xref: /freebsd/sys/net/route.h (revision 27c43fe1f3795622c5bd4bbfc465a29a800c0799)
1 /*-
2  * Copyright (c) 1980, 1986, 1993
3  *	The Regents of the University of California.  All rights reserved.
4  *
5  * Redistribution and use in source and binary forms, with or without
6  * modification, are permitted provided that the following conditions
7  * are met:
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.
16  *
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
27  * SUCH DAMAGE.
28  *
29  *	@(#)route.h	8.4 (Berkeley) 1/9/95
30  * $FreeBSD$
31  */
32 
33 #ifndef _NET_ROUTE_H_
34 #define _NET_ROUTE_H_
35 
36 #include <sys/counter.h>
37 
38 /*
39  * Kernel resident routing tables.
40  *
41  * The routing tables are initialized when interface addresses
42  * are set by making entries for all directly connected interfaces.
43  */
44 
45 /*
46  * A route consists of a destination address, a reference
47  * to a routing entry, and a reference to an llentry.
48  * These are often held by protocols in their control
49  * blocks, e.g. inpcb.
50  */
51 struct route {
52 	struct	rtentry *ro_rt;
53 	struct	llentry *ro_lle;
54 	struct	in_ifaddr *ro_ia;
55 	int		ro_flags;
56 	struct	sockaddr ro_dst;
57 };
58 
59 #define	RT_CACHING_CONTEXT	0x1	/* XXX: not used anywhere */
60 #define	RT_NORTREF		0x2	/* doesn't hold reference on ro_rt */
61 
62 struct rt_metrics {
63 	u_long	rmx_locks;	/* Kernel must leave these values alone */
64 	u_long	rmx_mtu;	/* MTU for this path */
65 	u_long	rmx_hopcount;	/* max hops expected */
66 	u_long	rmx_expire;	/* lifetime for route, e.g. redirect */
67 	u_long	rmx_recvpipe;	/* inbound delay-bandwidth product */
68 	u_long	rmx_sendpipe;	/* outbound delay-bandwidth product */
69 	u_long	rmx_ssthresh;	/* outbound gateway buffer limit */
70 	u_long	rmx_rtt;	/* estimated round trip time */
71 	u_long	rmx_rttvar;	/* estimated rtt variance */
72 	u_long	rmx_pksent;	/* packets sent using this route */
73 	u_long	rmx_weight;	/* route weight */
74 	u_long	rmx_filler[3];	/* will be used for T/TCP later */
75 };
76 
77 /*
78  * rmx_rtt and rmx_rttvar are stored as microseconds;
79  * RTTTOPRHZ(rtt) converts to a value suitable for use
80  * by a protocol slowtimo counter.
81  */
82 #define	RTM_RTTUNIT	1000000	/* units for rtt, rttvar, as units per sec */
83 #define	RTTTOPRHZ(r)	((r) / (RTM_RTTUNIT / PR_SLOWHZ))
84 
85 #define	RT_DEFAULT_FIB	0	/* Explicitly mark fib=0 restricted cases */
86 #define	RT_ALL_FIBS	-1	/* Announce event for every fib */
87 extern u_int rt_numfibs;	/* number of usable routing tables */
88 extern u_int rt_add_addr_allfibs;	/* Announce interfaces to all fibs */
89 /*
90  * XXX kernel function pointer `rt_output' is visible to applications.
91  */
92 struct mbuf;
93 
94 /*
95  * We distinguish between routes to hosts and routes to networks,
96  * preferring the former if available.  For each route we infer
97  * the interface to use from the gateway address supplied when
98  * the route was entered.  Routes that forward packets through
99  * gateways are marked so that the output routines know to address the
100  * gateway rather than the ultimate destination.
101  */
102 #ifndef RNF_NORMAL
103 #include <net/radix.h>
104 #ifdef RADIX_MPATH
105 #include <net/radix_mpath.h>
106 #endif
107 #endif
108 
109 #if defined(_KERNEL) || defined(_WANT_RTENTRY)
110 struct rtentry {
111 	struct	radix_node rt_nodes[2];	/* tree glue, and other values */
112 	/*
113 	 * XXX struct rtentry must begin with a struct radix_node (or two!)
114 	 * because the code does some casts of a 'struct radix_node *'
115 	 * to a 'struct rtentry *'
116 	 */
117 #define	rt_key(r)	(*((struct sockaddr **)(&(r)->rt_nodes->rn_key)))
118 #define	rt_mask(r)	(*((struct sockaddr **)(&(r)->rt_nodes->rn_mask)))
119 	struct	sockaddr *rt_gateway;	/* value */
120 	struct	ifnet *rt_ifp;		/* the answer: interface to use */
121 	struct	ifaddr *rt_ifa;		/* the answer: interface address to use */
122 	int		rt_flags;	/* up/down?, host/net */
123 	int		rt_refcnt;	/* # held references */
124 	u_int		rt_fibnum;	/* which FIB */
125 	u_long		rt_mtu;		/* MTU for this path */
126 	u_long		rt_weight;	/* absolute weight */
127 	u_long		rt_expire;	/* lifetime for route, e.g. redirect */
128 #define	rt_endzero	rt_pksent
129 	counter_u64_t	rt_pksent;	/* packets sent using this route */
130 	struct mtx	rt_mtx;		/* mutex for routing entry */
131 };
132 #endif /* _KERNEL || _WANT_RTENTRY */
133 
134 /*
135  * Following structure necessary for 4.3 compatibility;
136  * We should eventually move it to a compat file.
137  */
138 struct ortentry {
139 	u_long	rt_hash;		/* to speed lookups */
140 	struct	sockaddr rt_dst;	/* key */
141 	struct	sockaddr rt_gateway;	/* value */
142 	short	rt_flags;		/* up/down?, host/net */
143 	short	rt_refcnt;		/* # held references */
144 	u_long	rt_use;			/* raw # packets forwarded */
145 	struct	ifnet *rt_ifp;		/* the answer: interface to use */
146 };
147 
148 #define	RTF_UP		0x1		/* route usable */
149 #define	RTF_GATEWAY	0x2		/* destination is a gateway */
150 #define	RTF_HOST	0x4		/* host entry (net otherwise) */
151 #define	RTF_REJECT	0x8		/* host or net unreachable */
152 #define	RTF_DYNAMIC	0x10		/* created dynamically (by redirect) */
153 #define	RTF_MODIFIED	0x20		/* modified dynamically (by redirect) */
154 #define RTF_DONE	0x40		/* message confirmed */
155 /*			0x80		   unused, was RTF_DELCLONE */
156 /*			0x100		   unused, was RTF_CLONING */
157 #define RTF_XRESOLVE	0x200		/* external daemon resolves name */
158 #define RTF_LLINFO	0x400		/* DEPRECATED - exists ONLY for backward
159 					   compatibility */
160 #define RTF_LLDATA	0x400		/* used by apps to add/del L2 entries */
161 #define RTF_STATIC	0x800		/* manually added */
162 #define RTF_BLACKHOLE	0x1000		/* just discard pkts (during updates) */
163 #define RTF_PROTO2	0x4000		/* protocol specific routing flag */
164 #define RTF_PROTO1	0x8000		/* protocol specific routing flag */
165 
166 /* XXX: temporary to stay API/ABI compatible with userland */
167 #ifndef _KERNEL
168 #define RTF_PRCLONING	0x10000		/* unused, for compatibility */
169 #endif
170 
171 /*			0x20000		   unused, was RTF_WASCLONED */
172 #define RTF_PROTO3	0x40000		/* protocol specific routing flag */
173 /*			0x80000		   unused */
174 #define RTF_PINNED	0x100000	/* route is immutable */
175 #define	RTF_LOCAL	0x200000 	/* route represents a local address */
176 #define	RTF_BROADCAST	0x400000	/* route represents a bcast address */
177 #define	RTF_MULTICAST	0x800000	/* route represents a mcast address */
178 					/* 0x8000000 and up unassigned */
179 #define	RTF_STICKY	 0x10000000	/* always route dst->src */
180 
181 #define	RTF_RNH_LOCKED	 0x40000000	/* radix node head is locked */
182 
183 #define	RTF_GWFLAG_COMPAT 0x80000000	/* a compatibility bit for interacting
184 					   with existing routing apps */
185 
186 /* Mask of RTF flags that are allowed to be modified by RTM_CHANGE. */
187 #define RTF_FMASK	\
188 	(RTF_PROTO1 | RTF_PROTO2 | RTF_PROTO3 | RTF_BLACKHOLE | \
189 	 RTF_REJECT | RTF_STATIC | RTF_STICKY)
190 
191 /*
192  * Routing statistics.
193  */
194 struct	rtstat {
195 	short	rts_badredirect;	/* bogus redirect calls */
196 	short	rts_dynamic;		/* routes created by redirects */
197 	short	rts_newgateway;		/* routes modified by redirects */
198 	short	rts_unreach;		/* lookups which failed */
199 	short	rts_wildcard;		/* lookups satisfied by a wildcard */
200 };
201 /*
202  * Structures for routing messages.
203  */
204 struct rt_msghdr {
205 	u_short	rtm_msglen;	/* to skip over non-understood messages */
206 	u_char	rtm_version;	/* future binary compatibility */
207 	u_char	rtm_type;	/* message type */
208 	u_short	rtm_index;	/* index for associated ifp */
209 	int	rtm_flags;	/* flags, incl. kern & message, e.g. DONE */
210 	int	rtm_addrs;	/* bitmask identifying sockaddrs in msg */
211 	pid_t	rtm_pid;	/* identify sender */
212 	int	rtm_seq;	/* for sender to identify action */
213 	int	rtm_errno;	/* why failed */
214 	int	rtm_fmask;	/* bitmask used in RTM_CHANGE message */
215 	u_long	rtm_inits;	/* which metrics we are initializing */
216 	struct	rt_metrics rtm_rmx; /* metrics themselves */
217 };
218 
219 #define RTM_VERSION	5	/* Up the ante and ignore older versions */
220 
221 /*
222  * Message types.
223  */
224 #define RTM_ADD		0x1	/* Add Route */
225 #define RTM_DELETE	0x2	/* Delete Route */
226 #define RTM_CHANGE	0x3	/* Change Metrics or flags */
227 #define RTM_GET		0x4	/* Report Metrics */
228 #define RTM_LOSING	0x5	/* Kernel Suspects Partitioning */
229 #define RTM_REDIRECT	0x6	/* Told to use different route */
230 #define RTM_MISS	0x7	/* Lookup failed on this address */
231 #define RTM_LOCK	0x8	/* fix specified metrics */
232 #define RTM_OLDADD	0x9	/* caused by SIOCADDRT */
233 #define RTM_OLDDEL	0xa	/* caused by SIOCDELRT */
234 #define RTM_RESOLVE	0xb	/* req to resolve dst to LL addr */
235 #define RTM_NEWADDR	0xc	/* address being added to iface */
236 #define RTM_DELADDR	0xd	/* address being removed from iface */
237 #define RTM_IFINFO	0xe	/* iface going up/down etc. */
238 #define	RTM_NEWMADDR	0xf	/* mcast group membership being added to if */
239 #define	RTM_DELMADDR	0x10	/* mcast group membership being deleted */
240 #define	RTM_IFANNOUNCE	0x11	/* iface arrival/departure */
241 #define	RTM_IEEE80211	0x12	/* IEEE80211 wireless event */
242 
243 /*
244  * Bitmask values for rtm_inits and rmx_locks.
245  */
246 #define RTV_MTU		0x1	/* init or lock _mtu */
247 #define RTV_HOPCOUNT	0x2	/* init or lock _hopcount */
248 #define RTV_EXPIRE	0x4	/* init or lock _expire */
249 #define RTV_RPIPE	0x8	/* init or lock _recvpipe */
250 #define RTV_SPIPE	0x10	/* init or lock _sendpipe */
251 #define RTV_SSTHRESH	0x20	/* init or lock _ssthresh */
252 #define RTV_RTT		0x40	/* init or lock _rtt */
253 #define RTV_RTTVAR	0x80	/* init or lock _rttvar */
254 #define RTV_WEIGHT	0x100	/* init or lock _weight */
255 
256 /*
257  * Bitmask values for rtm_addrs.
258  */
259 #define RTA_DST		0x1	/* destination sockaddr present */
260 #define RTA_GATEWAY	0x2	/* gateway sockaddr present */
261 #define RTA_NETMASK	0x4	/* netmask sockaddr present */
262 #define RTA_GENMASK	0x8	/* cloning mask sockaddr present */
263 #define RTA_IFP		0x10	/* interface name sockaddr present */
264 #define RTA_IFA		0x20	/* interface addr sockaddr present */
265 #define RTA_AUTHOR	0x40	/* sockaddr for author of redirect */
266 #define RTA_BRD		0x80	/* for NEWADDR, broadcast or p-p dest addr */
267 
268 /*
269  * Index offsets for sockaddr array for alternate internal encoding.
270  */
271 #define RTAX_DST	0	/* destination sockaddr present */
272 #define RTAX_GATEWAY	1	/* gateway sockaddr present */
273 #define RTAX_NETMASK	2	/* netmask sockaddr present */
274 #define RTAX_GENMASK	3	/* cloning mask sockaddr present */
275 #define RTAX_IFP	4	/* interface name sockaddr present */
276 #define RTAX_IFA	5	/* interface addr sockaddr present */
277 #define RTAX_AUTHOR	6	/* sockaddr for author of redirect */
278 #define RTAX_BRD	7	/* for NEWADDR, broadcast or p-p dest addr */
279 #define RTAX_MAX	8	/* size of array to allocate */
280 
281 struct rt_addrinfo {
282 	int	rti_addrs;
283 	struct	sockaddr *rti_info[RTAX_MAX];
284 	int	rti_flags;
285 	struct	ifaddr *rti_ifa;
286 	struct	ifnet *rti_ifp;
287 };
288 
289 /*
290  * This macro returns the size of a struct sockaddr when passed
291  * through a routing socket. Basically we round up sa_len to
292  * a multiple of sizeof(long), with a minimum of sizeof(long).
293  * The check for a NULL pointer is just a convenience, probably never used.
294  * The case sa_len == 0 should only apply to empty structures.
295  */
296 #define SA_SIZE(sa)						\
297     (  (!(sa) || ((struct sockaddr *)(sa))->sa_len == 0) ?	\
298 	sizeof(long)		:				\
299 	1 + ( (((struct sockaddr *)(sa))->sa_len - 1) | (sizeof(long) - 1) ) )
300 
301 #ifdef _KERNEL
302 
303 #define RT_LINK_IS_UP(ifp)	(!((ifp)->if_capabilities & IFCAP_LINKSTATE) \
304 				 || (ifp)->if_link_state == LINK_STATE_UP)
305 
306 #define	RT_LOCK_INIT(_rt) \
307 	mtx_init(&(_rt)->rt_mtx, "rtentry", NULL, MTX_DEF | MTX_DUPOK)
308 #define	RT_LOCK(_rt)		mtx_lock(&(_rt)->rt_mtx)
309 #define	RT_UNLOCK(_rt)		mtx_unlock(&(_rt)->rt_mtx)
310 #define	RT_LOCK_DESTROY(_rt)	mtx_destroy(&(_rt)->rt_mtx)
311 #define	RT_LOCK_ASSERT(_rt)	mtx_assert(&(_rt)->rt_mtx, MA_OWNED)
312 #define	RT_UNLOCK_COND(_rt)	do {				\
313 	if (mtx_owned(&(_rt)->rt_mtx))				\
314 		mtx_unlock(&(_rt)->rt_mtx);			\
315 } while (0)
316 
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++;					\
322 } while (0)
323 
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--;					\
329 } while (0)
330 
331 #define	RTFREE_LOCKED(_rt) do {					\
332 	if ((_rt)->rt_refcnt <= 1)				\
333 		rtfree(_rt);					\
334 	else {							\
335 		RT_REMREF(_rt);					\
336 		RT_UNLOCK(_rt);					\
337 	}							\
338 	/* guard against invalid refs */			\
339 	_rt = 0;						\
340 } while (0)
341 
342 #define	RTFREE(_rt) do {					\
343 	RT_LOCK(_rt);						\
344 	RTFREE_LOCKED(_rt);					\
345 } while (0)
346 
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;			\
352 		} else {					\
353 			RT_LOCK((_ro)->ro_rt);			\
354 			RTFREE_LOCKED((_ro)->ro_rt);		\
355 		}						\
356 	}							\
357 } while (0)
358 
359 struct radix_node_head *rt_tables_get_rnh(int, int);
360 
361 struct ifmultiaddr;
362 
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 *);
375 
376 int	rtsock_addrmsg(int, struct ifaddr *, int);
377 int	rtsock_routemsg(int, struct ifnet *ifp, int, struct rtentry *, int);
378 
379 /*
380  * Note the following locking behavior:
381  *
382  *    rtalloc_ign() and rtalloc() return ro->ro_rt unlocked
383  *
384  *    rtalloc1() returns a locked rtentry
385  *
386  *    rtfree() and RTFREE_LOCKED() require a locked rtentry
387  *
388  *    RTFREE() uses an unlocked entry.
389  */
390 
391 int	 rtexpunge(struct rtentry *);
392 void	 rtfree(struct rtentry *);
393 int	 rt_check(struct rtentry **, struct rtentry **, struct sockaddr *);
394 
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 **);
407 
408 #ifndef BURN_BRIDGES
409 /* defaults to "all" FIBs */
410 int	 rtinit_fib(struct ifaddr *, int, int);
411 #endif
412 
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..
416  */
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);
427 
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);
434 #endif
435 
436 #endif
437