balloc.c 22.9 KB
Newer Older
1
/*
2
 *  linux/fs/ext4/balloc.c
3 4 5 6 7 8 9 10 11 12 13 14 15 16
 *
 * Copyright (C) 1992, 1993, 1994, 1995
 * Remy Card (card@masi.ibp.fr)
 * Laboratoire MASI - Institut Blaise Pascal
 * Universite Pierre et Marie Curie (Paris VI)
 *
 *  Enhanced block allocation by Stephen Tweedie (sct@redhat.com), 1993
 *  Big-endian to little-endian byte-swapping/bitmaps by
 *        David S. Miller (davem@caip.rutgers.edu), 1995
 */

#include <linux/time.h>
#include <linux/capability.h>
#include <linux/fs.h>
17
#include <linux/jbd2.h>
18 19
#include <linux/quotaops.h>
#include <linux/buffer_head.h>
20 21
#include "ext4.h"
#include "ext4_jbd2.h"
22
#include "mballoc.h"
23

24 25
#include <trace/events/ext4.h>

E
Eric Sandeen 已提交
26 27
static unsigned ext4_num_base_meta_clusters(struct super_block *sb,
					    ext4_group_t block_group);
28 29 30 31
/*
 * balloc.c contains the blocks allocation and deallocation routines
 */

32
/*
33 34
 * Calculate the block group number and offset into the block/cluster
 * allocation bitmap, given a block number
35 36
 */
void ext4_get_group_no_and_offset(struct super_block *sb, ext4_fsblk_t blocknr,
37
		ext4_group_t *blockgrpp, ext4_grpblk_t *offsetp)
38
{
D
Dave Kleikamp 已提交
39
	struct ext4_super_block *es = EXT4_SB(sb)->s_es;
40 41
	ext4_grpblk_t offset;

D
Dave Kleikamp 已提交
42
	blocknr = blocknr - le32_to_cpu(es->s_first_data_block);
43 44
	offset = do_div(blocknr, EXT4_BLOCKS_PER_GROUP(sb)) >>
		EXT4_SB(sb)->s_cluster_bits;
45 46 47
	if (offsetp)
		*offsetp = offset;
	if (blockgrpp)
D
Dave Kleikamp 已提交
48
		*blockgrpp = blocknr;
49 50 51

}

52 53 54 55
static int ext4_block_in_group(struct super_block *sb, ext4_fsblk_t block,
			ext4_group_t block_group)
{
	ext4_group_t actual_group;
A
Aneesh Kumar K.V 已提交
56
	ext4_get_group_no_and_offset(sb, block, &actual_group, NULL);
57 58 59 60 61
	if (actual_group == block_group)
		return 1;
	return 0;
}

62 63 64 65 66 67
/* Return the number of clusters used for file system metadata; this
 * represents the overhead needed by the file system.
 */
unsigned ext4_num_overhead_clusters(struct super_block *sb,
				    ext4_group_t block_group,
				    struct ext4_group_desc *gdp)
68
{
69 70 71 72
	unsigned num_clusters;
	int block_cluster = -1, inode_cluster = -1, itbl_cluster = -1, i, c;
	ext4_fsblk_t start = ext4_group_first_block_no(sb, block_group);
	ext4_fsblk_t itbl_blk;
73 74
	struct ext4_sb_info *sbi = EXT4_SB(sb);

75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92
	/* This is the number of clusters used by the superblock,
	 * block group descriptors, and reserved block group
	 * descriptor blocks */
	num_clusters = ext4_num_base_meta_clusters(sb, block_group);

	/*
	 * For the allocation bitmaps and inode table, we first need
	 * to check to see if the block is in the block group.  If it
	 * is, then check to see if the cluster is already accounted
	 * for in the clusters used for the base metadata cluster, or
	 * if we can increment the base metadata cluster to include
	 * that block.  Otherwise, we will have to track the cluster
	 * used for the allocation bitmap or inode table explicitly.
	 * Normally all of these blocks are contiguous, so the special
	 * case handling shouldn't be necessary except for *very*
	 * unusual file system layouts.
	 */
	if (ext4_block_in_group(sb, ext4_block_bitmap(sb, gdp), block_group)) {
93 94
		block_cluster = EXT4_B2C(sbi,
					 ext4_block_bitmap(sb, gdp) - start);
95 96 97 98 99 100 101 102 103 104
		if (block_cluster < num_clusters)
			block_cluster = -1;
		else if (block_cluster == num_clusters) {
			num_clusters++;
			block_cluster = -1;
		}
	}

	if (ext4_block_in_group(sb, ext4_inode_bitmap(sb, gdp), block_group)) {
		inode_cluster = EXT4_B2C(sbi,
105
					 ext4_inode_bitmap(sb, gdp) - start);
106 107 108 109 110 111 112 113 114 115 116
		if (inode_cluster < num_clusters)
			inode_cluster = -1;
		else if (inode_cluster == num_clusters) {
			num_clusters++;
			inode_cluster = -1;
		}
	}

	itbl_blk = ext4_inode_table(sb, gdp);
	for (i = 0; i < sbi->s_itb_per_group; i++) {
		if (ext4_block_in_group(sb, itbl_blk + i, block_group)) {
117
			c = EXT4_B2C(sbi, itbl_blk + i - start);
118 119 120 121 122 123 124 125 126
			if ((c < num_clusters) || (c == inode_cluster) ||
			    (c == block_cluster) || (c == itbl_cluster))
				continue;
			if (c == num_clusters) {
				num_clusters++;
				continue;
			}
			num_clusters++;
			itbl_cluster = c;
127 128
		}
	}
129 130 131 132 133 134 135

	if (block_cluster != -1)
		num_clusters++;
	if (inode_cluster != -1)
		num_clusters++;

	return num_clusters;
136
}
137

138 139
static unsigned int num_clusters_in_group(struct super_block *sb,
					  ext4_group_t block_group)
140
{
141 142
	unsigned int blocks;

143 144 145 146 147 148 149
	if (block_group == ext4_get_groups_count(sb) - 1) {
		/*
		 * Even though mke2fs always initializes the first and
		 * last group, just in case some other tool was used,
		 * we need to make sure we calculate the right free
		 * blocks.
		 */
150
		blocks = ext4_blocks_count(EXT4_SB(sb)->s_es) -
151 152
			ext4_group_first_block_no(sb, block_group);
	} else
153 154
		blocks = EXT4_BLOCKS_PER_GROUP(sb);
	return EXT4_NUM_B2C(EXT4_SB(sb), blocks);
155 156
}

157 158 159 160
/* Initializes an uninitialized block bitmap */
void ext4_init_block_bitmap(struct super_block *sb, struct buffer_head *bh,
			    ext4_group_t block_group,
			    struct ext4_group_desc *gdp)
A
Andreas Dilger 已提交
161
{
162
	unsigned int bit, bit_max;
A
Andreas Dilger 已提交
163
	struct ext4_sb_info *sbi = EXT4_SB(sb);
164 165 166 167 168 169 170
	ext4_fsblk_t start, tmp;
	int flex_bg = 0;

	J_ASSERT_BH(bh, buffer_locked(bh));

	/* If checksum is bad mark all blocks used to prevent allocation
	 * essentially implementing a per-group read-only flag. */
171
	if (!ext4_group_desc_csum_verify(sb, block_group, gdp)) {
172
		ext4_error(sb, "Checksum bad for group %u", block_group);
173
		ext4_free_group_clusters_set(sb, gdp, 0);
174 175 176
		ext4_free_inodes_set(sb, gdp, 0);
		ext4_itable_unused_set(sb, gdp, 0);
		memset(bh->b_data, 0xff, sb->s_blocksize);
177 178
		ext4_block_bitmap_csum_set(sb, block_group, gdp, bh,
					   EXT4_BLOCKS_PER_GROUP(sb) / 8);
179
		return;
A
Andreas Dilger 已提交
180
	}
181
	memset(bh->b_data, 0, sb->s_blocksize);
A
Andreas Dilger 已提交
182

183
	bit_max = ext4_num_base_meta_clusters(sb, block_group);
184 185
	for (bit = 0; bit < bit_max; bit++)
		ext4_set_bit(bit, bh->b_data);
186

187
	start = ext4_group_first_block_no(sb, block_group);
A
Andreas Dilger 已提交
188

189 190
	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_FLEX_BG))
		flex_bg = 1;
A
Andreas Dilger 已提交
191

192 193 194
	/* Set bits for block and inode bitmaps, and inode table */
	tmp = ext4_block_bitmap(sb, gdp);
	if (!flex_bg || ext4_block_in_group(sb, tmp, block_group))
195
		ext4_set_bit(EXT4_B2C(sbi, tmp - start), bh->b_data);
A
Andreas Dilger 已提交
196

197 198
	tmp = ext4_inode_bitmap(sb, gdp);
	if (!flex_bg || ext4_block_in_group(sb, tmp, block_group))
199
		ext4_set_bit(EXT4_B2C(sbi, tmp - start), bh->b_data);
200

201 202 203
	tmp = ext4_inode_table(sb, gdp);
	for (; tmp < ext4_inode_table(sb, gdp) +
		     sbi->s_itb_per_group; tmp++) {
204
		if (!flex_bg || ext4_block_in_group(sb, tmp, block_group))
205
			ext4_set_bit(EXT4_B2C(sbi, tmp - start), bh->b_data);
A
Andreas Dilger 已提交
206
	}
207

208 209 210 211 212
	/*
	 * Also if the number of blocks within the group is less than
	 * the blocksize * 8 ( which is the size of bitmap ), set rest
	 * of the block bitmap to 1
	 */
213
	ext4_mark_bitmap_end(num_clusters_in_group(sb, block_group),
214
			     sb->s_blocksize * 8, bh->b_data);
215 216
	ext4_block_bitmap_csum_set(sb, block_group, gdp, bh,
				   EXT4_BLOCKS_PER_GROUP(sb) / 8);
217
	ext4_group_desc_csum_set(sb, block_group, gdp);
A
Andreas Dilger 已提交
218 219
}

220 221 222
/* Return the number of free blocks in a block group.  It is used when
 * the block bitmap is uninitialized, so we can't just count the bits
 * in the bitmap. */
223 224 225
unsigned ext4_free_clusters_after_init(struct super_block *sb,
				       ext4_group_t block_group,
				       struct ext4_group_desc *gdp)
226
{
227 228
	return num_clusters_in_group(sb, block_group) - 
		ext4_num_overhead_clusters(sb, block_group, gdp);
229
}
A
Andreas Dilger 已提交
230

231 232 233 234 235 236 237 238
/*
 * The free blocks are managed by bitmaps.  A file system contains several
 * blocks groups.  Each group contains 1 bitmap block for blocks, 1 bitmap
 * block for inodes, N blocks for the inode table and data blocks.
 *
 * The file system contains group descriptors which are located after the
 * super block.  Each descriptor contains the number of the bitmap block and
 * the free blocks count in the block.  The descriptors are loaded in memory
239
 * when a file system is mounted (see ext4_fill_super).
240 241 242
 */

/**
243
 * ext4_get_group_desc() -- load group descriptor from disk
244 245 246 247 248
 * @sb:			super block
 * @block_group:	given block group
 * @bh:			pointer to the buffer head to store the block
 *			group descriptor
 */
249
struct ext4_group_desc * ext4_get_group_desc(struct super_block *sb,
250
					     ext4_group_t block_group,
251
					     struct buffer_head **bh)
252
{
253 254
	unsigned int group_desc;
	unsigned int offset;
255
	ext4_group_t ngroups = ext4_get_groups_count(sb);
256
	struct ext4_group_desc *desc;
257
	struct ext4_sb_info *sbi = EXT4_SB(sb);
258

259
	if (block_group >= ngroups) {
260 261
		ext4_error(sb, "block_group >= groups_count - block_group = %u,"
			   " groups_count = %u", block_group, ngroups);
262 263 264 265

		return NULL;
	}

266 267
	group_desc = block_group >> EXT4_DESC_PER_BLOCK_BITS(sb);
	offset = block_group & (EXT4_DESC_PER_BLOCK(sb) - 1);
268
	if (!sbi->s_group_desc[group_desc]) {
269
		ext4_error(sb, "Group descriptor not loaded - "
270
			   "block_group = %u, group_desc = %u, desc = %u",
271
			   block_group, group_desc, offset);
272 273 274
		return NULL;
	}

275 276 277
	desc = (struct ext4_group_desc *)(
		(__u8 *)sbi->s_group_desc[group_desc]->b_data +
		offset * EXT4_DESC_SIZE(sb));
278 279
	if (bh)
		*bh = sbi->s_group_desc[group_desc];
280
	return desc;
281 282
}

283
static int ext4_valid_block_bitmap(struct super_block *sb,
284 285 286
				   struct ext4_group_desc *desc,
				   unsigned int block_group,
				   struct buffer_head *bh)
287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328
{
	ext4_grpblk_t offset;
	ext4_grpblk_t next_zero_bit;
	ext4_fsblk_t bitmap_blk;
	ext4_fsblk_t group_first_block;

	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_FLEX_BG)) {
		/* with FLEX_BG, the inode/block bitmaps and itable
		 * blocks may not be in the group at all
		 * so the bitmap validation will be skipped for those groups
		 * or it has to also read the block group where the bitmaps
		 * are located to verify they are set.
		 */
		return 1;
	}
	group_first_block = ext4_group_first_block_no(sb, block_group);

	/* check whether block bitmap block number is set */
	bitmap_blk = ext4_block_bitmap(sb, desc);
	offset = bitmap_blk - group_first_block;
	if (!ext4_test_bit(offset, bh->b_data))
		/* bad block bitmap */
		goto err_out;

	/* check whether the inode bitmap block number is set */
	bitmap_blk = ext4_inode_bitmap(sb, desc);
	offset = bitmap_blk - group_first_block;
	if (!ext4_test_bit(offset, bh->b_data))
		/* bad block bitmap */
		goto err_out;

	/* check whether the inode table block number is set */
	bitmap_blk = ext4_inode_table(sb, desc);
	offset = bitmap_blk - group_first_block;
	next_zero_bit = ext4_find_next_zero_bit(bh->b_data,
				offset + EXT4_SB(sb)->s_itb_per_group,
				offset);
	if (next_zero_bit >= offset + EXT4_SB(sb)->s_itb_per_group)
		/* good bitmap for inode tables */
		return 1;

err_out:
329
	ext4_error(sb, "Invalid block bitmap - block_group = %d, block = %llu",
330 331 332
			block_group, bitmap_blk);
	return 0;
}
333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349

void ext4_validate_block_bitmap(struct super_block *sb,
			       struct ext4_group_desc *desc,
			       unsigned int block_group,
			       struct buffer_head *bh)
{
	if (buffer_verified(bh))
		return;

	ext4_lock_group(sb, block_group);
	if (ext4_valid_block_bitmap(sb, desc, block_group, bh) &&
	    ext4_block_bitmap_csum_verify(sb, block_group, desc, bh,
					  EXT4_BLOCKS_PER_GROUP(sb) / 8))
		set_buffer_verified(bh);
	ext4_unlock_group(sb, block_group);
}

350
/**
351
 * ext4_read_block_bitmap()
352 353 354
 * @sb:			super block
 * @block_group:	given block group
 *
355 356
 * Read the bitmap for a given block_group,and validate the
 * bits for block/inode/inode tables are set in the bitmaps
357 358 359
 *
 * Return buffer_head on success or NULL in case of failure.
 */
A
Andreas Dilger 已提交
360
struct buffer_head *
361
ext4_read_block_bitmap_nowait(struct super_block *sb, ext4_group_t block_group)
362
{
363
	struct ext4_group_desc *desc;
364
	struct buffer_head *bh;
365
	ext4_fsblk_t bitmap_blk;
366

A
Andreas Dilger 已提交
367
	desc = ext4_get_group_desc(sb, block_group, NULL);
368
	if (!desc)
369 370
		return NULL;
	bitmap_blk = ext4_block_bitmap(sb, desc);
371 372
	bh = sb_getblk(sb, bitmap_blk);
	if (unlikely(!bh)) {
373 374 375
		ext4_error(sb, "Cannot get buffer for block bitmap - "
			   "block_group = %u, block_bitmap = %llu",
			   block_group, bitmap_blk);
376 377
		return NULL;
	}
378 379

	if (bitmap_uptodate(bh))
380
		goto verify;
381

382
	lock_buffer(bh);
383 384
	if (bitmap_uptodate(bh)) {
		unlock_buffer(bh);
385
		goto verify;
386
	}
387
	ext4_lock_group(sb, block_group);
A
Andreas Dilger 已提交
388
	if (desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) {
389
		ext4_init_block_bitmap(sb, bh, block_group, desc);
390
		set_bitmap_uptodate(bh);
391
		set_buffer_uptodate(bh);
392
		ext4_unlock_group(sb, block_group);
A
Aneesh Kumar K.V 已提交
393
		unlock_buffer(bh);
394
		return bh;
A
Andreas Dilger 已提交
395
	}
396
	ext4_unlock_group(sb, block_group);
397 398 399 400 401 402 403
	if (buffer_uptodate(bh)) {
		/*
		 * if not uninit if bh is uptodate,
		 * bitmap is also uptodate
		 */
		set_bitmap_uptodate(bh);
		unlock_buffer(bh);
404
		goto verify;
405 406
	}
	/*
407
	 * submit the buffer_head for reading
408
	 */
409
	set_buffer_new(bh);
410
	trace_ext4_read_block_bitmap_load(sb, block_group);
411 412 413 414
	bh->b_end_io = ext4_end_bitmap_read;
	get_bh(bh);
	submit_bh(READ, bh);
	return bh;
415 416 417
verify:
	ext4_validate_block_bitmap(sb, desc, block_group, bh);
	return bh;
418 419 420 421 422 423 424 425 426 427 428 429 430 431 432
}

/* Returns 0 on success, 1 on error */
int ext4_wait_block_bitmap(struct super_block *sb, ext4_group_t block_group,
			   struct buffer_head *bh)
{
	struct ext4_group_desc *desc;

	if (!buffer_new(bh))
		return 0;
	desc = ext4_get_group_desc(sb, block_group, NULL);
	if (!desc)
		return 1;
	wait_on_buffer(bh);
	if (!buffer_uptodate(bh)) {
433
		ext4_error(sb, "Cannot read block bitmap - "
434
			   "block_group = %u, block_bitmap = %llu",
435
			   block_group, (unsigned long long) bh->b_blocknr);
436
		return 1;
437
	}
438 439
	clear_buffer_new(bh);
	/* Panic or remount fs read-only if block bitmap is invalid */
440
	ext4_validate_block_bitmap(sb, desc, block_group, bh);
441 442 443 444 445 446 447 448 449 450 451 452 453
	return 0;
}

struct buffer_head *
ext4_read_block_bitmap(struct super_block *sb, ext4_group_t block_group)
{
	struct buffer_head *bh;

	bh = ext4_read_block_bitmap_nowait(sb, block_group);
	if (ext4_wait_block_bitmap(sb, block_group, bh)) {
		put_bh(bh);
		return NULL;
	}
454 455 456
	return bh;
}

457
/**
458
 * ext4_has_free_clusters()
459
 * @sbi:	in-core super block structure.
460 461
 * @nclusters:	number of needed blocks
 * @flags:	flags from ext4_mb_new_blocks()
462
 *
463
 * Check if filesystem has nclusters free & available for allocation.
464 465
 * On success return 1, return 0 on failure.
 */
466 467
static int ext4_has_free_clusters(struct ext4_sb_info *sbi,
				  s64 nclusters, unsigned int flags)
468
{
469
	s64 free_clusters, dirty_clusters, root_clusters;
470
	struct percpu_counter *fcc = &sbi->s_freeclusters_counter;
471
	struct percpu_counter *dcc = &sbi->s_dirtyclusters_counter;
472

473 474 475
	free_clusters  = percpu_counter_read_positive(fcc);
	dirty_clusters = percpu_counter_read_positive(dcc);
	root_clusters = EXT4_B2C(sbi, ext4_r_blocks_count(sbi->s_es));
476

477 478 479 480
	if (free_clusters - (nclusters + root_clusters + dirty_clusters) <
					EXT4_FREECLUSTERS_WATERMARK) {
		free_clusters  = EXT4_C2B(sbi, percpu_counter_sum_positive(fcc));
		dirty_clusters = percpu_counter_sum_positive(dcc);
481
	}
482 483
	/* Check whether we have space after accounting for current
	 * dirty clusters & root reserved clusters.
484
	 */
485
	if (free_clusters >= ((root_clusters + nclusters) + dirty_clusters))
486
		return 1;
487

488
	/* Hm, nope.  Are (enough) root reserved clusters available? */
489 490
	if (uid_eq(sbi->s_resuid, current_fsuid()) ||
	    (!gid_eq(sbi->s_resgid, GLOBAL_ROOT_GID) && in_group_p(sbi->s_resgid)) ||
491 492 493
	    capable(CAP_SYS_RESOURCE) ||
		(flags & EXT4_MB_USE_ROOT_BLOCKS)) {

494
		if (free_clusters >= (nclusters + dirty_clusters))
495 496 497 498
			return 1;
	}

	return 0;
499 500
}

501 502
int ext4_claim_free_clusters(struct ext4_sb_info *sbi,
			     s64 nclusters, unsigned int flags)
503
{
504
	if (ext4_has_free_clusters(sbi, nclusters, flags)) {
505
		percpu_counter_add(&sbi->s_dirtyclusters_counter, nclusters);
506
		return 0;
507 508
	} else
		return -ENOSPC;
509
}
510

511
/**
512
 * ext4_should_retry_alloc()
513 514 515
 * @sb:			super block
 * @retries		number of attemps has been made
 *
516
 * ext4_should_retry_alloc() is called when ENOSPC is returned, and if
517
 * it is profitable to retry the operation, this function will wait
L
Lucas De Marchi 已提交
518
 * for the current or committing transaction to complete, and then
519 520 521 522
 * return TRUE.
 *
 * if the total number of retries exceed three times, return FALSE.
 */
523
int ext4_should_retry_alloc(struct super_block *sb, int *retries)
524
{
525
	if (!ext4_has_free_clusters(EXT4_SB(sb), 1, 0) ||
526 527
	    (*retries)++ > 3 ||
	    !EXT4_SB(sb)->s_journal)
528 529 530 531
		return 0;

	jbd_debug(1, "%s: retrying operation after ENOSPC\n", sb->s_id);

532
	return jbd2_journal_force_commit_nested(EXT4_SB(sb)->s_journal);
533 534
}

A
Aneesh Kumar K.V 已提交
535
/*
536
 * ext4_new_meta_blocks() -- allocate block for meta data (indexing) blocks
A
Aneesh Kumar K.V 已提交
537 538 539 540
 *
 * @handle:             handle to this transaction
 * @inode:              file inode
 * @goal:               given target block(filesystem wide)
541
 * @count:		pointer to total number of clusters needed
A
Aneesh Kumar K.V 已提交
542 543
 * @errp:               error code
 *
T
Theodore Ts'o 已提交
544
 * Return 1st allocated block number on success, *count stores total account
545
 * error stores in errp pointer
A
Aneesh Kumar K.V 已提交
546
 */
547
ext4_fsblk_t ext4_new_meta_blocks(handle_t *handle, struct inode *inode,
548 549
				  ext4_fsblk_t goal, unsigned int flags,
				  unsigned long *count, int *errp)
A
Aneesh Kumar K.V 已提交
550
{
T
Theodore Ts'o 已提交
551
	struct ext4_allocation_request ar;
552
	ext4_fsblk_t ret;
T
Theodore Ts'o 已提交
553 554 555 556 557 558

	memset(&ar, 0, sizeof(ar));
	/* Fill with neighbour allocated blocks */
	ar.inode = inode;
	ar.goal = goal;
	ar.len = count ? *count : 1;
559
	ar.flags = flags;
T
Theodore Ts'o 已提交
560 561 562 563

	ret = ext4_mb_new_blocks(handle, &ar, errp);
	if (count)
		*count = ar.len;
564
	/*
565 566
	 * Account for the allocated meta blocks.  We will never
	 * fail EDQUOT for metdata, but we do account for it.
567
	 */
568 569
	if (!(*errp) &&
	    ext4_test_inode_state(inode, EXT4_STATE_DELALLOC_RESERVED)) {
570
		spin_lock(&EXT4_I(inode)->i_block_reservation_lock);
T
Theodore Ts'o 已提交
571
		EXT4_I(inode)->i_allocated_meta_blocks += ar.len;
572
		spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
573 574
		dquot_alloc_block_nofail(inode,
				EXT4_C2B(EXT4_SB(inode->i_sb), ar.len));
575 576
	}
	return ret;
A
Aneesh Kumar K.V 已提交
577 578
}

579
/**
580
 * ext4_count_free_clusters() -- count filesystem free clusters
581 582
 * @sb:		superblock
 *
583
 * Adds up the number of free clusters from each block group.
584
 */
585
ext4_fsblk_t ext4_count_free_clusters(struct super_block *sb)
586
{
587 588
	ext4_fsblk_t desc_count;
	struct ext4_group_desc *gdp;
589
	ext4_group_t i;
590
	ext4_group_t ngroups = ext4_get_groups_count(sb);
591 592 593
#ifdef EXT4FS_DEBUG
	struct ext4_super_block *es;
	ext4_fsblk_t bitmap_count;
594
	unsigned int x;
595 596
	struct buffer_head *bitmap_bh = NULL;

597
	es = EXT4_SB(sb)->s_es;
598 599 600 601 602
	desc_count = 0;
	bitmap_count = 0;
	gdp = NULL;

	for (i = 0; i < ngroups; i++) {
603
		gdp = ext4_get_group_desc(sb, i, NULL);
604 605
		if (!gdp)
			continue;
606
		desc_count += ext4_free_group_clusters(sb, gdp);
607
		brelse(bitmap_bh);
608
		bitmap_bh = ext4_read_block_bitmap(sb, i);
609 610 611
		if (bitmap_bh == NULL)
			continue;

612 613
		x = ext4_count_free(bitmap_bh->b_data,
				    EXT4_BLOCKS_PER_GROUP(sb) / 8);
614
		printk(KERN_DEBUG "group %u: stored = %d, counted = %u\n",
615
			i, ext4_free_group_clusters(sb, gdp), x);
616 617 618
		bitmap_count += x;
	}
	brelse(bitmap_bh);
619 620
	printk(KERN_DEBUG "ext4_count_free_clusters: stored = %llu"
	       ", computed = %llu, %llu\n",
621
	       EXT4_B2C(EXT4_SB(sb), ext4_free_blocks_count(es)),
622
	       desc_count, bitmap_count);
623 624 625 626
	return bitmap_count;
#else
	desc_count = 0;
	for (i = 0; i < ngroups; i++) {
627
		gdp = ext4_get_group_desc(sb, i, NULL);
628 629
		if (!gdp)
			continue;
630
		desc_count += ext4_free_group_clusters(sb, gdp);
631 632 633 634 635 636
	}

	return desc_count;
#endif
}

637
static inline int test_root(ext4_group_t a, int b)
638 639 640 641 642 643 644 645
{
	int num = b;

	while (a > num)
		num *= b;
	return num == a;
}

646
static int ext4_group_sparse(ext4_group_t group)
647 648 649 650 651 652 653 654 655 656
{
	if (group <= 1)
		return 1;
	if (!(group & 1))
		return 0;
	return (test_root(group, 7) || test_root(group, 5) ||
		test_root(group, 3));
}

/**
657
 *	ext4_bg_has_super - number of blocks used by the superblock in group
658 659 660 661 662 663
 *	@sb: superblock for filesystem
 *	@group: group number to check
 *
 *	Return the number of blocks used by the superblock (primary or backup)
 *	in this group.  Currently this will be only 0 or 1.
 */
664
int ext4_bg_has_super(struct super_block *sb, ext4_group_t group)
665
{
666 667 668
	if (EXT4_HAS_RO_COMPAT_FEATURE(sb,
				EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER) &&
			!ext4_group_sparse(group))
669 670 671 672
		return 0;
	return 1;
}

673 674
static unsigned long ext4_bg_num_gdb_meta(struct super_block *sb,
					ext4_group_t group)
675
{
676
	unsigned long metagroup = group / EXT4_DESC_PER_BLOCK(sb);
677 678
	ext4_group_t first = metagroup * EXT4_DESC_PER_BLOCK(sb);
	ext4_group_t last = first + EXT4_DESC_PER_BLOCK(sb) - 1;
679 680 681 682 683 684

	if (group == first || group == first + 1 || group == last)
		return 1;
	return 0;
}

685 686
static unsigned long ext4_bg_num_gdb_nometa(struct super_block *sb,
					ext4_group_t group)
687
{
688 689 690 691 692 693 694
	if (!ext4_bg_has_super(sb, group))
		return 0;

	if (EXT4_HAS_INCOMPAT_FEATURE(sb,EXT4_FEATURE_INCOMPAT_META_BG))
		return le32_to_cpu(EXT4_SB(sb)->s_es->s_first_meta_bg);
	else
		return EXT4_SB(sb)->s_gdb_count;
695 696 697
}

/**
698
 *	ext4_bg_num_gdb - number of blocks used by the group table in group
699 700 701 702 703 704 705
 *	@sb: superblock for filesystem
 *	@group: group number to check
 *
 *	Return the number of blocks used by the group descriptor table
 *	(primary or backup) in this group.  In the future there may be a
 *	different number of descriptor blocks in each group.
 */
706
unsigned long ext4_bg_num_gdb(struct super_block *sb, ext4_group_t group)
707 708
{
	unsigned long first_meta_bg =
709 710
			le32_to_cpu(EXT4_SB(sb)->s_es->s_first_meta_bg);
	unsigned long metagroup = group / EXT4_DESC_PER_BLOCK(sb);
711

712
	if (!EXT4_HAS_INCOMPAT_FEATURE(sb,EXT4_FEATURE_INCOMPAT_META_BG) ||
713
			metagroup < first_meta_bg)
714
		return ext4_bg_num_gdb_nometa(sb, group);
715

716
	return ext4_bg_num_gdb_meta(sb,group);
717 718

}
719

720
/*
721
 * This function returns the number of file system metadata clusters at
722 723
 * the beginning of a block group, including the reserved gdt blocks.
 */
E
Eric Sandeen 已提交
724
static unsigned ext4_num_base_meta_clusters(struct super_block *sb,
725
				     ext4_group_t block_group)
726 727
{
	struct ext4_sb_info *sbi = EXT4_SB(sb);
728
	unsigned num;
729 730 731 732 733 734 735 736 737 738 739 740 741 742

	/* Check for superblock and gdt backups in this group */
	num = ext4_bg_has_super(sb, block_group);

	if (!EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_META_BG) ||
	    block_group < le32_to_cpu(sbi->s_es->s_first_meta_bg) *
			  sbi->s_desc_per_block) {
		if (num) {
			num += ext4_bg_num_gdb(sb, block_group);
			num += le16_to_cpu(sbi->s_es->s_reserved_gdt_blocks);
		}
	} else { /* For META_BG_BLOCK_GROUPS */
		num += ext4_bg_num_gdb(sb, block_group);
	}
743
	return EXT4_NUM_B2C(sbi, num);
744
}
745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792
/**
 *	ext4_inode_to_goal_block - return a hint for block allocation
 *	@inode: inode for block allocation
 *
 *	Return the ideal location to start allocating blocks for a
 *	newly created inode.
 */
ext4_fsblk_t ext4_inode_to_goal_block(struct inode *inode)
{
	struct ext4_inode_info *ei = EXT4_I(inode);
	ext4_group_t block_group;
	ext4_grpblk_t colour;
	int flex_size = ext4_flex_bg_size(EXT4_SB(inode->i_sb));
	ext4_fsblk_t bg_start;
	ext4_fsblk_t last_block;

	block_group = ei->i_block_group;
	if (flex_size >= EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME) {
		/*
		 * If there are at least EXT4_FLEX_SIZE_DIR_ALLOC_SCHEME
		 * block groups per flexgroup, reserve the first block
		 * group for directories and special files.  Regular
		 * files will start at the second block group.  This
		 * tends to speed up directory access and improves
		 * fsck times.
		 */
		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);
	last_block = ext4_blocks_count(EXT4_SB(inode->i_sb)->s_es) - 1;

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

	if (bg_start + EXT4_BLOCKS_PER_GROUP(inode->i_sb) <= last_block)
		colour = (current->pid % 16) *
			(EXT4_BLOCKS_PER_GROUP(inode->i_sb) / 16);
	else
		colour = (current->pid % 16) * ((last_block - bg_start) / 16);
	return bg_start + colour;
}