route.h revision 263478
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: stable/10/sys/net/route.h 263478 2014-03-21 15:15:30Z glebius $
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 */
51struct 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
62struct 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 */
86extern u_int rt_numfibs;	/* number fo usable routing tables */
87/*
88 * XXX kernel function pointer `rt_output' is visible to applications.
89 */
90struct mbuf;
91
92/*
93 * We distinguish between routes to hosts and routes to networks,
94 * preferring the former if available.  For each route we infer
95 * the interface to use from the gateway address supplied when
96 * the route was entered.  Routes that forward packets through
97 * gateways are marked so that the output routines know to address the
98 * gateway rather than the ultimate destination.
99 */
100#ifndef RNF_NORMAL
101#include <net/radix.h>
102#ifdef RADIX_MPATH
103#include <net/radix_mpath.h>
104#endif
105#endif
106
107#if defined(_KERNEL) || defined(_WANT_RTENTRY)
108struct rtentry {
109	struct	radix_node rt_nodes[2];	/* tree glue, and other values */
110	/*
111	 * XXX struct rtentry must begin with a struct radix_node (or two!)
112	 * because the code does some casts of a 'struct radix_node *'
113	 * to a 'struct rtentry *'
114	 */
115#define	rt_key(r)	(*((struct sockaddr **)(&(r)->rt_nodes->rn_key)))
116#define	rt_mask(r)	(*((struct sockaddr **)(&(r)->rt_nodes->rn_mask)))
117	struct	sockaddr *rt_gateway;	/* value */
118	struct	ifnet *rt_ifp;		/* the answer: interface to use */
119	struct	ifaddr *rt_ifa;		/* the answer: interface address to use */
120	int		rt_flags;	/* up/down?, host/net */
121	int		rt_refcnt;	/* # held references */
122	u_int		rt_fibnum;	/* which FIB */
123	u_long		rt_mtu;		/* MTU for this path */
124	u_long		rt_weight;	/* absolute weight */
125	u_long		rt_expire;	/* lifetime for route, e.g. redirect */
126#define	rt_endzero	rt_pksent
127	counter_u64_t	rt_pksent;	/* packets sent using this route */
128	struct mtx	rt_mtx;		/* mutex for routing entry */
129};
130#endif /* _KERNEL || _WANT_RTENTRY */
131
132/*
133 * Following structure necessary for 4.3 compatibility;
134 * We should eventually move it to a compat file.
135 */
136struct ortentry {
137	u_long	rt_hash;		/* to speed lookups */
138	struct	sockaddr rt_dst;	/* key */
139	struct	sockaddr rt_gateway;	/* value */
140	short	rt_flags;		/* up/down?, host/net */
141	short	rt_refcnt;		/* # held references */
142	u_long	rt_use;			/* raw # packets forwarded */
143	struct	ifnet *rt_ifp;		/* the answer: interface to use */
144};
145
146#define	RTF_UP		0x1		/* route usable */
147#define	RTF_GATEWAY	0x2		/* destination is a gateway */
148#define	RTF_HOST	0x4		/* host entry (net otherwise) */
149#define	RTF_REJECT	0x8		/* host or net unreachable */
150#define	RTF_DYNAMIC	0x10		/* created dynamically (by redirect) */
151#define	RTF_MODIFIED	0x20		/* modified dynamically (by redirect) */
152#define RTF_DONE	0x40		/* message confirmed */
153/*			0x80		   unused, was RTF_DELCLONE */
154/*			0x100		   unused, was RTF_CLONING */
155#define RTF_XRESOLVE	0x200		/* external daemon resolves name */
156#define RTF_LLINFO	0x400		/* DEPRECATED - exists ONLY for backward
157					   compatibility */
158#define RTF_LLDATA	0x400		/* used by apps to add/del L2 entries */
159#define RTF_STATIC	0x800		/* manually added */
160#define RTF_BLACKHOLE	0x1000		/* just discard pkts (during updates) */
161#define RTF_PROTO2	0x4000		/* protocol specific routing flag */
162#define RTF_PROTO1	0x8000		/* protocol specific routing flag */
163
164/* XXX: temporary to stay API/ABI compatible with userland */
165#ifndef _KERNEL
166#define RTF_PRCLONING	0x10000		/* unused, for compatibility */
167#endif
168
169/*			0x20000		   unused, was RTF_WASCLONED */
170#define RTF_PROTO3	0x40000		/* protocol specific routing flag */
171/*			0x80000		   unused */
172#define RTF_PINNED	0x100000	/* route is immutable */
173#define	RTF_LOCAL	0x200000 	/* route represents a local address */
174#define	RTF_BROADCAST	0x400000	/* route represents a bcast address */
175#define	RTF_MULTICAST	0x800000	/* route represents a mcast address */
176					/* 0x8000000 and up unassigned */
177#define	RTF_STICKY	 0x10000000	/* always route dst->src */
178
179#define	RTF_RNH_LOCKED	 0x40000000	/* radix node head is locked */
180
181#define	RTF_GWFLAG_COMPAT 0x80000000	/* a compatibility bit for interacting
182					   with existing routing apps */
183
184/* Mask of RTF flags that are allowed to be modified by RTM_CHANGE. */
185#define RTF_FMASK	\
186	(RTF_PROTO1 | RTF_PROTO2 | RTF_PROTO3 | RTF_BLACKHOLE | \
187	 RTF_REJECT | RTF_STATIC | RTF_STICKY)
188
189/*
190 * Routing statistics.
191 */
192struct	rtstat {
193	short	rts_badredirect;	/* bogus redirect calls */
194	short	rts_dynamic;		/* routes created by redirects */
195	short	rts_newgateway;		/* routes modified by redirects */
196	short	rts_unreach;		/* lookups which failed */
197	short	rts_wildcard;		/* lookups satisfied by a wildcard */
198};
199/*
200 * Structures for routing messages.
201 */
202struct rt_msghdr {
203	u_short	rtm_msglen;	/* to skip over non-understood messages */
204	u_char	rtm_version;	/* future binary compatibility */
205	u_char	rtm_type;	/* message type */
206	u_short	rtm_index;	/* index for associated ifp */
207	int	rtm_flags;	/* flags, incl. kern & message, e.g. DONE */
208	int	rtm_addrs;	/* bitmask identifying sockaddrs in msg */
209	pid_t	rtm_pid;	/* identify sender */
210	int	rtm_seq;	/* for sender to identify action */
211	int	rtm_errno;	/* why failed */
212	int	rtm_fmask;	/* bitmask used in RTM_CHANGE message */
213	u_long	rtm_inits;	/* which metrics we are initializing */
214	struct	rt_metrics rtm_rmx; /* metrics themselves */
215};
216
217#define RTM_VERSION	5	/* Up the ante and ignore older versions */
218
219/*
220 * Message types.
221 */
222#define RTM_ADD		0x1	/* Add Route */
223#define RTM_DELETE	0x2	/* Delete Route */
224#define RTM_CHANGE	0x3	/* Change Metrics or flags */
225#define RTM_GET		0x4	/* Report Metrics */
226#define RTM_LOSING	0x5	/* Kernel Suspects Partitioning */
227#define RTM_REDIRECT	0x6	/* Told to use different route */
228#define RTM_MISS	0x7	/* Lookup failed on this address */
229#define RTM_LOCK	0x8	/* fix specified metrics */
230#define RTM_OLDADD	0x9	/* caused by SIOCADDRT */
231#define RTM_OLDDEL	0xa	/* caused by SIOCDELRT */
232#define RTM_RESOLVE	0xb	/* req to resolve dst to LL addr */
233#define RTM_NEWADDR	0xc	/* address being added to iface */
234#define RTM_DELADDR	0xd	/* address being removed from iface */
235#define RTM_IFINFO	0xe	/* iface going up/down etc. */
236#define	RTM_NEWMADDR	0xf	/* mcast group membership being added to if */
237#define	RTM_DELMADDR	0x10	/* mcast group membership being deleted */
238#define	RTM_IFANNOUNCE	0x11	/* iface arrival/departure */
239#define	RTM_IEEE80211	0x12	/* IEEE80211 wireless event */
240
241/*
242 * Bitmask values for rtm_inits and rmx_locks.
243 */
244#define RTV_MTU		0x1	/* init or lock _mtu */
245#define RTV_HOPCOUNT	0x2	/* init or lock _hopcount */
246#define RTV_EXPIRE	0x4	/* init or lock _expire */
247#define RTV_RPIPE	0x8	/* init or lock _recvpipe */
248#define RTV_SPIPE	0x10	/* init or lock _sendpipe */
249#define RTV_SSTHRESH	0x20	/* init or lock _ssthresh */
250#define RTV_RTT		0x40	/* init or lock _rtt */
251#define RTV_RTTVAR	0x80	/* init or lock _rttvar */
252#define RTV_WEIGHT	0x100	/* init or lock _weight */
253
254/*
255 * Bitmask values for rtm_addrs.
256 */
257#define RTA_DST		0x1	/* destination sockaddr present */
258#define RTA_GATEWAY	0x2	/* gateway sockaddr present */
259#define RTA_NETMASK	0x4	/* netmask sockaddr present */
260#define RTA_GENMASK	0x8	/* cloning mask sockaddr present */
261#define RTA_IFP		0x10	/* interface name sockaddr present */
262#define RTA_IFA		0x20	/* interface addr sockaddr present */
263#define RTA_AUTHOR	0x40	/* sockaddr for author of redirect */
264#define RTA_BRD		0x80	/* for NEWADDR, broadcast or p-p dest addr */
265
266/*
267 * Index offsets for sockaddr array for alternate internal encoding.
268 */
269#define RTAX_DST	0	/* destination sockaddr present */
270#define RTAX_GATEWAY	1	/* gateway sockaddr present */
271#define RTAX_NETMASK	2	/* netmask sockaddr present */
272#define RTAX_GENMASK	3	/* cloning mask sockaddr present */
273#define RTAX_IFP	4	/* interface name sockaddr present */
274#define RTAX_IFA	5	/* interface addr sockaddr present */
275#define RTAX_AUTHOR	6	/* sockaddr for author of redirect */
276#define RTAX_BRD	7	/* for NEWADDR, broadcast or p-p dest addr */
277#define RTAX_MAX	8	/* size of array to allocate */
278
279struct rt_addrinfo {
280	int	rti_addrs;
281	struct	sockaddr *rti_info[RTAX_MAX];
282	int	rti_flags;
283	struct	ifaddr *rti_ifa;
284	struct	ifnet *rti_ifp;
285};
286
287/*
288 * This macro returns the size of a struct sockaddr when passed
289 * through a routing socket. Basically we round up sa_len to
290 * a multiple of sizeof(long), with a minimum of sizeof(long).
291 * The check for a NULL pointer is just a convenience, probably never used.
292 * The case sa_len == 0 should only apply to empty structures.
293 */
294#define SA_SIZE(sa)						\
295    (  (!(sa) || ((struct sockaddr *)(sa))->sa_len == 0) ?	\
296	sizeof(long)		:				\
297	1 + ( (((struct sockaddr *)(sa))->sa_len - 1) | (sizeof(long) - 1) ) )
298
299#ifdef _KERNEL
300
301#define RT_LINK_IS_UP(ifp)	(!((ifp)->if_capabilities & IFCAP_LINKSTATE) \
302				 || (ifp)->if_link_state == LINK_STATE_UP)
303
304#define	RT_LOCK_INIT(_rt) \
305	mtx_init(&(_rt)->rt_mtx, "rtentry", NULL, MTX_DEF | MTX_DUPOK)
306#define	RT_LOCK(_rt)		mtx_lock(&(_rt)->rt_mtx)
307#define	RT_UNLOCK(_rt)		mtx_unlock(&(_rt)->rt_mtx)
308#define	RT_LOCK_DESTROY(_rt)	mtx_destroy(&(_rt)->rt_mtx)
309#define	RT_LOCK_ASSERT(_rt)	mtx_assert(&(_rt)->rt_mtx, MA_OWNED)
310#define	RT_UNLOCK_COND(_rt)	do {				\
311	if (mtx_owned(&(_rt)->rt_mtx))				\
312		mtx_unlock(&(_rt)->rt_mtx);			\
313} while (0)
314
315#define	RT_ADDREF(_rt)	do {					\
316	RT_LOCK_ASSERT(_rt);					\
317	KASSERT((_rt)->rt_refcnt >= 0,				\
318		("negative refcnt %d", (_rt)->rt_refcnt));	\
319	(_rt)->rt_refcnt++;					\
320} while (0)
321
322#define	RT_REMREF(_rt)	do {					\
323	RT_LOCK_ASSERT(_rt);					\
324	KASSERT((_rt)->rt_refcnt > 0,				\
325		("bogus refcnt %d", (_rt)->rt_refcnt));	\
326	(_rt)->rt_refcnt--;					\
327} while (0)
328
329#define	RTFREE_LOCKED(_rt) do {					\
330	if ((_rt)->rt_refcnt <= 1)				\
331		rtfree(_rt);					\
332	else {							\
333		RT_REMREF(_rt);					\
334		RT_UNLOCK(_rt);					\
335	}							\
336	/* guard against invalid refs */			\
337	_rt = 0;						\
338} while (0)
339
340#define	RTFREE(_rt) do {					\
341	RT_LOCK(_rt);						\
342	RTFREE_LOCKED(_rt);					\
343} while (0)
344
345#define	RO_RTFREE(_ro) do {					\
346	if ((_ro)->ro_rt) {					\
347		if ((_ro)->ro_flags & RT_NORTREF) {		\
348			(_ro)->ro_flags &= ~RT_NORTREF;		\
349			(_ro)->ro_rt = NULL;			\
350		} else {					\
351			RT_LOCK((_ro)->ro_rt);			\
352			RTFREE_LOCKED((_ro)->ro_rt);		\
353		}						\
354	}							\
355} while (0)
356
357struct radix_node_head *rt_tables_get_rnh(int, int);
358
359struct ifmultiaddr;
360
361void	 rt_ieee80211msg(struct ifnet *, int, void *, size_t);
362void	 rt_ifannouncemsg(struct ifnet *, int);
363void	 rt_ifmsg(struct ifnet *);
364void	 rt_missmsg(int, struct rt_addrinfo *, int, int);
365void	 rt_missmsg_fib(int, struct rt_addrinfo *, int, int, int);
366void	 rt_newaddrmsg(int, struct ifaddr *, int, struct rtentry *);
367void	 rt_newaddrmsg_fib(int, struct ifaddr *, int, struct rtentry *, int);
368void	 rt_newmaddrmsg(int, struct ifmultiaddr *);
369int	 rt_setgate(struct rtentry *, struct sockaddr *, struct sockaddr *);
370void 	 rt_maskedcopy(struct sockaddr *, struct sockaddr *, struct sockaddr *);
371
372/*
373 * Note the following locking behavior:
374 *
375 *    rtalloc_ign() and rtalloc() return ro->ro_rt unlocked
376 *
377 *    rtalloc1() returns a locked rtentry
378 *
379 *    rtfree() and RTFREE_LOCKED() require a locked rtentry
380 *
381 *    RTFREE() uses an unlocked entry.
382 */
383
384int	 rtexpunge(struct rtentry *);
385void	 rtfree(struct rtentry *);
386int	 rt_check(struct rtentry **, struct rtentry **, struct sockaddr *);
387
388/* XXX MRT COMPAT VERSIONS THAT SET UNIVERSE to 0 */
389/* Thes are used by old code not yet converted to use multiple FIBS */
390int	 rt_getifa(struct rt_addrinfo *);
391void	 rtalloc_ign(struct route *ro, u_long ignflags);
392void	 rtalloc(struct route *ro); /* XXX deprecated, use rtalloc_ign(ro, 0) */
393struct rtentry *rtalloc1(struct sockaddr *, int, u_long);
394int	 rtinit(struct ifaddr *, int, int);
395int	 rtioctl(u_long, caddr_t);
396void	 rtredirect(struct sockaddr *, struct sockaddr *,
397	    struct sockaddr *, int, struct sockaddr *);
398int	 rtrequest(int, struct sockaddr *,
399	    struct sockaddr *, struct sockaddr *, int, struct rtentry **);
400
401#ifndef BURN_BRIDGES
402/* defaults to "all" FIBs */
403int	 rtinit_fib(struct ifaddr *, int, int);
404#endif
405
406/* XXX MRT NEW VERSIONS THAT USE FIBs
407 * For now the protocol indepedent versions are the same as the AF_INET ones
408 * but this will change..
409 */
410int	 rt_getifa_fib(struct rt_addrinfo *, u_int fibnum);
411void	 rtalloc_ign_fib(struct route *ro, u_long ignflags, u_int fibnum);
412void	 rtalloc_fib(struct route *ro, u_int fibnum);
413struct rtentry *rtalloc1_fib(struct sockaddr *, int, u_long, u_int);
414int	 rtioctl_fib(u_long, caddr_t, u_int);
415void	 rtredirect_fib(struct sockaddr *, struct sockaddr *,
416	    struct sockaddr *, int, struct sockaddr *, u_int);
417int	 rtrequest_fib(int, struct sockaddr *,
418	    struct sockaddr *, struct sockaddr *, int, struct rtentry **, u_int);
419int	 rtrequest1_fib(int, struct rt_addrinfo *, struct rtentry **, u_int);
420
421#include <sys/eventhandler.h>
422typedef void (*rtevent_arp_update_fn)(void *, struct rtentry *, uint8_t *, struct sockaddr *);
423typedef void (*rtevent_redirect_fn)(void *, struct rtentry *, struct rtentry *, struct sockaddr *);
424/* route_arp_update_event is no longer generated; see arp_update_event */
425EVENTHANDLER_DECLARE(route_arp_update_event, rtevent_arp_update_fn);
426EVENTHANDLER_DECLARE(route_redirect_event, rtevent_redirect_fn);
427#endif
428
429#endif
430