1SOCKET(2) Linux Programmer's Manual SOCKET(2)
2
6 socket - create an endpoint for communication
7
9 #include <sys/types.h> /* See NOTES */
10 #include <sys/socket.h>
11 int socket(int domain, int type, int protocol);
13
15 socket() creates an endpoint for communication and returns a file
16 descriptor that refers to that endpoint. The file descriptor returned
17 by a successful call will be the lowest-numbered file descriptor not
18 currently open for the process.
19 The domain argument specifies a communication domain; this selects the
21 protocol family which will be used for communication. These families
22 are defined in <sys/socket.h>. The formats currently understood by the
23 Linux kernel include:
24 Name Purpose Man page
26 AF_UNIX Local communication unix(7)
27 AF_LOCAL Synonym for AF_UNIX
28 AF_INET IPv4 Internet protocols ip(7)
29 AF_AX25 Amateur radio AX.25 protocol ax25(4)
30 AF_IPX IPX - Novell protocols
31 AF_APPLETALK AppleTalk ddp(7)
32 AF_X25 ITU-T X.25 / ISO-8208 protocol x25(7)
33 AF_INET6 IPv6 Internet protocols ipv6(7)
34 AF_DECnet DECet protocol sockets
35 AF_KEY Key management protocol, originally
36 developed for usage with IPsec
37 AF_NETLINK Kernel user interface device netlink(7)
38 AF_PACKET Low-level packet interface packet(7)
39 AF_RDS Reliable Datagram Sockets (RDS) protocol rds(7)
40 rds-rdma(7)
41 AF_PPPOX Generic PPP transport layer, for setting
42 up L2 tunnels (L2TP and PPPoE)
43 AF_LLC Logical link control (IEEE 802.2 LLC)
44 protocol
45 AF_IB InfiniBand native addressing
46 AF_MPLS Multiprotocol Label Switching
47 AF_CAN Controller Area Network automotive bus
48 protocol
49 AF_TIPC TIPC, "cluster domain sockets" protocol
50 AF_BLUETOOTH Bluetooth low-level socket protocol
51 AF_ALG Interface to kernel crypto API
52 AF_VSOCK VSOCK (originally "VMWare VSockets") vsock(7)
53 protocol for hypervisor-guest communica‐
54 tion
55 AF_KCM KCM (kernel connection multiplexor)
56 interface
57 AF_XDP XDP (express data path) interface
58 Further details of the above address families, as well as information
60 on several other address families, can be found in address_families(7).
61 The socket has the indicated type, which specifies the communication
63 semantics. Currently defined types are:
64 SOCK_STREAM Provides sequenced, reliable, two-way, connection-based
66 byte streams. An out-of-band data transmission mecha‐
67 nism may be supported.
68 SOCK_DGRAM Supports datagrams (connectionless, unreliable messages
70 of a fixed maximum length).
71 SOCK_SEQPACKET Provides a sequenced, reliable, two-way connection-
73 based data transmission path for datagrams of fixed
74 maximum length; a consumer is required to read an
75 entire packet with each input system call.
76 SOCK_RAW Provides raw network protocol access.
78 SOCK_RDM Provides a reliable datagram layer that does not guar‐
80 antee ordering.
81 SOCK_PACKET Obsolete and should not be used in new programs; see
83 packet(7).
84 Some socket types may not be implemented by all protocol families.
86 Since Linux 2.6.27, the type argument serves a second purpose: in addi‐
88 tion to specifying a socket type, it may include the bitwise OR of any
89 of the following values, to modify the behavior of socket():
90 SOCK_NONBLOCK Set the O_NONBLOCK file status flag on the open file
92 description (see open(2)) referred to by the new file
93 descriptor. Using this flag saves extra calls to
94 fcntl(2) to achieve the same result.
95 SOCK_CLOEXEC Set the close-on-exec (FD_CLOEXEC) flag on the new file
97 descriptor. See the description of the O_CLOEXEC flag
98 in open(2) for reasons why this may be useful.
99 The protocol specifies a particular protocol to be used with the
101 socket. Normally only a single protocol exists to support a particular
102 socket type within a given protocol family, in which case protocol can
103 be specified as 0. However, it is possible that many protocols may
104 exist, in which case a particular protocol must be specified in this
105 manner. The protocol number to use is specific to the “communication
106 domain” in which communication is to take place; see protocols(5). See
107 getprotoent(3) on how to map protocol name strings to protocol numbers.
108 Sockets of type SOCK_STREAM are full-duplex byte streams. They do not
110 preserve record boundaries. A stream socket must be in a connected
111 state before any data may be sent or received on it. A connection to
112 another socket is created with a connect(2) call. Once connected, data
113 may be transferred using read(2) and write(2) calls or some variant of
114 the send(2) and recv(2) calls. When a session has been completed a
115 close(2) may be performed. Out-of-band data may also be transmitted as
116 described in send(2) and received as described in recv(2).
117 The communications protocols which implement a SOCK_STREAM ensure that
119 data is not lost or duplicated. If a piece of data for which the peer
120 protocol has buffer space cannot be successfully transmitted within a
121 reasonable length of time, then the connection is considered to be
122 dead. When SO_KEEPALIVE is enabled on the socket the protocol checks
123 in a protocol-specific manner if the other end is still alive. A SIG‐
124 PIPE signal is raised if a process sends or receives on a broken
125 stream; this causes naive processes, which do not handle the signal, to
126 exit. SOCK_SEQPACKET sockets employ the same system calls as
127 SOCK_STREAM sockets. The only difference is that read(2) calls will
128 return only the amount of data requested, and any data remaining in the
129 arriving packet will be discarded. Also all message boundaries in
130 incoming datagrams are preserved.
131 SOCK_DGRAM and SOCK_RAW sockets allow sending of datagrams to corre‐
133 spondents named in sendto(2) calls. Datagrams are generally received
134 with recvfrom(2), which returns the next datagram along with the
135 address of its sender.
136 SOCK_PACKET is an obsolete socket type to receive raw packets directly
138 from the device driver. Use packet(7) instead.
139 An fcntl(2) F_SETOWN operation can be used to specify a process or
141 process group to receive a SIGURG signal when the out-of-band data
142 arrives or SIGPIPE signal when a SOCK_STREAM connection breaks unex‐
143 pectedly. This operation may also be used to set the process or
144 process group that receives the I/O and asynchronous notification of
145 I/O events via SIGIO. Using F_SETOWN is equivalent to an ioctl(2) call
146 with the FIOSETOWN or SIOCSPGRP argument.
147 When the network signals an error condition to the protocol module
149 (e.g., using an ICMP message for IP) the pending error flag is set for
150 the socket. The next operation on this socket will return the error
151 code of the pending error. For some protocols it is possible to enable
152 a per-socket error queue to retrieve detailed information about the
153 error; see IP_RECVERR in ip(7).
154 The operation of sockets is controlled by socket level options. These
156 options are defined in <sys/socket.h>. The functions setsockopt(2) and
157 getsockopt(2) are used to set and get options.
158
160 On success, a file descriptor for the new socket is returned. On
161 error, -1 is returned, and errno is set appropriately.
162
164 EACCES Permission to create a socket of the specified type and/or pro‐
165 tocol is denied.
166 EAFNOSUPPORT
168 The implementation does not support the specified address fam‐
169 ily.
170 EINVAL Unknown protocol, or protocol family not available.
172 EINVAL Invalid flags in type.
174 EMFILE The per-process limit on the number of open file descriptors has
176 been reached.
177 ENFILE The system-wide limit on the total number of open files has been
179 reached.
180 ENOBUFS or ENOMEM
182 Insufficient memory is available. The socket cannot be created
183 until sufficient resources are freed.
184 EPROTONOSUPPORT
186 The protocol type or the specified protocol is not supported
187 within this domain.
188 Other errors may be generated by the underlying protocol modules.
190
192 POSIX.1-2001, POSIX.1-2008, 4.4BSD.
193 The SOCK_NONBLOCK and SOCK_CLOEXEC flags are Linux-specific.
195 socket() appeared in 4.2BSD. It is generally portable to/from non-BSD
197 systems supporting clones of the BSD socket layer (including System V
198 variants).
199
201 POSIX.1 does not require the inclusion of <sys/types.h>, and this
202 header file is not required on Linux. However, some historical (BSD)
203 implementations required this header file, and portable applications
204 are probably wise to include it.
205 The manifest constants used under 4.x BSD for protocol families are
207 PF_UNIX, PF_INET, and so on, while AF_UNIX, AF_INET, and so on are used
208 for address families. However, already the BSD man page promises: "The
209 protocol family generally is the same as the address family", and sub‐
210 sequent standards use AF_* everywhere.
211
213 An example of the use of socket() is shown in getaddrinfo(3).
214
216 accept(2), bind(2), close(2), connect(2), fcntl(2), getpeername(2),
217 getsockname(2), getsockopt(2), ioctl(2), listen(2), read(2), recv(2),
218 select(2), send(2), shutdown(2), socketpair(2), write(2), getpro‐
219 toent(3), address_families(7), ip(7), socket(7), tcp(7), udp(7),
220 unix(7)
221 “An Introductory 4.3BSD Interprocess Communication Tutorial” and “BSD
223 Interprocess Communication Tutorial”, reprinted in UNIX Programmer's
224 Supplementary Documents Volume 1.
225
227 This page is part of release 5.07 of the Linux man-pages project. A
228 description of the project, information about reporting bugs, and the
229 latest version of this page, can be found at
230 https://www.kernel.org/doc/man-pages/.
231Linux 2020-06-09 SOCKET(2)