xref: /freebsd/sys/netinet/in_fib.c (revision 78adacd4eab39a3508bd8c65f0aba94fc6b907ce)
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_route.h"
35 
36 #include <sys/param.h>
37 #include <sys/systm.h>
38 #include <sys/lock.h>
39 #include <sys/rmlock.h>
40 #include <sys/malloc.h>
41 #include <sys/mbuf.h>
42 #include <sys/socket.h>
43 #include <sys/sysctl.h>
44 #include <sys/kernel.h>
45 
46 #include <net/if.h>
47 #include <net/if_var.h>
48 #include <net/if_dl.h>
49 #include <net/route.h>
50 #include <net/route/route_ctl.h>
51 #include <net/route/route_var.h>
52 #include <net/route/nhop.h>
53 #include <net/toeplitz.h>
54 #include <net/vnet.h>
55 
56 #include <netinet/in.h>
57 #include <netinet/in_var.h>
58 #include <netinet/in_fib.h>
59 
60 #ifdef INET
61 
62 /* Verify struct route compatiblity */
63 /* Assert 'struct route_in' is compatible with 'struct route' */
64 CHK_STRUCT_ROUTE_COMPAT(struct route_in, ro_dst4);
65 
66 #ifdef ROUTE_MPATH
67 struct _hash_5tuple_ipv4 {
68 	struct in_addr src;
69 	struct in_addr dst;
70 	unsigned short src_port;
71 	unsigned short dst_port;
72 	char proto;
73 	char spare[3];
74 };
75 _Static_assert(sizeof(struct _hash_5tuple_ipv4) == 16,
76     "_hash_5tuple_ipv4 size is wrong");
77 
78 uint32_t
79 fib4_calc_software_hash(struct in_addr src, struct in_addr dst,
80     unsigned short src_port, unsigned short dst_port, char proto,
81     uint32_t *phashtype)
82 {
83 	struct _hash_5tuple_ipv4 data;
84 
85 	data.src = src;
86 	data.dst = dst;
87 	data.src_port = src_port;
88 	data.dst_port = dst_port;
89 	data.proto = proto;
90 	data.spare[0] = data.spare[1] = data.spare[2] = 0;
91 
92 	*phashtype = M_HASHTYPE_OPAQUE;
93 
94 	return (toeplitz_hash(MPATH_ENTROPY_KEY_LEN, mpath_entropy_key,
95 	  sizeof(data), (uint8_t *)&data));
96 }
97 #endif
98 
99 /*
100  * Looks up path in fib @fibnum specified by @dst.
101  * Returns path nexthop on success. Nexthop is safe to use
102  *  within the current network epoch. If longer lifetime is required,
103  *  one needs to pass NHR_REF as a flag. This will return referenced
104  *  nexthop.
105  */
106 struct nhop_object *
107 fib4_lookup(uint32_t fibnum, struct in_addr dst, uint32_t scopeid,
108     uint32_t flags, uint32_t flowid)
109 {
110 	RIB_RLOCK_TRACKER;
111 	struct rib_head *rh;
112 	struct radix_node *rn;
113 	struct nhop_object *nh;
114 
115 	KASSERT((fibnum < rt_numfibs), ("fib4_lookup: bad fibnum"));
116 	rh = rt_tables_get_rnh(fibnum, AF_INET);
117 	if (rh == NULL)
118 		return (NULL);
119 
120 	/* Prepare lookup key */
121 	struct sockaddr_in sin4 = {
122 		.sin_family = AF_INET,
123 		.sin_len = sizeof(struct sockaddr_in),
124 		.sin_addr = dst,
125 	};
126 
127 	nh = NULL;
128 	RIB_RLOCK(rh);
129 	rn = rh->rnh_matchaddr((void *)&sin4, &rh->head);
130 	if (rn != NULL && ((rn->rn_flags & RNF_ROOT) == 0)) {
131 		nh = nhop_select((RNTORT(rn))->rt_nhop, flowid);
132 		/* Ensure route & ifp is UP */
133 		if (RT_LINK_IS_UP(nh->nh_ifp)) {
134 			if (flags & NHR_REF)
135 				nhop_ref_object(nh);
136 			RIB_RUNLOCK(rh);
137 			return (nh);
138 		}
139 	}
140 	RIB_RUNLOCK(rh);
141 
142 	RTSTAT_INC(rts_unreach);
143 	return (NULL);
144 }
145 
146 inline static int
147 check_urpf_nhop(const struct nhop_object *nh, uint32_t flags,
148     const struct ifnet *src_if)
149 {
150 
151 	if (src_if != NULL && nh->nh_aifp == src_if) {
152 		return (1);
153 	}
154 	if (src_if == NULL) {
155 		if ((flags & NHR_NODEFAULT) == 0)
156 			return (1);
157 		else if ((nh->nh_flags & NHF_DEFAULT) == 0)
158 			return (1);
159 	}
160 
161 	return (0);
162 }
163 
164 static int
165 check_urpf(struct nhop_object *nh, uint32_t flags,
166     const struct ifnet *src_if)
167 {
168 #ifdef ROUTE_MPATH
169 	if (NH_IS_NHGRP(nh)) {
170 		struct weightened_nhop *wn;
171 		uint32_t num_nhops;
172 		wn = nhgrp_get_nhops((struct nhgrp_object *)nh, &num_nhops);
173 			for (int i = 0; i < num_nhops; i++) {
174 				if (check_urpf_nhop(wn[i].nh, flags, src_if) != 0)
175 				return (1);
176 		}
177 		return (0);
178 	} else
179 #endif
180 		return (check_urpf_nhop(nh, flags, src_if));
181 }
182 
183 static struct nhop_object *
184 lookup_nhop(uint32_t fibnum, struct in_addr dst, uint32_t scopeid)
185 {
186 	RIB_RLOCK_TRACKER;
187 	struct rib_head *rh;
188 	struct radix_node *rn;
189 	struct nhop_object *nh;
190 
191 	KASSERT((fibnum < rt_numfibs), ("fib4_check_urpf: bad fibnum"));
192 	rh = rt_tables_get_rnh(fibnum, AF_INET);
193 	if (rh == NULL)
194 		return (NULL);
195 
196 	/* Prepare lookup key */
197 	struct sockaddr_in sin4;
198 	memset(&sin4, 0, sizeof(sin4));
199 	sin4.sin_len = sizeof(struct sockaddr_in);
200 	sin4.sin_addr = dst;
201 
202 	nh = NULL;
203 	RIB_RLOCK(rh);
204 	rn = rh->rnh_matchaddr((void *)&sin4, &rh->head);
205 	if (rn != NULL && ((rn->rn_flags & RNF_ROOT) == 0))
206 		nh = RNTORT(rn)->rt_nhop;
207 	RIB_RUNLOCK(rh);
208 
209 	return (nh);
210 }
211 
212 /*
213  * Performs reverse path forwarding lookup.
214  * If @src_if is non-zero, verifies that at least 1 path goes via
215  *   this interface.
216  * If @src_if is zero, verifies that route exist.
217  * if @flags contains NHR_NOTDEFAULT, do not consider default route.
218  *
219  * Returns 1 if route matching conditions is found, 0 otherwise.
220  */
221 int
222 fib4_check_urpf(uint32_t fibnum, struct in_addr dst, uint32_t scopeid,
223   uint32_t flags, const struct ifnet *src_if)
224 {
225 	struct nhop_object *nh;
226 
227 	nh = lookup_nhop(fibnum, dst, scopeid);
228 	if (nh != NULL)
229 		return (check_urpf(nh, flags, src_if));
230 
231 	return (0);
232 }
233 
234 /*
235  * Function returning prefix match data along with the nexthop data.
236  * Intended to be used by the control plane code.
237  * Supported flags:
238  *  NHR_UNLOCKED: do not lock radix during lookup.
239  * Returns pointer to rtentry and raw nexthop in @rnd. Both rtentry
240  *  and nexthop are safe to use within current epoch. Note:
241  * Note: rnd_nhop can actually be the nexthop group.
242  */
243 struct rtentry *
244 fib4_lookup_rt(uint32_t fibnum, struct in_addr dst, uint32_t scopeid,
245     uint32_t flags, struct route_nhop_data *rnd)
246 {
247 	RIB_RLOCK_TRACKER;
248 	struct rib_head *rh;
249 	struct radix_node *rn;
250 	struct rtentry *rt;
251 
252 	KASSERT((fibnum < rt_numfibs), ("fib4_lookup_rt: bad fibnum"));
253 	rh = rt_tables_get_rnh(fibnum, AF_INET);
254 	if (rh == NULL)
255 		return (NULL);
256 
257 	/* Prepare lookup key */
258 	struct sockaddr_in sin4 = {
259 		.sin_family = AF_INET,
260 		.sin_len = sizeof(struct sockaddr_in),
261 		.sin_addr = dst,
262 	};
263 
264 	rt = NULL;
265 	if (!(flags & NHR_UNLOCKED))
266 		RIB_RLOCK(rh);
267 	rn = rh->rnh_matchaddr((void *)&sin4, &rh->head);
268 	if (rn != NULL && ((rn->rn_flags & RNF_ROOT) == 0)) {
269 		rt = (struct rtentry *)rn;
270 		rnd->rnd_nhop = rt->rt_nhop;
271 		rnd->rnd_weight = rt->rt_weight;
272 	}
273 	if (!(flags & NHR_UNLOCKED))
274 		RIB_RUNLOCK(rh);
275 
276 	return (rt);
277 }
278 
279 struct nhop_object *
280 fib4_lookup_debugnet(uint32_t fibnum, struct in_addr dst, uint32_t scopeid,
281     uint32_t flags)
282 {
283 	struct rtentry *rt;
284 	struct route_nhop_data rnd;
285 
286 	rt = fib4_lookup_rt(fibnum, dst, scopeid, NHR_UNLOCKED, &rnd);
287 	if (rt != NULL) {
288 		struct nhop_object *nh = nhop_select(rnd.rnd_nhop, 0);
289 		/* Ensure route & ifp is UP */
290 		if (RT_LINK_IS_UP(nh->nh_ifp))
291 			return (nh);
292 	}
293 
294 	return (NULL);
295 }
296 
297 #endif
298