Update.

[glibc.git] / manual / socket.texi

@node Sockets, Low-Level Terminal Interface, Pipes and FIFOs, Top
@chapter Sockets

This chapter describes the GNU facilities for interprocess
communication using sockets.

@cindex socket
@cindex interprocess communication, with sockets
A @dfn{socket} is a generalized interprocess communication channel.
Like a pipe, a socket is represented as a file descriptor.  But,
unlike pipes, sockets support communication between unrelated
processes, and even between processes running on different machines
that communicate over a network.  Sockets are the primary means of
communicating with other machines; @code{telnet}, @code{rlogin},
@code{ftp}, @code{talk}, and the other familiar network programs use
sockets.

Not all operating systems support sockets.  In the GNU library, the
header file @file{sys/socket.h} exists regardless of the operating
system, and the socket functions always exist, but if the system does
not really support sockets, these functions always fail.

@strong{Incomplete:} We do not currently document the facilities for
broadcast messages or for configuring Internet interfaces.

@menu
* Socket Concepts::     Basic concepts you need to know about.
* Communication Styles::Stream communication, datagrams, and other styles.
* Socket Addresses::    How socket names (``addresses'') work.
* File Namespace::      Details about the file namespace.
* Internet Namespace::  Details about the Internet namespace.
* Misc Namespaces::     Other namespaces not documented fully here.
* Open/Close Sockets::  Creating sockets and destroying them.
* Connections::         Operations on sockets with connection state.
* Datagrams::           Operations on datagram sockets.
* Inetd::               Inetd is a daemon that starts servers on request.
                           The most convenient way to write a server
                           is to make it work with Inetd.
* Socket Options::      Miscellaneous low-level socket options.
* Networks Database::   Accessing the database of network names.
@end menu

@node Socket Concepts
@section Socket Concepts

@cindex communication style (of a socket)
@cindex style of communication (of a socket)
When you create a socket, you must specify the style of communication
you want to use and the type of protocol that should implement it.
The @dfn{communication style} of a socket defines the user-level
semantics of sending and receiving data on the socket.  Choosing a
communication style specifies the answers to questions such as these:

@itemize @bullet
@item
@cindex packet
@cindex byte stream
@cindex stream (sockets)
@strong{What are the units of data transmission?}  Some communication
styles regard the data as a sequence of bytes, with no larger
structure; others group the bytes into records (which are known in
this context as @dfn{packets}).

@item
@cindex loss of data on sockets
@cindex data loss on sockets
@strong{Can data be lost during normal operation?}  Some communication
styles guarantee that all the data sent arrives in the order it was
sent (barring system or network crashes); other styles occasionally
lose data as a normal part of operation, and may sometimes deliver
packets more than once or in the wrong order.

Designing a program to use unreliable communication styles usually
involves taking precautions to detect lost or misordered packets and
to retransmit data as needed.

@item
@strong{Is communication entirely with one partner?}  Some
communication styles are like a telephone call---you make a
@dfn{connection} with one remote socket, and then exchange data
freely.  Other styles are like mailing letters---you specify a
destination address for each message you send.
@end itemize

@cindex namespace (of socket)
@cindex domain (of socket)
@cindex socket namespace
@cindex socket domain
You must also choose a @dfn{namespace} for naming the socket.  A socket
name (``address'') is meaningful only in the context of a particular
namespace.  In fact, even the data type to use for a socket name may
depend on the namespace.  Namespaces are also called ``domains'', but we
avoid that word as it can be confused with other usage of the same
term.  Each namespace has a symbolic name that starts with @samp{PF_}.
A corresponding symbolic name starting with @samp{AF_} designates the
address format for that namespace.

@cindex network protocol
@cindex protocol (of socket)
@cindex socket protocol
@cindex protocol family
Finally you must choose the @dfn{protocol} to carry out the
communication.  The protocol determines what low-level mechanism is used
to transmit and receive data.  Each protocol is valid for a particular
namespace and communication style; a namespace is sometimes called a
@dfn{protocol family} because of this, which is why the namespace names
start with @samp{PF_}.

The rules of a protocol apply to the data passing between two programs,
perhaps on different computers; most of these rules are handled by the
operating system, and you need not know about them.  What you do need to
know about protocols is this:

@itemize @bullet
@item
In order to have communication between two sockets, they must specify
the @emph{same} protocol.

@item
Each protocol is meaningful with particular style/namespace
combinations and cannot be used with inappropriate combinations.  For
example, the TCP protocol fits only the byte stream style of
communication and the Internet namespace.

@item
For each combination of style and namespace, there is a @dfn{default
protocol} which you can request by specifying 0 as the protocol
number.  And that's what you should normally do---use the default.
@end itemize

Throughout the following description at various places
variables/parameters to denote sizes are required.  And here the trouble
starts.  In the first implementations the type of these variables was
simply @code{int}.  This type was on almost all machines of this time 32
bits wide and so a de-factor standard required 32 bit variables.  This
is important since references to variables of this type are passed to
the kernel.

But now the POSIX people came and unified the interface with their words
"all size values are of type @code{size_t}".  But on 64 bit machines
@code{size_t} is 64 bits wide and so variable references are not anymore
possible.

A solution provides the Unix98 specification which finally introduces a
type @code{socklen_t}.  This type is used in all of the cases that were
previously changed to use @code{size_t}.  The only requirement of this
type is that it is an unsigned type of at least 32 bits.  Therefore,
implementations which require references to 32 bit variables be passed
can be as happy as implementations which use right from the start 64 bit
values.


@node Communication Styles
@section Communication Styles

The GNU library includes support for several different kinds of sockets,
each with different characteristics.  This section describes the
supported socket types.  The symbolic constants listed here are
defined in @file{sys/socket.h}.
@pindex sys/socket.h

@comment sys/socket.h
@comment BSD
@deftypevr Macro int SOCK_STREAM
The @code{SOCK_STREAM} style is like a pipe (@pxref{Pipes and FIFOs});
it operates over a connection with a particular remote socket, and
transmits data reliably as a stream of bytes.

Use of this style is covered in detail in @ref{Connections}.
@end deftypevr

@comment sys/socket.h
@comment BSD
@deftypevr Macro int SOCK_DGRAM
The @code{SOCK_DGRAM} style is used for sending
individually-addressed packets, unreliably.
It is the diametrical opposite of @code{SOCK_STREAM}.

Each time you write data to a socket of this kind, that data becomes
one packet.  Since @code{SOCK_DGRAM} sockets do not have connections,
you must specify the recipient address with each packet.

The only guarantee that the system makes about your requests to
transmit data is that it will try its best to deliver each packet you
send.  It may succeed with the sixth packet after failing with the
fourth and fifth packets; the seventh packet may arrive before the
sixth, and may arrive a second time after the sixth.

The typical use for @code{SOCK_DGRAM} is in situations where it is
acceptable to simply resend a packet if no response is seen in a
reasonable amount of time.

@xref{Datagrams}, for detailed information about how to use datagram
sockets.
@end deftypevr

@ignore
@c This appears to be only for the NS domain, which we aren't
@c discussing and probably won't support either.
@comment sys/socket.h
@comment BSD
@deftypevr Macro int SOCK_SEQPACKET
This style is like @code{SOCK_STREAM} except that the data is
structured into packets.

A program that receives data over a @code{SOCK_SEQPACKET} socket
should be prepared to read the entire message packet in a single call
to @code{read}; if it only reads part of the message, the remainder of
the message is simply discarded instead of being available for
subsequent calls to @code{read}.

Many protocols do not support this communication style.
@end deftypevr
@end ignore

@ignore
@comment sys/socket.h
@comment BSD
@deftypevr Macro int SOCK_RDM
This style is a reliable version of @code{SOCK_DGRAM}: it sends
individually addressed packets, but guarantees that each packet sent
arrives exactly once.

@strong{Warning:} It is not clear this is actually supported
by any operating system.
@end deftypevr
@end ignore

@comment sys/socket.h
@comment BSD
@deftypevr Macro int SOCK_RAW
This style provides access to low-level network protocols and
interfaces.  Ordinary user programs usually have no need to use this
style.
@end deftypevr

@node Socket Addresses
@section Socket Addresses

@cindex address of socket
@cindex name of socket
@cindex binding a socket address
@cindex socket address (name) binding
The name of a socket is normally called an @dfn{address}.  The
functions and symbols for dealing with socket addresses were named
inconsistently, sometimes using the term ``name'' and sometimes using
``address''.  You can regard these terms as synonymous where sockets
are concerned.

A socket newly created with the @code{socket} function has no
address.  Other processes can find it for communication only if you
give it an address.  We call this @dfn{binding} the address to the
socket, and the way to do it is with the @code{bind} function.

You need be concerned with the address of a socket if other processes
are to find it and start communicating with it.  You can specify an
address for other sockets, but this is usually pointless; the first time
you send data from a socket, or use it to initiate a connection, the
system assigns an address automatically if you have not specified one.

Occasionally a client needs to specify an address because the server
discriminates based on addresses; for example, the rsh and rlogin
protocols look at the client's socket address and don't bypass password
checking unless it is less than @code{IPPORT_RESERVED} (@pxref{Ports}).

The details of socket addresses vary depending on what namespace you are
using.  @xref{File Namespace}, or @ref{Internet Namespace}, for specific
information.

Regardless of the namespace, you use the same functions @code{bind} and
@code{getsockname} to set and examine a socket's address.  These
functions use a phony data type, @code{struct sockaddr *}, to accept the
address.  In practice, the address lives in a structure of some other
data type appropriate to the address format you are using, but you cast
its address to @code{struct sockaddr *} when you pass it to
@code{bind}.

@menu
* Address Formats::             About @code{struct sockaddr}.
* Setting Address::             Binding an address to a socket.
* Reading Address::             Reading the address of a socket.
@end menu

@node Address Formats
@subsection Address Formats

The functions @code{bind} and @code{getsockname} use the generic data
type @code{struct sockaddr *} to represent a pointer to a socket
address.  You can't use this data type effectively to interpret an
address or construct one; for that, you must use the proper data type
for the socket's namespace.

Thus, the usual practice is to construct an address in the proper
namespace-specific type, then cast a pointer to @code{struct sockaddr *}
when you call @code{bind} or @code{getsockname}.

The one piece of information that you can get from the @code{struct
sockaddr} data type is the @dfn{address format} designator which tells
you which data type to use to understand the address fully.

@pindex sys/socket.h
The symbols in this section are defined in the header file
@file{sys/socket.h}.

@comment sys/socket.h
@comment BSD
@deftp {Date Type} {struct sockaddr}
The @code{struct sockaddr} type itself has the following members:

@table @code
@item short int sa_family
This is the code for the address format of this address.  It
identifies the format of the data which follows.

@item char sa_data[14]
This is the actual socket address data, which is format-dependent.  Its
length also depends on the format, and may well be more than 14.  The
length 14 of @code{sa_data} is essentially arbitrary.
@end table
@end deftp

Each address format has a symbolic name which starts with @samp{AF_}.
Each of them corresponds to a @samp{PF_} symbol which designates the
corresponding namespace.  Here is a list of address format names:

@table @code
@comment sys/socket.h
@comment GNU
@item AF_FILE
@vindex AF_FILE
This designates the address format that goes with the file namespace.
(@code{PF_FILE} is the name of that namespace.)  @xref{File Namespace
Details}, for information about this address format.

@comment sys/socket.h
@comment BSD
@item AF_UNIX
@vindex AF_UNIX
This is a synonym for @code{AF_FILE}, for compatibility.
(@code{PF_UNIX} is likewise a synonym for @code{PF_FILE}.)

@comment sys/socket.h
@comment BSD
@item AF_INET
@vindex AF_INET
This designates the address format that goes with the Internet
namespace.  (@code{PF_INET} is the name of that namespace.)
@xref{Internet Address Formats}.

@comment sys/socket.h
@comment IPv6 Basic API
@item AF_INET6
This is similar to @code{AF_INET}, but refers to the IPv6 protocol.
(@code{PF_INET6} is the name of the corresponding namespace.)

@comment sys/socket.h
@comment BSD
@item AF_UNSPEC
@vindex AF_UNSPEC
This designates no particular address format.  It is used only in rare
cases, such as to clear out the default destination address of a
``connected'' datagram socket.  @xref{Sending Datagrams}.

The corresponding namespace designator symbol @code{PF_UNSPEC} exists
for completeness, but there is no reason to use it in a program.
@end table

@file{sys/socket.h} defines symbols starting with @samp{AF_} for many
different kinds of networks, all or most of which are not actually
implemented.  We will document those that really work, as we receive
information about how to use them.

@node Setting Address
@subsection Setting the Address of a Socket

@pindex sys/socket.h
Use the @code{bind} function to assign an address to a socket.  The
prototype for @code{bind} is in the header file @file{sys/socket.h}.
For examples of use, see @ref{File Namespace}, or see @ref{Inet Example}.

@comment sys/socket.h
@comment BSD
@deftypefun int bind (int @var{socket}, struct sockaddr *@var{addr}, socklen_t @var{length})
The @code{bind} function assigns an address to the socket
@var{socket}.  The @var{addr} and @var{length} arguments specify the
address; the detailed format of the address depends on the namespace.
The first part of the address is always the format designator, which
specifies a namespace, and says that the address is in the format for
that namespace.

The return value is @code{0} on success and @code{-1} on failure.  The
following @code{errno} error conditions are defined for this function:

@table @code
@item EBADF
The @var{socket} argument is not a valid file descriptor.

@item ENOTSOCK
The descriptor @var{socket} is not a socket.

@item EADDRNOTAVAIL
The specified address is not available on this machine.

@item EADDRINUSE
Some other socket is already using the specified address.

@item EINVAL
The socket @var{socket} already has an address.

@item EACCES
You do not have permission to access the requested address.  (In the
Internet domain, only the super-user is allowed to specify a port number
in the range 0 through @code{IPPORT_RESERVED} minus one; see
@ref{Ports}.)
@end table

Additional conditions may be possible depending on the particular namespace
of the socket.
@end deftypefun

@node Reading Address
@subsection Reading the Address of a Socket

@pindex sys/socket.h
Use the function @code{getsockname} to examine the address of an
Internet socket.  The prototype for this function is in the header file
@file{sys/socket.h}.

@comment sys/socket.h
@comment BSD
@deftypefun int getsockname (int @var{socket}, struct sockaddr *@var{addr}, socklen_t *@var{length-ptr})
The @code{getsockname} function returns information about the
address of the socket @var{socket} in the locations specified by the
@var{addr} and @var{length-ptr} arguments.  Note that the
@var{length-ptr} is a pointer; you should initialize it to be the
allocation size of @var{addr}, and on return it contains the actual
size of the address data.

The format of the address data depends on the socket namespace.  The
length of the information is usually fixed for a given namespace, so
normally you can know exactly how much space is needed and can provide
that much.  The usual practice is to allocate a place for the value
using the proper data type for the socket's namespace, then cast its
address to @code{struct sockaddr *} to pass it to @code{getsockname}.

The return value is @code{0} on success and @code{-1} on error.  The
following @code{errno} error conditions are defined for this function:

@table @code
@item EBADF
The @var{socket} argument is not a valid file descriptor.

@item ENOTSOCK
The descriptor @var{socket} is not a socket.

@item ENOBUFS
There are not enough internal buffers available for the operation.
@end table
@end deftypefun

You can't read the address of a socket in the file namespace.  This is
consistent with the rest of the system; in general, there's no way to
find a file's name from a descriptor for that file.

@node File Namespace
@section The File Namespace
@cindex file namespace, for sockets

This section describes the details of the file namespace, whose
symbolic name (required when you create a socket) is @code{PF_FILE}.

@menu
* Concepts: File Namespace Concepts.    What you need to understand.
* Details: File Namespace Details.      Address format, symbolic names, etc.
* Example: File Socket Example.         Example of creating a socket.
@end menu

@node File Namespace Concepts
@subsection File Namespace Concepts

In the file namespace, socket addresses are file names.  You can specify
any file name you want as the address of the socket, but you must have
write permission on the directory containing it.  In order to connect to
a socket, you must have read permission for it.  It's common to put
these files in the @file{/tmp} directory.

One peculiarity of the file namespace is that the name is only used when
opening the connection; once that is over with, the address is not
meaningful and may not exist.

Another peculiarity is that you cannot connect to such a socket from
another machine--not even if the other machine shares the file system
which contains the name of the socket.  You can see the socket in a
directory listing, but connecting to it never succeeds.  Some programs
take advantage of this, such as by asking the client to send its own
process ID, and using the process IDs to distinguish between clients.
However, we recommend you not use this method in protocols you design,
as we might someday permit connections from other machines that mount
the same file systems.  Instead, send each new client an identifying
number if you want it to have one.

After you close a socket in the file namespace, you should delete the
file name from the file system.  Use @code{unlink} or @code{remove} to
do this; see @ref{Deleting Files}.

The file namespace supports just one protocol for any communication
style; it is protocol number @code{0}.

@node File Namespace Details
@subsection Details of File Namespace

@pindex sys/socket.h
To create a socket in the file namespace, use the constant
@code{PF_FILE} as the @var{namespace} argument to @code{socket} or
@code{socketpair}.  This constant is defined in @file{sys/socket.h}.

@comment sys/socket.h
@comment GNU
@deftypevr Macro int PF_FILE
This designates the file namespace, in which socket addresses are file
names, and its associated family of protocols.
@end deftypevr

@comment sys/socket.h
@comment BSD
@deftypevr Macro int PF_UNIX
This is a synonym for @code{PF_FILE}, for compatibility's sake.
@end deftypevr

The structure for specifying socket names in the file namespace is
defined in the header file @file{sys/un.h}:
@pindex sys/un.h

@comment sys/un.h
@comment BSD
@deftp {Data Type} {struct sockaddr_un}
This structure is used to specify file namespace socket addresses.  It has
the following members:

@table @code
@item short int sun_family
This identifies the address family or format of the socket address.
You should store the value @code{AF_FILE} to designate the file
namespace.  @xref{Socket Addresses}.

@item char sun_path[108]
This is the file name to use.

@strong{Incomplete:}  Why is 108 a magic number?  RMS suggests making
this a zero-length array and tweaking the example following to use
@code{alloca} to allocate an appropriate amount of storage based on
the length of the filename.
@end table
@end deftp

You should compute the @var{length} parameter for a socket address in
the file namespace as the sum of the size of the @code{sun_family}
component and the string length (@emph{not} the allocation size!) of
the file name string.

@node File Socket Example
@subsection Example of File-Namespace Sockets

Here is an example showing how to create and name a socket in the file
namespace.

@smallexample
@include mkfsock.c.texi
@end smallexample

@node Internet Namespace
@section The Internet Namespace
@cindex Internet namespace, for sockets

This section describes the details the protocols and socket naming
conventions used in the Internet namespace.

To create a socket in the Internet namespace, use the symbolic name
@code{PF_INET} of this namespace as the @var{namespace} argument to
@code{socket} or @code{socketpair}.  This macro is defined in
@file{sys/socket.h}.
@pindex sys/socket.h

@comment sys/socket.h
@comment BSD
@deftypevr Macro int PF_INET
This designates the Internet namespace and associated family of
protocols.
@end deftypevr

A socket address for the Internet namespace includes the following components:

@itemize @bullet
@item
The address of the machine you want to connect to.  Internet addresses
can be specified in several ways; these are discussed in @ref{Internet
Address Formats}, @ref{Host Addresses}, and @ref{Host Names}.

@item
A port number for that machine.  @xref{Ports}.
@end itemize

You must ensure that the address and port number are represented in a
canonical format called @dfn{network byte order}.  @xref{Byte Order},
for information about this.

@menu
* Internet Address Formats::    How socket addresses are specified in the
                                 Internet namespace.
* Host Addresses::              All about host addresses of internet host.
* Protocols Database::          Referring to protocols by name.
* Ports::                       Internet port numbers.
* Services Database::           Ports may have symbolic names.
* Byte Order::                  Different hosts may use different byte
                                 ordering conventions; you need to
                                 canonicalize host address and port number.
* Inet Example::                Putting it all together.
@end menu

@node Internet Address Formats
@subsection Internet Socket Address Formats

In the Internet namespace, for both IPv4 (@code{AF_INET}) and IPv6
(@code{AF_INET6}), a socket address consists of a host address
and a port on that host.  In addition, the protocol you choose serves
effectively as a part of the address because local port numbers are
meaningful only within a particular protocol.

The data types for representing socket addresses in the Internet namespace
are defined in the header file @file{netinet/in.h}.
@pindex netinet/in.h

@comment netinet/in.h
@comment BSD
@deftp {Data Type} {struct sockaddr_in}
This is the data type used to represent socket addresses in the
Internet namespace.  It has the following members:

@table @code
@item short int sin_family
This identifies the address family or format of the socket address.
You should store the value of @code{AF_INET} in this member.
@xref{Socket Addresses}.

@item struct in_addr sin_addr
This is the Internet address of the host machine.  @xref{Host
Addresses}, and @ref{Host Names}, for how to get a value to store
here.

@item unsigned short int sin_port
This is the port number.  @xref{Ports}.
@end table
@end deftp

When you call @code{bind} or @code{getsockname}, you should specify
@code{sizeof (struct sockaddr_in)} as the @var{length} parameter if
you are using an Internet namespace socket address.

@deftp {Data Type} {struct sockaddr_in6}
This is the data type used to represent socket addresses in the IPv6
namespace.  It has the following members:

@table @code
@item short int sin6_family
This identifies the address family or format of the socket address.
You should store the value of @code{AF_INET6} in this member.
@xref{Socket Addresses}.

@item struct in6_addr sin6_addr
This is the IPv6 address of the host machine.  @xref{Host
Addresses}, and @ref{Host Names}, for how to get a value to store
here.

@item uint32_t sin6_flowinfo
This is a currently unimplemented field.

@item uint16_t sin6_port
This is the port number.  @xref{Ports}.

@end table
@end deftp

@node Host Addresses
@subsection Host Addresses

Each computer on the Internet has one or more @dfn{Internet addresses},
numbers which identify that computer among all those on the Internet.
Users typically write IPv4 numeric host addresses as sequences of four
numbers, separated by periods, as in @samp{128.52.46.32}, and IPv6
numeric host addresses as sequences of up to eight numbers separated by
colons, as in @samp{5f03:1200:836f:c100::1}.

Each computer also has one or more @dfn{host names}, which are strings
of words separated by periods, as in @samp{churchy.gnu.ai.mit.edu}.

Programs that let the user specify a host typically accept both numeric
addresses and host names.  But the program needs a numeric address to
open a connection; to use a host name, you must convert it to the
numeric address it stands for.

@menu
* Abstract Host Addresses::     What a host number consists of.
* Data type: Host Address Data Type.    Data type for a host number.
* Functions: Host Address Functions.    Functions to operate on them.
* Names: Host Names.            Translating host names to host numbers.
@end menu

@node Abstract Host Addresses
@subsubsection Internet Host Addresses
@cindex host address, Internet
@cindex Internet host address

@ifinfo
Each computer on the Internet has one or more Internet addresses,
numbers which identify that computer among all those on the Internet.
@end ifinfo

@c I think this whole section could possibly be removed.  It is slightly
@c misleading these days.

@cindex network number
@cindex local network address number
An Internet host address is a number containing four bytes of data.
These are divided into two parts, a @dfn{network number} and a
@dfn{local network address number} within that network.  The network
number consists of the first one, two or three bytes; the rest of the
bytes are the local address.

Network numbers are registered with the Network Information Center
(NIC), and are divided into three classes---A, B, and C.  The local
network address numbers of individual machines are registered with the
administrator of the particular network.

Class A networks have single-byte numbers in the range 0 to 127.  There
are only a small number of Class A networks, but they can each support a
very large number of hosts.  Medium-sized Class B networks have two-byte
network numbers, with the first byte in the range 128 to 191.  Class C
networks are the smallest; they have three-byte network numbers, with
the first byte in the range 192-255.  Thus, the first 1, 2, or 3 bytes
of an Internet address specifies a network.  The remaining bytes of the
Internet address specify the address within that network.

The Class A network 0 is reserved for broadcast to all networks.  In
addition, the host number 0 within each network is reserved for broadcast
to all hosts in that network.

The Class A network 127 is reserved for loopback; you can always use
the Internet address @samp{127.0.0.1} to refer to the host machine.

Since a single machine can be a member of multiple networks, it can
have multiple Internet host addresses.  However, there is never
supposed to be more than one machine with the same host address.

@c !!! this section could document the IN_CLASS* macros in <netinet/in.h>.

@cindex standard dot notation, for Internet addresses
@cindex dot notation, for Internet addresses
There are four forms of the @dfn{standard numbers-and-dots notation}
for Internet addresses:

@table @code
@item @var{a}.@var{b}.@var{c}.@var{d}
This specifies all four bytes of the address individually.

@item @var{a}.@var{b}.@var{c}
The last part of the address, @var{c}, is interpreted as a 2-byte quantity.
This is useful for specifying host addresses in a Class B network with
network address number @code{@var{a}.@var{b}}.

@item @var{a}.@var{b}
The last part of the address, @var{c}, is interpreted as a 3-byte quantity.
This is useful for specifying host addresses in a Class A network with
network address number @var{a}.

@item @var{a}
If only one part is given, this corresponds directly to the host address
number.
@end table

Within each part of the address, the usual C conventions for specifying
the radix apply.  In other words, a leading @samp{0x} or @samp{0X} implies
hexadecimal radix; a leading @samp{0} implies octal; and otherwise decimal
radix is assumed.

@node Host Address Data Type
@subsubsection Host Address Data Type

Internet host addresses are represented in some contexts as integers
(type @code{unsigned long int}).  In other contexts, the integer is
packaged inside a structure of type @code{struct in_addr}.  It would
be better if the usage were made consistent, but it is not hard to extract
the integer from the structure or put the integer into a structure.

The following basic definitions for Internet addresses appear in the
header file@*@file{netinet/in.h}:
@pindex netinet/in.h

@comment netinet/in.h
@comment BSD
@deftp {Data Type} {struct in_addr}
This data type is used in certain contexts to contain an Internet host
address.  It has just one field, named @code{s_addr}, which records the
host address number as an @code{unsigned long int}.
@end deftp

@comment netinet/in.h
@comment BSD
@deftypevr Macro {unsigned int} INADDR_LOOPBACK
You can use this constant to stand for ``the address of this machine,''
instead of finding its actual address.  It is the Internet address
@samp{127.0.0.1}, which is usually called @samp{localhost}.  This
special constant saves you the trouble of looking up the address of your
own machine.  Also, the system usually implements @code{INADDR_LOOPBACK}
specially, avoiding any network traffic for the case of one machine
talking to itself.
@end deftypevr

@comment netinet/in.h
@comment BSD
@deftypevr Macro {unsigned int} INADDR_ANY
You can use this constant to stand for ``any incoming address,'' when
binding to an address.  @xref{Setting Address}.  This is the usual
address to give in the @code{sin_addr} member of @w{@code{struct
sockaddr_in}} when you want to accept Internet connections.
@end deftypevr

@comment netinet/in.h
@comment BSD
@deftypevr Macro {unsigned int} INADDR_BROADCAST
This constant is the address you use to send a broadcast message.
@c !!! broadcast needs further documented
@end deftypevr

@comment netinet/in.h
@comment BSD
@deftypevr Macro {unsigned int} INADDR_NONE
This constant is returned by some functions to indicate an error.
@end deftypevr

@comment netinet/in.h
@comment IPv6 basic API
@deftp {Data Type} {struct in6_addr}
This data type is used to store an IPv6 address.  It stores 128 bits of
data, which can be accessed (via a union) in a variety of ways.
@end deftp

@comment netinet/in.h
@comment IPv6 basic API
@deftypevr Constant {struct in6_addr} in6addr_loopback.
This constant is the IPv6 address @samp{::1}, the loopback address.  See
above for a description of what this means.  The macro
@code{IN6ADDR_LOOPBACK_INIT} is provided to allow you to initialise your
own variables to this value.
@end deftypevr

@comment netinet/in.h
@comment IPv6 basic API
@deftypevr Constant {struct in6_addr} in6addr_any
This constant is the IPv6 address @samp{::}, the unspecified address.  See
above for a description of what this means.  The macro
@code{IN6ADDR_ANY_INIT} is provided to allow you to initialise your
own variables to this value.
@end deftypevr

@node Host Address Functions
@subsubsection Host Address Functions

@pindex arpa/inet.h
@noindent
These additional functions for manipulating Internet addresses are
declared in@*@file{arpa/inet.h}.  They represent Internet addresses in
network byte order; they represent network numbers and
local-address-within-network numbers in host byte order.
@xref{Byte Order}, for an explanation of network and host byte order.

@comment arpa/inet.h
@comment BSD
@deftypefun int inet_aton (const char *@var{name}, struct in_addr *@var{addr})
This function converts the Internet host address @var{name}
from the standard numbers-and-dots notation into binary data and stores
it in the @code{struct in_addr} that @var{addr} points to.
@code{inet_aton} returns nonzero if the address is valid, zero if not.
@end deftypefun

@comment arpa/inet.h
@comment BSD
@deftypefun {unsigned long int} inet_addr (const char *@var{name})
This function converts the Internet host address @var{name} from the
standard numbers-and-dots notation into binary data.  If the input is
not valid, @code{inet_addr} returns @code{INADDR_NONE}.  This is an
obsolete interface to @code{inet_aton}, described immediately above; it
is obsolete because @code{INADDR_NONE} is a valid address
(255.255.255.255), and @code{inet_aton} provides a cleaner way to
indicate error return.
@end deftypefun

@comment arpa/inet.h
@comment BSD
@deftypefun {unsigned long int} inet_network (const char *@var{name})
This function extracts the network number from the address @var{name},
given in the standard numbers-and-dots notation. The returned address is
in host order. If the input is not valid, @code{inet_network} returns
@code{-1}.
@end deftypefun

@comment arpa/inet.h
@comment BSD
@deftypefun {char *} inet_ntoa (struct in_addr @var{addr})
This function converts the Internet host address @var{addr} to a
string in the standard numbers-and-dots notation.  The return value is
a pointer into a statically-allocated buffer.  Subsequent calls will
overwrite the same buffer, so you should copy the string if you need
to save it.

In multi-threaded programs each thread has an own statically-allocated
buffer.  But still subsequent calls of @code{inet_ntoa} in the same
thread will overwrite the result of the last call.
@end deftypefun

@comment arpa/inet.h
@comment BSD
@deftypefun {struct in_addr} inet_makeaddr (int @var{net}, int @var{local})
This function makes an Internet host address by combining the network
number @var{net} with the local-address-within-network number
@var{local}.
@end deftypefun

@comment arpa/inet.h
@comment BSD
@deftypefun int inet_lnaof (struct in_addr @var{addr})
This function returns the local-address-within-network part of the
Internet host address @var{addr}.
@end deftypefun

@comment arpa/inet.h
@comment BSD
@deftypefun int inet_netof (struct in_addr @var{addr})
This function returns the network number part of the Internet host
address @var{addr}.
@end deftypefun

@comment arpa/inet.h
@comment IPv6 basic API
@deftypefun int inet_pton (int @var{af}, const char *@var{cp}, void *@var{buf})
This function converts an Internet address (either IPv4 or IPv6) from
presentation (textual) to network (binary) format.  @var{af} should be
either @code{AF_INET} or @code{AF_INET6}, as appropriate for the type of
address being converted.  @var{cp} is a pointer to the input string, and
@var{buf} is a pointer to a buffer for the result.  It is the caller's
responsibility to make sure the buffer is large enough.
@end deftypefun

@comment arpa/inet.h
@comment IPv6 basic API
@deftypefun {char *} inet_ntop (int @var{af}, const void *@var{cp}, char *@var{buf}, size_t @var{len})
This function converts an Internet address (either IPv4 or IPv6) from
network (binary) to presentation (textual) form.  @var{af} should be
either @code{AF_INET} or @code{AF_INET6}, as appropriate.  @var{cp} is a
pointer to the address to be converted.  @var{buf} should be a pointer
to a buffer to hold the result, and @var{len} is the length of this
buffer.  The return value from the function will be this buffer address.
@end deftypefun

@node Host Names
@subsubsection Host Names
@cindex hosts database
@cindex converting host name to address
@cindex converting host address to name

Besides the standard numbers-and-dots notation for Internet addresses,
you can also refer to a host by a symbolic name.  The advantage of a
symbolic name is that it is usually easier to remember.  For example,
the machine with Internet address @samp{128.52.46.32} is also known as
@samp{churchy.gnu.ai.mit.edu}; and other machines in the @samp{gnu.ai.mit.edu}
domain can refer to it simply as @samp{churchy}.

@pindex /etc/hosts
@pindex netdb.h
Internally, the system uses a database to keep track of the mapping
between host names and host numbers.  This database is usually either
the file @file{/etc/hosts} or an equivalent provided by a name server.
The functions and other symbols for accessing this database are declared
in @file{netdb.h}.  They are BSD features, defined unconditionally if
you include @file{netdb.h}.

@comment netdb.h
@comment BSD
@deftp {Data Type} {struct hostent}
This data type is used to represent an entry in the hosts database.  It
has the following members:

@table @code
@item char *h_name
This is the ``official'' name of the host.

@item char **h_aliases
These are alternative names for the host, represented as a null-terminated
vector of strings.

@item int h_addrtype
This is the host address type; in practice, its value is always either
@code{AF_INET} or @code{AF_INET6}, with the latter being used for IPv6
hosts.  In principle other kinds of addresses could be represented in
the data base as well as Internet addresses; if this were done, you
might find a value in this field other than @code{AF_INET} or
@code{AF_INET6}.  @xref{Socket Addresses}.

@item int h_length
This is the length, in bytes, of each address.

@item char **h_addr_list
This is the vector of addresses for the host.  (Recall that the host
might be connected to multiple networks and have different addresses on
each one.)  The vector is terminated by a null pointer.

@item char *h_addr
This is a synonym for @code{h_addr_list[0]}; in other words, it is the
first host address.
@end table
@end deftp

As far as the host database is concerned, each address is just a block
of memory @code{h_length} bytes long.  But in other contexts there is an
implicit assumption that you can convert this to a @code{struct in_addr} or
an @code{unsigned long int}.  Host addresses in a @code{struct hostent}
structure are always given in network byte order; see @ref{Byte Order}.

You can use @code{gethostbyname}, @code{gethostbyname2} or
@code{gethostbyaddr} to search the hosts database for information about
a particular host.  The information is returned in a
statically-allocated structure; you must copy the information if you
need to save it across calls.  You can also use @code{getaddrinfo} and
@code{getnameinfo} to obtain this information.

@comment netdb.h
@comment BSD
@deftypefun {struct hostent *} gethostbyname (const char *@var{name})
The @code{gethostbyname} function returns information about the host
named @var{name}.  If the lookup fails, it returns a null pointer.
@end deftypefun

@comment netdb.h
@comment IPv6 Basic API
@deftypefun {struct hostent *} gethostbyname2 (const char *@var{name}, int @var{af})
The @code{gethostbyname2} function is like @code{gethostbyname}, but
allows the caller to specify the desired address family (e.g.@:
@code{AF_INET} or @code{AF_INET6}) for the result.
@end deftypefun

@comment netdb.h
@comment BSD
@deftypefun {struct hostent *} gethostbyaddr (const char *@var{addr}, int @var{length}, int @var{format})
The @code{gethostbyaddr} function returns information about the host
with Internet address @var{addr}.  The @var{length} argument is the
size (in bytes) of the address at @var{addr}.  @var{format} specifies
the address format; for an Internet address, specify a value of
@code{AF_INET}.

If the lookup fails, @code{gethostbyaddr} returns a null pointer.
@end deftypefun

@vindex h_errno
If the name lookup by @code{gethostbyname} or @code{gethostbyaddr}
fails, you can find out the reason by looking at the value of the
variable @code{h_errno}.  (It would be cleaner design for these
functions to set @code{errno}, but use of @code{h_errno} is compatible
with other systems.)  Before using @code{h_errno}, you must declare it
like this:

@smallexample
extern int h_errno;
@end smallexample

Here are the error codes that you may find in @code{h_errno}:

@table @code
@comment netdb.h
@comment BSD
@item HOST_NOT_FOUND
@vindex HOST_NOT_FOUND
No such host is known in the data base.

@comment netdb.h
@comment BSD
@item TRY_AGAIN
@vindex TRY_AGAIN
This condition happens when the name server could not be contacted.  If
you try again later, you may succeed then.

@comment netdb.h
@comment BSD
@item NO_RECOVERY
@vindex NO_RECOVERY
A non-recoverable error occurred.

@comment netdb.h
@comment BSD
@item NO_ADDRESS
@vindex NO_ADDRESS
The host database contains an entry for the name, but it doesn't have an
associated Internet address.
@end table

You can also scan the entire hosts database one entry at a time using
@code{sethostent}, @code{gethostent}, and @code{endhostent}.  Be careful
in using these functions, because they are not reentrant.

@comment netdb.h
@comment BSD
@deftypefun void sethostent (int @var{stayopen})
This function opens the hosts database to begin scanning it.  You can
then call @code{gethostent} to read the entries.

@c There was a rumor that this flag has different meaning if using the DNS,
@c but it appears this description is accurate in that case also.
If the @var{stayopen} argument is nonzero, this sets a flag so that
subsequent calls to @code{gethostbyname} or @code{gethostbyaddr} will
not close the database (as they usually would).  This makes for more
efficiency if you call those functions several times, by avoiding
reopening the database for each call.
@end deftypefun

@comment netdb.h
@comment BSD
@deftypefun {struct hostent *} gethostent ()
This function returns the next entry in the hosts database.  It
returns a null pointer if there are no more entries.
@end deftypefun

@comment netdb.h
@comment BSD
@deftypefun void endhostent ()
This function closes the hosts database.
@end deftypefun

@node Ports
@subsection Internet Ports
@cindex port number

A socket address in the Internet namespace consists of a machine's
Internet address plus a @dfn{port number} which distinguishes the
sockets on a given machine (for a given protocol).  Port numbers range
from 0 to 65,535.

Port numbers less than @code{IPPORT_RESERVED} are reserved for standard
servers, such as @code{finger} and @code{telnet}.  There is a database
that keeps track of these, and you can use the @code{getservbyname}
function to map a service name onto a port number; see @ref{Services
Database}.

If you write a server that is not one of the standard ones defined in
the database, you must choose a port number for it.  Use a number
greater than @code{IPPORT_USERRESERVED}; such numbers are reserved for
servers and won't ever be generated automatically by the system.
Avoiding conflicts with servers being run by other users is up to you.

When you use a socket without specifying its address, the system
generates a port number for it.  This number is between
@code{IPPORT_RESERVED} and @code{IPPORT_USERRESERVED}.

On the Internet, it is actually legitimate to have two different
sockets with the same port number, as long as they never both try to
communicate with the same socket address (host address plus port
number).  You shouldn't duplicate a port number except in special
circumstances where a higher-level protocol requires it.  Normally,
the system won't let you do it; @code{bind} normally insists on
distinct port numbers.  To reuse a port number, you must set the
socket option @code{SO_REUSEADDR}.  @xref{Socket-Level Options}.

@pindex netinet/in.h
These macros are defined in the header file @file{netinet/in.h}.

@comment netinet/in.h
@comment BSD
@deftypevr Macro int IPPORT_RESERVED
Port numbers less than @code{IPPORT_RESERVED} are reserved for
superuser use.
@end deftypevr

@comment netinet/in.h
@comment BSD
@deftypevr Macro int IPPORT_USERRESERVED
Port numbers greater than or equal to @code{IPPORT_USERRESERVED} are
reserved for explicit use; they will never be allocated automatically.
@end deftypevr

@node Services Database
@subsection The Services Database
@cindex services database
@cindex converting service name to port number
@cindex converting port number to service name

@pindex /etc/services
The database that keeps track of ``well-known'' services is usually
either the file @file{/etc/services} or an equivalent from a name server.
You can use these utilities, declared in @file{netdb.h}, to access
the services database.
@pindex netdb.h

@comment netdb.h
@comment BSD
@deftp {Data Type} {struct servent}
This data type holds information about entries from the services database.
It has the following members:

@table @code
@item char *s_name
This is the ``official'' name of the service.

@item char **s_aliases
These are alternate names for the service, represented as an array of
strings.  A null pointer terminates the array.

@item int s_port
This is the port number for the service.  Port numbers are given in
network byte order; see @ref{Byte Order}.

@item char *s_proto
This is the name of the protocol to use with this service.
@xref{Protocols Database}.
@end table
@end deftp

To get information about a particular service, use the
@code{getservbyname} or @code{getservbyport} functions.  The information
is returned in a statically-allocated structure; you must copy the
information if you need to save it across calls.

@comment netdb.h
@comment BSD
@deftypefun {struct servent *} getservbyname (const char *@var{name}, const char *@var{proto})
The @code{getservbyname} function returns information about the
service named @var{name} using protocol @var{proto}.  If it can't find
such a service, it returns a null pointer.

This function is useful for servers as well as for clients; servers
use it to determine which port they should listen on (@pxref{Listening}).
@end deftypefun

@comment netdb.h
@comment BSD
@deftypefun {struct servent *} getservbyport (int @var{port}, const char *@var{proto})
The @code{getservbyport} function returns information about the
service at port @var{port} using protocol @var{proto}.  If it can't
find such a service, it returns a null pointer.
@end deftypefun

@noindent
You can also scan the services database using @code{setservent},
@code{getservent}, and @code{endservent}.  Be careful in using these
functions, because they are not reentrant.

@comment netdb.h
@comment BSD
@deftypefun void setservent (int @var{stayopen})
This function opens the services database to begin scanning it.

If the @var{stayopen} argument is nonzero, this sets a flag so that
subsequent calls to @code{getservbyname} or @code{getservbyport} will
not close the database (as they usually would).  This makes for more
efficiency if you call those functions several times, by avoiding
reopening the database for each call.
@end deftypefun

@comment netdb.h
@comment BSD
@deftypefun {struct servent *} getservent (void)
This function returns the next entry in the services database.  If
there are no more entries, it returns a null pointer.
@end deftypefun

@comment netdb.h
@comment BSD
@deftypefun void endservent (void)
This function closes the services database.
@end deftypefun

@node Byte Order
@subsection Byte Order Conversion
@cindex byte order conversion, for socket
@cindex converting byte order

@cindex big-endian
@cindex little-endian
Different kinds of computers use different conventions for the
ordering of bytes within a word.  Some computers put the most
significant byte within a word first (this is called ``big-endian''
order), and others put it last (``little-endian'' order).

@cindex network byte order
So that machines with different byte order conventions can
communicate, the Internet protocols specify a canonical byte order
convention for data transmitted over the network.  This is known
as the @dfn{network byte order}.

When establishing an Internet socket connection, you must make sure that
the data in the @code{sin_port} and @code{sin_addr} members of the
@code{sockaddr_in} structure are represented in the network byte order.
If you are encoding integer data in the messages sent through the
socket, you should convert this to network byte order too.  If you don't
do this, your program may fail when running on or talking to other kinds
of machines.

If you use @code{getservbyname} and @code{gethostbyname} or
@code{inet_addr} to get the port number and host address, the values are
already in the network byte order, and you can copy them directly into
the @code{sockaddr_in} structure.

Otherwise, you have to convert the values explicitly.  Use
@code{htons} and @code{ntohs} to convert values for the @code{sin_port}
member.  Use @code{htonl} and @code{ntohl} to convert values for the
@code{sin_addr} member.  (Remember, @code{struct in_addr} is equivalent
to @code{unsigned long int}.)  These functions are declared in
@file{netinet/in.h}.
@pindex netinet/in.h

@comment netinet/in.h
@comment BSD
@deftypefun {unsigned short int} htons (unsigned short int @var{hostshort})
This function converts the @code{short} integer @var{hostshort} from
host byte order to network byte order.
@end deftypefun

@comment netinet/in.h
@comment BSD
@deftypefun {unsigned short int} ntohs (unsigned short int @var{netshort})
This function converts the @code{short} integer @var{netshort} from
network byte order to host byte order.
@end deftypefun

@comment netinet/in.h
@comment BSD
@deftypefun {unsigned long int} htonl (unsigned long int @var{hostlong})
This function converts the @code{long} integer @var{hostlong} from
host byte order to network byte order.
@end deftypefun

@comment netinet/in.h
@comment BSD
@deftypefun {unsigned long int} ntohl (unsigned long int @var{netlong})
This function converts the @code{long} integer @var{netlong} from
network byte order to host byte order.
@end deftypefun

@node Protocols Database
@subsection Protocols Database
@cindex protocols database

The communications protocol used with a socket controls low-level
details of how data is exchanged.  For example, the protocol implements
things like checksums to detect errors in transmissions, and routing
instructions for messages.  Normal user programs have little reason to
mess with these details directly.

@cindex TCP (Internet protocol)
The default communications protocol for the Internet namespace depends on
the communication style.  For stream communication, the default is TCP
(``transmission control protocol'').  For datagram communication, the
default is UDP (``user datagram protocol'').  For reliable datagram
communication, the default is RDP (``reliable datagram protocol'').
You should nearly always use the default.

@pindex /etc/protocols
Internet protocols are generally specified by a name instead of a
number.  The network protocols that a host knows about are stored in a
database.  This is usually either derived from the file
@file{/etc/protocols}, or it may be an equivalent provided by a name
server.  You look up the protocol number associated with a named
protocol in the database using the @code{getprotobyname} function.

Here are detailed descriptions of the utilities for accessing the
protocols database.  These are declared in @file{netdb.h}.
@pindex netdb.h

@comment netdb.h
@comment BSD
@deftp {Data Type} {struct protoent}
This data type is used to represent entries in the network protocols
database.  It has the following members:

@table @code
@item char *p_name
This is the official name of the protocol.

@item char **p_aliases
These are alternate names for the protocol, specified as an array of
strings.  The last element of the array is a null pointer.

@item int p_proto
This is the protocol number (in host byte order); use this member as the
@var{protocol} argument to @code{socket}.
@end table
@end deftp

You can use @code{getprotobyname} and @code{getprotobynumber} to search
the protocols database for a specific protocol.  The information is
returned in a statically-allocated structure; you must copy the
information if you need to save it across calls.

@comment netdb.h
@comment BSD
@deftypefun {struct protoent *} getprotobyname (const char *@var{name})
The @code{getprotobyname} function returns information about the
network protocol named @var{name}.  If there is no such protocol, it
returns a null pointer.
@end deftypefun

@comment netdb.h
@comment BSD
@deftypefun {struct protoent *} getprotobynumber (int @var{protocol})
The @code{getprotobynumber} function returns information about the
network protocol with number @var{protocol}.  If there is no such
protocol, it returns a null pointer.
@end deftypefun

You can also scan the whole protocols database one protocol at a time by
using @code{setprotoent}, @code{getprotoent}, and @code{endprotoent}.
Be careful in using these functions, because they are not reentrant.

@comment netdb.h
@comment BSD
@deftypefun void setprotoent (int @var{stayopen})
This function opens the protocols database to begin scanning it.

If the @var{stayopen} argument is nonzero, this sets a flag so that
subsequent calls to @code{getprotobyname} or @code{getprotobynumber} will
not close the database (as they usually would).  This makes for more
efficiency if you call those functions several times, by avoiding
reopening the database for each call.
@end deftypefun

@comment netdb.h
@comment BSD
@deftypefun {struct protoent *} getprotoent (void)
This function returns the next entry in the protocols database.  It
returns a null pointer if there are no more entries.
@end deftypefun

@comment netdb.h
@comment BSD
@deftypefun void endprotoent (void)
This function closes the protocols database.
@end deftypefun

@node Inet Example
@subsection Internet Socket Example

Here is an example showing how to create and name a socket in the
Internet namespace.  The newly created socket exists on the machine that
the program is running on.  Rather than finding and using the machine's
Internet address, this example specifies @code{INADDR_ANY} as the host
address; the system replaces that with the machine's actual address.

@smallexample
@include mkisock.c.texi
@end smallexample

Here is another example, showing how you can fill in a @code{sockaddr_in}
structure, given a host name string and a port number:

@smallexample
@include isockad.c.texi
@end smallexample

@node Misc Namespaces
@section Other Namespaces

@vindex PF_NS
@vindex PF_ISO
@vindex PF_CCITT
@vindex PF_IMPLINK
@vindex PF_ROUTE
Certain other namespaces and associated protocol families are supported
but not documented yet because they are not often used.  @code{PF_NS}
refers to the Xerox Network Software protocols.  @code{PF_ISO} stands
for Open Systems Interconnect.  @code{PF_CCITT} refers to protocols from
CCITT.  @file{socket.h} defines these symbols and others naming protocols
not actually implemented.

@code{PF_IMPLINK} is used for communicating between hosts and Internet
Message Processors.  For information on this, and on @code{PF_ROUTE}, an
occasionally-used local area routing protocol, see the GNU Hurd Manual
(to appear in the future).

@node Open/Close Sockets
@section Opening and Closing Sockets

This section describes the actual library functions for opening and
closing sockets.  The same functions work for all namespaces and
connection styles.

@menu
* Creating a Socket::           How to open a socket.
* Closing a Socket::            How to close a socket.
* Socket Pairs::                These are created like pipes.
@end menu

@node Creating a Socket
@subsection Creating a Socket
@cindex creating a socket
@cindex socket, creating
@cindex opening a socket

The primitive for creating a socket is the @code{socket} function,
declared in @file{sys/socket.h}.
@pindex sys/socket.h

@comment sys/socket.h
@comment BSD
@deftypefun int socket (int @var{namespace}, int @var{style}, int @var{protocol})
This function creates a socket and specifies communication style
@var{style}, which should be one of the socket styles listed in
@ref{Communication Styles}.  The @var{namespace} argument specifies
the namespace; it must be @code{PF_FILE} (@pxref{File Namespace}) or
@code{PF_INET} (@pxref{Internet Namespace}).  @var{protocol}
designates the specific protocol (@pxref{Socket Concepts}); zero is
usually right for @var{protocol}.

The return value from @code{socket} is the file descriptor for the new
socket, or @code{-1} in case of error.  The following @code{errno} error
conditions are defined for this function:

@table @code
@item EPROTONOSUPPORT
The @var{protocol} or @var{style} is not supported by the
@var{namespace} specified.

@item EMFILE
The process already has too many file descriptors open.

@item ENFILE
The system already has too many file descriptors open.

@item EACCESS
The process does not have privilege to create a socket of the specified
@var{style} or @var{protocol}.

@item ENOBUFS
The system ran out of internal buffer space.
@end table

The file descriptor returned by the @code{socket} function supports both
read and write operations.  But, like pipes, sockets do not support file
positioning operations.
@end deftypefun

For examples of how to call the @code{socket} function,
see @ref{File Namespace}, or @ref{Inet Example}.


@node Closing a Socket
@subsection Closing a Socket
@cindex socket, closing
@cindex closing a socket
@cindex shutting down a socket
@cindex socket shutdown

When you are finished using a socket, you can simply close its
file descriptor with @code{close}; see @ref{Opening and Closing Files}.
If there is still data waiting to be transmitted over the connection,
normally @code{close} tries to complete this transmission.  You
can control this behavior using the @code{SO_LINGER} socket option to
specify a timeout period; see @ref{Socket Options}.

@pindex sys/socket.h
You can also shut down only reception or only transmission on a
connection by calling @code{shutdown}, which is declared in
@file{sys/socket.h}.

@comment sys/socket.h
@comment BSD
@deftypefun int shutdown (int @var{socket}, int @var{how})
The @code{shutdown} function shuts down the connection of socket
@var{socket}.  The argument @var{how} specifies what action to
perform:

@table @code
@item 0
Stop receiving data for this socket.  If further data arrives,
reject it.

@item 1
Stop trying to transmit data from this socket.  Discard any data
waiting to be sent.  Stop looking for acknowledgement of data already
sent; don't retransmit it if it is lost.

@item 2
Stop both reception and transmission.
@end table

The return value is @code{0} on success and @code{-1} on failure.  The
following @code{errno} error conditions are defined for this function:

@table @code
@item EBADF
@var{socket} is not a valid file descriptor.

@item ENOTSOCK
@var{socket} is not a socket.

@item ENOTCONN
@var{socket} is not connected.
@end table
@end deftypefun

@node Socket Pairs
@subsection Socket Pairs
@cindex creating a socket pair
@cindex socket pair
@cindex opening a socket pair

@pindex sys/socket.h
A @dfn{socket pair} consists of a pair of connected (but unnamed)
sockets.  It is very similar to a pipe and is used in much the same
way.  Socket pairs are created with the @code{socketpair} function,
declared in @file{sys/socket.h}.  A socket pair is much like a pipe; the
main difference is that the socket pair is bidirectional, whereas the
pipe has one input-only end and one output-only end (@pxref{Pipes and
FIFOs}).

@comment sys/socket.h
@comment BSD
@deftypefun int socketpair (int @var{namespace}, int @var{style}, int @var{protocol}, int @var{filedes}@t{[2]})
This function creates a socket pair, returning the file descriptors in
@code{@var{filedes}[0]} and @code{@var{filedes}[1]}.  The socket pair
is a full-duplex communications channel, so that both reading and writing
may be performed at either end.

The @var{namespace}, @var{style}, and @var{protocol} arguments are
interpreted as for the @code{socket} function.  @var{style} should be
one of the communication styles listed in @ref{Communication Styles}.
The @var{namespace} argument specifies the namespace, which must be
@code{AF_FILE} (@pxref{File Namespace}); @var{protocol} specifies the
communications protocol, but zero is the only meaningful value.

If @var{style} specifies a connectionless communication style, then
the two sockets you get are not @emph{connected}, strictly speaking,
but each of them knows the other as the default destination address,
so they can send packets to each other.

The @code{socketpair} function returns @code{0} on success and @code{-1}
on failure.  The following @code{errno} error conditions are defined
for this function:

@table @code
@item EMFILE
The process has too many file descriptors open.

@item EAFNOSUPPORT
The specified namespace is not supported.

@item EPROTONOSUPPORT
The specified protocol is not supported.

@item EOPNOTSUPP
The specified protocol does not support the creation of socket pairs.
@end table
@end deftypefun

@node Connections
@section Using Sockets with Connections

@cindex connection
@cindex client
@cindex server
The most common communication styles involve making a connection to a
particular other socket, and then exchanging data with that socket
over and over.  Making a connection is asymmetric; one side (the
@dfn{client}) acts to request a connection, while the other side (the
@dfn{server}) makes a socket and waits for the connection request.

@iftex
@itemize @bullet
@item
@ref{Connecting}, describes what the client program must do to
initiate a connection with a server.

@item
@ref{Listening}, and @ref{Accepting Connections}, describe what the
server program must do to wait for and act upon connection requests
from clients.

@item
@ref{Transferring Data}, describes how data is transferred through the
connected socket.
@end itemize
@end iftex

@menu
* Connecting::               What the client program must do.
* Listening::                How a server program waits for requests.
* Accepting Connections::    What the server does when it gets a request.
* Who is Connected::         Getting the address of the
                                other side of a connection.
* Transferring Data::        How to send and receive data.
* Byte Stream Example::      An example program: a client for communicating
                              over a byte stream socket in the Internet namespace.
* Server Example::           A corresponding server program.
* Out-of-Band Data::         This is an advanced feature.
@end menu

@node Connecting
@subsection Making a Connection
@cindex connecting a socket
@cindex socket, connecting
@cindex socket, initiating a connection
@cindex socket, client actions

In making a connection, the client makes a connection while the server
waits for and accepts the connection.  Here we discuss what the client
program must do, using the @code{connect} function, which is declared in
@file{sys/socket.h}.

@comment sys/socket.h
@comment BSD
@deftypefun int connect (int @var{socket}, struct sockaddr *@var{addr}, socklen_t @var{length})
The @code{connect} function initiates a connection from the socket
with file descriptor @var{socket} to the socket whose address is
specified by the @var{addr} and @var{length} arguments.  (This socket
is typically on another machine, and it must be already set up as a
server.)  @xref{Socket Addresses}, for information about how these
arguments are interpreted.

Normally, @code{connect} waits until the server responds to the request
before it returns.  You can set nonblocking mode on the socket
@var{socket} to make @code{connect} return immediately without waiting
for the response.  @xref{File Status Flags}, for information about
nonblocking mode.
@c !!! how do you tell when it has finished connecting?  I suspect the
@c way you do it is select for writing.

The normal return value from @code{connect} is @code{0}.  If an error
occurs, @code{connect} returns @code{-1}.  The following @code{errno}
error conditions are defined for this function:

@table @code
@item EBADF
The socket @var{socket} is not a valid file descriptor.

@item ENOTSOCK
File descriptor @var{socket} is not a socket.

@item EADDRNOTAVAIL
The specified address is not available on the remote machine.

@item EAFNOSUPPORT
The namespace of the @var{addr} is not supported by this socket.

@item EISCONN
The socket @var{socket} is already connected.

@item ETIMEDOUT
The attempt to establish the connection timed out.

@item ECONNREFUSED
The server has actively refused to establish the connection.

@item ENETUNREACH
The network of the given @var{addr} isn't reachable from this host.

@item EADDRINUSE
The socket address of the given @var{addr} is already in use.

@item EINPROGRESS
The socket @var{socket} is non-blocking and the connection could not be
established immediately.  You can determine when the connection is
completely established with @code{select}; @pxref{Waiting for I/O}.
Another @code{connect} call on the same socket, before the connection is
completely established, will fail with @code{EALREADY}.

@item EALREADY
The socket @var{socket} is non-blocking and already has a pending
connection in progress (see @code{EINPROGRESS} above).
@end table

This function is defined as a cancelation point in multi-threaded
programs.  So one has to be prepared for this and make sure that
possibly allocated resources (like memory, files descriptors,
semaphores or whatever) are freed even if the thread is cancel.
@c @xref{pthread_cleanup_push}, for a method how to do this.
@end deftypefun

@node Listening
@subsection Listening for Connections
@cindex listening (sockets)
@cindex sockets, server actions
@cindex sockets, listening

Now let us consider what the server process must do to accept
connections on a socket.  First it must use the @code{listen} function
to enable connection requests on the socket, and then accept each
incoming connection with a call to @code{accept} (@pxref{Accepting
Connections}).  Once connection requests are enabled on a server socket,
the @code{select} function reports when the socket has a connection
ready to be accepted (@pxref{Waiting for I/O}).

The @code{listen} function is not allowed for sockets using
connectionless communication styles.

You can write a network server that does not even start running until a
connection to it is requested.  @xref{Inetd Servers}.

In the Internet namespace, there are no special protection mechanisms
for controlling access to connect to a port; any process on any machine
can make a connection to your server.  If you want to restrict access to
your server, make it examine the addresses associated with connection
requests or implement some other handshaking or identification
protocol.

In the File namespace, the ordinary file protection bits control who has
access to connect to the socket.

@comment sys/socket.h
@comment BSD
@deftypefun int listen (int @var{socket}, unsigned int @var{n})
The @code{listen} function enables the socket @var{socket} to accept
connections, thus making it a server socket.

The argument @var{n} specifies the length of the queue for pending
connections.  When the queue fills, new clients attempting to connect
fail with @code{ECONNREFUSED} until the server calls @code{accept} to
accept a connection from the queue.

The @code{listen} function returns @code{0} on success and @code{-1}
on failure.  The following @code{errno} error conditions are defined
for this function:

@table @code
@item EBADF
The argument @var{socket} is not a valid file descriptor.

@item ENOTSOCK
The argument @var{socket} is not a socket.

@item EOPNOTSUPP
The socket @var{socket} does not support this operation.
@end table
@end deftypefun

@node Accepting Connections
@subsection Accepting Connections
@cindex sockets, accepting connections
@cindex accepting connections

When a server receives a connection request, it can complete the
connection by accepting the request.  Use the function @code{accept}
to do this.

A socket that has been established as a server can accept connection
requests from multiple clients.  The server's original socket
@emph{does not become part} of the connection; instead, @code{accept}
makes a new socket which participates in the connection.
@code{accept} returns the descriptor for this socket.  The server's
original socket remains available for listening for further connection
requests.

The number of pending connection requests on a server socket is finite.
If connection requests arrive from clients faster than the server can
act upon them, the queue can fill up and additional requests are refused
with a @code{ECONNREFUSED} error.  You can specify the maximum length of
this queue as an argument to the @code{listen} function, although the
system may also impose its own internal limit on the length of this
queue.

@comment sys/socket.h
@comment BSD
@deftypefun int accept (int @var{socket}, struct sockaddr *@var{addr}, socklen_t *@var{length-ptr})
This function is used to accept a connection request on the server
socket @var{socket}.

The @code{accept} function waits if there are no connections pending,
unless the socket @var{socket} has nonblocking mode set.  (You can use
@code{select} to wait for a pending connection, with a nonblocking
socket.)  @xref{File Status Flags}, for information about nonblocking
mode.

The @var{addr} and @var{length-ptr} arguments are used to return
information about the name of the client socket that initiated the
connection.  @xref{Socket Addresses}, for information about the format
of the information.

Accepting a connection does not make @var{socket} part of the
connection.  Instead, it creates a new socket which becomes
connected.  The normal return value of @code{accept} is the file
descriptor for the new socket.

After @code{accept}, the original socket @var{socket} remains open and
unconnected, and continues listening until you close it.  You can
accept further connections with @var{socket} by calling @code{accept}
again.

If an error occurs, @code{accept} returns @code{-1}.  The following
@code{errno} error conditions are defined for this function:

@table @code
@item EBADF
The @var{socket} argument is not a valid file descriptor.

@item ENOTSOCK
The descriptor @var{socket} argument is not a socket.

@item EOPNOTSUPP
The descriptor @var{socket} does not support this operation.

@item EWOULDBLOCK
@var{socket} has nonblocking mode set, and there are no pending
connections immediately available.
@end table

This function is defined as a cancelation point in multi-threaded
programs.  So one has to be prepared for this and make sure that
possibly allocated resources (like memory, files descriptors,
semaphores or whatever) are freed even if the thread is cancel.
@c @xref{pthread_cleanup_push}, for a method how to do this.
@end deftypefun

The @code{accept} function is not allowed for sockets using
connectionless communication styles.

@node Who is Connected
@subsection Who is Connected to Me?

@comment sys/socket.h
@comment BSD
@deftypefun int getpeername (int @var{socket}, struct sockaddr *@var{addr}, size_t *@var{length-ptr})
The @code{getpeername} function returns the address of the socket that
@var{socket} is connected to; it stores the address in the memory space
specified by @var{addr} and @var{length-ptr}.  It stores the length of
the address in @code{*@var{length-ptr}}.

@xref{Socket Addresses}, for information about the format of the
address.  In some operating systems, @code{getpeername} works only for
sockets in the Internet domain.

The return value is @code{0} on success and @code{-1} on error.  The
following @code{errno} error conditions are defined for this function:

@table @code
@item EBADF
The argument @var{socket} is not a valid file descriptor.

@item ENOTSOCK
The descriptor @var{socket} is not a socket.

@item ENOTCONN
The socket @var{socket} is not connected.

@item ENOBUFS
There are not enough internal buffers available.
@end table
@end deftypefun


@node Transferring Data
@subsection Transferring Data
@cindex reading from a socket
@cindex writing to a socket

Once a socket has been connected to a peer, you can use the ordinary
@code{read} and @code{write} operations (@pxref{I/O Primitives}) to
transfer data.  A socket is a two-way communications channel, so read
and write operations can be performed at either end.

There are also some I/O modes that are specific to socket operations.
In order to specify these modes, you must use the @code{recv} and
@code{send} functions instead of the more generic @code{read} and
@code{write} functions.  The @code{recv} and @code{send} functions take
an additional argument which you can use to specify various flags to
control the special I/O modes.  For example, you can specify the
@code{MSG_OOB} flag to read or write out-of-band data, the
@code{MSG_PEEK} flag to peek at input, or the @code{MSG_DONTROUTE} flag
to control inclusion of routing information on output.

@menu
* Sending Data::                Sending data with @code{send}.
* Receiving Data::              Reading data with @code{recv}.
* Socket Data Options::         Using @code{send} and @code{recv}.
@end menu

@node Sending Data
@subsubsection Sending Data

@pindex sys/socket.h
The @code{send} function is declared in the header file
@file{sys/socket.h}.  If your @var{flags} argument is zero, you can just
as well use @code{write} instead of @code{send}; see @ref{I/O
Primitives}.  If the socket was connected but the connection has broken,
you get a @code{SIGPIPE} signal for any use of @code{send} or
@code{write} (@pxref{Miscellaneous Signals}).

@comment sys/socket.h
@comment BSD
@deftypefun int send (int @var{socket}, void *@var{buffer}, size_t @var{size}, int @var{flags})
The @code{send} function is like @code{write}, but with the additional
flags @var{flags}.  The possible values of @var{flags} are described
in @ref{Socket Data Options}.

This function returns the number of bytes transmitted, or @code{-1} on
failure.  If the socket is nonblocking, then @code{send} (like
@code{write}) can return after sending just part of the data.
@xref{File Status Flags}, for information about nonblocking mode.

Note, however, that a successful return value merely indicates that
the message has been sent without error, not necessarily that it has
been received without error.

The following @code{errno} error conditions are defined for this function:

@table @code
@item EBADF
The @var{socket} argument is not a valid file descriptor.

@item EINTR
The operation was interrupted by a signal before any data was sent.
@xref{Interrupted Primitives}.

@item ENOTSOCK
The descriptor @var{socket} is not a socket.

@item EMSGSIZE
The socket type requires that the message be sent atomically, but the
message is too large for this to be possible.

@item EWOULDBLOCK
Nonblocking mode has been set on the socket, and the write operation
would block.  (Normally @code{send} blocks until the operation can be
completed.)

@item ENOBUFS
There is not enough internal buffer space available.

@item ENOTCONN
You never connected this socket.

@item EPIPE
This socket was connected but the connection is now broken.  In this
case, @code{send} generates a @code{SIGPIPE} signal first; if that
signal is ignored or blocked, or if its handler returns, then
@code{send} fails with @code{EPIPE}.
@end table

This function is defined as a cancelation point in multi-threaded
programs.  So one has to be prepared for this and make sure that
possibly allocated resources (like memory, files descriptors,
semaphores or whatever) are freed even if the thread is cancel.
@c @xref{pthread_cleanup_push}, for a method how to do this.
@end deftypefun

@node Receiving Data
@subsubsection Receiving Data

@pindex sys/socket.h
The @code{recv} function is declared in the header file
@file{sys/socket.h}.  If your @var{flags} argument is zero, you can
just as well use @code{read} instead of @code{recv}; see @ref{I/O
Primitives}.

@comment sys/socket.h
@comment BSD
@deftypefun int recv (int @var{socket}, void *@var{buffer}, size_t @var{size}, int @var{flags})
The @code{recv} function is like @code{read}, but with the additional
flags @var{flags}.  The possible values of @var{flags} are described
In @ref{Socket Data Options}.

If nonblocking mode is set for @var{socket}, and no data is available to
be read, @code{recv} fails immediately rather than waiting.  @xref{File
Status Flags}, for information about nonblocking mode.

This function returns the number of bytes received, or @code{-1} on failure.
The following @code{errno} error conditions are defined for this function:

@table @code
@item EBADF
The @var{socket} argument is not a valid file descriptor.

@item ENOTSOCK
The descriptor @var{socket} is not a socket.

@item EWOULDBLOCK
Nonblocking mode has been set on the socket, and the read operation
would block.  (Normally, @code{recv} blocks until there is input
available to be read.)

@item EINTR
The operation was interrupted by a signal before any data was read.
@xref{Interrupted Primitives}.

@item ENOTCONN
You never connected this socket.
@end table

This function is defined as a cancelation point in multi-threaded
programs.  So one has to be prepared for this and make sure that
possibly allocated resources (like memory, files descriptors,
semaphores or whatever) are freed even if the thread is cancel.
@c @xref{pthread_cleanup_push}, for a method how to do this.
@end deftypefun

@node Socket Data Options
@subsubsection Socket Data Options

@pindex sys/socket.h
The @var{flags} argument to @code{send} and @code{recv} is a bit
mask.  You can bitwise-OR the values of the following macros together
to obtain a value for this argument.  All are defined in the header
file @file{sys/socket.h}.

@comment sys/socket.h
@comment BSD
@deftypevr Macro int MSG_OOB
Send or receive out-of-band data.  @xref{Out-of-Band Data}.
@end deftypevr

@comment sys/socket.h
@comment BSD
@deftypevr Macro int MSG_PEEK
Look at the data but don't remove it from the input queue.  This is
only meaningful with input functions such as @code{recv}, not with
@code{send}.
@end deftypevr

@comment sys/socket.h
@comment BSD
@deftypevr Macro int MSG_DONTROUTE
Don't include routing information in the message.  This is only
meaningful with output operations, and is usually only of interest for
diagnostic or routing programs.  We don't try to explain it here.
@end deftypevr

@node Byte Stream Example
@subsection Byte Stream Socket Example

Here is an example client program that makes a connection for a byte
stream socket in the Internet namespace.  It doesn't do anything
particularly interesting once it has connected to the server; it just
sends a text string to the server and exits.

@smallexample
@include inetcli.c.texi
@end smallexample

@node Server Example
@subsection Byte Stream Connection Server Example

The server end is much more complicated.  Since we want to allow
multiple clients to be connected to the server at the same time, it
would be incorrect to wait for input from a single client by simply
calling @code{read} or @code{recv}.  Instead, the right thing to do is
to use @code{select} (@pxref{Waiting for I/O}) to wait for input on
all of the open sockets.  This also allows the server to deal with
additional connection requests.

This particular server doesn't do anything interesting once it has
gotten a message from a client.  It does close the socket for that
client when it detects an end-of-file condition (resulting from the
client shutting down its end of the connection).

This program uses @code{make_socket} and @code{init_sockaddr} to set
up the socket address; see @ref{Inet Example}.

@smallexample
@include inetsrv.c.texi
@end smallexample

@node Out-of-Band Data
@subsection Out-of-Band Data

@cindex out-of-band data
@cindex high-priority data
Streams with connections permit @dfn{out-of-band} data that is
delivered with higher priority than ordinary data.  Typically the
reason for sending out-of-band data is to send notice of an
exceptional condition.  The way to send out-of-band data is using
@code{send}, specifying the flag @code{MSG_OOB} (@pxref{Sending
Data}).

Out-of-band data is received with higher priority because the
receiving process need not read it in sequence; to read the next
available out-of-band data, use @code{recv} with the @code{MSG_OOB}
flag (@pxref{Receiving Data}).  Ordinary read operations do not read
out-of-band data; they read only the ordinary data.

@cindex urgent socket condition
When a socket finds that out-of-band data is on its way, it sends a
@code{SIGURG} signal to the owner process or process group of the
socket.  You can specify the owner using the @code{F_SETOWN} command
to the @code{fcntl} function; see @ref{Interrupt Input}.  You must
also establish a handler for this signal, as described in @ref{Signal
Handling}, in order to take appropriate action such as reading the
out-of-band data.

Alternatively, you can test for pending out-of-band data, or wait
until there is out-of-band data, using the @code{select} function; it
can wait for an exceptional condition on the socket.  @xref{Waiting
for I/O}, for more information about @code{select}.

Notification of out-of-band data (whether with @code{SIGURG} or with
@code{select}) indicates that out-of-band data is on the way; the data
may not actually arrive until later.  If you try to read the
out-of-band data before it arrives, @code{recv} fails with an
@code{EWOULDBLOCK} error.

Sending out-of-band data automatically places a ``mark'' in the stream
of ordinary data, showing where in the sequence the out-of-band data
``would have been''.  This is useful when the meaning of out-of-band
data is ``cancel everything sent so far''.  Here is how you can test,
in the receiving process, whether any ordinary data was sent before
the mark:

@smallexample
success = ioctl (socket, SIOCATMARK, &result);
@end smallexample

Here's a function to discard any ordinary data preceding the
out-of-band mark:

@smallexample
int
discard_until_mark (int socket)
@{
  while (1)
    @{
      /* @r{This is not an arbitrary limit; any size will do.}  */
      char buffer[1024];
      int result, success;

      /* @r{If we have reached the mark, return.}  */
      success = ioctl (socket, SIOCATMARK, &result);
      if (success < 0)
        perror ("ioctl");
      if (result)
        return;

      /* @r{Otherwise, read a bunch of ordinary data and discard it.}
         @r{This is guaranteed not to read past the mark}
         @r{if it starts before the mark.}  */
      success = read (socket, buffer, sizeof buffer);
      if (success < 0)
        perror ("read");
    @}
@}
@end smallexample

If you don't want to discard the ordinary data preceding the mark, you
may need to read some of it anyway, to make room in internal system
buffers for the out-of-band data.  If you try to read out-of-band data
and get an @code{EWOULDBLOCK} error, try reading some ordinary data
(saving it so that you can use it when you want it) and see if that
makes room.  Here is an example:

@smallexample
struct buffer
@{
  char *buffer;
  int size;
  struct buffer *next;
@};

/* @r{Read the out-of-band data from SOCKET and return it}
   @r{as a `struct buffer', which records the address of the data}
   @r{and its size.}

   @r{It may be necessary to read some ordinary data}
   @r{in order to make room for the out-of-band data.}
   @r{If so, the ordinary data is saved as a chain of buffers}
   @r{found in the `next' field of the value.}  */

struct buffer *
read_oob (int socket)
@{
  struct buffer *tail = 0;
  struct buffer *list = 0;

  while (1)
    @{
      /* @r{This is an arbitrary limit.}
         @r{Does anyone know how to do this without a limit?}  */
      char *buffer = (char *) xmalloc (1024);
      struct buffer *link;
      int success;
      int result;

      /* @r{Try again to read the out-of-band data.}  */
      success = recv (socket, buffer, sizeof buffer, MSG_OOB);
      if (success >= 0)
        @{
          /* @r{We got it, so return it.}  */
          struct buffer *link
            = (struct buffer *) xmalloc (sizeof (struct buffer));
          link->buffer = buffer;
          link->size = success;
          link->next = list;
          return link;
        @}

      /* @r{If we fail, see if we are at the mark.}  */
      success = ioctl (socket, SIOCATMARK, &result);
      if (success < 0)
        perror ("ioctl");
      if (result)
        @{
          /* @r{At the mark; skipping past more ordinary data cannot help.}
             @r{So just wait a while.}  */
          sleep (1);
          continue;
        @}

      /* @r{Otherwise, read a bunch of ordinary data and save it.}
         @r{This is guaranteed not to read past the mark}
         @r{if it starts before the mark.}  */
      success = read (socket, buffer, sizeof buffer);
      if (success < 0)
        perror ("read");

      /* @r{Save this data in the buffer list.}  */
      @{
        struct buffer *link
          = (struct buffer *) xmalloc (sizeof (struct buffer));
        link->buffer = buffer;
        link->size = success;

        /* @r{Add the new link to the end of the list.}  */
        if (tail)
          tail->next = link;
        else
          list = link;
        tail = link;
      @}
    @}
@}
@end smallexample

@node Datagrams
@section Datagram Socket Operations

@cindex datagram socket
This section describes how to use communication styles that don't use
connections (styles @code{SOCK_DGRAM} and @code{SOCK_RDM}).  Using
these styles, you group data into packets and each packet is an
independent communication.  You specify the destination for each
packet individually.

Datagram packets are like letters: you send each one independently,
with its own destination address, and they may arrive in the wrong
order or not at all.

The @code{listen} and @code{accept} functions are not allowed for
sockets using connectionless communication styles.

@menu
* Sending Datagrams::    Sending packets on a datagram socket.
* Receiving Datagrams::  Receiving packets on a datagram socket.
* Datagram Example::     An example program: packets sent over a
                           datagram socket in the file namespace.
* Example Receiver::     Another program, that receives those packets.
@end menu

@node Sending Datagrams
@subsection Sending Datagrams
@cindex sending a datagram
@cindex transmitting datagrams
@cindex datagrams, transmitting

@pindex sys/socket.h
The normal way of sending data on a datagram socket is by using the
@code{sendto} function, declared in @file{sys/socket.h}.

You can call @code{connect} on a datagram socket, but this only
specifies a default destination for further data transmission on the
socket.  When a socket has a default destination, then you can use
@code{send} (@pxref{Sending Data}) or even @code{write} (@pxref{I/O
Primitives}) to send a packet there.  You can cancel the default
destination by calling @code{connect} using an address format of
@code{AF_UNSPEC} in the @var{addr} argument.  @xref{Connecting}, for
more information about the @code{connect} function.

@comment sys/socket.h
@comment BSD
@deftypefun int sendto (int @var{socket}, void *@var{buffer}. size_t @var{size}, int @var{flags}, struct sockaddr *@var{addr}, socklen_t @var{length})
The @code{sendto} function transmits the data in the @var{buffer}
through the socket @var{socket} to the destination address specified
by the @var{addr} and @var{length} arguments.  The @var{size} argument
specifies the number of bytes to be transmitted.

The @var{flags} are interpreted the same way as for @code{send}; see
@ref{Socket Data Options}.

The return value and error conditions are also the same as for
@code{send}, but you cannot rely on the system to detect errors and
report them; the most common error is that the packet is lost or there
is no one at the specified address to receive it, and the operating
system on your machine usually does not know this.

It is also possible for one call to @code{sendto} to report an error
due to a problem related to a previous call.

This function is defined as a cancelation point in multi-threaded
programs.  So one has to be prepared for this and make sure that
possibly allocated resources (like memory, files descriptors,
semaphores or whatever) are freed even if the thread is cancel.
@c @xref{pthread_cleanup_push}, for a method how to do this.
@end deftypefun

@node Receiving Datagrams
@subsection Receiving Datagrams
@cindex receiving datagrams

The @code{recvfrom} function reads a packet from a datagram socket and
also tells you where it was sent from.  This function is declared in
@file{sys/socket.h}.

@comment sys/socket.h
@comment BSD
@deftypefun int recvfrom (int @var{socket}, void *@var{buffer}, size_t @var{size}, int @var{flags}, struct sockaddr *@var{addr}, socklen_t *@var{length-ptr})
The @code{recvfrom} function reads one packet from the socket
@var{socket} into the buffer @var{buffer}.  The @var{size} argument
specifies the maximum number of bytes to be read.

If the packet is longer than @var{size} bytes, then you get the first
@var{size} bytes of the packet, and the rest of the packet is lost.
There's no way to read the rest of the packet.  Thus, when you use a
packet protocol, you must always know how long a packet to expect.

The @var{addr} and @var{length-ptr} arguments are used to return the
address where the packet came from.  @xref{Socket Addresses}.  For a
socket in the file domain, the address information won't be meaningful,
since you can't read the address of such a socket (@pxref{File
Namespace}).  You can specify a null pointer as the @var{addr} argument
if you are not interested in this information.

The @var{flags} are interpreted the same way as for @code{recv}
(@pxref{Socket Data Options}).  The return value and error conditions
are also the same as for @code{recv}.

This function is defined as a cancelation point in multi-threaded
programs.  So one has to be prepared for this and make sure that
possibly allocated resources (like memory, files descriptors,
semaphores or whatever) are freed even if the thread is cancel.
@c @xref{pthread_cleanup_push}, for a method how to do this.
@end deftypefun

You can use plain @code{recv} (@pxref{Receiving Data}) instead of
@code{recvfrom} if you know don't need to find out who sent the packet
(either because you know where it should come from or because you
treat all possible senders alike).  Even @code{read} can be used if
you don't want to specify @var{flags} (@pxref{I/O Primitives}).

@ignore
@c sendmsg and recvmsg are like readv and writev in that they
@c use a series of buffers.  It's not clear this is worth
@c supporting or that we support them.
@c !!! they can do more; it is hairy

@comment sys/socket.h
@comment BSD
@deftp {Data Type} {struct msghdr}
@end deftp

@comment sys/socket.h
@comment BSD
@deftypefun int sendmsg (int @var{socket}, const struct msghdr *@var{message}, int @var{flags})

This function is defined as a cancelation point in multi-threaded
programs.  So one has to be prepared for this and make sure that
possibly allocated resources (like memory, files descriptors,
semaphores or whatever) are freed even if the thread is cancel.
@c @xref{pthread_cleanup_push}, for a method how to do this.
@end deftypefun

@comment sys/socket.h
@comment BSD
@deftypefun int recvmsg (int @var{socket}, struct msghdr *@var{message}, int @var{flags})

This function is defined as a cancelation point in multi-threaded
programs.  So one has to be prepared for this and make sure that
possibly allocated resources (like memory, files descriptors,
semaphores or whatever) are freed even if the thread is cancel.
@c @xref{pthread_cleanup_push}, for a method how to do this.
@end deftypefun
@end ignore

@node Datagram Example
@subsection Datagram Socket Example

Here is a set of example programs that send messages over a datagram
stream in the file namespace.  Both the client and server programs use the
@code{make_named_socket} function that was presented in @ref{File
Namespace}, to create and name their sockets.

First, here is the server program.  It sits in a loop waiting for
messages to arrive, bouncing each message back to the sender.
Obviously, this isn't a particularly useful program, but it does show
the general ideas involved.

@smallexample
@include filesrv.c.texi
@end smallexample

@node Example Receiver
@subsection Example of Reading Datagrams

Here is the client program corresponding to the server above.

It sends a datagram to the server and then waits for a reply.  Notice
that the socket for the client (as well as for the server) in this
example has to be given a name.  This is so that the server can direct
a message back to the client.  Since the socket has no associated
connection state, the only way the server can do this is by
referencing the name of the client.

@smallexample
@include filecli.c.texi
@end smallexample

Keep in mind that datagram socket communications are unreliable.  In
this example, the client program waits indefinitely if the message
never reaches the server or if the server's response never comes
back.  It's up to the user running the program to kill it and restart
it, if desired.  A more automatic solution could be to use
@code{select} (@pxref{Waiting for I/O}) to establish a timeout period
for the reply, and in case of timeout either resend the message or
shut down the socket and exit.

@node Inetd
@section The @code{inetd} Daemon

We've explained above how to write a server program that does its own
listening.  Such a server must already be running in order for anyone
to connect to it.

Another way to provide service for an Internet port is to let the daemon
program @code{inetd} do the listening.  @code{inetd} is a program that
runs all the time and waits (using @code{select}) for messages on a
specified set of ports.  When it receives a message, it accepts the
connection (if the socket style calls for connections) and then forks a
child process to run the corresponding server program.  You specify the
ports and their programs in the file @file{/etc/inetd.conf}.

@menu
* Inetd Servers::
* Configuring Inetd::
@end menu

@node Inetd Servers
@subsection @code{inetd} Servers

Writing a server program to be run by @code{inetd} is very simple.  Each time
someone requests a connection to the appropriate port, a new server
process starts.  The connection already exists at this time; the
socket is available as the standard input descriptor and as the
standard output descriptor (descriptors 0 and 1) in the server
process.  So the server program can begin reading and writing data
right away.  Often the program needs only the ordinary I/O facilities;
in fact, a general-purpose filter program that knows nothing about
sockets can work as a byte stream server run by @code{inetd}.

You can also use @code{inetd} for servers that use connectionless
communication styles.  For these servers, @code{inetd} does not try to accept
a connection, since no connection is possible.  It just starts the
server program, which can read the incoming datagram packet from
descriptor 0.  The server program can handle one request and then
exit, or you can choose to write it to keep reading more requests
until no more arrive, and then exit.  You must specify which of these
two techniques the server uses, when you configure @code{inetd}.

@node Configuring Inetd
@subsection Configuring @code{inetd}

The file @file{/etc/inetd.conf} tells @code{inetd} which ports to listen to
and what server programs to run for them.  Normally each entry in the
file is one line, but you can split it onto multiple lines provided
all but the first line of the entry start with whitespace.  Lines that
start with @samp{#} are comments.

Here are two standard entries in @file{/etc/inetd.conf}:

@smallexample
ftp     stream  tcp     nowait  root    /libexec/ftpd   ftpd
talk    dgram   udp     wait    root    /libexec/talkd  talkd
@end smallexample

An entry has this format:

@smallexample
@var{service} @var{style} @var{protocol} @var{wait} @var{username} @var{program} @var{arguments}
@end smallexample

The @var{service} field says which service this program provides.  It
should be the name of a service defined in @file{/etc/services}.
@code{inetd} uses @var{service} to decide which port to listen on for
this entry.

The fields @var{style} and @var{protocol} specify the communication
style and the protocol to use for the listening socket.  The style
should be the name of a communication style, converted to lower case
and with @samp{SOCK_} deleted---for example, @samp{stream} or
@samp{dgram}.  @var{protocol} should be one of the protocols listed in
@file{/etc/protocols}.  The typical protocol names are @samp{tcp} for
byte stream connections and @samp{udp} for unreliable datagrams.

The @var{wait} field should be either @samp{wait} or @samp{nowait}.
Use @samp{wait} if @var{style} is a connectionless style and the
server, once started, handles multiple requests, as many as come in.
Use @samp{nowait} if @code{inetd} should start a new process for each message
or request that comes in.  If @var{style} uses connections, then
@var{wait} @strong{must} be @samp{nowait}.

@var{user} is the user name that the server should run as.  @code{inetd} runs
as root, so it can set the user ID of its children arbitrarily.  It's
best to avoid using @samp{root} for @var{user} if you can; but some
servers, such as Telnet and FTP, read a username and password
themselves.  These servers need to be root initially so they can log
in as commanded by the data coming over the network.

@var{program} together with @var{arguments} specifies the command to
run to start the server.  @var{program} should be an absolute file
name specifying the executable file to run.  @var{arguments} consists
of any number of whitespace-separated words, which become the
command-line arguments of @var{program}.  The first word in
@var{arguments} is argument zero, which should by convention be the
program name itself (sans directories).

If you edit @file{/etc/inetd.conf}, you can tell @code{inetd} to reread the
file and obey its new contents by sending the @code{inetd} process the
@code{SIGHUP} signal.  You'll have to use @code{ps} to determine the
process ID of the @code{inetd} process, as it is not fixed.

@c !!! could document /etc/inetd.sec

@node Socket Options
@section Socket Options
@cindex socket options

This section describes how to read or set various options that modify
the behavior of sockets and their underlying communications protocols.

@cindex level, for socket options
@cindex socket option level
When you are manipulating a socket option, you must specify which
@dfn{level} the option pertains to.  This describes whether the option
applies to the socket interface, or to a lower-level communications
protocol interface.

@menu
* Socket Option Functions::     The basic functions for setting and getting
                                 socket options.
* Socket-Level Options::        Details of the options at the socket level.
@end menu

@node Socket Option Functions
@subsection Socket Option Functions

@pindex sys/socket.h
Here are the functions for examining and modifying socket options.
They are declared in @file{sys/socket.h}.

@comment sys/socket.h
@comment BSD
@deftypefun int getsockopt (int @var{socket}, int @var{level}, int @var{optname}, void *@var{optval}, socklen_t *@var{optlen-ptr})
The @code{getsockopt} function gets information about the value of
option @var{optname} at level @var{level} for socket @var{socket}.

The option value is stored in a buffer that @var{optval} points to.
Before the call, you should supply in @code{*@var{optlen-ptr}} the
size of this buffer; on return, it contains the number of bytes of
information actually stored in the buffer.

Most options interpret the @var{optval} buffer as a single @code{int}
value.

The actual return value of @code{getsockopt} is @code{0} on success
and @code{-1} on failure.  The following @code{errno} error conditions
are defined:

@table @code
@item EBADF
The @var{socket} argument is not a valid file descriptor.

@item ENOTSOCK
The descriptor @var{socket} is not a socket.

@item ENOPROTOOPT
The @var{optname} doesn't make sense for the given @var{level}.
@end table
@end deftypefun

@comment sys/socket.h
@comment BSD
@deftypefun int setsockopt (int @var{socket}, int @var{level}, int @var{optname}, void *@var{optval}, socklen_t @var{optlen})
This function is used to set the socket option @var{optname} at level
@var{level} for socket @var{socket}.  The value of the option is passed
in the buffer @var{optval}, which has size @var{optlen}.

@c Argh. -zw
@iftex
@hfuzz 6pt
The return value and error codes for @code{setsockopt} are the same as
for @code{getsockopt}.
@end iftex
@ifinfo
The return value and error codes for @code{setsockopt} are the same as
for @code{getsockopt}.
@end ifinfo

@end deftypefun

@node Socket-Level Options
@subsection Socket-Level Options

@comment sys/socket.h
@comment BSD
@deftypevr Constant int SOL_SOCKET
Use this constant as the @var{level} argument to @code{getsockopt} or
@code{setsockopt} to manipulate the socket-level options described in
this section.
@end deftypevr

@pindex sys/socket.h
@noindent
Here is a table of socket-level option names; all are defined in the
header file@*@file{sys/socket.h}.

@table @code
@comment sys/socket.h
@comment BSD
@item SO_DEBUG
@c Extra blank line here makes the table look better.

This option toggles recording of debugging information in the underlying
protocol modules.  The value has type @code{int}; a nonzero value means
``yes''.
@c !!! should say how this is used
@c Ok, anyone who knows, please explain.

@comment sys/socket.h
@comment BSD
@item SO_REUSEADDR
This option controls whether @code{bind} (@pxref{Setting Address})
should permit reuse of local addresses for this socket.  If you enable
this option, you can actually have two sockets with the same Internet
port number; but the system won't allow you to use the two
identically-named sockets in a way that would confuse the Internet.  The
reason for this option is that some higher-level Internet protocols,
including FTP, require you to keep reusing the same socket number.

The value has type @code{int}; a nonzero value means ``yes''.

@comment sys/socket.h
@comment BSD
@item SO_KEEPALIVE
This option controls whether the underlying protocol should
periodically transmit messages on a connected socket.  If the peer
fails to respond to these messages, the connection is considered
broken.  The value has type @code{int}; a nonzero value means
``yes''.

@comment sys/socket.h
@comment BSD
@item SO_DONTROUTE
This option controls whether outgoing messages bypass the normal
message routing facilities.  If set, messages are sent directly to the
network interface instead.  The value has type @code{int}; a nonzero
value means ``yes''.

@comment sys/socket.h
@comment BSD
@item SO_LINGER
This option specifies what should happen when the socket of a type
that promises reliable delivery still has untransmitted messages when
it is closed; see @ref{Closing a Socket}.  The value has type
@code{struct linger}.

@comment sys/socket.h
@comment BSD
@deftp {Data Type} {struct linger}
This structure type has the following members:

@table @code
@item int l_onoff
This field is interpreted as a boolean.  If nonzero, @code{close}
blocks until the data is transmitted or the timeout period has expired.

@item int l_linger
This specifies the timeout period, in seconds.
@end table
@end deftp

@comment sys/socket.h
@comment BSD
@item SO_BROADCAST
This option controls whether datagrams may be broadcast from the socket.
The value has type @code{int}; a nonzero value means ``yes''.

@comment sys/socket.h
@comment BSD
@item SO_OOBINLINE
If this option is set, out-of-band data received on the socket is
placed in the normal input queue.  This permits it to be read using
@code{read} or @code{recv} without specifying the @code{MSG_OOB}
flag.  @xref{Out-of-Band Data}.  The value has type @code{int}; a
nonzero value means ``yes''.

@comment sys/socket.h
@comment BSD
@item SO_SNDBUF
This option gets or sets the size of the output buffer.  The value is a
@code{size_t}, which is the size in bytes.

@comment sys/socket.h
@comment BSD
@item SO_RCVBUF
This option gets or sets the size of the input buffer.  The value is a
@code{size_t}, which is the size in bytes.

@comment sys/socket.h
@comment GNU
@item SO_STYLE
@comment sys/socket.h
@comment BSD
@itemx SO_TYPE
This option can be used with @code{getsockopt} only.  It is used to
get the socket's communication style.  @code{SO_TYPE} is the
historical name, and @code{SO_STYLE} is the preferred name in GNU.
The value has type @code{int} and its value designates a communication
style; see @ref{Communication Styles}.

@comment sys/socket.h
@comment BSD
@item SO_ERROR
@c Extra blank line here makes the table look better.

This option can be used with @code{getsockopt} only.  It is used to reset
the error status of the socket.  The value is an @code{int}, which represents
the previous error status.
@c !!! what is "socket error status"?  this is never defined.
@end table

@node Networks Database
@section Networks Database
@cindex networks database
@cindex converting network number to network name
@cindex converting network name to network number

@pindex /etc/networks
@pindex netdb.h
Many systems come with a database that records a list of networks known
to the system developer.  This is usually kept either in the file
@file{/etc/networks} or in an equivalent from a name server.  This data
base is useful for routing programs such as @code{route}, but it is not
useful for programs that simply communicate over the network.  We
provide functions to access this data base, which are declared in
@file{netdb.h}.

@comment netdb.h
@comment BSD
@deftp {Data Type} {struct netent}
This data type is used to represent information about entries in the
networks database.  It has the following members:

@table @code
@item char *n_name
This is the ``official'' name of the network.

@item char **n_aliases
These are alternative names for the network, represented as a vector
of strings.  A null pointer terminates the array.

@item int n_addrtype
This is the type of the network number; this is always equal to
@code{AF_INET} for Internet networks.

@item unsigned long int n_net
This is the network number.  Network numbers are returned in host
byte order; see @ref{Byte Order}.
@end table
@end deftp

Use the @code{getnetbyname} or @code{getnetbyaddr} functions to search
the networks database for information about a specific network.  The
information is returned in a statically-allocated structure; you must
copy the information if you need to save it.

@comment netdb.h
@comment BSD
@deftypefun {struct netent *} getnetbyname (const char *@var{name})
The @code{getnetbyname} function returns information about the network
named @var{name}.  It returns a null pointer if there is no such
network.
@end deftypefun

@comment netdb.h
@comment BSD
@deftypefun {struct netent *} getnetbyaddr (long @var{net}, int @var{type})
The @code{getnetbyaddr} function returns information about the network
of type @var{type} with number @var{net}.  You should specify a value of
@code{AF_INET} for the @var{type} argument for Internet networks.

@code{getnetbyaddr} returns a null pointer if there is no such
network.
@end deftypefun

You can also scan the networks database using @code{setnetent},
@code{getnetent}, and @code{endnetent}.  Be careful in using these
functions, because they are not reentrant.

@comment netdb.h
@comment BSD
@deftypefun void setnetent (int @var{stayopen})
This function opens and rewinds the networks database.

If the @var{stayopen} argument is nonzero, this sets a flag so that
subsequent calls to @code{getnetbyname} or @code{getnetbyaddr} will
not close the database (as they usually would).  This makes for more
efficiency if you call those functions several times, by avoiding
reopening the database for each call.
@end deftypefun

@comment netdb.h
@comment BSD
@deftypefun {struct netent *} getnetent (void)
This function returns the next entry in the networks database.  It
returns a null pointer if there are no more entries.
@end deftypefun

@comment netdb.h
@comment BSD
@deftypefun void endnetent (void)
This function closes the networks database.
@end deftypefun