xref: /freebsd/sys/netinet6/in6_rmx.c (revision 1c6d60de932c8553af44629218cb9697bc0f2ef1)
1 /*-
2  * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
3  * 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  * 3. Neither the name of the project 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 PROJECT 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 PROJECT 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  *	$KAME: in6_rmx.c,v 1.11 2001/07/26 06:53:16 jinmei Exp $
30  */
31 
32 /*-
33  * Copyright 1994, 1995 Massachusetts Institute of Technology
34  *
35  * Permission to use, copy, modify, and distribute this software and
36  * its documentation for any purpose and without fee is hereby
37  * granted, provided that both the above copyright notice and this
38  * permission notice appear in all copies, that both the above
39  * copyright notice and this permission notice appear in all
40  * supporting documentation, and that the name of M.I.T. not be used
41  * in advertising or publicity pertaining to distribution of the
42  * software without specific, written prior permission.  M.I.T. makes
43  * no representations about the suitability of this software for any
44  * purpose.  It is provided "as is" without express or implied
45  * warranty.
46  *
47  * THIS SOFTWARE IS PROVIDED BY M.I.T. ``AS IS''.  M.I.T. DISCLAIMS
48  * ALL EXPRESS OR IMPLIED WARRANTIES WITH REGARD TO THIS SOFTWARE,
49  * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
50  * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT
51  * SHALL M.I.T. BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
52  * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
53  * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
54  * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
55  * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
56  * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
57  * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
58  * SUCH DAMAGE.
59  *
60  */
61 
62 #include <sys/cdefs.h>
63 __FBSDID("$FreeBSD$");
64 
65 #include <sys/param.h>
66 #include <sys/systm.h>
67 #include <sys/kernel.h>
68 #include <sys/lock.h>
69 #include <sys/sysctl.h>
70 #include <sys/queue.h>
71 #include <sys/socket.h>
72 #include <sys/socketvar.h>
73 #include <sys/mbuf.h>
74 #include <sys/rwlock.h>
75 #include <sys/syslog.h>
76 #include <sys/callout.h>
77 
78 #include <net/if.h>
79 #include <net/if_var.h>
80 #include <net/route.h>
81 
82 #include <netinet/in.h>
83 #include <netinet/ip_var.h>
84 #include <netinet/in_var.h>
85 
86 #include <netinet/ip6.h>
87 #include <netinet6/ip6_var.h>
88 
89 #include <netinet/icmp6.h>
90 #include <netinet6/nd6.h>
91 
92 #include <netinet/tcp.h>
93 #include <netinet/tcp_seq.h>
94 #include <netinet/tcp_timer.h>
95 #include <netinet/tcp_var.h>
96 
97 extern int	in6_inithead(void **head, int off);
98 #ifdef VIMAGE
99 extern int	in6_detachhead(void **head, int off);
100 #endif
101 
102 /*
103  * Do what we need to do when inserting a route.
104  */
105 static struct radix_node *
106 in6_addroute(void *v_arg, void *n_arg, struct radix_node_head *head,
107     struct radix_node *treenodes)
108 {
109 	struct rtentry *rt = (struct rtentry *)treenodes;
110 	struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)rt_key(rt);
111 	struct radix_node *ret;
112 
113 	RADIX_NODE_HEAD_WLOCK_ASSERT(head);
114 	if (IN6_IS_ADDR_MULTICAST(&sin6->sin6_addr))
115 		rt->rt_flags |= RTF_MULTICAST;
116 
117 	/*
118 	 * A little bit of help for both IPv6 output and input:
119 	 *   For local addresses, we make sure that RTF_LOCAL is set,
120 	 *   with the thought that this might one day be used to speed up
121 	 *   ip_input().
122 	 *
123 	 * We also mark routes to multicast addresses as such, because
124 	 * it's easy to do and might be useful (but this is much more
125 	 * dubious since it's so easy to inspect the address).  (This
126 	 * is done above.)
127 	 *
128 	 * XXX
129 	 * should elaborate the code.
130 	 */
131 	if (rt->rt_flags & RTF_HOST) {
132 		if (IN6_ARE_ADDR_EQUAL(&satosin6(rt->rt_ifa->ifa_addr)
133 					->sin6_addr,
134 				       &sin6->sin6_addr)) {
135 			rt->rt_flags |= RTF_LOCAL;
136 		}
137 	}
138 
139 	if (!rt->rt_rmx.rmx_mtu && rt->rt_ifp)
140 		rt->rt_rmx.rmx_mtu = IN6_LINKMTU(rt->rt_ifp);
141 
142 	ret = rn_addroute(v_arg, n_arg, head, treenodes);
143 	if (ret == NULL) {
144 		struct rtentry *rt2;
145 		/*
146 		 * We are trying to add a net route, but can't.
147 		 * The following case should be allowed, so we'll make a
148 		 * special check for this:
149 		 *	Two IPv6 addresses with the same prefix is assigned
150 		 *	to a single interrface.
151 		 *	# ifconfig if0 inet6 3ffe:0501::1 prefix 64 alias (*1)
152 		 *	# ifconfig if0 inet6 3ffe:0501::2 prefix 64 alias (*2)
153 		 *	In this case, (*1) and (*2) want to add the same
154 		 *	net route entry, 3ffe:0501:: -> if0.
155 		 *	This case should not raise an error.
156 		 */
157 		rt2 = in6_rtalloc1((struct sockaddr *)sin6, 0, RTF_RNH_LOCKED,
158 		    rt->rt_fibnum);
159 		if (rt2) {
160 			if (((rt2->rt_flags & (RTF_HOST|RTF_GATEWAY)) == 0)
161 			 && rt2->rt_gateway
162 			 && rt2->rt_gateway->sa_family == AF_LINK
163 			 && rt2->rt_ifp == rt->rt_ifp) {
164 				ret = rt2->rt_nodes;
165 			}
166 			RTFREE_LOCKED(rt2);
167 		}
168 	}
169 	return (ret);
170 }
171 
172 SYSCTL_DECL(_net_inet6_ip6);
173 
174 static VNET_DEFINE(int, rtq_toomany6) = 128;
175 	/* 128 cached routes is ``too many'' */
176 #define	V_rtq_toomany6			VNET(rtq_toomany6)
177 SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_RTMAXCACHE, rtmaxcache, CTLFLAG_RW,
178     &VNET_NAME(rtq_toomany6) , 0, "");
179 
180 struct rtqk_arg {
181 	struct radix_node_head *rnh;
182 	int mode;
183 	int updating;
184 	int draining;
185 	int killed;
186 	int found;
187 	time_t nextstop;
188 };
189 
190 /*
191  * Age old PMTUs.
192  */
193 struct mtuex_arg {
194 	struct radix_node_head *rnh;
195 	time_t nextstop;
196 };
197 static VNET_DEFINE(struct callout, rtq_mtutimer);
198 #define	V_rtq_mtutimer			VNET(rtq_mtutimer)
199 
200 static int
201 in6_mtuexpire(struct radix_node *rn, void *rock)
202 {
203 	struct rtentry *rt = (struct rtentry *)rn;
204 	struct mtuex_arg *ap = rock;
205 
206 	/* sanity */
207 	if (!rt)
208 		panic("rt == NULL in in6_mtuexpire");
209 
210 	if (rt->rt_rmx.rmx_expire && !(rt->rt_flags & RTF_PROBEMTU)) {
211 		if (rt->rt_rmx.rmx_expire <= time_uptime) {
212 			rt->rt_flags |= RTF_PROBEMTU;
213 		} else {
214 			ap->nextstop = lmin(ap->nextstop,
215 					rt->rt_rmx.rmx_expire);
216 		}
217 	}
218 
219 	return 0;
220 }
221 
222 #define	MTUTIMO_DEFAULT	(60*1)
223 
224 static void
225 in6_mtutimo_one(struct radix_node_head *rnh)
226 {
227 	struct mtuex_arg arg;
228 
229 	arg.rnh = rnh;
230 	arg.nextstop = time_uptime + MTUTIMO_DEFAULT;
231 	RADIX_NODE_HEAD_LOCK(rnh);
232 	rnh->rnh_walktree(rnh, in6_mtuexpire, &arg);
233 	RADIX_NODE_HEAD_UNLOCK(rnh);
234 }
235 
236 static void
237 in6_mtutimo(void *rock)
238 {
239 	CURVNET_SET_QUIET((struct vnet *) rock);
240 	struct radix_node_head *rnh;
241 	struct timeval atv;
242 	u_int fibnum;
243 
244 	for (fibnum = 0; fibnum < rt_numfibs; fibnum++) {
245 		rnh = rt_tables_get_rnh(fibnum, AF_INET6);
246 		if (rnh != NULL)
247 			in6_mtutimo_one(rnh);
248 	}
249 
250 	atv.tv_sec = MTUTIMO_DEFAULT;
251 	atv.tv_usec = 0;
252 	callout_reset(&V_rtq_mtutimer, tvtohz(&atv), in6_mtutimo, rock);
253 	CURVNET_RESTORE();
254 }
255 
256 /*
257  * Initialize our routing tree.
258  * XXX MRT When off == 0, we are being called from vfs_export.c
259  * so just set up their table and leave. (we know what the correct
260  * value should be so just use that).. FIX AFTER RELENG_7 is MFC'd
261  * see also comments in in_inithead() vfs_export.c and domain.h
262  */
263 static VNET_DEFINE(int, _in6_rt_was_here);
264 #define	V__in6_rt_was_here	VNET(_in6_rt_was_here)
265 
266 int
267 in6_inithead(void **head, int off)
268 {
269 	struct radix_node_head *rnh;
270 
271 	if (!rn_inithead(head, offsetof(struct sockaddr_in6, sin6_addr) << 3))
272 		return 0;		/* See above */
273 
274 	if (off == 0)		/* See above */
275 		return 1;	/* only do the rest for the real thing */
276 
277 	rnh = *head;
278 	rnh->rnh_addaddr = in6_addroute;
279 
280 	if (V__in6_rt_was_here == 0) {
281 		callout_init(&V_rtq_mtutimer, CALLOUT_MPSAFE);
282 		in6_mtutimo(curvnet);	/* kick off timeout first time */
283 		V__in6_rt_was_here = 1;
284 	}
285 
286 	return 1;
287 }
288 
289 #ifdef VIMAGE
290 int
291 in6_detachhead(void **head, int off)
292 {
293 
294 	callout_drain(&V_rtq_mtutimer);
295 	return (1);
296 }
297 #endif
298 
299 /*
300  * Extended API for IPv6 FIB support.
301  */
302 void
303 in6_rtredirect(struct sockaddr *dst, struct sockaddr *gw, struct sockaddr *nm,
304     int flags, struct sockaddr *src, u_int fibnum)
305 {
306 
307 	rtredirect_fib(dst, gw, nm, flags, src, fibnum);
308 }
309 
310 int
311 in6_rtrequest(int req, struct sockaddr *dst, struct sockaddr *gw,
312     struct sockaddr *mask, int flags, struct rtentry **ret_nrt, u_int fibnum)
313 {
314 
315 	return (rtrequest_fib(req, dst, gw, mask, flags, ret_nrt, fibnum));
316 }
317 
318 void
319 in6_rtalloc(struct route_in6 *ro, u_int fibnum)
320 {
321 
322 	rtalloc_ign_fib((struct route *)ro, 0ul, fibnum);
323 }
324 
325 void
326 in6_rtalloc_ign(struct route_in6 *ro, u_long ignflags, u_int fibnum)
327 {
328 
329 	rtalloc_ign_fib((struct route *)ro, ignflags, fibnum);
330 }
331 
332 struct rtentry *
333 in6_rtalloc1(struct sockaddr *dst, int report, u_long ignflags, u_int fibnum)
334 {
335 
336 	return (rtalloc1_fib(dst, report, ignflags, fibnum));
337 }
338