disk-io.c 122.1 KB
Newer Older
C
Chris Mason 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
/*
 * Copyright (C) 2007 Oracle.  All rights reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License v2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

C
Chris Mason 已提交
19
#include <linux/fs.h>
20
#include <linux/blkdev.h>
C
Chris Mason 已提交
21
#include <linux/scatterlist.h>
C
Chris Mason 已提交
22
#include <linux/swap.h>
23
#include <linux/radix-tree.h>
C
Chris Mason 已提交
24
#include <linux/writeback.h>
C
Chris Mason 已提交
25
#include <linux/buffer_head.h>
26
#include <linux/workqueue.h>
27
#include <linux/kthread.h>
28
#include <linux/slab.h>
29
#include <linux/migrate.h>
30
#include <linux/ratelimit.h>
31
#include <linux/uuid.h>
S
Stefan Behrens 已提交
32
#include <linux/semaphore.h>
33
#include <asm/unaligned.h>
34 35
#include "ctree.h"
#include "disk-io.h"
36
#include "hash.h"
37
#include "transaction.h"
38
#include "btrfs_inode.h"
39
#include "volumes.h"
40
#include "print-tree.h"
41
#include "locking.h"
42
#include "tree-log.h"
43
#include "free-space-cache.h"
44
#include "free-space-tree.h"
45
#include "inode-map.h"
46
#include "check-integrity.h"
47
#include "rcu-string.h"
48
#include "dev-replace.h"
D
David Woodhouse 已提交
49
#include "raid56.h"
50
#include "sysfs.h"
J
Josef Bacik 已提交
51
#include "qgroup.h"
52
#include "compression.h"
53

54 55 56 57
#ifdef CONFIG_X86
#include <asm/cpufeature.h>
#endif

58 59 60 61 62 63
#define BTRFS_SUPER_FLAG_SUPP	(BTRFS_HEADER_FLAG_WRITTEN |\
				 BTRFS_HEADER_FLAG_RELOC |\
				 BTRFS_SUPER_FLAG_ERROR |\
				 BTRFS_SUPER_FLAG_SEEDING |\
				 BTRFS_SUPER_FLAG_METADUMP)

64
static const struct extent_io_ops btree_extent_io_ops;
65
static void end_workqueue_fn(struct btrfs_work *work);
66
static void free_fs_root(struct btrfs_root *root);
67
static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info,
L
liubo 已提交
68
				    int read_only);
69
static void btrfs_destroy_ordered_extents(struct btrfs_root *root);
L
liubo 已提交
70 71
static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
				      struct btrfs_root *root);
72
static void btrfs_destroy_delalloc_inodes(struct btrfs_root *root);
L
liubo 已提交
73 74 75 76 77
static int btrfs_destroy_marked_extents(struct btrfs_root *root,
					struct extent_io_tree *dirty_pages,
					int mark);
static int btrfs_destroy_pinned_extent(struct btrfs_root *root,
				       struct extent_io_tree *pinned_extents);
78 79
static int btrfs_cleanup_transaction(struct btrfs_root *root);
static void btrfs_error_commit_super(struct btrfs_root *root);
80

C
Chris Mason 已提交
81
/*
82 83
 * btrfs_end_io_wq structs are used to do processing in task context when an IO
 * is complete.  This is used during reads to verify checksums, and it is used
C
Chris Mason 已提交
84 85
 * by writes to insert metadata for new file extents after IO is complete.
 */
86
struct btrfs_end_io_wq {
87 88 89 90 91
	struct bio *bio;
	bio_end_io_t *end_io;
	void *private;
	struct btrfs_fs_info *info;
	int error;
92
	enum btrfs_wq_endio_type metadata;
93
	struct list_head list;
94
	struct btrfs_work work;
95
};
96

97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
static struct kmem_cache *btrfs_end_io_wq_cache;

int __init btrfs_end_io_wq_init(void)
{
	btrfs_end_io_wq_cache = kmem_cache_create("btrfs_end_io_wq",
					sizeof(struct btrfs_end_io_wq),
					0,
					SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD,
					NULL);
	if (!btrfs_end_io_wq_cache)
		return -ENOMEM;
	return 0;
}

void btrfs_end_io_wq_exit(void)
{
113
	kmem_cache_destroy(btrfs_end_io_wq_cache);
114 115
}

C
Chris Mason 已提交
116 117 118 119 120
/*
 * async submit bios are used to offload expensive checksumming
 * onto the worker threads.  They checksum file and metadata bios
 * just before they are sent down the IO stack.
 */
121 122 123 124
struct async_submit_bio {
	struct inode *inode;
	struct bio *bio;
	struct list_head list;
C
Chris Mason 已提交
125 126
	extent_submit_bio_hook_t *submit_bio_start;
	extent_submit_bio_hook_t *submit_bio_done;
127 128
	int rw;
	int mirror_num;
C
Chris Mason 已提交
129
	unsigned long bio_flags;
130 131 132 133 134
	/*
	 * bio_offset is optional, can be used if the pages in the bio
	 * can't tell us where in the file the bio should go
	 */
	u64 bio_offset;
135
	struct btrfs_work work;
136
	int error;
137 138
};

139 140 141 142 143 144 145 146 147 148
/*
 * Lockdep class keys for extent_buffer->lock's in this root.  For a given
 * eb, the lockdep key is determined by the btrfs_root it belongs to and
 * the level the eb occupies in the tree.
 *
 * Different roots are used for different purposes and may nest inside each
 * other and they require separate keysets.  As lockdep keys should be
 * static, assign keysets according to the purpose of the root as indicated
 * by btrfs_root->objectid.  This ensures that all special purpose roots
 * have separate keysets.
149
 *
150 151 152
 * Lock-nesting across peer nodes is always done with the immediate parent
 * node locked thus preventing deadlock.  As lockdep doesn't know this, use
 * subclass to avoid triggering lockdep warning in such cases.
153
 *
154 155 156
 * The key is set by the readpage_end_io_hook after the buffer has passed
 * csum validation but before the pages are unlocked.  It is also set by
 * btrfs_init_new_buffer on freshly allocated blocks.
157
 *
158 159 160
 * We also add a check to make sure the highest level of the tree is the
 * same as our lockdep setup here.  If BTRFS_MAX_LEVEL changes, this code
 * needs update as well.
161 162 163 164 165
 */
#ifdef CONFIG_DEBUG_LOCK_ALLOC
# if BTRFS_MAX_LEVEL != 8
#  error
# endif
166 167 168 169 170 171 172 173 174 175 176 177 178

static struct btrfs_lockdep_keyset {
	u64			id;		/* root objectid */
	const char		*name_stem;	/* lock name stem */
	char			names[BTRFS_MAX_LEVEL + 1][20];
	struct lock_class_key	keys[BTRFS_MAX_LEVEL + 1];
} btrfs_lockdep_keysets[] = {
	{ .id = BTRFS_ROOT_TREE_OBJECTID,	.name_stem = "root"	},
	{ .id = BTRFS_EXTENT_TREE_OBJECTID,	.name_stem = "extent"	},
	{ .id = BTRFS_CHUNK_TREE_OBJECTID,	.name_stem = "chunk"	},
	{ .id = BTRFS_DEV_TREE_OBJECTID,	.name_stem = "dev"	},
	{ .id = BTRFS_FS_TREE_OBJECTID,		.name_stem = "fs"	},
	{ .id = BTRFS_CSUM_TREE_OBJECTID,	.name_stem = "csum"	},
179
	{ .id = BTRFS_QUOTA_TREE_OBJECTID,	.name_stem = "quota"	},
180 181 182
	{ .id = BTRFS_TREE_LOG_OBJECTID,	.name_stem = "log"	},
	{ .id = BTRFS_TREE_RELOC_OBJECTID,	.name_stem = "treloc"	},
	{ .id = BTRFS_DATA_RELOC_TREE_OBJECTID,	.name_stem = "dreloc"	},
183
	{ .id = BTRFS_UUID_TREE_OBJECTID,	.name_stem = "uuid"	},
184
	{ .id = BTRFS_FREE_SPACE_TREE_OBJECTID,	.name_stem = "free-space" },
185
	{ .id = 0,				.name_stem = "tree"	},
186
};
187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217

void __init btrfs_init_lockdep(void)
{
	int i, j;

	/* initialize lockdep class names */
	for (i = 0; i < ARRAY_SIZE(btrfs_lockdep_keysets); i++) {
		struct btrfs_lockdep_keyset *ks = &btrfs_lockdep_keysets[i];

		for (j = 0; j < ARRAY_SIZE(ks->names); j++)
			snprintf(ks->names[j], sizeof(ks->names[j]),
				 "btrfs-%s-%02d", ks->name_stem, j);
	}
}

void btrfs_set_buffer_lockdep_class(u64 objectid, struct extent_buffer *eb,
				    int level)
{
	struct btrfs_lockdep_keyset *ks;

	BUG_ON(level >= ARRAY_SIZE(ks->keys));

	/* find the matching keyset, id 0 is the default entry */
	for (ks = btrfs_lockdep_keysets; ks->id; ks++)
		if (ks->id == objectid)
			break;

	lockdep_set_class_and_name(&eb->lock,
				   &ks->keys[level], ks->names[level]);
}

218 219
#endif

C
Chris Mason 已提交
220 221 222 223
/*
 * extents on the btree inode are pretty simple, there's one extent
 * that covers the entire device
 */
224
static struct extent_map *btree_get_extent(struct inode *inode,
225
		struct page *page, size_t pg_offset, u64 start, u64 len,
226
		int create)
227
{
228 229 230 231
	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
	struct extent_map *em;
	int ret;

232
	read_lock(&em_tree->lock);
233
	em = lookup_extent_mapping(em_tree, start, len);
234 235 236
	if (em) {
		em->bdev =
			BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev;
237
		read_unlock(&em_tree->lock);
238
		goto out;
239
	}
240
	read_unlock(&em_tree->lock);
241

242
	em = alloc_extent_map();
243 244 245 246 247
	if (!em) {
		em = ERR_PTR(-ENOMEM);
		goto out;
	}
	em->start = 0;
248
	em->len = (u64)-1;
C
Chris Mason 已提交
249
	em->block_len = (u64)-1;
250
	em->block_start = 0;
251
	em->bdev = BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev;
252

253
	write_lock(&em_tree->lock);
J
Josef Bacik 已提交
254
	ret = add_extent_mapping(em_tree, em, 0);
255 256
	if (ret == -EEXIST) {
		free_extent_map(em);
257
		em = lookup_extent_mapping(em_tree, start, len);
258
		if (!em)
259
			em = ERR_PTR(-EIO);
260
	} else if (ret) {
261
		free_extent_map(em);
262
		em = ERR_PTR(ret);
263
	}
264
	write_unlock(&em_tree->lock);
265

266 267
out:
	return em;
268 269
}

270
u32 btrfs_csum_data(char *data, u32 seed, size_t len)
271
{
272
	return btrfs_crc32c(seed, data, len);
273 274 275 276
}

void btrfs_csum_final(u32 crc, char *result)
{
277
	put_unaligned_le32(~crc, result);
278 279
}

C
Chris Mason 已提交
280 281 282 283
/*
 * compute the csum for a btree block, and either verify it or write it
 * into the csum field of the block.
 */
284 285
static int csum_tree_block(struct btrfs_fs_info *fs_info,
			   struct extent_buffer *buf,
286 287
			   int verify)
{
288
	u16 csum_size = btrfs_super_csum_size(fs_info->super_copy);
289
	char *result = NULL;
290 291 292 293 294 295 296 297
	unsigned long len;
	unsigned long cur_len;
	unsigned long offset = BTRFS_CSUM_SIZE;
	char *kaddr;
	unsigned long map_start;
	unsigned long map_len;
	int err;
	u32 crc = ~(u32)0;
298
	unsigned long inline_result;
299 300

	len = buf->len - offset;
C
Chris Mason 已提交
301
	while (len > 0) {
302
		err = map_private_extent_buffer(buf, offset, 32,
303
					&kaddr, &map_start, &map_len);
C
Chris Mason 已提交
304
		if (err)
305
			return err;
306
		cur_len = min(len, map_len - (offset - map_start));
307
		crc = btrfs_csum_data(kaddr + offset - map_start,
308 309 310 311
				      crc, cur_len);
		len -= cur_len;
		offset += cur_len;
	}
312
	if (csum_size > sizeof(inline_result)) {
313
		result = kzalloc(csum_size, GFP_NOFS);
314
		if (!result)
315
			return -ENOMEM;
316 317 318 319
	} else {
		result = (char *)&inline_result;
	}

320 321 322
	btrfs_csum_final(crc, result);

	if (verify) {
323
		if (memcmp_extent_buffer(buf, result, 0, csum_size)) {
324 325
			u32 val;
			u32 found = 0;
326
			memcpy(&found, result, csum_size);
327

328
			read_extent_buffer(buf, &val, 0, csum_size);
329 330 331
			btrfs_warn_rl(fs_info,
				"%s checksum verify failed on %llu wanted %X found %X "
				"level %d",
332
				fs_info->sb->s_id, buf->start,
333
				val, found, btrfs_header_level(buf));
334 335
			if (result != (char *)&inline_result)
				kfree(result);
336
			return -EUCLEAN;
337 338
		}
	} else {
339
		write_extent_buffer(buf, result, 0, csum_size);
340
	}
341 342
	if (result != (char *)&inline_result)
		kfree(result);
343 344 345
	return 0;
}

C
Chris Mason 已提交
346 347 348 349 350 351
/*
 * we can't consider a given block up to date unless the transid of the
 * block matches the transid in the parent node's pointer.  This is how we
 * detect blocks that either didn't get written at all or got written
 * in the wrong place.
 */
352
static int verify_parent_transid(struct extent_io_tree *io_tree,
353 354
				 struct extent_buffer *eb, u64 parent_transid,
				 int atomic)
355
{
356
	struct extent_state *cached_state = NULL;
357
	int ret;
358
	bool need_lock = (current->journal_info == BTRFS_SEND_TRANS_STUB);
359 360 361 362

	if (!parent_transid || btrfs_header_generation(eb) == parent_transid)
		return 0;

363 364 365
	if (atomic)
		return -EAGAIN;

366 367 368 369 370
	if (need_lock) {
		btrfs_tree_read_lock(eb);
		btrfs_set_lock_blocking_rw(eb, BTRFS_READ_LOCK);
	}

371
	lock_extent_bits(io_tree, eb->start, eb->start + eb->len - 1,
372
			 &cached_state);
373
	if (extent_buffer_uptodate(eb) &&
374 375 376 377
	    btrfs_header_generation(eb) == parent_transid) {
		ret = 0;
		goto out;
	}
378 379 380
	btrfs_err_rl(eb->fs_info,
		"parent transid verify failed on %llu wanted %llu found %llu",
			eb->start,
381
			parent_transid, btrfs_header_generation(eb));
382
	ret = 1;
383 384 385 386 387 388 389 390 391 392 393

	/*
	 * Things reading via commit roots that don't have normal protection,
	 * like send, can have a really old block in cache that may point at a
	 * block that has been free'd and re-allocated.  So don't clear uptodate
	 * if we find an eb that is under IO (dirty/writeback) because we could
	 * end up reading in the stale data and then writing it back out and
	 * making everybody very sad.
	 */
	if (!extent_buffer_under_io(eb))
		clear_extent_buffer_uptodate(eb);
C
Chris Mason 已提交
394
out:
395 396
	unlock_extent_cached(io_tree, eb->start, eb->start + eb->len - 1,
			     &cached_state, GFP_NOFS);
397 398
	if (need_lock)
		btrfs_tree_read_unlock_blocking(eb);
399 400 401
	return ret;
}

D
David Sterba 已提交
402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431
/*
 * Return 0 if the superblock checksum type matches the checksum value of that
 * algorithm. Pass the raw disk superblock data.
 */
static int btrfs_check_super_csum(char *raw_disk_sb)
{
	struct btrfs_super_block *disk_sb =
		(struct btrfs_super_block *)raw_disk_sb;
	u16 csum_type = btrfs_super_csum_type(disk_sb);
	int ret = 0;

	if (csum_type == BTRFS_CSUM_TYPE_CRC32) {
		u32 crc = ~(u32)0;
		const int csum_size = sizeof(crc);
		char result[csum_size];

		/*
		 * The super_block structure does not span the whole
		 * BTRFS_SUPER_INFO_SIZE range, we expect that the unused space
		 * is filled with zeros and is included in the checkum.
		 */
		crc = btrfs_csum_data(raw_disk_sb + BTRFS_CSUM_SIZE,
				crc, BTRFS_SUPER_INFO_SIZE - BTRFS_CSUM_SIZE);
		btrfs_csum_final(crc, result);

		if (memcmp(raw_disk_sb, result, csum_size))
			ret = 1;
	}

	if (csum_type >= ARRAY_SIZE(btrfs_csum_sizes)) {
432
		printk(KERN_ERR "BTRFS: unsupported checksum algorithm %u\n",
D
David Sterba 已提交
433 434 435 436 437 438 439
				csum_type);
		ret = 1;
	}

	return ret;
}

C
Chris Mason 已提交
440 441 442 443
/*
 * helper to read a given tree block, doing retries as required when
 * the checksums don't match and we have alternate mirrors to try.
 */
444 445
static int btree_read_extent_buffer_pages(struct btrfs_root *root,
					  struct extent_buffer *eb,
446
					  u64 start, u64 parent_transid)
447 448
{
	struct extent_io_tree *io_tree;
449
	int failed = 0;
450 451 452
	int ret;
	int num_copies = 0;
	int mirror_num = 0;
453
	int failed_mirror = 0;
454

455
	clear_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags);
456 457
	io_tree = &BTRFS_I(root->fs_info->btree_inode)->io_tree;
	while (1) {
458 459
		ret = read_extent_buffer_pages(io_tree, eb, start,
					       WAIT_COMPLETE,
460
					       btree_get_extent, mirror_num);
461 462
		if (!ret) {
			if (!verify_parent_transid(io_tree, eb,
463
						   parent_transid, 0))
464 465 466 467
				break;
			else
				ret = -EIO;
		}
C
Chris Mason 已提交
468

469 470 471 472 473 474
		/*
		 * This buffer's crc is fine, but its contents are corrupted, so
		 * there is no reason to read the other copies, they won't be
		 * any less wrong.
		 */
		if (test_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags))
475 476
			break;

477
		num_copies = btrfs_num_copies(root->fs_info,
478
					      eb->start, eb->len);
C
Chris Mason 已提交
479
		if (num_copies == 1)
480
			break;
C
Chris Mason 已提交
481

482 483 484 485 486
		if (!failed_mirror) {
			failed = 1;
			failed_mirror = eb->read_mirror;
		}

487
		mirror_num++;
488 489 490
		if (mirror_num == failed_mirror)
			mirror_num++;

C
Chris Mason 已提交
491
		if (mirror_num > num_copies)
492
			break;
493
	}
494

495
	if (failed && !ret && failed_mirror)
496 497 498
		repair_eb_io_failure(root, eb, failed_mirror);

	return ret;
499
}
500

C
Chris Mason 已提交
501
/*
C
Chris Mason 已提交
502 503
 * checksum a dirty tree block before IO.  This has extra checks to make sure
 * we only fill in the checksum field in the first page of a multi-page block
C
Chris Mason 已提交
504
 */
C
Chris Mason 已提交
505

506
static int csum_dirty_buffer(struct btrfs_fs_info *fs_info, struct page *page)
507
{
M
Miao Xie 已提交
508
	u64 start = page_offset(page);
509 510
	u64 found_start;
	struct extent_buffer *eb;
511

J
Josef Bacik 已提交
512 513 514
	eb = (struct extent_buffer *)page->private;
	if (page != eb->pages[0])
		return 0;
515

516
	found_start = btrfs_header_bytenr(eb);
517 518 519 520 521 522 523 524 525 526 527 528
	/*
	 * Please do not consolidate these warnings into a single if.
	 * It is useful to know what went wrong.
	 */
	if (WARN_ON(found_start != start))
		return -EUCLEAN;
	if (WARN_ON(!PageUptodate(page)))
		return -EUCLEAN;

	ASSERT(memcmp_extent_buffer(eb, fs_info->fsid,
			btrfs_header_fsid(), BTRFS_FSID_SIZE) == 0);

529
	return csum_tree_block(fs_info, eb, 0);
530 531
}

532
static int check_tree_block_fsid(struct btrfs_fs_info *fs_info,
Y
Yan Zheng 已提交
533 534
				 struct extent_buffer *eb)
{
535
	struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
Y
Yan Zheng 已提交
536 537 538
	u8 fsid[BTRFS_UUID_SIZE];
	int ret = 1;

539
	read_extent_buffer(eb, fsid, btrfs_header_fsid(), BTRFS_FSID_SIZE);
Y
Yan Zheng 已提交
540 541 542 543 544 545 546 547 548 549
	while (fs_devices) {
		if (!memcmp(fsid, fs_devices->fsid, BTRFS_FSID_SIZE)) {
			ret = 0;
			break;
		}
		fs_devices = fs_devices->seed;
	}
	return ret;
}

550
#define CORRUPT(reason, eb, root, slot)				\
551 552
	btrfs_crit(root->fs_info, "corrupt leaf, %s: block=%llu,"	\
		   "root=%llu, slot=%d", reason,			\
553
	       btrfs_header_bytenr(eb),	root->objectid, slot)
554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615

static noinline int check_leaf(struct btrfs_root *root,
			       struct extent_buffer *leaf)
{
	struct btrfs_key key;
	struct btrfs_key leaf_key;
	u32 nritems = btrfs_header_nritems(leaf);
	int slot;

	if (nritems == 0)
		return 0;

	/* Check the 0 item */
	if (btrfs_item_offset_nr(leaf, 0) + btrfs_item_size_nr(leaf, 0) !=
	    BTRFS_LEAF_DATA_SIZE(root)) {
		CORRUPT("invalid item offset size pair", leaf, root, 0);
		return -EIO;
	}

	/*
	 * Check to make sure each items keys are in the correct order and their
	 * offsets make sense.  We only have to loop through nritems-1 because
	 * we check the current slot against the next slot, which verifies the
	 * next slot's offset+size makes sense and that the current's slot
	 * offset is correct.
	 */
	for (slot = 0; slot < nritems - 1; slot++) {
		btrfs_item_key_to_cpu(leaf, &leaf_key, slot);
		btrfs_item_key_to_cpu(leaf, &key, slot + 1);

		/* Make sure the keys are in the right order */
		if (btrfs_comp_cpu_keys(&leaf_key, &key) >= 0) {
			CORRUPT("bad key order", leaf, root, slot);
			return -EIO;
		}

		/*
		 * Make sure the offset and ends are right, remember that the
		 * item data starts at the end of the leaf and grows towards the
		 * front.
		 */
		if (btrfs_item_offset_nr(leaf, slot) !=
			btrfs_item_end_nr(leaf, slot + 1)) {
			CORRUPT("slot offset bad", leaf, root, slot);
			return -EIO;
		}

		/*
		 * Check to make sure that we don't point outside of the leaf,
		 * just incase all the items are consistent to eachother, but
		 * all point outside of the leaf.
		 */
		if (btrfs_item_end_nr(leaf, slot) >
		    BTRFS_LEAF_DATA_SIZE(root)) {
			CORRUPT("slot end outside of leaf", leaf, root, slot);
			return -EIO;
		}
	}

	return 0;
}

616 617 618
static int btree_readpage_end_io_hook(struct btrfs_io_bio *io_bio,
				      u64 phy_offset, struct page *page,
				      u64 start, u64 end, int mirror)
619 620 621 622 623
{
	u64 found_start;
	int found_level;
	struct extent_buffer *eb;
	struct btrfs_root *root = BTRFS_I(page->mapping->host)->root;
624
	struct btrfs_fs_info *fs_info = root->fs_info;
625
	int ret = 0;
626
	int reads_done;
627 628 629

	if (!page->private)
		goto out;
C
Chris Mason 已提交
630

J
Josef Bacik 已提交
631
	eb = (struct extent_buffer *)page->private;
C
Chris Mason 已提交
632

633 634 635 636 637 638
	/* the pending IO might have been the only thing that kept this buffer
	 * in memory.  Make sure we have a ref for all this other checks
	 */
	extent_buffer_get(eb);

	reads_done = atomic_dec_and_test(&eb->io_pages);
639 640
	if (!reads_done)
		goto err;
641

642
	eb->read_mirror = mirror;
643
	if (test_bit(EXTENT_BUFFER_READ_ERR, &eb->bflags)) {
644 645 646 647
		ret = -EIO;
		goto err;
	}

648
	found_start = btrfs_header_bytenr(eb);
649
	if (found_start != eb->start) {
650 651
		btrfs_err_rl(fs_info, "bad tree block start %llu %llu",
			     found_start, eb->start);
652
		ret = -EIO;
653 654
		goto err;
	}
655 656 657
	if (check_tree_block_fsid(fs_info, eb)) {
		btrfs_err_rl(fs_info, "bad fsid on block %llu",
			     eb->start);
658 659 660
		ret = -EIO;
		goto err;
	}
661
	found_level = btrfs_header_level(eb);
662
	if (found_level >= BTRFS_MAX_LEVEL) {
663 664
		btrfs_err(fs_info, "bad tree block level %d",
			  (int)btrfs_header_level(eb));
665 666 667
		ret = -EIO;
		goto err;
	}
668

669 670
	btrfs_set_buffer_lockdep_class(btrfs_header_owner(eb),
				       eb, found_level);
671

672
	ret = csum_tree_block(fs_info, eb, 1);
673
	if (ret)
674 675 676 677 678 679 680 681 682 683 684
		goto err;

	/*
	 * If this is a leaf block and it is corrupt, set the corrupt bit so
	 * that we don't try and read the other copies of this block, just
	 * return -EIO.
	 */
	if (found_level == 0 && check_leaf(root, eb)) {
		set_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags);
		ret = -EIO;
	}
685

686 687
	if (!ret)
		set_extent_buffer_uptodate(eb);
688
err:
689 690
	if (reads_done &&
	    test_and_clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags))
691
		btree_readahead_hook(fs_info, eb, eb->start, ret);
A
Arne Jansen 已提交
692

D
David Woodhouse 已提交
693 694 695 696 697 698 699
	if (ret) {
		/*
		 * our io error hook is going to dec the io pages
		 * again, we have to make sure it has something
		 * to decrement
		 */
		atomic_inc(&eb->io_pages);
700
		clear_extent_buffer_uptodate(eb);
D
David Woodhouse 已提交
701
	}
702
	free_extent_buffer(eb);
703
out:
704
	return ret;
705 706
}

707
static int btree_io_failed_hook(struct page *page, int failed_mirror)
A
Arne Jansen 已提交
708 709 710
{
	struct extent_buffer *eb;

J
Josef Bacik 已提交
711
	eb = (struct extent_buffer *)page->private;
712
	set_bit(EXTENT_BUFFER_READ_ERR, &eb->bflags);
713
	eb->read_mirror = failed_mirror;
D
David Woodhouse 已提交
714
	atomic_dec(&eb->io_pages);
715
	if (test_and_clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags))
716
		btree_readahead_hook(eb->fs_info, eb, eb->start, -EIO);
A
Arne Jansen 已提交
717 718 719
	return -EIO;	/* we fixed nothing */
}

720
static void end_workqueue_bio(struct bio *bio)
721
{
722
	struct btrfs_end_io_wq *end_io_wq = bio->bi_private;
723
	struct btrfs_fs_info *fs_info;
724 725
	struct btrfs_workqueue *wq;
	btrfs_work_func_t func;
726 727

	fs_info = end_io_wq->info;
728
	end_io_wq->error = bio->bi_error;
729

730
	if (bio->bi_rw & REQ_WRITE) {
731 732 733 734 735 736 737 738 739 740 741 742 743
		if (end_io_wq->metadata == BTRFS_WQ_ENDIO_METADATA) {
			wq = fs_info->endio_meta_write_workers;
			func = btrfs_endio_meta_write_helper;
		} else if (end_io_wq->metadata == BTRFS_WQ_ENDIO_FREE_SPACE) {
			wq = fs_info->endio_freespace_worker;
			func = btrfs_freespace_write_helper;
		} else if (end_io_wq->metadata == BTRFS_WQ_ENDIO_RAID56) {
			wq = fs_info->endio_raid56_workers;
			func = btrfs_endio_raid56_helper;
		} else {
			wq = fs_info->endio_write_workers;
			func = btrfs_endio_write_helper;
		}
744
	} else {
745 746 747 748 749
		if (unlikely(end_io_wq->metadata ==
			     BTRFS_WQ_ENDIO_DIO_REPAIR)) {
			wq = fs_info->endio_repair_workers;
			func = btrfs_endio_repair_helper;
		} else if (end_io_wq->metadata == BTRFS_WQ_ENDIO_RAID56) {
750 751 752 753 754 755 756 757 758
			wq = fs_info->endio_raid56_workers;
			func = btrfs_endio_raid56_helper;
		} else if (end_io_wq->metadata) {
			wq = fs_info->endio_meta_workers;
			func = btrfs_endio_meta_helper;
		} else {
			wq = fs_info->endio_workers;
			func = btrfs_endio_helper;
		}
759
	}
760 761 762

	btrfs_init_work(&end_io_wq->work, func, end_workqueue_fn, NULL, NULL);
	btrfs_queue_work(wq, &end_io_wq->work);
763 764
}

765
int btrfs_bio_wq_end_io(struct btrfs_fs_info *info, struct bio *bio,
766
			enum btrfs_wq_endio_type metadata)
767
{
768
	struct btrfs_end_io_wq *end_io_wq;
769

770
	end_io_wq = kmem_cache_alloc(btrfs_end_io_wq_cache, GFP_NOFS);
771 772 773 774 775
	if (!end_io_wq)
		return -ENOMEM;

	end_io_wq->private = bio->bi_private;
	end_io_wq->end_io = bio->bi_end_io;
776
	end_io_wq->info = info;
777 778
	end_io_wq->error = 0;
	end_io_wq->bio = bio;
779
	end_io_wq->metadata = metadata;
780 781 782

	bio->bi_private = end_io_wq;
	bio->bi_end_io = end_workqueue_bio;
783 784 785
	return 0;
}

786
unsigned long btrfs_async_submit_limit(struct btrfs_fs_info *info)
787
{
788
	unsigned long limit = min_t(unsigned long,
789
				    info->thread_pool_size,
790 791 792
				    info->fs_devices->open_devices);
	return 256 * limit;
}
793

C
Chris Mason 已提交
794 795 796
static void run_one_async_start(struct btrfs_work *work)
{
	struct async_submit_bio *async;
797
	int ret;
C
Chris Mason 已提交
798 799

	async = container_of(work, struct  async_submit_bio, work);
800 801 802 803 804
	ret = async->submit_bio_start(async->inode, async->rw, async->bio,
				      async->mirror_num, async->bio_flags,
				      async->bio_offset);
	if (ret)
		async->error = ret;
C
Chris Mason 已提交
805 806 807
}

static void run_one_async_done(struct btrfs_work *work)
808 809 810
{
	struct btrfs_fs_info *fs_info;
	struct async_submit_bio *async;
811
	int limit;
812 813 814

	async = container_of(work, struct  async_submit_bio, work);
	fs_info = BTRFS_I(async->inode)->root->fs_info;
815

816
	limit = btrfs_async_submit_limit(fs_info);
817 818
	limit = limit * 2 / 3;

819 820 821
	/*
	 * atomic_dec_return implies a barrier for waitqueue_active
	 */
822
	if (atomic_dec_return(&fs_info->nr_async_submits) < limit &&
823
	    waitqueue_active(&fs_info->async_submit_wait))
824 825
		wake_up(&fs_info->async_submit_wait);

826
	/* If an error occurred we just want to clean up the bio and move on */
827
	if (async->error) {
828 829
		async->bio->bi_error = async->error;
		bio_endio(async->bio);
830 831 832
		return;
	}

C
Chris Mason 已提交
833
	async->submit_bio_done(async->inode, async->rw, async->bio,
834 835
			       async->mirror_num, async->bio_flags,
			       async->bio_offset);
C
Chris Mason 已提交
836 837 838 839 840 841 842
}

static void run_one_async_free(struct btrfs_work *work)
{
	struct async_submit_bio *async;

	async = container_of(work, struct  async_submit_bio, work);
843 844 845
	kfree(async);
}

846 847
int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode,
			int rw, struct bio *bio, int mirror_num,
C
Chris Mason 已提交
848
			unsigned long bio_flags,
849
			u64 bio_offset,
C
Chris Mason 已提交
850 851
			extent_submit_bio_hook_t *submit_bio_start,
			extent_submit_bio_hook_t *submit_bio_done)
852 853 854 855 856 857 858 859 860 861 862
{
	struct async_submit_bio *async;

	async = kmalloc(sizeof(*async), GFP_NOFS);
	if (!async)
		return -ENOMEM;

	async->inode = inode;
	async->rw = rw;
	async->bio = bio;
	async->mirror_num = mirror_num;
C
Chris Mason 已提交
863 864 865
	async->submit_bio_start = submit_bio_start;
	async->submit_bio_done = submit_bio_done;

866
	btrfs_init_work(&async->work, btrfs_worker_helper, run_one_async_start,
867
			run_one_async_done, run_one_async_free);
C
Chris Mason 已提交
868

C
Chris Mason 已提交
869
	async->bio_flags = bio_flags;
870
	async->bio_offset = bio_offset;
871

872 873
	async->error = 0;

874
	atomic_inc(&fs_info->nr_async_submits);
875

876
	if (rw & REQ_SYNC)
877
		btrfs_set_work_high_priority(&async->work);
878

879
	btrfs_queue_work(fs_info->workers, &async->work);
880

C
Chris Mason 已提交
881
	while (atomic_read(&fs_info->async_submit_draining) &&
882 883 884 885 886
	      atomic_read(&fs_info->nr_async_submits)) {
		wait_event(fs_info->async_submit_wait,
			   (atomic_read(&fs_info->nr_async_submits) == 0));
	}

887 888 889
	return 0;
}

890 891
static int btree_csum_one_bio(struct bio *bio)
{
892
	struct bio_vec *bvec;
893
	struct btrfs_root *root;
894
	int i, ret = 0;
895

896
	bio_for_each_segment_all(bvec, bio, i) {
897
		root = BTRFS_I(bvec->bv_page->mapping->host)->root;
898
		ret = csum_dirty_buffer(root->fs_info, bvec->bv_page);
899 900
		if (ret)
			break;
901
	}
902

903
	return ret;
904 905
}

C
Chris Mason 已提交
906 907
static int __btree_submit_bio_start(struct inode *inode, int rw,
				    struct bio *bio, int mirror_num,
908 909
				    unsigned long bio_flags,
				    u64 bio_offset)
910
{
911 912
	/*
	 * when we're called for a write, we're already in the async
913
	 * submission context.  Just jump into btrfs_map_bio
914
	 */
915
	return btree_csum_one_bio(bio);
C
Chris Mason 已提交
916
}
917

C
Chris Mason 已提交
918
static int __btree_submit_bio_done(struct inode *inode, int rw, struct bio *bio,
919 920
				 int mirror_num, unsigned long bio_flags,
				 u64 bio_offset)
C
Chris Mason 已提交
921
{
922 923
	int ret;

924
	/*
C
Chris Mason 已提交
925 926
	 * when we're called for a write, we're already in the async
	 * submission context.  Just jump into btrfs_map_bio
927
	 */
928
	ret = btrfs_map_bio(BTRFS_I(inode)->root, rw, bio, mirror_num, 1);
929 930 931 932
	if (ret) {
		bio->bi_error = ret;
		bio_endio(bio);
	}
933
	return ret;
934 935
}

936 937 938 939 940
static int check_async_write(struct inode *inode, unsigned long bio_flags)
{
	if (bio_flags & EXTENT_BIO_TREE_LOG)
		return 0;
#ifdef CONFIG_X86
941
	if (static_cpu_has(X86_FEATURE_XMM4_2))
942 943 944 945 946
		return 0;
#endif
	return 1;
}

947
static int btree_submit_bio_hook(struct inode *inode, int rw, struct bio *bio,
948 949
				 int mirror_num, unsigned long bio_flags,
				 u64 bio_offset)
950
{
951
	int async = check_async_write(inode, bio_flags);
952 953
	int ret;

954
	if (!(rw & REQ_WRITE)) {
C
Chris Mason 已提交
955 956 957 958
		/*
		 * called for a read, do the setup so that checksum validation
		 * can happen in the async kernel threads
		 */
959
		ret = btrfs_bio_wq_end_io(BTRFS_I(inode)->root->fs_info,
960
					  bio, BTRFS_WQ_ENDIO_METADATA);
961
		if (ret)
962 963 964
			goto out_w_error;
		ret = btrfs_map_bio(BTRFS_I(inode)->root, rw, bio,
				    mirror_num, 0);
965 966 967
	} else if (!async) {
		ret = btree_csum_one_bio(bio);
		if (ret)
968 969 970 971 972 973 974 975 976 977 978 979 980
			goto out_w_error;
		ret = btrfs_map_bio(BTRFS_I(inode)->root, rw, bio,
				    mirror_num, 0);
	} else {
		/*
		 * kthread helpers are used to submit writes so that
		 * checksumming can happen in parallel across all CPUs
		 */
		ret = btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info,
					  inode, rw, bio, mirror_num, 0,
					  bio_offset,
					  __btree_submit_bio_start,
					  __btree_submit_bio_done);
981
	}
982

983 984 985 986
	if (ret)
		goto out_w_error;
	return 0;

987
out_w_error:
988 989
	bio->bi_error = ret;
	bio_endio(bio);
990
	return ret;
991 992
}

J
Jan Beulich 已提交
993
#ifdef CONFIG_MIGRATION
994
static int btree_migratepage(struct address_space *mapping,
995 996
			struct page *newpage, struct page *page,
			enum migrate_mode mode)
997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010
{
	/*
	 * we can't safely write a btree page from here,
	 * we haven't done the locking hook
	 */
	if (PageDirty(page))
		return -EAGAIN;
	/*
	 * Buffers may be managed in a filesystem specific way.
	 * We must have no buffers or drop them.
	 */
	if (page_has_private(page) &&
	    !try_to_release_page(page, GFP_KERNEL))
		return -EAGAIN;
1011
	return migrate_page(mapping, newpage, page, mode);
1012
}
J
Jan Beulich 已提交
1013
#endif
1014

1015 1016 1017 1018

static int btree_writepages(struct address_space *mapping,
			    struct writeback_control *wbc)
{
1019 1020 1021
	struct btrfs_fs_info *fs_info;
	int ret;

1022
	if (wbc->sync_mode == WB_SYNC_NONE) {
1023 1024 1025 1026

		if (wbc->for_kupdate)
			return 0;

1027
		fs_info = BTRFS_I(mapping->host)->root->fs_info;
1028
		/* this is a bit racy, but that's ok */
1029 1030 1031
		ret = percpu_counter_compare(&fs_info->dirty_metadata_bytes,
					     BTRFS_DIRTY_METADATA_THRESH);
		if (ret < 0)
1032 1033
			return 0;
	}
1034
	return btree_write_cache_pages(mapping, wbc);
1035 1036
}

1037
static int btree_readpage(struct file *file, struct page *page)
1038
{
1039 1040
	struct extent_io_tree *tree;
	tree = &BTRFS_I(page->mapping->host)->io_tree;
1041
	return extent_read_full_page(tree, page, btree_get_extent, 0);
1042
}
C
Chris Mason 已提交
1043

1044
static int btree_releasepage(struct page *page, gfp_t gfp_flags)
1045
{
1046
	if (PageWriteback(page) || PageDirty(page))
C
Chris Mason 已提交
1047
		return 0;
1048

1049
	return try_release_extent_buffer(page);
1050 1051
}

1052 1053
static void btree_invalidatepage(struct page *page, unsigned int offset,
				 unsigned int length)
1054
{
1055 1056
	struct extent_io_tree *tree;
	tree = &BTRFS_I(page->mapping->host)->io_tree;
1057 1058
	extent_invalidatepage(tree, page, offset);
	btree_releasepage(page, GFP_NOFS);
1059
	if (PagePrivate(page)) {
1060 1061 1062
		btrfs_warn(BTRFS_I(page->mapping->host)->root->fs_info,
			   "page private not zero on page %llu",
			   (unsigned long long)page_offset(page));
1063 1064
		ClearPagePrivate(page);
		set_page_private(page, 0);
1065
		put_page(page);
1066
	}
1067 1068
}

1069 1070
static int btree_set_page_dirty(struct page *page)
{
1071
#ifdef DEBUG
1072 1073 1074 1075 1076 1077 1078 1079
	struct extent_buffer *eb;

	BUG_ON(!PagePrivate(page));
	eb = (struct extent_buffer *)page->private;
	BUG_ON(!eb);
	BUG_ON(!test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags));
	BUG_ON(!atomic_read(&eb->refs));
	btrfs_assert_tree_locked(eb);
1080
#endif
1081 1082 1083
	return __set_page_dirty_nobuffers(page);
}

1084
static const struct address_space_operations btree_aops = {
1085
	.readpage	= btree_readpage,
1086
	.writepages	= btree_writepages,
1087 1088
	.releasepage	= btree_releasepage,
	.invalidatepage = btree_invalidatepage,
1089
#ifdef CONFIG_MIGRATION
1090
	.migratepage	= btree_migratepage,
1091
#endif
1092
	.set_page_dirty = btree_set_page_dirty,
1093 1094
};

1095
void readahead_tree_block(struct btrfs_root *root, u64 bytenr)
C
Chris Mason 已提交
1096
{
1097 1098
	struct extent_buffer *buf = NULL;
	struct inode *btree_inode = root->fs_info->btree_inode;
C
Chris Mason 已提交
1099

1100
	buf = btrfs_find_create_tree_block(root, bytenr);
1101
	if (!buf)
1102
		return;
1103
	read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree,
1104
				 buf, 0, WAIT_NONE, btree_get_extent, 0);
1105
	free_extent_buffer(buf);
C
Chris Mason 已提交
1106 1107
}

1108
int reada_tree_block_flagged(struct btrfs_root *root, u64 bytenr,
1109 1110 1111 1112 1113 1114 1115
			 int mirror_num, struct extent_buffer **eb)
{
	struct extent_buffer *buf = NULL;
	struct inode *btree_inode = root->fs_info->btree_inode;
	struct extent_io_tree *io_tree = &BTRFS_I(btree_inode)->io_tree;
	int ret;

1116
	buf = btrfs_find_create_tree_block(root, bytenr);
1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131
	if (!buf)
		return 0;

	set_bit(EXTENT_BUFFER_READAHEAD, &buf->bflags);

	ret = read_extent_buffer_pages(io_tree, buf, 0, WAIT_PAGE_LOCK,
				       btree_get_extent, mirror_num);
	if (ret) {
		free_extent_buffer(buf);
		return ret;
	}

	if (test_bit(EXTENT_BUFFER_CORRUPT, &buf->bflags)) {
		free_extent_buffer(buf);
		return -EIO;
1132
	} else if (extent_buffer_uptodate(buf)) {
1133 1134 1135 1136 1137 1138 1139
		*eb = buf;
	} else {
		free_extent_buffer(buf);
	}
	return 0;
}

1140
struct extent_buffer *btrfs_find_tree_block(struct btrfs_fs_info *fs_info,
1141
					    u64 bytenr)
1142
{
1143
	return find_extent_buffer(fs_info, bytenr);
1144 1145 1146
}

struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root,
1147
						 u64 bytenr)
1148
{
1149
	if (btrfs_test_is_dummy_root(root))
1150 1151
		return alloc_test_extent_buffer(root->fs_info, bytenr);
	return alloc_extent_buffer(root->fs_info, bytenr);
1152 1153 1154
}


1155 1156
int btrfs_write_tree_block(struct extent_buffer *buf)
{
1157
	return filemap_fdatawrite_range(buf->pages[0]->mapping, buf->start,
1158
					buf->start + buf->len - 1);
1159 1160 1161 1162
}

int btrfs_wait_tree_block_writeback(struct extent_buffer *buf)
{
1163
	return filemap_fdatawait_range(buf->pages[0]->mapping,
1164
				       buf->start, buf->start + buf->len - 1);
1165 1166
}

1167
struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
1168
				      u64 parent_transid)
1169 1170 1171 1172
{
	struct extent_buffer *buf = NULL;
	int ret;

1173
	buf = btrfs_find_create_tree_block(root, bytenr);
1174
	if (!buf)
1175
		return ERR_PTR(-ENOMEM);
1176

1177
	ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid);
1178 1179
	if (ret) {
		free_extent_buffer(buf);
1180
		return ERR_PTR(ret);
1181
	}
1182
	return buf;
1183

1184 1185
}

1186 1187
void clean_tree_block(struct btrfs_trans_handle *trans,
		      struct btrfs_fs_info *fs_info,
1188
		      struct extent_buffer *buf)
1189
{
1190
	if (btrfs_header_generation(buf) ==
1191
	    fs_info->running_transaction->transid) {
1192
		btrfs_assert_tree_locked(buf);
1193

1194
		if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &buf->bflags)) {
1195 1196 1197
			__percpu_counter_add(&fs_info->dirty_metadata_bytes,
					     -buf->len,
					     fs_info->dirty_metadata_batch);
1198 1199 1200 1201
			/* ugh, clear_extent_buffer_dirty needs to lock the page */
			btrfs_set_lock_blocking(buf);
			clear_extent_buffer_dirty(buf);
		}
1202
	}
1203 1204
}

1205 1206 1207 1208 1209 1210 1211 1212 1213
static struct btrfs_subvolume_writers *btrfs_alloc_subvolume_writers(void)
{
	struct btrfs_subvolume_writers *writers;
	int ret;

	writers = kmalloc(sizeof(*writers), GFP_NOFS);
	if (!writers)
		return ERR_PTR(-ENOMEM);

1214
	ret = percpu_counter_init(&writers->counter, 0, GFP_KERNEL);
1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230
	if (ret < 0) {
		kfree(writers);
		return ERR_PTR(ret);
	}

	init_waitqueue_head(&writers->wait);
	return writers;
}

static void
btrfs_free_subvolume_writers(struct btrfs_subvolume_writers *writers)
{
	percpu_counter_destroy(&writers->counter);
	kfree(writers);
}

1231 1232
static void __setup_root(u32 nodesize, u32 sectorsize, u32 stripesize,
			 struct btrfs_root *root, struct btrfs_fs_info *fs_info,
1233
			 u64 objectid)
1234
{
C
Chris Mason 已提交
1235
	root->node = NULL;
1236
	root->commit_root = NULL;
1237 1238
	root->sectorsize = sectorsize;
	root->nodesize = nodesize;
1239
	root->stripesize = stripesize;
1240
	root->state = 0;
1241
	root->orphan_cleanup_state = 0;
1242

1243 1244
	root->objectid = objectid;
	root->last_trans = 0;
1245
	root->highest_objectid = 0;
1246
	root->nr_delalloc_inodes = 0;
1247
	root->nr_ordered_extents = 0;
1248
	root->name = NULL;
1249
	root->inode_tree = RB_ROOT;
1250
	INIT_RADIX_TREE(&root->delayed_nodes_tree, GFP_ATOMIC);
1251
	root->block_rsv = NULL;
1252
	root->orphan_block_rsv = NULL;
1253 1254

	INIT_LIST_HEAD(&root->dirty_list);
1255
	INIT_LIST_HEAD(&root->root_list);
1256 1257
	INIT_LIST_HEAD(&root->delalloc_inodes);
	INIT_LIST_HEAD(&root->delalloc_root);
1258 1259
	INIT_LIST_HEAD(&root->ordered_extents);
	INIT_LIST_HEAD(&root->ordered_root);
1260 1261
	INIT_LIST_HEAD(&root->logged_list[0]);
	INIT_LIST_HEAD(&root->logged_list[1]);
1262
	spin_lock_init(&root->orphan_lock);
1263
	spin_lock_init(&root->inode_lock);
1264
	spin_lock_init(&root->delalloc_lock);
1265
	spin_lock_init(&root->ordered_extent_lock);
1266
	spin_lock_init(&root->accounting_lock);
1267 1268
	spin_lock_init(&root->log_extents_lock[0]);
	spin_lock_init(&root->log_extents_lock[1]);
1269
	mutex_init(&root->objectid_mutex);
1270
	mutex_init(&root->log_mutex);
1271
	mutex_init(&root->ordered_extent_mutex);
1272
	mutex_init(&root->delalloc_mutex);
Y
Yan Zheng 已提交
1273 1274 1275
	init_waitqueue_head(&root->log_writer_wait);
	init_waitqueue_head(&root->log_commit_wait[0]);
	init_waitqueue_head(&root->log_commit_wait[1]);
1276 1277
	INIT_LIST_HEAD(&root->log_ctxs[0]);
	INIT_LIST_HEAD(&root->log_ctxs[1]);
Y
Yan Zheng 已提交
1278 1279 1280
	atomic_set(&root->log_commit[0], 0);
	atomic_set(&root->log_commit[1], 0);
	atomic_set(&root->log_writers, 0);
M
Miao Xie 已提交
1281
	atomic_set(&root->log_batch, 0);
1282
	atomic_set(&root->orphan_inodes, 0);
1283
	atomic_set(&root->refs, 1);
1284
	atomic_set(&root->will_be_snapshoted, 0);
1285
	atomic_set(&root->qgroup_meta_rsv, 0);
Y
Yan Zheng 已提交
1286
	root->log_transid = 0;
1287
	root->log_transid_committed = -1;
1288
	root->last_log_commit = 0;
1289 1290 1291
	if (fs_info)
		extent_io_tree_init(&root->dirty_log_pages,
				     fs_info->btree_inode->i_mapping);
C
Chris Mason 已提交
1292

1293 1294
	memset(&root->root_key, 0, sizeof(root->root_key));
	memset(&root->root_item, 0, sizeof(root->root_item));
1295
	memset(&root->defrag_progress, 0, sizeof(root->defrag_progress));
1296 1297 1298 1299
	if (fs_info)
		root->defrag_trans_start = fs_info->generation;
	else
		root->defrag_trans_start = 0;
1300
	root->root_key.objectid = objectid;
1301
	root->anon_dev = 0;
1302

1303
	spin_lock_init(&root->root_item_lock);
1304 1305
}

1306 1307
static struct btrfs_root *btrfs_alloc_root(struct btrfs_fs_info *fs_info,
		gfp_t flags)
A
Al Viro 已提交
1308
{
1309
	struct btrfs_root *root = kzalloc(sizeof(*root), flags);
A
Al Viro 已提交
1310 1311 1312 1313 1314
	if (root)
		root->fs_info = fs_info;
	return root;
}

1315 1316 1317 1318 1319 1320
#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
/* Should only be used by the testing infrastructure */
struct btrfs_root *btrfs_alloc_dummy_root(void)
{
	struct btrfs_root *root;

1321
	root = btrfs_alloc_root(NULL, GFP_KERNEL);
1322 1323
	if (!root)
		return ERR_PTR(-ENOMEM);
1324
	__setup_root(4096, 4096, 4096, root, NULL, 1);
1325
	set_bit(BTRFS_ROOT_DUMMY_ROOT, &root->state);
1326
	root->alloc_bytenr = 0;
1327 1328 1329 1330 1331

	return root;
}
#endif

1332 1333 1334 1335 1336 1337 1338 1339 1340
struct btrfs_root *btrfs_create_tree(struct btrfs_trans_handle *trans,
				     struct btrfs_fs_info *fs_info,
				     u64 objectid)
{
	struct extent_buffer *leaf;
	struct btrfs_root *tree_root = fs_info->tree_root;
	struct btrfs_root *root;
	struct btrfs_key key;
	int ret = 0;
1341
	uuid_le uuid;
1342

1343
	root = btrfs_alloc_root(fs_info, GFP_KERNEL);
1344 1345 1346
	if (!root)
		return ERR_PTR(-ENOMEM);

1347 1348
	__setup_root(tree_root->nodesize, tree_root->sectorsize,
		tree_root->stripesize, root, fs_info, objectid);
1349 1350 1351 1352
	root->root_key.objectid = objectid;
	root->root_key.type = BTRFS_ROOT_ITEM_KEY;
	root->root_key.offset = 0;

1353
	leaf = btrfs_alloc_tree_block(trans, root, 0, objectid, NULL, 0, 0, 0);
1354 1355
	if (IS_ERR(leaf)) {
		ret = PTR_ERR(leaf);
1356
		leaf = NULL;
1357 1358 1359 1360 1361 1362 1363 1364 1365 1366
		goto fail;
	}

	memset_extent_buffer(leaf, 0, 0, sizeof(struct btrfs_header));
	btrfs_set_header_bytenr(leaf, leaf->start);
	btrfs_set_header_generation(leaf, trans->transid);
	btrfs_set_header_backref_rev(leaf, BTRFS_MIXED_BACKREF_REV);
	btrfs_set_header_owner(leaf, objectid);
	root->node = leaf;

1367
	write_extent_buffer(leaf, fs_info->fsid, btrfs_header_fsid(),
1368 1369
			    BTRFS_FSID_SIZE);
	write_extent_buffer(leaf, fs_info->chunk_tree_uuid,
1370
			    btrfs_header_chunk_tree_uuid(leaf),
1371 1372 1373 1374
			    BTRFS_UUID_SIZE);
	btrfs_mark_buffer_dirty(leaf);

	root->commit_root = btrfs_root_node(root);
1375
	set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
1376 1377 1378 1379 1380 1381 1382 1383 1384 1385

	root->root_item.flags = 0;
	root->root_item.byte_limit = 0;
	btrfs_set_root_bytenr(&root->root_item, leaf->start);
	btrfs_set_root_generation(&root->root_item, trans->transid);
	btrfs_set_root_level(&root->root_item, 0);
	btrfs_set_root_refs(&root->root_item, 1);
	btrfs_set_root_used(&root->root_item, leaf->len);
	btrfs_set_root_last_snapshot(&root->root_item, 0);
	btrfs_set_root_dirid(&root->root_item, 0);
1386 1387
	uuid_le_gen(&uuid);
	memcpy(root->root_item.uuid, uuid.b, BTRFS_UUID_SIZE);
1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398
	root->root_item.drop_level = 0;

	key.objectid = objectid;
	key.type = BTRFS_ROOT_ITEM_KEY;
	key.offset = 0;
	ret = btrfs_insert_root(trans, tree_root, &key, &root->root_item);
	if (ret)
		goto fail;

	btrfs_tree_unlock(leaf);

1399 1400
	return root;

1401
fail:
1402 1403
	if (leaf) {
		btrfs_tree_unlock(leaf);
1404
		free_extent_buffer(root->commit_root);
1405 1406 1407
		free_extent_buffer(leaf);
	}
	kfree(root);
1408

1409
	return ERR_PTR(ret);
1410 1411
}

Y
Yan Zheng 已提交
1412 1413
static struct btrfs_root *alloc_log_tree(struct btrfs_trans_handle *trans,
					 struct btrfs_fs_info *fs_info)
1414 1415 1416
{
	struct btrfs_root *root;
	struct btrfs_root *tree_root = fs_info->tree_root;
Y
Yan Zheng 已提交
1417
	struct extent_buffer *leaf;
1418

1419
	root = btrfs_alloc_root(fs_info, GFP_NOFS);
1420
	if (!root)
Y
Yan Zheng 已提交
1421
		return ERR_PTR(-ENOMEM);
1422

1423 1424 1425
	__setup_root(tree_root->nodesize, tree_root->sectorsize,
		     tree_root->stripesize, root, fs_info,
		     BTRFS_TREE_LOG_OBJECTID);
1426 1427 1428 1429

	root->root_key.objectid = BTRFS_TREE_LOG_OBJECTID;
	root->root_key.type = BTRFS_ROOT_ITEM_KEY;
	root->root_key.offset = BTRFS_TREE_LOG_OBJECTID;
1430

Y
Yan Zheng 已提交
1431
	/*
1432 1433
	 * DON'T set REF_COWS for log trees
	 *
Y
Yan Zheng 已提交
1434 1435 1436 1437 1438
	 * log trees do not get reference counted because they go away
	 * before a real commit is actually done.  They do store pointers
	 * to file data extents, and those reference counts still get
	 * updated (along with back refs to the log tree).
	 */
1439

1440 1441
	leaf = btrfs_alloc_tree_block(trans, root, 0, BTRFS_TREE_LOG_OBJECTID,
			NULL, 0, 0, 0);
Y
Yan Zheng 已提交
1442 1443 1444 1445
	if (IS_ERR(leaf)) {
		kfree(root);
		return ERR_CAST(leaf);
	}
1446

1447 1448 1449 1450 1451
	memset_extent_buffer(leaf, 0, 0, sizeof(struct btrfs_header));
	btrfs_set_header_bytenr(leaf, leaf->start);
	btrfs_set_header_generation(leaf, trans->transid);
	btrfs_set_header_backref_rev(leaf, BTRFS_MIXED_BACKREF_REV);
	btrfs_set_header_owner(leaf, BTRFS_TREE_LOG_OBJECTID);
Y
Yan Zheng 已提交
1452
	root->node = leaf;
1453 1454

	write_extent_buffer(root->node, root->fs_info->fsid,
1455
			    btrfs_header_fsid(), BTRFS_FSID_SIZE);
1456 1457
	btrfs_mark_buffer_dirty(root->node);
	btrfs_tree_unlock(root->node);
Y
Yan Zheng 已提交
1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487
	return root;
}

int btrfs_init_log_root_tree(struct btrfs_trans_handle *trans,
			     struct btrfs_fs_info *fs_info)
{
	struct btrfs_root *log_root;

	log_root = alloc_log_tree(trans, fs_info);
	if (IS_ERR(log_root))
		return PTR_ERR(log_root);
	WARN_ON(fs_info->log_root_tree);
	fs_info->log_root_tree = log_root;
	return 0;
}

int btrfs_add_log_tree(struct btrfs_trans_handle *trans,
		       struct btrfs_root *root)
{
	struct btrfs_root *log_root;
	struct btrfs_inode_item *inode_item;

	log_root = alloc_log_tree(trans, root->fs_info);
	if (IS_ERR(log_root))
		return PTR_ERR(log_root);

	log_root->last_trans = trans->transid;
	log_root->root_key.offset = root->root_key.objectid;

	inode_item = &log_root->root_item.inode;
1488 1489 1490
	btrfs_set_stack_inode_generation(inode_item, 1);
	btrfs_set_stack_inode_size(inode_item, 3);
	btrfs_set_stack_inode_nlink(inode_item, 1);
1491
	btrfs_set_stack_inode_nbytes(inode_item, root->nodesize);
1492
	btrfs_set_stack_inode_mode(inode_item, S_IFDIR | 0755);
Y
Yan Zheng 已提交
1493

1494
	btrfs_set_root_node(&log_root->root_item, log_root->node);
Y
Yan Zheng 已提交
1495 1496 1497 1498

	WARN_ON(root->log_root);
	root->log_root = log_root;
	root->log_transid = 0;
1499
	root->log_transid_committed = -1;
1500
	root->last_log_commit = 0;
1501 1502 1503
	return 0;
}

1504 1505
static struct btrfs_root *btrfs_read_tree_root(struct btrfs_root *tree_root,
					       struct btrfs_key *key)
1506 1507 1508
{
	struct btrfs_root *root;
	struct btrfs_fs_info *fs_info = tree_root->fs_info;
1509
	struct btrfs_path *path;
1510
	u64 generation;
1511
	int ret;
1512

1513 1514
	path = btrfs_alloc_path();
	if (!path)
1515
		return ERR_PTR(-ENOMEM);
1516

1517
	root = btrfs_alloc_root(fs_info, GFP_NOFS);
1518 1519 1520
	if (!root) {
		ret = -ENOMEM;
		goto alloc_fail;
1521 1522
	}

1523 1524
	__setup_root(tree_root->nodesize, tree_root->sectorsize,
		tree_root->stripesize, root, fs_info, key->objectid);
1525

1526 1527
	ret = btrfs_find_root(tree_root, key, path,
			      &root->root_item, &root->root_key);
1528
	if (ret) {
1529 1530
		if (ret > 0)
			ret = -ENOENT;
1531
		goto find_fail;
1532
	}
1533

1534
	generation = btrfs_root_generation(&root->root_item);
1535
	root->node = read_tree_block(root, btrfs_root_bytenr(&root->root_item),
1536
				     generation);
1537 1538
	if (IS_ERR(root->node)) {
		ret = PTR_ERR(root->node);
1539 1540 1541
		goto find_fail;
	} else if (!btrfs_buffer_uptodate(root->node, generation, 0)) {
		ret = -EIO;
1542 1543
		free_extent_buffer(root->node);
		goto find_fail;
1544
	}
1545
	root->commit_root = btrfs_root_node(root);
1546
out:
1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566
	btrfs_free_path(path);
	return root;

find_fail:
	kfree(root);
alloc_fail:
	root = ERR_PTR(ret);
	goto out;
}

struct btrfs_root *btrfs_read_fs_root(struct btrfs_root *tree_root,
				      struct btrfs_key *location)
{
	struct btrfs_root *root;

	root = btrfs_read_tree_root(tree_root, location);
	if (IS_ERR(root))
		return root;

	if (root->root_key.objectid != BTRFS_TREE_LOG_OBJECTID) {
1567
		set_bit(BTRFS_ROOT_REF_COWS, &root->state);
1568 1569
		btrfs_check_and_init_root_item(&root->root_item);
	}
1570

1571 1572 1573
	return root;
}

1574 1575 1576
int btrfs_init_fs_root(struct btrfs_root *root)
{
	int ret;
1577
	struct btrfs_subvolume_writers *writers;
1578 1579 1580 1581 1582 1583 1584 1585 1586

	root->free_ino_ctl = kzalloc(sizeof(*root->free_ino_ctl), GFP_NOFS);
	root->free_ino_pinned = kzalloc(sizeof(*root->free_ino_pinned),
					GFP_NOFS);
	if (!root->free_ino_pinned || !root->free_ino_ctl) {
		ret = -ENOMEM;
		goto fail;
	}

1587 1588 1589 1590 1591 1592 1593
	writers = btrfs_alloc_subvolume_writers();
	if (IS_ERR(writers)) {
		ret = PTR_ERR(writers);
		goto fail;
	}
	root->subv_writers = writers;

1594
	btrfs_init_free_ino_ctl(root);
1595 1596
	spin_lock_init(&root->ino_cache_lock);
	init_waitqueue_head(&root->ino_cache_wait);
1597 1598 1599

	ret = get_anon_bdev(&root->anon_dev);
	if (ret)
1600
		goto free_writers;
1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613

	mutex_lock(&root->objectid_mutex);
	ret = btrfs_find_highest_objectid(root,
					&root->highest_objectid);
	if (ret) {
		mutex_unlock(&root->objectid_mutex);
		goto free_root_dev;
	}

	ASSERT(root->highest_objectid <= BTRFS_LAST_FREE_OBJECTID);

	mutex_unlock(&root->objectid_mutex);

1614
	return 0;
1615

1616 1617
free_root_dev:
	free_anon_bdev(root->anon_dev);
1618 1619
free_writers:
	btrfs_free_subvolume_writers(root->subv_writers);
1620 1621 1622 1623 1624 1625
fail:
	kfree(root->free_ino_ctl);
	kfree(root->free_ino_pinned);
	return ret;
}

1626 1627
static struct btrfs_root *btrfs_lookup_fs_root(struct btrfs_fs_info *fs_info,
					       u64 root_id)
1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651
{
	struct btrfs_root *root;

	spin_lock(&fs_info->fs_roots_radix_lock);
	root = radix_tree_lookup(&fs_info->fs_roots_radix,
				 (unsigned long)root_id);
	spin_unlock(&fs_info->fs_roots_radix_lock);
	return root;
}

int btrfs_insert_fs_root(struct btrfs_fs_info *fs_info,
			 struct btrfs_root *root)
{
	int ret;

	ret = radix_tree_preload(GFP_NOFS & ~__GFP_HIGHMEM);
	if (ret)
		return ret;

	spin_lock(&fs_info->fs_roots_radix_lock);
	ret = radix_tree_insert(&fs_info->fs_roots_radix,
				(unsigned long)root->root_key.objectid,
				root);
	if (ret == 0)
1652
		set_bit(BTRFS_ROOT_IN_RADIX, &root->state);
1653 1654 1655 1656 1657 1658
	spin_unlock(&fs_info->fs_roots_radix_lock);
	radix_tree_preload_end();

	return ret;
}

1659 1660 1661
struct btrfs_root *btrfs_get_fs_root(struct btrfs_fs_info *fs_info,
				     struct btrfs_key *location,
				     bool check_ref)
1662 1663
{
	struct btrfs_root *root;
1664
	struct btrfs_path *path;
1665
	struct btrfs_key key;
1666 1667
	int ret;

1668 1669 1670 1671
	if (location->objectid == BTRFS_ROOT_TREE_OBJECTID)
		return fs_info->tree_root;
	if (location->objectid == BTRFS_EXTENT_TREE_OBJECTID)
		return fs_info->extent_root;
1672 1673 1674 1675
	if (location->objectid == BTRFS_CHUNK_TREE_OBJECTID)
		return fs_info->chunk_root;
	if (location->objectid == BTRFS_DEV_TREE_OBJECTID)
		return fs_info->dev_root;
1676 1677
	if (location->objectid == BTRFS_CSUM_TREE_OBJECTID)
		return fs_info->csum_root;
1678 1679 1680
	if (location->objectid == BTRFS_QUOTA_TREE_OBJECTID)
		return fs_info->quota_root ? fs_info->quota_root :
					     ERR_PTR(-ENOENT);
1681 1682 1683
	if (location->objectid == BTRFS_UUID_TREE_OBJECTID)
		return fs_info->uuid_root ? fs_info->uuid_root :
					    ERR_PTR(-ENOENT);
1684 1685 1686
	if (location->objectid == BTRFS_FREE_SPACE_TREE_OBJECTID)
		return fs_info->free_space_root ? fs_info->free_space_root :
						  ERR_PTR(-ENOENT);
1687
again:
1688
	root = btrfs_lookup_fs_root(fs_info, location->objectid);
1689
	if (root) {
1690
		if (check_ref && btrfs_root_refs(&root->root_item) == 0)
1691
			return ERR_PTR(-ENOENT);
1692
		return root;
1693
	}
1694

1695
	root = btrfs_read_fs_root(fs_info->tree_root, location);
1696 1697
	if (IS_ERR(root))
		return root;
1698

1699
	if (check_ref && btrfs_root_refs(&root->root_item) == 0) {
1700
		ret = -ENOENT;
1701
		goto fail;
1702
	}
1703

1704
	ret = btrfs_init_fs_root(root);
1705 1706
	if (ret)
		goto fail;
1707

1708 1709 1710 1711 1712
	path = btrfs_alloc_path();
	if (!path) {
		ret = -ENOMEM;
		goto fail;
	}
1713 1714 1715 1716 1717
	key.objectid = BTRFS_ORPHAN_OBJECTID;
	key.type = BTRFS_ORPHAN_ITEM_KEY;
	key.offset = location->objectid;

	ret = btrfs_search_slot(NULL, fs_info->tree_root, &key, path, 0, 0);
1718
	btrfs_free_path(path);
1719 1720 1721
	if (ret < 0)
		goto fail;
	if (ret == 0)
1722
		set_bit(BTRFS_ROOT_ORPHAN_ITEM_INSERTED, &root->state);
1723

1724
	ret = btrfs_insert_fs_root(fs_info, root);
1725
	if (ret) {
1726 1727 1728 1729 1730
		if (ret == -EEXIST) {
			free_fs_root(root);
			goto again;
		}
		goto fail;
1731
	}
1732
	return root;
1733 1734 1735
fail:
	free_fs_root(root);
	return ERR_PTR(ret);
1736 1737
}

C
Chris Mason 已提交
1738 1739 1740 1741 1742 1743
static int btrfs_congested_fn(void *congested_data, int bdi_bits)
{
	struct btrfs_fs_info *info = (struct btrfs_fs_info *)congested_data;
	int ret = 0;
	struct btrfs_device *device;
	struct backing_dev_info *bdi;
C
Chris Mason 已提交
1744

1745 1746
	rcu_read_lock();
	list_for_each_entry_rcu(device, &info->fs_devices->devices, dev_list) {
1747 1748
		if (!device->bdev)
			continue;
C
Chris Mason 已提交
1749
		bdi = blk_get_backing_dev_info(device->bdev);
1750
		if (bdi_congested(bdi, bdi_bits)) {
C
Chris Mason 已提交
1751 1752 1753 1754
			ret = 1;
			break;
		}
	}
1755
	rcu_read_unlock();
C
Chris Mason 已提交
1756 1757 1758 1759 1760
	return ret;
}

static int setup_bdi(struct btrfs_fs_info *info, struct backing_dev_info *bdi)
{
1761 1762
	int err;

1763
	err = bdi_setup_and_register(bdi, "btrfs");
1764 1765 1766
	if (err)
		return err;

1767
	bdi->ra_pages = VM_MAX_READAHEAD * 1024 / PAGE_SIZE;
C
Chris Mason 已提交
1768 1769
	bdi->congested_fn	= btrfs_congested_fn;
	bdi->congested_data	= info;
1770
	bdi->capabilities |= BDI_CAP_CGROUP_WRITEBACK;
C
Chris Mason 已提交
1771 1772 1773
	return 0;
}

1774 1775 1776 1777 1778
/*
 * called by the kthread helper functions to finally call the bio end_io
 * functions.  This is where read checksum verification actually happens
 */
static void end_workqueue_fn(struct btrfs_work *work)
1779 1780
{
	struct bio *bio;
1781
	struct btrfs_end_io_wq *end_io_wq;
1782

1783
	end_io_wq = container_of(work, struct btrfs_end_io_wq, work);
1784
	bio = end_io_wq->bio;
1785

1786
	bio->bi_error = end_io_wq->error;
1787 1788
	bio->bi_private = end_io_wq->private;
	bio->bi_end_io = end_io_wq->end_io;
1789
	kmem_cache_free(btrfs_end_io_wq_cache, end_io_wq);
1790
	bio_endio(bio);
1791 1792
}

1793 1794 1795
static int cleaner_kthread(void *arg)
{
	struct btrfs_root *root = arg;
1796
	int again;
1797
	struct btrfs_trans_handle *trans;
1798 1799

	do {
1800
		again = 0;
1801

1802
		/* Make the cleaner go to sleep early. */
1803
		if (btrfs_need_cleaner_sleep(root))
1804 1805 1806 1807 1808
			goto sleep;

		if (!mutex_trylock(&root->fs_info->cleaner_mutex))
			goto sleep;

1809 1810 1811 1812
		/*
		 * Avoid the problem that we change the status of the fs
		 * during the above check and trylock.
		 */
1813
		if (btrfs_need_cleaner_sleep(root)) {
1814 1815
			mutex_unlock(&root->fs_info->cleaner_mutex);
			goto sleep;
1816
		}
1817

1818
		mutex_lock(&root->fs_info->cleaner_delayed_iput_mutex);
1819
		btrfs_run_delayed_iputs(root);
1820 1821
		mutex_unlock(&root->fs_info->cleaner_delayed_iput_mutex);

1822 1823 1824 1825
		again = btrfs_clean_one_deleted_snapshot(root);
		mutex_unlock(&root->fs_info->cleaner_mutex);

		/*
1826 1827
		 * The defragger has dealt with the R/O remount and umount,
		 * needn't do anything special here.
1828 1829
		 */
		btrfs_run_defrag_inodes(root->fs_info);
1830 1831 1832 1833 1834 1835 1836 1837 1838 1839

		/*
		 * Acquires fs_info->delete_unused_bgs_mutex to avoid racing
		 * with relocation (btrfs_relocate_chunk) and relocation
		 * acquires fs_info->cleaner_mutex (btrfs_relocate_block_group)
		 * after acquiring fs_info->delete_unused_bgs_mutex. So we
		 * can't hold, nor need to, fs_info->cleaner_mutex when deleting
		 * unused block groups.
		 */
		btrfs_delete_unused_bgs(root->fs_info);
1840
sleep:
1841
		if (!again) {
1842
			set_current_state(TASK_INTERRUPTIBLE);
1843 1844
			if (!kthread_should_stop())
				schedule();
1845 1846 1847
			__set_current_state(TASK_RUNNING);
		}
	} while (!kthread_should_stop());
1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875

	/*
	 * Transaction kthread is stopped before us and wakes us up.
	 * However we might have started a new transaction and COWed some
	 * tree blocks when deleting unused block groups for example. So
	 * make sure we commit the transaction we started to have a clean
	 * shutdown when evicting the btree inode - if it has dirty pages
	 * when we do the final iput() on it, eviction will trigger a
	 * writeback for it which will fail with null pointer dereferences
	 * since work queues and other resources were already released and
	 * destroyed by the time the iput/eviction/writeback is made.
	 */
	trans = btrfs_attach_transaction(root);
	if (IS_ERR(trans)) {
		if (PTR_ERR(trans) != -ENOENT)
			btrfs_err(root->fs_info,
				  "cleaner transaction attach returned %ld",
				  PTR_ERR(trans));
	} else {
		int ret;

		ret = btrfs_commit_transaction(trans, root);
		if (ret)
			btrfs_err(root->fs_info,
				  "cleaner open transaction commit returned %d",
				  ret);
	}

1876 1877 1878 1879 1880 1881 1882 1883
	return 0;
}

static int transaction_kthread(void *arg)
{
	struct btrfs_root *root = arg;
	struct btrfs_trans_handle *trans;
	struct btrfs_transaction *cur;
1884
	u64 transid;
1885 1886
	unsigned long now;
	unsigned long delay;
1887
	bool cannot_commit;
1888 1889

	do {
1890
		cannot_commit = false;
1891
		delay = HZ * root->fs_info->commit_interval;
1892 1893
		mutex_lock(&root->fs_info->transaction_kthread_mutex);

J
Josef Bacik 已提交
1894
		spin_lock(&root->fs_info->trans_lock);
1895 1896
		cur = root->fs_info->running_transaction;
		if (!cur) {
J
Josef Bacik 已提交
1897
			spin_unlock(&root->fs_info->trans_lock);
1898 1899
			goto sleep;
		}
Y
Yan Zheng 已提交
1900

1901
		now = get_seconds();
1902
		if (cur->state < TRANS_STATE_BLOCKED &&
1903 1904
		    (now < cur->start_time ||
		     now - cur->start_time < root->fs_info->commit_interval)) {
J
Josef Bacik 已提交
1905
			spin_unlock(&root->fs_info->trans_lock);
1906 1907 1908
			delay = HZ * 5;
			goto sleep;
		}
1909
		transid = cur->transid;
J
Josef Bacik 已提交
1910
		spin_unlock(&root->fs_info->trans_lock);
1911

1912
		/* If the file system is aborted, this will always fail. */
1913
		trans = btrfs_attach_transaction(root);
1914
		if (IS_ERR(trans)) {
1915 1916
			if (PTR_ERR(trans) != -ENOENT)
				cannot_commit = true;
1917
			goto sleep;
1918
		}
1919
		if (transid == trans->transid) {
1920
			btrfs_commit_transaction(trans, root);
1921 1922 1923
		} else {
			btrfs_end_transaction(trans, root);
		}
1924 1925 1926 1927
sleep:
		wake_up_process(root->fs_info->cleaner_kthread);
		mutex_unlock(&root->fs_info->transaction_kthread_mutex);

J
Josef Bacik 已提交
1928 1929 1930
		if (unlikely(test_bit(BTRFS_FS_STATE_ERROR,
				      &root->fs_info->fs_state)))
			btrfs_cleanup_transaction(root);
1931 1932 1933 1934 1935 1936
		set_current_state(TASK_INTERRUPTIBLE);
		if (!kthread_should_stop() &&
				(!btrfs_transaction_blocked(root->fs_info) ||
				 cannot_commit))
			schedule_timeout(delay);
		__set_current_state(TASK_RUNNING);
1937 1938 1939 1940
	} while (!kthread_should_stop());
	return 0;
}

C
Chris Mason 已提交
1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046
/*
 * this will find the highest generation in the array of
 * root backups.  The index of the highest array is returned,
 * or -1 if we can't find anything.
 *
 * We check to make sure the array is valid by comparing the
 * generation of the latest  root in the array with the generation
 * in the super block.  If they don't match we pitch it.
 */
static int find_newest_super_backup(struct btrfs_fs_info *info, u64 newest_gen)
{
	u64 cur;
	int newest_index = -1;
	struct btrfs_root_backup *root_backup;
	int i;

	for (i = 0; i < BTRFS_NUM_BACKUP_ROOTS; i++) {
		root_backup = info->super_copy->super_roots + i;
		cur = btrfs_backup_tree_root_gen(root_backup);
		if (cur == newest_gen)
			newest_index = i;
	}

	/* check to see if we actually wrapped around */
	if (newest_index == BTRFS_NUM_BACKUP_ROOTS - 1) {
		root_backup = info->super_copy->super_roots;
		cur = btrfs_backup_tree_root_gen(root_backup);
		if (cur == newest_gen)
			newest_index = 0;
	}
	return newest_index;
}


/*
 * find the oldest backup so we know where to store new entries
 * in the backup array.  This will set the backup_root_index
 * field in the fs_info struct
 */
static void find_oldest_super_backup(struct btrfs_fs_info *info,
				     u64 newest_gen)
{
	int newest_index = -1;

	newest_index = find_newest_super_backup(info, newest_gen);
	/* if there was garbage in there, just move along */
	if (newest_index == -1) {
		info->backup_root_index = 0;
	} else {
		info->backup_root_index = (newest_index + 1) % BTRFS_NUM_BACKUP_ROOTS;
	}
}

/*
 * copy all the root pointers into the super backup array.
 * this will bump the backup pointer by one when it is
 * done
 */
static void backup_super_roots(struct btrfs_fs_info *info)
{
	int next_backup;
	struct btrfs_root_backup *root_backup;
	int last_backup;

	next_backup = info->backup_root_index;
	last_backup = (next_backup + BTRFS_NUM_BACKUP_ROOTS - 1) %
		BTRFS_NUM_BACKUP_ROOTS;

	/*
	 * just overwrite the last backup if we're at the same generation
	 * this happens only at umount
	 */
	root_backup = info->super_for_commit->super_roots + last_backup;
	if (btrfs_backup_tree_root_gen(root_backup) ==
	    btrfs_header_generation(info->tree_root->node))
		next_backup = last_backup;

	root_backup = info->super_for_commit->super_roots + next_backup;

	/*
	 * make sure all of our padding and empty slots get zero filled
	 * regardless of which ones we use today
	 */
	memset(root_backup, 0, sizeof(*root_backup));

	info->backup_root_index = (next_backup + 1) % BTRFS_NUM_BACKUP_ROOTS;

	btrfs_set_backup_tree_root(root_backup, info->tree_root->node->start);
	btrfs_set_backup_tree_root_gen(root_backup,
			       btrfs_header_generation(info->tree_root->node));

	btrfs_set_backup_tree_root_level(root_backup,
			       btrfs_header_level(info->tree_root->node));

	btrfs_set_backup_chunk_root(root_backup, info->chunk_root->node->start);
	btrfs_set_backup_chunk_root_gen(root_backup,
			       btrfs_header_generation(info->chunk_root->node));
	btrfs_set_backup_chunk_root_level(root_backup,
			       btrfs_header_level(info->chunk_root->node));

	btrfs_set_backup_extent_root(root_backup, info->extent_root->node->start);
	btrfs_set_backup_extent_root_gen(root_backup,
			       btrfs_header_generation(info->extent_root->node));
	btrfs_set_backup_extent_root_level(root_backup,
			       btrfs_header_level(info->extent_root->node));

2047 2048 2049 2050 2051 2052 2053 2054
	/*
	 * we might commit during log recovery, which happens before we set
	 * the fs_root.  Make sure it is valid before we fill it in.
	 */
	if (info->fs_root && info->fs_root->node) {
		btrfs_set_backup_fs_root(root_backup,
					 info->fs_root->node->start);
		btrfs_set_backup_fs_root_gen(root_backup,
C
Chris Mason 已提交
2055
			       btrfs_header_generation(info->fs_root->node));
2056
		btrfs_set_backup_fs_root_level(root_backup,
C
Chris Mason 已提交
2057
			       btrfs_header_level(info->fs_root->node));
2058
	}
C
Chris Mason 已提交
2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139

	btrfs_set_backup_dev_root(root_backup, info->dev_root->node->start);
	btrfs_set_backup_dev_root_gen(root_backup,
			       btrfs_header_generation(info->dev_root->node));
	btrfs_set_backup_dev_root_level(root_backup,
				       btrfs_header_level(info->dev_root->node));

	btrfs_set_backup_csum_root(root_backup, info->csum_root->node->start);
	btrfs_set_backup_csum_root_gen(root_backup,
			       btrfs_header_generation(info->csum_root->node));
	btrfs_set_backup_csum_root_level(root_backup,
			       btrfs_header_level(info->csum_root->node));

	btrfs_set_backup_total_bytes(root_backup,
			     btrfs_super_total_bytes(info->super_copy));
	btrfs_set_backup_bytes_used(root_backup,
			     btrfs_super_bytes_used(info->super_copy));
	btrfs_set_backup_num_devices(root_backup,
			     btrfs_super_num_devices(info->super_copy));

	/*
	 * if we don't copy this out to the super_copy, it won't get remembered
	 * for the next commit
	 */
	memcpy(&info->super_copy->super_roots,
	       &info->super_for_commit->super_roots,
	       sizeof(*root_backup) * BTRFS_NUM_BACKUP_ROOTS);
}

/*
 * this copies info out of the root backup array and back into
 * the in-memory super block.  It is meant to help iterate through
 * the array, so you send it the number of backups you've already
 * tried and the last backup index you used.
 *
 * this returns -1 when it has tried all the backups
 */
static noinline int next_root_backup(struct btrfs_fs_info *info,
				     struct btrfs_super_block *super,
				     int *num_backups_tried, int *backup_index)
{
	struct btrfs_root_backup *root_backup;
	int newest = *backup_index;

	if (*num_backups_tried == 0) {
		u64 gen = btrfs_super_generation(super);

		newest = find_newest_super_backup(info, gen);
		if (newest == -1)
			return -1;

		*backup_index = newest;
		*num_backups_tried = 1;
	} else if (*num_backups_tried == BTRFS_NUM_BACKUP_ROOTS) {
		/* we've tried all the backups, all done */
		return -1;
	} else {
		/* jump to the next oldest backup */
		newest = (*backup_index + BTRFS_NUM_BACKUP_ROOTS - 1) %
			BTRFS_NUM_BACKUP_ROOTS;
		*backup_index = newest;
		*num_backups_tried += 1;
	}
	root_backup = super->super_roots + newest;

	btrfs_set_super_generation(super,
				   btrfs_backup_tree_root_gen(root_backup));
	btrfs_set_super_root(super, btrfs_backup_tree_root(root_backup));
	btrfs_set_super_root_level(super,
				   btrfs_backup_tree_root_level(root_backup));
	btrfs_set_super_bytes_used(super, btrfs_backup_bytes_used(root_backup));

	/*
	 * fixme: the total bytes and num_devices need to match or we should
	 * need a fsck
	 */
	btrfs_set_super_total_bytes(super, btrfs_backup_total_bytes(root_backup));
	btrfs_set_super_num_devices(super, btrfs_backup_num_devices(root_backup));
	return 0;
}

L
Liu Bo 已提交
2140 2141 2142
/* helper to cleanup workers */
static void btrfs_stop_all_workers(struct btrfs_fs_info *fs_info)
{
2143
	btrfs_destroy_workqueue(fs_info->fixup_workers);
2144
	btrfs_destroy_workqueue(fs_info->delalloc_workers);
2145
	btrfs_destroy_workqueue(fs_info->workers);
2146 2147 2148
	btrfs_destroy_workqueue(fs_info->endio_workers);
	btrfs_destroy_workqueue(fs_info->endio_meta_workers);
	btrfs_destroy_workqueue(fs_info->endio_raid56_workers);
2149
	btrfs_destroy_workqueue(fs_info->endio_repair_workers);
2150
	btrfs_destroy_workqueue(fs_info->rmw_workers);
2151 2152 2153
	btrfs_destroy_workqueue(fs_info->endio_meta_write_workers);
	btrfs_destroy_workqueue(fs_info->endio_write_workers);
	btrfs_destroy_workqueue(fs_info->endio_freespace_worker);
2154
	btrfs_destroy_workqueue(fs_info->submit_workers);
2155
	btrfs_destroy_workqueue(fs_info->delayed_workers);
2156
	btrfs_destroy_workqueue(fs_info->caching_workers);
2157
	btrfs_destroy_workqueue(fs_info->readahead_workers);
2158
	btrfs_destroy_workqueue(fs_info->flush_workers);
2159
	btrfs_destroy_workqueue(fs_info->qgroup_rescan_workers);
C
Chris Mason 已提交
2160
	btrfs_destroy_workqueue(fs_info->extent_workers);
L
Liu Bo 已提交
2161 2162
}

2163 2164 2165 2166 2167 2168 2169 2170 2171 2172
static void free_root_extent_buffers(struct btrfs_root *root)
{
	if (root) {
		free_extent_buffer(root->node);
		free_extent_buffer(root->commit_root);
		root->node = NULL;
		root->commit_root = NULL;
	}
}

C
Chris Mason 已提交
2173 2174 2175
/* helper to cleanup tree roots */
static void free_root_pointers(struct btrfs_fs_info *info, int chunk_root)
{
2176
	free_root_extent_buffers(info->tree_root);
2177

2178 2179 2180 2181 2182 2183 2184
	free_root_extent_buffers(info->dev_root);
	free_root_extent_buffers(info->extent_root);
	free_root_extent_buffers(info->csum_root);
	free_root_extent_buffers(info->quota_root);
	free_root_extent_buffers(info->uuid_root);
	if (chunk_root)
		free_root_extent_buffers(info->chunk_root);
2185
	free_root_extent_buffers(info->free_space_root);
C
Chris Mason 已提交
2186 2187
}

2188
void btrfs_free_fs_roots(struct btrfs_fs_info *fs_info)
2189 2190 2191 2192 2193 2194 2195 2196 2197 2198
{
	int ret;
	struct btrfs_root *gang[8];
	int i;

	while (!list_empty(&fs_info->dead_roots)) {
		gang[0] = list_entry(fs_info->dead_roots.next,
				     struct btrfs_root, root_list);
		list_del(&gang[0]->root_list);

2199
		if (test_bit(BTRFS_ROOT_IN_RADIX, &gang[0]->state)) {
2200
			btrfs_drop_and_free_fs_root(fs_info, gang[0]);
2201 2202 2203
		} else {
			free_extent_buffer(gang[0]->node);
			free_extent_buffer(gang[0]->commit_root);
2204
			btrfs_put_fs_root(gang[0]);
2205 2206 2207 2208 2209 2210 2211 2212 2213 2214
		}
	}

	while (1) {
		ret = radix_tree_gang_lookup(&fs_info->fs_roots_radix,
					     (void **)gang, 0,
					     ARRAY_SIZE(gang));
		if (!ret)
			break;
		for (i = 0; i < ret; i++)
2215
			btrfs_drop_and_free_fs_root(fs_info, gang[i]);
2216
	}
2217 2218 2219 2220 2221 2222

	if (test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state)) {
		btrfs_free_log_root_tree(NULL, fs_info);
		btrfs_destroy_pinned_extent(fs_info->tree_root,
					    fs_info->pinned_extents);
	}
2223
}
C
Chris Mason 已提交
2224

2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235
static void btrfs_init_scrub(struct btrfs_fs_info *fs_info)
{
	mutex_init(&fs_info->scrub_lock);
	atomic_set(&fs_info->scrubs_running, 0);
	atomic_set(&fs_info->scrub_pause_req, 0);
	atomic_set(&fs_info->scrubs_paused, 0);
	atomic_set(&fs_info->scrub_cancel_req, 0);
	init_waitqueue_head(&fs_info->scrub_pause_wait);
	fs_info->scrub_workers_refcnt = 0;
}

2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246
static void btrfs_init_balance(struct btrfs_fs_info *fs_info)
{
	spin_lock_init(&fs_info->balance_lock);
	mutex_init(&fs_info->balance_mutex);
	atomic_set(&fs_info->balance_running, 0);
	atomic_set(&fs_info->balance_pause_req, 0);
	atomic_set(&fs_info->balance_cancel_req, 0);
	fs_info->balance_ctl = NULL;
	init_waitqueue_head(&fs_info->balance_wait_q);
}

2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275
static void btrfs_init_btree_inode(struct btrfs_fs_info *fs_info,
				   struct btrfs_root *tree_root)
{
	fs_info->btree_inode->i_ino = BTRFS_BTREE_INODE_OBJECTID;
	set_nlink(fs_info->btree_inode, 1);
	/*
	 * we set the i_size on the btree inode to the max possible int.
	 * the real end of the address space is determined by all of
	 * the devices in the system
	 */
	fs_info->btree_inode->i_size = OFFSET_MAX;
	fs_info->btree_inode->i_mapping->a_ops = &btree_aops;

	RB_CLEAR_NODE(&BTRFS_I(fs_info->btree_inode)->rb_node);
	extent_io_tree_init(&BTRFS_I(fs_info->btree_inode)->io_tree,
			     fs_info->btree_inode->i_mapping);
	BTRFS_I(fs_info->btree_inode)->io_tree.track_uptodate = 0;
	extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree);

	BTRFS_I(fs_info->btree_inode)->io_tree.ops = &btree_extent_io_ops;

	BTRFS_I(fs_info->btree_inode)->root = tree_root;
	memset(&BTRFS_I(fs_info->btree_inode)->location, 0,
	       sizeof(struct btrfs_key));
	set_bit(BTRFS_INODE_DUMMY,
		&BTRFS_I(fs_info->btree_inode)->runtime_flags);
	btrfs_insert_inode_hash(fs_info->btree_inode);
}

2276 2277 2278 2279 2280
static void btrfs_init_dev_replace_locks(struct btrfs_fs_info *fs_info)
{
	fs_info->dev_replace.lock_owner = 0;
	atomic_set(&fs_info->dev_replace.nesting_level, 0);
	mutex_init(&fs_info->dev_replace.lock_finishing_cancel_unmount);
2281 2282 2283
	rwlock_init(&fs_info->dev_replace.lock);
	atomic_set(&fs_info->dev_replace.read_locks, 0);
	atomic_set(&fs_info->dev_replace.blocking_readers, 0);
2284
	init_waitqueue_head(&fs_info->replace_wait);
2285
	init_waitqueue_head(&fs_info->dev_replace.read_lock_wq);
2286 2287
}

2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301
static void btrfs_init_qgroup(struct btrfs_fs_info *fs_info)
{
	spin_lock_init(&fs_info->qgroup_lock);
	mutex_init(&fs_info->qgroup_ioctl_lock);
	fs_info->qgroup_tree = RB_ROOT;
	fs_info->qgroup_op_tree = RB_ROOT;
	INIT_LIST_HEAD(&fs_info->dirty_qgroups);
	fs_info->qgroup_seq = 1;
	fs_info->quota_enabled = 0;
	fs_info->pending_quota_state = 0;
	fs_info->qgroup_ulist = NULL;
	mutex_init(&fs_info->qgroup_rescan_lock);
}

2302 2303 2304 2305
static int btrfs_init_workqueues(struct btrfs_fs_info *fs_info,
		struct btrfs_fs_devices *fs_devices)
{
	int max_active = fs_info->thread_pool_size;
2306
	unsigned int flags = WQ_MEM_RECLAIM | WQ_FREEZABLE | WQ_UNBOUND;
2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381

	fs_info->workers =
		btrfs_alloc_workqueue("worker", flags | WQ_HIGHPRI,
				      max_active, 16);

	fs_info->delalloc_workers =
		btrfs_alloc_workqueue("delalloc", flags, max_active, 2);

	fs_info->flush_workers =
		btrfs_alloc_workqueue("flush_delalloc", flags, max_active, 0);

	fs_info->caching_workers =
		btrfs_alloc_workqueue("cache", flags, max_active, 0);

	/*
	 * a higher idle thresh on the submit workers makes it much more
	 * likely that bios will be send down in a sane order to the
	 * devices
	 */
	fs_info->submit_workers =
		btrfs_alloc_workqueue("submit", flags,
				      min_t(u64, fs_devices->num_devices,
					    max_active), 64);

	fs_info->fixup_workers =
		btrfs_alloc_workqueue("fixup", flags, 1, 0);

	/*
	 * endios are largely parallel and should have a very
	 * low idle thresh
	 */
	fs_info->endio_workers =
		btrfs_alloc_workqueue("endio", flags, max_active, 4);
	fs_info->endio_meta_workers =
		btrfs_alloc_workqueue("endio-meta", flags, max_active, 4);
	fs_info->endio_meta_write_workers =
		btrfs_alloc_workqueue("endio-meta-write", flags, max_active, 2);
	fs_info->endio_raid56_workers =
		btrfs_alloc_workqueue("endio-raid56", flags, max_active, 4);
	fs_info->endio_repair_workers =
		btrfs_alloc_workqueue("endio-repair", flags, 1, 0);
	fs_info->rmw_workers =
		btrfs_alloc_workqueue("rmw", flags, max_active, 2);
	fs_info->endio_write_workers =
		btrfs_alloc_workqueue("endio-write", flags, max_active, 2);
	fs_info->endio_freespace_worker =
		btrfs_alloc_workqueue("freespace-write", flags, max_active, 0);
	fs_info->delayed_workers =
		btrfs_alloc_workqueue("delayed-meta", flags, max_active, 0);
	fs_info->readahead_workers =
		btrfs_alloc_workqueue("readahead", flags, max_active, 2);
	fs_info->qgroup_rescan_workers =
		btrfs_alloc_workqueue("qgroup-rescan", flags, 1, 0);
	fs_info->extent_workers =
		btrfs_alloc_workqueue("extent-refs", flags,
				      min_t(u64, fs_devices->num_devices,
					    max_active), 8);

	if (!(fs_info->workers && fs_info->delalloc_workers &&
	      fs_info->submit_workers && fs_info->flush_workers &&
	      fs_info->endio_workers && fs_info->endio_meta_workers &&
	      fs_info->endio_meta_write_workers &&
	      fs_info->endio_repair_workers &&
	      fs_info->endio_write_workers && fs_info->endio_raid56_workers &&
	      fs_info->endio_freespace_worker && fs_info->rmw_workers &&
	      fs_info->caching_workers && fs_info->readahead_workers &&
	      fs_info->fixup_workers && fs_info->delayed_workers &&
	      fs_info->extent_workers &&
	      fs_info->qgroup_rescan_workers)) {
		return -ENOMEM;
	}

	return 0;
}

2382 2383 2384 2385 2386 2387 2388 2389 2390 2391
static int btrfs_replay_log(struct btrfs_fs_info *fs_info,
			    struct btrfs_fs_devices *fs_devices)
{
	int ret;
	struct btrfs_root *tree_root = fs_info->tree_root;
	struct btrfs_root *log_tree_root;
	struct btrfs_super_block *disk_super = fs_info->super_copy;
	u64 bytenr = btrfs_super_log_root(disk_super);

	if (fs_devices->rw_devices == 0) {
2392
		btrfs_warn(fs_info, "log replay required on RO media");
2393 2394 2395
		return -EIO;
	}

2396
	log_tree_root = btrfs_alloc_root(fs_info, GFP_KERNEL);
2397 2398 2399 2400 2401 2402 2403 2404 2405
	if (!log_tree_root)
		return -ENOMEM;

	__setup_root(tree_root->nodesize, tree_root->sectorsize,
			tree_root->stripesize, log_tree_root, fs_info,
			BTRFS_TREE_LOG_OBJECTID);

	log_tree_root->node = read_tree_block(tree_root, bytenr,
			fs_info->generation + 1);
2406
	if (IS_ERR(log_tree_root->node)) {
2407
		btrfs_warn(fs_info, "failed to read log tree");
2408
		ret = PTR_ERR(log_tree_root->node);
2409
		kfree(log_tree_root);
2410
		return ret;
2411
	} else if (!extent_buffer_uptodate(log_tree_root->node)) {
2412
		btrfs_err(fs_info, "failed to read log tree");
2413 2414 2415 2416 2417 2418 2419
		free_extent_buffer(log_tree_root->node);
		kfree(log_tree_root);
		return -EIO;
	}
	/* returns with log_tree_root freed on success */
	ret = btrfs_recover_log_trees(log_tree_root);
	if (ret) {
2420
		btrfs_handle_fs_error(tree_root->fs_info, ret,
2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435
			    "Failed to recover log tree");
		free_extent_buffer(log_tree_root->node);
		kfree(log_tree_root);
		return ret;
	}

	if (fs_info->sb->s_flags & MS_RDONLY) {
		ret = btrfs_commit_super(tree_root);
		if (ret)
			return ret;
	}

	return 0;
}

2436 2437 2438
static int btrfs_read_roots(struct btrfs_fs_info *fs_info,
			    struct btrfs_root *tree_root)
{
2439
	struct btrfs_root *root;
2440 2441 2442 2443 2444 2445 2446
	struct btrfs_key location;
	int ret;

	location.objectid = BTRFS_EXTENT_TREE_OBJECTID;
	location.type = BTRFS_ROOT_ITEM_KEY;
	location.offset = 0;

2447 2448 2449 2450 2451
	root = btrfs_read_tree_root(tree_root, &location);
	if (IS_ERR(root))
		return PTR_ERR(root);
	set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
	fs_info->extent_root = root;
2452 2453

	location.objectid = BTRFS_DEV_TREE_OBJECTID;
2454 2455 2456 2457 2458
	root = btrfs_read_tree_root(tree_root, &location);
	if (IS_ERR(root))
		return PTR_ERR(root);
	set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
	fs_info->dev_root = root;
2459 2460 2461
	btrfs_init_devices_late(fs_info);

	location.objectid = BTRFS_CSUM_TREE_OBJECTID;
2462 2463 2464 2465 2466
	root = btrfs_read_tree_root(tree_root, &location);
	if (IS_ERR(root))
		return PTR_ERR(root);
	set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
	fs_info->csum_root = root;
2467 2468

	location.objectid = BTRFS_QUOTA_TREE_OBJECTID;
2469 2470 2471
	root = btrfs_read_tree_root(tree_root, &location);
	if (!IS_ERR(root)) {
		set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
2472 2473
		fs_info->quota_enabled = 1;
		fs_info->pending_quota_state = 1;
2474
		fs_info->quota_root = root;
2475 2476 2477
	}

	location.objectid = BTRFS_UUID_TREE_OBJECTID;
2478 2479 2480
	root = btrfs_read_tree_root(tree_root, &location);
	if (IS_ERR(root)) {
		ret = PTR_ERR(root);
2481 2482 2483
		if (ret != -ENOENT)
			return ret;
	} else {
2484 2485
		set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
		fs_info->uuid_root = root;
2486 2487
	}

2488 2489 2490 2491 2492 2493 2494 2495 2496
	if (btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE)) {
		location.objectid = BTRFS_FREE_SPACE_TREE_OBJECTID;
		root = btrfs_read_tree_root(tree_root, &location);
		if (IS_ERR(root))
			return PTR_ERR(root);
		set_bit(BTRFS_ROOT_TRACK_DIRTY, &root->state);
		fs_info->free_space_root = root;
	}

2497 2498 2499
	return 0;
}

A
Al Viro 已提交
2500 2501 2502
int open_ctree(struct super_block *sb,
	       struct btrfs_fs_devices *fs_devices,
	       char *options)
2503
{
2504 2505
	u32 sectorsize;
	u32 nodesize;
2506
	u32 stripesize;
2507
	u64 generation;
2508
	u64 features;
2509
	struct btrfs_key location;
2510
	struct buffer_head *bh;
2511
	struct btrfs_super_block *disk_super;
2512
	struct btrfs_fs_info *fs_info = btrfs_sb(sb);
2513
	struct btrfs_root *tree_root;
2514
	struct btrfs_root *chunk_root;
2515
	int ret;
2516
	int err = -EINVAL;
C
Chris Mason 已提交
2517 2518
	int num_backups_tried = 0;
	int backup_index = 0;
2519
	int max_active;
2520

2521 2522
	tree_root = fs_info->tree_root = btrfs_alloc_root(fs_info, GFP_KERNEL);
	chunk_root = fs_info->chunk_root = btrfs_alloc_root(fs_info, GFP_KERNEL);
2523
	if (!tree_root || !chunk_root) {
C
Chris Mason 已提交
2524 2525 2526
		err = -ENOMEM;
		goto fail;
	}
2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539

	ret = init_srcu_struct(&fs_info->subvol_srcu);
	if (ret) {
		err = ret;
		goto fail;
	}

	ret = setup_bdi(fs_info, &fs_info->bdi);
	if (ret) {
		err = ret;
		goto fail_srcu;
	}

2540
	ret = percpu_counter_init(&fs_info->dirty_metadata_bytes, 0, GFP_KERNEL);
2541 2542 2543 2544
	if (ret) {
		err = ret;
		goto fail_bdi;
	}
2545
	fs_info->dirty_metadata_batch = PAGE_SIZE *
2546 2547
					(1 + ilog2(nr_cpu_ids));

2548
	ret = percpu_counter_init(&fs_info->delalloc_bytes, 0, GFP_KERNEL);
2549 2550 2551 2552 2553
	if (ret) {
		err = ret;
		goto fail_dirty_metadata_bytes;
	}

2554
	ret = percpu_counter_init(&fs_info->bio_counter, 0, GFP_KERNEL);
2555 2556 2557 2558 2559
	if (ret) {
		err = ret;
		goto fail_delalloc_bytes;
	}

2560 2561 2562
	fs_info->btree_inode = new_inode(sb);
	if (!fs_info->btree_inode) {
		err = -ENOMEM;
2563
		goto fail_bio_counter;
2564 2565
	}

2566
	mapping_set_gfp_mask(fs_info->btree_inode->i_mapping, GFP_NOFS);
2567

2568
	INIT_RADIX_TREE(&fs_info->fs_roots_radix, GFP_ATOMIC);
2569
	INIT_RADIX_TREE(&fs_info->buffer_radix, GFP_ATOMIC);
C
Chris Mason 已提交
2570
	INIT_LIST_HEAD(&fs_info->trans_list);
2571
	INIT_LIST_HEAD(&fs_info->dead_roots);
Y
Yan, Zheng 已提交
2572
	INIT_LIST_HEAD(&fs_info->delayed_iputs);
2573
	INIT_LIST_HEAD(&fs_info->delalloc_roots);
2574
	INIT_LIST_HEAD(&fs_info->caching_block_groups);
2575
	spin_lock_init(&fs_info->delalloc_root_lock);
J
Josef Bacik 已提交
2576
	spin_lock_init(&fs_info->trans_lock);
2577
	spin_lock_init(&fs_info->fs_roots_radix_lock);
Y
Yan, Zheng 已提交
2578
	spin_lock_init(&fs_info->delayed_iput_lock);
C
Chris Mason 已提交
2579
	spin_lock_init(&fs_info->defrag_inodes_lock);
2580
	spin_lock_init(&fs_info->free_chunk_lock);
J
Jan Schmidt 已提交
2581
	spin_lock_init(&fs_info->tree_mod_seq_lock);
2582
	spin_lock_init(&fs_info->super_lock);
J
Josef Bacik 已提交
2583
	spin_lock_init(&fs_info->qgroup_op_lock);
2584
	spin_lock_init(&fs_info->buffer_lock);
2585
	spin_lock_init(&fs_info->unused_bgs_lock);
J
Jan Schmidt 已提交
2586
	rwlock_init(&fs_info->tree_mod_log_lock);
2587
	mutex_init(&fs_info->unused_bg_unpin_mutex);
2588
	mutex_init(&fs_info->delete_unused_bgs_mutex);
C
Chris Mason 已提交
2589
	mutex_init(&fs_info->reloc_mutex);
2590
	mutex_init(&fs_info->delalloc_root_mutex);
2591
	mutex_init(&fs_info->cleaner_delayed_iput_mutex);
2592
	seqlock_init(&fs_info->profiles_lock);
2593

2594
	INIT_LIST_HEAD(&fs_info->dirty_cowonly_roots);
2595
	INIT_LIST_HEAD(&fs_info->space_info);
J
Jan Schmidt 已提交
2596
	INIT_LIST_HEAD(&fs_info->tree_mod_seq_list);
2597
	INIT_LIST_HEAD(&fs_info->unused_bgs);
2598
	btrfs_mapping_init(&fs_info->mapping_tree);
2599 2600 2601 2602 2603 2604 2605 2606 2607
	btrfs_init_block_rsv(&fs_info->global_block_rsv,
			     BTRFS_BLOCK_RSV_GLOBAL);
	btrfs_init_block_rsv(&fs_info->delalloc_block_rsv,
			     BTRFS_BLOCK_RSV_DELALLOC);
	btrfs_init_block_rsv(&fs_info->trans_block_rsv, BTRFS_BLOCK_RSV_TRANS);
	btrfs_init_block_rsv(&fs_info->chunk_block_rsv, BTRFS_BLOCK_RSV_CHUNK);
	btrfs_init_block_rsv(&fs_info->empty_block_rsv, BTRFS_BLOCK_RSV_EMPTY);
	btrfs_init_block_rsv(&fs_info->delayed_block_rsv,
			     BTRFS_BLOCK_RSV_DELOPS);
2608
	atomic_set(&fs_info->nr_async_submits, 0);
2609
	atomic_set(&fs_info->async_delalloc_pages, 0);
2610
	atomic_set(&fs_info->async_submit_draining, 0);
2611
	atomic_set(&fs_info->nr_async_bios, 0);
C
Chris Mason 已提交
2612
	atomic_set(&fs_info->defrag_running, 0);
J
Josef Bacik 已提交
2613
	atomic_set(&fs_info->qgroup_op_seq, 0);
Z
Zhao Lei 已提交
2614
	atomic_set(&fs_info->reada_works_cnt, 0);
2615
	atomic64_set(&fs_info->tree_mod_seq, 0);
C
Chris Mason 已提交
2616
	fs_info->sb = sb;
2617
	fs_info->max_inline = BTRFS_DEFAULT_MAX_INLINE;
J
Josef Bacik 已提交
2618
	fs_info->metadata_ratio = 0;
C
Chris Mason 已提交
2619
	fs_info->defrag_inodes = RB_ROOT;
2620
	fs_info->free_chunk_space = 0;
J
Jan Schmidt 已提交
2621
	fs_info->tree_mod_log = RB_ROOT;
2622
	fs_info->commit_interval = BTRFS_DEFAULT_COMMIT_INTERVAL;
2623
	fs_info->avg_delayed_ref_runtime = NSEC_PER_SEC >> 6; /* div by 64 */
2624
	/* readahead state */
2625
	INIT_RADIX_TREE(&fs_info->reada_tree, GFP_NOFS & ~__GFP_DIRECT_RECLAIM);
2626
	spin_lock_init(&fs_info->reada_lock);
C
Chris Mason 已提交
2627

2628 2629
	fs_info->thread_pool_size = min_t(unsigned long,
					  num_online_cpus() + 2, 8);
2630

2631 2632
	INIT_LIST_HEAD(&fs_info->ordered_roots);
	spin_lock_init(&fs_info->ordered_root_lock);
2633
	fs_info->delayed_root = kmalloc(sizeof(struct btrfs_delayed_root),
2634
					GFP_KERNEL);
2635 2636 2637 2638 2639
	if (!fs_info->delayed_root) {
		err = -ENOMEM;
		goto fail_iput;
	}
	btrfs_init_delayed_root(fs_info->delayed_root);
2640

2641
	btrfs_init_scrub(fs_info);
2642 2643 2644
#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
	fs_info->check_integrity_print_mask = 0;
#endif
2645
	btrfs_init_balance(fs_info);
2646
	btrfs_init_async_reclaim_work(&fs_info->async_reclaim_work);
A
Arne Jansen 已提交
2647

2648 2649
	sb->s_blocksize = 4096;
	sb->s_blocksize_bits = blksize_bits(4096);
J
Jens Axboe 已提交
2650
	sb->s_bdi = &fs_info->bdi;
2651

2652
	btrfs_init_btree_inode(fs_info, tree_root);
2653

J
Josef Bacik 已提交
2654
	spin_lock_init(&fs_info->block_group_cache_lock);
2655
	fs_info->block_group_cache_tree = RB_ROOT;
2656
	fs_info->first_logical_byte = (u64)-1;
J
Josef Bacik 已提交
2657

2658
	extent_io_tree_init(&fs_info->freed_extents[0],
2659
			     fs_info->btree_inode->i_mapping);
2660
	extent_io_tree_init(&fs_info->freed_extents[1],
2661
			     fs_info->btree_inode->i_mapping);
2662
	fs_info->pinned_extents = &fs_info->freed_extents[0];
2663
	fs_info->do_barriers = 1;
2664

C
Chris Mason 已提交
2665

2666
	mutex_init(&fs_info->ordered_operations_mutex);
2667
	mutex_init(&fs_info->tree_log_mutex);
2668
	mutex_init(&fs_info->chunk_mutex);
2669 2670
	mutex_init(&fs_info->transaction_kthread_mutex);
	mutex_init(&fs_info->cleaner_mutex);
2671
	mutex_init(&fs_info->volume_mutex);
2672
	mutex_init(&fs_info->ro_block_group_mutex);
2673
	init_rwsem(&fs_info->commit_root_sem);
2674
	init_rwsem(&fs_info->cleanup_work_sem);
2675
	init_rwsem(&fs_info->subvol_sem);
S
Stefan Behrens 已提交
2676
	sema_init(&fs_info->uuid_tree_rescan_sem, 1);
2677

2678
	btrfs_init_dev_replace_locks(fs_info);
2679
	btrfs_init_qgroup(fs_info);
2680

2681 2682 2683
	btrfs_init_free_cluster(&fs_info->meta_alloc_cluster);
	btrfs_init_free_cluster(&fs_info->data_alloc_cluster);

2684
	init_waitqueue_head(&fs_info->transaction_throttle);
2685
	init_waitqueue_head(&fs_info->transaction_wait);
S
Sage Weil 已提交
2686
	init_waitqueue_head(&fs_info->transaction_blocked_wait);
2687
	init_waitqueue_head(&fs_info->async_submit_wait);
2688

2689 2690
	INIT_LIST_HEAD(&fs_info->pinned_chunks);

D
David Woodhouse 已提交
2691 2692
	ret = btrfs_alloc_stripe_hash_table(fs_info);
	if (ret) {
2693
		err = ret;
D
David Woodhouse 已提交
2694 2695 2696
		goto fail_alloc;
	}

2697
	__setup_root(4096, 4096, 4096, tree_root,
C
Chris Mason 已提交
2698
		     fs_info, BTRFS_ROOT_TREE_OBJECTID);
2699

2700
	invalidate_bdev(fs_devices->latest_bdev);
D
David Sterba 已提交
2701 2702 2703 2704

	/*
	 * Read super block and check the signature bytes only
	 */
Y
Yan Zheng 已提交
2705
	bh = btrfs_read_dev_super(fs_devices->latest_bdev);
2706 2707
	if (IS_ERR(bh)) {
		err = PTR_ERR(bh);
2708
		goto fail_alloc;
2709
	}
C
Chris Mason 已提交
2710

D
David Sterba 已提交
2711 2712 2713 2714 2715
	/*
	 * We want to check superblock checksum, the type is stored inside.
	 * Pass the whole disk block of size BTRFS_SUPER_INFO_SIZE (4k).
	 */
	if (btrfs_check_super_csum(bh->b_data)) {
2716
		printk(KERN_ERR "BTRFS: superblock checksum mismatch\n");
D
David Sterba 已提交
2717
		err = -EINVAL;
2718
		brelse(bh);
D
David Sterba 已提交
2719 2720 2721 2722 2723 2724 2725 2726
		goto fail_alloc;
	}

	/*
	 * super_copy is zeroed at allocation time and we never touch the
	 * following bytes up to INFO_SIZE, the checksum is calculated from
	 * the whole block of INFO_SIZE
	 */
2727 2728 2729
	memcpy(fs_info->super_copy, bh->b_data, sizeof(*fs_info->super_copy));
	memcpy(fs_info->super_for_commit, fs_info->super_copy,
	       sizeof(*fs_info->super_for_commit));
2730
	brelse(bh);
2731

2732
	memcpy(fs_info->fsid, fs_info->super_copy->fsid, BTRFS_FSID_SIZE);
2733

D
David Sterba 已提交
2734 2735
	ret = btrfs_check_super_valid(fs_info, sb->s_flags & MS_RDONLY);
	if (ret) {
2736
		printk(KERN_ERR "BTRFS: superblock contains fatal errors\n");
D
David Sterba 已提交
2737 2738 2739 2740
		err = -EINVAL;
		goto fail_alloc;
	}

2741
	disk_super = fs_info->super_copy;
2742
	if (!btrfs_super_root(disk_super))
2743
		goto fail_alloc;
2744

L
liubo 已提交
2745
	/* check FS state, whether FS is broken. */
2746 2747
	if (btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_ERROR)
		set_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state);
L
liubo 已提交
2748

C
Chris Mason 已提交
2749 2750 2751 2752 2753 2754 2755
	/*
	 * run through our array of backup supers and setup
	 * our ring pointer to the oldest one
	 */
	generation = btrfs_super_generation(disk_super);
	find_oldest_super_backup(fs_info, generation);

2756 2757 2758 2759 2760 2761
	/*
	 * In the long term, we'll store the compression type in the super
	 * block, and it'll be used for per file compression control.
	 */
	fs_info->compress_type = BTRFS_COMPRESS_ZLIB;

2762
	ret = btrfs_parse_options(tree_root, options, sb->s_flags);
Y
Yan Zheng 已提交
2763 2764
	if (ret) {
		err = ret;
2765
		goto fail_alloc;
Y
Yan Zheng 已提交
2766
	}
2767

2768 2769 2770 2771 2772
	features = btrfs_super_incompat_flags(disk_super) &
		~BTRFS_FEATURE_INCOMPAT_SUPP;
	if (features) {
		printk(KERN_ERR "BTRFS: couldn't mount because of "
		       "unsupported optional features (%Lx).\n",
2773
		       features);
2774
		err = -EINVAL;
2775
		goto fail_alloc;
2776 2777
	}

2778
	features = btrfs_super_incompat_flags(disk_super);
L
Li Zefan 已提交
2779
	features |= BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF;
L
Li Zefan 已提交
2780
	if (tree_root->fs_info->compress_type == BTRFS_COMPRESS_LZO)
L
Li Zefan 已提交
2781
		features |= BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO;
2782

2783
	if (features & BTRFS_FEATURE_INCOMPAT_SKINNY_METADATA)
2784
		printk(KERN_INFO "BTRFS: has skinny extents\n");
2785

2786 2787 2788 2789
	/*
	 * flag our filesystem as having big metadata blocks if
	 * they are bigger than the page size
	 */
2790
	if (btrfs_super_nodesize(disk_super) > PAGE_SIZE) {
2791
		if (!(features & BTRFS_FEATURE_INCOMPAT_BIG_METADATA))
2792
			printk(KERN_INFO "BTRFS: flagging fs with big metadata feature\n");
2793 2794 2795
		features |= BTRFS_FEATURE_INCOMPAT_BIG_METADATA;
	}

2796 2797 2798
	nodesize = btrfs_super_nodesize(disk_super);
	sectorsize = btrfs_super_sectorsize(disk_super);
	stripesize = btrfs_super_stripesize(disk_super);
2799
	fs_info->dirty_metadata_batch = nodesize * (1 + ilog2(nr_cpu_ids));
2800
	fs_info->delalloc_batch = sectorsize * 512 * (1 + ilog2(nr_cpu_ids));
2801 2802 2803 2804 2805 2806

	/*
	 * mixed block groups end up with duplicate but slightly offset
	 * extent buffers for the same range.  It leads to corruptions
	 */
	if ((features & BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS) &&
2807
	    (sectorsize != nodesize)) {
2808
		printk(KERN_ERR "BTRFS: unequal leaf/node/sector sizes "
2809 2810 2811 2812 2813
				"are not allowed for mixed block groups on %s\n",
				sb->s_id);
		goto fail_alloc;
	}

2814 2815 2816 2817
	/*
	 * Needn't use the lock because there is no other task which will
	 * update the flag.
	 */
L
Li Zefan 已提交
2818
	btrfs_set_super_incompat_flags(disk_super, features);
2819

2820 2821 2822 2823 2824
	features = btrfs_super_compat_ro_flags(disk_super) &
		~BTRFS_FEATURE_COMPAT_RO_SUPP;
	if (!(sb->s_flags & MS_RDONLY) && features) {
		printk(KERN_ERR "BTRFS: couldn't mount RDWR because of "
		       "unsupported option features (%Lx).\n",
2825
		       features);
2826
		err = -EINVAL;
2827
		goto fail_alloc;
2828
	}
2829

2830
	max_active = fs_info->thread_pool_size;
2831

2832 2833 2834
	ret = btrfs_init_workqueues(fs_info, fs_devices);
	if (ret) {
		err = ret;
2835 2836
		goto fail_sb_buffer;
	}
2837

2838
	fs_info->bdi.ra_pages *= btrfs_super_num_devices(disk_super);
C
Chris Mason 已提交
2839
	fs_info->bdi.ra_pages = max(fs_info->bdi.ra_pages,
2840
				    SZ_4M / PAGE_SIZE);
2841

2842 2843
	tree_root->nodesize = nodesize;
	tree_root->sectorsize = sectorsize;
2844
	tree_root->stripesize = stripesize;
2845 2846 2847

	sb->s_blocksize = sectorsize;
	sb->s_blocksize_bits = blksize_bits(sectorsize);
2848

2849
	mutex_lock(&fs_info->chunk_mutex);
Y
Yan Zheng 已提交
2850
	ret = btrfs_read_sys_array(tree_root);
2851
	mutex_unlock(&fs_info->chunk_mutex);
2852
	if (ret) {
2853
		printk(KERN_ERR "BTRFS: failed to read the system "
C
Chris Mason 已提交
2854
		       "array on %s\n", sb->s_id);
2855
		goto fail_sb_buffer;
2856
	}
2857

2858
	generation = btrfs_super_chunk_root_generation(disk_super);
2859

2860 2861
	__setup_root(nodesize, sectorsize, stripesize, chunk_root,
		     fs_info, BTRFS_CHUNK_TREE_OBJECTID);
2862 2863 2864

	chunk_root->node = read_tree_block(chunk_root,
					   btrfs_super_chunk_root(disk_super),
2865
					   generation);
2866 2867
	if (IS_ERR(chunk_root->node) ||
	    !extent_buffer_uptodate(chunk_root->node)) {
2868
		printk(KERN_ERR "BTRFS: failed to read chunk root on %s\n",
2869
		       sb->s_id);
2870 2871
		if (!IS_ERR(chunk_root->node))
			free_extent_buffer(chunk_root->node);
2872
		chunk_root->node = NULL;
C
Chris Mason 已提交
2873
		goto fail_tree_roots;
2874
	}
2875 2876
	btrfs_set_root_node(&chunk_root->root_item, chunk_root->node);
	chunk_root->commit_root = btrfs_root_node(chunk_root);
2877

2878
	read_extent_buffer(chunk_root->node, fs_info->chunk_tree_uuid,
2879
	   btrfs_header_chunk_tree_uuid(chunk_root->node), BTRFS_UUID_SIZE);
2880

2881
	ret = btrfs_read_chunk_tree(chunk_root);
Y
Yan Zheng 已提交
2882
	if (ret) {
2883
		printk(KERN_ERR "BTRFS: failed to read chunk tree on %s\n",
C
Chris Mason 已提交
2884
		       sb->s_id);
C
Chris Mason 已提交
2885
		goto fail_tree_roots;
Y
Yan Zheng 已提交
2886
	}
2887

2888 2889 2890 2891
	/*
	 * keep the device that is marked to be the target device for the
	 * dev_replace procedure
	 */
2892
	btrfs_close_extra_devices(fs_devices, 0);
2893

2894
	if (!fs_devices->latest_bdev) {
2895
		printk(KERN_ERR "BTRFS: failed to read devices on %s\n",
2896 2897 2898 2899
		       sb->s_id);
		goto fail_tree_roots;
	}

C
Chris Mason 已提交
2900
retry_root_backup:
2901
	generation = btrfs_super_generation(disk_super);
2902

C
Chris Mason 已提交
2903
	tree_root->node = read_tree_block(tree_root,
2904
					  btrfs_super_root(disk_super),
2905
					  generation);
2906 2907
	if (IS_ERR(tree_root->node) ||
	    !extent_buffer_uptodate(tree_root->node)) {
2908
		printk(KERN_WARNING "BTRFS: failed to read tree root on %s\n",
2909
		       sb->s_id);
2910 2911
		if (!IS_ERR(tree_root->node))
			free_extent_buffer(tree_root->node);
2912
		tree_root->node = NULL;
C
Chris Mason 已提交
2913
		goto recovery_tree_root;
2914
	}
C
Chris Mason 已提交
2915

2916 2917
	btrfs_set_root_node(&tree_root->root_item, tree_root->node);
	tree_root->commit_root = btrfs_root_node(tree_root);
2918
	btrfs_set_root_refs(&tree_root->root_item, 1);
2919

2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931
	mutex_lock(&tree_root->objectid_mutex);
	ret = btrfs_find_highest_objectid(tree_root,
					&tree_root->highest_objectid);
	if (ret) {
		mutex_unlock(&tree_root->objectid_mutex);
		goto recovery_tree_root;
	}

	ASSERT(tree_root->highest_objectid <= BTRFS_LAST_FREE_OBJECTID);

	mutex_unlock(&tree_root->objectid_mutex);

2932 2933
	ret = btrfs_read_roots(fs_info, tree_root);
	if (ret)
C
Chris Mason 已提交
2934
		goto recovery_tree_root;
2935

2936 2937 2938
	fs_info->generation = generation;
	fs_info->last_trans_committed = generation;

2939 2940
	ret = btrfs_recover_balance(fs_info);
	if (ret) {
2941
		printk(KERN_ERR "BTRFS: failed to recover balance\n");
2942 2943 2944
		goto fail_block_groups;
	}

2945 2946
	ret = btrfs_init_dev_stats(fs_info);
	if (ret) {
2947
		printk(KERN_ERR "BTRFS: failed to init dev_stats: %d\n",
2948 2949 2950 2951
		       ret);
		goto fail_block_groups;
	}

2952 2953
	ret = btrfs_init_dev_replace(fs_info);
	if (ret) {
2954
		pr_err("BTRFS: failed to init dev_replace: %d\n", ret);
2955 2956 2957
		goto fail_block_groups;
	}

2958
	btrfs_close_extra_devices(fs_devices, 1);
2959

2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971
	ret = btrfs_sysfs_add_fsid(fs_devices, NULL);
	if (ret) {
		pr_err("BTRFS: failed to init sysfs fsid interface: %d\n", ret);
		goto fail_block_groups;
	}

	ret = btrfs_sysfs_add_device(fs_devices);
	if (ret) {
		pr_err("BTRFS: failed to init sysfs device interface: %d\n", ret);
		goto fail_fsdev_sysfs;
	}

2972
	ret = btrfs_sysfs_add_mounted(fs_info);
2973
	if (ret) {
2974
		pr_err("BTRFS: failed to init sysfs interface: %d\n", ret);
2975
		goto fail_fsdev_sysfs;
2976 2977 2978 2979
	}

	ret = btrfs_init_space_info(fs_info);
	if (ret) {
2980
		printk(KERN_ERR "BTRFS: Failed to initial space info: %d\n", ret);
2981
		goto fail_sysfs;
2982 2983
	}

2984
	ret = btrfs_read_block_groups(fs_info->extent_root);
2985
	if (ret) {
2986
		printk(KERN_ERR "BTRFS: Failed to read block groups: %d\n", ret);
2987
		goto fail_sysfs;
2988
	}
2989 2990
	fs_info->num_tolerated_disk_barrier_failures =
		btrfs_calc_num_tolerated_disk_barrier_failures(fs_info);
2991 2992 2993
	if (fs_info->fs_devices->missing_devices >
	     fs_info->num_tolerated_disk_barrier_failures &&
	    !(sb->s_flags & MS_RDONLY)) {
2994 2995 2996
		pr_warn("BTRFS: missing devices(%llu) exceeds the limit(%d), writeable mount is not allowed\n",
			fs_info->fs_devices->missing_devices,
			fs_info->num_tolerated_disk_barrier_failures);
2997
		goto fail_sysfs;
2998
	}
C
Chris Mason 已提交
2999

3000 3001
	fs_info->cleaner_kthread = kthread_run(cleaner_kthread, tree_root,
					       "btrfs-cleaner");
3002
	if (IS_ERR(fs_info->cleaner_kthread))
3003
		goto fail_sysfs;
3004 3005 3006 3007

	fs_info->transaction_kthread = kthread_run(transaction_kthread,
						   tree_root,
						   "btrfs-transaction");
3008
	if (IS_ERR(fs_info->transaction_kthread))
3009
		goto fail_cleaner;
3010

C
Chris Mason 已提交
3011 3012 3013
	if (!btrfs_test_opt(tree_root, SSD) &&
	    !btrfs_test_opt(tree_root, NOSSD) &&
	    !fs_info->fs_devices->rotating) {
3014
		printk(KERN_INFO "BTRFS: detected SSD devices, enabling SSD "
C
Chris Mason 已提交
3015 3016 3017 3018
		       "mode\n");
		btrfs_set_opt(fs_info->mount_opt, SSD);
	}

3019 3020 3021 3022 3023
	/*
	 * Mount does not set all options immediatelly, we can do it now and do
	 * not have to wait for transaction commit
	 */
	btrfs_apply_pending_changes(fs_info);
3024

3025 3026 3027 3028 3029 3030 3031 3032
#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
	if (btrfs_test_opt(tree_root, CHECK_INTEGRITY)) {
		ret = btrfsic_mount(tree_root, fs_devices,
				    btrfs_test_opt(tree_root,
					CHECK_INTEGRITY_INCLUDING_EXTENT_DATA) ?
				    1 : 0,
				    fs_info->check_integrity_print_mask);
		if (ret)
3033
			printk(KERN_WARNING "BTRFS: failed to initialize"
3034 3035 3036
			       " integrity check module %s\n", sb->s_id);
	}
#endif
3037 3038 3039
	ret = btrfs_read_qgroup_config(fs_info);
	if (ret)
		goto fail_trans_kthread;
3040

3041 3042 3043
	/* do not make disk changes in broken FS or nologreplay is given */
	if (btrfs_super_log_root(disk_super) != 0 &&
	    !btrfs_test_opt(tree_root, NOLOGREPLAY)) {
3044
		ret = btrfs_replay_log(fs_info, fs_devices);
3045
		if (ret) {
3046
			err = ret;
3047
			goto fail_qgroup;
3048
		}
3049
	}
Z
Zheng Yan 已提交
3050

3051
	ret = btrfs_find_orphan_roots(tree_root);
3052
	if (ret)
3053
		goto fail_qgroup;
3054

3055
	if (!(sb->s_flags & MS_RDONLY)) {
3056
		ret = btrfs_cleanup_fs_roots(fs_info);
3057
		if (ret)
3058
			goto fail_qgroup;
3059

3060
		mutex_lock(&fs_info->cleaner_mutex);
3061
		ret = btrfs_recover_relocation(tree_root);
3062
		mutex_unlock(&fs_info->cleaner_mutex);
3063 3064
		if (ret < 0) {
			printk(KERN_WARNING
3065
			       "BTRFS: failed to recover relocation\n");
3066
			err = -EINVAL;
3067
			goto fail_qgroup;
3068
		}
3069
	}
Z
Zheng Yan 已提交
3070

3071 3072
	location.objectid = BTRFS_FS_TREE_OBJECTID;
	location.type = BTRFS_ROOT_ITEM_KEY;
3073
	location.offset = 0;
3074 3075

	fs_info->fs_root = btrfs_read_fs_root_no_name(fs_info, &location);
3076 3077
	if (IS_ERR(fs_info->fs_root)) {
		err = PTR_ERR(fs_info->fs_root);
3078
		goto fail_qgroup;
3079
	}
C
Chris Mason 已提交
3080

3081 3082
	if (sb->s_flags & MS_RDONLY)
		return 0;
I
Ilya Dryomov 已提交
3083

3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095
	if (btrfs_test_opt(tree_root, FREE_SPACE_TREE) &&
	    !btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE)) {
		pr_info("BTRFS: creating free space tree\n");
		ret = btrfs_create_free_space_tree(fs_info);
		if (ret) {
			pr_warn("BTRFS: failed to create free space tree %d\n",
				ret);
			close_ctree(tree_root);
			return ret;
		}
	}

3096 3097 3098
	down_read(&fs_info->cleanup_work_sem);
	if ((ret = btrfs_orphan_cleanup(fs_info->fs_root)) ||
	    (ret = btrfs_orphan_cleanup(fs_info->tree_root))) {
3099
		up_read(&fs_info->cleanup_work_sem);
3100 3101 3102 3103
		close_ctree(tree_root);
		return ret;
	}
	up_read(&fs_info->cleanup_work_sem);
I
Ilya Dryomov 已提交
3104

3105 3106
	ret = btrfs_resume_balance_async(fs_info);
	if (ret) {
3107
		printk(KERN_WARNING "BTRFS: failed to resume balance\n");
3108 3109
		close_ctree(tree_root);
		return ret;
3110 3111
	}

3112 3113
	ret = btrfs_resume_dev_replace_async(fs_info);
	if (ret) {
3114
		pr_warn("BTRFS: failed to resume dev_replace\n");
3115 3116 3117 3118
		close_ctree(tree_root);
		return ret;
	}

3119 3120
	btrfs_qgroup_rescan_resume(fs_info);

3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132
	if (btrfs_test_opt(tree_root, CLEAR_CACHE) &&
	    btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE)) {
		pr_info("BTRFS: clearing free space tree\n");
		ret = btrfs_clear_free_space_tree(fs_info);
		if (ret) {
			pr_warn("BTRFS: failed to clear free space tree %d\n",
				ret);
			close_ctree(tree_root);
			return ret;
		}
	}

3133
	if (!fs_info->uuid_root) {
3134
		pr_info("BTRFS: creating UUID tree\n");
3135 3136
		ret = btrfs_create_uuid_tree(fs_info);
		if (ret) {
3137
			pr_warn("BTRFS: failed to create the UUID tree %d\n",
3138 3139 3140 3141
				ret);
			close_ctree(tree_root);
			return ret;
		}
3142 3143 3144
	} else if (btrfs_test_opt(tree_root, RESCAN_UUID_TREE) ||
		   fs_info->generation !=
				btrfs_super_uuid_tree_generation(disk_super)) {
3145
		pr_info("BTRFS: checking UUID tree\n");
3146 3147
		ret = btrfs_check_uuid_tree(fs_info);
		if (ret) {
3148
			pr_warn("BTRFS: failed to check the UUID tree %d\n",
3149 3150 3151 3152 3153 3154
				ret);
			close_ctree(tree_root);
			return ret;
		}
	} else {
		fs_info->update_uuid_tree_gen = 1;
3155 3156
	}

3157 3158
	fs_info->open = 1;

3159 3160 3161 3162 3163 3164
	/*
	 * backuproot only affect mount behavior, and if open_ctree succeeded,
	 * no need to keep the flag
	 */
	btrfs_clear_opt(fs_info->mount_opt, USEBACKUPROOT);

A
Al Viro 已提交
3165
	return 0;
C
Chris Mason 已提交
3166

3167 3168
fail_qgroup:
	btrfs_free_qgroup_config(fs_info);
3169 3170
fail_trans_kthread:
	kthread_stop(fs_info->transaction_kthread);
J
Josef Bacik 已提交
3171
	btrfs_cleanup_transaction(fs_info->tree_root);
3172
	btrfs_free_fs_roots(fs_info);
3173
fail_cleaner:
3174
	kthread_stop(fs_info->cleaner_kthread);
3175 3176 3177 3178 3179 3180 3181

	/*
	 * make sure we're done with the btree inode before we stop our
	 * kthreads
	 */
	filemap_write_and_wait(fs_info->btree_inode->i_mapping);

3182
fail_sysfs:
3183
	btrfs_sysfs_remove_mounted(fs_info);
3184

3185 3186 3187
fail_fsdev_sysfs:
	btrfs_sysfs_remove_fsid(fs_info->fs_devices);

3188
fail_block_groups:
J
Josef Bacik 已提交
3189
	btrfs_put_block_group_cache(fs_info);
3190
	btrfs_free_block_groups(fs_info);
C
Chris Mason 已提交
3191 3192 3193

fail_tree_roots:
	free_root_pointers(fs_info, 1);
3194
	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
C
Chris Mason 已提交
3195

C
Chris Mason 已提交
3196
fail_sb_buffer:
L
Liu Bo 已提交
3197
	btrfs_stop_all_workers(fs_info);
3198
fail_alloc:
3199
fail_iput:
3200 3201
	btrfs_mapping_tree_free(&fs_info->mapping_tree);

3202
	iput(fs_info->btree_inode);
3203 3204
fail_bio_counter:
	percpu_counter_destroy(&fs_info->bio_counter);
3205 3206
fail_delalloc_bytes:
	percpu_counter_destroy(&fs_info->delalloc_bytes);
3207 3208
fail_dirty_metadata_bytes:
	percpu_counter_destroy(&fs_info->dirty_metadata_bytes);
3209
fail_bdi:
3210
	bdi_destroy(&fs_info->bdi);
3211 3212
fail_srcu:
	cleanup_srcu_struct(&fs_info->subvol_srcu);
3213
fail:
D
David Woodhouse 已提交
3214
	btrfs_free_stripe_hash_table(fs_info);
3215
	btrfs_close_devices(fs_info->fs_devices);
A
Al Viro 已提交
3216
	return err;
C
Chris Mason 已提交
3217 3218

recovery_tree_root:
3219
	if (!btrfs_test_opt(tree_root, USEBACKUPROOT))
C
Chris Mason 已提交
3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234
		goto fail_tree_roots;

	free_root_pointers(fs_info, 0);

	/* don't use the log in recovery mode, it won't be valid */
	btrfs_set_super_log_root(disk_super, 0);

	/* we can't trust the free space cache either */
	btrfs_set_opt(fs_info->mount_opt, CLEAR_CACHE);

	ret = next_root_backup(fs_info, fs_info->super_copy,
			       &num_backups_tried, &backup_index);
	if (ret == -1)
		goto fail_block_groups;
	goto retry_root_backup;
3235 3236
}

3237 3238 3239 3240 3241
static void btrfs_end_buffer_write_sync(struct buffer_head *bh, int uptodate)
{
	if (uptodate) {
		set_buffer_uptodate(bh);
	} else {
3242 3243 3244
		struct btrfs_device *device = (struct btrfs_device *)
			bh->b_private;

3245 3246
		btrfs_warn_rl_in_rcu(device->dev_root->fs_info,
				"lost page write due to IO error on %s",
3247
					  rcu_str_deref(device->name));
3248 3249 3250
		/* note, we dont' set_buffer_write_io_error because we have
		 * our own ways of dealing with the IO errors
		 */
3251
		clear_buffer_uptodate(bh);
3252
		btrfs_dev_stat_inc_and_print(device, BTRFS_DEV_STAT_WRITE_ERRS);
3253 3254 3255 3256 3257
	}
	unlock_buffer(bh);
	put_bh(bh);
}

3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288
int btrfs_read_dev_one_super(struct block_device *bdev, int copy_num,
			struct buffer_head **bh_ret)
{
	struct buffer_head *bh;
	struct btrfs_super_block *super;
	u64 bytenr;

	bytenr = btrfs_sb_offset(copy_num);
	if (bytenr + BTRFS_SUPER_INFO_SIZE >= i_size_read(bdev->bd_inode))
		return -EINVAL;

	bh = __bread(bdev, bytenr / 4096, BTRFS_SUPER_INFO_SIZE);
	/*
	 * If we fail to read from the underlying devices, as of now
	 * the best option we have is to mark it EIO.
	 */
	if (!bh)
		return -EIO;

	super = (struct btrfs_super_block *)bh->b_data;
	if (btrfs_super_bytenr(super) != bytenr ||
		    btrfs_super_magic(super) != BTRFS_MAGIC) {
		brelse(bh);
		return -EINVAL;
	}

	*bh_ret = bh;
	return 0;
}


Y
Yan Zheng 已提交
3289 3290 3291 3292 3293 3294 3295
struct buffer_head *btrfs_read_dev_super(struct block_device *bdev)
{
	struct buffer_head *bh;
	struct buffer_head *latest = NULL;
	struct btrfs_super_block *super;
	int i;
	u64 transid = 0;
3296
	int ret = -EINVAL;
Y
Yan Zheng 已提交
3297 3298 3299 3300 3301 3302 3303

	/* we would like to check all the supers, but that would make
	 * a btrfs mount succeed after a mkfs from a different FS.
	 * So, we need to add a special mount option to scan for
	 * later supers, using BTRFS_SUPER_MIRROR_MAX instead
	 */
	for (i = 0; i < 1; i++) {
3304 3305
		ret = btrfs_read_dev_one_super(bdev, i, &bh);
		if (ret)
Y
Yan Zheng 已提交
3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317
			continue;

		super = (struct btrfs_super_block *)bh->b_data;

		if (!latest || btrfs_super_generation(super) > transid) {
			brelse(latest);
			latest = bh;
			transid = btrfs_super_generation(super);
		} else {
			brelse(bh);
		}
	}
3318 3319 3320 3321

	if (!latest)
		return ERR_PTR(ret);

Y
Yan Zheng 已提交
3322 3323 3324
	return latest;
}

3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335
/*
 * this should be called twice, once with wait == 0 and
 * once with wait == 1.  When wait == 0 is done, all the buffer heads
 * we write are pinned.
 *
 * They are released when wait == 1 is done.
 * max_mirrors must be the same for both runs, and it indicates how
 * many supers on this one device should be written.
 *
 * max_mirrors == 0 means to write them all.
 */
Y
Yan Zheng 已提交
3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351
static int write_dev_supers(struct btrfs_device *device,
			    struct btrfs_super_block *sb,
			    int do_barriers, int wait, int max_mirrors)
{
	struct buffer_head *bh;
	int i;
	int ret;
	int errors = 0;
	u32 crc;
	u64 bytenr;

	if (max_mirrors == 0)
		max_mirrors = BTRFS_SUPER_MIRROR_MAX;

	for (i = 0; i < max_mirrors; i++) {
		bytenr = btrfs_sb_offset(i);
3352 3353
		if (bytenr + BTRFS_SUPER_INFO_SIZE >=
		    device->commit_total_bytes)
Y
Yan Zheng 已提交
3354 3355 3356 3357 3358
			break;

		if (wait) {
			bh = __find_get_block(device->bdev, bytenr / 4096,
					      BTRFS_SUPER_INFO_SIZE);
3359 3360 3361 3362
			if (!bh) {
				errors++;
				continue;
			}
Y
Yan Zheng 已提交
3363
			wait_on_buffer(bh);
3364 3365 3366 3367 3368 3369 3370 3371 3372
			if (!buffer_uptodate(bh))
				errors++;

			/* drop our reference */
			brelse(bh);

			/* drop the reference from the wait == 0 run */
			brelse(bh);
			continue;
Y
Yan Zheng 已提交
3373 3374 3375 3376
		} else {
			btrfs_set_super_bytenr(sb, bytenr);

			crc = ~(u32)0;
3377
			crc = btrfs_csum_data((char *)sb +
Y
Yan Zheng 已提交
3378 3379 3380 3381 3382
					      BTRFS_CSUM_SIZE, crc,
					      BTRFS_SUPER_INFO_SIZE -
					      BTRFS_CSUM_SIZE);
			btrfs_csum_final(crc, sb->csum);

3383 3384 3385 3386
			/*
			 * one reference for us, and we leave it for the
			 * caller
			 */
Y
Yan Zheng 已提交
3387 3388
			bh = __getblk(device->bdev, bytenr / 4096,
				      BTRFS_SUPER_INFO_SIZE);
3389
			if (!bh) {
3390 3391 3392
				btrfs_err(device->dev_root->fs_info,
				    "couldn't get super buffer head for bytenr %llu",
				    bytenr);
3393 3394 3395 3396
				errors++;
				continue;
			}

Y
Yan Zheng 已提交
3397 3398
			memcpy(bh->b_data, sb, BTRFS_SUPER_INFO_SIZE);

3399
			/* one reference for submit_bh */
Y
Yan Zheng 已提交
3400
			get_bh(bh);
3401 3402

			set_buffer_uptodate(bh);
Y
Yan Zheng 已提交
3403 3404
			lock_buffer(bh);
			bh->b_end_io = btrfs_end_buffer_write_sync;
3405
			bh->b_private = device;
Y
Yan Zheng 已提交
3406 3407
		}

C
Chris Mason 已提交
3408 3409 3410 3411
		/*
		 * we fua the first super.  The others we allow
		 * to go down lazy.
		 */
3412 3413 3414 3415
		if (i == 0)
			ret = btrfsic_submit_bh(WRITE_FUA, bh);
		else
			ret = btrfsic_submit_bh(WRITE_SYNC, bh);
3416
		if (ret)
Y
Yan Zheng 已提交
3417 3418 3419 3420 3421
			errors++;
	}
	return errors < i ? 0 : -1;
}

C
Chris Mason 已提交
3422 3423 3424 3425
/*
 * endio for the write_dev_flush, this will wake anyone waiting
 * for the barrier when it is done
 */
3426
static void btrfs_end_empty_barrier(struct bio *bio)
C
Chris Mason 已提交
3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454
{
	if (bio->bi_private)
		complete(bio->bi_private);
	bio_put(bio);
}

/*
 * trigger flushes for one the devices.  If you pass wait == 0, the flushes are
 * sent down.  With wait == 1, it waits for the previous flush.
 *
 * any device where the flush fails with eopnotsupp are flagged as not-barrier
 * capable
 */
static int write_dev_flush(struct btrfs_device *device, int wait)
{
	struct bio *bio;
	int ret = 0;

	if (device->nobarriers)
		return 0;

	if (wait) {
		bio = device->flush_bio;
		if (!bio)
			return 0;

		wait_for_completion(&device->flush_wait);

3455 3456
		if (bio->bi_error) {
			ret = bio->bi_error;
3457 3458
			btrfs_dev_stat_inc_and_print(device,
				BTRFS_DEV_STAT_FLUSH_ERRS);
C
Chris Mason 已提交
3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471
		}

		/* drop the reference from the wait == 0 run */
		bio_put(bio);
		device->flush_bio = NULL;

		return ret;
	}

	/*
	 * one reference for us, and we leave it for the
	 * caller
	 */
3472
	device->flush_bio = NULL;
3473
	bio = btrfs_io_bio_alloc(GFP_NOFS, 0);
C
Chris Mason 已提交
3474 3475 3476 3477 3478 3479 3480 3481 3482 3483
	if (!bio)
		return -ENOMEM;

	bio->bi_end_io = btrfs_end_empty_barrier;
	bio->bi_bdev = device->bdev;
	init_completion(&device->flush_wait);
	bio->bi_private = &device->flush_wait;
	device->flush_bio = bio;

	bio_get(bio);
3484
	btrfsic_submit_bio(WRITE_FLUSH, bio);
C
Chris Mason 已提交
3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496

	return 0;
}

/*
 * send an empty flush down to each device in parallel,
 * then wait for them
 */
static int barrier_all_devices(struct btrfs_fs_info *info)
{
	struct list_head *head;
	struct btrfs_device *dev;
3497 3498
	int errors_send = 0;
	int errors_wait = 0;
C
Chris Mason 已提交
3499 3500 3501 3502 3503
	int ret;

	/* send down all the barriers */
	head = &info->fs_devices->devices;
	list_for_each_entry_rcu(dev, head, dev_list) {
3504 3505
		if (dev->missing)
			continue;
C
Chris Mason 已提交
3506
		if (!dev->bdev) {
3507
			errors_send++;
C
Chris Mason 已提交
3508 3509 3510 3511 3512 3513 3514
			continue;
		}
		if (!dev->in_fs_metadata || !dev->writeable)
			continue;

		ret = write_dev_flush(dev, 0);
		if (ret)
3515
			errors_send++;
C
Chris Mason 已提交
3516 3517 3518 3519
	}

	/* wait for all the barriers */
	list_for_each_entry_rcu(dev, head, dev_list) {
3520 3521
		if (dev->missing)
			continue;
C
Chris Mason 已提交
3522
		if (!dev->bdev) {
3523
			errors_wait++;
C
Chris Mason 已提交
3524 3525 3526 3527 3528 3529 3530
			continue;
		}
		if (!dev->in_fs_metadata || !dev->writeable)
			continue;

		ret = write_dev_flush(dev, 1);
		if (ret)
3531
			errors_wait++;
C
Chris Mason 已提交
3532
	}
3533 3534
	if (errors_send > info->num_tolerated_disk_barrier_failures ||
	    errors_wait > info->num_tolerated_disk_barrier_failures)
C
Chris Mason 已提交
3535 3536 3537 3538
		return -EIO;
	return 0;
}

3539 3540
int btrfs_get_num_tolerated_disk_barrier_failures(u64 flags)
{
3541 3542
	int raid_type;
	int min_tolerated = INT_MAX;
3543

3544 3545 3546 3547 3548
	if ((flags & BTRFS_BLOCK_GROUP_PROFILE_MASK) == 0 ||
	    (flags & BTRFS_AVAIL_ALLOC_BIT_SINGLE))
		min_tolerated = min(min_tolerated,
				    btrfs_raid_array[BTRFS_RAID_SINGLE].
				    tolerated_failures);
3549

3550 3551 3552 3553 3554 3555 3556 3557 3558
	for (raid_type = 0; raid_type < BTRFS_NR_RAID_TYPES; raid_type++) {
		if (raid_type == BTRFS_RAID_SINGLE)
			continue;
		if (!(flags & btrfs_raid_group[raid_type]))
			continue;
		min_tolerated = min(min_tolerated,
				    btrfs_raid_array[raid_type].
				    tolerated_failures);
	}
3559

3560 3561 3562 3563 3564 3565
	if (min_tolerated == INT_MAX) {
		pr_warn("BTRFS: unknown raid flag: %llu\n", flags);
		min_tolerated = 0;
	}

	return min_tolerated;
3566 3567
}

3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581
int btrfs_calc_num_tolerated_disk_barrier_failures(
	struct btrfs_fs_info *fs_info)
{
	struct btrfs_ioctl_space_info space;
	struct btrfs_space_info *sinfo;
	u64 types[] = {BTRFS_BLOCK_GROUP_DATA,
		       BTRFS_BLOCK_GROUP_SYSTEM,
		       BTRFS_BLOCK_GROUP_METADATA,
		       BTRFS_BLOCK_GROUP_DATA | BTRFS_BLOCK_GROUP_METADATA};
	int i;
	int c;
	int num_tolerated_disk_barrier_failures =
		(int)fs_info->fs_devices->num_devices;

3582
	for (i = 0; i < ARRAY_SIZE(types); i++) {
3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599
		struct btrfs_space_info *tmp;

		sinfo = NULL;
		rcu_read_lock();
		list_for_each_entry_rcu(tmp, &fs_info->space_info, list) {
			if (tmp->flags == types[i]) {
				sinfo = tmp;
				break;
			}
		}
		rcu_read_unlock();

		if (!sinfo)
			continue;

		down_read(&sinfo->groups_sem);
		for (c = 0; c < BTRFS_NR_RAID_TYPES; c++) {
3600 3601 3602 3603 3604 3605 3606 3607 3608 3609
			u64 flags;

			if (list_empty(&sinfo->block_groups[c]))
				continue;

			btrfs_get_block_group_info(&sinfo->block_groups[c],
						   &space);
			if (space.total_bytes == 0 || space.used_bytes == 0)
				continue;
			flags = space.flags;
3610 3611 3612 3613 3614

			num_tolerated_disk_barrier_failures = min(
				num_tolerated_disk_barrier_failures,
				btrfs_get_num_tolerated_disk_barrier_failures(
					flags));
3615 3616 3617 3618 3619 3620 3621
		}
		up_read(&sinfo->groups_sem);
	}

	return num_tolerated_disk_barrier_failures;
}

3622
static int write_all_supers(struct btrfs_root *root, int max_mirrors)
3623
{
3624
	struct list_head *head;
3625
	struct btrfs_device *dev;
3626
	struct btrfs_super_block *sb;
3627 3628 3629
	struct btrfs_dev_item *dev_item;
	int ret;
	int do_barriers;
3630 3631
	int max_errors;
	int total_errors = 0;
3632
	u64 flags;
3633 3634

	do_barriers = !btrfs_test_opt(root, NOBARRIER);
C
Chris Mason 已提交
3635
	backup_super_roots(root->fs_info);
3636

3637
	sb = root->fs_info->super_for_commit;
3638
	dev_item = &sb->dev_item;
3639

3640
	mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
3641
	head = &root->fs_info->fs_devices->devices;
3642
	max_errors = btrfs_super_num_devices(root->fs_info->super_copy) - 1;
C
Chris Mason 已提交
3643

3644 3645 3646 3647 3648
	if (do_barriers) {
		ret = barrier_all_devices(root->fs_info);
		if (ret) {
			mutex_unlock(
				&root->fs_info->fs_devices->device_list_mutex);
3649
			btrfs_handle_fs_error(root->fs_info, ret,
3650 3651 3652 3653
				    "errors while submitting device barriers.");
			return ret;
		}
	}
C
Chris Mason 已提交
3654

3655
	list_for_each_entry_rcu(dev, head, dev_list) {
3656 3657 3658 3659
		if (!dev->bdev) {
			total_errors++;
			continue;
		}
Y
Yan Zheng 已提交
3660
		if (!dev->in_fs_metadata || !dev->writeable)
3661 3662
			continue;

Y
Yan Zheng 已提交
3663
		btrfs_set_stack_device_generation(dev_item, 0);
3664 3665
		btrfs_set_stack_device_type(dev_item, dev->type);
		btrfs_set_stack_device_id(dev_item, dev->devid);
3666
		btrfs_set_stack_device_total_bytes(dev_item,
3667
						   dev->commit_total_bytes);
3668 3669
		btrfs_set_stack_device_bytes_used(dev_item,
						  dev->commit_bytes_used);
3670 3671 3672 3673
		btrfs_set_stack_device_io_align(dev_item, dev->io_align);
		btrfs_set_stack_device_io_width(dev_item, dev->io_width);
		btrfs_set_stack_device_sector_size(dev_item, dev->sector_size);
		memcpy(dev_item->uuid, dev->uuid, BTRFS_UUID_SIZE);
Y
Yan Zheng 已提交
3674
		memcpy(dev_item->fsid, dev->fs_devices->fsid, BTRFS_UUID_SIZE);
Y
Yan Zheng 已提交
3675

3676 3677 3678
		flags = btrfs_super_flags(sb);
		btrfs_set_super_flags(sb, flags | BTRFS_HEADER_FLAG_WRITTEN);

Y
Yan Zheng 已提交
3679
		ret = write_dev_supers(dev, sb, do_barriers, 0, max_mirrors);
3680 3681
		if (ret)
			total_errors++;
3682
	}
3683
	if (total_errors > max_errors) {
3684
		btrfs_err(root->fs_info, "%d errors while writing supers",
C
Chris Mason 已提交
3685
		       total_errors);
3686
		mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
3687

3688
		/* FUA is masked off if unsupported and can't be the reason */
3689
		btrfs_handle_fs_error(root->fs_info, -EIO,
3690 3691
			    "%d errors while writing supers", total_errors);
		return -EIO;
3692
	}
3693

Y
Yan Zheng 已提交
3694
	total_errors = 0;
3695
	list_for_each_entry_rcu(dev, head, dev_list) {
3696 3697
		if (!dev->bdev)
			continue;
Y
Yan Zheng 已提交
3698
		if (!dev->in_fs_metadata || !dev->writeable)
3699 3700
			continue;

Y
Yan Zheng 已提交
3701 3702 3703
		ret = write_dev_supers(dev, sb, do_barriers, 1, max_mirrors);
		if (ret)
			total_errors++;
3704
	}
3705
	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
3706
	if (total_errors > max_errors) {
3707
		btrfs_handle_fs_error(root->fs_info, -EIO,
3708 3709
			    "%d errors while writing supers", total_errors);
		return -EIO;
3710
	}
3711 3712 3713
	return 0;
}

Y
Yan Zheng 已提交
3714 3715
int write_ctree_super(struct btrfs_trans_handle *trans,
		      struct btrfs_root *root, int max_mirrors)
3716
{
3717
	return write_all_supers(root, max_mirrors);
C
Chris Mason 已提交
3718 3719
}

3720 3721 3722
/* Drop a fs root from the radix tree and free it. */
void btrfs_drop_and_free_fs_root(struct btrfs_fs_info *fs_info,
				  struct btrfs_root *root)
C
Chris Mason 已提交
3723
{
3724
	spin_lock(&fs_info->fs_roots_radix_lock);
C
Chris Mason 已提交
3725 3726
	radix_tree_delete(&fs_info->fs_roots_radix,
			  (unsigned long)root->root_key.objectid);
3727
	spin_unlock(&fs_info->fs_roots_radix_lock);
3728 3729 3730 3731

	if (btrfs_root_refs(&root->root_item) == 0)
		synchronize_srcu(&fs_info->subvol_srcu);

3732
	if (test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state))
L
Liu Bo 已提交
3733 3734
		btrfs_free_log(NULL, root);

3735 3736 3737 3738
	if (root->free_ino_pinned)
		__btrfs_remove_free_space_cache(root->free_ino_pinned);
	if (root->free_ino_ctl)
		__btrfs_remove_free_space_cache(root->free_ino_ctl);
3739 3740 3741 3742 3743
	free_fs_root(root);
}

static void free_fs_root(struct btrfs_root *root)
{
3744
	iput(root->ino_cache_inode);
3745
	WARN_ON(!RB_EMPTY_ROOT(&root->inode_tree));
3746 3747
	btrfs_free_block_rsv(root, root->orphan_block_rsv);
	root->orphan_block_rsv = NULL;
3748 3749
	if (root->anon_dev)
		free_anon_bdev(root->anon_dev);
3750 3751
	if (root->subv_writers)
		btrfs_free_subvolume_writers(root->subv_writers);
3752 3753
	free_extent_buffer(root->node);
	free_extent_buffer(root->commit_root);
3754 3755
	kfree(root->free_ino_ctl);
	kfree(root->free_ino_pinned);
C
Chris Mason 已提交
3756
	kfree(root->name);
3757
	btrfs_put_fs_root(root);
C
Chris Mason 已提交
3758 3759
}

3760 3761 3762
void btrfs_free_fs_root(struct btrfs_root *root)
{
	free_fs_root(root);
C
Chris Mason 已提交
3763 3764
}

Y
Yan Zheng 已提交
3765
int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info)
C
Chris Mason 已提交
3766
{
Y
Yan Zheng 已提交
3767 3768
	u64 root_objectid = 0;
	struct btrfs_root *gang[8];
3769 3770 3771 3772
	int i = 0;
	int err = 0;
	unsigned int ret = 0;
	int index;
3773

Y
Yan Zheng 已提交
3774
	while (1) {
3775
		index = srcu_read_lock(&fs_info->subvol_srcu);
Y
Yan Zheng 已提交
3776 3777 3778
		ret = radix_tree_gang_lookup(&fs_info->fs_roots_radix,
					     (void **)gang, root_objectid,
					     ARRAY_SIZE(gang));
3779 3780
		if (!ret) {
			srcu_read_unlock(&fs_info->subvol_srcu, index);
Y
Yan Zheng 已提交
3781
			break;
3782
		}
3783
		root_objectid = gang[ret - 1]->root_key.objectid + 1;
3784

Y
Yan Zheng 已提交
3785
		for (i = 0; i < ret; i++) {
3786 3787 3788 3789 3790 3791 3792 3793 3794
			/* Avoid to grab roots in dead_roots */
			if (btrfs_root_refs(&gang[i]->root_item) == 0) {
				gang[i] = NULL;
				continue;
			}
			/* grab all the search result for later use */
			gang[i] = btrfs_grab_fs_root(gang[i]);
		}
		srcu_read_unlock(&fs_info->subvol_srcu, index);
3795

3796 3797 3798
		for (i = 0; i < ret; i++) {
			if (!gang[i])
				continue;
Y
Yan Zheng 已提交
3799
			root_objectid = gang[i]->root_key.objectid;
3800 3801
			err = btrfs_orphan_cleanup(gang[i]);
			if (err)
3802 3803
				break;
			btrfs_put_fs_root(gang[i]);
Y
Yan Zheng 已提交
3804 3805 3806
		}
		root_objectid++;
	}
3807 3808 3809 3810 3811 3812 3813

	/* release the uncleaned roots due to error */
	for (; i < ret; i++) {
		if (gang[i])
			btrfs_put_fs_root(gang[i]);
	}
	return err;
Y
Yan Zheng 已提交
3814
}
3815

Y
Yan Zheng 已提交
3816 3817 3818
int btrfs_commit_super(struct btrfs_root *root)
{
	struct btrfs_trans_handle *trans;
3819

Y
Yan Zheng 已提交
3820
	mutex_lock(&root->fs_info->cleaner_mutex);
Y
Yan, Zheng 已提交
3821
	btrfs_run_delayed_iputs(root);
Y
Yan Zheng 已提交
3822
	mutex_unlock(&root->fs_info->cleaner_mutex);
D
David Sterba 已提交
3823
	wake_up_process(root->fs_info->cleaner_kthread);
3824 3825 3826 3827 3828

	/* wait until ongoing cleanup work done */
	down_write(&root->fs_info->cleanup_work_sem);
	up_write(&root->fs_info->cleanup_work_sem);

3829
	trans = btrfs_join_transaction(root);
3830 3831
	if (IS_ERR(trans))
		return PTR_ERR(trans);
3832
	return btrfs_commit_transaction(trans, root);
Y
Yan Zheng 已提交
3833 3834
}

3835
void close_ctree(struct btrfs_root *root)
Y
Yan Zheng 已提交
3836 3837 3838 3839 3840 3841 3842
{
	struct btrfs_fs_info *fs_info = root->fs_info;
	int ret;

	fs_info->closing = 1;
	smp_mb();

3843 3844 3845
	/* wait for the qgroup rescan worker to stop */
	btrfs_qgroup_wait_for_completion(fs_info);

S
Stefan Behrens 已提交
3846 3847 3848 3849 3850
	/* wait for the uuid_scan task to finish */
	down(&fs_info->uuid_tree_rescan_sem);
	/* avoid complains from lockdep et al., set sem back to initial state */
	up(&fs_info->uuid_tree_rescan_sem);

3851
	/* pause restriper - we want to resume on mount */
3852
	btrfs_pause_balance(fs_info);
3853

3854 3855
	btrfs_dev_replace_suspend_for_unmount(fs_info);

3856
	btrfs_scrub_cancel(fs_info);
C
Chris Mason 已提交
3857 3858 3859 3860 3861 3862

	/* wait for any defraggers to finish */
	wait_event(fs_info->transaction_wait,
		   (atomic_read(&fs_info->defrag_running) == 0));

	/* clear out the rbtree of defraggable inodes */
3863
	btrfs_cleanup_defrag_inodes(fs_info);
C
Chris Mason 已提交
3864

3865 3866
	cancel_work_sync(&fs_info->async_reclaim_work);

Y
Yan Zheng 已提交
3867
	if (!(fs_info->sb->s_flags & MS_RDONLY)) {
3868 3869 3870 3871 3872 3873 3874
		/*
		 * If the cleaner thread is stopped and there are
		 * block groups queued for removal, the deletion will be
		 * skipped when we quit the cleaner thread.
		 */
		btrfs_delete_unused_bgs(root->fs_info);

L
liubo 已提交
3875 3876
		ret = btrfs_commit_super(root);
		if (ret)
3877
			btrfs_err(fs_info, "commit super ret %d", ret);
L
liubo 已提交
3878 3879
	}

3880
	if (test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state))
3881
		btrfs_error_commit_super(root);
3882

A
Al Viro 已提交
3883 3884
	kthread_stop(fs_info->transaction_kthread);
	kthread_stop(fs_info->cleaner_kthread);
3885

3886 3887 3888
	fs_info->closing = 2;
	smp_mb();

3889
	btrfs_free_qgroup_config(fs_info);
3890

3891
	if (percpu_counter_sum(&fs_info->delalloc_bytes)) {
3892
		btrfs_info(fs_info, "at unmount delalloc count %lld",
3893
		       percpu_counter_sum(&fs_info->delalloc_bytes));
C
Chris Mason 已提交
3894
	}
3895

3896
	btrfs_sysfs_remove_mounted(fs_info);
3897
	btrfs_sysfs_remove_fsid(fs_info->fs_devices);
3898

3899
	btrfs_free_fs_roots(fs_info);
3900

3901 3902
	btrfs_put_block_group_cache(fs_info);

3903 3904
	btrfs_free_block_groups(fs_info);

3905 3906 3907 3908 3909
	/*
	 * we must make sure there is not any read request to
	 * submit after we stopping all workers.
	 */
	invalidate_inode_pages2(fs_info->btree_inode->i_mapping);
3910 3911
	btrfs_stop_all_workers(fs_info);

3912
	fs_info->open = 0;
3913
	free_root_pointers(fs_info, 1);
3914

3915
	iput(fs_info->btree_inode);
3916

3917 3918 3919 3920 3921
#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
	if (btrfs_test_opt(root, CHECK_INTEGRITY))
		btrfsic_unmount(root, fs_info->fs_devices);
#endif

3922
	btrfs_close_devices(fs_info->fs_devices);
3923
	btrfs_mapping_tree_free(&fs_info->mapping_tree);
3924

3925
	percpu_counter_destroy(&fs_info->dirty_metadata_bytes);
3926
	percpu_counter_destroy(&fs_info->delalloc_bytes);
3927
	percpu_counter_destroy(&fs_info->bio_counter);
C
Chris Mason 已提交
3928
	bdi_destroy(&fs_info->bdi);
3929
	cleanup_srcu_struct(&fs_info->subvol_srcu);
3930

D
David Woodhouse 已提交
3931 3932
	btrfs_free_stripe_hash_table(fs_info);

3933
	__btrfs_free_block_rsv(root->orphan_block_rsv);
3934
	root->orphan_block_rsv = NULL;
3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945

	lock_chunks(root);
	while (!list_empty(&fs_info->pinned_chunks)) {
		struct extent_map *em;

		em = list_first_entry(&fs_info->pinned_chunks,
				      struct extent_map, list);
		list_del_init(&em->list);
		free_extent_map(em);
	}
	unlock_chunks(root);
3946 3947
}

3948 3949
int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid,
			  int atomic)
3950
{
3951
	int ret;
3952
	struct inode *btree_inode = buf->pages[0]->mapping->host;
3953

3954
	ret = extent_buffer_uptodate(buf);
3955 3956 3957 3958
	if (!ret)
		return ret;

	ret = verify_parent_transid(&BTRFS_I(btree_inode)->io_tree, buf,
3959 3960 3961
				    parent_transid, atomic);
	if (ret == -EAGAIN)
		return ret;
3962
	return !ret;
3963 3964 3965 3966
}

void btrfs_mark_buffer_dirty(struct extent_buffer *buf)
{
3967
	struct btrfs_root *root;
3968
	u64 transid = btrfs_header_generation(buf);
3969
	int was_dirty;
3970

3971 3972 3973 3974 3975 3976 3977 3978 3979 3980
#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
	/*
	 * This is a fast path so only do this check if we have sanity tests
	 * enabled.  Normal people shouldn't be marking dummy buffers as dirty
	 * outside of the sanity tests.
	 */
	if (unlikely(test_bit(EXTENT_BUFFER_DUMMY, &buf->bflags)))
		return;
#endif
	root = BTRFS_I(buf->pages[0]->mapping->host)->root;
3981
	btrfs_assert_tree_locked(buf);
J
Julia Lawall 已提交
3982 3983
	if (transid != root->fs_info->generation)
		WARN(1, KERN_CRIT "btrfs transid mismatch buffer %llu, "
C
Chris Mason 已提交
3984
		       "found %llu running %llu\n",
3985
			buf->start, transid, root->fs_info->generation);
3986
	was_dirty = set_extent_buffer_dirty(buf);
3987 3988 3989 3990
	if (!was_dirty)
		__percpu_counter_add(&root->fs_info->dirty_metadata_bytes,
				     buf->len,
				     root->fs_info->dirty_metadata_batch);
3991 3992 3993 3994 3995 3996
#ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY
	if (btrfs_header_level(buf) == 0 && check_leaf(root, buf)) {
		btrfs_print_leaf(root, buf);
		ASSERT(0);
	}
#endif
3997 3998
}

3999 4000
static void __btrfs_btree_balance_dirty(struct btrfs_root *root,
					int flush_delayed)
4001 4002 4003 4004 4005
{
	/*
	 * looks as though older kernels can get into trouble with
	 * this code, they end up stuck in balance_dirty_pages forever
	 */
4006
	int ret;
4007 4008 4009 4010

	if (current->flags & PF_MEMALLOC)
		return;

4011 4012
	if (flush_delayed)
		btrfs_balance_delayed_items(root);
4013

4014 4015 4016
	ret = percpu_counter_compare(&root->fs_info->dirty_metadata_bytes,
				     BTRFS_DIRTY_METADATA_THRESH);
	if (ret > 0) {
4017 4018
		balance_dirty_pages_ratelimited(
				   root->fs_info->btree_inode->i_mapping);
4019 4020 4021
	}
}

4022
void btrfs_btree_balance_dirty(struct btrfs_root *root)
C
Chris Mason 已提交
4023
{
4024 4025
	__btrfs_btree_balance_dirty(root, 1);
}
4026

4027 4028 4029
void btrfs_btree_balance_dirty_nodelay(struct btrfs_root *root)
{
	__btrfs_btree_balance_dirty(root, 0);
C
Chris Mason 已提交
4030
}
4031

4032
int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid)
4033
{
4034
	struct btrfs_root *root = BTRFS_I(buf->pages[0]->mapping->host)->root;
4035
	return btree_read_extent_buffer_pages(root, buf, 0, parent_transid);
4036
}
4037

4038
static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info,
L
liubo 已提交
4039 4040
			      int read_only)
{
D
David Sterba 已提交
4041
	struct btrfs_super_block *sb = fs_info->super_copy;
4042 4043
	u64 nodesize = btrfs_super_nodesize(sb);
	u64 sectorsize = btrfs_super_sectorsize(sb);
D
David Sterba 已提交
4044 4045
	int ret = 0;

4046 4047 4048 4049 4050 4051 4052
	if (btrfs_super_magic(sb) != BTRFS_MAGIC) {
		printk(KERN_ERR "BTRFS: no valid FS found\n");
		ret = -EINVAL;
	}
	if (btrfs_super_flags(sb) & ~BTRFS_SUPER_FLAG_SUPP)
		printk(KERN_WARNING "BTRFS: unrecognized super flag: %llu\n",
				btrfs_super_flags(sb) & ~BTRFS_SUPER_FLAG_SUPP);
4053 4054 4055
	if (btrfs_super_root_level(sb) >= BTRFS_MAX_LEVEL) {
		printk(KERN_ERR "BTRFS: tree_root level too big: %d >= %d\n",
				btrfs_super_root_level(sb), BTRFS_MAX_LEVEL);
D
David Sterba 已提交
4056 4057
		ret = -EINVAL;
	}
4058 4059 4060
	if (btrfs_super_chunk_root_level(sb) >= BTRFS_MAX_LEVEL) {
		printk(KERN_ERR "BTRFS: chunk_root level too big: %d >= %d\n",
				btrfs_super_chunk_root_level(sb), BTRFS_MAX_LEVEL);
D
David Sterba 已提交
4061 4062
		ret = -EINVAL;
	}
4063 4064 4065
	if (btrfs_super_log_root_level(sb) >= BTRFS_MAX_LEVEL) {
		printk(KERN_ERR "BTRFS: log_root level too big: %d >= %d\n",
				btrfs_super_log_root_level(sb), BTRFS_MAX_LEVEL);
D
David Sterba 已提交
4066 4067 4068
		ret = -EINVAL;
	}

D
David Sterba 已提交
4069
	/*
4070 4071
	 * Check sectorsize and nodesize first, other check will need it.
	 * Check all possible sectorsize(4K, 8K, 16K, 32K, 64K) here.
D
David Sterba 已提交
4072
	 */
4073 4074 4075 4076 4077 4078
	if (!is_power_of_2(sectorsize) || sectorsize < 4096 ||
	    sectorsize > BTRFS_MAX_METADATA_BLOCKSIZE) {
		printk(KERN_ERR "BTRFS: invalid sectorsize %llu\n", sectorsize);
		ret = -EINVAL;
	}
	/* Only PAGE SIZE is supported yet */
4079
	if (sectorsize != PAGE_SIZE) {
4080
		printk(KERN_ERR "BTRFS: sectorsize %llu not supported yet, only support %lu\n",
4081
				sectorsize, PAGE_SIZE);
4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097
		ret = -EINVAL;
	}
	if (!is_power_of_2(nodesize) || nodesize < sectorsize ||
	    nodesize > BTRFS_MAX_METADATA_BLOCKSIZE) {
		printk(KERN_ERR "BTRFS: invalid nodesize %llu\n", nodesize);
		ret = -EINVAL;
	}
	if (nodesize != le32_to_cpu(sb->__unused_leafsize)) {
		printk(KERN_ERR "BTRFS: invalid leafsize %u, should be %llu\n",
				le32_to_cpu(sb->__unused_leafsize),
				nodesize);
		ret = -EINVAL;
	}

	/* Root alignment check */
	if (!IS_ALIGNED(btrfs_super_root(sb), sectorsize)) {
D
David Sterba 已提交
4098
		printk(KERN_WARNING "BTRFS: tree_root block unaligned: %llu\n",
4099
				btrfs_super_root(sb));
4100 4101 4102
		ret = -EINVAL;
	}
	if (!IS_ALIGNED(btrfs_super_chunk_root(sb), sectorsize)) {
4103 4104
		printk(KERN_WARNING "BTRFS: chunk_root block unaligned: %llu\n",
				btrfs_super_chunk_root(sb));
4105 4106
		ret = -EINVAL;
	}
4107 4108 4109
	if (!IS_ALIGNED(btrfs_super_log_root(sb), sectorsize)) {
		printk(KERN_WARNING "BTRFS: log_root block unaligned: %llu\n",
				btrfs_super_log_root(sb));
4110 4111 4112
		ret = -EINVAL;
	}

D
David Sterba 已提交
4113 4114 4115 4116 4117 4118 4119 4120 4121 4122
	if (memcmp(fs_info->fsid, sb->dev_item.fsid, BTRFS_UUID_SIZE) != 0) {
		printk(KERN_ERR "BTRFS: dev_item UUID does not match fsid: %pU != %pU\n",
				fs_info->fsid, sb->dev_item.fsid);
		ret = -EINVAL;
	}

	/*
	 * Hint to catch really bogus numbers, bitflips or so, more exact checks are
	 * done later
	 */
4123
	if (btrfs_super_num_devices(sb) > (1UL << 31))
D
David Sterba 已提交
4124
		printk(KERN_WARNING "BTRFS: suspicious number of devices: %llu\n",
4125
				btrfs_super_num_devices(sb));
4126 4127 4128 4129
	if (btrfs_super_num_devices(sb) == 0) {
		printk(KERN_ERR "BTRFS: number of devices is 0\n");
		ret = -EINVAL;
	}
D
David Sterba 已提交
4130

4131
	if (btrfs_super_bytenr(sb) != BTRFS_SUPER_INFO_OFFSET) {
D
David Sterba 已提交
4132
		printk(KERN_ERR "BTRFS: super offset mismatch %llu != %u\n",
4133
				btrfs_super_bytenr(sb), BTRFS_SUPER_INFO_OFFSET);
D
David Sterba 已提交
4134 4135 4136
		ret = -EINVAL;
	}

4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148
	/*
	 * Obvious sys_chunk_array corruptions, it must hold at least one key
	 * and one chunk
	 */
	if (btrfs_super_sys_array_size(sb) > BTRFS_SYSTEM_CHUNK_ARRAY_SIZE) {
		printk(KERN_ERR "BTRFS: system chunk array too big %u > %u\n",
				btrfs_super_sys_array_size(sb),
				BTRFS_SYSTEM_CHUNK_ARRAY_SIZE);
		ret = -EINVAL;
	}
	if (btrfs_super_sys_array_size(sb) < sizeof(struct btrfs_disk_key)
			+ sizeof(struct btrfs_chunk)) {
4149
		printk(KERN_ERR "BTRFS: system chunk array too small %u < %zu\n",
4150 4151 4152 4153 4154 4155
				btrfs_super_sys_array_size(sb),
				sizeof(struct btrfs_disk_key)
				+ sizeof(struct btrfs_chunk));
		ret = -EINVAL;
	}

D
David Sterba 已提交
4156 4157 4158 4159
	/*
	 * The generation is a global counter, we'll trust it more than the others
	 * but it's still possible that it's the one that's wrong.
	 */
4160
	if (btrfs_super_generation(sb) < btrfs_super_chunk_root_generation(sb))
D
David Sterba 已提交
4161 4162
		printk(KERN_WARNING
			"BTRFS: suspicious: generation < chunk_root_generation: %llu < %llu\n",
4163 4164 4165
			btrfs_super_generation(sb), btrfs_super_chunk_root_generation(sb));
	if (btrfs_super_generation(sb) < btrfs_super_cache_generation(sb)
	    && btrfs_super_cache_generation(sb) != (u64)-1)
D
David Sterba 已提交
4166 4167
		printk(KERN_WARNING
			"BTRFS: suspicious: generation < cache_generation: %llu < %llu\n",
4168
			btrfs_super_generation(sb), btrfs_super_cache_generation(sb));
D
David Sterba 已提交
4169 4170

	return ret;
L
liubo 已提交
4171 4172
}

4173
static void btrfs_error_commit_super(struct btrfs_root *root)
L
liubo 已提交
4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185
{
	mutex_lock(&root->fs_info->cleaner_mutex);
	btrfs_run_delayed_iputs(root);
	mutex_unlock(&root->fs_info->cleaner_mutex);

	down_write(&root->fs_info->cleanup_work_sem);
	up_write(&root->fs_info->cleanup_work_sem);

	/* cleanup FS via transaction */
	btrfs_cleanup_transaction(root);
}

4186
static void btrfs_destroy_ordered_extents(struct btrfs_root *root)
L
liubo 已提交
4187 4188 4189
{
	struct btrfs_ordered_extent *ordered;

4190
	spin_lock(&root->ordered_extent_lock);
4191 4192 4193 4194
	/*
	 * This will just short circuit the ordered completion stuff which will
	 * make sure the ordered extent gets properly cleaned up.
	 */
4195
	list_for_each_entry(ordered, &root->ordered_extents,
4196 4197
			    root_extent_list)
		set_bit(BTRFS_ORDERED_IOERR, &ordered->flags);
4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212
	spin_unlock(&root->ordered_extent_lock);
}

static void btrfs_destroy_all_ordered_extents(struct btrfs_fs_info *fs_info)
{
	struct btrfs_root *root;
	struct list_head splice;

	INIT_LIST_HEAD(&splice);

	spin_lock(&fs_info->ordered_root_lock);
	list_splice_init(&fs_info->ordered_roots, &splice);
	while (!list_empty(&splice)) {
		root = list_first_entry(&splice, struct btrfs_root,
					ordered_root);
4213 4214
		list_move_tail(&root->ordered_root,
			       &fs_info->ordered_roots);
4215

4216
		spin_unlock(&fs_info->ordered_root_lock);
4217 4218
		btrfs_destroy_ordered_extents(root);

4219 4220
		cond_resched();
		spin_lock(&fs_info->ordered_root_lock);
4221 4222
	}
	spin_unlock(&fs_info->ordered_root_lock);
L
liubo 已提交
4223 4224
}

4225 4226
static int btrfs_destroy_delayed_refs(struct btrfs_transaction *trans,
				      struct btrfs_root *root)
L
liubo 已提交
4227 4228 4229 4230 4231 4232 4233 4234 4235
{
	struct rb_node *node;
	struct btrfs_delayed_ref_root *delayed_refs;
	struct btrfs_delayed_ref_node *ref;
	int ret = 0;

	delayed_refs = &trans->delayed_refs;

	spin_lock(&delayed_refs->lock);
4236
	if (atomic_read(&delayed_refs->num_entries) == 0) {
4237
		spin_unlock(&delayed_refs->lock);
4238
		btrfs_info(root->fs_info, "delayed_refs has NO entry");
L
liubo 已提交
4239 4240 4241
		return ret;
	}

4242 4243
	while ((node = rb_first(&delayed_refs->href_root)) != NULL) {
		struct btrfs_delayed_ref_head *head;
4244
		struct btrfs_delayed_ref_node *tmp;
4245
		bool pin_bytes = false;
L
liubo 已提交
4246

4247 4248 4249 4250 4251
		head = rb_entry(node, struct btrfs_delayed_ref_head,
				href_node);
		if (!mutex_trylock(&head->mutex)) {
			atomic_inc(&head->node.refs);
			spin_unlock(&delayed_refs->lock);
4252

4253
			mutex_lock(&head->mutex);
4254
			mutex_unlock(&head->mutex);
4255 4256 4257 4258 4259
			btrfs_put_delayed_ref(&head->node);
			spin_lock(&delayed_refs->lock);
			continue;
		}
		spin_lock(&head->lock);
4260 4261
		list_for_each_entry_safe_reverse(ref, tmp, &head->ref_list,
						 list) {
4262
			ref->in_tree = 0;
4263
			list_del(&ref->list);
4264 4265
			atomic_dec(&delayed_refs->num_entries);
			btrfs_put_delayed_ref(ref);
4266
		}
4267 4268 4269 4270 4271 4272 4273 4274 4275 4276 4277 4278
		if (head->must_insert_reserved)
			pin_bytes = true;
		btrfs_free_delayed_extent_op(head->extent_op);
		delayed_refs->num_heads--;
		if (head->processing == 0)
			delayed_refs->num_heads_ready--;
		atomic_dec(&delayed_refs->num_entries);
		head->node.in_tree = 0;
		rb_erase(&head->href_node, &delayed_refs->href_root);
		spin_unlock(&head->lock);
		spin_unlock(&delayed_refs->lock);
		mutex_unlock(&head->mutex);
L
liubo 已提交
4279

4280 4281 4282 4283
		if (pin_bytes)
			btrfs_pin_extent(root, head->node.bytenr,
					 head->node.num_bytes, 1);
		btrfs_put_delayed_ref(&head->node);
L
liubo 已提交
4284 4285 4286 4287 4288 4289 4290 4291 4292
		cond_resched();
		spin_lock(&delayed_refs->lock);
	}

	spin_unlock(&delayed_refs->lock);

	return ret;
}

4293
static void btrfs_destroy_delalloc_inodes(struct btrfs_root *root)
L
liubo 已提交
4294 4295 4296 4297 4298 4299
{
	struct btrfs_inode *btrfs_inode;
	struct list_head splice;

	INIT_LIST_HEAD(&splice);

4300 4301
	spin_lock(&root->delalloc_lock);
	list_splice_init(&root->delalloc_inodes, &splice);
L
liubo 已提交
4302 4303

	while (!list_empty(&splice)) {
4304 4305
		btrfs_inode = list_first_entry(&splice, struct btrfs_inode,
					       delalloc_inodes);
L
liubo 已提交
4306 4307

		list_del_init(&btrfs_inode->delalloc_inodes);
4308 4309
		clear_bit(BTRFS_INODE_IN_DELALLOC_LIST,
			  &btrfs_inode->runtime_flags);
4310
		spin_unlock(&root->delalloc_lock);
L
liubo 已提交
4311 4312

		btrfs_invalidate_inodes(btrfs_inode->root);
4313

4314
		spin_lock(&root->delalloc_lock);
L
liubo 已提交
4315 4316
	}

4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333 4334 4335 4336 4337 4338 4339 4340 4341 4342
	spin_unlock(&root->delalloc_lock);
}

static void btrfs_destroy_all_delalloc_inodes(struct btrfs_fs_info *fs_info)
{
	struct btrfs_root *root;
	struct list_head splice;

	INIT_LIST_HEAD(&splice);

	spin_lock(&fs_info->delalloc_root_lock);
	list_splice_init(&fs_info->delalloc_roots, &splice);
	while (!list_empty(&splice)) {
		root = list_first_entry(&splice, struct btrfs_root,
					 delalloc_root);
		list_del_init(&root->delalloc_root);
		root = btrfs_grab_fs_root(root);
		BUG_ON(!root);
		spin_unlock(&fs_info->delalloc_root_lock);

		btrfs_destroy_delalloc_inodes(root);
		btrfs_put_fs_root(root);

		spin_lock(&fs_info->delalloc_root_lock);
	}
	spin_unlock(&fs_info->delalloc_root_lock);
L
liubo 已提交
4343 4344 4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355
}

static int btrfs_destroy_marked_extents(struct btrfs_root *root,
					struct extent_io_tree *dirty_pages,
					int mark)
{
	int ret;
	struct extent_buffer *eb;
	u64 start = 0;
	u64 end;

	while (1) {
		ret = find_first_extent_bit(dirty_pages, start, &start, &end,
4356
					    mark, NULL);
L
liubo 已提交
4357 4358 4359 4360 4361
		if (ret)
			break;

		clear_extent_bits(dirty_pages, start, end, mark, GFP_NOFS);
		while (start <= end) {
4362
			eb = btrfs_find_tree_block(root->fs_info, start);
4363
			start += root->nodesize;
4364
			if (!eb)
L
liubo 已提交
4365
				continue;
4366
			wait_on_extent_buffer_writeback(eb);
L
liubo 已提交
4367

4368 4369 4370 4371
			if (test_and_clear_bit(EXTENT_BUFFER_DIRTY,
					       &eb->bflags))
				clear_extent_buffer_dirty(eb);
			free_extent_buffer_stale(eb);
L
liubo 已提交
4372 4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383 4384
		}
	}

	return ret;
}

static int btrfs_destroy_pinned_extent(struct btrfs_root *root,
				       struct extent_io_tree *pinned_extents)
{
	struct extent_io_tree *unpin;
	u64 start;
	u64 end;
	int ret;
4385
	bool loop = true;
L
liubo 已提交
4386 4387

	unpin = pinned_extents;
4388
again:
L
liubo 已提交
4389 4390
	while (1) {
		ret = find_first_extent_bit(unpin, 0, &start, &end,
4391
					    EXTENT_DIRTY, NULL);
L
liubo 已提交
4392 4393 4394 4395 4396 4397 4398 4399
		if (ret)
			break;

		clear_extent_dirty(unpin, start, end, GFP_NOFS);
		btrfs_error_unpin_extent_range(root, start, end);
		cond_resched();
	}

4400 4401 4402 4403 4404 4405 4406 4407 4408
	if (loop) {
		if (unpin == &root->fs_info->freed_extents[0])
			unpin = &root->fs_info->freed_extents[1];
		else
			unpin = &root->fs_info->freed_extents[0];
		loop = false;
		goto again;
	}

L
liubo 已提交
4409 4410 4411
	return 0;
}

4412 4413 4414 4415 4416
void btrfs_cleanup_one_transaction(struct btrfs_transaction *cur_trans,
				   struct btrfs_root *root)
{
	btrfs_destroy_delayed_refs(cur_trans, root);

4417
	cur_trans->state = TRANS_STATE_COMMIT_START;
4418
	wake_up(&root->fs_info->transaction_blocked_wait);
4419

4420
	cur_trans->state = TRANS_STATE_UNBLOCKED;
4421
	wake_up(&root->fs_info->transaction_wait);
4422

4423 4424
	btrfs_destroy_delayed_inodes(root);
	btrfs_assert_delayed_root_empty(root);
4425 4426 4427

	btrfs_destroy_marked_extents(root, &cur_trans->dirty_pages,
				     EXTENT_DIRTY);
4428 4429
	btrfs_destroy_pinned_extent(root,
				    root->fs_info->pinned_extents);
4430

4431 4432 4433
	cur_trans->state =TRANS_STATE_COMPLETED;
	wake_up(&cur_trans->commit_wait);

4434 4435 4436 4437 4438 4439
	/*
	memset(cur_trans, 0, sizeof(*cur_trans));
	kmem_cache_free(btrfs_transaction_cachep, cur_trans);
	*/
}

4440
static int btrfs_cleanup_transaction(struct btrfs_root *root)
L
liubo 已提交
4441 4442 4443 4444 4445
{
	struct btrfs_transaction *t;

	mutex_lock(&root->fs_info->transaction_kthread_mutex);

J
Josef Bacik 已提交
4446
	spin_lock(&root->fs_info->trans_lock);
4447 4448 4449 4450 4451 4452 4453 4454 4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468 4469 4470
	while (!list_empty(&root->fs_info->trans_list)) {
		t = list_first_entry(&root->fs_info->trans_list,
				     struct btrfs_transaction, list);
		if (t->state >= TRANS_STATE_COMMIT_START) {
			atomic_inc(&t->use_count);
			spin_unlock(&root->fs_info->trans_lock);
			btrfs_wait_for_commit(root, t->transid);
			btrfs_put_transaction(t);
			spin_lock(&root->fs_info->trans_lock);
			continue;
		}
		if (t == root->fs_info->running_transaction) {
			t->state = TRANS_STATE_COMMIT_DOING;
			spin_unlock(&root->fs_info->trans_lock);
			/*
			 * We wait for 0 num_writers since we don't hold a trans
			 * handle open currently for this transaction.
			 */
			wait_event(t->writer_wait,
				   atomic_read(&t->num_writers) == 0);
		} else {
			spin_unlock(&root->fs_info->trans_lock);
		}
		btrfs_cleanup_one_transaction(t, root);
4471

4472 4473 4474
		spin_lock(&root->fs_info->trans_lock);
		if (t == root->fs_info->running_transaction)
			root->fs_info->running_transaction = NULL;
L
liubo 已提交
4475
		list_del_init(&t->list);
4476
		spin_unlock(&root->fs_info->trans_lock);
L
liubo 已提交
4477

4478 4479 4480 4481 4482 4483 4484 4485 4486 4487
		btrfs_put_transaction(t);
		trace_btrfs_transaction_commit(root);
		spin_lock(&root->fs_info->trans_lock);
	}
	spin_unlock(&root->fs_info->trans_lock);
	btrfs_destroy_all_ordered_extents(root->fs_info);
	btrfs_destroy_delayed_inodes(root);
	btrfs_assert_delayed_root_empty(root);
	btrfs_destroy_pinned_extent(root, root->fs_info->pinned_extents);
	btrfs_destroy_all_delalloc_inodes(root->fs_info);
L
liubo 已提交
4488 4489 4490 4491 4492
	mutex_unlock(&root->fs_info->transaction_kthread_mutex);

	return 0;
}

4493
static const struct extent_io_ops btree_extent_io_ops = {
4494
	.readpage_end_io_hook = btree_readpage_end_io_hook,
A
Arne Jansen 已提交
4495
	.readpage_io_failed_hook = btree_io_failed_hook,
4496
	.submit_bio_hook = btree_submit_bio_hook,
4497 4498
	/* note we're sharing with inode.c for the merge bio hook */
	.merge_bio_hook = btrfs_merge_bio_hook,
4499
};