1/*	$NetBSD: if_ethersubr.c,v 1.188.8.3 2012/10/31 16:07:46 riz 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) 1982, 1989, 1993
34 *	The Regents of the University of California.  All rights reserved.
35 *
36 * Redistribution and use in source and binary forms, with or without
37 * modification, are permitted provided that the following conditions
38 * are met:
39 * 1. Redistributions of source code must retain the above copyright
40 *    notice, this list of conditions and the following disclaimer.
41 * 2. Redistributions in binary form must reproduce the above copyright
42 *    notice, this list of conditions and the following disclaimer in the
43 *    documentation and/or other materials provided with the distribution.
44 * 3. Neither the name of the University nor the names of its contributors
45 *    may be used to endorse or promote products derived from this software
46 *    without specific prior written permission.
47 *
48 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
49 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
50 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
51 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
52 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
53 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
54 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
55 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
56 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
57 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
58 * SUCH DAMAGE.
59 *
60 *	@(#)if_ethersubr.c	8.2 (Berkeley) 4/4/96
61 */
62
63#include <sys/cdefs.h>
64__KERNEL_RCSID(0, "$NetBSD: if_ethersubr.c,v 1.188.8.3 2012/10/31 16:07:46 riz Exp $");
65
66#include "opt_inet.h"
67#include "opt_atalk.h"
68#include "opt_iso.h"
69#include "opt_ipx.h"
70#include "opt_mbuftrace.h"
71#include "opt_mpls.h"
72#include "opt_gateway.h"
73#include "opt_pfil_hooks.h"
74#include "opt_pppoe.h"
75#include "vlan.h"
76#include "pppoe.h"
77#include "bridge.h"
78#include "arp.h"
79#include "agr.h"
80
81#include <sys/param.h>
82#include <sys/systm.h>
83#include <sys/kernel.h>
84#include <sys/callout.h>
85#include <sys/malloc.h>
86#include <sys/mbuf.h>
87#include <sys/protosw.h>
88#include <sys/socket.h>
89#include <sys/ioctl.h>
90#include <sys/errno.h>
91#include <sys/syslog.h>
92#include <sys/kauth.h>
93#include <sys/cpu.h>
94#include <sys/intr.h>
95#include <sys/device.h>
96
97#include <net/if.h>
98#include <net/netisr.h>
99#include <net/route.h>
100#include <net/if_llc.h>
101#include <net/if_dl.h>
102#include <net/if_types.h>
103
104#include <net/if_media.h>
105#include <dev/mii/mii.h>
106#include <dev/mii/miivar.h>
107
108#if NARP == 0
109/*
110 * XXX there should really be a way to issue this warning from within config(8)
111 */
112#error You have included NETATALK or a pseudo-device in your configuration that depends on the presence of ethernet interfaces, but have no such interfaces configured. Check if you really need pseudo-device bridge, pppoe, vlan or options NETATALK.
113#endif
114
115#include <net/bpf.h>
116
117#include <net/if_ether.h>
118#include <net/if_vlanvar.h>
119
120#if NPPPOE > 0
121#include <net/if_pppoe.h>
122#endif
123
124#if NAGR > 0
125#include <net/agr/ieee8023_slowprotocols.h>	/* XXX */
126#include <net/agr/ieee8023ad.h>
127#include <net/agr/if_agrvar.h>
128#endif
129
130#if NBRIDGE > 0
131#include <net/if_bridgevar.h>
132#endif
133
134#include <netinet/in.h>
135#ifdef INET
136#include <netinet/in_var.h>
137#endif
138#include <netinet/if_inarp.h>
139
140#ifdef INET6
141#ifndef INET
142#include <netinet/in.h>
143#endif
144#include <netinet6/in6_var.h>
145#include <netinet6/nd6.h>
146#endif
147
148
149#include "carp.h"
150#if NCARP > 0
151#include <netinet/ip_carp.h>
152#endif
153
154#ifdef IPX
155#include <netipx/ipx.h>
156#include <netipx/ipx_if.h>
157#endif
158
159#ifdef ISO
160#include <netiso/argo_debug.h>
161#include <netiso/iso.h>
162#include <netiso/iso_var.h>
163#include <netiso/iso_snpac.h>
164#endif
165
166
167
168#ifdef NETATALK
169#include <netatalk/at.h>
170#include <netatalk/at_var.h>
171#include <netatalk/at_extern.h>
172
173#define llc_snap_org_code llc_un.type_snap.org_code
174#define llc_snap_ether_type llc_un.type_snap.ether_type
175
176extern u_char	at_org_code[3];
177extern u_char	aarp_org_code[3];
178#endif /* NETATALK */
179
180#ifdef MPLS
181#include <netmpls/mpls.h>
182#include <netmpls/mpls_var.h>
183#endif
184
185static struct timeval bigpktppslim_last;
186static int bigpktppslim = 2;	/* XXX */
187static int bigpktpps_count;
188
189
190const uint8_t etherbroadcastaddr[ETHER_ADDR_LEN] =
191    { 0xff, 0xff, 0xff, 0xff, 0xff, 0xff };
192const uint8_t ethermulticastaddr_slowprotocols[ETHER_ADDR_LEN] =
193    { 0x01, 0x80, 0xc2, 0x00, 0x00, 0x02 };
194#define senderr(e) { error = (e); goto bad;}
195
196static	int ether_output(struct ifnet *, struct mbuf *,
197	    const struct sockaddr *, struct rtentry *);
198
199/*
200 * Ethernet output routine.
201 * Encapsulate a packet of type family for the local net.
202 * Assumes that ifp is actually pointer to ethercom structure.
203 */
204static int
205ether_output(struct ifnet * const ifp0, struct mbuf * const m0,
206	const struct sockaddr * const dst,
207	struct rtentry *rt0)
208{
209	uint16_t etype = 0;
210	int error = 0, hdrcmplt = 0;
211 	uint8_t esrc[6], edst[6];
212	struct mbuf *m = m0;
213	struct rtentry *rt;
214	struct mbuf *mcopy = NULL;
215	struct ether_header *eh;
216	struct ifnet *ifp = ifp0;
217	ALTQ_DECL(struct altq_pktattr pktattr;)
218#ifdef INET
219	struct arphdr *ah;
220#endif /* INET */
221#ifdef NETATALK
222	struct at_ifaddr *aa;
223#endif /* NETATALK */
224
225	KASSERT(KERNEL_LOCKED_P());
226
227#ifdef MBUFTRACE
228	m_claimm(m, ifp->if_mowner);
229#endif
230
231#if NCARP > 0
232	if (ifp->if_type == IFT_CARP) {
233		struct ifaddr *ifa;
234
235		/* loop back if this is going to the carp interface */
236		if (dst != NULL && ifp0->if_link_state == LINK_STATE_UP &&
237		    (ifa = ifa_ifwithaddr(dst)) != NULL &&
238		    ifa->ifa_ifp == ifp0)
239			return looutput(ifp0, m, dst, rt0);
240
241		ifp = ifp->if_carpdev;
242		/* ac = (struct arpcom *)ifp; */
243
244		if ((ifp0->if_flags & (IFF_UP|IFF_RUNNING)) !=
245		    (IFF_UP|IFF_RUNNING))
246			senderr(ENETDOWN);
247	}
248#endif /* NCARP > 0 */
249
250	if ((ifp->if_flags & (IFF_UP|IFF_RUNNING)) != (IFF_UP|IFF_RUNNING))
251		senderr(ENETDOWN);
252	if ((rt = rt0) != NULL) {
253		if ((rt->rt_flags & RTF_UP) == 0) {
254			if ((rt0 = rt = rtalloc1(dst, 1)) != NULL) {
255				rt->rt_refcnt--;
256				if (rt->rt_ifp != ifp)
257					return (*rt->rt_ifp->if_output)
258							(ifp, m0, dst, rt);
259			} else
260				senderr(EHOSTUNREACH);
261		}
262		if ((rt->rt_flags & RTF_GATEWAY) && dst->sa_family != AF_NS) {
263			if (rt->rt_gwroute == NULL)
264				goto lookup;
265			if (((rt = rt->rt_gwroute)->rt_flags & RTF_UP) == 0) {
266				rtfree(rt); rt = rt0;
267			lookup: rt->rt_gwroute = rtalloc1(rt->rt_gateway, 1);
268				if ((rt = rt->rt_gwroute) == NULL)
269					senderr(EHOSTUNREACH);
270				/* the "G" test below also prevents rt == rt0 */
271				if ((rt->rt_flags & RTF_GATEWAY) ||
272				    (rt->rt_ifp != ifp)) {
273					rt->rt_refcnt--;
274					rt0->rt_gwroute = NULL;
275					senderr(EHOSTUNREACH);
276				}
277			}
278		}
279		if (rt->rt_flags & RTF_REJECT)
280			if (rt->rt_rmx.rmx_expire == 0 ||
281			    (u_long) time_second < rt->rt_rmx.rmx_expire)
282				senderr(rt == rt0 ? EHOSTDOWN : EHOSTUNREACH);
283	}
284
285	switch (dst->sa_family) {
286
287#ifdef INET
288	case AF_INET:
289		if (m->m_flags & M_BCAST)
290			(void)memcpy(edst, etherbroadcastaddr, sizeof(edst));
291		else if (m->m_flags & M_MCAST)
292			ETHER_MAP_IP_MULTICAST(&satocsin(dst)->sin_addr, edst);
293		else if (!arpresolve(ifp, rt, m, dst, edst))
294			return (0);	/* if not yet resolved */
295		/* If broadcasting on a simplex interface, loopback a copy */
296		if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX))
297			mcopy = m_copy(m, 0, (int)M_COPYALL);
298		etype = htons(ETHERTYPE_IP);
299		break;
300
301	case AF_ARP:
302		ah = mtod(m, struct arphdr *);
303		if (m->m_flags & M_BCAST)
304			(void)memcpy(edst, etherbroadcastaddr, sizeof(edst));
305		else {
306			void *tha = ar_tha(ah);
307
308			if (tha == NULL) {
309				/* fake with ARPHDR_IEEE1394 */
310				return 0;
311			}
312			memcpy(edst, tha, sizeof(edst));
313		}
314
315		ah->ar_hrd = htons(ARPHRD_ETHER);
316
317		switch (ntohs(ah->ar_op)) {
318		case ARPOP_REVREQUEST:
319		case ARPOP_REVREPLY:
320			etype = htons(ETHERTYPE_REVARP);
321			break;
322
323		case ARPOP_REQUEST:
324		case ARPOP_REPLY:
325		default:
326			etype = htons(ETHERTYPE_ARP);
327		}
328
329		break;
330#endif
331#ifdef INET6
332	case AF_INET6:
333		if (!nd6_storelladdr(ifp, rt, m, dst, edst, sizeof(edst))){
334			/* something bad happened */
335			return (0);
336		}
337		etype = htons(ETHERTYPE_IPV6);
338		break;
339#endif
340#ifdef NETATALK
341    case AF_APPLETALK:
342		if (!aarpresolve(ifp, m, (const struct sockaddr_at *)dst, edst)) {
343#ifdef NETATALKDEBUG
344			printf("aarpresolv failed\n");
345#endif /* NETATALKDEBUG */
346			return (0);
347		}
348		/*
349		 * ifaddr is the first thing in at_ifaddr
350		 */
351		aa = (struct at_ifaddr *) at_ifawithnet(
352		    (const struct sockaddr_at *)dst, ifp);
353		if (aa == NULL)
354		    goto bad;
355
356		/*
357		 * In the phase 2 case, we need to prepend an mbuf for the
358		 * llc header.  Since we must preserve the value of m,
359		 * which is passed to us by value, we m_copy() the first
360		 * mbuf, and use it for our llc header.
361		 */
362		if (aa->aa_flags & AFA_PHASE2) {
363			struct llc llc;
364
365			M_PREPEND(m, sizeof(struct llc), M_DONTWAIT);
366			llc.llc_dsap = llc.llc_ssap = LLC_SNAP_LSAP;
367			llc.llc_control = LLC_UI;
368			memcpy(llc.llc_snap_org_code, at_org_code,
369			    sizeof(llc.llc_snap_org_code));
370			llc.llc_snap_ether_type = htons(ETHERTYPE_ATALK);
371			memcpy(mtod(m, void *), &llc, sizeof(struct llc));
372		} else {
373			etype = htons(ETHERTYPE_ATALK);
374		}
375		break;
376#endif /* NETATALK */
377#ifdef IPX
378	case AF_IPX:
379		etype = htons(ETHERTYPE_IPX);
380 		memcpy(edst,
381		    &(((const struct sockaddr_ipx *)dst)->sipx_addr.x_host),
382		    sizeof(edst));
383		/* If broadcasting on a simplex interface, loopback a copy */
384		if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX))
385			mcopy = m_copy(m, 0, (int)M_COPYALL);
386		break;
387#endif
388#ifdef	ISO
389	case AF_ISO: {
390		int	snpalen;
391		struct	llc *l;
392		const struct sockaddr_dl *sdl;
393
394		if (rt && (sdl = satocsdl(rt->rt_gateway)) &&
395		    sdl->sdl_family == AF_LINK && sdl->sdl_alen > 0) {
396			memcpy(edst, CLLADDR(sdl), sizeof(edst));
397		} else {
398			error = iso_snparesolve(ifp,
399			    (const struct sockaddr_iso *)dst,
400						(char *)edst, &snpalen);
401			if (error)
402				goto bad; /* Not Resolved */
403		}
404		/* If broadcasting on a simplex interface, loopback a copy */
405		if (*edst & 1)
406			m->m_flags |= (M_BCAST|M_MCAST);
407		if ((m->m_flags & M_BCAST) && (ifp->if_flags & IFF_SIMPLEX) &&
408		    (mcopy = m_copy(m, 0, (int)M_COPYALL))) {
409			M_PREPEND(mcopy, sizeof (*eh), M_DONTWAIT);
410			if (mcopy) {
411				eh = mtod(mcopy, struct ether_header *);
412				memcpy(eh->ether_dhost, edst, sizeof(edst));
413				memcpy(eh->ether_shost, CLLADDR(ifp->if_sadl),
414				    sizeof(edst));
415			}
416		}
417		M_PREPEND(m, 3, M_DONTWAIT);
418		if (m == NULL)
419			return (0);
420		l = mtod(m, struct llc *);
421		l->llc_dsap = l->llc_ssap = LLC_ISO_LSAP;
422		l->llc_control = LLC_UI;
423#ifdef ARGO_DEBUG
424		if (argo_debug[D_ETHER]) {
425			int i;
426			printf("unoutput: sending pkt to: ");
427			for (i=0; i<6; i++)
428				printf("%x ", edst[i] & 0xff);
429			printf("\n");
430		}
431#endif
432		} break;
433#endif /* ISO */
434
435	case pseudo_AF_HDRCMPLT:
436		hdrcmplt = 1;
437		memcpy(esrc,
438		    ((const struct ether_header *)dst->sa_data)->ether_shost,
439		    sizeof(esrc));
440		/* FALLTHROUGH */
441
442	case AF_UNSPEC:
443 		memcpy(edst,
444		    ((const struct ether_header *)dst->sa_data)->ether_dhost,
445		    sizeof(edst));
446		/* AF_UNSPEC doesn't swap the byte order of the ether_type. */
447		etype = ((const struct ether_header *)dst->sa_data)->ether_type;
448		break;
449
450	default:
451		printf("%s: can't handle af%d\n", ifp->if_xname,
452			dst->sa_family);
453		senderr(EAFNOSUPPORT);
454	}
455
456#ifdef MPLS
457	if (rt0 != NULL && rt_gettag(rt0) != NULL &&
458	    rt_gettag(rt0)->sa_family == AF_MPLS &&
459	    (m->m_flags & (M_MCAST | M_BCAST)) == 0) {
460		union mpls_shim msh;
461		msh.s_addr = MPLS_GETSADDR(rt0);
462		if (msh.shim.label != MPLS_LABEL_IMPLNULL)
463			etype = htons(ETHERTYPE_MPLS);
464	}
465#endif
466
467	if (mcopy)
468		(void)looutput(ifp, mcopy, dst, rt);
469
470	/* If no ether type is set, this must be a 802.2 formatted packet.
471	 */
472	if (etype == 0)
473		etype = htons(m->m_pkthdr.len);
474	/*
475	 * Add local net header.  If no space in first mbuf,
476	 * allocate another.
477	 */
478	M_PREPEND(m, sizeof (struct ether_header), M_DONTWAIT);
479	if (m == 0)
480		senderr(ENOBUFS);
481	eh = mtod(m, struct ether_header *);
482	/* Note: etype is already in network byte order. */
483	(void)memcpy(&eh->ether_type, &etype, sizeof(eh->ether_type));
484 	memcpy(eh->ether_dhost, edst, sizeof(edst));
485	if (hdrcmplt)
486		memcpy(eh->ether_shost, esrc, sizeof(eh->ether_shost));
487	else
488	 	memcpy(eh->ether_shost, CLLADDR(ifp->if_sadl),
489		    sizeof(eh->ether_shost));
490
491#if NCARP > 0
492	if (ifp0 != ifp && ifp0->if_type == IFT_CARP) {
493	 	memcpy(eh->ether_shost, CLLADDR(ifp0->if_sadl),
494		    sizeof(eh->ether_shost));
495	}
496#endif /* NCARP > 0 */
497
498#ifdef PFIL_HOOKS
499	if ((error = pfil_run_hooks(&ifp->if_pfil, &m, ifp, PFIL_OUT)) != 0)
500		return (error);
501	if (m == NULL)
502		return (0);
503#endif
504
505#if NBRIDGE > 0
506	/*
507	 * Bridges require special output handling.
508	 */
509	if (ifp->if_bridge)
510		return (bridge_output(ifp, m, NULL, NULL));
511#endif
512
513#if NCARP > 0
514	if (ifp != ifp0)
515		ifp0->if_obytes += m->m_pkthdr.len + ETHER_HDR_LEN;
516#endif /* NCARP > 0 */
517
518#ifdef ALTQ
519	/*
520	 * If ALTQ is enabled on the parent interface, do
521	 * classification; the queueing discipline might not
522	 * require classification, but might require the
523	 * address family/header pointer in the pktattr.
524	 */
525	if (ALTQ_IS_ENABLED(&ifp->if_snd))
526		altq_etherclassify(&ifp->if_snd, m, &pktattr);
527#endif
528	return ifq_enqueue(ifp, m ALTQ_COMMA ALTQ_DECL(&pktattr));
529
530bad:
531	if (m)
532		m_freem(m);
533	return (error);
534}
535
536#ifdef ALTQ
537/*
538 * This routine is a slight hack to allow a packet to be classified
539 * if the Ethernet headers are present.  It will go away when ALTQ's
540 * classification engine understands link headers.
541 */
542void
543altq_etherclassify(struct ifaltq *ifq, struct mbuf *m,
544    struct altq_pktattr *pktattr)
545{
546	struct ether_header *eh;
547	uint16_t ether_type;
548	int hlen, af, hdrsize;
549	void *hdr;
550
551	hlen = ETHER_HDR_LEN;
552	eh = mtod(m, struct ether_header *);
553
554	ether_type = htons(eh->ether_type);
555
556	if (ether_type < ETHERMTU) {
557		/* LLC/SNAP */
558		struct llc *llc = (struct llc *)(eh + 1);
559		hlen += 8;
560
561		if (m->m_len < hlen ||
562		    llc->llc_dsap != LLC_SNAP_LSAP ||
563		    llc->llc_ssap != LLC_SNAP_LSAP ||
564		    llc->llc_control != LLC_UI) {
565			/* Not SNAP. */
566			goto bad;
567		}
568
569		ether_type = htons(llc->llc_un.type_snap.ether_type);
570	}
571
572	switch (ether_type) {
573	case ETHERTYPE_IP:
574		af = AF_INET;
575		hdrsize = 20;		/* sizeof(struct ip) */
576		break;
577
578	case ETHERTYPE_IPV6:
579		af = AF_INET6;
580		hdrsize = 40;		/* sizeof(struct ip6_hdr) */
581		break;
582
583	default:
584		af = AF_UNSPEC;
585		hdrsize = 0;
586		break;
587	}
588
589	while (m->m_len <= hlen) {
590		hlen -= m->m_len;
591		m = m->m_next;
592	}
593	if (m->m_len < (hlen + hdrsize)) {
594		/*
595		 * protocol header not in a single mbuf.
596		 * We can't cope with this situation right
597		 * now (but it shouldn't ever happen, really, anyhow).
598		 */
599#ifdef DEBUG
600		printf("altq_etherclassify: headers span multiple mbufs: "
601		    "%d < %d\n", m->m_len, (hlen + hdrsize));
602#endif
603		goto bad;
604	}
605
606	m->m_data += hlen;
607	m->m_len -= hlen;
608
609	hdr = mtod(m, void *);
610
611	if (ALTQ_NEEDS_CLASSIFY(ifq))
612		pktattr->pattr_class =
613		    (*ifq->altq_classify)(ifq->altq_clfier, m, af);
614	pktattr->pattr_af = af;
615	pktattr->pattr_hdr = hdr;
616
617	m->m_data -= hlen;
618	m->m_len += hlen;
619
620	return;
621
622 bad:
623	pktattr->pattr_class = NULL;
624	pktattr->pattr_hdr = NULL;
625	pktattr->pattr_af = AF_UNSPEC;
626}
627#endif /* ALTQ */
628
629/*
630 * Process a received Ethernet packet;
631 * the packet is in the mbuf chain m with
632 * the ether header.
633 */
634void
635ether_input(struct ifnet *ifp, struct mbuf *m)
636{
637	struct ethercom *ec = (struct ethercom *) ifp;
638	struct ifqueue *inq;
639	uint16_t etype;
640	struct ether_header *eh;
641	size_t ehlen;
642#if defined (ISO) || defined (LLC) || defined(NETATALK)
643	struct llc *l;
644#endif
645
646	if ((ifp->if_flags & IFF_UP) == 0) {
647		m_freem(m);
648		return;
649	}
650
651#ifdef MBUFTRACE
652	m_claimm(m, &ec->ec_rx_mowner);
653#endif
654	eh = mtod(m, struct ether_header *);
655	etype = ntohs(eh->ether_type);
656	ehlen = sizeof(*eh);
657
658	/*
659	 * Determine if the packet is within its size limits.
660	 */
661	if (etype != ETHERTYPE_MPLS && m->m_pkthdr.len >
662	    ETHER_MAX_FRAME(ifp, etype, m->m_flags & M_HASFCS)) {
663		if (ppsratecheck(&bigpktppslim_last, &bigpktpps_count,
664			    bigpktppslim)) {
665			printf("%s: discarding oversize frame (len=%d)\n",
666			    ifp->if_xname, m->m_pkthdr.len);
667		}
668		m_freem(m);
669		return;
670	}
671
672	if (ETHER_IS_MULTICAST(eh->ether_dhost)) {
673		/*
674		 * If this is not a simplex interface, drop the packet
675		 * if it came from us.
676		 */
677		if ((ifp->if_flags & IFF_SIMPLEX) == 0 &&
678		    memcmp(CLLADDR(ifp->if_sadl), eh->ether_shost,
679		    ETHER_ADDR_LEN) == 0) {
680			m_freem(m);
681			return;
682		}
683
684		if (memcmp(etherbroadcastaddr,
685		    eh->ether_dhost, ETHER_ADDR_LEN) == 0)
686			m->m_flags |= M_BCAST;
687		else
688			m->m_flags |= M_MCAST;
689		ifp->if_imcasts++;
690	}
691
692	/* If the CRC is still on the packet, trim it off. */
693	if (m->m_flags & M_HASFCS) {
694		m_adj(m, -ETHER_CRC_LEN);
695		m->m_flags &= ~M_HASFCS;
696	}
697
698	ifp->if_ibytes += m->m_pkthdr.len;
699
700#if NBRIDGE > 0
701	/*
702	 * Tap the packet off here for a bridge.  bridge_input()
703	 * will return NULL if it has consumed the packet, otherwise
704	 * it gets processed as normal.  Note that bridge_input()
705	 * will always return the original packet if we need to
706	 * process it locally.
707	 */
708	if (ifp->if_bridge) {
709		/* clear M_PROMISC, in case the packets comes from a vlan */
710		m->m_flags &= ~M_PROMISC;
711		m = bridge_input(ifp, m);
712		if (m == NULL)
713			return;
714
715		/*
716		 * Bridge has determined that the packet is for us.
717		 * Update our interface pointer -- we may have had
718		 * to "bridge" the packet locally.
719		 */
720		ifp = m->m_pkthdr.rcvif;
721	} else
722#endif /* NBRIDGE > 0 */
723	{
724
725#if NCARP > 0
726		if (__predict_false(ifp->if_carp && ifp->if_type != IFT_CARP)) {
727			/*
728			 * clear M_PROMISC, in case the packets comes from a
729			 * vlan
730			 */
731			m->m_flags &= ~M_PROMISC;
732			if (carp_input(m, (uint8_t *)&eh->ether_shost,
733			    (uint8_t *)&eh->ether_dhost, eh->ether_type) == 0)
734				return;
735		}
736#endif /* NCARP > 0 */
737		if ((m->m_flags & (M_BCAST|M_MCAST|M_PROMISC)) == 0 &&
738		    (ifp->if_flags & IFF_PROMISC) != 0 &&
739		    memcmp(CLLADDR(ifp->if_sadl), eh->ether_dhost,
740			   ETHER_ADDR_LEN) != 0) {
741			m->m_flags |= M_PROMISC;
742		}
743	}
744
745#ifdef PFIL_HOOKS
746	if ((m->m_flags & M_PROMISC) == 0) {
747		if (pfil_run_hooks(&ifp->if_pfil, &m, ifp, PFIL_IN) != 0)
748			return;
749		if (m == NULL)
750			return;
751
752		eh = mtod(m, struct ether_header *);
753		etype = ntohs(eh->ether_type);
754		ehlen = sizeof(*eh);
755	}
756#endif
757
758#if NAGR > 0
759	if (ifp->if_agrprivate &&
760	    __predict_true(etype != ETHERTYPE_SLOWPROTOCOLS)) {
761		m->m_flags &= ~M_PROMISC;
762		agr_input(ifp, m);
763		return;
764	}
765#endif /* NAGR > 0 */
766
767	/*
768	 * If VLANs are configured on the interface, check to
769	 * see if the device performed the decapsulation and
770	 * provided us with the tag.
771	 */
772	if (ec->ec_nvlans && m_tag_find(m, PACKET_TAG_VLAN, NULL) != NULL) {
773#if NVLAN > 0
774		/*
775		 * vlan_input() will either recursively call ether_input()
776		 * or drop the packet.
777		 */
778		vlan_input(ifp, m);
779#else
780		m_freem(m);
781#endif
782		return;
783	}
784
785	/*
786	 * Handle protocols that expect to have the Ethernet header
787	 * (and possibly FCS) intact.
788	 */
789	switch (etype) {
790	case ETHERTYPE_VLAN: {
791		struct ether_vlan_header *evl = (void *)eh;
792		/*
793		 * If there is a tag of 0, then the VLAN header was probably
794		 * just being used to store the priority.  Extract the ether
795		 * type, and if IP or IPV6, let them deal with it.
796		 */
797		if (m->m_len <= sizeof(*evl)
798		    && EVL_VLANOFTAG(evl->evl_tag) == 0) {
799			etype = ntohs(evl->evl_proto);
800			ehlen = sizeof(*evl);
801			if ((m->m_flags & M_PROMISC) == 0
802			    && (etype == ETHERTYPE_IP
803				|| etype == ETHERTYPE_IPV6))
804				break;
805		}
806#if NVLAN > 0
807		/*
808		 * vlan_input() will either recursively call ether_input()
809		 * or drop the packet.
810		 */
811		if (((struct ethercom *)ifp)->ec_nvlans != 0)
812			vlan_input(ifp, m);
813		else
814#endif /* NVLAN > 0 */
815			m_freem(m);
816		return;
817	}
818#if NPPPOE > 0
819	case ETHERTYPE_PPPOEDISC:
820	case ETHERTYPE_PPPOE:
821		if (m->m_flags & M_PROMISC) {
822			m_freem(m);
823			return;
824		}
825#ifndef PPPOE_SERVER
826		if (m->m_flags & (M_MCAST | M_BCAST)) {
827			m_freem(m);
828			return;
829		}
830#endif
831
832		if (etype == ETHERTYPE_PPPOEDISC)
833			inq = &ppoediscinq;
834		else
835			inq = &ppoeinq;
836		if (IF_QFULL(inq)) {
837			IF_DROP(inq);
838			m_freem(m);
839		} else
840			IF_ENQUEUE(inq, m);
841		softint_schedule(pppoe_softintr);
842		return;
843#endif /* NPPPOE > 0 */
844	case ETHERTYPE_SLOWPROTOCOLS: {
845		uint8_t subtype;
846
847#if defined(DIAGNOSTIC)
848		if (m->m_pkthdr.len < sizeof(*eh) + sizeof(subtype)) {
849			panic("ether_input: too short slow protocol packet");
850		}
851#endif
852		m_copydata(m, sizeof(*eh), sizeof(subtype), &subtype);
853		switch (subtype) {
854#if NAGR > 0
855		case SLOWPROTOCOLS_SUBTYPE_LACP:
856			if (ifp->if_agrprivate) {
857				ieee8023ad_lacp_input(ifp, m);
858				return;
859			}
860			break;
861
862		case SLOWPROTOCOLS_SUBTYPE_MARKER:
863			if (ifp->if_agrprivate) {
864				ieee8023ad_marker_input(ifp, m);
865				return;
866			}
867			break;
868#endif /* NAGR > 0 */
869		default:
870			if (subtype == 0 || subtype > 10) {
871				/* illegal value */
872				m_freem(m);
873				return;
874			}
875			/* unknown subtype */
876			break;
877		}
878		/* FALLTHROUGH */
879	}
880	default:
881		if (m->m_flags & M_PROMISC) {
882			m_freem(m);
883			return;
884		}
885	}
886
887	/* If the CRC is still on the packet, trim it off. */
888	if (m->m_flags & M_HASFCS) {
889		m_adj(m, -ETHER_CRC_LEN);
890		m->m_flags &= ~M_HASFCS;
891	}
892
893	if (etype > ETHERMTU + sizeof (struct ether_header)) {
894		/* Strip off the Ethernet header. */
895		m_adj(m, ehlen);
896
897		switch (etype) {
898#ifdef INET
899		case ETHERTYPE_IP:
900#ifdef GATEWAY
901			if (ipflow_fastforward(m))
902				return;
903#endif
904			schednetisr(NETISR_IP);
905			inq = &ipintrq;
906			break;
907
908		case ETHERTYPE_ARP:
909			schednetisr(NETISR_ARP);
910			inq = &arpintrq;
911			break;
912
913		case ETHERTYPE_REVARP:
914			revarpinput(m);	/* XXX queue? */
915			return;
916#endif
917#ifdef INET6
918		case ETHERTYPE_IPV6:
919#ifdef GATEWAY
920			if (ip6flow_fastforward(&m))
921				return;
922#endif
923			schednetisr(NETISR_IPV6);
924			inq = &ip6intrq;
925			break;
926#endif
927#ifdef IPX
928		case ETHERTYPE_IPX:
929			schednetisr(NETISR_IPX);
930			inq = &ipxintrq;
931			break;
932#endif
933#ifdef NETATALK
934		case ETHERTYPE_ATALK:
935			schednetisr(NETISR_ATALK);
936			inq = &atintrq1;
937			break;
938		case ETHERTYPE_AARP:
939			/* probably this should be done with a NETISR as well */
940			aarpinput(ifp, m); /* XXX */
941			return;
942#endif /* NETATALK */
943#ifdef MPLS
944		case ETHERTYPE_MPLS:
945			schednetisr(NETISR_MPLS);
946			inq = &mplsintrq;
947			break;
948#endif
949		default:
950			m_freem(m);
951			return;
952		}
953	} else {
954#if defined (ISO) || defined (LLC) || defined (NETATALK)
955		l = (struct llc *)(eh+1);
956		switch (l->llc_dsap) {
957#ifdef NETATALK
958		case LLC_SNAP_LSAP:
959			switch (l->llc_control) {
960			case LLC_UI:
961				if (l->llc_ssap != LLC_SNAP_LSAP) {
962					goto dropanyway;
963				}
964
965				if (memcmp(&(l->llc_snap_org_code)[0],
966				    at_org_code, sizeof(at_org_code)) == 0 &&
967				    ntohs(l->llc_snap_ether_type) ==
968				    ETHERTYPE_ATALK) {
969					inq = &atintrq2;
970					m_adj(m, sizeof(struct ether_header)
971					    + sizeof(struct llc));
972					schednetisr(NETISR_ATALK);
973					break;
974				}
975
976				if (memcmp(&(l->llc_snap_org_code)[0],
977				    aarp_org_code,
978				    sizeof(aarp_org_code)) == 0 &&
979				    ntohs(l->llc_snap_ether_type) ==
980				    ETHERTYPE_AARP) {
981					m_adj( m, sizeof(struct ether_header)
982					    + sizeof(struct llc));
983					aarpinput(ifp, m); /* XXX */
984				    return;
985				}
986
987			default:
988				goto dropanyway;
989			}
990			break;
991#endif /* NETATALK */
992#ifdef	ISO
993		case LLC_ISO_LSAP:
994			switch (l->llc_control) {
995			case LLC_UI:
996				/* LLC_UI_P forbidden in class 1 service */
997				if ((l->llc_dsap == LLC_ISO_LSAP) &&	/* XXX? case tested */
998				    (l->llc_ssap == LLC_ISO_LSAP)) {
999					/* LSAP for ISO */
1000					/* XXX length computation?? */
1001					if (m->m_pkthdr.len > etype + sizeof(struct ether_header))
1002						m_adj(m, etype - m->m_pkthdr.len);
1003
1004#ifdef ARGO_DEBUG
1005					if (argo_debug[D_ETHER])
1006						printf("clnp packet");
1007#endif
1008					schednetisr(NETISR_ISO);
1009					inq = &clnlintrq;
1010					break;
1011				}
1012				goto dropanyway;
1013
1014			case LLC_XID:
1015			case LLC_XID_P:
1016				if(m->m_len < LLC_XID_BASIC_MINLEN + sizeof(struct ether_header))
1017					/* XXX m_pullup? */
1018					goto dropanyway;
1019				l->llc_window = 0;
1020				l->llc_fid = LLC_XID_FORMAT_BASIC;
1021				l->llc_class = LLC_XID_CLASS_I;
1022				l->llc_dsap = l->llc_ssap = 0;
1023				/* Fall through to */
1024			case LLC_TEST:
1025			case LLC_TEST_P:
1026			{
1027				struct sockaddr sa;
1028				struct ether_header *eh2;
1029				int i;
1030				u_char c = l->llc_dsap;
1031
1032				l->llc_dsap = l->llc_ssap;
1033				l->llc_ssap = c;
1034				m_adj(m, sizeof(struct ether_header));
1035				/* XXX we can optimize here? */
1036				if (m->m_flags & (M_BCAST | M_MCAST))
1037					memcpy(eh->ether_dhost,
1038					    CLLADDR(ifp->if_sadl),
1039					    ETHER_ADDR_LEN);
1040				sa.sa_family = AF_UNSPEC;
1041				sa.sa_len = sizeof(sa);
1042				eh2 = (struct ether_header *)sa.sa_data;
1043				for (i = 0; i < 6; i++) {
1044					eh2->ether_shost[i] = c =
1045					    eh->ether_dhost[i];
1046					eh2->ether_dhost[i] =
1047					    eh->ether_dhost[i] =
1048					    eh->ether_shost[i];
1049					eh->ether_shost[i] = c;
1050				}
1051				ifp->if_output(ifp, m, &sa, NULL);
1052				return;
1053			}
1054			default:
1055				m_freem(m);
1056				return;
1057			}
1058			break;
1059#endif /* ISO */
1060#if defined (ISO) || defined (NETATALK)
1061		dropanyway:
1062#endif
1063		default:
1064			m_freem(m);
1065			return;
1066		}
1067#else /* ISO || LLC || NETATALK*/
1068		m_freem(m);
1069		return;
1070#endif /* ISO || LLC || NETATALK*/
1071	}
1072
1073	if (IF_QFULL(inq)) {
1074		IF_DROP(inq);
1075		m_freem(m);
1076	} else
1077		IF_ENQUEUE(inq, m);
1078}
1079
1080/*
1081 * Convert Ethernet address to printable (loggable) representation.
1082 */
1083char *
1084ether_sprintf(const u_char *ap)
1085{
1086	static char etherbuf[3 * ETHER_ADDR_LEN];
1087	return ether_snprintf(etherbuf, sizeof(etherbuf), ap);
1088}
1089
1090char *
1091ether_snprintf(char *buf, size_t len, const u_char *ap)
1092{
1093	char *cp = buf;
1094	size_t i;
1095
1096	for (i = 0; i < len / 3; i++) {
1097		*cp++ = hexdigits[*ap >> 4];
1098		*cp++ = hexdigits[*ap++ & 0xf];
1099		*cp++ = ':';
1100	}
1101	*--cp = '\0';
1102	return buf;
1103}
1104
1105/*
1106 * Perform common duties while attaching to interface list
1107 */
1108void
1109ether_ifattach(struct ifnet *ifp, const uint8_t *lla)
1110{
1111	struct ethercom *ec = (struct ethercom *)ifp;
1112
1113	ifp->if_type = IFT_ETHER;
1114	ifp->if_hdrlen = ETHER_HDR_LEN;
1115	ifp->if_dlt = DLT_EN10MB;
1116	ifp->if_mtu = ETHERMTU;
1117	ifp->if_output = ether_output;
1118	ifp->if_input = ether_input;
1119	if (ifp->if_baudrate == 0)
1120		ifp->if_baudrate = IF_Mbps(10);		/* just a default */
1121
1122	if_set_sadl(ifp, lla, ETHER_ADDR_LEN, !ETHER_IS_LOCAL(lla));
1123
1124	LIST_INIT(&ec->ec_multiaddrs);
1125	ifp->if_broadcastaddr = etherbroadcastaddr;
1126	bpf_attach(ifp, DLT_EN10MB, sizeof(struct ether_header));
1127#ifdef MBUFTRACE
1128	strlcpy(ec->ec_tx_mowner.mo_name, ifp->if_xname,
1129	    sizeof(ec->ec_tx_mowner.mo_name));
1130	strlcpy(ec->ec_tx_mowner.mo_descr, "tx",
1131	    sizeof(ec->ec_tx_mowner.mo_descr));
1132	strlcpy(ec->ec_rx_mowner.mo_name, ifp->if_xname,
1133	    sizeof(ec->ec_rx_mowner.mo_name));
1134	strlcpy(ec->ec_rx_mowner.mo_descr, "rx",
1135	    sizeof(ec->ec_rx_mowner.mo_descr));
1136	MOWNER_ATTACH(&ec->ec_tx_mowner);
1137	MOWNER_ATTACH(&ec->ec_rx_mowner);
1138	ifp->if_mowner = &ec->ec_tx_mowner;
1139#endif
1140}
1141
1142void
1143ether_ifdetach(struct ifnet *ifp)
1144{
1145	struct ethercom *ec = (void *) ifp;
1146	struct ether_multi *enm;
1147	int s;
1148
1149	/*
1150	 * Prevent further calls to ioctl (for example turning off
1151	 * promiscuous mode from the bridge code), which eventually can
1152	 * call if_init() which can cause panics because the interface
1153	 * is in the process of being detached. Return device not configured
1154	 * instead.
1155	 */
1156	ifp->if_ioctl = (int (*)(struct ifnet *, u_long, void *))enxio;
1157
1158#if NBRIDGE > 0
1159	if (ifp->if_bridge)
1160		bridge_ifdetach(ifp);
1161#endif
1162
1163	bpf_detach(ifp);
1164
1165#if NVLAN > 0
1166	if (ec->ec_nvlans)
1167		vlan_ifdetach(ifp);
1168#endif
1169
1170	s = splnet();
1171	while ((enm = LIST_FIRST(&ec->ec_multiaddrs)) != NULL) {
1172		LIST_REMOVE(enm, enm_list);
1173		free(enm, M_IFMADDR);
1174		ec->ec_multicnt--;
1175	}
1176	splx(s);
1177
1178#if 0	/* done in if_detach() */
1179	if_free_sadl(ifp);
1180#endif
1181
1182	ifp->if_mowner = NULL;
1183	MOWNER_DETACH(&ec->ec_rx_mowner);
1184	MOWNER_DETACH(&ec->ec_tx_mowner);
1185}
1186
1187#if 0
1188/*
1189 * This is for reference.  We have a table-driven version
1190 * of the little-endian crc32 generator, which is faster
1191 * than the double-loop.
1192 */
1193uint32_t
1194ether_crc32_le(const uint8_t *buf, size_t len)
1195{
1196	uint32_t c, crc, carry;
1197	size_t i, j;
1198
1199	crc = 0xffffffffU;	/* initial value */
1200
1201	for (i = 0; i < len; i++) {
1202		c = buf[i];
1203		for (j = 0; j < 8; j++) {
1204			carry = ((crc & 0x01) ? 1 : 0) ^ (c & 0x01);
1205			crc >>= 1;
1206			c >>= 1;
1207			if (carry)
1208				crc = (crc ^ ETHER_CRC_POLY_LE);
1209		}
1210	}
1211
1212	return (crc);
1213}
1214#else
1215uint32_t
1216ether_crc32_le(const uint8_t *buf, size_t len)
1217{
1218	static const uint32_t crctab[] = {
1219		0x00000000, 0x1db71064, 0x3b6e20c8, 0x26d930ac,
1220		0x76dc4190, 0x6b6b51f4, 0x4db26158, 0x5005713c,
1221		0xedb88320, 0xf00f9344, 0xd6d6a3e8, 0xcb61b38c,
1222		0x9b64c2b0, 0x86d3d2d4, 0xa00ae278, 0xbdbdf21c
1223	};
1224	uint32_t crc;
1225	size_t i;
1226
1227	crc = 0xffffffffU;	/* initial value */
1228
1229	for (i = 0; i < len; i++) {
1230		crc ^= buf[i];
1231		crc = (crc >> 4) ^ crctab[crc & 0xf];
1232		crc = (crc >> 4) ^ crctab[crc & 0xf];
1233	}
1234
1235	return (crc);
1236}
1237#endif
1238
1239uint32_t
1240ether_crc32_be(const uint8_t *buf, size_t len)
1241{
1242	uint32_t c, crc, carry;
1243	size_t i, j;
1244
1245	crc = 0xffffffffU;	/* initial value */
1246
1247	for (i = 0; i < len; i++) {
1248		c = buf[i];
1249		for (j = 0; j < 8; j++) {
1250			carry = ((crc & 0x80000000U) ? 1 : 0) ^ (c & 0x01);
1251			crc <<= 1;
1252			c >>= 1;
1253			if (carry)
1254				crc = (crc ^ ETHER_CRC_POLY_BE) | carry;
1255		}
1256	}
1257
1258	return (crc);
1259}
1260
1261#ifdef INET
1262const uint8_t ether_ipmulticast_min[ETHER_ADDR_LEN] =
1263    { 0x01, 0x00, 0x5e, 0x00, 0x00, 0x00 };
1264const uint8_t ether_ipmulticast_max[ETHER_ADDR_LEN] =
1265    { 0x01, 0x00, 0x5e, 0x7f, 0xff, 0xff };
1266#endif
1267#ifdef INET6
1268const uint8_t ether_ip6multicast_min[ETHER_ADDR_LEN] =
1269    { 0x33, 0x33, 0x00, 0x00, 0x00, 0x00 };
1270const uint8_t ether_ip6multicast_max[ETHER_ADDR_LEN] =
1271    { 0x33, 0x33, 0xff, 0xff, 0xff, 0xff };
1272#endif
1273
1274/*
1275 * ether_aton implementation, not using a static buffer.
1276 */
1277int
1278ether_aton_r(u_char *dest, size_t len, const char *str)
1279{
1280        const u_char *cp = (const void *)str;
1281	u_char *ep;
1282
1283#define atox(c)	(((c) <= '9') ? ((c) - '0') : ((toupper(c) - 'A') + 10))
1284
1285	if (len < ETHER_ADDR_LEN)
1286		return ENOSPC;
1287
1288	ep = dest + ETHER_ADDR_LEN;
1289
1290	while (*cp) {
1291                if (!isxdigit(*cp))
1292                        return EINVAL;
1293		*dest = atox(*cp);
1294		cp++;
1295                if (isxdigit(*cp)) {
1296                        *dest = (*dest << 4) | atox(*cp);
1297			dest++;
1298			cp++;
1299                } else
1300			dest++;
1301		if (dest == ep)
1302			return *cp == '\0' ? 0 : ENAMETOOLONG;
1303		switch (*cp) {
1304		case ':':
1305		case '-':
1306		case '.':
1307			cp++;
1308			break;
1309		}
1310        }
1311	return ENOBUFS;
1312}
1313
1314/*
1315 * Convert a sockaddr into an Ethernet address or range of Ethernet
1316 * addresses.
1317 */
1318int
1319ether_multiaddr(const struct sockaddr *sa, uint8_t addrlo[ETHER_ADDR_LEN],
1320    uint8_t addrhi[ETHER_ADDR_LEN])
1321{
1322#ifdef INET
1323	const struct sockaddr_in *sin;
1324#endif /* INET */
1325#ifdef INET6
1326	const struct sockaddr_in6 *sin6;
1327#endif /* INET6 */
1328
1329	switch (sa->sa_family) {
1330
1331	case AF_UNSPEC:
1332		memcpy(addrlo, sa->sa_data, ETHER_ADDR_LEN);
1333		memcpy(addrhi, addrlo, ETHER_ADDR_LEN);
1334		break;
1335
1336#ifdef INET
1337	case AF_INET:
1338		sin = satocsin(sa);
1339		if (sin->sin_addr.s_addr == INADDR_ANY) {
1340			/*
1341			 * An IP address of INADDR_ANY means listen to
1342			 * or stop listening to all of the Ethernet
1343			 * multicast addresses used for IP.
1344			 * (This is for the sake of IP multicast routers.)
1345			 */
1346			memcpy(addrlo, ether_ipmulticast_min, ETHER_ADDR_LEN);
1347			memcpy(addrhi, ether_ipmulticast_max, ETHER_ADDR_LEN);
1348		}
1349		else {
1350			ETHER_MAP_IP_MULTICAST(&sin->sin_addr, addrlo);
1351			memcpy(addrhi, addrlo, ETHER_ADDR_LEN);
1352		}
1353		break;
1354#endif
1355#ifdef INET6
1356	case AF_INET6:
1357		sin6 = satocsin6(sa);
1358		if (IN6_IS_ADDR_UNSPECIFIED(&sin6->sin6_addr)) {
1359			/*
1360			 * An IP6 address of 0 means listen to or stop
1361			 * listening to all of the Ethernet multicast
1362			 * address used for IP6.
1363			 * (This is used for multicast routers.)
1364			 */
1365			memcpy(addrlo, ether_ip6multicast_min, ETHER_ADDR_LEN);
1366			memcpy(addrhi, ether_ip6multicast_max, ETHER_ADDR_LEN);
1367		} else {
1368			ETHER_MAP_IPV6_MULTICAST(&sin6->sin6_addr, addrlo);
1369			memcpy(addrhi, addrlo, ETHER_ADDR_LEN);
1370		}
1371		break;
1372#endif
1373
1374	default:
1375		return EAFNOSUPPORT;
1376	}
1377	return 0;
1378}
1379
1380/*
1381 * Add an Ethernet multicast address or range of addresses to the list for a
1382 * given interface.
1383 */
1384int
1385ether_addmulti(const struct sockaddr *sa, struct ethercom *ec)
1386{
1387	struct ether_multi *enm;
1388	u_char addrlo[ETHER_ADDR_LEN];
1389	u_char addrhi[ETHER_ADDR_LEN];
1390	int s = splnet(), error;
1391
1392	error = ether_multiaddr(sa, addrlo, addrhi);
1393	if (error != 0) {
1394		splx(s);
1395		return error;
1396	}
1397
1398	/*
1399	 * Verify that we have valid Ethernet multicast addresses.
1400	 */
1401	if (!ETHER_IS_MULTICAST(addrlo) || !ETHER_IS_MULTICAST(addrhi)) {
1402		splx(s);
1403		return EINVAL;
1404	}
1405	/*
1406	 * See if the address range is already in the list.
1407	 */
1408	ETHER_LOOKUP_MULTI(addrlo, addrhi, ec, enm);
1409	if (enm != NULL) {
1410		/*
1411		 * Found it; just increment the reference count.
1412		 */
1413		++enm->enm_refcount;
1414		splx(s);
1415		return 0;
1416	}
1417	/*
1418	 * New address or range; malloc a new multicast record
1419	 * and link it into the interface's multicast list.
1420	 */
1421	enm = (struct ether_multi *)malloc(sizeof(*enm), M_IFMADDR, M_NOWAIT);
1422	if (enm == NULL) {
1423		splx(s);
1424		return ENOBUFS;
1425	}
1426	memcpy(enm->enm_addrlo, addrlo, 6);
1427	memcpy(enm->enm_addrhi, addrhi, 6);
1428	enm->enm_refcount = 1;
1429	LIST_INSERT_HEAD(&ec->ec_multiaddrs, enm, enm_list);
1430	ec->ec_multicnt++;
1431	splx(s);
1432	/*
1433	 * Return ENETRESET to inform the driver that the list has changed
1434	 * and its reception filter should be adjusted accordingly.
1435	 */
1436	return ENETRESET;
1437}
1438
1439/*
1440 * Delete a multicast address record.
1441 */
1442int
1443ether_delmulti(const struct sockaddr *sa, struct ethercom *ec)
1444{
1445	struct ether_multi *enm;
1446	u_char addrlo[ETHER_ADDR_LEN];
1447	u_char addrhi[ETHER_ADDR_LEN];
1448	int s = splnet(), error;
1449
1450	error = ether_multiaddr(sa, addrlo, addrhi);
1451	if (error != 0) {
1452		splx(s);
1453		return (error);
1454	}
1455
1456	/*
1457	 * Look ur the address in our list.
1458	 */
1459	ETHER_LOOKUP_MULTI(addrlo, addrhi, ec, enm);
1460	if (enm == NULL) {
1461		splx(s);
1462		return (ENXIO);
1463	}
1464	if (--enm->enm_refcount != 0) {
1465		/*
1466		 * Still some claims to this record.
1467		 */
1468		splx(s);
1469		return (0);
1470	}
1471	/*
1472	 * No remaining claims to this record; unlink and free it.
1473	 */
1474	LIST_REMOVE(enm, enm_list);
1475	free(enm, M_IFMADDR);
1476	ec->ec_multicnt--;
1477	splx(s);
1478	/*
1479	 * Return ENETRESET to inform the driver that the list has changed
1480	 * and its reception filter should be adjusted accordingly.
1481	 */
1482	return (ENETRESET);
1483}
1484
1485void
1486ether_set_ifflags_cb(struct ethercom *ec, ether_cb_t cb)
1487{
1488	ec->ec_ifflags_cb = cb;
1489}
1490
1491/*
1492 * Common ioctls for Ethernet interfaces.  Note, we must be
1493 * called at splnet().
1494 */
1495int
1496ether_ioctl(struct ifnet *ifp, u_long cmd, void *data)
1497{
1498	struct ethercom *ec = (void *) ifp;
1499	struct ifreq *ifr = (struct ifreq *)data;
1500	struct if_laddrreq *iflr = data;
1501	const struct sockaddr_dl *sdl;
1502	static const uint8_t zero[ETHER_ADDR_LEN];
1503	int error;
1504
1505	switch (cmd) {
1506	case SIOCINITIFADDR:
1507		if ((ifp->if_flags & (IFF_UP|IFF_RUNNING)) !=
1508		    (IFF_UP|IFF_RUNNING)) {
1509			ifp->if_flags |= IFF_UP;
1510			if ((error = (*ifp->if_init)(ifp)) != 0)
1511				return error;
1512		}
1513#ifdef INET
1514		{
1515			struct ifaddr *ifa = (struct ifaddr *)data;
1516
1517			if (ifa->ifa_addr->sa_family == AF_INET)
1518				arp_ifinit(ifp, ifa);
1519		}
1520#endif /* INET */
1521		return 0;
1522
1523	case SIOCSIFMTU:
1524	    {
1525		int maxmtu;
1526
1527		if (ec->ec_capabilities & ETHERCAP_JUMBO_MTU)
1528			maxmtu = ETHERMTU_JUMBO;
1529		else
1530			maxmtu = ETHERMTU;
1531
1532		if (ifr->ifr_mtu < ETHERMIN || ifr->ifr_mtu > maxmtu)
1533			return EINVAL;
1534		else if ((error = ifioctl_common(ifp, cmd, data)) != ENETRESET)
1535			return error;
1536		else if (ifp->if_flags & IFF_UP) {
1537			/* Make sure the device notices the MTU change. */
1538			return (*ifp->if_init)(ifp);
1539		} else
1540			return 0;
1541	    }
1542
1543	case SIOCSIFFLAGS:
1544		if ((error = ifioctl_common(ifp, cmd, data)) != 0)
1545			return error;
1546		switch (ifp->if_flags & (IFF_UP|IFF_RUNNING)) {
1547		case IFF_RUNNING:
1548			/*
1549			 * If interface is marked down and it is running,
1550			 * then stop and disable it.
1551			 */
1552			(*ifp->if_stop)(ifp, 1);
1553			break;
1554		case IFF_UP:
1555			/*
1556			 * If interface is marked up and it is stopped, then
1557			 * start it.
1558			 */
1559			return (*ifp->if_init)(ifp);
1560		case IFF_UP|IFF_RUNNING:
1561			error = 0;
1562			if (ec->ec_ifflags_cb == NULL ||
1563			    (error = (*ec->ec_ifflags_cb)(ec)) == ENETRESET) {
1564				/*
1565				 * Reset the interface to pick up
1566				 * changes in any other flags that
1567				 * affect the hardware state.
1568				 */
1569				return (*ifp->if_init)(ifp);
1570			} else
1571				return error;
1572		case 0:
1573			break;
1574		}
1575		return 0;
1576	case SIOCADDMULTI:
1577		return ether_addmulti(ifreq_getaddr(cmd, ifr), ec);
1578	case SIOCDELMULTI:
1579		return ether_delmulti(ifreq_getaddr(cmd, ifr), ec);
1580	case SIOCSIFMEDIA:
1581	case SIOCGIFMEDIA:
1582		if (ec->ec_mii == NULL)
1583			return ENOTTY;
1584		return ifmedia_ioctl(ifp, ifr, &ec->ec_mii->mii_media, cmd);
1585	case SIOCALIFADDR:
1586		sdl = satocsdl(sstocsa(&iflr->addr));
1587		if (sdl->sdl_family != AF_LINK)
1588			;
1589		else if (ETHER_IS_MULTICAST(CLLADDR(sdl)))
1590			return EINVAL;
1591		else if (memcmp(zero, CLLADDR(sdl), sizeof(zero)) == 0)
1592			return EINVAL;
1593		/*FALLTHROUGH*/
1594	default:
1595		return ifioctl_common(ifp, cmd, data);
1596	}
1597	return 0;
1598}
1599