Index: head/sys/net/if_gre.c =================================================================== --- head/sys/net/if_gre.c (revision 316698) +++ head/sys/net/if_gre.c (revision 316699) @@ -1,1001 +1,995 @@ /*- * Copyright (c) 1998 The NetBSD Foundation, Inc. * Copyright (c) 2014 Andrey V. Elsukov * All rights reserved. * * This code is derived from software contributed to The NetBSD Foundation * by Heiko W.Rupp * * IPv6-over-GRE contributed by Gert Doering * * 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 NETBSD FOUNDATION, INC. AND CONTRIBUTORS * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE * POSSIBILITY OF SUCH DAMAGE. * * $NetBSD: if_gre.c,v 1.49 2003/12/11 00:22:29 itojun Exp $ */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef INET #include #include #include #include #endif #ifdef INET6 #include #include #include #include #endif #include #include #include #include #include -#define GREMTU 1500 +#define GREMTU 1476 static const char grename[] = "gre"; static MALLOC_DEFINE(M_GRE, grename, "Generic Routing Encapsulation"); static VNET_DEFINE(struct mtx, gre_mtx); #define V_gre_mtx VNET(gre_mtx) #define GRE_LIST_LOCK_INIT(x) mtx_init(&V_gre_mtx, "gre_mtx", NULL, \ MTX_DEF) #define GRE_LIST_LOCK_DESTROY(x) mtx_destroy(&V_gre_mtx) #define GRE_LIST_LOCK(x) mtx_lock(&V_gre_mtx) #define GRE_LIST_UNLOCK(x) mtx_unlock(&V_gre_mtx) static VNET_DEFINE(LIST_HEAD(, gre_softc), gre_softc_list); #define V_gre_softc_list VNET(gre_softc_list) static struct sx gre_ioctl_sx; SX_SYSINIT(gre_ioctl_sx, &gre_ioctl_sx, "gre_ioctl"); static int gre_clone_create(struct if_clone *, int, caddr_t); static void gre_clone_destroy(struct ifnet *); static VNET_DEFINE(struct if_clone *, gre_cloner); #define V_gre_cloner VNET(gre_cloner) static void gre_qflush(struct ifnet *); static int gre_transmit(struct ifnet *, struct mbuf *); static int gre_ioctl(struct ifnet *, u_long, caddr_t); static int gre_output(struct ifnet *, struct mbuf *, const struct sockaddr *, struct route *); static void gre_updatehdr(struct gre_softc *); static int gre_set_tunnel(struct ifnet *, struct sockaddr *, struct sockaddr *); static void gre_delete_tunnel(struct ifnet *); SYSCTL_DECL(_net_link); static SYSCTL_NODE(_net_link, IFT_TUNNEL, gre, CTLFLAG_RW, 0, "Generic Routing Encapsulation"); #ifndef MAX_GRE_NEST /* * This macro controls the default upper limitation on nesting of gre tunnels. * Since, setting a large value to this macro with a careless configuration * may introduce system crash, we don't allow any nestings by default. * If you need to configure nested gre tunnels, you can define this macro * in your kernel configuration file. However, if you do so, please be * careful to configure the tunnels so that it won't make a loop. */ #define MAX_GRE_NEST 1 #endif static VNET_DEFINE(int, max_gre_nesting) = MAX_GRE_NEST; #define V_max_gre_nesting VNET(max_gre_nesting) SYSCTL_INT(_net_link_gre, OID_AUTO, max_nesting, CTLFLAG_RW | CTLFLAG_VNET, &VNET_NAME(max_gre_nesting), 0, "Max nested tunnels"); static void vnet_gre_init(const void *unused __unused) { LIST_INIT(&V_gre_softc_list); GRE_LIST_LOCK_INIT(); V_gre_cloner = if_clone_simple(grename, gre_clone_create, gre_clone_destroy, 0); } VNET_SYSINIT(vnet_gre_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY, vnet_gre_init, NULL); static void vnet_gre_uninit(const void *unused __unused) { if_clone_detach(V_gre_cloner); GRE_LIST_LOCK_DESTROY(); } VNET_SYSUNINIT(vnet_gre_uninit, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY, vnet_gre_uninit, NULL); static int gre_clone_create(struct if_clone *ifc, int unit, caddr_t params) { struct gre_softc *sc; sc = malloc(sizeof(struct gre_softc), M_GRE, M_WAITOK | M_ZERO); sc->gre_fibnum = curthread->td_proc->p_fibnum; GRE2IFP(sc) = if_alloc(IFT_TUNNEL); GRE_LOCK_INIT(sc); GRE2IFP(sc)->if_softc = sc; if_initname(GRE2IFP(sc), grename, unit); - GRE2IFP(sc)->if_mtu = sc->gre_mtu = GREMTU; + GRE2IFP(sc)->if_mtu = GREMTU; GRE2IFP(sc)->if_flags = IFF_POINTOPOINT|IFF_MULTICAST; GRE2IFP(sc)->if_output = gre_output; GRE2IFP(sc)->if_ioctl = gre_ioctl; GRE2IFP(sc)->if_transmit = gre_transmit; GRE2IFP(sc)->if_qflush = gre_qflush; GRE2IFP(sc)->if_capabilities |= IFCAP_LINKSTATE; GRE2IFP(sc)->if_capenable |= IFCAP_LINKSTATE; if_attach(GRE2IFP(sc)); bpfattach(GRE2IFP(sc), DLT_NULL, sizeof(u_int32_t)); GRE_LIST_LOCK(); LIST_INSERT_HEAD(&V_gre_softc_list, sc, gre_list); GRE_LIST_UNLOCK(); return (0); } static void gre_clone_destroy(struct ifnet *ifp) { struct gre_softc *sc; sx_xlock(&gre_ioctl_sx); sc = ifp->if_softc; gre_delete_tunnel(ifp); GRE_LIST_LOCK(); LIST_REMOVE(sc, gre_list); GRE_LIST_UNLOCK(); bpfdetach(ifp); if_detach(ifp); ifp->if_softc = NULL; sx_xunlock(&gre_ioctl_sx); if_free(ifp); GRE_LOCK_DESTROY(sc); free(sc, M_GRE); } static int gre_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data) { GRE_RLOCK_TRACKER; struct ifreq *ifr = (struct ifreq *)data; struct sockaddr *src, *dst; struct gre_softc *sc; #ifdef INET struct sockaddr_in *sin = NULL; #endif #ifdef INET6 struct sockaddr_in6 *sin6 = NULL; #endif uint32_t opt; int error; switch (cmd) { case SIOCSIFMTU: /* XXX: */ if (ifr->ifr_mtu < 576) return (EINVAL); - break; + ifp->if_mtu = ifr->ifr_mtu; + return (0); case SIOCSIFADDR: ifp->if_flags |= IFF_UP; case SIOCSIFFLAGS: case SIOCADDMULTI: case SIOCDELMULTI: return (0); case GRESADDRS: case GRESADDRD: case GREGADDRS: case GREGADDRD: case GRESPROTO: case GREGPROTO: return (EOPNOTSUPP); } src = dst = NULL; sx_xlock(&gre_ioctl_sx); sc = ifp->if_softc; if (sc == NULL) { error = ENXIO; goto end; } error = 0; switch (cmd) { - case SIOCSIFMTU: - GRE_WLOCK(sc); - sc->gre_mtu = ifr->ifr_mtu; - gre_updatehdr(sc); - GRE_WUNLOCK(sc); - goto end; case SIOCSIFPHYADDR: #ifdef INET6 case SIOCSIFPHYADDR_IN6: #endif error = EINVAL; switch (cmd) { #ifdef INET case SIOCSIFPHYADDR: src = (struct sockaddr *) &(((struct in_aliasreq *)data)->ifra_addr); dst = (struct sockaddr *) &(((struct in_aliasreq *)data)->ifra_dstaddr); break; #endif #ifdef INET6 case SIOCSIFPHYADDR_IN6: src = (struct sockaddr *) &(((struct in6_aliasreq *)data)->ifra_addr); dst = (struct sockaddr *) &(((struct in6_aliasreq *)data)->ifra_dstaddr); break; #endif default: error = EAFNOSUPPORT; goto end; } /* sa_family must be equal */ if (src->sa_family != dst->sa_family || src->sa_len != dst->sa_len) goto end; /* validate sa_len */ switch (src->sa_family) { #ifdef INET case AF_INET: if (src->sa_len != sizeof(struct sockaddr_in)) goto end; break; #endif #ifdef INET6 case AF_INET6: if (src->sa_len != sizeof(struct sockaddr_in6)) goto end; break; #endif default: error = EAFNOSUPPORT; goto end; } /* check sa_family looks sane for the cmd */ error = EAFNOSUPPORT; switch (cmd) { #ifdef INET case SIOCSIFPHYADDR: if (src->sa_family == AF_INET) break; goto end; #endif #ifdef INET6 case SIOCSIFPHYADDR_IN6: if (src->sa_family == AF_INET6) break; goto end; #endif } error = EADDRNOTAVAIL; switch (src->sa_family) { #ifdef INET case AF_INET: if (satosin(src)->sin_addr.s_addr == INADDR_ANY || satosin(dst)->sin_addr.s_addr == INADDR_ANY) goto end; break; #endif #ifdef INET6 case AF_INET6: if (IN6_IS_ADDR_UNSPECIFIED(&satosin6(src)->sin6_addr) || IN6_IS_ADDR_UNSPECIFIED(&satosin6(dst)->sin6_addr)) goto end; /* * Check validity of the scope zone ID of the * addresses, and convert it into the kernel * internal form if necessary. */ error = sa6_embedscope(satosin6(src), 0); if (error != 0) goto end; error = sa6_embedscope(satosin6(dst), 0); if (error != 0) goto end; #endif } error = gre_set_tunnel(ifp, src, dst); break; case SIOCDIFPHYADDR: gre_delete_tunnel(ifp); break; case SIOCGIFPSRCADDR: case SIOCGIFPDSTADDR: #ifdef INET6 case SIOCGIFPSRCADDR_IN6: case SIOCGIFPDSTADDR_IN6: #endif if (sc->gre_family == 0) { error = EADDRNOTAVAIL; break; } GRE_RLOCK(sc); switch (cmd) { #ifdef INET case SIOCGIFPSRCADDR: case SIOCGIFPDSTADDR: if (sc->gre_family != AF_INET) { error = EADDRNOTAVAIL; break; } sin = (struct sockaddr_in *)&ifr->ifr_addr; memset(sin, 0, sizeof(*sin)); sin->sin_family = AF_INET; sin->sin_len = sizeof(*sin); break; #endif #ifdef INET6 case SIOCGIFPSRCADDR_IN6: case SIOCGIFPDSTADDR_IN6: if (sc->gre_family != AF_INET6) { error = EADDRNOTAVAIL; break; } sin6 = (struct sockaddr_in6 *) &(((struct in6_ifreq *)data)->ifr_addr); memset(sin6, 0, sizeof(*sin6)); sin6->sin6_family = AF_INET6; sin6->sin6_len = sizeof(*sin6); break; #endif } if (error == 0) { switch (cmd) { #ifdef INET case SIOCGIFPSRCADDR: sin->sin_addr = sc->gre_oip.ip_src; break; case SIOCGIFPDSTADDR: sin->sin_addr = sc->gre_oip.ip_dst; break; #endif #ifdef INET6 case SIOCGIFPSRCADDR_IN6: sin6->sin6_addr = sc->gre_oip6.ip6_src; break; case SIOCGIFPDSTADDR_IN6: sin6->sin6_addr = sc->gre_oip6.ip6_dst; break; #endif } } GRE_RUNLOCK(sc); if (error != 0) break; switch (cmd) { #ifdef INET case SIOCGIFPSRCADDR: case SIOCGIFPDSTADDR: error = prison_if(curthread->td_ucred, (struct sockaddr *)sin); if (error != 0) memset(sin, 0, sizeof(*sin)); break; #endif #ifdef INET6 case SIOCGIFPSRCADDR_IN6: case SIOCGIFPDSTADDR_IN6: error = prison_if(curthread->td_ucred, (struct sockaddr *)sin6); if (error == 0) error = sa6_recoverscope(sin6); if (error != 0) memset(sin6, 0, sizeof(*sin6)); #endif } break; case SIOCGTUNFIB: ifr->ifr_fib = sc->gre_fibnum; break; case SIOCSTUNFIB: if ((error = priv_check(curthread, PRIV_NET_GRE)) != 0) break; if (ifr->ifr_fib >= rt_numfibs) error = EINVAL; else sc->gre_fibnum = ifr->ifr_fib; break; case GRESKEY: if ((error = priv_check(curthread, PRIV_NET_GRE)) != 0) break; if ((error = copyin(ifr->ifr_data, &opt, sizeof(opt))) != 0) break; if (sc->gre_key != opt) { GRE_WLOCK(sc); sc->gre_key = opt; gre_updatehdr(sc); GRE_WUNLOCK(sc); } break; case GREGKEY: error = copyout(&sc->gre_key, ifr->ifr_data, sizeof(sc->gre_key)); break; case GRESOPTS: if ((error = priv_check(curthread, PRIV_NET_GRE)) != 0) break; if ((error = copyin(ifr->ifr_data, &opt, sizeof(opt))) != 0) break; if (opt & ~GRE_OPTMASK) error = EINVAL; else { if (sc->gre_options != opt) { GRE_WLOCK(sc); sc->gre_options = opt; gre_updatehdr(sc); GRE_WUNLOCK(sc); } } break; case GREGOPTS: error = copyout(&sc->gre_options, ifr->ifr_data, sizeof(sc->gre_options)); break; default: error = EINVAL; break; } end: sx_xunlock(&gre_ioctl_sx); return (error); } static void gre_updatehdr(struct gre_softc *sc) { struct grehdr *gh = NULL; uint32_t *opts; uint16_t flags; GRE_WLOCK_ASSERT(sc); switch (sc->gre_family) { #ifdef INET case AF_INET: sc->gre_hlen = sizeof(struct greip); sc->gre_oip.ip_v = IPPROTO_IPV4; sc->gre_oip.ip_hl = sizeof(struct ip) >> 2; sc->gre_oip.ip_p = IPPROTO_GRE; gh = &sc->gre_gihdr->gi_gre; break; #endif #ifdef INET6 case AF_INET6: sc->gre_hlen = sizeof(struct greip6); sc->gre_oip6.ip6_vfc = IPV6_VERSION; sc->gre_oip6.ip6_nxt = IPPROTO_GRE; gh = &sc->gre_gi6hdr->gi6_gre; break; #endif default: return; } flags = 0; opts = gh->gre_opts; if (sc->gre_options & GRE_ENABLE_CSUM) { flags |= GRE_FLAGS_CP; sc->gre_hlen += 2 * sizeof(uint16_t); *opts++ = 0; } if (sc->gre_key != 0) { flags |= GRE_FLAGS_KP; sc->gre_hlen += sizeof(uint32_t); *opts++ = htonl(sc->gre_key); } if (sc->gre_options & GRE_ENABLE_SEQ) { flags |= GRE_FLAGS_SP; sc->gre_hlen += sizeof(uint32_t); *opts++ = 0; } else sc->gre_oseq = 0; gh->gre_flags = htons(flags); - GRE2IFP(sc)->if_mtu = sc->gre_mtu - sc->gre_hlen; } static void gre_detach(struct gre_softc *sc) { sx_assert(&gre_ioctl_sx, SA_XLOCKED); if (sc->gre_ecookie != NULL) encap_detach(sc->gre_ecookie); sc->gre_ecookie = NULL; } static int gre_set_tunnel(struct ifnet *ifp, struct sockaddr *src, struct sockaddr *dst) { struct gre_softc *sc, *tsc; #ifdef INET6 struct ip6_hdr *ip6; #endif #ifdef INET struct ip *ip; #endif void *hdr; int error; sx_assert(&gre_ioctl_sx, SA_XLOCKED); GRE_LIST_LOCK(); sc = ifp->if_softc; LIST_FOREACH(tsc, &V_gre_softc_list, gre_list) { if (tsc == sc || tsc->gre_family != src->sa_family) continue; #ifdef INET if (tsc->gre_family == AF_INET && tsc->gre_oip.ip_src.s_addr == satosin(src)->sin_addr.s_addr && tsc->gre_oip.ip_dst.s_addr == satosin(dst)->sin_addr.s_addr) { GRE_LIST_UNLOCK(); return (EADDRNOTAVAIL); } #endif #ifdef INET6 if (tsc->gre_family == AF_INET6 && IN6_ARE_ADDR_EQUAL(&tsc->gre_oip6.ip6_src, &satosin6(src)->sin6_addr) && IN6_ARE_ADDR_EQUAL(&tsc->gre_oip6.ip6_dst, &satosin6(dst)->sin6_addr)) { GRE_LIST_UNLOCK(); return (EADDRNOTAVAIL); } #endif } GRE_LIST_UNLOCK(); switch (src->sa_family) { #ifdef INET case AF_INET: hdr = ip = malloc(sizeof(struct greip) + 3 * sizeof(uint32_t), M_GRE, M_WAITOK | M_ZERO); ip->ip_src = satosin(src)->sin_addr; ip->ip_dst = satosin(dst)->sin_addr; break; #endif #ifdef INET6 case AF_INET6: hdr = ip6 = malloc(sizeof(struct greip6) + 3 * sizeof(uint32_t), M_GRE, M_WAITOK | M_ZERO); ip6->ip6_src = satosin6(src)->sin6_addr; ip6->ip6_dst = satosin6(dst)->sin6_addr; break; #endif default: return (EAFNOSUPPORT); } if (sc->gre_family != 0) gre_detach(sc); GRE_WLOCK(sc); if (sc->gre_family != 0) free(sc->gre_hdr, M_GRE); sc->gre_family = src->sa_family; sc->gre_hdr = hdr; sc->gre_oseq = 0; sc->gre_iseq = UINT32_MAX; gre_updatehdr(sc); GRE_WUNLOCK(sc); error = 0; switch (src->sa_family) { #ifdef INET case AF_INET: error = in_gre_attach(sc); break; #endif #ifdef INET6 case AF_INET6: error = in6_gre_attach(sc); break; #endif } if (error == 0) { ifp->if_drv_flags |= IFF_DRV_RUNNING; if_link_state_change(ifp, LINK_STATE_UP); } return (error); } static void gre_delete_tunnel(struct ifnet *ifp) { struct gre_softc *sc = ifp->if_softc; int family; GRE_WLOCK(sc); family = sc->gre_family; sc->gre_family = 0; GRE_WUNLOCK(sc); if (family != 0) { gre_detach(sc); free(sc->gre_hdr, M_GRE); } ifp->if_drv_flags &= ~IFF_DRV_RUNNING; if_link_state_change(ifp, LINK_STATE_DOWN); } int gre_input(struct mbuf **mp, int *offp, int proto) { struct gre_softc *sc; struct grehdr *gh; struct ifnet *ifp; struct mbuf *m; uint32_t *opts; #ifdef notyet uint32_t key; #endif uint16_t flags; int hlen, isr, af; m = *mp; sc = encap_getarg(m); KASSERT(sc != NULL, ("encap_getarg returned NULL")); ifp = GRE2IFP(sc); hlen = *offp + sizeof(struct grehdr) + 4 * sizeof(uint32_t); if (m->m_pkthdr.len < hlen) goto drop; if (m->m_len < hlen) { m = m_pullup(m, hlen); if (m == NULL) goto drop; } gh = (struct grehdr *)mtodo(m, *offp); flags = ntohs(gh->gre_flags); if (flags & ~GRE_FLAGS_MASK) goto drop; opts = gh->gre_opts; hlen = 2 * sizeof(uint16_t); if (flags & GRE_FLAGS_CP) { /* reserved1 field must be zero */ if (((uint16_t *)opts)[1] != 0) goto drop; if (in_cksum_skip(m, m->m_pkthdr.len, *offp) != 0) goto drop; hlen += 2 * sizeof(uint16_t); opts++; } if (flags & GRE_FLAGS_KP) { #ifdef notyet /* * XXX: The current implementation uses the key only for outgoing * packets. But we can check the key value here, or even in the * encapcheck function. */ key = ntohl(*opts); #endif hlen += sizeof(uint32_t); opts++; } #ifdef notyet } else key = 0; if (sc->gre_key != 0 && (key != sc->gre_key || key != 0)) goto drop; #endif if (flags & GRE_FLAGS_SP) { #ifdef notyet seq = ntohl(*opts); #endif hlen += sizeof(uint32_t); } switch (ntohs(gh->gre_proto)) { case ETHERTYPE_WCCP: /* * For WCCP skip an additional 4 bytes if after GRE header * doesn't follow an IP header. */ if (flags == 0 && (*(uint8_t *)gh->gre_opts & 0xF0) != 0x40) hlen += sizeof(uint32_t); /* FALLTHROUGH */ case ETHERTYPE_IP: isr = NETISR_IP; af = AF_INET; break; case ETHERTYPE_IPV6: isr = NETISR_IPV6; af = AF_INET6; break; default: goto drop; } m_adj(m, *offp + hlen); m_clrprotoflags(m); m->m_pkthdr.rcvif = ifp; M_SETFIB(m, ifp->if_fib); #ifdef MAC mac_ifnet_create_mbuf(ifp, m); #endif BPF_MTAP2(ifp, &af, sizeof(af), m); if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1); if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len); if ((ifp->if_flags & IFF_MONITOR) != 0) m_freem(m); else netisr_dispatch(isr, m); return (IPPROTO_DONE); drop: if_inc_counter(ifp, IFCOUNTER_IERRORS, 1); m_freem(m); return (IPPROTO_DONE); } #define MTAG_GRE 1307983903 static int gre_check_nesting(struct ifnet *ifp, struct mbuf *m) { struct m_tag *mtag; int count; count = 1; mtag = NULL; while ((mtag = m_tag_locate(m, MTAG_GRE, 0, mtag)) != NULL) { if (*(struct ifnet **)(mtag + 1) == ifp) { log(LOG_NOTICE, "%s: loop detected\n", ifp->if_xname); return (EIO); } count++; } if (count > V_max_gre_nesting) { log(LOG_NOTICE, "%s: if_output recursively called too many times(%d)\n", ifp->if_xname, count); return (EIO); } mtag = m_tag_alloc(MTAG_GRE, 0, sizeof(struct ifnet *), M_NOWAIT); if (mtag == NULL) return (ENOMEM); *(struct ifnet **)(mtag + 1) = ifp; m_tag_prepend(m, mtag); return (0); } static int gre_output(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst, struct route *ro) { uint32_t af; int error; #ifdef MAC error = mac_ifnet_check_transmit(ifp, m); if (error != 0) goto drop; #endif if ((ifp->if_flags & IFF_MONITOR) != 0 || (ifp->if_flags & IFF_UP) == 0) { error = ENETDOWN; goto drop; } error = gre_check_nesting(ifp, m); if (error != 0) goto drop; m->m_flags &= ~(M_BCAST|M_MCAST); if (dst->sa_family == AF_UNSPEC) bcopy(dst->sa_data, &af, sizeof(af)); else af = dst->sa_family; BPF_MTAP2(ifp, &af, sizeof(af), m); m->m_pkthdr.csum_data = af; /* save af for if_transmit */ return (ifp->if_transmit(ifp, m)); drop: m_freem(m); if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); return (error); } static void gre_setseqn(struct grehdr *gh, uint32_t seq) { uint32_t *opts; uint16_t flags; opts = gh->gre_opts; flags = ntohs(gh->gre_flags); KASSERT((flags & GRE_FLAGS_SP) != 0, ("gre_setseqn called, but GRE_FLAGS_SP isn't set ")); if (flags & GRE_FLAGS_CP) opts++; if (flags & GRE_FLAGS_KP) opts++; *opts = htonl(seq); } static int gre_transmit(struct ifnet *ifp, struct mbuf *m) { GRE_RLOCK_TRACKER; struct gre_softc *sc; struct grehdr *gh; uint32_t iaf, oaf, oseq; int error, hlen, olen, plen; int want_seq, want_csum; plen = 0; sc = ifp->if_softc; if (sc == NULL) { error = ENETDOWN; m_freem(m); goto drop; } GRE_RLOCK(sc); if (sc->gre_family == 0) { GRE_RUNLOCK(sc); error = ENETDOWN; m_freem(m); goto drop; } iaf = m->m_pkthdr.csum_data; oaf = sc->gre_family; hlen = sc->gre_hlen; want_seq = (sc->gre_options & GRE_ENABLE_SEQ) != 0; if (want_seq) oseq = sc->gre_oseq++; /* XXX */ else oseq = 0; /* Make compiler happy. */ want_csum = (sc->gre_options & GRE_ENABLE_CSUM) != 0; M_SETFIB(m, sc->gre_fibnum); M_PREPEND(m, hlen, M_NOWAIT); if (m == NULL) { GRE_RUNLOCK(sc); error = ENOBUFS; goto drop; } bcopy(sc->gre_hdr, mtod(m, void *), hlen); GRE_RUNLOCK(sc); switch (oaf) { #ifdef INET case AF_INET: olen = sizeof(struct ip); break; #endif #ifdef INET6 case AF_INET6: olen = sizeof(struct ip6_hdr); break; #endif default: error = ENETDOWN; goto drop; } gh = (struct grehdr *)mtodo(m, olen); switch (iaf) { #ifdef INET case AF_INET: gh->gre_proto = htons(ETHERTYPE_IP); break; #endif #ifdef INET6 case AF_INET6: gh->gre_proto = htons(ETHERTYPE_IPV6); break; #endif default: error = ENETDOWN; goto drop; } if (want_seq) gre_setseqn(gh, oseq); if (want_csum) { *(uint16_t *)gh->gre_opts = in_cksum_skip(m, m->m_pkthdr.len, olen); } plen = m->m_pkthdr.len - hlen; switch (oaf) { #ifdef INET case AF_INET: error = in_gre_output(m, iaf, hlen); break; #endif #ifdef INET6 case AF_INET6: error = in6_gre_output(m, iaf, hlen); break; #endif default: m_freem(m); error = ENETDOWN; } drop: if (error) if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); else { if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1); if_inc_counter(ifp, IFCOUNTER_OBYTES, plen); } return (error); } static void gre_qflush(struct ifnet *ifp __unused) { } static int gremodevent(module_t mod, int type, void *data) { switch (type) { case MOD_LOAD: case MOD_UNLOAD: break; default: return (EOPNOTSUPP); } return (0); } static moduledata_t gre_mod = { "if_gre", gremodevent, 0 }; DECLARE_MODULE(if_gre, gre_mod, SI_SUB_PSEUDO, SI_ORDER_ANY); MODULE_VERSION(if_gre, 1); Index: head/sys/net/if_gre.h =================================================================== --- head/sys/net/if_gre.h (revision 316698) +++ head/sys/net/if_gre.h (revision 316699) @@ -1,136 +1,135 @@ /*- * Copyright (c) 1998 The NetBSD Foundation, Inc. * Copyright (c) 2014 Andrey V. Elsukov * All rights reserved * * This code is derived from software contributed to The NetBSD Foundation * by Heiko W.Rupp * * 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 NETBSD FOUNDATION, INC. AND CONTRIBUTORS * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE * POSSIBILITY OF SUCH DAMAGE. * * $NetBSD: if_gre.h,v 1.13 2003/11/10 08:51:52 wiz Exp $ * $FreeBSD$ */ #ifndef _NET_IF_GRE_H_ #define _NET_IF_GRE_H_ #ifdef _KERNEL /* GRE header according to RFC 2784 and RFC 2890 */ struct grehdr { uint16_t gre_flags; /* GRE flags */ #define GRE_FLAGS_CP 0x8000 /* checksum present */ #define GRE_FLAGS_KP 0x2000 /* key present */ #define GRE_FLAGS_SP 0x1000 /* sequence present */ #define GRE_FLAGS_MASK (GRE_FLAGS_CP|GRE_FLAGS_KP|GRE_FLAGS_SP) uint16_t gre_proto; /* protocol type */ uint32_t gre_opts[0]; /* optional fields */ } __packed; #ifdef INET struct greip { struct ip gi_ip; struct grehdr gi_gre; } __packed; #endif #ifdef INET6 struct greip6 { struct ip6_hdr gi6_ip6; struct grehdr gi6_gre; } __packed; #endif struct gre_softc { struct ifnet *gre_ifp; LIST_ENTRY(gre_softc) gre_list; struct rmlock gre_lock; int gre_family; /* AF of delivery header */ uint32_t gre_iseq; uint32_t gre_oseq; uint32_t gre_key; uint32_t gre_options; - uint32_t gre_mtu; u_int gre_fibnum; u_int gre_hlen; /* header size */ union { void *hdr; #ifdef INET struct greip *gihdr; #endif #ifdef INET6 struct greip6 *gi6hdr; #endif } gre_uhdr; const struct encaptab *gre_ecookie; }; #define GRE2IFP(sc) ((sc)->gre_ifp) #define GRE_LOCK_INIT(sc) rm_init(&(sc)->gre_lock, "gre softc") #define GRE_LOCK_DESTROY(sc) rm_destroy(&(sc)->gre_lock) #define GRE_RLOCK_TRACKER struct rm_priotracker gre_tracker #define GRE_RLOCK(sc) rm_rlock(&(sc)->gre_lock, &gre_tracker) #define GRE_RUNLOCK(sc) rm_runlock(&(sc)->gre_lock, &gre_tracker) #define GRE_RLOCK_ASSERT(sc) rm_assert(&(sc)->gre_lock, RA_RLOCKED) #define GRE_WLOCK(sc) rm_wlock(&(sc)->gre_lock) #define GRE_WUNLOCK(sc) rm_wunlock(&(sc)->gre_lock) #define GRE_WLOCK_ASSERT(sc) rm_assert(&(sc)->gre_lock, RA_WLOCKED) #define gre_hdr gre_uhdr.hdr #define gre_gihdr gre_uhdr.gihdr #define gre_gi6hdr gre_uhdr.gi6hdr #define gre_oip gre_gihdr->gi_ip #define gre_oip6 gre_gi6hdr->gi6_ip6 int gre_input(struct mbuf **, int *, int); #ifdef INET int in_gre_attach(struct gre_softc *); int in_gre_output(struct mbuf *, int, int); #endif #ifdef INET6 int in6_gre_attach(struct gre_softc *); int in6_gre_output(struct mbuf *, int, int); #endif /* * CISCO uses special type for GRE tunnel created as part of WCCP * connection, while in fact those packets are just IPv4 encapsulated * into GRE. */ #define ETHERTYPE_WCCP 0x883E #endif /* _KERNEL */ #define GRESADDRS _IOW('i', 101, struct ifreq) #define GRESADDRD _IOW('i', 102, struct ifreq) #define GREGADDRS _IOWR('i', 103, struct ifreq) #define GREGADDRD _IOWR('i', 104, struct ifreq) #define GRESPROTO _IOW('i' , 105, struct ifreq) #define GREGPROTO _IOWR('i', 106, struct ifreq) #define GREGKEY _IOWR('i', 107, struct ifreq) #define GRESKEY _IOW('i', 108, struct ifreq) #define GREGOPTS _IOWR('i', 109, struct ifreq) #define GRESOPTS _IOW('i', 110, struct ifreq) #define GRE_ENABLE_CSUM 0x0001 #define GRE_ENABLE_SEQ 0x0002 #define GRE_OPTMASK (GRE_ENABLE_CSUM|GRE_ENABLE_SEQ) #endif /* _NET_IF_GRE_H_ */ Index: head/sys/net/if_me.c =================================================================== --- head/sys/net/if_me.c (revision 316698) +++ head/sys/net/if_me.c (revision 316699) @@ -1,664 +1,664 @@ /*- * Copyright (c) 2014 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 #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include -#define MEMTU 1500 +#define MEMTU (1500 - sizeof(struct mobhdr)) static const char mename[] = "me"; static MALLOC_DEFINE(M_IFME, mename, "Minimal Encapsulation for IP"); static VNET_DEFINE(struct mtx, me_mtx); #define V_me_mtx VNET(me_mtx) /* Minimal forwarding header RFC 2004 */ struct mobhdr { uint8_t mob_proto; /* protocol */ uint8_t mob_flags; /* flags */ #define MOB_FLAGS_SP 0x80 /* source present */ uint16_t mob_csum; /* header checksum */ struct in_addr mob_dst; /* original destination address */ struct in_addr mob_src; /* original source addr (optional) */ } __packed; struct me_softc { struct ifnet *me_ifp; LIST_ENTRY(me_softc) me_list; struct rmlock me_lock; u_int me_fibnum; const struct encaptab *me_ecookie; struct in_addr me_src; struct in_addr me_dst; }; #define ME2IFP(sc) ((sc)->me_ifp) #define ME_READY(sc) ((sc)->me_src.s_addr != 0) #define ME_LOCK_INIT(sc) rm_init(&(sc)->me_lock, "me softc") #define ME_LOCK_DESTROY(sc) rm_destroy(&(sc)->me_lock) #define ME_RLOCK_TRACKER struct rm_priotracker me_tracker #define ME_RLOCK(sc) rm_rlock(&(sc)->me_lock, &me_tracker) #define ME_RUNLOCK(sc) rm_runlock(&(sc)->me_lock, &me_tracker) #define ME_RLOCK_ASSERT(sc) rm_assert(&(sc)->me_lock, RA_RLOCKED) #define ME_WLOCK(sc) rm_wlock(&(sc)->me_lock) #define ME_WUNLOCK(sc) rm_wunlock(&(sc)->me_lock) #define ME_WLOCK_ASSERT(sc) rm_assert(&(sc)->me_lock, RA_WLOCKED) #define ME_LIST_LOCK_INIT(x) mtx_init(&V_me_mtx, "me_mtx", NULL, MTX_DEF) #define ME_LIST_LOCK_DESTROY(x) mtx_destroy(&V_me_mtx) #define ME_LIST_LOCK(x) mtx_lock(&V_me_mtx) #define ME_LIST_UNLOCK(x) mtx_unlock(&V_me_mtx) static VNET_DEFINE(LIST_HEAD(, me_softc), me_softc_list); #define V_me_softc_list VNET(me_softc_list) static struct sx me_ioctl_sx; SX_SYSINIT(me_ioctl_sx, &me_ioctl_sx, "me_ioctl"); static int me_clone_create(struct if_clone *, int, caddr_t); static void me_clone_destroy(struct ifnet *); static VNET_DEFINE(struct if_clone *, me_cloner); #define V_me_cloner VNET(me_cloner) static void me_qflush(struct ifnet *); static int me_transmit(struct ifnet *, struct mbuf *); static int me_ioctl(struct ifnet *, u_long, caddr_t); static int me_output(struct ifnet *, struct mbuf *, const struct sockaddr *, struct route *); static int me_input(struct mbuf **, int *, int); static int me_set_tunnel(struct ifnet *, struct sockaddr_in *, struct sockaddr_in *); static void me_delete_tunnel(struct ifnet *); SYSCTL_DECL(_net_link); static SYSCTL_NODE(_net_link, IFT_TUNNEL, me, CTLFLAG_RW, 0, "Minimal Encapsulation for IP (RFC 2004)"); #ifndef MAX_ME_NEST #define MAX_ME_NEST 1 #endif static VNET_DEFINE(int, max_me_nesting) = MAX_ME_NEST; #define V_max_me_nesting VNET(max_me_nesting) SYSCTL_INT(_net_link_me, OID_AUTO, max_nesting, CTLFLAG_RW | CTLFLAG_VNET, &VNET_NAME(max_me_nesting), 0, "Max nested tunnels"); extern struct domain inetdomain; static const struct protosw in_mobile_protosw = { .pr_type = SOCK_RAW, .pr_domain = &inetdomain, .pr_protocol = IPPROTO_MOBILE, .pr_flags = PR_ATOMIC|PR_ADDR, .pr_input = me_input, .pr_output = rip_output, .pr_ctlinput = rip_ctlinput, .pr_ctloutput = rip_ctloutput, .pr_usrreqs = &rip_usrreqs }; static void vnet_me_init(const void *unused __unused) { LIST_INIT(&V_me_softc_list); ME_LIST_LOCK_INIT(); V_me_cloner = if_clone_simple(mename, me_clone_create, me_clone_destroy, 0); } VNET_SYSINIT(vnet_me_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY, vnet_me_init, NULL); static void vnet_me_uninit(const void *unused __unused) { if_clone_detach(V_me_cloner); ME_LIST_LOCK_DESTROY(); } VNET_SYSUNINIT(vnet_me_uninit, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY, vnet_me_uninit, NULL); static int me_clone_create(struct if_clone *ifc, int unit, caddr_t params) { struct me_softc *sc; sc = malloc(sizeof(struct me_softc), M_IFME, M_WAITOK | M_ZERO); sc->me_fibnum = curthread->td_proc->p_fibnum; ME2IFP(sc) = if_alloc(IFT_TUNNEL); ME_LOCK_INIT(sc); ME2IFP(sc)->if_softc = sc; if_initname(ME2IFP(sc), mename, unit); - ME2IFP(sc)->if_mtu = MEMTU - sizeof(struct mobhdr); + ME2IFP(sc)->if_mtu = MEMTU;; ME2IFP(sc)->if_flags = IFF_POINTOPOINT|IFF_MULTICAST; ME2IFP(sc)->if_output = me_output; ME2IFP(sc)->if_ioctl = me_ioctl; ME2IFP(sc)->if_transmit = me_transmit; ME2IFP(sc)->if_qflush = me_qflush; ME2IFP(sc)->if_capabilities |= IFCAP_LINKSTATE; ME2IFP(sc)->if_capenable |= IFCAP_LINKSTATE; if_attach(ME2IFP(sc)); bpfattach(ME2IFP(sc), DLT_NULL, sizeof(u_int32_t)); ME_LIST_LOCK(); LIST_INSERT_HEAD(&V_me_softc_list, sc, me_list); ME_LIST_UNLOCK(); return (0); } static void me_clone_destroy(struct ifnet *ifp) { struct me_softc *sc; sx_xlock(&me_ioctl_sx); sc = ifp->if_softc; me_delete_tunnel(ifp); ME_LIST_LOCK(); LIST_REMOVE(sc, me_list); ME_LIST_UNLOCK(); bpfdetach(ifp); if_detach(ifp); ifp->if_softc = NULL; sx_xunlock(&me_ioctl_sx); if_free(ifp); ME_LOCK_DESTROY(sc); free(sc, M_IFME); } static int me_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data) { ME_RLOCK_TRACKER; struct ifreq *ifr = (struct ifreq *)data; struct sockaddr_in *src, *dst; struct me_softc *sc; int error; switch (cmd) { case SIOCSIFMTU: if (ifr->ifr_mtu < 576) return (EINVAL); - ifp->if_mtu = ifr->ifr_mtu - sizeof(struct mobhdr); + ifp->if_mtu = ifr->ifr_mtu; return (0); case SIOCSIFADDR: ifp->if_flags |= IFF_UP; case SIOCSIFFLAGS: case SIOCADDMULTI: case SIOCDELMULTI: return (0); } sx_xlock(&me_ioctl_sx); sc = ifp->if_softc; if (sc == NULL) { error = ENXIO; goto end; } error = 0; switch (cmd) { case SIOCSIFPHYADDR: src = (struct sockaddr_in *) &(((struct in_aliasreq *)data)->ifra_addr); dst = (struct sockaddr_in *) &(((struct in_aliasreq *)data)->ifra_dstaddr); if (src->sin_family != dst->sin_family || src->sin_family != AF_INET || src->sin_len != dst->sin_len || src->sin_len != sizeof(struct sockaddr_in)) { error = EINVAL; break; } if (src->sin_addr.s_addr == INADDR_ANY || dst->sin_addr.s_addr == INADDR_ANY) { error = EADDRNOTAVAIL; break; } error = me_set_tunnel(ifp, src, dst); break; case SIOCDIFPHYADDR: me_delete_tunnel(ifp); break; case SIOCGIFPSRCADDR: case SIOCGIFPDSTADDR: ME_RLOCK(sc); if (!ME_READY(sc)) { error = EADDRNOTAVAIL; ME_RUNLOCK(sc); break; } src = (struct sockaddr_in *)&ifr->ifr_addr; memset(src, 0, sizeof(*src)); src->sin_family = AF_INET; src->sin_len = sizeof(*src); switch (cmd) { case SIOCGIFPSRCADDR: src->sin_addr = sc->me_src; break; case SIOCGIFPDSTADDR: src->sin_addr = sc->me_dst; break; } ME_RUNLOCK(sc); error = prison_if(curthread->td_ucred, sintosa(src)); if (error != 0) memset(src, 0, sizeof(*src)); break; case SIOCGTUNFIB: ifr->ifr_fib = sc->me_fibnum; break; case SIOCSTUNFIB: if ((error = priv_check(curthread, PRIV_NET_GRE)) != 0) break; if (ifr->ifr_fib >= rt_numfibs) error = EINVAL; else sc->me_fibnum = ifr->ifr_fib; break; default: error = EINVAL; break; } end: sx_xunlock(&me_ioctl_sx); return (error); } static int me_encapcheck(const struct mbuf *m, int off, int proto, void *arg) { ME_RLOCK_TRACKER; struct me_softc *sc; struct ip *ip; int ret; sc = (struct me_softc *)arg; if ((ME2IFP(sc)->if_flags & IFF_UP) == 0) return (0); M_ASSERTPKTHDR(m); if (m->m_pkthdr.len < sizeof(struct ip) + sizeof(struct mobhdr) - sizeof(struct in_addr)) return (0); ret = 0; ME_RLOCK(sc); if (ME_READY(sc)) { ip = mtod(m, struct ip *); if (sc->me_src.s_addr == ip->ip_dst.s_addr && sc->me_dst.s_addr == ip->ip_src.s_addr) ret = 32 * 2; } ME_RUNLOCK(sc); return (ret); } static int me_set_tunnel(struct ifnet *ifp, struct sockaddr_in *src, struct sockaddr_in *dst) { struct me_softc *sc, *tsc; sx_assert(&me_ioctl_sx, SA_XLOCKED); ME_LIST_LOCK(); sc = ifp->if_softc; LIST_FOREACH(tsc, &V_me_softc_list, me_list) { if (tsc == sc || !ME_READY(tsc)) continue; if (tsc->me_src.s_addr == src->sin_addr.s_addr && tsc->me_dst.s_addr == dst->sin_addr.s_addr) { ME_LIST_UNLOCK(); return (EADDRNOTAVAIL); } } ME_LIST_UNLOCK(); ME_WLOCK(sc); sc->me_dst = dst->sin_addr; sc->me_src = src->sin_addr; ME_WUNLOCK(sc); if (sc->me_ecookie == NULL) sc->me_ecookie = encap_attach_func(AF_INET, IPPROTO_MOBILE, me_encapcheck, &in_mobile_protosw, sc); if (sc->me_ecookie != NULL) { ifp->if_drv_flags |= IFF_DRV_RUNNING; if_link_state_change(ifp, LINK_STATE_UP); } return (0); } static void me_delete_tunnel(struct ifnet *ifp) { struct me_softc *sc = ifp->if_softc; sx_assert(&me_ioctl_sx, SA_XLOCKED); if (sc->me_ecookie != NULL) encap_detach(sc->me_ecookie); sc->me_ecookie = NULL; ME_WLOCK(sc); sc->me_src.s_addr = 0; sc->me_dst.s_addr = 0; ME_WUNLOCK(sc); ifp->if_drv_flags &= ~IFF_DRV_RUNNING; if_link_state_change(ifp, LINK_STATE_DOWN); } static uint16_t me_in_cksum(uint16_t *p, int nwords) { uint32_t sum = 0; while (nwords-- > 0) sum += *p++; sum = (sum >> 16) + (sum & 0xffff); sum += (sum >> 16); return (~sum); } int me_input(struct mbuf **mp, int *offp, int proto) { struct me_softc *sc; struct mobhdr *mh; struct ifnet *ifp; struct mbuf *m; struct ip *ip; int hlen; m = *mp; sc = encap_getarg(m); KASSERT(sc != NULL, ("encap_getarg returned NULL")); ifp = ME2IFP(sc); /* checks for short packets */ hlen = sizeof(struct mobhdr); if (m->m_pkthdr.len < sizeof(struct ip) + hlen) hlen -= sizeof(struct in_addr); if (m->m_len < sizeof(struct ip) + hlen) m = m_pullup(m, sizeof(struct ip) + hlen); if (m == NULL) goto drop; mh = (struct mobhdr *)mtodo(m, sizeof(struct ip)); /* check for wrong flags */ if (mh->mob_flags & (~MOB_FLAGS_SP)) { m_freem(m); goto drop; } if (mh->mob_flags) { if (hlen != sizeof(struct mobhdr)) { m_freem(m); goto drop; } } else hlen = sizeof(struct mobhdr) - sizeof(struct in_addr); /* check mobile header checksum */ if (me_in_cksum((uint16_t *)mh, hlen / sizeof(uint16_t)) != 0) { m_freem(m); goto drop; } #ifdef MAC mac_ifnet_create_mbuf(ifp, m); #endif ip = mtod(m, struct ip *); ip->ip_dst = mh->mob_dst; ip->ip_p = mh->mob_proto; ip->ip_sum = 0; ip->ip_len = htons(m->m_pkthdr.len - hlen); if (mh->mob_flags) ip->ip_src = mh->mob_src; memmove(mtodo(m, hlen), ip, sizeof(struct ip)); m_adj(m, hlen); m_clrprotoflags(m); m->m_pkthdr.rcvif = ifp; m->m_pkthdr.csum_flags |= (CSUM_IP_CHECKED | CSUM_IP_VALID); M_SETFIB(m, ifp->if_fib); hlen = AF_INET; BPF_MTAP2(ifp, &hlen, sizeof(hlen), m); if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1); if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len); if ((ifp->if_flags & IFF_MONITOR) != 0) m_freem(m); else netisr_dispatch(NETISR_IP, m); return (IPPROTO_DONE); drop: if_inc_counter(ifp, IFCOUNTER_IERRORS, 1); return (IPPROTO_DONE); } #define MTAG_ME 1414491977 static int me_check_nesting(struct ifnet *ifp, struct mbuf *m) { struct m_tag *mtag; int count; count = 1; mtag = NULL; while ((mtag = m_tag_locate(m, MTAG_ME, 0, mtag)) != NULL) { if (*(struct ifnet **)(mtag + 1) == ifp) { log(LOG_NOTICE, "%s: loop detected\n", ifp->if_xname); return (EIO); } count++; } if (count > V_max_me_nesting) { log(LOG_NOTICE, "%s: if_output recursively called too many times(%d)\n", ifp->if_xname, count); return (EIO); } mtag = m_tag_alloc(MTAG_ME, 0, sizeof(struct ifnet *), M_NOWAIT); if (mtag == NULL) return (ENOMEM); *(struct ifnet **)(mtag + 1) = ifp; m_tag_prepend(m, mtag); return (0); } static int me_output(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst, struct route *ro) { uint32_t af; int error; #ifdef MAC error = mac_ifnet_check_transmit(ifp, m); if (error != 0) goto drop; #endif if ((ifp->if_flags & IFF_MONITOR) != 0 || (ifp->if_flags & IFF_UP) == 0) { error = ENETDOWN; goto drop; } error = me_check_nesting(ifp, m); if (error != 0) goto drop; m->m_flags &= ~(M_BCAST|M_MCAST); if (dst->sa_family == AF_UNSPEC) bcopy(dst->sa_data, &af, sizeof(af)); else af = dst->sa_family; if (af != AF_INET) { error = EAFNOSUPPORT; goto drop; } BPF_MTAP2(ifp, &af, sizeof(af), m); return (ifp->if_transmit(ifp, m)); drop: m_freem(m); if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); return (error); } static int me_transmit(struct ifnet *ifp, struct mbuf *m) { ME_RLOCK_TRACKER; struct mobhdr mh; struct me_softc *sc; struct ip *ip; int error, hlen, plen; sc = ifp->if_softc; if (sc == NULL) { error = ENETDOWN; m_freem(m); goto drop; } if (m->m_len < sizeof(struct ip)) m = m_pullup(m, sizeof(struct ip)); if (m == NULL) { error = ENOBUFS; goto drop; } ip = mtod(m, struct ip *); /* Fragmented datagramms shouldn't be encapsulated */ if (ip->ip_off & htons(IP_MF | IP_OFFMASK)) { error = EINVAL; m_freem(m); goto drop; } mh.mob_proto = ip->ip_p; mh.mob_src = ip->ip_src; mh.mob_dst = ip->ip_dst; ME_RLOCK(sc); if (!ME_READY(sc)) { ME_RUNLOCK(sc); error = ENETDOWN; m_freem(m); goto drop; } if (in_hosteq(sc->me_src, ip->ip_src)) { hlen = sizeof(struct mobhdr) - sizeof(struct in_addr); mh.mob_flags = 0; } else { hlen = sizeof(struct mobhdr); mh.mob_flags = MOB_FLAGS_SP; } plen = m->m_pkthdr.len; ip->ip_src = sc->me_src; ip->ip_dst = sc->me_dst; M_SETFIB(m, sc->me_fibnum); ME_RUNLOCK(sc); M_PREPEND(m, hlen, M_NOWAIT); if (m == NULL) { error = ENOBUFS; goto drop; } if (m->m_len < sizeof(struct ip) + hlen) m = m_pullup(m, sizeof(struct ip) + hlen); if (m == NULL) { error = ENOBUFS; goto drop; } memmove(mtod(m, void *), mtodo(m, hlen), sizeof(struct ip)); ip = mtod(m, struct ip *); ip->ip_len = htons(m->m_pkthdr.len); ip->ip_p = IPPROTO_MOBILE; ip->ip_sum = 0; mh.mob_csum = 0; mh.mob_csum = me_in_cksum((uint16_t *)&mh, hlen / sizeof(uint16_t)); bcopy(&mh, mtodo(m, sizeof(struct ip)), hlen); error = ip_output(m, NULL, NULL, IP_FORWARDING, NULL, NULL); drop: if (error) if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); else { if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1); if_inc_counter(ifp, IFCOUNTER_OBYTES, plen); } return (error); } static void me_qflush(struct ifnet *ifp __unused) { } static int memodevent(module_t mod, int type, void *data) { switch (type) { case MOD_LOAD: case MOD_UNLOAD: break; default: return (EOPNOTSUPP); } return (0); } static moduledata_t me_mod = { "if_me", memodevent, 0 }; DECLARE_MODULE(if_me, me_mod, SI_SUB_PSEUDO, SI_ORDER_ANY); MODULE_VERSION(if_me, 1);