summaryrefslogtreecommitdiff
path: root/fs/jbd2/recovery.c
blob: 52054a83e717ff3dbd37f7bf4853137ca0aac393 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
/*
 * linux/fs/recovery.c
 *
 * Written by Stephen C. Tweedie <sct@redhat.com>, 1999
 *
 * Copyright 1999-2000 Red Hat Software --- All Rights Reserved
 *
 * This file is part of the Linux kernel and is made available under
 * the terms of the GNU General Public License, version 2, or at your
 * option, any later version, incorporated herein by reference.
 *
 * Journal recovery routines for the generic filesystem journaling code;
 * part of the ext2fs journaling system.
 */

#ifndef __KERNEL__
#include "jfs_user.h"
#else
#include <linux/time.h>
#include <linux/fs.h>
#include <linux/jbd2.h>
#include <linux/errno.h>
#include <linux/slab.h>
#endif

/*
 * Maintain information about the progress of the recovery job, so that
 * the different passes can carry information between them.
 */
struct recovery_info
{
	tid_t		start_transaction;
	tid_t		end_transaction;

	int		nr_replays;
	int		nr_revokes;
	int		nr_revoke_hits;
};

enum passtype {PASS_SCAN, PASS_REVOKE, PASS_REPLAY};
static int do_one_pass(journal_t *journal,
				struct recovery_info *info, enum passtype pass);
static int scan_revoke_records(journal_t *, struct buffer_head *,
				tid_t, struct recovery_info *);

#ifdef __KERNEL__

/* Release readahead buffers after use */
static void journal_brelse_array(struct buffer_head *b[], int n)
{
	while (--n >= 0)
		brelse (b[n]);
}


/*
 * When reading from the journal, we are going through the block device
 * layer directly and so there is no readahead being done for us.  We
 * need to implement any readahead ourselves if we want it to happen at
 * all.  Recovery is basically one long sequential read, so make sure we
 * do the IO in reasonably large chunks.
 *
 * This is not so critical that we need to be enormously clever about
 * the readahead size, though.  128K is a purely arbitrary, good-enough
 * fixed value.
 */

#define MAXBUF 8
static int do_readahead(journal_t *journal, unsigned int start)
{
	int err;
	unsigned int max, nbufs, next;
	sector_t blocknr;
	struct buffer_head *bh;

	struct buffer_head * bufs[MAXBUF];

	/* Do up to 128K of readahead */
	max = start + (128 * 1024 / journal->j_blocksize);
	if (max > journal->j_maxlen)
		max = journal->j_maxlen;

	/* Do the readahead itself.  We'll submit MAXBUF buffer_heads at
	 * a time to the block device IO layer. */

	nbufs = 0;

	for (next = start; next < max; next++) {
		err = jbd2_journal_bmap(journal, next, &blocknr);

		if (err) {
			printk (KERN_ERR "JBD: bad block at offset %u\n",
				next);
			goto failed;
		}

		bh = __getblk(journal->j_dev, blocknr, journal->j_blocksize);
		if (!bh) {
			err = -ENOMEM;
			goto failed;
		}

		if (!buffer_uptodate(bh) && !buffer_locked(bh)) {
			bufs[nbufs++] = bh;
			if (nbufs == MAXBUF) {
				ll_rw_block(READ, nbufs, bufs);
				journal_brelse_array(bufs, nbufs);
				nbufs = 0;
			}
		} else
			brelse(bh);
	}

	if (nbufs)
		ll_rw_block(READ, nbufs, bufs);
	err = 0;

failed:
	if (nbufs)
		journal_brelse_array(bufs, nbufs);
	return err;
}

#endif /* __KERNEL__ */


/*
 * Read a block from the journal
 */

static int jread(struct buffer_head **bhp, journal_t *journal,
		 unsigned int offset)
{
	int err;
	sector_t blocknr;
	struct buffer_head *bh;

	*bhp = NULL;

	if (offset >= journal->j_maxlen) {
		printk(KERN_ERR "JBD: corrupted journal superblock\n");
		return -EIO;
	}

	err = jbd2_journal_bmap(journal, offset, &blocknr);

	if (err) {
		printk (KERN_ERR "JBD: bad block at offset %u\n",
			offset);
		return err;
	}

	bh = __getblk(journal->j_dev, blocknr, journal->j_blocksize);
	if (!bh)
		return -ENOMEM;

	if (!buffer_uptodate(bh)) {
		/* If this is a brand new buffer, start readahead.
                   Otherwise, we assume we are already reading it.  */
		if (!buffer_req(bh))
			do_readahead(journal, offset);
		wait_on_buffer(bh);
	}

	if (!buffer_uptodate(bh)) {
		printk (KERN_ERR "JBD: Failed to read block at offset %u\n",
			offset);
		brelse(bh);
		return -EIO;
	}

	*bhp = bh;
	return 0;
}


/*
 * Count the number of in-use tags in a journal descriptor block.
 */

static int count_tags(journal_t *journal, struct buffer_head *bh)
{
	char *			tagp;
	journal_block_tag_t *	tag;
	int			nr = 0, size = journal->j_blocksize;
	int			tag_bytes = journal_tag_bytes(journal);

	tagp = &bh->b_data[sizeof(journal_header_t)];

	while ((tagp - bh->b_data + tag_bytes) <= size) {
		tag = (journal_block_tag_t *) tagp;

		nr++;
		tagp += tag_bytes;
		if (!(tag->t_flags & cpu_to_be32(JBD2_FLAG_SAME_UUID)))
			tagp += 16;

		if (tag->t_flags & cpu_to_be32(JBD2_FLAG_LAST_TAG))
			break;
	}

	return nr;
}


/* Make sure we wrap around the log correctly! */
#define wrap(journal, var)						\
do {									\
	if (var >= (journal)->j_last)					\
		var -= ((journal)->j_last - (journal)->j_first);	\
} while (0)

/**
 * jbd2_journal_recover - recovers a on-disk journal
 * @journal: the journal to recover
 *
 * The primary function for recovering the log contents when mounting a
 * journaled device.
 *
 * Recovery is done in three passes.  In the first pass, we look for the
 * end of the log.  In the second, we assemble the list of revoke
 * blocks.  In the third and final pass, we replay any un-revoked blocks
 * in the log.
 */
int jbd2_journal_recover(journal_t *journal)
{
	int			err;
	journal_superblock_t *	sb;

	struct recovery_info	info;

	memset(&info, 0, sizeof(info));
	sb = journal->j_superblock;

	/*
	 * The journal superblock's s_start field (the current log head)
	 * is always zero if, and only if, the journal was cleanly
	 * unmounted.
	 */

	if (!sb->s_start) {
		jbd_debug(1, "No recovery required, last transaction %d\n",
			  be32_to_cpu(sb->s_sequence));
		journal->j_transaction_sequence = be32_to_cpu(sb->s_sequence) + 1;
		return 0;
	}

	err = do_one_pass(journal, &info, PASS_SCAN);
	if (!err)
		err = do_one_pass(journal, &info, PASS_REVOKE);
	if (!err)
		err = do_one_pass(journal, &info, PASS_REPLAY);

	jbd_debug(0, "JBD: recovery, exit status %d, "
		  "recovered transactions %u to %u\n",
		  err, info.start_transaction, info.end_transaction);
	jbd_debug(0, "JBD: Replayed %d and revoked %d/%d blocks\n",
		  info.nr_replays, info.nr_revoke_hits, info.nr_revokes);

	/* Restart the log at the next transaction ID, thus invalidating
	 * any existing commit records in the log. */
	journal->j_transaction_sequence = ++info.end_transaction;

	jbd2_journal_clear_revoke(journal);
	sync_blockdev(journal->j_fs_dev);
	return err;
}

/**
 * jbd2_journal_skip_recovery - Start journal and wipe exiting records
 * @journal: journal to startup
 *
 * Locate any valid recovery information from the journal and set up the
 * journal structures in memory to ignore it (presumably because the
 * caller has evidence that it is out of date).
 * This function does'nt appear to be exorted..
 *
 * We perform one pass over the journal to allow us to tell the user how
 * much recovery information is being erased, and to let us initialise
 * the journal transaction sequence numbers to the next unused ID.
 */
int jbd2_journal_skip_recovery(journal_t *journal)
{
	int			err;
	journal_superblock_t *	sb;

	struct recovery_info	info;

	memset (&info, 0, sizeof(info));
	sb = journal->j_superblock;

	err = do_one_pass(journal, &info, PASS_SCAN);

	if (err) {
		printk(KERN_ERR "JBD: error %d scanning journal\n", err);
		++journal->j_transaction_sequence;
	} else {
#ifdef CONFIG_JBD_DEBUG
		int dropped = info.end_transaction - be32_to_cpu(sb->s_sequence);
#endif
		jbd_debug(0,
			  "JBD: ignoring %d transaction%s from the journal.\n",
			  dropped, (dropped == 1) ? "" : "s");
		journal->j_transaction_sequence = ++info.end_transaction;
	}

	journal->j_tail = 0;
	return err;
}

static inline sector_t read_tag_block(int tag_bytes, journal_block_tag_t *tag)
{
	sector_t block = be32_to_cpu(tag->t_blocknr);
	if (tag_bytes > JBD_TAG_SIZE32)
		block |= (u64)be32_to_cpu(tag->t_blocknr_high) << 32;
	return block;
}

static int do_one_pass(journal_t *journal,
			struct recovery_info *info, enum passtype pass)
{
	unsigned int		first_commit_ID, next_commit_ID;
	unsigned long		next_log_block;
	int			err, success = 0;
	journal_superblock_t *	sb;
	journal_header_t *	tmp;
	struct buffer_head *	bh;
	unsigned int		sequence;
	int			blocktype;
	int			tag_bytes = journal_tag_bytes(journal);

	/* Precompute the maximum metadata descriptors in a descriptor block */
	int			MAX_BLOCKS_PER_DESC;
	MAX_BLOCKS_PER_DESC = ((journal->j_blocksize-sizeof(journal_header_t))
			       / tag_bytes);

	/*
	 * First thing is to establish what we expect to find in the log
	 * (in terms of transaction IDs), and where (in terms of log
	 * block offsets): query the superblock.
	 */

	sb = journal->j_superblock;
	next_commit_ID = be32_to_cpu(sb->s_sequence);
	next_log_block = be32_to_cpu(sb->s_start);

	first_commit_ID = next_commit_ID;
	if (pass == PASS_SCAN)
		info->start_transaction = first_commit_ID;

	jbd_debug(1, "Starting recovery pass %d\n", pass);

	/*
	 * Now we walk through the log, transaction by transaction,
	 * making sure that each transaction has a commit block in the
	 * expected place.  Each complete transaction gets replayed back
	 * into the main filesystem.
	 */

	while (1) {
		int			flags;
		char *			tagp;
		journal_block_tag_t *	tag;
		struct buffer_head *	obh;
		struct buffer_head *	nbh;

		cond_resched();		/* We're under lock_kernel() */

		/* If we already know where to stop the log traversal,
		 * check right now that we haven't gone past the end of
		 * the log. */

		if (pass != PASS_SCAN)
			if (tid_geq(next_commit_ID, info->end_transaction))
				break;

		jbd_debug(2, "Scanning for sequence ID %u at %lu/%lu\n",
			  next_commit_ID, next_log_block, journal->j_last);

		/* Skip over each chunk of the transaction looking
		 * either the next descriptor block or the final commit
		 * record. */

		jbd_debug(3, "JBD: checking block %ld\n", next_log_block);
		err = jread(&bh, journal, next_log_block);
		if (err)
			goto failed;

		next_log_block++;
		wrap(journal, next_log_block);

		/* What kind of buffer is it?
		 *
		 * If it is a descriptor block, check that it has the
		 * expected sequence number.  Otherwise, we're all done
		 * here. */

		tmp = (journal_header_t *)bh->b_data;

		if (tmp->h_magic != cpu_to_be32(JBD2_MAGIC_NUMBER)) {
			brelse(bh);
			break;
		}

		blocktype = be32_to_cpu(tmp->h_blocktype);
		sequence = be32_to_cpu(tmp->h_sequence);
		jbd_debug(3, "Found magic %d, sequence %d\n",
			  blocktype, sequence);

		if (sequence != next_commit_ID) {
			brelse(bh);
			break;
		}

		/* OK, we have a valid descriptor block which matches
		 * all of the sequence number checks.  What are we going
		 * to do with it?  That depends on the pass... */

		switch(blocktype) {
		case JBD2_DESCRIPTOR_BLOCK:
			/* If it is a valid descriptor block, replay it
			 * in pass REPLAY; otherwise, just skip over the
			 * blocks it describes. */
			if (pass != PASS_REPLAY) {
				next_log_block += count_tags(journal, bh);
				wrap(journal, next_log_block);
				brelse(bh);
				continue;
			}

			/* A descriptor block: we can now write all of
			 * the data blocks.  Yay, useful work is finally
			 * getting done here! */

			tagp = &bh->b_data[sizeof(journal_header_t)];
			while ((tagp - bh->b_data + tag_bytes)
			       <= journal->j_blocksize) {
				unsigned long io_block;

				tag = (journal_block_tag_t *) tagp;
				flags = be32_to_cpu(tag->t_flags);

				io_block = next_log_block++;
				wrap(journal, next_log_block);
				err = jread(&obh, journal, io_block);
				if (err) {
					/* Recover what we can, but
					 * report failure at the end. */
					success = err;
					printk (KERN_ERR
						"JBD: IO error %d recovering "
						"block %ld in log\n",
						err, io_block);
				} else {
					sector_t blocknr;

					J_ASSERT(obh != NULL);
					blocknr = read_tag_block(tag_bytes,
								 tag);

					/* If the block has been
					 * revoked, then we're all done
					 * here. */
					if (jbd2_journal_test_revoke
					    (journal, blocknr,
					     next_commit_ID)) {
						brelse(obh);
						++info->nr_revoke_hits;
						goto skip_write;
					}

					/* Find a buffer for the new
					 * data being restored */
					nbh = __getblk(journal->j_fs_dev,
							blocknr,
							journal->j_blocksize);
					if (nbh == NULL) {
						printk(KERN_ERR
						       "JBD: Out of memory "
						       "during recovery.\n");
						err = -ENOMEM;
						brelse(bh);
						brelse(obh);
						goto failed;
					}

					lock_buffer(nbh);
					memcpy(nbh->b_data, obh->b_data,
							journal->j_blocksize);
					if (flags & JBD2_FLAG_ESCAPE) {
						*((__be32 *)bh->b_data) =
						cpu_to_be32(JBD2_MAGIC_NUMBER);
					}

					BUFFER_TRACE(nbh, "marking dirty");
					set_buffer_uptodate(nbh);
					mark_buffer_dirty(nbh);
					BUFFER_TRACE(nbh, "marking uptodate");
					++info->nr_replays;
					/* ll_rw_block(WRITE, 1, &nbh); */
					unlock_buffer(nbh);
					brelse(obh);
					brelse(nbh);
				}

			skip_write:
				tagp += tag_bytes;
				if (!(flags & JBD2_FLAG_SAME_UUID))
					tagp += 16;

				if (flags & JBD2_FLAG_LAST_TAG)
					break;
			}

			brelse(bh);
			continue;

		case JBD2_COMMIT_BLOCK:
			/* Found an expected commit block: not much to
			 * do other than move on to the next sequence
			 * number. */
			brelse(bh);
			next_commit_ID++;
			continue;

		case JBD2_REVOKE_BLOCK:
			/* If we aren't in the REVOKE pass, then we can
			 * just skip over this block. */
			if (pass != PASS_REVOKE) {
				brelse(bh);
				continue;
			}

			err = scan_revoke_records(journal, bh,
						  next_commit_ID, info);
			brelse(bh);
			if (err)
				goto failed;
			continue;

		default:
			jbd_debug(3, "Unrecognised magic %d, end of scan.\n",
				  blocktype);
			brelse(bh);
			goto done;
		}
	}

 done:
	/*
	 * We broke out of the log scan loop: either we came to the
	 * known end of the log or we found an unexpected block in the
	 * log.  If the latter happened, then we know that the "current"
	 * transaction marks the end of the valid log.
	 */

	if (pass == PASS_SCAN)
		info->end_transaction = next_commit_ID;
	else {
		/* It's really bad news if different passes end up at
		 * different places (but possible due to IO errors). */
		if (info->end_transaction != next_commit_ID) {
			printk (KERN_ERR "JBD: recovery pass %d ended at "
				"transaction %u, expected %u\n",
				pass, next_commit_ID, info->end_transaction);
			if (!success)
				success = -EIO;
		}
	}

	return success;

 failed:
	return err;
}


/* Scan a revoke record, marking all blocks mentioned as revoked. */

static int scan_revoke_records(journal_t *journal, struct buffer_head *bh,
			       tid_t sequence, struct recovery_info *info)
{
	jbd2_journal_revoke_header_t *header;
	int offset, max;
	int record_len = 4;

	header = (jbd2_journal_revoke_header_t *) bh->b_data;
	offset = sizeof(jbd2_journal_revoke_header_t);
	max = be32_to_cpu(header->r_count);

	if (JBD2_HAS_INCOMPAT_FEATURE(journal, JBD2_FEATURE_INCOMPAT_64BIT))
		record_len = 8;

	while (offset + record_len <= max) {
		sector_t blocknr;
		int err;

		if (record_len == 4)
			blocknr = be32_to_cpu(* ((__be32 *) (bh->b_data+offset)));
		else
			blocknr = be64_to_cpu(* ((__be64 *) (bh->b_data+offset)));
		offset += record_len;
		err = jbd2_journal_set_revoke(journal, blocknr, sequence);
		if (err)
			return err;
		++info->nr_revokes;
	}
	return 0;
}