xref: /openbmc/linux/fs/ubifs/orphan.c (revision 367b8112)
1 /*
2  * This file is part of UBIFS.
3  *
4  * Copyright (C) 2006-2008 Nokia Corporation.
5  *
6  * This program is free software; you can redistribute it and/or modify it
7  * under the terms of the GNU General Public License version 2 as published by
8  * the Free Software Foundation.
9  *
10  * This program is distributed in the hope that it will be useful, but WITHOUT
11  * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
12  * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
13  * more details.
14  *
15  * You should have received a copy of the GNU General Public License along with
16  * this program; if not, write to the Free Software Foundation, Inc., 51
17  * Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
18  *
19  * Author: Adrian Hunter
20  */
21 
22 #include "ubifs.h"
23 
24 /*
25  * An orphan is an inode number whose inode node has been committed to the index
26  * with a link count of zero. That happens when an open file is deleted
27  * (unlinked) and then a commit is run. In the normal course of events the inode
28  * would be deleted when the file is closed. However in the case of an unclean
29  * unmount, orphans need to be accounted for. After an unclean unmount, the
30  * orphans' inodes must be deleted which means either scanning the entire index
31  * looking for them, or keeping a list on flash somewhere. This unit implements
32  * the latter approach.
33  *
34  * The orphan area is a fixed number of LEBs situated between the LPT area and
35  * the main area. The number of orphan area LEBs is specified when the file
36  * system is created. The minimum number is 1. The size of the orphan area
37  * should be so that it can hold the maximum number of orphans that are expected
38  * to ever exist at one time.
39  *
40  * The number of orphans that can fit in a LEB is:
41  *
42  *         (c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64)
43  *
44  * For example: a 15872 byte LEB can fit 1980 orphans so 1 LEB may be enough.
45  *
46  * Orphans are accumulated in a rb-tree. When an inode's link count drops to
47  * zero, the inode number is added to the rb-tree. It is removed from the tree
48  * when the inode is deleted.  Any new orphans that are in the orphan tree when
49  * the commit is run, are written to the orphan area in 1 or more orph nodes.
50  * If the orphan area is full, it is consolidated to make space.  There is
51  * always enough space because validation prevents the user from creating more
52  * than the maximum number of orphans allowed.
53  */
54 
55 #ifdef CONFIG_UBIFS_FS_DEBUG
56 static int dbg_check_orphans(struct ubifs_info *c);
57 #else
58 #define dbg_check_orphans(c) 0
59 #endif
60 
61 /**
62  * ubifs_add_orphan - add an orphan.
63  * @c: UBIFS file-system description object
64  * @inum: orphan inode number
65  *
66  * Add an orphan. This function is called when an inodes link count drops to
67  * zero.
68  */
69 int ubifs_add_orphan(struct ubifs_info *c, ino_t inum)
70 {
71 	struct ubifs_orphan *orphan, *o;
72 	struct rb_node **p, *parent = NULL;
73 
74 	orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_NOFS);
75 	if (!orphan)
76 		return -ENOMEM;
77 	orphan->inum = inum;
78 	orphan->new = 1;
79 
80 	spin_lock(&c->orphan_lock);
81 	if (c->tot_orphans >= c->max_orphans) {
82 		spin_unlock(&c->orphan_lock);
83 		kfree(orphan);
84 		return -ENFILE;
85 	}
86 	p = &c->orph_tree.rb_node;
87 	while (*p) {
88 		parent = *p;
89 		o = rb_entry(parent, struct ubifs_orphan, rb);
90 		if (inum < o->inum)
91 			p = &(*p)->rb_left;
92 		else if (inum > o->inum)
93 			p = &(*p)->rb_right;
94 		else {
95 			dbg_err("orphaned twice");
96 			spin_unlock(&c->orphan_lock);
97 			kfree(orphan);
98 			return 0;
99 		}
100 	}
101 	c->tot_orphans += 1;
102 	c->new_orphans += 1;
103 	rb_link_node(&orphan->rb, parent, p);
104 	rb_insert_color(&orphan->rb, &c->orph_tree);
105 	list_add_tail(&orphan->list, &c->orph_list);
106 	list_add_tail(&orphan->new_list, &c->orph_new);
107 	spin_unlock(&c->orphan_lock);
108 	dbg_gen("ino %lu", inum);
109 	return 0;
110 }
111 
112 /**
113  * ubifs_delete_orphan - delete an orphan.
114  * @c: UBIFS file-system description object
115  * @inum: orphan inode number
116  *
117  * Delete an orphan. This function is called when an inode is deleted.
118  */
119 void ubifs_delete_orphan(struct ubifs_info *c, ino_t inum)
120 {
121 	struct ubifs_orphan *o;
122 	struct rb_node *p;
123 
124 	spin_lock(&c->orphan_lock);
125 	p = c->orph_tree.rb_node;
126 	while (p) {
127 		o = rb_entry(p, struct ubifs_orphan, rb);
128 		if (inum < o->inum)
129 			p = p->rb_left;
130 		else if (inum > o->inum)
131 			p = p->rb_right;
132 		else {
133 			if (o->dnext) {
134 				spin_unlock(&c->orphan_lock);
135 				dbg_gen("deleted twice ino %lu", inum);
136 				return;
137 			}
138 			if (o->cnext) {
139 				o->dnext = c->orph_dnext;
140 				c->orph_dnext = o;
141 				spin_unlock(&c->orphan_lock);
142 				dbg_gen("delete later ino %lu", inum);
143 				return;
144 			}
145 			rb_erase(p, &c->orph_tree);
146 			list_del(&o->list);
147 			c->tot_orphans -= 1;
148 			if (o->new) {
149 				list_del(&o->new_list);
150 				c->new_orphans -= 1;
151 			}
152 			spin_unlock(&c->orphan_lock);
153 			kfree(o);
154 			dbg_gen("inum %lu", inum);
155 			return;
156 		}
157 	}
158 	spin_unlock(&c->orphan_lock);
159 	dbg_err("missing orphan ino %lu", inum);
160 	dbg_dump_stack();
161 }
162 
163 /**
164  * ubifs_orphan_start_commit - start commit of orphans.
165  * @c: UBIFS file-system description object
166  *
167  * Start commit of orphans.
168  */
169 int ubifs_orphan_start_commit(struct ubifs_info *c)
170 {
171 	struct ubifs_orphan *orphan, **last;
172 
173 	spin_lock(&c->orphan_lock);
174 	last = &c->orph_cnext;
175 	list_for_each_entry(orphan, &c->orph_new, new_list) {
176 		ubifs_assert(orphan->new);
177 		orphan->new = 0;
178 		*last = orphan;
179 		last = &orphan->cnext;
180 	}
181 	*last = orphan->cnext;
182 	c->cmt_orphans = c->new_orphans;
183 	c->new_orphans = 0;
184 	dbg_cmt("%d orphans to commit", c->cmt_orphans);
185 	INIT_LIST_HEAD(&c->orph_new);
186 	if (c->tot_orphans == 0)
187 		c->no_orphs = 1;
188 	else
189 		c->no_orphs = 0;
190 	spin_unlock(&c->orphan_lock);
191 	return 0;
192 }
193 
194 /**
195  * avail_orphs - calculate available space.
196  * @c: UBIFS file-system description object
197  *
198  * This function returns the number of orphans that can be written in the
199  * available space.
200  */
201 static int avail_orphs(struct ubifs_info *c)
202 {
203 	int avail_lebs, avail, gap;
204 
205 	avail_lebs = c->orph_lebs - (c->ohead_lnum - c->orph_first) - 1;
206 	avail = avail_lebs *
207 	       ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
208 	gap = c->leb_size - c->ohead_offs;
209 	if (gap >= UBIFS_ORPH_NODE_SZ + sizeof(__le64))
210 		avail += (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
211 	return avail;
212 }
213 
214 /**
215  * tot_avail_orphs - calculate total space.
216  * @c: UBIFS file-system description object
217  *
218  * This function returns the number of orphans that can be written in half
219  * the total space. That leaves half the space for adding new orphans.
220  */
221 static int tot_avail_orphs(struct ubifs_info *c)
222 {
223 	int avail_lebs, avail;
224 
225 	avail_lebs = c->orph_lebs;
226 	avail = avail_lebs *
227 	       ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
228 	return avail / 2;
229 }
230 
231 /**
232  * do_write_orph_node - write a node
233  * @c: UBIFS file-system description object
234  * @len: length of node
235  * @atomic: write atomically
236  *
237  * This function writes a node to the orphan head from the orphan buffer. If
238  * %atomic is not zero, then the write is done atomically. On success, %0 is
239  * returned, otherwise a negative error code is returned.
240  */
241 static int do_write_orph_node(struct ubifs_info *c, int len, int atomic)
242 {
243 	int err = 0;
244 
245 	if (atomic) {
246 		ubifs_assert(c->ohead_offs == 0);
247 		ubifs_prepare_node(c, c->orph_buf, len, 1);
248 		len = ALIGN(len, c->min_io_size);
249 		err = ubifs_leb_change(c, c->ohead_lnum, c->orph_buf, len,
250 				       UBI_SHORTTERM);
251 	} else {
252 		if (c->ohead_offs == 0) {
253 			/* Ensure LEB has been unmapped */
254 			err = ubifs_leb_unmap(c, c->ohead_lnum);
255 			if (err)
256 				return err;
257 		}
258 		err = ubifs_write_node(c, c->orph_buf, len, c->ohead_lnum,
259 				       c->ohead_offs, UBI_SHORTTERM);
260 	}
261 	return err;
262 }
263 
264 /**
265  * write_orph_node - write an orph node
266  * @c: UBIFS file-system description object
267  * @atomic: write atomically
268  *
269  * This function builds an orph node from the cnext list and writes it to the
270  * orphan head. On success, %0 is returned, otherwise a negative error code
271  * is returned.
272  */
273 static int write_orph_node(struct ubifs_info *c, int atomic)
274 {
275 	struct ubifs_orphan *orphan, *cnext;
276 	struct ubifs_orph_node *orph;
277 	int gap, err, len, cnt, i;
278 
279 	ubifs_assert(c->cmt_orphans > 0);
280 	gap = c->leb_size - c->ohead_offs;
281 	if (gap < UBIFS_ORPH_NODE_SZ + sizeof(__le64)) {
282 		c->ohead_lnum += 1;
283 		c->ohead_offs = 0;
284 		gap = c->leb_size;
285 		if (c->ohead_lnum > c->orph_last) {
286 			/*
287 			 * We limit the number of orphans so that this should
288 			 * never happen.
289 			 */
290 			ubifs_err("out of space in orphan area");
291 			return -EINVAL;
292 		}
293 	}
294 	cnt = (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
295 	if (cnt > c->cmt_orphans)
296 		cnt = c->cmt_orphans;
297 	len = UBIFS_ORPH_NODE_SZ + cnt * sizeof(__le64);
298 	ubifs_assert(c->orph_buf);
299 	orph = c->orph_buf;
300 	orph->ch.node_type = UBIFS_ORPH_NODE;
301 	spin_lock(&c->orphan_lock);
302 	cnext = c->orph_cnext;
303 	for (i = 0; i < cnt; i++) {
304 		orphan = cnext;
305 		orph->inos[i] = cpu_to_le64(orphan->inum);
306 		cnext = orphan->cnext;
307 		orphan->cnext = NULL;
308 	}
309 	c->orph_cnext = cnext;
310 	c->cmt_orphans -= cnt;
311 	spin_unlock(&c->orphan_lock);
312 	if (c->cmt_orphans)
313 		orph->cmt_no = cpu_to_le64(c->cmt_no);
314 	else
315 		/* Mark the last node of the commit */
316 		orph->cmt_no = cpu_to_le64((c->cmt_no) | (1ULL << 63));
317 	ubifs_assert(c->ohead_offs + len <= c->leb_size);
318 	ubifs_assert(c->ohead_lnum >= c->orph_first);
319 	ubifs_assert(c->ohead_lnum <= c->orph_last);
320 	err = do_write_orph_node(c, len, atomic);
321 	c->ohead_offs += ALIGN(len, c->min_io_size);
322 	c->ohead_offs = ALIGN(c->ohead_offs, 8);
323 	return err;
324 }
325 
326 /**
327  * write_orph_nodes - write orph nodes until there are no more to commit
328  * @c: UBIFS file-system description object
329  * @atomic: write atomically
330  *
331  * This function writes orph nodes for all the orphans to commit. On success,
332  * %0 is returned, otherwise a negative error code is returned.
333  */
334 static int write_orph_nodes(struct ubifs_info *c, int atomic)
335 {
336 	int err;
337 
338 	while (c->cmt_orphans > 0) {
339 		err = write_orph_node(c, atomic);
340 		if (err)
341 			return err;
342 	}
343 	if (atomic) {
344 		int lnum;
345 
346 		/* Unmap any unused LEBs after consolidation */
347 		lnum = c->ohead_lnum + 1;
348 		for (lnum = c->ohead_lnum + 1; lnum <= c->orph_last; lnum++) {
349 			err = ubifs_leb_unmap(c, lnum);
350 			if (err)
351 				return err;
352 		}
353 	}
354 	return 0;
355 }
356 
357 /**
358  * consolidate - consolidate the orphan area.
359  * @c: UBIFS file-system description object
360  *
361  * This function enables consolidation by putting all the orphans into the list
362  * to commit. The list is in the order that the orphans were added, and the
363  * LEBs are written atomically in order, so at no time can orphans be lost by
364  * an unclean unmount.
365  *
366  * This function returns %0 on success and a negative error code on failure.
367  */
368 static int consolidate(struct ubifs_info *c)
369 {
370 	int tot_avail = tot_avail_orphs(c), err = 0;
371 
372 	spin_lock(&c->orphan_lock);
373 	dbg_cmt("there is space for %d orphans and there are %d",
374 		tot_avail, c->tot_orphans);
375 	if (c->tot_orphans - c->new_orphans <= tot_avail) {
376 		struct ubifs_orphan *orphan, **last;
377 		int cnt = 0;
378 
379 		/* Change the cnext list to include all non-new orphans */
380 		last = &c->orph_cnext;
381 		list_for_each_entry(orphan, &c->orph_list, list) {
382 			if (orphan->new)
383 				continue;
384 			*last = orphan;
385 			last = &orphan->cnext;
386 			cnt += 1;
387 		}
388 		*last = orphan->cnext;
389 		ubifs_assert(cnt == c->tot_orphans - c->new_orphans);
390 		c->cmt_orphans = cnt;
391 		c->ohead_lnum = c->orph_first;
392 		c->ohead_offs = 0;
393 	} else {
394 		/*
395 		 * We limit the number of orphans so that this should
396 		 * never happen.
397 		 */
398 		ubifs_err("out of space in orphan area");
399 		err = -EINVAL;
400 	}
401 	spin_unlock(&c->orphan_lock);
402 	return err;
403 }
404 
405 /**
406  * commit_orphans - commit orphans.
407  * @c: UBIFS file-system description object
408  *
409  * This function commits orphans to flash. On success, %0 is returned,
410  * otherwise a negative error code is returned.
411  */
412 static int commit_orphans(struct ubifs_info *c)
413 {
414 	int avail, atomic = 0, err;
415 
416 	ubifs_assert(c->cmt_orphans > 0);
417 	avail = avail_orphs(c);
418 	if (avail < c->cmt_orphans) {
419 		/* Not enough space to write new orphans, so consolidate */
420 		err = consolidate(c);
421 		if (err)
422 			return err;
423 		atomic = 1;
424 	}
425 	err = write_orph_nodes(c, atomic);
426 	return err;
427 }
428 
429 /**
430  * erase_deleted - erase the orphans marked for deletion.
431  * @c: UBIFS file-system description object
432  *
433  * During commit, the orphans being committed cannot be deleted, so they are
434  * marked for deletion and deleted by this function. Also, the recovery
435  * adds killed orphans to the deletion list, and therefore they are deleted
436  * here too.
437  */
438 static void erase_deleted(struct ubifs_info *c)
439 {
440 	struct ubifs_orphan *orphan, *dnext;
441 
442 	spin_lock(&c->orphan_lock);
443 	dnext = c->orph_dnext;
444 	while (dnext) {
445 		orphan = dnext;
446 		dnext = orphan->dnext;
447 		ubifs_assert(!orphan->new);
448 		rb_erase(&orphan->rb, &c->orph_tree);
449 		list_del(&orphan->list);
450 		c->tot_orphans -= 1;
451 		dbg_gen("deleting orphan ino %lu", orphan->inum);
452 		kfree(orphan);
453 	}
454 	c->orph_dnext = NULL;
455 	spin_unlock(&c->orphan_lock);
456 }
457 
458 /**
459  * ubifs_orphan_end_commit - end commit of orphans.
460  * @c: UBIFS file-system description object
461  *
462  * End commit of orphans.
463  */
464 int ubifs_orphan_end_commit(struct ubifs_info *c)
465 {
466 	int err;
467 
468 	if (c->cmt_orphans != 0) {
469 		err = commit_orphans(c);
470 		if (err)
471 			return err;
472 	}
473 	erase_deleted(c);
474 	err = dbg_check_orphans(c);
475 	return err;
476 }
477 
478 /**
479  * clear_orphans - erase all LEBs used for orphans.
480  * @c: UBIFS file-system description object
481  *
482  * If recovery is not required, then the orphans from the previous session
483  * are not needed. This function locates the LEBs used to record
484  * orphans, and un-maps them.
485  */
486 static int clear_orphans(struct ubifs_info *c)
487 {
488 	int lnum, err;
489 
490 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
491 		err = ubifs_leb_unmap(c, lnum);
492 		if (err)
493 			return err;
494 	}
495 	c->ohead_lnum = c->orph_first;
496 	c->ohead_offs = 0;
497 	return 0;
498 }
499 
500 /**
501  * insert_dead_orphan - insert an orphan.
502  * @c: UBIFS file-system description object
503  * @inum: orphan inode number
504  *
505  * This function is a helper to the 'do_kill_orphans()' function. The orphan
506  * must be kept until the next commit, so it is added to the rb-tree and the
507  * deletion list.
508  */
509 static int insert_dead_orphan(struct ubifs_info *c, ino_t inum)
510 {
511 	struct ubifs_orphan *orphan, *o;
512 	struct rb_node **p, *parent = NULL;
513 
514 	orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_KERNEL);
515 	if (!orphan)
516 		return -ENOMEM;
517 	orphan->inum = inum;
518 
519 	p = &c->orph_tree.rb_node;
520 	while (*p) {
521 		parent = *p;
522 		o = rb_entry(parent, struct ubifs_orphan, rb);
523 		if (inum < o->inum)
524 			p = &(*p)->rb_left;
525 		else if (inum > o->inum)
526 			p = &(*p)->rb_right;
527 		else {
528 			/* Already added - no problem */
529 			kfree(orphan);
530 			return 0;
531 		}
532 	}
533 	c->tot_orphans += 1;
534 	rb_link_node(&orphan->rb, parent, p);
535 	rb_insert_color(&orphan->rb, &c->orph_tree);
536 	list_add_tail(&orphan->list, &c->orph_list);
537 	orphan->dnext = c->orph_dnext;
538 	c->orph_dnext = orphan;
539 	dbg_mnt("ino %lu, new %d, tot %d",
540 		inum, c->new_orphans, c->tot_orphans);
541 	return 0;
542 }
543 
544 /**
545  * do_kill_orphans - remove orphan inodes from the index.
546  * @c: UBIFS file-system description object
547  * @sleb: scanned LEB
548  * @last_cmt_no: cmt_no of last orph node read is passed and returned here
549  * @outofdate: whether the LEB is out of date is returned here
550  * @last_flagged: whether the end orph node is encountered
551  *
552  * This function is a helper to the 'kill_orphans()' function. It goes through
553  * every orphan node in a LEB and for every inode number recorded, removes
554  * all keys for that inode from the TNC.
555  */
556 static int do_kill_orphans(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
557 			   unsigned long long *last_cmt_no, int *outofdate,
558 			   int *last_flagged)
559 {
560 	struct ubifs_scan_node *snod;
561 	struct ubifs_orph_node *orph;
562 	unsigned long long cmt_no;
563 	ino_t inum;
564 	int i, n, err, first = 1;
565 
566 	list_for_each_entry(snod, &sleb->nodes, list) {
567 		if (snod->type != UBIFS_ORPH_NODE) {
568 			ubifs_err("invalid node type %d in orphan area at "
569 				  "%d:%d", snod->type, sleb->lnum, snod->offs);
570 			dbg_dump_node(c, snod->node);
571 			return -EINVAL;
572 		}
573 
574 		orph = snod->node;
575 
576 		/* Check commit number */
577 		cmt_no = le64_to_cpu(orph->cmt_no) & LLONG_MAX;
578 		/*
579 		 * The commit number on the master node may be less, because
580 		 * of a failed commit. If there are several failed commits in a
581 		 * row, the commit number written on orph nodes will continue to
582 		 * increase (because the commit number is adjusted here) even
583 		 * though the commit number on the master node stays the same
584 		 * because the master node has not been re-written.
585 		 */
586 		if (cmt_no > c->cmt_no)
587 			c->cmt_no = cmt_no;
588 		if (cmt_no < *last_cmt_no && *last_flagged) {
589 			/*
590 			 * The last orph node had a higher commit number and was
591 			 * flagged as the last written for that commit number.
592 			 * That makes this orph node, out of date.
593 			 */
594 			if (!first) {
595 				ubifs_err("out of order commit number %llu in "
596 					  "orphan node at %d:%d",
597 					  cmt_no, sleb->lnum, snod->offs);
598 				dbg_dump_node(c, snod->node);
599 				return -EINVAL;
600 			}
601 			dbg_rcvry("out of date LEB %d", sleb->lnum);
602 			*outofdate = 1;
603 			return 0;
604 		}
605 
606 		if (first)
607 			first = 0;
608 
609 		n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
610 		for (i = 0; i < n; i++) {
611 			inum = le64_to_cpu(orph->inos[i]);
612 			dbg_rcvry("deleting orphaned inode %lu", inum);
613 			err = ubifs_tnc_remove_ino(c, inum);
614 			if (err)
615 				return err;
616 			err = insert_dead_orphan(c, inum);
617 			if (err)
618 				return err;
619 		}
620 
621 		*last_cmt_no = cmt_no;
622 		if (le64_to_cpu(orph->cmt_no) & (1ULL << 63)) {
623 			dbg_rcvry("last orph node for commit %llu at %d:%d",
624 				  cmt_no, sleb->lnum, snod->offs);
625 			*last_flagged = 1;
626 		} else
627 			*last_flagged = 0;
628 	}
629 
630 	return 0;
631 }
632 
633 /**
634  * kill_orphans - remove all orphan inodes from the index.
635  * @c: UBIFS file-system description object
636  *
637  * If recovery is required, then orphan inodes recorded during the previous
638  * session (which ended with an unclean unmount) must be deleted from the index.
639  * This is done by updating the TNC, but since the index is not updated until
640  * the next commit, the LEBs where the orphan information is recorded are not
641  * erased until the next commit.
642  */
643 static int kill_orphans(struct ubifs_info *c)
644 {
645 	unsigned long long last_cmt_no = 0;
646 	int lnum, err = 0, outofdate = 0, last_flagged = 0;
647 
648 	c->ohead_lnum = c->orph_first;
649 	c->ohead_offs = 0;
650 	/* Check no-orphans flag and skip this if no orphans */
651 	if (c->no_orphs) {
652 		dbg_rcvry("no orphans");
653 		return 0;
654 	}
655 	/*
656 	 * Orph nodes always start at c->orph_first and are written to each
657 	 * successive LEB in turn. Generally unused LEBs will have been unmapped
658 	 * but may contain out of date orph nodes if the unmap didn't go
659 	 * through. In addition, the last orph node written for each commit is
660 	 * marked (top bit of orph->cmt_no is set to 1). It is possible that
661 	 * there are orph nodes from the next commit (i.e. the commit did not
662 	 * complete successfully). In that case, no orphans will have been lost
663 	 * due to the way that orphans are written, and any orphans added will
664 	 * be valid orphans anyway and so can be deleted.
665 	 */
666 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
667 		struct ubifs_scan_leb *sleb;
668 
669 		dbg_rcvry("LEB %d", lnum);
670 		sleb = ubifs_scan(c, lnum, 0, c->sbuf);
671 		if (IS_ERR(sleb)) {
672 			sleb = ubifs_recover_leb(c, lnum, 0, c->sbuf, 0);
673 			if (IS_ERR(sleb)) {
674 				err = PTR_ERR(sleb);
675 				break;
676 			}
677 		}
678 		err = do_kill_orphans(c, sleb, &last_cmt_no, &outofdate,
679 				      &last_flagged);
680 		if (err || outofdate) {
681 			ubifs_scan_destroy(sleb);
682 			break;
683 		}
684 		if (sleb->endpt) {
685 			c->ohead_lnum = lnum;
686 			c->ohead_offs = sleb->endpt;
687 		}
688 		ubifs_scan_destroy(sleb);
689 	}
690 	return err;
691 }
692 
693 /**
694  * ubifs_mount_orphans - delete orphan inodes and erase LEBs that recorded them.
695  * @c: UBIFS file-system description object
696  * @unclean: indicates recovery from unclean unmount
697  * @read_only: indicates read only mount
698  *
699  * This function is called when mounting to erase orphans from the previous
700  * session. If UBIFS was not unmounted cleanly, then the inodes recorded as
701  * orphans are deleted.
702  */
703 int ubifs_mount_orphans(struct ubifs_info *c, int unclean, int read_only)
704 {
705 	int err = 0;
706 
707 	c->max_orphans = tot_avail_orphs(c);
708 
709 	if (!read_only) {
710 		c->orph_buf = vmalloc(c->leb_size);
711 		if (!c->orph_buf)
712 			return -ENOMEM;
713 	}
714 
715 	if (unclean)
716 		err = kill_orphans(c);
717 	else if (!read_only)
718 		err = clear_orphans(c);
719 
720 	return err;
721 }
722 
723 #ifdef CONFIG_UBIFS_FS_DEBUG
724 
725 struct check_orphan {
726 	struct rb_node rb;
727 	ino_t inum;
728 };
729 
730 struct check_info {
731 	unsigned long last_ino;
732 	unsigned long tot_inos;
733 	unsigned long missing;
734 	unsigned long long leaf_cnt;
735 	struct ubifs_ino_node *node;
736 	struct rb_root root;
737 };
738 
739 static int dbg_find_orphan(struct ubifs_info *c, ino_t inum)
740 {
741 	struct ubifs_orphan *o;
742 	struct rb_node *p;
743 
744 	spin_lock(&c->orphan_lock);
745 	p = c->orph_tree.rb_node;
746 	while (p) {
747 		o = rb_entry(p, struct ubifs_orphan, rb);
748 		if (inum < o->inum)
749 			p = p->rb_left;
750 		else if (inum > o->inum)
751 			p = p->rb_right;
752 		else {
753 			spin_unlock(&c->orphan_lock);
754 			return 1;
755 		}
756 	}
757 	spin_unlock(&c->orphan_lock);
758 	return 0;
759 }
760 
761 static int dbg_ins_check_orphan(struct rb_root *root, ino_t inum)
762 {
763 	struct check_orphan *orphan, *o;
764 	struct rb_node **p, *parent = NULL;
765 
766 	orphan = kzalloc(sizeof(struct check_orphan), GFP_NOFS);
767 	if (!orphan)
768 		return -ENOMEM;
769 	orphan->inum = inum;
770 
771 	p = &root->rb_node;
772 	while (*p) {
773 		parent = *p;
774 		o = rb_entry(parent, struct check_orphan, rb);
775 		if (inum < o->inum)
776 			p = &(*p)->rb_left;
777 		else if (inum > o->inum)
778 			p = &(*p)->rb_right;
779 		else {
780 			kfree(orphan);
781 			return 0;
782 		}
783 	}
784 	rb_link_node(&orphan->rb, parent, p);
785 	rb_insert_color(&orphan->rb, root);
786 	return 0;
787 }
788 
789 static int dbg_find_check_orphan(struct rb_root *root, ino_t inum)
790 {
791 	struct check_orphan *o;
792 	struct rb_node *p;
793 
794 	p = root->rb_node;
795 	while (p) {
796 		o = rb_entry(p, struct check_orphan, rb);
797 		if (inum < o->inum)
798 			p = p->rb_left;
799 		else if (inum > o->inum)
800 			p = p->rb_right;
801 		else
802 			return 1;
803 	}
804 	return 0;
805 }
806 
807 static void dbg_free_check_tree(struct rb_root *root)
808 {
809 	struct rb_node *this = root->rb_node;
810 	struct check_orphan *o;
811 
812 	while (this) {
813 		if (this->rb_left) {
814 			this = this->rb_left;
815 			continue;
816 		} else if (this->rb_right) {
817 			this = this->rb_right;
818 			continue;
819 		}
820 		o = rb_entry(this, struct check_orphan, rb);
821 		this = rb_parent(this);
822 		if (this) {
823 			if (this->rb_left == &o->rb)
824 				this->rb_left = NULL;
825 			else
826 				this->rb_right = NULL;
827 		}
828 		kfree(o);
829 	}
830 }
831 
832 static int dbg_orphan_check(struct ubifs_info *c, struct ubifs_zbranch *zbr,
833 			    void *priv)
834 {
835 	struct check_info *ci = priv;
836 	ino_t inum;
837 	int err;
838 
839 	inum = key_inum(c, &zbr->key);
840 	if (inum != ci->last_ino) {
841 		/* Lowest node type is the inode node, so it comes first */
842 		if (key_type(c, &zbr->key) != UBIFS_INO_KEY)
843 			ubifs_err("found orphan node ino %lu, type %d", inum,
844 				  key_type(c, &zbr->key));
845 		ci->last_ino = inum;
846 		ci->tot_inos += 1;
847 		err = ubifs_tnc_read_node(c, zbr, ci->node);
848 		if (err) {
849 			ubifs_err("node read failed, error %d", err);
850 			return err;
851 		}
852 		if (ci->node->nlink == 0)
853 			/* Must be recorded as an orphan */
854 			if (!dbg_find_check_orphan(&ci->root, inum) &&
855 			    !dbg_find_orphan(c, inum)) {
856 				ubifs_err("missing orphan, ino %lu", inum);
857 				ci->missing += 1;
858 			}
859 	}
860 	ci->leaf_cnt += 1;
861 	return 0;
862 }
863 
864 static int dbg_read_orphans(struct check_info *ci, struct ubifs_scan_leb *sleb)
865 {
866 	struct ubifs_scan_node *snod;
867 	struct ubifs_orph_node *orph;
868 	ino_t inum;
869 	int i, n, err;
870 
871 	list_for_each_entry(snod, &sleb->nodes, list) {
872 		cond_resched();
873 		if (snod->type != UBIFS_ORPH_NODE)
874 			continue;
875 		orph = snod->node;
876 		n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
877 		for (i = 0; i < n; i++) {
878 			inum = le64_to_cpu(orph->inos[i]);
879 			err = dbg_ins_check_orphan(&ci->root, inum);
880 			if (err)
881 				return err;
882 		}
883 	}
884 	return 0;
885 }
886 
887 static int dbg_scan_orphans(struct ubifs_info *c, struct check_info *ci)
888 {
889 	int lnum, err = 0;
890 
891 	/* Check no-orphans flag and skip this if no orphans */
892 	if (c->no_orphs)
893 		return 0;
894 
895 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
896 		struct ubifs_scan_leb *sleb;
897 
898 		sleb = ubifs_scan(c, lnum, 0, c->dbg_buf);
899 		if (IS_ERR(sleb)) {
900 			err = PTR_ERR(sleb);
901 			break;
902 		}
903 
904 		err = dbg_read_orphans(ci, sleb);
905 		ubifs_scan_destroy(sleb);
906 		if (err)
907 			break;
908 	}
909 
910 	return err;
911 }
912 
913 static int dbg_check_orphans(struct ubifs_info *c)
914 {
915 	struct check_info ci;
916 	int err;
917 
918 	if (!(ubifs_chk_flags & UBIFS_CHK_ORPH))
919 		return 0;
920 
921 	ci.last_ino = 0;
922 	ci.tot_inos = 0;
923 	ci.missing  = 0;
924 	ci.leaf_cnt = 0;
925 	ci.root = RB_ROOT;
926 	ci.node = kmalloc(UBIFS_MAX_INO_NODE_SZ, GFP_NOFS);
927 	if (!ci.node) {
928 		ubifs_err("out of memory");
929 		return -ENOMEM;
930 	}
931 
932 	err = dbg_scan_orphans(c, &ci);
933 	if (err)
934 		goto out;
935 
936 	err = dbg_walk_index(c, &dbg_orphan_check, NULL, &ci);
937 	if (err) {
938 		ubifs_err("cannot scan TNC, error %d", err);
939 		goto out;
940 	}
941 
942 	if (ci.missing) {
943 		ubifs_err("%lu missing orphan(s)", ci.missing);
944 		err = -EINVAL;
945 		goto out;
946 	}
947 
948 	dbg_cmt("last inode number is %lu", ci.last_ino);
949 	dbg_cmt("total number of inodes is %lu", ci.tot_inos);
950 	dbg_cmt("total number of leaf nodes is %llu", ci.leaf_cnt);
951 
952 out:
953 	dbg_free_check_tree(&ci.root);
954 	kfree(ci.node);
955 	return err;
956 }
957 
958 #endif /* CONFIG_UBIFS_FS_DEBUG */
959