inode.c 143 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
37
38
#include <linux/mpage.h>
#include <linux/uio.h>
#include <linux/bio.h>
39
#include "ext4_jbd2.h"
40
41
#include "xattr.h"
#include "acl.h"
42
#include "ext4_extents.h"
43

44
45
#define MPAGE_DA_EXTENT_TAIL 0x01

46
47
48
49
50
51
52
static inline int ext4_begin_ordered_truncate(struct inode *inode,
					      loff_t new_size)
{
	return jbd2_journal_begin_ordered_truncate(&EXT4_I(inode)->jinode,
						   new_size);
}

53
54
static void ext4_invalidatepage(struct page *page, unsigned long offset);

55
56
57
/*
 * Test whether an inode is a fast symlink.
 */
58
static int ext4_inode_is_fast_symlink(struct inode *inode)
59
{
60
	int ea_blocks = EXT4_I(inode)->i_file_acl ?
61
62
63
64
65
66
		(inode->i_sb->s_blocksize >> 9) : 0;

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

/*
67
 * The ext4 forget function must perform a revoke if we are freeing data
68
69
70
71
72
73
74
 * 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.
 */
75
76
int ext4_forget(handle_t *handle, int is_metadata, struct inode *inode,
			struct buffer_head *bh, ext4_fsblk_t blocknr)
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
{
	int err;

	might_sleep();

	BUFFER_TRACE(bh, "enter");

	jbd_debug(4, "forgetting bh %p: is_metadata = %d, mode %o, "
		  "data mode %lx\n",
		  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. */

94
95
	if (test_opt(inode->i_sb, DATA_FLAGS) == EXT4_MOUNT_JOURNAL_DATA ||
	    (!is_metadata && !ext4_should_journal_data(inode))) {
96
		if (bh) {
97
			BUFFER_TRACE(bh, "call jbd2_journal_forget");
98
			return ext4_journal_forget(handle, bh);
99
100
101
102
103
104
105
		}
		return 0;
	}

	/*
	 * data!=journal && (is_metadata || should_journal_data(inode))
	 */
106
107
	BUFFER_TRACE(bh, "call ext4_journal_revoke");
	err = ext4_journal_revoke(handle, blocknr, bh);
108
	if (err)
109
		ext4_abort(inode->i_sb, __func__,
110
111
112
113
114
115
116
117
118
119
120
			   "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
121
	ext4_lblk_t needed;
122
123
124
125
126
127

	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
128
	 * like a regular file for ext4 to try to delete it.  Things
129
130
131
132
133
134
135
	 * 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. */
136
137
	if (needed > EXT4_MAX_TRANS_DATA)
		needed = EXT4_MAX_TRANS_DATA;
138

139
	return EXT4_DATA_TRANS_BLOCKS(inode->i_sb) + needed;
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
}

/*
 * 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;

156
	result = ext4_journal_start(inode, blocks_for_truncate(inode));
157
158
159
	if (!IS_ERR(result))
		return result;

160
	ext4_std_error(inode->i_sb, PTR_ERR(result));
161
162
163
164
165
166
167
168
169
170
171
	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)
{
172
	if (handle->h_buffer_credits > EXT4_RESERVE_TRANS_BLOCKS)
173
		return 0;
174
	if (!ext4_journal_extend(handle, blocks_for_truncate(inode)))
175
176
177
178
179
180
181
182
183
		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.
 */
184
static int ext4_journal_test_restart(handle_t *handle, struct inode *inode)
185
186
{
	jbd_debug(2, "restarting handle %p\n", handle);
187
	return ext4_journal_restart(handle, blocks_for_truncate(inode));
188
189
190
191
192
}

/*
 * Called at the last iput() if i_nlink is zero.
 */
193
void ext4_delete_inode(struct inode *inode)
194
195
{
	handle_t *handle;
196
	int err;
197

198
199
	if (ext4_should_order_data(inode))
		ext4_begin_ordered_truncate(inode, 0);
200
201
202
203
204
	truncate_inode_pages(&inode->i_data, 0);

	if (is_bad_inode(inode))
		goto no_delete;

205
	handle = ext4_journal_start(inode, blocks_for_truncate(inode)+3);
206
	if (IS_ERR(handle)) {
207
		ext4_std_error(inode->i_sb, PTR_ERR(handle));
208
209
210
211
212
		/*
		 * 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.
		 */
213
		ext4_orphan_del(NULL, inode);
214
215
216
217
218
219
		goto no_delete;
	}

	if (IS_SYNC(inode))
		handle->h_sync = 1;
	inode->i_size = 0;
220
221
222
223
224
225
	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;
	}
226
	if (inode->i_blocks)
227
		ext4_truncate(inode);
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247

	/*
	 * 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.
	 */
	if (handle->h_buffer_credits < 3) {
		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;
		}
	}

248
	/*
249
	 * Kill off the orphan record which ext4_truncate created.
250
	 * AKPM: I think this can be inside the above `if'.
251
	 * Note that ext4_orphan_del() has to be able to cope with the
252
	 * deletion of a non-existent orphan - this is because we don't
253
	 * know if ext4_truncate() actually created an orphan record.
254
255
	 * (Well, we could do this if we need to, but heck - it works)
	 */
256
257
	ext4_orphan_del(handle, inode);
	EXT4_I(inode)->i_dtime	= get_seconds();
258
259
260
261
262
263
264
265

	/*
	 * 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.
	 */
266
	if (ext4_mark_inode_dirty(handle, inode))
267
268
269
		/* If that failed, just do the required in-core inode clear. */
		clear_inode(inode);
	else
270
271
		ext4_free_inode(handle, inode);
	ext4_journal_stop(handle);
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
	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;
}

/**
290
 *	ext4_block_to_path - parse the block number into array of offsets
291
292
293
 *	@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
294
295
 *	@boundary: set this non-zero if the referred-to block is likely to be
 *	       followed (on disk) by an indirect block.
296
 *
297
 *	To store the locations of file's data ext4 uses a data structure common
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
 *	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.
 */

320
static int ext4_block_to_path(struct inode *inode,
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
321
322
			ext4_lblk_t i_block,
			ext4_lblk_t offsets[4], int *boundary)
323
{
324
325
326
	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,
327
328
329
330
331
332
		indirect_blocks = ptrs,
		double_blocks = (1 << (ptrs_bits * 2));
	int n = 0;
	int final = 0;

	if (i_block < 0) {
333
		ext4_warning(inode->i_sb, "ext4_block_to_path", "block < 0");
334
335
336
	} else if (i_block < direct_blocks) {
		offsets[n++] = i_block;
		final = direct_blocks;
337
	} else if ((i_block -= direct_blocks) < indirect_blocks) {
338
		offsets[n++] = EXT4_IND_BLOCK;
339
340
341
		offsets[n++] = i_block;
		final = ptrs;
	} else if ((i_block -= indirect_blocks) < double_blocks) {
342
		offsets[n++] = EXT4_DIND_BLOCK;
343
344
345
346
		offsets[n++] = i_block >> ptrs_bits;
		offsets[n++] = i_block & (ptrs - 1);
		final = ptrs;
	} else if (((i_block -= double_blocks) >> (ptrs_bits * 2)) < ptrs) {
347
		offsets[n++] = EXT4_TIND_BLOCK;
348
349
350
351
352
		offsets[n++] = i_block >> (ptrs_bits * 2);
		offsets[n++] = (i_block >> ptrs_bits) & (ptrs - 1);
		offsets[n++] = i_block & (ptrs - 1);
		final = ptrs;
	} else {
353
		ext4_warning(inode->i_sb, "ext4_block_to_path",
354
				"block %lu > max",
355
356
				i_block + direct_blocks +
				indirect_blocks + double_blocks);
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
407
408
409
	if (!p->key)
		goto no_block;
	while (--depth) {
		bh = sb_bread(sb, le32_to_cpu(p->key));
		if (!bh)
			goto failure;
410
		add_chain(++p, bh, (__le32 *)bh->b_data + *++offsets);
411
412
413
414
415
416
417
418
419
420
421
422
423
		/* Reader: end */
		if (!p->key)
			goto no_block;
	}
	return NULL;

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

/**
424
 *	ext4_find_near - find a place for allocation with sufficient locality
425
426
427
 *	@inode: owner
 *	@ind: descriptor of indirect block.
 *
428
 *	This function returns the preferred place for block allocation.
429
430
431
432
433
434
435
436
437
438
439
440
441
442
 *	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.
 */
443
static ext4_fsblk_t ext4_find_near(struct inode *inode, Indirect *ind)
444
{
445
	struct ext4_inode_info *ei = EXT4_I(inode);
446
	__le32 *start = ind->bh ? (__le32 *) ind->bh->b_data : ei->i_data;
447
	__le32 *p;
448
	ext4_fsblk_t bg_start;
449
	ext4_fsblk_t last_block;
450
	ext4_grpblk_t colour;
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465

	/* 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.
	 */
466
	bg_start = ext4_group_first_block_no(inode->i_sb, ei->i_block_group);
467
468
469
470
	last_block = ext4_blocks_count(EXT4_SB(inode->i_sb)->s_es) - 1;

	if (bg_start + EXT4_BLOCKS_PER_GROUP(inode->i_sb) <= last_block)
		colour = (current->pid % 16) *
471
			(EXT4_BLOCKS_PER_GROUP(inode->i_sb) / 16);
472
473
	else
		colour = (current->pid % 16) * ((last_block - bg_start) / 16);
474
475
476
477
	return bg_start + colour;
}

/**
478
 *	ext4_find_goal - find a preferred place for allocation.
479
480
481
482
 *	@inode: owner
 *	@block:  block we want
 *	@partial: pointer to the last triple within a chain
 *
483
 *	Normally this function find the preferred place for block allocation,
484
 *	returns it.
485
 */
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
486
static ext4_fsblk_t ext4_find_goal(struct inode *inode, ext4_lblk_t block,
487
		Indirect *partial)
488
489
{
	/*
490
	 * XXX need to get goal block from mballoc's data structures
491
492
	 */

493
	return ext4_find_near(inode, partial);
494
495
496
}

/**
497
 *	ext4_blks_to_allocate: Look up the block map and count the number
498
499
500
501
502
503
504
505
506
507
 *	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.
 */
508
static int ext4_blks_to_allocate(Indirect *branch, int k, unsigned long blks,
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
		int blocks_to_boundary)
{
	unsigned long count = 0;

	/*
	 * 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;
}

/**
535
 *	ext4_alloc_blocks: multiple allocate blocks needed for a branch
536
537
538
539
540
541
542
543
 *	@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
 */
544
static int ext4_alloc_blocks(handle_t *handle, struct inode *inode,
545
546
547
				ext4_lblk_t iblock, ext4_fsblk_t goal,
				int indirect_blks, int blks,
				ext4_fsblk_t new_blocks[4], int *err)
548
549
{
	int target, i;
550
	unsigned long count = 0, blk_allocated = 0;
551
	int index = 0;
552
	ext4_fsblk_t current_block = 0;
553
554
555
556
557
558
559
560
561
562
	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)
	 */
563
564
565
	/* first we try to allocate the indirect blocks */
	target = indirect_blks;
	while (target > 0) {
566
567
		count = target;
		/* allocating blocks for indirect blocks and direct blocks */
568
569
		current_block = ext4_new_meta_blocks(handle, inode,
							goal, &count, err);
570
571
572
573
574
575
576
577
578
		if (*err)
			goto failed_out;

		target -= count;
		/* allocate blocks for indirect blocks */
		while (index < indirect_blks && count) {
			new_blocks[index++] = current_block++;
			count--;
		}
579
580
581
582
583
584
585
586
587
		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);
588
			break;
589
		}
590
591
	}

592
593
594
595
596
597
	target = blks - count ;
	blk_allocated = count;
	if (!target)
		goto allocated;
	/* Now allocate data blocks */
	count = target;
598
	/* allocating blocks for data blocks */
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
	current_block = ext4_new_blocks(handle, inode, iblock,
						goal, &count, err);
	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;
		}
		blk_allocated += count;
	}
allocated:
619
	/* total number of blocks allocated for direct blocks */
620
	ret = blk_allocated;
621
622
623
	*err = 0;
	return ret;
failed_out:
624
	for (i = 0; i < index; i++)
625
		ext4_free_blocks(handle, inode, new_blocks[i], 1, 0);
626
627
628
629
	return ret;
}

/**
630
 *	ext4_alloc_branch - allocate and set up a chain of blocks.
631
632
633
634
635
636
637
638
639
640
 *	@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
641
 *	the same format as ext4_get_branch() would do. We are calling it after
642
643
 *	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
644
 *	picture as after the successful ext4_get_block(), except that in one
645
646
647
648
649
650
 *	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
651
 *	ext4_alloc_block() (normally -ENOSPC). Otherwise we set the chain
652
653
 *	as described above and return 0.
 */
654
static int ext4_alloc_branch(handle_t *handle, struct inode *inode,
655
656
657
				ext4_lblk_t iblock, int indirect_blks,
				int *blks, ext4_fsblk_t goal,
				ext4_lblk_t *offsets, Indirect *branch)
658
659
660
661
662
663
{
	int blocksize = inode->i_sb->s_blocksize;
	int i, n = 0;
	int err = 0;
	struct buffer_head *bh;
	int num;
664
665
	ext4_fsblk_t new_blocks[4];
	ext4_fsblk_t current_block;
666

667
	num = ext4_alloc_blocks(handle, inode, iblock, goal, indirect_blks,
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
				*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");
686
		err = ext4_journal_get_create_access(handle, bh);
687
688
689
690
691
692
693
694
695
696
		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;
697
		if (n == indirect_blks) {
698
699
700
701
702
703
704
705
706
707
708
709
710
			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);

711
712
		BUFFER_TRACE(bh, "call ext4_journal_dirty_metadata");
		err = ext4_journal_dirty_metadata(handle, bh);
713
714
715
716
717
718
719
720
		if (err)
			goto failed;
	}
	*blks = num;
	return err;
failed:
	/* Allocation failed, free what we already allocated */
	for (i = 1; i <= n ; i++) {
721
		BUFFER_TRACE(branch[i].bh, "call jbd2_journal_forget");
722
		ext4_journal_forget(handle, branch[i].bh);
723
	}
724
	for (i = 0; i < indirect_blks; i++)
725
		ext4_free_blocks(handle, inode, new_blocks[i], 1, 0);
726

727
	ext4_free_blocks(handle, inode, new_blocks[i], num, 0);
728
729
730
731
732

	return err;
}

/**
733
 * ext4_splice_branch - splice the allocated branch onto inode.
734
735
736
 * @inode: owner
 * @block: (logical) number of block we are adding
 * @chain: chain of indirect blocks (with a missing link - see
737
 *	ext4_alloc_branch)
738
739
740
741
742
743
744
745
 * @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.
 */
746
static int ext4_splice_branch(handle_t *handle, struct inode *inode,
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
747
			ext4_lblk_t block, Indirect *where, int num, int blks)
748
749
750
{
	int i;
	int err = 0;
751
	ext4_fsblk_t current_block;
752
753
754
755
756
757
758
759

	/*
	 * 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");
760
		err = ext4_journal_get_write_access(handle, where->bh);
761
762
763
764
765
766
767
768
769
770
771
772
773
774
		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++)
775
			*(where->p + i) = cpu_to_le32(current_block++);
776
777
778
779
	}

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

Kalpak Shah's avatar
Kalpak Shah committed
780
	inode->i_ctime = ext4_current_time(inode);
781
	ext4_mark_inode_dirty(handle, inode);
782
783
784
785
786
787
788
789
790

	/* 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
791
		 * generic_commit_write->__mark_inode_dirty->ext4_dirty_inode.
792
793
		 */
		jbd_debug(5, "splicing indirect only\n");
794
795
		BUFFER_TRACE(where->bh, "call ext4_journal_dirty_metadata");
		err = ext4_journal_dirty_metadata(handle, where->bh);
796
797
798
799
800
801
802
803
804
805
806
807
808
		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++) {
809
		BUFFER_TRACE(where[i].bh, "call jbd2_journal_forget");
810
		ext4_journal_forget(handle, where[i].bh);
811
812
		ext4_free_blocks(handle, inode,
					le32_to_cpu(where[i-1].key), 1, 0);
813
	}
814
	ext4_free_blocks(handle, inode, le32_to_cpu(where[num].key), blks, 0);
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835

	return err;
}

/*
 * 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.
836
837
838
 *
 *
 * Need to be called with
839
840
 * down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system block
 * (ie, create is zero). Otherwise down_write(&EXT4_I(inode)->i_data_sem)
841
 */
842
int ext4_get_blocks_handle(handle_t *handle, struct inode *inode,
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
843
		ext4_lblk_t iblock, unsigned long maxblocks,
844
845
846
847
		struct buffer_head *bh_result,
		int create, int extend_disksize)
{
	int err = -EIO;
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
848
	ext4_lblk_t offsets[4];
849
850
	Indirect chain[4];
	Indirect *partial;
851
	ext4_fsblk_t goal;
852
853
854
	int indirect_blks;
	int blocks_to_boundary = 0;
	int depth;
855
	struct ext4_inode_info *ei = EXT4_I(inode);
856
	int count = 0;
857
	ext4_fsblk_t first_block = 0;
858
	loff_t disksize;
859
860


861
	J_ASSERT(!(EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL));
862
	J_ASSERT(handle != NULL || create == 0);
Aneesh Kumar K.V's avatar
Aneesh Kumar K.V committed
863
864
	depth = ext4_block_to_path(inode, iblock, offsets,
					&blocks_to_boundary);
865
866
867
868

	if (depth == 0)
		goto out;

869
	partial = ext4_get_branch(inode, depth, offsets, chain, &err);
870
871
872
873
874
875
876
877

	/* 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) {
878
			ext4_fsblk_t blk;
879
880
881
882
883
884
885
886

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

			if (blk == first_block + count)
				count++;
			else
				break;
		}
887
		goto got_it;
888
889
890
891
892
893
894
	}

	/* Next simple case - plain lookup or failed read of indirect block */
	if (!create || err == -EIO)
		goto cleanup;

	/*
895
	 * Okay, we need to do block allocation.
896
	*/
897
	goal = ext4_find_goal(inode, iblock, partial);
898
899
900
901
902
903
904
905

	/* 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.
	 */
906
	count = ext4_blks_to_allocate(partial, indirect_blks,
907
908
					maxblocks, blocks_to_boundary);
	/*
909
	 * Block out ext4_truncate while we alter the tree
910
	 */
911
912
913
	err = ext4_alloc_branch(handle, inode, iblock, indirect_blks,
					&count, goal,
					offsets + (partial - chain), partial);
914
915

	/*
916
	 * The ext4_splice_branch call will free and forget any buffers
917
918
919
920
921
922
	 * 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
	 * may need to return -EAGAIN upwards in the worst case.  --sct
	 */
	if (!err)
923
		err = ext4_splice_branch(handle, inode, iblock,
924
925
					partial, indirect_blks, count);
	/*
926
	 * i_disksize growing is protected by i_data_sem.  Don't forget to
927
	 * protect it if you're about to implement concurrent
928
	 * ext4_get_block() -bzzz
929
	*/
930
931
932
933
934
935
936
	if (!err && extend_disksize) {
		disksize = ((loff_t) iblock + count) << inode->i_blkbits;
		if (disksize > i_size_read(inode))
			disksize = i_size_read(inode);
		if (disksize > ei->i_disksize)
			ei->i_disksize = disksize;
	}
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
	if (err)
		goto cleanup;

	set_buffer_new(bh_result);
got_it:
	map_bh(bh_result, inode->i_sb, le32_to_cpu(chain[depth-1].key));
	if (count > blocks_to_boundary)
		set_buffer_boundary(bh_result);
	err = count;
	/* Clean up and exit */
	partial = chain + depth - 1;	/* the whole chain */
cleanup:
	while (partial > chain) {
		BUFFER_TRACE(partial->bh, "call brelse");
		brelse(partial->bh);
		partial--;
	}
	BUFFER_TRACE(bh_result, "returned");
out:
	return err;
}

959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
/*
 * Calculate the number of metadata blocks need to reserve
 * to allocate @blocks for non extent file based file
 */
static int ext4_indirect_calc_metadata_amount(struct inode *inode, int blocks)
{
	int icap = EXT4_ADDR_PER_BLOCK(inode->i_sb);
	int ind_blks, dind_blks, tind_blks;

	/* number of new indirect blocks needed */
	ind_blks = (blocks + icap - 1) / icap;

	dind_blks = (ind_blks + icap - 1) / icap;

	tind_blks = 1;

	return ind_blks + dind_blks + tind_blks;
}

/*
 * Calculate the number of metadata blocks need to reserve
 * to allocate given number of blocks
 */
static int ext4_calc_metadata_amount(struct inode *inode, int blocks)
{
984
985
986
	if (!blocks)
		return 0;

987
988
989
990
991
992
993
994
995
996
997
998
999
1000
	if (EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL)
		return ext4_ext_calc_metadata_amount(inode, blocks);

	return ext4_indirect_calc_metadata_amount(inode, blocks);
}

static void ext4_da_update_reserve_space(struct inode *inode, int used)
{
	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
	int total, mdb, mdb_free;

	spin_lock(&EXT4_I(inode)->i_block_reservation_lock);
	/* recalculate the number of metablocks still need to be reserved */
	total = EXT4_I(inode)->i_reserved_data_blocks - used;
For faster browsing, not all history is shown. View entire blame