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 1.65 retrieving revision 1.90 diff -u -p -r1.65 -r1.90 --- src/sys/netinet6/ip6_input.c 2003/08/07 16:33:26 1.65 +++ src/sys/netinet6/ip6_input.c 2006/11/16 01:33:45 1.90 @@ -1,4 +1,4 @@ -/* $NetBSD: ip6_input.c,v 1.65 2003/08/07 16:33:26 agc Exp $ */ +/* $NetBSD: ip6_input.c,v 1.90 2006/11/16 01:33:45 christos Exp $ */ /* $KAME: ip6_input.c,v 1.188 2001/03/29 05:34:31 itojun Exp $ */ /* @@ -62,9 +62,10 @@ */ #include -__KERNEL_RCSID(0, "$NetBSD: ip6_input.c,v 1.65 2003/08/07 16:33:26 agc Exp $"); +__KERNEL_RCSID(0, "$NetBSD: ip6_input.c,v 1.90 2006/11/16 01:33:45 christos Exp $"); #include "opt_inet.h" +#include "opt_inet6.h" #include "opt_ipsec.h" #include "opt_pfil_hooks.h" @@ -103,6 +104,7 @@ __KERNEL_RCSID(0, "$NetBSD: ip6_input.c, #include #include #include +#include #include #include @@ -112,11 +114,8 @@ __KERNEL_RCSID(0, "$NetBSD: ip6_input.c, #include -/* we need it for NLOOP. */ -#include "loop.h" #include "faith.h" #include "gif.h" -#include "bpfilter.h" #if NGIF > 0 #include @@ -131,7 +130,8 @@ static int ip6qmaxlen = IFQ_MAXLEN; struct in6_ifaddr *in6_ifaddr; struct ifqueue ip6intrq; -extern struct ifnet loif[NLOOP]; +extern struct callout in6_tmpaddrtimer_ch; + int ip6_forward_srcrt; /* XXX */ int ip6_sourcecheck; /* XXX */ int ip6_sourcecheck_interval; /* XXX */ @@ -143,6 +143,7 @@ struct pfil_head inet6_pfil_hook; struct ip6stat ip6stat; static void ip6_init2 __P((void *)); +static struct m_tag *ip6_setdstifaddr __P((struct mbuf *, struct in6_ifaddr *)); static int ip6_hopopts_input __P((u_int32_t *, u_int32_t *, struct mbuf **, int *)); static struct mbuf *ip6_pullexthdr __P((struct mbuf *, size_t, int)); @@ -154,23 +155,25 @@ static struct mbuf *ip6_pullexthdr __P(( void ip6_init() { - struct ip6protosw *pr; + 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(); - ip6_flow_seq = arc4random(); + ip6_desync_factor = arc4random() % MAX_TEMP_DESYNC_FACTOR; ip6_init2((void *)0); @@ -186,13 +189,19 @@ ip6_init() } static void -ip6_init2(dummy) - void *dummy; +ip6_init2(void *dummy) { /* nd6_timer_init */ callout_init(&nd6_timer_ch); callout_reset(&nd6_timer_ch, hz, nd6_timer, NULL); + + /* timer for regeneranation of temporary addresses randomize ID */ + callout_init(&in6_tmpaddrtimer_ch); + callout_reset(&in6_tmpaddrtimer_ch, + (ip6_temp_preferred_lifetime - ip6_desync_factor - + ip6_temp_regen_advance) * hz, + in6_tmpaddrtimer, NULL); } /* @@ -210,6 +219,11 @@ ip6intr() splx(s); if (m == 0) return; + /* 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); + return; + } ip6_input(m); } } @@ -238,6 +252,11 @@ ip6_input(m) #endif /* + * make sure we don't have onion peering information into m_tag. + */ + ip6_delaux(m); + + /* * mbuf statistics */ if (m->m_flags & M_EXT) { @@ -249,7 +268,7 @@ ip6_input(m) #define M2MMAX (sizeof(ip6stat.ip6s_m2m)/sizeof(ip6stat.ip6s_m2m[0])) if (m->m_next) { if (m->m_flags & M_LOOP) { - ip6stat.ip6s_m2m[loif[0].if_index]++; /* XXX */ + ip6stat.ip6s_m2m[lo0ifp->if_index]++; /* XXX */ } else if (m->m_pkthdr.rcvif->if_index < M2MMAX) ip6stat.ip6s_m2m[m->m_pkthdr.rcvif->if_index]++; else @@ -265,7 +284,7 @@ ip6_input(m) /* * 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 + * 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, caddr_t)) == 0) { @@ -380,61 +399,28 @@ ip6_input(m) } #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; - } - } - - /* drop packets if interface ID portion is already filled */ - if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0) { - if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_src) && - ip6->ip6_src.s6_addr16[1]) { - ip6stat.ip6s_badscope++; - goto bad; - } - if (IN6_IS_SCOPE_LINKLOCAL(&ip6->ip6_dst) && - ip6->ip6_dst.s6_addr16[1]) { - ip6stat.ip6s_badscope++; - goto bad; - } + /* + * 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 (__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)) { + ip6stat.ip6s_badscope++; /* XXX */ + goto bad; } - - 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); - - /* - * We use rt->rt_ifp to determine if the address is ours or not. - * If rt_ifp is lo0, the address is ours. - * The problem here is, rt->rt_ifp for fe80::%lo0/64 is set to lo0, - * so any address under fe80::%lo0/64 will be mistakenly considered - * local. The special case is supplied to handle the case properly - * by actually looking at interface addresses - * (using in6ifa_ifpwithaddr). - */ - if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) != 0 && - IN6_IS_ADDR_LINKLOCAL(&ip6->ip6_dst)) { - if (!in6ifa_ifpwithaddr(m->m_pkthdr.rcvif, &ip6->ip6_dst)) { - icmp6_error(m, ICMP6_DST_UNREACH, - ICMP6_DST_UNREACH_ADDR, 0); - /* m is already freed */ - return; - } - - ours = 1; - deliverifp = m->m_pkthdr.rcvif; - goto hbhcheck; + if (in6_setscope(&ip6->ip6_src, m->m_pkthdr.rcvif, NULL) || + in6_setscope(&ip6->ip6_dst, m->m_pkthdr.rcvif, NULL)) { + ip6stat.ip6s_badscope++; + goto bad; } /* @@ -467,7 +453,7 @@ ip6_input(m) if (ip6_forward_rt.ro_rt != NULL && (ip6_forward_rt.ro_rt->rt_flags & RTF_UP) != 0 && IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst, - &((struct sockaddr_in6 *)(&ip6_forward_rt.ro_dst))->sin6_addr)) + &((struct sockaddr_in6 *)(&ip6_forward_rt.ro_dst))->sin6_addr)) ip6stat.ip6s_forward_cachehit++; else { struct sockaddr_in6 *dst6; @@ -542,8 +528,8 @@ ip6_input(m) */ #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 (ip6_forward_rt.ro_rt && ip6_forward_rt.ro_rt->rt_ifp && + ip6_forward_rt.ro_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 */ @@ -588,6 +574,27 @@ ip6_input(m) 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)) { + struct in6_ifaddr *ia6; + + ia6 = in6_ifawithifp(deliverifp, &ip6->ip6_dst); + if (ia6) { + if (!ip6_setdstifaddr(m, ia6)) { + /* + * 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. @@ -614,7 +621,7 @@ ip6_input(m) if (ip6->ip6_plen == 0 && plen == 0) { /* * Note that if a valid jumbo payload option is - * contained, ip6_hoptops_input() must set a valid + * contained, ip6_hopopts_input() must set a valid * (non-zero) payload length to the variable plen. */ ip6stat.ip6s_badoptions++; @@ -758,6 +765,35 @@ ip6_input(m) } /* + * set/grab in6_ifaddr correspond to IPv6 destination address. + */ +static struct m_tag * +ip6_setdstifaddr(m, ia6) + struct mbuf *m; + struct in6_ifaddr *ia6; +{ + struct m_tag *mtag; + + mtag = ip6_addaux(m); + if (mtag) + ((struct ip6aux *)(mtag + 1))->ip6a_dstia6 = ia6; + return mtag; /* NULL if failed to set */ +} + +struct in6_ifaddr * +ip6_getdstifaddr(m) + struct mbuf *m; +{ + struct m_tag *mtag; + + mtag = ip6_findaux(m); + if (mtag) + return ((struct ip6aux *)(mtag + 1))->ip6a_dstia6; + 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. */ @@ -771,7 +807,6 @@ ip6_hopopts_input(plenp, rtalertp, mp, o struct mbuf *m = *mp; int off = *offp, hbhlen; struct ip6_hbh *hbh; - u_int8_t *opt; /* validation of the length of the header */ IP6_EXTHDR_GET(hbh, struct ip6_hbh *, m, @@ -790,7 +825,6 @@ ip6_hopopts_input(plenp, rtalertp, mp, o 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) @@ -1000,12 +1034,11 @@ ip6_savecontrol(in6p, mp, ip6, m) struct ip6_hdr *ip6; struct mbuf *m; { - struct proc *p = curproc; /* XXX */ - int privileged; - - privileged = 0; - if (p && !suser(p->p_ucred, &p->p_acflag)) - privileged++; +#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) { @@ -1018,52 +1051,58 @@ ip6_savecontrol(in6p, mp, ip6, m) mp = &(*mp)->m_next; } #endif - if (in6p->in6p_flags & IN6P_RECVDSTADDR) { - *mp = sbcreatecontrol((caddr_t) &ip6->ip6_dst, - sizeof(struct in6_addr), IPV6_RECVDSTADDR, IPPROTO_IPV6); - 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) != 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; + in6_clearscope(&pi6.ipi6_addr); /* XXX */ + pi6.ipi6_ifindex = m->m_pkthdr.rcvif ? + m->m_pkthdr.rcvif->if_index : 0; *mp = sbcreatecontrol((caddr_t) &pi6, - sizeof(struct in6_pktinfo), IPV6_PKTINFO, IPPROTO_IPV6); + 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); + 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((caddr_t)&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) != 0 && 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 @@ -1071,14 +1110,14 @@ 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; ext = ip6_pullexthdr(m, sizeof(struct ip6_hdr), - ip6->ip6_nxt); + xip6->ip6_nxt); if (ext == NULL) { ip6stat.ip6s_tooshort++; return; @@ -1095,10 +1134,11 @@ ip6_savecontrol(in6p, mp, ip6, m) * 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); + IS2292(IPV6_2292HOPOPTS, IPV6_HOPOPTS), + IPPROTO_IPV6); if (*mp) mp = &(*mp)->m_next; m_freem(ext); @@ -1107,8 +1147,8 @@ ip6_savecontrol(in6p, mp, ip6, m) /* 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 @@ -1117,7 +1157,7 @@ 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? */ + for (;;) { /* is explicit loop prevention necessary? */ struct ip6_ext *ip6e = NULL; int elen; struct mbuf *ext = NULL; @@ -1158,16 +1198,9 @@ ip6_savecontrol(in6p, mp, ip6, m) 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); + IS2292(IPV6_2292DSTOPTS, IPV6_DSTOPTS), + IPPROTO_IPV6); if (*mp) mp = &(*mp)->m_next; break; @@ -1177,7 +1210,8 @@ ip6_savecontrol(in6p, mp, ip6, m) break; *mp = sbcreatecontrol((caddr_t)ip6e, elen, - IPV6_RTHDR, IPPROTO_IPV6); + IS2292(IPV6_2292RTHDR, IPV6_RTHDR), + IPPROTO_IPV6); if (*mp) mp = &(*mp)->m_next; break; @@ -1209,6 +1243,45 @@ ip6_savecontrol(in6p, mp, ip6, m) ; } } +#undef IS2292 + + +void +ip6_notify_pmtu(struct in6pcb *in6p, 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((caddr_t)&mtuctl, sizeof(mtuctl), + IPV6_PATHMTU, IPPROTO_IPV6)) == NULL) + return; + + if (sbappendaddr(&so->so_rcv, (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 @@ -1336,6 +1409,9 @@ 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); @@ -1352,7 +1428,7 @@ 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) + if ((fh.ip6f_offlg & IP6F_OFF_MASK) != 0) return -1; if (nxtp) *nxtp = fh.ip6f_nxt; @@ -1411,7 +1487,7 @@ 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; @@ -1425,6 +1501,45 @@ ip6_lasthdr(m, off, proto, nxtp) } } +struct m_tag * +ip6_addaux(m) + 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); + bzero(mtag + 1, sizeof(struct ip6aux)); + } + } + return mtag; +} + +struct m_tag * +ip6_findaux(m) + struct mbuf *m; +{ + struct m_tag *mtag; + + mtag = m_tag_find(m, PACKET_TAG_INET6, NULL); + return mtag; +} + +void +ip6_delaux(m) + struct mbuf *m; +{ + struct m_tag *mtag; + + mtag = m_tag_find(m, PACKET_TAG_INET6, NULL); + if (mtag) + m_tag_delete(m, mtag); +} + /* * System control for IP6 */ @@ -1438,127 +1553,274 @@ u_char inet6ctlerrmap[PRC_NCMDS] = { ENOPROTOOPT }; -int -ip6_sysctl(name, namelen, oldp, oldlenp, newp, newlen) - int *name; - u_int namelen; - void *oldp; - size_t *oldlenp; - void *newp; - size_t newlen; +SYSCTL_SETUP(sysctl_net_inet6_ip6_setup, "sysctl net.inet6.ip6 subtree setup") { - 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); -#if NGIF > 0 - case IPV6CTL_GIF_HLIM: - return sysctl_int(oldp, oldlenp, newp, newlen, - &ip6_gif_hlim); -#endif - 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); - case IPV6CTL_V6ONLY: -#ifdef INET6_BINDV6ONLY - return sysctl_rdint(oldp, oldlenp, newp, ip6_v6only); +#ifdef RFC2292 +#define IS2292(x, y) ((in6p->in6p_flags & IN6P_RFC2292) ? (x) : (y)) #else - return sysctl_int(oldp, oldlenp, newp, newlen, &ip6_v6only); -#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 -#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 +#define IS2292(x, y) (y) #endif - ) { - ip6_anonportmax = old; - return (EINVAL); - } - return (error); + + 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 + 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); -#endif - case IPV6CTL_MAXFRAGS: - return sysctl_int(oldp, oldlenp, newp, newlen, &ip6_maxfrags); - default: - return EOPNOTSUPP; - } - /* NOTREACHED */ + 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"), + NULL, 0, &ip6stat, sizeof(ip6stat), + 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); }