xref: /linux/fs/btrfs/uuid-tree.c (revision 7ec462100ef9142344ddbf86f2c3008b97acddbe)
1 // SPDX-License-Identifier: GPL-2.0
2 /*
3  * Copyright (C) STRATO AG 2013.  All rights reserved.
4  */
5 
6 #include <linux/kthread.h>
7 #include <linux/uuid.h>
8 #include <linux/unaligned.h>
9 #include "messages.h"
10 #include "ctree.h"
11 #include "transaction.h"
12 #include "disk-io.h"
13 #include "fs.h"
14 #include "accessors.h"
15 #include "uuid-tree.h"
16 #include "ioctl.h"
17 
btrfs_uuid_to_key(const u8 * uuid,u8 type,struct btrfs_key * key)18 static void btrfs_uuid_to_key(const u8 *uuid, u8 type, struct btrfs_key *key)
19 {
20 	key->type = type;
21 	key->objectid = get_unaligned_le64(uuid);
22 	key->offset = get_unaligned_le64(uuid + sizeof(u64));
23 }
24 
25 /* return -ENOENT for !found, < 0 for errors, or 0 if an item was found */
btrfs_uuid_tree_lookup(struct btrfs_root * uuid_root,const u8 * uuid,u8 type,u64 subid)26 static int btrfs_uuid_tree_lookup(struct btrfs_root *uuid_root, const u8 *uuid,
27 				  u8 type, u64 subid)
28 {
29 	int ret;
30 	struct btrfs_path *path = NULL;
31 	struct extent_buffer *eb;
32 	int slot;
33 	u32 item_size;
34 	unsigned long offset;
35 	struct btrfs_key key;
36 
37 	if (WARN_ON_ONCE(!uuid_root)) {
38 		ret = -ENOENT;
39 		goto out;
40 	}
41 
42 	path = btrfs_alloc_path();
43 	if (!path) {
44 		ret = -ENOMEM;
45 		goto out;
46 	}
47 
48 	btrfs_uuid_to_key(uuid, type, &key);
49 	ret = btrfs_search_slot(NULL, uuid_root, &key, path, 0, 0);
50 	if (ret < 0) {
51 		goto out;
52 	} else if (ret > 0) {
53 		ret = -ENOENT;
54 		goto out;
55 	}
56 
57 	eb = path->nodes[0];
58 	slot = path->slots[0];
59 	item_size = btrfs_item_size(eb, slot);
60 	offset = btrfs_item_ptr_offset(eb, slot);
61 	ret = -ENOENT;
62 
63 	if (!IS_ALIGNED(item_size, sizeof(u64))) {
64 		btrfs_warn(uuid_root->fs_info,
65 			   "uuid item with illegal size %lu!",
66 			   (unsigned long)item_size);
67 		goto out;
68 	}
69 	while (item_size) {
70 		__le64 data;
71 
72 		read_extent_buffer(eb, &data, offset, sizeof(data));
73 		if (le64_to_cpu(data) == subid) {
74 			ret = 0;
75 			break;
76 		}
77 		offset += sizeof(data);
78 		item_size -= sizeof(data);
79 	}
80 
81 out:
82 	btrfs_free_path(path);
83 	return ret;
84 }
85 
btrfs_uuid_tree_add(struct btrfs_trans_handle * trans,const u8 * uuid,u8 type,u64 subid_cpu)86 int btrfs_uuid_tree_add(struct btrfs_trans_handle *trans, const u8 *uuid, u8 type,
87 			u64 subid_cpu)
88 {
89 	struct btrfs_fs_info *fs_info = trans->fs_info;
90 	struct btrfs_root *uuid_root = fs_info->uuid_root;
91 	int ret;
92 	struct btrfs_path *path = NULL;
93 	struct btrfs_key key;
94 	struct extent_buffer *eb;
95 	int slot;
96 	unsigned long offset;
97 	__le64 subid_le;
98 
99 	ret = btrfs_uuid_tree_lookup(uuid_root, uuid, type, subid_cpu);
100 	if (ret != -ENOENT)
101 		return ret;
102 
103 	if (WARN_ON_ONCE(!uuid_root)) {
104 		ret = -EINVAL;
105 		goto out;
106 	}
107 
108 	btrfs_uuid_to_key(uuid, type, &key);
109 
110 	path = btrfs_alloc_path();
111 	if (!path) {
112 		ret = -ENOMEM;
113 		goto out;
114 	}
115 
116 	ret = btrfs_insert_empty_item(trans, uuid_root, path, &key,
117 				      sizeof(subid_le));
118 	if (ret == 0) {
119 		/* Add an item for the type for the first time */
120 		eb = path->nodes[0];
121 		slot = path->slots[0];
122 		offset = btrfs_item_ptr_offset(eb, slot);
123 	} else if (ret == -EEXIST) {
124 		/*
125 		 * An item with that type already exists.
126 		 * Extend the item and store the new subid at the end.
127 		 */
128 		btrfs_extend_item(trans, path, sizeof(subid_le));
129 		eb = path->nodes[0];
130 		slot = path->slots[0];
131 		offset = btrfs_item_ptr_offset(eb, slot);
132 		offset += btrfs_item_size(eb, slot) - sizeof(subid_le);
133 	} else {
134 		btrfs_warn(fs_info,
135 			   "insert uuid item failed %d (0x%016llx, 0x%016llx) type %u!",
136 			   ret, key.objectid, key.offset, type);
137 		goto out;
138 	}
139 
140 	ret = 0;
141 	subid_le = cpu_to_le64(subid_cpu);
142 	write_extent_buffer(eb, &subid_le, offset, sizeof(subid_le));
143 	btrfs_mark_buffer_dirty(trans, eb);
144 
145 out:
146 	btrfs_free_path(path);
147 	return ret;
148 }
149 
btrfs_uuid_tree_remove(struct btrfs_trans_handle * trans,const u8 * uuid,u8 type,u64 subid)150 int btrfs_uuid_tree_remove(struct btrfs_trans_handle *trans, const u8 *uuid, u8 type,
151 			u64 subid)
152 {
153 	struct btrfs_fs_info *fs_info = trans->fs_info;
154 	struct btrfs_root *uuid_root = fs_info->uuid_root;
155 	int ret;
156 	struct btrfs_path *path = NULL;
157 	struct btrfs_key key;
158 	struct extent_buffer *eb;
159 	int slot;
160 	unsigned long offset;
161 	u32 item_size;
162 	unsigned long move_dst;
163 	unsigned long move_src;
164 	unsigned long move_len;
165 
166 	if (WARN_ON_ONCE(!uuid_root)) {
167 		ret = -EINVAL;
168 		goto out;
169 	}
170 
171 	btrfs_uuid_to_key(uuid, type, &key);
172 
173 	path = btrfs_alloc_path();
174 	if (!path) {
175 		ret = -ENOMEM;
176 		goto out;
177 	}
178 
179 	ret = btrfs_search_slot(trans, uuid_root, &key, path, -1, 1);
180 	if (ret < 0) {
181 		btrfs_warn(fs_info, "error %d while searching for uuid item!",
182 			   ret);
183 		goto out;
184 	}
185 	if (ret > 0) {
186 		ret = -ENOENT;
187 		goto out;
188 	}
189 
190 	eb = path->nodes[0];
191 	slot = path->slots[0];
192 	offset = btrfs_item_ptr_offset(eb, slot);
193 	item_size = btrfs_item_size(eb, slot);
194 	if (!IS_ALIGNED(item_size, sizeof(u64))) {
195 		btrfs_warn(fs_info, "uuid item with illegal size %lu!",
196 			   (unsigned long)item_size);
197 		ret = -ENOENT;
198 		goto out;
199 	}
200 	while (item_size) {
201 		__le64 read_subid;
202 
203 		read_extent_buffer(eb, &read_subid, offset, sizeof(read_subid));
204 		if (le64_to_cpu(read_subid) == subid)
205 			break;
206 		offset += sizeof(read_subid);
207 		item_size -= sizeof(read_subid);
208 	}
209 
210 	if (!item_size) {
211 		ret = -ENOENT;
212 		goto out;
213 	}
214 
215 	item_size = btrfs_item_size(eb, slot);
216 	if (item_size == sizeof(subid)) {
217 		ret = btrfs_del_item(trans, uuid_root, path);
218 		goto out;
219 	}
220 
221 	move_dst = offset;
222 	move_src = offset + sizeof(subid);
223 	move_len = item_size - (move_src - btrfs_item_ptr_offset(eb, slot));
224 	memmove_extent_buffer(eb, move_dst, move_src, move_len);
225 	btrfs_truncate_item(trans, path, item_size - sizeof(subid), 1);
226 
227 out:
228 	btrfs_free_path(path);
229 	return ret;
230 }
231 
btrfs_uuid_iter_rem(struct btrfs_root * uuid_root,u8 * uuid,u8 type,u64 subid)232 static int btrfs_uuid_iter_rem(struct btrfs_root *uuid_root, u8 *uuid, u8 type,
233 			       u64 subid)
234 {
235 	struct btrfs_trans_handle *trans;
236 	int ret;
237 
238 	/* 1 - for the uuid item */
239 	trans = btrfs_start_transaction(uuid_root, 1);
240 	if (IS_ERR(trans)) {
241 		ret = PTR_ERR(trans);
242 		goto out;
243 	}
244 
245 	ret = btrfs_uuid_tree_remove(trans, uuid, type, subid);
246 	btrfs_end_transaction(trans);
247 
248 out:
249 	return ret;
250 }
251 
252 /*
253  * Check if there's an matching subvolume for given UUID
254  *
255  * Return:
256  * 0	check succeeded, the entry is not outdated
257  * > 0	if the check failed, the caller should remove the entry
258  * < 0	if an error occurred
259  */
btrfs_check_uuid_tree_entry(struct btrfs_fs_info * fs_info,const u8 * uuid,u8 type,u64 subvolid)260 static int btrfs_check_uuid_tree_entry(struct btrfs_fs_info *fs_info,
261 				       const u8 *uuid, u8 type, u64 subvolid)
262 {
263 	int ret = 0;
264 	struct btrfs_root *subvol_root;
265 
266 	if (type != BTRFS_UUID_KEY_SUBVOL &&
267 	    type != BTRFS_UUID_KEY_RECEIVED_SUBVOL)
268 		goto out;
269 
270 	subvol_root = btrfs_get_fs_root(fs_info, subvolid, true);
271 	if (IS_ERR(subvol_root)) {
272 		ret = PTR_ERR(subvol_root);
273 		if (ret == -ENOENT)
274 			ret = 1;
275 		goto out;
276 	}
277 
278 	switch (type) {
279 	case BTRFS_UUID_KEY_SUBVOL:
280 		if (memcmp(uuid, subvol_root->root_item.uuid, BTRFS_UUID_SIZE))
281 			ret = 1;
282 		break;
283 	case BTRFS_UUID_KEY_RECEIVED_SUBVOL:
284 		if (memcmp(uuid, subvol_root->root_item.received_uuid,
285 			   BTRFS_UUID_SIZE))
286 			ret = 1;
287 		break;
288 	}
289 	btrfs_put_root(subvol_root);
290 out:
291 	return ret;
292 }
293 
btrfs_uuid_tree_iterate(struct btrfs_fs_info * fs_info)294 int btrfs_uuid_tree_iterate(struct btrfs_fs_info *fs_info)
295 {
296 	struct btrfs_root *root = fs_info->uuid_root;
297 	struct btrfs_key key;
298 	struct btrfs_path *path;
299 	int ret = 0;
300 	struct extent_buffer *leaf;
301 	int slot;
302 	u32 item_size;
303 	unsigned long offset;
304 
305 	path = btrfs_alloc_path();
306 	if (!path) {
307 		ret = -ENOMEM;
308 		goto out;
309 	}
310 
311 	key.objectid = 0;
312 	key.type = 0;
313 	key.offset = 0;
314 
315 again_search_slot:
316 	ret = btrfs_search_forward(root, &key, path, BTRFS_OLDEST_GENERATION);
317 	if (ret) {
318 		if (ret > 0)
319 			ret = 0;
320 		goto out;
321 	}
322 
323 	while (1) {
324 		if (btrfs_fs_closing(fs_info)) {
325 			ret = -EINTR;
326 			goto out;
327 		}
328 		cond_resched();
329 		leaf = path->nodes[0];
330 		slot = path->slots[0];
331 		btrfs_item_key_to_cpu(leaf, &key, slot);
332 
333 		if (key.type != BTRFS_UUID_KEY_SUBVOL &&
334 		    key.type != BTRFS_UUID_KEY_RECEIVED_SUBVOL)
335 			goto skip;
336 
337 		offset = btrfs_item_ptr_offset(leaf, slot);
338 		item_size = btrfs_item_size(leaf, slot);
339 		if (!IS_ALIGNED(item_size, sizeof(u64))) {
340 			btrfs_warn(fs_info,
341 				   "uuid item with illegal size %lu!",
342 				   (unsigned long)item_size);
343 			goto skip;
344 		}
345 		while (item_size) {
346 			u8 uuid[BTRFS_UUID_SIZE];
347 			__le64 subid_le;
348 			u64 subid_cpu;
349 
350 			put_unaligned_le64(key.objectid, uuid);
351 			put_unaligned_le64(key.offset, uuid + sizeof(u64));
352 			read_extent_buffer(leaf, &subid_le, offset,
353 					   sizeof(subid_le));
354 			subid_cpu = le64_to_cpu(subid_le);
355 			ret = btrfs_check_uuid_tree_entry(fs_info, uuid,
356 							  key.type, subid_cpu);
357 			if (ret < 0)
358 				goto out;
359 			if (ret > 0) {
360 				btrfs_release_path(path);
361 				ret = btrfs_uuid_iter_rem(root, uuid, key.type,
362 							  subid_cpu);
363 				if (ret == 0) {
364 					/*
365 					 * this might look inefficient, but the
366 					 * justification is that it is an
367 					 * exception that check_func returns 1,
368 					 * and that in the regular case only one
369 					 * entry per UUID exists.
370 					 */
371 					goto again_search_slot;
372 				}
373 				if (ret < 0 && ret != -ENOENT)
374 					goto out;
375 				key.offset++;
376 				goto again_search_slot;
377 			}
378 			item_size -= sizeof(subid_le);
379 			offset += sizeof(subid_le);
380 		}
381 
382 skip:
383 		ret = btrfs_next_item(root, path);
384 		if (ret == 0)
385 			continue;
386 		else if (ret > 0)
387 			ret = 0;
388 		break;
389 	}
390 
391 out:
392 	btrfs_free_path(path);
393 	return ret;
394 }
395 
btrfs_uuid_scan_kthread(void * data)396 int btrfs_uuid_scan_kthread(void *data)
397 {
398 	struct btrfs_fs_info *fs_info = data;
399 	struct btrfs_root *root = fs_info->tree_root;
400 	struct btrfs_key key;
401 	struct btrfs_path *path = NULL;
402 	int ret = 0;
403 	struct extent_buffer *eb;
404 	int slot;
405 	struct btrfs_root_item root_item;
406 	u32 item_size;
407 	struct btrfs_trans_handle *trans = NULL;
408 	bool closing = false;
409 
410 	path = btrfs_alloc_path();
411 	if (!path) {
412 		ret = -ENOMEM;
413 		goto out;
414 	}
415 
416 	key.objectid = 0;
417 	key.type = BTRFS_ROOT_ITEM_KEY;
418 	key.offset = 0;
419 
420 	while (1) {
421 		if (btrfs_fs_closing(fs_info)) {
422 			closing = true;
423 			break;
424 		}
425 		ret = btrfs_search_forward(root, &key, path,
426 				BTRFS_OLDEST_GENERATION);
427 		if (ret) {
428 			if (ret > 0)
429 				ret = 0;
430 			break;
431 		}
432 
433 		if (key.type != BTRFS_ROOT_ITEM_KEY ||
434 		    (key.objectid < BTRFS_FIRST_FREE_OBJECTID &&
435 		     key.objectid != BTRFS_FS_TREE_OBJECTID) ||
436 		    key.objectid > BTRFS_LAST_FREE_OBJECTID)
437 			goto skip;
438 
439 		eb = path->nodes[0];
440 		slot = path->slots[0];
441 		item_size = btrfs_item_size(eb, slot);
442 		if (item_size < sizeof(root_item))
443 			goto skip;
444 
445 		read_extent_buffer(eb, &root_item,
446 				   btrfs_item_ptr_offset(eb, slot),
447 				   (int)sizeof(root_item));
448 		if (btrfs_root_refs(&root_item) == 0)
449 			goto skip;
450 
451 		if (!btrfs_is_empty_uuid(root_item.uuid) ||
452 		    !btrfs_is_empty_uuid(root_item.received_uuid)) {
453 			if (trans)
454 				goto update_tree;
455 
456 			btrfs_release_path(path);
457 			/*
458 			 * 1 - subvol uuid item
459 			 * 1 - received_subvol uuid item
460 			 */
461 			trans = btrfs_start_transaction(fs_info->uuid_root, 2);
462 			if (IS_ERR(trans)) {
463 				ret = PTR_ERR(trans);
464 				break;
465 			}
466 			continue;
467 		} else {
468 			goto skip;
469 		}
470 update_tree:
471 		btrfs_release_path(path);
472 		if (!btrfs_is_empty_uuid(root_item.uuid)) {
473 			ret = btrfs_uuid_tree_add(trans, root_item.uuid,
474 						  BTRFS_UUID_KEY_SUBVOL,
475 						  key.objectid);
476 			if (ret < 0) {
477 				btrfs_warn(fs_info, "uuid_tree_add failed %d",
478 					ret);
479 				break;
480 			}
481 		}
482 
483 		if (!btrfs_is_empty_uuid(root_item.received_uuid)) {
484 			ret = btrfs_uuid_tree_add(trans,
485 						  root_item.received_uuid,
486 						 BTRFS_UUID_KEY_RECEIVED_SUBVOL,
487 						  key.objectid);
488 			if (ret < 0) {
489 				btrfs_warn(fs_info, "uuid_tree_add failed %d",
490 					ret);
491 				break;
492 			}
493 		}
494 
495 skip:
496 		btrfs_release_path(path);
497 		if (trans) {
498 			ret = btrfs_end_transaction(trans);
499 			trans = NULL;
500 			if (ret)
501 				break;
502 		}
503 
504 		if (key.offset < (u64)-1) {
505 			key.offset++;
506 		} else if (key.type < BTRFS_ROOT_ITEM_KEY) {
507 			key.offset = 0;
508 			key.type = BTRFS_ROOT_ITEM_KEY;
509 		} else if (key.objectid < (u64)-1) {
510 			key.offset = 0;
511 			key.type = BTRFS_ROOT_ITEM_KEY;
512 			key.objectid++;
513 		} else {
514 			break;
515 		}
516 		cond_resched();
517 	}
518 
519 out:
520 	btrfs_free_path(path);
521 	if (trans && !IS_ERR(trans))
522 		btrfs_end_transaction(trans);
523 	if (ret)
524 		btrfs_warn(fs_info, "btrfs_uuid_scan_kthread failed %d", ret);
525 	else if (!closing)
526 		set_bit(BTRFS_FS_UPDATE_UUID_TREE_GEN, &fs_info->flags);
527 	up(&fs_info->uuid_tree_rescan_sem);
528 	return 0;
529 }
530 
btrfs_create_uuid_tree(struct btrfs_fs_info * fs_info)531 int btrfs_create_uuid_tree(struct btrfs_fs_info *fs_info)
532 {
533 	struct btrfs_trans_handle *trans;
534 	struct btrfs_root *tree_root = fs_info->tree_root;
535 	struct btrfs_root *uuid_root;
536 	struct task_struct *task;
537 	int ret;
538 
539 	/*
540 	 * 1 - root node
541 	 * 1 - root item
542 	 */
543 	trans = btrfs_start_transaction(tree_root, 2);
544 	if (IS_ERR(trans))
545 		return PTR_ERR(trans);
546 
547 	uuid_root = btrfs_create_tree(trans, BTRFS_UUID_TREE_OBJECTID);
548 	if (IS_ERR(uuid_root)) {
549 		ret = PTR_ERR(uuid_root);
550 		btrfs_abort_transaction(trans, ret);
551 		btrfs_end_transaction(trans);
552 		return ret;
553 	}
554 
555 	fs_info->uuid_root = uuid_root;
556 
557 	ret = btrfs_commit_transaction(trans);
558 	if (ret)
559 		return ret;
560 
561 	down(&fs_info->uuid_tree_rescan_sem);
562 	task = kthread_run(btrfs_uuid_scan_kthread, fs_info, "btrfs-uuid");
563 	if (IS_ERR(task)) {
564 		/* fs_info->update_uuid_tree_gen remains 0 in all error case */
565 		btrfs_warn(fs_info, "failed to start uuid_scan task");
566 		up(&fs_info->uuid_tree_rescan_sem);
567 		return PTR_ERR(task);
568 	}
569 
570 	return 0;
571 }
572