xref: /openbmc/linux/fs/btrfs/ref-verify.c (revision 9be08a27)
1 // SPDX-License-Identifier: GPL-2.0
2 /*
3  * Copyright (C) 2014 Facebook.  All rights reserved.
4  */
5 
6 #include <linux/sched.h>
7 #include <linux/stacktrace.h>
8 #include "ctree.h"
9 #include "disk-io.h"
10 #include "locking.h"
11 #include "delayed-ref.h"
12 #include "ref-verify.h"
13 
14 /*
15  * Used to keep track the roots and number of refs each root has for a given
16  * bytenr.  This just tracks the number of direct references, no shared
17  * references.
18  */
19 struct root_entry {
20 	u64 root_objectid;
21 	u64 num_refs;
22 	struct rb_node node;
23 };
24 
25 /*
26  * These are meant to represent what should exist in the extent tree, these can
27  * be used to verify the extent tree is consistent as these should all match
28  * what the extent tree says.
29  */
30 struct ref_entry {
31 	u64 root_objectid;
32 	u64 parent;
33 	u64 owner;
34 	u64 offset;
35 	u64 num_refs;
36 	struct rb_node node;
37 };
38 
39 #define MAX_TRACE	16
40 
41 /*
42  * Whenever we add/remove a reference we record the action.  The action maps
43  * back to the delayed ref action.  We hold the ref we are changing in the
44  * action so we can account for the history properly, and we record the root we
45  * were called with since it could be different from ref_root.  We also store
46  * stack traces because thats how I roll.
47  */
48 struct ref_action {
49 	int action;
50 	u64 root;
51 	struct ref_entry ref;
52 	struct list_head list;
53 	unsigned long trace[MAX_TRACE];
54 	unsigned int trace_len;
55 };
56 
57 /*
58  * One of these for every block we reference, it holds the roots and references
59  * to it as well as all of the ref actions that have occured to it.  We never
60  * free it until we unmount the file system in order to make sure re-allocations
61  * are happening properly.
62  */
63 struct block_entry {
64 	u64 bytenr;
65 	u64 len;
66 	u64 num_refs;
67 	int metadata;
68 	int from_disk;
69 	struct rb_root roots;
70 	struct rb_root refs;
71 	struct rb_node node;
72 	struct list_head actions;
73 };
74 
75 static struct block_entry *insert_block_entry(struct rb_root *root,
76 					      struct block_entry *be)
77 {
78 	struct rb_node **p = &root->rb_node;
79 	struct rb_node *parent_node = NULL;
80 	struct block_entry *entry;
81 
82 	while (*p) {
83 		parent_node = *p;
84 		entry = rb_entry(parent_node, struct block_entry, node);
85 		if (entry->bytenr > be->bytenr)
86 			p = &(*p)->rb_left;
87 		else if (entry->bytenr < be->bytenr)
88 			p = &(*p)->rb_right;
89 		else
90 			return entry;
91 	}
92 
93 	rb_link_node(&be->node, parent_node, p);
94 	rb_insert_color(&be->node, root);
95 	return NULL;
96 }
97 
98 static struct block_entry *lookup_block_entry(struct rb_root *root, u64 bytenr)
99 {
100 	struct rb_node *n;
101 	struct block_entry *entry = NULL;
102 
103 	n = root->rb_node;
104 	while (n) {
105 		entry = rb_entry(n, struct block_entry, node);
106 		if (entry->bytenr < bytenr)
107 			n = n->rb_right;
108 		else if (entry->bytenr > bytenr)
109 			n = n->rb_left;
110 		else
111 			return entry;
112 	}
113 	return NULL;
114 }
115 
116 static struct root_entry *insert_root_entry(struct rb_root *root,
117 					    struct root_entry *re)
118 {
119 	struct rb_node **p = &root->rb_node;
120 	struct rb_node *parent_node = NULL;
121 	struct root_entry *entry;
122 
123 	while (*p) {
124 		parent_node = *p;
125 		entry = rb_entry(parent_node, struct root_entry, node);
126 		if (entry->root_objectid > re->root_objectid)
127 			p = &(*p)->rb_left;
128 		else if (entry->root_objectid < re->root_objectid)
129 			p = &(*p)->rb_right;
130 		else
131 			return entry;
132 	}
133 
134 	rb_link_node(&re->node, parent_node, p);
135 	rb_insert_color(&re->node, root);
136 	return NULL;
137 
138 }
139 
140 static int comp_refs(struct ref_entry *ref1, struct ref_entry *ref2)
141 {
142 	if (ref1->root_objectid < ref2->root_objectid)
143 		return -1;
144 	if (ref1->root_objectid > ref2->root_objectid)
145 		return 1;
146 	if (ref1->parent < ref2->parent)
147 		return -1;
148 	if (ref1->parent > ref2->parent)
149 		return 1;
150 	if (ref1->owner < ref2->owner)
151 		return -1;
152 	if (ref1->owner > ref2->owner)
153 		return 1;
154 	if (ref1->offset < ref2->offset)
155 		return -1;
156 	if (ref1->offset > ref2->offset)
157 		return 1;
158 	return 0;
159 }
160 
161 static struct ref_entry *insert_ref_entry(struct rb_root *root,
162 					  struct ref_entry *ref)
163 {
164 	struct rb_node **p = &root->rb_node;
165 	struct rb_node *parent_node = NULL;
166 	struct ref_entry *entry;
167 	int cmp;
168 
169 	while (*p) {
170 		parent_node = *p;
171 		entry = rb_entry(parent_node, struct ref_entry, node);
172 		cmp = comp_refs(entry, ref);
173 		if (cmp > 0)
174 			p = &(*p)->rb_left;
175 		else if (cmp < 0)
176 			p = &(*p)->rb_right;
177 		else
178 			return entry;
179 	}
180 
181 	rb_link_node(&ref->node, parent_node, p);
182 	rb_insert_color(&ref->node, root);
183 	return NULL;
184 
185 }
186 
187 static struct root_entry *lookup_root_entry(struct rb_root *root, u64 objectid)
188 {
189 	struct rb_node *n;
190 	struct root_entry *entry = NULL;
191 
192 	n = root->rb_node;
193 	while (n) {
194 		entry = rb_entry(n, struct root_entry, node);
195 		if (entry->root_objectid < objectid)
196 			n = n->rb_right;
197 		else if (entry->root_objectid > objectid)
198 			n = n->rb_left;
199 		else
200 			return entry;
201 	}
202 	return NULL;
203 }
204 
205 #ifdef CONFIG_STACKTRACE
206 static void __save_stack_trace(struct ref_action *ra)
207 {
208 	struct stack_trace stack_trace;
209 
210 	stack_trace.max_entries = MAX_TRACE;
211 	stack_trace.nr_entries = 0;
212 	stack_trace.entries = ra->trace;
213 	stack_trace.skip = 2;
214 	save_stack_trace(&stack_trace);
215 	ra->trace_len = stack_trace.nr_entries;
216 }
217 
218 static void __print_stack_trace(struct btrfs_fs_info *fs_info,
219 				struct ref_action *ra)
220 {
221 	struct stack_trace trace;
222 
223 	if (ra->trace_len == 0) {
224 		btrfs_err(fs_info, "  ref-verify: no stacktrace");
225 		return;
226 	}
227 	trace.nr_entries = ra->trace_len;
228 	trace.entries = ra->trace;
229 	print_stack_trace(&trace, 2);
230 }
231 #else
232 static void inline __save_stack_trace(struct ref_action *ra)
233 {
234 }
235 
236 static void inline __print_stack_trace(struct btrfs_fs_info *fs_info,
237 				       struct ref_action *ra)
238 {
239 	btrfs_err(fs_info, "  ref-verify: no stacktrace support");
240 }
241 #endif
242 
243 static void free_block_entry(struct block_entry *be)
244 {
245 	struct root_entry *re;
246 	struct ref_entry *ref;
247 	struct ref_action *ra;
248 	struct rb_node *n;
249 
250 	while ((n = rb_first(&be->roots))) {
251 		re = rb_entry(n, struct root_entry, node);
252 		rb_erase(&re->node, &be->roots);
253 		kfree(re);
254 	}
255 
256 	while((n = rb_first(&be->refs))) {
257 		ref = rb_entry(n, struct ref_entry, node);
258 		rb_erase(&ref->node, &be->refs);
259 		kfree(ref);
260 	}
261 
262 	while (!list_empty(&be->actions)) {
263 		ra = list_first_entry(&be->actions, struct ref_action,
264 				      list);
265 		list_del(&ra->list);
266 		kfree(ra);
267 	}
268 	kfree(be);
269 }
270 
271 static struct block_entry *add_block_entry(struct btrfs_fs_info *fs_info,
272 					   u64 bytenr, u64 len,
273 					   u64 root_objectid)
274 {
275 	struct block_entry *be = NULL, *exist;
276 	struct root_entry *re = NULL;
277 
278 	re = kzalloc(sizeof(struct root_entry), GFP_KERNEL);
279 	be = kzalloc(sizeof(struct block_entry), GFP_KERNEL);
280 	if (!be || !re) {
281 		kfree(re);
282 		kfree(be);
283 		return ERR_PTR(-ENOMEM);
284 	}
285 	be->bytenr = bytenr;
286 	be->len = len;
287 
288 	re->root_objectid = root_objectid;
289 	re->num_refs = 0;
290 
291 	spin_lock(&fs_info->ref_verify_lock);
292 	exist = insert_block_entry(&fs_info->block_tree, be);
293 	if (exist) {
294 		if (root_objectid) {
295 			struct root_entry *exist_re;
296 
297 			exist_re = insert_root_entry(&exist->roots, re);
298 			if (exist_re)
299 				kfree(re);
300 		}
301 		kfree(be);
302 		return exist;
303 	}
304 
305 	be->num_refs = 0;
306 	be->metadata = 0;
307 	be->from_disk = 0;
308 	be->roots = RB_ROOT;
309 	be->refs = RB_ROOT;
310 	INIT_LIST_HEAD(&be->actions);
311 	if (root_objectid)
312 		insert_root_entry(&be->roots, re);
313 	else
314 		kfree(re);
315 	return be;
316 }
317 
318 static int add_tree_block(struct btrfs_fs_info *fs_info, u64 ref_root,
319 			  u64 parent, u64 bytenr, int level)
320 {
321 	struct block_entry *be;
322 	struct root_entry *re;
323 	struct ref_entry *ref = NULL, *exist;
324 
325 	ref = kmalloc(sizeof(struct ref_entry), GFP_KERNEL);
326 	if (!ref)
327 		return -ENOMEM;
328 
329 	if (parent)
330 		ref->root_objectid = 0;
331 	else
332 		ref->root_objectid = ref_root;
333 	ref->parent = parent;
334 	ref->owner = level;
335 	ref->offset = 0;
336 	ref->num_refs = 1;
337 
338 	be = add_block_entry(fs_info, bytenr, fs_info->nodesize, ref_root);
339 	if (IS_ERR(be)) {
340 		kfree(ref);
341 		return PTR_ERR(be);
342 	}
343 	be->num_refs++;
344 	be->from_disk = 1;
345 	be->metadata = 1;
346 
347 	if (!parent) {
348 		ASSERT(ref_root);
349 		re = lookup_root_entry(&be->roots, ref_root);
350 		ASSERT(re);
351 		re->num_refs++;
352 	}
353 	exist = insert_ref_entry(&be->refs, ref);
354 	if (exist) {
355 		exist->num_refs++;
356 		kfree(ref);
357 	}
358 	spin_unlock(&fs_info->ref_verify_lock);
359 
360 	return 0;
361 }
362 
363 static int add_shared_data_ref(struct btrfs_fs_info *fs_info,
364 			       u64 parent, u32 num_refs, u64 bytenr,
365 			       u64 num_bytes)
366 {
367 	struct block_entry *be;
368 	struct ref_entry *ref;
369 
370 	ref = kzalloc(sizeof(struct ref_entry), GFP_KERNEL);
371 	if (!ref)
372 		return -ENOMEM;
373 	be = add_block_entry(fs_info, bytenr, num_bytes, 0);
374 	if (IS_ERR(be)) {
375 		kfree(ref);
376 		return PTR_ERR(be);
377 	}
378 	be->num_refs += num_refs;
379 
380 	ref->parent = parent;
381 	ref->num_refs = num_refs;
382 	if (insert_ref_entry(&be->refs, ref)) {
383 		spin_unlock(&fs_info->ref_verify_lock);
384 		btrfs_err(fs_info, "existing shared ref when reading from disk?");
385 		kfree(ref);
386 		return -EINVAL;
387 	}
388 	spin_unlock(&fs_info->ref_verify_lock);
389 	return 0;
390 }
391 
392 static int add_extent_data_ref(struct btrfs_fs_info *fs_info,
393 			       struct extent_buffer *leaf,
394 			       struct btrfs_extent_data_ref *dref,
395 			       u64 bytenr, u64 num_bytes)
396 {
397 	struct block_entry *be;
398 	struct ref_entry *ref;
399 	struct root_entry *re;
400 	u64 ref_root = btrfs_extent_data_ref_root(leaf, dref);
401 	u64 owner = btrfs_extent_data_ref_objectid(leaf, dref);
402 	u64 offset = btrfs_extent_data_ref_offset(leaf, dref);
403 	u32 num_refs = btrfs_extent_data_ref_count(leaf, dref);
404 
405 	ref = kzalloc(sizeof(struct ref_entry), GFP_KERNEL);
406 	if (!ref)
407 		return -ENOMEM;
408 	be = add_block_entry(fs_info, bytenr, num_bytes, ref_root);
409 	if (IS_ERR(be)) {
410 		kfree(ref);
411 		return PTR_ERR(be);
412 	}
413 	be->num_refs += num_refs;
414 
415 	ref->parent = 0;
416 	ref->owner = owner;
417 	ref->root_objectid = ref_root;
418 	ref->offset = offset;
419 	ref->num_refs = num_refs;
420 	if (insert_ref_entry(&be->refs, ref)) {
421 		spin_unlock(&fs_info->ref_verify_lock);
422 		btrfs_err(fs_info, "existing ref when reading from disk?");
423 		kfree(ref);
424 		return -EINVAL;
425 	}
426 
427 	re = lookup_root_entry(&be->roots, ref_root);
428 	if (!re) {
429 		spin_unlock(&fs_info->ref_verify_lock);
430 		btrfs_err(fs_info, "missing root in new block entry?");
431 		return -EINVAL;
432 	}
433 	re->num_refs += num_refs;
434 	spin_unlock(&fs_info->ref_verify_lock);
435 	return 0;
436 }
437 
438 static int process_extent_item(struct btrfs_fs_info *fs_info,
439 			       struct btrfs_path *path, struct btrfs_key *key,
440 			       int slot, int *tree_block_level)
441 {
442 	struct btrfs_extent_item *ei;
443 	struct btrfs_extent_inline_ref *iref;
444 	struct btrfs_extent_data_ref *dref;
445 	struct btrfs_shared_data_ref *sref;
446 	struct extent_buffer *leaf = path->nodes[0];
447 	u32 item_size = btrfs_item_size_nr(leaf, slot);
448 	unsigned long end, ptr;
449 	u64 offset, flags, count;
450 	int type, ret;
451 
452 	ei = btrfs_item_ptr(leaf, slot, struct btrfs_extent_item);
453 	flags = btrfs_extent_flags(leaf, ei);
454 
455 	if ((key->type == BTRFS_EXTENT_ITEM_KEY) &&
456 	    flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
457 		struct btrfs_tree_block_info *info;
458 
459 		info = (struct btrfs_tree_block_info *)(ei + 1);
460 		*tree_block_level = btrfs_tree_block_level(leaf, info);
461 		iref = (struct btrfs_extent_inline_ref *)(info + 1);
462 	} else {
463 		if (key->type == BTRFS_METADATA_ITEM_KEY)
464 			*tree_block_level = key->offset;
465 		iref = (struct btrfs_extent_inline_ref *)(ei + 1);
466 	}
467 
468 	ptr = (unsigned long)iref;
469 	end = (unsigned long)ei + item_size;
470 	while (ptr < end) {
471 		iref = (struct btrfs_extent_inline_ref *)ptr;
472 		type = btrfs_extent_inline_ref_type(leaf, iref);
473 		offset = btrfs_extent_inline_ref_offset(leaf, iref);
474 		switch (type) {
475 		case BTRFS_TREE_BLOCK_REF_KEY:
476 			ret = add_tree_block(fs_info, offset, 0, key->objectid,
477 					     *tree_block_level);
478 			break;
479 		case BTRFS_SHARED_BLOCK_REF_KEY:
480 			ret = add_tree_block(fs_info, 0, offset, key->objectid,
481 					     *tree_block_level);
482 			break;
483 		case BTRFS_EXTENT_DATA_REF_KEY:
484 			dref = (struct btrfs_extent_data_ref *)(&iref->offset);
485 			ret = add_extent_data_ref(fs_info, leaf, dref,
486 						  key->objectid, key->offset);
487 			break;
488 		case BTRFS_SHARED_DATA_REF_KEY:
489 			sref = (struct btrfs_shared_data_ref *)(iref + 1);
490 			count = btrfs_shared_data_ref_count(leaf, sref);
491 			ret = add_shared_data_ref(fs_info, offset, count,
492 						  key->objectid, key->offset);
493 			break;
494 		default:
495 			btrfs_err(fs_info, "invalid key type in iref");
496 			ret = -EINVAL;
497 			break;
498 		}
499 		if (ret)
500 			break;
501 		ptr += btrfs_extent_inline_ref_size(type);
502 	}
503 	return ret;
504 }
505 
506 static int process_leaf(struct btrfs_root *root,
507 			struct btrfs_path *path, u64 *bytenr, u64 *num_bytes)
508 {
509 	struct btrfs_fs_info *fs_info = root->fs_info;
510 	struct extent_buffer *leaf = path->nodes[0];
511 	struct btrfs_extent_data_ref *dref;
512 	struct btrfs_shared_data_ref *sref;
513 	u32 count;
514 	int i = 0, tree_block_level = 0, ret;
515 	struct btrfs_key key;
516 	int nritems = btrfs_header_nritems(leaf);
517 
518 	for (i = 0; i < nritems; i++) {
519 		btrfs_item_key_to_cpu(leaf, &key, i);
520 		switch (key.type) {
521 		case BTRFS_EXTENT_ITEM_KEY:
522 			*num_bytes = key.offset;
523 		case BTRFS_METADATA_ITEM_KEY:
524 			*bytenr = key.objectid;
525 			ret = process_extent_item(fs_info, path, &key, i,
526 						  &tree_block_level);
527 			break;
528 		case BTRFS_TREE_BLOCK_REF_KEY:
529 			ret = add_tree_block(fs_info, key.offset, 0,
530 					     key.objectid, tree_block_level);
531 			break;
532 		case BTRFS_SHARED_BLOCK_REF_KEY:
533 			ret = add_tree_block(fs_info, 0, key.offset,
534 					     key.objectid, tree_block_level);
535 			break;
536 		case BTRFS_EXTENT_DATA_REF_KEY:
537 			dref = btrfs_item_ptr(leaf, i,
538 					      struct btrfs_extent_data_ref);
539 			ret = add_extent_data_ref(fs_info, leaf, dref, *bytenr,
540 						  *num_bytes);
541 			break;
542 		case BTRFS_SHARED_DATA_REF_KEY:
543 			sref = btrfs_item_ptr(leaf, i,
544 					      struct btrfs_shared_data_ref);
545 			count = btrfs_shared_data_ref_count(leaf, sref);
546 			ret = add_shared_data_ref(fs_info, key.offset, count,
547 						  *bytenr, *num_bytes);
548 			break;
549 		default:
550 			break;
551 		}
552 		if (ret)
553 			break;
554 	}
555 	return ret;
556 }
557 
558 /* Walk down to the leaf from the given level */
559 static int walk_down_tree(struct btrfs_root *root, struct btrfs_path *path,
560 			  int level, u64 *bytenr, u64 *num_bytes)
561 {
562 	struct btrfs_fs_info *fs_info = root->fs_info;
563 	struct extent_buffer *eb;
564 	u64 block_bytenr, gen;
565 	int ret = 0;
566 
567 	while (level >= 0) {
568 		if (level) {
569 			struct btrfs_key first_key;
570 
571 			block_bytenr = btrfs_node_blockptr(path->nodes[level],
572 							   path->slots[level]);
573 			gen = btrfs_node_ptr_generation(path->nodes[level],
574 							path->slots[level]);
575 			btrfs_node_key_to_cpu(path->nodes[level], &first_key,
576 					      path->slots[level]);
577 			eb = read_tree_block(fs_info, block_bytenr, gen,
578 					     level - 1, &first_key);
579 			if (IS_ERR(eb))
580 				return PTR_ERR(eb);
581 			if (!extent_buffer_uptodate(eb)) {
582 				free_extent_buffer(eb);
583 				return -EIO;
584 			}
585 			btrfs_tree_read_lock(eb);
586 			btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
587 			path->nodes[level-1] = eb;
588 			path->slots[level-1] = 0;
589 			path->locks[level-1] = BTRFS_READ_LOCK_BLOCKING;
590 		} else {
591 			ret = process_leaf(root, path, bytenr, num_bytes);
592 			if (ret)
593 				break;
594 		}
595 		level--;
596 	}
597 	return ret;
598 }
599 
600 /* Walk up to the next node that needs to be processed */
601 static int walk_up_tree(struct btrfs_path *path, int *level)
602 {
603 	int l;
604 
605 	for (l = 0; l < BTRFS_MAX_LEVEL; l++) {
606 		if (!path->nodes[l])
607 			continue;
608 		if (l) {
609 			path->slots[l]++;
610 			if (path->slots[l] <
611 			    btrfs_header_nritems(path->nodes[l])) {
612 				*level = l;
613 				return 0;
614 			}
615 		}
616 		btrfs_tree_unlock_rw(path->nodes[l], path->locks[l]);
617 		free_extent_buffer(path->nodes[l]);
618 		path->nodes[l] = NULL;
619 		path->slots[l] = 0;
620 		path->locks[l] = 0;
621 	}
622 
623 	return 1;
624 }
625 
626 static void dump_ref_action(struct btrfs_fs_info *fs_info,
627 			    struct ref_action *ra)
628 {
629 	btrfs_err(fs_info,
630 "  Ref action %d, root %llu, ref_root %llu, parent %llu, owner %llu, offset %llu, num_refs %llu",
631 		  ra->action, ra->root, ra->ref.root_objectid, ra->ref.parent,
632 		  ra->ref.owner, ra->ref.offset, ra->ref.num_refs);
633 	__print_stack_trace(fs_info, ra);
634 }
635 
636 /*
637  * Dumps all the information from the block entry to printk, it's going to be
638  * awesome.
639  */
640 static void dump_block_entry(struct btrfs_fs_info *fs_info,
641 			     struct block_entry *be)
642 {
643 	struct ref_entry *ref;
644 	struct root_entry *re;
645 	struct ref_action *ra;
646 	struct rb_node *n;
647 
648 	btrfs_err(fs_info,
649 "dumping block entry [%llu %llu], num_refs %llu, metadata %d, from disk %d",
650 		  be->bytenr, be->len, be->num_refs, be->metadata,
651 		  be->from_disk);
652 
653 	for (n = rb_first(&be->refs); n; n = rb_next(n)) {
654 		ref = rb_entry(n, struct ref_entry, node);
655 		btrfs_err(fs_info,
656 "  ref root %llu, parent %llu, owner %llu, offset %llu, num_refs %llu",
657 			  ref->root_objectid, ref->parent, ref->owner,
658 			  ref->offset, ref->num_refs);
659 	}
660 
661 	for (n = rb_first(&be->roots); n; n = rb_next(n)) {
662 		re = rb_entry(n, struct root_entry, node);
663 		btrfs_err(fs_info, "  root entry %llu, num_refs %llu",
664 			  re->root_objectid, re->num_refs);
665 	}
666 
667 	list_for_each_entry(ra, &be->actions, list)
668 		dump_ref_action(fs_info, ra);
669 }
670 
671 /*
672  * btrfs_ref_tree_mod: called when we modify a ref for a bytenr
673  * @root: the root we are making this modification from.
674  * @bytenr: the bytenr we are modifying.
675  * @num_bytes: number of bytes.
676  * @parent: the parent bytenr.
677  * @ref_root: the original root owner of the bytenr.
678  * @owner: level in the case of metadata, inode in the case of data.
679  * @offset: 0 for metadata, file offset for data.
680  * @action: the action that we are doing, this is the same as the delayed ref
681  *	action.
682  *
683  * This will add an action item to the given bytenr and do sanity checks to make
684  * sure we haven't messed something up.  If we are making a new allocation and
685  * this block entry has history we will delete all previous actions as long as
686  * our sanity checks pass as they are no longer needed.
687  */
688 int btrfs_ref_tree_mod(struct btrfs_root *root, u64 bytenr, u64 num_bytes,
689 		       u64 parent, u64 ref_root, u64 owner, u64 offset,
690 		       int action)
691 {
692 	struct btrfs_fs_info *fs_info = root->fs_info;
693 	struct ref_entry *ref = NULL, *exist;
694 	struct ref_action *ra = NULL;
695 	struct block_entry *be = NULL;
696 	struct root_entry *re = NULL;
697 	int ret = 0;
698 	bool metadata = owner < BTRFS_FIRST_FREE_OBJECTID;
699 
700 	if (!btrfs_test_opt(root->fs_info, REF_VERIFY))
701 		return 0;
702 
703 	ref = kzalloc(sizeof(struct ref_entry), GFP_NOFS);
704 	ra = kmalloc(sizeof(struct ref_action), GFP_NOFS);
705 	if (!ra || !ref) {
706 		kfree(ref);
707 		kfree(ra);
708 		ret = -ENOMEM;
709 		goto out;
710 	}
711 
712 	if (parent) {
713 		ref->parent = parent;
714 	} else {
715 		ref->root_objectid = ref_root;
716 		ref->owner = owner;
717 		ref->offset = offset;
718 	}
719 	ref->num_refs = (action == BTRFS_DROP_DELAYED_REF) ? -1 : 1;
720 
721 	memcpy(&ra->ref, ref, sizeof(struct ref_entry));
722 	/*
723 	 * Save the extra info from the delayed ref in the ref action to make it
724 	 * easier to figure out what is happening.  The real ref's we add to the
725 	 * ref tree need to reflect what we save on disk so it matches any
726 	 * on-disk refs we pre-loaded.
727 	 */
728 	ra->ref.owner = owner;
729 	ra->ref.offset = offset;
730 	ra->ref.root_objectid = ref_root;
731 	__save_stack_trace(ra);
732 
733 	INIT_LIST_HEAD(&ra->list);
734 	ra->action = action;
735 	ra->root = root->objectid;
736 
737 	/*
738 	 * This is an allocation, preallocate the block_entry in case we haven't
739 	 * used it before.
740 	 */
741 	ret = -EINVAL;
742 	if (action == BTRFS_ADD_DELAYED_EXTENT) {
743 		/*
744 		 * For subvol_create we'll just pass in whatever the parent root
745 		 * is and the new root objectid, so let's not treat the passed
746 		 * in root as if it really has a ref for this bytenr.
747 		 */
748 		be = add_block_entry(root->fs_info, bytenr, num_bytes, ref_root);
749 		if (IS_ERR(be)) {
750 			kfree(ra);
751 			ret = PTR_ERR(be);
752 			goto out;
753 		}
754 		be->num_refs++;
755 		if (metadata)
756 			be->metadata = 1;
757 
758 		if (be->num_refs != 1) {
759 			btrfs_err(fs_info,
760 			"re-allocated a block that still has references to it!");
761 			dump_block_entry(fs_info, be);
762 			dump_ref_action(fs_info, ra);
763 			goto out_unlock;
764 		}
765 
766 		while (!list_empty(&be->actions)) {
767 			struct ref_action *tmp;
768 
769 			tmp = list_first_entry(&be->actions, struct ref_action,
770 					       list);
771 			list_del(&tmp->list);
772 			kfree(tmp);
773 		}
774 	} else {
775 		struct root_entry *tmp;
776 
777 		if (!parent) {
778 			re = kmalloc(sizeof(struct root_entry), GFP_NOFS);
779 			if (!re) {
780 				kfree(ref);
781 				kfree(ra);
782 				ret = -ENOMEM;
783 				goto out;
784 			}
785 			/*
786 			 * This is the root that is modifying us, so it's the
787 			 * one we want to lookup below when we modify the
788 			 * re->num_refs.
789 			 */
790 			ref_root = root->objectid;
791 			re->root_objectid = root->objectid;
792 			re->num_refs = 0;
793 		}
794 
795 		spin_lock(&root->fs_info->ref_verify_lock);
796 		be = lookup_block_entry(&root->fs_info->block_tree, bytenr);
797 		if (!be) {
798 			btrfs_err(fs_info,
799 "trying to do action %d to bytenr %llu num_bytes %llu but there is no existing entry!",
800 				  action, (unsigned long long)bytenr,
801 				  (unsigned long long)num_bytes);
802 			dump_ref_action(fs_info, ra);
803 			kfree(ref);
804 			kfree(ra);
805 			goto out_unlock;
806 		}
807 
808 		if (!parent) {
809 			tmp = insert_root_entry(&be->roots, re);
810 			if (tmp) {
811 				kfree(re);
812 				re = tmp;
813 			}
814 		}
815 	}
816 
817 	exist = insert_ref_entry(&be->refs, ref);
818 	if (exist) {
819 		if (action == BTRFS_DROP_DELAYED_REF) {
820 			if (exist->num_refs == 0) {
821 				btrfs_err(fs_info,
822 "dropping a ref for a existing root that doesn't have a ref on the block");
823 				dump_block_entry(fs_info, be);
824 				dump_ref_action(fs_info, ra);
825 				kfree(ra);
826 				goto out_unlock;
827 			}
828 			exist->num_refs--;
829 			if (exist->num_refs == 0) {
830 				rb_erase(&exist->node, &be->refs);
831 				kfree(exist);
832 			}
833 		} else if (!be->metadata) {
834 			exist->num_refs++;
835 		} else {
836 			btrfs_err(fs_info,
837 "attempting to add another ref for an existing ref on a tree block");
838 			dump_block_entry(fs_info, be);
839 			dump_ref_action(fs_info, ra);
840 			kfree(ra);
841 			goto out_unlock;
842 		}
843 		kfree(ref);
844 	} else {
845 		if (action == BTRFS_DROP_DELAYED_REF) {
846 			btrfs_err(fs_info,
847 "dropping a ref for a root that doesn't have a ref on the block");
848 			dump_block_entry(fs_info, be);
849 			dump_ref_action(fs_info, ra);
850 			kfree(ra);
851 			goto out_unlock;
852 		}
853 	}
854 
855 	if (!parent && !re) {
856 		re = lookup_root_entry(&be->roots, ref_root);
857 		if (!re) {
858 			/*
859 			 * This shouldn't happen because we will add our re
860 			 * above when we lookup the be with !parent, but just in
861 			 * case catch this case so we don't panic because I
862 			 * didn't thik of some other corner case.
863 			 */
864 			btrfs_err(fs_info, "failed to find root %llu for %llu",
865 				  root->objectid, be->bytenr);
866 			dump_block_entry(fs_info, be);
867 			dump_ref_action(fs_info, ra);
868 			kfree(ra);
869 			goto out_unlock;
870 		}
871 	}
872 	if (action == BTRFS_DROP_DELAYED_REF) {
873 		if (re)
874 			re->num_refs--;
875 		be->num_refs--;
876 	} else if (action == BTRFS_ADD_DELAYED_REF) {
877 		be->num_refs++;
878 		if (re)
879 			re->num_refs++;
880 	}
881 	list_add_tail(&ra->list, &be->actions);
882 	ret = 0;
883 out_unlock:
884 	spin_unlock(&root->fs_info->ref_verify_lock);
885 out:
886 	if (ret)
887 		btrfs_clear_opt(fs_info->mount_opt, REF_VERIFY);
888 	return ret;
889 }
890 
891 /* Free up the ref cache */
892 void btrfs_free_ref_cache(struct btrfs_fs_info *fs_info)
893 {
894 	struct block_entry *be;
895 	struct rb_node *n;
896 
897 	if (!btrfs_test_opt(fs_info, REF_VERIFY))
898 		return;
899 
900 	spin_lock(&fs_info->ref_verify_lock);
901 	while ((n = rb_first(&fs_info->block_tree))) {
902 		be = rb_entry(n, struct block_entry, node);
903 		rb_erase(&be->node, &fs_info->block_tree);
904 		free_block_entry(be);
905 		cond_resched_lock(&fs_info->ref_verify_lock);
906 	}
907 	spin_unlock(&fs_info->ref_verify_lock);
908 }
909 
910 void btrfs_free_ref_tree_range(struct btrfs_fs_info *fs_info, u64 start,
911 			       u64 len)
912 {
913 	struct block_entry *be = NULL, *entry;
914 	struct rb_node *n;
915 
916 	if (!btrfs_test_opt(fs_info, REF_VERIFY))
917 		return;
918 
919 	spin_lock(&fs_info->ref_verify_lock);
920 	n = fs_info->block_tree.rb_node;
921 	while (n) {
922 		entry = rb_entry(n, struct block_entry, node);
923 		if (entry->bytenr < start) {
924 			n = n->rb_right;
925 		} else if (entry->bytenr > start) {
926 			n = n->rb_left;
927 		} else {
928 			be = entry;
929 			break;
930 		}
931 		/* We want to get as close to start as possible */
932 		if (be == NULL ||
933 		    (entry->bytenr < start && be->bytenr > start) ||
934 		    (entry->bytenr < start && entry->bytenr > be->bytenr))
935 			be = entry;
936 	}
937 
938 	/*
939 	 * Could have an empty block group, maybe have something to check for
940 	 * this case to verify we were actually empty?
941 	 */
942 	if (!be) {
943 		spin_unlock(&fs_info->ref_verify_lock);
944 		return;
945 	}
946 
947 	n = &be->node;
948 	while (n) {
949 		be = rb_entry(n, struct block_entry, node);
950 		n = rb_next(n);
951 		if (be->bytenr < start && be->bytenr + be->len > start) {
952 			btrfs_err(fs_info,
953 				"block entry overlaps a block group [%llu,%llu]!",
954 				start, len);
955 			dump_block_entry(fs_info, be);
956 			continue;
957 		}
958 		if (be->bytenr < start)
959 			continue;
960 		if (be->bytenr >= start + len)
961 			break;
962 		if (be->bytenr + be->len > start + len) {
963 			btrfs_err(fs_info,
964 				"block entry overlaps a block group [%llu,%llu]!",
965 				start, len);
966 			dump_block_entry(fs_info, be);
967 		}
968 		rb_erase(&be->node, &fs_info->block_tree);
969 		free_block_entry(be);
970 	}
971 	spin_unlock(&fs_info->ref_verify_lock);
972 }
973 
974 /* Walk down all roots and build the ref tree, meant to be called at mount */
975 int btrfs_build_ref_tree(struct btrfs_fs_info *fs_info)
976 {
977 	struct btrfs_path *path;
978 	struct extent_buffer *eb;
979 	u64 bytenr = 0, num_bytes = 0;
980 	int ret, level;
981 
982 	if (!btrfs_test_opt(fs_info, REF_VERIFY))
983 		return 0;
984 
985 	path = btrfs_alloc_path();
986 	if (!path)
987 		return -ENOMEM;
988 
989 	eb = btrfs_read_lock_root_node(fs_info->extent_root);
990 	btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
991 	level = btrfs_header_level(eb);
992 	path->nodes[level] = eb;
993 	path->slots[level] = 0;
994 	path->locks[level] = BTRFS_READ_LOCK_BLOCKING;
995 
996 	while (1) {
997 		/*
998 		 * We have to keep track of the bytenr/num_bytes we last hit
999 		 * because we could have run out of space for an inline ref, and
1000 		 * would have had to added a ref key item which may appear on a
1001 		 * different leaf from the original extent item.
1002 		 */
1003 		ret = walk_down_tree(fs_info->extent_root, path, level,
1004 				     &bytenr, &num_bytes);
1005 		if (ret)
1006 			break;
1007 		ret = walk_up_tree(path, &level);
1008 		if (ret < 0)
1009 			break;
1010 		if (ret > 0) {
1011 			ret = 0;
1012 			break;
1013 		}
1014 	}
1015 	if (ret) {
1016 		btrfs_clear_opt(fs_info->mount_opt, REF_VERIFY);
1017 		btrfs_free_ref_cache(fs_info);
1018 	}
1019 	btrfs_free_path(path);
1020 	return ret;
1021 }
1022