1STAT(2) Linux Programmer's Manual STAT(2)
2
6 stat, fstat, lstat - get file status
7
9 #include <sys/types.h>
10 #include <sys/stat.h>
11 #include <unistd.h>
12 int stat(const char *path, struct stat *buf);
14 int fstat(int filedes, struct stat *buf);
15 int lstat(const char *path, struct stat *buf);
16
18 These functions return information about a file. No permissions are
19 required on the file itself, but — in the case of stat() and lstat() —
20 execute (search) permission is required on all of the directories in
21 path that lead to the file.
22 stat() stats the file pointed to by path and fills in buf.
24 lstat() is identical to stat(), except that if path is a symbolic link,
26 then the link itself is stat-ed, not the file that it refers to.
27 fstat() is identical to stat(), except that the file to be stat-ed is
29 specified by the file descriptor filedes.
30 All of these system calls return a stat structure, which contains the
32 following fields:
33 struct stat {
35 dev_t st_dev; /* ID of device containing file */
36 ino_t st_ino; /* inode number */
37 mode_t st_mode; /* protection */
38 nlink_t st_nlink; /* number of hard links */
39 uid_t st_uid; /* user ID of owner */
40 gid_t st_gid; /* group ID of owner */
41 dev_t st_rdev; /* device ID (if special file) */
42 off_t st_size; /* total size, in bytes */
43 blksize_t st_blksize; /* blocksize for filesystem I/O */
44 blkcnt_t st_blocks; /* number of blocks allocated */
45 time_t st_atime; /* time of last access */
46 time_t st_mtime; /* time of last modification */
47 time_t st_ctime; /* time of last status change */
48 };
49 The st_dev field describes the device on which this file resides.
51 The st_rdev field describes the device that this file (inode) repre‐
53 sents.
54 The st_size field gives the size of the file (if it is a regular file
56 or a symbolic link) in bytes. The size of a symlink is the length of
57 the pathname it contains, without a trailing null byte.
58 The st_blocks field indicates the number of blocks allocated to the
60 file, 512-byte units. (This may be smaller than st_size/512, for exam‐
61 ple, when the file has holes.)
62 The st_blksize field gives the "preferred" blocksize for efficient file
64 system I/O. (Writing to a file in smaller chunks may cause an ineffi‐
65 cient read-modify-rewrite.)
66 Not all of the Linux filesystems implement all of the time fields.
68 Some file system types allow mounting in such a way that file accesses
69 do not cause an update of the st_atime field. (See `noatime' in
70 mount(8).)
71 The field st_atime is changed by file accesses, e.g. by execve(2),
73 mknod(2), pipe(2), utime(2) and read(2) (of more than zero bytes).
74 Other routines, like mmap(2), may or may not update st_atime.
75 The field st_mtime is changed by file modifications, e.g. by mknod(2),
77 truncate(2), utime(2) and write(2) (of more than zero bytes). More‐
78 over, st_mtime of a directory is changed by the creation or deletion of
79 files in that directory. The st_mtime field is not changed for changes
80 in owner, group, hard link count, or mode.
81 The field st_ctime is changed by writing or by setting inode informa‐
83 tion (i.e., owner, group, link count, mode, etc.).
84 The following POSIX macros are defined to check the file type using the
86 st_mode field:
87 S_ISREG(m) is it a regular file?
89 S_ISDIR(m) directory?
91 S_ISCHR(m) character device?
93 S_ISBLK(m) block device?
95 S_ISFIFO(m) FIFO (named pipe)?
97 S_ISLNK(m) symbolic link? (Not in POSIX.1-1996.)
99 S_ISSOCK(m) socket? (Not in POSIX.1-1996.)
101 The following flags are defined for the st_mode field:
103 S_IFMT 0170000 bitmask for the file type bitfields
105 S_IFSOCK 0140000 socket
106 S_IFLNK 0120000 symbolic link
107 S_IFREG 0100000 regular file
108 S_IFBLK 0060000 block device
109 S_IFDIR 0040000 directory
110 S_IFCHR 0020000 character device
111 S_IFIFO 0010000 FIFO
112 S_ISUID 0004000 set UID bit
113 S_ISGID 0002000 set-group-ID bit (see below)
114 S_ISVTX 0001000 sticky bit (see below)
115 S_IRWXU 00700 mask for file owner permissions
116 S_IRUSR 00400 owner has read permission
117 S_IWUSR 00200 owner has write permission
118 S_IXUSR 00100 owner has execute permission
119 S_IRWXG 00070 mask for group permissions
120 S_IRGRP 00040 group has read permission
121 S_IWGRP 00020 group has write permission
122 S_IXGRP 00010 group has execute permission
123 S_IRWXO 00007 mask for permissions for others (not in group)
124 S_IROTH 00004 others have read permission
125 S_IWOTH 00002 others have write permission
126 S_IXOTH 00001 others have execute permission
127 The set-group-ID bit (S_ISGID) has several special uses. For a direc‐
129 tory it indicates that BSD semantics is to be used for that directory:
130 files created there inherit their group ID from the directory, not from
131 the effective group ID of the creating process, and directories created
132 there will also get the S_ISGID bit set. For a file that does not have
133 the group execution bit (S_IXGRP) set, the set-group-ID bit indicates
134 mandatory file/record locking.
135 The `sticky' bit (S_ISVTX) on a directory means that a file in that
137 directory can be renamed or deleted only by the owner of the file, by
138 the owner of the directory, and by a privileged process.
139
141 Since kernel 2.5.48, the stat structure supports nanosecond resolution
142 for the three file timestamp fields. Glibc exposes the nanosecond com‐
143 ponent of each field using names either of the form st_atim.tv_nsec, if
144 the _BSD_SOURCE or _SVID_SOURCE feature test macro is defined, or of
145 the form st_atimensec, if neither of these macros is defined. On file
146 systems that do not support sub-second timestamps, these nanosecond
147 fields are returned with the value 0.
148 For most files under the /proc directory, stat() does not return the
150 file size in the st_size field; instead the field is returned with the
151 value 0.
152
154 On success, zero is returned. On error, -1 is returned, and errno is
155 set appropriately.
156
158 EACCES Search permission is denied for one of the directories in the
159 path prefix of path. (See also path_resolution(2).)
160 EBADF filedes is bad.
162 EFAULT Bad address.
164 ELOOP Too many symbolic links encountered while traversing the path.
166 ENAMETOOLONG
168 File name too long.
169 ENOENT A component of the path path does not exist, or the path is an
171 empty string.
172 ENOMEM Out of memory (i.e. kernel memory).
174 ENOTDIR
176 A component of the path is not a directory.
177
179 These system calls conform to SVr4, 4.3BSD, POSIX.1-2001.
180 Use of the st_blocks and st_blksize fields may be less portable. (They
182 were introduced in BSD. The interpretation differs between systems,
183 and possibly on a single system when NFS mounts are involved.)
184 POSIX does not describe the S_IFMT, S_IFSOCK, S_IFLNK, S_IFREG,
186 S_IFBLK, S_IFDIR, S_IFCHR, S_IFIFO, S_ISVTX bits, but instead demands
187 the use of the macros S_ISDIR(), etc. The S_ISLNK and S_ISSOCK macros
188 are not in POSIX.1-1996, but both are present in POSIX.1-2001; the for‐
189 mer is from SVID 4, the latter from SUSv2.
190 Unix V7 (and later systems) had S_IREAD, S_IWRITE, S_IEXEC, where POSIX
192 prescribes the synonyms S_IRUSR, S_IWUSR, S_IXUSR.
193
195 Values that have been (or are) in use on various systems:
196 hex name ls octal description
198 f000 S_IFMT 170000 mask for file type
200 0000 000000 SCO out-of-service inode, BSD unknown type
201 SVID-v2 and XPG2 have both 0 and 0100000 for ordinary file
202 1000 S_IFIFO p| 010000 FIFO (named pipe)
203 2000 S_IFCHR c 020000 character special (V7)
204 3000 S_IFMPC 030000 multiplexed character special (V7)
205 4000 S_IFDIR d/ 040000 directory (V7)
206 5000 S_IFNAM 050000 XENIX named special file
207 with two subtypes, distinguished by st_rdev values 1, 2:
208 0001 S_INSEM s 000001 XENIX semaphore subtype of IFNAM
209 0002 S_INSHD m 000002 XENIX shared data subtype of IFNAM
210 6000 S_IFBLK b 060000 block special (V7)
211 7000 S_IFMPB 070000 multiplexed block special (V7)
212 8000 S_IFREG - 100000 regular (V7)
213 9000 S_IFCMP 110000 VxFS compressed
214 9000 S_IFNWK n 110000 network special (HP-UX)
215 a000 S_IFLNK l@ 120000 symbolic link (BSD)
216 b000 S_IFSHAD 130000 Solaris shadow inode for ACL (not seen by userspace)
217 c000 S_IFSOCK s= 140000 socket (BSD; also "S_IFSOC" on VxFS)
218 d000 S_IFDOOR D> 150000 Solaris door
219 e000 S_IFWHT w% 160000 BSD whiteout (not used for inode)
220 0200 S_ISVTX 001000 `sticky bit': save swapped text even after use (V7)
222 reserved (SVID-v2)
223 On non-directories: don't cache this file (SunOS)
224 On directories: restricted deletion flag (SVID-v4.2)
225 0400 S_ISGID 002000 set-group-ID on execution (V7)
226 for directories: use BSD semantics for propagation of GID
227 0400 S_ENFMT 002000 SysV file locking enforcement (shared with S_ISGID)
228 0800 S_ISUID 004000 set-user-ID on execution (V7)
229 0800 S_CDF 004000 directory is a context dependent file (HP-UX)
230 A sticky command appeared in Version 32V AT&T UNIX.
232
234 access(2), chmod(2), chown(2), fstatat(2), readlink(2), utime(2), capa‐
235 bilities(7)
236Linux 2.6.7 2004-06-23 STAT(2)