ctree.h 65.9 KB
Newer Older
Chris Mason's avatar
Chris Mason committed
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.
 */

Chris Mason's avatar
Chris Mason committed
19
20
#ifndef __BTRFS_CTREE__
#define __BTRFS_CTREE__
21

22
#include <linux/version.h>
23
24
#include <linux/mm.h>
#include <linux/highmem.h>
Chris Mason's avatar
Chris Mason committed
25
#include <linux/fs.h>
26
#include <linux/completion.h>
Chris Mason's avatar
Chris Mason committed
27
#include <linux/backing-dev.h>
28
#include <linux/wait.h>
29
#include <asm/kmap_types.h>
30
#include "extent_io.h"
31
#include "extent_map.h"
32
#include "async-thread.h"
Chris Mason's avatar
Chris Mason committed
33

34
struct btrfs_trans_handle;
Chris Mason's avatar
Chris Mason committed
35
struct btrfs_transaction;
36
37
38
extern struct kmem_cache *btrfs_trans_handle_cachep;
extern struct kmem_cache *btrfs_transaction_cachep;
extern struct kmem_cache *btrfs_bit_radix_cachep;
Chris Mason's avatar
Chris Mason committed
39
extern struct kmem_cache *btrfs_path_cachep;
40
struct btrfs_ordered_sum;
41

42
#define BTRFS_MAGIC "_BDRfS_M"
43

Josef Bacik's avatar
Josef Bacik committed
44
45
#define BTRFS_ACL_NOT_CACHED    ((void *)-1)

46
47
48
49
50
#ifdef CONFIG_LOCKDEP
# define BTRFS_MAX_LEVEL 7
#else
# define BTRFS_MAX_LEVEL 8
#endif
51
52

/* holds pointers to all of the tree roots */
53
#define BTRFS_ROOT_TREE_OBJECTID 1ULL
54
55

/* stores information about which extents are in use, and reference counts */
Chris Mason's avatar
Chris Mason committed
56
#define BTRFS_EXTENT_TREE_OBJECTID 2ULL
57
58
59
60
61

/*
 * chunk tree stores translations from logical -> physical block numbering
 * the super block points to the chunk tree
 */
62
#define BTRFS_CHUNK_TREE_OBJECTID 3ULL
63
64
65
66
67

/*
 * stores information about which areas of a given device are in use.
 * one per device.  The tree of tree roots points to the device tree
 */
68
69
70
71
72
73
74
#define BTRFS_DEV_TREE_OBJECTID 4ULL

/* one per subvolume, storing files and directories */
#define BTRFS_FS_TREE_OBJECTID 5ULL

/* directory objectid inside the root tree */
#define BTRFS_ROOT_TREE_DIR_OBJECTID 6ULL
75

76
77
78
/* orhpan objectid for tracking unlinked/truncated files */
#define BTRFS_ORPHAN_OBJECTID -5ULL

79
80
81
82
/* does write ahead logging to speed up fsyncs */
#define BTRFS_TREE_LOG_OBJECTID -6ULL
#define BTRFS_TREE_LOG_FIXUP_OBJECTID -7ULL

Zheng Yan's avatar
Zheng Yan committed
83
84
85
86
/* for space balancing */
#define BTRFS_TREE_RELOC_OBJECTID -8ULL
#define BTRFS_DATA_RELOC_TREE_OBJECTID -9ULL

Zheng Yan's avatar
Zheng Yan committed
87
88
89
/* dummy objectid represents multiple objectids */
#define BTRFS_MULTIPLE_OBJECTIDS -255ULL

90
/*
91
 * All files have objectids in this range.
92
 */
93
#define BTRFS_FIRST_FREE_OBJECTID 256ULL
94
#define BTRFS_LAST_FREE_OBJECTID -256ULL
95
#define BTRFS_FIRST_CHUNK_TREE_OBJECTID 256ULL
96

97
98
99
100
101
102
103

/*
 * the device items go into the chunk tree.  The key is in the form
 * [ 1 BTRFS_DEV_ITEM_KEY device_id ]
 */
#define BTRFS_DEV_ITEMS_OBJECTID 1ULL

Chris Mason's avatar
Chris Mason committed
104
105
106
107
108
109
/*
 * we can actually store much bigger names, but lets not confuse the rest
 * of linux
 */
#define BTRFS_NAME_LEN 255

Chris Mason's avatar
Chris Mason committed
110
111
/* 32 bytes in various csum fields */
#define BTRFS_CSUM_SIZE 32
112
113
/* four bytes for CRC32 */
#define BTRFS_CRC32_SIZE 4
114
#define BTRFS_EMPTY_DIR_SIZE 0
Chris Mason's avatar
Chris Mason committed
115

Chris Mason's avatar
Chris Mason committed
116
117
118
119
120
121
122
123
#define BTRFS_FT_UNKNOWN	0
#define BTRFS_FT_REG_FILE	1
#define BTRFS_FT_DIR		2
#define BTRFS_FT_CHRDEV		3
#define BTRFS_FT_BLKDEV		4
#define BTRFS_FT_FIFO		5
#define BTRFS_FT_SOCK		6
#define BTRFS_FT_SYMLINK	7
Josef Bacik's avatar
Josef Bacik committed
124
125
#define BTRFS_FT_XATTR		8
#define BTRFS_FT_MAX		9
Chris Mason's avatar
Chris Mason committed
126

127
128
129
130
131
132
133
134
135
/*
 * the key defines the order in the tree, and so it also defines (optimal)
 * block layout.  objectid corresonds to the inode number.  The flags
 * tells us things about the object, and is a kind of stream selector.
 * so for a given inode, keys with flags of 1 might refer to the inode
 * data, flags of 2 may point to file data in the btree and flags == 3
 * may point to extents.
 *
 * offset is the starting byte offset for this key in the stream.
Chris Mason's avatar
Chris Mason committed
136
137
138
139
 *
 * btrfs_disk_key is in disk byte order.  struct btrfs_key is always
 * in cpu native order.  Otherwise they are identical and their sizes
 * should be the same (ie both packed)
140
 */
Chris Mason's avatar
Chris Mason committed
141
142
struct btrfs_disk_key {
	__le64 objectid;
143
	u8 type;
144
	__le64 offset;
Chris Mason's avatar
Chris Mason committed
145
146
147
} __attribute__ ((__packed__));

struct btrfs_key {
148
	u64 objectid;
149
	u8 type;
150
	u64 offset;
151
152
} __attribute__ ((__packed__));

153
154
155
156
struct btrfs_mapping_tree {
	struct extent_map_tree map_tree;
};

157
#define BTRFS_UUID_SIZE 16
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
struct btrfs_dev_item {
	/* the internal btrfs device id */
	__le64 devid;

	/* size of the device */
	__le64 total_bytes;

	/* bytes used */
	__le64 bytes_used;

	/* optimal io alignment for this device */
	__le32 io_align;

	/* optimal io width for this device */
	__le32 io_width;

	/* minimal io size for this device */
	__le32 sector_size;

	/* type and info about this device */
	__le64 type;

Yan Zheng's avatar
Yan Zheng committed
180
181
182
	/* expected generation for this device */
	__le64 generation;

183
184
185
186
187
188
189
190
191
	/* grouping information for allocation decisions */
	__le32 dev_group;

	/* seek speed 0-100 where 100 is fastest */
	u8 seek_speed;

	/* bandwidth 0-100 where 100 is fastest */
	u8 bandwidth;

192
	/* btrfs generated uuid for this device */
193
	u8 uuid[BTRFS_UUID_SIZE];
Yan Zheng's avatar
Yan Zheng committed
194
195
196

	/* uuid of FS who owns this device */
	u8 fsid[BTRFS_UUID_SIZE];
197
198
199
200
201
} __attribute__ ((__packed__));

struct btrfs_stripe {
	__le64 devid;
	__le64 offset;
202
	u8 dev_uuid[BTRFS_UUID_SIZE];
203
204
205
} __attribute__ ((__packed__));

struct btrfs_chunk {
206
207
208
209
	/* size of this chunk in bytes */
	__le64 length;

	/* objectid of the root referencing this chunk */
210
	__le64 owner;
211

212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
	__le64 stripe_len;
	__le64 type;

	/* optimal io alignment for this chunk */
	__le32 io_align;

	/* optimal io width for this chunk */
	__le32 io_width;

	/* minimal io size for this chunk */
	__le32 sector_size;

	/* 2^16 stripes is quite a lot, a second limit is the size of a single
	 * item in the btree
	 */
	__le16 num_stripes;
Chris Mason's avatar
Chris Mason committed
228
229
230

	/* sub stripes only matter for raid10 */
	__le16 sub_stripes;
231
232
233
234
235
236
237
238
239
240
241
	struct btrfs_stripe stripe;
	/* additional stripes go here */
} __attribute__ ((__packed__));

static inline unsigned long btrfs_chunk_item_size(int num_stripes)
{
	BUG_ON(num_stripes == 0);
	return sizeof(struct btrfs_chunk) +
		sizeof(struct btrfs_stripe) * (num_stripes - 1);
}

242
#define BTRFS_FSID_SIZE 16
243
244
#define BTRFS_HEADER_FLAG_WRITTEN (1 << 0)

245
246
247
/*
 * every tree block (leaf or node) starts with this header.
 */
248
struct btrfs_header {
249
	/* these first four must match the super block */
Chris Mason's avatar
Chris Mason committed
250
	u8 csum[BTRFS_CSUM_SIZE];
251
	u8 fsid[BTRFS_FSID_SIZE]; /* FS specific uuid */
252
	__le64 bytenr; /* which block this node is supposed to live in */
253
	__le64 flags;
254
255
256

	/* allowed to be different from the super from here on down */
	u8 chunk_tree_uuid[BTRFS_UUID_SIZE];
257
	__le64 generation;
258
	__le64 owner;
259
	__le32 nritems;
260
	u8 level;
261
262
} __attribute__ ((__packed__));

263
#define BTRFS_NODEPTRS_PER_BLOCK(r) (((r)->nodesize - \
264
			        sizeof(struct btrfs_header)) / \
265
			        sizeof(struct btrfs_key_ptr))
266
#define __BTRFS_LEAF_DATA_SIZE(bs) ((bs) - sizeof(struct btrfs_header))
267
#define BTRFS_LEAF_DATA_SIZE(r) (__BTRFS_LEAF_DATA_SIZE(r->leafsize))
268
269
270
#define BTRFS_MAX_INLINE_DATA_SIZE(r) (BTRFS_LEAF_DATA_SIZE(r) - \
					sizeof(struct btrfs_item) - \
					sizeof(struct btrfs_file_extent_item))
271

Yan Zheng's avatar
Yan Zheng committed
272
#define BTRFS_SUPER_FLAG_SEEDING (1ULL << 32)
273
274
275
276
277
278

/*
 * this is a very generous portion of the super block, giving us
 * room to translate 14 chunks with 3 stripes each.
 */
#define BTRFS_SYSTEM_CHUNK_ARRAY_SIZE 2048
279
#define BTRFS_LABEL_SIZE 256
280

281
282
283
284
/*
 * the super block basically lists the main trees of the FS
 * it currently lacks any block count etc etc
 */
285
struct btrfs_super_block {
Chris Mason's avatar
Chris Mason committed
286
	u8 csum[BTRFS_CSUM_SIZE];
287
	/* the first 4 fields must match struct btrfs_header */
Yan Zheng's avatar
Yan Zheng committed
288
	u8 fsid[BTRFS_FSID_SIZE];    /* FS specific uuid */
289
	__le64 bytenr; /* this block number */
290
	__le64 flags;
291
292

	/* allowed to be different from the btrfs_header from here own down */
293
294
295
	__le64 magic;
	__le64 generation;
	__le64 root;
296
	__le64 chunk_root;
297
	__le64 log_root;
298
299
	__le64 total_bytes;
	__le64 bytes_used;
300
	__le64 root_dir_objectid;
301
	__le64 num_devices;
302
303
304
	__le32 sectorsize;
	__le32 nodesize;
	__le32 leafsize;
305
	__le32 stripesize;
306
	__le32 sys_chunk_array_size;
307
	__le64 chunk_root_generation;
308
	u8 root_level;
309
	u8 chunk_root_level;
310
	u8 log_root_level;
311
	struct btrfs_dev_item dev_item;
312
	char label[BTRFS_LABEL_SIZE];
313
	u8 sys_chunk_array[BTRFS_SYSTEM_CHUNK_ARRAY_SIZE];
Chris Mason's avatar
Chris Mason committed
314
315
} __attribute__ ((__packed__));

316
/*
317
 * A leaf is full of items. offset and size tell us where to find
318
319
 * the item in the leaf (relative to the start of the data area)
 */
Chris Mason's avatar
Chris Mason committed
320
struct btrfs_item {
Chris Mason's avatar
Chris Mason committed
321
	struct btrfs_disk_key key;
322
	__le32 offset;
323
	__le32 size;
324
325
} __attribute__ ((__packed__));

326
327
328
329
330
331
332
/*
 * leaves have an item area and a data area:
 * [item0, item1....itemN] [free space] [dataN...data1, data0]
 *
 * The data is separate from the items to get the keys closer together
 * during searches.
 */
333
struct btrfs_leaf {
334
	struct btrfs_header header;
335
	struct btrfs_item items[];
336
337
} __attribute__ ((__packed__));

338
339
340
341
/*
 * all non-leaf blocks are nodes, they hold only keys and pointers to
 * other blocks
 */
342
343
344
struct btrfs_key_ptr {
	struct btrfs_disk_key key;
	__le64 blockptr;
345
	__le64 generation;
346
347
} __attribute__ ((__packed__));

348
struct btrfs_node {
349
	struct btrfs_header header;
350
	struct btrfs_key_ptr ptrs[];
351
352
} __attribute__ ((__packed__));

353
/*
354
355
 * btrfs_paths remember the path taken from the root down to the leaf.
 * level 0 is always the leaf, and nodes[1...BTRFS_MAX_LEVEL] will point
356
357
358
359
360
 * to any other levels that are present.
 *
 * The slots array records the index of the item or block pointer
 * used while walking the tree.
 */
361
struct btrfs_path {
362
	struct extent_buffer *nodes[BTRFS_MAX_LEVEL];
363
	int slots[BTRFS_MAX_LEVEL];
364
365
	/* if there is real range locking, this locks field will change */
	int locks[BTRFS_MAX_LEVEL];
366
	int reada;
367
368
	/* keep some upper locks as we walk down */
	int keep_locks;
369
	int skip_locking;
370
	int lowest_level;
371
};
Chris Mason's avatar
Chris Mason committed
372

373
374
375
376
377
378
/*
 * items in the extent btree are used to record the objectid of the
 * owner of the block and the number of references
 */
struct btrfs_extent_item {
	__le32 refs;
379
380
381
382
383
384
} __attribute__ ((__packed__));

struct btrfs_extent_ref {
	__le64 root;
	__le64 generation;
	__le64 objectid;
Zheng Yan's avatar
Zheng Yan committed
385
	__le32 num_refs;
386
387
} __attribute__ ((__packed__));

388
389
/* dev extents record free space on individual devices.  The owner
 * field points back to the chunk allocation mapping tree that allocated
390
 * the extent.  The chunk tree uuid field is a way to double check the owner
391
392
 */
struct btrfs_dev_extent {
393
394
395
	__le64 chunk_tree;
	__le64 chunk_objectid;
	__le64 chunk_offset;
396
	__le64 length;
397
	u8 chunk_tree_uuid[BTRFS_UUID_SIZE];
398
399
} __attribute__ ((__packed__));

400
struct btrfs_inode_ref {
401
	__le64 index;
402
403
404
405
	__le16 name_len;
	/* name goes here */
} __attribute__ ((__packed__));

406
struct btrfs_timespec {
Chris Mason's avatar
Chris Mason committed
407
	__le64 sec;
Chris Mason's avatar
Chris Mason committed
408
409
410
	__le32 nsec;
} __attribute__ ((__packed__));

411
412
413
414
415
416
417
418
419
420
421
422
typedef enum {
	BTRFS_COMPRESS_NONE = 0,
	BTRFS_COMPRESS_ZLIB = 1,
	BTRFS_COMPRESS_LAST = 2,
} btrfs_compression_type;

/* we don't understand any encryption methods right now */
typedef enum {
	BTRFS_ENCRYPTION_NONE = 0,
	BTRFS_ENCRYPTION_LAST = 1,
} btrfs_encryption_type;

Chris Mason's avatar
Chris Mason committed
423
struct btrfs_inode_item {
424
	/* nfs style generation number */
Chris Mason's avatar
Chris Mason committed
425
	__le64 generation;
426
427
	/* transid that last touched this inode */
	__le64 transid;
Chris Mason's avatar
Chris Mason committed
428
	__le64 size;
429
	__le64 nbytes;
430
	__le64 block_group;
Chris Mason's avatar
Chris Mason committed
431
432
433
434
	__le32 nlink;
	__le32 uid;
	__le32 gid;
	__le32 mode;
435
	__le64 rdev;
Chris Mason's avatar
Chris Mason committed
436
437
	__le16 flags;
	__le16 compat_flags;
438

439
440
441
442
	struct btrfs_timespec atime;
	struct btrfs_timespec ctime;
	struct btrfs_timespec mtime;
	struct btrfs_timespec otime;
Chris Mason's avatar
Chris Mason committed
443
444
} __attribute__ ((__packed__));

445
446
447
448
struct btrfs_dir_log_item {
	__le64 end;
} __attribute__ ((__packed__));

449
struct btrfs_dir_item {
450
	struct btrfs_disk_key location;
451
	__le64 transid;
Josef Bacik's avatar
Josef Bacik committed
452
	__le16 data_len;
453
	__le16 name_len;
454
455
456
457
	u8 type;
} __attribute__ ((__packed__));

struct btrfs_root_item {
458
	struct btrfs_inode_item inode;
459
	__le64 generation;
460
	__le64 root_dirid;
461
462
463
	__le64 bytenr;
	__le64 byte_limit;
	__le64 bytes_used;
Yan Zheng's avatar
Yan Zheng committed
464
	__le64 last_snapshot;
465
	__le32 flags;
466
	__le32 refs;
467
468
	struct btrfs_disk_key drop_progress;
	u8 drop_level;
469
	u8 level;
470
} __attribute__ ((__packed__));
471

472
473
474
475
476
477
478
479
480
/*
 * this is used for both forward and backward root refs
 */
struct btrfs_root_ref {
	__le64 dirid;
	__le64 sequence;
	__le16 name_len;
} __attribute__ ((__packed__));

Yan Zheng's avatar
Yan Zheng committed
481
482
483
#define BTRFS_FILE_EXTENT_INLINE 0
#define BTRFS_FILE_EXTENT_REG 1
#define BTRFS_FILE_EXTENT_PREALLOC 2
484

485
struct btrfs_file_extent_item {
486
487
488
	/*
	 * transaction id that created this extent
	 */
489
	__le64 generation;
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
	/*
	 * max number of bytes to hold this extent in ram
	 * when we split a compressed extent we can't know how big
	 * each of the resulting pieces will be.  So, this is
	 * an upper limit on the size of the extent in ram instead of
	 * an exact limit.
	 */
	__le64 ram_bytes;

	/*
	 * 32 bits for the various ways we might encode the data,
	 * including compression and encryption.  If any of these
	 * are set to something a given disk format doesn't understand
	 * it is treated like an incompat flag for reading and writing,
	 * but not for stat.
	 */
	u8 compression;
	u8 encryption;
	__le16 other_encoding; /* spare for later use */

	/* are we inline data or a real extent? */
511
	u8 type;
512

513
514
515
516
	/*
	 * disk space consumed by the extent, checksum blocks are included
	 * in these numbers
	 */
517
518
	__le64 disk_bytenr;
	__le64 disk_num_bytes;
519
	/*
Chris Mason's avatar
Chris Mason committed
520
	 * the logical offset in file blocks (no csums)
521
522
523
524
525
526
527
	 * this extent record is for.  This allows a file extent to point
	 * into the middle of an existing extent on disk, sharing it
	 * between two snapshots (useful if some bytes in the middle of the
	 * extent have changed
	 */
	__le64 offset;
	/*
528
529
	 * the logical number of file blocks (no csums included).  This
	 * always reflects the size uncompressed and without encoding.
530
	 */
531
	__le64 num_bytes;
532

533
534
} __attribute__ ((__packed__));

Chris Mason's avatar
Chris Mason committed
535
struct btrfs_csum_item {
536
	u8 csum;
Chris Mason's avatar
Chris Mason committed
537
538
} __attribute__ ((__packed__));

539
540
541
542
/* different types of block groups (and chunks) */
#define BTRFS_BLOCK_GROUP_DATA     (1 << 0)
#define BTRFS_BLOCK_GROUP_SYSTEM   (1 << 1)
#define BTRFS_BLOCK_GROUP_METADATA (1 << 2)
543
#define BTRFS_BLOCK_GROUP_RAID0    (1 << 3)
544
#define BTRFS_BLOCK_GROUP_RAID1    (1 << 4)
545
#define BTRFS_BLOCK_GROUP_DUP	   (1 << 5)
Chris Mason's avatar
Chris Mason committed
546
#define BTRFS_BLOCK_GROUP_RAID10   (1 << 6)
Chris Mason's avatar
Chris Mason committed
547

Chris Mason's avatar
Chris Mason committed
548
549
struct btrfs_block_group_item {
	__le64 used;
550
551
	__le64 chunk_objectid;
	__le64 flags;
Chris Mason's avatar
Chris Mason committed
552
553
} __attribute__ ((__packed__));

554
555
556
557
558
struct btrfs_space_info {
	u64 flags;
	u64 total_bytes;
	u64 bytes_used;
	u64 bytes_pinned;
559
	u64 bytes_reserved;
560
	u64 bytes_readonly;
561
	int full;
562
	int force_alloc;
563
	struct list_head list;
564
565
566
567

	/* for block groups in our same type */
	struct list_head block_groups;
	spinlock_t lock;
568
	struct rw_semaphore groups_sem;
569
570
571
572
573
574
575
};

struct btrfs_free_space {
	struct rb_node bytes_index;
	struct rb_node offset_index;
	u64 offset;
	u64 bytes;
576
577
};

Chris Mason's avatar
Chris Mason committed
578
579
580
struct btrfs_block_group_cache {
	struct btrfs_key key;
	struct btrfs_block_group_item item;
581
	spinlock_t lock;
582
	struct mutex alloc_mutex;
583
	u64 pinned;
584
	u64 reserved;
585
586
	u64 flags;
	int cached;
587
	int ro;
588
589
590
591
592
593
594
595
596
597
598
599
600
	int dirty;

	struct btrfs_space_info *space_info;

	/* free space cache stuff */
	struct rb_root free_space_bytes;
	struct rb_root free_space_offset;

	/* block group cache stuff */
	struct rb_node cache_node;

	/* for block groups in the same raid type */
	struct list_head list;
Chris Mason's avatar
Chris Mason committed
601
};
602

Zheng Yan's avatar
Zheng Yan committed
603
604
605
606
607
608
struct btrfs_leaf_ref_tree {
	struct rb_root root;
	struct list_head list;
	spinlock_t lock;
};

609
struct btrfs_device;
610
struct btrfs_fs_devices;
611
struct btrfs_fs_info {
612
	u8 fsid[BTRFS_FSID_SIZE];
613
	u8 chunk_tree_uuid[BTRFS_UUID_SIZE];
614
615
	struct btrfs_root *extent_root;
	struct btrfs_root *tree_root;
616
617
	struct btrfs_root *chunk_root;
	struct btrfs_root *dev_root;
618
	struct btrfs_root *fs_root;
619
620
621

	/* the log root tree is a directory of all the other log roots */
	struct btrfs_root *log_root_tree;
622
	struct radix_tree_root fs_roots_radix;
623

624
625
626
627
	/* block group cache stuff */
	spinlock_t block_group_cache_lock;
	struct rb_root block_group_cache_tree;

628
629
630
	struct extent_io_tree pinned_extents;
	struct extent_io_tree pending_del;
	struct extent_io_tree extent_ins;
631

632
633
634
	/* logical->physical extent mapping */
	struct btrfs_mapping_tree mapping_tree;

635
	u64 generation;
636
	u64 last_trans_committed;
637
	u64 last_trans_new_blockgroup;
638
	u64 open_ioctl_trans;
639
	unsigned long mount_opt;
640
	u64 max_extent;
641
	u64 max_inline;
642
	u64 alloc_start;
Chris Mason's avatar
Chris Mason committed
643
	struct btrfs_transaction *running_transaction;
644
	wait_queue_head_t transaction_throttle;
645
	wait_queue_head_t transaction_wait;
646

647
	wait_queue_head_t async_submit_wait;
648
649
	wait_queue_head_t tree_log_wait;

650
	struct btrfs_super_block super_copy;
651
	struct btrfs_super_block super_for_commit;
652
	struct block_device *__bdev;
Chris Mason's avatar
Chris Mason committed
653
	struct super_block *sb;
654
	struct inode *btree_inode;
Chris Mason's avatar
Chris Mason committed
655
	struct backing_dev_info bdi;
656
	spinlock_t hash_lock;
Chris Mason's avatar
Chris Mason committed
657
	struct mutex trans_mutex;
658
	struct mutex tree_log_mutex;
659
660
	struct mutex transaction_kthread_mutex;
	struct mutex cleaner_mutex;
661
662
	struct mutex extent_ins_mutex;
	struct mutex pinned_mutex;
663
	struct mutex chunk_mutex;
664
	struct mutex drop_mutex;
665
	struct mutex volume_mutex;
Zheng Yan's avatar
Zheng Yan committed
666
	struct mutex tree_reloc_mutex;
Chris Mason's avatar
Chris Mason committed
667
	struct list_head trans_list;
668
	struct list_head hashers;
669
	struct list_head dead_roots;
670

671
	atomic_t nr_async_submits;
672
	atomic_t async_submit_draining;
673
	atomic_t nr_async_bios;
674
	atomic_t async_delalloc_pages;
675
676
677
678
	atomic_t tree_log_writers;
	atomic_t tree_log_commit;
	unsigned long tree_log_batch;
	u64 tree_log_transid;
679

680
681
682
683
684
685
	/*
	 * this is used by the balancing code to wait for all the pending
	 * ordered extents
	 */
	spinlock_t ordered_extent_lock;
	struct list_head ordered_extents;
686
	struct list_head delalloc_inodes;
687

688
689
690
691
692
693
	/*
	 * there is a pool of worker threads for checksumming during writes
	 * and a pool for checksumming after reads.  This is because readers
	 * can run with FS locks held, and the writers may be waiting for
	 * those locks.  We don't want ordering in the pending list to cause
	 * deadlocks, and so the two are serviced separately.
694
695
696
	 *
	 * A third pool does submit_bio to avoid deadlocking with the other
	 * two
697
698
	 */
	struct btrfs_workers workers;
699
	struct btrfs_workers delalloc_workers;
700
	struct btrfs_workers endio_workers;
701
	struct btrfs_workers endio_write_workers;
702
	struct btrfs_workers submit_workers;
703
704
705
706
707
708
	/*
	 * fixup workers take dirty pages that didn't properly go through
	 * the cow mechanism and make them safe to write.  It happens
	 * for the sys_munmap function call path
	 */
	struct btrfs_workers fixup_workers;
709
710
	struct task_struct *transaction_kthread;
	struct task_struct *cleaner_kthread;
711
	int thread_pool_size;
712

Zheng Yan's avatar
Zheng Yan committed
713
714
715
	/* tree relocation relocated fields */
	struct list_head dead_reloc_roots;
	struct btrfs_leaf_ref_tree reloc_ref_tree;
Zheng Yan's avatar
Zheng Yan committed
716
717
	struct btrfs_leaf_ref_tree shared_ref_tree;

718
719
	struct kobject super_kobj;
	struct completion kobj_unregister;
720
	int do_barriers;
721
	int closing;
722
	int log_root_recovering;
723
	atomic_t throttles;
724
	atomic_t throttle_gen;
725

726
	u64 total_pinned;
727
728
	struct list_head dirty_cowonly_roots;

729
	struct btrfs_fs_devices *fs_devices;
730
	struct list_head space_info;
731
	spinlock_t delalloc_lock;
732
	spinlock_t new_trans_lock;
733
	u64 delalloc_bytes;
734
	u64 last_alloc;
735
	u64 last_data_alloc;
736

Yan Zheng's avatar
Yan Zheng committed
737
738
739
	spinlock_t ref_cache_lock;
	u64 total_ref_cache_size;

740
741
742
743
744
745
	u64 avail_data_alloc_bits;
	u64 avail_metadata_alloc_bits;
	u64 avail_system_alloc_bits;
	u64 data_alloc_profile;
	u64 metadata_alloc_profile;
	u64 system_alloc_profile;
746
747

	void *bdev_holder;
748
};
749

750
751
/*
 * in ram representation of the tree.  extent_root is used for all allocations
752
 * and for the extent tree extent_root root.
753
 */
754
struct btrfs_dirty_root;
755
struct btrfs_root {
756
	struct extent_buffer *node;
757
758
759
760

	/* the node lock is held while changing the node pointer */
	spinlock_t node_lock;

761
	struct extent_buffer *commit_root;
Yan Zheng's avatar
Yan Zheng committed
762
	struct btrfs_leaf_ref_tree *ref_tree;
763
	struct btrfs_leaf_ref_tree ref_tree_struct;
764
	struct btrfs_dirty_root *dirty_root;
765
	struct btrfs_root *log_root;
Zheng Yan's avatar
Zheng Yan committed
766
	struct btrfs_root *reloc_root;
Yan Zheng's avatar
Yan Zheng committed
767

768
769
	struct btrfs_root_item root_item;
	struct btrfs_key root_key;
770
	struct btrfs_fs_info *fs_info;
771
772
	struct extent_io_tree dirty_log_pages;

773
774
	struct kobject root_kobj;
	struct completion kobj_unregister;
775
	struct mutex objectid_mutex;
776
	struct mutex log_mutex;
777

778
779
	u64 objectid;
	u64 last_trans;
780
781
782
783
784
785
786
787
788
789

	/* data allocations are done in sectorsize units */
	u32 sectorsize;

	/* node allocations are done in nodesize units */
	u32 nodesize;

	/* leaf allocations are done in leafsize units */
	u32 leafsize;

790
791
	u32 stripesize;

792
	u32 type;
Chris Mason's avatar
Chris Mason committed
793
794
	u64 highest_inode;
	u64 last_inode_alloc;
795
	int ref_cows;
796
	int track_dirty;
797
	u64 defrag_trans_start;
798
	struct btrfs_key defrag_progress;
799
	struct btrfs_key defrag_max;
800
801
	int defrag_running;
	int defrag_level;
802
	char *name;
803
	int in_sysfs;
804
805
806

	/* the dirty list is only used by non-reference counted roots */
	struct list_head dirty_list;
807

808
809
	spinlock_t list_lock;
	struct list_head dead_list;
810
	struct list_head orphan_list;
811
812
813
814
815
816

	/*
	 * right now this just gets used so that a root has its own devid
	 * for stat.  It may be used for more later
	 */
	struct super_block anon_super;
817
818
};

Chris Mason's avatar
Chris Mason committed
819
/*
820

Chris Mason's avatar
Chris Mason committed
821
822
823
824
 * inode items have the data typically returned from stat and store other
 * info about object characteristics.  There is one for every file and dir in
 * the FS
 */
Chris Mason's avatar
Chris Mason committed
825
#define BTRFS_INODE_ITEM_KEY		1
826
827
828
#define BTRFS_INODE_REF_KEY		12
#define BTRFS_XATTR_ITEM_KEY		24
#define BTRFS_ORPHAN_ITEM_KEY		48
Chris Mason's avatar
Chris Mason committed
829
/* reserve 2-15 close to the inode for later flexibility */
Chris Mason's avatar
Chris Mason committed
830
831
832
833
834

/*
 * dir items are the name -> inode pointers in a directory.  There is one
 * for every name in a directory.
 */
835
836
837
838
#define BTRFS_DIR_LOG_ITEM_KEY  60
#define BTRFS_DIR_LOG_INDEX_KEY 72
#define BTRFS_DIR_ITEM_KEY	84
#define BTRFS_DIR_INDEX_KEY	96
Chris Mason's avatar
Chris Mason committed
839
/*
Chris Mason's avatar
Chris Mason committed
840
 * extent data is for file data
Chris Mason's avatar
Chris Mason committed
841
 */
842
#define BTRFS_EXTENT_DATA_KEY	108
Chris Mason's avatar
Chris Mason committed
843
844
845
/*
 * csum items have the checksums for data in the extents
 */
846
#define BTRFS_CSUM_ITEM_KEY	120
Chris Mason's avatar
Chris Mason committed
847

848
849

/* reserve 21-31 for other file/dir stuff */
Chris Mason's avatar
Chris Mason committed
850

Chris Mason's avatar
Chris Mason committed
851
852
853
854
/*
 * root items point to tree roots.  There are typically in the root
 * tree used by the super block to find all the other trees
 */
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
#define BTRFS_ROOT_ITEM_KEY	132

/*
 * root backrefs tie subvols and snapshots to the directory entries that
 * reference them
 */
#define BTRFS_ROOT_BACKREF_KEY	144

/*
 * root refs make a fast index for listing all of the snapshots and
 * subvolumes referenced by a given root.  They point directly to the
 * directory item in the root that references the subvol
 */
#define BTRFS_ROOT_REF_KEY	156

Chris Mason's avatar
Chris Mason committed
870
871
872
873
/*
 * extent items are in the extent map tree.  These record which blocks
 * are used, and how many references there are to each block
 */
874
875
#define BTRFS_EXTENT_ITEM_KEY	168
#define BTRFS_EXTENT_REF_KEY	180
Chris Mason's avatar
Chris Mason committed
876
877
878
879
880

/*
 * block groups give us hints into the extent allocation trees.  Which
 * blocks are free etc etc
 */
881
#define BTRFS_BLOCK_GROUP_ITEM_KEY 192
882

883
884
885
#define BTRFS_DEV_EXTENT_KEY	204
#define BTRFS_DEV_ITEM_KEY	216
#define BTRFS_CHUNK_ITEM_KEY	228
886

Chris Mason's avatar
Chris Mason committed
887
888
889
890
/*
 * string items are for debugging.  They just store a short string of
 * data in the FS
 */
Chris Mason's avatar
Chris Mason committed
891
892
#define BTRFS_STRING_ITEM_KEY	253

893
894
895
#define BTRFS_MOUNT_NODATASUM		(1 << 0)
#define BTRFS_MOUNT_NODATACOW		(1 << 1)
#define BTRFS_MOUNT_NOBARRIER		(1 << 2)
896
#define BTRFS_MOUNT_SSD			(1 << 3)
897
#define BTRFS_MOUNT_DEGRADED		(1 << 4)
898
#define BTRFS_MOUNT_COMPRESS		(1 << 5)
899
900
901
902
903

#define btrfs_clear_opt(o, opt)		((o) &= ~BTRFS_MOUNT_##opt)
#define btrfs_set_opt(o, opt)		((o) |= BTRFS_MOUNT_##opt)
#define btrfs_test_opt(root, opt)	((root)->fs_info->mount_opt & \
					 BTRFS_MOUNT_##opt)
Yan's avatar
Yan committed
904
905
906
/*
 * Inode flags
 */
Yan's avatar
Yan committed
907
908
909
#define BTRFS_INODE_NODATASUM		(1 << 0)
#define BTRFS_INODE_NODATACOW		(1 << 1)
#define BTRFS_INODE_READONLY		(1 << 2)
910
#define BTRFS_INODE_NOCOMPRESS		(1 << 3)
Yan Zheng's avatar
Yan Zheng committed
911
#define BTRFS_INODE_PREALLOC		(1 << 4)
Yan's avatar
Yan committed
912
913
914
915
916
917
#define btrfs_clear_flag(inode, flag)	(BTRFS_I(inode)->flags &= \
					 ~BTRFS_INODE_##flag)
#define btrfs_set_flag(inode, flag)	(BTRFS_I(inode)->flags |= \
					 BTRFS_INODE_##flag)
#define btrfs_test_flag(inode, flag)	(BTRFS_I(inode)->flags & \
					 BTRFS_INODE_##flag)
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
/* some macros to generate set/get funcs for the struct fields.  This
 * assumes there is a lefoo_to_cpu for every type, so lets make a simple
 * one for u8:
 */
#define le8_to_cpu(v) (v)
#define cpu_to_le8(v) (v)
#define __le8 u8

#define read_eb_member(eb, ptr, type, member, result) (			\
	read_extent_buffer(eb, (char *)(result),			\
			   ((unsigned long)(ptr)) +			\
			    offsetof(type, member),			\
			   sizeof(((type *)0)->member)))

#define write_eb_member(eb, ptr, type, member, result) (		\
	write_extent_buffer(eb, (char *)(result),			\
			   ((unsigned long)(ptr)) +			\
			    offsetof(type, member),			\
			   sizeof(((type *)0)->member)))

938
#ifndef BTRFS_SETGET_FUNCS
939
#define BTRFS_SETGET_FUNCS(name, type, member, bits)			\
940
941
942
u##bits btrfs_##name(struct extent_buffer *eb, type *s);		\
void btrfs_set_##name(struct extent_buffer *eb, type *s, u##bits val);
#endif
943
944
945
946

#define BTRFS_SETGET_HEADER_FUNCS(name, type, member, bits)		\
static inline u##bits btrfs_##name(struct extent_buffer *eb)		\
{									\
David Miller's avatar
David Miller committed
947
948
949
	type *p = kmap_atomic(eb->first_page, KM_USER0);		\
	u##bits res = le##bits##_to_cpu(p->member);			\
	kunmap_atomic(p, KM_USER0);					\
950
	return res;							\
951
952
953
954
}									\
static inline void btrfs_set_##name(struct extent_buffer *eb,		\
				    u##bits val)			\
{									\
David Miller's avatar
David Miller committed
955
956
957
	type *p = kmap_atomic(eb->first_page, KM_USER0);		\
	p->member = cpu_to_le##bits(val);				\
	kunmap_atomic(p, KM_USER0);					\
958
}
Chris Mason's avatar
Chris Mason committed
959

960
961
962
963
964
965
966
967
#define BTRFS_SETGET_STACK_FUNCS(name, type, member, bits)		\
static inline u##bits btrfs_##name(type *s)				\
{									\
	return le##bits##_to_cpu(s->member);				\
}									\
static inline void btrfs_set_##name(type *s, u##bits val)		\
{									\
	s->member = cpu_to_le##bits(val);				\
Chris Mason's avatar
Chris Mason committed
968
969
}

970
971
972
973
974
975
976
BTRFS_SETGET_FUNCS(device_type, struct btrfs_dev_item, type, 64);
BTRFS_SETGET_FUNCS(device_total_bytes, struct btrfs_dev_item, total_bytes, 64);
BTRFS_SETGET_FUNCS(device_bytes_used, struct btrfs_dev_item, bytes_used, 64);
BTRFS_SETGET_FUNCS(device_io_align, struct btrfs_dev_item, io_align, 32);
BTRFS_SETGET_FUNCS(device_io_width, struct btrfs_dev_item, io_width, 32);
BTRFS_SETGET_FUNCS(device_sector_size, struct btrfs_dev_item, sector_size, 32);
BTRFS_SETGET_FUNCS(device_id, struct btrfs_dev_item, devid, 64);
977
978
979
BTRFS_SETGET_FUNCS(device_group, struct btrfs_dev_item, dev_group, 32);
BTRFS_SETGET_FUNCS(device_seek_speed, struct btrfs_dev_item, seek_speed, 8);
BTRFS_SETGET_FUNCS(device_bandwidth, struct btrfs_dev_item, bandwidth, 8);
Yan Zheng's avatar
Yan Zheng committed
980
BTRFS_SETGET_FUNCS(device_generation, struct btrfs_dev_item, generation, 64);
981

982
983
984
985
986
987
988
989
990
991
992
993
BTRFS_SETGET_STACK_FUNCS(stack_device_type, struct btrfs_dev_item, type, 64);
BTRFS_SETGET_STACK_FUNCS(stack_device_total_bytes, struct btrfs_dev_item,
			 total_bytes, 64);
BTRFS_SETGET_STACK_FUNCS(stack_device_bytes_used, struct btrfs_dev_item,
			 bytes_used, 64);
BTRFS_SETGET_STACK_FUNCS(stack_device_io_align, struct btrfs_dev_item,
			 io_align, 32);
BTRFS_SETGET_STACK_FUNCS(stack_device_io_width, struct btrfs_dev_item,
			 io_width, 32);
BTRFS_SETGET_STACK_FUNCS(stack_device_sector_size, struct btrfs_dev_item,
			 sector_size, 32);
BTRFS_SETGET_STACK_FUNCS(stack_device_id, struct btrfs_dev_item, devid, 64);
994
995
996
997
998
999
BTRFS_SETGET_STACK_FUNCS(stack_device_group, struct btrfs_dev_item,
			 dev_group, 32);
BTRFS_SETGET_STACK_FUNCS(stack_device_seek_speed, struct btrfs_dev_item,
			 seek_speed, 8);
BTRFS_SETGET_STACK_FUNCS(stack_device_bandwidth, struct btrfs_dev_item,
			 bandwidth, 8);
Yan Zheng's avatar
Yan Zheng committed
1000
1001
BTRFS_SETGET_STACK_FUNCS(stack_device_generation, struct btrfs_dev_item,
			 generation, 64);
1002

1003
1004
1005
1006
1007
static inline char *btrfs_device_uuid(struct btrfs_dev_item *d)
{
	return (char *)d + offsetof(struct btrfs_dev_item, uuid);
}

Yan Zheng's avatar
Yan Zheng committed
1008