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 1.105 retrieving revision 1.114.4.9 diff -u -p -r1.105 -r1.114.4.9 --- src/sys/netinet/ip_input.c 2000/03/03 13:07:42 1.105 +++ src/sys/netinet/ip_input.c 2002/11/10 19:58:05 1.114.4.9 @@ -1,4 +1,4 @@ -/* $NetBSD: ip_input.c,v 1.105 2000/03/03 13:07:42 itojun Exp $ */ +/* $NetBSD: ip_input.c,v 1.114.4.9 2002/11/10 19:58:05 itojun Exp $ */ /* * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. @@ -138,6 +138,12 @@ /* just for gif_ttl */ #include #include "gif.h" +#include +#include "gre.h" + +#ifdef MROUTING +#include +#endif #ifdef IPSEC #include @@ -200,10 +206,11 @@ struct in_ifaddrhashhead *in_ifaddrhasht struct ifqueue ipintrq; struct ipstat ipstat; u_int16_t ip_id; -int ip_defttl; struct ipqhead ipq; int ipq_locked; +int ip_nfragpackets = 0; +int ip_maxfragpackets = 200; static __inline int ipq_lock_try __P((void)); static __inline void ipq_unlock __P((void)); @@ -281,8 +288,8 @@ static void save_rte __P((u_char *, stru void ip_init() { - register struct protosw *pr; - register int i; + struct protosw *pr; + int i; pool_init(&ipqent_pool, sizeof(struct ipqent), 0, 0, 0, "ipqepl", 0, NULL, NULL, M_IPQ); @@ -303,9 +310,7 @@ ip_init() TAILQ_INIT(&in_ifaddr); in_ifaddrhashtbl = hashinit(IN_IFADDR_HASH_SIZE, M_IFADDR, M_WAITOK, &in_ifaddrhash); - if (ip_mtudisc != 0) - ip_mtudisc_timeout_q = - rt_timer_queue_create(ip_mtudisc_timeout); + ip_mtudisc_timeout_q = rt_timer_queue_create(ip_mtudisc_timeout); #ifdef GATEWAY ipflow_init(); #endif @@ -340,10 +345,10 @@ ipintr() void ip_input(struct mbuf *m) { - register struct ip *ip = NULL; - register struct ipq *fp; - register struct in_ifaddr *ia; - register struct ifaddr *ifa; + struct ip *ip = NULL; + struct ipq *fp; + struct in_ifaddr *ia; + struct ifaddr *ifa; struct ipqent *ipqe; int hlen = 0, mff, len; int downmatch; @@ -456,6 +461,14 @@ ip_input(struct mbuf *m) #endif #ifdef PFIL_HOOKS +#ifdef IPSEC + /* + * let ipfilter look at packet on the wire, + * not the decapsulated packet. + */ + if (ipsec_gethist(m, NULL)) + goto nofilt; +#endif /* * Run through list of hooks for input packets. If there are any * filters which require that additional packets in the flow are @@ -476,6 +489,9 @@ ip_input(struct mbuf *m) return; ip = mtod(m, struct ip *); } +#ifdef IPSEC +nofilt:; +#endif #endif /* PFIL_HOOKS */ /* @@ -606,6 +622,13 @@ ip_input(struct mbuf *m) ipstat.ips_cantforward++; return; } +#ifdef IPSEC + if (ipsec4_in_reject(m, NULL)) { + ipsecstat.in_polvio++; + goto bad; + } +#endif + ip_forward(m, 0); } return; @@ -684,6 +707,19 @@ found: IPQ_UNLOCK(); } +#ifdef IPSEC + /* + * enforce IPsec policy checking if we are seeing last header. + * note that we do not visit this with protocols with pcb layer + * code - like udp/tcp/raw ip. + */ + if ((inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0 && + ipsec4_in_reject(m, NULL)) { + ipsecstat.in_polvio++; + goto bad; + } +#endif + /* * Switch out to protocol's input routine. */ @@ -709,11 +745,11 @@ bad: */ struct mbuf * ip_reass(ipqe, fp) - register struct ipqent *ipqe; - register struct ipq *fp; + struct ipqent *ipqe; + struct ipq *fp; { - register struct mbuf *m = ipqe->ipqe_m; - register struct ipqent *nq, *p, *q; + struct mbuf *m = ipqe->ipqe_m; + struct ipqent *nq, *p, *q; struct ip *ip; struct mbuf *t; int hlen = ipqe->ipqe_ip->ip_hl << 2; @@ -732,6 +768,17 @@ ip_reass(ipqe, fp) * If first fragment to arrive, create a reassembly queue. */ if (fp == 0) { + /* + * Enforce upper bound on number of fragmented packets + * for which we attempt reassembly; + * If maxfrag is 0, never accept fragments. + * If maxfrag is -1, accept all fragments without limitation. + */ + if (ip_maxfragpackets < 0) + ; + else if (ip_nfragpackets >= ip_maxfragpackets) + goto dropfrag; + ip_nfragpackets++; MALLOC(fp, struct ipq *, sizeof (struct ipq), M_FTABLE, M_NOWAIT); if (fp == NULL) @@ -847,11 +894,12 @@ insert: ip->ip_dst = fp->ipq_dst; LIST_REMOVE(fp, ipq_q); FREE(fp, M_FTABLE); + ip_nfragpackets--; 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; + int plen = 0; for (t = m; t; t = t->m_next) plen += t->m_len; m->m_pkthdr.len = plen; @@ -873,7 +921,7 @@ void ip_freef(fp) struct ipq *fp; { - register struct ipqent *q, *p; + struct ipqent *q, *p; IPQ_LOCK_CHECK(); @@ -885,6 +933,7 @@ ip_freef(fp) } LIST_REMOVE(fp, ipq_q); FREE(fp, M_FTABLE); + ip_nfragpackets--; } /* @@ -895,7 +944,7 @@ ip_freef(fp) void ip_slowtimo() { - register struct ipq *fp, *nfp; + struct ipq *fp, *nfp; int s = splsoftnet(); IPQ_LOCK(); @@ -906,6 +955,17 @@ ip_slowtimo() ip_freef(fp); } } + /* + * If we are over the maximum number of fragments + * (due to the limit being lowered), drain off + * enough to get down to the new limit. + */ + if (ip_maxfragpackets < 0) + ; + else { + while (ip_nfragpackets > ip_maxfragpackets && ipq.lh_first) + ip_freef(ipq.lh_first); + } IPQ_UNLOCK(); #ifdef GATEWAY ipflow_slowtimo(); @@ -946,10 +1006,10 @@ int ip_dooptions(m) struct mbuf *m; { - register struct ip *ip = mtod(m, struct ip *); - register u_char *cp, *cp0; - register struct ip_timestamp *ipt; - register struct in_ifaddr *ia; + struct ip *ip = mtod(m, struct ip *); + u_char *cp, *cp0; + struct ip_timestamp *ipt; + struct in_ifaddr *ia; int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0; struct in_addr dst; n_time ntime; @@ -964,8 +1024,12 @@ ip_dooptions(m) if (opt == IPOPT_NOP) optlen = 1; else { + if (cnt < IPOPT_OLEN + sizeof(*cp)) { + code = &cp[IPOPT_OLEN] - (u_char *)ip; + goto bad; + } optlen = cp[IPOPT_OLEN]; - if (optlen <= 0 || optlen > cnt) { + if (optlen < IPOPT_OLEN + sizeof(*cp) || optlen > cnt) { code = &cp[IPOPT_OLEN] - (u_char *)ip; goto bad; } @@ -991,6 +1055,10 @@ ip_dooptions(m) code = ICMP_UNREACH_NET_PROHIB; goto bad; } + if (optlen < IPOPT_OFFSET + sizeof(*cp)) { + code = &cp[IPOPT_OLEN] - (u_char *)ip; + goto bad; + } if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) { code = &cp[IPOPT_OFFSET] - (u_char *)ip; goto bad; @@ -1010,7 +1078,7 @@ ip_dooptions(m) break; } off--; /* 0 origin */ - if (off > optlen - sizeof(struct in_addr)) { + if ((off + sizeof(struct in_addr)) > optlen) { /* * End of source route. Should be for us. */ @@ -1042,6 +1110,10 @@ ip_dooptions(m) break; case IPOPT_RR: + if (optlen < IPOPT_OFFSET + sizeof(*cp)) { + code = &cp[IPOPT_OLEN] - (u_char *)ip; + goto bad; + } if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) { code = &cp[IPOPT_OFFSET] - (u_char *)ip; goto bad; @@ -1050,7 +1122,7 @@ ip_dooptions(m) * If no space remains, ignore. */ off--; /* 0 origin */ - if (off > optlen - sizeof(struct in_addr)) + if ((off + sizeof(struct in_addr)) > optlen) break; bcopy((caddr_t)(&ip->ip_dst), (caddr_t)&ipaddr.sin_addr, sizeof(ipaddr.sin_addr)); @@ -1073,11 +1145,20 @@ ip_dooptions(m) case IPOPT_TS: code = cp - (u_char *)ip; ipt = (struct ip_timestamp *)cp; - if (ipt->ipt_len < 5) + if (ipt->ipt_len < 4 || ipt->ipt_len > 40) { + code = (u_char *)&ipt->ipt_len - (u_char *)ip; + goto bad; + } + if (ipt->ipt_ptr < 5) { + code = (u_char *)&ipt->ipt_ptr - (u_char *)ip; goto bad; + } if (ipt->ipt_ptr > ipt->ipt_len - sizeof (int32_t)) { - if (++ipt->ipt_oflw == 0) + if (++ipt->ipt_oflw == 0) { + code = (u_char *)&ipt->ipt_ptr - + (u_char *)ip; goto bad; + } break; } cp0 = (cp + ipt->ipt_ptr - 1); @@ -1088,8 +1169,11 @@ ip_dooptions(m) case IPOPT_TS_TSANDADDR: if (ipt->ipt_ptr - 1 + sizeof(n_time) + - sizeof(struct in_addr) > ipt->ipt_len) + sizeof(struct in_addr) > ipt->ipt_len) { + code = (u_char *)&ipt->ipt_ptr - + (u_char *)ip; goto bad; + } ipaddr.sin_addr = dst; ia = ifatoia(ifaof_ifpforaddr(sintosa(&ipaddr), m->m_pkthdr.rcvif)); @@ -1102,8 +1186,11 @@ ip_dooptions(m) case IPOPT_TS_PRESPEC: if (ipt->ipt_ptr - 1 + sizeof(n_time) + - sizeof(struct in_addr) > ipt->ipt_len) + sizeof(struct in_addr) > ipt->ipt_len) { + code = (u_char *)&ipt->ipt_ptr - + (u_char *)ip; goto bad; + } bcopy(cp0, &ipaddr.sin_addr, sizeof(struct in_addr)); if (ifatoia(ifa_ifwithaddr(sintosa(&ipaddr))) @@ -1113,10 +1200,13 @@ ip_dooptions(m) break; default: + /* XXX can't take &ipt->ipt_flg */ + code = (u_char *)&ipt->ipt_ptr - + (u_char *)ip + 1; goto bad; } ntime = iptime(); - cp0 = (u_char *) &ntime; /* XXX GCC BUG */ + cp0 = (u_char *) &ntime; /* XXX grumble, GCC... */ bcopy(cp0, (caddr_t)cp + ipt->ipt_ptr - 1, sizeof(n_time)); ipt->ipt_ptr += sizeof(n_time); @@ -1146,7 +1236,7 @@ struct in_ifaddr * ip_rtaddr(dst) struct in_addr dst; { - register struct sockaddr_in *sin; + struct sockaddr_in *sin; sin = satosin(&ipforward_rt.ro_dst); @@ -1197,8 +1287,8 @@ save_rte(option, dst) struct mbuf * ip_srcroute() { - register struct in_addr *p, *q; - register struct mbuf *m; + struct in_addr *p, *q; + struct mbuf *m; if (ip_nhops == 0) return ((struct mbuf *)0); @@ -1267,12 +1357,12 @@ ip_srcroute() */ void ip_stripoptions(m, mopt) - register struct mbuf *m; + struct mbuf *m; struct mbuf *mopt; { - register int i; + int i; struct ip *ip = mtod(m, struct ip *); - register caddr_t opts; + caddr_t opts; int olen; olen = (ip->ip_hl << 2) - sizeof (struct ip); @@ -1314,9 +1404,9 @@ ip_forward(m, srcrt) struct mbuf *m; int srcrt; { - register struct ip *ip = mtod(m, struct ip *); - register struct sockaddr_in *sin; - register struct rtentry *rt; + struct ip *ip = mtod(m, struct ip *); + struct sockaddr_in *sin; + struct rtentry *rt; int error, type = 0, code = 0; struct mbuf *mcopy; n_long dest; @@ -1365,8 +1455,11 @@ ip_forward(m, srcrt) /* * Save at most 68 bytes of the packet in case * we need to generate an ICMP message to the src. + * Pullup to avoid sharing mbuf cluster between m and mcopy. */ - mcopy = m_copy(m, 0, imin((int)ip->ip_len, 68)); + mcopy = m_copym(m, 0, imin((int)ip->ip_len, 68), M_DONTWAIT); + if (mcopy) + mcopy = m_pullup(mcopy, ip->ip_hl << 2); /* * If forwarding packet using same interface that it came in on, @@ -1403,7 +1496,7 @@ ip_forward(m, srcrt) #ifdef IPSEC /* Don't lookup socket in forwading case */ - ipsec_setsocket(m, NULL); + (void)ipsec_setsocket(m, NULL); #endif error = ip_output(m, (struct mbuf *)0, &ipforward_rt, (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)), 0); @@ -1514,10 +1607,10 @@ ip_forward(m, srcrt) void ip_savecontrol(inp, mp, ip, m) - register struct inpcb *inp; - register struct mbuf **mp; - register struct ip *ip; - register struct mbuf *m; + struct inpcb *inp; + struct mbuf **mp; + struct ip *ip; + struct mbuf *m; { if (inp->inp_socket->so_options & SO_TIMESTAMP) { @@ -1632,7 +1725,8 @@ ip_sysctl(name, namelen, oldp, oldlenp, case IPCTL_ANONPORTMIN: old = anonportmin; error = sysctl_int(oldp, oldlenp, newp, newlen, &anonportmin); - if (anonportmin >= anonportmax || anonportmin > 65535 + if (anonportmin >= anonportmax || anonportmin < 0 + || anonportmin > 65535 #ifndef IPNOPRIVPORTS || anonportmin < IPPORT_RESERVED #endif @@ -1644,7 +1738,8 @@ ip_sysctl(name, namelen, oldp, oldlenp, case IPCTL_ANONPORTMAX: old = anonportmax; error = sysctl_int(oldp, oldlenp, newp, newlen, &anonportmax); - if (anonportmin >= anonportmax || anonportmax > 65535 + if (anonportmin >= anonportmax || anonportmax < 0 + || anonportmax > 65535 #ifndef IPNOPRIVPORTS || anonportmax < IPPORT_RESERVED #endif @@ -1682,6 +1777,41 @@ ip_sysctl(name, namelen, oldp, oldlenp, &ip_gif_ttl)); #endif +#if NGRE > 0 + case IPCTL_GRE_TTL: + return(sysctl_int(oldp, oldlenp, newp, newlen, + &ip_gre_ttl)); +#endif + +#ifndef IPNOPRIVPORTS + case IPCTL_LOWPORTMIN: + old = lowportmin; + error = sysctl_int(oldp, oldlenp, newp, newlen, &lowportmin); + if (lowportmin >= lowportmax + || lowportmin > IPPORT_RESERVEDMAX + || lowportmin < IPPORT_RESERVEDMIN + ) { + lowportmin = old; + return (EINVAL); + } + return (error); + case IPCTL_LOWPORTMAX: + old = lowportmax; + error = sysctl_int(oldp, oldlenp, newp, newlen, &lowportmax); + if (lowportmin >= lowportmax + || lowportmax > IPPORT_RESERVEDMAX + || lowportmax < IPPORT_RESERVEDMIN + ) { + lowportmax = old; + return (EINVAL); + } + return (error); +#endif + + case IPCTL_MAXFRAGPACKETS: + return (sysctl_int(oldp, oldlenp, newp, newlen, + &ip_maxfragpackets)); + default: return (EOPNOTSUPP); }