2 * Copyright (C) 1998 WIDE Project.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
13 * 3. Neither the name of the project nor the names of its contributors
14 * may be used to endorse or promote products derived from this software
15 * without specific prior written permission.
17 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * $KAME: mld6.c,v 1.27 2001/04/04 05:17:30 itojun Exp $
33 * Copyright (c) 1988 Stephen Deering.
34 * Copyright (c) 1992, 1993
35 * The Regents of the University of California. All rights reserved.
37 * This code is derived from software contributed to Berkeley by
38 * Stephen Deering of Stanford University.
40 * Redistribution and use in source and binary forms, with or without
41 * modification, are permitted provided that the following conditions
43 * 1. Redistributions of source code must retain the above copyright
44 * notice, this list of conditions and the following disclaimer.
45 * 2. Redistributions in binary form must reproduce the above copyright
46 * notice, this list of conditions and the following disclaimer in the
47 * documentation and/or other materials provided with the distribution.
48 * 4. Neither the name of the University nor the names of its contributors
49 * may be used to endorse or promote products derived from this software
50 * without specific prior written permission.
52 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
53 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
54 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
55 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
56 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
57 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
58 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
59 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
60 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
61 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
64 * @(#)igmp.c 8.1 (Berkeley) 7/19/93
67 #include <sys/cdefs.h>
68 __FBSDID("$FreeBSD$");
71 #include "opt_inet6.h"
73 #include <sys/param.h>
74 #include <sys/systm.h>
76 #include <sys/socket.h>
77 #include <sys/protosw.h>
78 #include <sys/syslog.h>
79 #include <sys/kernel.h>
80 #include <sys/callout.h>
81 #include <sys/malloc.h>
82 #include <sys/vimage.h>
86 #include <netinet/in.h>
87 #include <netinet/in_var.h>
88 #include <netinet6/in6_var.h>
89 #include <netinet/ip6.h>
90 #include <netinet6/ip6_var.h>
91 #include <netinet6/scope6_var.h>
92 #include <netinet/icmp6.h>
93 #include <netinet6/mld6_var.h>
99 /* denotes that the MLD max response delay field specifies time in milliseconds */
100 #define MLD_TIMER_SCALE 1000
102 * time between repetitions of a node's initial report of interest in a
103 * multicast address(in seconds)
105 #define MLD_UNSOLICITED_REPORT_INTERVAL 10
107 static struct ip6_pktopts ip6_opts;
109 static void mld6_sendpkt(struct in6_multi *, int, const struct in6_addr *);
110 static void mld_starttimer(struct in6_multi *);
111 static void mld_stoptimer(struct in6_multi *);
112 static void mld_timeo(struct in6_multi *);
113 static u_long mld_timerresid(struct in6_multi *);
118 INIT_VNET_INET6(curvnet);
119 static u_int8_t hbh_buf[8];
120 struct ip6_hbh *hbh = (struct ip6_hbh *)hbh_buf;
121 u_int16_t rtalert_code = htons((u_int16_t)IP6OPT_RTALERT_MLD);
123 /* ip6h_nxt will be fill in later */
124 hbh->ip6h_len = 0; /* (8 >> 3) - 1 */
126 /* XXX: grotty hard coding... */
127 hbh_buf[2] = IP6OPT_PADN; /* 2 byte padding */
129 hbh_buf[4] = IP6OPT_ROUTER_ALERT;
130 hbh_buf[5] = IP6OPT_RTALERT_LEN - 2;
131 bcopy((caddr_t)&rtalert_code, &hbh_buf[6], sizeof(u_int16_t));
133 ip6_initpktopts(&V_ip6_opts);
134 V_ip6_opts.ip6po_hbh = hbh;
138 mld_starttimer(struct in6_multi *in6m)
143 in6m->in6m_timer_expire.tv_sec = now.tv_sec + in6m->in6m_timer / hz;
144 in6m->in6m_timer_expire.tv_usec = now.tv_usec +
145 (in6m->in6m_timer % hz) * (1000000 / hz);
146 if (in6m->in6m_timer_expire.tv_usec > 1000000) {
147 in6m->in6m_timer_expire.tv_sec++;
148 in6m->in6m_timer_expire.tv_usec -= 1000000;
151 /* start or restart the timer */
152 callout_reset(in6m->in6m_timer_ch, in6m->in6m_timer,
153 (void (*)(void *))mld_timeo, in6m);
157 mld_stoptimer(struct in6_multi *in6m)
159 if (in6m->in6m_timer == IN6M_TIMER_UNDEF)
162 callout_stop(in6m->in6m_timer_ch);
163 in6m->in6m_timer = IN6M_TIMER_UNDEF;
167 mld_timeo(struct in6_multi *in6m)
171 in6m->in6m_timer = IN6M_TIMER_UNDEF;
173 callout_stop(in6m->in6m_timer_ch);
175 switch (in6m->in6m_state) {
176 case MLD_REPORTPENDING:
177 mld6_start_listening(in6m);
180 mld6_sendpkt(in6m, MLD_LISTENER_REPORT, NULL);
188 mld_timerresid(struct in6_multi *in6m)
190 struct timeval now, diff;
194 if (now.tv_sec > in6m->in6m_timer_expire.tv_sec ||
195 (now.tv_sec == in6m->in6m_timer_expire.tv_sec &&
196 now.tv_usec > in6m->in6m_timer_expire.tv_usec)) {
199 diff = in6m->in6m_timer_expire;
200 diff.tv_sec -= now.tv_sec;
201 diff.tv_usec -= now.tv_usec;
202 if (diff.tv_usec < 0) {
204 diff.tv_usec += 1000000;
207 /* return the remaining time in milliseconds */
208 return (diff.tv_sec * 1000 + diff.tv_usec / 1000);
212 mld6_start_listening(struct in6_multi *in6m)
214 struct in6_addr all_in6;
219 * The node never sends a Report or Done for the link-scope all-nodes
221 * MLD messages are never sent for multicast addresses whose scope is 0
222 * (reserved) or 1 (node-local).
224 all_in6 = in6addr_linklocal_allnodes;
225 if (in6_setscope(&all_in6, in6m->in6m_ifp, NULL)) {
226 /* XXX: this should not happen! */
227 in6m->in6m_timer = 0;
228 in6m->in6m_state = MLD_OTHERLISTENER;
230 if (IN6_ARE_ADDR_EQUAL(&in6m->in6m_addr, &all_in6) ||
231 IPV6_ADDR_MC_SCOPE(&in6m->in6m_addr) <
232 IPV6_ADDR_SCOPE_LINKLOCAL) {
233 in6m->in6m_timer = 0;
234 in6m->in6m_state = MLD_OTHERLISTENER;
236 mld6_sendpkt(in6m, MLD_LISTENER_REPORT, NULL);
237 in6m->in6m_timer = arc4random() %
238 MLD_UNSOLICITED_REPORT_INTERVAL * hz;
239 in6m->in6m_state = MLD_IREPORTEDLAST;
241 mld_starttimer(in6m);
247 mld6_stop_listening(struct in6_multi *in6m)
249 struct in6_addr allnode, allrouter;
251 allnode = in6addr_linklocal_allnodes;
252 if (in6_setscope(&allnode, in6m->in6m_ifp, NULL)) {
253 /* XXX: this should not happen! */
256 allrouter = in6addr_linklocal_allrouters;
257 if (in6_setscope(&allrouter, in6m->in6m_ifp, NULL)) {
261 if (in6m->in6m_state == MLD_IREPORTEDLAST &&
262 !IN6_ARE_ADDR_EQUAL(&in6m->in6m_addr, &allnode) &&
263 IPV6_ADDR_MC_SCOPE(&in6m->in6m_addr) >
264 IPV6_ADDR_SCOPE_INTFACELOCAL) {
265 mld6_sendpkt(in6m, MLD_LISTENER_DONE, &allrouter);
270 mld6_input(struct mbuf *m, int off)
272 INIT_VNET_INET6(curvnet);
273 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
274 struct mld_hdr *mldh;
275 struct ifnet *ifp = m->m_pkthdr.rcvif;
276 struct in6_multi *in6m;
277 struct in6_addr mld_addr, all_in6;
278 struct in6_ifaddr *ia;
279 struct ifmultiaddr *ifma;
280 u_long timer; /* timer value in the MLD query header */
282 #ifndef PULLDOWN_TEST
283 IP6_EXTHDR_CHECK(m, off, sizeof(*mldh),);
284 mldh = (struct mld_hdr *)(mtod(m, caddr_t) + off);
286 IP6_EXTHDR_GET(mldh, struct mld_hdr *, m, off, sizeof(*mldh));
288 V_icmp6stat.icp6s_tooshort++;
293 /* source address validation */
294 ip6 = mtod(m, struct ip6_hdr *); /* in case mpullup */
295 if (!IN6_IS_ADDR_LINKLOCAL(&ip6->ip6_src)) {
296 char ip6bufs[INET6_ADDRSTRLEN], ip6bufg[INET6_ADDRSTRLEN];
298 "mld6_input: src %s is not link-local (grp=%s)\n",
299 ip6_sprintf(ip6bufs, &ip6->ip6_src),
300 ip6_sprintf(ip6bufg, &mldh->mld_addr));
302 * spec (RFC2710) does not explicitly
303 * specify to discard the packet from a non link-local
304 * source address. But we believe it's expected to do so.
305 * XXX: do we have to allow :: as source?
312 * make a copy for local work (in6_setscope() may modify the 1st arg)
314 mld_addr = mldh->mld_addr;
315 if (in6_setscope(&mld_addr, ifp, NULL)) {
316 /* XXX: this should not happen! */
322 * In the MLD6 specification, there are 3 states and a flag.
324 * In Non-Listener state, we simply don't have a membership record.
325 * In Delaying Listener state, our timer is running (in6m->in6m_timer)
326 * In Idle Listener state, our timer is not running
327 * (in6m->in6m_timer==IN6M_TIMER_UNDEF)
329 * The flag is in6m->in6m_state, it is set to MLD_OTHERLISTENER if
330 * we have heard a report from another member, or MLD_IREPORTEDLAST
331 * if we sent the last report.
333 switch(mldh->mld_type) {
334 case MLD_LISTENER_QUERY:
335 if (ifp->if_flags & IFF_LOOPBACK)
338 if (!IN6_IS_ADDR_UNSPECIFIED(&mld_addr) &&
339 !IN6_IS_ADDR_MULTICAST(&mld_addr))
340 break; /* print error or log stat? */
342 all_in6 = in6addr_linklocal_allnodes;
343 if (in6_setscope(&all_in6, ifp, NULL)) {
344 /* XXX: this should not happen! */
349 * - Start the timers in all of our membership records
350 * that the query applies to for the interface on
351 * which the query arrived excl. those that belong
352 * to the "all-nodes" group (ff02::1).
353 * - Restart any timer that is already running but has
354 * A value longer than the requested timeout.
355 * - Use the value specified in the query message as
356 * the maximum timeout.
358 timer = ntohs(mldh->mld_maxdelay);
365 * XXX: System timer resolution is too low to handle Max
366 * Response Delay, so set 1 to the internal timer even if
367 * the calculated value equals to zero when Max Response
370 timer = ntohs(mldh->mld_maxdelay) * PR_FASTHZ / MLD_TIMER_SCALE;
371 if (timer == 0 && mldh->mld_maxdelay)
375 TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) {
376 if (ifma->ifma_addr->sa_family != AF_INET6)
378 in6m = (struct in6_multi *)ifma->ifma_protospec;
380 if (IN6_ARE_ADDR_EQUAL(&in6m->in6m_addr, &all_in6) ||
381 IPV6_ADDR_MC_SCOPE(&in6m->in6m_addr) <
382 IPV6_ADDR_SCOPE_LINKLOCAL)
385 if (IN6_IS_ADDR_UNSPECIFIED(&mld_addr) ||
386 IN6_ARE_ADDR_EQUAL(&mld_addr, &in6m->in6m_addr)) {
388 /* send a report immediately */
390 mld6_sendpkt(in6m, MLD_LISTENER_REPORT,
392 in6m->in6m_timer = 0; /* reset timer */
393 in6m->in6m_state = MLD_IREPORTEDLAST;
395 else if (in6m->in6m_timer == IN6M_TIMER_UNDEF ||
396 mld_timerresid(in6m) > timer) {
398 1 + (arc4random() % timer) * hz / 1000;
399 mld_starttimer(in6m);
406 case MLD_LISTENER_REPORT:
408 * For fast leave to work, we have to know that we are the
409 * last person to send a report for this group. Reports
410 * can potentially get looped back if we are a multicast
411 * router, so discard reports sourced by me.
412 * Note that it is impossible to check IFF_LOOPBACK flag of
413 * ifp for this purpose, since ip6_mloopback pass the physical
414 * interface to looutput.
416 if (m->m_flags & M_LOOP) /* XXX: grotty flag, but efficient */
419 if (!IN6_IS_ADDR_MULTICAST(&mld_addr))
423 * If we belong to the group being reported, stop
424 * our timer for that group.
426 IN6_LOOKUP_MULTI(mld_addr, ifp, in6m);
428 in6m->in6m_timer = 0; /* transit to idle state */
429 in6m->in6m_state = MLD_OTHERLISTENER; /* clear flag */
432 default: /* this is impossible */
433 log(LOG_ERR, "mld6_input: illegal type(%d)", mldh->mld_type);
441 mld6_sendpkt(struct in6_multi *in6m, int type, const struct in6_addr *dst)
443 INIT_VNET_INET6(curvnet);
444 struct mbuf *mh, *md;
445 struct mld_hdr *mldh;
447 struct ip6_moptions im6o;
448 struct in6_ifaddr *ia;
449 struct ifnet *ifp = in6m->in6m_ifp;
450 struct ifnet *outif = NULL;
453 * At first, find a link local address on the outgoing interface
454 * to use as the source address of the MLD packet.
456 if ((ia = in6ifa_ifpforlinklocal(ifp, IN6_IFF_NOTREADY|IN6_IFF_ANYCAST))
461 * Allocate mbufs to store ip6 header and MLD header.
462 * We allocate 2 mbufs and make chain in advance because
463 * it is more convenient when inserting the hop-by-hop option later.
465 MGETHDR(mh, M_DONTWAIT, MT_HEADER);
468 MGET(md, M_DONTWAIT, MT_DATA);
475 mh->m_pkthdr.rcvif = NULL;
476 mh->m_pkthdr.len = sizeof(struct ip6_hdr) + sizeof(struct mld_hdr);
477 mh->m_len = sizeof(struct ip6_hdr);
478 MH_ALIGN(mh, sizeof(struct ip6_hdr));
480 /* fill in the ip6 header */
481 ip6 = mtod(mh, struct ip6_hdr *);
483 ip6->ip6_vfc &= ~IPV6_VERSION_MASK;
484 ip6->ip6_vfc |= IPV6_VERSION;
485 /* ip6_plen will be set later */
486 ip6->ip6_nxt = IPPROTO_ICMPV6;
487 /* ip6_hlim will be set by im6o.im6o_multicast_hlim */
488 ip6->ip6_src = ia->ia_addr.sin6_addr;
489 ip6->ip6_dst = dst ? *dst : in6m->in6m_addr;
491 /* fill in the MLD header */
492 md->m_len = sizeof(struct mld_hdr);
493 mldh = mtod(md, struct mld_hdr *);
494 mldh->mld_type = type;
497 /* XXX: we assume the function will not be called for query messages */
498 mldh->mld_maxdelay = 0;
499 mldh->mld_reserved = 0;
500 mldh->mld_addr = in6m->in6m_addr;
501 in6_clearscope(&mldh->mld_addr); /* XXX */
502 mldh->mld_cksum = in6_cksum(mh, IPPROTO_ICMPV6, sizeof(struct ip6_hdr),
503 sizeof(struct mld_hdr));
505 /* construct multicast option */
506 bzero(&im6o, sizeof(im6o));
507 im6o.im6o_multicast_ifp = ifp;
508 im6o.im6o_multicast_hlim = 1;
511 * Request loopback of the report if we are acting as a multicast
512 * router, so that the process-level routing daemon can hear it.
514 im6o.im6o_multicast_loop = (ip6_mrouter != NULL);
516 /* increment output statictics */
517 V_icmp6stat.icp6s_outhist[type]++;
519 ip6_output(mh, &V_ip6_opts, NULL, 0, &im6o, &outif, NULL);
521 icmp6_ifstat_inc(outif, ifs6_out_msg);
523 case MLD_LISTENER_QUERY:
524 icmp6_ifstat_inc(outif, ifs6_out_mldquery);
526 case MLD_LISTENER_REPORT:
527 icmp6_ifstat_inc(outif, ifs6_out_mldreport);
529 case MLD_LISTENER_DONE:
530 icmp6_ifstat_inc(outif, ifs6_out_mlddone);
537 * Add an address to the list of IP6 multicast addresses for a given interface.
538 * Add source addresses to the list also, if upstream router is MLDv2 capable
539 * and the number of source is not 0.
542 in6_addmulti(struct in6_addr *maddr6, struct ifnet *ifp,
543 int *errorp, int delay)
545 struct in6_multi *in6m;
551 /*IN6_MULTI_LOCK();*/
553 IN6_LOOKUP_MULTI(*maddr6, ifp, in6m);
556 * If we already joined this group, just bump the
557 * refcount and return it.
559 KASSERT(in6m->in6m_refcount >= 1,
560 ("%s: bad refcount %d", __func__, in6m->in6m_refcount));
561 ++in6m->in6m_refcount;
563 struct in6_multi *nin6m;
564 struct ifmultiaddr *ifma;
565 struct sockaddr_in6 sa6;
567 bzero(&sa6, sizeof(sa6));
568 sa6.sin6_family = AF_INET6;
569 sa6.sin6_len = sizeof(struct sockaddr_in6);
570 sa6.sin6_addr = *maddr6;
572 *errorp = if_addmulti(ifp, (struct sockaddr *)&sa6, &ifma);
577 * If ifma->ifma_protospec is null, then if_addmulti() created
578 * a new record. Otherwise, bump refcount, and we are done.
580 if (ifma->ifma_protospec != NULL) {
581 in6m = ifma->ifma_protospec;
582 ++in6m->in6m_refcount;
586 nin6m = malloc(sizeof(*nin6m), M_IP6MADDR, M_NOWAIT | M_ZERO);
588 if_delmulti_ifma(ifma);
592 nin6m->in6m_addr = *maddr6;
593 nin6m->in6m_ifp = ifp;
594 nin6m->in6m_refcount = 1;
595 nin6m->in6m_ifma = ifma;
596 ifma->ifma_protospec = nin6m;
598 nin6m->in6m_timer_ch = malloc(sizeof(*nin6m->in6m_timer_ch),
599 M_IP6MADDR, M_NOWAIT);
600 if (nin6m->in6m_timer_ch == NULL) {
601 free(nin6m, M_IP6MADDR);
602 if_delmulti_ifma(ifma);
606 LIST_INSERT_HEAD(&in6_multihead, nin6m, in6m_entry);
608 callout_init(nin6m->in6m_timer_ch, 0);
609 nin6m->in6m_timer = delay;
610 if (nin6m->in6m_timer > 0) {
611 nin6m->in6m_state = MLD_REPORTPENDING;
612 mld_starttimer(nin6m);
615 mld6_start_listening(nin6m);
621 /*IN6_MULTI_UNLOCK();*/
622 IFF_UNLOCKGIANT(ifp);
628 * Delete a multicast address record.
630 * TODO: Locking, as per netinet.
633 in6_delmulti(struct in6_multi *in6m)
635 struct ifmultiaddr *ifma;
637 KASSERT(in6m->in6m_refcount >= 1, ("%s: freeing freed in6m", __func__));
639 if (--in6m->in6m_refcount == 0) {
641 mld6_stop_listening(in6m);
643 ifma = in6m->in6m_ifma;
644 KASSERT(ifma->ifma_protospec == in6m,
645 ("%s: ifma_protospec != in6m", __func__));
646 ifma->ifma_protospec = NULL;
648 LIST_REMOVE(in6m, in6m_entry);
649 free(in6m->in6m_timer_ch, M_IP6MADDR);
650 free(in6m, M_IP6MADDR);
652 if_delmulti_ifma(ifma);