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 1.135.2.3 retrieving revision 1.165 diff -u -p -r1.135.2.3 -r1.165 --- src/sys/netinet/ip_input.c 2002/03/16 16:02:12 1.135.2.3 +++ src/sys/netinet/ip_input.c 2003/04/11 19:41:37 1.165 @@ -1,9 +1,9 @@ -/* $NetBSD: ip_input.c,v 1.135.2.3 2002/03/16 16:02:12 jdolecek Exp $ */ +/* $NetBSD: ip_input.c,v 1.165 2003/04/11 19:41:37 christos Exp $ */ /* * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. - * + * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: @@ -15,7 +15,7 @@ * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. - * + * * THIS SOFTWARE IS PROVIDED BY THE PROJECT 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 @@ -102,7 +102,7 @@ */ #include -__KERNEL_RCSID(0, "$NetBSD: ip_input.c,v 1.135.2.3 2002/03/16 16:02:12 jdolecek Exp $"); +__KERNEL_RCSID(0, "$NetBSD: ip_input.c,v 1.165 2003/04/11 19:41:37 christos Exp $"); #include "opt_gateway.h" #include "opt_pfil_hooks.h" @@ -168,7 +168,7 @@ __KERNEL_RCSID(0, "$NetBSD: ip_input.c,v #define IPALLOWSRCRT 1 /* allow source-routed packets */ #endif #ifndef IPMTUDISC -#define IPMTUDISC 0 +#define IPMTUDISC 1 #endif #ifndef IPMTUDISCTIMEOUT #define IPMTUDISCTIMEOUT (10 * 60) /* as per RFC 1191 */ @@ -192,15 +192,32 @@ int ip_forwsrcrt = IPFORWSRCRT; int ip_directedbcast = IPDIRECTEDBCAST; int ip_allowsrcrt = IPALLOWSRCRT; int ip_mtudisc = IPMTUDISC; -u_int ip_mtudisc_timeout = IPMTUDISCTIMEOUT; +int ip_mtudisc_timeout = IPMTUDISCTIMEOUT; #ifdef DIAGNOSTIC int ipprintfs = 0; #endif +/* + * XXX - Setting ip_checkinterface mostly implements the receive side of + * the Strong ES model described in RFC 1122, but since the routing table + * and transmit implementation do not implement the Strong ES model, + * setting this to 1 results in an odd hybrid. + * + * XXX - ip_checkinterface currently must be disabled if you use ipnat + * to translate the destination address to another local interface. + * + * XXX - ip_checkinterface must be disabled if you add IP aliases + * to the loopback interface instead of the interface where the + * packets for those addresses are received. + */ +int ip_checkinterface = 0; + struct rttimer_queue *ip_mtudisc_timeout_q = NULL; extern struct domain inetdomain; int ipqmaxlen = IFQ_MAXLEN; +u_long in_ifaddrhash; /* size of hash table - 1 */ +int in_ifaddrentries; /* total number of addrs */ struct in_ifaddrhead in_ifaddr; struct in_ifaddrhashhead *in_ifaddrhashtbl; struct ifqueue ipintrq; @@ -225,7 +242,7 @@ ipq_lock_try() int s; /* - * Use splvm() -- we're bloking things that would cause + * Use splvm() -- we're blocking things that would cause * mbuf allocation. */ s = splvm(); @@ -255,14 +272,14 @@ do { \ printf("%s:%d: ipq already locked\n", __FILE__, __LINE__); \ panic("ipq_lock"); \ } \ -} while (0) +} while (/*CONSTCOND*/ 0) #define IPQ_LOCK_CHECK() \ do { \ if (ipq_locked == 0) { \ printf("%s:%d: ipq lock not held\n", __FILE__, __LINE__); \ panic("ipq lock check"); \ } \ -} while (0) +} while (/*CONSTCOND*/ 0) #else #define IPQ_LOCK() (void) ipq_lock_try() #define IPQ_LOCK_CHECK() /* nothing */ @@ -307,6 +324,11 @@ static struct ip_srcrt { static void save_rte __P((u_char *, struct in_addr)); +#ifdef MBUFTRACE +struct mowner ip_rx_mowner = { "internet", "rx" }; +struct mowner ip_tx_mowner = { "internet", "tx" }; +#endif + /* * IP initialization: fill in IP protocol switch table. * All protocols not implemented in kernel go to raw IP protocol handler. @@ -336,9 +358,7 @@ ip_init() TAILQ_INIT(&in_ifaddr); in_ifaddrhashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, 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 @@ -358,6 +378,11 @@ ip_init() evcnt_attach_static(&ip_hwcsum_ok); evcnt_attach_static(&ip_swcsum); #endif /* INET_CSUM_COUNTERS */ + +#ifdef MBUFTRACE + MOWNER_ATTACH(&ip_tx_mowner); + MOWNER_ATTACH(&ip_rx_mowner); +#endif /* MBUFTRACE */ } struct sockaddr_in ipaddr = { sizeof(ipaddr), AF_INET }; @@ -378,6 +403,7 @@ ipintr() splx(s); if (m == 0) return; + MCLAIM(m, &ip_rx_mowner); ip_input(m); } } @@ -396,7 +422,9 @@ ip_input(struct mbuf *m) struct ipqent *ipqe; int hlen = 0, mff, len; int downmatch; + int checkif; + MCLAIM(m, &ip_rx_mowner); #ifdef DIAGNOSTIC if ((m->m_flags & M_PKTHDR) == 0) panic("ipintr no HDR"); @@ -411,6 +439,7 @@ ip_input(struct mbuf *m) m->m_flags &= ~M_AUTHIPDGM; } #endif + /* * If no IP addresses have been set yet but the interfaces * are receiving, can't do anything with incoming packets yet. @@ -418,10 +447,24 @@ ip_input(struct mbuf *m) if (TAILQ_FIRST(&in_ifaddr) == 0) goto bad; ipstat.ips_total++; - if (m->m_len < sizeof (struct ip) && - (m = m_pullup(m, sizeof (struct ip))) == 0) { - ipstat.ips_toosmall++; - return; + /* + * If the IP 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 IP header is in the first mbuf of the chain. + */ + if (IP_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) { + if ((m = m_copyup(m, sizeof(struct ip), + (max_linkhdr + 3) & ~3)) == NULL) { + /* XXXJRT new stat, please */ + ipstat.ips_toosmall++; + return; + } + } else if (__predict_false(m->m_len < sizeof (struct ip))) { + if ((m = m_pullup(m, sizeof (struct ip))) == NULL) { + ipstat.ips_toosmall++; + return; + } } ip = mtod(m, struct ip *); if (ip->ip_v != IPVERSION) { @@ -509,7 +552,7 @@ ip_input(struct mbuf *m) } #ifdef IPSEC - /* ipflow (IP fast fowarding) is not compatible with IPsec. */ + /* ipflow (IP fast forwarding) is not compatible with IPsec. */ m->m_flags &= ~M_CANFASTFWD; #else /* @@ -556,12 +599,6 @@ ip_input(struct mbuf *m) #endif /* - * Convert fields to host representation. - */ - NTOHS(ip->ip_len); - NTOHS(ip->ip_off); - - /* * Process options and, if not destined for us, * ship it on. ip_dooptions returns 1 when an * error was detected (causing an icmp message @@ -572,6 +609,26 @@ ip_input(struct mbuf *m) return; /* + * Enable a consistency check between the destination address + * and the arrival interface for a unicast packet (the RFC 1122 + * strong ES model) if IP forwarding is disabled and the packet + * is not locally generated. + * + * XXX - Checking also should be disabled if the destination + * address is ipnat'ed to a different interface. + * + * XXX - Checking is incompatible with IP aliases added + * to the loopback interface instead of the interface where + * the packets are received. + * + * XXX - We need to add a per ifaddr flag for this so that + * we get finer grain control. + */ + checkif = ip_checkinterface && (ipforwarding == 0) && + (m->m_pkthdr.rcvif != NULL) && + ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0); + + /* * Check our list of addresses, to see if the packet is for us. * * Traditional 4.4BSD did not consult IFF_UP at all. @@ -581,6 +638,8 @@ ip_input(struct mbuf *m) downmatch = 0; LIST_FOREACH(ia, &IN_IFADDR_HASH(ip->ip_dst.s_addr), ia_hash) { if (in_hosteq(ia->ia_addr.sin_addr, ip->ip_dst)) { + if (checkif && ia->ia_ifp != m->m_pkthdr.rcvif) + continue; if ((ia->ia_ifp->if_flags & IFF_UP) != 0) break; else @@ -616,7 +675,7 @@ ip_input(struct mbuf *m) #ifdef MROUTING extern struct socket *ip_mrouter; - if (m->m_flags & M_EXT) { + if (M_READONLY(m)) { if ((m = m_pullup(m, hlen)) == 0) { ipstat.ips_toosmall++; return; @@ -706,7 +765,15 @@ 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|IP_RF)) { + if (ip->ip_off & ~htons(IP_DF|IP_RF)) { + if (M_READONLY(m)) { + if ((m = m_pullup(m, hlen)) == NULL) { + ipstat.ips_toosmall++; + goto bad; + } + ip = mtod(m, struct ip *); + } + /* * Look for queue of fragments * of this datagram. @@ -726,27 +793,28 @@ found: * set ipqe_mff if more fragments are expected, * convert offset of this to bytes. */ - ip->ip_len -= hlen; - mff = (ip->ip_off & IP_MF) != 0; + ip->ip_len = htons(ntohs(ip->ip_len) - hlen); + mff = (ip->ip_off & htons(IP_MF)) != 0; if (mff) { /* * Make sure that fragments have a data length * that's a non-zero multiple of 8 bytes. */ - if (ip->ip_len == 0 || (ip->ip_len & 0x7) != 0) { + if (ntohs(ip->ip_len) == 0 || + (ntohs(ip->ip_len) & 0x7) != 0) { ipstat.ips_badfrags++; IPQ_UNLOCK(); goto bad; } } - ip->ip_off <<= 3; + ip->ip_off = htons((ntohs(ip->ip_off) & IP_OFFMASK) << 3); /* * If datagram marked as having more fragments * or if this is not the first fragment, * attempt reassembly; if it succeeds, proceed. */ - if (mff || ip->ip_off) { + if (mff || ip->ip_off != htons(0)) { ipstat.ips_fragments++; ipqe = pool_get(&ipqent_pool, PR_NOWAIT); if (ipqe == NULL) { @@ -765,7 +833,7 @@ found: ipstat.ips_reassembled++; ip = mtod(m, struct ip *); hlen = ip->ip_hl << 2; - ip->ip_len += hlen; + ip->ip_len = htons(ntohs(ip->ip_len) + hlen); } else if (fp) ip_freef(fp); @@ -790,7 +858,7 @@ found: */ #if IFA_STATS if (ia && ip) - ia->ia_ifa.ifa_data.ifad_inbytes += ip->ip_len; + ia->ia_ifa.ifa_data.ifad_inbytes += ntohs(ip->ip_len); #endif ipstat.ips_delivered++; { @@ -858,7 +926,7 @@ ip_reass(ipqe, fp) fp->ipq_ttl = IPFRAGTTL; fp->ipq_p = ipqe->ipqe_ip->ip_p; fp->ipq_id = ipqe->ipqe_ip->ip_id; - LIST_INIT(&fp->ipq_fragq); + TAILQ_INIT(&fp->ipq_fragq); fp->ipq_src = ipqe->ipqe_ip->ip_src; fp->ipq_dst = ipqe->ipqe_ip->ip_dst; p = NULL; @@ -868,9 +936,9 @@ ip_reass(ipqe, fp) /* * Find a segment which begins after this one does. */ - for (p = NULL, q = LIST_FIRST(&fp->ipq_fragq); q != NULL; - p = q, q = LIST_NEXT(q, ipqe_q)) - if (q->ipqe_ip->ip_off > ipqe->ipqe_ip->ip_off) + for (p = NULL, q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL; + p = q, q = TAILQ_NEXT(q, ipqe_q)) + if (ntohs(q->ipqe_ip->ip_off) > ntohs(ipqe->ipqe_ip->ip_off)) break; /* @@ -879,14 +947,16 @@ ip_reass(ipqe, fp) * segment. If it provides all of our data, drop us. */ if (p != NULL) { - i = p->ipqe_ip->ip_off + p->ipqe_ip->ip_len - - ipqe->ipqe_ip->ip_off; + i = ntohs(p->ipqe_ip->ip_off) + ntohs(p->ipqe_ip->ip_len) - + ntohs(ipqe->ipqe_ip->ip_off); if (i > 0) { - if (i >= ipqe->ipqe_ip->ip_len) + if (i >= ntohs(ipqe->ipqe_ip->ip_len)) goto dropfrag; m_adj(ipqe->ipqe_m, i); - ipqe->ipqe_ip->ip_off += i; - ipqe->ipqe_ip->ip_len -= i; + ipqe->ipqe_ip->ip_off = + htons(ntohs(ipqe->ipqe_ip->ip_off) + i); + ipqe->ipqe_ip->ip_len = + htons(ntohs(ipqe->ipqe_ip->ip_len) - i); } } @@ -894,19 +964,22 @@ ip_reass(ipqe, fp) * While we overlap succeeding segments trim them or, * if they are completely covered, dequeue them. */ - for (; q != NULL && ipqe->ipqe_ip->ip_off + ipqe->ipqe_ip->ip_len > - q->ipqe_ip->ip_off; q = nq) { - i = (ipqe->ipqe_ip->ip_off + ipqe->ipqe_ip->ip_len) - - q->ipqe_ip->ip_off; - if (i < q->ipqe_ip->ip_len) { - q->ipqe_ip->ip_len -= i; - q->ipqe_ip->ip_off += i; + for (; q != NULL && + ntohs(ipqe->ipqe_ip->ip_off) + ntohs(ipqe->ipqe_ip->ip_len) > + ntohs(q->ipqe_ip->ip_off); q = nq) { + i = (ntohs(ipqe->ipqe_ip->ip_off) + + ntohs(ipqe->ipqe_ip->ip_len)) - ntohs(q->ipqe_ip->ip_off); + if (i < ntohs(q->ipqe_ip->ip_len)) { + q->ipqe_ip->ip_len = + htons(ntohs(q->ipqe_ip->ip_len) - i); + q->ipqe_ip->ip_off = + htons(ntohs(q->ipqe_ip->ip_off) + i); m_adj(q->ipqe_m, i); break; } - nq = LIST_NEXT(q, ipqe_q); + nq = TAILQ_NEXT(q, ipqe_q); m_freem(q->ipqe_m); - LIST_REMOVE(q, ipqe_q); + TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q); pool_put(&ipqent_pool, q); } @@ -916,16 +989,16 @@ insert: * check for complete reassembly. */ if (p == NULL) { - LIST_INSERT_HEAD(&fp->ipq_fragq, ipqe, ipqe_q); + TAILQ_INSERT_HEAD(&fp->ipq_fragq, ipqe, ipqe_q); } else { - LIST_INSERT_AFTER(p, ipqe, ipqe_q); + TAILQ_INSERT_AFTER(&fp->ipq_fragq, p, ipqe, ipqe_q); } next = 0; - for (p = NULL, q = LIST_FIRST(&fp->ipq_fragq); q != NULL; - p = q, q = LIST_NEXT(q, ipqe_q)) { - if (q->ipqe_ip->ip_off != next) + for (p = NULL, q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL; + p = q, q = TAILQ_NEXT(q, ipqe_q)) { + if (ntohs(q->ipqe_ip->ip_off) != next) return (0); - next += q->ipqe_ip->ip_len; + next += ntohs(q->ipqe_ip->ip_len); } if (p->ipqe_mff) return (0); @@ -934,7 +1007,7 @@ insert: * Reassembly is complete. Check for a bogus message size and * concatenate fragments. */ - q = LIST_FIRST(&fp->ipq_fragq); + q = TAILQ_FIRST(&fp->ipq_fragq); ip = q->ipqe_ip; if ((next + (ip->ip_hl << 2)) > IP_MAXPACKET) { ipstat.ips_toolong++; @@ -945,11 +1018,11 @@ insert: t = m->m_next; m->m_next = 0; m_cat(m, t); - nq = LIST_NEXT(q, ipqe_q); + nq = TAILQ_NEXT(q, ipqe_q); pool_put(&ipqent_pool, q); for (q = nq; q != NULL; q = nq) { t = q->ipqe_m; - nq = LIST_NEXT(q, ipqe_q); + nq = TAILQ_NEXT(q, ipqe_q); pool_put(&ipqent_pool, q); m_cat(m, t); } @@ -960,7 +1033,7 @@ insert: * dequeue and discard fragment reassembly header. * Make header visible. */ - ip->ip_len = next; + ip->ip_len = htons(next); ip->ip_src = fp->ipq_src; ip->ip_dst = fp->ipq_dst; LIST_REMOVE(fp, ipq_q); @@ -996,10 +1069,10 @@ ip_freef(fp) IPQ_LOCK_CHECK(); - for (q = LIST_FIRST(&fp->ipq_fragq); q != NULL; q = p) { - p = LIST_NEXT(q, ipqe_q); + for (q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL; q = p) { + p = TAILQ_NEXT(q, ipqe_q); m_freem(q->ipqe_m); - LIST_REMOVE(q, ipqe_q); + TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q); pool_put(&ipqent_pool, q); } LIST_REMOVE(fp, ipq_q); @@ -1367,6 +1440,7 @@ ip_srcroute() if (m == 0) return ((struct mbuf *)0); + MCLAIM(m, &inetdomain.dom_mowner); #define OPTSIZ (sizeof(ip_srcrt.nop) + sizeof(ip_srcrt.srcopt)) /* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */ @@ -1443,7 +1517,7 @@ ip_stripoptions(m, mopt) m->m_len -= olen; if (m->m_flags & M_PKTHDR) m->m_pkthdr.len -= olen; - ip->ip_len -= olen; + ip->ip_len = htons(ntohs(ip->ip_len) - olen); ip->ip_hl = sizeof (struct ip) >> 2; } @@ -1487,6 +1561,11 @@ ip_forward(m, srcrt) #endif /* + * We are now in the output path. + */ + MCLAIM(m, &ip_tx_mowner); + + /* * Clear any in-bound checksum flags for this packet. */ m->m_pkthdr.csum_flags = 0; @@ -1533,7 +1612,7 @@ ip_forward(m, srcrt) * we need to generate an ICMP message to the src. * Pullup to avoid sharing mbuf cluster between m and mcopy. */ - mcopy = m_copym(m, 0, imin((int)ip->ip_len, 68), M_DONTWAIT); + mcopy = m_copym(m, 0, imin(ntohs(ip->ip_len), 68), M_DONTWAIT); if (mcopy) mcopy = m_pullup(mcopy, ip->ip_hl << 2); @@ -1660,6 +1739,8 @@ ip_forward(m, srcrt) ro = &sp->req->sav->sah->sa_route; if (ro->ro_rt && ro->ro_rt->rt_ifp) { dummyifp.if_mtu = + ro->ro_rt->rt_rmx.rmx_mtu ? + ro->ro_rt->rt_rmx.rmx_mtu : ro->ro_rt->rt_ifp->if_mtu; dummyifp.if_mtu -= ipsechdr; destifp = &dummyifp; @@ -1679,7 +1760,7 @@ ip_forward(m, srcrt) * a router should not generate ICMP_SOURCEQUENCH as * required in RFC1812 Requirements for IP Version 4 Routers. * source quench could be a big problem under DoS attacks, - * or the underlying interface is rate-limited. + * or if the underlying interface is rate-limited. */ if (mcopy) m_freem(mcopy); @@ -1802,13 +1883,8 @@ ip_sysctl(name, namelen, oldp, oldlenp, case IPCTL_MTUDISC: error = sysctl_int(oldp, oldlenp, newp, newlen, &ip_mtudisc); - if (ip_mtudisc != 0 && ip_mtudisc_timeout_q == NULL) { - ip_mtudisc_timeout_q = - rt_timer_queue_create(ip_mtudisc_timeout); - } else if (ip_mtudisc == 0 && ip_mtudisc_timeout_q != NULL) { - rt_timer_queue_destroy(ip_mtudisc_timeout_q, TRUE); - ip_mtudisc_timeout_q = NULL; - } + if (error == 0 && ip_mtudisc == 0) + rt_timer_queue_remove_all(ip_mtudisc_timeout_q, TRUE); return error; case IPCTL_ANONPORTMIN: old = anonportmin; @@ -1837,10 +1913,15 @@ ip_sysctl(name, namelen, oldp, oldlenp, } return (error); case IPCTL_MTUDISCTIMEOUT: + old = ip_mtudisc_timeout; error = sysctl_int(oldp, oldlenp, newp, newlen, &ip_mtudisc_timeout); - if (ip_mtudisc_timeout_q != NULL) - rt_timer_queue_change(ip_mtudisc_timeout_q, + if (ip_mtudisc_timeout < 0) { + ip_mtudisc_timeout = old; + return (EINVAL); + } + if (error == 0) + rt_timer_queue_change(ip_mtudisc_timeout_q, ip_mtudisc_timeout); return (error); #ifdef GATEWAY @@ -1861,13 +1942,13 @@ ip_sysctl(name, namelen, oldp, oldlenp, &hostzeroisbroadcast)); #if NGIF > 0 case IPCTL_GIF_TTL: - return(sysctl_int(oldp, oldlenp, newp, newlen, + return (sysctl_int(oldp, oldlenp, newp, newlen, &ip_gif_ttl)); #endif #if NGRE > 0 case IPCTL_GRE_TTL: - return(sysctl_int(oldp, oldlenp, newp, newlen, + return (sysctl_int(oldp, oldlenp, newp, newlen, &ip_gre_ttl)); #endif @@ -1900,6 +1981,9 @@ ip_sysctl(name, namelen, oldp, oldlenp, return (sysctl_int(oldp, oldlenp, newp, newlen, &ip_maxfragpackets)); + case IPCTL_CHECKINTERFACE: + return (sysctl_int(oldp, oldlenp, newp, newlen, + &ip_checkinterface)); default: return (EOPNOTSUPP); }