1 .\" Copyright (c) 1983, 1991 The Regents of the University of California.
2 .\" All rights reserved.
4 .\" Redistribution and use in source and binary forms, with or without
5 .\" modification, are permitted provided that the following conditions
7 .\" 1. Redistributions of source code must retain the above copyright
8 .\" notice, this list of conditions and the following disclaimer.
9 .\" 2. Redistributions in binary form must reproduce the above copyright
10 .\" notice, this list of conditions and the following disclaimer in the
11 .\" documentation and/or other materials provided with the distribution.
12 .\" 3. All advertising materials mentioning features or use of this software
13 .\" must display the following acknowledgement:
14 .\" This product includes software developed by the University of
15 .\" California, Berkeley and its contributors.
16 .\" 4. Neither the name of the University nor the names of its contributors
17 .\" may be used to endorse or promote products derived from this software
18 .\" without specific prior written permission.
20 .\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 .\" Modified 1993-07-24 by Rik Faith <faith@cs.unc.edu>
33 .\" Modified 1996-10-22 by Eric S. Raymond <esr@thyrsus.com>
34 .\" Modified Oct 1998 by Andi Kleen
35 .\" Modified Oct 2003 by aeb
36 .\" Modified 2004-07-01 by mtk
38 .TH SEND 2 2012-02-27 "Linux" "Linux Programmer's Manual"
40 send, sendto, sendmsg \- send a message on a socket
43 .B #include <sys/types.h>
44 .B #include <sys/socket.h>
46 .BI "ssize_t send(int " sockfd ", const void *" buf ", size_t " len \
49 .BI "ssize_t sendto(int " sockfd ", const void *" buf ", size_t " len \
51 .BI " const struct sockaddr *" dest_addr ", socklen_t " addrlen );
53 .BI "ssize_t sendmsg(int " sockfd ", const struct msghdr *" msg \
62 are used to transmit a message to another socket.
66 call may be used only when the socket is in a
68 state (so that the intended recipient is known).
69 The only difference between
81 Also, the following call
83 send(sockfd, buf, len, flags);
87 sendto(sockfd, buf, len, flags, NULL, 0);
91 is the file descriptor of the sending socket.
95 is used on a connection-mode
102 are ignored (and the error
104 may be returned when they are
105 not NULL and 0), and the error
107 is returned when the socket was not actually connected.
108 Otherwise, the address of the target is given by
115 the address of the target is given by
125 the message is found in
131 the message is pointed to by the elements of the array
135 call also allows sending ancillary data (also known as control information).
137 If the message is too long to pass atomically through the
138 underlying protocol, the error
140 is returned, and the message is not transmitted.
142 No indication of failure to deliver is implicit in a
144 Locally detected errors are indicated by a return value of \-1.
146 When the message does not fit into the send buffer of the socket,
148 normally blocks, unless the socket has been placed in nonblocking I/O
150 In nonblocking mode it would fail with the error
157 call may be used to determine when it is possible to send more data.
161 argument is the bitwise OR
162 of zero or more of the following flags.
163 .\" FIXME ? document MSG_PROXY (which went away in 2.3.15)
165 .BR MSG_CONFIRM " (Since Linux 2.3.15)"
166 Tell the link layer that forward progress happened: you got a successful
167 reply from the other side.
168 If the link layer doesn't get this
169 it will regularly reprobe the neighbor (e.g., via a unicast ARP).
174 sockets and currently only implemented for IPv4 and IPv6.
180 Don't use a gateway to send out the packet, only send to hosts on
181 directly connected networks.
182 This is usually used only
183 by diagnostic or routing programs.
184 This is only defined for protocol
185 families that route; packet sockets don't.
187 .BR MSG_DONTWAIT " (since Linux 2.2)"
188 Enables nonblocking operation; if the operation would block,
192 is returned (this can also be enabled using the
198 .BR MSG_EOR " (since Linux 2.2)"
199 Terminates a record (when this notion is supported, as for sockets of type
200 .BR SOCK_SEQPACKET ).
202 .BR MSG_MORE " (Since Linux 2.4.4)"
203 The caller has more data to send.
204 This flag is used with TCP sockets to obtain the same effect
209 with the difference that this flag can be set on a per-call basis.
211 Since Linux 2.6, this flag is also supported for UDP sockets, and informs
212 the kernel to package all of the data sent in calls with this flag set
213 into a single datagram which is only transmitted when a call is performed
214 that does not specify this flag.
217 socket option described in
220 .BR MSG_NOSIGNAL " (since Linux 2.2)"
223 on errors on stream oriented sockets when the other end breaks the
227 error is still returned.
232 data on sockets that support this notion (e.g., of type
234 the underlying protocol must also support
238 The definition of the
243 and below for an exact description of its fields.
248 void *msg_name; /* optional address */
249 socklen_t msg_namelen; /* size of address */
250 struct iovec *msg_iov; /* scatter/gather array */
251 size_t msg_iovlen; /* # elements in msg_iov */
252 void *msg_control; /* ancillary data, see below */
253 size_t msg_controllen; /* ancillary data buffer len */
254 int msg_flags; /* flags on received message */
259 You may send control information using the
264 The maximum control buffer length the kernel can process is limited
265 per socket by the value in
266 .IR /proc/sys/net/core/optmem_max ;
269 .\" Still to be documented:
270 .\" Send file descriptors and user credentials using the
271 .\" msg_control* fields.
272 .\" The flags returned in msg_flags.
274 On success, these calls return the number of characters sent.
275 On error, \-1 is returned, and
277 is set appropriately.
279 These are some standard errors generated by the socket layer.
281 may be generated and returned from the underlying protocol modules;
282 see their respective manual pages.
285 (For UNIX domain sockets, which are identified by pathname)
286 Write permission is denied on the destination socket file,
287 or search permission is denied for one of the directories
290 .BR path_resolution (7).)
292 .BR EAGAIN " or " EWOULDBLOCK
293 .\" Actually EAGAIN on Linux
294 The socket is marked nonblocking and the requested operation
296 POSIX.1-2001 allows either error to be returned for this case,
297 and does not require these constants to have the same value,
298 so a portable application should check for both possibilities.
301 An invalid descriptor was specified.
304 Connection reset by peer.
307 The socket is not connection-mode, and no peer address is set.
310 An invalid user space address was specified for an argument.
313 A signal occurred before any data was transmitted; see
317 Invalid argument passed.
320 The connection-mode socket was connected already but a
321 recipient was specified.
322 (Now either this error is returned, or the recipient specification
327 .\" (e.g., SOCK_DGRAM )
328 requires that message be sent atomically, and the size
329 of the message to be sent made this impossible.
332 The output queue for a network interface was full.
333 This generally indicates that the interface has stopped sending,
334 but may be caused by transient congestion.
335 (Normally, this does not occur in Linux.
336 Packets are just silently dropped
337 when a device queue overflows.)
343 The socket is not connected, and no target has been given.
353 argument is inappropriate for the socket type.
356 The local end has been shut down on a connection oriented socket.
357 In this case the process
364 4.4BSD, SVr4, POSIX.1-2001.
365 These function calls appeared in 4.2BSD.
367 POSIX.1-2001 only describes the
372 POSIX.1-2008 adds a specification of
376 flag is a Linux extension.
378 The prototypes given above follow the Single UNIX Specification,
379 as glibc2 also does; the
381 argument was \fIint\fP in 4.x BSD, but \fIunsigned int\fP in libc4 and libc5;
384 argument was \fIint\fP in 4.x BSD and libc4, but \fIsize_t\fP in libc5;
387 argument was \fIint\fP in 4.x BSD and libc4 and libc5.
391 According to POSIX.1-2001, the
395 structure should be typed as
397 but glibc currently types it as
399 .\" glibc bug raised 12 Mar 2006
400 .\" http://sourceware.org/bugzilla/show_bug.cgi?id=2448
401 .\" The problem is an underlying kernel issue: the size of the
402 .\" __kernel_size_t type used to type this field varies
403 .\" across architectures, but socklen_t is always 32 bits.
407 for information about a Linux-specific system call
408 that can be used to transmit multiple datagrams in a single call.
415 An example of the use of