Index: head/sys/netinet6/icmp6.c =================================================================== --- head/sys/netinet6/icmp6.c (revision 253969) +++ head/sys/netinet6/icmp6.c (revision 253970) @@ -1,2912 +1,2912 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $KAME: icmp6.c,v 1.211 2001/04/04 05:56:20 itojun Exp $ */ /*- * Copyright (c) 1982, 1986, 1988, 1993 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)ip_icmp.c 8.2 (Berkeley) 1/4/94 */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include "opt_ipsec.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef IPSEC #include #include #endif extern struct domain inet6domain; VNET_PCPUSTAT_DEFINE(struct icmp6stat, icmp6stat); VNET_PCPUSTAT_SYSINIT(icmp6stat); #ifdef VIMAGE VNET_PCPUSTAT_SYSUNINIT(icmp6stat); #endif /* VIMAGE */ VNET_DECLARE(struct inpcbinfo, ripcbinfo); VNET_DECLARE(struct inpcbhead, ripcb); VNET_DECLARE(int, icmp6errppslim); static VNET_DEFINE(int, icmp6errpps_count) = 0; static VNET_DEFINE(struct timeval, icmp6errppslim_last); VNET_DECLARE(int, icmp6_nodeinfo); #define V_ripcbinfo VNET(ripcbinfo) #define V_ripcb VNET(ripcb) #define V_icmp6errppslim VNET(icmp6errppslim) #define V_icmp6errpps_count VNET(icmp6errpps_count) #define V_icmp6errppslim_last VNET(icmp6errppslim_last) #define V_icmp6_nodeinfo VNET(icmp6_nodeinfo) static void icmp6_errcount(int, int); static int icmp6_rip6_input(struct mbuf **, int); static int icmp6_ratelimit(const struct in6_addr *, const int, const int); static const char *icmp6_redirect_diag(struct in6_addr *, struct in6_addr *, struct in6_addr *); static struct mbuf *ni6_input(struct mbuf *, int); static struct mbuf *ni6_nametodns(const char *, int, int); static int ni6_dnsmatch(const char *, int, const char *, int); static int ni6_addrs(struct icmp6_nodeinfo *, struct mbuf *, struct ifnet **, struct in6_addr *); static int ni6_store_addrs(struct icmp6_nodeinfo *, struct icmp6_nodeinfo *, struct ifnet *, int); static int icmp6_notify_error(struct mbuf **, int, int, int); /* * Kernel module interface for updating icmp6stat. The argument is an index * into icmp6stat treated as an array of u_quad_t. While this encodes the * general layout of icmp6stat into the caller, it doesn't encode its * location, so that future changes to add, for example, per-CPU stats * support won't cause binary compatibility problems for kernel modules. */ void kmod_icmp6stat_inc(int statnum) { counter_u64_add(VNET(icmp6stat)[statnum], 1); } static void icmp6_errcount(int type, int code) { switch (type) { case ICMP6_DST_UNREACH: switch (code) { case ICMP6_DST_UNREACH_NOROUTE: ICMP6STAT_INC(icp6s_odst_unreach_noroute); return; case ICMP6_DST_UNREACH_ADMIN: ICMP6STAT_INC(icp6s_odst_unreach_admin); return; case ICMP6_DST_UNREACH_BEYONDSCOPE: ICMP6STAT_INC(icp6s_odst_unreach_beyondscope); return; case ICMP6_DST_UNREACH_ADDR: ICMP6STAT_INC(icp6s_odst_unreach_addr); return; case ICMP6_DST_UNREACH_NOPORT: ICMP6STAT_INC(icp6s_odst_unreach_noport); return; } break; case ICMP6_PACKET_TOO_BIG: ICMP6STAT_INC(icp6s_opacket_too_big); return; case ICMP6_TIME_EXCEEDED: switch (code) { case ICMP6_TIME_EXCEED_TRANSIT: ICMP6STAT_INC(icp6s_otime_exceed_transit); return; case ICMP6_TIME_EXCEED_REASSEMBLY: ICMP6STAT_INC(icp6s_otime_exceed_reassembly); return; } break; case ICMP6_PARAM_PROB: switch (code) { case ICMP6_PARAMPROB_HEADER: ICMP6STAT_INC(icp6s_oparamprob_header); return; case ICMP6_PARAMPROB_NEXTHEADER: ICMP6STAT_INC(icp6s_oparamprob_nextheader); return; case ICMP6_PARAMPROB_OPTION: ICMP6STAT_INC(icp6s_oparamprob_option); return; } break; case ND_REDIRECT: ICMP6STAT_INC(icp6s_oredirect); return; } ICMP6STAT_INC(icp6s_ounknown); } /* * A wrapper function for icmp6_error() necessary when the erroneous packet * may not contain enough scope zone information. */ void icmp6_error2(struct mbuf *m, int type, int code, int param, struct ifnet *ifp) { struct ip6_hdr *ip6; if (ifp == NULL) return; #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, 0, sizeof(struct ip6_hdr), ); #else if (m->m_len < sizeof(struct ip6_hdr)) { m = m_pullup(m, sizeof(struct ip6_hdr)); if (m == NULL) return; } #endif ip6 = mtod(m, struct ip6_hdr *); if (in6_setscope(&ip6->ip6_src, ifp, NULL) != 0) return; if (in6_setscope(&ip6->ip6_dst, ifp, NULL) != 0) return; icmp6_error(m, type, code, param); } /* * Generate an error packet of type error in response to bad IP6 packet. */ void icmp6_error(struct mbuf *m, int type, int code, int param) { struct ip6_hdr *oip6, *nip6; struct icmp6_hdr *icmp6; u_int preplen; int off; int nxt; ICMP6STAT_INC(icp6s_error); /* count per-type-code statistics */ icmp6_errcount(type, code); #ifdef M_DECRYPTED /*not openbsd*/ if (m->m_flags & M_DECRYPTED) { ICMP6STAT_INC(icp6s_canterror); goto freeit; } #endif #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, 0, sizeof(struct ip6_hdr), ); #else if (m->m_len < sizeof(struct ip6_hdr)) { m = m_pullup(m, sizeof(struct ip6_hdr)); if (m == NULL) return; } #endif oip6 = mtod(m, struct ip6_hdr *); /* * If the destination address of the erroneous packet is a multicast * address, or the packet was sent using link-layer multicast, * we should basically suppress sending an error (RFC 2463, Section * 2.4). * We have two exceptions (the item e.2 in that section): * - the Packet Too Big message can be sent for path MTU discovery. * - the Parameter Problem Message that can be allowed an icmp6 error * in the option type field. This check has been done in * ip6_unknown_opt(), so we can just check the type and code. */ if ((m->m_flags & (M_BCAST|M_MCAST) || IN6_IS_ADDR_MULTICAST(&oip6->ip6_dst)) && (type != ICMP6_PACKET_TOO_BIG && (type != ICMP6_PARAM_PROB || code != ICMP6_PARAMPROB_OPTION))) goto freeit; /* * RFC 2463, 2.4 (e.5): source address check. * XXX: the case of anycast source? */ if (IN6_IS_ADDR_UNSPECIFIED(&oip6->ip6_src) || IN6_IS_ADDR_MULTICAST(&oip6->ip6_src)) goto freeit; /* * If we are about to send ICMPv6 against ICMPv6 error/redirect, * don't do it. */ nxt = -1; off = ip6_lasthdr(m, 0, IPPROTO_IPV6, &nxt); if (off >= 0 && nxt == IPPROTO_ICMPV6) { struct icmp6_hdr *icp; #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, 0, off + sizeof(struct icmp6_hdr), ); icp = (struct icmp6_hdr *)(mtod(m, caddr_t) + off); #else IP6_EXTHDR_GET(icp, struct icmp6_hdr *, m, off, sizeof(*icp)); if (icp == NULL) { ICMP6STAT_INC(icp6s_tooshort); return; } #endif if (icp->icmp6_type < ICMP6_ECHO_REQUEST || icp->icmp6_type == ND_REDIRECT) { /* * ICMPv6 error * Special case: for redirect (which is * informational) we must not send icmp6 error. */ ICMP6STAT_INC(icp6s_canterror); goto freeit; } else { /* ICMPv6 informational - send the error */ } } else { /* non-ICMPv6 - send the error */ } oip6 = mtod(m, struct ip6_hdr *); /* adjust pointer */ /* Finally, do rate limitation check. */ if (icmp6_ratelimit(&oip6->ip6_src, type, code)) { ICMP6STAT_INC(icp6s_toofreq); goto freeit; } /* * OK, ICMP6 can be generated. */ if (m->m_pkthdr.len >= ICMPV6_PLD_MAXLEN) m_adj(m, ICMPV6_PLD_MAXLEN - m->m_pkthdr.len); preplen = sizeof(struct ip6_hdr) + sizeof(struct icmp6_hdr); M_PREPEND(m, preplen, M_NOWAIT); /* FIB is also copied over. */ if (m && m->m_len < preplen) m = m_pullup(m, preplen); if (m == NULL) { nd6log((LOG_DEBUG, "ENOBUFS in icmp6_error %d\n", __LINE__)); return; } nip6 = mtod(m, struct ip6_hdr *); nip6->ip6_src = oip6->ip6_src; nip6->ip6_dst = oip6->ip6_dst; in6_clearscope(&oip6->ip6_src); in6_clearscope(&oip6->ip6_dst); icmp6 = (struct icmp6_hdr *)(nip6 + 1); icmp6->icmp6_type = type; icmp6->icmp6_code = code; icmp6->icmp6_pptr = htonl((u_int32_t)param); /* * icmp6_reflect() is designed to be in the input path. * icmp6_error() can be called from both input and output path, * and if we are in output path rcvif could contain bogus value. * clear m->m_pkthdr.rcvif for safety, we should have enough scope * information in ip header (nip6). */ m->m_pkthdr.rcvif = NULL; ICMP6STAT_INC(icp6s_outhist[type]); icmp6_reflect(m, sizeof(struct ip6_hdr)); /* header order: IPv6 - ICMPv6 */ return; freeit: /* * If we can't tell whether or not we can generate ICMP6, free it. */ m_freem(m); } /* * Process a received ICMP6 message. */ int icmp6_input(struct mbuf **mp, int *offp, int proto) { struct mbuf *m = *mp, *n; struct ifnet *ifp; struct ip6_hdr *ip6, *nip6; struct icmp6_hdr *icmp6, *nicmp6; int off = *offp; int icmp6len = m->m_pkthdr.len - *offp; int code, sum, noff; char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN]; int ip6len, error; ifp = m->m_pkthdr.rcvif; #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, off, sizeof(struct icmp6_hdr), IPPROTO_DONE); /* m might change if M_LOOP. So, call mtod after this */ #endif /* * Locate icmp6 structure in mbuf, and check * that not corrupted and of at least minimum length */ ip6 = mtod(m, struct ip6_hdr *); ip6len = sizeof(struct ip6_hdr) + ntohs(ip6->ip6_plen); if (icmp6len < sizeof(struct icmp6_hdr)) { ICMP6STAT_INC(icp6s_tooshort); goto freeit; } /* * Check multicast group membership. * Note: SSM filters are not applied for ICMPv6 traffic. */ if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) { struct in6_multi *inm; inm = in6m_lookup(ifp, &ip6->ip6_dst); if (inm == NULL) { IP6STAT_INC(ip6s_notmember); in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard); goto freeit; } } /* * calculate the checksum */ #ifndef PULLDOWN_TEST icmp6 = (struct icmp6_hdr *)((caddr_t)ip6 + off); #else IP6_EXTHDR_GET(icmp6, struct icmp6_hdr *, m, off, sizeof(*icmp6)); if (icmp6 == NULL) { ICMP6STAT_INC(icp6s_tooshort); return IPPROTO_DONE; } #endif code = icmp6->icmp6_code; if ((sum = in6_cksum(m, IPPROTO_ICMPV6, off, icmp6len)) != 0) { nd6log((LOG_ERR, "ICMP6 checksum error(%d|%x) %s\n", icmp6->icmp6_type, sum, ip6_sprintf(ip6bufs, &ip6->ip6_src))); ICMP6STAT_INC(icp6s_checksum); goto freeit; } if (faithprefix_p != NULL && (*faithprefix_p)(&ip6->ip6_dst)) { /* * Deliver very specific ICMP6 type only. * This is important to deliver TOOBIG. Otherwise PMTUD * will not work. */ switch (icmp6->icmp6_type) { case ICMP6_DST_UNREACH: case ICMP6_PACKET_TOO_BIG: case ICMP6_TIME_EXCEEDED: break; default: goto freeit; } } ICMP6STAT_INC(icp6s_inhist[icmp6->icmp6_type]); icmp6_ifstat_inc(ifp, ifs6_in_msg); if (icmp6->icmp6_type < ICMP6_INFOMSG_MASK) icmp6_ifstat_inc(ifp, ifs6_in_error); switch (icmp6->icmp6_type) { case ICMP6_DST_UNREACH: icmp6_ifstat_inc(ifp, ifs6_in_dstunreach); switch (code) { case ICMP6_DST_UNREACH_NOROUTE: code = PRC_UNREACH_NET; break; case ICMP6_DST_UNREACH_ADMIN: icmp6_ifstat_inc(ifp, ifs6_in_adminprohib); code = PRC_UNREACH_PROTOCOL; /* is this a good code? */ break; case ICMP6_DST_UNREACH_ADDR: code = PRC_HOSTDEAD; break; case ICMP6_DST_UNREACH_BEYONDSCOPE: /* I mean "source address was incorrect." */ code = PRC_PARAMPROB; break; case ICMP6_DST_UNREACH_NOPORT: code = PRC_UNREACH_PORT; break; default: goto badcode; } goto deliver; break; case ICMP6_PACKET_TOO_BIG: icmp6_ifstat_inc(ifp, ifs6_in_pkttoobig); /* validation is made in icmp6_mtudisc_update */ code = PRC_MSGSIZE; /* * Updating the path MTU will be done after examining * intermediate extension headers. */ goto deliver; break; case ICMP6_TIME_EXCEEDED: icmp6_ifstat_inc(ifp, ifs6_in_timeexceed); switch (code) { case ICMP6_TIME_EXCEED_TRANSIT: code = PRC_TIMXCEED_INTRANS; break; case ICMP6_TIME_EXCEED_REASSEMBLY: code = PRC_TIMXCEED_REASS; break; default: goto badcode; } goto deliver; break; case ICMP6_PARAM_PROB: icmp6_ifstat_inc(ifp, ifs6_in_paramprob); switch (code) { case ICMP6_PARAMPROB_NEXTHEADER: code = PRC_UNREACH_PROTOCOL; break; case ICMP6_PARAMPROB_HEADER: case ICMP6_PARAMPROB_OPTION: code = PRC_PARAMPROB; break; default: goto badcode; } goto deliver; break; case ICMP6_ECHO_REQUEST: icmp6_ifstat_inc(ifp, ifs6_in_echo); if (code != 0) goto badcode; if ((n = m_copy(m, 0, M_COPYALL)) == NULL) { /* Give up remote */ break; } if ((n->m_flags & M_EXT) != 0 || n->m_len < off + sizeof(struct icmp6_hdr)) { struct mbuf *n0 = n; int n0len; CTASSERT(sizeof(*nip6) + sizeof(*nicmp6) <= MHLEN); n = m_gethdr(M_NOWAIT, n0->m_type); if (n == NULL) { /* Give up remote */ m_freem(n0); break; } m_move_pkthdr(n, n0); /* FIB copied. */ n0len = n0->m_pkthdr.len; /* save for use below */ /* * Copy IPv6 and ICMPv6 only. */ nip6 = mtod(n, struct ip6_hdr *); bcopy(ip6, nip6, sizeof(struct ip6_hdr)); nicmp6 = (struct icmp6_hdr *)(nip6 + 1); bcopy(icmp6, nicmp6, sizeof(struct icmp6_hdr)); noff = sizeof(struct ip6_hdr); /* new mbuf contains only ipv6+icmpv6 headers */ n->m_len = noff + sizeof(struct icmp6_hdr); /* * Adjust mbuf. ip6_plen will be adjusted in * ip6_output(). */ m_adj(n0, off + sizeof(struct icmp6_hdr)); /* recalculate complete packet size */ n->m_pkthdr.len = n0len + (noff - off); n->m_next = n0; } else { nip6 = mtod(n, struct ip6_hdr *); IP6_EXTHDR_GET(nicmp6, struct icmp6_hdr *, n, off, sizeof(*nicmp6)); noff = off; } nicmp6->icmp6_type = ICMP6_ECHO_REPLY; nicmp6->icmp6_code = 0; if (n) { ICMP6STAT_INC(icp6s_reflect); ICMP6STAT_INC(icp6s_outhist[ICMP6_ECHO_REPLY]); icmp6_reflect(n, noff); } break; case ICMP6_ECHO_REPLY: icmp6_ifstat_inc(ifp, ifs6_in_echoreply); if (code != 0) goto badcode; break; case MLD_LISTENER_QUERY: case MLD_LISTENER_REPORT: case MLD_LISTENER_DONE: case MLDV2_LISTENER_REPORT: /* * Drop MLD traffic which is not link-local, has a hop limit * of greater than 1 hop, or which does not have the * IPv6 HBH Router Alert option. * As IPv6 HBH options are stripped in ip6_input() we must * check an mbuf header flag. * XXX Should we also sanity check that these messages * were directed to a link-local multicast prefix? */ if ((ip6->ip6_hlim != 1) || (m->m_flags & M_RTALERT_MLD) == 0) goto freeit; if (mld_input(m, off, icmp6len) != 0) return (IPPROTO_DONE); /* m stays. */ break; case ICMP6_WRUREQUEST: /* ICMP6_FQDN_QUERY */ { enum { WRU, FQDN } mode; if (!V_icmp6_nodeinfo) break; if (icmp6len == sizeof(struct icmp6_hdr) + 4) mode = WRU; else if (icmp6len >= sizeof(struct icmp6_nodeinfo)) mode = FQDN; else goto badlen; if (mode == FQDN) { #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, off, sizeof(struct icmp6_nodeinfo), IPPROTO_DONE); #endif n = m_copy(m, 0, M_COPYALL); if (n) n = ni6_input(n, off); /* XXX meaningless if n == NULL */ noff = sizeof(struct ip6_hdr); } else { struct prison *pr; u_char *p; int maxhlen, hlen; /* * XXX: this combination of flags is pointless, * but should we keep this for compatibility? */ if ((V_icmp6_nodeinfo & 5) != 5) break; if (code != 0) goto badcode; CTASSERT(sizeof(*nip6) + sizeof(*nicmp6) + 4 <= MHLEN); n = m_gethdr(M_NOWAIT, m->m_type); if (n == NULL) { /* Give up remote */ break; } if (!m_dup_pkthdr(n, m, M_NOWAIT)) { /* * Previous code did a blind M_COPY_PKTHDR * and said "just for rcvif". If true, then * we could tolerate the dup failing (due to * the deep copy of the tag chain). For now * be conservative and just fail. */ m_free(n); n = NULL; } maxhlen = M_TRAILINGSPACE(n) - (sizeof(*nip6) + sizeof(*nicmp6) + 4); pr = curthread->td_ucred->cr_prison; mtx_lock(&pr->pr_mtx); hlen = strlen(pr->pr_hostname); if (maxhlen > hlen) maxhlen = hlen; /* * Copy IPv6 and ICMPv6 only. */ nip6 = mtod(n, struct ip6_hdr *); bcopy(ip6, nip6, sizeof(struct ip6_hdr)); nicmp6 = (struct icmp6_hdr *)(nip6 + 1); bcopy(icmp6, nicmp6, sizeof(struct icmp6_hdr)); p = (u_char *)(nicmp6 + 1); bzero(p, 4); /* meaningless TTL */ bcopy(pr->pr_hostname, p + 4, maxhlen); mtx_unlock(&pr->pr_mtx); noff = sizeof(struct ip6_hdr); n->m_pkthdr.len = n->m_len = sizeof(struct ip6_hdr) + sizeof(struct icmp6_hdr) + 4 + maxhlen; nicmp6->icmp6_type = ICMP6_WRUREPLY; nicmp6->icmp6_code = 0; } if (n) { ICMP6STAT_INC(icp6s_reflect); ICMP6STAT_INC(icp6s_outhist[ICMP6_WRUREPLY]); icmp6_reflect(n, noff); } break; } case ICMP6_WRUREPLY: if (code != 0) goto badcode; break; case ND_ROUTER_SOLICIT: icmp6_ifstat_inc(ifp, ifs6_in_routersolicit); if (code != 0) goto badcode; if (icmp6len < sizeof(struct nd_router_solicit)) goto badlen; if ((n = m_copym(m, 0, M_COPYALL, M_NOWAIT)) == NULL) { /* give up local */ /* Send incoming SeND packet to user space. */ if (send_sendso_input_hook != NULL) { IP6_EXTHDR_CHECK(m, off, icmp6len, IPPROTO_DONE); error = send_sendso_input_hook(m, ifp, SND_IN, ip6len); /* -1 == no app on SEND socket */ if (error == 0) return (IPPROTO_DONE); nd6_rs_input(m, off, icmp6len); } else nd6_rs_input(m, off, icmp6len); m = NULL; goto freeit; } if (send_sendso_input_hook != NULL) { IP6_EXTHDR_CHECK(n, off, icmp6len, IPPROTO_DONE); error = send_sendso_input_hook(n, ifp, SND_IN, ip6len); if (error == 0) goto freeit; /* -1 == no app on SEND socket */ nd6_rs_input(n, off, icmp6len); } else nd6_rs_input(n, off, icmp6len); /* m stays. */ break; case ND_ROUTER_ADVERT: icmp6_ifstat_inc(ifp, ifs6_in_routeradvert); if (code != 0) goto badcode; if (icmp6len < sizeof(struct nd_router_advert)) goto badlen; if ((n = m_copym(m, 0, M_COPYALL, M_NOWAIT)) == NULL) { /* Send incoming SeND-protected/ND packet to user space. */ if (send_sendso_input_hook != NULL) { error = send_sendso_input_hook(m, ifp, SND_IN, ip6len); if (error == 0) return (IPPROTO_DONE); nd6_ra_input(m, off, icmp6len); } else nd6_ra_input(m, off, icmp6len); m = NULL; goto freeit; } if (send_sendso_input_hook != NULL) { error = send_sendso_input_hook(n, ifp, SND_IN, ip6len); if (error == 0) goto freeit; nd6_ra_input(n, off, icmp6len); } else nd6_ra_input(n, off, icmp6len); /* m stays. */ break; case ND_NEIGHBOR_SOLICIT: icmp6_ifstat_inc(ifp, ifs6_in_neighborsolicit); if (code != 0) goto badcode; if (icmp6len < sizeof(struct nd_neighbor_solicit)) goto badlen; if ((n = m_copym(m, 0, M_COPYALL, M_NOWAIT)) == NULL) { if (send_sendso_input_hook != NULL) { error = send_sendso_input_hook(m, ifp, SND_IN, ip6len); if (error == 0) return (IPPROTO_DONE); nd6_ns_input(m, off, icmp6len); } else nd6_ns_input(m, off, icmp6len); m = NULL; goto freeit; } if (send_sendso_input_hook != NULL) { error = send_sendso_input_hook(n, ifp, SND_IN, ip6len); if (error == 0) goto freeit; nd6_ns_input(n, off, icmp6len); } else nd6_ns_input(n, off, icmp6len); /* m stays. */ break; case ND_NEIGHBOR_ADVERT: icmp6_ifstat_inc(ifp, ifs6_in_neighboradvert); if (code != 0) goto badcode; if (icmp6len < sizeof(struct nd_neighbor_advert)) goto badlen; if ((n = m_copym(m, 0, M_COPYALL, M_NOWAIT)) == NULL) { /* Send incoming SeND-protected/ND packet to user space. */ if (send_sendso_input_hook != NULL) { error = send_sendso_input_hook(m, ifp, SND_IN, ip6len); if (error == 0) return (IPPROTO_DONE); nd6_na_input(m, off, icmp6len); } else nd6_na_input(m, off, icmp6len); m = NULL; goto freeit; } if (send_sendso_input_hook != NULL) { error = send_sendso_input_hook(n, ifp, SND_IN, ip6len); if (error == 0) goto freeit; nd6_na_input(n, off, icmp6len); } else nd6_na_input(n, off, icmp6len); /* m stays. */ break; case ND_REDIRECT: icmp6_ifstat_inc(ifp, ifs6_in_redirect); if (code != 0) goto badcode; if (icmp6len < sizeof(struct nd_redirect)) goto badlen; if ((n = m_copym(m, 0, M_COPYALL, M_NOWAIT)) == NULL) { if (send_sendso_input_hook != NULL) { error = send_sendso_input_hook(m, ifp, SND_IN, ip6len); if (error == 0) return (IPPROTO_DONE); icmp6_redirect_input(m, off); } else icmp6_redirect_input(m, off); m = NULL; goto freeit; } if (send_sendso_input_hook != NULL) { error = send_sendso_input_hook(n, ifp, SND_IN, ip6len); if (error == 0) goto freeit; icmp6_redirect_input(n, off); } else icmp6_redirect_input(n, off); /* m stays. */ break; case ICMP6_ROUTER_RENUMBERING: if (code != ICMP6_ROUTER_RENUMBERING_COMMAND && code != ICMP6_ROUTER_RENUMBERING_RESULT) goto badcode; if (icmp6len < sizeof(struct icmp6_router_renum)) goto badlen; break; default: nd6log((LOG_DEBUG, "icmp6_input: unknown type %d(src=%s, dst=%s, ifid=%d)\n", icmp6->icmp6_type, ip6_sprintf(ip6bufs, &ip6->ip6_src), ip6_sprintf(ip6bufd, &ip6->ip6_dst), ifp ? ifp->if_index : 0)); if (icmp6->icmp6_type < ICMP6_ECHO_REQUEST) { /* ICMPv6 error: MUST deliver it by spec... */ code = PRC_NCMDS; /* deliver */ } else { /* ICMPv6 informational: MUST not deliver */ break; } deliver: if (icmp6_notify_error(&m, off, icmp6len, code) != 0) { /* In this case, m should've been freed. */ return (IPPROTO_DONE); } break; badcode: ICMP6STAT_INC(icp6s_badcode); break; badlen: ICMP6STAT_INC(icp6s_badlen); break; } /* deliver the packet to appropriate sockets */ icmp6_rip6_input(&m, *offp); return IPPROTO_DONE; freeit: m_freem(m); return IPPROTO_DONE; } static int icmp6_notify_error(struct mbuf **mp, int off, int icmp6len, int code) { struct mbuf *m = *mp; struct icmp6_hdr *icmp6; struct ip6_hdr *eip6; u_int32_t notifymtu; struct sockaddr_in6 icmp6src, icmp6dst; if (icmp6len < sizeof(struct icmp6_hdr) + sizeof(struct ip6_hdr)) { ICMP6STAT_INC(icp6s_tooshort); goto freeit; } #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, off, sizeof(struct icmp6_hdr) + sizeof(struct ip6_hdr), -1); icmp6 = (struct icmp6_hdr *)(mtod(m, caddr_t) + off); #else IP6_EXTHDR_GET(icmp6, struct icmp6_hdr *, m, off, sizeof(*icmp6) + sizeof(struct ip6_hdr)); if (icmp6 == NULL) { ICMP6STAT_INC(icp6s_tooshort); return (-1); } #endif eip6 = (struct ip6_hdr *)(icmp6 + 1); /* Detect the upper level protocol */ { void (*ctlfunc)(int, struct sockaddr *, void *); u_int8_t nxt = eip6->ip6_nxt; int eoff = off + sizeof(struct icmp6_hdr) + sizeof(struct ip6_hdr); struct ip6ctlparam ip6cp; struct in6_addr *finaldst = NULL; int icmp6type = icmp6->icmp6_type; struct ip6_frag *fh; struct ip6_rthdr *rth; struct ip6_rthdr0 *rth0; int rthlen; while (1) { /* XXX: should avoid infinite loop explicitly? */ struct ip6_ext *eh; switch (nxt) { case IPPROTO_HOPOPTS: case IPPROTO_DSTOPTS: case IPPROTO_AH: #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, 0, eoff + sizeof(struct ip6_ext), -1); eh = (struct ip6_ext *)(mtod(m, caddr_t) + eoff); #else IP6_EXTHDR_GET(eh, struct ip6_ext *, m, eoff, sizeof(*eh)); if (eh == NULL) { ICMP6STAT_INC(icp6s_tooshort); return (-1); } #endif if (nxt == IPPROTO_AH) eoff += (eh->ip6e_len + 2) << 2; else eoff += (eh->ip6e_len + 1) << 3; nxt = eh->ip6e_nxt; break; case IPPROTO_ROUTING: /* * When the erroneous packet contains a * routing header, we should examine the * header to determine the final destination. * Otherwise, we can't properly update * information that depends on the final * destination (e.g. path MTU). */ #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, 0, eoff + sizeof(*rth), -1); rth = (struct ip6_rthdr *) (mtod(m, caddr_t) + eoff); #else IP6_EXTHDR_GET(rth, struct ip6_rthdr *, m, eoff, sizeof(*rth)); if (rth == NULL) { ICMP6STAT_INC(icp6s_tooshort); return (-1); } #endif rthlen = (rth->ip6r_len + 1) << 3; /* * XXX: currently there is no * officially defined type other * than type-0. * Note that if the segment left field * is 0, all intermediate hops must * have been passed. */ if (rth->ip6r_segleft && rth->ip6r_type == IPV6_RTHDR_TYPE_0) { int hops; #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, 0, eoff + rthlen, -1); rth0 = (struct ip6_rthdr0 *) (mtod(m, caddr_t) + eoff); #else IP6_EXTHDR_GET(rth0, struct ip6_rthdr0 *, m, eoff, rthlen); if (rth0 == NULL) { ICMP6STAT_INC(icp6s_tooshort); return (-1); } #endif /* just ignore a bogus header */ if ((rth0->ip6r0_len % 2) == 0 && (hops = rth0->ip6r0_len/2)) finaldst = (struct in6_addr *)(rth0 + 1) + (hops - 1); } eoff += rthlen; nxt = rth->ip6r_nxt; break; case IPPROTO_FRAGMENT: #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, 0, eoff + sizeof(struct ip6_frag), -1); fh = (struct ip6_frag *)(mtod(m, caddr_t) + eoff); #else IP6_EXTHDR_GET(fh, struct ip6_frag *, m, eoff, sizeof(*fh)); if (fh == NULL) { ICMP6STAT_INC(icp6s_tooshort); return (-1); } #endif /* * Data after a fragment header is meaningless * unless it is the first fragment, but * we'll go to the notify label for path MTU * discovery. */ if (fh->ip6f_offlg & IP6F_OFF_MASK) goto notify; eoff += sizeof(struct ip6_frag); nxt = fh->ip6f_nxt; break; default: /* * This case includes ESP and the No Next * Header. In such cases going to the notify * label does not have any meaning * (i.e. ctlfunc will be NULL), but we go * anyway since we might have to update * path MTU information. */ goto notify; } } notify: #ifndef PULLDOWN_TEST icmp6 = (struct icmp6_hdr *)(mtod(m, caddr_t) + off); #else IP6_EXTHDR_GET(icmp6, struct icmp6_hdr *, m, off, sizeof(*icmp6) + sizeof(struct ip6_hdr)); if (icmp6 == NULL) { ICMP6STAT_INC(icp6s_tooshort); return (-1); } #endif /* * retrieve parameters from the inner IPv6 header, and convert * them into sockaddr structures. * XXX: there is no guarantee that the source or destination * addresses of the inner packet are in the same scope as * the addresses of the icmp packet. But there is no other * way to determine the zone. */ eip6 = (struct ip6_hdr *)(icmp6 + 1); bzero(&icmp6dst, sizeof(icmp6dst)); icmp6dst.sin6_len = sizeof(struct sockaddr_in6); icmp6dst.sin6_family = AF_INET6; if (finaldst == NULL) icmp6dst.sin6_addr = eip6->ip6_dst; else icmp6dst.sin6_addr = *finaldst; if (in6_setscope(&icmp6dst.sin6_addr, m->m_pkthdr.rcvif, NULL)) goto freeit; bzero(&icmp6src, sizeof(icmp6src)); icmp6src.sin6_len = sizeof(struct sockaddr_in6); icmp6src.sin6_family = AF_INET6; icmp6src.sin6_addr = eip6->ip6_src; if (in6_setscope(&icmp6src.sin6_addr, m->m_pkthdr.rcvif, NULL)) goto freeit; icmp6src.sin6_flowinfo = (eip6->ip6_flow & IPV6_FLOWLABEL_MASK); if (finaldst == NULL) finaldst = &eip6->ip6_dst; ip6cp.ip6c_m = m; ip6cp.ip6c_icmp6 = icmp6; ip6cp.ip6c_ip6 = (struct ip6_hdr *)(icmp6 + 1); ip6cp.ip6c_off = eoff; ip6cp.ip6c_finaldst = finaldst; ip6cp.ip6c_src = &icmp6src; ip6cp.ip6c_nxt = nxt; if (icmp6type == ICMP6_PACKET_TOO_BIG) { notifymtu = ntohl(icmp6->icmp6_mtu); ip6cp.ip6c_cmdarg = (void *)¬ifymtu; icmp6_mtudisc_update(&ip6cp, 1); /*XXX*/ } ctlfunc = (void (*)(int, struct sockaddr *, void *)) (inet6sw[ip6_protox[nxt]].pr_ctlinput); if (ctlfunc) { (void) (*ctlfunc)(code, (struct sockaddr *)&icmp6dst, &ip6cp); } } *mp = m; return (0); freeit: m_freem(m); return (-1); } void icmp6_mtudisc_update(struct ip6ctlparam *ip6cp, int validated) { struct in6_addr *dst = ip6cp->ip6c_finaldst; struct icmp6_hdr *icmp6 = ip6cp->ip6c_icmp6; struct mbuf *m = ip6cp->ip6c_m; /* will be necessary for scope issue */ u_int mtu = ntohl(icmp6->icmp6_mtu); struct in_conninfo inc; #if 0 /* * RFC2460 section 5, last paragraph. * even though minimum link MTU for IPv6 is IPV6_MMTU, * we may see ICMPv6 too big with mtu < IPV6_MMTU * due to packet translator in the middle. * see ip6_output() and ip6_getpmtu() "alwaysfrag" case for * special handling. */ if (mtu < IPV6_MMTU) return; #endif /* * we reject ICMPv6 too big with abnormally small value. * XXX what is the good definition of "abnormally small"? */ if (mtu < sizeof(struct ip6_hdr) + sizeof(struct ip6_frag) + 8) return; if (!validated) return; /* * In case the suggested mtu is less than IPV6_MMTU, we * only need to remember that it was for above mentioned * "alwaysfrag" case. * Try to be as close to the spec as possible. */ if (mtu < IPV6_MMTU) mtu = IPV6_MMTU - 8; bzero(&inc, sizeof(inc)); inc.inc_flags |= INC_ISIPV6; inc.inc6_faddr = *dst; if (in6_setscope(&inc.inc6_faddr, m->m_pkthdr.rcvif, NULL)) return; if (mtu < tcp_maxmtu6(&inc, NULL)) { tcp_hc_updatemtu(&inc, mtu); ICMP6STAT_INC(icp6s_pmtuchg); } } /* * Process a Node Information Query packet, based on * draft-ietf-ipngwg-icmp-name-lookups-07. * * Spec incompatibilities: * - IPv6 Subject address handling * - IPv4 Subject address handling support missing * - Proxy reply (answer even if it's not for me) * - joins NI group address at in6_ifattach() time only, does not cope * with hostname changes by sethostname(3) */ static struct mbuf * ni6_input(struct mbuf *m, int off) { struct icmp6_nodeinfo *ni6, *nni6; struct mbuf *n = NULL; struct prison *pr; u_int16_t qtype; int subjlen; int replylen = sizeof(struct ip6_hdr) + sizeof(struct icmp6_nodeinfo); struct ni_reply_fqdn *fqdn; int addrs; /* for NI_QTYPE_NODEADDR */ struct ifnet *ifp = NULL; /* for NI_QTYPE_NODEADDR */ struct in6_addr in6_subj; /* subject address */ struct ip6_hdr *ip6; int oldfqdn = 0; /* if 1, return pascal string (03 draft) */ char *subj = NULL; struct in6_ifaddr *ia6 = NULL; ip6 = mtod(m, struct ip6_hdr *); #ifndef PULLDOWN_TEST ni6 = (struct icmp6_nodeinfo *)(mtod(m, caddr_t) + off); #else IP6_EXTHDR_GET(ni6, struct icmp6_nodeinfo *, m, off, sizeof(*ni6)); if (ni6 == NULL) { /* m is already reclaimed */ return (NULL); } #endif /* * Validate IPv6 source address. * The default configuration MUST be to refuse answering queries from * global-scope addresses according to RFC4602. * Notes: * - it's not very clear what "refuse" means; this implementation * simply drops it. * - it's not very easy to identify global-scope (unicast) addresses * since there are many prefixes for them. It should be safer * and in practice sufficient to check "all" but loopback and * link-local (note that site-local unicast was deprecated and * ULA is defined as global scope-wise) */ if ((V_icmp6_nodeinfo & ICMP6_NODEINFO_GLOBALOK) == 0 && !IN6_IS_ADDR_LOOPBACK(&ip6->ip6_src) && !IN6_IS_ADDR_LINKLOCAL(&ip6->ip6_src)) goto bad; /* * Validate IPv6 destination address. * * The Responder must discard the Query without further processing * unless it is one of the Responder's unicast or anycast addresses, or * a link-local scope multicast address which the Responder has joined. * [RFC4602, Section 5.] */ if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) { if (!IN6_IS_ADDR_MC_LINKLOCAL(&ip6->ip6_dst)) goto bad; /* else it's a link-local multicast, fine */ } else { /* unicast or anycast */ if ((ia6 = ip6_getdstifaddr(m)) == NULL) goto bad; /* XXX impossible */ if ((ia6->ia6_flags & IN6_IFF_TEMPORARY) && !(V_icmp6_nodeinfo & ICMP6_NODEINFO_TMPADDROK)) { ifa_free(&ia6->ia_ifa); nd6log((LOG_DEBUG, "ni6_input: ignore node info to " "a temporary address in %s:%d", __FILE__, __LINE__)); goto bad; } ifa_free(&ia6->ia_ifa); } /* validate query Subject field. */ qtype = ntohs(ni6->ni_qtype); subjlen = m->m_pkthdr.len - off - sizeof(struct icmp6_nodeinfo); switch (qtype) { case NI_QTYPE_NOOP: case NI_QTYPE_SUPTYPES: /* 07 draft */ if (ni6->ni_code == ICMP6_NI_SUBJ_FQDN && subjlen == 0) break; /* FALLTHROUGH */ case NI_QTYPE_FQDN: case NI_QTYPE_NODEADDR: case NI_QTYPE_IPV4ADDR: switch (ni6->ni_code) { case ICMP6_NI_SUBJ_IPV6: #if ICMP6_NI_SUBJ_IPV6 != 0 case 0: #endif /* * backward compatibility - try to accept 03 draft * format, where no Subject is present. */ if (qtype == NI_QTYPE_FQDN && ni6->ni_code == 0 && subjlen == 0) { oldfqdn++; break; } #if ICMP6_NI_SUBJ_IPV6 != 0 if (ni6->ni_code != ICMP6_NI_SUBJ_IPV6) goto bad; #endif if (subjlen != sizeof(struct in6_addr)) goto bad; /* * Validate Subject address. * * Not sure what exactly "address belongs to the node" * means in the spec, is it just unicast, or what? * * At this moment we consider Subject address as * "belong to the node" if the Subject address equals * to the IPv6 destination address; validation for * IPv6 destination address should have done enough * check for us. * * We do not do proxy at this moment. */ /* m_pulldown instead of copy? */ m_copydata(m, off + sizeof(struct icmp6_nodeinfo), subjlen, (caddr_t)&in6_subj); if (in6_setscope(&in6_subj, m->m_pkthdr.rcvif, NULL)) goto bad; subj = (char *)&in6_subj; if (IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, &in6_subj)) break; /* * XXX if we are to allow other cases, we should really * be careful about scope here. * basically, we should disallow queries toward IPv6 * destination X with subject Y, * if scope(X) > scope(Y). * if we allow scope(X) > scope(Y), it will result in * information leakage across scope boundary. */ goto bad; case ICMP6_NI_SUBJ_FQDN: /* * Validate Subject name with gethostname(3). * * The behavior may need some debate, since: * - we are not sure if the node has FQDN as * hostname (returned by gethostname(3)). * - the code does wildcard match for truncated names. * however, we are not sure if we want to perform * wildcard match, if gethostname(3) side has * truncated hostname. */ pr = curthread->td_ucred->cr_prison; mtx_lock(&pr->pr_mtx); n = ni6_nametodns(pr->pr_hostname, strlen(pr->pr_hostname), 0); mtx_unlock(&pr->pr_mtx); if (!n || n->m_next || n->m_len == 0) goto bad; IP6_EXTHDR_GET(subj, char *, m, off + sizeof(struct icmp6_nodeinfo), subjlen); if (subj == NULL) goto bad; if (!ni6_dnsmatch(subj, subjlen, mtod(n, const char *), n->m_len)) { goto bad; } m_freem(n); n = NULL; break; case ICMP6_NI_SUBJ_IPV4: /* XXX: to be implemented? */ default: goto bad; } break; } /* refuse based on configuration. XXX ICMP6_NI_REFUSED? */ switch (qtype) { case NI_QTYPE_FQDN: if ((V_icmp6_nodeinfo & ICMP6_NODEINFO_FQDNOK) == 0) goto bad; break; case NI_QTYPE_NODEADDR: case NI_QTYPE_IPV4ADDR: if ((V_icmp6_nodeinfo & ICMP6_NODEINFO_NODEADDROK) == 0) goto bad; break; } /* guess reply length */ switch (qtype) { case NI_QTYPE_NOOP: break; /* no reply data */ case NI_QTYPE_SUPTYPES: replylen += sizeof(u_int32_t); break; case NI_QTYPE_FQDN: /* XXX will append an mbuf */ replylen += offsetof(struct ni_reply_fqdn, ni_fqdn_namelen); break; case NI_QTYPE_NODEADDR: addrs = ni6_addrs(ni6, m, &ifp, (struct in6_addr *)subj); if ((replylen += addrs * (sizeof(struct in6_addr) + sizeof(u_int32_t))) > MCLBYTES) replylen = MCLBYTES; /* XXX: will truncate pkt later */ break; case NI_QTYPE_IPV4ADDR: /* unsupported - should respond with unknown Qtype? */ break; default: /* * XXX: We must return a reply with the ICMP6 code * `unknown Qtype' in this case. However we regard the case * as an FQDN query for backward compatibility. * Older versions set a random value to this field, * so it rarely varies in the defined qtypes. * But the mechanism is not reliable... * maybe we should obsolete older versions. */ qtype = NI_QTYPE_FQDN; /* XXX will append an mbuf */ replylen += offsetof(struct ni_reply_fqdn, ni_fqdn_namelen); oldfqdn++; break; } /* Allocate an mbuf to reply. */ if (replylen > MCLBYTES) { /* * XXX: should we try to allocate more? But MCLBYTES * is probably much larger than IPV6_MMTU... */ goto bad; } if (replylen > MHLEN) n = m_getcl(M_NOWAIT, m->m_type, M_PKTHDR); else n = m_gethdr(M_NOWAIT, m->m_type); if (n == NULL) { m_freem(m); return (NULL); } m_move_pkthdr(n, m); /* just for recvif and FIB */ n->m_pkthdr.len = n->m_len = replylen; /* copy mbuf header and IPv6 + Node Information base headers */ bcopy(mtod(m, caddr_t), mtod(n, caddr_t), sizeof(struct ip6_hdr)); nni6 = (struct icmp6_nodeinfo *)(mtod(n, struct ip6_hdr *) + 1); bcopy((caddr_t)ni6, (caddr_t)nni6, sizeof(struct icmp6_nodeinfo)); /* qtype dependent procedure */ switch (qtype) { case NI_QTYPE_NOOP: nni6->ni_code = ICMP6_NI_SUCCESS; nni6->ni_flags = 0; break; case NI_QTYPE_SUPTYPES: { u_int32_t v; nni6->ni_code = ICMP6_NI_SUCCESS; nni6->ni_flags = htons(0x0000); /* raw bitmap */ /* supports NOOP, SUPTYPES, FQDN, and NODEADDR */ v = (u_int32_t)htonl(0x0000000f); bcopy(&v, nni6 + 1, sizeof(u_int32_t)); break; } case NI_QTYPE_FQDN: nni6->ni_code = ICMP6_NI_SUCCESS; fqdn = (struct ni_reply_fqdn *)(mtod(n, caddr_t) + sizeof(struct ip6_hdr) + sizeof(struct icmp6_nodeinfo)); nni6->ni_flags = 0; /* XXX: meaningless TTL */ fqdn->ni_fqdn_ttl = 0; /* ditto. */ /* * XXX do we really have FQDN in hostname? */ pr = curthread->td_ucred->cr_prison; mtx_lock(&pr->pr_mtx); n->m_next = ni6_nametodns(pr->pr_hostname, strlen(pr->pr_hostname), oldfqdn); mtx_unlock(&pr->pr_mtx); if (n->m_next == NULL) goto bad; /* XXX we assume that n->m_next is not a chain */ if (n->m_next->m_next != NULL) goto bad; n->m_pkthdr.len += n->m_next->m_len; break; case NI_QTYPE_NODEADDR: { int lenlim, copied; nni6->ni_code = ICMP6_NI_SUCCESS; n->m_pkthdr.len = n->m_len = sizeof(struct ip6_hdr) + sizeof(struct icmp6_nodeinfo); lenlim = M_TRAILINGSPACE(n); copied = ni6_store_addrs(ni6, nni6, ifp, lenlim); /* XXX: reset mbuf length */ n->m_pkthdr.len = n->m_len = sizeof(struct ip6_hdr) + sizeof(struct icmp6_nodeinfo) + copied; break; } default: break; /* XXX impossible! */ } nni6->ni_type = ICMP6_NI_REPLY; m_freem(m); return (n); bad: m_freem(m); if (n) m_freem(n); return (NULL); } /* * make a mbuf with DNS-encoded string. no compression support. * * XXX names with less than 2 dots (like "foo" or "foo.section") will be * treated as truncated name (two \0 at the end). this is a wild guess. * * old - return pascal string if non-zero */ static struct mbuf * ni6_nametodns(const char *name, int namelen, int old) { struct mbuf *m; char *cp, *ep; const char *p, *q; int i, len, nterm; if (old) len = namelen + 1; else len = MCLBYTES; /* Because MAXHOSTNAMELEN is usually 256, we use cluster mbuf. */ if (len > MLEN) m = m_getcl(M_NOWAIT, MT_DATA, 0); else m = m_get(M_NOWAIT, MT_DATA); if (m == NULL) goto fail; if (old) { m->m_len = len; *mtod(m, char *) = namelen; bcopy(name, mtod(m, char *) + 1, namelen); return m; } else { m->m_len = 0; cp = mtod(m, char *); ep = mtod(m, char *) + M_TRAILINGSPACE(m); /* if not certain about my name, return empty buffer */ if (namelen == 0) return m; /* * guess if it looks like shortened hostname, or FQDN. * shortened hostname needs two trailing "\0". */ i = 0; for (p = name; p < name + namelen; p++) { if (*p && *p == '.') i++; } if (i < 2) nterm = 2; else nterm = 1; p = name; while (cp < ep && p < name + namelen) { i = 0; for (q = p; q < name + namelen && *q && *q != '.'; q++) i++; /* result does not fit into mbuf */ if (cp + i + 1 >= ep) goto fail; /* * DNS label length restriction, RFC1035 page 8. * "i == 0" case is included here to avoid returning * 0-length label on "foo..bar". */ if (i <= 0 || i >= 64) goto fail; *cp++ = i; bcopy(p, cp, i); cp += i; p = q; if (p < name + namelen && *p == '.') p++; } /* termination */ if (cp + nterm >= ep) goto fail; while (nterm-- > 0) *cp++ = '\0'; m->m_len = cp - mtod(m, char *); return m; } panic("should not reach here"); /* NOTREACHED */ fail: if (m) m_freem(m); return NULL; } /* * check if two DNS-encoded string matches. takes care of truncated * form (with \0\0 at the end). no compression support. * XXX upper/lowercase match (see RFC2065) */ static int ni6_dnsmatch(const char *a, int alen, const char *b, int blen) { const char *a0, *b0; int l; /* simplest case - need validation? */ if (alen == blen && bcmp(a, b, alen) == 0) return 1; a0 = a; b0 = b; /* termination is mandatory */ if (alen < 2 || blen < 2) return 0; if (a0[alen - 1] != '\0' || b0[blen - 1] != '\0') return 0; alen--; blen--; while (a - a0 < alen && b - b0 < blen) { if (a - a0 + 1 > alen || b - b0 + 1 > blen) return 0; if ((signed char)a[0] < 0 || (signed char)b[0] < 0) return 0; /* we don't support compression yet */ if (a[0] >= 64 || b[0] >= 64) return 0; /* truncated case */ if (a[0] == 0 && a - a0 == alen - 1) return 1; if (b[0] == 0 && b - b0 == blen - 1) return 1; if (a[0] == 0 || b[0] == 0) return 0; if (a[0] != b[0]) return 0; l = a[0]; if (a - a0 + 1 + l > alen || b - b0 + 1 + l > blen) return 0; if (bcmp(a + 1, b + 1, l) != 0) return 0; a += 1 + l; b += 1 + l; } if (a - a0 == alen && b - b0 == blen) return 1; else return 0; } /* * calculate the number of addresses to be returned in the node info reply. */ static int ni6_addrs(struct icmp6_nodeinfo *ni6, struct mbuf *m, struct ifnet **ifpp, struct in6_addr *subj) { struct ifnet *ifp; struct in6_ifaddr *ifa6; struct ifaddr *ifa; int addrs = 0, addrsofif, iffound = 0; int niflags = ni6->ni_flags; if ((niflags & NI_NODEADDR_FLAG_ALL) == 0) { switch (ni6->ni_code) { case ICMP6_NI_SUBJ_IPV6: if (subj == NULL) /* must be impossible... */ return (0); break; default: /* * XXX: we only support IPv6 subject address for * this Qtype. */ return (0); } } IFNET_RLOCK_NOSLEEP(); TAILQ_FOREACH(ifp, &V_ifnet, if_list) { addrsofif = 0; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; ifa6 = (struct in6_ifaddr *)ifa; if ((niflags & NI_NODEADDR_FLAG_ALL) == 0 && IN6_ARE_ADDR_EQUAL(subj, &ifa6->ia_addr.sin6_addr)) iffound = 1; /* * IPv4-mapped addresses can only be returned by a * Node Information proxy, since they represent * addresses of IPv4-only nodes, which perforce do * not implement this protocol. * [icmp-name-lookups-07, Section 5.4] * So we don't support NI_NODEADDR_FLAG_COMPAT in * this function at this moment. */ /* What do we have to do about ::1? */ switch (in6_addrscope(&ifa6->ia_addr.sin6_addr)) { case IPV6_ADDR_SCOPE_LINKLOCAL: if ((niflags & NI_NODEADDR_FLAG_LINKLOCAL) == 0) continue; break; case IPV6_ADDR_SCOPE_SITELOCAL: if ((niflags & NI_NODEADDR_FLAG_SITELOCAL) == 0) continue; break; case IPV6_ADDR_SCOPE_GLOBAL: if ((niflags & NI_NODEADDR_FLAG_GLOBAL) == 0) continue; break; default: continue; } /* * check if anycast is okay. * XXX: just experimental. not in the spec. */ if ((ifa6->ia6_flags & IN6_IFF_ANYCAST) != 0 && (niflags & NI_NODEADDR_FLAG_ANYCAST) == 0) continue; /* we need only unicast addresses */ if ((ifa6->ia6_flags & IN6_IFF_TEMPORARY) != 0 && (V_icmp6_nodeinfo & ICMP6_NODEINFO_TMPADDROK) == 0) { continue; } addrsofif++; /* count the address */ } IF_ADDR_RUNLOCK(ifp); if (iffound) { *ifpp = ifp; IFNET_RUNLOCK_NOSLEEP(); return (addrsofif); } addrs += addrsofif; } IFNET_RUNLOCK_NOSLEEP(); return (addrs); } static int ni6_store_addrs(struct icmp6_nodeinfo *ni6, struct icmp6_nodeinfo *nni6, struct ifnet *ifp0, int resid) { struct ifnet *ifp; struct in6_ifaddr *ifa6; struct ifaddr *ifa; struct ifnet *ifp_dep = NULL; int copied = 0, allow_deprecated = 0; u_char *cp = (u_char *)(nni6 + 1); int niflags = ni6->ni_flags; u_int32_t ltime; if (ifp0 == NULL && !(niflags & NI_NODEADDR_FLAG_ALL)) return (0); /* needless to copy */ IFNET_RLOCK_NOSLEEP(); ifp = ifp0 ? ifp0 : TAILQ_FIRST(&V_ifnet); again: for (; ifp; ifp = TAILQ_NEXT(ifp, if_list)) { IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; ifa6 = (struct in6_ifaddr *)ifa; if ((ifa6->ia6_flags & IN6_IFF_DEPRECATED) != 0 && allow_deprecated == 0) { /* * prefererred address should be put before * deprecated addresses. */ /* record the interface for later search */ if (ifp_dep == NULL) ifp_dep = ifp; continue; } else if ((ifa6->ia6_flags & IN6_IFF_DEPRECATED) == 0 && allow_deprecated != 0) continue; /* we now collect deprecated addrs */ /* What do we have to do about ::1? */ switch (in6_addrscope(&ifa6->ia_addr.sin6_addr)) { case IPV6_ADDR_SCOPE_LINKLOCAL: if ((niflags & NI_NODEADDR_FLAG_LINKLOCAL) == 0) continue; break; case IPV6_ADDR_SCOPE_SITELOCAL: if ((niflags & NI_NODEADDR_FLAG_SITELOCAL) == 0) continue; break; case IPV6_ADDR_SCOPE_GLOBAL: if ((niflags & NI_NODEADDR_FLAG_GLOBAL) == 0) continue; break; default: continue; } /* * check if anycast is okay. * XXX: just experimental. not in the spec. */ if ((ifa6->ia6_flags & IN6_IFF_ANYCAST) != 0 && (niflags & NI_NODEADDR_FLAG_ANYCAST) == 0) continue; if ((ifa6->ia6_flags & IN6_IFF_TEMPORARY) != 0 && (V_icmp6_nodeinfo & ICMP6_NODEINFO_TMPADDROK) == 0) { continue; } /* now we can copy the address */ if (resid < sizeof(struct in6_addr) + sizeof(u_int32_t)) { IF_ADDR_RUNLOCK(ifp); /* * We give up much more copy. * Set the truncate flag and return. */ nni6->ni_flags |= NI_NODEADDR_FLAG_TRUNCATE; IFNET_RUNLOCK_NOSLEEP(); return (copied); } /* * Set the TTL of the address. * The TTL value should be one of the following * according to the specification: * * 1. The remaining lifetime of a DHCP lease on the * address, or * 2. The remaining Valid Lifetime of a prefix from * which the address was derived through Stateless * Autoconfiguration. * * Note that we currently do not support stateful * address configuration by DHCPv6, so the former * case can't happen. */ if (ifa6->ia6_lifetime.ia6t_expire == 0) ltime = ND6_INFINITE_LIFETIME; else { if (ifa6->ia6_lifetime.ia6t_expire > - time_second) - ltime = htonl(ifa6->ia6_lifetime.ia6t_expire - time_second); + time_uptime) + ltime = htonl(ifa6->ia6_lifetime.ia6t_expire - time_uptime); else ltime = 0; } bcopy(<ime, cp, sizeof(u_int32_t)); cp += sizeof(u_int32_t); /* copy the address itself */ bcopy(&ifa6->ia_addr.sin6_addr, cp, sizeof(struct in6_addr)); in6_clearscope((struct in6_addr *)cp); /* XXX */ cp += sizeof(struct in6_addr); resid -= (sizeof(struct in6_addr) + sizeof(u_int32_t)); copied += (sizeof(struct in6_addr) + sizeof(u_int32_t)); } IF_ADDR_RUNLOCK(ifp); if (ifp0) /* we need search only on the specified IF */ break; } if (allow_deprecated == 0 && ifp_dep != NULL) { ifp = ifp_dep; allow_deprecated = 1; goto again; } IFNET_RUNLOCK_NOSLEEP(); return (copied); } /* * XXX almost dup'ed code with rip6_input. */ static int icmp6_rip6_input(struct mbuf **mp, int off) { struct mbuf *m = *mp; struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); struct inpcb *in6p; struct inpcb *last = NULL; struct sockaddr_in6 fromsa; struct icmp6_hdr *icmp6; struct mbuf *opts = NULL; #ifndef PULLDOWN_TEST /* this is assumed to be safe. */ icmp6 = (struct icmp6_hdr *)((caddr_t)ip6 + off); #else IP6_EXTHDR_GET(icmp6, struct icmp6_hdr *, m, off, sizeof(*icmp6)); if (icmp6 == NULL) { /* m is already reclaimed */ return (IPPROTO_DONE); } #endif /* * XXX: the address may have embedded scope zone ID, which should be * hidden from applications. */ bzero(&fromsa, sizeof(fromsa)); fromsa.sin6_family = AF_INET6; fromsa.sin6_len = sizeof(struct sockaddr_in6); fromsa.sin6_addr = ip6->ip6_src; if (sa6_recoverscope(&fromsa)) { m_freem(m); return (IPPROTO_DONE); } INP_INFO_RLOCK(&V_ripcbinfo); LIST_FOREACH(in6p, &V_ripcb, inp_list) { if ((in6p->inp_vflag & INP_IPV6) == 0) continue; if (in6p->inp_ip_p != IPPROTO_ICMPV6) continue; if (!IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_laddr) && !IN6_ARE_ADDR_EQUAL(&in6p->in6p_laddr, &ip6->ip6_dst)) continue; if (!IN6_IS_ADDR_UNSPECIFIED(&in6p->in6p_faddr) && !IN6_ARE_ADDR_EQUAL(&in6p->in6p_faddr, &ip6->ip6_src)) continue; INP_RLOCK(in6p); if (ICMP6_FILTER_WILLBLOCK(icmp6->icmp6_type, in6p->in6p_icmp6filt)) { INP_RUNLOCK(in6p); continue; } if (last != NULL) { struct mbuf *n = NULL; /* * Recent network drivers tend to allocate a single * mbuf cluster, rather than to make a couple of * mbufs without clusters. Also, since the IPv6 code * path tries to avoid m_pullup(), it is highly * probable that we still have an mbuf cluster here * even though the necessary length can be stored in an * mbuf's internal buffer. * Meanwhile, the default size of the receive socket * buffer for raw sockets is not so large. This means * the possibility of packet loss is relatively higher * than before. To avoid this scenario, we copy the * received data to a separate mbuf that does not use * a cluster, if possible. * XXX: it is better to copy the data after stripping * intermediate headers. */ if ((m->m_flags & M_EXT) && m->m_next == NULL && m->m_len <= MHLEN) { n = m_get(M_NOWAIT, m->m_type); if (n != NULL) { if (m_dup_pkthdr(n, m, M_NOWAIT)) { bcopy(m->m_data, n->m_data, m->m_len); n->m_len = m->m_len; } else { m_free(n); n = NULL; } } } if (n != NULL || (n = m_copy(m, 0, (int)M_COPYALL)) != NULL) { if (last->inp_flags & INP_CONTROLOPTS) ip6_savecontrol(last, n, &opts); /* strip intermediate headers */ m_adj(n, off); SOCKBUF_LOCK(&last->inp_socket->so_rcv); if (sbappendaddr_locked( &last->inp_socket->so_rcv, (struct sockaddr *)&fromsa, n, opts) == 0) { /* should notify about lost packet */ m_freem(n); if (opts) { m_freem(opts); } SOCKBUF_UNLOCK( &last->inp_socket->so_rcv); } else sorwakeup_locked(last->inp_socket); opts = NULL; } INP_RUNLOCK(last); } last = in6p; } INP_INFO_RUNLOCK(&V_ripcbinfo); if (last != NULL) { if (last->inp_flags & INP_CONTROLOPTS) ip6_savecontrol(last, m, &opts); /* strip intermediate headers */ m_adj(m, off); /* avoid using mbuf clusters if possible (see above) */ if ((m->m_flags & M_EXT) && m->m_next == NULL && m->m_len <= MHLEN) { struct mbuf *n; n = m_get(M_NOWAIT, m->m_type); if (n != NULL) { if (m_dup_pkthdr(n, m, M_NOWAIT)) { bcopy(m->m_data, n->m_data, m->m_len); n->m_len = m->m_len; m_freem(m); m = n; } else { m_freem(n); n = NULL; } } } SOCKBUF_LOCK(&last->inp_socket->so_rcv); if (sbappendaddr_locked(&last->inp_socket->so_rcv, (struct sockaddr *)&fromsa, m, opts) == 0) { m_freem(m); if (opts) m_freem(opts); SOCKBUF_UNLOCK(&last->inp_socket->so_rcv); } else sorwakeup_locked(last->inp_socket); INP_RUNLOCK(last); } else { m_freem(m); IP6STAT_DEC(ip6s_delivered); } return IPPROTO_DONE; } /* * Reflect the ip6 packet back to the source. * OFF points to the icmp6 header, counted from the top of the mbuf. */ void icmp6_reflect(struct mbuf *m, size_t off) { struct ip6_hdr *ip6; struct icmp6_hdr *icmp6; struct in6_ifaddr *ia = NULL; int plen; int type, code; struct ifnet *outif = NULL; struct in6_addr origdst, src, *srcp = NULL; /* too short to reflect */ if (off < sizeof(struct ip6_hdr)) { nd6log((LOG_DEBUG, "sanity fail: off=%lx, sizeof(ip6)=%lx in %s:%d\n", (u_long)off, (u_long)sizeof(struct ip6_hdr), __FILE__, __LINE__)); goto bad; } /* * If there are extra headers between IPv6 and ICMPv6, strip * off that header first. */ #ifdef DIAGNOSTIC if (sizeof(struct ip6_hdr) + sizeof(struct icmp6_hdr) > MHLEN) panic("assumption failed in icmp6_reflect"); #endif if (off > sizeof(struct ip6_hdr)) { size_t l; struct ip6_hdr nip6; l = off - sizeof(struct ip6_hdr); m_copydata(m, 0, sizeof(nip6), (caddr_t)&nip6); m_adj(m, l); l = sizeof(struct ip6_hdr) + sizeof(struct icmp6_hdr); if (m->m_len < l) { if ((m = m_pullup(m, l)) == NULL) return; } bcopy((caddr_t)&nip6, mtod(m, caddr_t), sizeof(nip6)); } else /* off == sizeof(struct ip6_hdr) */ { size_t l; l = sizeof(struct ip6_hdr) + sizeof(struct icmp6_hdr); if (m->m_len < l) { if ((m = m_pullup(m, l)) == NULL) return; } } plen = m->m_pkthdr.len - sizeof(struct ip6_hdr); ip6 = mtod(m, struct ip6_hdr *); ip6->ip6_nxt = IPPROTO_ICMPV6; icmp6 = (struct icmp6_hdr *)(ip6 + 1); type = icmp6->icmp6_type; /* keep type for statistics */ code = icmp6->icmp6_code; /* ditto. */ origdst = ip6->ip6_dst; /* * ip6_input() drops a packet if its src is multicast. * So, the src is never multicast. */ ip6->ip6_dst = ip6->ip6_src; /* * If the incoming packet was addressed directly to us (i.e. unicast), * use dst as the src for the reply. * The IN6_IFF_NOTREADY case should be VERY rare, but is possible * (for example) when we encounter an error while forwarding procedure * destined to a duplicated address of ours. * Note that ip6_getdstifaddr() may fail if we are in an error handling * procedure of an outgoing packet of our own, in which case we need * to search in the ifaddr list. */ if (!IN6_IS_ADDR_MULTICAST(&origdst)) { if ((ia = ip6_getdstifaddr(m))) { if (!(ia->ia6_flags & (IN6_IFF_ANYCAST|IN6_IFF_NOTREADY))) srcp = &ia->ia_addr.sin6_addr; } else { struct sockaddr_in6 d; bzero(&d, sizeof(d)); d.sin6_family = AF_INET6; d.sin6_len = sizeof(d); d.sin6_addr = origdst; ia = (struct in6_ifaddr *) ifa_ifwithaddr((struct sockaddr *)&d); if (ia && !(ia->ia6_flags & (IN6_IFF_ANYCAST|IN6_IFF_NOTREADY))) { srcp = &ia->ia_addr.sin6_addr; } } } if (srcp == NULL) { int e; struct sockaddr_in6 sin6; struct route_in6 ro; /* * This case matches to multicasts, our anycast, or unicasts * that we do not own. Select a source address based on the * source address of the erroneous packet. */ bzero(&sin6, sizeof(sin6)); sin6.sin6_family = AF_INET6; sin6.sin6_len = sizeof(sin6); sin6.sin6_addr = ip6->ip6_dst; /* zone ID should be embedded */ bzero(&ro, sizeof(ro)); e = in6_selectsrc(&sin6, NULL, NULL, &ro, NULL, &outif, &src); if (ro.ro_rt) RTFREE(ro.ro_rt); /* XXX: we could use this */ if (e) { char ip6buf[INET6_ADDRSTRLEN]; nd6log((LOG_DEBUG, "icmp6_reflect: source can't be determined: " "dst=%s, error=%d\n", ip6_sprintf(ip6buf, &sin6.sin6_addr), e)); goto bad; } srcp = &src; } ip6->ip6_src = *srcp; ip6->ip6_flow = 0; ip6->ip6_vfc &= ~IPV6_VERSION_MASK; ip6->ip6_vfc |= IPV6_VERSION; ip6->ip6_nxt = IPPROTO_ICMPV6; if (outif) ip6->ip6_hlim = ND_IFINFO(outif)->chlim; else if (m->m_pkthdr.rcvif) { /* XXX: This may not be the outgoing interface */ ip6->ip6_hlim = ND_IFINFO(m->m_pkthdr.rcvif)->chlim; } else ip6->ip6_hlim = V_ip6_defhlim; icmp6->icmp6_cksum = 0; icmp6->icmp6_cksum = in6_cksum(m, IPPROTO_ICMPV6, sizeof(struct ip6_hdr), plen); /* * XXX option handling */ m->m_flags &= ~(M_BCAST|M_MCAST); ip6_output(m, NULL, NULL, 0, NULL, &outif, NULL); if (outif) icmp6_ifoutstat_inc(outif, type, code); if (ia != NULL) ifa_free(&ia->ia_ifa); return; bad: if (ia != NULL) ifa_free(&ia->ia_ifa); m_freem(m); return; } void icmp6_fasttimo(void) { mld_fasttimo(); } void icmp6_slowtimo(void) { mld_slowtimo(); } static const char * icmp6_redirect_diag(struct in6_addr *src6, struct in6_addr *dst6, struct in6_addr *tgt6) { static char buf[1024]; char ip6bufs[INET6_ADDRSTRLEN]; char ip6bufd[INET6_ADDRSTRLEN]; char ip6buft[INET6_ADDRSTRLEN]; snprintf(buf, sizeof(buf), "(src=%s dst=%s tgt=%s)", ip6_sprintf(ip6bufs, src6), ip6_sprintf(ip6bufd, dst6), ip6_sprintf(ip6buft, tgt6)); return buf; } void icmp6_redirect_input(struct mbuf *m, int off) { struct ifnet *ifp; struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); struct nd_redirect *nd_rd; int icmp6len = ntohs(ip6->ip6_plen); char *lladdr = NULL; int lladdrlen = 0; struct rtentry *rt = NULL; int is_router; int is_onlink; struct in6_addr src6 = ip6->ip6_src; struct in6_addr redtgt6; struct in6_addr reddst6; union nd_opts ndopts; char ip6buf[INET6_ADDRSTRLEN]; M_ASSERTPKTHDR(m); KASSERT(m->m_pkthdr.rcvif != NULL, ("%s: no rcvif", __func__)); ifp = m->m_pkthdr.rcvif; /* XXX if we are router, we don't update route by icmp6 redirect */ if (V_ip6_forwarding) goto freeit; if (!V_icmp6_rediraccept) goto freeit; #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, off, icmp6len,); nd_rd = (struct nd_redirect *)((caddr_t)ip6 + off); #else IP6_EXTHDR_GET(nd_rd, struct nd_redirect *, m, off, icmp6len); if (nd_rd == NULL) { ICMP6STAT_INC(icp6s_tooshort); return; } #endif redtgt6 = nd_rd->nd_rd_target; reddst6 = nd_rd->nd_rd_dst; if (in6_setscope(&redtgt6, m->m_pkthdr.rcvif, NULL) || in6_setscope(&reddst6, m->m_pkthdr.rcvif, NULL)) { goto freeit; } /* validation */ if (!IN6_IS_ADDR_LINKLOCAL(&src6)) { nd6log((LOG_ERR, "ICMP6 redirect sent from %s rejected; " "must be from linklocal\n", ip6_sprintf(ip6buf, &src6))); goto bad; } if (ip6->ip6_hlim != 255) { nd6log((LOG_ERR, "ICMP6 redirect sent from %s rejected; " "hlim=%d (must be 255)\n", ip6_sprintf(ip6buf, &src6), ip6->ip6_hlim)); goto bad; } { /* ip6->ip6_src must be equal to gw for icmp6->icmp6_reddst */ struct sockaddr_in6 sin6; struct in6_addr *gw6; bzero(&sin6, sizeof(sin6)); sin6.sin6_family = AF_INET6; sin6.sin6_len = sizeof(struct sockaddr_in6); bcopy(&reddst6, &sin6.sin6_addr, sizeof(reddst6)); rt = in6_rtalloc1((struct sockaddr *)&sin6, 0, 0UL, RT_DEFAULT_FIB); if (rt) { if (rt->rt_gateway == NULL || rt->rt_gateway->sa_family != AF_INET6) { RTFREE_LOCKED(rt); nd6log((LOG_ERR, "ICMP6 redirect rejected; no route " "with inet6 gateway found for redirect dst: %s\n", icmp6_redirect_diag(&src6, &reddst6, &redtgt6))); goto bad; } gw6 = &(((struct sockaddr_in6 *)rt->rt_gateway)->sin6_addr); if (bcmp(&src6, gw6, sizeof(struct in6_addr)) != 0) { RTFREE_LOCKED(rt); nd6log((LOG_ERR, "ICMP6 redirect rejected; " "not equal to gw-for-src=%s (must be same): " "%s\n", ip6_sprintf(ip6buf, gw6), icmp6_redirect_diag(&src6, &reddst6, &redtgt6))); goto bad; } } else { nd6log((LOG_ERR, "ICMP6 redirect rejected; " "no route found for redirect dst: %s\n", icmp6_redirect_diag(&src6, &reddst6, &redtgt6))); goto bad; } RTFREE_LOCKED(rt); rt = NULL; } if (IN6_IS_ADDR_MULTICAST(&reddst6)) { nd6log((LOG_ERR, "ICMP6 redirect rejected; " "redirect dst must be unicast: %s\n", icmp6_redirect_diag(&src6, &reddst6, &redtgt6))); goto bad; } is_router = is_onlink = 0; if (IN6_IS_ADDR_LINKLOCAL(&redtgt6)) is_router = 1; /* router case */ if (bcmp(&redtgt6, &reddst6, sizeof(redtgt6)) == 0) is_onlink = 1; /* on-link destination case */ if (!is_router && !is_onlink) { nd6log((LOG_ERR, "ICMP6 redirect rejected; " "neither router case nor onlink case: %s\n", icmp6_redirect_diag(&src6, &reddst6, &redtgt6))); goto bad; } /* validation passed */ icmp6len -= sizeof(*nd_rd); nd6_option_init(nd_rd + 1, icmp6len, &ndopts); if (nd6_options(&ndopts) < 0) { nd6log((LOG_INFO, "%s: invalid ND option, rejected: %s\n", __func__, icmp6_redirect_diag(&src6, &reddst6, &redtgt6))); /* nd6_options have incremented stats */ goto freeit; } if (ndopts.nd_opts_tgt_lladdr) { lladdr = (char *)(ndopts.nd_opts_tgt_lladdr + 1); lladdrlen = ndopts.nd_opts_tgt_lladdr->nd_opt_len << 3; } if (lladdr && ((ifp->if_addrlen + 2 + 7) & ~7) != lladdrlen) { nd6log((LOG_INFO, "%s: lladdrlen mismatch for %s " "(if %d, icmp6 packet %d): %s\n", __func__, ip6_sprintf(ip6buf, &redtgt6), ifp->if_addrlen, lladdrlen - 2, icmp6_redirect_diag(&src6, &reddst6, &redtgt6))); goto bad; } /* RFC 2461 8.3 */ nd6_cache_lladdr(ifp, &redtgt6, lladdr, lladdrlen, ND_REDIRECT, is_onlink ? ND_REDIRECT_ONLINK : ND_REDIRECT_ROUTER); if (!is_onlink) { /* better router case. perform rtredirect. */ /* perform rtredirect */ struct sockaddr_in6 sdst; struct sockaddr_in6 sgw; struct sockaddr_in6 ssrc; u_int fibnum; bzero(&sdst, sizeof(sdst)); bzero(&sgw, sizeof(sgw)); bzero(&ssrc, sizeof(ssrc)); sdst.sin6_family = sgw.sin6_family = ssrc.sin6_family = AF_INET6; sdst.sin6_len = sgw.sin6_len = ssrc.sin6_len = sizeof(struct sockaddr_in6); bcopy(&redtgt6, &sgw.sin6_addr, sizeof(struct in6_addr)); bcopy(&reddst6, &sdst.sin6_addr, sizeof(struct in6_addr)); bcopy(&src6, &ssrc.sin6_addr, sizeof(struct in6_addr)); for (fibnum = 0; fibnum < rt_numfibs; fibnum++) in6_rtredirect((struct sockaddr *)&sdst, (struct sockaddr *)&sgw, (struct sockaddr *)NULL, RTF_GATEWAY | RTF_HOST, (struct sockaddr *)&ssrc, fibnum); } /* finally update cached route in each socket via pfctlinput */ { struct sockaddr_in6 sdst; bzero(&sdst, sizeof(sdst)); sdst.sin6_family = AF_INET6; sdst.sin6_len = sizeof(struct sockaddr_in6); bcopy(&reddst6, &sdst.sin6_addr, sizeof(struct in6_addr)); pfctlinput(PRC_REDIRECT_HOST, (struct sockaddr *)&sdst); #ifdef IPSEC key_sa_routechange((struct sockaddr *)&sdst); #endif /* IPSEC */ } freeit: m_freem(m); return; bad: ICMP6STAT_INC(icp6s_badredirect); m_freem(m); } void icmp6_redirect_output(struct mbuf *m0, struct rtentry *rt) { struct ifnet *ifp; /* my outgoing interface */ struct in6_addr *ifp_ll6; struct in6_addr *router_ll6; struct ip6_hdr *sip6; /* m0 as struct ip6_hdr */ struct mbuf *m = NULL; /* newly allocated one */ struct m_tag *mtag; struct ip6_hdr *ip6; /* m as struct ip6_hdr */ struct nd_redirect *nd_rd; struct llentry *ln = NULL; size_t maxlen; u_char *p; struct ifnet *outif = NULL; struct sockaddr_in6 src_sa; icmp6_errcount(ND_REDIRECT, 0); /* if we are not router, we don't send icmp6 redirect */ if (!V_ip6_forwarding) goto fail; /* sanity check */ if (!m0 || !rt || !(rt->rt_flags & RTF_UP) || !(ifp = rt->rt_ifp)) goto fail; /* * Address check: * the source address must identify a neighbor, and * the destination address must not be a multicast address * [RFC 2461, sec 8.2] */ sip6 = mtod(m0, struct ip6_hdr *); bzero(&src_sa, sizeof(src_sa)); src_sa.sin6_family = AF_INET6; src_sa.sin6_len = sizeof(src_sa); src_sa.sin6_addr = sip6->ip6_src; if (nd6_is_addr_neighbor(&src_sa, ifp) == 0) goto fail; if (IN6_IS_ADDR_MULTICAST(&sip6->ip6_dst)) goto fail; /* what should we do here? */ /* rate limit */ if (icmp6_ratelimit(&sip6->ip6_src, ND_REDIRECT, 0)) goto fail; /* * Since we are going to append up to 1280 bytes (= IPV6_MMTU), * we almost always ask for an mbuf cluster for simplicity. * (MHLEN < IPV6_MMTU is almost always true) */ #if IPV6_MMTU >= MCLBYTES # error assumption failed about IPV6_MMTU and MCLBYTES #endif m = m_getcl(M_NOWAIT, MT_DATA, M_PKTHDR); if (m == NULL) goto fail; M_SETFIB(m, rt->rt_fibnum); maxlen = M_TRAILINGSPACE(m); maxlen = min(IPV6_MMTU, maxlen); /* just for safety */ if (maxlen < sizeof(struct ip6_hdr) + sizeof(struct icmp6_hdr) + ((sizeof(struct nd_opt_hdr) + ifp->if_addrlen + 7) & ~7)) { goto fail; } { /* get ip6 linklocal address for ifp(my outgoing interface). */ struct in6_ifaddr *ia; if ((ia = in6ifa_ifpforlinklocal(ifp, IN6_IFF_NOTREADY| IN6_IFF_ANYCAST)) == NULL) goto fail; ifp_ll6 = &ia->ia_addr.sin6_addr; /* XXXRW: reference released prematurely. */ ifa_free(&ia->ia_ifa); } /* get ip6 linklocal address for the router. */ if (rt->rt_gateway && (rt->rt_flags & RTF_GATEWAY)) { struct sockaddr_in6 *sin6; sin6 = (struct sockaddr_in6 *)rt->rt_gateway; router_ll6 = &sin6->sin6_addr; if (!IN6_IS_ADDR_LINKLOCAL(router_ll6)) router_ll6 = (struct in6_addr *)NULL; } else router_ll6 = (struct in6_addr *)NULL; /* ip6 */ ip6 = mtod(m, struct ip6_hdr *); ip6->ip6_flow = 0; ip6->ip6_vfc &= ~IPV6_VERSION_MASK; ip6->ip6_vfc |= IPV6_VERSION; /* ip6->ip6_plen will be set later */ ip6->ip6_nxt = IPPROTO_ICMPV6; ip6->ip6_hlim = 255; /* ip6->ip6_src must be linklocal addr for my outgoing if. */ bcopy(ifp_ll6, &ip6->ip6_src, sizeof(struct in6_addr)); bcopy(&sip6->ip6_src, &ip6->ip6_dst, sizeof(struct in6_addr)); /* ND Redirect */ nd_rd = (struct nd_redirect *)(ip6 + 1); nd_rd->nd_rd_type = ND_REDIRECT; nd_rd->nd_rd_code = 0; nd_rd->nd_rd_reserved = 0; if (rt->rt_flags & RTF_GATEWAY) { /* * nd_rd->nd_rd_target must be a link-local address in * better router cases. */ if (!router_ll6) goto fail; bcopy(router_ll6, &nd_rd->nd_rd_target, sizeof(nd_rd->nd_rd_target)); bcopy(&sip6->ip6_dst, &nd_rd->nd_rd_dst, sizeof(nd_rd->nd_rd_dst)); } else { /* make sure redtgt == reddst */ bcopy(&sip6->ip6_dst, &nd_rd->nd_rd_target, sizeof(nd_rd->nd_rd_target)); bcopy(&sip6->ip6_dst, &nd_rd->nd_rd_dst, sizeof(nd_rd->nd_rd_dst)); } p = (u_char *)(nd_rd + 1); if (!router_ll6) goto nolladdropt; { /* target lladdr option */ int len; struct nd_opt_hdr *nd_opt; char *lladdr; IF_AFDATA_RLOCK(ifp); ln = nd6_lookup(router_ll6, 0, ifp); IF_AFDATA_RUNLOCK(ifp); if (ln == NULL) goto nolladdropt; len = sizeof(*nd_opt) + ifp->if_addrlen; len = (len + 7) & ~7; /* round by 8 */ /* safety check */ if (len + (p - (u_char *)ip6) > maxlen) goto nolladdropt; if (ln->la_flags & LLE_VALID) { nd_opt = (struct nd_opt_hdr *)p; nd_opt->nd_opt_type = ND_OPT_TARGET_LINKADDR; nd_opt->nd_opt_len = len >> 3; lladdr = (char *)(nd_opt + 1); bcopy(&ln->ll_addr, lladdr, ifp->if_addrlen); p += len; } } nolladdropt: if (ln != NULL) LLE_RUNLOCK(ln); m->m_pkthdr.len = m->m_len = p - (u_char *)ip6; /* just to be safe */ #ifdef M_DECRYPTED /*not openbsd*/ if (m0->m_flags & M_DECRYPTED) goto noredhdropt; #endif if (p - (u_char *)ip6 > maxlen) goto noredhdropt; { /* redirected header option */ int len; struct nd_opt_rd_hdr *nd_opt_rh; /* * compute the maximum size for icmp6 redirect header option. * XXX room for auth header? */ len = maxlen - (p - (u_char *)ip6); len &= ~7; /* This is just for simplicity. */ if (m0->m_pkthdr.len != m0->m_len) { if (m0->m_next) { m_freem(m0->m_next); m0->m_next = NULL; } m0->m_pkthdr.len = m0->m_len; } /* * Redirected header option spec (RFC2461 4.6.3) talks nothing * about padding/truncate rule for the original IP packet. * From the discussion on IPv6imp in Feb 1999, * the consensus was: * - "attach as much as possible" is the goal * - pad if not aligned (original size can be guessed by * original ip6 header) * Following code adds the padding if it is simple enough, * and truncates if not. */ if (m0->m_next || m0->m_pkthdr.len != m0->m_len) panic("assumption failed in %s:%d", __FILE__, __LINE__); if (len - sizeof(*nd_opt_rh) < m0->m_pkthdr.len) { /* not enough room, truncate */ m0->m_pkthdr.len = m0->m_len = len - sizeof(*nd_opt_rh); } else { /* enough room, pad or truncate */ size_t extra; extra = m0->m_pkthdr.len % 8; if (extra) { /* pad if easy enough, truncate if not */ if (8 - extra <= M_TRAILINGSPACE(m0)) { /* pad */ m0->m_len += (8 - extra); m0->m_pkthdr.len += (8 - extra); } else { /* truncate */ m0->m_pkthdr.len -= extra; m0->m_len -= extra; } } len = m0->m_pkthdr.len + sizeof(*nd_opt_rh); m0->m_pkthdr.len = m0->m_len = len - sizeof(*nd_opt_rh); } nd_opt_rh = (struct nd_opt_rd_hdr *)p; bzero(nd_opt_rh, sizeof(*nd_opt_rh)); nd_opt_rh->nd_opt_rh_type = ND_OPT_REDIRECTED_HEADER; nd_opt_rh->nd_opt_rh_len = len >> 3; p += sizeof(*nd_opt_rh); m->m_pkthdr.len = m->m_len = p - (u_char *)ip6; /* connect m0 to m */ m_tag_delete_chain(m0, NULL); m0->m_flags &= ~M_PKTHDR; m->m_next = m0; m->m_pkthdr.len = m->m_len + m0->m_len; m0 = NULL; } noredhdropt:; if (m0) { m_freem(m0); m0 = NULL; } /* XXX: clear embedded link IDs in the inner header */ in6_clearscope(&sip6->ip6_src); in6_clearscope(&sip6->ip6_dst); in6_clearscope(&nd_rd->nd_rd_target); in6_clearscope(&nd_rd->nd_rd_dst); ip6->ip6_plen = htons(m->m_pkthdr.len - sizeof(struct ip6_hdr)); nd_rd->nd_rd_cksum = 0; nd_rd->nd_rd_cksum = in6_cksum(m, IPPROTO_ICMPV6, sizeof(*ip6), ntohs(ip6->ip6_plen)); if (send_sendso_input_hook != NULL) { mtag = m_tag_get(PACKET_TAG_ND_OUTGOING, sizeof(unsigned short), M_NOWAIT); if (mtag == NULL) goto fail; *(unsigned short *)(mtag + 1) = nd_rd->nd_rd_type; m_tag_prepend(m, mtag); } /* send the packet to outside... */ ip6_output(m, NULL, NULL, 0, NULL, &outif, NULL); if (outif) { icmp6_ifstat_inc(outif, ifs6_out_msg); icmp6_ifstat_inc(outif, ifs6_out_redirect); } ICMP6STAT_INC(icp6s_outhist[ND_REDIRECT]); return; fail: if (m) m_freem(m); if (m0) m_freem(m0); } /* * ICMPv6 socket option processing. */ int icmp6_ctloutput(struct socket *so, struct sockopt *sopt) { int error = 0; int optlen; struct inpcb *inp = sotoinpcb(so); int level, op, optname; if (sopt) { level = sopt->sopt_level; op = sopt->sopt_dir; optname = sopt->sopt_name; optlen = sopt->sopt_valsize; } else level = op = optname = optlen = 0; if (level != IPPROTO_ICMPV6) { return EINVAL; } switch (op) { case PRCO_SETOPT: switch (optname) { case ICMP6_FILTER: { struct icmp6_filter ic6f; if (optlen != sizeof(ic6f)) { error = EMSGSIZE; break; } error = sooptcopyin(sopt, &ic6f, optlen, optlen); if (error == 0) { INP_WLOCK(inp); *inp->in6p_icmp6filt = ic6f; INP_WUNLOCK(inp); } break; } default: error = ENOPROTOOPT; break; } break; case PRCO_GETOPT: switch (optname) { case ICMP6_FILTER: { struct icmp6_filter ic6f; INP_RLOCK(inp); ic6f = *inp->in6p_icmp6filt; INP_RUNLOCK(inp); error = sooptcopyout(sopt, &ic6f, sizeof(ic6f)); break; } default: error = ENOPROTOOPT; break; } break; } return (error); } /* * Perform rate limit check. * Returns 0 if it is okay to send the icmp6 packet. * Returns 1 if the router SHOULD NOT send this icmp6 packet due to rate * limitation. * * XXX per-destination/type check necessary? * * dst - not used at this moment * type - not used at this moment * code - not used at this moment */ static int icmp6_ratelimit(const struct in6_addr *dst, const int type, const int code) { int ret; ret = 0; /* okay to send */ /* PPS limit */ if (!ppsratecheck(&V_icmp6errppslim_last, &V_icmp6errpps_count, V_icmp6errppslim)) { /* The packet is subject to rate limit */ ret++; } return ret; } Index: head/sys/netinet6/in6.c =================================================================== --- head/sys/netinet6/in6.c (revision 253969) +++ head/sys/netinet6/in6.c (revision 253970) @@ -1,2859 +1,2859 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $KAME: in6.c,v 1.259 2002/01/21 11:37:50 keiichi Exp $ */ /*- * Copyright (c) 1982, 1986, 1991, 1993 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)in.c 8.2 (Berkeley) 11/15/93 */ #include __FBSDID("$FreeBSD$"); #include "opt_compat.h" #include "opt_inet.h" #include "opt_inet6.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include VNET_DECLARE(int, icmp6_nodeinfo_oldmcprefix); #define V_icmp6_nodeinfo_oldmcprefix VNET(icmp6_nodeinfo_oldmcprefix) /* * Definitions of some costant IP6 addresses. */ const struct in6_addr in6addr_any = IN6ADDR_ANY_INIT; const struct in6_addr in6addr_loopback = IN6ADDR_LOOPBACK_INIT; const struct in6_addr in6addr_nodelocal_allnodes = IN6ADDR_NODELOCAL_ALLNODES_INIT; const struct in6_addr in6addr_linklocal_allnodes = IN6ADDR_LINKLOCAL_ALLNODES_INIT; const struct in6_addr in6addr_linklocal_allrouters = IN6ADDR_LINKLOCAL_ALLROUTERS_INIT; const struct in6_addr in6addr_linklocal_allv2routers = IN6ADDR_LINKLOCAL_ALLV2ROUTERS_INIT; const struct in6_addr in6mask0 = IN6MASK0; const struct in6_addr in6mask32 = IN6MASK32; const struct in6_addr in6mask64 = IN6MASK64; const struct in6_addr in6mask96 = IN6MASK96; const struct in6_addr in6mask128 = IN6MASK128; const struct sockaddr_in6 sa6_any = { sizeof(sa6_any), AF_INET6, 0, 0, IN6ADDR_ANY_INIT, 0 }; static int in6_lifaddr_ioctl(struct socket *, u_long, caddr_t, struct ifnet *, struct thread *); static int in6_ifinit(struct ifnet *, struct in6_ifaddr *, struct sockaddr_in6 *, int); static void in6_unlink_ifa(struct in6_ifaddr *, struct ifnet *); int (*faithprefix_p)(struct in6_addr *); #define ifa2ia6(ifa) ((struct in6_ifaddr *)(ifa)) #define ia62ifa(ia6) (&((ia6)->ia_ifa)) void in6_ifaddloop(struct ifaddr *ifa) { struct sockaddr_dl gateway; struct sockaddr_in6 mask, addr; struct rtentry rt; struct in6_ifaddr *ia; struct ifnet *ifp; struct llentry *ln; ia = ifa2ia6(ifa); ifp = ifa->ifa_ifp; IF_AFDATA_LOCK(ifp); ifa->ifa_rtrequest = nd6_rtrequest; ln = lla_lookup(LLTABLE6(ifp), (LLE_CREATE | LLE_IFADDR | LLE_EXCLUSIVE), (struct sockaddr *)&ia->ia_addr); IF_AFDATA_UNLOCK(ifp); if (ln != NULL) { ln->la_expire = 0; /* for IPv6 this means permanent */ ln->ln_state = ND6_LLINFO_REACHABLE; /* * initialize for rtmsg generation */ bzero(&gateway, sizeof(gateway)); gateway.sdl_len = sizeof(gateway); gateway.sdl_family = AF_LINK; gateway.sdl_nlen = 0; gateway.sdl_alen = 6; memcpy(gateway.sdl_data, &ln->ll_addr.mac_aligned, sizeof(ln->ll_addr)); LLE_WUNLOCK(ln); } bzero(&rt, sizeof(rt)); rt.rt_gateway = (struct sockaddr *)&gateway; memcpy(&mask, &ia->ia_prefixmask, sizeof(ia->ia_prefixmask)); memcpy(&addr, &ia->ia_addr, sizeof(ia->ia_addr)); rt_mask(&rt) = (struct sockaddr *)&mask; rt_key(&rt) = (struct sockaddr *)&addr; rt.rt_flags = RTF_UP | RTF_HOST | RTF_STATIC; /* Announce arrival of local address to all FIBs. */ rt_newaddrmsg(RTM_ADD, ifa, 0, &rt); } void in6_ifremloop(struct ifaddr *ifa) { struct sockaddr_dl gateway; struct sockaddr_in6 mask, addr; struct rtentry rt0; struct in6_ifaddr *ia; struct ifnet *ifp; ia = ifa2ia6(ifa); ifp = ifa->ifa_ifp; memcpy(&addr, &ia->ia_addr, sizeof(ia->ia_addr)); memcpy(&mask, &ia->ia_prefixmask, sizeof(ia->ia_prefixmask)); lltable_prefix_free(AF_INET6, (struct sockaddr *)&addr, (struct sockaddr *)&mask, LLE_STATIC); /* * initialize for rtmsg generation */ bzero(&gateway, sizeof(gateway)); gateway.sdl_len = sizeof(gateway); gateway.sdl_family = AF_LINK; gateway.sdl_nlen = 0; gateway.sdl_alen = ifp->if_addrlen; bzero(&rt0, sizeof(rt0)); rt0.rt_gateway = (struct sockaddr *)&gateway; rt_mask(&rt0) = (struct sockaddr *)&mask; rt_key(&rt0) = (struct sockaddr *)&addr; rt0.rt_flags = RTF_HOST | RTF_STATIC; /* Announce removal of local address to all FIBs. */ rt_newaddrmsg(RTM_DELETE, ifa, 0, &rt0); } int in6_mask2len(struct in6_addr *mask, u_char *lim0) { int x = 0, y; u_char *lim = lim0, *p; /* ignore the scope_id part */ if (lim0 == NULL || lim0 - (u_char *)mask > sizeof(*mask)) lim = (u_char *)mask + sizeof(*mask); for (p = (u_char *)mask; p < lim; x++, p++) { if (*p != 0xff) break; } y = 0; if (p < lim) { for (y = 0; y < 8; y++) { if ((*p & (0x80 >> y)) == 0) break; } } /* * when the limit pointer is given, do a stricter check on the * remaining bits. */ if (p < lim) { if (y != 0 && (*p & (0x00ff >> y)) != 0) return (-1); for (p = p + 1; p < lim; p++) if (*p != 0) return (-1); } return x * 8 + y; } #ifdef COMPAT_FREEBSD32 struct in6_ndifreq32 { char ifname[IFNAMSIZ]; uint32_t ifindex; }; #define SIOCGDEFIFACE32_IN6 _IOWR('i', 86, struct in6_ndifreq32) #endif int in6_control(struct socket *so, u_long cmd, caddr_t data, struct ifnet *ifp, struct thread *td) { struct in6_ifreq *ifr = (struct in6_ifreq *)data; struct in6_ifaddr *ia = NULL; struct in6_aliasreq *ifra = (struct in6_aliasreq *)data; struct sockaddr_in6 *sa6; int carp_attached = 0; int error; u_long ocmd = cmd; /* * Compat to make pre-10.x ifconfig(8) operable. */ if (cmd == OSIOCAIFADDR_IN6) cmd = SIOCAIFADDR_IN6; switch (cmd) { case SIOCGETSGCNT_IN6: case SIOCGETMIFCNT_IN6: /* * XXX mrt_ioctl has a 3rd, unused, FIB argument in route.c. * We cannot see how that would be needed, so do not adjust the * KPI blindly; more likely should clean up the IPv4 variant. */ return (mrt6_ioctl ? mrt6_ioctl(cmd, data) : EOPNOTSUPP); } switch(cmd) { case SIOCAADDRCTL_POLICY: case SIOCDADDRCTL_POLICY: if (td != NULL) { error = priv_check(td, PRIV_NETINET_ADDRCTRL6); if (error) return (error); } return (in6_src_ioctl(cmd, data)); } if (ifp == NULL) return (EOPNOTSUPP); switch (cmd) { case SIOCSNDFLUSH_IN6: case SIOCSPFXFLUSH_IN6: case SIOCSRTRFLUSH_IN6: case SIOCSDEFIFACE_IN6: case SIOCSIFINFO_FLAGS: case SIOCSIFINFO_IN6: if (td != NULL) { error = priv_check(td, PRIV_NETINET_ND6); if (error) return (error); } /* FALLTHROUGH */ case OSIOCGIFINFO_IN6: case SIOCGIFINFO_IN6: case SIOCGDRLST_IN6: case SIOCGPRLST_IN6: case SIOCGNBRINFO_IN6: case SIOCGDEFIFACE_IN6: return (nd6_ioctl(cmd, data, ifp)); #ifdef COMPAT_FREEBSD32 case SIOCGDEFIFACE32_IN6: { struct in6_ndifreq ndif; struct in6_ndifreq32 *ndif32; error = nd6_ioctl(SIOCGDEFIFACE_IN6, (caddr_t)&ndif, ifp); if (error) return (error); ndif32 = (struct in6_ndifreq32 *)data; ndif32->ifindex = ndif.ifindex; return (0); } #endif } switch (cmd) { case SIOCSIFPREFIX_IN6: case SIOCDIFPREFIX_IN6: case SIOCAIFPREFIX_IN6: case SIOCCIFPREFIX_IN6: case SIOCSGIFPREFIX_IN6: case SIOCGIFPREFIX_IN6: log(LOG_NOTICE, "prefix ioctls are now invalidated. " "please use ifconfig.\n"); return (EOPNOTSUPP); } switch (cmd) { case SIOCSSCOPE6: if (td != NULL) { error = priv_check(td, PRIV_NETINET_SCOPE6); if (error) return (error); } return (scope6_set(ifp, (struct scope6_id *)ifr->ifr_ifru.ifru_scope_id)); case SIOCGSCOPE6: return (scope6_get(ifp, (struct scope6_id *)ifr->ifr_ifru.ifru_scope_id)); case SIOCGSCOPE6DEF: return (scope6_get_default((struct scope6_id *) ifr->ifr_ifru.ifru_scope_id)); } switch (cmd) { case SIOCALIFADDR: if (td != NULL) { error = priv_check(td, PRIV_NET_ADDIFADDR); if (error) return (error); } return in6_lifaddr_ioctl(so, cmd, data, ifp, td); case SIOCDLIFADDR: if (td != NULL) { error = priv_check(td, PRIV_NET_DELIFADDR); if (error) return (error); } /* FALLTHROUGH */ case SIOCGLIFADDR: return in6_lifaddr_ioctl(so, cmd, data, ifp, td); } /* * Find address for this interface, if it exists. * * In netinet code, we have checked ifra_addr in SIOCSIF*ADDR operation * only, and used the first interface address as the target of other * operations (without checking ifra_addr). This was because netinet * code/API assumed at most 1 interface address per interface. * Since IPv6 allows a node to assign multiple addresses * on a single interface, we almost always look and check the * presence of ifra_addr, and reject invalid ones here. * It also decreases duplicated code among SIOC*_IN6 operations. */ switch (cmd) { case SIOCAIFADDR_IN6: case SIOCSIFPHYADDR_IN6: sa6 = &ifra->ifra_addr; break; case SIOCSIFADDR_IN6: case SIOCGIFADDR_IN6: case SIOCSIFDSTADDR_IN6: case SIOCSIFNETMASK_IN6: case SIOCGIFDSTADDR_IN6: case SIOCGIFNETMASK_IN6: case SIOCDIFADDR_IN6: case SIOCGIFPSRCADDR_IN6: case SIOCGIFPDSTADDR_IN6: case SIOCGIFAFLAG_IN6: case SIOCSNDFLUSH_IN6: case SIOCSPFXFLUSH_IN6: case SIOCSRTRFLUSH_IN6: case SIOCGIFALIFETIME_IN6: case SIOCSIFALIFETIME_IN6: case SIOCGIFSTAT_IN6: case SIOCGIFSTAT_ICMP6: sa6 = &ifr->ifr_addr; break; default: sa6 = NULL; break; } if (sa6 && sa6->sin6_family == AF_INET6) { if (sa6->sin6_scope_id != 0) error = sa6_embedscope(sa6, 0); else error = in6_setscope(&sa6->sin6_addr, ifp, NULL); if (error != 0) return (error); if (td != NULL && (error = prison_check_ip6(td->td_ucred, &sa6->sin6_addr)) != 0) return (error); ia = in6ifa_ifpwithaddr(ifp, &sa6->sin6_addr); } else ia = NULL; switch (cmd) { case SIOCSIFADDR_IN6: case SIOCSIFDSTADDR_IN6: case SIOCSIFNETMASK_IN6: /* * Since IPv6 allows a node to assign multiple addresses * on a single interface, SIOCSIFxxx ioctls are deprecated. */ /* we decided to obsolete this command (20000704) */ error = EINVAL; goto out; case SIOCDIFADDR_IN6: /* * for IPv4, we look for existing in_ifaddr here to allow * "ifconfig if0 delete" to remove the first IPv4 address on * the interface. For IPv6, as the spec allows multiple * interface address from the day one, we consider "remove the * first one" semantics to be not preferable. */ if (ia == NULL) { error = EADDRNOTAVAIL; goto out; } /* FALLTHROUGH */ case SIOCAIFADDR_IN6: /* * We always require users to specify a valid IPv6 address for * the corresponding operation. */ if (ifra->ifra_addr.sin6_family != AF_INET6 || ifra->ifra_addr.sin6_len != sizeof(struct sockaddr_in6)) { error = EAFNOSUPPORT; goto out; } if (td != NULL) { error = priv_check(td, (cmd == SIOCDIFADDR_IN6) ? PRIV_NET_DELIFADDR : PRIV_NET_ADDIFADDR); if (error) goto out; } break; case SIOCGIFADDR_IN6: /* This interface is basically deprecated. use SIOCGIFCONF. */ /* FALLTHROUGH */ case SIOCGIFAFLAG_IN6: case SIOCGIFNETMASK_IN6: case SIOCGIFDSTADDR_IN6: case SIOCGIFALIFETIME_IN6: /* must think again about its semantics */ if (ia == NULL) { error = EADDRNOTAVAIL; goto out; } break; case SIOCSIFALIFETIME_IN6: { struct in6_addrlifetime *lt; if (td != NULL) { error = priv_check(td, PRIV_NETINET_ALIFETIME6); if (error) goto out; } if (ia == NULL) { error = EADDRNOTAVAIL; goto out; } /* sanity for overflow - beware unsigned */ lt = &ifr->ifr_ifru.ifru_lifetime; if (lt->ia6t_vltime != ND6_INFINITE_LIFETIME && - lt->ia6t_vltime + time_second < time_second) { + lt->ia6t_vltime + time_uptime < time_uptime) { error = EINVAL; goto out; } if (lt->ia6t_pltime != ND6_INFINITE_LIFETIME && - lt->ia6t_pltime + time_second < time_second) { + lt->ia6t_pltime + time_uptime < time_uptime) { error = EINVAL; goto out; } break; } } switch (cmd) { case SIOCGIFADDR_IN6: ifr->ifr_addr = ia->ia_addr; if ((error = sa6_recoverscope(&ifr->ifr_addr)) != 0) goto out; break; case SIOCGIFDSTADDR_IN6: if ((ifp->if_flags & IFF_POINTOPOINT) == 0) { error = EINVAL; goto out; } /* * XXX: should we check if ifa_dstaddr is NULL and return * an error? */ ifr->ifr_dstaddr = ia->ia_dstaddr; if ((error = sa6_recoverscope(&ifr->ifr_dstaddr)) != 0) goto out; break; case SIOCGIFNETMASK_IN6: ifr->ifr_addr = ia->ia_prefixmask; break; case SIOCGIFAFLAG_IN6: ifr->ifr_ifru.ifru_flags6 = ia->ia6_flags; break; case SIOCGIFSTAT_IN6: if (ifp == NULL) { error = EINVAL; goto out; } COUNTER_ARRAY_COPY(((struct in6_ifextra *) ifp->if_afdata[AF_INET6])->in6_ifstat, &ifr->ifr_ifru.ifru_stat, sizeof(struct in6_ifstat) / sizeof(uint64_t)); break; case SIOCGIFSTAT_ICMP6: if (ifp == NULL) { error = EINVAL; goto out; } COUNTER_ARRAY_COPY(((struct in6_ifextra *) ifp->if_afdata[AF_INET6])->icmp6_ifstat, &ifr->ifr_ifru.ifru_icmp6stat, sizeof(struct icmp6_ifstat) / sizeof(uint64_t)); break; case SIOCGIFALIFETIME_IN6: ifr->ifr_ifru.ifru_lifetime = ia->ia6_lifetime; if (ia->ia6_lifetime.ia6t_vltime != ND6_INFINITE_LIFETIME) { time_t maxexpire; struct in6_addrlifetime *retlt = &ifr->ifr_ifru.ifru_lifetime; /* * XXX: adjust expiration time assuming time_t is * signed. */ maxexpire = (-1) & ~((time_t)1 << ((sizeof(maxexpire) * 8) - 1)); if (ia->ia6_lifetime.ia6t_vltime < maxexpire - ia->ia6_updatetime) { retlt->ia6t_expire = ia->ia6_updatetime + ia->ia6_lifetime.ia6t_vltime; } else retlt->ia6t_expire = maxexpire; } if (ia->ia6_lifetime.ia6t_pltime != ND6_INFINITE_LIFETIME) { time_t maxexpire; struct in6_addrlifetime *retlt = &ifr->ifr_ifru.ifru_lifetime; /* * XXX: adjust expiration time assuming time_t is * signed. */ maxexpire = (-1) & ~((time_t)1 << ((sizeof(maxexpire) * 8) - 1)); if (ia->ia6_lifetime.ia6t_pltime < maxexpire - ia->ia6_updatetime) { retlt->ia6t_preferred = ia->ia6_updatetime + ia->ia6_lifetime.ia6t_pltime; } else retlt->ia6t_preferred = maxexpire; } break; case SIOCSIFALIFETIME_IN6: ia->ia6_lifetime = ifr->ifr_ifru.ifru_lifetime; /* for sanity */ if (ia->ia6_lifetime.ia6t_vltime != ND6_INFINITE_LIFETIME) { ia->ia6_lifetime.ia6t_expire = - time_second + ia->ia6_lifetime.ia6t_vltime; + time_uptime + ia->ia6_lifetime.ia6t_vltime; } else ia->ia6_lifetime.ia6t_expire = 0; if (ia->ia6_lifetime.ia6t_pltime != ND6_INFINITE_LIFETIME) { ia->ia6_lifetime.ia6t_preferred = - time_second + ia->ia6_lifetime.ia6t_pltime; + time_uptime + ia->ia6_lifetime.ia6t_pltime; } else ia->ia6_lifetime.ia6t_preferred = 0; break; case SIOCAIFADDR_IN6: { int i; struct nd_prefixctl pr0; struct nd_prefix *pr; /* * first, make or update the interface address structure, * and link it to the list. */ if ((error = in6_update_ifa(ifp, ifra, ia, 0)) != 0) goto out; if (ia != NULL) ifa_free(&ia->ia_ifa); if ((ia = in6ifa_ifpwithaddr(ifp, &ifra->ifra_addr.sin6_addr)) == NULL) { /* * this can happen when the user specify the 0 valid * lifetime. */ break; } if (cmd == ocmd && ifra->ifra_vhid > 0) { if (carp_attach_p != NULL) error = (*carp_attach_p)(&ia->ia_ifa, ifra->ifra_vhid); else error = EPROTONOSUPPORT; if (error) goto out; else carp_attached = 1; } /* * then, make the prefix on-link on the interface. * XXX: we'd rather create the prefix before the address, but * we need at least one address to install the corresponding * interface route, so we configure the address first. */ /* * convert mask to prefix length (prefixmask has already * been validated in in6_update_ifa(). */ bzero(&pr0, sizeof(pr0)); pr0.ndpr_ifp = ifp; pr0.ndpr_plen = in6_mask2len(&ifra->ifra_prefixmask.sin6_addr, NULL); if (pr0.ndpr_plen == 128) { break; /* we don't need to install a host route. */ } pr0.ndpr_prefix = ifra->ifra_addr; /* apply the mask for safety. */ for (i = 0; i < 4; i++) { pr0.ndpr_prefix.sin6_addr.s6_addr32[i] &= ifra->ifra_prefixmask.sin6_addr.s6_addr32[i]; } /* * XXX: since we don't have an API to set prefix (not address) * lifetimes, we just use the same lifetimes as addresses. * The (temporarily) installed lifetimes can be overridden by * later advertised RAs (when accept_rtadv is non 0), which is * an intended behavior. */ pr0.ndpr_raf_onlink = 1; /* should be configurable? */ pr0.ndpr_raf_auto = ((ifra->ifra_flags & IN6_IFF_AUTOCONF) != 0); pr0.ndpr_vltime = ifra->ifra_lifetime.ia6t_vltime; pr0.ndpr_pltime = ifra->ifra_lifetime.ia6t_pltime; /* add the prefix if not yet. */ if ((pr = nd6_prefix_lookup(&pr0)) == NULL) { /* * nd6_prelist_add will install the corresponding * interface route. */ if ((error = nd6_prelist_add(&pr0, NULL, &pr)) != 0) { if (carp_attached) (*carp_detach_p)(&ia->ia_ifa); goto out; } if (pr == NULL) { if (carp_attached) (*carp_detach_p)(&ia->ia_ifa); log(LOG_ERR, "nd6_prelist_add succeeded but " "no prefix\n"); error = EINVAL; goto out; } } /* relate the address to the prefix */ if (ia->ia6_ndpr == NULL) { ia->ia6_ndpr = pr; pr->ndpr_refcnt++; /* * If this is the first autoconf address from the * prefix, create a temporary address as well * (when required). */ if ((ia->ia6_flags & IN6_IFF_AUTOCONF) && V_ip6_use_tempaddr && pr->ndpr_refcnt == 1) { int e; if ((e = in6_tmpifadd(ia, 1, 0)) != 0) { log(LOG_NOTICE, "in6_control: failed " "to create a temporary address, " "errno=%d\n", e); } } } /* * this might affect the status of autoconfigured addresses, * that is, this address might make other addresses detached. */ pfxlist_onlink_check(); if (error == 0 && ia) { if (ND_IFINFO(ifp)->flags & ND6_IFF_IFDISABLED) { /* * Try to clear the flag when a new * IPv6 address is added onto an * IFDISABLED interface and it * succeeds. */ struct in6_ndireq nd; memset(&nd, 0, sizeof(nd)); nd.ndi.flags = ND_IFINFO(ifp)->flags; nd.ndi.flags &= ~ND6_IFF_IFDISABLED; if (nd6_ioctl(SIOCSIFINFO_FLAGS, (caddr_t)&nd, ifp) < 0) log(LOG_NOTICE, "SIOCAIFADDR_IN6: " "SIOCSIFINFO_FLAGS for -ifdisabled " "failed."); /* * Ignore failure of clearing the flag * intentionally. The failure means * address duplication was detected. */ } EVENTHANDLER_INVOKE(ifaddr_event, ifp); } break; } case SIOCDIFADDR_IN6: { struct nd_prefix *pr; /* * If the address being deleted is the only one that owns * the corresponding prefix, expire the prefix as well. * XXX: theoretically, we don't have to worry about such * relationship, since we separate the address management * and the prefix management. We do this, however, to provide * as much backward compatibility as possible in terms of * the ioctl operation. * Note that in6_purgeaddr() will decrement ndpr_refcnt. */ pr = ia->ia6_ndpr; in6_purgeaddr(&ia->ia_ifa); if (pr && pr->ndpr_refcnt == 0) prelist_remove(pr); EVENTHANDLER_INVOKE(ifaddr_event, ifp); break; } default: if (ifp == NULL || ifp->if_ioctl == 0) { error = EOPNOTSUPP; goto out; } error = (*ifp->if_ioctl)(ifp, cmd, data); goto out; } error = 0; out: if (ia != NULL) ifa_free(&ia->ia_ifa); return (error); } /* * Join necessary multicast groups. Factored out from in6_update_ifa(). * This entire work should only be done once, for the default FIB. */ static int in6_update_ifa_join_mc(struct ifnet *ifp, struct in6_aliasreq *ifra, struct in6_ifaddr *ia, int flags, struct in6_multi **in6m_sol) { char ip6buf[INET6_ADDRSTRLEN]; struct sockaddr_in6 mltaddr, mltmask; struct in6_addr llsol; struct in6_multi_mship *imm; struct rtentry *rt; int delay, error; KASSERT(in6m_sol != NULL, ("%s: in6m_sol is NULL", __func__)); /* Join solicited multicast addr for new host id. */ bzero(&llsol, sizeof(struct in6_addr)); llsol.s6_addr32[0] = IPV6_ADDR_INT32_MLL; llsol.s6_addr32[1] = 0; llsol.s6_addr32[2] = htonl(1); llsol.s6_addr32[3] = ifra->ifra_addr.sin6_addr.s6_addr32[3]; llsol.s6_addr8[12] = 0xff; if ((error = in6_setscope(&llsol, ifp, NULL)) != 0) { /* XXX: should not happen */ log(LOG_ERR, "%s: in6_setscope failed\n", __func__); goto cleanup; } delay = 0; if ((flags & IN6_IFAUPDATE_DADDELAY)) { /* * We need a random delay for DAD on the address being * configured. It also means delaying transmission of the * corresponding MLD report to avoid report collision. * [RFC 4861, Section 6.3.7] */ delay = arc4random() % (MAX_RTR_SOLICITATION_DELAY * hz); } imm = in6_joingroup(ifp, &llsol, &error, delay); if (imm == NULL) { nd6log((LOG_WARNING, "%s: addmulti failed for %s on %s " "(errno=%d)\n", __func__, ip6_sprintf(ip6buf, &llsol), if_name(ifp), error)); goto cleanup; } LIST_INSERT_HEAD(&ia->ia6_memberships, imm, i6mm_chain); *in6m_sol = imm->i6mm_maddr; bzero(&mltmask, sizeof(mltmask)); mltmask.sin6_len = sizeof(struct sockaddr_in6); mltmask.sin6_family = AF_INET6; mltmask.sin6_addr = in6mask32; #define MLTMASK_LEN 4 /* mltmask's masklen (=32bit=4octet) */ /* * Join link-local all-nodes address. */ bzero(&mltaddr, sizeof(mltaddr)); mltaddr.sin6_len = sizeof(struct sockaddr_in6); mltaddr.sin6_family = AF_INET6; mltaddr.sin6_addr = in6addr_linklocal_allnodes; if ((error = in6_setscope(&mltaddr.sin6_addr, ifp, NULL)) != 0) goto cleanup; /* XXX: should not fail */ /* * XXX: do we really need this automatic routes? We should probably * reconsider this stuff. Most applications actually do not need the * routes, since they usually specify the outgoing interface. */ rt = in6_rtalloc1((struct sockaddr *)&mltaddr, 0, 0UL, RT_DEFAULT_FIB); if (rt != NULL) { /* XXX: only works in !SCOPEDROUTING case. */ if (memcmp(&mltaddr.sin6_addr, &((struct sockaddr_in6 *)rt_key(rt))->sin6_addr, MLTMASK_LEN)) { RTFREE_LOCKED(rt); rt = NULL; } } if (rt == NULL) { error = in6_rtrequest(RTM_ADD, (struct sockaddr *)&mltaddr, (struct sockaddr *)&ia->ia_addr, (struct sockaddr *)&mltmask, RTF_UP, (struct rtentry **)0, RT_DEFAULT_FIB); if (error) goto cleanup; } else RTFREE_LOCKED(rt); imm = in6_joingroup(ifp, &mltaddr.sin6_addr, &error, 0); if (imm == NULL) { nd6log((LOG_WARNING, "%s: addmulti failed for %s on %s " "(errno=%d)\n", __func__, ip6_sprintf(ip6buf, &mltaddr.sin6_addr), if_name(ifp), error)); goto cleanup; } LIST_INSERT_HEAD(&ia->ia6_memberships, imm, i6mm_chain); /* * Join node information group address. */ delay = 0; if ((flags & IN6_IFAUPDATE_DADDELAY)) { /* * The spec does not say anything about delay for this group, * but the same logic should apply. */ delay = arc4random() % (MAX_RTR_SOLICITATION_DELAY * hz); } if (in6_nigroup(ifp, NULL, -1, &mltaddr.sin6_addr) == 0) { /* XXX jinmei */ imm = in6_joingroup(ifp, &mltaddr.sin6_addr, &error, delay); if (imm == NULL) nd6log((LOG_WARNING, "%s: addmulti failed for %s on %s " "(errno=%d)\n", __func__, ip6_sprintf(ip6buf, &mltaddr.sin6_addr), if_name(ifp), error)); /* XXX not very fatal, go on... */ else LIST_INSERT_HEAD(&ia->ia6_memberships, imm, i6mm_chain); } if (V_icmp6_nodeinfo_oldmcprefix && in6_nigroup_oldmcprefix(ifp, NULL, -1, &mltaddr.sin6_addr) == 0) { imm = in6_joingroup(ifp, &mltaddr.sin6_addr, &error, delay); if (imm == NULL) nd6log((LOG_WARNING, "%s: addmulti failed for %s on %s " "(errno=%d)\n", __func__, ip6_sprintf(ip6buf, &mltaddr.sin6_addr), if_name(ifp), error)); /* XXX not very fatal, go on... */ else LIST_INSERT_HEAD(&ia->ia6_memberships, imm, i6mm_chain); } /* * Join interface-local all-nodes address. * (ff01::1%ifN, and ff01::%ifN/32) */ mltaddr.sin6_addr = in6addr_nodelocal_allnodes; if ((error = in6_setscope(&mltaddr.sin6_addr, ifp, NULL)) != 0) goto cleanup; /* XXX: should not fail */ /* XXX: again, do we really need the route? */ rt = in6_rtalloc1((struct sockaddr *)&mltaddr, 0, 0UL, RT_DEFAULT_FIB); if (rt != NULL) { if (memcmp(&mltaddr.sin6_addr, &((struct sockaddr_in6 *)rt_key(rt))->sin6_addr, MLTMASK_LEN)) { RTFREE_LOCKED(rt); rt = NULL; } } if (rt == NULL) { error = in6_rtrequest(RTM_ADD, (struct sockaddr *)&mltaddr, (struct sockaddr *)&ia->ia_addr, (struct sockaddr *)&mltmask, RTF_UP, (struct rtentry **)0, RT_DEFAULT_FIB); if (error) goto cleanup; } else RTFREE_LOCKED(rt); imm = in6_joingroup(ifp, &mltaddr.sin6_addr, &error, 0); if (imm == NULL) { nd6log((LOG_WARNING, "%s: addmulti failed for %s on %s " "(errno=%d)\n", __func__, ip6_sprintf(ip6buf, &mltaddr.sin6_addr), if_name(ifp), error)); goto cleanup; } LIST_INSERT_HEAD(&ia->ia6_memberships, imm, i6mm_chain); #undef MLTMASK_LEN cleanup: return (error); } /* * Update parameters of an IPv6 interface address. * If necessary, a new entry is created and linked into address chains. * This function is separated from in6_control(). */ int in6_update_ifa(struct ifnet *ifp, struct in6_aliasreq *ifra, struct in6_ifaddr *ia, int flags) { int error = 0, hostIsNew = 0, plen = -1; struct sockaddr_in6 dst6; struct in6_addrlifetime *lt; struct in6_multi *in6m_sol; int delay; char ip6buf[INET6_ADDRSTRLEN]; /* Validate parameters */ if (ifp == NULL || ifra == NULL) /* this maybe redundant */ return (EINVAL); /* * The destination address for a p2p link must have a family * of AF_UNSPEC or AF_INET6. */ if ((ifp->if_flags & IFF_POINTOPOINT) != 0 && ifra->ifra_dstaddr.sin6_family != AF_INET6 && ifra->ifra_dstaddr.sin6_family != AF_UNSPEC) return (EAFNOSUPPORT); /* * validate ifra_prefixmask. don't check sin6_family, netmask * does not carry fields other than sin6_len. */ if (ifra->ifra_prefixmask.sin6_len > sizeof(struct sockaddr_in6)) return (EINVAL); /* * Because the IPv6 address architecture is classless, we require * users to specify a (non 0) prefix length (mask) for a new address. * We also require the prefix (when specified) mask is valid, and thus * reject a non-consecutive mask. */ if (ia == NULL && ifra->ifra_prefixmask.sin6_len == 0) return (EINVAL); if (ifra->ifra_prefixmask.sin6_len != 0) { plen = in6_mask2len(&ifra->ifra_prefixmask.sin6_addr, (u_char *)&ifra->ifra_prefixmask + ifra->ifra_prefixmask.sin6_len); if (plen <= 0) return (EINVAL); } else { /* * In this case, ia must not be NULL. We just use its prefix * length. */ plen = in6_mask2len(&ia->ia_prefixmask.sin6_addr, NULL); } /* * If the destination address on a p2p interface is specified, * and the address is a scoped one, validate/set the scope * zone identifier. */ dst6 = ifra->ifra_dstaddr; if ((ifp->if_flags & (IFF_POINTOPOINT|IFF_LOOPBACK)) != 0 && (dst6.sin6_family == AF_INET6)) { struct in6_addr in6_tmp; u_int32_t zoneid; in6_tmp = dst6.sin6_addr; if (in6_setscope(&in6_tmp, ifp, &zoneid)) return (EINVAL); /* XXX: should be impossible */ if (dst6.sin6_scope_id != 0) { if (dst6.sin6_scope_id != zoneid) return (EINVAL); } else /* user omit to specify the ID. */ dst6.sin6_scope_id = zoneid; /* convert into the internal form */ if (sa6_embedscope(&dst6, 0)) return (EINVAL); /* XXX: should be impossible */ } /* * The destination address can be specified only for a p2p or a * loopback interface. If specified, the corresponding prefix length * must be 128. */ if (ifra->ifra_dstaddr.sin6_family == AF_INET6) { if ((ifp->if_flags & (IFF_POINTOPOINT|IFF_LOOPBACK)) == 0) { /* XXX: noisy message */ nd6log((LOG_INFO, "in6_update_ifa: a destination can " "be specified for a p2p or a loopback IF only\n")); return (EINVAL); } if (plen != 128) { nd6log((LOG_INFO, "in6_update_ifa: prefixlen should " "be 128 when dstaddr is specified\n")); return (EINVAL); } } /* lifetime consistency check */ lt = &ifra->ifra_lifetime; if (lt->ia6t_pltime > lt->ia6t_vltime) return (EINVAL); if (lt->ia6t_vltime == 0) { /* * the following log might be noisy, but this is a typical * configuration mistake or a tool's bug. */ nd6log((LOG_INFO, "in6_update_ifa: valid lifetime is 0 for %s\n", ip6_sprintf(ip6buf, &ifra->ifra_addr.sin6_addr))); if (ia == NULL) return (0); /* there's nothing to do */ } /* * If this is a new address, allocate a new ifaddr and link it * into chains. */ if (ia == NULL) { hostIsNew = 1; /* * When in6_update_ifa() is called in a process of a received * RA, it is called under an interrupt context. So, we should * call malloc with M_NOWAIT. */ ia = (struct in6_ifaddr *) malloc(sizeof(*ia), M_IFADDR, M_NOWAIT); if (ia == NULL) return (ENOBUFS); bzero((caddr_t)ia, sizeof(*ia)); ifa_init(&ia->ia_ifa); LIST_INIT(&ia->ia6_memberships); /* Initialize the address and masks, and put time stamp */ ia->ia_ifa.ifa_addr = (struct sockaddr *)&ia->ia_addr; ia->ia_addr.sin6_family = AF_INET6; ia->ia_addr.sin6_len = sizeof(ia->ia_addr); - ia->ia6_createtime = time_second; + ia->ia6_createtime = time_uptime; if ((ifp->if_flags & (IFF_POINTOPOINT | IFF_LOOPBACK)) != 0) { /* * XXX: some functions expect that ifa_dstaddr is not * NULL for p2p interfaces. */ ia->ia_ifa.ifa_dstaddr = (struct sockaddr *)&ia->ia_dstaddr; } else { ia->ia_ifa.ifa_dstaddr = NULL; } ia->ia_ifa.ifa_netmask = (struct sockaddr *)&ia->ia_prefixmask; ia->ia_ifp = ifp; ifa_ref(&ia->ia_ifa); /* if_addrhead */ IF_ADDR_WLOCK(ifp); TAILQ_INSERT_TAIL(&ifp->if_addrhead, &ia->ia_ifa, ifa_link); IF_ADDR_WUNLOCK(ifp); ifa_ref(&ia->ia_ifa); /* in6_ifaddrhead */ IN6_IFADDR_WLOCK(); TAILQ_INSERT_TAIL(&V_in6_ifaddrhead, ia, ia_link); LIST_INSERT_HEAD(IN6ADDR_HASH(&ifra->ifra_addr.sin6_addr), ia, ia6_hash); IN6_IFADDR_WUNLOCK(); } /* update timestamp */ - ia->ia6_updatetime = time_second; + ia->ia6_updatetime = time_uptime; /* set prefix mask */ if (ifra->ifra_prefixmask.sin6_len) { /* * We prohibit changing the prefix length of an existing * address, because * + such an operation should be rare in IPv6, and * + the operation would confuse prefix management. */ if (ia->ia_prefixmask.sin6_len && in6_mask2len(&ia->ia_prefixmask.sin6_addr, NULL) != plen) { nd6log((LOG_INFO, "in6_update_ifa: the prefix length of an" " existing (%s) address should not be changed\n", ip6_sprintf(ip6buf, &ia->ia_addr.sin6_addr))); error = EINVAL; goto unlink; } ia->ia_prefixmask = ifra->ifra_prefixmask; ia->ia_prefixmask.sin6_family = AF_INET6; } /* * If a new destination address is specified, scrub the old one and * install the new destination. Note that the interface must be * p2p or loopback (see the check above.) */ if (dst6.sin6_family == AF_INET6 && !IN6_ARE_ADDR_EQUAL(&dst6.sin6_addr, &ia->ia_dstaddr.sin6_addr)) { int e; if ((ia->ia_flags & IFA_ROUTE) != 0 && (e = rtinit(&(ia->ia_ifa), (int)RTM_DELETE, RTF_HOST)) != 0) { nd6log((LOG_ERR, "in6_update_ifa: failed to remove " "a route to the old destination: %s\n", ip6_sprintf(ip6buf, &ia->ia_addr.sin6_addr))); /* proceed anyway... */ } else ia->ia_flags &= ~IFA_ROUTE; ia->ia_dstaddr = dst6; } /* * Set lifetimes. We do not refer to ia6t_expire and ia6t_preferred * to see if the address is deprecated or invalidated, but initialize * these members for applications. */ ia->ia6_lifetime = ifra->ifra_lifetime; if (ia->ia6_lifetime.ia6t_vltime != ND6_INFINITE_LIFETIME) { ia->ia6_lifetime.ia6t_expire = - time_second + ia->ia6_lifetime.ia6t_vltime; + time_uptime + ia->ia6_lifetime.ia6t_vltime; } else ia->ia6_lifetime.ia6t_expire = 0; if (ia->ia6_lifetime.ia6t_pltime != ND6_INFINITE_LIFETIME) { ia->ia6_lifetime.ia6t_preferred = - time_second + ia->ia6_lifetime.ia6t_pltime; + time_uptime + ia->ia6_lifetime.ia6t_pltime; } else ia->ia6_lifetime.ia6t_preferred = 0; /* reset the interface and routing table appropriately. */ if ((error = in6_ifinit(ifp, ia, &ifra->ifra_addr, hostIsNew)) != 0) goto unlink; /* * configure address flags. */ ia->ia6_flags = ifra->ifra_flags; /* * backward compatibility - if IN6_IFF_DEPRECATED is set from the * userland, make it deprecated. */ if ((ifra->ifra_flags & IN6_IFF_DEPRECATED) != 0) { ia->ia6_lifetime.ia6t_pltime = 0; - ia->ia6_lifetime.ia6t_preferred = time_second; + ia->ia6_lifetime.ia6t_preferred = time_uptime; } /* * Make the address tentative before joining multicast addresses, * so that corresponding MLD responses would not have a tentative * source address. */ ia->ia6_flags &= ~IN6_IFF_DUPLICATED; /* safety */ if (hostIsNew && in6if_do_dad(ifp)) ia->ia6_flags |= IN6_IFF_TENTATIVE; /* DAD should be performed after ND6_IFF_IFDISABLED is cleared. */ if (ND_IFINFO(ifp)->flags & ND6_IFF_IFDISABLED) ia->ia6_flags |= IN6_IFF_TENTATIVE; /* * We are done if we have simply modified an existing address. */ if (!hostIsNew) return (error); /* * Beyond this point, we should call in6_purgeaddr upon an error, * not just go to unlink. */ /* Join necessary multicast groups. */ in6m_sol = NULL; if ((ifp->if_flags & IFF_MULTICAST) != 0) { error = in6_update_ifa_join_mc(ifp, ifra, ia, flags, &in6m_sol); if (error) goto cleanup; } /* * Perform DAD, if needed. * XXX It may be of use, if we can administratively disable DAD. */ if (in6if_do_dad(ifp) && ((ifra->ifra_flags & IN6_IFF_NODAD) == 0) && (ia->ia6_flags & IN6_IFF_TENTATIVE)) { int mindelay, maxdelay; delay = 0; if ((flags & IN6_IFAUPDATE_DADDELAY)) { /* * We need to impose a delay before sending an NS * for DAD. Check if we also needed a delay for the * corresponding MLD message. If we did, the delay * should be larger than the MLD delay (this could be * relaxed a bit, but this simple logic is at least * safe). * XXX: Break data hiding guidelines and look at * state for the solicited multicast group. */ mindelay = 0; if (in6m_sol != NULL && in6m_sol->in6m_state == MLD_REPORTING_MEMBER) { mindelay = in6m_sol->in6m_timer; } maxdelay = MAX_RTR_SOLICITATION_DELAY * hz; if (maxdelay - mindelay == 0) delay = 0; else { delay = (arc4random() % (maxdelay - mindelay)) + mindelay; } } nd6_dad_start((struct ifaddr *)ia, delay); } KASSERT(hostIsNew, ("in6_update_ifa: !hostIsNew")); ifa_free(&ia->ia_ifa); return (error); unlink: /* * XXX: if a change of an existing address failed, keep the entry * anyway. */ if (hostIsNew) { in6_unlink_ifa(ia, ifp); ifa_free(&ia->ia_ifa); } return (error); cleanup: KASSERT(hostIsNew, ("in6_update_ifa: cleanup: !hostIsNew")); ifa_free(&ia->ia_ifa); in6_purgeaddr(&ia->ia_ifa); return error; } /* * Leave multicast groups. Factored out from in6_purgeaddr(). * This entire work should only be done once, for the default FIB. */ static int in6_purgeaddr_mc(struct ifnet *ifp, struct in6_ifaddr *ia, struct ifaddr *ifa0) { struct sockaddr_in6 mltaddr, mltmask; struct in6_multi_mship *imm; struct rtentry *rt; struct sockaddr_in6 sin6; int error; /* * Leave from multicast groups we have joined for the interface. */ while ((imm = LIST_FIRST(&ia->ia6_memberships)) != NULL) { LIST_REMOVE(imm, i6mm_chain); in6_leavegroup(imm); } /* * Remove the link-local all-nodes address. */ bzero(&mltmask, sizeof(mltmask)); mltmask.sin6_len = sizeof(struct sockaddr_in6); mltmask.sin6_family = AF_INET6; mltmask.sin6_addr = in6mask32; bzero(&mltaddr, sizeof(mltaddr)); mltaddr.sin6_len = sizeof(struct sockaddr_in6); mltaddr.sin6_family = AF_INET6; mltaddr.sin6_addr = in6addr_linklocal_allnodes; if ((error = in6_setscope(&mltaddr.sin6_addr, ifp, NULL)) != 0) return (error); /* * As for the mltaddr above, proactively prepare the sin6 to avoid * rtentry un- and re-locking. */ if (ifa0 != NULL) { bzero(&sin6, sizeof(sin6)); sin6.sin6_len = sizeof(sin6); sin6.sin6_family = AF_INET6; memcpy(&sin6.sin6_addr, &satosin6(ifa0->ifa_addr)->sin6_addr, sizeof(sin6.sin6_addr)); error = in6_setscope(&sin6.sin6_addr, ifa0->ifa_ifp, NULL); if (error != 0) return (error); } rt = in6_rtalloc1((struct sockaddr *)&mltaddr, 0, 0UL, RT_DEFAULT_FIB); if (rt != NULL && rt->rt_gateway != NULL && (memcmp(&satosin6(rt->rt_gateway)->sin6_addr, &ia->ia_addr.sin6_addr, sizeof(ia->ia_addr.sin6_addr)) == 0)) { /* * If no more IPv6 address exists on this interface then * remove the multicast address route. */ if (ifa0 == NULL) { memcpy(&mltaddr.sin6_addr, &satosin6(rt_key(rt))->sin6_addr, sizeof(mltaddr.sin6_addr)); RTFREE_LOCKED(rt); error = in6_rtrequest(RTM_DELETE, (struct sockaddr *)&mltaddr, (struct sockaddr *)&ia->ia_addr, (struct sockaddr *)&mltmask, RTF_UP, (struct rtentry **)0, RT_DEFAULT_FIB); if (error) log(LOG_INFO, "%s: link-local all-nodes " "multicast address deletion error\n", __func__); } else { /* * Replace the gateway of the route. */ memcpy(rt->rt_gateway, &sin6, sizeof(sin6)); RTFREE_LOCKED(rt); } } else { if (rt != NULL) RTFREE_LOCKED(rt); } /* * Remove the node-local all-nodes address. */ mltaddr.sin6_addr = in6addr_nodelocal_allnodes; if ((error = in6_setscope(&mltaddr.sin6_addr, ifp, NULL)) != 0) return (error); rt = in6_rtalloc1((struct sockaddr *)&mltaddr, 0, 0UL, RT_DEFAULT_FIB); if (rt != NULL && rt->rt_gateway != NULL && (memcmp(&satosin6(rt->rt_gateway)->sin6_addr, &ia->ia_addr.sin6_addr, sizeof(ia->ia_addr.sin6_addr)) == 0)) { /* * If no more IPv6 address exists on this interface then * remove the multicast address route. */ if (ifa0 == NULL) { memcpy(&mltaddr.sin6_addr, &satosin6(rt_key(rt))->sin6_addr, sizeof(mltaddr.sin6_addr)); RTFREE_LOCKED(rt); error = in6_rtrequest(RTM_DELETE, (struct sockaddr *)&mltaddr, (struct sockaddr *)&ia->ia_addr, (struct sockaddr *)&mltmask, RTF_UP, (struct rtentry **)0, RT_DEFAULT_FIB); if (error) log(LOG_INFO, "%s: node-local all-nodes" "multicast address deletion error\n", __func__); } else { /* * Replace the gateway of the route. */ memcpy(rt->rt_gateway, &sin6, sizeof(sin6)); RTFREE_LOCKED(rt); } } else { if (rt != NULL) RTFREE_LOCKED(rt); } return (0); } void in6_purgeaddr(struct ifaddr *ifa) { struct ifnet *ifp = ifa->ifa_ifp; struct in6_ifaddr *ia = (struct in6_ifaddr *) ifa; int plen, error; struct ifaddr *ifa0; if (ifa->ifa_carp) (*carp_detach_p)(ifa); /* * find another IPv6 address as the gateway for the * link-local and node-local all-nodes multicast * address routes */ IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa0, &ifp->if_addrhead, ifa_link) { if ((ifa0->ifa_addr->sa_family != AF_INET6) || memcmp(&satosin6(ifa0->ifa_addr)->sin6_addr, &ia->ia_addr.sin6_addr, sizeof(struct in6_addr)) == 0) continue; else break; } if (ifa0 != NULL) ifa_ref(ifa0); IF_ADDR_RUNLOCK(ifp); /* * Remove the loopback route to the interface address. * The check for the current setting of "nd6_useloopback" * is not needed. */ if (ia->ia_flags & IFA_RTSELF) { error = ifa_del_loopback_route((struct ifaddr *)ia, (struct sockaddr *)&ia->ia_addr); if (error == 0) ia->ia_flags &= ~IFA_RTSELF; } /* stop DAD processing */ nd6_dad_stop(ifa); /* Remove local address entry from lltable. */ in6_ifremloop(ifa); /* Leave multicast groups. */ error = in6_purgeaddr_mc(ifp, ia, ifa0); if (ifa0 != NULL) ifa_free(ifa0); plen = in6_mask2len(&ia->ia_prefixmask.sin6_addr, NULL); /* XXX */ if ((ia->ia_flags & IFA_ROUTE) && plen == 128) { error = rtinit(&(ia->ia_ifa), RTM_DELETE, ia->ia_flags | (ia->ia_dstaddr.sin6_family == AF_INET6) ? RTF_HOST : 0); if (error != 0) log(LOG_INFO, "%s: err=%d, destination address delete " "failed\n", __func__, error); ia->ia_flags &= ~IFA_ROUTE; } in6_unlink_ifa(ia, ifp); } static void in6_unlink_ifa(struct in6_ifaddr *ia, struct ifnet *ifp) { IF_ADDR_WLOCK(ifp); TAILQ_REMOVE(&ifp->if_addrhead, &ia->ia_ifa, ifa_link); IF_ADDR_WUNLOCK(ifp); ifa_free(&ia->ia_ifa); /* if_addrhead */ /* * Defer the release of what might be the last reference to the * in6_ifaddr so that it can't be freed before the remainder of the * cleanup. */ IN6_IFADDR_WLOCK(); TAILQ_REMOVE(&V_in6_ifaddrhead, ia, ia_link); LIST_REMOVE(ia, ia6_hash); IN6_IFADDR_WUNLOCK(); /* * Release the reference to the base prefix. There should be a * positive reference. */ if (ia->ia6_ndpr == NULL) { nd6log((LOG_NOTICE, "in6_unlink_ifa: autoconf'ed address " "%p has no prefix\n", ia)); } else { ia->ia6_ndpr->ndpr_refcnt--; ia->ia6_ndpr = NULL; } /* * Also, if the address being removed is autoconf'ed, call * pfxlist_onlink_check() since the release might affect the status of * other (detached) addresses. */ if ((ia->ia6_flags & IN6_IFF_AUTOCONF)) { pfxlist_onlink_check(); } ifa_free(&ia->ia_ifa); /* in6_ifaddrhead */ } void in6_purgeif(struct ifnet *ifp) { struct ifaddr *ifa, *nifa; TAILQ_FOREACH_SAFE(ifa, &ifp->if_addrhead, ifa_link, nifa) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; in6_purgeaddr(ifa); } in6_ifdetach(ifp); } /* * SIOC[GAD]LIFADDR. * SIOCGLIFADDR: get first address. (?) * SIOCGLIFADDR with IFLR_PREFIX: * get first address that matches the specified prefix. * SIOCALIFADDR: add the specified address. * SIOCALIFADDR with IFLR_PREFIX: * add the specified prefix, filling hostid part from * the first link-local address. prefixlen must be <= 64. * SIOCDLIFADDR: delete the specified address. * SIOCDLIFADDR with IFLR_PREFIX: * delete the first address that matches the specified prefix. * return values: * EINVAL on invalid parameters * EADDRNOTAVAIL on prefix match failed/specified address not found * other values may be returned from in6_ioctl() * * NOTE: SIOCALIFADDR(with IFLR_PREFIX set) allows prefixlen less than 64. * this is to accomodate address naming scheme other than RFC2374, * in the future. * RFC2373 defines interface id to be 64bit, but it allows non-RFC2374 * address encoding scheme. (see figure on page 8) */ static int in6_lifaddr_ioctl(struct socket *so, u_long cmd, caddr_t data, struct ifnet *ifp, struct thread *td) { struct if_laddrreq *iflr = (struct if_laddrreq *)data; struct ifaddr *ifa; struct sockaddr *sa; /* sanity checks */ if (!data || !ifp) { panic("invalid argument to in6_lifaddr_ioctl"); /* NOTREACHED */ } switch (cmd) { case SIOCGLIFADDR: /* address must be specified on GET with IFLR_PREFIX */ if ((iflr->flags & IFLR_PREFIX) == 0) break; /* FALLTHROUGH */ case SIOCALIFADDR: case SIOCDLIFADDR: /* address must be specified on ADD and DELETE */ sa = (struct sockaddr *)&iflr->addr; if (sa->sa_family != AF_INET6) return EINVAL; if (sa->sa_len != sizeof(struct sockaddr_in6)) return EINVAL; /* XXX need improvement */ sa = (struct sockaddr *)&iflr->dstaddr; if (sa->sa_family && sa->sa_family != AF_INET6) return EINVAL; if (sa->sa_len && sa->sa_len != sizeof(struct sockaddr_in6)) return EINVAL; break; default: /* shouldn't happen */ #if 0 panic("invalid cmd to in6_lifaddr_ioctl"); /* NOTREACHED */ #else return EOPNOTSUPP; #endif } if (sizeof(struct in6_addr) * 8 < iflr->prefixlen) return EINVAL; switch (cmd) { case SIOCALIFADDR: { struct in6_aliasreq ifra; struct in6_addr *hostid = NULL; int prefixlen; ifa = NULL; if ((iflr->flags & IFLR_PREFIX) != 0) { struct sockaddr_in6 *sin6; /* * hostid is to fill in the hostid part of the * address. hostid points to the first link-local * address attached to the interface. */ ifa = (struct ifaddr *)in6ifa_ifpforlinklocal(ifp, 0); if (!ifa) return EADDRNOTAVAIL; hostid = IFA_IN6(ifa); /* prefixlen must be <= 64. */ if (64 < iflr->prefixlen) { if (ifa != NULL) ifa_free(ifa); return EINVAL; } prefixlen = iflr->prefixlen; /* hostid part must be zero. */ sin6 = (struct sockaddr_in6 *)&iflr->addr; if (sin6->sin6_addr.s6_addr32[2] != 0 || sin6->sin6_addr.s6_addr32[3] != 0) { if (ifa != NULL) ifa_free(ifa); return EINVAL; } } else prefixlen = iflr->prefixlen; /* copy args to in6_aliasreq, perform ioctl(SIOCAIFADDR_IN6). */ bzero(&ifra, sizeof(ifra)); bcopy(iflr->iflr_name, ifra.ifra_name, sizeof(ifra.ifra_name)); bcopy(&iflr->addr, &ifra.ifra_addr, ((struct sockaddr *)&iflr->addr)->sa_len); if (hostid) { /* fill in hostid part */ ifra.ifra_addr.sin6_addr.s6_addr32[2] = hostid->s6_addr32[2]; ifra.ifra_addr.sin6_addr.s6_addr32[3] = hostid->s6_addr32[3]; } if (((struct sockaddr *)&iflr->dstaddr)->sa_family) { /* XXX */ bcopy(&iflr->dstaddr, &ifra.ifra_dstaddr, ((struct sockaddr *)&iflr->dstaddr)->sa_len); if (hostid) { ifra.ifra_dstaddr.sin6_addr.s6_addr32[2] = hostid->s6_addr32[2]; ifra.ifra_dstaddr.sin6_addr.s6_addr32[3] = hostid->s6_addr32[3]; } } if (ifa != NULL) ifa_free(ifa); ifra.ifra_prefixmask.sin6_len = sizeof(struct sockaddr_in6); in6_prefixlen2mask(&ifra.ifra_prefixmask.sin6_addr, prefixlen); ifra.ifra_flags = iflr->flags & ~IFLR_PREFIX; return in6_control(so, SIOCAIFADDR_IN6, (caddr_t)&ifra, ifp, td); } case SIOCGLIFADDR: case SIOCDLIFADDR: { struct in6_ifaddr *ia; struct in6_addr mask, candidate, match; struct sockaddr_in6 *sin6; int cmp; bzero(&mask, sizeof(mask)); if (iflr->flags & IFLR_PREFIX) { /* lookup a prefix rather than address. */ in6_prefixlen2mask(&mask, iflr->prefixlen); sin6 = (struct sockaddr_in6 *)&iflr->addr; bcopy(&sin6->sin6_addr, &match, sizeof(match)); match.s6_addr32[0] &= mask.s6_addr32[0]; match.s6_addr32[1] &= mask.s6_addr32[1]; match.s6_addr32[2] &= mask.s6_addr32[2]; match.s6_addr32[3] &= mask.s6_addr32[3]; /* if you set extra bits, that's wrong */ if (bcmp(&match, &sin6->sin6_addr, sizeof(match))) return EINVAL; cmp = 1; } else { if (cmd == SIOCGLIFADDR) { /* on getting an address, take the 1st match */ cmp = 0; /* XXX */ } else { /* on deleting an address, do exact match */ in6_prefixlen2mask(&mask, 128); sin6 = (struct sockaddr_in6 *)&iflr->addr; bcopy(&sin6->sin6_addr, &match, sizeof(match)); cmp = 1; } } IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; if (!cmp) break; /* * XXX: this is adhoc, but is necessary to allow * a user to specify fe80::/64 (not /10) for a * link-local address. */ bcopy(IFA_IN6(ifa), &candidate, sizeof(candidate)); in6_clearscope(&candidate); candidate.s6_addr32[0] &= mask.s6_addr32[0]; candidate.s6_addr32[1] &= mask.s6_addr32[1]; candidate.s6_addr32[2] &= mask.s6_addr32[2]; candidate.s6_addr32[3] &= mask.s6_addr32[3]; if (IN6_ARE_ADDR_EQUAL(&candidate, &match)) break; } if (ifa != NULL) ifa_ref(ifa); IF_ADDR_RUNLOCK(ifp); if (!ifa) return EADDRNOTAVAIL; ia = ifa2ia6(ifa); if (cmd == SIOCGLIFADDR) { int error; /* fill in the if_laddrreq structure */ bcopy(&ia->ia_addr, &iflr->addr, ia->ia_addr.sin6_len); error = sa6_recoverscope( (struct sockaddr_in6 *)&iflr->addr); if (error != 0) { ifa_free(ifa); return (error); } if ((ifp->if_flags & IFF_POINTOPOINT) != 0) { bcopy(&ia->ia_dstaddr, &iflr->dstaddr, ia->ia_dstaddr.sin6_len); error = sa6_recoverscope( (struct sockaddr_in6 *)&iflr->dstaddr); if (error != 0) { ifa_free(ifa); return (error); } } else bzero(&iflr->dstaddr, sizeof(iflr->dstaddr)); iflr->prefixlen = in6_mask2len(&ia->ia_prefixmask.sin6_addr, NULL); iflr->flags = ia->ia6_flags; /* XXX */ ifa_free(ifa); return 0; } else { struct in6_aliasreq ifra; /* fill in6_aliasreq and do ioctl(SIOCDIFADDR_IN6) */ bzero(&ifra, sizeof(ifra)); bcopy(iflr->iflr_name, ifra.ifra_name, sizeof(ifra.ifra_name)); bcopy(&ia->ia_addr, &ifra.ifra_addr, ia->ia_addr.sin6_len); if ((ifp->if_flags & IFF_POINTOPOINT) != 0) { bcopy(&ia->ia_dstaddr, &ifra.ifra_dstaddr, ia->ia_dstaddr.sin6_len); } else { bzero(&ifra.ifra_dstaddr, sizeof(ifra.ifra_dstaddr)); } bcopy(&ia->ia_prefixmask, &ifra.ifra_dstaddr, ia->ia_prefixmask.sin6_len); ifra.ifra_flags = ia->ia6_flags; ifa_free(ifa); return in6_control(so, SIOCDIFADDR_IN6, (caddr_t)&ifra, ifp, td); } } } return EOPNOTSUPP; /* just for safety */ } /* * Initialize an interface's IPv6 address and routing table entry. */ static int in6_ifinit(struct ifnet *ifp, struct in6_ifaddr *ia, struct sockaddr_in6 *sin6, int newhost) { int error = 0, plen, ifacount = 0; struct ifaddr *ifa; /* * Give the interface a chance to initialize * if this is its first address, * and to validate the address if necessary. */ IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; ifacount++; } IF_ADDR_RUNLOCK(ifp); ia->ia_addr = *sin6; if (ifacount <= 1 && ifp->if_ioctl) { error = (*ifp->if_ioctl)(ifp, SIOCSIFADDR, (caddr_t)ia); if (error) return (error); } ia->ia_ifa.ifa_metric = ifp->if_metric; /* we could do in(6)_socktrim here, but just omit it at this moment. */ /* * Special case: * If a new destination address is specified for a point-to-point * interface, install a route to the destination as an interface * direct route. * XXX: the logic below rejects assigning multiple addresses on a p2p * interface that share the same destination. */ plen = in6_mask2len(&ia->ia_prefixmask.sin6_addr, NULL); /* XXX */ if (!(ia->ia_flags & IFA_ROUTE) && plen == 128 && ia->ia_dstaddr.sin6_family == AF_INET6) { int rtflags = RTF_UP | RTF_HOST; error = rtinit(&ia->ia_ifa, RTM_ADD, ia->ia_flags | rtflags); if (error) return (error); ia->ia_flags |= IFA_ROUTE; /* * Handle the case for ::1 . */ if (ifp->if_flags & IFF_LOOPBACK) ia->ia_flags |= IFA_RTSELF; } /* * add a loopback route to self */ if (!(ia->ia_flags & IFA_RTSELF) && V_nd6_useloopback) { error = ifa_add_loopback_route((struct ifaddr *)ia, (struct sockaddr *)&ia->ia_addr); if (error == 0) ia->ia_flags |= IFA_RTSELF; } /* Add local address to lltable, if necessary (ex. on p2p link). */ if (newhost) in6_ifaddloop(&(ia->ia_ifa)); return (error); } /* * Find an IPv6 interface link-local address specific to an interface. * ifaddr is returned referenced. */ struct in6_ifaddr * in6ifa_ifpforlinklocal(struct ifnet *ifp, int ignoreflags) { struct ifaddr *ifa; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; if (IN6_IS_ADDR_LINKLOCAL(IFA_IN6(ifa))) { if ((((struct in6_ifaddr *)ifa)->ia6_flags & ignoreflags) != 0) continue; ifa_ref(ifa); break; } } IF_ADDR_RUNLOCK(ifp); return ((struct in6_ifaddr *)ifa); } /* * find the internet address corresponding to a given interface and address. * ifaddr is returned referenced. */ struct in6_ifaddr * in6ifa_ifpwithaddr(struct ifnet *ifp, struct in6_addr *addr) { struct ifaddr *ifa; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; if (IN6_ARE_ADDR_EQUAL(addr, IFA_IN6(ifa))) { ifa_ref(ifa); break; } } IF_ADDR_RUNLOCK(ifp); return ((struct in6_ifaddr *)ifa); } /* * Find a link-local scoped address on ifp and return it if any. */ struct in6_ifaddr * in6ifa_llaonifp(struct ifnet *ifp) { struct sockaddr_in6 *sin6; struct ifaddr *ifa; if (ND_IFINFO(ifp)->flags & ND6_IFF_IFDISABLED) return (NULL); if_addr_rlock(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; sin6 = (struct sockaddr_in6 *)ifa->ifa_addr; if (IN6_IS_SCOPE_LINKLOCAL(&sin6->sin6_addr) || IN6_IS_ADDR_MC_INTFACELOCAL(&sin6->sin6_addr) || IN6_IS_ADDR_MC_NODELOCAL(&sin6->sin6_addr)) break; } if_addr_runlock(ifp); return ((struct in6_ifaddr *)ifa); } /* * Convert IP6 address to printable (loggable) representation. Caller * has to make sure that ip6buf is at least INET6_ADDRSTRLEN long. */ static char digits[] = "0123456789abcdef"; char * ip6_sprintf(char *ip6buf, const struct in6_addr *addr) { int i, cnt = 0, maxcnt = 0, idx = 0, index = 0; char *cp; const u_int16_t *a = (const u_int16_t *)addr; const u_int8_t *d; int dcolon = 0, zero = 0; cp = ip6buf; for (i = 0; i < 8; i++) { if (*(a + i) == 0) { cnt++; if (cnt == 1) idx = i; } else if (maxcnt < cnt) { maxcnt = cnt; index = idx; cnt = 0; } } if (maxcnt < cnt) { maxcnt = cnt; index = idx; } for (i = 0; i < 8; i++) { if (dcolon == 1) { if (*a == 0) { if (i == 7) *cp++ = ':'; a++; continue; } else dcolon = 2; } if (*a == 0) { if (dcolon == 0 && *(a + 1) == 0 && i == index) { if (i == 0) *cp++ = ':'; *cp++ = ':'; dcolon = 1; } else { *cp++ = '0'; *cp++ = ':'; } a++; continue; } d = (const u_char *)a; /* Try to eliminate leading zeros in printout like in :0001. */ zero = 1; *cp = digits[*d >> 4]; if (*cp != '0') { zero = 0; cp++; } *cp = digits[*d++ & 0xf]; if (zero == 0 || (*cp != '0')) { zero = 0; cp++; } *cp = digits[*d >> 4]; if (zero == 0 || (*cp != '0')) { zero = 0; cp++; } *cp++ = digits[*d & 0xf]; *cp++ = ':'; a++; } *--cp = '\0'; return (ip6buf); } int in6_localaddr(struct in6_addr *in6) { struct in6_ifaddr *ia; if (IN6_IS_ADDR_LOOPBACK(in6) || IN6_IS_ADDR_LINKLOCAL(in6)) return 1; IN6_IFADDR_RLOCK(); TAILQ_FOREACH(ia, &V_in6_ifaddrhead, ia_link) { if (IN6_ARE_MASKED_ADDR_EQUAL(in6, &ia->ia_addr.sin6_addr, &ia->ia_prefixmask.sin6_addr)) { IN6_IFADDR_RUNLOCK(); return 1; } } IN6_IFADDR_RUNLOCK(); return (0); } /* * Return 1 if an internet address is for the local host and configured * on one of its interfaces. */ int in6_localip(struct in6_addr *in6) { struct in6_ifaddr *ia; IN6_IFADDR_RLOCK(); LIST_FOREACH(ia, IN6ADDR_HASH(in6), ia6_hash) { if (IN6_ARE_ADDR_EQUAL(in6, &ia->ia_addr.sin6_addr)) { IN6_IFADDR_RUNLOCK(); return (1); } } IN6_IFADDR_RUNLOCK(); return (0); } int in6_is_addr_deprecated(struct sockaddr_in6 *sa6) { struct in6_ifaddr *ia; IN6_IFADDR_RLOCK(); LIST_FOREACH(ia, IN6ADDR_HASH(&sa6->sin6_addr), ia6_hash) { if (IN6_ARE_ADDR_EQUAL(IA6_IN6(ia), &sa6->sin6_addr)) { if (ia->ia6_flags & IN6_IFF_DEPRECATED) { IN6_IFADDR_RUNLOCK(); return (1); /* true */ } break; } } IN6_IFADDR_RUNLOCK(); return (0); /* false */ } /* * return length of part which dst and src are equal * hard coding... */ int in6_matchlen(struct in6_addr *src, struct in6_addr *dst) { int match = 0; u_char *s = (u_char *)src, *d = (u_char *)dst; u_char *lim = s + 16, r; while (s < lim) if ((r = (*d++ ^ *s++)) != 0) { while (r < 128) { match++; r <<= 1; } break; } else match += 8; return match; } /* XXX: to be scope conscious */ int in6_are_prefix_equal(struct in6_addr *p1, struct in6_addr *p2, int len) { int bytelen, bitlen; /* sanity check */ if (0 > len || len > 128) { log(LOG_ERR, "in6_are_prefix_equal: invalid prefix length(%d)\n", len); return (0); } bytelen = len / 8; bitlen = len % 8; if (bcmp(&p1->s6_addr, &p2->s6_addr, bytelen)) return (0); if (bitlen != 0 && p1->s6_addr[bytelen] >> (8 - bitlen) != p2->s6_addr[bytelen] >> (8 - bitlen)) return (0); return (1); } void in6_prefixlen2mask(struct in6_addr *maskp, int len) { u_char maskarray[8] = {0x80, 0xc0, 0xe0, 0xf0, 0xf8, 0xfc, 0xfe, 0xff}; int bytelen, bitlen, i; /* sanity check */ if (0 > len || len > 128) { log(LOG_ERR, "in6_prefixlen2mask: invalid prefix length(%d)\n", len); return; } bzero(maskp, sizeof(*maskp)); bytelen = len / 8; bitlen = len % 8; for (i = 0; i < bytelen; i++) maskp->s6_addr[i] = 0xff; if (bitlen) maskp->s6_addr[bytelen] = maskarray[bitlen - 1]; } /* * return the best address out of the same scope. if no address was * found, return the first valid address from designated IF. */ struct in6_ifaddr * in6_ifawithifp(struct ifnet *ifp, struct in6_addr *dst) { int dst_scope = in6_addrscope(dst), blen = -1, tlen; struct ifaddr *ifa; struct in6_ifaddr *besta = 0; struct in6_ifaddr *dep[2]; /* last-resort: deprecated */ dep[0] = dep[1] = NULL; /* * We first look for addresses in the same scope. * If there is one, return it. * If two or more, return one which matches the dst longest. * If none, return one of global addresses assigned other ifs. */ IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_ANYCAST) continue; /* XXX: is there any case to allow anycast? */ if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_NOTREADY) continue; /* don't use this interface */ if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_DETACHED) continue; if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_DEPRECATED) { if (V_ip6_use_deprecated) dep[0] = (struct in6_ifaddr *)ifa; continue; } if (dst_scope == in6_addrscope(IFA_IN6(ifa))) { /* * call in6_matchlen() as few as possible */ if (besta) { if (blen == -1) blen = in6_matchlen(&besta->ia_addr.sin6_addr, dst); tlen = in6_matchlen(IFA_IN6(ifa), dst); if (tlen > blen) { blen = tlen; besta = (struct in6_ifaddr *)ifa; } } else besta = (struct in6_ifaddr *)ifa; } } if (besta) { ifa_ref(&besta->ia_ifa); IF_ADDR_RUNLOCK(ifp); return (besta); } TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_ANYCAST) continue; /* XXX: is there any case to allow anycast? */ if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_NOTREADY) continue; /* don't use this interface */ if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_DETACHED) continue; if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_DEPRECATED) { if (V_ip6_use_deprecated) dep[1] = (struct in6_ifaddr *)ifa; continue; } if (ifa != NULL) ifa_ref(ifa); IF_ADDR_RUNLOCK(ifp); return (struct in6_ifaddr *)ifa; } /* use the last-resort values, that are, deprecated addresses */ if (dep[0]) { ifa_ref((struct ifaddr *)dep[0]); IF_ADDR_RUNLOCK(ifp); return dep[0]; } if (dep[1]) { ifa_ref((struct ifaddr *)dep[1]); IF_ADDR_RUNLOCK(ifp); return dep[1]; } IF_ADDR_RUNLOCK(ifp); return NULL; } /* * perform DAD when interface becomes IFF_UP. */ void in6_if_up(struct ifnet *ifp) { struct ifaddr *ifa; struct in6_ifaddr *ia; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; ia = (struct in6_ifaddr *)ifa; if (ia->ia6_flags & IN6_IFF_TENTATIVE) { /* * The TENTATIVE flag was likely set by hand * beforehand, implicitly indicating the need for DAD. * We may be able to skip the random delay in this * case, but we impose delays just in case. */ nd6_dad_start(ifa, arc4random() % (MAX_RTR_SOLICITATION_DELAY * hz)); } } IF_ADDR_RUNLOCK(ifp); /* * special cases, like 6to4, are handled in in6_ifattach */ in6_ifattach(ifp, NULL); } int in6if_do_dad(struct ifnet *ifp) { if ((ifp->if_flags & IFF_LOOPBACK) != 0) return (0); if (ND_IFINFO(ifp)->flags & ND6_IFF_IFDISABLED) return (0); switch (ifp->if_type) { #ifdef IFT_DUMMY case IFT_DUMMY: #endif case IFT_FAITH: /* * These interfaces do not have the IFF_LOOPBACK flag, * but loop packets back. We do not have to do DAD on such * interfaces. We should even omit it, because loop-backed * NS would confuse the DAD procedure. */ return (0); default: /* * Our DAD routine requires the interface up and running. * However, some interfaces can be up before the RUNNING * status. Additionaly, users may try to assign addresses * before the interface becomes up (or running). * We simply skip DAD in such a case as a work around. * XXX: we should rather mark "tentative" on such addresses, * and do DAD after the interface becomes ready. */ if (!((ifp->if_flags & IFF_UP) && (ifp->if_drv_flags & IFF_DRV_RUNNING))) return (0); return (1); } } /* * Calculate max IPv6 MTU through all the interfaces and store it * to in6_maxmtu. */ void in6_setmaxmtu(void) { unsigned long maxmtu = 0; struct ifnet *ifp; IFNET_RLOCK_NOSLEEP(); TAILQ_FOREACH(ifp, &V_ifnet, if_list) { /* this function can be called during ifnet initialization */ if (!ifp->if_afdata[AF_INET6]) continue; if ((ifp->if_flags & IFF_LOOPBACK) == 0 && IN6_LINKMTU(ifp) > maxmtu) maxmtu = IN6_LINKMTU(ifp); } IFNET_RUNLOCK_NOSLEEP(); if (maxmtu) /* update only when maxmtu is positive */ V_in6_maxmtu = maxmtu; } /* * Provide the length of interface identifiers to be used for the link attached * to the given interface. The length should be defined in "IPv6 over * xxx-link" document. Note that address architecture might also define * the length for a particular set of address prefixes, regardless of the * link type. As clarified in rfc2462bis, those two definitions should be * consistent, and those really are as of August 2004. */ int in6_if2idlen(struct ifnet *ifp) { switch (ifp->if_type) { case IFT_ETHER: /* RFC2464 */ #ifdef IFT_PROPVIRTUAL case IFT_PROPVIRTUAL: /* XXX: no RFC. treat it as ether */ #endif #ifdef IFT_L2VLAN case IFT_L2VLAN: /* ditto */ #endif #ifdef IFT_IEEE80211 case IFT_IEEE80211: /* ditto */ #endif #ifdef IFT_MIP case IFT_MIP: /* ditto */ #endif case IFT_INFINIBAND: return (64); case IFT_FDDI: /* RFC2467 */ return (64); case IFT_ISO88025: /* RFC2470 (IPv6 over Token Ring) */ return (64); case IFT_PPP: /* RFC2472 */ return (64); case IFT_ARCNET: /* RFC2497 */ return (64); case IFT_FRELAY: /* RFC2590 */ return (64); case IFT_IEEE1394: /* RFC3146 */ return (64); case IFT_GIF: return (64); /* draft-ietf-v6ops-mech-v2-07 */ case IFT_LOOP: return (64); /* XXX: is this really correct? */ default: /* * Unknown link type: * It might be controversial to use the today's common constant * of 64 for these cases unconditionally. For full compliance, * we should return an error in this case. On the other hand, * if we simply miss the standard for the link type or a new * standard is defined for a new link type, the IFID length * is very likely to be the common constant. As a compromise, * we always use the constant, but make an explicit notice * indicating the "unknown" case. */ printf("in6_if2idlen: unknown link type (%d)\n", ifp->if_type); return (64); } } #include struct in6_llentry { struct llentry base; struct sockaddr_in6 l3_addr6; }; /* * Deletes an address from the address table. * This function is called by the timer functions * such as arptimer() and nd6_llinfo_timer(), and * the caller does the locking. */ static void in6_lltable_free(struct lltable *llt, struct llentry *lle) { LLE_WUNLOCK(lle); LLE_LOCK_DESTROY(lle); free(lle, M_LLTABLE); } static struct llentry * in6_lltable_new(const struct sockaddr *l3addr, u_int flags) { struct in6_llentry *lle; lle = malloc(sizeof(struct in6_llentry), M_LLTABLE, M_NOWAIT | M_ZERO); if (lle == NULL) /* NB: caller generates msg */ return NULL; lle->l3_addr6 = *(const struct sockaddr_in6 *)l3addr; lle->base.lle_refcnt = 1; lle->base.lle_free = in6_lltable_free; LLE_LOCK_INIT(&lle->base); callout_init_rw(&lle->base.ln_timer_ch, &lle->base.lle_lock, CALLOUT_RETURNUNLOCKED); return (&lle->base); } static void in6_lltable_prefix_free(struct lltable *llt, const struct sockaddr *prefix, const struct sockaddr *mask, u_int flags) { const struct sockaddr_in6 *pfx = (const struct sockaddr_in6 *)prefix; const struct sockaddr_in6 *msk = (const struct sockaddr_in6 *)mask; struct llentry *lle, *next; int i; /* * (flags & LLE_STATIC) means deleting all entries * including static ND6 entries. */ IF_AFDATA_WLOCK(llt->llt_ifp); for (i = 0; i < LLTBL_HASHTBL_SIZE; i++) { LIST_FOREACH_SAFE(lle, &llt->lle_head[i], lle_next, next) { if (IN6_ARE_MASKED_ADDR_EQUAL( &satosin6(L3_ADDR(lle))->sin6_addr, &pfx->sin6_addr, &msk->sin6_addr) && ((flags & LLE_STATIC) || !(lle->la_flags & LLE_STATIC))) { LLE_WLOCK(lle); if (callout_stop(&lle->la_timer)) LLE_REMREF(lle); llentry_free(lle); } } } IF_AFDATA_WUNLOCK(llt->llt_ifp); } static int in6_lltable_rtcheck(struct ifnet *ifp, u_int flags, const struct sockaddr *l3addr) { struct rtentry *rt; char ip6buf[INET6_ADDRSTRLEN]; KASSERT(l3addr->sa_family == AF_INET6, ("sin_family %d", l3addr->sa_family)); /* Our local addresses are always only installed on the default FIB. */ /* XXX rtalloc1 should take a const param */ rt = in6_rtalloc1(__DECONST(struct sockaddr *, l3addr), 0, 0, RT_DEFAULT_FIB); if (rt == NULL || (rt->rt_flags & RTF_GATEWAY) || rt->rt_ifp != ifp) { struct ifaddr *ifa; /* * Create an ND6 cache for an IPv6 neighbor * that is not covered by our own prefix. */ /* XXX ifaof_ifpforaddr should take a const param */ ifa = ifaof_ifpforaddr(__DECONST(struct sockaddr *, l3addr), ifp); if (ifa != NULL) { ifa_free(ifa); if (rt != NULL) RTFREE_LOCKED(rt); return 0; } log(LOG_INFO, "IPv6 address: \"%s\" is not on the network\n", ip6_sprintf(ip6buf, &((const struct sockaddr_in6 *)l3addr)->sin6_addr)); if (rt != NULL) RTFREE_LOCKED(rt); return EINVAL; } RTFREE_LOCKED(rt); return 0; } static struct llentry * in6_lltable_lookup(struct lltable *llt, u_int flags, const struct sockaddr *l3addr) { const struct sockaddr_in6 *sin6 = (const struct sockaddr_in6 *)l3addr; struct ifnet *ifp = llt->llt_ifp; struct llentry *lle; struct llentries *lleh; u_int hashkey; IF_AFDATA_LOCK_ASSERT(ifp); KASSERT(l3addr->sa_family == AF_INET6, ("sin_family %d", l3addr->sa_family)); hashkey = sin6->sin6_addr.s6_addr32[3]; lleh = &llt->lle_head[LLATBL_HASH(hashkey, LLTBL_HASHMASK)]; LIST_FOREACH(lle, lleh, lle_next) { struct sockaddr_in6 *sa6 = (struct sockaddr_in6 *)L3_ADDR(lle); if (lle->la_flags & LLE_DELETED) continue; if (bcmp(&sa6->sin6_addr, &sin6->sin6_addr, sizeof(struct in6_addr)) == 0) break; } if (lle == NULL) { if (!(flags & LLE_CREATE)) return (NULL); /* * A route that covers the given address must have * been installed 1st because we are doing a resolution, * verify this. */ if (!(flags & LLE_IFADDR) && in6_lltable_rtcheck(ifp, flags, l3addr) != 0) return NULL; lle = in6_lltable_new(l3addr, flags); if (lle == NULL) { log(LOG_INFO, "lla_lookup: new lle malloc failed\n"); return NULL; } lle->la_flags = flags & ~LLE_CREATE; if ((flags & (LLE_CREATE | LLE_IFADDR)) == (LLE_CREATE | LLE_IFADDR)) { bcopy(IF_LLADDR(ifp), &lle->ll_addr, ifp->if_addrlen); lle->la_flags |= (LLE_VALID | LLE_STATIC); } lle->lle_tbl = llt; lle->lle_head = lleh; lle->la_flags |= LLE_LINKED; LIST_INSERT_HEAD(lleh, lle, lle_next); } else if (flags & LLE_DELETE) { if (!(lle->la_flags & LLE_IFADDR) || (flags & LLE_IFADDR)) { LLE_WLOCK(lle); lle->la_flags |= LLE_DELETED; #ifdef DIAGNOSTIC log(LOG_INFO, "ifaddr cache = %p is deleted\n", lle); #endif if ((lle->la_flags & (LLE_STATIC | LLE_IFADDR)) == LLE_STATIC) llentry_free(lle); else LLE_WUNLOCK(lle); } lle = (void *)-1; } if (LLE_IS_VALID(lle)) { if (flags & LLE_EXCLUSIVE) LLE_WLOCK(lle); else LLE_RLOCK(lle); } return (lle); } static int in6_lltable_dump(struct lltable *llt, struct sysctl_req *wr) { struct ifnet *ifp = llt->llt_ifp; struct llentry *lle; /* XXX stack use */ struct { struct rt_msghdr rtm; struct sockaddr_in6 sin6; /* * ndp.c assumes that sdl is word aligned */ #ifdef __LP64__ uint32_t pad; #endif struct sockaddr_dl sdl; } ndpc; int i, error; if (ifp->if_flags & IFF_LOOPBACK) return 0; LLTABLE_LOCK_ASSERT(); error = 0; for (i = 0; i < LLTBL_HASHTBL_SIZE; i++) { LIST_FOREACH(lle, &llt->lle_head[i], lle_next) { struct sockaddr_dl *sdl; /* skip deleted or invalid entries */ if ((lle->la_flags & (LLE_DELETED|LLE_VALID)) != LLE_VALID) continue; /* Skip if jailed and not a valid IP of the prison. */ if (prison_if(wr->td->td_ucred, L3_ADDR(lle)) != 0) continue; /* * produce a msg made of: * struct rt_msghdr; * struct sockaddr_in6 (IPv6) * struct sockaddr_dl; */ bzero(&ndpc, sizeof(ndpc)); ndpc.rtm.rtm_msglen = sizeof(ndpc); ndpc.rtm.rtm_version = RTM_VERSION; ndpc.rtm.rtm_type = RTM_GET; ndpc.rtm.rtm_flags = RTF_UP; ndpc.rtm.rtm_addrs = RTA_DST | RTA_GATEWAY; ndpc.sin6.sin6_family = AF_INET6; ndpc.sin6.sin6_len = sizeof(ndpc.sin6); bcopy(L3_ADDR(lle), &ndpc.sin6, L3_ADDR_LEN(lle)); if (V_deembed_scopeid) sa6_recoverscope(&ndpc.sin6); /* publish */ if (lle->la_flags & LLE_PUB) ndpc.rtm.rtm_flags |= RTF_ANNOUNCE; sdl = &ndpc.sdl; sdl->sdl_family = AF_LINK; sdl->sdl_len = sizeof(*sdl); sdl->sdl_alen = ifp->if_addrlen; sdl->sdl_index = ifp->if_index; sdl->sdl_type = ifp->if_type; bcopy(&lle->ll_addr, LLADDR(sdl), ifp->if_addrlen); ndpc.rtm.rtm_rmx.rmx_expire = lle->la_flags & LLE_STATIC ? 0 : lle->la_expire; ndpc.rtm.rtm_flags |= (RTF_HOST | RTF_LLDATA); if (lle->la_flags & LLE_STATIC) ndpc.rtm.rtm_flags |= RTF_STATIC; ndpc.rtm.rtm_index = ifp->if_index; error = SYSCTL_OUT(wr, &ndpc, sizeof(ndpc)); if (error) break; } } return error; } void * in6_domifattach(struct ifnet *ifp) { struct in6_ifextra *ext; /* There are not IPv6-capable interfaces. */ switch (ifp->if_type) { case IFT_PFLOG: case IFT_PFSYNC: case IFT_USB: return (NULL); } ext = (struct in6_ifextra *)malloc(sizeof(*ext), M_IFADDR, M_WAITOK); bzero(ext, sizeof(*ext)); ext->in6_ifstat = malloc(sizeof(counter_u64_t) * sizeof(struct in6_ifstat) / sizeof(uint64_t), M_IFADDR, M_WAITOK); COUNTER_ARRAY_ALLOC(ext->in6_ifstat, sizeof(struct in6_ifstat) / sizeof(uint64_t), M_WAITOK); ext->icmp6_ifstat = malloc(sizeof(counter_u64_t) * sizeof(struct icmp6_ifstat) / sizeof(uint64_t), M_IFADDR, M_WAITOK); COUNTER_ARRAY_ALLOC(ext->icmp6_ifstat, sizeof(struct icmp6_ifstat) / sizeof(uint64_t), M_WAITOK); ext->nd_ifinfo = nd6_ifattach(ifp); ext->scope6_id = scope6_ifattach(ifp); ext->lltable = lltable_init(ifp, AF_INET6); if (ext->lltable != NULL) { ext->lltable->llt_prefix_free = in6_lltable_prefix_free; ext->lltable->llt_lookup = in6_lltable_lookup; ext->lltable->llt_dump = in6_lltable_dump; } ext->mld_ifinfo = mld_domifattach(ifp); return ext; } void in6_domifdetach(struct ifnet *ifp, void *aux) { struct in6_ifextra *ext = (struct in6_ifextra *)aux; mld_domifdetach(ifp); scope6_ifdetach(ext->scope6_id); nd6_ifdetach(ext->nd_ifinfo); lltable_free(ext->lltable); COUNTER_ARRAY_FREE(ext->in6_ifstat, sizeof(struct in6_ifstat) / sizeof(uint64_t)); free(ext->in6_ifstat, M_IFADDR); COUNTER_ARRAY_FREE(ext->icmp6_ifstat, sizeof(struct icmp6_ifstat) / sizeof(uint64_t)); free(ext->icmp6_ifstat, M_IFADDR); free(ext, M_IFADDR); } /* * Convert sockaddr_in6 to sockaddr_in. Original sockaddr_in6 must be * v4 mapped addr or v4 compat addr */ void in6_sin6_2_sin(struct sockaddr_in *sin, struct sockaddr_in6 *sin6) { bzero(sin, sizeof(*sin)); sin->sin_len = sizeof(struct sockaddr_in); sin->sin_family = AF_INET; sin->sin_port = sin6->sin6_port; sin->sin_addr.s_addr = sin6->sin6_addr.s6_addr32[3]; } /* Convert sockaddr_in to sockaddr_in6 in v4 mapped addr format. */ void in6_sin_2_v4mapsin6(struct sockaddr_in *sin, struct sockaddr_in6 *sin6) { bzero(sin6, sizeof(*sin6)); sin6->sin6_len = sizeof(struct sockaddr_in6); sin6->sin6_family = AF_INET6; sin6->sin6_port = sin->sin_port; sin6->sin6_addr.s6_addr32[0] = 0; sin6->sin6_addr.s6_addr32[1] = 0; sin6->sin6_addr.s6_addr32[2] = IPV6_ADDR_INT32_SMP; sin6->sin6_addr.s6_addr32[3] = sin->sin_addr.s_addr; } /* Convert sockaddr_in6 into sockaddr_in. */ void in6_sin6_2_sin_in_sock(struct sockaddr *nam) { struct sockaddr_in *sin_p; struct sockaddr_in6 sin6; /* * Save original sockaddr_in6 addr and convert it * to sockaddr_in. */ sin6 = *(struct sockaddr_in6 *)nam; sin_p = (struct sockaddr_in *)nam; in6_sin6_2_sin(sin_p, &sin6); } /* Convert sockaddr_in into sockaddr_in6 in v4 mapped addr format. */ void in6_sin_2_v4mapsin6_in_sock(struct sockaddr **nam) { struct sockaddr_in *sin_p; struct sockaddr_in6 *sin6_p; sin6_p = malloc(sizeof *sin6_p, M_SONAME, M_WAITOK); sin_p = (struct sockaddr_in *)*nam; in6_sin_2_v4mapsin6(sin_p, sin6_p); free(*nam, M_SONAME); *nam = (struct sockaddr *)sin6_p; } Index: head/sys/netinet6/in6.h =================================================================== --- head/sys/netinet6/in6.h (revision 253969) +++ head/sys/netinet6/in6.h (revision 253970) @@ -1,718 +1,718 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $KAME: in6.h,v 1.89 2001/05/27 13:28:35 itojun Exp $ */ /*- * Copyright (c) 1982, 1986, 1990, 1993 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)in.h 8.3 (Berkeley) 1/3/94 * $FreeBSD$ */ #ifndef __KAME_NETINET_IN_H_INCLUDED_ #error "do not include netinet6/in6.h directly, include netinet/in.h. see RFC2553" #endif #ifndef _NETINET6_IN6_H_ #define _NETINET6_IN6_H_ /* * Identification of the network protocol stack * for *BSD-current/release: http://www.kame.net/dev/cvsweb.cgi/kame/COVERAGE * has the table of implementation/integration differences. */ #define __KAME__ #define __KAME_VERSION "FreeBSD" /* * IPv6 port allocation rules should mirror the IPv4 rules and are controlled * by the net.inet.ip.portrange sysctl tree. The following defines exist * for compatibility with userland applications that need them. */ #if __BSD_VISIBLE #define IPV6PORT_RESERVED 1024 #define IPV6PORT_ANONMIN 49152 #define IPV6PORT_ANONMAX 65535 #define IPV6PORT_RESERVEDMIN 600 #define IPV6PORT_RESERVEDMAX (IPV6PORT_RESERVED-1) #endif /* * IPv6 address */ struct in6_addr { union { uint8_t __u6_addr8[16]; uint16_t __u6_addr16[8]; uint32_t __u6_addr32[4]; } __u6_addr; /* 128-bit IP6 address */ }; #define s6_addr __u6_addr.__u6_addr8 #ifdef _KERNEL /* XXX nonstandard */ #define s6_addr8 __u6_addr.__u6_addr8 #define s6_addr16 __u6_addr.__u6_addr16 #define s6_addr32 __u6_addr.__u6_addr32 #endif #define INET6_ADDRSTRLEN 46 /* * XXX missing POSIX.1-2001 macro IPPROTO_IPV6. */ /* * Socket address for IPv6 */ #if __BSD_VISIBLE #define SIN6_LEN #endif struct sockaddr_in6 { uint8_t sin6_len; /* length of this struct */ sa_family_t sin6_family; /* AF_INET6 */ in_port_t sin6_port; /* Transport layer port # */ uint32_t sin6_flowinfo; /* IP6 flow information */ struct in6_addr sin6_addr; /* IP6 address */ uint32_t sin6_scope_id; /* scope zone index */ }; /* * Local definition for masks */ #ifdef _KERNEL /* XXX nonstandard */ #define IN6MASK0 {{{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }}} #define IN6MASK32 {{{ 0xff, 0xff, 0xff, 0xff, 0x00, 0x00, 0x00, 0x00, \ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 }}} #define IN6MASK64 {{{ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, \ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 }}} #define IN6MASK96 {{{ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, \ 0xff, 0xff, 0xff, 0xff, 0x00, 0x00, 0x00, 0x00 }}} #define IN6MASK128 {{{ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, \ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff }}} #endif #ifdef _KERNEL extern const struct sockaddr_in6 sa6_any; extern const struct in6_addr in6mask0; extern const struct in6_addr in6mask32; extern const struct in6_addr in6mask64; extern const struct in6_addr in6mask96; extern const struct in6_addr in6mask128; #endif /* _KERNEL */ /* * Macros started with IPV6_ADDR is KAME local */ #ifdef _KERNEL /* XXX nonstandard */ #if _BYTE_ORDER == _BIG_ENDIAN #define IPV6_ADDR_INT32_ONE 1 #define IPV6_ADDR_INT32_TWO 2 #define IPV6_ADDR_INT32_MNL 0xff010000 #define IPV6_ADDR_INT32_MLL 0xff020000 #define IPV6_ADDR_INT32_SMP 0x0000ffff #define IPV6_ADDR_INT16_ULL 0xfe80 #define IPV6_ADDR_INT16_USL 0xfec0 #define IPV6_ADDR_INT16_MLL 0xff02 #elif _BYTE_ORDER == _LITTLE_ENDIAN #define IPV6_ADDR_INT32_ONE 0x01000000 #define IPV6_ADDR_INT32_TWO 0x02000000 #define IPV6_ADDR_INT32_MNL 0x000001ff #define IPV6_ADDR_INT32_MLL 0x000002ff #define IPV6_ADDR_INT32_SMP 0xffff0000 #define IPV6_ADDR_INT16_ULL 0x80fe #define IPV6_ADDR_INT16_USL 0xc0fe #define IPV6_ADDR_INT16_MLL 0x02ff #endif #endif /* * Definition of some useful macros to handle IP6 addresses */ #if __BSD_VISIBLE #define IN6ADDR_ANY_INIT \ {{{ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, \ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 }}} #define IN6ADDR_LOOPBACK_INIT \ {{{ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, \ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01 }}} #define IN6ADDR_NODELOCAL_ALLNODES_INIT \ {{{ 0xff, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, \ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01 }}} #define IN6ADDR_INTFACELOCAL_ALLNODES_INIT \ {{{ 0xff, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, \ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01 }}} #define IN6ADDR_LINKLOCAL_ALLNODES_INIT \ {{{ 0xff, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, \ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01 }}} #define IN6ADDR_LINKLOCAL_ALLROUTERS_INIT \ {{{ 0xff, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, \ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02 }}} #define IN6ADDR_LINKLOCAL_ALLV2ROUTERS_INIT \ {{{ 0xff, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, \ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x16 }}} #endif extern const struct in6_addr in6addr_any; extern const struct in6_addr in6addr_loopback; #if __BSD_VISIBLE extern const struct in6_addr in6addr_nodelocal_allnodes; extern const struct in6_addr in6addr_linklocal_allnodes; extern const struct in6_addr in6addr_linklocal_allrouters; extern const struct in6_addr in6addr_linklocal_allv2routers; #endif /* * Equality * NOTE: Some of kernel programming environment (for example, openbsd/sparc) * does not supply memcmp(). For userland memcmp() is preferred as it is * in ANSI standard. */ #ifdef _KERNEL #define IN6_ARE_ADDR_EQUAL(a, b) \ (bcmp(&(a)->s6_addr[0], &(b)->s6_addr[0], sizeof(struct in6_addr)) == 0) #else #if __BSD_VISIBLE #define IN6_ARE_ADDR_EQUAL(a, b) \ (memcmp(&(a)->s6_addr[0], &(b)->s6_addr[0], sizeof(struct in6_addr)) == 0) #endif #endif /* * Unspecified */ #define IN6_IS_ADDR_UNSPECIFIED(a) \ ((a)->__u6_addr.__u6_addr32[0] == 0 && \ (a)->__u6_addr.__u6_addr32[1] == 0 && \ (a)->__u6_addr.__u6_addr32[2] == 0 && \ (a)->__u6_addr.__u6_addr32[3] == 0) /* * Loopback */ #define IN6_IS_ADDR_LOOPBACK(a) \ ((a)->__u6_addr.__u6_addr32[0] == 0 && \ (a)->__u6_addr.__u6_addr32[1] == 0 && \ (a)->__u6_addr.__u6_addr32[2] == 0 && \ (a)->__u6_addr.__u6_addr32[3] == ntohl(1)) /* * IPv4 compatible */ #define IN6_IS_ADDR_V4COMPAT(a) \ ((a)->__u6_addr.__u6_addr32[0] == 0 && \ (a)->__u6_addr.__u6_addr32[1] == 0 && \ (a)->__u6_addr.__u6_addr32[2] == 0 && \ (a)->__u6_addr.__u6_addr32[3] != 0 && \ (a)->__u6_addr.__u6_addr32[3] != ntohl(1)) /* * Mapped */ #define IN6_IS_ADDR_V4MAPPED(a) \ ((a)->__u6_addr.__u6_addr32[0] == 0 && \ (a)->__u6_addr.__u6_addr32[1] == 0 && \ (a)->__u6_addr.__u6_addr32[2] == ntohl(0x0000ffff)) /* * KAME Scope Values */ #ifdef _KERNEL /* XXX nonstandard */ #define IPV6_ADDR_SCOPE_NODELOCAL 0x01 #define IPV6_ADDR_SCOPE_INTFACELOCAL 0x01 #define IPV6_ADDR_SCOPE_LINKLOCAL 0x02 #define IPV6_ADDR_SCOPE_SITELOCAL 0x05 #define IPV6_ADDR_SCOPE_ORGLOCAL 0x08 /* just used in this file */ #define IPV6_ADDR_SCOPE_GLOBAL 0x0e #else #define __IPV6_ADDR_SCOPE_NODELOCAL 0x01 #define __IPV6_ADDR_SCOPE_INTFACELOCAL 0x01 #define __IPV6_ADDR_SCOPE_LINKLOCAL 0x02 #define __IPV6_ADDR_SCOPE_SITELOCAL 0x05 #define __IPV6_ADDR_SCOPE_ORGLOCAL 0x08 /* just used in this file */ #define __IPV6_ADDR_SCOPE_GLOBAL 0x0e #endif /* * Unicast Scope * Note that we must check topmost 10 bits only, not 16 bits (see RFC2373). */ #define IN6_IS_ADDR_LINKLOCAL(a) \ (((a)->s6_addr[0] == 0xfe) && (((a)->s6_addr[1] & 0xc0) == 0x80)) #define IN6_IS_ADDR_SITELOCAL(a) \ (((a)->s6_addr[0] == 0xfe) && (((a)->s6_addr[1] & 0xc0) == 0xc0)) /* * Multicast */ #define IN6_IS_ADDR_MULTICAST(a) ((a)->s6_addr[0] == 0xff) #ifdef _KERNEL /* XXX nonstandard */ #define IPV6_ADDR_MC_SCOPE(a) ((a)->s6_addr[1] & 0x0f) #else #define __IPV6_ADDR_MC_SCOPE(a) ((a)->s6_addr[1] & 0x0f) #endif /* * Multicast Scope */ #ifdef _KERNEL /* refers nonstandard items */ #define IN6_IS_ADDR_MC_NODELOCAL(a) \ (IN6_IS_ADDR_MULTICAST(a) && \ (IPV6_ADDR_MC_SCOPE(a) == IPV6_ADDR_SCOPE_NODELOCAL)) #define IN6_IS_ADDR_MC_INTFACELOCAL(a) \ (IN6_IS_ADDR_MULTICAST(a) && \ (IPV6_ADDR_MC_SCOPE(a) == IPV6_ADDR_SCOPE_INTFACELOCAL)) #define IN6_IS_ADDR_MC_LINKLOCAL(a) \ (IN6_IS_ADDR_MULTICAST(a) && \ (IPV6_ADDR_MC_SCOPE(a) == IPV6_ADDR_SCOPE_LINKLOCAL)) #define IN6_IS_ADDR_MC_SITELOCAL(a) \ (IN6_IS_ADDR_MULTICAST(a) && \ (IPV6_ADDR_MC_SCOPE(a) == IPV6_ADDR_SCOPE_SITELOCAL)) #define IN6_IS_ADDR_MC_ORGLOCAL(a) \ (IN6_IS_ADDR_MULTICAST(a) && \ (IPV6_ADDR_MC_SCOPE(a) == IPV6_ADDR_SCOPE_ORGLOCAL)) #define IN6_IS_ADDR_MC_GLOBAL(a) \ (IN6_IS_ADDR_MULTICAST(a) && \ (IPV6_ADDR_MC_SCOPE(a) == IPV6_ADDR_SCOPE_GLOBAL)) #else #define IN6_IS_ADDR_MC_NODELOCAL(a) \ (IN6_IS_ADDR_MULTICAST(a) && \ (__IPV6_ADDR_MC_SCOPE(a) == __IPV6_ADDR_SCOPE_NODELOCAL)) #define IN6_IS_ADDR_MC_LINKLOCAL(a) \ (IN6_IS_ADDR_MULTICAST(a) && \ (__IPV6_ADDR_MC_SCOPE(a) == __IPV6_ADDR_SCOPE_LINKLOCAL)) #define IN6_IS_ADDR_MC_SITELOCAL(a) \ (IN6_IS_ADDR_MULTICAST(a) && \ (__IPV6_ADDR_MC_SCOPE(a) == __IPV6_ADDR_SCOPE_SITELOCAL)) #define IN6_IS_ADDR_MC_ORGLOCAL(a) \ (IN6_IS_ADDR_MULTICAST(a) && \ (__IPV6_ADDR_MC_SCOPE(a) == __IPV6_ADDR_SCOPE_ORGLOCAL)) #define IN6_IS_ADDR_MC_GLOBAL(a) \ (IN6_IS_ADDR_MULTICAST(a) && \ (__IPV6_ADDR_MC_SCOPE(a) == __IPV6_ADDR_SCOPE_GLOBAL)) #endif #ifdef _KERNEL /* nonstandard */ /* * KAME Scope */ #define IN6_IS_SCOPE_LINKLOCAL(a) \ ((IN6_IS_ADDR_LINKLOCAL(a)) || \ (IN6_IS_ADDR_MC_LINKLOCAL(a))) #define IN6_IS_SCOPE_EMBED(a) \ ((IN6_IS_ADDR_LINKLOCAL(a)) || \ (IN6_IS_ADDR_MC_LINKLOCAL(a)) || \ (IN6_IS_ADDR_MC_INTFACELOCAL(a))) #define IFA6_IS_DEPRECATED(a) \ ((a)->ia6_lifetime.ia6t_pltime != ND6_INFINITE_LIFETIME && \ - (u_int32_t)((time_second - (a)->ia6_updatetime)) > \ + (u_int32_t)((time_uptime - (a)->ia6_updatetime)) > \ (a)->ia6_lifetime.ia6t_pltime) #define IFA6_IS_INVALID(a) \ ((a)->ia6_lifetime.ia6t_vltime != ND6_INFINITE_LIFETIME && \ - (u_int32_t)((time_second - (a)->ia6_updatetime)) > \ + (u_int32_t)((time_uptime - (a)->ia6_updatetime)) > \ (a)->ia6_lifetime.ia6t_vltime) #endif /* _KERNEL */ /* * IP6 route structure */ #if __BSD_VISIBLE struct route_in6 { struct rtentry *ro_rt; struct llentry *ro_lle; struct in6_addr *ro_ia6; int ro_flags; struct sockaddr_in6 ro_dst; }; #endif /* * Options for use with [gs]etsockopt at the IPV6 level. * First word of comment is data type; bool is stored in int. */ /* no hdrincl */ #if 0 /* the followings are relic in IPv4 and hence are disabled */ #define IPV6_OPTIONS 1 /* buf/ip6_opts; set/get IP6 options */ #define IPV6_RECVOPTS 5 /* bool; receive all IP6 opts w/dgram */ #define IPV6_RECVRETOPTS 6 /* bool; receive IP6 opts for response */ #define IPV6_RECVDSTADDR 7 /* bool; receive IP6 dst addr w/dgram */ #define IPV6_RETOPTS 8 /* ip6_opts; set/get IP6 options */ #endif #define IPV6_SOCKOPT_RESERVED1 3 /* reserved for future use */ #define IPV6_UNICAST_HOPS 4 /* int; IP6 hops */ #define IPV6_MULTICAST_IF 9 /* u_int; set/get IP6 multicast i/f */ #define IPV6_MULTICAST_HOPS 10 /* int; set/get IP6 multicast hops */ #define IPV6_MULTICAST_LOOP 11 /* u_int; set/get IP6 multicast loopback */ #define IPV6_JOIN_GROUP 12 /* ipv6_mreq; join a group membership */ #define IPV6_LEAVE_GROUP 13 /* ipv6_mreq; leave a group membership */ #define IPV6_PORTRANGE 14 /* int; range to choose for unspec port */ #define ICMP6_FILTER 18 /* icmp6_filter; icmp6 filter */ /* RFC2292 options */ #ifdef _KERNEL #define IPV6_2292PKTINFO 19 /* bool; send/recv if, src/dst addr */ #define IPV6_2292HOPLIMIT 20 /* bool; hop limit */ #define IPV6_2292NEXTHOP 21 /* bool; next hop addr */ #define IPV6_2292HOPOPTS 22 /* bool; hop-by-hop option */ #define IPV6_2292DSTOPTS 23 /* bool; destinaion option */ #define IPV6_2292RTHDR 24 /* bool; routing header */ #define IPV6_2292PKTOPTIONS 25 /* buf/cmsghdr; set/get IPv6 options */ #endif #define IPV6_CHECKSUM 26 /* int; checksum offset for raw socket */ #define IPV6_V6ONLY 27 /* bool; make AF_INET6 sockets v6 only */ #ifndef _KERNEL #define IPV6_BINDV6ONLY IPV6_V6ONLY #endif #if 1 /* IPSEC */ #define IPV6_IPSEC_POLICY 28 /* struct; get/set security policy */ #endif /* IPSEC */ #define IPV6_FAITH 29 /* bool; accept FAITH'ed connections */ #if 1 /* IPV6FIREWALL */ #define IPV6_FW_ADD 30 /* add a firewall rule to chain */ #define IPV6_FW_DEL 31 /* delete a firewall rule from chain */ #define IPV6_FW_FLUSH 32 /* flush firewall rule chain */ #define IPV6_FW_ZERO 33 /* clear single/all firewall counter(s) */ #define IPV6_FW_GET 34 /* get entire firewall rule chain */ #endif /* new socket options introduced in RFC3542 */ #define IPV6_RTHDRDSTOPTS 35 /* ip6_dest; send dst option before rthdr */ #define IPV6_RECVPKTINFO 36 /* bool; recv if, dst addr */ #define IPV6_RECVHOPLIMIT 37 /* bool; recv hop limit */ #define IPV6_RECVRTHDR 38 /* bool; recv routing header */ #define IPV6_RECVHOPOPTS 39 /* bool; recv hop-by-hop option */ #define IPV6_RECVDSTOPTS 40 /* bool; recv dst option after rthdr */ #ifdef _KERNEL #define IPV6_RECVRTHDRDSTOPTS 41 /* bool; recv dst option before rthdr */ #endif #define IPV6_USE_MIN_MTU 42 /* bool; send packets at the minimum MTU */ #define IPV6_RECVPATHMTU 43 /* bool; notify an according MTU */ #define IPV6_PATHMTU 44 /* mtuinfo; get the current path MTU (sopt), 4 bytes int; MTU notification (cmsg) */ #if 0 /*obsoleted during 2292bis -> 3542*/ #define IPV6_REACHCONF 45 /* no data; ND reachability confirm (cmsg only/not in of RFC3542) */ #endif /* more new socket options introduced in RFC3542 */ #define IPV6_PKTINFO 46 /* in6_pktinfo; send if, src addr */ #define IPV6_HOPLIMIT 47 /* int; send hop limit */ #define IPV6_NEXTHOP 48 /* sockaddr; next hop addr */ #define IPV6_HOPOPTS 49 /* ip6_hbh; send hop-by-hop option */ #define IPV6_DSTOPTS 50 /* ip6_dest; send dst option befor rthdr */ #define IPV6_RTHDR 51 /* ip6_rthdr; send routing header */ #if 0 #define IPV6_PKTOPTIONS 52 /* buf/cmsghdr; set/get IPv6 options */ /* obsoleted by RFC3542 */ #endif #define IPV6_RECVTCLASS 57 /* bool; recv traffic class values */ #define IPV6_AUTOFLOWLABEL 59 /* bool; attach flowlabel automagically */ #define IPV6_TCLASS 61 /* int; send traffic class value */ #define IPV6_DONTFRAG 62 /* bool; disable IPv6 fragmentation */ #define IPV6_PREFER_TEMPADDR 63 /* int; prefer temporary addresses as * the source address. */ #define IPV6_BINDANY 64 /* bool: allow bind to any address */ /* * The following option is private; do not use it from user applications. * It is deliberately defined to the same value as IP_MSFILTER. */ #define IPV6_MSFILTER 74 /* struct __msfilterreq; * set/get multicast source filter list. */ /* to define items, should talk with KAME guys first, for *BSD compatibility */ #define IPV6_RTHDR_LOOSE 0 /* this hop need not be a neighbor. XXX old spec */ #define IPV6_RTHDR_STRICT 1 /* this hop must be a neighbor. XXX old spec */ #define IPV6_RTHDR_TYPE_0 0 /* IPv6 routing header type 0 */ /* * Defaults and limits for options */ #define IPV6_DEFAULT_MULTICAST_HOPS 1 /* normally limit m'casts to 1 hop */ #define IPV6_DEFAULT_MULTICAST_LOOP 1 /* normally hear sends if a member */ /* * The im6o_membership vector for each socket is now dynamically allocated at * run-time, bounded by USHRT_MAX, and is reallocated when needed, sized * according to a power-of-two increment. */ #define IPV6_MIN_MEMBERSHIPS 31 #define IPV6_MAX_MEMBERSHIPS 4095 /* * Default resource limits for IPv6 multicast source filtering. * These may be modified by sysctl. */ #define IPV6_MAX_GROUP_SRC_FILTER 512 /* sources per group */ #define IPV6_MAX_SOCK_SRC_FILTER 128 /* sources per socket/group */ /* * Argument structure for IPV6_JOIN_GROUP and IPV6_LEAVE_GROUP. */ struct ipv6_mreq { struct in6_addr ipv6mr_multiaddr; unsigned int ipv6mr_interface; }; /* * IPV6_PKTINFO: Packet information(RFC2292 sec 5) */ struct in6_pktinfo { struct in6_addr ipi6_addr; /* src/dst IPv6 address */ unsigned int ipi6_ifindex; /* send/recv interface index */ }; /* * Control structure for IPV6_RECVPATHMTU socket option. */ struct ip6_mtuinfo { struct sockaddr_in6 ip6m_addr; /* or sockaddr_storage? */ uint32_t ip6m_mtu; }; /* * Argument for IPV6_PORTRANGE: * - which range to search when port is unspecified at bind() or connect() */ #define IPV6_PORTRANGE_DEFAULT 0 /* default range */ #define IPV6_PORTRANGE_HIGH 1 /* "high" - request firewall bypass */ #define IPV6_PORTRANGE_LOW 2 /* "low" - vouchsafe security */ #if __BSD_VISIBLE /* * Definitions for inet6 sysctl operations. * * Third level is protocol number. * Fourth level is desired variable within that protocol. */ #define IPV6PROTO_MAXID (IPPROTO_PIM + 1) /* don't list to IPV6PROTO_MAX */ /* * Names for IP sysctl objects */ #define IPV6CTL_FORWARDING 1 /* act as router */ #define IPV6CTL_SENDREDIRECTS 2 /* may send redirects when forwarding*/ #define IPV6CTL_DEFHLIM 3 /* default Hop-Limit */ #ifdef notyet #define IPV6CTL_DEFMTU 4 /* default MTU */ #endif #define IPV6CTL_FORWSRCRT 5 /* forward source-routed dgrams */ #define IPV6CTL_STATS 6 /* stats */ #define IPV6CTL_MRTSTATS 7 /* multicast forwarding stats */ #define IPV6CTL_MRTPROTO 8 /* multicast routing protocol */ #define IPV6CTL_MAXFRAGPACKETS 9 /* max packets reassembly queue */ #define IPV6CTL_SOURCECHECK 10 /* verify source route and intf */ #define IPV6CTL_SOURCECHECK_LOGINT 11 /* minimume logging interval */ #define IPV6CTL_ACCEPT_RTADV 12 #define IPV6CTL_KEEPFAITH 13 #define IPV6CTL_LOG_INTERVAL 14 #define IPV6CTL_HDRNESTLIMIT 15 #define IPV6CTL_DAD_COUNT 16 #define IPV6CTL_AUTO_FLOWLABEL 17 #define IPV6CTL_DEFMCASTHLIM 18 #define IPV6CTL_GIF_HLIM 19 /* default HLIM for gif encap packet */ #define IPV6CTL_KAME_VERSION 20 #define IPV6CTL_USE_DEPRECATED 21 /* use deprecated addr (RFC2462 5.5.4) */ #define IPV6CTL_RR_PRUNE 22 /* walk timer for router renumbering */ #if 0 /* obsolete */ #define IPV6CTL_MAPPED_ADDR 23 #endif #define IPV6CTL_V6ONLY 24 #define IPV6CTL_RTEXPIRE 25 /* cloned route expiration time */ #define IPV6CTL_RTMINEXPIRE 26 /* min value for expiration time */ #define IPV6CTL_RTMAXCACHE 27 /* trigger level for dynamic expire */ #define IPV6CTL_USETEMPADDR 32 /* use temporary addresses (RFC3041) */ #define IPV6CTL_TEMPPLTIME 33 /* preferred lifetime for tmpaddrs */ #define IPV6CTL_TEMPVLTIME 34 /* valid lifetime for tmpaddrs */ #define IPV6CTL_AUTO_LINKLOCAL 35 /* automatic link-local addr assign */ #define IPV6CTL_RIP6STATS 36 /* raw_ip6 stats */ #define IPV6CTL_PREFER_TEMPADDR 37 /* prefer temporary addr as src */ #define IPV6CTL_ADDRCTLPOLICY 38 /* get/set address selection policy */ #define IPV6CTL_USE_DEFAULTZONE 39 /* use default scope zone */ #define IPV6CTL_MAXFRAGS 41 /* max fragments */ #if 0 #define IPV6CTL_IFQ 42 /* ip6intrq node */ #define IPV6CTL_ISATAPRTR 43 /* isatap router */ #endif #define IPV6CTL_MCAST_PMTU 44 /* enable pMTU discovery for multicast? */ /* New entries should be added here from current IPV6CTL_MAXID value. */ /* to define items, should talk with KAME guys first, for *BSD compatibility */ #define IPV6CTL_STEALTH 45 #define ICMPV6CTL_ND6_ONLINKNSRFC4861 47 #define IPV6CTL_NO_RADR 48 /* No defroute from RA */ #define IPV6CTL_NORBIT_RAIF 49 /* Disable R-bit in NA on RA * receiving IF. */ #define IPV6CTL_RFC6204W3 50 /* Accept defroute even when forwarding enabled */ #define IPV6CTL_MAXID 51 #endif /* __BSD_VISIBLE */ /* * Redefinition of mbuf flags */ #define M_AUTHIPHDR M_PROTO2 #define M_DECRYPTED M_PROTO3 #define M_LOOP M_PROTO4 #define M_AUTHIPDGM M_PROTO5 #define M_RTALERT_MLD M_PROTO6 #ifdef _KERNEL struct cmsghdr; struct ip6_hdr; int in6_cksum_pseudo(struct ip6_hdr *, uint32_t, uint8_t, uint16_t); int in6_cksum(struct mbuf *, u_int8_t, u_int32_t, u_int32_t); int in6_localaddr(struct in6_addr *); int in6_localip(struct in6_addr *); int in6_addrscope(struct in6_addr *); struct in6_ifaddr *in6_ifawithifp(struct ifnet *, struct in6_addr *); extern void in6_if_up(struct ifnet *); struct sockaddr; extern u_char ip6_protox[]; void in6_sin6_2_sin(struct sockaddr_in *sin, struct sockaddr_in6 *sin6); void in6_sin_2_v4mapsin6(struct sockaddr_in *sin, struct sockaddr_in6 *sin6); void in6_sin6_2_sin_in_sock(struct sockaddr *nam); void in6_sin_2_v4mapsin6_in_sock(struct sockaddr **nam); extern void addrsel_policy_init(void); #define satosin6(sa) ((struct sockaddr_in6 *)(sa)) #define sin6tosa(sin6) ((struct sockaddr *)(sin6)) #define ifatoia6(ifa) ((struct in6_ifaddr *)(ifa)) extern int (*faithprefix_p)(struct in6_addr *); #endif /* _KERNEL */ #ifndef _SIZE_T_DECLARED typedef __size_t size_t; #define _SIZE_T_DECLARED #endif #ifndef _SOCKLEN_T_DECLARED typedef __socklen_t socklen_t; #define _SOCKLEN_T_DECLARED #endif #if __BSD_VISIBLE __BEGIN_DECLS struct cmsghdr; extern int inet6_option_space(int); extern int inet6_option_init(void *, struct cmsghdr **, int); extern int inet6_option_append(struct cmsghdr *, const uint8_t *, int, int); extern uint8_t *inet6_option_alloc(struct cmsghdr *, int, int, int); extern int inet6_option_next(const struct cmsghdr *, uint8_t **); extern int inet6_option_find(const struct cmsghdr *, uint8_t **, int); extern size_t inet6_rthdr_space(int, int); extern struct cmsghdr *inet6_rthdr_init(void *, int); extern int inet6_rthdr_add(struct cmsghdr *, const struct in6_addr *, unsigned int); extern int inet6_rthdr_lasthop(struct cmsghdr *, unsigned int); #if 0 /* not implemented yet */ extern int inet6_rthdr_reverse(const struct cmsghdr *, struct cmsghdr *); #endif extern int inet6_rthdr_segments(const struct cmsghdr *); extern struct in6_addr *inet6_rthdr_getaddr(struct cmsghdr *, int); extern int inet6_rthdr_getflags(const struct cmsghdr *, int); extern int inet6_opt_init(void *, socklen_t); extern int inet6_opt_append(void *, socklen_t, int, uint8_t, socklen_t, uint8_t, void **); extern int inet6_opt_finish(void *, socklen_t, int); extern int inet6_opt_set_val(void *, int, void *, socklen_t); extern int inet6_opt_next(void *, socklen_t, int, uint8_t *, socklen_t *, void **); extern int inet6_opt_find(void *, socklen_t, int, uint8_t, socklen_t *, void **); extern int inet6_opt_get_val(void *, int, void *, socklen_t); extern socklen_t inet6_rth_space(int, int); extern void *inet6_rth_init(void *, socklen_t, int, int); extern int inet6_rth_add(void *, const struct in6_addr *); extern int inet6_rth_reverse(const void *, void *); extern int inet6_rth_segments(const void *); extern struct in6_addr *inet6_rth_getaddr(const void *, int); __END_DECLS #endif /* __BSD_VISIBLE */ #endif /* !_NETINET6_IN6_H_ */ Index: head/sys/netinet6/ip6_forward.c =================================================================== --- head/sys/netinet6/ip6_forward.c (revision 253969) +++ head/sys/netinet6/ip6_forward.c (revision 253970) @@ -1,681 +1,681 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $KAME: ip6_forward.c,v 1.69 2001/05/17 03:48:30 itojun Exp $ */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include "opt_ipfw.h" #include "opt_ipsec.h" #include "opt_ipstealth.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef IPSEC #include #include #include #endif /* IPSEC */ #include /* * Forward a packet. If some error occurs return the sender * an icmp packet. Note we can't always generate a meaningful * icmp message because icmp doesn't have a large enough repertoire * of codes and types. * * If not forwarding, just drop the packet. This could be confusing * if ipforwarding was zero but some routing protocol was advancing * us as a gateway to somewhere. However, we must let the routing * protocol deal with that. * */ void ip6_forward(struct mbuf *m, int srcrt) { struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); struct sockaddr_in6 *dst = NULL; struct rtentry *rt = NULL; struct route_in6 rin6; int error, type = 0, code = 0; struct mbuf *mcopy = NULL; struct ifnet *origifp; /* maybe unnecessary */ u_int32_t inzone, outzone; struct in6_addr src_in6, dst_in6, odst; #ifdef IPSEC struct secpolicy *sp = NULL; int ipsecrt = 0; #endif #ifdef SCTP int sw_csum; #endif struct m_tag *fwd_tag; char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN]; #ifdef IPSEC /* * Check AH/ESP integrity. */ /* * Don't increment ip6s_cantforward because this is the check * before forwarding packet actually. */ if (ipsec6_in_reject(m, NULL)) { IPSEC6STAT_INC(ips_in_polvio); m_freem(m); return; } #endif /* IPSEC */ /* * Do not forward packets to multicast destination (should be handled * by ip6_mforward(). * Do not forward packets with unspecified source. It was discussed * in July 2000, on the ipngwg mailing list. */ if ((m->m_flags & (M_BCAST|M_MCAST)) != 0 || IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) || IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src)) { IP6STAT_INC(ip6s_cantforward); /* XXX in6_ifstat_inc(rt->rt_ifp, ifs6_in_discard) */ - if (V_ip6_log_time + V_ip6_log_interval < time_second) { - V_ip6_log_time = time_second; + if (V_ip6_log_time + V_ip6_log_interval < time_uptime) { + V_ip6_log_time = time_uptime; log(LOG_DEBUG, "cannot forward " "from %s to %s nxt %d received on %s\n", ip6_sprintf(ip6bufs, &ip6->ip6_src), ip6_sprintf(ip6bufd, &ip6->ip6_dst), ip6->ip6_nxt, if_name(m->m_pkthdr.rcvif)); } m_freem(m); return; } #ifdef IPSTEALTH if (!V_ip6stealth) { #endif if (ip6->ip6_hlim <= IPV6_HLIMDEC) { /* XXX in6_ifstat_inc(rt->rt_ifp, ifs6_in_discard) */ icmp6_error(m, ICMP6_TIME_EXCEEDED, ICMP6_TIME_EXCEED_TRANSIT, 0); return; } ip6->ip6_hlim -= IPV6_HLIMDEC; #ifdef IPSTEALTH } #endif /* * Save at most ICMPV6_PLD_MAXLEN (= the min IPv6 MTU - * size of IPv6 + ICMPv6 headers) bytes of the packet in case * we need to generate an ICMP6 message to the src. * Thanks to M_EXT, in most cases copy will not occur. * * It is important to save it before IPsec processing as IPsec * processing may modify the mbuf. */ mcopy = m_copy(m, 0, imin(m->m_pkthdr.len, ICMPV6_PLD_MAXLEN)); #ifdef IPSEC /* get a security policy for this packet */ sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_OUTBOUND, IP_FORWARDING, &error); if (sp == NULL) { IPSEC6STAT_INC(ips_out_inval); IP6STAT_INC(ip6s_cantforward); if (mcopy) { #if 0 /* XXX: what icmp ? */ #else m_freem(mcopy); #endif } m_freem(m); return; } error = 0; /* check policy */ switch (sp->policy) { case IPSEC_POLICY_DISCARD: /* * This packet is just discarded. */ IPSEC6STAT_INC(ips_out_polvio); IP6STAT_INC(ip6s_cantforward); KEY_FREESP(&sp); if (mcopy) { #if 0 /* XXX: what icmp ? */ #else m_freem(mcopy); #endif } m_freem(m); return; case IPSEC_POLICY_BYPASS: case IPSEC_POLICY_NONE: /* no need to do IPsec. */ KEY_FREESP(&sp); goto skip_ipsec; case IPSEC_POLICY_IPSEC: if (sp->req == NULL) { /* XXX should be panic ? */ printf("ip6_forward: No IPsec request specified.\n"); IP6STAT_INC(ip6s_cantforward); KEY_FREESP(&sp); if (mcopy) { #if 0 /* XXX: what icmp ? */ #else m_freem(mcopy); #endif } m_freem(m); return; } /* do IPsec */ break; case IPSEC_POLICY_ENTRUST: default: /* should be panic ?? */ printf("ip6_forward: Invalid policy found. %d\n", sp->policy); KEY_FREESP(&sp); goto skip_ipsec; } { struct ipsecrequest *isr = NULL; struct ipsec_output_state state; /* * when the kernel forwards a packet, it is not proper to apply * IPsec transport mode to the packet is not proper. this check * avoid from this. * at present, if there is even a transport mode SA request in the * security policy, the kernel does not apply IPsec to the packet. * this check is not enough because the following case is valid. * ipsec esp/tunnel/xxx-xxx/require esp/transport//require; */ for (isr = sp->req; isr; isr = isr->next) { if (isr->saidx.mode == IPSEC_MODE_ANY) goto doipsectunnel; if (isr->saidx.mode == IPSEC_MODE_TUNNEL) goto doipsectunnel; } /* * if there's no need for tunnel mode IPsec, skip. */ if (!isr) goto skip_ipsec; doipsectunnel: /* * All the extension headers will become inaccessible * (since they can be encrypted). * Don't panic, we need no more updates to extension headers * on inner IPv6 packet (since they are now encapsulated). * * IPv6 [ESP|AH] IPv6 [extension headers] payload */ bzero(&state, sizeof(state)); state.m = m; state.ro = NULL; /* update at ipsec6_output_tunnel() */ state.dst = NULL; /* update at ipsec6_output_tunnel() */ error = ipsec6_output_tunnel(&state, sp, 0); m = state.m; KEY_FREESP(&sp); if (error) { /* mbuf is already reclaimed in ipsec6_output_tunnel. */ switch (error) { case EHOSTUNREACH: case ENETUNREACH: case EMSGSIZE: case ENOBUFS: case ENOMEM: break; default: printf("ip6_output (ipsec): error code %d\n", error); /* FALLTHROUGH */ case ENOENT: /* don't show these error codes to the user */ break; } IP6STAT_INC(ip6s_cantforward); if (mcopy) { #if 0 /* XXX: what icmp ? */ #else m_freem(mcopy); #endif } m_freem(m); return; } else { /* * In the FAST IPSec case we have already * re-injected the packet and it has been freed * by the ipsec_done() function. So, just clean * up after ourselves. */ m = NULL; goto freecopy; } if ((m != NULL) && (ip6 != mtod(m, struct ip6_hdr *)) ){ /* * now tunnel mode headers are added. we are originating * packet instead of forwarding the packet. */ ip6_output(m, NULL, NULL, IPV6_FORWARDING/*XXX*/, NULL, NULL, NULL); goto freecopy; } /* adjust pointer */ dst = (struct sockaddr_in6 *)state.dst; rt = state.ro ? state.ro->ro_rt : NULL; if (dst != NULL && rt != NULL) ipsecrt = 1; } if (ipsecrt) goto skip_routing; skip_ipsec: #endif again: bzero(&rin6, sizeof(struct route_in6)); dst = (struct sockaddr_in6 *)&rin6.ro_dst; dst->sin6_len = sizeof(struct sockaddr_in6); dst->sin6_family = AF_INET6; dst->sin6_addr = ip6->ip6_dst; again2: rin6.ro_rt = in6_rtalloc1((struct sockaddr *)dst, 0, 0, M_GETFIB(m)); if (rin6.ro_rt != NULL) RT_UNLOCK(rin6.ro_rt); else { IP6STAT_INC(ip6s_noroute); in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_noroute); if (mcopy) { icmp6_error(mcopy, ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_NOROUTE, 0); } goto bad; } rt = rin6.ro_rt; #ifdef IPSEC skip_routing: #endif /* * Source scope check: if a packet can't be delivered to its * destination for the reason that the destination is beyond the scope * of the source address, discard the packet and return an icmp6 * destination unreachable error with Code 2 (beyond scope of source * address). We use a local copy of ip6_src, since in6_setscope() * will possibly modify its first argument. * [draft-ietf-ipngwg-icmp-v3-04.txt, Section 3.1] */ src_in6 = ip6->ip6_src; if (in6_setscope(&src_in6, rt->rt_ifp, &outzone)) { /* XXX: this should not happen */ IP6STAT_INC(ip6s_cantforward); IP6STAT_INC(ip6s_badscope); goto bad; } if (in6_setscope(&src_in6, m->m_pkthdr.rcvif, &inzone)) { IP6STAT_INC(ip6s_cantforward); IP6STAT_INC(ip6s_badscope); goto bad; } if (inzone != outzone #ifdef IPSEC && !ipsecrt #endif ) { IP6STAT_INC(ip6s_cantforward); IP6STAT_INC(ip6s_badscope); in6_ifstat_inc(rt->rt_ifp, ifs6_in_discard); - if (V_ip6_log_time + V_ip6_log_interval < time_second) { - V_ip6_log_time = time_second; + if (V_ip6_log_time + V_ip6_log_interval < time_uptime) { + V_ip6_log_time = time_uptime; log(LOG_DEBUG, "cannot forward " "src %s, dst %s, nxt %d, rcvif %s, outif %s\n", ip6_sprintf(ip6bufs, &ip6->ip6_src), ip6_sprintf(ip6bufd, &ip6->ip6_dst), ip6->ip6_nxt, if_name(m->m_pkthdr.rcvif), if_name(rt->rt_ifp)); } if (mcopy) icmp6_error(mcopy, ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_BEYONDSCOPE, 0); goto bad; } /* * Destination scope check: if a packet is going to break the scope * zone of packet's destination address, discard it. This case should * usually be prevented by appropriately-configured routing table, but * we need an explicit check because we may mistakenly forward the * packet to a different zone by (e.g.) a default route. */ dst_in6 = ip6->ip6_dst; if (in6_setscope(&dst_in6, m->m_pkthdr.rcvif, &inzone) != 0 || in6_setscope(&dst_in6, rt->rt_ifp, &outzone) != 0 || inzone != outzone) { IP6STAT_INC(ip6s_cantforward); IP6STAT_INC(ip6s_badscope); goto bad; } if (m->m_pkthdr.len > IN6_LINKMTU(rt->rt_ifp)) { in6_ifstat_inc(rt->rt_ifp, ifs6_in_toobig); if (mcopy) { u_long mtu; #ifdef IPSEC struct secpolicy *sp; int ipsecerror; size_t ipsechdrsiz; #endif /* IPSEC */ mtu = IN6_LINKMTU(rt->rt_ifp); #ifdef IPSEC /* * When we do IPsec tunnel ingress, we need to play * with the link value (decrement IPsec header size * from mtu value). The code is much simpler than v4 * case, as we have the outgoing interface for * encapsulated packet as "rt->rt_ifp". */ sp = ipsec_getpolicybyaddr(mcopy, IPSEC_DIR_OUTBOUND, IP_FORWARDING, &ipsecerror); if (sp) { ipsechdrsiz = ipsec_hdrsiz(mcopy, IPSEC_DIR_OUTBOUND, NULL); if (ipsechdrsiz < mtu) mtu -= ipsechdrsiz; } /* * if mtu becomes less than minimum MTU, * tell minimum MTU (and I'll need to fragment it). */ if (mtu < IPV6_MMTU) mtu = IPV6_MMTU; #endif /* IPSEC */ icmp6_error(mcopy, ICMP6_PACKET_TOO_BIG, 0, mtu); } goto bad; } if (rt->rt_flags & RTF_GATEWAY) dst = (struct sockaddr_in6 *)rt->rt_gateway; /* * If we are to forward the packet using the same interface * as one we got the packet from, perhaps we should send a redirect * to sender to shortcut a hop. * Only send redirect if source is sending directly to us, * and if packet was not source routed (or has any options). * Also, don't send redirect if forwarding using a route * modified by a redirect. */ if (V_ip6_sendredirects && rt->rt_ifp == m->m_pkthdr.rcvif && !srcrt && #ifdef IPSEC !ipsecrt && #endif /* IPSEC */ (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0) { if ((rt->rt_ifp->if_flags & IFF_POINTOPOINT) != 0) { /* * If the incoming interface is equal to the outgoing * one, and the link attached to the interface is * point-to-point, then it will be highly probable * that a routing loop occurs. Thus, we immediately * drop the packet and send an ICMPv6 error message. * * type/code is based on suggestion by Rich Draves. * not sure if it is the best pick. */ icmp6_error(mcopy, ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_ADDR, 0); goto bad; } type = ND_REDIRECT; } /* * Fake scoped addresses. Note that even link-local source or * destinaion can appear, if the originating node just sends the * packet to us (without address resolution for the destination). * Since both icmp6_error and icmp6_redirect_output fill the embedded * link identifiers, we can do this stuff after making a copy for * returning an error. */ if ((rt->rt_ifp->if_flags & IFF_LOOPBACK) != 0) { /* * See corresponding comments in ip6_output. * XXX: but is it possible that ip6_forward() sends a packet * to a loopback interface? I don't think so, and thus * I bark here. (jinmei@kame.net) * XXX: it is common to route invalid packets to loopback. * also, the codepath will be visited on use of ::1 in * rthdr. (itojun) */ #if 1 if (0) #else if ((rt->rt_flags & (RTF_BLACKHOLE|RTF_REJECT)) == 0) #endif { printf("ip6_forward: outgoing interface is loopback. " "src %s, dst %s, nxt %d, rcvif %s, outif %s\n", ip6_sprintf(ip6bufs, &ip6->ip6_src), ip6_sprintf(ip6bufd, &ip6->ip6_dst), ip6->ip6_nxt, if_name(m->m_pkthdr.rcvif), if_name(rt->rt_ifp)); } /* we can just use rcvif in forwarding. */ origifp = m->m_pkthdr.rcvif; } else origifp = rt->rt_ifp; /* * clear embedded scope identifiers if necessary. * in6_clearscope will touch the addresses only when necessary. */ in6_clearscope(&ip6->ip6_src); in6_clearscope(&ip6->ip6_dst); /* Jump over all PFIL processing if hooks are not active. */ if (!PFIL_HOOKED(&V_inet6_pfil_hook)) goto pass; odst = ip6->ip6_dst; /* Run through list of hooks for output packets. */ error = pfil_run_hooks(&V_inet6_pfil_hook, &m, rt->rt_ifp, PFIL_OUT, NULL); if (error != 0) goto senderr; if (m == NULL) goto freecopy; ip6 = mtod(m, struct ip6_hdr *); /* See if destination IP address was changed by packet filter. */ if (!IN6_ARE_ADDR_EQUAL(&odst, &ip6->ip6_dst)) { m->m_flags |= M_SKIP_FIREWALL; /* If destination is now ourself drop to ip6_input(). */ if (in6_localip(&ip6->ip6_dst)) { m->m_flags |= M_FASTFWD_OURS; if (m->m_pkthdr.rcvif == NULL) m->m_pkthdr.rcvif = V_loif; if (m->m_pkthdr.csum_flags & CSUM_DELAY_DATA_IPV6) { m->m_pkthdr.csum_flags |= CSUM_DATA_VALID_IPV6 | CSUM_PSEUDO_HDR; m->m_pkthdr.csum_data = 0xffff; } #ifdef SCTP if (m->m_pkthdr.csum_flags & CSUM_SCTP_IPV6) m->m_pkthdr.csum_flags |= CSUM_SCTP_VALID; #endif error = netisr_queue(NETISR_IPV6, m); goto out; } else goto again; /* Redo the routing table lookup. */ } /* See if local, if yes, send it to netisr. */ if (m->m_flags & M_FASTFWD_OURS) { if (m->m_pkthdr.rcvif == NULL) m->m_pkthdr.rcvif = V_loif; if (m->m_pkthdr.csum_flags & CSUM_DELAY_DATA_IPV6) { m->m_pkthdr.csum_flags |= CSUM_DATA_VALID_IPV6 | CSUM_PSEUDO_HDR; m->m_pkthdr.csum_data = 0xffff; } #ifdef SCTP if (m->m_pkthdr.csum_flags & CSUM_SCTP_IPV6) m->m_pkthdr.csum_flags |= CSUM_SCTP_VALID; #endif error = netisr_queue(NETISR_IPV6, m); goto out; } /* Or forward to some other address? */ if ((m->m_flags & M_IP6_NEXTHOP) && (fwd_tag = m_tag_find(m, PACKET_TAG_IPFORWARD, NULL)) != NULL) { dst = (struct sockaddr_in6 *)&rin6.ro_dst; bcopy((fwd_tag+1), dst, sizeof(struct sockaddr_in6)); m->m_flags |= M_SKIP_FIREWALL; m->m_flags &= ~M_IP6_NEXTHOP; m_tag_delete(m, fwd_tag); goto again2; } pass: error = nd6_output(rt->rt_ifp, origifp, m, dst, rt); if (error) { in6_ifstat_inc(rt->rt_ifp, ifs6_out_discard); IP6STAT_INC(ip6s_cantforward); } else { IP6STAT_INC(ip6s_forward); in6_ifstat_inc(rt->rt_ifp, ifs6_out_forward); if (type) IP6STAT_INC(ip6s_redirectsent); else { if (mcopy) goto freecopy; } } senderr: if (mcopy == NULL) goto out; switch (error) { case 0: if (type == ND_REDIRECT) { icmp6_redirect_output(mcopy, rt); goto out; } goto freecopy; case EMSGSIZE: /* xxx MTU is constant in PPP? */ goto freecopy; case ENOBUFS: /* Tell source to slow down like source quench in IP? */ goto freecopy; case ENETUNREACH: /* shouldn't happen, checked above */ case EHOSTUNREACH: case ENETDOWN: case EHOSTDOWN: default: type = ICMP6_DST_UNREACH; code = ICMP6_DST_UNREACH_ADDR; break; } icmp6_error(mcopy, type, code, 0); goto out; freecopy: m_freem(mcopy); goto out; bad: m_freem(m); out: if (rt != NULL #ifdef IPSEC && !ipsecrt #endif ) RTFREE(rt); } Index: head/sys/netinet6/ip6_id.c =================================================================== --- head/sys/netinet6/ip6_id.c (revision 253969) +++ head/sys/netinet6/ip6_id.c (revision 253970) @@ -1,267 +1,267 @@ /*- * Copyright (C) 2003 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $KAME: ip6_id.c,v 1.13 2003/09/16 09:11:19 itojun Exp $ */ /*- * Copyright 1998 Niels Provos * All rights reserved. * * Theo de Raadt came up with the idea of using * such a mathematical system to generate more random (yet non-repeating) * ids to solve the resolver/named problem. But Niels designed the * actual system based on the constraints. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. All advertising materials mentioning features or use of this software * must display the following acknowledgement: * This product includes software developed by Niels Provos. * 4. The name of the author may not be used to endorse or promote products * derived from this software without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * $OpenBSD: ip_id.c,v 1.6 2002/03/15 18:19:52 millert Exp $ */ #include __FBSDID("$FreeBSD$"); /* * seed = random (bits - 1) bit * n = prime, g0 = generator to n, * j = random so that gcd(j,n-1) == 1 * g = g0^j mod n will be a generator again. * * X[0] = random seed. * X[n] = a*X[n-1]+b mod m is a Linear Congruential Generator * with a = 7^(even random) mod m, * b = random with gcd(b,m) == 1 * m = constant and a maximal period of m-1. * * The transaction id is determined by: * id[n] = seed xor (g^X[n] mod n) * * Effectivly the id is restricted to the lower (bits - 1) bits, thus * yielding two different cycles by toggling the msb on and off. * This avoids reuse issues caused by reseeding. */ #include #include #include #include #include #include #include #include #include #include #ifndef INT32_MAX #define INT32_MAX 0x7fffffffU #endif struct randomtab { const int ru_bits; /* resulting bits */ const long ru_out; /* Time after wich will be reseeded */ const u_int32_t ru_max; /* Uniq cycle, avoid blackjack prediction */ const u_int32_t ru_gen; /* Starting generator */ const u_int32_t ru_n; /* ru_n: prime, ru_n - 1: product of pfacts[] */ const u_int32_t ru_agen; /* determine ru_a as ru_agen^(2*rand) */ const u_int32_t ru_m; /* ru_m = 2^x*3^y */ const u_int32_t pfacts[4]; /* factors of ru_n */ u_int32_t ru_counter; u_int32_t ru_msb; u_int32_t ru_x; u_int32_t ru_seed, ru_seed2; u_int32_t ru_a, ru_b; u_int32_t ru_g; long ru_reseed; }; static struct randomtab randomtab_32 = { 32, /* resulting bits */ 180, /* Time after wich will be reseeded */ 1000000000, /* Uniq cycle, avoid blackjack prediction */ 2, /* Starting generator */ 2147483629, /* RU_N-1 = 2^2*3^2*59652323 */ 7, /* determine ru_a as RU_AGEN^(2*rand) */ 1836660096, /* RU_M = 2^7*3^15 - don't change */ { 2, 3, 59652323, 0 }, /* factors of ru_n */ }; static struct randomtab randomtab_20 = { 20, /* resulting bits */ 180, /* Time after wich will be reseeded */ 200000, /* Uniq cycle, avoid blackjack prediction */ 2, /* Starting generator */ 524269, /* RU_N-1 = 2^2*3^2*14563 */ 7, /* determine ru_a as RU_AGEN^(2*rand) */ 279936, /* RU_M = 2^7*3^7 - don't change */ { 2, 3, 14563, 0 }, /* factors of ru_n */ }; static u_int32_t pmod(u_int32_t, u_int32_t, u_int32_t); static void initid(struct randomtab *); static u_int32_t randomid(struct randomtab *); /* * Do a fast modular exponation, returned value will be in the range * of 0 - (mod-1) */ static u_int32_t pmod(u_int32_t gen, u_int32_t expo, u_int32_t mod) { u_int64_t s, t, u; s = 1; t = gen; u = expo; while (u) { if (u & 1) s = (s * t) % mod; u >>= 1; t = (t * t) % mod; } return (s); } /* * Initalizes the seed and chooses a suitable generator. Also toggles * the msb flag. The msb flag is used to generate two distinct * cycles of random numbers and thus avoiding reuse of ids. * * This function is called from id_randomid() when needed, an * application does not have to worry about it. */ static void initid(struct randomtab *p) { u_int32_t j, i; int noprime = 1; p->ru_x = arc4random() % p->ru_m; /* (bits - 1) bits of random seed */ p->ru_seed = arc4random() & (~0U >> (32 - p->ru_bits + 1)); p->ru_seed2 = arc4random() & (~0U >> (32 - p->ru_bits + 1)); /* Determine the LCG we use */ p->ru_b = (arc4random() & (~0U >> (32 - p->ru_bits))) | 1; p->ru_a = pmod(p->ru_agen, (arc4random() & (~0U >> (32 - p->ru_bits))) & (~1U), p->ru_m); while (p->ru_b % 3 == 0) p->ru_b += 2; j = arc4random() % p->ru_n; /* * Do a fast gcd(j, RU_N - 1), so we can find a j with * gcd(j, RU_N - 1) == 1, giving a new generator for * RU_GEN^j mod RU_N */ while (noprime) { for (i = 0; p->pfacts[i] > 0; i++) if (j % p->pfacts[i] == 0) break; if (p->pfacts[i] == 0) noprime = 0; else j = (j + 1) % p->ru_n; } p->ru_g = pmod(p->ru_gen, j, p->ru_n); p->ru_counter = 0; - p->ru_reseed = time_second + p->ru_out; + p->ru_reseed = time_uptime + p->ru_out; p->ru_msb = p->ru_msb ? 0 : (1U << (p->ru_bits - 1)); } static u_int32_t randomid(struct randomtab *p) { int i, n; u_int32_t tmp; - if (p->ru_counter >= p->ru_max || time_second > p->ru_reseed) + if (p->ru_counter >= p->ru_max || time_uptime > p->ru_reseed) initid(p); tmp = arc4random(); /* Skip a random number of ids */ n = tmp & 0x3; tmp = tmp >> 2; if (p->ru_counter + n >= p->ru_max) initid(p); for (i = 0; i <= n; i++) { /* Linear Congruential Generator */ p->ru_x = (u_int32_t)((u_int64_t)p->ru_a * p->ru_x + p->ru_b) % p->ru_m; } p->ru_counter += i; return (p->ru_seed ^ pmod(p->ru_g, p->ru_seed2 ^ p->ru_x, p->ru_n)) | p->ru_msb; } u_int32_t ip6_randomid(void) { return randomid(&randomtab_32); } u_int32_t ip6_randomflowlabel(void) { return randomid(&randomtab_20) & 0xfffff; } Index: head/sys/netinet6/ip6_mroute.c =================================================================== --- head/sys/netinet6/ip6_mroute.c (revision 253969) +++ head/sys/netinet6/ip6_mroute.c (revision 253970) @@ -1,2065 +1,2065 @@ /*- * Copyright (C) 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $KAME: ip6_mroute.c,v 1.58 2001/12/18 02:36:31 itojun Exp $ */ /*- * Copyright (c) 1989 Stephen Deering * Copyright (c) 1992, 1993 * The Regents of the University of California. All rights reserved. * * This code is derived from software contributed to Berkeley by * Stephen Deering of Stanford University. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)ip_mroute.c 8.2 (Berkeley) 11/15/93 * BSDI ip_mroute.c,v 2.10 1996/11/14 00:29:52 jch Exp */ /* * IP multicast forwarding procedures * * Written by David Waitzman, BBN Labs, August 1988. * Modified by Steve Deering, Stanford, February 1989. * Modified by Mark J. Steiglitz, Stanford, May, 1991 * Modified by Van Jacobson, LBL, January 1993 * Modified by Ajit Thyagarajan, PARC, August 1993 * Modified by Bill Fenner, PARC, April 1994 * * MROUTING Revision: 3.5.1.2 + PIM-SMv2 (pimd) Support */ #include __FBSDID("$FreeBSD$"); #include "opt_inet6.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include static MALLOC_DEFINE(M_MRTABLE6, "mf6c", "multicast forwarding cache entry"); /* XXX: this is a very common idiom; move to ? */ #define M_HASCL(m) ((m)->m_flags & M_EXT) static int ip6_mdq(struct mbuf *, struct ifnet *, struct mf6c *); static void phyint_send(struct ip6_hdr *, struct mif6 *, struct mbuf *); static int register_send(struct ip6_hdr *, struct mif6 *, struct mbuf *); static int set_pim6(int *); static int socket_send(struct socket *, struct mbuf *, struct sockaddr_in6 *); extern int in6_mcast_loop; extern struct domain inet6domain; static const struct encaptab *pim6_encap_cookie; static const struct ip6protosw in6_pim_protosw = { .pr_type = SOCK_RAW, .pr_domain = &inet6domain, .pr_protocol = IPPROTO_PIM, .pr_flags = PR_ATOMIC|PR_ADDR|PR_LASTHDR, .pr_input = pim6_input, .pr_output = rip6_output, .pr_ctloutput = rip6_ctloutput, .pr_usrreqs = &rip6_usrreqs }; static int pim6_encapcheck(const struct mbuf *, int, int, void *); static VNET_DEFINE(int, ip6_mrouter_ver) = 0; #define V_ip6_mrouter_ver VNET(ip6_mrouter_ver) SYSCTL_DECL(_net_inet6); SYSCTL_DECL(_net_inet6_ip6); static SYSCTL_NODE(_net_inet6, IPPROTO_PIM, pim, CTLFLAG_RW, 0, "PIM"); static struct mrt6stat mrt6stat; SYSCTL_STRUCT(_net_inet6_ip6, OID_AUTO, mrt6stat, CTLFLAG_RW, &mrt6stat, mrt6stat, "Multicast Routing Statistics (struct mrt6stat, netinet6/ip6_mroute.h)"); #define MRT6STAT_INC(name) mrt6stat.name += 1 #define NO_RTE_FOUND 0x1 #define RTE_FOUND 0x2 static struct mtx mrouter6_mtx; #define MROUTER6_LOCK() mtx_lock(&mrouter6_mtx) #define MROUTER6_UNLOCK() mtx_unlock(&mrouter6_mtx) #define MROUTER6_LOCK_ASSERT() do { \ mtx_assert(&mrouter6_mtx, MA_OWNED); \ NET_ASSERT_GIANT(); \ } while (0) #define MROUTER6_LOCK_INIT() \ mtx_init(&mrouter6_mtx, "IPv6 multicast forwarding", NULL, MTX_DEF) #define MROUTER6_LOCK_DESTROY() mtx_destroy(&mrouter6_mtx) static struct mf6c *mf6ctable[MF6CTBLSIZ]; SYSCTL_OPAQUE(_net_inet6_ip6, OID_AUTO, mf6ctable, CTLFLAG_RD, &mf6ctable, sizeof(mf6ctable), "S,*mf6ctable[MF6CTBLSIZ]", "IPv6 Multicast Forwarding Table (struct *mf6ctable[MF6CTBLSIZ], " "netinet6/ip6_mroute.h)"); static struct mtx mfc6_mtx; #define MFC6_LOCK() mtx_lock(&mfc6_mtx) #define MFC6_UNLOCK() mtx_unlock(&mfc6_mtx) #define MFC6_LOCK_ASSERT() do { \ mtx_assert(&mfc6_mtx, MA_OWNED); \ NET_ASSERT_GIANT(); \ } while (0) #define MFC6_LOCK_INIT() \ mtx_init(&mfc6_mtx, "IPv6 multicast forwarding cache", NULL, MTX_DEF) #define MFC6_LOCK_DESTROY() mtx_destroy(&mfc6_mtx) static u_char n6expire[MF6CTBLSIZ]; static struct mif6 mif6table[MAXMIFS]; SYSCTL_OPAQUE(_net_inet6_ip6, OID_AUTO, mif6table, CTLFLAG_RD, &mif6table, sizeof(mif6table), "S,mif6[MAXMIFS]", "IPv6 Multicast Interfaces (struct mif6[MAXMIFS], netinet6/ip6_mroute.h)"); static struct mtx mif6_mtx; #define MIF6_LOCK() mtx_lock(&mif6_mtx) #define MIF6_UNLOCK() mtx_unlock(&mif6_mtx) #define MIF6_LOCK_ASSERT() mtx_assert(&mif6_mtx, MA_OWNED) #define MIF6_LOCK_INIT() \ mtx_init(&mif6_mtx, "IPv6 multicast interfaces", NULL, MTX_DEF) #define MIF6_LOCK_DESTROY() mtx_destroy(&mif6_mtx) #ifdef MRT6DEBUG static VNET_DEFINE(u_int, mrt6debug) = 0; /* debug level */ #define V_mrt6debug VNET(mrt6debug) #define DEBUG_MFC 0x02 #define DEBUG_FORWARD 0x04 #define DEBUG_EXPIRE 0x08 #define DEBUG_XMIT 0x10 #define DEBUG_REG 0x20 #define DEBUG_PIM 0x40 #endif static void expire_upcalls(void *); #define EXPIRE_TIMEOUT (hz / 4) /* 4x / second */ #define UPCALL_EXPIRE 6 /* number of timeouts */ /* * XXX TODO: maintain a count to if_allmulti() calls in struct ifnet. */ /* * 'Interfaces' associated with decapsulator (so we can tell * packets that went through it from ones that get reflected * by a broken gateway). Different from IPv4 register_if, * these interfaces are linked into the system ifnet list, * because per-interface IPv6 statistics are maintained in * ifp->if_afdata. But it does not have any routes point * to them. I.e., packets can't be sent this way. They * only exist as a placeholder for multicast source * verification. */ static struct ifnet *multicast_register_if6; #define ENCAP_HOPS 64 /* * Private variables. */ static mifi_t nummifs = 0; static mifi_t reg_mif_num = (mifi_t)-1; static struct pim6stat pim6stat; SYSCTL_STRUCT(_net_inet6_pim, PIM6CTL_STATS, stats, CTLFLAG_RW, &pim6stat, pim6stat, "PIM Statistics (struct pim6stat, netinet6/pim6_var.h)"); #define PIM6STAT_INC(name) pim6stat.name += 1 static VNET_DEFINE(int, pim6); #define V_pim6 VNET(pim6) /* * Hash function for a source, group entry */ #define MF6CHASH(a, g) MF6CHASHMOD((a).s6_addr32[0] ^ (a).s6_addr32[1] ^ \ (a).s6_addr32[2] ^ (a).s6_addr32[3] ^ \ (g).s6_addr32[0] ^ (g).s6_addr32[1] ^ \ (g).s6_addr32[2] ^ (g).s6_addr32[3]) /* * Find a route for a given origin IPv6 address and Multicast group address. */ #define MF6CFIND(o, g, rt) do { \ struct mf6c *_rt = mf6ctable[MF6CHASH(o,g)]; \ rt = NULL; \ MRT6STAT_INC(mrt6s_mfc_lookups); \ while (_rt) { \ if (IN6_ARE_ADDR_EQUAL(&_rt->mf6c_origin.sin6_addr, &(o)) && \ IN6_ARE_ADDR_EQUAL(&_rt->mf6c_mcastgrp.sin6_addr, &(g)) && \ (_rt->mf6c_stall == NULL)) { \ rt = _rt; \ break; \ } \ _rt = _rt->mf6c_next; \ } \ if (rt == NULL) { \ MRT6STAT_INC(mrt6s_mfc_misses); \ } \ } while (/*CONSTCOND*/ 0) /* * Macros to compute elapsed time efficiently * Borrowed from Van Jacobson's scheduling code * XXX: replace with timersub() ? */ #define TV_DELTA(a, b, delta) do { \ int xxs; \ \ delta = (a).tv_usec - (b).tv_usec; \ if ((xxs = (a).tv_sec - (b).tv_sec)) { \ switch (xxs) { \ case 2: \ delta += 1000000; \ /* FALLTHROUGH */ \ case 1: \ delta += 1000000; \ break; \ default: \ delta += (1000000 * xxs); \ } \ } \ } while (/*CONSTCOND*/ 0) /* XXX: replace with timercmp(a, b, <) ? */ #define TV_LT(a, b) (((a).tv_usec < (b).tv_usec && \ (a).tv_sec <= (b).tv_sec) || (a).tv_sec < (b).tv_sec) #ifdef UPCALL_TIMING #define UPCALL_MAX 50 static u_long upcall_data[UPCALL_MAX + 1]; static void collate(); #endif /* UPCALL_TIMING */ static int ip6_mrouter_init(struct socket *, int, int); static int add_m6fc(struct mf6cctl *); static int add_m6if(struct mif6ctl *); static int del_m6fc(struct mf6cctl *); static int del_m6if(mifi_t *); static int del_m6if_locked(mifi_t *); static int get_mif6_cnt(struct sioc_mif_req6 *); static int get_sg_cnt(struct sioc_sg_req6 *); static struct callout expire_upcalls_ch; int X_ip6_mforward(struct ip6_hdr *, struct ifnet *, struct mbuf *); int X_ip6_mrouter_done(void); int X_ip6_mrouter_set(struct socket *, struct sockopt *); int X_ip6_mrouter_get(struct socket *, struct sockopt *); int X_mrt6_ioctl(u_long, caddr_t); /* * Handle MRT setsockopt commands to modify the multicast routing tables. */ int X_ip6_mrouter_set(struct socket *so, struct sockopt *sopt) { int error = 0; int optval; struct mif6ctl mifc; struct mf6cctl mfcc; mifi_t mifi; if (so != V_ip6_mrouter && sopt->sopt_name != MRT6_INIT) return (EACCES); switch (sopt->sopt_name) { case MRT6_INIT: #ifdef MRT6_OINIT case MRT6_OINIT: #endif error = sooptcopyin(sopt, &optval, sizeof(optval), sizeof(optval)); if (error) break; error = ip6_mrouter_init(so, optval, sopt->sopt_name); break; case MRT6_DONE: error = X_ip6_mrouter_done(); break; case MRT6_ADD_MIF: error = sooptcopyin(sopt, &mifc, sizeof(mifc), sizeof(mifc)); if (error) break; error = add_m6if(&mifc); break; case MRT6_ADD_MFC: error = sooptcopyin(sopt, &mfcc, sizeof(mfcc), sizeof(mfcc)); if (error) break; error = add_m6fc(&mfcc); break; case MRT6_DEL_MFC: error = sooptcopyin(sopt, &mfcc, sizeof(mfcc), sizeof(mfcc)); if (error) break; error = del_m6fc(&mfcc); break; case MRT6_DEL_MIF: error = sooptcopyin(sopt, &mifi, sizeof(mifi), sizeof(mifi)); if (error) break; error = del_m6if(&mifi); break; case MRT6_PIM: error = sooptcopyin(sopt, &optval, sizeof(optval), sizeof(optval)); if (error) break; error = set_pim6(&optval); break; default: error = EOPNOTSUPP; break; } return (error); } /* * Handle MRT getsockopt commands */ int X_ip6_mrouter_get(struct socket *so, struct sockopt *sopt) { int error = 0; if (so != V_ip6_mrouter) return (EACCES); switch (sopt->sopt_name) { case MRT6_PIM: error = sooptcopyout(sopt, &V_pim6, sizeof(V_pim6)); break; } return (error); } /* * Handle ioctl commands to obtain information from the cache */ int X_mrt6_ioctl(u_long cmd, caddr_t data) { int ret; ret = EINVAL; switch (cmd) { case SIOCGETSGCNT_IN6: ret = get_sg_cnt((struct sioc_sg_req6 *)data); break; case SIOCGETMIFCNT_IN6: ret = get_mif6_cnt((struct sioc_mif_req6 *)data); break; default: break; } return (ret); } /* * returns the packet, byte, rpf-failure count for the source group provided */ static int get_sg_cnt(struct sioc_sg_req6 *req) { struct mf6c *rt; int ret; ret = 0; MFC6_LOCK(); MF6CFIND(req->src.sin6_addr, req->grp.sin6_addr, rt); if (rt == NULL) { ret = ESRCH; } else { req->pktcnt = rt->mf6c_pkt_cnt; req->bytecnt = rt->mf6c_byte_cnt; req->wrong_if = rt->mf6c_wrong_if; } MFC6_UNLOCK(); return (ret); } /* * returns the input and output packet and byte counts on the mif provided */ static int get_mif6_cnt(struct sioc_mif_req6 *req) { mifi_t mifi; int ret; ret = 0; mifi = req->mifi; MIF6_LOCK(); if (mifi >= nummifs) { ret = EINVAL; } else { req->icount = mif6table[mifi].m6_pkt_in; req->ocount = mif6table[mifi].m6_pkt_out; req->ibytes = mif6table[mifi].m6_bytes_in; req->obytes = mif6table[mifi].m6_bytes_out; } MIF6_UNLOCK(); return (ret); } static int set_pim6(int *i) { if ((*i != 1) && (*i != 0)) return (EINVAL); V_pim6 = *i; return (0); } /* * Enable multicast routing */ static int ip6_mrouter_init(struct socket *so, int v, int cmd) { #ifdef MRT6DEBUG if (V_mrt6debug) log(LOG_DEBUG, "ip6_mrouter_init: so_type = %d, pr_protocol = %d\n", so->so_type, so->so_proto->pr_protocol); #endif if (so->so_type != SOCK_RAW || so->so_proto->pr_protocol != IPPROTO_ICMPV6) return (EOPNOTSUPP); if (v != 1) return (ENOPROTOOPT); MROUTER6_LOCK(); if (V_ip6_mrouter != NULL) { MROUTER6_UNLOCK(); return (EADDRINUSE); } V_ip6_mrouter = so; V_ip6_mrouter_ver = cmd; bzero((caddr_t)mf6ctable, sizeof(mf6ctable)); bzero((caddr_t)n6expire, sizeof(n6expire)); V_pim6 = 0;/* used for stubbing out/in pim stuff */ callout_init(&expire_upcalls_ch, 0); callout_reset(&expire_upcalls_ch, EXPIRE_TIMEOUT, expire_upcalls, NULL); MROUTER6_UNLOCK(); #ifdef MRT6DEBUG if (V_mrt6debug) log(LOG_DEBUG, "ip6_mrouter_init\n"); #endif return (0); } /* * Disable IPv6 multicast forwarding. */ int X_ip6_mrouter_done(void) { mifi_t mifi; int i; struct mf6c *rt; struct rtdetq *rte; MROUTER6_LOCK(); if (V_ip6_mrouter == NULL) { MROUTER6_UNLOCK(); return (EINVAL); } /* * For each phyint in use, disable promiscuous reception of all IPv6 * multicasts. */ for (mifi = 0; mifi < nummifs; mifi++) { if (mif6table[mifi].m6_ifp && !(mif6table[mifi].m6_flags & MIFF_REGISTER)) { if_allmulti(mif6table[mifi].m6_ifp, 0); } } bzero((caddr_t)mif6table, sizeof(mif6table)); nummifs = 0; V_pim6 = 0; /* used to stub out/in pim specific code */ callout_stop(&expire_upcalls_ch); /* * Free all multicast forwarding cache entries. */ MFC6_LOCK(); for (i = 0; i < MF6CTBLSIZ; i++) { rt = mf6ctable[i]; while (rt) { struct mf6c *frt; for (rte = rt->mf6c_stall; rte != NULL; ) { struct rtdetq *n = rte->next; m_free(rte->m); free(rte, M_MRTABLE6); rte = n; } frt = rt; rt = rt->mf6c_next; free(frt, M_MRTABLE6); } } bzero((caddr_t)mf6ctable, sizeof(mf6ctable)); MFC6_UNLOCK(); /* * Reset register interface */ if (reg_mif_num != (mifi_t)-1 && multicast_register_if6 != NULL) { if_detach(multicast_register_if6); if_free(multicast_register_if6); reg_mif_num = (mifi_t)-1; multicast_register_if6 = NULL; } V_ip6_mrouter = NULL; V_ip6_mrouter_ver = 0; MROUTER6_UNLOCK(); #ifdef MRT6DEBUG if (V_mrt6debug) log(LOG_DEBUG, "ip6_mrouter_done\n"); #endif return (0); } static struct sockaddr_in6 sin6 = { sizeof(sin6), AF_INET6 }; /* * Add a mif to the mif table */ static int add_m6if(struct mif6ctl *mifcp) { struct mif6 *mifp; struct ifnet *ifp; int error; MIF6_LOCK(); if (mifcp->mif6c_mifi >= MAXMIFS) { MIF6_UNLOCK(); return (EINVAL); } mifp = mif6table + mifcp->mif6c_mifi; if (mifp->m6_ifp != NULL) { MIF6_UNLOCK(); return (EADDRINUSE); /* XXX: is it appropriate? */ } if (mifcp->mif6c_pifi == 0 || mifcp->mif6c_pifi > V_if_index) { MIF6_UNLOCK(); return (ENXIO); } ifp = ifnet_byindex(mifcp->mif6c_pifi); if (mifcp->mif6c_flags & MIFF_REGISTER) { if (reg_mif_num == (mifi_t)-1) { ifp = if_alloc(IFT_OTHER); if_initname(ifp, "register_mif", 0); ifp->if_flags |= IFF_LOOPBACK; if_attach(ifp); multicast_register_if6 = ifp; reg_mif_num = mifcp->mif6c_mifi; /* * it is impossible to guess the ifindex of the * register interface. So mif6c_pifi is automatically * calculated. */ mifcp->mif6c_pifi = ifp->if_index; } else { ifp = multicast_register_if6; } } else { /* Make sure the interface supports multicast */ if ((ifp->if_flags & IFF_MULTICAST) == 0) { MIF6_UNLOCK(); return (EOPNOTSUPP); } error = if_allmulti(ifp, 1); if (error) { MIF6_UNLOCK(); return (error); } } mifp->m6_flags = mifcp->mif6c_flags; mifp->m6_ifp = ifp; /* initialize per mif pkt counters */ mifp->m6_pkt_in = 0; mifp->m6_pkt_out = 0; mifp->m6_bytes_in = 0; mifp->m6_bytes_out = 0; /* Adjust nummifs up if the mifi is higher than nummifs */ if (nummifs <= mifcp->mif6c_mifi) nummifs = mifcp->mif6c_mifi + 1; MIF6_UNLOCK(); #ifdef MRT6DEBUG if (V_mrt6debug) log(LOG_DEBUG, "add_mif #%d, phyint %s\n", mifcp->mif6c_mifi, ifp->if_xname); #endif return (0); } /* * Delete a mif from the mif table */ static int del_m6if_locked(mifi_t *mifip) { struct mif6 *mifp = mif6table + *mifip; mifi_t mifi; struct ifnet *ifp; MIF6_LOCK_ASSERT(); if (*mifip >= nummifs) return (EINVAL); if (mifp->m6_ifp == NULL) return (EINVAL); if (!(mifp->m6_flags & MIFF_REGISTER)) { /* XXX: TODO: Maintain an ALLMULTI refcount in struct ifnet. */ ifp = mifp->m6_ifp; if_allmulti(ifp, 0); } else { if (reg_mif_num != (mifi_t)-1 && multicast_register_if6 != NULL) { if_detach(multicast_register_if6); if_free(multicast_register_if6); reg_mif_num = (mifi_t)-1; multicast_register_if6 = NULL; } } bzero((caddr_t)mifp, sizeof(*mifp)); /* Adjust nummifs down */ for (mifi = nummifs; mifi > 0; mifi--) if (mif6table[mifi - 1].m6_ifp) break; nummifs = mifi; #ifdef MRT6DEBUG if (V_mrt6debug) log(LOG_DEBUG, "del_m6if %d, nummifs %d\n", *mifip, nummifs); #endif return (0); } static int del_m6if(mifi_t *mifip) { int cc; MIF6_LOCK(); cc = del_m6if_locked(mifip); MIF6_UNLOCK(); return (cc); } /* * Add an mfc entry */ static int add_m6fc(struct mf6cctl *mfccp) { struct mf6c *rt; u_long hash; struct rtdetq *rte; u_short nstl; char ip6bufo[INET6_ADDRSTRLEN], ip6bufg[INET6_ADDRSTRLEN]; MFC6_LOCK(); MF6CFIND(mfccp->mf6cc_origin.sin6_addr, mfccp->mf6cc_mcastgrp.sin6_addr, rt); /* If an entry already exists, just update the fields */ if (rt) { #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_MFC) { log(LOG_DEBUG, "add_m6fc no upcall h %d o %s g %s p %x\n", ip6_sprintf(ip6bufo, &mfccp->mf6cc_origin.sin6_addr), ip6_sprintf(ip6bufg, &mfccp->mf6cc_mcastgrp.sin6_addr), mfccp->mf6cc_parent); } #endif rt->mf6c_parent = mfccp->mf6cc_parent; rt->mf6c_ifset = mfccp->mf6cc_ifset; MFC6_UNLOCK(); return (0); } /* * Find the entry for which the upcall was made and update */ hash = MF6CHASH(mfccp->mf6cc_origin.sin6_addr, mfccp->mf6cc_mcastgrp.sin6_addr); for (rt = mf6ctable[hash], nstl = 0; rt; rt = rt->mf6c_next) { if (IN6_ARE_ADDR_EQUAL(&rt->mf6c_origin.sin6_addr, &mfccp->mf6cc_origin.sin6_addr) && IN6_ARE_ADDR_EQUAL(&rt->mf6c_mcastgrp.sin6_addr, &mfccp->mf6cc_mcastgrp.sin6_addr) && (rt->mf6c_stall != NULL)) { if (nstl++) log(LOG_ERR, "add_m6fc: %s o %s g %s p %x dbx %p\n", "multiple kernel entries", ip6_sprintf(ip6bufo, &mfccp->mf6cc_origin.sin6_addr), ip6_sprintf(ip6bufg, &mfccp->mf6cc_mcastgrp.sin6_addr), mfccp->mf6cc_parent, rt->mf6c_stall); #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_MFC) log(LOG_DEBUG, "add_m6fc o %s g %s p %x dbg %x\n", ip6_sprintf(ip6bufo, &mfccp->mf6cc_origin.sin6_addr), ip6_sprintf(ip6bufg, &mfccp->mf6cc_mcastgrp.sin6_addr), mfccp->mf6cc_parent, rt->mf6c_stall); #endif rt->mf6c_origin = mfccp->mf6cc_origin; rt->mf6c_mcastgrp = mfccp->mf6cc_mcastgrp; rt->mf6c_parent = mfccp->mf6cc_parent; rt->mf6c_ifset = mfccp->mf6cc_ifset; /* initialize pkt counters per src-grp */ rt->mf6c_pkt_cnt = 0; rt->mf6c_byte_cnt = 0; rt->mf6c_wrong_if = 0; rt->mf6c_expire = 0; /* Don't clean this guy up */ n6expire[hash]--; /* free packets Qed at the end of this entry */ for (rte = rt->mf6c_stall; rte != NULL; ) { struct rtdetq *n = rte->next; ip6_mdq(rte->m, rte->ifp, rt); m_freem(rte->m); #ifdef UPCALL_TIMING collate(&(rte->t)); #endif /* UPCALL_TIMING */ free(rte, M_MRTABLE6); rte = n; } rt->mf6c_stall = NULL; } } /* * It is possible that an entry is being inserted without an upcall */ if (nstl == 0) { #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_MFC) log(LOG_DEBUG, "add_mfc no upcall h %d o %s g %s p %x\n", hash, ip6_sprintf(ip6bufo, &mfccp->mf6cc_origin.sin6_addr), ip6_sprintf(ip6bufg, &mfccp->mf6cc_mcastgrp.sin6_addr), mfccp->mf6cc_parent); #endif for (rt = mf6ctable[hash]; rt; rt = rt->mf6c_next) { if (IN6_ARE_ADDR_EQUAL(&rt->mf6c_origin.sin6_addr, &mfccp->mf6cc_origin.sin6_addr)&& IN6_ARE_ADDR_EQUAL(&rt->mf6c_mcastgrp.sin6_addr, &mfccp->mf6cc_mcastgrp.sin6_addr)) { rt->mf6c_origin = mfccp->mf6cc_origin; rt->mf6c_mcastgrp = mfccp->mf6cc_mcastgrp; rt->mf6c_parent = mfccp->mf6cc_parent; rt->mf6c_ifset = mfccp->mf6cc_ifset; /* initialize pkt counters per src-grp */ rt->mf6c_pkt_cnt = 0; rt->mf6c_byte_cnt = 0; rt->mf6c_wrong_if = 0; if (rt->mf6c_expire) n6expire[hash]--; rt->mf6c_expire = 0; } } if (rt == NULL) { /* no upcall, so make a new entry */ rt = (struct mf6c *)malloc(sizeof(*rt), M_MRTABLE6, M_NOWAIT); if (rt == NULL) { MFC6_UNLOCK(); return (ENOBUFS); } /* insert new entry at head of hash chain */ rt->mf6c_origin = mfccp->mf6cc_origin; rt->mf6c_mcastgrp = mfccp->mf6cc_mcastgrp; rt->mf6c_parent = mfccp->mf6cc_parent; rt->mf6c_ifset = mfccp->mf6cc_ifset; /* initialize pkt counters per src-grp */ rt->mf6c_pkt_cnt = 0; rt->mf6c_byte_cnt = 0; rt->mf6c_wrong_if = 0; rt->mf6c_expire = 0; rt->mf6c_stall = NULL; /* link into table */ rt->mf6c_next = mf6ctable[hash]; mf6ctable[hash] = rt; } } MFC6_UNLOCK(); return (0); } #ifdef UPCALL_TIMING /* * collect delay statistics on the upcalls */ static void collate(struct timeval *t) { u_long d; struct timeval tp; u_long delta; GET_TIME(tp); if (TV_LT(*t, tp)) { TV_DELTA(tp, *t, delta); d = delta >> 10; if (d > UPCALL_MAX) d = UPCALL_MAX; ++upcall_data[d]; } } #endif /* UPCALL_TIMING */ /* * Delete an mfc entry */ static int del_m6fc(struct mf6cctl *mfccp) { struct sockaddr_in6 origin; struct sockaddr_in6 mcastgrp; struct mf6c *rt; struct mf6c **nptr; u_long hash; origin = mfccp->mf6cc_origin; mcastgrp = mfccp->mf6cc_mcastgrp; hash = MF6CHASH(origin.sin6_addr, mcastgrp.sin6_addr); #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_MFC) { char ip6bufo[INET6_ADDRSTRLEN], ip6bufg[INET6_ADDRSTRLEN]; log(LOG_DEBUG,"del_m6fc orig %s mcastgrp %s\n", ip6_sprintf(ip6bufo, &origin.sin6_addr), ip6_sprintf(ip6bufg, &mcastgrp.sin6_addr)); } #endif MFC6_LOCK(); nptr = &mf6ctable[hash]; while ((rt = *nptr) != NULL) { if (IN6_ARE_ADDR_EQUAL(&origin.sin6_addr, &rt->mf6c_origin.sin6_addr) && IN6_ARE_ADDR_EQUAL(&mcastgrp.sin6_addr, &rt->mf6c_mcastgrp.sin6_addr) && rt->mf6c_stall == NULL) break; nptr = &rt->mf6c_next; } if (rt == NULL) { MFC6_UNLOCK(); return (EADDRNOTAVAIL); } *nptr = rt->mf6c_next; free(rt, M_MRTABLE6); MFC6_UNLOCK(); return (0); } static int socket_send(struct socket *s, struct mbuf *mm, struct sockaddr_in6 *src) { if (s) { if (sbappendaddr(&s->so_rcv, (struct sockaddr *)src, mm, (struct mbuf *)0) != 0) { sorwakeup(s); return (0); } } m_freem(mm); return (-1); } /* * IPv6 multicast forwarding function. This function assumes that the packet * pointed to by "ip6" has arrived on (or is about to be sent to) the interface * pointed to by "ifp", and the packet is to be relayed to other networks * that have members of the packet's destination IPv6 multicast group. * * The packet is returned unscathed to the caller, unless it is * erroneous, in which case a non-zero return value tells the caller to * discard it. * * NOTE: this implementation assumes that m->m_pkthdr.rcvif is NULL iff * this function is called in the originating context (i.e., not when * forwarding a packet from other node). ip6_output(), which is currently the * only function that calls this function is called in the originating context, * explicitly ensures this condition. It is caller's responsibility to ensure * that if this function is called from somewhere else in the originating * context in the future. */ int X_ip6_mforward(struct ip6_hdr *ip6, struct ifnet *ifp, struct mbuf *m) { struct mf6c *rt; struct mif6 *mifp; struct mbuf *mm; mifi_t mifi; char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN]; #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_FORWARD) log(LOG_DEBUG, "ip6_mforward: src %s, dst %s, ifindex %d\n", ip6_sprintf(ip6bufs, &ip6->ip6_src), ip6_sprintf(ip6bufd, &ip6->ip6_dst), ifp->if_index); #endif /* * Don't forward a packet with Hop limit of zero or one, * or a packet destined to a local-only group. */ if (ip6->ip6_hlim <= 1 || IN6_IS_ADDR_MC_INTFACELOCAL(&ip6->ip6_dst) || IN6_IS_ADDR_MC_LINKLOCAL(&ip6->ip6_dst)) return (0); ip6->ip6_hlim--; /* * Source address check: do not forward packets with unspecified * source. It was discussed in July 2000, on ipngwg mailing list. * This is rather more serious than unicast cases, because some * MLD packets can be sent with the unspecified source address * (although such packets must normally set 1 to the hop limit field). */ if (IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src)) { IP6STAT_INC(ip6s_cantforward); - if (V_ip6_log_time + V_ip6_log_interval < time_second) { - V_ip6_log_time = time_second; + if (V_ip6_log_time + V_ip6_log_interval < time_uptime) { + V_ip6_log_time = time_uptime; log(LOG_DEBUG, "cannot forward " "from %s to %s nxt %d received on %s\n", ip6_sprintf(ip6bufs, &ip6->ip6_src), ip6_sprintf(ip6bufd, &ip6->ip6_dst), ip6->ip6_nxt, if_name(m->m_pkthdr.rcvif)); } return (0); } MFC6_LOCK(); /* * Determine forwarding mifs from the forwarding cache table */ MF6CFIND(ip6->ip6_src, ip6->ip6_dst, rt); /* Entry exists, so forward if necessary */ if (rt) { MFC6_UNLOCK(); return (ip6_mdq(m, ifp, rt)); } else { /* * If we don't have a route for packet's origin, * Make a copy of the packet & * send message to routing daemon */ struct mbuf *mb0; struct rtdetq *rte; u_long hash; /* int i, npkts;*/ #ifdef UPCALL_TIMING struct timeval tp; GET_TIME(tp); #endif /* UPCALL_TIMING */ MRT6STAT_INC(mrt6s_no_route); #ifdef MRT6DEBUG if (V_mrt6debug & (DEBUG_FORWARD | DEBUG_MFC)) log(LOG_DEBUG, "ip6_mforward: no rte s %s g %s\n", ip6_sprintf(ip6bufs, &ip6->ip6_src), ip6_sprintf(ip6bufd, &ip6->ip6_dst)); #endif /* * Allocate mbufs early so that we don't do extra work if we * are just going to fail anyway. */ rte = (struct rtdetq *)malloc(sizeof(*rte), M_MRTABLE6, M_NOWAIT); if (rte == NULL) { MFC6_UNLOCK(); return (ENOBUFS); } mb0 = m_copy(m, 0, M_COPYALL); /* * Pullup packet header if needed before storing it, * as other references may modify it in the meantime. */ if (mb0 && (M_HASCL(mb0) || mb0->m_len < sizeof(struct ip6_hdr))) mb0 = m_pullup(mb0, sizeof(struct ip6_hdr)); if (mb0 == NULL) { free(rte, M_MRTABLE6); MFC6_UNLOCK(); return (ENOBUFS); } /* is there an upcall waiting for this packet? */ hash = MF6CHASH(ip6->ip6_src, ip6->ip6_dst); for (rt = mf6ctable[hash]; rt; rt = rt->mf6c_next) { if (IN6_ARE_ADDR_EQUAL(&ip6->ip6_src, &rt->mf6c_origin.sin6_addr) && IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, &rt->mf6c_mcastgrp.sin6_addr) && (rt->mf6c_stall != NULL)) break; } if (rt == NULL) { struct mrt6msg *im; #ifdef MRT6_OINIT struct omrt6msg *oim; #endif /* no upcall, so make a new entry */ rt = (struct mf6c *)malloc(sizeof(*rt), M_MRTABLE6, M_NOWAIT); if (rt == NULL) { free(rte, M_MRTABLE6); m_freem(mb0); MFC6_UNLOCK(); return (ENOBUFS); } /* * Make a copy of the header to send to the user * level process */ mm = m_copy(mb0, 0, sizeof(struct ip6_hdr)); if (mm == NULL) { free(rte, M_MRTABLE6); m_freem(mb0); free(rt, M_MRTABLE6); MFC6_UNLOCK(); return (ENOBUFS); } /* * Send message to routing daemon */ sin6.sin6_addr = ip6->ip6_src; im = NULL; #ifdef MRT6_OINIT oim = NULL; #endif switch (V_ip6_mrouter_ver) { #ifdef MRT6_OINIT case MRT6_OINIT: oim = mtod(mm, struct omrt6msg *); oim->im6_msgtype = MRT6MSG_NOCACHE; oim->im6_mbz = 0; break; #endif case MRT6_INIT: im = mtod(mm, struct mrt6msg *); im->im6_msgtype = MRT6MSG_NOCACHE; im->im6_mbz = 0; break; default: free(rte, M_MRTABLE6); m_freem(mb0); free(rt, M_MRTABLE6); MFC6_UNLOCK(); return (EINVAL); } #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_FORWARD) log(LOG_DEBUG, "getting the iif info in the kernel\n"); #endif for (mifp = mif6table, mifi = 0; mifi < nummifs && mifp->m6_ifp != ifp; mifp++, mifi++) ; switch (V_ip6_mrouter_ver) { #ifdef MRT6_OINIT case MRT6_OINIT: oim->im6_mif = mifi; break; #endif case MRT6_INIT: im->im6_mif = mifi; break; } if (socket_send(V_ip6_mrouter, mm, &sin6) < 0) { log(LOG_WARNING, "ip6_mforward: ip6_mrouter " "socket queue full\n"); MRT6STAT_INC(mrt6s_upq_sockfull); free(rte, M_MRTABLE6); m_freem(mb0); free(rt, M_MRTABLE6); MFC6_UNLOCK(); return (ENOBUFS); } MRT6STAT_INC(mrt6s_upcalls); /* insert new entry at head of hash chain */ bzero(rt, sizeof(*rt)); rt->mf6c_origin.sin6_family = AF_INET6; rt->mf6c_origin.sin6_len = sizeof(struct sockaddr_in6); rt->mf6c_origin.sin6_addr = ip6->ip6_src; rt->mf6c_mcastgrp.sin6_family = AF_INET6; rt->mf6c_mcastgrp.sin6_len = sizeof(struct sockaddr_in6); rt->mf6c_mcastgrp.sin6_addr = ip6->ip6_dst; rt->mf6c_expire = UPCALL_EXPIRE; n6expire[hash]++; rt->mf6c_parent = MF6C_INCOMPLETE_PARENT; /* link into table */ rt->mf6c_next = mf6ctable[hash]; mf6ctable[hash] = rt; /* Add this entry to the end of the queue */ rt->mf6c_stall = rte; } else { /* determine if q has overflowed */ struct rtdetq **p; int npkts = 0; for (p = &rt->mf6c_stall; *p != NULL; p = &(*p)->next) if (++npkts > MAX_UPQ6) { MRT6STAT_INC(mrt6s_upq_ovflw); free(rte, M_MRTABLE6); m_freem(mb0); MFC6_UNLOCK(); return (0); } /* Add this entry to the end of the queue */ *p = rte; } rte->next = NULL; rte->m = mb0; rte->ifp = ifp; #ifdef UPCALL_TIMING rte->t = tp; #endif /* UPCALL_TIMING */ MFC6_UNLOCK(); return (0); } } /* * Clean up cache entries if upcalls are not serviced * Call from the Slow Timeout mechanism, every half second. */ static void expire_upcalls(void *unused) { struct rtdetq *rte; struct mf6c *mfc, **nptr; int i; MFC6_LOCK(); for (i = 0; i < MF6CTBLSIZ; i++) { if (n6expire[i] == 0) continue; nptr = &mf6ctable[i]; while ((mfc = *nptr) != NULL) { rte = mfc->mf6c_stall; /* * Skip real cache entries * Make sure it wasn't marked to not expire (shouldn't happen) * If it expires now */ if (rte != NULL && mfc->mf6c_expire != 0 && --mfc->mf6c_expire == 0) { #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_EXPIRE) { char ip6bufo[INET6_ADDRSTRLEN]; char ip6bufg[INET6_ADDRSTRLEN]; log(LOG_DEBUG, "expire_upcalls: expiring (%s %s)\n", ip6_sprintf(ip6bufo, &mfc->mf6c_origin.sin6_addr), ip6_sprintf(ip6bufg, &mfc->mf6c_mcastgrp.sin6_addr)); } #endif /* * drop all the packets * free the mbuf with the pkt, if, timing info */ do { struct rtdetq *n = rte->next; m_freem(rte->m); free(rte, M_MRTABLE6); rte = n; } while (rte != NULL); MRT6STAT_INC(mrt6s_cache_cleanups); n6expire[i]--; *nptr = mfc->mf6c_next; free(mfc, M_MRTABLE6); } else { nptr = &mfc->mf6c_next; } } } MFC6_UNLOCK(); callout_reset(&expire_upcalls_ch, EXPIRE_TIMEOUT, expire_upcalls, NULL); } /* * Packet forwarding routine once entry in the cache is made */ static int ip6_mdq(struct mbuf *m, struct ifnet *ifp, struct mf6c *rt) { struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); mifi_t mifi, iif; struct mif6 *mifp; int plen = m->m_pkthdr.len; struct in6_addr src0, dst0; /* copies for local work */ u_int32_t iszone, idzone, oszone, odzone; int error = 0; /* * Macro to send packet on mif. Since RSVP packets don't get counted on * input, they shouldn't get counted on output, so statistics keeping is * separate. */ #define MC6_SEND(ip6, mifp, m) do { \ if ((mifp)->m6_flags & MIFF_REGISTER) \ register_send((ip6), (mifp), (m)); \ else \ phyint_send((ip6), (mifp), (m)); \ } while (/*CONSTCOND*/ 0) /* * Don't forward if it didn't arrive from the parent mif * for its origin. */ mifi = rt->mf6c_parent; if ((mifi >= nummifs) || (mif6table[mifi].m6_ifp != ifp)) { /* came in the wrong interface */ #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_FORWARD) log(LOG_DEBUG, "wrong if: ifid %d mifi %d mififid %x\n", ifp->if_index, mifi, mif6table[mifi].m6_ifp->if_index); #endif MRT6STAT_INC(mrt6s_wrong_if); rt->mf6c_wrong_if++; /* * If we are doing PIM processing, and we are forwarding * packets on this interface, send a message to the * routing daemon. */ /* have to make sure this is a valid mif */ if (mifi < nummifs && mif6table[mifi].m6_ifp) if (V_pim6 && (m->m_flags & M_LOOP) == 0) { /* * Check the M_LOOP flag to avoid an * unnecessary PIM assert. * XXX: M_LOOP is an ad-hoc hack... */ static struct sockaddr_in6 sin6 = { sizeof(sin6), AF_INET6 }; struct mbuf *mm; struct mrt6msg *im; #ifdef MRT6_OINIT struct omrt6msg *oim; #endif mm = m_copy(m, 0, sizeof(struct ip6_hdr)); if (mm && (M_HASCL(mm) || mm->m_len < sizeof(struct ip6_hdr))) mm = m_pullup(mm, sizeof(struct ip6_hdr)); if (mm == NULL) return (ENOBUFS); #ifdef MRT6_OINIT oim = NULL; #endif im = NULL; switch (V_ip6_mrouter_ver) { #ifdef MRT6_OINIT case MRT6_OINIT: oim = mtod(mm, struct omrt6msg *); oim->im6_msgtype = MRT6MSG_WRONGMIF; oim->im6_mbz = 0; break; #endif case MRT6_INIT: im = mtod(mm, struct mrt6msg *); im->im6_msgtype = MRT6MSG_WRONGMIF; im->im6_mbz = 0; break; default: m_freem(mm); return (EINVAL); } for (mifp = mif6table, iif = 0; iif < nummifs && mifp && mifp->m6_ifp != ifp; mifp++, iif++) ; switch (V_ip6_mrouter_ver) { #ifdef MRT6_OINIT case MRT6_OINIT: oim->im6_mif = iif; sin6.sin6_addr = oim->im6_src; break; #endif case MRT6_INIT: im->im6_mif = iif; sin6.sin6_addr = im->im6_src; break; } MRT6STAT_INC(mrt6s_upcalls); if (socket_send(V_ip6_mrouter, mm, &sin6) < 0) { #ifdef MRT6DEBUG if (V_mrt6debug) log(LOG_WARNING, "mdq, ip6_mrouter socket queue full\n"); #endif MRT6STAT_INC(mrt6s_upq_sockfull); return (ENOBUFS); } /* if socket Q full */ } /* if PIM */ return (0); } /* if wrong iif */ /* If I sourced this packet, it counts as output, else it was input. */ if (m->m_pkthdr.rcvif == NULL) { /* XXX: is rcvif really NULL when output?? */ mif6table[mifi].m6_pkt_out++; mif6table[mifi].m6_bytes_out += plen; } else { mif6table[mifi].m6_pkt_in++; mif6table[mifi].m6_bytes_in += plen; } rt->mf6c_pkt_cnt++; rt->mf6c_byte_cnt += plen; /* * For each mif, forward a copy of the packet if there are group * members downstream on the interface. */ src0 = ip6->ip6_src; dst0 = ip6->ip6_dst; if ((error = in6_setscope(&src0, ifp, &iszone)) != 0 || (error = in6_setscope(&dst0, ifp, &idzone)) != 0) { IP6STAT_INC(ip6s_badscope); return (error); } for (mifp = mif6table, mifi = 0; mifi < nummifs; mifp++, mifi++) { if (IF_ISSET(mifi, &rt->mf6c_ifset)) { /* * check if the outgoing packet is going to break * a scope boundary. * XXX For packets through PIM register tunnel * interface, we believe a routing daemon. */ if (!(mif6table[rt->mf6c_parent].m6_flags & MIFF_REGISTER) && !(mif6table[mifi].m6_flags & MIFF_REGISTER)) { if (in6_setscope(&src0, mif6table[mifi].m6_ifp, &oszone) || in6_setscope(&dst0, mif6table[mifi].m6_ifp, &odzone) || iszone != oszone || idzone != odzone) { IP6STAT_INC(ip6s_badscope); continue; } } mifp->m6_pkt_out++; mifp->m6_bytes_out += plen; MC6_SEND(ip6, mifp, m); } } return (0); } static void phyint_send(struct ip6_hdr *ip6, struct mif6 *mifp, struct mbuf *m) { struct mbuf *mb_copy; struct ifnet *ifp = mifp->m6_ifp; int error = 0; u_long linkmtu; /* * Make a new reference to the packet; make sure that * the IPv6 header is actually copied, not just referenced, * so that ip6_output() only scribbles on the copy. */ mb_copy = m_copy(m, 0, M_COPYALL); if (mb_copy && (M_HASCL(mb_copy) || mb_copy->m_len < sizeof(struct ip6_hdr))) mb_copy = m_pullup(mb_copy, sizeof(struct ip6_hdr)); if (mb_copy == NULL) { return; } /* set MCAST flag to the outgoing packet */ mb_copy->m_flags |= M_MCAST; /* * If we sourced the packet, call ip6_output since we may devide * the packet into fragments when the packet is too big for the * outgoing interface. * Otherwise, we can simply send the packet to the interface * sending queue. */ if (m->m_pkthdr.rcvif == NULL) { struct ip6_moptions im6o; im6o.im6o_multicast_ifp = ifp; /* XXX: ip6_output will override ip6->ip6_hlim */ im6o.im6o_multicast_hlim = ip6->ip6_hlim; im6o.im6o_multicast_loop = 1; error = ip6_output(mb_copy, NULL, NULL, IPV6_FORWARDING, &im6o, NULL, NULL); #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_XMIT) log(LOG_DEBUG, "phyint_send on mif %d err %d\n", mifp - mif6table, error); #endif return; } /* * If configured to loop back multicasts by default, * loop back a copy now. */ if (in6_mcast_loop) { struct sockaddr_in6 dst6; bzero(&dst6, sizeof(dst6)); dst6.sin6_len = sizeof(struct sockaddr_in6); dst6.sin6_family = AF_INET6; dst6.sin6_addr = ip6->ip6_dst; ip6_mloopback(ifp, m, &dst6); } /* * Put the packet into the sending queue of the outgoing interface * if it would fit in the MTU of the interface. */ linkmtu = IN6_LINKMTU(ifp); if (mb_copy->m_pkthdr.len <= linkmtu || linkmtu < IPV6_MMTU) { struct sockaddr_in6 dst6; bzero(&dst6, sizeof(dst6)); dst6.sin6_len = sizeof(struct sockaddr_in6); dst6.sin6_family = AF_INET6; dst6.sin6_addr = ip6->ip6_dst; /* * We just call if_output instead of nd6_output here, since * we need no ND for a multicast forwarded packet...right? */ error = (*ifp->if_output)(ifp, mb_copy, (struct sockaddr *)&dst6, NULL); #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_XMIT) log(LOG_DEBUG, "phyint_send on mif %d err %d\n", mifp - mif6table, error); #endif } else { /* * pMTU discovery is intentionally disabled by default, since * various router may notify pMTU in multicast, which can be * a DDoS to a router */ if (V_ip6_mcast_pmtu) icmp6_error(mb_copy, ICMP6_PACKET_TOO_BIG, 0, linkmtu); else { #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_XMIT) { char ip6bufs[INET6_ADDRSTRLEN]; char ip6bufd[INET6_ADDRSTRLEN]; log(LOG_DEBUG, "phyint_send: packet too big on %s o %s " "g %s size %d(discarded)\n", if_name(ifp), ip6_sprintf(ip6bufs, &ip6->ip6_src), ip6_sprintf(ip6bufd, &ip6->ip6_dst), mb_copy->m_pkthdr.len); } #endif /* MRT6DEBUG */ m_freem(mb_copy); /* simply discard the packet */ } } } static int register_send(struct ip6_hdr *ip6, struct mif6 *mif, struct mbuf *m) { struct mbuf *mm; int i, len = m->m_pkthdr.len; static struct sockaddr_in6 sin6 = { sizeof(sin6), AF_INET6 }; struct mrt6msg *im6; #ifdef MRT6DEBUG if (V_mrt6debug) { char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN]; log(LOG_DEBUG, "** IPv6 register_send **\n src %s dst %s\n", ip6_sprintf(ip6bufs, &ip6->ip6_src), ip6_sprintf(ip6bufd, &ip6->ip6_dst)); } #endif PIM6STAT_INC(pim6s_snd_registers); /* Make a copy of the packet to send to the user level process. */ mm = m_gethdr(M_NOWAIT, MT_DATA); if (mm == NULL) return (ENOBUFS); mm->m_data += max_linkhdr; mm->m_len = sizeof(struct ip6_hdr); if ((mm->m_next = m_copy(m, 0, M_COPYALL)) == NULL) { m_freem(mm); return (ENOBUFS); } i = MHLEN - M_LEADINGSPACE(mm); if (i > len) i = len; mm = m_pullup(mm, i); if (mm == NULL) return (ENOBUFS); /* TODO: check it! */ mm->m_pkthdr.len = len + sizeof(struct ip6_hdr); /* * Send message to routing daemon */ sin6.sin6_addr = ip6->ip6_src; im6 = mtod(mm, struct mrt6msg *); im6->im6_msgtype = MRT6MSG_WHOLEPKT; im6->im6_mbz = 0; im6->im6_mif = mif - mif6table; /* iif info is not given for reg. encap.n */ MRT6STAT_INC(mrt6s_upcalls); if (socket_send(V_ip6_mrouter, mm, &sin6) < 0) { #ifdef MRT6DEBUG if (V_mrt6debug) log(LOG_WARNING, "register_send: ip6_mrouter socket queue full\n"); #endif MRT6STAT_INC(mrt6s_upq_sockfull); return (ENOBUFS); } return (0); } /* * pim6_encapcheck() is called by the encap6_input() path at runtime to * determine if a packet is for PIM; allowing PIM to be dynamically loaded * into the kernel. */ static int pim6_encapcheck(const struct mbuf *m, int off, int proto, void *arg) { #ifdef DIAGNOSTIC KASSERT(proto == IPPROTO_PIM, ("not for IPPROTO_PIM")); #endif if (proto != IPPROTO_PIM) return 0; /* not for us; reject the datagram. */ return 64; /* claim the datagram. */ } /* * PIM sparse mode hook * Receives the pim control messages, and passes them up to the listening * socket, using rip6_input. * The only message processed is the REGISTER pim message; the pim header * is stripped off, and the inner packet is passed to register_mforward. */ int pim6_input(struct mbuf **mp, int *offp, int proto) { struct pim *pim; /* pointer to a pim struct */ struct ip6_hdr *ip6; int pimlen; struct mbuf *m = *mp; int minlen; int off = *offp; PIM6STAT_INC(pim6s_rcv_total); ip6 = mtod(m, struct ip6_hdr *); pimlen = m->m_pkthdr.len - *offp; /* * Validate lengths */ if (pimlen < PIM_MINLEN) { PIM6STAT_INC(pim6s_rcv_tooshort); #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_PIM) log(LOG_DEBUG,"pim6_input: PIM packet too short\n"); #endif m_freem(m); return (IPPROTO_DONE); } /* * if the packet is at least as big as a REGISTER, go ahead * and grab the PIM REGISTER header size, to avoid another * possible m_pullup() later. * * PIM_MINLEN == pimhdr + u_int32 == 8 * PIM6_REG_MINLEN == pimhdr + reghdr + eip6hdr == 4 + 4 + 40 */ minlen = (pimlen >= PIM6_REG_MINLEN) ? PIM6_REG_MINLEN : PIM_MINLEN; /* * Make sure that the IP6 and PIM headers in contiguous memory, and * possibly the PIM REGISTER header */ #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, off, minlen, IPPROTO_DONE); /* adjust pointer */ ip6 = mtod(m, struct ip6_hdr *); /* adjust mbuf to point to the PIM header */ pim = (struct pim *)((caddr_t)ip6 + off); #else IP6_EXTHDR_GET(pim, struct pim *, m, off, minlen); if (pim == NULL) { PIM6STAT_INC(pim6s_rcv_tooshort); return (IPPROTO_DONE); } #endif #define PIM6_CHECKSUM #ifdef PIM6_CHECKSUM { int cksumlen; /* * Validate checksum. * If PIM REGISTER, exclude the data packet */ if (pim->pim_type == PIM_REGISTER) cksumlen = PIM_MINLEN; else cksumlen = pimlen; if (in6_cksum(m, IPPROTO_PIM, off, cksumlen)) { PIM6STAT_INC(pim6s_rcv_badsum); #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_PIM) log(LOG_DEBUG, "pim6_input: invalid checksum\n"); #endif m_freem(m); return (IPPROTO_DONE); } } #endif /* PIM_CHECKSUM */ /* PIM version check */ if (pim->pim_ver != PIM_VERSION) { PIM6STAT_INC(pim6s_rcv_badversion); #ifdef MRT6DEBUG log(LOG_ERR, "pim6_input: incorrect version %d, expecting %d\n", pim->pim_ver, PIM_VERSION); #endif m_freem(m); return (IPPROTO_DONE); } if (pim->pim_type == PIM_REGISTER) { /* * since this is a REGISTER, we'll make a copy of the register * headers ip6+pim+u_int32_t+encap_ip6, to be passed up to the * routing daemon. */ static struct sockaddr_in6 dst = { sizeof(dst), AF_INET6 }; struct mbuf *mcp; struct ip6_hdr *eip6; u_int32_t *reghdr; int rc; #ifdef MRT6DEBUG char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN]; #endif PIM6STAT_INC(pim6s_rcv_registers); if ((reg_mif_num >= nummifs) || (reg_mif_num == (mifi_t) -1)) { #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_PIM) log(LOG_DEBUG, "pim6_input: register mif not set: %d\n", reg_mif_num); #endif m_freem(m); return (IPPROTO_DONE); } reghdr = (u_int32_t *)(pim + 1); if ((ntohl(*reghdr) & PIM_NULL_REGISTER)) goto pim6_input_to_daemon; /* * Validate length */ if (pimlen < PIM6_REG_MINLEN) { PIM6STAT_INC(pim6s_rcv_tooshort); PIM6STAT_INC(pim6s_rcv_badregisters); #ifdef MRT6DEBUG log(LOG_ERR, "pim6_input: register packet size too " "small %d from %s\n", pimlen, ip6_sprintf(ip6bufs, &ip6->ip6_src)); #endif m_freem(m); return (IPPROTO_DONE); } eip6 = (struct ip6_hdr *) (reghdr + 1); #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_PIM) log(LOG_DEBUG, "pim6_input[register], eip6: %s -> %s, " "eip6 plen %d\n", ip6_sprintf(ip6bufs, &eip6->ip6_src), ip6_sprintf(ip6bufd, &eip6->ip6_dst), ntohs(eip6->ip6_plen)); #endif /* verify the version number of the inner packet */ if ((eip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) { PIM6STAT_INC(pim6s_rcv_badregisters); #ifdef MRT6DEBUG log(LOG_DEBUG, "pim6_input: invalid IP version (%d) " "of the inner packet\n", (eip6->ip6_vfc & IPV6_VERSION)); #endif m_freem(m); return (IPPROTO_NONE); } /* verify the inner packet is destined to a mcast group */ if (!IN6_IS_ADDR_MULTICAST(&eip6->ip6_dst)) { PIM6STAT_INC(pim6s_rcv_badregisters); #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_PIM) log(LOG_DEBUG, "pim6_input: inner packet of register " "is not multicast %s\n", ip6_sprintf(ip6bufd, &eip6->ip6_dst)); #endif m_freem(m); return (IPPROTO_DONE); } /* * make a copy of the whole header to pass to the daemon later. */ mcp = m_copy(m, 0, off + PIM6_REG_MINLEN); if (mcp == NULL) { #ifdef MRT6DEBUG log(LOG_ERR, "pim6_input: pim register: " "could not copy register head\n"); #endif m_freem(m); return (IPPROTO_DONE); } /* * forward the inner ip6 packet; point m_data at the inner ip6. */ m_adj(m, off + PIM_MINLEN); #ifdef MRT6DEBUG if (V_mrt6debug & DEBUG_PIM) { log(LOG_DEBUG, "pim6_input: forwarding decapsulated register: " "src %s, dst %s, mif %d\n", ip6_sprintf(ip6bufs, &eip6->ip6_src), ip6_sprintf(ip6bufd, &eip6->ip6_dst), reg_mif_num); } #endif rc = if_simloop(mif6table[reg_mif_num].m6_ifp, m, dst.sin6_family, 0); /* prepare the register head to send to the mrouting daemon */ m = mcp; } /* * Pass the PIM message up to the daemon; if it is a register message * pass the 'head' only up to the daemon. This includes the * encapsulator ip6 header, pim header, register header and the * encapsulated ip6 header. */ pim6_input_to_daemon: rip6_input(&m, offp, proto); return (IPPROTO_DONE); } static int ip6_mroute_modevent(module_t mod, int type, void *unused) { switch (type) { case MOD_LOAD: MROUTER6_LOCK_INIT(); MFC6_LOCK_INIT(); MIF6_LOCK_INIT(); pim6_encap_cookie = encap_attach_func(AF_INET6, IPPROTO_PIM, pim6_encapcheck, (const struct protosw *)&in6_pim_protosw, NULL); if (pim6_encap_cookie == NULL) { printf("ip6_mroute: unable to attach pim6 encap\n"); MIF6_LOCK_DESTROY(); MFC6_LOCK_DESTROY(); MROUTER6_LOCK_DESTROY(); return (EINVAL); } ip6_mforward = X_ip6_mforward; ip6_mrouter_done = X_ip6_mrouter_done; ip6_mrouter_get = X_ip6_mrouter_get; ip6_mrouter_set = X_ip6_mrouter_set; mrt6_ioctl = X_mrt6_ioctl; break; case MOD_UNLOAD: if (V_ip6_mrouter != NULL) return EINVAL; if (pim6_encap_cookie) { encap_detach(pim6_encap_cookie); pim6_encap_cookie = NULL; } X_ip6_mrouter_done(); ip6_mforward = NULL; ip6_mrouter_done = NULL; ip6_mrouter_get = NULL; ip6_mrouter_set = NULL; mrt6_ioctl = NULL; MIF6_LOCK_DESTROY(); MFC6_LOCK_DESTROY(); MROUTER6_LOCK_DESTROY(); break; default: return (EOPNOTSUPP); } return (0); } static moduledata_t ip6_mroutemod = { "ip6_mroute", ip6_mroute_modevent, 0 }; DECLARE_MODULE(ip6_mroute, ip6_mroutemod, SI_SUB_PSEUDO, SI_ORDER_ANY); Index: head/sys/netinet6/nd6.c =================================================================== --- head/sys/netinet6/nd6.c (revision 253969) +++ head/sys/netinet6/nd6.c (revision 253970) @@ -1,2365 +1,2365 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $KAME: nd6.c,v 1.144 2001/05/24 07:44:00 itojun Exp $ */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define L3_ADDR_SIN6(le) ((struct sockaddr_in6 *) L3_ADDR(le)) #include #include #include #include #include #include #include #include #include #include #include #define ND6_SLOWTIMER_INTERVAL (60 * 60) /* 1 hour */ #define ND6_RECALC_REACHTM_INTERVAL (60 * 120) /* 2 hours */ #define SIN6(s) ((const struct sockaddr_in6 *)(s)) /* timer values */ VNET_DEFINE(int, nd6_prune) = 1; /* walk list every 1 seconds */ VNET_DEFINE(int, nd6_delay) = 5; /* delay first probe time 5 second */ VNET_DEFINE(int, nd6_umaxtries) = 3; /* maximum unicast query */ VNET_DEFINE(int, nd6_mmaxtries) = 3; /* maximum multicast query */ VNET_DEFINE(int, nd6_useloopback) = 1; /* use loopback interface for * local traffic */ VNET_DEFINE(int, nd6_gctimer) = (60 * 60 * 24); /* 1 day: garbage * collection timer */ /* preventing too many loops in ND option parsing */ static VNET_DEFINE(int, nd6_maxndopt) = 10; /* max # of ND options allowed */ VNET_DEFINE(int, nd6_maxnudhint) = 0; /* max # of subsequent upper * layer hints */ static VNET_DEFINE(int, nd6_maxqueuelen) = 1; /* max pkts cached in unresolved * ND entries */ #define V_nd6_maxndopt VNET(nd6_maxndopt) #define V_nd6_maxqueuelen VNET(nd6_maxqueuelen) #ifdef ND6_DEBUG VNET_DEFINE(int, nd6_debug) = 1; #else VNET_DEFINE(int, nd6_debug) = 0; #endif /* for debugging? */ #if 0 static int nd6_inuse, nd6_allocated; #endif VNET_DEFINE(struct nd_drhead, nd_defrouter); VNET_DEFINE(struct nd_prhead, nd_prefix); VNET_DEFINE(int, nd6_recalc_reachtm_interval) = ND6_RECALC_REACHTM_INTERVAL; #define V_nd6_recalc_reachtm_interval VNET(nd6_recalc_reachtm_interval) int (*send_sendso_input_hook)(struct mbuf *, struct ifnet *, int, int); static int nd6_is_new_addr_neighbor(struct sockaddr_in6 *, struct ifnet *); static void nd6_setmtu0(struct ifnet *, struct nd_ifinfo *); static void nd6_slowtimo(void *); static int regen_tmpaddr(struct in6_ifaddr *); static struct llentry *nd6_free(struct llentry *, int); static void nd6_llinfo_timer(void *); static void clear_llinfo_pqueue(struct llentry *); static VNET_DEFINE(struct callout, nd6_slowtimo_ch); #define V_nd6_slowtimo_ch VNET(nd6_slowtimo_ch) VNET_DEFINE(struct callout, nd6_timer_ch); void nd6_init(void) { LIST_INIT(&V_nd_prefix); /* initialization of the default router list */ TAILQ_INIT(&V_nd_defrouter); /* start timer */ callout_init(&V_nd6_slowtimo_ch, 0); callout_reset(&V_nd6_slowtimo_ch, ND6_SLOWTIMER_INTERVAL * hz, nd6_slowtimo, curvnet); } #ifdef VIMAGE void nd6_destroy() { callout_drain(&V_nd6_slowtimo_ch); callout_drain(&V_nd6_timer_ch); } #endif struct nd_ifinfo * nd6_ifattach(struct ifnet *ifp) { struct nd_ifinfo *nd; nd = (struct nd_ifinfo *)malloc(sizeof(*nd), M_IP6NDP, M_WAITOK|M_ZERO); nd->initialized = 1; nd->chlim = IPV6_DEFHLIM; nd->basereachable = REACHABLE_TIME; nd->reachable = ND_COMPUTE_RTIME(nd->basereachable); nd->retrans = RETRANS_TIMER; nd->flags = ND6_IFF_PERFORMNUD; /* A loopback interface always has ND6_IFF_AUTO_LINKLOCAL. * XXXHRS: Clear ND6_IFF_AUTO_LINKLOCAL on an IFT_BRIDGE interface by * default regardless of the V_ip6_auto_linklocal configuration to * give a reasonable default behavior. */ if ((V_ip6_auto_linklocal && ifp->if_type != IFT_BRIDGE) || (ifp->if_flags & IFF_LOOPBACK)) nd->flags |= ND6_IFF_AUTO_LINKLOCAL; /* * A loopback interface does not need to accept RTADV. * XXXHRS: Clear ND6_IFF_ACCEPT_RTADV on an IFT_BRIDGE interface by * default regardless of the V_ip6_accept_rtadv configuration to * prevent the interface from accepting RA messages arrived * on one of the member interfaces with ND6_IFF_ACCEPT_RTADV. */ if (V_ip6_accept_rtadv && !(ifp->if_flags & IFF_LOOPBACK) && (ifp->if_type != IFT_BRIDGE)) nd->flags |= ND6_IFF_ACCEPT_RTADV; if (V_ip6_no_radr && !(ifp->if_flags & IFF_LOOPBACK)) nd->flags |= ND6_IFF_NO_RADR; /* XXX: we cannot call nd6_setmtu since ifp is not fully initialized */ nd6_setmtu0(ifp, nd); return nd; } void nd6_ifdetach(struct nd_ifinfo *nd) { free(nd, M_IP6NDP); } /* * Reset ND level link MTU. This function is called when the physical MTU * changes, which means we might have to adjust the ND level MTU. */ void nd6_setmtu(struct ifnet *ifp) { nd6_setmtu0(ifp, ND_IFINFO(ifp)); } /* XXX todo: do not maintain copy of ifp->if_mtu in ndi->maxmtu */ void nd6_setmtu0(struct ifnet *ifp, struct nd_ifinfo *ndi) { u_int32_t omaxmtu; omaxmtu = ndi->maxmtu; switch (ifp->if_type) { case IFT_ARCNET: ndi->maxmtu = MIN(ARC_PHDS_MAXMTU, ifp->if_mtu); /* RFC2497 */ break; case IFT_FDDI: ndi->maxmtu = MIN(FDDIIPMTU, ifp->if_mtu); /* RFC2467 */ break; case IFT_ISO88025: ndi->maxmtu = MIN(ISO88025_MAX_MTU, ifp->if_mtu); break; default: ndi->maxmtu = ifp->if_mtu; break; } /* * Decreasing the interface MTU under IPV6 minimum MTU may cause * undesirable situation. We thus notify the operator of the change * explicitly. The check for omaxmtu is necessary to restrict the * log to the case of changing the MTU, not initializing it. */ if (omaxmtu >= IPV6_MMTU && ndi->maxmtu < IPV6_MMTU) { log(LOG_NOTICE, "nd6_setmtu0: " "new link MTU on %s (%lu) is too small for IPv6\n", if_name(ifp), (unsigned long)ndi->maxmtu); } if (ndi->maxmtu > V_in6_maxmtu) in6_setmaxmtu(); /* check all interfaces just in case */ } void nd6_option_init(void *opt, int icmp6len, union nd_opts *ndopts) { bzero(ndopts, sizeof(*ndopts)); ndopts->nd_opts_search = (struct nd_opt_hdr *)opt; ndopts->nd_opts_last = (struct nd_opt_hdr *)(((u_char *)opt) + icmp6len); if (icmp6len == 0) { ndopts->nd_opts_done = 1; ndopts->nd_opts_search = NULL; } } /* * Take one ND option. */ struct nd_opt_hdr * nd6_option(union nd_opts *ndopts) { struct nd_opt_hdr *nd_opt; int olen; KASSERT(ndopts != NULL, ("%s: ndopts == NULL", __func__)); KASSERT(ndopts->nd_opts_last != NULL, ("%s: uninitialized ndopts", __func__)); if (ndopts->nd_opts_search == NULL) return NULL; if (ndopts->nd_opts_done) return NULL; nd_opt = ndopts->nd_opts_search; /* make sure nd_opt_len is inside the buffer */ if ((caddr_t)&nd_opt->nd_opt_len >= (caddr_t)ndopts->nd_opts_last) { bzero(ndopts, sizeof(*ndopts)); return NULL; } olen = nd_opt->nd_opt_len << 3; if (olen == 0) { /* * Message validation requires that all included * options have a length that is greater than zero. */ bzero(ndopts, sizeof(*ndopts)); return NULL; } ndopts->nd_opts_search = (struct nd_opt_hdr *)((caddr_t)nd_opt + olen); if (ndopts->nd_opts_search > ndopts->nd_opts_last) { /* option overruns the end of buffer, invalid */ bzero(ndopts, sizeof(*ndopts)); return NULL; } else if (ndopts->nd_opts_search == ndopts->nd_opts_last) { /* reached the end of options chain */ ndopts->nd_opts_done = 1; ndopts->nd_opts_search = NULL; } return nd_opt; } /* * Parse multiple ND options. * This function is much easier to use, for ND routines that do not need * multiple options of the same type. */ int nd6_options(union nd_opts *ndopts) { struct nd_opt_hdr *nd_opt; int i = 0; KASSERT(ndopts != NULL, ("%s: ndopts == NULL", __func__)); KASSERT(ndopts->nd_opts_last != NULL, ("%s: uninitialized ndopts", __func__)); if (ndopts->nd_opts_search == NULL) return 0; while (1) { nd_opt = nd6_option(ndopts); if (nd_opt == NULL && ndopts->nd_opts_last == NULL) { /* * Message validation requires that all included * options have a length that is greater than zero. */ ICMP6STAT_INC(icp6s_nd_badopt); bzero(ndopts, sizeof(*ndopts)); return -1; } if (nd_opt == NULL) goto skip1; switch (nd_opt->nd_opt_type) { case ND_OPT_SOURCE_LINKADDR: case ND_OPT_TARGET_LINKADDR: case ND_OPT_MTU: case ND_OPT_REDIRECTED_HEADER: if (ndopts->nd_opt_array[nd_opt->nd_opt_type]) { nd6log((LOG_INFO, "duplicated ND6 option found (type=%d)\n", nd_opt->nd_opt_type)); /* XXX bark? */ } else { ndopts->nd_opt_array[nd_opt->nd_opt_type] = nd_opt; } break; case ND_OPT_PREFIX_INFORMATION: if (ndopts->nd_opt_array[nd_opt->nd_opt_type] == 0) { ndopts->nd_opt_array[nd_opt->nd_opt_type] = nd_opt; } ndopts->nd_opts_pi_end = (struct nd_opt_prefix_info *)nd_opt; break; /* What about ND_OPT_ROUTE_INFO? RFC 4191 */ case ND_OPT_RDNSS: /* RFC 6106 */ case ND_OPT_DNSSL: /* RFC 6106 */ /* * Silently ignore options we know and do not care about * in the kernel. */ break; default: /* * Unknown options must be silently ignored, * to accomodate future extension to the protocol. */ nd6log((LOG_DEBUG, "nd6_options: unsupported option %d - " "option ignored\n", nd_opt->nd_opt_type)); } skip1: i++; if (i > V_nd6_maxndopt) { ICMP6STAT_INC(icp6s_nd_toomanyopt); nd6log((LOG_INFO, "too many loop in nd opt\n")); break; } if (ndopts->nd_opts_done) break; } return 0; } /* * ND6 timer routine to handle ND6 entries */ void nd6_llinfo_settimer_locked(struct llentry *ln, long tick) { int canceled; LLE_WLOCK_ASSERT(ln); if (tick < 0) { ln->la_expire = 0; ln->ln_ntick = 0; canceled = callout_stop(&ln->ln_timer_ch); } else { - ln->la_expire = time_second + tick / hz; + ln->la_expire = time_uptime + tick / hz; LLE_ADDREF(ln); if (tick > INT_MAX) { ln->ln_ntick = tick - INT_MAX; canceled = callout_reset(&ln->ln_timer_ch, INT_MAX, nd6_llinfo_timer, ln); } else { ln->ln_ntick = 0; canceled = callout_reset(&ln->ln_timer_ch, tick, nd6_llinfo_timer, ln); } } if (canceled) LLE_REMREF(ln); } void nd6_llinfo_settimer(struct llentry *ln, long tick) { LLE_WLOCK(ln); nd6_llinfo_settimer_locked(ln, tick); LLE_WUNLOCK(ln); } static void nd6_llinfo_timer(void *arg) { struct llentry *ln; struct in6_addr *dst; struct ifnet *ifp; struct nd_ifinfo *ndi = NULL; KASSERT(arg != NULL, ("%s: arg NULL", __func__)); ln = (struct llentry *)arg; LLE_WLOCK_ASSERT(ln); ifp = ln->lle_tbl->llt_ifp; CURVNET_SET(ifp->if_vnet); if (ln->ln_ntick > 0) { if (ln->ln_ntick > INT_MAX) { ln->ln_ntick -= INT_MAX; nd6_llinfo_settimer_locked(ln, INT_MAX); } else { ln->ln_ntick = 0; nd6_llinfo_settimer_locked(ln, ln->ln_ntick); } goto done; } ndi = ND_IFINFO(ifp); dst = &L3_ADDR_SIN6(ln)->sin6_addr; if (ln->la_flags & LLE_STATIC) { goto done; } if (ln->la_flags & LLE_DELETED) { (void)nd6_free(ln, 0); ln = NULL; goto done; } switch (ln->ln_state) { case ND6_LLINFO_INCOMPLETE: if (ln->la_asked < V_nd6_mmaxtries) { ln->la_asked++; nd6_llinfo_settimer_locked(ln, (long)ndi->retrans * hz / 1000); LLE_WUNLOCK(ln); nd6_ns_output(ifp, NULL, dst, ln, 0); LLE_WLOCK(ln); } else { struct mbuf *m = ln->la_hold; if (m) { struct mbuf *m0; /* * assuming every packet in la_hold has the * same IP header. Send error after unlock. */ m0 = m->m_nextpkt; m->m_nextpkt = NULL; ln->la_hold = m0; clear_llinfo_pqueue(ln); } EVENTHANDLER_INVOKE(lle_event, ln, LLENTRY_TIMEDOUT); (void)nd6_free(ln, 0); ln = NULL; if (m != NULL) icmp6_error2(m, ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_ADDR, 0, ifp); } break; case ND6_LLINFO_REACHABLE: if (!ND6_LLINFO_PERMANENT(ln)) { ln->ln_state = ND6_LLINFO_STALE; nd6_llinfo_settimer_locked(ln, (long)V_nd6_gctimer * hz); } break; case ND6_LLINFO_STALE: /* Garbage Collection(RFC 2461 5.3) */ if (!ND6_LLINFO_PERMANENT(ln)) { EVENTHANDLER_INVOKE(lle_event, ln, LLENTRY_EXPIRED); (void)nd6_free(ln, 1); ln = NULL; } break; case ND6_LLINFO_DELAY: if (ndi && (ndi->flags & ND6_IFF_PERFORMNUD) != 0) { /* We need NUD */ ln->la_asked = 1; ln->ln_state = ND6_LLINFO_PROBE; nd6_llinfo_settimer_locked(ln, (long)ndi->retrans * hz / 1000); LLE_WUNLOCK(ln); nd6_ns_output(ifp, dst, dst, ln, 0); LLE_WLOCK(ln); } else { ln->ln_state = ND6_LLINFO_STALE; /* XXX */ nd6_llinfo_settimer_locked(ln, (long)V_nd6_gctimer * hz); } break; case ND6_LLINFO_PROBE: if (ln->la_asked < V_nd6_umaxtries) { ln->la_asked++; nd6_llinfo_settimer_locked(ln, (long)ndi->retrans * hz / 1000); LLE_WUNLOCK(ln); nd6_ns_output(ifp, dst, dst, ln, 0); LLE_WLOCK(ln); } else { EVENTHANDLER_INVOKE(lle_event, ln, LLENTRY_EXPIRED); (void)nd6_free(ln, 0); ln = NULL; } break; default: panic("%s: paths in a dark night can be confusing: %d", __func__, ln->ln_state); } done: if (ln != NULL) LLE_FREE_LOCKED(ln); CURVNET_RESTORE(); } /* * ND6 timer routine to expire default route list and prefix list */ void nd6_timer(void *arg) { CURVNET_SET((struct vnet *) arg); struct nd_defrouter *dr, *ndr; struct nd_prefix *pr, *npr; struct in6_ifaddr *ia6, *nia6; callout_reset(&V_nd6_timer_ch, V_nd6_prune * hz, nd6_timer, curvnet); /* expire default router list */ TAILQ_FOREACH_SAFE(dr, &V_nd_defrouter, dr_entry, ndr) { - if (dr->expire && dr->expire < time_second) + if (dr->expire && dr->expire < time_uptime) defrtrlist_del(dr); } /* * expire interface addresses. * in the past the loop was inside prefix expiry processing. * However, from a stricter speci-confrmance standpoint, we should * rather separate address lifetimes and prefix lifetimes. * * XXXRW: in6_ifaddrhead locking. */ addrloop: TAILQ_FOREACH_SAFE(ia6, &V_in6_ifaddrhead, ia_link, nia6) { /* check address lifetime */ if (IFA6_IS_INVALID(ia6)) { int regen = 0; /* * If the expiring address is temporary, try * regenerating a new one. This would be useful when * we suspended a laptop PC, then turned it on after a * period that could invalidate all temporary * addresses. Although we may have to restart the * loop (see below), it must be after purging the * address. Otherwise, we'd see an infinite loop of * regeneration. */ if (V_ip6_use_tempaddr && (ia6->ia6_flags & IN6_IFF_TEMPORARY) != 0) { if (regen_tmpaddr(ia6) == 0) regen = 1; } in6_purgeaddr(&ia6->ia_ifa); if (regen) goto addrloop; /* XXX: see below */ } else if (IFA6_IS_DEPRECATED(ia6)) { int oldflags = ia6->ia6_flags; ia6->ia6_flags |= IN6_IFF_DEPRECATED; /* * If a temporary address has just become deprecated, * regenerate a new one if possible. */ if (V_ip6_use_tempaddr && (ia6->ia6_flags & IN6_IFF_TEMPORARY) != 0 && (oldflags & IN6_IFF_DEPRECATED) == 0) { if (regen_tmpaddr(ia6) == 0) { /* * A new temporary address is * generated. * XXX: this means the address chain * has changed while we are still in * the loop. Although the change * would not cause disaster (because * it's not a deletion, but an * addition,) we'd rather restart the * loop just for safety. Or does this * significantly reduce performance?? */ goto addrloop; } } } else { /* * A new RA might have made a deprecated address * preferred. */ ia6->ia6_flags &= ~IN6_IFF_DEPRECATED; } } /* expire prefix list */ LIST_FOREACH_SAFE(pr, &V_nd_prefix, ndpr_entry, npr) { /* * check prefix lifetime. * since pltime is just for autoconf, pltime processing for * prefix is not necessary. */ if (pr->ndpr_vltime != ND6_INFINITE_LIFETIME && - time_second - pr->ndpr_lastupdate > pr->ndpr_vltime) { + time_uptime - pr->ndpr_lastupdate > pr->ndpr_vltime) { /* * address expiration and prefix expiration are * separate. NEVER perform in6_purgeaddr here. */ prelist_remove(pr); } } CURVNET_RESTORE(); } /* * ia6 - deprecated/invalidated temporary address */ static int regen_tmpaddr(struct in6_ifaddr *ia6) { struct ifaddr *ifa; struct ifnet *ifp; struct in6_ifaddr *public_ifa6 = NULL; ifp = ia6->ia_ifa.ifa_ifp; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { struct in6_ifaddr *it6; if (ifa->ifa_addr->sa_family != AF_INET6) continue; it6 = (struct in6_ifaddr *)ifa; /* ignore no autoconf addresses. */ if ((it6->ia6_flags & IN6_IFF_AUTOCONF) == 0) continue; /* ignore autoconf addresses with different prefixes. */ if (it6->ia6_ndpr == NULL || it6->ia6_ndpr != ia6->ia6_ndpr) continue; /* * Now we are looking at an autoconf address with the same * prefix as ours. If the address is temporary and is still * preferred, do not create another one. It would be rare, but * could happen, for example, when we resume a laptop PC after * a long period. */ if ((it6->ia6_flags & IN6_IFF_TEMPORARY) != 0 && !IFA6_IS_DEPRECATED(it6)) { public_ifa6 = NULL; break; } /* * This is a public autoconf address that has the same prefix * as ours. If it is preferred, keep it. We can't break the * loop here, because there may be a still-preferred temporary * address with the prefix. */ if (!IFA6_IS_DEPRECATED(it6)) public_ifa6 = it6; if (public_ifa6 != NULL) ifa_ref(&public_ifa6->ia_ifa); } IF_ADDR_RUNLOCK(ifp); if (public_ifa6 != NULL) { int e; if ((e = in6_tmpifadd(public_ifa6, 0, 0)) != 0) { ifa_free(&public_ifa6->ia_ifa); log(LOG_NOTICE, "regen_tmpaddr: failed to create a new" " tmp addr,errno=%d\n", e); return (-1); } ifa_free(&public_ifa6->ia_ifa); return (0); } return (-1); } /* * Nuke neighbor cache/prefix/default router management table, right before * ifp goes away. */ void nd6_purge(struct ifnet *ifp) { struct nd_defrouter *dr, *ndr; struct nd_prefix *pr, *npr; /* * Nuke default router list entries toward ifp. * We defer removal of default router list entries that is installed * in the routing table, in order to keep additional side effects as * small as possible. */ TAILQ_FOREACH_SAFE(dr, &V_nd_defrouter, dr_entry, ndr) { if (dr->installed) continue; if (dr->ifp == ifp) defrtrlist_del(dr); } TAILQ_FOREACH_SAFE(dr, &V_nd_defrouter, dr_entry, ndr) { if (!dr->installed) continue; if (dr->ifp == ifp) defrtrlist_del(dr); } /* Nuke prefix list entries toward ifp */ LIST_FOREACH_SAFE(pr, &V_nd_prefix, ndpr_entry, npr) { if (pr->ndpr_ifp == ifp) { /* * Because if_detach() does *not* release prefixes * while purging addresses the reference count will * still be above zero. We therefore reset it to * make sure that the prefix really gets purged. */ pr->ndpr_refcnt = 0; /* * Previously, pr->ndpr_addr is removed as well, * but I strongly believe we don't have to do it. * nd6_purge() is only called from in6_ifdetach(), * which removes all the associated interface addresses * by itself. * (jinmei@kame.net 20010129) */ prelist_remove(pr); } } /* cancel default outgoing interface setting */ if (V_nd6_defifindex == ifp->if_index) nd6_setdefaultiface(0); if (ND_IFINFO(ifp)->flags & ND6_IFF_ACCEPT_RTADV) { /* Refresh default router list. */ defrouter_select(); } /* XXXXX * We do not nuke the neighbor cache entries here any more * because the neighbor cache is kept in if_afdata[AF_INET6]. * nd6_purge() is invoked by in6_ifdetach() which is called * from if_detach() where everything gets purged. So let * in6_domifdetach() do the actual L2 table purging work. */ } /* * the caller acquires and releases the lock on the lltbls * Returns the llentry locked */ struct llentry * nd6_lookup(struct in6_addr *addr6, int flags, struct ifnet *ifp) { struct sockaddr_in6 sin6; struct llentry *ln; int llflags; bzero(&sin6, sizeof(sin6)); sin6.sin6_len = sizeof(struct sockaddr_in6); sin6.sin6_family = AF_INET6; sin6.sin6_addr = *addr6; IF_AFDATA_LOCK_ASSERT(ifp); llflags = 0; if (flags & ND6_CREATE) llflags |= LLE_CREATE; if (flags & ND6_EXCLUSIVE) llflags |= LLE_EXCLUSIVE; ln = lla_lookup(LLTABLE6(ifp), llflags, (struct sockaddr *)&sin6); if ((ln != NULL) && (llflags & LLE_CREATE)) ln->ln_state = ND6_LLINFO_NOSTATE; return (ln); } /* * Test whether a given IPv6 address is a neighbor or not, ignoring * the actual neighbor cache. The neighbor cache is ignored in order * to not reenter the routing code from within itself. */ static int nd6_is_new_addr_neighbor(struct sockaddr_in6 *addr, struct ifnet *ifp) { struct nd_prefix *pr; struct ifaddr *dstaddr; /* * A link-local address is always a neighbor. * XXX: a link does not necessarily specify a single interface. */ if (IN6_IS_ADDR_LINKLOCAL(&addr->sin6_addr)) { struct sockaddr_in6 sin6_copy; u_int32_t zone; /* * We need sin6_copy since sa6_recoverscope() may modify the * content (XXX). */ sin6_copy = *addr; if (sa6_recoverscope(&sin6_copy)) return (0); /* XXX: should be impossible */ if (in6_setscope(&sin6_copy.sin6_addr, ifp, &zone)) return (0); if (sin6_copy.sin6_scope_id == zone) return (1); else return (0); } /* * If the address matches one of our addresses, * it should be a neighbor. * If the address matches one of our on-link prefixes, it should be a * neighbor. */ LIST_FOREACH(pr, &V_nd_prefix, ndpr_entry) { if (pr->ndpr_ifp != ifp) continue; if (!(pr->ndpr_stateflags & NDPRF_ONLINK)) { struct rtentry *rt; /* Always use the default FIB here. */ rt = in6_rtalloc1((struct sockaddr *)&pr->ndpr_prefix, 0, 0, RT_DEFAULT_FIB); if (rt == NULL) continue; /* * This is the case where multiple interfaces * have the same prefix, but only one is installed * into the routing table and that prefix entry * is not the one being examined here. In the case * where RADIX_MPATH is enabled, multiple route * entries (of the same rt_key value) will be * installed because the interface addresses all * differ. */ if (!IN6_ARE_ADDR_EQUAL(&pr->ndpr_prefix.sin6_addr, &((struct sockaddr_in6 *)rt_key(rt))->sin6_addr)) { RTFREE_LOCKED(rt); continue; } RTFREE_LOCKED(rt); } if (IN6_ARE_MASKED_ADDR_EQUAL(&pr->ndpr_prefix.sin6_addr, &addr->sin6_addr, &pr->ndpr_mask)) return (1); } /* * If the address is assigned on the node of the other side of * a p2p interface, the address should be a neighbor. */ dstaddr = ifa_ifwithdstaddr((struct sockaddr *)addr); if (dstaddr != NULL) { if (dstaddr->ifa_ifp == ifp) { ifa_free(dstaddr); return (1); } ifa_free(dstaddr); } /* * If the default router list is empty, all addresses are regarded * as on-link, and thus, as a neighbor. */ if (ND_IFINFO(ifp)->flags & ND6_IFF_ACCEPT_RTADV && TAILQ_EMPTY(&V_nd_defrouter) && V_nd6_defifindex == ifp->if_index) { return (1); } return (0); } /* * Detect if a given IPv6 address identifies a neighbor on a given link. * XXX: should take care of the destination of a p2p link? */ int nd6_is_addr_neighbor(struct sockaddr_in6 *addr, struct ifnet *ifp) { struct llentry *lle; int rc = 0; IF_AFDATA_UNLOCK_ASSERT(ifp); if (nd6_is_new_addr_neighbor(addr, ifp)) return (1); /* * Even if the address matches none of our addresses, it might be * in the neighbor cache. */ IF_AFDATA_RLOCK(ifp); if ((lle = nd6_lookup(&addr->sin6_addr, 0, ifp)) != NULL) { LLE_RUNLOCK(lle); rc = 1; } IF_AFDATA_RUNLOCK(ifp); return (rc); } /* * Free an nd6 llinfo entry. * Since the function would cause significant changes in the kernel, DO NOT * make it global, unless you have a strong reason for the change, and are sure * that the change is safe. */ static struct llentry * nd6_free(struct llentry *ln, int gc) { struct llentry *next; struct nd_defrouter *dr; struct ifnet *ifp; LLE_WLOCK_ASSERT(ln); /* * we used to have pfctlinput(PRC_HOSTDEAD) here. * even though it is not harmful, it was not really necessary. */ /* cancel timer */ nd6_llinfo_settimer_locked(ln, -1); ifp = ln->lle_tbl->llt_ifp; if (ND_IFINFO(ifp)->flags & ND6_IFF_ACCEPT_RTADV) { dr = defrouter_lookup(&L3_ADDR_SIN6(ln)->sin6_addr, ifp); if (dr != NULL && dr->expire && ln->ln_state == ND6_LLINFO_STALE && gc) { /* * If the reason for the deletion is just garbage * collection, and the neighbor is an active default * router, do not delete it. Instead, reset the GC * timer using the router's lifetime. * Simply deleting the entry would affect default * router selection, which is not necessarily a good * thing, especially when we're using router preference * values. * XXX: the check for ln_state would be redundant, * but we intentionally keep it just in case. */ - if (dr->expire > time_second) + if (dr->expire > time_uptime) nd6_llinfo_settimer_locked(ln, - (dr->expire - time_second) * hz); + (dr->expire - time_uptime) * hz); else nd6_llinfo_settimer_locked(ln, (long)V_nd6_gctimer * hz); next = LIST_NEXT(ln, lle_next); LLE_REMREF(ln); LLE_WUNLOCK(ln); return (next); } if (dr) { /* * Unreachablity of a router might affect the default * router selection and on-link detection of advertised * prefixes. */ /* * Temporarily fake the state to choose a new default * router and to perform on-link determination of * prefixes correctly. * Below the state will be set correctly, * or the entry itself will be deleted. */ ln->ln_state = ND6_LLINFO_INCOMPLETE; } if (ln->ln_router || dr) { /* * We need to unlock to avoid a LOR with rt6_flush() with the * rnh and for the calls to pfxlist_onlink_check() and * defrouter_select() in the block further down for calls * into nd6_lookup(). We still hold a ref. */ LLE_WUNLOCK(ln); /* * rt6_flush must be called whether or not the neighbor * is in the Default Router List. * See a corresponding comment in nd6_na_input(). */ rt6_flush(&L3_ADDR_SIN6(ln)->sin6_addr, ifp); } if (dr) { /* * Since defrouter_select() does not affect the * on-link determination and MIP6 needs the check * before the default router selection, we perform * the check now. */ pfxlist_onlink_check(); /* * Refresh default router list. */ defrouter_select(); } if (ln->ln_router || dr) LLE_WLOCK(ln); } /* * Before deleting the entry, remember the next entry as the * return value. We need this because pfxlist_onlink_check() above * might have freed other entries (particularly the old next entry) as * a side effect (XXX). */ next = LIST_NEXT(ln, lle_next); /* * Save to unlock. We still hold an extra reference and will not * free(9) in llentry_free() if someone else holds one as well. */ LLE_WUNLOCK(ln); IF_AFDATA_LOCK(ifp); LLE_WLOCK(ln); /* Guard against race with other llentry_free(). */ if (ln->la_flags & LLE_LINKED) { LLE_REMREF(ln); llentry_free(ln); } else LLE_FREE_LOCKED(ln); IF_AFDATA_UNLOCK(ifp); return (next); } /* * Upper-layer reachability hint for Neighbor Unreachability Detection. * * XXX cost-effective methods? */ void nd6_nud_hint(struct rtentry *rt, struct in6_addr *dst6, int force) { struct llentry *ln; struct ifnet *ifp; if ((dst6 == NULL) || (rt == NULL)) return; ifp = rt->rt_ifp; IF_AFDATA_LOCK(ifp); ln = nd6_lookup(dst6, ND6_EXCLUSIVE, NULL); IF_AFDATA_UNLOCK(ifp); if (ln == NULL) return; if (ln->ln_state < ND6_LLINFO_REACHABLE) goto done; /* * if we get upper-layer reachability confirmation many times, * it is possible we have false information. */ if (!force) { ln->ln_byhint++; if (ln->ln_byhint > V_nd6_maxnudhint) { goto done; } } ln->ln_state = ND6_LLINFO_REACHABLE; if (!ND6_LLINFO_PERMANENT(ln)) { nd6_llinfo_settimer_locked(ln, (long)ND_IFINFO(rt->rt_ifp)->reachable * hz); } done: LLE_WUNLOCK(ln); } /* * Rejuvenate this function for routing operations related * processing. */ void nd6_rtrequest(int req, struct rtentry *rt, struct rt_addrinfo *info) { struct sockaddr_in6 *gateway; struct nd_defrouter *dr; struct ifnet *ifp; RT_LOCK_ASSERT(rt); gateway = (struct sockaddr_in6 *)rt->rt_gateway; ifp = rt->rt_ifp; switch (req) { case RTM_ADD: break; case RTM_DELETE: if (!ifp) return; /* * Only indirect routes are interesting. */ if ((rt->rt_flags & RTF_GATEWAY) == 0) return; /* * check for default route */ if (IN6_ARE_ADDR_EQUAL(&in6addr_any, &SIN6(rt_key(rt))->sin6_addr)) { dr = defrouter_lookup(&gateway->sin6_addr, ifp); if (dr != NULL) dr->installed = 0; } break; } } int nd6_ioctl(u_long cmd, caddr_t data, struct ifnet *ifp) { struct in6_drlist *drl = (struct in6_drlist *)data; struct in6_oprlist *oprl = (struct in6_oprlist *)data; struct in6_ndireq *ndi = (struct in6_ndireq *)data; struct in6_nbrinfo *nbi = (struct in6_nbrinfo *)data; struct in6_ndifreq *ndif = (struct in6_ndifreq *)data; struct nd_defrouter *dr; struct nd_prefix *pr; int i = 0, error = 0; if (ifp->if_afdata[AF_INET6] == NULL) return (EPFNOSUPPORT); switch (cmd) { case SIOCGDRLST_IN6: /* * obsolete API, use sysctl under net.inet6.icmp6 */ bzero(drl, sizeof(*drl)); TAILQ_FOREACH(dr, &V_nd_defrouter, dr_entry) { if (i >= DRLSTSIZ) break; drl->defrouter[i].rtaddr = dr->rtaddr; in6_clearscope(&drl->defrouter[i].rtaddr); drl->defrouter[i].flags = dr->flags; drl->defrouter[i].rtlifetime = dr->rtlifetime; drl->defrouter[i].expire = dr->expire; drl->defrouter[i].if_index = dr->ifp->if_index; i++; } break; case SIOCGPRLST_IN6: /* * obsolete API, use sysctl under net.inet6.icmp6 * * XXX the structure in6_prlist was changed in backward- * incompatible manner. in6_oprlist is used for SIOCGPRLST_IN6, * in6_prlist is used for nd6_sysctl() - fill_prlist(). */ /* * XXX meaning of fields, especialy "raflags", is very * differnet between RA prefix list and RR/static prefix list. * how about separating ioctls into two? */ bzero(oprl, sizeof(*oprl)); LIST_FOREACH(pr, &V_nd_prefix, ndpr_entry) { struct nd_pfxrouter *pfr; int j; if (i >= PRLSTSIZ) break; oprl->prefix[i].prefix = pr->ndpr_prefix.sin6_addr; oprl->prefix[i].raflags = pr->ndpr_raf; oprl->prefix[i].prefixlen = pr->ndpr_plen; oprl->prefix[i].vltime = pr->ndpr_vltime; oprl->prefix[i].pltime = pr->ndpr_pltime; oprl->prefix[i].if_index = pr->ndpr_ifp->if_index; if (pr->ndpr_vltime == ND6_INFINITE_LIFETIME) oprl->prefix[i].expire = 0; else { time_t maxexpire; /* XXX: we assume time_t is signed. */ maxexpire = (-1) & ~((time_t)1 << ((sizeof(maxexpire) * 8) - 1)); if (pr->ndpr_vltime < maxexpire - pr->ndpr_lastupdate) { oprl->prefix[i].expire = pr->ndpr_lastupdate + pr->ndpr_vltime; } else oprl->prefix[i].expire = maxexpire; } j = 0; LIST_FOREACH(pfr, &pr->ndpr_advrtrs, pfr_entry) { if (j < DRLSTSIZ) { #define RTRADDR oprl->prefix[i].advrtr[j] RTRADDR = pfr->router->rtaddr; in6_clearscope(&RTRADDR); #undef RTRADDR } j++; } oprl->prefix[i].advrtrs = j; oprl->prefix[i].origin = PR_ORIG_RA; i++; } break; case OSIOCGIFINFO_IN6: #define ND ndi->ndi /* XXX: old ndp(8) assumes a positive value for linkmtu. */ bzero(&ND, sizeof(ND)); ND.linkmtu = IN6_LINKMTU(ifp); ND.maxmtu = ND_IFINFO(ifp)->maxmtu; ND.basereachable = ND_IFINFO(ifp)->basereachable; ND.reachable = ND_IFINFO(ifp)->reachable; ND.retrans = ND_IFINFO(ifp)->retrans; ND.flags = ND_IFINFO(ifp)->flags; ND.recalctm = ND_IFINFO(ifp)->recalctm; ND.chlim = ND_IFINFO(ifp)->chlim; break; case SIOCGIFINFO_IN6: ND = *ND_IFINFO(ifp); break; case SIOCSIFINFO_IN6: /* * used to change host variables from userland. * intented for a use on router to reflect RA configurations. */ /* 0 means 'unspecified' */ if (ND.linkmtu != 0) { if (ND.linkmtu < IPV6_MMTU || ND.linkmtu > IN6_LINKMTU(ifp)) { error = EINVAL; break; } ND_IFINFO(ifp)->linkmtu = ND.linkmtu; } if (ND.basereachable != 0) { int obasereachable = ND_IFINFO(ifp)->basereachable; ND_IFINFO(ifp)->basereachable = ND.basereachable; if (ND.basereachable != obasereachable) ND_IFINFO(ifp)->reachable = ND_COMPUTE_RTIME(ND.basereachable); } if (ND.retrans != 0) ND_IFINFO(ifp)->retrans = ND.retrans; if (ND.chlim != 0) ND_IFINFO(ifp)->chlim = ND.chlim; /* FALLTHROUGH */ case SIOCSIFINFO_FLAGS: { struct ifaddr *ifa; struct in6_ifaddr *ia; if ((ND_IFINFO(ifp)->flags & ND6_IFF_IFDISABLED) && !(ND.flags & ND6_IFF_IFDISABLED)) { /* ifdisabled 1->0 transision */ /* * If the interface is marked as ND6_IFF_IFDISABLED and * has an link-local address with IN6_IFF_DUPLICATED, * do not clear ND6_IFF_IFDISABLED. * See RFC 4862, Section 5.4.5. */ int duplicated_linklocal = 0; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; ia = (struct in6_ifaddr *)ifa; if ((ia->ia6_flags & IN6_IFF_DUPLICATED) && IN6_IS_ADDR_LINKLOCAL(IA6_IN6(ia))) { duplicated_linklocal = 1; break; } } IF_ADDR_RUNLOCK(ifp); if (duplicated_linklocal) { ND.flags |= ND6_IFF_IFDISABLED; log(LOG_ERR, "Cannot enable an interface" " with a link-local address marked" " duplicate.\n"); } else { ND_IFINFO(ifp)->flags &= ~ND6_IFF_IFDISABLED; if (ifp->if_flags & IFF_UP) in6_if_up(ifp); } } else if (!(ND_IFINFO(ifp)->flags & ND6_IFF_IFDISABLED) && (ND.flags & ND6_IFF_IFDISABLED)) { /* ifdisabled 0->1 transision */ /* Mark all IPv6 address as tentative. */ ND_IFINFO(ifp)->flags |= ND6_IFF_IFDISABLED; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; ia = (struct in6_ifaddr *)ifa; ia->ia6_flags |= IN6_IFF_TENTATIVE; } IF_ADDR_RUNLOCK(ifp); } if (ND.flags & ND6_IFF_AUTO_LINKLOCAL) { if (!(ND_IFINFO(ifp)->flags & ND6_IFF_AUTO_LINKLOCAL)) { /* auto_linklocal 0->1 transision */ /* If no link-local address on ifp, configure */ ND_IFINFO(ifp)->flags |= ND6_IFF_AUTO_LINKLOCAL; in6_ifattach(ifp, NULL); } else if (!(ND.flags & ND6_IFF_IFDISABLED) && ifp->if_flags & IFF_UP) { /* * When the IF already has * ND6_IFF_AUTO_LINKLOCAL, no link-local * address is assigned, and IFF_UP, try to * assign one. */ int haslinklocal = 0; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; ia = (struct in6_ifaddr *)ifa; if (IN6_IS_ADDR_LINKLOCAL(IA6_IN6(ia))) { haslinklocal = 1; break; } } IF_ADDR_RUNLOCK(ifp); if (!haslinklocal) in6_ifattach(ifp, NULL); } } } ND_IFINFO(ifp)->flags = ND.flags; break; #undef ND case SIOCSNDFLUSH_IN6: /* XXX: the ioctl name is confusing... */ /* sync kernel routing table with the default router list */ defrouter_reset(); defrouter_select(); break; case SIOCSPFXFLUSH_IN6: { /* flush all the prefix advertised by routers */ struct nd_prefix *pr, *next; LIST_FOREACH_SAFE(pr, &V_nd_prefix, ndpr_entry, next) { struct in6_ifaddr *ia, *ia_next; if (IN6_IS_ADDR_LINKLOCAL(&pr->ndpr_prefix.sin6_addr)) continue; /* XXX */ /* do we really have to remove addresses as well? */ /* XXXRW: in6_ifaddrhead locking. */ TAILQ_FOREACH_SAFE(ia, &V_in6_ifaddrhead, ia_link, ia_next) { if ((ia->ia6_flags & IN6_IFF_AUTOCONF) == 0) continue; if (ia->ia6_ndpr == pr) in6_purgeaddr(&ia->ia_ifa); } prelist_remove(pr); } break; } case SIOCSRTRFLUSH_IN6: { /* flush all the default routers */ struct nd_defrouter *dr, *next; defrouter_reset(); TAILQ_FOREACH_SAFE(dr, &V_nd_defrouter, dr_entry, next) { defrtrlist_del(dr); } defrouter_select(); break; } case SIOCGNBRINFO_IN6: { struct llentry *ln; struct in6_addr nb_addr = nbi->addr; /* make local for safety */ if ((error = in6_setscope(&nb_addr, ifp, NULL)) != 0) return (error); IF_AFDATA_RLOCK(ifp); ln = nd6_lookup(&nb_addr, 0, ifp); IF_AFDATA_RUNLOCK(ifp); if (ln == NULL) { error = EINVAL; break; } nbi->state = ln->ln_state; nbi->asked = ln->la_asked; nbi->isrouter = ln->ln_router; nbi->expire = ln->la_expire; LLE_RUNLOCK(ln); break; } case SIOCGDEFIFACE_IN6: /* XXX: should be implemented as a sysctl? */ ndif->ifindex = V_nd6_defifindex; break; case SIOCSDEFIFACE_IN6: /* XXX: should be implemented as a sysctl? */ return (nd6_setdefaultiface(ndif->ifindex)); } return (error); } /* * Create neighbor cache entry and cache link-layer address, * on reception of inbound ND6 packets. (RS/RA/NS/redirect) * * type - ICMP6 type * code - type dependent information * * XXXXX * The caller of this function already acquired the ndp * cache table lock because the cache entry is returned. */ struct llentry * nd6_cache_lladdr(struct ifnet *ifp, struct in6_addr *from, char *lladdr, int lladdrlen, int type, int code) { struct llentry *ln = NULL; int is_newentry; int do_update; int olladdr; int llchange; int flags; int newstate = 0; uint16_t router = 0; struct sockaddr_in6 sin6; struct mbuf *chain = NULL; int static_route = 0; IF_AFDATA_UNLOCK_ASSERT(ifp); KASSERT(ifp != NULL, ("%s: ifp == NULL", __func__)); KASSERT(from != NULL, ("%s: from == NULL", __func__)); /* nothing must be updated for unspecified address */ if (IN6_IS_ADDR_UNSPECIFIED(from)) return NULL; /* * Validation about ifp->if_addrlen and lladdrlen must be done in * the caller. * * XXX If the link does not have link-layer adderss, what should * we do? (ifp->if_addrlen == 0) * Spec says nothing in sections for RA, RS and NA. There's small * description on it in NS section (RFC 2461 7.2.3). */ flags = lladdr ? ND6_EXCLUSIVE : 0; IF_AFDATA_LOCK(ifp); ln = nd6_lookup(from, flags, ifp); if (ln == NULL) { flags |= ND6_EXCLUSIVE; ln = nd6_lookup(from, flags | ND6_CREATE, ifp); IF_AFDATA_UNLOCK(ifp); is_newentry = 1; } else { IF_AFDATA_UNLOCK(ifp); /* do nothing if static ndp is set */ if (ln->la_flags & LLE_STATIC) { static_route = 1; goto done; } is_newentry = 0; } if (ln == NULL) return (NULL); olladdr = (ln->la_flags & LLE_VALID) ? 1 : 0; if (olladdr && lladdr) { llchange = bcmp(lladdr, &ln->ll_addr, ifp->if_addrlen); } else llchange = 0; /* * newentry olladdr lladdr llchange (*=record) * 0 n n -- (1) * 0 y n -- (2) * 0 n y -- (3) * STALE * 0 y y n (4) * * 0 y y y (5) * STALE * 1 -- n -- (6) NOSTATE(= PASSIVE) * 1 -- y -- (7) * STALE */ if (lladdr) { /* (3-5) and (7) */ /* * Record source link-layer address * XXX is it dependent to ifp->if_type? */ bcopy(lladdr, &ln->ll_addr, ifp->if_addrlen); ln->la_flags |= LLE_VALID; EVENTHANDLER_INVOKE(lle_event, ln, LLENTRY_RESOLVED); } if (!is_newentry) { if ((!olladdr && lladdr != NULL) || /* (3) */ (olladdr && lladdr != NULL && llchange)) { /* (5) */ do_update = 1; newstate = ND6_LLINFO_STALE; } else /* (1-2,4) */ do_update = 0; } else { do_update = 1; if (lladdr == NULL) /* (6) */ newstate = ND6_LLINFO_NOSTATE; else /* (7) */ newstate = ND6_LLINFO_STALE; } if (do_update) { /* * Update the state of the neighbor cache. */ ln->ln_state = newstate; if (ln->ln_state == ND6_LLINFO_STALE) { /* * XXX: since nd6_output() below will cause * state tansition to DELAY and reset the timer, * we must set the timer now, although it is actually * meaningless. */ nd6_llinfo_settimer_locked(ln, (long)V_nd6_gctimer * hz); if (ln->la_hold) { struct mbuf *m_hold, *m_hold_next; /* * reset the la_hold in advance, to explicitly * prevent a la_hold lookup in nd6_output() * (wouldn't happen, though...) */ for (m_hold = ln->la_hold, ln->la_hold = NULL; m_hold; m_hold = m_hold_next) { m_hold_next = m_hold->m_nextpkt; m_hold->m_nextpkt = NULL; /* * we assume ifp is not a p2p here, so * just set the 2nd argument as the * 1st one. */ nd6_output_lle(ifp, ifp, m_hold, L3_ADDR_SIN6(ln), NULL, ln, &chain); } /* * If we have mbufs in the chain we need to do * deferred transmit. Copy the address from the * llentry before dropping the lock down below. */ if (chain != NULL) memcpy(&sin6, L3_ADDR_SIN6(ln), sizeof(sin6)); } } else if (ln->ln_state == ND6_LLINFO_INCOMPLETE) { /* probe right away */ nd6_llinfo_settimer_locked((void *)ln, 0); } } /* * ICMP6 type dependent behavior. * * NS: clear IsRouter if new entry * RS: clear IsRouter * RA: set IsRouter if there's lladdr * redir: clear IsRouter if new entry * * RA case, (1): * The spec says that we must set IsRouter in the following cases: * - If lladdr exist, set IsRouter. This means (1-5). * - If it is old entry (!newentry), set IsRouter. This means (7). * So, based on the spec, in (1-5) and (7) cases we must set IsRouter. * A quetion arises for (1) case. (1) case has no lladdr in the * neighbor cache, this is similar to (6). * This case is rare but we figured that we MUST NOT set IsRouter. * * newentry olladdr lladdr llchange NS RS RA redir * D R * 0 n n -- (1) c ? s * 0 y n -- (2) c s s * 0 n y -- (3) c s s * 0 y y n (4) c s s * 0 y y y (5) c s s * 1 -- n -- (6) c c c s * 1 -- y -- (7) c c s c s * * (c=clear s=set) */ switch (type & 0xff) { case ND_NEIGHBOR_SOLICIT: /* * New entry must have is_router flag cleared. */ if (is_newentry) /* (6-7) */ ln->ln_router = 0; break; case ND_REDIRECT: /* * If the icmp is a redirect to a better router, always set the * is_router flag. Otherwise, if the entry is newly created, * clear the flag. [RFC 2461, sec 8.3] */ if (code == ND_REDIRECT_ROUTER) ln->ln_router = 1; else if (is_newentry) /* (6-7) */ ln->ln_router = 0; break; case ND_ROUTER_SOLICIT: /* * is_router flag must always be cleared. */ ln->ln_router = 0; break; case ND_ROUTER_ADVERT: /* * Mark an entry with lladdr as a router. */ if ((!is_newentry && (olladdr || lladdr)) || /* (2-5) */ (is_newentry && lladdr)) { /* (7) */ ln->ln_router = 1; } break; } if (ln != NULL) { static_route = (ln->la_flags & LLE_STATIC); router = ln->ln_router; if (flags & ND6_EXCLUSIVE) LLE_WUNLOCK(ln); else LLE_RUNLOCK(ln); if (static_route) ln = NULL; } if (chain) nd6_output_flush(ifp, ifp, chain, &sin6, NULL); /* * When the link-layer address of a router changes, select the * best router again. In particular, when the neighbor entry is newly * created, it might affect the selection policy. * Question: can we restrict the first condition to the "is_newentry" * case? * XXX: when we hear an RA from a new router with the link-layer * address option, defrouter_select() is called twice, since * defrtrlist_update called the function as well. However, I believe * we can compromise the overhead, since it only happens the first * time. * XXX: although defrouter_select() should not have a bad effect * for those are not autoconfigured hosts, we explicitly avoid such * cases for safety. */ if (do_update && router && ND_IFINFO(ifp)->flags & ND6_IFF_ACCEPT_RTADV) { /* * guaranteed recursion */ defrouter_select(); } return (ln); done: if (ln != NULL) { if (flags & ND6_EXCLUSIVE) LLE_WUNLOCK(ln); else LLE_RUNLOCK(ln); if (static_route) ln = NULL; } return (ln); } static void nd6_slowtimo(void *arg) { CURVNET_SET((struct vnet *) arg); struct nd_ifinfo *nd6if; struct ifnet *ifp; callout_reset(&V_nd6_slowtimo_ch, ND6_SLOWTIMER_INTERVAL * hz, nd6_slowtimo, curvnet); IFNET_RLOCK_NOSLEEP(); TAILQ_FOREACH(ifp, &V_ifnet, if_list) { if (ifp->if_afdata[AF_INET6] == NULL) continue; nd6if = ND_IFINFO(ifp); if (nd6if->basereachable && /* already initialized */ (nd6if->recalctm -= ND6_SLOWTIMER_INTERVAL) <= 0) { /* * Since reachable time rarely changes by router * advertisements, we SHOULD insure that a new random * value gets recomputed at least once every few hours. * (RFC 2461, 6.3.4) */ nd6if->recalctm = V_nd6_recalc_reachtm_interval; nd6if->reachable = ND_COMPUTE_RTIME(nd6if->basereachable); } } IFNET_RUNLOCK_NOSLEEP(); CURVNET_RESTORE(); } int nd6_output(struct ifnet *ifp, struct ifnet *origifp, struct mbuf *m0, struct sockaddr_in6 *dst, struct rtentry *rt0) { return (nd6_output_lle(ifp, origifp, m0, dst, rt0, NULL, NULL)); } /* * Note that I'm not enforcing any global serialization * lle state or asked changes here as the logic is too * complicated to avoid having to always acquire an exclusive * lock * KMM * */ #define senderr(e) { error = (e); goto bad;} int nd6_output_lle(struct ifnet *ifp, struct ifnet *origifp, struct mbuf *m0, struct sockaddr_in6 *dst, struct rtentry *rt0, struct llentry *lle, struct mbuf **chain) { struct mbuf *m = m0; struct m_tag *mtag; struct llentry *ln = lle; struct ip6_hdr *ip6; int error = 0; int flags = 0; int ip6len; #ifdef INVARIANTS if (lle != NULL) { LLE_WLOCK_ASSERT(lle); KASSERT(chain != NULL, (" lle locked but no mbuf chain pointer passed")); } #endif if (IN6_IS_ADDR_MULTICAST(&dst->sin6_addr)) goto sendpkt; if (nd6_need_cache(ifp) == 0) goto sendpkt; /* * next hop determination. This routine is derived from ether_output. */ /* * Address resolution or Neighbor Unreachability Detection * for the next hop. * At this point, the destination of the packet must be a unicast * or an anycast address(i.e. not a multicast). */ flags = ((m != NULL) || (lle != NULL)) ? LLE_EXCLUSIVE : 0; if (ln == NULL) { retry: IF_AFDATA_LOCK(ifp); ln = lla_lookup(LLTABLE6(ifp), flags, (struct sockaddr *)dst); IF_AFDATA_UNLOCK(ifp); if ((ln == NULL) && nd6_is_addr_neighbor(dst, ifp)) { /* * Since nd6_is_addr_neighbor() internally calls nd6_lookup(), * the condition below is not very efficient. But we believe * it is tolerable, because this should be a rare case. */ flags = ND6_CREATE | (m ? ND6_EXCLUSIVE : 0); IF_AFDATA_LOCK(ifp); ln = nd6_lookup(&dst->sin6_addr, flags, ifp); IF_AFDATA_UNLOCK(ifp); } } if (ln == NULL) { if ((ifp->if_flags & IFF_POINTOPOINT) == 0 && !(ND_IFINFO(ifp)->flags & ND6_IFF_PERFORMNUD)) { char ip6buf[INET6_ADDRSTRLEN]; log(LOG_DEBUG, "nd6_output: can't allocate llinfo for %s " "(ln=%p)\n", ip6_sprintf(ip6buf, &dst->sin6_addr), ln); senderr(EIO); /* XXX: good error? */ } goto sendpkt; /* send anyway */ } /* We don't have to do link-layer address resolution on a p2p link. */ if ((ifp->if_flags & IFF_POINTOPOINT) != 0 && ln->ln_state < ND6_LLINFO_REACHABLE) { if ((flags & LLE_EXCLUSIVE) == 0) { flags |= LLE_EXCLUSIVE; goto retry; } ln->ln_state = ND6_LLINFO_STALE; nd6_llinfo_settimer_locked(ln, (long)V_nd6_gctimer * hz); } /* * The first time we send a packet to a neighbor whose entry is * STALE, we have to change the state to DELAY and a sets a timer to * expire in DELAY_FIRST_PROBE_TIME seconds to ensure do * neighbor unreachability detection on expiration. * (RFC 2461 7.3.3) */ if (ln->ln_state == ND6_LLINFO_STALE) { if ((flags & LLE_EXCLUSIVE) == 0) { flags |= LLE_EXCLUSIVE; LLE_RUNLOCK(ln); goto retry; } ln->la_asked = 0; ln->ln_state = ND6_LLINFO_DELAY; nd6_llinfo_settimer_locked(ln, (long)V_nd6_delay * hz); } /* * If the neighbor cache entry has a state other than INCOMPLETE * (i.e. its link-layer address is already resolved), just * send the packet. */ if (ln->ln_state > ND6_LLINFO_INCOMPLETE) goto sendpkt; /* * There is a neighbor cache entry, but no ethernet address * response yet. Append this latest packet to the end of the * packet queue in the mbuf, unless the number of the packet * does not exceed nd6_maxqueuelen. When it exceeds nd6_maxqueuelen, * the oldest packet in the queue will be removed. */ if (ln->ln_state == ND6_LLINFO_NOSTATE) ln->ln_state = ND6_LLINFO_INCOMPLETE; if ((flags & LLE_EXCLUSIVE) == 0) { flags |= LLE_EXCLUSIVE; LLE_RUNLOCK(ln); goto retry; } LLE_WLOCK_ASSERT(ln); if (ln->la_hold) { struct mbuf *m_hold; int i; i = 0; for (m_hold = ln->la_hold; m_hold; m_hold = m_hold->m_nextpkt) { i++; if (m_hold->m_nextpkt == NULL) { m_hold->m_nextpkt = m; break; } } while (i >= V_nd6_maxqueuelen) { m_hold = ln->la_hold; ln->la_hold = ln->la_hold->m_nextpkt; m_freem(m_hold); i--; } } else { ln->la_hold = m; } /* * If there has been no NS for the neighbor after entering the * INCOMPLETE state, send the first solicitation. */ if (!ND6_LLINFO_PERMANENT(ln) && ln->la_asked == 0) { ln->la_asked++; nd6_llinfo_settimer_locked(ln, (long)ND_IFINFO(ifp)->retrans * hz / 1000); LLE_WUNLOCK(ln); nd6_ns_output(ifp, NULL, &dst->sin6_addr, ln, 0); if (lle != NULL && ln == lle) LLE_WLOCK(lle); } else if (lle == NULL || ln != lle) { /* * We did the lookup (no lle arg) so we * need to do the unlock here. */ LLE_WUNLOCK(ln); } return (0); sendpkt: /* discard the packet if IPv6 operation is disabled on the interface */ if ((ND_IFINFO(ifp)->flags & ND6_IFF_IFDISABLED)) { error = ENETDOWN; /* better error? */ goto bad; } /* * ln is valid and the caller did not pass in * an llentry */ if ((ln != NULL) && (lle == NULL)) { if (flags & LLE_EXCLUSIVE) LLE_WUNLOCK(ln); else LLE_RUNLOCK(ln); } #ifdef MAC mac_netinet6_nd6_send(ifp, m); #endif /* * If called from nd6_ns_output() (NS), nd6_na_output() (NA), * icmp6_redirect_output() (REDIRECT) or from rip6_output() (RS, RA * as handled by rtsol and rtadvd), mbufs will be tagged for SeND * to be diverted to user space. When re-injected into the kernel, * send_output() will directly dispatch them to the outgoing interface. */ if (send_sendso_input_hook != NULL) { mtag = m_tag_find(m, PACKET_TAG_ND_OUTGOING, NULL); if (mtag != NULL) { ip6 = mtod(m, struct ip6_hdr *); ip6len = sizeof(struct ip6_hdr) + ntohs(ip6->ip6_plen); /* Use the SEND socket */ error = send_sendso_input_hook(m, ifp, SND_OUT, ip6len); /* -1 == no app on SEND socket */ if (error == 0 || error != -1) return (error); } } /* * We were passed in a pointer to an lle with the lock held * this means that we can't call if_output as we will * recurse on the lle lock - so what we do is we create * a list of mbufs to send and transmit them in the caller * after the lock is dropped */ if (lle != NULL) { if (*chain == NULL) *chain = m; else { struct mbuf *mb; /* * append mbuf to end of deferred chain */ mb = *chain; while (mb->m_nextpkt != NULL) mb = mb->m_nextpkt; mb->m_nextpkt = m; } return (error); } /* Reset layer specific mbuf flags to avoid confusing lower layers. */ m->m_flags &= ~(M_PROTOFLAGS); if ((ifp->if_flags & IFF_LOOPBACK) != 0) { return ((*ifp->if_output)(origifp, m, (struct sockaddr *)dst, NULL)); } error = (*ifp->if_output)(ifp, m, (struct sockaddr *)dst, NULL); return (error); bad: /* * ln is valid and the caller did not pass in * an llentry */ if ((ln != NULL) && (lle == NULL)) { if (flags & LLE_EXCLUSIVE) LLE_WUNLOCK(ln); else LLE_RUNLOCK(ln); } if (m) m_freem(m); return (error); } #undef senderr int nd6_output_flush(struct ifnet *ifp, struct ifnet *origifp, struct mbuf *chain, struct sockaddr_in6 *dst, struct route *ro) { struct mbuf *m, *m_head; struct ifnet *outifp; int error = 0; m_head = chain; if ((ifp->if_flags & IFF_LOOPBACK) != 0) outifp = origifp; else outifp = ifp; while (m_head) { m = m_head; m_head = m_head->m_nextpkt; error = (*ifp->if_output)(ifp, m, (struct sockaddr *)dst, ro); } /* * XXX * note that intermediate errors are blindly ignored - but this is * the same convention as used with nd6_output when called by * nd6_cache_lladdr */ return (error); } int nd6_need_cache(struct ifnet *ifp) { /* * XXX: we currently do not make neighbor cache on any interface * other than ARCnet, Ethernet, FDDI and GIF. * * RFC2893 says: * - unidirectional tunnels needs no ND */ switch (ifp->if_type) { case IFT_ARCNET: case IFT_ETHER: case IFT_FDDI: case IFT_IEEE1394: #ifdef IFT_L2VLAN case IFT_L2VLAN: #endif #ifdef IFT_IEEE80211 case IFT_IEEE80211: #endif case IFT_INFINIBAND: case IFT_GIF: /* XXX need more cases? */ case IFT_PPP: case IFT_TUNNEL: case IFT_BRIDGE: case IFT_PROPVIRTUAL: return (1); default: return (0); } } /* * the callers of this function need to be re-worked to drop * the lle lock, drop here for now */ int nd6_storelladdr(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst, u_char *desten, struct llentry **lle) { struct llentry *ln; *lle = NULL; IF_AFDATA_UNLOCK_ASSERT(ifp); if (m != NULL && m->m_flags & M_MCAST) { int i; switch (ifp->if_type) { case IFT_ETHER: case IFT_FDDI: #ifdef IFT_L2VLAN case IFT_L2VLAN: #endif #ifdef IFT_IEEE80211 case IFT_IEEE80211: #endif case IFT_BRIDGE: case IFT_ISO88025: ETHER_MAP_IPV6_MULTICAST(&SIN6(dst)->sin6_addr, desten); return (0); case IFT_IEEE1394: /* * netbsd can use if_broadcastaddr, but we don't do so * to reduce # of ifdef. */ for (i = 0; i < ifp->if_addrlen; i++) desten[i] = ~0; return (0); case IFT_ARCNET: *desten = 0; return (0); default: m_freem(m); return (EAFNOSUPPORT); } } /* * the entry should have been created in nd6_store_lladdr */ IF_AFDATA_RLOCK(ifp); ln = lla_lookup(LLTABLE6(ifp), 0, dst); IF_AFDATA_RUNLOCK(ifp); if ((ln == NULL) || !(ln->la_flags & LLE_VALID)) { if (ln != NULL) LLE_RUNLOCK(ln); /* this could happen, if we could not allocate memory */ m_freem(m); return (1); } bcopy(&ln->ll_addr, desten, ifp->if_addrlen); *lle = ln; LLE_RUNLOCK(ln); /* * A *small* use after free race exists here */ return (0); } static void clear_llinfo_pqueue(struct llentry *ln) { struct mbuf *m_hold, *m_hold_next; for (m_hold = ln->la_hold; m_hold; m_hold = m_hold_next) { m_hold_next = m_hold->m_nextpkt; m_freem(m_hold); } ln->la_hold = NULL; return; } static int nd6_sysctl_drlist(SYSCTL_HANDLER_ARGS); static int nd6_sysctl_prlist(SYSCTL_HANDLER_ARGS); #ifdef SYSCTL_DECL SYSCTL_DECL(_net_inet6_icmp6); #endif SYSCTL_NODE(_net_inet6_icmp6, ICMPV6CTL_ND6_DRLIST, nd6_drlist, CTLFLAG_RD, nd6_sysctl_drlist, ""); SYSCTL_NODE(_net_inet6_icmp6, ICMPV6CTL_ND6_PRLIST, nd6_prlist, CTLFLAG_RD, nd6_sysctl_prlist, ""); SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_MAXQLEN, nd6_maxqueuelen, CTLFLAG_RW, &VNET_NAME(nd6_maxqueuelen), 1, ""); static int nd6_sysctl_drlist(SYSCTL_HANDLER_ARGS) { struct in6_defrouter d; struct nd_defrouter *dr; int error; if (req->newptr) return (EPERM); bzero(&d, sizeof(d)); d.rtaddr.sin6_family = AF_INET6; d.rtaddr.sin6_len = sizeof(d.rtaddr); /* * XXX locking */ TAILQ_FOREACH(dr, &V_nd_defrouter, dr_entry) { d.rtaddr.sin6_addr = dr->rtaddr; error = sa6_recoverscope(&d.rtaddr); if (error != 0) return (error); d.flags = dr->flags; d.rtlifetime = dr->rtlifetime; d.expire = dr->expire; d.if_index = dr->ifp->if_index; error = SYSCTL_OUT(req, &d, sizeof(d)); if (error != 0) return (error); } return (0); } static int nd6_sysctl_prlist(SYSCTL_HANDLER_ARGS) { struct in6_prefix p; struct sockaddr_in6 s6; struct nd_prefix *pr; struct nd_pfxrouter *pfr; time_t maxexpire; int error; char ip6buf[INET6_ADDRSTRLEN]; if (req->newptr) return (EPERM); bzero(&p, sizeof(p)); p.origin = PR_ORIG_RA; bzero(&s6, sizeof(s6)); s6.sin6_family = AF_INET6; s6.sin6_len = sizeof(s6); /* * XXX locking */ LIST_FOREACH(pr, &V_nd_prefix, ndpr_entry) { p.prefix = pr->ndpr_prefix; if (sa6_recoverscope(&p.prefix)) { log(LOG_ERR, "scope error in prefix list (%s)\n", ip6_sprintf(ip6buf, &p.prefix.sin6_addr)); /* XXX: press on... */ } p.raflags = pr->ndpr_raf; p.prefixlen = pr->ndpr_plen; p.vltime = pr->ndpr_vltime; p.pltime = pr->ndpr_pltime; p.if_index = pr->ndpr_ifp->if_index; if (pr->ndpr_vltime == ND6_INFINITE_LIFETIME) p.expire = 0; else { /* XXX: we assume time_t is signed. */ maxexpire = (-1) & ~((time_t)1 << ((sizeof(maxexpire) * 8) - 1)); if (pr->ndpr_vltime < maxexpire - pr->ndpr_lastupdate) p.expire = pr->ndpr_lastupdate + pr->ndpr_vltime; else p.expire = maxexpire; } p.refcnt = pr->ndpr_refcnt; p.flags = pr->ndpr_stateflags; p.advrtrs = 0; LIST_FOREACH(pfr, &pr->ndpr_advrtrs, pfr_entry) p.advrtrs++; error = SYSCTL_OUT(req, &p, sizeof(p)); if (error != 0) return (error); LIST_FOREACH(pfr, &pr->ndpr_advrtrs, pfr_entry) { s6.sin6_addr = pfr->router->rtaddr; if (sa6_recoverscope(&s6)) log(LOG_ERR, "scope error in prefix list (%s)\n", ip6_sprintf(ip6buf, &pfr->router->rtaddr)); error = SYSCTL_OUT(req, &s6, sizeof(s6)); if (error != 0) return (error); } } return (0); } Index: head/sys/netinet6/nd6_rtr.c =================================================================== --- head/sys/netinet6/nd6_rtr.c (revision 253969) +++ head/sys/netinet6/nd6_rtr.c (revision 253970) @@ -1,2173 +1,2173 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $KAME: nd6_rtr.c,v 1.111 2001/04/27 01:37:15 jinmei Exp $ */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include static int rtpref(struct nd_defrouter *); static struct nd_defrouter *defrtrlist_update(struct nd_defrouter *); static int prelist_update(struct nd_prefixctl *, struct nd_defrouter *, struct mbuf *, int); static struct in6_ifaddr *in6_ifadd(struct nd_prefixctl *, int); static struct nd_pfxrouter *pfxrtr_lookup(struct nd_prefix *, struct nd_defrouter *); static void pfxrtr_add(struct nd_prefix *, struct nd_defrouter *); static void pfxrtr_del(struct nd_pfxrouter *); static struct nd_pfxrouter *find_pfxlist_reachable_router (struct nd_prefix *); static void defrouter_delreq(struct nd_defrouter *); static void nd6_rtmsg(int, struct rtentry *); static int in6_init_prefix_ltimes(struct nd_prefix *); static void in6_init_address_ltimes(struct nd_prefix *, struct in6_addrlifetime *); static int nd6_prefix_onlink(struct nd_prefix *); static int nd6_prefix_offlink(struct nd_prefix *); static int rt6_deleteroute(struct radix_node *, void *); VNET_DECLARE(int, nd6_recalc_reachtm_interval); #define V_nd6_recalc_reachtm_interval VNET(nd6_recalc_reachtm_interval) static VNET_DEFINE(struct ifnet *, nd6_defifp); VNET_DEFINE(int, nd6_defifindex); #define V_nd6_defifp VNET(nd6_defifp) VNET_DEFINE(int, ip6_use_tempaddr) = 0; VNET_DEFINE(int, ip6_desync_factor); VNET_DEFINE(u_int32_t, ip6_temp_preferred_lifetime) = DEF_TEMP_PREFERRED_LIFETIME; VNET_DEFINE(u_int32_t, ip6_temp_valid_lifetime) = DEF_TEMP_VALID_LIFETIME; VNET_DEFINE(int, ip6_temp_regen_advance) = TEMPADDR_REGEN_ADVANCE; /* RTPREF_MEDIUM has to be 0! */ #define RTPREF_HIGH 1 #define RTPREF_MEDIUM 0 #define RTPREF_LOW (-1) #define RTPREF_RESERVED (-2) #define RTPREF_INVALID (-3) /* internal */ /* * Receive Router Solicitation Message - just for routers. * Router solicitation/advertisement is mostly managed by userland program * (rtadvd) so here we have no function like nd6_ra_output(). * * Based on RFC 2461 */ void nd6_rs_input(struct mbuf *m, int off, int icmp6len) { struct ifnet *ifp = m->m_pkthdr.rcvif; struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); struct nd_router_solicit *nd_rs; struct in6_addr saddr6 = ip6->ip6_src; char *lladdr = NULL; int lladdrlen = 0; union nd_opts ndopts; char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN]; /* * Accept RS only when V_ip6_forwarding=1 and the interface has * no ND6_IFF_ACCEPT_RTADV. */ if (!V_ip6_forwarding || ND_IFINFO(ifp)->flags & ND6_IFF_ACCEPT_RTADV) goto freeit; /* Sanity checks */ if (ip6->ip6_hlim != 255) { nd6log((LOG_ERR, "nd6_rs_input: invalid hlim (%d) from %s to %s on %s\n", ip6->ip6_hlim, ip6_sprintf(ip6bufs, &ip6->ip6_src), ip6_sprintf(ip6bufd, &ip6->ip6_dst), if_name(ifp))); goto bad; } /* * Don't update the neighbor cache, if src = ::. * This indicates that the src has no IP address assigned yet. */ if (IN6_IS_ADDR_UNSPECIFIED(&saddr6)) goto freeit; #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, off, icmp6len,); nd_rs = (struct nd_router_solicit *)((caddr_t)ip6 + off); #else IP6_EXTHDR_GET(nd_rs, struct nd_router_solicit *, m, off, icmp6len); if (nd_rs == NULL) { ICMP6STAT_INC(icp6s_tooshort); return; } #endif icmp6len -= sizeof(*nd_rs); nd6_option_init(nd_rs + 1, icmp6len, &ndopts); if (nd6_options(&ndopts) < 0) { nd6log((LOG_INFO, "nd6_rs_input: invalid ND option, ignored\n")); /* nd6_options have incremented stats */ goto freeit; } if (ndopts.nd_opts_src_lladdr) { lladdr = (char *)(ndopts.nd_opts_src_lladdr + 1); lladdrlen = ndopts.nd_opts_src_lladdr->nd_opt_len << 3; } if (lladdr && ((ifp->if_addrlen + 2 + 7) & ~7) != lladdrlen) { nd6log((LOG_INFO, "nd6_rs_input: lladdrlen mismatch for %s " "(if %d, RS packet %d)\n", ip6_sprintf(ip6bufs, &saddr6), ifp->if_addrlen, lladdrlen - 2)); goto bad; } nd6_cache_lladdr(ifp, &saddr6, lladdr, lladdrlen, ND_ROUTER_SOLICIT, 0); freeit: m_freem(m); return; bad: ICMP6STAT_INC(icp6s_badrs); m_freem(m); } /* * Receive Router Advertisement Message. * * Based on RFC 2461 * TODO: on-link bit on prefix information * TODO: ND_RA_FLAG_{OTHER,MANAGED} processing */ void nd6_ra_input(struct mbuf *m, int off, int icmp6len) { struct ifnet *ifp = m->m_pkthdr.rcvif; struct nd_ifinfo *ndi = ND_IFINFO(ifp); struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); struct nd_router_advert *nd_ra; struct in6_addr saddr6 = ip6->ip6_src; int mcast = 0; union nd_opts ndopts; struct nd_defrouter *dr; char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN]; /* * We only accept RAs only when the per-interface flag * ND6_IFF_ACCEPT_RTADV is on the receiving interface. */ if (!(ndi->flags & ND6_IFF_ACCEPT_RTADV)) goto freeit; if (ip6->ip6_hlim != 255) { nd6log((LOG_ERR, "nd6_ra_input: invalid hlim (%d) from %s to %s on %s\n", ip6->ip6_hlim, ip6_sprintf(ip6bufs, &ip6->ip6_src), ip6_sprintf(ip6bufd, &ip6->ip6_dst), if_name(ifp))); goto bad; } if (!IN6_IS_ADDR_LINKLOCAL(&saddr6)) { nd6log((LOG_ERR, "nd6_ra_input: src %s is not link-local\n", ip6_sprintf(ip6bufs, &saddr6))); goto bad; } #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, off, icmp6len,); nd_ra = (struct nd_router_advert *)((caddr_t)ip6 + off); #else IP6_EXTHDR_GET(nd_ra, struct nd_router_advert *, m, off, icmp6len); if (nd_ra == NULL) { ICMP6STAT_INC(icp6s_tooshort); return; } #endif icmp6len -= sizeof(*nd_ra); nd6_option_init(nd_ra + 1, icmp6len, &ndopts); if (nd6_options(&ndopts) < 0) { nd6log((LOG_INFO, "nd6_ra_input: invalid ND option, ignored\n")); /* nd6_options have incremented stats */ goto freeit; } { struct nd_defrouter dr0; u_int32_t advreachable = nd_ra->nd_ra_reachable; /* remember if this is a multicasted advertisement */ if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) mcast = 1; bzero(&dr0, sizeof(dr0)); dr0.rtaddr = saddr6; dr0.flags = nd_ra->nd_ra_flags_reserved; /* * Effectively-disable routes from RA messages when * ND6_IFF_NO_RADR enabled on the receiving interface or * (ip6.forwarding == 1 && ip6.rfc6204w3 != 1). */ if (ndi->flags & ND6_IFF_NO_RADR) dr0.rtlifetime = 0; else if (V_ip6_forwarding && !V_ip6_rfc6204w3) dr0.rtlifetime = 0; else dr0.rtlifetime = ntohs(nd_ra->nd_ra_router_lifetime); - dr0.expire = time_second + dr0.rtlifetime; + dr0.expire = time_uptime + dr0.rtlifetime; dr0.ifp = ifp; /* unspecified or not? (RFC 2461 6.3.4) */ if (advreachable) { advreachable = ntohl(advreachable); if (advreachable <= MAX_REACHABLE_TIME && ndi->basereachable != advreachable) { ndi->basereachable = advreachable; ndi->reachable = ND_COMPUTE_RTIME(ndi->basereachable); ndi->recalctm = V_nd6_recalc_reachtm_interval; /* reset */ } } if (nd_ra->nd_ra_retransmit) ndi->retrans = ntohl(nd_ra->nd_ra_retransmit); if (nd_ra->nd_ra_curhoplimit) ndi->chlim = nd_ra->nd_ra_curhoplimit; dr = defrtrlist_update(&dr0); } /* * prefix */ if (ndopts.nd_opts_pi) { struct nd_opt_hdr *pt; struct nd_opt_prefix_info *pi = NULL; struct nd_prefixctl pr; for (pt = (struct nd_opt_hdr *)ndopts.nd_opts_pi; pt <= (struct nd_opt_hdr *)ndopts.nd_opts_pi_end; pt = (struct nd_opt_hdr *)((caddr_t)pt + (pt->nd_opt_len << 3))) { if (pt->nd_opt_type != ND_OPT_PREFIX_INFORMATION) continue; pi = (struct nd_opt_prefix_info *)pt; if (pi->nd_opt_pi_len != 4) { nd6log((LOG_INFO, "nd6_ra_input: invalid option " "len %d for prefix information option, " "ignored\n", pi->nd_opt_pi_len)); continue; } if (128 < pi->nd_opt_pi_prefix_len) { nd6log((LOG_INFO, "nd6_ra_input: invalid prefix " "len %d for prefix information option, " "ignored\n", pi->nd_opt_pi_prefix_len)); continue; } if (IN6_IS_ADDR_MULTICAST(&pi->nd_opt_pi_prefix) || IN6_IS_ADDR_LINKLOCAL(&pi->nd_opt_pi_prefix)) { nd6log((LOG_INFO, "nd6_ra_input: invalid prefix " "%s, ignored\n", ip6_sprintf(ip6bufs, &pi->nd_opt_pi_prefix))); continue; } bzero(&pr, sizeof(pr)); pr.ndpr_prefix.sin6_family = AF_INET6; pr.ndpr_prefix.sin6_len = sizeof(pr.ndpr_prefix); pr.ndpr_prefix.sin6_addr = pi->nd_opt_pi_prefix; pr.ndpr_ifp = (struct ifnet *)m->m_pkthdr.rcvif; pr.ndpr_raf_onlink = (pi->nd_opt_pi_flags_reserved & ND_OPT_PI_FLAG_ONLINK) ? 1 : 0; pr.ndpr_raf_auto = (pi->nd_opt_pi_flags_reserved & ND_OPT_PI_FLAG_AUTO) ? 1 : 0; pr.ndpr_plen = pi->nd_opt_pi_prefix_len; pr.ndpr_vltime = ntohl(pi->nd_opt_pi_valid_time); pr.ndpr_pltime = ntohl(pi->nd_opt_pi_preferred_time); (void)prelist_update(&pr, dr, m, mcast); } } /* * MTU */ if (ndopts.nd_opts_mtu && ndopts.nd_opts_mtu->nd_opt_mtu_len == 1) { u_long mtu; u_long maxmtu; mtu = (u_long)ntohl(ndopts.nd_opts_mtu->nd_opt_mtu_mtu); /* lower bound */ if (mtu < IPV6_MMTU) { nd6log((LOG_INFO, "nd6_ra_input: bogus mtu option " "mtu=%lu sent from %s, ignoring\n", mtu, ip6_sprintf(ip6bufs, &ip6->ip6_src))); goto skip; } /* upper bound */ maxmtu = (ndi->maxmtu && ndi->maxmtu < ifp->if_mtu) ? ndi->maxmtu : ifp->if_mtu; if (mtu <= maxmtu) { int change = (ndi->linkmtu != mtu); ndi->linkmtu = mtu; if (change) /* in6_maxmtu may change */ in6_setmaxmtu(); } else { nd6log((LOG_INFO, "nd6_ra_input: bogus mtu " "mtu=%lu sent from %s; " "exceeds maxmtu %lu, ignoring\n", mtu, ip6_sprintf(ip6bufs, &ip6->ip6_src), maxmtu)); } } skip: /* * Source link layer address */ { char *lladdr = NULL; int lladdrlen = 0; if (ndopts.nd_opts_src_lladdr) { lladdr = (char *)(ndopts.nd_opts_src_lladdr + 1); lladdrlen = ndopts.nd_opts_src_lladdr->nd_opt_len << 3; } if (lladdr && ((ifp->if_addrlen + 2 + 7) & ~7) != lladdrlen) { nd6log((LOG_INFO, "nd6_ra_input: lladdrlen mismatch for %s " "(if %d, RA packet %d)\n", ip6_sprintf(ip6bufs, &saddr6), ifp->if_addrlen, lladdrlen - 2)); goto bad; } nd6_cache_lladdr(ifp, &saddr6, lladdr, lladdrlen, ND_ROUTER_ADVERT, 0); /* * Installing a link-layer address might change the state of the * router's neighbor cache, which might also affect our on-link * detection of adveritsed prefixes. */ pfxlist_onlink_check(); } freeit: m_freem(m); return; bad: ICMP6STAT_INC(icp6s_badra); m_freem(m); } /* * default router list proccessing sub routines */ /* tell the change to user processes watching the routing socket. */ static void nd6_rtmsg(int cmd, struct rtentry *rt) { struct rt_addrinfo info; struct ifnet *ifp; struct ifaddr *ifa; bzero((caddr_t)&info, sizeof(info)); info.rti_info[RTAX_DST] = rt_key(rt); info.rti_info[RTAX_GATEWAY] = rt->rt_gateway; info.rti_info[RTAX_NETMASK] = rt_mask(rt); ifp = rt->rt_ifp; if (ifp != NULL) { IF_ADDR_RLOCK(ifp); ifa = TAILQ_FIRST(&ifp->if_addrhead); info.rti_info[RTAX_IFP] = ifa->ifa_addr; ifa_ref(ifa); IF_ADDR_RUNLOCK(ifp); info.rti_info[RTAX_IFA] = rt->rt_ifa->ifa_addr; } else ifa = NULL; rt_missmsg_fib(cmd, &info, rt->rt_flags, 0, rt->rt_fibnum); if (ifa != NULL) ifa_free(ifa); } static void defrouter_addreq(struct nd_defrouter *new) { struct sockaddr_in6 def, mask, gate; struct rtentry *newrt = NULL; int error; bzero(&def, sizeof(def)); bzero(&mask, sizeof(mask)); bzero(&gate, sizeof(gate)); def.sin6_len = mask.sin6_len = gate.sin6_len = sizeof(struct sockaddr_in6); def.sin6_family = gate.sin6_family = AF_INET6; gate.sin6_addr = new->rtaddr; error = in6_rtrequest(RTM_ADD, (struct sockaddr *)&def, (struct sockaddr *)&gate, (struct sockaddr *)&mask, RTF_GATEWAY, &newrt, RT_DEFAULT_FIB); if (newrt) { nd6_rtmsg(RTM_ADD, newrt); /* tell user process */ RTFREE(newrt); } if (error == 0) new->installed = 1; return; } struct nd_defrouter * defrouter_lookup(struct in6_addr *addr, struct ifnet *ifp) { struct nd_defrouter *dr; TAILQ_FOREACH(dr, &V_nd_defrouter, dr_entry) { if (dr->ifp == ifp && IN6_ARE_ADDR_EQUAL(addr, &dr->rtaddr)) return (dr); } return (NULL); /* search failed */ } /* * Remove the default route for a given router. * This is just a subroutine function for defrouter_select(), and should * not be called from anywhere else. */ static void defrouter_delreq(struct nd_defrouter *dr) { struct sockaddr_in6 def, mask, gate; struct rtentry *oldrt = NULL; bzero(&def, sizeof(def)); bzero(&mask, sizeof(mask)); bzero(&gate, sizeof(gate)); def.sin6_len = mask.sin6_len = gate.sin6_len = sizeof(struct sockaddr_in6); def.sin6_family = gate.sin6_family = AF_INET6; gate.sin6_addr = dr->rtaddr; in6_rtrequest(RTM_DELETE, (struct sockaddr *)&def, (struct sockaddr *)&gate, (struct sockaddr *)&mask, RTF_GATEWAY, &oldrt, RT_DEFAULT_FIB); if (oldrt) { nd6_rtmsg(RTM_DELETE, oldrt); RTFREE(oldrt); } dr->installed = 0; } /* * remove all default routes from default router list */ void defrouter_reset(void) { struct nd_defrouter *dr; TAILQ_FOREACH(dr, &V_nd_defrouter, dr_entry) defrouter_delreq(dr); /* * XXX should we also nuke any default routers in the kernel, by * going through them by rtalloc1()? */ } void defrtrlist_del(struct nd_defrouter *dr) { struct nd_defrouter *deldr = NULL; struct nd_prefix *pr; /* * Flush all the routing table entries that use the router * as a next hop. */ if (ND_IFINFO(dr->ifp)->flags & ND6_IFF_ACCEPT_RTADV) rt6_flush(&dr->rtaddr, dr->ifp); if (dr->installed) { deldr = dr; defrouter_delreq(dr); } TAILQ_REMOVE(&V_nd_defrouter, dr, dr_entry); /* * Also delete all the pointers to the router in each prefix lists. */ LIST_FOREACH(pr, &V_nd_prefix, ndpr_entry) { struct nd_pfxrouter *pfxrtr; if ((pfxrtr = pfxrtr_lookup(pr, dr)) != NULL) pfxrtr_del(pfxrtr); } pfxlist_onlink_check(); /* * If the router is the primary one, choose a new one. * Note that defrouter_select() will remove the current gateway * from the routing table. */ if (deldr) defrouter_select(); free(dr, M_IP6NDP); } /* * Default Router Selection according to Section 6.3.6 of RFC 2461 and * draft-ietf-ipngwg-router-selection: * 1) Routers that are reachable or probably reachable should be preferred. * If we have more than one (probably) reachable router, prefer ones * with the highest router preference. * 2) When no routers on the list are known to be reachable or * probably reachable, routers SHOULD be selected in a round-robin * fashion, regardless of router preference values. * 3) If the Default Router List is empty, assume that all * destinations are on-link. * * We assume nd_defrouter is sorted by router preference value. * Since the code below covers both with and without router preference cases, * we do not need to classify the cases by ifdef. * * At this moment, we do not try to install more than one default router, * even when the multipath routing is available, because we're not sure about * the benefits for stub hosts comparing to the risk of making the code * complicated and the possibility of introducing bugs. */ void defrouter_select(void) { struct nd_defrouter *dr, *selected_dr = NULL, *installed_dr = NULL; struct llentry *ln = NULL; /* * Let's handle easy case (3) first: * If default router list is empty, there's nothing to be done. */ if (TAILQ_EMPTY(&V_nd_defrouter)) return; /* * Search for a (probably) reachable router from the list. * We just pick up the first reachable one (if any), assuming that * the ordering rule of the list described in defrtrlist_update(). */ TAILQ_FOREACH(dr, &V_nd_defrouter, dr_entry) { IF_AFDATA_RLOCK(dr->ifp); if (selected_dr == NULL && (ln = nd6_lookup(&dr->rtaddr, 0, dr->ifp)) && ND6_IS_LLINFO_PROBREACH(ln)) { selected_dr = dr; } IF_AFDATA_RUNLOCK(dr->ifp); if (ln != NULL) { LLE_RUNLOCK(ln); ln = NULL; } if (dr->installed && installed_dr == NULL) installed_dr = dr; else if (dr->installed && installed_dr) { /* this should not happen. warn for diagnosis. */ log(LOG_ERR, "defrouter_select: more than one router" " is installed\n"); } } /* * If none of the default routers was found to be reachable, * round-robin the list regardless of preference. * Otherwise, if we have an installed router, check if the selected * (reachable) router should really be preferred to the installed one. * We only prefer the new router when the old one is not reachable * or when the new one has a really higher preference value. */ if (selected_dr == NULL) { if (installed_dr == NULL || !TAILQ_NEXT(installed_dr, dr_entry)) selected_dr = TAILQ_FIRST(&V_nd_defrouter); else selected_dr = TAILQ_NEXT(installed_dr, dr_entry); } else if (installed_dr) { IF_AFDATA_RLOCK(installed_dr->ifp); if ((ln = nd6_lookup(&installed_dr->rtaddr, 0, installed_dr->ifp)) && ND6_IS_LLINFO_PROBREACH(ln) && rtpref(selected_dr) <= rtpref(installed_dr)) { selected_dr = installed_dr; } IF_AFDATA_RUNLOCK(installed_dr->ifp); if (ln != NULL) LLE_RUNLOCK(ln); } /* * If the selected router is different than the installed one, * remove the installed router and install the selected one. * Note that the selected router is never NULL here. */ if (installed_dr != selected_dr) { if (installed_dr) defrouter_delreq(installed_dr); defrouter_addreq(selected_dr); } return; } /* * for default router selection * regards router-preference field as a 2-bit signed integer */ static int rtpref(struct nd_defrouter *dr) { switch (dr->flags & ND_RA_FLAG_RTPREF_MASK) { case ND_RA_FLAG_RTPREF_HIGH: return (RTPREF_HIGH); case ND_RA_FLAG_RTPREF_MEDIUM: case ND_RA_FLAG_RTPREF_RSV: return (RTPREF_MEDIUM); case ND_RA_FLAG_RTPREF_LOW: return (RTPREF_LOW); default: /* * This case should never happen. If it did, it would mean a * serious bug of kernel internal. We thus always bark here. * Or, can we even panic? */ log(LOG_ERR, "rtpref: impossible RA flag %x\n", dr->flags); return (RTPREF_INVALID); } /* NOTREACHED */ } static struct nd_defrouter * defrtrlist_update(struct nd_defrouter *new) { struct nd_defrouter *dr, *n; if ((dr = defrouter_lookup(&new->rtaddr, new->ifp)) != NULL) { /* entry exists */ if (new->rtlifetime == 0) { defrtrlist_del(dr); dr = NULL; } else { int oldpref = rtpref(dr); /* override */ dr->flags = new->flags; /* xxx flag check */ dr->rtlifetime = new->rtlifetime; dr->expire = new->expire; /* * If the preference does not change, there's no need * to sort the entries. Also make sure the selected * router is still installed in the kernel. */ if (dr->installed && rtpref(new) == oldpref) return (dr); /* * preferred router may be changed, so relocate * this router. * XXX: calling TAILQ_REMOVE directly is a bad manner. * However, since defrtrlist_del() has many side * effects, we intentionally do so here. * defrouter_select() below will handle routing * changes later. */ TAILQ_REMOVE(&V_nd_defrouter, dr, dr_entry); n = dr; goto insert; } return (dr); } /* entry does not exist */ if (new->rtlifetime == 0) return (NULL); n = (struct nd_defrouter *)malloc(sizeof(*n), M_IP6NDP, M_NOWAIT); if (n == NULL) return (NULL); bzero(n, sizeof(*n)); *n = *new; insert: /* * Insert the new router in the Default Router List; * The Default Router List should be in the descending order * of router-preferece. Routers with the same preference are * sorted in the arriving time order. */ /* insert at the end of the group */ TAILQ_FOREACH(dr, &V_nd_defrouter, dr_entry) { if (rtpref(n) > rtpref(dr)) break; } if (dr) TAILQ_INSERT_BEFORE(dr, n, dr_entry); else TAILQ_INSERT_TAIL(&V_nd_defrouter, n, dr_entry); defrouter_select(); return (n); } static struct nd_pfxrouter * pfxrtr_lookup(struct nd_prefix *pr, struct nd_defrouter *dr) { struct nd_pfxrouter *search; LIST_FOREACH(search, &pr->ndpr_advrtrs, pfr_entry) { if (search->router == dr) break; } return (search); } static void pfxrtr_add(struct nd_prefix *pr, struct nd_defrouter *dr) { struct nd_pfxrouter *new; new = (struct nd_pfxrouter *)malloc(sizeof(*new), M_IP6NDP, M_NOWAIT); if (new == NULL) return; bzero(new, sizeof(*new)); new->router = dr; LIST_INSERT_HEAD(&pr->ndpr_advrtrs, new, pfr_entry); pfxlist_onlink_check(); } static void pfxrtr_del(struct nd_pfxrouter *pfr) { LIST_REMOVE(pfr, pfr_entry); free(pfr, M_IP6NDP); } struct nd_prefix * nd6_prefix_lookup(struct nd_prefixctl *key) { struct nd_prefix *search; LIST_FOREACH(search, &V_nd_prefix, ndpr_entry) { if (key->ndpr_ifp == search->ndpr_ifp && key->ndpr_plen == search->ndpr_plen && in6_are_prefix_equal(&key->ndpr_prefix.sin6_addr, &search->ndpr_prefix.sin6_addr, key->ndpr_plen)) { break; } } return (search); } int nd6_prelist_add(struct nd_prefixctl *pr, struct nd_defrouter *dr, struct nd_prefix **newp) { struct nd_prefix *new = NULL; int error = 0; int i; char ip6buf[INET6_ADDRSTRLEN]; new = (struct nd_prefix *)malloc(sizeof(*new), M_IP6NDP, M_NOWAIT); if (new == NULL) return(ENOMEM); bzero(new, sizeof(*new)); new->ndpr_ifp = pr->ndpr_ifp; new->ndpr_prefix = pr->ndpr_prefix; new->ndpr_plen = pr->ndpr_plen; new->ndpr_vltime = pr->ndpr_vltime; new->ndpr_pltime = pr->ndpr_pltime; new->ndpr_flags = pr->ndpr_flags; if ((error = in6_init_prefix_ltimes(new)) != 0) { free(new, M_IP6NDP); return(error); } - new->ndpr_lastupdate = time_second; + new->ndpr_lastupdate = time_uptime; if (newp != NULL) *newp = new; /* initialization */ LIST_INIT(&new->ndpr_advrtrs); in6_prefixlen2mask(&new->ndpr_mask, new->ndpr_plen); /* make prefix in the canonical form */ for (i = 0; i < 4; i++) new->ndpr_prefix.sin6_addr.s6_addr32[i] &= new->ndpr_mask.s6_addr32[i]; /* link ndpr_entry to nd_prefix list */ LIST_INSERT_HEAD(&V_nd_prefix, new, ndpr_entry); /* ND_OPT_PI_FLAG_ONLINK processing */ if (new->ndpr_raf_onlink) { int e; if ((e = nd6_prefix_onlink(new)) != 0) { nd6log((LOG_ERR, "nd6_prelist_add: failed to make " "the prefix %s/%d on-link on %s (errno=%d)\n", ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), pr->ndpr_plen, if_name(pr->ndpr_ifp), e)); /* proceed anyway. XXX: is it correct? */ } } if (dr) pfxrtr_add(new, dr); return 0; } void prelist_remove(struct nd_prefix *pr) { struct nd_pfxrouter *pfr, *next; int e; char ip6buf[INET6_ADDRSTRLEN]; /* make sure to invalidate the prefix until it is really freed. */ pr->ndpr_vltime = 0; pr->ndpr_pltime = 0; /* * Though these flags are now meaningless, we'd rather keep the value * of pr->ndpr_raf_onlink and pr->ndpr_raf_auto not to confuse users * when executing "ndp -p". */ if ((pr->ndpr_stateflags & NDPRF_ONLINK) != 0 && (e = nd6_prefix_offlink(pr)) != 0) { nd6log((LOG_ERR, "prelist_remove: failed to make %s/%d offlink " "on %s, errno=%d\n", ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), pr->ndpr_plen, if_name(pr->ndpr_ifp), e)); /* what should we do? */ } if (pr->ndpr_refcnt > 0) return; /* notice here? */ /* unlink ndpr_entry from nd_prefix list */ LIST_REMOVE(pr, ndpr_entry); /* free list of routers that adversed the prefix */ LIST_FOREACH_SAFE(pfr, &pr->ndpr_advrtrs, pfr_entry, next) { free(pfr, M_IP6NDP); } free(pr, M_IP6NDP); pfxlist_onlink_check(); } /* * dr - may be NULL */ static int prelist_update(struct nd_prefixctl *new, struct nd_defrouter *dr, struct mbuf *m, int mcast) { struct in6_ifaddr *ia6 = NULL, *ia6_match = NULL; struct ifaddr *ifa; struct ifnet *ifp = new->ndpr_ifp; struct nd_prefix *pr; int error = 0; int newprefix = 0; int auth; struct in6_addrlifetime lt6_tmp; char ip6buf[INET6_ADDRSTRLEN]; auth = 0; if (m) { /* * Authenticity for NA consists authentication for * both IP header and IP datagrams, doesn't it ? */ #if defined(M_AUTHIPHDR) && defined(M_AUTHIPDGM) auth = ((m->m_flags & M_AUTHIPHDR) && (m->m_flags & M_AUTHIPDGM)); #endif } if ((pr = nd6_prefix_lookup(new)) != NULL) { /* * nd6_prefix_lookup() ensures that pr and new have the same * prefix on a same interface. */ /* * Update prefix information. Note that the on-link (L) bit * and the autonomous (A) bit should NOT be changed from 1 * to 0. */ if (new->ndpr_raf_onlink == 1) pr->ndpr_raf_onlink = 1; if (new->ndpr_raf_auto == 1) pr->ndpr_raf_auto = 1; if (new->ndpr_raf_onlink) { pr->ndpr_vltime = new->ndpr_vltime; pr->ndpr_pltime = new->ndpr_pltime; (void)in6_init_prefix_ltimes(pr); /* XXX error case? */ - pr->ndpr_lastupdate = time_second; + pr->ndpr_lastupdate = time_uptime; } if (new->ndpr_raf_onlink && (pr->ndpr_stateflags & NDPRF_ONLINK) == 0) { int e; if ((e = nd6_prefix_onlink(pr)) != 0) { nd6log((LOG_ERR, "prelist_update: failed to make " "the prefix %s/%d on-link on %s " "(errno=%d)\n", ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), pr->ndpr_plen, if_name(pr->ndpr_ifp), e)); /* proceed anyway. XXX: is it correct? */ } } if (dr && pfxrtr_lookup(pr, dr) == NULL) pfxrtr_add(pr, dr); } else { struct nd_prefix *newpr = NULL; newprefix = 1; if (new->ndpr_vltime == 0) goto end; if (new->ndpr_raf_onlink == 0 && new->ndpr_raf_auto == 0) goto end; error = nd6_prelist_add(new, dr, &newpr); if (error != 0 || newpr == NULL) { nd6log((LOG_NOTICE, "prelist_update: " "nd6_prelist_add failed for %s/%d on %s " "errno=%d, returnpr=%p\n", ip6_sprintf(ip6buf, &new->ndpr_prefix.sin6_addr), new->ndpr_plen, if_name(new->ndpr_ifp), error, newpr)); goto end; /* we should just give up in this case. */ } /* * XXX: from the ND point of view, we can ignore a prefix * with the on-link bit being zero. However, we need a * prefix structure for references from autoconfigured * addresses. Thus, we explicitly make sure that the prefix * itself expires now. */ if (newpr->ndpr_raf_onlink == 0) { newpr->ndpr_vltime = 0; newpr->ndpr_pltime = 0; in6_init_prefix_ltimes(newpr); } pr = newpr; } /* * Address autoconfiguration based on Section 5.5.3 of RFC 2462. * Note that pr must be non NULL at this point. */ /* 5.5.3 (a). Ignore the prefix without the A bit set. */ if (!new->ndpr_raf_auto) goto end; /* * 5.5.3 (b). the link-local prefix should have been ignored in * nd6_ra_input. */ /* 5.5.3 (c). Consistency check on lifetimes: pltime <= vltime. */ if (new->ndpr_pltime > new->ndpr_vltime) { error = EINVAL; /* XXX: won't be used */ goto end; } /* * 5.5.3 (d). If the prefix advertised is not equal to the prefix of * an address configured by stateless autoconfiguration already in the * list of addresses associated with the interface, and the Valid * Lifetime is not 0, form an address. We first check if we have * a matching prefix. * Note: we apply a clarification in rfc2462bis-02 here. We only * consider autoconfigured addresses while RFC2462 simply said * "address". */ IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { struct in6_ifaddr *ifa6; u_int32_t remaininglifetime; if (ifa->ifa_addr->sa_family != AF_INET6) continue; ifa6 = (struct in6_ifaddr *)ifa; /* * We only consider autoconfigured addresses as per rfc2462bis. */ if (!(ifa6->ia6_flags & IN6_IFF_AUTOCONF)) continue; /* * Spec is not clear here, but I believe we should concentrate * on unicast (i.e. not anycast) addresses. * XXX: other ia6_flags? detached or duplicated? */ if ((ifa6->ia6_flags & IN6_IFF_ANYCAST) != 0) continue; /* * Ignore the address if it is not associated with a prefix * or is associated with a prefix that is different from this * one. (pr is never NULL here) */ if (ifa6->ia6_ndpr != pr) continue; if (ia6_match == NULL) /* remember the first one */ ia6_match = ifa6; /* * An already autoconfigured address matched. Now that we * are sure there is at least one matched address, we can * proceed to 5.5.3. (e): update the lifetimes according to the * "two hours" rule and the privacy extension. * We apply some clarifications in rfc2462bis: * - use remaininglifetime instead of storedlifetime as a * variable name * - remove the dead code in the "two-hour" rule */ #define TWOHOUR (120*60) lt6_tmp = ifa6->ia6_lifetime; if (lt6_tmp.ia6t_vltime == ND6_INFINITE_LIFETIME) remaininglifetime = ND6_INFINITE_LIFETIME; - else if (time_second - ifa6->ia6_updatetime > + else if (time_uptime - ifa6->ia6_updatetime > lt6_tmp.ia6t_vltime) { /* * The case of "invalid" address. We should usually * not see this case. */ remaininglifetime = 0; } else remaininglifetime = lt6_tmp.ia6t_vltime - - (time_second - ifa6->ia6_updatetime); + (time_uptime - ifa6->ia6_updatetime); /* when not updating, keep the current stored lifetime. */ lt6_tmp.ia6t_vltime = remaininglifetime; if (TWOHOUR < new->ndpr_vltime || remaininglifetime < new->ndpr_vltime) { lt6_tmp.ia6t_vltime = new->ndpr_vltime; } else if (remaininglifetime <= TWOHOUR) { if (auth) { lt6_tmp.ia6t_vltime = new->ndpr_vltime; } } else { /* * new->ndpr_vltime <= TWOHOUR && * TWOHOUR < remaininglifetime */ lt6_tmp.ia6t_vltime = TWOHOUR; } /* The 2 hour rule is not imposed for preferred lifetime. */ lt6_tmp.ia6t_pltime = new->ndpr_pltime; in6_init_address_ltimes(pr, <6_tmp); /* * We need to treat lifetimes for temporary addresses * differently, according to * draft-ietf-ipv6-privacy-addrs-v2-01.txt 3.3 (1); * we only update the lifetimes when they are in the maximum * intervals. */ if ((ifa6->ia6_flags & IN6_IFF_TEMPORARY) != 0) { u_int32_t maxvltime, maxpltime; if (V_ip6_temp_valid_lifetime > - (u_int32_t)((time_second - ifa6->ia6_createtime) + + (u_int32_t)((time_uptime - ifa6->ia6_createtime) + V_ip6_desync_factor)) { maxvltime = V_ip6_temp_valid_lifetime - - (time_second - ifa6->ia6_createtime) - + (time_uptime - ifa6->ia6_createtime) - V_ip6_desync_factor; } else maxvltime = 0; if (V_ip6_temp_preferred_lifetime > - (u_int32_t)((time_second - ifa6->ia6_createtime) + + (u_int32_t)((time_uptime - ifa6->ia6_createtime) + V_ip6_desync_factor)) { maxpltime = V_ip6_temp_preferred_lifetime - - (time_second - ifa6->ia6_createtime) - + (time_uptime - ifa6->ia6_createtime) - V_ip6_desync_factor; } else maxpltime = 0; if (lt6_tmp.ia6t_vltime == ND6_INFINITE_LIFETIME || lt6_tmp.ia6t_vltime > maxvltime) { lt6_tmp.ia6t_vltime = maxvltime; } if (lt6_tmp.ia6t_pltime == ND6_INFINITE_LIFETIME || lt6_tmp.ia6t_pltime > maxpltime) { lt6_tmp.ia6t_pltime = maxpltime; } } ifa6->ia6_lifetime = lt6_tmp; - ifa6->ia6_updatetime = time_second; + ifa6->ia6_updatetime = time_uptime; } IF_ADDR_RUNLOCK(ifp); if (ia6_match == NULL && new->ndpr_vltime) { int ifidlen; /* * 5.5.3 (d) (continued) * No address matched and the valid lifetime is non-zero. * Create a new address. */ /* * Prefix Length check: * If the sum of the prefix length and interface identifier * length does not equal 128 bits, the Prefix Information * option MUST be ignored. The length of the interface * identifier is defined in a separate link-type specific * document. */ ifidlen = in6_if2idlen(ifp); if (ifidlen < 0) { /* this should not happen, so we always log it. */ log(LOG_ERR, "prelist_update: IFID undefined (%s)\n", if_name(ifp)); goto end; } if (ifidlen + pr->ndpr_plen != 128) { nd6log((LOG_INFO, "prelist_update: invalid prefixlen " "%d for %s, ignored\n", pr->ndpr_plen, if_name(ifp))); goto end; } if ((ia6 = in6_ifadd(new, mcast)) != NULL) { /* * note that we should use pr (not new) for reference. */ pr->ndpr_refcnt++; ia6->ia6_ndpr = pr; /* * RFC 3041 3.3 (2). * When a new public address is created as described * in RFC2462, also create a new temporary address. * * RFC 3041 3.5. * When an interface connects to a new link, a new * randomized interface identifier should be generated * immediately together with a new set of temporary * addresses. Thus, we specifiy 1 as the 2nd arg of * in6_tmpifadd(). */ if (V_ip6_use_tempaddr) { int e; if ((e = in6_tmpifadd(ia6, 1, 1)) != 0) { nd6log((LOG_NOTICE, "prelist_update: " "failed to create a temporary " "address, errno=%d\n", e)); } } ifa_free(&ia6->ia_ifa); /* * A newly added address might affect the status * of other addresses, so we check and update it. * XXX: what if address duplication happens? */ pfxlist_onlink_check(); } else { /* just set an error. do not bark here. */ error = EADDRNOTAVAIL; /* XXX: might be unused. */ } } end: return error; } /* * A supplement function used in the on-link detection below; * detect if a given prefix has a (probably) reachable advertising router. * XXX: lengthy function name... */ static struct nd_pfxrouter * find_pfxlist_reachable_router(struct nd_prefix *pr) { struct nd_pfxrouter *pfxrtr; struct llentry *ln; int canreach; LIST_FOREACH(pfxrtr, &pr->ndpr_advrtrs, pfr_entry) { IF_AFDATA_RLOCK(pfxrtr->router->ifp); ln = nd6_lookup(&pfxrtr->router->rtaddr, 0, pfxrtr->router->ifp); IF_AFDATA_RUNLOCK(pfxrtr->router->ifp); if (ln == NULL) continue; canreach = ND6_IS_LLINFO_PROBREACH(ln); LLE_RUNLOCK(ln); if (canreach) break; } return (pfxrtr); } /* * Check if each prefix in the prefix list has at least one available router * that advertised the prefix (a router is "available" if its neighbor cache * entry is reachable or probably reachable). * If the check fails, the prefix may be off-link, because, for example, * we have moved from the network but the lifetime of the prefix has not * expired yet. So we should not use the prefix if there is another prefix * that has an available router. * But, if there is no prefix that has an available router, we still regards * all the prefixes as on-link. This is because we can't tell if all the * routers are simply dead or if we really moved from the network and there * is no router around us. */ void pfxlist_onlink_check() { struct nd_prefix *pr; struct in6_ifaddr *ifa; struct nd_defrouter *dr; struct nd_pfxrouter *pfxrtr = NULL; /* * Check if there is a prefix that has a reachable advertising * router. */ LIST_FOREACH(pr, &V_nd_prefix, ndpr_entry) { if (pr->ndpr_raf_onlink && find_pfxlist_reachable_router(pr)) break; } /* * If we have no such prefix, check whether we still have a router * that does not advertise any prefixes. */ if (pr == NULL) { TAILQ_FOREACH(dr, &V_nd_defrouter, dr_entry) { struct nd_prefix *pr0; LIST_FOREACH(pr0, &V_nd_prefix, ndpr_entry) { if ((pfxrtr = pfxrtr_lookup(pr0, dr)) != NULL) break; } if (pfxrtr != NULL) break; } } if (pr != NULL || (!TAILQ_EMPTY(&V_nd_defrouter) && pfxrtr == NULL)) { /* * There is at least one prefix that has a reachable router, * or at least a router which probably does not advertise * any prefixes. The latter would be the case when we move * to a new link where we have a router that does not provide * prefixes and we configure an address by hand. * Detach prefixes which have no reachable advertising * router, and attach other prefixes. */ LIST_FOREACH(pr, &V_nd_prefix, ndpr_entry) { /* XXX: a link-local prefix should never be detached */ if (IN6_IS_ADDR_LINKLOCAL(&pr->ndpr_prefix.sin6_addr)) continue; /* * we aren't interested in prefixes without the L bit * set. */ if (pr->ndpr_raf_onlink == 0) continue; if (pr->ndpr_raf_auto == 0) continue; if ((pr->ndpr_stateflags & NDPRF_DETACHED) == 0 && find_pfxlist_reachable_router(pr) == NULL) pr->ndpr_stateflags |= NDPRF_DETACHED; if ((pr->ndpr_stateflags & NDPRF_DETACHED) != 0 && find_pfxlist_reachable_router(pr) != 0) pr->ndpr_stateflags &= ~NDPRF_DETACHED; } } else { /* there is no prefix that has a reachable router */ LIST_FOREACH(pr, &V_nd_prefix, ndpr_entry) { if (IN6_IS_ADDR_LINKLOCAL(&pr->ndpr_prefix.sin6_addr)) continue; if (pr->ndpr_raf_onlink == 0) continue; if (pr->ndpr_raf_auto == 0) continue; if ((pr->ndpr_stateflags & NDPRF_DETACHED) != 0) pr->ndpr_stateflags &= ~NDPRF_DETACHED; } } /* * Remove each interface route associated with a (just) detached * prefix, and reinstall the interface route for a (just) attached * prefix. Note that all attempt of reinstallation does not * necessarily success, when a same prefix is shared among multiple * interfaces. Such cases will be handled in nd6_prefix_onlink, * so we don't have to care about them. */ LIST_FOREACH(pr, &V_nd_prefix, ndpr_entry) { int e; char ip6buf[INET6_ADDRSTRLEN]; if (IN6_IS_ADDR_LINKLOCAL(&pr->ndpr_prefix.sin6_addr)) continue; if (pr->ndpr_raf_onlink == 0) continue; if (pr->ndpr_raf_auto == 0) continue; if ((pr->ndpr_stateflags & NDPRF_DETACHED) != 0 && (pr->ndpr_stateflags & NDPRF_ONLINK) != 0) { if ((e = nd6_prefix_offlink(pr)) != 0) { nd6log((LOG_ERR, "pfxlist_onlink_check: failed to " "make %s/%d offlink, errno=%d\n", ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), pr->ndpr_plen, e)); } } if ((pr->ndpr_stateflags & NDPRF_DETACHED) == 0 && (pr->ndpr_stateflags & NDPRF_ONLINK) == 0 && pr->ndpr_raf_onlink) { if ((e = nd6_prefix_onlink(pr)) != 0) { nd6log((LOG_ERR, "pfxlist_onlink_check: failed to " "make %s/%d onlink, errno=%d\n", ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), pr->ndpr_plen, e)); } } } /* * Changes on the prefix status might affect address status as well. * Make sure that all addresses derived from an attached prefix are * attached, and that all addresses derived from a detached prefix are * detached. Note, however, that a manually configured address should * always be attached. * The precise detection logic is same as the one for prefixes. * * XXXRW: in6_ifaddrhead locking. */ TAILQ_FOREACH(ifa, &V_in6_ifaddrhead, ia_link) { if (!(ifa->ia6_flags & IN6_IFF_AUTOCONF)) continue; if (ifa->ia6_ndpr == NULL) { /* * This can happen when we first configure the address * (i.e. the address exists, but the prefix does not). * XXX: complicated relationships... */ continue; } if (find_pfxlist_reachable_router(ifa->ia6_ndpr)) break; } if (ifa) { TAILQ_FOREACH(ifa, &V_in6_ifaddrhead, ia_link) { if ((ifa->ia6_flags & IN6_IFF_AUTOCONF) == 0) continue; if (ifa->ia6_ndpr == NULL) /* XXX: see above. */ continue; if (find_pfxlist_reachable_router(ifa->ia6_ndpr)) { if (ifa->ia6_flags & IN6_IFF_DETACHED) { ifa->ia6_flags &= ~IN6_IFF_DETACHED; ifa->ia6_flags |= IN6_IFF_TENTATIVE; nd6_dad_start((struct ifaddr *)ifa, 0); } } else { ifa->ia6_flags |= IN6_IFF_DETACHED; } } } else { TAILQ_FOREACH(ifa, &V_in6_ifaddrhead, ia_link) { if ((ifa->ia6_flags & IN6_IFF_AUTOCONF) == 0) continue; if (ifa->ia6_flags & IN6_IFF_DETACHED) { ifa->ia6_flags &= ~IN6_IFF_DETACHED; ifa->ia6_flags |= IN6_IFF_TENTATIVE; /* Do we need a delay in this case? */ nd6_dad_start((struct ifaddr *)ifa, 0); } } } } static int nd6_prefix_onlink_rtrequest(struct nd_prefix *pr, struct ifaddr *ifa) { static struct sockaddr_dl null_sdl = {sizeof(null_sdl), AF_LINK}; struct radix_node_head *rnh; struct rtentry *rt; struct sockaddr_in6 mask6; u_long rtflags; int error, a_failure, fibnum; /* * in6_ifinit() sets nd6_rtrequest to ifa_rtrequest for all ifaddrs. * ifa->ifa_rtrequest = nd6_rtrequest; */ bzero(&mask6, sizeof(mask6)); mask6.sin6_len = sizeof(mask6); mask6.sin6_addr = pr->ndpr_mask; rtflags = (ifa->ifa_flags & ~IFA_RTSELF) | RTF_UP; a_failure = 0; for (fibnum = 0; fibnum < rt_numfibs; fibnum++) { rt = NULL; error = in6_rtrequest(RTM_ADD, (struct sockaddr *)&pr->ndpr_prefix, ifa->ifa_addr, (struct sockaddr *)&mask6, rtflags, &rt, fibnum); if (error == 0) { KASSERT(rt != NULL, ("%s: in6_rtrequest return no " "error(%d) but rt is NULL, pr=%p, ifa=%p", __func__, error, pr, ifa)); rnh = rt_tables_get_rnh(rt->rt_fibnum, AF_INET6); /* XXX what if rhn == NULL? */ RADIX_NODE_HEAD_LOCK(rnh); RT_LOCK(rt); if (rt_setgate(rt, rt_key(rt), (struct sockaddr *)&null_sdl) == 0) { struct sockaddr_dl *dl; dl = (struct sockaddr_dl *)rt->rt_gateway; dl->sdl_type = rt->rt_ifp->if_type; dl->sdl_index = rt->rt_ifp->if_index; } RADIX_NODE_HEAD_UNLOCK(rnh); nd6_rtmsg(RTM_ADD, rt); RT_UNLOCK(rt); pr->ndpr_stateflags |= NDPRF_ONLINK; } else { char ip6buf[INET6_ADDRSTRLEN]; char ip6bufg[INET6_ADDRSTRLEN]; char ip6bufm[INET6_ADDRSTRLEN]; struct sockaddr_in6 *sin6; sin6 = (struct sockaddr_in6 *)ifa->ifa_addr; nd6log((LOG_ERR, "nd6_prefix_onlink: failed to add " "route for a prefix (%s/%d) on %s, gw=%s, mask=%s, " "flags=%lx errno = %d\n", ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), pr->ndpr_plen, if_name(pr->ndpr_ifp), ip6_sprintf(ip6bufg, &sin6->sin6_addr), ip6_sprintf(ip6bufm, &mask6.sin6_addr), rtflags, error)); /* Save last error to return, see rtinit(). */ a_failure = error; } if (rt != NULL) { RT_LOCK(rt); RT_REMREF(rt); RT_UNLOCK(rt); } } /* Return the last error we got. */ return (a_failure); } static int nd6_prefix_onlink(struct nd_prefix *pr) { struct ifaddr *ifa; struct ifnet *ifp = pr->ndpr_ifp; struct nd_prefix *opr; int error = 0; char ip6buf[INET6_ADDRSTRLEN]; /* sanity check */ if ((pr->ndpr_stateflags & NDPRF_ONLINK) != 0) { nd6log((LOG_ERR, "nd6_prefix_onlink: %s/%d is already on-link\n", ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), pr->ndpr_plen)); return (EEXIST); } /* * Add the interface route associated with the prefix. Before * installing the route, check if there's the same prefix on another * interface, and the prefix has already installed the interface route. * Although such a configuration is expected to be rare, we explicitly * allow it. */ LIST_FOREACH(opr, &V_nd_prefix, ndpr_entry) { if (opr == pr) continue; if ((opr->ndpr_stateflags & NDPRF_ONLINK) == 0) continue; if (opr->ndpr_plen == pr->ndpr_plen && in6_are_prefix_equal(&pr->ndpr_prefix.sin6_addr, &opr->ndpr_prefix.sin6_addr, pr->ndpr_plen)) return (0); } /* * We prefer link-local addresses as the associated interface address. */ /* search for a link-local addr */ ifa = (struct ifaddr *)in6ifa_ifpforlinklocal(ifp, IN6_IFF_NOTREADY | IN6_IFF_ANYCAST); if (ifa == NULL) { /* XXX: freebsd does not have ifa_ifwithaf */ IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family == AF_INET6) break; } if (ifa != NULL) ifa_ref(ifa); IF_ADDR_RUNLOCK(ifp); /* should we care about ia6_flags? */ } if (ifa == NULL) { /* * This can still happen, when, for example, we receive an RA * containing a prefix with the L bit set and the A bit clear, * after removing all IPv6 addresses on the receiving * interface. This should, of course, be rare though. */ nd6log((LOG_NOTICE, "nd6_prefix_onlink: failed to find any ifaddr" " to add route for a prefix(%s/%d) on %s\n", ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), pr->ndpr_plen, if_name(ifp))); return (0); } error = nd6_prefix_onlink_rtrequest(pr, ifa); if (ifa != NULL) ifa_free(ifa); return (error); } static int nd6_prefix_offlink(struct nd_prefix *pr) { int error = 0; struct ifnet *ifp = pr->ndpr_ifp; struct nd_prefix *opr; struct sockaddr_in6 sa6, mask6; struct rtentry *rt; char ip6buf[INET6_ADDRSTRLEN]; int fibnum, a_failure; /* sanity check */ if ((pr->ndpr_stateflags & NDPRF_ONLINK) == 0) { nd6log((LOG_ERR, "nd6_prefix_offlink: %s/%d is already off-link\n", ip6_sprintf(ip6buf, &pr->ndpr_prefix.sin6_addr), pr->ndpr_plen)); return (EEXIST); } bzero(&sa6, sizeof(sa6)); sa6.sin6_family = AF_INET6; sa6.sin6_len = sizeof(sa6); bcopy(&pr->ndpr_prefix.sin6_addr, &sa6.sin6_addr, sizeof(struct in6_addr)); bzero(&mask6, sizeof(mask6)); mask6.sin6_family = AF_INET6; mask6.sin6_len = sizeof(sa6); bcopy(&pr->ndpr_mask, &mask6.sin6_addr, sizeof(struct in6_addr)); a_failure = 0; for (fibnum = 0; fibnum < rt_numfibs; fibnum++) { rt = NULL; error = in6_rtrequest(RTM_DELETE, (struct sockaddr *)&sa6, NULL, (struct sockaddr *)&mask6, 0, &rt, fibnum); if (error == 0) { /* report the route deletion to the routing socket. */ if (rt != NULL) nd6_rtmsg(RTM_DELETE, rt); } else { /* Save last error to return, see rtinit(). */ a_failure = error; } if (rt != NULL) { RTFREE(rt); } } error = a_failure; a_failure = 1; if (error == 0) { pr->ndpr_stateflags &= ~NDPRF_ONLINK; /* * There might be the same prefix on another interface, * the prefix which could not be on-link just because we have * the interface route (see comments in nd6_prefix_onlink). * If there's one, try to make the prefix on-link on the * interface. */ LIST_FOREACH(opr, &V_nd_prefix, ndpr_entry) { if (opr == pr) continue; if ((opr->ndpr_stateflags & NDPRF_ONLINK) != 0) continue; /* * KAME specific: detached prefixes should not be * on-link. */ if ((opr->ndpr_stateflags & NDPRF_DETACHED) != 0) continue; if (opr->ndpr_plen == pr->ndpr_plen && in6_are_prefix_equal(&pr->ndpr_prefix.sin6_addr, &opr->ndpr_prefix.sin6_addr, pr->ndpr_plen)) { int e; if ((e = nd6_prefix_onlink(opr)) != 0) { nd6log((LOG_ERR, "nd6_prefix_offlink: failed to " "recover a prefix %s/%d from %s " "to %s (errno = %d)\n", ip6_sprintf(ip6buf, &opr->ndpr_prefix.sin6_addr), opr->ndpr_plen, if_name(ifp), if_name(opr->ndpr_ifp), e)); } else a_failure = 0; } } } else { /* XXX: can we still set the NDPRF_ONLINK flag? */ nd6log((LOG_ERR, "nd6_prefix_offlink: failed to delete route: " "%s/%d on %s (errno = %d)\n", ip6_sprintf(ip6buf, &sa6.sin6_addr), pr->ndpr_plen, if_name(ifp), error)); } if (a_failure) lltable_prefix_free(AF_INET6, (struct sockaddr *)&sa6, (struct sockaddr *)&mask6, LLE_STATIC); return (error); } static struct in6_ifaddr * in6_ifadd(struct nd_prefixctl *pr, int mcast) { struct ifnet *ifp = pr->ndpr_ifp; struct ifaddr *ifa; struct in6_aliasreq ifra; struct in6_ifaddr *ia, *ib; int error, plen0; struct in6_addr mask; int prefixlen = pr->ndpr_plen; int updateflags; char ip6buf[INET6_ADDRSTRLEN]; in6_prefixlen2mask(&mask, prefixlen); /* * find a link-local address (will be interface ID). * Is it really mandatory? Theoretically, a global or a site-local * address can be configured without a link-local address, if we * have a unique interface identifier... * * it is not mandatory to have a link-local address, we can generate * interface identifier on the fly. we do this because: * (1) it should be the easiest way to find interface identifier. * (2) RFC2462 5.4 suggesting the use of the same interface identifier * for multiple addresses on a single interface, and possible shortcut * of DAD. we omitted DAD for this reason in the past. * (3) a user can prevent autoconfiguration of global address * by removing link-local address by hand (this is partly because we * don't have other way to control the use of IPv6 on an interface. * this has been our design choice - cf. NRL's "ifconfig auto"). * (4) it is easier to manage when an interface has addresses * with the same interface identifier, than to have multiple addresses * with different interface identifiers. */ ifa = (struct ifaddr *)in6ifa_ifpforlinklocal(ifp, 0); /* 0 is OK? */ if (ifa) ib = (struct in6_ifaddr *)ifa; else return NULL; /* prefixlen + ifidlen must be equal to 128 */ plen0 = in6_mask2len(&ib->ia_prefixmask.sin6_addr, NULL); if (prefixlen != plen0) { ifa_free(ifa); nd6log((LOG_INFO, "in6_ifadd: wrong prefixlen for %s " "(prefix=%d ifid=%d)\n", if_name(ifp), prefixlen, 128 - plen0)); return NULL; } /* make ifaddr */ bzero(&ifra, sizeof(ifra)); /* * in6_update_ifa() does not use ifra_name, but we accurately set it * for safety. */ strncpy(ifra.ifra_name, if_name(ifp), sizeof(ifra.ifra_name)); ifra.ifra_addr.sin6_family = AF_INET6; ifra.ifra_addr.sin6_len = sizeof(struct sockaddr_in6); /* prefix */ ifra.ifra_addr.sin6_addr = pr->ndpr_prefix.sin6_addr; ifra.ifra_addr.sin6_addr.s6_addr32[0] &= mask.s6_addr32[0]; ifra.ifra_addr.sin6_addr.s6_addr32[1] &= mask.s6_addr32[1]; ifra.ifra_addr.sin6_addr.s6_addr32[2] &= mask.s6_addr32[2]; ifra.ifra_addr.sin6_addr.s6_addr32[3] &= mask.s6_addr32[3]; /* interface ID */ ifra.ifra_addr.sin6_addr.s6_addr32[0] |= (ib->ia_addr.sin6_addr.s6_addr32[0] & ~mask.s6_addr32[0]); ifra.ifra_addr.sin6_addr.s6_addr32[1] |= (ib->ia_addr.sin6_addr.s6_addr32[1] & ~mask.s6_addr32[1]); ifra.ifra_addr.sin6_addr.s6_addr32[2] |= (ib->ia_addr.sin6_addr.s6_addr32[2] & ~mask.s6_addr32[2]); ifra.ifra_addr.sin6_addr.s6_addr32[3] |= (ib->ia_addr.sin6_addr.s6_addr32[3] & ~mask.s6_addr32[3]); ifa_free(ifa); /* new prefix mask. */ ifra.ifra_prefixmask.sin6_len = sizeof(struct sockaddr_in6); ifra.ifra_prefixmask.sin6_family = AF_INET6; bcopy(&mask, &ifra.ifra_prefixmask.sin6_addr, sizeof(ifra.ifra_prefixmask.sin6_addr)); /* lifetimes. */ ifra.ifra_lifetime.ia6t_vltime = pr->ndpr_vltime; ifra.ifra_lifetime.ia6t_pltime = pr->ndpr_pltime; /* XXX: scope zone ID? */ ifra.ifra_flags |= IN6_IFF_AUTOCONF; /* obey autoconf */ /* * Make sure that we do not have this address already. This should * usually not happen, but we can still see this case, e.g., if we * have manually configured the exact address to be configured. */ ifa = (struct ifaddr *)in6ifa_ifpwithaddr(ifp, &ifra.ifra_addr.sin6_addr); if (ifa != NULL) { ifa_free(ifa); /* this should be rare enough to make an explicit log */ log(LOG_INFO, "in6_ifadd: %s is already configured\n", ip6_sprintf(ip6buf, &ifra.ifra_addr.sin6_addr)); return (NULL); } /* * Allocate ifaddr structure, link into chain, etc. * If we are going to create a new address upon receiving a multicasted * RA, we need to impose a random delay before starting DAD. * [draft-ietf-ipv6-rfc2462bis-02.txt, Section 5.4.2] */ updateflags = 0; if (mcast) updateflags |= IN6_IFAUPDATE_DADDELAY; if ((error = in6_update_ifa(ifp, &ifra, NULL, updateflags)) != 0) { nd6log((LOG_ERR, "in6_ifadd: failed to make ifaddr %s on %s (errno=%d)\n", ip6_sprintf(ip6buf, &ifra.ifra_addr.sin6_addr), if_name(ifp), error)); return (NULL); /* ifaddr must not have been allocated. */ } ia = in6ifa_ifpwithaddr(ifp, &ifra.ifra_addr.sin6_addr); /* * XXXRW: Assumption of non-NULLness here might not be true with * fine-grained locking -- should we validate it? Or just return * earlier ifa rather than looking it up again? */ return (ia); /* this is always non-NULL and referenced. */ } /* * ia0 - corresponding public address */ int in6_tmpifadd(const struct in6_ifaddr *ia0, int forcegen, int delay) { struct ifnet *ifp = ia0->ia_ifa.ifa_ifp; struct in6_ifaddr *newia, *ia; struct in6_aliasreq ifra; int i, error; int trylimit = 3; /* XXX: adhoc value */ int updateflags; u_int32_t randid[2]; time_t vltime0, pltime0; bzero(&ifra, sizeof(ifra)); strncpy(ifra.ifra_name, if_name(ifp), sizeof(ifra.ifra_name)); ifra.ifra_addr = ia0->ia_addr; /* copy prefix mask */ ifra.ifra_prefixmask = ia0->ia_prefixmask; /* clear the old IFID */ for (i = 0; i < 4; i++) { ifra.ifra_addr.sin6_addr.s6_addr32[i] &= ifra.ifra_prefixmask.sin6_addr.s6_addr32[i]; } again: if (in6_get_tmpifid(ifp, (u_int8_t *)randid, (const u_int8_t *)&ia0->ia_addr.sin6_addr.s6_addr[8], forcegen)) { nd6log((LOG_NOTICE, "in6_tmpifadd: failed to find a good " "random IFID\n")); return (EINVAL); } ifra.ifra_addr.sin6_addr.s6_addr32[2] |= (randid[0] & ~(ifra.ifra_prefixmask.sin6_addr.s6_addr32[2])); ifra.ifra_addr.sin6_addr.s6_addr32[3] |= (randid[1] & ~(ifra.ifra_prefixmask.sin6_addr.s6_addr32[3])); /* * in6_get_tmpifid() quite likely provided a unique interface ID. * However, we may still have a chance to see collision, because * there may be a time lag between generation of the ID and generation * of the address. So, we'll do one more sanity check. */ IN6_IFADDR_RLOCK(); TAILQ_FOREACH(ia, &V_in6_ifaddrhead, ia_link) { if (IN6_ARE_ADDR_EQUAL(&ia->ia_addr.sin6_addr, &ifra.ifra_addr.sin6_addr)) { if (trylimit-- == 0) { IN6_IFADDR_RUNLOCK(); /* * Give up. Something strange should have * happened. */ nd6log((LOG_NOTICE, "in6_tmpifadd: failed to " "find a unique random IFID\n")); return (EEXIST); } IN6_IFADDR_RUNLOCK(); forcegen = 1; goto again; } } IN6_IFADDR_RUNLOCK(); /* * The Valid Lifetime is the lower of the Valid Lifetime of the * public address or TEMP_VALID_LIFETIME. * The Preferred Lifetime is the lower of the Preferred Lifetime * of the public address or TEMP_PREFERRED_LIFETIME - * DESYNC_FACTOR. */ if (ia0->ia6_lifetime.ia6t_vltime != ND6_INFINITE_LIFETIME) { vltime0 = IFA6_IS_INVALID(ia0) ? 0 : (ia0->ia6_lifetime.ia6t_vltime - - (time_second - ia0->ia6_updatetime)); + (time_uptime - ia0->ia6_updatetime)); if (vltime0 > V_ip6_temp_valid_lifetime) vltime0 = V_ip6_temp_valid_lifetime; } else vltime0 = V_ip6_temp_valid_lifetime; if (ia0->ia6_lifetime.ia6t_pltime != ND6_INFINITE_LIFETIME) { pltime0 = IFA6_IS_DEPRECATED(ia0) ? 0 : (ia0->ia6_lifetime.ia6t_pltime - - (time_second - ia0->ia6_updatetime)); + (time_uptime - ia0->ia6_updatetime)); if (pltime0 > V_ip6_temp_preferred_lifetime - V_ip6_desync_factor){ pltime0 = V_ip6_temp_preferred_lifetime - V_ip6_desync_factor; } } else pltime0 = V_ip6_temp_preferred_lifetime - V_ip6_desync_factor; ifra.ifra_lifetime.ia6t_vltime = vltime0; ifra.ifra_lifetime.ia6t_pltime = pltime0; /* * A temporary address is created only if this calculated Preferred * Lifetime is greater than REGEN_ADVANCE time units. */ if (ifra.ifra_lifetime.ia6t_pltime <= V_ip6_temp_regen_advance) return (0); /* XXX: scope zone ID? */ ifra.ifra_flags |= (IN6_IFF_AUTOCONF|IN6_IFF_TEMPORARY); /* allocate ifaddr structure, link into chain, etc. */ updateflags = 0; if (delay) updateflags |= IN6_IFAUPDATE_DADDELAY; if ((error = in6_update_ifa(ifp, &ifra, NULL, updateflags)) != 0) return (error); newia = in6ifa_ifpwithaddr(ifp, &ifra.ifra_addr.sin6_addr); if (newia == NULL) { /* XXX: can it happen? */ nd6log((LOG_ERR, "in6_tmpifadd: ifa update succeeded, but we got " "no ifaddr\n")); return (EINVAL); /* XXX */ } newia->ia6_ndpr = ia0->ia6_ndpr; newia->ia6_ndpr->ndpr_refcnt++; ifa_free(&newia->ia_ifa); /* * A newly added address might affect the status of other addresses. * XXX: when the temporary address is generated with a new public * address, the onlink check is redundant. However, it would be safe * to do the check explicitly everywhere a new address is generated, * and, in fact, we surely need the check when we create a new * temporary address due to deprecation of an old temporary address. */ pfxlist_onlink_check(); return (0); } static int in6_init_prefix_ltimes(struct nd_prefix *ndpr) { if (ndpr->ndpr_pltime == ND6_INFINITE_LIFETIME) ndpr->ndpr_preferred = 0; else - ndpr->ndpr_preferred = time_second + ndpr->ndpr_pltime; + ndpr->ndpr_preferred = time_uptime + ndpr->ndpr_pltime; if (ndpr->ndpr_vltime == ND6_INFINITE_LIFETIME) ndpr->ndpr_expire = 0; else - ndpr->ndpr_expire = time_second + ndpr->ndpr_vltime; + ndpr->ndpr_expire = time_uptime + ndpr->ndpr_vltime; return 0; } static void in6_init_address_ltimes(struct nd_prefix *new, struct in6_addrlifetime *lt6) { /* init ia6t_expire */ if (lt6->ia6t_vltime == ND6_INFINITE_LIFETIME) lt6->ia6t_expire = 0; else { - lt6->ia6t_expire = time_second; + lt6->ia6t_expire = time_uptime; lt6->ia6t_expire += lt6->ia6t_vltime; } /* init ia6t_preferred */ if (lt6->ia6t_pltime == ND6_INFINITE_LIFETIME) lt6->ia6t_preferred = 0; else { - lt6->ia6t_preferred = time_second; + lt6->ia6t_preferred = time_uptime; lt6->ia6t_preferred += lt6->ia6t_pltime; } } /* * Delete all the routing table entries that use the specified gateway. * XXX: this function causes search through all entries of routing table, so * it shouldn't be called when acting as a router. */ void rt6_flush(struct in6_addr *gateway, struct ifnet *ifp) { struct radix_node_head *rnh; u_int fibnum; /* We'll care only link-local addresses */ if (!IN6_IS_ADDR_LINKLOCAL(gateway)) return; /* XXX Do we really need to walk any but the default FIB? */ for (fibnum = 0; fibnum < rt_numfibs; fibnum++) { rnh = rt_tables_get_rnh(fibnum, AF_INET6); if (rnh == NULL) continue; RADIX_NODE_HEAD_LOCK(rnh); rnh->rnh_walktree(rnh, rt6_deleteroute, (void *)gateway); RADIX_NODE_HEAD_UNLOCK(rnh); } } static int rt6_deleteroute(struct radix_node *rn, void *arg) { #define SIN6(s) ((struct sockaddr_in6 *)s) struct rtentry *rt = (struct rtentry *)rn; struct in6_addr *gate = (struct in6_addr *)arg; if (rt->rt_gateway == NULL || rt->rt_gateway->sa_family != AF_INET6) return (0); if (!IN6_ARE_ADDR_EQUAL(gate, &SIN6(rt->rt_gateway)->sin6_addr)) { return (0); } /* * Do not delete a static route. * XXX: this seems to be a bit ad-hoc. Should we consider the * 'cloned' bit instead? */ if ((rt->rt_flags & RTF_STATIC) != 0) return (0); /* * We delete only host route. This means, in particular, we don't * delete default route. */ if ((rt->rt_flags & RTF_HOST) == 0) return (0); return (in6_rtrequest(RTM_DELETE, rt_key(rt), rt->rt_gateway, rt_mask(rt), rt->rt_flags, NULL, rt->rt_fibnum)); #undef SIN6 } int nd6_setdefaultiface(int ifindex) { int error = 0; if (ifindex < 0 || V_if_index < ifindex) return (EINVAL); if (ifindex != 0 && !ifnet_byindex(ifindex)) return (EINVAL); if (V_nd6_defifindex != ifindex) { V_nd6_defifindex = ifindex; if (V_nd6_defifindex > 0) V_nd6_defifp = ifnet_byindex(V_nd6_defifindex); else V_nd6_defifp = NULL; /* * Our current implementation assumes one-to-one maping between * interfaces and links, so it would be natural to use the * default interface as the default link. */ scope6_setdefault(V_nd6_defifp); } return (error); } Index: head/sys/sys/param.h =================================================================== --- head/sys/sys/param.h (revision 253969) +++ head/sys/sys/param.h (revision 253970) @@ -1,345 +1,345 @@ /*- * Copyright (c) 1982, 1986, 1989, 1993 * The Regents of the University of California. All rights reserved. * (c) UNIX System Laboratories, Inc. * All or some portions of this file are derived from material licensed * to the University of California by American Telephone and Telegraph * Co. or Unix System Laboratories, Inc. and are reproduced herein with * the permission of UNIX System Laboratories, Inc. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)param.h 8.3 (Berkeley) 4/4/95 * $FreeBSD$ */ #ifndef _SYS_PARAM_H_ #define _SYS_PARAM_H_ #include #define BSD 199506 /* System version (year & month). */ #define BSD4_3 1 #define BSD4_4 1 /* * __FreeBSD_version numbers are documented in the Porter's Handbook. * If you bump the version for any reason, you should update the documentation * there. * Currently this lives here in the doc/ repository: * * head/en_US.ISO8859-1/books/porters-handbook/book.xml * * scheme is: Rxx * 'R' is in the range 0 to 4 if this is a release branch or * x.0-CURRENT before RELENG_*_0 is created, otherwise 'R' is * in the range 5 to 9. */ #undef __FreeBSD_version -#define __FreeBSD_version 1000040 /* Master, propagated to newvers */ +#define __FreeBSD_version 1000041 /* Master, propagated to newvers */ /* * __FreeBSD_kernel__ indicates that this system uses the kernel of FreeBSD, * which by definition is always true on FreeBSD. This macro is also defined * on other systems that use the kernel of FreeBSD, such as GNU/kFreeBSD. * * It is tempting to use this macro in userland code when we want to enable * kernel-specific routines, and in fact it's fine to do this in code that * is part of FreeBSD itself. However, be aware that as presence of this * macro is still not widespread (e.g. older FreeBSD versions, 3rd party * compilers, etc), it is STRONGLY DISCOURAGED to check for this macro in * external applications without also checking for __FreeBSD__ as an * alternative. */ #undef __FreeBSD_kernel__ #define __FreeBSD_kernel__ #ifdef _KERNEL #define P_OSREL_SIGWAIT 700000 #define P_OSREL_SIGSEGV 700004 #define P_OSREL_MAP_ANON 800104 #endif #ifndef LOCORE #include #endif /* * Machine-independent constants (some used in following include files). * Redefined constants are from POSIX 1003.1 limits file. * * MAXCOMLEN should be >= sizeof(ac_comm) (see ) */ #include #define MAXCOMLEN 19 /* max command name remembered */ #define MAXINTERP PATH_MAX /* max interpreter file name length */ #define MAXLOGNAME 33 /* max login name length (incl. NUL) */ #define MAXUPRC CHILD_MAX /* max simultaneous processes */ #define NCARGS ARG_MAX /* max bytes for an exec function */ #define NGROUPS (NGROUPS_MAX+1) /* max number groups */ #define NOFILE OPEN_MAX /* max open files per process */ #define NOGROUP 65535 /* marker for empty group set member */ #define MAXHOSTNAMELEN 256 /* max hostname size */ #define SPECNAMELEN 63 /* max length of devicename */ /* More types and definitions used throughout the kernel. */ #ifdef _KERNEL #include #include #ifndef LOCORE #include #include #endif #ifndef FALSE #define FALSE 0 #endif #ifndef TRUE #define TRUE 1 #endif #endif #ifndef _KERNEL /* Signals. */ #include #endif /* Machine type dependent parameters. */ #include #ifndef _KERNEL #include #endif #ifndef DEV_BSHIFT #define DEV_BSHIFT 9 /* log2(DEV_BSIZE) */ #endif #define DEV_BSIZE (1<>PAGE_SHIFT) #endif /* * btodb() is messy and perhaps slow because `bytes' may be an off_t. We * want to shift an unsigned type to avoid sign extension and we don't * want to widen `bytes' unnecessarily. Assume that the result fits in * a daddr_t. */ #ifndef btodb #define btodb(bytes) /* calculates (bytes / DEV_BSIZE) */ \ (sizeof (bytes) > sizeof(long) \ ? (daddr_t)((unsigned long long)(bytes) >> DEV_BSHIFT) \ : (daddr_t)((unsigned long)(bytes) >> DEV_BSHIFT)) #endif #ifndef dbtob #define dbtob(db) /* calculates (db * DEV_BSIZE) */ \ ((off_t)(db) << DEV_BSHIFT) #endif #define PRIMASK 0x0ff #define PCATCH 0x100 /* OR'd with pri for tsleep to check signals */ #define PDROP 0x200 /* OR'd with pri to stop re-entry of interlock mutex */ #define NZERO 0 /* default "nice" */ #define NBBY 8 /* number of bits in a byte */ #define NBPW sizeof(int) /* number of bytes per word (integer) */ #define CMASK 022 /* default file mask: S_IWGRP|S_IWOTH */ #define NODEV (dev_t)(-1) /* non-existent device */ /* * File system parameters and macros. * * MAXBSIZE - Filesystems are made out of blocks of at most MAXBSIZE bytes * per block. MAXBSIZE may be made larger without effecting * any existing filesystems as long as it does not exceed MAXPHYS, * and may be made smaller at the risk of not being able to use * filesystems which require a block size exceeding MAXBSIZE. * * BKVASIZE - Nominal buffer space per buffer, in bytes. BKVASIZE is the * minimum KVM memory reservation the kernel is willing to make. * Filesystems can of course request smaller chunks. Actual * backing memory uses a chunk size of a page (PAGE_SIZE). * * If you make BKVASIZE too small you risk seriously fragmenting * the buffer KVM map which may slow things down a bit. If you * make it too big the kernel will not be able to optimally use * the KVM memory reserved for the buffer cache and will wind * up with too-few buffers. * * The default is 16384, roughly 2x the block size used by a * normal UFS filesystem. */ #define MAXBSIZE 65536 /* must be power of 2 */ #define BKVASIZE 16384 /* must be power of 2 */ #define BKVAMASK (BKVASIZE-1) /* * MAXPATHLEN defines the longest permissible path length after expanding * symbolic links. It is used to allocate a temporary buffer from the buffer * pool in which to do the name expansion, hence should be a power of two, * and must be less than or equal to MAXBSIZE. MAXSYMLINKS defines the * maximum number of symbolic links that may be expanded in a path name. * It should be set high enough to allow all legitimate uses, but halt * infinite loops reasonably quickly. */ #define MAXPATHLEN PATH_MAX #define MAXSYMLINKS 32 /* Bit map related macros. */ #define setbit(a,i) (((unsigned char *)(a))[(i)/NBBY] |= 1<<((i)%NBBY)) #define clrbit(a,i) (((unsigned char *)(a))[(i)/NBBY] &= ~(1<<((i)%NBBY))) #define isset(a,i) \ (((const unsigned char *)(a))[(i)/NBBY] & (1<<((i)%NBBY))) #define isclr(a,i) \ ((((const unsigned char *)(a))[(i)/NBBY] & (1<<((i)%NBBY))) == 0) /* Macros for counting and rounding. */ #ifndef howmany #define howmany(x, y) (((x)+((y)-1))/(y)) #endif #define nitems(x) (sizeof((x)) / sizeof((x)[0])) #define rounddown(x, y) (((x)/(y))*(y)) #define rounddown2(x, y) ((x)&(~((y)-1))) /* if y is power of two */ #define roundup(x, y) ((((x)+((y)-1))/(y))*(y)) /* to any y */ #define roundup2(x, y) (((x)+((y)-1))&(~((y)-1))) /* if y is powers of two */ #define powerof2(x) ((((x)-1)&(x))==0) /* Macros for min/max. */ #define MIN(a,b) (((a)<(b))?(a):(b)) #define MAX(a,b) (((a)>(b))?(a):(b)) #ifdef _KERNEL /* * Basic byte order function prototypes for non-inline functions. */ #ifndef LOCORE #ifndef _BYTEORDER_PROTOTYPED #define _BYTEORDER_PROTOTYPED __BEGIN_DECLS __uint32_t htonl(__uint32_t); __uint16_t htons(__uint16_t); __uint32_t ntohl(__uint32_t); __uint16_t ntohs(__uint16_t); __END_DECLS #endif #endif #ifndef lint #ifndef _BYTEORDER_FUNC_DEFINED #define _BYTEORDER_FUNC_DEFINED #define htonl(x) __htonl(x) #define htons(x) __htons(x) #define ntohl(x) __ntohl(x) #define ntohs(x) __ntohs(x) #endif /* !_BYTEORDER_FUNC_DEFINED */ #endif /* lint */ #endif /* _KERNEL */ /* * Scale factor for scaled integers used to count %cpu time and load avgs. * * The number of CPU `tick's that map to a unique `%age' can be expressed * by the formula (1 / (2 ^ (FSHIFT - 11))). The maximum load average that * can be calculated (assuming 32 bits) can be closely approximated using * the formula (2 ^ (2 * (16 - FSHIFT))) for (FSHIFT < 15). * * For the scheduler to maintain a 1:1 mapping of CPU `tick' to `%age', * FSHIFT must be at least 11; this gives us a maximum load avg of ~1024. */ #define FSHIFT 11 /* bits to right of fixed binary point */ #define FSCALE (1<> (PAGE_SHIFT - DEV_BSHIFT)) #define ctodb(db) /* calculates pages to devblks */ \ ((db) << (PAGE_SHIFT - DEV_BSHIFT)) /* * Old spelling of __containerof(). */ #define member2struct(s, m, x) \ ((struct s *)(void *)((char *)(x) - offsetof(struct s, m))) /* * Access a variable length array that has been declared as a fixed * length array. */ #define __PAST_END(array, offset) (((__typeof__(*(array)) *)(array))[offset]) #endif /* _SYS_PARAM_H_ */ Index: head/usr.sbin/ndp/ndp.c =================================================================== --- head/usr.sbin/ndp/ndp.c (revision 253969) +++ head/usr.sbin/ndp/ndp.c (revision 253970) @@ -1,1591 +1,1607 @@ /* $FreeBSD$ */ /* $KAME: ndp.c,v 1.104 2003/06/27 07:48:39 itojun Exp $ */ /* * Copyright (C) 1995, 1996, 1997, 1998, and 1999 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ /* * Copyright (c) 1984, 1993 * The Regents of the University of California. All rights reserved. * * This code is derived from software contributed to Berkeley by * Sun Microsystems, Inc. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ /* * Based on: * "@(#) Copyright (c) 1984, 1993\n\ * The Regents of the University of California. All rights reserved.\n"; * * "@(#)arp.c 8.2 (Berkeley) 1/2/94"; */ /* * ndp - display, set, delete and flush neighbor cache */ #include #include #include #include #include -#include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include +#include #include #include #include "gmt2local.h" /* packing rule for routing socket */ #define ROUNDUP(a) \ ((a) > 0 ? (1 + (((a) - 1) | (sizeof(long) - 1))) : sizeof(long)) #define ADVANCE(x, n) (x += ROUNDUP((n)->sa_len)) #define NEXTADDR(w, s) \ if (rtm->rtm_addrs & (w)) { \ bcopy((char *)&s, cp, sizeof(s)); cp += SA_SIZE(&s);} static pid_t pid; static int nflag; static int tflag; static int32_t thiszone; /* time difference with gmt */ static int s = -1; static int repeat = 0; +static struct timespec ts, ts0; char ntop_buf[INET6_ADDRSTRLEN]; /* inet_ntop() */ char host_buf[NI_MAXHOST]; /* getnameinfo() */ char ifix_buf[IFNAMSIZ]; /* if_indextoname() */ int main(int, char **); int file(char *); void getsocket(void); int set(int, char **); void get(char *); int delete(char *); void dump(struct in6_addr *, int); static struct in6_nbrinfo *getnbrinfo(struct in6_addr *, int, int); static char *ether_str(struct sockaddr_dl *); int ndp_ether_aton(char *, u_char *); void usage(void); int rtmsg(int); void ifinfo(char *, int, char **); void rtrlist(void); void plist(void); void pfx_flush(void); void rtr_flush(void); void harmonize_rtr(void); #ifdef SIOCSDEFIFACE_IN6 /* XXX: check SIOCGDEFIFACE_IN6 as well? */ static void getdefif(void); static void setdefif(char *); #endif static char *sec2str(time_t); -static void ts_print(const struct timeval *); +static void ts_print(const struct timespec *); #ifdef ICMPV6CTL_ND6_DRLIST static char *rtpref_str[] = { "medium", /* 00 */ "high", /* 01 */ "rsv", /* 10 */ "low" /* 11 */ }; #endif +#define TS_SUB(tsp, usp, vsp) \ + do { \ + (vsp)->tv_sec = (tsp)->tv_sec - (usp)->tv_sec; \ + (vsp)->tv_nsec = (tsp)->tv_nsec - (usp)->tv_nsec; \ + if ((vsp)->tv_nsec < 0) { \ + (vsp)->tv_sec--; \ + (vsp)->tv_nsec += 1000000000L; \ + } \ + } while (0) + int mode = 0; char *arg = NULL; int main(argc, argv) int argc; char **argv; { + struct timespec now; int ch; pid = getpid(); thiszone = gmt2local(0); + clock_gettime(CLOCK_REALTIME_FAST, &now); + clock_gettime(CLOCK_MONOTONIC_FAST, &ts); + TS_SUB(&now, &ts, &ts0); while ((ch = getopt(argc, argv, "acd:f:Ii:nprstA:HPR")) != -1) switch (ch) { case 'a': case 'c': case 'p': case 'r': case 'H': case 'P': case 'R': case 's': case 'I': if (mode) { usage(); /*NOTREACHED*/ } mode = ch; arg = NULL; break; case 'd': case 'f': case 'i' : if (mode) { usage(); /*NOTREACHED*/ } mode = ch; arg = optarg; break; case 'n': nflag = 1; break; case 't': tflag = 1; break; case 'A': if (mode) { usage(); /*NOTREACHED*/ } mode = 'a'; repeat = atoi(optarg); if (repeat < 0) { usage(); /*NOTREACHED*/ } break; default: usage(); } argc -= optind; argv += optind; switch (mode) { case 'a': case 'c': if (argc != 0) { usage(); /*NOTREACHED*/ } dump(0, mode == 'c'); break; case 'd': if (argc != 0) { usage(); /*NOTREACHED*/ } delete(arg); break; case 'I': #ifdef SIOCSDEFIFACE_IN6 /* XXX: check SIOCGDEFIFACE_IN6 as well? */ if (argc > 1) { usage(); /*NOTREACHED*/ } else if (argc == 1) { if (strcmp(*argv, "delete") == 0 || if_nametoindex(*argv)) setdefif(*argv); else errx(1, "invalid interface %s", *argv); } getdefif(); /* always call it to print the result */ break; #else errx(1, "not supported yet"); /*NOTREACHED*/ #endif case 'p': if (argc != 0) { usage(); /*NOTREACHED*/ } plist(); break; case 'i': ifinfo(arg, argc, argv); break; case 'r': if (argc != 0) { usage(); /*NOTREACHED*/ } rtrlist(); break; case 's': if (argc < 2 || argc > 4) usage(); exit(set(argc, argv) ? 1 : 0); case 'H': if (argc != 0) { usage(); /*NOTREACHED*/ } harmonize_rtr(); break; case 'P': if (argc != 0) { usage(); /*NOTREACHED*/ } pfx_flush(); break; case 'R': if (argc != 0) { usage(); /*NOTREACHED*/ } rtr_flush(); break; case 0: if (argc != 1) { usage(); /*NOTREACHED*/ } get(argv[0]); break; } exit(0); } /* * Process a file to set standard ndp entries */ int file(name) char *name; { FILE *fp; int i, retval; char line[100], arg[5][50], *args[5]; if ((fp = fopen(name, "r")) == NULL) { fprintf(stderr, "ndp: cannot open %s\n", name); exit(1); } args[0] = &arg[0][0]; args[1] = &arg[1][0]; args[2] = &arg[2][0]; args[3] = &arg[3][0]; args[4] = &arg[4][0]; retval = 0; while (fgets(line, sizeof(line), fp) != NULL) { i = sscanf(line, "%49s %49s %49s %49s %49s", arg[0], arg[1], arg[2], arg[3], arg[4]); if (i < 2) { fprintf(stderr, "ndp: bad line: %s\n", line); retval = 1; continue; } if (set(i, args)) retval = 1; } fclose(fp); return (retval); } void getsocket() { if (s < 0) { s = socket(PF_ROUTE, SOCK_RAW, 0); if (s < 0) { err(1, "socket"); /* NOTREACHED */ } } } struct sockaddr_in6 so_mask = {sizeof(so_mask), AF_INET6 }; struct sockaddr_in6 blank_sin = {sizeof(blank_sin), AF_INET6 }, sin_m; struct sockaddr_dl blank_sdl = {sizeof(blank_sdl), AF_LINK }, sdl_m; -int expire_time, flags, found_entry; +static time_t expire_time; +static int flags, found_entry; struct { struct rt_msghdr m_rtm; char m_space[512]; } m_rtmsg; /* * Set an individual neighbor cache entry */ int set(argc, argv) int argc; char **argv; { register struct sockaddr_in6 *sin = &sin_m; register struct sockaddr_dl *sdl; register struct rt_msghdr *rtm = &(m_rtmsg.m_rtm); struct addrinfo hints, *res; int gai_error; u_char *ea; char *host = argv[0], *eaddr = argv[1]; getsocket(); argc -= 2; argv += 2; sdl_m = blank_sdl; sin_m = blank_sin; bzero(&hints, sizeof(hints)); hints.ai_family = AF_INET6; gai_error = getaddrinfo(host, NULL, &hints, &res); if (gai_error) { fprintf(stderr, "ndp: %s: %s\n", host, gai_strerror(gai_error)); return 1; } sin->sin6_addr = ((struct sockaddr_in6 *)res->ai_addr)->sin6_addr; sin->sin6_scope_id = ((struct sockaddr_in6 *)res->ai_addr)->sin6_scope_id; ea = (u_char *)LLADDR(&sdl_m); if (ndp_ether_aton(eaddr, ea) == 0) sdl_m.sdl_alen = 6; flags = expire_time = 0; while (argc-- > 0) { if (strncmp(argv[0], "temp", 4) == 0) { - struct timeval time; + struct timespec now; - gettimeofday(&time, 0); - expire_time = time.tv_sec + 20 * 60; + clock_gettime(CLOCK_MONOTONIC_FAST, &now); + expire_time = now.tv_sec + 20 * 60; } else if (strncmp(argv[0], "proxy", 5) == 0) flags |= RTF_ANNOUNCE; argv++; } if (rtmsg(RTM_GET) < 0) { errx(1, "RTM_GET(%s) failed", host); /* NOTREACHED */ } sin = (struct sockaddr_in6 *)(rtm + 1); sdl = (struct sockaddr_dl *)(ROUNDUP(sin->sin6_len) + (char *)sin); if (IN6_ARE_ADDR_EQUAL(&sin->sin6_addr, &sin_m.sin6_addr)) { if (sdl->sdl_family == AF_LINK && !(rtm->rtm_flags & RTF_GATEWAY)) { switch (sdl->sdl_type) { case IFT_ETHER: case IFT_FDDI: case IFT_ISO88023: case IFT_ISO88024: case IFT_ISO88025: case IFT_L2VLAN: case IFT_BRIDGE: goto overwrite; } } fprintf(stderr, "set: cannot configure a new entry\n"); return 1; } overwrite: if (sdl->sdl_family != AF_LINK) { printf("cannot intuit interface index and type for %s\n", host); return (1); } sdl_m.sdl_type = sdl->sdl_type; sdl_m.sdl_index = sdl->sdl_index; return (rtmsg(RTM_ADD)); } /* * Display an individual neighbor cache entry */ void get(host) char *host; { struct sockaddr_in6 *sin = &sin_m; struct addrinfo hints, *res; int gai_error; sin_m = blank_sin; bzero(&hints, sizeof(hints)); hints.ai_family = AF_INET6; gai_error = getaddrinfo(host, NULL, &hints, &res); if (gai_error) { fprintf(stderr, "ndp: %s: %s\n", host, gai_strerror(gai_error)); return; } sin->sin6_addr = ((struct sockaddr_in6 *)res->ai_addr)->sin6_addr; dump(&sin->sin6_addr, 0); if (found_entry == 0) { getnameinfo((struct sockaddr *)sin, sin->sin6_len, host_buf, sizeof(host_buf), NULL ,0, (nflag ? NI_NUMERICHOST : 0)); printf("%s (%s) -- no entry\n", host, host_buf); exit(1); } } /* * Delete a neighbor cache entry */ int delete(host) char *host; { struct sockaddr_in6 *sin = &sin_m; register struct rt_msghdr *rtm = &m_rtmsg.m_rtm; register char *cp = m_rtmsg.m_space; struct sockaddr_dl *sdl; struct addrinfo hints, *res; int gai_error; getsocket(); sin_m = blank_sin; bzero(&hints, sizeof(hints)); hints.ai_family = AF_INET6; gai_error = getaddrinfo(host, NULL, &hints, &res); if (gai_error) { fprintf(stderr, "ndp: %s: %s\n", host, gai_strerror(gai_error)); return 1; } sin->sin6_addr = ((struct sockaddr_in6 *)res->ai_addr)->sin6_addr; sin->sin6_scope_id = ((struct sockaddr_in6 *)res->ai_addr)->sin6_scope_id; if (rtmsg(RTM_GET) < 0) { errx(1, "RTM_GET(%s) failed", host); /* NOTREACHED */ } sin = (struct sockaddr_in6 *)(rtm + 1); sdl = (struct sockaddr_dl *)(ROUNDUP(sin->sin6_len) + (char *)sin); if (IN6_ARE_ADDR_EQUAL(&sin->sin6_addr, &sin_m.sin6_addr)) { if (sdl->sdl_family == AF_LINK && !(rtm->rtm_flags & RTF_GATEWAY)) { goto delete; } fprintf(stderr, "delete: cannot delete non-NDP entry\n"); return 1; } delete: if (sdl->sdl_family != AF_LINK) { printf("cannot locate %s\n", host); return (1); } /* * need to reinit the field because it has rt_key * but we want the actual address */ NEXTADDR(RTA_DST, sin_m); rtm->rtm_flags |= RTF_LLDATA; if (rtmsg(RTM_DELETE) == 0) { getnameinfo((struct sockaddr *)sin, sin->sin6_len, host_buf, sizeof(host_buf), NULL, 0, (nflag ? NI_NUMERICHOST : 0)); printf("%s (%s) deleted\n", host, host_buf); } return 0; } #define W_ADDR 36 #define W_LL 17 #define W_IF 6 /* * Dump the entire neighbor cache */ void dump(addr, cflag) struct in6_addr *addr; int cflag; { int mib[6]; size_t needed; char *lim, *buf, *next; struct rt_msghdr *rtm; struct sockaddr_in6 *sin; struct sockaddr_dl *sdl; extern int h_errno; struct in6_nbrinfo *nbi; - struct timeval time; + struct timespec now; int addrwidth; int llwidth; int ifwidth; char flgbuf[8]; char *ifname; /* Print header */ if (!tflag && !cflag) printf("%-*.*s %-*.*s %*.*s %-9.9s %1s %5s\n", W_ADDR, W_ADDR, "Neighbor", W_LL, W_LL, "Linklayer Address", W_IF, W_IF, "Netif", "Expire", "S", "Flags"); again:; mib[0] = CTL_NET; mib[1] = PF_ROUTE; mib[2] = 0; mib[3] = AF_INET6; mib[4] = NET_RT_FLAGS; #ifdef RTF_LLINFO mib[5] = RTF_LLINFO; #else mib[5] = 0; #endif if (sysctl(mib, 6, NULL, &needed, NULL, 0) < 0) err(1, "sysctl(PF_ROUTE estimate)"); if (needed > 0) { if ((buf = malloc(needed)) == NULL) err(1, "malloc"); if (sysctl(mib, 6, buf, &needed, NULL, 0) < 0) err(1, "sysctl(PF_ROUTE, NET_RT_FLAGS)"); lim = buf + needed; } else buf = lim = NULL; for (next = buf; next && next < lim; next += rtm->rtm_msglen) { int isrouter = 0, prbs = 0; rtm = (struct rt_msghdr *)next; sin = (struct sockaddr_in6 *)(rtm + 1); sdl = (struct sockaddr_dl *)((char *)sin + ROUNDUP(sin->sin6_len)); /* * Some OSes can produce a route that has the LINK flag but * has a non-AF_LINK gateway (e.g. fe80::xx%lo0 on FreeBSD * and BSD/OS, where xx is not the interface identifier on * lo0). Such routes entry would annoy getnbrinfo() below, * so we skip them. * XXX: such routes should have the GATEWAY flag, not the * LINK flag. However, there is rotten routing software * that advertises all routes that have the GATEWAY flag. * Thus, KAME kernel intentionally does not set the LINK flag. * What is to be fixed is not ndp, but such routing software * (and the kernel workaround)... */ if (sdl->sdl_family != AF_LINK) continue; if (!(rtm->rtm_flags & RTF_HOST)) continue; if (addr) { if (!IN6_ARE_ADDR_EQUAL(addr, &sin->sin6_addr)) continue; found_entry = 1; } else if (IN6_IS_ADDR_MULTICAST(&sin->sin6_addr)) continue; if (IN6_IS_ADDR_LINKLOCAL(&sin->sin6_addr) || IN6_IS_ADDR_MC_LINKLOCAL(&sin->sin6_addr)) { /* XXX: should scope id be filled in the kernel? */ if (sin->sin6_scope_id == 0) sin->sin6_scope_id = sdl->sdl_index; } getnameinfo((struct sockaddr *)sin, sin->sin6_len, host_buf, sizeof(host_buf), NULL, 0, (nflag ? NI_NUMERICHOST : 0)); if (cflag) { #ifdef RTF_WASCLONED if (rtm->rtm_flags & RTF_WASCLONED) delete(host_buf); #elif defined(RTF_CLONED) if (rtm->rtm_flags & RTF_CLONED) delete(host_buf); #else delete(host_buf); #endif continue; } - gettimeofday(&time, 0); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); if (tflag) - ts_print(&time); + ts_print(&now); addrwidth = strlen(host_buf); if (addrwidth < W_ADDR) addrwidth = W_ADDR; llwidth = strlen(ether_str(sdl)); if (W_ADDR + W_LL - addrwidth > llwidth) llwidth = W_ADDR + W_LL - addrwidth; ifname = if_indextoname(sdl->sdl_index, ifix_buf); if (!ifname) ifname = "?"; ifwidth = strlen(ifname); if (W_ADDR + W_LL + W_IF - addrwidth - llwidth > ifwidth) ifwidth = W_ADDR + W_LL + W_IF - addrwidth - llwidth; printf("%-*.*s %-*.*s %*.*s", addrwidth, addrwidth, host_buf, llwidth, llwidth, ether_str(sdl), ifwidth, ifwidth, ifname); /* Print neighbor discovery specific informations */ nbi = getnbrinfo(&sin->sin6_addr, sdl->sdl_index, 1); if (nbi) { - if (nbi->expire > time.tv_sec) { + if (nbi->expire > now.tv_sec) { printf(" %-9.9s", - sec2str(nbi->expire - time.tv_sec)); + sec2str(nbi->expire - now.tv_sec)); } else if (nbi->expire == 0) printf(" %-9.9s", "permanent"); else printf(" %-9.9s", "expired"); switch (nbi->state) { case ND6_LLINFO_NOSTATE: printf(" N"); break; #ifdef ND6_LLINFO_WAITDELETE case ND6_LLINFO_WAITDELETE: printf(" W"); break; #endif case ND6_LLINFO_INCOMPLETE: printf(" I"); break; case ND6_LLINFO_REACHABLE: printf(" R"); break; case ND6_LLINFO_STALE: printf(" S"); break; case ND6_LLINFO_DELAY: printf(" D"); break; case ND6_LLINFO_PROBE: printf(" P"); break; default: printf(" ?"); break; } isrouter = nbi->isrouter; prbs = nbi->asked; } else { warnx("failed to get neighbor information"); printf(" "); } /* * other flags. R: router, P: proxy, W: ?? */ if ((rtm->rtm_addrs & RTA_NETMASK) == 0) { snprintf(flgbuf, sizeof(flgbuf), "%s%s", isrouter ? "R" : "", (rtm->rtm_flags & RTF_ANNOUNCE) ? "p" : ""); } else { sin = (struct sockaddr_in6 *) (sdl->sdl_len + (char *)sdl); #if 0 /* W and P are mystery even for us */ snprintf(flgbuf, sizeof(flgbuf), "%s%s%s%s", isrouter ? "R" : "", !IN6_IS_ADDR_UNSPECIFIED(&sin->sin6_addr) ? "P" : "", (sin->sin6_len != sizeof(struct sockaddr_in6)) ? "W" : "", (rtm->rtm_flags & RTF_ANNOUNCE) ? "p" : ""); #else snprintf(flgbuf, sizeof(flgbuf), "%s%s", isrouter ? "R" : "", (rtm->rtm_flags & RTF_ANNOUNCE) ? "p" : ""); #endif } printf(" %s", flgbuf); if (prbs) printf(" %d", prbs); printf("\n"); } if (buf != NULL) free(buf); if (repeat) { printf("\n"); fflush(stdout); sleep(repeat); goto again; } } static struct in6_nbrinfo * getnbrinfo(addr, ifindex, warning) struct in6_addr *addr; int ifindex; int warning; { static struct in6_nbrinfo nbi; int s; if ((s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) err(1, "socket"); bzero(&nbi, sizeof(nbi)); if_indextoname(ifindex, nbi.ifname); nbi.addr = *addr; if (ioctl(s, SIOCGNBRINFO_IN6, (caddr_t)&nbi) < 0) { if (warning) warn("ioctl(SIOCGNBRINFO_IN6)"); close(s); return(NULL); } close(s); return(&nbi); } static char * ether_str(struct sockaddr_dl *sdl) { static char hbuf[NI_MAXHOST]; char *cp; if (sdl->sdl_alen == ETHER_ADDR_LEN) { strlcpy(hbuf, ether_ntoa((struct ether_addr *)LLADDR(sdl)), sizeof(hbuf)); } else if (sdl->sdl_alen) { int n = sdl->sdl_nlen > 0 ? sdl->sdl_nlen + 1 : 0; snprintf(hbuf, sizeof(hbuf), "%s", link_ntoa(sdl) + n); } else snprintf(hbuf, sizeof(hbuf), "(incomplete)"); return(hbuf); } int ndp_ether_aton(a, n) char *a; u_char *n; { int i, o[6]; i = sscanf(a, "%x:%x:%x:%x:%x:%x", &o[0], &o[1], &o[2], &o[3], &o[4], &o[5]); if (i != 6) { fprintf(stderr, "ndp: invalid Ethernet address '%s'\n", a); return (1); } for (i = 0; i < 6; i++) n[i] = o[i]; return (0); } void usage() { printf("usage: ndp [-nt] hostname\n"); printf(" ndp [-nt] -a | -c | -p | -r | -H | -P | -R\n"); printf(" ndp [-nt] -A wait\n"); printf(" ndp [-nt] -d hostname\n"); printf(" ndp [-nt] -f filename\n"); printf(" ndp [-nt] -i interface [flags...]\n"); #ifdef SIOCSDEFIFACE_IN6 printf(" ndp [-nt] -I [interface|delete]\n"); #endif printf(" ndp [-nt] -s nodename etheraddr [temp] [proxy]\n"); exit(1); } int rtmsg(cmd) int cmd; { static int seq; int rlen; register struct rt_msghdr *rtm = &m_rtmsg.m_rtm; register char *cp = m_rtmsg.m_space; register int l; errno = 0; if (cmd == RTM_DELETE) goto doit; bzero((char *)&m_rtmsg, sizeof(m_rtmsg)); rtm->rtm_flags = flags; rtm->rtm_version = RTM_VERSION; switch (cmd) { default: fprintf(stderr, "ndp: internal wrong cmd\n"); exit(1); case RTM_ADD: rtm->rtm_addrs |= RTA_GATEWAY; if (expire_time) { rtm->rtm_rmx.rmx_expire = expire_time; rtm->rtm_inits = RTV_EXPIRE; } rtm->rtm_flags |= (RTF_HOST | RTF_STATIC | RTF_LLDATA); #if 0 /* we don't support ipv6addr/128 type proxying */ if (rtm->rtm_flags & RTF_ANNOUNCE) { rtm->rtm_flags &= ~RTF_HOST; rtm->rtm_addrs |= RTA_NETMASK; } #endif /* FALLTHROUGH */ case RTM_GET: rtm->rtm_addrs |= RTA_DST; } NEXTADDR(RTA_DST, sin_m); NEXTADDR(RTA_GATEWAY, sdl_m); #if 0 /* we don't support ipv6addr/128 type proxying */ memset(&so_mask.sin6_addr, 0xff, sizeof(so_mask.sin6_addr)); NEXTADDR(RTA_NETMASK, so_mask); #endif rtm->rtm_msglen = cp - (char *)&m_rtmsg; doit: l = rtm->rtm_msglen; rtm->rtm_seq = ++seq; rtm->rtm_type = cmd; if ((rlen = write(s, (char *)&m_rtmsg, l)) < 0) { if (errno != ESRCH || cmd != RTM_DELETE) { err(1, "writing to routing socket"); /* NOTREACHED */ } } do { l = read(s, (char *)&m_rtmsg, sizeof(m_rtmsg)); } while (l > 0 && (rtm->rtm_seq != seq || rtm->rtm_pid != pid)); if (l < 0) (void) fprintf(stderr, "ndp: read from routing socket: %s\n", strerror(errno)); return (0); } void ifinfo(ifname, argc, argv) char *ifname; int argc; char **argv; { struct in6_ndireq nd; int i, s; u_int32_t newflags; #ifdef IPV6CTL_USETEMPADDR u_int8_t nullbuf[8]; #endif if ((s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) { err(1, "socket"); /* NOTREACHED */ } bzero(&nd, sizeof(nd)); strlcpy(nd.ifname, ifname, sizeof(nd.ifname)); if (ioctl(s, SIOCGIFINFO_IN6, (caddr_t)&nd) < 0) { err(1, "ioctl(SIOCGIFINFO_IN6)"); /* NOTREACHED */ } #define ND nd.ndi newflags = ND.flags; for (i = 0; i < argc; i++) { int clear = 0; char *cp = argv[i]; if (*cp == '-') { clear = 1; cp++; } #define SETFLAG(s, f) \ do {\ if (strcmp(cp, (s)) == 0) {\ if (clear)\ newflags &= ~(f);\ else\ newflags |= (f);\ }\ } while (0) /* * XXX: this macro is not 100% correct, in that it matches "nud" against * "nudbogus". But we just let it go since this is minor. */ #define SETVALUE(f, v) \ do { \ char *valptr; \ unsigned long newval; \ v = 0; /* unspecified */ \ if (strncmp(cp, f, strlen(f)) == 0) { \ valptr = strchr(cp, '='); \ if (valptr == NULL) \ err(1, "syntax error in %s field", (f)); \ errno = 0; \ newval = strtoul(++valptr, NULL, 0); \ if (errno) \ err(1, "syntax error in %s's value", (f)); \ v = newval; \ } \ } while (0) SETFLAG("disabled", ND6_IFF_IFDISABLED); SETFLAG("nud", ND6_IFF_PERFORMNUD); #ifdef ND6_IFF_ACCEPT_RTADV SETFLAG("accept_rtadv", ND6_IFF_ACCEPT_RTADV); #endif #ifdef ND6_IFF_AUTO_LINKLOCAL SETFLAG("auto_linklocal", ND6_IFF_AUTO_LINKLOCAL); #endif #ifdef ND6_IFF_NO_PREFER_IFACE SETFLAG("no_prefer_iface", ND6_IFF_NO_PREFER_IFACE); #endif SETVALUE("basereachable", ND.basereachable); SETVALUE("retrans", ND.retrans); SETVALUE("curhlim", ND.chlim); ND.flags = newflags; if (ioctl(s, SIOCSIFINFO_IN6, (caddr_t)&nd) < 0) { err(1, "ioctl(SIOCSIFINFO_IN6)"); /* NOTREACHED */ } #undef SETFLAG #undef SETVALUE } if (!ND.initialized) { errx(1, "%s: not initialized yet", ifname); /* NOTREACHED */ } if (ioctl(s, SIOCGIFINFO_IN6, (caddr_t)&nd) < 0) { err(1, "ioctl(SIOCGIFINFO_IN6)"); /* NOTREACHED */ } printf("linkmtu=%d", ND.linkmtu); printf(", maxmtu=%d", ND.maxmtu); printf(", curhlim=%d", ND.chlim); printf(", basereachable=%ds%dms", ND.basereachable / 1000, ND.basereachable % 1000); printf(", reachable=%ds", ND.reachable); printf(", retrans=%ds%dms", ND.retrans / 1000, ND.retrans % 1000); #ifdef IPV6CTL_USETEMPADDR memset(nullbuf, 0, sizeof(nullbuf)); if (memcmp(nullbuf, ND.randomid, sizeof(nullbuf)) != 0) { int j; u_int8_t *rbuf; for (i = 0; i < 3; i++) { switch (i) { case 0: printf("\nRandom seed(0): "); rbuf = ND.randomseed0; break; case 1: printf("\nRandom seed(1): "); rbuf = ND.randomseed1; break; case 2: printf("\nRandom ID: "); rbuf = ND.randomid; break; default: errx(1, "impossible case for tempaddr display"); } for (j = 0; j < 8; j++) printf("%02x", rbuf[j]); } } #endif if (ND.flags) { printf("\nFlags: "); #ifdef ND6_IFF_IFDISABLED if ((ND.flags & ND6_IFF_IFDISABLED)) printf("disabled "); #endif if ((ND.flags & ND6_IFF_PERFORMNUD)) printf("nud "); #ifdef ND6_IFF_ACCEPT_RTADV if ((ND.flags & ND6_IFF_ACCEPT_RTADV)) printf("accept_rtadv "); #endif #ifdef ND6_IFF_AUTO_LINKLOCAL if ((ND.flags & ND6_IFF_AUTO_LINKLOCAL)) printf("auto_linklocal "); #endif #ifdef ND6_IFF_NO_PREFER_IFACE if ((ND.flags & ND6_IFF_NO_PREFER_IFACE)) printf("no_prefer_iface "); #endif } putc('\n', stdout); #undef ND close(s); } #ifndef ND_RA_FLAG_RTPREF_MASK /* XXX: just for compilation on *BSD release */ #define ND_RA_FLAG_RTPREF_MASK 0x18 /* 00011000 */ #endif void rtrlist() { #ifdef ICMPV6CTL_ND6_DRLIST int mib[] = { CTL_NET, PF_INET6, IPPROTO_ICMPV6, ICMPV6CTL_ND6_DRLIST }; char *buf; struct in6_defrouter *p, *ep; size_t l; - struct timeval time; + struct timespec now; if (sysctl(mib, sizeof(mib) / sizeof(mib[0]), NULL, &l, NULL, 0) < 0) { err(1, "sysctl(ICMPV6CTL_ND6_DRLIST)"); /*NOTREACHED*/ } if (l == 0) return; buf = malloc(l); if (!buf) { err(1, "malloc"); /*NOTREACHED*/ } if (sysctl(mib, sizeof(mib) / sizeof(mib[0]), buf, &l, NULL, 0) < 0) { err(1, "sysctl(ICMPV6CTL_ND6_DRLIST)"); /*NOTREACHED*/ } ep = (struct in6_defrouter *)(buf + l); for (p = (struct in6_defrouter *)buf; p < ep; p++) { int rtpref; if (getnameinfo((struct sockaddr *)&p->rtaddr, p->rtaddr.sin6_len, host_buf, sizeof(host_buf), NULL, 0, (nflag ? NI_NUMERICHOST : 0)) != 0) strlcpy(host_buf, "?", sizeof(host_buf)); printf("%s if=%s", host_buf, if_indextoname(p->if_index, ifix_buf)); printf(", flags=%s%s", p->flags & ND_RA_FLAG_MANAGED ? "M" : "", p->flags & ND_RA_FLAG_OTHER ? "O" : ""); rtpref = ((p->flags & ND_RA_FLAG_RTPREF_MASK) >> 3) & 0xff; printf(", pref=%s", rtpref_str[rtpref]); - gettimeofday(&time, 0); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); if (p->expire == 0) printf(", expire=Never\n"); else printf(", expire=%s\n", - sec2str(p->expire - time.tv_sec)); + sec2str(p->expire - now.tv_sec)); } free(buf); #else struct in6_drlist dr; int s, i; - struct timeval time; + struct timespec now; if ((s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) { err(1, "socket"); /* NOTREACHED */ } bzero(&dr, sizeof(dr)); strlcpy(dr.ifname, "lo0", sizeof(dr.ifname)); /* dummy */ if (ioctl(s, SIOCGDRLST_IN6, (caddr_t)&dr) < 0) { err(1, "ioctl(SIOCGDRLST_IN6)"); /* NOTREACHED */ } #define DR dr.defrouter[i] for (i = 0 ; DR.if_index && i < DRLSTSIZ ; i++) { struct sockaddr_in6 sin6; bzero(&sin6, sizeof(sin6)); sin6.sin6_family = AF_INET6; sin6.sin6_len = sizeof(sin6); sin6.sin6_addr = DR.rtaddr; getnameinfo((struct sockaddr *)&sin6, sin6.sin6_len, host_buf, sizeof(host_buf), NULL, 0, (nflag ? NI_NUMERICHOST : 0)); printf("%s if=%s", host_buf, if_indextoname(DR.if_index, ifix_buf)); printf(", flags=%s%s", DR.flags & ND_RA_FLAG_MANAGED ? "M" : "", DR.flags & ND_RA_FLAG_OTHER ? "O" : ""); - gettimeofday(&time, 0); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); if (DR.expire == 0) printf(", expire=Never\n"); else printf(", expire=%s\n", - sec2str(DR.expire - time.tv_sec)); + sec2str(DR.expire - now.tv_sec)); } #undef DR close(s); #endif } void plist() { #ifdef ICMPV6CTL_ND6_PRLIST int mib[] = { CTL_NET, PF_INET6, IPPROTO_ICMPV6, ICMPV6CTL_ND6_PRLIST }; char *buf; struct in6_prefix *p, *ep, *n; struct sockaddr_in6 *advrtr; size_t l; - struct timeval time; + struct timespec now; const int niflags = NI_NUMERICHOST; int ninflags = nflag ? NI_NUMERICHOST : 0; char namebuf[NI_MAXHOST]; if (sysctl(mib, sizeof(mib) / sizeof(mib[0]), NULL, &l, NULL, 0) < 0) { err(1, "sysctl(ICMPV6CTL_ND6_PRLIST)"); /*NOTREACHED*/ } buf = malloc(l); if (!buf) { err(1, "malloc"); /*NOTREACHED*/ } if (sysctl(mib, sizeof(mib) / sizeof(mib[0]), buf, &l, NULL, 0) < 0) { err(1, "sysctl(ICMPV6CTL_ND6_PRLIST)"); /*NOTREACHED*/ } ep = (struct in6_prefix *)(buf + l); for (p = (struct in6_prefix *)buf; p < ep; p = n) { advrtr = (struct sockaddr_in6 *)(p + 1); n = (struct in6_prefix *)&advrtr[p->advrtrs]; if (getnameinfo((struct sockaddr *)&p->prefix, p->prefix.sin6_len, namebuf, sizeof(namebuf), NULL, 0, niflags) != 0) strlcpy(namebuf, "?", sizeof(namebuf)); printf("%s/%d if=%s\n", namebuf, p->prefixlen, if_indextoname(p->if_index, ifix_buf)); - gettimeofday(&time, 0); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); /* * meaning of fields, especially flags, is very different * by origin. notify the difference to the users. */ printf("flags=%s%s%s%s%s", p->raflags.onlink ? "L" : "", p->raflags.autonomous ? "A" : "", (p->flags & NDPRF_ONLINK) != 0 ? "O" : "", (p->flags & NDPRF_DETACHED) != 0 ? "D" : "", #ifdef NDPRF_HOME (p->flags & NDPRF_HOME) != 0 ? "H" : "" #else "" #endif ); if (p->vltime == ND6_INFINITE_LIFETIME) printf(" vltime=infinity"); else printf(" vltime=%lu", (unsigned long)p->vltime); if (p->pltime == ND6_INFINITE_LIFETIME) printf(", pltime=infinity"); else printf(", pltime=%lu", (unsigned long)p->pltime); if (p->expire == 0) printf(", expire=Never"); - else if (p->expire >= time.tv_sec) + else if (p->expire >= now.tv_sec) printf(", expire=%s", - sec2str(p->expire - time.tv_sec)); + sec2str(p->expire - now.tv_sec)); else printf(", expired"); printf(", ref=%d", p->refcnt); printf("\n"); /* * "advertising router" list is meaningful only if the prefix * information is from RA. */ if (p->advrtrs) { int j; struct sockaddr_in6 *sin6; sin6 = advrtr; printf(" advertised by\n"); for (j = 0; j < p->advrtrs; j++) { struct in6_nbrinfo *nbi; if (getnameinfo((struct sockaddr *)sin6, sin6->sin6_len, namebuf, sizeof(namebuf), NULL, 0, ninflags) != 0) strlcpy(namebuf, "?", sizeof(namebuf)); printf(" %s", namebuf); nbi = getnbrinfo(&sin6->sin6_addr, p->if_index, 0); if (nbi) { switch (nbi->state) { case ND6_LLINFO_REACHABLE: case ND6_LLINFO_STALE: case ND6_LLINFO_DELAY: case ND6_LLINFO_PROBE: printf(" (reachable)\n"); break; default: printf(" (unreachable)\n"); } } else printf(" (no neighbor state)\n"); sin6++; } } else printf(" No advertising router\n"); } free(buf); #else struct in6_prlist pr; int s, i; - struct timeval time; + struct timespec now; - gettimeofday(&time, 0); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); if ((s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) { err(1, "socket"); /* NOTREACHED */ } bzero(&pr, sizeof(pr)); strlcpy(pr.ifname, "lo0", sizeof(pr.ifname)); /* dummy */ if (ioctl(s, SIOCGPRLST_IN6, (caddr_t)&pr) < 0) { err(1, "ioctl(SIOCGPRLST_IN6)"); /* NOTREACHED */ } #define PR pr.prefix[i] for (i = 0; PR.if_index && i < PRLSTSIZ ; i++) { struct sockaddr_in6 p6; char namebuf[NI_MAXHOST]; int niflags; #ifdef NDPRF_ONLINK p6 = PR.prefix; #else memset(&p6, 0, sizeof(p6)); p6.sin6_family = AF_INET6; p6.sin6_len = sizeof(p6); p6.sin6_addr = PR.prefix; #endif niflags = NI_NUMERICHOST; if (getnameinfo((struct sockaddr *)&p6, sizeof(p6), namebuf, sizeof(namebuf), NULL, 0, niflags)) { warnx("getnameinfo failed"); continue; } printf("%s/%d if=%s\n", namebuf, PR.prefixlen, if_indextoname(PR.if_index, ifix_buf)); - gettimeofday(&time, 0); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); /* * meaning of fields, especially flags, is very different * by origin. notify the difference to the users. */ #if 0 printf(" %s", PR.origin == PR_ORIG_RA ? "" : "advertise: "); #endif #ifdef NDPRF_ONLINK printf("flags=%s%s%s%s%s", PR.raflags.onlink ? "L" : "", PR.raflags.autonomous ? "A" : "", (PR.flags & NDPRF_ONLINK) != 0 ? "O" : "", (PR.flags & NDPRF_DETACHED) != 0 ? "D" : "", #ifdef NDPRF_HOME (PR.flags & NDPRF_HOME) != 0 ? "H" : "" #else "" #endif ); #else printf("flags=%s%s", PR.raflags.onlink ? "L" : "", PR.raflags.autonomous ? "A" : ""); #endif if (PR.vltime == ND6_INFINITE_LIFETIME) printf(" vltime=infinity"); else printf(" vltime=%lu", PR.vltime); if (PR.pltime == ND6_INFINITE_LIFETIME) printf(", pltime=infinity"); else printf(", pltime=%lu", PR.pltime); if (PR.expire == 0) printf(", expire=Never"); - else if (PR.expire >= time.tv_sec) + else if (PR.expire >= now.tv_sec) printf(", expire=%s", - sec2str(PR.expire - time.tv_sec)); + sec2str(PR.expire - now.tv_sec)); else printf(", expired"); #ifdef NDPRF_ONLINK printf(", ref=%d", PR.refcnt); #endif #if 0 switch (PR.origin) { case PR_ORIG_RA: printf(", origin=RA"); break; case PR_ORIG_RR: printf(", origin=RR"); break; case PR_ORIG_STATIC: printf(", origin=static"); break; case PR_ORIG_KERNEL: printf(", origin=kernel"); break; default: printf(", origin=?"); break; } #endif printf("\n"); /* * "advertising router" list is meaningful only if the prefix * information is from RA. */ if (0 && /* prefix origin is almost obsolted */ PR.origin != PR_ORIG_RA) ; else if (PR.advrtrs) { int j; printf(" advertised by\n"); for (j = 0; j < PR.advrtrs; j++) { struct sockaddr_in6 sin6; struct in6_nbrinfo *nbi; bzero(&sin6, sizeof(sin6)); sin6.sin6_family = AF_INET6; sin6.sin6_len = sizeof(sin6); sin6.sin6_addr = PR.advrtr[j]; sin6.sin6_scope_id = PR.if_index; /* XXX */ getnameinfo((struct sockaddr *)&sin6, sin6.sin6_len, host_buf, sizeof(host_buf), NULL, 0, (nflag ? NI_NUMERICHOST : 0)); printf(" %s", host_buf); nbi = getnbrinfo(&sin6.sin6_addr, PR.if_index, 0); if (nbi) { switch (nbi->state) { case ND6_LLINFO_REACHABLE: case ND6_LLINFO_STALE: case ND6_LLINFO_DELAY: case ND6_LLINFO_PROBE: printf(" (reachable)\n"); break; default: printf(" (unreachable)\n"); } } else printf(" (no neighbor state)\n"); } if (PR.advrtrs > DRLSTSIZ) printf(" and %d routers\n", PR.advrtrs - DRLSTSIZ); } else printf(" No advertising router\n"); } #undef PR close(s); #endif } void pfx_flush() { char dummyif[IFNAMSIZ+8]; int s; if ((s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) err(1, "socket"); strlcpy(dummyif, "lo0", sizeof(dummyif)); /* dummy */ if (ioctl(s, SIOCSPFXFLUSH_IN6, (caddr_t)&dummyif) < 0) err(1, "ioctl(SIOCSPFXFLUSH_IN6)"); } void rtr_flush() { char dummyif[IFNAMSIZ+8]; int s; if ((s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) err(1, "socket"); strlcpy(dummyif, "lo0", sizeof(dummyif)); /* dummy */ if (ioctl(s, SIOCSRTRFLUSH_IN6, (caddr_t)&dummyif) < 0) err(1, "ioctl(SIOCSRTRFLUSH_IN6)"); close(s); } void harmonize_rtr() { char dummyif[IFNAMSIZ+8]; int s; if ((s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) err(1, "socket"); strlcpy(dummyif, "lo0", sizeof(dummyif)); /* dummy */ if (ioctl(s, SIOCSNDFLUSH_IN6, (caddr_t)&dummyif) < 0) err(1, "ioctl(SIOCSNDFLUSH_IN6)"); close(s); } #ifdef SIOCSDEFIFACE_IN6 /* XXX: check SIOCGDEFIFACE_IN6 as well? */ static void setdefif(ifname) char *ifname; { struct in6_ndifreq ndifreq; unsigned int ifindex; if (strcasecmp(ifname, "delete") == 0) ifindex = 0; else { if ((ifindex = if_nametoindex(ifname)) == 0) err(1, "failed to resolve i/f index for %s", ifname); } if ((s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) err(1, "socket"); strlcpy(ndifreq.ifname, "lo0", sizeof(ndifreq.ifname)); /* dummy */ ndifreq.ifindex = ifindex; if (ioctl(s, SIOCSDEFIFACE_IN6, (caddr_t)&ndifreq) < 0) err(1, "ioctl(SIOCSDEFIFACE_IN6)"); close(s); } static void getdefif() { struct in6_ndifreq ndifreq; char ifname[IFNAMSIZ+8]; if ((s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) err(1, "socket"); memset(&ndifreq, 0, sizeof(ndifreq)); strlcpy(ndifreq.ifname, "lo0", sizeof(ndifreq.ifname)); /* dummy */ if (ioctl(s, SIOCGDEFIFACE_IN6, (caddr_t)&ndifreq) < 0) err(1, "ioctl(SIOCGDEFIFACE_IN6)"); if (ndifreq.ifindex == 0) printf("No default interface.\n"); else { if ((if_indextoname(ndifreq.ifindex, ifname)) == NULL) err(1, "failed to resolve ifname for index %lu", ndifreq.ifindex); printf("ND default interface = %s\n", ifname); } close(s); } #endif static char * sec2str(total) time_t total; { static char result[256]; int days, hours, mins, secs; int first = 1; char *p = result; char *ep = &result[sizeof(result)]; int n; days = total / 3600 / 24; hours = (total / 3600) % 24; mins = (total / 60) % 60; secs = total % 60; if (days) { first = 0; n = snprintf(p, ep - p, "%dd", days); if (n < 0 || n >= ep - p) return "?"; p += n; } if (!first || hours) { first = 0; n = snprintf(p, ep - p, "%dh", hours); if (n < 0 || n >= ep - p) return "?"; p += n; } if (!first || mins) { first = 0; n = snprintf(p, ep - p, "%dm", mins); if (n < 0 || n >= ep - p) return "?"; p += n; } snprintf(p, ep - p, "%ds", secs); return(result); } /* * Print the timestamp * from tcpdump/util.c */ static void -ts_print(tvp) - const struct timeval *tvp; +ts_print(tsp) + const struct timespec *tsp; { int s; /* Default */ - s = (tvp->tv_sec + thiszone) % 86400; + s = (tsp->tv_sec + thiszone + ts0.tv_sec) % 86400; (void)printf("%02d:%02d:%02d.%06u ", - s / 3600, (s % 3600) / 60, s % 60, (u_int32_t)tvp->tv_usec); + s / 3600, (s % 3600) / 60, s % 60, (u_int32_t)tsp->tv_nsec / 1000); } #undef NEXTADDR Index: head/usr.sbin/rtadvctl/rtadvctl.c =================================================================== --- head/usr.sbin/rtadvctl/rtadvctl.c (revision 253969) +++ head/usr.sbin/rtadvctl/rtadvctl.c (revision 253970) @@ -1,925 +1,939 @@ /*- * Copyright (C) 2011 Hiroki Sato * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR * BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * $FreeBSD$ * */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include +#include #include #include "pathnames.h" #include "rtadvd.h" #include "if.h" #include "timer_subr.h" #include "timer.h" #include "control.h" #include "control_client.h" #define RA_IFSTATUS_INACTIVE 0 #define RA_IFSTATUS_RA_RECV 1 #define RA_IFSTATUS_RA_SEND 2 static int vflag = LOG_ERR; static void usage(void); static int action_propset(char *); static int action_propget(char *, struct ctrl_msg_pl *); static int action_plgeneric(int, char *, char *); static int action_enable(int, char **); static int action_disable(int, char **); static int action_reload(int, char **); static int action_echo(int, char **); static int action_version(int, char **); static int action_shutdown(int, char **); static int action_show(int, char **); static int action_show_prefix(struct prefix *); static int action_show_rtinfo(struct rtinfo *); static int action_show_rdnss(void *); static int action_show_dnssl(void *); static int csock_client_open(struct sockinfo *); static size_t dname_labeldec(char *, size_t, const char *); static void mysyslog(int, const char *, ...); static const char *rtpref_str[] = { "medium", /* 00 */ "high", /* 01 */ "rsv", /* 10 */ "low" /* 11 */ }; static struct dispatch_table { const char *dt_comm; int (*dt_act)(int, char **); } dtable[] = { { "show", action_show }, { "reload", action_reload }, { "shutdown", action_shutdown }, { "enable", action_enable }, { "disable", action_disable }, { NULL, NULL }, { "echo", action_echo }, { "version", action_version }, { NULL, NULL }, }; static char errmsgbuf[1024]; static char *errmsg = NULL; static void mysyslog(int priority, const char * restrict fmt, ...) { va_list ap; if (vflag >= priority) { va_start(ap, fmt); vfprintf(stderr, fmt, ap); fprintf(stderr, "\n"); va_end(ap); } } static void usage(void) { int i; for (i = 0; (size_t)i < sizeof(dtable)/sizeof(dtable[0]); i++) { if (dtable[i].dt_comm == NULL) break; printf("%s\n", dtable[i].dt_comm); } exit(1); } int main(int argc, char *argv[]) { int i; int ch; int (*action)(int, char **) = NULL; int error; while ((ch = getopt(argc, argv, "Dv")) != -1) { switch (ch) { case 'D': vflag = LOG_DEBUG; break; case 'v': vflag++; break; default: usage(); } } argc -= optind; argv += optind; if (argc == 0) usage(); for (i = 0; (size_t)i < sizeof(dtable)/sizeof(dtable[0]); i++) { if (dtable[i].dt_comm == NULL || strcmp(dtable[i].dt_comm, argv[0]) == 0) { action = dtable[i].dt_act; break; } } if (action == NULL) usage(); error = (dtable[i].dt_act)(--argc, ++argv); if (error) { fprintf(stderr, "%s failed", dtable[i].dt_comm); if (errmsg != NULL) fprintf(stderr, ": %s", errmsg); fprintf(stderr, ".\n"); } return (error); } static int csock_client_open(struct sockinfo *s) { struct sockaddr_un sun; if ((s->si_fd = socket(PF_UNIX, SOCK_STREAM, 0)) == -1) err(1, "cannot open control socket."); memset(&sun, 0, sizeof(sun)); sun.sun_family = AF_UNIX; sun.sun_len = sizeof(sun); strlcpy(sun.sun_path, s->si_name, sizeof(sun.sun_path)); if (connect(s->si_fd, (struct sockaddr *)&sun, sizeof(sun)) == -1) err(1, "connect: %s", s->si_name); mysyslog(LOG_DEBUG, "<%s> connected to %s", __func__, sun.sun_path); return (0); } static int action_plgeneric(int action, char *plstr, char *buf) { struct ctrl_msg_hdr *cm; struct ctrl_msg_pl cp; struct sockinfo *s; char *msg; char *p; char *q; s = &ctrlsock; csock_client_open(s); cm = (struct ctrl_msg_hdr *)buf; msg = (char *)buf + sizeof(*cm); cm->cm_version = CM_VERSION; cm->cm_type = action; cm->cm_len = sizeof(*cm); if (plstr != NULL) { memset(&cp, 0, sizeof(cp)); p = strchr(plstr, ':'); q = strchr(plstr, '='); if (p != NULL && q != NULL && p > q) return (1); if (p == NULL) { /* No : */ cp.cp_ifname = NULL; cp.cp_key = plstr; } else if (p == plstr) { /* empty */ cp.cp_ifname = NULL; cp.cp_key = plstr + 1; } else { *p++ = '\0'; cp.cp_ifname = plstr; cp.cp_key = p; } if (q == NULL) cp.cp_val = NULL; else { *q++ = '\0'; cp.cp_val = q; } cm->cm_len += cm_pl2bin(msg, &cp); mysyslog(LOG_DEBUG, "<%s> key=%s, val_len=%d, ifname=%s", __func__,cp.cp_key, cp.cp_val_len, cp.cp_ifname); } return (cm_handler_client(s->si_fd, CM_STATE_MSG_DISPATCH, buf)); } static int action_propget(char *argv, struct ctrl_msg_pl *cp) { int error; struct ctrl_msg_hdr *cm; char buf[CM_MSG_MAXLEN]; char *msg; memset(cp, 0, sizeof(*cp)); cm = (struct ctrl_msg_hdr *)buf; msg = (char *)buf + sizeof(*cm); error = action_plgeneric(CM_TYPE_REQ_GET_PROP, argv, buf); if (error || cm->cm_len <= sizeof(*cm)) return (1); cm_bin2pl(msg, cp); mysyslog(LOG_DEBUG, "<%s> type=%d, len=%d", __func__, cm->cm_type, cm->cm_len); mysyslog(LOG_DEBUG, "<%s> key=%s, val_len=%d, ifname=%s", __func__,cp->cp_key, cp->cp_val_len, cp->cp_ifname); return (0); } static int action_propset(char *argv) { char buf[CM_MSG_MAXLEN]; return (action_plgeneric(CM_TYPE_REQ_SET_PROP, argv, buf)); } static int action_disable(int argc, char **argv) { char *action_argv; char argv_disable[IFNAMSIZ + sizeof(":disable=")]; int i; int error; if (argc < 1) return (1); error = 0; for (i = 0; i < argc; i++) { sprintf(argv_disable, "%s:disable=", argv[i]); action_argv = argv_disable; error += action_propset(action_argv); } return (error); } static int action_enable(int argc, char **argv) { char *action_argv; char argv_enable[IFNAMSIZ + sizeof(":enable=")]; int i; int error; if (argc < 1) return (1); error = 0; for (i = 0; i < argc; i++) { sprintf(argv_enable, "%s:enable=", argv[i]); action_argv = argv_enable; error += action_propset(action_argv); } return (error); } static int action_reload(int argc, char **argv) { char *action_argv; char argv_reload[IFNAMSIZ + sizeof(":reload=")]; int i; int error; if (argc == 0) { action_argv = strdup(":reload="); return (action_propset(action_argv)); } error = 0; for (i = 0; i < argc; i++) { sprintf(argv_reload, "%s:reload=", argv[i]); action_argv = argv_reload; error += action_propset(action_argv); } return (error); } static int action_echo(int argc __unused, char **argv __unused) { char *action_argv; action_argv = strdup("echo"); return (action_propset(action_argv)); } static int action_shutdown(int argc __unused, char **argv __unused) { char *action_argv; action_argv = strdup("shutdown"); return (action_propset(action_argv)); } /* XXX */ static int action_version(int argc __unused, char **argv __unused) { char *action_argv; struct ctrl_msg_pl cp; int error; action_argv = strdup(":version="); error = action_propget(action_argv, &cp); if (error) return (error); printf("version=%s\n", cp.cp_val); return (0); } static int action_show(int argc, char **argv) { char *action_argv; char argv_ifilist[sizeof(":ifilist=")] = ":ifilist="; char argv_ifi[IFNAMSIZ + sizeof(":ifi=")]; char argv_rai[IFNAMSIZ + sizeof(":rai=")]; char argv_rti[IFNAMSIZ + sizeof(":rti=")]; char argv_pfx[IFNAMSIZ + sizeof(":pfx=")]; char argv_ifi_ra_timer[IFNAMSIZ + sizeof(":ifi_ra_timer=")]; char argv_rdnss[IFNAMSIZ + sizeof(":rdnss=")]; char argv_dnssl[IFNAMSIZ + sizeof(":dnssl=")]; char ssbuf[SSBUFLEN]; + struct timespec now, ts0, ts; struct ctrl_msg_pl cp; struct ifinfo *ifi; TAILQ_HEAD(, ifinfo) ifl = TAILQ_HEAD_INITIALIZER(ifl); char *endp; char *p; int error; int i; int len; if (argc == 0) { action_argv = argv_ifilist; error = action_propget(action_argv, &cp); if (error) return (error); p = cp.cp_val; endp = p + cp.cp_val_len; while (p < endp) { ifi = malloc(sizeof(*ifi)); if (ifi == NULL) return (1); memset(ifi, 0, sizeof(*ifi)); strcpy(ifi->ifi_ifname, p); ifi->ifi_ifindex = if_nametoindex(ifi->ifi_ifname); TAILQ_INSERT_TAIL(&ifl, ifi, ifi_next); p += strlen(ifi->ifi_ifname) + 1; } } else { for (i = 0; i < argc; i++) { ifi = malloc(sizeof(*ifi)); if (ifi == NULL) return (1); memset(ifi, 0, sizeof(*ifi)); strcpy(ifi->ifi_ifname, argv[i]); ifi->ifi_ifindex = if_nametoindex(ifi->ifi_ifname); if (ifi->ifi_ifindex == 0) { sprintf(errmsgbuf, "invalid interface %s", ifi->ifi_ifname); errmsg = errmsgbuf; return (1); } TAILQ_INSERT_TAIL(&ifl, ifi, ifi_next); } } + clock_gettime(CLOCK_REALTIME_FAST, &now); + clock_gettime(CLOCK_MONOTONIC_FAST, &ts); + TS_SUB(&now, &ts, &ts0); + TAILQ_FOREACH(ifi, &ifl, ifi_next) { struct ifinfo *ifi_s; struct rtadvd_timer *rat; struct rainfo *rai; struct rtinfo *rti; struct prefix *pfx; int c; int ra_ifstatus; sprintf(argv_ifi, "%s:ifi=", ifi->ifi_ifname); action_argv = argv_ifi; error = action_propget(action_argv, &cp); if (error) return (error); ifi_s = (struct ifinfo *)cp.cp_val; if (!(ifi_s->ifi_persist) && vflag < LOG_NOTICE) continue; printf("%s: flags=<", ifi->ifi_ifname); c = 0; if (ifi_s->ifi_ifindex == 0) c += printf("NONEXISTENT"); else c += printf("%s", (ifi_s->ifi_flags & IFF_UP) ? "UP" : "DOWN"); switch (ifi_s->ifi_state) { case IFI_STATE_CONFIGURED: c += printf("%s%s", (c) ? "," : "", "CONFIGURED"); break; case IFI_STATE_TRANSITIVE: c += printf("%s%s", (c) ? "," : "", "TRANSITIVE"); break; } if (ifi_s->ifi_persist) c += printf("%s%s", (c) ? "," : "", "PERSIST"); printf(">"); ra_ifstatus = RA_IFSTATUS_INACTIVE; if ((ifi_s->ifi_flags & IFF_UP) && ((ifi_s->ifi_state == IFI_STATE_CONFIGURED) || (ifi_s->ifi_state == IFI_STATE_TRANSITIVE))) { #if (__FreeBSD_version < 900000) /* * RA_RECV: !ip6.forwarding && ip6.accept_rtadv * RA_SEND: ip6.forwarding */ if (getinet6sysctl(IPV6CTL_FORWARDING) == 0) { if (getinet6sysctl(IPV6CTL_ACCEPT_RTADV)) ra_ifstatus = RA_IFSTATUS_RA_RECV; else ra_ifstatus = RA_IFSTATUS_INACTIVE; } else ra_ifstatus = RA_IFSTATUS_RA_SEND; #else /* * RA_RECV: ND6_IFF_ACCEPT_RTADV * RA_SEND: ip6.forwarding */ if (ifi_s->ifi_nd_flags & ND6_IFF_ACCEPT_RTADV) ra_ifstatus = RA_IFSTATUS_RA_RECV; else if (getinet6sysctl(IPV6CTL_FORWARDING)) ra_ifstatus = RA_IFSTATUS_RA_SEND; else ra_ifstatus = RA_IFSTATUS_INACTIVE; #endif } c = 0; printf(" status=<"); if (ra_ifstatus == RA_IFSTATUS_INACTIVE) printf("%s%s", (c) ? "," : "", "INACTIVE"); else if (ra_ifstatus == RA_IFSTATUS_RA_RECV) printf("%s%s", (c) ? "," : "", "RA_RECV"); else if (ra_ifstatus == RA_IFSTATUS_RA_SEND) printf("%s%s", (c) ? "," : "", "RA_SEND"); printf("> "); switch (ifi_s->ifi_state) { case IFI_STATE_CONFIGURED: case IFI_STATE_TRANSITIVE: break; default: printf("\n"); continue; } printf("mtu %d\n", ifi_s->ifi_phymtu); sprintf(argv_rai, "%s:rai=", ifi->ifi_ifname); action_argv = argv_rai; error = action_propget(action_argv, &cp); if (error) continue; rai = (struct rainfo *)cp.cp_val; printf("\tDefaultLifetime: %s", sec2str(rai->rai_lifetime, ssbuf)); if (ra_ifstatus != RA_IFSTATUS_RA_SEND && rai->rai_lifetime == 0) printf(" (RAs will be sent with zero lifetime)"); printf("\n"); printf("\tMinAdvInterval/MaxAdvInterval: "); printf("%s/", sec2str(rai->rai_mininterval, ssbuf)); printf("%s\n", sec2str(rai->rai_maxinterval, ssbuf)); if (rai->rai_linkmtu) printf("\tAdvLinkMTU: %d", rai->rai_linkmtu); else printf("\tAdvLinkMTU: "); printf(", "); printf("Flags: "); if (rai->rai_managedflg || rai->rai_otherflg) { printf("%s", rai->rai_managedflg ? "M" : ""); printf("%s", rai->rai_otherflg ? "O" : ""); } else printf(""); printf(", "); printf("Preference: %s\n", rtpref_str[(rai->rai_rtpref >> 3) & 0xff]); printf("\tReachableTime: %s, ", sec2str(rai->rai_reachabletime, ssbuf)); printf("RetransTimer: %s, " "CurHopLimit: %d\n", sec2str(rai->rai_retranstimer, ssbuf), rai->rai_hoplimit); printf("\tAdvIfPrefixes: %s\n", rai->rai_advifprefix ? "yes" : "no"); /* RA timer */ rat = NULL; if (ifi_s->ifi_ra_timer != NULL) { sprintf(argv_ifi_ra_timer, "%s:ifi_ra_timer=", ifi->ifi_ifname); action_argv = argv_ifi_ra_timer; error = action_propget(action_argv, &cp); if (error) return (error); rat = (struct rtadvd_timer *)cp.cp_val; } - printf("\tNext RA send: %s", - (rat == NULL) ? "never\n" : - ctime((time_t *)&rat->rat_tm.tv_sec)); - printf("\tLast RA sent: %s", - (ifi_s->ifi_ra_lastsent.tv_sec == 0) ? "never\n" : - ctime((time_t *)&ifi_s->ifi_ra_lastsent.tv_sec)); + printf("\tNext RA send: "); + if (rat == NULL) + printf("never\n"); + else { + ts.tv_sec = rat->rat_tm.tv_sec + ts0.tv_sec; + printf("%s", ctime(&ts.tv_sec)); + } + printf("\tLast RA send: "); + if (ifi_s->ifi_ra_lastsent.tv_sec == 0) + printf("never\n"); + else { + ts.tv_sec = ifi_s->ifi_ra_lastsent.tv_sec + ts0.tv_sec; + printf("%s", ctime(&ts.tv_sec)); + } if (rai->rai_clockskew) printf("\tClock skew: %" PRIu16 "sec\n", rai->rai_clockskew); if (vflag < LOG_WARNING) continue; /* route information */ sprintf(argv_rti, "%s:rti=", ifi->ifi_ifname); action_argv = argv_rti; error = action_propget(action_argv, &cp); if (error) return (error); rti = (struct rtinfo *)cp.cp_val; len = cp.cp_val_len / sizeof(*rti); if (len > 0) { printf("\tRoute Info:\n"); for (i = 0; i < len; i++) action_show_rtinfo(&rti[i]); } /* prefix information */ sprintf(argv_pfx, "%s:pfx=", ifi->ifi_ifname); action_argv = argv_pfx; error = action_propget(action_argv, &cp); if (error) continue; pfx = (struct prefix *)cp.cp_val; len = cp.cp_val_len / sizeof(*pfx); if (len > 0) { printf("\tPrefixes (%d):\n", len); for (i = 0; i < len; i++) action_show_prefix(&pfx[i]); } /* RDNSS information */ sprintf(argv_rdnss, "%s:rdnss=", ifi->ifi_ifname); action_argv = argv_rdnss; error = action_propget(action_argv, &cp); if (error) continue; len = *((uint16_t *)cp.cp_val); if (len > 0) { printf("\tRDNSS entries:\n"); action_show_rdnss(cp.cp_val); } /* DNSSL information */ sprintf(argv_dnssl, "%s:dnssl=", ifi->ifi_ifname); action_argv = argv_dnssl; error = action_propget(action_argv, &cp); if (error) continue; len = *((uint16_t *)cp.cp_val); if (len > 0) { printf("\tDNSSL entries:\n"); action_show_dnssl(cp.cp_val); } if (vflag < LOG_NOTICE) continue; printf("\n"); printf("\tCounters\n" "\t RA burst counts: %" PRIu16 " (interval: %s)\n" "\t RS wait counts: %" PRIu16 "\n", ifi_s->ifi_burstcount, sec2str(ifi_s->ifi_burstinterval, ssbuf), ifi_s->ifi_rs_waitcount); printf("\tOutputs\n" "\t RA: %" PRIu64 "\n", ifi_s->ifi_raoutput); printf("\tInputs\n" "\t RA: %" PRIu64 " (normal)\n" "\t RA: %" PRIu64 " (inconsistent)\n" "\t RS: %" PRIu64 "\n", ifi_s->ifi_rainput, ifi_s->ifi_rainconsistent, ifi_s->ifi_rsinput); printf("\n"); #if 0 /* Not implemented yet */ printf("\tReceived RAs:\n"); #endif } return (0); } static int action_show_rtinfo(struct rtinfo *rti) { char ntopbuf[INET6_ADDRSTRLEN]; char ssbuf[SSBUFLEN]; printf("\t %s/%d (pref: %s, ltime: %s)\n", inet_ntop(AF_INET6, &rti->rti_prefix, ntopbuf, sizeof(ntopbuf)), rti->rti_prefixlen, rtpref_str[0xff & (rti->rti_rtpref >> 3)], (rti->rti_ltime == ND6_INFINITE_LIFETIME) ? "infinity" : sec2str(rti->rti_ltime, ssbuf)); return (0); } static int action_show_prefix(struct prefix *pfx) { char ntopbuf[INET6_ADDRSTRLEN]; char ssbuf[SSBUFLEN]; - struct timeval now; + struct timespec now; - gettimeofday(&now, NULL); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); printf("\t %s/%d", inet_ntop(AF_INET6, &pfx->pfx_prefix, ntopbuf, sizeof(ntopbuf)), pfx->pfx_prefixlen); printf(" ("); switch (pfx->pfx_origin) { case PREFIX_FROM_KERNEL: printf("KERNEL"); break; case PREFIX_FROM_CONFIG: printf("CONFIG"); break; case PREFIX_FROM_DYNAMIC: printf("DYNAMIC"); break; } printf(","); printf(" vltime=%s", (pfx->pfx_validlifetime == ND6_INFINITE_LIFETIME) ? "infinity" : sec2str(pfx->pfx_validlifetime, ssbuf)); if (pfx->pfx_vltimeexpire > 0) printf("(expire: %s)", ((long)pfx->pfx_vltimeexpire > now.tv_sec) ? sec2str(pfx->pfx_vltimeexpire - now.tv_sec, ssbuf) : "0"); printf(","); printf(" pltime=%s", (pfx->pfx_preflifetime == ND6_INFINITE_LIFETIME) ? "infinity" : sec2str(pfx->pfx_preflifetime, ssbuf)); if (pfx->pfx_pltimeexpire > 0) printf("(expire %s)", ((long)pfx->pfx_pltimeexpire > now.tv_sec) ? sec2str(pfx->pfx_pltimeexpire - now.tv_sec, ssbuf) : "0"); printf(","); printf(" flags="); if (pfx->pfx_onlinkflg || pfx->pfx_autoconfflg) { printf("%s", pfx->pfx_onlinkflg ? "L" : ""); printf("%s", pfx->pfx_autoconfflg ? "A" : ""); } else printf(""); if (pfx->pfx_timer) { - struct timeval *rest; + struct timespec *rest; rest = rtadvd_timer_rest(pfx->pfx_timer); if (rest) { /* XXX: what if not? */ printf(" expire=%s", sec2str(rest->tv_sec, ssbuf)); } } printf(")\n"); return (0); } static int action_show_rdnss(void *msg) { struct rdnss *rdn; struct rdnss_addr *rda; uint16_t *rdn_cnt; uint16_t *rda_cnt; int i; int j; char *p; uint32_t ltime; char ntopbuf[INET6_ADDRSTRLEN]; char ssbuf[SSBUFLEN]; p = msg; rdn_cnt = (uint16_t *)p; p += sizeof(*rdn_cnt); if (*rdn_cnt > 0) { for (i = 0; i < *rdn_cnt; i++) { rdn = (struct rdnss *)p; ltime = rdn->rd_ltime; p += sizeof(*rdn); rda_cnt = (uint16_t *)p; p += sizeof(*rda_cnt); if (*rda_cnt > 0) for (j = 0; j < *rda_cnt; j++) { rda = (struct rdnss_addr *)p; printf("\t %s (ltime=%s)\n", inet_ntop(AF_INET6, &rda->ra_dns, ntopbuf, sizeof(ntopbuf)), sec2str(ltime, ssbuf)); p += sizeof(*rda); } } } return (0); } static int action_show_dnssl(void *msg) { struct dnssl *dns; struct dnssl_addr *dna; uint16_t *dns_cnt; uint16_t *dna_cnt; int i; int j; char *p; uint32_t ltime; char hbuf[NI_MAXHOST]; char ssbuf[SSBUFLEN]; p = msg; dns_cnt = (uint16_t *)p; p += sizeof(*dns_cnt); if (*dns_cnt > 0) { for (i = 0; i < *dns_cnt; i++) { dns = (struct dnssl *)p; ltime = dns->dn_ltime; p += sizeof(*dns); dna_cnt = (uint16_t *)p; p += sizeof(*dna_cnt); if (*dna_cnt > 0) for (j = 0; j < *dna_cnt; j++) { dna = (struct dnssl_addr *)p; dname_labeldec(hbuf, sizeof(hbuf), dna->da_dom); printf("\t %s (ltime=%s)\n", hbuf, sec2str(ltime, ssbuf)); p += sizeof(*dna); } } } return (0); } /* Decode domain name label encoding in RFC 1035 Section 3.1 */ static size_t dname_labeldec(char *dst, size_t dlen, const char *src) { size_t len; const char *src_origin; const char *src_last; const char *dst_origin; src_origin = src; src_last = strchr(src, '\0'); dst_origin = dst; memset(dst, '\0', dlen); while (src && (len = (uint8_t)(*src++) & 0x3f) && (src + len) <= src_last) { if (dst != dst_origin) *dst++ = '.'; mysyslog(LOG_DEBUG, "<%s> labellen = %zd", __func__, len); memcpy(dst, src, len); src += len; dst += len; } *dst = '\0'; return (src - src_origin); } Index: head/usr.sbin/rtadvd/config.c =================================================================== --- head/usr.sbin/rtadvd/config.c (revision 253969) +++ head/usr.sbin/rtadvd/config.c (revision 253970) @@ -1,1529 +1,1531 @@ /* $FreeBSD$ */ /* $KAME: config.c,v 1.84 2003/08/05 12:34:23 itojun Exp $ */ /* * Copyright (C) 1998 WIDE Project. * Copyright (C) 2011 Hiroki Sato * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #include #include #include -#include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include +#include #include #include #include "rtadvd.h" #include "advcap.h" #include "timer.h" #include "if.h" #include "config.h" /* label of tcapcode + number + domain name + zero octet */ static char entbuf[10 + 3 + NI_MAXHOST + 1]; static char oentbuf[10 + 3 + NI_MAXHOST + 1]; static char abuf[DNAME_LABELENC_MAXLEN]; static time_t prefix_timo = (60 * 120); /* 2 hours. * XXX: should be configurable. */ static struct rtadvd_timer *prefix_timeout(void *); static void makeentry(char *, size_t, int, const char *); static size_t dname_labelenc(char *, const char *); /* Encode domain name label encoding in RFC 1035 Section 3.1 */ static size_t dname_labelenc(char *dst, const char *src) { char *dst_origin; char *p; size_t len; dst_origin = dst; len = strlen(src); /* Length fields per 63 octets + '\0' (<= DNAME_LABELENC_MAXLEN) */ memset(dst, 0, len + len / 64 + 1 + 1); syslog(LOG_DEBUG, "<%s> labelenc = %s", __func__, src); while (src && (len = strlen(src)) != 0) { /* Put a length field with 63 octet limitation first. */ p = strchr(src, '.'); if (p == NULL) *dst++ = len = MIN(63, len); else *dst++ = len = MIN(63, p - src); /* Copy 63 octets at most. */ memcpy(dst, src, len); dst += len; if (p == NULL) /* the last label */ break; src = p + 1; } /* Always need a 0-length label at the tail. */ *dst++ = '\0'; syslog(LOG_DEBUG, "<%s> labellen = %td", __func__, dst - dst_origin); return (dst - dst_origin); } #define MUSTHAVE(var, cap) \ do { \ int64_t t; \ if ((t = agetnum(cap)) < 0) { \ fprintf(stderr, "rtadvd: need %s for interface %s\n", \ cap, intface); \ exit(1); \ } \ var = t; \ } while (0) #define MAYHAVE(var, cap, def) \ do { \ if ((var = agetnum(cap)) < 0) \ var = def; \ } while (0) int loadconfig_index(int idx) { char ifname[IFNAMSIZ]; syslog(LOG_DEBUG, "<%s> enter", __func__); if (if_indextoname(idx, ifname) != NULL) return (loadconfig_ifname(ifname)); else return (1); } int loadconfig_ifname(char *ifname) { struct ifinfo *ifi; syslog(LOG_DEBUG, "<%s> enter", __func__); update_ifinfo(&ifilist, UPDATE_IFINFO_ALL); TAILQ_FOREACH(ifi, &ifilist, ifi_next) { /* NULL means all IFs will be processed. */ if (ifname != NULL && strcmp(ifi->ifi_ifname, ifname) != 0) continue; if (!ifi->ifi_persist) { syslog(LOG_INFO, "<%s> %s is not a target interface. " "Ignored at this moment.", __func__, ifi->ifi_ifname); continue; } if (ifi->ifi_ifindex == 0) { syslog(LOG_ERR, "<%s> %s not found. " "Ignored at this moment.", __func__, ifi->ifi_ifname); continue; } if (getconfig(ifi) == NULL) { syslog(LOG_ERR, "<%s> invalid configuration for %s. " "Ignored at this moment.", __func__, ifi->ifi_ifname); continue; } } return (0); } int rm_ifinfo_index(int idx) { struct ifinfo *ifi; ifi = if_indextoifinfo(idx); if (ifi == NULL) { syslog(LOG_ERR, "<%s>: ifinfo not found (idx=%d)", __func__, idx); return (-1); } return (rm_ifinfo(ifi)); } int rm_ifinfo(struct ifinfo *ifi) { int error; syslog(LOG_DEBUG, "<%s> enter (%s).", __func__, ifi->ifi_ifname); switch (ifi->ifi_state) { case IFI_STATE_UNCONFIGURED: return (0); break; default: ifi->ifi_state = IFI_STATE_UNCONFIGURED; syslog(LOG_DEBUG, "<%s> ifname=%s marked as UNCONFIGURED.", __func__, ifi->ifi_ifname); /* XXX: No MC leaving here because index is disappeared */ /* Inactivate timer */ rtadvd_remove_timer(ifi->ifi_ra_timer); ifi->ifi_ra_timer = NULL; break; } /* clean up ifi */ if (!ifi->ifi_persist) { TAILQ_REMOVE(&ifilist, ifi, ifi_next); syslog(LOG_DEBUG, "<%s>: ifinfo (idx=%d) removed.", __func__, ifi->ifi_ifindex); free(ifi); } else { /* recreate an empty entry */ update_persist_ifinfo(&ifilist, ifi->ifi_ifname); syslog(LOG_DEBUG, "<%s>: ifname=%s is persistent.", __func__, ifi->ifi_ifname); } /* clean up rai if any */ switch (ifi->ifi_state) { case IFI_STATE_CONFIGURED: if (ifi->ifi_rainfo != NULL) { error = rm_rainfo(ifi->ifi_rainfo); if (error) return (error); ifi->ifi_rainfo = NULL; } break; case IFI_STATE_TRANSITIVE: if (ifi->ifi_rainfo == ifi->ifi_rainfo_trans) { if (ifi->ifi_rainfo != NULL) { error = rm_rainfo(ifi->ifi_rainfo); if (error) return (error); ifi->ifi_rainfo = NULL; ifi->ifi_rainfo_trans = NULL; } } else { if (ifi->ifi_rainfo != NULL) { error = rm_rainfo(ifi->ifi_rainfo); if (error) return (error); ifi->ifi_rainfo = NULL; } if (ifi->ifi_rainfo_trans != NULL) { error = rm_rainfo(ifi->ifi_rainfo_trans); if (error) return (error); ifi->ifi_rainfo_trans = NULL; } } } syslog(LOG_DEBUG, "<%s> leave (%s).", __func__, ifi->ifi_ifname); return (0); } int rm_rainfo(struct rainfo *rai) { struct prefix *pfx; struct soliciter *sol; struct rdnss *rdn; struct rdnss_addr *rdna; struct dnssl *dns; struct rtinfo *rti; syslog(LOG_DEBUG, "<%s>: enter", __func__); TAILQ_REMOVE(&railist, rai, rai_next); if (rai->rai_ifinfo != NULL) syslog(LOG_DEBUG, "<%s>: rainfo (idx=%d) removed.", __func__, rai->rai_ifinfo->ifi_ifindex); if (rai->rai_ra_data != NULL) free(rai->rai_ra_data); while ((pfx = TAILQ_FIRST(&rai->rai_prefix)) != NULL) { TAILQ_REMOVE(&rai->rai_prefix, pfx, pfx_next); free(pfx); } while ((sol = TAILQ_FIRST(&rai->rai_soliciter)) != NULL) { TAILQ_REMOVE(&rai->rai_soliciter, sol, sol_next); free(sol); } while ((rdn = TAILQ_FIRST(&rai->rai_rdnss)) != NULL) { TAILQ_REMOVE(&rai->rai_rdnss, rdn, rd_next); while ((rdna = TAILQ_FIRST(&rdn->rd_list)) != NULL) { TAILQ_REMOVE(&rdn->rd_list, rdna, ra_next); free(rdna); } free(rdn); } while ((dns = TAILQ_FIRST(&rai->rai_dnssl)) != NULL) { TAILQ_REMOVE(&rai->rai_dnssl, dns, dn_next); free(dns); } while ((rti = TAILQ_FIRST(&rai->rai_route)) != NULL) { TAILQ_REMOVE(&rai->rai_route, rti, rti_next); free(rti); } free(rai); syslog(LOG_DEBUG, "<%s>: leave", __func__); return (0); } struct ifinfo * getconfig(struct ifinfo *ifi) { int stat, i; int error; char tbuf[BUFSIZ]; struct rainfo *rai; struct rainfo *rai_old; int32_t val; int64_t val64; char buf[BUFSIZ]; char *bp = buf; char *addr, *flagstr; if (ifi == NULL) /* if does not exist */ return (NULL); if (ifi->ifi_state == IFI_STATE_TRANSITIVE && ifi->ifi_rainfo == NULL) { syslog(LOG_INFO, "<%s> %s is shutting down. Skipped.", __func__, ifi->ifi_ifname); return (NULL); } if ((stat = agetent(tbuf, ifi->ifi_ifname)) <= 0) { memset(tbuf, 0, sizeof(tbuf)); syslog(LOG_INFO, "<%s> %s isn't defined in the configuration file" " or the configuration file doesn't exist." " Treat it as default", __func__, ifi->ifi_ifname); } ELM_MALLOC(rai, exit(1)); TAILQ_INIT(&rai->rai_prefix); TAILQ_INIT(&rai->rai_route); TAILQ_INIT(&rai->rai_rdnss); TAILQ_INIT(&rai->rai_dnssl); TAILQ_INIT(&rai->rai_soliciter); rai->rai_ifinfo = ifi; /* gather on-link prefixes from the network interfaces. */ if (agetflag("noifprefix")) rai->rai_advifprefix = 0; else rai->rai_advifprefix = 1; /* get interface information */ if (agetflag("nolladdr")) rai->rai_advlinkopt = 0; else rai->rai_advlinkopt = 1; if (rai->rai_advlinkopt) { if (ifi->ifi_sdl.sdl_type == 0) { syslog(LOG_ERR, "<%s> can't get information of %s", __func__, ifi->ifi_ifname); goto getconfig_free_rai; } } /* * set router configuration variables. */ MAYHAVE(val, "maxinterval", DEF_MAXRTRADVINTERVAL); if (val < MIN_MAXINTERVAL || val > MAX_MAXINTERVAL) { syslog(LOG_ERR, "<%s> maxinterval (%" PRIu32 ") on %s is invalid " "(must be between %u and %u)", __func__, val, ifi->ifi_ifname, MIN_MAXINTERVAL, MAX_MAXINTERVAL); goto getconfig_free_rai; } rai->rai_maxinterval = (uint16_t)val; MAYHAVE(val, "mininterval", rai->rai_maxinterval/3); if ((uint16_t)val < MIN_MININTERVAL || (uint16_t)val > (rai->rai_maxinterval * 3) / 4) { syslog(LOG_ERR, "<%s> mininterval (%" PRIu32 ") on %s is invalid " "(must be between %d and %d)", __func__, val, ifi->ifi_ifname, MIN_MININTERVAL, (rai->rai_maxinterval * 3) / 4); goto getconfig_free_rai; } rai->rai_mininterval = (uint16_t)val; MAYHAVE(val, "chlim", DEF_ADVCURHOPLIMIT); rai->rai_hoplimit = val & 0xff; if ((flagstr = (char *)agetstr("raflags", &bp))) { val = 0; if (strchr(flagstr, 'm')) val |= ND_RA_FLAG_MANAGED; if (strchr(flagstr, 'o')) val |= ND_RA_FLAG_OTHER; if (strchr(flagstr, 'h')) val |= ND_RA_FLAG_RTPREF_HIGH; if (strchr(flagstr, 'l')) { if ((val & ND_RA_FLAG_RTPREF_HIGH)) { syslog(LOG_ERR, "<%s> the \'h\' and \'l\'" " router flags are exclusive", __func__); goto getconfig_free_rai; } val |= ND_RA_FLAG_RTPREF_LOW; } } else MAYHAVE(val, "raflags", 0); rai->rai_managedflg = val & ND_RA_FLAG_MANAGED; rai->rai_otherflg = val & ND_RA_FLAG_OTHER; #ifndef ND_RA_FLAG_RTPREF_MASK #define ND_RA_FLAG_RTPREF_MASK 0x18 /* 00011000 */ #define ND_RA_FLAG_RTPREF_RSV 0x10 /* 00010000 */ #endif rai->rai_rtpref = val & ND_RA_FLAG_RTPREF_MASK; if (rai->rai_rtpref == ND_RA_FLAG_RTPREF_RSV) { syslog(LOG_ERR, "<%s> invalid router preference (%02x) on %s", __func__, rai->rai_rtpref, ifi->ifi_ifname); goto getconfig_free_rai; } MAYHAVE(val, "rltime", rai->rai_maxinterval * 3); if ((uint16_t)val && ((uint16_t)val < rai->rai_maxinterval || (uint16_t)val > MAXROUTERLIFETIME)) { syslog(LOG_ERR, "<%s> router lifetime (%" PRIu32 ") on %s is invalid " "(must be 0 or between %d and %d)", __func__, val, ifi->ifi_ifname, rai->rai_maxinterval, MAXROUTERLIFETIME); goto getconfig_free_rai; } rai->rai_lifetime = val & 0xffff; MAYHAVE(val, "rtime", DEF_ADVREACHABLETIME); if (val < 0 || val > MAXREACHABLETIME) { syslog(LOG_ERR, "<%s> reachable time (%" PRIu32 ") on %s is invalid " "(must be no greater than %d)", __func__, val, ifi->ifi_ifname, MAXREACHABLETIME); goto getconfig_free_rai; } rai->rai_reachabletime = (uint32_t)val; MAYHAVE(val64, "retrans", DEF_ADVRETRANSTIMER); if (val64 < 0 || val64 > 0xffffffff) { syslog(LOG_ERR, "<%s> retrans time (%" PRIu64 ") on %s out of range", __func__, val64, ifi->ifi_ifname); goto getconfig_free_rai; } rai->rai_retranstimer = (uint32_t)val64; if (agetnum("hapref") != -1 || agetnum("hatime") != -1) { syslog(LOG_ERR, "<%s> mobile-ip6 configuration not supported", __func__); goto getconfig_free_rai; } /* prefix information */ /* * This is an implementation specific parameter to consider * link propagation delays and poorly synchronized clocks when * checking consistency of advertised lifetimes. */ MAYHAVE(val, "clockskew", 0); rai->rai_clockskew = val; rai->rai_pfxs = 0; for (i = -1; i < MAXPREFIX; i++) { struct prefix *pfx; makeentry(entbuf, sizeof(entbuf), i, "addr"); addr = (char *)agetstr(entbuf, &bp); if (addr == NULL) continue; /* allocate memory to store prefix information */ ELM_MALLOC(pfx, exit(1)); pfx->pfx_rainfo = rai; pfx->pfx_origin = PREFIX_FROM_CONFIG; if (inet_pton(AF_INET6, addr, &pfx->pfx_prefix) != 1) { syslog(LOG_ERR, "<%s> inet_pton failed for %s", __func__, addr); goto getconfig_free_pfx; } if (IN6_IS_ADDR_MULTICAST(&pfx->pfx_prefix)) { syslog(LOG_ERR, "<%s> multicast prefix (%s) must " "not be advertised on %s", __func__, addr, ifi->ifi_ifname); goto getconfig_free_pfx; } if (IN6_IS_ADDR_LINKLOCAL(&pfx->pfx_prefix)) syslog(LOG_NOTICE, "<%s> link-local prefix (%s) will be" " advertised on %s", __func__, addr, ifi->ifi_ifname); makeentry(entbuf, sizeof(entbuf), i, "prefixlen"); MAYHAVE(val, entbuf, 64); if (val < 0 || val > 128) { syslog(LOG_ERR, "<%s> prefixlen (%" PRIu32 ") for %s " "on %s out of range", __func__, val, addr, ifi->ifi_ifname); goto getconfig_free_pfx; } pfx->pfx_prefixlen = (int)val; makeentry(entbuf, sizeof(entbuf), i, "pinfoflags"); if ((flagstr = (char *)agetstr(entbuf, &bp))) { val = 0; if (strchr(flagstr, 'l')) val |= ND_OPT_PI_FLAG_ONLINK; if (strchr(flagstr, 'a')) val |= ND_OPT_PI_FLAG_AUTO; } else { MAYHAVE(val, entbuf, (ND_OPT_PI_FLAG_ONLINK|ND_OPT_PI_FLAG_AUTO)); } pfx->pfx_onlinkflg = val & ND_OPT_PI_FLAG_ONLINK; pfx->pfx_autoconfflg = val & ND_OPT_PI_FLAG_AUTO; makeentry(entbuf, sizeof(entbuf), i, "vltime"); MAYHAVE(val64, entbuf, DEF_ADVVALIDLIFETIME); if (val64 < 0 || val64 > 0xffffffff) { syslog(LOG_ERR, "<%s> vltime (%" PRIu64 ") for " "%s/%d on %s is out of range", __func__, val64, addr, pfx->pfx_prefixlen, ifi->ifi_ifname); goto getconfig_free_pfx; } pfx->pfx_validlifetime = (uint32_t)val64; makeentry(entbuf, sizeof(entbuf), i, "vltimedecr"); if (agetflag(entbuf)) { - struct timeval now; - gettimeofday(&now, 0); + struct timespec now; + + clock_gettime(CLOCK_MONOTONIC_FAST, &now); pfx->pfx_vltimeexpire = now.tv_sec + pfx->pfx_validlifetime; } makeentry(entbuf, sizeof(entbuf), i, "pltime"); MAYHAVE(val64, entbuf, DEF_ADVPREFERREDLIFETIME); if (val64 < 0 || val64 > 0xffffffff) { syslog(LOG_ERR, "<%s> pltime (%" PRIu64 ") for %s/%d on %s " "is out of range", __func__, val64, addr, pfx->pfx_prefixlen, ifi->ifi_ifname); goto getconfig_free_pfx; } pfx->pfx_preflifetime = (uint32_t)val64; makeentry(entbuf, sizeof(entbuf), i, "pltimedecr"); if (agetflag(entbuf)) { - struct timeval now; - gettimeofday(&now, 0); + struct timespec now; + + clock_gettime(CLOCK_MONOTONIC_FAST, &now); pfx->pfx_pltimeexpire = now.tv_sec + pfx->pfx_preflifetime; } /* link into chain */ TAILQ_INSERT_TAIL(&rai->rai_prefix, pfx, pfx_next); rai->rai_pfxs++; continue; getconfig_free_pfx: free(pfx); } if (rai->rai_advifprefix && rai->rai_pfxs == 0) get_prefix(rai); MAYHAVE(val64, "mtu", 0); if (val < 0 || val64 > 0xffffffff) { syslog(LOG_ERR, "<%s> mtu (%" PRIu64 ") on %s out of range", __func__, val64, ifi->ifi_ifname); goto getconfig_free_rai; } rai->rai_linkmtu = (uint32_t)val64; if (rai->rai_linkmtu == 0) { char *mtustr; if ((mtustr = (char *)agetstr("mtu", &bp)) && strcmp(mtustr, "auto") == 0) rai->rai_linkmtu = ifi->ifi_phymtu; } else if (rai->rai_linkmtu < IPV6_MMTU || rai->rai_linkmtu > ifi->ifi_phymtu) { syslog(LOG_ERR, "<%s> advertised link mtu (%" PRIu32 ") on %s is invalid (must " "be between least MTU (%d) and physical link MTU (%d)", __func__, rai->rai_linkmtu, ifi->ifi_ifname, IPV6_MMTU, ifi->ifi_phymtu); goto getconfig_free_rai; } #ifdef SIOCSIFINFO_IN6 { struct in6_ndireq ndi; int s; if ((s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) { syslog(LOG_ERR, "<%s> socket: %s", __func__, strerror(errno)); exit(1); } memset(&ndi, 0, sizeof(ndi)); strncpy(ndi.ifname, ifi->ifi_ifname, sizeof(ndi.ifname)); if (ioctl(s, SIOCGIFINFO_IN6, (caddr_t)&ndi) < 0) syslog(LOG_INFO, "<%s> ioctl:SIOCGIFINFO_IN6 at %s: %s", __func__, ifi->ifi_ifname, strerror(errno)); /* reflect the RA info to the host variables in kernel */ ndi.ndi.chlim = rai->rai_hoplimit; ndi.ndi.retrans = rai->rai_retranstimer; ndi.ndi.basereachable = rai->rai_reachabletime; if (ioctl(s, SIOCSIFINFO_IN6, (caddr_t)&ndi) < 0) syslog(LOG_INFO, "<%s> ioctl:SIOCSIFINFO_IN6 at %s: %s", __func__, ifi->ifi_ifname, strerror(errno)); close(s); } #endif /* route information */ rai->rai_routes = 0; for (i = -1; i < MAXROUTE; i++) { struct rtinfo *rti; makeentry(entbuf, sizeof(entbuf), i, "rtprefix"); addr = (char *)agetstr(entbuf, &bp); if (addr == NULL) { makeentry(oentbuf, sizeof(oentbuf), i, "rtrprefix"); addr = (char *)agetstr(oentbuf, &bp); if (addr) fprintf(stderr, "%s was obsoleted. Use %s.\n", oentbuf, entbuf); } if (addr == NULL) continue; /* allocate memory to store prefix information */ ELM_MALLOC(rti, exit(1)); if (inet_pton(AF_INET6, addr, &rti->rti_prefix) != 1) { syslog(LOG_ERR, "<%s> inet_pton failed for %s", __func__, addr); goto getconfig_free_rti; } #if 0 /* * XXX: currently there's no restriction in route information * prefix according to * draft-ietf-ipngwg-router-selection-00.txt. * However, I think the similar restriction be necessary. */ MAYHAVE(val64, entbuf, DEF_ADVVALIDLIFETIME); if (IN6_IS_ADDR_MULTICAST(&rti->prefix)) { syslog(LOG_ERR, "<%s> multicast route (%s) must " "not be advertised on %s", __func__, addr, ifi->ifi_ifname); goto getconfig_free_rti; } if (IN6_IS_ADDR_LINKLOCAL(&rti->prefix)) { syslog(LOG_NOTICE, "<%s> link-local route (%s) will " "be advertised on %s", __func__, addr, ifi->ifi_ifname); goto getconfig_free_rti; } #endif makeentry(entbuf, sizeof(entbuf), i, "rtplen"); /* XXX: 256 is a magic number for compatibility check. */ MAYHAVE(val, entbuf, 256); if (val == 256) { makeentry(oentbuf, sizeof(oentbuf), i, "rtrplen"); MAYHAVE(val, oentbuf, 256); if (val != 256) fprintf(stderr, "%s was obsoleted. Use %s.\n", oentbuf, entbuf); else val = 64; } if (val < 0 || val > 128) { syslog(LOG_ERR, "<%s> prefixlen (%" PRIu32 ") for %s on %s " "out of range", __func__, val, addr, ifi->ifi_ifname); goto getconfig_free_rti; } rti->rti_prefixlen = (int)val; makeentry(entbuf, sizeof(entbuf), i, "rtflags"); if ((flagstr = (char *)agetstr(entbuf, &bp))) { val = 0; if (strchr(flagstr, 'h')) val |= ND_RA_FLAG_RTPREF_HIGH; if (strchr(flagstr, 'l')) { if ((val & ND_RA_FLAG_RTPREF_HIGH)) { syslog(LOG_ERR, "<%s> the \'h\' and \'l\' route" " preferences are exclusive", __func__); goto getconfig_free_rti; } val |= ND_RA_FLAG_RTPREF_LOW; } } else MAYHAVE(val, entbuf, 256); /* XXX */ if (val == 256) { makeentry(oentbuf, sizeof(oentbuf), i, "rtrflags"); MAYHAVE(val, oentbuf, 256); if (val != 256) { fprintf(stderr, "%s was obsoleted. Use %s.\n", oentbuf, entbuf); } else val = 0; } rti->rti_rtpref = val & ND_RA_FLAG_RTPREF_MASK; if (rti->rti_rtpref == ND_RA_FLAG_RTPREF_RSV) { syslog(LOG_ERR, "<%s> invalid route preference (%02x) " "for %s/%d on %s", __func__, rti->rti_rtpref, addr, rti->rti_prefixlen, ifi->ifi_ifname); goto getconfig_free_rti; } /* * Since the spec does not a default value, we should make * this entry mandatory. However, FreeBSD 4.4 has shipped * with this field being optional, we use the router lifetime * as an ad-hoc default value with a warning message. */ makeentry(entbuf, sizeof(entbuf), i, "rtltime"); MAYHAVE(val64, entbuf, -1); if (val64 == -1) { makeentry(oentbuf, sizeof(oentbuf), i, "rtrltime"); MAYHAVE(val64, oentbuf, -1); if (val64 != -1) fprintf(stderr, "%s was obsoleted. Use %s.\n", oentbuf, entbuf); else { fprintf(stderr, "%s should be specified " "for interface %s.\n", entbuf, ifi->ifi_ifname); val64 = rai->rai_lifetime; } } if (val64 < 0 || val64 > 0xffffffff) { syslog(LOG_ERR, "<%s> route lifetime (%" PRIu64 ") for " "%s/%d on %s out of range", __func__, val64, addr, rti->rti_prefixlen, ifi->ifi_ifname); goto getconfig_free_rti; } rti->rti_ltime = (uint32_t)val64; /* link into chain */ TAILQ_INSERT_TAIL(&rai->rai_route, rti, rti_next); rai->rai_routes++; continue; getconfig_free_rti: free(rti); } /* DNS server and DNS search list information */ for (i = -1; i < MAXRDNSSENT ; i++) { struct rdnss *rdn; struct rdnss_addr *rdna; char *ap; int c; makeentry(entbuf, sizeof(entbuf), i, "rdnss"); addr = (char *)agetstr(entbuf, &bp); if (addr == NULL) break; ELM_MALLOC(rdn, exit(1)); TAILQ_INIT(&rdn->rd_list); for (ap = addr; ap - addr < (ssize_t)strlen(addr); ap += c+1) { c = strcspn(ap, ","); strncpy(abuf, ap, c); abuf[c] = '\0'; ELM_MALLOC(rdna, goto getconfig_free_rdn); if (inet_pton(AF_INET6, abuf, &rdna->ra_dns) != 1) { syslog(LOG_ERR, "<%s> inet_pton failed for %s", __func__, abuf); free(rdna); goto getconfig_free_rdn; } TAILQ_INSERT_TAIL(&rdn->rd_list, rdna, ra_next); } makeentry(entbuf, sizeof(entbuf), i, "rdnssltime"); MAYHAVE(val, entbuf, (rai->rai_maxinterval * 3 / 2)); if ((uint16_t)val < rai->rai_maxinterval || (uint16_t)val > rai->rai_maxinterval * 2) { syslog(LOG_ERR, "%s (%" PRIu16 ") on %s is invalid " "(must be between %d and %d)", entbuf, val, ifi->ifi_ifname, rai->rai_maxinterval, rai->rai_maxinterval * 2); goto getconfig_free_rdn; } rdn->rd_ltime = val; /* link into chain */ TAILQ_INSERT_TAIL(&rai->rai_rdnss, rdn, rd_next); continue; getconfig_free_rdn: while ((rdna = TAILQ_FIRST(&rdn->rd_list)) != NULL) { TAILQ_REMOVE(&rdn->rd_list, rdna, ra_next); free(rdna); } free(rdn); } for (i = -1; i < MAXDNSSLENT ; i++) { struct dnssl *dns; struct dnssl_addr *dnsa; char *ap; int c; makeentry(entbuf, sizeof(entbuf), i, "dnssl"); addr = (char *)agetstr(entbuf, &bp); if (addr == NULL) break; ELM_MALLOC(dns, exit(1)); TAILQ_INIT(&dns->dn_list); for (ap = addr; ap - addr < (ssize_t)strlen(addr); ap += c+1) { c = strcspn(ap, ","); strncpy(abuf, ap, c); abuf[c] = '\0'; ELM_MALLOC(dnsa, goto getconfig_free_dns); dnsa->da_len = dname_labelenc(dnsa->da_dom, abuf); syslog(LOG_DEBUG, "<%s>: dnsa->da_len = %d", __func__, dnsa->da_len); TAILQ_INSERT_TAIL(&dns->dn_list, dnsa, da_next); } makeentry(entbuf, sizeof(entbuf), i, "dnsslltime"); MAYHAVE(val, entbuf, (rai->rai_maxinterval * 3 / 2)); if ((uint16_t)val < rai->rai_maxinterval || (uint16_t)val > rai->rai_maxinterval * 2) { syslog(LOG_ERR, "%s (%" PRIu16 ") on %s is invalid " "(must be between %d and %d)", entbuf, val, ifi->ifi_ifname, rai->rai_maxinterval, rai->rai_maxinterval * 2); goto getconfig_free_dns; } dns->dn_ltime = val; /* link into chain */ TAILQ_INSERT_TAIL(&rai->rai_dnssl, dns, dn_next); continue; getconfig_free_dns: while ((dnsa = TAILQ_FIRST(&dns->dn_list)) != NULL) { TAILQ_REMOVE(&dns->dn_list, dnsa, da_next); free(dnsa); } free(dns); } /* construct the sending packet */ make_packet(rai); /* * If an entry with the same ifindex exists, remove it first. * Before the removal, RDNSS and DNSSL options with * zero-lifetime will be sent. */ switch (ifi->ifi_state) { case IFI_STATE_UNCONFIGURED: /* UNCONFIGURED -> TRANSITIVE */ error = sock_mc_join(&sock, ifi->ifi_ifindex); if (error) exit(1); ifi->ifi_state = IFI_STATE_TRANSITIVE; ifi->ifi_burstcount = MAX_INITIAL_RTR_ADVERTISEMENTS; ifi->ifi_burstinterval = MAX_INITIAL_RTR_ADVERT_INTERVAL; /* The same two rai mean initial burst */ ifi->ifi_rainfo = rai; ifi->ifi_rainfo_trans = rai; TAILQ_INSERT_TAIL(&railist, rai, rai_next); if (ifi->ifi_ra_timer == NULL) ifi->ifi_ra_timer = rtadvd_add_timer(ra_timeout, ra_timer_update, ifi, ifi); ra_timer_update(ifi, &ifi->ifi_ra_timer->rat_tm); rtadvd_set_timer(&ifi->ifi_ra_timer->rat_tm, ifi->ifi_ra_timer); syslog(LOG_DEBUG, "<%s> ifname=%s marked as TRANSITIVE (initial burst).", __func__, ifi->ifi_ifname); break; case IFI_STATE_CONFIGURED: /* CONFIGURED -> TRANSITIVE */ rai_old = ifi->ifi_rainfo; if (rai_old == NULL) { syslog(LOG_ERR, "<%s> ifi_rainfo is NULL" " in IFI_STATE_CONFIGURED.", __func__); ifi = NULL; break; } else { struct rdnss *rdn; struct dnssl *dns; rai_old->rai_lifetime = 0; TAILQ_FOREACH(rdn, &rai_old->rai_rdnss, rd_next) rdn->rd_ltime = 0; TAILQ_FOREACH(dns, &rai_old->rai_dnssl, dn_next) dns->dn_ltime = 0; ifi->ifi_rainfo_trans = rai_old; ifi->ifi_state = IFI_STATE_TRANSITIVE; ifi->ifi_burstcount = MAX_FINAL_RTR_ADVERTISEMENTS; ifi->ifi_burstinterval = MIN_DELAY_BETWEEN_RAS; ra_timer_update(ifi, &ifi->ifi_ra_timer->rat_tm); rtadvd_set_timer(&ifi->ifi_ra_timer->rat_tm, ifi->ifi_ra_timer); syslog(LOG_DEBUG, "<%s> ifname=%s marked as TRANSITIVE" " (transitional burst)", __func__, ifi->ifi_ifname); } ifi->ifi_rainfo = rai; TAILQ_INSERT_TAIL(&railist, rai, rai_next); break; case IFI_STATE_TRANSITIVE: if (ifi->ifi_rainfo != NULL) { if (ifi->ifi_rainfo == ifi->ifi_rainfo_trans) { /* Reinitialize initial burst */ rm_rainfo(ifi->ifi_rainfo); ifi->ifi_rainfo = rai; ifi->ifi_rainfo_trans = rai; ifi->ifi_burstcount = MAX_INITIAL_RTR_ADVERTISEMENTS; ifi->ifi_burstinterval = MAX_INITIAL_RTR_ADVERT_INTERVAL; } else { /* Replace ifi_rainfo with the new one */ rm_rainfo(ifi->ifi_rainfo); ifi->ifi_rainfo = rai; } TAILQ_INSERT_TAIL(&railist, rai, rai_next); ra_timer_update(ifi, &ifi->ifi_ra_timer->rat_tm); rtadvd_set_timer(&ifi->ifi_ra_timer->rat_tm, ifi->ifi_ra_timer); } else { /* XXX: NOTREACHED. Being shut down. */ syslog(LOG_ERR, "<%s> %s is shutting down. Skipped.", __func__, ifi->ifi_ifname); rm_rainfo(rai); return (NULL); } break; } return (ifi); getconfig_free_rai: free(rai); return (NULL); } void get_prefix(struct rainfo *rai) { struct ifaddrs *ifap, *ifa; struct prefix *pfx; struct in6_addr *a; struct ifinfo *ifi; char *p, *ep, *m, *lim; char ntopbuf[INET6_ADDRSTRLEN]; if (getifaddrs(&ifap) < 0) { syslog(LOG_ERR, "<%s> can't get interface addresses", __func__); exit(1); } ifi = rai->rai_ifinfo; for (ifa = ifap; ifa; ifa = ifa->ifa_next) { int plen; if (strcmp(ifa->ifa_name, ifi->ifi_ifname) != 0) continue; if (ifa->ifa_addr->sa_family != AF_INET6) continue; a = &((struct sockaddr_in6 *)ifa->ifa_addr)->sin6_addr; if (IN6_IS_ADDR_LINKLOCAL(a)) continue; /* get prefix length */ m = (char *)&((struct sockaddr_in6 *)ifa->ifa_netmask)->sin6_addr; lim = (char *)(ifa->ifa_netmask) + ifa->ifa_netmask->sa_len; plen = prefixlen(m, lim); if (plen <= 0 || plen > 128) { syslog(LOG_ERR, "<%s> failed to get prefixlen " "or prefix is invalid", __func__); exit(1); } if (plen == 128) /* XXX */ continue; if (find_prefix(rai, a, plen)) { /* ignore a duplicated prefix. */ continue; } /* allocate memory to store prefix info. */ ELM_MALLOC(pfx, exit(1)); /* set prefix, sweep bits outside of prefixlen */ pfx->pfx_prefixlen = plen; memcpy(&pfx->pfx_prefix, a, sizeof(*a)); p = (char *)&pfx->pfx_prefix; ep = (char *)(&pfx->pfx_prefix + 1); while (m < lim && p < ep) *p++ &= *m++; while (p < ep) *p++ = 0x00; if (!inet_ntop(AF_INET6, &pfx->pfx_prefix, ntopbuf, sizeof(ntopbuf))) { syslog(LOG_ERR, "<%s> inet_ntop failed", __func__); exit(1); } syslog(LOG_DEBUG, "<%s> add %s/%d to prefix list on %s", __func__, ntopbuf, pfx->pfx_prefixlen, ifi->ifi_ifname); /* set other fields with protocol defaults */ pfx->pfx_validlifetime = DEF_ADVVALIDLIFETIME; pfx->pfx_preflifetime = DEF_ADVPREFERREDLIFETIME; pfx->pfx_onlinkflg = 1; pfx->pfx_autoconfflg = 1; pfx->pfx_origin = PREFIX_FROM_KERNEL; pfx->pfx_rainfo = rai; /* link into chain */ TAILQ_INSERT_TAIL(&rai->rai_prefix, pfx, pfx_next); /* counter increment */ rai->rai_pfxs++; } freeifaddrs(ifap); } static void makeentry(char *buf, size_t len, int id, const char *string) { if (id < 0) strlcpy(buf, string, len); else snprintf(buf, len, "%s%d", string, id); } /* * Add a prefix to the list of specified interface and reconstruct * the outgoing packet. * The prefix must not be in the list. * XXX: other parameters of the prefix (e.g. lifetime) should be * able to be specified. */ static void add_prefix(struct rainfo *rai, struct in6_prefixreq *ipr) { struct prefix *pfx; struct ifinfo *ifi; char ntopbuf[INET6_ADDRSTRLEN]; ifi = rai->rai_ifinfo; ELM_MALLOC(pfx, return); pfx->pfx_prefix = ipr->ipr_prefix.sin6_addr; pfx->pfx_prefixlen = ipr->ipr_plen; pfx->pfx_validlifetime = ipr->ipr_vltime; pfx->pfx_preflifetime = ipr->ipr_pltime; pfx->pfx_onlinkflg = ipr->ipr_raf_onlink; pfx->pfx_autoconfflg = ipr->ipr_raf_auto; pfx->pfx_origin = PREFIX_FROM_DYNAMIC; TAILQ_INSERT_TAIL(&rai->rai_prefix, pfx, pfx_next); pfx->pfx_rainfo = rai; syslog(LOG_DEBUG, "<%s> new prefix %s/%d was added on %s", __func__, inet_ntop(AF_INET6, &ipr->ipr_prefix.sin6_addr, ntopbuf, sizeof(ntopbuf)), ipr->ipr_plen, ifi->ifi_ifname); rai->rai_pfxs++; } /* * Delete a prefix to the list of specified interface and reconstruct * the outgoing packet. * The prefix must be in the list. */ void delete_prefix(struct prefix *pfx) { struct rainfo *rai; struct ifinfo *ifi; char ntopbuf[INET6_ADDRSTRLEN]; rai = pfx->pfx_rainfo; ifi = rai->rai_ifinfo; TAILQ_REMOVE(&rai->rai_prefix, pfx, pfx_next); syslog(LOG_DEBUG, "<%s> prefix %s/%d was deleted on %s", __func__, inet_ntop(AF_INET6, &pfx->pfx_prefix, ntopbuf, sizeof(ntopbuf)), pfx->pfx_prefixlen, ifi->ifi_ifname); if (pfx->pfx_timer) rtadvd_remove_timer(pfx->pfx_timer); free(pfx); rai->rai_pfxs--; } void invalidate_prefix(struct prefix *pfx) { - struct timeval timo; + struct timespec timo; struct rainfo *rai; struct ifinfo *ifi; char ntopbuf[INET6_ADDRSTRLEN]; rai = pfx->pfx_rainfo; ifi = rai->rai_ifinfo; if (pfx->pfx_timer) { /* sanity check */ syslog(LOG_ERR, "<%s> assumption failure: timer already exists", __func__); exit(1); } syslog(LOG_DEBUG, "<%s> prefix %s/%d was invalidated on %s, " "will expire in %ld seconds", __func__, inet_ntop(AF_INET6, &pfx->pfx_prefix, ntopbuf, sizeof(ntopbuf)), pfx->pfx_prefixlen, ifi->ifi_ifname, (long)prefix_timo); /* set the expiration timer */ pfx->pfx_timer = rtadvd_add_timer(prefix_timeout, NULL, pfx, NULL); if (pfx->pfx_timer == NULL) { syslog(LOG_ERR, "<%s> failed to add a timer for a prefix. " "remove the prefix", __func__); delete_prefix(pfx); } timo.tv_sec = prefix_timo; - timo.tv_usec = 0; + timo.tv_nsec = 0; rtadvd_set_timer(&timo, pfx->pfx_timer); } static struct rtadvd_timer * prefix_timeout(void *arg) { delete_prefix((struct prefix *)arg); return (NULL); } void update_prefix(struct prefix *pfx) { struct rainfo *rai; struct ifinfo *ifi; char ntopbuf[INET6_ADDRSTRLEN]; rai = pfx->pfx_rainfo; ifi = rai->rai_ifinfo; if (pfx->pfx_timer == NULL) { /* sanity check */ syslog(LOG_ERR, "<%s> assumption failure: timer does not exist", __func__); exit(1); } syslog(LOG_DEBUG, "<%s> prefix %s/%d was re-enabled on %s", __func__, inet_ntop(AF_INET6, &pfx->pfx_prefix, ntopbuf, sizeof(ntopbuf)), pfx->pfx_prefixlen, ifi->ifi_ifname); /* stop the expiration timer */ rtadvd_remove_timer(pfx->pfx_timer); pfx->pfx_timer = NULL; } /* * Try to get an in6_prefixreq contents for a prefix which matches * ipr->ipr_prefix and ipr->ipr_plen and belongs to * the interface whose name is ipr->ipr_name[]. */ static int init_prefix(struct in6_prefixreq *ipr) { #if 0 int s; if ((s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) { syslog(LOG_ERR, "<%s> socket: %s", __func__, strerror(errno)); exit(1); } if (ioctl(s, SIOCGIFPREFIX_IN6, (caddr_t)ipr) < 0) { syslog(LOG_INFO, "<%s> ioctl:SIOCGIFPREFIX %s", __func__, strerror(errno)); ipr->ipr_vltime = DEF_ADVVALIDLIFETIME; ipr->ipr_pltime = DEF_ADVPREFERREDLIFETIME; ipr->ipr_raf_onlink = 1; ipr->ipr_raf_auto = 1; /* omit other field initialization */ } else if (ipr->ipr_origin < PR_ORIG_RR) { char ntopbuf[INET6_ADDRSTRLEN]; syslog(LOG_WARNING, "<%s> Added prefix(%s)'s origin %d is" "lower than PR_ORIG_RR(router renumbering)." "This should not happen if I am router", __func__, inet_ntop(AF_INET6, &ipr->ipr_prefix.sin6_addr, ntopbuf, sizeof(ntopbuf)), ipr->ipr_origin); close(s); return (1); } close(s); return (0); #else ipr->ipr_vltime = DEF_ADVVALIDLIFETIME; ipr->ipr_pltime = DEF_ADVPREFERREDLIFETIME; ipr->ipr_raf_onlink = 1; ipr->ipr_raf_auto = 1; return (0); #endif } void make_prefix(struct rainfo *rai, int ifindex, struct in6_addr *addr, int plen) { struct in6_prefixreq ipr; memset(&ipr, 0, sizeof(ipr)); if (if_indextoname(ifindex, ipr.ipr_name) == NULL) { syslog(LOG_ERR, "<%s> Prefix added interface No.%d doesn't " "exist. This should not happen! %s", __func__, ifindex, strerror(errno)); exit(1); } ipr.ipr_prefix.sin6_len = sizeof(ipr.ipr_prefix); ipr.ipr_prefix.sin6_family = AF_INET6; ipr.ipr_prefix.sin6_addr = *addr; ipr.ipr_plen = plen; if (init_prefix(&ipr)) return; /* init failed by some error */ add_prefix(rai, &ipr); } void make_packet(struct rainfo *rai) { size_t packlen, lladdroptlen = 0; char *buf; struct nd_router_advert *ra; struct nd_opt_prefix_info *ndopt_pi; struct nd_opt_mtu *ndopt_mtu; struct nd_opt_route_info *ndopt_rti; struct rtinfo *rti; struct nd_opt_rdnss *ndopt_rdnss; struct rdnss *rdn; struct nd_opt_dnssl *ndopt_dnssl; struct dnssl *dns; size_t len; struct prefix *pfx; struct ifinfo *ifi; ifi = rai->rai_ifinfo; /* calculate total length */ packlen = sizeof(struct nd_router_advert); if (rai->rai_advlinkopt) { if ((lladdroptlen = lladdropt_length(&ifi->ifi_sdl)) == 0) { syslog(LOG_INFO, "<%s> link-layer address option has" " null length on %s. Treat as not included.", __func__, ifi->ifi_ifname); rai->rai_advlinkopt = 0; } packlen += lladdroptlen; } if (rai->rai_pfxs) packlen += sizeof(struct nd_opt_prefix_info) * rai->rai_pfxs; if (rai->rai_linkmtu) packlen += sizeof(struct nd_opt_mtu); TAILQ_FOREACH(rti, &rai->rai_route, rti_next) packlen += sizeof(struct nd_opt_route_info) + ((rti->rti_prefixlen + 0x3f) >> 6) * 8; TAILQ_FOREACH(rdn, &rai->rai_rdnss, rd_next) { struct rdnss_addr *rdna; packlen += sizeof(struct nd_opt_rdnss); TAILQ_FOREACH(rdna, &rdn->rd_list, ra_next) packlen += sizeof(rdna->ra_dns); } TAILQ_FOREACH(dns, &rai->rai_dnssl, dn_next) { struct dnssl_addr *dnsa; packlen += sizeof(struct nd_opt_dnssl); len = 0; TAILQ_FOREACH(dnsa, &dns->dn_list, da_next) len += dnsa->da_len; /* A zero octet and 8 octet boundary */ len++; len += (len % 8) ? 8 - len % 8 : 0; packlen += len; } /* allocate memory for the packet */ if ((buf = malloc(packlen)) == NULL) { syslog(LOG_ERR, "<%s> can't get enough memory for an RA packet", __func__); exit(1); } memset(buf, 0, packlen); if (rai->rai_ra_data) /* Free old data if any. */ free(rai->rai_ra_data); rai->rai_ra_data = buf; /* XXX: what if packlen > 576? */ rai->rai_ra_datalen = packlen; /* * construct the packet */ ra = (struct nd_router_advert *)buf; ra->nd_ra_type = ND_ROUTER_ADVERT; ra->nd_ra_code = 0; ra->nd_ra_cksum = 0; ra->nd_ra_curhoplimit = (uint8_t)(0xff & rai->rai_hoplimit); ra->nd_ra_flags_reserved = 0; /* just in case */ /* * XXX: the router preference field, which is a 2-bit field, should be * initialized before other fields. */ ra->nd_ra_flags_reserved = 0xff & rai->rai_rtpref; ra->nd_ra_flags_reserved |= rai->rai_managedflg ? ND_RA_FLAG_MANAGED : 0; ra->nd_ra_flags_reserved |= rai->rai_otherflg ? ND_RA_FLAG_OTHER : 0; ra->nd_ra_router_lifetime = htons(rai->rai_lifetime); ra->nd_ra_reachable = htonl(rai->rai_reachabletime); ra->nd_ra_retransmit = htonl(rai->rai_retranstimer); buf += sizeof(*ra); if (rai->rai_advlinkopt) { lladdropt_fill(&ifi->ifi_sdl, (struct nd_opt_hdr *)buf); buf += lladdroptlen; } if (rai->rai_linkmtu) { ndopt_mtu = (struct nd_opt_mtu *)buf; ndopt_mtu->nd_opt_mtu_type = ND_OPT_MTU; ndopt_mtu->nd_opt_mtu_len = 1; ndopt_mtu->nd_opt_mtu_reserved = 0; ndopt_mtu->nd_opt_mtu_mtu = htonl(rai->rai_linkmtu); buf += sizeof(struct nd_opt_mtu); } TAILQ_FOREACH(pfx, &rai->rai_prefix, pfx_next) { uint32_t vltime, pltime; - struct timeval now; + struct timespec now; ndopt_pi = (struct nd_opt_prefix_info *)buf; ndopt_pi->nd_opt_pi_type = ND_OPT_PREFIX_INFORMATION; ndopt_pi->nd_opt_pi_len = 4; ndopt_pi->nd_opt_pi_prefix_len = pfx->pfx_prefixlen; ndopt_pi->nd_opt_pi_flags_reserved = 0; if (pfx->pfx_onlinkflg) ndopt_pi->nd_opt_pi_flags_reserved |= ND_OPT_PI_FLAG_ONLINK; if (pfx->pfx_autoconfflg) ndopt_pi->nd_opt_pi_flags_reserved |= ND_OPT_PI_FLAG_AUTO; if (pfx->pfx_timer) vltime = 0; else { if (pfx->pfx_vltimeexpire || pfx->pfx_pltimeexpire) - gettimeofday(&now, NULL); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); if (pfx->pfx_vltimeexpire == 0) vltime = pfx->pfx_validlifetime; else vltime = ((time_t)pfx->pfx_vltimeexpire > now.tv_sec) ? pfx->pfx_vltimeexpire - now.tv_sec : 0; } if (pfx->pfx_timer) pltime = 0; else { if (pfx->pfx_pltimeexpire == 0) pltime = pfx->pfx_preflifetime; else pltime = ((time_t)pfx->pfx_pltimeexpire > now.tv_sec) ? pfx->pfx_pltimeexpire - now.tv_sec : 0; } if (vltime < pltime) { /* * this can happen if vltime is decrement but pltime * is not. */ pltime = vltime; } ndopt_pi->nd_opt_pi_valid_time = htonl(vltime); ndopt_pi->nd_opt_pi_preferred_time = htonl(pltime); ndopt_pi->nd_opt_pi_reserved2 = 0; ndopt_pi->nd_opt_pi_prefix = pfx->pfx_prefix; buf += sizeof(struct nd_opt_prefix_info); } TAILQ_FOREACH(rti, &rai->rai_route, rti_next) { uint8_t psize = (rti->rti_prefixlen + 0x3f) >> 6; ndopt_rti = (struct nd_opt_route_info *)buf; ndopt_rti->nd_opt_rti_type = ND_OPT_ROUTE_INFO; ndopt_rti->nd_opt_rti_len = 1 + psize; ndopt_rti->nd_opt_rti_prefixlen = rti->rti_prefixlen; ndopt_rti->nd_opt_rti_flags = 0xff & rti->rti_rtpref; ndopt_rti->nd_opt_rti_lifetime = htonl(rti->rti_ltime); memcpy(ndopt_rti + 1, &rti->rti_prefix, psize * 8); buf += sizeof(struct nd_opt_route_info) + psize * 8; } TAILQ_FOREACH(rdn, &rai->rai_rdnss, rd_next) { struct rdnss_addr *rdna; ndopt_rdnss = (struct nd_opt_rdnss *)buf; ndopt_rdnss->nd_opt_rdnss_type = ND_OPT_RDNSS; ndopt_rdnss->nd_opt_rdnss_len = 0; ndopt_rdnss->nd_opt_rdnss_reserved = 0; ndopt_rdnss->nd_opt_rdnss_lifetime = htonl(rdn->rd_ltime); buf += sizeof(struct nd_opt_rdnss); TAILQ_FOREACH(rdna, &rdn->rd_list, ra_next) { memcpy(buf, &rdna->ra_dns, sizeof(rdna->ra_dns)); buf += sizeof(rdna->ra_dns); } /* Length field should be in 8 octets */ ndopt_rdnss->nd_opt_rdnss_len = (buf - (char *)ndopt_rdnss) / 8; syslog(LOG_DEBUG, "<%s>: nd_opt_dnss_len = %d", __func__, ndopt_rdnss->nd_opt_rdnss_len); } TAILQ_FOREACH(dns, &rai->rai_dnssl, dn_next) { struct dnssl_addr *dnsa; ndopt_dnssl = (struct nd_opt_dnssl *)buf; ndopt_dnssl->nd_opt_dnssl_type = ND_OPT_DNSSL; ndopt_dnssl->nd_opt_dnssl_len = 0; ndopt_dnssl->nd_opt_dnssl_reserved = 0; ndopt_dnssl->nd_opt_dnssl_lifetime = htonl(dns->dn_ltime); buf += sizeof(*ndopt_dnssl); TAILQ_FOREACH(dnsa, &dns->dn_list, da_next) { memcpy(buf, dnsa->da_dom, dnsa->da_len); buf += dnsa->da_len; } /* A zero octet after encoded DNS server list. */ *buf++ = '\0'; /* Padding to next 8 octets boundary */ len = buf - (char *)ndopt_dnssl; len += (len % 8) ? 8 - len % 8 : 0; /* Length field must be in 8 octets */ ndopt_dnssl->nd_opt_dnssl_len = len / 8; syslog(LOG_DEBUG, "<%s>: nd_opt_dnssl_len = %d", __func__, ndopt_dnssl->nd_opt_dnssl_len); } return; } Index: head/usr.sbin/rtadvd/rrenum.c =================================================================== --- head/usr.sbin/rtadvd/rrenum.c (revision 253969) +++ head/usr.sbin/rtadvd/rrenum.c (revision 253970) @@ -1,500 +1,503 @@ /* $FreeBSD$ */ /* $KAME: rrenum.c,v 1.12 2002/06/10 19:59:47 itojun Exp $ */ /* * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include +#include #include #include "rtadvd.h" #include "rrenum.h" #include "if.h" #define RR_ISSET_SEGNUM(segnum_bits, segnum) \ ((((segnum_bits)[(segnum) >> 5]) & (1 << ((segnum) & 31))) != 0) #define RR_SET_SEGNUM(segnum_bits, segnum) \ (((segnum_bits)[(segnum) >> 5]) |= (1 << ((segnum) & 31))) struct rr_operation { u_long rro_seqnum; u_long rro_segnum_bits[8]; }; static struct rr_operation rro; static int rr_rcvifindex; static int rrcmd2pco[RPM_PCO_MAX] = { 0, SIOCAIFPREFIX_IN6, SIOCCIFPREFIX_IN6, SIOCSGIFPREFIX_IN6 }; static int s = -1; /* * Check validity of a Prefix Control Operation(PCO). * return 0 on success, 1 on failure. */ static int rr_pco_check(int len, struct rr_pco_match *rpm) { struct rr_pco_use *rpu, *rpulim; int checklen; /* rpm->rpm_len must be (4N * 3) as router-renum-05.txt */ if ((rpm->rpm_len - 3) < 0 || /* must be at least 3 */ (rpm->rpm_len - 3) & 0x3) { /* must be multiple of 4 */ syslog(LOG_WARNING, "<%s> rpm_len %d is not 4N * 3", __func__, rpm->rpm_len); return (1); } /* rpm->rpm_code must be valid value */ switch (rpm->rpm_code) { case RPM_PCO_ADD: case RPM_PCO_CHANGE: case RPM_PCO_SETGLOBAL: break; default: syslog(LOG_WARNING, "<%s> unknown rpm_code %d", __func__, rpm->rpm_code); return (1); } /* rpm->rpm_matchlen must be 0 to 128 inclusive */ if (rpm->rpm_matchlen > 128) { syslog(LOG_WARNING, "<%s> rpm_matchlen %d is over 128", __func__, rpm->rpm_matchlen); return (1); } /* * rpu->rpu_uselen, rpu->rpu_keeplen, and sum of them must be * between 0 and 128 inclusive */ for (rpu = (struct rr_pco_use *)(rpm + 1), rpulim = (struct rr_pco_use *)((char *)rpm + len); rpu < rpulim; rpu += 1) { checklen = rpu->rpu_uselen; checklen += rpu->rpu_keeplen; /* * omit these check, because either of rpu_uselen * and rpu_keeplen is unsigned char * (128 > rpu_uselen > 0) * (128 > rpu_keeplen > 0) * (rpu_uselen + rpu_keeplen > 0) */ if (checklen > 128) { syslog(LOG_WARNING, "<%s> sum of rpu_uselen %d and" " rpu_keeplen %d is %d(over 128)", __func__, rpu->rpu_uselen, rpu->rpu_keeplen, rpu->rpu_uselen + rpu->rpu_keeplen); return (1); } } return (0); } static void do_use_prefix(int len, struct rr_pco_match *rpm, struct in6_rrenumreq *irr, int ifindex) { struct rr_pco_use *rpu, *rpulim; struct rainfo *rai; struct ifinfo *ifi; struct prefix *pfx; rpu = (struct rr_pco_use *)(rpm + 1); rpulim = (struct rr_pco_use *)((char *)rpm + len); if (rpu == rpulim) { /* no use prefix */ if (rpm->rpm_code == RPM_PCO_ADD) return; irr->irr_u_uselen = 0; irr->irr_u_keeplen = 0; irr->irr_raf_mask_onlink = 0; irr->irr_raf_mask_auto = 0; irr->irr_vltime = 0; irr->irr_pltime = 0; memset(&irr->irr_flags, 0, sizeof(irr->irr_flags)); irr->irr_useprefix.sin6_len = 0; /* let it mean, no addition */ irr->irr_useprefix.sin6_family = 0; irr->irr_useprefix.sin6_addr = in6addr_any; if (ioctl(s, rrcmd2pco[rpm->rpm_code], (caddr_t)irr) < 0 && errno != EADDRNOTAVAIL) syslog(LOG_ERR, "<%s> ioctl: %s", __func__, strerror(errno)); return; } for (rpu = (struct rr_pco_use *)(rpm + 1), rpulim = (struct rr_pco_use *)((char *)rpm + len); rpu < rpulim; rpu += 1) { /* init in6_rrenumreq fields */ irr->irr_u_uselen = rpu->rpu_uselen; irr->irr_u_keeplen = rpu->rpu_keeplen; irr->irr_raf_mask_onlink = !!(rpu->rpu_ramask & ICMP6_RR_PCOUSE_RAFLAGS_ONLINK); irr->irr_raf_mask_auto = !!(rpu->rpu_ramask & ICMP6_RR_PCOUSE_RAFLAGS_AUTO); irr->irr_vltime = ntohl(rpu->rpu_vltime); irr->irr_pltime = ntohl(rpu->rpu_pltime); irr->irr_raf_onlink = (rpu->rpu_raflags & ICMP6_RR_PCOUSE_RAFLAGS_ONLINK) == 0 ? 0 : 1; irr->irr_raf_auto = (rpu->rpu_raflags & ICMP6_RR_PCOUSE_RAFLAGS_AUTO) == 0 ? 0 : 1; irr->irr_rrf_decrvalid = (rpu->rpu_flags & ICMP6_RR_PCOUSE_FLAGS_DECRVLTIME) == 0 ? 0 : 1; irr->irr_rrf_decrprefd = (rpu->rpu_flags & ICMP6_RR_PCOUSE_FLAGS_DECRPLTIME) == 0 ? 0 : 1; irr->irr_useprefix.sin6_len = sizeof(irr->irr_useprefix); irr->irr_useprefix.sin6_family = AF_INET6; irr->irr_useprefix.sin6_addr = rpu->rpu_prefix; if (ioctl(s, rrcmd2pco[rpm->rpm_code], (caddr_t)irr) < 0 && errno != EADDRNOTAVAIL) syslog(LOG_ERR, "<%s> ioctl: %s", __func__, strerror(errno)); /* very adhoc: should be rewritten */ if (rpm->rpm_code == RPM_PCO_CHANGE && IN6_ARE_ADDR_EQUAL(&rpm->rpm_prefix, &rpu->rpu_prefix) && rpm->rpm_matchlen == rpu->rpu_uselen && rpu->rpu_uselen == rpu->rpu_keeplen) { ifi = if_indextoifinfo(ifindex); if (ifi == NULL || ifi->ifi_rainfo == NULL) continue; /* non-advertising IF */ rai = ifi->ifi_rainfo; TAILQ_FOREACH(pfx, &rai->rai_prefix, pfx_next) { - struct timeval now; + struct timespec now; if (prefix_match(&pfx->pfx_prefix, pfx->pfx_prefixlen, &rpm->rpm_prefix, rpm->rpm_matchlen)) { /* change parameters */ pfx->pfx_validlifetime = ntohl(rpu->rpu_vltime); pfx->pfx_preflifetime = ntohl(rpu->rpu_pltime); if (irr->irr_rrf_decrvalid) { - gettimeofday(&now, 0); + clock_gettime(CLOCK_MONOTONIC_FAST, + &now); pfx->pfx_vltimeexpire = now.tv_sec + pfx->pfx_validlifetime; } else pfx->pfx_vltimeexpire = 0; if (irr->irr_rrf_decrprefd) { - gettimeofday(&now, 0); + clock_gettime(CLOCK_MONOTONIC_FAST, + &now); pfx->pfx_pltimeexpire = now.tv_sec + pfx->pfx_preflifetime; } else pfx->pfx_pltimeexpire = 0; } } } } } /* * process a Prefix Control Operation(PCO). * return 0 on success, 1 on failure */ static int do_pco(struct icmp6_router_renum *rr, int len, struct rr_pco_match *rpm) { int ifindex = 0; struct in6_rrenumreq irr; struct ifinfo *ifi; if ((rr_pco_check(len, rpm) != 0)) return (1); if (s == -1 && (s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) { syslog(LOG_ERR, "<%s> socket: %s", __func__, strerror(errno)); exit(1); } memset(&irr, 0, sizeof(irr)); irr.irr_origin = PR_ORIG_RR; irr.irr_m_len = rpm->rpm_matchlen; irr.irr_m_minlen = rpm->rpm_minlen; irr.irr_m_maxlen = rpm->rpm_maxlen; irr.irr_matchprefix.sin6_len = sizeof(irr.irr_matchprefix); irr.irr_matchprefix.sin6_family = AF_INET6; irr.irr_matchprefix.sin6_addr = rpm->rpm_prefix; while (if_indextoname(++ifindex, irr.irr_name)) { ifi = if_indextoifinfo(ifindex); if (ifi == NULL) { syslog(LOG_ERR, "<%s> ifindex not found.", __func__); return (1); } /* * if ICMP6_RR_FLAGS_FORCEAPPLY(A flag) is 0 and * IFF_UP is off, the interface is not applied */ if ((rr->rr_flags & ICMP6_RR_FLAGS_FORCEAPPLY) == 0 && (ifi->ifi_flags & IFF_UP) == 0) continue; /* TODO: interface scope check */ do_use_prefix(len, rpm, &irr, ifindex); } if (errno == ENXIO) return (0); else if (errno) { syslog(LOG_ERR, "<%s> if_indextoname: %s", __func__, strerror(errno)); return (1); } return (0); } /* * call do_pco() for each Prefix Control Operations(PCOs) in a received * Router Renumbering Command packet. * return 0 on success, 1 on failure */ static int do_rr(int len, struct icmp6_router_renum *rr) { struct rr_pco_match *rpm; char *cp, *lim; lim = (char *)rr + len; cp = (char *)(rr + 1); len -= sizeof(struct icmp6_router_renum); update_ifinfo(&ifilist, UPDATE_IFINFO_ALL); while (cp < lim) { int rpmlen; rpm = (struct rr_pco_match *)cp; if ((size_t)len < sizeof(struct rr_pco_match)) { tooshort: syslog(LOG_ERR, "<%s> pkt too short. left len = %d. " "garbage at end of pkt?", __func__, len); return (1); } rpmlen = rpm->rpm_len << 3; if (len < rpmlen) goto tooshort; if (do_pco(rr, rpmlen, rpm)) { syslog(LOG_WARNING, "<%s> invalid PCO", __func__); goto next; } next: cp += rpmlen; len -= rpmlen; } return (0); } /* * check validity of a router renumbering command packet * return 0 on success, 1 on failure */ static int rr_command_check(int len, struct icmp6_router_renum *rr, struct in6_addr *from, struct in6_addr *dst) { u_char ntopbuf[INET6_ADDRSTRLEN]; /* omit rr minimal length check. hope kernel have done it. */ /* rr_command length check */ if ((size_t)len < (sizeof(struct icmp6_router_renum) + sizeof(struct rr_pco_match))) { syslog(LOG_ERR, "<%s> rr_command len %d is too short", __func__, len); return (1); } /* destination check. only for multicast. omit unicast check. */ if (IN6_IS_ADDR_MULTICAST(dst) && !IN6_IS_ADDR_MC_LINKLOCAL(dst) && !IN6_IS_ADDR_MC_SITELOCAL(dst)) { syslog(LOG_ERR, "<%s> dst mcast addr %s is illegal", __func__, inet_ntop(AF_INET6, dst, ntopbuf, sizeof(ntopbuf))); return (1); } /* seqnum and segnum check */ if (rro.rro_seqnum > rr->rr_seqnum) { syslog(LOG_WARNING, "<%s> rcvd old seqnum %d from %s", __func__, (u_int32_t)ntohl(rr->rr_seqnum), inet_ntop(AF_INET6, from, ntopbuf, sizeof(ntopbuf))); return (1); } if (rro.rro_seqnum == rr->rr_seqnum && (rr->rr_flags & ICMP6_RR_FLAGS_TEST) == 0 && RR_ISSET_SEGNUM(rro.rro_segnum_bits, rr->rr_segnum)) { if ((rr->rr_flags & ICMP6_RR_FLAGS_REQRESULT) != 0) syslog(LOG_WARNING, "<%s> rcvd duped segnum %d from %s", __func__, rr->rr_segnum, inet_ntop(AF_INET6, from, ntopbuf, sizeof(ntopbuf))); return (0); } /* update seqnum */ if (rro.rro_seqnum != rr->rr_seqnum) { /* then must be "<" */ /* init rro_segnum_bits */ memset(rro.rro_segnum_bits, 0, sizeof(rro.rro_segnum_bits)); } rro.rro_seqnum = rr->rr_seqnum; return (0); } static void rr_command_input(int len, struct icmp6_router_renum *rr, struct in6_addr *from, struct in6_addr *dst) { /* rr_command validity check */ if (rr_command_check(len, rr, from, dst)) goto failed; if ((rr->rr_flags & (ICMP6_RR_FLAGS_TEST|ICMP6_RR_FLAGS_REQRESULT)) == ICMP6_RR_FLAGS_TEST) return; /* do router renumbering */ if (do_rr(len, rr)) goto failed; /* update segnum */ RR_SET_SEGNUM(rro.rro_segnum_bits, rr->rr_segnum); return; failed: syslog(LOG_ERR, "<%s> received RR was invalid", __func__); return; } void rr_input(int len, struct icmp6_router_renum *rr, struct in6_pktinfo *pi, struct sockaddr_in6 *from, struct in6_addr *dst) { u_char ntopbuf[2][INET6_ADDRSTRLEN], ifnamebuf[IFNAMSIZ]; syslog(LOG_DEBUG, "<%s> RR received from %s to %s on %s", __func__, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf[0] ,sizeof(ntopbuf[0])), inet_ntop(AF_INET6, &dst, ntopbuf[1], sizeof(ntopbuf[1])), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); /* packet validation based on Section 4.1 of RFC2894 */ if ((size_t)len < sizeof(struct icmp6_router_renum)) { syslog(LOG_NOTICE, "<%s>: RR short message (size %d) from %s to %s on %s", __func__, len, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf[0], sizeof(ntopbuf[0])), inet_ntop(AF_INET6, &dst, ntopbuf[1], sizeof(ntopbuf[1])), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } /* * If the IPv6 destination address is neither an All Routers multicast * address [AARCH] nor one of the receiving router's unicast addresses, * the message MUST be discarded and SHOULD be logged to network * management. * We rely on the kernel input routine for unicast addresses, and thus * check multicast destinations only. */ if (IN6_IS_ADDR_MULTICAST(&pi->ipi6_addr) && !IN6_ARE_ADDR_EQUAL( &sin6_sitelocal_allrouters.sin6_addr, &pi->ipi6_addr)) { syslog(LOG_NOTICE, "<%s>: RR message with invalid destination (%s) " "from %s on %s", __func__, inet_ntop(AF_INET6, &dst, ntopbuf[0], sizeof(ntopbuf[0])), inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf[1], sizeof(ntopbuf[1])), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } rr_rcvifindex = pi->ipi6_ifindex; switch (rr->rr_code) { case ICMP6_ROUTER_RENUMBERING_COMMAND: rr_command_input(len, rr, &from->sin6_addr, dst); /* TODO: send reply msg */ break; case ICMP6_ROUTER_RENUMBERING_RESULT: /* RESULT will be processed by rrenumd */ break; case ICMP6_ROUTER_RENUMBERING_SEQNUM_RESET: /* TODO: sequence number reset */ break; default: syslog(LOG_ERR, "<%s> received unknown code %d", __func__, rr->rr_code); break; } return; } Index: head/usr.sbin/rtadvd/rtadvd.c =================================================================== --- head/usr.sbin/rtadvd/rtadvd.c (revision 253969) +++ head/usr.sbin/rtadvd/rtadvd.c (revision 253970) @@ -1,1927 +1,1926 @@ /* $FreeBSD$ */ /* $KAME: rtadvd.c,v 1.82 2003/08/05 12:34:23 itojun Exp $ */ /* * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * Copyright (C) 2011 Hiroki Sato * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #include #include #include #include -#include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "pathnames.h" #include "rtadvd.h" #include "if.h" #include "rrenum.h" #include "advcap.h" #include "timer_subr.h" #include "timer.h" #include "config.h" #include "control.h" #include "control_server.h" #define RTADV_TYPE2BITMASK(type) (0x1 << type) struct msghdr rcvmhdr; static char *rcvcmsgbuf; static size_t rcvcmsgbuflen; static char *sndcmsgbuf = NULL; static size_t sndcmsgbuflen; struct msghdr sndmhdr; struct iovec rcviov[2]; struct iovec sndiov[2]; struct sockaddr_in6 rcvfrom; static const char *pidfilename = _PATH_RTADVDPID; const char *conffile = _PATH_RTADVDCONF; static struct pidfh *pfh; static int dflag, sflag; static int wait_shutdown; #define PFD_RAWSOCK 0 #define PFD_RTSOCK 1 #define PFD_CSOCK 2 #define PFD_MAX 3 struct railist_head_t railist = TAILQ_HEAD_INITIALIZER(railist); struct ifilist_head_t ifilist = TAILQ_HEAD_INITIALIZER(ifilist); struct nd_optlist { TAILQ_ENTRY(nd_optlist) nol_next; struct nd_opt_hdr *nol_opt; }; union nd_opt { struct nd_opt_hdr *opt_array[9]; struct { struct nd_opt_hdr *zero; struct nd_opt_hdr *src_lladdr; struct nd_opt_hdr *tgt_lladdr; struct nd_opt_prefix_info *pi; struct nd_opt_rd_hdr *rh; struct nd_opt_mtu *mtu; TAILQ_HEAD(, nd_optlist) opt_list; } nd_opt_each; }; #define opt_src_lladdr nd_opt_each.src_lladdr #define opt_tgt_lladdr nd_opt_each.tgt_lladdr #define opt_pi nd_opt_each.pi #define opt_rh nd_opt_each.rh #define opt_mtu nd_opt_each.mtu #define opt_list nd_opt_each.opt_list #define NDOPT_FLAG_SRCLINKADDR (1 << 0) #define NDOPT_FLAG_TGTLINKADDR (1 << 1) #define NDOPT_FLAG_PREFIXINFO (1 << 2) #define NDOPT_FLAG_RDHDR (1 << 3) #define NDOPT_FLAG_MTU (1 << 4) #define NDOPT_FLAG_RDNSS (1 << 5) #define NDOPT_FLAG_DNSSL (1 << 6) static uint32_t ndopt_flags[] = { [ND_OPT_SOURCE_LINKADDR] = NDOPT_FLAG_SRCLINKADDR, [ND_OPT_TARGET_LINKADDR] = NDOPT_FLAG_TGTLINKADDR, [ND_OPT_PREFIX_INFORMATION] = NDOPT_FLAG_PREFIXINFO, [ND_OPT_REDIRECTED_HEADER] = NDOPT_FLAG_RDHDR, [ND_OPT_MTU] = NDOPT_FLAG_MTU, [ND_OPT_RDNSS] = NDOPT_FLAG_RDNSS, [ND_OPT_DNSSL] = NDOPT_FLAG_DNSSL, }; static void rtadvd_shutdown(void); static void sock_open(struct sockinfo *); static void rtsock_open(struct sockinfo *); static void rtadvd_input(struct sockinfo *); static void rs_input(int, struct nd_router_solicit *, struct in6_pktinfo *, struct sockaddr_in6 *); static void ra_input(int, struct nd_router_advert *, struct in6_pktinfo *, struct sockaddr_in6 *); static int prefix_check(struct nd_opt_prefix_info *, struct rainfo *, struct sockaddr_in6 *); static int nd6_options(struct nd_opt_hdr *, int, union nd_opt *, uint32_t); static void free_ndopts(union nd_opt *); static void rtmsg_input(struct sockinfo *); static void set_short_delay(struct ifinfo *); static int check_accept_rtadv(int); static void usage(void) { fprintf(stderr, "usage: rtadvd [-dDfRs] " "[-c configfile] [-C ctlsock] [-M ifname] [-p pidfile]\n"); exit(1); } int main(int argc, char *argv[]) { struct pollfd set[PFD_MAX]; - struct timeval *timeout; + struct timespec *timeout; int i, ch; int fflag = 0, logopt; int error; pid_t pid, otherpid; /* get command line options and arguments */ while ((ch = getopt(argc, argv, "c:C:dDfhM:p:Rs")) != -1) { switch (ch) { case 'c': conffile = optarg; break; case 'C': ctrlsock.si_name = optarg; break; case 'd': dflag++; break; case 'D': dflag += 3; break; case 'f': fflag = 1; break; case 'M': mcastif = optarg; break; case 'R': fprintf(stderr, "rtadvd: " "the -R option is currently ignored.\n"); /* accept_rr = 1; */ /* run anyway... */ break; case 's': sflag = 1; break; case 'p': pidfilename = optarg; break; default: usage(); } } argc -= optind; argv += optind; logopt = LOG_NDELAY | LOG_PID; if (fflag) logopt |= LOG_PERROR; openlog("rtadvd", logopt, LOG_DAEMON); /* set log level */ if (dflag > 2) (void)setlogmask(LOG_UPTO(LOG_DEBUG)); else if (dflag > 1) (void)setlogmask(LOG_UPTO(LOG_INFO)); else if (dflag > 0) (void)setlogmask(LOG_UPTO(LOG_NOTICE)); else (void)setlogmask(LOG_UPTO(LOG_ERR)); /* timer initialization */ rtadvd_timer_init(); #ifndef HAVE_ARC4RANDOM /* random value initialization */ #ifdef __FreeBSD__ srandomdev(); #else srandom((unsigned long)time(NULL)); #endif #endif pfh = pidfile_open(pidfilename, 0600, &otherpid); if (pfh == NULL) { if (errno == EEXIST) errx(1, "%s already running, pid: %d", getprogname(), otherpid); syslog(LOG_ERR, "failed to open the pid file %s, run anyway.", pidfilename); } if (!fflag) daemon(1, 0); sock_open(&sock); update_ifinfo(&ifilist, UPDATE_IFINFO_ALL); for (i = 0; i < argc; i++) update_persist_ifinfo(&ifilist, argv[i]); csock_open(&ctrlsock, S_IRUSR|S_IWUSR|S_IRGRP|S_IROTH); if (ctrlsock.si_fd == -1) { syslog(LOG_ERR, "cannot open control socket: %s", strerror(errno)); exit(1); } /* record the current PID */ pid = getpid(); pidfile_write(pfh); set[PFD_RAWSOCK].fd = sock.si_fd; set[PFD_RAWSOCK].events = POLLIN; if (sflag == 0) { rtsock_open(&rtsock); set[PFD_RTSOCK].fd = rtsock.si_fd; set[PFD_RTSOCK].events = POLLIN; } else set[PFD_RTSOCK].fd = -1; set[PFD_CSOCK].fd = ctrlsock.si_fd; set[PFD_CSOCK].events = POLLIN; signal(SIGTERM, set_do_shutdown); signal(SIGINT, set_do_shutdown); signal(SIGHUP, set_do_reload); error = csock_listen(&ctrlsock); if (error) { syslog(LOG_ERR, "cannot listen control socket: %s", strerror(errno)); exit(1); } /* load configuration file */ set_do_reload(0); while (1) { if (is_do_shutdown()) rtadvd_shutdown(); if (is_do_reload()) { loadconfig_ifname(reload_ifname()); if (reload_ifname() == NULL) syslog(LOG_INFO, "configuration file reloaded."); else syslog(LOG_INFO, "configuration file for %s reloaded.", reload_ifname()); reset_do_reload(); } /* timeout handler update for active interfaces */ rtadvd_update_timeout_handler(); /* timer expiration check and reset the timer */ timeout = rtadvd_check_timer(); if (timeout != NULL) { syslog(LOG_DEBUG, "<%s> set timer to %ld:%ld. waiting for " "inputs or timeout", __func__, (long int)timeout->tv_sec, - (long int)timeout->tv_usec); + (long int)timeout->tv_nsec / 1000); } else { syslog(LOG_DEBUG, "<%s> there's no timer. waiting for inputs", __func__); } if ((i = poll(set, sizeof(set)/sizeof(set[0]), timeout ? (timeout->tv_sec * 1000 + - timeout->tv_usec / 1000) : INFTIM)) < 0) { + timeout->tv_nsec / 1000 / 1000) : INFTIM)) < 0) { /* EINTR would occur if a signal was delivered */ if (errno != EINTR) syslog(LOG_ERR, "poll() failed: %s", strerror(errno)); continue; } if (i == 0) /* timeout */ continue; if (rtsock.si_fd != -1 && set[PFD_RTSOCK].revents & POLLIN) rtmsg_input(&rtsock); if (set[PFD_RAWSOCK].revents & POLLIN) rtadvd_input(&sock); if (set[PFD_CSOCK].revents & POLLIN) { int fd; fd = csock_accept(&ctrlsock); if (fd == -1) syslog(LOG_ERR, "cannot accept() control socket: %s", strerror(errno)); else { cm_handler_server(fd); close(fd); } } } exit(0); /* NOTREACHED */ } static void rtadvd_shutdown(void) { struct ifinfo *ifi; struct rainfo *rai; struct rdnss *rdn; struct dnssl *dns; if (wait_shutdown) { syslog(LOG_INFO, "waiting expiration of the all RA timers."); TAILQ_FOREACH(ifi, &ifilist, ifi_next) { /* * Ignore !IFF_UP interfaces in waiting for shutdown. */ if (!(ifi->ifi_flags & IFF_UP) && ifi->ifi_ra_timer != NULL) { ifi->ifi_state = IFI_STATE_UNCONFIGURED; rtadvd_remove_timer(ifi->ifi_ra_timer); ifi->ifi_ra_timer = NULL; syslog(LOG_DEBUG, "<%s> %s(idx=%d) is down. " "Timer removed and marked as UNCONFIGURED.", __func__, ifi->ifi_ifname, ifi->ifi_ifindex); } } TAILQ_FOREACH(ifi, &ifilist, ifi_next) { if (ifi->ifi_ra_timer != NULL) break; } if (ifi == NULL) { syslog(LOG_NOTICE, "gracefully terminated."); exit(0); } sleep(1); return; } syslog(LOG_DEBUG, "<%s> cease to be an advertising router", __func__); wait_shutdown = 1; TAILQ_FOREACH(rai, &railist, rai_next) { rai->rai_lifetime = 0; TAILQ_FOREACH(rdn, &rai->rai_rdnss, rd_next) rdn->rd_ltime = 0; TAILQ_FOREACH(dns, &rai->rai_dnssl, dn_next) dns->dn_ltime = 0; } TAILQ_FOREACH(ifi, &ifilist, ifi_next) { if (!ifi->ifi_persist) continue; if (ifi->ifi_state == IFI_STATE_UNCONFIGURED) continue; if (ifi->ifi_ra_timer == NULL) continue; if (ifi->ifi_ra_lastsent.tv_sec == 0 && - ifi->ifi_ra_lastsent.tv_usec == 0 && + ifi->ifi_ra_lastsent.tv_nsec == 0 && ifi->ifi_ra_timer != NULL) { /* * When RA configured but never sent, * ignore the IF immediately. */ rtadvd_remove_timer(ifi->ifi_ra_timer); ifi->ifi_ra_timer = NULL; ifi->ifi_state = IFI_STATE_UNCONFIGURED; continue; } ifi->ifi_state = IFI_STATE_TRANSITIVE; /* Mark as the shut-down state. */ ifi->ifi_rainfo_trans = ifi->ifi_rainfo; ifi->ifi_rainfo = NULL; ifi->ifi_burstcount = MAX_FINAL_RTR_ADVERTISEMENTS; ifi->ifi_burstinterval = MIN_DELAY_BETWEEN_RAS; ra_timer_update(ifi, &ifi->ifi_ra_timer->rat_tm); rtadvd_set_timer(&ifi->ifi_ra_timer->rat_tm, ifi->ifi_ra_timer); } syslog(LOG_NOTICE, "final RA transmission started."); pidfile_remove(pfh); csock_close(&ctrlsock); } static void rtmsg_input(struct sockinfo *s) { int n, type, ifindex = 0, plen; size_t len; char msg[2048], *next, *lim; char ifname[IFNAMSIZ]; struct if_announcemsghdr *ifan; struct rt_msghdr *rtm; struct prefix *pfx; struct rainfo *rai; struct in6_addr *addr; struct ifinfo *ifi; char addrbuf[INET6_ADDRSTRLEN]; int prefixchange = 0; if (s == NULL) { syslog(LOG_ERR, "<%s> internal error", __func__); exit(1); } n = read(s->si_fd, msg, sizeof(msg)); rtm = (struct rt_msghdr *)msg; syslog(LOG_DEBUG, "<%s> received a routing message " "(type = %d, len = %d)", __func__, rtm->rtm_type, n); if (n > rtm->rtm_msglen) { /* * This usually won't happen for messages received on * a routing socket. */ syslog(LOG_DEBUG, "<%s> received data length is larger than " "1st routing message len. multiple messages? " "read %d bytes, but 1st msg len = %d", __func__, n, rtm->rtm_msglen); #if 0 /* adjust length */ n = rtm->rtm_msglen; #endif } lim = msg + n; for (next = msg; next < lim; next += len) { int oldifflags; next = get_next_msg(next, lim, 0, &len, RTADV_TYPE2BITMASK(RTM_ADD) | RTADV_TYPE2BITMASK(RTM_DELETE) | RTADV_TYPE2BITMASK(RTM_NEWADDR) | RTADV_TYPE2BITMASK(RTM_DELADDR) | RTADV_TYPE2BITMASK(RTM_IFINFO) | RTADV_TYPE2BITMASK(RTM_IFANNOUNCE)); if (len == 0) break; type = ((struct rt_msghdr *)next)->rtm_type; switch (type) { case RTM_ADD: case RTM_DELETE: ifindex = get_rtm_ifindex(next); break; case RTM_NEWADDR: case RTM_DELADDR: ifindex = (int)((struct ifa_msghdr *)next)->ifam_index; break; case RTM_IFINFO: ifindex = (int)((struct if_msghdr *)next)->ifm_index; break; case RTM_IFANNOUNCE: ifan = (struct if_announcemsghdr *)next; switch (ifan->ifan_what) { case IFAN_ARRIVAL: case IFAN_DEPARTURE: break; default: syslog(LOG_DEBUG, "<%s:%d> unknown ifan msg (ifan_what=%d)", __func__, __LINE__, ifan->ifan_what); continue; } syslog(LOG_DEBUG, "<%s>: if_announcemsg (idx=%d:%d)", __func__, ifan->ifan_index, ifan->ifan_what); switch (ifan->ifan_what) { case IFAN_ARRIVAL: syslog(LOG_NOTICE, "interface added (idx=%d)", ifan->ifan_index); update_ifinfo(&ifilist, ifan->ifan_index); loadconfig_index(ifan->ifan_index); break; case IFAN_DEPARTURE: syslog(LOG_NOTICE, "interface removed (idx=%d)", ifan->ifan_index); rm_ifinfo_index(ifan->ifan_index); /* Clear ifi_ifindex */ TAILQ_FOREACH(ifi, &ifilist, ifi_next) { if (ifi->ifi_ifindex == ifan->ifan_index) { ifi->ifi_ifindex = 0; break; } } update_ifinfo(&ifilist, ifan->ifan_index); break; } continue; default: /* should not reach here */ syslog(LOG_DEBUG, "<%s:%d> unknown rtmsg %d on %s", __func__, __LINE__, type, if_indextoname(ifindex, ifname)); continue; } ifi = if_indextoifinfo(ifindex); if (ifi == NULL) { syslog(LOG_DEBUG, "<%s> ifinfo not found for idx=%d. Why?", __func__, ifindex); continue; } rai = ifi->ifi_rainfo; if (rai == NULL) { syslog(LOG_DEBUG, "<%s> route changed on " "non advertising interface(%s)", __func__, ifi->ifi_ifname); continue; } oldifflags = ifi->ifi_flags; /* init ifflags because it may have changed */ update_ifinfo(&ifilist, ifindex); switch (type) { case RTM_ADD: if (sflag) break; /* we aren't interested in prefixes */ addr = get_addr(msg); plen = get_prefixlen(msg); /* sanity check for plen */ /* as RFC2373, prefixlen is at least 4 */ if (plen < 4 || plen > 127) { syslog(LOG_INFO, "<%s> new interface route's" "plen %d is invalid for a prefix", __func__, plen); break; } pfx = find_prefix(rai, addr, plen); if (pfx) { if (pfx->pfx_timer) { /* * If the prefix has been invalidated, * make it available again. */ update_prefix(pfx); prefixchange = 1; } else syslog(LOG_DEBUG, "<%s> new prefix(%s/%d) " "added on %s, " "but it was already in list", __func__, inet_ntop(AF_INET6, addr, (char *)addrbuf, sizeof(addrbuf)), plen, ifi->ifi_ifname); break; } make_prefix(rai, ifindex, addr, plen); prefixchange = 1; break; case RTM_DELETE: if (sflag) break; addr = get_addr(msg); plen = get_prefixlen(msg); /* sanity check for plen */ /* as RFC2373, prefixlen is at least 4 */ if (plen < 4 || plen > 127) { syslog(LOG_INFO, "<%s> deleted interface route's " "plen %d is invalid for a prefix", __func__, plen); break; } pfx = find_prefix(rai, addr, plen); if (pfx == NULL) { syslog(LOG_DEBUG, "<%s> prefix(%s/%d) was deleted on %s, " "but it was not in list", __func__, inet_ntop(AF_INET6, addr, (char *)addrbuf, sizeof(addrbuf)), plen, ifi->ifi_ifname); break; } invalidate_prefix(pfx); prefixchange = 1; break; case RTM_NEWADDR: case RTM_DELADDR: case RTM_IFINFO: break; default: /* should not reach here */ syslog(LOG_DEBUG, "<%s:%d> unknown rtmsg %d on %s", __func__, __LINE__, type, if_indextoname(ifindex, ifname)); return; } /* check if an interface flag is changed */ if ((oldifflags & IFF_UP) && /* UP to DOWN */ !(ifi->ifi_flags & IFF_UP)) { syslog(LOG_NOTICE, "ifi_ifname); rtadvd_remove_timer(ifi->ifi_ra_timer); ifi->ifi_ra_timer = NULL; } else if (!(oldifflags & IFF_UP) && /* DOWN to UP */ (ifi->ifi_flags & IFF_UP)) { syslog(LOG_NOTICE, "interface %s becomes up. restart timer.", ifi->ifi_ifname); ifi->ifi_state = IFI_STATE_TRANSITIVE; ifi->ifi_burstcount = MAX_INITIAL_RTR_ADVERTISEMENTS; ifi->ifi_burstinterval = MAX_INITIAL_RTR_ADVERT_INTERVAL; ifi->ifi_ra_timer = rtadvd_add_timer(ra_timeout, ra_timer_update, ifi, ifi); ra_timer_update(ifi, &ifi->ifi_ra_timer->rat_tm); rtadvd_set_timer(&ifi->ifi_ra_timer->rat_tm, ifi->ifi_ra_timer); } else if (prefixchange && (ifi->ifi_flags & IFF_UP)) { /* * An advertised prefix has been added or invalidated. * Will notice the change in a short delay. */ set_short_delay(ifi); } } return; } void rtadvd_input(struct sockinfo *s) { ssize_t i; int *hlimp = NULL; #ifdef OLDRAWSOCKET struct ip6_hdr *ip; #endif struct icmp6_hdr *icp; int ifindex = 0; struct cmsghdr *cm; struct in6_pktinfo *pi = NULL; char ntopbuf[INET6_ADDRSTRLEN], ifnamebuf[IFNAMSIZ]; struct in6_addr dst = in6addr_any; struct ifinfo *ifi; syslog(LOG_DEBUG, "<%s> enter", __func__); if (s == NULL) { syslog(LOG_ERR, "<%s> internal error", __func__); exit(1); } /* * Get message. We reset msg_controllen since the field could * be modified if we had received a message before setting * receive options. */ rcvmhdr.msg_controllen = rcvcmsgbuflen; if ((i = recvmsg(s->si_fd, &rcvmhdr, 0)) < 0) return; /* extract optional information via Advanced API */ for (cm = (struct cmsghdr *)CMSG_FIRSTHDR(&rcvmhdr); cm; cm = (struct cmsghdr *)CMSG_NXTHDR(&rcvmhdr, cm)) { if (cm->cmsg_level == IPPROTO_IPV6 && cm->cmsg_type == IPV6_PKTINFO && cm->cmsg_len == CMSG_LEN(sizeof(struct in6_pktinfo))) { pi = (struct in6_pktinfo *)(CMSG_DATA(cm)); ifindex = pi->ipi6_ifindex; dst = pi->ipi6_addr; } if (cm->cmsg_level == IPPROTO_IPV6 && cm->cmsg_type == IPV6_HOPLIMIT && cm->cmsg_len == CMSG_LEN(sizeof(int))) hlimp = (int *)CMSG_DATA(cm); } if (ifindex == 0) { syslog(LOG_ERR, "failed to get receiving interface"); return; } if (hlimp == NULL) { syslog(LOG_ERR, "failed to get receiving hop limit"); return; } /* * If we happen to receive data on an interface which is now gone * or down, just discard the data. */ ifi = if_indextoifinfo(pi->ipi6_ifindex); if (ifi == NULL || !(ifi->ifi_flags & IFF_UP)) { syslog(LOG_INFO, "<%s> received data on a disabled interface (%s)", __func__, (ifi == NULL) ? "[gone]" : ifi->ifi_ifname); return; } #ifdef OLDRAWSOCKET if ((size_t)i < sizeof(struct ip6_hdr) + sizeof(struct icmp6_hdr)) { syslog(LOG_ERR, "packet size(%d) is too short", i); return; } ip = (struct ip6_hdr *)rcvmhdr.msg_iov[0].iov_base; icp = (struct icmp6_hdr *)(ip + 1); /* XXX: ext. hdr? */ #else if ((size_t)i < sizeof(struct icmp6_hdr)) { syslog(LOG_ERR, "packet size(%zd) is too short", i); return; } icp = (struct icmp6_hdr *)rcvmhdr.msg_iov[0].iov_base; #endif switch (icp->icmp6_type) { case ND_ROUTER_SOLICIT: /* * Message verification - RFC 4861 6.1.1 * XXX: these checks must be done in the kernel as well, * but we can't completely rely on them. */ if (*hlimp != 255) { syslog(LOG_NOTICE, "RS with invalid hop limit(%d) " "received from %s on %s", *hlimp, inet_ntop(AF_INET6, &rcvfrom.sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } if (icp->icmp6_code) { syslog(LOG_NOTICE, "RS with invalid ICMP6 code(%d) " "received from %s on %s", icp->icmp6_code, inet_ntop(AF_INET6, &rcvfrom.sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } if ((size_t)i < sizeof(struct nd_router_solicit)) { syslog(LOG_NOTICE, "RS from %s on %s does not have enough " "length (len = %zd)", inet_ntop(AF_INET6, &rcvfrom.sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf), i); return; } rs_input(i, (struct nd_router_solicit *)icp, pi, &rcvfrom); break; case ND_ROUTER_ADVERT: /* * Message verification - RFC 4861 6.1.2 * XXX: there's the same dilemma as above... */ if (!IN6_IS_ADDR_LINKLOCAL(&rcvfrom.sin6_addr)) { syslog(LOG_NOTICE, "RA with non-linklocal source address " "received from %s on %s", inet_ntop(AF_INET6, &rcvfrom.sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } if (*hlimp != 255) { syslog(LOG_NOTICE, "RA with invalid hop limit(%d) " "received from %s on %s", *hlimp, inet_ntop(AF_INET6, &rcvfrom.sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } if (icp->icmp6_code) { syslog(LOG_NOTICE, "RA with invalid ICMP6 code(%d) " "received from %s on %s", icp->icmp6_code, inet_ntop(AF_INET6, &rcvfrom.sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } if ((size_t)i < sizeof(struct nd_router_advert)) { syslog(LOG_NOTICE, "RA from %s on %s does not have enough " "length (len = %zd)", inet_ntop(AF_INET6, &rcvfrom.sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf), i); return; } ra_input(i, (struct nd_router_advert *)icp, pi, &rcvfrom); break; case ICMP6_ROUTER_RENUMBERING: if (mcastif == NULL) { syslog(LOG_ERR, "received a router renumbering " "message, but not allowed to be accepted"); break; } rr_input(i, (struct icmp6_router_renum *)icp, pi, &rcvfrom, &dst); break; default: /* * Note that this case is POSSIBLE, especially just * after invocation of the daemon. This is because we * could receive message after opening the socket and * before setting ICMP6 type filter(see sock_open()). */ syslog(LOG_ERR, "invalid icmp type(%d)", icp->icmp6_type); return; } return; } static void rs_input(int len, struct nd_router_solicit *rs, struct in6_pktinfo *pi, struct sockaddr_in6 *from) { char ntopbuf[INET6_ADDRSTRLEN]; char ifnamebuf[IFNAMSIZ]; union nd_opt ndopts; struct rainfo *rai; struct ifinfo *ifi; struct soliciter *sol; syslog(LOG_DEBUG, "<%s> RS received from %s on %s", __func__, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); /* ND option check */ memset(&ndopts, 0, sizeof(ndopts)); TAILQ_INIT(&ndopts.opt_list); if (nd6_options((struct nd_opt_hdr *)(rs + 1), len - sizeof(struct nd_router_solicit), &ndopts, NDOPT_FLAG_SRCLINKADDR)) { syslog(LOG_INFO, "<%s> ND option check failed for an RS from %s on %s", __func__, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } /* * If the IP source address is the unspecified address, there * must be no source link-layer address option in the message. * (RFC 4861 6.1.1) */ if (IN6_IS_ADDR_UNSPECIFIED(&from->sin6_addr) && ndopts.opt_src_lladdr) { syslog(LOG_INFO, "<%s> RS from unspecified src on %s has a link-layer" " address option", __func__, if_indextoname(pi->ipi6_ifindex, ifnamebuf)); goto done; } ifi = if_indextoifinfo(pi->ipi6_ifindex); if (ifi == NULL) { syslog(LOG_INFO, "<%s> if (idx=%d) not found. Why?", __func__, pi->ipi6_ifindex); goto done; } rai = ifi->ifi_rainfo; if (rai == NULL) { syslog(LOG_INFO, "<%s> RS received on non advertising interface(%s)", __func__, if_indextoname(pi->ipi6_ifindex, ifnamebuf)); goto done; } rai->rai_ifinfo->ifi_rsinput++; /* * Decide whether to send RA according to the rate-limit * consideration. */ /* record sockaddr waiting for RA, if possible */ sol = (struct soliciter *)malloc(sizeof(*sol)); if (sol) { sol->sol_addr = *from; /* XXX RFC 2553 need clarification on flowinfo */ sol->sol_addr.sin6_flowinfo = 0; TAILQ_INSERT_TAIL(&rai->rai_soliciter, sol, sol_next); } /* * If there is already a waiting RS packet, don't * update the timer. */ if (ifi->ifi_rs_waitcount++) goto done; set_short_delay(ifi); done: free_ndopts(&ndopts); return; } static void set_short_delay(struct ifinfo *ifi) { long delay; /* must not be greater than 1000000 */ - struct timeval interval, now, min_delay, tm_tmp, *rest; + struct timespec interval, now, min_delay, tm_tmp, *rest; if (ifi->ifi_ra_timer == NULL) return; /* * Compute a random delay. If the computed value * corresponds to a time later than the time the next * multicast RA is scheduled to be sent, ignore the random * delay and send the advertisement at the * already-scheduled time. RFC 4861 6.2.6 */ #ifdef HAVE_ARC4RANDOM delay = arc4random_uniform(MAX_RA_DELAY_TIME); #else delay = random() % MAX_RA_DELAY_TIME; #endif interval.tv_sec = 0; - interval.tv_usec = delay; + interval.tv_nsec = delay * 1000; rest = rtadvd_timer_rest(ifi->ifi_ra_timer); - if (TIMEVAL_LT(rest, &interval)) { + if (TS_CMP(rest, &interval, <)) { syslog(LOG_DEBUG, "<%s> random delay is larger than " "the rest of the current timer", __func__); interval = *rest; } /* * If we sent a multicast Router Advertisement within * the last MIN_DELAY_BETWEEN_RAS seconds, schedule * the advertisement to be sent at a time corresponding to * MIN_DELAY_BETWEEN_RAS plus the random value after the * previous advertisement was sent. */ - gettimeofday(&now, NULL); - TIMEVAL_SUB(&now, &ifi->ifi_ra_lastsent, &tm_tmp); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); + TS_SUB(&now, &ifi->ifi_ra_lastsent, &tm_tmp); min_delay.tv_sec = MIN_DELAY_BETWEEN_RAS; - min_delay.tv_usec = 0; - if (TIMEVAL_LT(&tm_tmp, &min_delay)) { - TIMEVAL_SUB(&min_delay, &tm_tmp, &min_delay); - TIMEVAL_ADD(&min_delay, &interval, &interval); + min_delay.tv_nsec = 0; + if (TS_CMP(&tm_tmp, &min_delay, <)) { + TS_SUB(&min_delay, &tm_tmp, &min_delay); + TS_ADD(&min_delay, &interval, &interval); } rtadvd_set_timer(&interval, ifi->ifi_ra_timer); } static int check_accept_rtadv(int idx) { struct ifinfo *ifi; TAILQ_FOREACH(ifi, &ifilist, ifi_next) { if (ifi->ifi_ifindex == idx) break; } if (ifi == NULL) { syslog(LOG_DEBUG, "<%s> if (idx=%d) not found. Why?", __func__, idx); return (0); } #if (__FreeBSD_version < 900000) /* * RA_RECV: !ip6.forwarding && ip6.accept_rtadv * RA_SEND: ip6.forwarding */ return ((getinet6sysctl(IPV6CTL_FORWARDING) == 0) && (getinet6sysctl(IPV6CTL_ACCEPT_RTADV) == 1)); #else /* * RA_RECV: ND6_IFF_ACCEPT_RTADV * RA_SEND: ip6.forwarding */ if (update_ifinfo_nd_flags(ifi) != 0) { syslog(LOG_ERR, "cannot get nd6 flags (idx=%d)", idx); return (0); } return (ifi->ifi_nd_flags & ND6_IFF_ACCEPT_RTADV); #endif } static void ra_input(int len, struct nd_router_advert *nra, struct in6_pktinfo *pi, struct sockaddr_in6 *from) { struct rainfo *rai; struct ifinfo *ifi; char ntopbuf[INET6_ADDRSTRLEN]; char ifnamebuf[IFNAMSIZ]; union nd_opt ndopts; const char *on_off[] = {"OFF", "ON"}; uint32_t reachabletime, retranstimer, mtu; int inconsistent = 0; int error; syslog(LOG_DEBUG, "<%s> RA received from %s on %s", __func__, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); /* ND option check */ memset(&ndopts, 0, sizeof(ndopts)); TAILQ_INIT(&ndopts.opt_list); error = nd6_options((struct nd_opt_hdr *)(nra + 1), len - sizeof(struct nd_router_advert), &ndopts, NDOPT_FLAG_SRCLINKADDR | NDOPT_FLAG_PREFIXINFO | NDOPT_FLAG_MTU | NDOPT_FLAG_RDNSS | NDOPT_FLAG_DNSSL); if (error) { syslog(LOG_INFO, "<%s> ND option check failed for an RA from %s on %s", __func__, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } /* * RA consistency check according to RFC 4861 6.2.7 */ ifi = if_indextoifinfo(pi->ipi6_ifindex); if (ifi->ifi_rainfo == NULL) { syslog(LOG_INFO, "<%s> received RA from %s on non-advertising" " interface(%s)", __func__, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); goto done; } rai = ifi->ifi_rainfo; ifi->ifi_rainput++; syslog(LOG_DEBUG, "<%s> ifi->ifi_rainput = %" PRIu64, __func__, ifi->ifi_rainput); /* Cur Hop Limit value */ if (nra->nd_ra_curhoplimit && rai->rai_hoplimit && nra->nd_ra_curhoplimit != rai->rai_hoplimit) { syslog(LOG_NOTICE, "CurHopLimit inconsistent on %s:" " %d from %s, %d from us", ifi->ifi_ifname, nra->nd_ra_curhoplimit, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), rai->rai_hoplimit); inconsistent++; } /* M flag */ if ((nra->nd_ra_flags_reserved & ND_RA_FLAG_MANAGED) != rai->rai_managedflg) { syslog(LOG_NOTICE, "M flag inconsistent on %s:" " %s from %s, %s from us", ifi->ifi_ifname, on_off[!rai->rai_managedflg], inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), on_off[rai->rai_managedflg]); inconsistent++; } /* O flag */ if ((nra->nd_ra_flags_reserved & ND_RA_FLAG_OTHER) != rai->rai_otherflg) { syslog(LOG_NOTICE, "O flag inconsistent on %s:" " %s from %s, %s from us", ifi->ifi_ifname, on_off[!rai->rai_otherflg], inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), on_off[rai->rai_otherflg]); inconsistent++; } /* Reachable Time */ reachabletime = ntohl(nra->nd_ra_reachable); if (reachabletime && rai->rai_reachabletime && reachabletime != rai->rai_reachabletime) { syslog(LOG_NOTICE, "ReachableTime inconsistent on %s:" " %d from %s, %d from us", ifi->ifi_ifname, reachabletime, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), rai->rai_reachabletime); inconsistent++; } /* Retrans Timer */ retranstimer = ntohl(nra->nd_ra_retransmit); if (retranstimer && rai->rai_retranstimer && retranstimer != rai->rai_retranstimer) { syslog(LOG_NOTICE, "RetranceTimer inconsistent on %s:" " %d from %s, %d from us", ifi->ifi_ifname, retranstimer, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), rai->rai_retranstimer); inconsistent++; } /* Values in the MTU options */ if (ndopts.opt_mtu) { mtu = ntohl(ndopts.opt_mtu->nd_opt_mtu_mtu); if (mtu && rai->rai_linkmtu && mtu != rai->rai_linkmtu) { syslog(LOG_NOTICE, "MTU option value inconsistent on %s:" " %d from %s, %d from us", ifi->ifi_ifname, mtu, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), rai->rai_linkmtu); inconsistent++; } } /* Preferred and Valid Lifetimes for prefixes */ { struct nd_optlist *nol; if (ndopts.opt_pi) if (prefix_check(ndopts.opt_pi, rai, from)) inconsistent++; TAILQ_FOREACH(nol, &ndopts.opt_list, nol_next) if (prefix_check((struct nd_opt_prefix_info *)nol->nol_opt, rai, from)) inconsistent++; } if (inconsistent) ifi->ifi_rainconsistent++; done: free_ndopts(&ndopts); return; } /* return a non-zero value if the received prefix is inconsitent with ours */ static int prefix_check(struct nd_opt_prefix_info *pinfo, struct rainfo *rai, struct sockaddr_in6 *from) { struct ifinfo *ifi; uint32_t preferred_time, valid_time; struct prefix *pfx; int inconsistent = 0; char ntopbuf[INET6_ADDRSTRLEN]; char prefixbuf[INET6_ADDRSTRLEN]; - struct timeval now; + struct timespec now; #if 0 /* impossible */ if (pinfo->nd_opt_pi_type != ND_OPT_PREFIX_INFORMATION) return (0); #endif ifi = rai->rai_ifinfo; /* * log if the adveritsed prefix has link-local scope(sanity check?) */ if (IN6_IS_ADDR_LINKLOCAL(&pinfo->nd_opt_pi_prefix)) syslog(LOG_INFO, "<%s> link-local prefix %s/%d is advertised " "from %s on %s", __func__, inet_ntop(AF_INET6, &pinfo->nd_opt_pi_prefix, prefixbuf, sizeof(prefixbuf)), pinfo->nd_opt_pi_prefix_len, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), ifi->ifi_ifname); if ((pfx = find_prefix(rai, &pinfo->nd_opt_pi_prefix, pinfo->nd_opt_pi_prefix_len)) == NULL) { syslog(LOG_INFO, "<%s> prefix %s/%d from %s on %s is not in our list", __func__, inet_ntop(AF_INET6, &pinfo->nd_opt_pi_prefix, prefixbuf, sizeof(prefixbuf)), pinfo->nd_opt_pi_prefix_len, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), ifi->ifi_ifname); return (0); } preferred_time = ntohl(pinfo->nd_opt_pi_preferred_time); if (pfx->pfx_pltimeexpire) { /* * The lifetime is decremented in real time, so we should * compare the expiration time. * (RFC 2461 Section 6.2.7.) * XXX: can we really expect that all routers on the link * have synchronized clocks? */ - gettimeofday(&now, NULL); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); preferred_time += now.tv_sec; if (!pfx->pfx_timer && rai->rai_clockskew && abs(preferred_time - pfx->pfx_pltimeexpire) > rai->rai_clockskew) { syslog(LOG_INFO, "<%s> preferred lifetime for %s/%d" " (decr. in real time) inconsistent on %s:" " %" PRIu32 " from %s, %" PRIu32 " from us", __func__, inet_ntop(AF_INET6, &pinfo->nd_opt_pi_prefix, prefixbuf, sizeof(prefixbuf)), pinfo->nd_opt_pi_prefix_len, ifi->ifi_ifname, preferred_time, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), pfx->pfx_pltimeexpire); inconsistent++; } } else if (!pfx->pfx_timer && preferred_time != pfx->pfx_preflifetime) syslog(LOG_INFO, "<%s> preferred lifetime for %s/%d" " inconsistent on %s:" " %d from %s, %d from us", __func__, inet_ntop(AF_INET6, &pinfo->nd_opt_pi_prefix, prefixbuf, sizeof(prefixbuf)), pinfo->nd_opt_pi_prefix_len, ifi->ifi_ifname, preferred_time, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), pfx->pfx_preflifetime); valid_time = ntohl(pinfo->nd_opt_pi_valid_time); if (pfx->pfx_vltimeexpire) { - gettimeofday(&now, NULL); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); valid_time += now.tv_sec; if (!pfx->pfx_timer && rai->rai_clockskew && abs(valid_time - pfx->pfx_vltimeexpire) > rai->rai_clockskew) { syslog(LOG_INFO, "<%s> valid lifetime for %s/%d" " (decr. in real time) inconsistent on %s:" " %d from %s, %" PRIu32 " from us", __func__, inet_ntop(AF_INET6, &pinfo->nd_opt_pi_prefix, prefixbuf, sizeof(prefixbuf)), pinfo->nd_opt_pi_prefix_len, ifi->ifi_ifname, preferred_time, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), pfx->pfx_vltimeexpire); inconsistent++; } } else if (!pfx->pfx_timer && valid_time != pfx->pfx_validlifetime) { syslog(LOG_INFO, "<%s> valid lifetime for %s/%d" " inconsistent on %s:" " %d from %s, %d from us", __func__, inet_ntop(AF_INET6, &pinfo->nd_opt_pi_prefix, prefixbuf, sizeof(prefixbuf)), pinfo->nd_opt_pi_prefix_len, ifi->ifi_ifname, valid_time, inet_ntop(AF_INET6, &from->sin6_addr, ntopbuf, sizeof(ntopbuf)), pfx->pfx_validlifetime); inconsistent++; } return (inconsistent); } struct prefix * find_prefix(struct rainfo *rai, struct in6_addr *prefix, int plen) { struct prefix *pfx; int bytelen, bitlen; char bitmask; TAILQ_FOREACH(pfx, &rai->rai_prefix, pfx_next) { if (plen != pfx->pfx_prefixlen) continue; bytelen = plen / 8; bitlen = plen % 8; bitmask = 0xff << (8 - bitlen); if (memcmp((void *)prefix, (void *)&pfx->pfx_prefix, bytelen)) continue; if (bitlen == 0 || ((prefix->s6_addr[bytelen] & bitmask) == (pfx->pfx_prefix.s6_addr[bytelen] & bitmask))) { return (pfx); } } return (NULL); } /* check if p0/plen0 matches p1/plen1; return 1 if matches, otherwise 0. */ int prefix_match(struct in6_addr *p0, int plen0, struct in6_addr *p1, int plen1) { int bytelen, bitlen; char bitmask; if (plen0 < plen1) return (0); bytelen = plen1 / 8; bitlen = plen1 % 8; bitmask = 0xff << (8 - bitlen); if (memcmp((void *)p0, (void *)p1, bytelen)) return (0); if (bitlen == 0 || ((p0->s6_addr[bytelen] & bitmask) == (p1->s6_addr[bytelen] & bitmask))) { return (1); } return (0); } static int nd6_options(struct nd_opt_hdr *hdr, int limit, union nd_opt *ndopts, uint32_t optflags) { int optlen = 0; for (; limit > 0; limit -= optlen) { if ((size_t)limit < sizeof(struct nd_opt_hdr)) { syslog(LOG_INFO, "<%s> short option header", __func__); goto bad; } hdr = (struct nd_opt_hdr *)((caddr_t)hdr + optlen); if (hdr->nd_opt_len == 0) { syslog(LOG_INFO, "<%s> bad ND option length(0) (type = %d)", __func__, hdr->nd_opt_type); goto bad; } optlen = hdr->nd_opt_len << 3; if (optlen > limit) { syslog(LOG_INFO, "<%s> short option", __func__); goto bad; } if (hdr->nd_opt_type > ND_OPT_MTU && hdr->nd_opt_type != ND_OPT_RDNSS && hdr->nd_opt_type != ND_OPT_DNSSL) { syslog(LOG_INFO, "<%s> unknown ND option(type %d)", __func__, hdr->nd_opt_type); continue; } if ((ndopt_flags[hdr->nd_opt_type] & optflags) == 0) { syslog(LOG_INFO, "<%s> unexpected ND option(type %d)", __func__, hdr->nd_opt_type); continue; } /* * Option length check. Do it here for all fixed-length * options. */ switch (hdr->nd_opt_type) { case ND_OPT_MTU: if (optlen == sizeof(struct nd_opt_mtu)) break; goto skip; case ND_OPT_RDNSS: if (optlen >= 24 && (optlen - sizeof(struct nd_opt_rdnss)) % 16 == 0) break; goto skip; case ND_OPT_DNSSL: if (optlen >= 16 && (optlen - sizeof(struct nd_opt_dnssl)) % 8 == 0) break; goto skip; case ND_OPT_PREFIX_INFORMATION: if (optlen == sizeof(struct nd_opt_prefix_info)) break; skip: syslog(LOG_INFO, "<%s> invalid option length", __func__); continue; } switch (hdr->nd_opt_type) { case ND_OPT_TARGET_LINKADDR: case ND_OPT_REDIRECTED_HEADER: case ND_OPT_RDNSS: case ND_OPT_DNSSL: break; /* we don't care about these options */ case ND_OPT_SOURCE_LINKADDR: case ND_OPT_MTU: if (ndopts->opt_array[hdr->nd_opt_type]) { syslog(LOG_INFO, "<%s> duplicated ND option (type = %d)", __func__, hdr->nd_opt_type); } ndopts->opt_array[hdr->nd_opt_type] = hdr; break; case ND_OPT_PREFIX_INFORMATION: { struct nd_optlist *nol; if (ndopts->opt_pi == 0) { ndopts->opt_pi = (struct nd_opt_prefix_info *)hdr; continue; } nol = malloc(sizeof(*nol)); if (nol == NULL) { syslog(LOG_ERR, "<%s> can't allocate memory", __func__); goto bad; } nol->nol_opt = hdr; TAILQ_INSERT_TAIL(&(ndopts->opt_list), nol, nol_next); break; } default: /* impossible */ break; } } return (0); bad: free_ndopts(ndopts); return (-1); } static void free_ndopts(union nd_opt *ndopts) { struct nd_optlist *nol; while ((nol = TAILQ_FIRST(&ndopts->opt_list)) != NULL) { TAILQ_REMOVE(&ndopts->opt_list, nol, nol_next); free(nol); } } void sock_open(struct sockinfo *s) { struct icmp6_filter filt; int on; /* XXX: should be max MTU attached to the node */ static char answer[1500]; syslog(LOG_DEBUG, "<%s> enter", __func__); if (s == NULL) { syslog(LOG_ERR, "<%s> internal error", __func__); exit(1); } rcvcmsgbuflen = CMSG_SPACE(sizeof(struct in6_pktinfo)) + CMSG_SPACE(sizeof(int)); rcvcmsgbuf = (char *)malloc(rcvcmsgbuflen); if (rcvcmsgbuf == NULL) { syslog(LOG_ERR, "<%s> not enough core", __func__); exit(1); } sndcmsgbuflen = CMSG_SPACE(sizeof(struct in6_pktinfo)) + CMSG_SPACE(sizeof(int)); sndcmsgbuf = (char *)malloc(sndcmsgbuflen); if (sndcmsgbuf == NULL) { syslog(LOG_ERR, "<%s> not enough core", __func__); exit(1); } if ((s->si_fd = socket(AF_INET6, SOCK_RAW, IPPROTO_ICMPV6)) < 0) { syslog(LOG_ERR, "<%s> socket: %s", __func__, strerror(errno)); exit(1); } /* specify to tell receiving interface */ on = 1; if (setsockopt(s->si_fd, IPPROTO_IPV6, IPV6_RECVPKTINFO, &on, sizeof(on)) < 0) { syslog(LOG_ERR, "<%s> IPV6_RECVPKTINFO: %s", __func__, strerror(errno)); exit(1); } on = 1; /* specify to tell value of hoplimit field of received IP6 hdr */ if (setsockopt(s->si_fd, IPPROTO_IPV6, IPV6_RECVHOPLIMIT, &on, sizeof(on)) < 0) { syslog(LOG_ERR, "<%s> IPV6_RECVHOPLIMIT: %s", __func__, strerror(errno)); exit(1); } ICMP6_FILTER_SETBLOCKALL(&filt); ICMP6_FILTER_SETPASS(ND_ROUTER_SOLICIT, &filt); ICMP6_FILTER_SETPASS(ND_ROUTER_ADVERT, &filt); if (mcastif != NULL) ICMP6_FILTER_SETPASS(ICMP6_ROUTER_RENUMBERING, &filt); if (setsockopt(s->si_fd, IPPROTO_ICMPV6, ICMP6_FILTER, &filt, sizeof(filt)) < 0) { syslog(LOG_ERR, "<%s> IICMP6_FILTER: %s", __func__, strerror(errno)); exit(1); } /* initialize msghdr for receiving packets */ rcviov[0].iov_base = (caddr_t)answer; rcviov[0].iov_len = sizeof(answer); rcvmhdr.msg_name = (caddr_t)&rcvfrom; rcvmhdr.msg_namelen = sizeof(rcvfrom); rcvmhdr.msg_iov = rcviov; rcvmhdr.msg_iovlen = 1; rcvmhdr.msg_control = (caddr_t) rcvcmsgbuf; rcvmhdr.msg_controllen = rcvcmsgbuflen; /* initialize msghdr for sending packets */ sndmhdr.msg_namelen = sizeof(struct sockaddr_in6); sndmhdr.msg_iov = sndiov; sndmhdr.msg_iovlen = 1; sndmhdr.msg_control = (caddr_t)sndcmsgbuf; sndmhdr.msg_controllen = sndcmsgbuflen; return; } /* open a routing socket to watch the routing table */ static void rtsock_open(struct sockinfo *s) { if (s == NULL) { syslog(LOG_ERR, "<%s> internal error", __func__); exit(1); } if ((s->si_fd = socket(PF_ROUTE, SOCK_RAW, 0)) < 0) { syslog(LOG_ERR, "<%s> socket: %s", __func__, strerror(errno)); exit(1); } } struct ifinfo * if_indextoifinfo(int idx) { struct ifinfo *ifi; char *name, name0[IFNAMSIZ]; /* Check if the interface has a valid name or not. */ if (if_indextoname(idx, name0) == NULL) return (NULL); TAILQ_FOREACH(ifi, &ifilist, ifi_next) { if (ifi->ifi_ifindex == idx) return (ifi); } if (ifi != NULL) syslog(LOG_DEBUG, "<%s> ifi found (idx=%d)", __func__, idx); else syslog(LOG_DEBUG, "<%s> ifi not found (idx=%d)", __func__, idx); return (NULL); /* search failed */ } void ra_output(struct ifinfo *ifi) { int i; struct cmsghdr *cm; struct in6_pktinfo *pi; struct soliciter *sol; struct rainfo *rai; switch (ifi->ifi_state) { case IFI_STATE_CONFIGURED: rai = ifi->ifi_rainfo; break; case IFI_STATE_TRANSITIVE: rai = ifi->ifi_rainfo_trans; break; case IFI_STATE_UNCONFIGURED: syslog(LOG_DEBUG, "<%s> %s is unconfigured. " "Skip sending RAs.", __func__, ifi->ifi_ifname); return; default: rai = NULL; } if (rai == NULL) { syslog(LOG_DEBUG, "<%s> rainfo is NULL on %s." "Skip sending RAs.", __func__, ifi->ifi_ifname); return; } if (!(ifi->ifi_flags & IFF_UP)) { syslog(LOG_DEBUG, "<%s> %s is not up. " "Skip sending RAs.", __func__, ifi->ifi_ifname); return; } /* * Check lifetime, ACCEPT_RTADV flag, and ip6.forwarding. * * (lifetime == 0) = output * (lifetime != 0 && (check_accept_rtadv()) = no output * * Basically, hosts MUST NOT send Router Advertisement * messages at any time (RFC 4861, Section 6.2.3). However, it * would sometimes be useful to allow hosts to advertise some * parameters such as prefix information and link MTU. Thus, * we allow hosts to invoke rtadvd only when router lifetime * (on every advertising interface) is explicitly set * zero. (see also the above section) */ syslog(LOG_DEBUG, "<%s> check lifetime=%d, ACCEPT_RTADV=%d, ip6.forwarding=%d " "on %s", __func__, rai->rai_lifetime, check_accept_rtadv(ifi->ifi_ifindex), getinet6sysctl(IPV6CTL_FORWARDING), ifi->ifi_ifname); if (rai->rai_lifetime != 0) { if (getinet6sysctl(IPV6CTL_FORWARDING) == 0) { syslog(LOG_ERR, "non-zero lifetime RA " "but net.inet6.ip6.forwarding=0. " "Ignored."); return; } if (check_accept_rtadv(ifi->ifi_ifindex)) { syslog(LOG_ERR, "non-zero lifetime RA " "on RA receiving interface %s." " Ignored.", ifi->ifi_ifname); return; } } make_packet(rai); /* XXX: inefficient */ sndmhdr.msg_name = (caddr_t)&sin6_linklocal_allnodes; sndmhdr.msg_iov[0].iov_base = (caddr_t)rai->rai_ra_data; sndmhdr.msg_iov[0].iov_len = rai->rai_ra_datalen; cm = CMSG_FIRSTHDR(&sndmhdr); /* specify the outgoing interface */ cm->cmsg_level = IPPROTO_IPV6; cm->cmsg_type = IPV6_PKTINFO; cm->cmsg_len = CMSG_LEN(sizeof(struct in6_pktinfo)); pi = (struct in6_pktinfo *)CMSG_DATA(cm); memset(&pi->ipi6_addr, 0, sizeof(pi->ipi6_addr)); /*XXX*/ pi->ipi6_ifindex = ifi->ifi_ifindex; /* specify the hop limit of the packet */ { int hoplimit = 255; cm = CMSG_NXTHDR(&sndmhdr, cm); cm->cmsg_level = IPPROTO_IPV6; cm->cmsg_type = IPV6_HOPLIMIT; cm->cmsg_len = CMSG_LEN(sizeof(int)); memcpy(CMSG_DATA(cm), &hoplimit, sizeof(int)); } syslog(LOG_DEBUG, "<%s> send RA on %s, # of RS waitings = %d", __func__, ifi->ifi_ifname, ifi->ifi_rs_waitcount); i = sendmsg(sock.si_fd, &sndmhdr, 0); if (i < 0 || (size_t)i != rai->rai_ra_datalen) { if (i < 0) { syslog(LOG_ERR, "<%s> sendmsg on %s: %s", __func__, ifi->ifi_ifname, strerror(errno)); } } /* * unicast advertisements * XXX commented out. reason: though spec does not forbit it, unicast * advert does not really help */ while ((sol = TAILQ_FIRST(&rai->rai_soliciter)) != NULL) { TAILQ_REMOVE(&rai->rai_soliciter, sol, sol_next); free(sol); } /* update timestamp */ - gettimeofday(&ifi->ifi_ra_lastsent, NULL); + clock_gettime(CLOCK_MONOTONIC_FAST, &ifi->ifi_ra_lastsent); /* update counter */ ifi->ifi_rs_waitcount = 0; ifi->ifi_raoutput++; switch (ifi->ifi_state) { case IFI_STATE_CONFIGURED: if (ifi->ifi_burstcount > 0) ifi->ifi_burstcount--; break; case IFI_STATE_TRANSITIVE: ifi->ifi_burstcount--; if (ifi->ifi_burstcount == 0) { if (ifi->ifi_rainfo == ifi->ifi_rainfo_trans) { /* Initial burst finished. */ if (ifi->ifi_rainfo_trans != NULL) ifi->ifi_rainfo_trans = NULL; } /* Remove burst RA information */ if (ifi->ifi_rainfo_trans != NULL) { rm_rainfo(ifi->ifi_rainfo_trans); ifi->ifi_rainfo_trans = NULL; } if (ifi->ifi_rainfo != NULL) { /* * TRANSITIVE -> CONFIGURED * * After initial burst or transition from * one configuration to another, * ifi_rainfo always points to the next RA * information. */ ifi->ifi_state = IFI_STATE_CONFIGURED; syslog(LOG_DEBUG, "<%s> ifname=%s marked as " "CONFIGURED.", __func__, ifi->ifi_ifname); } else { /* * TRANSITIVE -> UNCONFIGURED * * If ifi_rainfo points to NULL, this * interface is shutting down. * */ int error; ifi->ifi_state = IFI_STATE_UNCONFIGURED; syslog(LOG_DEBUG, "<%s> ifname=%s marked as " "UNCONFIGURED.", __func__, ifi->ifi_ifname); error = sock_mc_leave(&sock, ifi->ifi_ifindex); if (error) exit(1); } } break; } } /* process RA timer */ struct rtadvd_timer * ra_timeout(void *arg) { struct ifinfo *ifi; ifi = (struct ifinfo *)arg; syslog(LOG_DEBUG, "<%s> RA timer on %s is expired", __func__, ifi->ifi_ifname); ra_output(ifi); return (ifi->ifi_ra_timer); } /* update RA timer */ void -ra_timer_update(void *arg, struct timeval *tm) +ra_timer_update(void *arg, struct timespec *tm) { uint16_t interval; struct rainfo *rai; struct ifinfo *ifi; ifi = (struct ifinfo *)arg; rai = ifi->ifi_rainfo; interval = 0; switch (ifi->ifi_state) { case IFI_STATE_UNCONFIGURED: return; break; case IFI_STATE_CONFIGURED: /* * Whenever a multicast advertisement is sent from * an interface, the timer is reset to a * uniformly-distributed random value between the * interface's configured MinRtrAdvInterval and * MaxRtrAdvInterval (RFC4861 6.2.4). */ interval = rai->rai_mininterval; #ifdef HAVE_ARC4RANDOM interval += arc4random_uniform(rai->rai_maxinterval - rai->rai_mininterval); #else interval += random() % (rai->rai_maxinterval - rai->rai_mininterval); #endif break; case IFI_STATE_TRANSITIVE: /* * For the first few advertisements (up to * MAX_INITIAL_RTR_ADVERTISEMENTS), if the randomly chosen * interval is greater than * MAX_INITIAL_RTR_ADVERT_INTERVAL, the timer SHOULD be * set to MAX_INITIAL_RTR_ADVERT_INTERVAL instead. (RFC * 4861 6.2.4) * * In such cases, the router SHOULD transmit one or more * (but not more than MAX_FINAL_RTR_ADVERTISEMENTS) final * multicast Router Advertisements on the interface with a * Router Lifetime field of zero. (RFC 4861 6.2.5) */ interval = ifi->ifi_burstinterval; break; } tm->tv_sec = interval; - tm->tv_usec = 0; + tm->tv_nsec = 0; syslog(LOG_DEBUG, "<%s> RA timer on %s is set to %ld:%ld", __func__, ifi->ifi_ifname, - (long int)tm->tv_sec, (long int)tm->tv_usec); + (long int)tm->tv_sec, (long int)tm->tv_nsec / 1000); return; } Index: head/usr.sbin/rtadvd/rtadvd.h =================================================================== --- head/usr.sbin/rtadvd/rtadvd.h (revision 253969) +++ head/usr.sbin/rtadvd/rtadvd.h (revision 253970) @@ -1,298 +1,298 @@ /* $FreeBSD$ */ /* $KAME: rtadvd.h,v 1.26 2003/08/05 12:34:23 itojun Exp $ */ /* * Copyright (C) 1998 WIDE Project. * Copyright (C) 2011 Hiroki Sato * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #define ELM_MALLOC(p,error_action) \ do { \ p = malloc(sizeof(*p)); \ if (p == NULL) { \ syslog(LOG_ERR, "<%s> malloc failed: %s", \ __func__, strerror(errno)); \ error_action; \ } \ memset(p, 0, sizeof(*p)); \ } while(0) #define IN6ADDR_LINKLOCAL_ALLNODES_INIT \ {{{ 0xff, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, \ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01 }}} #define IN6ADDR_LINKLOCAL_ALLROUTERS_INIT \ {{{ 0xff, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, \ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02 }}} #define IN6ADDR_SITELOCAL_ALLROUTERS_INIT \ {{{ 0xff, 0x05, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, \ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02 }}} extern struct sockaddr_in6 sin6_linklocal_allnodes; extern struct sockaddr_in6 sin6_linklocal_allrouters; extern struct sockaddr_in6 sin6_sitelocal_allrouters; /* * RFC 3542 API deprecates IPV6_PKTINFO in favor of * IPV6_RECVPKTINFO */ #ifndef IPV6_RECVPKTINFO #ifdef IPV6_PKTINFO #define IPV6_RECVPKTINFO IPV6_PKTINFO #endif #endif /* * RFC 3542 API deprecates IPV6_HOPLIMIT in favor of * IPV6_RECVHOPLIMIT */ #ifndef IPV6_RECVHOPLIMIT #ifdef IPV6_HOPLIMIT #define IPV6_RECVHOPLIMIT IPV6_HOPLIMIT #endif #endif /* protocol constants and default values */ #define DEF_MAXRTRADVINTERVAL 600 #define DEF_ADVLINKMTU 0 #define DEF_ADVREACHABLETIME 0 #define DEF_ADVRETRANSTIMER 0 #define DEF_ADVCURHOPLIMIT 64 #define DEF_ADVVALIDLIFETIME 2592000 #define DEF_ADVPREFERREDLIFETIME 604800 #define MAXROUTERLIFETIME 9000 #define MIN_MAXINTERVAL 4 #define MAX_MAXINTERVAL 1800 #define MIN_MININTERVAL 3 #define MAXREACHABLETIME 3600000 #define MAX_INITIAL_RTR_ADVERT_INTERVAL 16 #define MAX_INITIAL_RTR_ADVERTISEMENTS 3 #define MAX_FINAL_RTR_ADVERTISEMENTS 3 #define MIN_DELAY_BETWEEN_RAS 3 #define MAX_RA_DELAY_TIME 500000 /* usec */ #define PREFIX_FROM_KERNEL 1 #define PREFIX_FROM_CONFIG 2 #define PREFIX_FROM_DYNAMIC 3 struct prefix { TAILQ_ENTRY(prefix) pfx_next; struct rainfo *pfx_rainfo; /* back pointer to the interface */ /* * Expiration timer. This is used when a prefix derived from * the kernel is deleted. */ struct rtadvd_timer *pfx_timer; uint32_t pfx_validlifetime; /* AdvValidLifetime */ uint32_t pfx_vltimeexpire; /* Expiration of vltime */ uint32_t pfx_preflifetime; /* AdvPreferredLifetime */ uint32_t pfx_pltimeexpire; /* Expiration of pltime */ int pfx_onlinkflg; /* bool: AdvOnLinkFlag */ int pfx_autoconfflg; /* bool: AdvAutonomousFlag */ int pfx_prefixlen; int pfx_origin; /* From kernel or config */ struct in6_addr pfx_prefix; }; struct rtinfo { TAILQ_ENTRY(rtinfo) rti_next; uint32_t rti_ltime; /* route lifetime */ int rti_rtpref; /* route preference */ int rti_prefixlen; struct in6_addr rti_prefix; }; struct rdnss_addr { TAILQ_ENTRY(rdnss_addr) ra_next; struct in6_addr ra_dns; /* DNS server entry */ }; struct rdnss { TAILQ_ENTRY(rdnss) rd_next; TAILQ_HEAD(, rdnss_addr) rd_list; /* list of DNS servers */ uint32_t rd_ltime; /* number of seconds valid */ }; /* * The maximum length of a domain name in a DNS search list is calculated * by a domain name + length fields per 63 octets + a zero octet at * the tail and adding 8 octet boundary padding. */ #define _DNAME_LABELENC_MAXLEN \ (NI_MAXHOST + (NI_MAXHOST / 64 + 1) + 1) #define DNAME_LABELENC_MAXLEN \ (_DNAME_LABELENC_MAXLEN + 8 - _DNAME_LABELENC_MAXLEN % 8) struct dnssl_addr { TAILQ_ENTRY(dnssl_addr) da_next; int da_len; /* length of entry */ char da_dom[DNAME_LABELENC_MAXLEN]; /* search domain name entry */ }; struct dnssl { TAILQ_ENTRY(dnssl) dn_next; TAILQ_HEAD(, dnssl_addr) dn_list; /* list of search domains */ uint32_t dn_ltime; /* number of seconds valid */ }; struct soliciter { TAILQ_ENTRY(soliciter) sol_next; struct sockaddr_in6 sol_addr; }; struct rainfo { /* pointer for list */ TAILQ_ENTRY(rainfo) rai_next; /* interface information */ struct ifinfo *rai_ifinfo; int rai_advlinkopt; /* bool: whether include link-layer addr opt */ int rai_advifprefix; /* bool: gather IF prefixes? */ /* Router configuration variables */ uint16_t rai_lifetime; /* AdvDefaultLifetime */ uint16_t rai_maxinterval; /* MaxRtrAdvInterval */ uint16_t rai_mininterval; /* MinRtrAdvInterval */ int rai_managedflg; /* AdvManagedFlag */ int rai_otherflg; /* AdvOtherConfigFlag */ int rai_rtpref; /* router preference */ uint32_t rai_linkmtu; /* AdvLinkMTU */ uint32_t rai_reachabletime; /* AdvReachableTime */ uint32_t rai_retranstimer; /* AdvRetransTimer */ uint8_t rai_hoplimit; /* AdvCurHopLimit */ TAILQ_HEAD(, prefix) rai_prefix;/* AdvPrefixList(link head) */ int rai_pfxs; /* number of prefixes */ uint16_t rai_clockskew; /* used for consisitency check of lifetimes */ TAILQ_HEAD(, rdnss) rai_rdnss; /* DNS server list */ TAILQ_HEAD(, dnssl) rai_dnssl; /* search domain list */ TAILQ_HEAD(, rtinfo) rai_route; /* route information option (link head) */ int rai_routes; /* number of route information options */ /* actual RA packet data and its length */ size_t rai_ra_datalen; char *rai_ra_data; /* info about soliciter */ TAILQ_HEAD(, soliciter) rai_soliciter; /* recent solication source */ }; /* RA information list */ extern TAILQ_HEAD(railist_head_t, rainfo) railist; /* * ifi_state: * * (INIT) * | * | update_ifinfo() * | update_persist_ifinfo() * v * UNCONFIGURED * | ^ * loadconfig()| |rm_ifinfo(), ra_output() * (MC join)| |(MC leave) * | | * | | * v | * TRANSITIVE * | ^ * ra_output()| |getconfig() * | | * | | * | | * v | * CONFIGURED * * */ #define IFI_STATE_UNCONFIGURED 0 #define IFI_STATE_CONFIGURED 1 #define IFI_STATE_TRANSITIVE 2 struct ifinfo { TAILQ_ENTRY(ifinfo) ifi_next; uint16_t ifi_state; uint16_t ifi_persist; uint16_t ifi_ifindex; char ifi_ifname[IFNAMSIZ]; uint8_t ifi_type; uint16_t ifi_flags; uint32_t ifi_nd_flags; uint32_t ifi_phymtu; struct sockaddr_dl ifi_sdl; struct rainfo *ifi_rainfo; struct rainfo *ifi_rainfo_trans; uint16_t ifi_burstcount; uint32_t ifi_burstinterval; struct rtadvd_timer *ifi_ra_timer; /* timestamp when the latest RA was sent */ - struct timeval ifi_ra_lastsent; + struct timespec ifi_ra_lastsent; uint16_t ifi_rs_waitcount; /* statistics */ uint64_t ifi_raoutput; /* # of RAs sent */ uint64_t ifi_rainput; /* # of RAs received */ uint64_t ifi_rainconsistent; /* # of inconsistent recv'd RAs */ uint64_t ifi_rsinput; /* # of RSs received */ }; /* Interface list */ extern TAILQ_HEAD(ifilist_head_t, ifinfo) ifilist; extern char *mcastif; struct rtadvd_timer *ra_timeout(void *); -void ra_timer_update(void *, struct timeval *); +void ra_timer_update(void *, struct timespec *); void ra_output(struct ifinfo *); int prefix_match(struct in6_addr *, int, struct in6_addr *, int); struct ifinfo *if_indextoifinfo(int); struct prefix *find_prefix(struct rainfo *, struct in6_addr *, int); void rtadvd_set_reload(int); void rtadvd_set_shutdown(int); Index: head/usr.sbin/rtadvd/timer.c =================================================================== --- head/usr.sbin/rtadvd/timer.c (revision 253969) +++ head/usr.sbin/rtadvd/timer.c (revision 253970) @@ -1,197 +1,202 @@ /* $FreeBSD$ */ /* $KAME: timer.c,v 1.9 2002/06/10 19:59:47 itojun Exp $ */ /* * Copyright (C) 1998 WIDE Project. * Copyright (C) 2011 Hiroki Sato * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ -#include #include #include #include #include #include #include #include #include #include #include +#include #include #include "rtadvd.h" #include "timer_subr.h" #include "timer.h" struct rtadvd_timer_head_t ra_timer = TAILQ_HEAD_INITIALIZER(ra_timer); -static struct timeval tm_limit = {0x7fffffff, 0x7fffffff}; -static struct timeval tm_max; +static struct timespec tm_limit; +static struct timespec tm_max; void rtadvd_timer_init(void) { + /* Generate maximum time in timespec. */ + memset(&tm_limit.tv_sec, 0xff, sizeof(tm_limit.tv_sec)); + memset(&tm_limit.tv_nsec, 0xff, sizeof(tm_limit.tv_nsec)); + tm_limit.tv_sec &= ~(1UL << (sizeof(tm_limit.tv_sec) * 8 - 1)); + tm_limit.tv_nsec &= ~(1UL << (sizeof(tm_limit.tv_nsec) * 8 - 1)); tm_max = tm_limit; TAILQ_INIT(&ra_timer); } void rtadvd_update_timeout_handler(void) { struct ifinfo *ifi; TAILQ_FOREACH(ifi, &ifilist, ifi_next) { switch (ifi->ifi_state) { case IFI_STATE_CONFIGURED: case IFI_STATE_TRANSITIVE: if (ifi->ifi_ra_timer != NULL) continue; syslog(LOG_DEBUG, "<%s> add timer for %s (idx=%d)", __func__, ifi->ifi_ifname, ifi->ifi_ifindex); ifi->ifi_ra_timer = rtadvd_add_timer(ra_timeout, ra_timer_update, ifi, ifi); ra_timer_update((void *)ifi, &ifi->ifi_ra_timer->rat_tm); rtadvd_set_timer(&ifi->ifi_ra_timer->rat_tm, ifi->ifi_ra_timer); break; case IFI_STATE_UNCONFIGURED: if (ifi->ifi_ra_timer == NULL) continue; syslog(LOG_DEBUG, "<%s> remove timer for %s (idx=%d)", __func__, ifi->ifi_ifname, ifi->ifi_ifindex); rtadvd_remove_timer(ifi->ifi_ra_timer); ifi->ifi_ra_timer = NULL; break; } } return; } struct rtadvd_timer * rtadvd_add_timer(struct rtadvd_timer *(*timeout)(void *), - void (*update)(void *, struct timeval *), + void (*update)(void *, struct timespec *), void *timeodata, void *updatedata) { struct rtadvd_timer *rat; if (timeout == NULL) { syslog(LOG_ERR, "<%s> timeout function unspecified", __func__); exit(1); } rat = malloc(sizeof(*rat)); if (rat == NULL) { syslog(LOG_ERR, "<%s> can't allocate memory", __func__); exit(1); } memset(rat, 0, sizeof(*rat)); rat->rat_expire = timeout; rat->rat_update = update; rat->rat_expire_data = timeodata; rat->rat_update_data = updatedata; rat->rat_tm = tm_max; /* link into chain */ TAILQ_INSERT_TAIL(&ra_timer, rat, rat_next); return (rat); } void rtadvd_remove_timer(struct rtadvd_timer *rat) { if (rat == NULL) return; TAILQ_REMOVE(&ra_timer, rat, rat_next); free(rat); } /* * Check expiration for each timer. If a timer expires, * call the expire function for the timer and update the timer. * Return the next interval for select() call. */ -struct timeval * +struct timespec * rtadvd_check_timer(void) { - static struct timeval returnval; - struct timeval now; + static struct timespec returnval; + struct timespec now; struct rtadvd_timer *rat; - gettimeofday(&now, NULL); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); tm_max = tm_limit; TAILQ_FOREACH(rat, &ra_timer, rat_next) { - if (TIMEVAL_LEQ(&rat->rat_tm, &now)) { + if (TS_CMP(&rat->rat_tm, &now, <=)) { if (((*rat->rat_expire)(rat->rat_expire_data) == NULL)) continue; /* the timer was removed */ if (rat->rat_update) (*rat->rat_update)(rat->rat_update_data, &rat->rat_tm); - TIMEVAL_ADD(&rat->rat_tm, &now, &rat->rat_tm); + TS_ADD(&rat->rat_tm, &now, &rat->rat_tm); } - if (TIMEVAL_LT(&rat->rat_tm, &tm_max)) + if (TS_CMP(&rat->rat_tm, &tm_max, <)) tm_max = rat->rat_tm; } - if (TIMEVAL_EQUAL(&tm_max, &tm_limit)) { + if (TS_CMP(&tm_max, &tm_limit, ==)) { /* no need to timeout */ return (NULL); - } else if (TIMEVAL_LT(&tm_max, &now)) { + } else if (TS_CMP(&tm_max, &now, <)) { /* this may occur when the interval is too small */ - returnval.tv_sec = returnval.tv_usec = 0; + returnval.tv_sec = returnval.tv_nsec = 0; } else - TIMEVAL_SUB(&tm_max, &now, &returnval); + TS_SUB(&tm_max, &now, &returnval); return (&returnval); } void -rtadvd_set_timer(struct timeval *tm, struct rtadvd_timer *rat) +rtadvd_set_timer(struct timespec *tm, struct rtadvd_timer *rat) { - struct timeval now; + struct timespec now; /* reset the timer */ - gettimeofday(&now, NULL); - TIMEVAL_ADD(&now, tm, &rat->rat_tm); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); + TS_ADD(&now, tm, &rat->rat_tm); /* update the next expiration time */ - if (TIMEVAL_LT(&rat->rat_tm, &tm_max)) + if (TS_CMP(&rat->rat_tm, &tm_max, <)) tm_max = rat->rat_tm; return; } Index: head/usr.sbin/rtadvd/timer.h =================================================================== --- head/usr.sbin/rtadvd/timer.h (revision 253969) +++ head/usr.sbin/rtadvd/timer.h (revision 253970) @@ -1,52 +1,52 @@ /* $FreeBSD$ */ /* $KAME: timer.h,v 1.5 2002/05/31 13:30:38 jinmei Exp $ */ /* * Copyright (C) 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ extern TAILQ_HEAD(rtadvd_timer_head_t, rtadvd_timer) ra_timer; struct rtadvd_timer { TAILQ_ENTRY(rtadvd_timer) rat_next; struct rainfo *rat_rai; - struct timeval rat_tm; + struct timespec rat_tm; struct rtadvd_timer *(*rat_expire)(void *); void *rat_expire_data; - void (*rat_update)(void *, struct timeval *); + void (*rat_update)(void *, struct timespec *); void *rat_update_data; }; void rtadvd_timer_init(void); void rtadvd_update_timeout_handler(void); struct rtadvd_timer *rtadvd_add_timer(struct rtadvd_timer *(*)(void *), - void (*)(void *, struct timeval *), void *, void *); -void rtadvd_set_timer(struct timeval *, + void (*)(void *, struct timespec *), void *, void *); +void rtadvd_set_timer(struct timespec *, struct rtadvd_timer *); void rtadvd_remove_timer(struct rtadvd_timer *); -struct timeval *rtadvd_check_timer(void); +struct timespec *rtadvd_check_timer(void); Index: head/usr.sbin/rtadvd/timer_subr.c =================================================================== --- head/usr.sbin/rtadvd/timer_subr.c (revision 253969) +++ head/usr.sbin/rtadvd/timer_subr.c (revision 253970) @@ -1,126 +1,91 @@ /* $FreeBSD$ */ /* $KAME: timer.c,v 1.9 2002/06/10 19:59:47 itojun Exp $ */ /* * Copyright (C) 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ -#include #include #include #include #include #include +#include #include "timer.h" #include "timer_subr.h" -struct timeval * +struct timespec * rtadvd_timer_rest(struct rtadvd_timer *rat) { - static struct timeval returnval, now; + static struct timespec returnval, now; - gettimeofday(&now, NULL); - if (TIMEVAL_LEQ(&rat->rat_tm, &now)) { + clock_gettime(CLOCK_MONOTONIC_FAST, &now); + if (TS_CMP(&rat->rat_tm, &now, <=)) { syslog(LOG_DEBUG, "<%s> a timer must be expired, but not yet", __func__); - returnval.tv_sec = returnval.tv_usec = 0; + returnval.tv_sec = returnval.tv_nsec = 0; } else - TIMEVAL_SUB(&rat->rat_tm, &now, &returnval); + TS_SUB(&rat->rat_tm, &now, &returnval); return (&returnval); -} - -/* result = a + b */ -void -TIMEVAL_ADD(struct timeval *a, struct timeval *b, struct timeval *result) -{ - long l; - - if ((l = a->tv_usec + b->tv_usec) < MILLION) { - result->tv_usec = l; - result->tv_sec = a->tv_sec + b->tv_sec; - } - else { - result->tv_usec = l - MILLION; - result->tv_sec = a->tv_sec + b->tv_sec + 1; - } -} - -/* - * result = a - b - * XXX: this function assumes that a >= b. - */ -void -TIMEVAL_SUB(struct timeval *a, struct timeval *b, struct timeval *result) -{ - long l; - - if ((l = a->tv_usec - b->tv_usec) >= 0) { - result->tv_usec = l; - result->tv_sec = a->tv_sec - b->tv_sec; - } - else { - result->tv_usec = MILLION + l; - result->tv_sec = a->tv_sec - b->tv_sec - 1; - } } char * sec2str(uint32_t s, char *buf) { uint32_t day; uint32_t hour; uint32_t min; uint32_t sec; char *p; min = s / 60; sec = s % 60; hour = min / 60; min = min % 60; day = hour / 24; hour = hour % 24; p = buf; if (day > 0) p += sprintf(p, "%" PRIu32 "d", day); if (hour > 0) p += sprintf(p, "%" PRIu32 "h", hour); if (min > 0) p += sprintf(p, "%" PRIu32 "m", min); if ((p == buf) || (sec > 0 && p > buf)) sprintf(p, "%" PRIu32 "s", sec); return (buf); } Index: head/usr.sbin/rtadvd/timer_subr.h =================================================================== --- head/usr.sbin/rtadvd/timer_subr.h (revision 253969) +++ head/usr.sbin/rtadvd/timer_subr.h (revision 253970) @@ -1,57 +1,59 @@ /* $FreeBSD$ */ /* $KAME: timer.h,v 1.5 2002/05/31 13:30:38 jinmei Exp $ */ /* * Copyright (C) 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #define SSBUFLEN 1024 -#define MILLION 1000000 -/* a < b */ -#define TIMEVAL_LT(a, b) \ - (((a)->tv_sec < (b)->tv_sec) || \ - (((a)->tv_sec == (b)->tv_sec) && \ - ((a)->tv_usec < (b)->tv_usec))) +#define TS_CMP(tsp, usp, cmp) \ + (((tsp)->tv_sec == (usp)->tv_sec) ? \ + ((tsp)->tv_nsec cmp (usp)->tv_nsec) : \ + ((tsp)->tv_sec cmp (usp)->tv_sec)) +#define TS_ADD(tsp, usp, vsp) \ + do { \ + (vsp)->tv_sec = (tsp)->tv_sec + (usp)->tv_sec; \ + (vsp)->tv_nsec = (tsp)->tv_nsec + (usp)->tv_nsec; \ + if ((vsp)->tv_nsec >= 1000000000L) { \ + (vsp)->tv_sec++; \ + (vsp)->tv_nsec -= 1000000000L; \ + } \ + } while (0) +#define TS_SUB(tsp, usp, vsp) \ + do { \ + (vsp)->tv_sec = (tsp)->tv_sec - (usp)->tv_sec; \ + (vsp)->tv_nsec = (tsp)->tv_nsec - (usp)->tv_nsec; \ + if ((vsp)->tv_nsec < 0) { \ + (vsp)->tv_sec--; \ + (vsp)->tv_nsec += 1000000000L; \ + } \ + } while (0) -/* a <= b */ -#define TIMEVAL_LEQ(a, b) \ - (((a)->tv_sec < (b)->tv_sec) || \ - (((a)->tv_sec == (b)->tv_sec) && \ - ((a)->tv_usec <= (b)->tv_usec))) - -#define TIMEVAL_EQUAL(a,b) \ - (((a)->tv_sec == (b)->tv_sec) && \ - ((a)->tv_usec == (b)->tv_usec)) - -struct timeval *rtadvd_timer_rest(struct rtadvd_timer *); -void TIMEVAL_ADD(struct timeval *, struct timeval *, - struct timeval *); -void TIMEVAL_SUB(struct timeval *, struct timeval *, - struct timeval *); +struct timespec *rtadvd_timer_rest(struct rtadvd_timer *); char *sec2str(uint32_t, char *buf); Index: head/usr.sbin/rtsold/dump.c =================================================================== --- head/usr.sbin/rtsold/dump.c (revision 253969) +++ head/usr.sbin/rtsold/dump.c (revision 253970) @@ -1,186 +1,185 @@ /* $KAME: dump.c,v 1.13 2003/10/05 00:09:36 itojun Exp $ */ /* * Copyright (C) 1999 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $FreeBSD$ */ #include -#include #include #include #include #include #include #include #include #include #include #include #include #include "rtsold.h" static FILE *fp; extern struct ifinfo *iflist; static void dump_interface_status(void); static const char * const ifstatstr[] = {"IDLE", "DELAY", "PROBE", "DOWN", "TENTATIVE"}; static void dump_interface_status(void) { struct ifinfo *ifi; struct rainfo *rai; struct ra_opt *rao; - struct timeval now; + struct timespec now; char ntopbuf[INET6_ADDRSTRLEN]; - gettimeofday(&now, NULL); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); TAILQ_FOREACH(ifi, &ifinfo_head, ifi_next) { fprintf(fp, "Interface %s\n", ifi->ifname); fprintf(fp, " probe interval: "); if (ifi->probeinterval) { fprintf(fp, "%d\n", ifi->probeinterval); fprintf(fp, " probe timer: %d\n", ifi->probetimer); } else { fprintf(fp, "infinity\n"); fprintf(fp, " no probe timer\n"); } fprintf(fp, " interface status: %s\n", ifi->active > 0 ? "active" : "inactive"); fprintf(fp, " other config: %s\n", ifi->otherconfig ? "on" : "off"); fprintf(fp, " rtsold status: %s\n", ifstatstr[ifi->state]); fprintf(fp, " carrier detection: %s\n", ifi->mediareqok ? "available" : "unavailable"); fprintf(fp, " probes: %d, dadcount = %d\n", ifi->probes, ifi->dadcount); if (ifi->timer.tv_sec == tm_max.tv_sec && - ifi->timer.tv_usec == tm_max.tv_usec) + ifi->timer.tv_nsec == tm_max.tv_nsec) fprintf(fp, " no timer\n"); else { fprintf(fp, " timer: interval=%d:%d, expire=%s\n", (int)ifi->timer.tv_sec, - (int)ifi->timer.tv_usec, + (int)ifi->timer.tv_nsec / 1000, (ifi->expire.tv_sec < now.tv_sec) ? "expired" : sec2str(&ifi->expire)); } fprintf(fp, " number of valid RAs: %d\n", ifi->racnt); TAILQ_FOREACH(rai, &ifi->ifi_rainfo, rai_next) { fprintf(fp, " RA from %s\n", inet_ntop(AF_INET6, &rai->rai_saddr.sin6_addr, ntopbuf, sizeof(ntopbuf))); TAILQ_FOREACH(rao, &rai->rai_ra_opt, rao_next) { fprintf(fp, " option: "); switch (rao->rao_type) { case ND_OPT_RDNSS: fprintf(fp, "RDNSS: %s (expire: %s)\n", (char *)rao->rao_msg, sec2str(&rao->rao_expire)); break; case ND_OPT_DNSSL: fprintf(fp, "DNSSL: %s (expire: %s)\n", (char *)rao->rao_msg, sec2str(&rao->rao_expire)); break; default: break; } } fprintf(fp, "\n"); } } } void rtsold_dump_file(const char *dumpfile) { if ((fp = fopen(dumpfile, "w")) == NULL) { warnmsg(LOG_WARNING, __func__, "open a dump file(%s): %s", dumpfile, strerror(errno)); return; } dump_interface_status(); fclose(fp); } const char * -sec2str(const struct timeval *total) +sec2str(const struct timespec *total) { static char result[256]; int days, hours, mins, secs; int first = 1; char *p = result; char *ep = &result[sizeof(result)]; int n; - struct timeval now; + struct timespec now; time_t tsec; - gettimeofday(&now, NULL); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); tsec = total->tv_sec; - tsec += total->tv_usec / 1000000; + tsec += total->tv_nsec / 1000 / 1000000; tsec -= now.tv_sec; - tsec -= now.tv_usec / 1000000; + tsec -= now.tv_nsec / 1000 / 1000000; days = tsec / 3600 / 24; hours = (tsec / 3600) % 24; mins = (tsec / 60) % 60; secs = tsec % 60; if (days) { first = 0; n = snprintf(p, ep - p, "%dd", days); if (n < 0 || n >= ep - p) return "?"; p += n; } if (!first || hours) { first = 0; n = snprintf(p, ep - p, "%dh", hours); if (n < 0 || n >= ep - p) return "?"; p += n; } if (!first || mins) { first = 0; n = snprintf(p, ep - p, "%dm", mins); if (n < 0 || n >= ep - p) return "?"; p += n; } snprintf(p, ep - p, "%ds", secs); return (result); } Index: head/usr.sbin/rtsold/rtsol.c =================================================================== --- head/usr.sbin/rtsold/rtsol.c (revision 253969) +++ head/usr.sbin/rtsold/rtsol.c (revision 253970) @@ -1,954 +1,954 @@ /* $KAME: rtsol.c,v 1.27 2003/10/05 00:09:36 itojun Exp $ */ /* * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * Copyright (C) 2011 Hiroki Sato * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $FreeBSD$ */ #include #include #include -#include #include #include #include #include #include #include #define __BSD_VISIBLE 1 /* IN6ADDR_LINKLOCAL_ALLROUTERS_INIT */ #include #undef __BSD_VISIBLE #include #include #include #include #include #include #include #include #include +#include #include #include #include #include #include #include "rtsold.h" static struct msghdr rcvmhdr; static struct msghdr sndmhdr; static struct iovec rcviov[2]; static struct iovec sndiov[2]; static struct sockaddr_in6 from; static int rcvcmsglen; int rssock; static char rsid[IFNAMSIZ + 1 + sizeof(DNSINFO_ORIGIN_LABEL) + 1 + NI_MAXHOST]; struct ifinfo_head_t ifinfo_head = TAILQ_HEAD_INITIALIZER(ifinfo_head); static const struct sockaddr_in6 sin6_allrouters = { .sin6_len = sizeof(sin6_allrouters), .sin6_family = AF_INET6, .sin6_addr = IN6ADDR_LINKLOCAL_ALLROUTERS_INIT, }; static void call_script(const int, const char *const *, struct script_msg_head_t *); static size_t dname_labeldec(char *, size_t, const char *); static int safefile(const char *); static struct ra_opt *find_raopt(struct rainfo *, int, void *, size_t); static int ra_opt_rdnss_dispatch(struct ifinfo *, struct rainfo *, struct script_msg_head_t *, struct script_msg_head_t *); static char *make_rsid(const char *, const char *, struct rainfo *); #define _ARGS_OTHER otherconf_script, ifi->ifname #define _ARGS_RESADD resolvconf_script, "-a", rsid #define _ARGS_RESDEL resolvconf_script, "-d", rsid #define CALL_SCRIPT(name, sm_head) \ do { \ const char *const sarg[] = { _ARGS_##name, NULL }; \ call_script(sizeof(sarg), sarg, sm_head); \ } while(0) #define ELM_MALLOC(p,error_action) \ do { \ p = malloc(sizeof(*p)); \ if (p == NULL) { \ warnmsg(LOG_ERR, __func__, "malloc failed: %s", \ strerror(errno)); \ error_action; \ } \ memset(p, 0, sizeof(*p)); \ } while(0) int sockopen(void) { static u_char *rcvcmsgbuf = NULL, *sndcmsgbuf = NULL; int sndcmsglen, on; static u_char answer[1500]; struct icmp6_filter filt; sndcmsglen = rcvcmsglen = CMSG_SPACE(sizeof(struct in6_pktinfo)) + CMSG_SPACE(sizeof(int)); if (rcvcmsgbuf == NULL && (rcvcmsgbuf = malloc(rcvcmsglen)) == NULL) { warnmsg(LOG_ERR, __func__, "malloc for receive msghdr failed"); return (-1); } if (sndcmsgbuf == NULL && (sndcmsgbuf = malloc(sndcmsglen)) == NULL) { warnmsg(LOG_ERR, __func__, "malloc for send msghdr failed"); return (-1); } if ((rssock = socket(AF_INET6, SOCK_RAW, IPPROTO_ICMPV6)) < 0) { warnmsg(LOG_ERR, __func__, "socket: %s", strerror(errno)); return (-1); } /* specify to tell receiving interface */ on = 1; if (setsockopt(rssock, IPPROTO_IPV6, IPV6_RECVPKTINFO, &on, sizeof(on)) < 0) { warnmsg(LOG_ERR, __func__, "IPV6_RECVPKTINFO: %s", strerror(errno)); exit(1); } /* specify to tell value of hoplimit field of received IP6 hdr */ on = 1; if (setsockopt(rssock, IPPROTO_IPV6, IPV6_RECVHOPLIMIT, &on, sizeof(on)) < 0) { warnmsg(LOG_ERR, __func__, "IPV6_RECVHOPLIMIT: %s", strerror(errno)); exit(1); } /* specfiy to accept only router advertisements on the socket */ ICMP6_FILTER_SETBLOCKALL(&filt); ICMP6_FILTER_SETPASS(ND_ROUTER_ADVERT, &filt); if (setsockopt(rssock, IPPROTO_ICMPV6, ICMP6_FILTER, &filt, sizeof(filt)) == -1) { warnmsg(LOG_ERR, __func__, "setsockopt(ICMP6_FILTER): %s", strerror(errno)); return(-1); } /* initialize msghdr for receiving packets */ rcviov[0].iov_base = (caddr_t)answer; rcviov[0].iov_len = sizeof(answer); rcvmhdr.msg_name = (caddr_t)&from; rcvmhdr.msg_iov = rcviov; rcvmhdr.msg_iovlen = 1; rcvmhdr.msg_control = (caddr_t) rcvcmsgbuf; /* initialize msghdr for sending packets */ sndmhdr.msg_namelen = sizeof(struct sockaddr_in6); sndmhdr.msg_iov = sndiov; sndmhdr.msg_iovlen = 1; sndmhdr.msg_control = (caddr_t)sndcmsgbuf; sndmhdr.msg_controllen = sndcmsglen; return (rssock); } void sendpacket(struct ifinfo *ifi) { struct in6_pktinfo *pi; struct cmsghdr *cm; int hoplimit = 255; ssize_t i; struct sockaddr_in6 dst; dst = sin6_allrouters; dst.sin6_scope_id = ifi->linkid; sndmhdr.msg_name = (caddr_t)&dst; sndmhdr.msg_iov[0].iov_base = (caddr_t)ifi->rs_data; sndmhdr.msg_iov[0].iov_len = ifi->rs_datalen; cm = CMSG_FIRSTHDR(&sndmhdr); /* specify the outgoing interface */ cm->cmsg_level = IPPROTO_IPV6; cm->cmsg_type = IPV6_PKTINFO; cm->cmsg_len = CMSG_LEN(sizeof(struct in6_pktinfo)); pi = (struct in6_pktinfo *)CMSG_DATA(cm); memset(&pi->ipi6_addr, 0, sizeof(pi->ipi6_addr)); /*XXX*/ pi->ipi6_ifindex = ifi->sdl->sdl_index; /* specify the hop limit of the packet */ cm = CMSG_NXTHDR(&sndmhdr, cm); cm->cmsg_level = IPPROTO_IPV6; cm->cmsg_type = IPV6_HOPLIMIT; cm->cmsg_len = CMSG_LEN(sizeof(int)); memcpy(CMSG_DATA(cm), &hoplimit, sizeof(int)); warnmsg(LOG_DEBUG, __func__, "send RS on %s, whose state is %d", ifi->ifname, ifi->state); i = sendmsg(rssock, &sndmhdr, 0); if (i < 0 || (size_t)i != ifi->rs_datalen) { /* * ENETDOWN is not so serious, especially when using several * network cards on a mobile node. We ignore it. */ if (errno != ENETDOWN || dflag > 0) warnmsg(LOG_ERR, __func__, "sendmsg on %s: %s", ifi->ifname, strerror(errno)); } /* update counter */ ifi->probes++; } void rtsol_input(int s) { char ntopbuf[INET6_ADDRSTRLEN], ifnamebuf[IFNAMSIZ]; int l, ifindex = 0, *hlimp = NULL; ssize_t msglen; struct in6_pktinfo *pi = NULL; struct ifinfo *ifi = NULL; struct ra_opt *rao = NULL; struct icmp6_hdr *icp; struct nd_router_advert *nd_ra; struct cmsghdr *cm; struct rainfo *rai; char *raoptp; char *p; struct in6_addr *addr; struct nd_opt_hdr *ndo; struct nd_opt_rdnss *rdnss; struct nd_opt_dnssl *dnssl; size_t len; char nsbuf[INET6_ADDRSTRLEN + 1 + IFNAMSIZ + 1]; char dname[NI_MAXHOST]; - struct timeval now; - struct timeval lifetime; + struct timespec now; + struct timespec lifetime; int newent_rai; int newent_rao; /* get message. namelen and controllen must always be initialized. */ rcvmhdr.msg_namelen = sizeof(from); rcvmhdr.msg_controllen = rcvcmsglen; if ((msglen = recvmsg(s, &rcvmhdr, 0)) < 0) { warnmsg(LOG_ERR, __func__, "recvmsg: %s", strerror(errno)); return; } /* extract optional information via Advanced API */ for (cm = (struct cmsghdr *)CMSG_FIRSTHDR(&rcvmhdr); cm; cm = (struct cmsghdr *)CMSG_NXTHDR(&rcvmhdr, cm)) { if (cm->cmsg_level == IPPROTO_IPV6 && cm->cmsg_type == IPV6_PKTINFO && cm->cmsg_len == CMSG_LEN(sizeof(struct in6_pktinfo))) { pi = (struct in6_pktinfo *)(CMSG_DATA(cm)); ifindex = pi->ipi6_ifindex; } if (cm->cmsg_level == IPPROTO_IPV6 && cm->cmsg_type == IPV6_HOPLIMIT && cm->cmsg_len == CMSG_LEN(sizeof(int))) hlimp = (int *)CMSG_DATA(cm); } if (ifindex == 0) { warnmsg(LOG_ERR, __func__, "failed to get receiving interface"); return; } if (hlimp == NULL) { warnmsg(LOG_ERR, __func__, "failed to get receiving hop limit"); return; } if ((size_t)msglen < sizeof(struct nd_router_advert)) { warnmsg(LOG_INFO, __func__, "packet size(%zd) is too short", msglen); return; } icp = (struct icmp6_hdr *)rcvmhdr.msg_iov[0].iov_base; if (icp->icmp6_type != ND_ROUTER_ADVERT) { /* * this should not happen because we configured a filter * that only passes RAs on the receiving socket. */ warnmsg(LOG_ERR, __func__, "invalid icmp type(%d) from %s on %s", icp->icmp6_type, inet_ntop(AF_INET6, &from.sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } if (icp->icmp6_code != 0) { warnmsg(LOG_INFO, __func__, "invalid icmp code(%d) from %s on %s", icp->icmp6_code, inet_ntop(AF_INET6, &from.sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } if (*hlimp != 255) { warnmsg(LOG_INFO, __func__, "invalid RA with hop limit(%d) from %s on %s", *hlimp, inet_ntop(AF_INET6, &from.sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } if (pi && !IN6_IS_ADDR_LINKLOCAL(&from.sin6_addr)) { warnmsg(LOG_INFO, __func__, "invalid RA with non link-local source from %s on %s", inet_ntop(AF_INET6, &from.sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } /* xxx: more validation? */ if ((ifi = find_ifinfo(pi->ipi6_ifindex)) == NULL) { warnmsg(LOG_INFO, __func__, "received RA from %s on an unexpected IF(%s)", inet_ntop(AF_INET6, &from.sin6_addr, ntopbuf, sizeof(ntopbuf)), if_indextoname(pi->ipi6_ifindex, ifnamebuf)); return; } warnmsg(LOG_DEBUG, __func__, "received RA from %s on %s, state is %d", inet_ntop(AF_INET6, &from.sin6_addr, ntopbuf, sizeof(ntopbuf)), ifi->ifname, ifi->state); nd_ra = (struct nd_router_advert *)icp; /* * Process the "O bit." * If the value of OtherConfigFlag changes from FALSE to TRUE, the * host should invoke the stateful autoconfiguration protocol, * requesting information. * [RFC 2462 Section 5.5.3] */ if (((nd_ra->nd_ra_flags_reserved) & ND_RA_FLAG_OTHER) && !ifi->otherconfig) { warnmsg(LOG_DEBUG, __func__, "OtherConfigFlag on %s is turned on", ifi->ifname); ifi->otherconfig = 1; CALL_SCRIPT(OTHER, NULL); } - gettimeofday(&now, NULL); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); newent_rai = 0; rai = find_rainfo(ifi, &from); if (rai == NULL) { ELM_MALLOC(rai, exit(1)); rai->rai_ifinfo = ifi; TAILQ_INIT(&rai->rai_ra_opt); rai->rai_saddr.sin6_family = AF_INET6; rai->rai_saddr.sin6_len = sizeof(rai->rai_saddr); memcpy(&rai->rai_saddr.sin6_addr, &from.sin6_addr, sizeof(rai->rai_saddr.sin6_addr)); newent_rai = 1; } #define RA_OPT_NEXT_HDR(x) (struct nd_opt_hdr *)((char *)x + \ (((struct nd_opt_hdr *)x)->nd_opt_len * 8)) /* Process RA options. */ warnmsg(LOG_DEBUG, __func__, "Processing RA"); raoptp = (char *)icp + sizeof(struct nd_router_advert); while (raoptp < (char *)icp + msglen) { ndo = (struct nd_opt_hdr *)raoptp; warnmsg(LOG_DEBUG, __func__, "ndo = %p", raoptp); warnmsg(LOG_DEBUG, __func__, "ndo->nd_opt_type = %d", ndo->nd_opt_type); warnmsg(LOG_DEBUG, __func__, "ndo->nd_opt_len = %d", ndo->nd_opt_len); switch (ndo->nd_opt_type) { case ND_OPT_RDNSS: rdnss = (struct nd_opt_rdnss *)raoptp; /* Optlen sanity check (Section 5.3.1 in RFC 6106) */ if (rdnss->nd_opt_rdnss_len < 3) { warnmsg(LOG_INFO, __func__, "too short RDNSS option" "in RA from %s was ignored.", inet_ntop(AF_INET6, &from.sin6_addr, ntopbuf, sizeof(ntopbuf))); break; } addr = (struct in6_addr *)(raoptp + sizeof(*rdnss)); while ((char *)addr < (char *)RA_OPT_NEXT_HDR(raoptp)) { if (inet_ntop(AF_INET6, addr, ntopbuf, sizeof(ntopbuf)) == NULL) { warnmsg(LOG_INFO, __func__, "an invalid address in RDNSS option" " in RA from %s was ignored.", inet_ntop(AF_INET6, &from.sin6_addr, ntopbuf, sizeof(ntopbuf))); addr++; continue; } if (IN6_IS_ADDR_LINKLOCAL(addr)) /* XXX: % has to be escaped here */ l = snprintf(nsbuf, sizeof(nsbuf), "%s%c%s", ntopbuf, SCOPE_DELIMITER, ifi->ifname); else l = snprintf(nsbuf, sizeof(nsbuf), "%s", ntopbuf); if (l < 0 || (size_t)l >= sizeof(nsbuf)) { warnmsg(LOG_ERR, __func__, "address copying error in " "RDNSS option: %d.", l); addr++; continue; } warnmsg(LOG_DEBUG, __func__, "nsbuf = %s", nsbuf); newent_rao = 0; rao = find_raopt(rai, ndo->nd_opt_type, nsbuf, strlen(nsbuf)); if (rao == NULL) { ELM_MALLOC(rao, break); rao->rao_type = ndo->nd_opt_type; rao->rao_len = strlen(nsbuf); rao->rao_msg = strdup(nsbuf); if (rao->rao_msg == NULL) { warnmsg(LOG_ERR, __func__, "strdup failed: %s", strerror(errno)); free(rao); addr++; continue; } newent_rao = 1; } /* Set expiration timer */ memset(&rao->rao_expire, 0, sizeof(rao->rao_expire)); memset(&lifetime, 0, sizeof(lifetime)); lifetime.tv_sec = ntohl(rdnss->nd_opt_rdnss_lifetime); - timeradd(&now, &lifetime, &rao->rao_expire); + TS_ADD(&now, &lifetime, &rao->rao_expire); if (newent_rao) TAILQ_INSERT_TAIL(&rai->rai_ra_opt, rao, rao_next); addr++; } break; case ND_OPT_DNSSL: dnssl = (struct nd_opt_dnssl *)raoptp; /* Optlen sanity check (Section 5.3.1 in RFC 6106) */ if (dnssl->nd_opt_dnssl_len < 2) { warnmsg(LOG_INFO, __func__, "too short DNSSL option" "in RA from %s was ignored.", inet_ntop(AF_INET6, &from.sin6_addr, ntopbuf, sizeof(ntopbuf))); break; } /* * Ensure NUL-termination in DNSSL in case of * malformed field. */ p = (char *)RA_OPT_NEXT_HDR(raoptp); *(p - 1) = '\0'; p = raoptp + sizeof(*dnssl); while (1 < (len = dname_labeldec(dname, sizeof(dname), p))) { /* length == 1 means empty string */ warnmsg(LOG_DEBUG, __func__, "dname = %s", dname); newent_rao = 0; rao = find_raopt(rai, ndo->nd_opt_type, dname, strlen(dname)); if (rao == NULL) { ELM_MALLOC(rao, break); rao->rao_type = ndo->nd_opt_type; rao->rao_len = strlen(dname); rao->rao_msg = strdup(dname); if (rao->rao_msg == NULL) { warnmsg(LOG_ERR, __func__, "strdup failed: %s", strerror(errno)); free(rao); addr++; continue; } newent_rao = 1; } /* Set expiration timer */ memset(&rao->rao_expire, 0, sizeof(rao->rao_expire)); memset(&lifetime, 0, sizeof(lifetime)); lifetime.tv_sec = ntohl(dnssl->nd_opt_dnssl_lifetime); - timeradd(&now, &lifetime, &rao->rao_expire); + TS_ADD(&now, &lifetime, &rao->rao_expire); if (newent_rao) TAILQ_INSERT_TAIL(&rai->rai_ra_opt, rao, rao_next); p += len; } break; default: /* nothing to do for other options */ break; } raoptp = (char *)RA_OPT_NEXT_HDR(raoptp); } if (newent_rai) TAILQ_INSERT_TAIL(&ifi->ifi_rainfo, rai, rai_next); ra_opt_handler(ifi); ifi->racnt++; switch (ifi->state) { case IFS_IDLE: /* should be ignored */ case IFS_DELAY: /* right? */ break; case IFS_PROBE: ifi->state = IFS_IDLE; ifi->probes = 0; rtsol_timer_update(ifi); break; } } static char resstr_ns_prefix[] = "nameserver "; static char resstr_sh_prefix[] = "search "; static char resstr_nl[] = "\n"; static char resstr_sp[] = " "; int ra_opt_handler(struct ifinfo *ifi) { struct ra_opt *rao; struct rainfo *rai; struct script_msg *smp1, *smp2, *smp3; - struct timeval now; + struct timespec now; struct script_msg_head_t sm_rdnss_head = TAILQ_HEAD_INITIALIZER(sm_rdnss_head); struct script_msg_head_t sm_dnssl_head = TAILQ_HEAD_INITIALIZER(sm_dnssl_head); int dcount, dlen; dcount = 0; dlen = strlen(resstr_sh_prefix) + strlen(resstr_nl); - gettimeofday(&now, NULL); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); /* * All options from multiple RAs with the same or different * source addresses on a single interface will be gathered and * handled, not overridden. [RFC 4861 6.3.4] */ TAILQ_FOREACH(rai, &ifi->ifi_rainfo, rai_next) { TAILQ_FOREACH(rao, &rai->rai_ra_opt, rao_next) { switch (rao->rao_type) { case ND_OPT_RDNSS: - if (timercmp(&now, &rao->rao_expire, >)) { + if (TS_CMP(&now, &rao->rao_expire, >)) { warnmsg(LOG_INFO, __func__, "expired rdnss entry: %s", (char *)rao->rao_msg); break; } ELM_MALLOC(smp1, continue); ELM_MALLOC(smp2, goto free1); ELM_MALLOC(smp3, goto free2); smp1->sm_msg = resstr_ns_prefix; TAILQ_INSERT_TAIL(&sm_rdnss_head, smp1, sm_next); smp2->sm_msg = rao->rao_msg; TAILQ_INSERT_TAIL(&sm_rdnss_head, smp2, sm_next); smp3->sm_msg = resstr_nl; TAILQ_INSERT_TAIL(&sm_rdnss_head, smp3, sm_next); ifi->ifi_rdnss = IFI_DNSOPT_STATE_RECEIVED; break; case ND_OPT_DNSSL: - if (timercmp(&now, &rao->rao_expire, >)) { + if (TS_CMP(&now, &rao->rao_expire, >)) { warnmsg(LOG_INFO, __func__, "expired dnssl entry: %s", (char *)rao->rao_msg); break; } dcount++; /* Check resolv.conf(5) restrictions. */ if (dcount > 6) { warnmsg(LOG_INFO, __func__, "dnssl entry exceeding maximum count (%d>6)" ": %s", dcount, (char *)rao->rao_msg); break; } if (256 < dlen + strlen(rao->rao_msg) + strlen(resstr_sp)) { warnmsg(LOG_INFO, __func__, "dnssl entry exceeding maximum length " "(>256): %s", (char *)rao->rao_msg); break; } ELM_MALLOC(smp1, continue); ELM_MALLOC(smp2, goto free1); if (TAILQ_EMPTY(&sm_dnssl_head)) { ELM_MALLOC(smp3, goto free2); smp3->sm_msg = resstr_sh_prefix; TAILQ_INSERT_TAIL(&sm_dnssl_head, smp3, sm_next); } smp1->sm_msg = rao->rao_msg; TAILQ_INSERT_TAIL(&sm_dnssl_head, smp1, sm_next); smp2->sm_msg = resstr_sp; TAILQ_INSERT_TAIL(&sm_dnssl_head, smp2, sm_next); dlen += strlen(rao->rao_msg) + strlen(resstr_sp); break; ifi->ifi_dnssl = IFI_DNSOPT_STATE_RECEIVED; default: break; } continue; free2: free(smp2); free1: free(smp1); } /* Call the script for each information source. */ if (uflag) ra_opt_rdnss_dispatch(ifi, rai, &sm_rdnss_head, &sm_dnssl_head); } /* Call the script for each interface. */ if (!uflag) ra_opt_rdnss_dispatch(ifi, NULL, &sm_rdnss_head, &sm_dnssl_head); return (0); } char * make_rsid(const char *ifname, const char *origin, struct rainfo *rai) { char hbuf[NI_MAXHOST]; if (rai == NULL) sprintf(rsid, "%s:%s", ifname, origin); else { if (!IN6_IS_ADDR_LINKLOCAL(&rai->rai_saddr.sin6_addr)) return (NULL); if (getnameinfo((struct sockaddr *)&rai->rai_saddr, rai->rai_saddr.sin6_len, hbuf, sizeof(hbuf), NULL, 0, NI_NUMERICHOST) != 0) return (NULL); sprintf(rsid, "%s:%s:[%s]", ifname, origin, hbuf); } warnmsg(LOG_DEBUG, __func__, "rsid = [%s]", rsid); return (rsid); } int ra_opt_rdnss_dispatch(struct ifinfo *ifi, struct rainfo *rai, struct script_msg_head_t *sm_rdnss_head, struct script_msg_head_t *sm_dnssl_head) { const char *r; struct script_msg *smp1; int error; error = 0; /* Add \n for DNSSL list. */ if (!TAILQ_EMPTY(sm_dnssl_head)) { ELM_MALLOC(smp1, goto ra_opt_rdnss_freeit); smp1->sm_msg = resstr_nl; TAILQ_INSERT_TAIL(sm_dnssl_head, smp1, sm_next); } TAILQ_CONCAT(sm_rdnss_head, sm_dnssl_head, sm_next); if (rai != NULL && uflag) r = make_rsid(ifi->ifname, DNSINFO_ORIGIN_LABEL, rai); else r = make_rsid(ifi->ifname, DNSINFO_ORIGIN_LABEL, NULL); if (r == NULL) { warnmsg(LOG_ERR, __func__, "make_rsid() failed. " "Script was not invoked."); error = 1; goto ra_opt_rdnss_freeit; } if (!TAILQ_EMPTY(sm_rdnss_head)) CALL_SCRIPT(RESADD, sm_rdnss_head); else if (ifi->ifi_rdnss == IFI_DNSOPT_STATE_RECEIVED || ifi->ifi_dnssl == IFI_DNSOPT_STATE_RECEIVED) { CALL_SCRIPT(RESDEL, NULL); ifi->ifi_rdnss = IFI_DNSOPT_STATE_NOINFO; ifi->ifi_dnssl = IFI_DNSOPT_STATE_NOINFO; } ra_opt_rdnss_freeit: /* Clear script message queue. */ if (!TAILQ_EMPTY(sm_rdnss_head)) { while ((smp1 = TAILQ_FIRST(sm_rdnss_head)) != NULL) { TAILQ_REMOVE(sm_rdnss_head, smp1, sm_next); free(smp1); } } if (!TAILQ_EMPTY(sm_dnssl_head)) { while ((smp1 = TAILQ_FIRST(sm_dnssl_head)) != NULL) { TAILQ_REMOVE(sm_dnssl_head, smp1, sm_next); free(smp1); } } return (error); } static struct ra_opt * find_raopt(struct rainfo *rai, int type, void *msg, size_t len) { struct ra_opt *rao; TAILQ_FOREACH(rao, &rai->rai_ra_opt, rao_next) { if (rao->rao_type == type && rao->rao_len == strlen(msg) && memcmp(rao->rao_msg, msg, len) == 0) break; } return (rao); } static void call_script(const int argc, const char *const argv[], struct script_msg_head_t *sm_head) { const char *scriptpath; int fd[2]; int error; pid_t pid, wpid; if ((scriptpath = argv[0]) == NULL) return; fd[0] = fd[1] = -1; if (sm_head != NULL && !TAILQ_EMPTY(sm_head)) { error = pipe(fd); if (error) { warnmsg(LOG_ERR, __func__, "failed to create a pipe: %s", strerror(errno)); return; } } /* launch the script */ pid = fork(); if (pid < 0) { warnmsg(LOG_ERR, __func__, "failed to fork: %s", strerror(errno)); return; } else if (pid) { /* parent */ int wstatus; if (fd[0] != -1) { /* Send message to the child if any. */ ssize_t len; struct script_msg *smp; close(fd[0]); TAILQ_FOREACH(smp, sm_head, sm_next) { len = strlen(smp->sm_msg); warnmsg(LOG_DEBUG, __func__, "write to child = %s(%zd)", smp->sm_msg, len); if (write(fd[1], smp->sm_msg, len) != len) { warnmsg(LOG_ERR, __func__, "write to child failed: %s", strerror(errno)); break; } } close(fd[1]); } do { wpid = wait(&wstatus); } while (wpid != pid && wpid > 0); if (wpid < 0) warnmsg(LOG_ERR, __func__, "wait: %s", strerror(errno)); else warnmsg(LOG_DEBUG, __func__, "script \"%s\" terminated", scriptpath); } else { /* child */ int nullfd; char **_argv; if (safefile(scriptpath)) { warnmsg(LOG_ERR, __func__, "script \"%s\" cannot be executed safely", scriptpath); exit(1); } nullfd = open("/dev/null", O_RDWR); if (nullfd < 0) { warnmsg(LOG_ERR, __func__, "open /dev/null: %s", strerror(errno)); exit(1); } if (fd[0] != -1) { /* Receive message from STDIN if any. */ close(fd[1]); if (fd[0] != STDIN_FILENO) { /* Connect a pipe read-end to child's STDIN. */ if (dup2(fd[0], STDIN_FILENO) != STDIN_FILENO) { warnmsg(LOG_ERR, __func__, "dup2 STDIN: %s", strerror(errno)); exit(1); } close(fd[0]); } } else dup2(nullfd, STDIN_FILENO); dup2(nullfd, STDOUT_FILENO); dup2(nullfd, STDERR_FILENO); if (nullfd > STDERR_FILENO) close(nullfd); _argv = malloc(sizeof(*_argv) * argc); if (_argv == NULL) { warnmsg(LOG_ERR, __func__, "malloc: %s", strerror(errno)); exit(1); } memcpy(_argv, argv, (size_t)argc); execv(scriptpath, (char *const *)_argv); warnmsg(LOG_ERR, __func__, "child: exec failed: %s", strerror(errno)); exit(1); } return; } static int safefile(const char *path) { struct stat s; uid_t myuid; /* no setuid */ if (getuid() != geteuid()) { warnmsg(LOG_NOTICE, __func__, "setuid'ed execution not allowed\n"); return (-1); } if (lstat(path, &s) != 0) { warnmsg(LOG_NOTICE, __func__, "lstat failed: %s", strerror(errno)); return (-1); } /* the file must be owned by the running uid */ myuid = getuid(); if (s.st_uid != myuid) { warnmsg(LOG_NOTICE, __func__, "%s has invalid owner uid\n", path); return (-1); } switch (s.st_mode & S_IFMT) { case S_IFREG: break; default: warnmsg(LOG_NOTICE, __func__, "%s is an invalid file type 0x%o\n", path, (s.st_mode & S_IFMT)); return (-1); } return (0); } /* Decode domain name label encoding in RFC 1035 Section 3.1 */ static size_t dname_labeldec(char *dst, size_t dlen, const char *src) { size_t len; const char *src_origin; const char *src_last; const char *dst_origin; src_origin = src; src_last = strchr(src, '\0'); dst_origin = dst; memset(dst, '\0', dlen); while (src && (len = (uint8_t)(*src++) & 0x3f) && (src + len) <= src_last) { if (dst != dst_origin) *dst++ = '.'; warnmsg(LOG_DEBUG, __func__, "labellen = %zd", len); memcpy(dst, src, len); src += len; dst += len; } *dst = '\0'; /* * XXX validate that domain name only contains valid characters * for two reasons: 1) correctness, 2) we do not want to pass * possible malicious, unescaped characters like `` to a script * or program that could be exploited that way. */ return (src - src_origin); } Index: head/usr.sbin/rtsold/rtsold.c =================================================================== --- head/usr.sbin/rtsold/rtsold.c (revision 253969) +++ head/usr.sbin/rtsold/rtsold.c (revision 253970) @@ -1,984 +1,990 @@ /* $KAME: rtsold.c,v 1.67 2003/05/17 18:16:15 itojun Exp $ */ /* * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $FreeBSD$ */ #include #include -#include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include +#include #include #include #include #include #ifdef HAVE_POLL_H #include #endif #include "rtsold.h" #define RTSOL_DUMPFILE "/var/run/rtsold.dump"; #define RTSOL_PIDFILE "/var/run/rtsold.pid"; struct ifinfo *iflist; -struct timeval tm_max = {0x7fffffff, 0x7fffffff}; +struct timespec tm_max; static int log_upto = 999; static int fflag = 0; int Fflag = 0; /* force setting sysctl parameters */ int aflag = 0; int dflag = 0; int uflag = 0; const char *otherconf_script; const char *resolvconf_script = "/sbin/resolvconf"; /* protocol constants */ #define MAX_RTR_SOLICITATION_DELAY 1 /* second */ #define RTR_SOLICITATION_INTERVAL 4 /* seconds */ #define MAX_RTR_SOLICITATIONS 3 /* times */ /* * implementation dependent constants in seconds * XXX: should be configurable */ #define PROBE_INTERVAL 60 /* static variables and functions */ static int mobile_node = 0; static const char *pidfilename = RTSOL_PIDFILE; #ifndef SMALL static int do_dump; static const char *dumpfilename = RTSOL_DUMPFILE; #endif #if 0 static int ifreconfig(char *); #endif static int make_packet(struct ifinfo *); -static struct timeval *rtsol_check_timer(void); +static struct timespec *rtsol_check_timer(void); #ifndef SMALL static void rtsold_set_dump_file(int); #endif static void usage(void); int main(int argc, char **argv) { int s, ch, once = 0; - struct timeval *timeout; + struct timespec *timeout; const char *opts; #ifdef HAVE_POLL_H struct pollfd set[2]; #else fd_set *fdsetp, *selectfdp; int fdmasks; int maxfd; #endif int rtsock; char *argv0; #ifndef SMALL /* rtsold */ opts = "adDfFm1O:p:R:u"; #else /* rtsol */ opts = "adDFO:R:u"; fflag = 1; once = 1; #endif argv0 = argv[0]; while ((ch = getopt(argc, argv, opts)) != -1) { switch (ch) { case 'a': aflag = 1; break; case 'd': dflag += 1; break; case 'D': dflag += 2; break; case 'f': fflag = 1; break; case 'F': Fflag = 1; break; case 'm': mobile_node = 1; break; case '1': once = 1; break; case 'O': otherconf_script = optarg; break; case 'p': pidfilename = optarg; break; case 'R': resolvconf_script = optarg; break; case 'u': uflag = 1; break; default: usage(); exit(1); } } argc -= optind; argv += optind; if ((!aflag && argc == 0) || (aflag && argc != 0)) { usage(); exit(1); } + /* Generate maximum time in timespec. */ + memset(&tm_max.tv_sec, 0xff, sizeof(tm_max.tv_sec)); + memset(&tm_max.tv_nsec, 0xff, sizeof(tm_max.tv_nsec)); + tm_max.tv_sec &= ~(1UL << (sizeof(tm_max.tv_sec) * 8 - 1)); + tm_max.tv_nsec &= ~(1UL << (sizeof(tm_max.tv_nsec) * 8 - 1)); + /* set log level */ if (dflag > 1) log_upto = LOG_DEBUG; else if (dflag > 0) log_upto = LOG_INFO; else log_upto = LOG_NOTICE; if (!fflag) { char *ident; ident = strrchr(argv0, '/'); if (!ident) ident = argv0; else ident++; openlog(ident, LOG_NDELAY|LOG_PID, LOG_DAEMON); if (log_upto >= 0) setlogmask(LOG_UPTO(log_upto)); } if (otherconf_script && *otherconf_script != '/') { errx(1, "configuration script (%s) must be an absolute path", otherconf_script); } if (resolvconf_script && *resolvconf_script != '/') { errx(1, "configuration script (%s) must be an absolute path", resolvconf_script); } if (pidfilename && *pidfilename != '/') { errx(1, "pid filename (%s) must be an absolute path", pidfilename); } #ifndef HAVE_ARC4RANDOM /* random value initialization */ srandom((u_long)time(NULL)); #endif #if (__FreeBSD_version < 900000) if (Fflag) { setinet6sysctl(IPV6CTL_FORWARDING, 0); } else { /* warn if forwarding is up */ if (getinet6sysctl(IPV6CTL_FORWARDING)) warnx("kernel is configured as a router, not a host"); } #endif #ifndef SMALL /* initialization to dump internal status to a file */ signal(SIGUSR1, rtsold_set_dump_file); #endif if (!fflag) daemon(0, 0); /* act as a daemon */ /* * Open a socket for sending RS and receiving RA. * This should be done before calling ifinit(), since the function * uses the socket. */ if ((s = sockopen()) < 0) { warnmsg(LOG_ERR, __func__, "failed to open a socket"); exit(1); } #ifdef HAVE_POLL_H set[0].fd = s; set[0].events = POLLIN; #else maxfd = s; #endif #ifdef HAVE_POLL_H set[1].fd = -1; #endif if ((rtsock = rtsock_open()) < 0) { warnmsg(LOG_ERR, __func__, "failed to open a socket"); exit(1); } #ifdef HAVE_POLL_H set[1].fd = rtsock; set[1].events = POLLIN; #else if (rtsock > maxfd) maxfd = rtsock; #endif #ifndef HAVE_POLL_H fdmasks = howmany(maxfd + 1, NFDBITS) * sizeof(fd_mask); if ((fdsetp = malloc(fdmasks)) == NULL) { warnmsg(LOG_ERR, __func__, "malloc"); exit(1); } if ((selectfdp = malloc(fdmasks)) == NULL) { warnmsg(LOG_ERR, __func__, "malloc"); exit(1); } #endif /* configuration per interface */ if (ifinit()) { warnmsg(LOG_ERR, __func__, "failed to initialize interfaces"); exit(1); } if (aflag) argv = autoifprobe(); while (argv && *argv) { if (ifconfig(*argv)) { warnmsg(LOG_ERR, __func__, "failed to initialize %s", *argv); exit(1); } argv++; } /* setup for probing default routers */ if (probe_init()) { warnmsg(LOG_ERR, __func__, "failed to setup for probing routers"); exit(1); /*NOTREACHED*/ } /* dump the current pid */ if (!once) { pid_t pid = getpid(); FILE *fp; if ((fp = fopen(pidfilename, "w")) == NULL) warnmsg(LOG_ERR, __func__, "failed to open a pid log file(%s): %s", pidfilename, strerror(errno)); else { fprintf(fp, "%d\n", pid); fclose(fp); } } #ifndef HAVE_POLL_H memset(fdsetp, 0, fdmasks); FD_SET(s, fdsetp); FD_SET(rtsock, fdsetp); #endif while (1) { /* main loop */ int e; #ifndef HAVE_POLL_H memcpy(selectfdp, fdsetp, fdmasks); #endif #ifndef SMALL if (do_dump) { /* SIGUSR1 */ do_dump = 0; rtsold_dump_file(dumpfilename); } #endif timeout = rtsol_check_timer(); if (once) { struct ifinfo *ifi; /* if we have no timeout, we are done (or failed) */ if (timeout == NULL) break; /* if all interfaces have got RA packet, we are done */ TAILQ_FOREACH(ifi, &ifinfo_head, ifi_next) { if (ifi->state != IFS_DOWN && ifi->racnt == 0) break; } if (ifi == NULL) break; } #ifdef HAVE_POLL_H - e = poll(set, 2, timeout ? (timeout->tv_sec * 1000 + timeout->tv_usec / 1000) : INFTIM); + e = poll(set, 2, timeout ? (timeout->tv_sec * 1000 + timeout->tv_nsec / 1000 / 1000) : INFTIM); #else e = select(maxfd + 1, selectfdp, NULL, NULL, timeout); #endif if (e < 1) { if (e < 0 && errno != EINTR) { warnmsg(LOG_ERR, __func__, "select: %s", strerror(errno)); } continue; } /* packet reception */ #ifdef HAVE_POLL_H if (set[1].revents & POLLIN) #else if (FD_ISSET(rtsock, selectfdp)) #endif rtsock_input(rtsock); #ifdef HAVE_POLL_H if (set[0].revents & POLLIN) #else if (FD_ISSET(s, selectfdp)) #endif rtsol_input(s); } /* NOTREACHED */ return (0); } int ifconfig(char *ifname) { struct ifinfo *ifi; struct sockaddr_dl *sdl; int flags; if ((sdl = if_nametosdl(ifname)) == NULL) { warnmsg(LOG_ERR, __func__, "failed to get link layer information for %s", ifname); return (-1); } if (find_ifinfo(sdl->sdl_index)) { warnmsg(LOG_ERR, __func__, "interface %s was already configured", ifname); free(sdl); return (-1); } if (Fflag) { struct in6_ndireq nd; int s; if ((s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) { warnmsg(LOG_ERR, __func__, "socket() failed."); return (-1); } memset(&nd, 0, sizeof(nd)); strlcpy(nd.ifname, ifname, sizeof(nd.ifname)); if (ioctl(s, SIOCGIFINFO_IN6, (caddr_t)&nd) < 0) { warnmsg(LOG_ERR, __func__, "cannot get accept_rtadv flag"); close(s); return (-1); } nd.ndi.flags |= ND6_IFF_ACCEPT_RTADV; if (ioctl(s, SIOCSIFINFO_IN6, (caddr_t)&nd) < 0) { warnmsg(LOG_ERR, __func__, "cannot set accept_rtadv flag"); close(s); return (-1); } close(s); } if ((ifi = malloc(sizeof(*ifi))) == NULL) { warnmsg(LOG_ERR, __func__, "memory allocation failed"); free(sdl); return (-1); } memset(ifi, 0, sizeof(*ifi)); ifi->sdl = sdl; ifi->ifi_rdnss = IFI_DNSOPT_STATE_NOINFO; ifi->ifi_dnssl = IFI_DNSOPT_STATE_NOINFO; TAILQ_INIT(&ifi->ifi_rainfo); strlcpy(ifi->ifname, ifname, sizeof(ifi->ifname)); /* construct a router solicitation message */ if (make_packet(ifi)) goto bad; /* set link ID of this interface. */ #ifdef HAVE_SCOPELIB if (inet_zoneid(AF_INET6, 2, ifname, &ifi->linkid)) goto bad; #else /* XXX: assume interface IDs as link IDs */ ifi->linkid = ifi->sdl->sdl_index; #endif /* * check if the interface is available. * also check if SIOCGIFMEDIA ioctl is OK on the interface. */ ifi->mediareqok = 1; ifi->active = interface_status(ifi); if (!ifi->mediareqok) { /* * probe routers periodically even if the link status * does not change. */ ifi->probeinterval = PROBE_INTERVAL; } /* activate interface: interface_up returns 0 on success */ flags = interface_up(ifi->ifname); if (flags == 0) ifi->state = IFS_DELAY; else if (flags == IFS_TENTATIVE) ifi->state = IFS_TENTATIVE; else ifi->state = IFS_DOWN; rtsol_timer_update(ifi); TAILQ_INSERT_TAIL(&ifinfo_head, ifi, ifi_next); return (0); bad: free(ifi->sdl); free(ifi); return (-1); } void iflist_init(void) { struct ifinfo *ifi; while ((ifi = TAILQ_FIRST(&ifinfo_head)) != NULL) { TAILQ_REMOVE(&ifinfo_head, ifi, ifi_next); if (ifi->sdl != NULL) free(ifi->sdl); if (ifi->rs_data != NULL) free(ifi->rs_data); free(ifi); } } #if 0 static int ifreconfig(char *ifname) { struct ifinfo *ifi, *prev; int rv; prev = NULL; TAILQ_FOREACH(ifi, &ifinfo_head, ifi_next) { if (strncmp(ifi->ifname, ifname, sizeof(ifi->ifname)) == 0) break; prev = ifi; } prev->next = ifi->next; rv = ifconfig(ifname); /* reclaim it after ifconfig() in case ifname is pointer inside ifi */ if (ifi->rs_data) free(ifi->rs_data); free(ifi->sdl); free(ifi); return (rv); } #endif struct rainfo * find_rainfo(struct ifinfo *ifi, struct sockaddr_in6 *sin6) { struct rainfo *rai; TAILQ_FOREACH(rai, &ifi->ifi_rainfo, rai_next) if (memcmp(&rai->rai_saddr.sin6_addr, &sin6->sin6_addr, sizeof(rai->rai_saddr.sin6_addr)) == 0) return (rai); return (NULL); } struct ifinfo * find_ifinfo(int ifindex) { struct ifinfo *ifi; TAILQ_FOREACH(ifi, &ifinfo_head, ifi_next) { if (ifi->sdl->sdl_index == ifindex) return (ifi); } return (NULL); } static int make_packet(struct ifinfo *ifi) { size_t packlen = sizeof(struct nd_router_solicit), lladdroptlen = 0; struct nd_router_solicit *rs; char *buf; if ((lladdroptlen = lladdropt_length(ifi->sdl)) == 0) { warnmsg(LOG_INFO, __func__, "link-layer address option has null length" " on %s. Treat as not included.", ifi->ifname); } packlen += lladdroptlen; ifi->rs_datalen = packlen; /* allocate buffer */ if ((buf = malloc(packlen)) == NULL) { warnmsg(LOG_ERR, __func__, "memory allocation failed for %s", ifi->ifname); return (-1); } ifi->rs_data = buf; /* fill in the message */ rs = (struct nd_router_solicit *)buf; rs->nd_rs_type = ND_ROUTER_SOLICIT; rs->nd_rs_code = 0; rs->nd_rs_cksum = 0; rs->nd_rs_reserved = 0; buf += sizeof(*rs); /* fill in source link-layer address option */ if (lladdroptlen) lladdropt_fill(ifi->sdl, (struct nd_opt_hdr *)buf); return (0); } -static struct timeval * +static struct timespec * rtsol_check_timer(void) { - static struct timeval returnval; - struct timeval now, rtsol_timer; + static struct timespec returnval; + struct timespec now, rtsol_timer; struct ifinfo *ifi; struct rainfo *rai; struct ra_opt *rao; int flags; - gettimeofday(&now, NULL); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); rtsol_timer = tm_max; TAILQ_FOREACH(ifi, &ifinfo_head, ifi_next) { - if (timercmp(&ifi->expire, &now, <=)) { + if (TS_CMP(&ifi->expire, &now, <=)) { warnmsg(LOG_DEBUG, __func__, "timer expiration on %s, " "state = %d", ifi->ifname, ifi->state); while((rai = TAILQ_FIRST(&ifi->ifi_rainfo)) != NULL) { /* Remove all RA options. */ TAILQ_REMOVE(&ifi->ifi_rainfo, rai, rai_next); while ((rao = TAILQ_FIRST(&rai->rai_ra_opt)) != NULL) { TAILQ_REMOVE(&rai->rai_ra_opt, rao, rao_next); if (rao->rao_msg != NULL) free(rao->rao_msg); free(rao); } free(rai); } switch (ifi->state) { case IFS_DOWN: case IFS_TENTATIVE: /* interface_up returns 0 on success */ flags = interface_up(ifi->ifname); if (flags == 0) ifi->state = IFS_DELAY; else if (flags == IFS_TENTATIVE) ifi->state = IFS_TENTATIVE; else ifi->state = IFS_DOWN; break; case IFS_IDLE: { int oldstatus = ifi->active; int probe = 0; ifi->active = interface_status(ifi); if (oldstatus != ifi->active) { warnmsg(LOG_DEBUG, __func__, "%s status is changed" " from %d to %d", ifi->ifname, oldstatus, ifi->active); probe = 1; ifi->state = IFS_DELAY; } else if (ifi->probeinterval && (ifi->probetimer -= ifi->timer.tv_sec) <= 0) { /* probe timer expired */ ifi->probetimer = ifi->probeinterval; probe = 1; ifi->state = IFS_PROBE; } /* * If we need a probe, clear the previous * status wrt the "other" configuration. */ if (probe) ifi->otherconfig = 0; if (probe && mobile_node) defrouter_probe(ifi); break; } case IFS_DELAY: ifi->state = IFS_PROBE; sendpacket(ifi); break; case IFS_PROBE: if (ifi->probes < MAX_RTR_SOLICITATIONS) sendpacket(ifi); else { warnmsg(LOG_INFO, __func__, "No answer after sending %d RSs", ifi->probes); ifi->probes = 0; ifi->state = IFS_IDLE; } break; } rtsol_timer_update(ifi); } else { /* Expiration check for RA options. */ int expire = 0; TAILQ_FOREACH(rai, &ifi->ifi_rainfo, rai_next) { TAILQ_FOREACH(rao, &rai->rai_ra_opt, rao_next) { warnmsg(LOG_DEBUG, __func__, "RA expiration timer: " "type=%d, msg=%s, expire=%s", rao->rao_type, (char *)rao->rao_msg, sec2str(&rao->rao_expire)); - if (timercmp(&now, &rao->rao_expire, + if (TS_CMP(&now, &rao->rao_expire, >=)) { warnmsg(LOG_DEBUG, __func__, "RA expiration timer: " "expired."); TAILQ_REMOVE(&rai->rai_ra_opt, rao, rao_next); if (rao->rao_msg != NULL) free(rao->rao_msg); free(rao); expire = 1; } } } if (expire) ra_opt_handler(ifi); } - if (timercmp(&ifi->expire, &rtsol_timer, <)) + if (TS_CMP(&ifi->expire, &rtsol_timer, <)) rtsol_timer = ifi->expire; } - if (timercmp(&rtsol_timer, &tm_max, ==)) { + if (TS_CMP(&rtsol_timer, &tm_max, ==)) { warnmsg(LOG_DEBUG, __func__, "there is no timer"); return (NULL); - } else if (timercmp(&rtsol_timer, &now, <)) + } else if (TS_CMP(&rtsol_timer, &now, <)) /* this may occur when the interval is too small */ - returnval.tv_sec = returnval.tv_usec = 0; + returnval.tv_sec = returnval.tv_nsec = 0; else - timersub(&rtsol_timer, &now, &returnval); + TS_SUB(&rtsol_timer, &now, &returnval); now.tv_sec += returnval.tv_sec; - now.tv_usec += returnval.tv_usec; + now.tv_nsec += returnval.tv_nsec; warnmsg(LOG_DEBUG, __func__, "New timer is %s", sec2str(&now)); return (&returnval); } void rtsol_timer_update(struct ifinfo *ifi) { #define MILLION 1000000 #define DADRETRY 10 /* XXX: adhoc */ long interval; - struct timeval now; + struct timespec now; bzero(&ifi->timer, sizeof(ifi->timer)); switch (ifi->state) { case IFS_DOWN: case IFS_TENTATIVE: if (++ifi->dadcount > DADRETRY) { ifi->dadcount = 0; ifi->timer.tv_sec = PROBE_INTERVAL; } else ifi->timer.tv_sec = 1; break; case IFS_IDLE: if (mobile_node) { /* XXX should be configurable */ ifi->timer.tv_sec = 3; } else ifi->timer = tm_max; /* stop timer(valid?) */ break; case IFS_DELAY: #ifndef HAVE_ARC4RANDOM interval = random() % (MAX_RTR_SOLICITATION_DELAY * MILLION); #else interval = arc4random_uniform(MAX_RTR_SOLICITATION_DELAY * MILLION); #endif ifi->timer.tv_sec = interval / MILLION; - ifi->timer.tv_usec = interval % MILLION; + ifi->timer.tv_nsec = (interval % MILLION) * 1000; break; case IFS_PROBE: if (ifi->probes < MAX_RTR_SOLICITATIONS) ifi->timer.tv_sec = RTR_SOLICITATION_INTERVAL; else { /* * After sending MAX_RTR_SOLICITATIONS solicitations, * we're just waiting for possible replies; there * will be no more solicitation. Thus, we change * the timer value to MAX_RTR_SOLICITATION_DELAY based * on RFC 2461, Section 6.3.7. */ ifi->timer.tv_sec = MAX_RTR_SOLICITATION_DELAY; } break; default: warnmsg(LOG_ERR, __func__, "illegal interface state(%d) on %s", ifi->state, ifi->ifname); return; } /* reset the timer */ - if (timercmp(&ifi->timer, &tm_max, ==)) { + if (TS_CMP(&ifi->timer, &tm_max, ==)) { ifi->expire = tm_max; warnmsg(LOG_DEBUG, __func__, "stop timer for %s", ifi->ifname); } else { - gettimeofday(&now, NULL); - timeradd(&now, &ifi->timer, &ifi->expire); + clock_gettime(CLOCK_MONOTONIC_FAST, &now); + TS_ADD(&now, &ifi->timer, &ifi->expire); now.tv_sec += ifi->timer.tv_sec; - now.tv_usec += ifi->timer.tv_usec; + now.tv_nsec += ifi->timer.tv_nsec; warnmsg(LOG_DEBUG, __func__, "set timer for %s to %s", ifi->ifname, sec2str(&now)); } #undef MILLION } /* timer related utility functions */ #define MILLION 1000000 #ifndef SMALL static void rtsold_set_dump_file(int sig __unused) { do_dump = 1; } #endif static void usage(void) { #ifndef SMALL fprintf(stderr, "usage: rtsold [-adDfFm1] [-O script-name] " "[-P pidfile] [-R script-name] interfaces...\n"); fprintf(stderr, "usage: rtsold [-dDfFm1] [-O script-name] " "[-P pidfile] [-R script-name] -a\n"); #else fprintf(stderr, "usage: rtsol [-dDF] [-O script-name] " "[-P pidfile] [-R script-name] interfaces...\n"); fprintf(stderr, "usage: rtsol [-dDF] [-O script-name] " "[-P pidfile] [-R script-name] -a\n"); #endif } void warnmsg(int priority, const char *func, const char *msg, ...) { va_list ap; char buf[BUFSIZ]; va_start(ap, msg); if (fflag) { if (priority <= log_upto) { (void)vfprintf(stderr, msg, ap); (void)fprintf(stderr, "\n"); } } else { snprintf(buf, sizeof(buf), "<%s> %s", func, msg); msg = buf; vsyslog(priority, msg, ap); } va_end(ap); } /* * return a list of interfaces which is suitable to sending an RS. */ char ** autoifprobe(void) { static char **argv = NULL; static int n = 0; char **a; int s = 0, i, found; struct ifaddrs *ifap, *ifa; struct in6_ndireq nd; /* initialize */ while (n--) free(argv[n]); if (argv) { free(argv); argv = NULL; } n = 0; if (getifaddrs(&ifap) != 0) return (NULL); if (!Fflag && (s = socket(AF_INET6, SOCK_DGRAM, 0)) < 0) { warnmsg(LOG_ERR, __func__, "socket"); exit(1); } /* find an ethernet */ for (ifa = ifap; ifa; ifa = ifa->ifa_next) { if ((ifa->ifa_flags & IFF_UP) == 0) continue; if ((ifa->ifa_flags & IFF_POINTOPOINT) != 0) continue; if ((ifa->ifa_flags & IFF_LOOPBACK) != 0) continue; if ((ifa->ifa_flags & IFF_MULTICAST) == 0) continue; if (ifa->ifa_addr->sa_family != AF_INET6) continue; found = 0; for (i = 0; i < n; i++) { if (strcmp(argv[i], ifa->ifa_name) == 0) { found++; break; } } if (found) continue; /* * Skip the interfaces which IPv6 and/or accepting RA * is disabled. */ if (!Fflag) { memset(&nd, 0, sizeof(nd)); strlcpy(nd.ifname, ifa->ifa_name, sizeof(nd.ifname)); if (ioctl(s, SIOCGIFINFO_IN6, (caddr_t)&nd) < 0) { warnmsg(LOG_ERR, __func__, "ioctl(SIOCGIFINFO_IN6)"); exit(1); } if ((nd.ndi.flags & ND6_IFF_IFDISABLED)) continue; if (!(nd.ndi.flags & ND6_IFF_ACCEPT_RTADV)) continue; } /* if we find multiple candidates, just warn. */ if (n != 0 && dflag > 1) warnmsg(LOG_WARNING, __func__, "multiple interfaces found"); a = (char **)realloc(argv, (n + 1) * sizeof(char **)); if (a == NULL) { warnmsg(LOG_ERR, __func__, "realloc"); exit(1); } argv = a; argv[n] = strdup(ifa->ifa_name); if (!argv[n]) { warnmsg(LOG_ERR, __func__, "malloc"); exit(1); } n++; } if (n) { a = (char **)realloc(argv, (n + 1) * sizeof(char **)); if (a == NULL) { warnmsg(LOG_ERR, __func__, "realloc"); exit(1); } argv = a; argv[n] = NULL; if (dflag > 0) { for (i = 0; i < n; i++) warnmsg(LOG_WARNING, __func__, "probing %s", argv[i]); } } if (!Fflag) close(s); freeifaddrs(ifap); return (argv); } Index: head/usr.sbin/rtsold/rtsold.h =================================================================== --- head/usr.sbin/rtsold/rtsold.h (revision 253969) +++ head/usr.sbin/rtsold/rtsold.h (revision 253970) @@ -1,170 +1,193 @@ /* $KAME: rtsold.h,v 1.19 2003/04/16 09:48:15 itojun Exp $ */ /* * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $FreeBSD$ */ struct script_msg { TAILQ_ENTRY(script_msg) sm_next; char *sm_msg; }; TAILQ_HEAD(script_msg_head_t, script_msg); struct ra_opt { TAILQ_ENTRY(ra_opt) rao_next; u_int8_t rao_type; - struct timeval rao_expire; + struct timespec rao_expire; size_t rao_len; void *rao_msg; }; TAILQ_HEAD(rainfo_head, ra_opt); struct rainfo { TAILQ_ENTRY(rainfo) rai_next; struct ifinfo *rai_ifinfo; struct sockaddr_in6 rai_saddr; TAILQ_HEAD(, ra_opt) rai_ra_opt; }; struct ifinfo { TAILQ_ENTRY(ifinfo) ifi_next; /* pointer to the next interface */ struct sockaddr_dl *sdl; /* link-layer address */ char ifname[IFNAMSIZ]; /* interface name */ u_int32_t linkid; /* link ID of this interface */ int active; /* interface status */ int probeinterval; /* interval of probe timer (if necessary) */ int probetimer; /* rest of probe timer */ int mediareqok; /* wheter the IF supports SIOCGIFMEDIA */ int otherconfig; /* need a separate protocol for the "other" * configuration */ int state; int probes; int dadcount; - struct timeval timer; - struct timeval expire; + struct timespec timer; + struct timespec expire; int errors; /* # of errors we've got - detect wedge */ #define IFI_DNSOPT_STATE_NOINFO 0 #define IFI_DNSOPT_STATE_RECEIVED 1 int ifi_rdnss; /* RDNSS option state */ int ifi_dnssl; /* DNSSL option state */ int racnt; /* total # of valid RAs it have got */ TAILQ_HEAD(, rainfo) ifi_rainfo; size_t rs_datalen; u_char *rs_data; }; /* per interface status */ #define IFS_IDLE 0 #define IFS_DELAY 1 #define IFS_PROBE 2 #define IFS_DOWN 3 #define IFS_TENTATIVE 4 /* Interface list */ extern TAILQ_HEAD(ifinfo_head_t, ifinfo) ifinfo_head; #define DNSINFO_ORIGIN_LABEL "slaac" /* * RFC 3542 API deprecates IPV6_PKTINFO in favor of * IPV6_RECVPKTINFO */ #ifndef IPV6_RECVPKTINFO #ifdef IPV6_PKTINFO #define IPV6_RECVPKTINFO IPV6_PKTINFO #endif #endif /* * RFC 3542 API deprecates IPV6_HOPLIMIT in favor of * IPV6_RECVHOPLIMIT */ #ifndef IPV6_RECVHOPLIMIT #ifdef IPV6_HOPLIMIT #define IPV6_RECVHOPLIMIT IPV6_HOPLIMIT #endif #endif #ifndef IN6ADDR_LINKLOCAL_ALLROUTERS_INIT #define IN6ADDR_LINKLOCAL_ALLROUTERS_INIT \ {{{ 0xff, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, \ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02 }}} #endif +#define TS_CMP(tsp, usp, cmp) \ + (((tsp)->tv_sec == (usp)->tv_sec) ? \ + ((tsp)->tv_nsec cmp (usp)->tv_nsec) : \ + ((tsp)->tv_sec cmp (usp)->tv_sec)) +#define TS_ADD(tsp, usp, vsp) \ + do { \ + (vsp)->tv_sec = (tsp)->tv_sec + (usp)->tv_sec; \ + (vsp)->tv_nsec = (tsp)->tv_nsec + (usp)->tv_nsec; \ + if ((vsp)->tv_nsec >= 1000000000L) { \ + (vsp)->tv_sec++; \ + (vsp)->tv_nsec -= 1000000000L; \ + } \ + } while (0) +#define TS_SUB(tsp, usp, vsp) \ + do { \ + (vsp)->tv_sec = (tsp)->tv_sec - (usp)->tv_sec; \ + (vsp)->tv_nsec = (tsp)->tv_nsec - (usp)->tv_nsec; \ + if ((vsp)->tv_nsec < 0) { \ + (vsp)->tv_sec--; \ + (vsp)->tv_nsec += 1000000000L; \ + } \ + } while (0) + /* rtsold.c */ -extern struct timeval tm_max; +extern struct timespec tm_max; extern int dflag; extern int aflag; extern int Fflag; extern int uflag; extern const char *otherconf_script; extern const char *resolvconf_script; extern int ifconfig(char *); extern void iflist_init(void); struct ifinfo *find_ifinfo(int); struct rainfo *find_rainfo(struct ifinfo *, struct sockaddr_in6 *); void rtsol_timer_update(struct ifinfo *); extern void warnmsg(int, const char *, const char *, ...) __attribute__((__format__(__printf__, 3, 4))); extern char **autoifprobe(void); extern int ra_opt_handler(struct ifinfo *); /* if.c */ extern int ifinit(void); extern int interface_up(char *); extern int interface_status(struct ifinfo *); extern int lladdropt_length(struct sockaddr_dl *); extern void lladdropt_fill(struct sockaddr_dl *, struct nd_opt_hdr *); extern struct sockaddr_dl *if_nametosdl(char *); extern int getinet6sysctl(int); extern int setinet6sysctl(int, int); /* rtsol.c */ extern int sockopen(void); extern void sendpacket(struct ifinfo *); extern void rtsol_input(int); /* probe.c */ extern int probe_init(void); extern void defrouter_probe(struct ifinfo *); /* dump.c */ extern void rtsold_dump_file(const char *); -extern const char *sec2str(const struct timeval *); +extern const char *sec2str(const struct timespec *); /* rtsock.c */ extern int rtsock_open(void); extern int rtsock_input(int);