xref: /netbsd-src/sys/net/if.c (revision 7f21db1c0118155e0dd40b75182e30c589d9f63e)
1 /*	$NetBSD: if.c,v 1.242 2010/01/28 14:12:11 mbalmer Exp $	*/
2 
3 /*-
4  * Copyright (c) 1999, 2000, 2001, 2008 The NetBSD Foundation, Inc.
5  * All rights reserved.
6  *
7  * This code is derived from software contributed to The NetBSD Foundation
8  * by William Studenmund and Jason R. Thorpe.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
20  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
21  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
22  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
23  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
24  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
25  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
26  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
27  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
28  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
29  * POSSIBILITY OF SUCH DAMAGE.
30  */
31 
32 /*
33  * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
34  * 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 project 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 PROJECT 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 PROJECT 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 
61 /*
62  * Copyright (c) 1980, 1986, 1993
63  *	The Regents of the University of California.  All rights reserved.
64  *
65  * Redistribution and use in source and binary forms, with or without
66  * modification, are permitted provided that the following conditions
67  * are met:
68  * 1. Redistributions of source code must retain the above copyright
69  *    notice, this list of conditions and the following disclaimer.
70  * 2. Redistributions in binary form must reproduce the above copyright
71  *    notice, this list of conditions and the following disclaimer in the
72  *    documentation and/or other materials provided with the distribution.
73  * 3. Neither the name of the University nor the names of its contributors
74  *    may be used to endorse or promote products derived from this software
75  *    without specific prior written permission.
76  *
77  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
78  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
79  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
80  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
81  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
82  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
83  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
84  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
85  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
86  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
87  * SUCH DAMAGE.
88  *
89  *	@(#)if.c	8.5 (Berkeley) 1/9/95
90  */
91 
92 #include <sys/cdefs.h>
93 __KERNEL_RCSID(0, "$NetBSD: if.c,v 1.242 2010/01/28 14:12:11 mbalmer Exp $");
94 
95 #include "opt_inet.h"
96 
97 #include "opt_atalk.h"
98 #include "opt_natm.h"
99 #include "opt_pfil_hooks.h"
100 
101 #include <sys/param.h>
102 #include <sys/mbuf.h>
103 #include <sys/systm.h>
104 #include <sys/callout.h>
105 #include <sys/proc.h>
106 #include <sys/socket.h>
107 #include <sys/socketvar.h>
108 #include <sys/domain.h>
109 #include <sys/protosw.h>
110 #include <sys/kernel.h>
111 #include <sys/ioctl.h>
112 #include <sys/sysctl.h>
113 #include <sys/syslog.h>
114 #include <sys/kauth.h>
115 
116 #include <net/if.h>
117 #include <net/if_dl.h>
118 #include <net/if_ether.h>
119 #include <net/if_media.h>
120 #include <net80211/ieee80211.h>
121 #include <net80211/ieee80211_ioctl.h>
122 #include <net/if_types.h>
123 #include <net/radix.h>
124 #include <net/route.h>
125 #include <net/netisr.h>
126 #ifdef NETATALK
127 #include <netatalk/at_extern.h>
128 #include <netatalk/at.h>
129 #endif
130 #include <net/pfil.h>
131 
132 #ifdef INET6
133 #include <netinet/in.h>
134 #include <netinet6/in6_var.h>
135 #include <netinet6/nd6.h>
136 #endif
137 
138 #include "carp.h"
139 #if NCARP > 0
140 #include <netinet/ip_carp.h>
141 #endif
142 
143 #include <compat/sys/sockio.h>
144 #include <compat/sys/socket.h>
145 
146 MALLOC_DEFINE(M_IFADDR, "ifaddr", "interface address");
147 MALLOC_DEFINE(M_IFMADDR, "ether_multi", "link-level multicast address");
148 
149 int	ifqmaxlen = IFQ_MAXLEN;
150 callout_t if_slowtimo_ch;
151 
152 int netisr;			/* scheduling bits for network */
153 
154 static int	if_rt_walktree(struct rtentry *, void *);
155 
156 static struct if_clone *if_clone_lookup(const char *, int *);
157 static int	if_clone_list(struct if_clonereq *);
158 
159 static LIST_HEAD(, if_clone) if_cloners = LIST_HEAD_INITIALIZER(if_cloners);
160 static int if_cloners_count;
161 
162 static uint64_t index_gen;
163 static kmutex_t index_gen_mtx;
164 
165 #ifdef PFIL_HOOKS
166 struct pfil_head if_pfil;	/* packet filtering hook for interfaces */
167 #endif
168 
169 static kauth_listener_t if_listener;
170 
171 static void if_detach_queues(struct ifnet *, struct ifqueue *);
172 static void sysctl_sndq_setup(struct sysctllog **, const char *,
173     struct ifaltq *);
174 
175 #if defined(INET) || defined(INET6)
176 static void sysctl_net_ifq_setup(struct sysctllog **, int, const char *,
177 				 int, const char *, int, struct ifqueue *);
178 #endif
179 
180 static int
181 if_listener_cb(kauth_cred_t cred, kauth_action_t action, void *cookie,
182     void *arg0, void *arg1, void *arg2, void *arg3)
183 {
184 	int result;
185 	enum kauth_network_req req;
186 
187 	result = KAUTH_RESULT_DEFER;
188 	req = (enum kauth_network_req)arg1;
189 
190 	if (action != KAUTH_NETWORK_INTERFACE)
191 		return result;
192 
193 	if ((req == KAUTH_REQ_NETWORK_INTERFACE_GET) ||
194 	    (req == KAUTH_REQ_NETWORK_INTERFACE_SET))
195 		result = KAUTH_RESULT_ALLOW;
196 
197 	return result;
198 }
199 
200 /*
201  * Network interface utility routines.
202  *
203  * Routines with ifa_ifwith* names take sockaddr *'s as
204  * parameters.
205  */
206 void
207 ifinit(void)
208 {
209 #ifdef INET
210 	{extern struct ifqueue ipintrq;
211 	sysctl_net_ifq_setup(NULL, PF_INET, "inet", IPPROTO_IP, "ip",
212 			     IPCTL_IFQ, &ipintrq);}
213 #endif /* INET */
214 #ifdef INET6
215 	{extern struct ifqueue ip6intrq;
216 	sysctl_net_ifq_setup(NULL, PF_INET6, "inet6", IPPROTO_IPV6, "ip6",
217 			     IPV6CTL_IFQ, &ip6intrq);}
218 #endif /* INET6 */
219 
220 	callout_init(&if_slowtimo_ch, 0);
221 	if_slowtimo(NULL);
222 
223 	if_listener = kauth_listen_scope(KAUTH_SCOPE_NETWORK,
224 	    if_listener_cb, NULL);
225 }
226 
227 /*
228  * XXX Initialization before configure().
229  * XXX hack to get pfil_add_hook working in autoconf.
230  */
231 void
232 ifinit1(void)
233 {
234 
235 	mutex_init(&index_gen_mtx, MUTEX_DEFAULT, IPL_NONE);
236 
237 #ifdef PFIL_HOOKS
238 	if_pfil.ph_type = PFIL_TYPE_IFNET;
239 	if_pfil.ph_ifnet = NULL;
240 	if (pfil_head_register(&if_pfil) != 0)
241 		printf("WARNING: unable to register pfil hook\n");
242 #endif
243 }
244 
245 struct ifnet *
246 if_alloc(u_char type)
247 {
248 	return malloc(sizeof(struct ifnet), M_DEVBUF, M_WAITOK|M_ZERO);
249 }
250 
251 void
252 if_initname(struct ifnet *ifp, const char *name, int unit)
253 {
254 	(void)snprintf(ifp->if_xname, sizeof(ifp->if_xname),
255 	    "%s%d", name, unit);
256 }
257 
258 /*
259  * Null routines used while an interface is going away.  These routines
260  * just return an error.
261  */
262 
263 int
264 if_nulloutput(struct ifnet *ifp, struct mbuf *m,
265     const struct sockaddr *so, struct rtentry *rt)
266 {
267 
268 	return ENXIO;
269 }
270 
271 void
272 if_nullinput(struct ifnet *ifp, struct mbuf *m)
273 {
274 
275 	/* Nothing. */
276 }
277 
278 void
279 if_nullstart(struct ifnet *ifp)
280 {
281 
282 	/* Nothing. */
283 }
284 
285 int
286 if_nullioctl(struct ifnet *ifp, u_long cmd, void *data)
287 {
288 
289 	return ENXIO;
290 }
291 
292 int
293 if_nullinit(struct ifnet *ifp)
294 {
295 
296 	return ENXIO;
297 }
298 
299 void
300 if_nullstop(struct ifnet *ifp, int disable)
301 {
302 
303 	/* Nothing. */
304 }
305 
306 void
307 if_nullwatchdog(struct ifnet *ifp)
308 {
309 
310 	/* Nothing. */
311 }
312 
313 void
314 if_nulldrain(struct ifnet *ifp)
315 {
316 
317 	/* Nothing. */
318 }
319 
320 static u_int if_index = 1;
321 struct ifnet_head ifnet;
322 size_t if_indexlim = 0;
323 struct ifaddr **ifnet_addrs = NULL;
324 struct ifnet **ifindex2ifnet = NULL;
325 struct ifnet *lo0ifp;
326 
327 void
328 if_set_sadl(struct ifnet *ifp, const void *lla, u_char addrlen, bool factory)
329 {
330 	struct ifaddr *ifa;
331 	struct sockaddr_dl *sdl;
332 
333 	ifp->if_addrlen = addrlen;
334 	if_alloc_sadl(ifp);
335 	ifa = ifp->if_dl;
336 	sdl = satosdl(ifa->ifa_addr);
337 
338 	(void)sockaddr_dl_setaddr(sdl, sdl->sdl_len, lla, ifp->if_addrlen);
339 	if (factory) {
340 		ifp->if_hwdl = ifp->if_dl;
341 		IFAREF(ifp->if_hwdl);
342 	}
343 	/* TBD routing socket */
344 }
345 
346 struct ifaddr *
347 if_dl_create(const struct ifnet *ifp, const struct sockaddr_dl **sdlp)
348 {
349 	unsigned socksize, ifasize;
350 	int addrlen, namelen;
351 	struct sockaddr_dl *mask, *sdl;
352 	struct ifaddr *ifa;
353 
354 	namelen = strlen(ifp->if_xname);
355 	addrlen = ifp->if_addrlen;
356 	socksize = roundup(sockaddr_dl_measure(namelen, addrlen), sizeof(long));
357 	ifasize = sizeof(*ifa) + 2 * socksize;
358 	ifa = (struct ifaddr *)malloc(ifasize, M_IFADDR, M_WAITOK|M_ZERO);
359 
360 	sdl = (struct sockaddr_dl *)(ifa + 1);
361 	mask = (struct sockaddr_dl *)(socksize + (char *)sdl);
362 
363 	sockaddr_dl_init(sdl, socksize, ifp->if_index, ifp->if_type,
364 	    ifp->if_xname, namelen, NULL, addrlen);
365 	mask->sdl_len = sockaddr_dl_measure(namelen, 0);
366 	memset(&mask->sdl_data[0], 0xff, namelen);
367 	ifa->ifa_rtrequest = link_rtrequest;
368 	ifa->ifa_addr = (struct sockaddr *)sdl;
369 	ifa->ifa_netmask = (struct sockaddr *)mask;
370 
371 	*sdlp = sdl;
372 
373 	return ifa;
374 }
375 
376 static void
377 if_sadl_setrefs(struct ifnet *ifp, struct ifaddr *ifa)
378 {
379 	const struct sockaddr_dl *sdl;
380 	ifnet_addrs[ifp->if_index] = ifa;
381 	IFAREF(ifa);
382 	ifp->if_dl = ifa;
383 	IFAREF(ifa);
384 	sdl = satosdl(ifa->ifa_addr);
385 	ifp->if_sadl = sdl;
386 }
387 
388 /*
389  * Allocate the link level name for the specified interface.  This
390  * is an attachment helper.  It must be called after ifp->if_addrlen
391  * is initialized, which may not be the case when if_attach() is
392  * called.
393  */
394 void
395 if_alloc_sadl(struct ifnet *ifp)
396 {
397 	struct ifaddr *ifa;
398 	const struct sockaddr_dl *sdl;
399 
400 	/*
401 	 * If the interface already has a link name, release it
402 	 * now.  This is useful for interfaces that can change
403 	 * link types, and thus switch link names often.
404 	 */
405 	if (ifp->if_sadl != NULL)
406 		if_free_sadl(ifp);
407 
408 	ifa = if_dl_create(ifp, &sdl);
409 
410 	ifa_insert(ifp, ifa);
411 	if_sadl_setrefs(ifp, ifa);
412 }
413 
414 static void
415 if_deactivate_sadl(struct ifnet *ifp)
416 {
417 	struct ifaddr *ifa;
418 
419 	KASSERT(ifp->if_dl != NULL);
420 
421 	ifa = ifp->if_dl;
422 
423 	ifp->if_sadl = NULL;
424 
425 	ifnet_addrs[ifp->if_index] = NULL;
426 	IFAFREE(ifa);
427 	ifp->if_dl = NULL;
428 	IFAFREE(ifa);
429 }
430 
431 void
432 if_activate_sadl(struct ifnet *ifp, struct ifaddr *ifa,
433     const struct sockaddr_dl *sdl)
434 {
435 	int s;
436 
437 	s = splnet();
438 
439 	if_deactivate_sadl(ifp);
440 
441 	if_sadl_setrefs(ifp, ifa);
442 	IFADDR_FOREACH(ifa, ifp)
443 		rtinit(ifa, RTM_LLINFO_UPD, 0);
444 	splx(s);
445 }
446 
447 /*
448  * Free the link level name for the specified interface.  This is
449  * a detach helper.  This is called from if_detach() or from
450  * link layer type specific detach functions.
451  */
452 void
453 if_free_sadl(struct ifnet *ifp)
454 {
455 	struct ifaddr *ifa;
456 	int s;
457 
458 	ifa = ifnet_addrs[ifp->if_index];
459 	if (ifa == NULL) {
460 		KASSERT(ifp->if_sadl == NULL);
461 		KASSERT(ifp->if_dl == NULL);
462 		return;
463 	}
464 
465 	KASSERT(ifp->if_sadl != NULL);
466 	KASSERT(ifp->if_dl != NULL);
467 
468 	s = splnet();
469 	rtinit(ifa, RTM_DELETE, 0);
470 	ifa_remove(ifp, ifa);
471 	if_deactivate_sadl(ifp);
472 	if (ifp->if_hwdl == ifa) {
473 		IFAFREE(ifa);
474 		ifp->if_hwdl = NULL;
475 	}
476 	splx(s);
477 }
478 
479 /*
480  * Attach an interface to the
481  * list of "active" interfaces.
482  */
483 void
484 if_attach(struct ifnet *ifp)
485 {
486 	int indexlim = 0;
487 
488 	if (if_indexlim == 0) {
489 		TAILQ_INIT(&ifnet);
490 		if_indexlim = 8;
491 	}
492 	TAILQ_INIT(&ifp->if_addrlist);
493 	TAILQ_INSERT_TAIL(&ifnet, ifp, if_list);
494 	if (ifp->if_ioctl == NULL)
495 		ifp->if_ioctl = ifioctl_common;
496 
497 	mutex_enter(&index_gen_mtx);
498 	ifp->if_index_gen = index_gen++;
499 	mutex_exit(&index_gen_mtx);
500 
501 	ifp->if_index = if_index;
502 	if (ifindex2ifnet == NULL)
503 		if_index++;
504 	else
505 		while (ifp->if_index < if_indexlim &&
506 		    ifindex2ifnet[ifp->if_index] != NULL) {
507 			++if_index;
508 			if (if_index == 0)
509 				if_index = 1;
510 			/*
511 			 * If we hit USHRT_MAX, we skip back to 0 since
512 			 * there are a number of places where the value
513 			 * of if_index or if_index itself is compared
514 			 * to or stored in an unsigned short.  By
515 			 * jumping back, we won't botch those assignments
516 			 * or comparisons.
517 			 */
518 			else if (if_index == USHRT_MAX) {
519 				/*
520 				 * However, if we have to jump back to
521 				 * zero *twice* without finding an empty
522 				 * slot in ifindex2ifnet[], then there
523 				 * there are too many (>65535) interfaces.
524 				 */
525 				if (indexlim++)
526 					panic("too many interfaces");
527 				else
528 					if_index = 1;
529 			}
530 			ifp->if_index = if_index;
531 		}
532 
533 	/*
534 	 * We have some arrays that should be indexed by if_index.
535 	 * since if_index will grow dynamically, they should grow too.
536 	 *	struct ifadd **ifnet_addrs
537 	 *	struct ifnet **ifindex2ifnet
538 	 */
539 	if (ifnet_addrs == NULL || ifindex2ifnet == NULL ||
540 	    ifp->if_index >= if_indexlim) {
541 		size_t m, n, oldlim;
542 		void *q;
543 
544 		oldlim = if_indexlim;
545 		while (ifp->if_index >= if_indexlim)
546 			if_indexlim <<= 1;
547 
548 		/* grow ifnet_addrs */
549 		m = oldlim * sizeof(struct ifaddr *);
550 		n = if_indexlim * sizeof(struct ifaddr *);
551 		q = malloc(n, M_IFADDR, M_WAITOK|M_ZERO);
552 		if (ifnet_addrs != NULL) {
553 			memcpy(q, ifnet_addrs, m);
554 			free(ifnet_addrs, M_IFADDR);
555 		}
556 		ifnet_addrs = (struct ifaddr **)q;
557 
558 		/* grow ifindex2ifnet */
559 		m = oldlim * sizeof(struct ifnet *);
560 		n = if_indexlim * sizeof(struct ifnet *);
561 		q = malloc(n, M_IFADDR, M_WAITOK|M_ZERO);
562 		if (ifindex2ifnet != NULL) {
563 			memcpy(q, ifindex2ifnet, m);
564 			free(ifindex2ifnet, M_IFADDR);
565 		}
566 		ifindex2ifnet = (struct ifnet **)q;
567 	}
568 
569 	ifindex2ifnet[ifp->if_index] = ifp;
570 
571 	/*
572 	 * Link level name is allocated later by a separate call to
573 	 * if_alloc_sadl().
574 	 */
575 
576 	if (ifp->if_snd.ifq_maxlen == 0)
577 		ifp->if_snd.ifq_maxlen = ifqmaxlen;
578 
579 	sysctl_sndq_setup(&ifp->if_sysctl_log, ifp->if_xname, &ifp->if_snd);
580 
581 	ifp->if_broadcastaddr = 0; /* reliably crash if used uninitialized */
582 
583 	ifp->if_link_state = LINK_STATE_UNKNOWN;
584 
585 	ifp->if_capenable = 0;
586 	ifp->if_csum_flags_tx = 0;
587 	ifp->if_csum_flags_rx = 0;
588 
589 #ifdef ALTQ
590 	ifp->if_snd.altq_type = 0;
591 	ifp->if_snd.altq_disc = NULL;
592 	ifp->if_snd.altq_flags &= ALTQF_CANTCHANGE;
593 	ifp->if_snd.altq_tbr  = NULL;
594 	ifp->if_snd.altq_ifp  = ifp;
595 #endif
596 
597 #ifdef PFIL_HOOKS
598 	ifp->if_pfil.ph_type = PFIL_TYPE_IFNET;
599 	ifp->if_pfil.ph_ifnet = ifp;
600 	if (pfil_head_register(&ifp->if_pfil) != 0)
601 		printf("%s: WARNING: unable to register pfil hook\n",
602 		    ifp->if_xname);
603 	(void)pfil_run_hooks(&if_pfil,
604 	    (struct mbuf **)PFIL_IFNET_ATTACH, ifp, PFIL_IFNET);
605 #endif
606 
607 	if (!STAILQ_EMPTY(&domains))
608 		if_attachdomain1(ifp);
609 
610 	/* Announce the interface. */
611 	rt_ifannouncemsg(ifp, IFAN_ARRIVAL);
612 }
613 
614 void
615 if_attachdomain(void)
616 {
617 	struct ifnet *ifp;
618 	int s;
619 
620 	s = splnet();
621 	IFNET_FOREACH(ifp)
622 		if_attachdomain1(ifp);
623 	splx(s);
624 }
625 
626 void
627 if_attachdomain1(struct ifnet *ifp)
628 {
629 	struct domain *dp;
630 	int s;
631 
632 	s = splnet();
633 
634 	/* address family dependent data region */
635 	memset(ifp->if_afdata, 0, sizeof(ifp->if_afdata));
636 	DOMAIN_FOREACH(dp) {
637 		if (dp->dom_ifattach != NULL)
638 			ifp->if_afdata[dp->dom_family] =
639 			    (*dp->dom_ifattach)(ifp);
640 	}
641 
642 	splx(s);
643 }
644 
645 /*
646  * Deactivate an interface.  This points all of the procedure
647  * handles at error stubs.  May be called from interrupt context.
648  */
649 void
650 if_deactivate(struct ifnet *ifp)
651 {
652 	int s;
653 
654 	s = splnet();
655 
656 	ifp->if_output	 = if_nulloutput;
657 	ifp->if_input	 = if_nullinput;
658 	ifp->if_start	 = if_nullstart;
659 	ifp->if_ioctl	 = if_nullioctl;
660 	ifp->if_init	 = if_nullinit;
661 	ifp->if_stop	 = if_nullstop;
662 	ifp->if_watchdog = if_nullwatchdog;
663 	ifp->if_drain	 = if_nulldrain;
664 
665 	/* No more packets may be enqueued. */
666 	ifp->if_snd.ifq_maxlen = 0;
667 
668 	splx(s);
669 }
670 
671 void
672 if_purgeaddrs(struct ifnet *ifp, int family, void (*purgeaddr)(struct ifaddr *))
673 {
674 	struct ifaddr *ifa, *nifa;
675 
676 	for (ifa = IFADDR_FIRST(ifp); ifa != NULL; ifa = nifa) {
677 		nifa = IFADDR_NEXT(ifa);
678 		if (ifa->ifa_addr->sa_family != family)
679 			continue;
680 		(*purgeaddr)(ifa);
681 	}
682 }
683 
684 /*
685  * Detach an interface from the list of "active" interfaces,
686  * freeing any resources as we go along.
687  *
688  * NOTE: This routine must be called with a valid thread context,
689  * as it may block.
690  */
691 void
692 if_detach(struct ifnet *ifp)
693 {
694 	struct socket so;
695 	struct ifaddr *ifa;
696 #ifdef IFAREF_DEBUG
697 	struct ifaddr *last_ifa = NULL;
698 #endif
699 	struct domain *dp;
700 	const struct protosw *pr;
701 	int s, i, family, purged;
702 
703 	/*
704 	 * XXX It's kind of lame that we have to have the
705 	 * XXX socket structure...
706 	 */
707 	memset(&so, 0, sizeof(so));
708 
709 	s = splnet();
710 
711 	/*
712 	 * Do an if_down() to give protocols a chance to do something.
713 	 */
714 	if_down(ifp);
715 
716 #ifdef ALTQ
717 	if (ALTQ_IS_ENABLED(&ifp->if_snd))
718 		altq_disable(&ifp->if_snd);
719 	if (ALTQ_IS_ATTACHED(&ifp->if_snd))
720 		altq_detach(&ifp->if_snd);
721 #endif
722 
723 	sysctl_teardown(&ifp->if_sysctl_log);
724 
725 #if NCARP > 0
726 	/* Remove the interface from any carp group it is a part of.  */
727 	if (ifp->if_carp != NULL && ifp->if_type != IFT_CARP)
728 		carp_ifdetach(ifp);
729 #endif
730 
731 	/*
732 	 * Rip all the addresses off the interface.  This should make
733 	 * all of the routes go away.
734 	 *
735 	 * pr_usrreq calls can remove an arbitrary number of ifaddrs
736 	 * from the list, including our "cursor", ifa.  For safety,
737 	 * and to honor the TAILQ abstraction, I just restart the
738 	 * loop after each removal.  Note that the loop will exit
739 	 * when all of the remaining ifaddrs belong to the AF_LINK
740 	 * family.  I am counting on the historical fact that at
741 	 * least one pr_usrreq in each address domain removes at
742 	 * least one ifaddr.
743 	 */
744 again:
745 	IFADDR_FOREACH(ifa, ifp) {
746 		family = ifa->ifa_addr->sa_family;
747 #ifdef IFAREF_DEBUG
748 		printf("if_detach: ifaddr %p, family %d, refcnt %d\n",
749 		    ifa, family, ifa->ifa_refcnt);
750 		if (last_ifa != NULL && ifa == last_ifa)
751 			panic("if_detach: loop detected");
752 		last_ifa = ifa;
753 #endif
754 		if (family == AF_LINK)
755 			continue;
756 		dp = pffinddomain(family);
757 #ifdef DIAGNOSTIC
758 		if (dp == NULL)
759 			panic("if_detach: no domain for AF %d",
760 			    family);
761 #endif
762 		/*
763 		 * XXX These PURGEIF calls are redundant with the
764 		 * purge-all-families calls below, but are left in for
765 		 * now both to make a smaller change, and to avoid
766 		 * unplanned interactions with clearing of
767 		 * ifp->if_addrlist.
768 		 */
769 		purged = 0;
770 		for (pr = dp->dom_protosw;
771 		     pr < dp->dom_protoswNPROTOSW; pr++) {
772 			so.so_proto = pr;
773 			if (pr->pr_usrreq != NULL) {
774 				(void) (*pr->pr_usrreq)(&so,
775 				    PRU_PURGEIF, NULL, NULL,
776 				    (struct mbuf *) ifp, curlwp);
777 				purged = 1;
778 			}
779 		}
780 		if (purged == 0) {
781 			/*
782 			 * XXX What's really the best thing to do
783 			 * XXX here?  --thorpej@NetBSD.org
784 			 */
785 			printf("if_detach: WARNING: AF %d not purged\n",
786 			    family);
787 			ifa_remove(ifp, ifa);
788 		}
789 		goto again;
790 	}
791 
792 	if_free_sadl(ifp);
793 
794 	/* Walk the routing table looking for stragglers. */
795 	for (i = 0; i <= AF_MAX; i++)
796 		(void)rt_walktree(i, if_rt_walktree, ifp);
797 
798 	DOMAIN_FOREACH(dp) {
799 		if (dp->dom_ifdetach != NULL && ifp->if_afdata[dp->dom_family])
800 			(*dp->dom_ifdetach)(ifp,
801 			    ifp->if_afdata[dp->dom_family]);
802 
803 		/*
804 		 * One would expect multicast memberships (INET and
805 		 * INET6) on UDP sockets to be purged by the PURGEIF
806 		 * calls above, but if all addresses were removed from
807 		 * the interface prior to destruction, the calls will
808 		 * not be made (e.g. ppp, for which pppd(8) generally
809 		 * removes addresses before destroying the interface).
810 		 * Because there is no invariant that multicast
811 		 * memberships only exist for interfaces with IPv4
812 		 * addresses, we must call PURGEIF regardless of
813 		 * addresses.  (Protocols which might store ifnet
814 		 * pointers are marked with PR_PURGEIF.)
815 		 */
816 		for (pr = dp->dom_protosw; pr < dp->dom_protoswNPROTOSW; pr++) {
817 			so.so_proto = pr;
818 			if (pr->pr_usrreq != NULL && pr->pr_flags & PR_PURGEIF)
819 				(void)(*pr->pr_usrreq)(&so, PRU_PURGEIF, NULL,
820 				    NULL, (struct mbuf *)ifp, curlwp);
821 		}
822 	}
823 
824 #ifdef PFIL_HOOKS
825 	(void)pfil_run_hooks(&if_pfil,
826 	    (struct mbuf **)PFIL_IFNET_DETACH, ifp, PFIL_IFNET);
827 	(void)pfil_head_unregister(&ifp->if_pfil);
828 #endif
829 
830 	/* Announce that the interface is gone. */
831 	rt_ifannouncemsg(ifp, IFAN_DEPARTURE);
832 
833 	ifindex2ifnet[ifp->if_index] = NULL;
834 
835 	TAILQ_REMOVE(&ifnet, ifp, if_list);
836 
837 	/*
838 	 * remove packets that came from ifp, from software interrupt queues.
839 	 */
840 	DOMAIN_FOREACH(dp) {
841 		for (i = 0; i < __arraycount(dp->dom_ifqueues); i++) {
842 			if (dp->dom_ifqueues[i] == NULL)
843 				break;
844 			if_detach_queues(ifp, dp->dom_ifqueues[i]);
845 		}
846 	}
847 
848 	splx(s);
849 }
850 
851 static void
852 if_detach_queues(struct ifnet *ifp, struct ifqueue *q)
853 {
854 	struct mbuf *m, *prev, *next;
855 
856 	prev = NULL;
857 	for (m = q->ifq_head; m != NULL; m = next) {
858 		next = m->m_nextpkt;
859 #ifdef DIAGNOSTIC
860 		if ((m->m_flags & M_PKTHDR) == 0) {
861 			prev = m;
862 			continue;
863 		}
864 #endif
865 		if (m->m_pkthdr.rcvif != ifp) {
866 			prev = m;
867 			continue;
868 		}
869 
870 		if (prev != NULL)
871 			prev->m_nextpkt = m->m_nextpkt;
872 		else
873 			q->ifq_head = m->m_nextpkt;
874 		if (q->ifq_tail == m)
875 			q->ifq_tail = prev;
876 		q->ifq_len--;
877 
878 		m->m_nextpkt = NULL;
879 		m_freem(m);
880 		IF_DROP(q);
881 	}
882 }
883 
884 /*
885  * Callback for a radix tree walk to delete all references to an
886  * ifnet.
887  */
888 static int
889 if_rt_walktree(struct rtentry *rt, void *v)
890 {
891 	struct ifnet *ifp = (struct ifnet *)v;
892 	int error;
893 
894 	if (rt->rt_ifp != ifp)
895 		return 0;
896 
897 	/* Delete the entry. */
898 	++rt->rt_refcnt;
899 	error = rtrequest(RTM_DELETE, rt_getkey(rt), rt->rt_gateway,
900 	    rt_mask(rt), rt->rt_flags, NULL);
901 	KASSERT((rt->rt_flags & RTF_UP) == 0);
902 	rt->rt_ifp = NULL;
903 	RTFREE(rt);
904 	if (error != 0)
905 		printf("%s: warning: unable to delete rtentry @ %p, "
906 		    "error = %d\n", ifp->if_xname, rt, error);
907 	return 0;
908 }
909 
910 /*
911  * Create a clone network interface.
912  */
913 int
914 if_clone_create(const char *name)
915 {
916 	struct if_clone *ifc;
917 	int unit;
918 
919 	ifc = if_clone_lookup(name, &unit);
920 	if (ifc == NULL)
921 		return EINVAL;
922 
923 	if (ifunit(name) != NULL)
924 		return EEXIST;
925 
926 	return (*ifc->ifc_create)(ifc, unit);
927 }
928 
929 /*
930  * Destroy a clone network interface.
931  */
932 int
933 if_clone_destroy(const char *name)
934 {
935 	struct if_clone *ifc;
936 	struct ifnet *ifp;
937 
938 	ifc = if_clone_lookup(name, NULL);
939 	if (ifc == NULL)
940 		return EINVAL;
941 
942 	ifp = ifunit(name);
943 	if (ifp == NULL)
944 		return ENXIO;
945 
946 	if (ifc->ifc_destroy == NULL)
947 		return EOPNOTSUPP;
948 
949 	return (*ifc->ifc_destroy)(ifp);
950 }
951 
952 /*
953  * Look up a network interface cloner.
954  */
955 static struct if_clone *
956 if_clone_lookup(const char *name, int *unitp)
957 {
958 	struct if_clone *ifc;
959 	const char *cp;
960 	int unit;
961 
962 	/* separate interface name from unit */
963 	for (cp = name;
964 	    cp - name < IFNAMSIZ && *cp && (*cp < '0' || *cp > '9');
965 	    cp++)
966 		continue;
967 
968 	if (cp == name || cp - name == IFNAMSIZ || !*cp)
969 		return NULL;	/* No name or unit number */
970 
971 	LIST_FOREACH(ifc, &if_cloners, ifc_list) {
972 		if (strlen(ifc->ifc_name) == cp - name &&
973 		    strncmp(name, ifc->ifc_name, cp - name) == 0)
974 			break;
975 	}
976 
977 	if (ifc == NULL)
978 		return NULL;
979 
980 	unit = 0;
981 	while (cp - name < IFNAMSIZ && *cp) {
982 		if (*cp < '0' || *cp > '9' || unit > INT_MAX / 10) {
983 			/* Bogus unit number. */
984 			return NULL;
985 		}
986 		unit = (unit * 10) + (*cp++ - '0');
987 	}
988 
989 	if (unitp != NULL)
990 		*unitp = unit;
991 	return ifc;
992 }
993 
994 /*
995  * Register a network interface cloner.
996  */
997 void
998 if_clone_attach(struct if_clone *ifc)
999 {
1000 
1001 	LIST_INSERT_HEAD(&if_cloners, ifc, ifc_list);
1002 	if_cloners_count++;
1003 }
1004 
1005 /*
1006  * Unregister a network interface cloner.
1007  */
1008 void
1009 if_clone_detach(struct if_clone *ifc)
1010 {
1011 
1012 	LIST_REMOVE(ifc, ifc_list);
1013 	if_cloners_count--;
1014 }
1015 
1016 /*
1017  * Provide list of interface cloners to userspace.
1018  */
1019 static int
1020 if_clone_list(struct if_clonereq *ifcr)
1021 {
1022 	char outbuf[IFNAMSIZ], *dst;
1023 	struct if_clone *ifc;
1024 	int count, error = 0;
1025 
1026 	ifcr->ifcr_total = if_cloners_count;
1027 	if ((dst = ifcr->ifcr_buffer) == NULL) {
1028 		/* Just asking how many there are. */
1029 		return 0;
1030 	}
1031 
1032 	if (ifcr->ifcr_count < 0)
1033 		return EINVAL;
1034 
1035 	count = (if_cloners_count < ifcr->ifcr_count) ?
1036 	    if_cloners_count : ifcr->ifcr_count;
1037 
1038 	for (ifc = LIST_FIRST(&if_cloners); ifc != NULL && count != 0;
1039 	     ifc = LIST_NEXT(ifc, ifc_list), count--, dst += IFNAMSIZ) {
1040 		(void)strncpy(outbuf, ifc->ifc_name, sizeof(outbuf));
1041 		if (outbuf[sizeof(outbuf) - 1] != '\0')
1042 			return ENAMETOOLONG;
1043 		error = copyout(outbuf, dst, sizeof(outbuf));
1044 		if (error != 0)
1045 			break;
1046 	}
1047 
1048 	return error;
1049 }
1050 
1051 void
1052 ifa_insert(struct ifnet *ifp, struct ifaddr *ifa)
1053 {
1054 	ifa->ifa_ifp = ifp;
1055 	TAILQ_INSERT_TAIL(&ifp->if_addrlist, ifa, ifa_list);
1056 	IFAREF(ifa);
1057 }
1058 
1059 void
1060 ifa_remove(struct ifnet *ifp, struct ifaddr *ifa)
1061 {
1062 	KASSERT(ifa->ifa_ifp == ifp);
1063 	TAILQ_REMOVE(&ifp->if_addrlist, ifa, ifa_list);
1064 	IFAFREE(ifa);
1065 }
1066 
1067 static inline int
1068 equal(const struct sockaddr *sa1, const struct sockaddr *sa2)
1069 {
1070 	return sockaddr_cmp(sa1, sa2) == 0;
1071 }
1072 
1073 /*
1074  * Locate an interface based on a complete address.
1075  */
1076 /*ARGSUSED*/
1077 struct ifaddr *
1078 ifa_ifwithaddr(const struct sockaddr *addr)
1079 {
1080 	struct ifnet *ifp;
1081 	struct ifaddr *ifa;
1082 
1083 	IFNET_FOREACH(ifp) {
1084 		if (ifp->if_output == if_nulloutput)
1085 			continue;
1086 		IFADDR_FOREACH(ifa, ifp) {
1087 			if (ifa->ifa_addr->sa_family != addr->sa_family)
1088 				continue;
1089 			if (equal(addr, ifa->ifa_addr))
1090 				return ifa;
1091 			if ((ifp->if_flags & IFF_BROADCAST) &&
1092 			    ifa->ifa_broadaddr &&
1093 			    /* IP6 doesn't have broadcast */
1094 			    ifa->ifa_broadaddr->sa_len != 0 &&
1095 			    equal(ifa->ifa_broadaddr, addr))
1096 				return ifa;
1097 		}
1098 	}
1099 	return NULL;
1100 }
1101 
1102 /*
1103  * Locate the point to point interface with a given destination address.
1104  */
1105 /*ARGSUSED*/
1106 struct ifaddr *
1107 ifa_ifwithdstaddr(const struct sockaddr *addr)
1108 {
1109 	struct ifnet *ifp;
1110 	struct ifaddr *ifa;
1111 
1112 	IFNET_FOREACH(ifp) {
1113 		if (ifp->if_output == if_nulloutput)
1114 			continue;
1115 		if ((ifp->if_flags & IFF_POINTOPOINT) == 0)
1116 			continue;
1117 		IFADDR_FOREACH(ifa, ifp) {
1118 			if (ifa->ifa_addr->sa_family != addr->sa_family ||
1119 			    ifa->ifa_dstaddr == NULL)
1120 				continue;
1121 			if (equal(addr, ifa->ifa_dstaddr))
1122 				return ifa;
1123 		}
1124 	}
1125 	return NULL;
1126 }
1127 
1128 /*
1129  * Find an interface on a specific network.  If many, choice
1130  * is most specific found.
1131  */
1132 struct ifaddr *
1133 ifa_ifwithnet(const struct sockaddr *addr)
1134 {
1135 	struct ifnet *ifp;
1136 	struct ifaddr *ifa;
1137 	const struct sockaddr_dl *sdl;
1138 	struct ifaddr *ifa_maybe = 0;
1139 	u_int af = addr->sa_family;
1140 	const char *addr_data = addr->sa_data, *cplim;
1141 
1142 	if (af == AF_LINK) {
1143 		sdl = satocsdl(addr);
1144 		if (sdl->sdl_index && sdl->sdl_index < if_indexlim &&
1145 		    ifindex2ifnet[sdl->sdl_index] &&
1146 		    ifindex2ifnet[sdl->sdl_index]->if_output != if_nulloutput)
1147 			return ifnet_addrs[sdl->sdl_index];
1148 	}
1149 #ifdef NETATALK
1150 	if (af == AF_APPLETALK) {
1151 		const struct sockaddr_at *sat, *sat2;
1152 		sat = (const struct sockaddr_at *)addr;
1153 		IFNET_FOREACH(ifp) {
1154 			if (ifp->if_output == if_nulloutput)
1155 				continue;
1156 			ifa = at_ifawithnet((const struct sockaddr_at *)addr, ifp);
1157 			if (ifa == NULL)
1158 				continue;
1159 			sat2 = (struct sockaddr_at *)ifa->ifa_addr;
1160 			if (sat2->sat_addr.s_net == sat->sat_addr.s_net)
1161 				return ifa; /* exact match */
1162 			if (ifa_maybe == NULL) {
1163 				/* else keep the if with the right range */
1164 				ifa_maybe = ifa;
1165 			}
1166 		}
1167 		return ifa_maybe;
1168 	}
1169 #endif
1170 	IFNET_FOREACH(ifp) {
1171 		if (ifp->if_output == if_nulloutput)
1172 			continue;
1173 		IFADDR_FOREACH(ifa, ifp) {
1174 			const char *cp, *cp2, *cp3;
1175 
1176 			if (ifa->ifa_addr->sa_family != af ||
1177 			    ifa->ifa_netmask == NULL)
1178  next:				continue;
1179 			cp = addr_data;
1180 			cp2 = ifa->ifa_addr->sa_data;
1181 			cp3 = ifa->ifa_netmask->sa_data;
1182 			cplim = (const char *)ifa->ifa_netmask +
1183 			    ifa->ifa_netmask->sa_len;
1184 			while (cp3 < cplim) {
1185 				if ((*cp++ ^ *cp2++) & *cp3++) {
1186 					/* want to continue for() loop */
1187 					goto next;
1188 				}
1189 			}
1190 			if (ifa_maybe == NULL ||
1191 			    rn_refines((void *)ifa->ifa_netmask,
1192 			    (void *)ifa_maybe->ifa_netmask))
1193 				ifa_maybe = ifa;
1194 		}
1195 	}
1196 	return ifa_maybe;
1197 }
1198 
1199 /*
1200  * Find the interface of the addresss.
1201  */
1202 struct ifaddr *
1203 ifa_ifwithladdr(const struct sockaddr *addr)
1204 {
1205 	struct ifaddr *ia;
1206 
1207 	if ((ia = ifa_ifwithaddr(addr)) || (ia = ifa_ifwithdstaddr(addr)) ||
1208 	    (ia = ifa_ifwithnet(addr)))
1209 		return ia;
1210 	return NULL;
1211 }
1212 
1213 /*
1214  * Find an interface using a specific address family
1215  */
1216 struct ifaddr *
1217 ifa_ifwithaf(int af)
1218 {
1219 	struct ifnet *ifp;
1220 	struct ifaddr *ifa;
1221 
1222 	IFNET_FOREACH(ifp) {
1223 		if (ifp->if_output == if_nulloutput)
1224 			continue;
1225 		IFADDR_FOREACH(ifa, ifp) {
1226 			if (ifa->ifa_addr->sa_family == af)
1227 				return ifa;
1228 		}
1229 	}
1230 	return NULL;
1231 }
1232 
1233 /*
1234  * Find an interface address specific to an interface best matching
1235  * a given address.
1236  */
1237 struct ifaddr *
1238 ifaof_ifpforaddr(const struct sockaddr *addr, struct ifnet *ifp)
1239 {
1240 	struct ifaddr *ifa;
1241 	const char *cp, *cp2, *cp3;
1242 	const char *cplim;
1243 	struct ifaddr *ifa_maybe = 0;
1244 	u_int af = addr->sa_family;
1245 
1246 	if (ifp->if_output == if_nulloutput)
1247 		return NULL;
1248 
1249 	if (af >= AF_MAX)
1250 		return NULL;
1251 
1252 	IFADDR_FOREACH(ifa, ifp) {
1253 		if (ifa->ifa_addr->sa_family != af)
1254 			continue;
1255 		ifa_maybe = ifa;
1256 		if (ifa->ifa_netmask == NULL) {
1257 			if (equal(addr, ifa->ifa_addr) ||
1258 			    (ifa->ifa_dstaddr &&
1259 			     equal(addr, ifa->ifa_dstaddr)))
1260 				return ifa;
1261 			continue;
1262 		}
1263 		cp = addr->sa_data;
1264 		cp2 = ifa->ifa_addr->sa_data;
1265 		cp3 = ifa->ifa_netmask->sa_data;
1266 		cplim = ifa->ifa_netmask->sa_len + (char *)ifa->ifa_netmask;
1267 		for (; cp3 < cplim; cp3++) {
1268 			if ((*cp++ ^ *cp2++) & *cp3)
1269 				break;
1270 		}
1271 		if (cp3 == cplim)
1272 			return ifa;
1273 	}
1274 	return ifa_maybe;
1275 }
1276 
1277 /*
1278  * Default action when installing a route with a Link Level gateway.
1279  * Lookup an appropriate real ifa to point to.
1280  * This should be moved to /sys/net/link.c eventually.
1281  */
1282 void
1283 link_rtrequest(int cmd, struct rtentry *rt, const struct rt_addrinfo *info)
1284 {
1285 	struct ifaddr *ifa;
1286 	const struct sockaddr *dst;
1287 	struct ifnet *ifp;
1288 
1289 	if (cmd != RTM_ADD || (ifa = rt->rt_ifa) == NULL ||
1290 	    (ifp = ifa->ifa_ifp) == NULL || (dst = rt_getkey(rt)) == NULL)
1291 		return;
1292 	if ((ifa = ifaof_ifpforaddr(dst, ifp)) != NULL) {
1293 		rt_replace_ifa(rt, ifa);
1294 		if (ifa->ifa_rtrequest && ifa->ifa_rtrequest != link_rtrequest)
1295 			ifa->ifa_rtrequest(cmd, rt, info);
1296 	}
1297 }
1298 
1299 /*
1300  * Handle a change in the interface link state.
1301  */
1302 void
1303 if_link_state_change(struct ifnet *ifp, int link_state)
1304 {
1305 	if (ifp->if_link_state == link_state)
1306 		return;
1307 	ifp->if_link_state = link_state;
1308 	/* Notify that the link state has changed. */
1309 	rt_ifmsg(ifp);
1310 #if NCARP > 0
1311 	if (ifp->if_carp)
1312 		carp_carpdev_state(ifp);
1313 #endif
1314 }
1315 
1316 /*
1317  * Mark an interface down and notify protocols of
1318  * the transition.
1319  * NOTE: must be called at splsoftnet or equivalent.
1320  */
1321 void
1322 if_down(struct ifnet *ifp)
1323 {
1324 	struct ifaddr *ifa;
1325 
1326 	ifp->if_flags &= ~IFF_UP;
1327 	nanotime(&ifp->if_lastchange);
1328 	IFADDR_FOREACH(ifa, ifp)
1329 		pfctlinput(PRC_IFDOWN, ifa->ifa_addr);
1330 	IFQ_PURGE(&ifp->if_snd);
1331 #if NCARP > 0
1332 	if (ifp->if_carp)
1333 		carp_carpdev_state(ifp);
1334 #endif
1335 	rt_ifmsg(ifp);
1336 }
1337 
1338 /*
1339  * Mark an interface up and notify protocols of
1340  * the transition.
1341  * NOTE: must be called at splsoftnet or equivalent.
1342  */
1343 void
1344 if_up(struct ifnet *ifp)
1345 {
1346 #ifdef notyet
1347 	struct ifaddr *ifa;
1348 #endif
1349 
1350 	ifp->if_flags |= IFF_UP;
1351 	nanotime(&ifp->if_lastchange);
1352 #ifdef notyet
1353 	/* this has no effect on IP, and will kill all ISO connections XXX */
1354 	IFADDR_FOREACH(ifa, ifp)
1355 		pfctlinput(PRC_IFUP, ifa->ifa_addr);
1356 #endif
1357 #if NCARP > 0
1358 	if (ifp->if_carp)
1359 		carp_carpdev_state(ifp);
1360 #endif
1361 	rt_ifmsg(ifp);
1362 #ifdef INET6
1363 	in6_if_up(ifp);
1364 #endif
1365 }
1366 
1367 /*
1368  * Handle interface watchdog timer routines.  Called
1369  * from softclock, we decrement timers (if set) and
1370  * call the appropriate interface routine on expiration.
1371  */
1372 void
1373 if_slowtimo(void *arg)
1374 {
1375 	struct ifnet *ifp;
1376 	int s = splnet();
1377 
1378 	IFNET_FOREACH(ifp) {
1379 		if (ifp->if_timer == 0 || --ifp->if_timer)
1380 			continue;
1381 		if (ifp->if_watchdog != NULL)
1382 			(*ifp->if_watchdog)(ifp);
1383 	}
1384 	splx(s);
1385 	callout_reset(&if_slowtimo_ch, hz / IFNET_SLOWHZ, if_slowtimo, NULL);
1386 }
1387 
1388 /*
1389  * Set/clear promiscuous mode on interface ifp based on the truth value
1390  * of pswitch.  The calls are reference counted so that only the first
1391  * "on" request actually has an effect, as does the final "off" request.
1392  * Results are undefined if the "off" and "on" requests are not matched.
1393  */
1394 int
1395 ifpromisc(struct ifnet *ifp, int pswitch)
1396 {
1397 	int pcount, ret;
1398 	short flags;
1399 	struct ifreq ifr;
1400 
1401 	pcount = ifp->if_pcount;
1402 	flags = ifp->if_flags;
1403 	if (pswitch) {
1404 		/*
1405 		 * Allow the device to be "placed" into promiscuous
1406 		 * mode even if it is not configured up.  It will
1407 		 * consult IFF_PROMISC when it is brought up.
1408 		 */
1409 		if (ifp->if_pcount++ != 0)
1410 			return 0;
1411 		ifp->if_flags |= IFF_PROMISC;
1412 		if ((ifp->if_flags & IFF_UP) == 0)
1413 			return 0;
1414 	} else {
1415 		if (--ifp->if_pcount > 0)
1416 			return 0;
1417 		ifp->if_flags &= ~IFF_PROMISC;
1418 		/*
1419 		 * If the device is not configured up, we should not need to
1420 		 * turn off promiscuous mode (device should have turned it
1421 		 * off when interface went down; and will look at IFF_PROMISC
1422 		 * again next time interface comes up).
1423 		 */
1424 		if ((ifp->if_flags & IFF_UP) == 0)
1425 			return 0;
1426 	}
1427 	memset(&ifr, 0, sizeof(ifr));
1428 	ifr.ifr_flags = ifp->if_flags;
1429 	ret = (*ifp->if_ioctl)(ifp, SIOCSIFFLAGS, &ifr);
1430 	/* Restore interface state if not successful. */
1431 	if (ret != 0) {
1432 		ifp->if_pcount = pcount;
1433 		ifp->if_flags = flags;
1434 	}
1435 	return ret;
1436 }
1437 
1438 /*
1439  * Map interface name to
1440  * interface structure pointer.
1441  */
1442 struct ifnet *
1443 ifunit(const char *name)
1444 {
1445 	struct ifnet *ifp;
1446 	const char *cp = name;
1447 	u_int unit = 0;
1448 	u_int i;
1449 
1450 	/*
1451 	 * If the entire name is a number, treat it as an ifindex.
1452 	 */
1453 	for (i = 0; i < IFNAMSIZ && *cp >= '0' && *cp <= '9'; i++, cp++) {
1454 		unit = unit * 10 + (*cp - '0');
1455 	}
1456 
1457 	/*
1458 	 * If the number took all of the name, then it's a valid ifindex.
1459 	 */
1460 	if (i == IFNAMSIZ || (cp != name && *cp == '\0')) {
1461 		if (unit >= if_indexlim)
1462 			return NULL;
1463 		ifp = ifindex2ifnet[unit];
1464 		if (ifp == NULL || ifp->if_output == if_nulloutput)
1465 			return NULL;
1466 		return ifp;
1467 	}
1468 
1469 	IFNET_FOREACH(ifp) {
1470 		if (ifp->if_output == if_nulloutput)
1471 			continue;
1472 	 	if (strcmp(ifp->if_xname, name) == 0)
1473 			return ifp;
1474 	}
1475 	return NULL;
1476 }
1477 
1478 /* common */
1479 int
1480 ifioctl_common(struct ifnet *ifp, u_long cmd, void *data)
1481 {
1482 	int s;
1483 	struct ifreq *ifr;
1484 	struct ifcapreq *ifcr;
1485 	struct ifdatareq *ifdr;
1486 
1487 	switch (cmd) {
1488 	case SIOCSIFCAP:
1489 		ifcr = data;
1490 		if ((ifcr->ifcr_capenable & ~ifp->if_capabilities) != 0)
1491 			return EINVAL;
1492 
1493 		if (ifcr->ifcr_capenable == ifp->if_capenable)
1494 			return 0;
1495 
1496 		ifp->if_capenable = ifcr->ifcr_capenable;
1497 
1498 		/* Pre-compute the checksum flags mask. */
1499 		ifp->if_csum_flags_tx = 0;
1500 		ifp->if_csum_flags_rx = 0;
1501 		if (ifp->if_capenable & IFCAP_CSUM_IPv4_Tx) {
1502 			ifp->if_csum_flags_tx |= M_CSUM_IPv4;
1503 		}
1504 		if (ifp->if_capenable & IFCAP_CSUM_IPv4_Rx) {
1505 			ifp->if_csum_flags_rx |= M_CSUM_IPv4;
1506 		}
1507 
1508 		if (ifp->if_capenable & IFCAP_CSUM_TCPv4_Tx) {
1509 			ifp->if_csum_flags_tx |= M_CSUM_TCPv4;
1510 		}
1511 		if (ifp->if_capenable & IFCAP_CSUM_TCPv4_Rx) {
1512 			ifp->if_csum_flags_rx |= M_CSUM_TCPv4;
1513 		}
1514 
1515 		if (ifp->if_capenable & IFCAP_CSUM_UDPv4_Tx) {
1516 			ifp->if_csum_flags_tx |= M_CSUM_UDPv4;
1517 		}
1518 		if (ifp->if_capenable & IFCAP_CSUM_UDPv4_Rx) {
1519 			ifp->if_csum_flags_rx |= M_CSUM_UDPv4;
1520 		}
1521 
1522 		if (ifp->if_capenable & IFCAP_CSUM_TCPv6_Tx) {
1523 			ifp->if_csum_flags_tx |= M_CSUM_TCPv6;
1524 		}
1525 		if (ifp->if_capenable & IFCAP_CSUM_TCPv6_Rx) {
1526 			ifp->if_csum_flags_rx |= M_CSUM_TCPv6;
1527 		}
1528 
1529 		if (ifp->if_capenable & IFCAP_CSUM_UDPv6_Tx) {
1530 			ifp->if_csum_flags_tx |= M_CSUM_UDPv6;
1531 		}
1532 		if (ifp->if_capenable & IFCAP_CSUM_UDPv6_Rx) {
1533 			ifp->if_csum_flags_rx |= M_CSUM_UDPv6;
1534 		}
1535 		if (ifp->if_flags & IFF_UP)
1536 			return ENETRESET;
1537 		return 0;
1538 	case SIOCSIFFLAGS:
1539 		ifr = data;
1540 		if (ifp->if_flags & IFF_UP && (ifr->ifr_flags & IFF_UP) == 0) {
1541 			s = splnet();
1542 			if_down(ifp);
1543 			splx(s);
1544 		}
1545 		if (ifr->ifr_flags & IFF_UP && (ifp->if_flags & IFF_UP) == 0) {
1546 			s = splnet();
1547 			if_up(ifp);
1548 			splx(s);
1549 		}
1550 		ifp->if_flags = (ifp->if_flags & IFF_CANTCHANGE) |
1551 			(ifr->ifr_flags &~ IFF_CANTCHANGE);
1552 		break;
1553 	case SIOCGIFFLAGS:
1554 		ifr = data;
1555 		ifr->ifr_flags = ifp->if_flags;
1556 		break;
1557 
1558 	case SIOCGIFMETRIC:
1559 		ifr = data;
1560 		ifr->ifr_metric = ifp->if_metric;
1561 		break;
1562 
1563 	case SIOCGIFMTU:
1564 		ifr = data;
1565 		ifr->ifr_mtu = ifp->if_mtu;
1566 		break;
1567 
1568 	case SIOCGIFDLT:
1569 		ifr = data;
1570 		ifr->ifr_dlt = ifp->if_dlt;
1571 		break;
1572 
1573 	case SIOCGIFCAP:
1574 		ifcr = data;
1575 		ifcr->ifcr_capabilities = ifp->if_capabilities;
1576 		ifcr->ifcr_capenable = ifp->if_capenable;
1577 		break;
1578 
1579 	case SIOCSIFMETRIC:
1580 		ifr = data;
1581 		ifp->if_metric = ifr->ifr_metric;
1582 		break;
1583 
1584 	case SIOCGIFDATA:
1585 		ifdr = data;
1586 		ifdr->ifdr_data = ifp->if_data;
1587 		break;
1588 
1589 	case SIOCZIFDATA:
1590 		ifdr = data;
1591 		ifdr->ifdr_data = ifp->if_data;
1592 		/*
1593 		 * Assumes that the volatile counters that can be
1594 		 * zero'ed are at the end of if_data.
1595 		 */
1596 		memset(&ifp->if_data.ifi_ipackets, 0, sizeof(ifp->if_data) -
1597 		    offsetof(struct if_data, ifi_ipackets));
1598 		break;
1599 	case SIOCSIFMTU:
1600 		ifr = data;
1601 		if (ifp->if_mtu == ifr->ifr_mtu)
1602 			break;
1603 		ifp->if_mtu = ifr->ifr_mtu;
1604 		/*
1605 		 * If the link MTU changed, do network layer specific procedure.
1606 		 */
1607 #ifdef INET6
1608 		nd6_setmtu(ifp);
1609 #endif
1610 		return ENETRESET;
1611 	default:
1612 		return ENOTTY;
1613 	}
1614 	return 0;
1615 }
1616 
1617 int
1618 ifaddrpref_ioctl(struct socket *so, u_long cmd, void *data, struct ifnet *ifp,
1619     lwp_t *l)
1620 {
1621 	struct if_addrprefreq *ifap = (struct if_addrprefreq *)data;
1622 	struct ifaddr *ifa;
1623 	const struct sockaddr *any, *sa;
1624 	union {
1625 		struct sockaddr sa;
1626 		struct sockaddr_storage ss;
1627 	} u, v;
1628 
1629 	switch (cmd) {
1630 	case SIOCSIFADDRPREF:
1631 		if (kauth_authorize_network(l->l_cred, KAUTH_NETWORK_INTERFACE,
1632 		    KAUTH_REQ_NETWORK_INTERFACE_SETPRIV, ifp, (void *)cmd,
1633 		    NULL) != 0)
1634 			return EPERM;
1635 	case SIOCGIFADDRPREF:
1636 		break;
1637 	default:
1638 		return EOPNOTSUPP;
1639 	}
1640 
1641 	/* sanity checks */
1642 	if (data == NULL || ifp == NULL) {
1643 		panic("invalid argument to %s", __func__);
1644 		/*NOTREACHED*/
1645 	}
1646 
1647 	/* address must be specified on ADD and DELETE */
1648 	sa = sstocsa(&ifap->ifap_addr);
1649 	if (sa->sa_family != sofamily(so))
1650 		return EINVAL;
1651 	if ((any = sockaddr_any(sa)) == NULL || sa->sa_len != any->sa_len)
1652 		return EINVAL;
1653 
1654 	sockaddr_externalize(&v.sa, sizeof(v.ss), sa);
1655 
1656 	IFADDR_FOREACH(ifa, ifp) {
1657 		if (ifa->ifa_addr->sa_family != sa->sa_family)
1658 			continue;
1659 		sockaddr_externalize(&u.sa, sizeof(u.ss), ifa->ifa_addr);
1660 		if (sockaddr_cmp(&u.sa, &v.sa) == 0)
1661 			break;
1662 	}
1663 	if (ifa == NULL)
1664 		return EADDRNOTAVAIL;
1665 
1666 	switch (cmd) {
1667 	case SIOCSIFADDRPREF:
1668 		ifa->ifa_preference = ifap->ifap_preference;
1669 		return 0;
1670 	case SIOCGIFADDRPREF:
1671 		/* fill in the if_laddrreq structure */
1672 		(void)sockaddr_copy(sstosa(&ifap->ifap_addr),
1673 		    sizeof(ifap->ifap_addr), ifa->ifa_addr);
1674 		ifap->ifap_preference = ifa->ifa_preference;
1675 		return 0;
1676 	default:
1677 		return EOPNOTSUPP;
1678 	}
1679 }
1680 
1681 /*
1682  * Interface ioctls.
1683  */
1684 int
1685 ifioctl(struct socket *so, u_long cmd, void *data, struct lwp *l)
1686 {
1687 	struct ifnet *ifp;
1688 	struct ifreq *ifr;
1689 	struct ifcapreq *ifcr;
1690 	struct ifdatareq *ifdr;
1691 	int error = 0;
1692 #if defined(COMPAT_OSOCK) || defined(COMPAT_OIFREQ)
1693 	u_long ocmd = cmd;
1694 #endif
1695 	short oif_flags;
1696 #ifdef COMPAT_OIFREQ
1697 	struct ifreq ifrb;
1698 	struct oifreq *oifr = NULL;
1699 #endif
1700 
1701 	switch (cmd) {
1702 #ifdef COMPAT_OIFREQ
1703 	case OSIOCGIFCONF:
1704 	case OOSIOCGIFCONF:
1705 		return compat_ifconf(cmd, data);
1706 #endif
1707 #ifdef COMPAT_OIFDATA
1708 	case OSIOCGIFDATA:
1709 	case OSIOCZIFDATA:
1710 		return compat_ifdatareq(l, cmd, data);
1711 #endif
1712 	case SIOCGIFCONF:
1713 		return ifconf(cmd, data);
1714 	case SIOCINITIFADDR:
1715 		return EPERM;
1716 	}
1717 
1718 #ifdef COMPAT_OIFREQ
1719 	cmd = compat_cvtcmd(cmd);
1720 	if (cmd != ocmd) {
1721 		oifr = data;
1722 		data = ifr = &ifrb;
1723 		ifreqo2n(oifr, ifr);
1724 	} else
1725 #endif
1726 		ifr = data;
1727 	ifcr = data;
1728 	ifdr = data;
1729 
1730 	ifp = ifunit(ifr->ifr_name);
1731 
1732 	switch (cmd) {
1733 	case SIOCIFCREATE:
1734 	case SIOCIFDESTROY:
1735 		if (l != NULL) {
1736 			error = kauth_authorize_network(l->l_cred,
1737 			    KAUTH_NETWORK_INTERFACE,
1738 			    KAUTH_REQ_NETWORK_INTERFACE_SETPRIV, ifp,
1739 			    (void *)cmd, NULL);
1740 			if (error != 0)
1741 				return error;
1742 		}
1743 		return (cmd == SIOCIFCREATE) ?
1744 			if_clone_create(ifr->ifr_name) :
1745 			if_clone_destroy(ifr->ifr_name);
1746 
1747 	case SIOCIFGCLONERS:
1748 		return if_clone_list((struct if_clonereq *)data);
1749 	}
1750 
1751 	if (ifp == NULL)
1752 		return ENXIO;
1753 
1754 	switch (cmd) {
1755 	case SIOCALIFADDR:
1756 	case SIOCDLIFADDR:
1757 	case SIOCSIFADDRPREF:
1758 	case SIOCSIFFLAGS:
1759 	case SIOCSIFCAP:
1760 	case SIOCSIFMETRIC:
1761 	case SIOCZIFDATA:
1762 	case SIOCSIFMTU:
1763 	case SIOCSIFPHYADDR:
1764 	case SIOCDIFPHYADDR:
1765 #ifdef INET6
1766 	case SIOCSIFPHYADDR_IN6:
1767 #endif
1768 	case SIOCSLIFPHYADDR:
1769 	case SIOCADDMULTI:
1770 	case SIOCDELMULTI:
1771 	case SIOCSIFMEDIA:
1772 	case SIOCSDRVSPEC:
1773 	case SIOCG80211:
1774 	case SIOCS80211:
1775 	case SIOCS80211NWID:
1776 	case SIOCS80211NWKEY:
1777 	case SIOCS80211POWER:
1778 	case SIOCS80211BSSID:
1779 	case SIOCS80211CHANNEL:
1780 		if (l != NULL) {
1781 			error = kauth_authorize_network(l->l_cred,
1782 			    KAUTH_NETWORK_INTERFACE,
1783 			    KAUTH_REQ_NETWORK_INTERFACE_SETPRIV, ifp,
1784 			    (void *)cmd, NULL);
1785 			if (error != 0)
1786 				return error;
1787 		}
1788 	}
1789 
1790 	oif_flags = ifp->if_flags;
1791 
1792 	error = (*ifp->if_ioctl)(ifp, cmd, data);
1793 	if (error != ENOTTY)
1794 		;
1795 	else if (so->so_proto == NULL)
1796 		return EOPNOTSUPP;
1797 	else {
1798 #ifdef COMPAT_OSOCK
1799 		error = compat_ifioctl(so, ocmd, cmd, data, l);
1800 #else
1801 		error = (*so->so_proto->pr_usrreq)(so, PRU_CONTROL,
1802 		    (struct mbuf *)cmd, (struct mbuf *)data,
1803 		    (struct mbuf *)ifp, l);
1804 #endif
1805 	}
1806 
1807 	if (((oif_flags ^ ifp->if_flags) & IFF_UP) != 0) {
1808 #ifdef INET6
1809 		if ((ifp->if_flags & IFF_UP) != 0) {
1810 			int s = splnet();
1811 			in6_if_up(ifp);
1812 			splx(s);
1813 		}
1814 #endif
1815 	}
1816 #ifdef COMPAT_OIFREQ
1817 	if (cmd != ocmd)
1818 		ifreqn2o(oifr, ifr);
1819 #endif
1820 
1821 	return error;
1822 }
1823 
1824 /*
1825  * Return interface configuration
1826  * of system.  List may be used
1827  * in later ioctl's (above) to get
1828  * other information.
1829  *
1830  * Each record is a struct ifreq.  Before the addition of
1831  * sockaddr_storage, the API rule was that sockaddr flavors that did
1832  * not fit would extend beyond the struct ifreq, with the next struct
1833  * ifreq starting sa_len beyond the struct sockaddr.  Because the
1834  * union in struct ifreq includes struct sockaddr_storage, every kind
1835  * of sockaddr must fit.  Thus, there are no longer any overlength
1836  * records.
1837  *
1838  * Records are added to the user buffer if they fit, and ifc_len is
1839  * adjusted to the length that was written.  Thus, the user is only
1840  * assured of getting the complete list if ifc_len on return is at
1841  * least sizeof(struct ifreq) less than it was on entry.
1842  *
1843  * If the user buffer pointer is NULL, this routine copies no data and
1844  * returns the amount of space that would be needed.
1845  *
1846  * Invariants:
1847  * ifrp points to the next part of the user's buffer to be used.  If
1848  * ifrp != NULL, space holds the number of bytes remaining that we may
1849  * write at ifrp.  Otherwise, space holds the number of bytes that
1850  * would have been written had there been adequate space.
1851  */
1852 /*ARGSUSED*/
1853 int
1854 ifconf(u_long cmd, void *data)
1855 {
1856 	struct ifconf *ifc = (struct ifconf *)data;
1857 	struct ifnet *ifp;
1858 	struct ifaddr *ifa;
1859 	struct ifreq ifr, *ifrp;
1860 	int space, error = 0;
1861 	const int sz = (int)sizeof(struct ifreq);
1862 
1863 	if ((ifrp = ifc->ifc_req) == NULL)
1864 		space = 0;
1865 	else
1866 		space = ifc->ifc_len;
1867 	IFNET_FOREACH(ifp) {
1868 		(void)strncpy(ifr.ifr_name, ifp->if_xname,
1869 		    sizeof(ifr.ifr_name));
1870 		if (ifr.ifr_name[sizeof(ifr.ifr_name) - 1] != '\0')
1871 			return ENAMETOOLONG;
1872 		if (IFADDR_EMPTY(ifp)) {
1873 			/* Interface with no addresses - send zero sockaddr. */
1874 			memset(&ifr.ifr_addr, 0, sizeof(ifr.ifr_addr));
1875 			if (ifrp == NULL) {
1876 				space += sz;
1877 				continue;
1878 			}
1879 			if (space >= sz) {
1880 				error = copyout(&ifr, ifrp, sz);
1881 				if (error != 0)
1882 					return error;
1883 				ifrp++;
1884 				space -= sz;
1885 			}
1886 		}
1887 
1888 		IFADDR_FOREACH(ifa, ifp) {
1889 			struct sockaddr *sa = ifa->ifa_addr;
1890 			/* all sockaddrs must fit in sockaddr_storage */
1891 			KASSERT(sa->sa_len <= sizeof(ifr.ifr_ifru));
1892 
1893 			if (ifrp == NULL) {
1894 				space += sz;
1895 				continue;
1896 			}
1897 			memcpy(&ifr.ifr_space, sa, sa->sa_len);
1898 			if (space >= sz) {
1899 				error = copyout(&ifr, ifrp, sz);
1900 				if (error != 0)
1901 					return (error);
1902 				ifrp++; space -= sz;
1903 			}
1904 		}
1905 	}
1906 	if (ifrp != NULL) {
1907 		KASSERT(0 <= space && space <= ifc->ifc_len);
1908 		ifc->ifc_len -= space;
1909 	} else {
1910 		KASSERT(space >= 0);
1911 		ifc->ifc_len = space;
1912 	}
1913 	return (0);
1914 }
1915 
1916 int
1917 ifreq_setaddr(const u_long cmd, struct ifreq *ifr, const struct sockaddr *sa)
1918 {
1919 	uint8_t len;
1920 	u_long ncmd;
1921 
1922 	if ((ncmd = compat_cvtcmd(cmd)) != cmd)
1923 		len = sizeof(ifr->ifr_addr);
1924 	else
1925 		len = sizeof(ifr->ifr_ifru.ifru_space);
1926 	if (len < sa->sa_len)
1927 		return EFBIG;
1928 	memset(&ifr->ifr_addr, 0, len);
1929 	sockaddr_copy(&ifr->ifr_addr, len, sa);
1930 	return 0;
1931 }
1932 
1933 /*
1934  * Queue message on interface, and start output if interface
1935  * not yet active.
1936  */
1937 int
1938 ifq_enqueue(struct ifnet *ifp, struct mbuf *m
1939     ALTQ_COMMA ALTQ_DECL(struct altq_pktattr *pktattr))
1940 {
1941 	int len = m->m_pkthdr.len;
1942 	int mflags = m->m_flags;
1943 	int s = splnet();
1944 	int error;
1945 
1946 	IFQ_ENQUEUE(&ifp->if_snd, m, pktattr, error);
1947 	if (error != 0)
1948 		goto out;
1949 	ifp->if_obytes += len;
1950 	if (mflags & M_MCAST)
1951 		ifp->if_omcasts++;
1952 	if ((ifp->if_flags & IFF_OACTIVE) == 0)
1953 		(*ifp->if_start)(ifp);
1954 out:
1955 	splx(s);
1956 	return error;
1957 }
1958 
1959 /*
1960  * Queue message on interface, possibly using a second fast queue
1961  */
1962 int
1963 ifq_enqueue2(struct ifnet *ifp, struct ifqueue *ifq, struct mbuf *m
1964     ALTQ_COMMA ALTQ_DECL(struct altq_pktattr *pktattr))
1965 {
1966 	int error = 0;
1967 
1968 	if (ifq != NULL
1969 #ifdef ALTQ
1970 	    && ALTQ_IS_ENABLED(&ifp->if_snd) == 0
1971 #endif
1972 	    ) {
1973 		if (IF_QFULL(ifq)) {
1974 			IF_DROP(&ifp->if_snd);
1975 			m_freem(m);
1976 			if (error == 0)
1977 				error = ENOBUFS;
1978 		} else
1979 			IF_ENQUEUE(ifq, m);
1980 	} else
1981 		IFQ_ENQUEUE(&ifp->if_snd, m, pktattr, error);
1982 	if (error != 0) {
1983 		++ifp->if_oerrors;
1984 		return error;
1985 	}
1986 	return 0;
1987 }
1988 
1989 
1990 static void
1991 sysctl_sndq_setup(struct sysctllog **clog, const char *ifname,
1992     struct ifaltq *ifq)
1993 {
1994 	const struct sysctlnode *cnode, *rnode;
1995 
1996 	if (sysctl_createv(clog, 0, NULL, &rnode,
1997 		       CTLFLAG_PERMANENT,
1998 		       CTLTYPE_NODE, "net", NULL,
1999 		       NULL, 0, NULL, 0,
2000 		       CTL_NET, CTL_EOL) != 0)
2001 		goto bad;
2002 
2003 	if (sysctl_createv(clog, 0, &rnode, &rnode,
2004 		       CTLFLAG_PERMANENT,
2005 		       CTLTYPE_NODE, "interfaces",
2006 		       SYSCTL_DESCR("Per-interface controls"),
2007 		       NULL, 0, NULL, 0,
2008 		       CTL_CREATE, CTL_EOL) != 0)
2009 		goto bad;
2010 
2011 	if (sysctl_createv(clog, 0, &rnode, &rnode,
2012 		       CTLFLAG_PERMANENT,
2013 		       CTLTYPE_NODE, ifname,
2014 		       SYSCTL_DESCR("Interface controls"),
2015 		       NULL, 0, NULL, 0,
2016 		       CTL_CREATE, CTL_EOL) != 0)
2017 		goto bad;
2018 
2019 	if (sysctl_createv(clog, 0, &rnode, &rnode,
2020 		       CTLFLAG_PERMANENT,
2021 		       CTLTYPE_NODE, "sndq",
2022 		       SYSCTL_DESCR("Interface output queue controls"),
2023 		       NULL, 0, NULL, 0,
2024 		       CTL_CREATE, CTL_EOL) != 0)
2025 		goto bad;
2026 
2027 	if (sysctl_createv(clog, 0, &rnode, &cnode,
2028 		       CTLFLAG_PERMANENT,
2029 		       CTLTYPE_INT, "len",
2030 		       SYSCTL_DESCR("Current output queue length"),
2031 		       NULL, 0, &ifq->ifq_len, 0,
2032 		       CTL_CREATE, CTL_EOL) != 0)
2033 		goto bad;
2034 
2035 	if (sysctl_createv(clog, 0, &rnode, &cnode,
2036 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
2037 		       CTLTYPE_INT, "maxlen",
2038 		       SYSCTL_DESCR("Maximum allowed output queue length"),
2039 		       NULL, 0, &ifq->ifq_maxlen, 0,
2040 		       CTL_CREATE, CTL_EOL) != 0)
2041 		goto bad;
2042 
2043 	if (sysctl_createv(clog, 0, &rnode, &cnode,
2044 		       CTLFLAG_PERMANENT,
2045 		       CTLTYPE_INT, "drops",
2046 		       SYSCTL_DESCR("Packets dropped due to full output queue"),
2047 		       NULL, 0, &ifq->ifq_drops, 0,
2048 		       CTL_CREATE, CTL_EOL) != 0)
2049 		goto bad;
2050 
2051 	return;
2052 bad:
2053 	printf("%s: could not attach sysctl nodes\n", ifname);
2054 	return;
2055 }
2056 
2057 #if defined(INET) || defined(INET6)
2058 static void
2059 sysctl_net_ifq_setup(struct sysctllog **clog,
2060 		     int pf, const char *pfname,
2061 		     int ipn, const char *ipname,
2062 		     int qid, struct ifqueue *ifq)
2063 {
2064 
2065 	sysctl_createv(clog, 0, NULL, NULL,
2066 		       CTLFLAG_PERMANENT,
2067 		       CTLTYPE_NODE, "net", NULL,
2068 		       NULL, 0, NULL, 0,
2069 		       CTL_NET, CTL_EOL);
2070 	sysctl_createv(clog, 0, NULL, NULL,
2071 		       CTLFLAG_PERMANENT,
2072 		       CTLTYPE_NODE, pfname, NULL,
2073 		       NULL, 0, NULL, 0,
2074 		       CTL_NET, pf, CTL_EOL);
2075 	sysctl_createv(clog, 0, NULL, NULL,
2076 		       CTLFLAG_PERMANENT,
2077 		       CTLTYPE_NODE, ipname, NULL,
2078 		       NULL, 0, NULL, 0,
2079 		       CTL_NET, pf, ipn, CTL_EOL);
2080 	sysctl_createv(clog, 0, NULL, NULL,
2081 		       CTLFLAG_PERMANENT,
2082 		       CTLTYPE_NODE, "ifq",
2083 		       SYSCTL_DESCR("Protocol input queue controls"),
2084 		       NULL, 0, NULL, 0,
2085 		       CTL_NET, pf, ipn, qid, CTL_EOL);
2086 
2087 	sysctl_createv(clog, 0, NULL, NULL,
2088 		       CTLFLAG_PERMANENT,
2089 		       CTLTYPE_INT, "len",
2090 		       SYSCTL_DESCR("Current input queue length"),
2091 		       NULL, 0, &ifq->ifq_len, 0,
2092 		       CTL_NET, pf, ipn, qid, IFQCTL_LEN, CTL_EOL);
2093 	sysctl_createv(clog, 0, NULL, NULL,
2094 		       CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
2095 		       CTLTYPE_INT, "maxlen",
2096 		       SYSCTL_DESCR("Maximum allowed input queue length"),
2097 		       NULL, 0, &ifq->ifq_maxlen, 0,
2098 		       CTL_NET, pf, ipn, qid, IFQCTL_MAXLEN, CTL_EOL);
2099 #ifdef notyet
2100 	sysctl_createv(clog, 0, NULL, NULL,
2101 		       CTLFLAG_PERMANENT,
2102 		       CTLTYPE_INT, "peak",
2103 		       SYSCTL_DESCR("Highest input queue length"),
2104 		       NULL, 0, &ifq->ifq_peak, 0,
2105 		       CTL_NET, pf, ipn, qid, IFQCTL_PEAK, CTL_EOL);
2106 #endif
2107 	sysctl_createv(clog, 0, NULL, NULL,
2108 		       CTLFLAG_PERMANENT,
2109 		       CTLTYPE_INT, "drops",
2110 		       SYSCTL_DESCR("Packets dropped due to full input queue"),
2111 		       NULL, 0, &ifq->ifq_drops, 0,
2112 		       CTL_NET, pf, ipn, qid, IFQCTL_DROPS, CTL_EOL);
2113 }
2114 #endif /* INET || INET6 */
2115