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;
135 struct in6_multi *in6m;
140 in6m->in6m_timer_expire.tv_sec = now.tv_sec + in6m->in6m_timer / hz;
141 in6m->in6m_timer_expire.tv_usec = now.tv_usec +
142 (in6m->in6m_timer % hz) * (1000000 / hz);
143 if (in6m->in6m_timer_expire.tv_usec > 1000000) {
144 in6m->in6m_timer_expire.tv_sec++;
145 in6m->in6m_timer_expire.tv_usec -= 1000000;
148 /* start or restart the timer */
149 callout_reset(in6m->in6m_timer_ch, in6m->in6m_timer,
150 (void (*) __P((void *)))mld_timeo, in6m);
155 struct in6_multi *in6m;
157 if (in6m->in6m_timer == IN6M_TIMER_UNDEF)
160 callout_stop(in6m->in6m_timer_ch);
161 in6m->in6m_timer = IN6M_TIMER_UNDEF;
166 struct in6_multi *in6m;
170 in6m->in6m_timer = IN6M_TIMER_UNDEF;
172 callout_stop(in6m->in6m_timer_ch);
174 switch (in6m->in6m_state) {
175 case MLD_REPORTPENDING:
176 mld6_start_listening(in6m);
179 mld6_sendpkt(in6m, MLD_LISTENER_REPORT, NULL);
188 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 (((u_long)(diff.tv_sec * 1000000 + diff.tv_usec)) / 1000);
212 mld6_start_listening(in6m)
213 struct in6_multi *in6m;
215 struct in6_addr all_in6;
220 * The node never sends a Report or Done for the link-scope all-nodes
222 * MLD messages are never sent for multicast addresses whose scope is 0
223 * (reserved) or 1 (node-local).
225 all_in6 = in6addr_linklocal_allnodes;
226 if (in6_setscope(&all_in6, in6m->in6m_ifp, NULL)) {
227 /* XXX: this should not happen! */
228 in6m->in6m_timer = 0;
229 in6m->in6m_state = MLD_OTHERLISTENER;
231 if (IN6_ARE_ADDR_EQUAL(&in6m->in6m_addr, &all_in6) ||
232 IPV6_ADDR_MC_SCOPE(&in6m->in6m_addr) <
233 IPV6_ADDR_SCOPE_LINKLOCAL) {
234 in6m->in6m_timer = 0;
235 in6m->in6m_state = MLD_OTHERLISTENER;
237 mld6_sendpkt(in6m, MLD_LISTENER_REPORT, NULL);
238 in6m->in6m_timer = arc4random() %
239 MLD_UNSOLICITED_REPORT_INTERVAL * hz;
240 in6m->in6m_state = MLD_IREPORTEDLAST;
242 mld_starttimer(in6m);
248 mld6_stop_listening(in6m)
249 struct in6_multi *in6m;
251 struct in6_addr allnode, allrouter;
253 allnode = in6addr_linklocal_allnodes;
254 if (in6_setscope(&allnode, in6m->in6m_ifp, NULL)) {
255 /* XXX: this should not happen! */
258 allrouter = in6addr_linklocal_allrouters;
259 if (in6_setscope(&allrouter, in6m->in6m_ifp, NULL)) {
263 if (in6m->in6m_state == MLD_IREPORTEDLAST &&
264 !IN6_ARE_ADDR_EQUAL(&in6m->in6m_addr, &allnode) &&
265 IPV6_ADDR_MC_SCOPE(&in6m->in6m_addr) >
266 IPV6_ADDR_SCOPE_INTFACELOCAL) {
267 mld6_sendpkt(in6m, MLD_LISTENER_DONE, &allrouter);
276 struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *);
277 struct mld_hdr *mldh;
278 struct ifnet *ifp = m->m_pkthdr.rcvif;
279 struct in6_multi *in6m;
280 struct in6_addr mld_addr, all_in6;
281 struct in6_ifaddr *ia;
282 struct ifmultiaddr *ifma;
283 int timer; /* timer value in the MLD query header */
285 #ifndef PULLDOWN_TEST
286 IP6_EXTHDR_CHECK(m, off, sizeof(*mldh),);
287 mldh = (struct mld_hdr *)(mtod(m, caddr_t) + off);
289 IP6_EXTHDR_GET(mldh, struct mld_hdr *, m, off, sizeof(*mldh));
291 icmp6stat.icp6s_tooshort++;
296 /* source address validation */
297 ip6 = mtod(m, struct ip6_hdr *); /* in case mpullup */
298 if (!IN6_IS_ADDR_LINKLOCAL(&ip6->ip6_src)) {
299 char ip6bufs[INET6_ADDRSTRLEN], ip6bufg[INET6_ADDRSTRLEN];
301 "mld6_input: src %s is not link-local (grp=%s)\n",
302 ip6_sprintf(ip6bufs, &ip6->ip6_src),
303 ip6_sprintf(ip6bufg, &mldh->mld_addr));
305 * spec (RFC2710) does not explicitly
306 * specify to discard the packet from a non link-local
307 * source address. But we believe it's expected to do so.
308 * XXX: do we have to allow :: as source?
315 * make a copy for local work (in6_setscope() may modify the 1st arg)
317 mld_addr = mldh->mld_addr;
318 if (in6_setscope(&mld_addr, ifp, NULL)) {
319 /* XXX: this should not happen! */
325 * In the MLD6 specification, there are 3 states and a flag.
327 * In Non-Listener state, we simply don't have a membership record.
328 * In Delaying Listener state, our timer is running (in6m->in6m_timer)
329 * In Idle Listener state, our timer is not running
330 * (in6m->in6m_timer==IN6M_TIMER_UNDEF)
332 * The flag is in6m->in6m_state, it is set to MLD_OTHERLISTENER if
333 * we have heard a report from another member, or MLD_IREPORTEDLAST
334 * if we sent the last report.
336 switch(mldh->mld_type) {
337 case MLD_LISTENER_QUERY:
338 if (ifp->if_flags & IFF_LOOPBACK)
341 if (!IN6_IS_ADDR_UNSPECIFIED(&mld_addr) &&
342 !IN6_IS_ADDR_MULTICAST(&mld_addr))
343 break; /* print error or log stat? */
345 all_in6 = in6addr_linklocal_allnodes;
346 if (in6_setscope(&all_in6, ifp, NULL)) {
347 /* XXX: this should not happen! */
352 * - Start the timers in all of our membership records
353 * that the query applies to for the interface on
354 * which the query arrived excl. those that belong
355 * to the "all-nodes" group (ff02::1).
356 * - Restart any timer that is already running but has
357 * A value longer than the requested timeout.
358 * - Use the value specified in the query message as
359 * the maximum timeout.
361 timer = ntohs(mldh->mld_maxdelay);
368 * XXX: System timer resolution is too low to handle Max
369 * Response Delay, so set 1 to the internal timer even if
370 * the calculated value equals to zero when Max Response
373 timer = ntohs(mldh->mld_maxdelay) * PR_FASTHZ / MLD_TIMER_SCALE;
374 if (timer == 0 && mldh->mld_maxdelay)
378 TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) {
379 if (ifma->ifma_addr->sa_family != AF_INET6)
381 in6m = (struct in6_multi *)ifma->ifma_protospec;
383 if (IN6_ARE_ADDR_EQUAL(&in6m->in6m_addr, &all_in6) ||
384 IPV6_ADDR_MC_SCOPE(&in6m->in6m_addr) <
385 IPV6_ADDR_SCOPE_LINKLOCAL)
388 if (IN6_IS_ADDR_UNSPECIFIED(&mld_addr) ||
389 IN6_ARE_ADDR_EQUAL(&mld_addr, &in6m->in6m_addr)) {
391 /* send a report immediately */
393 mld6_sendpkt(in6m, MLD_LISTENER_REPORT,
395 in6m->in6m_timer = 0; /* reset timer */
396 in6m->in6m_state = MLD_IREPORTEDLAST;
398 else if (in6m->in6m_timer == IN6M_TIMER_UNDEF ||
399 mld_timerresid(in6m) > (u_long)timer) {
400 in6m->in6m_timer = arc4random() %
401 (int)((long)(timer * hz) / 1000);
402 mld_starttimer(in6m);
409 case MLD_LISTENER_REPORT:
411 * For fast leave to work, we have to know that we are the
412 * last person to send a report for this group. Reports
413 * can potentially get looped back if we are a multicast
414 * router, so discard reports sourced by me.
415 * Note that it is impossible to check IFF_LOOPBACK flag of
416 * ifp for this purpose, since ip6_mloopback pass the physical
417 * interface to looutput.
419 if (m->m_flags & M_LOOP) /* XXX: grotty flag, but efficient */
422 if (!IN6_IS_ADDR_MULTICAST(&mld_addr))
426 * If we belong to the group being reported, stop
427 * our timer for that group.
429 IN6_LOOKUP_MULTI(mld_addr, ifp, in6m);
431 in6m->in6m_timer = 0; /* transit to idle state */
432 in6m->in6m_state = MLD_OTHERLISTENER; /* clear flag */
435 default: /* this is impossible */
436 log(LOG_ERR, "mld6_input: illegal type(%d)", mldh->mld_type);
444 mld6_sendpkt(in6m, type, dst)
445 struct in6_multi *in6m;
447 const struct in6_addr *dst;
449 struct mbuf *mh, *md;
450 struct mld_hdr *mldh;
452 struct ip6_moptions im6o;
453 struct in6_ifaddr *ia;
454 struct ifnet *ifp = in6m->in6m_ifp;
455 struct ifnet *outif = NULL;
458 * At first, find a link local address on the outgoing interface
459 * to use as the source address of the MLD packet.
461 if ((ia = in6ifa_ifpforlinklocal(ifp, IN6_IFF_NOTREADY|IN6_IFF_ANYCAST))
466 * Allocate mbufs to store ip6 header and MLD header.
467 * We allocate 2 mbufs and make chain in advance because
468 * it is more convenient when inserting the hop-by-hop option later.
470 MGETHDR(mh, M_DONTWAIT, MT_HEADER);
473 MGET(md, M_DONTWAIT, MT_DATA);
480 mh->m_pkthdr.rcvif = NULL;
481 mh->m_pkthdr.len = sizeof(struct ip6_hdr) + sizeof(struct mld_hdr);
482 mh->m_len = sizeof(struct ip6_hdr);
483 MH_ALIGN(mh, sizeof(struct ip6_hdr));
485 /* fill in the ip6 header */
486 ip6 = mtod(mh, struct ip6_hdr *);
488 ip6->ip6_vfc &= ~IPV6_VERSION_MASK;
489 ip6->ip6_vfc |= IPV6_VERSION;
490 /* ip6_plen will be set later */
491 ip6->ip6_nxt = IPPROTO_ICMPV6;
492 /* ip6_hlim will be set by im6o.im6o_multicast_hlim */
493 ip6->ip6_src = ia->ia_addr.sin6_addr;
494 ip6->ip6_dst = dst ? *dst : in6m->in6m_addr;
496 /* fill in the MLD header */
497 md->m_len = sizeof(struct mld_hdr);
498 mldh = mtod(md, struct mld_hdr *);
499 mldh->mld_type = type;
502 /* XXX: we assume the function will not be called for query messages */
503 mldh->mld_maxdelay = 0;
504 mldh->mld_reserved = 0;
505 mldh->mld_addr = in6m->in6m_addr;
506 in6_clearscope(&mldh->mld_addr); /* XXX */
507 mldh->mld_cksum = in6_cksum(mh, IPPROTO_ICMPV6, sizeof(struct ip6_hdr),
508 sizeof(struct mld_hdr));
510 /* construct multicast option */
511 bzero(&im6o, sizeof(im6o));
512 im6o.im6o_multicast_ifp = ifp;
513 im6o.im6o_multicast_hlim = 1;
516 * Request loopback of the report if we are acting as a multicast
517 * router, so that the process-level routing daemon can hear it.
519 im6o.im6o_multicast_loop = (ip6_mrouter != NULL);
521 /* increment output statictics */
522 icmp6stat.icp6s_outhist[type]++;
524 ip6_output(mh, &ip6_opts, NULL, 0, &im6o, &outif, NULL);
526 icmp6_ifstat_inc(outif, ifs6_out_msg);
528 case MLD_LISTENER_QUERY:
529 icmp6_ifstat_inc(outif, ifs6_out_mldquery);
531 case MLD_LISTENER_REPORT:
532 icmp6_ifstat_inc(outif, ifs6_out_mldreport);
534 case MLD_LISTENER_DONE:
535 icmp6_ifstat_inc(outif, ifs6_out_mlddone);
542 * Add an address to the list of IP6 multicast addresses for a given interface.
543 * Add source addresses to the list also, if upstream router is MLDv2 capable
544 * and the number of source is not 0.
547 in6_addmulti(maddr6, ifp, errorp, delay)
548 struct in6_addr *maddr6;
552 struct in6_multi *in6m;
558 /*IN6_MULTI_LOCK();*/
560 IN6_LOOKUP_MULTI(*maddr6, ifp, in6m);
563 * If we already joined this group, just bump the
564 * refcount and return it.
566 KASSERT(in6m->in6m_refcount >= 1,
567 ("%s: bad refcount %d", __func__, in6m->in6m_refcount));
568 ++in6m->in6m_refcount;
570 struct in6_multi *nin6m;
571 struct ifmultiaddr *ifma;
572 struct sockaddr_in6 sa6;
574 bzero(&sa6, sizeof(sa6));
575 sa6.sin6_family = AF_INET6;
576 sa6.sin6_len = sizeof(struct sockaddr_in6);
577 sa6.sin6_addr = *maddr6;
579 *errorp = if_addmulti(ifp, (struct sockaddr *)&sa6, &ifma);
584 * If ifma->ifma_protospec is null, then if_addmulti() created
585 * a new record. Otherwise, bump refcount, and we are done.
587 if (ifma->ifma_protospec != NULL) {
588 in6m = ifma->ifma_protospec;
589 ++in6m->in6m_refcount;
593 nin6m = malloc(sizeof(*nin6m), M_IP6MADDR, M_NOWAIT | M_ZERO);
595 if_delmulti_ifma(ifma);
599 nin6m->in6m_addr = *maddr6;
600 nin6m->in6m_ifp = ifp;
601 nin6m->in6m_refcount = 1;
602 nin6m->in6m_ifma = ifma;
603 ifma->ifma_protospec = nin6m;
605 nin6m->in6m_timer_ch = malloc(sizeof(*nin6m->in6m_timer_ch),
606 M_IP6MADDR, M_NOWAIT);
607 if (nin6m->in6m_timer_ch == NULL) {
608 free(nin6m, M_IP6MADDR);
609 if_delmulti_ifma(ifma);
613 LIST_INSERT_HEAD(&in6_multihead, nin6m, in6m_entry);
615 callout_init(nin6m->in6m_timer_ch, 0);
616 nin6m->in6m_timer = delay;
617 if (nin6m->in6m_timer > 0) {
618 nin6m->in6m_state = MLD_REPORTPENDING;
619 mld_starttimer(nin6m);
622 mld6_start_listening(nin6m);
628 /*IN6_MULTI_UNLOCK();*/
629 IFF_UNLOCKGIANT(ifp);
635 * Delete a multicast address record.
637 * TODO: Locking, as per netinet.
640 in6_delmulti(struct in6_multi *in6m)
642 struct ifmultiaddr *ifma;
644 KASSERT(in6m->in6m_refcount >= 1, ("%s: freeing freed in6m", __func__));
646 if (--in6m->in6m_refcount == 0) {
648 mld6_stop_listening(in6m);
650 ifma = in6m->in6m_ifma;
651 KASSERT(ifma->ifma_protospec == in6m,
652 ("%s: ifma_protospec != in6m", __func__));
653 ifma->ifma_protospec = NULL;
655 LIST_REMOVE(in6m, in6m_entry);
656 free(in6m->in6m_timer_ch, M_IP6MADDR);
657 free(in6m, M_IP6MADDR);
659 if_delmulti_ifma(ifma);