1NETINTRO(4) 386BSD Programmer's Manual NETINTRO(4) 2 3NNAAMMEE 4 nneettwwoorrkkiinngg - introduction to networking facilities 5 6SSYYNNOOPPSSIISS 7 ##iinncclluuddee <<ssyyss//ssoocckkeett..hh>> 8 ##iinncclluuddee <<nneett//rroouuttee..hh>> 9 ##iinncclluuddee <<nneett//iiff..hh>> 10 11DDEESSCCRRIIPPTTIIOONN 12 This section is a general introduction to the networking facilities 13 available in the system. Documentation in this part of section 4 is 14 broken up into three areas: _p_r_o_t_o_c_o_l _f_a_m_i_l_i_e_s (domains), _p_r_o_t_o_c_o_l_s, and 15 _n_e_t_w_o_r_k _i_n_t_e_r_f_a_c_e_s. 16 17 All network protocols are associated with a specific _p_r_o_t_o_c_o_l _f_a_m_i_l_y. A 18 protocol family provides basic services to the protocol implementation to 19 allow it to function within a specific network environment. These 20 services may include packet fragmentation and reassembly, routing, 21 addressing, and basic transport. A protocol family may support multiple 22 methods of addressing, though the current protocol implementations do 23 not. A protocol family is normally comprised of a number of protocols, 24 one per socket(2) type. It is not required that a protocol family 25 support all socket types. A protocol family may contain multiple 26 protocols supporting the same socket abstraction. 27 28 A protocol supports one of the socket abstractions detailed in socket(2). 29 A specific protocol may be accessed either by creating a socket of the 30 appropriate type and protocol family, or by requesting the protocol 31 explicitly when creating a socket. Protocols normally accept only one 32 type of address format, usually determined by the addressing structure 33 inherent in the design of the protocol family/network architecture. 34 Certain semantics of the basic socket abstractions are protocol specific. 35 All protocols are expected to support the basic model for their 36 particular socket type, but may, in addition, provide non-standard 37 facilities or extensions to a mechanism. For example, a protocol 38 supporting the SOCK_STREAM abstraction may allow more than one byte of 39 out-of-band data to be transmitted per out-of-band message. 40 41 A network interface is similar to a device interface. Network interfaces 42 comprise the lowest layer of the networking subsystem, interacting with 43 the actual transport hardware. An interface may support one or more 44 protocol families and/or address formats. The SYNOPSIS section of each 45 network interface entry gives a sample specification of the related 46 drivers for use in providing a system description to the config(8) 47 program. The DIAGNOSTICS section lists messages which may appear on the 48 console and/or in the system error log, /_v_a_r/_l_o_g/_m_e_s_s_a_g_e_s (see 49 syslogd(8)), due to errors in device operation. 50 51PPRROOTTOOCCOOLLSS 52 The system currently supports the DARPA Internet protocols, the Xerox 53 Network Systems(tm) protocols, and some of the ISO OSI protocols. Raw 54 socket interfaces are provided to the IP protocol layer of the DARPA 55 Internet, to the IMP link layer (1822), and to the IDP protocol of Xerox 56 NS. Consult the appropriate manual pages in this section for more 57 information regarding the support for each protocol family. 58 59AADDDDRREESSSSIINNGG 60 Associated with each protocol family is an address format. All network 61 address adhere to a general structure, called a sockaddr, described 62 below. However, each protocol imposes finer and more specific structure, 63 generally renaming the variant, which is discussed in the protocol family 64 manual page alluded to above. 65 66 struct sockaddr { 67 u_char sa_len; 68 u_char sa_family; 69 char sa_data[14]; 70 }; 71 72 The field _s_a__l_e_n contains the total length of the of the structure, which 73 may exceed 16 bytes. The following address values for _s_a__f_a_m_i_l_y are 74 known to the system (and additional formats are defined for possible 75 future implementation): 76 77 #define AF_UNIX 1 /* local to host (pipes, portals) */ 78 #define AF_INET 2 /* internetwork: UDP, TCP, etc. */ 79 #define AF_IMPLINK 3 /* arpanet imp addresses */ 80 #define AF_NS 6 /* Xerox NS protocols */ 81 #define AF_CCITT 10 /* CCITT protocols, X.25 etc */ 82 #define AF_HYLINK 15 /* NSC Hyperchannel */ 83 #define AF_ISO 18 /* ISO protocols */ 84 85RROOUUTTIINNGG 86 UNIX provides some packet routing facilities. The kernel maintains a 87 routing information database, which is used in selecting the appropriate 88 network interface when transmitting packets. 89 90 A user process (or possibly multiple co-operating processes) maintains 91 this database by sending messages over a special kind of socket. This 92 supplants fixed size ioctl(2) used in earlier releases. 93 94 This facility is described in route(4). 95 96IINNTTEERRFFAACCEESS 97 Each network interface in a system corresponds to a path through which 98 messages may be sent and received. A network interface usually has a 99 hardware device associated with it, though certain interfaces such as the 100 loopback interface, lo(4), do not. 101 102 The following ioctl calls may be used to manipulate network interfaces. 103 The ioctl is made on a socket (typically of type SOCK_DGRAM) in the 104 desired domain. Most of the requests supported in earlier releases take 105 an _i_f_r_e_q structure as its parameter. This structure has the form 106 107 struct ifreq { 108 #define IFNAMSIZ 16 109 char ifr_name[IFNAMSIZE]; /* if name, e.g. "en0" */ 110 union { 111 struct sockaddr ifru_addr; 112 struct sockaddr ifru_dstaddr; 113 struct sockaddr ifru_broadaddr; 114 short ifru_flags; 115 int ifru_metric; 116 caddr_t ifru_data; 117 } ifr_ifru; 118 #define ifr_addr ifr_ifru.ifru_addr /* address */ 119 #define ifr_dstaddr ifr_ifru.ifru_dstaddr /* other end of p-to-p link */ 120 #define ifr_broadaddr ifr_ifru.ifru_broadaddr /* broadcast address */ 121 #define ifr_flags ifr_ifru.ifru_flags /* flags */ 122 #define ifr_metric ifr_ifru.ifru_metric /* metric */ 123 #define ifr_data ifr_ifru.ifru_data /* for use by interface */ 124 }; 125 126 Calls which are now depricated are: 127 128 SIOCSIFADDR Set interface address for protocol family. Following the 129 address assignment, the ``initialization'' routine for 130 131 132 the interface is called. 133 134 SIOCSIFDSTADDR Set point to point address for protocol family and 135 interface. 136 137 SIOCSIFBRDADDR Set broadcast address for protocol family and interface. 138 139 Ioctl requests to obtain addresses and requests both to set and retreive 140 other data are still fully supported and use the _i_f_r_e_q structure: 141 142 SIOCGIFADDR Get interface address for protocol family. 143 144 SIOCGIFDSTADDR Get point to point address for protocol family and 145 interface. 146 147 SIOCGIFBRDADDR Get broadcast address for protocol family and interface. 148 149 SIOCSIFFLAGS Set interface flags field. If the interface is marked 150 down, any processes currently routing packets through the 151 interface are notified; some interfaces may be reset so 152 that incoming packets are no longer received. When 153 marked up again, the interface is reinitialized. 154 155 SIOCGIFFLAGS Get interface flags. 156 157 SIOCSIFMETRIC Set interface routing metric. The metric is used only by 158 user-level routers. 159 160 SIOCGIFMETRIC Get interface metric. 161 162 There are two requests that make use of a new structure: 163 164 SIOCAIFADDR An interface may have more than one address associated 165 with it in some protocols. This request provides a means 166 to add additional addresses (or modify characteristics of 167 the primary address if the default address for the 168 address family is specified). Rather than making 169 separate calls to set destination or broadcast addresses, 170 or network masks (now an integral feature of multiple 171 protocols) a separate structure is used to specify all 172 three facets simultaneously (see below). One would use a 173 slightly tailored version of this struct specific to each 174 family (replacing each sockaddr by one of the family- 175 specific type). Where the sockaddr itself is larger than 176 the default size, one needs to modify the ioctl 177 identifier itself to include the total size, as described 178 in ioctl. 179 180 SIOCDIFADDR This requests deletes the specified address from the list 181 associated with an interface. It also uses the 182 _i_f__a_l_i_a_s_r_e_q structure to allow for the possibility of 183 protocols allowing multiple masks or destination 184 addresses, and also adopts the convention that 185 specification of the default address means to delete the 186 first address for the interface belonging to the address 187 family in which the original socket was opened. 188 189 SIOCGIFCONF Get interface configuration list. This request takes an 190 _i_f_c_o_n_f structure (see below) as a value-result parameter. 191 The _i_f_c__l_e_n field should be initially set to the size of 192 the buffer pointed to by _i_f_c__b_u_f. On return it will 193 contain the length, in bytes, of the configuration list. 194 195 /* 196 * Structure used in SIOCAIFCONF request. 197 */ 198 struct ifaliasreq { 199 char ifra_name[IFNAMSIZ]; /* if name, e.g. "en0" */ 200 struct sockaddr ifra_addr; 201 struct sockaddr ifra_broadaddr; 202 struct sockaddr ifra_mask; 203 }; 204 205 /* 206 * Structure used in SIOCGIFCONF request. 207 * Used to retrieve interface configuration 208 * for machine (useful for programs which 209 * must know all networks accessible). 210 */ 211 struct ifconf { 212 int ifc_len; /* size of associated buffer */ 213 union { 214 caddr_t ifcu_buf; 215 struct ifreq *ifcu_req; 216 } ifc_ifcu; 217 #define ifc_buf ifc_ifcu.ifcu_buf /* buffer address */ 218 #define ifc_req ifc_ifcu.ifcu_req /* array of structures returned */ 219 }; 220 221SSEEEE AALLSSOO 222 socket(2), ioctl(2), intro(4), config(8), routed(8) 223 224HHIISSTTOORRYY 225 The nneettiinnttrroo manual appeared in 4.3BSD-Tahoe. 226 2274.2 Berkeley Distribution March 28, 1991 4 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265