On-disk Format 磁盘格式 

This document describes the Btrfs on‐disk format.
本文档描述了 Btrfs 的磁盘格式。

Note 注意

This document contains outdated and incomplete information and has been copied from the original btrfs.wiki.kernel.org with little review.
本文档包含过时和不完整的信息,并且是从原始的 btrfs.wiki.kernel.org 复制而来,经过了很少的审查。

Overview 概述

Aside from the superblock, Btrfs consists entirely of several trees. The trees use copy-on-write. Trees are stored in nodes, each of with belong to a level in the b-tree structure. Internal nodes contain references to other internal nodes on the next level, or to leaf nodes then the level reaches zero. Leaf nodes contain various types of data structures, depending on the tree.
除了超级块外,Btrfs 完全由几棵树组成。这些树使用写时复制。树存储在节点中,每个节点都属于 B 树结构中的一个级别。内部节点包含对下一级别的其他内部节点的引用,或者在级别达到零时对叶节点的引用。叶节点包含各种类型的数据结构,取决于树的类型。

Btrfs makes a distinction between logical and physical addresses. Logical addresses are used in the filesystem structures, while physical addresses are simply byte offsets on a disk. One logical address may correspond to physical addresses on any number of disks, depending on RAID settings. The chunk tree is used to convert from logical addresses to physical addresses; the dev tree can be used for the reverse.
Btrfs 在逻辑地址和物理地址之间进行区分。逻辑地址用于文件系统结构,而物理地址仅是磁盘上的字节偏移量。一个逻辑地址可能对应于任意数量的磁盘上的物理地址,这取决于 RAID 设置。块树用于将逻辑地址转换为物理地址;dev 树可用于反向操作。

For bootstrapping purposes, the superblock contains a subset of the chunk tree, specifically it contains “chunk items” for all system chunks. The superblock also contains a logical reference to root nodes in the root and chunk trees, which can then be used to locate all the other trees and data stored.
为引导目的,超级块包含块树的子集,具体来说,它包含所有系统块的“块项”。超级块还包含对根节点在根和块树中的逻辑引用,然后可以用于定位所有其他树和存储的数据。

TODO Subvolumes and snapshots.
待办事项:子卷和快照。

Basic Structures 基本结构

Note that the fields are unsigned, so object ID −1 will be treated as 0xffffffffffffffff and sorted to the end of the tree. Since Btrfs uses little‐endian, a simple byte‐by‐byte comparison of KEYs will not work.
请注意,字段是无符号的,因此对象 ID -1 将被视为 0xffffffffffffffff 并排序到树的末尾。由于 Btrfs 使用小端序,对 KEY 的简单逐字节比较将不起作用。

Off

Size

Type

Description

0

8

UINT

Object ID. Each tree has its own set of Object IDs.
对象 ID。每个树都有自己的对象 ID 集合。

8

1

UINT

Item type. 项目类型。

9

8

UINT

Offset. The meaning depends on the item type.
偏移。其含义取决于项目类型。

11

Btrfs uses Unix time. Btrfs 使用 Unix 时间。

Off

Size

Type

Description

0

8

SINT

Number of seconds since 1970-01-01T00:00:00Z.
从 1970-01-01T00:00:00Z 开始的秒数。

8

4

UINT

Number of nanoseconds since the beginning of the second.
从秒开始的纳秒数。

c

Superblock 超级块 

The primary superblock is located at 0x10000 (64KiB). Mirror copies of the superblock are located at physical addresses 0x4000000 (64 MiB) and 0x4000000000 (256GiB), if these locations are valid. Superblock copies are updated simultaneously. During mount btrfs’ kernel module reads only the first super block (at 64KiB), if an error is detected mounting fails.
主超级块位于 0x10000(64KiB)处。如果这些位置有效,则超级块的镜像副本位于物理地址 0x4000000(64 MiB)和 0x4000000000(256GiB)处。超级块副本同时更新。在挂载 btrfs 时,内核模块只读取第一个超级块(在 64KiB 处),如果检测到错误,则挂载失败。

Note that btrfs only recognizes disks with a valid 0x1 0000 superblock; otherwise, there would be confusion with other filesystems.
请注意,btrfs 仅识别具有有效 0x1 0000 超级块的磁盘;否则,将与其他文件系统混淆。

TODO 待办事项

Off

Size

Type

Description

0

20

CSUM

Checksum of everything past this field (from 20 to 1000)
从此字段开始的所有内容的校验和(从 20 到 1000)

20

10

UUID

FS UUID

30

8

UINT

physical address of this block (different for mirrors)
该块的物理地址(镜像不同)

38

8

flags

40

8

ASCII

magic (“_BHRfS_M”) 魔法(“_BHRfS_M”)

48

8

generation

50

8

logical address of the root tree root
根树根的逻辑地址

58

8

logical address of the chunk tree root
块树根的逻辑地址

60

8

logical address of the log tree root
日志树根的逻辑地址

68

8

log_root_transid

70

8

total_bytes

78

8

bytes_used

80

8

root_dir_objectid (usually 6)
root_dir_objectid(通常为 6)

88

8

num_devices

90

4

sectorsize

94

4

nodesize

98

4

leafsize

9c

4

stripesize

a0

4

sys_chunk_array_size

a4

8

chunk_root_generation

ac

8

compat_flags

b4

8

compat_ro_flags - only implementations that support the flags can write to the filesystem
compat_ro_flags - 只有支持这些标志的实现才能写入文件系统

bc

8

incompat_flags - only implementations that support the flags can use the filesystem
incompat_flags - 只有支持这些标志的实现才能使用文件系统

c4

2

csum_type - Btrfs currently uses the CRC32c little-endian hash function with seed -1.
csum_type - Btrfs 目前使用 CRC32c 小端哈希函数,种子为 -1。

c6

1

root_level

c7

1

chunk_root_level

c8

1

log_root_level

c9

62

DEV_ITEM data for this device
此设备的 DEV_ITEM 数据

12b

100

label (may not contain ‘/’ or ‘\\’)
标签(可能不包含 '/' 或 '\\')

22b

8

cache_generation

233

8

uuid_tree_generation

23b

f0

reserved /* future expansion */
保留 /* 未来扩展 */

32b

800

sys_chunk_array:(n bytes valid) Contains (KEY, CHUNK_ITEM) pairs for all SYSTEM chunks. This is needed to bootstrap the mapping from logical addresses to physical.
sys_chunk_array:(n 字节有效) 包含所有系统块的 (KEY, CHUNK_ITEM) 对。这对于引导从逻辑地址到物理地址的映射是必要的。

b2b

2a0

Contain super_roots (4 btrfs_root_backup)
包含 super_roots(4 个 btrfs_root_backup)

dcb

235

current unused 当前未使用

1000

Internal Node 内部节点 

In internal nodes, the node header is followed by a number of key pointers.
在内部节点中,节点头部后面跟着一些键指针。

Off

Size

Type

Description

0

11

KEY

key

11

8

UINT

block number 块编号

19

8

UINT

generation

21

header

key ptr 键指针

key ptr 键指针

key ptr 键指针

free space 空闲空间

Leaf Node 叶节点

In leaf nodes, the node header is followed by a number of items. The items’ data is stored at the end of the node, and the contents of the item data depends on the item type stored in the key.
在叶节点中,节点头部后面跟着一些项目。项目的数据存储在节点的末尾,项目数据的内容取决于键中存储的项目类型。

Off

Size

Type

Description

0

11

KEY

key

11

4

UINT

data offset relative to end of header (65)
相对于标头末尾的数据偏移量(65)

15

4

UINT

data size 数据大小

19

header

item 0 项目 0

item 1 项目 1

item N 项目 N

free space 空闲空间

data N 数据 N

data 1 数据 1

data 0 数据 0

Object Types 对象类型 

TODO 待办事项

Objects 对象 

ROOT_TREE (1)

The root tree holds ROOT_ITEMs, ROOT_REFs, and ROOT_BACKREFs for every tree other than itself. It is used to find the other trees and to determine the subvolume structure. It also holds the items for the root tree directory. The logical address of the root tree is stored in the superblock.
根树包含 ROOT_ITEMs、ROOT_REFs 和 ROOT_BACKREFs,用于存储除自身以外的每棵树。它用于查找其他树并确定子卷结构。它还保存了根树目录的项目。根树的逻辑地址存储在超级块中。

Reserved objectids 保留的对象 ID。

There are several well-known objectids that refer to internal trees.
有几个众所周知的 objectids 指的是内部树。

All root objectids between BTRFS_FIRST_FREE_OBJECTID = 256ULL and BTRFS_LAST_FREE_OBJECTID = -256ULL refer to file trees.
所有根 objectids 在 BTRFS_FIRST_FREE_OBJECTID = 256ULLBTRFS_LAST_FREE_OBJECTID = -256ULL 之间指的是文件树。

Otherwise, the objectid should be considered reserved for internal use.
否则,该 objectid 应被视为内部使用保留。

  • BTRFS_ROOT_TREE_OBJECTID = 1

    The object id that refers to the ROOT_TREE itself.
    指向 ROOT_TREE 本身的对象 ID。

  • BTRFS_EXTENT_TREE_OBJECTID = 2

    The objectid that refers to the EXTENT_TREE
    指向 EXTENT_TREE 的 objectid

  • BTRFS_CHUNK_TREE_OBJECTID = 3

    The objectid that refers to the root of the CHUNK_TREE
    指向 CHUNK_TREE 根目录的 objectid

  • BTRFS_DEV_TREE_OBJECTID = 4

    The objectid that refers to the root of the DEV_TREE
    指向 DEV_TREE 根目录的 objectid

  • BTRFS_FS_TREE_OBJECTID = 5

    The objectid that refers to the global FS_TREE root.
    指向全局 FS_TREE 根的 objectid。

  • BTRFS_CSUM_TREE_OBJECTID = 7

    The objectid that refers to the CSUM_TREE
    指向 CSUM_TREE 的 objectid

  • BTRFS_QUOTA_TREE_OBJECTID = 8

    The objectid that refers to the QUOTA_TREE
    指向 QUOTA_TREE 的 objectid

  • BTRFS_UUID_TREE_OBJECTID = 9

    The objectid that refers to the UUID_TREE.
    指向 UUID_TREE 的 objectid。

  • BTRFS_FREE_SPACE_TREE_OBJECTID = 10

    The objectid that refers to the FREE_SPACE_TREE.
    指向 FREE_SPACE_TREE 的 objectid。

  • BTRFS_TREE_LOG_OBJECTID = -7ULL

    The objectid that refers to the TREE_LOG tree.
    指向 TREE_LOG 树的 objectid。

  • BTRFS_TREE_RELOC_OBJECTID = -8ULL

    The objectid that refers to the TREE_RELOC tree.
    指向 TREE_RELOC 树的 objectid。

  • BTRFS_DATA_RELOC_TREE_OBJECTID = -9ULL

    The objectid that refers to the DATA_RELOC tree.
    指向 DATA_RELOC 树的 objectid。

The following are well-known objectids within the ROOT_TREE that do not refer to other trees.
以下是 ROOT_TREE 中众所周知的 objectids,它们不指向其他树。

  • BTRFS_ROOT_TREE_DIR_OBJECTID = 6

    The objectid that refers to the directory within the root tree. If it exists, it will have the usual items used to implement a directory associated with it. There will only be a single entry called default that points to a key to be used as the root directory on the file system instead of the FS_TREE.
    指向根树中目录的 objectid。如果存在,它将具有用于实现与之关联的目录的常用项目。只会有一个名为 default 的条目,指向一个用作文件系统根目录而不是 FS_TREE 的密钥。

  • BTRFS_ORPHAN_OBJECTID = -5ULL

    The objectid used for orphan root tracking.
    用于孤立根跟踪的 objectid。

Developer note: If implementing a feature that requires a new objectid in the reserved range, you must reserve the objectid via the mailing list before posting your code for general use. This is a disk format change.
开发者注意:如果要实现需要在保留范围内使用新 objectid 的功能,您必须通过邮件列表预留 objectid,然后再发布您的代码供一般使用。这是磁盘格式更改。

Orphans 孤儿

Removing a root is a multi-step process that may involve many transactions. References to every extent used by the tree must be decremented and, if they hit zero, the extents must be released. It is possible that the system crashes, loses power, or otherwise encounters an error during root removal. Without additional information, the file system could ultimately contain partially removed roots, which would make it inconsistent. When a root is removed, it performs several small operations in a single transaction in preparation for removal. This process should be familiar to those with an understanding of how orphans work when an inode is unlinked on any UNIX-style file system.
删除根是一个多步骤过程,可能涉及许多事务。必须递减树使用的每个范围的引用,并且如果它们达到零,则必须释放这些范围。在删除根时,系统可能会崩溃、断电或在其他情况下遇到错误。如果没有额外信息,文件系统最终可能包含部分删除的根,这将使其不一致。当删除根时,它会在准备删除的单个事务中执行几个小操作。对于了解当在任何 UNIX 风格文件系统上取消链接 inode 时孤儿如何工作的人来说,这个过程应该很熟悉。

  1. Unlink the root from the directory that contains it.
    从包含它的目录中取消链接根。

  2. Initialize the drop_progress and drop_level fields and set the refs field to 0 in the ROOT_ITEM.
    ROOT_ITEM 中初始化 drop_progressdrop_level 字段,并将 refs 字段设置为 0

  3. If an orphan key for this root has not already been inserted into the tree, insert one.
    如果此根节点的孤立键尚未插入树中,请插入一个。

  4. Remove the UUID entries for this root and any associated received root from the UUID_TREE.
    UUID_TREE 中删除此根节点的 UUID 条目以及任何相关的接收根节点。

Ultimately, the cleaner thread handles the reference count adjustments and, once that is complete, the root has been successfully removed and it removes the orphan key for that root. As the cleaner progresses, the drop_progress and drop_level fields are updated to reflect the most recently processed item.
最终,清理线程处理引用计数调整,一旦完成,根已成功移除,并移除该根的孤立键。随着清理的进行, drop_progressdrop_level 字段将更新以反映最近处理的项目。

This process may be interrupted at any time and it must be recoverable. The orphan key is how btrfs avoids inconsistencies when that occurs. The orphan key is located in the ROOT_TREE and is of the following form.
此过程可能随时中断,必须是可恢复的。孤立键是 btrfs 在发生中断时避免不一致性的方式。孤立键位于 ROOT_TREE 中,其形式如下。

struct btrfs_key 结构 btrfs_key

objectid

BTRFS_ORPHAN_OBJECTID [-5ULL]

  • There is no item body associated with this key. All required information is contained within the key itself and the ROOT_ITEM associated with the objectid contained in offset
    与此键关联的项目主体不存在。所有必需信息都包含在键本身和 offset 中包含的 objectid 相关联的 ROOT_ITEM 中。

When the file system is mounted again after failure, the ROOT_TREE is searched for all orphan keys and the process is resumed for each one using the drop_progress and drop_level fields in the ROOT_ITEM.
当文件系统在故障后再次挂载时,将搜索 ROOT_TREE 以查找所有孤立键,并使用 ROOT_ITEM 中的 drop_progressdrop_level 字段为每个键恢复进程。

EXTENT tree (2) EXTENT 树(2)

TODO 待办事项

  • Holds EXTENT_ITEMs, BLOCK_GROUP_ITEMs
    持有 EXTENT_ITEMs,BLOCK_GROUP_ITEMs

  • Pointed to by ROOT 被 ROOT 指向

EMPTY_SUBVOL dir (2) 空的子卷目录 (2) 

TODO 待办事项

CHUNK_TREE (3) 块树 (3) 

The chunk tree holds all DEV_ITEMs and CHUNK_ITEMs, making it possible to determine the device(s) and physical address(es) corresponding to a given logical address. It is therefore crucial for access to the contents of the filesystem.
块树包含所有的 DEV_ITEM 和 CHUNK_ITEM,从而可以确定与给定逻辑地址对应的设备和物理地址。因此,对于访问文件系统内容至关重要。

The chunk tree resides entirely in SYSTEM block groups, and will therefore be accessible from the CHUNK_ITEM array in the Superblock. It also has an entry in the ROOT tree.
块树完全驻留在系统块组中,因此可以从超级块中的 CHUNK_ITEM 数组中访问。它还在 ROOT 树中有一个条目。

Reserved objectids 保留的对象 ID。

  • BTRFS_FIRST_CHUNK_TREE_OBJECTID = 256

    This objectid indicates the first available objectid in this CHUNK_TREE. In practice, it is the only objectid used in the tree. The offset field of the key is the only component used to distinguish separate `CHUNK_ITEM <#CHUNK_ITEM>`__ items.
    此 objectid 指示此 CHUNK_TREE 中第一个可用的 objectid。实际上,它是树中唯一使用的 objectid。键的 offset 字段是用于区分单独的 `CHUNK_ITEM <#CHUNK_ITEM>`__ 项的唯一组件。

Dev tree (4) Dev 树 (4) 

The dev tree holds all DEV_EXTENTs, making it possible to determine the logical address corresponding to a given physical address. This is necessary when shrinking or removing devices. The dev tree has an entry in the root tree.
dev 树保存所有 DEV_EXTENT,从而可以确定给定物理地址对应的逻辑地址。在缩小或删除设备时,这是必要的。dev 树在根树中有一个条目。

FS_TREE (5) FS_TREE(5)

TODO 待办事项

  • Holds INODE_ITEM, INODE_REF, DIR_ITEM, DIR_INDEXen, XATTR_ITEMs, EXTENT_DATA for a filesystem
    为文件系统保存 INODE_ITEMINODE_REFDIR_ITEM ,DIR_INDEXen,XATTR_ITEMs, EXTENT_DATA

  • Pointed to by ROOT 被 ROOT 指向

  • TODO: “..” 待办事项: “..”

Root tree directory 根树目录

The root tree directory is stored in the root tree. It has an INODE_ITEM and a DIR_ITEM with name “default” pointing to the FS tree. There is also a corresponding INODE_REF, but no DIR_INDEX. The objectid of the root tree directory is stored in the superblock, but is currently always 6.
根树目录存储在根树中。它具有一个名为“default”的 INODE_ITEM 和 DIR_ITEM,指向 FS 树。还有一个相应的 INODE_REF,但没有 DIR_INDEX。根树目录的 objectid 存储在超级块中,但目前始终为 6。

Checksum tree (7) 校验树(7)

The checksum tree contains all the EXTENT_CSUMs. It has an entry in the root tree.
校验和树包含所有的 EXTENT_CSUMs。它在根树中有一个条目。

ORPHAN (-5) 孤立块 (-5)

TODO 待办事项

TREE_LOG (-6) 树木日志 (-6) 

TODO 待办事项

TREE_LOG_FIXUP (-7) 树木日志修正 (-7) 

TODO 待办事项

TREE_RELOC (-8)

TODO 待办事项

  • Just a copy of another tree
    只是另一棵树的副本

DATA_RELOC tree (-9) DATA_RELOC 树 (-9) 

TODO 待办事项

  • Holds 100 INODE_ITEM 0 拥有 100 个 INODE_ITEM 0

  • Holds 100 INODE_REF 100 0:’..’
    拥有 100 个 INODE_REF 100 0:’..’

  • Pointed to by ROOT 被 ROOT 指向

EXTENT_CSUM (-a)

TODO 待办事项

MULTIPLE_OBJECTIDS (-100)
多个对象 ID (-100) 

TODO 待办事项

Item Types 项目类型

INODE_ITEM (01) INODE_ITEM (01)

Location 位置 

INODE_ITEM items are located primarily in file trees but are also found in the ROOT_TREE to implement the free space cache (v1).
INODE_ITEM 项主要位于文件树中,但也可以在 ROOT_TREE 中找到,以实现空闲空间缓存(v1)。

Usage 用法 

struct btrfs_key 结构 btrfs_key

objectid

objectid (Used as inode number)
objectid(用作 inode 号)

Description 描述 

Contains the stat information for an inode; see stat(2).
包含 i 节点的 stat 信息;请参阅 stat(2)。

Item Contents 项目内容

INODE_ITEM items contain a single btrfs_inode_item structure.
INODE_ITEM 项包含一个单一的 btrfs_inode_item 结构。

INODE_REF (0c)

(inode_id, directory_id) TODO
(inode_id, directory_id) 待办事项

From an inode to a name in a directory.
从一个 inode 到目录中的名称。

Off

Size

Type

Description

0

8

UINT

index in the directory 目录中的索引

8

2

UINT

(n)

a

n

ASCII

name in the directory 目录中的名称

a+n

This structure can be repeated…?
这个结构可以重复吗?

INODE_EXTREF (0d)

(inode_id, hash of name [using directory object ID as seed]) TODO
(inode_id, 使用目录对象 ID 作为种子的名称哈希) 待办事项

From an inode to a name in a directory. Used if the regarding INODE_REF array ran out of space. This item requires the EXTENDED_IREF feature.
从一个 inode 到目录中的名称。如果相关的 INODE_REF 数组空间不足,则使用。此项需要 EXTENDED_IREF 功能。

Off

Size

Type

Description

0

8

UINT

directory object ID 目录对象标识符

8

8

UINT

index in the directory 目录中的索引

10

2

UINT

(n)

12

n

ASCII

name in the directory 目录中的名称

12+n

This structure can be repeated…?
这个结构可以重复吗?

XATTR_ITEM (18)

Location 位置 

XATTR_ITEM items are only located in file trees.
XATTR_ITEM 项仅位于文件树中。

Usage 用法 

struct btrfs_key

objectid

objectid of owning inode

Description 描述 

XATTR_ITEM items contain extended attributes. Each name is hashed using the name hash and that value is used in the key for locating the entry quickly. Each XATTR_ITEM item contains one or more extended attributes with names represented by the same hash. All extended attributes that share the same name hash must fit in a single leaf.
XATTR_ITEM 项包含扩展属性。每个名称都使用名称哈希进行哈希处理,该值用于在定位条目时快速查找键。每个 XATTR_ITEM 项包含一个或多个由相同哈希表示的名称的扩展属性。所有共享相同名称哈希的扩展属性必须适合单个叶子。

Item Contents 项目内容

XATTR_ITEM items consist of a series of one or more extended attribute entries with names that share a hash value. Each entry consists of a btrfs_dir_item structure immediately followed by the name and the attribute data. The length of each name is contained in btrfs_dir_item.name_len. The data payload begins immediately after the name. The data payload length is contained in btrfs_dir_item.data_len btrfs_dir_item.data_len.location is unused and must be zeroed. btrfs_dir_item.type contains a shorthand value referring to the type of item to which an entry refers it must always be be BTRFS_FT_XATTR when used to describe an extended attribute.
XATTR_ITEM 项由一系列一个或多个具有共享哈希值的名称的扩展属性条目组成。每个条目由一个 btrfs_dir_item 结构紧随其后,然后是名称和属性数据。每个名称的长度包含在 btrfs_dir_item.name_len 中。数据有效载荷紧随名称之后开始。数据有效载荷长度包含在 btrfs_dir_item.data_len 中。 btrfs_dir_item.data_len.location 未使用,必须清零。 btrfs_dir_item.type 包含一个简写值,指示条目所指的项目类型,用于描述扩展属性时必须始终为 BTRFS_FT_XATTR

When there is more than one entry for a single hash value, the offset of each entry must be calculating using the lengths of the preceding entries including names and data.
当一个哈希值有多个条目时,每个条目的偏移量必须使用前面条目的长度(包括名称和数据)进行计算。

For more details, please see: struct btrfs_dir_item and `DIR_ITEM.
更多详情,请参阅: struct btrfs_dir_item`DIR_ITEM

VERITY_DESC (24)

Location 位置 

VERITY_DESC items are located in the FS_TREE. TODO
VERITY_DESC 项位于 FS_TREE 中。待办事项

VERITY_MERKLE (25)

Location 位置 

VERITY_MERKLE items are located in the FS_TREE. TODO
VERITY_MERKLE 项位于 FS_TREE 中。待办事项

ORPHAN_ITEM (30) 孤立项 (30) 

(-5, objid of orphan inode) TODO
(-5,孤立索引节点的 objid) 待办事项

``   Empty.`` ``   空的。``

DIR_LOG_ITEM (3c)

(directory_id, first offset) TODO
(目录 ID,第一个偏移) 待办事项

``   The log is considered authoritative for ([first offset, end offset)]``
`` 日志被认为是([第一个偏移量,结束偏移量)]的权威来源``
``    0  8 UINT   end offset``
``    0  8 UINT   结束偏移``

DIR_LOG_INDEX (48)

(directory_id, first offset) TODO
(目录 ID,第一个偏移) 待办事项

``   Same as DIR_LOG_ITEM.``
``与 DIR_LOG_ITEM 相同。``

DIR_ITEM (54) DIR_ITEM(54)

Location 位置 

DIR_ITEM items are only located in file trees.
DIR_ITEM 项仅位于文件树中。

Usage 用法 

struct btrfs_key

objectid

objectid of owning inode

Description 描述 

DIR_ITEM items contain directory entries. Each name is hashed using the name hash and that value is used in the key for locating the entry quickly. Each DIR_ITEM item contains one or more directory entries with names represented by the same hash. All directory entries that share the same name hash must fit in a single leaf.
DIR_ITEM 项包含目录条目。每个名称都使用名称哈希进行哈希处理,该值用于快速定位条目的键。每个 DIR_ITEM 项包含一个或多个由相同哈希表示的名称的目录条目。所有共享相同名称哈希的目录条目必须适合单个叶子。

Item Contents 项目内容

DIR_ITEM items consist of a series of one or more directory entries with names that share a hash value. Each entry consists of a btrfs_dir_item structure immediately followed by the name. The length of each name is contained in btrfs_dir_item.name_len. The location of the item to which this entry refers is contained in btrfs_dir_item.location and must refer to a valid item in the same file tree. btrfs_dir_item.type contains a shorthand value referring to the type of item to which an entry refers. It will never be BTRFS_FT_XATTR when used in a standard directory. btrfs_dir_item.data_len is unused and must be 0.
DIR_ITEM 项由一系列共享哈希值的一个或多个目录条目组成。每个条目由紧随其后的名称组成的 btrfs_dir_item 结构组成。每个名称的长度包含在 btrfs_dir_item.name_len 中。此条目引用的项目的位置包含在 btrfs_dir_item.location 中,必须引用同一文件树中的有效项目。 btrfs_dir_item.type 包含一个简写值,指示条目引用的项目类型。在标准目录中使用时,它永远不会是 BTRFS_FT_XATTRbtrfs_dir_item.data_len 未使用,必须是 0

When there is more than one entry for a single hash value, the offset of each entry must be calculating using the lengths of the preceding entries including names.
当一个哈希值对应多个条目时,每个条目的偏移量必须使用前面条目的长度(包括名称)进行计算。

For more details, please see: struct btrfs_dir_item.
更多详情,请参阅: struct btrfs_dir_item

DIR_INDEX (60)

(parent objectid, 60, index in parent)
(父对象 id,60,父对象中的索引)

Allows looking up an item in a directory by index. Indices start at 2 (because of “.” and “..”); removed files can cause “holes” in the index space. DIR_INDEXen have the same contents as DIR_ITEM, but may contain only one entry.
允许通过索引在目录中查找项目。索引从 2 开始(因为“.”和“..”);已删除的文件可能会在索引空间中造成“空洞”。DIR_INDEXen 与 DIR_ITEM 具有相同的内容,但可能只包含一个条目。

EXTENT_DATA (6c) EXTENT_DATA(6c)

(inode id, 6c, offset in file) TODO
(inode id, 6c, 文件中的偏移) 待办事项

The contents of a file.
文件的内容

Off

Size

Type

Description

0

8

UINT

generation

8

8

UINT

(n) size of decoded extent
(n) 已解码范围的大小

10

1

UINT

compression (0=none, 1=zlib, 2=LZO)
压缩 (0=无, 1=zlib, 2=LZO)

11

1

UINT

encryption (0=none) 加密 (0=无)

12

2

UINT

other encoding (0=none) 其他编码 (0=无)

14

1

UINT

type (0=inline, 1=regular, 2=prealloc)
类型(0=内联,1=常规,2=预分配)

15

If the extent is inline, the remaining item bytes are the data bytes (n bytes in case no compression/encryption/other encoding is used).
如果范围是内联的,则剩余的项目字节是数据字节(如果没有使用压缩/加密/其他编码,则为 n 字节)。

Otherwise, the structure continues:
否则,结构继续:

Off

Size

Type

Description

15

8

UINT

(ea) logical address of extent. If this is zero, the extent is sparse and consists of all zeroes.
(ea) 扩展的逻辑地址。如果这个值为零,则该扩展是稀疏的,由全零组成。

1d

8

UINT

(es) size of extent (es) 扩展的大小

25

8

UINT

(o) offset within the extent
(o) 在扩展内的偏移量

2d

8

UINT

(s) logical number of bytes in file
(s) 文件中的逻辑字节数

35

ea and es must exactly match an EXTENT_ITEM. If the es bytes of data at logical address ea are decoded, n bytes will result. The file’s data contains the s bytes at offset o within the decoded bytes. In the simplest, uncompressed case, o=0 and n=es=s, so the file’s data simply contains the n bytes at logical address ea.
ea 和 es 必须与 EXTENT_ITEM 完全匹配。如果在逻辑地址 ea 处的 es 字节数据被解码,将得到 n 字节。文件的数据包含解码字节中偏移量为 o 处的 s 字节。在最简单的、未压缩的情况下,o=0 且 n=es=s,因此文件的数据简单地包含逻辑地址 ea 处的 n 字节。

EXTENT_CSUM (80)

(-a, logical address?) TODO
(-a,逻辑地址?) 待办事项

``   Contains one or more checksums of the type in the superblock for adjacent``
`` 包含相邻块从逻辑地址(块大小)开始的超级块中类型的一个或多个校验和``
``   blocks starting at logical address (blocksize).``
`` 块``

ROOT_ITEM (84) 根项目 (84) 

Location 位置 

ROOT_ITEM items are only located in the ROOT_TREE.
仅有 ROOT_ITEM 个项目位于根树中。

Usage 用法 

struct btrfs_key

objectid

objectid of root (TODO: document reserved objectids)

Description 描述 

A fundamental component of btrfs is the btree. ROOT_ITEM items define the location and parameters of the root of a btree.
btrfs 的一个基本组件是 btree。 ROOT_ITEM 项定义了 btree 根的位置和参数。

Item Contents 项目内容

ROOT_ITEM items contain a single btrfs_root_item structure.
ROOT_ITEM 项包含一个单一的 btrfs_root_item 结构。

ROOT_BACKREF (90)

(subtree id, 90, tree id) TODO
(子树 id, 90, 树 id) 待办事项

Same content as ROOT_REF.
与 ROOT_REF 相同的内容。

ROOT_REF (9c)

Location 位置 

ROOT_REF items are only located in the `ROOT_TREE <#ROOT_TREE>`__.
仅有 ROOT_REF 项位于 `ROOT_TREE <#ROOT_TREE>`__。

(tree id, subtree id) TODO
(树 ID,子树 ID) 待办事项

``    0  8 UINT   ID of directory in [tree id] that contains the subtree``
``    0  8 UINT   包含子树的[树 ID]中目录的 ID``
``    8  8 UINT   Sequence (index in tree) (even, starting at 2?)``
``    8  8 UINT   序列(树中的索引)(偶数,从 2 开始?)``
``   10  2 UINT   (n)`` ``   10  2 UINT   (n)``
``   12  n ASCII  name``
``   12  n ASCII  名称``

EXTENT_ITEM (a8)

Location 位置 

EXTENT_ITEM items are only located in the `EXTENT_TREE <#EXTENT_TREE>`__.
EXTENT_ITEM 项仅位于 `EXTENT_TREE <#EXTENT_TREE>`__。

Usage 用法 

struct btrfs_key

objectid

byte offset for start of extent

Description 描述 

EXTENT_ITEM items describe the space allocated for metadata tree nodes and leafs as well as data extents. The space is allocated from block groups that define the appropriate regions. In addition to functioning as basic allocation records, EXTENT_ITEM items also contain back references that can be used to repair the file system or resolve extent ownership back to a set of one or more file trees. Although EXTENT_ITEM items can be used to describe both DATA and TREE_BLOCK extents, newer file systems with the skinny metadata feature enabled at mkfs time use METADATA_ITEM items to represent metadata instead.
EXTENT_ITEM 项描述了为元数据树节点和叶子以及数据范围分配的空间。该空间是从定义适当区域的块组中分配的。除了作为基本分配记录外, EXTENT_ITEM 项还包含可以用于修复文件系统或将范围所有权解析回一组或多个文件树的反向引用。虽然 EXTENT_ITEM 项可用于描述 DATATREE_BLOCK 范围,但在 mkfs 时间启用了瘦元数据功能的新文件系统使用 METADATA_ITEM 项来代表元数据。

Item Contents 项目内容

EXTENT_ITEM items begin with the `btrfs_extent_item <Data_Structures#btrfs_extent_item>`__ structure and are followed by records that are defined by the flags field in that structure.
EXTENT_ITEM 项以 `btrfs_extent_item `__ 结构开头,后跟由该结构中的 flags 字段定义的记录。

METADATA_ITEM (a9) 元数据项 (a9) 

Location 位置 

METADATA_ITEM items are only located in the EXTENT_TREE.
仅有 METADATA_ITEM 项位于 EXTENT_TREE

Usage 用法 

struct btrfs_key

objectid

byte offset for start of extent

Description 描述 

METADATA_ITEM items describe the space allocated for metadata tree nodes and leafs. The space is allocated from block groups that define metadata regions. In addition to functioning as basic allocation records, METADATA_ITEM items also contain back references that can be used to repair the file system or resolve extent ownership back to a set of one or more file trees.
METADATA_ITEM 项描述了为元数据树节点和叶子节点分配的空间。该空间是从定义元数据区域的块组中分配的。除了作为基本分配记录的功能外, METADATA_ITEM 项还包含可以用于修复文件系统或将范围所有权解析回一组或多个文件树的反向引用。

Item Contents 项目内容

METADATA_ITEM items begin with the btrfs_extent_item structure and are followed by records that are defined by the flags field in that structure.
METADATA_ITEM 项以 btrfs_extent_item 结构开头,并由该结构中的 flags 字段定义的记录跟随。

TREE_BLOCK_REF (b0)

(logical address, b0, root object id) TODO
(逻辑地址,b0,根对象 ID) 待办事项

``    0   8 UINT   offset (the object ID of the tree)``
`` 0 8 UINT 偏移量(树的对象 ID)``

EXTENT_DATA_REF (b2) EXTENT_DATA_REF(b2)

(logical address, b2, hash of first three fields) TODO
(逻辑地址,b2,前三个字段的哈希) 待办事项

``    0   8 UINT   root objectid (id of tree contained in)``
`` 0 8 UINT 根对象 id(所包含树的 id)``
``    8   8 UINT   object id (owner)``
`` 8 8 UINT 对象 id(所有者)``
``   10   8 UINT   offset (in the file data)``
``   10   8 UINT   偏移量(在文件数据中)``
``   18   4 UINT   count (always 1?)``
``   18   4 UINT   计数(始终为 1?)``

EXTENT_REF_V0 (b4) EXTENT_REF_V0(b4)

TODO 待办事项

SHARED_BLOCK_REF (b6) 共享块引用(b6)

(logical address, b6, parent) TODO
(逻辑地址,b6,父级)待办事项

Off

Size

Type

Description

0

8

UINT

offset

8

SHARED_DATA_REF (b8) 共享数据引用 (b8) 

(logical address, b8, parent) TODO
(逻辑地址,b8,父级) 待办事项

Off

Size

Type

Description

0

8

UINT

offset

8

4

UINT

count (always 1?) 计数(始终为 1?)

c

BLOCK_GROUP_ITEM (c0) 区块组项目 (c0) 

Location 位置 

BLOCK_GROUP_ITEM items are only found in the EXTENT_TREE.
仅在 EXTENT_TREE 中找到 BLOCK_GROUP_ITEM 项。

Usage 用法 

struct btrfs_key

objectid

Starting offset in the space defined by the `EXTENT_TREE <#EXTENT_TREE>`__.
`EXTENT_TREE <#EXTENT_TREE>`__ 定义的空间中的起始偏移量。

Description 描述 

While the EXTENT_TREE defines the address space used for extent allocations for the entire file system, block groups allocate and define the parameters within that space. Every EXTENT_ITEM or METADATA_ITEM that describes an extent in use by the file system is apportioned from allocated block groups. Each block group can represent space used for SYSTEM objects (e.g. the CHUNK_TREE and primary super block), METADATA trees and items, or DATA extents. It is possible to combine METADATA and DATA allocations within a single block group, though it is not recommended. This mixed allocation policy is typically only seen on file systems smaller than approximately 10 GiB in size.
虽然 EXTENT_TREE 定义了整个文件系统用于范围分配的地址空间,但块组分配并定义了该空间内的参数。描述文件系统中正在使用的范围的每个 EXTENT_ITEMMETADATA_ITEM 都是从已分配的块组中分配的。每个块组可以表示用于 SYSTEM 对象(例如 CHUNK_TREE 和主超级块)、 METADATA 树和项目,或 DATA 范围的空间。可以在单个块组中组合 METADATADATA 分配,但不建议这样做。这种混合分配策略通常仅在大小小于约 10 GiB 的文件系统中看到。

Item Contents 项目内容

BTRFS_BLOCK_GROUP items contain a single struct btrfs_block_group_item.
BTRFS_BLOCK_GROUP 项目包含一个 struct btrfs_block_group_item

DEV_EXTENT (cc)

(device id, cc, physical address) TODO
(设备标识符, cc, 物理地址) 待办事项

Maps from physical address to logical.
从物理地址映射到逻辑地址。

Off

Size

Type

Description

0

8

UINT

chunk tree (always 3) 块树(始终为 3)

8

8

OBJID

chunk oid (always 256?) 块 OID(始终为 256?)

10

8

UINT

logical address 逻辑地址

18

8

UINT

size in bytes 字节大小

20

10

UUID

chunk tree UUID 块树 UUID

30

DEV_ITEM (d8)

(1, device id) TODO
(1, 设备标识) 待办事项

Contains information about one device.
包含有关一个设备的信息

Off

Size

Type

Description

0

8

UINT

device id 设备标识

8

8

UINT

number of bytes 字节数

10

8

UINT

number of bytes used 使用的字节数

18

4

UINT

optimal I/O align 最佳 I/O 对齐

1c

4

UINT

optimal I/O width 最佳 I/O 宽度

20

4

UINT

minimal I/O size (sector size)
最小 I/O 大小(扇区大小)

24

8

UINT

type

2c

8

UINT

generation

34

8

UINT

start offset 起始偏移量

3c

4

UINT

dev group 开发组

40

1

UINT

seek speed 寻找速度

41

1

UINT

bandwidth

42

10

UUID

device UUID 设备 UUID

52

10

UUID

FS UUID

62

CHUNK_ITEM (e4)

(100, logical address) TODO
(100, 逻辑地址) 待办事项

``   Maps logical address to physical.``
``将逻辑地址映射到物理地址。``
``    0  8 UINT   size of chunk (bytes)``
``0 8 UINT 每个块的大小(字节)``
``    8  8 OBJID  root referencing this chunk (2)``
``8 8 OBJID 引用此块的根(2)``
``   10  8 UINT   stripe length``
``   10  8 UINT   条带长度``
``   18  8 UINT   type (same as flags for block group?)``
``   18  8 UINT   类型(与块组标志相同?)``
``   20  4 UINT   optimal io alignment``
``   20  4 UINT   最佳 IO 对齐``
``   24  4 UINT   optimal io width``
``   24  4 UINT   最佳 io 宽度``
``   28  4 UINT   minimal io size (sector size)``
``   28  4 UINT   最小 io 大小(扇区大小)``
``   2c  2 UINT   number of stripes``
``   2c  2 UINT   条带数量``
``   2e  2 UINT   sub stripes``
``   2e  2 UINT   子条带``
``   30``
``   Stripes follow (for each number of stripes):``
``   对于每个条带,依次执行以下操作:``
``    0  8 OBJID  device id``
``    0  8 OBJID  设备标识符``
``    8  8 UINT   offset``
``    8  8 UINT   偏移量``
``   10 10 UUID   device UUID``
``   10 10 UUID   设备 UUID``
``   20``

STRING_ITEM (fd)

(anything, 0) (任何东西,0)

Contains a string; used for testing only.
仅用于测试的字符串。