xref: /openbsd-src/sys/net/rtable.c (revision ae3cb403620ab940fbaabb3055fac045a63d56b7)
1 /*	$OpenBSD: rtable.c,v 1.63 2017/09/05 11:15:39 mpi Exp $ */
2 
3 /*
4  * Copyright (c) 2014-2016 Martin Pieuchot
5  *
6  * Permission to use, copy, modify, and distribute this software for any
7  * purpose with or without fee is hereby granted, provided that the above
8  * copyright notice and this permission notice appear in all copies.
9  *
10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
13  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
17  */
18 
19 #ifndef _KERNEL
20 #include "kern_compat.h"
21 #else
22 #include <sys/param.h>
23 #include <sys/systm.h>
24 #include <sys/socket.h>
25 #include <sys/malloc.h>
26 #include <sys/queue.h>
27 #include <sys/domain.h>
28 #include <sys/srp.h>
29 #endif
30 
31 #include <net/rtable.h>
32 #include <net/route.h>
33 
34 /*
35  * Structures used by rtable_get() to retrieve the corresponding
36  * routing table for a given pair of ``af'' and ``rtableid''.
37  *
38  * Note that once allocated routing table heads are never freed.
39  * This way we do not need to reference count them.
40  *
41  *	afmap		    rtmap/dommp
42  *   -----------          ---------     -----
43  *   |   0     |--------> | 0 | 0 | ... | 0 |	Array mapping rtableid (=index)
44  *   -----------          ---------     -----   to rdomain/loopback (=value).
45  *   | AF_INET |.
46  *   ----------- `.       .---------.     .---------.
47  *       ...	   `----> | rtable0 | ... | rtableN |	Array of pointers for
48  *   -----------          '---------'     '---------'	IPv4 routing tables
49  *   | AF_MPLS |					indexed by ``rtableid''.
50  *   -----------
51  */
52 struct srp	  *afmap;
53 uint8_t		   af2idx[AF_MAX+1];	/* To only allocate supported AF */
54 uint8_t		   af2idx_max;
55 
56 /* Array of routing table pointers. */
57 struct rtmap {
58 	unsigned int	   limit;
59 	void		 **tbl;
60 };
61 
62 /*
63  * Array of rtableid -> rdomain mapping.
64  *
65  * Only used for the first index as describbed above.
66  */
67 struct dommp {
68 	unsigned int	   limit;
69 	/*
70 	 * Array to get the routing domain and loopback interface related to
71 	 * a routing table. Format:
72 	 *
73 	 * 8 unused bits | 16 bits for loopback index | 8 bits for rdomain
74 	 */
75 	unsigned int	  *value;
76 };
77 
78 unsigned int	   rtmap_limit = 0;
79 
80 void		   rtmap_init(void);
81 void		   rtmap_grow(unsigned int, sa_family_t);
82 void		   rtmap_dtor(void *, void *);
83 
84 struct srp_gc	   rtmap_gc = SRP_GC_INITIALIZER(rtmap_dtor, NULL);
85 
86 void		   rtable_init_backend(unsigned int);
87 void		  *rtable_alloc(unsigned int, unsigned int, unsigned int);
88 void		  *rtable_get(unsigned int, sa_family_t);
89 
90 void
91 rtmap_init(void)
92 {
93 	struct domain	*dp;
94 	int		 i;
95 
96 	/* Start with a single table for every domain that requires it. */
97 	for (i = 0; (dp = domains[i]) != NULL; i++) {
98 		if (dp->dom_rtoffset == 0)
99 			continue;
100 
101 		rtmap_grow(1, dp->dom_family);
102 	}
103 
104 	/* Initialize the rtableid->rdomain mapping table. */
105 	rtmap_grow(1, 0);
106 
107 	rtmap_limit = 1;
108 }
109 
110 /*
111  * Grow the size of the array of routing table for AF ``af'' to ``nlimit''.
112  */
113 void
114 rtmap_grow(unsigned int nlimit, sa_family_t af)
115 {
116 	struct rtmap	*map, *nmap;
117 	int		 i;
118 
119 	KERNEL_ASSERT_LOCKED();
120 
121 	KASSERT(nlimit > rtmap_limit);
122 
123 	nmap = malloc(sizeof(*nmap), M_RTABLE, M_WAITOK);
124 	nmap->limit = nlimit;
125 	nmap->tbl = mallocarray(nlimit, sizeof(*nmap[0].tbl), M_RTABLE,
126 	    M_WAITOK|M_ZERO);
127 
128 	map = srp_get_locked(&afmap[af2idx[af]]);
129 	if (map != NULL) {
130 		KASSERT(map->limit == rtmap_limit);
131 
132 		for (i = 0; i < map->limit; i++)
133 			nmap->tbl[i] = map->tbl[i];
134 	}
135 
136 	srp_update_locked(&rtmap_gc, &afmap[af2idx[af]], nmap);
137 }
138 
139 void
140 rtmap_dtor(void *null, void *xmap)
141 {
142 	struct rtmap	*map = xmap;
143 
144 	/*
145 	 * doesnt need to be serialized since this is the last reference
146 	 * to this map. there's nothing to race against.
147 	 */
148 	free(map->tbl, M_RTABLE, map->limit * sizeof(*map[0].tbl));
149 	free(map, M_RTABLE, sizeof(*map));
150 }
151 
152 void
153 rtable_init(void)
154 {
155 	struct domain	*dp;
156 	unsigned int	 keylen = 0;
157 	int		 i;
158 
159 	KASSERT(sizeof(struct rtmap) == sizeof(struct dommp));
160 
161 	/* We use index 0 for the rtable/rdomain map. */
162 	af2idx_max = 1;
163 	memset(af2idx, 0, sizeof(af2idx));
164 
165 	/*
166 	 * Compute the maximum supported key length in case the routing
167 	 * table backend needs it.
168 	 */
169 	for (i = 0; (dp = domains[i]) != NULL; i++) {
170 		if (dp->dom_rtoffset == 0)
171 			continue;
172 
173 		af2idx[dp->dom_family] = af2idx_max++;
174 		if (dp->dom_rtkeylen > keylen)
175 			keylen = dp->dom_rtkeylen;
176 
177 	}
178 	rtable_init_backend(keylen);
179 
180 	/*
181 	 * Allocate AF-to-id table now that we now how many AFs this
182 	 * kernel supports.
183 	 */
184 	afmap = mallocarray(af2idx_max + 1, sizeof(*afmap), M_RTABLE,
185 	    M_WAITOK|M_ZERO);
186 
187 	rtmap_init();
188 
189 	if (rtable_add(0) != 0)
190 		panic("unable to create default routing table");
191 }
192 
193 int
194 rtable_add(unsigned int id)
195 {
196 	struct domain	*dp;
197 	void		*tbl;
198 	struct rtmap	*map;
199 	struct dommp	*dmm;
200 	sa_family_t	 af;
201 	unsigned int	 off, alen;
202 	int		 i;
203 
204 	KERNEL_ASSERT_LOCKED();
205 
206 	if (id > RT_TABLEID_MAX)
207 		return (EINVAL);
208 
209 	if (rtable_exists(id))
210 		return (EEXIST);
211 
212 	for (i = 0; (dp = domains[i]) != NULL; i++) {
213 		if (dp->dom_rtoffset == 0)
214 			continue;
215 
216 		af = dp->dom_family;
217 		off = dp->dom_rtoffset;
218 		alen = dp->dom_maxplen;
219 
220 		if (id >= rtmap_limit)
221 			rtmap_grow(id + 1, af);
222 
223 		tbl = rtable_alloc(id, alen, off);
224 		if (tbl == NULL)
225 			return (ENOMEM);
226 
227 		map = srp_get_locked(&afmap[af2idx[af]]);
228 		map->tbl[id] = tbl;
229 	}
230 
231 	/* Reflect possible growth. */
232 	if (id >= rtmap_limit) {
233 		rtmap_grow(id + 1, 0);
234 		rtmap_limit = id + 1;
235 	}
236 
237 	/* Use main rtable/rdomain by default. */
238 	dmm = srp_get_locked(&afmap[0]);
239 	dmm->value[id] = 0;
240 
241 	return (0);
242 }
243 
244 void *
245 rtable_get(unsigned int rtableid, sa_family_t af)
246 {
247 	struct rtmap	*map;
248 	void		*tbl = NULL;
249 	struct srp_ref	 sr;
250 
251 	if (af >= nitems(af2idx) || af2idx[af] == 0)
252 		return (NULL);
253 
254 	map = srp_enter(&sr, &afmap[af2idx[af]]);
255 	if (rtableid < map->limit)
256 		tbl = map->tbl[rtableid];
257 	srp_leave(&sr);
258 
259 	return (tbl);
260 }
261 
262 int
263 rtable_exists(unsigned int rtableid)
264 {
265 	struct domain	*dp;
266 	void		*tbl;
267 	int		 i;
268 
269 	for (i = 0; (dp = domains[i]) != NULL; i++) {
270 		if (dp->dom_rtoffset == 0)
271 			continue;
272 
273 		tbl = rtable_get(rtableid, dp->dom_family);
274 		if (tbl != NULL)
275 			return (1);
276 	}
277 
278 	return (0);
279 }
280 
281 unsigned int
282 rtable_l2(unsigned int rtableid)
283 {
284 	struct dommp	*dmm;
285 	unsigned int	 rdomain = 0;
286 	struct srp_ref	 sr;
287 
288 	dmm = srp_enter(&sr, &afmap[0]);
289 	if (rtableid < dmm->limit)
290 		rdomain = (dmm->value[rtableid] & RT_TABLEID_MASK);
291 	srp_leave(&sr);
292 
293 	return (rdomain);
294 }
295 
296 unsigned int
297 rtable_loindex(unsigned int rtableid)
298 {
299 	struct dommp	*dmm;
300 	unsigned int	 loifidx = 0;
301 	struct srp_ref	 sr;
302 
303 	dmm = srp_enter(&sr, &afmap[0]);
304 	if (rtableid < dmm->limit)
305 		loifidx = (dmm->value[rtableid] >> RT_TABLEID_BITS);
306 	srp_leave(&sr);
307 
308 	return (loifidx);
309 }
310 
311 void
312 rtable_l2set(unsigned int rtableid, unsigned int rdomain, unsigned int loifidx)
313 {
314 	struct dommp	*dmm;
315 	unsigned int	 value;
316 
317 	KERNEL_ASSERT_LOCKED();
318 
319 	if (!rtable_exists(rtableid) || !rtable_exists(rdomain))
320 		return;
321 
322 	value = (rdomain & RT_TABLEID_MASK) | (loifidx << RT_TABLEID_BITS);
323 
324 	dmm = srp_get_locked(&afmap[0]);
325 	dmm->value[rtableid] = value;
326 }
327 
328 
329 static inline uint8_t	*satoaddr(struct art_root *, struct sockaddr *);
330 
331 int	an_match(struct art_node *, struct sockaddr *, int);
332 void	rtentry_ref(void *, void *);
333 void	rtentry_unref(void *, void *);
334 
335 void	rtable_mpath_insert(struct art_node *, struct rtentry *);
336 
337 struct srpl_rc rt_rc = SRPL_RC_INITIALIZER(rtentry_ref, rtentry_unref, NULL);
338 
339 void
340 rtable_init_backend(unsigned int keylen)
341 {
342 	art_init();
343 }
344 
345 void *
346 rtable_alloc(unsigned int rtableid, unsigned int alen, unsigned int off)
347 {
348 	return (art_alloc(rtableid, alen, off));
349 }
350 
351 struct rtentry *
352 rtable_lookup(unsigned int rtableid, struct sockaddr *dst,
353     struct sockaddr *mask, struct sockaddr *gateway, uint8_t prio)
354 {
355 	struct art_root			*ar;
356 	struct art_node			*an;
357 	struct rtentry			*rt = NULL;
358 	struct srp_ref			 sr, nsr;
359 	uint8_t				*addr;
360 	int				 plen;
361 
362 	ar = rtable_get(rtableid, dst->sa_family);
363 	if (ar == NULL)
364 		return (NULL);
365 
366 	addr = satoaddr(ar, dst);
367 
368 	/* No need for a perfect match. */
369 	if (mask == NULL) {
370 		an = art_match(ar, addr, &nsr);
371 		if (an == NULL)
372 			goto out;
373 	} else {
374 		plen = rtable_satoplen(dst->sa_family, mask);
375 		if (plen == -1)
376 			return (NULL);
377 
378 		an = art_lookup(ar, addr, plen, &nsr);
379 
380 		/* Make sure we've got a perfect match. */
381 		if (!an_match(an, dst, plen))
382 			goto out;
383 	}
384 
385 	SRPL_FOREACH(rt, &sr, &an->an_rtlist, rt_next) {
386 		if (prio != RTP_ANY &&
387 		    (rt->rt_priority & RTP_MASK) != (prio & RTP_MASK))
388 			continue;
389 
390 		if (gateway == NULL)
391 			break;
392 
393 		if (rt->rt_gateway->sa_len == gateway->sa_len &&
394 		    memcmp(rt->rt_gateway, gateway, gateway->sa_len) == 0)
395 			break;
396 	}
397 	if (rt != NULL)
398 		rtref(rt);
399 
400 	SRPL_LEAVE(&sr);
401 out:
402 	srp_leave(&nsr);
403 
404 	return (rt);
405 }
406 
407 struct rtentry *
408 rtable_match(unsigned int rtableid, struct sockaddr *dst, uint32_t *src)
409 {
410 	struct art_root			*ar;
411 	struct art_node			*an;
412 	struct rtentry			*rt = NULL;
413 	struct srp_ref			 sr, nsr;
414 	uint8_t				*addr;
415 	int				 hash;
416 
417 	ar = rtable_get(rtableid, dst->sa_family);
418 	if (ar == NULL)
419 		return (NULL);
420 
421 	addr = satoaddr(ar, dst);
422 
423 	an = art_match(ar, addr, &nsr);
424 	if (an == NULL)
425 		goto out;
426 
427 	rt = SRPL_FIRST(&sr, &an->an_rtlist);
428 	rtref(rt);
429 	SRPL_LEAVE(&sr);
430 
431 	/* Gateway selection by Hash-Threshold (RFC 2992) */
432 	if ((hash = rt_hash(rt, dst, src)) != -1) {
433 		struct rtentry		*mrt;
434 		int			 threshold, npaths = 0;
435 
436 		KASSERT(hash <= 0xffff);
437 
438 		SRPL_FOREACH(mrt, &sr, &an->an_rtlist, rt_next) {
439 			/* Only count nexthops with the same priority. */
440 			if (mrt->rt_priority == rt->rt_priority)
441 				npaths++;
442 		}
443 		SRPL_LEAVE(&sr);
444 
445 		threshold = (0xffff / npaths) + 1;
446 
447 		/*
448 		 * we have no protection against concurrent modification of the
449 		 * route list attached to the node, so we won't necessarily
450 		 * have the same number of routes.  for most modifications,
451 		 * we'll pick a route that we wouldn't have if we only saw the
452 		 * list before or after the change.  if we were going to use
453 		 * the last available route, but it got removed, we'll hit
454 		 * the end of the list and then pick the first route.
455 		 */
456 
457 		mrt = SRPL_FIRST(&sr, &an->an_rtlist);
458 		while (hash > threshold && mrt != NULL) {
459 			if (mrt->rt_priority == rt->rt_priority)
460 				hash -= threshold;
461 			mrt = SRPL_FOLLOW(&sr, mrt, rt_next);
462 		}
463 
464 		if (mrt != NULL) {
465 			rtref(mrt);
466 			rtfree(rt);
467 			rt = mrt;
468 		}
469 		SRPL_LEAVE(&sr);
470 	}
471 out:
472 	srp_leave(&nsr);
473 	return (rt);
474 }
475 
476 int
477 rtable_insert(unsigned int rtableid, struct sockaddr *dst,
478     struct sockaddr *mask, struct sockaddr *gateway, uint8_t prio,
479     struct rtentry *rt)
480 {
481 	struct rtentry			*mrt;
482 	struct srp_ref			 sr;
483 	struct art_root			*ar;
484 	struct art_node			*an, *prev;
485 	uint8_t				*addr;
486 	int				 plen;
487 	unsigned int			 rt_flags;
488 	int				 error = 0;
489 
490 	ar = rtable_get(rtableid, dst->sa_family);
491 	if (ar == NULL)
492 		return (EAFNOSUPPORT);
493 
494 	addr = satoaddr(ar, dst);
495 	plen = rtable_satoplen(dst->sa_family, mask);
496 	if (plen == -1)
497 		return (EINVAL);
498 
499 	rtref(rt); /* guarantee rtfree won't do anything during insert */
500 	rw_enter_write(&ar->ar_lock);
501 
502 	/* Do not permit exactly the same dst/mask/gw pair. */
503 	an = art_lookup(ar, addr, plen, &sr);
504 	srp_leave(&sr); /* an can't go away while we have the lock */
505 	if (an_match(an, dst, plen)) {
506 		struct rtentry  *mrt;
507 		int		 mpathok = ISSET(rt->rt_flags, RTF_MPATH);
508 
509 		SRPL_FOREACH_LOCKED(mrt, &an->an_rtlist, rt_next) {
510 			if (prio != RTP_ANY &&
511 			    (mrt->rt_priority & RTP_MASK) != (prio & RTP_MASK))
512 				continue;
513 
514 			if (!mpathok ||
515 			    (mrt->rt_gateway->sa_len == gateway->sa_len &&
516 			    !memcmp(mrt->rt_gateway, gateway, gateway->sa_len))){
517 				error = EEXIST;
518 				goto leave;
519 			}
520 		}
521 	}
522 
523 	an = art_get(dst, plen);
524 	if (an == NULL) {
525 		error = ENOBUFS;
526 		goto leave;
527 	}
528 
529 	/* prepare for immediate operation if insert succeeds */
530 	rt_flags = rt->rt_flags;
531 	rt->rt_flags &= ~RTF_MPATH;
532 	rt->rt_dest = dst;
533 	rt->rt_plen = plen;
534 	SRPL_INSERT_HEAD_LOCKED(&rt_rc, &an->an_rtlist, rt, rt_next);
535 
536 	prev = art_insert(ar, an, addr, plen);
537 	if (prev != an) {
538 		SRPL_REMOVE_LOCKED(&rt_rc, &an->an_rtlist, rt, rtentry,
539 		    rt_next);
540 		rt->rt_flags = rt_flags;
541 		art_put(an);
542 
543 		if (prev == NULL) {
544 			error = ESRCH;
545 			goto leave;
546 		}
547 
548 		an = prev;
549 
550 		mrt = SRPL_FIRST_LOCKED(&an->an_rtlist);
551 		KASSERT(mrt != NULL);
552 		KASSERT((rt->rt_flags & RTF_MPATH) || mrt->rt_priority != prio);
553 
554 		/*
555 		 * An ART node with the same destination/netmask already
556 		 * exists, MPATH conflict must have been already checked.
557 		 */
558 		if (rt->rt_flags & RTF_MPATH) {
559 			/*
560 			 * Only keep the RTF_MPATH flag if two routes have
561 			 * the same gateway.
562 			 */
563 			rt->rt_flags &= ~RTF_MPATH;
564 			SRPL_FOREACH_LOCKED(mrt, &an->an_rtlist, rt_next) {
565 				if (mrt->rt_priority == prio) {
566 					mrt->rt_flags |= RTF_MPATH;
567 					rt->rt_flags |= RTF_MPATH;
568 				}
569 			}
570 		}
571 
572 		/* Put newly inserted entry at the right place. */
573 		rtable_mpath_insert(an, rt);
574 	}
575 leave:
576 	rw_exit_write(&ar->ar_lock);
577 	rtfree(rt);
578 	return (error);
579 }
580 
581 int
582 rtable_delete(unsigned int rtableid, struct sockaddr *dst,
583     struct sockaddr *mask, struct rtentry *rt)
584 {
585 	struct art_root			*ar;
586 	struct art_node			*an;
587 	struct srp_ref			 sr;
588 	uint8_t				*addr;
589 	int				 plen;
590 	struct rtentry			*mrt;
591 	int				 npaths = 0;
592 	int				 error = 0;
593 
594 	ar = rtable_get(rtableid, dst->sa_family);
595 	if (ar == NULL)
596 		return (EAFNOSUPPORT);
597 
598 	addr = satoaddr(ar, dst);
599 	plen = rtable_satoplen(dst->sa_family, mask);
600 
601 	rtref(rt); /* guarantee rtfree won't do anything under ar_lock */
602 	rw_enter_write(&ar->ar_lock);
603 	an = art_lookup(ar, addr, plen, &sr);
604 	srp_leave(&sr); /* an can't go away while we have the lock */
605 
606 	/* Make sure we've got a perfect match. */
607 	if (!an_match(an, dst, plen)) {
608 		error = ESRCH;
609 		goto leave;
610 	}
611 
612 	/*
613 	 * If other multipath route entries are still attached to
614 	 * this ART node we only have to unlink it.
615 	 */
616 	SRPL_FOREACH_LOCKED(mrt, &an->an_rtlist, rt_next)
617 		npaths++;
618 
619 	if (npaths > 1) {
620 		KASSERT(rt->rt_refcnt >= 1);
621 		SRPL_REMOVE_LOCKED(&rt_rc, &an->an_rtlist, rt, rtentry,
622 		    rt_next);
623 
624 		mrt = SRPL_FIRST_LOCKED(&an->an_rtlist);
625 		if (npaths == 2)
626 			mrt->rt_flags &= ~RTF_MPATH;
627 
628 		goto leave;
629 	}
630 
631 	if (art_delete(ar, an, addr, plen) == NULL)
632 		panic("art_delete failed to find node %p", an);
633 
634 	KASSERT(rt->rt_refcnt >= 1);
635 	SRPL_REMOVE_LOCKED(&rt_rc, &an->an_rtlist, rt, rtentry, rt_next);
636 	art_put(an);
637 
638 leave:
639 	rw_exit_write(&ar->ar_lock);
640 	rtfree(rt);
641 
642 	return (error);
643 }
644 
645 struct rtable_walk_cookie {
646 	int		(*rwc_func)(struct rtentry *, void *, unsigned int);
647 	void		 *rwc_arg;
648 	unsigned int	  rwc_rid;
649 };
650 
651 /*
652  * Helper for rtable_walk to keep the ART code free from any "struct rtentry".
653  */
654 int
655 rtable_walk_helper(struct art_node *an, void *xrwc)
656 {
657 	struct srp_ref			 sr;
658 	struct rtable_walk_cookie	*rwc = xrwc;
659 	struct rtentry			*rt;
660 	int				 error = 0;
661 
662 	SRPL_FOREACH(rt, &sr, &an->an_rtlist, rt_next) {
663 		if ((error = (*rwc->rwc_func)(rt, rwc->rwc_arg, rwc->rwc_rid)))
664 			break;
665 	}
666 	SRPL_LEAVE(&sr);
667 
668 	return (error);
669 }
670 
671 int
672 rtable_walk(unsigned int rtableid, sa_family_t af,
673     int (*func)(struct rtentry *, void *, unsigned int), void *arg)
674 {
675 	struct art_root			*ar;
676 	struct rtable_walk_cookie	 rwc;
677 	int				 error;
678 
679 	ar = rtable_get(rtableid, af);
680 	if (ar == NULL)
681 		return (EAFNOSUPPORT);
682 
683 	rwc.rwc_func = func;
684 	rwc.rwc_arg = arg;
685 	rwc.rwc_rid = rtableid;
686 
687 	while ((error = art_walk(ar, rtable_walk_helper, &rwc)) == EAGAIN)
688 		continue;
689 
690 	return (error);
691 }
692 
693 struct rtentry *
694 rtable_iterate(struct rtentry *rt0)
695 {
696 	struct rtentry *rt = NULL;
697 	struct srp_ref sr;
698 
699 	rt = SRPL_NEXT(&sr, rt0, rt_next);
700 	if (rt != NULL)
701 		rtref(rt);
702 	SRPL_LEAVE(&sr);
703 	rtfree(rt0);
704 	return (rt);
705 }
706 
707 int
708 rtable_mpath_capable(unsigned int rtableid, sa_family_t af)
709 {
710 	return (1);
711 }
712 
713 int
714 rtable_mpath_reprio(unsigned int rtableid, struct sockaddr *dst,
715     struct sockaddr *mask, uint8_t prio, struct rtentry *rt)
716 {
717 	struct art_root			*ar;
718 	struct art_node			*an;
719 	struct srp_ref			 sr;
720 	uint8_t				*addr;
721 	int				 plen;
722 	int				 error = 0;
723 
724 	ar = rtable_get(rtableid, dst->sa_family);
725 	if (ar == NULL)
726 		return (EAFNOSUPPORT);
727 
728 	addr = satoaddr(ar, dst);
729 	plen = rtable_satoplen(dst->sa_family, mask);
730 
731 	rw_enter_write(&ar->ar_lock);
732 	an = art_lookup(ar, addr, plen, &sr);
733 	srp_leave(&sr); /* an can't go away while we have the lock */
734 
735 	/* Make sure we've got a perfect match. */
736 	if (!an_match(an, dst, plen)) {
737 		error = ESRCH;
738 	} else if (SRPL_FIRST_LOCKED(&an->an_rtlist) == rt &&
739 		SRPL_NEXT_LOCKED(rt, rt_next) == NULL) {
740 		/*
741 		 * If there's only one entry on the list do not go
742 		 * through an insert/remove cycle.  This is done to
743 		 * guarantee that ``an->an_rtlist''  is never empty
744 		 * when a node is in the tree.
745 		 */
746 		rt->rt_priority = prio;
747 	} else {
748 		rtref(rt); /* keep rt alive in between remove and insert */
749 		SRPL_REMOVE_LOCKED(&rt_rc, &an->an_rtlist,
750 		    rt, rtentry, rt_next);
751 		rt->rt_priority = prio;
752 		rtable_mpath_insert(an, rt);
753 		rtfree(rt);
754 		error = EAGAIN;
755 	}
756 	rw_exit_write(&ar->ar_lock);
757 
758 	return (error);
759 }
760 
761 void
762 rtable_mpath_insert(struct art_node *an, struct rtentry *rt)
763 {
764 	struct rtentry			*mrt, *prt = NULL;
765 	uint8_t				 prio = rt->rt_priority;
766 
767 	if ((mrt = SRPL_FIRST_LOCKED(&an->an_rtlist)) == NULL) {
768 		SRPL_INSERT_HEAD_LOCKED(&rt_rc, &an->an_rtlist, rt, rt_next);
769 		return;
770 	}
771 
772 	/* Iterate until we find the route to be placed after ``rt''. */
773 	while (mrt->rt_priority <= prio && SRPL_NEXT_LOCKED(mrt, rt_next)) {
774 		prt = mrt;
775 		mrt = SRPL_NEXT_LOCKED(mrt, rt_next);
776 	}
777 
778 	if (mrt->rt_priority <= prio) {
779 		SRPL_INSERT_AFTER_LOCKED(&rt_rc, mrt, rt, rt_next);
780 	} else if (prt != NULL) {
781 		SRPL_INSERT_AFTER_LOCKED(&rt_rc, prt, rt, rt_next);
782 	} else {
783 		SRPL_INSERT_HEAD_LOCKED(&rt_rc, &an->an_rtlist, rt, rt_next);
784 	}
785 }
786 
787 /*
788  * Returns 1 if ``an'' perfectly matches (``dst'', ``plen''), 0 otherwise.
789  */
790 int
791 an_match(struct art_node *an, struct sockaddr *dst, int plen)
792 {
793 	struct rtentry			*rt;
794 	struct srp_ref			 sr;
795 	int				 match;
796 
797 	if (an == NULL || an->an_plen != plen)
798 		return (0);
799 
800 	rt = SRPL_FIRST(&sr, &an->an_rtlist);
801 	match = (memcmp(rt->rt_dest, dst, dst->sa_len) == 0);
802 	SRPL_LEAVE(&sr);
803 
804 	return (match);
805 }
806 
807 void
808 rtentry_ref(void *null, void *xrt)
809 {
810 	struct rtentry *rt = xrt;
811 
812 	rtref(rt);
813 }
814 
815 void
816 rtentry_unref(void *null, void *xrt)
817 {
818 	struct rtentry *rt = xrt;
819 
820 	rtfree(rt);
821 }
822 
823 /*
824  * Return a pointer to the address (key).  This is an heritage from the
825  * BSD radix tree needed to skip the non-address fields from the flavor
826  * of "struct sockaddr" used by this routing table.
827  */
828 static inline uint8_t *
829 satoaddr(struct art_root *at, struct sockaddr *sa)
830 {
831 	return (((uint8_t *)sa) + at->ar_off);
832 }
833 
834 /*
835  * Return the prefix length of a mask.
836  */
837 int
838 rtable_satoplen(sa_family_t af, struct sockaddr *mask)
839 {
840 	struct domain	*dp;
841 	uint8_t		*ap, *ep;
842 	int		 mlen, plen = 0;
843 	int		 i;
844 
845 	for (i = 0; (dp = domains[i]) != NULL; i++) {
846 		if (dp->dom_rtoffset == 0)
847 			continue;
848 
849 		if (af == dp->dom_family)
850 			break;
851 	}
852 	if (dp == NULL)
853 		return (-1);
854 
855 	/* Host route */
856 	if (mask == NULL)
857 		return (dp->dom_maxplen);
858 
859 	mlen = mask->sa_len;
860 
861 	/* Default route */
862 	if (mlen == 0)
863 		return (0);
864 
865 	ap = (uint8_t *)((uint8_t *)mask) + dp->dom_rtoffset;
866 	ep = (uint8_t *)((uint8_t *)mask) + mlen;
867 	if (ap > ep)
868 		return (-1);
869 
870 	if (ap == ep)
871 		return (0);
872 
873 	/* "Beauty" adapted from sbin/route/show.c ... */
874 	while (ap < ep) {
875 		switch (*ap) {
876 		case 0xff:
877 			plen += 8;
878 			ap++;
879 			break;
880 		case 0xfe:
881 			plen += 7;
882 			ap++;
883 			goto out;
884 		case 0xfc:
885 			plen += 6;
886 			ap++;
887 			goto out;
888 		case 0xf8:
889 			plen += 5;
890 			ap++;
891 			goto out;
892 		case 0xf0:
893 			plen += 4;
894 			ap++;
895 			goto out;
896 		case 0xe0:
897 			plen += 3;
898 			ap++;
899 			goto out;
900 		case 0xc0:
901 			plen += 2;
902 			ap++;
903 			goto out;
904 		case 0x80:
905 			plen += 1;
906 			ap++;
907 			goto out;
908 		case 0x00:
909 			goto out;
910 		default:
911 			/* Non contiguous mask. */
912 			return (-1);
913 		}
914 
915 	}
916 
917 out:
918 #ifdef DIAGNOSTIC
919 	for (; ap < ep; ap++) {
920 		if (*ap != 0x00)
921 			return (-1);
922 	}
923 #endif /* DIAGNOSTIC */
924 
925 	return (plen);
926 }
927