1rename(2) System Calls Manual rename(2)
2
6 rename, renameat, renameat2 - change the name or location of a file
7
9 Standard C library (libc, -lc)
10
12 #include <stdio.h>
13 int rename(const char *oldpath, const char *newpath);
15 #include <fcntl.h> /* Definition of AT_* constants */
17 #include <stdio.h>
18 int renameat(int olddirfd, const char *oldpath,
20 int newdirfd, const char *newpath);
21 int renameat2(int olddirfd, const char *oldpath,
22 int newdirfd, const char *newpath, unsigned int flags);
23 Feature Test Macro Requirements for glibc (see feature_test_macros(7)):
25 renameat():
27 Since glibc 2.10:
28 _POSIX_C_SOURCE >= 200809L
29 Before glibc 2.10:
30 _ATFILE_SOURCE
31 renameat2():
33 _GNU_SOURCE
34
36 rename() renames a file, moving it between directories if required.
37 Any other hard links to the file (as created using link(2)) are unaf‐
38 fected. Open file descriptors for oldpath are also unaffected.
39 Various restrictions determine whether or not the rename operation suc‐
41 ceeds: see ERRORS below.
42 If newpath already exists, it will be atomically replaced, so that
44 there is no point at which another process attempting to access newpath
45 will find it missing. However, there will probably be a window in
46 which both oldpath and newpath refer to the file being renamed.
47 If oldpath and newpath are existing hard links referring to the same
49 file, then rename() does nothing, and returns a success status.
50 If newpath exists but the operation fails for some reason, rename()
52 guarantees to leave an instance of newpath in place.
53 oldpath can specify a directory. In this case, newpath must either not
55 exist, or it must specify an empty directory.
56 If oldpath refers to a symbolic link, the link is renamed; if newpath
58 refers to a symbolic link, the link will be overwritten.
59 renameat()
61 The renameat() system call operates in exactly the same way as re‐
62 name(), except for the differences described here.
63 If the pathname given in oldpath is relative, then it is interpreted
65 relative to the directory referred to by the file descriptor olddirfd
66 (rather than relative to the current working directory of the calling
67 process, as is done by rename() for a relative pathname).
68 If oldpath is relative and olddirfd is the special value AT_FDCWD, then
70 oldpath is interpreted relative to the current working directory of the
71 calling process (like rename()).
72 If oldpath is absolute, then olddirfd is ignored.
74 The interpretation of newpath is as for oldpath, except that a relative
76 pathname is interpreted relative to the directory referred to by the
77 file descriptor newdirfd.
78 See openat(2) for an explanation of the need for renameat().
80 renameat2()
82 renameat2() has an additional flags argument. A renameat2() call with
83 a zero flags argument is equivalent to renameat().
84 The flags argument is a bit mask consisting of zero or more of the fol‐
86 lowing flags:
87 RENAME_EXCHANGE
89 Atomically exchange oldpath and newpath. Both pathnames must
90 exist but may be of different types (e.g., one could be a non-
91 empty directory and the other a symbolic link).
92 RENAME_NOREPLACE
94 Don't overwrite newpath of the rename. Return an error if new‐
95 path already exists.
96 RENAME_NOREPLACE can't be employed together with RENAME_EX‐
98 CHANGE.
99 RENAME_NOREPLACE requires support from the underlying filesys‐
101 tem. Support for various filesystems was added as follows:
102 • ext4 (Linux 3.15);
104 • btrfs, tmpfs, and cifs (Linux 3.17);
106 • xfs (Linux 4.0);
108 • Support for many other filesystems was added in Linux 4.9,
110 including ext2, minix, reiserfs, jfs, vfat, and bpf.
111 RENAME_WHITEOUT (since Linux 3.18)
113 This operation makes sense only for overlay/union filesystem im‐
114 plementations.
115 Specifying RENAME_WHITEOUT creates a "whiteout" object at the
117 source of the rename at the same time as performing the rename.
118 The whole operation is atomic, so that if the rename succeeds
119 then the whiteout will also have been created.
120 A "whiteout" is an object that has special meaning in
122 union/overlay filesystem constructs. In these constructs, mul‐
123 tiple layers exist and only the top one is ever modified. A
124 whiteout on an upper layer will effectively hide a matching file
125 in the lower layer, making it appear as if the file didn't ex‐
126 ist.
127 When a file that exists on the lower layer is renamed, the file
129 is first copied up (if not already on the upper layer) and then
130 renamed on the upper, read-write layer. At the same time, the
131 source file needs to be "whiteouted" (so that the version of the
132 source file in the lower layer is rendered invisible). The
133 whole operation needs to be done atomically.
134 When not part of a union/overlay, the whiteout appears as a
136 character device with a {0,0} device number. (Note that other
137 union/overlay implementations may employ different methods for
138 storing whiteout entries; specifically, BSD union mount employs
139 a separate inode type, DT_WHT, which, while supported by some
140 filesystems available in Linux, such as CODA and XFS, is ignored
141 by the kernel's whiteout support code, as of Linux 4.19, at
142 least.)
143 RENAME_WHITEOUT requires the same privileges as creating a de‐
145 vice node (i.e., the CAP_MKNOD capability).
146 RENAME_WHITEOUT can't be employed together with RENAME_EXCHANGE.
148 RENAME_WHITEOUT requires support from the underlying filesystem.
150 Among the filesystems that support it are tmpfs (since Linux
151 3.18), ext4 (since Linux 3.18), XFS (since Linux 4.1), f2fs
152 (since Linux 4.2), btrfs (since Linux 4.7), and ubifs (since
153 Linux 4.9).
154
156 On success, zero is returned. On error, -1 is returned, and errno is
157 set to indicate the error.
158
160 EACCES Write permission is denied for the directory containing oldpath
161 or newpath, or, search permission is denied for one of the di‐
162 rectories in the path prefix of oldpath or newpath, or oldpath
163 is a directory and does not allow write permission (needed to
164 update the .. entry). (See also path_resolution(7).)
165 EBUSY The rename fails because oldpath or newpath is a directory that
167 is in use by some process (perhaps as current working directory,
168 or as root directory, or because it was open for reading) or is
169 in use by the system (for example as a mount point), while the
170 system considers this an error. (Note that there is no require‐
171 ment to return EBUSY in such cases—there is nothing wrong with
172 doing the rename anyway—but it is allowed to return EBUSY if the
173 system cannot otherwise handle such situations.)
174 EDQUOT The user's quota of disk blocks on the filesystem has been ex‐
176 hausted.
177 EFAULT oldpath or newpath points outside your accessible address space.
179 EINVAL The new pathname contained a path prefix of the old, or, more
181 generally, an attempt was made to make a directory a subdirec‐
182 tory of itself.
183 EISDIR newpath is an existing directory, but oldpath is not a direc‐
185 tory.
186 ELOOP Too many symbolic links were encountered in resolving oldpath or
188 newpath.
189 EMLINK oldpath already has the maximum number of links to it, or it was
191 a directory and the directory containing newpath has the maximum
192 number of links.
193 ENAMETOOLONG
195 oldpath or newpath was too long.
196 ENOENT The link named by oldpath does not exist; or, a directory compo‐
198 nent in newpath does not exist; or, oldpath or newpath is an
199 empty string.
200 ENOMEM Insufficient kernel memory was available.
202 ENOSPC The device containing the file has no room for the new directory
204 entry.
205 ENOTDIR
207 A component used as a directory in oldpath or newpath is not, in
208 fact, a directory. Or, oldpath is a directory, and newpath ex‐
209 ists but is not a directory.
210 ENOTEMPTY or EEXIST
212 newpath is a nonempty directory, that is, contains entries other
213 than "." and "..".
214 EPERM or EACCES
216 The directory containing oldpath has the sticky bit (S_ISVTX)
217 set and the process's effective user ID is neither the user ID
218 of the file to be deleted nor that of the directory containing
219 it, and the process is not privileged (Linux: does not have the
220 CAP_FOWNER capability); or newpath is an existing file and the
221 directory containing it has the sticky bit set and the process's
222 effective user ID is neither the user ID of the file to be re‐
223 placed nor that of the directory containing it, and the process
224 is not privileged (Linux: does not have the CAP_FOWNER capabil‐
225 ity); or the filesystem containing oldpath does not support re‐
226 naming of the type requested.
227 EROFS The file is on a read-only filesystem.
229 EXDEV oldpath and newpath are not on the same mounted filesystem.
231 (Linux permits a filesystem to be mounted at multiple points,
232 but rename() does not work across different mount points, even
233 if the same filesystem is mounted on both.)
234 The following additional errors can occur for renameat() and re‐
236 nameat2():
237 EBADF oldpath (newpath) is relative but olddirfd (newdirfd) is not a
239 valid file descriptor.
240 ENOTDIR
242 oldpath is relative and olddirfd is a file descriptor referring
243 to a file other than a directory; or similar for newpath and
244 newdirfd
245 The following additional errors can occur for renameat2():
247 EEXIST flags contains RENAME_NOREPLACE and newpath already exists.
249 EINVAL An invalid flag was specified in flags.
251 EINVAL Both RENAME_NOREPLACE and RENAME_EXCHANGE were specified in
253 flags.
254 EINVAL Both RENAME_WHITEOUT and RENAME_EXCHANGE were specified in
256 flags.
257 EINVAL The filesystem does not support one of the flags in flags.
259 ENOENT flags contains RENAME_EXCHANGE and newpath does not exist.
261 EPERM RENAME_WHITEOUT was specified in flags, but the caller does not
263 have the CAP_MKNOD capability.
264
266 rename()
267 C11, POSIX.1-2008.
268 renameat()
270 POSIX.1-2008.
271 renameat2()
273 Linux.
274
276 rename()
277 4.3BSD, C89, POSIX.1-2001.
278 renameat()
280 Linux 2.6.16, glibc 2.4.
281 renameat2()
283 Linux 3.15, glibc 2.28.
284 glibc notes
286 On older kernels where renameat() is unavailable, the glibc wrapper
287 function falls back to the use of rename(). When oldpath and newpath
288 are relative pathnames, glibc constructs pathnames based on the sym‐
289 bolic links in /proc/self/fd that correspond to the olddirfd and
290 newdirfd arguments.
291
293 On NFS filesystems, you can not assume that if the operation failed,
294 the file was not renamed. If the server does the rename operation and
295 then crashes, the retransmitted RPC which will be processed when the
296 server is up again causes a failure. The application is expected to
297 deal with this. See link(2) for a similar problem.
298
300 mv(1), rename(1), chmod(2), link(2), symlink(2), unlink(2), path_reso‐
301 lution(7), symlink(7)
302Linux man-pages 6.04 2023-03-30 rename(2)