2 * SPDX-License-Identifier: BSD-3-Clause
4 * Copyright (C) 1998 WIDE Project.
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 * 3. Neither the name of the project nor the names of its contributors
16 * may be used to endorse or promote products derived from this software
17 * without specific prior written permission.
19 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31 * $KAME: ip6_mroute.c,v 1.58 2001/12/18 02:36:31 itojun Exp $
35 * Copyright (c) 1989 Stephen Deering
36 * Copyright (c) 1992, 1993
37 * The Regents of the University of California. All rights reserved.
39 * This code is derived from software contributed to Berkeley by
40 * Stephen Deering of Stanford University.
42 * Redistribution and use in source and binary forms, with or without
43 * modification, are permitted provided that the following conditions
45 * 1. Redistributions of source code must retain the above copyright
46 * notice, this list of conditions and the following disclaimer.
47 * 2. Redistributions in binary form must reproduce the above copyright
48 * notice, this list of conditions and the following disclaimer in the
49 * documentation and/or other materials provided with the distribution.
50 * 3. Neither the name of the University nor the names of its contributors
51 * may be used to endorse or promote products derived from this software
52 * without specific prior written permission.
54 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
55 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
56 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
57 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
58 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
59 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
60 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
61 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
62 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
63 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
66 * @(#)ip_mroute.c 8.2 (Berkeley) 11/15/93
67 * BSDI ip_mroute.c,v 2.10 1996/11/14 00:29:52 jch Exp
71 * IP multicast forwarding procedures
73 * Written by David Waitzman, BBN Labs, August 1988.
74 * Modified by Steve Deering, Stanford, February 1989.
75 * Modified by Mark J. Steiglitz, Stanford, May, 1991
76 * Modified by Van Jacobson, LBL, January 1993
77 * Modified by Ajit Thyagarajan, PARC, August 1993
78 * Modified by Bill Fenner, PARC, April 1994
80 * MROUTING Revision: 3.5.1.2 + PIM-SMv2 (pimd) Support
83 #include <sys/cdefs.h>
84 __FBSDID("$FreeBSD$");
86 #include "opt_inet6.h"
88 #include <sys/param.h>
89 #include <sys/callout.h>
90 #include <sys/errno.h>
91 #include <sys/kernel.h>
93 #include <sys/malloc.h>
95 #include <sys/module.h>
96 #include <sys/domain.h>
97 #include <sys/protosw.h>
99 #include <sys/signalvar.h>
100 #include <sys/socket.h>
101 #include <sys/socketvar.h>
102 #include <sys/sockio.h>
104 #include <sys/sysctl.h>
105 #include <sys/syslog.h>
106 #include <sys/systm.h>
107 #include <sys/time.h>
110 #include <net/if_var.h>
111 #include <net/if_types.h>
112 #include <net/vnet.h>
114 #include <netinet/in.h>
115 #include <netinet/in_var.h>
116 #include <netinet/icmp6.h>
117 #include <netinet/ip_encap.h>
119 #include <netinet/ip6.h>
120 #include <netinet/in_kdtrace.h>
121 #include <netinet6/ip6_var.h>
122 #include <netinet6/scope6_var.h>
123 #include <netinet6/nd6.h>
124 #include <netinet6/ip6_mroute.h>
125 #include <netinet6/pim6.h>
126 #include <netinet6/pim6_var.h>
128 static MALLOC_DEFINE(M_MRTABLE6, "mf6c", "multicast forwarding cache entry");
130 static int ip6_mdq(struct mbuf *, struct ifnet *, struct mf6c *);
131 static void phyint_send(struct ip6_hdr *, struct mif6 *, struct mbuf *);
132 static int register_send(struct ip6_hdr *, struct mif6 *, struct mbuf *);
133 static int set_pim6(int *);
134 static int socket_send(struct socket *, struct mbuf *,
135 struct sockaddr_in6 *);
137 extern int in6_mcast_loop;
138 extern struct domain inet6domain;
140 static const struct encaptab *pim6_encap_cookie;
141 static int pim6_encapcheck(const struct mbuf *, int, int, void *);
142 static int pim6_input(struct mbuf *, int, int, void *);
144 static const struct encap_config ipv6_encap_cfg = {
145 .proto = IPPROTO_PIM,
146 .min_length = sizeof(struct ip6_hdr) + PIM_MINLEN,
148 .check = pim6_encapcheck,
152 VNET_DEFINE_STATIC(int, ip6_mrouter_ver) = 0;
153 #define V_ip6_mrouter_ver VNET(ip6_mrouter_ver)
155 SYSCTL_DECL(_net_inet6);
156 SYSCTL_DECL(_net_inet6_ip6);
157 static SYSCTL_NODE(_net_inet6, IPPROTO_PIM, pim,
158 CTLFLAG_RW | CTLFLAG_MPSAFE, 0,
161 static struct mrt6stat mrt6stat;
162 SYSCTL_STRUCT(_net_inet6_ip6, OID_AUTO, mrt6stat, CTLFLAG_RW,
164 "Multicast Routing Statistics (struct mrt6stat, netinet6/ip6_mroute.h)");
166 #define MRT6STAT_INC(name) mrt6stat.name += 1
167 #define NO_RTE_FOUND 0x1
168 #define RTE_FOUND 0x2
170 static struct mtx mrouter6_mtx;
171 #define MROUTER6_LOCK() mtx_lock(&mrouter6_mtx)
172 #define MROUTER6_UNLOCK() mtx_unlock(&mrouter6_mtx)
173 #define MROUTER6_LOCK_ASSERT() do { \
174 mtx_assert(&mrouter6_mtx, MA_OWNED); \
175 NET_ASSERT_GIANT(); \
177 #define MROUTER6_LOCK_INIT() \
178 mtx_init(&mrouter6_mtx, "IPv6 multicast forwarding", NULL, MTX_DEF)
179 #define MROUTER6_LOCK_DESTROY() mtx_destroy(&mrouter6_mtx)
181 static struct mf6c *mf6ctable[MF6CTBLSIZ];
182 SYSCTL_OPAQUE(_net_inet6_ip6, OID_AUTO, mf6ctable, CTLFLAG_RD,
183 &mf6ctable, sizeof(mf6ctable), "S,*mf6ctable[MF6CTBLSIZ]",
184 "IPv6 Multicast Forwarding Table (struct *mf6ctable[MF6CTBLSIZ], "
185 "netinet6/ip6_mroute.h)");
187 static struct mtx mfc6_mtx;
188 #define MFC6_LOCK() mtx_lock(&mfc6_mtx)
189 #define MFC6_UNLOCK() mtx_unlock(&mfc6_mtx)
190 #define MFC6_LOCK_ASSERT() do { \
191 mtx_assert(&mfc6_mtx, MA_OWNED); \
192 NET_ASSERT_GIANT(); \
194 #define MFC6_LOCK_INIT() \
195 mtx_init(&mfc6_mtx, "IPv6 multicast forwarding cache", NULL, MTX_DEF)
196 #define MFC6_LOCK_DESTROY() mtx_destroy(&mfc6_mtx)
198 static u_char n6expire[MF6CTBLSIZ];
200 static struct mif6 mif6table[MAXMIFS];
202 sysctl_mif6table(SYSCTL_HANDLER_ARGS)
204 struct mif6_sctl *out;
207 out = malloc(sizeof(struct mif6_sctl) * MAXMIFS, M_TEMP,
209 for (int i = 0; i < MAXMIFS; i++) {
210 out[i].m6_flags = mif6table[i].m6_flags;
211 out[i].m6_rate_limit = mif6table[i].m6_rate_limit;
212 out[i].m6_lcl_addr = mif6table[i].m6_lcl_addr;
213 if (mif6table[i].m6_ifp != NULL)
214 out[i].m6_ifp = mif6table[i].m6_ifp->if_index;
217 out[i].m6_pkt_in = mif6table[i].m6_pkt_in;
218 out[i].m6_pkt_out = mif6table[i].m6_pkt_out;
219 out[i].m6_bytes_in = mif6table[i].m6_bytes_in;
220 out[i].m6_bytes_out = mif6table[i].m6_bytes_out;
222 error = SYSCTL_OUT(req, out, sizeof(struct mif6_sctl) * MAXMIFS);
226 SYSCTL_PROC(_net_inet6_ip6, OID_AUTO, mif6table,
227 CTLTYPE_OPAQUE | CTLFLAG_RD | CTLFLAG_NEEDGIANT,
228 NULL, 0, sysctl_mif6table, "S,mif6_sctl[MAXMIFS]",
229 "IPv6 Multicast Interfaces (struct mif6_sctl[MAXMIFS], "
230 "netinet6/ip6_mroute.h)");
232 static struct mtx mif6_mtx;
233 #define MIF6_LOCK() mtx_lock(&mif6_mtx)
234 #define MIF6_UNLOCK() mtx_unlock(&mif6_mtx)
235 #define MIF6_LOCK_ASSERT() mtx_assert(&mif6_mtx, MA_OWNED)
236 #define MIF6_LOCK_INIT() \
237 mtx_init(&mif6_mtx, "IPv6 multicast interfaces", NULL, MTX_DEF)
238 #define MIF6_LOCK_DESTROY() mtx_destroy(&mif6_mtx)
241 VNET_DEFINE_STATIC(u_int, mrt6debug) = 0; /* debug level */
242 #define V_mrt6debug VNET(mrt6debug)
243 #define DEBUG_MFC 0x02
244 #define DEBUG_FORWARD 0x04
245 #define DEBUG_EXPIRE 0x08
246 #define DEBUG_XMIT 0x10
247 #define DEBUG_REG 0x20
248 #define DEBUG_PIM 0x40
249 #define DEBUG_ERR 0x80
250 #define DEBUG_ANY 0x7f
251 #define MRT6_DLOG(m, fmt, ...) \
252 if (V_mrt6debug & (m)) \
253 log(((m) & DEBUG_ERR) ? LOG_ERR: LOG_DEBUG, \
254 "%s: " fmt "\n", __func__, ##__VA_ARGS__)
256 #define MRT6_DLOG(m, fmt, ...)
259 static void expire_upcalls(void *);
260 #define EXPIRE_TIMEOUT (hz / 4) /* 4x / second */
261 #define UPCALL_EXPIRE 6 /* number of timeouts */
264 * XXX TODO: maintain a count to if_allmulti() calls in struct ifnet.
268 * 'Interfaces' associated with decapsulator (so we can tell
269 * packets that went through it from ones that get reflected
270 * by a broken gateway). Different from IPv4 register_if,
271 * these interfaces are linked into the system ifnet list,
272 * because per-interface IPv6 statistics are maintained in
273 * ifp->if_afdata. But it does not have any routes point
274 * to them. I.e., packets can't be sent this way. They
275 * only exist as a placeholder for multicast source
278 static struct ifnet *multicast_register_if6;
280 #define ENCAP_HOPS 64
285 static mifi_t nummifs = 0;
286 static mifi_t reg_mif_num = (mifi_t)-1;
288 static struct pim6stat pim6stat;
289 SYSCTL_STRUCT(_net_inet6_pim, PIM6CTL_STATS, stats, CTLFLAG_RW,
291 "PIM Statistics (struct pim6stat, netinet6/pim6_var.h)");
293 #define PIM6STAT_INC(name) pim6stat.name += 1
294 VNET_DEFINE_STATIC(int, pim6);
295 #define V_pim6 VNET(pim6)
298 * Hash function for a source, group entry
300 #define MF6CHASH(a, g) MF6CHASHMOD((a).s6_addr32[0] ^ (a).s6_addr32[1] ^ \
301 (a).s6_addr32[2] ^ (a).s6_addr32[3] ^ \
302 (g).s6_addr32[0] ^ (g).s6_addr32[1] ^ \
303 (g).s6_addr32[2] ^ (g).s6_addr32[3])
306 * Find a route for a given origin IPv6 address and Multicast group address.
308 #define MF6CFIND(o, g, rt) do { \
309 struct mf6c *_rt = mf6ctable[MF6CHASH(o,g)]; \
312 if (IN6_ARE_ADDR_EQUAL(&_rt->mf6c_origin.sin6_addr, &(o)) && \
313 IN6_ARE_ADDR_EQUAL(&_rt->mf6c_mcastgrp.sin6_addr, &(g)) && \
314 (_rt->mf6c_stall == NULL)) { \
318 _rt = _rt->mf6c_next; \
321 MRT6STAT_INC(mrt6s_mfc_misses); \
323 } while (/*CONSTCOND*/ 0)
326 * Macros to compute elapsed time efficiently
327 * Borrowed from Van Jacobson's scheduling code
328 * XXX: replace with timersub() ?
330 #define TV_DELTA(a, b, delta) do { \
333 delta = (a).tv_usec - (b).tv_usec; \
334 if ((xxs = (a).tv_sec - (b).tv_sec)) { \
343 delta += (1000000 * xxs); \
346 } while (/*CONSTCOND*/ 0)
348 /* XXX: replace with timercmp(a, b, <) ? */
349 #define TV_LT(a, b) (((a).tv_usec < (b).tv_usec && \
350 (a).tv_sec <= (b).tv_sec) || (a).tv_sec < (b).tv_sec)
353 #define UPCALL_MAX 50
354 static u_long upcall_data[UPCALL_MAX + 1];
355 static void collate();
356 #endif /* UPCALL_TIMING */
358 static int ip6_mrouter_init(struct socket *, int, int);
359 static int add_m6fc(struct mf6cctl *);
360 static int add_m6if(struct mif6ctl *);
361 static int del_m6fc(struct mf6cctl *);
362 static int del_m6if(mifi_t *);
363 static int del_m6if_locked(mifi_t *);
364 static int get_mif6_cnt(struct sioc_mif_req6 *);
365 static int get_sg_cnt(struct sioc_sg_req6 *);
367 static struct callout expire_upcalls_ch;
369 int X_ip6_mforward(struct ip6_hdr *, struct ifnet *, struct mbuf *);
370 int X_ip6_mrouter_done(void);
371 int X_ip6_mrouter_set(struct socket *, struct sockopt *);
372 int X_ip6_mrouter_get(struct socket *, struct sockopt *);
373 int X_mrt6_ioctl(u_long, caddr_t);
376 * Handle MRT setsockopt commands to modify the multicast routing tables.
379 X_ip6_mrouter_set(struct socket *so, struct sockopt *sopt)
387 if (so != V_ip6_mrouter && sopt->sopt_name != MRT6_INIT)
390 switch (sopt->sopt_name) {
395 error = sooptcopyin(sopt, &optval, sizeof(optval),
399 error = ip6_mrouter_init(so, optval, sopt->sopt_name);
402 error = X_ip6_mrouter_done();
405 error = sooptcopyin(sopt, &mifc, sizeof(mifc), sizeof(mifc));
408 error = add_m6if(&mifc);
411 error = sooptcopyin(sopt, &mfcc, sizeof(mfcc), sizeof(mfcc));
414 error = add_m6fc(&mfcc);
417 error = sooptcopyin(sopt, &mfcc, sizeof(mfcc), sizeof(mfcc));
420 error = del_m6fc(&mfcc);
423 error = sooptcopyin(sopt, &mifi, sizeof(mifi), sizeof(mifi));
426 error = del_m6if(&mifi);
429 error = sooptcopyin(sopt, &optval, sizeof(optval),
433 error = set_pim6(&optval);
444 * Handle MRT getsockopt commands
447 X_ip6_mrouter_get(struct socket *so, struct sockopt *sopt)
451 if (so != V_ip6_mrouter)
454 switch (sopt->sopt_name) {
456 error = sooptcopyout(sopt, &V_pim6, sizeof(V_pim6));
463 * Handle ioctl commands to obtain information from the cache
466 X_mrt6_ioctl(u_long cmd, caddr_t data)
473 case SIOCGETSGCNT_IN6:
474 ret = get_sg_cnt((struct sioc_sg_req6 *)data);
477 case SIOCGETMIFCNT_IN6:
478 ret = get_mif6_cnt((struct sioc_mif_req6 *)data);
489 * returns the packet, byte, rpf-failure count for the source group provided
492 get_sg_cnt(struct sioc_sg_req6 *req)
501 MF6CFIND(req->src.sin6_addr, req->grp.sin6_addr, rt);
505 req->pktcnt = rt->mf6c_pkt_cnt;
506 req->bytecnt = rt->mf6c_byte_cnt;
507 req->wrong_if = rt->mf6c_wrong_if;
516 * returns the input and output packet and byte counts on the mif provided
519 get_mif6_cnt(struct sioc_mif_req6 *req)
529 if (mifi >= nummifs) {
532 req->icount = mif6table[mifi].m6_pkt_in;
533 req->ocount = mif6table[mifi].m6_pkt_out;
534 req->ibytes = mif6table[mifi].m6_bytes_in;
535 req->obytes = mif6table[mifi].m6_bytes_out;
546 if ((*i != 1) && (*i != 0))
555 * Enable multicast routing
558 ip6_mrouter_init(struct socket *so, int v, int cmd)
561 MRT6_DLOG(DEBUG_ANY, "so_type = %d, pr_protocol = %d",
562 so->so_type, so->so_proto->pr_protocol);
564 if (so->so_type != SOCK_RAW ||
565 so->so_proto->pr_protocol != IPPROTO_ICMPV6)
569 return (ENOPROTOOPT);
573 if (V_ip6_mrouter != NULL) {
579 V_ip6_mrouter_ver = cmd;
581 bzero((caddr_t)mf6ctable, sizeof(mf6ctable));
582 bzero((caddr_t)n6expire, sizeof(n6expire));
584 V_pim6 = 0;/* used for stubbing out/in pim stuff */
586 callout_init(&expire_upcalls_ch, 0);
587 callout_reset(&expire_upcalls_ch, EXPIRE_TIMEOUT,
588 expire_upcalls, NULL);
591 MRT6_DLOG(DEBUG_ANY, "finished");
597 * Disable IPv6 multicast forwarding.
600 X_ip6_mrouter_done(void)
609 if (V_ip6_mrouter == NULL) {
615 * For each phyint in use, disable promiscuous reception of all IPv6
618 for (mifi = 0; mifi < nummifs; mifi++) {
619 if (mif6table[mifi].m6_ifp &&
620 !(mif6table[mifi].m6_flags & MIFF_REGISTER)) {
621 if_allmulti(mif6table[mifi].m6_ifp, 0);
624 bzero((caddr_t)mif6table, sizeof(mif6table));
627 V_pim6 = 0; /* used to stub out/in pim specific code */
629 callout_stop(&expire_upcalls_ch);
632 * Free all multicast forwarding cache entries.
635 for (i = 0; i < MF6CTBLSIZ; i++) {
640 for (rte = rt->mf6c_stall; rte != NULL; ) {
641 struct rtdetq *n = rte->next;
644 free(rte, M_MRTABLE6);
649 free(frt, M_MRTABLE6);
652 bzero((caddr_t)mf6ctable, sizeof(mf6ctable));
656 * Reset register interface
658 if (reg_mif_num != (mifi_t)-1 && multicast_register_if6 != NULL) {
659 if_detach(multicast_register_if6);
660 if_free(multicast_register_if6);
661 reg_mif_num = (mifi_t)-1;
662 multicast_register_if6 = NULL;
665 V_ip6_mrouter = NULL;
666 V_ip6_mrouter_ver = 0;
669 MRT6_DLOG(DEBUG_ANY, "finished");
674 static struct sockaddr_in6 sin6 = { sizeof(sin6), AF_INET6 };
677 * Add a mif to the mif table
680 add_m6if(struct mif6ctl *mifcp)
688 if (mifcp->mif6c_mifi >= MAXMIFS) {
692 mifp = mif6table + mifcp->mif6c_mifi;
693 if (mifp->m6_ifp != NULL) {
695 return (EADDRINUSE); /* XXX: is it appropriate? */
697 if (mifcp->mif6c_pifi == 0 || mifcp->mif6c_pifi > V_if_index) {
702 ifp = ifnet_byindex(mifcp->mif6c_pifi);
704 if (mifcp->mif6c_flags & MIFF_REGISTER) {
705 if (reg_mif_num == (mifi_t)-1) {
706 ifp = if_alloc(IFT_OTHER);
708 if_initname(ifp, "register_mif", 0);
709 ifp->if_flags |= IFF_LOOPBACK;
711 multicast_register_if6 = ifp;
712 reg_mif_num = mifcp->mif6c_mifi;
714 * it is impossible to guess the ifindex of the
715 * register interface. So mif6c_pifi is automatically
718 mifcp->mif6c_pifi = ifp->if_index;
720 ifp = multicast_register_if6;
723 /* Make sure the interface supports multicast */
724 if ((ifp->if_flags & IFF_MULTICAST) == 0) {
729 error = if_allmulti(ifp, 1);
736 mifp->m6_flags = mifcp->mif6c_flags;
739 /* initialize per mif pkt counters */
741 mifp->m6_pkt_out = 0;
742 mifp->m6_bytes_in = 0;
743 mifp->m6_bytes_out = 0;
745 /* Adjust nummifs up if the mifi is higher than nummifs */
746 if (nummifs <= mifcp->mif6c_mifi)
747 nummifs = mifcp->mif6c_mifi + 1;
750 MRT6_DLOG(DEBUG_ANY, "mif #%d, phyint %s", mifcp->mif6c_mifi,
757 * Delete a mif from the mif table
760 del_m6if_locked(mifi_t *mifip)
762 struct mif6 *mifp = mif6table + *mifip;
768 if (*mifip >= nummifs)
770 if (mifp->m6_ifp == NULL)
773 if (!(mifp->m6_flags & MIFF_REGISTER)) {
774 /* XXX: TODO: Maintain an ALLMULTI refcount in struct ifnet. */
778 if (reg_mif_num != (mifi_t)-1 &&
779 multicast_register_if6 != NULL) {
780 if_detach(multicast_register_if6);
781 if_free(multicast_register_if6);
782 reg_mif_num = (mifi_t)-1;
783 multicast_register_if6 = NULL;
787 bzero((caddr_t)mifp, sizeof(*mifp));
789 /* Adjust nummifs down */
790 for (mifi = nummifs; mifi > 0; mifi--)
791 if (mif6table[mifi - 1].m6_ifp)
794 MRT6_DLOG(DEBUG_ANY, "mif %d, nummifs %d", *mifip, nummifs);
800 del_m6if(mifi_t *mifip)
805 cc = del_m6if_locked(mifip);
815 add_m6fc(struct mf6cctl *mfccp)
821 char ip6bufo[INET6_ADDRSTRLEN], ip6bufg[INET6_ADDRSTRLEN];
825 MF6CFIND(mfccp->mf6cc_origin.sin6_addr,
826 mfccp->mf6cc_mcastgrp.sin6_addr, rt);
828 /* If an entry already exists, just update the fields */
830 MRT6_DLOG(DEBUG_MFC, "no upcall o %s g %s p %x",
831 ip6_sprintf(ip6bufo, &mfccp->mf6cc_origin.sin6_addr),
832 ip6_sprintf(ip6bufg, &mfccp->mf6cc_mcastgrp.sin6_addr),
833 mfccp->mf6cc_parent);
835 rt->mf6c_parent = mfccp->mf6cc_parent;
836 rt->mf6c_ifset = mfccp->mf6cc_ifset;
843 * Find the entry for which the upcall was made and update
845 hash = MF6CHASH(mfccp->mf6cc_origin.sin6_addr,
846 mfccp->mf6cc_mcastgrp.sin6_addr);
847 for (rt = mf6ctable[hash], nstl = 0; rt; rt = rt->mf6c_next) {
848 if (IN6_ARE_ADDR_EQUAL(&rt->mf6c_origin.sin6_addr,
849 &mfccp->mf6cc_origin.sin6_addr) &&
850 IN6_ARE_ADDR_EQUAL(&rt->mf6c_mcastgrp.sin6_addr,
851 &mfccp->mf6cc_mcastgrp.sin6_addr) &&
852 (rt->mf6c_stall != NULL)) {
855 "add_m6fc: %s o %s g %s p %x dbx %p\n",
856 "multiple kernel entries",
858 &mfccp->mf6cc_origin.sin6_addr),
860 &mfccp->mf6cc_mcastgrp.sin6_addr),
861 mfccp->mf6cc_parent, rt->mf6c_stall);
863 MRT6_DLOG(DEBUG_MFC, "o %s g %s p %x dbg %p",
865 &mfccp->mf6cc_origin.sin6_addr),
867 &mfccp->mf6cc_mcastgrp.sin6_addr),
868 mfccp->mf6cc_parent, rt->mf6c_stall);
870 rt->mf6c_origin = mfccp->mf6cc_origin;
871 rt->mf6c_mcastgrp = mfccp->mf6cc_mcastgrp;
872 rt->mf6c_parent = mfccp->mf6cc_parent;
873 rt->mf6c_ifset = mfccp->mf6cc_ifset;
874 /* initialize pkt counters per src-grp */
875 rt->mf6c_pkt_cnt = 0;
876 rt->mf6c_byte_cnt = 0;
877 rt->mf6c_wrong_if = 0;
879 rt->mf6c_expire = 0; /* Don't clean this guy up */
882 /* free packets Qed at the end of this entry */
883 for (rte = rt->mf6c_stall; rte != NULL; ) {
884 struct rtdetq *n = rte->next;
885 ip6_mdq(rte->m, rte->ifp, rt);
889 #endif /* UPCALL_TIMING */
890 free(rte, M_MRTABLE6);
893 rt->mf6c_stall = NULL;
898 * It is possible that an entry is being inserted without an upcall
901 MRT6_DLOG(DEBUG_MFC, "no upcall h %lu o %s g %s p %x", hash,
902 ip6_sprintf(ip6bufo, &mfccp->mf6cc_origin.sin6_addr),
903 ip6_sprintf(ip6bufg, &mfccp->mf6cc_mcastgrp.sin6_addr),
904 mfccp->mf6cc_parent);
906 for (rt = mf6ctable[hash]; rt; rt = rt->mf6c_next) {
907 if (IN6_ARE_ADDR_EQUAL(&rt->mf6c_origin.sin6_addr,
908 &mfccp->mf6cc_origin.sin6_addr)&&
909 IN6_ARE_ADDR_EQUAL(&rt->mf6c_mcastgrp.sin6_addr,
910 &mfccp->mf6cc_mcastgrp.sin6_addr)) {
911 rt->mf6c_origin = mfccp->mf6cc_origin;
912 rt->mf6c_mcastgrp = mfccp->mf6cc_mcastgrp;
913 rt->mf6c_parent = mfccp->mf6cc_parent;
914 rt->mf6c_ifset = mfccp->mf6cc_ifset;
915 /* initialize pkt counters per src-grp */
916 rt->mf6c_pkt_cnt = 0;
917 rt->mf6c_byte_cnt = 0;
918 rt->mf6c_wrong_if = 0;
926 /* no upcall, so make a new entry */
927 rt = (struct mf6c *)malloc(sizeof(*rt), M_MRTABLE6,
934 /* insert new entry at head of hash chain */
935 rt->mf6c_origin = mfccp->mf6cc_origin;
936 rt->mf6c_mcastgrp = mfccp->mf6cc_mcastgrp;
937 rt->mf6c_parent = mfccp->mf6cc_parent;
938 rt->mf6c_ifset = mfccp->mf6cc_ifset;
939 /* initialize pkt counters per src-grp */
940 rt->mf6c_pkt_cnt = 0;
941 rt->mf6c_byte_cnt = 0;
942 rt->mf6c_wrong_if = 0;
944 rt->mf6c_stall = NULL;
946 /* link into table */
947 rt->mf6c_next = mf6ctable[hash];
948 mf6ctable[hash] = rt;
958 * collect delay statistics on the upcalls
961 collate(struct timeval *t)
971 TV_DELTA(tp, *t, delta);
980 #endif /* UPCALL_TIMING */
983 * Delete an mfc entry
986 del_m6fc(struct mf6cctl *mfccp)
989 char ip6bufo[INET6_ADDRSTRLEN], ip6bufg[INET6_ADDRSTRLEN];
991 struct sockaddr_in6 origin;
992 struct sockaddr_in6 mcastgrp;
997 origin = mfccp->mf6cc_origin;
998 mcastgrp = mfccp->mf6cc_mcastgrp;
999 hash = MF6CHASH(origin.sin6_addr, mcastgrp.sin6_addr);
1001 MRT6_DLOG(DEBUG_MFC, "orig %s mcastgrp %s",
1002 ip6_sprintf(ip6bufo, &origin.sin6_addr),
1003 ip6_sprintf(ip6bufg, &mcastgrp.sin6_addr));
1007 nptr = &mf6ctable[hash];
1008 while ((rt = *nptr) != NULL) {
1009 if (IN6_ARE_ADDR_EQUAL(&origin.sin6_addr,
1010 &rt->mf6c_origin.sin6_addr) &&
1011 IN6_ARE_ADDR_EQUAL(&mcastgrp.sin6_addr,
1012 &rt->mf6c_mcastgrp.sin6_addr) &&
1013 rt->mf6c_stall == NULL)
1016 nptr = &rt->mf6c_next;
1020 return (EADDRNOTAVAIL);
1023 *nptr = rt->mf6c_next;
1024 free(rt, M_MRTABLE6);
1032 socket_send(struct socket *s, struct mbuf *mm, struct sockaddr_in6 *src)
1036 if (sbappendaddr(&s->so_rcv,
1037 (struct sockaddr *)src,
1038 mm, (struct mbuf *)0) != 0) {
1048 * IPv6 multicast forwarding function. This function assumes that the packet
1049 * pointed to by "ip6" has arrived on (or is about to be sent to) the interface
1050 * pointed to by "ifp", and the packet is to be relayed to other networks
1051 * that have members of the packet's destination IPv6 multicast group.
1053 * The packet is returned unscathed to the caller, unless it is
1054 * erroneous, in which case a non-zero return value tells the caller to
1057 * NOTE: this implementation assumes that m->m_pkthdr.rcvif is NULL iff
1058 * this function is called in the originating context (i.e., not when
1059 * forwarding a packet from other node). ip6_output(), which is currently the
1060 * only function that calls this function is called in the originating context,
1061 * explicitly ensures this condition. It is caller's responsibility to ensure
1062 * that if this function is called from somewhere else in the originating
1063 * context in the future.
1066 X_ip6_mforward(struct ip6_hdr *ip6, struct ifnet *ifp, struct mbuf *m)
1075 char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN];
1076 #ifdef UPCALL_TIMING
1080 #endif /* UPCALL_TIMING */
1082 MRT6_DLOG(DEBUG_FORWARD, "src %s, dst %s, ifindex %d",
1083 ip6_sprintf(ip6bufs, &ip6->ip6_src),
1084 ip6_sprintf(ip6bufd, &ip6->ip6_dst), ifp->if_index);
1087 * Don't forward a packet with Hop limit of zero or one,
1088 * or a packet destined to a local-only group.
1090 if (ip6->ip6_hlim <= 1 || IN6_IS_ADDR_MC_INTFACELOCAL(&ip6->ip6_dst) ||
1091 IN6_IS_ADDR_MC_LINKLOCAL(&ip6->ip6_dst))
1096 * Source address check: do not forward packets with unspecified
1097 * source. It was discussed in July 2000, on ipngwg mailing list.
1098 * This is rather more serious than unicast cases, because some
1099 * MLD packets can be sent with the unspecified source address
1100 * (although such packets must normally set 1 to the hop limit field).
1102 if (IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src)) {
1103 IP6STAT_INC(ip6s_cantforward);
1104 if (V_ip6_log_time + V_ip6_log_interval < time_uptime) {
1105 V_ip6_log_time = time_uptime;
1108 "from %s to %s nxt %d received on %s\n",
1109 ip6_sprintf(ip6bufs, &ip6->ip6_src),
1110 ip6_sprintf(ip6bufd, &ip6->ip6_dst),
1112 if_name(m->m_pkthdr.rcvif));
1120 * Determine forwarding mifs from the forwarding cache table
1122 MF6CFIND(ip6->ip6_src, ip6->ip6_dst, rt);
1123 MRT6STAT_INC(mrt6s_mfc_lookups);
1125 /* Entry exists, so forward if necessary */
1128 return (ip6_mdq(m, ifp, rt));
1132 * If we don't have a route for packet's origin,
1133 * Make a copy of the packet & send message to routing daemon.
1135 MRT6STAT_INC(mrt6s_no_route);
1136 MRT6_DLOG(DEBUG_FORWARD | DEBUG_MFC, "no rte s %s g %s",
1137 ip6_sprintf(ip6bufs, &ip6->ip6_src),
1138 ip6_sprintf(ip6bufd, &ip6->ip6_dst));
1141 * Allocate mbufs early so that we don't do extra work if we
1142 * are just going to fail anyway.
1144 rte = (struct rtdetq *)malloc(sizeof(*rte), M_MRTABLE6, M_NOWAIT);
1149 mb0 = m_copym(m, 0, M_COPYALL, M_NOWAIT);
1151 * Pullup packet header if needed before storing it,
1152 * as other references may modify it in the meantime.
1154 if (mb0 && (!M_WRITABLE(mb0) || mb0->m_len < sizeof(struct ip6_hdr)))
1155 mb0 = m_pullup(mb0, sizeof(struct ip6_hdr));
1157 free(rte, M_MRTABLE6);
1162 /* is there an upcall waiting for this packet? */
1163 hash = MF6CHASH(ip6->ip6_src, ip6->ip6_dst);
1164 for (rt = mf6ctable[hash]; rt; rt = rt->mf6c_next) {
1165 if (IN6_ARE_ADDR_EQUAL(&ip6->ip6_src,
1166 &rt->mf6c_origin.sin6_addr) &&
1167 IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst,
1168 &rt->mf6c_mcastgrp.sin6_addr) && (rt->mf6c_stall != NULL))
1175 struct omrt6msg *oim;
1177 /* no upcall, so make a new entry */
1178 rt = (struct mf6c *)malloc(sizeof(*rt), M_MRTABLE6, M_NOWAIT);
1180 free(rte, M_MRTABLE6);
1186 * Make a copy of the header to send to the user
1189 mm = m_copym(mb0, 0, sizeof(struct ip6_hdr), M_NOWAIT);
1191 free(rte, M_MRTABLE6);
1193 free(rt, M_MRTABLE6);
1199 * Send message to routing daemon
1201 sin6.sin6_addr = ip6->ip6_src;
1206 switch (V_ip6_mrouter_ver) {
1209 oim = mtod(mm, struct omrt6msg *);
1210 oim->im6_msgtype = MRT6MSG_NOCACHE;
1215 im = mtod(mm, struct mrt6msg *);
1216 im->im6_msgtype = MRT6MSG_NOCACHE;
1220 free(rte, M_MRTABLE6);
1222 free(rt, M_MRTABLE6);
1227 MRT6_DLOG(DEBUG_FORWARD, "getting the iif info in the kernel");
1228 for (mifp = mif6table, mifi = 0;
1229 mifi < nummifs && mifp->m6_ifp != ifp; mifp++, mifi++)
1232 switch (V_ip6_mrouter_ver) {
1235 oim->im6_mif = mifi;
1243 if (socket_send(V_ip6_mrouter, mm, &sin6) < 0) {
1244 log(LOG_WARNING, "ip6_mforward: ip6_mrouter "
1245 "socket queue full\n");
1246 MRT6STAT_INC(mrt6s_upq_sockfull);
1247 free(rte, M_MRTABLE6);
1249 free(rt, M_MRTABLE6);
1254 MRT6STAT_INC(mrt6s_upcalls);
1256 /* insert new entry at head of hash chain */
1257 bzero(rt, sizeof(*rt));
1258 rt->mf6c_origin.sin6_family = AF_INET6;
1259 rt->mf6c_origin.sin6_len = sizeof(struct sockaddr_in6);
1260 rt->mf6c_origin.sin6_addr = ip6->ip6_src;
1261 rt->mf6c_mcastgrp.sin6_family = AF_INET6;
1262 rt->mf6c_mcastgrp.sin6_len = sizeof(struct sockaddr_in6);
1263 rt->mf6c_mcastgrp.sin6_addr = ip6->ip6_dst;
1264 rt->mf6c_expire = UPCALL_EXPIRE;
1266 rt->mf6c_parent = MF6C_INCOMPLETE_PARENT;
1268 /* link into table */
1269 rt->mf6c_next = mf6ctable[hash];
1270 mf6ctable[hash] = rt;
1271 /* Add this entry to the end of the queue */
1272 rt->mf6c_stall = rte;
1274 /* determine if q has overflowed */
1278 for (p = &rt->mf6c_stall; *p != NULL; p = &(*p)->next)
1279 if (++npkts > MAX_UPQ6) {
1280 MRT6STAT_INC(mrt6s_upq_ovflw);
1281 free(rte, M_MRTABLE6);
1287 /* Add this entry to the end of the queue */
1294 #ifdef UPCALL_TIMING
1296 #endif /* UPCALL_TIMING */
1304 * Clean up cache entries if upcalls are not serviced
1305 * Call from the Slow Timeout mechanism, every half second.
1308 expire_upcalls(void *unused)
1311 char ip6bufo[INET6_ADDRSTRLEN], ip6bufg[INET6_ADDRSTRLEN];
1314 struct mf6c *mfc, **nptr;
1318 for (i = 0; i < MF6CTBLSIZ; i++) {
1319 if (n6expire[i] == 0)
1321 nptr = &mf6ctable[i];
1322 while ((mfc = *nptr) != NULL) {
1323 rte = mfc->mf6c_stall;
1325 * Skip real cache entries
1326 * Make sure it wasn't marked to not expire (shouldn't happen)
1330 mfc->mf6c_expire != 0 &&
1331 --mfc->mf6c_expire == 0) {
1332 MRT6_DLOG(DEBUG_EXPIRE, "expiring (%s %s)",
1333 ip6_sprintf(ip6bufo, &mfc->mf6c_origin.sin6_addr),
1334 ip6_sprintf(ip6bufg, &mfc->mf6c_mcastgrp.sin6_addr));
1336 * drop all the packets
1337 * free the mbuf with the pkt, if, timing info
1340 struct rtdetq *n = rte->next;
1342 free(rte, M_MRTABLE6);
1344 } while (rte != NULL);
1345 MRT6STAT_INC(mrt6s_cache_cleanups);
1348 *nptr = mfc->mf6c_next;
1349 free(mfc, M_MRTABLE6);
1351 nptr = &mfc->mf6c_next;
1356 callout_reset(&expire_upcalls_ch, EXPIRE_TIMEOUT,
1357 expire_upcalls, NULL);
1361 * Packet forwarding routine once entry in the cache is made
1364 ip6_mdq(struct mbuf *m, struct ifnet *ifp, struct mf6c *rt)
1366 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
1369 int plen = m->m_pkthdr.len;
1370 struct in6_addr src0, dst0; /* copies for local work */
1371 u_int32_t iszone, idzone, oszone, odzone;
1375 * Don't forward if it didn't arrive from the parent mif
1378 mifi = rt->mf6c_parent;
1379 if ((mifi >= nummifs) || (mif6table[mifi].m6_ifp != ifp)) {
1380 /* came in the wrong interface */
1381 MRT6_DLOG(DEBUG_FORWARD,
1382 "wrong if: ifid %d mifi %d mififid %x", ifp->if_index,
1383 mifi, mif6table[mifi].m6_ifp->if_index);
1384 MRT6STAT_INC(mrt6s_wrong_if);
1385 rt->mf6c_wrong_if++;
1387 * If we are doing PIM processing, and we are forwarding
1388 * packets on this interface, send a message to the
1391 /* have to make sure this is a valid mif */
1392 if (mifi < nummifs && mif6table[mifi].m6_ifp)
1393 if (V_pim6 && (m->m_flags & M_LOOP) == 0) {
1395 * Check the M_LOOP flag to avoid an
1396 * unnecessary PIM assert.
1397 * XXX: M_LOOP is an ad-hoc hack...
1399 static struct sockaddr_in6 sin6 =
1400 { sizeof(sin6), AF_INET6 };
1405 struct omrt6msg *oim;
1408 mm = m_copym(m, 0, sizeof(struct ip6_hdr),
1412 mm->m_len < sizeof(struct ip6_hdr)))
1413 mm = m_pullup(mm, sizeof(struct ip6_hdr));
1421 switch (V_ip6_mrouter_ver) {
1424 oim = mtod(mm, struct omrt6msg *);
1425 oim->im6_msgtype = MRT6MSG_WRONGMIF;
1430 im = mtod(mm, struct mrt6msg *);
1431 im->im6_msgtype = MRT6MSG_WRONGMIF;
1439 for (mifp = mif6table, iif = 0;
1440 iif < nummifs && mifp &&
1441 mifp->m6_ifp != ifp;
1445 switch (V_ip6_mrouter_ver) {
1449 sin6.sin6_addr = oim->im6_src;
1454 sin6.sin6_addr = im->im6_src;
1458 MRT6STAT_INC(mrt6s_upcalls);
1460 if (socket_send(V_ip6_mrouter, mm, &sin6) < 0) {
1461 MRT6_DLOG(DEBUG_ANY,
1462 "ip6_mrouter socket queue full");
1463 MRT6STAT_INC(mrt6s_upq_sockfull);
1465 } /* if socket Q full */
1468 } /* if wrong iif */
1470 /* If I sourced this packet, it counts as output, else it was input. */
1471 if (m->m_pkthdr.rcvif == NULL) {
1472 /* XXX: is rcvif really NULL when output?? */
1473 mif6table[mifi].m6_pkt_out++;
1474 mif6table[mifi].m6_bytes_out += plen;
1476 mif6table[mifi].m6_pkt_in++;
1477 mif6table[mifi].m6_bytes_in += plen;
1480 rt->mf6c_byte_cnt += plen;
1483 * For each mif, forward a copy of the packet if there are group
1484 * members downstream on the interface.
1486 src0 = ip6->ip6_src;
1487 dst0 = ip6->ip6_dst;
1488 if ((error = in6_setscope(&src0, ifp, &iszone)) != 0 ||
1489 (error = in6_setscope(&dst0, ifp, &idzone)) != 0) {
1490 IP6STAT_INC(ip6s_badscope);
1493 for (mifp = mif6table, mifi = 0; mifi < nummifs; mifp++, mifi++) {
1494 if (IF_ISSET(mifi, &rt->mf6c_ifset)) {
1496 * check if the outgoing packet is going to break
1498 * XXX For packets through PIM register tunnel
1499 * interface, we believe a routing daemon.
1501 if (!(mif6table[rt->mf6c_parent].m6_flags &
1503 !(mif6table[mifi].m6_flags & MIFF_REGISTER)) {
1504 if (in6_setscope(&src0, mif6table[mifi].m6_ifp,
1506 in6_setscope(&dst0, mif6table[mifi].m6_ifp,
1510 IP6STAT_INC(ip6s_badscope);
1516 mifp->m6_bytes_out += plen;
1517 if (mifp->m6_flags & MIFF_REGISTER)
1518 register_send(ip6, mifp, m);
1520 phyint_send(ip6, mifp, m);
1527 phyint_send(struct ip6_hdr *ip6, struct mif6 *mifp, struct mbuf *m)
1530 char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN];
1532 struct mbuf *mb_copy;
1533 struct ifnet *ifp = mifp->m6_ifp;
1538 * Make a new reference to the packet; make sure that
1539 * the IPv6 header is actually copied, not just referenced,
1540 * so that ip6_output() only scribbles on the copy.
1542 mb_copy = m_copym(m, 0, M_COPYALL, M_NOWAIT);
1544 (!M_WRITABLE(mb_copy) || mb_copy->m_len < sizeof(struct ip6_hdr)))
1545 mb_copy = m_pullup(mb_copy, sizeof(struct ip6_hdr));
1546 if (mb_copy == NULL) {
1549 /* set MCAST flag to the outgoing packet */
1550 mb_copy->m_flags |= M_MCAST;
1553 * If we sourced the packet, call ip6_output since we may devide
1554 * the packet into fragments when the packet is too big for the
1555 * outgoing interface.
1556 * Otherwise, we can simply send the packet to the interface
1559 if (m->m_pkthdr.rcvif == NULL) {
1560 struct ip6_moptions im6o;
1561 struct epoch_tracker et;
1563 im6o.im6o_multicast_ifp = ifp;
1564 /* XXX: ip6_output will override ip6->ip6_hlim */
1565 im6o.im6o_multicast_hlim = ip6->ip6_hlim;
1566 im6o.im6o_multicast_loop = 1;
1567 NET_EPOCH_ENTER(et);
1568 error = ip6_output(mb_copy, NULL, NULL, IPV6_FORWARDING, &im6o,
1572 MRT6_DLOG(DEBUG_XMIT, "mif %u err %d",
1573 (uint16_t)(mifp - mif6table), error);
1578 * If configured to loop back multicasts by default,
1579 * loop back a copy now.
1582 ip6_mloopback(ifp, m);
1585 * Put the packet into the sending queue of the outgoing interface
1586 * if it would fit in the MTU of the interface.
1588 linkmtu = IN6_LINKMTU(ifp);
1589 if (mb_copy->m_pkthdr.len <= linkmtu || linkmtu < IPV6_MMTU) {
1590 struct sockaddr_in6 dst6;
1592 bzero(&dst6, sizeof(dst6));
1593 dst6.sin6_len = sizeof(struct sockaddr_in6);
1594 dst6.sin6_family = AF_INET6;
1595 dst6.sin6_addr = ip6->ip6_dst;
1597 IP_PROBE(send, NULL, NULL, ip6, ifp, NULL, ip6);
1599 * We just call if_output instead of nd6_output here, since
1600 * we need no ND for a multicast forwarded packet...right?
1602 m_clrprotoflags(m); /* Avoid confusing lower layers. */
1603 error = (*ifp->if_output)(ifp, mb_copy,
1604 (struct sockaddr *)&dst6, NULL);
1605 MRT6_DLOG(DEBUG_XMIT, "mif %u err %d",
1606 (uint16_t)(mifp - mif6table), error);
1609 * pMTU discovery is intentionally disabled by default, since
1610 * various router may notify pMTU in multicast, which can be
1611 * a DDoS to a router
1613 if (V_ip6_mcast_pmtu)
1614 icmp6_error(mb_copy, ICMP6_PACKET_TOO_BIG, 0, linkmtu);
1616 MRT6_DLOG(DEBUG_XMIT, " packet too big on %s o %s "
1617 "g %s size %d (discarded)", if_name(ifp),
1618 ip6_sprintf(ip6bufs, &ip6->ip6_src),
1619 ip6_sprintf(ip6bufd, &ip6->ip6_dst),
1620 mb_copy->m_pkthdr.len);
1621 m_freem(mb_copy); /* simply discard the packet */
1627 register_send(struct ip6_hdr *ip6, struct mif6 *mif, struct mbuf *m)
1630 char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN];
1633 int i, len = m->m_pkthdr.len;
1634 static struct sockaddr_in6 sin6 = { sizeof(sin6), AF_INET6 };
1635 struct mrt6msg *im6;
1637 MRT6_DLOG(DEBUG_ANY, "src %s dst %s",
1638 ip6_sprintf(ip6bufs, &ip6->ip6_src),
1639 ip6_sprintf(ip6bufd, &ip6->ip6_dst));
1640 PIM6STAT_INC(pim6s_snd_registers);
1642 /* Make a copy of the packet to send to the user level process. */
1643 mm = m_gethdr(M_NOWAIT, MT_DATA);
1646 mm->m_data += max_linkhdr;
1647 mm->m_len = sizeof(struct ip6_hdr);
1649 if ((mm->m_next = m_copym(m, 0, M_COPYALL, M_NOWAIT)) == NULL) {
1653 i = MHLEN - M_LEADINGSPACE(mm);
1656 mm = m_pullup(mm, i);
1659 /* TODO: check it! */
1660 mm->m_pkthdr.len = len + sizeof(struct ip6_hdr);
1663 * Send message to routing daemon
1665 sin6.sin6_addr = ip6->ip6_src;
1667 im6 = mtod(mm, struct mrt6msg *);
1668 im6->im6_msgtype = MRT6MSG_WHOLEPKT;
1671 im6->im6_mif = mif - mif6table;
1673 /* iif info is not given for reg. encap.n */
1674 MRT6STAT_INC(mrt6s_upcalls);
1676 if (socket_send(V_ip6_mrouter, mm, &sin6) < 0) {
1677 MRT6_DLOG(DEBUG_ANY, "ip6_mrouter socket queue full");
1678 MRT6STAT_INC(mrt6s_upq_sockfull);
1685 * pim6_encapcheck() is called by the encap6_input() path at runtime to
1686 * determine if a packet is for PIM; allowing PIM to be dynamically loaded
1690 pim6_encapcheck(const struct mbuf *m __unused, int off __unused,
1691 int proto __unused, void *arg __unused)
1694 KASSERT(proto == IPPROTO_PIM, ("not for IPPROTO_PIM"));
1695 return (8); /* claim the datagram. */
1699 * PIM sparse mode hook
1700 * Receives the pim control messages, and passes them up to the listening
1701 * socket, using rip6_input.
1702 * The only message processed is the REGISTER pim message; the pim header
1703 * is stripped off, and the inner packet is passed to register_mforward.
1706 pim6_input(struct mbuf *m, int off, int proto, void *arg __unused)
1708 struct pim *pim; /* pointer to a pim struct */
1709 struct ip6_hdr *ip6;
1713 PIM6STAT_INC(pim6s_rcv_total);
1718 pimlen = m->m_pkthdr.len - off;
1719 if (pimlen < PIM_MINLEN) {
1720 PIM6STAT_INC(pim6s_rcv_tooshort);
1721 MRT6_DLOG(DEBUG_PIM, "PIM packet too short");
1723 return (IPPROTO_DONE);
1727 * if the packet is at least as big as a REGISTER, go ahead
1728 * and grab the PIM REGISTER header size, to avoid another
1729 * possible m_pullup() later.
1731 * PIM_MINLEN == pimhdr + u_int32 == 8
1732 * PIM6_REG_MINLEN == pimhdr + reghdr + eip6hdr == 4 + 4 + 40
1734 minlen = (pimlen >= PIM6_REG_MINLEN) ? PIM6_REG_MINLEN : PIM_MINLEN;
1737 * Make sure that the IP6 and PIM headers in contiguous memory, and
1738 * possibly the PIM REGISTER header
1740 if (m->m_len < off + minlen) {
1741 m = m_pullup(m, off + minlen);
1743 IP6STAT_INC(ip6s_exthdrtoolong);
1744 return (IPPROTO_DONE);
1747 ip6 = mtod(m, struct ip6_hdr *);
1748 pim = (struct pim *)((caddr_t)ip6 + off);
1750 #define PIM6_CHECKSUM
1751 #ifdef PIM6_CHECKSUM
1756 * Validate checksum.
1757 * If PIM REGISTER, exclude the data packet
1759 if (pim->pim_type == PIM_REGISTER)
1760 cksumlen = PIM_MINLEN;
1764 if (in6_cksum(m, IPPROTO_PIM, off, cksumlen)) {
1765 PIM6STAT_INC(pim6s_rcv_badsum);
1766 MRT6_DLOG(DEBUG_PIM, "invalid checksum");
1768 return (IPPROTO_DONE);
1771 #endif /* PIM_CHECKSUM */
1773 /* PIM version check */
1774 if (pim->pim_ver != PIM_VERSION) {
1775 PIM6STAT_INC(pim6s_rcv_badversion);
1776 MRT6_DLOG(DEBUG_ANY | DEBUG_ERR,
1777 "incorrect version %d, expecting %d",
1778 pim->pim_ver, PIM_VERSION);
1780 return (IPPROTO_DONE);
1783 if (pim->pim_type == PIM_REGISTER) {
1785 * since this is a REGISTER, we'll make a copy of the register
1786 * headers ip6+pim+u_int32_t+encap_ip6, to be passed up to the
1789 static struct sockaddr_in6 dst = { sizeof(dst), AF_INET6 };
1792 struct ip6_hdr *eip6;
1796 char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN];
1799 PIM6STAT_INC(pim6s_rcv_registers);
1801 if ((reg_mif_num >= nummifs) || (reg_mif_num == (mifi_t) -1)) {
1802 MRT6_DLOG(DEBUG_PIM, "register mif not set: %d",
1805 return (IPPROTO_DONE);
1808 reghdr = (u_int32_t *)(pim + 1);
1810 if ((ntohl(*reghdr) & PIM_NULL_REGISTER))
1811 goto pim6_input_to_daemon;
1816 if (pimlen < PIM6_REG_MINLEN) {
1817 PIM6STAT_INC(pim6s_rcv_tooshort);
1818 PIM6STAT_INC(pim6s_rcv_badregisters);
1819 MRT6_DLOG(DEBUG_ANY | DEBUG_ERR, "register packet "
1820 "size too small %d from %s",
1821 pimlen, ip6_sprintf(ip6bufs, &ip6->ip6_src));
1823 return (IPPROTO_DONE);
1826 eip6 = (struct ip6_hdr *) (reghdr + 1);
1827 MRT6_DLOG(DEBUG_PIM, "eip6: %s -> %s, eip6 plen %d",
1828 ip6_sprintf(ip6bufs, &eip6->ip6_src),
1829 ip6_sprintf(ip6bufd, &eip6->ip6_dst),
1830 ntohs(eip6->ip6_plen));
1832 /* verify the version number of the inner packet */
1833 if ((eip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) {
1834 PIM6STAT_INC(pim6s_rcv_badregisters);
1835 MRT6_DLOG(DEBUG_ANY, "invalid IP version (%d) "
1836 "of the inner packet",
1837 (eip6->ip6_vfc & IPV6_VERSION));
1839 return (IPPROTO_DONE);
1842 /* verify the inner packet is destined to a mcast group */
1843 if (!IN6_IS_ADDR_MULTICAST(&eip6->ip6_dst)) {
1844 PIM6STAT_INC(pim6s_rcv_badregisters);
1845 MRT6_DLOG(DEBUG_PIM, "inner packet of register "
1846 "is not multicast %s",
1847 ip6_sprintf(ip6bufd, &eip6->ip6_dst));
1849 return (IPPROTO_DONE);
1853 * make a copy of the whole header to pass to the daemon later.
1855 mcp = m_copym(m, 0, off + PIM6_REG_MINLEN, M_NOWAIT);
1857 MRT6_DLOG(DEBUG_ANY | DEBUG_ERR, "pim register: "
1858 "could not copy register head");
1860 return (IPPROTO_DONE);
1864 * forward the inner ip6 packet; point m_data at the inner ip6.
1866 m_adj(m, off + PIM_MINLEN);
1867 MRT6_DLOG(DEBUG_PIM, "forwarding decapsulated register: "
1868 "src %s, dst %s, mif %d",
1869 ip6_sprintf(ip6bufs, &eip6->ip6_src),
1870 ip6_sprintf(ip6bufd, &eip6->ip6_dst), reg_mif_num);
1872 rc = if_simloop(mif6table[reg_mif_num].m6_ifp, m,
1873 dst.sin6_family, 0);
1875 /* prepare the register head to send to the mrouting daemon */
1880 * Pass the PIM message up to the daemon; if it is a register message
1881 * pass the 'head' only up to the daemon. This includes the
1882 * encapsulator ip6 header, pim header, register header and the
1883 * encapsulated ip6 header.
1885 pim6_input_to_daemon:
1886 return (rip6_input(&m, &off, proto));
1890 ip6_mroute_modevent(module_t mod, int type, void *unused)
1895 MROUTER6_LOCK_INIT();
1899 pim6_encap_cookie = ip6_encap_attach(&ipv6_encap_cfg,
1901 if (pim6_encap_cookie == NULL) {
1902 printf("ip6_mroute: unable to attach pim6 encap\n");
1903 MIF6_LOCK_DESTROY();
1904 MFC6_LOCK_DESTROY();
1905 MROUTER6_LOCK_DESTROY();
1909 ip6_mforward = X_ip6_mforward;
1910 ip6_mrouter_done = X_ip6_mrouter_done;
1911 ip6_mrouter_get = X_ip6_mrouter_get;
1912 ip6_mrouter_set = X_ip6_mrouter_set;
1913 mrt6_ioctl = X_mrt6_ioctl;
1917 if (V_ip6_mrouter != NULL)
1920 if (pim6_encap_cookie) {
1921 ip6_encap_detach(pim6_encap_cookie);
1922 pim6_encap_cookie = NULL;
1924 X_ip6_mrouter_done();
1925 ip6_mforward = NULL;
1926 ip6_mrouter_done = NULL;
1927 ip6_mrouter_get = NULL;
1928 ip6_mrouter_set = NULL;
1931 MIF6_LOCK_DESTROY();
1932 MFC6_LOCK_DESTROY();
1933 MROUTER6_LOCK_DESTROY();
1937 return (EOPNOTSUPP);
1943 static moduledata_t ip6_mroutemod = {
1945 ip6_mroute_modevent,
1949 DECLARE_MODULE(ip6_mroute, ip6_mroutemod, SI_SUB_PROTO_MC, SI_ORDER_ANY);