Home | History | Annotate | Line # | Download | only in netinet
ip_input.c revision 1.319.4.7
      1 /*	$NetBSD: ip_input.c,v 1.319.4.7 2016/10/05 20:56:09 skrll Exp $	*/
      2 
      3 /*
      4  * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
      5  * All rights reserved.
      6  *
      7  * Redistribution and use in source and binary forms, with or without
      8  * modification, are permitted provided that the following conditions
      9  * are met:
     10  * 1. Redistributions of source code must retain the above copyright
     11  *    notice, this list of conditions and the following disclaimer.
     12  * 2. Redistributions in binary form must reproduce the above copyright
     13  *    notice, this list of conditions and the following disclaimer in the
     14  *    documentation and/or other materials provided with the distribution.
     15  * 3. Neither the name of the project nor the names of its contributors
     16  *    may be used to endorse or promote products derived from this software
     17  *    without specific prior written permission.
     18  *
     19  * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
     20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
     23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     29  * SUCH DAMAGE.
     30  */
     31 
     32 /*-
     33  * Copyright (c) 1998 The NetBSD Foundation, Inc.
     34  * All rights reserved.
     35  *
     36  * This code is derived from software contributed to The NetBSD Foundation
     37  * by Public Access Networks Corporation ("Panix").  It was developed under
     38  * contract to Panix by Eric Haszlakiewicz and Thor Lancelot Simon.
     39  *
     40  * Redistribution and use in source and binary forms, with or without
     41  * modification, are permitted provided that the following conditions
     42  * are met:
     43  * 1. Redistributions of source code must retain the above copyright
     44  *    notice, this list of conditions and the following disclaimer.
     45  * 2. Redistributions in binary form must reproduce the above copyright
     46  *    notice, this list of conditions and the following disclaimer in the
     47  *    documentation and/or other materials provided with the distribution.
     48  *
     49  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
     50  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
     51  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
     52  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
     53  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
     54  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
     55  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
     56  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
     57  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
     58  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
     59  * POSSIBILITY OF SUCH DAMAGE.
     60  */
     61 
     62 /*
     63  * Copyright (c) 1982, 1986, 1988, 1993
     64  *	The Regents of the University of California.  All rights reserved.
     65  *
     66  * Redistribution and use in source and binary forms, with or without
     67  * modification, are permitted provided that the following conditions
     68  * are met:
     69  * 1. Redistributions of source code must retain the above copyright
     70  *    notice, this list of conditions and the following disclaimer.
     71  * 2. Redistributions in binary form must reproduce the above copyright
     72  *    notice, this list of conditions and the following disclaimer in the
     73  *    documentation and/or other materials provided with the distribution.
     74  * 3. Neither the name of the University nor the names of its contributors
     75  *    may be used to endorse or promote products derived from this software
     76  *    without specific prior written permission.
     77  *
     78  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     79  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     80  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     81  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     82  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     83  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     84  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     85  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     86  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     87  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     88  * SUCH DAMAGE.
     89  *
     90  *	@(#)ip_input.c	8.2 (Berkeley) 1/4/94
     91  */
     92 
     93 #include <sys/cdefs.h>
     94 __KERNEL_RCSID(0, "$NetBSD: ip_input.c,v 1.319.4.7 2016/10/05 20:56:09 skrll Exp $");
     95 
     96 #ifdef _KERNEL_OPT
     97 #include "opt_inet.h"
     98 #include "opt_compat_netbsd.h"
     99 #include "opt_gateway.h"
    100 #include "opt_ipsec.h"
    101 #include "opt_mrouting.h"
    102 #include "opt_mbuftrace.h"
    103 #include "opt_inet_csum.h"
    104 #endif
    105 
    106 #include "arp.h"
    107 
    108 #include <sys/param.h>
    109 #include <sys/systm.h>
    110 #include <sys/cpu.h>
    111 #include <sys/mbuf.h>
    112 #include <sys/domain.h>
    113 #include <sys/protosw.h>
    114 #include <sys/socket.h>
    115 #include <sys/socketvar.h>
    116 #include <sys/errno.h>
    117 #include <sys/time.h>
    118 #include <sys/kernel.h>
    119 #include <sys/pool.h>
    120 #include <sys/sysctl.h>
    121 #include <sys/kauth.h>
    122 
    123 #include <net/if.h>
    124 #include <net/if_dl.h>
    125 #include <net/route.h>
    126 #include <net/pktqueue.h>
    127 #include <net/pfil.h>
    128 
    129 #include <netinet/in.h>
    130 #include <netinet/in_systm.h>
    131 #include <netinet/ip.h>
    132 #include <netinet/in_pcb.h>
    133 #include <netinet/in_proto.h>
    134 #include <netinet/in_var.h>
    135 #include <netinet/ip_var.h>
    136 #include <netinet/ip_private.h>
    137 #include <netinet/ip_icmp.h>
    138 /* just for gif_ttl */
    139 #include <netinet/in_gif.h>
    140 #include "gif.h"
    141 #include <net/if_gre.h>
    142 #include "gre.h"
    143 
    144 #ifdef MROUTING
    145 #include <netinet/ip_mroute.h>
    146 #endif
    147 #include <netinet/portalgo.h>
    148 
    149 #ifdef IPSEC
    150 #include <netipsec/ipsec.h>
    151 #endif
    152 
    153 #ifndef	IPFORWARDING
    154 #ifdef GATEWAY
    155 #define	IPFORWARDING	1	/* forward IP packets not for us */
    156 #else /* GATEWAY */
    157 #define	IPFORWARDING	0	/* don't forward IP packets not for us */
    158 #endif /* GATEWAY */
    159 #endif /* IPFORWARDING */
    160 #ifndef	IPSENDREDIRECTS
    161 #define	IPSENDREDIRECTS	1
    162 #endif
    163 #ifndef IPFORWSRCRT
    164 #define	IPFORWSRCRT	1	/* forward source-routed packets */
    165 #endif
    166 #ifndef IPALLOWSRCRT
    167 #define	IPALLOWSRCRT	1	/* allow source-routed packets */
    168 #endif
    169 #ifndef IPMTUDISC
    170 #define IPMTUDISC	1
    171 #endif
    172 #ifndef IPMTUDISCTIMEOUT
    173 #define IPMTUDISCTIMEOUT (10 * 60)	/* as per RFC 1191 */
    174 #endif
    175 
    176 #ifdef COMPAT_50
    177 #include <compat/sys/time.h>
    178 #include <compat/sys/socket.h>
    179 #endif
    180 
    181 /*
    182  * Note: DIRECTED_BROADCAST is handled this way so that previous
    183  * configuration using this option will Just Work.
    184  */
    185 #ifndef IPDIRECTEDBCAST
    186 #ifdef DIRECTED_BROADCAST
    187 #define IPDIRECTEDBCAST	1
    188 #else
    189 #define	IPDIRECTEDBCAST	0
    190 #endif /* DIRECTED_BROADCAST */
    191 #endif /* IPDIRECTEDBCAST */
    192 int	ipforwarding = IPFORWARDING;
    193 int	ipsendredirects = IPSENDREDIRECTS;
    194 int	ip_defttl = IPDEFTTL;
    195 int	ip_forwsrcrt = IPFORWSRCRT;
    196 int	ip_directedbcast = IPDIRECTEDBCAST;
    197 int	ip_allowsrcrt = IPALLOWSRCRT;
    198 int	ip_mtudisc = IPMTUDISC;
    199 int	ip_mtudisc_timeout = IPMTUDISCTIMEOUT;
    200 #ifdef DIAGNOSTIC
    201 int	ipprintfs = 0;
    202 #endif
    203 
    204 int	ip_do_randomid = 0;
    205 
    206 /*
    207  * XXX - Setting ip_checkinterface mostly implements the receive side of
    208  * the Strong ES model described in RFC 1122, but since the routing table
    209  * and transmit implementation do not implement the Strong ES model,
    210  * setting this to 1 results in an odd hybrid.
    211  *
    212  * XXX - ip_checkinterface currently must be disabled if you use ipnat
    213  * to translate the destination address to another local interface.
    214  *
    215  * XXX - ip_checkinterface must be disabled if you add IP aliases
    216  * to the loopback interface instead of the interface where the
    217  * packets for those addresses are received.
    218  */
    219 static int		ip_checkinterface	__read_mostly = 0;
    220 
    221 struct rttimer_queue *ip_mtudisc_timeout_q = NULL;
    222 
    223 pktqueue_t *		ip_pktq			__read_mostly;
    224 pfil_head_t *		inet_pfil_hook		__read_mostly;
    225 ipid_state_t *		ip_ids			__read_mostly;
    226 percpu_t *		ipstat_percpu		__read_mostly;
    227 
    228 static percpu_t		*ipforward_rt_percpu	__cacheline_aligned;
    229 
    230 uint16_t ip_id;
    231 
    232 #ifdef INET_CSUM_COUNTERS
    233 #include <sys/device.h>
    234 
    235 struct evcnt ip_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
    236     NULL, "inet", "hwcsum bad");
    237 struct evcnt ip_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
    238     NULL, "inet", "hwcsum ok");
    239 struct evcnt ip_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
    240     NULL, "inet", "swcsum");
    241 
    242 #define	INET_CSUM_COUNTER_INCR(ev)	(ev)->ev_count++
    243 
    244 EVCNT_ATTACH_STATIC(ip_hwcsum_bad);
    245 EVCNT_ATTACH_STATIC(ip_hwcsum_ok);
    246 EVCNT_ATTACH_STATIC(ip_swcsum);
    247 
    248 #else
    249 
    250 #define	INET_CSUM_COUNTER_INCR(ev)	/* nothing */
    251 
    252 #endif /* INET_CSUM_COUNTERS */
    253 
    254 /*
    255  * We need to save the IP options in case a protocol wants to respond
    256  * to an incoming packet over the same route if the packet got here
    257  * using IP source routing.  This allows connection establishment and
    258  * maintenance when the remote end is on a network that is not known
    259  * to us.
    260  */
    261 
    262 static int	ip_nhops = 0;
    263 
    264 static	struct ip_srcrt {
    265 	struct	in_addr dst;			/* final destination */
    266 	char	nop;				/* one NOP to align */
    267 	char	srcopt[IPOPT_OFFSET + 1];	/* OPTVAL, OLEN and OFFSET */
    268 	struct	in_addr route[MAX_IPOPTLEN/sizeof(struct in_addr)];
    269 } ip_srcrt;
    270 
    271 static int ip_drainwanted;
    272 
    273 struct	sockaddr_in ipaddr = {
    274 	.sin_len = sizeof(ipaddr),
    275 	.sin_family = AF_INET,
    276 };
    277 
    278 static void save_rte(u_char *, struct in_addr);
    279 
    280 #ifdef MBUFTRACE
    281 struct mowner ip_rx_mowner = MOWNER_INIT("internet", "rx");
    282 struct mowner ip_tx_mowner = MOWNER_INIT("internet", "tx");
    283 #endif
    284 
    285 static void		ipintr(void *);
    286 static void		ip_input(struct mbuf *);
    287 static void		ip_forward(struct mbuf *, int, struct ifnet *);
    288 static bool		ip_dooptions(struct mbuf *);
    289 static struct in_ifaddr *ip_rtaddr(struct in_addr);
    290 static void		sysctl_net_inet_ip_setup(struct sysctllog **);
    291 
    292 static struct in_ifaddr	*ip_match_our_address(struct ifnet *, struct ip *,
    293 			    int *);
    294 static struct in_ifaddr	*ip_match_our_address_broadcast(struct ifnet *,
    295 			    struct ip *);
    296 
    297 /* XXX: Not yet enabled. */
    298 #define	SOFTNET_LOCK()		KASSERT(mutex_owned(softnet_lock))
    299 #define	SOFTNET_UNLOCK()	KASSERT(mutex_owned(softnet_lock))
    300 
    301 /*
    302  * IP initialization: fill in IP protocol switch table.
    303  * All protocols not implemented in kernel go to raw IP protocol handler.
    304  */
    305 void
    306 ip_init(void)
    307 {
    308 	const struct protosw *pr;
    309 
    310 	in_init();
    311 	sysctl_net_inet_ip_setup(NULL);
    312 
    313 	pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW);
    314 	KASSERT(pr != NULL);
    315 
    316 	ip_pktq = pktq_create(IFQ_MAXLEN, ipintr, NULL);
    317 	KASSERT(ip_pktq != NULL);
    318 
    319 	for (u_int i = 0; i < IPPROTO_MAX; i++) {
    320 		ip_protox[i] = pr - inetsw;
    321 	}
    322 	for (pr = inetdomain.dom_protosw;
    323 	    pr < inetdomain.dom_protoswNPROTOSW; pr++)
    324 		if (pr->pr_domain->dom_family == PF_INET &&
    325 		    pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW)
    326 			ip_protox[pr->pr_protocol] = pr - inetsw;
    327 
    328 	ip_reass_init();
    329 
    330 	ip_ids = ip_id_init();
    331 	ip_id = time_uptime & 0xfffff;
    332 
    333 	ip_mtudisc_timeout_q = rt_timer_queue_create(ip_mtudisc_timeout);
    334 #ifdef GATEWAY
    335 	ipflow_init();
    336 #endif
    337 
    338 	/* Register our Packet Filter hook. */
    339 	inet_pfil_hook = pfil_head_create(PFIL_TYPE_AF, (void *)AF_INET);
    340 	KASSERT(inet_pfil_hook != NULL);
    341 
    342 #ifdef MBUFTRACE
    343 	MOWNER_ATTACH(&ip_tx_mowner);
    344 	MOWNER_ATTACH(&ip_rx_mowner);
    345 #endif /* MBUFTRACE */
    346 
    347 	ipstat_percpu = percpu_alloc(sizeof(uint64_t) * IP_NSTATS);
    348 
    349 	ipforward_rt_percpu = percpu_alloc(sizeof(struct route));
    350 	if (ipforward_rt_percpu == NULL)
    351 		panic("failed to allocate ipforward_rt_percpu");
    352 }
    353 
    354 static struct in_ifaddr *
    355 ip_match_our_address(struct ifnet *ifp, struct ip *ip, int *downmatch)
    356 {
    357 	struct in_ifaddr *ia = NULL;
    358 	int checkif;
    359 
    360 	/*
    361 	 * Enable a consistency check between the destination address
    362 	 * and the arrival interface for a unicast packet (the RFC 1122
    363 	 * strong ES model) if IP forwarding is disabled and the packet
    364 	 * is not locally generated.
    365 	 *
    366 	 * XXX - Checking also should be disabled if the destination
    367 	 * address is ipnat'ed to a different interface.
    368 	 *
    369 	 * XXX - Checking is incompatible with IP aliases added
    370 	 * to the loopback interface instead of the interface where
    371 	 * the packets are received.
    372 	 *
    373 	 * XXX - We need to add a per ifaddr flag for this so that
    374 	 * we get finer grain control.
    375 	 */
    376 	checkif = ip_checkinterface && (ipforwarding == 0) &&
    377 	    (ifp->if_flags & IFF_LOOPBACK) == 0;
    378 
    379 	IN_ADDRHASH_READER_FOREACH(ia, ip->ip_dst.s_addr) {
    380 		if (in_hosteq(ia->ia_addr.sin_addr, ip->ip_dst)) {
    381 			if (ia->ia4_flags & IN_IFF_NOTREADY)
    382 				continue;
    383 			if (checkif && ia->ia_ifp != ifp)
    384 				continue;
    385 			if ((ia->ia_ifp->if_flags & IFF_UP) != 0 &&
    386 			    (ia->ia4_flags & IN_IFF_DETACHED) == 0)
    387 				break;
    388 			else
    389 				(*downmatch)++;
    390 		}
    391 	}
    392 
    393 	return ia;
    394 }
    395 
    396 static struct in_ifaddr *
    397 ip_match_our_address_broadcast(struct ifnet *ifp, struct ip *ip)
    398 {
    399 	struct in_ifaddr *ia = NULL;
    400 	struct ifaddr *ifa;
    401 
    402 	IFADDR_READER_FOREACH(ifa, ifp) {
    403 		if (ifa->ifa_addr->sa_family != AF_INET)
    404 			continue;
    405 		ia = ifatoia(ifa);
    406 		if (ia->ia4_flags & (IN_IFF_NOTREADY | IN_IFF_DETACHED))
    407 			continue;
    408 		if (in_hosteq(ip->ip_dst, ia->ia_broadaddr.sin_addr) ||
    409 		    in_hosteq(ip->ip_dst, ia->ia_netbroadcast) ||
    410 		    /*
    411 		     * Look for all-0's host part (old broadcast addr),
    412 		     * either for subnet or net.
    413 		     */
    414 		    ip->ip_dst.s_addr == ia->ia_subnet ||
    415 		    ip->ip_dst.s_addr == ia->ia_net)
    416 			goto matched;
    417 		/*
    418 		 * An interface with IP address zero accepts
    419 		 * all packets that arrive on that interface.
    420 		 */
    421 		if (in_nullhost(ia->ia_addr.sin_addr))
    422 			goto matched;
    423 	}
    424 	ia = NULL;
    425 
    426 matched:
    427 	return ia;
    428 }
    429 
    430 /*
    431  * IP software interrupt routine.
    432  */
    433 static void
    434 ipintr(void *arg __unused)
    435 {
    436 	struct mbuf *m;
    437 
    438 	KASSERT(cpu_softintr_p());
    439 
    440 	mutex_enter(softnet_lock);
    441 	while ((m = pktq_dequeue(ip_pktq)) != NULL) {
    442 		ip_input(m);
    443 	}
    444 	mutex_exit(softnet_lock);
    445 }
    446 
    447 /*
    448  * IP input routine.  Checksum and byte swap header.  If fragmented
    449  * try to reassemble.  Process options.  Pass to next level.
    450  */
    451 static void
    452 ip_input(struct mbuf *m)
    453 {
    454 	struct ip *ip = NULL;
    455 	struct in_ifaddr *ia = NULL;
    456 	int hlen = 0, len;
    457 	int downmatch;
    458 	int srcrt = 0;
    459 	ifnet_t *ifp;
    460 	struct psref psref;
    461 	int s;
    462 
    463 	KASSERTMSG(cpu_softintr_p(), "ip_input: not in the software "
    464 	    "interrupt handler; synchronization assumptions violated");
    465 
    466 	MCLAIM(m, &ip_rx_mowner);
    467 	KASSERT((m->m_flags & M_PKTHDR) != 0);
    468 
    469 	ifp = m_get_rcvif_psref(m, &psref);
    470 	if (__predict_false(ifp == NULL))
    471 		goto out;
    472 
    473 	/*
    474 	 * If no IP addresses have been set yet but the interfaces
    475 	 * are receiving, can't do anything with incoming packets yet.
    476 	 * Note: we pre-check without locks held.
    477 	 */
    478 	if (IN_ADDRLIST_READER_EMPTY())
    479 		goto out;
    480 	IP_STATINC(IP_STAT_TOTAL);
    481 
    482 	/*
    483 	 * If the IP header is not aligned, slurp it up into a new
    484 	 * mbuf with space for link headers, in the event we forward
    485 	 * it.  Otherwise, if it is aligned, make sure the entire
    486 	 * base IP header is in the first mbuf of the chain.
    487 	 */
    488 	if (IP_HDR_ALIGNED_P(mtod(m, void *)) == 0) {
    489 		if ((m = m_copyup(m, sizeof(struct ip),
    490 				  (max_linkhdr + 3) & ~3)) == NULL) {
    491 			/* XXXJRT new stat, please */
    492 			IP_STATINC(IP_STAT_TOOSMALL);
    493 			goto out;
    494 		}
    495 	} else if (__predict_false(m->m_len < sizeof (struct ip))) {
    496 		if ((m = m_pullup(m, sizeof (struct ip))) == NULL) {
    497 			IP_STATINC(IP_STAT_TOOSMALL);
    498 			goto out;
    499 		}
    500 	}
    501 	ip = mtod(m, struct ip *);
    502 	if (ip->ip_v != IPVERSION) {
    503 		IP_STATINC(IP_STAT_BADVERS);
    504 		goto out;
    505 	}
    506 	hlen = ip->ip_hl << 2;
    507 	if (hlen < sizeof(struct ip)) {	/* minimum header length */
    508 		IP_STATINC(IP_STAT_BADHLEN);
    509 		goto out;
    510 	}
    511 	if (hlen > m->m_len) {
    512 		if ((m = m_pullup(m, hlen)) == NULL) {
    513 			IP_STATINC(IP_STAT_BADHLEN);
    514 			goto out;
    515 		}
    516 		ip = mtod(m, struct ip *);
    517 	}
    518 
    519 	/*
    520 	 * RFC1122: packets with a multicast source address are
    521 	 * not allowed.
    522 	 */
    523 	if (IN_MULTICAST(ip->ip_src.s_addr)) {
    524 		IP_STATINC(IP_STAT_BADADDR);
    525 		goto out;
    526 	}
    527 
    528 	/* 127/8 must not appear on wire - RFC1122 */
    529 	if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET ||
    530 	    (ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) {
    531 		if ((ifp->if_flags & IFF_LOOPBACK) == 0) {
    532 			IP_STATINC(IP_STAT_BADADDR);
    533 			goto out;
    534 		}
    535 	}
    536 
    537 	switch (m->m_pkthdr.csum_flags &
    538 		((ifp->if_csum_flags_rx & M_CSUM_IPv4) |
    539 		 M_CSUM_IPv4_BAD)) {
    540 	case M_CSUM_IPv4|M_CSUM_IPv4_BAD:
    541 		INET_CSUM_COUNTER_INCR(&ip_hwcsum_bad);
    542 		IP_STATINC(IP_STAT_BADSUM);
    543 		goto out;
    544 
    545 	case M_CSUM_IPv4:
    546 		/* Checksum was okay. */
    547 		INET_CSUM_COUNTER_INCR(&ip_hwcsum_ok);
    548 		break;
    549 
    550 	default:
    551 		/*
    552 		 * Must compute it ourselves.  Maybe skip checksum on
    553 		 * loopback interfaces.
    554 		 */
    555 		if (__predict_true(!(ifp->if_flags & IFF_LOOPBACK) ||
    556 		    ip_do_loopback_cksum)) {
    557 			INET_CSUM_COUNTER_INCR(&ip_swcsum);
    558 			if (in_cksum(m, hlen) != 0) {
    559 				IP_STATINC(IP_STAT_BADSUM);
    560 				goto out;
    561 			}
    562 		}
    563 		break;
    564 	}
    565 
    566 	/* Retrieve the packet length. */
    567 	len = ntohs(ip->ip_len);
    568 
    569 	/*
    570 	 * Check for additional length bogosity
    571 	 */
    572 	if (len < hlen) {
    573 		IP_STATINC(IP_STAT_BADLEN);
    574 		goto out;
    575 	}
    576 
    577 	/*
    578 	 * Check that the amount of data in the buffers
    579 	 * is as at least much as the IP header would have us expect.
    580 	 * Trim mbufs if longer than we expect.
    581 	 * Drop packet if shorter than we expect.
    582 	 */
    583 	if (m->m_pkthdr.len < len) {
    584 		IP_STATINC(IP_STAT_TOOSHORT);
    585 		goto out;
    586 	}
    587 	if (m->m_pkthdr.len > len) {
    588 		if (m->m_len == m->m_pkthdr.len) {
    589 			m->m_len = len;
    590 			m->m_pkthdr.len = len;
    591 		} else
    592 			m_adj(m, len - m->m_pkthdr.len);
    593 	}
    594 
    595 	/*
    596 	 * Assume that we can create a fast-forward IP flow entry
    597 	 * based on this packet.
    598 	 */
    599 	m->m_flags |= M_CANFASTFWD;
    600 
    601 	/*
    602 	 * Run through list of hooks for input packets.  If there are any
    603 	 * filters which require that additional packets in the flow are
    604 	 * not fast-forwarded, they must clear the M_CANFASTFWD flag.
    605 	 * Note that filters must _never_ set this flag, as another filter
    606 	 * in the list may have previously cleared it.
    607 	 */
    608 #if defined(IPSEC)
    609 	if (!ipsec_used || !ipsec_indone(m))
    610 #else
    611 	if (1)
    612 #endif
    613 	{
    614 		struct in_addr odst = ip->ip_dst;
    615 		bool freed;
    616 
    617 		SOFTNET_LOCK();
    618 		freed = pfil_run_hooks(inet_pfil_hook, &m, ifp, PFIL_IN) != 0;
    619 		SOFTNET_UNLOCK();
    620 		if (freed || m == NULL) {
    621 			goto out;
    622 		}
    623 		ip = mtod(m, struct ip *);
    624 		hlen = ip->ip_hl << 2;
    625 
    626 		/*
    627 		 * XXX The setting of "srcrt" here is to prevent ip_forward()
    628 		 * from generating ICMP redirects for packets that have
    629 		 * been redirected by a hook back out on to the same LAN that
    630 		 * they came from and is not an indication that the packet
    631 		 * is being inffluenced by source routing options.  This
    632 		 * allows things like
    633 		 * "rdr tlp0 0/0 port 80 -> 1.1.1.200 3128 tcp"
    634 		 * where tlp0 is both on the 1.1.1.0/24 network and is the
    635 		 * default route for hosts on 1.1.1.0/24.  Of course this
    636 		 * also requires a "map tlp0 ..." to complete the story.
    637 		 * One might argue whether or not this kind of network config.
    638 		 * should be supported in this manner...
    639 		 */
    640 		srcrt = (odst.s_addr != ip->ip_dst.s_addr);
    641 	}
    642 
    643 #ifdef ALTQ
    644 	/* XXX Temporary until ALTQ is changed to use a pfil hook */
    645 	if (altq_input) {
    646 		SOFTNET_LOCK();
    647 		if ((*altq_input)(m, AF_INET) == 0) {
    648 			/* Packet dropped by traffic conditioner. */
    649 			SOFTNET_UNLOCK();
    650 			goto out;
    651 		}
    652 		SOFTNET_UNLOCK();
    653 	}
    654 #endif
    655 
    656 	/*
    657 	 * Process options and, if not destined for us,
    658 	 * ship it on.  ip_dooptions returns 1 when an
    659 	 * error was detected (causing an icmp message
    660 	 * to be sent and the original packet to be freed).
    661 	 */
    662 	ip_nhops = 0;		/* for source routed packets */
    663 	if (hlen > sizeof (struct ip) && ip_dooptions(m))
    664 		goto out;
    665 
    666 	/*
    667 	 * Check our list of addresses, to see if the packet is for us.
    668 	 *
    669 	 * Traditional 4.4BSD did not consult IFF_UP at all.
    670 	 * The behavior here is to treat addresses on !IFF_UP interface
    671 	 * or IN_IFF_NOTREADY addresses as not mine.
    672 	 */
    673 	downmatch = 0;
    674 	s = pserialize_read_enter();
    675 	ia = ip_match_our_address(ifp, ip, &downmatch);
    676 	if (ia != NULL) {
    677 		pserialize_read_exit(s);
    678 		goto ours;
    679 	}
    680 
    681 	if (ifp->if_flags & IFF_BROADCAST) {
    682 		ia = ip_match_our_address_broadcast(ifp, ip);
    683 		if (ia != NULL) {
    684 			pserialize_read_exit(s);
    685 			goto ours;
    686 		}
    687 	}
    688 	pserialize_read_exit(s);
    689 
    690 	if (IN_MULTICAST(ip->ip_dst.s_addr)) {
    691 #ifdef MROUTING
    692 		extern struct socket *ip_mrouter;
    693 
    694 		if (ip_mrouter) {
    695 			/*
    696 			 * If we are acting as a multicast router, all
    697 			 * incoming multicast packets are passed to the
    698 			 * kernel-level multicast forwarding function.
    699 			 * The packet is returned (relatively) intact; if
    700 			 * ip_mforward() returns a non-zero value, the packet
    701 			 * must be discarded, else it may be accepted below.
    702 			 *
    703 			 * (The IP ident field is put in the same byte order
    704 			 * as expected when ip_mforward() is called from
    705 			 * ip_output().)
    706 			 */
    707 			SOFTNET_LOCK();
    708 			if (ip_mforward(m, ifp) != 0) {
    709 				SOFTNET_UNLOCK();
    710 				IP_STATINC(IP_STAT_CANTFORWARD);
    711 				goto out;
    712 			}
    713 			SOFTNET_UNLOCK();
    714 
    715 			/*
    716 			 * The process-level routing demon needs to receive
    717 			 * all multicast IGMP packets, whether or not this
    718 			 * host belongs to their destination groups.
    719 			 */
    720 			if (ip->ip_p == IPPROTO_IGMP) {
    721 				goto ours;
    722 			}
    723 			IP_STATINC(IP_STAT_CANTFORWARD);
    724 		}
    725 #endif
    726 		/*
    727 		 * See if we belong to the destination multicast group on the
    728 		 * arrival interface.
    729 		 */
    730 		if (!in_multi_group(ip->ip_dst, ifp, 0)) {
    731 			IP_STATINC(IP_STAT_CANTFORWARD);
    732 			goto out;
    733 		}
    734 		goto ours;
    735 	}
    736 	if (ip->ip_dst.s_addr == INADDR_BROADCAST ||
    737 	    in_nullhost(ip->ip_dst))
    738 		goto ours;
    739 
    740 	/*
    741 	 * Not for us; forward if possible and desirable.
    742 	 */
    743 	if (ipforwarding == 0) {
    744 		m_put_rcvif_psref(ifp, &psref);
    745 		IP_STATINC(IP_STAT_CANTFORWARD);
    746 		m_freem(m);
    747 	} else {
    748 		/*
    749 		 * If ip_dst matched any of my address on !IFF_UP interface,
    750 		 * and there's no IFF_UP interface that matches ip_dst,
    751 		 * send icmp unreach.  Forwarding it will result in in-kernel
    752 		 * forwarding loop till TTL goes to 0.
    753 		 */
    754 		if (downmatch) {
    755 			m_put_rcvif_psref(ifp, &psref);
    756 			icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, 0, 0);
    757 			IP_STATINC(IP_STAT_CANTFORWARD);
    758 			return;
    759 		}
    760 #ifdef IPSEC
    761 		/* Perform IPsec, if any. */
    762 		if (ipsec_used) {
    763 			SOFTNET_LOCK();
    764 			if (ipsec4_input(m, IP_FORWARDING |
    765 			    (ip_directedbcast ? IP_ALLOWBROADCAST : 0)) != 0) {
    766 				SOFTNET_UNLOCK();
    767 				goto out;
    768 			}
    769 			SOFTNET_UNLOCK();
    770 		}
    771 #endif
    772 		ip_forward(m, srcrt, ifp);
    773 		m_put_rcvif_psref(ifp, &psref);
    774 	}
    775 	return;
    776 
    777 ours:
    778 	m_put_rcvif_psref(ifp, &psref);
    779 	ifp = NULL;
    780 
    781 	/*
    782 	 * If offset or IP_MF are set, must reassemble.
    783 	 */
    784 	if (ip->ip_off & ~htons(IP_DF|IP_RF)) {
    785 		/*
    786 		 * Pass to IP reassembly mechanism.
    787 		 */
    788 		if (ip_reass_packet(&m, ip) != 0) {
    789 			/* Failed; invalid fragment(s) or packet. */
    790 			goto out;
    791 		}
    792 		if (m == NULL) {
    793 			/* More fragments should come; silently return. */
    794 			goto out;
    795 		}
    796 		/*
    797 		 * Reassembly is done, we have the final packet.
    798 		 * Updated cached data in local variable(s).
    799 		 */
    800 		ip = mtod(m, struct ip *);
    801 		hlen = ip->ip_hl << 2;
    802 	}
    803 
    804 #ifdef IPSEC
    805 	/*
    806 	 * Enforce IPsec policy checking if we are seeing last header.
    807 	 * Note that we do not visit this with protocols with PCB layer
    808 	 * code - like UDP/TCP/raw IP.
    809 	 */
    810 	if (ipsec_used &&
    811 	    (inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0) {
    812 		SOFTNET_LOCK();
    813 		if (ipsec4_input(m, 0) != 0) {
    814 			SOFTNET_UNLOCK();
    815 			goto out;
    816 		}
    817 		SOFTNET_UNLOCK();
    818 	}
    819 #endif
    820 
    821 	/*
    822 	 * Switch out to protocol's input routine.
    823 	 */
    824 #if IFA_STATS
    825 	if (ia && ip) {
    826 		struct in_ifaddr *_ia;
    827 		/*
    828 		 * Keep a reference from ip_match_our_address with psref
    829 		 * is expensive, so explore ia here again.
    830 		 */
    831 		s = pserialize_read_enter();
    832 		_ia = in_get_ia(ip->ip_dst.s_addr);
    833 		_ia->ia_ifa.ifa_data.ifad_inbytes += ntohs(ip->ip_len);
    834 		pserialize_read_exit(s);
    835 	}
    836 #endif
    837 	IP_STATINC(IP_STAT_DELIVERED);
    838 
    839 	const int off = hlen, nh = ip->ip_p;
    840 
    841 	SOFTNET_LOCK();
    842 	(*inetsw[ip_protox[nh]].pr_input)(m, off, nh);
    843 	SOFTNET_UNLOCK();
    844 	return;
    845 
    846 out:
    847 	m_put_rcvif_psref(ifp, &psref);
    848 	if (m != NULL)
    849 		m_freem(m);
    850 }
    851 
    852 /*
    853  * IP timer processing.
    854  */
    855 void
    856 ip_slowtimo(void)
    857 {
    858 
    859 	mutex_enter(softnet_lock);
    860 	KERNEL_LOCK(1, NULL);
    861 
    862 	ip_reass_slowtimo();
    863 
    864 	KERNEL_UNLOCK_ONE(NULL);
    865 	mutex_exit(softnet_lock);
    866 }
    867 
    868 /*
    869  * IP drain processing.
    870  */
    871 void
    872 ip_drain(void)
    873 {
    874 
    875 	KERNEL_LOCK(1, NULL);
    876 	ip_reass_drain();
    877 	KERNEL_UNLOCK_ONE(NULL);
    878 }
    879 
    880 /*
    881  * ip_dooptions: perform option processing on a datagram, possibly discarding
    882  * it if bad options are encountered, or forwarding it if source-routed.
    883  *
    884  * => Returns true if packet has been forwarded/freed.
    885  * => Returns false if the packet should be processed further.
    886  */
    887 static bool
    888 ip_dooptions(struct mbuf *m)
    889 {
    890 	struct ip *ip = mtod(m, struct ip *);
    891 	u_char *cp, *cp0;
    892 	struct ip_timestamp *ipt;
    893 	struct in_ifaddr *ia;
    894 	int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0;
    895 	struct in_addr dst;
    896 	n_time ntime;
    897 	struct ifaddr *ifa;
    898 	int s;
    899 
    900 	dst = ip->ip_dst;
    901 	cp = (u_char *)(ip + 1);
    902 	cnt = (ip->ip_hl << 2) - sizeof (struct ip);
    903 	for (; cnt > 0; cnt -= optlen, cp += optlen) {
    904 		opt = cp[IPOPT_OPTVAL];
    905 		if (opt == IPOPT_EOL)
    906 			break;
    907 		if (opt == IPOPT_NOP)
    908 			optlen = 1;
    909 		else {
    910 			if (cnt < IPOPT_OLEN + sizeof(*cp)) {
    911 				code = &cp[IPOPT_OLEN] - (u_char *)ip;
    912 				goto bad;
    913 			}
    914 			optlen = cp[IPOPT_OLEN];
    915 			if (optlen < IPOPT_OLEN + sizeof(*cp) || optlen > cnt) {
    916 				code = &cp[IPOPT_OLEN] - (u_char *)ip;
    917 				goto bad;
    918 			}
    919 		}
    920 		switch (opt) {
    921 
    922 		default:
    923 			break;
    924 
    925 		/*
    926 		 * Source routing with record.
    927 		 * Find interface with current destination address.
    928 		 * If none on this machine then drop if strictly routed,
    929 		 * or do nothing if loosely routed.
    930 		 * Record interface address and bring up next address
    931 		 * component.  If strictly routed make sure next
    932 		 * address is on directly accessible net.
    933 		 */
    934 		case IPOPT_LSRR:
    935 		case IPOPT_SSRR:
    936 			if (ip_allowsrcrt == 0) {
    937 				type = ICMP_UNREACH;
    938 				code = ICMP_UNREACH_NET_PROHIB;
    939 				goto bad;
    940 			}
    941 			if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
    942 				code = &cp[IPOPT_OLEN] - (u_char *)ip;
    943 				goto bad;
    944 			}
    945 			if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
    946 				code = &cp[IPOPT_OFFSET] - (u_char *)ip;
    947 				goto bad;
    948 			}
    949 			ipaddr.sin_addr = ip->ip_dst;
    950 
    951 			s = pserialize_read_enter();
    952 			ifa = ifa_ifwithaddr(sintosa(&ipaddr));
    953 			if (ifa == NULL) {
    954 				pserialize_read_exit(s);
    955 				if (opt == IPOPT_SSRR) {
    956 					type = ICMP_UNREACH;
    957 					code = ICMP_UNREACH_SRCFAIL;
    958 					goto bad;
    959 				}
    960 				/*
    961 				 * Loose routing, and not at next destination
    962 				 * yet; nothing to do except forward.
    963 				 */
    964 				break;
    965 			}
    966 			pserialize_read_exit(s);
    967 
    968 			off--;			/* 0 origin */
    969 			if ((off + sizeof(struct in_addr)) > optlen) {
    970 				/*
    971 				 * End of source route.  Should be for us.
    972 				 */
    973 				save_rte(cp, ip->ip_src);
    974 				break;
    975 			}
    976 			/*
    977 			 * locate outgoing interface
    978 			 */
    979 			memcpy((void *)&ipaddr.sin_addr, (void *)(cp + off),
    980 			    sizeof(ipaddr.sin_addr));
    981 			s = pserialize_read_enter();
    982 			if (opt == IPOPT_SSRR) {
    983 				ifa = ifa_ifwithladdr(sintosa(&ipaddr));
    984 				if (ifa != NULL)
    985 					ia = ifatoia(ifa);
    986 				else
    987 					ia = NULL;
    988 			} else {
    989 				ia = ip_rtaddr(ipaddr.sin_addr);
    990 			}
    991 			if (ia == NULL) {
    992 				type = ICMP_UNREACH;
    993 				code = ICMP_UNREACH_SRCFAIL;
    994 				pserialize_read_exit(s);
    995 				goto bad;
    996 			}
    997 			ip->ip_dst = ipaddr.sin_addr;
    998 			bcopy((void *)&ia->ia_addr.sin_addr,
    999 			    (void *)(cp + off), sizeof(struct in_addr));
   1000 			pserialize_read_exit(s);
   1001 			cp[IPOPT_OFFSET] += sizeof(struct in_addr);
   1002 			/*
   1003 			 * Let ip_intr's mcast routing check handle mcast pkts
   1004 			 */
   1005 			forward = !IN_MULTICAST(ip->ip_dst.s_addr);
   1006 			break;
   1007 
   1008 		case IPOPT_RR:
   1009 			if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
   1010 				code = &cp[IPOPT_OLEN] - (u_char *)ip;
   1011 				goto bad;
   1012 			}
   1013 			if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
   1014 				code = &cp[IPOPT_OFFSET] - (u_char *)ip;
   1015 				goto bad;
   1016 			}
   1017 			/*
   1018 			 * If no space remains, ignore.
   1019 			 */
   1020 			off--;			/* 0 origin */
   1021 			if ((off + sizeof(struct in_addr)) > optlen)
   1022 				break;
   1023 			memcpy((void *)&ipaddr.sin_addr, (void *)(&ip->ip_dst),
   1024 			    sizeof(ipaddr.sin_addr));
   1025 			/*
   1026 			 * locate outgoing interface; if we're the destination,
   1027 			 * use the incoming interface (should be same).
   1028 			 */
   1029 			s = pserialize_read_enter();
   1030 			ifa = ifa_ifwithaddr(sintosa(&ipaddr));
   1031 			if (ifa == NULL) {
   1032 				ia = ip_rtaddr(ipaddr.sin_addr);
   1033 				if (ia == NULL) {
   1034 					pserialize_read_exit(s);
   1035 					type = ICMP_UNREACH;
   1036 					code = ICMP_UNREACH_HOST;
   1037 					goto bad;
   1038 				}
   1039 			} else {
   1040 				ia = ifatoia(ifa);
   1041 			}
   1042 			bcopy((void *)&ia->ia_addr.sin_addr,
   1043 			    (void *)(cp + off), sizeof(struct in_addr));
   1044 			pserialize_read_exit(s);
   1045 			cp[IPOPT_OFFSET] += sizeof(struct in_addr);
   1046 			break;
   1047 
   1048 		case IPOPT_TS:
   1049 			code = cp - (u_char *)ip;
   1050 			ipt = (struct ip_timestamp *)cp;
   1051 			if (ipt->ipt_len < 4 || ipt->ipt_len > 40) {
   1052 				code = (u_char *)&ipt->ipt_len - (u_char *)ip;
   1053 				goto bad;
   1054 			}
   1055 			if (ipt->ipt_ptr < 5) {
   1056 				code = (u_char *)&ipt->ipt_ptr - (u_char *)ip;
   1057 				goto bad;
   1058 			}
   1059 			if (ipt->ipt_ptr > ipt->ipt_len - sizeof (int32_t)) {
   1060 				if (++ipt->ipt_oflw == 0) {
   1061 					code = (u_char *)&ipt->ipt_ptr -
   1062 					    (u_char *)ip;
   1063 					goto bad;
   1064 				}
   1065 				break;
   1066 			}
   1067 			cp0 = (cp + ipt->ipt_ptr - 1);
   1068 			switch (ipt->ipt_flg) {
   1069 
   1070 			case IPOPT_TS_TSONLY:
   1071 				break;
   1072 
   1073 			case IPOPT_TS_TSANDADDR: {
   1074 				struct ifnet *rcvif;
   1075 				int _s, _ss;
   1076 
   1077 				if (ipt->ipt_ptr - 1 + sizeof(n_time) +
   1078 				    sizeof(struct in_addr) > ipt->ipt_len) {
   1079 					code = (u_char *)&ipt->ipt_ptr -
   1080 					    (u_char *)ip;
   1081 					goto bad;
   1082 				}
   1083 				ipaddr.sin_addr = dst;
   1084 				_ss = pserialize_read_enter();
   1085 				rcvif = m_get_rcvif(m, &_s);
   1086 				ifa = ifaof_ifpforaddr(sintosa(&ipaddr), rcvif);
   1087 				m_put_rcvif(rcvif, &_s);
   1088 				if (ifa == NULL) {
   1089 					pserialize_read_exit(_ss);
   1090 					break;
   1091 				}
   1092 				ia = ifatoia(ifa);
   1093 				bcopy(&ia->ia_addr.sin_addr,
   1094 				    cp0, sizeof(struct in_addr));
   1095 				pserialize_read_exit(_ss);
   1096 				ipt->ipt_ptr += sizeof(struct in_addr);
   1097 				break;
   1098 			}
   1099 
   1100 			case IPOPT_TS_PRESPEC:
   1101 				if (ipt->ipt_ptr - 1 + sizeof(n_time) +
   1102 				    sizeof(struct in_addr) > ipt->ipt_len) {
   1103 					code = (u_char *)&ipt->ipt_ptr -
   1104 					    (u_char *)ip;
   1105 					goto bad;
   1106 				}
   1107 				memcpy(&ipaddr.sin_addr, cp0,
   1108 				    sizeof(struct in_addr));
   1109 				s = pserialize_read_enter();
   1110 				ifa = ifa_ifwithaddr(sintosa(&ipaddr));
   1111 				if (ifa == NULL) {
   1112 					pserialize_read_exit(s);
   1113 					continue;
   1114 				}
   1115 				pserialize_read_exit(s);
   1116 				ipt->ipt_ptr += sizeof(struct in_addr);
   1117 				break;
   1118 
   1119 			default:
   1120 				/* XXX can't take &ipt->ipt_flg */
   1121 				code = (u_char *)&ipt->ipt_ptr -
   1122 				    (u_char *)ip + 1;
   1123 				goto bad;
   1124 			}
   1125 			ntime = iptime();
   1126 			cp0 = (u_char *) &ntime; /* XXX grumble, GCC... */
   1127 			memmove((char *)cp + ipt->ipt_ptr - 1, cp0,
   1128 			    sizeof(n_time));
   1129 			ipt->ipt_ptr += sizeof(n_time);
   1130 		}
   1131 	}
   1132 	if (forward) {
   1133 		struct ifnet *rcvif;
   1134 		struct psref _psref;
   1135 
   1136 		if (ip_forwsrcrt == 0) {
   1137 			type = ICMP_UNREACH;
   1138 			code = ICMP_UNREACH_SRCFAIL;
   1139 			goto bad;
   1140 		}
   1141 
   1142 		rcvif = m_get_rcvif_psref(m, &_psref);
   1143 		if (__predict_false(rcvif == NULL)) {
   1144 			type = ICMP_UNREACH;
   1145 			code = ICMP_UNREACH_HOST;
   1146 			goto bad;
   1147 		}
   1148 		ip_forward(m, 1, rcvif);
   1149 		m_put_rcvif_psref(rcvif, &_psref);
   1150 		return true;
   1151 	}
   1152 	return false;
   1153 bad:
   1154 	icmp_error(m, type, code, 0, 0);
   1155 	IP_STATINC(IP_STAT_BADOPTIONS);
   1156 	return true;
   1157 }
   1158 
   1159 /*
   1160  * ip_rtaddr: given address of next destination (final or next hop),
   1161  * return internet address info of interface to be used to get there.
   1162  */
   1163 static struct in_ifaddr *
   1164 ip_rtaddr(struct in_addr dst)
   1165 {
   1166 	struct rtentry *rt;
   1167 	union {
   1168 		struct sockaddr		dst;
   1169 		struct sockaddr_in	dst4;
   1170 	} u;
   1171 	struct route *ro;
   1172 
   1173 	sockaddr_in_init(&u.dst4, &dst, 0);
   1174 
   1175 	SOFTNET_LOCK();
   1176 	ro = percpu_getref(ipforward_rt_percpu);
   1177 	rt = rtcache_lookup(ro, &u.dst);
   1178 	percpu_putref(ipforward_rt_percpu);
   1179 	SOFTNET_UNLOCK();
   1180 	if (rt == NULL)
   1181 		return NULL;
   1182 
   1183 	return ifatoia(rt->rt_ifa);
   1184 }
   1185 
   1186 /*
   1187  * save_rte: save incoming source route for use in replies, to be picked
   1188  * up later by ip_srcroute if the receiver is interested.
   1189  */
   1190 static void
   1191 save_rte(u_char *option, struct in_addr dst)
   1192 {
   1193 	unsigned olen;
   1194 
   1195 	olen = option[IPOPT_OLEN];
   1196 	if (olen > sizeof(ip_srcrt) - (1 + sizeof(dst)))
   1197 		return;
   1198 	memcpy((void *)ip_srcrt.srcopt, (void *)option, olen);
   1199 	ip_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr);
   1200 	ip_srcrt.dst = dst;
   1201 }
   1202 
   1203 /*
   1204  * Retrieve incoming source route for use in replies,
   1205  * in the same form used by setsockopt.
   1206  * The first hop is placed before the options, will be removed later.
   1207  */
   1208 struct mbuf *
   1209 ip_srcroute(void)
   1210 {
   1211 	struct in_addr *p, *q;
   1212 	struct mbuf *m;
   1213 
   1214 	if (ip_nhops == 0)
   1215 		return NULL;
   1216 	m = m_get(M_DONTWAIT, MT_SOOPTS);
   1217 	if (m == 0)
   1218 		return NULL;
   1219 
   1220 	MCLAIM(m, &inetdomain.dom_mowner);
   1221 #define OPTSIZ	(sizeof(ip_srcrt.nop) + sizeof(ip_srcrt.srcopt))
   1222 
   1223 	/* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */
   1224 	m->m_len = ip_nhops * sizeof(struct in_addr) + sizeof(struct in_addr) +
   1225 	    OPTSIZ;
   1226 
   1227 	/*
   1228 	 * First save first hop for return route
   1229 	 */
   1230 	p = &ip_srcrt.route[ip_nhops - 1];
   1231 	*(mtod(m, struct in_addr *)) = *p--;
   1232 
   1233 	/*
   1234 	 * Copy option fields and padding (nop) to mbuf.
   1235 	 */
   1236 	ip_srcrt.nop = IPOPT_NOP;
   1237 	ip_srcrt.srcopt[IPOPT_OFFSET] = IPOPT_MINOFF;
   1238 	memmove(mtod(m, char *) + sizeof(struct in_addr), &ip_srcrt.nop,
   1239 	    OPTSIZ);
   1240 	q = (struct in_addr *)(mtod(m, char *) +
   1241 	    sizeof(struct in_addr) + OPTSIZ);
   1242 #undef OPTSIZ
   1243 	/*
   1244 	 * Record return path as an IP source route,
   1245 	 * reversing the path (pointers are now aligned).
   1246 	 */
   1247 	while (p >= ip_srcrt.route) {
   1248 		*q++ = *p--;
   1249 	}
   1250 	/*
   1251 	 * Last hop goes to final destination.
   1252 	 */
   1253 	*q = ip_srcrt.dst;
   1254 	return (m);
   1255 }
   1256 
   1257 const int inetctlerrmap[PRC_NCMDS] = {
   1258 	[PRC_MSGSIZE] = EMSGSIZE,
   1259 	[PRC_HOSTDEAD] = EHOSTDOWN,
   1260 	[PRC_HOSTUNREACH] = EHOSTUNREACH,
   1261 	[PRC_UNREACH_NET] = EHOSTUNREACH,
   1262 	[PRC_UNREACH_HOST] = EHOSTUNREACH,
   1263 	[PRC_UNREACH_PROTOCOL] = ECONNREFUSED,
   1264 	[PRC_UNREACH_PORT] = ECONNREFUSED,
   1265 	[PRC_UNREACH_SRCFAIL] = EHOSTUNREACH,
   1266 	[PRC_PARAMPROB] = ENOPROTOOPT,
   1267 };
   1268 
   1269 void
   1270 ip_fasttimo(void)
   1271 {
   1272 	if (ip_drainwanted) {
   1273 		ip_drain();
   1274 		ip_drainwanted = 0;
   1275 	}
   1276 }
   1277 
   1278 void
   1279 ip_drainstub(void)
   1280 {
   1281 	ip_drainwanted = 1;
   1282 }
   1283 
   1284 /*
   1285  * Forward a packet.  If some error occurs return the sender
   1286  * an icmp packet.  Note we can't always generate a meaningful
   1287  * icmp message because icmp doesn't have a large enough repertoire
   1288  * of codes and types.
   1289  *
   1290  * If not forwarding, just drop the packet.  This could be confusing
   1291  * if ipforwarding was zero but some routing protocol was advancing
   1292  * us as a gateway to somewhere.  However, we must let the routing
   1293  * protocol deal with that.
   1294  *
   1295  * The srcrt parameter indicates whether the packet is being forwarded
   1296  * via a source route.
   1297  */
   1298 static void
   1299 ip_forward(struct mbuf *m, int srcrt, struct ifnet *rcvif)
   1300 {
   1301 	struct ip *ip = mtod(m, struct ip *);
   1302 	struct rtentry *rt;
   1303 	int error, type = 0, code = 0, destmtu = 0;
   1304 	struct mbuf *mcopy;
   1305 	n_long dest;
   1306 	union {
   1307 		struct sockaddr		dst;
   1308 		struct sockaddr_in	dst4;
   1309 	} u;
   1310 	uint64_t *ips;
   1311 	struct route *ro;
   1312 
   1313 	KASSERTMSG(cpu_softintr_p(), "ip_forward: not in the software "
   1314 	    "interrupt handler; synchronization assumptions violated");
   1315 
   1316 	/*
   1317 	 * We are now in the output path.
   1318 	 */
   1319 	MCLAIM(m, &ip_tx_mowner);
   1320 
   1321 	/*
   1322 	 * Clear any in-bound checksum flags for this packet.
   1323 	 */
   1324 	m->m_pkthdr.csum_flags = 0;
   1325 
   1326 	dest = 0;
   1327 	if (m->m_flags & (M_BCAST|M_MCAST) || in_canforward(ip->ip_dst) == 0) {
   1328 		IP_STATINC(IP_STAT_CANTFORWARD);
   1329 		m_freem(m);
   1330 		return;
   1331 	}
   1332 
   1333 	SOFTNET_LOCK();
   1334 
   1335 	if (ip->ip_ttl <= IPTTLDEC) {
   1336 		icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0);
   1337 		SOFTNET_UNLOCK();
   1338 		return;
   1339 	}
   1340 
   1341 	sockaddr_in_init(&u.dst4, &ip->ip_dst, 0);
   1342 
   1343 	ro = percpu_getref(ipforward_rt_percpu);
   1344 	if ((rt = rtcache_lookup(ro, &u.dst)) == NULL) {
   1345 		percpu_putref(ipforward_rt_percpu);
   1346 		icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_NET, dest, 0);
   1347 		SOFTNET_UNLOCK();
   1348 		return;
   1349 	}
   1350 
   1351 	/*
   1352 	 * Save at most 68 bytes of the packet in case
   1353 	 * we need to generate an ICMP message to the src.
   1354 	 * Pullup to avoid sharing mbuf cluster between m and mcopy.
   1355 	 */
   1356 	mcopy = m_copym(m, 0, imin(ntohs(ip->ip_len), 68), M_DONTWAIT);
   1357 	if (mcopy)
   1358 		mcopy = m_pullup(mcopy, ip->ip_hl << 2);
   1359 
   1360 	ip->ip_ttl -= IPTTLDEC;
   1361 
   1362 	/*
   1363 	 * If forwarding packet using same interface that it came in on,
   1364 	 * perhaps should send a redirect to sender to shortcut a hop.
   1365 	 * Only send redirect if source is sending directly to us,
   1366 	 * and if packet was not source routed (or has any options).
   1367 	 * Also, don't send redirect if forwarding using a default route
   1368 	 * or a route modified by a redirect.
   1369 	 */
   1370 	if (rt->rt_ifp == rcvif &&
   1371 	    (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 &&
   1372 	    !in_nullhost(satocsin(rt_getkey(rt))->sin_addr) &&
   1373 	    ipsendredirects && !srcrt) {
   1374 		if (rt->rt_ifa &&
   1375 		    (ip->ip_src.s_addr & ifatoia(rt->rt_ifa)->ia_subnetmask) ==
   1376 		    ifatoia(rt->rt_ifa)->ia_subnet) {
   1377 			if (rt->rt_flags & RTF_GATEWAY)
   1378 				dest = satosin(rt->rt_gateway)->sin_addr.s_addr;
   1379 			else
   1380 				dest = ip->ip_dst.s_addr;
   1381 			/*
   1382 			 * Router requirements says to only send host
   1383 			 * redirects.
   1384 			 */
   1385 			type = ICMP_REDIRECT;
   1386 			code = ICMP_REDIRECT_HOST;
   1387 		}
   1388 	}
   1389 
   1390 	error = ip_output(m, NULL, ro,
   1391 	    (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)),
   1392 	    NULL, NULL);
   1393 
   1394 	if (error) {
   1395 		IP_STATINC(IP_STAT_CANTFORWARD);
   1396 		goto error;
   1397 	}
   1398 
   1399 	ips = IP_STAT_GETREF();
   1400 	ips[IP_STAT_FORWARD]++;
   1401 
   1402 	if (type) {
   1403 		ips[IP_STAT_REDIRECTSENT]++;
   1404 		IP_STAT_PUTREF();
   1405 		goto redirect;
   1406 	}
   1407 
   1408 	IP_STAT_PUTREF();
   1409 	if (mcopy) {
   1410 #ifdef GATEWAY
   1411 		if (mcopy->m_flags & M_CANFASTFWD)
   1412 			ipflow_create(ro, mcopy);
   1413 #endif
   1414 		m_freem(mcopy);
   1415 	}
   1416 
   1417 	percpu_putref(ipforward_rt_percpu);
   1418 	SOFTNET_UNLOCK();
   1419 	return;
   1420 
   1421 redirect:
   1422 error:
   1423 	if (mcopy == NULL) {
   1424 		percpu_putref(ipforward_rt_percpu);
   1425 		SOFTNET_UNLOCK();
   1426 		return;
   1427 	}
   1428 
   1429 	switch (error) {
   1430 
   1431 	case 0:				/* forwarded, but need redirect */
   1432 		/* type, code set above */
   1433 		break;
   1434 
   1435 	case ENETUNREACH:		/* shouldn't happen, checked above */
   1436 	case EHOSTUNREACH:
   1437 	case ENETDOWN:
   1438 	case EHOSTDOWN:
   1439 	default:
   1440 		type = ICMP_UNREACH;
   1441 		code = ICMP_UNREACH_HOST;
   1442 		break;
   1443 
   1444 	case EMSGSIZE:
   1445 		type = ICMP_UNREACH;
   1446 		code = ICMP_UNREACH_NEEDFRAG;
   1447 
   1448 		if ((rt = rtcache_validate(ro)) != NULL)
   1449 			destmtu = rt->rt_ifp->if_mtu;
   1450 #ifdef IPSEC
   1451 		if (ipsec_used)
   1452 			(void)ipsec4_forward(mcopy, &destmtu);
   1453 #endif
   1454 		IP_STATINC(IP_STAT_CANTFRAG);
   1455 		break;
   1456 
   1457 	case ENOBUFS:
   1458 		/*
   1459 		 * Do not generate ICMP_SOURCEQUENCH as required in RFC 1812,
   1460 		 * Requirements for IP Version 4 Routers.  Source quench can
   1461 		 * big problem under DoS attacks or if the underlying
   1462 		 * interface is rate-limited.
   1463 		 */
   1464 		if (mcopy)
   1465 			m_freem(mcopy);
   1466 		percpu_putref(ipforward_rt_percpu);
   1467 		SOFTNET_UNLOCK();
   1468 		return;
   1469 	}
   1470 	icmp_error(mcopy, type, code, dest, destmtu);
   1471 	percpu_putref(ipforward_rt_percpu);
   1472 	SOFTNET_UNLOCK();
   1473 }
   1474 
   1475 void
   1476 ip_savecontrol(struct inpcb *inp, struct mbuf **mp, struct ip *ip,
   1477     struct mbuf *m)
   1478 {
   1479 	struct socket *so = inp->inp_socket;
   1480 	ifnet_t *ifp;
   1481 	int inpflags = inp->inp_flags;
   1482 	struct psref psref;
   1483 
   1484 	ifp = m_get_rcvif_psref(m, &psref);
   1485 	if (__predict_false(ifp == NULL))
   1486 		return; /* XXX should report error? */
   1487 
   1488 	if (so->so_options & SO_TIMESTAMP
   1489 #ifdef SO_OTIMESTAMP
   1490 	    || so->so_options & SO_OTIMESTAMP
   1491 #endif
   1492 	    ) {
   1493 		struct timeval tv;
   1494 
   1495 		microtime(&tv);
   1496 #ifdef SO_OTIMESTAMP
   1497 		if (so->so_options & SO_OTIMESTAMP) {
   1498 			struct timeval50 tv50;
   1499 			timeval_to_timeval50(&tv, &tv50);
   1500 			*mp = sbcreatecontrol((void *) &tv50, sizeof(tv50),
   1501 			    SCM_OTIMESTAMP, SOL_SOCKET);
   1502 		} else
   1503 #endif
   1504 		*mp = sbcreatecontrol((void *) &tv, sizeof(tv),
   1505 		    SCM_TIMESTAMP, SOL_SOCKET);
   1506 		if (*mp)
   1507 			mp = &(*mp)->m_next;
   1508 	}
   1509 	if (inpflags & INP_RECVDSTADDR) {
   1510 		*mp = sbcreatecontrol((void *) &ip->ip_dst,
   1511 		    sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP);
   1512 		if (*mp)
   1513 			mp = &(*mp)->m_next;
   1514 	}
   1515 	if (inpflags & INP_RECVPKTINFO) {
   1516 		struct in_pktinfo ipi;
   1517 		ipi.ipi_addr = ip->ip_src;
   1518 		ipi.ipi_ifindex = ifp->if_index;
   1519 		*mp = sbcreatecontrol((void *) &ipi,
   1520 		    sizeof(ipi), IP_RECVPKTINFO, IPPROTO_IP);
   1521 		if (*mp)
   1522 			mp = &(*mp)->m_next;
   1523 	}
   1524 	if (inpflags & INP_PKTINFO) {
   1525 		struct in_pktinfo ipi;
   1526 		ipi.ipi_addr = ip->ip_dst;
   1527 		ipi.ipi_ifindex = ifp->if_index;
   1528 		*mp = sbcreatecontrol((void *) &ipi,
   1529 		    sizeof(ipi), IP_PKTINFO, IPPROTO_IP);
   1530 		if (*mp)
   1531 			mp = &(*mp)->m_next;
   1532 	}
   1533 	if (inpflags & INP_RECVIF) {
   1534 		struct sockaddr_dl sdl;
   1535 
   1536 		sockaddr_dl_init(&sdl, sizeof(sdl), ifp->if_index, 0, NULL, 0,
   1537 		    NULL, 0);
   1538 		*mp = sbcreatecontrol(&sdl, sdl.sdl_len, IP_RECVIF, IPPROTO_IP);
   1539 		if (*mp)
   1540 			mp = &(*mp)->m_next;
   1541 	}
   1542 	if (inpflags & INP_RECVTTL) {
   1543 		*mp = sbcreatecontrol((void *) &ip->ip_ttl,
   1544 		    sizeof(uint8_t), IP_RECVTTL, IPPROTO_IP);
   1545 		if (*mp)
   1546 			mp = &(*mp)->m_next;
   1547 	}
   1548 	m_put_rcvif_psref(ifp, &psref);
   1549 }
   1550 
   1551 /*
   1552  * sysctl helper routine for net.inet.ip.forwsrcrt.
   1553  */
   1554 static int
   1555 sysctl_net_inet_ip_forwsrcrt(SYSCTLFN_ARGS)
   1556 {
   1557 	int error, tmp;
   1558 	struct sysctlnode node;
   1559 
   1560 	node = *rnode;
   1561 	tmp = ip_forwsrcrt;
   1562 	node.sysctl_data = &tmp;
   1563 	error = sysctl_lookup(SYSCTLFN_CALL(&node));
   1564 	if (error || newp == NULL)
   1565 		return (error);
   1566 
   1567 	error = kauth_authorize_network(l->l_cred, KAUTH_NETWORK_FORWSRCRT,
   1568 	    0, NULL, NULL, NULL);
   1569 	if (error)
   1570 		return (error);
   1571 
   1572 	ip_forwsrcrt = tmp;
   1573 
   1574 	return (0);
   1575 }
   1576 
   1577 /*
   1578  * sysctl helper routine for net.inet.ip.mtudisctimeout.  checks the
   1579  * range of the new value and tweaks timers if it changes.
   1580  */
   1581 static int
   1582 sysctl_net_inet_ip_pmtudto(SYSCTLFN_ARGS)
   1583 {
   1584 	int error, tmp;
   1585 	struct sysctlnode node;
   1586 
   1587 	node = *rnode;
   1588 	tmp = ip_mtudisc_timeout;
   1589 	node.sysctl_data = &tmp;
   1590 	error = sysctl_lookup(SYSCTLFN_CALL(&node));
   1591 	if (error || newp == NULL)
   1592 		return (error);
   1593 	if (tmp < 0)
   1594 		return (EINVAL);
   1595 
   1596 	mutex_enter(softnet_lock);
   1597 
   1598 	ip_mtudisc_timeout = tmp;
   1599 	rt_timer_queue_change(ip_mtudisc_timeout_q, ip_mtudisc_timeout);
   1600 
   1601 	mutex_exit(softnet_lock);
   1602 
   1603 	return (0);
   1604 }
   1605 
   1606 static int
   1607 sysctl_net_inet_ip_stats(SYSCTLFN_ARGS)
   1608 {
   1609 
   1610 	return (NETSTAT_SYSCTL(ipstat_percpu, IP_NSTATS));
   1611 }
   1612 
   1613 static void
   1614 sysctl_net_inet_ip_setup(struct sysctllog **clog)
   1615 {
   1616 	sysctl_createv(clog, 0, NULL, NULL,
   1617 		       CTLFLAG_PERMANENT,
   1618 		       CTLTYPE_NODE, "inet",
   1619 		       SYSCTL_DESCR("PF_INET related settings"),
   1620 		       NULL, 0, NULL, 0,
   1621 		       CTL_NET, PF_INET, CTL_EOL);
   1622 	sysctl_createv(clog, 0, NULL, NULL,
   1623 		       CTLFLAG_PERMANENT,
   1624 		       CTLTYPE_NODE, "ip",
   1625 		       SYSCTL_DESCR("IPv4 related settings"),
   1626 		       NULL, 0, NULL, 0,
   1627 		       CTL_NET, PF_INET, IPPROTO_IP, CTL_EOL);
   1628 
   1629 	sysctl_createv(clog, 0, NULL, NULL,
   1630 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1631 		       CTLTYPE_INT, "forwarding",
   1632 		       SYSCTL_DESCR("Enable forwarding of INET datagrams"),
   1633 		       NULL, 0, &ipforwarding, 0,
   1634 		       CTL_NET, PF_INET, IPPROTO_IP,
   1635 		       IPCTL_FORWARDING, CTL_EOL);
   1636 	sysctl_createv(clog, 0, NULL, NULL,
   1637 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1638 		       CTLTYPE_INT, "redirect",
   1639 		       SYSCTL_DESCR("Enable sending of ICMP redirect messages"),
   1640 		       NULL, 0, &ipsendredirects, 0,
   1641 		       CTL_NET, PF_INET, IPPROTO_IP,
   1642 		       IPCTL_SENDREDIRECTS, CTL_EOL);
   1643 	sysctl_createv(clog, 0, NULL, NULL,
   1644 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1645 		       CTLTYPE_INT, "ttl",
   1646 		       SYSCTL_DESCR("Default TTL for an INET datagram"),
   1647 		       NULL, 0, &ip_defttl, 0,
   1648 		       CTL_NET, PF_INET, IPPROTO_IP,
   1649 		       IPCTL_DEFTTL, CTL_EOL);
   1650 #ifdef IPCTL_DEFMTU
   1651 	sysctl_createv(clog, 0, NULL, NULL,
   1652 		       CTLFLAG_PERMANENT /* |CTLFLAG_READWRITE? */,
   1653 		       CTLTYPE_INT, "mtu",
   1654 		       SYSCTL_DESCR("Default MTA for an INET route"),
   1655 		       NULL, 0, &ip_mtu, 0,
   1656 		       CTL_NET, PF_INET, IPPROTO_IP,
   1657 		       IPCTL_DEFMTU, CTL_EOL);
   1658 #endif /* IPCTL_DEFMTU */
   1659 	sysctl_createv(clog, 0, NULL, NULL,
   1660 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1661 		       CTLTYPE_INT, "forwsrcrt",
   1662 		       SYSCTL_DESCR("Enable forwarding of source-routed "
   1663 				    "datagrams"),
   1664 		       sysctl_net_inet_ip_forwsrcrt, 0, &ip_forwsrcrt, 0,
   1665 		       CTL_NET, PF_INET, IPPROTO_IP,
   1666 		       IPCTL_FORWSRCRT, CTL_EOL);
   1667 	sysctl_createv(clog, 0, NULL, NULL,
   1668 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1669 		       CTLTYPE_INT, "directed-broadcast",
   1670 		       SYSCTL_DESCR("Enable forwarding of broadcast datagrams"),
   1671 		       NULL, 0, &ip_directedbcast, 0,
   1672 		       CTL_NET, PF_INET, IPPROTO_IP,
   1673 		       IPCTL_DIRECTEDBCAST, CTL_EOL);
   1674 	sysctl_createv(clog, 0, NULL, NULL,
   1675 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1676 		       CTLTYPE_INT, "allowsrcrt",
   1677 		       SYSCTL_DESCR("Accept source-routed datagrams"),
   1678 		       NULL, 0, &ip_allowsrcrt, 0,
   1679 		       CTL_NET, PF_INET, IPPROTO_IP,
   1680 		       IPCTL_ALLOWSRCRT, CTL_EOL);
   1681 
   1682 	sysctl_createv(clog, 0, NULL, NULL,
   1683 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1684 		       CTLTYPE_INT, "mtudisc",
   1685 		       SYSCTL_DESCR("Use RFC1191 Path MTU Discovery"),
   1686 		       NULL, 0, &ip_mtudisc, 0,
   1687 		       CTL_NET, PF_INET, IPPROTO_IP,
   1688 		       IPCTL_MTUDISC, CTL_EOL);
   1689 	sysctl_createv(clog, 0, NULL, NULL,
   1690 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1691 		       CTLTYPE_INT, "anonportmin",
   1692 		       SYSCTL_DESCR("Lowest ephemeral port number to assign"),
   1693 		       sysctl_net_inet_ip_ports, 0, &anonportmin, 0,
   1694 		       CTL_NET, PF_INET, IPPROTO_IP,
   1695 		       IPCTL_ANONPORTMIN, CTL_EOL);
   1696 	sysctl_createv(clog, 0, NULL, NULL,
   1697 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1698 		       CTLTYPE_INT, "anonportmax",
   1699 		       SYSCTL_DESCR("Highest ephemeral port number to assign"),
   1700 		       sysctl_net_inet_ip_ports, 0, &anonportmax, 0,
   1701 		       CTL_NET, PF_INET, IPPROTO_IP,
   1702 		       IPCTL_ANONPORTMAX, CTL_EOL);
   1703 	sysctl_createv(clog, 0, NULL, NULL,
   1704 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1705 		       CTLTYPE_INT, "mtudisctimeout",
   1706 		       SYSCTL_DESCR("Lifetime of a Path MTU Discovered route"),
   1707 		       sysctl_net_inet_ip_pmtudto, 0, (void *)&ip_mtudisc_timeout, 0,
   1708 		       CTL_NET, PF_INET, IPPROTO_IP,
   1709 		       IPCTL_MTUDISCTIMEOUT, CTL_EOL);
   1710 #ifndef IPNOPRIVPORTS
   1711 	sysctl_createv(clog, 0, NULL, NULL,
   1712 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1713 		       CTLTYPE_INT, "lowportmin",
   1714 		       SYSCTL_DESCR("Lowest privileged ephemeral port number "
   1715 				    "to assign"),
   1716 		       sysctl_net_inet_ip_ports, 0, &lowportmin, 0,
   1717 		       CTL_NET, PF_INET, IPPROTO_IP,
   1718 		       IPCTL_LOWPORTMIN, CTL_EOL);
   1719 	sysctl_createv(clog, 0, NULL, NULL,
   1720 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1721 		       CTLTYPE_INT, "lowportmax",
   1722 		       SYSCTL_DESCR("Highest privileged ephemeral port number "
   1723 				    "to assign"),
   1724 		       sysctl_net_inet_ip_ports, 0, &lowportmax, 0,
   1725 		       CTL_NET, PF_INET, IPPROTO_IP,
   1726 		       IPCTL_LOWPORTMAX, CTL_EOL);
   1727 #endif /* IPNOPRIVPORTS */
   1728 #if NGRE > 0
   1729 	sysctl_createv(clog, 0, NULL, NULL,
   1730 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1731 		       CTLTYPE_INT, "grettl",
   1732 		       SYSCTL_DESCR("Default TTL for a gre tunnel datagram"),
   1733 		       NULL, 0, &ip_gre_ttl, 0,
   1734 		       CTL_NET, PF_INET, IPPROTO_IP,
   1735 		       IPCTL_GRE_TTL, CTL_EOL);
   1736 #endif /* NGRE */
   1737 	sysctl_createv(clog, 0, NULL, NULL,
   1738 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1739 		       CTLTYPE_INT, "checkinterface",
   1740 		       SYSCTL_DESCR("Enable receive side of Strong ES model "
   1741 				    "from RFC1122"),
   1742 		       NULL, 0, &ip_checkinterface, 0,
   1743 		       CTL_NET, PF_INET, IPPROTO_IP,
   1744 		       IPCTL_CHECKINTERFACE, CTL_EOL);
   1745 	sysctl_createv(clog, 0, NULL, NULL,
   1746 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1747 		       CTLTYPE_INT, "random_id",
   1748 		       SYSCTL_DESCR("Assign random ip_id values"),
   1749 		       NULL, 0, &ip_do_randomid, 0,
   1750 		       CTL_NET, PF_INET, IPPROTO_IP,
   1751 		       IPCTL_RANDOMID, CTL_EOL);
   1752 	sysctl_createv(clog, 0, NULL, NULL,
   1753 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1754 		       CTLTYPE_INT, "do_loopback_cksum",
   1755 		       SYSCTL_DESCR("Perform IP checksum on loopback"),
   1756 		       NULL, 0, &ip_do_loopback_cksum, 0,
   1757 		       CTL_NET, PF_INET, IPPROTO_IP,
   1758 		       IPCTL_LOOPBACKCKSUM, CTL_EOL);
   1759 	sysctl_createv(clog, 0, NULL, NULL,
   1760 		       CTLFLAG_PERMANENT,
   1761 		       CTLTYPE_STRUCT, "stats",
   1762 		       SYSCTL_DESCR("IP statistics"),
   1763 		       sysctl_net_inet_ip_stats, 0, NULL, 0,
   1764 		       CTL_NET, PF_INET, IPPROTO_IP, IPCTL_STATS,
   1765 		       CTL_EOL);
   1766 #if NARP
   1767 	sysctl_createv(clog, 0, NULL, NULL,
   1768 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1769 		       CTLTYPE_INT, "dad_count",
   1770 		       SYSCTL_DESCR("Number of Duplicate Address Detection "
   1771 				    "probes to send"),
   1772 		       NULL, 0, &ip_dad_count, 0,
   1773 		       CTL_NET, PF_INET, IPPROTO_IP,
   1774 		       IPCTL_DAD_COUNT, CTL_EOL);
   1775 #endif
   1776 
   1777 	/* anonportalgo RFC6056 subtree */
   1778 	const struct sysctlnode *portalgo_node;
   1779 	sysctl_createv(clog, 0, NULL, &portalgo_node,
   1780 		       CTLFLAG_PERMANENT,
   1781 		       CTLTYPE_NODE, "anonportalgo",
   1782 		       SYSCTL_DESCR("Anonymous Port Algorithm Selection (RFC 6056)"),
   1783 	    	       NULL, 0, NULL, 0,
   1784 		       CTL_NET, PF_INET, IPPROTO_IP, CTL_CREATE, CTL_EOL);
   1785 	sysctl_createv(clog, 0, &portalgo_node, NULL,
   1786 		       CTLFLAG_PERMANENT,
   1787 		       CTLTYPE_STRING, "available",
   1788 		       SYSCTL_DESCR("available algorithms"),
   1789 		       sysctl_portalgo_available, 0, NULL, PORTALGO_MAXLEN,
   1790 		       CTL_CREATE, CTL_EOL);
   1791 	sysctl_createv(clog, 0, &portalgo_node, NULL,
   1792 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1793 		       CTLTYPE_STRING, "selected",
   1794 		       SYSCTL_DESCR("selected algorithm"),
   1795 		       sysctl_portalgo_selected4, 0, NULL, PORTALGO_MAXLEN,
   1796 		       CTL_CREATE, CTL_EOL);
   1797 	sysctl_createv(clog, 0, &portalgo_node, NULL,
   1798 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
   1799 		       CTLTYPE_STRUCT, "reserve",
   1800 		       SYSCTL_DESCR("bitmap of reserved ports"),
   1801 		       sysctl_portalgo_reserve4, 0, NULL, 0,
   1802 		       CTL_CREATE, CTL_EOL);
   1803 }
   1804 
   1805 void
   1806 ip_statinc(u_int stat)
   1807 {
   1808 
   1809 	KASSERT(stat < IP_NSTATS);
   1810 	IP_STATINC(stat);
   1811 }
   1812