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_private.h>
112 #include <net/if_types.h>
113 #include <net/vnet.h>
115 #include <netinet/in.h>
116 #include <netinet/in_var.h>
117 #include <netinet/icmp6.h>
118 #include <netinet/ip_encap.h>
120 #include <netinet/ip6.h>
121 #include <netinet/in_kdtrace.h>
122 #include <netinet6/ip6_var.h>
123 #include <netinet6/scope6_var.h>
124 #include <netinet6/nd6.h>
125 #include <netinet6/ip6_mroute.h>
126 #include <netinet6/pim6.h>
127 #include <netinet6/pim6_var.h>
129 static MALLOC_DEFINE(M_MRTABLE6, "mf6c", "multicast forwarding cache entry");
131 static int ip6_mdq(struct mbuf *, struct ifnet *, struct mf6c *);
132 static void phyint_send(struct ip6_hdr *, struct mif6 *, struct mbuf *);
133 static int register_send(struct ip6_hdr *, struct mif6 *, struct mbuf *);
134 static int set_pim6(int *);
135 static int socket_send(struct socket *, struct mbuf *,
136 struct sockaddr_in6 *);
138 extern int in6_mcast_loop;
139 extern struct domain inet6domain;
141 static const struct encaptab *pim6_encap_cookie;
142 static int pim6_encapcheck(const struct mbuf *, int, int, void *);
143 static int pim6_input(struct mbuf *, int, int, void *);
145 static const struct encap_config ipv6_encap_cfg = {
146 .proto = IPPROTO_PIM,
147 .min_length = sizeof(struct ip6_hdr) + PIM_MINLEN,
149 .check = pim6_encapcheck,
153 VNET_DEFINE_STATIC(int, ip6_mrouter_ver) = 0;
154 #define V_ip6_mrouter_ver VNET(ip6_mrouter_ver)
156 SYSCTL_DECL(_net_inet6);
157 SYSCTL_DECL(_net_inet6_ip6);
158 static SYSCTL_NODE(_net_inet6, IPPROTO_PIM, pim,
159 CTLFLAG_RW | CTLFLAG_MPSAFE, 0,
162 static struct mrt6stat mrt6stat;
163 SYSCTL_STRUCT(_net_inet6_ip6, OID_AUTO, mrt6stat, CTLFLAG_RW,
165 "Multicast Routing Statistics (struct mrt6stat, netinet6/ip6_mroute.h)");
167 #define MRT6STAT_INC(name) mrt6stat.name += 1
168 #define NO_RTE_FOUND 0x1
169 #define RTE_FOUND 0x2
171 static struct sx mrouter6_mtx;
172 #define MROUTER6_LOCKPTR() (&mrouter6_mtx)
173 #define MROUTER6_LOCK() sx_xlock(MROUTER6_LOCKPTR())
174 #define MROUTER6_UNLOCK() sx_xunlock(MROUTER6_LOCKPTR())
175 #define MROUTER6_LOCK_ASSERT() sx_assert(MROUTER6_LOCKPTR(), SA_XLOCKED
176 #define MROUTER6_LOCK_INIT() sx_init(MROUTER6_LOCKPTR(), "mrouter6")
177 #define MROUTER6_LOCK_DESTROY() sx_destroy(MROUTER6_LOCKPTR())
179 static struct mf6c *mf6ctable[MF6CTBLSIZ];
180 SYSCTL_OPAQUE(_net_inet6_ip6, OID_AUTO, mf6ctable, CTLFLAG_RD,
181 &mf6ctable, sizeof(mf6ctable), "S,*mf6ctable[MF6CTBLSIZ]",
182 "IPv6 Multicast Forwarding Table (struct *mf6ctable[MF6CTBLSIZ], "
183 "netinet6/ip6_mroute.h)");
185 static struct mtx mfc6_mtx;
186 #define MFC6_LOCKPTR() (&mfc6_mtx)
187 #define MFC6_LOCK() mtx_lock(MFC6_LOCKPTR())
188 #define MFC6_UNLOCK() mtx_unlock(MFC6_LOCKPTR())
189 #define MFC6_LOCK_ASSERT() mtx_assert(MFC6_LOCKPTR(), MA_OWNED)
190 #define MFC6_LOCK_INIT() mtx_init(MFC6_LOCKPTR(), \
191 "IPv6 multicast forwarding cache", \
193 #define MFC6_LOCK_DESTROY() mtx_destroy(MFC6_LOCKPTR())
195 static u_char n6expire[MF6CTBLSIZ];
197 static struct mif6 mif6table[MAXMIFS];
199 sysctl_mif6table(SYSCTL_HANDLER_ARGS)
201 struct mif6_sctl *out;
204 out = malloc(sizeof(struct mif6_sctl) * MAXMIFS, M_TEMP,
206 for (int i = 0; i < MAXMIFS; i++) {
207 out[i].m6_flags = mif6table[i].m6_flags;
208 out[i].m6_rate_limit = mif6table[i].m6_rate_limit;
209 out[i].m6_lcl_addr = mif6table[i].m6_lcl_addr;
210 if (mif6table[i].m6_ifp != NULL)
211 out[i].m6_ifp = mif6table[i].m6_ifp->if_index;
214 out[i].m6_pkt_in = mif6table[i].m6_pkt_in;
215 out[i].m6_pkt_out = mif6table[i].m6_pkt_out;
216 out[i].m6_bytes_in = mif6table[i].m6_bytes_in;
217 out[i].m6_bytes_out = mif6table[i].m6_bytes_out;
219 error = SYSCTL_OUT(req, out, sizeof(struct mif6_sctl) * MAXMIFS);
223 SYSCTL_PROC(_net_inet6_ip6, OID_AUTO, mif6table,
224 CTLTYPE_OPAQUE | CTLFLAG_RD | CTLFLAG_NEEDGIANT,
225 NULL, 0, sysctl_mif6table, "S,mif6_sctl[MAXMIFS]",
226 "IPv6 Multicast Interfaces (struct mif6_sctl[MAXMIFS], "
227 "netinet6/ip6_mroute.h)");
229 static struct mtx mif6_mtx;
230 #define MIF6_LOCKPTR() (&mif6_mtx)
231 #define MIF6_LOCK() mtx_lock(MIF6_LOCKPTR())
232 #define MIF6_UNLOCK() mtx_unlock(MIF6_LOCKPTR())
233 #define MIF6_LOCK_ASSERT() mtx_assert(MIF6_LOCKPTR(), MA_OWNED)
234 #define MIF6_LOCK_INIT() \
235 mtx_init(MIF6_LOCKPTR(), "IPv6 multicast interfaces", NULL, MTX_DEF)
236 #define MIF6_LOCK_DESTROY() mtx_destroy(MIF6_LOCKPTR())
239 VNET_DEFINE_STATIC(u_int, mrt6debug) = 0; /* debug level */
240 #define V_mrt6debug VNET(mrt6debug)
241 #define DEBUG_MFC 0x02
242 #define DEBUG_FORWARD 0x04
243 #define DEBUG_EXPIRE 0x08
244 #define DEBUG_XMIT 0x10
245 #define DEBUG_REG 0x20
246 #define DEBUG_PIM 0x40
247 #define DEBUG_ERR 0x80
248 #define DEBUG_ANY 0x7f
249 #define MRT6_DLOG(m, fmt, ...) \
250 if (V_mrt6debug & (m)) \
251 log(((m) & DEBUG_ERR) ? LOG_ERR: LOG_DEBUG, \
252 "%s: " fmt "\n", __func__, ##__VA_ARGS__)
254 #define MRT6_DLOG(m, fmt, ...)
257 static void expire_upcalls(void *);
258 #define EXPIRE_TIMEOUT (hz / 4) /* 4x / second */
259 #define UPCALL_EXPIRE 6 /* number of timeouts */
262 * XXX TODO: maintain a count to if_allmulti() calls in struct ifnet.
266 * 'Interfaces' associated with decapsulator (so we can tell
267 * packets that went through it from ones that get reflected
268 * by a broken gateway). Different from IPv4 register_if,
269 * these interfaces are linked into the system ifnet list,
270 * because per-interface IPv6 statistics are maintained in
271 * ifp->if_afdata. But it does not have any routes point
272 * to them. I.e., packets can't be sent this way. They
273 * only exist as a placeholder for multicast source
276 static struct ifnet *multicast_register_if6;
278 #define ENCAP_HOPS 64
283 static mifi_t nummifs = 0;
284 static mifi_t reg_mif_num = (mifi_t)-1;
286 static struct pim6stat pim6stat;
287 SYSCTL_STRUCT(_net_inet6_pim, PIM6CTL_STATS, stats, CTLFLAG_RW,
289 "PIM Statistics (struct pim6stat, netinet6/pim6_var.h)");
291 #define PIM6STAT_INC(name) pim6stat.name += 1
292 VNET_DEFINE_STATIC(int, pim6);
293 #define V_pim6 VNET(pim6)
296 * Hash function for a source, group entry
298 #define MF6CHASH(a, g) MF6CHASHMOD((a).s6_addr32[0] ^ (a).s6_addr32[1] ^ \
299 (a).s6_addr32[2] ^ (a).s6_addr32[3] ^ \
300 (g).s6_addr32[0] ^ (g).s6_addr32[1] ^ \
301 (g).s6_addr32[2] ^ (g).s6_addr32[3])
304 * Find a route for a given origin IPv6 address and Multicast group address.
306 #define MF6CFIND(o, g, rt) do { \
307 struct mf6c *_rt = mf6ctable[MF6CHASH(o,g)]; \
310 if (IN6_ARE_ADDR_EQUAL(&_rt->mf6c_origin.sin6_addr, &(o)) && \
311 IN6_ARE_ADDR_EQUAL(&_rt->mf6c_mcastgrp.sin6_addr, &(g)) && \
312 (_rt->mf6c_stall == NULL)) { \
316 _rt = _rt->mf6c_next; \
319 MRT6STAT_INC(mrt6s_mfc_misses); \
321 } while (/*CONSTCOND*/ 0)
324 * Macros to compute elapsed time efficiently
325 * Borrowed from Van Jacobson's scheduling code
326 * XXX: replace with timersub() ?
328 #define TV_DELTA(a, b, delta) do { \
331 delta = (a).tv_usec - (b).tv_usec; \
332 if ((xxs = (a).tv_sec - (b).tv_sec)) { \
341 delta += (1000000 * xxs); \
344 } while (/*CONSTCOND*/ 0)
346 /* XXX: replace with timercmp(a, b, <) ? */
347 #define TV_LT(a, b) (((a).tv_usec < (b).tv_usec && \
348 (a).tv_sec <= (b).tv_sec) || (a).tv_sec < (b).tv_sec)
351 #define UPCALL_MAX 50
352 static u_long upcall_data[UPCALL_MAX + 1];
353 static void collate();
354 #endif /* UPCALL_TIMING */
356 static int ip6_mrouter_init(struct socket *, int, int);
357 static int add_m6fc(struct mf6cctl *);
358 static int add_m6if(struct mif6ctl *);
359 static int del_m6fc(struct mf6cctl *);
360 static int del_m6if(mifi_t *);
361 static int del_m6if_locked(mifi_t *);
362 static int get_mif6_cnt(struct sioc_mif_req6 *);
363 static int get_sg_cnt(struct sioc_sg_req6 *);
365 static struct callout expire_upcalls_ch;
367 int X_ip6_mforward(struct ip6_hdr *, struct ifnet *, struct mbuf *);
368 int X_ip6_mrouter_done(void);
369 int X_ip6_mrouter_set(struct socket *, struct sockopt *);
370 int X_ip6_mrouter_get(struct socket *, struct sockopt *);
371 int X_mrt6_ioctl(u_long, caddr_t);
374 * Handle MRT setsockopt commands to modify the multicast routing tables.
377 X_ip6_mrouter_set(struct socket *so, struct sockopt *sopt)
385 if (so != V_ip6_mrouter && sopt->sopt_name != MRT6_INIT)
388 switch (sopt->sopt_name) {
393 error = sooptcopyin(sopt, &optval, sizeof(optval),
397 error = ip6_mrouter_init(so, optval, sopt->sopt_name);
400 error = X_ip6_mrouter_done();
403 error = sooptcopyin(sopt, &mifc, sizeof(mifc), sizeof(mifc));
406 error = add_m6if(&mifc);
409 error = sooptcopyin(sopt, &mfcc, sizeof(mfcc), sizeof(mfcc));
412 error = add_m6fc(&mfcc);
415 error = sooptcopyin(sopt, &mfcc, sizeof(mfcc), sizeof(mfcc));
418 error = del_m6fc(&mfcc);
421 error = sooptcopyin(sopt, &mifi, sizeof(mifi), sizeof(mifi));
424 error = del_m6if(&mifi);
427 error = sooptcopyin(sopt, &optval, sizeof(optval),
431 error = set_pim6(&optval);
442 * Handle MRT getsockopt commands
445 X_ip6_mrouter_get(struct socket *so, struct sockopt *sopt)
449 if (so != V_ip6_mrouter)
452 switch (sopt->sopt_name) {
454 error = sooptcopyout(sopt, &V_pim6, sizeof(V_pim6));
461 * Handle ioctl commands to obtain information from the cache
464 X_mrt6_ioctl(u_long cmd, caddr_t data)
471 case SIOCGETSGCNT_IN6:
472 ret = get_sg_cnt((struct sioc_sg_req6 *)data);
475 case SIOCGETMIFCNT_IN6:
476 ret = get_mif6_cnt((struct sioc_mif_req6 *)data);
487 * returns the packet, byte, rpf-failure count for the source group provided
490 get_sg_cnt(struct sioc_sg_req6 *req)
499 MF6CFIND(req->src.sin6_addr, req->grp.sin6_addr, rt);
503 req->pktcnt = rt->mf6c_pkt_cnt;
504 req->bytecnt = rt->mf6c_byte_cnt;
505 req->wrong_if = rt->mf6c_wrong_if;
514 * returns the input and output packet and byte counts on the mif provided
517 get_mif6_cnt(struct sioc_mif_req6 *req)
527 if (mifi >= nummifs) {
530 req->icount = mif6table[mifi].m6_pkt_in;
531 req->ocount = mif6table[mifi].m6_pkt_out;
532 req->ibytes = mif6table[mifi].m6_bytes_in;
533 req->obytes = mif6table[mifi].m6_bytes_out;
544 if ((*i != 1) && (*i != 0))
553 * Enable multicast routing
556 ip6_mrouter_init(struct socket *so, int v, int cmd)
559 MRT6_DLOG(DEBUG_ANY, "%s: socket %p", __func__, so);
562 return (ENOPROTOOPT);
566 if (V_ip6_mrouter != NULL) {
572 V_ip6_mrouter_ver = cmd;
574 bzero((caddr_t)mf6ctable, sizeof(mf6ctable));
575 bzero((caddr_t)n6expire, sizeof(n6expire));
577 V_pim6 = 0;/* used for stubbing out/in pim stuff */
579 callout_init_mtx(&expire_upcalls_ch, MFC6_LOCKPTR(), 0);
580 callout_reset(&expire_upcalls_ch, EXPIRE_TIMEOUT,
581 expire_upcalls, NULL);
585 MRT6_DLOG(DEBUG_ANY, "finished");
591 * Disable IPv6 multicast forwarding.
594 X_ip6_mrouter_done(void)
603 if (V_ip6_mrouter == NULL) {
609 * For each phyint in use, disable promiscuous reception of all IPv6
612 for (mifi = 0; mifi < nummifs; mifi++) {
613 if (mif6table[mifi].m6_ifp &&
614 !(mif6table[mifi].m6_flags & MIFF_REGISTER)) {
615 if_allmulti(mif6table[mifi].m6_ifp, 0);
618 bzero((caddr_t)mif6table, sizeof(mif6table));
621 V_pim6 = 0; /* used to stub out/in pim specific code */
624 * Free all multicast forwarding cache entries.
627 for (i = 0; i < MF6CTBLSIZ; i++) {
632 for (rte = rt->mf6c_stall; rte != NULL; ) {
633 struct rtdetq *n = rte->next;
636 free(rte, M_MRTABLE6);
641 free(frt, M_MRTABLE6);
644 bzero((caddr_t)mf6ctable, sizeof(mf6ctable));
647 callout_drain(&expire_upcalls_ch);
650 * Reset register interface
652 if (reg_mif_num != (mifi_t)-1 && multicast_register_if6 != NULL) {
653 if_detach(multicast_register_if6);
654 if_free(multicast_register_if6);
655 reg_mif_num = (mifi_t)-1;
656 multicast_register_if6 = NULL;
659 V_ip6_mrouter = NULL;
660 V_ip6_mrouter_ver = 0;
663 MRT6_DLOG(DEBUG_ANY, "finished");
668 static struct sockaddr_in6 sin6 = { sizeof(sin6), AF_INET6 };
671 * Add a mif to the mif table
674 add_m6if(struct mif6ctl *mifcp)
676 struct epoch_tracker et;
683 if (mifcp->mif6c_mifi >= MAXMIFS) {
687 mifp = mif6table + mifcp->mif6c_mifi;
688 if (mifp->m6_ifp != NULL) {
690 return (EADDRINUSE); /* XXX: is it appropriate? */
694 if ((ifp = ifnet_byindex(mifcp->mif6c_pifi)) == NULL) {
699 NET_EPOCH_EXIT(et); /* XXXGL: unsafe ifp */
701 if (mifcp->mif6c_flags & MIFF_REGISTER) {
702 if (reg_mif_num == (mifi_t)-1) {
703 ifp = if_alloc(IFT_OTHER);
705 if_initname(ifp, "register_mif", 0);
706 ifp->if_flags |= IFF_LOOPBACK;
708 multicast_register_if6 = ifp;
709 reg_mif_num = mifcp->mif6c_mifi;
711 * it is impossible to guess the ifindex of the
712 * register interface. So mif6c_pifi is automatically
715 mifcp->mif6c_pifi = ifp->if_index;
717 ifp = multicast_register_if6;
720 /* Make sure the interface supports multicast */
721 if ((ifp->if_flags & IFF_MULTICAST) == 0) {
726 error = if_allmulti(ifp, 1);
733 mifp->m6_flags = mifcp->mif6c_flags;
736 /* initialize per mif pkt counters */
738 mifp->m6_pkt_out = 0;
739 mifp->m6_bytes_in = 0;
740 mifp->m6_bytes_out = 0;
742 /* Adjust nummifs up if the mifi is higher than nummifs */
743 if (nummifs <= mifcp->mif6c_mifi)
744 nummifs = mifcp->mif6c_mifi + 1;
747 MRT6_DLOG(DEBUG_ANY, "mif #%d, phyint %s", mifcp->mif6c_mifi,
754 * Delete a mif from the mif table
757 del_m6if_locked(mifi_t *mifip)
759 struct mif6 *mifp = mif6table + *mifip;
765 if (*mifip >= nummifs)
767 if (mifp->m6_ifp == NULL)
770 if (!(mifp->m6_flags & MIFF_REGISTER)) {
771 /* XXX: TODO: Maintain an ALLMULTI refcount in struct ifnet. */
775 if (reg_mif_num != (mifi_t)-1 &&
776 multicast_register_if6 != NULL) {
777 if_detach(multicast_register_if6);
778 if_free(multicast_register_if6);
779 reg_mif_num = (mifi_t)-1;
780 multicast_register_if6 = NULL;
784 bzero((caddr_t)mifp, sizeof(*mifp));
786 /* Adjust nummifs down */
787 for (mifi = nummifs; mifi > 0; mifi--)
788 if (mif6table[mifi - 1].m6_ifp)
791 MRT6_DLOG(DEBUG_ANY, "mif %d, nummifs %d", *mifip, nummifs);
797 del_m6if(mifi_t *mifip)
802 cc = del_m6if_locked(mifip);
812 add_m6fc(struct mf6cctl *mfccp)
818 char ip6bufo[INET6_ADDRSTRLEN], ip6bufg[INET6_ADDRSTRLEN];
822 MF6CFIND(mfccp->mf6cc_origin.sin6_addr,
823 mfccp->mf6cc_mcastgrp.sin6_addr, rt);
825 /* If an entry already exists, just update the fields */
827 MRT6_DLOG(DEBUG_MFC, "no upcall o %s g %s p %x",
828 ip6_sprintf(ip6bufo, &mfccp->mf6cc_origin.sin6_addr),
829 ip6_sprintf(ip6bufg, &mfccp->mf6cc_mcastgrp.sin6_addr),
830 mfccp->mf6cc_parent);
832 rt->mf6c_parent = mfccp->mf6cc_parent;
833 rt->mf6c_ifset = mfccp->mf6cc_ifset;
840 * Find the entry for which the upcall was made and update
842 hash = MF6CHASH(mfccp->mf6cc_origin.sin6_addr,
843 mfccp->mf6cc_mcastgrp.sin6_addr);
844 for (rt = mf6ctable[hash], nstl = 0; rt; rt = rt->mf6c_next) {
845 if (IN6_ARE_ADDR_EQUAL(&rt->mf6c_origin.sin6_addr,
846 &mfccp->mf6cc_origin.sin6_addr) &&
847 IN6_ARE_ADDR_EQUAL(&rt->mf6c_mcastgrp.sin6_addr,
848 &mfccp->mf6cc_mcastgrp.sin6_addr) &&
849 (rt->mf6c_stall != NULL)) {
852 "add_m6fc: %s o %s g %s p %x dbx %p\n",
853 "multiple kernel entries",
855 &mfccp->mf6cc_origin.sin6_addr),
857 &mfccp->mf6cc_mcastgrp.sin6_addr),
858 mfccp->mf6cc_parent, rt->mf6c_stall);
860 MRT6_DLOG(DEBUG_MFC, "o %s g %s p %x dbg %p",
862 &mfccp->mf6cc_origin.sin6_addr),
864 &mfccp->mf6cc_mcastgrp.sin6_addr),
865 mfccp->mf6cc_parent, rt->mf6c_stall);
867 rt->mf6c_origin = mfccp->mf6cc_origin;
868 rt->mf6c_mcastgrp = mfccp->mf6cc_mcastgrp;
869 rt->mf6c_parent = mfccp->mf6cc_parent;
870 rt->mf6c_ifset = mfccp->mf6cc_ifset;
871 /* initialize pkt counters per src-grp */
872 rt->mf6c_pkt_cnt = 0;
873 rt->mf6c_byte_cnt = 0;
874 rt->mf6c_wrong_if = 0;
876 rt->mf6c_expire = 0; /* Don't clean this guy up */
879 /* free packets Qed at the end of this entry */
880 for (rte = rt->mf6c_stall; rte != NULL; ) {
881 struct rtdetq *n = rte->next;
882 ip6_mdq(rte->m, rte->ifp, rt);
886 #endif /* UPCALL_TIMING */
887 free(rte, M_MRTABLE6);
890 rt->mf6c_stall = NULL;
895 * It is possible that an entry is being inserted without an upcall
898 MRT6_DLOG(DEBUG_MFC, "no upcall h %lu o %s g %s p %x", hash,
899 ip6_sprintf(ip6bufo, &mfccp->mf6cc_origin.sin6_addr),
900 ip6_sprintf(ip6bufg, &mfccp->mf6cc_mcastgrp.sin6_addr),
901 mfccp->mf6cc_parent);
903 for (rt = mf6ctable[hash]; rt; rt = rt->mf6c_next) {
904 if (IN6_ARE_ADDR_EQUAL(&rt->mf6c_origin.sin6_addr,
905 &mfccp->mf6cc_origin.sin6_addr)&&
906 IN6_ARE_ADDR_EQUAL(&rt->mf6c_mcastgrp.sin6_addr,
907 &mfccp->mf6cc_mcastgrp.sin6_addr)) {
908 rt->mf6c_origin = mfccp->mf6cc_origin;
909 rt->mf6c_mcastgrp = mfccp->mf6cc_mcastgrp;
910 rt->mf6c_parent = mfccp->mf6cc_parent;
911 rt->mf6c_ifset = mfccp->mf6cc_ifset;
912 /* initialize pkt counters per src-grp */
913 rt->mf6c_pkt_cnt = 0;
914 rt->mf6c_byte_cnt = 0;
915 rt->mf6c_wrong_if = 0;
923 /* no upcall, so make a new entry */
924 rt = (struct mf6c *)malloc(sizeof(*rt), M_MRTABLE6,
931 /* insert new entry at head of hash chain */
932 rt->mf6c_origin = mfccp->mf6cc_origin;
933 rt->mf6c_mcastgrp = mfccp->mf6cc_mcastgrp;
934 rt->mf6c_parent = mfccp->mf6cc_parent;
935 rt->mf6c_ifset = mfccp->mf6cc_ifset;
936 /* initialize pkt counters per src-grp */
937 rt->mf6c_pkt_cnt = 0;
938 rt->mf6c_byte_cnt = 0;
939 rt->mf6c_wrong_if = 0;
941 rt->mf6c_stall = NULL;
943 /* link into table */
944 rt->mf6c_next = mf6ctable[hash];
945 mf6ctable[hash] = rt;
955 * collect delay statistics on the upcalls
958 collate(struct timeval *t)
968 TV_DELTA(tp, *t, delta);
977 #endif /* UPCALL_TIMING */
980 * Delete an mfc entry
983 del_m6fc(struct mf6cctl *mfccp)
986 char ip6bufo[INET6_ADDRSTRLEN], ip6bufg[INET6_ADDRSTRLEN];
988 struct sockaddr_in6 origin;
989 struct sockaddr_in6 mcastgrp;
994 origin = mfccp->mf6cc_origin;
995 mcastgrp = mfccp->mf6cc_mcastgrp;
996 hash = MF6CHASH(origin.sin6_addr, mcastgrp.sin6_addr);
998 MRT6_DLOG(DEBUG_MFC, "orig %s mcastgrp %s",
999 ip6_sprintf(ip6bufo, &origin.sin6_addr),
1000 ip6_sprintf(ip6bufg, &mcastgrp.sin6_addr));
1004 nptr = &mf6ctable[hash];
1005 while ((rt = *nptr) != NULL) {
1006 if (IN6_ARE_ADDR_EQUAL(&origin.sin6_addr,
1007 &rt->mf6c_origin.sin6_addr) &&
1008 IN6_ARE_ADDR_EQUAL(&mcastgrp.sin6_addr,
1009 &rt->mf6c_mcastgrp.sin6_addr) &&
1010 rt->mf6c_stall == NULL)
1013 nptr = &rt->mf6c_next;
1017 return (EADDRNOTAVAIL);
1020 *nptr = rt->mf6c_next;
1021 free(rt, M_MRTABLE6);
1029 socket_send(struct socket *s, struct mbuf *mm, struct sockaddr_in6 *src)
1033 if (sbappendaddr(&s->so_rcv,
1034 (struct sockaddr *)src,
1035 mm, (struct mbuf *)0) != 0) {
1046 * IPv6 multicast forwarding function. This function assumes that the packet
1047 * pointed to by "ip6" has arrived on (or is about to be sent to) the interface
1048 * pointed to by "ifp", and the packet is to be relayed to other networks
1049 * that have members of the packet's destination IPv6 multicast group.
1051 * The packet is returned unscathed to the caller, unless it is
1052 * erroneous, in which case a non-zero return value tells the caller to
1055 * NOTE: this implementation assumes that m->m_pkthdr.rcvif is NULL iff
1056 * this function is called in the originating context (i.e., not when
1057 * forwarding a packet from other node). ip6_output(), which is currently the
1058 * only function that calls this function is called in the originating context,
1059 * explicitly ensures this condition. It is caller's responsibility to ensure
1060 * that if this function is called from somewhere else in the originating
1061 * context in the future.
1064 X_ip6_mforward(struct ip6_hdr *ip6, struct ifnet *ifp, struct mbuf *m)
1073 char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN];
1074 #ifdef UPCALL_TIMING
1078 #endif /* UPCALL_TIMING */
1080 MRT6_DLOG(DEBUG_FORWARD, "src %s, dst %s, ifindex %d",
1081 ip6_sprintf(ip6bufs, &ip6->ip6_src),
1082 ip6_sprintf(ip6bufd, &ip6->ip6_dst), ifp->if_index);
1085 * Don't forward a packet with Hop limit of zero or one,
1086 * or a packet destined to a local-only group.
1088 if (ip6->ip6_hlim <= 1 || IN6_IS_ADDR_MC_INTFACELOCAL(&ip6->ip6_dst) ||
1089 IN6_IS_ADDR_MC_LINKLOCAL(&ip6->ip6_dst))
1094 * Source address check: do not forward packets with unspecified
1095 * source. It was discussed in July 2000, on ipngwg mailing list.
1096 * This is rather more serious than unicast cases, because some
1097 * MLD packets can be sent with the unspecified source address
1098 * (although such packets must normally set 1 to the hop limit field).
1100 if (IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src)) {
1101 IP6STAT_INC(ip6s_cantforward);
1102 if (V_ip6_log_time + V_ip6_log_interval < time_uptime) {
1103 V_ip6_log_time = time_uptime;
1106 "from %s to %s nxt %d received on %s\n",
1107 ip6_sprintf(ip6bufs, &ip6->ip6_src),
1108 ip6_sprintf(ip6bufd, &ip6->ip6_dst),
1110 if_name(m->m_pkthdr.rcvif));
1118 * Determine forwarding mifs from the forwarding cache table
1120 MF6CFIND(ip6->ip6_src, ip6->ip6_dst, rt);
1121 MRT6STAT_INC(mrt6s_mfc_lookups);
1123 /* Entry exists, so forward if necessary */
1126 return (ip6_mdq(m, ifp, rt));
1130 * If we don't have a route for packet's origin,
1131 * Make a copy of the packet & send message to routing daemon.
1133 MRT6STAT_INC(mrt6s_no_route);
1134 MRT6_DLOG(DEBUG_FORWARD | DEBUG_MFC, "no rte s %s g %s",
1135 ip6_sprintf(ip6bufs, &ip6->ip6_src),
1136 ip6_sprintf(ip6bufd, &ip6->ip6_dst));
1139 * Allocate mbufs early so that we don't do extra work if we
1140 * are just going to fail anyway.
1142 rte = (struct rtdetq *)malloc(sizeof(*rte), M_MRTABLE6, M_NOWAIT);
1147 mb0 = m_copym(m, 0, M_COPYALL, M_NOWAIT);
1149 * Pullup packet header if needed before storing it,
1150 * as other references may modify it in the meantime.
1152 if (mb0 && (!M_WRITABLE(mb0) || mb0->m_len < sizeof(struct ip6_hdr)))
1153 mb0 = m_pullup(mb0, sizeof(struct ip6_hdr));
1155 free(rte, M_MRTABLE6);
1160 /* is there an upcall waiting for this packet? */
1161 hash = MF6CHASH(ip6->ip6_src, ip6->ip6_dst);
1162 for (rt = mf6ctable[hash]; rt; rt = rt->mf6c_next) {
1163 if (IN6_ARE_ADDR_EQUAL(&ip6->ip6_src,
1164 &rt->mf6c_origin.sin6_addr) &&
1165 IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst,
1166 &rt->mf6c_mcastgrp.sin6_addr) && (rt->mf6c_stall != NULL))
1173 struct omrt6msg *oim;
1175 /* no upcall, so make a new entry */
1176 rt = (struct mf6c *)malloc(sizeof(*rt), M_MRTABLE6, M_NOWAIT);
1178 free(rte, M_MRTABLE6);
1184 * Make a copy of the header to send to the user
1187 mm = m_copym(mb0, 0, sizeof(struct ip6_hdr), M_NOWAIT);
1189 free(rte, M_MRTABLE6);
1191 free(rt, M_MRTABLE6);
1197 * Send message to routing daemon
1199 sin6.sin6_addr = ip6->ip6_src;
1204 switch (V_ip6_mrouter_ver) {
1207 oim = mtod(mm, struct omrt6msg *);
1208 oim->im6_msgtype = MRT6MSG_NOCACHE;
1213 im = mtod(mm, struct mrt6msg *);
1214 im->im6_msgtype = MRT6MSG_NOCACHE;
1218 free(rte, M_MRTABLE6);
1220 free(rt, M_MRTABLE6);
1225 MRT6_DLOG(DEBUG_FORWARD, "getting the iif info in the kernel");
1226 for (mifp = mif6table, mifi = 0;
1227 mifi < nummifs && mifp->m6_ifp != ifp; mifp++, mifi++)
1230 switch (V_ip6_mrouter_ver) {
1233 oim->im6_mif = mifi;
1241 if (socket_send(V_ip6_mrouter, mm, &sin6) < 0) {
1242 log(LOG_WARNING, "ip6_mforward: ip6_mrouter "
1243 "socket queue full\n");
1244 MRT6STAT_INC(mrt6s_upq_sockfull);
1245 free(rte, M_MRTABLE6);
1247 free(rt, M_MRTABLE6);
1252 MRT6STAT_INC(mrt6s_upcalls);
1254 /* insert new entry at head of hash chain */
1255 bzero(rt, sizeof(*rt));
1256 rt->mf6c_origin.sin6_family = AF_INET6;
1257 rt->mf6c_origin.sin6_len = sizeof(struct sockaddr_in6);
1258 rt->mf6c_origin.sin6_addr = ip6->ip6_src;
1259 rt->mf6c_mcastgrp.sin6_family = AF_INET6;
1260 rt->mf6c_mcastgrp.sin6_len = sizeof(struct sockaddr_in6);
1261 rt->mf6c_mcastgrp.sin6_addr = ip6->ip6_dst;
1262 rt->mf6c_expire = UPCALL_EXPIRE;
1264 rt->mf6c_parent = MF6C_INCOMPLETE_PARENT;
1266 /* link into table */
1267 rt->mf6c_next = mf6ctable[hash];
1268 mf6ctable[hash] = rt;
1269 /* Add this entry to the end of the queue */
1270 rt->mf6c_stall = rte;
1272 /* determine if q has overflowed */
1276 for (p = &rt->mf6c_stall; *p != NULL; p = &(*p)->next)
1277 if (++npkts > MAX_UPQ6) {
1278 MRT6STAT_INC(mrt6s_upq_ovflw);
1279 free(rte, M_MRTABLE6);
1285 /* Add this entry to the end of the queue */
1292 #ifdef UPCALL_TIMING
1294 #endif /* UPCALL_TIMING */
1302 * Clean up cache entries if upcalls are not serviced
1303 * Call from the Slow Timeout mechanism, every half second.
1306 expire_upcalls(void *unused)
1309 char ip6bufo[INET6_ADDRSTRLEN], ip6bufg[INET6_ADDRSTRLEN];
1312 struct mf6c *mfc, **nptr;
1317 for (i = 0; i < MF6CTBLSIZ; i++) {
1318 if (n6expire[i] == 0)
1320 nptr = &mf6ctable[i];
1321 while ((mfc = *nptr) != NULL) {
1322 rte = mfc->mf6c_stall;
1324 * Skip real cache entries
1325 * Make sure it wasn't marked to not expire (shouldn't happen)
1329 mfc->mf6c_expire != 0 &&
1330 --mfc->mf6c_expire == 0) {
1331 MRT6_DLOG(DEBUG_EXPIRE, "expiring (%s %s)",
1332 ip6_sprintf(ip6bufo, &mfc->mf6c_origin.sin6_addr),
1333 ip6_sprintf(ip6bufg, &mfc->mf6c_mcastgrp.sin6_addr));
1335 * drop all the packets
1336 * free the mbuf with the pkt, if, timing info
1339 struct rtdetq *n = rte->next;
1341 free(rte, M_MRTABLE6);
1343 } while (rte != NULL);
1344 MRT6STAT_INC(mrt6s_cache_cleanups);
1347 *nptr = mfc->mf6c_next;
1348 free(mfc, M_MRTABLE6);
1350 nptr = &mfc->mf6c_next;
1354 callout_reset(&expire_upcalls_ch, EXPIRE_TIMEOUT,
1355 expire_upcalls, NULL);
1359 * Packet forwarding routine once entry in the cache is made
1362 ip6_mdq(struct mbuf *m, struct ifnet *ifp, struct mf6c *rt)
1364 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
1367 int plen = m->m_pkthdr.len;
1368 struct in6_addr src0, dst0; /* copies for local work */
1369 u_int32_t iszone, idzone, oszone, odzone;
1373 * Don't forward if it didn't arrive from the parent mif
1376 mifi = rt->mf6c_parent;
1377 if ((mifi >= nummifs) || (mif6table[mifi].m6_ifp != ifp)) {
1378 /* came in the wrong interface */
1379 MRT6_DLOG(DEBUG_FORWARD,
1380 "wrong if: ifid %d mifi %d mififid %x", ifp->if_index,
1381 mifi, mif6table[mifi].m6_ifp->if_index);
1382 MRT6STAT_INC(mrt6s_wrong_if);
1383 rt->mf6c_wrong_if++;
1385 * If we are doing PIM processing, and we are forwarding
1386 * packets on this interface, send a message to the
1389 /* have to make sure this is a valid mif */
1390 if (mifi < nummifs && mif6table[mifi].m6_ifp)
1391 if (V_pim6 && (m->m_flags & M_LOOP) == 0) {
1393 * Check the M_LOOP flag to avoid an
1394 * unnecessary PIM assert.
1395 * XXX: M_LOOP is an ad-hoc hack...
1397 static struct sockaddr_in6 sin6 =
1398 { sizeof(sin6), AF_INET6 };
1403 struct omrt6msg *oim;
1406 mm = m_copym(m, 0, sizeof(struct ip6_hdr),
1410 mm->m_len < sizeof(struct ip6_hdr)))
1411 mm = m_pullup(mm, sizeof(struct ip6_hdr));
1419 switch (V_ip6_mrouter_ver) {
1422 oim = mtod(mm, struct omrt6msg *);
1423 oim->im6_msgtype = MRT6MSG_WRONGMIF;
1428 im = mtod(mm, struct mrt6msg *);
1429 im->im6_msgtype = MRT6MSG_WRONGMIF;
1437 for (mifp = mif6table, iif = 0;
1438 iif < nummifs && mifp &&
1439 mifp->m6_ifp != ifp;
1443 switch (V_ip6_mrouter_ver) {
1447 sin6.sin6_addr = oim->im6_src;
1452 sin6.sin6_addr = im->im6_src;
1456 MRT6STAT_INC(mrt6s_upcalls);
1458 if (socket_send(V_ip6_mrouter, mm, &sin6) < 0) {
1459 MRT6_DLOG(DEBUG_ANY,
1460 "ip6_mrouter socket queue full");
1461 MRT6STAT_INC(mrt6s_upq_sockfull);
1463 } /* if socket Q full */
1466 } /* if wrong iif */
1468 /* If I sourced this packet, it counts as output, else it was input. */
1469 if (m->m_pkthdr.rcvif == NULL) {
1470 /* XXX: is rcvif really NULL when output?? */
1471 mif6table[mifi].m6_pkt_out++;
1472 mif6table[mifi].m6_bytes_out += plen;
1474 mif6table[mifi].m6_pkt_in++;
1475 mif6table[mifi].m6_bytes_in += plen;
1478 rt->mf6c_byte_cnt += plen;
1481 * For each mif, forward a copy of the packet if there are group
1482 * members downstream on the interface.
1484 src0 = ip6->ip6_src;
1485 dst0 = ip6->ip6_dst;
1486 if ((error = in6_setscope(&src0, ifp, &iszone)) != 0 ||
1487 (error = in6_setscope(&dst0, ifp, &idzone)) != 0) {
1488 IP6STAT_INC(ip6s_badscope);
1491 for (mifp = mif6table, mifi = 0; mifi < nummifs; mifp++, mifi++) {
1492 if (IF_ISSET(mifi, &rt->mf6c_ifset)) {
1494 * check if the outgoing packet is going to break
1496 * XXX For packets through PIM register tunnel
1497 * interface, we believe a routing daemon.
1499 if (!(mif6table[rt->mf6c_parent].m6_flags &
1501 !(mif6table[mifi].m6_flags & MIFF_REGISTER)) {
1502 if (in6_setscope(&src0, mif6table[mifi].m6_ifp,
1504 in6_setscope(&dst0, mif6table[mifi].m6_ifp,
1508 IP6STAT_INC(ip6s_badscope);
1514 mifp->m6_bytes_out += plen;
1515 if (mifp->m6_flags & MIFF_REGISTER)
1516 register_send(ip6, mifp, m);
1518 phyint_send(ip6, mifp, m);
1525 phyint_send(struct ip6_hdr *ip6, struct mif6 *mifp, struct mbuf *m)
1528 char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN];
1530 struct mbuf *mb_copy;
1531 struct ifnet *ifp = mifp->m6_ifp;
1532 int error __unused = 0;
1536 * Make a new reference to the packet; make sure that
1537 * the IPv6 header is actually copied, not just referenced,
1538 * so that ip6_output() only scribbles on the copy.
1540 mb_copy = m_copym(m, 0, M_COPYALL, M_NOWAIT);
1542 (!M_WRITABLE(mb_copy) || mb_copy->m_len < sizeof(struct ip6_hdr)))
1543 mb_copy = m_pullup(mb_copy, sizeof(struct ip6_hdr));
1544 if (mb_copy == NULL) {
1547 /* set MCAST flag to the outgoing packet */
1548 mb_copy->m_flags |= M_MCAST;
1551 * If we sourced the packet, call ip6_output since we may devide
1552 * the packet into fragments when the packet is too big for the
1553 * outgoing interface.
1554 * Otherwise, we can simply send the packet to the interface
1557 if (m->m_pkthdr.rcvif == NULL) {
1558 struct ip6_moptions im6o;
1559 struct epoch_tracker et;
1561 im6o.im6o_multicast_ifp = ifp;
1562 /* XXX: ip6_output will override ip6->ip6_hlim */
1563 im6o.im6o_multicast_hlim = ip6->ip6_hlim;
1564 im6o.im6o_multicast_loop = 1;
1565 NET_EPOCH_ENTER(et);
1566 error = ip6_output(mb_copy, NULL, NULL, IPV6_FORWARDING, &im6o,
1570 MRT6_DLOG(DEBUG_XMIT, "mif %u err %d",
1571 (uint16_t)(mifp - mif6table), error);
1576 * If configured to loop back multicasts by default,
1577 * loop back a copy now.
1580 ip6_mloopback(ifp, m);
1583 * Put the packet into the sending queue of the outgoing interface
1584 * if it would fit in the MTU of the interface.
1586 linkmtu = IN6_LINKMTU(ifp);
1587 if (mb_copy->m_pkthdr.len <= linkmtu || linkmtu < IPV6_MMTU) {
1588 struct sockaddr_in6 dst6;
1590 bzero(&dst6, sizeof(dst6));
1591 dst6.sin6_len = sizeof(struct sockaddr_in6);
1592 dst6.sin6_family = AF_INET6;
1593 dst6.sin6_addr = ip6->ip6_dst;
1595 IP_PROBE(send, NULL, NULL, ip6, ifp, NULL, ip6);
1597 * We just call if_output instead of nd6_output here, since
1598 * we need no ND for a multicast forwarded packet...right?
1600 m_clrprotoflags(m); /* Avoid confusing lower layers. */
1601 error = (*ifp->if_output)(ifp, mb_copy,
1602 (struct sockaddr *)&dst6, NULL);
1603 MRT6_DLOG(DEBUG_XMIT, "mif %u err %d",
1604 (uint16_t)(mifp - mif6table), error);
1607 * pMTU discovery is intentionally disabled by default, since
1608 * various router may notify pMTU in multicast, which can be
1609 * a DDoS to a router
1611 if (V_ip6_mcast_pmtu)
1612 icmp6_error(mb_copy, ICMP6_PACKET_TOO_BIG, 0, linkmtu);
1614 MRT6_DLOG(DEBUG_XMIT, " packet too big on %s o %s "
1615 "g %s size %d (discarded)", if_name(ifp),
1616 ip6_sprintf(ip6bufs, &ip6->ip6_src),
1617 ip6_sprintf(ip6bufd, &ip6->ip6_dst),
1618 mb_copy->m_pkthdr.len);
1619 m_freem(mb_copy); /* simply discard the packet */
1625 register_send(struct ip6_hdr *ip6, struct mif6 *mif, struct mbuf *m)
1628 char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN];
1631 int i, len = m->m_pkthdr.len;
1632 static struct sockaddr_in6 sin6 = { sizeof(sin6), AF_INET6 };
1633 struct mrt6msg *im6;
1635 MRT6_DLOG(DEBUG_ANY, "src %s dst %s",
1636 ip6_sprintf(ip6bufs, &ip6->ip6_src),
1637 ip6_sprintf(ip6bufd, &ip6->ip6_dst));
1638 PIM6STAT_INC(pim6s_snd_registers);
1640 /* Make a copy of the packet to send to the user level process. */
1641 mm = m_gethdr(M_NOWAIT, MT_DATA);
1644 mm->m_data += max_linkhdr;
1645 mm->m_len = sizeof(struct ip6_hdr);
1647 if ((mm->m_next = m_copym(m, 0, M_COPYALL, M_NOWAIT)) == NULL) {
1651 i = MHLEN - M_LEADINGSPACE(mm);
1654 mm = m_pullup(mm, i);
1657 /* TODO: check it! */
1658 mm->m_pkthdr.len = len + sizeof(struct ip6_hdr);
1661 * Send message to routing daemon
1663 sin6.sin6_addr = ip6->ip6_src;
1665 im6 = mtod(mm, struct mrt6msg *);
1666 im6->im6_msgtype = MRT6MSG_WHOLEPKT;
1669 im6->im6_mif = mif - mif6table;
1671 /* iif info is not given for reg. encap.n */
1672 MRT6STAT_INC(mrt6s_upcalls);
1674 if (socket_send(V_ip6_mrouter, mm, &sin6) < 0) {
1675 MRT6_DLOG(DEBUG_ANY, "ip6_mrouter socket queue full");
1676 MRT6STAT_INC(mrt6s_upq_sockfull);
1683 * pim6_encapcheck() is called by the encap6_input() path at runtime to
1684 * determine if a packet is for PIM; allowing PIM to be dynamically loaded
1688 pim6_encapcheck(const struct mbuf *m __unused, int off __unused,
1689 int proto __unused, void *arg __unused)
1692 KASSERT(proto == IPPROTO_PIM, ("not for IPPROTO_PIM"));
1693 return (8); /* claim the datagram. */
1697 * PIM sparse mode hook
1698 * Receives the pim control messages, and passes them up to the listening
1699 * socket, using rip6_input.
1700 * The only message processed is the REGISTER pim message; the pim header
1701 * is stripped off, and the inner packet is passed to register_mforward.
1704 pim6_input(struct mbuf *m, int off, int proto, void *arg __unused)
1706 struct pim *pim; /* pointer to a pim struct */
1707 struct ip6_hdr *ip6;
1711 PIM6STAT_INC(pim6s_rcv_total);
1716 pimlen = m->m_pkthdr.len - off;
1717 if (pimlen < PIM_MINLEN) {
1718 PIM6STAT_INC(pim6s_rcv_tooshort);
1719 MRT6_DLOG(DEBUG_PIM, "PIM packet too short");
1721 return (IPPROTO_DONE);
1725 * if the packet is at least as big as a REGISTER, go ahead
1726 * and grab the PIM REGISTER header size, to avoid another
1727 * possible m_pullup() later.
1729 * PIM_MINLEN == pimhdr + u_int32 == 8
1730 * PIM6_REG_MINLEN == pimhdr + reghdr + eip6hdr == 4 + 4 + 40
1732 minlen = (pimlen >= PIM6_REG_MINLEN) ? PIM6_REG_MINLEN : PIM_MINLEN;
1735 * Make sure that the IP6 and PIM headers in contiguous memory, and
1736 * possibly the PIM REGISTER header
1738 if (m->m_len < off + minlen) {
1739 m = m_pullup(m, off + minlen);
1741 IP6STAT_INC(ip6s_exthdrtoolong);
1742 return (IPPROTO_DONE);
1745 ip6 = mtod(m, struct ip6_hdr *);
1746 pim = (struct pim *)((caddr_t)ip6 + off);
1748 #define PIM6_CHECKSUM
1749 #ifdef PIM6_CHECKSUM
1754 * Validate checksum.
1755 * If PIM REGISTER, exclude the data packet
1757 if (pim->pim_type == PIM_REGISTER)
1758 cksumlen = PIM_MINLEN;
1762 if (in6_cksum(m, IPPROTO_PIM, off, cksumlen)) {
1763 PIM6STAT_INC(pim6s_rcv_badsum);
1764 MRT6_DLOG(DEBUG_PIM, "invalid checksum");
1766 return (IPPROTO_DONE);
1769 #endif /* PIM_CHECKSUM */
1771 /* PIM version check */
1772 if (pim->pim_ver != PIM_VERSION) {
1773 PIM6STAT_INC(pim6s_rcv_badversion);
1774 MRT6_DLOG(DEBUG_ANY | DEBUG_ERR,
1775 "incorrect version %d, expecting %d",
1776 pim->pim_ver, PIM_VERSION);
1778 return (IPPROTO_DONE);
1781 if (pim->pim_type == PIM_REGISTER) {
1783 * since this is a REGISTER, we'll make a copy of the register
1784 * headers ip6+pim+u_int32_t+encap_ip6, to be passed up to the
1787 static struct sockaddr_in6 dst = { sizeof(dst), AF_INET6 };
1790 struct ip6_hdr *eip6;
1793 char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN];
1796 PIM6STAT_INC(pim6s_rcv_registers);
1798 if ((reg_mif_num >= nummifs) || (reg_mif_num == (mifi_t) -1)) {
1799 MRT6_DLOG(DEBUG_PIM, "register mif not set: %d",
1802 return (IPPROTO_DONE);
1805 reghdr = (u_int32_t *)(pim + 1);
1807 if ((ntohl(*reghdr) & PIM_NULL_REGISTER))
1808 goto pim6_input_to_daemon;
1813 if (pimlen < PIM6_REG_MINLEN) {
1814 PIM6STAT_INC(pim6s_rcv_tooshort);
1815 PIM6STAT_INC(pim6s_rcv_badregisters);
1816 MRT6_DLOG(DEBUG_ANY | DEBUG_ERR, "register packet "
1817 "size too small %d from %s",
1818 pimlen, ip6_sprintf(ip6bufs, &ip6->ip6_src));
1820 return (IPPROTO_DONE);
1823 eip6 = (struct ip6_hdr *) (reghdr + 1);
1824 MRT6_DLOG(DEBUG_PIM, "eip6: %s -> %s, eip6 plen %d",
1825 ip6_sprintf(ip6bufs, &eip6->ip6_src),
1826 ip6_sprintf(ip6bufd, &eip6->ip6_dst),
1827 ntohs(eip6->ip6_plen));
1829 /* verify the version number of the inner packet */
1830 if ((eip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) {
1831 PIM6STAT_INC(pim6s_rcv_badregisters);
1832 MRT6_DLOG(DEBUG_ANY, "invalid IP version (%d) "
1833 "of the inner packet",
1834 (eip6->ip6_vfc & IPV6_VERSION));
1836 return (IPPROTO_DONE);
1839 /* verify the inner packet is destined to a mcast group */
1840 if (!IN6_IS_ADDR_MULTICAST(&eip6->ip6_dst)) {
1841 PIM6STAT_INC(pim6s_rcv_badregisters);
1842 MRT6_DLOG(DEBUG_PIM, "inner packet of register "
1843 "is not multicast %s",
1844 ip6_sprintf(ip6bufd, &eip6->ip6_dst));
1846 return (IPPROTO_DONE);
1850 * make a copy of the whole header to pass to the daemon later.
1852 mcp = m_copym(m, 0, off + PIM6_REG_MINLEN, M_NOWAIT);
1854 MRT6_DLOG(DEBUG_ANY | DEBUG_ERR, "pim register: "
1855 "could not copy register head");
1857 return (IPPROTO_DONE);
1861 * forward the inner ip6 packet; point m_data at the inner ip6.
1863 m_adj(m, off + PIM_MINLEN);
1864 MRT6_DLOG(DEBUG_PIM, "forwarding decapsulated register: "
1865 "src %s, dst %s, mif %d",
1866 ip6_sprintf(ip6bufs, &eip6->ip6_src),
1867 ip6_sprintf(ip6bufd, &eip6->ip6_dst), reg_mif_num);
1869 if_simloop(mif6table[reg_mif_num].m6_ifp, m,
1870 dst.sin6_family, 0);
1872 /* prepare the register head to send to the mrouting daemon */
1877 * Pass the PIM message up to the daemon; if it is a register message
1878 * pass the 'head' only up to the daemon. This includes the
1879 * encapsulator ip6 header, pim header, register header and the
1880 * encapsulated ip6 header.
1882 pim6_input_to_daemon:
1883 return (rip6_input(&m, &off, proto));
1887 ip6_mroute_modevent(module_t mod, int type, void *unused)
1892 MROUTER6_LOCK_INIT();
1896 pim6_encap_cookie = ip6_encap_attach(&ipv6_encap_cfg,
1898 if (pim6_encap_cookie == NULL) {
1899 printf("ip6_mroute: unable to attach pim6 encap\n");
1900 MIF6_LOCK_DESTROY();
1901 MFC6_LOCK_DESTROY();
1902 MROUTER6_LOCK_DESTROY();
1906 ip6_mforward = X_ip6_mforward;
1907 ip6_mrouter_done = X_ip6_mrouter_done;
1908 ip6_mrouter_get = X_ip6_mrouter_get;
1909 ip6_mrouter_set = X_ip6_mrouter_set;
1910 mrt6_ioctl = X_mrt6_ioctl;
1914 if (V_ip6_mrouter != NULL)
1917 if (pim6_encap_cookie) {
1918 ip6_encap_detach(pim6_encap_cookie);
1919 pim6_encap_cookie = NULL;
1921 X_ip6_mrouter_done();
1922 ip6_mforward = NULL;
1923 ip6_mrouter_done = NULL;
1924 ip6_mrouter_get = NULL;
1925 ip6_mrouter_set = NULL;
1928 MIF6_LOCK_DESTROY();
1929 MFC6_LOCK_DESTROY();
1930 MROUTER6_LOCK_DESTROY();
1934 return (EOPNOTSUPP);
1940 static moduledata_t ip6_mroutemod = {
1942 ip6_mroute_modevent,
1946 DECLARE_MODULE(ip6_mroute, ip6_mroutemod, SI_SUB_PROTO_MC, SI_ORDER_ANY);