Index: sys/conf/files =================================================================== --- sys/conf/files +++ sys/conf/files @@ -4081,6 +4081,7 @@ netinet6/in6_rmx.c optional inet6 netinet6/in6_rss.c optional inet6 rss netinet6/in6_src.c optional inet6 +netinet6/ip6_fastfwd.c optional inet6 netinet6/ip6_forward.c optional inet6 netinet6/ip6_gre.c optional gre inet6 netinet6/ip6_id.c optional inet6 Index: sys/netinet6/in6_var.h =================================================================== --- sys/netinet6/in6_var.h +++ sys/netinet6/in6_var.h @@ -819,6 +819,7 @@ /* * Extended API for IPv6 FIB support. */ +struct mbuf *ip6_tryforward(struct mbuf *); void in6_rtredirect(struct sockaddr *, struct sockaddr *, struct sockaddr *, int, struct sockaddr *, u_int); int in6_rtrequest(int, struct sockaddr *, struct sockaddr *, Index: sys/netinet6/ip6_fastfwd.c =================================================================== --- /dev/null +++ sys/netinet6/ip6_fastfwd.c @@ -0,0 +1,283 @@ +/*- + * Copyright (c) 2014-2016 Andrey V. Elsukov + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES + * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. + * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, + * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT + * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF + * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include +__FBSDID("$FreeBSD$"); + +#include "opt_inet6.h" +#include "opt_ipstealth.h" + +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +static int +ip6_findroute(struct nhop6_basic *pnh, const struct sockaddr_in6 *dst, + struct mbuf *m) +{ + + if (fib6_lookup_nh_basic(M_GETFIB(m), &dst->sin6_addr, + dst->sin6_scope_id, 0, dst->sin6_flowinfo, pnh) != 0) { + IP6STAT_INC(ip6s_noroute); + IP6STAT_INC(ip6s_cantforward); + icmp6_error(m, ICMP6_DST_UNREACH, + ICMP6_DST_UNREACH_NOROUTE, 0); + return (EHOSTUNREACH); + } + if (pnh->nh_flags & NHF_BLACKHOLE) { + IP6STAT_INC(ip6s_cantforward); + m_freem(m); + return (EHOSTUNREACH); + } + + if (pnh->nh_flags & NHF_REJECT) { + IP6STAT_INC(ip6s_cantforward); + icmp6_error(m, ICMP6_DST_UNREACH, + ICMP6_DST_UNREACH_REJECT, 0); + return (EHOSTUNREACH); + } + return (0); +} + +struct mbuf* +ip6_tryforward(struct mbuf *m) +{ + struct sockaddr_in6 dst; + struct nhop6_basic nh; + struct m_tag *fwd_tag; + struct ip6_hdr *ip6; + struct ifnet *rcvif; + uint32_t plen; + int error; + + /* + * Fallback conditions to ip6_input for slow path processing. + */ + ip6 = mtod(m, struct ip6_hdr *); + if (ip6->ip6_nxt == IPPROTO_HOPOPTS || + IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) || + IN6_IS_ADDR_LINKLOCAL(&ip6->ip6_dst) || + IN6_IS_ADDR_LINKLOCAL(&ip6->ip6_src) || + IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src) || + in6_localip(&ip6->ip6_dst)) + return (m); + /* + * Check that the amount of data in the buffers + * is as at least much as the IPv6 header would have us expect. + * Trim mbufs if longer than we expect. + * Drop packet if shorter than we expect. + */ + rcvif = m->m_pkthdr.rcvif; + plen = ntohs(ip6->ip6_plen); + if (plen == 0) { + /* + * Jumbograms must have hop-by-hop header and go via + * slow path. + */ + IP6STAT_INC(ip6s_badoptions); + goto dropin; + } + if (m->m_pkthdr.len - sizeof(struct ip6_hdr) < plen) { + IP6STAT_INC(ip6s_tooshort); + in6_ifstat_inc(rcvif, ifs6_in_truncated); + goto dropin; + } + if (m->m_pkthdr.len > sizeof(struct ip6_hdr) + plen) { + if (m->m_len == m->m_pkthdr.len) { + m->m_len = sizeof(struct ip6_hdr) + plen; + m->m_pkthdr.len = sizeof(struct ip6_hdr) + plen; + } else + m_adj(m, sizeof(struct ip6_hdr) + plen - + m->m_pkthdr.len); + } + + /* + * Hop limit. + */ +#ifdef IPSTEALTH + if (!V_ip6stealth) +#endif + if (ip6->ip6_hlim <= IPV6_HLIMDEC) { + icmp6_error(m, ICMP6_TIME_EXCEEDED, + ICMP6_TIME_EXCEED_TRANSIT, 0); + m = NULL; + goto dropin; + } + + bzero(&dst, sizeof(dst)); + dst.sin6_family = AF_INET6; + dst.sin6_len = sizeof(dst); + dst.sin6_addr = ip6->ip6_dst; + + /* + * Incoming packet firewall processing. + */ + if (!PFIL_HOOKED(&V_inet6_pfil_hook)) + goto passin; + if (pfil_run_hooks(&V_inet6_pfil_hook, &m, rcvif, PFIL_IN, + NULL) != 0 || m == NULL) + goto dropin; + /* + * If packet filter sets the M_FASTFWD_OURS flag, this means + * that new destination or next hop is our local address. + * So, we can just go back to ip6_input. + * XXX: should we decrement ip6_hlim in such case? + * + * Also it can forward packet to another destination, e.g. + * M_IP6_NEXTHOP flag is set and fwd_tag is attached to mbuf. + */ + if (m->m_flags & M_FASTFWD_OURS) + return (m); + + ip6 = mtod(m, struct ip6_hdr *); + if ((m->m_flags & M_IP6_NEXTHOP) && + (fwd_tag = m_tag_find(m, PACKET_TAG_IPFORWARD, NULL)) != NULL) { + /* + * Now we will find route to forwarded by pfil destination. + */ + bcopy((fwd_tag + 1), &dst, sizeof(dst)); + m->m_flags &= ~M_IP6_NEXTHOP; + m_tag_delete(m, fwd_tag); + } else { + /* Update dst since pfil could change it */ + dst.sin6_addr = ip6->ip6_dst; + } +passin: + /* + * Find route to destination. + */ + if (ip6_findroute(&nh, &dst, m) != 0) { + m = NULL; + in6_ifstat_inc(rcvif, ifs6_in_noroute); + goto dropin; + } + /* + * We used slow path processing for packets with scoped addresses. + * So, scope checks aren't needed here. + */ + if (m->m_pkthdr.len > nh.nh_mtu) { + in6_ifstat_inc(nh.nh_ifp, ifs6_in_toobig); + icmp6_error(m, ICMP6_PACKET_TOO_BIG, 0, nh.nh_mtu); + m = NULL; + goto dropout; + } + + /* + * Outgoing packet firewall processing. + */ + if (!PFIL_HOOKED(&V_inet6_pfil_hook)) + goto passout; + if (pfil_run_hooks(&V_inet6_pfil_hook, &m, nh.nh_ifp, PFIL_OUT, + NULL) != 0 || m == NULL) + goto dropout; + /* + * If packet filter sets the M_FASTFWD_OURS flag, this means + * that new destination or next hop is our local address. + * So, we can just go back to ip6_input. + * + * Also it can forward packet to another destination, e.g. + * M_IP6_NEXTHOP flag is set and fwd_tag is attached to mbuf. + */ + if (m->m_flags & M_FASTFWD_OURS) { + /* + * XXX: we did one hop and should decrement hop limit. But + * now we are the destination and just don't pay attention. + */ + return (m); + } + /* + * Again. A packet filter could change the destination address. + */ + ip6 = mtod(m, struct ip6_hdr *); + if (m->m_flags & M_IP6_NEXTHOP) + fwd_tag = m_tag_find(m, PACKET_TAG_IPFORWARD, NULL); + else + fwd_tag = NULL; + + if (fwd_tag != NULL || + !IN6_ARE_ADDR_EQUAL(&dst.sin6_addr, &ip6->ip6_dst)) { + if (fwd_tag != NULL) { + bcopy((fwd_tag + 1), &dst, sizeof(dst)); + m->m_flags &= ~M_IP6_NEXTHOP; + m_tag_delete(m, fwd_tag); + } else + dst.sin6_addr = ip6->ip6_dst; + /* + * Redo route lookup with new destination address + */ + if (ip6_findroute(&nh, &dst, m) != 0) { + m = NULL; + goto dropout; + } + } +passout: +#ifdef IPSTEALTH + if (!V_ip6stealth) +#endif + { + ip6->ip6_hlim -= IPV6_HLIMDEC; + } + + error = (*nh.nh_ifp->if_output)(nh.nh_ifp, m, + (struct sockaddr *)&dst, NULL); + if (error != 0) { + in6_ifstat_inc(nh.nh_ifp, ifs6_out_discard); + IP6STAT_INC(ip6s_cantforward); + } else { + in6_ifstat_inc(nh.nh_ifp, ifs6_out_forward); + IP6STAT_INC(ip6s_forward); + } + return (NULL); +dropin: + in6_ifstat_inc(rcvif, ifs6_in_discard); + goto drop; +dropout: + in6_ifstat_inc(nh.nh_ifp, ifs6_out_discard); +drop: + if (m != NULL) + m_freem(m); + return (NULL); +} + Index: sys/netinet6/ip6_input.c =================================================================== --- sys/netinet6/ip6_input.c +++ sys/netinet6/ip6_input.c @@ -119,6 +119,7 @@ #include #ifdef IPSEC +#include #include #include #include @@ -548,12 +549,20 @@ struct in6_addr odst; struct ip6_hdr *ip6; struct in6_ifaddr *ia; + struct ifnet *rcvif; u_int32_t plen; u_int32_t rtalert = ~0; int off = sizeof(struct ip6_hdr), nest; int nxt, ours = 0; int srcrt = 0; + /* + * Drop the packet if IPv6 operation is disabled on the interface. + */ + rcvif = m->m_pkthdr.rcvif; + if ((ND_IFINFO(rcvif)->flags & ND6_IFF_IFDISABLED)) + goto bad; + #ifdef IPSEC /* * should the inner packet be considered authentic? @@ -566,18 +575,8 @@ #endif /* IPSEC */ - if (m->m_flags & M_FASTFWD_OURS) { - /* - * Firewall changed destination to local. - */ - m->m_flags &= ~M_FASTFWD_OURS; - ours = 1; - ip6 = mtod(m, struct ip6_hdr *); - goto hbhcheck; - } - /* - * mbuf statistics + * Update statistics. */ if (m->m_flags & M_EXT) { if (m->m_next) @@ -588,22 +587,27 @@ if (m->m_next) { if (m->m_flags & M_LOOP) { IP6STAT_INC(ip6s_m2m[V_loif->if_index]); - } else if (m->m_pkthdr.rcvif->if_index < IP6S_M2MMAX) - IP6STAT_INC( - ip6s_m2m[m->m_pkthdr.rcvif->if_index]); + } else if (rcvif->if_index < IP6S_M2MMAX) + IP6STAT_INC(ip6s_m2m[rcvif->if_index]); else IP6STAT_INC(ip6s_m2m[0]); } else IP6STAT_INC(ip6s_m1); } - /* drop the packet if IPv6 operation is disabled on the IF */ - if ((ND_IFINFO(m->m_pkthdr.rcvif)->flags & ND6_IFF_IFDISABLED)) - goto bad; - - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_receive); + in6_ifstat_inc(rcvif, ifs6_in_receive); IP6STAT_INC(ip6s_total); + if (m->m_flags & M_FASTFWD_OURS) { + /* + * Firewall changed destination to local. + */ + m->m_flags &= ~M_FASTFWD_OURS; + ours = 1; + ip6 = mtod(m, struct ip6_hdr *); + goto hbhcheck; + } + #ifndef PULLDOWN_TEST /* * L2 bridge code and some other code can return mbuf chain @@ -617,10 +621,8 @@ n = m_getcl(M_NOWAIT, MT_DATA, M_PKTHDR); else n = m_gethdr(M_NOWAIT, MT_DATA); - if (n == NULL) { - m_freem(m); - return; /* ENOBUFS */ - } + if (n == NULL) + goto bad; m_move_pkthdr(n, m); m_copydata(m, 0, n->m_pkthdr.len, mtod(n, caddr_t)); @@ -632,26 +634,22 @@ #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))) == NULL) { IP6STAT_INC(ip6s_toosmall); - in6_ifstat_inc(inifp, ifs6_in_hdrerr); - return; + in6_ifstat_inc(rcvif, ifs6_in_hdrerr); + goto bad; } } ip6 = mtod(m, struct ip6_hdr *); - if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) { IP6STAT_INC(ip6s_badvers); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr); + in6_ifstat_inc(rcvif, ifs6_in_hdrerr); goto bad; } IP6STAT_INC(ip6s_nxthist[ip6->ip6_nxt]); - - IP_PROBE(receive, NULL, NULL, ip6, m->m_pkthdr.rcvif, NULL, ip6); + IP_PROBE(receive, NULL, NULL, ip6, rcvif, NULL, ip6); /* * Check against address spoofing/corruption. @@ -662,7 +660,7 @@ * XXX: "badscope" is not very suitable for a multicast source. */ IP6STAT_INC(ip6s_badscope); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); + in6_ifstat_inc(rcvif, ifs6_in_addrerr); goto bad; } if (IN6_IS_ADDR_MC_INTFACELOCAL(&ip6->ip6_dst) && @@ -674,7 +672,7 @@ * as the outgoing/incoming interface. */ IP6STAT_INC(ip6s_badscope); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); + in6_ifstat_inc(rcvif, ifs6_in_addrerr); goto bad; } if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) && @@ -686,7 +684,7 @@ * a packet is received, it must be silently dropped. */ IP6STAT_INC(ip6s_badscope); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); + in6_ifstat_inc(rcvif, ifs6_in_addrerr); goto bad; } #ifdef ALTQ @@ -710,7 +708,7 @@ if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) || IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) { IP6STAT_INC(ip6s_badscope); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); + in6_ifstat_inc(rcvif, ifs6_in_addrerr); goto bad; } #if 0 @@ -724,16 +722,25 @@ if (IN6_IS_ADDR_V4COMPAT(&ip6->ip6_src) || IN6_IS_ADDR_V4COMPAT(&ip6->ip6_dst)) { IP6STAT_INC(ip6s_badscope); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); + in6_ifstat_inc(rcvif, ifs6_in_addrerr); goto bad; } #endif + /* Try to forward the packet, but if we fail continue */ #ifdef IPSEC + if (!key_havesp(IPSEC_DIR_INBOUND) && + !key_havesp(IPSEC_DIR_OUTBOUND) && V_ip6_forwarding != 0) + if (ip6_tryforward(m) == NULL) + return; /* * Bypass packet filtering for packets previously handled by IPsec. */ if (ip6_ipsec_filtertunnel(m)) goto passin; +#else + if (V_ip6_forwarding != 0) + if (ip6_tryforward(m) == NULL) + return; #endif /* IPSEC */ /* @@ -749,8 +756,7 @@ if (!PFIL_HOOKED(&V_inet6_pfil_hook)) goto passin; - if (pfil_run_hooks(&V_inet6_pfil_hook, &m, - m->m_pkthdr.rcvif, PFIL_IN, NULL)) + if (pfil_run_hooks(&V_inet6_pfil_hook, &m, rcvif, PFIL_IN, NULL)) return; if (m == NULL) /* consumed by filter */ return; @@ -788,8 +794,8 @@ IP6STAT_INC(ip6s_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)) { + if (in6_setscope(&ip6->ip6_src, rcvif, NULL) || + in6_setscope(&ip6->ip6_dst, rcvif, NULL)) { IP6STAT_INC(ip6s_badscope); goto bad; } @@ -799,7 +805,7 @@ */ if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) { ours = 1; - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_mcast); + in6_ifstat_inc(rcvif, ifs6_in_mcast); goto hbhcheck; } /* @@ -834,7 +840,7 @@ */ if (!V_ip6_forwarding) { IP6STAT_INC(ip6s_cantforward); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard); + in6_ifstat_inc(rcvif, ifs6_in_discard); goto bad; } @@ -866,7 +872,7 @@ */ if (m->m_pkthdr.len - sizeof(struct ip6_hdr) < plen) { IP6STAT_INC(ip6s_tooshort); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_truncated); + in6_ifstat_inc(rcvif, ifs6_in_truncated); goto bad; } if (m->m_pkthdr.len > sizeof(struct ip6_hdr) + plen) { @@ -894,9 +900,9 @@ * unnecessarily calling into ip6_mforward(). */ if (ip6_mforward && - ip6_mforward(ip6, m->m_pkthdr.rcvif, m)) { + ip6_mforward(ip6, rcvif, m)) { IP6STAT_INC(ip6s_cantforward); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_discard); + in6_ifstat_inc(rcvif, ifs6_in_discard); goto bad; } } else if (!ours) { @@ -918,7 +924,7 @@ if (IN6_IS_ADDR_V4MAPPED(&ip6->ip6_src) || IN6_IS_ADDR_V4MAPPED(&ip6->ip6_dst)) { IP6STAT_INC(ip6s_badscope); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_addrerr); + in6_ifstat_inc(rcvif, ifs6_in_addrerr); goto bad; } @@ -926,7 +932,7 @@ * Tell launch routine the next header */ IP6STAT_INC(ip6s_delivered); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_deliver); + in6_ifstat_inc(rcvif, ifs6_in_deliver); nest = 0; while (nxt != IPPROTO_DONE) { @@ -941,7 +947,7 @@ */ if (m->m_pkthdr.len < off) { IP6STAT_INC(ip6s_tooshort); - in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_truncated); + in6_ifstat_inc(rcvif, ifs6_in_truncated); goto bad; } @@ -959,7 +965,9 @@ } return; bad: - m_freem(m); + in6_ifstat_inc(rcvif, ifs6_in_discard); + if (m != NULL) + m_freem(m); } /*