xref: /linux/fs/ubifs/orphan.c (revision ebf68996de0ab250c5d520eb2291ab65643e9a1e)
1 // SPDX-License-Identifier: GPL-2.0-only
2 /*
3  * This file is part of UBIFS.
4  *
5  * Copyright (C) 2006-2008 Nokia Corporation.
6  *
7  * Author: Adrian Hunter
8  */
9 
10 #include "ubifs.h"
11 
12 /*
13  * An orphan is an inode number whose inode node has been committed to the index
14  * with a link count of zero. That happens when an open file is deleted
15  * (unlinked) and then a commit is run. In the normal course of events the inode
16  * would be deleted when the file is closed. However in the case of an unclean
17  * unmount, orphans need to be accounted for. After an unclean unmount, the
18  * orphans' inodes must be deleted which means either scanning the entire index
19  * looking for them, or keeping a list on flash somewhere. This unit implements
20  * the latter approach.
21  *
22  * The orphan area is a fixed number of LEBs situated between the LPT area and
23  * the main area. The number of orphan area LEBs is specified when the file
24  * system is created. The minimum number is 1. The size of the orphan area
25  * should be so that it can hold the maximum number of orphans that are expected
26  * to ever exist at one time.
27  *
28  * The number of orphans that can fit in a LEB is:
29  *
30  *         (c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64)
31  *
32  * For example: a 15872 byte LEB can fit 1980 orphans so 1 LEB may be enough.
33  *
34  * Orphans are accumulated in a rb-tree. When an inode's link count drops to
35  * zero, the inode number is added to the rb-tree. It is removed from the tree
36  * when the inode is deleted.  Any new orphans that are in the orphan tree when
37  * the commit is run, are written to the orphan area in 1 or more orphan nodes.
38  * If the orphan area is full, it is consolidated to make space.  There is
39  * always enough space because validation prevents the user from creating more
40  * than the maximum number of orphans allowed.
41  */
42 
43 static int dbg_check_orphans(struct ubifs_info *c);
44 
45 static struct ubifs_orphan *orphan_add(struct ubifs_info *c, ino_t inum,
46 				       struct ubifs_orphan *parent_orphan)
47 {
48 	struct ubifs_orphan *orphan, *o;
49 	struct rb_node **p, *parent = NULL;
50 
51 	orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_NOFS);
52 	if (!orphan)
53 		return ERR_PTR(-ENOMEM);
54 	orphan->inum = inum;
55 	orphan->new = 1;
56 	INIT_LIST_HEAD(&orphan->child_list);
57 
58 	spin_lock(&c->orphan_lock);
59 	if (c->tot_orphans >= c->max_orphans) {
60 		spin_unlock(&c->orphan_lock);
61 		kfree(orphan);
62 		return ERR_PTR(-ENFILE);
63 	}
64 	p = &c->orph_tree.rb_node;
65 	while (*p) {
66 		parent = *p;
67 		o = rb_entry(parent, struct ubifs_orphan, rb);
68 		if (inum < o->inum)
69 			p = &(*p)->rb_left;
70 		else if (inum > o->inum)
71 			p = &(*p)->rb_right;
72 		else {
73 			ubifs_err(c, "orphaned twice");
74 			spin_unlock(&c->orphan_lock);
75 			kfree(orphan);
76 			return ERR_PTR(-EINVAL);
77 		}
78 	}
79 	c->tot_orphans += 1;
80 	c->new_orphans += 1;
81 	rb_link_node(&orphan->rb, parent, p);
82 	rb_insert_color(&orphan->rb, &c->orph_tree);
83 	list_add_tail(&orphan->list, &c->orph_list);
84 	list_add_tail(&orphan->new_list, &c->orph_new);
85 
86 	if (parent_orphan) {
87 		list_add_tail(&orphan->child_list,
88 			      &parent_orphan->child_list);
89 	}
90 
91 	spin_unlock(&c->orphan_lock);
92 	dbg_gen("ino %lu", (unsigned long)inum);
93 	return orphan;
94 }
95 
96 static struct ubifs_orphan *lookup_orphan(struct ubifs_info *c, ino_t inum)
97 {
98 	struct ubifs_orphan *o;
99 	struct rb_node *p;
100 
101 	p = c->orph_tree.rb_node;
102 	while (p) {
103 		o = rb_entry(p, struct ubifs_orphan, rb);
104 		if (inum < o->inum)
105 			p = p->rb_left;
106 		else if (inum > o->inum)
107 			p = p->rb_right;
108 		else {
109 			return o;
110 		}
111 	}
112 	return NULL;
113 }
114 
115 static void __orphan_drop(struct ubifs_info *c, struct ubifs_orphan *o)
116 {
117 	rb_erase(&o->rb, &c->orph_tree);
118 	list_del(&o->list);
119 	c->tot_orphans -= 1;
120 
121 	if (o->new) {
122 		list_del(&o->new_list);
123 		c->new_orphans -= 1;
124 	}
125 
126 	kfree(o);
127 }
128 
129 static void orphan_delete(struct ubifs_info *c, ino_t inum)
130 {
131 	struct ubifs_orphan *orph, *child_orph, *tmp_o;
132 
133 	spin_lock(&c->orphan_lock);
134 
135 	orph = lookup_orphan(c, inum);
136 	if (!orph) {
137 		spin_unlock(&c->orphan_lock);
138 		ubifs_err(c, "missing orphan ino %lu", (unsigned long)inum);
139 		dump_stack();
140 
141 		return;
142 	}
143 
144 	if (orph->del) {
145 		spin_unlock(&c->orphan_lock);
146 		dbg_gen("deleted twice ino %lu",
147 			(unsigned long)inum);
148 		return;
149 	}
150 
151 	if (orph->cmt) {
152 		orph->del = 1;
153 		orph->dnext = c->orph_dnext;
154 		c->orph_dnext = orph;
155 		spin_unlock(&c->orphan_lock);
156 		dbg_gen("delete later ino %lu",
157 			(unsigned long)inum);
158 		return;
159 	}
160 
161 	list_for_each_entry_safe(child_orph, tmp_o, &orph->child_list, child_list) {
162 		list_del(&child_orph->child_list);
163 		__orphan_drop(c, child_orph);
164 	}
165 
166 	__orphan_drop(c, orph);
167 
168 	spin_unlock(&c->orphan_lock);
169 }
170 
171 /**
172  * ubifs_add_orphan - add an orphan.
173  * @c: UBIFS file-system description object
174  * @inum: orphan inode number
175  *
176  * Add an orphan. This function is called when an inodes link count drops to
177  * zero.
178  */
179 int ubifs_add_orphan(struct ubifs_info *c, ino_t inum)
180 {
181 	int err = 0;
182 	ino_t xattr_inum;
183 	union ubifs_key key;
184 	struct ubifs_dent_node *xent;
185 	struct fscrypt_name nm = {0};
186 	struct ubifs_orphan *xattr_orphan;
187 	struct ubifs_orphan *orphan;
188 
189 	orphan = orphan_add(c, inum, NULL);
190 	if (IS_ERR(orphan))
191 		return PTR_ERR(orphan);
192 
193 	lowest_xent_key(c, &key, inum);
194 	while (1) {
195 		xent = ubifs_tnc_next_ent(c, &key, &nm);
196 		if (IS_ERR(xent)) {
197 			err = PTR_ERR(xent);
198 			if (err == -ENOENT)
199 				break;
200 			return err;
201 		}
202 
203 		fname_name(&nm) = xent->name;
204 		fname_len(&nm) = le16_to_cpu(xent->nlen);
205 		xattr_inum = le64_to_cpu(xent->inum);
206 
207 		xattr_orphan = orphan_add(c, xattr_inum, orphan);
208 		if (IS_ERR(xattr_orphan))
209 			return PTR_ERR(xattr_orphan);
210 
211 		key_read(c, &xent->key, &key);
212 	}
213 
214 	return 0;
215 }
216 
217 /**
218  * ubifs_delete_orphan - delete an orphan.
219  * @c: UBIFS file-system description object
220  * @inum: orphan inode number
221  *
222  * Delete an orphan. This function is called when an inode is deleted.
223  */
224 void ubifs_delete_orphan(struct ubifs_info *c, ino_t inum)
225 {
226 	orphan_delete(c, inum);
227 }
228 
229 /**
230  * ubifs_orphan_start_commit - start commit of orphans.
231  * @c: UBIFS file-system description object
232  *
233  * Start commit of orphans.
234  */
235 int ubifs_orphan_start_commit(struct ubifs_info *c)
236 {
237 	struct ubifs_orphan *orphan, **last;
238 
239 	spin_lock(&c->orphan_lock);
240 	last = &c->orph_cnext;
241 	list_for_each_entry(orphan, &c->orph_new, new_list) {
242 		ubifs_assert(c, orphan->new);
243 		ubifs_assert(c, !orphan->cmt);
244 		orphan->new = 0;
245 		orphan->cmt = 1;
246 		*last = orphan;
247 		last = &orphan->cnext;
248 	}
249 	*last = NULL;
250 	c->cmt_orphans = c->new_orphans;
251 	c->new_orphans = 0;
252 	dbg_cmt("%d orphans to commit", c->cmt_orphans);
253 	INIT_LIST_HEAD(&c->orph_new);
254 	if (c->tot_orphans == 0)
255 		c->no_orphs = 1;
256 	else
257 		c->no_orphs = 0;
258 	spin_unlock(&c->orphan_lock);
259 	return 0;
260 }
261 
262 /**
263  * avail_orphs - calculate available space.
264  * @c: UBIFS file-system description object
265  *
266  * This function returns the number of orphans that can be written in the
267  * available space.
268  */
269 static int avail_orphs(struct ubifs_info *c)
270 {
271 	int avail_lebs, avail, gap;
272 
273 	avail_lebs = c->orph_lebs - (c->ohead_lnum - c->orph_first) - 1;
274 	avail = avail_lebs *
275 	       ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
276 	gap = c->leb_size - c->ohead_offs;
277 	if (gap >= UBIFS_ORPH_NODE_SZ + sizeof(__le64))
278 		avail += (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
279 	return avail;
280 }
281 
282 /**
283  * tot_avail_orphs - calculate total space.
284  * @c: UBIFS file-system description object
285  *
286  * This function returns the number of orphans that can be written in half
287  * the total space. That leaves half the space for adding new orphans.
288  */
289 static int tot_avail_orphs(struct ubifs_info *c)
290 {
291 	int avail_lebs, avail;
292 
293 	avail_lebs = c->orph_lebs;
294 	avail = avail_lebs *
295 	       ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
296 	return avail / 2;
297 }
298 
299 /**
300  * do_write_orph_node - write a node to the orphan head.
301  * @c: UBIFS file-system description object
302  * @len: length of node
303  * @atomic: write atomically
304  *
305  * This function writes a node to the orphan head from the orphan buffer. If
306  * %atomic is not zero, then the write is done atomically. On success, %0 is
307  * returned, otherwise a negative error code is returned.
308  */
309 static int do_write_orph_node(struct ubifs_info *c, int len, int atomic)
310 {
311 	int err = 0;
312 
313 	if (atomic) {
314 		ubifs_assert(c, c->ohead_offs == 0);
315 		ubifs_prepare_node(c, c->orph_buf, len, 1);
316 		len = ALIGN(len, c->min_io_size);
317 		err = ubifs_leb_change(c, c->ohead_lnum, c->orph_buf, len);
318 	} else {
319 		if (c->ohead_offs == 0) {
320 			/* Ensure LEB has been unmapped */
321 			err = ubifs_leb_unmap(c, c->ohead_lnum);
322 			if (err)
323 				return err;
324 		}
325 		err = ubifs_write_node(c, c->orph_buf, len, c->ohead_lnum,
326 				       c->ohead_offs);
327 	}
328 	return err;
329 }
330 
331 /**
332  * write_orph_node - write an orphan node.
333  * @c: UBIFS file-system description object
334  * @atomic: write atomically
335  *
336  * This function builds an orphan node from the cnext list and writes it to the
337  * orphan head. On success, %0 is returned, otherwise a negative error code
338  * is returned.
339  */
340 static int write_orph_node(struct ubifs_info *c, int atomic)
341 {
342 	struct ubifs_orphan *orphan, *cnext;
343 	struct ubifs_orph_node *orph;
344 	int gap, err, len, cnt, i;
345 
346 	ubifs_assert(c, c->cmt_orphans > 0);
347 	gap = c->leb_size - c->ohead_offs;
348 	if (gap < UBIFS_ORPH_NODE_SZ + sizeof(__le64)) {
349 		c->ohead_lnum += 1;
350 		c->ohead_offs = 0;
351 		gap = c->leb_size;
352 		if (c->ohead_lnum > c->orph_last) {
353 			/*
354 			 * We limit the number of orphans so that this should
355 			 * never happen.
356 			 */
357 			ubifs_err(c, "out of space in orphan area");
358 			return -EINVAL;
359 		}
360 	}
361 	cnt = (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
362 	if (cnt > c->cmt_orphans)
363 		cnt = c->cmt_orphans;
364 	len = UBIFS_ORPH_NODE_SZ + cnt * sizeof(__le64);
365 	ubifs_assert(c, c->orph_buf);
366 	orph = c->orph_buf;
367 	orph->ch.node_type = UBIFS_ORPH_NODE;
368 	spin_lock(&c->orphan_lock);
369 	cnext = c->orph_cnext;
370 	for (i = 0; i < cnt; i++) {
371 		orphan = cnext;
372 		ubifs_assert(c, orphan->cmt);
373 		orph->inos[i] = cpu_to_le64(orphan->inum);
374 		orphan->cmt = 0;
375 		cnext = orphan->cnext;
376 		orphan->cnext = NULL;
377 	}
378 	c->orph_cnext = cnext;
379 	c->cmt_orphans -= cnt;
380 	spin_unlock(&c->orphan_lock);
381 	if (c->cmt_orphans)
382 		orph->cmt_no = cpu_to_le64(c->cmt_no);
383 	else
384 		/* Mark the last node of the commit */
385 		orph->cmt_no = cpu_to_le64((c->cmt_no) | (1ULL << 63));
386 	ubifs_assert(c, c->ohead_offs + len <= c->leb_size);
387 	ubifs_assert(c, c->ohead_lnum >= c->orph_first);
388 	ubifs_assert(c, c->ohead_lnum <= c->orph_last);
389 	err = do_write_orph_node(c, len, atomic);
390 	c->ohead_offs += ALIGN(len, c->min_io_size);
391 	c->ohead_offs = ALIGN(c->ohead_offs, 8);
392 	return err;
393 }
394 
395 /**
396  * write_orph_nodes - write orphan nodes until there are no more to commit.
397  * @c: UBIFS file-system description object
398  * @atomic: write atomically
399  *
400  * This function writes orphan nodes for all the orphans to commit. On success,
401  * %0 is returned, otherwise a negative error code is returned.
402  */
403 static int write_orph_nodes(struct ubifs_info *c, int atomic)
404 {
405 	int err;
406 
407 	while (c->cmt_orphans > 0) {
408 		err = write_orph_node(c, atomic);
409 		if (err)
410 			return err;
411 	}
412 	if (atomic) {
413 		int lnum;
414 
415 		/* Unmap any unused LEBs after consolidation */
416 		for (lnum = c->ohead_lnum + 1; lnum <= c->orph_last; lnum++) {
417 			err = ubifs_leb_unmap(c, lnum);
418 			if (err)
419 				return err;
420 		}
421 	}
422 	return 0;
423 }
424 
425 /**
426  * consolidate - consolidate the orphan area.
427  * @c: UBIFS file-system description object
428  *
429  * This function enables consolidation by putting all the orphans into the list
430  * to commit. The list is in the order that the orphans were added, and the
431  * LEBs are written atomically in order, so at no time can orphans be lost by
432  * an unclean unmount.
433  *
434  * This function returns %0 on success and a negative error code on failure.
435  */
436 static int consolidate(struct ubifs_info *c)
437 {
438 	int tot_avail = tot_avail_orphs(c), err = 0;
439 
440 	spin_lock(&c->orphan_lock);
441 	dbg_cmt("there is space for %d orphans and there are %d",
442 		tot_avail, c->tot_orphans);
443 	if (c->tot_orphans - c->new_orphans <= tot_avail) {
444 		struct ubifs_orphan *orphan, **last;
445 		int cnt = 0;
446 
447 		/* Change the cnext list to include all non-new orphans */
448 		last = &c->orph_cnext;
449 		list_for_each_entry(orphan, &c->orph_list, list) {
450 			if (orphan->new)
451 				continue;
452 			orphan->cmt = 1;
453 			*last = orphan;
454 			last = &orphan->cnext;
455 			cnt += 1;
456 		}
457 		*last = NULL;
458 		ubifs_assert(c, cnt == c->tot_orphans - c->new_orphans);
459 		c->cmt_orphans = cnt;
460 		c->ohead_lnum = c->orph_first;
461 		c->ohead_offs = 0;
462 	} else {
463 		/*
464 		 * We limit the number of orphans so that this should
465 		 * never happen.
466 		 */
467 		ubifs_err(c, "out of space in orphan area");
468 		err = -EINVAL;
469 	}
470 	spin_unlock(&c->orphan_lock);
471 	return err;
472 }
473 
474 /**
475  * commit_orphans - commit orphans.
476  * @c: UBIFS file-system description object
477  *
478  * This function commits orphans to flash. On success, %0 is returned,
479  * otherwise a negative error code is returned.
480  */
481 static int commit_orphans(struct ubifs_info *c)
482 {
483 	int avail, atomic = 0, err;
484 
485 	ubifs_assert(c, c->cmt_orphans > 0);
486 	avail = avail_orphs(c);
487 	if (avail < c->cmt_orphans) {
488 		/* Not enough space to write new orphans, so consolidate */
489 		err = consolidate(c);
490 		if (err)
491 			return err;
492 		atomic = 1;
493 	}
494 	err = write_orph_nodes(c, atomic);
495 	return err;
496 }
497 
498 /**
499  * erase_deleted - erase the orphans marked for deletion.
500  * @c: UBIFS file-system description object
501  *
502  * During commit, the orphans being committed cannot be deleted, so they are
503  * marked for deletion and deleted by this function. Also, the recovery
504  * adds killed orphans to the deletion list, and therefore they are deleted
505  * here too.
506  */
507 static void erase_deleted(struct ubifs_info *c)
508 {
509 	struct ubifs_orphan *orphan, *dnext;
510 
511 	spin_lock(&c->orphan_lock);
512 	dnext = c->orph_dnext;
513 	while (dnext) {
514 		orphan = dnext;
515 		dnext = orphan->dnext;
516 		ubifs_assert(c, !orphan->new);
517 		ubifs_assert(c, orphan->del);
518 		rb_erase(&orphan->rb, &c->orph_tree);
519 		list_del(&orphan->list);
520 		c->tot_orphans -= 1;
521 		dbg_gen("deleting orphan ino %lu", (unsigned long)orphan->inum);
522 		kfree(orphan);
523 	}
524 	c->orph_dnext = NULL;
525 	spin_unlock(&c->orphan_lock);
526 }
527 
528 /**
529  * ubifs_orphan_end_commit - end commit of orphans.
530  * @c: UBIFS file-system description object
531  *
532  * End commit of orphans.
533  */
534 int ubifs_orphan_end_commit(struct ubifs_info *c)
535 {
536 	int err;
537 
538 	if (c->cmt_orphans != 0) {
539 		err = commit_orphans(c);
540 		if (err)
541 			return err;
542 	}
543 	erase_deleted(c);
544 	err = dbg_check_orphans(c);
545 	return err;
546 }
547 
548 /**
549  * ubifs_clear_orphans - erase all LEBs used for orphans.
550  * @c: UBIFS file-system description object
551  *
552  * If recovery is not required, then the orphans from the previous session
553  * are not needed. This function locates the LEBs used to record
554  * orphans, and un-maps them.
555  */
556 int ubifs_clear_orphans(struct ubifs_info *c)
557 {
558 	int lnum, err;
559 
560 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
561 		err = ubifs_leb_unmap(c, lnum);
562 		if (err)
563 			return err;
564 	}
565 	c->ohead_lnum = c->orph_first;
566 	c->ohead_offs = 0;
567 	return 0;
568 }
569 
570 /**
571  * insert_dead_orphan - insert an orphan.
572  * @c: UBIFS file-system description object
573  * @inum: orphan inode number
574  *
575  * This function is a helper to the 'do_kill_orphans()' function. The orphan
576  * must be kept until the next commit, so it is added to the rb-tree and the
577  * deletion list.
578  */
579 static int insert_dead_orphan(struct ubifs_info *c, ino_t inum)
580 {
581 	struct ubifs_orphan *orphan, *o;
582 	struct rb_node **p, *parent = NULL;
583 
584 	orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_KERNEL);
585 	if (!orphan)
586 		return -ENOMEM;
587 	orphan->inum = inum;
588 
589 	p = &c->orph_tree.rb_node;
590 	while (*p) {
591 		parent = *p;
592 		o = rb_entry(parent, struct ubifs_orphan, rb);
593 		if (inum < o->inum)
594 			p = &(*p)->rb_left;
595 		else if (inum > o->inum)
596 			p = &(*p)->rb_right;
597 		else {
598 			/* Already added - no problem */
599 			kfree(orphan);
600 			return 0;
601 		}
602 	}
603 	c->tot_orphans += 1;
604 	rb_link_node(&orphan->rb, parent, p);
605 	rb_insert_color(&orphan->rb, &c->orph_tree);
606 	list_add_tail(&orphan->list, &c->orph_list);
607 	orphan->del = 1;
608 	orphan->dnext = c->orph_dnext;
609 	c->orph_dnext = orphan;
610 	dbg_mnt("ino %lu, new %d, tot %d", (unsigned long)inum,
611 		c->new_orphans, c->tot_orphans);
612 	return 0;
613 }
614 
615 /**
616  * do_kill_orphans - remove orphan inodes from the index.
617  * @c: UBIFS file-system description object
618  * @sleb: scanned LEB
619  * @last_cmt_no: cmt_no of last orphan node read is passed and returned here
620  * @outofdate: whether the LEB is out of date is returned here
621  * @last_flagged: whether the end orphan node is encountered
622  *
623  * This function is a helper to the 'kill_orphans()' function. It goes through
624  * every orphan node in a LEB and for every inode number recorded, removes
625  * all keys for that inode from the TNC.
626  */
627 static int do_kill_orphans(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
628 			   unsigned long long *last_cmt_no, int *outofdate,
629 			   int *last_flagged)
630 {
631 	struct ubifs_scan_node *snod;
632 	struct ubifs_orph_node *orph;
633 	unsigned long long cmt_no;
634 	ino_t inum;
635 	int i, n, err, first = 1;
636 
637 	list_for_each_entry(snod, &sleb->nodes, list) {
638 		if (snod->type != UBIFS_ORPH_NODE) {
639 			ubifs_err(c, "invalid node type %d in orphan area at %d:%d",
640 				  snod->type, sleb->lnum, snod->offs);
641 			ubifs_dump_node(c, snod->node);
642 			return -EINVAL;
643 		}
644 
645 		orph = snod->node;
646 
647 		/* Check commit number */
648 		cmt_no = le64_to_cpu(orph->cmt_no) & LLONG_MAX;
649 		/*
650 		 * The commit number on the master node may be less, because
651 		 * of a failed commit. If there are several failed commits in a
652 		 * row, the commit number written on orphan nodes will continue
653 		 * to increase (because the commit number is adjusted here) even
654 		 * though the commit number on the master node stays the same
655 		 * because the master node has not been re-written.
656 		 */
657 		if (cmt_no > c->cmt_no)
658 			c->cmt_no = cmt_no;
659 		if (cmt_no < *last_cmt_no && *last_flagged) {
660 			/*
661 			 * The last orphan node had a higher commit number and
662 			 * was flagged as the last written for that commit
663 			 * number. That makes this orphan node, out of date.
664 			 */
665 			if (!first) {
666 				ubifs_err(c, "out of order commit number %llu in orphan node at %d:%d",
667 					  cmt_no, sleb->lnum, snod->offs);
668 				ubifs_dump_node(c, snod->node);
669 				return -EINVAL;
670 			}
671 			dbg_rcvry("out of date LEB %d", sleb->lnum);
672 			*outofdate = 1;
673 			return 0;
674 		}
675 
676 		if (first)
677 			first = 0;
678 
679 		n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
680 		for (i = 0; i < n; i++) {
681 			union ubifs_key key1, key2;
682 
683 			inum = le64_to_cpu(orph->inos[i]);
684 			dbg_rcvry("deleting orphaned inode %lu",
685 				  (unsigned long)inum);
686 
687 			lowest_ino_key(c, &key1, inum);
688 			highest_ino_key(c, &key2, inum);
689 
690 			err = ubifs_tnc_remove_range(c, &key1, &key2);
691 			if (err)
692 				return err;
693 			err = insert_dead_orphan(c, inum);
694 			if (err)
695 				return err;
696 		}
697 
698 		*last_cmt_no = cmt_no;
699 		if (le64_to_cpu(orph->cmt_no) & (1ULL << 63)) {
700 			dbg_rcvry("last orph node for commit %llu at %d:%d",
701 				  cmt_no, sleb->lnum, snod->offs);
702 			*last_flagged = 1;
703 		} else
704 			*last_flagged = 0;
705 	}
706 
707 	return 0;
708 }
709 
710 /**
711  * kill_orphans - remove all orphan inodes from the index.
712  * @c: UBIFS file-system description object
713  *
714  * If recovery is required, then orphan inodes recorded during the previous
715  * session (which ended with an unclean unmount) must be deleted from the index.
716  * This is done by updating the TNC, but since the index is not updated until
717  * the next commit, the LEBs where the orphan information is recorded are not
718  * erased until the next commit.
719  */
720 static int kill_orphans(struct ubifs_info *c)
721 {
722 	unsigned long long last_cmt_no = 0;
723 	int lnum, err = 0, outofdate = 0, last_flagged = 0;
724 
725 	c->ohead_lnum = c->orph_first;
726 	c->ohead_offs = 0;
727 	/* Check no-orphans flag and skip this if no orphans */
728 	if (c->no_orphs) {
729 		dbg_rcvry("no orphans");
730 		return 0;
731 	}
732 	/*
733 	 * Orph nodes always start at c->orph_first and are written to each
734 	 * successive LEB in turn. Generally unused LEBs will have been unmapped
735 	 * but may contain out of date orphan nodes if the unmap didn't go
736 	 * through. In addition, the last orphan node written for each commit is
737 	 * marked (top bit of orph->cmt_no is set to 1). It is possible that
738 	 * there are orphan nodes from the next commit (i.e. the commit did not
739 	 * complete successfully). In that case, no orphans will have been lost
740 	 * due to the way that orphans are written, and any orphans added will
741 	 * be valid orphans anyway and so can be deleted.
742 	 */
743 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
744 		struct ubifs_scan_leb *sleb;
745 
746 		dbg_rcvry("LEB %d", lnum);
747 		sleb = ubifs_scan(c, lnum, 0, c->sbuf, 1);
748 		if (IS_ERR(sleb)) {
749 			if (PTR_ERR(sleb) == -EUCLEAN)
750 				sleb = ubifs_recover_leb(c, lnum, 0,
751 							 c->sbuf, -1);
752 			if (IS_ERR(sleb)) {
753 				err = PTR_ERR(sleb);
754 				break;
755 			}
756 		}
757 		err = do_kill_orphans(c, sleb, &last_cmt_no, &outofdate,
758 				      &last_flagged);
759 		if (err || outofdate) {
760 			ubifs_scan_destroy(sleb);
761 			break;
762 		}
763 		if (sleb->endpt) {
764 			c->ohead_lnum = lnum;
765 			c->ohead_offs = sleb->endpt;
766 		}
767 		ubifs_scan_destroy(sleb);
768 	}
769 	return err;
770 }
771 
772 /**
773  * ubifs_mount_orphans - delete orphan inodes and erase LEBs that recorded them.
774  * @c: UBIFS file-system description object
775  * @unclean: indicates recovery from unclean unmount
776  * @read_only: indicates read only mount
777  *
778  * This function is called when mounting to erase orphans from the previous
779  * session. If UBIFS was not unmounted cleanly, then the inodes recorded as
780  * orphans are deleted.
781  */
782 int ubifs_mount_orphans(struct ubifs_info *c, int unclean, int read_only)
783 {
784 	int err = 0;
785 
786 	c->max_orphans = tot_avail_orphs(c);
787 
788 	if (!read_only) {
789 		c->orph_buf = vmalloc(c->leb_size);
790 		if (!c->orph_buf)
791 			return -ENOMEM;
792 	}
793 
794 	if (unclean)
795 		err = kill_orphans(c);
796 	else if (!read_only)
797 		err = ubifs_clear_orphans(c);
798 
799 	return err;
800 }
801 
802 /*
803  * Everything below is related to debugging.
804  */
805 
806 struct check_orphan {
807 	struct rb_node rb;
808 	ino_t inum;
809 };
810 
811 struct check_info {
812 	unsigned long last_ino;
813 	unsigned long tot_inos;
814 	unsigned long missing;
815 	unsigned long long leaf_cnt;
816 	struct ubifs_ino_node *node;
817 	struct rb_root root;
818 };
819 
820 static bool dbg_find_orphan(struct ubifs_info *c, ino_t inum)
821 {
822 	bool found = false;
823 
824 	spin_lock(&c->orphan_lock);
825 	found = !!lookup_orphan(c, inum);
826 	spin_unlock(&c->orphan_lock);
827 
828 	return found;
829 }
830 
831 static int dbg_ins_check_orphan(struct rb_root *root, ino_t inum)
832 {
833 	struct check_orphan *orphan, *o;
834 	struct rb_node **p, *parent = NULL;
835 
836 	orphan = kzalloc(sizeof(struct check_orphan), GFP_NOFS);
837 	if (!orphan)
838 		return -ENOMEM;
839 	orphan->inum = inum;
840 
841 	p = &root->rb_node;
842 	while (*p) {
843 		parent = *p;
844 		o = rb_entry(parent, struct check_orphan, rb);
845 		if (inum < o->inum)
846 			p = &(*p)->rb_left;
847 		else if (inum > o->inum)
848 			p = &(*p)->rb_right;
849 		else {
850 			kfree(orphan);
851 			return 0;
852 		}
853 	}
854 	rb_link_node(&orphan->rb, parent, p);
855 	rb_insert_color(&orphan->rb, root);
856 	return 0;
857 }
858 
859 static int dbg_find_check_orphan(struct rb_root *root, ino_t inum)
860 {
861 	struct check_orphan *o;
862 	struct rb_node *p;
863 
864 	p = root->rb_node;
865 	while (p) {
866 		o = rb_entry(p, struct check_orphan, rb);
867 		if (inum < o->inum)
868 			p = p->rb_left;
869 		else if (inum > o->inum)
870 			p = p->rb_right;
871 		else
872 			return 1;
873 	}
874 	return 0;
875 }
876 
877 static void dbg_free_check_tree(struct rb_root *root)
878 {
879 	struct check_orphan *o, *n;
880 
881 	rbtree_postorder_for_each_entry_safe(o, n, root, rb)
882 		kfree(o);
883 }
884 
885 static int dbg_orphan_check(struct ubifs_info *c, struct ubifs_zbranch *zbr,
886 			    void *priv)
887 {
888 	struct check_info *ci = priv;
889 	ino_t inum;
890 	int err;
891 
892 	inum = key_inum(c, &zbr->key);
893 	if (inum != ci->last_ino) {
894 		/* Lowest node type is the inode node, so it comes first */
895 		if (key_type(c, &zbr->key) != UBIFS_INO_KEY)
896 			ubifs_err(c, "found orphan node ino %lu, type %d",
897 				  (unsigned long)inum, key_type(c, &zbr->key));
898 		ci->last_ino = inum;
899 		ci->tot_inos += 1;
900 		err = ubifs_tnc_read_node(c, zbr, ci->node);
901 		if (err) {
902 			ubifs_err(c, "node read failed, error %d", err);
903 			return err;
904 		}
905 		if (ci->node->nlink == 0)
906 			/* Must be recorded as an orphan */
907 			if (!dbg_find_check_orphan(&ci->root, inum) &&
908 			    !dbg_find_orphan(c, inum)) {
909 				ubifs_err(c, "missing orphan, ino %lu",
910 					  (unsigned long)inum);
911 				ci->missing += 1;
912 			}
913 	}
914 	ci->leaf_cnt += 1;
915 	return 0;
916 }
917 
918 static int dbg_read_orphans(struct check_info *ci, struct ubifs_scan_leb *sleb)
919 {
920 	struct ubifs_scan_node *snod;
921 	struct ubifs_orph_node *orph;
922 	ino_t inum;
923 	int i, n, err;
924 
925 	list_for_each_entry(snod, &sleb->nodes, list) {
926 		cond_resched();
927 		if (snod->type != UBIFS_ORPH_NODE)
928 			continue;
929 		orph = snod->node;
930 		n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
931 		for (i = 0; i < n; i++) {
932 			inum = le64_to_cpu(orph->inos[i]);
933 			err = dbg_ins_check_orphan(&ci->root, inum);
934 			if (err)
935 				return err;
936 		}
937 	}
938 	return 0;
939 }
940 
941 static int dbg_scan_orphans(struct ubifs_info *c, struct check_info *ci)
942 {
943 	int lnum, err = 0;
944 	void *buf;
945 
946 	/* Check no-orphans flag and skip this if no orphans */
947 	if (c->no_orphs)
948 		return 0;
949 
950 	buf = __vmalloc(c->leb_size, GFP_NOFS, PAGE_KERNEL);
951 	if (!buf) {
952 		ubifs_err(c, "cannot allocate memory to check orphans");
953 		return 0;
954 	}
955 
956 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
957 		struct ubifs_scan_leb *sleb;
958 
959 		sleb = ubifs_scan(c, lnum, 0, buf, 0);
960 		if (IS_ERR(sleb)) {
961 			err = PTR_ERR(sleb);
962 			break;
963 		}
964 
965 		err = dbg_read_orphans(ci, sleb);
966 		ubifs_scan_destroy(sleb);
967 		if (err)
968 			break;
969 	}
970 
971 	vfree(buf);
972 	return err;
973 }
974 
975 static int dbg_check_orphans(struct ubifs_info *c)
976 {
977 	struct check_info ci;
978 	int err;
979 
980 	if (!dbg_is_chk_orph(c))
981 		return 0;
982 
983 	ci.last_ino = 0;
984 	ci.tot_inos = 0;
985 	ci.missing  = 0;
986 	ci.leaf_cnt = 0;
987 	ci.root = RB_ROOT;
988 	ci.node = kmalloc(UBIFS_MAX_INO_NODE_SZ, GFP_NOFS);
989 	if (!ci.node) {
990 		ubifs_err(c, "out of memory");
991 		return -ENOMEM;
992 	}
993 
994 	err = dbg_scan_orphans(c, &ci);
995 	if (err)
996 		goto out;
997 
998 	err = dbg_walk_index(c, &dbg_orphan_check, NULL, &ci);
999 	if (err) {
1000 		ubifs_err(c, "cannot scan TNC, error %d", err);
1001 		goto out;
1002 	}
1003 
1004 	if (ci.missing) {
1005 		ubifs_err(c, "%lu missing orphan(s)", ci.missing);
1006 		err = -EINVAL;
1007 		goto out;
1008 	}
1009 
1010 	dbg_cmt("last inode number is %lu", ci.last_ino);
1011 	dbg_cmt("total number of inodes is %lu", ci.tot_inos);
1012 	dbg_cmt("total number of leaf nodes is %llu", ci.leaf_cnt);
1013 
1014 out:
1015 	dbg_free_check_tree(&ci.root);
1016 	kfree(ci.node);
1017 	return err;
1018 }
1019