1INTRO(4N) INTRO(4N)
2
6 networking - introduction to networking facilities
7
9 #include <sys/socket.h>
10 #include <net/route.h>
11 #include <net/if.h>
12
14 This section briefly describes the networking facilities available in
15 the system. Documentation in this part of section 4 is broken up into
16 three areas: protocol families (domains), protocols, and network inter‐
17 faces. Entries describing a protocol family are marked ``4F,'' while
18 entries describing protocol use are marked ``4P.'' Hardware support
19 for network interfaces are found among the standard ``4'' entries.
20 All network protocols are associated with a specific protocol family.
22 A protocol family provides basic services to the protocol implementa‐
23 tion to allow it to function within a specific network environment.
24 These services may include packet fragmentation and reassembly, rout‐
25 ing, addressing, and basic transport. A protocol family may support
26 multiple methods of addressing, though the current protocol implementa‐
27 tions do not. A protocol family is normally comprised of a number of
28 protocols, one per socket(2) type. It is not required that a protocol
29 family support all socket types. A protocol family may contain multi‐
30 ple protocols supporting the same socket abstraction.
31 A protocol supports one of the socket abstractions detailed in
33 socket(2). A specific protocol may be accessed either by creating a
34 socket of the appropriate type and protocol family, or by requesting
35 the protocol explicitly when creating a socket. Protocols normally
36 accept only one type of address format, usually determined by the
37 addressing structure inherent in the design of the protocol family/net‐
38 work architecture. Certain semantics of the basic socket abstractions
39 are protocol specific. All protocols are expected to support the basic
40 model for their particular socket type, but may, in addition, provide
41 non-standard facilities or extensions to a mechanism. For example, a
42 protocol supporting the SOCK_STREAM abstraction may allow more than one
43 byte of out-of-band data to be transmitted per out-of-band message.
44 A network interface is similar to a device interface. Network inter‐
46 faces comprise the lowest layer of the networking subsystem, interact‐
47 ing with the actual transport hardware. An interface may support one
48 or more protocol families and/or address formats. The SYNOPSIS section
49 of each network interface entry gives a sample specification of the
50 related drivers for use in providing a system description to the
51 /sys/conf/config script. The DIAGNOSTICS section lists messages which
52 may appear on the console and/or in the system error log, /usr/adm/mes‐
53 sages (see syslogd(8)), due to errors in device operation.
54
56 The system currently supports the DARPA Internet protocols and the
57 Xerox Network Systems(tm) protocols. Raw socket interfaces are pro‐
58 vided to the IP protocol layer of the DARPA Internet, to the IMP link
59 layer (1822), and to the IDP protocol of Xerox NS. Consult the appro‐
60 priate manual pages in this section for more information regarding the
61 support for each protocol family.
62
64 Associated with each protocol family is an address format. The follow‐
65 ing address formats are used by the system (and additional formats are
66 defined for possible future implementation):
67 #define AF_UNIX 1 /* local to host (pipes, portals) */
69 #define AF_INET 2 /* internetwork: UDP, TCP, etc. */
70 #define AF_IMPLINK 3 /* arpanet imp addresses */
71 #define AF_PUP 4 /* pup protocols: e.g. BSP */
72 #define AF_NS 6 /* Xerox NS protocols */
73 #define AF_HYLINK 15 /* NSC Hyperchannel */
74
76 The network facilities provided limited packet routing. A simple set
77 of data structures comprise a ``routing table'' used in selecting the
78 appropriate network interface when transmitting packets. This table
79 contains a single entry for each route to a specific network or host.
80 A user process, the routing daemon, maintains this data base with the
81 aid of two socket-specific ioctl(2) commands, SIOCADDRT and SIOCDELRT.
82 The commands allow the addition and deletion of a single routing table
83 entry, respectively. Routing table manipulations may only be carried
84 out by super-user.
85 A routing table entry has the following form, as defined in
87 <net/route.h>;
88 struct rtentry {
90 u_long rt_hash;
91 struct sockaddr rt_dst;
92 struct sockaddr rt_gateway;
93 short rt_flags;
94 short rt_refcnt;
95 u_long rt_use;
96 struct ifnet *rt_ifp;
97 };
98 with rt_flags defined from,
100 #define RTF_UP 0x1 /* route usable */
102 #define RTF_GATEWAY 0x2 /* destination is a gateway */
103 #define RTF_HOST 0x4 /* host entry (net otherwise) */
104 #define RTF_DYNAMIC 0x10 /* created dynamically (by redirect) */
105 Routing table entries come in three flavors: for a specific host, for
107 all hosts on a specific network, for any destination not matched by
108 entries of the first two types (a wildcard route). When the system is
109 booted and addresses are assigned to the network interfaces, each pro‐
110 tocol family installs a routing table entry for each interface when it
111 is ready for traffic. Normally the protocol specifies the route
112 through each interface as a ``direct'' connection to the destination
113 host or network. If the route is direct, the transport layer of a pro‐
114 tocol family usually requests the packet be sent to the same host spec‐
115 ified in the packet. Otherwise, the interface is requested to address
116 the packet to the gateway listed in the routing entry (i.e. the packet
117 is forwarded).
118 Routing table entries installed by a user process may not specify the
120 hash, reference count, use, or interface fields; these are filled in by
121 the routing routines. If a route is in use when it is deleted
122 (rt_refcnt is non-zero), the routing entry will be marked down and
123 removed from the routing table, but the resources associated with it
124 will not be reclaimed until all references to it are released. The
125 routing code returns EEXIST if requested to duplicate an existing
126 entry, ESRCH if requested to delete a non-existent entry, or ENOBUFS if
127 insufficient resources were available to install a new route. User
128 processes read the routing tables through the /dev/kmem device. The
129 rt_use field contains the number of packets sent along the route.
130 When routing a packet, the kernel will first attempt to find a route to
132 the destination host. Failing that, a search is made for a route to
133 the network of the destination. Finally, any route to a default
134 (``wildcard'') gateway is chosen. If multiple routes are present in
135 the table, the first route found will be used. If no entry is found,
136 the destination is declared to be unreachable.
137 A wildcard routing entry is specified with a zero destination address
139 value. Wildcard routes are used only when the system fails to find a
140 route to the destination host and network. The combination of wildcard
141 routes and routing redirects can provide an economical mechanism for
142 routing traffic.
143
145 Each network interface in a system corresponds to a path through which
146 messages may be sent and received. A network interface usually has a
147 hardware device associated with it, though certain interfaces such as
148 the loopback interface, lo(4), do not.
149 The following ioctl calls may be used to manipulate network interfaces.
151 The ioctl is made on a socket (typically of type SOCK_DGRAM) in the
152 desired domain. Unless specified otherwise, the request takes an ifre‐
153 quest structure as its parameter. This structure has the form
154 struct ifreq {
156 #define IFNAMSIZ 16
157 char ifr_name[IFNAMSIZ]; /* if name, e.g. "en0" */
158 union {
159 struct sockaddr ifru_addr;
160 struct sockaddr ifru_dstaddr;
161 struct sockaddr ifru_broadaddr;
162 short ifru_flags;
163 int ifru_metric;
164 caddr_t ifru_data;
165 } ifr_ifru;
166 #define ifr_addr ifr_ifru.ifru_addr /* address */
167 #define ifr_dstaddr ifr_ifru.ifru_dstaddr /* other end of p-to-p link */
168 #define ifr_broadaddr ifr_ifru.ifru_broadaddr /* broadcast address */
169 #define ifr_flags ifr_ifru.ifru_flags /* flags */
170 #define ifr_metric ifr_ifru.ifru_metric /* metric */
171 #define ifr_data ifr_ifru.ifru_data /* for use by interface */
172 };
173 SIOCSIFADDR
175 Set interface address for protocol family. Following the
176 address assignment, the ``initialization'' routine for the
177 interface is called.
178 SIOCGIFADDR
180 Get interface address for protocol family.
181 SIOCSIFDSTADDR
183 Set point to point address for protocol family and interface.
184 SIOCGIFDSTADDR
186 Get point to point address for protocol family and interface.
187 SIOCSIFBRDADDR
189 Set broadcast address for protocol family and interface.
190 SIOCGIFBRDADDR
192 Get broadcast address for protocol family and interface.
193 SIOCSIFFLAGS
195 Set interface flags field. If the interface is marked down, any
196 processes currently routing packets through the interface are
197 notified; some interfaces may be reset so that incoming packets
198 are no longer received. When marked up again, the interface is
199 reinitialized.
200 SIOCGIFFLAGS
202 Get interface flags.
203 SIOCSIFMETRIC
205 Set interface routing metric. The metric is used only by user-
206 level routers.
207 SIOCGIFMETRIC
209 Get interface metric.
210 SIOCGIFCONF
212 Get interface configuration list. This request takes an ifconf
213 structure (see below) as a value-result parameter. The ifc_len
214 field should be initially set to the size of the buffer pointed
215 to by ifc_buf. On return it will contain the length, in bytes,
216 of the configuration list.
217 /*
219 * Structure used in SIOCGIFCONF request.
220 * Used to retrieve interface configuration
221 * for machine (useful for programs which
222 * must know all networks accessible).
223 */
224 struct ifconf {
225 int ifc_len; /* size of associated buffer */
226 union {
227 caddr_t ifcu_buf;
228 struct ifreq *ifcu_req;
229 } ifc_ifcu;
230 #define ifc_buf ifc_ifcu.ifcu_buf /* buffer address */
231 #define ifc_req ifc_ifcu.ifcu_req /* array of structures returned */
232 };
233
235 socket(2), ioctl(2), intro(4), config(8), routed(8C)
2364.2 Berkeley Distribution August 1, 1987 INTRO(4N)