1WRITE(3P) POSIX Programmer's Manual WRITE(3P)
2
6 This manual page is part of the POSIX Programmer's Manual. The Linux
7 implementation of this interface may differ (consult the corresponding
8 Linux manual page for details of Linux behavior), or the interface may
9 not be implemented on Linux.
10
13 pwrite, write — write on a file
14
16 #include <unistd.h>
17 ssize_t pwrite(int fildes, const void *buf, size_t nbyte,
19 off_t offset);
20 ssize_t write(int fildes, const void *buf, size_t nbyte);
21
23 The write() function shall attempt to write nbyte bytes from the buffer
24 pointed to by buf to the file associated with the open file descriptor,
25 fildes.
26 Before any action described below is taken, and if nbyte is zero and
28 the file is a regular file, the write() function may detect and return
29 errors as described below. In the absence of errors, or if error detec‐
30 tion is not performed, the write() function shall return zero and have
31 no other results. If nbyte is zero and the file is not a regular file,
32 the results are unspecified.
33 On a regular file or other file capable of seeking, the actual writing
35 of data shall proceed from the position in the file indicated by the
36 file offset associated with fildes. Before successful return from
37 write(), the file offset shall be incremented by the number of bytes
38 actually written. On a regular file, if the position of the last byte
39 written is greater than or equal to the length of the file, the length
40 of the file shall be set to this position plus one.
41 On a file not capable of seeking, writing shall always take place
43 starting at the current position. The value of a file offset associated
44 with such a device is undefined.
45 If the O_APPEND flag of the file status flags is set, the file offset
47 shall be set to the end of the file prior to each write and no inter‐
48 vening file modification operation shall occur between changing the
49 file offset and the write operation.
50 If a write() requests that more bytes be written than there is room for
52 (for example, the file size limit of the process or the physical end of
53 a medium), only as many bytes as there is room for shall be written.
54 For example, suppose there is space for 20 bytes more in a file before
55 reaching a limit. A write of 512 bytes will return 20. The next write
56 of a non-zero number of bytes would give a failure return (except as
57 noted below).
58 If the request would cause the file size to exceed the soft file size
60 limit for the process and there is no room for any bytes to be written,
61 the request shall fail and the implementation shall generate the
62 SIGXFSZ signal for the thread.
63 If write() is interrupted by a signal before it writes any data, it
65 shall return −1 with errno set to [EINTR].
66 If write() is interrupted by a signal after it successfully writes some
68 data, it shall return the number of bytes written.
69 If the value of nbyte is greater than {SSIZE_MAX}, the result is imple‐
71 mentation-defined.
72 After a write() to a regular file has successfully returned:
74 * Any successful read() from each byte position in the file that was
76 modified by that write shall return the data specified by the
77 write() for that position until such byte positions are again modi‐
78 fied.
79 * Any subsequent successful write() to the same byte position in the
81 file shall overwrite that file data.
82 Write requests to a pipe or FIFO shall be handled in the same way as a
84 regular file with the following exceptions:
85 * There is no file offset associated with a pipe, hence each write
87 request shall append to the end of the pipe.
88 * Write requests of {PIPE_BUF} bytes or less shall not be interleaved
90 with data from other processes doing writes on the same pipe.
91 Writes of greater than {PIPE_BUF} bytes may have data interleaved,
92 on arbitrary boundaries, with writes by other processes, whether or
93 not the O_NONBLOCK flag of the file status flags is set.
94 * If the O_NONBLOCK flag is clear, a write request may cause the
96 thread to block, but on normal completion it shall return nbyte.
97 * If the O_NONBLOCK flag is set, write() requests shall be handled
99 differently, in the following ways:
100 -- The write() function shall not block the thread.
102 -- A write request for {PIPE_BUF} or fewer bytes shall have the
104 following effect: if there is sufficient space available in the
105 pipe, write() shall transfer all the data and return the number
106 of bytes requested. Otherwise, write() shall transfer no data
107 and return −1 with errno set to [EAGAIN].
108 -- A write request for more than {PIPE_BUF} bytes shall cause one
110 of the following:
111 -- When at least one byte can be written, transfer what it can
113 and return the number of bytes written. When all data pre‐
114 viously written to the pipe is read, it shall transfer at
115 least {PIPE_BUF} bytes.
116 -- When no data can be written, transfer no data, and return
118 −1 with errno set to [EAGAIN].
119 When attempting to write to a file descriptor (other than a pipe or
121 FIFO) that supports non-blocking writes and cannot accept the data
122 immediately:
123 * If the O_NONBLOCK flag is clear, write() shall block the calling
125 thread until the data can be accepted.
126 * If the O_NONBLOCK flag is set, write() shall not block the thread.
128 If some data can be written without blocking the thread, write()
129 shall write what it can and return the number of bytes written.
130 Otherwise, it shall return −1 and set errno to [EAGAIN].
131 Upon successful completion, where nbyte is greater than 0, write()
133 shall mark for update the last data modification and last file status
134 change timestamps of the file, and if the file is a regular file, the
135 S_ISUID and S_ISGID bits of the file mode may be cleared.
136 For regular files, no data transfer shall occur past the offset maximum
138 established in the open file description associated with fildes.
139 If fildes refers to a socket, write() shall be equivalent to send()
141 with no flags set.
142 If the O_DSYNC bit has been set, write I/O operations on the file
144 descriptor shall complete as defined by synchronized I/O data integrity
145 completion.
146 If the O_SYNC bit has been set, write I/O operations on the file
148 descriptor shall complete as defined by synchronized I/O file integrity
149 completion.
150 If fildes refers to a shared memory object, the result of the write()
152 function is unspecified.
153 If fildes refers to a typed memory object, the result of the write()
155 function is unspecified.
156 If fildes refers to a STREAM, the operation of write() shall be deter‐
158 mined by the values of the minimum and maximum nbyte range (packet
159 size) accepted by the STREAM. These values are determined by the top‐
160 most STREAM module. If nbyte falls within the packet size range, nbyte
161 bytes shall be written. If nbyte does not fall within the range and the
162 minimum packet size value is 0, write() shall break the buffer into
163 maximum packet size segments prior to sending the data downstream (the
164 last segment may contain less than the maximum packet size). If nbyte
165 does not fall within the range and the minimum value is non-zero,
166 write() shall fail with errno set to [ERANGE]. Writing a zero-length
167 buffer (nbyte is 0) to a STREAMS device sends 0 bytes with 0 returned.
168 However, writing a zero-length buffer to a STREAMS-based pipe or FIFO
169 sends no message and 0 is returned. The process may issue I_SWROPT
170 ioctl() to enable zero-length messages to be sent across the pipe or
171 FIFO.
172 When writing to a STREAM, data messages are created with a priority
174 band of 0. When writing to a STREAM that is not a pipe or FIFO:
175 * If O_NONBLOCK is clear, and the STREAM cannot accept data (the
177 STREAM write queue is full due to internal flow control condi‐
178 tions), write() shall block until data can be accepted.
179 * If O_NONBLOCK is set and the STREAM cannot accept data, write()
181 shall return −1 and set errno to [EAGAIN].
182 * If O_NONBLOCK is set and part of the buffer has been written while
184 a condition in which the STREAM cannot accept additional data
185 occurs, write() shall terminate and return the number of bytes
186 written.
187 In addition, write() shall fail if the STREAM head has processed an
189 asynchronous error before the call. In this case, the value of errno
190 does not reflect the result of write(), but reflects the prior error.
191 The pwrite() function shall be equivalent to write(), except that it
193 writes into a given position and does not change the file offset
194 (regardless of whether O_APPEND is set). The first three arguments to
195 pwrite() are the same as write() with the addition of a fourth argument
196 offset for the desired position inside the file. An attempt to perform
197 a pwrite() on a file that is incapable of seeking shall result in an
198 error.
199
201 Upon successful completion, these functions shall return the number of
202 bytes actually written to the file associated with fildes. This number
203 shall never be greater than nbyte. Otherwise, −1 shall be returned and
204 errno set to indicate the error.
205
207 These functions shall fail if:
208 EAGAIN The file is neither a pipe, nor a FIFO, nor a socket, the O_NON‐
210 BLOCK flag is set for the file descriptor, and the thread would
211 be delayed in the write() operation.
212 EBADF The fildes argument is not a valid file descriptor open for
214 writing.
215 EFBIG An attempt was made to write a file that exceeds the implementa‐
217 tion-defined maximum file size or the file size limit of the
218 process, and there was no room for any bytes to be written.
219 EFBIG The file is a regular file, nbyte is greater than 0, and the
221 starting position is greater than or equal to the offset maximum
222 established in the open file description associated with fildes.
223 EINTR The write operation was terminated due to the receipt of a sig‐
225 nal, and no data was transferred.
226 EIO The process is a member of a background process group attempting
228 to write to its controlling terminal, TOSTOP is set, the calling
229 thread is not blocking SIGTTOU, the process is not ignoring
230 SIGTTOU, and the process group of the process is orphaned. This
231 error may also be returned under implementation-defined condi‐
232 tions.
233 ENOSPC There was no free space remaining on the device containing the
235 file.
236 ERANGE The transfer request size was outside the range supported by the
238 STREAMS file associated with fildes.
239 The pwrite() function shall fail if:
241 EINVAL The file is a regular file or block special file, and the offset
243 argument is negative. The file pointer shall remain unchanged.
244 ESPIPE The file is a pipe, FIFO, or socket.
246 The write() function shall fail if:
248 EAGAIN The file is a pipe or FIFO, the O_NONBLOCK flag is set for the
250 file descriptor, and the thread would be delayed in the write
251 operation.
252 EAGAIN or EWOULDBLOCK
254 The file is a socket, the O_NONBLOCK flag is set for the file
255 descriptor, and the thread would be delayed in the write opera‐
256 tion.
257 ECONNRESET
259 A write was attempted on a socket that is not connected.
260 EPIPE An attempt is made to write to a pipe or FIFO that is not open
262 for reading by any process, or that only has one end open. A
263 SIGPIPE signal shall also be sent to the thread.
264 EPIPE A write was attempted on a socket that is shut down for writing,
266 or is no longer connected. In the latter case, if the socket is
267 of type SOCK_STREAM, a SIGPIPE signal shall also be sent to the
268 thread.
269 These functions may fail if:
271 EINVAL The STREAM or multiplexer referenced by fildes is linked
273 (directly or indirectly) downstream from a multiplexer.
274 EIO A physical I/O error has occurred.
276 ENOBUFS
278 Insufficient resources were available in the system to perform
279 the operation.
280 ENXIO A request was made of a nonexistent device, or the request was
282 outside the capabilities of the device.
283 ENXIO A hangup occurred on the STREAM being written to.
285 A write to a STREAMS file may fail if an error message has been
287 received at the STREAM head. In this case, errno is set to the value
288 included in the error message.
289 The write() function may fail if:
291 EACCES A write was attempted on a socket and the calling process does
293 not have appropriate privileges.
294 ENETDOWN
296 A write was attempted on a socket and the local network inter‐
297 face used to reach the destination is down.
298 ENETUNREACH
300 A write was attempted on a socket and no route to the network is
301 present.
302 The following sections are informative.
304
306 Writing from a Buffer
307 The following example writes data from the buffer pointed to by buf to
308 the file associated with the file descriptor fd.
309 #include <sys/types.h>
311 #include <string.h>
312 ...
313 char buf[20];
314 size_t nbytes;
315 ssize_t bytes_written;
316 int fd;
317 ...
318 strcpy(buf, "This is a test\n");
319 nbytes = strlen(buf);
320 bytes_written = write(fd, buf, nbytes);
322 ...
323
325 None.
326
328 See also the RATIONALE section in read().
329 An attempt to write to a pipe or FIFO has several major characteris‐
331 tics:
332 * Atomic/non-atomic: A write is atomic if the whole amount written in
334 one operation is not interleaved with data from any other process.
335 This is useful when there are multiple writers sending data to a
336 single reader. Applications need to know how large a write request
337 can be expected to be performed atomically. This maximum is called
338 {PIPE_BUF}. This volume of POSIX.1‐2008 does not say whether write
339 requests for more than {PIPE_BUF} bytes are atomic, but requires
340 that writes of {PIPE_BUF} or fewer bytes shall be atomic.
341 * Blocking/immediate: Blocking is only possible with O_NONBLOCK
343 clear. If there is enough space for all the data requested to be
344 written immediately, the implementation should do so. Otherwise,
345 the calling thread may block; that is, pause until enough space is
346 available for writing. The effective size of a pipe or FIFO (the
347 maximum amount that can be written in one operation without block‐
348 ing) may vary dynamically, depending on the implementation, so it
349 is not possible to specify a fixed value for it.
350 * Complete/partial/deferred: A write request:
352 int fildes;
354 size_t nbyte;
355 ssize_t ret;
356 char *buf;
357 ret = write(fildes, buf, nbyte);
359 may return:
361 Complete ret=nbyte
363 Partial ret<nbyte
365 This shall never happen if nbyte≤{PIPE_BUF}. If it does
367 happen (with nbyte>{PIPE_BUF}), this volume of
368 POSIX.1‐2008 does not guarantee atomicity, even if
369 ret≤{PIPE_BUF}, because atomicity is guaranteed according
370 to the amount requested, not the amount written.
371 Deferred: ret=−1, errno=[EAGAIN]
373 This error indicates that a later request may succeed. It
375 does not indicate that it shall succeed, even if
376 nbyte≤{PIPE_BUF}, because if no process reads from the
377 pipe or FIFO, the write never succeeds. An application
378 could usefully count the number of times [EAGAIN] is
379 caused by a particular value of nbyte>{PIPE_BUF} and per‐
380 haps do later writes with a smaller value, on the assump‐
381 tion that the effective size of the pipe may have
382 decreased.
383 Partial and deferred writes are only possible with O_NONBLOCK set.
385 The relations of these properties are shown in the following tables:
387 ┌───────────────────────────────────────────────────────────────────────┐
389 │ Write to a Pipe or FIFO with O_NONBLOCK clear │
390 ├─────────────────────┬─────────────────────────────────────────────────┤
391 │Immediately Writable:│ None Some nbyte │
392 ├─────────────────────┼─────────────────────────────────────────────────┤
393 │nbyte≤{PIPE_BUF} │Atomic blocking Atomic blocking Atomic immediate │
394 │ │nbyte nbyte nbyte │
395 ├─────────────────────┼─────────────────────────────────────────────────┤
396 │nbyte>{PIPE_BUF} │Blocking nbyte Blocking nbyte Blocking nbyte │
397 └─────────────────────┴─────────────────────────────────────────────────┘
398 If the O_NONBLOCK flag is clear, a write request shall block if the
399 amount writable immediately is less than that requested. If the flag is
400 set (by fcntl()), a write request shall never block.
401 ┌───────────────────────────────────────────────────────────────┐
403 │ Write to a Pipe or FIFO with O_NONBLOCK set │
404 ├─────────────────────┬─────────────────────────────────────────┤
405 │Immediately Writable:│ None Some nbyte │
406 ├─────────────────────┼─────────────────────────────────────────┤
407 │nbyte≤{PIPE_BUF} │−1, [EAGAIN] −1, [EAGAIN] Atomic nbyte │
408 ├─────────────────────┼─────────────────────────────────────────┤
409 │nbyte>{PIPE_BUF} │−1, [EAGAIN] <nbyte or −1, ≤nbyte or −1, │
410 │ │ [EAGAIN] [EAGAIN] │
411 └─────────────────────┴─────────────────────────────────────────┘
412 There is no exception regarding partial writes when O_NONBLOCK is set.
413 With the exception of writing to an empty pipe, this volume of
414 POSIX.1‐2008 does not specify exactly when a partial write is performed
415 since that would require specifying internal details of the implementa‐
416 tion. Every application should be prepared to handle partial writes
417 when O_NONBLOCK is set and the requested amount is greater than
418 {PIPE_BUF}, just as every application should be prepared to handle par‐
419 tial writes on other kinds of file descriptors.
420 The intent of forcing writing at least one byte if any can be written
422 is to assure that each write makes progress if there is any room in the
423 pipe. If the pipe is empty, {PIPE_BUF} bytes must be written; if not,
424 at least some progress must have been made.
425 Where this volume of POSIX.1‐2008 requires −1 to be returned and errno
427 set to [EAGAIN], most historical implementations return zero (with the
428 O_NDELAY flag set, which is the historical predecessor of O_NONBLOCK,
429 but is not itself in this volume of POSIX.1‐2008). The error indica‐
430 tions in this volume of POSIX.1‐2008 were chosen so that an application
431 can distinguish these cases from end-of-file. While write() cannot
432 receive an indication of end-of-file, read() can, and the two functions
433 have similar return values. Also, some existing systems (for example,
434 Eighth Edition) permit a write of zero bytes to mean that the reader
435 should get an end-of-file indication; for those systems, a return value
436 of zero from write() indicates a successful write of an end-of-file
437 indication.
438 Implementations are allowed, but not required, to perform error check‐
440 ing for write() requests of zero bytes.
441 The concept of a {PIPE_MAX} limit (indicating the maximum number of
443 bytes that can be written to a pipe in a single operation) was consid‐
444 ered, but rejected, because this concept would unnecessarily limit
445 application writing.
446 See also the discussion of O_NONBLOCK in read().
448 Writes can be serialized with respect to other reads and writes. If a
450 read() of file data can be proven (by any means) to occur after a
451 write() of the data, it must reflect that write(), even if the calls
452 are made by different processes. A similar requirement applies to mul‐
453 tiple write operations to the same file position. This is needed to
454 guarantee the propagation of data from write() calls to subsequent
455 read() calls. This requirement is particularly significant for net‐
456 worked file systems, where some caching schemes violate these seman‐
457 tics.
458 Note that this is specified in terms of read() and write(). The XSI
460 extensions readv() and writev() also obey these semantics. A new
461 ``high-performance'' write analog that did not follow these serializa‐
462 tion requirements would also be permitted by this wording. This volume
463 of POSIX.1‐2008 is also silent about any effects of application-level
464 caching (such as that done by stdio).
465 This volume of POSIX.1‐2008 does not specify the value of the file off‐
467 set after an error is returned; there are too many cases. For program‐
468 ming errors, such as [EBADF], the concept is meaningless since no file
469 is involved. For errors that are detected immediately, such as
470 [EAGAIN], clearly the pointer should not change. After an interrupt or
471 hardware error, however, an updated value would be very useful and is
472 the behavior of many implementations.
473 This volume of POSIX.1‐2008 does not specify behavior of concurrent
475 writes to a file from multiple processes. Applications should use some
476 form of concurrency control.
477 This volume of POSIX.1‐2008 intentionally does not specify any pwrite()
479 errors related to pipes, FIFOs, and sockets other than [ESPIPE].
480
482 None.
483
485 chmod(), creat(), dup(), fcntl(), getrlimit(), lseek(), open(), pipe(),
486 read(), ulimit(), writev()
487 The Base Definitions volume of POSIX.1‐2008, <limits.h>, <stropts.h>,
489 <sys_uio.h>, <unistd.h>
490
492 Portions of this text are reprinted and reproduced in electronic form
493 from IEEE Std 1003.1, 2013 Edition, Standard for Information Technology
494 -- Portable Operating System Interface (POSIX), The Open Group Base
495 Specifications Issue 7, Copyright (C) 2013 by the Institute of Electri‐
496 cal and Electronics Engineers, Inc and The Open Group. (This is
497 POSIX.1-2008 with the 2013 Technical Corrigendum 1 applied.) In the
498 event of any discrepancy between this version and the original IEEE and
499 The Open Group Standard, the original IEEE and The Open Group Standard
500 is the referee document. The original Standard can be obtained online
501 at http://www.unix.org/online.html .
502 Any typographical or formatting errors that appear in this page are
504 most likely to have been introduced during the conversion of the source
505 files to man page format. To report such errors, see https://www.ker‐
506 nel.org/doc/man-pages/reporting_bugs.html .
507IEEE/The Open Group 2013 WRITE(3P)