1 .\" Copyright (c) 1983, 1990, 1991 The Regents of the University of California.
2 .\" All rights reserved.
4 .\" %%%LICENSE_START(BSD_4_CLAUSE_UCB)
5 .\" Redistribution and use in source and binary forms, with or without
6 .\" modification, are permitted provided that the following conditions
8 .\" 1. Redistributions of source code must retain the above copyright
9 .\" notice, this list of conditions and the following disclaimer.
10 .\" 2. Redistributions in binary form must reproduce the above copyright
11 .\" notice, this list of conditions and the following disclaimer in the
12 .\" documentation and/or other materials provided with the distribution.
13 .\" 3. All advertising materials mentioning features or use of this software
14 .\" must display the following acknowledgement:
15 .\" This product includes software developed by the University of
16 .\" California, Berkeley and its contributors.
17 .\" 4. Neither the name of the University nor the names of its contributors
18 .\" may be used to endorse or promote products derived from this software
19 .\" without specific prior written permission.
21 .\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 .\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25 .\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 .\" $Id: recv.2,v 1.3 1999/05/13 11:33:38 freitag Exp $
36 .\" Modified Sat Jul 24 00:22:20 1993 by Rik Faith <faith@cs.unc.edu>
37 .\" Modified Tue Oct 22 17:45:19 1996 by Eric S. Raymond <esr@thyrsus.com>
38 .\" Modified 1998,1999 by Andi Kleen
39 .\" 2001-06-19 corrected SO_EE_OFFENDER, bug report by James Hawtin
41 .TH RECV 2 2021-03-22 "Linux" "Linux Programmer's Manual"
43 recv, recvfrom, recvmsg \- receive a message from a socket
46 .B #include <sys/socket.h>
48 .BI "ssize_t recv(int " sockfd ", void *" buf ", size_t " len ", int " flags );
49 .BI "ssize_t recvfrom(int " sockfd ", void *restrict " buf ", size_t " len \
51 .BI " struct sockaddr *restrict " src_addr ,
52 .BI " socklen_t *restrict " addrlen );
53 .BI "ssize_t recvmsg(int " sockfd ", struct msghdr *" msg ", int " flags );
61 calls are used to receive messages from a socket.
63 to receive data on both connectionless and connection-oriented sockets.
64 This page first describes common features of all three system calls,
65 and then describes the differences between the calls.
67 The only difference between
77 is generally equivalent to
80 Also, the following call
82 recv(sockfd, buf, len, flags);
86 recvfrom(sockfd, buf, len, flags, NULL, NULL);
88 All three calls return the length of the message on successful
90 If a message is too long to fit in the supplied buffer, excess
91 bytes may be discarded depending on the type of socket the message is
94 If no messages are available at the socket, the receive calls wait for a
95 message to arrive, unless the socket is nonblocking (see
97 in which case the value \-1 is returned and
100 .BR EAGAIN " or " EWOULDBLOCK .
101 The receive calls normally return any data available, up to the requested
102 amount, rather than waiting for receipt of the full amount requested.
104 An application can use
109 to determine when more data arrives on a socket.
110 .SS The flags argument
113 argument is formed by ORing one or more of the following values:
115 .BR MSG_CMSG_CLOEXEC " (" recvmsg "() only; since Linux 2.6.23)"
116 Set the close-on-exec flag for the file descriptor received
117 via a UNIX domain file descriptor using the
119 operation (described in
121 This flag is useful for the same reasons as the
126 .BR MSG_DONTWAIT " (since Linux 2.2)"
127 Enables nonblocking operation; if the operation would block,
128 the call fails with the error
129 .BR EAGAIN " or " EWOULDBLOCK .
130 This provides similar behavior to setting the
135 operation), but differs in that
137 is a per-call option, whereas
139 is a setting on the open file description (see
141 which will affect all threads in the calling process
142 and as well as other processes that hold file descriptors
143 referring to the same open file description.
145 .BR MSG_ERRQUEUE " (since Linux 2.2)"
147 specifies that queued errors should be received from the socket error queue.
148 The error is passed in
149 an ancillary message with a type dependent on the protocol (for IPv4
151 The user should supply a buffer of sufficient size.
156 for more information.
157 The payload of the original packet that caused the error
158 is passed as normal data via
160 The original destination address of the datagram that caused the error
164 The error is supplied in a
170 #define SO_EE_ORIGIN_NONE 0
171 #define SO_EE_ORIGIN_LOCAL 1
172 #define SO_EE_ORIGIN_ICMP 2
173 #define SO_EE_ORIGIN_ICMP6 3
175 struct sock_extended_err
177 uint32_t ee_errno; /* Error number */
178 uint8_t ee_origin; /* Where the error originated */
179 uint8_t ee_type; /* Type */
180 uint8_t ee_code; /* Code */
181 uint8_t ee_pad; /* Padding */
182 uint32_t ee_info; /* Additional information */
183 uint32_t ee_data; /* Other data */
184 /* More data may follow */
187 struct sockaddr *SO_EE_OFFENDER(struct sock_extended_err *);
194 number of the queued error.
196 is the origin code of where the error originated.
197 The other fields are protocol-specific.
200 returns a pointer to the address of the network object
201 where the error originated from given a pointer to the ancillary message.
202 If this address is not known, the
208 and the other fields of the
211 The payload of the packet that caused the error is passed as normal data.
213 For local errors, no address is passed (this
214 can be checked with the
223 After an error has been passed, the pending socket error
224 is regenerated based on the next queued error and will be passed
225 on the next socket operation.
228 This flag requests receipt of out-of-band data that would not be received
229 in the normal data stream.
230 Some protocols place expedited data
231 at the head of the normal data queue, and thus this flag cannot
232 be used with such protocols.
235 This flag causes the receive operation to
236 return data from the beginning of the
237 receive queue without removing that data from the queue.
239 subsequent receive call will return the same data.
241 .BR MSG_TRUNC " (since Linux 2.2)"
244 Internet datagram (since Linux 2.4.27/2.6.8),
245 netlink (since Linux 2.6.22), and UNIX datagram
246 .\" commit 9f6f9af7694ede6314bed281eec74d588ba9474f
247 (since Linux 3.4) sockets:
248 return the real length of the packet or datagram,
249 even when it was longer than the passed buffer.
251 For use with Internet stream sockets, see
254 .BR MSG_WAITALL " (since Linux 2.2)"
255 This flag requests that the operation block until the full request is
257 However, the call may still return less data than requested if
258 a signal is caught, an error or disconnect occurs, or the next data to be
259 received is of a different type than that returned.
260 This flag has no effect for datagram sockets.
264 places the received message into the buffer
266 The caller must specify the size of the buffer in
272 and the underlying protocol provides the source address of the message,
273 that source address is placed in the buffer pointed to by
275 .\" (Note: for datagram sockets in both the UNIX and Internet domains,
279 .\" is also filled in for stream sockets in the UNIX domain, but is not
280 .\" filled in for stream sockets in the Internet domain.)
281 .\" [The above notes on AF_UNIX and AF_INET sockets apply as at
282 .\" Kernel 2.4.18. (MTK, 22 Jul 02)]
285 is a value-result argument.
287 it should be initialized to the size of the buffer associated with
291 is updated to contain the actual size of the source address.
292 The returned address is truncated if the buffer provided is too small;
295 will return a value greater than was supplied to the call.
297 If the caller is not interested in the source address,
301 should be specified as NULL.
306 call is normally used only on a
310 It is equivalent to the call:
312 recvfrom(fd, buf, len, flags, NULL, 0);
319 structure to minimize the number of directly supplied arguments.
320 This structure is defined as follows in
325 struct iovec { /* Scatter/gather array items */
326 void *iov_base; /* Starting address */
327 size_t iov_len; /* Number of bytes to transfer */
331 void *msg_name; /* Optional address */
332 socklen_t msg_namelen; /* Size of address */
333 struct iovec *msg_iov; /* Scatter/gather array */
334 size_t msg_iovlen; /* # elements in msg_iov */
335 void *msg_control; /* Ancillary data, see below */
336 size_t msg_controllen; /* Ancillary data buffer len */
337 int msg_flags; /* Flags on received message */
344 field points to a caller-allocated buffer that is used to
345 return the source address if the socket is unconnected.
346 The caller should set
348 to the size of this buffer before this call;
349 upon return from a successful call,
351 will contain the length of the returned address.
352 If the application does not need to know the source address,
354 can be specified as NULL.
360 describe scatter-gather locations, as discussed in
367 points to a buffer for other protocol control-related messages or
368 miscellaneous ancillary data.
373 should contain the length of the available buffer in
375 upon return from a successful call it will contain the length
376 of the control message sequence.
378 The messages are of the form:
383 size_t cmsg_len; /* Data byte count, including header
384 (type is socklen_t in POSIX) */
385 int cmsg_level; /* Originating protocol */
386 int cmsg_type; /* Protocol\-specific type */
388 unsigned char cmsg_data[]; */
393 Ancillary data should be accessed only by the macros defined in
396 As an example, Linux uses this ancillary data mechanism to pass extended
397 errors, IP options, or file descriptors over UNIX domain sockets.
398 For further information on the use of ancillary data in various
410 It can contain several flags:
413 indicates end-of-record; the data returned completed a record (generally
414 used with sockets of type
415 .BR SOCK_SEQPACKET ).
418 indicates that the trailing portion of a datagram was discarded because the
419 datagram was larger than the buffer supplied.
422 indicates that some control data was discarded due to lack of space in the
423 buffer for ancillary data.
426 is returned to indicate that expedited or out-of-band data was received.
429 indicates that no data was received but an extended error from the socket
432 These calls return the number of bytes received, or \-1
433 if an error occurred.
434 In the event of an error,
436 is set to indicate the error.
438 When a stream socket peer has performed an orderly shutdown,
439 the return value will be 0 (the traditional "end-of-file" return).
441 Datagram sockets in various domains (e.g., the UNIX and Internet domains)
442 permit zero-length datagrams.
443 When such a datagram is received, the return value is 0.
445 The value 0 may also be returned if the requested number of bytes
446 to receive from a stream socket was 0.
448 These are some standard errors generated by the socket layer.
450 may be generated and returned from the underlying protocol modules;
451 see their manual pages.
453 .BR EAGAIN " or " EWOULDBLOCK
454 .\" Actually EAGAIN on Linux
455 The socket is marked nonblocking and the receive operation
456 would block, or a receive timeout had been set and the timeout expired
457 before data was received.
458 POSIX.1 allows either error to be returned for this case,
459 and does not require these constants to have the same value,
460 so a portable application should check for both possibilities.
465 is an invalid file descriptor.
468 A remote host refused to allow the network connection (typically
469 because it is not running the requested service).
472 The receive buffer pointer(s) point outside the process's
476 The receive was interrupted by delivery of a signal before
477 any data was available; see
481 Invalid argument passed.
482 .\" e.g., msg_namelen < 0 for recvmsg() or addrlen < 0 for recvfrom()
485 Could not allocate memory for
489 The socket is associated with a connection-oriented protocol
490 and has not been connected (see
498 does not refer to a socket.
500 POSIX.1-2001, POSIX.1-2008,
501 4.4BSD (these interfaces first appeared in 4.2BSD).
503 POSIX.1 describes only the
510 If a zero-length datagram is pending,
516 argument of zero provide different behavior.
517 In this circumstance,
519 has no effect (the datagram remains pending), while
521 consumes the pending datagram.
525 type was invented by POSIX.
529 According to POSIX.1,
530 .\" POSIX.1-2001, POSIX.1-2008
535 structure should be typed as
539 field should be typed as
541 but glibc currently types both as
543 .\" glibc bug for msg_controllen raised 12 Mar 2006
544 .\" http://sourceware.org/bugzilla/show_bug.cgi?id=2448
545 .\" The problem is an underlying kernel issue: the size of the
546 .\" __kernel_size_t type used to type these fields varies
547 .\" across architectures, but socklen_t is always 32 bits,
548 .\" as (at least with GCC) is int.
552 for information about a Linux-specific system call
553 that can be used to receive multiple datagrams in a single call.
555 An example of the use of