xref: /freebsd/sys/net/route.h (revision 9e5787d2284e187abb5b654d924394a65772e004)
1 /*-
2  * SPDX-License-Identifier: BSD-3-Clause
3  *
4  * Copyright (c) 1980, 1986, 1993
5  *	The Regents of the University of California.  All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  * 3. Neither the name of the University nor the names of its contributors
16  *    may be used to endorse or promote products derived from this software
17  *    without specific prior written permission.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29  * SUCH DAMAGE.
30  *
31  *	@(#)route.h	8.4 (Berkeley) 1/9/95
32  * $FreeBSD$
33  */
34 
35 #ifndef _NET_ROUTE_H_
36 #define _NET_ROUTE_H_
37 
38 #include <net/vnet.h>
39 
40 /*
41  * Kernel resident routing tables.
42  *
43  * The routing tables are initialized when interface addresses
44  * are set by making entries for all directly connected interfaces.
45  */
46 
47 /*
48  * Struct route consiste of a destination address,
49  * a route entry pointer, link-layer prepend data pointer along
50  * with its length.
51  */
52 struct route {
53 	struct	nhop_object *ro_nh;
54 	struct	llentry *ro_lle;
55 	/*
56 	 * ro_prepend and ro_plen are only used for bpf to pass in a
57 	 * preformed header.  They are not cacheable.
58 	 */
59 	char		*ro_prepend;
60 	uint16_t	ro_plen;
61 	uint16_t	ro_flags;
62 	uint16_t	ro_mtu;	/* saved ro_rt mtu */
63 	uint16_t	spare;
64 	struct	sockaddr ro_dst;
65 };
66 
67 #define	RT_L2_ME_BIT		2	/* dst L2 addr is our address */
68 #define	RT_MAY_LOOP_BIT		3	/* dst may require loop copy */
69 #define	RT_HAS_HEADER_BIT	4	/* mbuf already have its header prepended */
70 
71 #define	RT_L2_ME		(1 << RT_L2_ME_BIT)		/* 0x0004 */
72 #define	RT_MAY_LOOP		(1 << RT_MAY_LOOP_BIT)		/* 0x0008 */
73 #define	RT_HAS_HEADER		(1 << RT_HAS_HEADER_BIT)	/* 0x0010 */
74 
75 #define	RT_REJECT		0x0020		/* Destination is reject */
76 #define	RT_BLACKHOLE		0x0040		/* Destination is blackhole */
77 #define	RT_HAS_GW		0x0080		/* Destination has GW  */
78 #define	RT_LLE_CACHE		0x0100		/* Cache link layer  */
79 
80 struct rt_metrics {
81 	u_long	rmx_locks;	/* Kernel must leave these values alone */
82 	u_long	rmx_mtu;	/* MTU for this path */
83 	u_long	rmx_hopcount;	/* max hops expected */
84 	u_long	rmx_expire;	/* lifetime for route, e.g. redirect */
85 	u_long	rmx_recvpipe;	/* inbound delay-bandwidth product */
86 	u_long	rmx_sendpipe;	/* outbound delay-bandwidth product */
87 	u_long	rmx_ssthresh;	/* outbound gateway buffer limit */
88 	u_long	rmx_rtt;	/* estimated round trip time */
89 	u_long	rmx_rttvar;	/* estimated rtt variance */
90 	u_long	rmx_pksent;	/* packets sent using this route */
91 	u_long	rmx_weight;	/* route weight */
92 	u_long	rmx_nhidx;	/* route nexhop index */
93 	u_long	rmx_filler[2];	/* will be used for T/TCP later */
94 };
95 
96 /*
97  * rmx_rtt and rmx_rttvar are stored as microseconds;
98  * RTTTOPRHZ(rtt) converts to a value suitable for use
99  * by a protocol slowtimo counter.
100  */
101 #define	RTM_RTTUNIT	1000000	/* units for rtt, rttvar, as units per sec */
102 #define	RTTTOPRHZ(r)	((r) / (RTM_RTTUNIT / PR_SLOWHZ))
103 
104 /* lle state is exported in rmx_state rt_metrics field */
105 #define	rmx_state	rmx_weight
106 
107 /*
108  * Keep a generation count of routing table, incremented on route addition,
109  * so we can invalidate caches.  This is accessed without a lock, as precision
110  * is not required.
111  */
112 typedef volatile u_int rt_gen_t;	/* tree generation (for adds) */
113 #define RT_GEN(fibnum, af)	rt_tables_get_gen(fibnum, af)
114 
115 #define	RT_DEFAULT_FIB	0	/* Explicitly mark fib=0 restricted cases */
116 #define	RT_ALL_FIBS	-1	/* Announce event for every fib */
117 #ifdef _KERNEL
118 VNET_DECLARE(uint32_t, _rt_numfibs);	/* number of existing route tables */
119 #define	V_rt_numfibs		VNET(_rt_numfibs)
120 /* temporary compat arg */
121 #define	rt_numfibs		V_rt_numfibs
122 VNET_DECLARE(u_int, rt_add_addr_allfibs); /* Announce interfaces to all fibs */
123 #define	V_rt_add_addr_allfibs	VNET(rt_add_addr_allfibs)
124 #endif
125 
126 /*
127  * We distinguish between routes to hosts and routes to networks,
128  * preferring the former if available.  For each route we infer
129  * the interface to use from the gateway address supplied when
130  * the route was entered.  Routes that forward packets through
131  * gateways are marked so that the output routines know to address the
132  * gateway rather than the ultimate destination.
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  * fib_ nexthop API flags.
174  */
175 
176 /* Consumer-visible nexthop info flags */
177 #define	NHF_REJECT		0x0010	/* RTF_REJECT */
178 #define	NHF_BLACKHOLE		0x0020	/* RTF_BLACKHOLE */
179 #define	NHF_REDIRECT		0x0040	/* RTF_DYNAMIC|RTF_MODIFIED */
180 #define	NHF_DEFAULT		0x0080	/* Default route */
181 #define	NHF_BROADCAST		0x0100	/* RTF_BROADCAST */
182 #define	NHF_GATEWAY		0x0200	/* RTF_GATEWAY */
183 #define	NHF_HOST		0x0400	/* RTF_HOST */
184 
185 /* Nexthop request flags */
186 #define	NHR_NONE		0x00	/* empty flags field */
187 #define	NHR_IFAIF		0x01	/* Return ifa_ifp interface */
188 #define	NHR_REF			0x02	/* For future use */
189 
190 /* uRPF */
191 #define	NHR_NODEFAULT		0x04	/* do not consider default route */
192 
193 /* Control plane route request flags */
194 #define	NHR_COPY		0x100	/* Copy rte data */
195 
196 /*
197  * Routing statistics.
198  */
199 struct rtstat {
200 	uint64_t rts_badredirect;	/* bogus redirect calls */
201 	uint64_t rts_dynamic;		/* routes created by redirects */
202 	uint64_t rts_newgateway;	/* routes modified by redirects */
203 	uint64_t rts_unreach;		/* lookups which failed */
204 	uint64_t rts_wildcard;		/* lookups satisfied by a wildcard */
205 	uint64_t rts_nh_idx_alloc_failure;	/* nexthop index alloc failure*/
206 	uint64_t rts_nh_alloc_failure;	/* nexthop allocation failure*/
207 };
208 
209 /*
210  * Structures for routing messages.
211  */
212 struct rt_msghdr {
213 	u_short	rtm_msglen;	/* to skip over non-understood messages */
214 	u_char	rtm_version;	/* future binary compatibility */
215 	u_char	rtm_type;	/* message type */
216 	u_short	rtm_index;	/* index for associated ifp */
217 	u_short _rtm_spare1;
218 	int	rtm_flags;	/* flags, incl. kern & message, e.g. DONE */
219 	int	rtm_addrs;	/* bitmask identifying sockaddrs in msg */
220 	pid_t	rtm_pid;	/* identify sender */
221 	int	rtm_seq;	/* for sender to identify action */
222 	int	rtm_errno;	/* why failed */
223 	int	rtm_fmask;	/* bitmask used in RTM_CHANGE message */
224 	u_long	rtm_inits;	/* which metrics we are initializing */
225 	struct	rt_metrics rtm_rmx; /* metrics themselves */
226 };
227 
228 #define RTM_VERSION	5	/* Up the ante and ignore older versions */
229 
230 /*
231  * Message types.
232  *
233  * The format for each message is annotated below using the following
234  * identifiers:
235  *
236  * (1) struct rt_msghdr
237  * (2) struct ifa_msghdr
238  * (3) struct if_msghdr
239  * (4) struct ifma_msghdr
240  * (5) struct if_announcemsghdr
241  *
242  */
243 #define	RTM_ADD		0x1	/* (1) Add Route */
244 #define	RTM_DELETE	0x2	/* (1) Delete Route */
245 #define	RTM_CHANGE	0x3	/* (1) Change Metrics or flags */
246 #define	RTM_GET		0x4	/* (1) Report Metrics */
247 #define	RTM_LOSING	0x5	/* (1) Kernel Suspects Partitioning */
248 #define	RTM_REDIRECT	0x6	/* (1) Told to use different route */
249 #define	RTM_MISS	0x7	/* (1) Lookup failed on this address */
250 #define	RTM_LOCK	0x8	/* (1) fix specified metrics */
251 		    /*	0x9  */
252 		    /*	0xa  */
253 #define	RTM_RESOLVE	0xb	/* (1) req to resolve dst to LL addr */
254 #define	RTM_NEWADDR	0xc	/* (2) address being added to iface */
255 #define	RTM_DELADDR	0xd	/* (2) address being removed from iface */
256 #define	RTM_IFINFO	0xe	/* (3) iface going up/down etc. */
257 #define	RTM_NEWMADDR	0xf	/* (4) mcast group membership being added to if */
258 #define	RTM_DELMADDR	0x10	/* (4) mcast group membership being deleted */
259 #define	RTM_IFANNOUNCE	0x11	/* (5) iface arrival/departure */
260 #define	RTM_IEEE80211	0x12	/* (5) IEEE80211 wireless event */
261 
262 /*
263  * Bitmask values for rtm_inits and rmx_locks.
264  */
265 #define RTV_MTU		0x1	/* init or lock _mtu */
266 #define RTV_HOPCOUNT	0x2	/* init or lock _hopcount */
267 #define RTV_EXPIRE	0x4	/* init or lock _expire */
268 #define RTV_RPIPE	0x8	/* init or lock _recvpipe */
269 #define RTV_SPIPE	0x10	/* init or lock _sendpipe */
270 #define RTV_SSTHRESH	0x20	/* init or lock _ssthresh */
271 #define RTV_RTT		0x40	/* init or lock _rtt */
272 #define RTV_RTTVAR	0x80	/* init or lock _rttvar */
273 #define RTV_WEIGHT	0x100	/* init or lock _weight */
274 
275 /*
276  * Bitmask values for rtm_addrs.
277  */
278 #define RTA_DST		0x1	/* destination sockaddr present */
279 #define RTA_GATEWAY	0x2	/* gateway sockaddr present */
280 #define RTA_NETMASK	0x4	/* netmask sockaddr present */
281 #define RTA_GENMASK	0x8	/* cloning mask sockaddr present */
282 #define RTA_IFP		0x10	/* interface name sockaddr present */
283 #define RTA_IFA		0x20	/* interface addr sockaddr present */
284 #define RTA_AUTHOR	0x40	/* sockaddr for author of redirect */
285 #define RTA_BRD		0x80	/* for NEWADDR, broadcast or p-p dest addr */
286 
287 /*
288  * Index offsets for sockaddr array for alternate internal encoding.
289  */
290 #define RTAX_DST	0	/* destination sockaddr present */
291 #define RTAX_GATEWAY	1	/* gateway sockaddr present */
292 #define RTAX_NETMASK	2	/* netmask sockaddr present */
293 #define RTAX_GENMASK	3	/* cloning mask sockaddr present */
294 #define RTAX_IFP	4	/* interface name sockaddr present */
295 #define RTAX_IFA	5	/* interface addr sockaddr present */
296 #define RTAX_AUTHOR	6	/* sockaddr for author of redirect */
297 #define RTAX_BRD	7	/* for NEWADDR, broadcast or p-p dest addr */
298 #define RTAX_MAX	8	/* size of array to allocate */
299 
300 struct rtentry;
301 struct nhop_object;
302 typedef int rt_filter_f_t(const struct rtentry *, const struct nhop_object *,
303     void *);
304 
305 struct rt_addrinfo {
306 	int	rti_addrs;			/* Route RTF_ flags */
307 	int	rti_flags;			/* Route RTF_ flags */
308 	struct	sockaddr *rti_info[RTAX_MAX];	/* Sockaddr data */
309 	struct	ifaddr *rti_ifa;		/* value of rt_ifa addr */
310 	struct	ifnet *rti_ifp;			/* route interface */
311 	rt_filter_f_t	*rti_filter;		/* filter function */
312 	void	*rti_filterdata;		/* filter paramenters */
313 	u_long	rti_mflags;			/* metrics RTV_ flags */
314 	u_long	rti_spare;			/* Will be used for fib */
315 	struct	rt_metrics *rti_rmx;		/* Pointer to route metrics */
316 };
317 
318 /*
319  * This macro returns the size of a struct sockaddr when passed
320  * through a routing socket. Basically we round up sa_len to
321  * a multiple of sizeof(long), with a minimum of sizeof(long).
322  * The case sa_len == 0 should only apply to empty structures.
323  */
324 #define SA_SIZE(sa)						\
325     (  (((struct sockaddr *)(sa))->sa_len == 0) ?		\
326 	sizeof(long)		:				\
327 	1 + ( (((struct sockaddr *)(sa))->sa_len - 1) | (sizeof(long) - 1) ) )
328 
329 #define	sa_equal(a, b) (	\
330     (((const struct sockaddr *)(a))->sa_len == ((const struct sockaddr *)(b))->sa_len) && \
331     (bcmp((a), (b), ((const struct sockaddr *)(b))->sa_len) == 0))
332 
333 #ifdef _KERNEL
334 
335 #define RT_LINK_IS_UP(ifp)	(!((ifp)->if_capabilities & IFCAP_LINKSTATE) \
336 				 || (ifp)->if_link_state == LINK_STATE_UP)
337 
338 #define	RO_NHFREE(_ro) do {					\
339 	if ((_ro)->ro_nh) {					\
340 		NH_FREE((_ro)->ro_nh);				\
341 		(_ro)->ro_nh = NULL;				\
342 	}							\
343 } while (0)
344 
345 #define	RO_INVALIDATE_CACHE(ro) do {					\
346 		if ((ro)->ro_lle != NULL) {				\
347 			LLE_FREE((ro)->ro_lle);				\
348 			(ro)->ro_lle = NULL;				\
349 		}							\
350 		if ((ro)->ro_nh != NULL) {				\
351 			NH_FREE((ro)->ro_nh);				\
352 			(ro)->ro_nh = NULL;				\
353 		}							\
354 	} while (0)
355 
356 /*
357  * Validate a cached route based on a supplied cookie.  If there is an
358  * out-of-date cache, simply free it.  Update the generation number
359  * for the new allocation
360  */
361 #define NH_VALIDATE(ro, cookiep, fibnum) do {				\
362 	rt_gen_t cookie = RT_GEN(fibnum, (ro)->ro_dst.sa_family);	\
363 	if (*(cookiep) != cookie) {					\
364 		RO_INVALIDATE_CACHE(ro);				\
365 		*(cookiep) = cookie;					\
366 	}								\
367 } while (0)
368 
369 struct ifmultiaddr;
370 struct rib_head;
371 
372 void	 rt_ieee80211msg(struct ifnet *, int, void *, size_t);
373 void	 rt_ifannouncemsg(struct ifnet *, int);
374 void	 rt_ifmsg(struct ifnet *);
375 void	 rt_missmsg(int, struct rt_addrinfo *, int, int);
376 void	 rt_missmsg_fib(int, struct rt_addrinfo *, int, int, int);
377 void	 rt_newaddrmsg_fib(int, struct ifaddr *, struct rtentry *, int);
378 int	 rt_addrmsg(int, struct ifaddr *, int);
379 int	 rt_routemsg(int, struct rtentry *, struct ifnet *ifp, int, int);
380 int	 rt_routemsg_info(int, struct rt_addrinfo *, int);
381 void	 rt_newmaddrmsg(int, struct ifmultiaddr *);
382 void 	 rt_maskedcopy(struct sockaddr *, struct sockaddr *, struct sockaddr *);
383 struct rib_head *rt_table_init(int, int, u_int);
384 void	rt_table_destroy(struct rib_head *);
385 u_int	rt_tables_get_gen(uint32_t table, sa_family_t family);
386 
387 int	rtsock_addrmsg(int, struct ifaddr *, int);
388 int	rtsock_routemsg(int, struct rtentry *, struct ifnet *ifp, int, int);
389 int	rtsock_routemsg_info(int, struct rt_addrinfo *, int);
390 
391 struct sockaddr *rtsock_fix_netmask(const struct sockaddr *dst,
392 	    const struct sockaddr *smask, struct sockaddr_storage *dmask);
393 
394 void	rt_updatemtu(struct ifnet *);
395 
396 void	rt_flushifroutes_af(struct ifnet *, int);
397 void	rt_flushifroutes(struct ifnet *ifp);
398 
399 /* XXX MRT COMPAT VERSIONS THAT SET UNIVERSE to 0 */
400 /* Thes are used by old code not yet converted to use multiple FIBS */
401 int	 rtinit(struct ifaddr *, int, int);
402 
403 /* XXX MRT NEW VERSIONS THAT USE FIBs
404  * For now the protocol indepedent versions are the same as the AF_INET ones
405  * but this will change..
406  */
407 int	 rtioctl_fib(u_long, caddr_t, u_int);
408 int	rib_lookup_info(uint32_t, const struct sockaddr *, uint32_t, uint32_t,
409 	    struct rt_addrinfo *);
410 void	rib_free_info(struct rt_addrinfo *info);
411 
412 /* New API */
413 struct nhop_object *rib_lookup(uint32_t fibnum, const struct sockaddr *dst,
414 	    uint32_t flags, uint32_t flowid);
415 #endif
416 
417 #endif
418