route.h revision 187094
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: head/sys/net/route.h 187094 2009-01-12 11:24:32Z qingli $
31 */
32
33#ifndef _NET_ROUTE_H_
34#define _NET_ROUTE_H_
35
36/*
37 * Kernel resident routing tables.
38 *
39 * The routing tables are initialized when interface addresses
40 * are set by making entries for all directly connected interfaces.
41 */
42
43/*
44 * A route consists of a destination address and a reference
45 * to a routing entry.  These are often held by protocols
46 * in their control blocks, e.g. inpcb.
47 */
48struct route {
49	struct	rtentry *ro_rt;
50	struct	sockaddr ro_dst;
51};
52
53/*
54 * These numbers are used by reliable protocols for determining
55 * retransmission behavior and are included in the routing structure.
56 */
57struct rt_metrics_lite {
58	u_long	rmx_mtu;	/* MTU for this path */
59	u_long	rmx_expire;	/* lifetime for route, e.g. redirect */
60	u_long	rmx_pksent;	/* packets sent using this route */
61};
62
63struct 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_filler[4];	/* 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/* MRT compile-time constants */
86#ifdef _KERNEL
87 #ifndef ROUTETABLES
88  #define RT_NUMFIBS 1
89  #define RT_MAXFIBS 1
90 #else
91  /* while we use 4 bits in the mbuf flags, we are limited to 16 */
92  #define RT_MAXFIBS 16
93  #if ROUTETABLES > RT_MAXFIBS
94   #define RT_NUMFIBS RT_MAXFIBS
95   #error "ROUTETABLES defined too big"
96  #else
97   #if ROUTETABLES == 0
98    #define RT_NUMFIBS 1
99   #else
100    #define RT_NUMFIBS ROUTETABLES
101   #endif
102  #endif
103 #endif
104#endif
105
106extern u_int rt_numfibs;	/* number fo usable routing tables */
107extern u_int tunnel_fib;	/* tunnels use these */
108extern u_int fwd_fib;		/* packets being forwarded use these routes */
109/*
110 * XXX kernel function pointer `rt_output' is visible to applications.
111 */
112struct mbuf;
113
114/*
115 * We distinguish between routes to hosts and routes to networks,
116 * preferring the former if available.  For each route we infer
117 * the interface to use from the gateway address supplied when
118 * the route was entered.  Routes that forward packets through
119 * gateways are marked so that the output routines know to address the
120 * gateway rather than the ultimate destination.
121 */
122#ifndef RNF_NORMAL
123#include <net/radix.h>
124#ifdef RADIX_MPATH
125#include <net/radix_mpath.h>
126#endif
127#endif
128struct rtentry {
129	struct	radix_node rt_nodes[2];	/* tree glue, and other values */
130	/*
131	 * XXX struct rtentry must begin with a struct radix_node (or two!)
132	 * because the code does some casts of a 'struct radix_node *'
133	 * to a 'struct rtentry *'
134	 */
135#define	rt_key(r)	(*((struct sockaddr **)(&(r)->rt_nodes->rn_key)))
136#define	rt_mask(r)	(*((struct sockaddr **)(&(r)->rt_nodes->rn_mask)))
137	struct	sockaddr *rt_gateway;	/* value */
138	int	rt_flags;		/* up/down?, host/net */
139	int	rt_refcnt;		/* # held references */
140	struct	ifnet *rt_ifp;		/* the answer: interface to use */
141	struct	ifaddr *rt_ifa;		/* the answer: interface address to use */
142	struct	rt_metrics_lite rt_rmx;	/* metrics used by rx'ing protocols */
143	u_int	rt_fibnum;		/* which FIB */
144#ifdef _KERNEL
145	/* XXX ugly, user apps use this definition but don't have a mtx def */
146	struct	mtx rt_mtx;		/* mutex for routing entry */
147#endif
148};
149
150/*
151 * Following structure necessary for 4.3 compatibility;
152 * We should eventually move it to a compat file.
153 */
154struct ortentry {
155	u_long	rt_hash;		/* to speed lookups */
156	struct	sockaddr rt_dst;	/* key */
157	struct	sockaddr rt_gateway;	/* value */
158	short	rt_flags;		/* up/down?, host/net */
159	short	rt_refcnt;		/* # held references */
160	u_long	rt_use;			/* raw # packets forwarded */
161	struct	ifnet *rt_ifp;		/* the answer: interface to use */
162};
163
164#define rt_use rt_rmx.rmx_pksent
165
166#define	RTF_UP		0x1		/* route usable */
167#define	RTF_GATEWAY	0x2		/* destination is a gateway */
168#define	RTF_HOST	0x4		/* host entry (net otherwise) */
169#define	RTF_REJECT	0x8		/* host or net unreachable */
170#define	RTF_DYNAMIC	0x10		/* created dynamically (by redirect) */
171#define	RTF_MODIFIED	0x20		/* modified dynamically (by redirect) */
172#define RTF_DONE	0x40		/* message confirmed */
173/*			0x80		   unused, was RTF_DELCLONE */
174/*			0x100		   unused, was RTF_CLONING */
175#define RTF_XRESOLVE	0x200		/* external daemon resolves name */
176#define RTF_LLINFO	0x400		/* DEPRECATED - exists ONLY for backward
177					   compatibility */
178#define RTF_LLDATA	0x400		/* used by apps to add/del L2 entries */
179#define RTF_STATIC	0x800		/* manually added */
180#define RTF_BLACKHOLE	0x1000		/* just discard pkts (during updates) */
181#define RTF_PROTO2	0x4000		/* protocol specific routing flag */
182#define RTF_PROTO1	0x8000		/* protocol specific routing flag */
183
184/* XXX: temporary to stay API/ABI compatible with userland */
185#ifndef _KERNEL
186#define RTF_PRCLONING	0x10000		/* unused, for compatibility */
187#endif
188
189/*			0x20000		   unused, was RTF_WASCLONED */
190#define RTF_PROTO3	0x40000		/* protocol specific routing flag */
191/*			0x80000		   unused */
192#define RTF_PINNED	0x100000	/* future use */
193#define	RTF_LOCAL	0x200000 	/* route represents a local address */
194#define	RTF_BROADCAST	0x400000	/* route represents a bcast address */
195#define	RTF_MULTICAST	0x800000	/* route represents a mcast address */
196					/* 0x1000000 and up unassigned */
197#define	RTF_RNH_LOCKED	 0x40000000	/* radix node head locked by caller */
198
199/* Mask of RTF flags that are allowed to be modified by RTM_CHANGE. */
200#define RTF_FMASK	\
201	(RTF_PROTO1 | RTF_PROTO2 | RTF_PROTO3 | RTF_BLACKHOLE | \
202	 RTF_REJECT | RTF_STATIC)
203
204/*
205 * Routing statistics.
206 */
207struct	rtstat {
208	short	rts_badredirect;	/* bogus redirect calls */
209	short	rts_dynamic;		/* routes created by redirects */
210	short	rts_newgateway;		/* routes modified by redirects */
211	short	rts_unreach;		/* lookups which failed */
212	short	rts_wildcard;		/* lookups satisfied by a wildcard */
213};
214/*
215 * Structures for routing messages.
216 */
217struct rt_msghdr {
218	u_short	rtm_msglen;	/* to skip over non-understood messages */
219	u_char	rtm_version;	/* future binary compatibility */
220	u_char	rtm_type;	/* message type */
221	u_short	rtm_index;	/* index for associated ifp */
222	int	rtm_flags;	/* flags, incl. kern & message, e.g. DONE */
223	int	rtm_addrs;	/* bitmask identifying sockaddrs in msg */
224	pid_t	rtm_pid;	/* identify sender */
225	int	rtm_seq;	/* for sender to identify action */
226	int	rtm_errno;	/* why failed */
227	int	rtm_fmask;	/* bitmask used in RTM_CHANGE message */
228#define	rtm_use	rtm_fmask	/* deprecated, use rtm_rmx->rmx_pksent */
229	u_long	rtm_inits;	/* which metrics we are initializing */
230	struct	rt_metrics rtm_rmx; /* metrics themselves */
231};
232
233#define RTM_VERSION	5	/* Up the ante and ignore older versions */
234
235/*
236 * Message types.
237 */
238#define RTM_ADD		0x1	/* Add Route */
239#define RTM_DELETE	0x2	/* Delete Route */
240#define RTM_CHANGE	0x3	/* Change Metrics or flags */
241#define RTM_GET		0x4	/* Report Metrics */
242#define RTM_LOSING	0x5	/* Kernel Suspects Partitioning */
243#define RTM_REDIRECT	0x6	/* Told to use different route */
244#define RTM_MISS	0x7	/* Lookup failed on this address */
245#define RTM_LOCK	0x8	/* fix specified metrics */
246#define RTM_OLDADD	0x9	/* caused by SIOCADDRT */
247#define RTM_OLDDEL	0xa	/* caused by SIOCDELRT */
248#define RTM_RESOLVE	0xb	/* req to resolve dst to LL addr */
249#define RTM_NEWADDR	0xc	/* address being added to iface */
250#define RTM_DELADDR	0xd	/* address being removed from iface */
251#define RTM_IFINFO	0xe	/* iface going up/down etc. */
252#define	RTM_NEWMADDR	0xf	/* mcast group membership being added to if */
253#define	RTM_DELMADDR	0x10	/* mcast group membership being deleted */
254#define	RTM_IFANNOUNCE	0x11	/* iface arrival/departure */
255#define	RTM_IEEE80211	0x12	/* IEEE80211 wireless event */
256
257/*
258 * Bitmask values for rtm_inits and rmx_locks.
259 */
260#define RTV_MTU		0x1	/* init or lock _mtu */
261#define RTV_HOPCOUNT	0x2	/* init or lock _hopcount */
262#define RTV_EXPIRE	0x4	/* init or lock _expire */
263#define RTV_RPIPE	0x8	/* init or lock _recvpipe */
264#define RTV_SPIPE	0x10	/* init or lock _sendpipe */
265#define RTV_SSTHRESH	0x20	/* init or lock _ssthresh */
266#define RTV_RTT		0x40	/* init or lock _rtt */
267#define RTV_RTTVAR	0x80	/* init or lock _rttvar */
268
269/*
270 * Bitmask values for rtm_addrs.
271 */
272#define RTA_DST		0x1	/* destination sockaddr present */
273#define RTA_GATEWAY	0x2	/* gateway sockaddr present */
274#define RTA_NETMASK	0x4	/* netmask sockaddr present */
275#define RTA_GENMASK	0x8	/* cloning mask sockaddr present */
276#define RTA_IFP		0x10	/* interface name sockaddr present */
277#define RTA_IFA		0x20	/* interface addr sockaddr present */
278#define RTA_AUTHOR	0x40	/* sockaddr for author of redirect */
279#define RTA_BRD		0x80	/* for NEWADDR, broadcast or p-p dest addr */
280
281/*
282 * Index offsets for sockaddr array for alternate internal encoding.
283 */
284#define RTAX_DST	0	/* destination sockaddr present */
285#define RTAX_GATEWAY	1	/* gateway sockaddr present */
286#define RTAX_NETMASK	2	/* netmask sockaddr present */
287#define RTAX_GENMASK	3	/* cloning mask sockaddr present */
288#define RTAX_IFP	4	/* interface name sockaddr present */
289#define RTAX_IFA	5	/* interface addr sockaddr present */
290#define RTAX_AUTHOR	6	/* sockaddr for author of redirect */
291#define RTAX_BRD	7	/* for NEWADDR, broadcast or p-p dest addr */
292#define RTAX_MAX	8	/* size of array to allocate */
293
294struct rt_addrinfo {
295	int	rti_addrs;
296	struct	sockaddr *rti_info[RTAX_MAX];
297	int	rti_flags;
298	struct	ifaddr *rti_ifa;
299	struct	ifnet *rti_ifp;
300};
301
302/*
303 * This macro returns the size of a struct sockaddr when passed
304 * through a routing socket. Basically we round up sa_len to
305 * a multiple of sizeof(long), with a minimum of sizeof(long).
306 * The check for a NULL pointer is just a convenience, probably never used.
307 * The case sa_len == 0 should only apply to empty structures.
308 */
309#define SA_SIZE(sa)						\
310    (  (!(sa) || ((struct sockaddr *)(sa))->sa_len == 0) ?	\
311	sizeof(long)		:				\
312	1 + ( (((struct sockaddr *)(sa))->sa_len - 1) | (sizeof(long) - 1) ) )
313
314#ifdef _KERNEL
315
316#define	RT_LOCK_INIT(_rt) \
317	mtx_init(&(_rt)->rt_mtx, "rtentry", NULL, MTX_DEF | MTX_DUPOK)
318#define	RT_LOCK(_rt)		mtx_lock(&(_rt)->rt_mtx)
319#define	RT_TRYLOCK(_rt)		mtx_trylock(&(_rt)->rt_mtx)
320#define	RT_UNLOCK(_rt)		mtx_unlock(&(_rt)->rt_mtx)
321#define	RT_LOCK_DESTROY(_rt)	mtx_destroy(&(_rt)->rt_mtx)
322#define	RT_LOCK_ASSERT(_rt)	mtx_assert(&(_rt)->rt_mtx, MA_OWNED)
323
324#define	RT_ADDREF(_rt)	do {					\
325	RT_LOCK_ASSERT(_rt);					\
326	KASSERT((_rt)->rt_refcnt >= 0,				\
327		("negative refcnt %d", (_rt)->rt_refcnt));	\
328	(_rt)->rt_refcnt++;					\
329} while (0)
330
331#define	RT_REMREF(_rt)	do {					\
332	RT_LOCK_ASSERT(_rt);					\
333	KASSERT((_rt)->rt_refcnt > 0,				\
334		("bogus refcnt %d", (_rt)->rt_refcnt));	\
335	(_rt)->rt_refcnt--;					\
336} while (0)
337
338#define	RTFREE_LOCKED(_rt) do {					\
339	if ((_rt)->rt_refcnt <= 1)				\
340		rtfree(_rt);					\
341	else {							\
342		RT_REMREF(_rt);					\
343		RT_UNLOCK(_rt);					\
344	}							\
345	/* guard against invalid refs */			\
346	_rt = 0;						\
347} while (0)
348
349#define	RTFREE(_rt) do {					\
350	RT_LOCK(_rt);						\
351	RTFREE_LOCKED(_rt);					\
352} while (0)
353
354#define RT_TEMP_UNLOCK(_rt) do {				\
355	RT_ADDREF(_rt);						\
356	RT_UNLOCK(_rt);						\
357} while (0)
358
359#define RT_RELOCK(_rt) do {					\
360	RT_LOCK(_rt);						\
361	if ((_rt)->rt_refcnt <= 1) {				\
362		rtfree(_rt);					\
363		_rt = 0; /*  signal that it went away */	\
364	} else {						\
365		RT_REMREF(_rt);					\
366		/* note that _rt is still valid */		\
367	}							\
368} while (0)
369
370extern struct radix_node_head *rt_tables[][AF_MAX+1];
371
372struct ifmultiaddr;
373
374void	 rt_ieee80211msg(struct ifnet *, int, void *, size_t);
375void	 rt_ifannouncemsg(struct ifnet *, int);
376void	 rt_ifmsg(struct ifnet *);
377void	 rt_missmsg(int, struct rt_addrinfo *, int, int);
378void	 rt_newaddrmsg(int, struct ifaddr *, int, struct rtentry *);
379void	 rt_newmaddrmsg(int, struct ifmultiaddr *);
380int	 rt_setgate(struct rtentry *, struct sockaddr *, struct sockaddr *);
381
382/*
383 * Note the following locking behavior:
384 *
385 *    rtalloc_ign() and rtalloc() return ro->ro_rt unlocked
386 *
387 *    rtalloc1() returns a locked rtentry
388 *
389 *    rtfree() and RTFREE_LOCKED() require a locked rtentry
390 *
391 *    RTFREE() uses an unlocked entry.
392 */
393
394int	 rtexpunge(struct rtentry *);
395void	 rtfree(struct rtentry *);
396int	 rt_check(struct rtentry **, struct rtentry **, struct sockaddr *);
397
398/* XXX MRT COMPAT VERSIONS THAT SET UNIVERSE to 0 */
399/* Thes are used by old code not yet converted to use multiple FIBS */
400int	 rt_getifa(struct rt_addrinfo *);
401void	 rtalloc_ign(struct route *ro, u_long ignflags);
402void	 rtalloc(struct route *ro); /* XXX deprecated, use rtalloc_ign(ro, 0) */
403struct rtentry *rtalloc1(struct sockaddr *, int, u_long);
404int	 rtinit(struct ifaddr *, int, int);
405int	 rtioctl(u_long, caddr_t);
406void	 rtredirect(struct sockaddr *, struct sockaddr *,
407	    struct sockaddr *, int, struct sockaddr *);
408int	 rtrequest(int, struct sockaddr *,
409	    struct sockaddr *, struct sockaddr *, int, struct rtentry **);
410
411/* defaults to "all" FIBs */
412int	 rtinit_fib(struct ifaddr *, int, int);
413
414/* XXX MRT NEW VERSIONS THAT USE FIBs
415 * For now the protocol indepedent versions are the same as the AF_INET ones
416 * but this will change..
417 */
418int	 rt_getifa_fib(struct rt_addrinfo *, u_int fibnum);
419void	 rtalloc_ign_fib(struct route *ro, u_long ignflags, u_int fibnum);
420void	 rtalloc_fib(struct route *ro, u_int fibnum);
421struct rtentry *rtalloc1_fib(struct sockaddr *, int, u_long, u_int);
422int	 rtioctl_fib(u_long, caddr_t, u_int);
423void	 rtredirect_fib(struct sockaddr *, struct sockaddr *,
424	    struct sockaddr *, int, struct sockaddr *, u_int);
425int	 rtrequest_fib(int, struct sockaddr *,
426	    struct sockaddr *, struct sockaddr *, int, struct rtentry **, u_int);
427int	 rtrequest1_fib(int, struct rt_addrinfo *, struct rtentry **, u_int);
428
429#include <sys/eventhandler.h>
430typedef void (*rtevent_arp_update_fn)(void *, struct rtentry *, uint8_t *, struct sockaddr *);
431typedef void (*rtevent_redirect_fn)(void *, struct rtentry *, struct rtentry *, struct sockaddr *);
432EVENTHANDLER_DECLARE(route_arp_update_event, rtevent_arp_update_fn);
433EVENTHANDLER_DECLARE(route_redirect_event, rtevent_redirect_fn);
434#endif
435
436#endif
437