?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
m68k: Skip futex_atomic_cmpxchg_inatomic() test
[linux/fpc-iii.git] / fs / btrfs / backref.c
blobaded3ef3d3d4abfa6d61fa078a51e64e4b80be65
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 void free_inode_elem_list(struct extent_inode_elem *eie)
70 {
71 struct extent_inode_elem *eie_next;
72
73 for (; eie; eie = eie_next) {
74 eie_next = eie->next;
75 kfree(eie);
76 }
77 }
78
79 static int find_extent_in_eb(struct extent_buffer *eb, u64 wanted_disk_byte,
80 u64 extent_item_pos,
81 struct extent_inode_elem **eie)
82 {
83 u64 disk_byte;
84 struct btrfs_key key;
85 struct btrfs_file_extent_item *fi;
86 int slot;
87 int nritems;
88 int extent_type;
89 int ret;
90
91 /*
92 * from the shared data ref, we only have the leaf but we need
93 * the key. thus, we must look into all items and see that we
94 * find one (some) with a reference to our extent item.
95 */
96 nritems = btrfs_header_nritems(eb);
97 for (slot = 0; slot < nritems; ++slot) {
98 btrfs_item_key_to_cpu(eb, &key, slot);
99 if (key.type != BTRFS_EXTENT_DATA_KEY)
100 continue;
101 fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
102 extent_type = btrfs_file_extent_type(eb, fi);
103 if (extent_type == BTRFS_FILE_EXTENT_INLINE)
104 continue;
105 /* don't skip BTRFS_FILE_EXTENT_PREALLOC, we can handle that */
106 disk_byte = btrfs_file_extent_disk_bytenr(eb, fi);
107 if (disk_byte != wanted_disk_byte)
108 continue;
109
110 ret = check_extent_in_eb(&key, eb, fi, extent_item_pos, eie);
111 if (ret < 0)
112 return ret;
113 }
114
115 return 0;
116 }
117
118 /*
119 * this structure records all encountered refs on the way up to the root
120 */
121 struct __prelim_ref {
122 struct list_head list;
123 u64 root_id;
124 struct btrfs_key key_for_search;
125 int level;
126 int count;
127 struct extent_inode_elem *inode_list;
128 u64 parent;
129 u64 wanted_disk_byte;
130 };
131
132 static struct kmem_cache *btrfs_prelim_ref_cache;
133
134 int __init btrfs_prelim_ref_init(void)
135 {
136 btrfs_prelim_ref_cache = kmem_cache_create("btrfs_prelim_ref",
137 sizeof(struct __prelim_ref),
138 0,
139 SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD,
140 NULL);
141 if (!btrfs_prelim_ref_cache)
142 return -ENOMEM;
143 return 0;
144 }
145
146 void btrfs_prelim_ref_exit(void)
147 {
148 if (btrfs_prelim_ref_cache)
149 kmem_cache_destroy(btrfs_prelim_ref_cache);
150 }
151
152 /*
153 * the rules for all callers of this function are:
154 * - obtaining the parent is the goal
155 * - if you add a key, you must know that it is a correct key
156 * - if you cannot add the parent or a correct key, then we will look into the
157 * block later to set a correct key
158 *
159 * delayed refs
160 * ============
161 * backref type | shared | indirect | shared | indirect
162 * information | tree | tree | data | data
163 * --------------------+--------+----------+--------+----------
164 * parent logical | y | - | - | -
165 * key to resolve | - | y | y | y
166 * tree block logical | - | - | - | -
167 * root for resolving | y | y | y | y
168 *
169 * - column 1: we've the parent -> done
170 * - column 2, 3, 4: we use the key to find the parent
171 *
172 * on disk refs (inline or keyed)
173 * ==============================
174 * backref type | shared | indirect | shared | indirect
175 * information | tree | tree | data | data
176 * --------------------+--------+----------+--------+----------
177 * parent logical | y | - | y | -
178 * key to resolve | - | - | - | y
179 * tree block logical | y | y | y | y
180 * root for resolving | - | y | y | y
181 *
182 * - column 1, 3: we've the parent -> done
183 * - column 2: we take the first key from the block to find the parent
184 * (see __add_missing_keys)
185 * - column 4: we use the key to find the parent
186 *
187 * additional information that's available but not required to find the parent
188 * block might help in merging entries to gain some speed.
189 */
190
191 static int __add_prelim_ref(struct list_head *head, u64 root_id,
192 struct btrfs_key *key, int level,
193 u64 parent, u64 wanted_disk_byte, int count,
194 gfp_t gfp_mask)
195 {
196 struct __prelim_ref *ref;
197
198 if (root_id == BTRFS_DATA_RELOC_TREE_OBJECTID)
199 return 0;
200
201 ref = kmem_cache_alloc(btrfs_prelim_ref_cache, gfp_mask);
202 if (!ref)
203 return -ENOMEM;
204
205 ref->root_id = root_id;
206 if (key)
207 ref->key_for_search = *key;
208 else
209 memset(&ref->key_for_search, 0, sizeof(ref->key_for_search));
210
211 ref->inode_list = NULL;
212 ref->level = level;
213 ref->count = count;
214 ref->parent = parent;
215 ref->wanted_disk_byte = wanted_disk_byte;
216 list_add_tail(&ref->list, head);
217
218 return 0;
219 }
220
221 static int add_all_parents(struct btrfs_root *root, struct btrfs_path *path,
222 struct ulist *parents, struct __prelim_ref *ref,
223 int level, u64 time_seq, const u64 *extent_item_pos)
224 {
225 int ret = 0;
226 int slot;
227 struct extent_buffer *eb;
228 struct btrfs_key key;
229 struct btrfs_key *key_for_search = &ref->key_for_search;
230 struct btrfs_file_extent_item *fi;
231 struct extent_inode_elem *eie = NULL, *old = NULL;
232 u64 disk_byte;
233 u64 wanted_disk_byte = ref->wanted_disk_byte;
234 u64 count = 0;
235
236 if (level != 0) {
237 eb = path->nodes[level];
238 ret = ulist_add(parents, eb->start, 0, GFP_NOFS);
239 if (ret < 0)
240 return ret;
241 return 0;
242 }
243
244 /*
245 * We normally enter this function with the path already pointing to
246 * the first item to check. But sometimes, we may enter it with
247 * slot==nritems. In that case, go to the next leaf before we continue.
248 */
249 if (path->slots[0] >= btrfs_header_nritems(path->nodes[0]))
250 ret = btrfs_next_old_leaf(root, path, time_seq);
251
252 while (!ret && count < ref->count) {
253 eb = path->nodes[0];
254 slot = path->slots[0];
255
256 btrfs_item_key_to_cpu(eb, &key, slot);
257
258 if (key.objectid != key_for_search->objectid ||
259 key.type != BTRFS_EXTENT_DATA_KEY)
260 break;
261
262 fi = btrfs_item_ptr(eb, slot, struct btrfs_file_extent_item);
263 disk_byte = btrfs_file_extent_disk_bytenr(eb, fi);
264
265 if (disk_byte == wanted_disk_byte) {
266 eie = NULL;
267 old = NULL;
268 count++;
269 if (extent_item_pos) {
270 ret = check_extent_in_eb(&key, eb, fi,
271 *extent_item_pos,
272 &eie);
273 if (ret < 0)
274 break;
275 }
276 if (ret > 0)
277 goto next;
278 ret = ulist_add_merge(parents, eb->start,
279 (uintptr_t)eie,
280 (u64 *)&old, GFP_NOFS);
281 if (ret < 0)
282 break;
283 if (!ret && extent_item_pos) {
284 while (old->next)
285 old = old->next;
286 old->next = eie;
287 }
288 eie = NULL;
289 }
290 next:
291 ret = btrfs_next_old_item(root, path, time_seq);
292 }
293
294 if (ret > 0)
295 ret = 0;
296 else if (ret < 0)
297 free_inode_elem_list(eie);
298 return ret;
299 }
300
301 /*
302 * resolve an indirect backref in the form (root_id, key, level)
303 * to a logical address
304 */
305 static int __resolve_indirect_ref(struct btrfs_fs_info *fs_info,
306 struct btrfs_path *path, u64 time_seq,
307 struct __prelim_ref *ref,
308 struct ulist *parents,
309 const u64 *extent_item_pos)
310 {
311 struct btrfs_root *root;
312 struct btrfs_key root_key;
313 struct extent_buffer *eb;
314 int ret = 0;
315 int root_level;
316 int level = ref->level;
317 int index;
318
319 root_key.objectid = ref->root_id;
320 root_key.type = BTRFS_ROOT_ITEM_KEY;
321 root_key.offset = (u64)-1;
322
323 index = srcu_read_lock(&fs_info->subvol_srcu);
324
325 root = btrfs_read_fs_root_no_name(fs_info, &root_key);
326 if (IS_ERR(root)) {
327 srcu_read_unlock(&fs_info->subvol_srcu, index);
328 ret = PTR_ERR(root);
329 goto out;
330 }
331
332 root_level = btrfs_old_root_level(root, time_seq);
333
334 if (root_level + 1 == level) {
335 srcu_read_unlock(&fs_info->subvol_srcu, index);
336 goto out;
337 }
338
339 path->lowest_level = level;
340 ret = btrfs_search_old_slot(root, &ref->key_for_search, path, time_seq);
341
342 /* root node has been locked, we can release @subvol_srcu safely here */
343 srcu_read_unlock(&fs_info->subvol_srcu, index);
344
345 pr_debug("search slot in root %llu (level %d, ref count %d) returned "
346 "%d for key (%llu %u %llu)\n",
347 ref->root_id, level, ref->count, ret,
348 ref->key_for_search.objectid, ref->key_for_search.type,
349 ref->key_for_search.offset);
350 if (ret < 0)
351 goto out;
352
353 eb = path->nodes[level];
354 while (!eb) {
355 if (WARN_ON(!level)) {
356 ret = 1;
357 goto out;
358 }
359 level--;
360 eb = path->nodes[level];
361 }
362
363 ret = add_all_parents(root, path, parents, ref, level, time_seq,
364 extent_item_pos);
365 out:
366 path->lowest_level = 0;
367 btrfs_release_path(path);
368 return ret;
369 }
370
371 /*
372 * resolve all indirect backrefs from the list
373 */
374 static int __resolve_indirect_refs(struct btrfs_fs_info *fs_info,
375 struct btrfs_path *path, u64 time_seq,
376 struct list_head *head,
377 const u64 *extent_item_pos)
378 {
379 int err;
380 int ret = 0;
381 struct __prelim_ref *ref;
382 struct __prelim_ref *ref_safe;
383 struct __prelim_ref *new_ref;
384 struct ulist *parents;
385 struct ulist_node *node;
386 struct ulist_iterator uiter;
387
388 parents = ulist_alloc(GFP_NOFS);
389 if (!parents)
390 return -ENOMEM;
391
392 /*
393 * _safe allows us to insert directly after the current item without
394 * iterating over the newly inserted items.
395 * we're also allowed to re-assign ref during iteration.
396 */
397 list_for_each_entry_safe(ref, ref_safe, head, list) {
398 if (ref->parent) /* already direct */
399 continue;
400 if (ref->count == 0)
401 continue;
402 err = __resolve_indirect_ref(fs_info, path, time_seq, ref,
403 parents, extent_item_pos);
404 /*
405 * we can only tolerate ENOENT,otherwise,we should catch error
406 * and return directly.
407 */
408 if (err == -ENOENT) {
409 continue;
410 } else if (err) {
411 ret = err;
412 goto out;
413 }
414
415 /* we put the first parent into the ref at hand */
416 ULIST_ITER_INIT(&uiter);
417 node = ulist_next(parents, &uiter);
418 ref->parent = node ? node->val : 0;
419 ref->inode_list = node ?
420 (struct extent_inode_elem *)(uintptr_t)node->aux : NULL;
421
422 /* additional parents require new refs being added here */
423 while ((node = ulist_next(parents, &uiter))) {
424 new_ref = kmem_cache_alloc(btrfs_prelim_ref_cache,
425 GFP_NOFS);
426 if (!new_ref) {
427 ret = -ENOMEM;
428 goto out;
429 }
430 memcpy(new_ref, ref, sizeof(*ref));
431 new_ref->parent = node->val;
432 new_ref->inode_list = (struct extent_inode_elem *)
433 (uintptr_t)node->aux;
434 list_add(&new_ref->list, &ref->list);
435 }
436 ulist_reinit(parents);
437 }
438 out:
439 ulist_free(parents);
440 return ret;
441 }
442
443 static inline int ref_for_same_block(struct __prelim_ref *ref1,
444 struct __prelim_ref *ref2)
445 {
446 if (ref1->level != ref2->level)
447 return 0;
448 if (ref1->root_id != ref2->root_id)
449 return 0;
450 if (ref1->key_for_search.type != ref2->key_for_search.type)
451 return 0;
452 if (ref1->key_for_search.objectid != ref2->key_for_search.objectid)
453 return 0;
454 if (ref1->key_for_search.offset != ref2->key_for_search.offset)
455 return 0;
456 if (ref1->parent != ref2->parent)
457 return 0;
458
459 return 1;
460 }
461
462 /*
463 * read tree blocks and add keys where required.
464 */
465 static int __add_missing_keys(struct btrfs_fs_info *fs_info,
466 struct list_head *head)
467 {
468 struct list_head *pos;
469 struct extent_buffer *eb;
470
471 list_for_each(pos, head) {
472 struct __prelim_ref *ref;
473 ref = list_entry(pos, struct __prelim_ref, list);
474
475 if (ref->parent)
476 continue;
477 if (ref->key_for_search.type)
478 continue;
479 BUG_ON(!ref->wanted_disk_byte);
480 eb = read_tree_block(fs_info->tree_root, ref->wanted_disk_byte,
481 fs_info->tree_root->leafsize, 0);
482 if (!eb || !extent_buffer_uptodate(eb)) {
483 free_extent_buffer(eb);
484 return -EIO;
485 }
486 btrfs_tree_read_lock(eb);
487 if (btrfs_header_level(eb) == 0)
488 btrfs_item_key_to_cpu(eb, &ref->key_for_search, 0);
489 else
490 btrfs_node_key_to_cpu(eb, &ref->key_for_search, 0);
491 btrfs_tree_read_unlock(eb);
492 free_extent_buffer(eb);
493 }
494 return 0;
495 }
496
497 /*
498 * merge two lists of backrefs and adjust counts accordingly
499 *
500 * mode = 1: merge identical keys, if key is set
501 * FIXME: if we add more keys in __add_prelim_ref, we can merge more here.
502 * additionally, we could even add a key range for the blocks we
503 * looked into to merge even more (-> replace unresolved refs by those
504 * having a parent).
505 * mode = 2: merge identical parents
506 */
507 static void __merge_refs(struct list_head *head, int mode)
508 {
509 struct list_head *pos1;
510
511 list_for_each(pos1, head) {
512 struct list_head *n2;
513 struct list_head *pos2;
514 struct __prelim_ref *ref1;
515
516 ref1 = list_entry(pos1, struct __prelim_ref, list);
517
518 for (pos2 = pos1->next, n2 = pos2->next; pos2 != head;
519 pos2 = n2, n2 = pos2->next) {
520 struct __prelim_ref *ref2;
521 struct __prelim_ref *xchg;
522 struct extent_inode_elem *eie;
523
524 ref2 = list_entry(pos2, struct __prelim_ref, list);
525
526 if (mode == 1) {
527 if (!ref_for_same_block(ref1, ref2))
528 continue;
529 if (!ref1->parent && ref2->parent) {
530 xchg = ref1;
531 ref1 = ref2;
532 ref2 = xchg;
533 }
534 } else {
535 if (ref1->parent != ref2->parent)
536 continue;
537 }
538
539 eie = ref1->inode_list;
540 while (eie && eie->next)
541 eie = eie->next;
542 if (eie)
543 eie->next = ref2->inode_list;
544 else
545 ref1->inode_list = ref2->inode_list;
546 ref1->count += ref2->count;
547
548 list_del(&ref2->list);
549 kmem_cache_free(btrfs_prelim_ref_cache, ref2);
550 }
551
552 }
553 }
554
555 /*
556 * add all currently queued delayed refs from this head whose seq nr is
557 * smaller or equal that seq to the list
558 */
559 static int __add_delayed_refs(struct btrfs_delayed_ref_head *head, u64 seq,
560 struct list_head *prefs)
561 {
562 struct btrfs_delayed_extent_op *extent_op = head->extent_op;
563 struct rb_node *n = &head->node.rb_node;
564 struct btrfs_key key;
565 struct btrfs_key op_key = {0};
566 int sgn;
567 int ret = 0;
568
569 if (extent_op && extent_op->update_key)
570 btrfs_disk_key_to_cpu(&op_key, &extent_op->key);
571
572 spin_lock(&head->lock);
573 n = rb_first(&head->ref_root);
574 while (n) {
575 struct btrfs_delayed_ref_node *node;
576 node = rb_entry(n, struct btrfs_delayed_ref_node,
577 rb_node);
578 n = rb_next(n);
579 if (node->seq > seq)
580 continue;
581
582 switch (node->action) {
583 case BTRFS_ADD_DELAYED_EXTENT:
584 case BTRFS_UPDATE_DELAYED_HEAD:
585 WARN_ON(1);
586 continue;
587 case BTRFS_ADD_DELAYED_REF:
588 sgn = 1;
589 break;
590 case BTRFS_DROP_DELAYED_REF:
591 sgn = -1;
592 break;
593 default:
594 BUG_ON(1);
595 }
596 switch (node->type) {
597 case BTRFS_TREE_BLOCK_REF_KEY: {
598 struct btrfs_delayed_tree_ref *ref;
599
600 ref = btrfs_delayed_node_to_tree_ref(node);
601 ret = __add_prelim_ref(prefs, ref->root, &op_key,
602 ref->level + 1, 0, node->bytenr,
603 node->ref_mod * sgn, GFP_ATOMIC);
604 break;
605 }
606 case BTRFS_SHARED_BLOCK_REF_KEY: {
607 struct btrfs_delayed_tree_ref *ref;
608
609 ref = btrfs_delayed_node_to_tree_ref(node);
610 ret = __add_prelim_ref(prefs, ref->root, NULL,
611 ref->level + 1, ref->parent,
612 node->bytenr,
613 node->ref_mod * sgn, GFP_ATOMIC);
614 break;
615 }
616 case BTRFS_EXTENT_DATA_REF_KEY: {
617 struct btrfs_delayed_data_ref *ref;
618 ref = btrfs_delayed_node_to_data_ref(node);
619
620 key.objectid = ref->objectid;
621 key.type = BTRFS_EXTENT_DATA_KEY;
622 key.offset = ref->offset;
623 ret = __add_prelim_ref(prefs, ref->root, &key, 0, 0,
624 node->bytenr,
625 node->ref_mod * sgn, GFP_ATOMIC);
626 break;
627 }
628 case BTRFS_SHARED_DATA_REF_KEY: {
629 struct btrfs_delayed_data_ref *ref;
630
631 ref = btrfs_delayed_node_to_data_ref(node);
632
633 key.objectid = ref->objectid;
634 key.type = BTRFS_EXTENT_DATA_KEY;
635 key.offset = ref->offset;
636 ret = __add_prelim_ref(prefs, ref->root, &key, 0,
637 ref->parent, node->bytenr,
638 node->ref_mod * sgn, GFP_ATOMIC);
639 break;
640 }
641 default:
642 WARN_ON(1);
643 }
644 if (ret)
645 break;
646 }
647 spin_unlock(&head->lock);
648 return ret;
649 }
650
651 /*
652 * add all inline backrefs for bytenr to the list
653 */
654 static int __add_inline_refs(struct btrfs_fs_info *fs_info,
655 struct btrfs_path *path, u64 bytenr,
656 int *info_level, struct list_head *prefs)
657 {
658 int ret = 0;
659 int slot;
660 struct extent_buffer *leaf;
661 struct btrfs_key key;
662 struct btrfs_key found_key;
663 unsigned long ptr;
664 unsigned long end;
665 struct btrfs_extent_item *ei;
666 u64 flags;
667 u64 item_size;
668
669 /*
670 * enumerate all inline refs
671 */
672 leaf = path->nodes[0];
673 slot = path->slots[0];
674
675 item_size = btrfs_item_size_nr(leaf, slot);
676 BUG_ON(item_size < sizeof(*ei));
677
678 ei = btrfs_item_ptr(leaf, slot, struct btrfs_extent_item);
679 flags = btrfs_extent_flags(leaf, ei);
680 btrfs_item_key_to_cpu(leaf, &found_key, slot);
681
682 ptr = (unsigned long)(ei + 1);
683 end = (unsigned long)ei + item_size;
684
685 if (found_key.type == BTRFS_EXTENT_ITEM_KEY &&
686 flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
687 struct btrfs_tree_block_info *info;
688
689 info = (struct btrfs_tree_block_info *)ptr;
690 *info_level = btrfs_tree_block_level(leaf, info);
691 ptr += sizeof(struct btrfs_tree_block_info);
692 BUG_ON(ptr > end);
693 } else if (found_key.type == BTRFS_METADATA_ITEM_KEY) {
694 *info_level = found_key.offset;
695 } else {
696 BUG_ON(!(flags & BTRFS_EXTENT_FLAG_DATA));
697 }
698
699 while (ptr < end) {
700 struct btrfs_extent_inline_ref *iref;
701 u64 offset;
702 int type;
703
704 iref = (struct btrfs_extent_inline_ref *)ptr;
705 type = btrfs_extent_inline_ref_type(leaf, iref);
706 offset = btrfs_extent_inline_ref_offset(leaf, iref);
707
708 switch (type) {
709 case BTRFS_SHARED_BLOCK_REF_KEY:
710 ret = __add_prelim_ref(prefs, 0, NULL,
711 *info_level + 1, offset,
712 bytenr, 1, GFP_NOFS);
713 break;
714 case BTRFS_SHARED_DATA_REF_KEY: {
715 struct btrfs_shared_data_ref *sdref;
716 int count;
717
718 sdref = (struct btrfs_shared_data_ref *)(iref + 1);
719 count = btrfs_shared_data_ref_count(leaf, sdref);
720 ret = __add_prelim_ref(prefs, 0, NULL, 0, offset,
721 bytenr, count, GFP_NOFS);
722 break;
723 }
724 case BTRFS_TREE_BLOCK_REF_KEY:
725 ret = __add_prelim_ref(prefs, offset, NULL,
726 *info_level + 1, 0,
727 bytenr, 1, GFP_NOFS);
728 break;
729 case BTRFS_EXTENT_DATA_REF_KEY: {
730 struct btrfs_extent_data_ref *dref;
731 int count;
732 u64 root;
733
734 dref = (struct btrfs_extent_data_ref *)(&iref->offset);
735 count = btrfs_extent_data_ref_count(leaf, dref);
736 key.objectid = btrfs_extent_data_ref_objectid(leaf,
737 dref);
738 key.type = BTRFS_EXTENT_DATA_KEY;
739 key.offset = btrfs_extent_data_ref_offset(leaf, dref);
740 root = btrfs_extent_data_ref_root(leaf, dref);
741 ret = __add_prelim_ref(prefs, root, &key, 0, 0,
742 bytenr, count, GFP_NOFS);
743 break;
744 }
745 default:
746 WARN_ON(1);
747 }
748 if (ret)
749 return ret;
750 ptr += btrfs_extent_inline_ref_size(type);
751 }
752
753 return 0;
754 }
755
756 /*
757 * add all non-inline backrefs for bytenr to the list
758 */
759 static int __add_keyed_refs(struct btrfs_fs_info *fs_info,
760 struct btrfs_path *path, u64 bytenr,
761 int info_level, struct list_head *prefs)
762 {
763 struct btrfs_root *extent_root = fs_info->extent_root;
764 int ret;
765 int slot;
766 struct extent_buffer *leaf;
767 struct btrfs_key key;
768
769 while (1) {
770 ret = btrfs_next_item(extent_root, path);
771 if (ret < 0)
772 break;
773 if (ret) {
774 ret = 0;
775 break;
776 }
777
778 slot = path->slots[0];
779 leaf = path->nodes[0];
780 btrfs_item_key_to_cpu(leaf, &key, slot);
781
782 if (key.objectid != bytenr)
783 break;
784 if (key.type < BTRFS_TREE_BLOCK_REF_KEY)
785 continue;
786 if (key.type > BTRFS_SHARED_DATA_REF_KEY)
787 break;
788
789 switch (key.type) {
790 case BTRFS_SHARED_BLOCK_REF_KEY:
791 ret = __add_prelim_ref(prefs, 0, NULL,
792 info_level + 1, key.offset,
793 bytenr, 1, GFP_NOFS);
794 break;
795 case BTRFS_SHARED_DATA_REF_KEY: {
796 struct btrfs_shared_data_ref *sdref;
797 int count;
798
799 sdref = btrfs_item_ptr(leaf, slot,
800 struct btrfs_shared_data_ref);
801 count = btrfs_shared_data_ref_count(leaf, sdref);
802 ret = __add_prelim_ref(prefs, 0, NULL, 0, key.offset,
803 bytenr, count, GFP_NOFS);
804 break;
805 }
806 case BTRFS_TREE_BLOCK_REF_KEY:
807 ret = __add_prelim_ref(prefs, key.offset, NULL,
808 info_level + 1, 0,
809 bytenr, 1, GFP_NOFS);
810 break;
811 case BTRFS_EXTENT_DATA_REF_KEY: {
812 struct btrfs_extent_data_ref *dref;
813 int count;
814 u64 root;
815
816 dref = btrfs_item_ptr(leaf, slot,
817 struct btrfs_extent_data_ref);
818 count = btrfs_extent_data_ref_count(leaf, dref);
819 key.objectid = btrfs_extent_data_ref_objectid(leaf,
820 dref);
821 key.type = BTRFS_EXTENT_DATA_KEY;
822 key.offset = btrfs_extent_data_ref_offset(leaf, dref);
823 root = btrfs_extent_data_ref_root(leaf, dref);
824 ret = __add_prelim_ref(prefs, root, &key, 0, 0,
825 bytenr, count, GFP_NOFS);
826 break;
827 }
828 default:
829 WARN_ON(1);
830 }
831 if (ret)
832 return ret;
833
834 }
835
836 return ret;
837 }
838
839 /*
840 * this adds all existing backrefs (inline backrefs, backrefs and delayed
841 * refs) for the given bytenr to the refs list, merges duplicates and resolves
842 * indirect refs to their parent bytenr.
843 * When roots are found, they're added to the roots list
844 *
845 * FIXME some caching might speed things up
846 */
847 static int find_parent_nodes(struct btrfs_trans_handle *trans,
848 struct btrfs_fs_info *fs_info, u64 bytenr,
849 u64 time_seq, struct ulist *refs,
850 struct ulist *roots, const u64 *extent_item_pos)
851 {
852 struct btrfs_key key;
853 struct btrfs_path *path;
854 struct btrfs_delayed_ref_root *delayed_refs = NULL;
855 struct btrfs_delayed_ref_head *head;
856 int info_level = 0;
857 int ret;
858 struct list_head prefs_delayed;
859 struct list_head prefs;
860 struct __prelim_ref *ref;
861 struct extent_inode_elem *eie = NULL;
862
863 INIT_LIST_HEAD(&prefs);
864 INIT_LIST_HEAD(&prefs_delayed);
865
866 key.objectid = bytenr;
867 key.offset = (u64)-1;
868 if (btrfs_fs_incompat(fs_info, SKINNY_METADATA))
869 key.type = BTRFS_METADATA_ITEM_KEY;
870 else
871 key.type = BTRFS_EXTENT_ITEM_KEY;
872
873 path = btrfs_alloc_path();
874 if (!path)
875 return -ENOMEM;
876 if (!trans)
877 path->search_commit_root = 1;
878
879 /*
880 * grab both a lock on the path and a lock on the delayed ref head.
881 * We need both to get a consistent picture of how the refs look
882 * at a specified point in time
883 */
884 again:
885 head = NULL;
886
887 ret = btrfs_search_slot(trans, fs_info->extent_root, &key, path, 0, 0);
888 if (ret < 0)
889 goto out;
890 BUG_ON(ret == 0);
891
892 if (trans) {
893 /*
894 * look if there are updates for this ref queued and lock the
895 * head
896 */
897 delayed_refs = &trans->transaction->delayed_refs;
898 spin_lock(&delayed_refs->lock);
899 head = btrfs_find_delayed_ref_head(trans, bytenr);
900 if (head) {
901 if (!mutex_trylock(&head->mutex)) {
902 atomic_inc(&head->node.refs);
903 spin_unlock(&delayed_refs->lock);
904
905 btrfs_release_path(path);
906
907 /*
908 * Mutex was contended, block until it's
909 * released and try again
910 */
911 mutex_lock(&head->mutex);
912 mutex_unlock(&head->mutex);
913 btrfs_put_delayed_ref(&head->node);
914 goto again;
915 }
916 spin_unlock(&delayed_refs->lock);
917 ret = __add_delayed_refs(head, time_seq,
918 &prefs_delayed);
919 mutex_unlock(&head->mutex);
920 if (ret)
921 goto out;
922 } else {
923 spin_unlock(&delayed_refs->lock);
924 }
925 }
926
927 if (path->slots[0]) {
928 struct extent_buffer *leaf;
929 int slot;
930
931 path->slots[0]--;
932 leaf = path->nodes[0];
933 slot = path->slots[0];
934 btrfs_item_key_to_cpu(leaf, &key, slot);
935 if (key.objectid == bytenr &&
936 (key.type == BTRFS_EXTENT_ITEM_KEY ||
937 key.type == BTRFS_METADATA_ITEM_KEY)) {
938 ret = __add_inline_refs(fs_info, path, bytenr,
939 &info_level, &prefs);
940 if (ret)
941 goto out;
942 ret = __add_keyed_refs(fs_info, path, bytenr,
943 info_level, &prefs);
944 if (ret)
945 goto out;
946 }
947 }
948 btrfs_release_path(path);
949
950 list_splice_init(&prefs_delayed, &prefs);
951
952 ret = __add_missing_keys(fs_info, &prefs);
953 if (ret)
954 goto out;
955
956 __merge_refs(&prefs, 1);
957
958 ret = __resolve_indirect_refs(fs_info, path, time_seq, &prefs,
959 extent_item_pos);
960 if (ret)
961 goto out;
962
963 __merge_refs(&prefs, 2);
964
965 while (!list_empty(&prefs)) {
966 ref = list_first_entry(&prefs, struct __prelim_ref, list);
967 WARN_ON(ref->count < 0);
968 if (ref->count && ref->root_id && ref->parent == 0) {
969 /* no parent == root of tree */
970 ret = ulist_add(roots, ref->root_id, 0, GFP_NOFS);
971 if (ret < 0)
972 goto out;
973 }
974 if (ref->count && ref->parent) {
975 if (extent_item_pos && !ref->inode_list) {
976 u32 bsz;
977 struct extent_buffer *eb;
978 bsz = btrfs_level_size(fs_info->extent_root,
979 info_level);
980 eb = read_tree_block(fs_info->extent_root,
981 ref->parent, bsz, 0);
982 if (!eb || !extent_buffer_uptodate(eb)) {
983 free_extent_buffer(eb);
984 ret = -EIO;
985 goto out;
986 }
987 ret = find_extent_in_eb(eb, bytenr,
988 *extent_item_pos, &eie);
989 free_extent_buffer(eb);
990 if (ret < 0)
991 goto out;
992 ref->inode_list = eie;
993 }
994 ret = ulist_add_merge(refs, ref->parent,
995 (uintptr_t)ref->inode_list,
996 (u64 *)&eie, GFP_NOFS);
997 if (ret < 0)
998 goto out;
999 if (!ret && extent_item_pos) {
1000 /*
1001 * we've recorded that parent, so we must extend
1002 * its inode list here
1003 */
1004 BUG_ON(!eie);
1005 while (eie->next)
1006 eie = eie->next;
1007 eie->next = ref->inode_list;
1008 }
1009 eie = NULL;
1010 }
1011 list_del(&ref->list);
1012 kmem_cache_free(btrfs_prelim_ref_cache, ref);
1013 }
1014
1015 out:
1016 btrfs_free_path(path);
1017 while (!list_empty(&prefs)) {
1018 ref = list_first_entry(&prefs, struct __prelim_ref, list);
1019 list_del(&ref->list);
1020 kmem_cache_free(btrfs_prelim_ref_cache, ref);
1021 }
1022 while (!list_empty(&prefs_delayed)) {
1023 ref = list_first_entry(&prefs_delayed, struct __prelim_ref,
1024 list);
1025 list_del(&ref->list);
1026 kmem_cache_free(btrfs_prelim_ref_cache, ref);
1027 }
1028 if (ret < 0)
1029 free_inode_elem_list(eie);
1030 return ret;
1031 }
1032
1033 static void free_leaf_list(struct ulist *blocks)
1034 {
1035 struct ulist_node *node = NULL;
1036 struct extent_inode_elem *eie;
1037 struct ulist_iterator uiter;
1038
1039 ULIST_ITER_INIT(&uiter);
1040 while ((node = ulist_next(blocks, &uiter))) {
1041 if (!node->aux)
1042 continue;
1043 eie = (struct extent_inode_elem *)(uintptr_t)node->aux;
1044 free_inode_elem_list(eie);
1045 node->aux = 0;
1046 }
1047
1048 ulist_free(blocks);
1049 }
1050
1051 /*
1052 * Finds all leafs with a reference to the specified combination of bytenr and
1053 * offset. key_list_head will point to a list of corresponding keys (caller must
1054 * free each list element). The leafs will be stored in the leafs ulist, which
1055 * must be freed with ulist_free.
1056 *
1057 * returns 0 on success, <0 on error
1058 */
1059 static int btrfs_find_all_leafs(struct btrfs_trans_handle *trans,
1060 struct btrfs_fs_info *fs_info, u64 bytenr,
1061 u64 time_seq, struct ulist **leafs,
1062 const u64 *extent_item_pos)
1063 {
1064 struct ulist *tmp;
1065 int ret;
1066
1067 tmp = ulist_alloc(GFP_NOFS);
1068 if (!tmp)
1069 return -ENOMEM;
1070 *leafs = ulist_alloc(GFP_NOFS);
1071 if (!*leafs) {
1072 ulist_free(tmp);
1073 return -ENOMEM;
1074 }
1075
1076 ret = find_parent_nodes(trans, fs_info, bytenr,
1077 time_seq, *leafs, tmp, extent_item_pos);
1078 ulist_free(tmp);
1079
1080 if (ret < 0 && ret != -ENOENT) {
1081 free_leaf_list(*leafs);
1082 return ret;
1083 }
1084
1085 return 0;
1086 }
1087
1088 /*
1089 * walk all backrefs for a given extent to find all roots that reference this
1090 * extent. Walking a backref means finding all extents that reference this
1091 * extent and in turn walk the backrefs of those, too. Naturally this is a
1092 * recursive process, but here it is implemented in an iterative fashion: We
1093 * find all referencing extents for the extent in question and put them on a
1094 * list. In turn, we find all referencing extents for those, further appending
1095 * to the list. The way we iterate the list allows adding more elements after
1096 * the current while iterating. The process stops when we reach the end of the
1097 * list. Found roots are added to the roots list.
1098 *
1099 * returns 0 on success, < 0 on error.
1100 */
1101 int btrfs_find_all_roots(struct btrfs_trans_handle *trans,
1102 struct btrfs_fs_info *fs_info, u64 bytenr,
1103 u64 time_seq, struct ulist **roots)
1104 {
1105 struct ulist *tmp;
1106 struct ulist_node *node = NULL;
1107 struct ulist_iterator uiter;
1108 int ret;
1109
1110 tmp = ulist_alloc(GFP_NOFS);
1111 if (!tmp)
1112 return -ENOMEM;
1113 *roots = ulist_alloc(GFP_NOFS);
1114 if (!*roots) {
1115 ulist_free(tmp);
1116 return -ENOMEM;
1117 }
1118
1119 ULIST_ITER_INIT(&uiter);
1120 while (1) {
1121 ret = find_parent_nodes(trans, fs_info, bytenr,
1122 time_seq, tmp, *roots, NULL);
1123 if (ret < 0 && ret != -ENOENT) {
1124 ulist_free(tmp);
1125 ulist_free(*roots);
1126 return ret;
1127 }
1128 node = ulist_next(tmp, &uiter);
1129 if (!node)
1130 break;
1131 bytenr = node->val;
1132 cond_resched();
1133 }
1134
1135 ulist_free(tmp);
1136 return 0;
1137 }
1138
1139 /*
1140 * this makes the path point to (inum INODE_ITEM ioff)
1141 */
1142 int inode_item_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
1143 struct btrfs_path *path)
1144 {
1145 struct btrfs_key key;
1146 return btrfs_find_item(fs_root, path, inum, ioff,
1147 BTRFS_INODE_ITEM_KEY, &key);
1148 }
1149
1150 static int inode_ref_info(u64 inum, u64 ioff, struct btrfs_root *fs_root,
1151 struct btrfs_path *path,
1152 struct btrfs_key *found_key)
1153 {
1154 return btrfs_find_item(fs_root, path, inum, ioff,
1155 BTRFS_INODE_REF_KEY, found_key);
1156 }
1157
1158 int btrfs_find_one_extref(struct btrfs_root *root, u64 inode_objectid,
1159 u64 start_off, struct btrfs_path *path,
1160 struct btrfs_inode_extref **ret_extref,
1161 u64 *found_off)
1162 {
1163 int ret, slot;
1164 struct btrfs_key key;
1165 struct btrfs_key found_key;
1166 struct btrfs_inode_extref *extref;
1167 struct extent_buffer *leaf;
1168 unsigned long ptr;
1169
1170 key.objectid = inode_objectid;
1171 btrfs_set_key_type(&key, BTRFS_INODE_EXTREF_KEY);
1172 key.offset = start_off;
1173
1174 ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
1175 if (ret < 0)
1176 return ret;
1177
1178 while (1) {
1179 leaf = path->nodes[0];
1180 slot = path->slots[0];
1181 if (slot >= btrfs_header_nritems(leaf)) {
1182 /*
1183 * If the item at offset is not found,
1184 * btrfs_search_slot will point us to the slot
1185 * where it should be inserted. In our case
1186 * that will be the slot directly before the
1187 * next INODE_REF_KEY_V2 item. In the case
1188 * that we're pointing to the last slot in a
1189 * leaf, we must move one leaf over.
1190 */
1191 ret = btrfs_next_leaf(root, path);
1192 if (ret) {
1193 if (ret >= 1)
1194 ret = -ENOENT;
1195 break;
1196 }
1197 continue;
1198 }
1199
1200 btrfs_item_key_to_cpu(leaf, &found_key, slot);
1201
1202 /*
1203 * Check that we're still looking at an extended ref key for
1204 * this particular objectid. If we have different
1205 * objectid or type then there are no more to be found
1206 * in the tree and we can exit.
1207 */
1208 ret = -ENOENT;
1209 if (found_key.objectid != inode_objectid)
1210 break;
1211 if (btrfs_key_type(&found_key) != BTRFS_INODE_EXTREF_KEY)
1212 break;
1213
1214 ret = 0;
1215 ptr = btrfs_item_ptr_offset(leaf, path->slots[0]);
1216 extref = (struct btrfs_inode_extref *)ptr;
1217 *ret_extref = extref;
1218 if (found_off)
1219 *found_off = found_key.offset;
1220 break;
1221 }
1222
1223 return ret;
1224 }
1225
1226 /*
1227 * this iterates to turn a name (from iref/extref) into a full filesystem path.
1228 * Elements of the path are separated by '/' and the path is guaranteed to be
1229 * 0-terminated. the path is only given within the current file system.
1230 * Therefore, it never starts with a '/'. the caller is responsible to provide
1231 * "size" bytes in "dest". the dest buffer will be filled backwards. finally,
1232 * the start point of the resulting string is returned. this pointer is within
1233 * dest, normally.
1234 * in case the path buffer would overflow, the pointer is decremented further
1235 * as if output was written to the buffer, though no more output is actually
1236 * generated. that way, the caller can determine how much space would be
1237 * required for the path to fit into the buffer. in that case, the returned
1238 * value will be smaller than dest. callers must check this!
1239 */
1240 char *btrfs_ref_to_path(struct btrfs_root *fs_root, struct btrfs_path *path,
1241 u32 name_len, unsigned long name_off,
1242 struct extent_buffer *eb_in, u64 parent,
1243 char *dest, u32 size)
1244 {
1245 int slot;
1246 u64 next_inum;
1247 int ret;
1248 s64 bytes_left = ((s64)size) - 1;
1249 struct extent_buffer *eb = eb_in;
1250 struct btrfs_key found_key;
1251 int leave_spinning = path->leave_spinning;
1252 struct btrfs_inode_ref *iref;
1253
1254 if (bytes_left >= 0)
1255 dest[bytes_left] = '\0';
1256
1257 path->leave_spinning = 1;
1258 while (1) {
1259 bytes_left -= name_len;
1260 if (bytes_left >= 0)
1261 read_extent_buffer(eb, dest + bytes_left,
1262 name_off, name_len);
1263 if (eb != eb_in) {
1264 btrfs_tree_read_unlock_blocking(eb);
1265 free_extent_buffer(eb);
1266 }
1267 ret = inode_ref_info(parent, 0, fs_root, path, &found_key);
1268 if (ret > 0)
1269 ret = -ENOENT;
1270 if (ret)
1271 break;
1272
1273 next_inum = found_key.offset;
1274
1275 /* regular exit ahead */
1276 if (parent == next_inum)
1277 break;
1278
1279 slot = path->slots[0];
1280 eb = path->nodes[0];
1281 /* make sure we can use eb after releasing the path */
1282 if (eb != eb_in) {
1283 atomic_inc(&eb->refs);
1284 btrfs_tree_read_lock(eb);
1285 btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1286 }
1287 btrfs_release_path(path);
1288 iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
1289
1290 name_len = btrfs_inode_ref_name_len(eb, iref);
1291 name_off = (unsigned long)(iref + 1);
1292
1293 parent = next_inum;
1294 --bytes_left;
1295 if (bytes_left >= 0)
1296 dest[bytes_left] = '/';
1297 }
1298
1299 btrfs_release_path(path);
1300 path->leave_spinning = leave_spinning;
1301
1302 if (ret)
1303 return ERR_PTR(ret);
1304
1305 return dest + bytes_left;
1306 }
1307
1308 /*
1309 * this makes the path point to (logical EXTENT_ITEM *)
1310 * returns BTRFS_EXTENT_FLAG_DATA for data, BTRFS_EXTENT_FLAG_TREE_BLOCK for
1311 * tree blocks and <0 on error.
1312 */
1313 int extent_from_logical(struct btrfs_fs_info *fs_info, u64 logical,
1314 struct btrfs_path *path, struct btrfs_key *found_key,
1315 u64 *flags_ret)
1316 {
1317 int ret;
1318 u64 flags;
1319 u64 size = 0;
1320 u32 item_size;
1321 struct extent_buffer *eb;
1322 struct btrfs_extent_item *ei;
1323 struct btrfs_key key;
1324
1325 if (btrfs_fs_incompat(fs_info, SKINNY_METADATA))
1326 key.type = BTRFS_METADATA_ITEM_KEY;
1327 else
1328 key.type = BTRFS_EXTENT_ITEM_KEY;
1329 key.objectid = logical;
1330 key.offset = (u64)-1;
1331
1332 ret = btrfs_search_slot(NULL, fs_info->extent_root, &key, path, 0, 0);
1333 if (ret < 0)
1334 return ret;
1335
1336 while (1) {
1337 u32 nritems;
1338 if (path->slots[0] == 0) {
1339 btrfs_set_path_blocking(path);
1340 ret = btrfs_prev_leaf(fs_info->extent_root, path);
1341 if (ret != 0) {
1342 if (ret > 0) {
1343 pr_debug("logical %llu is not within "
1344 "any extent\n", logical);
1345 ret = -ENOENT;
1346 }
1347 return ret;
1348 }
1349 } else {
1350 path->slots[0]--;
1351 }
1352 nritems = btrfs_header_nritems(path->nodes[0]);
1353 if (nritems == 0) {
1354 pr_debug("logical %llu is not within any extent\n",
1355 logical);
1356 return -ENOENT;
1357 }
1358 if (path->slots[0] == nritems)
1359 path->slots[0]--;
1360
1361 btrfs_item_key_to_cpu(path->nodes[0], found_key,
1362 path->slots[0]);
1363 if (found_key->type == BTRFS_EXTENT_ITEM_KEY ||
1364 found_key->type == BTRFS_METADATA_ITEM_KEY)
1365 break;
1366 }
1367
1368 if (found_key->type == BTRFS_METADATA_ITEM_KEY)
1369 size = fs_info->extent_root->leafsize;
1370 else if (found_key->type == BTRFS_EXTENT_ITEM_KEY)
1371 size = found_key->offset;
1372
1373 if (found_key->objectid > logical ||
1374 found_key->objectid + size <= logical) {
1375 pr_debug("logical %llu is not within any extent\n", logical);
1376 return -ENOENT;
1377 }
1378
1379 eb = path->nodes[0];
1380 item_size = btrfs_item_size_nr(eb, path->slots[0]);
1381 BUG_ON(item_size < sizeof(*ei));
1382
1383 ei = btrfs_item_ptr(eb, path->slots[0], struct btrfs_extent_item);
1384 flags = btrfs_extent_flags(eb, ei);
1385
1386 pr_debug("logical %llu is at position %llu within the extent (%llu "
1387 "EXTENT_ITEM %llu) flags %#llx size %u\n",
1388 logical, logical - found_key->objectid, found_key->objectid,
1389 found_key->offset, flags, item_size);
1390
1391 WARN_ON(!flags_ret);
1392 if (flags_ret) {
1393 if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)
1394 *flags_ret = BTRFS_EXTENT_FLAG_TREE_BLOCK;
1395 else if (flags & BTRFS_EXTENT_FLAG_DATA)
1396 *flags_ret = BTRFS_EXTENT_FLAG_DATA;
1397 else
1398 BUG_ON(1);
1399 return 0;
1400 }
1401
1402 return -EIO;
1403 }
1404
1405 /*
1406 * helper function to iterate extent inline refs. ptr must point to a 0 value
1407 * for the first call and may be modified. it is used to track state.
1408 * if more refs exist, 0 is returned and the next call to
1409 * __get_extent_inline_ref must pass the modified ptr parameter to get the
1410 * next ref. after the last ref was processed, 1 is returned.
1411 * returns <0 on error
1412 */
1413 static int __get_extent_inline_ref(unsigned long *ptr, struct extent_buffer *eb,
1414 struct btrfs_extent_item *ei, u32 item_size,
1415 struct btrfs_extent_inline_ref **out_eiref,
1416 int *out_type)
1417 {
1418 unsigned long end;
1419 u64 flags;
1420 struct btrfs_tree_block_info *info;
1421
1422 if (!*ptr) {
1423 /* first call */
1424 flags = btrfs_extent_flags(eb, ei);
1425 if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
1426 info = (struct btrfs_tree_block_info *)(ei + 1);
1427 *out_eiref =
1428 (struct btrfs_extent_inline_ref *)(info + 1);
1429 } else {
1430 *out_eiref = (struct btrfs_extent_inline_ref *)(ei + 1);
1431 }
1432 *ptr = (unsigned long)*out_eiref;
1433 if ((void *)*ptr >= (void *)ei + item_size)
1434 return -ENOENT;
1435 }
1436
1437 end = (unsigned long)ei + item_size;
1438 *out_eiref = (struct btrfs_extent_inline_ref *)*ptr;
1439 *out_type = btrfs_extent_inline_ref_type(eb, *out_eiref);
1440
1441 *ptr += btrfs_extent_inline_ref_size(*out_type);
1442 WARN_ON(*ptr > end);
1443 if (*ptr == end)
1444 return 1; /* last */
1445
1446 return 0;
1447 }
1448
1449 /*
1450 * reads the tree block backref for an extent. tree level and root are returned
1451 * through out_level and out_root. ptr must point to a 0 value for the first
1452 * call and may be modified (see __get_extent_inline_ref comment).
1453 * returns 0 if data was provided, 1 if there was no more data to provide or
1454 * <0 on error.
1455 */
1456 int tree_backref_for_extent(unsigned long *ptr, struct extent_buffer *eb,
1457 struct btrfs_extent_item *ei, u32 item_size,
1458 u64 *out_root, u8 *out_level)
1459 {
1460 int ret;
1461 int type;
1462 struct btrfs_tree_block_info *info;
1463 struct btrfs_extent_inline_ref *eiref;
1464
1465 if (*ptr == (unsigned long)-1)
1466 return 1;
1467
1468 while (1) {
1469 ret = __get_extent_inline_ref(ptr, eb, ei, item_size,
1470 &eiref, &type);
1471 if (ret < 0)
1472 return ret;
1473
1474 if (type == BTRFS_TREE_BLOCK_REF_KEY ||
1475 type == BTRFS_SHARED_BLOCK_REF_KEY)
1476 break;
1477
1478 if (ret == 1)
1479 return 1;
1480 }
1481
1482 /* we can treat both ref types equally here */
1483 info = (struct btrfs_tree_block_info *)(ei + 1);
1484 *out_root = btrfs_extent_inline_ref_offset(eb, eiref);
1485 *out_level = btrfs_tree_block_level(eb, info);
1486
1487 if (ret == 1)
1488 *ptr = (unsigned long)-1;
1489
1490 return 0;
1491 }
1492
1493 static int iterate_leaf_refs(struct extent_inode_elem *inode_list,
1494 u64 root, u64 extent_item_objectid,
1495 iterate_extent_inodes_t *iterate, void *ctx)
1496 {
1497 struct extent_inode_elem *eie;
1498 int ret = 0;
1499
1500 for (eie = inode_list; eie; eie = eie->next) {
1501 pr_debug("ref for %llu resolved, key (%llu EXTEND_DATA %llu), "
1502 "root %llu\n", extent_item_objectid,
1503 eie->inum, eie->offset, root);
1504 ret = iterate(eie->inum, eie->offset, root, ctx);
1505 if (ret) {
1506 pr_debug("stopping iteration for %llu due to ret=%d\n",
1507 extent_item_objectid, ret);
1508 break;
1509 }
1510 }
1511
1512 return ret;
1513 }
1514
1515 /*
1516 * calls iterate() for every inode that references the extent identified by
1517 * the given parameters.
1518 * when the iterator function returns a non-zero value, iteration stops.
1519 */
1520 int iterate_extent_inodes(struct btrfs_fs_info *fs_info,
1521 u64 extent_item_objectid, u64 extent_item_pos,
1522 int search_commit_root,
1523 iterate_extent_inodes_t *iterate, void *ctx)
1524 {
1525 int ret;
1526 struct btrfs_trans_handle *trans = NULL;
1527 struct ulist *refs = NULL;
1528 struct ulist *roots = NULL;
1529 struct ulist_node *ref_node = NULL;
1530 struct ulist_node *root_node = NULL;
1531 struct seq_list tree_mod_seq_elem = {};
1532 struct ulist_iterator ref_uiter;
1533 struct ulist_iterator root_uiter;
1534
1535 pr_debug("resolving all inodes for extent %llu\n",
1536 extent_item_objectid);
1537
1538 if (!search_commit_root) {
1539 trans = btrfs_join_transaction(fs_info->extent_root);
1540 if (IS_ERR(trans))
1541 return PTR_ERR(trans);
1542 btrfs_get_tree_mod_seq(fs_info, &tree_mod_seq_elem);
1543 }
1544
1545 ret = btrfs_find_all_leafs(trans, fs_info, extent_item_objectid,
1546 tree_mod_seq_elem.seq, &refs,
1547 &extent_item_pos);
1548 if (ret)
1549 goto out;
1550
1551 ULIST_ITER_INIT(&ref_uiter);
1552 while (!ret && (ref_node = ulist_next(refs, &ref_uiter))) {
1553 ret = btrfs_find_all_roots(trans, fs_info, ref_node->val,
1554 tree_mod_seq_elem.seq, &roots);
1555 if (ret)
1556 break;
1557 ULIST_ITER_INIT(&root_uiter);
1558 while (!ret && (root_node = ulist_next(roots, &root_uiter))) {
1559 pr_debug("root %llu references leaf %llu, data list "
1560 "%#llx\n", root_node->val, ref_node->val,
1561 ref_node->aux);
1562 ret = iterate_leaf_refs((struct extent_inode_elem *)
1563 (uintptr_t)ref_node->aux,
1564 root_node->val,
1565 extent_item_objectid,
1566 iterate, ctx);
1567 }
1568 ulist_free(roots);
1569 }
1570
1571 free_leaf_list(refs);
1572 out:
1573 if (!search_commit_root) {
1574 btrfs_put_tree_mod_seq(fs_info, &tree_mod_seq_elem);
1575 btrfs_end_transaction(trans, fs_info->extent_root);
1576 }
1577
1578 return ret;
1579 }
1580
1581 int iterate_inodes_from_logical(u64 logical, struct btrfs_fs_info *fs_info,
1582 struct btrfs_path *path,
1583 iterate_extent_inodes_t *iterate, void *ctx)
1584 {
1585 int ret;
1586 u64 extent_item_pos;
1587 u64 flags = 0;
1588 struct btrfs_key found_key;
1589 int search_commit_root = path->search_commit_root;
1590
1591 ret = extent_from_logical(fs_info, logical, path, &found_key, &flags);
1592 btrfs_release_path(path);
1593 if (ret < 0)
1594 return ret;
1595 if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)
1596 return -EINVAL;
1597
1598 extent_item_pos = logical - found_key.objectid;
1599 ret = iterate_extent_inodes(fs_info, found_key.objectid,
1600 extent_item_pos, search_commit_root,
1601 iterate, ctx);
1602
1603 return ret;
1604 }
1605
1606 typedef int (iterate_irefs_t)(u64 parent, u32 name_len, unsigned long name_off,
1607 struct extent_buffer *eb, void *ctx);
1608
1609 static int iterate_inode_refs(u64 inum, struct btrfs_root *fs_root,
1610 struct btrfs_path *path,
1611 iterate_irefs_t *iterate, void *ctx)
1612 {
1613 int ret = 0;
1614 int slot;
1615 u32 cur;
1616 u32 len;
1617 u32 name_len;
1618 u64 parent = 0;
1619 int found = 0;
1620 struct extent_buffer *eb;
1621 struct btrfs_item *item;
1622 struct btrfs_inode_ref *iref;
1623 struct btrfs_key found_key;
1624
1625 while (!ret) {
1626 ret = inode_ref_info(inum, parent ? parent+1 : 0, fs_root, path,
1627 &found_key);
1628 if (ret < 0)
1629 break;
1630 if (ret) {
1631 ret = found ? 0 : -ENOENT;
1632 break;
1633 }
1634 ++found;
1635
1636 parent = found_key.offset;
1637 slot = path->slots[0];
1638 eb = btrfs_clone_extent_buffer(path->nodes[0]);
1639 if (!eb) {
1640 ret = -ENOMEM;
1641 break;
1642 }
1643 extent_buffer_get(eb);
1644 btrfs_tree_read_lock(eb);
1645 btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1646 btrfs_release_path(path);
1647
1648 item = btrfs_item_nr(slot);
1649 iref = btrfs_item_ptr(eb, slot, struct btrfs_inode_ref);
1650
1651 for (cur = 0; cur < btrfs_item_size(eb, item); cur += len) {
1652 name_len = btrfs_inode_ref_name_len(eb, iref);
1653 /* path must be released before calling iterate()! */
1654 pr_debug("following ref at offset %u for inode %llu in "
1655 "tree %llu\n", cur, found_key.objectid,
1656 fs_root->objectid);
1657 ret = iterate(parent, name_len,
1658 (unsigned long)(iref + 1), eb, ctx);
1659 if (ret)
1660 break;
1661 len = sizeof(*iref) + name_len;
1662 iref = (struct btrfs_inode_ref *)((char *)iref + len);
1663 }
1664 btrfs_tree_read_unlock_blocking(eb);
1665 free_extent_buffer(eb);
1666 }
1667
1668 btrfs_release_path(path);
1669
1670 return ret;
1671 }
1672
1673 static int iterate_inode_extrefs(u64 inum, struct btrfs_root *fs_root,
1674 struct btrfs_path *path,
1675 iterate_irefs_t *iterate, void *ctx)
1676 {
1677 int ret;
1678 int slot;
1679 u64 offset = 0;
1680 u64 parent;
1681 int found = 0;
1682 struct extent_buffer *eb;
1683 struct btrfs_inode_extref *extref;
1684 struct extent_buffer *leaf;
1685 u32 item_size;
1686 u32 cur_offset;
1687 unsigned long ptr;
1688
1689 while (1) {
1690 ret = btrfs_find_one_extref(fs_root, inum, offset, path, &extref,
1691 &offset);
1692 if (ret < 0)
1693 break;
1694 if (ret) {
1695 ret = found ? 0 : -ENOENT;
1696 break;
1697 }
1698 ++found;
1699
1700 slot = path->slots[0];
1701 eb = btrfs_clone_extent_buffer(path->nodes[0]);
1702 if (!eb) {
1703 ret = -ENOMEM;
1704 break;
1705 }
1706 extent_buffer_get(eb);
1707
1708 btrfs_tree_read_lock(eb);
1709 btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
1710 btrfs_release_path(path);
1711
1712 leaf = path->nodes[0];
1713 item_size = btrfs_item_size_nr(leaf, slot);
1714 ptr = btrfs_item_ptr_offset(leaf, slot);
1715 cur_offset = 0;
1716
1717 while (cur_offset < item_size) {
1718 u32 name_len;
1719
1720 extref = (struct btrfs_inode_extref *)(ptr + cur_offset);
1721 parent = btrfs_inode_extref_parent(eb, extref);
1722 name_len = btrfs_inode_extref_name_len(eb, extref);
1723 ret = iterate(parent, name_len,
1724 (unsigned long)&extref->name, eb, ctx);
1725 if (ret)
1726 break;
1727
1728 cur_offset += btrfs_inode_extref_name_len(leaf, extref);
1729 cur_offset += sizeof(*extref);
1730 }
1731 btrfs_tree_read_unlock_blocking(eb);
1732 free_extent_buffer(eb);
1733
1734 offset++;
1735 }
1736
1737 btrfs_release_path(path);
1738
1739 return ret;
1740 }
1741
1742 static int iterate_irefs(u64 inum, struct btrfs_root *fs_root,
1743 struct btrfs_path *path, iterate_irefs_t *iterate,
1744 void *ctx)
1745 {
1746 int ret;
1747 int found_refs = 0;
1748
1749 ret = iterate_inode_refs(inum, fs_root, path, iterate, ctx);
1750 if (!ret)
1751 ++found_refs;
1752 else if (ret != -ENOENT)
1753 return ret;
1754
1755 ret = iterate_inode_extrefs(inum, fs_root, path, iterate, ctx);
1756 if (ret == -ENOENT && found_refs)
1757 return 0;
1758
1759 return ret;
1760 }
1761
1762 /*
1763 * returns 0 if the path could be dumped (probably truncated)
1764 * returns <0 in case of an error
1765 */
1766 static int inode_to_path(u64 inum, u32 name_len, unsigned long name_off,
1767 struct extent_buffer *eb, void *ctx)
1768 {
1769 struct inode_fs_paths *ipath = ctx;
1770 char *fspath;
1771 char *fspath_min;
1772 int i = ipath->fspath->elem_cnt;
1773 const int s_ptr = sizeof(char *);
1774 u32 bytes_left;
1775
1776 bytes_left = ipath->fspath->bytes_left > s_ptr ?
1777 ipath->fspath->bytes_left - s_ptr : 0;
1778
1779 fspath_min = (char *)ipath->fspath->val + (i + 1) * s_ptr;
1780 fspath = btrfs_ref_to_path(ipath->fs_root, ipath->btrfs_path, name_len,
1781 name_off, eb, inum, fspath_min, bytes_left);
1782 if (IS_ERR(fspath))
1783 return PTR_ERR(fspath);
1784
1785 if (fspath > fspath_min) {
1786 ipath->fspath->val[i] = (u64)(unsigned long)fspath;
1787 ++ipath->fspath->elem_cnt;
1788 ipath->fspath->bytes_left = fspath - fspath_min;
1789 } else {
1790 ++ipath->fspath->elem_missed;
1791 ipath->fspath->bytes_missing += fspath_min - fspath;
1792 ipath->fspath->bytes_left = 0;
1793 }
1794
1795 return 0;
1796 }
1797
1798 /*
1799 * this dumps all file system paths to the inode into the ipath struct, provided
1800 * is has been created large enough. each path is zero-terminated and accessed
1801 * from ipath->fspath->val[i].
1802 * when it returns, there are ipath->fspath->elem_cnt number of paths available
1803 * in ipath->fspath->val[]. when the allocated space wasn't sufficient, the
1804 * number of missed paths in recored in ipath->fspath->elem_missed, otherwise,
1805 * it's zero. ipath->fspath->bytes_missing holds the number of bytes that would
1806 * have been needed to return all paths.
1807 */
1808 int paths_from_inode(u64 inum, struct inode_fs_paths *ipath)
1809 {
1810 return iterate_irefs(inum, ipath->fs_root, ipath->btrfs_path,
1811 inode_to_path, ipath);
1812 }
1813
1814 struct btrfs_data_container *init_data_container(u32 total_bytes)
1815 {
1816 struct btrfs_data_container *data;
1817 size_t alloc_bytes;
1818
1819 alloc_bytes = max_t(size_t, total_bytes, sizeof(*data));
1820 data = vmalloc(alloc_bytes);
1821 if (!data)
1822 return ERR_PTR(-ENOMEM);
1823
1824 if (total_bytes >= sizeof(*data)) {
1825 data->bytes_left = total_bytes - sizeof(*data);
1826 data->bytes_missing = 0;
1827 } else {
1828 data->bytes_missing = sizeof(*data) - total_bytes;
1829 data->bytes_left = 0;
1830 }
1831
1832 data->elem_cnt = 0;
1833 data->elem_missed = 0;
1834
1835 return data;
1836 }
1837
1838 /*
1839 * allocates space to return multiple file system paths for an inode.
1840 * total_bytes to allocate are passed, note that space usable for actual path
1841 * information will be total_bytes - sizeof(struct inode_fs_paths).
1842 * the returned pointer must be freed with free_ipath() in the end.
1843 */
1844 struct inode_fs_paths *init_ipath(s32 total_bytes, struct btrfs_root *fs_root,
1845 struct btrfs_path *path)
1846 {
1847 struct inode_fs_paths *ifp;
1848 struct btrfs_data_container *fspath;
1849
1850 fspath = init_data_container(total_bytes);
1851 if (IS_ERR(fspath))
1852 return (void *)fspath;
1853
1854 ifp = kmalloc(sizeof(*ifp), GFP_NOFS);
1855 if (!ifp) {
1856 kfree(fspath);
1857 return ERR_PTR(-ENOMEM);
1858 }
1859
1860 ifp->btrfs_path = path;
1861 ifp->fspath = fspath;
1862 ifp->fs_root = fs_root;
1863
1864 return ifp;
1865 }
1866
1867 void free_ipath(struct inode_fs_paths *ipath)
1868 {
1869 if (!ipath)
1870 return;
1871 vfree(ipath->fspath);
1872 kfree(ipath);
1873 }
Linux with added FPC-III support