2 * Copyright (c) 2004, 2005 The DragonFly Project. All rights reserved.
4 * This code is derived from software contributed to The DragonFly Project
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
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 DragonFly Project nor the names of its
16 * contributors may be used to endorse or promote products derived
17 * from this software without specific, prior written permission.
19 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
22 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
23 * COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
24 * INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING,
25 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
26 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED
27 * AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
28 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
29 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * Copyright (c) 1982, 1986, 1988, 1993
35 * The Regents of the University of California. All rights reserved.
37 * Redistribution and use in source and binary forms, with or without
38 * modification, are permitted provided that the following conditions
40 * 1. Redistributions of source code must retain the above copyright
41 * notice, this list of conditions and the following disclaimer.
42 * 2. Redistributions in binary form must reproduce the above copyright
43 * notice, this list of conditions and the following disclaimer in the
44 * documentation and/or other materials provided with the distribution.
45 * 3. All advertising materials mentioning features or use of this software
46 * must display the following acknowledgement:
47 * This product includes software developed by the University of
48 * California, Berkeley and its contributors.
49 * 4. Neither the name of the University nor the names of its contributors
50 * may be used to endorse or promote products derived from this software
51 * without specific prior written permission.
53 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
54 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
55 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
56 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
57 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
58 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
59 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
60 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
61 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
62 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
65 * @(#)if_ether.c 8.1 (Berkeley) 6/10/93
66 * $FreeBSD: src/sys/netinet/if_ether.c,v 1.64.2.23 2003/04/11 07:23:15 fjoe Exp $
67 * $DragonFly: src/sys/netinet/if_ether.c,v 1.40 2007/08/27 13:15:14 hasso Exp $
71 * Ethernet address resolution protocol.
73 * add "inuse/lock" bit (or ref. count) along with valid bit
79 #include <sys/param.h>
80 #include <sys/kernel.h>
81 #include <sys/queue.h>
82 #include <sys/sysctl.h>
83 #include <sys/systm.h>
85 #include <sys/malloc.h>
86 #include <sys/socket.h>
87 #include <sys/syslog.h>
90 #include <net/if_dl.h>
91 #include <net/if_types.h>
92 #include <net/route.h>
93 #include <net/netisr.h>
94 #include <net/if_llc.h>
96 #include <netinet/in.h>
97 #include <netinet/in_var.h>
98 #include <netinet/if_ether.h>
100 #include <net/iso88025.h>
102 #include <sys/thread2.h>
103 #include <sys/msgport2.h>
104 #include <net/netmsg2.h>
108 #include <netinet/ip_carp.h>
111 #define SIN(s) ((struct sockaddr_in *)s)
112 #define SDL(s) ((struct sockaddr_dl *)s)
114 SYSCTL_DECL(_net_link_ether
);
115 SYSCTL_NODE(_net_link_ether
, PF_INET
, inet
, CTLFLAG_RW
, 0, "");
118 static int arpt_prune
= (5*60*1); /* walk list every 5 minutes */
119 static int arpt_keep
= (20*60); /* once resolved, good for 20 more minutes */
120 static int arpt_down
= 20; /* once declared down, don't send for 20 sec */
122 SYSCTL_INT(_net_link_ether_inet
, OID_AUTO
, prune_intvl
, CTLFLAG_RW
,
124 SYSCTL_INT(_net_link_ether_inet
, OID_AUTO
, max_age
, CTLFLAG_RW
,
126 SYSCTL_INT(_net_link_ether_inet
, OID_AUTO
, host_down_time
, CTLFLAG_RW
,
129 #define rt_expire rt_rmx.rmx_expire
132 LIST_ENTRY(llinfo_arp
) la_le
;
133 struct rtentry
*la_rt
;
134 struct mbuf
*la_hold
; /* last packet until resolved/timeout */
135 u_short la_preempt
; /* countdown for pre-expiry arps */
136 u_short la_asked
; /* #times we QUERIED following expiration */
139 static LIST_HEAD(, llinfo_arp
) llinfo_arp_list
[MAXCPU
];
141 static int arp_maxtries
= 5;
142 static int useloopback
= 1; /* use loopback interface for local traffic */
143 static int arp_proxyall
= 0;
145 SYSCTL_INT(_net_link_ether_inet
, OID_AUTO
, maxtries
, CTLFLAG_RW
,
146 &arp_maxtries
, 0, "");
147 SYSCTL_INT(_net_link_ether_inet
, OID_AUTO
, useloopback
, CTLFLAG_RW
,
148 &useloopback
, 0, "");
149 SYSCTL_INT(_net_link_ether_inet
, OID_AUTO
, proxyall
, CTLFLAG_RW
,
150 &arp_proxyall
, 0, "");
152 void arprequest_acces(struct ifnet
*ifp
, struct in_addr
*sip
, struct in_addr
*tip
, u_char
*enaddr
);
153 static void arp_rtrequest (int, struct rtentry
*, struct rt_addrinfo
*);
154 static void arprequest (struct ifnet
*,
155 struct in_addr
*, struct in_addr
*, u_char
*);
156 static void arpintr(struct netmsg
*);
157 static void arptfree (struct llinfo_arp
*);
158 static void arptimer (void *);
159 static struct llinfo_arp
160 *arplookup (in_addr_t addr
, boolean_t create
, boolean_t proxy
);
162 static void in_arpinput (struct mbuf
*);
165 static struct callout arptimer_ch
[MAXCPU
];
168 * Timeout routine. Age arp_tab entries periodically.
172 arptimer(void *ignored_arg
)
174 struct llinfo_arp
*la
, *nla
;
177 LIST_FOREACH_MUTABLE(la
, &llinfo_arp_list
[mycpuid
], la_le
, nla
) {
178 if (la
->la_rt
->rt_expire
&& la
->la_rt
->rt_expire
<= time_second
)
181 callout_reset(&arptimer_ch
[mycpuid
], arpt_prune
* hz
, arptimer
, NULL
);
186 * Parallel to llc_rtrequest.
189 arp_rtrequest(int req
, struct rtentry
*rt
, struct rt_addrinfo
*info
)
191 struct sockaddr
*gate
= rt
->rt_gateway
;
192 struct llinfo_arp
*la
= rt
->rt_llinfo
;
194 struct sockaddr_dl null_sdl
= { sizeof null_sdl
, AF_LINK
};
195 static boolean_t arpinit_done
[MAXCPU
];
197 if (!arpinit_done
[mycpuid
]) {
198 arpinit_done
[mycpuid
] = TRUE
;
199 callout_init(&arptimer_ch
[mycpuid
]);
200 callout_reset(&arptimer_ch
[mycpuid
], hz
, arptimer
, NULL
);
202 if (rt
->rt_flags
& RTF_GATEWAY
)
208 * XXX: If this is a manually added route to interface
209 * such as older version of routed or gated might provide,
210 * restore cloning bit.
212 if (!(rt
->rt_flags
& RTF_HOST
) &&
213 SIN(rt_mask(rt
))->sin_addr
.s_addr
!= 0xffffffff)
214 rt
->rt_flags
|= RTF_CLONING
;
215 if (rt
->rt_flags
& RTF_CLONING
) {
217 * Case 1: This route should come from a route to iface.
219 rt_setgate(rt
, rt_key(rt
),
220 (struct sockaddr
*)&null_sdl
);
221 gate
= rt
->rt_gateway
;
222 SDL(gate
)->sdl_type
= rt
->rt_ifp
->if_type
;
223 SDL(gate
)->sdl_index
= rt
->rt_ifp
->if_index
;
224 rt
->rt_expire
= time_second
;
227 /* Announce a new entry if requested. */
228 if (rt
->rt_flags
& RTF_ANNOUNCE
)
229 arprequest(rt
->rt_ifp
,
230 &SIN(rt_key(rt
))->sin_addr
,
231 &SIN(rt_key(rt
))->sin_addr
,
235 if (gate
->sa_family
!= AF_LINK
||
236 gate
->sa_len
< sizeof(struct sockaddr_dl
)) {
237 log(LOG_DEBUG
, "arp_rtrequest: bad gateway value\n");
240 SDL(gate
)->sdl_type
= rt
->rt_ifp
->if_type
;
241 SDL(gate
)->sdl_index
= rt
->rt_ifp
->if_index
;
243 break; /* This happens on a route change */
245 * Case 2: This route may come from cloning, or a manual route
246 * add with a LL address.
248 R_Malloc(la
, struct llinfo_arp
*, sizeof *la
);
251 log(LOG_DEBUG
, "arp_rtrequest: malloc failed\n");
254 bzero(la
, sizeof *la
);
256 rt
->rt_flags
|= RTF_LLINFO
;
257 LIST_INSERT_HEAD(&llinfo_arp_list
[mycpuid
], la
, la_le
);
261 * This keeps the multicast addresses from showing up
262 * in `arp -a' listings as unresolved. It's not actually
263 * functional. Then the same for broadcast.
265 if (IN_MULTICAST(ntohl(SIN(rt_key(rt
))->sin_addr
.s_addr
))) {
266 ETHER_MAP_IP_MULTICAST(&SIN(rt_key(rt
))->sin_addr
,
268 SDL(gate
)->sdl_alen
= 6;
271 if (in_broadcast(SIN(rt_key(rt
))->sin_addr
, rt
->rt_ifp
)) {
272 memcpy(LLADDR(SDL(gate
)), rt
->rt_ifp
->if_broadcastaddr
,
273 rt
->rt_ifp
->if_addrlen
);
274 SDL(gate
)->sdl_alen
= rt
->rt_ifp
->if_addrlen
;
279 if (SIN(rt_key(rt
))->sin_addr
.s_addr
==
280 (IA_SIN(rt
->rt_ifa
))->sin_addr
.s_addr
) {
282 * This test used to be
283 * if (loif.if_flags & IFF_UP)
284 * It allowed local traffic to be forced
285 * through the hardware by configuring the
286 * loopback down. However, it causes problems
287 * during network configuration for boards
288 * that can't receive packets they send. It
289 * is now necessary to clear "useloopback" and
290 * remove the route to force traffic out to
294 bcopy(IF_LLADDR(rt
->rt_ifp
), LLADDR(SDL(gate
)),
295 SDL(gate
)->sdl_alen
= rt
->rt_ifp
->if_addrlen
);
304 LIST_REMOVE(la
, la_le
);
305 rt
->rt_llinfo
= NULL
;
306 rt
->rt_flags
&= ~RTF_LLINFO
;
307 if (la
->la_hold
!= NULL
)
308 m_freem(la
->la_hold
);
314 * Broadcast an ARP request. Caller specifies:
315 * - arp header source ip address
316 * - arp header target ip address
317 * - arp header source ethernet address
320 arprequest(struct ifnet
*ifp
, struct in_addr
*sip
, struct in_addr
*tip
,
324 struct ether_header
*eh
;
327 static u_char llcx
[] = { 0x82, 0x40, LLC_SNAP_LSAP
, LLC_SNAP_LSAP
,
328 LLC_UI
, 0x00, 0x00, 0x00, 0x08, 0x06 };
331 if ((m
= m_gethdr(MB_DONTWAIT
, MT_DATA
)) == NULL
)
333 m
->m_pkthdr
.rcvif
= (struct ifnet
*)NULL
;
335 switch (ifp
->if_type
) {
337 ar_hrd
= htons(ARPHRD_IEEE802
);
339 m
->m_len
= (sizeof llcx
) +
340 arphdr_len2(ifp
->if_addrlen
, sizeof(struct in_addr
));
341 m
->m_pkthdr
.len
= m
->m_len
;
342 MH_ALIGN(m
, m
->m_len
);
344 memcpy(mtod(m
, caddr_t
), llcx
, sizeof llcx
);
345 memcpy(sa
.sa_data
, ifp
->if_broadcastaddr
, ifp
->if_addrlen
);
346 memcpy(sa
.sa_data
+ 6, enaddr
, 6);
347 sa
.sa_data
[6] |= TR_RII
;
348 sa
.sa_data
[12] = TR_AC
;
349 sa
.sa_data
[13] = TR_LLC_FRAME
;
351 ah
= (struct arphdr
*)(mtod(m
, char *) + sizeof llcx
);
356 * This may not be correct for types not explicitly
357 * listed, but this is our best guess
360 ar_hrd
= htons(ARPHRD_ETHER
);
362 m
->m_len
= arphdr_len2(ifp
->if_addrlen
, sizeof(struct in_addr
));
363 m
->m_pkthdr
.len
= m
->m_len
;
364 MH_ALIGN(m
, m
->m_len
);
366 eh
= (struct ether_header
*)sa
.sa_data
;
367 /* if_output() will not swap */
368 eh
->ether_type
= htons(ETHERTYPE_ARP
);
369 memcpy(eh
->ether_dhost
, ifp
->if_broadcastaddr
, ifp
->if_addrlen
);
371 ah
= mtod(m
, struct arphdr
*);
376 ah
->ar_pro
= htons(ETHERTYPE_IP
);
377 ah
->ar_hln
= ifp
->if_addrlen
; /* hardware address length */
378 ah
->ar_pln
= sizeof(struct in_addr
); /* protocol address length */
379 ah
->ar_op
= htons(ARPOP_REQUEST
);
380 memcpy(ar_sha(ah
), enaddr
, ah
->ar_hln
);
381 memset(ar_tha(ah
), 0, ah
->ar_hln
);
382 memcpy(ar_spa(ah
), sip
, ah
->ar_pln
);
383 memcpy(ar_tpa(ah
), tip
, ah
->ar_pln
);
385 sa
.sa_family
= AF_UNSPEC
;
386 sa
.sa_len
= sizeof sa
;
387 (*ifp
->if_output
)(ifp
, m
, &sa
, (struct rtentry
*)NULL
);
391 * Resolve an IP address into an ethernet address. If success,
392 * desten is filled in. If there is no entry in arptab,
393 * set one up and broadcast a request for the IP address.
394 * Hold onto this mbuf and resend it once the address
395 * is finally resolved. A return value of 1 indicates
396 * that desten has been filled in and the packet should be sent
397 * normally; a 0 return indicates that the packet has been
398 * taken over here, either now or for later transmission.
405 struct sockaddr
*dst
,
409 struct llinfo_arp
*la
= NULL
;
410 struct sockaddr_dl
*sdl
;
412 if (m
->m_flags
& M_BCAST
) { /* broadcast */
413 memcpy(desten
, ifp
->if_broadcastaddr
, ifp
->if_addrlen
);
416 if (m
->m_flags
& M_MCAST
) {/* multicast */
417 ETHER_MAP_IP_MULTICAST(&SIN(dst
)->sin_addr
, desten
);
421 if (rt_llroute(dst
, rt0
, &rt
) != 0) {
428 la
= arplookup(SIN(dst
)->sin_addr
.s_addr
, TRUE
, FALSE
);
432 if (la
== NULL
|| rt
== NULL
) {
433 log(LOG_DEBUG
, "arpresolve: can't allocate llinfo for %s%s%s\n",
434 inet_ntoa(SIN(dst
)->sin_addr
), la
? "la" : " ",
439 sdl
= SDL(rt
->rt_gateway
);
441 * Check the address family and length is valid, the address
442 * is resolved; otherwise, try to resolve.
444 if ((rt
->rt_expire
== 0 || rt
->rt_expire
> time_second
) &&
445 sdl
->sdl_family
== AF_LINK
&& sdl
->sdl_alen
!= 0) {
447 * If entry has an expiry time and it is approaching,
448 * see if we need to send an ARP request within this
449 * arpt_down interval.
451 if ((rt
->rt_expire
!= 0) &&
452 (time_second
+ la
->la_preempt
> rt
->rt_expire
)) {
454 &SIN(rt
->rt_ifa
->ifa_addr
)->sin_addr
,
460 bcopy(LLADDR(sdl
), desten
, sdl
->sdl_alen
);
464 * If ARP is disabled on this interface, stop.
466 * Probably should not allocate empty llinfo struct if we are
467 * not going to be sending out an arp request.
469 if (ifp
->if_flags
& IFF_NOARP
) {
474 * There is an arptab entry, but no ethernet address
475 * response yet. Replace the held mbuf with this
478 if (la
->la_hold
!= NULL
)
479 m_freem(la
->la_hold
);
481 if (rt
->rt_expire
|| ((rt
->rt_flags
& RTF_STATIC
) && !sdl
->sdl_alen
)) {
482 rt
->rt_flags
&= ~RTF_REJECT
;
483 if (la
->la_asked
== 0 || rt
->rt_expire
!= time_second
) {
484 rt
->rt_expire
= time_second
;
485 if (la
->la_asked
++ < arp_maxtries
) {
487 &SIN(rt
->rt_ifa
->ifa_addr
)->sin_addr
,
491 rt
->rt_flags
|= RTF_REJECT
;
492 rt
->rt_expire
+= arpt_down
;
494 la
->la_preempt
= arp_maxtries
;
503 * Common length and type checks are done here,
504 * then the protocol-specific routine is called.
507 arpintr(struct netmsg
*msg
)
509 struct mbuf
*m
= ((struct netmsg_packet
*)msg
)->nm_packet
;
513 if (m
->m_len
< sizeof(struct arphdr
) &&
514 ((m
= m_pullup(m
, sizeof(struct arphdr
))) == NULL
)) {
515 log(LOG_ERR
, "arp: runt packet -- m_pullup failed\n");
518 ar
= mtod(m
, struct arphdr
*);
520 ar_hrd
= ntohs(ar
->ar_hrd
);
521 if (ar_hrd
!= ARPHRD_ETHER
&&
522 ar_hrd
!= ARPHRD_IEEE802
) {
524 "arp: unknown hardware address format (0x%2D)\n",
525 (unsigned char *)&ar
->ar_hrd
, "");
529 if (m
->m_pkthdr
.len
< arphdr_len(ar
) &&
530 (m
= m_pullup(m
, arphdr_len(ar
))) == NULL
) {
531 log(LOG_ERR
, "arp: runt packet\n");
535 switch (ntohs(ar
->ar_pro
)) {
546 /* msg was embedded in the mbuf, do not reply! */
551 * ARP for Internet protocols on 10 Mb/s Ethernet.
552 * Algorithm is that given in RFC 826.
553 * In addition, a sanity check is performed on the sender
554 * protocol address, to catch impersonators.
555 * We no longer handle negotiations for use of trailer protocol:
556 * Formerly, ARP replied for protocol type ETHERTYPE_TRAIL sent
557 * along with IP replies if we wanted trailers sent to us,
558 * and also sent them in response to IP replies.
559 * This allowed either end to announce the desire to receive
561 * We no longer reply to requests for ETHERTYPE_TRAIL protocol either,
562 * but formerly didn't normally send requests.
564 static int log_arp_wrong_iface
= 1;
565 SYSCTL_INT(_net_link_ether_inet
, OID_AUTO
, log_arp_wrong_iface
, CTLFLAG_RW
,
566 &log_arp_wrong_iface
, 0,
567 "log arp packets arriving on the wrong interface");
570 arp_update_oncpu(struct mbuf
*m
, in_addr_t saddr
, boolean_t create
,
573 struct arphdr
*ah
= mtod(m
, struct arphdr
*);
574 struct ifnet
*ifp
= m
->m_pkthdr
.rcvif
;
575 struct llinfo_arp
*la
;
576 struct sockaddr_dl
*sdl
;
580 la
= arplookup(saddr
, create
, FALSE
);
581 if (la
&& (rt
= la
->la_rt
) && (sdl
= SDL(rt
->rt_gateway
))) {
582 struct in_addr isaddr
= { saddr
};
584 /* the following is not an error when doing bridging */
585 if (rt
->rt_ifp
!= ifp
) {
586 if (dologging
&& log_arp_wrong_iface
&& cpu
== 0) {
589 "but got reply from %*D on %s\n",
591 rt
->rt_ifp
->if_xname
,
592 ifp
->if_addrlen
, (u_char
*)ar_sha(ah
), ":",
598 bcmp(ar_sha(ah
), LLADDR(sdl
), sdl
->sdl_alen
)) {
599 if (rt
->rt_expire
!= 0) {
600 if (dologging
&& cpu
== 0) {
602 "arp: %s moved from %*D to %*D on %s\n",
604 ifp
->if_addrlen
, (u_char
*)LLADDR(sdl
),
605 ":", ifp
->if_addrlen
,
606 (u_char
*)ar_sha(ah
), ":",
610 if (dologging
&& cpu
== 0) {
612 "arp: %*D attempts to modify permanent entry for %s on %s\n",
613 ifp
->if_addrlen
, (u_char
*)ar_sha(ah
), ":",
614 inet_ntoa(isaddr
), ifp
->if_xname
);
620 * sanity check for the address length.
621 * XXX this does not work for protocols with variable address
624 if (dologging
&& sdl
->sdl_alen
&& sdl
->sdl_alen
!= ah
->ar_hln
&&
628 "arp from %*D: new addr len %d, was %d",
629 ifp
->if_addrlen
, (u_char
*) ar_sha(ah
), ":",
630 ah
->ar_hln
, sdl
->sdl_alen
);
632 if (ifp
->if_addrlen
!= ah
->ar_hln
) {
633 if (dologging
&& cpu
== 0) {
635 "arp from %*D: addr len: new %d, i/f %d (ignored)",
636 ifp
->if_addrlen
, (u_char
*) ar_sha(ah
), ":",
637 ah
->ar_hln
, ifp
->if_addrlen
);
641 memcpy(LLADDR(sdl
), ar_sha(ah
), sdl
->sdl_alen
= ah
->ar_hln
);
643 * If we receive an arp from a token-ring station over
644 * a token-ring nic then try to save the source
647 if (ifp
->if_type
== IFT_ISO88025
) {
648 struct iso88025_header
*th
=
649 (struct iso88025_header
*)m
->m_pkthdr
.header
;
650 struct iso88025_sockaddr_dl_data
*trld
=
654 rif_len
= TR_RCF_RIFLEN(th
->rcf
);
655 if ((th
->iso88025_shost
[0] & TR_RII
) &&
657 trld
->trld_rcf
= th
->rcf
;
658 trld
->trld_rcf
^= htons(TR_RCF_DIR
);
659 memcpy(trld
->trld_route
, th
->rd
, rif_len
- 2);
660 trld
->trld_rcf
&= ~htons(TR_RCF_BCST_MASK
);
662 * Set up source routing information for
665 m
->m_data
-= rif_len
;
667 m
->m_pkthdr
.len
+= rif_len
;
669 th
->iso88025_shost
[0] &= ~TR_RII
;
674 m
->m_pkthdr
.len
+= 8;
675 th
->rcf
= trld
->trld_rcf
;
677 if (rt
->rt_expire
!= 0)
678 rt
->rt_expire
= time_second
+ arpt_keep
;
679 rt
->rt_flags
&= ~RTF_REJECT
;
681 la
->la_preempt
= arp_maxtries
;
684 * This particular cpu might have been holding an mbuf
685 * pending ARP resolution. If so, transmit the mbuf now.
687 if (la
->la_hold
!= NULL
) {
688 m_adj(la
->la_hold
, sizeof(struct ether_header
));
689 lwkt_serialize_enter(ifp
->if_serializer
);
690 (*ifp
->if_output
)(ifp
, la
->la_hold
, rt_key(rt
), rt
);
691 lwkt_serialize_exit(ifp
->if_serializer
);
699 struct netmsg_arp_update
{
700 struct netmsg netmsg
;
706 static void arp_update_msghandler(struct netmsg
*netmsg
);
711 * Called from arpintr() - this routine is run from a single cpu.
714 in_arpinput(struct mbuf
*m
)
717 struct ifnet
*ifp
= m
->m_pkthdr
.rcvif
;
718 struct ether_header
*eh
;
719 struct iso88025_header
*th
= (struct iso88025_header
*)NULL
;
722 struct in_ifaddr
*ia
;
724 struct in_addr isaddr
, itaddr
, myaddr
;
726 struct netmsg_arp_update msg
;
728 u_int8_t
*enaddr
= NULL
;
732 req_len
= arphdr_len2(ifp
->if_addrlen
, sizeof(struct in_addr
));
733 if (m
->m_len
< req_len
&& (m
= m_pullup(m
, req_len
)) == NULL
) {
734 log(LOG_ERR
, "in_arp: runt packet -- m_pullup failed\n");
738 ah
= mtod(m
, struct arphdr
*);
739 op
= ntohs(ah
->ar_op
);
740 memcpy(&isaddr
, ar_spa(ah
), sizeof isaddr
);
741 memcpy(&itaddr
, ar_tpa(ah
), sizeof itaddr
);
743 * Check both target and sender IP addresses:
745 * If we receive the packet on the interface owning the address,
746 * then accept the address.
748 * For a bridge, we accept the address if the receive interface and
749 * the interface owning the address are on the same bridge.
750 * (This will change slightly when we have clusters of interfaces).
752 LIST_FOREACH(ia
, INADDR_HASH(itaddr
.s_addr
), ia_hash
) {
753 /* Skip all ia's which don't match */
754 if (itaddr
.s_addr
!= ia
->ia_addr
.sin_addr
.s_addr
)
757 if (ia
->ia_ifp
== ifp
)
760 if (ifp
->if_bridge
&& ia
->ia_ifp
&&
761 ifp
->if_bridge
== ia
->ia_ifp
->if_bridge
)
766 * If the interface does not match, but the recieving interface
767 * is part of carp, we call carp_iamatch to see if this is a
768 * request for the virtual host ip.
769 * XXX: This is really ugly!
771 if (ifp
->if_carp
!= NULL
&&
772 carp_iamatch(ifp
->if_carp
, ia
, &isaddr
, &enaddr
) &&
773 itaddr
.s_addr
== ia
->ia_addr
.sin_addr
.s_addr
)
777 LIST_FOREACH(ia
, INADDR_HASH(isaddr
.s_addr
), ia_hash
) {
778 /* Skip all ia's which don't match */
779 if (isaddr
.s_addr
!= ia
->ia_addr
.sin_addr
.s_addr
)
782 if (ia
->ia_ifp
== ifp
)
785 if (ifp
->if_bridge
&& ia
->ia_ifp
&&
786 ifp
->if_bridge
== ia
->ia_ifp
->if_bridge
)
790 * No match, use the first inet address on the receive interface
791 * as a dummy address for the rest of the function.
793 TAILQ_FOREACH(ifa
, &ifp
->if_addrhead
, ifa_link
) {
794 if (ifa
->ifa_addr
&& ifa
->ifa_addr
->sa_family
== AF_INET
) {
800 * If we got here, we didn't find any suitable interface,
801 * so drop the packet.
808 enaddr
= (u_int8_t
*)IF_LLADDR(ifp
);
809 myaddr
= ia
->ia_addr
.sin_addr
;
810 if (!bcmp(ar_sha(ah
), enaddr
, ifp
->if_addrlen
)) {
811 m_freem(m
); /* it's from me, ignore it. */
814 if (!bcmp(ar_sha(ah
), ifp
->if_broadcastaddr
, ifp
->if_addrlen
)) {
816 "arp: link address is broadcast for IP address %s!\n",
821 if (isaddr
.s_addr
== myaddr
.s_addr
&& myaddr
.s_addr
!= 0) {
823 "arp: %*D is using my IP address %s!\n",
824 ifp
->if_addrlen
, (u_char
*)ar_sha(ah
), ":",
830 netmsg_init(&msg
.netmsg
, &curthread
->td_msgport
, 0,
831 arp_update_msghandler
);
833 msg
.saddr
= isaddr
.s_addr
;
834 msg
.create
= (itaddr
.s_addr
== myaddr
.s_addr
);
835 lwkt_domsg(rtable_portfn(0), &msg
.netmsg
.nm_lmsg
, 0);
837 arp_update_oncpu(m
, isaddr
.s_addr
, (itaddr
.s_addr
== myaddr
.s_addr
),
840 if (op
!= ARPOP_REQUEST
) {
844 if (itaddr
.s_addr
== myaddr
.s_addr
) {
845 /* I am the target */
846 memcpy(ar_tha(ah
), ar_sha(ah
), ah
->ar_hln
);
847 memcpy(ar_sha(ah
), enaddr
, ah
->ar_hln
);
849 struct llinfo_arp
*la
;
851 la
= arplookup(itaddr
.s_addr
, FALSE
, SIN_PROXY
);
853 struct sockaddr_in sin
;
860 bzero(&sin
, sizeof sin
);
861 sin
.sin_family
= AF_INET
;
862 sin
.sin_len
= sizeof sin
;
863 sin
.sin_addr
= itaddr
;
865 rt
= rtpurelookup((struct sockaddr
*)&sin
);
872 * Don't send proxies for nodes on the same interface
873 * as this one came out of, or we'll get into a fight
874 * over who claims what Ether address.
876 if (rt
->rt_ifp
== ifp
) {
880 memcpy(ar_tha(ah
), ar_sha(ah
), ah
->ar_hln
);
881 memcpy(ar_sha(ah
), enaddr
, ah
->ar_hln
);
883 kprintf("arp: proxying for %s\n", inet_ntoa(itaddr
));
886 struct sockaddr_dl
*sdl
;
889 memcpy(ar_tha(ah
), ar_sha(ah
), ah
->ar_hln
);
890 sdl
= SDL(rt
->rt_gateway
);
891 memcpy(ar_sha(ah
), LLADDR(sdl
), ah
->ar_hln
);
895 memcpy(ar_tpa(ah
), ar_spa(ah
), ah
->ar_pln
);
896 memcpy(ar_spa(ah
), &itaddr
, ah
->ar_pln
);
897 ah
->ar_op
= htons(ARPOP_REPLY
);
898 ah
->ar_pro
= htons(ETHERTYPE_IP
); /* let's be sure! */
899 switch (ifp
->if_type
) {
901 /* Re-arrange the source/dest address */
902 memcpy(th
->iso88025_dhost
, th
->iso88025_shost
,
903 sizeof th
->iso88025_dhost
);
904 memcpy(th
->iso88025_shost
, IF_LLADDR(ifp
),
905 sizeof th
->iso88025_shost
);
906 /* Set the source routing bit if neccesary */
907 if (th
->iso88025_dhost
[0] & TR_RII
) {
908 th
->iso88025_dhost
[0] &= ~TR_RII
;
909 if (TR_RCF_RIFLEN(th
->rcf
) > 2)
910 th
->iso88025_shost
[0] |= TR_RII
;
912 /* Copy the addresses, ac and fc into sa_data */
913 memcpy(sa
.sa_data
, th
->iso88025_dhost
,
914 (sizeof th
->iso88025_dhost
) * 2);
915 sa
.sa_data
[(sizeof th
->iso88025_dhost
) * 2] = TR_AC
;
916 sa
.sa_data
[(sizeof th
->iso88025_dhost
) * 2 + 1] = TR_LLC_FRAME
;
921 * May not be correct for types not explictly
922 * listed, but it is our best guess.
925 eh
= (struct ether_header
*)sa
.sa_data
;
926 memcpy(eh
->ether_dhost
, ar_tha(ah
), sizeof eh
->ether_dhost
);
927 eh
->ether_type
= htons(ETHERTYPE_ARP
);
930 sa
.sa_family
= AF_UNSPEC
;
931 sa
.sa_len
= sizeof sa
;
932 lwkt_serialize_enter(ifp
->if_serializer
);
933 (*ifp
->if_output
)(ifp
, m
, &sa
, (struct rtentry
*)0);
934 lwkt_serialize_exit(ifp
->if_serializer
);
942 arp_update_msghandler(struct netmsg
*netmsg
)
944 struct netmsg_arp_update
*msg
= (struct netmsg_arp_update
*)netmsg
;
947 arp_update_oncpu(msg
->m
, msg
->saddr
, msg
->create
, FALSE
);
949 nextcpu
= mycpuid
+ 1;
950 if (nextcpu
< ncpus
) {
951 lwkt_forwardmsg(rtable_portfn(nextcpu
), &msg
->netmsg
.nm_lmsg
);
953 lwkt_replymsg(&msg
->netmsg
.nm_lmsg
, 0);
962 * Free an arp entry. If the arp entry is actively referenced or represents
963 * a static entry we only clear it back to an unresolved state, otherwise
964 * we destroy the entry entirely.
966 * Note that static entries are created when route add ... -interface is used
967 * to create an interface route to a (direct) destination.
970 arptfree(struct llinfo_arp
*la
)
972 struct rtentry
*rt
= la
->la_rt
;
973 struct sockaddr_dl
*sdl
;
977 sdl
= SDL(rt
->rt_gateway
);
979 ((rt
->rt_refcnt
> 0 && sdl
->sdl_family
== AF_LINK
) ||
980 (rt
->rt_flags
& RTF_STATIC
))) {
982 la
->la_preempt
= la
->la_asked
= 0;
983 rt
->rt_flags
&= ~RTF_REJECT
;
986 rtrequest(RTM_DELETE
, rt_key(rt
), NULL
, rt_mask(rt
), 0, NULL
);
990 * Lookup or enter a new address in arptab.
992 static struct llinfo_arp
*
993 arplookup(in_addr_t addr
, boolean_t create
, boolean_t proxy
)
996 struct sockaddr_inarp sin
= { sizeof sin
, AF_INET
};
997 const char *why
= NULL
;
999 sin
.sin_addr
.s_addr
= addr
;
1000 sin
.sin_other
= proxy
? SIN_PROXY
: 0;
1002 rt
= rtlookup((struct sockaddr
*)&sin
);
1004 rt
= rtpurelookup((struct sockaddr
*)&sin
);
1009 if (rt
->rt_flags
& RTF_GATEWAY
)
1010 why
= "host is not on local network";
1011 else if (!(rt
->rt_flags
& RTF_LLINFO
))
1012 why
= "could not allocate llinfo";
1013 else if (rt
->rt_gateway
->sa_family
!= AF_LINK
)
1014 why
= "gateway route is not ours";
1018 log(LOG_DEBUG
, "arplookup %s failed: %s\n",
1019 inet_ntoa(sin
.sin_addr
), why
);
1021 if (rt
->rt_refcnt
<= 0 && (rt
->rt_flags
& RTF_WASCLONED
)) {
1022 /* No references to this route. Purge it. */
1023 rtrequest(RTM_DELETE
, rt_key(rt
), rt
->rt_gateway
,
1024 rt_mask(rt
), rt
->rt_flags
, NULL
);
1028 return (rt
->rt_llinfo
);
1032 arp_ifinit(struct ifnet
*ifp
, struct ifaddr
*ifa
)
1034 if (IA_SIN(ifa
)->sin_addr
.s_addr
!= INADDR_ANY
)
1035 arprequest(ifp
, &IA_SIN(ifa
)->sin_addr
, &IA_SIN(ifa
)->sin_addr
,
1037 ifa
->ifa_rtrequest
= arp_rtrequest
;
1038 ifa
->ifa_flags
|= RTF_CLONING
;
1042 arp_ifinit2(struct ifnet
*ifp
, struct ifaddr
*ifa
, u_char
*enaddr
)
1044 if (IA_SIN(ifa
)->sin_addr
.s_addr
!= INADDR_ANY
)
1045 arprequest(ifp
, &IA_SIN(ifa
)->sin_addr
, &IA_SIN(ifa
)->sin_addr
,
1047 ifa
->ifa_rtrequest
= arp_rtrequest
;
1048 ifa
->ifa_flags
|= RTF_CLONING
;
1056 for (cpu
= 0; cpu
< ncpus2
; cpu
++)
1057 LIST_INIT(&llinfo_arp_list
[cpu
]);
1058 netisr_register(NETISR_ARP
, cpu0_portfn
, arpintr
);
1061 SYSINIT(arp
, SI_SUB_PROTO_DOMAIN
, SI_ORDER_ANY
, arp_init
, 0);