:man| Alphabetical   Categories   About us 
IPV6 (7) | Miscellanea | Unix Manual Pages | :man


ipv6, PF_INET6 - Linux IPv6 protocol implementation




#include <sys/socket.h>
#include <netinet/in.h>

tcp6_socket" = socket(PF_INET6, SOCK_STREAM, 0);"
raw6_socket" = socket(PF_INET6, SOCK_RAW, "protocol");"
udp6_socket" = socket(PF_INET6, SOCK_DGRAM, "protocol");"


Linux 2.2 optionally implements the Internet Protocol, version 6. This man page contains a description of the IPv6 basic API as implemented by the Linux kernel and glibc 2.1. The interface is based on the BSD sockets interface; see socket(7).

The IPv6 API aims to be mostly compatible with the ip(7) v4 API. Only differences are described in this man page.

To bind an AF_INET6 socket to any process the local address should be copied from the in6addr_any variable which has in6_addr type. In static initializations IN6ADDR_ANY_INIT may also be used, which expands to a constant expression. Both of them are in network order.

The IPv6 loopback address (::1) is available in the global in6addr_loopback variable. For initializations IN6ADDR_LOOPBACK_INIT should be used.

IPv4 connections can be handled with the v6 API by using the v4-mapped-on-v6 address type; thus a program only needs only to support this API type to support both protocols. This is handled transparently by the address handling functions in libc.

IPv4 and IPv6 share the local port space. When you get an IPv4 connection or packet to a IPv6 socket its source address will be mapped to v6 and it’ll be mapped to v6.


struct sockaddr_in6 {
u_int16_t sin6_family;/* AF_INET6 */
u_int16_t sin6_port;/* port number */
u_int32_t sin6_flowinfo;/* IPv6 flow information */
struct in6_addr sin6_addr;/* IPv6 address */
u_int32_t sin6_scope_id; /* Scope id (new in 2.4) */

struct in6_addr {
unsigned char s6_addr[16];/* IPv6 address */

sin6_family is always set to AF_INET6; sin6_port is the protocol port (see sin_port in ip(7)); sin6_flowinfo is the IPv6 flow identifier; sin6_addr is the 128bit IPv6 address. sin6_scope_id is an id of depending of on the scope of the address. It is new in Linux 2.4. Linux only supports it for link scope addresses, in that case sin6_scope_id contains the interface index (see netdevice(7))

IPv6 supports several address types: unicast to address a single host, multicast to address a group of hosts, anycast to address the nearest member of a group of hosts (not implemented in Linux), IPv4-on-IPv6 to address a IPv4 host, and other reserved address types.

The address notation for IPv6 is a group of 16 2 digit hexadecimal numbers, separated with a ’:’. ’::’ stands for a string of 0 bits. Special addresses are ::1 for loopback and ::FFFF:<IPv4 address> for IPv4-mapped-on-IPv6.

The port space of IPv6 is shared with IPv4.


IPv6 supports some protocol specific socket options that can be set with setsockopt(2) and read with getsockopt(2). The socket option level for IPv6 is IPPROTO_IPV6. A boolean integer flag is zero when it is false, otherwise true. -->
Set the unicast hop limit for the socket. Argument is an pointer to an integer. -1 in the value means use the route default, otherwise it should be between 0 and 255.
Set the multicast hop limit for the socket. Argument is a pointer to an integer. -1 in the value means use the route default, otherwise it should be between 0 and 255.
Set the device for outgoing multicast packets on the socket. This is only allowed for SOCK_DGRAM and SOCK_RAW socket. The argument is an pointer to an interface index (see netdevice(7)) in an integer.
Turn an AF_INET6 socket into a socket of a different address family. Only AF_INET is currently supported for that. It is only allowed for IPv6 sockets that are connected and bound to a v4-mapped-on-v6 address. The argument is a pointer to a integer containing AF_INET. This is useful to pass v4-mapped sockets as filedescriptors to programs that don’t know how to deal with the IPv6 API.
Set delivery of the IPV6_PKTINFO control message on incoming datagrams. Only allowed for SOCK_DGRAM or SOCK_RAW sockets. Argument is a pointer to a boolean value in an integer.
IPV6_RTHDR, IPV6_AUTHHDR, IPV6_DSTOPS, IPV6_HOPOPTS, IPV6_FLOWINFO, IPV6_HOPLIMIT Set delivery of control messages for incoming datagrams containing extension headers from the received packet. IPV6_RTHDR delivers the routing header, IPV6_AUTHHDR delivers the authentication header, IPV6_DSTOPTS delivers the destination options, IPV6_HOPOPTS delivers the hop options, IPV6_FLOWINFO delivers an integer containing the flow id, IPV6_HOPLIMIT delivers an integer containing the hop count of the packet. The control messages have the same type as the socket option. All these header options can also be set for outgoing packets by putting the appropriate control message into the control buffer of sendmsg(2). Only allowed for SOCK_DGRAM or SOCK_RAW sockets. Argument is a pointer to a boolean value.
Control whether the socket sees multicast packets that is has send itself. Argument is a pointer to boolean.
Control membership in multicast groups. Argument is a pointer to a struct ipv6_mreq structure.
Set the MTU to be used for the socket. The MTU is limited by the device MTU or the path mtu when path mtu discovery is enabled. Argument is a pointer to integer.
Control path mtu discovery on the socket. See IP_MTU_DISCOVER in ip(7) for details.
Control receiving of asynchronous error options. See IP_RECVERR in ip(7) for details. Argument is a pointer to boolean.
Pass all forwarded packets containing an router alert option to this socket. Only allowed for datagram sockets and for root. Argument is a pointer to boolean.


The older libinet6 libc5 based IPv6 API implementation for Linux is not described here and may vary in details.

Linux 2.4 will break binary compatibility for the sockaddr_in6 for 64bit hosts by changing the alignment of in6_addr and adding an additional sin6_scope_id field. The kernel interfaces stay compatible, but a program including sockaddr_in6 or in6_addr into other structures may not be. This is not a problem for 32bit hosts like i386.

The sin6_flowinfo field is new in Linux 2.4. It is transparently passed/read by the kernel when the passed address length contains it. Some programs that pass a longer address buffer and then check the outgoing address length may break.


The sockaddr_in6 structure is bigger than the generic sockaddr. Programs that assume that all address types can be stored safely in a struct sockaddr need to be changed to use struct sockaddr_storage for that instead.


The IPv6 extended API as in RFC2292 is currently only partly implemented; although the 2.2 kernel has near complete support for receiving options, the macros for generating IPv6 options are missing in glibc 2.1.

IPSec support for EH and AH headers is missing.

Flow label management is not complete and not documented here.

This man page is not complete.


ip(7), cmsg(3)

Share this page

     Follow us

Facebook Twitter Google+ LinkedIn

Created by Blin Media, 2008-2013