CACHE_SHIFT; | |
724 | ||
097b8a7c | 725 | tree_mod_log_read_lock(fs_info); |
bd989ba3 JS |
726 | tm_root = &fs_info->tree_mod_log; |
727 | node = tm_root->rb_node; | |
728 | while (node) { | |
729 | cur = container_of(node, struct tree_mod_elem, node); | |
730 | if (cur->index < index) { | |
731 | node = node->rb_left; | |
732 | } else if (cur->index > index) { | |
733 | node = node->rb_right; | |
097b8a7c | 734 | } else if (cur->seq < min_seq) { |
bd989ba3 JS |
735 | node = node->rb_left; |
736 | } else if (!smallest) { | |
737 | /* we want the node with the highest seq */ | |
738 | if (found) | |
097b8a7c | 739 | BUG_ON(found->seq > cur->seq); |
bd989ba3 JS |
740 | found = cur; |
741 | node = node->rb_left; | |
097b8a7c | 742 | } else if (cur->seq > min_seq) { |
bd989ba3 JS |
743 | /* we want the node with the smallest seq */ |
744 | if (found) | |
097b8a7c | 745 | BUG_ON(found->seq < cur->seq); |
bd989ba3 JS |
746 | found = cur; |
747 | node = node->rb_right; | |
748 | } else { | |
749 | found = cur; | |
750 | break; | |
751 | } | |
752 | } | |
097b8a7c | 753 | tree_mod_log_read_unlock(fs_info); |
bd989ba3 JS |
754 | |
755 | return found; | |
756 | } | |
757 | ||
758 | /* | |
759 | * this returns the element from the log with the smallest time sequence | |
760 | * value that's in the log (the oldest log item). any element with a time | |
761 | * sequence lower than min_seq will be ignored. | |
762 | */ | |
763 | static struct tree_mod_elem * | |
764 | tree_mod_log_search_oldest(struct btrfs_fs_info *fs_info, u64 start, | |
765 | u64 min_seq) | |
766 | { | |
767 | return __tree_mod_log_search(fs_info, start, min_seq, 1); | |
768 | } | |
769 | ||
770 | /* | |
771 | * this returns the element from the log with the largest time sequence | |
772 | * value that's in the log (the most recent log item). any element with | |
773 | * a time sequence lower than min_seq will be ignored. | |
774 | */ | |
775 | static struct tree_mod_elem * | |
776 | tree_mod_log_search(struct btrfs_fs_info *fs_info, u64 start, u64 min_seq) | |
777 | { | |
778 | return __tree_mod_log_search(fs_info, start, min_seq, 0); | |
779 | } | |
780 | ||
097b8a7c | 781 | static noinline void |
bd989ba3 JS |
782 | tree_mod_log_eb_copy(struct btrfs_fs_info *fs_info, struct extent_buffer *dst, |
783 | struct extent_buffer *src, unsigned long dst_offset, | |
90f8d62e | 784 | unsigned long src_offset, int nr_items) |
bd989ba3 JS |
785 | { |
786 | int ret; | |
787 | int i; | |
788 | ||
e9b7fd4d | 789 | if (tree_mod_dont_log(fs_info, NULL)) |
bd989ba3 JS |
790 | return; |
791 | ||
097b8a7c JS |
792 | if (btrfs_header_level(dst) == 0 && btrfs_header_level(src) == 0) { |
793 | tree_mod_log_write_unlock(fs_info); | |
bd989ba3 | 794 | return; |
097b8a7c | 795 | } |
bd989ba3 | 796 | |
bd989ba3 | 797 | for (i = 0; i < nr_items; i++) { |
90f8d62e JS |
798 | ret = tree_mod_log_insert_key_locked(fs_info, src, |
799 | i + src_offset, | |
800 | MOD_LOG_KEY_REMOVE); | |
801 | BUG_ON(ret < 0); | |
097b8a7c JS |
802 | ret = tree_mod_log_insert_key_locked(fs_info, dst, |
803 | i + dst_offset, | |
804 | MOD_LOG_KEY_ADD); | |
bd989ba3 JS |
805 | BUG_ON(ret < 0); |
806 | } | |
097b8a7c JS |
807 | |
808 | tree_mod_log_write_unlock(fs_info); | |
bd989ba3 JS |
809 | } |
810 | ||
811 | static inline void | |
812 | tree_mod_log_eb_move(struct btrfs_fs_info *fs_info, struct extent_buffer *dst, | |
813 | int dst_offset, int src_offset, int nr_items) | |
814 | { | |
815 | int ret; | |
816 | ret = tree_mod_log_insert_move(fs_info, dst, dst_offset, src_offset, | |
817 | nr_items, GFP_NOFS); | |
818 | BUG_ON(ret < 0); | |
819 | } | |
820 | ||
097b8a7c | 821 | static noinline void |
bd989ba3 | 822 | tree_mod_log_set_node_key(struct btrfs_fs_info *fs_info, |
32adf090 | 823 | struct extent_buffer *eb, int slot, int atomic) |
bd989ba3 JS |
824 | { |
825 | int ret; | |
826 | ||
827 | ret = tree_mod_log_insert_key_mask(fs_info, eb, slot, | |
828 | MOD_LOG_KEY_REPLACE, | |
829 | atomic ? GFP_ATOMIC : GFP_NOFS); | |
830 | BUG_ON(ret < 0); | |
831 | } | |
832 | ||
097b8a7c JS |
833 | static noinline void |
834 | tree_mod_log_free_eb(struct btrfs_fs_info *fs_info, struct extent_buffer *eb) | |
bd989ba3 | 835 | { |
e9b7fd4d | 836 | if (tree_mod_dont_log(fs_info, eb)) |
bd989ba3 JS |
837 | return; |
838 | ||
097b8a7c JS |
839 | __tree_mod_log_free_eb(fs_info, eb); |
840 | ||
841 | tree_mod_log_write_unlock(fs_info); | |
bd989ba3 JS |
842 | } |
843 | ||
097b8a7c | 844 | static noinline void |
bd989ba3 | 845 | tree_mod_log_set_root_pointer(struct btrfs_root *root, |
90f8d62e JS |
846 | struct extent_buffer *new_root_node, |
847 | int log_removal) | |
bd989ba3 JS |
848 | { |
849 | int ret; | |
bd989ba3 | 850 | ret = tree_mod_log_insert_root(root->fs_info, root->node, |
90f8d62e | 851 | new_root_node, GFP_NOFS, log_removal); |
bd989ba3 JS |
852 | BUG_ON(ret < 0); |
853 | } | |
854 | ||
5d4f98a2 YZ |
855 | /* |
856 | * check if the tree block can be shared by multiple trees | |
857 | */ | |
858 | int btrfs_block_can_be_shared(struct btrfs_root *root, | |
859 | struct extent_buffer *buf) | |
860 | { | |
861 | /* | |
862 | * Tree blocks not in refernece counted trees and tree roots | |
863 | * are never shared. If a block was allocated after the last | |
864 | * snapshot and the block was not allocated by tree relocation, | |
865 | * we know the block is not shared. | |
866 | */ | |
867 | if (root->ref_cows && | |
868 | buf != root->node && buf != root->commit_root && | |
869 | (btrfs_header_generation(buf) <= | |
870 | btrfs_root_last_snapshot(&root->root_item) || | |
871 | btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC))) | |
872 | return 1; | |
873 | #ifdef BTRFS_COMPAT_EXTENT_TREE_V0 | |
874 | if (root->ref_cows && | |
875 | btrfs_header_backref_rev(buf) < BTRFS_MIXED_BACKREF_REV) | |
876 | return 1; | |
877 | #endif | |
878 | return 0; | |
879 | } | |
880 | ||
881 | static noinline int update_ref_for_cow(struct btrfs_trans_handle *trans, | |
882 | struct btrfs_root *root, | |
883 | struct extent_buffer *buf, | |
f0486c68 YZ |
884 | struct extent_buffer *cow, |
885 | int *last_ref) | |
5d4f98a2 YZ |
886 | { |
887 | u64 refs; | |
888 | u64 owner; | |
889 | u64 flags; | |
890 | u64 new_flags = 0; | |
891 | int ret; | |
892 | ||
893 | /* | |
894 | * Backrefs update rules: | |
895 | * | |
896 | * Always use full backrefs for extent pointers in tree block | |
897 | * allocated by tree relocation. | |
898 | * | |
899 | * If a shared tree block is no longer referenced by its owner | |
900 | * tree (btrfs_header_owner(buf) == root->root_key.objectid), | |
901 | * use full backrefs for extent pointers in tree block. | |
902 | * | |
903 | * If a tree block is been relocating | |
904 | * (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID), | |
905 | * use full backrefs for extent pointers in tree block. | |
906 | * The reason for this is some operations (such as drop tree) | |
907 | * are only allowed for blocks use full backrefs. | |
908 | */ | |
909 | ||
910 | if (btrfs_block_can_be_shared(root, buf)) { | |
911 | ret = btrfs_lookup_extent_info(trans, root, buf->start, | |
3173a18f JB |
912 | btrfs_header_level(buf), 1, |
913 | &refs, &flags); | |
be1a5564 MF |
914 | if (ret) |
915 | return ret; | |
e5df9573 MF |
916 | if (refs == 0) { |
917 | ret = -EROFS; | |
918 | btrfs_std_error(root->fs_info, ret); | |
919 | return ret; | |
920 | } | |
5d4f98a2 YZ |
921 | } else { |
922 | refs = 1; | |
923 | if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID || | |
924 | btrfs_header_backref_rev(buf) < BTRFS_MIXED_BACKREF_REV) | |
925 | flags = BTRFS_BLOCK_FLAG_FULL_BACKREF; | |
926 | else | |
927 | flags = 0; | |
928 | } | |
929 | ||
930 | owner = btrfs_header_owner(buf); | |
931 | BUG_ON(owner == BTRFS_TREE_RELOC_OBJECTID && | |
932 | !(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)); | |
933 | ||
934 | if (refs > 1) { | |
935 | if ((owner == root->root_key.objectid || | |
936 | root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) && | |
937 | !(flags & BTRFS_BLOCK_FLAG_FULL_BACKREF)) { | |
66d7e7f0 | 938 | ret = btrfs_inc_ref(trans, root, buf, 1, 1); |
79787eaa | 939 | BUG_ON(ret); /* -ENOMEM */ |
5d4f98a2 YZ |
940 | |
941 | if (root->root_key.objectid == | |
942 | BTRFS_TREE_RELOC_OBJECTID) { | |
66d7e7f0 | 943 | ret = btrfs_dec_ref(trans, root, buf, 0, 1); |
79787eaa | 944 | BUG_ON(ret); /* -ENOMEM */ |
66d7e7f0 | 945 | ret = btrfs_inc_ref(trans, root, cow, 1, 1); |
79787eaa | 946 | BUG_ON(ret); /* -ENOMEM */ |
5d4f98a2 YZ |
947 | } |
948 | new_flags |= BTRFS_BLOCK_FLAG_FULL_BACKREF; | |
949 | } else { | |
950 | ||
951 | if (root->root_key.objectid == | |
952 | BTRFS_TREE_RELOC_OBJECTID) | |
66d7e7f0 | 953 | ret = btrfs_inc_ref(trans, root, cow, 1, 1); |
5d4f98a2 | 954 | else |
66d7e7f0 | 955 | ret = btrfs_inc_ref(trans, root, cow, 0, 1); |
79787eaa | 956 | BUG_ON(ret); /* -ENOMEM */ |
5d4f98a2 YZ |
957 | } |
958 | if (new_flags != 0) { | |
959 | ret = btrfs_set_disk_extent_flags(trans, root, | |
960 | buf->start, | |
961 | buf->len, | |
962 | new_flags, 0); | |
be1a5564 MF |
963 | if (ret) |
964 | return ret; | |
5d4f98a2 YZ |
965 | } |
966 | } else { | |
967 | if (flags & BTRFS_BLOCK_FLAG_FULL_BACKREF) { | |
968 | if (root->root_key.objectid == | |
969 | BTRFS_TREE_RELOC_OBJECTID) | |
66d7e7f0 | 970 | ret = btrfs_inc_ref(trans, root, cow, 1, 1); |
5d4f98a2 | 971 | else |
66d7e7f0 | 972 | ret = btrfs_inc_ref(trans, root, cow, 0, 1); |
79787eaa | 973 | BUG_ON(ret); /* -ENOMEM */ |
66d7e7f0 | 974 | ret = btrfs_dec_ref(trans, root, buf, 1, 1); |
79787eaa | 975 | BUG_ON(ret); /* -ENOMEM */ |
5d4f98a2 YZ |
976 | } |
977 | clean_tree_block(trans, root, buf); | |
f0486c68 | 978 | *last_ref = 1; |
5d4f98a2 YZ |
979 | } |
980 | return 0; | |
981 | } | |
982 | ||
d352ac68 | 983 | /* |
d397712b CM |
984 | * does the dirty work in cow of a single block. The parent block (if |
985 | * supplied) is updated to point to the new cow copy. The new buffer is marked | |
986 | * dirty and returned locked. If you modify the block it needs to be marked | |
987 | * dirty again. | |
d352ac68 CM |
988 | * |
989 | * search_start -- an allocation hint for the new block | |
990 | * | |
d397712b CM |
991 | * empty_size -- a hint that you plan on doing more cow. This is the size in |
992 | * bytes the allocator should try to find free next to the block it returns. | |
993 | * This is just a hint and may be ignored by the allocator. | |
d352ac68 | 994 | */ |
d397712b | 995 | static noinline int __btrfs_cow_block(struct btrfs_trans_handle *trans, |
5f39d397 CM |
996 | struct btrfs_root *root, |
997 | struct extent_buffer *buf, | |
998 | struct extent_buffer *parent, int parent_slot, | |
999 | struct extent_buffer **cow_ret, | |
9fa8cfe7 | 1000 | u64 search_start, u64 empty_size) |
02217ed2 | 1001 | { |
5d4f98a2 | 1002 | struct btrfs_disk_key disk_key; |
5f39d397 | 1003 | struct extent_buffer *cow; |
be1a5564 | 1004 | int level, ret; |
f0486c68 | 1005 | int last_ref = 0; |
925baedd | 1006 | int unlock_orig = 0; |
5d4f98a2 | 1007 | u64 parent_start; |
7bb86316 | 1008 | |
925baedd CM |
1009 | if (*cow_ret == buf) |
1010 | unlock_orig = 1; | |
1011 | ||
b9447ef8 | 1012 | btrfs_assert_tree_locked(buf); |
925baedd | 1013 | |
7bb86316 CM |
1014 | WARN_ON(root->ref_cows && trans->transid != |
1015 | root->fs_info->running_transaction->transid); | |
6702ed49 | 1016 | WARN_ON(root->ref_cows && trans->transid != root->last_trans); |
5f39d397 | 1017 | |
7bb86316 | 1018 | level = btrfs_header_level(buf); |
31840ae1 | 1019 | |
5d4f98a2 YZ |
1020 | if (level == 0) |
1021 | btrfs_item_key(buf, &disk_key, 0); | |
1022 | else | |
1023 | btrfs_node_key(buf, &disk_key, 0); | |
1024 | ||
1025 | if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) { | |
1026 | if (parent) | |
1027 | parent_start = parent->start; | |
1028 | else | |
1029 | parent_start = 0; | |
1030 | } else | |
1031 | parent_start = 0; | |
1032 | ||
1033 | cow = btrfs_alloc_free_block(trans, root, buf->len, parent_start, | |
1034 | root->root_key.objectid, &disk_key, | |
5581a51a | 1035 | level, search_start, empty_size); |
54aa1f4d CM |
1036 | if (IS_ERR(cow)) |
1037 | return PTR_ERR(cow); | |
6702ed49 | 1038 | |
b4ce94de CM |
1039 | /* cow is set to blocking by btrfs_init_new_buffer */ |
1040 | ||
5f39d397 | 1041 | copy_extent_buffer(cow, buf, 0, 0, cow->len); |
db94535d | 1042 | btrfs_set_header_bytenr(cow, cow->start); |
5f39d397 | 1043 | btrfs_set_header_generation(cow, trans->transid); |
5d4f98a2 YZ |
1044 | btrfs_set_header_backref_rev(cow, BTRFS_MIXED_BACKREF_REV); |
1045 | btrfs_clear_header_flag(cow, BTRFS_HEADER_FLAG_WRITTEN | | |
1046 | BTRFS_HEADER_FLAG_RELOC); | |
1047 | if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) | |
1048 | btrfs_set_header_flag(cow, BTRFS_HEADER_FLAG_RELOC); | |
1049 | else | |
1050 | btrfs_set_header_owner(cow, root->root_key.objectid); | |
6702ed49 | 1051 | |
2b82032c YZ |
1052 | write_extent_buffer(cow, root->fs_info->fsid, |
1053 | (unsigned long)btrfs_header_fsid(cow), | |
1054 | BTRFS_FSID_SIZE); | |
1055 | ||
be1a5564 | 1056 | ret = update_ref_for_cow(trans, root, buf, cow, &last_ref); |
b68dc2a9 | 1057 | if (ret) { |
79787eaa | 1058 | btrfs_abort_transaction(trans, root, ret); |
b68dc2a9 MF |
1059 | return ret; |
1060 | } | |
1a40e23b | 1061 | |
3fd0a558 YZ |
1062 | if (root->ref_cows) |
1063 | btrfs_reloc_cow_block(trans, root, buf, cow); | |
1064 | ||
02217ed2 | 1065 | if (buf == root->node) { |
925baedd | 1066 | WARN_ON(parent && parent != buf); |
5d4f98a2 YZ |
1067 | if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID || |
1068 | btrfs_header_backref_rev(buf) < BTRFS_MIXED_BACKREF_REV) | |
1069 | parent_start = buf->start; | |
1070 | else | |
1071 | parent_start = 0; | |
925baedd | 1072 | |
5f39d397 | 1073 | extent_buffer_get(cow); |
90f8d62e | 1074 | tree_mod_log_set_root_pointer(root, cow, 1); |
240f62c8 | 1075 | rcu_assign_pointer(root->node, cow); |
925baedd | 1076 | |
f0486c68 | 1077 | btrfs_free_tree_block(trans, root, buf, parent_start, |
5581a51a | 1078 | last_ref); |
5f39d397 | 1079 | free_extent_buffer(buf); |
0b86a832 | 1080 | add_root_to_dirty_list(root); |
02217ed2 | 1081 | } else { |
5d4f98a2 YZ |
1082 | if (root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID) |
1083 | parent_start = parent->start; | |
1084 | else | |
1085 | parent_start = 0; | |
1086 | ||
1087 | WARN_ON(trans->transid != btrfs_header_generation(parent)); | |
f230475e JS |
1088 | tree_mod_log_insert_key(root->fs_info, parent, parent_slot, |
1089 | MOD_LOG_KEY_REPLACE); | |
5f39d397 | 1090 | btrfs_set_node_blockptr(parent, parent_slot, |
db94535d | 1091 | cow->start); |
74493f7a CM |
1092 | btrfs_set_node_ptr_generation(parent, parent_slot, |
1093 | trans->transid); | |
d6025579 | 1094 | btrfs_mark_buffer_dirty(parent); |
d9abbf1c | 1095 | tree_mod_log_free_eb(root->fs_info, buf); |
f0486c68 | 1096 | btrfs_free_tree_block(trans, root, buf, parent_start, |
5581a51a | 1097 | last_ref); |
02217ed2 | 1098 | } |
925baedd CM |
1099 | if (unlock_orig) |
1100 | btrfs_tree_unlock(buf); | |
3083ee2e | 1101 | free_extent_buffer_stale(buf); |
ccd467d6 | 1102 | btrfs_mark_buffer_dirty(cow); |
2c90e5d6 | 1103 | *cow_ret = cow; |
02217ed2 CM |
1104 | return 0; |
1105 | } | |
1106 | ||
5d9e75c4 JS |
1107 | /* |
1108 | * returns the logical address of the oldest predecessor of the given root. | |
1109 | * entries older than time_seq are ignored. | |
1110 | */ | |
1111 | static struct tree_mod_elem * | |
1112 | __tree_mod_log_oldest_root(struct btrfs_fs_info *fs_info, | |
30b0463a | 1113 | struct extent_buffer *eb_root, u64 time_seq) |
5d9e75c4 JS |
1114 | { |
1115 | struct tree_mod_elem *tm; | |
1116 | struct tree_mod_elem *found = NULL; | |
30b0463a | 1117 | u64 root_logical = eb_root->start; |
5d9e75c4 JS |
1118 | int looped = 0; |
1119 | ||
1120 | if (!time_seq) | |
1121 | return 0; | |
1122 | ||
1123 | /* | |
1124 | * the very last operation that's logged for a root is the replacement | |
1125 | * operation (if it is replaced at all). this has the index of the *new* | |
1126 | * root, making it the very first operation that's logged for this root. | |
1127 | */ | |
1128 | while (1) { | |
1129 | tm = tree_mod_log_search_oldest(fs_info, root_logical, | |
1130 | time_seq); | |
1131 | if (!looped && !tm) | |
1132 | return 0; | |
1133 | /* | |
28da9fb4 JS |
1134 | * if there are no tree operation for the oldest root, we simply |
1135 | * return it. this should only happen if that (old) root is at | |
1136 | * level 0. | |
5d9e75c4 | 1137 | */ |
28da9fb4 JS |
1138 | if (!tm) |
1139 | break; | |
5d9e75c4 | 1140 | |
28da9fb4 JS |
1141 | /* |
1142 | * if there's an operation that's not a root replacement, we | |
1143 | * found the oldest version of our root. normally, we'll find a | |
1144 | * MOD_LOG_KEY_REMOVE_WHILE_FREEING operation here. | |
1145 | */ | |
5d9e75c4 JS |
1146 | if (tm->op != MOD_LOG_ROOT_REPLACE) |
1147 | break; | |
1148 | ||
1149 | found = tm; | |
1150 | root_logical = tm->old_root.logical; | |
5d9e75c4 JS |
1151 | looped = 1; |
1152 | } | |
1153 | ||
a95236d9 JS |
1154 | /* if there's no old root to return, return what we found instead */ |
1155 | if (!found) | |
1156 | found = tm; | |
1157 | ||
5d9e75c4 JS |
1158 | return found; |
1159 | } | |
1160 | ||
1161 | /* | |
1162 | * tm is a pointer to the first operation to rewind within eb. then, all | |
1163 | * previous operations will be rewinded (until we reach something older than | |
1164 | * time_seq). | |
1165 | */ | |
1166 | static void | |
1167 | __tree_mod_log_rewind(struct extent_buffer *eb, u64 time_seq, | |
1168 | struct tree_mod_elem *first_tm) | |
1169 | { | |
1170 | u32 n; | |
1171 | struct rb_node *next; | |
1172 | struct tree_mod_elem *tm = first_tm; | |
1173 | unsigned long o_dst; | |
1174 | unsigned long o_src; | |
1175 | unsigned long p_size = sizeof(struct btrfs_key_ptr); | |
1176 | ||
1177 | n = btrfs_header_nritems(eb); | |
097b8a7c | 1178 | while (tm && tm->seq >= time_seq) { |
5d9e75c4 JS |
1179 | /* |
1180 | * all the operations are recorded with the operator used for | |
1181 | * the modification. as we're going backwards, we do the | |
1182 | * opposite of each operation here. | |
1183 | */ | |
1184 | switch (tm->op) { | |
1185 | case MOD_LOG_KEY_REMOVE_WHILE_FREEING: | |
1186 | BUG_ON(tm->slot < n); | |
1c697d4a | 1187 | /* Fallthrough */ |
95c80bb1 | 1188 | case MOD_LOG_KEY_REMOVE_WHILE_MOVING: |
4c3e6969 | 1189 | case MOD_LOG_KEY_REMOVE: |
5d9e75c4 JS |
1190 | btrfs_set_node_key(eb, &tm->key, tm->slot); |
1191 | btrfs_set_node_blockptr(eb, tm->slot, tm->blockptr); | |
1192 | btrfs_set_node_ptr_generation(eb, tm->slot, | |
1193 | tm->generation); | |
4c3e6969 | 1194 | n++; |
5d9e75c4 JS |
1195 | break; |
1196 | case MOD_LOG_KEY_REPLACE: | |
1197 | BUG_ON(tm->slot >= n); | |
1198 | btrfs_set_node_key(eb, &tm->key, tm->slot); | |
1199 | btrfs_set_node_blockptr(eb, tm->slot, tm->blockptr); | |
1200 | btrfs_set_node_ptr_generation(eb, tm->slot, | |
1201 | tm->generation); | |
1202 | break; | |
1203 | case MOD_LOG_KEY_ADD: | |
19956c7e | 1204 | /* if a move operation is needed it's in the log */ |
5d9e75c4 JS |
1205 | n--; |
1206 | break; | |
1207 | case MOD_LOG_MOVE_KEYS: | |
c3193108 JS |
1208 | o_dst = btrfs_node_key_ptr_offset(tm->slot); |
1209 | o_src = btrfs_node_key_ptr_offset(tm->move.dst_slot); | |
1210 | memmove_extent_buffer(eb, o_dst, o_src, | |
5d9e75c4 JS |
1211 | tm->move.nr_items * p_size); |
1212 | break; | |
1213 | case MOD_LOG_ROOT_REPLACE: | |
1214 | /* | |
1215 | * this operation is special. for roots, this must be | |
1216 | * handled explicitly before rewinding. | |
1217 | * for non-roots, this operation may exist if the node | |
1218 | * was a root: root A -> child B; then A gets empty and | |
1219 | * B is promoted to the new root. in the mod log, we'll | |
1220 | * have a root-replace operation for B, a tree block | |
1221 | * that is no root. we simply ignore that operation. | |
1222 | */ | |
1223 | break; | |
1224 | } | |
1225 | next = rb_next(&tm->node); | |
1226 | if (!next) | |
1227 | break; | |
1228 | tm = container_of(next, struct tree_mod_elem, node); | |
1229 | if (tm->index != first_tm->index) | |
1230 | break; | |
1231 | } | |
1232 | btrfs_set_header_nritems(eb, n); | |
1233 | } | |
1234 | ||
47fb091f JS |
1235 | /* |
1236 | * Called with eb read locked. If the buffer cannot be rewinded, the same buffer | |
1237 | * is returned. If rewind operations happen, a fresh buffer is returned. The | |
1238 | * returned buffer is always read-locked. If the returned buffer is not the | |
1239 | * input buffer, the lock on the input buffer is released and the input buffer | |
1240 | * is freed (its refcount is decremented). | |
1241 | */ | |
5d9e75c4 JS |
1242 | static struct extent_buffer * |
1243 | tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb, | |
1244 | u64 time_seq) | |
1245 | { | |
1246 | struct extent_buffer *eb_rewin; | |
1247 | struct tree_mod_elem *tm; | |
1248 | ||
1249 | if (!time_seq) | |
1250 | return eb; | |
1251 | ||
1252 | if (btrfs_header_level(eb) == 0) | |
1253 | return eb; | |
1254 | ||
1255 | tm = tree_mod_log_search(fs_info, eb->start, time_seq); | |
1256 | if (!tm) | |
1257 | return eb; | |
1258 | ||
1259 | if (tm->op == MOD_LOG_KEY_REMOVE_WHILE_FREEING) { | |
1260 | BUG_ON(tm->slot != 0); | |
1261 | eb_rewin = alloc_dummy_extent_buffer(eb->start, | |
1262 | fs_info->tree_root->nodesize); | |
1263 | BUG_ON(!eb_rewin); | |
1264 | btrfs_set_header_bytenr(eb_rewin, eb->start); | |
1265 | btrfs_set_header_backref_rev(eb_rewin, | |
1266 | btrfs_header_backref_rev(eb)); | |
1267 | btrfs_set_header_owner(eb_rewin, btrfs_header_owner(eb)); | |
c3193108 | 1268 | btrfs_set_header_level(eb_rewin, btrfs_header_level(eb)); |
5d9e75c4 JS |
1269 | } else { |
1270 | eb_rewin = btrfs_clone_extent_buffer(eb); | |
1271 | BUG_ON(!eb_rewin); | |
1272 | } | |
1273 | ||
1274 | extent_buffer_get(eb_rewin); | |
47fb091f | 1275 | btrfs_tree_read_unlock(eb); |
5d9e75c4 JS |
1276 | free_extent_buffer(eb); |
1277 | ||
47fb091f JS |
1278 | extent_buffer_get(eb_rewin); |
1279 | btrfs_tree_read_lock(eb_rewin); | |
5d9e75c4 | 1280 | __tree_mod_log_rewind(eb_rewin, time_seq, tm); |
57911b8b | 1281 | WARN_ON(btrfs_header_nritems(eb_rewin) > |
2a745b14 | 1282 | BTRFS_NODEPTRS_PER_BLOCK(fs_info->tree_root)); |
5d9e75c4 JS |
1283 | |
1284 | return eb_rewin; | |
1285 | } | |
1286 | ||
8ba97a15 JS |
1287 | /* |
1288 | * get_old_root() rewinds the state of @root's root node to the given @time_seq | |
1289 | * value. If there are no changes, the current root->root_node is returned. If | |
1290 | * anything changed in between, there's a fresh buffer allocated on which the | |
1291 | * rewind operations are done. In any case, the returned buffer is read locked. | |
1292 | * Returns NULL on error (with no locks held). | |
1293 | */ | |
5d9e75c4 JS |
1294 | static inline struct extent_buffer * |
1295 | get_old_root(struct btrfs_root *root, u64 time_seq) | |
1296 | { | |
1297 | struct tree_mod_elem *tm; | |
30b0463a JS |
1298 | struct extent_buffer *eb = NULL; |
1299 | struct extent_buffer *eb_root; | |
7bfdcf7f | 1300 | struct extent_buffer *old; |
a95236d9 | 1301 | struct tree_mod_root *old_root = NULL; |
4325edd0 | 1302 | u64 old_generation = 0; |
a95236d9 | 1303 | u64 logical; |
834328a8 | 1304 | u32 blocksize; |
5d9e75c4 | 1305 | |
30b0463a JS |
1306 | eb_root = btrfs_read_lock_root_node(root); |
1307 | tm = __tree_mod_log_oldest_root(root->fs_info, eb_root, time_seq); | |
5d9e75c4 | 1308 | if (!tm) |
30b0463a | 1309 | return eb_root; |
5d9e75c4 | 1310 | |
a95236d9 JS |
1311 | if (tm->op == MOD_LOG_ROOT_REPLACE) { |
1312 | old_root = &tm->old_root; | |
1313 | old_generation = tm->generation; | |
1314 | logical = old_root->logical; | |
1315 | } else { | |
30b0463a | 1316 | logical = eb_root->start; |
a95236d9 | 1317 | } |
5d9e75c4 | 1318 | |
a95236d9 | 1319 | tm = tree_mod_log_search(root->fs_info, logical, time_seq); |
834328a8 | 1320 | if (old_root && tm && tm->op != MOD_LOG_KEY_REMOVE_WHILE_FREEING) { |
30b0463a JS |
1321 | btrfs_tree_read_unlock(eb_root); |
1322 | free_extent_buffer(eb_root); | |
834328a8 | 1323 | blocksize = btrfs_level_size(root, old_root->level); |
7bfdcf7f | 1324 | old = read_tree_block(root, logical, blocksize, 0); |
416bc658 JB |
1325 | if (!old || !extent_buffer_uptodate(old)) { |
1326 | free_extent_buffer(old); | |
834328a8 JS |
1327 | pr_warn("btrfs: failed to read tree block %llu from get_old_root\n", |
1328 | logical); | |
1329 | WARN_ON(1); | |
1330 | } else { | |
7bfdcf7f LB |
1331 | eb = btrfs_clone_extent_buffer(old); |
1332 | free_extent_buffer(old); | |
834328a8 JS |
1333 | } |
1334 | } else if (old_root) { | |
30b0463a JS |
1335 | btrfs_tree_read_unlock(eb_root); |
1336 | free_extent_buffer(eb_root); | |
28da9fb4 | 1337 | eb = alloc_dummy_extent_buffer(logical, root->nodesize); |
834328a8 | 1338 | } else { |
30b0463a JS |
1339 | eb = btrfs_clone_extent_buffer(eb_root); |
1340 | btrfs_tree_read_unlock(eb_root); | |
1341 | free_extent_buffer(eb_root); | |
834328a8 JS |
1342 | } |
1343 | ||
8ba97a15 JS |
1344 | if (!eb) |
1345 | return NULL; | |
d6381084 | 1346 | extent_buffer_get(eb); |
8ba97a15 | 1347 | btrfs_tree_read_lock(eb); |
a95236d9 | 1348 | if (old_root) { |
5d9e75c4 JS |
1349 | btrfs_set_header_bytenr(eb, eb->start); |
1350 | btrfs_set_header_backref_rev(eb, BTRFS_MIXED_BACKREF_REV); | |
30b0463a | 1351 | btrfs_set_header_owner(eb, btrfs_header_owner(eb_root)); |
a95236d9 JS |
1352 | btrfs_set_header_level(eb, old_root->level); |
1353 | btrfs_set_header_generation(eb, old_generation); | |
5d9e75c4 | 1354 | } |
28da9fb4 JS |
1355 | if (tm) |
1356 | __tree_mod_log_rewind(eb, time_seq, tm); | |
1357 | else | |
1358 | WARN_ON(btrfs_header_level(eb) != 0); | |
57911b8b | 1359 | WARN_ON(btrfs_header_nritems(eb) > BTRFS_NODEPTRS_PER_BLOCK(root)); |
5d9e75c4 JS |
1360 | |
1361 | return eb; | |
1362 | } | |
1363 | ||
5b6602e7 JS |
1364 | int btrfs_old_root_level(struct btrfs_root *root, u64 time_seq) |
1365 | { | |
1366 | struct tree_mod_elem *tm; | |
1367 | int level; | |
30b0463a | 1368 | struct extent_buffer *eb_root = btrfs_root_node(root); |
5b6602e7 | 1369 | |
30b0463a | 1370 | tm = __tree_mod_log_oldest_root(root->fs_info, eb_root, time_seq); |
5b6602e7 JS |
1371 | if (tm && tm->op == MOD_LOG_ROOT_REPLACE) { |
1372 | level = tm->old_root.level; | |
1373 | } else { | |
30b0463a | 1374 | level = btrfs_header_level(eb_root); |
5b6602e7 | 1375 | } |
30b0463a | 1376 | free_extent_buffer(eb_root); |
5b6602e7 JS |
1377 | |
1378 | return level; | |
1379 | } | |
1380 | ||
5d4f98a2 YZ |
1381 | static inline int should_cow_block(struct btrfs_trans_handle *trans, |
1382 | struct btrfs_root *root, | |
1383 | struct extent_buffer *buf) | |
1384 | { | |
f1ebcc74 LB |
1385 | /* ensure we can see the force_cow */ |
1386 | smp_rmb(); | |
1387 | ||
1388 | /* | |
1389 | * We do not need to cow a block if | |
1390 | * 1) this block is not created or changed in this transaction; | |
1391 | * 2) this block does not belong to TREE_RELOC tree; | |
1392 | * 3) the root is not forced COW. | |
1393 | * | |
1394 | * What is forced COW: | |
1395 | * when we create snapshot during commiting the transaction, | |
1396 | * after we've finished coping src root, we must COW the shared | |
1397 | * block to ensure the metadata consistency. | |
1398 | */ | |
5d4f98a2 YZ |
1399 | if (btrfs_header_generation(buf) == trans->transid && |
1400 | !btrfs_header_flag(buf, BTRFS_HEADER_FLAG_WRITTEN) && | |
1401 | !(root->root_key.objectid != BTRFS_TREE_RELOC_OBJECTID && | |
f1ebcc74 LB |
1402 | btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC)) && |
1403 | !root->force_cow) | |
5d4f98a2 YZ |
1404 | return 0; |
1405 | return 1; | |
1406 | } | |
1407 | ||
d352ac68 CM |
1408 | /* |
1409 | * cows a single block, see __btrfs_cow_block for the real work. | |
1410 | * This version of it has extra checks so that a block isn't cow'd more than | |
1411 | * once per transaction, as long as it hasn't been written yet | |
1412 | */ | |
d397712b | 1413 | noinline int btrfs_cow_block(struct btrfs_trans_handle *trans, |
5f39d397 CM |
1414 | struct btrfs_root *root, struct extent_buffer *buf, |
1415 | struct extent_buffer *parent, int parent_slot, | |
9fa8cfe7 | 1416 | struct extent_buffer **cow_ret) |
6702ed49 CM |
1417 | { |
1418 | u64 search_start; | |
f510cfec | 1419 | int ret; |
dc17ff8f | 1420 | |
31b1a2bd JL |
1421 | if (trans->transaction != root->fs_info->running_transaction) |
1422 | WARN(1, KERN_CRIT "trans %llu running %llu\n", | |
d397712b CM |
1423 | (unsigned long long)trans->transid, |
1424 | (unsigned long long) | |
6702ed49 | 1425 | root->fs_info->running_transaction->transid); |
31b1a2bd JL |
1426 | |
1427 | if (trans->transid != root->fs_info->generation) | |
1428 | WARN(1, KERN_CRIT "trans %llu running %llu\n", | |
d397712b CM |
1429 | (unsigned long long)trans->transid, |
1430 | (unsigned long long)root->fs_info->generation); | |
dc17ff8f | 1431 | |
5d4f98a2 | 1432 | if (!should_cow_block(trans, root, buf)) { |
6702ed49 CM |
1433 | *cow_ret = buf; |
1434 | return 0; | |
1435 | } | |
c487685d | 1436 | |
0b86a832 | 1437 | search_start = buf->start & ~((u64)(1024 * 1024 * 1024) - 1); |
b4ce94de CM |
1438 | |
1439 | if (parent) | |
1440 | btrfs_set_lock_blocking(parent); | |
1441 | btrfs_set_lock_blocking(buf); | |
1442 | ||
f510cfec | 1443 | ret = __btrfs_cow_block(trans, root, buf, parent, |
9fa8cfe7 | 1444 | parent_slot, cow_ret, search_start, 0); |
1abe9b8a | 1445 | |
1446 | trace_btrfs_cow_block(root, buf, *cow_ret); | |
1447 | ||
f510cfec | 1448 | return ret; |
6702ed49 CM |
1449 | } |
1450 | ||
d352ac68 CM |
1451 | /* |
1452 | * helper function for defrag to decide if two blocks pointed to by a | |
1453 | * node are actually close by | |
1454 | */ | |
6b80053d | 1455 | static int close_blocks(u64 blocknr, u64 other, u32 blocksize) |
6702ed49 | 1456 | { |
6b80053d | 1457 | if (blocknr < other && other - (blocknr + blocksize) < 32768) |
6702ed49 | 1458 | return 1; |
6b80053d | 1459 | if (blocknr > other && blocknr - (other + blocksize) < 32768) |
6702ed49 CM |
1460 | return 1; |
1461 | return 0; | |
1462 | } | |
1463 | ||
081e9573 CM |
1464 | /* |
1465 | * compare two keys in a memcmp fashion | |
1466 | */ | |
1467 | static int comp_keys(struct btrfs_disk_key *disk, struct btrfs_key *k2) | |
1468 | { | |
1469 | struct btrfs_key k1; | |
1470 | ||
1471 | btrfs_disk_key_to_cpu(&k1, disk); | |
1472 | ||
20736aba | 1473 | return btrfs_comp_cpu_keys(&k1, k2); |
081e9573 CM |
1474 | } |
1475 | ||
f3465ca4 JB |
1476 | /* |
1477 | * same as comp_keys only with two btrfs_key's | |
1478 | */ | |
5d4f98a2 | 1479 | int btrfs_comp_cpu_keys(struct btrfs_key *k1, struct btrfs_key *k2) |
f3465ca4 JB |
1480 | { |
1481 | if (k1->objectid > k2->objectid) | |
1482 | return 1; | |
1483 | if (k1->objectid < k2->objectid) | |
1484 | return -1; | |
1485 | if (k1->type > k2->type) | |
1486 | return 1; | |
1487 | if (k1->type < k2->type) | |
1488 | return -1; | |
1489 | if (k1->offset > k2->offset) | |
1490 | return 1; | |
1491 | if (k1->offset < k2->offset) | |
1492 | return -1; | |
1493 | return 0; | |
1494 | } | |
081e9573 | 1495 | |
d352ac68 CM |
1496 | /* |
1497 | * this is used by the defrag code to go through all the | |
1498 | * leaves pointed to by a node and reallocate them so that | |
1499 | * disk order is close to key order | |
1500 | */ | |
6702ed49 | 1501 | int btrfs_realloc_node(struct btrfs_trans_handle *trans, |
5f39d397 | 1502 | struct btrfs_root *root, struct extent_buffer *parent, |
de78b51a | 1503 | int start_slot, u64 *last_ret, |
a6b6e75e | 1504 | struct btrfs_key *progress) |
6702ed49 | 1505 | { |
6b80053d | 1506 | struct extent_buffer *cur; |
6702ed49 | 1507 | u64 blocknr; |
ca7a79ad | 1508 | u64 gen; |
e9d0b13b CM |
1509 | u64 search_start = *last_ret; |
1510 | u64 last_block = 0; | |
6702ed49 CM |
1511 | u64 other; |
1512 | u32 parent_nritems; | |
6702ed49 CM |
1513 | int end_slot; |
1514 | int i; | |
1515 | int err = 0; | |
f2183bde | 1516 | int parent_level; |
6b80053d CM |
1517 | int uptodate; |
1518 | u32 blocksize; | |
081e9573 CM |
1519 | int progress_passed = 0; |
1520 | struct btrfs_disk_key disk_key; | |
6702ed49 | 1521 | |
5708b959 | 1522 | parent_level = btrfs_header_level(parent); |
5708b959 | 1523 | |
6c1500f2 JL |
1524 | WARN_ON(trans->transaction != root->fs_info->running_transaction); |
1525 | WARN_ON(trans->transid != root->fs_info->generation); | |
86479a04 | 1526 | |
6b80053d | 1527 | parent_nritems = btrfs_header_nritems(parent); |
6b80053d | 1528 | blocksize = btrfs_level_size(root, parent_level - 1); |
6702ed49 CM |
1529 | end_slot = parent_nritems; |
1530 | ||
1531 | if (parent_nritems == 1) | |
1532 | return 0; | |
1533 | ||
b4ce94de CM |
1534 | btrfs_set_lock_blocking(parent); |
1535 | ||
6702ed49 CM |
1536 | for (i = start_slot; i < end_slot; i++) { |
1537 | int close = 1; | |
a6b6e75e | 1538 | |
081e9573 CM |
1539 | btrfs_node_key(parent, &disk_key, i); |
1540 | if (!progress_passed && comp_keys(&disk_key, progress) < 0) | |
1541 | continue; | |
1542 | ||
1543 | progress_passed = 1; | |
6b80053d | 1544 | blocknr = btrfs_node_blockptr(parent, i); |
ca7a79ad | 1545 | gen = btrfs_node_ptr_generation(parent, i); |
e9d0b13b CM |
1546 | if (last_block == 0) |
1547 | last_block = blocknr; | |
5708b959 | 1548 | |
6702ed49 | 1549 | if (i > 0) { |
6b80053d CM |
1550 | other = btrfs_node_blockptr(parent, i - 1); |
1551 | close = close_blocks(blocknr, other, blocksize); | |
6702ed49 | 1552 | } |
0ef3e66b | 1553 | if (!close && i < end_slot - 2) { |
6b80053d CM |
1554 | other = btrfs_node_blockptr(parent, i + 1); |
1555 | close = close_blocks(blocknr, other, blocksize); | |
6702ed49 | 1556 | } |
e9d0b13b CM |
1557 | if (close) { |
1558 | last_block = blocknr; | |
6702ed49 | 1559 | continue; |
e9d0b13b | 1560 | } |
6702ed49 | 1561 | |
6b80053d CM |
1562 | cur = btrfs_find_tree_block(root, blocknr, blocksize); |
1563 | if (cur) | |
b9fab919 | 1564 | uptodate = btrfs_buffer_uptodate(cur, gen, 0); |
6b80053d CM |
1565 | else |
1566 | uptodate = 0; | |
5708b959 | 1567 | if (!cur || !uptodate) { |
6b80053d CM |
1568 | if (!cur) { |
1569 | cur = read_tree_block(root, blocknr, | |
ca7a79ad | 1570 | blocksize, gen); |
416bc658 JB |
1571 | if (!cur || !extent_buffer_uptodate(cur)) { |
1572 | free_extent_buffer(cur); | |
97d9a8a4 | 1573 | return -EIO; |
416bc658 | 1574 | } |
6b80053d | 1575 | } else if (!uptodate) { |
018642a1 TI |
1576 | err = btrfs_read_buffer(cur, gen); |
1577 | if (err) { | |
1578 | free_extent_buffer(cur); | |
1579 | return err; | |
1580 | } | |
f2183bde | 1581 | } |
6702ed49 | 1582 | } |
e9d0b13b | 1583 | if (search_start == 0) |
6b80053d | 1584 | search_start = last_block; |
e9d0b13b | 1585 | |
e7a84565 | 1586 | btrfs_tree_lock(cur); |
b4ce94de | 1587 | btrfs_set_lock_blocking(cur); |
6b80053d | 1588 | err = __btrfs_cow_block(trans, root, cur, parent, i, |
e7a84565 | 1589 | &cur, search_start, |
6b80053d | 1590 | min(16 * blocksize, |
9fa8cfe7 | 1591 | (end_slot - i) * blocksize)); |
252c38f0 | 1592 | if (err) { |
e7a84565 | 1593 | btrfs_tree_unlock(cur); |
6b80053d | 1594 | free_extent_buffer(cur); |
6702ed49 | 1595 | break; |
252c38f0 | 1596 | } |
e7a84565 CM |
1597 | search_start = cur->start; |
1598 | last_block = cur->start; | |
f2183bde | 1599 | *last_ret = search_start; |
e7a84565 CM |
1600 | btrfs_tree_unlock(cur); |
1601 | free_extent_buffer(cur); | |
6702ed49 CM |
1602 | } |
1603 | return err; | |
1604 | } | |
1605 | ||
74123bd7 CM |
1606 | /* |
1607 | * The leaf data grows from end-to-front in the node. | |
1608 | * this returns the address of the start of the last item, | |
1609 | * which is the stop of the leaf data stack | |
1610 | */ | |
123abc88 | 1611 | static inline unsigned int leaf_data_end(struct btrfs_root *root, |
5f39d397 | 1612 | struct extent_buffer *leaf) |
be0e5c09 | 1613 | { |
5f39d397 | 1614 | u32 nr = btrfs_header_nritems(leaf); |
be0e5c09 | 1615 | if (nr == 0) |
123abc88 | 1616 | return BTRFS_LEAF_DATA_SIZE(root); |
5f39d397 | 1617 | return btrfs_item_offset_nr(leaf, nr - 1); |
be0e5c09 CM |
1618 | } |
1619 | ||
aa5d6bed | 1620 | |
74123bd7 | 1621 | /* |
5f39d397 CM |
1622 | * search for key in the extent_buffer. The items start at offset p, |
1623 | * and they are item_size apart. There are 'max' items in p. | |
1624 | * | |
74123bd7 CM |
1625 | * the slot in the array is returned via slot, and it points to |
1626 | * the place where you would insert key if it is not found in | |
1627 | * the array. | |
1628 | * | |
1629 | * slot may point to max if the key is bigger than all of the keys | |
1630 | */ | |
e02119d5 CM |
1631 | static noinline int generic_bin_search(struct extent_buffer *eb, |
1632 | unsigned long p, | |
1633 | int item_size, struct btrfs_key *key, | |
1634 | int max, int *slot) | |
be0e5c09 CM |
1635 | { |
1636 | int low = 0; | |
1637 | int high = max; | |
1638 | int mid; | |
1639 | int ret; | |
479965d6 | 1640 | struct btrfs_disk_key *tmp = NULL; |
5f39d397 CM |
1641 | struct btrfs_disk_key unaligned; |
1642 | unsigned long offset; | |
5f39d397 CM |
1643 | char *kaddr = NULL; |
1644 | unsigned long map_start = 0; | |
1645 | unsigned long map_len = 0; | |
479965d6 | 1646 | int err; |
be0e5c09 | 1647 | |
d397712b | 1648 | while (low < high) { |
be0e5c09 | 1649 | mid = (low + high) / 2; |
5f39d397 CM |
1650 | offset = p + mid * item_size; |
1651 | ||
a6591715 | 1652 | if (!kaddr || offset < map_start || |
5f39d397 CM |
1653 | (offset + sizeof(struct btrfs_disk_key)) > |
1654 | map_start + map_len) { | |
934d375b CM |
1655 | |
1656 | err = map_private_extent_buffer(eb, offset, | |
479965d6 | 1657 | sizeof(struct btrfs_disk_key), |
a6591715 | 1658 | &kaddr, &map_start, &map_len); |
479965d6 CM |
1659 | |
1660 | if (!err) { | |
1661 | tmp = (struct btrfs_disk_key *)(kaddr + offset - | |
1662 | map_start); | |
1663 | } else { | |
1664 | read_extent_buffer(eb, &unaligned, | |
1665 | offset, sizeof(unaligned)); | |
1666 | tmp = &unaligned; | |
1667 | } | |
5f39d397 | 1668 | |
5f39d397 CM |
1669 | } else { |
1670 | tmp = (struct btrfs_disk_key *)(kaddr + offset - | |
1671 | map_start); | |
1672 | } | |
be0e5c09 CM |
1673 | ret = comp_keys(tmp, key); |
1674 | ||
1675 | if (ret < 0) | |
1676 | low = mid + 1; | |
1677 | else if (ret > 0) | |
1678 | high = mid; | |
1679 | else { | |
1680 | *slot = mid; | |
1681 | return 0; | |
1682 | } | |
1683 | } | |
1684 | *slot = low; | |
1685 | return 1; | |
1686 | } | |
1687 | ||
97571fd0 CM |
1688 | /* |
1689 | * simple bin_search frontend that does the right thing for | |
1690 | * leaves vs nodes | |
1691 | */ | |
5f39d397 CM |
1692 | static int bin_search(struct extent_buffer *eb, struct btrfs_key *key, |
1693 | int level, int *slot) | |
be0e5c09 | 1694 | { |
f775738f | 1695 | if (level == 0) |
5f39d397 CM |
1696 | return generic_bin_search(eb, |
1697 | offsetof(struct btrfs_leaf, items), | |
0783fcfc | 1698 | sizeof(struct btrfs_item), |
5f39d397 | 1699 | key, btrfs_header_nritems(eb), |
7518a238 | 1700 | slot); |
f775738f | 1701 | else |
5f39d397 CM |
1702 | return generic_bin_search(eb, |
1703 | offsetof(struct btrfs_node, ptrs), | |
123abc88 | 1704 | sizeof(struct btrfs_key_ptr), |
5f39d397 | 1705 | key, btrfs_header_nritems(eb), |
7518a238 | 1706 | slot); |
be0e5c09 CM |
1707 | } |
1708 | ||
5d4f98a2 YZ |
1709 | int btrfs_bin_search(struct extent_buffer *eb, struct btrfs_key *key, |
1710 | int level, int *slot) | |
1711 | { | |
1712 | return bin_search(eb, key, level, slot); | |
1713 | } | |
1714 | ||
f0486c68 YZ |
1715 | static void root_add_used(struct btrfs_root *root, u32 size) |
1716 | { | |
1717 | spin_lock(&root->accounting_lock); | |
1718 | btrfs_set_root_used(&root->root_item, | |
1719 | btrfs_root_used(&root->root_item) + size); | |
1720 | spin_unlock(&root->accounting_lock); | |
1721 | } | |
1722 | ||
1723 | static void root_sub_used(struct btrfs_root *root, u32 size) | |
1724 | { | |
1725 | spin_lock(&root->accounting_lock); | |
1726 | btrfs_set_root_used(&root->root_item, | |
1727 | btrfs_root_used(&root->root_item) - size); | |
1728 | spin_unlock(&root->accounting_lock); | |
1729 | } | |
1730 | ||
d352ac68 CM |
1731 | /* given a node and slot number, this reads the blocks it points to. The |
1732 | * extent buffer is returned with a reference taken (but unlocked). | |
1733 | * NULL is returned on error. | |
1734 | */ | |
e02119d5 | 1735 | static noinline struct extent_buffer *read_node_slot(struct btrfs_root *root, |
5f39d397 | 1736 | struct extent_buffer *parent, int slot) |
bb803951 | 1737 | { |
ca7a79ad | 1738 | int level = btrfs_header_level(parent); |
416bc658 JB |
1739 | struct extent_buffer *eb; |
1740 | ||
bb803951 CM |
1741 | if (slot < 0) |
1742 | return NULL; | |
5f39d397 | 1743 | if (slot >= btrfs_header_nritems(parent)) |
bb803951 | 1744 | return NULL; |
ca7a79ad CM |
1745 | |
1746 | BUG_ON(level == 0); | |
1747 | ||
416bc658 JB |
1748 | eb = read_tree_block(root, btrfs_node_blockptr(parent, slot), |
1749 | btrfs_level_size(root, level - 1), | |
1750 | btrfs_node_ptr_generation(parent, slot)); | |
1751 | if (eb && !extent_buffer_uptodate(eb)) { | |
1752 | free_extent_buffer(eb); | |
1753 | eb = NULL; | |
1754 | } | |
1755 | ||
1756 | return eb; | |
bb803951 CM |
1757 | } |
1758 | ||
d352ac68 CM |
1759 | /* |
1760 | * node level balancing, used to make sure nodes are in proper order for | |
1761 | * item deletion. We balance from the top down, so we have to make sure | |
1762 | * that a deletion won't leave an node completely empty later on. | |
1763 | */ | |
e02119d5 | 1764 | static noinline int balance_level(struct btrfs_trans_handle *trans, |
98ed5174 CM |
1765 | struct btrfs_root *root, |
1766 | struct btrfs_path *path, int level) | |
bb803951 | 1767 | { |
5f39d397 CM |
1768 | struct extent_buffer *right = NULL; |
1769 | struct extent_buffer *mid; | |
1770 | struct extent_buffer *left = NULL; | |
1771 | struct extent_buffer *parent = NULL; | |
bb803951 CM |
1772 | int ret = 0; |
1773 | int wret; | |
1774 | int pslot; | |
bb803951 | 1775 | int orig_slot = path->slots[level]; |
79f95c82 | 1776 | u64 orig_ptr; |
bb803951 CM |
1777 | |
1778 | if (level == 0) | |
1779 | return 0; | |
1780 | ||
5f39d397 | 1781 | mid = path->nodes[level]; |
b4ce94de | 1782 | |
bd681513 CM |
1783 | WARN_ON(path->locks[level] != BTRFS_WRITE_LOCK && |
1784 | path->locks[level] != BTRFS_WRITE_LOCK_BLOCKING); | |
7bb86316 CM |
1785 | WARN_ON(btrfs_header_generation(mid) != trans->transid); |
1786 | ||
1d4f8a0c | 1787 | orig_ptr = btrfs_node_blockptr(mid, orig_slot); |
79f95c82 | 1788 | |
a05a9bb1 | 1789 | if (level < BTRFS_MAX_LEVEL - 1) { |
5f39d397 | 1790 | parent = path->nodes[level + 1]; |
a05a9bb1 LZ |
1791 | pslot = path->slots[level + 1]; |
1792 | } | |
bb803951 | 1793 | |
40689478 CM |
1794 | /* |
1795 | * deal with the case where there is only one pointer in the root | |
1796 | * by promoting the node below to a root | |
1797 | */ | |
5f39d397 CM |
1798 | if (!parent) { |
1799 | struct extent_buffer *child; | |
bb803951 | 1800 | |
5f39d397 | 1801 | if (btrfs_header_nritems(mid) != 1) |
bb803951 CM |
1802 | return 0; |
1803 | ||
1804 | /* promote the child to a root */ | |
5f39d397 | 1805 | child = read_node_slot(root, mid, 0); |
305a26af MF |
1806 | if (!child) { |
1807 | ret = -EROFS; | |
1808 | btrfs_std_error(root->fs_info, ret); | |
1809 | goto enospc; | |
1810 | } | |
1811 | ||
925baedd | 1812 | btrfs_tree_lock(child); |
b4ce94de | 1813 | btrfs_set_lock_blocking(child); |
9fa8cfe7 | 1814 | ret = btrfs_cow_block(trans, root, child, mid, 0, &child); |
f0486c68 YZ |
1815 | if (ret) { |
1816 | btrfs_tree_unlock(child); | |
1817 | free_extent_buffer(child); | |
1818 | goto enospc; | |
1819 | } | |
2f375ab9 | 1820 | |
90f8d62e | 1821 | tree_mod_log_set_root_pointer(root, child, 1); |
240f62c8 | 1822 | rcu_assign_pointer(root->node, child); |
925baedd | 1823 | |
0b86a832 | 1824 | add_root_to_dirty_list(root); |
925baedd | 1825 | btrfs_tree_unlock(child); |
b4ce94de | 1826 | |
925baedd | 1827 | path->locks[level] = 0; |
bb803951 | 1828 | path->nodes[level] = NULL; |
5f39d397 | 1829 | clean_tree_block(trans, root, mid); |
925baedd | 1830 | btrfs_tree_unlock(mid); |
bb803951 | 1831 | /* once for the path */ |
5f39d397 | 1832 | free_extent_buffer(mid); |
f0486c68 YZ |
1833 | |
1834 | root_sub_used(root, mid->len); | |
5581a51a | 1835 | btrfs_free_tree_block(trans, root, mid, 0, 1); |
bb803951 | 1836 | /* once for the root ptr */ |
3083ee2e | 1837 | free_extent_buffer_stale(mid); |
f0486c68 | 1838 | return 0; |
bb803951 | 1839 | } |
5f39d397 | 1840 | if (btrfs_header_nritems(mid) > |
123abc88 | 1841 | BTRFS_NODEPTRS_PER_BLOCK(root) / 4) |
bb803951 CM |
1842 | return 0; |
1843 | ||
5f39d397 CM |
1844 | left = read_node_slot(root, parent, pslot - 1); |
1845 | if (left) { | |
925baedd | 1846 | btrfs_tree_lock(left); |
b4ce94de | 1847 | btrfs_set_lock_blocking(left); |
5f39d397 | 1848 | wret = btrfs_cow_block(trans, root, left, |
9fa8cfe7 | 1849 | parent, pslot - 1, &left); |
54aa1f4d CM |
1850 | if (wret) { |
1851 | ret = wret; | |
1852 | goto enospc; | |
1853 | } | |
2cc58cf2 | 1854 | } |
5f39d397 CM |
1855 | right = read_node_slot(root, parent, pslot + 1); |
1856 | if (right) { | |
925baedd | 1857 | btrfs_tree_lock(right); |
b4ce94de | 1858 | btrfs_set_lock_blocking(right); |
5f39d397 | 1859 | wret = btrfs_cow_block(trans, root, right, |
9fa8cfe7 | 1860 | parent, pslot + 1, &right); |
2cc58cf2 CM |
1861 | if (wret) { |
1862 | ret = wret; | |
1863 | goto enospc; | |
1864 | } | |
1865 | } | |
1866 | ||
1867 | /* first, try to make some room in the middle buffer */ | |
5f39d397 CM |
1868 | if (left) { |
1869 | orig_slot += btrfs_header_nritems(left); | |
bce4eae9 | 1870 | wret = push_node_left(trans, root, left, mid, 1); |
79f95c82 CM |
1871 | if (wret < 0) |
1872 | ret = wret; | |
bb803951 | 1873 | } |
79f95c82 CM |
1874 | |
1875 | /* | |
1876 | * then try to empty the right most buffer into the middle | |
1877 | */ | |
5f39d397 | 1878 | if (right) { |
971a1f66 | 1879 | wret = push_node_left(trans, root, mid, right, 1); |
54aa1f4d | 1880 | if (wret < 0 && wret != -ENOSPC) |
79f95c82 | 1881 | ret = wret; |
5f39d397 | 1882 | if (btrfs_header_nritems(right) == 0) { |
5f39d397 | 1883 | clean_tree_block(trans, root, right); |
925baedd | 1884 | btrfs_tree_unlock(right); |
afe5fea7 | 1885 | del_ptr(root, path, level + 1, pslot + 1); |
f0486c68 | 1886 | root_sub_used(root, right->len); |
5581a51a | 1887 | btrfs_free_tree_block(trans, root, right, 0, 1); |
3083ee2e | 1888 | free_extent_buffer_stale(right); |
f0486c68 | 1889 | right = NULL; |
bb803951 | 1890 | } else { |
5f39d397 CM |
1891 | struct btrfs_disk_key right_key; |
1892 | btrfs_node_key(right, &right_key, 0); | |
f230475e | 1893 | tree_mod_log_set_node_key(root->fs_info, parent, |
32adf090 | 1894 | pslot + 1, 0); |
5f39d397 CM |
1895 | btrfs_set_node_key(parent, &right_key, pslot + 1); |
1896 | btrfs_mark_buffer_dirty(parent); | |
bb803951 CM |
1897 | } |
1898 | } | |
5f39d397 | 1899 | if (btrfs_header_nritems(mid) == 1) { |
79f95c82 CM |
1900 | /* |
1901 | * we're not allowed to leave a node with one item in the | |
1902 | * tree during a delete. A deletion from lower in the tree | |
1903 | * could try to delete the only pointer in this node. | |
1904 | * So, pull some keys from the left. | |
1905 | * There has to be a left pointer at this point because | |
1906 | * otherwise we would have pulled some pointers from the | |
1907 | * right | |
1908 | */ | |
305a26af MF |
1909 | if (!left) { |
1910 | ret = -EROFS; | |
1911 | btrfs_std_error(root->fs_info, ret); | |
1912 | goto enospc; | |
1913 | } | |
5f39d397 | 1914 | wret = balance_node_right(trans, root, mid, left); |
54aa1f4d | 1915 | if (wret < 0) { |
79f95c82 | 1916 | ret = wret; |
54aa1f4d CM |
1917 | goto enospc; |
1918 | } | |
bce4eae9 CM |
1919 | if (wret == 1) { |
1920 | wret = push_node_left(trans, root, left, mid, 1); | |
1921 | if (wret < 0) | |
1922 | ret = wret; | |
1923 | } | |
79f95c82 CM |
1924 | BUG_ON(wret == 1); |
1925 | } | |
5f39d397 | 1926 | if (btrfs_header_nritems(mid) == 0) { |
5f39d397 | 1927 | clean_tree_block(trans, root, mid); |
925baedd | 1928 | btrfs_tree_unlock(mid); |
afe5fea7 | 1929 | del_ptr(root, path, level + 1, pslot); |
f0486c68 | 1930 | root_sub_used(root, mid->len); |
5581a51a | 1931 | btrfs_free_tree_block(trans, root, mid, 0, 1); |
3083ee2e | 1932 | free_extent_buffer_stale(mid); |
f0486c68 | 1933 | mid = NULL; |
79f95c82 CM |
1934 | } else { |
1935 | /* update the parent key to reflect our changes */ | |
5f39d397 CM |
1936 | struct btrfs_disk_key mid_key; |
1937 | btrfs_node_key(mid, &mid_key, 0); | |
32adf090 | 1938 | tree_mod_log_set_node_key(root->fs_info, parent, |
f230475e | 1939 | pslot, 0); |
5f39d397 CM |
1940 | btrfs_set_node_key(parent, &mid_key, pslot); |
1941 | btrfs_mark_buffer_dirty(parent); | |
79f95c82 | 1942 | } |
bb803951 | 1943 | |
79f95c82 | 1944 | /* update the path */ |
5f39d397 CM |
1945 | if (left) { |
1946 | if (btrfs_header_nritems(left) > orig_slot) { | |
1947 | extent_buffer_get(left); | |
925baedd | 1948 | /* left was locked after cow */ |
5f39d397 | 1949 | path->nodes[level] = left; |
bb803951 CM |
1950 | path->slots[level + 1] -= 1; |
1951 | path->slots[level] = orig_slot; | |
925baedd CM |
1952 | if (mid) { |
1953 | btrfs_tree_unlock(mid); | |
5f39d397 | 1954 | free_extent_buffer(mid); |
925baedd | 1955 | } |
bb803951 | 1956 | } else { |
5f39d397 | 1957 | orig_slot -= btrfs_header_nritems(left); |
bb803951 CM |
1958 | path->slots[level] = orig_slot; |
1959 | } | |
1960 | } | |
79f95c82 | 1961 | /* double check we haven't messed things up */ |
e20d96d6 | 1962 | if (orig_ptr != |
5f39d397 | 1963 | btrfs_node_blockptr(path->nodes[level], path->slots[level])) |
79f95c82 | 1964 | BUG(); |
54aa1f4d | 1965 | enospc: |
925baedd CM |
1966 | if (right) { |
1967 | btrfs_tree_unlock(right); | |
5f39d397 | 1968 | free_extent_buffer(right); |
925baedd CM |
1969 | } |
1970 | if (left) { | |
1971 | if (path->nodes[level] != left) | |
1972 | btrfs_tree_unlock(left); | |
5f39d397 | 1973 | free_extent_buffer(left); |
925baedd | 1974 | } |
bb803951 CM |
1975 | return ret; |
1976 | } | |
1977 | ||
d352ac68 CM |
1978 | /* Node balancing for insertion. Here we only split or push nodes around |
1979 | * when they are completely full. This is also done top down, so we | |
1980 | * have to be pessimistic. | |
1981 | */ | |
d397712b | 1982 | static noinline int push_nodes_for_insert(struct btrfs_trans_handle *trans, |
98ed5174 CM |
1983 | struct btrfs_root *root, |
1984 | struct btrfs_path *path, int level) | |
e66f709b | 1985 | { |
5f39d397 CM |
1986 | struct extent_buffer *right = NULL; |
1987 | struct extent_buffer *mid; | |
1988 | struct extent_buffer *left = NULL; | |
1989 | struct extent_buffer *parent = NULL; | |
e66f709b CM |
1990 | int ret = 0; |
1991 | int wret; | |
1992 | int pslot; | |
1993 | int orig_slot = path->slots[level]; | |
e66f709b CM |
1994 | |
1995 | if (level == 0) | |
1996 | return 1; | |
1997 | ||
5f39d397 | 1998 | mid = path->nodes[level]; |
7bb86316 | 1999 | WARN_ON(btrfs_header_generation(mid) != trans->transid); |
e66f709b | 2000 | |
a05a9bb1 | 2001 | if (level < BTRFS_MAX_LEVEL - 1) { |
5f39d397 | 2002 | parent = path->nodes[level + 1]; |
a05a9bb1 LZ |
2003 | pslot = path->slots[level + 1]; |
2004 | } | |
e66f709b | 2005 | |
5f39d397 | 2006 | if (!parent) |
e66f709b | 2007 | return 1; |
e66f709b | 2008 | |
5f39d397 | 2009 | left = read_node_slot(root, parent, pslot - 1); |
e66f709b CM |
2010 | |
2011 | /* first, try to make some room in the middle buffer */ | |
5f39d397 | 2012 | if (left) { |
e66f709b | 2013 | u32 left_nr; |
925baedd CM |
2014 | |
2015 | btrfs_tree_lock(left); | |
b4ce94de CM |
2016 | btrfs_set_lock_blocking(left); |
2017 | ||
5f39d397 | 2018 | left_nr = btrfs_header_nritems(left); |
33ade1f8 CM |
2019 | if (left_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) { |
2020 | wret = 1; | |
2021 | } else { | |
5f39d397 | 2022 | ret = btrfs_cow_block(trans, root, left, parent, |
9fa8cfe7 | 2023 | pslot - 1, &left); |
54aa1f4d CM |
2024 | if (ret) |
2025 | wret = 1; | |
2026 | else { | |
54aa1f4d | 2027 | wret = push_node_left(trans, root, |
971a1f66 | 2028 | left, mid, 0); |
54aa1f4d | 2029 | } |
33ade1f8 | 2030 | } |
e66f709b CM |
2031 | if (wret < 0) |
2032 | ret = wret; | |
2033 | if (wret == 0) { | |
5f39d397 | 2034 | struct btrfs_disk_key disk_key; |
e66f709b | 2035 | orig_slot += left_nr; |
5f39d397 | 2036 | btrfs_node_key(mid, &disk_key, 0); |
f230475e | 2037 | tree_mod_log_set_node_key(root->fs_info, parent, |
32adf090 | 2038 | pslot, 0); |
5f39d397 CM |
2039 | btrfs_set_node_key(parent, &disk_key, pslot); |
2040 | btrfs_mark_buffer_dirty(parent); | |
2041 | if (btrfs_header_nritems(left) > orig_slot) { | |
2042 | path->nodes[level] = left; | |
e66f709b CM |
2043 | path->slots[level + 1] -= 1; |
2044 | path->slots[level] = orig_slot; | |
925baedd | 2045 | btrfs_tree_unlock(mid); |
5f39d397 | 2046 | free_extent_buffer(mid); |
e66f709b CM |
2047 | } else { |
2048 | orig_slot -= | |
5f39d397 | 2049 | btrfs_header_nritems(left); |
e66f709b | 2050 | path->slots[level] = orig_slot; |
925baedd | 2051 | btrfs_tree_unlock(left); |
5f39d397 | 2052 | free_extent_buffer(left); |
e66f709b | 2053 | } |
e66f709b CM |
2054 | return 0; |
2055 | } | |
925baedd | 2056 | btrfs_tree_unlock(left); |
5f39d397 | 2057 | free_extent_buffer(left); |
e66f709b | 2058 | } |
925baedd | 2059 | right = read_node_slot(root, parent, pslot + 1); |
e66f709b CM |
2060 | |
2061 | /* | |
2062 | * then try to empty the right most buffer into the middle | |
2063 | */ | |
5f39d397 | 2064 | if (right) { |
33ade1f8 | 2065 | u32 right_nr; |
b4ce94de | 2066 | |
925baedd | 2067 | btrfs_tree_lock(right); |
b4ce94de CM |
2068 | btrfs_set_lock_blocking(right); |
2069 | ||
5f39d397 | 2070 | right_nr = btrfs_header_nritems(right); |
33ade1f8 CM |
2071 | if (right_nr >= BTRFS_NODEPTRS_PER_BLOCK(root) - 1) { |
2072 | wret = 1; | |
2073 | } else { | |
5f39d397 CM |
2074 | ret = btrfs_cow_block(trans, root, right, |
2075 | parent, pslot + 1, | |
9fa8cfe7 | 2076 | &right); |
54aa1f4d CM |
2077 | if (ret) |
2078 | wret = 1; | |
2079 | else { | |
54aa1f4d | 2080 | wret = balance_node_right(trans, root, |
5f39d397 | 2081 | right, mid); |
54aa1f4d | 2082 | } |
33ade1f8 | 2083 | } |
e66f709b CM |
2084 | if (wret < 0) |
2085 | ret = wret; | |
2086 | if (wret == 0) { | |
5f39d397 CM |
2087 | struct btrfs_disk_key disk_key; |
2088 | ||
2089 | btrfs_node_key(right, &disk_key, 0); | |
f230475e | 2090 | tree_mod_log_set_node_key(root->fs_info, parent, |
32adf090 | 2091 | pslot + 1, 0); |
5f39d397 CM |
2092 | btrfs_set_node_key(parent, &disk_key, pslot + 1); |
2093 | btrfs_mark_buffer_dirty(parent); | |
2094 | ||
2095 | if (btrfs_header_nritems(mid) <= orig_slot) { | |
2096 | path->nodes[level] = right; | |
e66f709b CM |
2097 | path->slots[level + 1] += 1; |
2098 | path->slots[level] = orig_slot - | |
5f39d397 | 2099 | btrfs_header_nritems(mid); |
925baedd | 2100 | btrfs_tree_unlock(mid); |
5f39d397 | 2101 | free_extent_buffer(mid); |
e66f709b | 2102 | } else { |
925baedd | 2103 | btrfs_tree_unlock(right); |
5f39d397 | 2104 | free_extent_buffer(right); |
e66f709b | 2105 | } |
e66f709b CM |
2106 | return 0; |
2107 | } | |
925baedd | 2108 | btrfs_tree_unlock(right); |
5f39d397 | 2109 | free_extent_buffer(right); |
e66f709b | 2110 | } |
e66f709b CM |
2111 | return 1; |
2112 | } | |
2113 | ||
3c69faec | 2114 | /* |
d352ac68 CM |
2115 | * readahead one full node of leaves, finding things that are close |
2116 | * to the block in 'slot', and triggering ra on them. | |
3c69faec | 2117 | */ |
c8c42864 CM |
2118 | static void reada_for_search(struct btrfs_root *root, |
2119 | struct btrfs_path *path, | |
2120 | int level, int slot, u64 objectid) | |
3c69faec | 2121 | { |
5f39d397 | 2122 | struct extent_buffer *node; |
01f46658 | 2123 | struct btrfs_disk_key disk_key; |
3c69faec | 2124 | u32 nritems; |
3c69faec | 2125 | u64 search; |
a7175319 | 2126 | u64 target; |
6b80053d | 2127 | u64 nread = 0; |
cb25c2ea | 2128 | u64 gen; |
3c69faec | 2129 | int direction = path->reada; |
5f39d397 | 2130 | struct extent_buffer *eb; |
6b80053d CM |
2131 | u32 nr; |
2132 | u32 blocksize; | |
2133 | u32 nscan = 0; | |
db94535d | 2134 | |
a6b6e75e | 2135 | if (level != 1) |
6702ed49 CM |
2136 | return; |
2137 | ||
2138 | if (!path->nodes[level]) | |
3c69faec CM |
2139 | return; |
2140 | ||
5f39d397 | 2141 | node = path->nodes[level]; |
925baedd | 2142 | |
3c69faec | 2143 | search = btrfs_node_blockptr(node, slot); |
6b80053d CM |
2144 | blocksize = btrfs_level_size(root, level - 1); |
2145 | eb = btrfs_find_tree_block(root, search, blocksize); | |
5f39d397 CM |
2146 | if (eb) { |
2147 | free_extent_buffer(eb); | |
3c69faec CM |
2148 | return; |
2149 | } | |
2150 | ||
a7175319 | 2151 | target = search; |
6b80053d | 2152 | |
5f39d397 | 2153 | nritems = btrfs_header_nritems(node); |
6b80053d | 2154 | nr = slot; |
25b8b936 | 2155 | |
d397712b | 2156 | while (1) { |
6b80053d CM |
2157 | if (direction < 0) { |
2158 | if (nr == 0) | |
2159 | break; | |
2160 | nr--; | |
2161 | } else if (direction > 0) { | |
2162 | nr++; | |
2163 | if (nr >= nritems) | |
2164 | break; | |
3c69faec | 2165 | } |
01f46658 CM |
2166 | if (path->reada < 0 && objectid) { |
2167 | btrfs_node_key(node, &disk_key, nr); | |
2168 | if (btrfs_disk_key_objectid(&disk_key) != objectid) | |
2169 | break; | |
2170 | } | |
6b80053d | 2171 | search = btrfs_node_blockptr(node, nr); |
a7175319 CM |
2172 | if ((search <= target && target - search <= 65536) || |
2173 | (search > target && search - target <= 65536)) { | |
cb25c2ea | 2174 | gen = btrfs_node_ptr_generation(node, nr); |
cb25c2ea | 2175 | readahead_tree_block(root, search, blocksize, gen); |
6b80053d CM |
2176 | nread += blocksize; |
2177 | } | |
2178 | nscan++; | |
a7175319 | 2179 | if ((nread > 65536 || nscan > 32)) |
6b80053d | 2180 | break; |
3c69faec CM |
2181 | } |
2182 | } | |
925baedd | 2183 | |
b4ce94de CM |
2184 | /* |
2185 | * returns -EAGAIN if it had to drop the path, or zero if everything was in | |
2186 | * cache | |
2187 | */ | |
2188 | static noinline int reada_for_balance(struct btrfs_root *root, | |
2189 | struct btrfs_path *path, int level) | |
2190 | { | |
2191 | int slot; | |
2192 | int nritems; | |
2193 | struct extent_buffer *parent; | |
2194 | struct extent_buffer *eb; | |
2195 | u64 gen; | |
2196 | u64 block1 = 0; | |
2197 | u64 block2 = 0; | |
2198 | int ret = 0; | |
2199 | int blocksize; | |
2200 | ||
8c594ea8 | 2201 | parent = path->nodes[level + 1]; |
b4ce94de CM |
2202 | if (!parent) |
2203 | return 0; | |
2204 | ||
2205 | nritems = btrfs_header_nritems(parent); | |
8c594ea8 | 2206 | slot = path->slots[level + 1]; |
b4ce94de CM |
2207 | blocksize = btrfs_level_size(root, level); |
2208 | ||
2209 | if (slot > 0) { | |
2210 | block1 = btrfs_node_blockptr(parent, slot - 1); | |
2211 | gen = btrfs_node_ptr_generation(parent, slot - 1); | |
2212 | eb = btrfs_find_tree_block(root, block1, blocksize); | |
b9fab919 CM |
2213 | /* |
2214 | * if we get -eagain from btrfs_buffer_uptodate, we | |
2215 | * don't want to return eagain here. That will loop | |
2216 | * forever | |
2217 | */ | |
2218 | if (eb && btrfs_buffer_uptodate(eb, gen, 1) != 0) | |
b4ce94de CM |
2219 | block1 = 0; |
2220 | free_extent_buffer(eb); | |
2221 | } | |
8c594ea8 | 2222 | if (slot + 1 < nritems) { |
b4ce94de CM |
2223 | block2 = btrfs_node_blockptr(parent, slot + 1); |
2224 | gen = btrfs_node_ptr_generation(parent, slot + 1); | |
2225 | eb = btrfs_find_tree_block(root, block2, blocksize); | |
b9fab919 | 2226 | if (eb && btrfs_buffer_uptodate(eb, gen, 1) != 0) |
b4ce94de CM |
2227 | block2 = 0; |
2228 | free_extent_buffer(eb); | |
2229 | } | |
2230 | if (block1 || block2) { | |
2231 | ret = -EAGAIN; | |
8c594ea8 CM |
2232 | |
2233 | /* release the whole path */ | |
b3b4aa74 | 2234 | btrfs_release_path(path); |
8c594ea8 CM |
2235 | |
2236 | /* read the blocks */ | |
b4ce94de CM |
2237 | if (block1) |
2238 | readahead_tree_block(root, block1, blocksize, 0); | |
2239 | if (block2) | |
2240 | readahead_tree_block(root, block2, blocksize, 0); | |
2241 | ||
2242 | if (block1) { | |
2243 | eb = read_tree_block(root, block1, blocksize, 0); | |
2244 | free_extent_buffer(eb); | |
2245 | } | |
8c594ea8 | 2246 | if (block2) { |
b4ce94de CM |
2247 | eb = read_tree_block(root, block2, blocksize, 0); |
2248 | free_extent_buffer(eb); | |
2249 | } | |
2250 | } | |
2251 | return ret; | |
2252 | } | |
2253 | ||
2254 | ||
d352ac68 | 2255 | /* |
d397712b CM |
2256 | * when we walk down the tree, it is usually safe to unlock the higher layers |
2257 | * in the tree. The exceptions are when our path goes through slot 0, because | |
2258 | * operations on the tree might require changing key pointers higher up in the | |
2259 | * tree. | |
d352ac68 | 2260 | * |
d397712b CM |
2261 | * callers might also have set path->keep_locks, which tells this code to keep |
2262 | * the lock if the path points to the last slot in the block. This is part of | |
2263 | * walking through the tree, and selecting the next slot in the higher block. | |
d352ac68 | 2264 | * |
d397712b CM |
2265 | * lowest_unlock sets the lowest level in the tree we're allowed to unlock. so |
2266 | * if lowest_unlock is 1, level 0 won't be unlocked | |
d352ac68 | 2267 | */ |
e02119d5 | 2268 | static noinline void unlock_up(struct btrfs_path *path, int level, |
f7c79f30 CM |
2269 | int lowest_unlock, int min_write_lock_level, |
2270 | int *write_lock_level) | |
925baedd CM |
2271 | { |
2272 | int i; | |
2273 | int skip_level = level; | |
051e1b9f | 2274 | int no_skips = 0; |
925baedd CM |
2275 | struct extent_buffer *t; |
2276 | ||
2277 | for (i = level; i < BTRFS_MAX_LEVEL; i++) { | |
2278 | if (!path->nodes[i]) | |
2279 | break; | |
2280 | if (!path->locks[i]) | |
2281 | break; | |
051e1b9f | 2282 | if (!no_skips && path->slots[i] == 0) { |
925baedd CM |
2283 | skip_level = i + 1; |
2284 | continue; | |
2285 | } | |
051e1b9f | 2286 | if (!no_skips && path->keep_locks) { |
925baedd CM |
2287 | u32 nritems; |
2288 | t = path->nodes[i]; | |
2289 | nritems = btrfs_header_nritems(t); | |
051e1b9f | 2290 | if (nritems < 1 || path->slots[i] >= nritems - 1) { |
925baedd CM |
2291 | skip_level = i + 1; |
2292 | continue; | |
2293 | } | |
2294 | } | |
051e1b9f CM |
2295 | if (skip_level < i && i >= lowest_unlock) |
2296 | no_skips = 1; | |
2297 | ||
925baedd CM |
2298 | t = path->nodes[i]; |
2299 | if (i >= lowest_unlock && i > skip_level && path->locks[i]) { | |
bd681513 | 2300 | btrfs_tree_unlock_rw(t, path->locks[i]); |
925baedd | 2301 | path->locks[i] = 0; |
f7c79f30 CM |
2302 | if (write_lock_level && |
2303 | i > min_write_lock_level && | |
2304 | i <= *write_lock_level) { | |
2305 | *write_lock_level = i - 1; | |
2306 | } | |
925baedd CM |
2307 | } |
2308 | } | |
2309 | } | |
2310 | ||
b4ce94de CM |
2311 | /* |
2312 | * This releases any locks held in the path starting at level and | |
2313 | * going all the way up to the root. | |
2314 | * | |
2315 | * btrfs_search_slot will keep the lock held on higher nodes in a few | |
2316 | * corner cases, such as COW of the block at slot zero in the node. This | |
2317 | * ignores those rules, and it should only be called when there are no | |
2318 | * more updates to be done higher up in the tree. | |
2319 | */ | |
2320 | noinline void btrfs_unlock_up_safe(struct btrfs_path *path, int level) | |
2321 | { | |
2322 | int i; | |
2323 | ||
09a2a8f9 | 2324 | if (path->keep_locks) |
b4ce94de CM |
2325 | return; |
2326 | ||
2327 | for (i = level; i < BTRFS_MAX_LEVEL; i++) { | |
2328 | if (!path->nodes[i]) | |
12f4dacc | 2329 | continue; |
b4ce94de | 2330 | if (!path->locks[i]) |
12f4dacc | 2331 | continue; |
bd681513 | 2332 | btrfs_tree_unlock_rw(path->nodes[i], path->locks[i]); |
b4ce94de CM |
2333 | path->locks[i] = 0; |
2334 | } | |
2335 | } | |
2336 | ||
c8c42864 CM |
2337 | /* |
2338 | * helper function for btrfs_search_slot. The goal is to find a block | |
2339 | * in cache without setting the path to blocking. If we find the block | |
2340 | * we return zero and the path is unchanged. | |
2341 | * | |
2342 | * If we can't find the block, we set the path blocking and do some | |
2343 | * reada. -EAGAIN is returned and the search must be repeated. | |
2344 | */ | |
2345 | static int | |
2346 | read_block_for_search(struct btrfs_trans_handle *trans, | |
2347 | struct btrfs_root *root, struct btrfs_path *p, | |
2348 | struct extent_buffer **eb_ret, int level, int slot, | |
5d9e75c4 | 2349 | struct btrfs_key *key, u64 time_seq) |
c8c42864 CM |
2350 | { |
2351 | u64 blocknr; | |
2352 | u64 gen; | |
2353 | u32 blocksize; | |
2354 | struct extent_buffer *b = *eb_ret; | |
2355 | struct extent_buffer *tmp; | |
76a05b35 | 2356 | int ret; |
c8c42864 CM |
2357 | |
2358 | blocknr = btrfs_node_blockptr(b, slot); | |
2359 | gen = btrfs_node_ptr_generation(b, slot); | |
2360 | blocksize = btrfs_level_size(root, level - 1); | |
2361 | ||
2362 | tmp = btrfs_find_tree_block(root, blocknr, blocksize); | |
cb44921a | 2363 | if (tmp) { |
b9fab919 CM |
2364 | /* first we do an atomic uptodate check */ |
2365 | if (btrfs_buffer_uptodate(tmp, 0, 1) > 0) { | |
2366 | if (btrfs_buffer_uptodate(tmp, gen, 1) > 0) { | |
cb44921a CM |
2367 | /* |
2368 | * we found an up to date block without | |
2369 | * sleeping, return | |
2370 | * right away | |
2371 | */ | |
2372 | *eb_ret = tmp; | |
2373 | return 0; | |
2374 | } | |
2375 | /* the pages were up to date, but we failed | |
2376 | * the generation number check. Do a full | |
2377 | * read for the generation number that is correct. | |
2378 | * We must do this without dropping locks so | |
2379 | * we can trust our generation number | |
2380 | */ | |
2381 | free_extent_buffer(tmp); | |
bd681513 CM |
2382 | btrfs_set_path_blocking(p); |
2383 | ||
b9fab919 | 2384 | /* now we're allowed to do a blocking uptodate check */ |
cb44921a | 2385 | tmp = read_tree_block(root, blocknr, blocksize, gen); |
b9fab919 | 2386 | if (tmp && btrfs_buffer_uptodate(tmp, gen, 0) > 0) { |
cb44921a CM |
2387 | *eb_ret = tmp; |
2388 | return 0; | |
2389 | } | |
2390 | free_extent_buffer(tmp); | |
b3b4aa74 | 2391 | btrfs_release_path(p); |
cb44921a CM |
2392 | return -EIO; |
2393 | } | |
c8c42864 CM |
2394 | } |
2395 | ||
2396 | /* | |
2397 | * reduce lock contention at high levels | |
2398 | * of the btree by dropping locks before | |
76a05b35 CM |
2399 | * we read. Don't release the lock on the current |
2400 | * level because we need to walk this node to figure | |
2401 | * out which blocks to read. | |
c8c42864 | 2402 | */ |
8c594ea8 CM |
2403 | btrfs_unlock_up_safe(p, level + 1); |
2404 | btrfs_set_path_blocking(p); | |
2405 | ||
cb44921a | 2406 | free_extent_buffer(tmp); |
c8c42864 CM |
2407 | if (p->reada) |
2408 | reada_for_search(root, p, level, slot, key->objectid); | |
2409 | ||
b3b4aa74 | 2410 | btrfs_release_path(p); |
76a05b35 CM |
2411 | |
2412 | ret = -EAGAIN; | |
5bdd3536 | 2413 | tmp = read_tree_block(root, blocknr, blocksize, 0); |
76a05b35 CM |
2414 | if (tmp) { |
2415 | /* | |
2416 | * If the read above didn't mark this buffer up to date, | |
2417 | * it will never end up being up to date. Set ret to EIO now | |
2418 | * and give up so that our caller doesn't loop forever | |
2419 | * on our EAGAINs. | |
2420 | */ | |
b9fab919 | 2421 | if (!btrfs_buffer_uptodate(tmp, 0, 0)) |
76a05b35 | 2422 | ret = -EIO; |
c8c42864 | 2423 | free_extent_buffer(tmp); |
76a05b35 CM |
2424 | } |
2425 | return ret; | |
c8c42864 CM |
2426 | } |
2427 | ||
2428 | /* | |
2429 | * helper function for btrfs_search_slot. This does all of the checks | |
2430 | * for node-level blocks and does any balancing required based on | |
2431 | * the ins_len. | |
2432 | * | |
2433 | * If no extra work was required, zero is returned. If we had to | |
2434 | * drop the path, -EAGAIN is returned and btrfs_search_slot must | |
2435 | * start over | |
2436 | */ | |
2437 | static int | |
2438 | setup_nodes_for_search(struct btrfs_trans_handle *trans, | |
2439 | struct btrfs_root *root, struct btrfs_path *p, | |
bd681513 CM |
2440 | struct extent_buffer *b, int level, int ins_len, |
2441 | int *write_lock_level) | |
c8c42864 CM |
2442 | { |
2443 | int ret; | |
2444 | if ((p->search_for_split || ins_len > 0) && btrfs_header_nritems(b) >= | |
2445 | BTRFS_NODEPTRS_PER_BLOCK(root) - 3) { | |
2446 | int sret; | |
2447 | ||
bd681513 CM |
2448 | if (*write_lock_level < level + 1) { |
2449 | *write_lock_level = level + 1; | |
2450 | btrfs_release_path(p); | |
2451 | goto again; | |
2452 | } | |
2453 | ||
c8c42864 CM |
2454 | sret = reada_for_balance(root, p, level); |
2455 | if (sret) | |
2456 | goto again; | |
2457 | ||
2458 | btrfs_set_path_blocking(p); | |
2459 | sret = split_node(trans, root, p, level); | |
bd681513 | 2460 | btrfs_clear_path_blocking(p, NULL, 0); |
c8c42864 CM |
2461 | |
2462 | BUG_ON(sret > 0); | |
2463 | if (sret) { | |
2464 | ret = sret; | |
2465 | goto done; | |
2466 | } | |
2467 | b = p->nodes[level]; | |
2468 | } else if (ins_len < 0 && btrfs_header_nritems(b) < | |
cfbb9308 | 2469 | BTRFS_NODEPTRS_PER_BLOCK(root) / 2) { |
c8c42864 CM |
2470 | int sret; |
2471 | ||
bd681513 CM |
2472 | if (*write_lock_level < level + 1) { |
2473 | *write_lock_level = level + 1; | |
2474 | btrfs_release_path(p); | |
2475 | goto again; | |
2476 | } | |
2477 | ||
c8c42864 CM |
2478 | sret = reada_for_balance(root, p, level); |
2479 | if (sret) | |
2480 | goto again; | |
2481 | ||
2482 | btrfs_set_path_blocking(p); | |
2483 | sret = balance_level(trans, root, p, level); | |
bd681513 | 2484 | btrfs_clear_path_blocking(p, NULL, 0); |
c8c42864 CM |
2485 | |
2486 | if (sret) { | |
2487 | ret = sret; | |
2488 | goto done; | |
2489 | } | |
2490 | b = p->nodes[level]; | |
2491 | if (!b) { | |
b3b4aa74 | 2492 | btrfs_release_path(p); |
c8c42864 CM |
2493 | goto again; |
2494 | } | |
2495 | BUG_ON(btrfs_header_nritems(b) == 1); | |
2496 | } | |
2497 | return 0; | |
2498 | ||
2499 | again: | |
2500 | ret = -EAGAIN; | |
2501 | done: | |
2502 | return ret; | |
2503 | } | |
2504 | ||
74123bd7 CM |
2505 | /* |
2506 | * look for key in the tree. path is filled in with nodes along the way | |
2507 | * if key is found, we return zero and you can find the item in the leaf | |
2508 | * level of the path (level 0) | |
2509 | * | |
2510 | * If the key isn't found, the path points to the slot where it should | |
aa5d6bed CM |
2511 | * be inserted, and 1 is returned. If there are other errors during the |
2512 | * search a negative error number is returned. | |
97571fd0 CM |
2513 | * |
2514 | * if ins_len > 0, nodes and leaves will be split as we walk down the | |
2515 | * tree. if ins_len < 0, nodes will be merged as we walk down the tree (if | |
2516 | * possible) | |
74123bd7 | 2517 | */ |
e089f05c CM |
2518 | int btrfs_search_slot(struct btrfs_trans_handle *trans, struct btrfs_root |
2519 | *root, struct btrfs_key *key, struct btrfs_path *p, int | |
2520 | ins_len, int cow) | |
be0e5c09 | 2521 | { |
5f39d397 | 2522 | struct extent_buffer *b; |
be0e5c09 CM |
2523 | int slot; |
2524 | int ret; | |
33c66f43 | 2525 | int err; |
be0e5c09 | 2526 | int level; |
925baedd | 2527 | int lowest_unlock = 1; |
bd681513 CM |
2528 | int root_lock; |
2529 | /* everything at write_lock_level or lower must be write locked */ | |
2530 | int write_lock_level = 0; | |
9f3a7427 | 2531 | u8 lowest_level = 0; |
f7c79f30 | 2532 | int min_write_lock_level; |
9f3a7427 | 2533 | |
6702ed49 | 2534 | lowest_level = p->lowest_level; |
323ac95b | 2535 | WARN_ON(lowest_level && ins_len > 0); |
22b0ebda | 2536 | WARN_ON(p->nodes[0] != NULL); |
25179201 | 2537 | |
bd681513 | 2538 | if (ins_len < 0) { |
925baedd | 2539 | lowest_unlock = 2; |
65b51a00 | 2540 | |
bd681513 CM |
2541 | /* when we are removing items, we might have to go up to level |
2542 | * two as we update tree pointers Make sure we keep write | |
2543 | * for those levels as well | |
2544 | */ | |
2545 | write_lock_level = 2; | |
2546 | } else if (ins_len > 0) { | |
2547 | /* | |
2548 | * for inserting items, make sure we have a write lock on | |
2549 | * level 1 so we can update keys | |
2550 | */ | |
2551 | write_lock_level = 1; | |
2552 | } | |
2553 | ||
2554 | if (!cow) | |
2555 | write_lock_level = -1; | |
2556 | ||
09a2a8f9 | 2557 | if (cow && (p->keep_locks || p->lowest_level)) |
bd681513 CM |
2558 | write_lock_level = BTRFS_MAX_LEVEL; |
2559 | ||
f7c79f30 CM |
2560 | min_write_lock_level = write_lock_level; |
2561 | ||
bb803951 | 2562 | again: |
bd681513 CM |
2563 | /* |
2564 | * we try very hard to do read locks on the root | |
2565 | */ | |
2566 | root_lock = BTRFS_READ_LOCK; | |
2567 | level = 0; | |
5d4f98a2 | 2568 | if (p->search_commit_root) { |
bd681513 CM |
2569 | /* |
2570 | * the commit roots are read only | |
2571 | * so we always do read locks | |
2572 | */ | |
5d4f98a2 YZ |
2573 | b = root->commit_root; |
2574 | extent_buffer_get(b); | |
bd681513 | 2575 | level = btrfs_header_level(b); |
5d4f98a2 | 2576 | if (!p->skip_locking) |
bd681513 | 2577 | btrfs_tree_read_lock(b); |
5d4f98a2 | 2578 | } else { |
bd681513 | 2579 | if (p->skip_locking) { |
5d4f98a2 | 2580 | b = btrfs_root_node(root); |
bd681513 CM |
2581 | level = btrfs_header_level(b); |
2582 | } else { | |
2583 | /* we don't know the level of the root node | |
2584 | * until we actually have it read locked | |
2585 | */ | |
2586 | b = btrfs_read_lock_root_node(root); | |
2587 | level = btrfs_header_level(b); | |
2588 | if (level <= write_lock_level) { | |
2589 | /* whoops, must trade for write lock */ | |
2590 | btrfs_tree_read_unlock(b); | |
2591 | free_extent_buffer(b); | |
2592 | b = btrfs_lock_root_node(root); | |
2593 | root_lock = BTRFS_WRITE_LOCK; | |
2594 | ||
2595 | /* the level might have changed, check again */ | |
2596 | level = btrfs_header_level(b); | |
2597 | } | |
2598 | } | |
5d4f98a2 | 2599 | } |
bd681513 CM |
2600 | p->nodes[level] = b; |
2601 | if (!p->skip_locking) | |
2602 | p->locks[level] = root_lock; | |
925baedd | 2603 | |
eb60ceac | 2604 | while (b) { |
5f39d397 | 2605 | level = btrfs_header_level(b); |
65b51a00 CM |
2606 | |
2607 | /* | |
2608 | * setup the path here so we can release it under lock | |
2609 | * contention with the cow code | |
2610 | */ | |
02217ed2 | 2611 | if (cow) { |
c8c42864 CM |
2612 | /* |
2613 | * if we don't really need to cow this block | |
2614 | * then we don't want to set the path blocking, | |
2615 | * so we test it here | |
2616 | */ | |
5d4f98a2 | 2617 | if (!should_cow_block(trans, root, b)) |
65b51a00 | 2618 | goto cow_done; |
5d4f98a2 | 2619 | |
b4ce94de CM |
2620 | btrfs_set_path_blocking(p); |
2621 | ||
bd681513 CM |
2622 | /* |
2623 | * must have write locks on this node and the | |
2624 | * parent | |
2625 | */ | |
5124e00e JB |
2626 | if (level > write_lock_level || |
2627 | (level + 1 > write_lock_level && | |
2628 | level + 1 < BTRFS_MAX_LEVEL && | |
2629 | p->nodes[level + 1])) { | |
bd681513 CM |
2630 | write_lock_level = level + 1; |
2631 | btrfs_release_path(p); | |
2632 | goto again; | |
2633 | } | |
2634 | ||
33c66f43 YZ |
2635 | err = btrfs_cow_block(trans, root, b, |
2636 | p->nodes[level + 1], | |
2637 | p->slots[level + 1], &b); | |
2638 | if (err) { | |
33c66f43 | 2639 | ret = err; |
65b51a00 | 2640 | goto done; |
54aa1f4d | 2641 | } |
02217ed2 | 2642 | } |
65b51a00 | 2643 | cow_done: |
02217ed2 | 2644 | BUG_ON(!cow && ins_len); |
65b51a00 | 2645 | |
eb60ceac | 2646 | p->nodes[level] = b; |
bd681513 | 2647 | btrfs_clear_path_blocking(p, NULL, 0); |
b4ce94de CM |
2648 | |
2649 | /* | |
2650 | * we have a lock on b and as long as we aren't changing | |
2651 | * the tree, there is no way to for the items in b to change. | |
2652 | * It is safe to drop the lock on our parent before we | |
2653 | * go through the expensive btree search on b. | |
2654 | * | |
2655 | * If cow is true, then we might be changing slot zero, | |
2656 | * which may require changing the parent. So, we can't | |
2657 | * drop the lock until after we know which slot we're | |
2658 | * operating on. | |
2659 | */ | |
2660 | if (!cow) | |
2661 | btrfs_unlock_up_safe(p, level + 1); | |
2662 | ||
5f39d397 | 2663 | ret = bin_search(b, key, level, &slot); |
b4ce94de | 2664 | |
5f39d397 | 2665 | if (level != 0) { |
33c66f43 YZ |
2666 | int dec = 0; |
2667 | if (ret && slot > 0) { | |
2668 | dec = 1; | |
be0e5c09 | 2669 | slot -= 1; |
33c66f43 | 2670 | } |
be0e5c09 | 2671 | p->slots[level] = slot; |
33c66f43 | 2672 | err = setup_nodes_for_search(trans, root, p, b, level, |
bd681513 | 2673 | ins_len, &write_lock_level); |
33c66f43 | 2674 | if (err == -EAGAIN) |
c8c42864 | 2675 | goto again; |
33c66f43 YZ |
2676 | if (err) { |
2677 | ret = err; | |
c8c42864 | 2678 | goto done; |
33c66f43 | 2679 | } |
c8c42864 CM |
2680 | b = p->nodes[level]; |
2681 | slot = p->slots[level]; | |
b4ce94de | 2682 | |
bd681513 CM |
2683 | /* |
2684 | * slot 0 is special, if we change the key | |
2685 | * we have to update the parent pointer | |
2686 | * which means we must have a write lock | |
2687 | * on the parent | |
2688 | */ | |
2689 | if (slot == 0 && cow && | |
2690 | write_lock_level < level + 1) { | |
2691 | write_lock_level = level + 1; | |
2692 | btrfs_release_path(p); | |
2693 | goto again; | |
2694 | } | |
2695 | ||
f7c79f30 CM |
2696 | unlock_up(p, level, lowest_unlock, |
2697 | min_write_lock_level, &write_lock_level); | |
f9efa9c7 | 2698 | |
925baedd | 2699 | if (level == lowest_level) { |
33c66f43 YZ |
2700 | if (dec) |
2701 | p->slots[level]++; | |
5b21f2ed | 2702 | goto done; |
925baedd | 2703 | } |
ca7a79ad | 2704 | |
33c66f43 | 2705 | err = read_block_for_search(trans, root, p, |
5d9e75c4 | 2706 | &b, level, slot, key, 0); |
33c66f43 | 2707 | if (err == -EAGAIN) |
c8c42864 | 2708 | goto again; |
33c66f43 YZ |
2709 | if (err) { |
2710 | ret = err; | |
76a05b35 | 2711 | goto done; |
33c66f43 | 2712 | } |
76a05b35 | 2713 | |
b4ce94de | 2714 | if (!p->skip_locking) { |
bd681513 CM |
2715 | level = btrfs_header_level(b); |
2716 | if (level <= write_lock_level) { | |
2717 | err = btrfs_try_tree_write_lock(b); | |
2718 | if (!err) { | |
2719 | btrfs_set_path_blocking(p); | |
2720 | btrfs_tree_lock(b); | |
2721 | btrfs_clear_path_blocking(p, b, | |
2722 | BTRFS_WRITE_LOCK); | |
2723 | } | |
2724 | p->locks[level] = BTRFS_WRITE_LOCK; | |
2725 | } else { | |
2726 | err = btrfs_try_tree_read_lock(b); | |
2727 | if (!err) { | |
2728 | btrfs_set_path_blocking(p); | |
2729 | btrfs_tree_read_lock(b); | |
2730 | btrfs_clear_path_blocking(p, b, | |
2731 | BTRFS_READ_LOCK); | |
2732 | } | |
2733 | p->locks[level] = BTRFS_READ_LOCK; | |
b4ce94de | 2734 | } |
bd681513 | 2735 | p->nodes[level] = b; |
b4ce94de | 2736 | } |
be0e5c09 CM |
2737 | } else { |
2738 | p->slots[level] = slot; | |
87b29b20 YZ |
2739 | if (ins_len > 0 && |
2740 | btrfs_leaf_free_space(root, b) < ins_len) { | |
bd681513 CM |
2741 | if (write_lock_level < 1) { |
2742 | write_lock_level = 1; | |
2743 | btrfs_release_path(p); | |
2744 | goto again; | |
2745 | } | |
2746 | ||
b4ce94de | 2747 | btrfs_set_path_blocking(p); |
33c66f43 YZ |
2748 | err = split_leaf(trans, root, key, |
2749 | p, ins_len, ret == 0); | |
bd681513 | 2750 | btrfs_clear_path_blocking(p, NULL, 0); |
b4ce94de | 2751 | |
33c66f43 YZ |
2752 | BUG_ON(err > 0); |
2753 | if (err) { | |
2754 | ret = err; | |
65b51a00 CM |
2755 | goto done; |
2756 | } | |
5c680ed6 | 2757 | } |
459931ec | 2758 | if (!p->search_for_split) |
f7c79f30 CM |
2759 | unlock_up(p, level, lowest_unlock, |
2760 | min_write_lock_level, &write_lock_level); | |
65b51a00 | 2761 | goto done; |
be0e5c09 CM |
2762 | } |
2763 | } | |
65b51a00 CM |
2764 | ret = 1; |
2765 | done: | |
b4ce94de CM |
2766 | /* |
2767 | * we don't really know what they plan on doing with the path | |
2768 | * from here on, so for now just mark it as blocking | |
2769 | */ | |
b9473439 CM |
2770 | if (!p->leave_spinning) |
2771 | btrfs_set_path_blocking(p); | |
76a05b35 | 2772 | if (ret < 0) |
b3b4aa74 | 2773 | btrfs_release_path(p); |
65b51a00 | 2774 | return ret; |
be0e5c09 CM |
2775 | } |
2776 | ||
5d9e75c4 JS |
2777 | /* |
2778 | * Like btrfs_search_slot, this looks for a key in the given tree. It uses the | |
2779 | * current state of the tree together with the operations recorded in the tree | |
2780 | * modification log to search for the key in a previous version of this tree, as | |
2781 | * denoted by the time_seq parameter. | |
2782 | * | |
2783 | * Naturally, there is no support for insert, delete or cow operations. | |
2784 | * | |
2785 | * The resulting path and return value will be set up as if we called | |
2786 | * btrfs_search_slot at that point in time with ins_len and cow both set to 0. | |
2787 | */ | |
2788 | int btrfs_search_old_slot(struct btrfs_root *root, struct btrfs_key *key, | |
2789 | struct btrfs_path *p, u64 time_seq) | |
2790 | { | |
2791 | struct extent_buffer *b; | |
2792 | int slot; | |
2793 | int ret; | |
2794 | int err; | |
2795 | int level; | |
2796 | int lowest_unlock = 1; | |
2797 | u8 lowest_level = 0; | |
2798 | ||
2799 | lowest_level = p->lowest_level; | |
2800 | WARN_ON(p->nodes[0] != NULL); | |
2801 | ||
2802 | if (p->search_commit_root) { | |
2803 | BUG_ON(time_seq); | |
2804 | return btrfs_search_slot(NULL, root, key, p, 0, 0); | |
2805 | } | |
2806 | ||
2807 | again: | |
5d9e75c4 | 2808 | b = get_old_root(root, time_seq); |
5d9e75c4 | 2809 | level = btrfs_header_level(b); |
5d9e75c4 JS |
2810 | p->locks[level] = BTRFS_READ_LOCK; |
2811 | ||
2812 | while (b) { | |
2813 | level = btrfs_header_level(b); | |
2814 | p->nodes[level] = b; | |
2815 | btrfs_clear_path_blocking(p, NULL, 0); | |
2816 | ||
2817 | /* | |
2818 | * we have a lock on b and as long as we aren't changing | |
2819 | * the tree, there is no way to for the items in b to change. | |
2820 | * It is safe to drop the lock on our parent before we | |
2821 | * go through the expensive btree search on b. | |
2822 | */ | |
2823 | btrfs_unlock_up_safe(p, level + 1); | |
2824 | ||
2825 | ret = bin_search(b, key, level, &slot); | |
2826 | ||
2827 | if (level != 0) { | |
2828 | int dec = 0; | |
2829 | if (ret && slot > 0) { | |
2830 | dec = 1; | |
2831 | slot -= 1; | |
2832 | } | |
2833 | p->slots[level] = slot; | |
2834 | unlock_up(p, level, lowest_unlock, 0, NULL); | |
2835 | ||
2836 | if (level == lowest_level) { | |
2837 | if (dec) | |
2838 | p->slots[level]++; | |
2839 | goto done; | |
2840 | } | |
2841 | ||
2842 | err = read_block_for_search(NULL, root, p, &b, level, | |
2843 | slot, key, time_seq); | |
2844 | if (err == -EAGAIN) | |
2845 | goto again; | |
2846 | if (err) { | |
2847 | ret = err; | |
2848 | goto done; | |
2849 | } | |
2850 | ||
2851 | level = btrfs_header_level(b); | |
2852 | err = btrfs_try_tree_read_lock(b); | |
2853 | if (!err) { | |
2854 | btrfs_set_path_blocking(p); | |
2855 | btrfs_tree_read_lock(b); | |
2856 | btrfs_clear_path_blocking(p, b, | |
2857 | BTRFS_READ_LOCK); | |
2858 | } | |
47fb091f | 2859 | b = tree_mod_log_rewind(root->fs_info, b, time_seq); |
5d9e75c4 JS |
2860 | p->locks[level] = BTRFS_READ_LOCK; |
2861 | p->nodes[level] = b; | |
5d9e75c4 JS |
2862 | } else { |
2863 | p->slots[level] = slot; | |
2864 | unlock_up(p, level, lowest_unlock, 0, NULL); | |
2865 | goto done; | |
2866 | } | |
2867 | } | |
2868 | ret = 1; | |
2869 | done: | |
2870 | if (!p->leave_spinning) | |
2871 | btrfs_set_path_blocking(p); | |
2872 | if (ret < 0) | |
2873 | btrfs_release_path(p); | |
2874 | ||
2875 | return ret; | |
2876 | } | |
2877 | ||
2f38b3e1 AJ |
2878 | /* |
2879 | * helper to use instead of search slot if no exact match is needed but | |
2880 | * instead the next or previous item should be returned. | |
2881 | * When find_higher is true, the next higher item is returned, the next lower | |
2882 | * otherwise. | |
2883 | * When return_any and find_higher are both true, and no higher item is found, | |
2884 | * return the next lower instead. | |
2885 | * When return_any is true and find_higher is false, and no lower item is found, | |
2886 | * return the next higher instead. | |
2887 | * It returns 0 if any item is found, 1 if none is found (tree empty), and | |
2888 | * < 0 on error | |
2889 | */ | |
2890 | int btrfs_search_slot_for_read(struct btrfs_root *root, | |
2891 | struct btrfs_key *key, struct btrfs_path *p, | |
2892 | int find_higher, int return_any) | |
2893 | { | |
2894 | int ret; | |
2895 | struct extent_buffer *leaf; | |
2896 | ||
2897 | again: | |
2898 | ret = btrfs_search_slot(NULL, root, key, p, 0, 0); | |
2899 | if (ret <= 0) | |
2900 | return ret; | |
2901 | /* | |
2902 | * a return value of 1 means the path is at the position where the | |
2903 | * item should be inserted. Normally this is the next bigger item, | |
2904 | * but in case the previous item is the last in a leaf, path points | |
2905 | * to the first free slot in the previous leaf, i.e. at an invalid | |
2906 | * item. | |
2907 | */ | |
2908 | leaf = p->nodes[0]; | |
2909 | ||
2910 | if (find_higher) { | |
2911 | if (p->slots[0] >= btrfs_header_nritems(leaf)) { | |
2912 | ret = btrfs_next_leaf(root, p); | |
2913 | if (ret <= 0) | |
2914 | return ret; | |
2915 | if (!return_any) | |
2916 | return 1; | |
2917 | /* | |
2918 | * no higher item found, return the next | |
2919 | * lower instead | |
2920 | */ | |
2921 | return_any = 0; | |
2922 | find_higher = 0; | |
2923 | btrfs_release_path(p); | |
2924 | goto again; | |
2925 | } | |
2926 | } else { | |
e6793769 AJ |
2927 | if (p->slots[0] == 0) { |
2928 | ret = btrfs_prev_leaf(root, p); | |
2929 | if (ret < 0) | |
2930 | return ret; | |
2931 | if (!ret) { | |
2932 | p->slots[0] = btrfs_header_nritems(leaf) - 1; | |
2933 | return 0; | |
2f38b3e1 | 2934 | } |
e6793769 AJ |
2935 | if (!return_any) |
2936 | return 1; | |
2937 | /* | |
2938 | * no lower item found, return the next | |
2939 | * higher instead | |
2940 | */ | |
2941 | return_any = 0; | |
2942 | find_higher = 1; | |
2943 | btrfs_release_path(p); | |
2944 | goto again; | |
2945 | } else { | |
2f38b3e1 AJ |
2946 | --p->slots[0]; |
2947 | } | |
2948 | } | |
2949 | return 0; | |
2950 | } | |
2951 | ||
74123bd7 CM |
2952 | /* |
2953 | * adjust the pointers going up the tree, starting at level | |
2954 | * making sure the right key of each node is points to 'key'. | |
2955 | * This is used after shifting pointers to the left, so it stops | |
2956 | * fixing up pointers when a given leaf/node is not in slot 0 of the | |
2957 | * higher levels | |
aa5d6bed | 2958 | * |
74123bd7 | 2959 | */ |
d6a0a126 | 2960 | static void fixup_low_keys(struct btrfs_root *root, struct btrfs_path *path, |
143bede5 | 2961 | struct btrfs_disk_key *key, int level) |
be0e5c09 CM |
2962 | { |
2963 | int i; | |
5f39d397 CM |
2964 | struct extent_buffer *t; |
2965 | ||
234b63a0 | 2966 | for (i = level; i < BTRFS_MAX_LEVEL; i++) { |
be0e5c09 | 2967 | int tslot = path->slots[i]; |
eb60ceac | 2968 | if (!path->nodes[i]) |
be0e5c09 | 2969 | break; |
5f39d397 | 2970 | t = path->nodes[i]; |
32adf090 | 2971 | tree_mod_log_set_node_key(root->fs_info, t, tslot, 1); |
5f39d397 | 2972 | btrfs_set_node_key(t, key, tslot); |
d6025579 | 2973 | btrfs_mark_buffer_dirty(path->nodes[i]); |
be0e5c09 CM |
2974 | if (tslot != 0) |
2975 | break; | |
2976 | } | |
2977 | } | |
2978 | ||
31840ae1 ZY |
2979 | /* |
2980 | * update item key. | |
2981 | * | |
2982 | * This function isn't completely safe. It's the caller's responsibility | |
2983 | * that the new key won't break the order | |
2984 | */ | |
afe5fea7 | 2985 | void btrfs_set_item_key_safe(struct btrfs_root *root, struct btrfs_path *path, |
143bede5 | 2986 | struct btrfs_key *new_key) |
31840ae1 ZY |
2987 | { |
2988 | struct btrfs_disk_key disk_key; | |
2989 | struct extent_buffer *eb; | |
2990 | int slot; | |
2991 | ||
2992 | eb = path->nodes[0]; | |
2993 | slot = path->slots[0]; | |
2994 | if (slot > 0) { | |
2995 | btrfs_item_key(eb, &disk_key, slot - 1); | |
143bede5 | 2996 | BUG_ON(comp_keys(&disk_key, new_key) >= 0); |
31840ae1 ZY |
2997 | } |
2998 | if (slot < btrfs_header_nritems(eb) - 1) { | |
2999 | btrfs_item_key(eb, &disk_key, slot + 1); | |
143bede5 | 3000 | BUG_ON(comp_keys(&disk_key, new_key) <= 0); |
31840ae1 ZY |
3001 | } |
3002 | ||
3003 | btrfs_cpu_key_to_disk(&disk_key, new_key); | |
3004 | btrfs_set_item_key(eb, &disk_key, slot); | |
3005 | btrfs_mark_buffer_dirty(eb); | |
3006 | if (slot == 0) | |
d6a0a126 | 3007 | fixup_low_keys(root, path, &disk_key, 1); |
31840ae1 ZY |
3008 | } |
3009 | ||
74123bd7 CM |
3010 | /* |
3011 | * try to push data from one node into the next node left in the | |
79f95c82 | 3012 | * tree. |
aa5d6bed CM |
3013 | * |
3014 | * returns 0 if some ptrs were pushed left, < 0 if there was some horrible | |
3015 | * error, and > 0 if there was no room in the left hand block. | |
74123bd7 | 3016 | */ |
98ed5174 CM |
3017 | static int push_node_left(struct btrfs_trans_handle *trans, |
3018 | struct btrfs_root *root, struct extent_buffer *dst, | |
971a1f66 | 3019 | struct extent_buffer *src, int empty) |
be0e5c09 | 3020 | { |
be0e5c09 | 3021 | int push_items = 0; |
bb803951 CM |
3022 | int src_nritems; |
3023 | int dst_nritems; | |
aa5d6bed | 3024 | int ret = 0; |
be0e5c09 | 3025 | |
5f39d397 CM |
3026 | src_nritems = btrfs_header_nritems(src); |
3027 | dst_nritems = btrfs_header_nritems(dst); | |
123abc88 | 3028 | push_items = BTRFS_NODEPTRS_PER_BLOCK(root) - dst_nritems; |
7bb86316 CM |
3029 | WARN_ON(btrfs_header_generation(src) != trans->transid); |
3030 | WARN_ON(btrfs_header_generation(dst) != trans->transid); | |
54aa1f4d | 3031 | |
bce4eae9 | 3032 | if (!empty && src_nritems <= 8) |
971a1f66 CM |
3033 | return 1; |
3034 | ||
d397712b | 3035 | if (push_items <= 0) |
be0e5c09 CM |
3036 | return 1; |
3037 | ||
bce4eae9 | 3038 | if (empty) { |
971a1f66 | 3039 | push_items = min(src_nritems, push_items); |
bce4eae9 CM |
3040 | if (push_items < src_nritems) { |
3041 | /* leave at least 8 pointers in the node if | |
3042 | * we aren't going to empty it | |
3043 | */ | |
3044 | if (src_nritems - push_items < 8) { | |
3045 | if (push_items <= 8) | |
3046 | return 1; | |
3047 | push_items -= 8; | |
3048 | } | |
3049 | } | |
3050 | } else | |
3051 | push_items = min(src_nritems - 8, push_items); | |
79f95c82 | 3052 | |
f230475e | 3053 | tree_mod_log_eb_copy(root->fs_info, dst, src, dst_nritems, 0, |
90f8d62e | 3054 | push_items); |
5f39d397 CM |
3055 | copy_extent_buffer(dst, src, |
3056 | btrfs_node_key_ptr_offset(dst_nritems), | |
3057 | btrfs_node_key_ptr_offset(0), | |
d397712b | 3058 | push_items * sizeof(struct btrfs_key_ptr)); |
5f39d397 | 3059 | |
bb803951 | 3060 | if (push_items < src_nritems) { |
57911b8b JS |
3061 | /* |
3062 | * don't call tree_mod_log_eb_move here, key removal was already | |
3063 | * fully logged by tree_mod_log_eb_copy above. | |
3064 | */ | |
5f39d397 CM |
3065 | memmove_extent_buffer(src, btrfs_node_key_ptr_offset(0), |
3066 | btrfs_node_key_ptr_offset(push_items), | |
3067 | (src_nritems - push_items) * | |
3068 | sizeof(struct btrfs_key_ptr)); | |
3069 | } | |
3070 | btrfs_set_header_nritems(src, src_nritems - push_items); | |
3071 | btrfs_set_header_nritems(dst, dst_nritems + push_items); | |
3072 | btrfs_mark_buffer_dirty(src); | |
3073 | btrfs_mark_buffer_dirty(dst); | |
31840ae1 | 3074 | |
79f95c82 CM |
3075 | return ret; |
3076 | } | |
3077 | ||
3078 | /* | |
3079 | * try to push data from one node into the next node right in the | |
3080 | * tree. | |
3081 | * | |
3082 | * returns 0 if some ptrs were pushed, < 0 if there was some horrible | |
3083 | * error, and > 0 if there was no room in the right hand block. | |
3084 | * | |
3085 | * this will only push up to 1/2 the contents of the left node over | |
3086 | */ | |
5f39d397 CM |
3087 | static int balance_node_right(struct btrfs_trans_handle *trans, |
3088 | struct btrfs_root *root, | |
3089 | struct extent_buffer *dst, | |
3090 | struct extent_buffer *src) | |
79f95c82 | 3091 | { |
79f95c82 CM |
3092 | int push_items = 0; |
3093 | int max_push; | |
3094 | int src_nritems; | |
3095 | int dst_nritems; | |
3096 | int ret = 0; | |
79f95c82 | 3097 | |
7bb86316 CM |
3098 | WARN_ON(btrfs_header_generation(src) != trans->transid); |
3099 | WARN_ON(btrfs_header_generation(dst) != trans->transid); | |
3100 | ||
5f39d397 CM |
3101 | src_nritems = btrfs_header_nritems(src); |
3102 | dst_nritems = btrfs_header_nritems(dst); | |
123abc88 | 3103 | push_items = BTRFS_NODEPTRS_PER_BLOCK(root) - dst_nritems; |
d397712b | 3104 | if (push_items <= 0) |
79f95c82 | 3105 | return 1; |
bce4eae9 | 3106 | |
d397712b | 3107 | if (src_nritems < 4) |
bce4eae9 | 3108 | return 1; |
79f95c82 CM |
3109 | |
3110 | max_push = src_nritems / 2 + 1; | |
3111 | /* don't try to empty the node */ | |
d397712b | 3112 | if (max_push >= src_nritems) |
79f95c82 | 3113 | return 1; |
252c38f0 | 3114 | |
79f95c82 CM |
3115 | if (max_push < push_items) |
3116 | push_items = max_push; | |
3117 | ||
f230475e | 3118 | tree_mod_log_eb_move(root->fs_info, dst, push_items, 0, dst_nritems); |
5f39d397 CM |
3119 | memmove_extent_buffer(dst, btrfs_node_key_ptr_offset(push_items), |
3120 | btrfs_node_key_ptr_offset(0), | |
3121 | (dst_nritems) * | |
3122 | sizeof(struct btrfs_key_ptr)); | |
d6025579 | 3123 | |
f230475e | 3124 | tree_mod_log_eb_copy(root->fs_info, dst, src, 0, |
90f8d62e | 3125 | src_nritems - push_items, push_items); |
5f39d397 CM |
3126 | copy_extent_buffer(dst, src, |
3127 | btrfs_node_key_ptr_offset(0), | |
3128 | btrfs_node_key_ptr_offset(src_nritems - push_items), | |
d397712b | 3129 | push_items * sizeof(struct btrfs_key_ptr)); |
79f95c82 | 3130 | |
5f39d397 CM |
3131 | btrfs_set_header_nritems(src, src_nritems - push_items); |
3132 | btrfs_set_header_nritems(dst, dst_nritems + push_items); | |
79f95c82 | 3133 | |
5f39d397 CM |
3134 | btrfs_mark_buffer_dirty(src); |
3135 | btrfs_mark_buffer_dirty(dst); | |
31840ae1 | 3136 | |
aa5d6bed | 3137 | return ret; |
be0e5c09 CM |
3138 | } |
3139 | ||
97571fd0 CM |
3140 | /* |
3141 | * helper function to insert a new root level in the tree. | |
3142 | * A new node is allocated, and a single item is inserted to | |
3143 | * point to the existing root | |
aa5d6bed CM |
3144 | * |
3145 | * returns zero on success or < 0 on failure. | |
97571fd0 | 3146 | */ |
d397712b | 3147 | static noinline int insert_new_root(struct btrfs_trans_handle *trans, |
5f39d397 | 3148 | struct btrfs_root *root, |
90f8d62e | 3149 | struct btrfs_path *path, int level, int log_removal) |
5c680ed6 | 3150 | { |
7bb86316 | 3151 | u64 lower_gen; |
5f39d397 CM |
3152 | struct extent_buffer *lower; |
3153 | struct extent_buffer *c; | |
925baedd | 3154 | struct extent_buffer *old; |
5f39d397 | 3155 | struct btrfs_disk_key lower_key; |
5c680ed6 CM |
3156 | |
3157 | BUG_ON(path->nodes[level]); | |
3158 | BUG_ON(path->nodes[level-1] != root->node); | |
3159 | ||
7bb86316 CM |
3160 | lower = path->nodes[level-1]; |
3161 | if (level == 1) | |
3162 | btrfs_item_key(lower, &lower_key, 0); | |
3163 | else | |
3164 | btrfs_node_key(lower, &lower_key, 0); | |
3165 | ||
31840ae1 | 3166 | c = btrfs_alloc_free_block(trans, root, root->nodesize, 0, |
5d4f98a2 | 3167 | root->root_key.objectid, &lower_key, |
5581a51a | 3168 | level, root->node->start, 0); |
5f39d397 CM |
3169 | if (IS_ERR(c)) |
3170 | return PTR_ERR(c); | |
925baedd | 3171 | |
f0486c68 YZ |
3172 | root_add_used(root, root->nodesize); |
3173 | ||
5d4f98a2 | 3174 | memset_extent_buffer(c, 0, 0, sizeof(struct btrfs_header)); |
5f39d397 CM |
3175 | btrfs_set_header_nritems(c, 1); |
3176 | btrfs_set_header_level(c, level); | |
db94535d | 3177 | btrfs_set_header_bytenr(c, c->start); |
5f39d397 | 3178 | btrfs_set_header_generation(c, trans->transid); |
5d4f98a2 | 3179 | btrfs_set_header_backref_rev(c, BTRFS_MIXED_BACKREF_REV); |
5f39d397 | 3180 | btrfs_set_header_owner(c, root->root_key.objectid); |
5f39d397 CM |
3181 | |
3182 | write_extent_buffer(c, root->fs_info->fsid, | |
3183 | (unsigned long)btrfs_header_fsid(c), | |
3184 | BTRFS_FSID_SIZE); | |
e17cade2 CM |
3185 | |
3186 | write_extent_buffer(c, root->fs_info->chunk_tree_uuid, | |
3187 | (unsigned long)btrfs_header_chunk_tree_uuid(c), | |
3188 | BTRFS_UUID_SIZE); | |
3189 | ||
5f39d397 | 3190 | btrfs_set_node_key(c, &lower_key, 0); |
db94535d | 3191 | btrfs_set_node_blockptr(c, 0, lower->start); |
7bb86316 | 3192 | lower_gen = btrfs_header_generation(lower); |
31840ae1 | 3193 | WARN_ON(lower_gen != trans->transid); |
7bb86316 CM |
3194 | |
3195 | btrfs_set_node_ptr_generation(c, 0, lower_gen); | |
d5719762 | 3196 | |
5f39d397 | 3197 | btrfs_mark_buffer_dirty(c); |
d5719762 | 3198 | |
925baedd | 3199 | old = root->node; |
90f8d62e | 3200 | tree_mod_log_set_root_pointer(root, c, log_removal); |
240f62c8 | 3201 | rcu_assign_pointer(root->node, c); |
925baedd CM |
3202 | |
3203 | /* the super has an extra ref to root->node */ | |
3204 | free_extent_buffer(old); | |
3205 | ||
0b86a832 | 3206 | add_root_to_dirty_list(root); |
5f39d397 CM |
3207 | extent_buffer_get(c); |
3208 | path->nodes[level] = c; | |
bd681513 | 3209 | path->locks[level] = BTRFS_WRITE_LOCK; |
5c680ed6 CM |
3210 | path->slots[level] = 0; |
3211 | return 0; | |
3212 | } | |
3213 | ||
74123bd7 CM |
3214 | /* |
3215 | * worker function to insert a single pointer in a node. | |
3216 | * the node should have enough room for the pointer already | |
97571fd0 | 3217 | * |
74123bd7 CM |
3218 | * slot and level indicate where you want the key to go, and |
3219 | * blocknr is the block the key points to. | |
3220 | */ | |
143bede5 JM |
3221 | static void insert_ptr(struct btrfs_trans_handle *trans, |
3222 | struct btrfs_root *root, struct btrfs_path *path, | |
3223 | struct btrfs_disk_key *key, u64 bytenr, | |
c3e06965 | 3224 | int slot, int level) |
74123bd7 | 3225 | { |
5f39d397 | 3226 | struct extent_buffer *lower; |
74123bd7 | 3227 | int nritems; |
f3ea38da | 3228 | int ret; |
5c680ed6 CM |
3229 | |
3230 | BUG_ON(!path->nodes[level]); | |
f0486c68 | 3231 | btrfs_assert_tree_locked(path->nodes[level]); |
5f39d397 CM |
3232 | lower = path->nodes[level]; |
3233 | nritems = btrfs_header_nritems(lower); | |
c293498b | 3234 | BUG_ON(slot > nritems); |
143bede5 | 3235 | BUG_ON(nritems == BTRFS_NODEPTRS_PER_BLOCK(root)); |
74123bd7 | 3236 | if (slot != nritems) { |
c3e06965 | 3237 | if (level) |
f3ea38da JS |
3238 | tree_mod_log_eb_move(root->fs_info, lower, slot + 1, |
3239 | slot, nritems - slot); | |
5f39d397 CM |
3240 | memmove_extent_buffer(lower, |
3241 | btrfs_node_key_ptr_offset(slot + 1), | |
3242 | btrfs_node_key_ptr_offset(slot), | |
d6025579 | 3243 | (nritems - slot) * sizeof(struct btrfs_key_ptr)); |
74123bd7 | 3244 | } |
c3e06965 | 3245 | if (level) { |
f3ea38da JS |
3246 | ret = tree_mod_log_insert_key(root->fs_info, lower, slot, |
3247 | MOD_LOG_KEY_ADD); | |
3248 | BUG_ON(ret < 0); | |
3249 | } | |
5f39d397 | 3250 | btrfs_set_node_key(lower, key, slot); |
db94535d | 3251 | btrfs_set_node_blockptr(lower, slot, bytenr); |
74493f7a CM |
3252 | WARN_ON(trans->transid == 0); |
3253 | btrfs_set_node_ptr_generation(lower, slot, trans->transid); | |
5f39d397 CM |
3254 | btrfs_set_header_nritems(lower, nritems + 1); |
3255 | btrfs_mark_buffer_dirty(lower); | |
74123bd7 CM |
3256 | } |
3257 | ||
97571fd0 CM |
3258 | /* |
3259 | * split the node at the specified level in path in two. | |
3260 | * The path is corrected to point to the appropriate node after the split | |
3261 | * | |
3262 | * Before splitting this tries to make some room in the node by pushing | |
3263 | * left and right, if either one works, it returns right away. | |
aa5d6bed CM |
3264 | * |
3265 | * returns 0 on success and < 0 on failure | |
97571fd0 | 3266 | */ |
e02119d5 CM |
3267 | static noinline int split_node(struct btrfs_trans_handle *trans, |
3268 | struct btrfs_root *root, | |
3269 | struct btrfs_path *path, int level) | |
be0e5c09 | 3270 | { |
5f39d397 CM |
3271 | struct extent_buffer *c; |
3272 | struct extent_buffer *split; | |
3273 | struct btrfs_disk_key disk_key; | |
be0e5c09 | 3274 | int mid; |
5c680ed6 | 3275 | int ret; |
7518a238 | 3276 | u32 c_nritems; |
eb60ceac | 3277 | |
5f39d397 | 3278 | c = path->nodes[level]; |
7bb86316 | 3279 | WARN_ON(btrfs_header_generation(c) != trans->transid); |
5f39d397 | 3280 | if (c == root->node) { |
d9abbf1c | 3281 | /* |
90f8d62e JS |
3282 | * trying to split the root, lets make a new one |
3283 | * | |
3284 | * tree mod log: We pass 0 as log_removal parameter to | |
3285 | * insert_new_root, because that root buffer will be kept as a | |
3286 | * normal node. We are going to log removal of half of the | |
3287 | * elements below with tree_mod_log_eb_copy. We're holding a | |
3288 | * tree lock on the buffer, which is why we cannot race with | |
3289 | * other tree_mod_log users. | |
d9abbf1c | 3290 | */ |
90f8d62e | 3291 | ret = insert_new_root(trans, root, path, level + 1, 0); |
5c680ed6 CM |
3292 | if (ret) |
3293 | return ret; | |
b3612421 | 3294 | } else { |
e66f709b | 3295 | ret = push_nodes_for_insert(trans, root, path, level); |
5f39d397 CM |
3296 | c = path->nodes[level]; |
3297 | if (!ret && btrfs_header_nritems(c) < | |
c448acf0 | 3298 | BTRFS_NODEPTRS_PER_BLOCK(root) - 3) |
e66f709b | 3299 | return 0; |
54aa1f4d CM |
3300 | if (ret < 0) |
3301 | return ret; | |
be0e5c09 | 3302 | } |
e66f709b | 3303 | |
5f39d397 | 3304 | c_nritems = btrfs_header_nritems(c); |
5d4f98a2 YZ |
3305 | mid = (c_nritems + 1) / 2; |
3306 | btrfs_node_key(c, &disk_key, mid); | |
7bb86316 | 3307 | |
5d4f98a2 | 3308 | split = btrfs_alloc_free_block(trans, root, root->nodesize, 0, |
31840ae1 | 3309 | root->root_key.objectid, |
5581a51a | 3310 | &disk_key, level, c->start, 0); |
5f39d397 CM |
3311 | if (IS_ERR(split)) |
3312 | return PTR_ERR(split); | |
3313 | ||
f0486c68 YZ |
3314 | root_add_used(root, root->nodesize); |
3315 | ||
5d4f98a2 | 3316 | memset_extent_buffer(split, 0, 0, sizeof(struct btrfs_header)); |
5f39d397 | 3317 | btrfs_set_header_level(split, btrfs_header_level(c)); |
db94535d | 3318 | btrfs_set_header_bytenr(split, split->start); |
5f39d397 | 3319 | btrfs_set_header_generation(split, trans->transid); |
5d4f98a2 | 3320 | btrfs_set_header_backref_rev(split, BTRFS_MIXED_BACKREF_REV); |
5f39d397 CM |
3321 | btrfs_set_header_owner(split, root->root_key.objectid); |
3322 | write_extent_buffer(split, root->fs_info->fsid, | |
3323 | (unsigned long)btrfs_header_fsid(split), | |
3324 | BTRFS_FSID_SIZE); | |
e17cade2 CM |
3325 | write_extent_buffer(split, root->fs_info->chunk_tree_uuid, |
3326 | (unsigned long)btrfs_header_chunk_tree_uuid(split), | |
3327 | BTRFS_UUID_SIZE); | |
54aa1f4d | 3328 | |
90f8d62e | 3329 | tree_mod_log_eb_copy(root->fs_info, split, c, 0, mid, c_nritems - mid); |
5f39d397 CM |
3330 | copy_extent_buffer(split, c, |
3331 | btrfs_node_key_ptr_offset(0), | |
3332 | btrfs_node_key_ptr_offset(mid), | |
3333 | (c_nritems - mid) * sizeof(struct btrfs_key_ptr)); | |
3334 | btrfs_set_header_nritems(split, c_nritems - mid); | |
3335 | btrfs_set_header_nritems(c, mid); | |
aa5d6bed CM |
3336 | ret = 0; |
3337 | ||
5f39d397 CM |
3338 | btrfs_mark_buffer_dirty(c); |
3339 | btrfs_mark_buffer_dirty(split); | |
3340 | ||
143bede5 | 3341 | insert_ptr(trans, root, path, &disk_key, split->start, |
c3e06965 | 3342 | path->slots[level + 1] + 1, level + 1); |
aa5d6bed | 3343 | |
5de08d7d | 3344 | if (path->slots[level] >= mid) { |
5c680ed6 | 3345 | path->slots[level] -= mid; |
925baedd | 3346 | btrfs_tree_unlock(c); |
5f39d397 CM |
3347 | free_extent_buffer(c); |
3348 | path->nodes[level] = split; | |
5c680ed6 CM |
3349 | path->slots[level + 1] += 1; |
3350 | } else { | |
925baedd | 3351 | btrfs_tree_unlock(split); |
5f39d397 | 3352 | free_extent_buffer(split); |
be0e5c09 | 3353 | } |
aa5d6bed | 3354 | return ret; |
be0e5c09 CM |
3355 | } |
3356 | ||
74123bd7 CM |
3357 | /* |
3358 | * how many bytes are required to store the items in a leaf. start | |
3359 | * and nr indicate which items in the leaf to check. This totals up the | |
3360 | * space used both by the item structs and the item data | |
3361 | */ | |
5f39d397 | 3362 | static int leaf_space_used(struct extent_buffer *l, int start, int nr) |
be0e5c09 | 3363 | { |
41be1f3b JB |
3364 | struct btrfs_item *start_item; |
3365 | struct btrfs_item *end_item; | |
3366 | struct btrfs_map_token token; | |
be0e5c09 | 3367 | int data_len; |
5f39d397 | 3368 | int nritems = btrfs_header_nritems(l); |
d4dbff95 | 3369 | int end = min(nritems, start + nr) - 1; |
be0e5c09 CM |
3370 | |
3371 | if (!nr) | |
3372 | return 0; | |
41be1f3b JB |
3373 | btrfs_init_map_token(&token); |
3374 | start_item = btrfs_item_nr(l, start); | |
3375 | end_item = btrfs_item_nr(l, end); | |
3376 | data_len = btrfs_token_item_offset(l, start_item, &token) + | |
3377 | btrfs_token_item_size(l, start_item, &token); | |
3378 | data_len = data_len - btrfs_token_item_offset(l, end_item, &token); | |
0783fcfc | 3379 | data_len += sizeof(struct btrfs_item) * nr; |
d4dbff95 | 3380 | WARN_ON(data_len < 0); |
be0e5c09 CM |
3381 | return data_len; |
3382 | } | |
3383 | ||
d4dbff95 CM |
3384 | /* |
3385 | * The space between the end of the leaf items and | |
3386 | * the start of the leaf data. IOW, how much room | |
3387 | * the leaf has left for both items and data | |
3388 | */ | |
d397712b | 3389 | noinline int btrfs_leaf_free_space(struct btrfs_root *root, |
e02119d5 | 3390 | struct extent_buffer *leaf) |
d4dbff95 | 3391 | { |
5f39d397 CM |
3392 | int nritems = btrfs_header_nritems(leaf); |
3393 | int ret; | |
3394 | ret = BTRFS_LEAF_DATA_SIZE(root) - leaf_space_used(leaf, 0, nritems); | |
3395 | if (ret < 0) { | |
d397712b CM |
3396 | printk(KERN_CRIT "leaf free space ret %d, leaf data size %lu, " |
3397 | "used %d nritems %d\n", | |
ae2f5411 | 3398 | ret, (unsigned long) BTRFS_LEAF_DATA_SIZE(root), |
5f39d397 CM |
3399 | leaf_space_used(leaf, 0, nritems), nritems); |
3400 | } | |
3401 | return ret; | |
d4dbff95 CM |
3402 | } |
3403 | ||
99d8f83c CM |
3404 | /* |
3405 | * min slot controls the lowest index we're willing to push to the | |
3406 | * right. We'll push up to and including min_slot, but no lower | |
3407 | */ | |
44871b1b CM |
3408 | static noinline int __push_leaf_right(struct btrfs_trans_handle *trans, |
3409 | struct btrfs_root *root, | |
3410 | struct btrfs_path *path, | |
3411 | int data_size, int empty, | |
3412 | struct extent_buffer *right, | |
99d8f83c CM |
3413 | int free_space, u32 left_nritems, |
3414 | u32 min_slot) | |
00ec4c51 | 3415 | { |
5f39d397 | 3416 | struct extent_buffer *left = path->nodes[0]; |
44871b1b | 3417 | struct extent_buffer *upper = path->nodes[1]; |
cfed81a0 | 3418 | struct btrfs_map_token token; |
5f39d397 | 3419 | struct btrfs_disk_key disk_key; |
00ec4c51 | 3420 | int slot; |
34a38218 | 3421 | u32 i; |
00ec4c51 CM |
3422 | int push_space = 0; |
3423 | int push_items = 0; | |
0783fcfc | 3424 | struct btrfs_item *item; |
34a38218 | 3425 | u32 nr; |
7518a238 | 3426 | u32 right_nritems; |
5f39d397 | 3427 | u32 data_end; |
db94535d | 3428 | u32 this_item_size; |
00ec4c51 | 3429 | |
cfed81a0 CM |
3430 | btrfs_init_map_token(&token); |
3431 | ||
34a38218 CM |
3432 | if (empty) |
3433 | nr = 0; | |
3434 | else | |
99d8f83c | 3435 | nr = max_t(u32, 1, min_slot); |
34a38218 | 3436 | |
31840ae1 | 3437 | if (path->slots[0] >= left_nritems) |
87b29b20 | 3438 | push_space += data_size; |
31840ae1 | 3439 | |
44871b1b | 3440 | slot = path->slots[1]; |
34a38218 CM |
3441 | i = left_nritems - 1; |
3442 | while (i >= nr) { | |
5f39d397 | 3443 | item = btrfs_item_nr(left, i); |
db94535d | 3444 | |
31840ae1 ZY |
3445 | if (!empty && push_items > 0) { |
3446 | if (path->slots[0] > i) | |
3447 | break; | |
3448 | if (path->slots[0] == i) { | |
3449 | int space = btrfs_leaf_free_space(root, left); | |
3450 | if (space + push_space * 2 > free_space) | |
3451 | break; | |
3452 | } | |
3453 | } | |
3454 | ||
00ec4c51 | 3455 | if (path->slots[0] == i) |
87b29b20 | 3456 | push_space += data_size; |
db94535d | 3457 | |
db94535d CM |
3458 | this_item_size = btrfs_item_size(left, item); |
3459 | if (this_item_size + sizeof(*item) + push_space > free_space) | |
00ec4c51 | 3460 | break; |
31840ae1 | 3461 | |
00ec4c51 | 3462 | push_items++; |
db94535d | 3463 | push_space += this_item_size + sizeof(*item); |
34a38218 CM |
3464 | if (i == 0) |
3465 | break; | |
3466 | i--; | |
db94535d | 3467 | } |
5f39d397 | 3468 | |
925baedd CM |
3469 | if (push_items == 0) |
3470 | goto out_unlock; | |
5f39d397 | 3471 | |
6c1500f2 | 3472 | WARN_ON(!empty && push_items == left_nritems); |
5f39d397 | 3473 | |
00ec4c51 | 3474 | /* push left to right */ |
5f39d397 | 3475 | right_nritems = btrfs_header_nritems(right); |
34a38218 | 3476 | |
5f39d397 | 3477 | push_space = btrfs_item_end_nr(left, left_nritems - push_items); |
123abc88 | 3478 | push_space -= leaf_data_end(root, left); |
5f39d397 | 3479 | |
00ec4c51 | 3480 | /* make room in the right data area */ |
5f39d397 CM |
3481 | data_end = leaf_data_end(root, right); |
3482 | memmove_extent_buffer(right, | |
3483 | btrfs_leaf_data(right) + data_end - push_space, | |
3484 | btrfs_leaf_data(right) + data_end, | |
3485 | BTRFS_LEAF_DATA_SIZE(root) - data_end); | |
3486 | ||
00ec4c51 | 3487 | /* copy from the left data area */ |
5f39d397 | 3488 | copy_extent_buffer(right, left, btrfs_leaf_data(right) + |
d6025579 CM |
3489 | BTRFS_LEAF_DATA_SIZE(root) - push_space, |
3490 | btrfs_leaf_data(left) + leaf_data_end(root, left), | |
3491 | push_space); | |
5f39d397 CM |
3492 | |
3493 | memmove_extent_buffer(right, btrfs_item_nr_offset(push_items), | |
3494 | btrfs_item_nr_offset(0), | |
3495 | right_nritems * sizeof(struct btrfs_item)); | |
3496 | ||
00ec4c51 | 3497 | /* copy the items from left to right */ |
5f39d397 CM |
3498 | copy_extent_buffer(right, left, btrfs_item_nr_offset(0), |
3499 | btrfs_item_nr_offset(left_nritems - push_items), | |
3500 | push_items * sizeof(struct btrfs_item)); | |
00ec4c51 CM |
3501 | |
3502 | /* update the item pointers */ | |
7518a238 | 3503 | right_nritems += push_items; |
5f39d397 | 3504 | btrfs_set_header_nritems(right, right_nritems); |
123abc88 | 3505 | push_space = BTRFS_LEAF_DATA_SIZE(root); |
7518a238 | 3506 | for (i = 0; i < right_nritems; i++) { |
5f39d397 | 3507 | item = btrfs_item_nr(right, i); |
cfed81a0 CM |
3508 | push_space -= btrfs_token_item_size(right, item, &token); |
3509 | btrfs_set_token_item_offset(right, item, push_space, &token); | |
db94535d CM |
3510 | } |
3511 | ||
7518a238 | 3512 | left_nritems -= push_items; |
5f39d397 | 3513 | btrfs_set_header_nritems(left, left_nritems); |
00ec4c51 | 3514 | |
34a38218 CM |
3515 | if (left_nritems) |
3516 | btrfs_mark_buffer_dirty(left); | |
f0486c68 YZ |
3517 | else |
3518 | clean_tree_block(trans, root, left); | |
3519 | ||
5f39d397 | 3520 | btrfs_mark_buffer_dirty(right); |
a429e513 | 3521 | |
5f39d397 CM |
3522 | btrfs_item_key(right, &disk_key, 0); |
3523 | btrfs_set_node_key(upper, &disk_key, slot + 1); | |
d6025579 | 3524 | btrfs_mark_buffer_dirty(upper); |
02217ed2 | 3525 | |
00ec4c51 | 3526 | /* then fixup the leaf pointer in the path */ |
7518a238 CM |
3527 | if (path->slots[0] >= left_nritems) { |
3528 | path->slots[0] -= left_nritems; | |
925baedd CM |
3529 | if (btrfs_header_nritems(path->nodes[0]) == 0) |
3530 | clean_tree_block(trans, root, path->nodes[0]); | |
3531 | btrfs_tree_unlock(path->nodes[0]); | |
5f39d397 CM |
3532 | free_extent_buffer(path->nodes[0]); |
3533 | path->nodes[0] = right; | |
00ec4c51 CM |
3534 | path->slots[1] += 1; |
3535 | } else { | |
925baedd | 3536 | btrfs_tree_unlock(right); |
5f39d397 | 3537 | free_extent_buffer(right); |
00ec4c51 CM |
3538 | } |
3539 | return 0; | |
925baedd CM |
3540 | |
3541 | out_unlock: | |
3542 | btrfs_tree_unlock(right); | |
3543 | free_extent_buffer(right); | |
3544 | return 1; | |
00ec4c51 | 3545 | } |
925baedd | 3546 | |
44871b1b CM |
3547 | /* |
3548 | * push some data in the path leaf to the right, trying to free up at | |
3549 | * least data_size bytes. returns zero if the push worked, nonzero otherwise | |
3550 | * | |
3551 | * returns 1 if the push failed because the other node didn't have enough | |
3552 | * room, 0 if everything worked out and < 0 if there were major errors. | |
99d8f83c CM |
3553 | * |
3554 | * this will push starting from min_slot to the end of the leaf. It won't | |
3555 | * push any slot lower than min_slot | |
44871b1b CM |
3556 | */ |
3557 | static int push_leaf_right(struct btrfs_trans_handle *trans, struct btrfs_root | |
99d8f83c CM |
3558 | *root, struct btrfs_path *path, |
3559 | int min_data_size, int data_size, | |
3560 | int empty, u32 min_slot) | |
44871b1b CM |
3561 | { |
3562 | struct extent_buffer *left = path->nodes[0]; | |
3563 | struct extent_buffer *right; | |
3564 | struct extent_buffer *upper; | |
3565 | int slot; | |
3566 | int free_space; | |
3567 | u32 left_nritems; | |
3568 | int ret; | |
3569 | ||
3570 | if (!path->nodes[1]) | |
3571 | return 1; | |
3572 | ||
3573 | slot = path->slots[1]; | |
3574 | upper = path->nodes[1]; | |
3575 | if (slot >= btrfs_header_nritems(upper) - 1) | |
3576 | return 1; | |
3577 | ||
3578 | btrfs_assert_tree_locked(path->nodes[1]); | |
3579 | ||
3580 | right = read_node_slot(root, upper, slot + 1); | |
91ca338d TI |
3581 | if (right == NULL) |
3582 | return 1; | |
3583 | ||
44871b1b CM |
3584 | btrfs_tree_lock(right); |
3585 | btrfs_set_lock_blocking(right); | |
3586 | ||
3587 | free_space = btrfs_leaf_free_space(root, right); | |
3588 | if (free_space < data_size) | |
3589 | goto out_unlock; | |
3590 | ||
3591 | /* cow and double check */ | |
3592 | ret = btrfs_cow_block(trans, root, right, upper, | |
3593 | slot + 1, &right); | |
3594 | if (ret) | |
3595 | goto out_unlock; | |
3596 | ||
3597 | free_space = btrfs_leaf_free_space(root, right); | |
3598 | if (free_space < data_size) | |
3599 | goto out_unlock; | |
3600 | ||
3601 | left_nritems = btrfs_header_nritems(left); | |
3602 | if (left_nritems == 0) | |
3603 | goto out_unlock; | |
3604 | ||
99d8f83c CM |
3605 | return __push_leaf_right(trans, root, path, min_data_size, empty, |
3606 | right, free_space, left_nritems, min_slot); | |
44871b1b CM |
3607 | out_unlock: |
3608 | btrfs_tree_unlock(right); | |
3609 | free_extent_buffer(right); | |
3610 | return 1; | |
3611 | } | |
3612 | ||
74123bd7 CM |
3613 | /* |
3614 | * push some data in the path leaf to the left, trying to free up at | |
3615 | * least data_size bytes. returns zero if the push worked, nonzero otherwise | |
99d8f83c CM |
3616 | * |
3617 | * max_slot can put a limit on how far into the leaf we'll push items. The | |
3618 | * item at 'max_slot' won't be touched. Use (u32)-1 to make us do all the | |
3619 | * items | |
74123bd7 | 3620 | */ |
44871b1b CM |
3621 | static noinline int __push_leaf_left(struct btrfs_trans_handle *trans, |
3622 | struct btrfs_root *root, | |
3623 | struct btrfs_path *path, int data_size, | |
3624 | int empty, struct extent_buffer *left, | |
99d8f83c CM |
3625 | int free_space, u32 right_nritems, |
3626 | u32 max_slot) | |
be0e5c09 | 3627 | { |
5f39d397 CM |
3628 | struct btrfs_disk_key disk_key; |
3629 | struct extent_buffer *right = path->nodes[0]; | |
be0e5c09 | 3630 | int i; |
be0e5c09 CM |
3631 | int push_space = 0; |
3632 | int push_items = 0; | |
0783fcfc | 3633 | struct btrfs_item *item; |
7518a238 | 3634 | u32 old_left_nritems; |
34a38218 | 3635 | u32 nr; |
aa5d6bed | 3636 | int ret = 0; |
db94535d CM |
3637 | u32 this_item_size; |
3638 | u32 old_left_item_size; | |
cfed81a0 CM |
3639 | struct btrfs_map_token token; |
3640 | ||
3641 | btrfs_init_map_token(&token); | |
be0e5c09 | 3642 | |
34a38218 | 3643 | if (empty) |
99d8f83c | 3644 | nr = min(right_nritems, max_slot); |
34a38218 | 3645 | else |
99d8f83c | 3646 | nr = min(right_nritems - 1, max_slot); |
34a38218 CM |
3647 | |
3648 | for (i = 0; i < nr; i++) { | |
5f39d397 | 3649 | item = btrfs_item_nr(right, i); |
db94535d | 3650 | |
31840ae1 ZY |
3651 | if (!empty && push_items > 0) { |
3652 | if (path->slots[0] < i) | |
3653 | break; | |
3654 | if (path->slots[0] == i) { | |
3655 | int space = btrfs_leaf_free_space(root, right); | |
3656 | if (space + push_space * 2 > free_space) | |
3657 | break; | |
3658 | } | |
3659 | } | |
3660 | ||
be0e5c09 | 3661 | if (path->slots[0] == i) |
87b29b20 | 3662 | push_space += data_size; |
db94535d CM |
3663 | |
3664 | this_item_size = btrfs_item_size(right, item); | |
3665 | if (this_item_size + sizeof(*item) + push_space > free_space) | |
be0e5c09 | 3666 | break; |
db94535d | 3667 | |
be0e5c09 | 3668 | push_items++; |
db94535d CM |
3669 | push_space += this_item_size + sizeof(*item); |
3670 | } | |
3671 | ||
be0e5c09 | 3672 | if (push_items == 0) { |
925baedd CM |
3673 | ret = 1; |
3674 | goto out; | |
be0e5c09 | 3675 | } |
34a38218 | 3676 | if (!empty && push_items == btrfs_header_nritems(right)) |
a429e513 | 3677 | WARN_ON(1); |
5f39d397 | 3678 | |
be0e5c09 | 3679 | /* push data from right to left */ |
5f39d397 CM |
3680 | copy_extent_buffer(left, right, |
3681 | btrfs_item_nr_offset(btrfs_header_nritems(left)), | |
3682 | btrfs_item_nr_offset(0), | |
3683 | push_items * sizeof(struct btrfs_item)); | |
3684 | ||
123abc88 | 3685 | push_space = BTRFS_LEAF_DATA_SIZE(root) - |
d397712b | 3686 | btrfs_item_offset_nr(right, push_items - 1); |
5f39d397 CM |
3687 | |
3688 | copy_extent_buffer(left, right, btrfs_leaf_data(left) + | |
d6025579 CM |
3689 | leaf_data_end(root, left) - push_space, |
3690 | btrfs_leaf_data(right) + | |
5f39d397 | 3691 | btrfs_item_offset_nr(right, push_items - 1), |
d6025579 | 3692 | push_space); |
5f39d397 | 3693 | old_left_nritems = btrfs_header_nritems(left); |
87b29b20 | 3694 | BUG_ON(old_left_nritems <= 0); |
eb60ceac | 3695 | |
db94535d | 3696 | old_left_item_size = btrfs_item_offset_nr(left, old_left_nritems - 1); |
0783fcfc | 3697 | for (i = old_left_nritems; i < old_left_nritems + push_items; i++) { |
5f39d397 | 3698 | u32 ioff; |
db94535d | 3699 | |
5f39d397 | 3700 | item = btrfs_item_nr(left, i); |
db94535d | 3701 | |
cfed81a0 CM |
3702 | ioff = btrfs_token_item_offset(left, item, &token); |
3703 | btrfs_set_token_item_offset(left, item, | |
3704 | ioff - (BTRFS_LEAF_DATA_SIZE(root) - old_left_item_size), | |
3705 | &token); | |
be0e5c09 | 3706 | } |
5f39d397 | 3707 | btrfs_set_header_nritems(left, old_left_nritems + push_items); |
be0e5c09 CM |
3708 | |
3709 | /* fixup right node */ | |
31b1a2bd JL |
3710 | if (push_items > right_nritems) |
3711 | WARN(1, KERN_CRIT "push items %d nr %u\n", push_items, | |
d397712b | 3712 | right_nritems); |
34a38218 CM |
3713 | |
3714 | if (push_items < right_nritems) { | |
3715 | push_space = btrfs_item_offset_nr(right, push_items - 1) - | |
3716 | leaf_data_end(root, right); | |
3717 | memmove_extent_buffer(right, btrfs_leaf_data(right) + | |
3718 | BTRFS_LEAF_DATA_SIZE(root) - push_space, | |
3719 | btrfs_leaf_data(right) + | |
3720 | leaf_data_end(root, right), push_space); | |
3721 | ||
3722 | memmove_extent_buffer(right, btrfs_item_nr_offset(0), | |
5f39d397 CM |
3723 | btrfs_item_nr_offset(push_items), |
3724 | (btrfs_header_nritems(right) - push_items) * | |
3725 | sizeof(struct btrfs_item)); | |
34a38218 | 3726 | } |
eef1c494 Y |
3727 | right_nritems -= push_items; |
3728 | btrfs_set_header_nritems(right, right_nritems); | |
123abc88 | 3729 | push_space = BTRFS_LEAF_DATA_SIZE(root); |
5f39d397 CM |
3730 | for (i = 0; i < right_nritems; i++) { |
3731 | item = btrfs_item_nr(right, i); | |
db94535d | 3732 | |
cfed81a0 CM |
3733 | push_space = push_space - btrfs_token_item_size(right, |
3734 | item, &token); | |
3735 | btrfs_set_token_item_offset(right, item, push_space, &token); | |
db94535d | 3736 | } |
eb60ceac | 3737 | |
5f39d397 | 3738 | btrfs_mark_buffer_dirty(left); |
34a38218 CM |
3739 | if (right_nritems) |
3740 | btrfs_mark_buffer_dirty(right); | |
f0486c68 YZ |
3741 | else |
3742 | clean_tree_block(trans, root, right); | |
098f59c2 | 3743 | |
5f39d397 | 3744 | btrfs_item_key(right, &disk_key, 0); |
d6a0a126 | 3745 | fixup_low_keys(root, path, &disk_key, 1); |
be0e5c09 CM |
3746 | |
3747 | /* then fixup the leaf pointer in the path */ | |
3748 | if (path->slots[0] < push_items) { | |
3749 | path->slots[0] += old_left_nritems; | |
925baedd | 3750 | btrfs_tree_unlock(path->nodes[0]); |
5f39d397 CM |
3751 | free_extent_buffer(path->nodes[0]); |
3752 | path->nodes[0] = left; | |
be0e5c09 CM |
3753 | path->slots[1] -= 1; |
3754 | } else { | |
925baedd | 3755 | btrfs_tree_unlock(left); |
5f39d397 | 3756 | free_extent_buffer(left); |
be0e5c09 CM |
3757 | path->slots[0] -= push_items; |
3758 | } | |
eb60ceac | 3759 | BUG_ON(path->slots[0] < 0); |
aa5d6bed | 3760 | return ret; |
925baedd CM |
3761 | out: |
3762 | btrfs_tree_unlock(left); | |
3763 | free_extent_buffer(left); | |
3764 | return ret; | |
be0e5c09 CM |
3765 | } |
3766 | ||
44871b1b CM |
3767 | /* |
3768 | * push some data in the path leaf to the left, trying to free up at | |
3769 | * least data_size bytes. returns zero if the push worked, nonzero otherwise | |
99d8f83c CM |
3770 | * |
3771 | * max_slot can put a limit on how far into the leaf we'll push items. The | |
3772 | * item at 'max_slot' won't be touched. Use (u32)-1 to make us push all the | |
3773 | * items | |
44871b1b CM |
3774 | */ |
3775 | static int push_leaf_left(struct btrfs_trans_handle *trans, struct btrfs_root | |
99d8f83c CM |
3776 | *root, struct btrfs_path *path, int min_data_size, |
3777 | int data_size, int empty, u32 max_slot) | |
44871b1b CM |
3778 | { |
3779 | struct extent_buffer *right = path->nodes[0]; | |
3780 | struct extent_buffer *left; | |
3781 | int slot; | |
3782 | int free_space; | |
3783 | u32 right_nritems; | |
3784 | int ret = 0; | |
3785 | ||
3786 | slot = path->slots[1]; | |
3787 | if (slot == 0) | |
3788 | return 1; | |
3789 | if (!path->nodes[1]) | |
3790 | return 1; | |
3791 | ||
3792 | right_nritems = btrfs_header_nritems(right); | |
3793 | if (right_nritems == 0) | |
3794 | return 1; | |
3795 | ||
3796 | btrfs_assert_tree_locked(path->nodes[1]); | |
3797 | ||
3798 | left = read_node_slot(root, path->nodes[1], slot - 1); | |
91ca338d TI |
3799 | if (left == NULL) |
3800 | return 1; | |
3801 | ||
44871b1b CM |
3802 | btrfs_tree_lock(left); |
3803 | btrfs_set_lock_blocking(left); | |
3804 | ||
3805 | free_space = btrfs_leaf_free_space(root, left); | |
3806 | if (free_space < data_size) { | |
3807 | ret = 1; | |
3808 | goto out; | |
3809 | } | |
3810 | ||
3811 | /* cow and double check */ | |
3812 | ret = btrfs_cow_block(trans, root, left, | |
3813 | path->nodes[1], slot - 1, &left); | |
3814 | if (ret) { | |
3815 | /* we hit -ENOSPC, but it isn't fatal here */ | |
79787eaa JM |
3816 | if (ret == -ENOSPC) |
3817 | ret = 1; | |
44871b1b CM |
3818 | goto out; |
3819 | } | |
3820 | ||
3821 | free_space = btrfs_leaf_free_space(root, left); | |
3822 | if (free_space < data_size) { | |
3823 | ret = 1; | |
3824 | goto out; | |
3825 | } | |
3826 | ||
99d8f83c CM |
3827 | return __push_leaf_left(trans, root, path, min_data_size, |
3828 | empty, left, free_space, right_nritems, | |
3829 | max_slot); | |
44871b1b CM |
3830 | out: |
3831 | btrfs_tree_unlock(left); | |
3832 | free_extent_buffer(left); | |
3833 | return ret; | |
3834 | } | |
3835 | ||
3836 | /* | |
3837 | * split the path's leaf in two, making sure there is at least data_size | |
3838 | * available for the resulting leaf level of the path. | |
44871b1b | 3839 | */ |
143bede5 JM |
3840 | static noinline void copy_for_split(struct btrfs_trans_handle *trans, |
3841 | struct btrfs_root *root, | |
3842 | struct btrfs_path *path, | |
3843 | struct extent_buffer *l, | |
3844 | struct extent_buffer *right, | |
3845 | int slot, int mid, int nritems) | |
44871b1b CM |
3846 | { |
3847 | int data_copy_size; | |
3848 | int rt_data_off; | |
3849 | int i; | |
44871b1b | 3850 | struct btrfs_disk_key disk_key; |
cfed81a0 CM |
3851 | struct btrfs_map_token token; |
3852 | ||
3853 | btrfs_init_map_token(&token); | |
44871b1b CM |
3854 | |
3855 | nritems = nritems - mid; | |
3856 | btrfs_set_header_nritems(right, nritems); | |
3857 | data_copy_size = btrfs_item_end_nr(l, mid) - leaf_data_end(root, l); | |
3858 | ||
3859 | copy_extent_buffer(right, l, btrfs_item_nr_offset(0), | |
3860 | btrfs_item_nr_offset(mid), | |
3861 | nritems * sizeof(struct btrfs_item)); | |
3862 | ||
3863 | copy_extent_buffer(right, l, | |
3864 | btrfs_leaf_data(right) + BTRFS_LEAF_DATA_SIZE(root) - | |
3865 | data_copy_size, btrfs_leaf_data(l) + | |
3866 | leaf_data_end(root, l), data_copy_size); | |
3867 | ||
3868 | rt_data_off = BTRFS_LEAF_DATA_SIZE(root) - | |
3869 | btrfs_item_end_nr(l, mid); | |
3870 | ||
3871 | for (i = 0; i < nritems; i++) { | |
3872 | struct btrfs_item *item = btrfs_item_nr(right, i); | |
3873 | u32 ioff; | |
3874 | ||
cfed81a0 CM |
3875 | ioff = btrfs_token_item_offset(right, item, &token); |
3876 | btrfs_set_token_item_offset(right, item, | |
3877 | ioff + rt_data_off, &token); | |
44871b1b CM |
3878 | } |
3879 | ||
44871b1b | 3880 | btrfs_set_header_nritems(l, mid); |
44871b1b | 3881 | btrfs_item_key(right, &disk_key, 0); |
143bede5 | 3882 | insert_ptr(trans, root, path, &disk_key, right->start, |
c3e06965 | 3883 | path->slots[1] + 1, 1); |
44871b1b CM |
3884 | |
3885 | btrfs_mark_buffer_dirty(right); | |
3886 | btrfs_mark_buffer_dirty(l); | |
3887 | BUG_ON(path->slots[0] != slot); | |
3888 | ||
44871b1b CM |
3889 | if (mid <= slot) { |
3890 | btrfs_tree_unlock(path->nodes[0]); | |
3891 | free_extent_buffer(path->nodes[0]); | |
3892 | path->nodes[0] = right; | |
3893 | path->slots[0] -= mid; | |
3894 | path->slots[1] += 1; | |
3895 | } else { | |
3896 | btrfs_tree_unlock(right); | |
3897 | free_extent_buffer(right); | |
3898 | } | |
3899 | ||
3900 | BUG_ON(path->slots[0] < 0); | |
44871b1b CM |
3901 | } |
3902 | ||
99d8f83c CM |
3903 | /* |
3904 | * double splits happen when we need to insert a big item in the middle | |
3905 | * of a leaf. A double split can leave us with 3 mostly empty leaves: | |
3906 | * leaf: [ slots 0 - N] [ our target ] [ N + 1 - total in leaf ] | |
3907 | * A B C | |
3908 | * | |
3909 | * We avoid this by trying to push the items on either side of our target | |
3910 | * into the adjacent leaves. If all goes well we can avoid the double split | |
3911 | * completely. | |
3912 | */ | |
3913 | static noinline int push_for_double_split(struct btrfs_trans_handle *trans, | |
3914 | struct btrfs_root *root, | |
3915 | struct btrfs_path *path, | |
3916 | int data_size) | |
3917 | { | |
3918 | int ret; | |
3919 | int progress = 0; | |
3920 | int slot; | |
3921 | u32 nritems; | |
3922 | ||
3923 | slot = path->slots[0]; | |
3924 | ||
3925 | /* | |
3926 | * try to push all the items after our slot into the | |
3927 | * right leaf | |
3928 | */ | |
3929 | ret = push_leaf_right(trans, root, path, 1, data_size, 0, slot); | |
3930 | if (ret < 0) | |
3931 | return ret; | |
3932 | ||
3933 | if (ret == 0) | |
3934 | progress++; | |
3935 | ||
3936 | nritems = btrfs_header_nritems(path->nodes[0]); | |
3937 | /* | |
3938 | * our goal is to get our slot at the start or end of a leaf. If | |
3939 | * we've done so we're done | |
3940 | */ | |
3941 | if (path->slots[0] == 0 || path->slots[0] == nritems) | |
3942 | return 0; | |
3943 | ||
3944 | if (btrfs_leaf_free_space(root, path->nodes[0]) >= data_size) | |
3945 | return 0; | |
3946 | ||
3947 | /* try to push all the items before our slot into the next leaf */ | |
3948 | slot = path->slots[0]; | |
3949 | ret = push_leaf_left(trans, root, path, 1, data_size, 0, slot); | |
3950 | if (ret < 0) | |
3951 | return ret; | |
3952 | ||
3953 | if (ret == 0) | |
3954 | progress++; | |
3955 | ||
3956 | if (progress) | |
3957 | return 0; | |
3958 | return 1; | |
3959 | } | |
3960 | ||
74123bd7 CM |
3961 | /* |
3962 | * split the path's leaf in two, making sure there is at least data_size | |
3963 | * available for the resulting leaf level of the path. | |
aa5d6bed CM |
3964 | * |
3965 | * returns 0 if all went well and < 0 on failure. | |
74123bd7 | 3966 | */ |
e02119d5 CM |
3967 | static noinline int split_leaf(struct btrfs_trans_handle *trans, |
3968 | struct btrfs_root *root, | |
3969 | struct btrfs_key *ins_key, | |
3970 | struct btrfs_path *path, int data_size, | |
3971 | int extend) | |
be0e5c09 | 3972 | { |
5d4f98a2 | 3973 | struct btrfs_disk_key disk_key; |
5f39d397 | 3974 | struct extent_buffer *l; |
7518a238 | 3975 | u32 nritems; |
eb60ceac CM |
3976 | int mid; |
3977 | int slot; | |
5f39d397 | 3978 | struct extent_buffer *right; |
d4dbff95 | 3979 | int ret = 0; |
aa5d6bed | 3980 | int wret; |
5d4f98a2 | 3981 | int split; |
cc0c5538 | 3982 | int num_doubles = 0; |
99d8f83c | 3983 | int tried_avoid_double = 0; |
aa5d6bed | 3984 | |
a5719521 YZ |
3985 | l = path->nodes[0]; |
3986 | slot = path->slots[0]; | |
3987 | if (extend && data_size + btrfs_item_size_nr(l, slot) + | |
3988 | sizeof(struct btrfs_item) > BTRFS_LEAF_DATA_SIZE(root)) | |
3989 | return -EOVERFLOW; | |
3990 | ||
40689478 | 3991 | /* first try to make some room by pushing left and right */ |
99d8f83c CM |
3992 | if (data_size) { |
3993 | wret = push_leaf_right(trans, root, path, data_size, | |
3994 | data_size, 0, 0); | |
d397712b | 3995 | if (wret < 0) |
eaee50e8 | 3996 | return wret; |
3685f791 | 3997 | if (wret) { |
99d8f83c CM |
3998 | wret = push_leaf_left(trans, root, path, data_size, |
3999 | data_size, 0, (u32)-1); | |
3685f791 CM |
4000 | if (wret < 0) |
4001 | return wret; | |
4002 | } | |
4003 | l = path->nodes[0]; | |
aa5d6bed | 4004 | |
3685f791 | 4005 | /* did the pushes work? */ |
87b29b20 | 4006 | if (btrfs_leaf_free_space(root, l) >= data_size) |
3685f791 | 4007 | return 0; |
3326d1b0 | 4008 | } |
aa5d6bed | 4009 | |
5c680ed6 | 4010 | if (!path->nodes[1]) { |
90f8d62e | 4011 | ret = insert_new_root(trans, root, path, 1, 1); |
5c680ed6 CM |
4012 | if (ret) |
4013 | return ret; | |
4014 | } | |
cc0c5538 | 4015 | again: |
5d4f98a2 | 4016 | split = 1; |
cc0c5538 | 4017 | l = path->nodes[0]; |
eb60ceac | 4018 | slot = path->slots[0]; |
5f39d397 | 4019 | nritems = btrfs_header_nritems(l); |
d397712b | 4020 | mid = (nritems + 1) / 2; |
54aa1f4d | 4021 | |
5d4f98a2 YZ |
4022 | if (mid <= slot) { |
4023 | if (nritems == 1 || | |
4024 | leaf_space_used(l, mid, nritems - mid) + data_size > | |
4025 | BTRFS_LEAF_DATA_SIZE(root)) { | |
4026 | if (slot >= nritems) { | |
4027 | split = 0; | |
4028 | } else { | |
4029 | mid = slot; | |
4030 | if (mid != nritems && | |
4031 | leaf_space_used(l, mid, nritems - mid) + | |
4032 | data_size > BTRFS_LEAF_DATA_SIZE(root)) { | |
99d8f83c CM |
4033 | if (data_size && !tried_avoid_double) |
4034 | goto push_for_double; | |
5d4f98a2 YZ |
4035 | split = 2; |
4036 | } | |
4037 | } | |
4038 | } | |
4039 | } else { | |
4040 | if (leaf_space_used(l, 0, mid) + data_size > | |
4041 | BTRFS_LEAF_DATA_SIZE(root)) { | |
4042 | if (!extend && data_size && slot == 0) { | |
4043 | split = 0; | |
4044 | } else if ((extend || !data_size) && slot == 0) { | |
4045 | mid = 1; | |
4046 | } else { | |
4047 | mid = slot; | |
4048 | if (mid != nritems && | |
4049 | leaf_space_used(l, mid, nritems - mid) + | |
4050 | data_size > BTRFS_LEAF_DATA_SIZE(root)) { | |
99d8f83c CM |
4051 | if (data_size && !tried_avoid_double) |
4052 | goto push_for_double; | |
5d4f98a2 YZ |
4053 | split = 2 ; |
4054 | } | |
4055 | } | |
4056 | } | |
4057 | } | |
4058 | ||
4059 | if (split == 0) | |
4060 | btrfs_cpu_key_to_disk(&disk_key, ins_key); | |
4061 | else | |
4062 | btrfs_item_key(l, &disk_key, mid); | |
4063 | ||
4064 | right = btrfs_alloc_free_block(trans, root, root->leafsize, 0, | |
31840ae1 | 4065 | root->root_key.objectid, |
5581a51a | 4066 | &disk_key, 0, l->start, 0); |
f0486c68 | 4067 | if (IS_ERR(right)) |
5f39d397 | 4068 | return PTR_ERR(right); |
f0486c68 YZ |
4069 | |
4070 | root_add_used(root, root->leafsize); | |
5f39d397 CM |
4071 | |
4072 | memset_extent_buffer(right, 0, 0, sizeof(struct btrfs_header)); | |
db94535d | 4073 | btrfs_set_header_bytenr(right, right->start); |
5f39d397 | 4074 | btrfs_set_header_generation(right, trans->transid); |
5d4f98a2 | 4075 | btrfs_set_header_backref_rev(right, BTRFS_MIXED_BACKREF_REV); |
5f39d397 CM |
4076 | btrfs_set_header_owner(right, root->root_key.objectid); |
4077 | btrfs_set_header_level(right, 0); | |
4078 | write_extent_buffer(right, root->fs_info->fsid, | |
4079 | (unsigned long)btrfs_header_fsid(right), | |
4080 | BTRFS_FSID_SIZE); | |
e17cade2 CM |
4081 | |
4082 | write_extent_buffer(right, root->fs_info->chunk_tree_uuid, | |
4083 | (unsigned long)btrfs_header_chunk_tree_uuid(right), | |
4084 | BTRFS_UUID_SIZE); | |
44871b1b | 4085 | |
5d4f98a2 YZ |
4086 | if (split == 0) { |
4087 | if (mid <= slot) { | |
4088 | btrfs_set_header_nritems(right, 0); | |
143bede5 | 4089 | insert_ptr(trans, root, path, &disk_key, right->start, |
c3e06965 | 4090 | path->slots[1] + 1, 1); |
5d4f98a2 YZ |
4091 | btrfs_tree_unlock(path->nodes[0]); |
4092 | free_extent_buffer(path->nodes[0]); | |
4093 | path->nodes[0] = right; | |
4094 | path->slots[0] = 0; | |
4095 | path->slots[1] += 1; | |
4096 | } else { | |
4097 | btrfs_set_header_nritems(right, 0); | |
143bede5 | 4098 | insert_ptr(trans, root, path, &disk_key, right->start, |
c3e06965 | 4099 | path->slots[1], 1); |
5d4f98a2 YZ |
4100 | btrfs_tree_unlock(path->nodes[0]); |
4101 | free_extent_buffer(path->nodes[0]); | |
4102 | path->nodes[0] = right; | |
4103 | path->slots[0] = 0; | |
143bede5 | 4104 | if (path->slots[1] == 0) |
d6a0a126 | 4105 | fixup_low_keys(root, path, &disk_key, 1); |
d4dbff95 | 4106 | } |
5d4f98a2 YZ |
4107 | btrfs_mark_buffer_dirty(right); |
4108 | return ret; | |
d4dbff95 | 4109 | } |
74123bd7 | 4110 | |
143bede5 | 4111 | copy_for_split(trans, root, path, l, right, slot, mid, nritems); |
31840ae1 | 4112 | |
5d4f98a2 | 4113 | if (split == 2) { |
cc0c5538 CM |
4114 | BUG_ON(num_doubles != 0); |
4115 | num_doubles++; | |
4116 | goto again; | |
a429e513 | 4117 | } |
44871b1b | 4118 | |
143bede5 | 4119 | return 0; |
99d8f83c CM |
4120 | |
4121 | push_for_double: | |
4122 | push_for_double_split(trans, root, path, data_size); | |
4123 | tried_avoid_double = 1; | |
4124 | if (btrfs_leaf_free_space(root, path->nodes[0]) >= data_size) | |
4125 | return 0; | |
4126 | goto again; | |
be0e5c09 CM |
4127 | } |
4128 | ||
ad48fd75 YZ |
4129 | static noinline int setup_leaf_for_split(struct btrfs_trans_handle *trans, |
4130 | struct btrfs_root *root, | |
4131 | struct btrfs_path *path, int ins_len) | |
459931ec | 4132 | { |
ad48fd75 | 4133 | struct btrfs_key key; |
459931ec | 4134 | struct extent_buffer *leaf; |
ad48fd75 YZ |
4135 | struct btrfs_file_extent_item *fi; |
4136 | u64 extent_len = 0; | |
4137 | u32 item_size; | |
4138 | int ret; | |
459931ec CM |
4139 | |
4140 | leaf = path->nodes[0]; | |
ad48fd75 YZ |
4141 | btrfs_item_key_to_cpu(leaf, &key, path->slots[0]); |
4142 | ||
4143 | BUG_ON(key.type != BTRFS_EXTENT_DATA_KEY && | |
4144 | key.type != BTRFS_EXTENT_CSUM_KEY); | |
4145 | ||
4146 | if (btrfs_leaf_free_space(root, leaf) >= ins_len) | |
4147 | return 0; | |
459931ec CM |
4148 | |
4149 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); | |
ad48fd75 YZ |
4150 | if (key.type == BTRFS_EXTENT_DATA_KEY) { |
4151 | fi = btrfs_item_ptr(leaf, path->slots[0], | |
4152 | struct btrfs_file_extent_item); | |
4153 | extent_len = btrfs_file_extent_num_bytes(leaf, fi); | |
4154 | } | |
b3b4aa74 | 4155 | btrfs_release_path(path); |
459931ec | 4156 | |
459931ec | 4157 | path->keep_locks = 1; |
ad48fd75 YZ |
4158 | path->search_for_split = 1; |
4159 | ret = btrfs_search_slot(trans, root, &key, path, 0, 1); | |
459931ec | 4160 | path->search_for_split = 0; |
ad48fd75 YZ |
4161 | if (ret < 0) |
4162 | goto err; | |
459931ec | 4163 | |
ad48fd75 YZ |
4164 | ret = -EAGAIN; |
4165 | leaf = path->nodes[0]; | |
459931ec | 4166 | /* if our item isn't there or got smaller, return now */ |
ad48fd75 YZ |
4167 | if (ret > 0 || item_size != btrfs_item_size_nr(leaf, path->slots[0])) |
4168 | goto err; | |
4169 | ||
109f6aef CM |
4170 | /* the leaf has changed, it now has room. return now */ |
4171 | if (btrfs_leaf_free_space(root, path->nodes[0]) >= ins_len) | |
4172 | goto err; | |
4173 | ||
ad48fd75 YZ |
4174 | if (key.type == BTRFS_EXTENT_DATA_KEY) { |
4175 | fi = btrfs_item_ptr(leaf, path->slots[0], | |
4176 | struct btrfs_file_extent_item); | |
4177 | if (extent_len != btrfs_file_extent_num_bytes(leaf, fi)) | |
4178 | goto err; | |
459931ec CM |
4179 | } |
4180 | ||
b9473439 | 4181 | btrfs_set_path_blocking(path); |
ad48fd75 | 4182 | ret = split_leaf(trans, root, &key, path, ins_len, 1); |
f0486c68 YZ |
4183 | if (ret) |
4184 | goto err; | |
459931ec | 4185 | |
ad48fd75 | 4186 | path->keep_locks = 0; |
b9473439 | 4187 | btrfs_unlock_up_safe(path, 1); |
ad48fd75 YZ |
4188 | return 0; |
4189 | err: | |
4190 | path->keep_locks = 0; | |
4191 | return ret; | |
4192 | } | |
4193 | ||
4194 | static noinline int split_item(struct btrfs_trans_handle *trans, | |
4195 | struct btrfs_root *root, | |
4196 | struct btrfs_path *path, | |
4197 | struct btrfs_key *new_key, | |
4198 | unsigned long split_offset) | |
4199 | { | |
4200 | struct extent_buffer *leaf; | |
4201 | struct btrfs_item *item; | |
4202 | struct btrfs_item *new_item; | |
4203 | int slot; | |
4204 | char *buf; | |
4205 | u32 nritems; | |
4206 | u32 item_size; | |
4207 | u32 orig_offset; | |
4208 | struct btrfs_disk_key disk_key; | |
4209 | ||
b9473439 CM |
4210 | leaf = path->nodes[0]; |
4211 | BUG_ON(btrfs_leaf_free_space(root, leaf) < sizeof(struct btrfs_item)); | |
4212 | ||
b4ce94de CM |
4213 | btrfs_set_path_blocking(path); |
4214 | ||
459931ec CM |
4215 | item = btrfs_item_nr(leaf, path->slots[0]); |
4216 | orig_offset = btrfs_item_offset(leaf, item); | |
4217 | item_size = btrfs_item_size(leaf, item); | |
4218 | ||
459931ec | 4219 | buf = kmalloc(item_size, GFP_NOFS); |
ad48fd75 YZ |
4220 | if (!buf) |
4221 | return -ENOMEM; | |
4222 | ||
459931ec CM |
4223 | read_extent_buffer(leaf, buf, btrfs_item_ptr_offset(leaf, |
4224 | path->slots[0]), item_size); | |
459931ec | 4225 | |
ad48fd75 | 4226 | slot = path->slots[0] + 1; |
459931ec | 4227 | nritems = btrfs_header_nritems(leaf); |
459931ec CM |
4228 | if (slot != nritems) { |
4229 | /* shift the items */ | |
4230 | memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + 1), | |
ad48fd75 YZ |
4231 | btrfs_item_nr_offset(slot), |
4232 | (nritems - slot) * sizeof(struct btrfs_item)); | |
459931ec CM |
4233 | } |
4234 | ||
4235 | btrfs_cpu_key_to_disk(&disk_key, new_key); | |
4236 | btrfs_set_item_key(leaf, &disk_key, slot); | |
4237 | ||
4238 | new_item = btrfs_item_nr(leaf, slot); | |
4239 | ||
4240 | btrfs_set_item_offset(leaf, new_item, orig_offset); | |
4241 | btrfs_set_item_size(leaf, new_item, item_size - split_offset); | |
4242 | ||
4243 | btrfs_set_item_offset(leaf, item, | |
4244 | orig_offset + item_size - split_offset); | |
4245 | btrfs_set_item_size(leaf, item, split_offset); | |
4246 | ||
4247 | btrfs_set_header_nritems(leaf, nritems + 1); | |
4248 | ||
4249 | /* write the data for the start of the original item */ | |
4250 | write_extent_buffer(leaf, buf, | |
4251 | btrfs_item_ptr_offset(leaf, path->slots[0]), | |
4252 | split_offset); | |
4253 | ||
4254 | /* write the data for the new item */ | |
4255 | write_extent_buffer(leaf, buf + split_offset, | |
4256 | btrfs_item_ptr_offset(leaf, slot), | |
4257 | item_size - split_offset); | |
4258 | btrfs_mark_buffer_dirty(leaf); | |
4259 | ||
ad48fd75 | 4260 | BUG_ON(btrfs_leaf_free_space(root, leaf) < 0); |
459931ec | 4261 | kfree(buf); |
ad48fd75 YZ |
4262 | return 0; |
4263 | } | |
4264 | ||
4265 | /* | |
4266 | * This function splits a single item into two items, | |
4267 | * giving 'new_key' to the new item and splitting the | |
4268 | * old one at split_offset (from the start of the item). | |
4269 | * | |
4270 | * The path may be released by this operation. After | |
4271 | * the split, the path is pointing to the old item. The | |
4272 | * new item is going to be in the same node as the old one. | |
4273 | * | |
4274 | * Note, the item being split must be smaller enough to live alone on | |
4275 | * a tree block with room for one extra struct btrfs_item | |
4276 | * | |
4277 | * This allows us to split the item in place, keeping a lock on the | |
4278 | * leaf the entire time. | |
4279 | */ | |
4280 | int btrfs_split_item(struct btrfs_trans_handle *trans, | |
4281 | struct btrfs_root *root, | |
4282 | struct btrfs_path *path, | |
4283 | struct btrfs_key *new_key, | |
4284 | unsigned long split_offset) | |
4285 | { | |
4286 | int ret; | |
4287 | ret = setup_leaf_for_split(trans, root, path, | |
4288 | sizeof(struct btrfs_item)); | |
4289 | if (ret) | |
4290 | return ret; | |
4291 | ||
4292 | ret = split_item(trans, root, path, new_key, split_offset); | |
459931ec CM |
4293 | return ret; |
4294 | } | |
4295 | ||
ad48fd75 YZ |
4296 | /* |
4297 | * This function duplicate a item, giving 'new_key' to the new item. | |
4298 | * It guarantees both items live in the same tree leaf and the new item | |
4299 | * is contiguous with the original item. | |
4300 | * | |
4301 | * This allows us to split file extent in place, keeping a lock on the | |
4302 | * leaf the entire time. | |
4303 | */ | |
4304 | int btrfs_duplicate_item(struct btrfs_trans_handle *trans, | |
4305 | struct btrfs_root *root, | |
4306 | struct btrfs_path *path, | |
4307 | struct btrfs_key *new_key) | |
4308 | { | |
4309 | struct extent_buffer *leaf; | |
4310 | int ret; | |
4311 | u32 item_size; | |
4312 | ||
4313 | leaf = path->nodes[0]; | |
4314 | item_size = btrfs_item_size_nr(leaf, path->slots[0]); | |
4315 | ret = setup_leaf_for_split(trans, root, path, | |
4316 | item_size + sizeof(struct btrfs_item)); | |
4317 | if (ret) | |
4318 | return ret; | |
4319 | ||
4320 | path->slots[0]++; | |
afe5fea7 | 4321 | setup_items_for_insert(root, path, new_key, &item_size, |
143bede5 JM |
4322 | item_size, item_size + |
4323 | sizeof(struct btrfs_item), 1); | |
ad48fd75 YZ |
4324 | leaf = path->nodes[0]; |
4325 | memcpy_extent_buffer(leaf, | |
4326 | btrfs_item_ptr_offset(leaf, path->slots[0]), | |
4327 | btrfs_item_ptr_offset(leaf, path->slots[0] - 1), | |
4328 | item_size); | |
4329 | return 0; | |
4330 | } | |
4331 | ||
d352ac68 CM |
4332 | /* |
4333 | * make the item pointed to by the path smaller. new_size indicates | |
4334 | * how small to make it, and from_end tells us if we just chop bytes | |
4335 | * off the end of the item or if we shift the item to chop bytes off | |
4336 | * the front. | |
4337 | */ | |
afe5fea7 | 4338 | void btrfs_truncate_item(struct btrfs_root *root, struct btrfs_path *path, |
143bede5 | 4339 | u32 new_size, int from_end) |
b18c6685 | 4340 | { |
b18c6685 | 4341 | int slot; |
5f39d397 CM |
4342 | struct extent_buffer *leaf; |
4343 | struct btrfs_item *item; | |
b18c6685 CM |
4344 | u32 nritems; |
4345 | unsigned int data_end; | |
4346 | unsigned int old_data_start; | |
4347 | unsigned int old_size; | |
4348 | unsigned int size_diff; | |
4349 | int i; | |
cfed81a0 CM |
4350 | struct btrfs_map_token token; |
4351 | ||
4352 | btrfs_init_map_token(&token); | |
b18c6685 | 4353 | |
5f39d397 | 4354 | leaf = path->nodes[0]; |
179e29e4 CM |
4355 | slot = path->slots[0]; |
4356 | ||
4357 | old_size = btrfs_item_size_nr(leaf, slot); | |
4358 | if (old_size == new_size) | |
143bede5 | 4359 | return; |
b18c6685 | 4360 | |
5f39d397 | 4361 | nritems = btrfs_header_nritems(leaf); |
b18c6685 CM |
4362 | data_end = leaf_data_end(root, leaf); |
4363 | ||
5f39d397 | 4364 | old_data_start = btrfs_item_offset_nr(leaf, slot); |
179e29e4 | 4365 | |
b18c6685 CM |
4366 | size_diff = old_size - new_size; |
4367 | ||
4368 | BUG_ON(slot < 0); | |
4369 | BUG_ON(slot >= nritems); | |
4370 | ||
4371 | /* | |
4372 | * item0..itemN ... dataN.offset..dataN.size .. data0.size | |
4373 | */ | |
4374 | /* first correct the data pointers */ | |
4375 | for (i = slot; i < nritems; i++) { | |
5f39d397 CM |
4376 | u32 ioff; |
4377 | item = btrfs_item_nr(leaf, i); | |
db94535d | 4378 | |
cfed81a0 CM |
4379 | ioff = btrfs_token_item_offset(leaf, item, &token); |
4380 | btrfs_set_token_item_offset(leaf, item, | |
4381 | ioff + size_diff, &token); | |
b18c6685 | 4382 | } |
db94535d | 4383 | |
b18c6685 | 4384 | /* shift the data */ |
179e29e4 CM |
4385 | if (from_end) { |
4386 | memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) + | |
4387 | data_end + size_diff, btrfs_leaf_data(leaf) + | |
4388 | data_end, old_data_start + new_size - data_end); | |
4389 | } else { | |
4390 | struct btrfs_disk_key disk_key; | |
4391 | u64 offset; | |
4392 | ||
4393 | btrfs_item_key(leaf, &disk_key, slot); | |
4394 | ||
4395 | if (btrfs_disk_key_type(&disk_key) == BTRFS_EXTENT_DATA_KEY) { | |
4396 | unsigned long ptr; | |
4397 | struct btrfs_file_extent_item *fi; | |
4398 | ||
4399 | fi = btrfs_item_ptr(leaf, slot, | |
4400 | struct btrfs_file_extent_item); | |
4401 | fi = (struct btrfs_file_extent_item *)( | |
4402 | (unsigned long)fi - size_diff); | |
4403 | ||
4404 | if (btrfs_file_extent_type(leaf, fi) == | |
4405 | BTRFS_FILE_EXTENT_INLINE) { | |
4406 | ptr = btrfs_item_ptr_offset(leaf, slot); | |
4407 | memmove_extent_buffer(leaf, ptr, | |
d397712b CM |
4408 | (unsigned long)fi, |
4409 | offsetof(struct btrfs_file_extent_item, | |
179e29e4 CM |
4410 | disk_bytenr)); |
4411 | } | |
4412 | } | |
4413 | ||
4414 | memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) + | |
4415 | data_end + size_diff, btrfs_leaf_data(leaf) + | |
4416 | data_end, old_data_start - data_end); | |
4417 | ||
4418 | offset = btrfs_disk_key_offset(&disk_key); | |
4419 | btrfs_set_disk_key_offset(&disk_key, offset + size_diff); | |
4420 | btrfs_set_item_key(leaf, &disk_key, slot); | |
4421 | if (slot == 0) | |
d6a0a126 | 4422 | fixup_low_keys(root, path, &disk_key, 1); |
179e29e4 | 4423 | } |
5f39d397 CM |
4424 | |
4425 | item = btrfs_item_nr(leaf, slot); | |
4426 | btrfs_set_item_size(leaf, item, new_size); | |
4427 | btrfs_mark_buffer_dirty(leaf); | |
b18c6685 | 4428 | |
5f39d397 CM |
4429 | if (btrfs_leaf_free_space(root, leaf) < 0) { |
4430 | btrfs_print_leaf(root, leaf); | |
b18c6685 | 4431 | BUG(); |
5f39d397 | 4432 | } |
b18c6685 CM |
4433 | } |
4434 | ||
d352ac68 CM |
4435 | /* |
4436 | * make the item pointed to by the path bigger, data_size is the new size. | |
4437 | */ | |
4b90c680 | 4438 | void btrfs_extend_item(struct btrfs_root *root, struct btrfs_path *path, |
143bede5 | 4439 | u32 data_size) |
6567e837 | 4440 | { |
6567e837 | 4441 | int slot; |
5f39d397 CM |
4442 | struct extent_buffer *leaf; |
4443 | struct btrfs_item *item; | |
6567e837 CM |
4444 | u32 nritems; |
4445 | unsigned int data_end; | |
4446 | unsigned int old_data; | |
4447 | unsigned int old_size; | |
4448 | int i; | |
cfed81a0 CM |
4449 | struct btrfs_map_token token; |
4450 | ||
4451 | btrfs_init_map_token(&token); | |
6567e837 | 4452 | |
5f39d397 | 4453 | leaf = path->nodes[0]; |
6567e837 | 4454 | |
5f39d397 | 4455 | nritems = btrfs_header_nritems(leaf); |
6567e837 CM |
4456 | data_end = leaf_data_end(root, leaf); |
4457 | ||
5f39d397 CM |
4458 | if (btrfs_leaf_free_space(root, leaf) < data_size) { |
4459 | btrfs_print_leaf(root, leaf); | |
6567e837 | 4460 | BUG(); |
5f39d397 | 4461 | } |
6567e837 | 4462 | slot = path->slots[0]; |
5f39d397 | 4463 | old_data = btrfs_item_end_nr(leaf, slot); |
6567e837 CM |
4464 | |
4465 | BUG_ON(slot < 0); | |
3326d1b0 CM |
4466 | if (slot >= nritems) { |
4467 | btrfs_print_leaf(root, leaf); | |
d397712b CM |
4468 | printk(KERN_CRIT "slot %d too large, nritems %d\n", |
4469 | slot, nritems); | |
3326d1b0 CM |
4470 | BUG_ON(1); |
4471 | } | |
6567e837 CM |
4472 | |
4473 | /* | |
4474 | * item0..itemN ... dataN.offset..dataN.size .. data0.size | |
4475 | */ | |
4476 | /* first correct the data pointers */ | |
4477 | for (i = slot; i < nritems; i++) { | |
5f39d397 CM |
4478 | u32 ioff; |
4479 | item = btrfs_item_nr(leaf, i); | |
db94535d | 4480 | |
cfed81a0 CM |
4481 | ioff = btrfs_token_item_offset(leaf, item, &token); |
4482 | btrfs_set_token_item_offset(leaf, item, | |
4483 | ioff - data_size, &token); | |
6567e837 | 4484 | } |
5f39d397 | 4485 | |
6567e837 | 4486 | /* shift the data */ |
5f39d397 | 4487 | memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) + |
6567e837 CM |
4488 | data_end - data_size, btrfs_leaf_data(leaf) + |
4489 | data_end, old_data - data_end); | |
5f39d397 | 4490 | |
6567e837 | 4491 | data_end = old_data; |
5f39d397 CM |
4492 | old_size = btrfs_item_size_nr(leaf, slot); |
4493 | item = btrfs_item_nr(leaf, slot); | |
4494 | btrfs_set_item_size(leaf, item, old_size + data_size); | |
4495 | btrfs_mark_buffer_dirty(leaf); | |
6567e837 | 4496 | |
5f39d397 CM |
4497 | if (btrfs_leaf_free_space(root, leaf) < 0) { |
4498 | btrfs_print_leaf(root, leaf); | |
6567e837 | 4499 | BUG(); |
5f39d397 | 4500 | } |
6567e837 CM |
4501 | } |
4502 | ||
74123bd7 | 4503 | /* |
44871b1b CM |
4504 | * this is a helper for btrfs_insert_empty_items, the main goal here is |
4505 | * to save stack depth by doing the bulk of the work in a function | |
4506 | * that doesn't call btrfs_search_slot | |
74123bd7 | 4507 | */ |
afe5fea7 | 4508 | void setup_items_for_insert(struct btrfs_root *root, struct btrfs_path *path, |
143bede5 JM |
4509 | struct btrfs_key *cpu_key, u32 *data_size, |
4510 | u32 total_data, u32 total_size, int nr) | |
be0e5c09 | 4511 | { |
5f39d397 | 4512 | struct btrfs_item *item; |
9c58309d | 4513 | int i; |
7518a238 | 4514 | u32 nritems; |
be0e5c09 | 4515 | unsigned int data_end; |
e2fa7227 | 4516 | struct btrfs_disk_key disk_key; |
44871b1b CM |
4517 | struct extent_buffer *leaf; |
4518 | int slot; | |
cfed81a0 CM |
4519 | struct btrfs_map_token token; |
4520 | ||
4521 | btrfs_init_map_token(&token); | |
e2fa7227 | 4522 | |
5f39d397 | 4523 | leaf = path->nodes[0]; |
44871b1b | 4524 | slot = path->slots[0]; |
74123bd7 | 4525 | |
5f39d397 | 4526 | nritems = btrfs_header_nritems(leaf); |
123abc88 | 4527 | data_end = leaf_data_end(root, leaf); |
eb60ceac | 4528 | |
f25956cc | 4529 | if (btrfs_leaf_free_space(root, leaf) < total_size) { |
3326d1b0 | 4530 | btrfs_print_leaf(root, leaf); |
d397712b | 4531 | printk(KERN_CRIT "not enough freespace need %u have %d\n", |
9c58309d | 4532 | total_size, btrfs_leaf_free_space(root, leaf)); |
be0e5c09 | 4533 | BUG(); |
d4dbff95 | 4534 | } |
5f39d397 | 4535 | |
be0e5c09 | 4536 | if (slot != nritems) { |
5f39d397 | 4537 | unsigned int old_data = btrfs_item_end_nr(leaf, slot); |
be0e5c09 | 4538 | |
5f39d397 CM |
4539 | if (old_data < data_end) { |
4540 | btrfs_print_leaf(root, leaf); | |
d397712b | 4541 | printk(KERN_CRIT "slot %d old_data %d data_end %d\n", |
5f39d397 CM |
4542 | slot, old_data, data_end); |
4543 | BUG_ON(1); | |
4544 | } | |
be0e5c09 CM |
4545 | /* |
4546 | * item0..itemN ... dataN.offset..dataN.size .. data0.size | |
4547 | */ | |
4548 | /* first correct the data pointers */ | |
0783fcfc | 4549 | for (i = slot; i < nritems; i++) { |
5f39d397 | 4550 | u32 ioff; |
db94535d | 4551 | |
5f39d397 | 4552 | item = btrfs_item_nr(leaf, i); |
cfed81a0 CM |
4553 | ioff = btrfs_token_item_offset(leaf, item, &token); |
4554 | btrfs_set_token_item_offset(leaf, item, | |
4555 | ioff - total_data, &token); | |
0783fcfc | 4556 | } |
be0e5c09 | 4557 | /* shift the items */ |
9c58309d | 4558 | memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot + nr), |
5f39d397 | 4559 | btrfs_item_nr_offset(slot), |
d6025579 | 4560 | (nritems - slot) * sizeof(struct btrfs_item)); |
be0e5c09 CM |
4561 | |
4562 | /* shift the data */ | |
5f39d397 | 4563 | memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) + |
9c58309d | 4564 | data_end - total_data, btrfs_leaf_data(leaf) + |
d6025579 | 4565 | data_end, old_data - data_end); |
be0e5c09 CM |
4566 | data_end = old_data; |
4567 | } | |
5f39d397 | 4568 | |
62e2749e | 4569 | /* setup the item for the new data */ |
9c58309d CM |
4570 | for (i = 0; i < nr; i++) { |
4571 | btrfs_cpu_key_to_disk(&disk_key, cpu_key + i); | |
4572 | btrfs_set_item_key(leaf, &disk_key, slot + i); | |
4573 | item = btrfs_item_nr(leaf, slot + i); | |
cfed81a0 CM |
4574 | btrfs_set_token_item_offset(leaf, item, |
4575 | data_end - data_size[i], &token); | |
9c58309d | 4576 | data_end -= data_size[i]; |
cfed81a0 | 4577 | btrfs_set_token_item_size(leaf, item, data_size[i], &token); |
9c58309d | 4578 | } |
44871b1b | 4579 | |
9c58309d | 4580 | btrfs_set_header_nritems(leaf, nritems + nr); |
aa5d6bed | 4581 | |
5a01a2e3 CM |
4582 | if (slot == 0) { |
4583 | btrfs_cpu_key_to_disk(&disk_key, cpu_key); | |
d6a0a126 | 4584 | fixup_low_keys(root, path, &disk_key, 1); |
5a01a2e3 | 4585 | } |
b9473439 CM |
4586 | btrfs_unlock_up_safe(path, 1); |
4587 | btrfs_mark_buffer_dirty(leaf); | |
aa5d6bed | 4588 | |
5f39d397 CM |
4589 | if (btrfs_leaf_free_space(root, leaf) < 0) { |
4590 | btrfs_print_leaf(root, leaf); | |
be0e5c09 | 4591 | BUG(); |
5f39d397 | 4592 | } |
44871b1b CM |
4593 | } |
4594 | ||
4595 | /* | |
4596 | * Given a key and some data, insert items into the tree. | |
4597 | * This does all the path init required, making room in the tree if needed. | |
4598 | */ | |
4599 | int btrfs_insert_empty_items(struct btrfs_trans_handle *trans, | |
4600 | struct btrfs_root *root, | |
4601 | struct btrfs_path *path, | |
4602 | struct btrfs_key *cpu_key, u32 *data_size, | |
4603 | int nr) | |
4604 | { | |
44871b1b CM |
4605 | int ret = 0; |
4606 | int slot; | |
4607 | int i; | |
4608 | u32 total_size = 0; | |
4609 | u32 total_data = 0; | |
4610 | ||
4611 | for (i = 0; i < nr; i++) | |
4612 | total_data += data_size[i]; | |
4613 | ||
4614 | total_size = total_data + (nr * sizeof(struct btrfs_item)); | |
4615 | ret = btrfs_search_slot(trans, root, cpu_key, path, total_size, 1); | |
4616 | if (ret == 0) | |
4617 | return -EEXIST; | |
4618 | if (ret < 0) | |
143bede5 | 4619 | return ret; |
44871b1b | 4620 | |
44871b1b CM |
4621 | slot = path->slots[0]; |
4622 | BUG_ON(slot < 0); | |
4623 | ||
afe5fea7 | 4624 | setup_items_for_insert(root, path, cpu_key, data_size, |
44871b1b | 4625 | total_data, total_size, nr); |
143bede5 | 4626 | return 0; |
62e2749e CM |
4627 | } |
4628 | ||
4629 | /* | |
4630 | * Given a key and some data, insert an item into the tree. | |
4631 | * This does all the path init required, making room in the tree if needed. | |
4632 | */ | |
e089f05c CM |
4633 | int btrfs_insert_item(struct btrfs_trans_handle *trans, struct btrfs_root |
4634 | *root, struct btrfs_key *cpu_key, void *data, u32 | |
4635 | data_size) | |
62e2749e CM |
4636 | { |
4637 | int ret = 0; | |
2c90e5d6 | 4638 | struct btrfs_path *path; |
5f39d397 CM |
4639 | struct extent_buffer *leaf; |
4640 | unsigned long ptr; | |
62e2749e | 4641 | |
2c90e5d6 | 4642 | path = btrfs_alloc_path(); |
db5b493a TI |
4643 | if (!path) |
4644 | return -ENOMEM; | |
2c90e5d6 | 4645 | ret = btrfs_insert_empty_item(trans, root, path, cpu_key, data_size); |
62e2749e | 4646 | if (!ret) { |
5f39d397 CM |
4647 | leaf = path->nodes[0]; |
4648 | ptr = btrfs_item_ptr_offset(leaf, path->slots[0]); | |
4649 | write_extent_buffer(leaf, data, ptr, data_size); | |
4650 | btrfs_mark_buffer_dirty(leaf); | |
62e2749e | 4651 | } |
2c90e5d6 | 4652 | btrfs_free_path(path); |
aa5d6bed | 4653 | return ret; |
be0e5c09 CM |
4654 | } |
4655 | ||
74123bd7 | 4656 | /* |
5de08d7d | 4657 | * delete the pointer from a given node. |
74123bd7 | 4658 | * |
d352ac68 CM |
4659 | * the tree should have been previously balanced so the deletion does not |
4660 | * empty a node. | |
74123bd7 | 4661 | */ |
afe5fea7 TI |
4662 | static void del_ptr(struct btrfs_root *root, struct btrfs_path *path, |
4663 | int level, int slot) | |
be0e5c09 | 4664 | { |
5f39d397 | 4665 | struct extent_buffer *parent = path->nodes[level]; |
7518a238 | 4666 | u32 nritems; |
f3ea38da | 4667 | int ret; |
be0e5c09 | 4668 | |
5f39d397 | 4669 | nritems = btrfs_header_nritems(parent); |
d397712b | 4670 | if (slot != nritems - 1) { |
0e411ece | 4671 | if (level) |
f3ea38da JS |
4672 | tree_mod_log_eb_move(root->fs_info, parent, slot, |
4673 | slot + 1, nritems - slot - 1); | |
5f39d397 CM |
4674 | memmove_extent_buffer(parent, |
4675 | btrfs_node_key_ptr_offset(slot), | |
4676 | btrfs_node_key_ptr_offset(slot + 1), | |
d6025579 CM |
4677 | sizeof(struct btrfs_key_ptr) * |
4678 | (nritems - slot - 1)); | |
57ba86c0 CM |
4679 | } else if (level) { |
4680 | ret = tree_mod_log_insert_key(root->fs_info, parent, slot, | |
4681 | MOD_LOG_KEY_REMOVE); | |
4682 | BUG_ON(ret < 0); | |
bb803951 | 4683 | } |
f3ea38da | 4684 | |
7518a238 | 4685 | nritems--; |
5f39d397 | 4686 | btrfs_set_header_nritems(parent, nritems); |
7518a238 | 4687 | if (nritems == 0 && parent == root->node) { |
5f39d397 | 4688 | BUG_ON(btrfs_header_level(root->node) != 1); |
bb803951 | 4689 | /* just turn the root into a leaf and break */ |
5f39d397 | 4690 | btrfs_set_header_level(root->node, 0); |
bb803951 | 4691 | } else if (slot == 0) { |
5f39d397 CM |
4692 | struct btrfs_disk_key disk_key; |
4693 | ||
4694 | btrfs_node_key(parent, &disk_key, 0); | |
d6a0a126 | 4695 | fixup_low_keys(root, path, &disk_key, level + 1); |
be0e5c09 | 4696 | } |
d6025579 | 4697 | btrfs_mark_buffer_dirty(parent); |
be0e5c09 CM |
4698 | } |
4699 | ||
323ac95b CM |
4700 | /* |
4701 | * a helper function to delete the leaf pointed to by path->slots[1] and | |
5d4f98a2 | 4702 | * path->nodes[1]. |
323ac95b CM |
4703 | * |
4704 | * This deletes the pointer in path->nodes[1] and frees the leaf | |
4705 | * block extent. zero is returned if it all worked out, < 0 otherwise. | |
4706 | * | |
4707 | * The path must have already been setup for deleting the leaf, including | |
4708 | * all the proper balancing. path->nodes[1] must be locked. | |
4709 | */ | |
143bede5 JM |
4710 | static noinline void btrfs_del_leaf(struct btrfs_trans_handle *trans, |
4711 | struct btrfs_root *root, | |
4712 | struct btrfs_path *path, | |
4713 | struct extent_buffer *leaf) | |
323ac95b | 4714 | { |
5d4f98a2 | 4715 | WARN_ON(btrfs_header_generation(leaf) != trans->transid); |
afe5fea7 | 4716 | del_ptr(root, path, 1, path->slots[1]); |
323ac95b | 4717 | |
4d081c41 CM |
4718 | /* |
4719 | * btrfs_free_extent is expensive, we want to make sure we | |
4720 | * aren't holding any locks when we call it | |
4721 | */ | |
4722 | btrfs_unlock_up_safe(path, 0); | |
4723 | ||
f0486c68 YZ |
4724 | root_sub_used(root, leaf->len); |
4725 | ||
3083ee2e | 4726 | extent_buffer_get(leaf); |
5581a51a | 4727 | btrfs_free_tree_block(trans, root, leaf, 0, 1); |
3083ee2e | 4728 | free_extent_buffer_stale(leaf); |
323ac95b | 4729 | } |
74123bd7 CM |
4730 | /* |
4731 | * delete the item at the leaf level in path. If that empties | |
4732 | * the leaf, remove it from the tree | |
4733 | */ | |
85e21bac CM |
4734 | int btrfs_del_items(struct btrfs_trans_handle *trans, struct btrfs_root *root, |
4735 | struct btrfs_path *path, int slot, int nr) | |
be0e5c09 | 4736 | { |
5f39d397 CM |
4737 | struct extent_buffer *leaf; |
4738 | struct btrfs_item *item; | |
85e21bac CM |
4739 | int last_off; |
4740 | int dsize = 0; | |
aa5d6bed CM |
4741 | int ret = 0; |
4742 | int wret; | |
85e21bac | 4743 | int i; |
7518a238 | 4744 | u32 nritems; |
cfed81a0 CM |
4745 | struct btrfs_map_token token; |
4746 | ||
4747 | btrfs_init_map_token(&token); | |
be0e5c09 | 4748 | |
5f39d397 | 4749 | leaf = path->nodes[0]; |
85e21bac CM |
4750 | last_off = btrfs_item_offset_nr(leaf, slot + nr - 1); |
4751 | ||
4752 | for (i = 0; i < nr; i++) | |
4753 | dsize += btrfs_item_size_nr(leaf, slot + i); | |
4754 | ||
5f39d397 | 4755 | nritems = btrfs_header_nritems(leaf); |
be0e5c09 | 4756 | |
85e21bac | 4757 | if (slot + nr != nritems) { |
123abc88 | 4758 | int data_end = leaf_data_end(root, leaf); |
5f39d397 CM |
4759 | |
4760 | memmove_extent_buffer(leaf, btrfs_leaf_data(leaf) + | |
d6025579 CM |
4761 | data_end + dsize, |
4762 | btrfs_leaf_data(leaf) + data_end, | |
85e21bac | 4763 | last_off - data_end); |
5f39d397 | 4764 | |
85e21bac | 4765 | for (i = slot + nr; i < nritems; i++) { |
5f39d397 | 4766 | u32 ioff; |
db94535d | 4767 | |
5f39d397 | 4768 | item = btrfs_item_nr(leaf, i); |
cfed81a0 CM |
4769 | ioff = btrfs_token_item_offset(leaf, item, &token); |
4770 | btrfs_set_token_item_offset(leaf, item, | |
4771 | ioff + dsize, &token); | |
0783fcfc | 4772 | } |
db94535d | 4773 | |
5f39d397 | 4774 | memmove_extent_buffer(leaf, btrfs_item_nr_offset(slot), |
85e21bac | 4775 | btrfs_item_nr_offset(slot + nr), |
d6025579 | 4776 | sizeof(struct btrfs_item) * |
85e21bac | 4777 | (nritems - slot - nr)); |
be0e5c09 | 4778 | } |
85e21bac CM |
4779 | btrfs_set_header_nritems(leaf, nritems - nr); |
4780 | nritems -= nr; | |
5f39d397 | 4781 | |
74123bd7 | 4782 | /* delete the leaf if we've emptied it */ |
7518a238 | 4783 | if (nritems == 0) { |
5f39d397 CM |
4784 | if (leaf == root->node) { |
4785 | btrfs_set_header_level(leaf, 0); | |
9a8dd150 | 4786 | } else { |
f0486c68 YZ |
4787 | btrfs_set_path_blocking(path); |
4788 | clean_tree_block(trans, root, leaf); | |
143bede5 | 4789 | btrfs_del_leaf(trans, root, path, leaf); |
9a8dd150 | 4790 | } |
be0e5c09 | 4791 | } else { |
7518a238 | 4792 | int used = leaf_space_used(leaf, 0, nritems); |
aa5d6bed | 4793 | if (slot == 0) { |
5f39d397 CM |
4794 | struct btrfs_disk_key disk_key; |
4795 | ||
4796 | btrfs_item_key(leaf, &disk_key, 0); | |
d6a0a126 | 4797 | fixup_low_keys(root, path, &disk_key, 1); |
aa5d6bed | 4798 | } |
aa5d6bed | 4799 | |
74123bd7 | 4800 | /* delete the leaf if it is mostly empty */ |
d717aa1d | 4801 | if (used < BTRFS_LEAF_DATA_SIZE(root) / 3) { |
be0e5c09 CM |
4802 | /* push_leaf_left fixes the path. |
4803 | * make sure the path still points to our leaf | |
4804 | * for possible call to del_ptr below | |
4805 | */ | |
4920c9ac | 4806 | slot = path->slots[1]; |
5f39d397 CM |
4807 | extent_buffer_get(leaf); |
4808 | ||
b9473439 | 4809 | btrfs_set_path_blocking(path); |
99d8f83c CM |
4810 | wret = push_leaf_left(trans, root, path, 1, 1, |
4811 | 1, (u32)-1); | |
54aa1f4d | 4812 | if (wret < 0 && wret != -ENOSPC) |
aa5d6bed | 4813 | ret = wret; |
5f39d397 CM |
4814 | |
4815 | if (path->nodes[0] == leaf && | |
4816 | btrfs_header_nritems(leaf)) { | |
99d8f83c CM |
4817 | wret = push_leaf_right(trans, root, path, 1, |
4818 | 1, 1, 0); | |
54aa1f4d | 4819 | if (wret < 0 && wret != -ENOSPC) |
aa5d6bed CM |
4820 | ret = wret; |
4821 | } | |
5f39d397 CM |
4822 | |
4823 | if (btrfs_header_nritems(leaf) == 0) { | |
323ac95b | 4824 | path->slots[1] = slot; |
143bede5 | 4825 | btrfs_del_leaf(trans, root, path, leaf); |
5f39d397 | 4826 | free_extent_buffer(leaf); |
143bede5 | 4827 | ret = 0; |
5de08d7d | 4828 | } else { |
925baedd CM |
4829 | /* if we're still in the path, make sure |
4830 | * we're dirty. Otherwise, one of the | |
4831 | * push_leaf functions must have already | |
4832 | * dirtied this buffer | |
4833 | */ | |
4834 | if (path->nodes[0] == leaf) | |
4835 | btrfs_mark_buffer_dirty(leaf); | |
5f39d397 | 4836 | free_extent_buffer(leaf); |
be0e5c09 | 4837 | } |
d5719762 | 4838 | } else { |
5f39d397 | 4839 | btrfs_mark_buffer_dirty(leaf); |
be0e5c09 CM |
4840 | } |
4841 | } | |
aa5d6bed | 4842 | return ret; |
be0e5c09 CM |
4843 | } |
4844 | ||
7bb86316 | 4845 | /* |
925baedd | 4846 | * search the tree again to find a leaf with lesser keys |
7bb86316 CM |
4847 | * returns 0 if it found something or 1 if there are no lesser leaves. |
4848 | * returns < 0 on io errors. | |
d352ac68 CM |
4849 | * |
4850 | * This may release the path, and so you may lose any locks held at the | |
4851 | * time you call it. | |
7bb86316 CM |
4852 | */ |
4853 | int btrfs_prev_leaf(struct btrfs_root *root, struct btrfs_path *path) | |
4854 | { | |
925baedd CM |
4855 | struct btrfs_key key; |
4856 | struct btrfs_disk_key found_key; | |
4857 | int ret; | |
7bb86316 | 4858 | |
925baedd | 4859 | btrfs_item_key_to_cpu(path->nodes[0], &key, 0); |
7bb86316 | 4860 | |
925baedd CM |
4861 | if (key.offset > 0) |
4862 | key.offset--; | |
4863 | else if (key.type > 0) | |
4864 | key.type--; | |
4865 | else if (key.objectid > 0) | |
4866 | key.objectid--; | |
4867 | else | |
4868 | return 1; | |
7bb86316 | 4869 | |
b3b4aa74 | 4870 | btrfs_release_path(path); |
925baedd CM |
4871 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); |
4872 | if (ret < 0) | |
4873 | return ret; | |
4874 | btrfs_item_key(path->nodes[0], &found_key, 0); | |
4875 | ret = comp_keys(&found_key, &key); | |
4876 | if (ret < 0) | |
4877 | return 0; | |
4878 | return 1; | |
7bb86316 CM |
4879 | } |
4880 | ||
3f157a2f CM |
4881 | /* |
4882 | * A helper function to walk down the tree starting at min_key, and looking | |
de78b51a ES |
4883 | * for nodes or leaves that are have a minimum transaction id. |
4884 | * This is used by the btree defrag code, and tree logging | |
3f157a2f CM |
4885 | * |
4886 | * This does not cow, but it does stuff the starting key it finds back | |
4887 | * into min_key, so you can call btrfs_search_slot with cow=1 on the | |
4888 | * key and get a writable path. | |
4889 | * | |
4890 | * This does lock as it descends, and path->keep_locks should be set | |
4891 | * to 1 by the caller. | |
4892 | * | |
4893 | * This honors path->lowest_level to prevent descent past a given level | |
4894 | * of the tree. | |
4895 | * | |
d352ac68 CM |
4896 | * min_trans indicates the oldest transaction that you are interested |
4897 | * in walking through. Any nodes or leaves older than min_trans are | |
4898 | * skipped over (without reading them). | |
4899 | * | |
3f157a2f CM |
4900 | * returns zero if something useful was found, < 0 on error and 1 if there |
4901 | * was nothing in the tree that matched the search criteria. | |
4902 | */ | |
4903 | int btrfs_search_forward(struct btrfs_root *root, struct btrfs_key *min_key, | |
e02119d5 | 4904 | struct btrfs_key *max_key, |
de78b51a | 4905 | struct btrfs_path *path, |
3f157a2f CM |
4906 | u64 min_trans) |
4907 | { | |
4908 | struct extent_buffer *cur; | |
4909 | struct btrfs_key found_key; | |
4910 | int slot; | |
9652480b | 4911 | int sret; |
3f157a2f CM |
4912 | u32 nritems; |
4913 | int level; | |
4914 | int ret = 1; | |
4915 | ||
934d375b | 4916 | WARN_ON(!path->keep_locks); |
3f157a2f | 4917 | again: |
bd681513 | 4918 | cur = btrfs_read_lock_root_node(root); |
3f157a2f | 4919 | level = btrfs_header_level(cur); |
e02119d5 | 4920 | WARN_ON(path->nodes[level]); |
3f157a2f | 4921 | path->nodes[level] = cur; |
bd681513 | 4922 | path->locks[level] = BTRFS_READ_LOCK; |
3f157a2f CM |
4923 | |
4924 | if (btrfs_header_generation(cur) < min_trans) { | |
4925 | ret = 1; | |
4926 | goto out; | |
4927 | } | |
d397712b | 4928 | while (1) { |
3f157a2f CM |
4929 | nritems = btrfs_header_nritems(cur); |
4930 | level = btrfs_header_level(cur); | |
9652480b | 4931 | sret = bin_search(cur, min_key, level, &slot); |
3f157a2f | 4932 | |
323ac95b CM |
4933 | /* at the lowest level, we're done, setup the path and exit */ |
4934 | if (level == path->lowest_level) { | |
e02119d5 CM |
4935 | if (slot >= nritems) |
4936 | goto find_next_key; | |
3f157a2f CM |
4937 | ret = 0; |
4938 | path->slots[level] = slot; | |
4939 | btrfs_item_key_to_cpu(cur, &found_key, slot); | |
4940 | goto out; | |
4941 | } | |
9652480b Y |
4942 | if (sret && slot > 0) |
4943 | slot--; | |
3f157a2f | 4944 | /* |
de78b51a ES |
4945 | * check this node pointer against the min_trans parameters. |
4946 | * If it is too old, old, skip to the next one. | |
3f157a2f | 4947 | */ |
d397712b | 4948 | while (slot < nritems) { |
3f157a2f CM |
4949 | u64 blockptr; |
4950 | u64 gen; | |
e02119d5 | 4951 | |
3f157a2f CM |
4952 | blockptr = btrfs_node_blockptr(cur, slot); |
4953 | gen = btrfs_node_ptr_generation(cur, slot); | |
4954 | if (gen < min_trans) { | |
4955 | slot++; | |
4956 | continue; | |
4957 | } | |
de78b51a | 4958 | break; |
3f157a2f | 4959 | } |
e02119d5 | 4960 | find_next_key: |
3f157a2f CM |
4961 | /* |
4962 | * we didn't find a candidate key in this node, walk forward | |
4963 | * and find another one | |
4964 | */ | |
4965 | if (slot >= nritems) { | |
e02119d5 | 4966 | path->slots[level] = slot; |
b4ce94de | 4967 | btrfs_set_path_blocking(path); |
e02119d5 | 4968 | sret = btrfs_find_next_key(root, path, min_key, level, |
de78b51a | 4969 | min_trans); |
e02119d5 | 4970 | if (sret == 0) { |
b3b4aa74 | 4971 | btrfs_release_path(path); |
3f157a2f CM |
4972 | goto again; |
4973 | } else { | |
4974 | goto out; | |
4975 | } | |
4976 | } | |
4977 | /* save our key for returning back */ | |
4978 | btrfs_node_key_to_cpu(cur, &found_key, slot); | |
4979 | path->slots[level] = slot; | |
4980 | if (level == path->lowest_level) { | |
4981 | ret = 0; | |
f7c79f30 | 4982 | unlock_up(path, level, 1, 0, NULL); |
3f157a2f CM |
4983 | goto out; |
4984 | } | |
b4ce94de | 4985 | btrfs_set_path_blocking(path); |
3f157a2f | 4986 | cur = read_node_slot(root, cur, slot); |
79787eaa | 4987 | BUG_ON(!cur); /* -ENOMEM */ |
3f157a2f | 4988 | |
bd681513 | 4989 | btrfs_tree_read_lock(cur); |
b4ce94de | 4990 | |
bd681513 | 4991 | path->locks[level - 1] = BTRFS_READ_LOCK; |
3f157a2f | 4992 | path->nodes[level - 1] = cur; |
f7c79f30 | 4993 | unlock_up(path, level, 1, 0, NULL); |
bd681513 | 4994 | btrfs_clear_path_blocking(path, NULL, 0); |
3f157a2f CM |
4995 | } |
4996 | out: | |
4997 | if (ret == 0) | |
4998 | memcpy(min_key, &found_key, sizeof(found_key)); | |
b4ce94de | 4999 | btrfs_set_path_blocking(path); |
3f157a2f CM |
5000 | return ret; |
5001 | } | |
5002 | ||
7069830a AB |
5003 | static void tree_move_down(struct btrfs_root *root, |
5004 | struct btrfs_path *path, | |
5005 | int *level, int root_level) | |
5006 | { | |
74dd17fb | 5007 | BUG_ON(*level == 0); |
7069830a AB |
5008 | path->nodes[*level - 1] = read_node_slot(root, path->nodes[*level], |
5009 | path->slots[*level]); | |
5010 | path->slots[*level - 1] = 0; | |
5011 | (*level)--; | |
5012 | } | |
5013 | ||
5014 | static int tree_move_next_or_upnext(struct btrfs_root *root, | |
5015 | struct btrfs_path *path, | |
5016 | int *level, int root_level) | |
5017 | { | |
5018 | int ret = 0; | |
5019 | int nritems; | |
5020 | nritems = btrfs_header_nritems(path->nodes[*level]); | |
5021 | ||
5022 | path->slots[*level]++; | |
5023 | ||
74dd17fb | 5024 | while (path->slots[*level] >= nritems) { |
7069830a AB |
5025 | if (*level == root_level) |
5026 | return -1; | |
5027 | ||
5028 | /* move upnext */ | |
5029 | path->slots[*level] = 0; | |
5030 | free_extent_buffer(path->nodes[*level]); | |
5031 | path->nodes[*level] = NULL; | |
5032 | (*level)++; | |
5033 | path->slots[*level]++; | |
5034 | ||
5035 | nritems = btrfs_header_nritems(path->nodes[*level]); | |
5036 | ret = 1; | |
5037 | } | |
5038 | return ret; | |
5039 | } | |
5040 | ||
5041 | /* | |
5042 | * Returns 1 if it had to move up and next. 0 is returned if it moved only next | |
5043 | * or down. | |
5044 | */ | |
5045 | static int tree_advance(struct btrfs_root *root, | |
5046 | struct btrfs_path *path, | |
5047 | int *level, int root_level, | |
5048 | int allow_down, | |
5049 | struct btrfs_key *key) | |
5050 | { | |
5051 | int ret; | |
5052 | ||
5053 | if (*level == 0 || !allow_down) { | |
5054 | ret = tree_move_next_or_upnext(root, path, level, root_level); | |
5055 | } else { | |
5056 | tree_move_down(root, path, level, root_level); | |
5057 | ret = 0; | |
5058 | } | |
5059 | if (ret >= 0) { | |
5060 | if (*level == 0) | |
5061 | btrfs_item_key_to_cpu(path->nodes[*level], key, | |
5062 | path->slots[*level]); | |
5063 | else | |
5064 | btrfs_node_key_to_cpu(path->nodes[*level], key, | |
5065 | path->slots[*level]); | |
5066 | } | |
5067 | return ret; | |
5068 | } | |
5069 | ||
5070 | static int tree_compare_item(struct btrfs_root *left_root, | |
5071 | struct btrfs_path *left_path, | |
5072 | struct btrfs_path *right_path, | |
5073 | char *tmp_buf) | |
5074 | { | |
5075 | int cmp; | |
5076 | int len1, len2; | |
5077 | unsigned long off1, off2; | |
5078 | ||
5079 | len1 = btrfs_item_size_nr(left_path->nodes[0], left_path->slots[0]); | |
5080 | len2 = btrfs_item_size_nr(right_path->nodes[0], right_path->slots[0]); | |
5081 | if (len1 != len2) | |
5082 | return 1; | |
5083 | ||
5084 | off1 = btrfs_item_ptr_offset(left_path->nodes[0], left_path->slots[0]); | |
5085 | off2 = btrfs_item_ptr_offset(right_path->nodes[0], | |
5086 | right_path->slots[0]); | |
5087 | ||
5088 | read_extent_buffer(left_path->nodes[0], tmp_buf, off1, len1); | |
5089 | ||
5090 | cmp = memcmp_extent_buffer(right_path->nodes[0], tmp_buf, off2, len1); | |
5091 | if (cmp) | |
5092 | return 1; | |
5093 | return 0; | |
5094 | } | |
5095 | ||
5096 | #define ADVANCE 1 | |
5097 | #define ADVANCE_ONLY_NEXT -1 | |
5098 | ||
5099 | /* | |
5100 | * This function compares two trees and calls the provided callback for | |
5101 | * every changed/new/deleted item it finds. | |
5102 | * If shared tree blocks are encountered, whole subtrees are skipped, making | |
5103 | * the compare pretty fast on snapshotted subvolumes. | |
5104 | * | |
5105 | * This currently works on commit roots only. As commit roots are read only, | |
5106 | * we don't do any locking. The commit roots are protected with transactions. | |
5107 | * Transactions are ended and rejoined when a commit is tried in between. | |
5108 | * | |
5109 | * This function checks for modifications done to the trees while comparing. | |
5110 | * If it detects a change, it aborts immediately. | |
5111 | */ | |
5112 | int btrfs_compare_trees(struct btrfs_root *left_root, | |
5113 | struct btrfs_root *right_root, | |
5114 | btrfs_changed_cb_t changed_cb, void *ctx) | |
5115 | { | |
5116 | int ret; | |
5117 | int cmp; | |
5118 | struct btrfs_trans_handle *trans = NULL; | |
5119 | struct btrfs_path *left_path = NULL; | |
5120 | struct btrfs_path *right_path = NULL; | |
5121 | struct btrfs_key left_key; | |
5122 | struct btrfs_key right_key; | |
5123 | char *tmp_buf = NULL; | |
5124 | int left_root_level; | |
5125 | int right_root_level; | |
5126 | int left_level; | |
5127 | int right_level; | |
5128 | int left_end_reached; | |
5129 | int right_end_reached; | |
5130 | int advance_left; | |
5131 | int advance_right; | |
5132 | u64 left_blockptr; | |
5133 | u64 right_blockptr; | |
5134 | u64 left_start_ctransid; | |
5135 | u64 right_start_ctransid; | |
5136 | u64 ctransid; | |
5137 | ||
5138 | left_path = btrfs_alloc_path(); | |
5139 | if (!left_path) { | |
5140 | ret = -ENOMEM; | |
5141 | goto out; | |
5142 | } | |
5143 | right_path = btrfs_alloc_path(); | |
5144 | if (!right_path) { | |
5145 | ret = -ENOMEM; | |
5146 | goto out; | |
5147 | } | |
5148 | ||
5149 | tmp_buf = kmalloc(left_root->leafsize, GFP_NOFS); | |
5150 | if (!tmp_buf) { | |
5151 | ret = -ENOMEM; | |
5152 | goto out; | |
5153 | } | |
5154 | ||
5155 | left_path->search_commit_root = 1; | |
5156 | left_path->skip_locking = 1; | |
5157 | right_path->search_commit_root = 1; | |
5158 | right_path->skip_locking = 1; | |
5159 | ||
5f3ab90a | 5160 | spin_lock(&left_root->root_item_lock); |
7069830a | 5161 | left_start_ctransid = btrfs_root_ctransid(&left_root->root_item); |
5f3ab90a | 5162 | spin_unlock(&left_root->root_item_lock); |
7069830a | 5163 | |
5f3ab90a | 5164 | spin_lock(&right_root->root_item_lock); |
7069830a | 5165 | right_start_ctransid = btrfs_root_ctransid(&right_root->root_item); |
5f3ab90a | 5166 | spin_unlock(&right_root->root_item_lock); |
7069830a AB |
5167 | |
5168 | trans = btrfs_join_transaction(left_root); | |
5169 | if (IS_ERR(trans)) { | |
5170 | ret = PTR_ERR(trans); | |
5171 | trans = NULL; | |
5172 | goto out; | |
5173 | } | |
5174 | ||
5175 | /* | |
5176 | * Strategy: Go to the first items of both trees. Then do | |
5177 | * | |
5178 | * If both trees are at level 0 | |
5179 | * Compare keys of current items | |
5180 | * If left < right treat left item as new, advance left tree | |
5181 | * and repeat | |
5182 | * If left > right treat right item as deleted, advance right tree | |
5183 | * and repeat | |
5184 | * If left == right do deep compare of items, treat as changed if | |
5185 | * needed, advance both trees and repeat | |
5186 | * If both trees are at the same level but not at level 0 | |
5187 | * Compare keys of current nodes/leafs | |
5188 | * If left < right advance left tree and repeat | |
5189 | * If left > right advance right tree and repeat | |
5190 | * If left == right compare blockptrs of the next nodes/leafs | |
5191 | * If they match advance both trees but stay at the same level | |
5192 | * and repeat | |
5193 | * If they don't match advance both trees while allowing to go | |
5194 | * deeper and repeat | |
5195 | * If tree levels are different | |
5196 | * Advance the tree that needs it and repeat | |
5197 | * | |
5198 | * Advancing a tree means: | |
5199 | * If we are at level 0, try to go to the next slot. If that's not | |
5200 | * possible, go one level up and repeat. Stop when we found a level | |
5201 | * where we could go to the next slot. We may at this point be on a | |
5202 | * node or a leaf. | |
5203 | * | |
5204 | * If we are not at level 0 and not on shared tree blocks, go one | |
5205 | * level deeper. | |
5206 | * | |
5207 | * If we are not at level 0 and on shared tree blocks, go one slot to | |
5208 | * the right if possible or go up and right. | |
5209 | */ | |
5210 | ||
5211 | left_level = btrfs_header_level(left_root->commit_root); | |
5212 | left_root_level = left_level; | |
5213 | left_path->nodes[left_level] = left_root->commit_root; | |
5214 | extent_buffer_get(left_path->nodes[left_level]); | |
5215 | ||
5216 | right_level = btrfs_header_level(right_root->commit_root); | |
5217 | right_root_level = right_level; | |
5218 | right_path->nodes[right_level] = right_root->commit_root; | |
5219 | extent_buffer_get(right_path->nodes[right_level]); | |
5220 | ||
5221 | if (left_level == 0) | |
5222 | btrfs_item_key_to_cpu(left_path->nodes[left_level], | |
5223 | &left_key, left_path->slots[left_level]); | |
5224 | else | |
5225 | btrfs_node_key_to_cpu(left_path->nodes[left_level], | |
5226 | &left_key, left_path->slots[left_level]); | |
5227 | if (right_level == 0) | |
5228 | btrfs_item_key_to_cpu(right_path->nodes[right_level], | |
5229 | &right_key, right_path->slots[right_level]); | |
5230 | else | |
5231 | btrfs_node_key_to_cpu(right_path->nodes[right_level], | |
5232 | &right_key, right_path->slots[right_level]); | |
5233 | ||
5234 | left_end_reached = right_end_reached = 0; | |
5235 | advance_left = advance_right = 0; | |
5236 | ||
5237 | while (1) { | |
5238 | /* | |
5239 | * We need to make sure the transaction does not get committed | |
5240 | * while we do anything on commit roots. This means, we need to | |
5241 | * join and leave transactions for every item that we process. | |
5242 | */ | |
5243 | if (trans && btrfs_should_end_transaction(trans, left_root)) { | |
5244 | btrfs_release_path(left_path); | |
5245 | btrfs_release_path(right_path); | |
5246 | ||
5247 | ret = btrfs_end_transaction(trans, left_root); | |
5248 | trans = NULL; | |
5249 | if (ret < 0) | |
5250 | goto out; | |
5251 | } | |
5252 | /* now rejoin the transaction */ | |
5253 | if (!trans) { | |
5254 | trans = btrfs_join_transaction(left_root); | |
5255 | if (IS_ERR(trans)) { | |
5256 | ret = PTR_ERR(trans); | |
5257 | trans = NULL; | |
5258 | goto out; | |
5259 | } | |
5260 | ||
5f3ab90a | 5261 | spin_lock(&left_root->root_item_lock); |
7069830a | 5262 | ctransid = btrfs_root_ctransid(&left_root->root_item); |
5f3ab90a | 5263 | spin_unlock(&left_root->root_item_lock); |
7069830a AB |
5264 | if (ctransid != left_start_ctransid) |
5265 | left_start_ctransid = 0; | |
5266 | ||
5f3ab90a | 5267 | spin_lock(&right_root->root_item_lock); |
7069830a | 5268 | ctransid = btrfs_root_ctransid(&right_root->root_item); |
5f3ab90a | 5269 | spin_unlock(&right_root->root_item_lock); |
7069830a AB |
5270 | if (ctransid != right_start_ctransid) |
5271 | right_start_ctransid = 0; | |
5272 | ||
5273 | if (!left_start_ctransid || !right_start_ctransid) { | |
5274 | WARN(1, KERN_WARNING | |
5275 | "btrfs: btrfs_compare_tree detected " | |
5276 | "a change in one of the trees while " | |
5277 | "iterating. This is probably a " | |
5278 | "bug.\n"); | |
5279 | ret = -EIO; | |
5280 | goto out; | |
5281 | } | |
5282 | ||
5283 | /* | |
5284 | * the commit root may have changed, so start again | |
5285 | * where we stopped | |
5286 | */ | |
5287 | left_path->lowest_level = left_level; | |
5288 | right_path->lowest_level = right_level; | |
5289 | ret = btrfs_search_slot(NULL, left_root, | |
5290 | &left_key, left_path, 0, 0); | |
5291 | if (ret < 0) | |
5292 | goto out; | |
5293 | ret = btrfs_search_slot(NULL, right_root, | |
5294 | &right_key, right_path, 0, 0); | |
5295 | if (ret < 0) | |
5296 | goto out; | |
5297 | } | |
5298 | ||
5299 | if (advance_left && !left_end_reached) { | |
5300 | ret = tree_advance(left_root, left_path, &left_level, | |
5301 | left_root_level, | |
5302 | advance_left != ADVANCE_ONLY_NEXT, | |
5303 | &left_key); | |
5304 | if (ret < 0) | |
5305 | left_end_reached = ADVANCE; | |
5306 | advance_left = 0; | |
5307 | } | |
5308 | if (advance_right && !right_end_reached) { | |
5309 | ret = tree_advance(right_root, right_path, &right_level, | |
5310 | right_root_level, | |
5311 | advance_right != ADVANCE_ONLY_NEXT, | |
5312 | &right_key); | |
5313 | if (ret < 0) | |
5314 | right_end_reached = ADVANCE; | |
5315 | advance_right = 0; | |
5316 | } | |
5317 | ||
5318 | if (left_end_reached && right_end_reached) { | |
5319 | ret = 0; | |
5320 | goto out; | |
5321 | } else if (left_end_reached) { | |
5322 | if (right_level == 0) { | |
5323 | ret = changed_cb(left_root, right_root, | |
5324 | left_path, right_path, | |
5325 | &right_key, | |
5326 | BTRFS_COMPARE_TREE_DELETED, | |
5327 | ctx); | |
5328 | if (ret < 0) | |
5329 | goto out; | |
5330 | } | |
5331 | advance_right = ADVANCE; | |
5332 | continue; | |
5333 | } else if (right_end_reached) { | |
5334 | if (left_level == 0) { | |
5335 | ret = changed_cb(left_root, right_root, | |
5336 | left_path, right_path, | |
5337 | &left_key, | |
5338 | BTRFS_COMPARE_TREE_NEW, | |
5339 | ctx); | |
5340 | if (ret < 0) | |
5341 | goto out; | |
5342 | } | |
5343 | advance_left = ADVANCE; | |
5344 | continue; | |
5345 | } | |
5346 | ||
5347 | if (left_level == 0 && right_level == 0) { | |
5348 | cmp = btrfs_comp_cpu_keys(&left_key, &right_key); | |
5349 | if (cmp < 0) { | |
5350 | ret = changed_cb(left_root, right_root, | |
5351 | left_path, right_path, | |
5352 | &left_key, | |
5353 | BTRFS_COMPARE_TREE_NEW, | |
5354 | ctx); | |
5355 | if (ret < 0) | |
5356 | goto out; | |
5357 | advance_left = ADVANCE; | |
5358 | } else if (cmp > 0) { | |
5359 | ret = changed_cb(left_root, right_root, | |
5360 | left_path, right_path, | |
5361 | &right_key, | |
5362 | BTRFS_COMPARE_TREE_DELETED, | |
5363 | ctx); | |
5364 | if (ret < 0) | |
5365 | goto out; | |
5366 | advance_right = ADVANCE; | |
5367 | } else { | |
74dd17fb | 5368 | WARN_ON(!extent_buffer_uptodate(left_path->nodes[0])); |
7069830a AB |
5369 | ret = tree_compare_item(left_root, left_path, |
5370 | right_path, tmp_buf); | |
5371 | if (ret) { | |
74dd17fb | 5372 | WARN_ON(!extent_buffer_uptodate(left_path->nodes[0])); |
7069830a AB |
5373 | ret = changed_cb(left_root, right_root, |
5374 | left_path, right_path, | |
5375 | &left_key, | |
5376 | BTRFS_COMPARE_TREE_CHANGED, | |
5377 | ctx); | |
5378 | if (ret < 0) | |
5379 | goto out; | |
5380 | } | |
5381 | advance_left = ADVANCE; | |
5382 | advance_right = ADVANCE; | |
5383 | } | |
5384 | } else if (left_level == right_level) { | |
5385 | cmp = btrfs_comp_cpu_keys(&left_key, &right_key); | |
5386 | if (cmp < 0) { | |
5387 | advance_left = ADVANCE; | |
5388 | } else if (cmp > 0) { | |
5389 | advance_right = ADVANCE; | |
5390 | } else { | |
5391 | left_blockptr = btrfs_node_blockptr( | |
5392 | left_path->nodes[left_level], | |
5393 | left_path->slots[left_level]); | |
5394 | right_blockptr = btrfs_node_blockptr( | |
5395 | right_path->nodes[right_level], | |
5396 | right_path->slots[right_level]); | |
5397 | if (left_blockptr == right_blockptr) { | |
5398 | /* | |
5399 | * As we're on a shared block, don't | |
5400 | * allow to go deeper. | |
5401 | */ | |
5402 | advance_left = ADVANCE_ONLY_NEXT; | |
5403 | advance_right = ADVANCE_ONLY_NEXT; | |
5404 | } else { | |
5405 | advance_left = ADVANCE; | |
5406 | advance_right = ADVANCE; | |
5407 | } | |
5408 | } | |
5409 | } else if (left_level < right_level) { | |
5410 | advance_right = ADVANCE; | |
5411 | } else { | |
5412 | advance_left = ADVANCE; | |
5413 | } | |
5414 | } | |
5415 | ||
5416 | out: | |
5417 | btrfs_free_path(left_path); | |
5418 | btrfs_free_path(right_path); | |
5419 | kfree(tmp_buf); | |
5420 | ||
5421 | if (trans) { | |
5422 | if (!ret) | |
5423 | ret = btrfs_end_transaction(trans, left_root); | |
5424 | else | |
5425 | btrfs_end_transaction(trans, left_root); | |
5426 | } | |
5427 | ||
5428 | return ret; | |
5429 | } | |
5430 | ||
3f157a2f CM |
5431 | /* |
5432 | * this is similar to btrfs_next_leaf, but does not try to preserve | |
5433 | * and fixup the path. It looks for and returns the next key in the | |
de78b51a | 5434 | * tree based on the current path and the min_trans parameters. |
3f157a2f CM |
5435 | * |
5436 | * 0 is returned if another key is found, < 0 if there are any errors | |
5437 | * and 1 is returned if there are no higher keys in the tree | |
5438 | * | |
5439 | * path->keep_locks should be set to 1 on the search made before | |
5440 | * calling this function. | |
5441 | */ | |
e7a84565 | 5442 | int btrfs_find_next_key(struct btrfs_root *root, struct btrfs_path *path, |
de78b51a | 5443 | struct btrfs_key *key, int level, u64 min_trans) |
e7a84565 | 5444 | { |
e7a84565 CM |
5445 | int slot; |
5446 | struct extent_buffer *c; | |
5447 | ||
934d375b | 5448 | WARN_ON(!path->keep_locks); |
d397712b | 5449 | while (level < BTRFS_MAX_LEVEL) { |
e7a84565 CM |
5450 | if (!path->nodes[level]) |
5451 | return 1; | |
5452 | ||
5453 | slot = path->slots[level] + 1; | |
5454 | c = path->nodes[level]; | |
3f157a2f | 5455 | next: |
e7a84565 | 5456 | if (slot >= btrfs_header_nritems(c)) { |
33c66f43 YZ |
5457 | int ret; |
5458 | int orig_lowest; | |
5459 | struct btrfs_key cur_key; | |
5460 | if (level + 1 >= BTRFS_MAX_LEVEL || | |
5461 | !path->nodes[level + 1]) | |
e7a84565 | 5462 | return 1; |
33c66f43 YZ |
5463 | |
5464 | if (path->locks[level + 1]) { | |
5465 | level++; | |
5466 | continue; | |
5467 | } | |
5468 | ||
5469 | slot = btrfs_header_nritems(c) - 1; | |
5470 | if (level == 0) | |
5471 | btrfs_item_key_to_cpu(c, &cur_key, slot); | |
5472 | else | |
5473 | btrfs_node_key_to_cpu(c, &cur_key, slot); | |
5474 | ||
5475 | orig_lowest = path->lowest_level; | |
b3b4aa74 | 5476 | btrfs_release_path(path); |
33c66f43 YZ |
5477 | path->lowest_level = level; |
5478 | ret = btrfs_search_slot(NULL, root, &cur_key, path, | |
5479 | 0, 0); | |
5480 | path->lowest_level = orig_lowest; | |
5481 | if (ret < 0) | |
5482 | return ret; | |
5483 | ||
5484 | c = path->nodes[level]; | |
5485 | slot = path->slots[level]; | |
5486 | if (ret == 0) | |
5487 | slot++; | |
5488 | goto next; | |
e7a84565 | 5489 | } |
33c66f43 | 5490 | |
e7a84565 CM |
5491 | if (level == 0) |
5492 | btrfs_item_key_to_cpu(c, key, slot); | |
3f157a2f | 5493 | else { |
3f157a2f CM |
5494 | u64 gen = btrfs_node_ptr_generation(c, slot); |
5495 | ||
3f157a2f CM |
5496 | if (gen < min_trans) { |
5497 | slot++; | |
5498 | goto next; | |
5499 | } | |
e7a84565 | 5500 | btrfs_node_key_to_cpu(c, key, slot); |
3f157a2f | 5501 | } |
e7a84565 CM |
5502 | return 0; |
5503 | } | |
5504 | return 1; | |
5505 | } | |
5506 | ||
97571fd0 | 5507 | /* |
925baedd | 5508 | * search the tree again to find a leaf with greater keys |
0f70abe2 CM |
5509 | * returns 0 if it found something or 1 if there are no greater leaves. |
5510 | * returns < 0 on io errors. | |
97571fd0 | 5511 | */ |
234b63a0 | 5512 | int btrfs_next_leaf(struct btrfs_root *root, struct btrfs_path *path) |
3d7806ec JS |
5513 | { |
5514 | return btrfs_next_old_leaf(root, path, 0); | |
5515 | } | |
5516 | ||
5517 | int btrfs_next_old_leaf(struct btrfs_root *root, struct btrfs_path *path, | |
5518 | u64 time_seq) | |
d97e63b6 CM |
5519 | { |
5520 | int slot; | |
8e73f275 | 5521 | int level; |
5f39d397 | 5522 | struct extent_buffer *c; |
8e73f275 | 5523 | struct extent_buffer *next; |
925baedd CM |
5524 | struct btrfs_key key; |
5525 | u32 nritems; | |
5526 | int ret; | |
8e73f275 | 5527 | int old_spinning = path->leave_spinning; |
bd681513 | 5528 | int next_rw_lock = 0; |
925baedd CM |
5529 | |
5530 | nritems = btrfs_header_nritems(path->nodes[0]); | |
d397712b | 5531 | if (nritems == 0) |
925baedd | 5532 | return 1; |
925baedd | 5533 | |
8e73f275 CM |
5534 | btrfs_item_key_to_cpu(path->nodes[0], &key, nritems - 1); |
5535 | again: | |
5536 | level = 1; | |
5537 | next = NULL; | |
bd681513 | 5538 | next_rw_lock = 0; |
b3b4aa74 | 5539 | btrfs_release_path(path); |
8e73f275 | 5540 | |
a2135011 | 5541 | path->keep_locks = 1; |
31533fb2 | 5542 | path->leave_spinning = 1; |
8e73f275 | 5543 | |
3d7806ec JS |
5544 | if (time_seq) |
5545 | ret = btrfs_search_old_slot(root, &key, path, time_seq); | |
5546 | else | |
5547 | ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); | |
925baedd CM |
5548 | path->keep_locks = 0; |
5549 | ||
5550 | if (ret < 0) | |
5551 | return ret; | |
5552 | ||
a2135011 | 5553 | nritems = btrfs_header_nritems(path->nodes[0]); |
168fd7d2 CM |
5554 | /* |
5555 | * by releasing the path above we dropped all our locks. A balance | |
5556 | * could have added more items next to the key that used to be | |
5557 | * at the very end of the block. So, check again here and | |
5558 | * advance the path if there are now more items available. | |
5559 | */ | |
a2135011 | 5560 | if (nritems > 0 && path->slots[0] < nritems - 1) { |
e457afec YZ |
5561 | if (ret == 0) |
5562 | path->slots[0]++; | |
8e73f275 | 5563 | ret = 0; |
925baedd CM |
5564 | goto done; |
5565 | } | |
d97e63b6 | 5566 | |
d397712b | 5567 | while (level < BTRFS_MAX_LEVEL) { |
8e73f275 CM |
5568 | if (!path->nodes[level]) { |
5569 | ret = 1; | |
5570 | goto done; | |
5571 | } | |
5f39d397 | 5572 | |
d97e63b6 CM |
5573 | slot = path->slots[level] + 1; |
5574 | c = path->nodes[level]; | |
5f39d397 | 5575 | if (slot >= btrfs_header_nritems(c)) { |
d97e63b6 | 5576 | level++; |
8e73f275 CM |
5577 | if (level == BTRFS_MAX_LEVEL) { |
5578 | ret = 1; | |
5579 | goto done; | |
5580 | } | |
d97e63b6 CM |
5581 | continue; |
5582 | } | |
5f39d397 | 5583 | |
925baedd | 5584 | if (next) { |
bd681513 | 5585 | btrfs_tree_unlock_rw(next, next_rw_lock); |
5f39d397 | 5586 | free_extent_buffer(next); |
925baedd | 5587 | } |
5f39d397 | 5588 | |
8e73f275 | 5589 | next = c; |
bd681513 | 5590 | next_rw_lock = path->locks[level]; |
8e73f275 | 5591 | ret = read_block_for_search(NULL, root, path, &next, level, |
5d9e75c4 | 5592 | slot, &key, 0); |
8e73f275 CM |
5593 | if (ret == -EAGAIN) |
5594 | goto again; | |
5f39d397 | 5595 | |
76a05b35 | 5596 | if (ret < 0) { |
b3b4aa74 | 5597 | btrfs_release_path(path); |
76a05b35 CM |
5598 | goto done; |
5599 | } | |
5600 | ||
5cd57b2c | 5601 | if (!path->skip_locking) { |
bd681513 | 5602 | ret = btrfs_try_tree_read_lock(next); |
d42244a0 JS |
5603 | if (!ret && time_seq) { |
5604 | /* | |
5605 | * If we don't get the lock, we may be racing | |
5606 | * with push_leaf_left, holding that lock while | |
5607 | * itself waiting for the leaf we've currently | |
5608 | * locked. To solve this situation, we give up | |
5609 | * on our lock and cycle. | |
5610 | */ | |
cf538830 | 5611 | free_extent_buffer(next); |
d42244a0 JS |
5612 | btrfs_release_path(path); |
5613 | cond_resched(); | |
5614 | goto again; | |
5615 | } | |
8e73f275 CM |
5616 | if (!ret) { |
5617 | btrfs_set_path_blocking(path); | |
bd681513 | 5618 | btrfs_tree_read_lock(next); |
31533fb2 | 5619 | btrfs_clear_path_blocking(path, next, |
bd681513 | 5620 | BTRFS_READ_LOCK); |
8e73f275 | 5621 | } |
31533fb2 | 5622 | next_rw_lock = BTRFS_READ_LOCK; |
5cd57b2c | 5623 | } |
d97e63b6 CM |
5624 | break; |
5625 | } | |
5626 | path->slots[level] = slot; | |
d397712b | 5627 | while (1) { |
d97e63b6 CM |
5628 | level--; |
5629 | c = path->nodes[level]; | |
925baedd | 5630 | if (path->locks[level]) |
bd681513 | 5631 | btrfs_tree_unlock_rw(c, path->locks[level]); |
8e73f275 | 5632 | |
5f39d397 | 5633 | free_extent_buffer(c); |
d97e63b6 CM |
5634 | path->nodes[level] = next; |
5635 | path->slots[level] = 0; | |
a74a4b97 | 5636 | if (!path->skip_locking) |
bd681513 | 5637 | path->locks[level] = next_rw_lock; |
d97e63b6 CM |
5638 | if (!level) |
5639 | break; | |
b4ce94de | 5640 | |
8e73f275 | 5641 | ret = read_block_for_search(NULL, root, path, &next, level, |
5d9e75c4 | 5642 | 0, &key, 0); |
8e73f275 CM |
5643 | if (ret == -EAGAIN) |
5644 | goto again; | |
5645 | ||
76a05b35 | 5646 | if (ret < 0) { |
b3b4aa74 | 5647 | btrfs_release_path(path); |
76a05b35 CM |
5648 | goto done; |
5649 | } | |
5650 | ||
5cd57b2c | 5651 | if (!path->skip_locking) { |
bd681513 | 5652 | ret = btrfs_try_tree_read_lock(next); |
8e73f275 CM |
5653 | if (!ret) { |
5654 | btrfs_set_path_blocking(path); | |
bd681513 | 5655 | btrfs_tree_read_lock(next); |
31533fb2 | 5656 | btrfs_clear_path_blocking(path, next, |
bd681513 CM |
5657 | BTRFS_READ_LOCK); |
5658 | } | |
31533fb2 | 5659 | next_rw_lock = BTRFS_READ_LOCK; |
5cd57b2c | 5660 | } |
d97e63b6 | 5661 | } |
8e73f275 | 5662 | ret = 0; |
925baedd | 5663 | done: |
f7c79f30 | 5664 | unlock_up(path, 0, 1, 0, NULL); |
8e73f275 CM |
5665 | path->leave_spinning = old_spinning; |
5666 | if (!old_spinning) | |
5667 | btrfs_set_path_blocking(path); | |
5668 | ||
5669 | return ret; | |
d97e63b6 | 5670 | } |
0b86a832 | 5671 | |
3f157a2f CM |
5672 | /* |
5673 | * this uses btrfs_prev_leaf to walk backwards in the tree, and keeps | |
5674 | * searching until it gets past min_objectid or finds an item of 'type' | |
5675 | * | |
5676 | * returns 0 if something is found, 1 if nothing was found and < 0 on error | |
5677 | */ | |
0b86a832 CM |
5678 | int btrfs_previous_item(struct btrfs_root *root, |
5679 | struct btrfs_path *path, u64 min_objectid, | |
5680 | int type) | |
5681 | { | |
5682 | struct btrfs_key found_key; | |
5683 | struct extent_buffer *leaf; | |
e02119d5 | 5684 | u32 nritems; |
0b86a832 CM |
5685 | int ret; |
5686 | ||
d397712b | 5687 | while (1) { |
0b86a832 | 5688 | if (path->slots[0] == 0) { |
b4ce94de | 5689 | btrfs_set_path_blocking(path); |
0b86a832 CM |
5690 | ret = btrfs_prev_leaf(root, path); |
5691 | if (ret != 0) | |
5692 | return ret; | |
5693 | } else { | |
5694 | path->slots[0]--; | |
5695 | } | |
5696 | leaf = path->nodes[0]; | |
e02119d5 CM |
5697 | nritems = btrfs_header_nritems(leaf); |
5698 | if (nritems == 0) | |
5699 | return 1; | |
5700 | if (path->slots[0] == nritems) | |
5701 | path->slots[0]--; | |
5702 | ||
0b86a832 | 5703 | btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]); |
e02119d5 CM |
5704 | if (found_key.objectid < min_objectid) |
5705 | break; | |
0a4eefbb YZ |
5706 | if (found_key.type == type) |
5707 | return 0; | |
e02119d5 CM |
5708 | if (found_key.objectid == min_objectid && |
5709 | found_key.type < type) | |
5710 | break; | |
0b86a832 CM |
5711 | } |
5712 | return 1; | |
5713 | } |