balloc.c 21.0 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 26 27
/*
 * balloc.c contains the blocks allocation and deallocation routines
 */

28 29 30 31
/*
 * Calculate the block group number and offset, given a block number
 */
void ext4_get_group_no_and_offset(struct super_block *sb, ext4_fsblk_t blocknr,
32
		ext4_group_t *blockgrpp, ext4_grpblk_t *offsetp)
33
{
D
Dave Kleikamp 已提交
34
	struct ext4_super_block *es = EXT4_SB(sb)->s_es;
35 36
	ext4_grpblk_t offset;

D
Dave Kleikamp 已提交
37
	blocknr = blocknr - le32_to_cpu(es->s_first_data_block);
A
Andrew Morton 已提交
38
	offset = do_div(blocknr, EXT4_BLOCKS_PER_GROUP(sb));
39 40 41
	if (offsetp)
		*offsetp = offset;
	if (blockgrpp)
D
Dave Kleikamp 已提交
42
		*blockgrpp = blocknr;
43 44 45

}

46 47 48 49
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
50
	ext4_get_group_no_and_offset(sb, block, &actual_group, NULL);
51 52 53 54 55 56
	if (actual_group == block_group)
		return 1;
	return 0;
}

static int ext4_group_used_meta_blocks(struct super_block *sb,
57 58
				       ext4_group_t block_group,
				       struct ext4_group_desc *gdp)
59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82
{
	ext4_fsblk_t tmp;
	struct ext4_sb_info *sbi = EXT4_SB(sb);
	/* block bitmap, inode bitmap, and inode table blocks */
	int used_blocks = sbi->s_itb_per_group + 2;

	if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_FLEX_BG)) {
		if (!ext4_block_in_group(sb, ext4_block_bitmap(sb, gdp),
					block_group))
			used_blocks--;

		if (!ext4_block_in_group(sb, ext4_inode_bitmap(sb, gdp),
					block_group))
			used_blocks--;

		tmp = ext4_inode_table(sb, gdp);
		for (; tmp < ext4_inode_table(sb, gdp) +
				sbi->s_itb_per_group; tmp++) {
			if (!ext4_block_in_group(sb, tmp, block_group))
				used_blocks -= 1;
		}
	}
	return used_blocks;
}
83

A
Andreas Dilger 已提交
84 85 86
/* Initializes an uninitialized block bitmap if given, and returns the
 * number of blocks free in the group. */
unsigned ext4_init_block_bitmap(struct super_block *sb, struct buffer_head *bh,
87
		 ext4_group_t block_group, struct ext4_group_desc *gdp)
A
Andreas Dilger 已提交
88 89
{
	int bit, bit_max;
90
	ext4_group_t ngroups = ext4_get_groups_count(sb);
A
Andreas Dilger 已提交
91 92 93 94 95 96 97 98 99
	unsigned free_blocks, group_blocks;
	struct ext4_sb_info *sbi = EXT4_SB(sb);

	if (bh) {
		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. */
		if (!ext4_group_desc_csum_verify(sbi, block_group, gdp)) {
100 101
			ext4_error(sb, "Checksum bad for group %u",
					block_group);
102 103 104
			ext4_free_blks_set(sb, gdp, 0);
			ext4_free_inodes_set(sb, gdp, 0);
			ext4_itable_unused_set(sb, gdp, 0);
A
Andreas Dilger 已提交
105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
			memset(bh->b_data, 0xff, sb->s_blocksize);
			return 0;
		}
		memset(bh->b_data, 0, sb->s_blocksize);
	}

	/* Check for superblock and gdt backups in this group */
	bit_max = 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 (bit_max) {
			bit_max += ext4_bg_num_gdb(sb, block_group);
			bit_max +=
				le16_to_cpu(sbi->s_es->s_reserved_gdt_blocks);
		}
	} else { /* For META_BG_BLOCK_GROUPS */
123
		bit_max += ext4_bg_num_gdb(sb, block_group);
A
Andreas Dilger 已提交
124 125
	}

126
	if (block_group == ngroups - 1) {
A
Andreas Dilger 已提交
127 128 129 130 131 132
		/*
		 * Even though mke2fs always initialize first and last group
		 * if some other tool enabled the EXT4_BG_BLOCK_UNINIT we need
		 * to make sure we calculate the right free blocks
		 */
		group_blocks = ext4_blocks_count(sbi->s_es) -
133
			ext4_group_first_block_no(sb, ngroups - 1);
A
Andreas Dilger 已提交
134 135 136 137 138 139 140
	} else {
		group_blocks = EXT4_BLOCKS_PER_GROUP(sb);
	}

	free_blocks = group_blocks - bit_max;

	if (bh) {
141 142
		ext4_fsblk_t start, tmp;
		int flex_bg = 0;
143

A
Andreas Dilger 已提交
144 145 146
		for (bit = 0; bit < bit_max; bit++)
			ext4_set_bit(bit, bh->b_data);

147
		start = ext4_group_first_block_no(sb, block_group);
A
Andreas Dilger 已提交
148

149 150 151
		if (EXT4_HAS_INCOMPAT_FEATURE(sb,
					      EXT4_FEATURE_INCOMPAT_FLEX_BG))
			flex_bg = 1;
A
Andreas Dilger 已提交
152

153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168
		/* 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))
			ext4_set_bit(tmp - start, bh->b_data);

		tmp = ext4_inode_bitmap(sb, gdp);
		if (!flex_bg || ext4_block_in_group(sb, tmp, block_group))
			ext4_set_bit(tmp - start, bh->b_data);

		tmp = ext4_inode_table(sb, gdp);
		for (; tmp < ext4_inode_table(sb, gdp) +
				sbi->s_itb_per_group; tmp++) {
			if (!flex_bg ||
				ext4_block_in_group(sb, tmp, block_group))
				ext4_set_bit(tmp - start, bh->b_data);
		}
A
Andreas Dilger 已提交
169 170 171 172 173 174 175
		/*
		 * 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
		 */
		mark_bitmap_end(group_blocks, sb->s_blocksize * 8, bh->b_data);
	}
176
	return free_blocks - ext4_group_used_meta_blocks(sb, block_group, gdp);
A
Andreas Dilger 已提交
177 178 179
}


180 181 182 183 184 185 186 187
/*
 * 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
188
 * when a file system is mounted (see ext4_fill_super).
189 190 191
 */

/**
192
 * ext4_get_group_desc() -- load group descriptor from disk
193 194 195 196 197
 * @sb:			super block
 * @block_group:	given block group
 * @bh:			pointer to the buffer head to store the block
 *			group descriptor
 */
198
struct ext4_group_desc * ext4_get_group_desc(struct super_block *sb,
199
					     ext4_group_t block_group,
200
					     struct buffer_head **bh)
201
{
202 203
	unsigned int group_desc;
	unsigned int offset;
204
	ext4_group_t ngroups = ext4_get_groups_count(sb);
205
	struct ext4_group_desc *desc;
206
	struct ext4_sb_info *sbi = EXT4_SB(sb);
207

208
	if (block_group >= ngroups) {
209 210
		ext4_error(sb, "block_group >= groups_count - block_group = %u,"
			   " groups_count = %u", block_group, ngroups);
211 212 213 214

		return NULL;
	}

215 216
	group_desc = block_group >> EXT4_DESC_PER_BLOCK_BITS(sb);
	offset = block_group & (EXT4_DESC_PER_BLOCK(sb) - 1);
217
	if (!sbi->s_group_desc[group_desc]) {
218
		ext4_error(sb, "Group descriptor not loaded - "
219
			   "block_group = %u, group_desc = %u, desc = %u",
220
			   block_group, group_desc, offset);
221 222 223
		return NULL;
	}

224 225 226
	desc = (struct ext4_group_desc *)(
		(__u8 *)sbi->s_group_desc[group_desc]->b_data +
		offset * EXT4_DESC_SIZE(sb));
227 228
	if (bh)
		*bh = sbi->s_group_desc[group_desc];
229
	return desc;
230 231
}

232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277
static int ext4_valid_block_bitmap(struct super_block *sb,
					struct ext4_group_desc *desc,
					unsigned int block_group,
					struct buffer_head *bh)
{
	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:
278
	ext4_error(sb, "Invalid block bitmap - block_group = %d, block = %llu",
279 280 281
			block_group, bitmap_blk);
	return 0;
}
282
/**
283
 * ext4_read_block_bitmap()
284 285 286
 * @sb:			super block
 * @block_group:	given block group
 *
287 288
 * Read the bitmap for a given block_group,and validate the
 * bits for block/inode/inode tables are set in the bitmaps
289 290 291
 *
 * Return buffer_head on success or NULL in case of failure.
 */
A
Andreas Dilger 已提交
292
struct buffer_head *
293
ext4_read_block_bitmap(struct super_block *sb, ext4_group_t block_group)
294
{
295 296
	struct ext4_group_desc *desc;
	struct buffer_head *bh = NULL;
297
	ext4_fsblk_t bitmap_blk;
298

A
Andreas Dilger 已提交
299
	desc = ext4_get_group_desc(sb, block_group, NULL);
300
	if (!desc)
301 302
		return NULL;
	bitmap_blk = ext4_block_bitmap(sb, desc);
303 304
	bh = sb_getblk(sb, bitmap_blk);
	if (unlikely(!bh)) {
305
		ext4_error(sb, "Cannot read block bitmap - "
306
			    "block_group = %u, block_bitmap = %llu",
307
			    block_group, bitmap_blk);
308 309
		return NULL;
	}
310 311

	if (bitmap_uptodate(bh))
312 313
		return bh;

314
	lock_buffer(bh);
315 316 317 318
	if (bitmap_uptodate(bh)) {
		unlock_buffer(bh);
		return bh;
	}
319
	ext4_lock_group(sb, block_group);
A
Andreas Dilger 已提交
320
	if (desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) {
321
		ext4_init_block_bitmap(sb, bh, block_group, desc);
322
		set_bitmap_uptodate(bh);
323
		set_buffer_uptodate(bh);
324
		ext4_unlock_group(sb, block_group);
A
Aneesh Kumar K.V 已提交
325
		unlock_buffer(bh);
326
		return bh;
A
Andreas Dilger 已提交
327
	}
328
	ext4_unlock_group(sb, block_group);
329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344
	if (buffer_uptodate(bh)) {
		/*
		 * if not uninit if bh is uptodate,
		 * bitmap is also uptodate
		 */
		set_bitmap_uptodate(bh);
		unlock_buffer(bh);
		return bh;
	}
	/*
	 * submit the buffer_head for read. We can
	 * safely mark the bitmap as uptodate now.
	 * We do it here so the bitmap uptodate bit
	 * get set with buffer lock held.
	 */
	set_bitmap_uptodate(bh);
345 346
	if (bh_submit_read(bh) < 0) {
		put_bh(bh);
347
		ext4_error(sb, "Cannot read block bitmap - "
348
			    "block_group = %u, block_bitmap = %llu",
349
			    block_group, bitmap_blk);
350 351
		return NULL;
	}
352 353 354 355 356
	ext4_valid_block_bitmap(sb, desc, block_group, bh);
	/*
	 * file system mounted not to panic on error,
	 * continue with corrupt bitmap
	 */
357 358 359 360
	return bh;
}

/**
361
 * ext4_add_groupblocks() -- Add given blocks to an existing group
362 363
 * @handle:			handle to this transaction
 * @sb:				super block
364
 * @block:			start physcial block to add to the block group
365
 * @count:			number of blocks to free
366
 *
367 368 369
 * This marks the blocks as free in the bitmap. We ask the
 * mballoc to reload the buddy after this by setting group
 * EXT4_GROUP_INFO_NEED_INIT_BIT flag
370
 */
371 372
void ext4_add_groupblocks(handle_t *handle, struct super_block *sb,
			 ext4_fsblk_t block, unsigned long count)
373 374 375
{
	struct buffer_head *bitmap_bh = NULL;
	struct buffer_head *gd_bh;
376
	ext4_group_t block_group;
377
	ext4_grpblk_t bit;
378
	unsigned int i;
379
	struct ext4_group_desc *desc;
T
Theodore Ts'o 已提交
380
	struct ext4_sb_info *sbi = EXT4_SB(sb);
381
	int err = 0, ret, blk_free_count;
382 383
	ext4_grpblk_t blocks_freed;
	struct ext4_group_info *grp;
384

385
	ext4_debug("Adding block(s) %llu-%llu\n", block, block + count - 1);
386

387
	ext4_get_group_no_and_offset(sb, block, &block_group, &bit);
388
	grp = ext4_get_group_info(sb, block_group);
389 390 391 392
	/*
	 * Check to see if we are freeing blocks across a group
	 * boundary.
	 */
393
	if (bit + count > EXT4_BLOCKS_PER_GROUP(sb)) {
394
		goto error_return;
395
	}
396
	bitmap_bh = ext4_read_block_bitmap(sb, block_group);
397 398
	if (!bitmap_bh)
		goto error_return;
399
	desc = ext4_get_group_desc(sb, block_group, &gd_bh);
400 401 402
	if (!desc)
		goto error_return;

403 404 405 406
	if (in_range(ext4_block_bitmap(sb, desc), block, count) ||
	    in_range(ext4_inode_bitmap(sb, desc), block, count) ||
	    in_range(block, ext4_inode_table(sb, desc), sbi->s_itb_per_group) ||
	    in_range(block + count - 1, ext4_inode_table(sb, desc),
407
		     sbi->s_itb_per_group)) {
408
		ext4_error(sb, "Adding blocks in system zones - "
409 410
			   "Block = %llu, count = %lu",
			   block, count);
411 412
		goto error_return;
	}
413 414

	/*
415
	 * We are about to add blocks to the bitmap,
416 417 418
	 * so we need undo access.
	 */
	BUFFER_TRACE(bitmap_bh, "getting undo access");
419
	err = ext4_journal_get_undo_access(handle, bitmap_bh);
420 421 422 423 424 425 426 427 428
	if (err)
		goto error_return;

	/*
	 * We are about to modify some metadata.  Call the journal APIs
	 * to unshare ->b_data if a currently-committing transaction is
	 * using it
	 */
	BUFFER_TRACE(gd_bh, "get_write_access");
429
	err = ext4_journal_get_write_access(handle, gd_bh);
430 431
	if (err)
		goto error_return;
432 433 434 435 436
	/*
	 * make sure we don't allow a parallel init on other groups in the
	 * same buddy cache
	 */
	down_write(&grp->alloc_sem);
437
	for (i = 0, blocks_freed = 0; i < count; i++) {
438
		BUFFER_TRACE(bitmap_bh, "clear bit");
439
		if (!ext4_clear_bit_atomic(ext4_group_lock_ptr(sb, block_group),
440
						bit + i, bitmap_bh->b_data)) {
441
			ext4_error(sb, "bit already cleared for block %llu",
L
Laurent Vivier 已提交
442
				   (ext4_fsblk_t)(block + i));
443 444
			BUFFER_TRACE(bitmap_bh, "bit already cleared");
		} else {
445
			blocks_freed++;
446 447
		}
	}
448
	ext4_lock_group(sb, block_group);
449 450
	blk_free_count = blocks_freed + ext4_free_blks_count(sb, desc);
	ext4_free_blks_set(sb, desc, blk_free_count);
A
Andreas Dilger 已提交
451
	desc->bg_checksum = ext4_group_desc_csum(sbi, block_group, desc);
452
	ext4_unlock_group(sb, block_group);
453
	percpu_counter_add(&sbi->s_freeblocks_counter, blocks_freed);
454

455 456
	if (sbi->s_log_groups_per_flex) {
		ext4_group_t flex_group = ext4_flex_group(sbi, block_group);
457 458
		atomic_add(blocks_freed,
			   &sbi->s_flex_groups[flex_group].free_blocks);
459
	}
460 461 462 463 464
	/*
	 * request to reload the buddy with the
	 * new bitmap information
	 */
	set_bit(EXT4_GROUP_INFO_NEED_INIT_BIT, &(grp->bb_state));
465
	grp->bb_free += blocks_freed;
466
	up_write(&grp->alloc_sem);
467

468 469
	/* We dirtied the bitmap block */
	BUFFER_TRACE(bitmap_bh, "dirtied bitmap block");
470
	err = ext4_handle_dirty_metadata(handle, NULL, bitmap_bh);
471 472 473

	/* And the group descriptor block */
	BUFFER_TRACE(gd_bh, "dirtied group descriptor block");
474
	ret = ext4_handle_dirty_metadata(handle, NULL, gd_bh);
475 476 477
	if (!err)
		err = ret;

478 479
error_return:
	brelse(bitmap_bh);
480
	ext4_std_error(sb, err);
481 482 483
	return;
}

484 485 486 487 488 489 490 491
/**
 * ext4_has_free_blocks()
 * @sbi:	in-core super block structure.
 * @nblocks:	number of needed blocks
 *
 * Check if filesystem has nblocks free & available for allocation.
 * On success return 1, return 0 on failure.
 */
492
static int ext4_has_free_blocks(struct ext4_sb_info *sbi, s64 nblocks)
493
{
494
	s64 free_blocks, dirty_blocks, root_blocks;
495
	struct percpu_counter *fbc = &sbi->s_freeblocks_counter;
496
	struct percpu_counter *dbc = &sbi->s_dirtyblocks_counter;
497

498 499
	free_blocks  = percpu_counter_read_positive(fbc);
	dirty_blocks = percpu_counter_read_positive(dbc);
500
	root_blocks = ext4_r_blocks_count(sbi->s_es);
501

502 503
	if (free_blocks - (nblocks + root_blocks + dirty_blocks) <
						EXT4_FREEBLOCKS_WATERMARK) {
504 505
		free_blocks  = percpu_counter_sum_positive(fbc);
		dirty_blocks = percpu_counter_sum_positive(dbc);
506 507 508
		if (dirty_blocks < 0) {
			printk(KERN_CRIT "Dirty block accounting "
					"went wrong %lld\n",
509
					(long long)dirty_blocks);
510 511 512
		}
	}
	/* Check whether we have space after
513
	 * accounting for current dirty blocks & root reserved blocks.
514
	 */
515 516
	if (free_blocks >= ((root_blocks + nblocks) + dirty_blocks))
		return 1;
517

518
	/* Hm, nope.  Are (enough) root reserved blocks available? */
519
	if (sbi->s_resuid == current_fsuid() ||
520 521 522 523 524 525 526
	    ((sbi->s_resgid != 0) && in_group_p(sbi->s_resgid)) ||
	    capable(CAP_SYS_RESOURCE)) {
		if (free_blocks >= (nblocks + dirty_blocks))
			return 1;
	}

	return 0;
527 528
}

529
int ext4_claim_free_blocks(struct ext4_sb_info *sbi,
A
Aneesh Kumar K.V 已提交
530
						s64 nblocks)
531
{
532 533
	if (ext4_has_free_blocks(sbi, nblocks)) {
		percpu_counter_add(&sbi->s_dirtyblocks_counter, nblocks);
534
		return 0;
535 536
	} else
		return -ENOSPC;
537
}
538

539
/**
540
 * ext4_should_retry_alloc()
541 542 543
 * @sb:			super block
 * @retries		number of attemps has been made
 *
544
 * ext4_should_retry_alloc() is called when ENOSPC is returned, and if
545 546 547 548 549 550
 * it is profitable to retry the operation, this function will wait
 * for the current or commiting transaction to complete, and then
 * return TRUE.
 *
 * if the total number of retries exceed three times, return FALSE.
 */
551
int ext4_should_retry_alloc(struct super_block *sb, int *retries)
552
{
553 554 555
	if (!ext4_has_free_blocks(EXT4_SB(sb), 1) ||
	    (*retries)++ > 3 ||
	    !EXT4_SB(sb)->s_journal)
556 557 558 559
		return 0;

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

560
	return jbd2_journal_force_commit_nested(EXT4_SB(sb)->s_journal);
561 562
}

A
Aneesh Kumar K.V 已提交
563
/*
564
 * ext4_new_meta_blocks() -- allocate block for meta data (indexing) blocks
A
Aneesh Kumar K.V 已提交
565 566 567 568
 *
 * @handle:             handle to this transaction
 * @inode:              file inode
 * @goal:               given target block(filesystem wide)
T
Theodore Ts'o 已提交
569
 * @count:		pointer to total number of blocks needed
A
Aneesh Kumar K.V 已提交
570 571
 * @errp:               error code
 *
T
Theodore Ts'o 已提交
572
 * Return 1st allocated block number on success, *count stores total account
573
 * error stores in errp pointer
A
Aneesh Kumar K.V 已提交
574
 */
575 576
ext4_fsblk_t ext4_new_meta_blocks(handle_t *handle, struct inode *inode,
		ext4_fsblk_t goal, unsigned long *count, int *errp)
A
Aneesh Kumar K.V 已提交
577
{
T
Theodore Ts'o 已提交
578
	struct ext4_allocation_request ar;
579
	ext4_fsblk_t ret;
T
Theodore Ts'o 已提交
580 581 582 583 584 585 586 587 588 589

	memset(&ar, 0, sizeof(ar));
	/* Fill with neighbour allocated blocks */
	ar.inode = inode;
	ar.goal = goal;
	ar.len = count ? *count : 1;

	ret = ext4_mb_new_blocks(handle, &ar, errp);
	if (count)
		*count = ar.len;
590
	/*
591 592
	 * Account for the allocated meta blocks.  We will never
	 * fail EDQUOT for metdata, but we do account for it.
593
	 */
594
	if (!(*errp) && EXT4_I(inode)->i_delalloc_reserved_flag) {
595
		spin_lock(&EXT4_I(inode)->i_block_reservation_lock);
T
Theodore Ts'o 已提交
596
		EXT4_I(inode)->i_allocated_meta_blocks += ar.len;
597
		spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
598
		dquot_alloc_block_nofail(inode, ar.len);
599 600
	}
	return ret;
A
Aneesh Kumar K.V 已提交
601 602
}

603
/**
604
 * ext4_count_free_blocks() -- count filesystem free blocks
605 606 607 608
 * @sb:		superblock
 *
 * Adds up the number of free blocks from each block group.
 */
609
ext4_fsblk_t ext4_count_free_blocks(struct super_block *sb)
610
{
611 612
	ext4_fsblk_t desc_count;
	struct ext4_group_desc *gdp;
613
	ext4_group_t i;
614
	ext4_group_t ngroups = ext4_get_groups_count(sb);
615 616 617
#ifdef EXT4FS_DEBUG
	struct ext4_super_block *es;
	ext4_fsblk_t bitmap_count;
618
	unsigned int x;
619 620
	struct buffer_head *bitmap_bh = NULL;

621
	es = EXT4_SB(sb)->s_es;
622 623 624 625 626
	desc_count = 0;
	bitmap_count = 0;
	gdp = NULL;

	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_blks_count(sb, gdp);
631
		brelse(bitmap_bh);
632
		bitmap_bh = ext4_read_block_bitmap(sb, i);
633 634 635
		if (bitmap_bh == NULL)
			continue;

636
		x = ext4_count_free(bitmap_bh, sb->s_blocksize);
637 638
		printk(KERN_DEBUG "group %u: stored = %d, counted = %u\n",
			i, ext4_free_blks_count(sb, gdp), x);
639 640 641
		bitmap_count += x;
	}
	brelse(bitmap_bh);
642 643 644
	printk(KERN_DEBUG "ext4_count_free_blocks: stored = %llu"
		", computed = %llu, %llu\n", ext4_free_blocks_count(es),
	       desc_count, bitmap_count);
645 646 647 648
	return bitmap_count;
#else
	desc_count = 0;
	for (i = 0; i < ngroups; i++) {
649
		gdp = ext4_get_group_desc(sb, i, NULL);
650 651
		if (!gdp)
			continue;
652
		desc_count += ext4_free_blks_count(sb, gdp);
653 654 655 656 657 658
	}

	return desc_count;
#endif
}

659
static inline int test_root(ext4_group_t a, int b)
660 661 662 663 664 665 666 667
{
	int num = b;

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

668
static int ext4_group_sparse(ext4_group_t group)
669 670 671 672 673 674 675 676 677 678
{
	if (group <= 1)
		return 1;
	if (!(group & 1))
		return 0;
	return (test_root(group, 7) || test_root(group, 5) ||
		test_root(group, 3));
}

/**
679
 *	ext4_bg_has_super - number of blocks used by the superblock in group
680 681 682 683 684 685
 *	@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.
 */
686
int ext4_bg_has_super(struct super_block *sb, ext4_group_t group)
687
{
688 689 690
	if (EXT4_HAS_RO_COMPAT_FEATURE(sb,
				EXT4_FEATURE_RO_COMPAT_SPARSE_SUPER) &&
			!ext4_group_sparse(group))
691 692 693 694
		return 0;
	return 1;
}

695 696
static unsigned long ext4_bg_num_gdb_meta(struct super_block *sb,
					ext4_group_t group)
697
{
698
	unsigned long metagroup = group / EXT4_DESC_PER_BLOCK(sb);
699 700
	ext4_group_t first = metagroup * EXT4_DESC_PER_BLOCK(sb);
	ext4_group_t last = first + EXT4_DESC_PER_BLOCK(sb) - 1;
701 702 703 704 705 706

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

707 708
static unsigned long ext4_bg_num_gdb_nometa(struct super_block *sb,
					ext4_group_t group)
709
{
710 711 712 713 714 715 716
	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;
717 718 719
}

/**
720
 *	ext4_bg_num_gdb - number of blocks used by the group table in group
721 722 723 724 725 726 727
 *	@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.
 */
728
unsigned long ext4_bg_num_gdb(struct super_block *sb, ext4_group_t group)
729 730
{
	unsigned long first_meta_bg =
731 732
			le32_to_cpu(EXT4_SB(sb)->s_es->s_first_meta_bg);
	unsigned long metagroup = group / EXT4_DESC_PER_BLOCK(sb);
733

734
	if (!EXT4_HAS_INCOMPAT_FEATURE(sb,EXT4_FEATURE_INCOMPAT_META_BG) ||
735
			metagroup < first_meta_bg)
736
		return ext4_bg_num_gdb_nometa(sb, group);
737

738
	return ext4_bg_num_gdb_meta(sb,group);
739 740

}
741