xref: /linux/fs/ubifs/orphan.c (revision 861e10be08c69808065d755d3e3cab5d520a2d32)
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 orphan 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 static int dbg_check_orphans(struct ubifs_info *c);
56 
57 /**
58  * ubifs_add_orphan - add an orphan.
59  * @c: UBIFS file-system description object
60  * @inum: orphan inode number
61  *
62  * Add an orphan. This function is called when an inodes link count drops to
63  * zero.
64  */
65 int ubifs_add_orphan(struct ubifs_info *c, ino_t inum)
66 {
67 	struct ubifs_orphan *orphan, *o;
68 	struct rb_node **p, *parent = NULL;
69 
70 	orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_NOFS);
71 	if (!orphan)
72 		return -ENOMEM;
73 	orphan->inum = inum;
74 	orphan->new = 1;
75 
76 	spin_lock(&c->orphan_lock);
77 	if (c->tot_orphans >= c->max_orphans) {
78 		spin_unlock(&c->orphan_lock);
79 		kfree(orphan);
80 		return -ENFILE;
81 	}
82 	p = &c->orph_tree.rb_node;
83 	while (*p) {
84 		parent = *p;
85 		o = rb_entry(parent, struct ubifs_orphan, rb);
86 		if (inum < o->inum)
87 			p = &(*p)->rb_left;
88 		else if (inum > o->inum)
89 			p = &(*p)->rb_right;
90 		else {
91 			ubifs_err("orphaned twice");
92 			spin_unlock(&c->orphan_lock);
93 			kfree(orphan);
94 			return 0;
95 		}
96 	}
97 	c->tot_orphans += 1;
98 	c->new_orphans += 1;
99 	rb_link_node(&orphan->rb, parent, p);
100 	rb_insert_color(&orphan->rb, &c->orph_tree);
101 	list_add_tail(&orphan->list, &c->orph_list);
102 	list_add_tail(&orphan->new_list, &c->orph_new);
103 	spin_unlock(&c->orphan_lock);
104 	dbg_gen("ino %lu", (unsigned long)inum);
105 	return 0;
106 }
107 
108 /**
109  * ubifs_delete_orphan - delete an orphan.
110  * @c: UBIFS file-system description object
111  * @inum: orphan inode number
112  *
113  * Delete an orphan. This function is called when an inode is deleted.
114  */
115 void ubifs_delete_orphan(struct ubifs_info *c, ino_t inum)
116 {
117 	struct ubifs_orphan *o;
118 	struct rb_node *p;
119 
120 	spin_lock(&c->orphan_lock);
121 	p = c->orph_tree.rb_node;
122 	while (p) {
123 		o = rb_entry(p, struct ubifs_orphan, rb);
124 		if (inum < o->inum)
125 			p = p->rb_left;
126 		else if (inum > o->inum)
127 			p = p->rb_right;
128 		else {
129 			if (o->dnext) {
130 				spin_unlock(&c->orphan_lock);
131 				dbg_gen("deleted twice ino %lu",
132 					(unsigned long)inum);
133 				return;
134 			}
135 			if (o->cnext) {
136 				o->dnext = c->orph_dnext;
137 				c->orph_dnext = o;
138 				spin_unlock(&c->orphan_lock);
139 				dbg_gen("delete later ino %lu",
140 					(unsigned long)inum);
141 				return;
142 			}
143 			rb_erase(p, &c->orph_tree);
144 			list_del(&o->list);
145 			c->tot_orphans -= 1;
146 			if (o->new) {
147 				list_del(&o->new_list);
148 				c->new_orphans -= 1;
149 			}
150 			spin_unlock(&c->orphan_lock);
151 			kfree(o);
152 			dbg_gen("inum %lu", (unsigned long)inum);
153 			return;
154 		}
155 	}
156 	spin_unlock(&c->orphan_lock);
157 	ubifs_err("missing orphan ino %lu", (unsigned long)inum);
158 	dump_stack();
159 }
160 
161 /**
162  * ubifs_orphan_start_commit - start commit of orphans.
163  * @c: UBIFS file-system description object
164  *
165  * Start commit of orphans.
166  */
167 int ubifs_orphan_start_commit(struct ubifs_info *c)
168 {
169 	struct ubifs_orphan *orphan, **last;
170 
171 	spin_lock(&c->orphan_lock);
172 	last = &c->orph_cnext;
173 	list_for_each_entry(orphan, &c->orph_new, new_list) {
174 		ubifs_assert(orphan->new);
175 		orphan->new = 0;
176 		*last = orphan;
177 		last = &orphan->cnext;
178 	}
179 	*last = NULL;
180 	c->cmt_orphans = c->new_orphans;
181 	c->new_orphans = 0;
182 	dbg_cmt("%d orphans to commit", c->cmt_orphans);
183 	INIT_LIST_HEAD(&c->orph_new);
184 	if (c->tot_orphans == 0)
185 		c->no_orphs = 1;
186 	else
187 		c->no_orphs = 0;
188 	spin_unlock(&c->orphan_lock);
189 	return 0;
190 }
191 
192 /**
193  * avail_orphs - calculate available space.
194  * @c: UBIFS file-system description object
195  *
196  * This function returns the number of orphans that can be written in the
197  * available space.
198  */
199 static int avail_orphs(struct ubifs_info *c)
200 {
201 	int avail_lebs, avail, gap;
202 
203 	avail_lebs = c->orph_lebs - (c->ohead_lnum - c->orph_first) - 1;
204 	avail = avail_lebs *
205 	       ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
206 	gap = c->leb_size - c->ohead_offs;
207 	if (gap >= UBIFS_ORPH_NODE_SZ + sizeof(__le64))
208 		avail += (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
209 	return avail;
210 }
211 
212 /**
213  * tot_avail_orphs - calculate total space.
214  * @c: UBIFS file-system description object
215  *
216  * This function returns the number of orphans that can be written in half
217  * the total space. That leaves half the space for adding new orphans.
218  */
219 static int tot_avail_orphs(struct ubifs_info *c)
220 {
221 	int avail_lebs, avail;
222 
223 	avail_lebs = c->orph_lebs;
224 	avail = avail_lebs *
225 	       ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
226 	return avail / 2;
227 }
228 
229 /**
230  * do_write_orph_node - write a node to the orphan head.
231  * @c: UBIFS file-system description object
232  * @len: length of node
233  * @atomic: write atomically
234  *
235  * This function writes a node to the orphan head from the orphan buffer. If
236  * %atomic is not zero, then the write is done atomically. On success, %0 is
237  * returned, otherwise a negative error code is returned.
238  */
239 static int do_write_orph_node(struct ubifs_info *c, int len, int atomic)
240 {
241 	int err = 0;
242 
243 	if (atomic) {
244 		ubifs_assert(c->ohead_offs == 0);
245 		ubifs_prepare_node(c, c->orph_buf, len, 1);
246 		len = ALIGN(len, c->min_io_size);
247 		err = ubifs_leb_change(c, c->ohead_lnum, c->orph_buf, len);
248 	} else {
249 		if (c->ohead_offs == 0) {
250 			/* Ensure LEB has been unmapped */
251 			err = ubifs_leb_unmap(c, c->ohead_lnum);
252 			if (err)
253 				return err;
254 		}
255 		err = ubifs_write_node(c, c->orph_buf, len, c->ohead_lnum,
256 				       c->ohead_offs);
257 	}
258 	return err;
259 }
260 
261 /**
262  * write_orph_node - write an orphan node.
263  * @c: UBIFS file-system description object
264  * @atomic: write atomically
265  *
266  * This function builds an orphan node from the cnext list and writes it to the
267  * orphan head. On success, %0 is returned, otherwise a negative error code
268  * is returned.
269  */
270 static int write_orph_node(struct ubifs_info *c, int atomic)
271 {
272 	struct ubifs_orphan *orphan, *cnext;
273 	struct ubifs_orph_node *orph;
274 	int gap, err, len, cnt, i;
275 
276 	ubifs_assert(c->cmt_orphans > 0);
277 	gap = c->leb_size - c->ohead_offs;
278 	if (gap < UBIFS_ORPH_NODE_SZ + sizeof(__le64)) {
279 		c->ohead_lnum += 1;
280 		c->ohead_offs = 0;
281 		gap = c->leb_size;
282 		if (c->ohead_lnum > c->orph_last) {
283 			/*
284 			 * We limit the number of orphans so that this should
285 			 * never happen.
286 			 */
287 			ubifs_err("out of space in orphan area");
288 			return -EINVAL;
289 		}
290 	}
291 	cnt = (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
292 	if (cnt > c->cmt_orphans)
293 		cnt = c->cmt_orphans;
294 	len = UBIFS_ORPH_NODE_SZ + cnt * sizeof(__le64);
295 	ubifs_assert(c->orph_buf);
296 	orph = c->orph_buf;
297 	orph->ch.node_type = UBIFS_ORPH_NODE;
298 	spin_lock(&c->orphan_lock);
299 	cnext = c->orph_cnext;
300 	for (i = 0; i < cnt; i++) {
301 		orphan = cnext;
302 		orph->inos[i] = cpu_to_le64(orphan->inum);
303 		cnext = orphan->cnext;
304 		orphan->cnext = NULL;
305 	}
306 	c->orph_cnext = cnext;
307 	c->cmt_orphans -= cnt;
308 	spin_unlock(&c->orphan_lock);
309 	if (c->cmt_orphans)
310 		orph->cmt_no = cpu_to_le64(c->cmt_no);
311 	else
312 		/* Mark the last node of the commit */
313 		orph->cmt_no = cpu_to_le64((c->cmt_no) | (1ULL << 63));
314 	ubifs_assert(c->ohead_offs + len <= c->leb_size);
315 	ubifs_assert(c->ohead_lnum >= c->orph_first);
316 	ubifs_assert(c->ohead_lnum <= c->orph_last);
317 	err = do_write_orph_node(c, len, atomic);
318 	c->ohead_offs += ALIGN(len, c->min_io_size);
319 	c->ohead_offs = ALIGN(c->ohead_offs, 8);
320 	return err;
321 }
322 
323 /**
324  * write_orph_nodes - write orphan nodes until there are no more to commit.
325  * @c: UBIFS file-system description object
326  * @atomic: write atomically
327  *
328  * This function writes orphan nodes for all the orphans to commit. On success,
329  * %0 is returned, otherwise a negative error code is returned.
330  */
331 static int write_orph_nodes(struct ubifs_info *c, int atomic)
332 {
333 	int err;
334 
335 	while (c->cmt_orphans > 0) {
336 		err = write_orph_node(c, atomic);
337 		if (err)
338 			return err;
339 	}
340 	if (atomic) {
341 		int lnum;
342 
343 		/* Unmap any unused LEBs after consolidation */
344 		lnum = c->ohead_lnum + 1;
345 		for (lnum = c->ohead_lnum + 1; lnum <= c->orph_last; lnum++) {
346 			err = ubifs_leb_unmap(c, lnum);
347 			if (err)
348 				return err;
349 		}
350 	}
351 	return 0;
352 }
353 
354 /**
355  * consolidate - consolidate the orphan area.
356  * @c: UBIFS file-system description object
357  *
358  * This function enables consolidation by putting all the orphans into the list
359  * to commit. The list is in the order that the orphans were added, and the
360  * LEBs are written atomically in order, so at no time can orphans be lost by
361  * an unclean unmount.
362  *
363  * This function returns %0 on success and a negative error code on failure.
364  */
365 static int consolidate(struct ubifs_info *c)
366 {
367 	int tot_avail = tot_avail_orphs(c), err = 0;
368 
369 	spin_lock(&c->orphan_lock);
370 	dbg_cmt("there is space for %d orphans and there are %d",
371 		tot_avail, c->tot_orphans);
372 	if (c->tot_orphans - c->new_orphans <= tot_avail) {
373 		struct ubifs_orphan *orphan, **last;
374 		int cnt = 0;
375 
376 		/* Change the cnext list to include all non-new orphans */
377 		last = &c->orph_cnext;
378 		list_for_each_entry(orphan, &c->orph_list, list) {
379 			if (orphan->new)
380 				continue;
381 			*last = orphan;
382 			last = &orphan->cnext;
383 			cnt += 1;
384 		}
385 		*last = NULL;
386 		ubifs_assert(cnt == c->tot_orphans - c->new_orphans);
387 		c->cmt_orphans = cnt;
388 		c->ohead_lnum = c->orph_first;
389 		c->ohead_offs = 0;
390 	} else {
391 		/*
392 		 * We limit the number of orphans so that this should
393 		 * never happen.
394 		 */
395 		ubifs_err("out of space in orphan area");
396 		err = -EINVAL;
397 	}
398 	spin_unlock(&c->orphan_lock);
399 	return err;
400 }
401 
402 /**
403  * commit_orphans - commit orphans.
404  * @c: UBIFS file-system description object
405  *
406  * This function commits orphans to flash. On success, %0 is returned,
407  * otherwise a negative error code is returned.
408  */
409 static int commit_orphans(struct ubifs_info *c)
410 {
411 	int avail, atomic = 0, err;
412 
413 	ubifs_assert(c->cmt_orphans > 0);
414 	avail = avail_orphs(c);
415 	if (avail < c->cmt_orphans) {
416 		/* Not enough space to write new orphans, so consolidate */
417 		err = consolidate(c);
418 		if (err)
419 			return err;
420 		atomic = 1;
421 	}
422 	err = write_orph_nodes(c, atomic);
423 	return err;
424 }
425 
426 /**
427  * erase_deleted - erase the orphans marked for deletion.
428  * @c: UBIFS file-system description object
429  *
430  * During commit, the orphans being committed cannot be deleted, so they are
431  * marked for deletion and deleted by this function. Also, the recovery
432  * adds killed orphans to the deletion list, and therefore they are deleted
433  * here too.
434  */
435 static void erase_deleted(struct ubifs_info *c)
436 {
437 	struct ubifs_orphan *orphan, *dnext;
438 
439 	spin_lock(&c->orphan_lock);
440 	dnext = c->orph_dnext;
441 	while (dnext) {
442 		orphan = dnext;
443 		dnext = orphan->dnext;
444 		ubifs_assert(!orphan->new);
445 		rb_erase(&orphan->rb, &c->orph_tree);
446 		list_del(&orphan->list);
447 		c->tot_orphans -= 1;
448 		dbg_gen("deleting orphan ino %lu", (unsigned long)orphan->inum);
449 		kfree(orphan);
450 	}
451 	c->orph_dnext = NULL;
452 	spin_unlock(&c->orphan_lock);
453 }
454 
455 /**
456  * ubifs_orphan_end_commit - end commit of orphans.
457  * @c: UBIFS file-system description object
458  *
459  * End commit of orphans.
460  */
461 int ubifs_orphan_end_commit(struct ubifs_info *c)
462 {
463 	int err;
464 
465 	if (c->cmt_orphans != 0) {
466 		err = commit_orphans(c);
467 		if (err)
468 			return err;
469 	}
470 	erase_deleted(c);
471 	err = dbg_check_orphans(c);
472 	return err;
473 }
474 
475 /**
476  * ubifs_clear_orphans - erase all LEBs used for orphans.
477  * @c: UBIFS file-system description object
478  *
479  * If recovery is not required, then the orphans from the previous session
480  * are not needed. This function locates the LEBs used to record
481  * orphans, and un-maps them.
482  */
483 int ubifs_clear_orphans(struct ubifs_info *c)
484 {
485 	int lnum, err;
486 
487 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
488 		err = ubifs_leb_unmap(c, lnum);
489 		if (err)
490 			return err;
491 	}
492 	c->ohead_lnum = c->orph_first;
493 	c->ohead_offs = 0;
494 	return 0;
495 }
496 
497 /**
498  * insert_dead_orphan - insert an orphan.
499  * @c: UBIFS file-system description object
500  * @inum: orphan inode number
501  *
502  * This function is a helper to the 'do_kill_orphans()' function. The orphan
503  * must be kept until the next commit, so it is added to the rb-tree and the
504  * deletion list.
505  */
506 static int insert_dead_orphan(struct ubifs_info *c, ino_t inum)
507 {
508 	struct ubifs_orphan *orphan, *o;
509 	struct rb_node **p, *parent = NULL;
510 
511 	orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_KERNEL);
512 	if (!orphan)
513 		return -ENOMEM;
514 	orphan->inum = inum;
515 
516 	p = &c->orph_tree.rb_node;
517 	while (*p) {
518 		parent = *p;
519 		o = rb_entry(parent, struct ubifs_orphan, rb);
520 		if (inum < o->inum)
521 			p = &(*p)->rb_left;
522 		else if (inum > o->inum)
523 			p = &(*p)->rb_right;
524 		else {
525 			/* Already added - no problem */
526 			kfree(orphan);
527 			return 0;
528 		}
529 	}
530 	c->tot_orphans += 1;
531 	rb_link_node(&orphan->rb, parent, p);
532 	rb_insert_color(&orphan->rb, &c->orph_tree);
533 	list_add_tail(&orphan->list, &c->orph_list);
534 	orphan->dnext = c->orph_dnext;
535 	c->orph_dnext = orphan;
536 	dbg_mnt("ino %lu, new %d, tot %d", (unsigned long)inum,
537 		c->new_orphans, c->tot_orphans);
538 	return 0;
539 }
540 
541 /**
542  * do_kill_orphans - remove orphan inodes from the index.
543  * @c: UBIFS file-system description object
544  * @sleb: scanned LEB
545  * @last_cmt_no: cmt_no of last orphan node read is passed and returned here
546  * @outofdate: whether the LEB is out of date is returned here
547  * @last_flagged: whether the end orphan node is encountered
548  *
549  * This function is a helper to the 'kill_orphans()' function. It goes through
550  * every orphan node in a LEB and for every inode number recorded, removes
551  * all keys for that inode from the TNC.
552  */
553 static int do_kill_orphans(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
554 			   unsigned long long *last_cmt_no, int *outofdate,
555 			   int *last_flagged)
556 {
557 	struct ubifs_scan_node *snod;
558 	struct ubifs_orph_node *orph;
559 	unsigned long long cmt_no;
560 	ino_t inum;
561 	int i, n, err, first = 1;
562 
563 	list_for_each_entry(snod, &sleb->nodes, list) {
564 		if (snod->type != UBIFS_ORPH_NODE) {
565 			ubifs_err("invalid node type %d in orphan area at %d:%d",
566 				  snod->type, sleb->lnum, snod->offs);
567 			ubifs_dump_node(c, snod->node);
568 			return -EINVAL;
569 		}
570 
571 		orph = snod->node;
572 
573 		/* Check commit number */
574 		cmt_no = le64_to_cpu(orph->cmt_no) & LLONG_MAX;
575 		/*
576 		 * The commit number on the master node may be less, because
577 		 * of a failed commit. If there are several failed commits in a
578 		 * row, the commit number written on orphan nodes will continue
579 		 * to increase (because the commit number is adjusted here) even
580 		 * though the commit number on the master node stays the same
581 		 * because the master node has not been re-written.
582 		 */
583 		if (cmt_no > c->cmt_no)
584 			c->cmt_no = cmt_no;
585 		if (cmt_no < *last_cmt_no && *last_flagged) {
586 			/*
587 			 * The last orphan node had a higher commit number and
588 			 * was flagged as the last written for that commit
589 			 * number. That makes this orphan node, out of date.
590 			 */
591 			if (!first) {
592 				ubifs_err("out of order commit number %llu in orphan node at %d:%d",
593 					  cmt_no, sleb->lnum, snod->offs);
594 				ubifs_dump_node(c, snod->node);
595 				return -EINVAL;
596 			}
597 			dbg_rcvry("out of date LEB %d", sleb->lnum);
598 			*outofdate = 1;
599 			return 0;
600 		}
601 
602 		if (first)
603 			first = 0;
604 
605 		n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
606 		for (i = 0; i < n; i++) {
607 			inum = le64_to_cpu(orph->inos[i]);
608 			dbg_rcvry("deleting orphaned inode %lu",
609 				  (unsigned long)inum);
610 			err = ubifs_tnc_remove_ino(c, inum);
611 			if (err)
612 				return err;
613 			err = insert_dead_orphan(c, inum);
614 			if (err)
615 				return err;
616 		}
617 
618 		*last_cmt_no = cmt_no;
619 		if (le64_to_cpu(orph->cmt_no) & (1ULL << 63)) {
620 			dbg_rcvry("last orph node for commit %llu at %d:%d",
621 				  cmt_no, sleb->lnum, snod->offs);
622 			*last_flagged = 1;
623 		} else
624 			*last_flagged = 0;
625 	}
626 
627 	return 0;
628 }
629 
630 /**
631  * kill_orphans - remove all orphan inodes from the index.
632  * @c: UBIFS file-system description object
633  *
634  * If recovery is required, then orphan inodes recorded during the previous
635  * session (which ended with an unclean unmount) must be deleted from the index.
636  * This is done by updating the TNC, but since the index is not updated until
637  * the next commit, the LEBs where the orphan information is recorded are not
638  * erased until the next commit.
639  */
640 static int kill_orphans(struct ubifs_info *c)
641 {
642 	unsigned long long last_cmt_no = 0;
643 	int lnum, err = 0, outofdate = 0, last_flagged = 0;
644 
645 	c->ohead_lnum = c->orph_first;
646 	c->ohead_offs = 0;
647 	/* Check no-orphans flag and skip this if no orphans */
648 	if (c->no_orphs) {
649 		dbg_rcvry("no orphans");
650 		return 0;
651 	}
652 	/*
653 	 * Orph nodes always start at c->orph_first and are written to each
654 	 * successive LEB in turn. Generally unused LEBs will have been unmapped
655 	 * but may contain out of date orphan nodes if the unmap didn't go
656 	 * through. In addition, the last orphan node written for each commit is
657 	 * marked (top bit of orph->cmt_no is set to 1). It is possible that
658 	 * there are orphan nodes from the next commit (i.e. the commit did not
659 	 * complete successfully). In that case, no orphans will have been lost
660 	 * due to the way that orphans are written, and any orphans added will
661 	 * be valid orphans anyway and so can be deleted.
662 	 */
663 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
664 		struct ubifs_scan_leb *sleb;
665 
666 		dbg_rcvry("LEB %d", lnum);
667 		sleb = ubifs_scan(c, lnum, 0, c->sbuf, 1);
668 		if (IS_ERR(sleb)) {
669 			if (PTR_ERR(sleb) == -EUCLEAN)
670 				sleb = ubifs_recover_leb(c, lnum, 0,
671 							 c->sbuf, -1);
672 			if (IS_ERR(sleb)) {
673 				err = PTR_ERR(sleb);
674 				break;
675 			}
676 		}
677 		err = do_kill_orphans(c, sleb, &last_cmt_no, &outofdate,
678 				      &last_flagged);
679 		if (err || outofdate) {
680 			ubifs_scan_destroy(sleb);
681 			break;
682 		}
683 		if (sleb->endpt) {
684 			c->ohead_lnum = lnum;
685 			c->ohead_offs = sleb->endpt;
686 		}
687 		ubifs_scan_destroy(sleb);
688 	}
689 	return err;
690 }
691 
692 /**
693  * ubifs_mount_orphans - delete orphan inodes and erase LEBs that recorded them.
694  * @c: UBIFS file-system description object
695  * @unclean: indicates recovery from unclean unmount
696  * @read_only: indicates read only mount
697  *
698  * This function is called when mounting to erase orphans from the previous
699  * session. If UBIFS was not unmounted cleanly, then the inodes recorded as
700  * orphans are deleted.
701  */
702 int ubifs_mount_orphans(struct ubifs_info *c, int unclean, int read_only)
703 {
704 	int err = 0;
705 
706 	c->max_orphans = tot_avail_orphs(c);
707 
708 	if (!read_only) {
709 		c->orph_buf = vmalloc(c->leb_size);
710 		if (!c->orph_buf)
711 			return -ENOMEM;
712 	}
713 
714 	if (unclean)
715 		err = kill_orphans(c);
716 	else if (!read_only)
717 		err = ubifs_clear_orphans(c);
718 
719 	return err;
720 }
721 
722 /*
723  * Everything below is related to debugging.
724  */
725 
726 struct check_orphan {
727 	struct rb_node rb;
728 	ino_t inum;
729 };
730 
731 struct check_info {
732 	unsigned long last_ino;
733 	unsigned long tot_inos;
734 	unsigned long missing;
735 	unsigned long long leaf_cnt;
736 	struct ubifs_ino_node *node;
737 	struct rb_root root;
738 };
739 
740 static int dbg_find_orphan(struct ubifs_info *c, ino_t inum)
741 {
742 	struct ubifs_orphan *o;
743 	struct rb_node *p;
744 
745 	spin_lock(&c->orphan_lock);
746 	p = c->orph_tree.rb_node;
747 	while (p) {
748 		o = rb_entry(p, struct ubifs_orphan, rb);
749 		if (inum < o->inum)
750 			p = p->rb_left;
751 		else if (inum > o->inum)
752 			p = p->rb_right;
753 		else {
754 			spin_unlock(&c->orphan_lock);
755 			return 1;
756 		}
757 	}
758 	spin_unlock(&c->orphan_lock);
759 	return 0;
760 }
761 
762 static int dbg_ins_check_orphan(struct rb_root *root, ino_t inum)
763 {
764 	struct check_orphan *orphan, *o;
765 	struct rb_node **p, *parent = NULL;
766 
767 	orphan = kzalloc(sizeof(struct check_orphan), GFP_NOFS);
768 	if (!orphan)
769 		return -ENOMEM;
770 	orphan->inum = inum;
771 
772 	p = &root->rb_node;
773 	while (*p) {
774 		parent = *p;
775 		o = rb_entry(parent, struct check_orphan, rb);
776 		if (inum < o->inum)
777 			p = &(*p)->rb_left;
778 		else if (inum > o->inum)
779 			p = &(*p)->rb_right;
780 		else {
781 			kfree(orphan);
782 			return 0;
783 		}
784 	}
785 	rb_link_node(&orphan->rb, parent, p);
786 	rb_insert_color(&orphan->rb, root);
787 	return 0;
788 }
789 
790 static int dbg_find_check_orphan(struct rb_root *root, ino_t inum)
791 {
792 	struct check_orphan *o;
793 	struct rb_node *p;
794 
795 	p = root->rb_node;
796 	while (p) {
797 		o = rb_entry(p, struct check_orphan, rb);
798 		if (inum < o->inum)
799 			p = p->rb_left;
800 		else if (inum > o->inum)
801 			p = p->rb_right;
802 		else
803 			return 1;
804 	}
805 	return 0;
806 }
807 
808 static void dbg_free_check_tree(struct rb_root *root)
809 {
810 	struct rb_node *this = root->rb_node;
811 	struct check_orphan *o;
812 
813 	while (this) {
814 		if (this->rb_left) {
815 			this = this->rb_left;
816 			continue;
817 		} else if (this->rb_right) {
818 			this = this->rb_right;
819 			continue;
820 		}
821 		o = rb_entry(this, struct check_orphan, rb);
822 		this = rb_parent(this);
823 		if (this) {
824 			if (this->rb_left == &o->rb)
825 				this->rb_left = NULL;
826 			else
827 				this->rb_right = NULL;
828 		}
829 		kfree(o);
830 	}
831 }
832 
833 static int dbg_orphan_check(struct ubifs_info *c, struct ubifs_zbranch *zbr,
834 			    void *priv)
835 {
836 	struct check_info *ci = priv;
837 	ino_t inum;
838 	int err;
839 
840 	inum = key_inum(c, &zbr->key);
841 	if (inum != ci->last_ino) {
842 		/* Lowest node type is the inode node, so it comes first */
843 		if (key_type(c, &zbr->key) != UBIFS_INO_KEY)
844 			ubifs_err("found orphan node ino %lu, type %d",
845 				  (unsigned long)inum, key_type(c, &zbr->key));
846 		ci->last_ino = inum;
847 		ci->tot_inos += 1;
848 		err = ubifs_tnc_read_node(c, zbr, ci->node);
849 		if (err) {
850 			ubifs_err("node read failed, error %d", err);
851 			return err;
852 		}
853 		if (ci->node->nlink == 0)
854 			/* Must be recorded as an orphan */
855 			if (!dbg_find_check_orphan(&ci->root, inum) &&
856 			    !dbg_find_orphan(c, inum)) {
857 				ubifs_err("missing orphan, ino %lu",
858 					  (unsigned long)inum);
859 				ci->missing += 1;
860 			}
861 	}
862 	ci->leaf_cnt += 1;
863 	return 0;
864 }
865 
866 static int dbg_read_orphans(struct check_info *ci, struct ubifs_scan_leb *sleb)
867 {
868 	struct ubifs_scan_node *snod;
869 	struct ubifs_orph_node *orph;
870 	ino_t inum;
871 	int i, n, err;
872 
873 	list_for_each_entry(snod, &sleb->nodes, list) {
874 		cond_resched();
875 		if (snod->type != UBIFS_ORPH_NODE)
876 			continue;
877 		orph = snod->node;
878 		n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
879 		for (i = 0; i < n; i++) {
880 			inum = le64_to_cpu(orph->inos[i]);
881 			err = dbg_ins_check_orphan(&ci->root, inum);
882 			if (err)
883 				return err;
884 		}
885 	}
886 	return 0;
887 }
888 
889 static int dbg_scan_orphans(struct ubifs_info *c, struct check_info *ci)
890 {
891 	int lnum, err = 0;
892 	void *buf;
893 
894 	/* Check no-orphans flag and skip this if no orphans */
895 	if (c->no_orphs)
896 		return 0;
897 
898 	buf = __vmalloc(c->leb_size, GFP_NOFS, PAGE_KERNEL);
899 	if (!buf) {
900 		ubifs_err("cannot allocate memory to check orphans");
901 		return 0;
902 	}
903 
904 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
905 		struct ubifs_scan_leb *sleb;
906 
907 		sleb = ubifs_scan(c, lnum, 0, buf, 0);
908 		if (IS_ERR(sleb)) {
909 			err = PTR_ERR(sleb);
910 			break;
911 		}
912 
913 		err = dbg_read_orphans(ci, sleb);
914 		ubifs_scan_destroy(sleb);
915 		if (err)
916 			break;
917 	}
918 
919 	vfree(buf);
920 	return err;
921 }
922 
923 static int dbg_check_orphans(struct ubifs_info *c)
924 {
925 	struct check_info ci;
926 	int err;
927 
928 	if (!dbg_is_chk_orph(c))
929 		return 0;
930 
931 	ci.last_ino = 0;
932 	ci.tot_inos = 0;
933 	ci.missing  = 0;
934 	ci.leaf_cnt = 0;
935 	ci.root = RB_ROOT;
936 	ci.node = kmalloc(UBIFS_MAX_INO_NODE_SZ, GFP_NOFS);
937 	if (!ci.node) {
938 		ubifs_err("out of memory");
939 		return -ENOMEM;
940 	}
941 
942 	err = dbg_scan_orphans(c, &ci);
943 	if (err)
944 		goto out;
945 
946 	err = dbg_walk_index(c, &dbg_orphan_check, NULL, &ci);
947 	if (err) {
948 		ubifs_err("cannot scan TNC, error %d", err);
949 		goto out;
950 	}
951 
952 	if (ci.missing) {
953 		ubifs_err("%lu missing orphan(s)", ci.missing);
954 		err = -EINVAL;
955 		goto out;
956 	}
957 
958 	dbg_cmt("last inode number is %lu", ci.last_ino);
959 	dbg_cmt("total number of inodes is %lu", ci.tot_inos);
960 	dbg_cmt("total number of leaf nodes is %llu", ci.leaf_cnt);
961 
962 out:
963 	dbg_free_check_tree(&ci.root);
964 	kfree(ci.node);
965 	return err;
966 }
967