1ACCESS(2) Linux Programmer's Manual ACCESS(2)
2
6 access, faccessat - check user's permissions for a file
7
9 #include <unistd.h>
10 int access(const char *pathname, int mode);
12 #include <fcntl.h> /* Definition of AT_* constants */
14 #include <unistd.h>
15 int faccessat(int dirfd, const char *pathname, int mode, int flags);
17 Feature Test Macro Requirements for glibc (see feature_test_macros(7)):
19 faccessat():
21 Since glibc 2.10:
22 _POSIX_C_SOURCE >= 200809L
23 Before glibc 2.10:
24 _ATFILE_SOURCE
25
27 access() checks whether the calling process can access the file path‐
28 name. If pathname is a symbolic link, it is dereferenced.
29 The mode specifies the accessibility check(s) to be performed, and is
31 either the value F_OK, or a mask consisting of the bitwise OR of one or
32 more of R_OK, W_OK, and X_OK. F_OK tests for the existence of the
33 file. R_OK, W_OK, and X_OK test whether the file exists and grants
34 read, write, and execute permissions, respectively.
35 The check is done using the calling process's real UID and GID, rather
37 than the effective IDs as is done when actually attempting an operation
38 (e.g., open(2)) on the file. Similarly, for the root user, the check
39 uses the set of permitted capabilities rather than the set of effective
40 capabilities; and for non-root users, the check uses an empty set of
41 capabilities.
42 This allows set-user-ID programs and capability-endowed programs to
44 easily determine the invoking user's authority. In other words,
45 access() does not answer the "can I read/write/execute this file?"
46 question. It answers a slightly different question: "(assuming I'm a
47 setuid binary) can the user who invoked me read/write/execute this
48 file?", which gives set-user-ID programs the possibility to prevent
49 malicious users from causing them to read files which users shouldn't
50 be able to read.
51 If the calling process is privileged (i.e., its real UID is zero), then
53 an X_OK check is successful for a regular file if execute permission is
54 enabled for any of the file owner, group, or other.
55 faccessat()
57 The faccessat() system call operates in exactly the same way as
58 access(), except for the differences described here.
59 If the pathname given in pathname is relative, then it is interpreted
61 relative to the directory referred to by the file descriptor dirfd
62 (rather than relative to the current working directory of the calling
63 process, as is done by access() for a relative pathname).
64 If pathname is relative and dirfd is the special value AT_FDCWD, then
66 pathname is interpreted relative to the current working directory of
67 the calling process (like access()).
68 If pathname is absolute, then dirfd is ignored.
70 flags is constructed by ORing together zero or more of the following
72 values:
73 AT_EACCESS
75 Perform access checks using the effective user and group IDs.
76 By default, faccessat() uses the real IDs (like access()).
77 AT_SYMLINK_NOFOLLOW
79 If pathname is a symbolic link, do not dereference it: instead
80 return information about the link itself.
81 See openat(2) for an explanation of the need for faccessat().
83
85 On success (all requested permissions granted, or mode is F_OK and the
86 file exists), zero is returned. On error (at least one bit in mode
87 asked for a permission that is denied, or mode is F_OK and the file
88 does not exist, or some other error occurred), -1 is returned, and
89 errno is set appropriately.
90
92 access() and faccessat() shall fail if:
93 EACCES The requested access would be denied to the file, or search per‐
95 mission is denied for one of the directories in the path prefix
96 of pathname. (See also path_resolution(7).)
97 ELOOP Too many symbolic links were encountered in resolving pathname.
99 ENAMETOOLONG
101 pathname is too long.
102 ENOENT A component of pathname does not exist or is a dangling symbolic
104 link.
105 ENOTDIR
107 A component used as a directory in pathname is not, in fact, a
108 directory.
109 EROFS Write permission was requested for a file on a read-only
111 filesystem.
112 access() and faccessat() may fail if:
114 EFAULT pathname points outside your accessible address space.
116 EINVAL mode was incorrectly specified.
118 EIO An I/O error occurred.
120 ENOMEM Insufficient kernel memory was available.
122 ETXTBSY
124 Write access was requested to an executable which is being exe‐
125 cuted.
126 The following additional errors can occur for faccessat():
128 EBADF dirfd is not a valid file descriptor.
130 EINVAL Invalid flag specified in flags.
132 ENOTDIR
134 pathname is relative and dirfd is a file descriptor referring to
135 a file other than a directory.
136
138 faccessat() was added to Linux in kernel 2.6.16; library support was
139 added to glibc in version 2.4.
140
142 access(): SVr4, 4.3BSD, POSIX.1-2001, POSIX.1-2008.
143 faccessat(): POSIX.1-2008.
145
147 Warning: Using these calls to check if a user is authorized to, for
148 example, open a file before actually doing so using open(2) creates a
149 security hole, because the user might exploit the short time interval
150 between checking and opening the file to manipulate it. For this rea‐
151 son, the use of this system call should be avoided. (In the example
152 just described, a safer alternative would be to temporarily switch the
153 process's effective user ID to the real ID and then call open(2).)
154 access() always dereferences symbolic links. If you need to check the
156 permissions on a symbolic link, use faccessat() with the flag AT_SYM‐
157 LINK_NOFOLLOW.
158 These calls return an error if any of the access types in mode is
160 denied, even if some of the other access types in mode are permitted.
161 If the calling process has appropriate privileges (i.e., is superuser),
163 POSIX.1-2001 permits an implementation to indicate success for an X_OK
164 check even if none of the execute file permission bits are set. Linux
165 does not do this.
166 A file is accessible only if the permissions on each of the directories
168 in the path prefix of pathname grant search (i.e., execute) access. If
169 any directory is inaccessible, then the access() call fails, regardless
170 of the permissions on the file itself.
171 Only access bits are checked, not the file type or contents. There‐
173 fore, if a directory is found to be writable, it probably means that
174 files can be created in the directory, and not that the directory can
175 be written as a file. Similarly, a DOS file may be found to be "exe‐
176 cutable," but the execve(2) call will still fail.
177 These calls may not work correctly on NFSv2 filesystems with UID map‐
179 ping enabled, because UID mapping is done on the server and hidden from
180 the client, which checks permissions. (NFS versions 3 and higher per‐
181 form the check on the server.) Similar problems can occur to FUSE
182 mounts.
183 C library/kernel differences
185 The raw faccessat() system call takes only the first three arguments.
186 The AT_EACCESS and AT_SYMLINK_NOFOLLOW flags are actually implemented
187 within the glibc wrapper function for faccessat(). If either of these
188 flags is specified, then the wrapper function employs fstatat(2) to
189 determine access permissions.
190 Glibc notes
192 On older kernels where faccessat() is unavailable (and when the AT_EAC‐
193 CESS and AT_SYMLINK_NOFOLLOW flags are not specified), the glibc wrap‐
194 per function falls back to the use of access(). When pathname is a
195 relative pathname, glibc constructs a pathname based on the symbolic
196 link in /proc/self/fd that corresponds to the dirfd argument.
197
199 In kernel 2.4 (and earlier) there is some strangeness in the handling
200 of X_OK tests for superuser. If all categories of execute permission
201 are disabled for a nondirectory file, then the only access() test that
202 returns -1 is when mode is specified as just X_OK; if R_OK or W_OK is
203 also specified in mode, then access() returns 0 for such files. Early
204 2.6 kernels (up to and including 2.6.3) also behaved in the same way as
205 kernel 2.4.
206 In kernels before 2.6.20, these calls ignored the effect of the
208 MS_NOEXEC flag if it was used to mount(2) the underlying filesystem.
209 Since kernel 2.6.20, the MS_NOEXEC flag is honored.
210
212 chmod(2), chown(2), open(2), setgid(2), setuid(2), stat(2), euidac‐
213 cess(3), credentials(7), path_resolution(7), symlink(7)
214
216 This page is part of release 5.04 of the Linux man-pages project. A
217 description of the project, information about reporting bugs, and the
218 latest version of this page, can be found at
219 https://www.kernel.org/doc/man-pages/.
220Linux 2016-03-15 ACCESS(2)