2 /* $KAME: mld6.c,v 1.27 2001/04/04 05:17:30 itojun Exp $ */
5 * Copyright (C) 1998 WIDE Project.
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in the
15 * documentation and/or other materials provided with the distribution.
16 * 3. Neither the name of the project nor the names of its contributors
17 * may be used to endorse or promote products derived from this software
18 * without specific prior written permission.
20 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * Copyright (c) 1988 Stephen Deering.
35 * Copyright (c) 1992, 1993
36 * The Regents of the University of California. All rights reserved.
38 * This code is derived from software contributed to Berkeley by
39 * Stephen Deering of Stanford University.
41 * Redistribution and use in source and binary forms, with or without
42 * modification, are permitted provided that the following conditions
44 * 1. Redistributions of source code must retain the above copyright
45 * notice, this list of conditions and the following disclaimer.
46 * 2. Redistributions in binary form must reproduce the above copyright
47 * notice, this list of conditions and the following disclaimer in the
48 * documentation and/or other materials provided with the distribution.
49 * 4. Neither the name of the University nor the names of its contributors
50 * may be used to endorse or promote products derived from this software
51 * without specific prior written permission.
53 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
54 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
55 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
56 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
57 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
58 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
59 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
60 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
61 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
62 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
65 * @(#)igmp.c 8.1 (Berkeley) 7/19/93
69 #include "opt_inet6.h"
71 #include <sys/param.h>
72 #include <sys/systm.h>
74 #include <sys/socket.h>
75 #include <sys/protosw.h>
76 #include <sys/syslog.h>
77 #include <sys/kernel.h>
78 #include <sys/callout.h>
79 #include <sys/malloc.h>
83 #include <netinet/in.h>
84 #include <netinet/in_var.h>
85 #include <netinet6/in6_var.h>
86 #include <netinet/ip6.h>
87 #include <netinet6/ip6_var.h>
88 #include <netinet6/scope6_var.h>
89 #include <netinet/icmp6.h>
90 #include <netinet6/mld6_var.h>
96 /* denotes that the MLD max response delay field specifies time in milliseconds */
97 #define MLD_TIMER_SCALE 1000
99 * time between repetitions of a node's initial report of interest in a
100 * multicast address(in seconds)
102 #define MLD_UNSOLICITED_REPORT_INTERVAL 10
104 static struct ip6_pktopts ip6_opts;
106 static void mld6_sendpkt(struct in6_multi *, int, const struct in6_addr *);
107 static void mld_starttimer(struct in6_multi *);
108 static void mld_stoptimer(struct in6_multi *);
109 static void mld_timeo(struct in6_multi *);
110 static u_long mld_timerresid(struct in6_multi *);
115 static u_int8_t hbh_buf[8];
116 struct ip6_hbh *hbh = (struct ip6_hbh *)hbh_buf;
117 u_int16_t rtalert_code = htons((u_int16_t)IP6OPT_RTALERT_MLD);
119 /* ip6h_nxt will be fill in later */
120 hbh->ip6h_len = 0; /* (8 >> 3) - 1 */
122 /* XXX: grotty hard coding... */
123 hbh_buf[2] = IP6OPT_PADN; /* 2 byte padding */
125 hbh_buf[4] = IP6OPT_ROUTER_ALERT;
126 hbh_buf[5] = IP6OPT_RTALERT_LEN - 2;
127 bcopy((caddr_t)&rtalert_code, &hbh_buf[6], sizeof(u_int16_t));
129 ip6_initpktopts(&ip6_opts);
130 ip6_opts.ip6po_hbh = hbh;
134 mld_starttimer(struct in6_multi *in6m)
139 in6m->in6m_timer_expire.tv_sec = now.tv_sec + in6m->in6m_timer / hz;
140 in6m->in6m_timer_expire.tv_usec = now.tv_usec +
141 (in6m->in6m_timer % hz) * (1000000 / hz);
142 if (in6m->in6m_timer_expire.tv_usec > 1000000) {
143 in6m->in6m_timer_expire.tv_sec++;
144 in6m->in6m_timer_expire.tv_usec -= 1000000;
147 /* start or restart the timer */
148 callout_reset(in6m->in6m_timer_ch, in6m->in6m_timer,
149 (void (*) __P((void *)))mld_timeo, in6m);
153 mld_stoptimer(struct in6_multi *in6m)
155 if (in6m->in6m_timer == IN6M_TIMER_UNDEF)
158 callout_stop(in6m->in6m_timer_ch);
159 in6m->in6m_timer = IN6M_TIMER_UNDEF;
163 mld_timeo(struct in6_multi *in6m)
167 in6m->in6m_timer = IN6M_TIMER_UNDEF;
169 callout_stop(in6m->in6m_timer_ch);
171 switch (in6m->in6m_state) {
172 case MLD_REPORTPENDING:
173 mld6_start_listening(in6m);
176 mld6_sendpkt(in6m, MLD_LISTENER_REPORT, NULL);
184 mld_timerresid(struct in6_multi *in6m)
186 struct timeval now, diff;
190 if (now.tv_sec > in6m->in6m_timer_expire.tv_sec ||
191 (now.tv_sec == in6m->in6m_timer_expire.tv_sec &&
192 now.tv_usec > in6m->in6m_timer_expire.tv_usec)) {
195 diff = in6m->in6m_timer_expire;
196 diff.tv_sec -= now.tv_sec;
197 diff.tv_usec -= now.tv_usec;
198 if (diff.tv_usec < 0) {
200 diff.tv_usec += 1000000;
203 /* return the remaining time in milliseconds */
204 return (((u_long)(diff.tv_sec * 1000000 + diff.tv_usec)) / 1000);
208 mld6_start_listening(struct in6_multi *in6m)
210 struct in6_addr all_in6;
215 * The node never sends a Report or Done for the link-scope all-nodes
217 * MLD messages are never sent for multicast addresses whose scope is 0
218 * (reserved) or 1 (node-local).
220 all_in6 = in6addr_linklocal_allnodes;
221 if (in6_setscope(&all_in6, in6m->in6m_ifp, NULL)) {
222 /* XXX: this should not happen! */
223 in6m->in6m_timer = 0;
224 in6m->in6m_state = MLD_OTHERLISTENER;
226 if (IN6_ARE_ADDR_EQUAL(&in6m->in6m_addr, &all_in6) ||
227 IPV6_ADDR_MC_SCOPE(&in6m->in6m_addr) <
228 IPV6_ADDR_SCOPE_LINKLOCAL) {
229 in6m->in6m_timer = 0;
230 in6m->in6m_state = MLD_OTHERLISTENER;
232 mld6_sendpkt(in6m, MLD_LISTENER_REPORT, NULL);
233 in6m->in6m_timer = arc4random() %
234 MLD_UNSOLICITED_REPORT_INTERVAL * hz;
235 in6m->in6m_state = MLD_IREPORTEDLAST;
237 mld_starttimer(in6m);
243 mld6_stop_listening(struct in6_multi *in6m)
245 struct in6_addr allnode, allrouter;
247 allnode = in6addr_linklocal_allnodes;
248 if (in6_setscope(&allnode, in6m->in6m_ifp, NULL)) {
249 /* XXX: this should not happen! */
252 allrouter = in6addr_linklocal_allrouters;
253 if (in6_setscope(&allrouter, in6m->in6m_ifp, NULL)) {
257 if (in6m->in6m_state == MLD_IREPORTEDLAST &&
258 !IN6_ARE_ADDR_EQUAL(&in6m->in6m_addr, &allnode) &&
259 IPV6_ADDR_MC_SCOPE(&in6m->in6m_addr) >
260 IPV6_ADDR_SCOPE_INTFACELOCAL) {
261 mld6_sendpkt(in6m, MLD_LISTENER_DONE, &allrouter);
266 mld6_input(struct mbuf *m, int off)
268 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
269 struct mld_hdr *mldh;
270 struct ifnet *ifp = m->m_pkthdr.rcvif;
271 struct in6_multi *in6m;
272 struct in6_addr mld_addr, all_in6;
273 struct in6_ifaddr *ia;
274 struct ifmultiaddr *ifma;
275 int timer; /* timer value in the MLD query header */
277 #ifndef PULLDOWN_TEST
278 IP6_EXTHDR_CHECK(m, off, sizeof(*mldh),);
279 mldh = (struct mld_hdr *)(mtod(m, caddr_t) + off);
281 IP6_EXTHDR_GET(mldh, struct mld_hdr *, m, off, sizeof(*mldh));
283 icmp6stat.icp6s_tooshort++;
288 /* source address validation */
289 ip6 = mtod(m, struct ip6_hdr *); /* in case mpullup */
290 if (!IN6_IS_ADDR_LINKLOCAL(&ip6->ip6_src)) {
291 char ip6bufs[INET6_ADDRSTRLEN], ip6bufg[INET6_ADDRSTRLEN];
293 "mld6_input: src %s is not link-local (grp=%s)\n",
294 ip6_sprintf(ip6bufs, &ip6->ip6_src),
295 ip6_sprintf(ip6bufg, &mldh->mld_addr));
297 * spec (RFC2710) does not explicitly
298 * specify to discard the packet from a non link-local
299 * source address. But we believe it's expected to do so.
300 * XXX: do we have to allow :: as source?
307 * make a copy for local work (in6_setscope() may modify the 1st arg)
309 mld_addr = mldh->mld_addr;
310 if (in6_setscope(&mld_addr, ifp, NULL)) {
311 /* XXX: this should not happen! */
317 * In the MLD6 specification, there are 3 states and a flag.
319 * In Non-Listener state, we simply don't have a membership record.
320 * In Delaying Listener state, our timer is running (in6m->in6m_timer)
321 * In Idle Listener state, our timer is not running
322 * (in6m->in6m_timer==IN6M_TIMER_UNDEF)
324 * The flag is in6m->in6m_state, it is set to MLD_OTHERLISTENER if
325 * we have heard a report from another member, or MLD_IREPORTEDLAST
326 * if we sent the last report.
328 switch(mldh->mld_type) {
329 case MLD_LISTENER_QUERY:
330 if (ifp->if_flags & IFF_LOOPBACK)
333 if (!IN6_IS_ADDR_UNSPECIFIED(&mld_addr) &&
334 !IN6_IS_ADDR_MULTICAST(&mld_addr))
335 break; /* print error or log stat? */
337 all_in6 = in6addr_linklocal_allnodes;
338 if (in6_setscope(&all_in6, ifp, NULL)) {
339 /* XXX: this should not happen! */
344 * - Start the timers in all of our membership records
345 * that the query applies to for the interface on
346 * which the query arrived excl. those that belong
347 * to the "all-nodes" group (ff02::1).
348 * - Restart any timer that is already running but has
349 * A value longer than the requested timeout.
350 * - Use the value specified in the query message as
351 * the maximum timeout.
353 timer = ntohs(mldh->mld_maxdelay);
360 * XXX: System timer resolution is too low to handle Max
361 * Response Delay, so set 1 to the internal timer even if
362 * the calculated value equals to zero when Max Response
365 timer = ntohs(mldh->mld_maxdelay) * PR_FASTHZ / MLD_TIMER_SCALE;
366 if (timer == 0 && mldh->mld_maxdelay)
370 TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) {
371 if (ifma->ifma_addr->sa_family != AF_INET6)
373 in6m = (struct in6_multi *)ifma->ifma_protospec;
375 if (IN6_ARE_ADDR_EQUAL(&in6m->in6m_addr, &all_in6) ||
376 IPV6_ADDR_MC_SCOPE(&in6m->in6m_addr) <
377 IPV6_ADDR_SCOPE_LINKLOCAL)
380 if (IN6_IS_ADDR_UNSPECIFIED(&mld_addr) ||
381 IN6_ARE_ADDR_EQUAL(&mld_addr, &in6m->in6m_addr)) {
383 /* send a report immediately */
385 mld6_sendpkt(in6m, MLD_LISTENER_REPORT,
387 in6m->in6m_timer = 0; /* reset timer */
388 in6m->in6m_state = MLD_IREPORTEDLAST;
390 else if (in6m->in6m_timer == IN6M_TIMER_UNDEF ||
391 mld_timerresid(in6m) > (u_long)timer) {
392 in6m->in6m_timer = arc4random() %
393 (int)((long)(timer * hz) / 1000);
394 mld_starttimer(in6m);
401 case MLD_LISTENER_REPORT:
403 * For fast leave to work, we have to know that we are the
404 * last person to send a report for this group. Reports
405 * can potentially get looped back if we are a multicast
406 * router, so discard reports sourced by me.
407 * Note that it is impossible to check IFF_LOOPBACK flag of
408 * ifp for this purpose, since ip6_mloopback pass the physical
409 * interface to looutput.
411 if (m->m_flags & M_LOOP) /* XXX: grotty flag, but efficient */
414 if (!IN6_IS_ADDR_MULTICAST(&mld_addr))
418 * If we belong to the group being reported, stop
419 * our timer for that group.
421 IN6_LOOKUP_MULTI(mld_addr, ifp, in6m);
423 in6m->in6m_timer = 0; /* transit to idle state */
424 in6m->in6m_state = MLD_OTHERLISTENER; /* clear flag */
427 default: /* this is impossible */
428 log(LOG_ERR, "mld6_input: illegal type(%d)", mldh->mld_type);
436 mld6_sendpkt(struct in6_multi *in6m, int type, const struct in6_addr *dst)
438 struct mbuf *mh, *md;
439 struct mld_hdr *mldh;
441 struct ip6_moptions im6o;
442 struct in6_ifaddr *ia;
443 struct ifnet *ifp = in6m->in6m_ifp;
444 struct ifnet *outif = NULL;
447 * At first, find a link local address on the outgoing interface
448 * to use as the source address of the MLD packet.
450 if ((ia = in6ifa_ifpforlinklocal(ifp, IN6_IFF_NOTREADY|IN6_IFF_ANYCAST))
455 * Allocate mbufs to store ip6 header and MLD header.
456 * We allocate 2 mbufs and make chain in advance because
457 * it is more convenient when inserting the hop-by-hop option later.
459 MGETHDR(mh, M_DONTWAIT, MT_HEADER);
462 MGET(md, M_DONTWAIT, MT_DATA);
469 mh->m_pkthdr.rcvif = NULL;
470 mh->m_pkthdr.len = sizeof(struct ip6_hdr) + sizeof(struct mld_hdr);
471 mh->m_len = sizeof(struct ip6_hdr);
472 MH_ALIGN(mh, sizeof(struct ip6_hdr));
474 /* fill in the ip6 header */
475 ip6 = mtod(mh, struct ip6_hdr *);
477 ip6->ip6_vfc &= ~IPV6_VERSION_MASK;
478 ip6->ip6_vfc |= IPV6_VERSION;
479 /* ip6_plen will be set later */
480 ip6->ip6_nxt = IPPROTO_ICMPV6;
481 /* ip6_hlim will be set by im6o.im6o_multicast_hlim */
482 ip6->ip6_src = ia->ia_addr.sin6_addr;
483 ip6->ip6_dst = dst ? *dst : in6m->in6m_addr;
485 /* fill in the MLD header */
486 md->m_len = sizeof(struct mld_hdr);
487 mldh = mtod(md, struct mld_hdr *);
488 mldh->mld_type = type;
491 /* XXX: we assume the function will not be called for query messages */
492 mldh->mld_maxdelay = 0;
493 mldh->mld_reserved = 0;
494 mldh->mld_addr = in6m->in6m_addr;
495 in6_clearscope(&mldh->mld_addr); /* XXX */
496 mldh->mld_cksum = in6_cksum(mh, IPPROTO_ICMPV6, sizeof(struct ip6_hdr),
497 sizeof(struct mld_hdr));
499 /* construct multicast option */
500 bzero(&im6o, sizeof(im6o));
501 im6o.im6o_multicast_ifp = ifp;
502 im6o.im6o_multicast_hlim = 1;
505 * Request loopback of the report if we are acting as a multicast
506 * router, so that the process-level routing daemon can hear it.
508 im6o.im6o_multicast_loop = (ip6_mrouter != NULL);
510 /* increment output statictics */
511 icmp6stat.icp6s_outhist[type]++;
513 ip6_output(mh, &ip6_opts, NULL, 0, &im6o, &outif, NULL);
515 icmp6_ifstat_inc(outif, ifs6_out_msg);
517 case MLD_LISTENER_QUERY:
518 icmp6_ifstat_inc(outif, ifs6_out_mldquery);
520 case MLD_LISTENER_REPORT:
521 icmp6_ifstat_inc(outif, ifs6_out_mldreport);
523 case MLD_LISTENER_DONE:
524 icmp6_ifstat_inc(outif, ifs6_out_mlddone);
531 * Add an address to the list of IP6 multicast addresses for a given interface.
532 * Add source addresses to the list also, if upstream router is MLDv2 capable
533 * and the number of source is not 0.
536 in6_addmulti(struct in6_addr *maddr6, struct ifnet *ifp,
537 int *errorp, int delay)
539 struct in6_multi *in6m;
545 /*IN6_MULTI_LOCK();*/
547 IN6_LOOKUP_MULTI(*maddr6, ifp, in6m);
550 * If we already joined this group, just bump the
551 * refcount and return it.
553 KASSERT(in6m->in6m_refcount >= 1,
554 ("%s: bad refcount %d", __func__, in6m->in6m_refcount));
555 ++in6m->in6m_refcount;
557 struct in6_multi *nin6m;
558 struct ifmultiaddr *ifma;
559 struct sockaddr_in6 sa6;
561 bzero(&sa6, sizeof(sa6));
562 sa6.sin6_family = AF_INET6;
563 sa6.sin6_len = sizeof(struct sockaddr_in6);
564 sa6.sin6_addr = *maddr6;
566 *errorp = if_addmulti(ifp, (struct sockaddr *)&sa6, &ifma);
571 * If ifma->ifma_protospec is null, then if_addmulti() created
572 * a new record. Otherwise, bump refcount, and we are done.
574 if (ifma->ifma_protospec != NULL) {
575 in6m = ifma->ifma_protospec;
576 ++in6m->in6m_refcount;
580 nin6m = malloc(sizeof(*nin6m), M_IP6MADDR, M_NOWAIT | M_ZERO);
582 if_delmulti_ifma(ifma);
586 nin6m->in6m_addr = *maddr6;
587 nin6m->in6m_ifp = ifp;
588 nin6m->in6m_refcount = 1;
589 nin6m->in6m_ifma = ifma;
590 ifma->ifma_protospec = nin6m;
592 nin6m->in6m_timer_ch = malloc(sizeof(*nin6m->in6m_timer_ch),
593 M_IP6MADDR, M_NOWAIT);
594 if (nin6m->in6m_timer_ch == NULL) {
595 free(nin6m, M_IP6MADDR);
596 if_delmulti_ifma(ifma);
600 LIST_INSERT_HEAD(&in6_multihead, nin6m, in6m_entry);
602 callout_init(nin6m->in6m_timer_ch, 0);
603 nin6m->in6m_timer = delay;
604 if (nin6m->in6m_timer > 0) {
605 nin6m->in6m_state = MLD_REPORTPENDING;
606 mld_starttimer(nin6m);
609 mld6_start_listening(nin6m);
615 /*IN6_MULTI_UNLOCK();*/
616 IFF_UNLOCKGIANT(ifp);
622 * Delete a multicast address record.
624 * TODO: Locking, as per netinet.
627 in6_delmulti(struct in6_multi *in6m)
629 struct ifmultiaddr *ifma;
631 KASSERT(in6m->in6m_refcount >= 1, ("%s: freeing freed in6m", __func__));
633 if (--in6m->in6m_refcount == 0) {
635 mld6_stop_listening(in6m);
637 ifma = in6m->in6m_ifma;
638 KASSERT(ifma->ifma_protospec == in6m,
639 ("%s: ifma_protospec != in6m", __func__));
640 ifma->ifma_protospec = NULL;
642 LIST_REMOVE(in6m, in6m_entry);
643 free(in6m->in6m_timer_ch, M_IP6MADDR);
644 free(in6m, M_IP6MADDR);
646 if_delmulti_ifma(ifma);