1 /* $KAME: radix_mpath.c,v 1.17 2004/11/08 10:29:39 itojun Exp $ */
4 * SPDX-License-Identifier: BSD-3-Clause
6 * Copyright (C) 2001 WIDE Project.
9 * Redistribution and use in source and binary forms, with or without
10 * modification, are permitted provided that the following conditions
12 * 1. Redistributions of source code must retain the above copyright
13 * notice, this list of conditions and the following disclaimer.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 * notice, this list of conditions and the following disclaimer in the
16 * documentation and/or other materials provided with the distribution.
17 * 3. Neither the name of the project nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
21 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
32 * THE AUTHORS DO NOT GUARANTEE THAT THIS SOFTWARE DOES NOT INFRINGE
33 * ANY OTHERS' INTELLECTUAL PROPERTIES. IN NO EVENT SHALL THE AUTHORS
34 * BE LIABLE FOR ANY INFRINGEMENT OF ANY OTHERS' INTELLECTUAL
38 #include <sys/cdefs.h>
39 __FBSDID("$FreeBSD$");
42 #include "opt_inet6.h"
43 #include "opt_mpath.h"
45 #include <sys/param.h>
46 #include <sys/systm.h>
48 #include <sys/malloc.h>
49 #include <sys/mutex.h>
50 #include <sys/socket.h>
51 #include <sys/domain.h>
52 #include <sys/kernel.h>
53 #include <sys/syslog.h>
54 #include <net/radix.h>
55 #include <net/radix_mpath.h>
56 #include <sys/rmlock.h>
57 #include <net/route.h>
58 #include <net/route/nhop.h>
59 #include <net/route/shared.h>
60 #include <net/route/route_var.h>
61 #include <net/route/nhop.h>
63 #include <net/if_var.h>
66 * give some jitter to hash, to avoid synchronization between routers
68 static uint32_t hashjitter;
71 rt_mpath_capable(struct rib_head *rnh)
74 return rnh->rnh_multipath;
78 rn_mpath_capable(struct radix_head *rh)
81 return (rt_mpath_capable((struct rib_head *)rh));
85 rn_mpath_next(struct radix_node *rn)
87 struct radix_node *next;
91 next = rn->rn_dupedkey;
92 if (rn->rn_mask == next->rn_mask)
99 rn_mpath_count(struct radix_node *rn)
105 rt = (struct rtentry *)rn;
107 rn = rn_mpath_next(rn);
113 rt_mpath_matchgate(struct rtentry *rt, struct sockaddr *gate)
115 struct radix_node *rn;
116 struct nhop_object *nh;
121 /* beyond here, we use rn as the master copy */
122 rn = (struct radix_node *)rt;
124 rt = (struct rtentry *)rn;
127 * we are removing an address alias that has
128 * the same prefix as another address
129 * we need to compare the interface address because
130 * gateway is a special sockaddr_dl structure
132 if (nh->gw_sa.sa_family == AF_LINK) {
133 if (!memcmp(nh->nh_ifa->ifa_addr, gate, gate->sa_len))
138 * Check for other options:
139 * 1) Routes with 'real' IPv4/IPv6 gateway
140 * 2) Loopback host routes (another AF_LINK/sockadd_dl check)
142 if (nh->gw_sa.sa_len == gate->sa_len &&
143 !memcmp(&nh->gw_sa, gate, gate->sa_len))
145 } while ((rn = rn_mpath_next(rn)) != NULL);
147 return (struct rtentry *)rn;
151 * go through the chain and unlink "rt" from the list
152 * the caller will free "rt"
155 rt_mpath_deldup(struct rtentry *headrt, struct rtentry *rt)
157 struct radix_node *t, *tt;
161 t = (struct radix_node *)headrt;
162 tt = rn_mpath_next(t);
164 if (tt == (struct radix_node *)rt) {
165 t->rn_dupedkey = tt->rn_dupedkey;
166 tt->rn_dupedkey = NULL;
167 tt->rn_flags &= ~RNF_ACTIVE;
168 tt[1].rn_flags &= ~RNF_ACTIVE;
172 tt = rn_mpath_next((struct radix_node *)t);
178 * check if we have the same key/mask/gateway on the table already.
179 * Assume @rt rt_key host bits are cleared according to @netmask
182 rt_mpath_conflict(struct rib_head *rnh, struct rtentry *rt,
183 struct sockaddr *netmask)
185 struct radix_node *rn, *rn1;
186 struct nhop_object *nh, *nh1;
189 rn = (struct radix_node *)rt;
190 rn1 = rnh->rnh_lookup(rt_key(rt), netmask, &rnh->head);
191 if (!rn1 || rn1->rn_flags & RNF_ROOT)
194 /* key/mask are the same. compare gateway for all multipaths */
196 rt1 = (struct rtentry *)rn1;
198 /* sanity: no use in comparing the same thing */
205 if (nh1->gw_sa.sa_family == AF_LINK) {
206 if (nh1->nh_ifa->ifa_addr->sa_len != nh->nh_ifa->ifa_addr->sa_len ||
207 bcmp(nh1->nh_ifa->ifa_addr, nh->nh_ifa->ifa_addr,
208 nh1->nh_ifa->ifa_addr->sa_len))
211 if (nh1->gw_sa.sa_len != nh->gw_sa.sa_len ||
212 bcmp(&nh1->gw_sa, &nh->gw_sa, nh1->gw_sa.sa_len))
216 /* all key/mask/gateway are the same. conflicting entry. */
218 } while ((rn1 = rn_mpath_next(rn1)) != NULL);
224 rt_mpath_selectrte(struct rtentry *rte, uint32_t hash)
226 struct radix_node *rn0, *rn;
227 uint32_t total_weight;
231 /* beyond here, we use rn as the master copy */
232 rn0 = rn = (struct radix_node *)rte;
235 /* gw selection by Modulo-N Hash (RFC2991) XXX need improvement? */
236 total_weight = rn_mpath_count(rn0);
238 hash %= total_weight;
239 for (weight = abs((int32_t)hash);
240 rt != NULL && weight >= rt->rt_weight;
241 weight -= (rt == NULL) ? 0 : rt->rt_weight) {
243 /* stay within the multipath routes */
244 if (rn->rn_dupedkey && rn->rn_mask != rn->rn_dupedkey->rn_mask)
246 rn = rn->rn_dupedkey;
247 rt = (struct rtentry *)rn;
254 rt_mpath_select(struct rtentry *rte, uint32_t hash)
256 if (rn_mpath_next((struct radix_node *)rte) == NULL)
259 return (rt_mpath_selectrte(rte, hash));
263 rt_mpath_init_rnh(struct rib_head *rnh)
266 rnh->rnh_multipath = 1;
274 hashjitter = arc4random();
276 SYSINIT(mpath_init, SI_SUB_LAST, SI_ORDER_ANY, mpath_init, NULL);