Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files. =================================================================== RCS file: /ftp/cvs/cvsroot/src/sys/netinet6/ip6_input.c,v rcsdiff: /ftp/cvs/cvsroot/src/sys/netinet6/ip6_input.c,v: warning: Unknown phrases like `commitid ...;' are present. retrieving revision 1.22.2.4 retrieving revision 1.118.2.2 diff -u -p -r1.22.2.4 -r1.118.2.2 --- src/sys/netinet6/ip6_input.c 2001/04/06 01:37:35 1.22.2.4 +++ src/sys/netinet6/ip6_input.c 2009/05/04 08:14:18 1.118.2.2 @@ -1,5 +1,5 @@ -/* $NetBSD: ip6_input.c,v 1.22.2.4 2001/04/06 01:37:35 he Exp $ */ -/* $KAME: ip6_input.c,v 1.119 2000/08/26 10:00:45 itojun Exp $ */ +/* $NetBSD: ip6_input.c,v 1.118.2.2 2009/05/04 08:14:18 yamt Exp $ */ +/* $KAME: ip6_input.c,v 1.188 2001/03/29 05:34:31 itojun Exp $ */ /* * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. @@ -42,11 +42,7 @@ * 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 the University of - * California, Berkeley and its contributors. - * 4. Neither the name of the University nor the names of its contributors + * 3. 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. * @@ -65,9 +61,14 @@ * @(#)ip_input.c 8.2 (Berkeley) 1/4/94 */ +#include +__KERNEL_RCSID(0, "$NetBSD: ip6_input.c,v 1.118.2.2 2009/05/04 08:14:18 yamt Exp $"); + #include "opt_inet.h" +#include "opt_inet6.h" #include "opt_ipsec.h" #include "opt_pfil_hooks.h" +#include "opt_compat_netbsd.h" #include #include @@ -82,6 +83,7 @@ #include #include #include +#include #include #include @@ -97,32 +99,41 @@ #ifdef INET #include #include -#endif /*INET*/ +#endif /* INET */ #include #include #include +#include #include #include +#include #include #include -#include - -#ifdef IPV6FIREWALL -#include -#endif #ifdef IPSEC #include +#include +#endif + +#ifdef FAST_IPSEC +#include +#include +#include +#endif /* FAST_IPSEC */ + +#ifdef COMPAT_50 +#include +#include #endif #include -/* we need it for NLOOP. */ -#include "loop.h" #include "faith.h" - #include "gif.h" -#include "bpfilter.h" + +#if NGIF > 0 +#include +#endif #include @@ -133,159 +144,196 @@ static int ip6qmaxlen = IFQ_MAXLEN; struct in6_ifaddr *in6_ifaddr; struct ifqueue ip6intrq; -extern struct ifnet loif[NLOOP]; +extern callout_t in6_tmpaddrtimer_ch; + int ip6_forward_srcrt; /* XXX */ int ip6_sourcecheck; /* XXX */ int ip6_sourcecheck_interval; /* XXX */ -#ifdef IPV6FIREWALL -/* firewall hooks */ -ip6_fw_chk_t *ip6_fw_chk_ptr; -ip6_fw_ctl_t *ip6_fw_ctl_ptr; +#ifdef PFIL_HOOKS +struct pfil_head inet6_pfil_hook; #endif -struct ip6stat ip6stat; +percpu_t *ip6stat_percpu; -static void ip6_init2 __P((void *)); +static void ip6_init2(void *); +static struct m_tag *ip6_setdstifaddr(struct mbuf *, const struct in6_ifaddr *); -static int ip6_hopopts_input __P((u_int32_t *, u_int32_t *, struct mbuf **, int *)); +static int ip6_hopopts_input(u_int32_t *, u_int32_t *, struct mbuf **, int *); +static struct mbuf *ip6_pullexthdr(struct mbuf *, size_t, int); /* * IP6 initialization: fill in IP6 protocol switch table. * All protocols not implemented in kernel go to raw IP6 protocol handler. */ void -ip6_init() +ip6_init(void) { - register struct ip6protosw *pr; - register int i; - struct timeval tv; + const struct ip6protosw *pr; + int i; - pr = (struct ip6protosw *)pffindproto(PF_INET6, IPPROTO_RAW, SOCK_RAW); + pr = (const struct ip6protosw *)pffindproto(PF_INET6, IPPROTO_RAW, SOCK_RAW); if (pr == 0) panic("ip6_init"); for (i = 0; i < IPPROTO_MAX; i++) ip6_protox[i] = pr - inet6sw; - for (pr = (struct ip6protosw *)inet6domain.dom_protosw; - pr < (struct ip6protosw *)inet6domain.dom_protoswNPROTOSW; pr++) + for (pr = (const struct ip6protosw *)inet6domain.dom_protosw; + pr < (const struct ip6protosw *)inet6domain.dom_protoswNPROTOSW; pr++) if (pr->pr_domain->dom_family == PF_INET6 && pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW) ip6_protox[pr->pr_protocol] = pr - inet6sw; ip6intrq.ifq_maxlen = ip6qmaxlen; + scope6_init(); + addrsel_policy_init(); nd6_init(); frag6_init(); -#ifdef IPV6FIREWALL - ip6_fw_init(); -#endif - /* - * in many cases, random() here does NOT return random number - * as initialization during bootstrap time occur in fixed order. - */ - microtime(&tv); - ip6_flow_seq = random() ^ tv.tv_usec; + ip6_desync_factor = arc4random() % MAX_TEMP_DESYNC_FACTOR; ip6_init2((void *)0); +#ifdef GATEWAY + ip6flow_init(ip6_hashsize); +#endif + +#ifdef PFIL_HOOKS + /* Register our Packet Filter hook. */ + inet6_pfil_hook.ph_type = PFIL_TYPE_AF; + inet6_pfil_hook.ph_af = AF_INET6; + i = pfil_head_register(&inet6_pfil_hook); + if (i != 0) + printf("ip6_init: WARNING: unable to register pfil hook, " + "error %d\n", i); +#endif /* PFIL_HOOKS */ + + ip6stat_percpu = percpu_alloc(sizeof(uint64_t) * IP6_NSTATS); } static void -ip6_init2(dummy) - void *dummy; +ip6_init2(void *dummy) { - /* - * to route local address of p2p link to loopback, - * assign loopback address first. - */ - in6_ifattach(&loif[0], NULL); /* nd6_timer_init */ - callout_init(&nd6_timer_ch); + callout_init(&nd6_timer_ch, CALLOUT_MPSAFE); callout_reset(&nd6_timer_ch, hz, nd6_timer, NULL); - /* router renumbering prefix list maintenance */ - callout_init(&in6_rr_timer_ch); - callout_reset(&in6_rr_timer_ch, hz, in6_rr_timer, NULL); + + /* timer for regeneranation of temporary addresses randomize ID */ + callout_init(&in6_tmpaddrtimer_ch, CALLOUT_MPSAFE); + callout_reset(&in6_tmpaddrtimer_ch, + (ip6_temp_preferred_lifetime - ip6_desync_factor - + ip6_temp_regen_advance) * hz, + in6_tmpaddrtimer, NULL); } /* * IP6 input interrupt handling. Just pass the packet to ip6_input. */ void -ip6intr() +ip6intr(void) { int s; struct mbuf *m; + mutex_enter(softnet_lock); + KERNEL_LOCK(1, NULL); for (;;) { - s = splimp(); + s = splnet(); IF_DEQUEUE(&ip6intrq, m); splx(s); if (m == 0) - return; + break; + /* drop the packet if IPv6 operation is disabled on the IF */ + if ((ND_IFINFO(m->m_pkthdr.rcvif)->flags & ND6_IFF_IFDISABLED)) { + m_freem(m); + break; + } ip6_input(m); } + KERNEL_UNLOCK_ONE(NULL); + mutex_exit(softnet_lock); } -extern struct route_in6 ip6_forward_rt; +extern struct route ip6_forward_rt; void -ip6_input(m) - struct mbuf *m; +ip6_input(struct mbuf *m) { struct ip6_hdr *ip6; - int off = sizeof(struct ip6_hdr), nest; + int hit, off = sizeof(struct ip6_hdr), nest; u_int32_t plen; u_int32_t rtalert = ~0; - int nxt, ours = 0; + int nxt, ours = 0, rh_present = 0; struct ifnet *deliverifp = NULL; -#ifdef PFIL_HOOKS - struct packet_filter_hook *pfh; - struct mbuf *m0; - int rv; -#endif /* PFIL_HOOKS */ + int srcrt = 0; + const struct rtentry *rt; + union { + struct sockaddr dst; + struct sockaddr_in6 dst6; + } u; +#ifdef FAST_IPSEC + struct m_tag *mtag; + struct tdb_ident *tdbi; + struct secpolicy *sp; + int s, error; +#endif #ifdef IPSEC /* * should the inner packet be considered authentic? * see comment in ah4_input(). */ - if (m) { - m->m_flags &= ~M_AUTHIPHDR; - m->m_flags &= ~M_AUTHIPDGM; - } + m->m_flags &= ~M_AUTHIPHDR; + m->m_flags &= ~M_AUTHIPDGM; #endif /* - * mbuf statistics by kazu + * make sure we don't have onion peering information into m_tag. + */ + ip6_delaux(m); + + /* + * mbuf statistics */ if (m->m_flags & M_EXT) { if (m->m_next) - ip6stat.ip6s_mext2m++; + IP6_STATINC(IP6_STAT_MEXT2M); else - ip6stat.ip6s_mext1++; + IP6_STATINC(IP6_STAT_MEXT1); } else { +#define M2MMAX 32 if (m->m_next) { if (m->m_flags & M_LOOP) { - ip6stat.ip6s_m2m[loif[0].if_index]++; /*XXX*/ - } else if (m->m_pkthdr.rcvif->if_index <= 31) - ip6stat.ip6s_m2m[m->m_pkthdr.rcvif->if_index]++; - else - ip6stat.ip6s_m2m[0]++; + /*XXX*/ IP6_STATINC(IP6_STAT_M2M + lo0ifp->if_index); + } else if (m->m_pkthdr.rcvif->if_index < M2MMAX) { + IP6_STATINC(IP6_STAT_M2M + + m->m_pkthdr.rcvif->if_index); + } else + IP6_STATINC(IP6_STAT_M2M); } else - ip6stat.ip6s_m1++; + IP6_STATINC(IP6_STAT_M1); +#undef M2MMAX } in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_receive); - ip6stat.ip6s_total++; + IP6_STATINC(IP6_STAT_TOTAL); -#ifndef PULLDOWN_TEST - /* XXX is the line really necessary? */ - IP6_EXTHDR_CHECK(m, 0, sizeof(struct ip6_hdr), /*nothing*/); -#endif - - if (m->m_len < sizeof(struct ip6_hdr)) { - struct ifnet *inifp; - inifp = m->m_pkthdr.rcvif; - if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == 0) { - ip6stat.ip6s_toosmall++; + /* + * If the IPv6 header is not aligned, slurp it up into a new + * mbuf with space for link headers, in the event we forward + * it. Otherwise, if it is aligned, make sure the entire base + * IPv6 header is in the first mbuf of the chain. + */ + if (IP6_HDR_ALIGNED_P(mtod(m, void *)) == 0) { + struct ifnet *inifp = m->m_pkthdr.rcvif; + if ((m = m_copyup(m, sizeof(struct ip6_hdr), + (max_linkhdr + 3) & ~3)) == NULL) { + /* XXXJRT new stat, please */ + IP6_STATINC(IP6_STAT_TOOSMALL); + in6_ifstat_inc(inifp, ifs6_in_hdrerr); + return; + } + } else if (__predict_false(m->m_len < sizeof(struct ip6_hdr))) { + struct ifnet *inifp = m->m_pkthdr.rcvif; + if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) { + IP6_STATINC(IP6_STAT_TOOSMALL); in6_ifstat_inc(inifp, ifs6_in_hdrerr); return; } @@ -294,11 +342,22 @@ ip6_input(m) ip6 = mtod(m, struct ip6_hdr *); if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) { - ip6stat.ip6s_badvers++; + IP6_STATINC(IP6_STAT_BADVERS); in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr); goto bad; } +#if defined(IPSEC) + /* IPv6 fast forwarding is not compatible with IPsec. */ + m->m_flags &= ~M_CANFASTFWD; +#else + /* + * Assume that we can create a fast-forward IP flow entry + * based on this packet. + */ + m->m_flags |= M_CANFASTFWD; +#endif + #ifdef PFIL_HOOKS /* * Run through list of hooks for input packets. If there are any @@ -307,67 +366,67 @@ ip6_input(m) * Note that filters must _never_ set this flag, as another filter * in the list may have previously cleared it. */ - m0 = m; + /* + * let ipfilter look at packet on the wire, + * not the decapsulated packet. + */ #ifdef IPSEC - if (ipsec_gethist(m, NULL)) - pfh = NULL; - else - pfh = pfil_hook_get(PFIL_IN, - &inetsw[ip_protox[IPPROTO_IPV6]].pr_pfh); + if (!ipsec_getnhist(m)) +#elif defined(FAST_IPSEC) + if (!ipsec_indone(m)) #else - pfh = pfil_hook_get(PFIL_IN, &inetsw[ip_protox[IPPROTO_IPV6]].pr_pfh); + if (1) #endif - for (; pfh; pfh = pfh->pfil_link.tqe_next) - if (pfh->pfil_func) { - rv = pfh->pfil_func(ip6, sizeof(*ip6), - m->m_pkthdr.rcvif, 0, &m0); - if (rv) - return; - m = m0; - if (m == NULL) - return; - ip6 = mtod(m, struct ip6_hdr *); - } -#endif /* PFIL_HOOKS */ + { + struct in6_addr odst; + odst = ip6->ip6_dst; + if (pfil_run_hooks(&inet6_pfil_hook, &m, m->m_pkthdr.rcvif, + PFIL_IN) != 0) + return; + if (m == NULL) + return; + ip6 = mtod(m, struct ip6_hdr *); + srcrt = !IN6_ARE_ADDR_EQUAL(&odst, &ip6->ip6_dst); + } +#endif /* PFIL_HOOKS */ - ip6stat.ip6s_nxthist[ip6->ip6_nxt]++; + IP6_STATINC(IP6_STAT_NXTHIST + ip6->ip6_nxt); -#ifdef IPV6FIREWALL - /* - * Check with the firewall... - */ - if (ip6_fw_chk_ptr) { - u_short port = 0; - /* If ipfw says divert, we have to just drop packet */ - /* use port as a dummy argument */ - if ((*ip6_fw_chk_ptr)(&ip6, NULL, &port, &m)) { - m_freem(m); - m = NULL; - } - if (!m) - return; +#ifdef ALTQ + if (altq_input != NULL && (*altq_input)(m, AF_INET6) == 0) { + /* packet is dropped by traffic conditioner */ + return; } #endif /* - * Scope check + * Check against address spoofing/corruption. */ if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_src) || IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_dst)) { - ip6stat.ip6s_badscope++; + /* + * XXX: "badscope" is not very suitable for a multicast source. + */ + IP6_STATINC(IP6_STAT_BADSCOPE); in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); goto bad; } /* - * The following check is not documented in the spec. Malicious party - * may be able to use IPv4 mapped addr to confuse tcp/udp stack and - * bypass security checks (act as if it was from 127.0.0.1 by using - * IPv6 src ::ffff:127.0.0.1). Be cautious. + * The following check is not documented in specs. A malicious + * party may be able to use IPv4 mapped addr to confuse tcp/udp stack + * and bypass security checks (act as if it was from 127.0.0.1 by using + * IPv6 src ::ffff:127.0.0.1). Be cautious. + * + * This check chokes if we are in an SIIT cloud. As none of BSDs + * support IPv4-less kernel compilation, we cannot support SIIT + * environment at all. So, it makes more sense for us to reject any + * malicious packets for non-SIIT environment, than try to do a + * partial support for SIIT environment. */ if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) || IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) { - ip6stat.ip6s_badscope++; + IP6_STATINC(IP6_STAT_BADSCOPE); in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); goto bad; } @@ -381,50 +440,34 @@ ip6_input(m) */ if (IN6_IS_ADDR_V4COMPAT(&ip6->ip6_src) || IN6_IS_ADDR_V4COMPAT(&ip6->ip6_dst)) { - ip6stat.ip6s_badscope++; + IP6_STATINC(IP6_STAT_BADSCOPE); in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); goto bad; } #endif - if (IN6_IS_ADDR_LOOPBACK(&ip6->ip6_src) || - IN6_IS_ADDR_LOOPBACK(&ip6->ip6_dst)) { - if (m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) { - ours = 1; - deliverifp = m->m_pkthdr.rcvif; - goto hbhcheck; - } else { - ip6stat.ip6s_badscope++; - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); - goto bad; - } - } - -#ifndef FAKE_LOOPBACK_IF - if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0) -#else - if (1) -#endif - { - if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_src)) - ip6->ip6_src.s6_addr16[1] - = htons(m->m_pkthdr.rcvif->if_index); - if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_dst)) - ip6->ip6_dst.s6_addr16[1] - = htons(m->m_pkthdr.rcvif->if_index); - } /* - * XXX we need this since we do not have "goto ours" hack route - * for some of our ifaddrs on loopback interface. - * we should correct it by changing in6_ifattach to install - * "goto ours" hack route. + * Disambiguate address scope zones (if there is ambiguity). + * We first make sure that the original source or destination address + * is not in our internal form for scoped addresses. Such addresses + * are not necessarily invalid spec-wise, but we cannot accept them due + * to the usage conflict. + * in6_setscope() then also checks and rejects the cases where src or + * dst are the loopback address and the receiving interface + * is not loopback. */ - if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) != 0) { - if (IN6_IS_ADDR_LINKLOCAL(&ip6->ip6_dst)) { - ours = 1; - deliverifp = m->m_pkthdr.rcvif; - goto hbhcheck; - } + if (__predict_false( + m_makewritable(&m, 0, sizeof(struct ip6_hdr), M_DONTWAIT))) + goto bad; + ip6 = mtod(m, struct ip6_hdr *); + if (in6_clearscope(&ip6->ip6_src) || in6_clearscope(&ip6->ip6_dst)) { + IP6_STATINC(IP6_STAT_BADSCOPE); /* XXX */ + goto bad; + } + if (in6_setscope(&ip6->ip6_src, m->m_pkthdr.rcvif, NULL) || + in6_setscope(&ip6->ip6_dst, m->m_pkthdr.rcvif, NULL)) { + IP6_STATINC(IP6_STAT_BADSCOPE); + goto bad; } /* @@ -442,8 +485,10 @@ ip6_input(m) if (in6m) ours = 1; else if (!ip6_mrouter) { - ip6stat.ip6s_notmember++; - ip6stat.ip6s_cantforward++; + uint64_t *ip6s = IP6_STAT_GETREF(); + ip6s[IP6_STAT_NOTMEMBER]++; + ip6s[IP6_STAT_CANTFORWARD]++; + IP6_STAT_PUTREF(); in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard); goto bad; } @@ -451,30 +496,18 @@ ip6_input(m) goto hbhcheck; } + sockaddr_in6_init(&u.dst6, &ip6->ip6_dst, 0, 0, 0); + /* * Unicast check */ - if (ip6_forward_rt.ro_rt != NULL && - (ip6_forward_rt.ro_rt->rt_flags & RTF_UP) != 0 && - IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, - &ip6_forward_rt.ro_dst.sin6_addr)) - ; /* cache hit */ - else { - if (ip6_forward_rt.ro_rt) { - /* route is down or destination is different */ - RTFREE(ip6_forward_rt.ro_rt); - ip6_forward_rt.ro_rt = 0; - } - - bzero(&ip6_forward_rt.ro_dst, sizeof(struct sockaddr_in6)); - ip6_forward_rt.ro_dst.sin6_len = sizeof(struct sockaddr_in6); - ip6_forward_rt.ro_dst.sin6_family = AF_INET6; - ip6_forward_rt.ro_dst.sin6_addr = ip6->ip6_dst; - - rtalloc((struct route *)&ip6_forward_rt); - } + rt = rtcache_lookup2(&ip6_forward_rt, &u.dst, 1, &hit); + if (hit) + IP6_STATINC(IP6_STAT_FORWARD_CACHEHIT); + else + IP6_STATINC(IP6_STAT_FORWARD_CACHEMISS); -#define rt6_key(r) ((struct sockaddr_in6 *)((r)->rt_nodes->rn_key)) +#define rt6_getkey(__rt) satocsin6(rt_getkey(__rt)) /* * Accept the packet if the forwarding interface to the destination @@ -485,21 +518,19 @@ ip6_input(m) * But we think it's even useful in some situations, e.g. when using * a special daemon which wants to intercept the packet. */ - if (ip6_forward_rt.ro_rt && - (ip6_forward_rt.ro_rt->rt_flags & - (RTF_HOST|RTF_GATEWAY)) == RTF_HOST && + if (rt != NULL && + (rt->rt_flags & (RTF_HOST|RTF_GATEWAY)) == RTF_HOST && + !(rt->rt_flags & RTF_CLONED) && #if 0 /* * The check below is redundant since the comparison of * the destination and the key of the rtentry has * already done through looking up the routing table. */ - IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, - &rt6_key(ip6_forward_rt.ro_rt)->sin6_addr) && + IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, &rt6_getkey(rt)->sin6_addr) && #endif - ip6_forward_rt.ro_rt->rt_ifp->if_type == IFT_LOOP) { - struct in6_ifaddr *ia6 = - (struct in6_ifaddr *)ip6_forward_rt.ro_rt->rt_ifa; + rt->rt_ifp->if_type == IFT_LOOP) { + struct in6_ifaddr *ia6 = (struct in6_ifaddr *)rt->rt_ifa; if (ia6->ia6_flags & IN6_IFF_ANYCAST) m->m_flags |= M_ANYCAST6; /* @@ -513,25 +544,25 @@ ip6_input(m) goto hbhcheck; } else { /* address is not ready, so discard the packet. */ - log(LOG_INFO, - "ip6_input: packet to an unready address %s->%s", + nd6log((LOG_INFO, + "ip6_input: packet to an unready address %s->%s\n", ip6_sprintf(&ip6->ip6_src), - ip6_sprintf(&ip6->ip6_dst)); + ip6_sprintf(&ip6->ip6_dst))); goto bad; } } /* - * FAITH(Firewall Aided Internet Translator) + * FAITH (Firewall Aided Internet Translator) */ #if defined(NFAITH) && 0 < NFAITH if (ip6_keepfaith) { - if (ip6_forward_rt.ro_rt && ip6_forward_rt.ro_rt->rt_ifp - && ip6_forward_rt.ro_rt->rt_ifp->if_type == IFT_FAITH) { + if (rt != NULL && rt->rt_ifp != NULL && + rt->rt_ifp->if_type == IFT_FAITH) { /* XXX do we need more sanity checks? */ ours = 1; - deliverifp = ip6_forward_rt.ro_rt->rt_ifp; /*faith*/ + deliverifp = rt->rt_ifp; /* faith */ goto hbhcheck; } } @@ -545,9 +576,7 @@ ip6_input(m) * working right. */ struct ifaddr *ifa; - for (ifa = m->m_pkthdr.rcvif->if_addrlist.tqh_first; - ifa; - ifa = ifa->ifa_list.tqe_next) { + IFADDR_FOREACH(ifa, m->m_pkthdr.rcvif) { if (ifa->ifa_addr == NULL) continue; /* just for safety */ if (ifa->ifa_addr->sa_family != AF_INET6) @@ -566,13 +595,32 @@ ip6_input(m) * and we're not a router. */ if (!ip6_forwarding) { - ip6stat.ip6s_cantforward++; + IP6_STATINC(IP6_STAT_CANTFORWARD); in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard); goto bad; } hbhcheck: /* + * record address information into m_tag, if we don't have one yet. + * note that we are unable to record it, if the address is not listed + * as our interface address (e.g. multicast addresses, addresses + * within FAITH prefixes and such). + */ + if (deliverifp && ip6_getdstifaddr(m) == NULL) { + struct in6_ifaddr *ia6; + + ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst); + if (ia6 != NULL && ip6_setdstifaddr(m, ia6) == NULL) { + /* + * XXX maybe we should drop the packet here, + * as we could not provide enough information + * to the upper layers. + */ + } + } + + /* * Process Hop-by-Hop options header if it's contained. * m may be modified in ip6_hopopts_input(). * If a JumboPayload option is included, plen will also be modified. @@ -592,35 +640,31 @@ ip6_input(m) ip6 = mtod(m, struct ip6_hdr *); /* - * if the payload length field is 0 and the next header field + * if the payload length field is 0 and the next header field * indicates Hop-by-Hop Options header, then a Jumbo Payload * option MUST be included. */ if (ip6->ip6_plen == 0 && plen == 0) { /* * Note that if a valid jumbo payload option is - * contained, ip6_hoptops_input() must set a valid - * (non-zero) payload length to the variable plen. + * contained, ip6_hopopts_input() must set a valid + * (non-zero) payload length to the variable plen. */ - ip6stat.ip6s_badoptions++; + IP6_STATINC(IP6_STAT_BADOPTIONS); in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard); in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr); icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_HEADER, - (caddr_t)&ip6->ip6_plen - (caddr_t)ip6); + (char *)&ip6->ip6_plen - (char *)ip6); return; } -#ifndef PULLDOWN_TEST - /* ip6_hopopts_input() ensures that mbuf is contiguous */ - hbh = (struct ip6_hbh *)(ip6 + 1); -#else IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr), sizeof(struct ip6_hbh)); if (hbh == NULL) { - ip6stat.ip6s_tooshort++; + IP6_STATINC(IP6_STAT_TOOSHORT); return; } -#endif + KASSERT(IP6_HDR_ALIGNED_P(hbh)); nxt = hbh->ip6h_nxt; /* @@ -639,7 +683,7 @@ ip6_input(m) * Drop packet if shorter than we expect. */ if (m->m_pkthdr.len - sizeof(struct ip6_hdr) < plen) { - ip6stat.ip6s_tooshort++; + IP6_STATINC(IP6_STAT_TOOSHORT); in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_truncated); goto bad; } @@ -664,7 +708,7 @@ ip6_input(m) * must be discarded, else it may be accepted below. */ if (ip6_mrouter && ip6_mforward(ip6, m->m_pkthdr.rcvif, m)) { - ip6stat.ip6s_cantforward++; + IP6_STATINC(IP6_STAT_CANTFORWARD); m_freem(m); return; } @@ -673,27 +717,48 @@ ip6_input(m) return; } } else if (!ours) { - ip6_forward(m, 0); + ip6_forward(m, srcrt); return; - } + } + + ip6 = mtod(m, struct ip6_hdr *); + + /* + * Malicious party may be able to use IPv4 mapped addr to confuse + * tcp/udp stack and bypass security checks (act as if it was from + * 127.0.0.1 by using IPv6 src ::ffff:127.0.0.1). Be cautious. + * + * For SIIT end node behavior, you may want to disable the check. + * However, you will become vulnerable to attacks using IPv4 mapped + * source. + */ + if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) || + IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) { + IP6_STATINC(IP6_STAT_BADSCOPE); + in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); + goto bad; + } /* * Tell launch routine the next header */ #ifdef IFA_STATS - if (IFA_STATS && deliverifp != NULL) { + if (deliverifp != NULL) { struct in6_ifaddr *ia6; ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst); if (ia6) ia6->ia_ifa.ifa_data.ifad_inbytes += m->m_pkthdr.len; } #endif - ip6stat.ip6s_delivered++; + IP6_STATINC(IP6_STAT_DELIVERED); in6_ifstat_inc(deliverifp, ifs6_in_deliver); nest = 0; + + rh_present = 0; while (nxt != IPPROTO_DONE) { if (ip6_hdrnestlimit && (++nest > ip6_hdrnestlimit)) { - ip6stat.ip6s_toomanyhdr++; + IP6_STATINC(IP6_STAT_TOOMANYHDR); + in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr); goto bad; } @@ -702,11 +767,20 @@ ip6_input(m) * more sanity checks in header chain processing. */ if (m->m_pkthdr.len < off) { - ip6stat.ip6s_tooshort++; + IP6_STATINC(IP6_STAT_TOOSHORT); in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_truncated); goto bad; } + if (nxt == IPPROTO_ROUTING) { + if (rh_present++) { + in6_ifstat_inc(m->m_pkthdr.rcvif, + ifs6_in_hdrerr); + IP6_STATINC(IP6_STAT_BADOPTIONS); + goto bad; + } + } + #ifdef IPSEC /* * enforce IPsec policy checking if we are seeing last header. @@ -715,11 +789,50 @@ ip6_input(m) */ if ((inet6sw[ip6_protox[nxt]].pr_flags & PR_LASTHDR) != 0 && ipsec6_in_reject(m, NULL)) { - ipsec6stat.in_polvio++; + IPSEC6_STATINC(IPSEC_STAT_IN_POLVIO); goto bad; } #endif - +#ifdef FAST_IPSEC + /* + * enforce IPsec policy checking if we are seeing last header. + * note that we do not visit this with protocols with pcb layer + * code - like udp/tcp/raw ip. + */ + if ((inet6sw[ip_protox[nxt]].pr_flags & PR_LASTHDR) != 0) { + /* + * Check if the packet has already had IPsec processing + * done. If so, then just pass it along. This tag gets + * set during AH, ESP, etc. input handling, before the + * packet is returned to the ip input queue for delivery. + */ + mtag = m_tag_find(m, PACKET_TAG_IPSEC_IN_DONE, NULL); + s = splsoftnet(); + if (mtag != NULL) { + tdbi = (struct tdb_ident *)(mtag + 1); + sp = ipsec_getpolicy(tdbi, IPSEC_DIR_INBOUND); + } else { + sp = ipsec_getpolicybyaddr(m, IPSEC_DIR_INBOUND, + IP_FORWARDING, &error); + } + if (sp != NULL) { + /* + * Check security policy against packet attributes. + */ + error = ipsec_in_reject(sp, m); + KEY_FREESP(&sp); + } else { + /* XXX error stat??? */ + error = EINVAL; + DPRINTF(("ip6_input: no SP, packet discarded\n"));/*XXX*/ + } + splx(s); + if (error) + goto bad; + } +#endif /* FAST_IPSEC */ + + nxt = (*inet6sw[ip6_protox[nxt]].pr_input)(&m, &off, nxt); } return; @@ -728,55 +841,76 @@ ip6_input(m) } /* + * set/grab in6_ifaddr correspond to IPv6 destination address. + */ +static struct m_tag * +ip6_setdstifaddr(struct mbuf *m, const struct in6_ifaddr *ia) +{ + struct m_tag *mtag; + + mtag = ip6_addaux(m); + if (mtag != NULL) { + struct ip6aux *ip6a; + + ip6a = (struct ip6aux *)(mtag + 1); + in6_setscope(&ip6a->ip6a_src, ia->ia_ifp, &ip6a->ip6a_scope_id); + ip6a->ip6a_src = ia->ia_addr.sin6_addr; + ip6a->ip6a_flags = ia->ia6_flags; + } + return mtag; /* NULL if failed to set */ +} + +const struct ip6aux * +ip6_getdstifaddr(struct mbuf *m) +{ + struct m_tag *mtag; + + mtag = ip6_findaux(m); + if (mtag != NULL) + return (struct ip6aux *)(mtag + 1); + else + return NULL; +} + +/* * Hop-by-Hop options header processing. If a valid jumbo payload option is * included, the real payload length will be stored in plenp. + * + * rtalertp - XXX: should be stored more smart way */ static int -ip6_hopopts_input(plenp, rtalertp, mp, offp) - u_int32_t *plenp; - u_int32_t *rtalertp; /* XXX: should be stored more smart way */ - struct mbuf **mp; - int *offp; +ip6_hopopts_input(u_int32_t *plenp, u_int32_t *rtalertp, + struct mbuf **mp, int *offp) { - register struct mbuf *m = *mp; + struct mbuf *m = *mp; int off = *offp, hbhlen; struct ip6_hbh *hbh; - u_int8_t *opt; /* validation of the length of the header */ -#ifndef PULLDOWN_TEST - IP6_EXTHDR_CHECK(m, off, sizeof(*hbh), -1); - hbh = (struct ip6_hbh *)(mtod(m, caddr_t) + off); - hbhlen = (hbh->ip6h_len + 1) << 3; - - IP6_EXTHDR_CHECK(m, off, hbhlen, -1); - hbh = (struct ip6_hbh *)(mtod(m, caddr_t) + off); -#else IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr), sizeof(struct ip6_hbh)); if (hbh == NULL) { - ip6stat.ip6s_tooshort++; + IP6_STATINC(IP6_STAT_TOOSHORT); return -1; } hbhlen = (hbh->ip6h_len + 1) << 3; IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, sizeof(struct ip6_hdr), hbhlen); if (hbh == NULL) { - ip6stat.ip6s_tooshort++; + IP6_STATINC(IP6_STAT_TOOSHORT); return -1; } -#endif + KASSERT(IP6_HDR_ALIGNED_P(hbh)); off += hbhlen; hbhlen -= sizeof(struct ip6_hbh); - opt = (u_int8_t *)hbh + sizeof(struct ip6_hbh); if (ip6_process_hopopts(m, (u_int8_t *)hbh + sizeof(struct ip6_hbh), hbhlen, rtalertp, plenp) < 0) - return(-1); + return (-1); *offp = off; *mp = m; - return(0); + return (0); } /* @@ -784,80 +918,84 @@ ip6_hopopts_input(plenp, rtalertp, mp, o * This function is separate from ip6_hopopts_input() in order to * handle a case where the sending node itself process its hop-by-hop * options header. In such a case, the function is called from ip6_output(). + * + * The function assumes that hbh header is located right after the IPv6 header + * (RFC2460 p7), opthead is pointer into data content in m, and opthead to + * opthead + hbhlen is located in continuous memory region. */ int -ip6_process_hopopts(m, opthead, hbhlen, rtalertp, plenp) - struct mbuf *m; - u_int8_t *opthead; - int hbhlen; - u_int32_t *rtalertp; - u_int32_t *plenp; +ip6_process_hopopts(struct mbuf *m, u_int8_t *opthead, int hbhlen, + u_int32_t *rtalertp, u_int32_t *plenp) { struct ip6_hdr *ip6; int optlen = 0; u_int8_t *opt = opthead; u_int16_t rtalert_val; u_int32_t jumboplen; + const int erroff = sizeof(struct ip6_hdr) + sizeof(struct ip6_hbh); for (; hbhlen > 0; hbhlen -= optlen, opt += optlen) { - switch(*opt) { - case IP6OPT_PAD1: - optlen = 1; - break; - case IP6OPT_PADN: - if (hbhlen < IP6OPT_MINLEN) { - ip6stat.ip6s_toosmall++; - goto bad; - } - optlen = *(opt + 1) + 2; - break; - case IP6OPT_RTALERT: - /* XXX may need check for alignment */ - if (hbhlen < IP6OPT_RTALERT_LEN) { - ip6stat.ip6s_toosmall++; - goto bad; - } - if (*(opt + 1) != IP6OPT_RTALERT_LEN - 2) - /* XXX: should we discard the packet? */ - log(LOG_ERR, "length of router alert opt is inconsitent(%d)", - *(opt + 1)); - optlen = IP6OPT_RTALERT_LEN; - bcopy((caddr_t)(opt + 2), (caddr_t)&rtalert_val, 2); - *rtalertp = ntohs(rtalert_val); - break; - case IP6OPT_JUMBO: + switch (*opt) { + case IP6OPT_PAD1: + optlen = 1; + break; + case IP6OPT_PADN: + if (hbhlen < IP6OPT_MINLEN) { + IP6_STATINC(IP6_STAT_TOOSMALL); + goto bad; + } + optlen = *(opt + 1) + 2; + break; + case IP6OPT_RTALERT: + /* XXX may need check for alignment */ + if (hbhlen < IP6OPT_RTALERT_LEN) { + IP6_STATINC(IP6_STAT_TOOSMALL); + goto bad; + } + if (*(opt + 1) != IP6OPT_RTALERT_LEN - 2) { + /* XXX stat */ + icmp6_error(m, ICMP6_PARAM_PROB, + ICMP6_PARAMPROB_HEADER, + erroff + opt + 1 - opthead); + return (-1); + } + optlen = IP6OPT_RTALERT_LEN; + memcpy((void *)&rtalert_val, (void *)(opt + 2), 2); + *rtalertp = ntohs(rtalert_val); + break; + case IP6OPT_JUMBO: /* XXX may need check for alignment */ if (hbhlen < IP6OPT_JUMBO_LEN) { - ip6stat.ip6s_toosmall++; + IP6_STATINC(IP6_STAT_TOOSMALL); goto bad; } - if (*(opt + 1) != IP6OPT_JUMBO_LEN - 2) - /* XXX: should we discard the packet? */ - log(LOG_ERR, "length of jumbopayload opt " - "is inconsistent(%d)", - *(opt + 1)); + if (*(opt + 1) != IP6OPT_JUMBO_LEN - 2) { + /* XXX stat */ + icmp6_error(m, ICMP6_PARAM_PROB, + ICMP6_PARAMPROB_HEADER, + erroff + opt + 1 - opthead); + return (-1); + } optlen = IP6OPT_JUMBO_LEN; /* * IPv6 packets that have non 0 payload length - * must not contain a jumbo paylod option. + * must not contain a jumbo payload option. */ ip6 = mtod(m, struct ip6_hdr *); if (ip6->ip6_plen) { - ip6stat.ip6s_badoptions++; + IP6_STATINC(IP6_STAT_BADOPTIONS); icmp6_error(m, ICMP6_PARAM_PROB, - ICMP6_PARAMPROB_HEADER, - sizeof(struct ip6_hdr) + - sizeof(struct ip6_hbh) + - opt - opthead); - return(-1); + ICMP6_PARAMPROB_HEADER, + erroff + opt - opthead); + return (-1); } /* * We may see jumbolen in unaligned location, so * we'd need to perform bcopy(). */ - bcopy(opt + 2, &jumboplen, sizeof(jumboplen)); + memcpy(&jumboplen, opt + 2, sizeof(jumboplen)); jumboplen = (u_int32_t)htonl(jumboplen); #if 1 @@ -870,13 +1008,11 @@ ip6_process_hopopts(m, opthead, hbhlen, * there's no explicit mention in specification. */ if (*plenp != 0) { - ip6stat.ip6s_badoptions++; + IP6_STATINC(IP6_STAT_BADOPTIONS); icmp6_error(m, ICMP6_PARAM_PROB, - ICMP6_PARAMPROB_HEADER, - sizeof(struct ip6_hdr) + - sizeof(struct ip6_hbh) + - opt + 2 - opthead); - return(-1); + ICMP6_PARAMPROB_HEADER, + erroff + opt + 2 - opthead); + return (-1); } #endif @@ -884,37 +1020,34 @@ ip6_process_hopopts(m, opthead, hbhlen, * jumbo payload length must be larger than 65535. */ if (jumboplen <= IPV6_MAXPACKET) { - ip6stat.ip6s_badoptions++; + IP6_STATINC(IP6_STAT_BADOPTIONS); icmp6_error(m, ICMP6_PARAM_PROB, - ICMP6_PARAMPROB_HEADER, - sizeof(struct ip6_hdr) + - sizeof(struct ip6_hbh) + - opt + 2 - opthead); - return(-1); + ICMP6_PARAMPROB_HEADER, + erroff + opt + 2 - opthead); + return (-1); } *plenp = jumboplen; break; - default: /* unknown option */ - if (hbhlen < IP6OPT_MINLEN) { - ip6stat.ip6s_toosmall++; - goto bad; - } - if ((optlen = ip6_unknown_opt(opt, m, - sizeof(struct ip6_hdr) + - sizeof(struct ip6_hbh) + - opt - opthead)) == -1) - return(-1); - optlen += 2; - break; + default: /* unknown option */ + if (hbhlen < IP6OPT_MINLEN) { + IP6_STATINC(IP6_STAT_TOOSMALL); + goto bad; + } + optlen = ip6_unknown_opt(opt, m, + erroff + opt - opthead); + if (optlen == -1) + return (-1); + optlen += 2; + break; } } - return(0); + return (0); bad: m_freem(m); - return(-1); + return (-1); } /* @@ -924,37 +1057,34 @@ ip6_process_hopopts(m, opthead, hbhlen, * is not continuous in order to return an ICMPv6 error. */ int -ip6_unknown_opt(optp, m, off) - u_int8_t *optp; - struct mbuf *m; - int off; +ip6_unknown_opt(u_int8_t *optp, struct mbuf *m, int off) { struct ip6_hdr *ip6; - switch(IP6OPT_TYPE(*optp)) { - case IP6OPT_TYPE_SKIP: /* ignore the option */ - return((int)*(optp + 1)); - case IP6OPT_TYPE_DISCARD: /* silently discard */ - m_freem(m); - return(-1); - case IP6OPT_TYPE_FORCEICMP: /* send ICMP even if multicasted */ - ip6stat.ip6s_badoptions++; - icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_OPTION, off); - return(-1); - case IP6OPT_TYPE_ICMP: /* send ICMP if not multicasted */ - ip6stat.ip6s_badoptions++; - ip6 = mtod(m, struct ip6_hdr *); - if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) || - (m->m_flags & (M_BCAST|M_MCAST))) - m_freem(m); - else - icmp6_error(m, ICMP6_PARAM_PROB, - ICMP6_PARAMPROB_OPTION, off); - return(-1); + switch (IP6OPT_TYPE(*optp)) { + case IP6OPT_TYPE_SKIP: /* ignore the option */ + return ((int)*(optp + 1)); + case IP6OPT_TYPE_DISCARD: /* silently discard */ + m_freem(m); + return (-1); + case IP6OPT_TYPE_FORCEICMP: /* send ICMP even if multicasted */ + IP6_STATINC(IP6_STAT_BADOPTIONS); + icmp6_error(m, ICMP6_PARAM_PROB, ICMP6_PARAMPROB_OPTION, off); + return (-1); + case IP6OPT_TYPE_ICMP: /* send ICMP if not multicasted */ + IP6_STATINC(IP6_STAT_BADOPTIONS); + ip6 = mtod(m, struct ip6_hdr *); + if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) || + (m->m_flags & (M_BCAST|M_MCAST))) + m_freem(m); + else + icmp6_error(m, ICMP6_PARAM_PROB, + ICMP6_PARAMPROB_OPTION, off); + return (-1); } m_freem(m); /* XXX: NOTREACHED */ - return(-1); + return (-1); } /* @@ -969,78 +1099,88 @@ ip6_unknown_opt(optp, m, off) * you are using IP6_EXTHDR_CHECK() not m_pulldown()) */ void -ip6_savecontrol(in6p, mp, ip6, m) - register struct in6pcb *in6p; - register struct mbuf **mp; - register struct ip6_hdr *ip6; - register struct mbuf *m; -{ - struct proc *p = curproc; /* XXX */ - int privileged; - - privileged = 0; - if (p && !suser(p->p_ucred, &p->p_acflag)) - privileged++; +ip6_savecontrol(struct in6pcb *in6p, struct mbuf **mp, + struct ip6_hdr *ip6, struct mbuf *m) +{ +#ifdef RFC2292 +#define IS2292(x, y) ((in6p->in6p_flags & IN6P_RFC2292) ? (x) : (y)) +#else +#define IS2292(x, y) (y) +#endif -#ifdef SO_TIMESTAMP - if (in6p->in6p_socket->so_options & SO_TIMESTAMP) { + if (in6p->in6p_socket->so_options & SO_TIMESTAMP +#ifdef SO_OTIMESTAMP + || in6p->in6p_socket->so_options & SO_OTIMESTAMP +#endif + ) { struct timeval tv; microtime(&tv); - *mp = sbcreatecontrol((caddr_t) &tv, sizeof(tv), - SCM_TIMESTAMP, SOL_SOCKET); - if (*mp) - mp = &(*mp)->m_next; - } +#ifdef SO_OTIMESTAMP + if (in6p->in6p_socket->so_options & SO_OTIMESTAMP) { + struct timeval50 tv50; + timeval_to_timeval50(&tv, &tv50); + *mp = sbcreatecontrol((void *) &tv50, sizeof(tv50), + SCM_OTIMESTAMP, SOL_SOCKET); + } else #endif - if (in6p->in6p_flags & IN6P_RECVDSTADDR) { - *mp = sbcreatecontrol((caddr_t) &ip6->ip6_dst, - sizeof(struct in6_addr), IPV6_RECVDSTADDR, - IPPROTO_IPV6); + *mp = sbcreatecontrol((void *) &tv, sizeof(tv), + SCM_TIMESTAMP, SOL_SOCKET); if (*mp) mp = &(*mp)->m_next; } -#ifdef noyet - /* options were tossed above */ - if (in6p->in6p_flags & IN6P_RECVOPTS) - /* broken */ - /* ip6_srcroute doesn't do what we want here, need to fix */ - if (in6p->in6p_flags & IPV6P_RECVRETOPTS) - /* broken */ -#endif + /* some OSes call this logic with IPv4 packet, for SO_TIMESTAMP */ + if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) + return; /* RFC 2292 sec. 5 */ - if (in6p->in6p_flags & IN6P_PKTINFO) { + if ((in6p->in6p_flags & IN6P_PKTINFO) != 0) { struct in6_pktinfo pi6; - bcopy(&ip6->ip6_dst, &pi6.ipi6_addr, sizeof(struct in6_addr)); - if (IN6_IS_SCOPE_LINKLOCAL(&pi6.ipi6_addr)) - pi6.ipi6_addr.s6_addr16[1] = 0; - pi6.ipi6_ifindex = (m && m->m_pkthdr.rcvif) - ? m->m_pkthdr.rcvif->if_index - : 0; - *mp = sbcreatecontrol((caddr_t) &pi6, - sizeof(struct in6_pktinfo), IPV6_PKTINFO, - IPPROTO_IPV6); + + memcpy(&pi6.ipi6_addr, &ip6->ip6_dst, sizeof(struct in6_addr)); + in6_clearscope(&pi6.ipi6_addr); /* XXX */ + pi6.ipi6_ifindex = m->m_pkthdr.rcvif ? + m->m_pkthdr.rcvif->if_index : 0; + *mp = sbcreatecontrol((void *) &pi6, + sizeof(struct in6_pktinfo), + IS2292(IPV6_2292PKTINFO, IPV6_PKTINFO), IPPROTO_IPV6); if (*mp) mp = &(*mp)->m_next; } + if (in6p->in6p_flags & IN6P_HOPLIMIT) { int hlim = ip6->ip6_hlim & 0xff; - *mp = sbcreatecontrol((caddr_t) &hlim, - sizeof(int), IPV6_HOPLIMIT, IPPROTO_IPV6); + + *mp = sbcreatecontrol((void *) &hlim, sizeof(int), + IS2292(IPV6_2292HOPLIMIT, IPV6_HOPLIMIT), IPPROTO_IPV6); + if (*mp) + mp = &(*mp)->m_next; + } + + if ((in6p->in6p_flags & IN6P_TCLASS) != 0) { + u_int32_t flowinfo; + int tclass; + + flowinfo = (u_int32_t)ntohl(ip6->ip6_flow & IPV6_FLOWINFO_MASK); + flowinfo >>= 20; + + tclass = flowinfo & 0xff; + *mp = sbcreatecontrol((void *)&tclass, sizeof(tclass), + IPV6_TCLASS, IPPROTO_IPV6); + if (*mp) mp = &(*mp)->m_next; } - /* IN6P_NEXTHOP - for outgoing packet only */ /* - * IPV6_HOPOPTS socket option. We require super-user privilege - * for the option, but it might be too strict, since there might - * be some hop-by-hop options which can be returned to normal user. - * See RFC 2292 section 6. + * IPV6_HOPOPTS socket option. Recall that we required super-user + * privilege for the option (see ip6_ctloutput), but it might be too + * strict, since there might be some hop-by-hop options which can be + * returned to normal user. + * See also RFC3542 section 8 (or RFC2292 section 6). */ - if ((in6p->in6p_flags & IN6P_HOPOPTS) && privileged) { + if ((in6p->in6p_flags & IN6P_HOPOPTS) != 0) { /* * Check if a hop-by-hop options header is contatined in the * received packet, and if so, store the options as ancillary @@ -1048,47 +1188,45 @@ ip6_savecontrol(in6p, mp, ip6, m) * just after the IPv6 header, which fact is assured through * the IPv6 input processing. */ - struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); - if (ip6->ip6_nxt == IPPROTO_HOPOPTS) { + struct ip6_hdr *xip6 = mtod(m, struct ip6_hdr *); + if (xip6->ip6_nxt == IPPROTO_HOPOPTS) { struct ip6_hbh *hbh; int hbhlen; + struct mbuf *ext; -#ifndef PULLDOWN_TEST - hbh = (struct ip6_hbh *)(ip6 + 1); - hbhlen = (hbh->ip6h_len + 1) << 3; -#else - IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, - sizeof(struct ip6_hdr), sizeof(struct ip6_hbh)); - if (hbh == NULL) { - ip6stat.ip6s_tooshort++; + ext = ip6_pullexthdr(m, sizeof(struct ip6_hdr), + xip6->ip6_nxt); + if (ext == NULL) { + IP6_STATINC(IP6_STAT_TOOSHORT); return; } + hbh = mtod(ext, struct ip6_hbh *); hbhlen = (hbh->ip6h_len + 1) << 3; - IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, - sizeof(struct ip6_hdr), hbhlen); - if (hbh == NULL) { - ip6stat.ip6s_tooshort++; + if (hbhlen != ext->m_len) { + m_freem(ext); + IP6_STATINC(IP6_STAT_TOOSHORT); return; } -#endif /* * XXX: We copy whole the header even if a jumbo * payload option is included, which option is to * be removed before returning in the RFC 2292. - * But it's too painful operation... + * Note: this constraint is removed in RFC3542. */ - *mp = sbcreatecontrol((caddr_t)hbh, hbhlen, - IPV6_HOPOPTS, IPPROTO_IPV6); + *mp = sbcreatecontrol((void *)hbh, hbhlen, + IS2292(IPV6_2292HOPOPTS, IPV6_HOPOPTS), + IPPROTO_IPV6); if (*mp) mp = &(*mp)->m_next; + m_freem(ext); } } /* IPV6_DSTOPTS and IPV6_RTHDR socket options */ if (in6p->in6p_flags & (IN6P_DSTOPTS | IN6P_RTHDR)) { - struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); - int nxt = ip6->ip6_nxt, off = sizeof(struct ip6_hdr);; + struct ip6_hdr *xip6 = mtod(m, struct ip6_hdr *); + int nxt = xip6->ip6_nxt, off = sizeof(struct ip6_hdr); /* * Search for destination options headers or routing @@ -1097,94 +1235,182 @@ ip6_savecontrol(in6p, mp, ip6, m) * Note that the order of the headers remains in * the chain of ancillary data. */ - while(1) { /* is explicit loop prevention necessary? */ - struct ip6_ext *ip6e; + for (;;) { /* is explicit loop prevention necessary? */ + struct ip6_ext *ip6e = NULL; int elen; + struct mbuf *ext = NULL; -#ifndef PULLDOWN_TEST - ip6e = (struct ip6_ext *)(mtod(m, caddr_t) + off); - if (nxt == IPPROTO_AH) - elen = (ip6e->ip6e_len + 2) << 2; - else - elen = (ip6e->ip6e_len + 1) << 3; -#else - IP6_EXTHDR_GET(ip6e, struct ip6_ext *, m, off, - sizeof(struct ip6_ext)); - if (ip6e == NULL) { - ip6stat.ip6s_tooshort++; + /* + * if it is not an extension header, don't try to + * pull it from the chain. + */ + switch (nxt) { + case IPPROTO_DSTOPTS: + case IPPROTO_ROUTING: + case IPPROTO_HOPOPTS: + case IPPROTO_AH: /* is it possible? */ + break; + default: + goto loopend; + } + + ext = ip6_pullexthdr(m, off, nxt); + if (ext == NULL) { + IP6_STATINC(IP6_STAT_TOOSHORT); return; } + ip6e = mtod(ext, struct ip6_ext *); if (nxt == IPPROTO_AH) elen = (ip6e->ip6e_len + 2) << 2; else elen = (ip6e->ip6e_len + 1) << 3; - IP6_EXTHDR_GET(ip6e, struct ip6_ext *, m, off, elen); - if (ip6e == NULL) { - ip6stat.ip6s_tooshort++; + if (elen != ext->m_len) { + m_freem(ext); + IP6_STATINC(IP6_STAT_TOOSHORT); return; } -#endif + KASSERT(IP6_HDR_ALIGNED_P(ip6e)); + + switch (nxt) { + case IPPROTO_DSTOPTS: + if (!(in6p->in6p_flags & IN6P_DSTOPTS)) + break; + + *mp = sbcreatecontrol((void *)ip6e, elen, + IS2292(IPV6_2292DSTOPTS, IPV6_DSTOPTS), + IPPROTO_IPV6); + if (*mp) + mp = &(*mp)->m_next; + break; + + case IPPROTO_ROUTING: + if (!(in6p->in6p_flags & IN6P_RTHDR)) + break; + + *mp = sbcreatecontrol((void *)ip6e, elen, + IS2292(IPV6_2292RTHDR, IPV6_RTHDR), + IPPROTO_IPV6); + if (*mp) + mp = &(*mp)->m_next; + break; + + case IPPROTO_HOPOPTS: + case IPPROTO_AH: /* is it possible? */ + break; + + default: + /* + * other cases have been filtered in the above. + * none will visit this case. here we supply + * the code just in case (nxt overwritten or + * other cases). + */ + m_freem(ext); + goto loopend; - switch(nxt) { - case IPPROTO_DSTOPTS: - if (!in6p->in6p_flags & IN6P_DSTOPTS) - break; - - /* - * We also require super-user privilege for - * the option. - * See the comments on IN6_HOPOPTS. - */ - if (!privileged) - break; - - *mp = sbcreatecontrol((caddr_t)ip6e, elen, - IPV6_DSTOPTS, - IPPROTO_IPV6); - if (*mp) - mp = &(*mp)->m_next; - break; - - case IPPROTO_ROUTING: - if (!in6p->in6p_flags & IN6P_RTHDR) - break; - - *mp = sbcreatecontrol((caddr_t)ip6e, elen, - IPV6_RTHDR, - IPPROTO_IPV6); - if (*mp) - mp = &(*mp)->m_next; - break; - - case IPPROTO_UDP: - case IPPROTO_TCP: - case IPPROTO_ICMPV6: - default: - /* - * stop search if we encounter an upper - * layer protocol headers. - */ - goto loopend; - - case IPPROTO_HOPOPTS: - case IPPROTO_AH: /* is it possible? */ - break; } /* proceed with the next header. */ off += elen; nxt = ip6e->ip6e_nxt; + ip6e = NULL; + m_freem(ext); + ext = NULL; } loopend: + ; + } +} +#undef IS2292 + + +void +ip6_notify_pmtu(struct in6pcb *in6p, const struct sockaddr_in6 *dst, + uint32_t *mtu) +{ + struct socket *so; + struct mbuf *m_mtu; + struct ip6_mtuinfo mtuctl; + + so = in6p->in6p_socket; + + if (mtu == NULL) + return; + +#ifdef DIAGNOSTIC + if (so == NULL) /* I believe this is impossible */ + panic("ip6_notify_pmtu: socket is NULL"); +#endif + + memset(&mtuctl, 0, sizeof(mtuctl)); /* zero-clear for safety */ + mtuctl.ip6m_mtu = *mtu; + mtuctl.ip6m_addr = *dst; + if (sa6_recoverscope(&mtuctl.ip6m_addr)) + return; + + if ((m_mtu = sbcreatecontrol((void *)&mtuctl, sizeof(mtuctl), + IPV6_PATHMTU, IPPROTO_IPV6)) == NULL) + return; + + if (sbappendaddr(&so->so_rcv, (const struct sockaddr *)dst, NULL, m_mtu) + == 0) { + m_freem(m_mtu); + /* XXX: should count statistics */ + } else + sorwakeup(so); + + return; +} + +/* + * pull single extension header from mbuf chain. returns single mbuf that + * contains the result, or NULL on error. + */ +static struct mbuf * +ip6_pullexthdr(struct mbuf *m, size_t off, int nxt) +{ + struct ip6_ext ip6e; + size_t elen; + struct mbuf *n; + +#ifdef DIAGNOSTIC + switch (nxt) { + case IPPROTO_DSTOPTS: + case IPPROTO_ROUTING: + case IPPROTO_HOPOPTS: + case IPPROTO_AH: /* is it possible? */ + break; + default: + printf("ip6_pullexthdr: invalid nxt=%d\n", nxt); } - if ((in6p->in6p_flags & IN6P_HOPOPTS) && privileged) { - /* to be done */ +#endif + + m_copydata(m, off, sizeof(ip6e), (void *)&ip6e); + if (nxt == IPPROTO_AH) + elen = (ip6e.ip6e_len + 2) << 2; + else + elen = (ip6e.ip6e_len + 1) << 3; + + MGET(n, M_DONTWAIT, MT_DATA); + if (n && elen >= MLEN) { + MCLGET(n, M_DONTWAIT); + if ((n->m_flags & M_EXT) == 0) { + m_free(n); + n = NULL; + } } - if ((in6p->in6p_flags & IN6P_DSTOPTS) && privileged) { - /* to be done */ + if (!n) + return NULL; + + n->m_len = 0; + if (elen >= M_TRAILINGSPACE(n)) { + m_free(n); + return NULL; } - /* IN6P_RTHDR - to be done */ + m_copydata(m, off, elen, mtod(n, void *)); + n->m_len = elen; + return n; } /* @@ -1199,15 +1425,13 @@ ip6_savecontrol(in6p, mp, ip6, m) * carefully. Moreover, it will not be used in the near future when * we develop `neater' mechanism to process extension headers. */ -char * -ip6_get_prevhdr(m, off) - struct mbuf *m; - int off; +u_int8_t * +ip6_get_prevhdr(struct mbuf *m, int off) { struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); if (off == sizeof(struct ip6_hdr)) - return(&ip6->ip6_nxt); + return (&ip6->ip6_nxt); else { int len, nxt; struct ip6_ext *ip6e = NULL; @@ -1215,9 +1439,9 @@ ip6_get_prevhdr(m, off) nxt = ip6->ip6_nxt; len = sizeof(struct ip6_hdr); while (len < off) { - ip6e = (struct ip6_ext *)(mtod(m, caddr_t) + len); + ip6e = (struct ip6_ext *)(mtod(m, char *) + len); - switch(nxt) { + switch (nxt) { case IPPROTO_FRAGMENT: len += sizeof(struct ip6_frag); break; @@ -1231,7 +1455,7 @@ ip6_get_prevhdr(m, off) nxt = ip6e->ip6e_nxt; } if (ip6e) - return(&ip6e->ip6e_nxt); + return (&ip6e->ip6e_nxt); else return NULL; } @@ -1241,11 +1465,7 @@ ip6_get_prevhdr(m, off) * get next header offset. m will be retained. */ int -ip6_nexthdr(m, off, proto, nxtp) - struct mbuf *m; - int off; - int proto; - int *nxtp; +ip6_nexthdr(struct mbuf *m, int off, int proto, int *nxtp) { struct ip6_hdr ip6; struct ip6_ext ip6e; @@ -1259,9 +1479,12 @@ ip6_nexthdr(m, off, proto, nxtp) switch (proto) { case IPPROTO_IPV6: + /* do not chase beyond intermediate IPv6 headers */ + if (off != 0) + return -1; if (m->m_pkthdr.len < off + sizeof(ip6)) return -1; - m_copydata(m, off, sizeof(ip6), (caddr_t)&ip6); + m_copydata(m, off, sizeof(ip6), (void *)&ip6); if (nxtp) *nxtp = ip6.ip6_nxt; off += sizeof(ip6); @@ -1274,8 +1497,8 @@ ip6_nexthdr(m, off, proto, nxtp) */ if (m->m_pkthdr.len < off + sizeof(fh)) return -1; - m_copydata(m, off, sizeof(fh), (caddr_t)&fh); - if ((ntohs(fh.ip6f_offlg) & IP6F_OFF_MASK) != 0) + m_copydata(m, off, sizeof(fh), (void *)&fh); + if ((fh.ip6f_offlg & IP6F_OFF_MASK) != 0) return -1; if (nxtp) *nxtp = fh.ip6f_nxt; @@ -1285,10 +1508,12 @@ ip6_nexthdr(m, off, proto, nxtp) case IPPROTO_AH: if (m->m_pkthdr.len < off + sizeof(ip6e)) return -1; - m_copydata(m, off, sizeof(ip6e), (caddr_t)&ip6e); + m_copydata(m, off, sizeof(ip6e), (void *)&ip6e); if (nxtp) *nxtp = ip6e.ip6e_nxt; off += (ip6e.ip6e_len + 2) << 2; + if (m->m_pkthdr.len < off) + return -1; return off; case IPPROTO_HOPOPTS: @@ -1296,10 +1521,12 @@ ip6_nexthdr(m, off, proto, nxtp) case IPPROTO_DSTOPTS: if (m->m_pkthdr.len < off + sizeof(ip6e)) return -1; - m_copydata(m, off, sizeof(ip6e), (caddr_t)&ip6e); + m_copydata(m, off, sizeof(ip6e), (void *)&ip6e); if (nxtp) *nxtp = ip6e.ip6e_nxt; off += (ip6e.ip6e_len + 1) << 3; + if (m->m_pkthdr.len < off) + return -1; return off; case IPPROTO_NONE: @@ -1311,19 +1538,13 @@ ip6_nexthdr(m, off, proto, nxtp) default: return -1; } - - return -1; } /* * get offset for the last header in the chain. m will be kept untainted. */ int -ip6_lasthdr(m, off, proto, nxtp) - struct mbuf *m; - int off; - int proto; - int *nxtp; +ip6_lasthdr(struct mbuf *m, int off, int proto, int *nxtp) { int newoff; int nxt; @@ -1332,7 +1553,7 @@ ip6_lasthdr(m, off, proto, nxtp) nxt = -1; nxtp = &nxt; } - while (1) { + for (;;) { newoff = ip6_nexthdr(m, off, proto, nxtp); if (newoff < 0) return off; @@ -1346,11 +1567,107 @@ ip6_lasthdr(m, off, proto, nxtp) } } +struct m_tag * +ip6_addaux(struct mbuf *m) +{ + struct m_tag *mtag; + + mtag = m_tag_find(m, PACKET_TAG_INET6, NULL); + if (!mtag) { + mtag = m_tag_get(PACKET_TAG_INET6, sizeof(struct ip6aux), + M_NOWAIT); + if (mtag) { + m_tag_prepend(m, mtag); + memset(mtag + 1, 0, sizeof(struct ip6aux)); + } + } + return mtag; +} + +struct m_tag * +ip6_findaux(struct mbuf *m) +{ + struct m_tag *mtag; + + mtag = m_tag_find(m, PACKET_TAG_INET6, NULL); + return mtag; +} + +void +ip6_delaux(struct mbuf *m) +{ + struct m_tag *mtag; + + mtag = m_tag_find(m, PACKET_TAG_INET6, NULL); + if (mtag) + m_tag_delete(m, mtag); +} + +#ifdef GATEWAY +/* + * sysctl helper routine for net.inet.ip6.maxflows. Since + * we could reduce this value, call ip6flow_reap(); + */ +static int +sysctl_net_inet6_ip6_maxflows(SYSCTLFN_ARGS) +{ + int error; + + error = sysctl_lookup(SYSCTLFN_CALL(rnode)); + if (error || newp == NULL) + return (error); + + mutex_enter(softnet_lock); + KERNEL_LOCK(1, NULL); + + ip6flow_reap(0); + + KERNEL_UNLOCK_ONE(NULL); + mutex_exit(softnet_lock); + + return (0); +} + +static int +sysctl_net_inet6_ip6_hashsize(SYSCTLFN_ARGS) +{ + int error, tmp; + struct sysctlnode node; + + node = *rnode; + tmp = ip6_hashsize; + node.sysctl_data = &tmp; + error = sysctl_lookup(SYSCTLFN_CALL(&node)); + if (error || newp == NULL) + return (error); + + if ((tmp & (tmp - 1)) == 0 && tmp != 0) { + /* + * Can only fail due to malloc() + */ + mutex_enter(softnet_lock); + KERNEL_LOCK(1, NULL); + + error = ip6flow_invalidate_all(tmp); + + KERNEL_UNLOCK_ONE(NULL); + mutex_exit(softnet_lock); + } else { + /* + * EINVAL if not a power of 2 + */ + error = EINVAL; + } + + return error; +} +#endif /* GATEWAY */ + /* * System control for IP6 */ -u_char inet6ctlerrmap[PRC_NCMDS] = { +const u_char inet6ctlerrmap[PRC_NCMDS] = { 0, 0, 0, 0, 0, EMSGSIZE, EHOSTDOWN, EHOSTUNREACH, EHOSTUNREACH, EHOSTUNREACH, ECONNREFUSED, ECONNREFUSED, @@ -1359,125 +1676,305 @@ u_char inet6ctlerrmap[PRC_NCMDS] = { ENOPROTOOPT }; -#include -#include +static int +sysctl_net_inet6_ip6_stats(SYSCTLFN_ARGS) +{ -int -ip6_sysctl(name, namelen, oldp, oldlenp, newp, newlen) - int *name; - u_int namelen; - void *oldp; - size_t *oldlenp; - void *newp; - size_t newlen; -{ - int old, error; - - /* All sysctl names at this level are terminal. */ - if (namelen != 1) - return ENOTDIR; - - switch (name[0]) { - - case IPV6CTL_FORWARDING: - return sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_forwarding); - case IPV6CTL_SENDREDIRECTS: - return sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_sendredirects); - case IPV6CTL_DEFHLIM: - return sysctl_int(oldp, oldlenp, newp, newlen, &ip6_defhlim); - case IPV6CTL_MAXFRAGPACKETS: - return sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_maxfragpackets); - case IPV6CTL_ACCEPT_RTADV: - return sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_accept_rtadv); - case IPV6CTL_KEEPFAITH: - return sysctl_int(oldp, oldlenp, newp, newlen, &ip6_keepfaith); - case IPV6CTL_LOG_INTERVAL: - return sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_log_interval); - case IPV6CTL_HDRNESTLIMIT: - return sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_hdrnestlimit); - case IPV6CTL_DAD_COUNT: - return sysctl_int(oldp, oldlenp, newp, newlen, &ip6_dad_count); - case IPV6CTL_AUTO_FLOWLABEL: - return sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_auto_flowlabel); - case IPV6CTL_DEFMCASTHLIM: - return sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_defmcasthlim); - case IPV6CTL_GIF_HLIM: - return sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_gif_hlim); - case IPV6CTL_KAME_VERSION: - return sysctl_rdstring(oldp, oldlenp, newp, __KAME_VERSION); - case IPV6CTL_USE_DEPRECATED: - return sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_use_deprecated); - case IPV6CTL_RR_PRUNE: - return sysctl_int(oldp, oldlenp, newp, newlen, &ip6_rr_prune); -#ifndef INET6_BINDV6ONLY - case IPV6CTL_BINDV6ONLY: - return sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_bindv6only); -#endif - case IPV6CTL_ANONPORTMIN: - old = ip6_anonportmin; - error = sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_anonportmin); - if (ip6_anonportmin >= ip6_anonportmax || ip6_anonportmin < 0 || - ip6_anonportmin > 65535 -#ifndef IPNOPRIVPORTS - || ip6_anonportmin < IPV6PORT_RESERVED + return (NETSTAT_SYSCTL(ip6stat_percpu, IP6_NSTATS)); +} + +SYSCTL_SETUP(sysctl_net_inet6_ip6_setup, "sysctl net.inet6.ip6 subtree setup") +{ +#ifdef RFC2292 +#define IS2292(x, y) ((in6p->in6p_flags & IN6P_RFC2292) ? (x) : (y)) +#else +#define IS2292(x, y) (y) #endif - ) { - ip6_anonportmin = old; - return (EINVAL); - } - return (error); - case IPV6CTL_ANONPORTMAX: - old = ip6_anonportmax; - error = sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_anonportmax); - if (ip6_anonportmin >= ip6_anonportmax || ip6_anonportmax < 0 || - ip6_anonportmax > 65535 -#ifndef IPNOPRIVPORTS - || ip6_anonportmax < IPV6PORT_RESERVED + + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT, + CTLTYPE_NODE, "net", NULL, + NULL, 0, NULL, 0, + CTL_NET, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT, + CTLTYPE_NODE, "inet6", + SYSCTL_DESCR("PF_INET6 related settings"), + NULL, 0, NULL, 0, + CTL_NET, PF_INET6, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT, + CTLTYPE_NODE, "ip6", + SYSCTL_DESCR("IPv6 related settings"), + NULL, 0, NULL, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, CTL_EOL); + + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "forwarding", + SYSCTL_DESCR("Enable forwarding of INET6 datagrams"), + NULL, 0, &ip6_forwarding, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_FORWARDING, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "redirect", + SYSCTL_DESCR("Enable sending of ICMPv6 redirect messages"), + NULL, 0, &ip6_sendredirects, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_SENDREDIRECTS, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "hlim", + SYSCTL_DESCR("Hop limit for an INET6 datagram"), + NULL, 0, &ip6_defhlim, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_DEFHLIM, CTL_EOL); +#ifdef notyet + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "mtu", NULL, + NULL, 0, &, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_DEFMTU, CTL_EOL); +#endif +#ifdef __no_idea__ + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "forwsrcrt", NULL, + NULL, 0, &?, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_FORWSRCRT, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_STRUCT, "mrtstats", NULL, + NULL, 0, &?, sizeof(?), + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_MRTSTATS, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_?, "mrtproto", NULL, + NULL, 0, &?, sizeof(?), + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_MRTPROTO, CTL_EOL); +#endif + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "maxfragpackets", + SYSCTL_DESCR("Maximum number of fragments to buffer " + "for reassembly"), + NULL, 0, &ip6_maxfragpackets, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_MAXFRAGPACKETS, CTL_EOL); +#ifdef __no_idea__ + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "sourcecheck", NULL, + NULL, 0, &?, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_SOURCECHECK, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "sourcecheck_logint", NULL, + NULL, 0, &?, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_SOURCECHECK_LOGINT, CTL_EOL); +#endif + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "accept_rtadv", + SYSCTL_DESCR("Accept router advertisements"), + NULL, 0, &ip6_accept_rtadv, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_ACCEPT_RTADV, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "keepfaith", + SYSCTL_DESCR("Activate faith interface"), + NULL, 0, &ip6_keepfaith, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_KEEPFAITH, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "log_interval", + SYSCTL_DESCR("Minumum interval between logging " + "unroutable packets"), + NULL, 0, &ip6_log_interval, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_LOG_INTERVAL, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "hdrnestlimit", + SYSCTL_DESCR("Maximum number of nested IPv6 headers"), + NULL, 0, &ip6_hdrnestlimit, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_HDRNESTLIMIT, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "dad_count", + SYSCTL_DESCR("Number of Duplicate Address Detection " + "probes to send"), + NULL, 0, &ip6_dad_count, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_DAD_COUNT, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "auto_flowlabel", + SYSCTL_DESCR("Assign random IPv6 flow labels"), + NULL, 0, &ip6_auto_flowlabel, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_AUTO_FLOWLABEL, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "defmcasthlim", + SYSCTL_DESCR("Default multicast hop limit"), + NULL, 0, &ip6_defmcasthlim, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_DEFMCASTHLIM, CTL_EOL); +#if NGIF > 0 + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "gifhlim", + SYSCTL_DESCR("Default hop limit for a gif tunnel datagram"), + NULL, 0, &ip6_gif_hlim, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_GIF_HLIM, CTL_EOL); +#endif /* NGIF */ + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT, + CTLTYPE_STRING, "kame_version", + SYSCTL_DESCR("KAME Version"), + NULL, 0, __UNCONST(__KAME_VERSION), 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_KAME_VERSION, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "use_deprecated", + SYSCTL_DESCR("Allow use of deprecated addresses as " + "source addresses"), + NULL, 0, &ip6_use_deprecated, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_USE_DEPRECATED, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "rr_prune", NULL, + NULL, 0, &ip6_rr_prune, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_RR_PRUNE, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT +#ifndef INET6_BINDV6ONLY + |CTLFLAG_READWRITE, #endif - ) { - ip6_anonportmax = old; - return (EINVAL); - } - return (error); + CTLTYPE_INT, "v6only", + SYSCTL_DESCR("Disallow PF_INET6 sockets from connecting " + "to PF_INET sockets"), + NULL, 0, &ip6_v6only, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_V6ONLY, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "anonportmin", + SYSCTL_DESCR("Lowest ephemeral port number to assign"), + sysctl_net_inet_ip_ports, 0, &ip6_anonportmin, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_ANONPORTMIN, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "anonportmax", + SYSCTL_DESCR("Highest ephemeral port number to assign"), + sysctl_net_inet_ip_ports, 0, &ip6_anonportmax, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_ANONPORTMAX, CTL_EOL); #ifndef IPNOPRIVPORTS - case IPV6CTL_LOWPORTMIN: - old = ip6_lowportmin; - error = sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_lowportmin); - if (ip6_lowportmin >= ip6_lowportmax || - ip6_lowportmin > IPV6PORT_RESERVEDMAX || - ip6_lowportmin < IPV6PORT_RESERVEDMIN) { - ip6_lowportmin = old; - return (EINVAL); - } - return (error); - case IPV6CTL_LOWPORTMAX: - old = ip6_lowportmax; - error = sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_lowportmax); - if (ip6_lowportmin >= ip6_lowportmax || - ip6_lowportmax > IPV6PORT_RESERVEDMAX || - ip6_lowportmax < IPV6PORT_RESERVEDMIN) { - ip6_lowportmax = old; - return (EINVAL); - } - return (error); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "lowportmin", + SYSCTL_DESCR("Lowest privileged ephemeral port number " + "to assign"), + sysctl_net_inet_ip_ports, 0, &ip6_lowportmin, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_LOWPORTMIN, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "lowportmax", + SYSCTL_DESCR("Highest privileged ephemeral port number " + "to assign"), + sysctl_net_inet_ip_ports, 0, &ip6_lowportmax, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_LOWPORTMAX, CTL_EOL); +#endif /* IPNOPRIVPORTS */ + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "use_tempaddr", + SYSCTL_DESCR("Use temporary address"), + NULL, 0, &ip6_use_tempaddr, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + CTL_CREATE, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "temppltime", + SYSCTL_DESCR("preferred lifetime of a temporary address"), + NULL, 0, &ip6_temp_preferred_lifetime, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + CTL_CREATE, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "tempvltime", + SYSCTL_DESCR("valid lifetime of a temporary address"), + NULL, 0, &ip6_temp_valid_lifetime, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + CTL_CREATE, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "maxfrags", + SYSCTL_DESCR("Maximum fragments in reassembly queue"), + NULL, 0, &ip6_maxfrags, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_MAXFRAGS, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT, + CTLTYPE_STRUCT, "stats", + SYSCTL_DESCR("IPv6 statistics"), + sysctl_net_inet6_ip6_stats, 0, NULL, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_STATS, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "use_defaultzone", + SYSCTL_DESCR("Whether to use the default scope zones"), + NULL, 0, &ip6_use_defzone, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + IPV6CTL_USE_DEFAULTZONE, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "mcast_pmtu", + SYSCTL_DESCR("Enable pMTU discovery for multicast packet"), + NULL, 0, &ip6_mcast_pmtu, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + CTL_CREATE, CTL_EOL); +#ifdef GATEWAY + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "maxflows", + SYSCTL_DESCR("Number of flows for fast forwarding (IPv6)"), + sysctl_net_inet6_ip6_maxflows, 0, &ip6_maxflows, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + CTL_CREATE, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "hashsize", + SYSCTL_DESCR("Size of hash table for fast forwarding (IPv6)"), + sysctl_net_inet6_ip6_hashsize, 0, &ip6_hashsize, 0, + CTL_NET, PF_INET6, IPPROTO_IPV6, + CTL_CREATE, CTL_EOL); #endif - default: - return EOPNOTSUPP; - } - /* NOTREACHED */ +} + +void +ip6_statinc(u_int stat) +{ + + KASSERT(stat < IP6_NSTATS); + IP6_STATINC(stat); }