?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Merge tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm...
[linux-2.6.git] / fs / btrfs / backref.c
blob3775947429b28ea07678c4f58877bfcbb5f8bed6
1 /*
2 * Copyright (C) 2011 STRATO. All rights reserved.
3 *
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU General Public
6 * License v2 as published by the Free Software Foundation.
7 *
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
11 * General Public License for more details.
12 *
13 * You should have received a copy of the GNU General Public
14 * License along with this program; if not, write to the
15 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
16 * Boston, MA 021110-1307, USA.
17 */
18
19 #include <linux/vmalloc.h>
20 #include "ctree.h"
21 #include "disk-io.h"
22 #include "backref.h"
23 #include "ulist.h"
24 #include "transaction.h"
25 #include "delayed-ref.h"
26 #include "locking.h"
27
28 struct extent_inode_elem {
29 u64 inum;
30 u64 offset;
31 struct extent_inode_elem *next;
32 };
33
34 static int check_extent_in_eb(struct btrfs_key *key, struct extent_buffer *eb,
35 struct btrfs_file_extent_item *fi,
36 u64 extent_item_pos,
37 struct extent_inode_elem **eie)
38 {
39 u64 offset = 0;
40 struct extent_inode_elem *e;
41
42 if (!btrfs_file_extent_compression(eb, fi) &&
43 !btrfs_file_extent_encryption(eb, fi) &&
44 !btrfs_file_extent_other_encoding(eb, fi)) {
45 u64 data_offset;
46 u64 data_len;
47
48 data_offset = btrfs_file_extent_offset(eb, fi);
49 data_len = btrfs_file_extent_num_bytes(eb, fi);
50
51 if (extent_item_pos < data_offset ||
52 extent_item_pos >= data_offset + data_len)
53 return 1;
54 offset = extent_item_pos - data_offset;
55 }
56
57 e = kmalloc(sizeof(*e), GFP_NOFS);
58 if (!e)
59 return -ENOMEM;
60
61 e->next = *eie;
62 e->inum = key->objectid;
63 e->offset = key->offset + offset;
64 *eie = e;
65
66 return 0;
67 }
68
69 static int find_extent_in_eb(struct extent_buffer *eb, u64 wanted_disk_byte,
70 u64 extent_item_pos,
71 struct extent_inode_elem **eie)
72 {
73 u64 disk_byte;
74 struct btrfs_key key;
75 struct btrfs_file_extent_item *fi;
76 int slot;
77 int nritems;
78 int extent_type;
79 int ret;
80
81 /*
82 * from the shared data ref, we only have the leaf but we need
83 * the key. thus, we must look into all items and see that we
84 * find one (some) with a reference to our extent item.
85 */
86 nritems = btrfs_header_nritems(eb);
87 for (slot = 0; slot < nritems; ++slot) {
88 btrfs_item_key_to_cpu(eb, &key, slot);
89 if (key.type != BTRFS_EXTENT_DATA_KEY)
90 continue;
91 fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
92 extent_type = btrfs_file_extent_type(eb, fi);
93 if (extent_type == BTRFS_FILE_EXTENT_INLINE)
94 continue;
95 /* don't skip BTRFS_FILE_EXTENT_PREALLOC, we can handle that */
96 disk_byte = btrfs_file_extent_disk_bytenr(eb, fi);
97 if (disk_byte != wanted_disk_byte)
98 continue;
99
100 ret = check_extent_in_eb(&key, eb, fi, extent_item_pos, eie);
101 if (ret < 0)
102 return ret;
103 }
104
105 return 0;
106 }
107
108 /*
109 * this structure records all encountered refs on the way up to the root
110 */
111 struct __prelim_ref {
112 struct list_head list;
113 u64 root_id;
114 struct btrfs_key key_for_search;
115 int level;
116 int count;
117 struct extent_inode_elem *inode_list;
118 u64 parent;
119 u64 wanted_disk_byte;
120 };
121
122 static struct kmem_cache *btrfs_prelim_ref_cache;
123
124 int __init btrfs_prelim_ref_init(void)
125 {
126 btrfs_prelim_ref_cache = kmem_cache_create("btrfs_prelim_ref",
127 sizeof(struct __prelim_ref),
128 0,
129 SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD,
130 NULL);
131 if (!btrfs_prelim_ref_cache)
132 return -ENOMEM;
133 return 0;
134 }
135
136 void btrfs_prelim_ref_exit(void)
137 {
138 if (btrfs_prelim_ref_cache)
139 kmem_cache_destroy(btrfs_prelim_ref_cache);
140 }
141
142 /*
143 * the rules for all callers of this function are:
144 * - obtaining the parent is the goal
145 * - if you add a key, you must know that it is a correct key
146 * - if you cannot add the parent or a correct key, then we will look into the
147 * block later to set a correct key
148 *
149 * delayed refs
150 * ============
151 * backref type | shared | indirect | shared | indirect
152 * information | tree | tree | data | data
153 * --------------------+--------+----------+--------+----------
154 * parent logical | y | - | - | -
155 * key to resolve | - | y | y | y
156 * tree block logical | - | - | - | -
157 * root for resolving | y | y | y | y
158 *
159 * - column 1: we've the parent -> done
160 * - column 2, 3, 4: we use the key to find the parent
161 *
162 * on disk refs (inline or keyed)
163 * ==============================
164 * backref type | shared | indirect | shared | indirect
165 * information | tree | tree | data | data
166 * --------------------+--------+----------+--------+----------
167 * parent logical | y | - | y | -
168 * key to resolve | - | - | - | y
169 * tree block logical | y | y | y | y
170 * root for resolving | - | y | y | y
171 *
172 * - column 1, 3: we've the parent -> done
173 * - column 2: we take the first key from the block to find the parent
174 * (see __add_missing_keys)
175 * - column 4: we use the key to find the parent
176 *
177 * additional information that's available but not required to find the parent
178 * block might help in merging entries to gain some speed.
179 */
180
181 static int __add_prelim_ref(struct list_head *head, u64 root_id,
182 struct btrfs_key *key, int level,
183 u64 parent, u64 wanted_disk_byte, int count,
184 gfp_t gfp_mask)
185 {
186 struct __prelim_ref *ref;
187
188 if (root_id == BTRFS_DATA_RELOC_TREE_OBJECTID)
189 return 0;
190
191 ref = kmem_cache_alloc(btrfs_prelim_ref_cache, gfp_mask);
192 if (!ref)
193 return -ENOMEM;
194
195 ref->root_id = root_id;
196 if (key)
197 ref->key_for_search = *key;
198 else
199 memset(&ref->key_for_search, 0, sizeof(ref->key_for_search));
200
201 ref->inode_list = NULL;
202 ref->level = level;
203 ref->count = count;
204 ref->parent = parent;
205 ref->wanted_disk_byte = wanted_disk_byte;
206 list_add_tail(&ref->list, head);
207
208 return 0;
209 }
210
211 static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path,
212 struct ulist *parents, int level,
213 struct btrfs_key *key_for_search, u64 time_seq,
214 u64 wanted_disk_byte,
215 const u64 *extent_item_pos)
216 {
217 int ret = 0;
218 int slot;
219 struct extent_buffer *eb;
220 struct btrfs_key key;
221 struct btrfs_file_extent_item *fi;
222 struct extent_inode_elem *eie = NULL, *old = NULL;
223 u64 disk_byte;
224
225 if (level != 0) {
226 eb = path->nodes[level];
227 ret = ulist_add(parents, eb->start, 0, GFP_NOFS);
228 if (ret < 0)
229 return ret;
230 return 0;
231 }
232
233 /*
234 * We normally enter this function with the path already pointing to
235 * the first item to check. But sometimes, we may enter it with
236 * slot==nritems. In that case, go to the next leaf before we continue.
237 */
238 if (path->slots[0] >= btrfs_header_nritems(path->nodes[0]))
239 ret = btrfs_next_old_leaf(root, path, time_seq);
240
241 while (!ret) {
242 eb = path->nodes[0];
243 slot = path->slots[0];
244
245 btrfs_item_key_to_cpu(eb, &key, slot);
246
247 if (key.objectid != key_for_search->objectid ||
248 key.type != BTRFS_EXTENT_DATA_KEY)
249 break;
250
251 fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
252 disk_byte = btrfs_file_extent_disk_bytenr(eb, fi);
253
254 if (disk_byte == wanted_disk_byte) {
255 eie = NULL;
256 old = NULL;
257 if (extent_item_pos) {
258 ret = check_extent_in_eb(&key, eb, fi,
259 *extent_item_pos,
260 &eie);
261 if (ret < 0)
262 break;
263 }
264 if (ret > 0)
265 goto next;
266 ret = ulist_add_merge(parents, eb->start,
267 (uintptr_t)eie,
268 (u64 *)&old, GFP_NOFS);
269 if (ret < 0)
270 break;
271 if (!ret && extent_item_pos) {
272 while (old->next)
273 old = old->next;
274 old->next = eie;
275 }
276 }
277 next:
278 ret = btrfs_next_old_item(root, path, time_seq);
279 }
280
281 if (ret > 0)
282 ret = 0;
283 return ret;
284 }
285
286 /*
287 * resolve an indirect backref in the form (root_id, key, level)
288 * to a logical address
289 */
290 static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info,
291 struct btrfs_path *path, u64 time_seq,
292 struct __prelim_ref *ref,
293 struct ulist *parents,
294 const u64 *extent_item_pos)
295 {
296 struct btrfs_root *root;
297 struct btrfs_key root_key;
298 struct extent_buffer *eb;
299 int ret = 0;
300 int root_level;
301 int level = ref->level;
302
303 root_key.objectid = ref->root_id;
304 root_key.type = BTRFS_ROOT_ITEM_KEY;
305 root_key.offset = (u64)-1;
306 root = btrfs_read_fs_root_no_name(fs_info, &root_key);
307 if (IS_ERR(root)) {
308 ret = PTR_ERR(root);
309 goto out;
310 }
311
312 root_level = btrfs_old_root_level(root, time_seq);
313
314 if (root_level + 1 == level)
315 goto out;
316
317 path->lowest_level = level;
318 ret = btrfs_search_old_slot(root, &ref->key_for_search, path, time_seq);
319 pr_debug("search slot in root %llu (level %d, ref count %d) returned "
320 "%d for key (%llu %u %llu)\n",
321 ref->root_id, level, ref->count, ret,
322 ref->key_for_search.objectid, ref->key_for_search.type,
323 ref->key_for_search.offset);
324 if (ret < 0)
325 goto out;
326
327 eb = path->nodes[level];
328 while (!eb) {
329 if (WARN_ON(!level)) {
330 ret = 1;
331 goto out;
332 }
333 level--;
334 eb = path->nodes[level];
335 }
336
337 ret = add_all_parents(root, path, parents, level, &ref->key_for_search,
338 time_seq, ref->wanted_disk_byte,
339 extent_item_pos);
340 out:
341 path->lowest_level = 0;
342 btrfs_release_path(path);
343 return ret;
344 }
345
346 /*
347 * resolve all indirect backrefs from the list
348 */
349 static int __resolve_indirect_refs(struct btrfs_fs_info *fs_info,
350 struct btrfs_path *path, u64 time_seq,
351 struct list_head *head,
352 const u64 *extent_item_pos)
353 {
354 int err;
355 int ret = 0;
356 struct __prelim_ref *ref;
357 struct __prelim_ref *ref_safe;
358 struct __prelim_ref *new_ref;
359 struct ulist *parents;
360 struct ulist_node *node;
361 struct ulist_iterator uiter;
362
363 parents = ulist_alloc(GFP_NOFS);
364 if (!parents)
365 return -ENOMEM;
366
367 /*
368 * _safe allows us to insert directly after the current item without
369 * iterating over the newly inserted items.
370 * we're also allowed to re-assign ref during iteration.
371 */
372 list_for_each_entry_safe(ref, ref_safe, head, list) {
373 if (ref->parent) /* already direct */
374 continue;
375 if (ref->count == 0)
376 continue;
377 err = __resolve_indirect_ref(fs_info, path, time_seq, ref,
378 parents, extent_item_pos);
379 if (err == -ENOMEM)
380 goto out;
381 if (err)
382 continue;
383
384 /* we put the first parent into the ref at hand */
385 ULIST_ITER_INIT(&uiter);
386 node = ulist_next(parents, &uiter);
387 ref->parent = node ? node->val : 0;
388 ref->inode_list = node ?
389 (struct extent_inode_elem *)(uintptr_t)node->aux : NULL;
390
391 /* additional parents require new refs being added here */
392 while ((node = ulist_next(parents, &uiter))) {
393 new_ref = kmem_cache_alloc(btrfs_prelim_ref_cache,
394 GFP_NOFS);
395 if (!new_ref) {
396 ret = -ENOMEM;
397 goto out;
398 }
399 memcpy(new_ref, ref, sizeof(*ref));
400 new_ref->parent = node->val;
401 new_ref->inode_list = (struct extent_inode_elem *)
402 (uintptr_t)node->aux;
403 list_add(&new_ref->list, &ref->list);
404 }
405 ulist_reinit(parents);
406 }
407 out:
408 ulist_free(parents);
409 return ret;
410 }
411
412 static inline int ref_for_same_block(struct __prelim_ref *ref1,
413 struct __prelim_ref *ref2)
414 {
415 if (ref1->level != ref2->level)
416 return 0;
417 if (ref1->root_id != ref2->root_id)
418 return 0;
419 if (ref1->key_for_search.type != ref2->key_for_search.type)
420 return 0;
421 if (ref1->key_for_search.objectid != ref2->key_for_search.objectid)
422 return 0;
423 if (ref1->key_for_search.offset != ref2->key_for_search.offset)
424 return 0;
425 if (ref1->parent != ref2->parent)
426 return 0;
427
428 return 1;
429 }
430
431 /*
432 * read tree blocks and add keys where required.
433 */
434 static int __add_missing_keys(struct btrfs_fs_info *fs_info,
435 struct list_head *head)
436 {
437 struct list_head *pos;
438 struct extent_buffer *eb;
439
440 list_for_each(pos, head) {
441 struct __prelim_ref *ref;
442 ref = list_entry(pos, struct __prelim_ref, list);
443
444 if (ref->parent)
445 continue;
446 if (ref->key_for_search.type)
447 continue;
448 BUG_ON(!ref->wanted_disk_byte);
449 eb = read_tree_block(fs_info->tree_root, ref->wanted_disk_byte,
450 fs_info->tree_root->leafsize, 0);
451 if (!eb || !extent_buffer_uptodate(eb)) {
452 free_extent_buffer(eb);
453 return -EIO;
454 }
455 btrfs_tree_read_lock(eb);
456 if (btrfs_header_level(eb) == 0)
457 btrfs_item_key_to_cpu(eb, &ref->key_for_search, 0);
458 else
459 btrfs_node_key_to_cpu(eb, &ref->key_for_search, 0);
460 btrfs_tree_read_unlock(eb);
461 free_extent_buffer(eb);
462 }
463 return 0;
464 }
465
466 /*
467 * merge two lists of backrefs and adjust counts accordingly
468 *
469 * mode = 1: merge identical keys, if key is set
470 * FIXME: if we add more keys in __add_prelim_ref, we can merge more here.
471 * additionally, we could even add a key range for the blocks we
472 * looked into to merge even more (-> replace unresolved refs by those
473 * having a parent).
474 * mode = 2: merge identical parents
475 */
476 static void __merge_refs(struct list_head *head, int mode)
477 {
478 struct list_head *pos1;
479
480 list_for_each(pos1, head) {
481 struct list_head *n2;
482 struct list_head *pos2;
483 struct __prelim_ref *ref1;
484
485 ref1 = list_entry(pos1, struct __prelim_ref, list);
486
487 for (pos2 = pos1->next, n2 = pos2->next; pos2 != head;
488 pos2 = n2, n2 = pos2->next) {
489 struct __prelim_ref *ref2;
490 struct __prelim_ref *xchg;
491 struct extent_inode_elem *eie;
492
493 ref2 = list_entry(pos2, struct __prelim_ref, list);
494
495 if (mode == 1) {
496 if (!ref_for_same_block(ref1, ref2))
497 continue;
498 if (!ref1->parent && ref2->parent) {
499 xchg = ref1;
500 ref1 = ref2;
501 ref2 = xchg;
502 }
503 } else {
504 if (ref1->parent != ref2->parent)
505 continue;
506 }
507
508 eie = ref1->inode_list;
509 while (eie && eie->next)
510 eie = eie->next;
511 if (eie)
512 eie->next = ref2->inode_list;
513 else
514 ref1->inode_list = ref2->inode_list;
515 ref1->count += ref2->count;
516
517 list_del(&ref2->list);
518 kmem_cache_free(btrfs_prelim_ref_cache, ref2);
519 }
520
521 }
522 }
523
524 /*
525 * add all currently queued delayed refs from this head whose seq nr is
526 * smaller or equal that seq to the list
527 */
528 static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq,
529 struct list_head *prefs)
530 {
531 struct btrfs_delayed_extent_op *extent_op = head->extent_op;
532 struct rb_node *n = &head->node.rb_node;
533 struct btrfs_key key;
534 struct btrfs_key op_key = {0};
535 int sgn;
536 int ret = 0;
537
538 if (extent_op && extent_op->update_key)
539 btrfs_disk_key_to_cpu(&op_key, &extent_op->key);
540
541 while ((n = rb_prev(n))) {
542 struct btrfs_delayed_ref_node *node;
543 node = rb_entry(n, struct btrfs_delayed_ref_node,
544 rb_node);
545 if (node->bytenr != head->node.bytenr)
546 break;
547 WARN_ON(node->is_head);
548
549 if (node->seq > seq)
550 continue;
551
552 switch (node->action) {
553 case BTRFS_ADD_DELAYED_EXTENT:
554 case BTRFS_UPDATE_DELAYED_HEAD:
555 WARN_ON(1);
556 continue;
557 case BTRFS_ADD_DELAYED_REF:
558 sgn = 1;
559 break;
560 case BTRFS_DROP_DELAYED_REF:
561 sgn = -1;
562 break;
563 default:
564 BUG_ON(1);
565 }
566 switch (node->type) {
567 case BTRFS_TREE_BLOCK_REF_KEY: {
568 struct btrfs_delayed_tree_ref *ref;
569
570 ref = btrfs_delayed_node_to_tree_ref(node);
571 ret = __add_prelim_ref(prefs, ref->root, &op_key,
572 ref->level + 1, 0, node->bytenr,
573 node->ref_mod * sgn, GFP_ATOMIC);
574 break;
575 }
576 case BTRFS_SHARED_BLOCK_REF_KEY: {
577 struct btrfs_delayed_tree_ref *ref;
578
579 ref = btrfs_delayed_node_to_tree_ref(node);
580 ret = __add_prelim_ref(prefs, ref->root, NULL,
581 ref->level + 1, ref->parent,
582 node->bytenr,
583 node->ref_mod * sgn, GFP_ATOMIC);
584 break;
585 }
586 case BTRFS_EXTENT_DATA_REF_KEY: {
587 struct btrfs_delayed_data_ref *ref;
588 ref = btrfs_delayed_node_to_data_ref(node);
589
590 key.objectid = ref->objectid;
591 key.type = BTRFS_EXTENT_DATA_KEY;
592 key.offset = ref->offset;
593 ret = __add_prelim_ref(prefs, ref->root, &key, 0, 0,
594 node->bytenr,
595 node->ref_mod * sgn, GFP_ATOMIC);
596 break;
597 }
598 case BTRFS_SHARED_DATA_REF_KEY: {
599 struct btrfs_delayed_data_ref *ref;
600
601 ref = btrfs_delayed_node_to_data_ref(node);
602
603 key.objectid = ref->objectid;
604 key.type = BTRFS_EXTENT_DATA_KEY;
605 key.offset = ref->offset;
606 ret = __add_prelim_ref(prefs, ref->root, &key, 0,
607 ref->parent, node->bytenr,
608 node->ref_mod * sgn, GFP_ATOMIC);
609 break;
610 }
611 default:
612 WARN_ON(1);
613 }
614 if (ret)
615 return ret;
616 }
617
618 return 0;
619 }
620
621 /*
622 * add all inline backrefs for bytenr to the list
623 */
624 static int __add_inline_refs(struct btrfs_fs_info *fs_info,
625 struct btrfs_path *path, u64 bytenr,
626 int *info_level, struct list_head *prefs)
627 {
628 int ret = 0;
629 int slot;
630 struct extent_buffer *leaf;
631 struct btrfs_key key;
632 struct btrfs_key found_key;
633 unsigned long ptr;
634 unsigned long end;
635 struct btrfs_extent_item *ei;
636 u64 flags;
637 u64 item_size;
638
639 /*
640 * enumerate all inline refs
641 */
642 leaf = path->nodes[0];
643 slot = path->slots[0];
644
645 item_size = btrfs_item_size_nr(leaf, slot);
646 BUG_ON(item_size < sizeof(*ei));
647
648 ei = btrfs_item_ptr(leaf, slot, struct btrfs_extent_item);
649 flags = btrfs_extent_flags(leaf, ei);
650 btrfs_item_key_to_cpu(leaf, &found_key, slot);
651
652 ptr = (unsigned long)(ei + 1);
653 end = (unsigned long)ei + item_size;
654
655 if (found_key.type == BTRFS_EXTENT_ITEM_KEY &&
656 flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
657 struct btrfs_tree_block_info *info;
658
659 info = (struct btrfs_tree_block_info *)ptr;
660 *info_level = btrfs_tree_block_level(leaf, info);
661 ptr += sizeof(struct btrfs_tree_block_info);
662 BUG_ON(ptr > end);
663 } else if (found_key.type == BTRFS_METADATA_ITEM_KEY) {
664 *info_level = found_key.offset;
665 } else {
666 BUG_ON(!(flags & BTRFS_EXTENT_FLAG_DATA));
667 }
668
669 while (ptr < end) {
670 struct btrfs_extent_inline_ref *iref;
671 u64 offset;
672 int type;
673
674 iref = (struct btrfs_extent_inline_ref *)ptr;
675 type = btrfs_extent_inline_ref_type(leaf, iref);
676 offset = btrfs_extent_inline_ref_offset(leaf, iref);
677
678 switch (type) {
679 case BTRFS_SHARED_BLOCK_REF_KEY:
680 ret = __add_prelim_ref(prefs, 0, NULL,
681 *info_level + 1, offset,
682 bytenr, 1, GFP_NOFS);
683 break;
684 case BTRFS_SHARED_DATA_REF_KEY: {
685 struct btrfs_shared_data_ref *sdref;
686 int count;
687
688 sdref = (struct btrfs_shared_data_ref *)(iref + 1);
689 count = btrfs_shared_data_ref_count(leaf, sdref);
690 ret = __add_prelim_ref(prefs, 0, NULL, 0, offset,
691 bytenr, count, GFP_NOFS);
692 break;
693 }
694 case BTRFS_TREE_BLOCK_REF_KEY:
695 ret = __add_prelim_ref(prefs, offset, NULL,
696 *info_level + 1, 0,
697 bytenr, 1, GFP_NOFS);
698 break;
699 case BTRFS_EXTENT_DATA_REF_KEY: {
700 struct btrfs_extent_data_ref *dref;
701 int count;
702 u64 root;
703
704 dref = (struct btrfs_extent_data_ref *)(&iref->offset);
705 count = btrfs_extent_data_ref_count(leaf, dref);
706 key.objectid = btrfs_extent_data_ref_objectid(leaf,
707 dref);
708 key.type = BTRFS_EXTENT_DATA_KEY;
709 key.offset = btrfs_extent_data_ref_offset(leaf, dref);
710 root = btrfs_extent_data_ref_root(leaf, dref);
711 ret = __add_prelim_ref(prefs, root, &key, 0, 0,
712 bytenr, count, GFP_NOFS);
713 break;
714 }
715 default:
716 WARN_ON(1);
717 }
718 if (ret)
719 return ret;
720 ptr += btrfs_extent_inline_ref_size(type);
721 }
722
723 return 0;
724 }
725
726 /*
727 * add all non-inline backrefs for bytenr to the list
728 */
729 static int __add_keyed_refs(struct btrfs_fs_info *fs_info,
730 struct btrfs_path *path, u64 bytenr,
731 int info_level, struct list_head *prefs)
732 {
733 struct btrfs_root *extent_root = fs_info->extent_root;
734 int ret;
735 int slot;
736 struct extent_buffer *leaf;
737 struct btrfs_key key;
738
739 while (1) {
740 ret = btrfs_next_item(extent_root, path);
741 if (ret < 0)
742 break;
743 if (ret) {
744 ret = 0;
745 break;
746 }
747
748 slot = path->slots[0];
749 leaf = path->nodes[0];
750 btrfs_item_key_to_cpu(leaf, &key, slot);
751
752 if (key.objectid != bytenr)
753 break;
754 if (key.type < BTRFS_TREE_BLOCK_REF_KEY)
755 continue;
756 if (key.type > BTRFS_SHARED_DATA_REF_KEY)
757 break;
758
759 switch (key.type) {
760 case BTRFS_SHARED_BLOCK_REF_KEY:
761 ret = __add_prelim_ref(prefs, 0, NULL,
762 info_level + 1, key.offset,
763 bytenr, 1, GFP_NOFS);
764 break;
765 case BTRFS_SHARED_DATA_REF_KEY: {
766 struct btrfs_shared_data_ref *sdref;
767 int count;
768
769 sdref = btrfs_item_ptr(leaf, slot,
770 struct btrfs_shared_data_ref);
771 count = btrfs_shared_data_ref_count(leaf, sdref);
772 ret = __add_prelim_ref(prefs, 0, NULL, 0, key.offset,
773 bytenr, count, GFP_NOFS);
774 break;
775 }
776 case BTRFS_TREE_BLOCK_REF_KEY:
777 ret = __add_prelim_ref(prefs, key.offset, NULL,
778 info_level + 1, 0,
779 bytenr, 1, GFP_NOFS);
780 break;
781 case BTRFS_EXTENT_DATA_REF_KEY: {
782 struct btrfs_extent_data_ref *dref;
783 int count;
784 u64 root;
785
786 dref = btrfs_item_ptr(leaf, slot,
787 struct btrfs_extent_data_ref);
788 count = btrfs_extent_data_ref_count(leaf, dref);
789 key.objectid = btrfs_extent_data_ref_objectid(leaf,
790 dref);
791 key.type = BTRFS_EXTENT_DATA_KEY;
792 key.offset = btrfs_extent_data_ref_offset(leaf, dref);
793 root = btrfs_extent_data_ref_root(leaf, dref);
794 ret = __add_prelim_ref(prefs, root, &key, 0, 0,
795 bytenr, count, GFP_NOFS);
796 break;
797 }
798 default:
799 WARN_ON(1);
800 }
801 if (ret)
802 return ret;
803
804 }
805
806 return ret;
807 }
808
809 /*
810 * this adds all existing backrefs (inline backrefs, backrefs and delayed
811 * refs) for the given bytenr to the refs list, merges duplicates and resolves
812 * indirect refs to their parent bytenr.
813 * When roots are found, they're added to the roots list
814 *
815 * FIXME some caching might speed things up
816 */
817 static int find_parent_nodes(struct btrfs_trans_handle *trans,
818 struct btrfs_fs_info *fs_info, u64 bytenr,
819 u64 time_seq, struct ulist *refs,
820 struct ulist *roots, const u64 *extent_item_pos)
821 {
822 struct btrfs_key key;
823 struct btrfs_path *path;
824 struct btrfs_delayed_ref_root *delayed_refs = NULL;
825 struct btrfs_delayed_ref_head *head;
826 int info_level = 0;
827 int ret;
828 struct list_head prefs_delayed;
829 struct list_head prefs;
830 struct __prelim_ref *ref;
831
832 INIT_LIST_HEAD(&prefs);
833 INIT_LIST_HEAD(&prefs_delayed);
834
835 key.objectid = bytenr;
836 key.offset = (u64)-1;
837 if (btrfs_fs_incompat(fs_info, SKINNY_METADATA))
838 key.type = BTRFS_METADATA_ITEM_KEY;
839 else
840 key.type = BTRFS_EXTENT_ITEM_KEY;
841
842 path = btrfs_alloc_path();
843 if (!path)
844 return -ENOMEM;
845 if (!trans)
846 path->search_commit_root = 1;
847
848 /*
849 * grab both a lock on the path and a lock on the delayed ref head.
850 * We need both to get a consistent picture of how the refs look
851 * at a specified point in time
852 */
853 again:
854 head = NULL;
855
856 ret = btrfs_search_slot(trans, fs_info->extent_root, &key, path, 0, 0);
857 if (ret < 0)
858 goto out;
859 BUG_ON(ret == 0);
860
861 if (trans) {
862 /*
863 * look if there are updates for this ref queued and lock the
864 * head
865 */
866 delayed_refs = &trans->transaction->delayed_refs;
867 spin_lock(&delayed_refs->lock);
868 head = btrfs_find_delayed_ref_head(trans, bytenr);
869 if (head) {
870 if (!mutex_trylock(&head->mutex)) {
871 atomic_inc(&head->node.refs);
872 spin_unlock(&delayed_refs->lock);
873
874 btrfs_release_path(path);
875
876 /*
877 * Mutex was contended, block until it's
878 * released and try again
879 */
880 mutex_lock(&head->mutex);
881 mutex_unlock(&head->mutex);
882 btrfs_put_delayed_ref(&head->node);
883 goto again;
884 }
885 ret = __add_delayed_refs(head, time_seq,
886 &prefs_delayed);
887 mutex_unlock(&head->mutex);
888 if (ret) {
889 spin_unlock(&delayed_refs->lock);
890 goto out;
891 }
892 }
893 spin_unlock(&delayed_refs->lock);
894 }
895
896 if (path->slots[0]) {
897 struct extent_buffer *leaf;
898 int slot;
899
900 path->slots[0]--;
901 leaf = path->nodes[0];
902 slot = path->slots[0];
903 btrfs_item_key_to_cpu(leaf, &key, slot);
904 if (key.objectid == bytenr &&
905 (key.type == BTRFS_EXTENT_ITEM_KEY ||
906 key.type == BTRFS_METADATA_ITEM_KEY)) {
907 ret = __add_inline_refs(fs_info, path, bytenr,
908 &info_level, &prefs);
909 if (ret)
910 goto out;
911 ret = __add_keyed_refs(fs_info, path, bytenr,
912 info_level, &prefs);
913 if (ret)
914 goto out;
915 }
916 }
917 btrfs_release_path(path);
918
919 list_splice_init(&prefs_delayed, &prefs);
920
921 ret = __add_missing_keys(fs_info, &prefs);
922 if (ret)
923 goto out;
924
925 __merge_refs(&prefs, 1);
926
927 ret = __resolve_indirect_refs(fs_info, path, time_seq, &prefs,
928 extent_item_pos);
929 if (ret)
930 goto out;
931
932 __merge_refs(&prefs, 2);
933
934 while (!list_empty(&prefs)) {
935 ref = list_first_entry(&prefs, struct __prelim_ref, list);
936 WARN_ON(ref->count < 0);
937 if (ref->count && ref->root_id && ref->parent == 0) {
938 /* no parent == root of tree */
939 ret = ulist_add(roots, ref->root_id, 0, GFP_NOFS);
940 if (ret < 0)
941 goto out;
942 }
943 if (ref->count && ref->parent) {
944 struct extent_inode_elem *eie = NULL;
945 if (extent_item_pos && !ref->inode_list) {
946 u32 bsz;
947 struct extent_buffer *eb;
948 bsz = btrfs_level_size(fs_info->extent_root,
949 info_level);
950 eb = read_tree_block(fs_info->extent_root,
951 ref->parent, bsz, 0);
952 if (!eb || !extent_buffer_uptodate(eb)) {
953 free_extent_buffer(eb);
954 ret = -EIO;
955 goto out;
956 }
957 ret = find_extent_in_eb(eb, bytenr,
958 *extent_item_pos, &eie);
959 free_extent_buffer(eb);
960 if (ret < 0)
961 goto out;
962 ref->inode_list = eie;
963 }
964 ret = ulist_add_merge(refs, ref->parent,
965 (uintptr_t)ref->inode_list,
966 (u64 *)&eie, GFP_NOFS);
967 if (ret < 0)
968 goto out;
969 if (!ret && extent_item_pos) {
970 /*
971 * we've recorded that parent, so we must extend
972 * its inode list here
973 */
974 BUG_ON(!eie);
975 while (eie->next)
976 eie = eie->next;
977 eie->next = ref->inode_list;
978 }
979 }
980 list_del(&ref->list);
981 kmem_cache_free(btrfs_prelim_ref_cache, ref);
982 }
983
984 out:
985 btrfs_free_path(path);
986 while (!list_empty(&prefs)) {
987 ref = list_first_entry(&prefs, struct __prelim_ref, list);
988 list_del(&ref->list);
989 kmem_cache_free(btrfs_prelim_ref_cache, ref);
990 }
991 while (!list_empty(&prefs_delayed)) {
992 ref = list_first_entry(&prefs_delayed, struct __prelim_ref,
993 list);
994 list_del(&ref->list);
995 kmem_cache_free(btrfs_prelim_ref_cache, ref);
996 }
997
998 return ret;
999 }
1000
1001 static void free_leaf_list(struct ulist *blocks)
1002 {
1003 struct ulist_node *node = NULL;
1004 struct extent_inode_elem *eie;
1005 struct extent_inode_elem *eie_next;
1006 struct ulist_iterator uiter;
1007
1008 ULIST_ITER_INIT(&uiter);
1009 while ((node = ulist_next(blocks, &uiter))) {
1010 if (!node->aux)
1011 continue;
1012 eie = (struct extent_inode_elem *)(uintptr_t)node->aux;
1013 for (; eie; eie = eie_next) {
1014 eie_next = eie->next;
1015 kfree(eie);
1016 }
1017 node->aux = 0;
1018 }
1019
1020 ulist_free(blocks);
1021 }
1022
1023 /*
1024 * Finds all leafs with a reference to the specified combination of bytenr and
1025 * offset. key_list_head will point to a list of corresponding keys (caller must
1026 * free each list element). The leafs will be stored in the leafs ulist, which
1027 * must be freed with ulist_free.
1028 *
1029 * returns 0 on success, <0 on error
1030 */
1031 static int btrfs_find_all_leafs(struct btrfs_trans_handle *trans,
1032 struct btrfs_fs_info *fs_info, u64 bytenr,
1033 u64 time_seq, struct ulist **leafs,
1034 const u64 *extent_item_pos)
1035 {
1036 struct ulist *tmp;
1037 int ret;
1038
1039 tmp = ulist_alloc(GFP_NOFS);
1040 if (!tmp)
1041 return -ENOMEM;
1042 *leafs = ulist_alloc(GFP_NOFS);
1043 if (!*leafs) {
1044 ulist_free(tmp);
1045 return -ENOMEM;
1046 }
1047
1048 ret = find_parent_nodes(trans, fs_info, bytenr,
1049 time_seq, *leafs, tmp, extent_item_pos);
1050 ulist_free(tmp);
1051
1052 if (ret < 0 && ret != -ENOENT) {
1053 free_leaf_list(*leafs);
1054 return ret;
1055 }
1056
1057 return 0;
1058 }
1059
1060 /*
1061 * walk all backrefs for a given extent to find all roots that reference this
1062 * extent. Walking a backref means finding all extents that reference this
1063 * extent and in turn walk the backrefs of those, too. Naturally this is a
1064 * recursive process, but here it is implemented in an iterative fashion: We
1065 * find all referencing extents for the extent in question and put them on a
1066 * list. In turn, we find all referencing extents for those, further appending
1067 * to the list. The way we iterate the list allows adding more elements after
1068 * the current while iterating. The process stops when we reach the end of the
1069 * list. Found roots are added to the roots list.
1070 *
1071 * returns 0 on success, < 0 on error.
1072 */
1073 int btrfs_find_all_roots(struct btrfs_trans_handle *trans,
1074 struct btrfs_fs_info *fs_info, u64 bytenr,
1075 u64 time_seq, struct ulist **roots)
1076 {
1077 struct ulist *tmp;
1078 struct ulist_node *node = NULL;
1079 struct ulist_iterator uiter;
1080 int ret;
1081
1082 tmp = ulist_alloc(GFP_NOFS);
1083 if (!tmp)
1084 return -ENOMEM;
1085 *roots = ulist_alloc(GFP_NOFS);
1086 if (!*roots) {
1087 ulist_free(tmp);
1088 return -ENOMEM;
1089 }
1090
1091 ULIST_ITER_INIT(&uiter);
1092 while (1) {
1093 ret = find_parent_nodes(trans, fs_info, bytenr,
1094 time_seq, tmp, *roots, NULL);
1095 if (ret < 0 && ret != -ENOENT) {
1096 ulist_free(tmp);
1097 ulist_free(*roots);
1098 return ret;
1099 }
1100 node = ulist_next(tmp, &uiter);
1101 if (!node)
1102 break;
1103 bytenr = node->val;
1104 }
1105
1106 ulist_free(tmp);
1107 return 0;
1108 }
1109
1110
1111 static int __inode_info(u64 inum, u64 ioff, u8 key_type,
1112 struct btrfs_root *fs_root, struct btrfs_path *path,
1113 struct btrfs_key *found_key)
1114 {
1115 int ret;
1116 struct btrfs_key key;
1117 struct extent_buffer *eb;
1118
1119 key.type = key_type;
1120 key.objectid = inum;
1121 key.offset = ioff;
1122
1123 ret = btrfs_search_slot(NULL, fs_root, &key, path, 0, 0);
1124 if (ret < 0)
1125 return ret;
1126
1127 eb = path->nodes[0];
1128 if (ret && path->slots[0] >= btrfs_header_nritems(eb)) {
1129 ret = btrfs_next_leaf(fs_root, path);
1130 if (ret)
1131 return ret;
1132 eb = path->nodes[0];
1133 }
1134
1135 btrfs_item_key_to_cpu(eb, found_key, path->slots[0]);
1136 if (found_key->type != key.type || found_key->objectid != key.objectid)
1137 return 1;
1138
1139 return 0;
1140 }
1141
1142 /*
1143 * this makes the path point to (inum INODE_ITEM ioff)
1144 */
1145 int inode_item_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
1146 struct btrfs_path *path)
1147 {
1148 struct btrfs_key key;
1149 return __inode_info(inum, ioff, BTRFS_INODE_ITEM_KEY, fs_root, path,
1150 &key);
1151 }
1152
1153 static int inode_ref_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
1154 struct btrfs_path *path,
1155 struct btrfs_key *found_key)
1156 {
1157 return __inode_info(inum, ioff, BTRFS_INODE_REF_KEY, fs_root, path,
1158 found_key);
1159 }
1160
1161 int btrfs_find_one_extref(struct btrfs_root *root, u64 inode_objectid,
1162 u64 start_off, struct btrfs_path *path,
1163 struct btrfs_inode_extref **ret_extref,
1164 u64 *found_off)
1165 {
1166 int ret, slot;
1167 struct btrfs_key key;
1168 struct btrfs_key found_key;
1169 struct btrfs_inode_extref *extref;
1170 struct extent_buffer *leaf;
1171 unsigned long ptr;
1172
1173 key.objectid = inode_objectid;
1174 btrfs_set_key_type(&key, BTRFS_INODE_EXTREF_KEY);
1175 key.offset = start_off;
1176
1177 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
1178 if (ret < 0)
1179 return ret;
1180
1181 while (1) {
1182 leaf = path->nodes[0];
1183 slot = path->slots[0];
1184 if (slot >= btrfs_header_nritems(leaf)) {
1185 /*
1186 * If the item at offset is not found,
1187 * btrfs_search_slot will point us to the slot
1188 * where it should be inserted. In our case
1189 * that will be the slot directly before the
1190 * next INODE_REF_KEY_V2 item. In the case
1191 * that we're pointing to the last slot in a
1192 * leaf, we must move one leaf over.
1193 */
1194 ret = btrfs_next_leaf(root, path);
1195 if (ret) {
1196 if (ret >= 1)
1197 ret = -ENOENT;
1198 break;
1199 }
1200 continue;
1201 }
1202
1203 btrfs_item_key_to_cpu(leaf, &found_key, slot);
1204
1205 /*
1206 * Check that we're still looking at an extended ref key for
1207 * this particular objectid. If we have different
1208 * objectid or type then there are no more to be found
1209 * in the tree and we can exit.
1210 */
1211 ret = -ENOENT;
1212 if (found_key.objectid != inode_objectid)
1213 break;
1214 if (btrfs_key_type(&found_key) != BTRFS_INODE_EXTREF_KEY)
1215 break;
1216
1217 ret = 0;
1218 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
1219 extref = (struct btrfs_inode_extref *)ptr;
1220 *ret_extref = extref;
1221 if (found_off)
1222 *found_off = found_key.offset;
1223 break;
1224 }
1225
1226 return ret;
1227 }
1228
1229 /*
1230 * this iterates to turn a name (from iref/extref) into a full filesystem path.
1231 * Elements of the path are separated by '/' and the path is guaranteed to be
1232 * 0-terminated. the path is only given within the current file system.
1233 * Therefore, it never starts with a '/'. the caller is responsible to provide
1234 * "size" bytes in "dest". the dest buffer will be filled backwards. finally,
1235 * the start point of the resulting string is returned. this pointer is within
1236 * dest, normally.
1237 * in case the path buffer would overflow, the pointer is decremented further
1238 * as if output was written to the buffer, though no more output is actually
1239 * generated. that way, the caller can determine how much space would be
1240 * required for the path to fit into the buffer. in that case, the returned
1241 * value will be smaller than dest. callers must check this!
1242 */
1243 char *btrfs_ref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
1244 u32 name_len, unsigned long name_off,
1245 struct extent_buffer *eb_in, u64 parent,
1246 char *dest, u32 size)
1247 {
1248 int slot;
1249 u64 next_inum;
1250 int ret;
1251 s64 bytes_left = ((s64)size) - 1;
1252 struct extent_buffer *eb = eb_in;
1253 struct btrfs_key found_key;
1254 int leave_spinning = path->leave_spinning;
1255 struct btrfs_inode_ref *iref;
1256
1257 if (bytes_left >= 0)
1258 dest[bytes_left] = '\0';
1259
1260 path->leave_spinning = 1;
1261 while (1) {
1262 bytes_left -= name_len;
1263 if (bytes_left >= 0)
1264 read_extent_buffer(eb, dest + bytes_left,
1265 name_off, name_len);
1266 if (eb != eb_in) {
1267 btrfs_tree_read_unlock_blocking(eb);
1268 free_extent_buffer(eb);
1269 }
1270 ret = inode_ref_info(parent, 0, fs_root, path, &found_key);
1271 if (ret > 0)
1272 ret = -ENOENT;
1273 if (ret)
1274 break;
1275
1276 next_inum = found_key.offset;
1277
1278 /* regular exit ahead */
1279 if (parent == next_inum)
1280 break;
1281
1282 slot = path->slots[0];
1283 eb = path->nodes[0];
1284 /* make sure we can use eb after releasing the path */
1285 if (eb != eb_in) {
1286 atomic_inc(&eb->refs);
1287 btrfs_tree_read_lock(eb);
1288 btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1289 }
1290 btrfs_release_path(path);
1291 iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
1292
1293 name_len = btrfs_inode_ref_name_len(eb, iref);
1294 name_off = (unsigned long)(iref + 1);
1295
1296 parent = next_inum;
1297 --bytes_left;
1298 if (bytes_left >= 0)
1299 dest[bytes_left] = '/';
1300 }
1301
1302 btrfs_release_path(path);
1303 path->leave_spinning = leave_spinning;
1304
1305 if (ret)
1306 return ERR_PTR(ret);
1307
1308 return dest + bytes_left;
1309 }
1310
1311 /*
1312 * this makes the path point to (logical EXTENT_ITEM *)
1313 * returns BTRFS_EXTENT_FLAG_DATA for data, BTRFS_EXTENT_FLAG_TREE_BLOCK for
1314 * tree blocks and <0 on error.
1315 */
1316 int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical,
1317 struct btrfs_path *path, struct btrfs_key *found_key,
1318 u64 *flags_ret)
1319 {
1320 int ret;
1321 u64 flags;
1322 u64 size = 0;
1323 u32 item_size;
1324 struct extent_buffer *eb;
1325 struct btrfs_extent_item *ei;
1326 struct btrfs_key key;
1327
1328 if (btrfs_fs_incompat(fs_info, SKINNY_METADATA))
1329 key.type = BTRFS_METADATA_ITEM_KEY;
1330 else
1331 key.type = BTRFS_EXTENT_ITEM_KEY;
1332 key.objectid = logical;
1333 key.offset = (u64)-1;
1334
1335 ret = btrfs_search_slot(NULL, fs_info->extent_root, &key, path, 0, 0);
1336 if (ret < 0)
1337 return ret;
1338 ret = btrfs_previous_item(fs_info->extent_root, path,
1339 0, BTRFS_EXTENT_ITEM_KEY);
1340 if (ret < 0)
1341 return ret;
1342
1343 btrfs_item_key_to_cpu(path->nodes[0], found_key, path->slots[0]);
1344 if (found_key->type == BTRFS_METADATA_ITEM_KEY)
1345 size = fs_info->extent_root->leafsize;
1346 else if (found_key->type == BTRFS_EXTENT_ITEM_KEY)
1347 size = found_key->offset;
1348
1349 if ((found_key->type != BTRFS_EXTENT_ITEM_KEY &&
1350 found_key->type != BTRFS_METADATA_ITEM_KEY) ||
1351 found_key->objectid > logical ||
1352 found_key->objectid + size <= logical) {
1353 pr_debug("logical %llu is not within any extent\n", logical);
1354 return -ENOENT;
1355 }
1356
1357 eb = path->nodes[0];
1358 item_size = btrfs_item_size_nr(eb, path->slots[0]);
1359 BUG_ON(item_size < sizeof(*ei));
1360
1361 ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item);
1362 flags = btrfs_extent_flags(eb, ei);
1363
1364 pr_debug("logical %llu is at position %llu within the extent (%llu "
1365 "EXTENT_ITEM %llu) flags %#llx size %u\n",
1366 logical, logical - found_key->objectid, found_key->objectid,
1367 found_key->offset, flags, item_size);
1368
1369 WARN_ON(!flags_ret);
1370 if (flags_ret) {
1371 if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)
1372 *flags_ret = BTRFS_EXTENT_FLAG_TREE_BLOCK;
1373 else if (flags & BTRFS_EXTENT_FLAG_DATA)
1374 *flags_ret = BTRFS_EXTENT_FLAG_DATA;
1375 else
1376 BUG_ON(1);
1377 return 0;
1378 }
1379
1380 return -EIO;
1381 }
1382
1383 /*
1384 * helper function to iterate extent inline refs. ptr must point to a 0 value
1385 * for the first call and may be modified. it is used to track state.
1386 * if more refs exist, 0 is returned and the next call to
1387 * __get_extent_inline_ref must pass the modified ptr parameter to get the
1388 * next ref. after the last ref was processed, 1 is returned.
1389 * returns <0 on error
1390 */
1391 static int __get_extent_inline_ref(unsigned long *ptr, struct extent_buffer *eb,
1392 struct btrfs_extent_item *ei, u32 item_size,
1393 struct btrfs_extent_inline_ref **out_eiref,
1394 int *out_type)
1395 {
1396 unsigned long end;
1397 u64 flags;
1398 struct btrfs_tree_block_info *info;
1399
1400 if (!*ptr) {
1401 /* first call */
1402 flags = btrfs_extent_flags(eb, ei);
1403 if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
1404 info = (struct btrfs_tree_block_info *)(ei + 1);
1405 *out_eiref =
1406 (struct btrfs_extent_inline_ref *)(info + 1);
1407 } else {
1408 *out_eiref = (struct btrfs_extent_inline_ref *)(ei + 1);
1409 }
1410 *ptr = (unsigned long)*out_eiref;
1411 if ((void *)*ptr >= (void *)ei + item_size)
1412 return -ENOENT;
1413 }
1414
1415 end = (unsigned long)ei + item_size;
1416 *out_eiref = (struct btrfs_extent_inline_ref *)*ptr;
1417 *out_type = btrfs_extent_inline_ref_type(eb, *out_eiref);
1418
1419 *ptr += btrfs_extent_inline_ref_size(*out_type);
1420 WARN_ON(*ptr > end);
1421 if (*ptr == end)
1422 return 1; /* last */
1423
1424 return 0;
1425 }
1426
1427 /*
1428 * reads the tree block backref for an extent. tree level and root are returned
1429 * through out_level and out_root. ptr must point to a 0 value for the first
1430 * call and may be modified (see __get_extent_inline_ref comment).
1431 * returns 0 if data was provided, 1 if there was no more data to provide or
1432 * <0 on error.
1433 */
1434 int tree_backref_for_extent(unsigned long *ptr, struct extent_buffer *eb,
1435 struct btrfs_extent_item *ei, u32 item_size,
1436 u64 *out_root, u8 *out_level)
1437 {
1438 int ret;
1439 int type;
1440 struct btrfs_tree_block_info *info;
1441 struct btrfs_extent_inline_ref *eiref;
1442
1443 if (*ptr == (unsigned long)-1)
1444 return 1;
1445
1446 while (1) {
1447 ret = __get_extent_inline_ref(ptr, eb, ei, item_size,
1448 &eiref, &type);
1449 if (ret < 0)
1450 return ret;
1451
1452 if (type == BTRFS_TREE_BLOCK_REF_KEY ||
1453 type == BTRFS_SHARED_BLOCK_REF_KEY)
1454 break;
1455
1456 if (ret == 1)
1457 return 1;
1458 }
1459
1460 /* we can treat both ref types equally here */
1461 info = (struct btrfs_tree_block_info *)(ei + 1);
1462 *out_root = btrfs_extent_inline_ref_offset(eb, eiref);
1463 *out_level = btrfs_tree_block_level(eb, info);
1464
1465 if (ret == 1)
1466 *ptr = (unsigned long)-1;
1467
1468 return 0;
1469 }
1470
1471 static int iterate_leaf_refs(struct extent_inode_elem *inode_list,
1472 u64 root, u64 extent_item_objectid,
1473 iterate_extent_inodes_t *iterate, void *ctx)
1474 {
1475 struct extent_inode_elem *eie;
1476 int ret = 0;
1477
1478 for (eie = inode_list; eie; eie = eie->next) {
1479 pr_debug("ref for %llu resolved, key (%llu EXTEND_DATA %llu), "
1480 "root %llu\n", extent_item_objectid,
1481 eie->inum, eie->offset, root);
1482 ret = iterate(eie->inum, eie->offset, root, ctx);
1483 if (ret) {
1484 pr_debug("stopping iteration for %llu due to ret=%d\n",
1485 extent_item_objectid, ret);
1486 break;
1487 }
1488 }
1489
1490 return ret;
1491 }
1492
1493 /*
1494 * calls iterate() for every inode that references the extent identified by
1495 * the given parameters.
1496 * when the iterator function returns a non-zero value, iteration stops.
1497 */
1498 int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
1499 u64 extent_item_objectid, u64 extent_item_pos,
1500 int search_commit_root,
1501 iterate_extent_inodes_t *iterate, void *ctx)
1502 {
1503 int ret;
1504 struct btrfs_trans_handle *trans = NULL;
1505 struct ulist *refs = NULL;
1506 struct ulist *roots = NULL;
1507 struct ulist_node *ref_node = NULL;
1508 struct ulist_node *root_node = NULL;
1509 struct seq_list tree_mod_seq_elem = {};
1510 struct ulist_iterator ref_uiter;
1511 struct ulist_iterator root_uiter;
1512
1513 pr_debug("resolving all inodes for extent %llu\n",
1514 extent_item_objectid);
1515
1516 if (!search_commit_root) {
1517 trans = btrfs_join_transaction(fs_info->extent_root);
1518 if (IS_ERR(trans))
1519 return PTR_ERR(trans);
1520 btrfs_get_tree_mod_seq(fs_info, &tree_mod_seq_elem);
1521 }
1522
1523 ret = btrfs_find_all_leafs(trans, fs_info, extent_item_objectid,
1524 tree_mod_seq_elem.seq, &refs,
1525 &extent_item_pos);
1526 if (ret)
1527 goto out;
1528
1529 ULIST_ITER_INIT(&ref_uiter);
1530 while (!ret && (ref_node = ulist_next(refs, &ref_uiter))) {
1531 ret = btrfs_find_all_roots(trans, fs_info, ref_node->val,
1532 tree_mod_seq_elem.seq, &roots);
1533 if (ret)
1534 break;
1535 ULIST_ITER_INIT(&root_uiter);
1536 while (!ret && (root_node = ulist_next(roots, &root_uiter))) {
1537 pr_debug("root %llu references leaf %llu, data list "
1538 "%#llx\n", root_node->val, ref_node->val,
1539 ref_node->aux);
1540 ret = iterate_leaf_refs((struct extent_inode_elem *)
1541 (uintptr_t)ref_node->aux,
1542 root_node->val,
1543 extent_item_objectid,
1544 iterate, ctx);
1545 }
1546 ulist_free(roots);
1547 }
1548
1549 free_leaf_list(refs);
1550 out:
1551 if (!search_commit_root) {
1552 btrfs_put_tree_mod_seq(fs_info, &tree_mod_seq_elem);
1553 btrfs_end_transaction(trans, fs_info->extent_root);
1554 }
1555
1556 return ret;
1557 }
1558
1559 int iterate_inodes_from_logical(u64 logical, struct btrfs_fs_info *fs_info,
1560 struct btrfs_path *path,
1561 iterate_extent_inodes_t *iterate, void *ctx)
1562 {
1563 int ret;
1564 u64 extent_item_pos;
1565 u64 flags = 0;
1566 struct btrfs_key found_key;
1567 int search_commit_root = path->search_commit_root;
1568
1569 ret = extent_from_logical(fs_info, logical, path, &found_key, &flags);
1570 btrfs_release_path(path);
1571 if (ret < 0)
1572 return ret;
1573 if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)
1574 return -EINVAL;
1575
1576 extent_item_pos = logical - found_key.objectid;
1577 ret = iterate_extent_inodes(fs_info, found_key.objectid,
1578 extent_item_pos, search_commit_root,
1579 iterate, ctx);
1580
1581 return ret;
1582 }
1583
1584 typedef int (iterate_irefs_t)(u64 parent, u32 name_len, unsigned long name_off,
1585 struct extent_buffer *eb, void *ctx);
1586
1587 static int iterate_inode_refs(u64 inum, struct btrfs_root *fs_root,
1588 struct btrfs_path *path,
1589 iterate_irefs_t *iterate, void *ctx)
1590 {
1591 int ret = 0;
1592 int slot;
1593 u32 cur;
1594 u32 len;
1595 u32 name_len;
1596 u64 parent = 0;
1597 int found = 0;
1598 struct extent_buffer *eb;
1599 struct btrfs_item *item;
1600 struct btrfs_inode_ref *iref;
1601 struct btrfs_key found_key;
1602
1603 while (!ret) {
1604 path->leave_spinning = 1;
1605 ret = inode_ref_info(inum, parent ? parent+1 : 0, fs_root, path,
1606 &found_key);
1607 if (ret < 0)
1608 break;
1609 if (ret) {
1610 ret = found ? 0 : -ENOENT;
1611 break;
1612 }
1613 ++found;
1614
1615 parent = found_key.offset;
1616 slot = path->slots[0];
1617 eb = path->nodes[0];
1618 /* make sure we can use eb after releasing the path */
1619 atomic_inc(&eb->refs);
1620 btrfs_tree_read_lock(eb);
1621 btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1622 btrfs_release_path(path);
1623
1624 item = btrfs_item_nr(slot);
1625 iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
1626
1627 for (cur = 0; cur < btrfs_item_size(eb, item); cur += len) {
1628 name_len = btrfs_inode_ref_name_len(eb, iref);
1629 /* path must be released before calling iterate()! */
1630 pr_debug("following ref at offset %u for inode %llu in "
1631 "tree %llu\n", cur, found_key.objectid,
1632 fs_root->objectid);
1633 ret = iterate(parent, name_len,
1634 (unsigned long)(iref + 1), eb, ctx);
1635 if (ret)
1636 break;
1637 len = sizeof(*iref) + name_len;
1638 iref = (struct btrfs_inode_ref *)((char *)iref + len);
1639 }
1640 btrfs_tree_read_unlock_blocking(eb);
1641 free_extent_buffer(eb);
1642 }
1643
1644 btrfs_release_path(path);
1645
1646 return ret;
1647 }
1648
1649 static int iterate_inode_extrefs(u64 inum, struct btrfs_root *fs_root,
1650 struct btrfs_path *path,
1651 iterate_irefs_t *iterate, void *ctx)
1652 {
1653 int ret;
1654 int slot;
1655 u64 offset = 0;
1656 u64 parent;
1657 int found = 0;
1658 struct extent_buffer *eb;
1659 struct btrfs_inode_extref *extref;
1660 struct extent_buffer *leaf;
1661 u32 item_size;
1662 u32 cur_offset;
1663 unsigned long ptr;
1664
1665 while (1) {
1666 ret = btrfs_find_one_extref(fs_root, inum, offset, path, &extref,
1667 &offset);
1668 if (ret < 0)
1669 break;
1670 if (ret) {
1671 ret = found ? 0 : -ENOENT;
1672 break;
1673 }
1674 ++found;
1675
1676 slot = path->slots[0];
1677 eb = path->nodes[0];
1678 /* make sure we can use eb after releasing the path */
1679 atomic_inc(&eb->refs);
1680
1681 btrfs_tree_read_lock(eb);
1682 btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1683 btrfs_release_path(path);
1684
1685 leaf = path->nodes[0];
1686 item_size = btrfs_item_size_nr(leaf, path->slots[0]);
1687 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
1688 cur_offset = 0;
1689
1690 while (cur_offset < item_size) {
1691 u32 name_len;
1692
1693 extref = (struct btrfs_inode_extref *)(ptr + cur_offset);
1694 parent = btrfs_inode_extref_parent(eb, extref);
1695 name_len = btrfs_inode_extref_name_len(eb, extref);
1696 ret = iterate(parent, name_len,
1697 (unsigned long)&extref->name, eb, ctx);
1698 if (ret)
1699 break;
1700
1701 cur_offset += btrfs_inode_extref_name_len(leaf, extref);
1702 cur_offset += sizeof(*extref);
1703 }
1704 btrfs_tree_read_unlock_blocking(eb);
1705 free_extent_buffer(eb);
1706
1707 offset++;
1708 }
1709
1710 btrfs_release_path(path);
1711
1712 return ret;
1713 }
1714
1715 static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
1716 struct btrfs_path *path, iterate_irefs_t *iterate,
1717 void *ctx)
1718 {
1719 int ret;
1720 int found_refs = 0;
1721
1722 ret = iterate_inode_refs(inum, fs_root, path, iterate, ctx);
1723 if (!ret)
1724 ++found_refs;
1725 else if (ret != -ENOENT)
1726 return ret;
1727
1728 ret = iterate_inode_extrefs(inum, fs_root, path, iterate, ctx);
1729 if (ret == -ENOENT && found_refs)
1730 return 0;
1731
1732 return ret;
1733 }
1734
1735 /*
1736 * returns 0 if the path could be dumped (probably truncated)
1737 * returns <0 in case of an error
1738 */
1739 static int inode_to_path(u64 inum, u32 name_len, unsigned long name_off,
1740 struct extent_buffer *eb, void *ctx)
1741 {
1742 struct inode_fs_paths *ipath = ctx;
1743 char *fspath;
1744 char *fspath_min;
1745 int i = ipath->fspath->elem_cnt;
1746 const int s_ptr = sizeof(char *);
1747 u32 bytes_left;
1748
1749 bytes_left = ipath->fspath->bytes_left > s_ptr ?
1750 ipath->fspath->bytes_left - s_ptr : 0;
1751
1752 fspath_min = (char *)ipath->fspath->val + (i + 1) * s_ptr;
1753 fspath = btrfs_ref_to_path(ipath->fs_root, ipath->btrfs_path, name_len,
1754 name_off, eb, inum, fspath_min, bytes_left);
1755 if (IS_ERR(fspath))
1756 return PTR_ERR(fspath);
1757
1758 if (fspath > fspath_min) {
1759 ipath->fspath->val[i] = (u64)(unsigned long)fspath;
1760 ++ipath->fspath->elem_cnt;
1761 ipath->fspath->bytes_left = fspath - fspath_min;
1762 } else {
1763 ++ipath->fspath->elem_missed;
1764 ipath->fspath->bytes_missing += fspath_min - fspath;
1765 ipath->fspath->bytes_left = 0;
1766 }
1767
1768 return 0;
1769 }
1770
1771 /*
1772 * this dumps all file system paths to the inode into the ipath struct, provided
1773 * is has been created large enough. each path is zero-terminated and accessed
1774 * from ipath->fspath->val[i].
1775 * when it returns, there are ipath->fspath->elem_cnt number of paths available
1776 * in ipath->fspath->val[]. when the allocated space wasn't sufficient, the
1777 * number of missed paths in recored in ipath->fspath->elem_missed, otherwise,
1778 * it's zero. ipath->fspath->bytes_missing holds the number of bytes that would
1779 * have been needed to return all paths.
1780 */
1781 int paths_from_inode(u64 inum, struct inode_fs_paths *ipath)
1782 {
1783 return iterate_irefs(inum, ipath->fs_root, ipath->btrfs_path,
1784 inode_to_path, ipath);
1785 }
1786
1787 struct btrfs_data_container *init_data_container(u32 total_bytes)
1788 {
1789 struct btrfs_data_container *data;
1790 size_t alloc_bytes;
1791
1792 alloc_bytes = max_t(size_t, total_bytes, sizeof(*data));
1793 data = vmalloc(alloc_bytes);
1794 if (!data)
1795 return ERR_PTR(-ENOMEM);
1796
1797 if (total_bytes >= sizeof(*data)) {
1798 data->bytes_left = total_bytes - sizeof(*data);
1799 data->bytes_missing = 0;
1800 } else {
1801 data->bytes_missing = sizeof(*data) - total_bytes;
1802 data->bytes_left = 0;
1803 }
1804
1805 data->elem_cnt = 0;
1806 data->elem_missed = 0;
1807
1808 return data;
1809 }
1810
1811 /*
1812 * allocates space to return multiple file system paths for an inode.
1813 * total_bytes to allocate are passed, note that space usable for actual path
1814 * information will be total_bytes - sizeof(struct inode_fs_paths).
1815 * the returned pointer must be freed with free_ipath() in the end.
1816 */
1817 struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root,
1818 struct btrfs_path *path)
1819 {
1820 struct inode_fs_paths *ifp;
1821 struct btrfs_data_container *fspath;
1822
1823 fspath = init_data_container(total_bytes);
1824 if (IS_ERR(fspath))
1825 return (void *)fspath;
1826
1827 ifp = kmalloc(sizeof(*ifp), GFP_NOFS);
1828 if (!ifp) {
1829 kfree(fspath);
1830 return ERR_PTR(-ENOMEM);
1831 }
1832
1833 ifp->btrfs_path = path;
1834 ifp->fspath = fspath;
1835 ifp->fs_root = fs_root;
1836
1837 return ifp;
1838 }
1839
1840 void free_ipath(struct inode_fs_paths *ipath)
1841 {
1842 if (!ipath)
1843 return;
1844 vfree(ipath->fspath);
1845 kfree(ipath);
1846 }
Linus' kernel tree