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/netinet/ip_input.c,v rcsdiff: /ftp/cvs/cvsroot/src/sys/netinet/ip_input.c,v: warning: Unknown phrases like `commitid ...;' are present. retrieving revision 1.25 retrieving revision 1.50.4.1 diff -u -p -r1.25 -r1.50.4.1 --- src/sys/netinet/ip_input.c 1995/11/21 01:07:34 1.25 +++ src/sys/netinet/ip_input.c 1997/09/01 21:00:39 1.50.4.1 @@ -1,4 +1,4 @@ -/* $NetBSD: ip_input.c,v 1.25 1995/11/21 01:07:34 cgd Exp $ */ +/* $NetBSD: ip_input.c,v 1.50.4.1 1997/09/01 21:00:39 thorpej Exp $ */ /* * Copyright (c) 1982, 1986, 1988, 1993 @@ -42,12 +42,19 @@ #include #include #include +#include #include #include #include +#include + +#include +#include #include +#include #include +#include #include #include @@ -57,6 +64,9 @@ #include #include +/* XXX should really put this in libkern.h */ +#define offsetof(type, member) ((size_t)(&((type *)0)->member)) + #ifndef IPFORWARDING #ifdef GATEWAY #define IPFORWARDING 1 /* forward IP packets not for us */ @@ -67,9 +77,29 @@ #ifndef IPSENDREDIRECTS #define IPSENDREDIRECTS 1 #endif +#ifndef IPFORWSRCRT +#define IPFORWSRCRT 1 /* forward source-routed packets */ +#endif +#ifndef IPALLOWSRCRT +#define IPALLOWSRCRT 1 /* allow source-routed packets */ +#endif +/* + * Note: DIRECTED_BROADCAST is handled this way so that previous + * configuration using this option will Just Work. + */ +#ifndef IPDIRECTEDBCAST +#ifdef DIRECTED_BROADCAST +#define IPDIRECTEDBCAST 1 +#else +#define IPDIRECTEDBCAST 0 +#endif /* DIRECTED_BROADCAST */ +#endif /* IPDIRECTEDBCAST */ int ipforwarding = IPFORWARDING; int ipsendredirects = IPSENDREDIRECTS; int ip_defttl = IPDEFTTL; +int ip_forwsrcrt = IPFORWSRCRT; +int ip_directedbcast = IPDIRECTEDBCAST; +int ip_allowsrcrt = IPALLOWSRCRT; #ifdef DIAGNOSTIC int ipprintfs = 0; #endif @@ -97,6 +127,7 @@ static struct ip_srcrt { } ip_srcrt; static void save_rte __P((u_char *, struct in_addr)); + /* * IP initialization: fill in IP protocol switch table. * All protocols not implemented in kernel go to raw IP protocol handler. @@ -133,12 +164,17 @@ struct route ipforward_rt; void ipintr() { - register struct ip *ip; + register struct ip *ip = NULL; register struct mbuf *m; register struct ipq *fp; register struct in_ifaddr *ia; struct ipqent *ipqe; - int hlen, mff, s; + int hlen = 0, mff, len, s; +#ifdef PFIL_HOOKS + struct packet_filter_hook *pfh; + struct mbuf *m0; + int rv; +#endif /* PFIL_HOOKS */ next: /* @@ -183,7 +219,7 @@ next: } ip = mtod(m, struct ip *); } - if (ip->ip_sum = in_cksum(m, hlen)) { + if ((ip->ip_sum = in_cksum(m, hlen)) != 0) { ipstat.ips_badsum++; goto bad; } @@ -192,12 +228,9 @@ next: * Convert fields to host representation. */ NTOHS(ip->ip_len); - if (ip->ip_len < hlen) { - ipstat.ips_badlen++; - goto bad; - } NTOHS(ip->ip_id); NTOHS(ip->ip_off); + len = ip->ip_len; /* * Check that the amount of data in the buffers @@ -205,18 +238,32 @@ next: * Trim mbufs if longer than we expect. * Drop packet if shorter than we expect. */ - if (m->m_pkthdr.len < ip->ip_len) { + if (m->m_pkthdr.len < len) { ipstat.ips_tooshort++; goto bad; } - if (m->m_pkthdr.len > ip->ip_len) { + if (m->m_pkthdr.len > len) { if (m->m_len == m->m_pkthdr.len) { - m->m_len = ip->ip_len; - m->m_pkthdr.len = ip->ip_len; + m->m_len = len; + m->m_pkthdr.len = len; } else - m_adj(m, ip->ip_len - m->m_pkthdr.len); + m_adj(m, len - m->m_pkthdr.len); } +#ifdef PFIL_HOOKS + /* + * Run through list of hooks for input packets. + */ + m0 = m; + for (pfh = pfil_hook_get(PFIL_IN); pfh; pfh = pfh->pfil_link.le_next) + if (pfh->pfil_func) { + rv = pfh->pfil_func(ip, hlen, m->m_pkthdr.rcvif, 0, &m0); + if (rv) + goto next; + ip = mtod(m = m0, struct ip *); + } +#endif /* PFIL_HOOKS */ + /* * Process options and, if not destined for us, * ship it on. ip_dooptions returns 1 when an @@ -231,15 +278,13 @@ next: * Check our list of addresses, to see if the packet is for us. */ for (ia = in_ifaddr.tqh_first; ia; ia = ia->ia_list.tqe_next) { - if (ip->ip_dst.s_addr == ia->ia_addr.sin_addr.s_addr) + if (in_hosteq(ip->ip_dst, ia->ia_addr.sin_addr)) goto ours; - if ( -#ifdef DIRECTED_BROADCAST - ia->ia_ifp == m->m_pkthdr.rcvif && -#endif + if (((ip_directedbcast == 0) || (ip_directedbcast && + ia->ia_ifp == m->m_pkthdr.rcvif)) && (ia->ia_ifp->if_flags & IFF_BROADCAST)) { - if (ip->ip_dst.s_addr == ia->ia_broadaddr.sin_addr.s_addr || - ip->ip_dst.s_addr == ia->ia_netbroadcast.s_addr || + if (in_hosteq(ip->ip_dst, ia->ia_broadaddr.sin_addr) || + in_hosteq(ip->ip_dst, ia->ia_netbroadcast) || /* * Look for all-0's host part (old broadcast addr), * either for subnet or net. @@ -248,6 +293,13 @@ next: ip->ip_dst.s_addr == ia->ia_net) goto ours; } + /* + * An interface with IP address zero accepts + * all packets that arrive on that interface. + */ + if ((ia->ia_ifp == m->m_pkthdr.rcvif) && + in_nullhost(ia->ia_addr.sin_addr)) + goto ours; } if (IN_MULTICAST(ip->ip_dst.s_addr)) { struct in_multi *inm; @@ -306,7 +358,7 @@ next: goto ours; } if (ip->ip_dst.s_addr == INADDR_BROADCAST || - ip->ip_dst.s_addr == INADDR_ANY) + in_nullhost(ip->ip_dst)) goto ours; /* @@ -327,22 +379,15 @@ ours: * if the packet was previously fragmented, * but it's not worth the time; just let them time out.) */ - if (ip->ip_off &~ IP_DF) { - if (m->m_flags & M_EXT) { /* XXX */ - if ((m = m_pullup(m, sizeof (struct ip))) == 0) { - ipstat.ips_toosmall++; - goto next; - } - ip = mtod(m, struct ip *); - } + if (ip->ip_off & ~(IP_DF|IP_RF)) { /* * Look for queue of fragments * of this datagram. */ for (fp = ipq.lh_first; fp != NULL; fp = fp->ipq_q.le_next) if (ip->ip_id == fp->ipq_id && - ip->ip_src.s_addr == fp->ipq_src.s_addr && - ip->ip_dst.s_addr == fp->ipq_dst.s_addr && + in_hosteq(ip->ip_src, fp->ipq_src) && + in_hosteq(ip->ip_dst, fp->ipq_dst) && ip->ip_p == fp->ipq_p) goto found; fp = 0; @@ -381,12 +426,13 @@ found: goto bad; } ipqe->ipqe_mff = mff; + ipqe->ipqe_m = m; ipqe->ipqe_ip = ip; - ip = ip_reass(ipqe, fp); - if (ip == 0) + m = ip_reass(ipqe, fp); + if (m == 0) goto next; ipstat.ips_reassembled++; - m = dtom(ip); + ip = mtod(m, struct ip *); } else if (fp) ip_freef(fp); @@ -410,12 +456,12 @@ bad: * reassembly of this datagram already exists, then it * is given as fp; otherwise have to make a chain. */ -struct ip * +struct mbuf * ip_reass(ipqe, fp) register struct ipqent *ipqe; register struct ipq *fp; { - register struct mbuf *m = dtom(ipqe->ipqe_ip); + register struct mbuf *m = ipqe->ipqe_m; register struct ipqent *nq, *p, *q; struct ip *ip; struct mbuf *t; @@ -433,9 +479,10 @@ ip_reass(ipqe, fp) * If first fragment to arrive, create a reassembly queue. */ if (fp == 0) { - if ((t = m_get(M_DONTWAIT, MT_FTABLE)) == NULL) + MALLOC(fp, struct ipq *, sizeof (struct ipq), + M_FTABLE, M_NOWAIT); + if (fp == NULL) goto dropfrag; - fp = mtod(t, struct ipq *); LIST_INSERT_HEAD(&ipq, fp, ipq_q); fp->ipq_ttl = IPFRAGTTL; fp->ipq_p = ipqe->ipqe_ip->ip_p; @@ -466,7 +513,7 @@ ip_reass(ipqe, fp) if (i > 0) { if (i >= ipqe->ipqe_ip->ip_len) goto dropfrag; - m_adj(dtom(ipqe->ipqe_ip), i); + m_adj(ipqe->ipqe_m, i); ipqe->ipqe_ip->ip_off += i; ipqe->ipqe_ip->ip_len -= i; } @@ -483,11 +530,11 @@ ip_reass(ipqe, fp) if (i < q->ipqe_ip->ip_len) { q->ipqe_ip->ip_len -= i; q->ipqe_ip->ip_off += i; - m_adj(dtom(q->ipqe_ip), i); + m_adj(q->ipqe_m, i); break; } nq = q->ipqe_q.le_next; - m_freem(dtom(q->ipqe_ip)); + m_freem(q->ipqe_m); LIST_REMOVE(q, ipqe_q); FREE(q, M_IPQ); } @@ -513,18 +560,24 @@ insert: return (0); /* - * Reassembly is complete; concatenate fragments. + * Reassembly is complete. Check for a bogus message size and + * concatenate fragments. */ q = fp->ipq_fragq.lh_first; ip = q->ipqe_ip; - m = dtom(q->ipqe_ip); + if ((next + (ip->ip_hl << 2)) > IP_MAXPACKET) { + ipstat.ips_toolong++; + ip_freef(fp); + return (0); + } + m = q->ipqe_m; t = m->m_next; m->m_next = 0; m_cat(m, t); nq = q->ipqe_q.le_next; FREE(q, M_IPQ); for (q = nq; q != NULL; q = nq) { - t = dtom(q->ipqe_ip); + t = q->ipqe_m; nq = q->ipqe_q.le_next; FREE(q, M_IPQ); m_cat(m, t); @@ -540,17 +593,17 @@ insert: ip->ip_src = fp->ipq_src; ip->ip_dst = fp->ipq_dst; LIST_REMOVE(fp, ipq_q); - (void) m_free(dtom(fp)); + FREE(fp, M_FTABLE); m->m_len += (ip->ip_hl << 2); m->m_data -= (ip->ip_hl << 2); /* some debugging cruft by sklower, below, will go away soon */ if (m->m_flags & M_PKTHDR) { /* XXX this should be done elsewhere */ register int plen = 0; - for (t = m; m; m = m->m_next) - plen += m->m_len; - t->m_pkthdr.len = plen; + for (t = m; t; t = t->m_next) + plen += t->m_len; + m->m_pkthdr.len = plen; } - return (ip); + return (m); dropfrag: ipstat.ips_fragdropped++; @@ -571,12 +624,12 @@ ip_freef(fp) for (q = fp->ipq_fragq.lh_first; q != NULL; q = p) { p = q->ipqe_q.le_next; - m_freem(dtom(q->ipqe_ip)); + m_freem(q->ipqe_m); LIST_REMOVE(q, ipqe_q); FREE(q, M_IPQ); } LIST_REMOVE(fp, ipq_q); - (void) m_free(dtom(fp)); + FREE(fp, M_FTABLE); } /* @@ -664,6 +717,11 @@ ip_dooptions(m) */ case IPOPT_LSRR: case IPOPT_SSRR: + if (ip_allowsrcrt == 0) { + type = ICMP_UNREACH; + code = ICMP_UNREACH_NET_PROHIB; + goto bad; + } if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) { code = &cp[IPOPT_OFFSET] - (u_char *)ip; goto bad; @@ -698,8 +756,7 @@ ip_dooptions(m) if (opt == IPOPT_SSRR) { #define INA struct in_ifaddr * #define SA struct sockaddr * - if ((ia = (INA)ifa_ifwithdstaddr((SA)&ipaddr)) == 0) - ia = (INA)ifa_ifwithnet((SA)&ipaddr); + ia = (INA)ifa_ifwithladdr((SA)&ipaddr); } else ia = ip_rtaddr(ipaddr.sin_addr); if (ia == 0) { @@ -796,6 +853,11 @@ ip_dooptions(m) } } if (forward) { + if (ip_forwsrcrt == 0) { + type = ICMP_UNREACH; + code = ICMP_UNREACH_SRCFAIL; + goto bad; + } ip_forward(m, 1); return (1); } @@ -819,7 +881,7 @@ ip_rtaddr(dst) sin = satosin(&ipforward_rt.ro_dst); - if (ipforward_rt.ro_rt == 0 || dst.s_addr != sin->sin_addr.s_addr) { + if (ipforward_rt.ro_rt == 0 || !in_hosteq(dst, sin->sin_addr)) { if (ipforward_rt.ro_rt) { RTFREE(ipforward_rt.ro_rt); ipforward_rt.ro_rt = 0; @@ -892,7 +954,7 @@ ip_srcroute() *(mtod(m, struct in_addr *)) = *p--; #ifdef DIAGNOSTIC if (ipprintfs) - printf(" hops %lx", ntohl(mtod(m, struct in_addr *)->s_addr)); + printf(" hops %x", ntohl(mtod(m, struct in_addr *)->s_addr)); #endif /* @@ -912,7 +974,7 @@ ip_srcroute() while (p >= ip_srcrt.route) { #ifdef DIAGNOSTIC if (ipprintfs) - printf(" %lx", ntohl(q->s_addr)); + printf(" %x", ntohl(q->s_addr)); #endif *q++ = *p--; } @@ -922,7 +984,7 @@ ip_srcroute() *q = ip_srcrt.dst; #ifdef DIAGNOSTIC if (ipprintfs) - printf(" %lx\n", ntohl(q->s_addr)); + printf(" %x\n", ntohl(q->s_addr)); #endif return (m); } @@ -985,7 +1047,7 @@ ip_forward(m, srcrt) register struct ip *ip = mtod(m, struct ip *); register struct sockaddr_in *sin; register struct rtentry *rt; - int error, type = 0, code; + int error, type = 0, code = 0; struct mbuf *mcopy; n_long dest; struct ifnet *destifp; @@ -993,8 +1055,8 @@ ip_forward(m, srcrt) dest = 0; #ifdef DIAGNOSTIC if (ipprintfs) - printf("forward: src %x dst %x ttl %x\n", ip->ip_src, - ip->ip_dst, ip->ip_ttl); + printf("forward: src %x dst %x ttl %x\n", + ip->ip_src.s_addr, ip->ip_dst.s_addr, ip->ip_ttl); #endif if (m->m_flags & M_BCAST || in_canforward(ip->ip_dst) == 0) { ipstat.ips_cantforward++; @@ -1010,13 +1072,13 @@ ip_forward(m, srcrt) sin = satosin(&ipforward_rt.ro_dst); if ((rt = ipforward_rt.ro_rt) == 0 || - ip->ip_dst.s_addr != sin->sin_addr.s_addr) { + !in_hosteq(ip->ip_dst, sin->sin_addr)) { if (ipforward_rt.ro_rt) { RTFREE(ipforward_rt.ro_rt); ipforward_rt.ro_rt = 0; } sin->sin_family = AF_INET; - sin->sin_len = sizeof(*sin); + sin->sin_len = sizeof(struct sockaddr_in); sin->sin_addr = ip->ip_dst; rtalloc(&ipforward_rt); @@ -1028,10 +1090,10 @@ ip_forward(m, srcrt) } /* - * Save at most 64 bytes of the packet in case + * Save at most 68 bytes of the packet in case * we need to generate an ICMP message to the src. */ - mcopy = m_copy(m, 0, imin((int)ip->ip_len, 64)); + mcopy = m_copy(m, 0, imin((int)ip->ip_len, 68)); /* * If forwarding packet using same interface that it came in on, @@ -1043,7 +1105,7 @@ ip_forward(m, srcrt) */ if (rt->rt_ifp == m->m_pkthdr.rcvif && (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 && - satosin(rt_key(rt))->sin_addr.s_addr != 0 && + !in_nullhost(satosin(rt_key(rt))->sin_addr) && ipsendredirects && !srcrt) { if (rt->rt_ifa && (ip->ip_src.s_addr & ifatoia(rt->rt_ifa)->ia_subnetmask) == @@ -1057,16 +1119,13 @@ ip_forward(m, srcrt) code = ICMP_REDIRECT_HOST; #ifdef DIAGNOSTIC if (ipprintfs) - printf("redirect (%d) to %lx\n", code, (u_int32_t)dest); + printf("redirect (%d) to %x\n", code, (u_int32_t)dest); #endif } } - error = ip_output(m, (struct mbuf *)0, &ipforward_rt, IP_FORWARDING -#ifdef DIRECTED_BROADCAST - | IP_ALLOWBROADCAST -#endif - , 0); + error = ip_output(m, (struct mbuf *)0, &ipforward_rt, + (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)), 0); if (error) ipstat.ips_cantforward++; else { @@ -1114,6 +1173,66 @@ ip_forward(m, srcrt) icmp_error(mcopy, type, code, dest, destifp); } +void +ip_savecontrol(inp, mp, ip, m) + register struct inpcb *inp; + register struct mbuf **mp; + register struct ip *ip; + register struct mbuf *m; +{ + + if (inp->inp_socket->so_options & SO_TIMESTAMP) { + struct timeval tv; + + microtime(&tv); + *mp = sbcreatecontrol((caddr_t) &tv, sizeof(tv), + SCM_TIMESTAMP, SOL_SOCKET); + if (*mp) + mp = &(*mp)->m_next; + } + if (inp->inp_flags & INP_RECVDSTADDR) { + *mp = sbcreatecontrol((caddr_t) &ip->ip_dst, + sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP); + if (*mp) + mp = &(*mp)->m_next; + } +#ifdef notyet + /* + * XXX + * Moving these out of udp_input() made them even more broken + * than they already were. + * - fenner@parc.xerox.com + */ + /* options were tossed already */ + if (inp->inp_flags & INP_RECVOPTS) { + *mp = sbcreatecontrol((caddr_t) opts_deleted_above, + sizeof(struct in_addr), IP_RECVOPTS, IPPROTO_IP); + if (*mp) + mp = &(*mp)->m_next; + } + /* ip_srcroute doesn't do what we want here, need to fix */ + if (inp->inp_flags & INP_RECVRETOPTS) { + *mp = sbcreatecontrol((caddr_t) ip_srcroute(), + sizeof(struct in_addr), IP_RECVRETOPTS, IPPROTO_IP); + if (*mp) + mp = &(*mp)->m_next; + } +#endif + if (inp->inp_flags & INP_RECVIF) { + struct sockaddr_dl sdl; + + sdl.sdl_len = offsetof(struct sockaddr_dl, sdl_data[0]); + sdl.sdl_family = AF_LINK; + sdl.sdl_index = m->m_pkthdr.rcvif ? + m->m_pkthdr.rcvif->if_index : 0; + sdl.sdl_nlen = sdl.sdl_alen = sdl.sdl_slen = 0; + *mp = sbcreatecontrol((caddr_t) &sdl, sdl.sdl_len, + IP_RECVIF, IPPROTO_IP); + if (*mp) + mp = &(*mp)->m_next; + } +} + int ip_sysctl(name, namelen, oldp, oldlenp, newp, newlen) int *name; @@ -1139,6 +1258,20 @@ ip_sysctl(name, namelen, oldp, oldlenp, case IPCTL_DEFMTU: return (sysctl_int(oldp, oldlenp, newp, newlen, &ip_mtu)); #endif + case IPCTL_FORWSRCRT: + /* Don't allow this to change in a secure environment. */ + if (securelevel > 0) + return (sysctl_rdint(oldp, oldlenp, newp, + ip_forwsrcrt)); + else + return (sysctl_int(oldp, oldlenp, newp, newlen, + &ip_forwsrcrt)); + case IPCTL_DIRECTEDBCAST: + return (sysctl_int(oldp, oldlenp, newp, newlen, + &ip_directedbcast)); + case IPCTL_ALLOWSRCRT: + return (sysctl_int(oldp, oldlenp, newp, newlen, + &ip_allowsrcrt)); default: return (EOPNOTSUPP); }