• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
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 
orphan_add(struct ubifs_info * c,ino_t inum,struct ubifs_orphan * parent_orphan)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 
lookup_orphan(struct ubifs_info * c,ino_t inum)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 
__orphan_drop(struct ubifs_info * c,struct ubifs_orphan * o)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 
orphan_delete(struct ubifs_info * c,struct ubifs_orphan * orph)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  */
ubifs_add_orphan(struct ubifs_info * c,ino_t inum)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 			kfree(pxent);
177 			return err;
178 		}
179 
180 		fname_name(&nm) = xent->name;
181 		fname_len(&nm) = le16_to_cpu(xent->nlen);
182 		xattr_inum = le64_to_cpu(xent->inum);
183 
184 		xattr_orphan = orphan_add(c, xattr_inum, orphan);
185 		if (IS_ERR(xattr_orphan)) {
186 			kfree(pxent);
187 			kfree(xent);
188 			return PTR_ERR(xattr_orphan);
189 		}
190 
191 		kfree(pxent);
192 		pxent = xent;
193 		key_read(c, &xent->key, &key);
194 	}
195 	kfree(pxent);
196 
197 	return 0;
198 }
199 
200 /**
201  * ubifs_delete_orphan - delete an orphan.
202  * @c: UBIFS file-system description object
203  * @inum: orphan inode number
204  *
205  * Delete an orphan. This function is called when an inode is deleted.
206  */
ubifs_delete_orphan(struct ubifs_info * c,ino_t inum)207 void ubifs_delete_orphan(struct ubifs_info *c, ino_t inum)
208 {
209 	struct ubifs_orphan *orph, *child_orph, *tmp_o;
210 
211 	spin_lock(&c->orphan_lock);
212 
213 	orph = lookup_orphan(c, inum);
214 	if (!orph) {
215 		spin_unlock(&c->orphan_lock);
216 		ubifs_err(c, "missing orphan ino %lu", (unsigned long)inum);
217 		dump_stack();
218 
219 		return;
220 	}
221 
222 	list_for_each_entry_safe(child_orph, tmp_o, &orph->child_list, child_list) {
223 		list_del(&child_orph->child_list);
224 		orphan_delete(c, child_orph);
225 	}
226 
227 	orphan_delete(c, orph);
228 
229 	spin_unlock(&c->orphan_lock);
230 }
231 
232 /**
233  * ubifs_orphan_start_commit - start commit of orphans.
234  * @c: UBIFS file-system description object
235  *
236  * Start commit of orphans.
237  */
ubifs_orphan_start_commit(struct ubifs_info * c)238 int ubifs_orphan_start_commit(struct ubifs_info *c)
239 {
240 	struct ubifs_orphan *orphan, **last;
241 
242 	spin_lock(&c->orphan_lock);
243 	last = &c->orph_cnext;
244 	list_for_each_entry(orphan, &c->orph_new, new_list) {
245 		ubifs_assert(c, orphan->new);
246 		ubifs_assert(c, !orphan->cmt);
247 		orphan->new = 0;
248 		orphan->cmt = 1;
249 		*last = orphan;
250 		last = &orphan->cnext;
251 	}
252 	*last = NULL;
253 	c->cmt_orphans = c->new_orphans;
254 	c->new_orphans = 0;
255 	dbg_cmt("%d orphans to commit", c->cmt_orphans);
256 	INIT_LIST_HEAD(&c->orph_new);
257 	if (c->tot_orphans == 0)
258 		c->no_orphs = 1;
259 	else
260 		c->no_orphs = 0;
261 	spin_unlock(&c->orphan_lock);
262 	return 0;
263 }
264 
265 /**
266  * avail_orphs - calculate available space.
267  * @c: UBIFS file-system description object
268  *
269  * This function returns the number of orphans that can be written in the
270  * available space.
271  */
avail_orphs(struct ubifs_info * c)272 static int avail_orphs(struct ubifs_info *c)
273 {
274 	int avail_lebs, avail, gap;
275 
276 	avail_lebs = c->orph_lebs - (c->ohead_lnum - c->orph_first) - 1;
277 	avail = avail_lebs *
278 	       ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
279 	gap = c->leb_size - c->ohead_offs;
280 	if (gap >= UBIFS_ORPH_NODE_SZ + sizeof(__le64))
281 		avail += (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
282 	return avail;
283 }
284 
285 /**
286  * tot_avail_orphs - calculate total space.
287  * @c: UBIFS file-system description object
288  *
289  * This function returns the number of orphans that can be written in half
290  * the total space. That leaves half the space for adding new orphans.
291  */
tot_avail_orphs(struct ubifs_info * c)292 static int tot_avail_orphs(struct ubifs_info *c)
293 {
294 	int avail_lebs, avail;
295 
296 	avail_lebs = c->orph_lebs;
297 	avail = avail_lebs *
298 	       ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
299 	return avail / 2;
300 }
301 
302 /**
303  * do_write_orph_node - write a node to the orphan head.
304  * @c: UBIFS file-system description object
305  * @len: length of node
306  * @atomic: write atomically
307  *
308  * This function writes a node to the orphan head from the orphan buffer. If
309  * %atomic is not zero, then the write is done atomically. On success, %0 is
310  * returned, otherwise a negative error code is returned.
311  */
do_write_orph_node(struct ubifs_info * c,int len,int atomic)312 static int do_write_orph_node(struct ubifs_info *c, int len, int atomic)
313 {
314 	int err = 0;
315 
316 	if (atomic) {
317 		ubifs_assert(c, c->ohead_offs == 0);
318 		ubifs_prepare_node(c, c->orph_buf, len, 1);
319 		len = ALIGN(len, c->min_io_size);
320 		err = ubifs_leb_change(c, c->ohead_lnum, c->orph_buf, len);
321 	} else {
322 		if (c->ohead_offs == 0) {
323 			/* Ensure LEB has been unmapped */
324 			err = ubifs_leb_unmap(c, c->ohead_lnum);
325 			if (err)
326 				return err;
327 		}
328 		err = ubifs_write_node(c, c->orph_buf, len, c->ohead_lnum,
329 				       c->ohead_offs);
330 	}
331 	return err;
332 }
333 
334 /**
335  * write_orph_node - write an orphan node.
336  * @c: UBIFS file-system description object
337  * @atomic: write atomically
338  *
339  * This function builds an orphan node from the cnext list and writes it to the
340  * orphan head. On success, %0 is returned, otherwise a negative error code
341  * is returned.
342  */
write_orph_node(struct ubifs_info * c,int atomic)343 static int write_orph_node(struct ubifs_info *c, int atomic)
344 {
345 	struct ubifs_orphan *orphan, *cnext;
346 	struct ubifs_orph_node *orph;
347 	int gap, err, len, cnt, i;
348 
349 	ubifs_assert(c, c->cmt_orphans > 0);
350 	gap = c->leb_size - c->ohead_offs;
351 	if (gap < UBIFS_ORPH_NODE_SZ + sizeof(__le64)) {
352 		c->ohead_lnum += 1;
353 		c->ohead_offs = 0;
354 		gap = c->leb_size;
355 		if (c->ohead_lnum > c->orph_last) {
356 			/*
357 			 * We limit the number of orphans so that this should
358 			 * never happen.
359 			 */
360 			ubifs_err(c, "out of space in orphan area");
361 			return -EINVAL;
362 		}
363 	}
364 	cnt = (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
365 	if (cnt > c->cmt_orphans)
366 		cnt = c->cmt_orphans;
367 	len = UBIFS_ORPH_NODE_SZ + cnt * sizeof(__le64);
368 	ubifs_assert(c, c->orph_buf);
369 	orph = c->orph_buf;
370 	orph->ch.node_type = UBIFS_ORPH_NODE;
371 	spin_lock(&c->orphan_lock);
372 	cnext = c->orph_cnext;
373 	for (i = 0; i < cnt; i++) {
374 		orphan = cnext;
375 		ubifs_assert(c, orphan->cmt);
376 		orph->inos[i] = cpu_to_le64(orphan->inum);
377 		orphan->cmt = 0;
378 		cnext = orphan->cnext;
379 		orphan->cnext = NULL;
380 	}
381 	c->orph_cnext = cnext;
382 	c->cmt_orphans -= cnt;
383 	spin_unlock(&c->orphan_lock);
384 	if (c->cmt_orphans)
385 		orph->cmt_no = cpu_to_le64(c->cmt_no);
386 	else
387 		/* Mark the last node of the commit */
388 		orph->cmt_no = cpu_to_le64((c->cmt_no) | (1ULL << 63));
389 	ubifs_assert(c, c->ohead_offs + len <= c->leb_size);
390 	ubifs_assert(c, c->ohead_lnum >= c->orph_first);
391 	ubifs_assert(c, c->ohead_lnum <= c->orph_last);
392 	err = do_write_orph_node(c, len, atomic);
393 	c->ohead_offs += ALIGN(len, c->min_io_size);
394 	c->ohead_offs = ALIGN(c->ohead_offs, 8);
395 	return err;
396 }
397 
398 /**
399  * write_orph_nodes - write orphan nodes until there are no more to commit.
400  * @c: UBIFS file-system description object
401  * @atomic: write atomically
402  *
403  * This function writes orphan nodes for all the orphans to commit. On success,
404  * %0 is returned, otherwise a negative error code is returned.
405  */
write_orph_nodes(struct ubifs_info * c,int atomic)406 static int write_orph_nodes(struct ubifs_info *c, int atomic)
407 {
408 	int err;
409 
410 	while (c->cmt_orphans > 0) {
411 		err = write_orph_node(c, atomic);
412 		if (err)
413 			return err;
414 	}
415 	if (atomic) {
416 		int lnum;
417 
418 		/* Unmap any unused LEBs after consolidation */
419 		for (lnum = c->ohead_lnum + 1; lnum <= c->orph_last; lnum++) {
420 			err = ubifs_leb_unmap(c, lnum);
421 			if (err)
422 				return err;
423 		}
424 	}
425 	return 0;
426 }
427 
428 /**
429  * consolidate - consolidate the orphan area.
430  * @c: UBIFS file-system description object
431  *
432  * This function enables consolidation by putting all the orphans into the list
433  * to commit. The list is in the order that the orphans were added, and the
434  * LEBs are written atomically in order, so at no time can orphans be lost by
435  * an unclean unmount.
436  *
437  * This function returns %0 on success and a negative error code on failure.
438  */
consolidate(struct ubifs_info * c)439 static int consolidate(struct ubifs_info *c)
440 {
441 	int tot_avail = tot_avail_orphs(c), err = 0;
442 
443 	spin_lock(&c->orphan_lock);
444 	dbg_cmt("there is space for %d orphans and there are %d",
445 		tot_avail, c->tot_orphans);
446 	if (c->tot_orphans - c->new_orphans <= tot_avail) {
447 		struct ubifs_orphan *orphan, **last;
448 		int cnt = 0;
449 
450 		/* Change the cnext list to include all non-new orphans */
451 		last = &c->orph_cnext;
452 		list_for_each_entry(orphan, &c->orph_list, list) {
453 			if (orphan->new)
454 				continue;
455 			orphan->cmt = 1;
456 			*last = orphan;
457 			last = &orphan->cnext;
458 			cnt += 1;
459 		}
460 		*last = NULL;
461 		ubifs_assert(c, cnt == c->tot_orphans - c->new_orphans);
462 		c->cmt_orphans = cnt;
463 		c->ohead_lnum = c->orph_first;
464 		c->ohead_offs = 0;
465 	} else {
466 		/*
467 		 * We limit the number of orphans so that this should
468 		 * never happen.
469 		 */
470 		ubifs_err(c, "out of space in orphan area");
471 		err = -EINVAL;
472 	}
473 	spin_unlock(&c->orphan_lock);
474 	return err;
475 }
476 
477 /**
478  * commit_orphans - commit orphans.
479  * @c: UBIFS file-system description object
480  *
481  * This function commits orphans to flash. On success, %0 is returned,
482  * otherwise a negative error code is returned.
483  */
commit_orphans(struct ubifs_info * c)484 static int commit_orphans(struct ubifs_info *c)
485 {
486 	int avail, atomic = 0, err;
487 
488 	ubifs_assert(c, c->cmt_orphans > 0);
489 	avail = avail_orphs(c);
490 	if (avail < c->cmt_orphans) {
491 		/* Not enough space to write new orphans, so consolidate */
492 		err = consolidate(c);
493 		if (err)
494 			return err;
495 		atomic = 1;
496 	}
497 	err = write_orph_nodes(c, atomic);
498 	return err;
499 }
500 
501 /**
502  * erase_deleted - erase the orphans marked for deletion.
503  * @c: UBIFS file-system description object
504  *
505  * During commit, the orphans being committed cannot be deleted, so they are
506  * marked for deletion and deleted by this function. Also, the recovery
507  * adds killed orphans to the deletion list, and therefore they are deleted
508  * here too.
509  */
erase_deleted(struct ubifs_info * c)510 static void erase_deleted(struct ubifs_info *c)
511 {
512 	struct ubifs_orphan *orphan, *dnext;
513 
514 	spin_lock(&c->orphan_lock);
515 	dnext = c->orph_dnext;
516 	while (dnext) {
517 		orphan = dnext;
518 		dnext = orphan->dnext;
519 		ubifs_assert(c, !orphan->new);
520 		ubifs_assert(c, orphan->del);
521 		rb_erase(&orphan->rb, &c->orph_tree);
522 		list_del(&orphan->list);
523 		c->tot_orphans -= 1;
524 		dbg_gen("deleting orphan ino %lu", (unsigned long)orphan->inum);
525 		kfree(orphan);
526 	}
527 	c->orph_dnext = NULL;
528 	spin_unlock(&c->orphan_lock);
529 }
530 
531 /**
532  * ubifs_orphan_end_commit - end commit of orphans.
533  * @c: UBIFS file-system description object
534  *
535  * End commit of orphans.
536  */
ubifs_orphan_end_commit(struct ubifs_info * c)537 int ubifs_orphan_end_commit(struct ubifs_info *c)
538 {
539 	int err;
540 
541 	if (c->cmt_orphans != 0) {
542 		err = commit_orphans(c);
543 		if (err)
544 			return err;
545 	}
546 	erase_deleted(c);
547 	err = dbg_check_orphans(c);
548 	return err;
549 }
550 
551 /**
552  * ubifs_clear_orphans - erase all LEBs used for orphans.
553  * @c: UBIFS file-system description object
554  *
555  * If recovery is not required, then the orphans from the previous session
556  * are not needed. This function locates the LEBs used to record
557  * orphans, and un-maps them.
558  */
ubifs_clear_orphans(struct ubifs_info * c)559 int ubifs_clear_orphans(struct ubifs_info *c)
560 {
561 	int lnum, err;
562 
563 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
564 		err = ubifs_leb_unmap(c, lnum);
565 		if (err)
566 			return err;
567 	}
568 	c->ohead_lnum = c->orph_first;
569 	c->ohead_offs = 0;
570 	return 0;
571 }
572 
573 /**
574  * insert_dead_orphan - insert an orphan.
575  * @c: UBIFS file-system description object
576  * @inum: orphan inode number
577  *
578  * This function is a helper to the 'do_kill_orphans()' function. The orphan
579  * must be kept until the next commit, so it is added to the rb-tree and the
580  * deletion list.
581  */
insert_dead_orphan(struct ubifs_info * c,ino_t inum)582 static int insert_dead_orphan(struct ubifs_info *c, ino_t inum)
583 {
584 	struct ubifs_orphan *orphan, *o;
585 	struct rb_node **p, *parent = NULL;
586 
587 	orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_KERNEL);
588 	if (!orphan)
589 		return -ENOMEM;
590 	orphan->inum = inum;
591 
592 	p = &c->orph_tree.rb_node;
593 	while (*p) {
594 		parent = *p;
595 		o = rb_entry(parent, struct ubifs_orphan, rb);
596 		if (inum < o->inum)
597 			p = &(*p)->rb_left;
598 		else if (inum > o->inum)
599 			p = &(*p)->rb_right;
600 		else {
601 			/* Already added - no problem */
602 			kfree(orphan);
603 			return 0;
604 		}
605 	}
606 	c->tot_orphans += 1;
607 	rb_link_node(&orphan->rb, parent, p);
608 	rb_insert_color(&orphan->rb, &c->orph_tree);
609 	list_add_tail(&orphan->list, &c->orph_list);
610 	orphan->del = 1;
611 	orphan->dnext = c->orph_dnext;
612 	c->orph_dnext = orphan;
613 	dbg_mnt("ino %lu, new %d, tot %d", (unsigned long)inum,
614 		c->new_orphans, c->tot_orphans);
615 	return 0;
616 }
617 
618 /**
619  * do_kill_orphans - remove orphan inodes from the index.
620  * @c: UBIFS file-system description object
621  * @sleb: scanned LEB
622  * @last_cmt_no: cmt_no of last orphan node read is passed and returned here
623  * @outofdate: whether the LEB is out of date is returned here
624  * @last_flagged: whether the end orphan node is encountered
625  *
626  * This function is a helper to the 'kill_orphans()' function. It goes through
627  * every orphan node in a LEB and for every inode number recorded, removes
628  * all keys for that inode from the TNC.
629  */
do_kill_orphans(struct ubifs_info * c,struct ubifs_scan_leb * sleb,unsigned long long * last_cmt_no,int * outofdate,int * last_flagged)630 static int do_kill_orphans(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
631 			   unsigned long long *last_cmt_no, int *outofdate,
632 			   int *last_flagged)
633 {
634 	struct ubifs_scan_node *snod;
635 	struct ubifs_orph_node *orph;
636 	struct ubifs_ino_node *ino = NULL;
637 	unsigned long long cmt_no;
638 	ino_t inum;
639 	int i, n, err, first = 1;
640 
641 	ino = kmalloc(UBIFS_MAX_INO_NODE_SZ, GFP_NOFS);
642 	if (!ino)
643 		return -ENOMEM;
644 
645 	list_for_each_entry(snod, &sleb->nodes, list) {
646 		if (snod->type != UBIFS_ORPH_NODE) {
647 			ubifs_err(c, "invalid node type %d in orphan area at %d:%d",
648 				  snod->type, sleb->lnum, snod->offs);
649 			ubifs_dump_node(c, snod->node,
650 					c->leb_size - snod->offs);
651 			err = -EINVAL;
652 			goto out_free;
653 		}
654 
655 		orph = snod->node;
656 
657 		/* Check commit number */
658 		cmt_no = le64_to_cpu(orph->cmt_no) & LLONG_MAX;
659 		/*
660 		 * The commit number on the master node may be less, because
661 		 * of a failed commit. If there are several failed commits in a
662 		 * row, the commit number written on orphan nodes will continue
663 		 * to increase (because the commit number is adjusted here) even
664 		 * though the commit number on the master node stays the same
665 		 * because the master node has not been re-written.
666 		 */
667 		if (cmt_no > c->cmt_no)
668 			c->cmt_no = cmt_no;
669 		if (cmt_no < *last_cmt_no && *last_flagged) {
670 			/*
671 			 * The last orphan node had a higher commit number and
672 			 * was flagged as the last written for that commit
673 			 * number. That makes this orphan node, out of date.
674 			 */
675 			if (!first) {
676 				ubifs_err(c, "out of order commit number %llu in orphan node at %d:%d",
677 					  cmt_no, sleb->lnum, snod->offs);
678 				ubifs_dump_node(c, snod->node,
679 						c->leb_size - snod->offs);
680 				err = -EINVAL;
681 				goto out_free;
682 			}
683 			dbg_rcvry("out of date LEB %d", sleb->lnum);
684 			*outofdate = 1;
685 			err = 0;
686 			goto out_free;
687 		}
688 
689 		if (first)
690 			first = 0;
691 
692 		n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
693 		for (i = 0; i < n; i++) {
694 			union ubifs_key key1, key2;
695 
696 			inum = le64_to_cpu(orph->inos[i]);
697 
698 			ino_key_init(c, &key1, inum);
699 			err = ubifs_tnc_lookup(c, &key1, ino);
700 			if (err && err != -ENOENT)
701 				goto out_free;
702 
703 			/*
704 			 * Check whether an inode can really get deleted.
705 			 * linkat() with O_TMPFILE allows rebirth of an inode.
706 			 */
707 			if (err == 0 && ino->nlink == 0) {
708 				dbg_rcvry("deleting orphaned inode %lu",
709 					  (unsigned long)inum);
710 
711 				lowest_ino_key(c, &key1, inum);
712 				highest_ino_key(c, &key2, inum);
713 
714 				err = ubifs_tnc_remove_range(c, &key1, &key2);
715 				if (err)
716 					goto out_ro;
717 			}
718 
719 			err = insert_dead_orphan(c, inum);
720 			if (err)
721 				goto out_free;
722 		}
723 
724 		*last_cmt_no = cmt_no;
725 		if (le64_to_cpu(orph->cmt_no) & (1ULL << 63)) {
726 			dbg_rcvry("last orph node for commit %llu at %d:%d",
727 				  cmt_no, sleb->lnum, snod->offs);
728 			*last_flagged = 1;
729 		} else
730 			*last_flagged = 0;
731 	}
732 
733 	err = 0;
734 out_free:
735 	kfree(ino);
736 	return err;
737 
738 out_ro:
739 	ubifs_ro_mode(c, err);
740 	kfree(ino);
741 	return err;
742 }
743 
744 /**
745  * kill_orphans - remove all orphan inodes from the index.
746  * @c: UBIFS file-system description object
747  *
748  * If recovery is required, then orphan inodes recorded during the previous
749  * session (which ended with an unclean unmount) must be deleted from the index.
750  * This is done by updating the TNC, but since the index is not updated until
751  * the next commit, the LEBs where the orphan information is recorded are not
752  * erased until the next commit.
753  */
kill_orphans(struct ubifs_info * c)754 static int kill_orphans(struct ubifs_info *c)
755 {
756 	unsigned long long last_cmt_no = 0;
757 	int lnum, err = 0, outofdate = 0, last_flagged = 0;
758 
759 	c->ohead_lnum = c->orph_first;
760 	c->ohead_offs = 0;
761 	/* Check no-orphans flag and skip this if no orphans */
762 	if (c->no_orphs) {
763 		dbg_rcvry("no orphans");
764 		return 0;
765 	}
766 	/*
767 	 * Orph nodes always start at c->orph_first and are written to each
768 	 * successive LEB in turn. Generally unused LEBs will have been unmapped
769 	 * but may contain out of date orphan nodes if the unmap didn't go
770 	 * through. In addition, the last orphan node written for each commit is
771 	 * marked (top bit of orph->cmt_no is set to 1). It is possible that
772 	 * there are orphan nodes from the next commit (i.e. the commit did not
773 	 * complete successfully). In that case, no orphans will have been lost
774 	 * due to the way that orphans are written, and any orphans added will
775 	 * be valid orphans anyway and so can be deleted.
776 	 */
777 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
778 		struct ubifs_scan_leb *sleb;
779 
780 		dbg_rcvry("LEB %d", lnum);
781 		sleb = ubifs_scan(c, lnum, 0, c->sbuf, 1);
782 		if (IS_ERR(sleb)) {
783 			if (PTR_ERR(sleb) == -EUCLEAN)
784 				sleb = ubifs_recover_leb(c, lnum, 0,
785 							 c->sbuf, -1);
786 			if (IS_ERR(sleb)) {
787 				err = PTR_ERR(sleb);
788 				break;
789 			}
790 		}
791 		err = do_kill_orphans(c, sleb, &last_cmt_no, &outofdate,
792 				      &last_flagged);
793 		if (err || outofdate) {
794 			ubifs_scan_destroy(sleb);
795 			break;
796 		}
797 		if (sleb->endpt) {
798 			c->ohead_lnum = lnum;
799 			c->ohead_offs = sleb->endpt;
800 		}
801 		ubifs_scan_destroy(sleb);
802 	}
803 	return err;
804 }
805 
806 /**
807  * ubifs_mount_orphans - delete orphan inodes and erase LEBs that recorded them.
808  * @c: UBIFS file-system description object
809  * @unclean: indicates recovery from unclean unmount
810  * @read_only: indicates read only mount
811  *
812  * This function is called when mounting to erase orphans from the previous
813  * session. If UBIFS was not unmounted cleanly, then the inodes recorded as
814  * orphans are deleted.
815  */
ubifs_mount_orphans(struct ubifs_info * c,int unclean,int read_only)816 int ubifs_mount_orphans(struct ubifs_info *c, int unclean, int read_only)
817 {
818 	int err = 0;
819 
820 	c->max_orphans = tot_avail_orphs(c);
821 
822 	if (!read_only) {
823 		c->orph_buf = vmalloc(c->leb_size);
824 		if (!c->orph_buf)
825 			return -ENOMEM;
826 	}
827 
828 	if (unclean)
829 		err = kill_orphans(c);
830 	else if (!read_only)
831 		err = ubifs_clear_orphans(c);
832 
833 	return err;
834 }
835 
836 /*
837  * Everything below is related to debugging.
838  */
839 
840 struct check_orphan {
841 	struct rb_node rb;
842 	ino_t inum;
843 };
844 
845 struct check_info {
846 	unsigned long last_ino;
847 	unsigned long tot_inos;
848 	unsigned long missing;
849 	unsigned long long leaf_cnt;
850 	struct ubifs_ino_node *node;
851 	struct rb_root root;
852 };
853 
dbg_find_orphan(struct ubifs_info * c,ino_t inum)854 static bool dbg_find_orphan(struct ubifs_info *c, ino_t inum)
855 {
856 	bool found = false;
857 
858 	spin_lock(&c->orphan_lock);
859 	found = !!lookup_orphan(c, inum);
860 	spin_unlock(&c->orphan_lock);
861 
862 	return found;
863 }
864 
dbg_ins_check_orphan(struct rb_root * root,ino_t inum)865 static int dbg_ins_check_orphan(struct rb_root *root, ino_t inum)
866 {
867 	struct check_orphan *orphan, *o;
868 	struct rb_node **p, *parent = NULL;
869 
870 	orphan = kzalloc(sizeof(struct check_orphan), GFP_NOFS);
871 	if (!orphan)
872 		return -ENOMEM;
873 	orphan->inum = inum;
874 
875 	p = &root->rb_node;
876 	while (*p) {
877 		parent = *p;
878 		o = rb_entry(parent, struct check_orphan, rb);
879 		if (inum < o->inum)
880 			p = &(*p)->rb_left;
881 		else if (inum > o->inum)
882 			p = &(*p)->rb_right;
883 		else {
884 			kfree(orphan);
885 			return 0;
886 		}
887 	}
888 	rb_link_node(&orphan->rb, parent, p);
889 	rb_insert_color(&orphan->rb, root);
890 	return 0;
891 }
892 
dbg_find_check_orphan(struct rb_root * root,ino_t inum)893 static int dbg_find_check_orphan(struct rb_root *root, ino_t inum)
894 {
895 	struct check_orphan *o;
896 	struct rb_node *p;
897 
898 	p = root->rb_node;
899 	while (p) {
900 		o = rb_entry(p, struct check_orphan, rb);
901 		if (inum < o->inum)
902 			p = p->rb_left;
903 		else if (inum > o->inum)
904 			p = p->rb_right;
905 		else
906 			return 1;
907 	}
908 	return 0;
909 }
910 
dbg_free_check_tree(struct rb_root * root)911 static void dbg_free_check_tree(struct rb_root *root)
912 {
913 	struct check_orphan *o, *n;
914 
915 	rbtree_postorder_for_each_entry_safe(o, n, root, rb)
916 		kfree(o);
917 }
918 
dbg_orphan_check(struct ubifs_info * c,struct ubifs_zbranch * zbr,void * priv)919 static int dbg_orphan_check(struct ubifs_info *c, struct ubifs_zbranch *zbr,
920 			    void *priv)
921 {
922 	struct check_info *ci = priv;
923 	ino_t inum;
924 	int err;
925 
926 	inum = key_inum(c, &zbr->key);
927 	if (inum != ci->last_ino) {
928 		/* Lowest node type is the inode node, so it comes first */
929 		if (key_type(c, &zbr->key) != UBIFS_INO_KEY)
930 			ubifs_err(c, "found orphan node ino %lu, type %d",
931 				  (unsigned long)inum, key_type(c, &zbr->key));
932 		ci->last_ino = inum;
933 		ci->tot_inos += 1;
934 		err = ubifs_tnc_read_node(c, zbr, ci->node);
935 		if (err) {
936 			ubifs_err(c, "node read failed, error %d", err);
937 			return err;
938 		}
939 		if (ci->node->nlink == 0)
940 			/* Must be recorded as an orphan */
941 			if (!dbg_find_check_orphan(&ci->root, inum) &&
942 			    !dbg_find_orphan(c, inum)) {
943 				ubifs_err(c, "missing orphan, ino %lu",
944 					  (unsigned long)inum);
945 				ci->missing += 1;
946 			}
947 	}
948 	ci->leaf_cnt += 1;
949 	return 0;
950 }
951 
dbg_read_orphans(struct check_info * ci,struct ubifs_scan_leb * sleb)952 static int dbg_read_orphans(struct check_info *ci, struct ubifs_scan_leb *sleb)
953 {
954 	struct ubifs_scan_node *snod;
955 	struct ubifs_orph_node *orph;
956 	ino_t inum;
957 	int i, n, err;
958 
959 	list_for_each_entry(snod, &sleb->nodes, list) {
960 		cond_resched();
961 		if (snod->type != UBIFS_ORPH_NODE)
962 			continue;
963 		orph = snod->node;
964 		n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
965 		for (i = 0; i < n; i++) {
966 			inum = le64_to_cpu(orph->inos[i]);
967 			err = dbg_ins_check_orphan(&ci->root, inum);
968 			if (err)
969 				return err;
970 		}
971 	}
972 	return 0;
973 }
974 
dbg_scan_orphans(struct ubifs_info * c,struct check_info * ci)975 static int dbg_scan_orphans(struct ubifs_info *c, struct check_info *ci)
976 {
977 	int lnum, err = 0;
978 	void *buf;
979 
980 	/* Check no-orphans flag and skip this if no orphans */
981 	if (c->no_orphs)
982 		return 0;
983 
984 	buf = __vmalloc(c->leb_size, GFP_NOFS);
985 	if (!buf) {
986 		ubifs_err(c, "cannot allocate memory to check orphans");
987 		return 0;
988 	}
989 
990 	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
991 		struct ubifs_scan_leb *sleb;
992 
993 		sleb = ubifs_scan(c, lnum, 0, buf, 0);
994 		if (IS_ERR(sleb)) {
995 			err = PTR_ERR(sleb);
996 			break;
997 		}
998 
999 		err = dbg_read_orphans(ci, sleb);
1000 		ubifs_scan_destroy(sleb);
1001 		if (err)
1002 			break;
1003 	}
1004 
1005 	vfree(buf);
1006 	return err;
1007 }
1008 
dbg_check_orphans(struct ubifs_info * c)1009 static int dbg_check_orphans(struct ubifs_info *c)
1010 {
1011 	struct check_info ci;
1012 	int err;
1013 
1014 	if (!dbg_is_chk_orph(c))
1015 		return 0;
1016 
1017 	ci.last_ino = 0;
1018 	ci.tot_inos = 0;
1019 	ci.missing  = 0;
1020 	ci.leaf_cnt = 0;
1021 	ci.root = RB_ROOT;
1022 	ci.node = kmalloc(UBIFS_MAX_INO_NODE_SZ, GFP_NOFS);
1023 	if (!ci.node) {
1024 		ubifs_err(c, "out of memory");
1025 		return -ENOMEM;
1026 	}
1027 
1028 	err = dbg_scan_orphans(c, &ci);
1029 	if (err)
1030 		goto out;
1031 
1032 	err = dbg_walk_index(c, &dbg_orphan_check, NULL, &ci);
1033 	if (err) {
1034 		ubifs_err(c, "cannot scan TNC, error %d", err);
1035 		goto out;
1036 	}
1037 
1038 	if (ci.missing) {
1039 		ubifs_err(c, "%lu missing orphan(s)", ci.missing);
1040 		err = -EINVAL;
1041 		goto out;
1042 	}
1043 
1044 	dbg_cmt("last inode number is %lu", ci.last_ino);
1045 	dbg_cmt("total number of inodes is %lu", ci.tot_inos);
1046 	dbg_cmt("total number of leaf nodes is %llu", ci.leaf_cnt);
1047 
1048 out:
1049 	dbg_free_check_tree(&ci.root);
1050 	kfree(ci.node);
1051 	return err;
1052 }
1053