1write(2)                      System Calls Manual                     write(2)


6       write - write to a file descriptor


9       Standard C library (libc, -lc)


12       #include <unistd.h>
14       ssize_t write(int fd, const void buf[.count], size_t count);


17       write() writes up to count bytes from the buffer starting at buf to the
18       file referred to by the file descriptor fd.
20       The number of bytes written may be less than  count  if,  for  example,
21       there  is  insufficient space on the underlying physical medium, or the
22       RLIMIT_FSIZE resource limit is encountered (see setrlimit(2)),  or  the
23       call was interrupted by a signal handler after having written less than
24       count bytes.  (See also pipe(7).)
26       For a seekable file (i.e., one to which lseek(2) may  be  applied,  for
27       example,  a  regular  file) writing takes place at the file offset, and
28       the file offset is incremented by the number of bytes actually written.
29       If  the  file was open(2)ed with O_APPEND, the file offset is first set
30       to the end of the file before writing.  The adjustment of the file off‐
31       set and the write operation are performed as an atomic step.
33       POSIX  requires  that  a  read(2)  that  can be proved to occur after a
34       write() has returned will return the  new  data.   Note  that  not  all
35       filesystems are POSIX conforming.
37       According to POSIX.1, if count is greater than SSIZE_MAX, the result is
38       implementation-defined; see NOTES for the upper limit on Linux.


41       On success, the number of bytes written is returned.  On error,  -1  is
42       returned, and errno is set to indicate the error.
44       Note  that  a  successful  write() may transfer fewer than count bytes.
45       Such partial writes can occur for various reasons; for example, because
46       there was insufficient space on the disk device to write all of the re‐
47       quested bytes, or because a blocked write() to a socket, pipe, or simi‐
48       lar  was interrupted by a signal handler after it had transferred some,
49       but before it had transferred all of the requested bytes.  In the event
50       of  a partial write, the caller can make another write() call to trans‐
51       fer the remaining bytes.  The subsequent call will either transfer fur‐
52       ther bytes or may result in an error (e.g., if the disk is now full).
54       If  count is zero and fd refers to a regular file, then write() may re‐
55       turn a failure status if one of the errors below is  detected.   If  no
56       errors are detected, or error detection is not performed, 0 is returned
57       without causing any other effect.  If count is zero and fd refers to  a
58       file other than a regular file, the results are not specified.


61       EAGAIN The  file descriptor fd refers to a file other than a socket and
62              has been marked nonblocking (O_NONBLOCK), and  the  write  would
63              block.  See open(2) for further details on the O_NONBLOCK flag.
66              The  file  descriptor  fd refers to a socket and has been marked
67              nonblocking   (O_NONBLOCK),   and   the   write   would   block.
68              POSIX.1-2001  allows  either error to be returned for this case,
69              and does not require these constants to have the same value,  so
70              a portable application should check for both possibilities.
72       EBADF  fd is not a valid file descriptor or is not open for writing.
75              fd  refers to a datagram socket for which a peer address has not
76              been set using connect(2).
78       EDQUOT The user's quota of disk blocks on the filesystem containing the
79              file referred to by fd has been exhausted.
81       EFAULT buf is outside your accessible address space.
83       EFBIG  An attempt was made to write a file that exceeds the implementa‐
84              tion-defined maximum file size or the process's file size limit,
85              or to write at a position past the maximum allowed offset.
87       EINTR  The  call  was interrupted by a signal before any data was writ‐
88              ten; see signal(7).
90       EINVAL fd is attached to an object which is unsuitable for writing;  or
91              the  file  was opened with the O_DIRECT flag, and either the ad‐
92              dress specified in buf, the value specified  in  count,  or  the
93              file offset is not suitably aligned.
95       EIO    A  low-level I/O error occurred while modifying the inode.  This
96              error may relate to the write-back of data written by an earlier
97              write(), which may have been issued to a different file descrip‐
98              tor on the same file.  Since Linux 4.13, errors from  write-back
99              come  with  a  promise  that they may be reported by subsequent.
100              write() requests, and will be reported by a subsequent  fsync(2)
101              (whether  or not they were also reported by write()).  An alter‐
102              nate cause of EIO on networked filesystems is when  an  advisory
103              lock had been taken out on the file descriptor and this lock has
104              been lost.  See the Lost locks section of fcntl(2)  for  further
105              details.
107       ENOSPC The device containing the file referred to by fd has no room for
108              the data.
110       EPERM  The operation was prevented by a file seal; see fcntl(2).
112       EPIPE  fd is connected to a pipe or socket whose reading end is closed.
113              When  this  happens the writing process will also receive a SIG‐
114              PIPE signal.  (Thus, the write return value is seen only if  the
115              program catches, blocks or ignores this signal.)
117       Other errors may occur, depending on the object connected to fd.


120       POSIX.1-2008.


123       SVr4, 4.3BSD, POSIX.1-2001.
125       Under  SVr4  a  write may be interrupted and return EINTR at any point,
126       not just before any data is written.


129       A successful return from write() does not make any guarantee that  data
130       has  been  committed  to  disk.  On some filesystems, including NFS, it
131       does not even guarantee that space has successfully been  reserved  for
132       the  data.   In  this case, some errors might be delayed until a future
133       write(), fsync(2), or even close(2).  The only way to  be  sure  is  to
134       call fsync(2) after you are done writing all your data.
136       If  a  write()  is interrupted by a signal handler before any bytes are
137       written, then the call fails with the error EINTR; if it is interrupted
138       after  at  least  one byte has been written, the call succeeds, and re‐
139       turns the number of bytes written.
141       On Linux, write() (and similar system  calls)  will  transfer  at  most
142       0x7ffff000  (2,147,479,552)  bytes, returning the number of bytes actu‐
143       ally transferred.  (This is true on both 32-bit and 64-bit systems.)
145       An error return value while performing write() using  direct  I/O  does
146       not  mean the entire write has failed.  Partial data may be written and
147       the data at the file offset on which the write() was  attempted  should
148       be considered inconsistent.


151       According to POSIX.1-2008/SUSv4 Section XSI 2.9.7 ("Thread Interactions
152       with Regular File Operations"):
154           All of the following functions shall be atomic with respect to each
155           other in the effects specified in POSIX.1-2008 when they operate on
156           regular files or symbolic links: ...
158       Among the APIs subsequently listed  are  write()  and  writev(2).   And
159       among  the effects that should be atomic across threads (and processes)
160       are updates of the file offset.  However, before Linux 3.14,  this  was
161       not the case: if two processes that share an open file description (see
162       open(2)) perform a write() (or writev(2)) at the same  time,  then  the
163       I/O  operations  were not atomic with respect to updating the file off‐
164       set, with the result that the blocks of data output  by  the  two  pro‐
165       cesses  might  (incorrectly)  overlap.  This problem was fixed in Linux
166       3.14.


169       close(2), fcntl(2), fsync(2), ioctl(2), lseek(2),  open(2),  pwrite(2),
170       read(2), select(2), writev(2), fwrite(3)
174Linux man-pages 6.05              2023-04-03                          write(2)