1/*	$NetBSD: ip_icmp.c,v 1.128 2012/01/09 14:31:22 liamjfoy 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, 2000 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 * This code is derived from software contributed to The NetBSD Foundation
41 * by Jason R. Thorpe of Zembu Labs, Inc.
42 *
43 * Redistribution and use in source and binary forms, with or without
44 * modification, are permitted provided that the following conditions
45 * are met:
46 * 1. Redistributions of source code must retain the above copyright
47 *    notice, this list of conditions and the following disclaimer.
48 * 2. Redistributions in binary form must reproduce the above copyright
49 *    notice, this list of conditions and the following disclaimer in the
50 *    documentation and/or other materials provided with the distribution.
51 *
52 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
53 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
54 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
55 * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
56 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
57 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
58 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
59 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
60 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
61 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
62 * POSSIBILITY OF SUCH DAMAGE.
63 */
64
65/*
66 * Copyright (c) 1982, 1986, 1988, 1993
67 *	The Regents of the University of California.  All rights reserved.
68 *
69 * Redistribution and use in source and binary forms, with or without
70 * modification, are permitted provided that the following conditions
71 * are met:
72 * 1. Redistributions of source code must retain the above copyright
73 *    notice, this list of conditions and the following disclaimer.
74 * 2. Redistributions in binary form must reproduce the above copyright
75 *    notice, this list of conditions and the following disclaimer in the
76 *    documentation and/or other materials provided with the distribution.
77 * 3. Neither the name of the University nor the names of its contributors
78 *    may be used to endorse or promote products derived from this software
79 *    without specific prior written permission.
80 *
81 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
82 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
83 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
84 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
85 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
86 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
87 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
88 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
89 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
90 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
91 * SUCH DAMAGE.
92 *
93 *	@(#)ip_icmp.c	8.2 (Berkeley) 1/4/94
94 */
95
96#include <sys/cdefs.h>
97__KERNEL_RCSID(0, "$NetBSD: ip_icmp.c,v 1.128 2012/01/09 14:31:22 liamjfoy Exp $");
98
99#include "opt_ipsec.h"
100
101#include <sys/param.h>
102#include <sys/systm.h>
103#include <sys/malloc.h>
104#include <sys/mbuf.h>
105#include <sys/protosw.h>
106#include <sys/socket.h>
107#include <sys/time.h>
108#include <sys/kernel.h>
109#include <sys/syslog.h>
110#include <sys/sysctl.h>
111
112#include <net/if.h>
113#include <net/route.h>
114
115#include <netinet/in.h>
116#include <netinet/in_systm.h>
117#include <netinet/in_var.h>
118#include <netinet/ip.h>
119#include <netinet/ip_icmp.h>
120#include <netinet/ip_var.h>
121#include <netinet/in_pcb.h>
122#include <netinet/in_proto.h>
123#include <netinet/icmp_var.h>
124#include <netinet/icmp_private.h>
125
126#ifdef KAME_IPSEC
127#include <netinet6/ipsec.h>
128#include <netkey/key.h>
129#endif
130
131#ifdef FAST_IPSEC
132#include <netipsec/ipsec.h>
133#include <netipsec/key.h>
134#endif	/* FAST_IPSEC*/
135
136/*
137 * ICMP routines: error generation, receive packet processing, and
138 * routines to turnaround packets back to the originator, and
139 * host table maintenance routines.
140 */
141
142int	icmpmaskrepl = 0;
143int	icmpbmcastecho = 0;
144#ifdef ICMPPRINTFS
145int	icmpprintfs = 0;
146#endif
147int	icmpreturndatabytes = 8;
148
149percpu_t *icmpstat_percpu;
150
151/*
152 * List of callbacks to notify when Path MTU changes are made.
153 */
154struct icmp_mtudisc_callback {
155	LIST_ENTRY(icmp_mtudisc_callback) mc_list;
156	void (*mc_func)(struct in_addr);
157};
158
159LIST_HEAD(, icmp_mtudisc_callback) icmp_mtudisc_callbacks =
160    LIST_HEAD_INITIALIZER(&icmp_mtudisc_callbacks);
161
162#if 0
163static u_int	ip_next_mtu(u_int, int);
164#else
165/*static*/ u_int	ip_next_mtu(u_int, int);
166#endif
167
168extern int icmperrppslim;
169static int icmperrpps_count = 0;
170static struct timeval icmperrppslim_last;
171static int icmp_rediraccept = 1;
172static int icmp_redirtimeout = 600;
173static struct rttimer_queue *icmp_redirect_timeout_q = NULL;
174
175static void icmp_mtudisc_timeout(struct rtentry *, struct rttimer *);
176static void icmp_redirect_timeout(struct rtentry *, struct rttimer *);
177
178static void sysctl_netinet_icmp_setup(struct sysctllog **);
179
180void
181icmp_init(void)
182{
183
184	sysctl_netinet_icmp_setup(NULL);
185
186	/*
187	 * This is only useful if the user initializes redirtimeout to
188	 * something other than zero.
189	 */
190	if (icmp_redirtimeout != 0) {
191		icmp_redirect_timeout_q =
192			rt_timer_queue_create(icmp_redirtimeout);
193	}
194
195	icmpstat_percpu = percpu_alloc(sizeof(uint64_t) * ICMP_NSTATS);
196}
197
198/*
199 * Register a Path MTU Discovery callback.
200 */
201void
202icmp_mtudisc_callback_register(void (*func)(struct in_addr))
203{
204	struct icmp_mtudisc_callback *mc;
205
206	for (mc = LIST_FIRST(&icmp_mtudisc_callbacks); mc != NULL;
207	     mc = LIST_NEXT(mc, mc_list)) {
208		if (mc->mc_func == func)
209			return;
210	}
211
212	mc = malloc(sizeof(*mc), M_PCB, M_NOWAIT);
213	if (mc == NULL)
214		panic("icmp_mtudisc_callback_register");
215
216	mc->mc_func = func;
217	LIST_INSERT_HEAD(&icmp_mtudisc_callbacks, mc, mc_list);
218}
219
220/*
221 * Generate an error packet of type error
222 * in response to bad packet ip.
223 */
224void
225icmp_error(struct mbuf *n, int type, int code, n_long dest,
226    int destmtu)
227{
228	struct ip *oip = mtod(n, struct ip *), *nip;
229	unsigned oiplen = oip->ip_hl << 2;
230	struct icmp *icp;
231	struct mbuf *m;
232	struct m_tag *mtag;
233	unsigned icmplen, mblen;
234
235#ifdef ICMPPRINTFS
236	if (icmpprintfs)
237		printf("icmp_error(%p, type:%d, code:%d)\n", oip, type, code);
238#endif
239	if (type != ICMP_REDIRECT)
240		ICMP_STATINC(ICMP_STAT_ERROR);
241	/*
242	 * Don't send error if the original packet was encrypted.
243	 * Don't send error if not the first fragment of message.
244	 * Don't error if the old packet protocol was ICMP
245	 * error message, only known informational types.
246	 */
247	if (n->m_flags & M_DECRYPTED)
248		goto freeit;
249	if (oip->ip_off &~ htons(IP_MF|IP_DF))
250		goto freeit;
251	if (oip->ip_p == IPPROTO_ICMP && type != ICMP_REDIRECT &&
252	  n->m_len >= oiplen + ICMP_MINLEN &&
253	  !ICMP_INFOTYPE(((struct icmp *)((char *)oip + oiplen))->icmp_type)) {
254		ICMP_STATINC(ICMP_STAT_OLDICMP);
255		goto freeit;
256	}
257	/* Don't send error in response to a multicast or broadcast packet */
258	if (n->m_flags & (M_BCAST|M_MCAST))
259		goto freeit;
260
261	/*
262	 * First, do a rate limitation check.
263	 */
264	if (icmp_ratelimit(&oip->ip_src, type, code)) {
265		/* XXX stat */
266		goto freeit;
267	}
268
269	/*
270	 * Now, formulate icmp message
271	 */
272	icmplen = oiplen + min(icmpreturndatabytes,
273	    ntohs(oip->ip_len) - oiplen);
274	/*
275	 * Defend against mbuf chains shorter than oip->ip_len - oiplen:
276	 */
277	mblen = 0;
278	for (m = n; m && (mblen < icmplen); m = m->m_next)
279		mblen += m->m_len;
280	icmplen = min(mblen, icmplen);
281
282	/*
283	 * As we are not required to return everything we have,
284	 * we return whatever we can return at ease.
285	 *
286	 * Note that ICMP datagrams longer than 576 octets are out of spec
287	 * according to RFC1812; the limit on icmpreturndatabytes below in
288	 * icmp_sysctl will keep things below that limit.
289	 */
290
291	KASSERT(ICMP_MINLEN <= MCLBYTES);
292
293	if (icmplen + ICMP_MINLEN > MCLBYTES)
294		icmplen = MCLBYTES - ICMP_MINLEN;
295
296	m = m_gethdr(M_DONTWAIT, MT_HEADER);
297	if (m && (icmplen + ICMP_MINLEN > MHLEN)) {
298		MCLGET(m, M_DONTWAIT);
299		if ((m->m_flags & M_EXT) == 0) {
300			m_freem(m);
301			m = NULL;
302		}
303	}
304	if (m == NULL)
305		goto freeit;
306	MCLAIM(m, n->m_owner);
307	m->m_len = icmplen + ICMP_MINLEN;
308	if ((m->m_flags & M_EXT) == 0)
309		MH_ALIGN(m, m->m_len);
310	else {
311		m->m_data += sizeof(struct ip);
312		m->m_len -= sizeof(struct ip);
313	}
314	icp = mtod(m, struct icmp *);
315	if ((u_int)type > ICMP_MAXTYPE)
316		panic("icmp_error");
317	ICMP_STATINC(ICMP_STAT_OUTHIST + type);
318	icp->icmp_type = type;
319	if (type == ICMP_REDIRECT)
320		icp->icmp_gwaddr.s_addr = dest;
321	else {
322		icp->icmp_void = 0;
323		/*
324		 * The following assignments assume an overlay with the
325		 * zeroed icmp_void field.
326		 */
327		if (type == ICMP_PARAMPROB) {
328			icp->icmp_pptr = code;
329			code = 0;
330		} else if (type == ICMP_UNREACH &&
331		    code == ICMP_UNREACH_NEEDFRAG && destmtu)
332			icp->icmp_nextmtu = htons(destmtu);
333	}
334
335	icp->icmp_code = code;
336	m_copydata(n, 0, icmplen, (void *)&icp->icmp_ip);
337
338	/*
339	 * Now, copy old ip header (without options)
340	 * in front of icmp message.
341	 */
342	if ((m->m_flags & M_EXT) == 0 &&
343	    m->m_data - sizeof(struct ip) < m->m_pktdat)
344		panic("icmp len");
345	m->m_data -= sizeof(struct ip);
346	m->m_len += sizeof(struct ip);
347	m->m_pkthdr.len = m->m_len;
348	m->m_pkthdr.rcvif = n->m_pkthdr.rcvif;
349	nip = mtod(m, struct ip *);
350	/* ip_v set in ip_output */
351	nip->ip_hl = sizeof(struct ip) >> 2;
352	nip->ip_tos = 0;
353	nip->ip_len = htons(m->m_len);
354	/* ip_id set in ip_output */
355	nip->ip_off = htons(0);
356	/* ip_ttl set in icmp_reflect */
357	nip->ip_p = IPPROTO_ICMP;
358	nip->ip_src = oip->ip_src;
359	nip->ip_dst = oip->ip_dst;
360	/* move PF m_tag to new packet, if it exists */
361	mtag = m_tag_find(n, PACKET_TAG_PF, NULL);
362	if (mtag != NULL) {
363		m_tag_unlink(n, mtag);
364		m_tag_prepend(m, mtag);
365	}
366	icmp_reflect(m);
367
368freeit:
369	m_freem(n);
370}
371
372struct sockaddr_in icmpsrc = {
373	.sin_len = sizeof (struct sockaddr_in),
374	.sin_family = AF_INET,
375};
376static struct sockaddr_in icmpdst = {
377	.sin_len = sizeof (struct sockaddr_in),
378	.sin_family = AF_INET,
379};
380static struct sockaddr_in icmpgw = {
381	.sin_len = sizeof (struct sockaddr_in),
382	.sin_family = AF_INET,
383};
384struct sockaddr_in icmpmask = {
385	.sin_len = 8,
386	.sin_family = 0,
387};
388
389/*
390 * Process a received ICMP message.
391 */
392void
393icmp_input(struct mbuf *m, ...)
394{
395	int proto;
396	struct icmp *icp;
397	struct ip *ip = mtod(m, struct ip *);
398	int icmplen;
399	int i;
400	struct in_ifaddr *ia;
401	void *(*ctlfunc)(int, const struct sockaddr *, void *);
402	int code;
403	int hlen;
404	va_list ap;
405	struct rtentry *rt;
406
407	va_start(ap, m);
408	hlen = va_arg(ap, int);
409	proto = va_arg(ap, int);
410	va_end(ap);
411
412	/*
413	 * Locate icmp structure in mbuf, and check
414	 * that not corrupted and of at least minimum length.
415	 */
416	icmplen = ntohs(ip->ip_len) - hlen;
417#ifdef ICMPPRINTFS
418	if (icmpprintfs) {
419		printf("icmp_input from `%s' to ", inet_ntoa(ip->ip_src));
420		printf("`%s', len %d\n", inet_ntoa(ip->ip_dst), icmplen);
421	}
422#endif
423	if (icmplen < ICMP_MINLEN) {
424		ICMP_STATINC(ICMP_STAT_TOOSHORT);
425		goto freeit;
426	}
427	i = hlen + min(icmplen, ICMP_ADVLENMIN);
428	if ((m->m_len < i || M_READONLY(m)) && (m = m_pullup(m, i)) == NULL) {
429		ICMP_STATINC(ICMP_STAT_TOOSHORT);
430		return;
431	}
432	ip = mtod(m, struct ip *);
433	m->m_len -= hlen;
434	m->m_data += hlen;
435	icp = mtod(m, struct icmp *);
436	/* Don't need to assert alignment, here. */
437	if (in_cksum(m, icmplen)) {
438		ICMP_STATINC(ICMP_STAT_CHECKSUM);
439		goto freeit;
440	}
441	m->m_len += hlen;
442	m->m_data -= hlen;
443
444#ifdef ICMPPRINTFS
445	/*
446	 * Message type specific processing.
447	 */
448	if (icmpprintfs)
449		printf("icmp_input(type:%d, code:%d)\n", icp->icmp_type,
450		    icp->icmp_code);
451#endif
452	if (icp->icmp_type > ICMP_MAXTYPE)
453		goto raw;
454	ICMP_STATINC(ICMP_STAT_INHIST + icp->icmp_type);
455	code = icp->icmp_code;
456	switch (icp->icmp_type) {
457
458	case ICMP_UNREACH:
459		switch (code) {
460		case ICMP_UNREACH_PROTOCOL:
461			code = PRC_UNREACH_PROTOCOL;
462			break;
463
464		case ICMP_UNREACH_PORT:
465			code = PRC_UNREACH_PORT;
466			break;
467
468		case ICMP_UNREACH_SRCFAIL:
469			code = PRC_UNREACH_SRCFAIL;
470			break;
471
472		case ICMP_UNREACH_NEEDFRAG:
473			code = PRC_MSGSIZE;
474			break;
475
476		case ICMP_UNREACH_NET:
477		case ICMP_UNREACH_NET_UNKNOWN:
478		case ICMP_UNREACH_NET_PROHIB:
479		case ICMP_UNREACH_TOSNET:
480			code = PRC_UNREACH_NET;
481			break;
482
483		case ICMP_UNREACH_HOST:
484		case ICMP_UNREACH_HOST_UNKNOWN:
485		case ICMP_UNREACH_ISOLATED:
486		case ICMP_UNREACH_HOST_PROHIB:
487		case ICMP_UNREACH_TOSHOST:
488		case ICMP_UNREACH_ADMIN_PROHIBIT:
489		case ICMP_UNREACH_HOST_PREC:
490		case ICMP_UNREACH_PREC_CUTOFF:
491			code = PRC_UNREACH_HOST;
492			break;
493
494		default:
495			goto badcode;
496		}
497		goto deliver;
498
499	case ICMP_TIMXCEED:
500		if (code > 1)
501			goto badcode;
502		code += PRC_TIMXCEED_INTRANS;
503		goto deliver;
504
505	case ICMP_PARAMPROB:
506		if (code > 1)
507			goto badcode;
508		code = PRC_PARAMPROB;
509		goto deliver;
510
511	case ICMP_SOURCEQUENCH:
512		if (code)
513			goto badcode;
514		code = PRC_QUENCH;
515		goto deliver;
516
517	deliver:
518		/*
519		 * Problem with datagram; advise higher level routines.
520		 */
521		if (icmplen < ICMP_ADVLENMIN || icmplen < ICMP_ADVLEN(icp) ||
522		    icp->icmp_ip.ip_hl < (sizeof(struct ip) >> 2)) {
523			ICMP_STATINC(ICMP_STAT_BADLEN);
524			goto freeit;
525		}
526		if (IN_MULTICAST(icp->icmp_ip.ip_dst.s_addr))
527			goto badcode;
528#ifdef ICMPPRINTFS
529		if (icmpprintfs)
530			printf("deliver to protocol %d\n", icp->icmp_ip.ip_p);
531#endif
532		icmpsrc.sin_addr = icp->icmp_ip.ip_dst;
533		ctlfunc = inetsw[ip_protox[icp->icmp_ip.ip_p]].pr_ctlinput;
534		if (ctlfunc)
535			(void) (*ctlfunc)(code, sintosa(&icmpsrc),
536			    &icp->icmp_ip);
537		break;
538
539	badcode:
540		ICMP_STATINC(ICMP_STAT_BADCODE);
541		break;
542
543	case ICMP_ECHO:
544		if (!icmpbmcastecho &&
545		    (m->m_flags & (M_MCAST | M_BCAST)) != 0)  {
546			ICMP_STATINC(ICMP_STAT_BMCASTECHO);
547			break;
548		}
549		icp->icmp_type = ICMP_ECHOREPLY;
550		goto reflect;
551
552	case ICMP_TSTAMP:
553		if (icmplen < ICMP_TSLEN) {
554			ICMP_STATINC(ICMP_STAT_BADLEN);
555			break;
556		}
557		if (!icmpbmcastecho &&
558		    (m->m_flags & (M_MCAST | M_BCAST)) != 0)  {
559			ICMP_STATINC(ICMP_STAT_BMCASTTSTAMP);
560			break;
561		}
562		icp->icmp_type = ICMP_TSTAMPREPLY;
563		icp->icmp_rtime = iptime();
564		icp->icmp_ttime = icp->icmp_rtime;	/* bogus, do later! */
565		goto reflect;
566
567	case ICMP_MASKREQ:
568		if (icmpmaskrepl == 0)
569			break;
570		/*
571		 * We are not able to respond with all ones broadcast
572		 * unless we receive it over a point-to-point interface.
573		 */
574		if (icmplen < ICMP_MASKLEN) {
575			ICMP_STATINC(ICMP_STAT_BADLEN);
576			break;
577		}
578		if (ip->ip_dst.s_addr == INADDR_BROADCAST ||
579		    in_nullhost(ip->ip_dst))
580			icmpdst.sin_addr = ip->ip_src;
581		else
582			icmpdst.sin_addr = ip->ip_dst;
583		ia = ifatoia(ifaof_ifpforaddr(sintosa(&icmpdst),
584		    m->m_pkthdr.rcvif));
585		if (ia == 0)
586			break;
587		icp->icmp_type = ICMP_MASKREPLY;
588		icp->icmp_mask = ia->ia_sockmask.sin_addr.s_addr;
589		if (in_nullhost(ip->ip_src)) {
590			if (ia->ia_ifp->if_flags & IFF_BROADCAST)
591				ip->ip_src = ia->ia_broadaddr.sin_addr;
592			else if (ia->ia_ifp->if_flags & IFF_POINTOPOINT)
593				ip->ip_src = ia->ia_dstaddr.sin_addr;
594		}
595reflect:
596		{
597			uint64_t *icps = percpu_getref(icmpstat_percpu);
598			icps[ICMP_STAT_REFLECT]++;
599			icps[ICMP_STAT_OUTHIST + icp->icmp_type]++;
600			percpu_putref(icmpstat_percpu);
601		}
602		icmp_reflect(m);
603		return;
604
605	case ICMP_REDIRECT:
606		if (code > 3)
607			goto badcode;
608		if (icmp_rediraccept == 0)
609			goto freeit;
610		if (icmplen < ICMP_ADVLENMIN || icmplen < ICMP_ADVLEN(icp) ||
611		    icp->icmp_ip.ip_hl < (sizeof(struct ip) >> 2)) {
612			ICMP_STATINC(ICMP_STAT_BADLEN);
613			break;
614		}
615		/*
616		 * Short circuit routing redirects to force
617		 * immediate change in the kernel's routing
618		 * tables.  The message is also handed to anyone
619		 * listening on a raw socket (e.g. the routing
620		 * daemon for use in updating its tables).
621		 */
622		icmpgw.sin_addr = ip->ip_src;
623		icmpdst.sin_addr = icp->icmp_gwaddr;
624#ifdef	ICMPPRINTFS
625		if (icmpprintfs) {
626			printf("redirect dst `%s' to `%s'\n",
627			    inet_ntoa(icp->icmp_ip.ip_dst),
628			    inet_ntoa(icp->icmp_gwaddr));
629		}
630#endif
631		icmpsrc.sin_addr = icp->icmp_ip.ip_dst;
632		rt = NULL;
633		rtredirect(sintosa(&icmpsrc), sintosa(&icmpdst),
634		    NULL, RTF_GATEWAY | RTF_HOST, sintosa(&icmpgw), &rt);
635		if (rt != NULL && icmp_redirtimeout != 0) {
636			i = rt_timer_add(rt, icmp_redirect_timeout,
637					 icmp_redirect_timeout_q);
638			if (i)
639				log(LOG_ERR, "ICMP:  redirect failed to "
640				    "register timeout for route to %x, "
641				    "code %d\n",
642				    icp->icmp_ip.ip_dst.s_addr, i);
643		}
644		if (rt != NULL)
645			rtfree(rt);
646
647		pfctlinput(PRC_REDIRECT_HOST, sintosa(&icmpsrc));
648#if defined(KAME_IPSEC) || defined(FAST_IPSEC)
649		key_sa_routechange((struct sockaddr *)&icmpsrc);
650#endif
651		break;
652
653	/*
654	 * No kernel processing for the following;
655	 * just fall through to send to raw listener.
656	 */
657	case ICMP_ECHOREPLY:
658	case ICMP_ROUTERADVERT:
659	case ICMP_ROUTERSOLICIT:
660	case ICMP_TSTAMPREPLY:
661	case ICMP_IREQREPLY:
662	case ICMP_MASKREPLY:
663	default:
664		break;
665	}
666
667raw:
668	rip_input(m, hlen, proto);
669	return;
670
671freeit:
672	m_freem(m);
673	return;
674}
675
676/*
677 * Reflect the ip packet back to the source
678 */
679void
680icmp_reflect(struct mbuf *m)
681{
682	struct ip *ip = mtod(m, struct ip *);
683	struct in_ifaddr *ia;
684	struct ifaddr *ifa;
685	struct sockaddr_in *sin = 0;
686	struct in_addr t;
687	struct mbuf *opts = 0;
688	int optlen = (ip->ip_hl << 2) - sizeof(struct ip);
689
690	if (!in_canforward(ip->ip_src) &&
691	    ((ip->ip_src.s_addr & IN_CLASSA_NET) !=
692	     htonl(IN_LOOPBACKNET << IN_CLASSA_NSHIFT))) {
693		m_freem(m);	/* Bad return address */
694		goto done;	/* ip_output() will check for broadcast */
695	}
696	t = ip->ip_dst;
697	ip->ip_dst = ip->ip_src;
698	/*
699	 * If the incoming packet was addressed directly to us, use
700	 * dst as the src for the reply.  Otherwise (broadcast or
701	 * anonymous), use an address which corresponds to the
702	 * incoming interface, with a preference for the address which
703	 * corresponds to the route to the destination of the ICMP.
704	 */
705
706	/* Look for packet addressed to us */
707	INADDR_TO_IA(t, ia);
708
709	/* look for packet sent to broadcast address */
710	if (ia == NULL && m->m_pkthdr.rcvif &&
711	    (m->m_pkthdr.rcvif->if_flags & IFF_BROADCAST)) {
712		IFADDR_FOREACH(ifa, m->m_pkthdr.rcvif) {
713			if (ifa->ifa_addr->sa_family != AF_INET)
714				continue;
715			if (in_hosteq(t,ifatoia(ifa)->ia_broadaddr.sin_addr)) {
716				ia = ifatoia(ifa);
717				break;
718			}
719		}
720	}
721
722	if (ia)
723		sin = &ia->ia_addr;
724
725	icmpdst.sin_addr = t;
726
727	/*
728	 * if the packet is addressed somewhere else, compute the
729	 * source address for packets routed back to the source, and
730	 * use that, if it's an address on the interface which
731	 * received the packet
732	 */
733	if (sin == NULL && m->m_pkthdr.rcvif) {
734		struct sockaddr_in sin_dst;
735		struct route icmproute;
736		int errornum;
737
738		sockaddr_in_init(&sin_dst, &ip->ip_dst, 0);
739		memset(&icmproute, 0, sizeof(icmproute));
740		errornum = 0;
741		sin = in_selectsrc(&sin_dst, &icmproute, 0, NULL, &errornum);
742		/* errornum is never used */
743		rtcache_free(&icmproute);
744		/* check to make sure sin is a source address on rcvif */
745		if (sin) {
746			t = sin->sin_addr;
747			sin = NULL;
748			INADDR_TO_IA(t, ia);
749			while (ia) {
750				if (ia->ia_ifp == m->m_pkthdr.rcvif) {
751					sin = &ia->ia_addr;
752					break;
753				}
754				NEXT_IA_WITH_SAME_ADDR(ia);
755			}
756		}
757	}
758
759	/*
760	 * if it was not addressed to us, but the route doesn't go out
761	 * the source interface, pick an address on the source
762	 * interface.  This can happen when routing is asymmetric, or
763	 * when the incoming packet was encapsulated
764	 */
765	if (sin == NULL && m->m_pkthdr.rcvif) {
766		IFADDR_FOREACH(ifa, m->m_pkthdr.rcvif) {
767			if (ifa->ifa_addr->sa_family != AF_INET)
768				continue;
769			sin = &(ifatoia(ifa)->ia_addr);
770			break;
771		}
772	}
773
774	/*
775	 * The following happens if the packet was not addressed to us,
776	 * and was received on an interface with no IP address:
777	 * We find the first AF_INET address on the first non-loopback
778	 * interface.
779	 */
780	if (sin == NULL)
781		TAILQ_FOREACH(ia, &in_ifaddrhead, ia_list) {
782			if (ia->ia_ifp->if_flags & IFF_LOOPBACK)
783				continue;
784			sin = &ia->ia_addr;
785			break;
786		}
787
788	/*
789	 * If we still didn't find an address, punt.  We could have an
790	 * interface up (and receiving packets) with no address.
791	 */
792	if (sin == NULL) {
793		m_freem(m);
794		goto done;
795	}
796
797	ip->ip_src = sin->sin_addr;
798	ip->ip_ttl = MAXTTL;
799
800	if (optlen > 0) {
801		u_char *cp;
802		int opt, cnt;
803		u_int len;
804
805		/*
806		 * Retrieve any source routing from the incoming packet;
807		 * add on any record-route or timestamp options.
808		 */
809		cp = (u_char *) (ip + 1);
810		if ((opts = ip_srcroute()) == 0 &&
811		    (opts = m_gethdr(M_DONTWAIT, MT_HEADER))) {
812			MCLAIM(opts, m->m_owner);
813			opts->m_len = sizeof(struct in_addr);
814			*mtod(opts, struct in_addr *) = zeroin_addr;
815		}
816		if (opts) {
817#ifdef ICMPPRINTFS
818		    if (icmpprintfs)
819			    printf("icmp_reflect optlen %d rt %d => ",
820				optlen, opts->m_len);
821#endif
822		    for (cnt = optlen; cnt > 0; cnt -= len, cp += len) {
823			    opt = cp[IPOPT_OPTVAL];
824			    if (opt == IPOPT_EOL)
825				    break;
826			    if (opt == IPOPT_NOP)
827				    len = 1;
828			    else {
829				    if (cnt < IPOPT_OLEN + sizeof(*cp))
830					    break;
831				    len = cp[IPOPT_OLEN];
832				    if (len < IPOPT_OLEN + sizeof(*cp) ||
833				        len > cnt)
834					    break;
835			    }
836			    /*
837			     * Should check for overflow, but it "can't happen"
838			     */
839			    if (opt == IPOPT_RR || opt == IPOPT_TS ||
840				opt == IPOPT_SECURITY) {
841				    memmove(mtod(opts, char *) + opts->m_len,
842					cp, len);
843				    opts->m_len += len;
844			    }
845		    }
846		    /* Terminate & pad, if necessary */
847		    if ((cnt = opts->m_len % 4) != 0) {
848			    for (; cnt < 4; cnt++) {
849				    *(mtod(opts, char *) + opts->m_len) =
850					IPOPT_EOL;
851				    opts->m_len++;
852			    }
853		    }
854#ifdef ICMPPRINTFS
855		    if (icmpprintfs)
856			    printf("%d\n", opts->m_len);
857#endif
858		}
859		/*
860		 * Now strip out original options by copying rest of first
861		 * mbuf's data back, and adjust the IP length.
862		 */
863		ip->ip_len = htons(ntohs(ip->ip_len) - optlen);
864		ip->ip_hl = sizeof(struct ip) >> 2;
865		m->m_len -= optlen;
866		if (m->m_flags & M_PKTHDR)
867			m->m_pkthdr.len -= optlen;
868		optlen += sizeof(struct ip);
869		memmove(ip + 1, (char *)ip + optlen,
870		    (unsigned)(m->m_len - sizeof(struct ip)));
871	}
872	m_tag_delete_nonpersistent(m);
873	m->m_flags &= ~(M_BCAST|M_MCAST);
874
875	/*
876	 * Clear any in-bound checksum flags for this packet.
877	 */
878	if (m->m_flags & M_PKTHDR)
879		m->m_pkthdr.csum_flags = 0;
880
881	icmp_send(m, opts);
882done:
883	if (opts)
884		(void)m_free(opts);
885}
886
887/*
888 * Send an icmp packet back to the ip level,
889 * after supplying a checksum.
890 */
891void
892icmp_send(struct mbuf *m, struct mbuf *opts)
893{
894	struct ip *ip = mtod(m, struct ip *);
895	int hlen;
896	struct icmp *icp;
897
898	hlen = ip->ip_hl << 2;
899	m->m_data += hlen;
900	m->m_len -= hlen;
901	icp = mtod(m, struct icmp *);
902	icp->icmp_cksum = 0;
903	icp->icmp_cksum = in_cksum(m, ntohs(ip->ip_len) - hlen);
904	m->m_data -= hlen;
905	m->m_len += hlen;
906#ifdef ICMPPRINTFS
907	if (icmpprintfs) {
908		printf("icmp_send to destination `%s' from `%s'\n",
909		    inet_ntoa(ip->ip_dst), inet_ntoa(ip->ip_src));
910	}
911#endif
912	(void)ip_output(m, opts, NULL, 0, NULL, NULL);
913}
914
915n_time
916iptime(void)
917{
918	struct timeval atv;
919	u_long t;
920
921	microtime(&atv);
922	t = (atv.tv_sec % (24*60*60)) * 1000 + atv.tv_usec / 1000;
923	return (htonl(t));
924}
925
926/*
927 * sysctl helper routine for net.inet.icmp.returndatabytes.  ensures
928 * that the new value is in the correct range.
929 */
930static int
931sysctl_net_inet_icmp_returndatabytes(SYSCTLFN_ARGS)
932{
933	int error, t;
934	struct sysctlnode node;
935
936	node = *rnode;
937	node.sysctl_data = &t;
938	t = icmpreturndatabytes;
939	error = sysctl_lookup(SYSCTLFN_CALL(&node));
940	if (error || newp == NULL)
941		return (error);
942
943	if (t < 8 || t > 512)
944		return (EINVAL);
945	icmpreturndatabytes = t;
946
947	return (0);
948}
949
950/*
951 * sysctl helper routine for net.inet.icmp.redirtimeout.  ensures that
952 * the given value is not less than zero and then resets the timeout
953 * queue.
954 */
955static int
956sysctl_net_inet_icmp_redirtimeout(SYSCTLFN_ARGS)
957{
958	int error, tmp;
959	struct sysctlnode node;
960
961	node = *rnode;
962	node.sysctl_data = &tmp;
963	tmp = icmp_redirtimeout;
964	error = sysctl_lookup(SYSCTLFN_CALL(&node));
965	if (error || newp == NULL)
966		return (error);
967	if (tmp < 0)
968		return (EINVAL);
969	icmp_redirtimeout = tmp;
970
971	/*
972	 * was it a *defined* side-effect that anyone even *reading*
973	 * this value causes these things to happen?
974	 */
975	if (icmp_redirect_timeout_q != NULL) {
976		if (icmp_redirtimeout == 0) {
977			rt_timer_queue_destroy(icmp_redirect_timeout_q,
978			    true);
979			icmp_redirect_timeout_q = NULL;
980		} else {
981			rt_timer_queue_change(icmp_redirect_timeout_q,
982			    icmp_redirtimeout);
983		}
984	} else if (icmp_redirtimeout > 0) {
985		icmp_redirect_timeout_q =
986		    rt_timer_queue_create(icmp_redirtimeout);
987	}
988
989	return (0);
990}
991
992static int
993sysctl_net_inet_icmp_stats(SYSCTLFN_ARGS)
994{
995
996	return (NETSTAT_SYSCTL(icmpstat_percpu, ICMP_NSTATS));
997}
998
999static void
1000sysctl_netinet_icmp_setup(struct sysctllog **clog)
1001{
1002
1003	sysctl_createv(clog, 0, NULL, NULL,
1004		       CTLFLAG_PERMANENT,
1005		       CTLTYPE_NODE, "net", NULL,
1006		       NULL, 0, NULL, 0,
1007		       CTL_NET, CTL_EOL);
1008	sysctl_createv(clog, 0, NULL, NULL,
1009		       CTLFLAG_PERMANENT,
1010		       CTLTYPE_NODE, "inet", NULL,
1011		       NULL, 0, NULL, 0,
1012		       CTL_NET, PF_INET, CTL_EOL);
1013	sysctl_createv(clog, 0, NULL, NULL,
1014		       CTLFLAG_PERMANENT,
1015		       CTLTYPE_NODE, "icmp",
1016		       SYSCTL_DESCR("ICMPv4 related settings"),
1017		       NULL, 0, NULL, 0,
1018		       CTL_NET, PF_INET, IPPROTO_ICMP, CTL_EOL);
1019
1020	sysctl_createv(clog, 0, NULL, NULL,
1021		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1022		       CTLTYPE_INT, "maskrepl",
1023		       SYSCTL_DESCR("Respond to ICMP_MASKREQ messages"),
1024		       NULL, 0, &icmpmaskrepl, 0,
1025		       CTL_NET, PF_INET, IPPROTO_ICMP,
1026		       ICMPCTL_MASKREPL, CTL_EOL);
1027	sysctl_createv(clog, 0, NULL, NULL,
1028		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1029		       CTLTYPE_INT, "returndatabytes",
1030		       SYSCTL_DESCR("Number of bytes to return in an ICMP "
1031				    "error message"),
1032		       sysctl_net_inet_icmp_returndatabytes, 0,
1033		       &icmpreturndatabytes, 0,
1034		       CTL_NET, PF_INET, IPPROTO_ICMP,
1035		       ICMPCTL_RETURNDATABYTES, CTL_EOL);
1036	sysctl_createv(clog, 0, NULL, NULL,
1037		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1038		       CTLTYPE_INT, "errppslimit",
1039		       SYSCTL_DESCR("Maximum number of outgoing ICMP error "
1040				    "messages per second"),
1041		       NULL, 0, &icmperrppslim, 0,
1042		       CTL_NET, PF_INET, IPPROTO_ICMP,
1043		       ICMPCTL_ERRPPSLIMIT, CTL_EOL);
1044	sysctl_createv(clog, 0, NULL, NULL,
1045		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1046		       CTLTYPE_INT, "rediraccept",
1047		       SYSCTL_DESCR("Accept ICMP_REDIRECT messages"),
1048		       NULL, 0, &icmp_rediraccept, 0,
1049		       CTL_NET, PF_INET, IPPROTO_ICMP,
1050		       ICMPCTL_REDIRACCEPT, CTL_EOL);
1051	sysctl_createv(clog, 0, NULL, NULL,
1052		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1053		       CTLTYPE_INT, "redirtimeout",
1054		       SYSCTL_DESCR("Lifetime of ICMP_REDIRECT generated "
1055				    "routes"),
1056		       sysctl_net_inet_icmp_redirtimeout, 0,
1057		       &icmp_redirtimeout, 0,
1058		       CTL_NET, PF_INET, IPPROTO_ICMP,
1059		       ICMPCTL_REDIRTIMEOUT, CTL_EOL);
1060	sysctl_createv(clog, 0, NULL, NULL,
1061		       CTLFLAG_PERMANENT,
1062		       CTLTYPE_STRUCT, "stats",
1063		       SYSCTL_DESCR("ICMP statistics"),
1064		       sysctl_net_inet_icmp_stats, 0, NULL, 0,
1065		       CTL_NET, PF_INET, IPPROTO_ICMP, ICMPCTL_STATS,
1066		       CTL_EOL);
1067	sysctl_createv(clog, 0, NULL, NULL,
1068		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
1069		       CTLTYPE_INT, "bmcastecho",
1070		       SYSCTL_DESCR("Respond to ICMP_ECHO or ICMP_TIMESTAMP "
1071				    "message to the broadcast or multicast"),
1072		       NULL, 0, &icmpbmcastecho, 0,
1073		       CTL_NET, PF_INET, IPPROTO_ICMP, ICMPCTL_BMCASTECHO,
1074		       CTL_EOL);
1075}
1076
1077void
1078icmp_statinc(u_int stat)
1079{
1080
1081	KASSERT(stat < ICMP_NSTATS);
1082	ICMP_STATINC(stat);
1083}
1084
1085/* Table of common MTUs: */
1086
1087static const u_int mtu_table[] = {
1088	65535, 65280, 32000, 17914, 9180, 8166,
1089	4352, 2002, 1492, 1006, 508, 296, 68, 0
1090};
1091
1092void
1093icmp_mtudisc(struct icmp *icp, struct in_addr faddr)
1094{
1095	struct icmp_mtudisc_callback *mc;
1096	struct sockaddr *dst = sintosa(&icmpsrc);
1097	struct rtentry *rt;
1098	u_long mtu = ntohs(icp->icmp_nextmtu);  /* Why a long?  IPv6 */
1099	int    error;
1100
1101	rt = rtalloc1(dst, 1);
1102	if (rt == 0)
1103		return;
1104
1105	/* If we didn't get a host route, allocate one */
1106
1107	if ((rt->rt_flags & RTF_HOST) == 0) {
1108		struct rtentry *nrt;
1109
1110		error = rtrequest((int) RTM_ADD, dst,
1111		    (struct sockaddr *) rt->rt_gateway, NULL,
1112		    RTF_GATEWAY | RTF_HOST | RTF_DYNAMIC, &nrt);
1113		if (error) {
1114			rtfree(rt);
1115			return;
1116		}
1117		nrt->rt_rmx = rt->rt_rmx;
1118		rtfree(rt);
1119		rt = nrt;
1120	}
1121	error = rt_timer_add(rt, icmp_mtudisc_timeout, ip_mtudisc_timeout_q);
1122	if (error) {
1123		rtfree(rt);
1124		return;
1125	}
1126
1127	if (mtu == 0) {
1128		int i = 0;
1129
1130		mtu = ntohs(icp->icmp_ip.ip_len);
1131		/* Some 4.2BSD-based routers incorrectly adjust the ip_len */
1132		if (mtu > rt->rt_rmx.rmx_mtu && rt->rt_rmx.rmx_mtu != 0)
1133			mtu -= (icp->icmp_ip.ip_hl << 2);
1134
1135		/* If we still can't guess a value, try the route */
1136
1137		if (mtu == 0) {
1138			mtu = rt->rt_rmx.rmx_mtu;
1139
1140			/* If no route mtu, default to the interface mtu */
1141
1142			if (mtu == 0)
1143				mtu = rt->rt_ifp->if_mtu;
1144		}
1145
1146		for (i = 0; i < sizeof(mtu_table) / sizeof(mtu_table[0]); i++)
1147			if (mtu > mtu_table[i]) {
1148				mtu = mtu_table[i];
1149				break;
1150			}
1151	}
1152
1153	/*
1154	 * XXX:   RTV_MTU is overloaded, since the admin can set it
1155	 *	  to turn off PMTU for a route, and the kernel can
1156	 *	  set it to indicate a serious problem with PMTU
1157	 *	  on a route.  We should be using a separate flag
1158	 *	  for the kernel to indicate this.
1159	 */
1160
1161	if ((rt->rt_rmx.rmx_locks & RTV_MTU) == 0) {
1162		if (mtu < 296 || mtu > rt->rt_ifp->if_mtu)
1163			rt->rt_rmx.rmx_locks |= RTV_MTU;
1164		else if (rt->rt_rmx.rmx_mtu > mtu ||
1165			 rt->rt_rmx.rmx_mtu == 0) {
1166			ICMP_STATINC(ICMP_STAT_PMTUCHG);
1167			rt->rt_rmx.rmx_mtu = mtu;
1168		}
1169	}
1170
1171	if (rt)
1172		rtfree(rt);
1173
1174	/*
1175	 * Notify protocols that the MTU for this destination
1176	 * has changed.
1177	 */
1178	for (mc = LIST_FIRST(&icmp_mtudisc_callbacks); mc != NULL;
1179	     mc = LIST_NEXT(mc, mc_list))
1180		(*mc->mc_func)(faddr);
1181}
1182
1183/*
1184 * Return the next larger or smaller MTU plateau (table from RFC 1191)
1185 * given current value MTU.  If DIR is less than zero, a larger plateau
1186 * is returned; otherwise, a smaller value is returned.
1187 */
1188u_int
1189ip_next_mtu(u_int mtu, int dir)	/* XXX */
1190{
1191	int i;
1192
1193	for (i = 0; i < (sizeof mtu_table) / (sizeof mtu_table[0]); i++) {
1194		if (mtu >= mtu_table[i])
1195			break;
1196	}
1197
1198	if (dir < 0) {
1199		if (i == 0) {
1200			return 0;
1201		} else {
1202			return mtu_table[i - 1];
1203		}
1204	} else {
1205		if (mtu_table[i] == 0) {
1206			return 0;
1207		} else if (mtu > mtu_table[i]) {
1208			return mtu_table[i];
1209		} else {
1210			return mtu_table[i + 1];
1211		}
1212	}
1213}
1214
1215static void
1216icmp_mtudisc_timeout(struct rtentry *rt, struct rttimer *r)
1217{
1218	if (rt == NULL)
1219		panic("icmp_mtudisc_timeout:  bad route to timeout");
1220	if ((rt->rt_flags & (RTF_DYNAMIC | RTF_HOST)) ==
1221	    (RTF_DYNAMIC | RTF_HOST)) {
1222		rtrequest((int) RTM_DELETE, rt_getkey(rt),
1223		    rt->rt_gateway, rt_mask(rt), rt->rt_flags, 0);
1224	} else {
1225		if ((rt->rt_rmx.rmx_locks & RTV_MTU) == 0) {
1226			rt->rt_rmx.rmx_mtu = 0;
1227		}
1228	}
1229}
1230
1231static void
1232icmp_redirect_timeout(struct rtentry *rt, struct rttimer *r)
1233{
1234	if (rt == NULL)
1235		panic("icmp_redirect_timeout:  bad route to timeout");
1236	if ((rt->rt_flags & (RTF_DYNAMIC | RTF_HOST)) ==
1237	    (RTF_DYNAMIC | RTF_HOST)) {
1238		rtrequest((int) RTM_DELETE, rt_getkey(rt),
1239		    rt->rt_gateway, rt_mask(rt), rt->rt_flags, 0);
1240	}
1241}
1242
1243/*
1244 * Perform rate limit check.
1245 * Returns 0 if it is okay to send the icmp packet.
1246 * Returns 1 if the router SHOULD NOT send this icmp packet due to rate
1247 * limitation.
1248 *
1249 * XXX per-destination/type check necessary?
1250 */
1251int
1252icmp_ratelimit(const struct in_addr *dst, const int type,
1253    const int code)
1254{
1255
1256	/* PPS limit */
1257	if (!ppsratecheck(&icmperrppslim_last, &icmperrpps_count,
1258	    icmperrppslim)) {
1259		/* The packet is subject to rate limit */
1260		return 1;
1261	}
1262
1263	/* okay to send */
1264	return 0;
1265}
1266