xref: /freebsd/sys/net/route.h (revision fcb560670601b2a4d87bb31d7531c8dcc37ee71b)
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 #include <net/vnet.h>
38 
39 /*
40  * Kernel resident routing tables.
41  *
42  * The routing tables are initialized when interface addresses
43  * are set by making entries for all directly connected interfaces.
44  */
45 
46 /*
47  * A route consists of a destination address, a reference
48  * to a routing entry, and a reference to an llentry.
49  * These are often held by protocols in their control
50  * blocks, e.g. inpcb.
51  */
52 struct route {
53 	struct	rtentry *ro_rt;
54 	struct	llentry *ro_lle;
55 	struct	in_ifaddr *ro_ia;
56 	int		ro_flags;
57 	struct	sockaddr ro_dst;
58 };
59 
60 #define	RT_CACHING_CONTEXT	0x1	/* XXX: not used anywhere */
61 #define	RT_NORTREF		0x2	/* doesn't hold reference on ro_rt */
62 
63 struct rt_metrics {
64 	u_long	rmx_locks;	/* Kernel must leave these values alone */
65 	u_long	rmx_mtu;	/* MTU for this path */
66 	u_long	rmx_hopcount;	/* max hops expected */
67 	u_long	rmx_expire;	/* lifetime for route, e.g. redirect */
68 	u_long	rmx_recvpipe;	/* inbound delay-bandwidth product */
69 	u_long	rmx_sendpipe;	/* outbound delay-bandwidth product */
70 	u_long	rmx_ssthresh;	/* outbound gateway buffer limit */
71 	u_long	rmx_rtt;	/* estimated round trip time */
72 	u_long	rmx_rttvar;	/* estimated rtt variance */
73 	u_long	rmx_pksent;	/* packets sent using this route */
74 	u_long	rmx_weight;	/* route weight */
75 	u_long	rmx_filler[3];	/* will be used for T/TCP later */
76 };
77 
78 /*
79  * rmx_rtt and rmx_rttvar are stored as microseconds;
80  * RTTTOPRHZ(rtt) converts to a value suitable for use
81  * by a protocol slowtimo counter.
82  */
83 #define	RTM_RTTUNIT	1000000	/* units for rtt, rttvar, as units per sec */
84 #define	RTTTOPRHZ(r)	((r) / (RTM_RTTUNIT / PR_SLOWHZ))
85 
86 #define	RT_DEFAULT_FIB	0	/* Explicitly mark fib=0 restricted cases */
87 #define	RT_ALL_FIBS	-1	/* Announce event for every fib */
88 #ifdef _KERNEL
89 extern u_int rt_numfibs;	/* number of usable routing tables */
90 VNET_DECLARE(u_int, rt_add_addr_allfibs); /* Announce interfaces to all fibs */
91 #define	V_rt_add_addr_allfibs	VNET(rt_add_addr_allfibs)
92 #endif
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 #define	RTF_UP		0x1		/* route usable */
135 #define	RTF_GATEWAY	0x2		/* destination is a gateway */
136 #define	RTF_HOST	0x4		/* host entry (net otherwise) */
137 #define	RTF_REJECT	0x8		/* host or net unreachable */
138 #define	RTF_DYNAMIC	0x10		/* created dynamically (by redirect) */
139 #define	RTF_MODIFIED	0x20		/* modified dynamically (by redirect) */
140 #define RTF_DONE	0x40		/* message confirmed */
141 /*			0x80		   unused, was RTF_DELCLONE */
142 /*			0x100		   unused, was RTF_CLONING */
143 #define RTF_XRESOLVE	0x200		/* external daemon resolves name */
144 #define RTF_LLINFO	0x400		/* DEPRECATED - exists ONLY for backward
145 					   compatibility */
146 #define RTF_LLDATA	0x400		/* used by apps to add/del L2 entries */
147 #define RTF_STATIC	0x800		/* manually added */
148 #define RTF_BLACKHOLE	0x1000		/* just discard pkts (during updates) */
149 #define RTF_PROTO2	0x4000		/* protocol specific routing flag */
150 #define RTF_PROTO1	0x8000		/* protocol specific routing flag */
151 /*			0x10000		   unused, was RTF_PRCLONING */
152 /*			0x20000		   unused, was RTF_WASCLONED */
153 #define RTF_PROTO3	0x40000		/* protocol specific routing flag */
154 #define	RTF_FIXEDMTU	0x80000		/* MTU was explicitly specified */
155 #define RTF_PINNED	0x100000	/* route is immutable */
156 #define	RTF_LOCAL	0x200000 	/* route represents a local address */
157 #define	RTF_BROADCAST	0x400000	/* route represents a bcast address */
158 #define	RTF_MULTICAST	0x800000	/* route represents a mcast address */
159 					/* 0x8000000 and up unassigned */
160 #define	RTF_STICKY	 0x10000000	/* always route dst->src */
161 
162 #define	RTF_RNH_LOCKED	 0x40000000	/* radix node head is locked */
163 
164 #define	RTF_GWFLAG_COMPAT 0x80000000	/* a compatibility bit for interacting
165 					   with existing routing apps */
166 
167 /* Mask of RTF flags that are allowed to be modified by RTM_CHANGE. */
168 #define RTF_FMASK	\
169 	(RTF_PROTO1 | RTF_PROTO2 | RTF_PROTO3 | RTF_BLACKHOLE | \
170 	 RTF_REJECT | RTF_STATIC | RTF_STICKY)
171 
172 /*
173  * Routing statistics.
174  */
175 struct	rtstat {
176 	short	rts_badredirect;	/* bogus redirect calls */
177 	short	rts_dynamic;		/* routes created by redirects */
178 	short	rts_newgateway;		/* routes modified by redirects */
179 	short	rts_unreach;		/* lookups which failed */
180 	short	rts_wildcard;		/* lookups satisfied by a wildcard */
181 };
182 /*
183  * Structures for routing messages.
184  */
185 struct rt_msghdr {
186 	u_short	rtm_msglen;	/* to skip over non-understood messages */
187 	u_char	rtm_version;	/* future binary compatibility */
188 	u_char	rtm_type;	/* message type */
189 	u_short	rtm_index;	/* index for associated ifp */
190 	int	rtm_flags;	/* flags, incl. kern & message, e.g. DONE */
191 	int	rtm_addrs;	/* bitmask identifying sockaddrs in msg */
192 	pid_t	rtm_pid;	/* identify sender */
193 	int	rtm_seq;	/* for sender to identify action */
194 	int	rtm_errno;	/* why failed */
195 	int	rtm_fmask;	/* bitmask used in RTM_CHANGE message */
196 	u_long	rtm_inits;	/* which metrics we are initializing */
197 	struct	rt_metrics rtm_rmx; /* metrics themselves */
198 };
199 
200 #define RTM_VERSION	5	/* Up the ante and ignore older versions */
201 
202 /*
203  * Message types.
204  */
205 #define RTM_ADD		0x1	/* Add Route */
206 #define RTM_DELETE	0x2	/* Delete Route */
207 #define RTM_CHANGE	0x3	/* Change Metrics or flags */
208 #define RTM_GET		0x4	/* Report Metrics */
209 #define RTM_LOSING	0x5	/* Kernel Suspects Partitioning */
210 #define RTM_REDIRECT	0x6	/* Told to use different route */
211 #define RTM_MISS	0x7	/* Lookup failed on this address */
212 #define RTM_LOCK	0x8	/* fix specified metrics */
213 		    /*	0x9  */
214 		    /*	0xa  */
215 #define RTM_RESOLVE	0xb	/* req to resolve dst to LL addr */
216 #define RTM_NEWADDR	0xc	/* address being added to iface */
217 #define RTM_DELADDR	0xd	/* address being removed from iface */
218 #define RTM_IFINFO	0xe	/* iface going up/down etc. */
219 #define	RTM_NEWMADDR	0xf	/* mcast group membership being added to if */
220 #define	RTM_DELMADDR	0x10	/* mcast group membership being deleted */
221 #define	RTM_IFANNOUNCE	0x11	/* iface arrival/departure */
222 #define	RTM_IEEE80211	0x12	/* IEEE80211 wireless event */
223 
224 /*
225  * Bitmask values for rtm_inits and rmx_locks.
226  */
227 #define RTV_MTU		0x1	/* init or lock _mtu */
228 #define RTV_HOPCOUNT	0x2	/* init or lock _hopcount */
229 #define RTV_EXPIRE	0x4	/* init or lock _expire */
230 #define RTV_RPIPE	0x8	/* init or lock _recvpipe */
231 #define RTV_SPIPE	0x10	/* init or lock _sendpipe */
232 #define RTV_SSTHRESH	0x20	/* init or lock _ssthresh */
233 #define RTV_RTT		0x40	/* init or lock _rtt */
234 #define RTV_RTTVAR	0x80	/* init or lock _rttvar */
235 #define RTV_WEIGHT	0x100	/* init or lock _weight */
236 
237 /*
238  * Bitmask values for rtm_addrs.
239  */
240 #define RTA_DST		0x1	/* destination sockaddr present */
241 #define RTA_GATEWAY	0x2	/* gateway sockaddr present */
242 #define RTA_NETMASK	0x4	/* netmask sockaddr present */
243 #define RTA_GENMASK	0x8	/* cloning mask sockaddr present */
244 #define RTA_IFP		0x10	/* interface name sockaddr present */
245 #define RTA_IFA		0x20	/* interface addr sockaddr present */
246 #define RTA_AUTHOR	0x40	/* sockaddr for author of redirect */
247 #define RTA_BRD		0x80	/* for NEWADDR, broadcast or p-p dest addr */
248 
249 /*
250  * Index offsets for sockaddr array for alternate internal encoding.
251  */
252 #define RTAX_DST	0	/* destination sockaddr present */
253 #define RTAX_GATEWAY	1	/* gateway sockaddr present */
254 #define RTAX_NETMASK	2	/* netmask sockaddr present */
255 #define RTAX_GENMASK	3	/* cloning mask sockaddr present */
256 #define RTAX_IFP	4	/* interface name sockaddr present */
257 #define RTAX_IFA	5	/* interface addr sockaddr present */
258 #define RTAX_AUTHOR	6	/* sockaddr for author of redirect */
259 #define RTAX_BRD	7	/* for NEWADDR, broadcast or p-p dest addr */
260 #define RTAX_MAX	8	/* size of array to allocate */
261 
262 struct rt_addrinfo {
263 	int	rti_addrs;
264 	struct	sockaddr *rti_info[RTAX_MAX];
265 	int	rti_flags;
266 	struct	ifaddr *rti_ifa;
267 	struct	ifnet *rti_ifp;
268 	u_long	rti_mflags;
269 	struct	rt_metrics *rti_rmx;
270 };
271 
272 /*
273  * This macro returns the size of a struct sockaddr when passed
274  * through a routing socket. Basically we round up sa_len to
275  * a multiple of sizeof(long), with a minimum of sizeof(long).
276  * The check for a NULL pointer is just a convenience, probably never used.
277  * The case sa_len == 0 should only apply to empty structures.
278  */
279 #define SA_SIZE(sa)						\
280     (  (!(sa) || ((struct sockaddr *)(sa))->sa_len == 0) ?	\
281 	sizeof(long)		:				\
282 	1 + ( (((struct sockaddr *)(sa))->sa_len - 1) | (sizeof(long) - 1) ) )
283 
284 #define	sa_equal(a, b) (	\
285     (((struct sockaddr *)(a))->sa_len == ((struct sockaddr *)(b))->sa_len) && \
286     (bcmp((a), (b), ((struct sockaddr *)(b))->sa_len) == 0))
287 
288 #ifdef _KERNEL
289 
290 #define RT_LINK_IS_UP(ifp)	(!((ifp)->if_capabilities & IFCAP_LINKSTATE) \
291 				 || (ifp)->if_link_state == LINK_STATE_UP)
292 
293 #define	RT_LOCK_INIT(_rt) \
294 	mtx_init(&(_rt)->rt_mtx, "rtentry", NULL, MTX_DEF | MTX_DUPOK)
295 #define	RT_LOCK(_rt)		mtx_lock(&(_rt)->rt_mtx)
296 #define	RT_UNLOCK(_rt)		mtx_unlock(&(_rt)->rt_mtx)
297 #define	RT_LOCK_DESTROY(_rt)	mtx_destroy(&(_rt)->rt_mtx)
298 #define	RT_LOCK_ASSERT(_rt)	mtx_assert(&(_rt)->rt_mtx, MA_OWNED)
299 #define	RT_UNLOCK_COND(_rt)	do {				\
300 	if (mtx_owned(&(_rt)->rt_mtx))				\
301 		mtx_unlock(&(_rt)->rt_mtx);			\
302 } while (0)
303 
304 #define	RT_ADDREF(_rt)	do {					\
305 	RT_LOCK_ASSERT(_rt);					\
306 	KASSERT((_rt)->rt_refcnt >= 0,				\
307 		("negative refcnt %d", (_rt)->rt_refcnt));	\
308 	(_rt)->rt_refcnt++;					\
309 } while (0)
310 
311 #define	RT_REMREF(_rt)	do {					\
312 	RT_LOCK_ASSERT(_rt);					\
313 	KASSERT((_rt)->rt_refcnt > 0,				\
314 		("bogus refcnt %d", (_rt)->rt_refcnt));	\
315 	(_rt)->rt_refcnt--;					\
316 } while (0)
317 
318 #define	RTFREE_LOCKED(_rt) do {					\
319 	if ((_rt)->rt_refcnt <= 1)				\
320 		rtfree(_rt);					\
321 	else {							\
322 		RT_REMREF(_rt);					\
323 		RT_UNLOCK(_rt);					\
324 	}							\
325 	/* guard against invalid refs */			\
326 	_rt = 0;						\
327 } while (0)
328 
329 #define	RTFREE(_rt) do {					\
330 	RT_LOCK(_rt);						\
331 	RTFREE_LOCKED(_rt);					\
332 } while (0)
333 
334 #define	RO_RTFREE(_ro) do {					\
335 	if ((_ro)->ro_rt) {					\
336 		if ((_ro)->ro_flags & RT_NORTREF) {		\
337 			(_ro)->ro_flags &= ~RT_NORTREF;		\
338 			(_ro)->ro_rt = NULL;			\
339 		} else {					\
340 			RT_LOCK((_ro)->ro_rt);			\
341 			RTFREE_LOCKED((_ro)->ro_rt);		\
342 		}						\
343 	}							\
344 } while (0)
345 
346 struct radix_node_head *rt_tables_get_rnh(int, int);
347 
348 struct ifmultiaddr;
349 
350 void	 rt_ieee80211msg(struct ifnet *, int, void *, size_t);
351 void	 rt_ifannouncemsg(struct ifnet *, int);
352 void	 rt_ifmsg(struct ifnet *);
353 void	 rt_missmsg(int, struct rt_addrinfo *, int, int);
354 void	 rt_missmsg_fib(int, struct rt_addrinfo *, int, int, int);
355 void	 rt_newaddrmsg(int, struct ifaddr *, int, struct rtentry *);
356 void	 rt_newaddrmsg_fib(int, struct ifaddr *, int, struct rtentry *, int);
357 int	 rt_addrmsg(int, struct ifaddr *, int);
358 int	 rt_routemsg(int, struct ifnet *ifp, int, struct rtentry *, int);
359 void	 rt_newmaddrmsg(int, struct ifmultiaddr *);
360 int	 rt_setgate(struct rtentry *, struct sockaddr *, struct sockaddr *);
361 void 	 rt_maskedcopy(struct sockaddr *, struct sockaddr *, struct sockaddr *);
362 
363 int	rtsock_addrmsg(int, struct ifaddr *, int);
364 int	rtsock_routemsg(int, struct ifnet *ifp, int, struct rtentry *, int);
365 
366 /*
367  * Note the following locking behavior:
368  *
369  *    rtalloc_ign() and rtalloc() return ro->ro_rt unlocked
370  *
371  *    rtalloc1() returns a locked rtentry
372  *
373  *    rtfree() and RTFREE_LOCKED() require a locked rtentry
374  *
375  *    RTFREE() uses an unlocked entry.
376  */
377 
378 int	 rt_expunge(struct radix_node_head *, struct rtentry *);
379 void	 rtfree(struct rtentry *);
380 int	 rt_check(struct rtentry **, struct rtentry **, struct sockaddr *);
381 void	rt_updatemtu(struct ifnet *);
382 
383 /* XXX MRT COMPAT VERSIONS THAT SET UNIVERSE to 0 */
384 /* Thes are used by old code not yet converted to use multiple FIBS */
385 int	 rt_getifa(struct rt_addrinfo *);
386 void	 rtalloc_ign(struct route *ro, u_long ignflags);
387 void	 rtalloc(struct route *ro); /* XXX deprecated, use rtalloc_ign(ro, 0) */
388 struct rtentry *rtalloc1(struct sockaddr *, int, u_long);
389 int	 rtinit(struct ifaddr *, int, int);
390 int	 rtioctl(u_long, caddr_t);
391 void	 rtredirect(struct sockaddr *, struct sockaddr *,
392 	    struct sockaddr *, int, struct sockaddr *);
393 int	 rtrequest(int, struct sockaddr *,
394 	    struct sockaddr *, struct sockaddr *, int, struct rtentry **);
395 
396 /* XXX MRT NEW VERSIONS THAT USE FIBs
397  * For now the protocol indepedent versions are the same as the AF_INET ones
398  * but this will change..
399  */
400 int	 rt_getifa_fib(struct rt_addrinfo *, u_int fibnum);
401 void	 rtalloc_ign_fib(struct route *ro, u_long ignflags, u_int fibnum);
402 void	 rtalloc_fib(struct route *ro, u_int fibnum);
403 struct rtentry *rtalloc1_fib(struct sockaddr *, int, u_long, u_int);
404 int	 rtioctl_fib(u_long, caddr_t, u_int);
405 void	 rtredirect_fib(struct sockaddr *, struct sockaddr *,
406 	    struct sockaddr *, int, struct sockaddr *, u_int);
407 int	 rtrequest_fib(int, struct sockaddr *,
408 	    struct sockaddr *, struct sockaddr *, int, struct rtentry **, u_int);
409 int	 rtrequest1_fib(int, struct rt_addrinfo *, struct rtentry **, u_int);
410 
411 #include <sys/eventhandler.h>
412 typedef void (*rtevent_redirect_fn)(void *, struct rtentry *, struct rtentry *, struct sockaddr *);
413 EVENTHANDLER_DECLARE(route_redirect_event, rtevent_redirect_fn);
414 #endif
415 
416 #endif
417