route.h revision 1.122
1/*	$NetBSD: route.h,v 1.122 2019/04/29 11:57:22 roy Exp $	*/
2
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	8.5 (Berkeley) 2/8/95
32 */
33
34#ifndef _NET_ROUTE_H_
35#define _NET_ROUTE_H_
36
37#include <sys/queue.h>
38#include <sys/socket.h>
39#include <sys/types.h>
40#include <net/if.h>
41#ifdef _KERNEL
42#include <sys/rwlock.h>
43#include <sys/condvar.h>
44#include <sys/pserialize.h>
45#endif
46#include <sys/psref.h>
47
48#if !(defined(_KERNEL) || defined(_STANDALONE))
49#include <stdbool.h>
50#endif
51
52/*
53 * Kernel resident routing tables.
54 *
55 * The routing tables are initialized when interface addresses
56 * are set by making entries for all directly connected interfaces.
57 */
58
59/*
60 * A route consists of a destination address and a reference
61 * to a routing entry.  These are often held by protocols
62 * in their control blocks, e.g. inpcb.
63 */
64struct route {
65	struct	rtentry		*_ro_rt;
66	struct	sockaddr	*ro_sa;
67	uint64_t		ro_rtcache_generation;
68	struct	psref		ro_psref;
69	int			ro_bound;
70};
71
72/*
73 * These numbers are used by reliable protocols for determining
74 * retransmission behavior and are included in the routing structure.
75 */
76struct rt_metrics {
77	uint64_t rmx_locks;	/* Kernel must leave these values alone */
78	uint64_t rmx_mtu;	/* MTU for this path */
79	uint64_t rmx_hopcount;	/* max hops expected */
80	uint64_t rmx_recvpipe;	/* inbound delay-bandwidth product */
81	uint64_t rmx_sendpipe;	/* outbound delay-bandwidth product */
82	uint64_t rmx_ssthresh;	/* outbound gateway buffer limit */
83	uint64_t rmx_rtt;	/* estimated round trip time */
84	uint64_t rmx_rttvar;	/* estimated rtt variance */
85	time_t	rmx_expire;	/* lifetime for route, e.g. redirect */
86	time_t	rmx_pksent;	/* packets sent using this route */
87};
88
89/*
90 * rmx_rtt and rmx_rttvar are stored as microseconds;
91 * RTTTOPRHZ(rtt) converts to a value suitable for use
92 * by a protocol slowtimo counter.
93 */
94#define	RTM_RTTUNIT	1000000	/* units for rtt, rttvar, as units per sec */
95#define	RTTTOPRHZ(r)	((r) / (RTM_RTTUNIT / PR_SLOWHZ))
96
97/*
98 * We distinguish between routes to hosts and routes to networks,
99 * preferring the former if available.  For each route we infer
100 * the interface to use from the gateway address supplied when
101 * the route was entered.  Routes that forward packets through
102 * gateways are marked so that the output routines know to address the
103 * gateway rather than the ultimate destination.
104 */
105#ifndef RNF_NORMAL
106#include <net/radix.h>
107#endif
108struct rtentry {
109	struct	radix_node rt_nodes[2];	/* tree glue, and other values */
110#define	rt_mask(r)	((const struct sockaddr *)((r)->rt_nodes->rn_mask))
111	struct	sockaddr *rt_gateway;	/* value */
112	int	rt_flags;		/* up/down?, host/net */
113	int	rt_refcnt;		/* # held references */
114	uint64_t rt_use;			/* raw # packets forwarded */
115	struct	ifnet *rt_ifp;		/* the answer: interface to use */
116	struct	ifaddr *rt_ifa;		/* the answer: interface to use */
117	uint32_t rt_ifa_seqno;
118	void *	rt_llinfo;		/* pointer to link level info cache */
119	struct	rt_metrics rt_rmx;	/* metrics used by rx'ing protocols */
120	struct	rtentry *rt_gwroute;	/* implied entry for gatewayed routes */
121	LIST_HEAD(, rttimer) rt_timer;  /* queue of timeouts for misc funcs */
122	struct	rtentry *rt_parent;	/* parent of cloned route */
123	struct	sockaddr *_rt_key;
124	struct	sockaddr *rt_tag;	/* route tagging info */
125#ifdef _KERNEL
126	kcondvar_t rt_cv;
127	struct psref_target rt_psref;
128	SLIST_ENTRY(rtentry) rt_free;	/* queue of deferred frees */
129#endif
130};
131
132static __inline const struct sockaddr *
133rt_getkey(const struct rtentry *rt)
134{
135	return rt->_rt_key;
136}
137
138/*
139 * Following structure necessary for 4.3 compatibility;
140 * We should eventually move it to a compat file.
141 */
142struct ortentry {
143	uint32_t rt_hash;		/* to speed lookups */
144	struct	sockaddr rt_dst;	/* key */
145	struct	sockaddr rt_gateway;	/* value */
146	int16_t	rt_flags;		/* up/down?, host/net */
147	int16_t	rt_refcnt;		/* # held references */
148	uint32_t rt_use;		/* raw # packets forwarded */
149	struct	ifnet *rt_ifp;		/* the answer: interface to use */
150};
151
152#define	RTF_UP		0x1		/* route usable */
153#define	RTF_GATEWAY	0x2		/* destination is a gateway */
154#define	RTF_HOST	0x4		/* host entry (net otherwise) */
155#define	RTF_REJECT	0x8		/* host or net unreachable */
156#define	RTF_DYNAMIC	0x10		/* created dynamically (by redirect) */
157#define	RTF_MODIFIED	0x20		/* modified dynamically (by redirect) */
158#define RTF_DONE	0x40		/* message confirmed */
159#define RTF_MASK	0x80		/* subnet mask present */
160// #define RTF_CLONING	0x100		/* generate new routes on use */
161#define RTF_CONNECTED	0x100		/* hosts on this route are neighbours */
162// #define RTF_XRESOLVE	0x200		/* external daemon resolves name */
163// #define RTF_LLINFO	0x400		/* generated by ARP or NDP */
164#define RTF_LLDATA	0x400		/* used by apps to add/del L2 entries */
165#define RTF_STATIC	0x800		/* manually added */
166#define RTF_BLACKHOLE	0x1000		/* just discard pkts (during updates) */
167// #define RTF_CLONED	0x2000		/* this is a cloned route */
168#define RTF_PROTO2	0x4000		/* protocol specific routing flag */
169#define RTF_PROTO1	0x8000		/* protocol specific routing flag */
170#define RTF_SRC		0x10000		/* route has fixed source address */
171#define RTF_ANNOUNCE	0x20000		/* announce new ARP or NDP entry */
172#define RTF_LOCAL	0x40000		/* route represents a local address */
173#define RTF_BROADCAST	0x80000		/* route represents a bcast address */
174#define RTF_UPDATING	0x100000	/* route is updating */
175/*
176 * The flag is nevert set to rt_flags.  It just tells rtrequest1 to set a passed
177 * ifa to rt_ifa (via rti_ifa) and not replace rt_ifa in ifa_rtrequest.
178 */
179#define RTF_DONTCHANGEIFA	0x200000	/* suppress rt_ifa replacement */
180
181/*
182 * 0x400 is exposed to userland just for backward compatibility. For that
183 * purpose, it should be shown as LLINFO.
184 */
185#define RTFBITS "\020\1UP\2GATEWAY\3HOST\4REJECT\5DYNAMIC\6MODIFIED\7DONE" \
186    "\010MASK_PRESENT\011CONNECTED\012XRESOLVE\013LLINFO\014STATIC" \
187    "\015BLACKHOLE\016CLONED\017PROTO2\020PROTO1\021SRC\022ANNOUNCE" \
188    "\023LOCAL\024BROADCAST\025UPDATING"
189
190
191/*
192 * Routing statistics.
193 */
194struct	rtstat {
195	uint64_t rts_badredirect;	/* bogus redirect calls */
196	uint64_t rts_dynamic;		/* routes created by redirects */
197	uint64_t rts_newgateway;	/* routes modified by redirects */
198	uint64_t rts_unreach;		/* lookups which failed */
199	uint64_t rts_wildcard;		/* lookups satisfied by a wildcard */
200};
201
202/*
203 * Structures for routing messages.  By forcing the first member to be aligned
204 * at a 64-bit boundary, we also force the size to be a multiple of 64-bits.
205 */
206
207#if !defined(_KERNEL) || !defined(COMPAT_RTSOCK)
208/*
209 * If we aren't being compiled for backwards compatiblity, enforce 64-bit
210 * alignment so any routing message is the same regardless if the kernel
211 * is an ILP32 or LP64 kernel.
212 */
213#define	__align64	__aligned(sizeof(uint64_t))
214#else
215#define	__align64
216#endif
217
218struct rt_msghdr {
219	u_short	rtm_msglen __align64;
220				/* to skip over non-understood messages */
221	u_char	rtm_version;	/* future binary compatibility */
222	u_char	rtm_type;	/* message type */
223	u_short	rtm_index;	/* index for associated ifp */
224	int	rtm_flags;	/* flags, incl. kern & message, e.g. DONE */
225	int	rtm_addrs;	/* bitmask identifying sockaddrs in msg */
226	pid_t	rtm_pid;	/* identify sender */
227	int	rtm_seq;	/* for sender to identify action */
228	int	rtm_errno;	/* why failed */
229	int	rtm_use;	/* from rtentry */
230	int	rtm_inits;	/* which metrics we are initializing */
231	struct	rt_metrics rtm_rmx __align64;
232				/* metrics themselves */
233};
234
235#undef __align64
236
237#define RTM_VERSION	4	/* Up the ante and ignore older versions */
238
239#define RTM_ADD		0x1	/* Add Route */
240#define RTM_DELETE	0x2	/* Delete Route */
241#define RTM_CHANGE	0x3	/* Change Metrics or flags */
242#define RTM_GET		0x4	/* Report Metrics */
243#define RTM_LOSING	0x5	/* Kernel Suspects Partitioning */
244#define RTM_REDIRECT	0x6	/* Told to use different route */
245#define RTM_MISS	0x7	/* Lookup failed on this address */
246#define RTM_LOCK	0x8	/* fix specified metrics */
247#define RTM_OLDADD	0x9	/* caused by SIOCADDRT */
248#define RTM_OLDDEL	0xa	/* caused by SIOCDELRT */
249// #define RTM_RESOLVE	0xb	/* req to resolve dst to LL addr */
250#define RTM_ONEWADDR	0xc	/* Old (pre-8.0) RTM_NEWADDR message */
251#define RTM_ODELADDR	0xd	/* Old (pre-8.0) RTM_DELADDR message */
252#define RTM_OOIFINFO	0xe	/* Old (pre-1.5) RTM_IFINFO message */
253#define RTM_OIFINFO	0xf	/* Old (pre-64bit time) RTM_IFINFO message */
254#define	RTM_IFANNOUNCE	0x10	/* iface arrival/departure */
255#define	RTM_IEEE80211	0x11	/* IEEE80211 wireless event */
256#define	RTM_SETGATE	0x12	/* set prototype gateway for clones
257				 * (see example in arp_rtrequest).
258				 */
259#define	RTM_LLINFO_UPD	0x13	/* indication to ARP/NDP/etc. that link-layer
260				 * address has changed
261				 */
262#define RTM_IFINFO	0x14	/* iface/link going up/down etc. */
263#define RTM_OCHGADDR	0x15	/* Old (pre-8.0) RTM_CHGADDR message */
264#define RTM_NEWADDR	0x16	/* address being added to iface */
265#define RTM_DELADDR	0x17	/* address being removed from iface */
266#define RTM_CHGADDR	0x18	/* address properties changed */
267
268/*
269 * setsockopt defines used for the filtering.
270 */
271#define	RO_MSGFILTER	1	/* array of which rtm_type to send to client */
272
273#define RTV_MTU		0x1	/* init or lock _mtu */
274#define RTV_HOPCOUNT	0x2	/* init or lock _hopcount */
275#define RTV_EXPIRE	0x4	/* init or lock _expire */
276#define RTV_RPIPE	0x8	/* init or lock _recvpipe */
277#define RTV_SPIPE	0x10	/* init or lock _sendpipe */
278#define RTV_SSTHRESH	0x20	/* init or lock _ssthresh */
279#define RTV_RTT		0x40	/* init or lock _rtt */
280#define RTV_RTTVAR	0x80	/* init or lock _rttvar */
281
282#define RTVBITS "\020\1MTU\2HOPCOUNT\3EXPIRE\4RECVPIPE\5SENDPIPE" \
283    "\6SSTHRESH\7RTT\010RTTVAR"
284
285/*
286 * Bitmask values for rtm_addr.
287 */
288#define RTA_DST		0x1	/* destination sockaddr present */
289#define RTA_GATEWAY	0x2	/* gateway sockaddr present */
290#define RTA_NETMASK	0x4	/* netmask sockaddr present */
291#define RTA_GENMASK	0x8	/* cloning mask sockaddr present */
292#define RTA_IFP		0x10	/* interface name sockaddr present */
293#define RTA_IFA		0x20	/* interface addr sockaddr present */
294#define RTA_AUTHOR	0x40	/* sockaddr for author of redirect */
295#define RTA_BRD		0x80	/* for NEWADDR, broadcast or p-p dest addr */
296#define RTA_TAG		0x100	/* route tag */
297
298#define RTABITS "\020\1DST\2GATEWAY\3NETMASK\4GENMASK\5IFP\6IFA\7AUTHOR" \
299    "\010BRD\011TAG"
300
301/*
302 * Index offsets for sockaddr array for alternate internal encoding.
303 */
304#define RTAX_DST	0	/* destination sockaddr present */
305#define RTAX_GATEWAY	1	/* gateway sockaddr present */
306#define RTAX_NETMASK	2	/* netmask sockaddr present */
307#define RTAX_GENMASK	3	/* cloning mask sockaddr present */
308#define RTAX_IFP	4	/* interface name sockaddr present */
309#define RTAX_IFA	5	/* interface addr sockaddr present */
310#define RTAX_AUTHOR	6	/* sockaddr for author of redirect */
311#define RTAX_BRD	7	/* for NEWADDR, broadcast or p-p dest addr */
312#define RTAX_TAG	8	/* route tag */
313#define RTAX_MAX	9	/* size of array to allocate */
314
315#define RT_ROUNDUP2(a, n)	((a) > 0 ? (1 + (((a) - 1U) | ((n) - 1))) : (n))
316#define RT_ROUNDUP(a)		RT_ROUNDUP2((a), sizeof(uint64_t))
317#define RT_ADVANCE(x, n)	(x += RT_ROUNDUP((n)->sa_len))
318
319struct rt_addrinfo {
320	int	rti_addrs;
321	const struct	sockaddr *rti_info[RTAX_MAX];
322	int	rti_flags;
323	struct	ifaddr *rti_ifa;
324	struct	ifnet *rti_ifp;
325};
326
327struct route_cb {
328	int	ip_count;
329	int	ip6_count;
330	int	unused1;
331	int	mpls_count;
332	int	any_count;
333};
334
335/*
336 * This structure, and the prototypes for the rt_timer_{init,remove_all,
337 * add,timer} functions all used with the kind permission of BSDI.
338 * These allow functions to be called for routes at specific times.
339 */
340
341struct rttimer {
342	TAILQ_ENTRY(rttimer)	rtt_next;  /* entry on timer queue */
343	LIST_ENTRY(rttimer) 	rtt_link;  /* multiple timers per rtentry */
344	struct rttimer_queue   *rtt_queue; /* back pointer to queue */
345	struct rtentry         *rtt_rt;    /* Back pointer to the route */
346	void		      (*rtt_func)(struct rtentry *, struct rttimer *);
347	time_t          	rtt_time;  /* When this timer was registered */
348};
349
350struct rttimer_queue {
351	long				rtq_timeout;
352	unsigned long			rtq_count;
353	TAILQ_HEAD(, rttimer)		rtq_head;
354	LIST_ENTRY(rttimer_queue)	rtq_link;
355};
356
357
358struct rtbl;
359typedef struct rtbl rtbl_t;
360
361#ifdef _KERNEL
362
363struct rtbl {
364	struct radix_node_head t_rnh;
365};
366
367struct rt_walkarg {
368	int	w_op;
369	int	w_arg;
370	int	w_given;
371	int	w_needed;
372	void *	w_where;
373	int	w_tmemsize;
374	int	w_tmemneeded;
375	void *	w_tmem;
376};
377
378#if 0
379#define	RT_DPRINTF(__fmt, ...)	do { } while (/*CONSTCOND*/0)
380#else
381#define	RT_DPRINTF(__fmt, ...)	/* do nothing */
382#endif
383
384struct rtwalk {
385	int (*rw_f)(struct rtentry *, void *);
386	void *rw_v;
387};
388
389/*
390 * Global data specific to the routing socket.
391 */
392struct route_info {
393	struct sockaddr ri_dst;
394	struct sockaddr ri_src;
395	struct route_cb ri_cb;
396	int ri_maxqlen;
397	struct ifqueue ri_intrq;
398	void *ri_sih;
399};
400
401extern	struct	route_info route_info;
402extern	struct	rtstat	rtstat;
403
404struct socket;
405
406void	rt_init(void);
407
408int	rt_timer_add(struct rtentry *,
409	    void(*)(struct rtentry *, struct rttimer *),
410	    struct rttimer_queue *);
411unsigned long
412	rt_timer_count(struct rttimer_queue *);
413void	rt_timer_queue_change(struct rttimer_queue *, long);
414struct rttimer_queue *
415	rt_timer_queue_create(u_int);
416void	rt_timer_queue_destroy(struct rttimer_queue *);
417
418void	rt_free(struct rtentry *);
419void	rt_unref(struct rtentry *);
420
421int	rt_update(struct rtentry *, struct rt_addrinfo *, void *);
422int	rt_update_prepare(struct rtentry *);
423void	rt_update_finish(struct rtentry *);
424
425void	rt_newmsg(const int, const struct rtentry *);
426struct rtentry *
427	rtalloc1(const struct sockaddr *, int);
428int	rtinit(struct ifaddr *, int, int);
429void	rtredirect(const struct sockaddr *, const struct sockaddr *,
430	    const struct sockaddr *, int, const struct sockaddr *,
431	    struct rtentry **);
432int	rtrequest(int, const struct sockaddr *,
433	    const struct sockaddr *, const struct sockaddr *, int,
434	    struct rtentry **);
435int	rtrequest1(int, struct rt_addrinfo *, struct rtentry **);
436int	rtrequest_newmsg(const int, const struct sockaddr *,
437	    const struct sockaddr *, const struct sockaddr *, const int);
438
439int	rt_ifa_addlocal(struct ifaddr *);
440int	rt_ifa_remlocal(struct ifaddr *, struct ifaddr *);
441struct ifaddr *
442	rt_get_ifa(struct rtentry *);
443void	rt_replace_ifa(struct rtentry *, struct ifaddr *);
444int	rt_setgate(struct rtentry *, const struct sockaddr *);
445
446const struct sockaddr *
447	rt_settag(struct rtentry *, const struct sockaddr *);
448struct sockaddr *
449	rt_gettag(const struct rtentry *);
450
451int	rt_check_reject_route(const struct rtentry *, const struct ifnet *);
452void	rt_delete_matched_entries(sa_family_t,
453	    int (*)(struct rtentry *, void *), void *);
454int	rt_walktree(sa_family_t, int (*)(struct rtentry *, void *), void *);
455
456static __inline void
457rt_assert_referenced(const struct rtentry *rt)
458{
459
460	KASSERT(rt->rt_refcnt > 0);
461}
462
463void	rtcache_copy(struct route *, struct route *);
464void	rtcache_free(struct route *);
465struct rtentry *
466	rtcache_init(struct route *);
467struct rtentry *
468	rtcache_init_noclone(struct route *);
469struct rtentry *
470	rtcache_lookup2(struct route *, const struct sockaddr *, int,
471	    int *);
472int	rtcache_setdst(struct route *, const struct sockaddr *);
473struct rtentry *
474	rtcache_update(struct route *, int);
475
476static __inline void
477rtcache_invariants(const struct route *ro)
478{
479
480	KASSERT(ro->ro_sa != NULL || ro->_ro_rt == NULL);
481}
482
483static __inline struct rtentry *
484rtcache_lookup1(struct route *ro, const struct sockaddr *dst, int clone)
485{
486	int hit;
487
488	return rtcache_lookup2(ro, dst, clone, &hit);
489}
490
491static __inline struct rtentry *
492rtcache_lookup(struct route *ro, const struct sockaddr *dst)
493{
494	return rtcache_lookup1(ro, dst, 1);
495}
496
497static __inline const struct sockaddr *
498rtcache_getdst(const struct route *ro)
499{
500
501	rtcache_invariants(ro);
502	return ro->ro_sa;
503}
504
505struct rtentry *
506	rtcache_validate(struct route *);
507
508void	rtcache_unref(struct rtentry *, struct route *);
509
510/* rtsock */
511void	rt_ieee80211msg(struct ifnet *, int, void *, size_t);
512void	rt_ifannouncemsg(struct ifnet *, int);
513void	rt_ifmsg(struct ifnet *);
514void	rt_missmsg(int, const struct rt_addrinfo *, int, int);
515struct mbuf *
516	rt_msg1(int, struct rt_addrinfo *, void *, int);
517int	rt_msg3(int, struct rt_addrinfo *, void *, struct rt_walkarg *, int *);
518void	rt_addrmsg(int, struct ifaddr *);
519void	rt_addrmsg_rt(int, struct ifaddr *, int, struct rtentry *);
520void	route_enqueue(struct mbuf *, int);
521
522struct llentry;
523void	rt_clonedmsg(const struct sockaddr *, const struct ifnet *,
524	    const struct rtentry *);
525
526void	rt_setmetrics(void *, struct rtentry *);
527
528/* rtbl */
529int	rt_addaddr(rtbl_t *, struct rtentry *, const struct sockaddr *);
530void	rt_assert_inactive(const struct rtentry *);
531struct rtentry *
532	rt_deladdr(rtbl_t *, const struct sockaddr *,
533	    const struct sockaddr *);
534rtbl_t *rt_gettable(sa_family_t);
535int	rt_inithead(rtbl_t **, int);
536struct rtentry *
537	rt_lookup(rtbl_t *, const struct sockaddr *,
538	    const struct sockaddr *);
539struct rtentry *
540	rt_matchaddr(rtbl_t *, const struct sockaddr *);
541int	rt_refines(const struct sockaddr *, const struct sockaddr *);
542int	rtbl_walktree(sa_family_t, int (*)(struct rtentry *, void *), void *);
543struct rtentry *
544	rtbl_search_matched_entry(sa_family_t,
545	    int (*)(struct rtentry *, void *), void *);
546void	rtbl_init(void);
547
548void sysctl_net_route_setup(struct sysctllog **, int, const char *);
549
550#endif /* _KERNEL */
551
552#endif /* !_NET_ROUTE_H_ */
553