1route(7P) Protocols route(7P)
2
6 route - kernel packet forwarding database
7
9 #include <sys/types.h>
10 #include <sys/socket.h>
11 #include <net/if.h>
12 #include <net/route.h>
13 int socket(PF_ROUTE, SOCK_RAW, int protocol);
15
18 UNIX provides some packet routing facilities. The kernel maintains a
19 routing information database, which is used in selecting the appropri‐
20 ate network interface when transmitting packets.
21 A user process (or possibly multiple co-operating processes) maintains
24 this database by sending messages over a special kind of socket. This
25 supplants fixed size ioctl(2)'s specified in routing(7P). Routing table
26 changes can only be carried out by the superuser.
27 The operating system might spontaneously emit routing messages in
30 response to external events, such as receipt of a re-direct, or failure
31 to locate a suitable route for a request. The message types are
32 described in greater detail below.
33 Routing database entries come in two flavors: entries for a specific
36 host, or entries for all hosts on a generic subnetwork (as specified by
37 a bit mask and value under the mask). The effect of wildcard or default
38 route can be achieved by using a mask of all zeros, and there can be
39 hierarchical routes.
40 When the system is booted and addresses are assigned to the network
43 interfaces, the internet protocol family installs a routing table entry
44 for each interface when it is ready for traffic. Normally the protocol
45 specifies the route through each interface as a direct connection to
46 the destination host or network. If the route is direct, the transport
47 layer of a protocol family usually requests the packet be sent to the
48 same host specified in the packet. Otherwise, the interface is
49 requested to address the packet to the gateway listed in the routing
50 entry, that is, the packet is forwarded.
51 When routing a packet, the kernel attempts to find the most specific
54 route matching the destination. If no entry is found, the destination
55 is declared to be unreachable, and a routing-miss message is generated
56 if there are any listeners on the routing control socket (described
57 below). If there are two different mask and value-under-the-mask pairs
58 that match, the more specific is the one with more bits in the mask. A
59 route to a host is regarded as being supplied with a mask of as many
60 ones as there are bits in the destination.
61 A wildcard routing entry is specified with a zero destination address
64 value, and a mask of all zeroes. Wildcard routes are used when the sys‐
65 tem fails to find other routes matching the destination. The combina‐
66 tion of wildcard routes and routing redirects can provide an economical
67 mechanism for routing traffic.
68 One opens the channel for passing routing control messages by using the
71 socket call. There can be more than one routing socket open per system.
72 Messages are formed by a header followed by a small number of sockad‐
75 drs, whose length depend on the address family. sockaddrs are inter‐
76 preted by position. An example of a type of message with three
77 addresses might be a CIDR prefix route: Destination, Netmask, and Gate‐
78 way. The interpretation of which addresses are present is given by a
79 bit mask within the header, and the sequence is least significant to
80 most significant bit within the vector.
81 Any messages sent to the kernel are returned, and copies are sent to
84 all interested listeners. The kernel provides the process ID of the
85 sender, and the sender can use an additional sequence field to distin‐
86 guish between outstanding messages. However, message replies can be
87 lost when kernel buffers are exhausted.
88 The protocol parameter specifies which messages an application listen‐
91 ing on the routing socket is interested in seeing, based on the the
92 address family of the sockaddrs present. Currently, you can specify
93 AF_INET and AF_INET6 to filter the messages seen by the listener, or
94 alternatively, you can specify AF_UNSPEC to indicate that the listener
95 is interested in all routing messages.
96 The kernel might reject certain messages, and indicates this by filling
99 in the rtm_errno field of the rt_msghdr struct (see below). The follow‐
100 ing codes are returned:
101 EEXIST If requested to duplicate an existing entry
103 ESRCH If requested to delete a non-existent entry
106 ENOBUFS If insufficient resources were available to install a new
109 route.
110 EPERM If the calling process does not have appropriate privileges
113 to alter the routing table.
114 In the current implementation, all routing processes run locally, and
118 the values for rtm_errno are available through the normal errno mecha‐
119 nism, even if the routing reply message is lost.
120 A process can avoid the expense of reading replies to its own messages
123 by issuing a setsockopt(3SOCKET) call indicating that the SO_USELOOP‐
124 BACK option at the SOL_SOCKET level is to be turned off. A process can
125 ignore all messages from the routing socket by doing a shut‐
126 down(3SOCKET) system call for further input.
127 By default, underlying IP interfaces in an IPMP group are not visible
130 to routing sockets. As such, routing sockets do not receive events
131 related to underlying IP interface in an IPMP group. For consistency,
132 when an IP interface is placed into an IPMP group, RTM_DELADDR messages
133 are generated for each IFF_UP address that is not migrated to the cor‐
134 responding IPMP IP interface and an RTM_IFINFO message is sent indicat‐
135 ing the interface is down. Similarly, when an underlying interface is
136 removed from an IPMP group, an RTM_IFINFO message is sent indicating
137 the interface is again up and RTM_NEWADDR messages are generated for
138 each IFF_UP address found on the interface.
139 The RT_AWARE socket option at the SOL_ROUTE level allows an application
142 to indicate its awareness of certain features, which control routing
143 socket behavior. The supported values are:
144 RTAW_DEFAULT Default awareness.
146 RTAW_UNDER_IPMP IPMP underlying interface awareness. When enabled,
149 underlying IP interfaces in an IPMP group remain
150 visible to the routing socket and events related to
151 them continue to be generated.
152 An RTM_ADD request tied to an underlying IP interface in an IPMP group
156 is translated to an RTM_ADD request for its corresponding IPMP IP
157 interface. All routing socket requests other than RTM_ADD and RTM_GET
158 fail when issued on an underlying IP interface in an IPMP group.
159 If a route is in use when it is deleted, the routing entry is marked
162 down and removed from the routing table, but the resources associated
163 with it are not reclaimed until all references to it are released.
164 The RTM_IFINFO, RTM_NEWADDR, and RTM_ADD messages associated with
167 interface configuration (setting the IFF_UP bit) are normally delayed
168 until after Duplicate Address Detection completes. Thus, applications
169 that configure interfaces and wish to wait until the interface is ready
170 can wait until RTM_IFINFO is returned and SIOCGLIFFLAGS shows that
171 IFF_DUPLICATE is not set.
172 Messages
174 User processes can obtain information about the routing entry to a spe‐
175 cific destination by using a RTM_GET message.
176 Messages include:
179 #define RTM_ADD 0x1 /* Add Route */
181 #define RTM_DELETE 0x2 /* Delete Route */
182 #define RTM_CHANGE 0x3 /* Change Metrics, Flags, or Gateway */
183 #define RTM_GET 0x4 /* Report Information */
184 #define RTM_LOSING 0x5 /* Kernel Suspects Partitioning */
185 #define RTM_REDIRECT 0x6 /* Told to use different route */
186 #define RTM_MISS 0x7 /* Lookup failed on this address */
187 #define RTM_LOCK 0x8 /* fix specified metrics */
188 #define RTM_OLDADD 0x9 /* caused by SIOCADDRT */
189 #define RTM_OLDDEL 0xa /* caused by SIOCDELRT */
190 #define RTM_RESOLVE 0xb /* request to resolve dst to LL addr */
191 #define RTM_NEWADDR 0xc /* address being added to iface */
192 #define RTM_DELADDR 0xd /* address being removed from iface */
193 #define RTM_IFINFO 0xe /* iface going up/down etc. */
194 A message header consists of:
198 struct rt_msghdr {
200 ushort_t rtm_msglen; /* to skip over non-understood messages */
201 uchar_t rtm_version; /* future binary compatibility */
202 uchar_t rtm_type; /* message type */
203 ushort_t rtm_index; /* index for associated ifp */
204 pid_t rtm_pid; /* identify sender */
205 int rtm_addrs; /* bitmask identifying sockaddrs in msg */
206 int rtm_seq; /* for sender to identify action */
207 int rtm_errno; /* why failed */
208 int rtm_flags; /* flags, incl kern & message, e.g., DONE */
209 int rtm_use; /* from rtentry */
210 uint_t rtm_inits; /* which values we are initializing */
211 struct rt_metrics rtm_rmx; /* metrics themselves */
213 };
214 where
218 struct rt_metrics {
220 uint32_t rmx_locks; /* Kernel must leave these values alone */
221 uint32_t rmx_mtu; /* MTU for this path */
222 uint32_t rmx_hopcount; /* max hops expected */
223 uint32_t rmx_expire; /* lifetime for route, e.g., redirect */
224 uint32_t rmx_recvpipe; /* inbound delay-bandwidth product */
225 uint32_t rmx_sendpipe; /* outbound delay-bandwidth product */
226 uint32_t rmx_ssthresh; /* outbound gateway buffer limit */
227 uint32_t rmx_rtt; /* estimated round trip time */
228 uint32_t rmx_rttvar; /* estimated rtt variance */
229 uint32_t rmx_pksent; /* packets sent using this route */
230 };
231 /* Flags include the values */
233 #define RTF_UP 0x1 /* route usable */
236 #define RTF_GATEWAY 0x2 /* destination is a gateway */
237 #define RTF_HOST 0x4 /* host entry (net otherwise) */
238 #define RTF_REJECT 0x8 /* host or net unreachable */
239 #define RTF_DYNAMIC 0x10 /* created dynamically(by redirect) */
240 #define RTF_MODIFIED 0x20 /* modified dynamically(by redirect) */
241 #define RTF_DONE 0x40 /* message confirmed */
242 #define RTF_MASK 0x80 /* subnet mask present */
243 #define RTF_CLONING 0x100 /* generate new routes on use */
244 #define RTF_XRESOLVE 0x200 /* external daemon resolves name */
245 #define RTF_LLINFO 0x400 /* generated by ARP */
246 #define RTF_STATIC 0x800 /* manually added */
247 #define RTF_BLACKHOLE 0x1000 /* just discard pkts (during updates) */
248 #define RTF_PRIVATE 0x2000 /* do not advertise this route */
249 #define RTF_PROTO2 0x4000 /* protocol specific routing flag #2 */
250 #define RTF_PROTO1 0x8000 /* protocol specific routing flag #1 */
251 /* Specifiers for metric values in rmx_locks and rtm_inits are */
253 #define RTV_MTU 0x1 /* init or lock _mtu */
255 #define RTV_HOPCOUNT 0x2 /* init or lock _hopcount */
256 #define RTV_EXPIRE 0x4 /* init or lock _expire */
257 #define RTV_RPIPE 0x8 /* init or lock _recvpipe */
258 #define RTV_SPIPE 0x10 /* init or lock _sendpipe */
259 #define RTV_SSTHRESH 0x20 /* init or lock _ssthresh */
260 #define RTV_RTT 0x40 /* init or lock _rtt */
261 #define RTV_RTTVAR 0x80 /* init or lock _rttvar */
262 /* Specifiers for which addresses are present in the messages are */
264 #define RTA_DST 0x1 /* destination sockaddr present */
266 #define RTA_GATEWAY 0x2 /* gateway sockaddr present */
267 #define RTA_NETMASK 0x4 /* netmask sockaddr present */
268 #define RTA_GENMASK 0x8 /* cloning mask sockaddr present */
269 #define RTA_IFP 0x10 /* interface name sockaddr present */
270 #define RTA_IFA 0x20 /* interface addr sockaddr present */
271 #define RTA_AUTHOR 0x40 /* sockaddr for author of redirect */
272 #define RTA_BRD 0x80 /* for NEWADDR, broadcast or p-p dest addr */
273
276 ioctl(2), setsockopt(3SOCKET), shutdown(3SOCKET), routing(7P)
277
279 Some of the metrics might not be implemented and return zero. The
280 implemented metrics are set in rtm_inits.
281SunOS 5.11 5 Feb 2009 route(7P)