1 OPEN(2) System Calls OPEN(2)
2
3 NAME
4 open, openat - open a file
5
6 SYNOPSIS
7 #include <sys/types.h>
8 #include <sys/stat.h>
9 #include <fcntl.h>
10
11 int
12 open(const char *path, int oflag [, mode_t mode]);
13
14 int
15 openat(int fildes, const char *path, int oflag [, mode_t mode]);
16
17 DESCRIPTION
18 The open() function establishes the connection between a file and a file
19 descriptor. It creates an open file description that refers to a file
20 and a file descriptor that refers to that open file description. The
21 file descriptor is used by other I/O functions to refer to that file.
22 The path argument points to a pathname naming the file.
23
24 The openat() function is identical to the open() function except that the
25 path argument is interpreted relative to the starting point implied by
26 the fildes argument. If the fildes argument has the special value
27 AT_FDCWD, a relative path argument will be resolved relative to the
28 current working directory. If the path argument is absolute, the fildes
29 argument is ignored.
30
31 The open() function returns a file descriptor for the named file that is
32 the lowest file descriptor not currently open for that process. The open
33 file description is new, and therefore the file descriptor does not share
34 it with any other process in the system. The FD_CLOEXEC file descriptor
35 flag associated with the new file descriptor is cleared.
36
37 The file offset used to mark the current position within the file is set
38 to the beginning of the file.
39
40 The file status flags and file access modes of the open file description
41 are set according to the value of oflag. The mode argument is used only
42 when O_CREAT is specified (see below).
43
44 Values for oflag are constructed by a bitwise-inclusive-OR of flags from
45 the following list, defined in fcntl.h(3HEAD). Applications must specify
46 exactly one of the first three values (file access modes) below in the
47 value of oflag:
48
49 O_RDONLY
50 Open for reading only.
51
52 O_WRONLY
53 Open for writing only.
54
55 O_RDWR Open for reading and writing. The result is undefined if this
56 flag is applied to a FIFO.
57
58 Any combination of the following may be used:
59
60 O_APPEND
61 If set, the file offset is set to the end of the file prior to
62 each write.
63
64 O_CREAT
65 Create the file if it does not exist. This flag requires that
66 the mode argument be specified.
67
68 If the file exists, this flag has no effect except as noted under
69 O_EXCL below. Otherwise, the file is created with the user ID of
70 the file set to the effective user ID of the process. The group
71 ID of the file is set to the effective group IDs of the process,
72 or if the S_ISGID bit is set in the directory in which the file
73 is being created, the file's group ID is set to the group ID of
74 its parent directory. If the group ID of the new file does not
75 match the effective group ID or one of the supplementary groups
76 IDs, the S_ISGID bit is cleared.
77
78 The access permission bits (see stat.h(3HEAD)) of the file mode
79 are set to the value of mode, modified as follows (see creat(2)):
80 a bitwise-AND is performed on the file-mode bits and the
81 corresponding bits in the complement of the process's file mode
82 creation mask. Thus, all bits set in the process's file mode
83 creation mask (see umask(2)) are correspondingly cleared in the
84 file's permission mask. The "save text image after execution
85 bit" of the mode is cleared (see chmod(2)). When bits other than
86 the file permission bits are set, the effect is unspecified. The
87 mode argument does not affect whether the file is open for
88 reading, writing or for both.
89
90 O_DIRECT
91 Indicates that the file data is not going to be reused in the
92 near future. When possible, data is read or written directly
93 between the application's memory and the device when the data is
94 accessed with read(2) and write(2) operations. See directio(3C)
95 for more details.
96
97 O_DIRECTORY
98 Indicates that attempts to open path should fail unless path is a
99 directory. If both O_CREAT and O_DIRECTORY are specified then
100 the call will fail if it would result in a file being created.
101 If a directory already exists at path then it will behave as if
102 the O_DIRECTORY flag had not been present. If the O_EXCL and
103 O_CREAT flags are specified, then the call will always fail as
104 they imply a file should always be created.
105
106 O_DSYNC
107 Write I/O operations on the file descriptor complete as defined
108 by synchronized I/O data integrity completion.
109
110 O_EXCL If O_CREAT and O_EXCL are set, open() fails if the file exists.
111 The check for the existence of the file and the creation of the
112 file if it does not exist is atomic with respect to other threads
113 executing open() naming the same filename in the same directory
114 with O_EXCL and O_CREAT set. If O_EXCL and O_CREAT are set, and
115 path names a symbolic link, open() fails and sets errno to
116 EEXIST, regardless of the contents of the symbolic link. If
117 O_EXCL is set and O_CREAT is not set, the result is undefined.
118
119 O_EXEC If set, indicates that the file should be opened for execute
120 permission. This option is only valid for regular files; an
121 error will be returned if the target is not a regular file.
122
123 O_LARGEFILE
124 If set, the offset maximum in the open file description is the
125 largest value that can be represented correctly in an object of
126 type off64_t.
127
128 O_NOCTTY
129 If set and path identifies a terminal device, open() does not
130 cause the terminal device to become the controlling terminal for
131 the process.
132
133 O_NOFOLLOW
134 If the path names a symbolic link, open() fails and sets errno to
135 ELOOP.
136
137 O_NOLINKS
138 If the link count of the named file is greater than 1, open()
139 fails and sets errno to EMLINK.
140
141 O_CLOEXEC
142 If set, the file descriptor returned will be closed prior to any
143 future exec(2) calls.
144
145 O_NONBLOCK O_NDELAY
146 These flags can affect subsequent reads and writes (see read(2)
147 and write(2)). If both O_NDELAY and O_NONBLOCK are set,
148 O_NONBLOCK takes precedence.
149
150 When opening a FIFO with O_RDONLY or O_WRONLY set:
151
152 o If O_NONBLOCK or O_NDELAY is set, an open() for reading only
153 returns without delay. An open() for writing only returns an
154 error if no process currently has the file open for reading.
155
156 o If O_NONBLOCK and O_NDELAY are clear, an open() for reading
157 only blocks until a thread opens the file for writing. An
158 open() for writing only blocks the calling thread until a
159 thread opens the file for reading.
160
161 After both ends of a FIFO have been opened once, there is no
162 guarantee that further calls to open() O_RDONLY (O_WRONLY) will
163 synchronize with later calls to open() O_WRONLY (O_RDONLY) until
164 both ends of the FIFO have been closed by all readers and
165 writers. Any data written into a FIFO will be lost if both ends
166 of the FIFO are closed before the data is read.
167
168 When opening a block special or character special file that
169 supports non-blocking opens:
170
171 o If O_NONBLOCK or O_NDELAY is set, the open() function returns
172 without blocking for the device to be ready or available.
173 Subsequent behavior of the device is device-specific.
174
175 o If O_NONBLOCK and O_NDELAY are clear, the open() function
176 blocks the calling thread until the device is ready or
177 available before returning.
178
179 Otherwise, the behavior of O_NONBLOCK and O_NDELAY is
180 unspecified.
181
182 O_RSYNC
183 Read I/O operations on the file descriptor complete at the same
184 level of integrity as specified by the O_DSYNC and O_SYNC flags.
185 If both O_DSYNC and O_RSYNC are set in oflag, all I/O operations
186 on the file descriptor complete as defined by synchronized I/O
187 data integrity completion. If both O_SYNC and O_RSYNC are set in
188 oflag, all I/O operations on the file descriptor complete as
189 defined by synchronized I/O file integrity completion.
190
191 O_SEARCH
192 If set, indicates that the directory should be opened for
193 searching. This option is only valid for a directory; an error
194 will be returned if the target is not a directory.
195
196 O_SYNC Write I/O operations on the file descriptor complete as defined
197 by synchronized I/O file integrity completion (see
198 fcntl.h(3HEAD)) definition of O_SYNC.
199
200 O_TRUNC
201 If the file exists and is a regular file, and the file is
202 successfully opened O_RDWR or O_WRONLY, its length is truncated
203 to 0 and the mode and owner are unchanged. It has no effect on
204 FIFO special files or terminal device files. Its effect on other
205 file types is implementation-dependent. The result of using
206 O_TRUNC with O_RDONLY is undefined.
207
208 O_XATTR
209 If set in openat(), a relative path argument is interpreted as a
210 reference to an extended attribute of the file associated with
211 the supplied file descriptor. This flag therefore requires the
212 presence of a legal fildes argument. If set in open(), the
213 implied file descriptor is that for the current working
214 directory. Extended attributes must be referenced with a
215 relative path; providing an absolute path results in a normal
216 file reference.
217
218 If O_CREAT is set and the file did not previously exist, upon successful
219 completion, open() marks for update the st_atime, st_ctime, and st_mtime
220 fields of the file and the st_ctime and st_mtime fields of the parent
221 directory.
222
223 If O_TRUNC is set and the file did previously exist, upon successful
224 completion, open() marks for update the st_ctime and st_mtime fields of
225 the file.
226
227 If both the O_SYNC and O_DSYNC flags are set, the effect is as if only
228 the O_SYNC flag was set.
229
230 If path refers to a STREAMS file, oflag may be constructed from
231 O_NONBLOCK or O_NODELAY OR-ed with either O_RDONLY, O_WRONLY, or O_RDWR.
232 Other flag values are not applicable to STREAMS devices and have no
233 effect on them. The values O_NONBLOCK and O_NODELAY affect the operation
234 of STREAMS drivers and certain functions (see read(2), getmsg(2),
235 putmsg(2), and write(2)) applied to file descriptors associated with
236 STREAMS files. For STREAMS drivers, the implementation of O_NONBLOCK and
237 O_NODELAY is device-specific.
238
239 When open() is invoked to open a named stream, and the connld(7M) module
240 has been pushed on the pipe, open() blocks until the server process has
241 issued an I_RECVFD ioctl(2) (see streamio(7I)) to receive the file
242 descriptor.
243
244 If path names the manager side of a pseudo-terminal device, then it is
245 unspecified whether open() locks the subsidiary side so that it cannot be
246 opened. Portable applications must call unlockpt(3C) before opening the
247 subsidiary side.
248
249 If the file is a regular file and the local file system is mounted with
250 the nbmand mount option, then a mandatory share reservation is
251 automatically obtained on the file. The share reservation is obtained as
252 if fcntl(2) were called with cmd F_SHARE_NBMAND and the fshare_t values
253 set as follows:
254
255 f_access
256 Set to the type of read/write access for which the file is
257 opened.
258
259 f_deny F_NODNY
260
261 f_id The file descriptor value returned from open().
262
263 If path is a symbolic link and O_CREAT and O_EXCL are set, the link is
264 not followed.
265
266 Certain flag values can be set following open() as described in fcntl(2).
267
268 The largest value that can be represented correctly in an object of type
269 off_t is established as the offset maximum in the open file description.
270
271 RETURN VALUES
272 The open() and openat() functions open the file and, if successful,
273 return a non-negative integer representing the lowest numbered unused
274 file descriptor; otherwise the value -1 is returned and the global
275 variable errno is set to indicate the error and no files are created or
276 modified.
277
278 EXAMPLES
279 Example 1 Open a file for writing by the owner.
280
281 The following example opens the file /tmp/file, either by creating it if
282 it does not already exist, or by truncating its length to 0 if it does
283 exist. If the call creates a new file, the access permission bits in the
284 file mode of the file are set to permit reading and writing by the owner,
285 and to permit reading only by group members and others.
286
287 If the call to open() is successful, the file is opened for writing.
288
289 #include <fcntl.h>
290 ...
291 int fd;
292 mode_t mode = S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH;
293 char *filename = "/tmp/file";
294 ...
295 fd = open(filename, O_WRONLY | O_CREAT | O_TRUNC, mode);
296 ...
297
298 Example 2 Open a file using an existence check.
299
300 The following example uses the open() function to try to create the
301 LOCKFILE file and open it for writing. Since the open() function
302 specifies the O_EXCL flag, the call fails if the file already exists. In
303 that case, the application assumes that someone else is updating the
304 password file and exits.
305
306 #include <fcntl.h>
307 #include <stdio.h>
308 #include <stdlib.h>
309 #include <err.h>
310 ...
311 #define LOCKFILE "/etc/ptmp"
312 ...
313 int pfd; /* Integer for file descriptor returned by open() call. */
314 ...
315 if ((pfd = open(LOCKFILE, O_WRONLY | O_CREAT | O_EXCL,
316 S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH)) < 0) {
317 err(1, "Cannot open %s. Try again later.", LOCKFILE);
318 }
319 ...
320
321 Example 3 Open a file for writing.
322
323 The following example opens a file for writing, creating the file if it
324 does not already exist. If the file does exist, the system truncates the
325 file to zero bytes.
326
327 #include <fcntl.h>
328 #include <stdio.h>
329 #include <stdlib.h>
330 #include <err.h>
331 ...
332 int pfd;
333 char filename[PATH_MAX+1];
334 ...
335 if ((pfd = open(filename, O_WRONLY | O_CREAT | O_TRUNC,
336 S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH)) < 0) {
337 err(1, "Cannot open output file");
338 }
339 ...
340
341 ERRORS
342 The open() and openat() functions will fail if:
343
344 EACCES Search permission is denied on a component of the path
345 prefix.
346
347 The file exists and the permissions specified by oflag
348 are denied.
349
350 The file does not exist and write permission is denied
351 for the parent directory of the file to be created.
352
353 O_TRUNC is specified and write permission is denied.
354
355 The {PRIV_FILE_DAC_SEARCH} privilege allows processes
356 to search directories regardless of permission bits.
357 The {PRIV_FILE_DAC_WRITE} privilege allows processes
358 to open files for writing regardless of permission
359 bits. See privileges(5) for special considerations
360 when opening files owned by user ID 0 for writing.
361 The {PRIV_FILE_DAC_READ} privilege allows processes to
362 open files for reading regardless of permission bits.
363
364 EAGAIN A mandatory share reservation could not be obtained
365 because the desired access conflicts with an existing
366 f_deny share reservation (see fcntl(2)).
367
368 EDQUOT The file does not exist, O_CREAT is specified, and
369 either the directory where the new file entry is being
370 placed cannot be extended because the user's quota of
371 disk blocks on that file system has been exhausted, or
372 the user's quota of inodes on the file system where
373 the file is being created has been exhausted.
374
375 EEXIST The O_CREAT and O_EXCL flags are set and the named
376 file already exists.
377
378 EILSEQ The path argument includes bytes that are not valid
379 UTF-8 characters, and the file system accepts only
380 file names where all characters are part of the UTF-8
381 character codeset.
382
383 EINTR A signal was caught during open().
384
385 EFAULT The path argument points to an illegal address.
386
387 EINVAL Either the system does not support synchronized or
388 direct I/O for this file, or the O_XATTR flag was
389 supplied and the underlying file system does not
390 support extended file attributes.
391
392 EIO The path argument names a STREAMS file and a hangup or
393 error occurred during the open().
394
395 EISDIR The named file is a directory and oflag includes
396 O_WRONLY or O_RDWR.
397
398 ELOOP Too many symbolic links were encountered in resolving
399 path.
400
401 A loop exists in symbolic links encountered during
402 resolution of the path argument.
403
404 The O_NOFOLLOW flag is set and the final component of
405 path is a symbolic link.
406
407 EMFILE There are currently {OPEN_MAX} file descriptors open
408 in the calling process.
409
410 EMLINK The O_NOLINKS flag is set and the named file has a
411 link count greater than 1.
412
413 EMULTIHOP Components of path require hopping to multiple remote
414 machines and the file system does not allow it.
415
416 ENAMETOOLONG The length of the path argument exceeds {PATH_MAX} or
417 a pathname component is longer than {NAME_MAX}.
418
419 ENFILE The maximum allowable number of files is currently
420 open in the system.
421
422 ENOENT The O_CREAT flag is not set and the named file does
423 not exist; or the O_CREAT flag is set and either the
424 path prefix does not exist or the path argument points
425 to an empty string.
426
427 The O_CREAT and O_DIRECTORY flags were both set and
428 path did not point to a file.
429
430 ENOEXEC The O_EXEC flag is set and path does not point to a
431 regular file.
432
433 ENOLINK The path argument points to a remote machine, and the
434 link to that machine is no longer active.
435
436 ENOSR Th path argument names a STREAMS-based file and the
437 system is unable to allocate a STREAM.
438
439 ENOSPC The directory or file system that would contain the
440 new file cannot be expanded, the file does not exist,
441 and O_CREAT is specified.
442
443 ENOSYS The device specified by path does not support the open
444 operation.
445
446 ENOTDIR A component of the path prefix is not a directory or a
447 relative path was supplied to openat(), the O_XATTR
448 flag was not supplied, and the file descriptor does
449 not refer to a directory. The O_SEARCH flag was
450 passed and path does not refer to a directory.
451
452 The O_DIRECTORY flag was set and the file was not a
453 directory.
454
455 ENXIO The O_NONBLOCK flag is set, the named file is a FIFO,
456 the O_WRONLY flag is set, and no process has the file
457 open for reading; or the named file is a character
458 special or block special file and the device
459 associated with this special file does not exist or
460 has been retired by the fault management framework.
461
462 EOPNOTSUPP An attempt was made to open a path that corresponds to
463 an AF_UNIX socket.
464
465 EOVERFLOW The named file is a regular file and either
466 O_LARGEFILE is not set and the size of the file cannot
467 be represented correctly in an object of type off_t or
468 O_LARGEFILE is set and the size of the file cannot be
469 represented correctly in an object of type off64_t.
470
471 EROFS The named file resides on a read-only file system and
472 either O_WRONLY, O_RDWR, O_CREAT (if file does not
473 exist), or O_TRUNC is set in the oflag argument.
474
475 The openat() function will fail if:
476
477 EBADF The fildes argument is not a valid open file
478 descriptor or is not AT_FTCWD.
479
480 The open() function may fail if:
481
482 EAGAIN The path argument names the subsidiary side of a
483 pseudo-terminal device that is locked.
484
485 EINVAL The value of the oflag argument is not valid.
486
487 ENAMETOOLONG Pathname resolution of a symbolic link produced an
488 intermediate result whose length exceeds {PATH_MAX}.
489
490 ENOMEM The path argument names a STREAMS file and the system
491 is unable to allocate resources.
492
493 ETXTBSY The file is a pure procedure (shared text) file that
494 is being executed and oflag is O_WRONLY or O_RDWR.
495
496 USAGE
497 The open() function has a transitional interface for 64-bit file offsets.
498 See lf64(5). Note that using open64() is equivalent to using open(with)
499 O_LARGEFILE set in oflag.
500
501 INTERFACE STABILITY
502 Committed
503
504 MT LEVEL
505 Async-Signal-Safe
506
507 SEE ALSO
508 chmod(2), close(2), creat(2), dup(2), exec(2), fcntl(2), getmsg(2),
509 getrlimit(2), Intro(2), lseek(2), putmsg(2), read(2), stat(2), umask(2),
510 write(2), attropen(3C), directio(3C), unlockpt(3C), fcntl.h(3HEAD),
511 stat.h(3HEAD), attributes(5), lf64(5), privileges(5), standards(5),
512 streamio(7I), connld(7M)
513
514 NOTES
515 Hierarchical Storage Management (HSM) file systems can sometimes cause
516 long delays when opening a file, since HSM files must be recalled from
517 secondary storage.
518
519 illumos February 5, 2022 illumos