1.\" $OpenBSD: netintro.4,v 1.33 2005/06/08 04:17:43 henning Exp $ 2.\" $NetBSD: netintro.4,v 1.4 1995/10/19 08:03:40 jtc Exp $ 3.\" 4.\" Copyright (c) 1983, 1990, 1991, 1993 5.\" The Regents of the University of California. All rights reserved. 6.\" 7.\" Redistribution and use in source and binary forms, with or without 8.\" modification, are permitted provided that the following conditions 9.\" are met: 10.\" 1. Redistributions of source code must retain the above copyright 11.\" notice, this list of conditions and the following disclaimer. 12.\" 2. Redistributions in binary form must reproduce the above copyright 13.\" notice, this list of conditions and the following disclaimer in the 14.\" documentation and/or other materials provided with the distribution. 15.\" 3. Neither the name of the University nor the names of its contributors 16.\" may be used to endorse or promote products derived from this software 17.\" without specific prior written permission. 18.\" 19.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 20.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 23.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29.\" SUCH DAMAGE. 30.\" 31.\" @(#)netintro.4 8.2 (Berkeley) 11/30/93 32.\" 33.Dd September 3, 1994 34.Dt NETINTRO 4 35.Os 36.Sh NAME 37.Nm netintro 38.Nd introduction to networking facilities 39.Sh SYNOPSIS 40.Fd #include <sys/socket.h> 41.Fd #include <net/route.h> 42.Fd #include <net/if.h> 43.Sh DESCRIPTION 44This section is a general introduction to the networking facilities 45available in the system. 46Documentation in this part of section 474 is broken up into three areas: 48.Em protocol families 49(domains), 50.Em protocols , 51and 52.Em network interfaces . 53.Pp 54All network protocols are associated with a specific 55.Em protocol family . 56A protocol family provides basic services to the protocol 57implementation to allow it to function within a specific 58network environment. 59These services may include packet fragmentation and reassembly, routing, 60addressing, and basic transport. 61A protocol family may support multiple methods of addressing, though 62the current protocol implementations do not. 63A protocol family is normally comprised of a number of protocols, one per 64.Xr socket 2 65type. 66It is not required that a protocol family support all socket types. 67A protocol family may contain multiple protocols supporting the same socket 68abstraction. 69.Pp 70A protocol supports one of the socket abstractions detailed in 71.Xr socket 2 . 72A specific protocol may be accessed either by creating a 73socket of the appropriate type and protocol family, or 74by requesting the protocol explicitly when creating a socket. 75Protocols normally accept only one type of address format, 76usually determined by the addressing structure inherent in 77the design of the protocol family/network architecture. 78Certain semantics of the basic socket abstractions are 79protocol specific. 80All protocols are expected to support the basic model for their particular 81socket type, but may, in addition, provide non-standard facilities or 82extensions to a mechanism. 83For example, a protocol supporting the 84.Dv SOCK_STREAM 85abstraction may allow more than one byte of out-of-band 86data to be transmitted per out-of-band message. 87.Pp 88A network interface is similar to a device interface. 89Network interfaces comprise the lowest layer of the 90networking subsystem, interacting with the actual transport 91hardware. 92An interface may support one or more protocol families and/or address formats. 93The 94.Sx SYNOPSIS 95section of each network interface entry gives a sample 96specification of the related drivers for use in providing a system description 97to the 98.Xr config 8 99program. 100The 101.Sx DIAGNOSTICS 102section lists messages which may appear on the console 103and/or in the system error log, 104.Pa /var/log/messages 105(see 106.Xr syslogd 8 ) , 107due to errors in device operation. 108.Pp 109Network interfaces may be collected together into interface groups. 110An interface group is a container that can be used generically when 111referring to any interface related by some criteria. 112When an action is performed on an interface group, such as packet 113filtering by the 114.Xr pf 4 115subsystem, the operation will be applied to each member interface in the 116group, if supported by the subsystem. 117The 118.Xr ifconfig 8 119utility can be used to view and assign membership of an interface to an 120interface group with the 121.Cm group 122modifier. 123.Sh PROTOCOLS 124The system currently supports the 125Internet protocols (IPv4 and IPv6), 126Appletalk, Novell's IPX protocols, 127and a few others. 128Raw socket interfaces are provided to the 129.Tn IP 130protocol 131layer of the 132Internet. 133Consult the appropriate manual pages in this section for more 134information regarding the support for each protocol family. 135.Sh ADDRESSING 136Associated with each protocol family is an address 137format. 138All network addresses adhere to a general structure, called a 139.Vt sockaddr , 140described below. 141However, each protocol imposes a finer, more specific structure, generally 142renaming the variant, which is discussed in the protocol family manual 143page alluded to above. 144.Bd -literal -offset indent 145struct sockaddr { 146 u_int8_t sa_len; /* total length */ 147 sa_family_t sa_family; /* address family */ 148 char sa_data[14]; /* actually longer */ 149}; 150.Ed 151.Pp 152The field 153.Va sa_len 154contains the total length of the structure, 155which may exceed 16 bytes. 156The following address values for 157.Va sa_family 158are known to the system 159(and additional formats are defined for possible future implementation): 160.Bd -literal 161#define AF_LOCAL 1 /* local to host (pipes, portals) */ 162#define AF_INET 2 /* internetwork: UDP, TCP, etc. */ 163#define AF_HYLINK 15 /* NSC Hyperchannel */ 164#define AF_APPLETALK 16 /* AppleTalk */ 165#define AF_IPX 23 /* Novell Internet Protocol */ 166#define AF_INET6 24 /* IPv6 */ 167#define AF_NATM 27 /* native ATM access */ 168.Ed 169.Pp 170The 171.Va sa_data 172field contains the actual address value. 173Note that it may be longer than 14 bytes. 174.Sh ROUTING 175.Ox 176provides some packet routing facilities. 177The kernel maintains a routing information database, which 178is used in selecting the appropriate network interface when 179transmitting packets. 180.Pp 181A user process (or possibly multiple co-operating processes) 182maintains this database by sending messages over a special kind 183of socket. 184This supplants fixed-size 185.Xr ioctl 2 's 186used in earlier releases. 187.Pp 188This facility is described in 189.Xr route 4 . 190.Sh INTERFACES 191Each network interface in a system corresponds to a 192path through which messages may be sent and received. 193A network interface usually has a hardware device associated with it, 194though certain interfaces such as the loopback interface, 195.Xr lo 4 , 196do not. 197.Pp 198The following 199.Xr ioctl 2 200calls may be used to manipulate network interfaces. 201The 202.Xr ioctl 2 203is made on a socket (typically of type 204.Dv SOCK_DGRAM ) 205in the desired domain. 206Most of the requests 207take an 208.Vt ifreq 209structure pointer as their parameter. 210This structure is as follows: 211.Bd -literal 212struct ifreq { 213#define IFNAMSIZ 16 214 char ifr_name[IFNAMSIZ]; /* if name, e.g. "en0" */ 215 union { 216 struct sockaddr ifru_addr; 217 struct sockaddr ifru_dstaddr; 218 struct sockaddr ifru_broadaddr; 219 short ifru_flags; 220 int ifru_metric; 221 caddr_t ifru_data; 222 } ifr_ifru; 223#define ifr_addr ifr_ifru.ifru_addr /* address */ 224#define ifr_dstaddr ifr_ifru.ifru_dstaddr /* p-to-p peer */ 225#define ifr_broadaddr ifr_ifru.ifru_broadaddr /* broadcast address */ 226#define ifr_flags ifr_ifru.ifru_flags /* flags */ 227#define ifr_metric ifr_ifru.ifru_metric /* metric */ 228#define ifr_mtu ifr_ifru.ifru_metric /* mtu (overload) */ 229#define ifr_media ifr_ifru.ifru_metric /* media options */ 230#define ifr_data ifr_ifru.ifru_data /* used by interface */ 231}; 232.Ed 233.Pp 234The supported 235.Xr ioctl 2 236requests are: 237.Bl -tag -width Ds 238.It Dv SIOCSIFADDR Fa "struct ifreq *" 239Set the interface address for a protocol family. 240Following the address assignment, the 241.Dq initialization 242routine for the 243interface is called. 244.Pp 245This call has been deprecated and superseded by the 246.Dv SIOCAIFADDR 247call, described below. 248.It Dv SIOCSIFDSTADDR Fa "struct ifreq *" 249Set the point-to-point address for a protocol family and interface. 250.Pp 251This call has been deprecated and superseded by the 252.Dv SIOCAIFADDR 253call, described below. 254.It Dv SIOCSIFBRDADDR Fa "struct ifreq *" 255Set the broadcast address for a protocol family and interface. 256.Pp 257This call has been deprecated and superseded by the 258.Dv SIOCAIFADDR 259call, described below. 260.It Dv SIOCGIFADDR Fa "struct ifreq *" 261Get the interface address for a protocol family. 262.It Dv SIOCGIFDSTADDR Fa "struct ifreq *" 263Get the point-to-point address for a protocol family and interface. 264.It Dv SIOCGIFBRDADDR Fa "struct ifreq *" 265Get the broadcast address for a protocol family and interface. 266.It Dv SIOCGIFDESCR Fa "struct ifreq *" 267Get the interface description, returned in the 268.Va ifru_data 269field. 270.It Dv SIOCSIFDESCR Fa "struct ifreq *" 271Set the interface description to the value of the 272.Va ifru_data 273field, limited to the size of 274.Dv IFDESCRSIZE . 275.It Dv SIOCSIFFLAGS Fa "struct ifreq *" 276Set the interface flags. 277If the interface is marked down, any processes currently routing packets 278through the interface are notified; some interfaces may be reset so that 279incoming packets are no longer received. 280When marked up again, the interface is reinitialized. 281.It Dv SIOCGIFFLAGS Fa "struct ifreq *" 282Get the interface flags. 283.It Dv SIOCSIFMEDIA Fa "struct ifreq *" 284Set the interface media settings. 285See 286.Xr ifmedia 4 287for possible values. 288.It Dv SIOCGIFMEDIA Fa "struct ifmediareq *" 289Get the interface media settings. 290The 291.Vt ifmediareq 292structure is as follows: 293.Bd -literal 294struct ifmediareq { 295 char ifm_name[IFNAMSIZ]; /* if name, e.g. "en0" */ 296 int ifm_current; /* current media options */ 297 int ifm_mask; /* don't care mask */ 298 int ifm_status; /* media status */ 299 int ifm_active; /* active options */ 300 int ifm_count; /* #entries in ifm_ulist array */ 301 int *ifm_ulist; /* media words */ 302}; 303.Ed 304.Pp 305See 306.Xr ifmedia 4 307for interpreting this value. 308.It Dv SIOCSIFMETRIC Fa "struct ifreq *" 309Set the interface routing metric. 310The metric is used only by user-level routers. 311.It Dv SIOCGIFMETRIC Fa "struct ifreq *" 312Get the interface metric. 313.It Dv SIOCAIFADDR Fa "struct ifaliasreq *" 314An interface may have more than one address associated with it 315in some protocols. 316This request provides a means to add additional addresses (or modify 317characteristics of the primary address if the default address for the 318address family is specified). 319.Pp 320Rather than making separate calls to set destination or broadcast addresses, 321or network masks (now an integral feature of multiple protocols), a separate 322structure, 323.Vt ifaliasreq , 324is used to specify all three facets simultaneously (see below). 325One would use a slightly tailored version of this structure specific 326to each family (replacing each 327.Vt sockaddr 328by one 329of the family-specific type). 330One should always set the length of a 331.Vt sockaddr , 332as described in 333.Xr ioctl 2 . 334.Pp 335The 336.Vt ifaliasreq 337structure is as follows: 338.Bd -literal 339struct ifaliasreq { 340 char ifra_name[IFNAMSIZ]; /* if name, e.g. "en0" */ 341 struct sockaddr ifra_addr; 342 struct sockaddr ifra_dstaddr; 343#define ifra_broadaddr ifra_dstaddr 344 struct sockaddr ifra_mask; 345}; 346.Ed 347.It Dv SIOCDIFADDR Fa "struct ifreq *" 348This request deletes the specified address from the list 349associated with an interface. 350It also uses the 351.Vt ifaliasreq 352structure to allow for the possibility of protocols allowing 353multiple masks or destination addresses, and also adopts the 354convention that specification of the default address means 355to delete the first address for the interface belonging to 356the address family in which the original socket was opened. 357.It Dv SIOCGIFCONF Fa "struct ifconf *" 358Get the interface configuration list. 359This request takes an 360.Vt ifconf 361structure (see below) as a value-result parameter. 362The 363.Va ifc_len 364field should be initially set to the size of the buffer 365pointed to by 366.Va ifc_buf . 367On return it will contain the length, in bytes, of the 368configuration list. 369.Pp 370Alternately, if the 371.Va ifc_len 372passed in is set to 0, 373.Dv SIOCGIFCONF 374will set 375.Va ifc_len 376to the size that 377.Va ifc_buf 378needs to be to fit the entire configuration list and will not 379fill in the other parameters. 380This is useful for determining the exact size that 381.Va ifc_buf 382needs to be in advance. 383Note, however, that this is an extension 384that not all operating systems support. 385.Bd -literal 386struct ifconf { 387 int ifc_len; /* size of associated buffer */ 388 union { 389 caddr_t ifcu_buf; 390 struct ifreq *ifcu_req; 391 } ifc_ifcu; 392#define ifc_buf ifc_ifcu.ifcu_buf /* buffer address */ 393#define ifc_req ifc_ifcu.ifcu_req /* array of structures ret'd */ 394}; 395.Ed 396.It Dv SIOCIFCREATE Fa "struct ifreq *" 397Attempt to create the specified interface. 398.It Dv SIOCIFDESTROY Fa "struct ifreq *" 399Attempt to destroy the specified interface. 400.It Dv SIOCIFGCLONERS Fa "struct if_clonereq *" 401Get the list of clonable interfaces. 402This request takes an 403.Vt if_clonereq 404structure pointer (see below) as a value-result parameter. 405The 406.Va ifcr_count 407field should be set to the number of 408.Dv IFNAMSIZ Ns -sized 409strings that can fit in the buffer pointed to by 410.Va ifcr_buffer . 411On return, 412.Va ifcr_total 413will be set to the number of clonable interfaces, and the buffer pointed 414to by 415.Va ifcr_buffer 416will be filled with the names of clonable interfaces aligned on 417.Dv IFNAMSIZ 418boundaries. 419.Pp 420The 421.Vt if_clonereq 422structure is as follows: 423.Bd -literal 424struct if_clonereq { 425 int ifcr_total; /* total cloners (out) */ 426 int ifcr_count; /* room for this many in user buf */ 427 char *ifcr_buffer; /* buffer for cloner names */ 428}; 429.Ed 430.It Dv SIOCAIFGROUP Fa "struct ifgroupreq *" 431Associate the interface named by 432.Va ifgr_name 433with the interface group named by 434.Va ifgr_group . 435The 436.Vt ifgroupreq 437structure is as follows: 438.Bd -literal 439struct ifg_req { 440 char ifgrq_group[IFNAMSIZ]; 441}; 442 443struct ifgroupreq { 444 char ifgr_name[IFNAMSIZ]; 445 u_int ifgr_len; 446 union { 447 char ifgru_group[IFNAMSIZ]; 448 struct ifg_req *ifgru_groups; 449 } ifgr_ifgru; 450#define ifgr_group ifgr_ifgru.ifgru_group 451#define ifgr_groups ifgr_ifgru.ifgru_groups 452}; 453.Ed 454.It Dv SIOCGIFGROUP Fa "struct ifgroupreq *" 455Retrieve the list of groups for which an interface is a member. 456The interface is named by 457.Va ifgr_name . 458On enter, the amount of memory in which the group names will 459be written is stored in 460.Va ifgr_len , 461and the group names themselves will be written to the memory 462pointed to by 463.Va ifgr_groups . 464On return, the amount of memory actually written is returned in 465.Va ifgr_len . 466.Pp 467Alternately, if the 468.Va ifgr_len 469passed in is set to 0, 470.Dv SIOCGIFGROUP 471will set 472.Va ifgr_len 473to the size that 474.Va ifgr_groups 475needs to be to fit the entire group list and will not 476fill in the other parameters. 477This is useful for determining the exact size that 478.Va ifgr_groups 479needs to be in advance. 480.It Dv SIOCDIFGROUP Fa "struct ifgroupreq *" 481Remove the membership of the interface named by 482.Va ifgr_name 483from the group 484.Va ifgr_group . 485.El 486.Sh SEE ALSO 487.Xr netstat 1 , 488.Xr ioctl 2 , 489.Xr socket 2 , 490.Xr inet 3 , 491.Xr ipx 3 , 492.Xr arp 4 , 493.Xr bridge 4 , 494.Xr ifmedia 4 , 495.Xr inet 4 , 496.Xr intro 4 , 497.Xr ip 4 , 498.Xr ip6 4 , 499.Xr lo 4 , 500.Xr pf 4 , 501.Xr tcp 4 , 502.Xr udp 4 , 503.Xr hosts 5 , 504.Xr networks 5 , 505.Xr config 8 , 506.Xr ifconfig 8 , 507.Xr netstart 8 , 508.Xr route 8 , 509.Xr routed 8 510.Sh HISTORY 511The 512.Nm 513manual appeared in 514.Bx 4.3 Tahoe . 515