xref: /openbmc/linux/fs/ubifs/orphan.c (revision 305c8388)
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, struct ubifs_orphan *orph)
130 {
131 	if (orph->del) {
132 		dbg_gen("deleted twice ino %lu", (unsigned long)orph->inum);
133 		return;
134 	}
135 
136 	if (orph->cmt) {
137 		orph->del = 1;
138 		orph->dnext = c->orph_dnext;
139 		c->orph_dnext = orph;
140 		dbg_gen("delete later ino %lu", (unsigned long)orph->inum);
141 		return;
142 	}
143 
144 	__orphan_drop(c, orph);
145 }
146 
147 /**
148  * ubifs_add_orphan - add an orphan.
149  * @c: UBIFS file-system description object
150  * @inum: orphan inode number
151  *
152  * Add an orphan. This function is called when an inodes link count drops to
153  * zero.
154  */
155 int ubifs_add_orphan(struct ubifs_info *c, ino_t inum)
156 {
157 	int err = 0;
158 	ino_t xattr_inum;
159 	union ubifs_key key;
160 	struct ubifs_dent_node *xent, *pxent = NULL;
161 	struct fscrypt_name nm = {0};
162 	struct ubifs_orphan *xattr_orphan;
163 	struct ubifs_orphan *orphan;
164 
165 	orphan = orphan_add(c, inum, NULL);
166 	if (IS_ERR(orphan))
167 		return PTR_ERR(orphan);
168 
169 	lowest_xent_key(c, &key, inum);
170 	while (1) {
171 		xent = ubifs_tnc_next_ent(c, &key, &nm);
172 		if (IS_ERR(xent)) {
173 			err = PTR_ERR(xent);
174 			if (err == -ENOENT)
175 				break;
176 			return err;
177 		}
178 
179 		fname_name(&nm) = xent->name;
180 		fname_len(&nm) = le16_to_cpu(xent->nlen);
181 		xattr_inum = le64_to_cpu(xent->inum);
182 
183 		xattr_orphan = orphan_add(c, xattr_inum, orphan);
184 		if (IS_ERR(xattr_orphan)) {
185 			kfree(xent);
186 			return PTR_ERR(xattr_orphan);
187 		}
188 
189 		kfree(pxent);
190 		pxent = xent;
191 		key_read(c, &xent->key, &key);
192 	}
193 	kfree(pxent);
194 
195 	return 0;
196 }
197 
198 /**
199  * ubifs_delete_orphan - delete an orphan.
200  * @c: UBIFS file-system description object
201  * @inum: orphan inode number
202  *
203  * Delete an orphan. This function is called when an inode is deleted.
204  */
205 void ubifs_delete_orphan(struct ubifs_info *c, ino_t inum)
206 {
207 	struct ubifs_orphan *orph, *child_orph, *tmp_o;
208 
209 	spin_lock(&c->orphan_lock);
210 
211 	orph = lookup_orphan(c, inum);
212 	if (!orph) {
213 		spin_unlock(&c->orphan_lock);
214 		ubifs_err(c, "missing orphan ino %lu", (unsigned long)inum);
215 		dump_stack();
216 
217 		return;
218 	}
219 
220 	list_for_each_entry_safe(child_orph, tmp_o, &orph->child_list, child_list) {
221 		list_del(&child_orph->child_list);
222 		orphan_delete(c, child_orph);
223 	}
224 
225 	orphan_delete(c, orph);
226 
227 	spin_unlock(&c->orphan_lock);
228 }
229 
230 /**
231  * ubifs_orphan_start_commit - start commit of orphans.
232  * @c: UBIFS file-system description object
233  *
234  * Start commit of orphans.
235  */
236 int ubifs_orphan_start_commit(struct ubifs_info *c)
237 {
238 	struct ubifs_orphan *orphan, **last;
239 
240 	spin_lock(&c->orphan_lock);
241 	last = &c->orph_cnext;
242 	list_for_each_entry(orphan, &c->orph_new, new_list) {
243 		ubifs_assert(c, orphan->new);
244 		ubifs_assert(c, !orphan->cmt);
245 		orphan->new = 0;
246 		orphan->cmt = 1;
247 		*last = orphan;
248 		last = &orphan->cnext;
249 	}
250 	*last = NULL;
251 	c->cmt_orphans = c->new_orphans;
252 	c->new_orphans = 0;
253 	dbg_cmt("%d orphans to commit", c->cmt_orphans);
254 	INIT_LIST_HEAD(&c->orph_new);
255 	if (c->tot_orphans == 0)
256 		c->no_orphs = 1;
257 	else
258 		c->no_orphs = 0;
259 	spin_unlock(&c->orphan_lock);
260 	return 0;
261 }
262 
263 /**
264  * avail_orphs - calculate available space.
265  * @c: UBIFS file-system description object
266  *
267  * This function returns the number of orphans that can be written in the
268  * available space.
269  */
270 static int avail_orphs(struct ubifs_info *c)
271 {
272 	int avail_lebs, avail, gap;
273 
274 	avail_lebs = c->orph_lebs - (c->ohead_lnum - c->orph_first) - 1;
275 	avail = avail_lebs *
276 	       ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
277 	gap = c->leb_size - c->ohead_offs;
278 	if (gap >= UBIFS_ORPH_NODE_SZ + sizeof(__le64))
279 		avail += (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
280 	return avail;
281 }
282 
283 /**
284  * tot_avail_orphs - calculate total space.
285  * @c: UBIFS file-system description object
286  *
287  * This function returns the number of orphans that can be written in half
288  * the total space. That leaves half the space for adding new orphans.
289  */
290 static int tot_avail_orphs(struct ubifs_info *c)
291 {
292 	int avail_lebs, avail;
293 
294 	avail_lebs = c->orph_lebs;
295 	avail = avail_lebs *
296 	       ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
297 	return avail / 2;
298 }
299 
300 /**
301  * do_write_orph_node - write a node to the orphan head.
302  * @c: UBIFS file-system description object
303  * @len: length of node
304  * @atomic: write atomically
305  *
306  * This function writes a node to the orphan head from the orphan buffer. If
307  * %atomic is not zero, then the write is done atomically. On success, %0 is
308  * returned, otherwise a negative error code is returned.
309  */
310 static int do_write_orph_node(struct ubifs_info *c, int len, int atomic)
311 {
312 	int err = 0;
313 
314 	if (atomic) {
315 		ubifs_assert(c, c->ohead_offs == 0);
316 		ubifs_prepare_node(c, c->orph_buf, len, 1);
317 		len = ALIGN(len, c->min_io_size);
318 		err = ubifs_leb_change(c, c->ohead_lnum, c->orph_buf, len);
319 	} else {
320 		if (c->ohead_offs == 0) {
321 			/* Ensure LEB has been unmapped */
322 			err = ubifs_leb_unmap(c, c->ohead_lnum);
323 			if (err)
324 				return err;
325 		}
326 		err = ubifs_write_node(c, c->orph_buf, len, c->ohead_lnum,
327 				       c->ohead_offs);
328 	}
329 	return err;
330 }
331 
332 /**
333  * write_orph_node - write an orphan node.
334  * @c: UBIFS file-system description object
335  * @atomic: write atomically
336  *
337  * This function builds an orphan node from the cnext list and writes it to the
338  * orphan head. On success, %0 is returned, otherwise a negative error code
339  * is returned.
340  */
341 static int write_orph_node(struct ubifs_info *c, int atomic)
342 {
343 	struct ubifs_orphan *orphan, *cnext;
344 	struct ubifs_orph_node *orph;
345 	int gap, err, len, cnt, i;
346 
347 	ubifs_assert(c, c->cmt_orphans > 0);
348 	gap = c->leb_size - c->ohead_offs;
349 	if (gap < UBIFS_ORPH_NODE_SZ + sizeof(__le64)) {
350 		c->ohead_lnum += 1;
351 		c->ohead_offs = 0;
352 		gap = c->leb_size;
353 		if (c->ohead_lnum > c->orph_last) {
354 			/*
355 			 * We limit the number of orphans so that this should
356 			 * never happen.
357 			 */
358 			ubifs_err(c, "out of space in orphan area");
359 			return -EINVAL;
360 		}
361 	}
362 	cnt = (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
363 	if (cnt > c->cmt_orphans)
364 		cnt = c->cmt_orphans;
365 	len = UBIFS_ORPH_NODE_SZ + cnt * sizeof(__le64);
366 	ubifs_assert(c, c->orph_buf);
367 	orph = c->orph_buf;
368 	orph->ch.node_type = UBIFS_ORPH_NODE;
369 	spin_lock(&c->orphan_lock);
370 	cnext = c->orph_cnext;
371 	for (i = 0; i < cnt; i++) {
372 		orphan = cnext;
373 		ubifs_assert(c, orphan->cmt);
374 		orph->inos[i] = cpu_to_le64(orphan->inum);
375 		orphan->cmt = 0;
376 		cnext = orphan->cnext;
377 		orphan->cnext = NULL;
378 	}
379 	c->orph_cnext = cnext;
380 	c->cmt_orphans -= cnt;
381 	spin_unlock(&c->orphan_lock);
382 	if (c->cmt_orphans)
383 		orph->cmt_no = cpu_to_le64(c->cmt_no);
384 	else
385 		/* Mark the last node of the commit */
386 		orph->cmt_no = cpu_to_le64((c->cmt_no) | (1ULL << 63));
387 	ubifs_assert(c, c->ohead_offs + len <= c->leb_size);
388 	ubifs_assert(c, c->ohead_lnum >= c->orph_first);
389 	ubifs_assert(c, c->ohead_lnum <= c->orph_last);
390 	err = do_write_orph_node(c, len, atomic);
391 	c->ohead_offs += ALIGN(len, c->min_io_size);
392 	c->ohead_offs = ALIGN(c->ohead_offs, 8);
393 	return err;
394 }
395 
396 /**
397  * write_orph_nodes - write orphan nodes until there are no more to commit.
398  * @c: UBIFS file-system description object
399  * @atomic: write atomically
400  *
401  * This function writes orphan nodes for all the orphans to commit. On success,
402  * %0 is returned, otherwise a negative error code is returned.
403  */
404 static int write_orph_nodes(struct ubifs_info *c, int atomic)
405 {
406 	int err;
407 
408 	while (c->cmt_orphans > 0) {
409 		err = write_orph_node(c, atomic);
410 		if (err)
411 			return err;
412 	}
413 	if (atomic) {
414 		int lnum;
415 
416 		/* Unmap any unused LEBs after consolidation */
417 		for (lnum = c->ohead_lnum + 1; lnum <= c->orph_last; lnum++) {
418 			err = ubifs_leb_unmap(c, lnum);
419 			if (err)
420 				return err;
421 		}
422 	}
423 	return 0;
424 }
425 
426 /**
427  * consolidate - consolidate the orphan area.
428  * @c: UBIFS file-system description object
429  *
430  * This function enables consolidation by putting all the orphans into the list
431  * to commit. The list is in the order that the orphans were added, and the
432  * LEBs are written atomically in order, so at no time can orphans be lost by
433  * an unclean unmount.
434  *
435  * This function returns %0 on success and a negative error code on failure.
436  */
437 static int consolidate(struct ubifs_info *c)
438 {
439 	int tot_avail = tot_avail_orphs(c), err = 0;
440 
441 	spin_lock(&c->orphan_lock);
442 	dbg_cmt("there is space for %d orphans and there are %d",
443 		tot_avail, c->tot_orphans);
444 	if (c->tot_orphans - c->new_orphans <= tot_avail) {
445 		struct ubifs_orphan *orphan, **last;
446 		int cnt = 0;
447 
448 		/* Change the cnext list to include all non-new orphans */
449 		last = &c->orph_cnext;
450 		list_for_each_entry(orphan, &c->orph_list, list) {
451 			if (orphan->new)
452 				continue;
453 			orphan->cmt = 1;
454 			*last = orphan;
455 			last = &orphan->cnext;
456 			cnt += 1;
457 		}
458 		*last = NULL;
459 		ubifs_assert(c, cnt == c->tot_orphans - c->new_orphans);
460 		c->cmt_orphans = cnt;
461 		c->ohead_lnum = c->orph_first;
462 		c->ohead_offs = 0;
463 	} else {
464 		/*
465 		 * We limit the number of orphans so that this should
466 		 * never happen.
467 		 */
468 		ubifs_err(c, "out of space in orphan area");
469 		err = -EINVAL;
470 	}
471 	spin_unlock(&c->orphan_lock);
472 	return err;
473 }
474 
475 /**
476  * commit_orphans - commit orphans.
477  * @c: UBIFS file-system description object
478  *
479  * This function commits orphans to flash. On success, %0 is returned,
480  * otherwise a negative error code is returned.
481  */
482 static int commit_orphans(struct ubifs_info *c)
483 {
484 	int avail, atomic = 0, err;
485 
486 	ubifs_assert(c, c->cmt_orphans > 0);
487 	avail = avail_orphs(c);
488 	if (avail < c->cmt_orphans) {
489 		/* Not enough space to write new orphans, so consolidate */
490 		err = consolidate(c);
491 		if (err)
492 			return err;
493 		atomic = 1;
494 	}
495 	err = write_orph_nodes(c, atomic);
496 	return err;
497 }
498 
499 /**
500  * erase_deleted - erase the orphans marked for deletion.
501  * @c: UBIFS file-system description object
502  *
503  * During commit, the orphans being committed cannot be deleted, so they are
504  * marked for deletion and deleted by this function. Also, the recovery
505  * adds killed orphans to the deletion list, and therefore they are deleted
506  * here too.
507  */
508 static void erase_deleted(struct ubifs_info *c)
509 {
510 	struct ubifs_orphan *orphan, *dnext;
511 
512 	spin_lock(&c->orphan_lock);
513 	dnext = c->orph_dnext;
514 	while (dnext) {
515 		orphan = dnext;
516 		dnext = orphan->dnext;
517 		ubifs_assert(c, !orphan->new);
518 		ubifs_assert(c, orphan->del);
519 		rb_erase(&orphan->rb, &c->orph_tree);
520 		list_del(&orphan->list);
521 		c->tot_orphans -= 1;
522 		dbg_gen("deleting orphan ino %lu", (unsigned long)orphan->inum);
523 		kfree(orphan);
524 	}
525 	c->orph_dnext = NULL;
526 	spin_unlock(&c->orphan_lock);
527 }
528 
529 /**
530  * ubifs_orphan_end_commit - end commit of orphans.
531  * @c: UBIFS file-system description object
532  *
533  * End commit of orphans.
534  */
535 int ubifs_orphan_end_commit(struct ubifs_info *c)
536 {
537 	int err;
538 
539 	if (c->cmt_orphans != 0) {
540 		err = commit_orphans(c);
541 		if (err)
542 			return err;
543 	}
544 	erase_deleted(c);
545 	err = dbg_check_orphans(c);
546 	return err;
547 }
548 
549 /**
550  * ubifs_clear_orphans - erase all LEBs used for orphans.
551  * @c: UBIFS file-system description object
552  *
553  * If recovery is not required, then the orphans from the previous session
554  * are not needed. This function locates the LEBs used to record
555  * orphans, and un-maps them.
556  */
557 int ubifs_clear_orphans(struct ubifs_info *c)
558 {
559 	int lnum, err;
560 
561 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
562 		err = ubifs_leb_unmap(c, lnum);
563 		if (err)
564 			return err;
565 	}
566 	c->ohead_lnum = c->orph_first;
567 	c->ohead_offs = 0;
568 	return 0;
569 }
570 
571 /**
572  * insert_dead_orphan - insert an orphan.
573  * @c: UBIFS file-system description object
574  * @inum: orphan inode number
575  *
576  * This function is a helper to the 'do_kill_orphans()' function. The orphan
577  * must be kept until the next commit, so it is added to the rb-tree and the
578  * deletion list.
579  */
580 static int insert_dead_orphan(struct ubifs_info *c, ino_t inum)
581 {
582 	struct ubifs_orphan *orphan, *o;
583 	struct rb_node **p, *parent = NULL;
584 
585 	orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_KERNEL);
586 	if (!orphan)
587 		return -ENOMEM;
588 	orphan->inum = inum;
589 
590 	p = &c->orph_tree.rb_node;
591 	while (*p) {
592 		parent = *p;
593 		o = rb_entry(parent, struct ubifs_orphan, rb);
594 		if (inum < o->inum)
595 			p = &(*p)->rb_left;
596 		else if (inum > o->inum)
597 			p = &(*p)->rb_right;
598 		else {
599 			/* Already added - no problem */
600 			kfree(orphan);
601 			return 0;
602 		}
603 	}
604 	c->tot_orphans += 1;
605 	rb_link_node(&orphan->rb, parent, p);
606 	rb_insert_color(&orphan->rb, &c->orph_tree);
607 	list_add_tail(&orphan->list, &c->orph_list);
608 	orphan->del = 1;
609 	orphan->dnext = c->orph_dnext;
610 	c->orph_dnext = orphan;
611 	dbg_mnt("ino %lu, new %d, tot %d", (unsigned long)inum,
612 		c->new_orphans, c->tot_orphans);
613 	return 0;
614 }
615 
616 /**
617  * do_kill_orphans - remove orphan inodes from the index.
618  * @c: UBIFS file-system description object
619  * @sleb: scanned LEB
620  * @last_cmt_no: cmt_no of last orphan node read is passed and returned here
621  * @outofdate: whether the LEB is out of date is returned here
622  * @last_flagged: whether the end orphan node is encountered
623  *
624  * This function is a helper to the 'kill_orphans()' function. It goes through
625  * every orphan node in a LEB and for every inode number recorded, removes
626  * all keys for that inode from the TNC.
627  */
628 static int do_kill_orphans(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
629 			   unsigned long long *last_cmt_no, int *outofdate,
630 			   int *last_flagged)
631 {
632 	struct ubifs_scan_node *snod;
633 	struct ubifs_orph_node *orph;
634 	struct ubifs_ino_node *ino = NULL;
635 	unsigned long long cmt_no;
636 	ino_t inum;
637 	int i, n, err, first = 1;
638 
639 	ino = kmalloc(UBIFS_MAX_INO_NODE_SZ, GFP_NOFS);
640 	if (!ino)
641 		return -ENOMEM;
642 
643 	list_for_each_entry(snod, &sleb->nodes, list) {
644 		if (snod->type != UBIFS_ORPH_NODE) {
645 			ubifs_err(c, "invalid node type %d in orphan area at %d:%d",
646 				  snod->type, sleb->lnum, snod->offs);
647 			ubifs_dump_node(c, snod->node);
648 			err = -EINVAL;
649 			goto out_free;
650 		}
651 
652 		orph = snod->node;
653 
654 		/* Check commit number */
655 		cmt_no = le64_to_cpu(orph->cmt_no) & LLONG_MAX;
656 		/*
657 		 * The commit number on the master node may be less, because
658 		 * of a failed commit. If there are several failed commits in a
659 		 * row, the commit number written on orphan nodes will continue
660 		 * to increase (because the commit number is adjusted here) even
661 		 * though the commit number on the master node stays the same
662 		 * because the master node has not been re-written.
663 		 */
664 		if (cmt_no > c->cmt_no)
665 			c->cmt_no = cmt_no;
666 		if (cmt_no < *last_cmt_no && *last_flagged) {
667 			/*
668 			 * The last orphan node had a higher commit number and
669 			 * was flagged as the last written for that commit
670 			 * number. That makes this orphan node, out of date.
671 			 */
672 			if (!first) {
673 				ubifs_err(c, "out of order commit number %llu in orphan node at %d:%d",
674 					  cmt_no, sleb->lnum, snod->offs);
675 				ubifs_dump_node(c, snod->node);
676 				err = -EINVAL;
677 				goto out_free;
678 			}
679 			dbg_rcvry("out of date LEB %d", sleb->lnum);
680 			*outofdate = 1;
681 			err = 0;
682 			goto out_free;
683 		}
684 
685 		if (first)
686 			first = 0;
687 
688 		n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
689 		for (i = 0; i < n; i++) {
690 			union ubifs_key key1, key2;
691 
692 			inum = le64_to_cpu(orph->inos[i]);
693 
694 			ino_key_init(c, &key1, inum);
695 			err = ubifs_tnc_lookup(c, &key1, ino);
696 			if (err && err != -ENOENT)
697 				goto out_free;
698 
699 			/*
700 			 * Check whether an inode can really get deleted.
701 			 * linkat() with O_TMPFILE allows rebirth of an inode.
702 			 */
703 			if (err == 0 && ino->nlink == 0) {
704 				dbg_rcvry("deleting orphaned inode %lu",
705 					  (unsigned long)inum);
706 
707 				lowest_ino_key(c, &key1, inum);
708 				highest_ino_key(c, &key2, inum);
709 
710 				err = ubifs_tnc_remove_range(c, &key1, &key2);
711 				if (err)
712 					goto out_ro;
713 			}
714 
715 			err = insert_dead_orphan(c, inum);
716 			if (err)
717 				goto out_free;
718 		}
719 
720 		*last_cmt_no = cmt_no;
721 		if (le64_to_cpu(orph->cmt_no) & (1ULL << 63)) {
722 			dbg_rcvry("last orph node for commit %llu at %d:%d",
723 				  cmt_no, sleb->lnum, snod->offs);
724 			*last_flagged = 1;
725 		} else
726 			*last_flagged = 0;
727 	}
728 
729 	err = 0;
730 out_free:
731 	kfree(ino);
732 	return err;
733 
734 out_ro:
735 	ubifs_ro_mode(c, err);
736 	kfree(ino);
737 	return err;
738 }
739 
740 /**
741  * kill_orphans - remove all orphan inodes from the index.
742  * @c: UBIFS file-system description object
743  *
744  * If recovery is required, then orphan inodes recorded during the previous
745  * session (which ended with an unclean unmount) must be deleted from the index.
746  * This is done by updating the TNC, but since the index is not updated until
747  * the next commit, the LEBs where the orphan information is recorded are not
748  * erased until the next commit.
749  */
750 static int kill_orphans(struct ubifs_info *c)
751 {
752 	unsigned long long last_cmt_no = 0;
753 	int lnum, err = 0, outofdate = 0, last_flagged = 0;
754 
755 	c->ohead_lnum = c->orph_first;
756 	c->ohead_offs = 0;
757 	/* Check no-orphans flag and skip this if no orphans */
758 	if (c->no_orphs) {
759 		dbg_rcvry("no orphans");
760 		return 0;
761 	}
762 	/*
763 	 * Orph nodes always start at c->orph_first and are written to each
764 	 * successive LEB in turn. Generally unused LEBs will have been unmapped
765 	 * but may contain out of date orphan nodes if the unmap didn't go
766 	 * through. In addition, the last orphan node written for each commit is
767 	 * marked (top bit of orph->cmt_no is set to 1). It is possible that
768 	 * there are orphan nodes from the next commit (i.e. the commit did not
769 	 * complete successfully). In that case, no orphans will have been lost
770 	 * due to the way that orphans are written, and any orphans added will
771 	 * be valid orphans anyway and so can be deleted.
772 	 */
773 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
774 		struct ubifs_scan_leb *sleb;
775 
776 		dbg_rcvry("LEB %d", lnum);
777 		sleb = ubifs_scan(c, lnum, 0, c->sbuf, 1);
778 		if (IS_ERR(sleb)) {
779 			if (PTR_ERR(sleb) == -EUCLEAN)
780 				sleb = ubifs_recover_leb(c, lnum, 0,
781 							 c->sbuf, -1);
782 			if (IS_ERR(sleb)) {
783 				err = PTR_ERR(sleb);
784 				break;
785 			}
786 		}
787 		err = do_kill_orphans(c, sleb, &last_cmt_no, &outofdate,
788 				      &last_flagged);
789 		if (err || outofdate) {
790 			ubifs_scan_destroy(sleb);
791 			break;
792 		}
793 		if (sleb->endpt) {
794 			c->ohead_lnum = lnum;
795 			c->ohead_offs = sleb->endpt;
796 		}
797 		ubifs_scan_destroy(sleb);
798 	}
799 	return err;
800 }
801 
802 /**
803  * ubifs_mount_orphans - delete orphan inodes and erase LEBs that recorded them.
804  * @c: UBIFS file-system description object
805  * @unclean: indicates recovery from unclean unmount
806  * @read_only: indicates read only mount
807  *
808  * This function is called when mounting to erase orphans from the previous
809  * session. If UBIFS was not unmounted cleanly, then the inodes recorded as
810  * orphans are deleted.
811  */
812 int ubifs_mount_orphans(struct ubifs_info *c, int unclean, int read_only)
813 {
814 	int err = 0;
815 
816 	c->max_orphans = tot_avail_orphs(c);
817 
818 	if (!read_only) {
819 		c->orph_buf = vmalloc(c->leb_size);
820 		if (!c->orph_buf)
821 			return -ENOMEM;
822 	}
823 
824 	if (unclean)
825 		err = kill_orphans(c);
826 	else if (!read_only)
827 		err = ubifs_clear_orphans(c);
828 
829 	return err;
830 }
831 
832 /*
833  * Everything below is related to debugging.
834  */
835 
836 struct check_orphan {
837 	struct rb_node rb;
838 	ino_t inum;
839 };
840 
841 struct check_info {
842 	unsigned long last_ino;
843 	unsigned long tot_inos;
844 	unsigned long missing;
845 	unsigned long long leaf_cnt;
846 	struct ubifs_ino_node *node;
847 	struct rb_root root;
848 };
849 
850 static bool dbg_find_orphan(struct ubifs_info *c, ino_t inum)
851 {
852 	bool found = false;
853 
854 	spin_lock(&c->orphan_lock);
855 	found = !!lookup_orphan(c, inum);
856 	spin_unlock(&c->orphan_lock);
857 
858 	return found;
859 }
860 
861 static int dbg_ins_check_orphan(struct rb_root *root, ino_t inum)
862 {
863 	struct check_orphan *orphan, *o;
864 	struct rb_node **p, *parent = NULL;
865 
866 	orphan = kzalloc(sizeof(struct check_orphan), GFP_NOFS);
867 	if (!orphan)
868 		return -ENOMEM;
869 	orphan->inum = inum;
870 
871 	p = &root->rb_node;
872 	while (*p) {
873 		parent = *p;
874 		o = rb_entry(parent, struct check_orphan, rb);
875 		if (inum < o->inum)
876 			p = &(*p)->rb_left;
877 		else if (inum > o->inum)
878 			p = &(*p)->rb_right;
879 		else {
880 			kfree(orphan);
881 			return 0;
882 		}
883 	}
884 	rb_link_node(&orphan->rb, parent, p);
885 	rb_insert_color(&orphan->rb, root);
886 	return 0;
887 }
888 
889 static int dbg_find_check_orphan(struct rb_root *root, ino_t inum)
890 {
891 	struct check_orphan *o;
892 	struct rb_node *p;
893 
894 	p = root->rb_node;
895 	while (p) {
896 		o = rb_entry(p, struct check_orphan, rb);
897 		if (inum < o->inum)
898 			p = p->rb_left;
899 		else if (inum > o->inum)
900 			p = p->rb_right;
901 		else
902 			return 1;
903 	}
904 	return 0;
905 }
906 
907 static void dbg_free_check_tree(struct rb_root *root)
908 {
909 	struct check_orphan *o, *n;
910 
911 	rbtree_postorder_for_each_entry_safe(o, n, root, rb)
912 		kfree(o);
913 }
914 
915 static int dbg_orphan_check(struct ubifs_info *c, struct ubifs_zbranch *zbr,
916 			    void *priv)
917 {
918 	struct check_info *ci = priv;
919 	ino_t inum;
920 	int err;
921 
922 	inum = key_inum(c, &zbr->key);
923 	if (inum != ci->last_ino) {
924 		/* Lowest node type is the inode node, so it comes first */
925 		if (key_type(c, &zbr->key) != UBIFS_INO_KEY)
926 			ubifs_err(c, "found orphan node ino %lu, type %d",
927 				  (unsigned long)inum, key_type(c, &zbr->key));
928 		ci->last_ino = inum;
929 		ci->tot_inos += 1;
930 		err = ubifs_tnc_read_node(c, zbr, ci->node);
931 		if (err) {
932 			ubifs_err(c, "node read failed, error %d", err);
933 			return err;
934 		}
935 		if (ci->node->nlink == 0)
936 			/* Must be recorded as an orphan */
937 			if (!dbg_find_check_orphan(&ci->root, inum) &&
938 			    !dbg_find_orphan(c, inum)) {
939 				ubifs_err(c, "missing orphan, ino %lu",
940 					  (unsigned long)inum);
941 				ci->missing += 1;
942 			}
943 	}
944 	ci->leaf_cnt += 1;
945 	return 0;
946 }
947 
948 static int dbg_read_orphans(struct check_info *ci, struct ubifs_scan_leb *sleb)
949 {
950 	struct ubifs_scan_node *snod;
951 	struct ubifs_orph_node *orph;
952 	ino_t inum;
953 	int i, n, err;
954 
955 	list_for_each_entry(snod, &sleb->nodes, list) {
956 		cond_resched();
957 		if (snod->type != UBIFS_ORPH_NODE)
958 			continue;
959 		orph = snod->node;
960 		n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
961 		for (i = 0; i < n; i++) {
962 			inum = le64_to_cpu(orph->inos[i]);
963 			err = dbg_ins_check_orphan(&ci->root, inum);
964 			if (err)
965 				return err;
966 		}
967 	}
968 	return 0;
969 }
970 
971 static int dbg_scan_orphans(struct ubifs_info *c, struct check_info *ci)
972 {
973 	int lnum, err = 0;
974 	void *buf;
975 
976 	/* Check no-orphans flag and skip this if no orphans */
977 	if (c->no_orphs)
978 		return 0;
979 
980 	buf = __vmalloc(c->leb_size, GFP_NOFS);
981 	if (!buf) {
982 		ubifs_err(c, "cannot allocate memory to check orphans");
983 		return 0;
984 	}
985 
986 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
987 		struct ubifs_scan_leb *sleb;
988 
989 		sleb = ubifs_scan(c, lnum, 0, buf, 0);
990 		if (IS_ERR(sleb)) {
991 			err = PTR_ERR(sleb);
992 			break;
993 		}
994 
995 		err = dbg_read_orphans(ci, sleb);
996 		ubifs_scan_destroy(sleb);
997 		if (err)
998 			break;
999 	}
1000 
1001 	vfree(buf);
1002 	return err;
1003 }
1004 
1005 static int dbg_check_orphans(struct ubifs_info *c)
1006 {
1007 	struct check_info ci;
1008 	int err;
1009 
1010 	if (!dbg_is_chk_orph(c))
1011 		return 0;
1012 
1013 	ci.last_ino = 0;
1014 	ci.tot_inos = 0;
1015 	ci.missing  = 0;
1016 	ci.leaf_cnt = 0;
1017 	ci.root = RB_ROOT;
1018 	ci.node = kmalloc(UBIFS_MAX_INO_NODE_SZ, GFP_NOFS);
1019 	if (!ci.node) {
1020 		ubifs_err(c, "out of memory");
1021 		return -ENOMEM;
1022 	}
1023 
1024 	err = dbg_scan_orphans(c, &ci);
1025 	if (err)
1026 		goto out;
1027 
1028 	err = dbg_walk_index(c, &dbg_orphan_check, NULL, &ci);
1029 	if (err) {
1030 		ubifs_err(c, "cannot scan TNC, error %d", err);
1031 		goto out;
1032 	}
1033 
1034 	if (ci.missing) {
1035 		ubifs_err(c, "%lu missing orphan(s)", ci.missing);
1036 		err = -EINVAL;
1037 		goto out;
1038 	}
1039 
1040 	dbg_cmt("last inode number is %lu", ci.last_ino);
1041 	dbg_cmt("total number of inodes is %lu", ci.tot_inos);
1042 	dbg_cmt("total number of leaf nodes is %llu", ci.leaf_cnt);
1043 
1044 out:
1045 	dbg_free_check_tree(&ci.root);
1046 	kfree(ci.node);
1047 	return err;
1048 }
1049