aboutsummaryrefslogtreecommitdiffstats
path: root/Documentation/filesystems
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2009-01-09 18:18:49 -0500
committerLinus Torvalds <torvalds@linux-foundation.org>2009-01-09 18:18:49 -0500
commit31aeb6c815549948571eec988ad9728c27d7a68d (patch)
treee155438be253924ebb1b792182e406947369b3eb /Documentation/filesystems
parentc40f6f8bbc4cbd2902671aacd587400ddca62627 (diff)
parentfc55584175589b70f4c30cb594f09f4bd6ad5d40 (diff)
Merge git://git.kernel.org/pub/scm/linux/kernel/git/pkl/squashfs-linus
* git://git.kernel.org/pub/scm/linux/kernel/git/pkl/squashfs-linus: MAINTAINERS: squashfs entry Squashfs: documentation Squashfs: initrd support Squashfs: Kconfig entry Squashfs: Makefiles Squashfs: header files Squashfs: block operations Squashfs: cache operations Squashfs: uid/gid lookup operations Squashfs: fragment block operations Squashfs: export operations Squashfs: super block operations Squashfs: symlink operations Squashfs: regular file operations Squashfs: directory readdir operations Squashfs: directory lookup operations Squashfs: inode operations
Diffstat (limited to 'Documentation/filesystems')
-rw-r--r--Documentation/filesystems/squashfs.txt225
1 files changed, 225 insertions, 0 deletions
diff --git a/Documentation/filesystems/squashfs.txt b/Documentation/filesystems/squashfs.txt
new file mode 100644
index 000000000000..3e79e4a7a392
--- /dev/null
+++ b/Documentation/filesystems/squashfs.txt
@@ -0,0 +1,225 @@
1SQUASHFS 4.0 FILESYSTEM
2=======================
3
4Squashfs is a compressed read-only filesystem for Linux.
5It uses zlib compression to compress files, inodes and directories.
6Inodes in the system are very small and all blocks are packed to minimise
7data overhead. Block sizes greater than 4K are supported up to a maximum
8of 1Mbytes (default block size 128K).
9
10Squashfs is intended for general read-only filesystem use, for archival
11use (i.e. in cases where a .tar.gz file may be used), and in constrained
12block device/memory systems (e.g. embedded systems) where low overhead is
13needed.
14
15Mailing list: squashfs-devel@lists.sourceforge.net
16Web site: www.squashfs.org
17
181. FILESYSTEM FEATURES
19----------------------
20
21Squashfs filesystem features versus Cramfs:
22
23 Squashfs Cramfs
24
25Max filesystem size: 2^64 16 MiB
26Max file size: ~ 2 TiB 16 MiB
27Max files: unlimited unlimited
28Max directories: unlimited unlimited
29Max entries per directory: unlimited unlimited
30Max block size: 1 MiB 4 KiB
31Metadata compression: yes no
32Directory indexes: yes no
33Sparse file support: yes no
34Tail-end packing (fragments): yes no
35Exportable (NFS etc.): yes no
36Hard link support: yes no
37"." and ".." in readdir: yes no
38Real inode numbers: yes no
3932-bit uids/gids: yes no
40File creation time: yes no
41Xattr and ACL support: no no
42
43Squashfs compresses data, inodes and directories. In addition, inode and
44directory data are highly compacted, and packed on byte boundaries. Each
45compressed inode is on average 8 bytes in length (the exact length varies on
46file type, i.e. regular file, directory, symbolic link, and block/char device
47inodes have different sizes).
48
492. USING SQUASHFS
50-----------------
51
52As squashfs is a read-only filesystem, the mksquashfs program must be used to
53create populated squashfs filesystems. This and other squashfs utilities
54can be obtained from http://www.squashfs.org. Usage instructions can be
55obtained from this site also.
56
57
583. SQUASHFS FILESYSTEM DESIGN
59-----------------------------
60
61A squashfs filesystem consists of seven parts, packed together on a byte
62alignment:
63
64 ---------------
65 | superblock |
66 |---------------|
67 | datablocks |
68 | & fragments |
69 |---------------|
70 | inode table |
71 |---------------|
72 | directory |
73 | table |
74 |---------------|
75 | fragment |
76 | table |
77 |---------------|
78 | export |
79 | table |
80 |---------------|
81 | uid/gid |
82 | lookup table |
83 ---------------
84
85Compressed data blocks are written to the filesystem as files are read from
86the source directory, and checked for duplicates. Once all file data has been
87written the completed inode, directory, fragment, export and uid/gid lookup
88tables are written.
89
903.1 Inodes
91----------
92
93Metadata (inodes and directories) are compressed in 8Kbyte blocks. Each
94compressed block is prefixed by a two byte length, the top bit is set if the
95block is uncompressed. A block will be uncompressed if the -noI option is set,
96or if the compressed block was larger than the uncompressed block.
97
98Inodes are packed into the metadata blocks, and are not aligned to block
99boundaries, therefore inodes overlap compressed blocks. Inodes are identified
100by a 48-bit number which encodes the location of the compressed metadata block
101containing the inode, and the byte offset into that block where the inode is
102placed (<block, offset>).
103
104To maximise compression there are different inodes for each file type
105(regular file, directory, device, etc.), the inode contents and length
106varying with the type.
107
108To further maximise compression, two types of regular file inode and
109directory inode are defined: inodes optimised for frequently occurring
110regular files and directories, and extended types where extra
111information has to be stored.
112
1133.2 Directories
114---------------
115
116Like inodes, directories are packed into compressed metadata blocks, stored
117in a directory table. Directories are accessed using the start address of
118the metablock containing the directory and the offset into the
119decompressed block (<block, offset>).
120
121Directories are organised in a slightly complex way, and are not simply
122a list of file names. The organisation takes advantage of the
123fact that (in most cases) the inodes of the files will be in the same
124compressed metadata block, and therefore, can share the start block.
125Directories are therefore organised in a two level list, a directory
126header containing the shared start block value, and a sequence of directory
127entries, each of which share the shared start block. A new directory header
128is written once/if the inode start block changes. The directory
129header/directory entry list is repeated as many times as necessary.
130
131Directories are sorted, and can contain a directory index to speed up
132file lookup. Directory indexes store one entry per metablock, each entry
133storing the index/filename mapping to the first directory header
134in each metadata block. Directories are sorted in alphabetical order,
135and at lookup the index is scanned linearly looking for the first filename
136alphabetically larger than the filename being looked up. At this point the
137location of the metadata block the filename is in has been found.
138The general idea of the index is ensure only one metadata block needs to be
139decompressed to do a lookup irrespective of the length of the directory.
140This scheme has the advantage that it doesn't require extra memory overhead
141and doesn't require much extra storage on disk.
142
1433.3 File data
144-------------
145
146Regular files consist of a sequence of contiguous compressed blocks, and/or a
147compressed fragment block (tail-end packed block). The compressed size
148of each datablock is stored in a block list contained within the
149file inode.
150
151To speed up access to datablocks when reading 'large' files (256 Mbytes or
152larger), the code implements an index cache that caches the mapping from
153block index to datablock location on disk.
154
155The index cache allows Squashfs to handle large files (up to 1.75 TiB) while
156retaining a simple and space-efficient block list on disk. The cache
157is split into slots, caching up to eight 224 GiB files (128 KiB blocks).
158Larger files use multiple slots, with 1.75 TiB files using all 8 slots.
159The index cache is designed to be memory efficient, and by default uses
16016 KiB.
161
1623.4 Fragment lookup table
163-------------------------
164
165Regular files can contain a fragment index which is mapped to a fragment
166location on disk and compressed size using a fragment lookup table. This
167fragment lookup table is itself stored compressed into metadata blocks.
168A second index table is used to locate these. This second index table for
169speed of access (and because it is small) is read at mount time and cached
170in memory.
171
1723.5 Uid/gid lookup table
173------------------------
174
175For space efficiency regular files store uid and gid indexes, which are
176converted to 32-bit uids/gids using an id look up table. This table is
177stored compressed into metadata blocks. A second index table is used to
178locate these. This second index table for speed of access (and because it
179is small) is read at mount time and cached in memory.
180
1813.6 Export table
182----------------
183
184To enable Squashfs filesystems to be exportable (via NFS etc.) filesystems
185can optionally (disabled with the -no-exports Mksquashfs option) contain
186an inode number to inode disk location lookup table. This is required to
187enable Squashfs to map inode numbers passed in filehandles to the inode
188location on disk, which is necessary when the export code reinstantiates
189expired/flushed inodes.
190
191This table is stored compressed into metadata blocks. A second index table is
192used to locate these. This second index table for speed of access (and because
193it is small) is read at mount time and cached in memory.
194
195
1964. TODOS AND OUTSTANDING ISSUES
197-------------------------------
198
1994.1 Todo list
200-------------
201
202Implement Xattr and ACL support. The Squashfs 4.0 filesystem layout has hooks
203for these but the code has not been written. Once the code has been written
204the existing layout should not require modification.
205
2064.2 Squashfs internal cache
207---------------------------
208
209Blocks in Squashfs are compressed. To avoid repeatedly decompressing
210recently accessed data Squashfs uses two small metadata and fragment caches.
211
212The cache is not used for file datablocks, these are decompressed and cached in
213the page-cache in the normal way. The cache is used to temporarily cache
214fragment and metadata blocks which have been read as a result of a metadata
215(i.e. inode or directory) or fragment access. Because metadata and fragments
216are packed together into blocks (to gain greater compression) the read of a
217particular piece of metadata or fragment will retrieve other metadata/fragments
218which have been packed with it, these because of locality-of-reference may be
219read in the near future. Temporarily caching them ensures they are available
220for near future access without requiring an additional read and decompress.
221
222In the future this internal cache may be replaced with an implementation which
223uses the kernel page cache. Because the page cache operates on page sized
224units this may introduce additional complexity in terms of locking and
225associated race conditions.