NAME

socket − create an endpoint for communication

SYNTAX

#include <sys/types.h>
#include <sys/socket.h>

int s = socket(af, type, protocol)
int af, type, protocol;

DESCRIPTION

The socket system call creates an endpoint for communication and returns a descriptor. 

The af specifies an address format with which addresses specified in later operations using the socket should be interpreted.  These formats are defined in the include file <sys/socket.h>.  The currently understood formats are:

AF_UNIX      (UNIX path names)
AF_INET      (ARPA Internet addresses)
AF_PUP      (Xerox PUP-I Internet addresses)
AF_IMPLINK      (IMP “host at IMP” addresses)

The socket has the indicated type which specifies the semantics of communication.  Currently, defined types are:

SOCK_STREAM
Provides sequenced, reliable, two-way connection based byte streams with an out-of-band data transmission mechanism.

SOCK_DGRAM
Supports datagrams which have connectionless, unreliable messages of a fixed (typically small) maximum length.

SOCK_RAW
Provides access to internal network interfaces. This type is available only to the superuser.

SOCK_SEQPACKET and SOCK_RDM
Currently are not implemented and, therefore, are not described here.

The specified protocol names the protocol to be used with the socket.  Normally, only a single protocol exists to support a particular socket type using a given address format.  However, it is possible that many protocols may exist in which case a particular protocol must be specified in this manner.  The protocol number to use is particular to the “communication domain” in which communication is to take place.  For further information, see services(5) and protocols(5). 

Similar to a pipe, a SOCK_STREAM is a full-duplex byte stream.  A stream socket must be in a connected state before any data can be sent or received on it.  A connection to another socket is created with a connect call.  Once connected, data can be transferred using either read and write calls or some variant of the send and recv calls.  When a session has been completed, a close call can be performed.  The “out-of-band” data can also be transmitted, as described in send(2), and received, as described in recv(2). 

The communications protocols used to implement a SOCK_STREAM insure that data is not lost or duplicated.  If a piece of data which the peer protocol has buffer space for cannot be successfully transmitted within a reasonable length of time, then the connection is considered broken, and calls will indicate an error with −1 returns and with ETIMEDOUT as the specific code in the global variable errno.  The protocols optionally keep sockets “warm” by forcing transmissions roughly every minute in the absence of other activity.  An error is then indicated if no response can be elicited on an otherwise idle connection for a extended period (for example, 5 minutes).  A SIGPIPE signal is raised if a process sends on a broken stream.  This causes naive processes, which do not handle the signal, to exit. 

Both SOCK_DGRAM and SOCK_RAW sockets allow sending of datagrams to correspondents named in send calls.  It is also possible to receive datagrams at such a socket with a recv call. 

A fcntl call can be used to specify a process group to receive a SIGURG signal when the out-of-band data arrives. 

The operation of sockets is controlled by socket level options.  These options are defined in the file <sys/socket.h> and explained below.  The setsockopt and getsockopt calls are used to set and get options, respectively. 

SO_DEBUG
Turns on recording of debugging information.

SO_REUSEADDR
Allows local address reuse.

SO_KEEPALIVE
Keeps connections alive.

SO_DONTROUTE
Does no apply routing on outgoing messages.

SO_LINGER
Lingers on closel, if data present.

SO_DONTLINGER
Does not linger on close.

•SO_DEBUG enables debugging in the underlying protocol modules. 

•SO_REUSEADDR indicates the rules used in validating addresses supplied in a bind call should allow reuse of local addresses. 

•SO_KEEPALIVE enables the periodic transmission of messages on a connected socket.  Should the connected party fail to respond to these messages, the connection is considered broken, and processes using the socket are notified via a SIGPIPE signal. 

•SO_DONTROUTE indicates that outgoing messages should bypass the standard routing facilities.  Instead, messages are directed to the appropriate network interface according to the network portion of the destination address. 

•SO_LINGER and SO_DONTLINGER control the actions taken when unsent messages are queued on socket and when a close call is performed.  If the socket promises reliable delivery of data and if SO_LINGER is set, the system will block the process on the close attempt either until it is able to transmit the data or until it decides it is unable to deliver the information (a timeout period, termed the linger interval, is specified in the setsockopt call when SO_LINGER is requested).  If SO_DONTLINGER is specified and if a close is issued, the system will process the close in a manner which allows the process to continue as quickly as possible. 

RETURN VALUE

If successful, returns a descriptor that references the socket.  If unsuccessful, returns a −1, and the global variable errno indicates the error code. 

DIAGNOSTICS

The socket call will fails if:

[EAFNOSUPPORT] The specified address family is not supported in this version of the system. 

[ESOCKTNOSUPPORT] The specified socket type is not supported in this address family. 

[EPROTONOSUPPORT] The specified protocol is not supported. 

[EMFILE] The per-process descriptor table is full. 

[ENOBUFS] No buffer space is available.  The socket cannot be created. 

SEE ALSO

accept(2), bind(2), connect(2), getsockname(2), getsockopt(2), ioctl(2), listen(2), recv(2), select(2), send(2), shutdown(2), socketpair(2)
“A 4.2BSD Interprocess Communication Primer”