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 fd, struct stat *buf);
15 int lstat(const char *path, struct stat *buf);
16 Feature Test Macro Requirements for glibc (see feature_test_macros(7)):
18 lstat(): _BSD_SOURCE || _XOPEN_SOURCE >= 500
20
22 These functions return information about a file. No permissions are
23 required on the file itself, but — in the case of stat() and lstat() —
24 execute (search) permission is required on all of the directories in
25 path that lead to the file.
26 stat() stats the file pointed to by path and fills in buf.
28 lstat() is identical to stat(), except that if path is a symbolic link,
30 then the link itself is stat-ed, not the file that it refers to.
31 fstat() is identical to stat(), except that the file to be stat-ed is
33 specified by the file descriptor fd.
34 All of these system calls return a stat structure, which contains the
36 following fields:
37 struct stat {
39 dev_t st_dev; /* ID of device containing file */
40 ino_t st_ino; /* inode number */
41 mode_t st_mode; /* protection */
42 nlink_t st_nlink; /* number of hard links */
43 uid_t st_uid; /* user ID of owner */
44 gid_t st_gid; /* group ID of owner */
45 dev_t st_rdev; /* device ID (if special file) */
46 off_t st_size; /* total size, in bytes */
47 blksize_t st_blksize; /* blocksize for file system I/O */
48 blkcnt_t st_blocks; /* number of 512B blocks allocated */
49 time_t st_atime; /* time of last access */
50 time_t st_mtime; /* time of last modification */
51 time_t st_ctime; /* time of last status change */
52 };
53 The st_dev field describes the device on which this file resides. (The
55 major(3) and minor(3) macros may be useful to decompose the device ID
56 in this field.)
57 The st_rdev field describes the device that this file (inode) repre‐
59 sents.
60 The st_size field gives the size of the file (if it is a regular file
62 or a symbolic link) in bytes. The size of a symlink is the length of
63 the pathname it contains, without a trailing null byte.
64 The st_blocks field indicates the number of blocks allocated to the
66 file, 512-byte units. (This may be smaller than st_size/512 when the
67 file has holes.)
68 The st_blksize field gives the "preferred" blocksize for efficient file
70 system I/O. (Writing to a file in smaller chunks may cause an ineffi‐
71 cient read-modify-rewrite.)
72 Not all of the Linux file systems implement all of the time fields.
74 Some file system types allow mounting in such a way that file and/or
75 directory accesses do not cause an update of the st_atime field. (See
76 noatime, nodiratime, and relatime in mount(8), and related information
77 in mount(2).) In addition, st_atime is not updated if a file is opened
78 with the O_NOATIME; see open(2).
79 The field st_atime is changed by file accesses, for example, by
81 execve(2), mknod(2), pipe(2), utime(2) and read(2) (of more than zero
82 bytes). Other routines, like mmap(2), may or may not update st_atime.
83 The field st_mtime is changed by file modifications, for example, by
85 mknod(2), truncate(2), utime(2) and write(2) (of more than zero bytes).
86 Moreover, st_mtime of a directory is changed by the creation or dele‐
87 tion of files in that directory. The st_mtime field is not changed for
88 changes in owner, group, hard link count, or mode.
89 The field st_ctime is changed by writing or by setting inode informa‐
91 tion (i.e., owner, group, link count, mode, etc.).
92 The following POSIX macros are defined to check the file type using the
94 st_mode field:
95 S_ISREG(m) is it a regular file?
97 S_ISDIR(m) directory?
99 S_ISCHR(m) character device?
101 S_ISBLK(m) block device?
103 S_ISFIFO(m) FIFO (named pipe)?
105 S_ISLNK(m) symbolic link? (Not in POSIX.1-1996.)
107 S_ISSOCK(m) socket? (Not in POSIX.1-1996.)
109 The following flags are defined for the st_mode field:
111 S_IFMT 0170000 bit mask for the file type bit fields
113 S_IFSOCK 0140000 socket
114 S_IFLNK 0120000 symbolic link
115 S_IFREG 0100000 regular file
116 S_IFBLK 0060000 block device
117 S_IFDIR 0040000 directory
118 S_IFCHR 0020000 character device
119 S_IFIFO 0010000 FIFO
120 S_ISUID 0004000 set UID bit
121 S_ISGID 0002000 set-group-ID bit (see below)
122 S_ISVTX 0001000 sticky bit (see below)
123 S_IRWXU 00700 mask for file owner permissions
124 S_IRUSR 00400 owner has read permission
125 S_IWUSR 00200 owner has write permission
126 S_IXUSR 00100 owner has execute permission
127 S_IRWXG 00070 mask for group permissions
128 S_IRGRP 00040 group has read permission
129 S_IWGRP 00020 group has write permission
130 S_IXGRP 00010 group has execute permission
131 S_IRWXO 00007 mask for permissions for others (not in group)
132 S_IROTH 00004 others have read permission
134 S_IWOTH 00002 others have write permission
135 S_IXOTH 00001 others have execute permission
136 The set-group-ID bit (S_ISGID) has several special uses. For a direc‐
138 tory it indicates that BSD semantics is to be used for that directory:
139 files created there inherit their group ID from the directory, not from
140 the effective group ID of the creating process, and directories created
141 there will also get the S_ISGID bit set. For a file that does not have
142 the group execution bit (S_IXGRP) set, the set-group-ID bit indicates
143 mandatory file/record locking.
144 The sticky bit (S_ISVTX) on a directory means that a file in that
146 directory can be renamed or deleted only by the owner of the file, by
147 the owner of the directory, and by a privileged process.
148
150 On success, zero is returned. On error, -1 is returned, and errno is
151 set appropriately.
152
154 EACCES Search permission is denied for one of the directories in the
155 path prefix of path. (See also path_resolution(7).)
156 EBADF fd is bad.
158 EFAULT Bad address.
160 ELOOP Too many symbolic links encountered while traversing the path.
162 ENAMETOOLONG
164 File name too long.
165 ENOENT A component of path does not exist, or path is an empty string.
167 ENOMEM Out of memory (i.e., kernel memory).
169 ENOTDIR
171 A component of the path prefix of path is not a directory.
172 EOVERFLOW
174 (stat()) path refers to a file whose size cannot be represented
175 in the type off_t. This can occur when an application compiled
176 on a 32-bit platform without -D_FILE_OFFSET_BITS=64 calls stat()
177 on a file whose size exceeds (2<<31)-1 bits.
178
180 These system calls conform to SVr4, 4.3BSD, POSIX.1-2001.
181 Use of the st_blocks and st_blksize fields may be less portable. (They
183 were introduced in BSD. The interpretation differs between systems,
184 and possibly on a single system when NFS mounts are involved.)
185 POSIX does not describe the S_IFMT, S_IFSOCK, S_IFLNK, S_IFREG,
187 S_IFBLK, S_IFDIR, S_IFCHR, S_IFIFO, S_ISVTX bits, but instead demands
188 the use of the macros S_ISDIR(), etc. The S_ISLNK() and S_ISSOCK()
189 macros are not in POSIX.1-1996, but both are present in POSIX.1-2001;
190 the former is from SVID 4, the latter from SUSv2.
191 Unix V7 (and later systems) had S_IREAD, S_IWRITE, S_IEXEC, where POSIX
193 prescribes the synonyms S_IRUSR, S_IWUSR, S_IXUSR.
194 Other Systems
196 Values that have been (or are) in use on various systems:
197 hex name ls octal description
200 f000 S_IFMT 170000 mask for file type
201 0000 000000 SCO out-of-service inode; BSD unknown
202 type; SVID-v2 and XPG2 have both
203 0 and 0100000 for ordinary file
204 1000 S_IFIFO p| 010000 FIFO (named pipe)
205 2000 S_IFCHR c 020000 character special (V7)
206 3000 S_IFMPC 030000 multiplexed character special (V7)
207 4000 S_IFDIR d/ 040000 directory (V7)
208 5000 S_IFNAM 050000 XENIX named special file
209 with two subtypes, distinguished by
210 st_rdev values 1, 2
211 0001 S_INSEM s 000001 XENIX semaphore subtype of IFNAM
212 0002 S_INSHD m 000002 XENIX shared data subtype of IFNAM
213 6000 S_IFBLK b 060000 block special (V7)
214 7000 S_IFMPB 070000 multiplexed block special (V7)
215 8000 S_IFREG - 100000 regular (V7)
216 9000 S_IFCMP 110000 VxFS compressed
217 9000 S_IFNWK n 110000 network special (HP-UX)
218 a000 S_IFLNK l@ 120000 symbolic link (BSD)
219 b000 S_IFSHAD 130000 Solaris shadow inode for ACL
220 (not seen by userspace)
221 c000 S_IFSOCK s= 140000 socket (BSD; also "S_IFSOC" on VxFS)
222 d000 S_IFDOOR D> 150000 Solaris door
223 e000 S_IFWHT w% 160000 BSD whiteout (not used for inode)
224 0200 S_ISVTX 001000 sticky bit: save swapped text even
225 after use (V7)
226 reserved (SVID-v2)
227 On nondirectories: don't cache this
228 file (SunOS)
229 On directories: restricted deletion
230 flag (SVID-v4.2)
231 0400 S_ISGID 002000 set-group-ID on execution (V7)
232 for directories: use BSD semantics for
233 propagation of GID
234 0400 S_ENFMT 002000 System V file locking enforcement (shared
235 with S_ISGID)
236 0800 S_ISUID 004000 set-user-ID on execution (V7)
237 0800 S_CDF 004000 directory is a context dependent
238 file (HP-UX)
239 A sticky command appeared in Version 32V AT&T UNIX.
241
243 Since kernel 2.5.48, the stat structure supports nanosecond resolution
244 for the three file timestamp fields. Glibc exposes the nanosecond com‐
245 ponent of each field using names either of the form st_atim.tv_nsec, if
246 the _BSD_SOURCE or _SVID_SOURCE feature test macro is defined, or of
247 the form st_atimensec, if neither of these macros is defined. On file
248 systems that do not support subsecond timestamps, these nanosecond
249 fields are returned with the value 0.
250 On Linux, lstat() will generally not trigger automounter action,
252 whereas stat() will.
253 For most files under the /proc directory, stat() does not return the
255 file size in the st_size field; instead the field is returned with the
256 value 0.
257 Underlying kernel interface
259 Over time, increases in the size of the stat structure have led to
260 three successive versions of stat(): sys_stat() (slot __NR_oldstat),
261 sys_newstat() (slot __NR_stat), and sys_stat64() (new in kernel 2.4;
262 slot __NR_stat64). The glibc stat() wrapper function hides these
263 details from applications, invoking the most recent version of the sys‐
264 tem call provided by the kernel, and repacking the returned information
265 if required for old binaries. Similar remarks apply for fstat() and
266 lstat().
267
269 The following program calls stat() and displays selected fields in the
270 returned stat structure.
271 #include <sys/types.h>
273 #include <sys/stat.h>
274 #include <time.h>
275 #include <stdio.h>
276 #include <stdlib.h>
277 int
279 main(int argc, char *argv[])
280 {
281 struct stat sb;
282 if (argc != 2) {
284 fprintf(stderr, "Usage: %s <pathname>\n", argv[0]);
285 exit(EXIT_FAILURE);
286 }
287 if (stat(argv[1], &sb) == -1) {
289 perror("stat");
290 exit(EXIT_FAILURE);
291 }
292 printf("File type: ");
294 switch (sb.st_mode & S_IFMT) {
296 case S_IFBLK: printf("block device\n"); break;
297 case S_IFCHR: printf("character device\n"); break;
298 case S_IFDIR: printf("directory\n"); break;
299 case S_IFIFO: printf("FIFO/pipe\n"); break;
300 case S_IFLNK: printf("symlink\n"); break;
301 case S_IFREG: printf("regular file\n"); break;
302 case S_IFSOCK: printf("socket\n"); break;
303 default: printf("unknown?\n"); break;
304 }
305 printf("I-node number: %ld\n", (long) sb.st_ino);
307 printf("Mode: %lo (octal)\n",
309 (unsigned long) sb.st_mode);
310 printf("Link count: %ld\n", (long) sb.st_nlink);
312 printf("Ownership: UID=%ld GID=%ld\n",
313 (long) sb.st_uid, (long) sb.st_gid);
314 printf("Preferred I/O block size: %ld bytes\n",
316 (long) sb.st_blksize);
317 printf("File size: %lld bytes\n",
318 (long long) sb.st_size);
319 printf("Blocks allocated: %lld\n",
320 (long long) sb.st_blocks);
321 printf("Last status change: %s", ctime(&sb.st_ctime));
323 printf("Last file access: %s", ctime(&sb.st_atime));
324 printf("Last file modification: %s", ctime(&sb.st_mtime));
325 exit(EXIT_SUCCESS);
327 }
328
330 access(2), chmod(2), chown(2), fstatat(2), readlink(2), utime(2), capa‐
331 bilities(7), symlink(7)
332
334 This page is part of release 3.25 of the Linux man-pages project. A
335 description of the project, and information about reporting bugs, can
336 be found at http://www.kernel.org/doc/man-pages/.
337Linux 2009-09-30 STAT(2)