xref: /freebsd/sys/netinet6/in6_rmx.c (revision 145992504973bd16cf3518af9ba5ce185fefa82a)
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/route.h>
80 
81 #include <netinet/in.h>
82 #include <netinet/ip_var.h>
83 #include <netinet/in_var.h>
84 
85 #include <netinet/ip6.h>
86 #include <netinet6/ip6_var.h>
87 
88 #include <netinet/icmp6.h>
89 #include <netinet6/nd6.h>
90 
91 #include <netinet/tcp.h>
92 #include <netinet/tcp_seq.h>
93 #include <netinet/tcp_timer.h>
94 #include <netinet/tcp_var.h>
95 
96 extern int	in6_inithead(void **head, int off);
97 #ifdef VIMAGE
98 extern int	in6_detachhead(void **head, int off);
99 #endif
100 
101 /*
102  * Do what we need to do when inserting a route.
103  */
104 static struct radix_node *
105 in6_addroute(void *v_arg, void *n_arg, struct radix_node_head *head,
106     struct radix_node *treenodes)
107 {
108 	struct rtentry *rt = (struct rtentry *)treenodes;
109 	struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)rt_key(rt);
110 	struct radix_node *ret;
111 
112 	RADIX_NODE_HEAD_WLOCK_ASSERT(head);
113 	if (IN6_IS_ADDR_MULTICAST(&sin6->sin6_addr))
114 		rt->rt_flags |= RTF_MULTICAST;
115 
116 	/*
117 	 * A little bit of help for both IPv6 output and input:
118 	 *   For local addresses, we make sure that RTF_LOCAL is set,
119 	 *   with the thought that this might one day be used to speed up
120 	 *   ip_input().
121 	 *
122 	 * We also mark routes to multicast addresses as such, because
123 	 * it's easy to do and might be useful (but this is much more
124 	 * dubious since it's so easy to inspect the address).  (This
125 	 * is done above.)
126 	 *
127 	 * XXX
128 	 * should elaborate the code.
129 	 */
130 	if (rt->rt_flags & RTF_HOST) {
131 		if (IN6_ARE_ADDR_EQUAL(&satosin6(rt->rt_ifa->ifa_addr)
132 					->sin6_addr,
133 				       &sin6->sin6_addr)) {
134 			rt->rt_flags |= RTF_LOCAL;
135 		}
136 	}
137 
138 	if (!rt->rt_rmx.rmx_mtu && rt->rt_ifp)
139 		rt->rt_rmx.rmx_mtu = IN6_LINKMTU(rt->rt_ifp);
140 
141 	ret = rn_addroute(v_arg, n_arg, head, treenodes);
142 	if (ret == NULL) {
143 		struct rtentry *rt2;
144 		/*
145 		 * We are trying to add a net route, but can't.
146 		 * The following case should be allowed, so we'll make a
147 		 * special check for this:
148 		 *	Two IPv6 addresses with the same prefix is assigned
149 		 *	to a single interrface.
150 		 *	# ifconfig if0 inet6 3ffe:0501::1 prefix 64 alias (*1)
151 		 *	# ifconfig if0 inet6 3ffe:0501::2 prefix 64 alias (*2)
152 		 *	In this case, (*1) and (*2) want to add the same
153 		 *	net route entry, 3ffe:0501:: -> if0.
154 		 *	This case should not raise an error.
155 		 */
156 		rt2 = in6_rtalloc1((struct sockaddr *)sin6, 0, RTF_RNH_LOCKED,
157 		    rt->rt_fibnum);
158 		if (rt2) {
159 			if (((rt2->rt_flags & (RTF_HOST|RTF_GATEWAY)) == 0)
160 			 && rt2->rt_gateway
161 			 && rt2->rt_gateway->sa_family == AF_LINK
162 			 && rt2->rt_ifp == rt->rt_ifp) {
163 				ret = rt2->rt_nodes;
164 			}
165 			RTFREE_LOCKED(rt2);
166 		}
167 	}
168 	return (ret);
169 }
170 
171 SYSCTL_DECL(_net_inet6_ip6);
172 
173 static VNET_DEFINE(int, rtq_toomany6) = 128;
174 	/* 128 cached routes is ``too many'' */
175 #define	V_rtq_toomany6			VNET(rtq_toomany6)
176 SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_RTMAXCACHE, rtmaxcache, CTLFLAG_RW,
177     &VNET_NAME(rtq_toomany6) , 0, "");
178 
179 struct rtqk_arg {
180 	struct radix_node_head *rnh;
181 	int mode;
182 	int updating;
183 	int draining;
184 	int killed;
185 	int found;
186 	time_t nextstop;
187 };
188 
189 /*
190  * Age old PMTUs.
191  */
192 struct mtuex_arg {
193 	struct radix_node_head *rnh;
194 	time_t nextstop;
195 };
196 static VNET_DEFINE(struct callout, rtq_mtutimer);
197 #define	V_rtq_mtutimer			VNET(rtq_mtutimer)
198 
199 static int
200 in6_mtuexpire(struct radix_node *rn, void *rock)
201 {
202 	struct rtentry *rt = (struct rtentry *)rn;
203 	struct mtuex_arg *ap = rock;
204 
205 	/* sanity */
206 	if (!rt)
207 		panic("rt == NULL in in6_mtuexpire");
208 
209 	if (rt->rt_rmx.rmx_expire && !(rt->rt_flags & RTF_PROBEMTU)) {
210 		if (rt->rt_rmx.rmx_expire <= time_uptime) {
211 			rt->rt_flags |= RTF_PROBEMTU;
212 		} else {
213 			ap->nextstop = lmin(ap->nextstop,
214 					rt->rt_rmx.rmx_expire);
215 		}
216 	}
217 
218 	return 0;
219 }
220 
221 #define	MTUTIMO_DEFAULT	(60*1)
222 
223 static void
224 in6_mtutimo_one(struct radix_node_head *rnh)
225 {
226 	struct mtuex_arg arg;
227 
228 	arg.rnh = rnh;
229 	arg.nextstop = time_uptime + MTUTIMO_DEFAULT;
230 	RADIX_NODE_HEAD_LOCK(rnh);
231 	rnh->rnh_walktree(rnh, in6_mtuexpire, &arg);
232 	RADIX_NODE_HEAD_UNLOCK(rnh);
233 }
234 
235 static void
236 in6_mtutimo(void *rock)
237 {
238 	CURVNET_SET_QUIET((struct vnet *) rock);
239 	struct radix_node_head *rnh;
240 	struct timeval atv;
241 	u_int fibnum;
242 
243 	for (fibnum = 0; fibnum < rt_numfibs; fibnum++) {
244 		rnh = rt_tables_get_rnh(fibnum, AF_INET6);
245 		if (rnh != NULL)
246 			in6_mtutimo_one(rnh);
247 	}
248 
249 	atv.tv_sec = MTUTIMO_DEFAULT;
250 	atv.tv_usec = 0;
251 	callout_reset(&V_rtq_mtutimer, tvtohz(&atv), in6_mtutimo, rock);
252 	CURVNET_RESTORE();
253 }
254 
255 /*
256  * Initialize our routing tree.
257  * XXX MRT When off == 0, we are being called from vfs_export.c
258  * so just set up their table and leave. (we know what the correct
259  * value should be so just use that).. FIX AFTER RELENG_7 is MFC'd
260  * see also comments in in_inithead() vfs_export.c and domain.h
261  */
262 static VNET_DEFINE(int, _in6_rt_was_here);
263 #define	V__in6_rt_was_here	VNET(_in6_rt_was_here)
264 
265 int
266 in6_inithead(void **head, int off)
267 {
268 	struct radix_node_head *rnh;
269 
270 	if (!rn_inithead(head, offsetof(struct sockaddr_in6, sin6_addr) << 3))
271 		return 0;		/* See above */
272 
273 	if (off == 0)		/* See above */
274 		return 1;	/* only do the rest for the real thing */
275 
276 	rnh = *head;
277 	rnh->rnh_addaddr = in6_addroute;
278 
279 	if (V__in6_rt_was_here == 0) {
280 		callout_init(&V_rtq_mtutimer, CALLOUT_MPSAFE);
281 		in6_mtutimo(curvnet);	/* kick off timeout first time */
282 		V__in6_rt_was_here = 1;
283 	}
284 
285 	return 1;
286 }
287 
288 #ifdef VIMAGE
289 int
290 in6_detachhead(void **head, int off)
291 {
292 
293 	callout_drain(&V_rtq_mtutimer);
294 	return (1);
295 }
296 #endif
297 
298 /*
299  * Extended API for IPv6 FIB support.
300  */
301 void
302 in6_rtredirect(struct sockaddr *dst, struct sockaddr *gw, struct sockaddr *nm,
303     int flags, struct sockaddr *src, u_int fibnum)
304 {
305 
306 	rtredirect_fib(dst, gw, nm, flags, src, fibnum);
307 }
308 
309 int
310 in6_rtrequest(int req, struct sockaddr *dst, struct sockaddr *gw,
311     struct sockaddr *mask, int flags, struct rtentry **ret_nrt, u_int fibnum)
312 {
313 
314 	return (rtrequest_fib(req, dst, gw, mask, flags, ret_nrt, fibnum));
315 }
316 
317 void
318 in6_rtalloc(struct route_in6 *ro, u_int fibnum)
319 {
320 
321 	rtalloc_ign_fib((struct route *)ro, 0ul, fibnum);
322 }
323 
324 void
325 in6_rtalloc_ign(struct route_in6 *ro, u_long ignflags, u_int fibnum)
326 {
327 
328 	rtalloc_ign_fib((struct route *)ro, ignflags, fibnum);
329 }
330 
331 struct rtentry *
332 in6_rtalloc1(struct sockaddr *dst, int report, u_long ignflags, u_int fibnum)
333 {
334 
335 	return (rtalloc1_fib(dst, report, ignflags, fibnum));
336 }
337