1CPIO(5) BSD File Formats Manual CPIO(5)
2
4 cpio — format of cpio archive files
5
7 The cpio archive format collects any number of files, directories, and
8 other file system objects (symbolic links, device nodes, etc.) into a
9 single stream of bytes.
10 General Format
12 Each file system object in a cpio archive comprises a header record with
13 basic numeric metadata followed by the full pathname of the entry and the
14 file data. The header record stores a series of integer values that gen‐
15 erally follow the fields in struct stat. (See stat(2) for details.) The
16 variants differ primarily in how they store those integers (binary,
17 octal, or hexadecimal). The header is followed by the pathname of the
18 entry (the length of the pathname is stored in the header) and any file
19 data. The end of the archive is indicated by a special record with the
20 pathname “TRAILER!!!”.
21 PWB format
23 XXX Any documentation of the original PWB/UNIX 1.0 format? XXX
24 Old Binary Format
26 The old binary cpio format stores numbers as 2-byte and 4-byte binary
27 values. Each entry begins with a header in the following format:
28 struct header_old_cpio {
30 unsigned short c_magic;
31 unsigned short c_dev;
32 unsigned short c_ino;
33 unsigned short c_mode;
34 unsigned short c_uid;
35 unsigned short c_gid;
36 unsigned short c_nlink;
37 unsigned short c_rdev;
38 unsigned short c_mtime[2];
39 unsigned short c_namesize;
40 unsigned short c_filesize[2];
41 };
42 The unsigned short fields here are 16-bit integer values; the unsigned
44 int fields are 32-bit integer values. The fields are as follows
45 magic The integer value octal 070707. This value can be used to deter‐
47 mine whether this archive is written with little-endian or big-
48 endian integers.
49 dev, ino
51 The device and inode numbers from the disk. These are used by
52 programs that read cpio archives to determine when two entries
53 refer to the same file. Programs that synthesize cpio archives
54 should be careful to set these to distinct values for each entry.
55 mode The mode specifies both the regular permissions and the file
57 type. It consists of several bit fields as follows:
58 0170000 This masks the file type bits.
59 0140000 File type value for sockets.
60 0120000 File type value for symbolic links. For symbolic links,
61 the link body is stored as file data.
62 0100000 File type value for regular files.
63 0060000 File type value for block special devices.
64 0040000 File type value for directories.
65 0020000 File type value for character special devices.
66 0010000 File type value for named pipes or FIFOs.
67 0004000 SUID bit.
68 0002000 SGID bit.
69 0001000 Sticky bit. On some systems, this modifies the behavior
70 of executables and/or directories.
71 0000777 The lower 9 bits specify read/write/execute permissions
72 for world, group, and user following standard POSIX con‐
73 ventions.
74 uid, gid
76 The numeric user id and group id of the owner.
77 nlink The number of links to this file. Directories always have a
79 value of at least two here. Note that hardlinked files include
80 file data with every copy in the archive.
81 rdev For block special and character special entries, this field con‐
83 tains the associated device number. For all other entry types,
84 it should be set to zero by writers and ignored by readers.
85 mtime Modification time of the file, indicated as the number of seconds
87 since the start of the epoch, 00:00:00 UTC January 1, 1970. The
88 four-byte integer is stored with the most-significant 16 bits
89 first followed by the least-significant 16 bits. Each of the two
90 16 bit values are stored in machine-native byte order.
91 namesize
93 The number of bytes in the pathname that follows the header.
94 This count includes the trailing NUL byte.
95 filesize
97 The size of the file. Note that this archive format is limited
98 to four gigabyte file sizes. See mtime above for a description
99 of the storage of four-byte integers.
100 The pathname immediately follows the fixed header. If the namesize is
102 odd, an additional NUL byte is added after the pathname. The file data
103 is then appended, padded with NUL bytes to an even length.
104 Hardlinked files are not given special treatment; the full file contents
106 are included with each copy of the file.
107 Portable ASCII Format
109 Version 2 of the Single UNIX Specification (“SUSv2”) standardized an
110 ASCII variant that is portable across all platforms. It is commonly
111 known as the “old character” format or as the “odc” format. It stores
112 the same numeric fields as the old binary format, but represents them as
113 6-character or 11-character octal values.
114 struct cpio_odc_header {
116 char c_magic[6];
117 char c_dev[6];
118 char c_ino[6];
119 char c_mode[6];
120 char c_uid[6];
121 char c_gid[6];
122 char c_nlink[6];
123 char c_rdev[6];
124 char c_mtime[11];
125 char c_namesize[6];
126 char c_filesize[11];
127 };
128 The fields are identical to those in the old binary format. The name and
130 file body follow the fixed header. Unlike the old binary format, there
131 is no additional padding after the pathname or file contents. If the
132 files being archived are themselves entirely ASCII, then the resulting
133 archive will be entirely ASCII, except for the NUL byte that terminates
134 the name field.
135 New ASCII Format
137 The "new" ASCII format uses 8-byte hexadecimal fields for all numbers and
138 separates device numbers into separate fields for major and minor num‐
139 bers.
140 struct cpio_newc_header {
142 char c_magic[6];
143 char c_ino[8];
144 char c_mode[8];
145 char c_uid[8];
146 char c_gid[8];
147 char c_nlink[8];
148 char c_mtime[8];
149 char c_filesize[8];
150 char c_devmajor[8];
151 char c_devminor[8];
152 char c_rdevmajor[8];
153 char c_rdevminor[8];
154 char c_namesize[8];
155 char c_check[8];
156 };
157 Except as specified below, the fields here match those specified for the
159 old binary format above.
160 magic The string “070701”.
162 check This field is always set to zero by writers and ignored by read‐
164 ers. See the next section for more details.
165 The pathname is followed by NUL bytes so that the total size of the fixed
167 header plus pathname is a multiple of four. Likewise, the file data is
168 padded to a multiple of four bytes. Note that this format supports only
169 4 gigabyte files (unlike the older ASCII format, which supports 8 giga‐
170 byte files).
171 In this format, hardlinked files are handled by setting the filesize to
173 zero for each entry except the first one that appears in the archive.
174 New CRC Format
176 The CRC format is identical to the new ASCII format described in the pre‐
177 vious section except that the magic field is set to “070702” and the
178 check field is set to the sum of all bytes in the file data. This sum is
179 computed treating all bytes as unsigned values and using unsigned arith‐
180 metic. Only the least-significant 32 bits of the sum are stored.
181 HP variants
183 The cpio implementation distributed with HPUX used XXXX but stored device
184 numbers differently XXX.
185 Other Extensions and Variants
187 Sun Solaris uses additional file types to store extended file data,
188 including ACLs and extended attributes, as special entries in cpio ar‐
189 chives.
190 XXX Others? XXX
192
194 cpio(1), tar(5)
195
197 The cpio utility is no longer a part of POSIX or the Single Unix Stan‐
198 dard. It last appeared in Version 2 of the Single UNIX Specification
199 (“SUSv2”). It has been supplanted in subsequent standards by pax(1).
200 The portable ASCII format is currently part of the specification for the
201 pax(1) utility.
202
204 The original cpio utility was written by Dick Haight while working in
205 AT&T's Unix Support Group. It appeared in 1977 as part of PWB/UNIX 1.0,
206 the “Programmer's Work Bench” derived from Version 6 AT&T UNIX that was
207 used internally at AT&T. Both the old binary and old character formats
208 were in use by 1980, according to the System III source released by SCO
209 under their “Ancient Unix” license. The character format was adopted as
210 part of IEEE Std 1003.1-1988 (“POSIX.1”). XXX when did "newc" appear?
211 Who invented it? When did HP come out with their variant? When did Sun
212 introduce ACLs and extended attributes? XXX
213
215 The “CRC” format is mis-named, as it uses a simple checksum and not a
216 cyclic redundancy check.
217 The old binary format is limited to 16 bits for user id, group id,
219 device, and inode numbers. It is limited to 4 gigabyte file sizes.
220 The old ASCII format is limited to 18 bits for the user id, group id,
222 device, and inode numbers. It is limited to 8 gigabyte file sizes.
223 The new ASCII format is limited to 4 gigabyte file sizes.
225 None of the cpio formats store user or group names, which are essential
227 when moving files between systems with dissimilar user or group number‐
228 ing.
229 Especially when writing older cpio variants, it may be necessary to map
231 actual device/inode values to synthesized values that fit the available
232 fields. With very large filesystems, this may be necessary even for the
233 newer formats.
234BSD December 23, 2011 BSD