xref: /dragonfly/sys/net/route.h (revision ffe53622)
1 /*
2  * Copyright (c) 2004, 2005 The DragonFly Project.  All rights reserved.
3  *
4  * This code is derived from software contributed to The DragonFly Project
5  * by Jeffrey M. Hsu.
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 DragonFly Project nor the names of its
16  *    contributors may be used to endorse or promote products derived
17  *    from this software without specific, prior written permission.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
22  * FOR A PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE
23  * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
24  * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
25  * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
26  * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
27  * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
28  * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
29  * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30  * SUCH DAMAGE.
31  */
32 
33 /*
34  * Copyright (c) 1980, 1986, 1993
35  *	The Regents of the University of California.  All rights reserved.
36  *
37  * Redistribution and use in source and binary forms, with or without
38  * modification, are permitted provided that the following conditions
39  * are met:
40  * 1. Redistributions of source code must retain the above copyright
41  *    notice, this list of conditions and the following disclaimer.
42  * 2. Redistributions in binary form must reproduce the above copyright
43  *    notice, this list of conditions and the following disclaimer in the
44  *    documentation and/or other materials provided with the distribution.
45  * 3. Neither the name of the University nor the names of its contributors
46  *    may be used to endorse or promote products derived from this software
47  *    without specific prior written permission.
48  *
49  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
50  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
51  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
52  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
53  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
54  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
55  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
56  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
57  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
58  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
59  * SUCH DAMAGE.
60  *
61  *	@(#)route.h	8.4 (Berkeley) 1/9/95
62  * $FreeBSD: src/sys/net/route.h,v 1.36.2.5 2002/02/01 11:48:01 ru Exp $
63  * $DragonFly: src/sys/net/route.h,v 1.24 2008/09/11 11:23:29 sephe Exp $
64  */
65 
66 #ifndef _NET_ROUTE_H_
67 #define _NET_ROUTE_H_
68 
69 #ifndef _SYS_TYPES_H_
70 #include <sys/types.h>
71 #endif
72 #ifndef _SYS_PARAM_H_
73 #include <sys/param.h>
74 #endif
75 #ifndef _SYS_SOCKET_H_
76 #include <sys/socket.h>
77 #endif
78 
79 /*
80  * Kernel resident routing tables.
81  *
82  * The routing tables are initialized when interface addresses
83  * are set by making entries for all directly connected interfaces.
84  */
85 
86 /*
87  * A route consists of a destination address and a reference
88  * to a routing entry.  These are often held by protocols
89  * in their control blocks, e.g. inpcb.
90  */
91 struct route {
92 	struct	rtentry *ro_rt;
93 	struct	sockaddr ro_dst;
94 };
95 
96 /*
97  * These numbers are used by reliable protocols for determining
98  * retransmission behavior and are included in the routing structure.
99  */
100 struct rt_metrics {	/* grouped for locality of reference */
101 	u_long	rmx_locks;	/* Kernel must leave these values alone */
102 	u_long	rmx_mtu;	/* MTU for this path */
103 	u_long	rmx_pksent;	/* packets sent using this route */
104 	u_long	rmx_expire;	/* lifetime for route */
105 
106 	u_long	rmx_sendpipe;	/* outbound delay-bandwidth product */
107 	u_long	rmx_ssthresh;	/* outbound gateway buffer limit */
108 	u_long	rmx_rtt;	/* estimated round trip time */
109 	u_long	rmx_rttvar;	/* estimated rtt variance */
110 
111 	u_long	rmx_recvpipe;	/* inbound delay-bandwidth product */
112 	u_long	rmx_hopcount;	/* max hops expected */
113 	u_short rmx_mssopt;	/* peer's cached MSS */
114 	u_short	rmx_pad;	/* explicit pad */
115 	u_long	rmx_msl;	/* maximum segment lifetime, unit: ms */
116 	u_long	rmx_iwmaxsegs;	/* IW segments max */
117 	u_long	rmx_iwcapsegs;	/* IW segments */
118 };
119 
120 /*
121  * rmx_rtt and rmx_rttvar are stored as microseconds;
122  * RTTTOPRHZ(rtt) converts to a value suitable for use
123  * by a protocol slowtimo counter.
124  */
125 #define	RTM_RTTUNIT	1000000	/* units for rtt, rttvar, as units per sec */
126 #define	RTTTOPRHZ(r)	((r) / (RTM_RTTUNIT / PR_SLOWHZ))
127 
128 /*
129  * XXX kernel function pointer `rt_output' is visible to applications.
130  */
131 struct mbuf;
132 
133 /*
134  * We distinguish between routes to hosts and routes to networks,
135  * preferring the former if available.  For each route we infer
136  * the interface to use from the gateway address supplied when
137  * the route was entered.  Routes that forward packets through
138  * gateways are marked so that the output routines know to address the
139  * gateway rather than the ultimate destination.
140  */
141 #ifndef RNF_NORMAL
142 #include <net/radix.h>
143 #endif
144 
145 struct rtentry {
146 	struct	radix_node rt_nodes[2];	/* tree glue, and other values */
147 #define	rt_key(r)	((struct sockaddr *)((r)->rt_nodes->rn_key))
148 #define	rt_mask(r)	((struct sockaddr *)((r)->rt_nodes->rn_mask))
149 	struct	sockaddr *rt_gateway;	/* value */
150 	long	rt_refcnt;		/* # held references */
151 	u_long	rt_flags;		/* up/down?, host/net */
152 	struct	ifnet *rt_ifp;		/* the answer: interface to use */
153 	struct	ifaddr *rt_ifa;		/* the answer: interface to use */
154 	struct	sockaddr *rt_genmask;	/* for generation of cloned routes */
155 	void	*rt_llinfo;		/* pointer to link level info cache */
156 	struct	rt_metrics rt_rmx;	/* metrics used by rx'ing protocols */
157 	struct	rtentry *rt_gwroute;	/* implied entry for gatewayed routes */
158 	int	(*rt_output) (struct ifnet *, struct mbuf *, struct sockaddr *,
159 			      struct rtentry *);
160 					/* output routine for this (rt,if) */
161 	struct	rtentry *rt_parent;	/* cloning parent of this route */
162 	int	rt_cpuid;		/* owner cpu */
163 	struct	sockaddr *rt_shim[3];	/* mpls label / operation array */
164 };
165 
166 #define rt_use rt_rmx.rmx_pksent
167 
168 #define	RTF_UP		0x1		/* route usable */
169 #define	RTF_GATEWAY	0x2		/* destination is a gateway */
170 #define	RTF_HOST	0x4		/* host entry (net otherwise) */
171 #define	RTF_REJECT	0x8		/* host or net unreachable */
172 #define	RTF_DYNAMIC	0x10		/* created dynamically (by redirect) */
173 #define	RTF_MODIFIED	0x20		/* modified dynamically (by redirect) */
174 #define RTF_DONE	0x40		/* message confirmed */
175 /*			0x80		   unused, was RTF_DELCLONE */
176 #define RTF_CLONING	0x100		/* generate new routes on use */
177 #define RTF_XRESOLVE	0x200		/* external daemon resolves name */
178 #define RTF_LLINFO	0x400		/* generated by link layer (e.g. ARP) */
179 #define RTF_STATIC	0x800		/* manually added */
180 #define RTF_BLACKHOLE	0x1000		/* just discard pkts (during updates) */
181 /*			0x2000		   unused */
182 #define RTF_PROTO2	0x4000		/* protocol specific routing flag */
183 #define RTF_PROTO1	0x8000		/* protocol specific routing flag */
184 
185 #define RTF_PRCLONING	0x10000		/* protocol requires cloning */
186 #define RTF_WASCLONED	0x20000		/* route generated through cloning */
187 #define RTF_PROTO3	0x40000		/* protocol specific routing flag */
188 /*			0x80000		   unused */
189 #define RTF_PINNED	0x100000	/* future use */
190 #define	RTF_LOCAL	0x200000	/* route represents a local address */
191 #define	RTF_BROADCAST	0x400000	/* route represents a bcast address */
192 #define	RTF_MULTICAST	0x800000	/* route represents a mcast address */
193 #define	RTF_MPLSOPS	0x1000000	/* route uses mpls label operations */
194 					/* 0x2000000 and up unassigned */
195 
196 /*
197  * Routing statistics.
198  */
199 struct rtstatistics {
200 	u_long	rts_badredirect;	/* bogus redirect calls */
201 	u_long	rts_dynamic;		/* routes created by redirects */
202 	u_long	rts_newgateway;		/* routes modified by redirects */
203 	u_long	rts_unreach;		/* lookups which failed */
204 	u_long	rts_wildcard;		/* lookups satisfied by a wildcard */
205 	u_long	rts_pad[3];		/* pad to cache line size (64B) */
206 };
207 #ifdef _KERNEL
208 CTASSERT((sizeof(struct rtstatistics) & __VM_CACHELINE_MASK) == 0);
209 #endif
210 
211 /*
212  * Structures for routing messages.
213  */
214 struct rt_msghdr {
215 	u_short	rtm_msglen;	/* to skip over non-understood messages */
216 	u_char	rtm_version;	/* future binary compatibility */
217 	u_char	rtm_type;	/* message type */
218 	u_short	rtm_index;	/* index for associated ifp */
219 	int	rtm_flags;	/* flags, incl. kern & message, e.g. DONE */
220 	int	rtm_addrs;	/* bitmask identifying sockaddrs in msg */
221 	pid_t	rtm_pid;	/* identify sender */
222 	int	rtm_seq;	/* for sender to identify action */
223 	int	rtm_errno;	/* why failed */
224 	int	rtm_use;	/* from rtentry */
225 	u_long	rtm_inits;	/* which metrics we are initializing */
226 	struct	rt_metrics rtm_rmx; /* metrics themselves */
227 };
228 
229 #define RTM_VERSION	6	/* Up the ante and ignore older versions */
230 
231 /*
232  * Message types.
233  */
234 #define RTM_ADD		0x1	/* Add Route */
235 #define RTM_DELETE	0x2	/* Delete Route */
236 #define RTM_CHANGE	0x3	/* Change Metrics or flags */
237 #define RTM_GET		0x4	/* Report Metrics */
238 #define RTM_LOSING	0x5	/* Kernel Suspects Partitioning */
239 #define RTM_REDIRECT	0x6	/* Told to use different route */
240 #define RTM_MISS	0x7	/* Lookup failed on this address */
241 #define RTM_LOCK	0x8	/* fix specified metrics */
242 /* 			0x9	   unused, was RTM_OLDADD */
243 /* 			0xa	   unused, was RTM_OLDDEL */
244 #define RTM_RESOLVE	0xb	/* req to resolve dst to LL addr */
245 #define RTM_NEWADDR	0xc	/* address being added to iface */
246 #define RTM_DELADDR	0xd	/* address being removed from iface */
247 #define RTM_IFINFO	0xe	/* iface going up/down etc. */
248 #define	RTM_NEWMADDR	0xf	/* mcast group membership being added to if */
249 #define	RTM_DELMADDR	0x10	/* mcast group membership being deleted */
250 #define	RTM_IFANNOUNCE	0x11	/* iface arrival/departure */
251 #define	RTM_IEEE80211	0x12	/* IEEE80211 wireless event */
252 
253 /*
254  * Bitmask values for rtm_inits and rmx_locks.
255  */
256 #define RTV_MTU		0x1	/* init or lock _mtu */
257 #define RTV_HOPCOUNT	0x2	/* init or lock _hopcount */
258 #define RTV_EXPIRE	0x4	/* init or lock _expire */
259 #define RTV_RPIPE	0x8	/* init or lock _recvpipe */
260 #define RTV_SPIPE	0x10	/* init or lock _sendpipe */
261 #define RTV_SSTHRESH	0x20	/* init or lock _ssthresh */
262 #define RTV_RTT		0x40	/* init or lock _rtt */
263 #define RTV_RTTVAR	0x80	/* init or lock _rttvar */
264 #define RTV_MSL		0x100	/* init or lock _msl */
265 #define RTV_IWMAXSEGS	0x200	/* init or lock _iwmaxsegs */
266 #define RTV_IWCAPSEGS	0x400	/* init or lock _iwcapsegs */
267 
268 /*
269  * Bitmask values for rtm_addrs.
270  */
271 #define RTA_DST		0x1	/* destination sockaddr present */
272 #define RTA_GATEWAY	0x2	/* gateway sockaddr present */
273 #define RTA_NETMASK	0x4	/* netmask sockaddr present */
274 #define RTA_GENMASK	0x8	/* cloning mask sockaddr present */
275 #define RTA_IFP		0x10	/* interface name sockaddr present */
276 #define RTA_IFA		0x20	/* interface addr sockaddr present */
277 #define RTA_AUTHOR	0x40	/* sockaddr for author of redirect */
278 #define RTA_BRD		0x80	/* for NEWADDR, broadcast or p-p dest addr */
279 #define RTA_MPLS1	0x100	/* mpls label and/or operation present */
280 #define RTA_MPLS2	0x200	/* mpls label and/or operation present */
281 #define RTA_MPLS3	0x400	/* mpls label and/or operation present */
282 
283 /*
284  * Index offsets for sockaddr array for alternate internal encoding.
285  */
286 #define RTAX_DST	0	/* destination sockaddr present */
287 #define RTAX_GATEWAY	1	/* gateway sockaddr present */
288 #define RTAX_NETMASK	2	/* netmask sockaddr present */
289 #define RTAX_GENMASK	3	/* cloning mask sockaddr present */
290 #define RTAX_IFP	4	/* interface name sockaddr present */
291 #define RTAX_IFA	5	/* interface addr sockaddr present */
292 #define RTAX_AUTHOR	6	/* sockaddr for author of redirect */
293 #define RTAX_BRD	7	/* for NEWADDR, broadcast or p-p dest addr */
294 #define RTAX_MPLS1	8	/* mpls label and/or operation present */
295 #define RTAX_MPLS2	9	/* mpls label and/or operation present */
296 #define RTAX_MPLS3	10	/* mpls label and/or operation present */
297 #define RTAX_MAX	11	/* size of array to allocate */
298 
299 /*
300  * Sockaddrs passed through the routing socket are padded up to a certain
301  * size, currently multiples of sizeof(long).  These macros used to be
302  * replicated in various user and kernel files, but are now made available
303  * here for convenience and consistency.
304  *
305  * FreeBSD uses a similar macro SA_SIZE which could be defined as
306  * #define SA_SIZE(sa)	RT_ROUNDUP((sa)->sa_len)
307  */
308 #define RT_ROUNDUP2(a, n)	((a) > 0 ? (1 + (((a) - 1) | ((n) - 1))) : (n))
309 #define RT_ROUNDUP(a)		RT_ROUNDUP2((a), sizeof(long))
310 #define RT_ADVANCE(x, n)	(x += RT_ROUNDUP((n)->sa_len))
311 
312 struct rt_addrinfo {
313 	int		 rti_addrs;
314 	struct sockaddr	*rti_info[RTAX_MAX];
315 	int		 rti_flags;
316 	struct ifaddr	*rti_ifa;
317 	struct ifnet	*rti_ifp;
318 };
319 
320 #ifdef _KERNEL
321 
322 #define	rti_dst		rti_info[RTAX_DST]
323 #define	rti_gateway	rti_info[RTAX_GATEWAY]
324 #define	rti_netmask	rti_info[RTAX_NETMASK]
325 #define	rti_genmask	rti_info[RTAX_GENMASK]
326 #define	rti_ifpaddr	rti_info[RTAX_IFP]
327 #define	rti_ifaaddr	rti_info[RTAX_IFA]
328 #define	rti_author	rti_info[RTAX_AUTHOR]
329 #define	rti_bcastaddr	rti_info[RTAX_BRD]
330 #define	rti_mpls1	rti_info[RTAX_MPLS1]
331 #define	rti_mpls2	rti_info[RTAX_MPLS2]
332 #define	rti_mpls3	rti_info[RTAX_MPLS3]
333 
334 extern struct radix_node_head *rt_tables[MAXCPU][AF_MAX+1];
335 
336 struct ifmultiaddr;
337 struct proc;
338 struct ucred;
339 
340 void	 route_init (void);
341 void	 rt_dstmsg(int type, struct sockaddr *dst, int error);
342 int	 rt_getifa (struct rt_addrinfo *);
343 void	 rt_ieee80211msg(struct ifnet *, int, void *, size_t);
344 void	 rt_ifannouncemsg (struct ifnet *, int);
345 void	 rt_ifmsg (struct ifnet *);
346 int	 rt_llroute (struct sockaddr *dst, struct rtentry *rt0,
347 	    struct rtentry **drt);
348 void	 rt_missmsg (int, struct rt_addrinfo *, int, int);
349 void	 rt_newaddrmsg (int, struct ifaddr *, int, struct rtentry *);
350 void	 rt_newmaddrmsg (int, struct ifmultiaddr *);
351 void	 rt_rtmsg(int cmd, struct rtentry *rt, struct ifnet *ifp, int error);
352 int	 rt_setgate (struct rtentry *,
353 	    struct sockaddr *, struct sockaddr *, boolean_t);
354 void	 rtalloc (struct route *);
355 void	 rtalloc_ign (struct route *, u_long);
356 
357 struct rtentry *
358 	 _rtlookup (struct sockaddr *, __boolean_t, u_long);
359 #define		RTL_REPORTMSG	TRUE
360 #define		RTL_DONTREPORT	FALSE
361 
362 /* flags to ignore */
363 #define		RTL_DOCLONE	0UL
364 #define		RTL_DONTCLONE	(RTF_CLONING | RTF_PRCLONING)
365 
366 /*
367  * Look up a route with no cloning side-effects or miss reports generated.
368  */
369 static __inline struct rtentry *
370 rtpurelookup(struct sockaddr *dst)
371 {
372 	return _rtlookup(dst, RTL_DONTREPORT, RTL_DONTCLONE);
373 }
374 
375 /*
376  * Do full route lookup with cloning and reporting on misses.
377  */
378 static __inline struct rtentry *
379 rtlookup(struct sockaddr *dst)
380 {
381 	return _rtlookup(dst, RTL_REPORTMSG, RTL_DOCLONE);
382 }
383 
384 typedef void (*rtrequest1_callback_func_t)(int, int, struct rt_addrinfo *,
385 				      struct rtentry *, void *);
386 typedef int (*rtsearch_callback_func_t)(int, struct rt_addrinfo *,
387 					struct rtentry *, void *, int);
388 
389 void	 rtfree (struct rtentry *);
390 void	 rtfree_async (struct rtentry *);
391 int	 rtinit (struct ifaddr *, int, int);
392 int	 rtchange (struct ifaddr *, struct ifaddr *);
393 int	 rtioctl (u_long, caddr_t, struct ucred *);
394 void	 rtredirect (struct sockaddr *, struct sockaddr *,
395 	    struct sockaddr *, int, struct sockaddr *);
396 int	 rtredirect_oncpu (struct sockaddr *, struct sockaddr *,
397 	    struct sockaddr *, int, struct sockaddr *);
398 int	 rtrequest (int, struct sockaddr *,
399 	    struct sockaddr *, struct sockaddr *, int, struct rtentry **);
400 int	 rtrequest_global (int, struct sockaddr *,
401 	    struct sockaddr *, struct sockaddr *, int);
402 int	 rtrequest1 (int, struct rt_addrinfo *, struct rtentry **);
403 int	 rtrequest1_global (int, struct rt_addrinfo *,
404 	    rtrequest1_callback_func_t, void *, boolean_t);
405 
406 #define RTS_EXACTMATCH		TRUE
407 #define RTS_NOEXACTMATCH	FALSE
408 
409 #define RTREQ_PRIO_HIGH		TRUE
410 #define RTREQ_PRIO_NORM		FALSE
411 
412 int	 rtsearch_global(int, struct rt_addrinfo *,
413 	    rtsearch_callback_func_t, void *, boolean_t, boolean_t);
414 
415 int	 rtmask_add_global(struct sockaddr *, boolean_t);
416 
417 struct sockaddr *_rtmask_lookup(struct sockaddr *, boolean_t);
418 
419 static __inline struct sockaddr *
420 rtmask_lookup(struct sockaddr *_mask)
421 {
422 	return _rtmask_lookup(_mask, FALSE);
423 }
424 
425 static __inline struct sockaddr *
426 rtmask_purelookup(struct sockaddr *_mask)
427 {
428 	return _rtmask_lookup(_mask, TRUE);
429 }
430 
431 void	rt_print(struct rt_addrinfo *, struct rtentry *);
432 void	rt_addrinfo_print(int cmd, struct rt_addrinfo *);
433 void	sockaddr_print(struct sockaddr *);
434 
435 #ifndef _SYS_GLOBALDATA_H_
436 #include <sys/globaldata.h>
437 #endif
438 #ifndef _NET_NETISR2_H_
439 #include <net/netisr2.h>
440 #endif
441 
442 static __inline void
443 RTFREE(struct rtentry *rt)
444 {
445 
446 	ASSERT_NETISR_NCPUS(rt->rt_cpuid);
447 	if (rt->rt_refcnt <= 1)
448 		rtfree(rt);
449 	else
450 		--rt->rt_refcnt;
451 }
452 
453 int	in_inithead(void **, int);
454 
455 #endif	/* _KERNEL */
456 
457 #endif	/* !_NET_ROUTE_H_ */
458