2 * Copyright (c) 1980, 1986, 1993
3 * The Regents of the University of California. All rights reserved.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
13 * 4. Neither the name of the University nor the names of its contributors
14 * may be used to endorse or promote products derived from this software
15 * without specific prior written permission.
17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * @(#)route.h 8.4 (Berkeley) 1/9/95
37 * Kernel resident routing tables.
39 * The routing tables are initialized when interface addresses
40 * are set by making entries for all directly connected interfaces.
44 * A route consists of a destination address, a reference
45 * to a routing entry, and a reference to an llentry.
46 * These are often held by protocols in their control
50 struct rtentry *ro_rt;
51 struct llentry *ro_lle;
52 struct in_ifaddr *ro_ia;
54 struct sockaddr ro_dst;
57 #define RT_CACHING_CONTEXT 0x1 /* XXX: not used anywhere */
58 #define RT_NORTREF 0x2 /* doesn't hold reference on ro_rt */
61 * These numbers are used by reliable protocols for determining
62 * retransmission behavior and are included in the routing structure.
64 struct rt_metrics_lite {
65 u_long rmx_mtu; /* MTU for this path */
66 u_long rmx_expire; /* lifetime for route, e.g. redirect */
67 u_long rmx_pksent; /* packets sent using this route */
68 u_long rmx_weight; /* absolute weight */
72 u_long rmx_locks; /* Kernel must leave these values alone */
73 u_long rmx_mtu; /* MTU for this path */
74 u_long rmx_hopcount; /* max hops expected */
75 u_long rmx_expire; /* lifetime for route, e.g. redirect */
76 u_long rmx_recvpipe; /* inbound delay-bandwidth product */
77 u_long rmx_sendpipe; /* outbound delay-bandwidth product */
78 u_long rmx_ssthresh; /* outbound gateway buffer limit */
79 u_long rmx_rtt; /* estimated round trip time */
80 u_long rmx_rttvar; /* estimated rtt variance */
81 u_long rmx_pksent; /* packets sent using this route */
82 u_long rmx_weight; /* route weight */
83 u_long rmx_filler[3]; /* will be used for T/TCP later */
87 * rmx_rtt and rmx_rttvar are stored as microseconds;
88 * RTTTOPRHZ(rtt) converts to a value suitable for use
89 * by a protocol slowtimo counter.
91 #define RTM_RTTUNIT 1000000 /* units for rtt, rttvar, as units per sec */
92 #define RTTTOPRHZ(r) ((r) / (RTM_RTTUNIT / PR_SLOWHZ))
94 #define RT_DEFAULT_FIB 0 /* Explicitly mark fib=0 restricted cases */
95 #define RT_ALL_FIBS -1 /* Announce event for every fib */
96 extern u_int rt_numfibs; /* number of usable routing tables */
97 extern u_int rt_add_addr_allfibs; /* Announce interfaces to all fibs */
99 * XXX kernel function pointer `rt_output' is visible to applications.
104 * We distinguish between routes to hosts and routes to networks,
105 * preferring the former if available. For each route we infer
106 * the interface to use from the gateway address supplied when
107 * the route was entered. Routes that forward packets through
108 * gateways are marked so that the output routines know to address the
109 * gateway rather than the ultimate destination.
112 #include <net/radix.h>
114 #include <net/radix_mpath.h>
118 struct radix_node rt_nodes[2]; /* tree glue, and other values */
120 * XXX struct rtentry must begin with a struct radix_node (or two!)
121 * because the code does some casts of a 'struct radix_node *'
122 * to a 'struct rtentry *'
124 #define rt_key(r) (*((struct sockaddr **)(&(r)->rt_nodes->rn_key)))
125 #define rt_mask(r) (*((struct sockaddr **)(&(r)->rt_nodes->rn_mask)))
126 struct sockaddr *rt_gateway; /* value */
127 int rt_flags; /* up/down?, host/net */
128 int rt_refcnt; /* # held references */
129 struct ifnet *rt_ifp; /* the answer: interface to use */
130 struct ifaddr *rt_ifa; /* the answer: interface address to use */
131 struct rt_metrics_lite rt_rmx; /* metrics used by rx'ing protocols */
132 u_int rt_fibnum; /* which FIB */
134 /* XXX ugly, user apps use this definition but don't have a mtx def */
135 struct mtx rt_mtx; /* mutex for routing entry */
140 * Following structure necessary for 4.3 compatibility;
141 * We should eventually move it to a compat file.
144 u_long rt_hash; /* to speed lookups */
145 struct sockaddr rt_dst; /* key */
146 struct sockaddr rt_gateway; /* value */
147 short rt_flags; /* up/down?, host/net */
148 short rt_refcnt; /* # held references */
149 u_long rt_use; /* raw # packets forwarded */
150 struct ifnet *rt_ifp; /* the answer: interface to use */
153 #define rt_use rt_rmx.rmx_pksent
155 #define RTF_UP 0x1 /* route usable */
156 #define RTF_GATEWAY 0x2 /* destination is a gateway */
157 #define RTF_HOST 0x4 /* host entry (net otherwise) */
158 #define RTF_REJECT 0x8 /* host or net unreachable */
159 #define RTF_DYNAMIC 0x10 /* created dynamically (by redirect) */
160 #define RTF_MODIFIED 0x20 /* modified dynamically (by redirect) */
161 #define RTF_DONE 0x40 /* message confirmed */
162 /* 0x80 unused, was RTF_DELCLONE */
163 /* 0x100 unused, was RTF_CLONING */
164 #define RTF_XRESOLVE 0x200 /* external daemon resolves name */
165 #define RTF_LLINFO 0x400 /* DEPRECATED - exists ONLY for backward
167 #define RTF_LLDATA 0x400 /* used by apps to add/del L2 entries */
168 #define RTF_STATIC 0x800 /* manually added */
169 #define RTF_BLACKHOLE 0x1000 /* just discard pkts (during updates) */
170 #define RTF_PROTO2 0x4000 /* protocol specific routing flag */
171 #define RTF_PROTO1 0x8000 /* protocol specific routing flag */
173 /* XXX: temporary to stay API/ABI compatible with userland */
175 #define RTF_PRCLONING 0x10000 /* unused, for compatibility */
178 /* 0x20000 unused, was RTF_WASCLONED */
179 #define RTF_PROTO3 0x40000 /* protocol specific routing flag */
181 #define RTF_PINNED 0x100000 /* route is immutable */
182 #define RTF_LOCAL 0x200000 /* route represents a local address */
183 #define RTF_BROADCAST 0x400000 /* route represents a bcast address */
184 #define RTF_MULTICAST 0x800000 /* route represents a mcast address */
185 /* 0x8000000 and up unassigned */
186 #define RTF_STICKY 0x10000000 /* always route dst->src */
188 #define RTF_RNH_LOCKED 0x40000000 /* radix node head is locked */
190 /* Mask of RTF flags that are allowed to be modified by RTM_CHANGE. */
192 (RTF_PROTO1 | RTF_PROTO2 | RTF_PROTO3 | RTF_BLACKHOLE | \
193 RTF_REJECT | RTF_STATIC | RTF_STICKY)
196 * Routing statistics.
199 short rts_badredirect; /* bogus redirect calls */
200 short rts_dynamic; /* routes created by redirects */
201 short rts_newgateway; /* routes modified by redirects */
202 short rts_unreach; /* lookups which failed */
203 short rts_wildcard; /* lookups satisfied by a wildcard */
206 * Structures for routing messages.
209 u_short rtm_msglen; /* to skip over non-understood messages */
210 u_char rtm_version; /* future binary compatibility */
211 u_char rtm_type; /* message type */
212 u_short rtm_index; /* index for associated ifp */
213 int rtm_flags; /* flags, incl. kern & message, e.g. DONE */
214 int rtm_addrs; /* bitmask identifying sockaddrs in msg */
215 pid_t rtm_pid; /* identify sender */
216 int rtm_seq; /* for sender to identify action */
217 int rtm_errno; /* why failed */
218 int rtm_fmask; /* bitmask used in RTM_CHANGE message */
219 u_long rtm_inits; /* which metrics we are initializing */
220 struct rt_metrics rtm_rmx; /* metrics themselves */
223 #define RTM_VERSION 5 /* Up the ante and ignore older versions */
228 #define RTM_ADD 0x1 /* Add Route */
229 #define RTM_DELETE 0x2 /* Delete Route */
230 #define RTM_CHANGE 0x3 /* Change Metrics or flags */
231 #define RTM_GET 0x4 /* Report Metrics */
232 #define RTM_LOSING 0x5 /* Kernel Suspects Partitioning */
233 #define RTM_REDIRECT 0x6 /* Told to use different route */
234 #define RTM_MISS 0x7 /* Lookup failed on this address */
235 #define RTM_LOCK 0x8 /* fix specified metrics */
236 #define RTM_OLDADD 0x9 /* caused by SIOCADDRT */
237 #define RTM_OLDDEL 0xa /* caused by SIOCDELRT */
238 #define RTM_RESOLVE 0xb /* req to resolve dst to LL addr */
239 #define RTM_NEWADDR 0xc /* address being added to iface */
240 #define RTM_DELADDR 0xd /* address being removed from iface */
241 #define RTM_IFINFO 0xe /* iface going up/down etc. */
242 #define RTM_NEWMADDR 0xf /* mcast group membership being added to if */
243 #define RTM_DELMADDR 0x10 /* mcast group membership being deleted */
244 #define RTM_IFANNOUNCE 0x11 /* iface arrival/departure */
245 #define RTM_IEEE80211 0x12 /* IEEE80211 wireless event */
248 * Bitmask values for rtm_inits and rmx_locks.
250 #define RTV_MTU 0x1 /* init or lock _mtu */
251 #define RTV_HOPCOUNT 0x2 /* init or lock _hopcount */
252 #define RTV_EXPIRE 0x4 /* init or lock _expire */
253 #define RTV_RPIPE 0x8 /* init or lock _recvpipe */
254 #define RTV_SPIPE 0x10 /* init or lock _sendpipe */
255 #define RTV_SSTHRESH 0x20 /* init or lock _ssthresh */
256 #define RTV_RTT 0x40 /* init or lock _rtt */
257 #define RTV_RTTVAR 0x80 /* init or lock _rttvar */
258 #define RTV_WEIGHT 0x100 /* init or lock _weight */
261 * Bitmask values for rtm_addrs.
263 #define RTA_DST 0x1 /* destination sockaddr present */
264 #define RTA_GATEWAY 0x2 /* gateway sockaddr present */
265 #define RTA_NETMASK 0x4 /* netmask sockaddr present */
266 #define RTA_GENMASK 0x8 /* cloning mask sockaddr present */
267 #define RTA_IFP 0x10 /* interface name sockaddr present */
268 #define RTA_IFA 0x20 /* interface addr sockaddr present */
269 #define RTA_AUTHOR 0x40 /* sockaddr for author of redirect */
270 #define RTA_BRD 0x80 /* for NEWADDR, broadcast or p-p dest addr */
273 * Index offsets for sockaddr array for alternate internal encoding.
275 #define RTAX_DST 0 /* destination sockaddr present */
276 #define RTAX_GATEWAY 1 /* gateway sockaddr present */
277 #define RTAX_NETMASK 2 /* netmask sockaddr present */
278 #define RTAX_GENMASK 3 /* cloning mask sockaddr present */
279 #define RTAX_IFP 4 /* interface name sockaddr present */
280 #define RTAX_IFA 5 /* interface addr sockaddr present */
281 #define RTAX_AUTHOR 6 /* sockaddr for author of redirect */
282 #define RTAX_BRD 7 /* for NEWADDR, broadcast or p-p dest addr */
283 #define RTAX_MAX 8 /* size of array to allocate */
287 struct sockaddr *rti_info[RTAX_MAX];
289 struct ifaddr *rti_ifa;
290 struct ifnet *rti_ifp;
294 * This macro returns the size of a struct sockaddr when passed
295 * through a routing socket. Basically we round up sa_len to
296 * a multiple of sizeof(long), with a minimum of sizeof(long).
297 * The check for a NULL pointer is just a convenience, probably never used.
298 * The case sa_len == 0 should only apply to empty structures.
300 #define SA_SIZE(sa) \
301 ( (!(sa) || ((struct sockaddr *)(sa))->sa_len == 0) ? \
303 1 + ( (((struct sockaddr *)(sa))->sa_len - 1) | (sizeof(long) - 1) ) )
307 #define RT_LINK_IS_UP(ifp) (!((ifp)->if_capabilities & IFCAP_LINKSTATE) \
308 || (ifp)->if_link_state == LINK_STATE_UP)
310 #define RT_LOCK_INIT(_rt) \
311 mtx_init(&(_rt)->rt_mtx, "rtentry", NULL, MTX_DEF | MTX_DUPOK)
312 #define RT_LOCK(_rt) mtx_lock(&(_rt)->rt_mtx)
313 #define RT_UNLOCK(_rt) mtx_unlock(&(_rt)->rt_mtx)
314 #define RT_LOCK_DESTROY(_rt) mtx_destroy(&(_rt)->rt_mtx)
315 #define RT_LOCK_ASSERT(_rt) mtx_assert(&(_rt)->rt_mtx, MA_OWNED)
317 #define RT_ADDREF(_rt) do { \
318 RT_LOCK_ASSERT(_rt); \
319 KASSERT((_rt)->rt_refcnt >= 0, \
320 ("negative refcnt %d", (_rt)->rt_refcnt)); \
321 (_rt)->rt_refcnt++; \
324 #define RT_REMREF(_rt) do { \
325 RT_LOCK_ASSERT(_rt); \
326 KASSERT((_rt)->rt_refcnt > 0, \
327 ("bogus refcnt %d", (_rt)->rt_refcnt)); \
328 (_rt)->rt_refcnt--; \
331 #define RTFREE_LOCKED(_rt) do { \
332 if ((_rt)->rt_refcnt <= 1) \
338 /* guard against invalid refs */ \
342 #define RTFREE(_rt) do { \
344 RTFREE_LOCKED(_rt); \
347 #define RO_RTFREE(_ro) do { \
348 if ((_ro)->ro_rt) { \
349 if ((_ro)->ro_flags & RT_NORTREF) { \
350 (_ro)->ro_flags &= ~RT_NORTREF; \
351 (_ro)->ro_rt = NULL; \
353 RT_LOCK((_ro)->ro_rt); \
354 RTFREE_LOCKED((_ro)->ro_rt); \
359 struct radix_node_head *rt_tables_get_rnh(int, int);
363 void rt_ieee80211msg(struct ifnet *, int, void *, size_t);
364 void rt_ifannouncemsg(struct ifnet *, int);
365 void rt_ifmsg(struct ifnet *);
366 void rt_missmsg(int, struct rt_addrinfo *, int, int);
367 void rt_missmsg_fib(int, struct rt_addrinfo *, int, int, int);
368 void rt_newaddrmsg(int, struct ifaddr *, int, struct rtentry *);
369 void rt_newaddrmsg_fib(int, struct ifaddr *, int, struct rtentry *, int);
370 int rt_addrmsg(int, struct ifaddr *, int);
371 int rt_routemsg(int, struct ifnet *ifp, int, struct rtentry *, int);
372 void rt_newmaddrmsg(int, struct ifmultiaddr *);
373 int rt_setgate(struct rtentry *, struct sockaddr *, struct sockaddr *);
374 void rt_maskedcopy(struct sockaddr *, struct sockaddr *, struct sockaddr *);
376 int rtsock_addrmsg(int, struct ifaddr *, int);
377 int rtsock_routemsg(int, struct ifnet *ifp, int, struct rtentry *, int);
380 * Note the following locking behavior:
382 * rtalloc_ign() and rtalloc() return ro->ro_rt unlocked
384 * rtalloc1() returns a locked rtentry
386 * rtfree() and RTFREE_LOCKED() require a locked rtentry
388 * RTFREE() uses an unlocked entry.
391 int rtexpunge(struct rtentry *);
392 void rtfree(struct rtentry *);
393 int rt_check(struct rtentry **, struct rtentry **, struct sockaddr *);
395 /* XXX MRT COMPAT VERSIONS THAT SET UNIVERSE to 0 */
396 /* Thes are used by old code not yet converted to use multiple FIBS */
397 int rt_getifa(struct rt_addrinfo *);
398 void rtalloc_ign(struct route *ro, u_long ignflags);
399 void rtalloc(struct route *ro); /* XXX deprecated, use rtalloc_ign(ro, 0) */
400 struct rtentry *rtalloc1(struct sockaddr *, int, u_long);
401 int rtinit(struct ifaddr *, int, int);
402 int rtioctl(u_long, caddr_t);
403 void rtredirect(struct sockaddr *, struct sockaddr *,
404 struct sockaddr *, int, struct sockaddr *);
405 int rtrequest(int, struct sockaddr *,
406 struct sockaddr *, struct sockaddr *, int, struct rtentry **);
409 /* defaults to "all" FIBs */
410 int rtinit_fib(struct ifaddr *, int, int);
413 /* XXX MRT NEW VERSIONS THAT USE FIBs
414 * For now the protocol indepedent versions are the same as the AF_INET ones
415 * but this will change..
417 int rt_getifa_fib(struct rt_addrinfo *, u_int fibnum);
418 void rtalloc_ign_fib(struct route *ro, u_long ignflags, u_int fibnum);
419 void rtalloc_fib(struct route *ro, u_int fibnum);
420 struct rtentry *rtalloc1_fib(struct sockaddr *, int, u_long, u_int);
421 int rtioctl_fib(u_long, caddr_t, u_int);
422 void rtredirect_fib(struct sockaddr *, struct sockaddr *,
423 struct sockaddr *, int, struct sockaddr *, u_int);
424 int rtrequest_fib(int, struct sockaddr *,
425 struct sockaddr *, struct sockaddr *, int, struct rtentry **, u_int);
426 int rtrequest1_fib(int, struct rt_addrinfo *, struct rtentry **, u_int);
428 #include <sys/eventhandler.h>
429 typedef void (*rtevent_arp_update_fn)(void *, struct rtentry *, uint8_t *, struct sockaddr *);
430 typedef void (*rtevent_redirect_fn)(void *, struct rtentry *, struct rtentry *, struct sockaddr *);
431 /* route_arp_update_event is no longer generated; see arp_update_event */
432 EVENTHANDLER_DECLARE(route_arp_update_event, rtevent_arp_update_fn);
433 EVENTHANDLER_DECLARE(route_redirect_event, rtevent_redirect_fn);