Home | History | Annotate | Line # | Download | only in netinet
ip_input.c revision 1.156
      1 /*	$NetBSD: ip_input.c,v 1.156 2002/09/11 02:41:19 itojun 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  * 3. All advertising materials mentioning features or use of this software
     49  *    must display the following acknowledgement:
     50  *	This product includes software developed by the NetBSD
     51  *	Foundation, Inc. and its contributors.
     52  * 4. Neither the name of The NetBSD Foundation nor the names of its
     53  *    contributors may be used to endorse or promote products derived
     54  *    from this software without specific prior written permission.
     55  *
     56  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
     57  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
     58  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
     59  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
     60  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
     61  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
     62  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
     63  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
     64  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
     65  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
     66  * POSSIBILITY OF SUCH DAMAGE.
     67  */
     68 
     69 /*
     70  * Copyright (c) 1982, 1986, 1988, 1993
     71  *	The Regents of the University of California.  All rights reserved.
     72  *
     73  * Redistribution and use in source and binary forms, with or without
     74  * modification, are permitted provided that the following conditions
     75  * are met:
     76  * 1. Redistributions of source code must retain the above copyright
     77  *    notice, this list of conditions and the following disclaimer.
     78  * 2. Redistributions in binary form must reproduce the above copyright
     79  *    notice, this list of conditions and the following disclaimer in the
     80  *    documentation and/or other materials provided with the distribution.
     81  * 3. All advertising materials mentioning features or use of this software
     82  *    must display the following acknowledgement:
     83  *	This product includes software developed by the University of
     84  *	California, Berkeley and its contributors.
     85  * 4. Neither the name of the University nor the names of its contributors
     86  *    may be used to endorse or promote products derived from this software
     87  *    without specific prior written permission.
     88  *
     89  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     90  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     91  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     92  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     93  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     94  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     95  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     96  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     97  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     98  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     99  * SUCH DAMAGE.
    100  *
    101  *	@(#)ip_input.c	8.2 (Berkeley) 1/4/94
    102  */
    103 
    104 #include <sys/cdefs.h>
    105 __KERNEL_RCSID(0, "$NetBSD: ip_input.c,v 1.156 2002/09/11 02:41:19 itojun Exp $");
    106 
    107 #include "opt_gateway.h"
    108 #include "opt_pfil_hooks.h"
    109 #include "opt_ipsec.h"
    110 #include "opt_mrouting.h"
    111 #include "opt_inet_csum.h"
    112 
    113 #include <sys/param.h>
    114 #include <sys/systm.h>
    115 #include <sys/malloc.h>
    116 #include <sys/mbuf.h>
    117 #include <sys/domain.h>
    118 #include <sys/protosw.h>
    119 #include <sys/socket.h>
    120 #include <sys/socketvar.h>
    121 #include <sys/errno.h>
    122 #include <sys/time.h>
    123 #include <sys/kernel.h>
    124 #include <sys/pool.h>
    125 #include <sys/sysctl.h>
    126 
    127 #include <net/if.h>
    128 #include <net/if_dl.h>
    129 #include <net/route.h>
    130 #include <net/pfil.h>
    131 
    132 #include <netinet/in.h>
    133 #include <netinet/in_systm.h>
    134 #include <netinet/ip.h>
    135 #include <netinet/in_pcb.h>
    136 #include <netinet/in_var.h>
    137 #include <netinet/ip_var.h>
    138 #include <netinet/ip_icmp.h>
    139 /* just for gif_ttl */
    140 #include <netinet/in_gif.h>
    141 #include "gif.h"
    142 #include <net/if_gre.h>
    143 #include "gre.h"
    144 
    145 #ifdef MROUTING
    146 #include <netinet/ip_mroute.h>
    147 #endif
    148 
    149 #ifdef IPSEC
    150 #include <netinet6/ipsec.h>
    151 #include <netkey/key.h>
    152 #endif
    153 
    154 #ifndef	IPFORWARDING
    155 #ifdef GATEWAY
    156 #define	IPFORWARDING	1	/* forward IP packets not for us */
    157 #else /* GATEWAY */
    158 #define	IPFORWARDING	0	/* don't forward IP packets not for us */
    159 #endif /* GATEWAY */
    160 #endif /* IPFORWARDING */
    161 #ifndef	IPSENDREDIRECTS
    162 #define	IPSENDREDIRECTS	1
    163 #endif
    164 #ifndef IPFORWSRCRT
    165 #define	IPFORWSRCRT	1	/* forward source-routed packets */
    166 #endif
    167 #ifndef IPALLOWSRCRT
    168 #define	IPALLOWSRCRT	1	/* allow source-routed packets */
    169 #endif
    170 #ifndef IPMTUDISC
    171 #define IPMTUDISC	1
    172 #endif
    173 #ifndef IPMTUDISCTIMEOUT
    174 #define IPMTUDISCTIMEOUT (10 * 60)	/* as per RFC 1191 */
    175 #endif
    176 
    177 /*
    178  * Note: DIRECTED_BROADCAST is handled this way so that previous
    179  * configuration using this option will Just Work.
    180  */
    181 #ifndef IPDIRECTEDBCAST
    182 #ifdef DIRECTED_BROADCAST
    183 #define IPDIRECTEDBCAST	1
    184 #else
    185 #define	IPDIRECTEDBCAST	0
    186 #endif /* DIRECTED_BROADCAST */
    187 #endif /* IPDIRECTEDBCAST */
    188 int	ipforwarding = IPFORWARDING;
    189 int	ipsendredirects = IPSENDREDIRECTS;
    190 int	ip_defttl = IPDEFTTL;
    191 int	ip_forwsrcrt = IPFORWSRCRT;
    192 int	ip_directedbcast = IPDIRECTEDBCAST;
    193 int	ip_allowsrcrt = IPALLOWSRCRT;
    194 int	ip_mtudisc = IPMTUDISC;
    195 int	ip_mtudisc_timeout = IPMTUDISCTIMEOUT;
    196 #ifdef DIAGNOSTIC
    197 int	ipprintfs = 0;
    198 #endif
    199 
    200 struct rttimer_queue *ip_mtudisc_timeout_q = NULL;
    201 
    202 extern	struct domain inetdomain;
    203 int	ipqmaxlen = IFQ_MAXLEN;
    204 u_long	in_ifaddrhash;				/* size of hash table - 1 */
    205 int	in_ifaddrentries;			/* total number of addrs */
    206 struct	in_ifaddrhead in_ifaddr;
    207 struct	in_ifaddrhashhead *in_ifaddrhashtbl;
    208 struct	ifqueue ipintrq;
    209 struct	ipstat	ipstat;
    210 u_int16_t	ip_id;
    211 
    212 #ifdef PFIL_HOOKS
    213 struct pfil_head inet_pfil_hook;
    214 #endif
    215 
    216 struct ipqhead ipq;
    217 int	ipq_locked;
    218 int	ip_nfragpackets = 0;
    219 int	ip_maxfragpackets = 200;
    220 
    221 static __inline int ipq_lock_try __P((void));
    222 static __inline void ipq_unlock __P((void));
    223 
    224 static __inline int
    225 ipq_lock_try()
    226 {
    227 	int s;
    228 
    229 	/*
    230 	 * Use splvm() -- we're blocking things that would cause
    231 	 * mbuf allocation.
    232 	 */
    233 	s = splvm();
    234 	if (ipq_locked) {
    235 		splx(s);
    236 		return (0);
    237 	}
    238 	ipq_locked = 1;
    239 	splx(s);
    240 	return (1);
    241 }
    242 
    243 static __inline void
    244 ipq_unlock()
    245 {
    246 	int s;
    247 
    248 	s = splvm();
    249 	ipq_locked = 0;
    250 	splx(s);
    251 }
    252 
    253 #ifdef DIAGNOSTIC
    254 #define	IPQ_LOCK()							\
    255 do {									\
    256 	if (ipq_lock_try() == 0) {					\
    257 		printf("%s:%d: ipq already locked\n", __FILE__, __LINE__); \
    258 		panic("ipq_lock");					\
    259 	}								\
    260 } while (0)
    261 #define	IPQ_LOCK_CHECK()						\
    262 do {									\
    263 	if (ipq_locked == 0) {						\
    264 		printf("%s:%d: ipq lock not held\n", __FILE__, __LINE__); \
    265 		panic("ipq lock check");				\
    266 	}								\
    267 } while (0)
    268 #else
    269 #define	IPQ_LOCK()		(void) ipq_lock_try()
    270 #define	IPQ_LOCK_CHECK()	/* nothing */
    271 #endif
    272 
    273 #define	IPQ_UNLOCK()		ipq_unlock()
    274 
    275 struct pool ipqent_pool;
    276 
    277 #ifdef INET_CSUM_COUNTERS
    278 #include <sys/device.h>
    279 
    280 struct evcnt ip_hwcsum_bad = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
    281     NULL, "inet", "hwcsum bad");
    282 struct evcnt ip_hwcsum_ok = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
    283     NULL, "inet", "hwcsum ok");
    284 struct evcnt ip_swcsum = EVCNT_INITIALIZER(EVCNT_TYPE_MISC,
    285     NULL, "inet", "swcsum");
    286 
    287 #define	INET_CSUM_COUNTER_INCR(ev)	(ev)->ev_count++
    288 
    289 #else
    290 
    291 #define	INET_CSUM_COUNTER_INCR(ev)	/* nothing */
    292 
    293 #endif /* INET_CSUM_COUNTERS */
    294 
    295 /*
    296  * We need to save the IP options in case a protocol wants to respond
    297  * to an incoming packet over the same route if the packet got here
    298  * using IP source routing.  This allows connection establishment and
    299  * maintenance when the remote end is on a network that is not known
    300  * to us.
    301  */
    302 int	ip_nhops = 0;
    303 static	struct ip_srcrt {
    304 	struct	in_addr dst;			/* final destination */
    305 	char	nop;				/* one NOP to align */
    306 	char	srcopt[IPOPT_OFFSET + 1];	/* OPTVAL, OLEN and OFFSET */
    307 	struct	in_addr route[MAX_IPOPTLEN/sizeof(struct in_addr)];
    308 } ip_srcrt;
    309 
    310 static void save_rte __P((u_char *, struct in_addr));
    311 
    312 /*
    313  * IP initialization: fill in IP protocol switch table.
    314  * All protocols not implemented in kernel go to raw IP protocol handler.
    315  */
    316 void
    317 ip_init()
    318 {
    319 	struct protosw *pr;
    320 	int i;
    321 
    322 	pool_init(&ipqent_pool, sizeof(struct ipqent), 0, 0, 0, "ipqepl",
    323 	    NULL);
    324 
    325 	pr = pffindproto(PF_INET, IPPROTO_RAW, SOCK_RAW);
    326 	if (pr == 0)
    327 		panic("ip_init");
    328 	for (i = 0; i < IPPROTO_MAX; i++)
    329 		ip_protox[i] = pr - inetsw;
    330 	for (pr = inetdomain.dom_protosw;
    331 	    pr < inetdomain.dom_protoswNPROTOSW; pr++)
    332 		if (pr->pr_domain->dom_family == PF_INET &&
    333 		    pr->pr_protocol && pr->pr_protocol != IPPROTO_RAW)
    334 			ip_protox[pr->pr_protocol] = pr - inetsw;
    335 	LIST_INIT(&ipq);
    336 	ip_id = time.tv_sec & 0xffff;
    337 	ipintrq.ifq_maxlen = ipqmaxlen;
    338 	TAILQ_INIT(&in_ifaddr);
    339 	in_ifaddrhashtbl = hashinit(IN_IFADDR_HASH_SIZE, HASH_LIST, M_IFADDR,
    340 	    M_WAITOK, &in_ifaddrhash);
    341 	if (ip_mtudisc != 0)
    342 		ip_mtudisc_timeout_q =
    343 		    rt_timer_queue_create(ip_mtudisc_timeout);
    344 #ifdef GATEWAY
    345 	ipflow_init();
    346 #endif
    347 
    348 #ifdef PFIL_HOOKS
    349 	/* Register our Packet Filter hook. */
    350 	inet_pfil_hook.ph_type = PFIL_TYPE_AF;
    351 	inet_pfil_hook.ph_af   = AF_INET;
    352 	i = pfil_head_register(&inet_pfil_hook);
    353 	if (i != 0)
    354 		printf("ip_init: WARNING: unable to register pfil hook, "
    355 		    "error %d\n", i);
    356 #endif /* PFIL_HOOKS */
    357 
    358 #ifdef INET_CSUM_COUNTERS
    359 	evcnt_attach_static(&ip_hwcsum_bad);
    360 	evcnt_attach_static(&ip_hwcsum_ok);
    361 	evcnt_attach_static(&ip_swcsum);
    362 #endif /* INET_CSUM_COUNTERS */
    363 }
    364 
    365 struct	sockaddr_in ipaddr = { sizeof(ipaddr), AF_INET };
    366 struct	route ipforward_rt;
    367 
    368 /*
    369  * IP software interrupt routine
    370  */
    371 void
    372 ipintr()
    373 {
    374 	int s;
    375 	struct mbuf *m;
    376 
    377 	while (1) {
    378 		s = splnet();
    379 		IF_DEQUEUE(&ipintrq, m);
    380 		splx(s);
    381 		if (m == 0)
    382 			return;
    383 		ip_input(m);
    384 	}
    385 }
    386 
    387 /*
    388  * Ip input routine.  Checksum and byte swap header.  If fragmented
    389  * try to reassemble.  Process options.  Pass to next level.
    390  */
    391 void
    392 ip_input(struct mbuf *m)
    393 {
    394 	struct ip *ip = NULL;
    395 	struct ipq *fp;
    396 	struct in_ifaddr *ia;
    397 	struct ifaddr *ifa;
    398 	struct ipqent *ipqe;
    399 	int hlen = 0, mff, len;
    400 	int downmatch;
    401 
    402 #ifdef	DIAGNOSTIC
    403 	if ((m->m_flags & M_PKTHDR) == 0)
    404 		panic("ipintr no HDR");
    405 #endif
    406 #ifdef IPSEC
    407 	/*
    408 	 * should the inner packet be considered authentic?
    409 	 * see comment in ah4_input().
    410 	 */
    411 	if (m) {
    412 		m->m_flags &= ~M_AUTHIPHDR;
    413 		m->m_flags &= ~M_AUTHIPDGM;
    414 	}
    415 #endif
    416 	/*
    417 	 * If no IP addresses have been set yet but the interfaces
    418 	 * are receiving, can't do anything with incoming packets yet.
    419 	 */
    420 	if (TAILQ_FIRST(&in_ifaddr) == 0)
    421 		goto bad;
    422 	ipstat.ips_total++;
    423 	/*
    424 	 * If the IP header is not aligned, slurp it up into a new
    425 	 * mbuf with space for link headers, in the event we forward
    426 	 * it.  Otherwise, if it is aligned, make sure the entire
    427 	 * base IP header is in the first mbuf of the chain.
    428 	 */
    429 	if (IP_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) {
    430 		if ((m = m_copyup(m, sizeof(struct ip),
    431 				  (max_linkhdr + 3) & ~3)) == NULL) {
    432 			/* XXXJRT new stat, please */
    433 			ipstat.ips_toosmall++;
    434 			return;
    435 		}
    436 	} else if (__predict_false(m->m_len < sizeof (struct ip))) {
    437 		if ((m = m_pullup(m, sizeof (struct ip))) == NULL) {
    438 			ipstat.ips_toosmall++;
    439 			return;
    440 		}
    441 	}
    442 	ip = mtod(m, struct ip *);
    443 	if (ip->ip_v != IPVERSION) {
    444 		ipstat.ips_badvers++;
    445 		goto bad;
    446 	}
    447 	hlen = ip->ip_hl << 2;
    448 	if (hlen < sizeof(struct ip)) {	/* minimum header length */
    449 		ipstat.ips_badhlen++;
    450 		goto bad;
    451 	}
    452 	if (hlen > m->m_len) {
    453 		if ((m = m_pullup(m, hlen)) == 0) {
    454 			ipstat.ips_badhlen++;
    455 			return;
    456 		}
    457 		ip = mtod(m, struct ip *);
    458 	}
    459 
    460 	/*
    461 	 * RFC1122: packets with a multicast source address are
    462 	 * not allowed.
    463 	 */
    464 	if (IN_MULTICAST(ip->ip_src.s_addr)) {
    465 		ipstat.ips_badaddr++;
    466 		goto bad;
    467 	}
    468 
    469 	/* 127/8 must not appear on wire - RFC1122 */
    470 	if ((ntohl(ip->ip_dst.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET ||
    471 	    (ntohl(ip->ip_src.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) {
    472 		if ((m->m_pkthdr.rcvif->if_flags & IFF_LOOPBACK) == 0) {
    473 			ipstat.ips_badaddr++;
    474 			goto bad;
    475 		}
    476 	}
    477 
    478 	switch (m->m_pkthdr.csum_flags &
    479 		((m->m_pkthdr.rcvif->if_csum_flags_rx & M_CSUM_IPv4) |
    480 		 M_CSUM_IPv4_BAD)) {
    481 	case M_CSUM_IPv4|M_CSUM_IPv4_BAD:
    482 		INET_CSUM_COUNTER_INCR(&ip_hwcsum_bad);
    483 		goto badcsum;
    484 
    485 	case M_CSUM_IPv4:
    486 		/* Checksum was okay. */
    487 		INET_CSUM_COUNTER_INCR(&ip_hwcsum_ok);
    488 		break;
    489 
    490 	default:
    491 		/* Must compute it ourselves. */
    492 		INET_CSUM_COUNTER_INCR(&ip_swcsum);
    493 		if (in_cksum(m, hlen) != 0)
    494 			goto bad;
    495 		break;
    496 	}
    497 
    498 	/* Retrieve the packet length. */
    499 	len = ntohs(ip->ip_len);
    500 
    501 	/*
    502 	 * Check for additional length bogosity
    503 	 */
    504 	if (len < hlen) {
    505 	 	ipstat.ips_badlen++;
    506 		goto bad;
    507 	}
    508 
    509 	/*
    510 	 * Check that the amount of data in the buffers
    511 	 * is as at least much as the IP header would have us expect.
    512 	 * Trim mbufs if longer than we expect.
    513 	 * Drop packet if shorter than we expect.
    514 	 */
    515 	if (m->m_pkthdr.len < len) {
    516 		ipstat.ips_tooshort++;
    517 		goto bad;
    518 	}
    519 	if (m->m_pkthdr.len > len) {
    520 		if (m->m_len == m->m_pkthdr.len) {
    521 			m->m_len = len;
    522 			m->m_pkthdr.len = len;
    523 		} else
    524 			m_adj(m, len - m->m_pkthdr.len);
    525 	}
    526 
    527 #ifdef IPSEC
    528 	/* ipflow (IP fast forwarding) is not compatible with IPsec. */
    529 	m->m_flags &= ~M_CANFASTFWD;
    530 #else
    531 	/*
    532 	 * Assume that we can create a fast-forward IP flow entry
    533 	 * based on this packet.
    534 	 */
    535 	m->m_flags |= M_CANFASTFWD;
    536 #endif
    537 
    538 #ifdef PFIL_HOOKS
    539 	/*
    540 	 * Run through list of hooks for input packets.  If there are any
    541 	 * filters which require that additional packets in the flow are
    542 	 * not fast-forwarded, they must clear the M_CANFASTFWD flag.
    543 	 * Note that filters must _never_ set this flag, as another filter
    544 	 * in the list may have previously cleared it.
    545 	 */
    546 	/*
    547 	 * let ipfilter look at packet on the wire,
    548 	 * not the decapsulated packet.
    549 	 */
    550 #ifdef IPSEC
    551 	if (!ipsec_getnhist(m))
    552 #else
    553 	if (1)
    554 #endif
    555 	{
    556 		if (pfil_run_hooks(&inet_pfil_hook, &m, m->m_pkthdr.rcvif,
    557 				   PFIL_IN) != 0)
    558 		return;
    559 		if (m == NULL)
    560 			return;
    561 		ip = mtod(m, struct ip *);
    562 		hlen = ip->ip_hl << 2;
    563 	}
    564 #endif /* PFIL_HOOKS */
    565 
    566 #ifdef ALTQ
    567 	/* XXX Temporary until ALTQ is changed to use a pfil hook */
    568 	if (altq_input != NULL && (*altq_input)(m, AF_INET) == 0) {
    569 		/* packet dropped by traffic conditioner */
    570 		return;
    571 	}
    572 #endif
    573 
    574 	/*
    575 	 * Process options and, if not destined for us,
    576 	 * ship it on.  ip_dooptions returns 1 when an
    577 	 * error was detected (causing an icmp message
    578 	 * to be sent and the original packet to be freed).
    579 	 */
    580 	ip_nhops = 0;		/* for source routed packets */
    581 	if (hlen > sizeof (struct ip) && ip_dooptions(m))
    582 		return;
    583 
    584 	/*
    585 	 * Check our list of addresses, to see if the packet is for us.
    586 	 *
    587 	 * Traditional 4.4BSD did not consult IFF_UP at all.
    588 	 * The behavior here is to treat addresses on !IFF_UP interface
    589 	 * as not mine.
    590 	 */
    591 	downmatch = 0;
    592 	LIST_FOREACH(ia, &IN_IFADDR_HASH(ip->ip_dst.s_addr), ia_hash) {
    593 		if (in_hosteq(ia->ia_addr.sin_addr, ip->ip_dst)) {
    594 			if ((ia->ia_ifp->if_flags & IFF_UP) != 0)
    595 				break;
    596 			else
    597 				downmatch++;
    598 		}
    599 	}
    600 	if (ia != NULL)
    601 		goto ours;
    602 	if (m->m_pkthdr.rcvif->if_flags & IFF_BROADCAST) {
    603 		TAILQ_FOREACH(ifa, &m->m_pkthdr.rcvif->if_addrlist, ifa_list) {
    604 			if (ifa->ifa_addr->sa_family != AF_INET)
    605 				continue;
    606 			ia = ifatoia(ifa);
    607 			if (in_hosteq(ip->ip_dst, ia->ia_broadaddr.sin_addr) ||
    608 			    in_hosteq(ip->ip_dst, ia->ia_netbroadcast) ||
    609 			    /*
    610 			     * Look for all-0's host part (old broadcast addr),
    611 			     * either for subnet or net.
    612 			     */
    613 			    ip->ip_dst.s_addr == ia->ia_subnet ||
    614 			    ip->ip_dst.s_addr == ia->ia_net)
    615 				goto ours;
    616 			/*
    617 			 * An interface with IP address zero accepts
    618 			 * all packets that arrive on that interface.
    619 			 */
    620 			if (in_nullhost(ia->ia_addr.sin_addr))
    621 				goto ours;
    622 		}
    623 	}
    624 	if (IN_MULTICAST(ip->ip_dst.s_addr)) {
    625 		struct in_multi *inm;
    626 #ifdef MROUTING
    627 		extern struct socket *ip_mrouter;
    628 
    629 		if (M_READONLY(m)) {
    630 			if ((m = m_pullup(m, hlen)) == 0) {
    631 				ipstat.ips_toosmall++;
    632 				return;
    633 			}
    634 			ip = mtod(m, struct ip *);
    635 		}
    636 
    637 		if (ip_mrouter) {
    638 			/*
    639 			 * If we are acting as a multicast router, all
    640 			 * incoming multicast packets are passed to the
    641 			 * kernel-level multicast forwarding function.
    642 			 * The packet is returned (relatively) intact; if
    643 			 * ip_mforward() returns a non-zero value, the packet
    644 			 * must be discarded, else it may be accepted below.
    645 			 *
    646 			 * (The IP ident field is put in the same byte order
    647 			 * as expected when ip_mforward() is called from
    648 			 * ip_output().)
    649 			 */
    650 			if (ip_mforward(m, m->m_pkthdr.rcvif) != 0) {
    651 				ipstat.ips_cantforward++;
    652 				m_freem(m);
    653 				return;
    654 			}
    655 
    656 			/*
    657 			 * The process-level routing demon needs to receive
    658 			 * all multicast IGMP packets, whether or not this
    659 			 * host belongs to their destination groups.
    660 			 */
    661 			if (ip->ip_p == IPPROTO_IGMP)
    662 				goto ours;
    663 			ipstat.ips_forward++;
    664 		}
    665 #endif
    666 		/*
    667 		 * See if we belong to the destination multicast group on the
    668 		 * arrival interface.
    669 		 */
    670 		IN_LOOKUP_MULTI(ip->ip_dst, m->m_pkthdr.rcvif, inm);
    671 		if (inm == NULL) {
    672 			ipstat.ips_cantforward++;
    673 			m_freem(m);
    674 			return;
    675 		}
    676 		goto ours;
    677 	}
    678 	if (ip->ip_dst.s_addr == INADDR_BROADCAST ||
    679 	    in_nullhost(ip->ip_dst))
    680 		goto ours;
    681 
    682 	/*
    683 	 * Not for us; forward if possible and desirable.
    684 	 */
    685 	if (ipforwarding == 0) {
    686 		ipstat.ips_cantforward++;
    687 		m_freem(m);
    688 	} else {
    689 		/*
    690 		 * If ip_dst matched any of my address on !IFF_UP interface,
    691 		 * and there's no IFF_UP interface that matches ip_dst,
    692 		 * send icmp unreach.  Forwarding it will result in in-kernel
    693 		 * forwarding loop till TTL goes to 0.
    694 		 */
    695 		if (downmatch) {
    696 			icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, 0, 0);
    697 			ipstat.ips_cantforward++;
    698 			return;
    699 		}
    700 #ifdef IPSEC
    701 		if (ipsec4_in_reject(m, NULL)) {
    702 			ipsecstat.in_polvio++;
    703 			goto bad;
    704 		}
    705 #endif
    706 
    707 		ip_forward(m, 0);
    708 	}
    709 	return;
    710 
    711 ours:
    712 	/*
    713 	 * If offset or IP_MF are set, must reassemble.
    714 	 * Otherwise, nothing need be done.
    715 	 * (We could look in the reassembly queue to see
    716 	 * if the packet was previously fragmented,
    717 	 * but it's not worth the time; just let them time out.)
    718 	 */
    719 	if (ip->ip_off & ~htons(IP_DF|IP_RF)) {
    720 		if (M_READONLY(m)) {
    721 			if ((m = m_pullup(m, hlen)) == NULL) {
    722 				ipstat.ips_toosmall++;
    723 				goto bad;
    724 			}
    725 			ip = mtod(m, struct ip *);
    726 		}
    727 
    728 		/*
    729 		 * Look for queue of fragments
    730 		 * of this datagram.
    731 		 */
    732 		IPQ_LOCK();
    733 		LIST_FOREACH(fp, &ipq, ipq_q)
    734 			if (ip->ip_id == fp->ipq_id &&
    735 			    in_hosteq(ip->ip_src, fp->ipq_src) &&
    736 			    in_hosteq(ip->ip_dst, fp->ipq_dst) &&
    737 			    ip->ip_p == fp->ipq_p)
    738 				goto found;
    739 		fp = 0;
    740 found:
    741 
    742 		/*
    743 		 * Adjust ip_len to not reflect header,
    744 		 * set ipqe_mff if more fragments are expected,
    745 		 * convert offset of this to bytes.
    746 		 */
    747 		ip->ip_len = htons(ntohs(ip->ip_len) - hlen);
    748 		mff = (ip->ip_off & htons(IP_MF)) != 0;
    749 		if (mff) {
    750 		        /*
    751 		         * Make sure that fragments have a data length
    752 			 * that's a non-zero multiple of 8 bytes.
    753 		         */
    754 			if (ntohs(ip->ip_len) == 0 ||
    755 			    (ntohs(ip->ip_len) & 0x7) != 0) {
    756 				ipstat.ips_badfrags++;
    757 				IPQ_UNLOCK();
    758 				goto bad;
    759 			}
    760 		}
    761 		ip->ip_off = htons((ntohs(ip->ip_off) & IP_OFFMASK) << 3);
    762 
    763 		/*
    764 		 * If datagram marked as having more fragments
    765 		 * or if this is not the first fragment,
    766 		 * attempt reassembly; if it succeeds, proceed.
    767 		 */
    768 		if (mff || ip->ip_off != htons(0)) {
    769 			ipstat.ips_fragments++;
    770 			ipqe = pool_get(&ipqent_pool, PR_NOWAIT);
    771 			if (ipqe == NULL) {
    772 				ipstat.ips_rcvmemdrop++;
    773 				IPQ_UNLOCK();
    774 				goto bad;
    775 			}
    776 			ipqe->ipqe_mff = mff;
    777 			ipqe->ipqe_m = m;
    778 			ipqe->ipqe_ip = ip;
    779 			m = ip_reass(ipqe, fp);
    780 			if (m == 0) {
    781 				IPQ_UNLOCK();
    782 				return;
    783 			}
    784 			ipstat.ips_reassembled++;
    785 			ip = mtod(m, struct ip *);
    786 			hlen = ip->ip_hl << 2;
    787 			ip->ip_len = htons(ntohs(ip->ip_len) + hlen);
    788 		} else
    789 			if (fp)
    790 				ip_freef(fp);
    791 		IPQ_UNLOCK();
    792 	}
    793 
    794 #ifdef IPSEC
    795 	/*
    796 	 * enforce IPsec policy checking if we are seeing last header.
    797 	 * note that we do not visit this with protocols with pcb layer
    798 	 * code - like udp/tcp/raw ip.
    799 	 */
    800 	if ((inetsw[ip_protox[ip->ip_p]].pr_flags & PR_LASTHDR) != 0 &&
    801 	    ipsec4_in_reject(m, NULL)) {
    802 		ipsecstat.in_polvio++;
    803 		goto bad;
    804 	}
    805 #endif
    806 
    807 	/*
    808 	 * Switch out to protocol's input routine.
    809 	 */
    810 #if IFA_STATS
    811 	if (ia && ip)
    812 		ia->ia_ifa.ifa_data.ifad_inbytes += ntohs(ip->ip_len);
    813 #endif
    814 	ipstat.ips_delivered++;
    815     {
    816 	int off = hlen, nh = ip->ip_p;
    817 
    818 	(*inetsw[ip_protox[nh]].pr_input)(m, off, nh);
    819 	return;
    820     }
    821 bad:
    822 	m_freem(m);
    823 	return;
    824 
    825 badcsum:
    826 	ipstat.ips_badsum++;
    827 	m_freem(m);
    828 }
    829 
    830 /*
    831  * Take incoming datagram fragment and try to
    832  * reassemble it into whole datagram.  If a chain for
    833  * reassembly of this datagram already exists, then it
    834  * is given as fp; otherwise have to make a chain.
    835  */
    836 struct mbuf *
    837 ip_reass(ipqe, fp)
    838 	struct ipqent *ipqe;
    839 	struct ipq *fp;
    840 {
    841 	struct mbuf *m = ipqe->ipqe_m;
    842 	struct ipqent *nq, *p, *q;
    843 	struct ip *ip;
    844 	struct mbuf *t;
    845 	int hlen = ipqe->ipqe_ip->ip_hl << 2;
    846 	int i, next;
    847 
    848 	IPQ_LOCK_CHECK();
    849 
    850 	/*
    851 	 * Presence of header sizes in mbufs
    852 	 * would confuse code below.
    853 	 */
    854 	m->m_data += hlen;
    855 	m->m_len -= hlen;
    856 
    857 	/*
    858 	 * If first fragment to arrive, create a reassembly queue.
    859 	 */
    860 	if (fp == 0) {
    861 		/*
    862 		 * Enforce upper bound on number of fragmented packets
    863 		 * for which we attempt reassembly;
    864 		 * If maxfrag is 0, never accept fragments.
    865 		 * If maxfrag is -1, accept all fragments without limitation.
    866 		 */
    867 		if (ip_maxfragpackets < 0)
    868 			;
    869 		else if (ip_nfragpackets >= ip_maxfragpackets)
    870 			goto dropfrag;
    871 		ip_nfragpackets++;
    872 		MALLOC(fp, struct ipq *, sizeof (struct ipq),
    873 		    M_FTABLE, M_NOWAIT);
    874 		if (fp == NULL)
    875 			goto dropfrag;
    876 		LIST_INSERT_HEAD(&ipq, fp, ipq_q);
    877 		fp->ipq_ttl = IPFRAGTTL;
    878 		fp->ipq_p = ipqe->ipqe_ip->ip_p;
    879 		fp->ipq_id = ipqe->ipqe_ip->ip_id;
    880 		TAILQ_INIT(&fp->ipq_fragq);
    881 		fp->ipq_src = ipqe->ipqe_ip->ip_src;
    882 		fp->ipq_dst = ipqe->ipqe_ip->ip_dst;
    883 		p = NULL;
    884 		goto insert;
    885 	}
    886 
    887 	/*
    888 	 * Find a segment which begins after this one does.
    889 	 */
    890 	for (p = NULL, q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL;
    891 	    p = q, q = TAILQ_NEXT(q, ipqe_q))
    892 		if (ntohs(q->ipqe_ip->ip_off) > ntohs(ipqe->ipqe_ip->ip_off))
    893 			break;
    894 
    895 	/*
    896 	 * If there is a preceding segment, it may provide some of
    897 	 * our data already.  If so, drop the data from the incoming
    898 	 * segment.  If it provides all of our data, drop us.
    899 	 */
    900 	if (p != NULL) {
    901 		i = ntohs(p->ipqe_ip->ip_off) + ntohs(p->ipqe_ip->ip_len) -
    902 		    ntohs(ipqe->ipqe_ip->ip_off);
    903 		if (i > 0) {
    904 			if (i >= ntohs(ipqe->ipqe_ip->ip_len))
    905 				goto dropfrag;
    906 			m_adj(ipqe->ipqe_m, i);
    907 			ipqe->ipqe_ip->ip_off =
    908 			    htons(ntohs(ipqe->ipqe_ip->ip_off) + i);
    909 			ipqe->ipqe_ip->ip_len =
    910 			    htons(ntohs(ipqe->ipqe_ip->ip_len) - i);
    911 		}
    912 	}
    913 
    914 	/*
    915 	 * While we overlap succeeding segments trim them or,
    916 	 * if they are completely covered, dequeue them.
    917 	 */
    918 	for (; q != NULL &&
    919 	    ntohs(ipqe->ipqe_ip->ip_off) + ntohs(ipqe->ipqe_ip->ip_len) >
    920 	    ntohs(q->ipqe_ip->ip_off); q = nq) {
    921 		i = (ntohs(ipqe->ipqe_ip->ip_off) +
    922 		    ntohs(ipqe->ipqe_ip->ip_len)) - ntohs(q->ipqe_ip->ip_off);
    923 		if (i < ntohs(q->ipqe_ip->ip_len)) {
    924 			q->ipqe_ip->ip_len =
    925 			    htons(ntohs(q->ipqe_ip->ip_len) - i);
    926 			q->ipqe_ip->ip_off =
    927 			    htons(ntohs(q->ipqe_ip->ip_off) + i);
    928 			m_adj(q->ipqe_m, i);
    929 			break;
    930 		}
    931 		nq = TAILQ_NEXT(q, ipqe_q);
    932 		m_freem(q->ipqe_m);
    933 		TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q);
    934 		pool_put(&ipqent_pool, q);
    935 	}
    936 
    937 insert:
    938 	/*
    939 	 * Stick new segment in its place;
    940 	 * check for complete reassembly.
    941 	 */
    942 	if (p == NULL) {
    943 		TAILQ_INSERT_HEAD(&fp->ipq_fragq, ipqe, ipqe_q);
    944 	} else {
    945 		TAILQ_INSERT_AFTER(&fp->ipq_fragq, p, ipqe, ipqe_q);
    946 	}
    947 	next = 0;
    948 	for (p = NULL, q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL;
    949 	    p = q, q = TAILQ_NEXT(q, ipqe_q)) {
    950 		if (ntohs(q->ipqe_ip->ip_off) != next)
    951 			return (0);
    952 		next += ntohs(q->ipqe_ip->ip_len);
    953 	}
    954 	if (p->ipqe_mff)
    955 		return (0);
    956 
    957 	/*
    958 	 * Reassembly is complete.  Check for a bogus message size and
    959 	 * concatenate fragments.
    960 	 */
    961 	q = TAILQ_FIRST(&fp->ipq_fragq);
    962 	ip = q->ipqe_ip;
    963 	if ((next + (ip->ip_hl << 2)) > IP_MAXPACKET) {
    964 		ipstat.ips_toolong++;
    965 		ip_freef(fp);
    966 		return (0);
    967 	}
    968 	m = q->ipqe_m;
    969 	t = m->m_next;
    970 	m->m_next = 0;
    971 	m_cat(m, t);
    972 	nq = TAILQ_NEXT(q, ipqe_q);
    973 	pool_put(&ipqent_pool, q);
    974 	for (q = nq; q != NULL; q = nq) {
    975 		t = q->ipqe_m;
    976 		nq = TAILQ_NEXT(q, ipqe_q);
    977 		pool_put(&ipqent_pool, q);
    978 		m_cat(m, t);
    979 	}
    980 
    981 	/*
    982 	 * Create header for new ip packet by
    983 	 * modifying header of first packet;
    984 	 * dequeue and discard fragment reassembly header.
    985 	 * Make header visible.
    986 	 */
    987 	ip->ip_len = htons(next);
    988 	ip->ip_src = fp->ipq_src;
    989 	ip->ip_dst = fp->ipq_dst;
    990 	LIST_REMOVE(fp, ipq_q);
    991 	FREE(fp, M_FTABLE);
    992 	ip_nfragpackets--;
    993 	m->m_len += (ip->ip_hl << 2);
    994 	m->m_data -= (ip->ip_hl << 2);
    995 	/* some debugging cruft by sklower, below, will go away soon */
    996 	if (m->m_flags & M_PKTHDR) { /* XXX this should be done elsewhere */
    997 		int plen = 0;
    998 		for (t = m; t; t = t->m_next)
    999 			plen += t->m_len;
   1000 		m->m_pkthdr.len = plen;
   1001 	}
   1002 	return (m);
   1003 
   1004 dropfrag:
   1005 	ipstat.ips_fragdropped++;
   1006 	m_freem(m);
   1007 	pool_put(&ipqent_pool, ipqe);
   1008 	return (0);
   1009 }
   1010 
   1011 /*
   1012  * Free a fragment reassembly header and all
   1013  * associated datagrams.
   1014  */
   1015 void
   1016 ip_freef(fp)
   1017 	struct ipq *fp;
   1018 {
   1019 	struct ipqent *q, *p;
   1020 
   1021 	IPQ_LOCK_CHECK();
   1022 
   1023 	for (q = TAILQ_FIRST(&fp->ipq_fragq); q != NULL; q = p) {
   1024 		p = TAILQ_NEXT(q, ipqe_q);
   1025 		m_freem(q->ipqe_m);
   1026 		TAILQ_REMOVE(&fp->ipq_fragq, q, ipqe_q);
   1027 		pool_put(&ipqent_pool, q);
   1028 	}
   1029 	LIST_REMOVE(fp, ipq_q);
   1030 	FREE(fp, M_FTABLE);
   1031 	ip_nfragpackets--;
   1032 }
   1033 
   1034 /*
   1035  * IP timer processing;
   1036  * if a timer expires on a reassembly
   1037  * queue, discard it.
   1038  */
   1039 void
   1040 ip_slowtimo()
   1041 {
   1042 	struct ipq *fp, *nfp;
   1043 	int s = splsoftnet();
   1044 
   1045 	IPQ_LOCK();
   1046 	for (fp = LIST_FIRST(&ipq); fp != NULL; fp = nfp) {
   1047 		nfp = LIST_NEXT(fp, ipq_q);
   1048 		if (--fp->ipq_ttl == 0) {
   1049 			ipstat.ips_fragtimeout++;
   1050 			ip_freef(fp);
   1051 		}
   1052 	}
   1053 	/*
   1054 	 * If we are over the maximum number of fragments
   1055 	 * (due to the limit being lowered), drain off
   1056 	 * enough to get down to the new limit.
   1057 	 */
   1058 	if (ip_maxfragpackets < 0)
   1059 		;
   1060 	else {
   1061 		while (ip_nfragpackets > ip_maxfragpackets && LIST_FIRST(&ipq))
   1062 			ip_freef(LIST_FIRST(&ipq));
   1063 	}
   1064 	IPQ_UNLOCK();
   1065 #ifdef GATEWAY
   1066 	ipflow_slowtimo();
   1067 #endif
   1068 	splx(s);
   1069 }
   1070 
   1071 /*
   1072  * Drain off all datagram fragments.
   1073  */
   1074 void
   1075 ip_drain()
   1076 {
   1077 
   1078 	/*
   1079 	 * We may be called from a device's interrupt context.  If
   1080 	 * the ipq is already busy, just bail out now.
   1081 	 */
   1082 	if (ipq_lock_try() == 0)
   1083 		return;
   1084 
   1085 	while (LIST_FIRST(&ipq) != NULL) {
   1086 		ipstat.ips_fragdropped++;
   1087 		ip_freef(LIST_FIRST(&ipq));
   1088 	}
   1089 
   1090 	IPQ_UNLOCK();
   1091 }
   1092 
   1093 /*
   1094  * Do option processing on a datagram,
   1095  * possibly discarding it if bad options are encountered,
   1096  * or forwarding it if source-routed.
   1097  * Returns 1 if packet has been forwarded/freed,
   1098  * 0 if the packet should be processed further.
   1099  */
   1100 int
   1101 ip_dooptions(m)
   1102 	struct mbuf *m;
   1103 {
   1104 	struct ip *ip = mtod(m, struct ip *);
   1105 	u_char *cp, *cp0;
   1106 	struct ip_timestamp *ipt;
   1107 	struct in_ifaddr *ia;
   1108 	int opt, optlen, cnt, off, code, type = ICMP_PARAMPROB, forward = 0;
   1109 	struct in_addr dst;
   1110 	n_time ntime;
   1111 
   1112 	dst = ip->ip_dst;
   1113 	cp = (u_char *)(ip + 1);
   1114 	cnt = (ip->ip_hl << 2) - sizeof (struct ip);
   1115 	for (; cnt > 0; cnt -= optlen, cp += optlen) {
   1116 		opt = cp[IPOPT_OPTVAL];
   1117 		if (opt == IPOPT_EOL)
   1118 			break;
   1119 		if (opt == IPOPT_NOP)
   1120 			optlen = 1;
   1121 		else {
   1122 			if (cnt < IPOPT_OLEN + sizeof(*cp)) {
   1123 				code = &cp[IPOPT_OLEN] - (u_char *)ip;
   1124 				goto bad;
   1125 			}
   1126 			optlen = cp[IPOPT_OLEN];
   1127 			if (optlen < IPOPT_OLEN + sizeof(*cp) || optlen > cnt) {
   1128 				code = &cp[IPOPT_OLEN] - (u_char *)ip;
   1129 				goto bad;
   1130 			}
   1131 		}
   1132 		switch (opt) {
   1133 
   1134 		default:
   1135 			break;
   1136 
   1137 		/*
   1138 		 * Source routing with record.
   1139 		 * Find interface with current destination address.
   1140 		 * If none on this machine then drop if strictly routed,
   1141 		 * or do nothing if loosely routed.
   1142 		 * Record interface address and bring up next address
   1143 		 * component.  If strictly routed make sure next
   1144 		 * address is on directly accessible net.
   1145 		 */
   1146 		case IPOPT_LSRR:
   1147 		case IPOPT_SSRR:
   1148 			if (ip_allowsrcrt == 0) {
   1149 				type = ICMP_UNREACH;
   1150 				code = ICMP_UNREACH_NET_PROHIB;
   1151 				goto bad;
   1152 			}
   1153 			if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
   1154 				code = &cp[IPOPT_OLEN] - (u_char *)ip;
   1155 				goto bad;
   1156 			}
   1157 			if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
   1158 				code = &cp[IPOPT_OFFSET] - (u_char *)ip;
   1159 				goto bad;
   1160 			}
   1161 			ipaddr.sin_addr = ip->ip_dst;
   1162 			ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr)));
   1163 			if (ia == 0) {
   1164 				if (opt == IPOPT_SSRR) {
   1165 					type = ICMP_UNREACH;
   1166 					code = ICMP_UNREACH_SRCFAIL;
   1167 					goto bad;
   1168 				}
   1169 				/*
   1170 				 * Loose routing, and not at next destination
   1171 				 * yet; nothing to do except forward.
   1172 				 */
   1173 				break;
   1174 			}
   1175 			off--;			/* 0 origin */
   1176 			if ((off + sizeof(struct in_addr)) > optlen) {
   1177 				/*
   1178 				 * End of source route.  Should be for us.
   1179 				 */
   1180 				save_rte(cp, ip->ip_src);
   1181 				break;
   1182 			}
   1183 			/*
   1184 			 * locate outgoing interface
   1185 			 */
   1186 			bcopy((caddr_t)(cp + off), (caddr_t)&ipaddr.sin_addr,
   1187 			    sizeof(ipaddr.sin_addr));
   1188 			if (opt == IPOPT_SSRR)
   1189 				ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr)));
   1190 			else
   1191 				ia = ip_rtaddr(ipaddr.sin_addr);
   1192 			if (ia == 0) {
   1193 				type = ICMP_UNREACH;
   1194 				code = ICMP_UNREACH_SRCFAIL;
   1195 				goto bad;
   1196 			}
   1197 			ip->ip_dst = ipaddr.sin_addr;
   1198 			bcopy((caddr_t)&ia->ia_addr.sin_addr,
   1199 			    (caddr_t)(cp + off), sizeof(struct in_addr));
   1200 			cp[IPOPT_OFFSET] += sizeof(struct in_addr);
   1201 			/*
   1202 			 * Let ip_intr's mcast routing check handle mcast pkts
   1203 			 */
   1204 			forward = !IN_MULTICAST(ip->ip_dst.s_addr);
   1205 			break;
   1206 
   1207 		case IPOPT_RR:
   1208 			if (optlen < IPOPT_OFFSET + sizeof(*cp)) {
   1209 				code = &cp[IPOPT_OLEN] - (u_char *)ip;
   1210 				goto bad;
   1211 			}
   1212 			if ((off = cp[IPOPT_OFFSET]) < IPOPT_MINOFF) {
   1213 				code = &cp[IPOPT_OFFSET] - (u_char *)ip;
   1214 				goto bad;
   1215 			}
   1216 			/*
   1217 			 * If no space remains, ignore.
   1218 			 */
   1219 			off--;			/* 0 origin */
   1220 			if ((off + sizeof(struct in_addr)) > optlen)
   1221 				break;
   1222 			bcopy((caddr_t)(&ip->ip_dst), (caddr_t)&ipaddr.sin_addr,
   1223 			    sizeof(ipaddr.sin_addr));
   1224 			/*
   1225 			 * locate outgoing interface; if we're the destination,
   1226 			 * use the incoming interface (should be same).
   1227 			 */
   1228 			if ((ia = ifatoia(ifa_ifwithaddr(sintosa(&ipaddr))))
   1229 			    == NULL &&
   1230 			    (ia = ip_rtaddr(ipaddr.sin_addr)) == NULL) {
   1231 				type = ICMP_UNREACH;
   1232 				code = ICMP_UNREACH_HOST;
   1233 				goto bad;
   1234 			}
   1235 			bcopy((caddr_t)&ia->ia_addr.sin_addr,
   1236 			    (caddr_t)(cp + off), sizeof(struct in_addr));
   1237 			cp[IPOPT_OFFSET] += sizeof(struct in_addr);
   1238 			break;
   1239 
   1240 		case IPOPT_TS:
   1241 			code = cp - (u_char *)ip;
   1242 			ipt = (struct ip_timestamp *)cp;
   1243 			if (ipt->ipt_len < 4 || ipt->ipt_len > 40) {
   1244 				code = (u_char *)&ipt->ipt_len - (u_char *)ip;
   1245 				goto bad;
   1246 			}
   1247 			if (ipt->ipt_ptr < 5) {
   1248 				code = (u_char *)&ipt->ipt_ptr - (u_char *)ip;
   1249 				goto bad;
   1250 			}
   1251 			if (ipt->ipt_ptr > ipt->ipt_len - sizeof (int32_t)) {
   1252 				if (++ipt->ipt_oflw == 0) {
   1253 					code = (u_char *)&ipt->ipt_ptr -
   1254 					    (u_char *)ip;
   1255 					goto bad;
   1256 				}
   1257 				break;
   1258 			}
   1259 			cp0 = (cp + ipt->ipt_ptr - 1);
   1260 			switch (ipt->ipt_flg) {
   1261 
   1262 			case IPOPT_TS_TSONLY:
   1263 				break;
   1264 
   1265 			case IPOPT_TS_TSANDADDR:
   1266 				if (ipt->ipt_ptr - 1 + sizeof(n_time) +
   1267 				    sizeof(struct in_addr) > ipt->ipt_len) {
   1268 					code = (u_char *)&ipt->ipt_ptr -
   1269 					    (u_char *)ip;
   1270 					goto bad;
   1271 				}
   1272 				ipaddr.sin_addr = dst;
   1273 				ia = ifatoia(ifaof_ifpforaddr(sintosa(&ipaddr),
   1274 				    m->m_pkthdr.rcvif));
   1275 				if (ia == 0)
   1276 					continue;
   1277 				bcopy(&ia->ia_addr.sin_addr,
   1278 				    cp0, sizeof(struct in_addr));
   1279 				ipt->ipt_ptr += sizeof(struct in_addr);
   1280 				break;
   1281 
   1282 			case IPOPT_TS_PRESPEC:
   1283 				if (ipt->ipt_ptr - 1 + sizeof(n_time) +
   1284 				    sizeof(struct in_addr) > ipt->ipt_len) {
   1285 					code = (u_char *)&ipt->ipt_ptr -
   1286 					    (u_char *)ip;
   1287 					goto bad;
   1288 				}
   1289 				bcopy(cp0, &ipaddr.sin_addr,
   1290 				    sizeof(struct in_addr));
   1291 				if (ifatoia(ifa_ifwithaddr(sintosa(&ipaddr)))
   1292 				    == NULL)
   1293 					continue;
   1294 				ipt->ipt_ptr += sizeof(struct in_addr);
   1295 				break;
   1296 
   1297 			default:
   1298 				/* XXX can't take &ipt->ipt_flg */
   1299 				code = (u_char *)&ipt->ipt_ptr -
   1300 				    (u_char *)ip + 1;
   1301 				goto bad;
   1302 			}
   1303 			ntime = iptime();
   1304 			cp0 = (u_char *) &ntime; /* XXX grumble, GCC... */
   1305 			bcopy(cp0, (caddr_t)cp + ipt->ipt_ptr - 1,
   1306 			    sizeof(n_time));
   1307 			ipt->ipt_ptr += sizeof(n_time);
   1308 		}
   1309 	}
   1310 	if (forward) {
   1311 		if (ip_forwsrcrt == 0) {
   1312 			type = ICMP_UNREACH;
   1313 			code = ICMP_UNREACH_SRCFAIL;
   1314 			goto bad;
   1315 		}
   1316 		ip_forward(m, 1);
   1317 		return (1);
   1318 	}
   1319 	return (0);
   1320 bad:
   1321 	icmp_error(m, type, code, 0, 0);
   1322 	ipstat.ips_badoptions++;
   1323 	return (1);
   1324 }
   1325 
   1326 /*
   1327  * Given address of next destination (final or next hop),
   1328  * return internet address info of interface to be used to get there.
   1329  */
   1330 struct in_ifaddr *
   1331 ip_rtaddr(dst)
   1332 	 struct in_addr dst;
   1333 {
   1334 	struct sockaddr_in *sin;
   1335 
   1336 	sin = satosin(&ipforward_rt.ro_dst);
   1337 
   1338 	if (ipforward_rt.ro_rt == 0 || !in_hosteq(dst, sin->sin_addr)) {
   1339 		if (ipforward_rt.ro_rt) {
   1340 			RTFREE(ipforward_rt.ro_rt);
   1341 			ipforward_rt.ro_rt = 0;
   1342 		}
   1343 		sin->sin_family = AF_INET;
   1344 		sin->sin_len = sizeof(*sin);
   1345 		sin->sin_addr = dst;
   1346 
   1347 		rtalloc(&ipforward_rt);
   1348 	}
   1349 	if (ipforward_rt.ro_rt == 0)
   1350 		return ((struct in_ifaddr *)0);
   1351 	return (ifatoia(ipforward_rt.ro_rt->rt_ifa));
   1352 }
   1353 
   1354 /*
   1355  * Save incoming source route for use in replies,
   1356  * to be picked up later by ip_srcroute if the receiver is interested.
   1357  */
   1358 void
   1359 save_rte(option, dst)
   1360 	u_char *option;
   1361 	struct in_addr dst;
   1362 {
   1363 	unsigned olen;
   1364 
   1365 	olen = option[IPOPT_OLEN];
   1366 #ifdef DIAGNOSTIC
   1367 	if (ipprintfs)
   1368 		printf("save_rte: olen %d\n", olen);
   1369 #endif /* 0 */
   1370 	if (olen > sizeof(ip_srcrt) - (1 + sizeof(dst)))
   1371 		return;
   1372 	bcopy((caddr_t)option, (caddr_t)ip_srcrt.srcopt, olen);
   1373 	ip_nhops = (olen - IPOPT_OFFSET - 1) / sizeof(struct in_addr);
   1374 	ip_srcrt.dst = dst;
   1375 }
   1376 
   1377 /*
   1378  * Retrieve incoming source route for use in replies,
   1379  * in the same form used by setsockopt.
   1380  * The first hop is placed before the options, will be removed later.
   1381  */
   1382 struct mbuf *
   1383 ip_srcroute()
   1384 {
   1385 	struct in_addr *p, *q;
   1386 	struct mbuf *m;
   1387 
   1388 	if (ip_nhops == 0)
   1389 		return ((struct mbuf *)0);
   1390 	m = m_get(M_DONTWAIT, MT_SOOPTS);
   1391 	if (m == 0)
   1392 		return ((struct mbuf *)0);
   1393 
   1394 #define OPTSIZ	(sizeof(ip_srcrt.nop) + sizeof(ip_srcrt.srcopt))
   1395 
   1396 	/* length is (nhops+1)*sizeof(addr) + sizeof(nop + srcrt header) */
   1397 	m->m_len = ip_nhops * sizeof(struct in_addr) + sizeof(struct in_addr) +
   1398 	    OPTSIZ;
   1399 #ifdef DIAGNOSTIC
   1400 	if (ipprintfs)
   1401 		printf("ip_srcroute: nhops %d mlen %d", ip_nhops, m->m_len);
   1402 #endif
   1403 
   1404 	/*
   1405 	 * First save first hop for return route
   1406 	 */
   1407 	p = &ip_srcrt.route[ip_nhops - 1];
   1408 	*(mtod(m, struct in_addr *)) = *p--;
   1409 #ifdef DIAGNOSTIC
   1410 	if (ipprintfs)
   1411 		printf(" hops %x", ntohl(mtod(m, struct in_addr *)->s_addr));
   1412 #endif
   1413 
   1414 	/*
   1415 	 * Copy option fields and padding (nop) to mbuf.
   1416 	 */
   1417 	ip_srcrt.nop = IPOPT_NOP;
   1418 	ip_srcrt.srcopt[IPOPT_OFFSET] = IPOPT_MINOFF;
   1419 	bcopy((caddr_t)&ip_srcrt.nop,
   1420 	    mtod(m, caddr_t) + sizeof(struct in_addr), OPTSIZ);
   1421 	q = (struct in_addr *)(mtod(m, caddr_t) +
   1422 	    sizeof(struct in_addr) + OPTSIZ);
   1423 #undef OPTSIZ
   1424 	/*
   1425 	 * Record return path as an IP source route,
   1426 	 * reversing the path (pointers are now aligned).
   1427 	 */
   1428 	while (p >= ip_srcrt.route) {
   1429 #ifdef DIAGNOSTIC
   1430 		if (ipprintfs)
   1431 			printf(" %x", ntohl(q->s_addr));
   1432 #endif
   1433 		*q++ = *p--;
   1434 	}
   1435 	/*
   1436 	 * Last hop goes to final destination.
   1437 	 */
   1438 	*q = ip_srcrt.dst;
   1439 #ifdef DIAGNOSTIC
   1440 	if (ipprintfs)
   1441 		printf(" %x\n", ntohl(q->s_addr));
   1442 #endif
   1443 	return (m);
   1444 }
   1445 
   1446 /*
   1447  * Strip out IP options, at higher
   1448  * level protocol in the kernel.
   1449  * Second argument is buffer to which options
   1450  * will be moved, and return value is their length.
   1451  * XXX should be deleted; last arg currently ignored.
   1452  */
   1453 void
   1454 ip_stripoptions(m, mopt)
   1455 	struct mbuf *m;
   1456 	struct mbuf *mopt;
   1457 {
   1458 	int i;
   1459 	struct ip *ip = mtod(m, struct ip *);
   1460 	caddr_t opts;
   1461 	int olen;
   1462 
   1463 	olen = (ip->ip_hl << 2) - sizeof (struct ip);
   1464 	opts = (caddr_t)(ip + 1);
   1465 	i = m->m_len - (sizeof (struct ip) + olen);
   1466 	bcopy(opts  + olen, opts, (unsigned)i);
   1467 	m->m_len -= olen;
   1468 	if (m->m_flags & M_PKTHDR)
   1469 		m->m_pkthdr.len -= olen;
   1470 	ip->ip_len = htons(ntohs(ip->ip_len) - olen);
   1471 	ip->ip_hl = sizeof (struct ip) >> 2;
   1472 }
   1473 
   1474 const int inetctlerrmap[PRC_NCMDS] = {
   1475 	0,		0,		0,		0,
   1476 	0,		EMSGSIZE,	EHOSTDOWN,	EHOSTUNREACH,
   1477 	EHOSTUNREACH,	EHOSTUNREACH,	ECONNREFUSED,	ECONNREFUSED,
   1478 	EMSGSIZE,	EHOSTUNREACH,	0,		0,
   1479 	0,		0,		0,		0,
   1480 	ENOPROTOOPT
   1481 };
   1482 
   1483 /*
   1484  * Forward a packet.  If some error occurs return the sender
   1485  * an icmp packet.  Note we can't always generate a meaningful
   1486  * icmp message because icmp doesn't have a large enough repertoire
   1487  * of codes and types.
   1488  *
   1489  * If not forwarding, just drop the packet.  This could be confusing
   1490  * if ipforwarding was zero but some routing protocol was advancing
   1491  * us as a gateway to somewhere.  However, we must let the routing
   1492  * protocol deal with that.
   1493  *
   1494  * The srcrt parameter indicates whether the packet is being forwarded
   1495  * via a source route.
   1496  */
   1497 void
   1498 ip_forward(m, srcrt)
   1499 	struct mbuf *m;
   1500 	int srcrt;
   1501 {
   1502 	struct ip *ip = mtod(m, struct ip *);
   1503 	struct sockaddr_in *sin;
   1504 	struct rtentry *rt;
   1505 	int error, type = 0, code = 0;
   1506 	struct mbuf *mcopy;
   1507 	n_long dest;
   1508 	struct ifnet *destifp;
   1509 #ifdef IPSEC
   1510 	struct ifnet dummyifp;
   1511 #endif
   1512 
   1513 	/*
   1514 	 * Clear any in-bound checksum flags for this packet.
   1515 	 */
   1516 	m->m_pkthdr.csum_flags = 0;
   1517 
   1518 	dest = 0;
   1519 #ifdef DIAGNOSTIC
   1520 	if (ipprintfs)
   1521 		printf("forward: src %2.2x dst %2.2x ttl %x\n",
   1522 		    ntohl(ip->ip_src.s_addr),
   1523 		    ntohl(ip->ip_dst.s_addr), ip->ip_ttl);
   1524 #endif
   1525 	if (m->m_flags & (M_BCAST|M_MCAST) || in_canforward(ip->ip_dst) == 0) {
   1526 		ipstat.ips_cantforward++;
   1527 		m_freem(m);
   1528 		return;
   1529 	}
   1530 	if (ip->ip_ttl <= IPTTLDEC) {
   1531 		icmp_error(m, ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS, dest, 0);
   1532 		return;
   1533 	}
   1534 	ip->ip_ttl -= IPTTLDEC;
   1535 
   1536 	sin = satosin(&ipforward_rt.ro_dst);
   1537 	if ((rt = ipforward_rt.ro_rt) == 0 ||
   1538 	    !in_hosteq(ip->ip_dst, sin->sin_addr)) {
   1539 		if (ipforward_rt.ro_rt) {
   1540 			RTFREE(ipforward_rt.ro_rt);
   1541 			ipforward_rt.ro_rt = 0;
   1542 		}
   1543 		sin->sin_family = AF_INET;
   1544 		sin->sin_len = sizeof(struct sockaddr_in);
   1545 		sin->sin_addr = ip->ip_dst;
   1546 
   1547 		rtalloc(&ipforward_rt);
   1548 		if (ipforward_rt.ro_rt == 0) {
   1549 			icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_HOST, dest, 0);
   1550 			return;
   1551 		}
   1552 		rt = ipforward_rt.ro_rt;
   1553 	}
   1554 
   1555 	/*
   1556 	 * Save at most 68 bytes of the packet in case
   1557 	 * we need to generate an ICMP message to the src.
   1558 	 * Pullup to avoid sharing mbuf cluster between m and mcopy.
   1559 	 */
   1560 	mcopy = m_copym(m, 0, imin(ntohs(ip->ip_len), 68), M_DONTWAIT);
   1561 	if (mcopy)
   1562 		mcopy = m_pullup(mcopy, ip->ip_hl << 2);
   1563 
   1564 	/*
   1565 	 * If forwarding packet using same interface that it came in on,
   1566 	 * perhaps should send a redirect to sender to shortcut a hop.
   1567 	 * Only send redirect if source is sending directly to us,
   1568 	 * and if packet was not source routed (or has any options).
   1569 	 * Also, don't send redirect if forwarding using a default route
   1570 	 * or a route modified by a redirect.
   1571 	 */
   1572 	if (rt->rt_ifp == m->m_pkthdr.rcvif &&
   1573 	    (rt->rt_flags & (RTF_DYNAMIC|RTF_MODIFIED)) == 0 &&
   1574 	    !in_nullhost(satosin(rt_key(rt))->sin_addr) &&
   1575 	    ipsendredirects && !srcrt) {
   1576 		if (rt->rt_ifa &&
   1577 		    (ip->ip_src.s_addr & ifatoia(rt->rt_ifa)->ia_subnetmask) ==
   1578 		    ifatoia(rt->rt_ifa)->ia_subnet) {
   1579 			if (rt->rt_flags & RTF_GATEWAY)
   1580 				dest = satosin(rt->rt_gateway)->sin_addr.s_addr;
   1581 			else
   1582 				dest = ip->ip_dst.s_addr;
   1583 			/*
   1584 			 * Router requirements says to only send host
   1585 			 * redirects.
   1586 			 */
   1587 			type = ICMP_REDIRECT;
   1588 			code = ICMP_REDIRECT_HOST;
   1589 #ifdef DIAGNOSTIC
   1590 			if (ipprintfs)
   1591 				printf("redirect (%d) to %x\n", code,
   1592 				    (u_int32_t)dest);
   1593 #endif
   1594 		}
   1595 	}
   1596 
   1597 #ifdef IPSEC
   1598 	/* Don't lookup socket in forwarding case */
   1599 	(void)ipsec_setsocket(m, NULL);
   1600 #endif
   1601 	error = ip_output(m, (struct mbuf *)0, &ipforward_rt,
   1602 	    (IP_FORWARDING | (ip_directedbcast ? IP_ALLOWBROADCAST : 0)), 0);
   1603 	if (error)
   1604 		ipstat.ips_cantforward++;
   1605 	else {
   1606 		ipstat.ips_forward++;
   1607 		if (type)
   1608 			ipstat.ips_redirectsent++;
   1609 		else {
   1610 			if (mcopy) {
   1611 #ifdef GATEWAY
   1612 				if (mcopy->m_flags & M_CANFASTFWD)
   1613 					ipflow_create(&ipforward_rt, mcopy);
   1614 #endif
   1615 				m_freem(mcopy);
   1616 			}
   1617 			return;
   1618 		}
   1619 	}
   1620 	if (mcopy == NULL)
   1621 		return;
   1622 	destifp = NULL;
   1623 
   1624 	switch (error) {
   1625 
   1626 	case 0:				/* forwarded, but need redirect */
   1627 		/* type, code set above */
   1628 		break;
   1629 
   1630 	case ENETUNREACH:		/* shouldn't happen, checked above */
   1631 	case EHOSTUNREACH:
   1632 	case ENETDOWN:
   1633 	case EHOSTDOWN:
   1634 	default:
   1635 		type = ICMP_UNREACH;
   1636 		code = ICMP_UNREACH_HOST;
   1637 		break;
   1638 
   1639 	case EMSGSIZE:
   1640 		type = ICMP_UNREACH;
   1641 		code = ICMP_UNREACH_NEEDFRAG;
   1642 #ifndef IPSEC
   1643 		if (ipforward_rt.ro_rt)
   1644 			destifp = ipforward_rt.ro_rt->rt_ifp;
   1645 #else
   1646 		/*
   1647 		 * If the packet is routed over IPsec tunnel, tell the
   1648 		 * originator the tunnel MTU.
   1649 		 *	tunnel MTU = if MTU - sizeof(IP) - ESP/AH hdrsiz
   1650 		 * XXX quickhack!!!
   1651 		 */
   1652 		if (ipforward_rt.ro_rt) {
   1653 			struct secpolicy *sp;
   1654 			int ipsecerror;
   1655 			size_t ipsechdr;
   1656 			struct route *ro;
   1657 
   1658 			sp = ipsec4_getpolicybyaddr(mcopy,
   1659 			                            IPSEC_DIR_OUTBOUND,
   1660 			                            IP_FORWARDING,
   1661 			                            &ipsecerror);
   1662 
   1663 			if (sp == NULL)
   1664 				destifp = ipforward_rt.ro_rt->rt_ifp;
   1665 			else {
   1666 				/* count IPsec header size */
   1667 				ipsechdr = ipsec4_hdrsiz(mcopy,
   1668 				                         IPSEC_DIR_OUTBOUND,
   1669 				                         NULL);
   1670 
   1671 				/*
   1672 				 * find the correct route for outer IPv4
   1673 				 * header, compute tunnel MTU.
   1674 				 *
   1675 				 * XXX BUG ALERT
   1676 				 * The "dummyifp" code relies upon the fact
   1677 				 * that icmp_error() touches only ifp->if_mtu.
   1678 				 */
   1679 				/*XXX*/
   1680 				destifp = NULL;
   1681 				if (sp->req != NULL
   1682 				 && sp->req->sav != NULL
   1683 				 && sp->req->sav->sah != NULL) {
   1684 					ro = &sp->req->sav->sah->sa_route;
   1685 					if (ro->ro_rt && ro->ro_rt->rt_ifp) {
   1686 						dummyifp.if_mtu =
   1687 						    ro->ro_rt->rt_rmx.rmx_mtu ?
   1688 						    ro->ro_rt->rt_rmx.rmx_mtu :
   1689 						    ro->ro_rt->rt_ifp->if_mtu;
   1690 						dummyifp.if_mtu -= ipsechdr;
   1691 						destifp = &dummyifp;
   1692 					}
   1693 				}
   1694 
   1695 				key_freesp(sp);
   1696 			}
   1697 		}
   1698 #endif /*IPSEC*/
   1699 		ipstat.ips_cantfrag++;
   1700 		break;
   1701 
   1702 	case ENOBUFS:
   1703 #if 1
   1704 		/*
   1705 		 * a router should not generate ICMP_SOURCEQUENCH as
   1706 		 * required in RFC1812 Requirements for IP Version 4 Routers.
   1707 		 * source quench could be a big problem under DoS attacks,
   1708 		 * or if the underlying interface is rate-limited.
   1709 		 */
   1710 		if (mcopy)
   1711 			m_freem(mcopy);
   1712 		return;
   1713 #else
   1714 		type = ICMP_SOURCEQUENCH;
   1715 		code = 0;
   1716 		break;
   1717 #endif
   1718 	}
   1719 	icmp_error(mcopy, type, code, dest, destifp);
   1720 }
   1721 
   1722 void
   1723 ip_savecontrol(inp, mp, ip, m)
   1724 	struct inpcb *inp;
   1725 	struct mbuf **mp;
   1726 	struct ip *ip;
   1727 	struct mbuf *m;
   1728 {
   1729 
   1730 	if (inp->inp_socket->so_options & SO_TIMESTAMP) {
   1731 		struct timeval tv;
   1732 
   1733 		microtime(&tv);
   1734 		*mp = sbcreatecontrol((caddr_t) &tv, sizeof(tv),
   1735 		    SCM_TIMESTAMP, SOL_SOCKET);
   1736 		if (*mp)
   1737 			mp = &(*mp)->m_next;
   1738 	}
   1739 	if (inp->inp_flags & INP_RECVDSTADDR) {
   1740 		*mp = sbcreatecontrol((caddr_t) &ip->ip_dst,
   1741 		    sizeof(struct in_addr), IP_RECVDSTADDR, IPPROTO_IP);
   1742 		if (*mp)
   1743 			mp = &(*mp)->m_next;
   1744 	}
   1745 #ifdef notyet
   1746 	/*
   1747 	 * XXX
   1748 	 * Moving these out of udp_input() made them even more broken
   1749 	 * than they already were.
   1750 	 *	- fenner (at) parc.xerox.com
   1751 	 */
   1752 	/* options were tossed already */
   1753 	if (inp->inp_flags & INP_RECVOPTS) {
   1754 		*mp = sbcreatecontrol((caddr_t) opts_deleted_above,
   1755 		    sizeof(struct in_addr), IP_RECVOPTS, IPPROTO_IP);
   1756 		if (*mp)
   1757 			mp = &(*mp)->m_next;
   1758 	}
   1759 	/* ip_srcroute doesn't do what we want here, need to fix */
   1760 	if (inp->inp_flags & INP_RECVRETOPTS) {
   1761 		*mp = sbcreatecontrol((caddr_t) ip_srcroute(),
   1762 		    sizeof(struct in_addr), IP_RECVRETOPTS, IPPROTO_IP);
   1763 		if (*mp)
   1764 			mp = &(*mp)->m_next;
   1765 	}
   1766 #endif
   1767 	if (inp->inp_flags & INP_RECVIF) {
   1768 		struct sockaddr_dl sdl;
   1769 
   1770 		sdl.sdl_len = offsetof(struct sockaddr_dl, sdl_data[0]);
   1771 		sdl.sdl_family = AF_LINK;
   1772 		sdl.sdl_index = m->m_pkthdr.rcvif ?
   1773 		    m->m_pkthdr.rcvif->if_index : 0;
   1774 		sdl.sdl_nlen = sdl.sdl_alen = sdl.sdl_slen = 0;
   1775 		*mp = sbcreatecontrol((caddr_t) &sdl, sdl.sdl_len,
   1776 		    IP_RECVIF, IPPROTO_IP);
   1777 		if (*mp)
   1778 			mp = &(*mp)->m_next;
   1779 	}
   1780 }
   1781 
   1782 int
   1783 ip_sysctl(name, namelen, oldp, oldlenp, newp, newlen)
   1784 	int *name;
   1785 	u_int namelen;
   1786 	void *oldp;
   1787 	size_t *oldlenp;
   1788 	void *newp;
   1789 	size_t newlen;
   1790 {
   1791 	extern int subnetsarelocal, hostzeroisbroadcast;
   1792 
   1793 	int error, old;
   1794 
   1795 	/* All sysctl names at this level are terminal. */
   1796 	if (namelen != 1)
   1797 		return (ENOTDIR);
   1798 
   1799 	switch (name[0]) {
   1800 	case IPCTL_FORWARDING:
   1801 		return (sysctl_int(oldp, oldlenp, newp, newlen, &ipforwarding));
   1802 	case IPCTL_SENDREDIRECTS:
   1803 		return (sysctl_int(oldp, oldlenp, newp, newlen,
   1804 			&ipsendredirects));
   1805 	case IPCTL_DEFTTL:
   1806 		return (sysctl_int(oldp, oldlenp, newp, newlen, &ip_defttl));
   1807 #ifdef notyet
   1808 	case IPCTL_DEFMTU:
   1809 		return (sysctl_int(oldp, oldlenp, newp, newlen, &ip_mtu));
   1810 #endif
   1811 	case IPCTL_FORWSRCRT:
   1812 		/* Don't allow this to change in a secure environment.  */
   1813 		if (securelevel > 0)
   1814 			return (sysctl_rdint(oldp, oldlenp, newp,
   1815 			    ip_forwsrcrt));
   1816 		else
   1817 			return (sysctl_int(oldp, oldlenp, newp, newlen,
   1818 			    &ip_forwsrcrt));
   1819 	case IPCTL_DIRECTEDBCAST:
   1820 		return (sysctl_int(oldp, oldlenp, newp, newlen,
   1821 		    &ip_directedbcast));
   1822 	case IPCTL_ALLOWSRCRT:
   1823 		return (sysctl_int(oldp, oldlenp, newp, newlen,
   1824 		    &ip_allowsrcrt));
   1825 	case IPCTL_SUBNETSARELOCAL:
   1826 		return (sysctl_int(oldp, oldlenp, newp, newlen,
   1827 		    &subnetsarelocal));
   1828 	case IPCTL_MTUDISC:
   1829 		error = sysctl_int(oldp, oldlenp, newp, newlen,
   1830 		    &ip_mtudisc);
   1831 		if (ip_mtudisc != 0 && ip_mtudisc_timeout_q == NULL) {
   1832 			ip_mtudisc_timeout_q =
   1833 			    rt_timer_queue_create(ip_mtudisc_timeout);
   1834 		} else if (ip_mtudisc == 0 && ip_mtudisc_timeout_q != NULL) {
   1835 			rt_timer_queue_destroy(ip_mtudisc_timeout_q, TRUE);
   1836 			ip_mtudisc_timeout_q = NULL;
   1837 		}
   1838 		return error;
   1839 	case IPCTL_ANONPORTMIN:
   1840 		old = anonportmin;
   1841 		error = sysctl_int(oldp, oldlenp, newp, newlen, &anonportmin);
   1842 		if (anonportmin >= anonportmax || anonportmin < 0
   1843 		    || anonportmin > 65535
   1844 #ifndef IPNOPRIVPORTS
   1845 		    || anonportmin < IPPORT_RESERVED
   1846 #endif
   1847 		    ) {
   1848 			anonportmin = old;
   1849 			return (EINVAL);
   1850 		}
   1851 		return (error);
   1852 	case IPCTL_ANONPORTMAX:
   1853 		old = anonportmax;
   1854 		error = sysctl_int(oldp, oldlenp, newp, newlen, &anonportmax);
   1855 		if (anonportmin >= anonportmax || anonportmax < 0
   1856 		    || anonportmax > 65535
   1857 #ifndef IPNOPRIVPORTS
   1858 		    || anonportmax < IPPORT_RESERVED
   1859 #endif
   1860 		    ) {
   1861 			anonportmax = old;
   1862 			return (EINVAL);
   1863 		}
   1864 		return (error);
   1865 	case IPCTL_MTUDISCTIMEOUT:
   1866 		error = sysctl_int(oldp, oldlenp, newp, newlen,
   1867 		   &ip_mtudisc_timeout);
   1868 		if (ip_mtudisc_timeout < 0)
   1869 			return (EINVAL);
   1870 		if (ip_mtudisc_timeout_q != NULL)
   1871 			rt_timer_queue_change(ip_mtudisc_timeout_q,
   1872 					      ip_mtudisc_timeout);
   1873 		return (error);
   1874 #ifdef GATEWAY
   1875 	case IPCTL_MAXFLOWS:
   1876 	    {
   1877 		int s;
   1878 
   1879 		error = sysctl_int(oldp, oldlenp, newp, newlen,
   1880 		   &ip_maxflows);
   1881 		s = splsoftnet();
   1882 		ipflow_reap(0);
   1883 		splx(s);
   1884 		return (error);
   1885 	    }
   1886 #endif
   1887 	case IPCTL_HOSTZEROBROADCAST:
   1888 		return (sysctl_int(oldp, oldlenp, newp, newlen,
   1889 		    &hostzeroisbroadcast));
   1890 #if NGIF > 0
   1891 	case IPCTL_GIF_TTL:
   1892 		return(sysctl_int(oldp, oldlenp, newp, newlen,
   1893 				  &ip_gif_ttl));
   1894 #endif
   1895 
   1896 #if NGRE > 0
   1897 	case IPCTL_GRE_TTL:
   1898 		return(sysctl_int(oldp, oldlenp, newp, newlen,
   1899 				  &ip_gre_ttl));
   1900 #endif
   1901 
   1902 #ifndef IPNOPRIVPORTS
   1903 	case IPCTL_LOWPORTMIN:
   1904 		old = lowportmin;
   1905 		error = sysctl_int(oldp, oldlenp, newp, newlen, &lowportmin);
   1906 		if (lowportmin >= lowportmax
   1907 		    || lowportmin > IPPORT_RESERVEDMAX
   1908 		    || lowportmin < IPPORT_RESERVEDMIN
   1909 		    ) {
   1910 			lowportmin = old;
   1911 			return (EINVAL);
   1912 		}
   1913 		return (error);
   1914 	case IPCTL_LOWPORTMAX:
   1915 		old = lowportmax;
   1916 		error = sysctl_int(oldp, oldlenp, newp, newlen, &lowportmax);
   1917 		if (lowportmin >= lowportmax
   1918 		    || lowportmax > IPPORT_RESERVEDMAX
   1919 		    || lowportmax < IPPORT_RESERVEDMIN
   1920 		    ) {
   1921 			lowportmax = old;
   1922 			return (EINVAL);
   1923 		}
   1924 		return (error);
   1925 #endif
   1926 
   1927 	case IPCTL_MAXFRAGPACKETS:
   1928 		return (sysctl_int(oldp, oldlenp, newp, newlen,
   1929 		    &ip_maxfragpackets));
   1930 
   1931 	default:
   1932 		return (EOPNOTSUPP);
   1933 	}
   1934 	/* NOTREACHED */
   1935 }
   1936