inode.c 171 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
#include <linux/workqueue.h>
41
#include <linux/kernel.h>
42
#include <linux/printk.h>
43
#include <linux/slab.h>
44
#include <linux/ratelimit.h>
45

46
#include "ext4_jbd2.h"
47
48
#include "xattr.h"
#include "acl.h"
49
#include "ext4_extents.h"
50

51
52
#include <trace/events/ext4.h>

53
54
#define MPAGE_DA_EXTENT_TAIL 0x01

55
56
57
static inline int ext4_begin_ordered_truncate(struct inode *inode,
					      loff_t new_size)
{
58
	trace_ext4_begin_ordered_truncate(inode, new_size);
59
60
61
62
63
64
65
66
67
68
69
	/*
	 * If jinode is zero, then we never opened the file for
	 * writing, so there's no need to call
	 * jbd2_journal_begin_ordered_truncate() since there's no
	 * outstanding writes we need to flush.
	 */
	if (!EXT4_I(inode)->jinode)
		return 0;
	return jbd2_journal_begin_ordered_truncate(EXT4_JOURNAL(inode),
						   EXT4_I(inode)->jinode,
						   new_size);
70
71
}

72
static void ext4_invalidatepage(struct page *page, unsigned long offset);
73
74
75
76
77
78
static int noalloc_get_block_write(struct inode *inode, sector_t iblock,
				   struct buffer_head *bh_result, int create);
static int ext4_set_bh_endio(struct buffer_head *bh, struct inode *inode);
static void ext4_end_io_buffer_write(struct buffer_head *bh, int uptodate);
static int __ext4_journalled_writepage(struct page *page, unsigned int len);
static int ext4_bh_delay_or_unwritten(handle_t *handle, struct buffer_head *bh);
79

80
81
82
/*
 * Test whether an inode is a fast symlink.
 */
83
static int ext4_inode_is_fast_symlink(struct inode *inode)
84
{
85
	int ea_blocks = EXT4_I(inode)->i_file_acl ?
86
87
88
89
90
91
92
93
94
95
96
		(inode->i_sb->s_blocksize >> 9) : 0;

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

/*
 * 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
97
	ext4_lblk_t needed;
98
99
100
101
102
103

	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
104
	 * like a regular file for ext4 to try to delete it.  Things
105
106
107
108
109
110
111
	 * 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. */
112
113
	if (needed > EXT4_MAX_TRANS_DATA)
		needed = EXT4_MAX_TRANS_DATA;
114

115
	return EXT4_DATA_TRANS_BLOCKS(inode->i_sb) + needed;
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
}

/*
 * 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;

132
	result = ext4_journal_start(inode, blocks_for_truncate(inode));
133
134
135
	if (!IS_ERR(result))
		return result;

136
	ext4_std_error(inode->i_sb, PTR_ERR(result));
137
138
139
140
141
142
143
144
145
146
147
	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)
{
148
149
150
	if (!ext4_handle_valid(handle))
		return 0;
	if (ext4_handle_has_enough_credits(handle, EXT4_RESERVE_TRANS_BLOCKS+1))
151
		return 0;
152
	if (!ext4_journal_extend(handle, blocks_for_truncate(inode)))
153
154
155
156
157
158
159
160
161
		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.
 */
162
int ext4_truncate_restart_trans(handle_t *handle, struct inode *inode,
163
				 int nblocks)
164
{
165
166
167
	int ret;

	/*
168
	 * Drop i_data_sem to avoid deadlock with ext4_map_blocks.  At this
169
170
171
172
	 * moment, get_block can be called only for blocks inside i_size since
	 * page cache has been already dropped and writes are blocked by
	 * i_mutex. So we can safely drop the i_data_sem here.
	 */
173
	BUG_ON(EXT4_JOURNAL(inode) == NULL);
174
	jbd_debug(2, "restarting handle %p\n", handle);
175
	up_write(&EXT4_I(inode)->i_data_sem);
176
	ret = ext4_journal_restart(handle, nblocks);
177
	down_write(&EXT4_I(inode)->i_data_sem);
178
	ext4_discard_preallocations(inode);
179
180

	return ret;
181
182
183
184
185
}

/*
 * Called at the last iput() if i_nlink is zero.
 */
Al Viro's avatar
Al Viro committed
186
void ext4_evict_inode(struct inode *inode)
187
188
{
	handle_t *handle;
189
	int err;
190

191
	trace_ext4_evict_inode(inode);
Al Viro's avatar
Al Viro committed
192
193
194
195
196
	if (inode->i_nlink) {
		truncate_inode_pages(&inode->i_data, 0);
		goto no_delete;
	}

197
	if (!is_bad_inode(inode))
198
		dquot_initialize(inode);
199

200
201
	if (ext4_should_order_data(inode))
		ext4_begin_ordered_truncate(inode, 0);
202
203
204
205
206
	truncate_inode_pages(&inode->i_data, 0);

	if (is_bad_inode(inode))
		goto no_delete;

207
	handle = ext4_journal_start(inode, blocks_for_truncate(inode)+3);
208
	if (IS_ERR(handle)) {
209
		ext4_std_error(inode->i_sb, PTR_ERR(handle));
210
211
212
213
214
		/*
		 * 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.
		 */
215
		ext4_orphan_del(NULL, inode);
216
217
218
219
		goto no_delete;
	}

	if (IS_SYNC(inode))
220
		ext4_handle_sync(handle);
221
	inode->i_size = 0;
222
223
	err = ext4_mark_inode_dirty(handle, inode);
	if (err) {
224
		ext4_warning(inode->i_sb,
225
226
227
			     "couldn't mark inode dirty (err %d)", err);
		goto stop_handle;
	}
228
	if (inode->i_blocks)
229
		ext4_truncate(inode);
230
231
232
233
234
235
236

	/*
	 * 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.
	 */
237
	if (!ext4_handle_has_enough_credits(handle, 3)) {
238
239
240
241
		err = ext4_journal_extend(handle, 3);
		if (err > 0)
			err = ext4_journal_restart(handle, 3);
		if (err != 0) {
242
			ext4_warning(inode->i_sb,
243
244
245
				     "couldn't extend journal (err %d)", err);
		stop_handle:
			ext4_journal_stop(handle);
246
			ext4_orphan_del(NULL, inode);
247
248
249
250
			goto no_delete;
		}
	}

251
	/*
252
	 * Kill off the orphan record which ext4_truncate created.
253
	 * AKPM: I think this can be inside the above `if'.
254
	 * Note that ext4_orphan_del() has to be able to cope with the
255
	 * deletion of a non-existent orphan - this is because we don't
256
	 * know if ext4_truncate() actually created an orphan record.
257
258
	 * (Well, we could do this if we need to, but heck - it works)
	 */
259
260
	ext4_orphan_del(handle, inode);
	EXT4_I(inode)->i_dtime	= get_seconds();
261
262
263
264
265
266
267
268

	/*
	 * 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.
	 */
269
	if (ext4_mark_inode_dirty(handle, inode))
270
		/* If that failed, just do the required in-core inode clear. */
Al Viro's avatar
Al Viro committed
271
		ext4_clear_inode(inode);
272
	else
273
274
		ext4_free_inode(handle, inode);
	ext4_journal_stop(handle);
275
276
	return;
no_delete:
Al Viro's avatar
Al Viro committed
277
	ext4_clear_inode(inode);	/* We must guarantee clearing of inode... */
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
}

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;
}

/**
293
 *	ext4_block_to_path - parse the block number into array of offsets
294
295
296
 *	@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
297
298
 *	@boundary: set this non-zero if the referred-to block is likely to be
 *	       followed (on disk) by an indirect block.
299
 *
300
 *	To store the locations of file's data ext4 uses a data structure common
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
 *	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.
 */

323
static int ext4_block_to_path(struct inode *inode,
324
325
			      ext4_lblk_t i_block,
			      ext4_lblk_t offsets[4], int *boundary)
326
{
327
328
329
	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,
330
331
332
333
334
		indirect_blocks = ptrs,
		double_blocks = (1 << (ptrs_bits * 2));
	int n = 0;
	int final = 0;

335
	if (i_block < direct_blocks) {
336
337
		offsets[n++] = i_block;
		final = direct_blocks;
338
	} else if ((i_block -= direct_blocks) < indirect_blocks) {
339
		offsets[n++] = EXT4_IND_BLOCK;
340
341
342
		offsets[n++] = i_block;
		final = ptrs;
	} else if ((i_block -= indirect_blocks) < double_blocks) {
343
		offsets[n++] = EXT4_DIND_BLOCK;
344
345
346
347
		offsets[n++] = i_block >> ptrs_bits;
		offsets[n++] = i_block & (ptrs - 1);
		final = ptrs;
	} else if (((i_block -= double_blocks) >> (ptrs_bits * 2)) < ptrs) {
348
		offsets[n++] = EXT4_TIND_BLOCK;
349
350
351
352
353
		offsets[n++] = i_block >> (ptrs_bits * 2);
		offsets[n++] = (i_block >> ptrs_bits) & (ptrs - 1);
		offsets[n++] = i_block & (ptrs - 1);
		final = ptrs;
	} else {
354
		ext4_warning(inode->i_sb, "block %lu > max in inode %lu",
355
356
			     i_block + direct_blocks +
			     indirect_blocks + double_blocks, inode->i_ino);
357
358
359
360
361
362
363
	}
	if (boundary)
		*boundary = final - 1 - (i_block & (ptrs - 1));
	return n;
}

/**
364
 *	ext4_get_branch - read the chain of indirect blocks leading to data
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
 *	@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).
389
390
 *
 *      Need to be called with
391
 *      down_read(&EXT4_I(inode)->i_data_sem)
392
 */
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
393
394
static Indirect *ext4_get_branch(struct inode *inode, int depth,
				 ext4_lblk_t  *offsets,
395
396
397
398
399
400
401
402
				 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 */
403
	add_chain(chain, NULL, EXT4_I(inode)->i_data + *offsets);
404
405
406
	if (!p->key)
		goto no_block;
	while (--depth) {
407
408
		bh = sb_getblk(sb, le32_to_cpu(p->key));
		if (unlikely(!bh))
409
			goto failure;
410

411
412
413
414
415
416
417
418
419
420
421
		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;
			}
		}
422

423
		add_chain(++p, bh, (__le32 *)bh->b_data + *++offsets);
424
425
426
427
428
429
430
431
432
433
434
435
436
		/* Reader: end */
		if (!p->key)
			goto no_block;
	}
	return NULL;

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

/**
437
 *	ext4_find_near - find a place for allocation with sufficient locality
438
439
440
 *	@inode: owner
 *	@ind: descriptor of indirect block.
 *
441
 *	This function returns the preferred place for block allocation.
442
443
444
445
446
447
448
449
450
451
452
453
454
455
 *	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.
 */
456
static ext4_fsblk_t ext4_find_near(struct inode *inode, Indirect *ind)
457
{
458
	struct ext4_inode_info *ei = EXT4_I(inode);
459
	__le32 *start = ind->bh ? (__le32 *) ind->bh->b_data : ei->i_data;
460
	__le32 *p;
461
	ext4_fsblk_t bg_start;
462
	ext4_fsblk_t last_block;
463
	ext4_grpblk_t colour;
464
465
	ext4_group_t block_group;
	int flex_size = ext4_flex_bg_size(EXT4_SB(inode->i_sb));
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480

	/* 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.
	 */
481
482
483
484
485
486
487
	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);
488
489
	last_block = ext4_blocks_count(EXT4_SB(inode->i_sb)->s_es) - 1;

490
491
492
493
494
495
496
	/*
	 * If we are doing delayed allocation, we don't need take
	 * colour into account.
	 */
	if (test_opt(inode->i_sb, DELALLOC))
		return bg_start;

497
498
	if (bg_start + EXT4_BLOCKS_PER_GROUP(inode->i_sb) <= last_block)
		colour = (current->pid % 16) *
499
			(EXT4_BLOCKS_PER_GROUP(inode->i_sb) / 16);
500
501
	else
		colour = (current->pid % 16) * ((last_block - bg_start) / 16);
502
503
504
505
	return bg_start + colour;
}

/**
506
 *	ext4_find_goal - find a preferred place for allocation.
507
508
509
510
 *	@inode: owner
 *	@block:  block we want
 *	@partial: pointer to the last triple within a chain
 *
511
 *	Normally this function find the preferred place for block allocation,
512
 *	returns it.
513
514
 *	Because this is only used for non-extent files, we limit the block nr
 *	to 32 bits.
515
 */
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
516
static ext4_fsblk_t ext4_find_goal(struct inode *inode, ext4_lblk_t block,
517
				   Indirect *partial)
518
{
519
520
	ext4_fsblk_t goal;

521
	/*
522
	 * XXX need to get goal block from mballoc's data structures
523
524
	 */

525
526
527
	goal = ext4_find_near(inode, partial);
	goal = goal & EXT4_MAX_BLOCK_FILE_PHYS;
	return goal;
528
529
530
}

/**
531
 *	ext4_blks_to_allocate - Look up the block map and count the number
532
533
534
535
536
537
538
539
540
541
 *	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.
 */
542
static int ext4_blks_to_allocate(Indirect *branch, int k, unsigned int blks,
543
				 int blocks_to_boundary)
544
{
545
	unsigned int count = 0;
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568

	/*
	 * 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;
}

/**
569
 *	ext4_alloc_blocks: multiple allocate blocks needed for a branch
570
571
572
573
 *	@handle: handle for this transaction
 *	@inode: inode which needs allocated blocks
 *	@iblock: the logical block to start allocated at
 *	@goal: preferred physical block of allocation
574
575
 *	@indirect_blks: the number of blocks need to allocate for indirect
 *			blocks
576
 *	@blks: number of desired blocks
577
578
 *	@new_blocks: on return it will store the new block numbers for
 *	the indirect blocks(if needed) and the first direct block,
579
580
581
582
 *	@err: on return it will store the error code
 *
 *	This function will return the number of blocks allocated as
 *	requested by the passed-in parameters.
583
 */
584
static int ext4_alloc_blocks(handle_t *handle, struct inode *inode,
585
586
587
			     ext4_lblk_t iblock, ext4_fsblk_t goal,
			     int indirect_blks, int blks,
			     ext4_fsblk_t new_blocks[4], int *err)
588
{
589
	struct ext4_allocation_request ar;
590
	int target, i;
591
	unsigned long count = 0, blk_allocated = 0;
592
	int index = 0;
593
	ext4_fsblk_t current_block = 0;
594
595
596
597
598
599
600
601
602
603
	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)
	 */
604
605
606
	/* first we try to allocate the indirect blocks */
	target = indirect_blks;
	while (target > 0) {
607
608
		count = target;
		/* allocating blocks for indirect blocks and direct blocks */
609
610
		current_block = ext4_new_meta_blocks(handle, inode, goal,
						     0, &count, err);
611
612
613
		if (*err)
			goto failed_out;

614
615
616
617
618
619
620
621
		if (unlikely(current_block + count > EXT4_MAX_BLOCK_FILE_PHYS)) {
			EXT4_ERROR_INODE(inode,
					 "current_block %llu + count %lu > %d!",
					 current_block, count,
					 EXT4_MAX_BLOCK_FILE_PHYS);
			*err = -EIO;
			goto failed_out;
		}
622

623
624
625
626
627
628
		target -= count;
		/* allocate blocks for indirect blocks */
		while (index < indirect_blks && count) {
			new_blocks[index++] = current_block++;
			count--;
		}
629
630
631
632
633
634
635
636
637
		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);
638
			break;
639
		}
640
641
	}

642
643
644
645
646
	target = blks - count ;
	blk_allocated = count;
	if (!target)
		goto allocated;
	/* Now allocate data blocks */
647
648
649
650
651
652
653
654
655
656
	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);
657
658
659
660
661
662
663
664
	if (unlikely(current_block + ar.len > EXT4_MAX_BLOCK_FILE_PHYS)) {
		EXT4_ERROR_INODE(inode,
				 "current_block %llu + ar.len %d > %d!",
				 current_block, ar.len,
				 EXT4_MAX_BLOCK_FILE_PHYS);
		*err = -EIO;
		goto failed_out;
	}
665

666
667
668
669
670
671
672
673
674
	if (*err && (target == blks)) {
		/*
		 * if the allocation failed and we didn't allocate
		 * any blocks before
		 */
		goto failed_out;
	}
	if (!*err) {
		if (target == blks) {
675
676
677
678
			/*
			 * save the new block number
			 * for the first direct block
			 */
679
680
			new_blocks[index] = current_block;
		}
681
		blk_allocated += ar.len;
682
683
	}
allocated:
684
	/* total number of blocks allocated for direct blocks */
685
	ret = blk_allocated;
686
687
688
	*err = 0;
	return ret;
failed_out:
689
	for (i = 0; i < index; i++)
690
		ext4_free_blocks(handle, inode, NULL, new_blocks[i], 1, 0);
691
692
693
694
	return ret;
}

/**
695
 *	ext4_alloc_branch - allocate and set up a chain of blocks.
696
 *	@handle: handle for this transaction
697
698
699
 *	@inode: owner
 *	@indirect_blks: number of allocated indirect blocks
 *	@blks: number of allocated direct blocks
700
 *	@goal: preferred place for allocation
701
702
703
704
705
706
707
 *	@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
708
 *	the same format as ext4_get_branch() would do. We are calling it after
709
710
 *	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
711
 *	picture as after the successful ext4_get_block(), except that in one
712
713
714
715
716
717
 *	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
718
 *	ext4_alloc_block() (normally -ENOSPC). Otherwise we set the chain
719
720
 *	as described above and return 0.
 */
721
static int ext4_alloc_branch(handle_t *handle, struct inode *inode,
722
723
724
			     ext4_lblk_t iblock, int indirect_blks,
			     int *blks, ext4_fsblk_t goal,
			     ext4_lblk_t *offsets, Indirect *branch)
725
726
727
728
729
730
{
	int blocksize = inode->i_sb->s_blocksize;
	int i, n = 0;
	int err = 0;
	struct buffer_head *bh;
	int num;
731
732
	ext4_fsblk_t new_blocks[4];
	ext4_fsblk_t current_block;
733

734
	num = ext4_alloc_blocks(handle, inode, iblock, goal, indirect_blks,
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
				*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]);
750
751
752
753
754
		if (unlikely(!bh)) {
			err = -EIO;
			goto failed;
		}

755
756
757
		branch[n].bh = bh;
		lock_buffer(bh);
		BUFFER_TRACE(bh, "call get_create_access");
758
		err = ext4_journal_get_create_access(handle, bh);
759
		if (err) {
760
761
			/* Don't brelse(bh) here; it's done in
			 * ext4_journal_forget() below */
762
763
764
765
766
767
768
769
			unlock_buffer(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;
770
		if (n == indirect_blks) {
771
772
773
774
775
776
			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
			 */
777
			for (i = 1; i < num; i++)
778
779
780
781
782
783
				*(branch[n].p + i) = cpu_to_le32(++current_block);
		}
		BUFFER_TRACE(bh, "marking uptodate");
		set_buffer_uptodate(bh);
		unlock_buffer(bh);

784
785
		BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
		err = ext4_handle_dirty_metadata(handle, inode, bh);
786
787
788
789
790
791
792
		if (err)
			goto failed;
	}
	*blks = num;
	return err;
failed:
	/* Allocation failed, free what we already allocated */
793
	ext4_free_blocks(handle, inode, NULL, new_blocks[0], 1, 0);
794
	for (i = 1; i <= n ; i++) {
795
		/*
796
797
798
		 * branch[i].bh is newly allocated, so there is no
		 * need to revoke the block, which is why we don't
		 * need to set EXT4_FREE_BLOCKS_METADATA.
799
		 */
800
		ext4_free_blocks(handle, inode, NULL, new_blocks[i], 1,
801
				 EXT4_FREE_BLOCKS_FORGET);
802
	}
803
	for (i = n+1; i < indirect_blks; i++)
804
		ext4_free_blocks(handle, inode, NULL, new_blocks[i], 1, 0);
805

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

	return err;
}

/**
812
 * ext4_splice_branch - splice the allocated branch onto inode.
813
 * @handle: handle for this transaction
814
815
816
 * @inode: owner
 * @block: (logical) number of block we are adding
 * @chain: chain of indirect blocks (with a missing link - see
817
 *	ext4_alloc_branch)
818
819
820
821
822
823
824
825
 * @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.
 */
826
static int ext4_splice_branch(handle_t *handle, struct inode *inode,
827
828
			      ext4_lblk_t block, Indirect *where, int num,
			      int blks)
829
830
831
{
	int i;
	int err = 0;
832
	ext4_fsblk_t current_block;
833
834
835
836
837
838
839
840

	/*
	 * 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");
841
		err = ext4_journal_get_write_access(handle, where->bh);
842
843
844
845
846
847
848
849
850
851
852
853
854
855
		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++)
856
			*(where->p + i) = cpu_to_le32(current_block++);
857
858
859
860
861
862
863
864
865
866
867
	}

	/* We are done with atomic stuff, now do the rest of housekeeping */
	/* 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
868
		 * generic_commit_write->__mark_inode_dirty->ext4_dirty_inode.
869
870
		 */
		jbd_debug(5, "splicing indirect only\n");
871
872
		BUFFER_TRACE(where->bh, "call ext4_handle_dirty_metadata");
		err = ext4_handle_dirty_metadata(handle, inode, where->bh);
873
874
875
876
877
878
		if (err)
			goto err_out;
	} else {
		/*
		 * OK, we spliced it into the inode itself on a direct block.
		 */
879
		ext4_mark_inode_dirty(handle, inode);
880
881
882
883
884
885
		jbd_debug(5, "splicing direct\n");
	}
	return err;

err_out:
	for (i = 1; i <= num; i++) {
886
		/*
887
888
889
		 * branch[i].bh is newly allocated, so there is no
		 * need to revoke the block, which is why we don't
		 * need to set EXT4_FREE_BLOCKS_METADATA.
890
		 */
891
892
		ext4_free_blocks(handle, inode, where[i].bh, 0, 1,
				 EXT4_FREE_BLOCKS_FORGET);
893
	}
894
	ext4_free_blocks(handle, inode, NULL, le32_to_cpu(where[num].key),
895
			 blks, 0);
896
897
898
899
900

	return err;
}

/*
901
 * The ext4_ind_map_blocks() function handles non-extents inodes
902
 * (i.e., using the traditional indirect/double-indirect i_blocks
903
 * scheme) for ext4_map_blocks().
904
 *
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
 * 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.
921
 *
922
923
924
925
926
 * 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.
927
 */
928
929
static int ext4_ind_map_blocks(handle_t *handle, struct inode *inode,
			       struct ext4_map_blocks *map,
930
			       int flags)
931
932
{
	int err = -EIO;
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
933
	ext4_lblk_t offsets[4];
934
935
	Indirect chain[4];
	Indirect *partial;
936
	ext4_fsblk_t goal;
937
938
939
940
	int indirect_blks;
	int blocks_to_boundary = 0;
	int depth;
	int count = 0;
941
	ext4_fsblk_t first_block = 0;
942

943
	trace_ext4_ind_map_blocks_enter(inode, map->m_lblk, map->m_len, flags);
944
	J_ASSERT(!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)));
945
	J_ASSERT(handle != NULL || (flags & EXT4_GET_BLOCKS_CREATE) == 0);
946
	depth = ext4_block_to_path(inode, map->m_lblk, offsets,
947
				   &blocks_to_boundary);
948
949
950
951

	if (depth == 0)
		goto out;

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

	/* Simplest case - block found, no allocation needed */
	if (!partial) {
		first_block = le32_to_cpu(chain[depth - 1].key);
		count++;
		/*map more blocks*/
959
		while (count < map->m_len && count <= blocks_to_boundary) {
960
			ext4_fsblk_t blk;
961
962
963
964
965
966
967
968

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

			if (blk == first_block + count)
				count++;
			else
				break;
		}
969
		goto got_it;
970
971
972
	}

	/* Next simple case - plain lookup or failed read of indirect block */
973
	if ((flags & EXT4_GET_BLOCKS_CREATE) == 0 || err == -EIO)
974
975
976
		goto cleanup;

	/*
977
	 * Okay, we need to do block allocation.
978
	*/
979
	goal = ext4_find_goal(inode, map->m_lblk, partial);
980
981
982
983
984
985
986
987

	/* 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.
	 */
988
	count = ext4_blks_to_allocate(partial, indirect_blks,
989
				      map->m_len, blocks_to_boundary);
990
	/*
991
	 * Block out ext4_truncate while we alter the tree
992
	 */
993
	err = ext4_alloc_branch(handle, inode, map->m_lblk, indirect_blks,
994
995
				&count, goal,
				offsets + (partial - chain), partial);
996
997

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