1chmod(2) System Calls Manual chmod(2)
2
6 chmod, fchmod, fchmodat - change permissions of a file
7
9 Standard C library (libc, -lc)
10
12 #include <sys/stat.h>
13 int chmod(const char *pathname, mode_t mode);
15 int fchmod(int fd, mode_t mode);
16 #include <fcntl.h> /* Definition of AT_* constants */
18 #include <sys/stat.h>
19 int fchmodat(int dirfd, const char *pathname, mode_t mode, int flags);
21 Feature Test Macro Requirements for glibc (see feature_test_macros(7)):
23 fchmod():
25 Since glibc 2.24:
26 _POSIX_C_SOURCE >= 199309L
27 glibc 2.19 to glibc 2.23
28 _POSIX_C_SOURCE
29 glibc 2.16 to glibc 2.19:
30 _BSD_SOURCE || _POSIX_C_SOURCE
31 glibc 2.12 to glibc 2.16:
32 _BSD_SOURCE || _XOPEN_SOURCE >= 500
33 || _POSIX_C_SOURCE >= 200809L
34 glibc 2.11 and earlier:
35 _BSD_SOURCE || _XOPEN_SOURCE >= 500
36 fchmodat():
38 Since glibc 2.10:
39 _POSIX_C_SOURCE >= 200809L
40 Before glibc 2.10:
41 _ATFILE_SOURCE
42
44 The chmod() and fchmod() system calls change a file's mode bits. (The
45 file mode consists of the file permission bits plus the set-user-ID,
46 set-group-ID, and sticky bits.) These system calls differ only in how
47 the file is specified:
48 • chmod() changes the mode of the file specified whose pathname is
50 given in pathname, which is dereferenced if it is a symbolic link.
51 • fchmod() changes the mode of the file referred to by the open file
53 descriptor fd.
54 The new file mode is specified in mode, which is a bit mask created by
56 ORing together zero or more of the following:
57 S_ISUID [22m(04000) set-user-ID (set process effective user ID on ex‐
59 ecve(2))
60 S_ISGID [22m(02000) set-group-ID (set process effective group ID on ex‐
62 ecve(2); mandatory locking, as described in fcntl(2);
63 take a new file's group from parent directory, as de‐
64 scribed in chown(2) and mkdir(2))
65 S_ISVTX [22m(01000) sticky bit (restricted deletion flag, as described in
67 unlink(2))
68 S_IRUSR [22m(00400) read by owner
70 S_IWUSR [22m(00200) write by owner
72 S_IXUSR [22m(00100) execute/search by owner ("search" applies for direc‐
74 tories, and means that entries within the directory
75 can be accessed)
76 S_IRGRP [22m(00040) read by group
78 S_IWGRP [22m(00020) write by group
80 S_IXGRP [22m(00010) execute/search by group
82 S_IROTH [22m(00004) read by others
84 S_IWOTH [22m(00002) write by others
86 S_IXOTH [22m(00001) execute/search by others
88 The effective UID of the calling process must match the owner of the
90 file, or the process must be privileged (Linux: it must have the
91 CAP_FOWNER capability).
92 If the calling process is not privileged (Linux: does not have the
94 CAP_FSETID capability), and the group of the file does not match the
95 effective group ID of the process or one of its supplementary group
96 IDs, the S_ISGID bit will be turned off, but this will not cause an er‐
97 ror to be returned.
98 As a security measure, depending on the filesystem, the set-user-ID and
100 set-group-ID execution bits may be turned off if a file is written.
101 (On Linux, this occurs if the writing process does not have the
102 CAP_FSETID capability.) On some filesystems, only the superuser can
103 set the sticky bit, which may have a special meaning. For the sticky
104 bit, and for set-user-ID and set-group-ID bits on directories, see in‐
105 ode(7).
106 On NFS filesystems, restricting the permissions will immediately influ‐
108 ence already open files, because the access control is done on the
109 server, but open files are maintained by the client. Widening the per‐
110 missions may be delayed for other clients if attribute caching is en‐
111 abled on them.
112 fchmodat()
114 The fchmodat() system call operates in exactly the same way as chmod(),
115 except for the differences described here.
116 If the pathname given in pathname is relative, then it is interpreted
118 relative to the directory referred to by the file descriptor dirfd
119 (rather than relative to the current working directory of the calling
120 process, as is done by chmod() for a relative pathname).
121 If pathname is relative and dirfd is the special value AT_FDCWD, then
123 pathname is interpreted relative to the current working directory of
124 the calling process (like chmod()).
125 If pathname is absolute, then dirfd is ignored.
127 flags can either be 0, or include the following flag:
129 AT_SYMLINK_NOFOLLOW
131 If pathname is a symbolic link, do not dereference it: instead
132 operate on the link itself. This flag is not currently imple‐
133 mented.
134 See openat(2) for an explanation of the need for fchmodat().
136
138 On success, zero is returned. On error, -1 is returned, and errno is
139 set to indicate the error.
140
142 Depending on the filesystem, errors other than those listed below can
143 be returned.
144 The more general errors for chmod() are listed below:
146 EACCES Search permission is denied on a component of the path prefix.
148 (See also path_resolution(7).)
149 EBADF (fchmod()) The file descriptor fd is not valid.
151 EBADF (fchmodat()) pathname is relative but dirfd is neither AT_FDCWD
153 nor a valid file descriptor.
154 EFAULT pathname points outside your accessible address space.
156 EINVAL (fchmodat()) Invalid flag specified in flags.
158 EIO An I/O error occurred.
160 ELOOP Too many symbolic links were encountered in resolving pathname.
162 ENAMETOOLONG
164 pathname is too long.
165 ENOENT The file does not exist.
167 ENOMEM Insufficient kernel memory was available.
169 ENOTDIR
171 A component of the path prefix is not a directory.
172 ENOTDIR
174 (fchmodat()) pathname is relative and dirfd is a file descriptor
175 referring to a file other than a directory.
176 ENOTSUP
178 (fchmodat()) flags specified AT_SYMLINK_NOFOLLOW, which is not
179 supported.
180 EPERM The effective UID does not match the owner of the file, and the
182 process is not privileged (Linux: it does not have the
183 CAP_FOWNER capability).
184 EPERM The file is marked immutable or append-only. (See
186 ioctl_iflags(2).)
187 EROFS The named file resides on a read-only filesystem.
189
191 C library/kernel differences
192 The GNU C library fchmodat() wrapper function implements the POSIX-
193 specified interface described in this page. This interface differs
194 from the underlying Linux system call, which does not have a flags ar‐
195 gument.
196 glibc notes
198 On older kernels where fchmodat() is unavailable, the glibc wrapper
199 function falls back to the use of chmod(). When pathname is a relative
200 pathname, glibc constructs a pathname based on the symbolic link in
201 /proc/self/fd that corresponds to the dirfd argument.
202
204 POSIX.1-2008.
205
207 chmod()
208 fchmod()
209 4.4BSD, SVr4, POSIX.1-2001.
210 fchmodat()
212 POSIX.1-2008. Linux 2.6.16, glibc 2.4.
213
215 chmod(1), chown(2), execve(2), open(2), stat(2), inode(7), path_resolu‐
216 tion(7), symlink(7)
217Linux man-pages 6.05 2023-03-30 chmod(2)