xref: /freebsd/sys/netinet6/in6_fib.c (revision 652a9748855320619e075c4e83aef2f5294412d2)
1 /*-
2  * Copyright (c) 2015
3  * 	Alexander V. Chernikov <melifaro@FreeBSD.org>
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 University 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 REGENTS 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 REGENTS 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 
30 #include <sys/cdefs.h>
31 __FBSDID("$FreeBSD$");
32 
33 #include "opt_inet.h"
34 #include "opt_inet6.h"
35 #include "opt_route.h"
36 #include "opt_mpath.h"
37 
38 #include <sys/param.h>
39 #include <sys/systm.h>
40 #include <sys/lock.h>
41 #include <sys/rmlock.h>
42 #include <sys/malloc.h>
43 #include <sys/mbuf.h>
44 #include <sys/socket.h>
45 #include <sys/sysctl.h>
46 #include <sys/kernel.h>
47 
48 #include <net/if.h>
49 #include <net/if_var.h>
50 #include <net/if_dl.h>
51 #include <net/route.h>
52 #include <net/route_var.h>
53 #include <net/vnet.h>
54 
55 #ifdef RADIX_MPATH
56 #include <net/radix_mpath.h>
57 #endif
58 
59 #include <netinet/in.h>
60 #include <netinet/in_var.h>
61 #include <netinet/ip_mroute.h>
62 #include <netinet/ip6.h>
63 #include <netinet6/in6_fib.h>
64 #include <netinet6/in6_var.h>
65 #include <netinet6/nd6.h>
66 #include <netinet6/scope6_var.h>
67 
68 #include <net/if_types.h>
69 
70 #ifdef INET6
71 static void fib6_rte_to_nh_extended(struct rtentry *rte,
72     const struct in6_addr *dst, uint32_t flags, struct nhop6_extended *pnh6);
73 static void fib6_rte_to_nh_basic(struct rtentry *rte, const struct in6_addr *dst,
74     uint32_t flags, struct nhop6_basic *pnh6);
75 static struct ifnet *fib6_get_ifaifp(struct rtentry *rte);
76 #define RNTORT(p)	((struct rtentry *)(p))
77 
78 #define	ifatoia6(ifa)	((struct in6_ifaddr *)(ifa))
79 
80 CHK_STRUCT_ROUTE_COMPAT(struct route_in6, ro_dst);
81 
82 /*
83  * Gets real interface for the @rte.
84  * Returns rt_ifp for !IFF_LOOPBACK routers.
85  * Extracts "real" address interface from interface address
86  * loopback routes.
87  */
88 static struct ifnet *
89 fib6_get_ifaifp(struct rtentry *rte)
90 {
91 	struct ifnet *ifp;
92 	struct sockaddr_dl *sdl;
93 
94 	ifp = rte->rt_ifp;
95 	if ((ifp->if_flags & IFF_LOOPBACK) &&
96 	    rte->rt_gateway->sa_family == AF_LINK) {
97 		sdl = (struct sockaddr_dl *)rte->rt_gateway;
98 		return (ifnet_byindex(sdl->sdl_index));
99 	}
100 
101 	return (ifp);
102 }
103 
104 static void
105 fib6_rte_to_nh_basic(struct rtentry *rte, const struct in6_addr *dst,
106     uint32_t flags, struct nhop6_basic *pnh6)
107 {
108 	struct sockaddr_in6 *gw;
109 
110 	/* Do explicit nexthop zero unless we're copying it */
111 	memset(pnh6, 0, sizeof(*pnh6));
112 
113 	if ((flags & NHR_IFAIF) != 0)
114 		pnh6->nh_ifp = fib6_get_ifaifp(rte);
115 	else
116 		pnh6->nh_ifp = rte->rt_ifp;
117 
118 	pnh6->nh_mtu = min(rte->rt_mtu, IN6_LINKMTU(rte->rt_ifp));
119 	if (rte->rt_flags & RTF_GATEWAY) {
120 		/* Return address with embedded scope. */
121 		gw = (struct sockaddr_in6 *)rte->rt_gateway;
122 		pnh6->nh_addr = gw->sin6_addr;
123 	} else
124 		pnh6->nh_addr = *dst;
125 	/* Set flags */
126 	pnh6->nh_flags = fib_rte_to_nh_flags(rte->rt_flags);
127 	gw = (struct sockaddr_in6 *)rt_key(rte);
128 	if (IN6_IS_ADDR_UNSPECIFIED(&gw->sin6_addr))
129 		pnh6->nh_flags |= NHF_DEFAULT;
130 }
131 
132 static void
133 fib6_rte_to_nh_extended(struct rtentry *rte, const struct in6_addr *dst,
134     uint32_t flags, struct nhop6_extended *pnh6)
135 {
136 	struct sockaddr_in6 *gw;
137 
138 	/* Do explicit nexthop zero unless we're copying it */
139 	memset(pnh6, 0, sizeof(*pnh6));
140 
141 	if ((flags & NHR_IFAIF) != 0)
142 		pnh6->nh_ifp = fib6_get_ifaifp(rte);
143 	else
144 		pnh6->nh_ifp = rte->rt_ifp;
145 
146 	pnh6->nh_mtu = min(rte->rt_mtu, IN6_LINKMTU(rte->rt_ifp));
147 	if (rte->rt_flags & RTF_GATEWAY) {
148 		/* Return address with embedded scope. */
149 		gw = (struct sockaddr_in6 *)rte->rt_gateway;
150 		pnh6->nh_addr = gw->sin6_addr;
151 	} else
152 		pnh6->nh_addr = *dst;
153 	/* Set flags */
154 	pnh6->nh_flags = fib_rte_to_nh_flags(rte->rt_flags);
155 	gw = (struct sockaddr_in6 *)rt_key(rte);
156 	if (IN6_IS_ADDR_UNSPECIFIED(&gw->sin6_addr))
157 		pnh6->nh_flags |= NHF_DEFAULT;
158 	pnh6->nh_ia = ifatoia6(rte->rt_ifa);
159 }
160 
161 /*
162  * Performs IPv6 route table lookup on @dst. Returns 0 on success.
163  * Stores basic nexthop info into provided @pnh6 structure.
164  * Note that
165  * - nh_ifp represents logical transmit interface (rt_ifp) by default
166  * - nh_ifp represents "address" interface if NHR_IFAIF flag is passed
167  * - mtu from logical transmit interface will be returned.
168  * - nh_ifp cannot be safely dereferenced
169  * - nh_ifp represents rt_ifp (e.g. if looking up address on
170  *   interface "ix0" pointer to "ix0" interface will be returned instead
171  *   of "lo0")
172  * - howewer mtu from "transmit" interface will be returned.
173  * - scope will be embedded in nh_addr
174  */
175 int
176 fib6_lookup_nh_basic(uint32_t fibnum, const struct in6_addr *dst, uint32_t scopeid,
177     uint32_t flags, uint32_t flowid, struct nhop6_basic *pnh6)
178 {
179 	RIB_RLOCK_TRACKER;
180 	struct rib_head *rh;
181 	struct radix_node *rn;
182 	struct sockaddr_in6 sin6;
183 	struct rtentry *rte;
184 
185 	KASSERT((fibnum < rt_numfibs), ("fib6_lookup_nh_basic: bad fibnum"));
186 	rh = rt_tables_get_rnh(fibnum, AF_INET6);
187 	if (rh == NULL)
188 		return (ENOENT);
189 
190 	/* Prepare lookup key */
191 	memset(&sin6, 0, sizeof(sin6));
192 	sin6.sin6_addr = *dst;
193 	sin6.sin6_len = sizeof(struct sockaddr_in6);
194 	/* Assume scopeid is valid and embed it directly */
195 	if (IN6_IS_SCOPE_LINKLOCAL(dst))
196 		sin6.sin6_addr.s6_addr16[1] = htons(scopeid & 0xffff);
197 
198 	RIB_RLOCK(rh);
199 	rn = rh->rnh_matchaddr((void *)&sin6, &rh->head);
200 	if (rn != NULL && ((rn->rn_flags & RNF_ROOT) == 0)) {
201 		rte = RNTORT(rn);
202 		/* Ensure route & ifp is UP */
203 		if (RT_LINK_IS_UP(rte->rt_ifp)) {
204 			fib6_rte_to_nh_basic(rte, &sin6.sin6_addr, flags, pnh6);
205 			RIB_RUNLOCK(rh);
206 			return (0);
207 		}
208 	}
209 	RIB_RUNLOCK(rh);
210 
211 	return (ENOENT);
212 }
213 
214 /*
215  * Performs IPv6 route table lookup on @dst. Returns 0 on success.
216  * Stores extended nexthop info into provided @pnh6 structure.
217  * Note that
218  * - nh_ifp cannot be safely dereferenced unless NHR_REF is specified.
219  * - in that case you need to call fib6_free_nh_ext()
220  * - nh_ifp represents logical transmit interface (rt_ifp) by default
221  * - nh_ifp represents "address" interface if NHR_IFAIF flag is passed
222  * - mtu from logical transmit interface will be returned.
223  * - scope will be embedded in nh_addr
224  */
225 int
226 fib6_lookup_nh_ext(uint32_t fibnum, const struct in6_addr *dst,uint32_t scopeid,
227     uint32_t flags, uint32_t flowid, struct nhop6_extended *pnh6)
228 {
229 	RIB_RLOCK_TRACKER;
230 	struct rib_head *rh;
231 	struct radix_node *rn;
232 	struct sockaddr_in6 sin6;
233 	struct rtentry *rte;
234 
235 	KASSERT((fibnum < rt_numfibs), ("fib6_lookup_nh_ext: bad fibnum"));
236 	rh = rt_tables_get_rnh(fibnum, AF_INET6);
237 	if (rh == NULL)
238 		return (ENOENT);
239 
240 	/* Prepare lookup key */
241 	memset(&sin6, 0, sizeof(sin6));
242 	sin6.sin6_len = sizeof(struct sockaddr_in6);
243 	sin6.sin6_addr = *dst;
244 	/* Assume scopeid is valid and embed it directly */
245 	if (IN6_IS_SCOPE_LINKLOCAL(dst))
246 		sin6.sin6_addr.s6_addr16[1] = htons(scopeid & 0xffff);
247 
248 	RIB_RLOCK(rh);
249 	rn = rh->rnh_matchaddr((void *)&sin6, &rh->head);
250 	if (rn != NULL && ((rn->rn_flags & RNF_ROOT) == 0)) {
251 		rte = RNTORT(rn);
252 #ifdef RADIX_MPATH
253 		rte = rt_mpath_select(rte, flowid);
254 		if (rte == NULL) {
255 			RIB_RUNLOCK(rh);
256 			return (ENOENT);
257 		}
258 #endif
259 		/* Ensure route & ifp is UP */
260 		if (RT_LINK_IS_UP(rte->rt_ifp)) {
261 			fib6_rte_to_nh_extended(rte, &sin6.sin6_addr, flags,
262 			    pnh6);
263 			if ((flags & NHR_REF) != 0) {
264 				/* TODO: Do lwref on egress ifp's */
265 			}
266 			RIB_RUNLOCK(rh);
267 
268 			return (0);
269 		}
270 	}
271 	RIB_RUNLOCK(rh);
272 
273 	return (ENOENT);
274 }
275 
276 void
277 fib6_free_nh_ext(uint32_t fibnum, struct nhop6_extended *pnh6)
278 {
279 
280 }
281 
282 #endif
283 
284