inode.c 157 KB
Newer Older
1
/*
2
 *  linux/fs/ext4/inode.c
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
 *
 * Copyright (C) 1992, 1993, 1994, 1995
 * Remy Card (card@masi.ibp.fr)
 * Laboratoire MASI - Institut Blaise Pascal
 * Universite Pierre et Marie Curie (Paris VI)
 *
 *  from
 *
 *  linux/fs/minix/inode.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 *
 *  Goal-directed block allocation by Stephen Tweedie
 *	(sct@redhat.com), 1993, 1998
 *  Big-endian to little-endian byte-swapping/bitmaps by
 *        David S. Miller (davem@caip.rutgers.edu), 1995
 *  64-bit file support on 64-bit platforms by Jakub Jelinek
 *	(jj@sunsite.ms.mff.cuni.cz)
 *
22
 *  Assorted race fixes, rewrite of ext4_get_block() by Al Viro, 2000
23
24
25
26
27
 */

#include <linux/module.h>
#include <linux/fs.h>
#include <linux/time.h>
28
#include <linux/jbd2.h>
29
30
31
32
33
34
#include <linux/highuid.h>
#include <linux/pagemap.h>
#include <linux/quotaops.h>
#include <linux/string.h>
#include <linux/buffer_head.h>
#include <linux/writeback.h>
35
#include <linux/pagevec.h>
36
#include <linux/mpage.h>
37
#include <linux/namei.h>
38
39
#include <linux/uio.h>
#include <linux/bio.h>
40

41
#include "ext4_jbd2.h"
42
43
#include "xattr.h"
#include "acl.h"
44
#include "ext4_extents.h"
45

46
47
#include <trace/events/ext4.h>

48
49
#define MPAGE_DA_EXTENT_TAIL 0x01

50
51
52
static inline int ext4_begin_ordered_truncate(struct inode *inode,
					      loff_t new_size)
{
53
54
55
56
	return jbd2_journal_begin_ordered_truncate(
					EXT4_SB(inode->i_sb)->s_journal,
					&EXT4_I(inode)->jinode,
					new_size);
57
58
}

59
60
static void ext4_invalidatepage(struct page *page, unsigned long offset);

61
62
63
/*
 * Test whether an inode is a fast symlink.
 */
64
static int ext4_inode_is_fast_symlink(struct inode *inode)
65
{
66
	int ea_blocks = EXT4_I(inode)->i_file_acl ?
67
68
69
70
71
72
		(inode->i_sb->s_blocksize >> 9) : 0;

	return (S_ISLNK(inode->i_mode) && inode->i_blocks - ea_blocks == 0);
}

/*
73
 * The ext4 forget function must perform a revoke if we are freeing data
74
75
76
77
78
79
 * which has been journaled.  Metadata (eg. indirect blocks) must be
 * revoked in all cases.
 *
 * "bh" may be NULL: a metadata block may have been freed from memory
 * but there may still be a record of it in the journal, and that record
 * still needs to be revoked.
80
81
 *
 * If the handle isn't valid we're not journaling so there's nothing to do.
82
 */
83
84
int ext4_forget(handle_t *handle, int is_metadata, struct inode *inode,
			struct buffer_head *bh, ext4_fsblk_t blocknr)
85
86
87
{
	int err;

88
89
90
	if (!ext4_handle_valid(handle))
		return 0;

91
92
93
94
95
	might_sleep();

	BUFFER_TRACE(bh, "enter");

	jbd_debug(4, "forgetting bh %p: is_metadata = %d, mode %o, "
96
		  "data mode %x\n",
97
98
99
100
101
102
103
104
		  bh, is_metadata, inode->i_mode,
		  test_opt(inode->i_sb, DATA_FLAGS));

	/* Never use the revoke function if we are doing full data
	 * journaling: there is no need to, and a V1 superblock won't
	 * support it.  Otherwise, only skip the revoke on un-journaled
	 * data blocks. */

105
106
	if (test_opt(inode->i_sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA ||
	    (!is_metadata && !ext4_should_journal_data(inode))) {
107
		if (bh) {
108
			BUFFER_TRACE(bh, "call jbd2_journal_forget");
109
			return ext4_journal_forget(handle, bh);
110
111
112
113
114
115
116
		}
		return 0;
	}

	/*
	 * data!=journal && (is_metadata || should_journal_data(inode))
	 */
117
118
	BUFFER_TRACE(bh, "call ext4_journal_revoke");
	err = ext4_journal_revoke(handle, blocknr, bh);
119
	if (err)
120
		ext4_abort(inode->i_sb, __func__,
121
122
123
124
125
126
127
128
129
130
131
			   "error %d when attempting revoke", err);
	BUFFER_TRACE(bh, "exit");
	return err;
}

/*
 * Work out how many blocks we need to proceed with the next chunk of a
 * truncate transaction.
 */
static unsigned long blocks_for_truncate(struct inode *inode)
{
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
132
	ext4_lblk_t needed;
133
134
135
136
137
138

	needed = inode->i_blocks >> (inode->i_sb->s_blocksize_bits - 9);

	/* Give ourselves just enough room to cope with inodes in which
	 * i_blocks is corrupt: we've seen disk corruptions in the past
	 * which resulted in random data in an inode which looked enough
139
	 * like a regular file for ext4 to try to delete it.  Things
140
141
142
143
144
145
146
	 * will go a bit crazy if that happens, but at least we should
	 * try not to panic the whole kernel. */
	if (needed < 2)
		needed = 2;

	/* But we need to bound the transaction so we don't overflow the
	 * journal. */
147
148
	if (needed > EXT4_MAX_TRANS_DATA)
		needed = EXT4_MAX_TRANS_DATA;
149

150
	return EXT4_DATA_TRANS_BLOCKS(inode->i_sb) + needed;
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
}

/*
 * Truncate transactions can be complex and absolutely huge.  So we need to
 * be able to restart the transaction at a conventient checkpoint to make
 * sure we don't overflow the journal.
 *
 * start_transaction gets us a new handle for a truncate transaction,
 * and extend_transaction tries to extend the existing one a bit.  If
 * extend fails, we need to propagate the failure up and restart the
 * transaction in the top-level truncate loop. --sct
 */
static handle_t *start_transaction(struct inode *inode)
{
	handle_t *result;

167
	result = ext4_journal_start(inode, blocks_for_truncate(inode));
168
169
170
	if (!IS_ERR(result))
		return result;

171
	ext4_std_error(inode->i_sb, PTR_ERR(result));
172
173
174
175
176
177
178
179
180
181
182
	return result;
}

/*
 * Try to extend this transaction for the purposes of truncation.
 *
 * Returns 0 if we managed to create more room.  If we can't create more
 * room, and the transaction must be restarted we return 1.
 */
static int try_to_extend_transaction(handle_t *handle, struct inode *inode)
{
183
184
185
	if (!ext4_handle_valid(handle))
		return 0;
	if (ext4_handle_has_enough_credits(handle, EXT4_RESERVE_TRANS_BLOCKS+1))
186
		return 0;
187
	if (!ext4_journal_extend(handle, blocks_for_truncate(inode)))
188
189
190
191
192
193
194
195
196
		return 0;
	return 1;
}

/*
 * Restart the transaction associated with *handle.  This does a commit,
 * so before we call here everything must be consistently dirtied against
 * this transaction.
 */
197
static int ext4_journal_test_restart(handle_t *handle, struct inode *inode)
198
{
199
	BUG_ON(EXT4_JOURNAL(inode) == NULL);
200
	jbd_debug(2, "restarting handle %p\n", handle);
201
	return ext4_journal_restart(handle, blocks_for_truncate(inode));
202
203
204
205
206
}

/*
 * Called at the last iput() if i_nlink is zero.
 */
207
void ext4_delete_inode(struct inode *inode)
208
209
{
	handle_t *handle;
210
	int err;
211

212
213
	if (ext4_should_order_data(inode))
		ext4_begin_ordered_truncate(inode, 0);
214
215
216
217
218
	truncate_inode_pages(&inode->i_data, 0);

	if (is_bad_inode(inode))
		goto no_delete;

219
	handle = ext4_journal_start(inode, blocks_for_truncate(inode)+3);
220
	if (IS_ERR(handle)) {
221
		ext4_std_error(inode->i_sb, PTR_ERR(handle));
222
223
224
225
226
		/*
		 * If we're going to skip the normal cleanup, we still need to
		 * make sure that the in-core orphan linked list is properly
		 * cleaned up.
		 */
227
		ext4_orphan_del(NULL, inode);
228
229
230
231
		goto no_delete;
	}

	if (IS_SYNC(inode))
232
		ext4_handle_sync(handle);
233
	inode->i_size = 0;
234
235
236
237
238
239
	err = ext4_mark_inode_dirty(handle, inode);
	if (err) {
		ext4_warning(inode->i_sb, __func__,
			     "couldn't mark inode dirty (err %d)", err);
		goto stop_handle;
	}
240
	if (inode->i_blocks)
241
		ext4_truncate(inode);
242
243
244
245
246
247
248

	/*
	 * ext4_ext_truncate() doesn't reserve any slop when it
	 * restarts journal transactions; therefore there may not be
	 * enough credits left in the handle to remove the inode from
	 * the orphan list and set the dtime field.
	 */
249
	if (!ext4_handle_has_enough_credits(handle, 3)) {
250
251
252
253
254
255
256
257
258
259
260
261
		err = ext4_journal_extend(handle, 3);
		if (err > 0)
			err = ext4_journal_restart(handle, 3);
		if (err != 0) {
			ext4_warning(inode->i_sb, __func__,
				     "couldn't extend journal (err %d)", err);
		stop_handle:
			ext4_journal_stop(handle);
			goto no_delete;
		}
	}

262
	/*
263
	 * Kill off the orphan record which ext4_truncate created.
264
	 * AKPM: I think this can be inside the above `if'.
265
	 * Note that ext4_orphan_del() has to be able to cope with the
266
	 * deletion of a non-existent orphan - this is because we don't
267
	 * know if ext4_truncate() actually created an orphan record.
268
269
	 * (Well, we could do this if we need to, but heck - it works)
	 */
270
271
	ext4_orphan_del(handle, inode);
	EXT4_I(inode)->i_dtime	= get_seconds();
272
273
274
275
276
277
278
279

	/*
	 * One subtle ordering requirement: if anything has gone wrong
	 * (transaction abort, IO errors, whatever), then we can still
	 * do these next steps (the fs will already have been marked as
	 * having errors), but we can't free the inode if the mark_dirty
	 * fails.
	 */
280
	if (ext4_mark_inode_dirty(handle, inode))
281
282
283
		/* If that failed, just do the required in-core inode clear. */
		clear_inode(inode);
	else
284
285
		ext4_free_inode(handle, inode);
	ext4_journal_stop(handle);
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
	return;
no_delete:
	clear_inode(inode);	/* We must guarantee clearing of inode... */
}

typedef struct {
	__le32	*p;
	__le32	key;
	struct buffer_head *bh;
} Indirect;

static inline void add_chain(Indirect *p, struct buffer_head *bh, __le32 *v)
{
	p->key = *(p->p = v);
	p->bh = bh;
}

/**
304
 *	ext4_block_to_path - parse the block number into array of offsets
305
306
307
 *	@inode: inode in question (we are only interested in its superblock)
 *	@i_block: block number to be parsed
 *	@offsets: array to store the offsets in
Dave Kleikamp's avatar
Dave Kleikamp committed
308
309
 *	@boundary: set this non-zero if the referred-to block is likely to be
 *	       followed (on disk) by an indirect block.
310
 *
311
 *	To store the locations of file's data ext4 uses a data structure common
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
 *	for UNIX filesystems - tree of pointers anchored in the inode, with
 *	data blocks at leaves and indirect blocks in intermediate nodes.
 *	This function translates the block number into path in that tree -
 *	return value is the path length and @offsets[n] is the offset of
 *	pointer to (n+1)th node in the nth one. If @block is out of range
 *	(negative or too large) warning is printed and zero returned.
 *
 *	Note: function doesn't find node addresses, so no IO is needed. All
 *	we need to know is the capacity of indirect blocks (taken from the
 *	inode->i_sb).
 */

/*
 * Portability note: the last comparison (check that we fit into triple
 * indirect block) is spelled differently, because otherwise on an
 * architecture with 32-bit longs and 8Kb pages we might get into trouble
 * if our filesystem had 8Kb blocks. We might use long long, but that would
 * kill us on x86. Oh, well, at least the sign propagation does not matter -
 * i_block would have to be negative in the very beginning, so we would not
 * get there at all.
 */

334
static int ext4_block_to_path(struct inode *inode,
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
335
336
			ext4_lblk_t i_block,
			ext4_lblk_t offsets[4], int *boundary)
337
{
338
339
340
	int ptrs = EXT4_ADDR_PER_BLOCK(inode->i_sb);
	int ptrs_bits = EXT4_ADDR_PER_BLOCK_BITS(inode->i_sb);
	const long direct_blocks = EXT4_NDIR_BLOCKS,
341
342
343
344
345
346
		indirect_blocks = ptrs,
		double_blocks = (1 << (ptrs_bits * 2));
	int n = 0;
	int final = 0;

	if (i_block < 0) {
347
		ext4_warning(inode->i_sb, "ext4_block_to_path", "block < 0");
348
349
350
	} else if (i_block < direct_blocks) {
		offsets[n++] = i_block;
		final = direct_blocks;
351
	} else if ((i_block -= direct_blocks) < indirect_blocks) {
352
		offsets[n++] = EXT4_IND_BLOCK;
353
354
355
		offsets[n++] = i_block;
		final = ptrs;
	} else if ((i_block -= indirect_blocks) < double_blocks) {
356
		offsets[n++] = EXT4_DIND_BLOCK;
357
358
359
360
		offsets[n++] = i_block >> ptrs_bits;
		offsets[n++] = i_block & (ptrs - 1);
		final = ptrs;
	} else if (((i_block -= double_blocks) >> (ptrs_bits * 2)) < ptrs) {
361
		offsets[n++] = EXT4_TIND_BLOCK;
362
363
364
365
366
		offsets[n++] = i_block >> (ptrs_bits * 2);
		offsets[n++] = (i_block >> ptrs_bits) & (ptrs - 1);
		offsets[n++] = i_block & (ptrs - 1);
		final = ptrs;
	} else {
367
		ext4_warning(inode->i_sb, "ext4_block_to_path",
368
				"block %lu > max in inode %lu",
369
				i_block + direct_blocks +
370
				indirect_blocks + double_blocks, inode->i_ino);
371
372
373
374
375
376
	}
	if (boundary)
		*boundary = final - 1 - (i_block & (ptrs - 1));
	return n;
}

377
static int __ext4_check_blockref(const char *function, struct inode *inode,
378
379
				 __le32 *p, unsigned int max)
{
380
	__le32 *bref = p;
381
382
	unsigned int blk;

383
	while (bref < p+max) {
384
385
386
387
		blk = le32_to_cpu(*bref++);
		if (blk && 
		    unlikely(!ext4_data_block_valid(EXT4_SB(inode->i_sb), 
						    blk, 1))) {
388
			ext4_error(inode->i_sb, function,
389
390
				   "invalid block reference %u "
				   "in inode #%lu", blk, inode->i_ino);
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
 			return -EIO;
 		}
 	}
 	return 0;
}


#define ext4_check_indirect_blockref(inode, bh)                         \
        __ext4_check_blockref(__func__, inode, (__le32 *)(bh)->b_data,  \
			      EXT4_ADDR_PER_BLOCK((inode)->i_sb))

#define ext4_check_inode_blockref(inode)                                \
        __ext4_check_blockref(__func__, inode, EXT4_I(inode)->i_data,   \
			      EXT4_NDIR_BLOCKS)

406
/**
407
 *	ext4_get_branch - read the chain of indirect blocks leading to data
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
 *	@inode: inode in question
 *	@depth: depth of the chain (1 - direct pointer, etc.)
 *	@offsets: offsets of pointers in inode/indirect blocks
 *	@chain: place to store the result
 *	@err: here we store the error value
 *
 *	Function fills the array of triples <key, p, bh> and returns %NULL
 *	if everything went OK or the pointer to the last filled triple
 *	(incomplete one) otherwise. Upon the return chain[i].key contains
 *	the number of (i+1)-th block in the chain (as it is stored in memory,
 *	i.e. little-endian 32-bit), chain[i].p contains the address of that
 *	number (it points into struct inode for i==0 and into the bh->b_data
 *	for i>0) and chain[i].bh points to the buffer_head of i-th indirect
 *	block for i>0 and NULL for i==0. In other words, it holds the block
 *	numbers of the chain, addresses they were taken from (and where we can
 *	verify that chain did not change) and buffer_heads hosting these
 *	numbers.
 *
 *	Function stops when it stumbles upon zero pointer (absent block)
 *		(pointer to last triple returned, *@err == 0)
 *	or when it gets an IO error reading an indirect block
 *		(ditto, *@err == -EIO)
 *	or when it reads all @depth-1 indirect blocks successfully and finds
 *	the whole chain, all way to the data (returns %NULL, *err == 0).
432
433
 *
 *      Need to be called with
434
 *      down_read(&EXT4_I(inode)->i_data_sem)
435
 */
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
436
437
static Indirect *ext4_get_branch(struct inode *inode, int depth,
				 ext4_lblk_t  *offsets,
438
439
440
441
442
443
444
445
				 Indirect chain[4], int *err)
{
	struct super_block *sb = inode->i_sb;
	Indirect *p = chain;
	struct buffer_head *bh;

	*err = 0;
	/* i_data is not going away, no lock needed */
446
	add_chain(chain, NULL, EXT4_I(inode)->i_data + *offsets);
447
448
449
	if (!p->key)
		goto no_block;
	while (--depth) {
450
451
		bh = sb_getblk(sb, le32_to_cpu(p->key));
		if (unlikely(!bh))
452
			goto failure;
453
454
455
456
457
458
459
460
461
462
463
464
465
                  
		if (!bh_uptodate_or_lock(bh)) {
			if (bh_submit_read(bh) < 0) {
				put_bh(bh);
				goto failure;
			}
			/* validate block references */
			if (ext4_check_indirect_blockref(inode, bh)) {
				put_bh(bh);
				goto failure;
			}
		}
		
466
		add_chain(++p, bh, (__le32 *)bh->b_data + *++offsets);
467
468
469
470
471
472
473
474
475
476
477
478
479
		/* Reader: end */
		if (!p->key)
			goto no_block;
	}
	return NULL;

failure:
	*err = -EIO;
no_block:
	return p;
}

/**
480
 *	ext4_find_near - find a place for allocation with sufficient locality
481
482
483
 *	@inode: owner
 *	@ind: descriptor of indirect block.
 *
484
 *	This function returns the preferred place for block allocation.
485
486
487
488
489
490
491
492
493
494
495
496
497
498
 *	It is used when heuristic for sequential allocation fails.
 *	Rules are:
 *	  + if there is a block to the left of our position - allocate near it.
 *	  + if pointer will live in indirect block - allocate near that block.
 *	  + if pointer will live in inode - allocate in the same
 *	    cylinder group.
 *
 * In the latter case we colour the starting block by the callers PID to
 * prevent it from clashing with concurrent allocations for a different inode
 * in the same block group.   The PID is used here so that functionally related
 * files will be close-by on-disk.
 *
 *	Caller must make sure that @ind is valid and will stay that way.
 */
499
static ext4_fsblk_t ext4_find_near(struct inode *inode, Indirect *ind)
500
{
501
	struct ext4_inode_info *ei = EXT4_I(inode);
502
	__le32 *start = ind->bh ? (__le32 *) ind->bh->b_data : ei->i_data;
503
	__le32 *p;
504
	ext4_fsblk_t bg_start;
505
	ext4_fsblk_t last_block;
506
	ext4_grpblk_t colour;
507
508
	ext4_group_t block_group;
	int flex_size = ext4_flex_bg_size(EXT4_SB(inode->i_sb));
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523

	/* Try to find previous block */
	for (p = ind->p - 1; p >= start; p--) {
		if (*p)
			return le32_to_cpu(*p);
	}

	/* No such thing, so let's try location of indirect block */
	if (ind->bh)
		return ind->bh->b_blocknr;

	/*
	 * It is going to be referred to from the inode itself? OK, just put it
	 * into the same cylinder group then.
	 */
524
525
526
527
528
529
530
	block_group = ei->i_block_group;
	if (flex_size >= EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME) {
		block_group &= ~(flex_size-1);
		if (S_ISREG(inode->i_mode))
			block_group++;
	}
	bg_start = ext4_group_first_block_no(inode->i_sb, block_group);
531
532
	last_block = ext4_blocks_count(EXT4_SB(inode->i_sb)->s_es) - 1;

533
534
535
536
537
538
539
	/*
	 * If we are doing delayed allocation, we don't need take
	 * colour into account.
	 */
	if (test_opt(inode->i_sb, DELALLOC))
		return bg_start;

540
541
	if (bg_start + EXT4_BLOCKS_PER_GROUP(inode->i_sb) <= last_block)
		colour = (current->pid % 16) *
542
			(EXT4_BLOCKS_PER_GROUP(inode->i_sb) / 16);
543
544
	else
		colour = (current->pid % 16) * ((last_block - bg_start) / 16);
545
546
547
548
	return bg_start + colour;
}

/**
549
 *	ext4_find_goal - find a preferred place for allocation.
550
551
552
553
 *	@inode: owner
 *	@block:  block we want
 *	@partial: pointer to the last triple within a chain
 *
554
 *	Normally this function find the preferred place for block allocation,
555
 *	returns it.
556
 */
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
557
static ext4_fsblk_t ext4_find_goal(struct inode *inode, ext4_lblk_t block,
558
		Indirect *partial)
559
560
{
	/*
561
	 * XXX need to get goal block from mballoc's data structures
562
563
	 */

564
	return ext4_find_near(inode, partial);
565
566
567
}

/**
568
 *	ext4_blks_to_allocate: Look up the block map and count the number
569
570
571
572
573
574
575
576
577
578
 *	of direct blocks need to be allocated for the given branch.
 *
 *	@branch: chain of indirect blocks
 *	@k: number of blocks need for indirect blocks
 *	@blks: number of data blocks to be mapped.
 *	@blocks_to_boundary:  the offset in the indirect block
 *
 *	return the total number of blocks to be allocate, including the
 *	direct and indirect blocks.
 */
579
static int ext4_blks_to_allocate(Indirect *branch, int k, unsigned int blks,
580
581
		int blocks_to_boundary)
{
582
	unsigned int count = 0;
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605

	/*
	 * Simple case, [t,d]Indirect block(s) has not allocated yet
	 * then it's clear blocks on that path have not allocated
	 */
	if (k > 0) {
		/* right now we don't handle cross boundary allocation */
		if (blks < blocks_to_boundary + 1)
			count += blks;
		else
			count += blocks_to_boundary + 1;
		return count;
	}

	count++;
	while (count < blks && count <= blocks_to_boundary &&
		le32_to_cpu(*(branch[0].p + count)) == 0) {
		count++;
	}
	return count;
}

/**
606
 *	ext4_alloc_blocks: multiple allocate blocks needed for a branch
607
608
609
610
611
612
613
614
 *	@indirect_blks: the number of blocks need to allocate for indirect
 *			blocks
 *
 *	@new_blocks: on return it will store the new block numbers for
 *	the indirect blocks(if needed) and the first direct block,
 *	@blks:	on return it will store the total number of allocated
 *		direct blocks
 */
615
static int ext4_alloc_blocks(handle_t *handle, struct inode *inode,
616
617
618
				ext4_lblk_t iblock, ext4_fsblk_t goal,
				int indirect_blks, int blks,
				ext4_fsblk_t new_blocks[4], int *err)
619
{
620
	struct ext4_allocation_request ar;
621
	int target, i;
622
	unsigned long count = 0, blk_allocated = 0;
623
	int index = 0;
624
	ext4_fsblk_t current_block = 0;
625
626
627
628
629
630
631
632
633
634
	int ret = 0;

	/*
	 * Here we try to allocate the requested multiple blocks at once,
	 * on a best-effort basis.
	 * To build a branch, we should allocate blocks for
	 * the indirect blocks(if not allocated yet), and at least
	 * the first direct block of this branch.  That's the
	 * minimum number of blocks need to allocate(required)
	 */
635
636
637
	/* first we try to allocate the indirect blocks */
	target = indirect_blks;
	while (target > 0) {
638
639
		count = target;
		/* allocating blocks for indirect blocks and direct blocks */
640
641
		current_block = ext4_new_meta_blocks(handle, inode,
							goal, &count, err);
642
643
644
645
646
647
648
649
650
		if (*err)
			goto failed_out;

		target -= count;
		/* allocate blocks for indirect blocks */
		while (index < indirect_blks && count) {
			new_blocks[index++] = current_block++;
			count--;
		}
651
652
653
654
655
656
657
658
659
		if (count > 0) {
			/*
			 * save the new block number
			 * for the first direct block
			 */
			new_blocks[index] = current_block;
			printk(KERN_INFO "%s returned more blocks than "
						"requested\n", __func__);
			WARN_ON(1);
660
			break;
661
		}
662
663
	}

664
665
666
667
668
	target = blks - count ;
	blk_allocated = count;
	if (!target)
		goto allocated;
	/* Now allocate data blocks */
669
670
671
672
673
674
675
676
677
678
679
	memset(&ar, 0, sizeof(ar));
	ar.inode = inode;
	ar.goal = goal;
	ar.len = target;
	ar.logical = iblock;
	if (S_ISREG(inode->i_mode))
		/* enable in-core preallocation only for regular files */
		ar.flags = EXT4_MB_HINT_DATA;

	current_block = ext4_mb_new_blocks(handle, &ar, err);

680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
	if (*err && (target == blks)) {
		/*
		 * if the allocation failed and we didn't allocate
		 * any blocks before
		 */
		goto failed_out;
	}
	if (!*err) {
		if (target == blks) {
		/*
		 * save the new block number
		 * for the first direct block
		 */
			new_blocks[index] = current_block;
		}
695
		blk_allocated += ar.len;
696
697
	}
allocated:
698
	/* total number of blocks allocated for direct blocks */
699
	ret = blk_allocated;
700
701
702
	*err = 0;
	return ret;
failed_out:
703
	for (i = 0; i < index; i++)
704
		ext4_free_blocks(handle, inode, new_blocks[i], 1, 0);
705
706
707
708
	return ret;
}

/**
709
 *	ext4_alloc_branch - allocate and set up a chain of blocks.
710
711
712
713
714
715
716
717
718
719
 *	@inode: owner
 *	@indirect_blks: number of allocated indirect blocks
 *	@blks: number of allocated direct blocks
 *	@offsets: offsets (in the blocks) to store the pointers to next.
 *	@branch: place to store the chain in.
 *
 *	This function allocates blocks, zeroes out all but the last one,
 *	links them into chain and (if we are synchronous) writes them to disk.
 *	In other words, it prepares a branch that can be spliced onto the
 *	inode. It stores the information about that chain in the branch[], in
720
 *	the same format as ext4_get_branch() would do. We are calling it after
721
722
 *	we had read the existing part of chain and partial points to the last
 *	triple of that (one with zero ->key). Upon the exit we have the same
723
 *	picture as after the successful ext4_get_block(), except that in one
724
725
726
727
728
729
 *	place chain is disconnected - *branch->p is still zero (we did not
 *	set the last link), but branch->key contains the number that should
 *	be placed into *branch->p to fill that gap.
 *
 *	If allocation fails we free all blocks we've allocated (and forget
 *	their buffer_heads) and return the error value the from failed
730
 *	ext4_alloc_block() (normally -ENOSPC). Otherwise we set the chain
731
732
 *	as described above and return 0.
 */
733
static int ext4_alloc_branch(handle_t *handle, struct inode *inode,
734
735
736
				ext4_lblk_t iblock, int indirect_blks,
				int *blks, ext4_fsblk_t goal,
				ext4_lblk_t *offsets, Indirect *branch)
737
738
739
740
741
742
{
	int blocksize = inode->i_sb->s_blocksize;
	int i, n = 0;
	int err = 0;
	struct buffer_head *bh;
	int num;
743
744
	ext4_fsblk_t new_blocks[4];
	ext4_fsblk_t current_block;
745

746
	num = ext4_alloc_blocks(handle, inode, iblock, goal, indirect_blks,
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
				*blks, new_blocks, &err);
	if (err)
		return err;

	branch[0].key = cpu_to_le32(new_blocks[0]);
	/*
	 * metadata blocks and data blocks are allocated.
	 */
	for (n = 1; n <= indirect_blks;  n++) {
		/*
		 * Get buffer_head for parent block, zero it out
		 * and set the pointer to new one, then send
		 * parent to disk.
		 */
		bh = sb_getblk(inode->i_sb, new_blocks[n-1]);
		branch[n].bh = bh;
		lock_buffer(bh);
		BUFFER_TRACE(bh, "call get_create_access");
765
		err = ext4_journal_get_create_access(handle, bh);
766
767
768
769
770
771
772
773
774
775
		if (err) {
			unlock_buffer(bh);
			brelse(bh);
			goto failed;
		}

		memset(bh->b_data, 0, blocksize);
		branch[n].p = (__le32 *) bh->b_data + offsets[n];
		branch[n].key = cpu_to_le32(new_blocks[n]);
		*branch[n].p = branch[n].key;
776
		if (n == indirect_blks) {
777
778
779
780
781
782
783
784
785
786
787
788
789
			current_block = new_blocks[n];
			/*
			 * End of chain, update the last new metablock of
			 * the chain to point to the new allocated
			 * data blocks numbers
			 */
			for (i=1; i < num; i++)
				*(branch[n].p + i) = cpu_to_le32(++current_block);
		}
		BUFFER_TRACE(bh, "marking uptodate");
		set_buffer_uptodate(bh);
		unlock_buffer(bh);

790
791
		BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
		err = ext4_handle_dirty_metadata(handle, inode, bh);
792
793
794
795
796
797
798
799
		if (err)
			goto failed;
	}
	*blks = num;
	return err;
failed:
	/* Allocation failed, free what we already allocated */
	for (i = 1; i <= n ; i++) {
800
		BUFFER_TRACE(branch[i].bh, "call jbd2_journal_forget");
801
		ext4_journal_forget(handle, branch[i].bh);
802
	}
803
	for (i = 0; i < indirect_blks; i++)
804
		ext4_free_blocks(handle, inode, new_blocks[i], 1, 0);
805

806
	ext4_free_blocks(handle, inode, new_blocks[i], num, 0);
807
808
809
810
811

	return err;
}

/**
812
 * ext4_splice_branch - splice the allocated branch onto inode.
813
814
815
 * @inode: owner
 * @block: (logical) number of block we are adding
 * @chain: chain of indirect blocks (with a missing link - see
816
 *	ext4_alloc_branch)
817
818
819
820
821
822
823
824
 * @where: location of missing link
 * @num:   number of indirect blocks we are adding
 * @blks:  number of direct blocks we are adding
 *
 * This function fills the missing link and does all housekeeping needed in
 * inode (->i_blocks, etc.). In case of success we end up with the full
 * chain to new block and return 0.
 */
825
static int ext4_splice_branch(handle_t *handle, struct inode *inode,
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
826
			ext4_lblk_t block, Indirect *where, int num, int blks)
827
828
829
{
	int i;
	int err = 0;
830
	ext4_fsblk_t current_block;
831
832
833
834
835
836
837
838

	/*
	 * If we're splicing into a [td]indirect block (as opposed to the
	 * inode) then we need to get write access to the [td]indirect block
	 * before the splice.
	 */
	if (where->bh) {
		BUFFER_TRACE(where->bh, "get_write_access");
839
		err = ext4_journal_get_write_access(handle, where->bh);
840
841
842
843
844
845
846
847
848
849
850
851
852
853
		if (err)
			goto err_out;
	}
	/* That's it */

	*where->p = where->key;

	/*
	 * Update the host buffer_head or inode to point to more just allocated
	 * direct blocks blocks
	 */
	if (num == 0 && blks > 1) {
		current_block = le32_to_cpu(where->key) + 1;
		for (i = 1; i < blks; i++)
854
			*(where->p + i) = cpu_to_le32(current_block++);
855
856
857
858
	}

	/* We are done with atomic stuff, now do the rest of housekeeping */

Kalpak Shah's avatar
Kalpak Shah committed
859
	inode->i_ctime = ext4_current_time(inode);
860
	ext4_mark_inode_dirty(handle, inode);
861
862
863
864
865
866
867
868
869

	/* had we spliced it onto indirect block? */
	if (where->bh) {
		/*
		 * If we spliced it onto an indirect block, we haven't
		 * altered the inode.  Note however that if it is being spliced
		 * onto an indirect block at the very end of the file (the
		 * file is growing) then we *will* alter the inode to reflect
		 * the new i_size.  But that is not done here - it is done in
870
		 * generic_commit_write->__mark_inode_dirty->ext4_dirty_inode.
871
872
		 */
		jbd_debug(5, "splicing indirect only\n");
873
874
		BUFFER_TRACE(where->bh, "call ext4_handle_dirty_metadata");
		err = ext4_handle_dirty_metadata(handle, inode, where->bh);
875
876
877
878
879
880
881
882
883
884
885
886
887
		if (err)
			goto err_out;
	} else {
		/*
		 * OK, we spliced it into the inode itself on a direct block.
		 * Inode was dirtied above.
		 */
		jbd_debug(5, "splicing direct\n");
	}
	return err;

err_out:
	for (i = 1; i <= num; i++) {
888
		BUFFER_TRACE(where[i].bh, "call jbd2_journal_forget");
889
		ext4_journal_forget(handle, where[i].bh);
890
891
		ext4_free_blocks(handle, inode,
					le32_to_cpu(where[i-1].key), 1, 0);
892
	}
893
	ext4_free_blocks(handle, inode, le32_to_cpu(where[num].key), blks, 0);
894
895
896
897
898

	return err;
}

/*
899
900
901
902
 * The ext4_ind_get_blocks() function handles non-extents inodes
 * (i.e., using the traditional indirect/double-indirect i_blocks
 * scheme) for ext4_get_blocks().
 *
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
 * Allocation strategy is simple: if we have to allocate something, we will
 * have to go the whole way to leaf. So let's do it before attaching anything
 * to tree, set linkage between the newborn blocks, write them if sync is
 * required, recheck the path, free and repeat if check fails, otherwise
 * set the last missing link (that will protect us from any truncate-generated
 * removals - all blocks on the path are immune now) and possibly force the
 * write on the parent block.
 * That has a nice additional property: no special recovery from the failed
 * allocations is needed - we simply release blocks and do not touch anything
 * reachable from inode.
 *
 * `handle' can be NULL if create == 0.
 *
 * return > 0, # of blocks mapped or allocated.
 * return = 0, if plain lookup failed.
 * return < 0, error case.
919
 *
920
921
922
923
924
 * The ext4_ind_get_blocks() function should be called with
 * down_write(&EXT4_I(inode)->i_data_sem) if allocating filesystem
 * blocks (i.e., flags has EXT4_GET_BLOCKS_CREATE set) or
 * down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system
 * blocks.
925
 */
926
static int ext4_ind_get_blocks(handle_t *handle, struct inode *inode,
927
928
				  ext4_lblk_t iblock, unsigned int maxblocks,
				  struct buffer_head *bh_result,
929
				  int flags)
930
931
{
	int err = -EIO;
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
932
	ext4_lblk_t offsets[4];
933
934
	Indirect chain[4];
	Indirect *partial;
935
	ext4_fsblk_t goal;
936
937
938
939
	int indirect_blks;
	int blocks_to_boundary = 0;
	int depth;
	int count = 0;
940
	ext4_fsblk_t first_block = 0;
941

942
	J_ASSERT(!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL));
943
	J_ASSERT(handle != NULL || (flags & EXT4_GET_BLOCKS_CREATE) == 0);
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
944
945
	depth = ext4_block_to_path(inode, iblock, offsets,
					&blocks_to_boundary);
946
947
948
949

	if (depth == 0)
		goto out;

950
	partial = ext4_get_branch(inode, depth, offsets, chain, &err);
951
952
953
954
955
956
957
958

	/* Simplest case - block found, no allocation needed */
	if (!partial) {
		first_block = le32_to_cpu(chain[depth - 1].key);
		clear_buffer_new(bh_result);
		count++;
		/*map more blocks*/
		while (count < maxblocks && count <= blocks_to_boundary) {
959
			ext4_fsblk_t blk;
960
961
962
963
964
965
966
967

			blk = le32_to_cpu(*(chain[depth-1].p + count));

			if (blk == first_block + count)
				count++;
			else
				break;
		}
968
		goto got_it;
969
970
971
	}

	/* Next simple case - plain lookup or failed read of indirect block */
972
	if ((flags & EXT4_GET_BLOCKS_CREATE) == 0 || err == -EIO)
973
974
975
		goto cleanup;

	/*
976
	 * Okay, we need to do block allocation.
977
	*/
978
	goal = ext4_find_goal(inode, iblock, partial);
979
980
981
982
983
984
985
986

	/* the number of blocks need to allocate for [d,t]indirect blocks */
	indirect_blks = (chain + depth) - partial - 1;

	/*
	 * Next look up the indirect map to count the totoal number of
	 * direct blocks to allocate for this branch.
	 */
987
	count = ext4_blks_to_allocate(partial, indirect_blks,
988
989
					maxblocks, blocks_to_boundary);
	/*
990
	 * Block out ext4_truncate while we alter the tree
991
	 */
992
993
994
	err = ext4_alloc_branch(handle, inode, iblock, indirect_blks,
					&count, goal,
					offsets + (partial - chain), partial);
995
996

	/*
997
	 * The ext4_splice_branch call will free and forget any buffers
998
999
1000
	 * on the new chain if there is a failure, but that risks using
	 * up transaction credits, especially for bitmaps where the
	 * credits cannot be returned.  Can we handle this somehow?  We
For faster browsing, not all history is shown. View entire blame