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 1.166 retrieving revision 1.376.2.3 diff -u -p -r1.166 -r1.376.2.3 --- src/sys/netinet/ip_input.c 2003/06/15 02:49:33 1.166 +++ src/sys/netinet/ip_input.c 2018/05/21 04:36:16 1.376.2.3 @@ -1,4 +1,4 @@ -/* $NetBSD: ip_input.c,v 1.166 2003/06/15 02:49:33 matt Exp $ */ +/* $NetBSD: ip_input.c,v 1.376.2.3 2018/05/21 04:36:16 pgoyette Exp $ */ /* * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. @@ -29,7 +29,7 @@ * SUCH DAMAGE. */ -/*- +/* * Copyright (c) 1998 The NetBSD Foundation, Inc. * All rights reserved. * @@ -45,13 +45,6 @@ * 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. - * 3. All advertising materials mentioning features or use of this software - * must display the following acknowledgement: - * This product includes software developed by the NetBSD - * Foundation, Inc. and its contributors. - * 4. Neither the name of The NetBSD Foundation 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 NETBSD FOUNDATION, INC. AND CONTRIBUTORS * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED @@ -78,11 +71,7 @@ * 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. - * 3. All advertising materials mentioning features or use of this software - * must display the following acknowledgement: - * This product includes software developed by the University of - * California, Berkeley and its contributors. - * 4. Neither the name of the University nor the names of its contributors + * 3. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * @@ -102,17 +91,23 @@ */ #include -__KERNEL_RCSID(0, "$NetBSD: ip_input.c,v 1.166 2003/06/15 02:49:33 matt Exp $"); +__KERNEL_RCSID(0, "$NetBSD: ip_input.c,v 1.376.2.3 2018/05/21 04:36:16 pgoyette Exp $"); +#ifdef _KERNEL_OPT +#include "opt_inet.h" #include "opt_gateway.h" -#include "opt_pfil_hooks.h" #include "opt_ipsec.h" #include "opt_mrouting.h" +#include "opt_mbuftrace.h" #include "opt_inet_csum.h" +#include "opt_net_mpsafe.h" +#endif + +#include "arp.h" #include #include -#include +#include #include #include #include @@ -123,18 +118,22 @@ __KERNEL_RCSID(0, "$NetBSD: ip_input.c,v #include #include #include +#include #include #include #include +#include #include #include #include #include #include +#include #include #include +#include #include /* just for gif_ttl */ #include @@ -145,57 +144,32 @@ __KERNEL_RCSID(0, "$NetBSD: ip_input.c,v #ifdef MROUTING #include #endif +#include #ifdef IPSEC -#include -#include +#include #endif #ifndef IPFORWARDING #ifdef GATEWAY #define IPFORWARDING 1 /* forward IP packets not for us */ -#else /* GATEWAY */ +#else #define IPFORWARDING 0 /* don't forward IP packets not for us */ -#endif /* GATEWAY */ -#endif /* IPFORWARDING */ -#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 -#ifndef IPMTUDISC -#define IPMTUDISC 1 -#endif -#ifndef IPMTUDISCTIMEOUT + #define IPMTUDISCTIMEOUT (10 * 60) /* as per RFC 1191 */ -#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; -int ip_mtudisc = IPMTUDISC; -int ip_mtudisc_timeout = IPMTUDISCTIMEOUT; -#ifdef DIAGNOSTIC -int ipprintfs = 0; -#endif +int ipforwarding = IPFORWARDING; +int ipsendredirects = 1; +int ip_defttl = IPDEFTTL; +int ip_forwsrcrt = 0; +int ip_directedbcast = 0; +int ip_allowsrcrt = 0; +int ip_mtudisc = 1; +int ip_mtudisc_timeout = IPMTUDISCTIMEOUT; +int ip_do_randomid = 0; + /* * XXX - Setting ip_checkinterface mostly implements the receive side of * the Strong ES model described in RFC 1122, but since the routing table @@ -209,90 +183,18 @@ int ipprintfs = 0; * to the loopback interface instead of the interface where the * packets for those addresses are received. */ -int ip_checkinterface = 0; - +static int ip_checkinterface __read_mostly = 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; -u_long in_multihash; /* size of hash table - 1 */ -int in_multientries; /* total number of addrs */ -struct in_multihead in_multi; -struct in_multihashhead *in_multihashtbl; -struct ifqueue ipintrq; -struct ipstat ipstat; -u_int16_t ip_id; - -#ifdef PFIL_HOOKS -struct pfil_head inet_pfil_hook; -#endif - -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)); - -static __inline int -ipq_lock_try() -{ - int s; +pktqueue_t * ip_pktq __read_mostly; +pfil_head_t * inet_pfil_hook __read_mostly; +ipid_state_t * ip_ids __read_mostly; +percpu_t * ipstat_percpu __read_mostly; - /* - * Use splvm() -- we're blocking things that would cause - * mbuf allocation. - */ - s = splvm(); - if (ipq_locked) { - splx(s); - return (0); - } - ipq_locked = 1; - splx(s); - return (1); -} +static percpu_t *ipforward_rt_percpu __cacheline_aligned; -static __inline void -ipq_unlock() -{ - int s; - - s = splvm(); - ipq_locked = 0; - splx(s); -} - -#ifdef DIAGNOSTIC -#define IPQ_LOCK() \ -do { \ - if (ipq_lock_try() == 0) { \ - printf("%s:%d: ipq already locked\n", __FILE__, __LINE__); \ - panic("ipq_lock"); \ - } \ -} 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 (/*CONSTCOND*/ 0) -#else -#define IPQ_LOCK() (void) ipq_lock_try() -#define IPQ_LOCK_CHECK() /* nothing */ -#endif - -#define IPQ_UNLOCK() ipq_unlock() - -struct pool inmulti_pool; -struct pool ipqent_pool; +uint16_t ip_id; #ifdef INET_CSUM_COUNTERS #include @@ -306,6 +208,10 @@ struct evcnt ip_swcsum = EVCNT_INITIALIZ #define INET_CSUM_COUNTER_INCR(ev) (ev)->ev_count++ +EVCNT_ATTACH_STATIC(ip_hwcsum_bad); +EVCNT_ATTACH_STATIC(ip_hwcsum_ok); +EVCNT_ATTACH_STATIC(ip_swcsum); + #else #define INET_CSUM_COUNTER_INCR(ev) /* nothing */ @@ -313,25 +219,47 @@ struct evcnt ip_swcsum = EVCNT_INITIALIZ #endif /* INET_CSUM_COUNTERS */ /* - * We need to save the IP options in case a protocol wants to respond + * Used to save the IP options in case a protocol wants to respond * to an incoming packet over the same route if the packet got here * using IP source routing. This allows connection establishment and * maintenance when the remote end is on a network that is not known * to us. */ -int ip_nhops = 0; -static struct ip_srcrt { - struct in_addr dst; /* final destination */ - char nop; /* one NOP to align */ - char srcopt[IPOPT_OFFSET + 1]; /* OPTVAL, OLEN and OFFSET */ - struct in_addr route[MAX_IPOPTLEN/sizeof(struct in_addr)]; -} ip_srcrt; +struct ip_srcrt { + int isr_nhops; /* number of hops */ + struct in_addr isr_dst; /* final destination */ + char isr_nop; /* one NOP to align */ + char isr_hdr[IPOPT_OFFSET + 1]; /* OPTVAL, OLEN & OFFSET */ + struct in_addr isr_routes[MAX_IPOPTLEN/sizeof(struct in_addr)]; +}; -static void save_rte __P((u_char *, struct in_addr)); +static int ip_drainwanted; + +static void save_rte(struct mbuf *, u_char *, struct in_addr); #ifdef MBUFTRACE -struct mowner ip_rx_mowner = { "internet", "rx" }; -struct mowner ip_tx_mowner = { "internet", "tx" }; +struct mowner ip_rx_mowner = MOWNER_INIT("internet", "rx"); +struct mowner ip_tx_mowner = MOWNER_INIT("internet", "tx"); +#endif + +static void ipintr(void *); +static void ip_input(struct mbuf *); +static void ip_forward(struct mbuf *, int, struct ifnet *); +static bool ip_dooptions(struct mbuf *); +static struct in_ifaddr *ip_rtaddr(struct in_addr, struct psref *); +static void sysctl_net_inet_ip_setup(struct sysctllog **); + +static struct in_ifaddr *ip_match_our_address(struct ifnet *, struct ip *, + int *); +static struct in_ifaddr *ip_match_our_address_broadcast(struct ifnet *, + struct ip *); + +#ifdef NET_MPSAFE +#define SOFTNET_LOCK() mutex_enter(softnet_lock) +#define SOFTNET_UNLOCK() mutex_exit(softnet_lock) +#else +#define SOFTNET_LOCK() KASSERT(mutex_owned(softnet_lock)) +#define SOFTNET_UNLOCK() KASSERT(mutex_owned(softnet_lock)) #endif /* @@ -339,156 +267,218 @@ struct mowner ip_tx_mowner = { "internet * All protocols not implemented in kernel go to raw IP protocol handler. */ void -ip_init() +ip_init(void) { - struct protosw *pr; - int i; + const struct protosw *pr; - pool_init(&inmulti_pool, sizeof(struct in_multi), 0, 0, 0, "inmltpl", - NULL); - pool_init(&ipqent_pool, sizeof(struct ipqent), 0, 0, 0, "ipqepl", - NULL); + in_init(); + sysctl_net_inet_ip_setup(NULL); pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW); - if (pr == 0) - panic("ip_init"); - for (i = 0; i < IPPROTO_MAX; i++) + KASSERT(pr != NULL); + + ip_pktq = pktq_create(IFQ_MAXLEN, ipintr, NULL); + KASSERT(ip_pktq != NULL); + + for (u_int i = 0; i < IPPROTO_MAX; i++) { ip_protox[i] = pr - inetsw; + } for (pr = inetdomain.dom_protosw; pr < inetdomain.dom_protoswNPROTOSW; pr++) if (pr->pr_domain->dom_family == PF_INET && pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW) ip_protox[pr->pr_protocol] = pr - inetsw; - LIST_INIT(&ipq); - ip_id = time.tv_sec & 0xffff; - ipintrq.ifq_maxlen = ipqmaxlen; - TAILQ_INIT(&in_ifaddr); - in_ifaddrhashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, M_IFADDR, - M_WAITOK, &in_ifaddrhash); - in_multihashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, M_IPMADDR, - M_WAITOK, &in_multihash); - ip_mtudisc_timeout_q = rt_timer_queue_create(ip_mtudisc_timeout); + + ip_reass_init(); + + ip_ids = ip_id_init(); + ip_id = time_uptime & 0xfffff; + #ifdef GATEWAY ipflow_init(); #endif -#ifdef PFIL_HOOKS /* Register our Packet Filter hook. */ - inet_pfil_hook.ph_type = PFIL_TYPE_AF; - inet_pfil_hook.ph_af = AF_INET; - i = pfil_head_register(&inet_pfil_hook); - if (i != 0) - printf("ip_init: WARNING: unable to register pfil hook, " - "error %d\n", i); -#endif /* PFIL_HOOKS */ - -#ifdef INET_CSUM_COUNTERS - evcnt_attach_static(&ip_hwcsum_bad); - evcnt_attach_static(&ip_hwcsum_ok); - evcnt_attach_static(&ip_swcsum); -#endif /* INET_CSUM_COUNTERS */ + inet_pfil_hook = pfil_head_create(PFIL_TYPE_AF, (void *)AF_INET); + KASSERT(inet_pfil_hook != NULL); #ifdef MBUFTRACE MOWNER_ATTACH(&ip_tx_mowner); MOWNER_ATTACH(&ip_rx_mowner); -#endif /* MBUFTRACE */ +#endif + + ipstat_percpu = percpu_alloc(sizeof(uint64_t) * IP_NSTATS); + ipforward_rt_percpu = percpu_alloc(sizeof(struct route)); + ip_mtudisc_timeout_q = rt_timer_queue_create(ip_mtudisc_timeout); } -struct sockaddr_in ipaddr = { sizeof(ipaddr), AF_INET }; -struct route ipforward_rt; +static struct in_ifaddr * +ip_match_our_address(struct ifnet *ifp, struct ip *ip, int *downmatch) +{ + struct in_ifaddr *ia = NULL; + int checkif; + + /* + * 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) && + (ifp->if_flags & IFF_LOOPBACK) == 0; + + IN_ADDRHASH_READER_FOREACH(ia, ip->ip_dst.s_addr) { + if (in_hosteq(ia->ia_addr.sin_addr, ip->ip_dst)) { + if (ia->ia4_flags & IN_IFF_NOTREADY) + continue; + if (checkif && ia->ia_ifp != ifp) + continue; + if ((ia->ia_ifp->if_flags & IFF_UP) == 0) { + (*downmatch)++; + continue; + } + if (ia->ia4_flags & IN_IFF_DETACHED && + (ifp->if_flags & IFF_LOOPBACK) == 0) + continue; + break; + } + } + + return ia; +} + +static struct in_ifaddr * +ip_match_our_address_broadcast(struct ifnet *ifp, struct ip *ip) +{ + struct in_ifaddr *ia = NULL; + struct ifaddr *ifa; + + IFADDR_READER_FOREACH(ifa, ifp) { + if (ifa->ifa_addr->sa_family != AF_INET) + continue; + ia = ifatoia(ifa); + if (ia->ia4_flags & IN_IFF_NOTREADY) + continue; + if (ia->ia4_flags & IN_IFF_DETACHED && + (ifp->if_flags & IFF_LOOPBACK) == 0) + continue; + 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. + */ + ip->ip_dst.s_addr == ia->ia_subnet || + ip->ip_dst.s_addr == ia->ia_net) + goto matched; + /* + * An interface with IP address zero accepts + * all packets that arrive on that interface. + */ + if (in_nullhost(ia->ia_addr.sin_addr)) + goto matched; + } + ia = NULL; + +matched: + return ia; +} /* - * IP software interrupt routine + * IP software interrupt routine. */ -void -ipintr() +static void +ipintr(void *arg __unused) { - int s; struct mbuf *m; - while (1) { - s = splnet(); - IF_DEQUEUE(&ipintrq, m); - splx(s); - if (m == 0) - return; - MCLAIM(m, &ip_rx_mowner); + KASSERT(cpu_softintr_p()); + + SOFTNET_KERNEL_LOCK_UNLESS_NET_MPSAFE(); + while ((m = pktq_dequeue(ip_pktq)) != NULL) { ip_input(m); } + SOFTNET_KERNEL_UNLOCK_UNLESS_NET_MPSAFE(); } /* - * Ip input routine. Checksum and byte swap header. If fragmented + * IP input routine. Checksum and byte swap header. If fragmented * try to reassemble. Process options. Pass to next level. */ -void +static void ip_input(struct mbuf *m) { struct ip *ip = NULL; - struct ipq *fp; - struct in_ifaddr *ia; - struct ifaddr *ifa; - struct ipqent *ipqe; - int hlen = 0, mff, len; + struct in_ifaddr *ia = NULL; + int hlen = 0, len; int downmatch; - int checkif; + int srcrt = 0; + ifnet_t *ifp; + struct psref psref; + int s; + + KASSERTMSG(cpu_softintr_p(), "ip_input: not in the software " + "interrupt handler; synchronization assumptions violated"); MCLAIM(m, &ip_rx_mowner); -#ifdef DIAGNOSTIC - if ((m->m_flags & M_PKTHDR) == 0) - panic("ipintr no HDR"); -#endif -#ifdef IPSEC - /* - * should the inner packet be considered authentic? - * see comment in ah4_input(). - */ - if (m) { - m->m_flags &= ~M_AUTHIPHDR; - m->m_flags &= ~M_AUTHIPDGM; - } -#endif + KASSERT((m->m_flags & M_PKTHDR) != 0); + + ifp = m_get_rcvif_psref(m, &psref); + if (__predict_false(ifp == NULL)) + goto out; /* * If no IP addresses have been set yet but the interfaces * are receiving, can't do anything with incoming packets yet. + * Note: we pre-check without locks held. */ - if (TAILQ_FIRST(&in_ifaddr) == 0) - goto bad; - ipstat.ips_total++; + if (IN_ADDRLIST_READER_EMPTY()) + goto out; + IP_STATINC(IP_STAT_TOTAL); + /* * 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 (IP_HDR_ALIGNED_P(mtod(m, void *)) == 0) { if ((m = m_copyup(m, sizeof(struct ip), - (max_linkhdr + 3) & ~3)) == NULL) { + (max_linkhdr + 3) & ~3)) == NULL) { /* XXXJRT new stat, please */ - ipstat.ips_toosmall++; - return; + IP_STATINC(IP_STAT_TOOSMALL); + goto out; } - } else if (__predict_false(m->m_len < sizeof (struct ip))) { - if ((m = m_pullup(m, sizeof (struct ip))) == NULL) { - ipstat.ips_toosmall++; - return; + } else if (__predict_false(m->m_len < sizeof(struct ip))) { + if ((m = m_pullup(m, sizeof(struct ip))) == NULL) { + IP_STATINC(IP_STAT_TOOSMALL); + goto out; } } ip = mtod(m, struct ip *); if (ip->ip_v != IPVERSION) { - ipstat.ips_badvers++; - goto bad; + IP_STATINC(IP_STAT_BADVERS); + goto out; } hlen = ip->ip_hl << 2; if (hlen < sizeof(struct ip)) { /* minimum header length */ - ipstat.ips_badhlen++; - goto bad; + IP_STATINC(IP_STAT_BADHLEN); + goto out; } if (hlen > m->m_len) { - if ((m = m_pullup(m, hlen)) == 0) { - ipstat.ips_badhlen++; - return; + if ((m = m_pullup(m, hlen)) == NULL) { + IP_STATINC(IP_STAT_BADHLEN); + goto out; } ip = mtod(m, struct ip *); } @@ -498,25 +488,25 @@ ip_input(struct mbuf *m) * not allowed. */ if (IN_MULTICAST(ip->ip_src.s_addr)) { - ipstat.ips_badaddr++; - goto bad; + IP_STATINC(IP_STAT_BADADDR); + goto out; } /* 127/8 must not appear on wire - RFC1122 */ if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET || (ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) { - if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0) { - ipstat.ips_badaddr++; - goto bad; + if ((ifp->if_flags & IFF_LOOPBACK) == 0) { + IP_STATINC(IP_STAT_BADADDR); + goto out; } } switch (m->m_pkthdr.csum_flags & - ((m->m_pkthdr.rcvif->if_csum_flags_rx & M_CSUM_IPv4) | - M_CSUM_IPv4_BAD)) { + ((ifp->if_csum_flags_rx & M_CSUM_IPv4) | M_CSUM_IPv4_BAD)) { case M_CSUM_IPv4|M_CSUM_IPv4_BAD: INET_CSUM_COUNTER_INCR(&ip_hwcsum_bad); - goto badcsum; + IP_STATINC(IP_STAT_BADSUM); + goto out; case M_CSUM_IPv4: /* Checksum was okay. */ @@ -524,10 +514,18 @@ ip_input(struct mbuf *m) break; default: - /* Must compute it ourselves. */ - INET_CSUM_COUNTER_INCR(&ip_swcsum); - if (in_cksum(m, hlen) != 0) - goto bad; + /* + * Must compute it ourselves. Maybe skip checksum on + * loopback interfaces. + */ + if (__predict_true(!(ifp->if_flags & IFF_LOOPBACK) || + ip_do_loopback_cksum)) { + INET_CSUM_COUNTER_INCR(&ip_swcsum); + if (in_cksum(m, hlen) != 0) { + IP_STATINC(IP_STAT_BADSUM); + goto out; + } + } break; } @@ -538,19 +536,18 @@ ip_input(struct mbuf *m) * Check for additional length bogosity */ if (len < hlen) { - ipstat.ips_badlen++; - goto bad; + IP_STATINC(IP_STAT_BADLEN); + goto out; } /* - * Check that the amount of data in the buffers - * is as at least much as the IP header would have us expect. - * Trim mbufs if longer than we expect. - * Drop packet if shorter than we expect. + * Check that the amount of data in the buffers is at least as much + * as the IP header would have us expect. Trim mbufs if longer than + * we expect. Drop packet if shorter than we expect. */ if (m->m_pkthdr.len < len) { - ipstat.ips_tooshort++; - goto bad; + IP_STATINC(IP_STAT_TOOSHORT); + goto out; } if (m->m_pkthdr.len > len) { if (m->m_len == m->m_pkthdr.len) { @@ -560,50 +557,69 @@ ip_input(struct mbuf *m) m_adj(m, len - m->m_pkthdr.len); } -#ifdef IPSEC - /* ipflow (IP fast forwarding) is not compatible with IPsec. */ - m->m_flags &= ~M_CANFASTFWD; -#else /* * Assume that we can create a fast-forward IP flow entry * based on this packet. */ m->m_flags |= M_CANFASTFWD; -#endif -#ifdef PFIL_HOOKS /* * Run through list of hooks for input packets. If there are any * filters which require that additional packets in the flow are * not fast-forwarded, they must clear the M_CANFASTFWD flag. * Note that filters must _never_ set this flag, as another filter * in the list may have previously cleared it. + * + * Don't call hooks if the packet has already been processed by + * IPsec (encapsulated, tunnel mode). */ - /* - * let ipfilter look at packet on the wire, - * not the decapsulated packet. - */ -#ifdef IPSEC - if (!ipsec_getnhist(m)) +#if defined(IPSEC) + if (!ipsec_used || !ipsec_indone(m)) #else if (1) #endif { - if (pfil_run_hooks(&inet_pfil_hook, &m, m->m_pkthdr.rcvif, - PFIL_IN) != 0) - return; - if (m == NULL) - return; + struct in_addr odst = ip->ip_dst; + bool freed; + + freed = pfil_run_hooks(inet_pfil_hook, &m, ifp, PFIL_IN) != 0; + if (freed || m == NULL) { + m = NULL; + goto out; + } + KASSERT(m->m_len >= sizeof(struct ip)); ip = mtod(m, struct ip *); hlen = ip->ip_hl << 2; + KASSERT(m->m_len >= hlen); + + /* + * XXX The setting of "srcrt" here is to prevent ip_forward() + * from generating ICMP redirects for packets that have + * been redirected by a hook back out on to the same LAN that + * they came from and is not an indication that the packet + * is being influenced by source routing options. This + * allows things like + * "rdr tlp0 0/0 port 80 -> 1.1.1.200 3128 tcp" + * where tlp0 is both on the 1.1.1.0/24 network and is the + * default route for hosts on 1.1.1.0/24. Of course this + * also requires a "map tlp0 ..." to complete the story. + * One might argue whether or not this kind of network config. + * should be supported in this manner... + */ + srcrt = (odst.s_addr != ip->ip_dst.s_addr); } -#endif /* PFIL_HOOKS */ #ifdef ALTQ /* XXX Temporary until ALTQ is changed to use a pfil hook */ - if (altq_input != NULL && (*altq_input)(m, AF_INET) == 0) { - /* packet dropped by traffic conditioner */ - return; + if (altq_input) { + SOFTNET_LOCK(); + if ((*altq_input)(m, AF_INET) == 0) { + /* Packet dropped by traffic conditioner. */ + SOFTNET_UNLOCK(); + m = NULL; + goto out; + } + SOFTNET_UNLOCK(); } #endif @@ -613,85 +629,39 @@ ip_input(struct mbuf *m) * error was detected (causing an icmp message * to be sent and the original packet to be freed). */ - ip_nhops = 0; /* for source routed packets */ - if (hlen > sizeof (struct ip) && ip_dooptions(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); + if (hlen > sizeof(struct ip) && ip_dooptions(m)) { + m = NULL; + goto out; + } /* * Check our list of addresses, to see if the packet is for us. * * Traditional 4.4BSD did not consult IFF_UP at all. * The behavior here is to treat addresses on !IFF_UP interface - * as not mine. + * or IN_IFF_NOTREADY addresses as not mine. */ 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 - downmatch++; - } - } - if (ia != NULL) + s = pserialize_read_enter(); + ia = ip_match_our_address(ifp, ip, &downmatch); + if (ia != NULL) { + pserialize_read_exit(s); goto ours; - if (m->m_pkthdr.rcvif->if_flags & IFF_BROADCAST) { - TAILQ_FOREACH(ifa, &m->m_pkthdr.rcvif->if_addrlist, ifa_list) { - if (ifa->ifa_addr->sa_family != AF_INET) - continue; - ia = ifatoia(ifa); - 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. - */ - ip->ip_dst.s_addr == ia->ia_subnet || - 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 (in_nullhost(ia->ia_addr.sin_addr)) - goto ours; + } + + if (ifp->if_flags & IFF_BROADCAST) { + ia = ip_match_our_address_broadcast(ifp, ip); + if (ia != NULL) { + pserialize_read_exit(s); + goto ours; } } + pserialize_read_exit(s); + if (IN_MULTICAST(ip->ip_dst.s_addr)) { - struct in_multi *inm; #ifdef MROUTING extern struct socket *ip_mrouter; - if (M_READONLY(m)) { - if ((m = m_pullup(m, hlen)) == 0) { - ipstat.ips_toosmall++; - return; - } - ip = mtod(m, struct ip *); - } - if (ip_mrouter) { /* * If we are acting as a multicast router, all @@ -705,31 +675,32 @@ ip_input(struct mbuf *m) * as expected when ip_mforward() is called from * ip_output().) */ - if (ip_mforward(m, m->m_pkthdr.rcvif) != 0) { - ipstat.ips_cantforward++; - m_freem(m); - return; + SOFTNET_LOCK(); + if (ip_mforward(m, ifp) != 0) { + SOFTNET_UNLOCK(); + IP_STATINC(IP_STAT_CANTFORWARD); + goto out; } + SOFTNET_UNLOCK(); /* * The process-level routing demon needs to receive * all multicast IGMP packets, whether or not this * host belongs to their destination groups. */ - if (ip->ip_p == IPPROTO_IGMP) + if (ip->ip_p == IPPROTO_IGMP) { goto ours; - ipstat.ips_forward++; + } + IP_STATINC(IP_STAT_CANTFORWARD); } #endif /* * See if we belong to the destination multicast group on the * arrival interface. */ - IN_LOOKUP_MULTI(ip->ip_dst, m->m_pkthdr.rcvif, inm); - if (inm == NULL) { - ipstat.ips_cantforward++; - m_freem(m); - return; + if (!in_multi_group(ip->ip_dst, ifp, 0)) { + IP_STATINC(IP_STAT_CANTFORWARD); + goto out; } goto ours; } @@ -741,7 +712,8 @@ ip_input(struct mbuf *m) * Not for us; forward if possible and desirable. */ if (ipforwarding == 0) { - ipstat.ips_cantforward++; + m_put_rcvif_psref(ifp, &psref); + IP_STATINC(IP_STAT_CANTFORWARD); m_freem(m); } else { /* @@ -751,114 +723,64 @@ ip_input(struct mbuf *m) * forwarding loop till TTL goes to 0. */ if (downmatch) { + m_put_rcvif_psref(ifp, &psref); icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, 0, 0); - ipstat.ips_cantforward++; + IP_STATINC(IP_STAT_CANTFORWARD); return; } #ifdef IPSEC - if (ipsec4_in_reject(m, NULL)) { - ipsecstat.in_polvio++; - goto bad; + /* Check the security policy (SP) for the packet */ + if (ipsec_used) { + if (ipsec_ip_input(m, true) != 0) { + goto out; + } } #endif - - ip_forward(m, 0); + ip_forward(m, srcrt, ifp); + m_put_rcvif_psref(ifp, &psref); } return; ours: + m_put_rcvif_psref(ifp, &psref); + ifp = NULL; + /* * If offset or IP_MF are set, must reassemble. - * Otherwise, nothing need be done. - * (We could look in the reassembly queue to see - * if the packet was previously fragmented, - * but it's not worth the time; just let them time out.) */ 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. - */ - IPQ_LOCK(); - LIST_FOREACH(fp, &ipq, ipq_q) - if (ip->ip_id == fp->ipq_id && - 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; -found: - /* - * Adjust ip_len to not reflect header, - * set ipqe_mff if more fragments are expected, - * convert offset of this to bytes. + * Pass to IP reassembly mechanism. */ - 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 (ntohs(ip->ip_len) == 0 || - (ntohs(ip->ip_len) & 0x7) != 0) { - ipstat.ips_badfrags++; - IPQ_UNLOCK(); - goto bad; - } + if (ip_reass_packet(&m, ip) != 0) { + /* Failed; invalid fragment(s) or packet. */ + goto out; + } + if (m == NULL) { + /* More fragments should come; silently return. */ + goto out; } - 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. + * Reassembly is done, we have the final packet. + * Update cached data in local variable(s). */ - if (mff || ip->ip_off != htons(0)) { - ipstat.ips_fragments++; - ipqe = pool_get(&ipqent_pool, PR_NOWAIT); - if (ipqe == NULL) { - ipstat.ips_rcvmemdrop++; - IPQ_UNLOCK(); - goto bad; - } - ipqe->ipqe_mff = mff; - ipqe->ipqe_m = m; - ipqe->ipqe_ip = ip; - m = ip_reass(ipqe, fp); - if (m == 0) { - IPQ_UNLOCK(); - return; - } - ipstat.ips_reassembled++; - ip = mtod(m, struct ip *); - hlen = ip->ip_hl << 2; - ip->ip_len = htons(ntohs(ip->ip_len) + hlen); - } else - if (fp) - ip_freef(fp); - IPQ_UNLOCK(); + ip = mtod(m, struct ip *); + hlen = ip->ip_hl << 2; } + M_VERIFY_PACKET(m); + #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; + * 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 (ipsec_used && + (inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0) { + if (ipsec_ip_input(m, false) != 0) { + goto out; + } } #endif @@ -866,310 +788,85 @@ found: * Switch out to protocol's input routine. */ #if IFA_STATS - if (ia && ip) - ia->ia_ifa.ifa_data.ifad_inbytes += ntohs(ip->ip_len); -#endif - ipstat.ips_delivered++; - { - int off = hlen, nh = ip->ip_p; - - (*inetsw[ip_protox[nh]].pr_input)(m, off, nh); - return; - } -bad: - m_freem(m); - return; - -badcsum: - ipstat.ips_badsum++; - m_freem(m); -} - -/* - * Take incoming datagram fragment and try to - * reassemble it into whole datagram. If a chain for - * reassembly of this datagram already exists, then it - * is given as fp; otherwise have to make a chain. - */ -struct mbuf * -ip_reass(ipqe, fp) - struct ipqent *ipqe; - struct ipq *fp; -{ - 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; - int i, next; - - IPQ_LOCK_CHECK(); - - /* - * Presence of header sizes in mbufs - * would confuse code below. - */ - m->m_data += hlen; - m->m_len -= hlen; - - /* - * If first fragment to arrive, create a reassembly queue. - */ - if (fp == 0) { + if (ia) { + struct in_ifaddr *_ia; /* - * 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. + * Keep a reference from ip_match_our_address with psref + * is expensive, so explore ia here again. */ - 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) - goto dropfrag; - LIST_INSERT_HEAD(&ipq, fp, ipq_q); - fp->ipq_ttl = IPFRAGTTL; - fp->ipq_p = ipqe->ipqe_ip->ip_p; - fp->ipq_id = ipqe->ipqe_ip->ip_id; - TAILQ_INIT(&fp->ipq_fragq); - fp->ipq_src = ipqe->ipqe_ip->ip_src; - fp->ipq_dst = ipqe->ipqe_ip->ip_dst; - p = NULL; - goto insert; - } - - /* - * Find a segment which begins after this one does. - */ - 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; - - /* - * If there is a preceding segment, it may provide some of - * our data already. If so, drop the data from the incoming - * segment. If it provides all of our data, drop us. - */ - if (p != NULL) { - i = ntohs(p->ipqe_ip->ip_off) + ntohs(p->ipqe_ip->ip_len) - - ntohs(ipqe->ipqe_ip->ip_off); - if (i > 0) { - if (i >= ntohs(ipqe->ipqe_ip->ip_len)) - goto dropfrag; - m_adj(ipqe->ipqe_m, 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); - } + s = pserialize_read_enter(); + _ia = in_get_ia(ip->ip_dst); + _ia->ia_ifa.ifa_data.ifad_inbytes += ntohs(ip->ip_len); + pserialize_read_exit(s); } +#endif + IP_STATINC(IP_STAT_DELIVERED); - /* - * While we overlap succeeding segments trim them or, - * if they are completely covered, dequeue them. - */ - 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 = TAILQ_NEXT(q, ipqe_q); - m_freem(q->ipqe_m); - TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q); - pool_put(&ipqent_pool, q); - } + const int off = hlen, nh = ip->ip_p; -insert: - /* - * Stick new segment in its place; - * check for complete reassembly. - */ - if (p == NULL) { - TAILQ_INSERT_HEAD(&fp->ipq_fragq, ipqe, ipqe_q); - } else { - TAILQ_INSERT_AFTER(&fp->ipq_fragq, p, ipqe, ipqe_q); - } - next = 0; - 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 += ntohs(q->ipqe_ip->ip_len); - } - if (p->ipqe_mff) - return (0); - - /* - * Reassembly is complete. Check for a bogus message size and - * concatenate fragments. - */ - q = TAILQ_FIRST(&fp->ipq_fragq); - ip = 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 = TAILQ_NEXT(q, ipqe_q); - pool_put(&ipqent_pool, q); - for (q = nq; q != NULL; q = nq) { - t = q->ipqe_m; - nq = TAILQ_NEXT(q, ipqe_q); - pool_put(&ipqent_pool, q); - m_cat(m, t); - } + (*inetsw[ip_protox[nh]].pr_input)(m, off, nh); + return; - /* - * Create header for new ip packet by - * modifying header of first packet; - * dequeue and discard fragment reassembly header. - * Make header visible. - */ - ip->ip_len = htons(next); - ip->ip_src = fp->ipq_src; - 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 */ - int plen = 0; - for (t = m; t; t = t->m_next) - plen += t->m_len; - m->m_pkthdr.len = plen; - } - return (m); - -dropfrag: - ipstat.ips_fragdropped++; - m_freem(m); - pool_put(&ipqent_pool, ipqe); - return (0); +out: + m_put_rcvif_psref(ifp, &psref); + if (m != NULL) + m_freem(m); } /* - * Free a fragment reassembly header and all - * associated datagrams. + * IP timer processing. */ void -ip_freef(fp) - struct ipq *fp; +ip_slowtimo(void) { - struct ipqent *q, *p; - IPQ_LOCK_CHECK(); + SOFTNET_KERNEL_LOCK_UNLESS_NET_MPSAFE(); - for (q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL; q = p) { - p = TAILQ_NEXT(q, ipqe_q); - m_freem(q->ipqe_m); - TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q); - pool_put(&ipqent_pool, q); - } - LIST_REMOVE(fp, ipq_q); - FREE(fp, M_FTABLE); - ip_nfragpackets--; -} + ip_reass_slowtimo(); -/* - * IP timer processing; - * if a timer expires on a reassembly - * queue, discard it. - */ -void -ip_slowtimo() -{ - struct ipq *fp, *nfp; - int s = splsoftnet(); - - IPQ_LOCK(); - for (fp = LIST_FIRST(&ipq); fp != NULL; fp = nfp) { - nfp = LIST_NEXT(fp, ipq_q); - if (--fp->ipq_ttl == 0) { - ipstat.ips_fragtimeout++; - 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 && LIST_FIRST(&ipq)) - ip_freef(LIST_FIRST(&ipq)); - } - IPQ_UNLOCK(); -#ifdef GATEWAY - ipflow_slowtimo(); -#endif - splx(s); + SOFTNET_KERNEL_UNLOCK_UNLESS_NET_MPSAFE(); } /* - * Drain off all datagram fragments. + * IP drain processing. */ void -ip_drain() +ip_drain(void) { - /* - * We may be called from a device's interrupt context. If - * the ipq is already busy, just bail out now. - */ - if (ipq_lock_try() == 0) - return; - - while (LIST_FIRST(&ipq) != NULL) { - ipstat.ips_fragdropped++; - ip_freef(LIST_FIRST(&ipq)); - } - - IPQ_UNLOCK(); + KERNEL_LOCK(1, NULL); + ip_reass_drain(); + KERNEL_UNLOCK_ONE(NULL); } /* - * Do option processing on a datagram, - * possibly discarding it if bad options are encountered, - * or forwarding it if source-routed. - * Returns 1 if packet has been forwarded/freed, - * 0 if the packet should be processed further. + * ip_dooptions: perform option processing on a datagram, possibly discarding + * it if bad options are encountered, or forwarding it if source-routed. + * + * => Returns true if packet has been forwarded/freed. + * => Returns false if the packet should be processed further. */ -int -ip_dooptions(m) - struct mbuf *m; +static bool +ip_dooptions(struct mbuf *m) { 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; + int srr_present, rr_present, ts_present; struct in_addr dst; n_time ntime; + struct ifaddr *ifa = NULL; + int s; + + srr_present = 0; + rr_present = 0; + ts_present = 0; dst = ip->ip_dst; cp = (u_char *)(ip + 1); - cnt = (ip->ip_hl << 2) - sizeof (struct ip); + cnt = (ip->ip_hl << 2) - sizeof(struct ip); for (; cnt > 0; cnt -= optlen, cp += optlen) { opt = cp[IPOPT_OPTVAL]; if (opt == IPOPT_EOL) @@ -1202,12 +899,22 @@ ip_dooptions(m) * address is on directly accessible net. */ case IPOPT_LSRR: - case IPOPT_SSRR: + case IPOPT_SSRR: { + struct psref psref; + struct sockaddr_in ipaddr = { + .sin_len = sizeof(ipaddr), + .sin_family = AF_INET, + }; + if (ip_allowsrcrt == 0) { type = ICMP_UNREACH; code = ICMP_UNREACH_NET_PROHIB; goto bad; } + if (srr_present++) { + code = &cp[IPOPT_OPTVAL] - (u_char *)ip; + goto bad; + } if (optlen < IPOPT_OFFSET + sizeof(*cp)) { code = &cp[IPOPT_OLEN] - (u_char *)ip; goto bad; @@ -1217,8 +924,11 @@ ip_dooptions(m) goto bad; } ipaddr.sin_addr = ip->ip_dst; - ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr))); - if (ia == 0) { + + s = pserialize_read_enter(); + ifa = ifa_ifwithaddr(sintosa(&ipaddr)); + if (ifa == NULL) { + pserialize_read_exit(s); if (opt == IPOPT_SSRR) { type = ICMP_UNREACH; code = ICMP_UNREACH_SRCFAIL; @@ -1230,39 +940,59 @@ ip_dooptions(m) */ break; } + pserialize_read_exit(s); + off--; /* 0 origin */ if ((off + sizeof(struct in_addr)) > optlen) { /* * End of source route. Should be for us. */ - save_rte(cp, ip->ip_src); + save_rte(m, cp, ip->ip_src); break; } /* * locate outgoing interface */ - bcopy((caddr_t)(cp + off), (caddr_t)&ipaddr.sin_addr, + memcpy((void *)&ipaddr.sin_addr, (void *)(cp + off), sizeof(ipaddr.sin_addr)); - if (opt == IPOPT_SSRR) - ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr))); - else - ia = ip_rtaddr(ipaddr.sin_addr); - if (ia == 0) { + if (opt == IPOPT_SSRR) { + ifa = ifa_ifwithladdr_psref(sintosa(&ipaddr), + &psref); + if (ifa != NULL) + ia = ifatoia(ifa); + else + ia = NULL; + } else { + ia = ip_rtaddr(ipaddr.sin_addr, &psref); + } + if (ia == NULL) { type = ICMP_UNREACH; code = ICMP_UNREACH_SRCFAIL; goto bad; } ip->ip_dst = ipaddr.sin_addr; - bcopy((caddr_t)&ia->ia_addr.sin_addr, - (caddr_t)(cp + off), sizeof(struct in_addr)); + memcpy(cp + off, &ia->ia_addr.sin_addr, + sizeof(struct in_addr)); + ia4_release(ia, &psref); cp[IPOPT_OFFSET] += sizeof(struct in_addr); /* * Let ip_intr's mcast routing check handle mcast pkts */ forward = !IN_MULTICAST(ip->ip_dst.s_addr); break; + } + + case IPOPT_RR: { + struct psref psref; + struct sockaddr_in ipaddr = { + .sin_len = sizeof(ipaddr), + .sin_family = AF_INET, + }; - case IPOPT_RR: + if (rr_present++) { + code = &cp[IPOPT_OPTVAL] - (u_char *)ip; + goto bad; + } if (optlen < IPOPT_OFFSET + sizeof(*cp)) { code = &cp[IPOPT_OLEN] - (u_char *)ip; goto bad; @@ -1277,27 +1007,37 @@ ip_dooptions(m) off--; /* 0 origin */ if ((off + sizeof(struct in_addr)) > optlen) break; - bcopy((caddr_t)(&ip->ip_dst), (caddr_t)&ipaddr.sin_addr, + memcpy((void *)&ipaddr.sin_addr, (void *)&ip->ip_dst, sizeof(ipaddr.sin_addr)); /* * locate outgoing interface; if we're the destination, * use the incoming interface (should be same). */ - if ((ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr)))) - == NULL && - (ia = ip_rtaddr(ipaddr.sin_addr)) == NULL) { - type = ICMP_UNREACH; - code = ICMP_UNREACH_HOST; - goto bad; + ifa = ifa_ifwithaddr_psref(sintosa(&ipaddr), &psref); + if (ifa == NULL) { + ia = ip_rtaddr(ipaddr.sin_addr, &psref); + if (ia == NULL) { + type = ICMP_UNREACH; + code = ICMP_UNREACH_HOST; + goto bad; + } + } else { + ia = ifatoia(ifa); } - bcopy((caddr_t)&ia->ia_addr.sin_addr, - (caddr_t)(cp + off), sizeof(struct in_addr)); + memcpy(cp + off, &ia->ia_addr.sin_addr, + sizeof(struct in_addr)); + ia4_release(ia, &psref); cp[IPOPT_OFFSET] += sizeof(struct in_addr); break; + } case IPOPT_TS: code = cp - (u_char *)ip; ipt = (struct ip_timestamp *)cp; + if (ts_present++) { + code = &cp[IPOPT_OPTVAL] - (u_char *)ip; + goto bad; + } if (ipt->ipt_len < 4 || ipt->ipt_len > 40) { code = (u_char *)&ipt->ipt_len - (u_char *)ip; goto bad; @@ -1306,7 +1046,7 @@ ip_dooptions(m) code = (u_char *)&ipt->ipt_ptr - (u_char *)ip; goto bad; } - if (ipt->ipt_ptr > ipt->ipt_len - sizeof (int32_t)) { + if (ipt->ipt_ptr > ipt->ipt_len - sizeof(int32_t)) { if (++ipt->ipt_oflw == 0) { code = (u_char *)&ipt->ipt_ptr - (u_char *)ip; @@ -1320,7 +1060,14 @@ ip_dooptions(m) case IPOPT_TS_TSONLY: break; - case IPOPT_TS_TSANDADDR: + case IPOPT_TS_TSANDADDR: { + struct ifnet *rcvif; + int _s, _ss; + struct sockaddr_in ipaddr = { + .sin_len = sizeof(ipaddr), + .sin_family = AF_INET, + }; + if (ipt->ipt_ptr - 1 + sizeof(n_time) + sizeof(struct in_addr) > ipt->ipt_len) { code = (u_char *)&ipt->ipt_ptr - @@ -1328,29 +1075,49 @@ ip_dooptions(m) goto bad; } ipaddr.sin_addr = dst; - ia = ifatoia(ifaof_ifpforaddr(sintosa(&ipaddr), - m->m_pkthdr.rcvif)); - if (ia == 0) - continue; - bcopy(&ia->ia_addr.sin_addr, - cp0, sizeof(struct in_addr)); + _ss = pserialize_read_enter(); + rcvif = m_get_rcvif(m, &_s); + if (__predict_true(rcvif != NULL)) { + ifa = ifaof_ifpforaddr(sintosa(&ipaddr), + rcvif); + } + m_put_rcvif(rcvif, &_s); + if (ifa == NULL) { + pserialize_read_exit(_ss); + break; + } + ia = ifatoia(ifa); + memcpy(cp0, &ia->ia_addr.sin_addr, + sizeof(struct in_addr)); + pserialize_read_exit(_ss); ipt->ipt_ptr += sizeof(struct in_addr); break; + } + + case IPOPT_TS_PRESPEC: { + struct sockaddr_in ipaddr = { + .sin_len = sizeof(ipaddr), + .sin_family = AF_INET, + }; - case IPOPT_TS_PRESPEC: if (ipt->ipt_ptr - 1 + sizeof(n_time) + sizeof(struct in_addr) > ipt->ipt_len) { code = (u_char *)&ipt->ipt_ptr - (u_char *)ip; goto bad; } - bcopy(cp0, &ipaddr.sin_addr, + memcpy(&ipaddr.sin_addr, cp0, sizeof(struct in_addr)); - if (ifatoia(ifa_ifwithaddr(sintosa(&ipaddr))) - == NULL) + s = pserialize_read_enter(); + ifa = ifa_ifwithaddr(sintosa(&ipaddr)); + if (ifa == NULL) { + pserialize_read_exit(s); continue; + } + pserialize_read_exit(s); ipt->ipt_ptr += sizeof(struct in_addr); break; + } default: /* XXX can't take &ipt->ipt_flg */ @@ -1360,76 +1127,92 @@ ip_dooptions(m) } ntime = iptime(); cp0 = (u_char *) &ntime; /* XXX grumble, GCC... */ - bcopy(cp0, (caddr_t)cp + ipt->ipt_ptr - 1, + memmove((char *)cp + ipt->ipt_ptr - 1, cp0, sizeof(n_time)); ipt->ipt_ptr += sizeof(n_time); } } if (forward) { + struct ifnet *rcvif; + struct psref _psref; + if (ip_forwsrcrt == 0) { type = ICMP_UNREACH; code = ICMP_UNREACH_SRCFAIL; goto bad; } - ip_forward(m, 1); - return (1); + + rcvif = m_get_rcvif_psref(m, &_psref); + if (__predict_false(rcvif == NULL)) { + type = ICMP_UNREACH; + code = ICMP_UNREACH_HOST; + goto bad; + } + ip_forward(m, 1, rcvif); + m_put_rcvif_psref(rcvif, &_psref); + return true; } - return (0); + return false; bad: icmp_error(m, type, code, 0, 0); - ipstat.ips_badoptions++; - return (1); + IP_STATINC(IP_STAT_BADOPTIONS); + return true; } /* - * Given address of next destination (final or next hop), + * ip_rtaddr: given address of next destination (final or next hop), * return internet address info of interface to be used to get there. */ -struct in_ifaddr * -ip_rtaddr(dst) - struct in_addr dst; +static struct in_ifaddr * +ip_rtaddr(struct in_addr dst, struct psref *psref) { - struct sockaddr_in *sin; - - sin = satosin(&ipforward_rt.ro_dst); - - 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; - } - sin->sin_family = AF_INET; - sin->sin_len = sizeof(*sin); - sin->sin_addr = dst; + struct rtentry *rt; + union { + struct sockaddr dst; + struct sockaddr_in dst4; + } u; + struct route *ro; + + sockaddr_in_init(&u.dst4, &dst, 0); + + ro = percpu_getref(ipforward_rt_percpu); + rt = rtcache_lookup(ro, &u.dst); + if (rt == NULL) { + percpu_putref(ipforward_rt_percpu); + return NULL; + } + + ia4_acquire(ifatoia(rt->rt_ifa), psref); + rtcache_unref(rt, ro); + percpu_putref(ipforward_rt_percpu); - rtalloc(&ipforward_rt); - } - if (ipforward_rt.ro_rt == 0) - return ((struct in_ifaddr *)0); - return (ifatoia(ipforward_rt.ro_rt->rt_ifa)); + return ifatoia(rt->rt_ifa); } /* - * Save incoming source route for use in replies, - * to be picked up later by ip_srcroute if the receiver is interested. + * save_rte: save incoming source route for use in replies, to be picked + * up later by ip_srcroute if the receiver is interested. */ -void -save_rte(option, dst) - u_char *option; - struct in_addr dst; +static void +save_rte(struct mbuf *m, u_char *option, struct in_addr dst) { + struct ip_srcrt *isr; + struct m_tag *mtag; unsigned olen; olen = option[IPOPT_OLEN]; -#ifdef DIAGNOSTIC - if (ipprintfs) - printf("save_rte: olen %d\n", olen); -#endif /* 0 */ - if (olen > sizeof(ip_srcrt) - (1 + sizeof(dst))) + if (olen > sizeof(isr->isr_hdr) + sizeof(isr->isr_routes)) return; - bcopy((caddr_t)option, (caddr_t)ip_srcrt.srcopt, olen); - ip_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr); - ip_srcrt.dst = dst; + + mtag = m_tag_get(PACKET_TAG_SRCROUTE, sizeof(*isr), M_NOWAIT); + if (mtag == NULL) + return; + isr = (struct ip_srcrt *)(mtag + 1); + + memcpy(isr->isr_hdr, option, olen); + isr->isr_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr); + isr->isr_dst = dst; + m_tag_prepend(m, mtag); } /* @@ -1438,106 +1221,88 @@ save_rte(option, dst) * The first hop is placed before the options, will be removed later. */ struct mbuf * -ip_srcroute() +ip_srcroute(struct mbuf *m0) { struct in_addr *p, *q; struct mbuf *m; + struct ip_srcrt *isr; + struct m_tag *mtag; + + mtag = m_tag_find(m0, PACKET_TAG_SRCROUTE, NULL); + if (mtag == NULL) + return NULL; + isr = (struct ip_srcrt *)(mtag + 1); + + if (isr->isr_nhops == 0) + return NULL; - if (ip_nhops == 0) - return ((struct mbuf *)0); m = m_get(M_DONTWAIT, MT_SOOPTS); - if (m == 0) - return ((struct mbuf *)0); + if (m == NULL) + return NULL; MCLAIM(m, &inetdomain.dom_mowner); -#define OPTSIZ (sizeof(ip_srcrt.nop) + sizeof(ip_srcrt.srcopt)) +#define OPTSIZ (sizeof(isr->isr_nop) + sizeof(isr->isr_hdr)) - /* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */ - m->m_len = ip_nhops * sizeof(struct in_addr) + sizeof(struct in_addr) + - OPTSIZ; -#ifdef DIAGNOSTIC - if (ipprintfs) - printf("ip_srcroute: nhops %d mlen %d", ip_nhops, m->m_len); -#endif + /* length is (nhops+1)*sizeof(addr) + sizeof(nop + header) */ + m->m_len = (isr->isr_nhops + 1) * sizeof(struct in_addr) + OPTSIZ; /* * First save first hop for return route */ - p = &ip_srcrt.route[ip_nhops - 1]; + p = &(isr->isr_routes[isr->isr_nhops - 1]); *(mtod(m, struct in_addr *)) = *p--; -#ifdef DIAGNOSTIC - if (ipprintfs) - printf(" hops %x", ntohl(mtod(m, struct in_addr *)->s_addr)); -#endif /* * Copy option fields and padding (nop) to mbuf. */ - ip_srcrt.nop = IPOPT_NOP; - ip_srcrt.srcopt[IPOPT_OFFSET] = IPOPT_MINOFF; - bcopy((caddr_t)&ip_srcrt.nop, - mtod(m, caddr_t) + sizeof(struct in_addr), OPTSIZ); - q = (struct in_addr *)(mtod(m, caddr_t) + + isr->isr_nop = IPOPT_NOP; + isr->isr_hdr[IPOPT_OFFSET] = IPOPT_MINOFF; + memmove(mtod(m, char *) + sizeof(struct in_addr), &isr->isr_nop, + OPTSIZ); + q = (struct in_addr *)(mtod(m, char *) + sizeof(struct in_addr) + OPTSIZ); #undef OPTSIZ /* * Record return path as an IP source route, * reversing the path (pointers are now aligned). */ - while (p >= ip_srcrt.route) { -#ifdef DIAGNOSTIC - if (ipprintfs) - printf(" %x", ntohl(q->s_addr)); -#endif + while (p >= isr->isr_routes) { *q++ = *p--; } /* * Last hop goes to final destination. */ - *q = ip_srcrt.dst; -#ifdef DIAGNOSTIC - if (ipprintfs) - printf(" %x\n", ntohl(q->s_addr)); -#endif - return (m); + *q = isr->isr_dst; + m_tag_delete(m0, mtag); + return m; } -/* - * Strip out IP options, at higher - * level protocol in the kernel. - * Second argument is buffer to which options - * will be moved, and return value is their length. - * XXX should be deleted; last arg currently ignored. - */ +const int inetctlerrmap[PRC_NCMDS] = { + [PRC_MSGSIZE] = EMSGSIZE, + [PRC_HOSTDEAD] = EHOSTDOWN, + [PRC_HOSTUNREACH] = EHOSTUNREACH, + [PRC_UNREACH_NET] = EHOSTUNREACH, + [PRC_UNREACH_HOST] = EHOSTUNREACH, + [PRC_UNREACH_PROTOCOL] = ECONNREFUSED, + [PRC_UNREACH_PORT] = ECONNREFUSED, + [PRC_UNREACH_SRCFAIL] = EHOSTUNREACH, + [PRC_PARAMPROB] = ENOPROTOOPT, +}; + void -ip_stripoptions(m, mopt) - struct mbuf *m; - struct mbuf *mopt; +ip_fasttimo(void) { - int i; - struct ip *ip = mtod(m, struct ip *); - caddr_t opts; - int olen; - - olen = (ip->ip_hl << 2) - sizeof (struct ip); - opts = (caddr_t)(ip + 1); - i = m->m_len - (sizeof (struct ip) + olen); - bcopy(opts + olen, opts, (unsigned)i); - m->m_len -= olen; - if (m->m_flags & M_PKTHDR) - m->m_pkthdr.len -= olen; - ip->ip_len = htons(ntohs(ip->ip_len) - olen); - ip->ip_hl = sizeof (struct ip) >> 2; + if (ip_drainwanted) { + ip_drain(); + ip_drainwanted = 0; + } } -const int inetctlerrmap[PRC_NCMDS] = { - 0, 0, 0, 0, - 0, EMSGSIZE, EHOSTDOWN, EHOSTUNREACH, - EHOSTUNREACH, EHOSTUNREACH, ECONNREFUSED, ECONNREFUSED, - EMSGSIZE, EHOSTUNREACH, 0, 0, - 0, 0, 0, 0, - ENOPROTOOPT -}; +void +ip_drainstub(void) +{ + ip_drainwanted = 1; +} /* * Forward a packet. If some error occurs return the sender @@ -1553,21 +1318,23 @@ const int inetctlerrmap[PRC_NCMDS] = { * The srcrt parameter indicates whether the packet is being forwarded * via a source route. */ -void -ip_forward(m, srcrt) - struct mbuf *m; - int srcrt; +static void +ip_forward(struct mbuf *m, int srcrt, struct ifnet *rcvif) { struct ip *ip = mtod(m, struct ip *); - struct sockaddr_in *sin; struct rtentry *rt; - int error, type = 0, code = 0; + int error, type = 0, code = 0, destmtu = 0; struct mbuf *mcopy; n_long dest; - struct ifnet *destifp; -#ifdef IPSEC - struct ifnet dummyifp; -#endif + union { + struct sockaddr dst; + struct sockaddr_in dst4; + } u; + uint64_t *ips; + struct route *ro; + + KASSERTMSG(cpu_softintr_p(), "ip_forward: not in the software " + "interrupt handler; synchronization assumptions violated"); /* * We are now in the output path. @@ -1580,40 +1347,25 @@ ip_forward(m, srcrt) m->m_pkthdr.csum_flags = 0; dest = 0; -#ifdef DIAGNOSTIC - if (ipprintfs) - printf("forward: src %2.2x dst %2.2x ttl %x\n", - ntohl(ip->ip_src.s_addr), - ntohl(ip->ip_dst.s_addr), ip->ip_ttl); -#endif if (m->m_flags & (M_BCAST|M_MCAST) || in_canforward(ip->ip_dst) == 0) { - ipstat.ips_cantforward++; + IP_STATINC(IP_STAT_CANTFORWARD); m_freem(m); return; } + if (ip->ip_ttl <= IPTTLDEC) { icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0); return; } - ip->ip_ttl -= IPTTLDEC; - sin = satosin(&ipforward_rt.ro_dst); - if ((rt = ipforward_rt.ro_rt) == 0 || - !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(struct sockaddr_in); - sin->sin_addr = ip->ip_dst; - - rtalloc(&ipforward_rt); - if (ipforward_rt.ro_rt == 0) { - icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, dest, 0); - return; - } - rt = ipforward_rt.ro_rt; + sockaddr_in_init(&u.dst4, &ip->ip_dst, 0); + + ro = percpu_getref(ipforward_rt_percpu); + rt = rtcache_lookup(ro, &u.dst); + if (rt == NULL) { + percpu_putref(ipforward_rt_percpu); + icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_NET, dest, 0); + return; } /* @@ -1625,6 +1377,8 @@ ip_forward(m, srcrt) if (mcopy) mcopy = m_pullup(mcopy, ip->ip_hl << 2); + ip->ip_ttl -= IPTTLDEC; + /* * If forwarding packet using same interface that it came in on, * perhaps should send a redirect to sender to shortcut a hop. @@ -1633,9 +1387,9 @@ ip_forward(m, srcrt) * Also, don't send redirect if forwarding using a default route * or a route modified by a redirect. */ - if (rt->rt_ifp == m->m_pkthdr.rcvif && + if (rt->rt_ifp == rcvif && (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 && - !in_nullhost(satosin(rt_key(rt))->sin_addr) && + !in_nullhost(satocsin(rt_getkey(rt))->sin_addr) && ipsendredirects && !srcrt) { if (rt->rt_ifa && (ip->ip_src.s_addr & ifatoia(rt->rt_ifa)->ia_subnetmask) == @@ -1650,40 +1404,46 @@ ip_forward(m, srcrt) */ type = ICMP_REDIRECT; code = ICMP_REDIRECT_HOST; -#ifdef DIAGNOSTIC - if (ipprintfs) - printf("redirect (%d) to %x\n", code, - (u_int32_t)dest); -#endif } } + rtcache_unref(rt, ro); -#ifdef IPSEC - /* Don't lookup socket in forwarding case */ - (void)ipsec_setsocket(m, NULL); -#endif - error = ip_output(m, (struct mbuf *)0, &ipforward_rt, - (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)), 0); - if (error) - ipstat.ips_cantforward++; - else { - ipstat.ips_forward++; - if (type) - ipstat.ips_redirectsent++; - else { - if (mcopy) { + error = ip_output(m, NULL, ro, + (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)), + NULL, NULL); + + if (error) { + IP_STATINC(IP_STAT_CANTFORWARD); + goto error; + } + + ips = IP_STAT_GETREF(); + ips[IP_STAT_FORWARD]++; + + if (type) { + ips[IP_STAT_REDIRECTSENT]++; + IP_STAT_PUTREF(); + goto redirect; + } + + IP_STAT_PUTREF(); + if (mcopy) { #ifdef GATEWAY - if (mcopy->m_flags & M_CANFASTFWD) - ipflow_create(&ipforward_rt, mcopy); + if (mcopy->m_flags & M_CANFASTFWD) + ipflow_create(ro, mcopy); #endif - m_freem(mcopy); - } - return; - } + m_freem(mcopy); } - if (mcopy == NULL) + + percpu_putref(ipforward_rt_percpu); + return; + +redirect: +error: + if (mcopy == NULL) { + percpu_putref(ipforward_rt_percpu); return; - destifp = NULL; + } switch (error) { @@ -1703,298 +1463,339 @@ ip_forward(m, srcrt) case EMSGSIZE: type = ICMP_UNREACH; code = ICMP_UNREACH_NEEDFRAG; -#ifndef IPSEC - if (ipforward_rt.ro_rt) - destifp = ipforward_rt.ro_rt->rt_ifp; -#else - /* - * If the packet is routed over IPsec tunnel, tell the - * originator the tunnel MTU. - * tunnel MTU = if MTU - sizeof(IP) - ESP/AH hdrsiz - * XXX quickhack!!! - */ - if (ipforward_rt.ro_rt) { - struct secpolicy *sp; - int ipsecerror; - size_t ipsechdr; - struct route *ro; - - sp = ipsec4_getpolicybyaddr(mcopy, - IPSEC_DIR_OUTBOUND, - IP_FORWARDING, - &ipsecerror); - - if (sp == NULL) - destifp = ipforward_rt.ro_rt->rt_ifp; - else { - /* count IPsec header size */ - ipsechdr = ipsec4_hdrsiz(mcopy, - IPSEC_DIR_OUTBOUND, - NULL); - /* - * find the correct route for outer IPv4 - * header, compute tunnel MTU. - * - * XXX BUG ALERT - * The "dummyifp" code relies upon the fact - * that icmp_error() touches only ifp->if_mtu. - */ - /*XXX*/ - destifp = NULL; - if (sp->req != NULL - && sp->req->sav != NULL - && sp->req->sav->sah != NULL) { - 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; - } - } - - key_freesp(sp); - } + if ((rt = rtcache_validate(ro)) != NULL) { + destmtu = rt->rt_ifp->if_mtu; + rtcache_unref(rt, ro); } -#endif /*IPSEC*/ - ipstat.ips_cantfrag++; +#ifdef IPSEC + if (ipsec_used) + ipsec_mtu(mcopy, &destmtu); +#endif + IP_STATINC(IP_STAT_CANTFRAG); break; case ENOBUFS: -#if 1 /* - * 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 if the underlying interface is rate-limited. + * Do not generate ICMP_SOURCEQUENCH as required in RFC 1812, + * Requirements for IP Version 4 Routers. Source quench can + * be a big problem under DoS attacks or if the underlying + * interface is rate-limited. */ if (mcopy) m_freem(mcopy); + percpu_putref(ipforward_rt_percpu); return; -#else - type = ICMP_SOURCEQUENCH; - code = 0; - break; -#endif } - icmp_error(mcopy, type, code, dest, destifp); + icmp_error(mcopy, type, code, dest, destmtu); + percpu_putref(ipforward_rt_percpu); } void -ip_savecontrol(inp, mp, ip, m) - struct inpcb *inp; - struct mbuf **mp; - struct ip *ip; - struct mbuf *m; +ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip, + struct mbuf *m) { + struct socket *so = inp->inp_socket; + int inpflags = inp->inp_flags; - if (inp->inp_socket->so_options & SO_TIMESTAMP) { - struct timeval tv; + if (SOOPT_TIMESTAMP(so->so_options)) + mp = sbsavetimestamp(so->so_options, mp); - 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, + if (inpflags & INP_RECVDSTADDR) { + *mp = sbcreatecontrol(&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 (inpflags & INP_RECVTTL) { + *mp = sbcreatecontrol(&ip->ip_ttl, + sizeof(uint8_t), IP_RECVTTL, 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); + + struct psref psref; + ifnet_t *ifp = m_get_rcvif_psref(m, &psref); + if (__predict_false(ifp == NULL)) { +#ifdef DIAGNOSTIC + printf("%s: missing receive interface\n", __func__); +#endif + return; /* XXX should report error? */ + } + + if (inpflags & INP_RECVPKTINFO) { + struct in_pktinfo ipi; + ipi.ipi_addr = ip->ip_dst; + ipi.ipi_ifindex = ifp->if_index; + *mp = sbcreatecontrol(&ipi, + sizeof(ipi), IP_PKTINFO, IPPROTO_IP); if (*mp) mp = &(*mp)->m_next; } -#endif - if (inp->inp_flags & INP_RECVIF) { + if (inpflags & 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); + sockaddr_dl_init(&sdl, sizeof(sdl), ifp->if_index, 0, NULL, 0, + NULL, 0); + *mp = sbcreatecontrol(&sdl, sdl.sdl_len, IP_RECVIF, IPPROTO_IP); if (*mp) mp = &(*mp)->m_next; } + m_put_rcvif_psref(ifp, &psref); } -int -ip_sysctl(name, namelen, oldp, oldlenp, newp, newlen) - int *name; - u_int namelen; - void *oldp; - size_t *oldlenp; - void *newp; - size_t newlen; +/* + * sysctl helper routine for net.inet.ip.forwsrcrt. + */ +static int +sysctl_net_inet_ip_forwsrcrt(SYSCTLFN_ARGS) { - extern int subnetsarelocal, hostzeroisbroadcast; - - int error, old; + int error, tmp; + struct sysctlnode node; - /* All sysctl names at this level are terminal. */ - if (namelen != 1) - return (ENOTDIR); - - switch (name[0]) { - case IPCTL_FORWARDING: - return (sysctl_int(oldp, oldlenp, newp, newlen, &ipforwarding)); - case IPCTL_SENDREDIRECTS: - return (sysctl_int(oldp, oldlenp, newp, newlen, - &ipsendredirects)); - case IPCTL_DEFTTL: - return (sysctl_int(oldp, oldlenp, newp, newlen, &ip_defttl)); -#ifdef notyet - 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)); - case IPCTL_SUBNETSARELOCAL: - return (sysctl_int(oldp, oldlenp, newp, newlen, - &subnetsarelocal)); - case IPCTL_MTUDISC: - error = sysctl_int(oldp, oldlenp, newp, newlen, - &ip_mtudisc); - if (error == 0 && ip_mtudisc == 0) - rt_timer_queue_remove_all(ip_mtudisc_timeout_q, TRUE); - return error; - case IPCTL_ANONPORTMIN: - old = anonportmin; - error = sysctl_int(oldp, oldlenp, newp, newlen, &anonportmin); - if (anonportmin >= anonportmax || anonportmin < 0 - || anonportmin > 65535 -#ifndef IPNOPRIVPORTS - || anonportmin < IPPORT_RESERVED -#endif - ) { - anonportmin = old; - return (EINVAL); - } + node = *rnode; + tmp = ip_forwsrcrt; + node.sysctl_data = &tmp; + error = sysctl_lookup(SYSCTLFN_CALL(&node)); + if (error || newp == NULL) return (error); - case IPCTL_ANONPORTMAX: - old = anonportmax; - error = sysctl_int(oldp, oldlenp, newp, newlen, &anonportmax); - if (anonportmin >= anonportmax || anonportmax < 0 - || anonportmax > 65535 -#ifndef IPNOPRIVPORTS - || anonportmax < IPPORT_RESERVED -#endif - ) { - anonportmax = old; - return (EINVAL); - } - return (error); - case IPCTL_MTUDISCTIMEOUT: - old = ip_mtudisc_timeout; - error = sysctl_int(oldp, oldlenp, newp, newlen, - &ip_mtudisc_timeout); - 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 - case IPCTL_MAXFLOWS: - { - int s; - - error = sysctl_int(oldp, oldlenp, newp, newlen, - &ip_maxflows); - s = splsoftnet(); - ipflow_reap(0); - splx(s); + + error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_FORWSRCRT, + 0, NULL, NULL, NULL); + if (error) return (error); - } -#endif - case IPCTL_HOSTZEROBROADCAST: - return (sysctl_int(oldp, oldlenp, newp, newlen, - &hostzeroisbroadcast)); -#if NGIF > 0 - case IPCTL_GIF_TTL: - 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, - &ip_gre_ttl)); -#endif + ip_forwsrcrt = tmp; + return (0); +} + +/* + * sysctl helper routine for net.inet.ip.mtudisctimeout. checks the + * range of the new value and tweaks timers if it changes. + */ +static int +sysctl_net_inet_ip_pmtudto(SYSCTLFN_ARGS) +{ + int error, tmp; + struct sysctlnode node; + + icmp_mtudisc_lock(); + + node = *rnode; + tmp = ip_mtudisc_timeout; + node.sysctl_data = &tmp; + error = sysctl_lookup(SYSCTLFN_CALL(&node)); + if (error || newp == NULL) + goto out; + if (tmp < 0) { + error = EINVAL; + goto out; + } + + ip_mtudisc_timeout = tmp; + rt_timer_queue_change(ip_mtudisc_timeout_q, ip_mtudisc_timeout); + error = 0; +out: + icmp_mtudisc_unlock(); + return error; +} + +static int +sysctl_net_inet_ip_stats(SYSCTLFN_ARGS) +{ + + return (NETSTAT_SYSCTL(ipstat_percpu, IP_NSTATS)); +} + +static void +sysctl_net_inet_ip_setup(struct sysctllog **clog) +{ + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT, + CTLTYPE_NODE, "inet", + SYSCTL_DESCR("PF_INET related settings"), + NULL, 0, NULL, 0, + CTL_NET, PF_INET, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT, + CTLTYPE_NODE, "ip", + SYSCTL_DESCR("IPv4 related settings"), + NULL, 0, NULL, 0, + CTL_NET, PF_INET, IPPROTO_IP, CTL_EOL); + + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "forwarding", + SYSCTL_DESCR("Enable forwarding of INET datagrams"), + NULL, 0, &ipforwarding, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_FORWARDING, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "redirect", + SYSCTL_DESCR("Enable sending of ICMP redirect messages"), + NULL, 0, &ipsendredirects, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_SENDREDIRECTS, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "ttl", + SYSCTL_DESCR("Default TTL for an INET datagram"), + NULL, 0, &ip_defttl, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_DEFTTL, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "forwsrcrt", + SYSCTL_DESCR("Enable forwarding of source-routed " + "datagrams"), + sysctl_net_inet_ip_forwsrcrt, 0, &ip_forwsrcrt, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_FORWSRCRT, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "directed-broadcast", + SYSCTL_DESCR("Enable forwarding of broadcast datagrams"), + NULL, 0, &ip_directedbcast, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_DIRECTEDBCAST, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "allowsrcrt", + SYSCTL_DESCR("Accept source-routed datagrams"), + NULL, 0, &ip_allowsrcrt, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_ALLOWSRCRT, CTL_EOL); + + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "mtudisc", + SYSCTL_DESCR("Use RFC1191 Path MTU Discovery"), + NULL, 0, &ip_mtudisc, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_MTUDISC, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "anonportmin", + SYSCTL_DESCR("Lowest ephemeral port number to assign"), + sysctl_net_inet_ip_ports, 0, &anonportmin, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_ANONPORTMIN, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "anonportmax", + SYSCTL_DESCR("Highest ephemeral port number to assign"), + sysctl_net_inet_ip_ports, 0, &anonportmax, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_ANONPORTMAX, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "mtudisctimeout", + SYSCTL_DESCR("Lifetime of a Path MTU Discovered route"), + sysctl_net_inet_ip_pmtudto, 0, (void *)&ip_mtudisc_timeout, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_MTUDISCTIMEOUT, CTL_EOL); #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 + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "lowportmin", + SYSCTL_DESCR("Lowest privileged ephemeral port number " + "to assign"), + sysctl_net_inet_ip_ports, 0, &lowportmin, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_LOWPORTMIN, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "lowportmax", + SYSCTL_DESCR("Highest privileged ephemeral port number " + "to assign"), + sysctl_net_inet_ip_ports, 0, &lowportmax, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_LOWPORTMAX, CTL_EOL); +#endif /* IPNOPRIVPORTS */ +#if NGRE > 0 + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "grettl", + SYSCTL_DESCR("Default TTL for a gre tunnel datagram"), + NULL, 0, &ip_gre_ttl, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_GRE_TTL, CTL_EOL); +#endif /* NGRE */ + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "checkinterface", + SYSCTL_DESCR("Enable receive side of Strong ES model " + "from RFC1122"), + NULL, 0, &ip_checkinterface, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_CHECKINTERFACE, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "random_id", + SYSCTL_DESCR("Assign random ip_id values"), + NULL, 0, &ip_do_randomid, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_RANDOMID, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "do_loopback_cksum", + SYSCTL_DESCR("Perform IP checksum on loopback"), + NULL, 0, &ip_do_loopback_cksum, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_LOOPBACKCKSUM, CTL_EOL); + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT, + CTLTYPE_STRUCT, "stats", + SYSCTL_DESCR("IP statistics"), + sysctl_net_inet_ip_stats, 0, NULL, 0, + CTL_NET, PF_INET, IPPROTO_IP, IPCTL_STATS, + CTL_EOL); +#if NARP + sysctl_createv(clog, 0, NULL, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_INT, "dad_count", + SYSCTL_DESCR("Number of Duplicate Address Detection " + "probes to send"), + NULL, 0, &ip_dad_count, 0, + CTL_NET, PF_INET, IPPROTO_IP, + IPCTL_DAD_COUNT, CTL_EOL); +#endif + + /* anonportalgo RFC6056 subtree */ + const struct sysctlnode *portalgo_node; + sysctl_createv(clog, 0, NULL, &portalgo_node, + CTLFLAG_PERMANENT, + CTLTYPE_NODE, "anonportalgo", + SYSCTL_DESCR("Anonymous Port Algorithm Selection (RFC 6056)"), + NULL, 0, NULL, 0, + CTL_NET, PF_INET, IPPROTO_IP, CTL_CREATE, CTL_EOL); + sysctl_createv(clog, 0, &portalgo_node, NULL, + CTLFLAG_PERMANENT, + CTLTYPE_STRING, "available", + SYSCTL_DESCR("available algorithms"), + sysctl_portalgo_available, 0, NULL, PORTALGO_MAXLEN, + CTL_CREATE, CTL_EOL); + sysctl_createv(clog, 0, &portalgo_node, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_STRING, "selected", + SYSCTL_DESCR("selected algorithm"), + sysctl_portalgo_selected4, 0, NULL, PORTALGO_MAXLEN, + CTL_CREATE, CTL_EOL); + sysctl_createv(clog, 0, &portalgo_node, NULL, + CTLFLAG_PERMANENT|CTLFLAG_READWRITE, + CTLTYPE_STRUCT, "reserve", + SYSCTL_DESCR("bitmap of reserved ports"), + sysctl_portalgo_reserve4, 0, NULL, 0, + CTL_CREATE, CTL_EOL); +} - case IPCTL_MAXFRAGPACKETS: - return (sysctl_int(oldp, oldlenp, newp, newlen, - &ip_maxfragpackets)); - - case IPCTL_CHECKINTERFACE: - return (sysctl_int(oldp, oldlenp, newp, newlen, - &ip_checkinterface)); - default: - return (EOPNOTSUPP); - } - /* NOTREACHED */ +void +ip_statinc(u_int stat) +{ + + KASSERT(stat < IP_NSTATS); + IP_STATINC(stat); }