Skip site navigation (1)Skip section navigation (2)

FreeBSD Manual Pages


home | help
INET(4)			 BSD Kernel Interfaces Manual		       INET(4)

     inet -- Internet protocol family

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

     The Internet protocol family is a collection of protocols layered atop
     the Internet Protocol (IP)	transport layer, and utilizing the Internet
     address format.  The Internet family provides protocol support for	the
     SOCK_STREAM, SOCK_DGRAM, and SOCK_RAW socket types; the SOCK_RAW inter-
     face provides access to the IP protocol.

     Internet addresses	are four byte quantities, stored in network standard
     format (on	little endian machines,	such as	the alpha, amd64, i386 and
     ia64 these	are word and byte reversed).  The include file <netinet/in.h>
     defines this address as a discriminated union.

     Sockets bound to the Internet protocol family utilize the following ad-
     dressing structure,

	   struct sockaddr_in {
		   uint8_t	   sin_len;
		   sa_family_t	   sin_family;
		   in_port_t	   sin_port;
		   struct in_addr  sin_addr;
		   char		   sin_zero[8];

     Sockets may be created with the local address INADDR_ANY to affect
     "wildcard"	matching on incoming messages.	The address in a connect(2) or
     sendto(2) call may	be given as INADDR_ANY to mean "this host".  The dis-
     tinguished	address	INADDR_BROADCAST is allowed as a shorthand for the
     broadcast address on the primary network if the first network configured
     supports broadcast.

     The Internet protocol family is comprised of the IP network protocol, In-
     ternet Control Message Protocol (ICMP), Internet Group Management Proto-
     col (IGMP), Transmission Control Protocol (TCP), and User Datagram	Proto-
     col (UDP).	 TCP is	used to	support	the SOCK_STREAM	abstraction while UDP
     is	used to	support	the SOCK_DGRAM abstraction.  A raw interface to	IP is
     available by creating an Internet socket of type SOCK_RAW.	 The ICMP mes-
     sage protocol is accessible from a	raw socket.

     The 32-bit	Internet address contains both network and host	parts.	How-
     ever, direct examination of addresses is discouraged.  For	those programs
     which absolutely need to break addresses into their component parts, the
     following ioctl(2)	commands are provided for a datagram socket in the In-
     ternet domain; they have the same form as the SIOCIFADDR command (see

     SIOCSIFNETMASK  Set interface network mask.  The network mask defines the
		     network part of the address; if it	contains more of the
		     address than the address type would indicate, then	sub-
		     nets are in use.

     SIOCGIFNETMASK  Get interface network mask.

     The current implementation	of Internet protocols includes some routing-
     table adaptations to provide enhanced caching of certain end-to-end in-
     formation necessary for Transaction TCP and Path MTU Discovery.  The fol-
     lowing changes are	the most significant:

     1.	  All IP routes, except	those with the RTF_CLONING flag	and those to
	  multicast destinations, have the RTF_PRCLONING flag forcibly enabled
	  (they	are thus said to be "protocol cloning").

     2.	  When the last	reference to an	IP route is dropped, the route is ex-
	  amined to determine if it was	created	by cloning such	a route.  If
	  this is the case, the	RTF_PROTO3 flag	is turned on, and the expira-
	  tion timer is	initialized to go off in net.inet.ip.rtexpire seconds.
	  If such a route is re-referenced, the	flag and expiration timer are

     3.	  A kernel timeout runs	once every ten minutes,	or sooner if there are
	  soon-to-expire routes	in the kernel routing table, and deletes the
	  expired routes.

     A dynamic process is in place to modify the value of net.inet.ip.rtexpire
     if	the number of cached routes grows too large.  If after an expiration
     run there are still more than net.inet.ip.rtmaxcache unreferenced routes
     remaining,	the rtexpire value is multiplied by 3/4, and any routes	which
     have longer expiration times have those times adjusted.  This process is
     damped somewhat by	specification of a minimum rtexpire value
     (net.inet.ip.rtminexpire),	and by restricting the reduction to once in a
     ten-minute	period.

     If	some external process deletes the original route from which a proto-
     col-cloned	route was generated, the "child	route" is deleted.  (This is
     actually a	generic	mechanism in the routing code support for protocol-re-
     quested cloning.)

     No	attempt	is made	to manage routes which were not	created	by protocol
     cloning; these are	assumed	to be static, under the	management of an ex-
     ternal routing process, or	under the management of	a link layer (e.g.,
     ARP for Ethernets).

     Only certain types	of network activity will result	in the cloning of a
     route using this mechanism.  Specifically,	those protocols	(such as TCP
     and UDP) which themselves cache a long-lasting reference to route for a
     destination will trigger the mechanism; whereas raw IP packets, whether
     locally-generated or forwarded, will not.

   MIB Variables
     A number of variables are implemented in the net.inet branch of the
     sysctl(3) MIB.  In	addition to the	variables supported by the transport
     protocols (for which the respective manual	pages may be consulted), the
     following general variables are defined:

     IPCTL_FORWARDING	   (ip.forwarding) Boolean: enable/disable forwarding
			   of IP packets.  Defaults to off.

     IPCTL_FASTFORWARDING  (ip.fastforwarding) Boolean:	enable/disable the use
			   of fast IP forwarding code.	Defaults to off.  When
			   fast	forwarding is enabled, IP packets are for-
			   warded directly to the appropriate network inter-
			   face	with a minimal validity	checking, which
			   greatly improves the	throughput.  On	the other
			   hand, they bypass the standard procedures, such as
			   IP option processing	and ipfirewall(4) checking.
			   It is not guaranteed	that every packet will be

     IPCTL_SENDREDIRECTS   (ip.redirect) Boolean: enable/disable sending of
			   ICMP	redirects in response to unforwardable IP
			   packets.  Defaults to on.

     IPCTL_DEFTTL	   (ip.ttl) Integer: default time-to-live ("TTL") to
			   use for outgoing IP packets.

			   (ip.accept_sourceroute) Boolean: enable/disable ac-
			   cepting of source-routed IP packets (default

     IPCTL_SOURCEROUTE	   (ip.sourceroute) Boolean: enable/disable forwarding
			   of source-routed IP packets (default	false).

     IPCTL_RTEXPIRE	   (ip.rtexpire) Integer: lifetime in seconds of pro-
			   tocol-cloned	IP routes after	the last reference
			   drops (default one hour).  This value varies	dynam-
			   ically as described above.

     IPCTL_RTMINEXPIRE	   (ip.rtminexpire) Integer: minimum value of ip.rtex-
			   pire	(default ten seconds).	This value has no ef-
			   fect	on user	modifications, but restricts the dy-
			   namic adaptation described above.

     IPCTL_RTMAXCACHE	   (ip.rtmaxcache) Integer: trigger level of cached,
			   unreferenced, protocol-cloned routes	which initi-
			   ates	dynamic	adaptation (default 128).

     ioctl(2), socket(2), sysctl(3), icmp(4), intro(4),	ip(4), ipfirewall(4),
     tcp(4), ttcp(4), udp(4)

     "An Introductory 4.3 BSD Interprocess Communication Tutorial", PS1, 7.

     "An Advanced 4.3 BSD Interprocess Communication Tutorial",	PS1, 8.

     The Internet protocol support is subject to change	as the Internet	proto-
     cols develop.  Users should not depend on details of the current imple-
     mentation,	but rather the services	exported.

     The inet protocol interface appeared in 4.2BSD.  The "protocol cloning"
     code appeared in FreeBSD 2.1.

BSD			       February	14, 1995			   BSD


Want to link to this manual page? Use this URL:

home | help