xref: /freebsd/sys/net/if_bridge.c (revision e8e8c939350bdf3c228a411caa9660c607c27a11)
1 /*	$NetBSD: if_bridge.c,v 1.31 2005/06/01 19:45:34 jdc Exp $	*/
2 
3 /*
4  * Copyright 2001 Wasabi Systems, Inc.
5  * All rights reserved.
6  *
7  * Written by Jason R. Thorpe for Wasabi Systems, Inc.
8  *
9  * Redistribution and use in source and binary forms, with or without
10  * modification, are permitted provided that the following conditions
11  * are met:
12  * 1. Redistributions of source code must retain the above copyright
13  *    notice, this list of conditions and the following disclaimer.
14  * 2. Redistributions in binary form must reproduce the above copyright
15  *    notice, this list of conditions and the following disclaimer in the
16  *    documentation and/or other materials provided with the distribution.
17  * 3. All advertising materials mentioning features or use of this software
18  *    must display the following acknowledgement:
19  *	This product includes software developed for the NetBSD Project by
20  *	Wasabi Systems, Inc.
21  * 4. The name of Wasabi Systems, Inc. may not be used to endorse
22  *    or promote products derived from this software without specific prior
23  *    written permission.
24  *
25  * THIS SOFTWARE IS PROVIDED BY WASABI SYSTEMS, INC. ``AS IS'' AND
26  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
27  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
28  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL WASABI SYSTEMS, INC
29  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
30  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
31  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
32  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
33  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
34  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
35  * POSSIBILITY OF SUCH DAMAGE.
36  */
37 
38 /*
39  * Copyright (c) 1999, 2000 Jason L. Wright (jason@thought.net)
40  * All rights reserved.
41  *
42  * Redistribution and use in source and binary forms, with or without
43  * modification, are permitted provided that the following conditions
44  * are met:
45  * 1. Redistributions of source code must retain the above copyright
46  *    notice, this list of conditions and the following disclaimer.
47  * 2. Redistributions in binary form must reproduce the above copyright
48  *    notice, this list of conditions and the following disclaimer in the
49  *    documentation and/or other materials provided with the distribution.
50  *
51  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
52  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
53  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
54  * DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT,
55  * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
56  * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
57  * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
58  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
59  * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
60  * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
61  * POSSIBILITY OF SUCH DAMAGE.
62  *
63  * OpenBSD: if_bridge.c,v 1.60 2001/06/15 03:38:33 itojun Exp
64  */
65 
66 /*
67  * Network interface bridge support.
68  *
69  * TODO:
70  *
71  *	- Currently only supports Ethernet-like interfaces (Ethernet,
72  *	  802.11, VLANs on Ethernet, etc.)  Figure out a nice way
73  *	  to bridge other types of interfaces (FDDI-FDDI, and maybe
74  *	  consider heterogenous bridges).
75  */
76 
77 #include <sys/cdefs.h>
78 __FBSDID("$FreeBSD$");
79 
80 #include "opt_inet.h"
81 #include "opt_inet6.h"
82 
83 #include <sys/param.h>
84 #include <sys/eventhandler.h>
85 #include <sys/mbuf.h>
86 #include <sys/malloc.h>
87 #include <sys/protosw.h>
88 #include <sys/systm.h>
89 #include <sys/jail.h>
90 #include <sys/time.h>
91 #include <sys/socket.h> /* for net/if.h */
92 #include <sys/sockio.h>
93 #include <sys/ctype.h>  /* string functions */
94 #include <sys/kernel.h>
95 #include <sys/random.h>
96 #include <sys/syslog.h>
97 #include <sys/sysctl.h>
98 #include <vm/uma.h>
99 #include <sys/module.h>
100 #include <sys/priv.h>
101 #include <sys/proc.h>
102 #include <sys/lock.h>
103 #include <sys/mutex.h>
104 
105 #include <net/bpf.h>
106 #include <net/if.h>
107 #include <net/if_clone.h>
108 #include <net/if_dl.h>
109 #include <net/if_types.h>
110 #include <net/if_var.h>
111 #include <net/pfil.h>
112 #include <net/vnet.h>
113 
114 #include <netinet/in.h>
115 #include <netinet/in_systm.h>
116 #include <netinet/in_var.h>
117 #include <netinet/ip.h>
118 #include <netinet/ip_var.h>
119 #ifdef INET6
120 #include <netinet/ip6.h>
121 #include <netinet6/ip6_var.h>
122 #include <netinet6/in6_ifattach.h>
123 #endif
124 #if defined(INET) || defined(INET6)
125 #include <netinet/ip_carp.h>
126 #endif
127 #include <machine/in_cksum.h>
128 #include <netinet/if_ether.h>
129 #include <net/bridgestp.h>
130 #include <net/if_bridgevar.h>
131 #include <net/if_llc.h>
132 #include <net/if_vlan_var.h>
133 
134 #include <net/route.h>
135 
136 /*
137  * Size of the route hash table.  Must be a power of two.
138  */
139 #ifndef BRIDGE_RTHASH_SIZE
140 #define	BRIDGE_RTHASH_SIZE		1024
141 #endif
142 
143 #define	BRIDGE_RTHASH_MASK		(BRIDGE_RTHASH_SIZE - 1)
144 
145 /*
146  * Default maximum number of addresses to cache.
147  */
148 #ifndef BRIDGE_RTABLE_MAX
149 #define	BRIDGE_RTABLE_MAX		2000
150 #endif
151 
152 /*
153  * Timeout (in seconds) for entries learned dynamically.
154  */
155 #ifndef BRIDGE_RTABLE_TIMEOUT
156 #define	BRIDGE_RTABLE_TIMEOUT		(20 * 60)	/* same as ARP */
157 #endif
158 
159 /*
160  * Number of seconds between walks of the route list.
161  */
162 #ifndef BRIDGE_RTABLE_PRUNE_PERIOD
163 #define	BRIDGE_RTABLE_PRUNE_PERIOD	(5 * 60)
164 #endif
165 
166 /*
167  * List of capabilities to possibly mask on the member interface.
168  */
169 #define	BRIDGE_IFCAPS_MASK		(IFCAP_TOE|IFCAP_TSO|IFCAP_TXCSUM)
170 
171 /*
172  * List of capabilities to strip
173  */
174 #define	BRIDGE_IFCAPS_STRIP		IFCAP_LRO
175 
176 /*
177  * Bridge interface list entry.
178  */
179 struct bridge_iflist {
180 	LIST_ENTRY(bridge_iflist) bif_next;
181 	struct ifnet		*bif_ifp;	/* member if */
182 	struct bstp_port	bif_stp;	/* STP state */
183 	uint32_t		bif_flags;	/* member if flags */
184 	int			bif_savedcaps;	/* saved capabilities */
185 	uint32_t		bif_addrmax;	/* max # of addresses */
186 	uint32_t		bif_addrcnt;	/* cur. # of addresses */
187 	uint32_t		bif_addrexceeded;/* # of address violations */
188 };
189 
190 /*
191  * Bridge route node.
192  */
193 struct bridge_rtnode {
194 	LIST_ENTRY(bridge_rtnode) brt_hash;	/* hash table linkage */
195 	LIST_ENTRY(bridge_rtnode) brt_list;	/* list linkage */
196 	struct bridge_iflist	*brt_dst;	/* destination if */
197 	unsigned long		brt_expire;	/* expiration time */
198 	uint8_t			brt_flags;	/* address flags */
199 	uint8_t			brt_addr[ETHER_ADDR_LEN];
200 	uint16_t		brt_vlan;	/* vlan id */
201 };
202 #define	brt_ifp			brt_dst->bif_ifp
203 
204 /*
205  * Software state for each bridge.
206  */
207 struct bridge_softc {
208 	struct ifnet		*sc_ifp;	/* make this an interface */
209 	LIST_ENTRY(bridge_softc) sc_list;
210 	struct mtx		sc_mtx;
211 	struct cv		sc_cv;
212 	uint32_t		sc_brtmax;	/* max # of addresses */
213 	uint32_t		sc_brtcnt;	/* cur. # of addresses */
214 	uint32_t		sc_brttimeout;	/* rt timeout in seconds */
215 	struct callout		sc_brcallout;	/* bridge callout */
216 	uint32_t		sc_iflist_ref;	/* refcount for sc_iflist */
217 	uint32_t		sc_iflist_xcnt;	/* refcount for sc_iflist */
218 	LIST_HEAD(, bridge_iflist) sc_iflist;	/* member interface list */
219 	LIST_HEAD(, bridge_rtnode) *sc_rthash;	/* our forwarding table */
220 	LIST_HEAD(, bridge_rtnode) sc_rtlist;	/* list version of above */
221 	uint32_t		sc_rthash_key;	/* key for hash */
222 	LIST_HEAD(, bridge_iflist) sc_spanlist;	/* span ports list */
223 	struct bstp_state	sc_stp;		/* STP state */
224 	uint32_t		sc_brtexceeded;	/* # of cache drops */
225 	struct ifnet		*sc_ifaddr;	/* member mac copied from */
226 	u_char			sc_defaddr[6];	/* Default MAC address */
227 };
228 
229 static VNET_DEFINE(struct mtx, bridge_list_mtx);
230 #define	V_bridge_list_mtx	VNET(bridge_list_mtx)
231 static eventhandler_tag bridge_detach_cookie;
232 
233 int	bridge_rtable_prune_period = BRIDGE_RTABLE_PRUNE_PERIOD;
234 
235 uma_zone_t bridge_rtnode_zone;
236 
237 static int	bridge_clone_create(struct if_clone *, int, caddr_t);
238 static void	bridge_clone_destroy(struct ifnet *);
239 
240 static int	bridge_ioctl(struct ifnet *, u_long, caddr_t);
241 static void	bridge_mutecaps(struct bridge_softc *);
242 static void	bridge_set_ifcap(struct bridge_softc *, struct bridge_iflist *,
243 		    int);
244 static void	bridge_ifdetach(void *arg __unused, struct ifnet *);
245 static void	bridge_init(void *);
246 static void	bridge_dummynet(struct mbuf *, struct ifnet *);
247 static void	bridge_stop(struct ifnet *, int);
248 static int	bridge_transmit(struct ifnet *, struct mbuf *);
249 static void	bridge_qflush(struct ifnet *);
250 static struct mbuf *bridge_input(struct ifnet *, struct mbuf *);
251 static int	bridge_output(struct ifnet *, struct mbuf *, struct sockaddr *,
252 		    struct rtentry *);
253 static int	bridge_enqueue(struct bridge_softc *, struct ifnet *,
254 		    struct mbuf *);
255 static void	bridge_rtdelete(struct bridge_softc *, struct ifnet *ifp, int);
256 
257 static void	bridge_forward(struct bridge_softc *, struct bridge_iflist *,
258 		    struct mbuf *m);
259 
260 static void	bridge_timer(void *);
261 
262 static void	bridge_broadcast(struct bridge_softc *, struct ifnet *,
263 		    struct mbuf *, int);
264 static void	bridge_span(struct bridge_softc *, struct mbuf *);
265 
266 static int	bridge_rtupdate(struct bridge_softc *, const uint8_t *,
267 		    uint16_t, struct bridge_iflist *, int, uint8_t);
268 static struct ifnet *bridge_rtlookup(struct bridge_softc *, const uint8_t *,
269 		    uint16_t);
270 static void	bridge_rttrim(struct bridge_softc *);
271 static void	bridge_rtage(struct bridge_softc *);
272 static void	bridge_rtflush(struct bridge_softc *, int);
273 static int	bridge_rtdaddr(struct bridge_softc *, const uint8_t *,
274 		    uint16_t);
275 
276 static void	bridge_rtable_init(struct bridge_softc *);
277 static void	bridge_rtable_fini(struct bridge_softc *);
278 
279 static int	bridge_rtnode_addr_cmp(const uint8_t *, const uint8_t *);
280 static struct bridge_rtnode *bridge_rtnode_lookup(struct bridge_softc *,
281 		    const uint8_t *, uint16_t);
282 static int	bridge_rtnode_insert(struct bridge_softc *,
283 		    struct bridge_rtnode *);
284 static void	bridge_rtnode_destroy(struct bridge_softc *,
285 		    struct bridge_rtnode *);
286 static void	bridge_rtable_expire(struct ifnet *, int);
287 static void	bridge_state_change(struct ifnet *, int);
288 
289 static struct bridge_iflist *bridge_lookup_member(struct bridge_softc *,
290 		    const char *name);
291 static struct bridge_iflist *bridge_lookup_member_if(struct bridge_softc *,
292 		    struct ifnet *ifp);
293 static void	bridge_delete_member(struct bridge_softc *,
294 		    struct bridge_iflist *, int);
295 static void	bridge_delete_span(struct bridge_softc *,
296 		    struct bridge_iflist *);
297 
298 static int	bridge_ioctl_add(struct bridge_softc *, void *);
299 static int	bridge_ioctl_del(struct bridge_softc *, void *);
300 static int	bridge_ioctl_gifflags(struct bridge_softc *, void *);
301 static int	bridge_ioctl_sifflags(struct bridge_softc *, void *);
302 static int	bridge_ioctl_scache(struct bridge_softc *, void *);
303 static int	bridge_ioctl_gcache(struct bridge_softc *, void *);
304 static int	bridge_ioctl_gifs(struct bridge_softc *, void *);
305 static int	bridge_ioctl_rts(struct bridge_softc *, void *);
306 static int	bridge_ioctl_saddr(struct bridge_softc *, void *);
307 static int	bridge_ioctl_sto(struct bridge_softc *, void *);
308 static int	bridge_ioctl_gto(struct bridge_softc *, void *);
309 static int	bridge_ioctl_daddr(struct bridge_softc *, void *);
310 static int	bridge_ioctl_flush(struct bridge_softc *, void *);
311 static int	bridge_ioctl_gpri(struct bridge_softc *, void *);
312 static int	bridge_ioctl_spri(struct bridge_softc *, void *);
313 static int	bridge_ioctl_ght(struct bridge_softc *, void *);
314 static int	bridge_ioctl_sht(struct bridge_softc *, void *);
315 static int	bridge_ioctl_gfd(struct bridge_softc *, void *);
316 static int	bridge_ioctl_sfd(struct bridge_softc *, void *);
317 static int	bridge_ioctl_gma(struct bridge_softc *, void *);
318 static int	bridge_ioctl_sma(struct bridge_softc *, void *);
319 static int	bridge_ioctl_sifprio(struct bridge_softc *, void *);
320 static int	bridge_ioctl_sifcost(struct bridge_softc *, void *);
321 static int	bridge_ioctl_sifmaxaddr(struct bridge_softc *, void *);
322 static int	bridge_ioctl_addspan(struct bridge_softc *, void *);
323 static int	bridge_ioctl_delspan(struct bridge_softc *, void *);
324 static int	bridge_ioctl_gbparam(struct bridge_softc *, void *);
325 static int	bridge_ioctl_grte(struct bridge_softc *, void *);
326 static int	bridge_ioctl_gifsstp(struct bridge_softc *, void *);
327 static int	bridge_ioctl_sproto(struct bridge_softc *, void *);
328 static int	bridge_ioctl_stxhc(struct bridge_softc *, void *);
329 static int	bridge_pfil(struct mbuf **, struct ifnet *, struct ifnet *,
330 		    int);
331 static int	bridge_ip_checkbasic(struct mbuf **mp);
332 #ifdef INET6
333 static int	bridge_ip6_checkbasic(struct mbuf **mp);
334 #endif /* INET6 */
335 static int	bridge_fragment(struct ifnet *, struct mbuf *,
336 		    struct ether_header *, int, struct llc *);
337 static void	bridge_linkstate(struct ifnet *ifp);
338 static void	bridge_linkcheck(struct bridge_softc *sc);
339 
340 extern void (*bridge_linkstate_p)(struct ifnet *ifp);
341 
342 /* The default bridge vlan is 1 (IEEE 802.1Q-2003 Table 9-2) */
343 #define	VLANTAGOF(_m)	\
344     (_m->m_flags & M_VLANTAG) ? EVL_VLANOFTAG(_m->m_pkthdr.ether_vtag) : 1
345 
346 static struct bstp_cb_ops bridge_ops = {
347 	.bcb_state = bridge_state_change,
348 	.bcb_rtage = bridge_rtable_expire
349 };
350 
351 SYSCTL_DECL(_net_link);
352 static SYSCTL_NODE(_net_link, IFT_BRIDGE, bridge, CTLFLAG_RW, 0, "Bridge");
353 
354 /* only pass IP[46] packets when pfil is enabled */
355 static VNET_DEFINE(int, pfil_onlyip) = 1;
356 #define	V_pfil_onlyip	VNET(pfil_onlyip)
357 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_onlyip,
358     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_onlyip), 0,
359     "Only pass IP packets when pfil is enabled");
360 
361 /* run pfil hooks on the bridge interface */
362 static VNET_DEFINE(int, pfil_bridge) = 1;
363 #define	V_pfil_bridge	VNET(pfil_bridge)
364 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_bridge,
365     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_bridge), 0,
366     "Packet filter on the bridge interface");
367 
368 /* layer2 filter with ipfw */
369 static VNET_DEFINE(int, pfil_ipfw);
370 #define	V_pfil_ipfw	VNET(pfil_ipfw)
371 
372 /* layer2 ARP filter with ipfw */
373 static VNET_DEFINE(int, pfil_ipfw_arp);
374 #define	V_pfil_ipfw_arp	VNET(pfil_ipfw_arp)
375 SYSCTL_INT(_net_link_bridge, OID_AUTO, ipfw_arp,
376     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_ipfw_arp), 0,
377     "Filter ARP packets through IPFW layer2");
378 
379 /* run pfil hooks on the member interface */
380 static VNET_DEFINE(int, pfil_member) = 1;
381 #define	V_pfil_member	VNET(pfil_member)
382 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_member,
383     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_member), 0,
384     "Packet filter on the member interface");
385 
386 /* run pfil hooks on the physical interface for locally destined packets */
387 static VNET_DEFINE(int, pfil_local_phys);
388 #define	V_pfil_local_phys	VNET(pfil_local_phys)
389 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_local_phys,
390     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_local_phys), 0,
391     "Packet filter on the physical interface for locally destined packets");
392 
393 /* log STP state changes */
394 static VNET_DEFINE(int, log_stp);
395 #define	V_log_stp	VNET(log_stp)
396 SYSCTL_INT(_net_link_bridge, OID_AUTO, log_stp,
397     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(log_stp), 0,
398     "Log STP state changes");
399 
400 /* share MAC with first bridge member */
401 static VNET_DEFINE(int, bridge_inherit_mac);
402 #define	V_bridge_inherit_mac	VNET(bridge_inherit_mac)
403 SYSCTL_INT(_net_link_bridge, OID_AUTO, inherit_mac,
404     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(bridge_inherit_mac), 0,
405     "Inherit MAC address from the first bridge member");
406 
407 static VNET_DEFINE(int, allow_llz_overlap) = 0;
408 #define	V_allow_llz_overlap	VNET(allow_llz_overlap)
409 SYSCTL_INT(_net_link_bridge, OID_AUTO, allow_llz_overlap,
410     CTLFLAG_VNET | CTLFLAG_RW | CTLFLAG_VNET, &VNET_NAME(allow_llz_overlap), 0,
411     "Allow overlap of link-local scope "
412     "zones of a bridge interface and the member interfaces");
413 
414 struct bridge_control {
415 	int	(*bc_func)(struct bridge_softc *, void *);
416 	int	bc_argsize;
417 	int	bc_flags;
418 };
419 
420 #define	BC_F_COPYIN		0x01	/* copy arguments in */
421 #define	BC_F_COPYOUT		0x02	/* copy arguments out */
422 #define	BC_F_SUSER		0x04	/* do super-user check */
423 
424 const struct bridge_control bridge_control_table[] = {
425 	{ bridge_ioctl_add,		sizeof(struct ifbreq),
426 	  BC_F_COPYIN|BC_F_SUSER },
427 	{ bridge_ioctl_del,		sizeof(struct ifbreq),
428 	  BC_F_COPYIN|BC_F_SUSER },
429 
430 	{ bridge_ioctl_gifflags,	sizeof(struct ifbreq),
431 	  BC_F_COPYIN|BC_F_COPYOUT },
432 	{ bridge_ioctl_sifflags,	sizeof(struct ifbreq),
433 	  BC_F_COPYIN|BC_F_SUSER },
434 
435 	{ bridge_ioctl_scache,		sizeof(struct ifbrparam),
436 	  BC_F_COPYIN|BC_F_SUSER },
437 	{ bridge_ioctl_gcache,		sizeof(struct ifbrparam),
438 	  BC_F_COPYOUT },
439 
440 	{ bridge_ioctl_gifs,		sizeof(struct ifbifconf),
441 	  BC_F_COPYIN|BC_F_COPYOUT },
442 	{ bridge_ioctl_rts,		sizeof(struct ifbaconf),
443 	  BC_F_COPYIN|BC_F_COPYOUT },
444 
445 	{ bridge_ioctl_saddr,		sizeof(struct ifbareq),
446 	  BC_F_COPYIN|BC_F_SUSER },
447 
448 	{ bridge_ioctl_sto,		sizeof(struct ifbrparam),
449 	  BC_F_COPYIN|BC_F_SUSER },
450 	{ bridge_ioctl_gto,		sizeof(struct ifbrparam),
451 	  BC_F_COPYOUT },
452 
453 	{ bridge_ioctl_daddr,		sizeof(struct ifbareq),
454 	  BC_F_COPYIN|BC_F_SUSER },
455 
456 	{ bridge_ioctl_flush,		sizeof(struct ifbreq),
457 	  BC_F_COPYIN|BC_F_SUSER },
458 
459 	{ bridge_ioctl_gpri,		sizeof(struct ifbrparam),
460 	  BC_F_COPYOUT },
461 	{ bridge_ioctl_spri,		sizeof(struct ifbrparam),
462 	  BC_F_COPYIN|BC_F_SUSER },
463 
464 	{ bridge_ioctl_ght,		sizeof(struct ifbrparam),
465 	  BC_F_COPYOUT },
466 	{ bridge_ioctl_sht,		sizeof(struct ifbrparam),
467 	  BC_F_COPYIN|BC_F_SUSER },
468 
469 	{ bridge_ioctl_gfd,		sizeof(struct ifbrparam),
470 	  BC_F_COPYOUT },
471 	{ bridge_ioctl_sfd,		sizeof(struct ifbrparam),
472 	  BC_F_COPYIN|BC_F_SUSER },
473 
474 	{ bridge_ioctl_gma,		sizeof(struct ifbrparam),
475 	  BC_F_COPYOUT },
476 	{ bridge_ioctl_sma,		sizeof(struct ifbrparam),
477 	  BC_F_COPYIN|BC_F_SUSER },
478 
479 	{ bridge_ioctl_sifprio,		sizeof(struct ifbreq),
480 	  BC_F_COPYIN|BC_F_SUSER },
481 
482 	{ bridge_ioctl_sifcost,		sizeof(struct ifbreq),
483 	  BC_F_COPYIN|BC_F_SUSER },
484 
485 	{ bridge_ioctl_addspan,		sizeof(struct ifbreq),
486 	  BC_F_COPYIN|BC_F_SUSER },
487 	{ bridge_ioctl_delspan,		sizeof(struct ifbreq),
488 	  BC_F_COPYIN|BC_F_SUSER },
489 
490 	{ bridge_ioctl_gbparam,		sizeof(struct ifbropreq),
491 	  BC_F_COPYOUT },
492 
493 	{ bridge_ioctl_grte,		sizeof(struct ifbrparam),
494 	  BC_F_COPYOUT },
495 
496 	{ bridge_ioctl_gifsstp,		sizeof(struct ifbpstpconf),
497 	  BC_F_COPYIN|BC_F_COPYOUT },
498 
499 	{ bridge_ioctl_sproto,		sizeof(struct ifbrparam),
500 	  BC_F_COPYIN|BC_F_SUSER },
501 
502 	{ bridge_ioctl_stxhc,		sizeof(struct ifbrparam),
503 	  BC_F_COPYIN|BC_F_SUSER },
504 
505 	{ bridge_ioctl_sifmaxaddr,	sizeof(struct ifbreq),
506 	  BC_F_COPYIN|BC_F_SUSER },
507 
508 };
509 const int bridge_control_table_size = nitems(bridge_control_table);
510 
511 static VNET_DEFINE(LIST_HEAD(, bridge_softc), bridge_list);
512 #define	V_bridge_list	VNET(bridge_list)
513 #define	BRIDGE_LIST_LOCK_INIT(x)	mtx_init(&V_bridge_list_mtx,	\
514 					    "if_bridge list", NULL, MTX_DEF)
515 #define	BRIDGE_LIST_LOCK_DESTROY(x)	mtx_destroy(&V_bridge_list_mtx)
516 #define	BRIDGE_LIST_LOCK(x)		mtx_lock(&V_bridge_list_mtx)
517 #define	BRIDGE_LIST_UNLOCK(x)		mtx_unlock(&V_bridge_list_mtx)
518 
519 static VNET_DEFINE(struct if_clone *, bridge_cloner);
520 #define	V_bridge_cloner	VNET(bridge_cloner)
521 
522 static const char bridge_name[] = "bridge";
523 
524 static void
525 vnet_bridge_init(const void *unused __unused)
526 {
527 
528 	BRIDGE_LIST_LOCK_INIT();
529 	LIST_INIT(&V_bridge_list);
530 	V_bridge_cloner = if_clone_simple(bridge_name,
531 	    bridge_clone_create, bridge_clone_destroy, 0);
532 }
533 VNET_SYSINIT(vnet_bridge_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
534     vnet_bridge_init, NULL);
535 
536 static void
537 vnet_bridge_uninit(const void *unused __unused)
538 {
539 
540 	if_clone_detach(V_bridge_cloner);
541 	V_bridge_cloner = NULL;
542 	BRIDGE_LIST_LOCK_DESTROY();
543 }
544 VNET_SYSUNINIT(vnet_bridge_uninit, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
545     vnet_bridge_uninit, NULL);
546 
547 static int
548 bridge_modevent(module_t mod, int type, void *data)
549 {
550 
551 	switch (type) {
552 	case MOD_LOAD:
553 		bridge_rtnode_zone = uma_zcreate("bridge_rtnode",
554 		    sizeof(struct bridge_rtnode), NULL, NULL, NULL, NULL,
555 		    UMA_ALIGN_PTR, 0);
556 		bridge_input_p = bridge_input;
557 		bridge_output_p = bridge_output;
558 		bridge_dn_p = bridge_dummynet;
559 		bridge_linkstate_p = bridge_linkstate;
560 		bridge_detach_cookie = EVENTHANDLER_REGISTER(
561 		    ifnet_departure_event, bridge_ifdetach, NULL,
562 		    EVENTHANDLER_PRI_ANY);
563 		break;
564 	case MOD_UNLOAD:
565 		EVENTHANDLER_DEREGISTER(ifnet_departure_event,
566 		    bridge_detach_cookie);
567 		uma_zdestroy(bridge_rtnode_zone);
568 		bridge_input_p = NULL;
569 		bridge_output_p = NULL;
570 		bridge_dn_p = NULL;
571 		bridge_linkstate_p = NULL;
572 		break;
573 	default:
574 		return (EOPNOTSUPP);
575 	}
576 	return (0);
577 }
578 
579 static moduledata_t bridge_mod = {
580 	"if_bridge",
581 	bridge_modevent,
582 	0
583 };
584 
585 DECLARE_MODULE(if_bridge, bridge_mod, SI_SUB_PSEUDO, SI_ORDER_ANY);
586 MODULE_DEPEND(if_bridge, bridgestp, 1, 1, 1);
587 
588 /*
589  * handler for net.link.bridge.ipfw
590  */
591 static int
592 sysctl_pfil_ipfw(SYSCTL_HANDLER_ARGS)
593 {
594 	int enable = V_pfil_ipfw;
595 	int error;
596 
597 	error = sysctl_handle_int(oidp, &enable, 0, req);
598 	enable &= 1;
599 
600 	if (enable != V_pfil_ipfw) {
601 		V_pfil_ipfw = enable;
602 
603 		/*
604 		 * Disable pfil so that ipfw doesnt run twice, if the user
605 		 * really wants both then they can re-enable pfil_bridge and/or
606 		 * pfil_member. Also allow non-ip packets as ipfw can filter by
607 		 * layer2 type.
608 		 */
609 		if (V_pfil_ipfw) {
610 			V_pfil_onlyip = 0;
611 			V_pfil_bridge = 0;
612 			V_pfil_member = 0;
613 		}
614 	}
615 
616 	return (error);
617 }
618 SYSCTL_PROC(_net_link_bridge, OID_AUTO, ipfw,
619     CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_VNET,
620     &VNET_NAME(pfil_ipfw), 0, &sysctl_pfil_ipfw, "I",
621     "Layer2 filter with IPFW");
622 
623 /*
624  * bridge_clone_create:
625  *
626  *	Create a new bridge instance.
627  */
628 static int
629 bridge_clone_create(struct if_clone *ifc, int unit, caddr_t params)
630 {
631 	struct bridge_softc *sc, *sc2;
632 	struct ifnet *bifp, *ifp;
633 	int fb, retry;
634 	unsigned long hostid;
635 
636 	sc = malloc(sizeof(*sc), M_DEVBUF, M_WAITOK|M_ZERO);
637 	ifp = sc->sc_ifp = if_alloc(IFT_ETHER);
638 	if (ifp == NULL) {
639 		free(sc, M_DEVBUF);
640 		return (ENOSPC);
641 	}
642 
643 	BRIDGE_LOCK_INIT(sc);
644 	sc->sc_brtmax = BRIDGE_RTABLE_MAX;
645 	sc->sc_brttimeout = BRIDGE_RTABLE_TIMEOUT;
646 
647 	/* Initialize our routing table. */
648 	bridge_rtable_init(sc);
649 
650 	callout_init_mtx(&sc->sc_brcallout, &sc->sc_mtx, 0);
651 
652 	LIST_INIT(&sc->sc_iflist);
653 	LIST_INIT(&sc->sc_spanlist);
654 
655 	ifp->if_softc = sc;
656 	if_initname(ifp, bridge_name, unit);
657 	ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
658 	ifp->if_ioctl = bridge_ioctl;
659 	ifp->if_transmit = bridge_transmit;
660 	ifp->if_qflush = bridge_qflush;
661 	ifp->if_init = bridge_init;
662 	ifp->if_type = IFT_BRIDGE;
663 
664 	/*
665 	 * Generate an ethernet address with a locally administered address.
666 	 *
667 	 * Since we are using random ethernet addresses for the bridge, it is
668 	 * possible that we might have address collisions, so make sure that
669 	 * this hardware address isn't already in use on another bridge.
670 	 * The first try uses the hostid and falls back to arc4rand().
671 	 */
672 	fb = 0;
673 	getcredhostid(curthread->td_ucred, &hostid);
674 	do {
675 		if (fb || hostid == 0) {
676 			arc4rand(sc->sc_defaddr, ETHER_ADDR_LEN, 1);
677 			sc->sc_defaddr[0] &= ~1;/* clear multicast bit */
678 			sc->sc_defaddr[0] |= 2;	/* set the LAA bit */
679 		} else {
680 			sc->sc_defaddr[0] = 0x2;
681 			sc->sc_defaddr[1] = (hostid >> 24) & 0xff;
682 			sc->sc_defaddr[2] = (hostid >> 16) & 0xff;
683 			sc->sc_defaddr[3] = (hostid >> 8 ) & 0xff;
684 			sc->sc_defaddr[4] =  hostid        & 0xff;
685 			sc->sc_defaddr[5] = ifp->if_dunit & 0xff;
686 		}
687 
688 		fb = 1;
689 		retry = 0;
690 		BRIDGE_LIST_LOCK();
691 		LIST_FOREACH(sc2, &V_bridge_list, sc_list) {
692 			bifp = sc2->sc_ifp;
693 			if (memcmp(sc->sc_defaddr,
694 			    IF_LLADDR(bifp), ETHER_ADDR_LEN) == 0) {
695 				retry = 1;
696 				break;
697 			}
698 		}
699 		BRIDGE_LIST_UNLOCK();
700 	} while (retry == 1);
701 
702 	bstp_attach(&sc->sc_stp, &bridge_ops);
703 	ether_ifattach(ifp, sc->sc_defaddr);
704 	/* Now undo some of the damage... */
705 	ifp->if_baudrate = 0;
706 	ifp->if_type = IFT_BRIDGE;
707 
708 	BRIDGE_LIST_LOCK();
709 	LIST_INSERT_HEAD(&V_bridge_list, sc, sc_list);
710 	BRIDGE_LIST_UNLOCK();
711 
712 	return (0);
713 }
714 
715 /*
716  * bridge_clone_destroy:
717  *
718  *	Destroy a bridge instance.
719  */
720 static void
721 bridge_clone_destroy(struct ifnet *ifp)
722 {
723 	struct bridge_softc *sc = ifp->if_softc;
724 	struct bridge_iflist *bif;
725 
726 	BRIDGE_LOCK(sc);
727 
728 	bridge_stop(ifp, 1);
729 	ifp->if_flags &= ~IFF_UP;
730 
731 	while ((bif = LIST_FIRST(&sc->sc_iflist)) != NULL)
732 		bridge_delete_member(sc, bif, 0);
733 
734 	while ((bif = LIST_FIRST(&sc->sc_spanlist)) != NULL) {
735 		bridge_delete_span(sc, bif);
736 	}
737 
738 	BRIDGE_UNLOCK(sc);
739 
740 	callout_drain(&sc->sc_brcallout);
741 
742 	BRIDGE_LIST_LOCK();
743 	LIST_REMOVE(sc, sc_list);
744 	BRIDGE_LIST_UNLOCK();
745 
746 	bstp_detach(&sc->sc_stp);
747 	ether_ifdetach(ifp);
748 	if_free(ifp);
749 
750 	/* Tear down the routing table. */
751 	bridge_rtable_fini(sc);
752 
753 	BRIDGE_LOCK_DESTROY(sc);
754 	free(sc, M_DEVBUF);
755 }
756 
757 /*
758  * bridge_ioctl:
759  *
760  *	Handle a control request from the operator.
761  */
762 static int
763 bridge_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
764 {
765 	struct bridge_softc *sc = ifp->if_softc;
766 	struct ifreq *ifr = (struct ifreq *)data;
767 	struct bridge_iflist *bif;
768 	struct thread *td = curthread;
769 	union {
770 		struct ifbreq ifbreq;
771 		struct ifbifconf ifbifconf;
772 		struct ifbareq ifbareq;
773 		struct ifbaconf ifbaconf;
774 		struct ifbrparam ifbrparam;
775 		struct ifbropreq ifbropreq;
776 	} args;
777 	struct ifdrv *ifd = (struct ifdrv *) data;
778 	const struct bridge_control *bc;
779 	int error = 0;
780 
781 	switch (cmd) {
782 
783 	case SIOCADDMULTI:
784 	case SIOCDELMULTI:
785 		break;
786 
787 	case SIOCGDRVSPEC:
788 	case SIOCSDRVSPEC:
789 		if (ifd->ifd_cmd >= bridge_control_table_size) {
790 			error = EINVAL;
791 			break;
792 		}
793 		bc = &bridge_control_table[ifd->ifd_cmd];
794 
795 		if (cmd == SIOCGDRVSPEC &&
796 		    (bc->bc_flags & BC_F_COPYOUT) == 0) {
797 			error = EINVAL;
798 			break;
799 		}
800 		else if (cmd == SIOCSDRVSPEC &&
801 		    (bc->bc_flags & BC_F_COPYOUT) != 0) {
802 			error = EINVAL;
803 			break;
804 		}
805 
806 		if (bc->bc_flags & BC_F_SUSER) {
807 			error = priv_check(td, PRIV_NET_BRIDGE);
808 			if (error)
809 				break;
810 		}
811 
812 		if (ifd->ifd_len != bc->bc_argsize ||
813 		    ifd->ifd_len > sizeof(args)) {
814 			error = EINVAL;
815 			break;
816 		}
817 
818 		bzero(&args, sizeof(args));
819 		if (bc->bc_flags & BC_F_COPYIN) {
820 			error = copyin(ifd->ifd_data, &args, ifd->ifd_len);
821 			if (error)
822 				break;
823 		}
824 
825 		BRIDGE_LOCK(sc);
826 		error = (*bc->bc_func)(sc, &args);
827 		BRIDGE_UNLOCK(sc);
828 		if (error)
829 			break;
830 
831 		if (bc->bc_flags & BC_F_COPYOUT)
832 			error = copyout(&args, ifd->ifd_data, ifd->ifd_len);
833 
834 		break;
835 
836 	case SIOCSIFFLAGS:
837 		if (!(ifp->if_flags & IFF_UP) &&
838 		    (ifp->if_drv_flags & IFF_DRV_RUNNING)) {
839 			/*
840 			 * If interface is marked down and it is running,
841 			 * then stop and disable it.
842 			 */
843 			BRIDGE_LOCK(sc);
844 			bridge_stop(ifp, 1);
845 			BRIDGE_UNLOCK(sc);
846 		} else if ((ifp->if_flags & IFF_UP) &&
847 		    !(ifp->if_drv_flags & IFF_DRV_RUNNING)) {
848 			/*
849 			 * If interface is marked up and it is stopped, then
850 			 * start it.
851 			 */
852 			(*ifp->if_init)(sc);
853 		}
854 		break;
855 
856 	case SIOCSIFMTU:
857 		if (ifr->ifr_mtu < 576) {
858 			error = EINVAL;
859 			break;
860 		}
861 		if (LIST_EMPTY(&sc->sc_iflist)) {
862 			sc->sc_ifp->if_mtu = ifr->ifr_mtu;
863 			break;
864 		}
865 		BRIDGE_LOCK(sc);
866 		LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
867 			if (bif->bif_ifp->if_mtu != ifr->ifr_mtu) {
868 				log(LOG_NOTICE, "%s: invalid MTU: %u(%s)"
869 				    " != %d\n", sc->sc_ifp->if_xname,
870 				    bif->bif_ifp->if_mtu,
871 				    bif->bif_ifp->if_xname, ifr->ifr_mtu);
872 				error = EINVAL;
873 				break;
874 			}
875 		}
876 		if (!error)
877 			sc->sc_ifp->if_mtu = ifr->ifr_mtu;
878 		BRIDGE_UNLOCK(sc);
879 		break;
880 	default:
881 		/*
882 		 * drop the lock as ether_ioctl() will call bridge_start() and
883 		 * cause the lock to be recursed.
884 		 */
885 		error = ether_ioctl(ifp, cmd, data);
886 		break;
887 	}
888 
889 	return (error);
890 }
891 
892 /*
893  * bridge_mutecaps:
894  *
895  *	Clear or restore unwanted capabilities on the member interface
896  */
897 static void
898 bridge_mutecaps(struct bridge_softc *sc)
899 {
900 	struct bridge_iflist *bif;
901 	int enabled, mask;
902 
903 	/* Initial bitmask of capabilities to test */
904 	mask = BRIDGE_IFCAPS_MASK;
905 
906 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
907 		/* Every member must support it or its disabled */
908 		mask &= bif->bif_savedcaps;
909 	}
910 
911 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
912 		enabled = bif->bif_ifp->if_capenable;
913 		enabled &= ~BRIDGE_IFCAPS_STRIP;
914 		/* strip off mask bits and enable them again if allowed */
915 		enabled &= ~BRIDGE_IFCAPS_MASK;
916 		enabled |= mask;
917 		bridge_set_ifcap(sc, bif, enabled);
918 	}
919 
920 }
921 
922 static void
923 bridge_set_ifcap(struct bridge_softc *sc, struct bridge_iflist *bif, int set)
924 {
925 	struct ifnet *ifp = bif->bif_ifp;
926 	struct ifreq ifr;
927 	int error;
928 
929 	bzero(&ifr, sizeof(ifr));
930 	ifr.ifr_reqcap = set;
931 
932 	if (ifp->if_capenable != set) {
933 		error = (*ifp->if_ioctl)(ifp, SIOCSIFCAP, (caddr_t)&ifr);
934 		if (error)
935 			if_printf(sc->sc_ifp,
936 			    "error setting interface capabilities on %s\n",
937 			    ifp->if_xname);
938 	}
939 }
940 
941 /*
942  * bridge_lookup_member:
943  *
944  *	Lookup a bridge member interface.
945  */
946 static struct bridge_iflist *
947 bridge_lookup_member(struct bridge_softc *sc, const char *name)
948 {
949 	struct bridge_iflist *bif;
950 	struct ifnet *ifp;
951 
952 	BRIDGE_LOCK_ASSERT(sc);
953 
954 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
955 		ifp = bif->bif_ifp;
956 		if (strcmp(ifp->if_xname, name) == 0)
957 			return (bif);
958 	}
959 
960 	return (NULL);
961 }
962 
963 /*
964  * bridge_lookup_member_if:
965  *
966  *	Lookup a bridge member interface by ifnet*.
967  */
968 static struct bridge_iflist *
969 bridge_lookup_member_if(struct bridge_softc *sc, struct ifnet *member_ifp)
970 {
971 	struct bridge_iflist *bif;
972 
973 	BRIDGE_LOCK_ASSERT(sc);
974 
975 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
976 		if (bif->bif_ifp == member_ifp)
977 			return (bif);
978 	}
979 
980 	return (NULL);
981 }
982 
983 /*
984  * bridge_delete_member:
985  *
986  *	Delete the specified member interface.
987  */
988 static void
989 bridge_delete_member(struct bridge_softc *sc, struct bridge_iflist *bif,
990     int gone)
991 {
992 	struct ifnet *ifs = bif->bif_ifp;
993 	struct ifnet *fif = NULL;
994 
995 	BRIDGE_LOCK_ASSERT(sc);
996 
997 	if (bif->bif_flags & IFBIF_STP)
998 		bstp_disable(&bif->bif_stp);
999 
1000 	ifs->if_bridge = NULL;
1001 	BRIDGE_XLOCK(sc);
1002 	LIST_REMOVE(bif, bif_next);
1003 	BRIDGE_XDROP(sc);
1004 
1005 	/*
1006 	 * If removing the interface that gave the bridge its mac address, set
1007 	 * the mac address of the bridge to the address of the next member, or
1008 	 * to its default address if no members are left.
1009 	 */
1010 	if (V_bridge_inherit_mac && sc->sc_ifaddr == ifs) {
1011 		if (LIST_EMPTY(&sc->sc_iflist)) {
1012 			bcopy(sc->sc_defaddr,
1013 			    IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN);
1014 			sc->sc_ifaddr = NULL;
1015 		} else {
1016 			fif = LIST_FIRST(&sc->sc_iflist)->bif_ifp;
1017 			bcopy(IF_LLADDR(fif),
1018 			    IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN);
1019 			sc->sc_ifaddr = fif;
1020 		}
1021 		EVENTHANDLER_INVOKE(iflladdr_event, sc->sc_ifp);
1022 	}
1023 
1024 	bridge_linkcheck(sc);
1025 	bridge_mutecaps(sc);	/* recalcuate now this interface is removed */
1026 	bridge_rtdelete(sc, ifs, IFBF_FLUSHALL);
1027 	KASSERT(bif->bif_addrcnt == 0,
1028 	    ("%s: %d bridge routes referenced", __func__, bif->bif_addrcnt));
1029 
1030 	BRIDGE_UNLOCK(sc);
1031 	if (!gone) {
1032 		switch (ifs->if_type) {
1033 		case IFT_ETHER:
1034 		case IFT_L2VLAN:
1035 			/*
1036 			 * Take the interface out of promiscuous mode, but only
1037 			 * if it was promiscuous in the first place. It might
1038 			 * not be if we're in the bridge_ioctl_add() error path.
1039 			 */
1040 			if (ifs->if_flags & IFF_PROMISC)
1041 				(void) ifpromisc(ifs, 0);
1042 			break;
1043 
1044 		case IFT_GIF:
1045 			break;
1046 
1047 		default:
1048 #ifdef DIAGNOSTIC
1049 			panic("bridge_delete_member: impossible");
1050 #endif
1051 			break;
1052 		}
1053 		/* reneable any interface capabilities */
1054 		bridge_set_ifcap(sc, bif, bif->bif_savedcaps);
1055 	}
1056 	bstp_destroy(&bif->bif_stp);	/* prepare to free */
1057 	BRIDGE_LOCK(sc);
1058 	free(bif, M_DEVBUF);
1059 }
1060 
1061 /*
1062  * bridge_delete_span:
1063  *
1064  *	Delete the specified span interface.
1065  */
1066 static void
1067 bridge_delete_span(struct bridge_softc *sc, struct bridge_iflist *bif)
1068 {
1069 	BRIDGE_LOCK_ASSERT(sc);
1070 
1071 	KASSERT(bif->bif_ifp->if_bridge == NULL,
1072 	    ("%s: not a span interface", __func__));
1073 
1074 	LIST_REMOVE(bif, bif_next);
1075 	free(bif, M_DEVBUF);
1076 }
1077 
1078 static int
1079 bridge_ioctl_add(struct bridge_softc *sc, void *arg)
1080 {
1081 	struct ifbreq *req = arg;
1082 	struct bridge_iflist *bif = NULL;
1083 	struct ifnet *ifs;
1084 	int error = 0;
1085 
1086 	ifs = ifunit(req->ifbr_ifsname);
1087 	if (ifs == NULL)
1088 		return (ENOENT);
1089 	if (ifs->if_ioctl == NULL)	/* must be supported */
1090 		return (EINVAL);
1091 
1092 	/* If it's in the span list, it can't be a member. */
1093 	LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1094 		if (ifs == bif->bif_ifp)
1095 			return (EBUSY);
1096 
1097 	if (ifs->if_bridge == sc)
1098 		return (EEXIST);
1099 
1100 	if (ifs->if_bridge != NULL)
1101 		return (EBUSY);
1102 
1103 	switch (ifs->if_type) {
1104 	case IFT_ETHER:
1105 	case IFT_L2VLAN:
1106 	case IFT_GIF:
1107 		/* permitted interface types */
1108 		break;
1109 	default:
1110 		return (EINVAL);
1111 	}
1112 
1113 #ifdef INET6
1114 	/*
1115 	 * Two valid inet6 addresses with link-local scope must not be
1116 	 * on the parent interface and the member interfaces at the
1117 	 * same time.  This restriction is needed to prevent violation
1118 	 * of link-local scope zone.  Attempts to add a member
1119 	 * interface which has inet6 addresses when the parent has
1120 	 * inet6 triggers removal of all inet6 addresses on the member
1121 	 * interface.
1122 	 */
1123 
1124 	/* Check if the parent interface has a link-local scope addr. */
1125 	if (V_allow_llz_overlap == 0 &&
1126 	    in6ifa_llaonifp(sc->sc_ifp) != NULL) {
1127 		/*
1128 		 * If any, remove all inet6 addresses from the member
1129 		 * interfaces.
1130 		 */
1131 		BRIDGE_XLOCK(sc);
1132 		LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
1133  			if (in6ifa_llaonifp(bif->bif_ifp)) {
1134 				BRIDGE_UNLOCK(sc);
1135 				in6_ifdetach(bif->bif_ifp);
1136 				BRIDGE_LOCK(sc);
1137 				if_printf(sc->sc_ifp,
1138 				    "IPv6 addresses on %s have been removed "
1139 				    "before adding it as a member to prevent "
1140 				    "IPv6 address scope violation.\n",
1141 				    bif->bif_ifp->if_xname);
1142 			}
1143 		}
1144 		BRIDGE_XDROP(sc);
1145 		if (in6ifa_llaonifp(ifs)) {
1146 			BRIDGE_UNLOCK(sc);
1147 			in6_ifdetach(ifs);
1148 			BRIDGE_LOCK(sc);
1149 			if_printf(sc->sc_ifp,
1150 			    "IPv6 addresses on %s have been removed "
1151 			    "before adding it as a member to prevent "
1152 			    "IPv6 address scope violation.\n",
1153 			    ifs->if_xname);
1154 		}
1155 	}
1156 #endif
1157 	/* Allow the first Ethernet member to define the MTU */
1158 	if (LIST_EMPTY(&sc->sc_iflist))
1159 		sc->sc_ifp->if_mtu = ifs->if_mtu;
1160 	else if (sc->sc_ifp->if_mtu != ifs->if_mtu) {
1161 		if_printf(sc->sc_ifp, "invalid MTU: %u(%s) != %u\n",
1162 		    ifs->if_mtu, ifs->if_xname, sc->sc_ifp->if_mtu);
1163 		return (EINVAL);
1164 	}
1165 
1166 	bif = malloc(sizeof(*bif), M_DEVBUF, M_NOWAIT|M_ZERO);
1167 	if (bif == NULL)
1168 		return (ENOMEM);
1169 
1170 	bif->bif_ifp = ifs;
1171 	bif->bif_flags = IFBIF_LEARNING | IFBIF_DISCOVER;
1172 	bif->bif_savedcaps = ifs->if_capenable;
1173 
1174 	/*
1175 	 * Assign the interface's MAC address to the bridge if it's the first
1176 	 * member and the MAC address of the bridge has not been changed from
1177 	 * the default randomly generated one.
1178 	 */
1179 	if (V_bridge_inherit_mac && LIST_EMPTY(&sc->sc_iflist) &&
1180 	    !memcmp(IF_LLADDR(sc->sc_ifp), sc->sc_defaddr, ETHER_ADDR_LEN)) {
1181 		bcopy(IF_LLADDR(ifs), IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN);
1182 		sc->sc_ifaddr = ifs;
1183 		EVENTHANDLER_INVOKE(iflladdr_event, sc->sc_ifp);
1184 	}
1185 
1186 	ifs->if_bridge = sc;
1187 	bstp_create(&sc->sc_stp, &bif->bif_stp, bif->bif_ifp);
1188 	/*
1189 	 * XXX: XLOCK HERE!?!
1190 	 *
1191 	 * NOTE: insert_***HEAD*** should be safe for the traversals.
1192 	 */
1193 	LIST_INSERT_HEAD(&sc->sc_iflist, bif, bif_next);
1194 
1195 	/* Set interface capabilities to the intersection set of all members */
1196 	bridge_mutecaps(sc);
1197 	bridge_linkcheck(sc);
1198 
1199 	/* Place the interface into promiscuous mode */
1200 	switch (ifs->if_type) {
1201 		case IFT_ETHER:
1202 		case IFT_L2VLAN:
1203 			BRIDGE_UNLOCK(sc);
1204 			error = ifpromisc(ifs, 1);
1205 			BRIDGE_LOCK(sc);
1206 			break;
1207 	}
1208 
1209 	if (error)
1210 		bridge_delete_member(sc, bif, 0);
1211 	return (error);
1212 }
1213 
1214 static int
1215 bridge_ioctl_del(struct bridge_softc *sc, void *arg)
1216 {
1217 	struct ifbreq *req = arg;
1218 	struct bridge_iflist *bif;
1219 
1220 	bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1221 	if (bif == NULL)
1222 		return (ENOENT);
1223 
1224 	bridge_delete_member(sc, bif, 0);
1225 
1226 	return (0);
1227 }
1228 
1229 static int
1230 bridge_ioctl_gifflags(struct bridge_softc *sc, void *arg)
1231 {
1232 	struct ifbreq *req = arg;
1233 	struct bridge_iflist *bif;
1234 	struct bstp_port *bp;
1235 
1236 	bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1237 	if (bif == NULL)
1238 		return (ENOENT);
1239 
1240 	bp = &bif->bif_stp;
1241 	req->ifbr_ifsflags = bif->bif_flags;
1242 	req->ifbr_state = bp->bp_state;
1243 	req->ifbr_priority = bp->bp_priority;
1244 	req->ifbr_path_cost = bp->bp_path_cost;
1245 	req->ifbr_portno = bif->bif_ifp->if_index & 0xfff;
1246 	req->ifbr_proto = bp->bp_protover;
1247 	req->ifbr_role = bp->bp_role;
1248 	req->ifbr_stpflags = bp->bp_flags;
1249 	req->ifbr_addrcnt = bif->bif_addrcnt;
1250 	req->ifbr_addrmax = bif->bif_addrmax;
1251 	req->ifbr_addrexceeded = bif->bif_addrexceeded;
1252 
1253 	/* Copy STP state options as flags */
1254 	if (bp->bp_operedge)
1255 		req->ifbr_ifsflags |= IFBIF_BSTP_EDGE;
1256 	if (bp->bp_flags & BSTP_PORT_AUTOEDGE)
1257 		req->ifbr_ifsflags |= IFBIF_BSTP_AUTOEDGE;
1258 	if (bp->bp_ptp_link)
1259 		req->ifbr_ifsflags |= IFBIF_BSTP_PTP;
1260 	if (bp->bp_flags & BSTP_PORT_AUTOPTP)
1261 		req->ifbr_ifsflags |= IFBIF_BSTP_AUTOPTP;
1262 	if (bp->bp_flags & BSTP_PORT_ADMEDGE)
1263 		req->ifbr_ifsflags |= IFBIF_BSTP_ADMEDGE;
1264 	if (bp->bp_flags & BSTP_PORT_ADMCOST)
1265 		req->ifbr_ifsflags |= IFBIF_BSTP_ADMCOST;
1266 	return (0);
1267 }
1268 
1269 static int
1270 bridge_ioctl_sifflags(struct bridge_softc *sc, void *arg)
1271 {
1272 	struct ifbreq *req = arg;
1273 	struct bridge_iflist *bif;
1274 	struct bstp_port *bp;
1275 	int error;
1276 
1277 	bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1278 	if (bif == NULL)
1279 		return (ENOENT);
1280 	bp = &bif->bif_stp;
1281 
1282 	if (req->ifbr_ifsflags & IFBIF_SPAN)
1283 		/* SPAN is readonly */
1284 		return (EINVAL);
1285 
1286 	if (req->ifbr_ifsflags & IFBIF_STP) {
1287 		if ((bif->bif_flags & IFBIF_STP) == 0) {
1288 			error = bstp_enable(&bif->bif_stp);
1289 			if (error)
1290 				return (error);
1291 		}
1292 	} else {
1293 		if ((bif->bif_flags & IFBIF_STP) != 0)
1294 			bstp_disable(&bif->bif_stp);
1295 	}
1296 
1297 	/* Pass on STP flags */
1298 	bstp_set_edge(bp, req->ifbr_ifsflags & IFBIF_BSTP_EDGE ? 1 : 0);
1299 	bstp_set_autoedge(bp, req->ifbr_ifsflags & IFBIF_BSTP_AUTOEDGE ? 1 : 0);
1300 	bstp_set_ptp(bp, req->ifbr_ifsflags & IFBIF_BSTP_PTP ? 1 : 0);
1301 	bstp_set_autoptp(bp, req->ifbr_ifsflags & IFBIF_BSTP_AUTOPTP ? 1 : 0);
1302 
1303 	/* Save the bits relating to the bridge */
1304 	bif->bif_flags = req->ifbr_ifsflags & IFBIFMASK;
1305 
1306 	return (0);
1307 }
1308 
1309 static int
1310 bridge_ioctl_scache(struct bridge_softc *sc, void *arg)
1311 {
1312 	struct ifbrparam *param = arg;
1313 
1314 	sc->sc_brtmax = param->ifbrp_csize;
1315 	bridge_rttrim(sc);
1316 
1317 	return (0);
1318 }
1319 
1320 static int
1321 bridge_ioctl_gcache(struct bridge_softc *sc, void *arg)
1322 {
1323 	struct ifbrparam *param = arg;
1324 
1325 	param->ifbrp_csize = sc->sc_brtmax;
1326 
1327 	return (0);
1328 }
1329 
1330 static int
1331 bridge_ioctl_gifs(struct bridge_softc *sc, void *arg)
1332 {
1333 	struct ifbifconf *bifc = arg;
1334 	struct bridge_iflist *bif;
1335 	struct ifbreq breq;
1336 	char *buf, *outbuf;
1337 	int count, buflen, len, error = 0;
1338 
1339 	count = 0;
1340 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next)
1341 		count++;
1342 	LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1343 		count++;
1344 
1345 	buflen = sizeof(breq) * count;
1346 	if (bifc->ifbic_len == 0) {
1347 		bifc->ifbic_len = buflen;
1348 		return (0);
1349 	}
1350 	BRIDGE_UNLOCK(sc);
1351 	outbuf = malloc(buflen, M_TEMP, M_WAITOK | M_ZERO);
1352 	BRIDGE_LOCK(sc);
1353 
1354 	count = 0;
1355 	buf = outbuf;
1356 	len = min(bifc->ifbic_len, buflen);
1357 	bzero(&breq, sizeof(breq));
1358 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
1359 		if (len < sizeof(breq))
1360 			break;
1361 
1362 		strlcpy(breq.ifbr_ifsname, bif->bif_ifp->if_xname,
1363 		    sizeof(breq.ifbr_ifsname));
1364 		/* Fill in the ifbreq structure */
1365 		error = bridge_ioctl_gifflags(sc, &breq);
1366 		if (error)
1367 			break;
1368 		memcpy(buf, &breq, sizeof(breq));
1369 		count++;
1370 		buf += sizeof(breq);
1371 		len -= sizeof(breq);
1372 	}
1373 	LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) {
1374 		if (len < sizeof(breq))
1375 			break;
1376 
1377 		strlcpy(breq.ifbr_ifsname, bif->bif_ifp->if_xname,
1378 		    sizeof(breq.ifbr_ifsname));
1379 		breq.ifbr_ifsflags = bif->bif_flags;
1380 		breq.ifbr_portno = bif->bif_ifp->if_index & 0xfff;
1381 		memcpy(buf, &breq, sizeof(breq));
1382 		count++;
1383 		buf += sizeof(breq);
1384 		len -= sizeof(breq);
1385 	}
1386 
1387 	BRIDGE_UNLOCK(sc);
1388 	bifc->ifbic_len = sizeof(breq) * count;
1389 	error = copyout(outbuf, bifc->ifbic_req, bifc->ifbic_len);
1390 	BRIDGE_LOCK(sc);
1391 	free(outbuf, M_TEMP);
1392 	return (error);
1393 }
1394 
1395 static int
1396 bridge_ioctl_rts(struct bridge_softc *sc, void *arg)
1397 {
1398 	struct ifbaconf *bac = arg;
1399 	struct bridge_rtnode *brt;
1400 	struct ifbareq bareq;
1401 	char *buf, *outbuf;
1402 	int count, buflen, len, error = 0;
1403 
1404 	if (bac->ifbac_len == 0)
1405 		return (0);
1406 
1407 	count = 0;
1408 	LIST_FOREACH(brt, &sc->sc_rtlist, brt_list)
1409 		count++;
1410 	buflen = sizeof(bareq) * count;
1411 
1412 	BRIDGE_UNLOCK(sc);
1413 	outbuf = malloc(buflen, M_TEMP, M_WAITOK | M_ZERO);
1414 	BRIDGE_LOCK(sc);
1415 
1416 	count = 0;
1417 	buf = outbuf;
1418 	len = min(bac->ifbac_len, buflen);
1419 	bzero(&bareq, sizeof(bareq));
1420 	LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) {
1421 		if (len < sizeof(bareq))
1422 			goto out;
1423 		strlcpy(bareq.ifba_ifsname, brt->brt_ifp->if_xname,
1424 		    sizeof(bareq.ifba_ifsname));
1425 		memcpy(bareq.ifba_dst, brt->brt_addr, sizeof(brt->brt_addr));
1426 		bareq.ifba_vlan = brt->brt_vlan;
1427 		if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC &&
1428 				time_uptime < brt->brt_expire)
1429 			bareq.ifba_expire = brt->brt_expire - time_uptime;
1430 		else
1431 			bareq.ifba_expire = 0;
1432 		bareq.ifba_flags = brt->brt_flags;
1433 
1434 		memcpy(buf, &bareq, sizeof(bareq));
1435 		count++;
1436 		buf += sizeof(bareq);
1437 		len -= sizeof(bareq);
1438 	}
1439 out:
1440 	BRIDGE_UNLOCK(sc);
1441 	bac->ifbac_len = sizeof(bareq) * count;
1442 	error = copyout(outbuf, bac->ifbac_req, bac->ifbac_len);
1443 	BRIDGE_LOCK(sc);
1444 	free(outbuf, M_TEMP);
1445 	return (error);
1446 }
1447 
1448 static int
1449 bridge_ioctl_saddr(struct bridge_softc *sc, void *arg)
1450 {
1451 	struct ifbareq *req = arg;
1452 	struct bridge_iflist *bif;
1453 	int error;
1454 
1455 	bif = bridge_lookup_member(sc, req->ifba_ifsname);
1456 	if (bif == NULL)
1457 		return (ENOENT);
1458 
1459 	error = bridge_rtupdate(sc, req->ifba_dst, req->ifba_vlan, bif, 1,
1460 	    req->ifba_flags);
1461 
1462 	return (error);
1463 }
1464 
1465 static int
1466 bridge_ioctl_sto(struct bridge_softc *sc, void *arg)
1467 {
1468 	struct ifbrparam *param = arg;
1469 
1470 	sc->sc_brttimeout = param->ifbrp_ctime;
1471 	return (0);
1472 }
1473 
1474 static int
1475 bridge_ioctl_gto(struct bridge_softc *sc, void *arg)
1476 {
1477 	struct ifbrparam *param = arg;
1478 
1479 	param->ifbrp_ctime = sc->sc_brttimeout;
1480 	return (0);
1481 }
1482 
1483 static int
1484 bridge_ioctl_daddr(struct bridge_softc *sc, void *arg)
1485 {
1486 	struct ifbareq *req = arg;
1487 
1488 	return (bridge_rtdaddr(sc, req->ifba_dst, req->ifba_vlan));
1489 }
1490 
1491 static int
1492 bridge_ioctl_flush(struct bridge_softc *sc, void *arg)
1493 {
1494 	struct ifbreq *req = arg;
1495 
1496 	bridge_rtflush(sc, req->ifbr_ifsflags);
1497 	return (0);
1498 }
1499 
1500 static int
1501 bridge_ioctl_gpri(struct bridge_softc *sc, void *arg)
1502 {
1503 	struct ifbrparam *param = arg;
1504 	struct bstp_state *bs = &sc->sc_stp;
1505 
1506 	param->ifbrp_prio = bs->bs_bridge_priority;
1507 	return (0);
1508 }
1509 
1510 static int
1511 bridge_ioctl_spri(struct bridge_softc *sc, void *arg)
1512 {
1513 	struct ifbrparam *param = arg;
1514 
1515 	return (bstp_set_priority(&sc->sc_stp, param->ifbrp_prio));
1516 }
1517 
1518 static int
1519 bridge_ioctl_ght(struct bridge_softc *sc, void *arg)
1520 {
1521 	struct ifbrparam *param = arg;
1522 	struct bstp_state *bs = &sc->sc_stp;
1523 
1524 	param->ifbrp_hellotime = bs->bs_bridge_htime >> 8;
1525 	return (0);
1526 }
1527 
1528 static int
1529 bridge_ioctl_sht(struct bridge_softc *sc, void *arg)
1530 {
1531 	struct ifbrparam *param = arg;
1532 
1533 	return (bstp_set_htime(&sc->sc_stp, param->ifbrp_hellotime));
1534 }
1535 
1536 static int
1537 bridge_ioctl_gfd(struct bridge_softc *sc, void *arg)
1538 {
1539 	struct ifbrparam *param = arg;
1540 	struct bstp_state *bs = &sc->sc_stp;
1541 
1542 	param->ifbrp_fwddelay = bs->bs_bridge_fdelay >> 8;
1543 	return (0);
1544 }
1545 
1546 static int
1547 bridge_ioctl_sfd(struct bridge_softc *sc, void *arg)
1548 {
1549 	struct ifbrparam *param = arg;
1550 
1551 	return (bstp_set_fdelay(&sc->sc_stp, param->ifbrp_fwddelay));
1552 }
1553 
1554 static int
1555 bridge_ioctl_gma(struct bridge_softc *sc, void *arg)
1556 {
1557 	struct ifbrparam *param = arg;
1558 	struct bstp_state *bs = &sc->sc_stp;
1559 
1560 	param->ifbrp_maxage = bs->bs_bridge_max_age >> 8;
1561 	return (0);
1562 }
1563 
1564 static int
1565 bridge_ioctl_sma(struct bridge_softc *sc, void *arg)
1566 {
1567 	struct ifbrparam *param = arg;
1568 
1569 	return (bstp_set_maxage(&sc->sc_stp, param->ifbrp_maxage));
1570 }
1571 
1572 static int
1573 bridge_ioctl_sifprio(struct bridge_softc *sc, void *arg)
1574 {
1575 	struct ifbreq *req = arg;
1576 	struct bridge_iflist *bif;
1577 
1578 	bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1579 	if (bif == NULL)
1580 		return (ENOENT);
1581 
1582 	return (bstp_set_port_priority(&bif->bif_stp, req->ifbr_priority));
1583 }
1584 
1585 static int
1586 bridge_ioctl_sifcost(struct bridge_softc *sc, void *arg)
1587 {
1588 	struct ifbreq *req = arg;
1589 	struct bridge_iflist *bif;
1590 
1591 	bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1592 	if (bif == NULL)
1593 		return (ENOENT);
1594 
1595 	return (bstp_set_path_cost(&bif->bif_stp, req->ifbr_path_cost));
1596 }
1597 
1598 static int
1599 bridge_ioctl_sifmaxaddr(struct bridge_softc *sc, void *arg)
1600 {
1601 	struct ifbreq *req = arg;
1602 	struct bridge_iflist *bif;
1603 
1604 	bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1605 	if (bif == NULL)
1606 		return (ENOENT);
1607 
1608 	bif->bif_addrmax = req->ifbr_addrmax;
1609 	return (0);
1610 }
1611 
1612 static int
1613 bridge_ioctl_addspan(struct bridge_softc *sc, void *arg)
1614 {
1615 	struct ifbreq *req = arg;
1616 	struct bridge_iflist *bif = NULL;
1617 	struct ifnet *ifs;
1618 
1619 	ifs = ifunit(req->ifbr_ifsname);
1620 	if (ifs == NULL)
1621 		return (ENOENT);
1622 
1623 	LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1624 		if (ifs == bif->bif_ifp)
1625 			return (EBUSY);
1626 
1627 	if (ifs->if_bridge != NULL)
1628 		return (EBUSY);
1629 
1630 	switch (ifs->if_type) {
1631 		case IFT_ETHER:
1632 		case IFT_GIF:
1633 		case IFT_L2VLAN:
1634 			break;
1635 		default:
1636 			return (EINVAL);
1637 	}
1638 
1639 	bif = malloc(sizeof(*bif), M_DEVBUF, M_NOWAIT|M_ZERO);
1640 	if (bif == NULL)
1641 		return (ENOMEM);
1642 
1643 	bif->bif_ifp = ifs;
1644 	bif->bif_flags = IFBIF_SPAN;
1645 
1646 	LIST_INSERT_HEAD(&sc->sc_spanlist, bif, bif_next);
1647 
1648 	return (0);
1649 }
1650 
1651 static int
1652 bridge_ioctl_delspan(struct bridge_softc *sc, void *arg)
1653 {
1654 	struct ifbreq *req = arg;
1655 	struct bridge_iflist *bif;
1656 	struct ifnet *ifs;
1657 
1658 	ifs = ifunit(req->ifbr_ifsname);
1659 	if (ifs == NULL)
1660 		return (ENOENT);
1661 
1662 	LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1663 		if (ifs == bif->bif_ifp)
1664 			break;
1665 
1666 	if (bif == NULL)
1667 		return (ENOENT);
1668 
1669 	bridge_delete_span(sc, bif);
1670 
1671 	return (0);
1672 }
1673 
1674 static int
1675 bridge_ioctl_gbparam(struct bridge_softc *sc, void *arg)
1676 {
1677 	struct ifbropreq *req = arg;
1678 	struct bstp_state *bs = &sc->sc_stp;
1679 	struct bstp_port *root_port;
1680 
1681 	req->ifbop_maxage = bs->bs_bridge_max_age >> 8;
1682 	req->ifbop_hellotime = bs->bs_bridge_htime >> 8;
1683 	req->ifbop_fwddelay = bs->bs_bridge_fdelay >> 8;
1684 
1685 	root_port = bs->bs_root_port;
1686 	if (root_port == NULL)
1687 		req->ifbop_root_port = 0;
1688 	else
1689 		req->ifbop_root_port = root_port->bp_ifp->if_index;
1690 
1691 	req->ifbop_holdcount = bs->bs_txholdcount;
1692 	req->ifbop_priority = bs->bs_bridge_priority;
1693 	req->ifbop_protocol = bs->bs_protover;
1694 	req->ifbop_root_path_cost = bs->bs_root_pv.pv_cost;
1695 	req->ifbop_bridgeid = bs->bs_bridge_pv.pv_dbridge_id;
1696 	req->ifbop_designated_root = bs->bs_root_pv.pv_root_id;
1697 	req->ifbop_designated_bridge = bs->bs_root_pv.pv_dbridge_id;
1698 	req->ifbop_last_tc_time.tv_sec = bs->bs_last_tc_time.tv_sec;
1699 	req->ifbop_last_tc_time.tv_usec = bs->bs_last_tc_time.tv_usec;
1700 
1701 	return (0);
1702 }
1703 
1704 static int
1705 bridge_ioctl_grte(struct bridge_softc *sc, void *arg)
1706 {
1707 	struct ifbrparam *param = arg;
1708 
1709 	param->ifbrp_cexceeded = sc->sc_brtexceeded;
1710 	return (0);
1711 }
1712 
1713 static int
1714 bridge_ioctl_gifsstp(struct bridge_softc *sc, void *arg)
1715 {
1716 	struct ifbpstpconf *bifstp = arg;
1717 	struct bridge_iflist *bif;
1718 	struct bstp_port *bp;
1719 	struct ifbpstpreq bpreq;
1720 	char *buf, *outbuf;
1721 	int count, buflen, len, error = 0;
1722 
1723 	count = 0;
1724 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
1725 		if ((bif->bif_flags & IFBIF_STP) != 0)
1726 			count++;
1727 	}
1728 
1729 	buflen = sizeof(bpreq) * count;
1730 	if (bifstp->ifbpstp_len == 0) {
1731 		bifstp->ifbpstp_len = buflen;
1732 		return (0);
1733 	}
1734 
1735 	BRIDGE_UNLOCK(sc);
1736 	outbuf = malloc(buflen, M_TEMP, M_WAITOK | M_ZERO);
1737 	BRIDGE_LOCK(sc);
1738 
1739 	count = 0;
1740 	buf = outbuf;
1741 	len = min(bifstp->ifbpstp_len, buflen);
1742 	bzero(&bpreq, sizeof(bpreq));
1743 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
1744 		if (len < sizeof(bpreq))
1745 			break;
1746 
1747 		if ((bif->bif_flags & IFBIF_STP) == 0)
1748 			continue;
1749 
1750 		bp = &bif->bif_stp;
1751 		bpreq.ifbp_portno = bif->bif_ifp->if_index & 0xfff;
1752 		bpreq.ifbp_fwd_trans = bp->bp_forward_transitions;
1753 		bpreq.ifbp_design_cost = bp->bp_desg_pv.pv_cost;
1754 		bpreq.ifbp_design_port = bp->bp_desg_pv.pv_port_id;
1755 		bpreq.ifbp_design_bridge = bp->bp_desg_pv.pv_dbridge_id;
1756 		bpreq.ifbp_design_root = bp->bp_desg_pv.pv_root_id;
1757 
1758 		memcpy(buf, &bpreq, sizeof(bpreq));
1759 		count++;
1760 		buf += sizeof(bpreq);
1761 		len -= sizeof(bpreq);
1762 	}
1763 
1764 	BRIDGE_UNLOCK(sc);
1765 	bifstp->ifbpstp_len = sizeof(bpreq) * count;
1766 	error = copyout(outbuf, bifstp->ifbpstp_req, bifstp->ifbpstp_len);
1767 	BRIDGE_LOCK(sc);
1768 	free(outbuf, M_TEMP);
1769 	return (error);
1770 }
1771 
1772 static int
1773 bridge_ioctl_sproto(struct bridge_softc *sc, void *arg)
1774 {
1775 	struct ifbrparam *param = arg;
1776 
1777 	return (bstp_set_protocol(&sc->sc_stp, param->ifbrp_proto));
1778 }
1779 
1780 static int
1781 bridge_ioctl_stxhc(struct bridge_softc *sc, void *arg)
1782 {
1783 	struct ifbrparam *param = arg;
1784 
1785 	return (bstp_set_holdcount(&sc->sc_stp, param->ifbrp_txhc));
1786 }
1787 
1788 /*
1789  * bridge_ifdetach:
1790  *
1791  *	Detach an interface from a bridge.  Called when a member
1792  *	interface is detaching.
1793  */
1794 static void
1795 bridge_ifdetach(void *arg __unused, struct ifnet *ifp)
1796 {
1797 	struct bridge_softc *sc = ifp->if_bridge;
1798 	struct bridge_iflist *bif;
1799 
1800 	if (ifp->if_flags & IFF_RENAMING)
1801 		return;
1802 	if (V_bridge_cloner == NULL) {
1803 		/*
1804 		 * This detach handler can be called after
1805 		 * vnet_bridge_uninit().  Just return in that case.
1806 		 */
1807 		return;
1808 	}
1809 	/* Check if the interface is a bridge member */
1810 	if (sc != NULL) {
1811 		BRIDGE_LOCK(sc);
1812 
1813 		bif = bridge_lookup_member_if(sc, ifp);
1814 		if (bif != NULL)
1815 			bridge_delete_member(sc, bif, 1);
1816 
1817 		BRIDGE_UNLOCK(sc);
1818 		return;
1819 	}
1820 
1821 	/* Check if the interface is a span port */
1822 	BRIDGE_LIST_LOCK();
1823 	LIST_FOREACH(sc, &V_bridge_list, sc_list) {
1824 		BRIDGE_LOCK(sc);
1825 		LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1826 			if (ifp == bif->bif_ifp) {
1827 				bridge_delete_span(sc, bif);
1828 				break;
1829 			}
1830 
1831 		BRIDGE_UNLOCK(sc);
1832 	}
1833 	BRIDGE_LIST_UNLOCK();
1834 }
1835 
1836 /*
1837  * bridge_init:
1838  *
1839  *	Initialize a bridge interface.
1840  */
1841 static void
1842 bridge_init(void *xsc)
1843 {
1844 	struct bridge_softc *sc = (struct bridge_softc *)xsc;
1845 	struct ifnet *ifp = sc->sc_ifp;
1846 
1847 	if (ifp->if_drv_flags & IFF_DRV_RUNNING)
1848 		return;
1849 
1850 	BRIDGE_LOCK(sc);
1851 	callout_reset(&sc->sc_brcallout, bridge_rtable_prune_period * hz,
1852 	    bridge_timer, sc);
1853 
1854 	ifp->if_drv_flags |= IFF_DRV_RUNNING;
1855 	bstp_init(&sc->sc_stp);		/* Initialize Spanning Tree */
1856 
1857 	BRIDGE_UNLOCK(sc);
1858 }
1859 
1860 /*
1861  * bridge_stop:
1862  *
1863  *	Stop the bridge interface.
1864  */
1865 static void
1866 bridge_stop(struct ifnet *ifp, int disable)
1867 {
1868 	struct bridge_softc *sc = ifp->if_softc;
1869 
1870 	BRIDGE_LOCK_ASSERT(sc);
1871 
1872 	if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0)
1873 		return;
1874 
1875 	callout_stop(&sc->sc_brcallout);
1876 	bstp_stop(&sc->sc_stp);
1877 
1878 	bridge_rtflush(sc, IFBF_FLUSHDYN);
1879 
1880 	ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
1881 }
1882 
1883 /*
1884  * bridge_enqueue:
1885  *
1886  *	Enqueue a packet on a bridge member interface.
1887  *
1888  */
1889 static int
1890 bridge_enqueue(struct bridge_softc *sc, struct ifnet *dst_ifp, struct mbuf *m)
1891 {
1892 	int len, err = 0;
1893 	short mflags;
1894 	struct mbuf *m0;
1895 
1896 	/* We may be sending a fragment so traverse the mbuf */
1897 	for (; m; m = m0) {
1898 		m0 = m->m_nextpkt;
1899 		m->m_nextpkt = NULL;
1900 		len = m->m_pkthdr.len;
1901 		mflags = m->m_flags;
1902 
1903 		/*
1904 		 * If underlying interface can not do VLAN tag insertion itself
1905 		 * then attach a packet tag that holds it.
1906 		 */
1907 		if ((m->m_flags & M_VLANTAG) &&
1908 		    (dst_ifp->if_capenable & IFCAP_VLAN_HWTAGGING) == 0) {
1909 			m = ether_vlanencap(m, m->m_pkthdr.ether_vtag);
1910 			if (m == NULL) {
1911 				if_printf(dst_ifp,
1912 				    "unable to prepend VLAN header\n");
1913 				if_inc_counter(dst_ifp, IFCOUNTER_OERRORS, 1);
1914 				continue;
1915 			}
1916 			m->m_flags &= ~M_VLANTAG;
1917 		}
1918 
1919 		if ((err = dst_ifp->if_transmit(dst_ifp, m))) {
1920 			m_freem(m0);
1921 			if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
1922 			break;
1923 		}
1924 
1925 		if_inc_counter(sc->sc_ifp, IFCOUNTER_OPACKETS, 1);
1926 		if_inc_counter(sc->sc_ifp, IFCOUNTER_OBYTES, len);
1927 		if (mflags & M_MCAST)
1928 			if_inc_counter(sc->sc_ifp, IFCOUNTER_OMCASTS, 1);
1929 	}
1930 
1931 	return (err);
1932 }
1933 
1934 /*
1935  * bridge_dummynet:
1936  *
1937  * 	Receive a queued packet from dummynet and pass it on to the output
1938  * 	interface.
1939  *
1940  *	The mbuf has the Ethernet header already attached.
1941  */
1942 static void
1943 bridge_dummynet(struct mbuf *m, struct ifnet *ifp)
1944 {
1945 	struct bridge_softc *sc;
1946 
1947 	sc = ifp->if_bridge;
1948 
1949 	/*
1950 	 * The packet didnt originate from a member interface. This should only
1951 	 * ever happen if a member interface is removed while packets are
1952 	 * queued for it.
1953 	 */
1954 	if (sc == NULL) {
1955 		m_freem(m);
1956 		return;
1957 	}
1958 
1959 	if (PFIL_HOOKED(&V_inet_pfil_hook)
1960 #ifdef INET6
1961 	    || PFIL_HOOKED(&V_inet6_pfil_hook)
1962 #endif
1963 	    ) {
1964 		if (bridge_pfil(&m, sc->sc_ifp, ifp, PFIL_OUT) != 0)
1965 			return;
1966 		if (m == NULL)
1967 			return;
1968 	}
1969 
1970 	bridge_enqueue(sc, ifp, m);
1971 }
1972 
1973 /*
1974  * bridge_output:
1975  *
1976  *	Send output from a bridge member interface.  This
1977  *	performs the bridging function for locally originated
1978  *	packets.
1979  *
1980  *	The mbuf has the Ethernet header already attached.  We must
1981  *	enqueue or free the mbuf before returning.
1982  */
1983 static int
1984 bridge_output(struct ifnet *ifp, struct mbuf *m, struct sockaddr *sa,
1985     struct rtentry *rt)
1986 {
1987 	struct ether_header *eh;
1988 	struct ifnet *dst_if;
1989 	struct bridge_softc *sc;
1990 	uint16_t vlan;
1991 
1992 	if (m->m_len < ETHER_HDR_LEN) {
1993 		m = m_pullup(m, ETHER_HDR_LEN);
1994 		if (m == NULL)
1995 			return (0);
1996 	}
1997 
1998 	eh = mtod(m, struct ether_header *);
1999 	sc = ifp->if_bridge;
2000 	vlan = VLANTAGOF(m);
2001 
2002 	BRIDGE_LOCK(sc);
2003 
2004 	/*
2005 	 * If bridge is down, but the original output interface is up,
2006 	 * go ahead and send out that interface.  Otherwise, the packet
2007 	 * is dropped below.
2008 	 */
2009 	if ((sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) {
2010 		dst_if = ifp;
2011 		goto sendunicast;
2012 	}
2013 
2014 	/*
2015 	 * If the packet is a multicast, or we don't know a better way to
2016 	 * get there, send to all interfaces.
2017 	 */
2018 	if (ETHER_IS_MULTICAST(eh->ether_dhost))
2019 		dst_if = NULL;
2020 	else
2021 		dst_if = bridge_rtlookup(sc, eh->ether_dhost, vlan);
2022 	if (dst_if == NULL) {
2023 		struct bridge_iflist *bif;
2024 		struct mbuf *mc;
2025 		int error = 0, used = 0;
2026 
2027 		bridge_span(sc, m);
2028 
2029 		BRIDGE_LOCK2REF(sc, error);
2030 		if (error) {
2031 			m_freem(m);
2032 			return (0);
2033 		}
2034 
2035 		LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
2036 			dst_if = bif->bif_ifp;
2037 
2038 			if (dst_if->if_type == IFT_GIF)
2039 				continue;
2040 			if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
2041 				continue;
2042 
2043 			/*
2044 			 * If this is not the original output interface,
2045 			 * and the interface is participating in spanning
2046 			 * tree, make sure the port is in a state that
2047 			 * allows forwarding.
2048 			 */
2049 			if (dst_if != ifp && (bif->bif_flags & IFBIF_STP) &&
2050 			    bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
2051 				continue;
2052 
2053 			if (LIST_NEXT(bif, bif_next) == NULL) {
2054 				used = 1;
2055 				mc = m;
2056 			} else {
2057 				mc = m_copypacket(m, M_NOWAIT);
2058 				if (mc == NULL) {
2059 					if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
2060 					continue;
2061 				}
2062 			}
2063 
2064 			bridge_enqueue(sc, dst_if, mc);
2065 		}
2066 		if (used == 0)
2067 			m_freem(m);
2068 		BRIDGE_UNREF(sc);
2069 		return (0);
2070 	}
2071 
2072 sendunicast:
2073 	/*
2074 	 * XXX Spanning tree consideration here?
2075 	 */
2076 
2077 	bridge_span(sc, m);
2078 	if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) {
2079 		m_freem(m);
2080 		BRIDGE_UNLOCK(sc);
2081 		return (0);
2082 	}
2083 
2084 	BRIDGE_UNLOCK(sc);
2085 	bridge_enqueue(sc, dst_if, m);
2086 	return (0);
2087 }
2088 
2089 /*
2090  * bridge_transmit:
2091  *
2092  *	Do output on a bridge.
2093  *
2094  */
2095 static int
2096 bridge_transmit(struct ifnet *ifp, struct mbuf *m)
2097 {
2098 	struct bridge_softc *sc;
2099 	struct ether_header *eh;
2100 	struct ifnet *dst_if;
2101 	int error = 0;
2102 
2103 	sc = ifp->if_softc;
2104 
2105 	ETHER_BPF_MTAP(ifp, m);
2106 
2107 	eh = mtod(m, struct ether_header *);
2108 
2109 	BRIDGE_LOCK(sc);
2110 	if (((m->m_flags & (M_BCAST|M_MCAST)) == 0) &&
2111 	    (dst_if = bridge_rtlookup(sc, eh->ether_dhost, 1)) != NULL) {
2112 		BRIDGE_UNLOCK(sc);
2113 		error = bridge_enqueue(sc, dst_if, m);
2114 	} else
2115 		bridge_broadcast(sc, ifp, m, 0);
2116 
2117 	return (error);
2118 }
2119 
2120 /*
2121  * The ifp->if_qflush entry point for if_bridge(4) is no-op.
2122  */
2123 static void
2124 bridge_qflush(struct ifnet *ifp __unused)
2125 {
2126 }
2127 
2128 /*
2129  * bridge_forward:
2130  *
2131  *	The forwarding function of the bridge.
2132  *
2133  *	NOTE: Releases the lock on return.
2134  */
2135 static void
2136 bridge_forward(struct bridge_softc *sc, struct bridge_iflist *sbif,
2137     struct mbuf *m)
2138 {
2139 	struct bridge_iflist *dbif;
2140 	struct ifnet *src_if, *dst_if, *ifp;
2141 	struct ether_header *eh;
2142 	uint16_t vlan;
2143 	uint8_t *dst;
2144 	int error;
2145 
2146 	src_if = m->m_pkthdr.rcvif;
2147 	ifp = sc->sc_ifp;
2148 
2149 	if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1);
2150 	if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len);
2151 	vlan = VLANTAGOF(m);
2152 
2153 	if ((sbif->bif_flags & IFBIF_STP) &&
2154 	    sbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
2155 		goto drop;
2156 
2157 	eh = mtod(m, struct ether_header *);
2158 	dst = eh->ether_dhost;
2159 
2160 	/* If the interface is learning, record the address. */
2161 	if (sbif->bif_flags & IFBIF_LEARNING) {
2162 		error = bridge_rtupdate(sc, eh->ether_shost, vlan,
2163 		    sbif, 0, IFBAF_DYNAMIC);
2164 		/*
2165 		 * If the interface has addresses limits then deny any source
2166 		 * that is not in the cache.
2167 		 */
2168 		if (error && sbif->bif_addrmax)
2169 			goto drop;
2170 	}
2171 
2172 	if ((sbif->bif_flags & IFBIF_STP) != 0 &&
2173 	    sbif->bif_stp.bp_state == BSTP_IFSTATE_LEARNING)
2174 		goto drop;
2175 
2176 	/*
2177 	 * At this point, the port either doesn't participate
2178 	 * in spanning tree or it is in the forwarding state.
2179 	 */
2180 
2181 	/*
2182 	 * If the packet is unicast, destined for someone on
2183 	 * "this" side of the bridge, drop it.
2184 	 */
2185 	if ((m->m_flags & (M_BCAST|M_MCAST)) == 0) {
2186 		dst_if = bridge_rtlookup(sc, dst, vlan);
2187 		if (src_if == dst_if)
2188 			goto drop;
2189 	} else {
2190 		/*
2191 		 * Check if its a reserved multicast address, any address
2192 		 * listed in 802.1D section 7.12.6 may not be forwarded by the
2193 		 * bridge.
2194 		 * This is currently 01-80-C2-00-00-00 to 01-80-C2-00-00-0F
2195 		 */
2196 		if (dst[0] == 0x01 && dst[1] == 0x80 &&
2197 		    dst[2] == 0xc2 && dst[3] == 0x00 &&
2198 		    dst[4] == 0x00 && dst[5] <= 0x0f)
2199 			goto drop;
2200 
2201 		/* ...forward it to all interfaces. */
2202 		if_inc_counter(ifp, IFCOUNTER_IMCASTS, 1);
2203 		dst_if = NULL;
2204 	}
2205 
2206 	/*
2207 	 * If we have a destination interface which is a member of our bridge,
2208 	 * OR this is a unicast packet, push it through the bpf(4) machinery.
2209 	 * For broadcast or multicast packets, don't bother because it will
2210 	 * be reinjected into ether_input. We do this before we pass the packets
2211 	 * through the pfil(9) framework, as it is possible that pfil(9) will
2212 	 * drop the packet, or possibly modify it, making it difficult to debug
2213 	 * firewall issues on the bridge.
2214 	 */
2215 	if (dst_if != NULL || (m->m_flags & (M_BCAST | M_MCAST)) == 0)
2216 		ETHER_BPF_MTAP(ifp, m);
2217 
2218 	/* run the packet filter */
2219 	if (PFIL_HOOKED(&V_inet_pfil_hook)
2220 #ifdef INET6
2221 	    || PFIL_HOOKED(&V_inet6_pfil_hook)
2222 #endif
2223 	    ) {
2224 		BRIDGE_UNLOCK(sc);
2225 		if (bridge_pfil(&m, ifp, src_if, PFIL_IN) != 0)
2226 			return;
2227 		if (m == NULL)
2228 			return;
2229 		BRIDGE_LOCK(sc);
2230 	}
2231 
2232 	if (dst_if == NULL) {
2233 		bridge_broadcast(sc, src_if, m, 1);
2234 		return;
2235 	}
2236 
2237 	/*
2238 	 * At this point, we're dealing with a unicast frame
2239 	 * going to a different interface.
2240 	 */
2241 	if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
2242 		goto drop;
2243 
2244 	dbif = bridge_lookup_member_if(sc, dst_if);
2245 	if (dbif == NULL)
2246 		/* Not a member of the bridge (anymore?) */
2247 		goto drop;
2248 
2249 	/* Private segments can not talk to each other */
2250 	if (sbif->bif_flags & dbif->bif_flags & IFBIF_PRIVATE)
2251 		goto drop;
2252 
2253 	if ((dbif->bif_flags & IFBIF_STP) &&
2254 	    dbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
2255 		goto drop;
2256 
2257 	BRIDGE_UNLOCK(sc);
2258 
2259 	if (PFIL_HOOKED(&V_inet_pfil_hook)
2260 #ifdef INET6
2261 	    || PFIL_HOOKED(&V_inet6_pfil_hook)
2262 #endif
2263 	    ) {
2264 		if (bridge_pfil(&m, ifp, dst_if, PFIL_OUT) != 0)
2265 			return;
2266 		if (m == NULL)
2267 			return;
2268 	}
2269 
2270 	bridge_enqueue(sc, dst_if, m);
2271 	return;
2272 
2273 drop:
2274 	BRIDGE_UNLOCK(sc);
2275 	m_freem(m);
2276 }
2277 
2278 /*
2279  * bridge_input:
2280  *
2281  *	Receive input from a member interface.  Queue the packet for
2282  *	bridging if it is not for us.
2283  */
2284 static struct mbuf *
2285 bridge_input(struct ifnet *ifp, struct mbuf *m)
2286 {
2287 	struct bridge_softc *sc = ifp->if_bridge;
2288 	struct bridge_iflist *bif, *bif2;
2289 	struct ifnet *bifp;
2290 	struct ether_header *eh;
2291 	struct mbuf *mc, *mc2;
2292 	uint16_t vlan;
2293 	int error;
2294 
2295 	if ((sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING) == 0)
2296 		return (m);
2297 
2298 	bifp = sc->sc_ifp;
2299 	vlan = VLANTAGOF(m);
2300 
2301 	/*
2302 	 * Implement support for bridge monitoring. If this flag has been
2303 	 * set on this interface, discard the packet once we push it through
2304 	 * the bpf(4) machinery, but before we do, increment the byte and
2305 	 * packet counters associated with this interface.
2306 	 */
2307 	if ((bifp->if_flags & IFF_MONITOR) != 0) {
2308 		m->m_pkthdr.rcvif  = bifp;
2309 		ETHER_BPF_MTAP(bifp, m);
2310 		if_inc_counter(bifp, IFCOUNTER_IPACKETS, 1);
2311 		if_inc_counter(bifp, IFCOUNTER_IBYTES, m->m_pkthdr.len);
2312 		m_freem(m);
2313 		return (NULL);
2314 	}
2315 	BRIDGE_LOCK(sc);
2316 	bif = bridge_lookup_member_if(sc, ifp);
2317 	if (bif == NULL) {
2318 		BRIDGE_UNLOCK(sc);
2319 		return (m);
2320 	}
2321 
2322 	eh = mtod(m, struct ether_header *);
2323 
2324 	bridge_span(sc, m);
2325 
2326 	if (m->m_flags & (M_BCAST|M_MCAST)) {
2327 		/* Tap off 802.1D packets; they do not get forwarded. */
2328 		if (memcmp(eh->ether_dhost, bstp_etheraddr,
2329 		    ETHER_ADDR_LEN) == 0) {
2330 			bstp_input(&bif->bif_stp, ifp, m); /* consumes mbuf */
2331 			BRIDGE_UNLOCK(sc);
2332 			return (NULL);
2333 		}
2334 
2335 		if ((bif->bif_flags & IFBIF_STP) &&
2336 		    bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) {
2337 			BRIDGE_UNLOCK(sc);
2338 			return (m);
2339 		}
2340 
2341 		/*
2342 		 * Make a deep copy of the packet and enqueue the copy
2343 		 * for bridge processing; return the original packet for
2344 		 * local processing.
2345 		 */
2346 		mc = m_dup(m, M_NOWAIT);
2347 		if (mc == NULL) {
2348 			BRIDGE_UNLOCK(sc);
2349 			return (m);
2350 		}
2351 
2352 		/* Perform the bridge forwarding function with the copy. */
2353 		bridge_forward(sc, bif, mc);
2354 
2355 		/*
2356 		 * Reinject the mbuf as arriving on the bridge so we have a
2357 		 * chance at claiming multicast packets. We can not loop back
2358 		 * here from ether_input as a bridge is never a member of a
2359 		 * bridge.
2360 		 */
2361 		KASSERT(bifp->if_bridge == NULL,
2362 		    ("loop created in bridge_input"));
2363 		mc2 = m_dup(m, M_NOWAIT);
2364 		if (mc2 != NULL) {
2365 			/* Keep the layer3 header aligned */
2366 			int i = min(mc2->m_pkthdr.len, max_protohdr);
2367 			mc2 = m_copyup(mc2, i, ETHER_ALIGN);
2368 		}
2369 		if (mc2 != NULL) {
2370 			mc2->m_pkthdr.rcvif = bifp;
2371 			(*bifp->if_input)(bifp, mc2);
2372 		}
2373 
2374 		/* Return the original packet for local processing. */
2375 		return (m);
2376 	}
2377 
2378 	if ((bif->bif_flags & IFBIF_STP) &&
2379 	    bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) {
2380 		BRIDGE_UNLOCK(sc);
2381 		return (m);
2382 	}
2383 
2384 #if (defined(INET) || defined(INET6))
2385 #   define OR_CARP_CHECK_WE_ARE_DST(iface) \
2386 	|| ((iface)->if_carp \
2387 	    && (*carp_forus_p)((iface), eh->ether_dhost))
2388 #   define OR_CARP_CHECK_WE_ARE_SRC(iface) \
2389 	|| ((iface)->if_carp \
2390 	    && (*carp_forus_p)((iface), eh->ether_shost))
2391 #else
2392 #   define OR_CARP_CHECK_WE_ARE_DST(iface)
2393 #   define OR_CARP_CHECK_WE_ARE_SRC(iface)
2394 #endif
2395 
2396 #ifdef INET6
2397 #   define OR_PFIL_HOOKED_INET6 \
2398 	|| PFIL_HOOKED(&V_inet6_pfil_hook)
2399 #else
2400 #   define OR_PFIL_HOOKED_INET6
2401 #endif
2402 
2403 #define GRAB_OUR_PACKETS(iface) \
2404 	if ((iface)->if_type == IFT_GIF) \
2405 		continue; \
2406 	/* It is destined for us. */ \
2407 	if (memcmp(IF_LLADDR((iface)), eh->ether_dhost,  ETHER_ADDR_LEN) == 0 \
2408 	    OR_CARP_CHECK_WE_ARE_DST((iface))				\
2409 	    ) {								\
2410 		if ((iface)->if_type == IFT_BRIDGE) {			\
2411 			ETHER_BPF_MTAP(iface, m);			\
2412 			if_inc_counter(iface, IFCOUNTER_IPACKETS, 1);				\
2413 			if_inc_counter(iface, IFCOUNTER_IBYTES, m->m_pkthdr.len);		\
2414 			/* Filter on the physical interface. */		\
2415 			if (V_pfil_local_phys &&			\
2416 			    (PFIL_HOOKED(&V_inet_pfil_hook)		\
2417 			     OR_PFIL_HOOKED_INET6)) {			\
2418 				if (bridge_pfil(&m, NULL, ifp,		\
2419 				    PFIL_IN) != 0 || m == NULL) {	\
2420 					BRIDGE_UNLOCK(sc);		\
2421 					return (NULL);			\
2422 				}					\
2423 				eh = mtod(m, struct ether_header *);	\
2424 			}						\
2425 		}							\
2426 		if (bif->bif_flags & IFBIF_LEARNING) {			\
2427 			error = bridge_rtupdate(sc, eh->ether_shost,	\
2428 			    vlan, bif, 0, IFBAF_DYNAMIC);		\
2429 			if (error && bif->bif_addrmax) {		\
2430 				BRIDGE_UNLOCK(sc);			\
2431 				m_freem(m);				\
2432 				return (NULL);				\
2433 			}						\
2434 		}							\
2435 		m->m_pkthdr.rcvif = iface;				\
2436 		BRIDGE_UNLOCK(sc);					\
2437 		return (m);						\
2438 	}								\
2439 									\
2440 	/* We just received a packet that we sent out. */		\
2441 	if (memcmp(IF_LLADDR((iface)), eh->ether_shost, ETHER_ADDR_LEN) == 0 \
2442 	    OR_CARP_CHECK_WE_ARE_SRC((iface))			\
2443 	    ) {								\
2444 		BRIDGE_UNLOCK(sc);					\
2445 		m_freem(m);						\
2446 		return (NULL);						\
2447 	}
2448 
2449 	/*
2450 	 * Unicast.  Make sure it's not for the bridge.
2451 	 */
2452 	do { GRAB_OUR_PACKETS(bifp) } while (0);
2453 
2454 	/*
2455 	 * Give a chance for ifp at first priority. This will help when	the
2456 	 * packet comes through the interface like VLAN's with the same MACs
2457 	 * on several interfaces from the same bridge. This also will save
2458 	 * some CPU cycles in case the destination interface and the input
2459 	 * interface (eq ifp) are the same.
2460 	 */
2461 	do { GRAB_OUR_PACKETS(ifp) } while (0);
2462 
2463 	/* Now check the all bridge members. */
2464 	LIST_FOREACH(bif2, &sc->sc_iflist, bif_next) {
2465 		GRAB_OUR_PACKETS(bif2->bif_ifp)
2466 	}
2467 
2468 #undef OR_CARP_CHECK_WE_ARE_DST
2469 #undef OR_CARP_CHECK_WE_ARE_SRC
2470 #undef OR_PFIL_HOOKED_INET6
2471 #undef GRAB_OUR_PACKETS
2472 
2473 	/* Perform the bridge forwarding function. */
2474 	bridge_forward(sc, bif, m);
2475 
2476 	return (NULL);
2477 }
2478 
2479 /*
2480  * bridge_broadcast:
2481  *
2482  *	Send a frame to all interfaces that are members of
2483  *	the bridge, except for the one on which the packet
2484  *	arrived.
2485  *
2486  *	NOTE: Releases the lock on return.
2487  */
2488 static void
2489 bridge_broadcast(struct bridge_softc *sc, struct ifnet *src_if,
2490     struct mbuf *m, int runfilt)
2491 {
2492 	struct bridge_iflist *dbif, *sbif;
2493 	struct mbuf *mc;
2494 	struct ifnet *dst_if;
2495 	int error = 0, used = 0, i;
2496 
2497 	sbif = bridge_lookup_member_if(sc, src_if);
2498 
2499 	BRIDGE_LOCK2REF(sc, error);
2500 	if (error) {
2501 		m_freem(m);
2502 		return;
2503 	}
2504 
2505 	/* Filter on the bridge interface before broadcasting */
2506 	if (runfilt && (PFIL_HOOKED(&V_inet_pfil_hook)
2507 #ifdef INET6
2508 	    || PFIL_HOOKED(&V_inet6_pfil_hook)
2509 #endif
2510 	    )) {
2511 		if (bridge_pfil(&m, sc->sc_ifp, NULL, PFIL_OUT) != 0)
2512 			goto out;
2513 		if (m == NULL)
2514 			goto out;
2515 	}
2516 
2517 	LIST_FOREACH(dbif, &sc->sc_iflist, bif_next) {
2518 		dst_if = dbif->bif_ifp;
2519 		if (dst_if == src_if)
2520 			continue;
2521 
2522 		/* Private segments can not talk to each other */
2523 		if (sbif && (sbif->bif_flags & dbif->bif_flags & IFBIF_PRIVATE))
2524 			continue;
2525 
2526 		if ((dbif->bif_flags & IFBIF_STP) &&
2527 		    dbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
2528 			continue;
2529 
2530 		if ((dbif->bif_flags & IFBIF_DISCOVER) == 0 &&
2531 		    (m->m_flags & (M_BCAST|M_MCAST)) == 0)
2532 			continue;
2533 
2534 		if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
2535 			continue;
2536 
2537 		if (LIST_NEXT(dbif, bif_next) == NULL) {
2538 			mc = m;
2539 			used = 1;
2540 		} else {
2541 			mc = m_dup(m, M_NOWAIT);
2542 			if (mc == NULL) {
2543 				if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
2544 				continue;
2545 			}
2546 		}
2547 
2548 		/*
2549 		 * Filter on the output interface. Pass a NULL bridge interface
2550 		 * pointer so we do not redundantly filter on the bridge for
2551 		 * each interface we broadcast on.
2552 		 */
2553 		if (runfilt && (PFIL_HOOKED(&V_inet_pfil_hook)
2554 #ifdef INET6
2555 		    || PFIL_HOOKED(&V_inet6_pfil_hook)
2556 #endif
2557 		    )) {
2558 			if (used == 0) {
2559 				/* Keep the layer3 header aligned */
2560 				i = min(mc->m_pkthdr.len, max_protohdr);
2561 				mc = m_copyup(mc, i, ETHER_ALIGN);
2562 				if (mc == NULL) {
2563 					if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
2564 					continue;
2565 				}
2566 			}
2567 			if (bridge_pfil(&mc, NULL, dst_if, PFIL_OUT) != 0)
2568 				continue;
2569 			if (mc == NULL)
2570 				continue;
2571 		}
2572 
2573 		bridge_enqueue(sc, dst_if, mc);
2574 	}
2575 	if (used == 0)
2576 		m_freem(m);
2577 
2578 out:
2579 	BRIDGE_UNREF(sc);
2580 }
2581 
2582 /*
2583  * bridge_span:
2584  *
2585  *	Duplicate a packet out one or more interfaces that are in span mode,
2586  *	the original mbuf is unmodified.
2587  */
2588 static void
2589 bridge_span(struct bridge_softc *sc, struct mbuf *m)
2590 {
2591 	struct bridge_iflist *bif;
2592 	struct ifnet *dst_if;
2593 	struct mbuf *mc;
2594 
2595 	if (LIST_EMPTY(&sc->sc_spanlist))
2596 		return;
2597 
2598 	LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) {
2599 		dst_if = bif->bif_ifp;
2600 
2601 		if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
2602 			continue;
2603 
2604 		mc = m_copypacket(m, M_NOWAIT);
2605 		if (mc == NULL) {
2606 			if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
2607 			continue;
2608 		}
2609 
2610 		bridge_enqueue(sc, dst_if, mc);
2611 	}
2612 }
2613 
2614 /*
2615  * bridge_rtupdate:
2616  *
2617  *	Add a bridge routing entry.
2618  */
2619 static int
2620 bridge_rtupdate(struct bridge_softc *sc, const uint8_t *dst, uint16_t vlan,
2621     struct bridge_iflist *bif, int setflags, uint8_t flags)
2622 {
2623 	struct bridge_rtnode *brt;
2624 	int error;
2625 
2626 	BRIDGE_LOCK_ASSERT(sc);
2627 
2628 	/* Check the source address is valid and not multicast. */
2629 	if (ETHER_IS_MULTICAST(dst) ||
2630 	    (dst[0] == 0 && dst[1] == 0 && dst[2] == 0 &&
2631 	     dst[3] == 0 && dst[4] == 0 && dst[5] == 0) != 0)
2632 		return (EINVAL);
2633 
2634 	/* 802.1p frames map to vlan 1 */
2635 	if (vlan == 0)
2636 		vlan = 1;
2637 
2638 	/*
2639 	 * A route for this destination might already exist.  If so,
2640 	 * update it, otherwise create a new one.
2641 	 */
2642 	if ((brt = bridge_rtnode_lookup(sc, dst, vlan)) == NULL) {
2643 		if (sc->sc_brtcnt >= sc->sc_brtmax) {
2644 			sc->sc_brtexceeded++;
2645 			return (ENOSPC);
2646 		}
2647 		/* Check per interface address limits (if enabled) */
2648 		if (bif->bif_addrmax && bif->bif_addrcnt >= bif->bif_addrmax) {
2649 			bif->bif_addrexceeded++;
2650 			return (ENOSPC);
2651 		}
2652 
2653 		/*
2654 		 * Allocate a new bridge forwarding node, and
2655 		 * initialize the expiration time and Ethernet
2656 		 * address.
2657 		 */
2658 		brt = uma_zalloc(bridge_rtnode_zone, M_NOWAIT | M_ZERO);
2659 		if (brt == NULL)
2660 			return (ENOMEM);
2661 
2662 		if (bif->bif_flags & IFBIF_STICKY)
2663 			brt->brt_flags = IFBAF_STICKY;
2664 		else
2665 			brt->brt_flags = IFBAF_DYNAMIC;
2666 
2667 		memcpy(brt->brt_addr, dst, ETHER_ADDR_LEN);
2668 		brt->brt_vlan = vlan;
2669 
2670 		if ((error = bridge_rtnode_insert(sc, brt)) != 0) {
2671 			uma_zfree(bridge_rtnode_zone, brt);
2672 			return (error);
2673 		}
2674 		brt->brt_dst = bif;
2675 		bif->bif_addrcnt++;
2676 	}
2677 
2678 	if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC &&
2679 	    brt->brt_dst != bif) {
2680 		brt->brt_dst->bif_addrcnt--;
2681 		brt->brt_dst = bif;
2682 		brt->brt_dst->bif_addrcnt++;
2683 	}
2684 
2685 	if ((flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
2686 		brt->brt_expire = time_uptime + sc->sc_brttimeout;
2687 	if (setflags)
2688 		brt->brt_flags = flags;
2689 
2690 	return (0);
2691 }
2692 
2693 /*
2694  * bridge_rtlookup:
2695  *
2696  *	Lookup the destination interface for an address.
2697  */
2698 static struct ifnet *
2699 bridge_rtlookup(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan)
2700 {
2701 	struct bridge_rtnode *brt;
2702 
2703 	BRIDGE_LOCK_ASSERT(sc);
2704 
2705 	if ((brt = bridge_rtnode_lookup(sc, addr, vlan)) == NULL)
2706 		return (NULL);
2707 
2708 	return (brt->brt_ifp);
2709 }
2710 
2711 /*
2712  * bridge_rttrim:
2713  *
2714  *	Trim the routine table so that we have a number
2715  *	of routing entries less than or equal to the
2716  *	maximum number.
2717  */
2718 static void
2719 bridge_rttrim(struct bridge_softc *sc)
2720 {
2721 	struct bridge_rtnode *brt, *nbrt;
2722 
2723 	BRIDGE_LOCK_ASSERT(sc);
2724 
2725 	/* Make sure we actually need to do this. */
2726 	if (sc->sc_brtcnt <= sc->sc_brtmax)
2727 		return;
2728 
2729 	/* Force an aging cycle; this might trim enough addresses. */
2730 	bridge_rtage(sc);
2731 	if (sc->sc_brtcnt <= sc->sc_brtmax)
2732 		return;
2733 
2734 	LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
2735 		if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) {
2736 			bridge_rtnode_destroy(sc, brt);
2737 			if (sc->sc_brtcnt <= sc->sc_brtmax)
2738 				return;
2739 		}
2740 	}
2741 }
2742 
2743 /*
2744  * bridge_timer:
2745  *
2746  *	Aging timer for the bridge.
2747  */
2748 static void
2749 bridge_timer(void *arg)
2750 {
2751 	struct bridge_softc *sc = arg;
2752 
2753 	BRIDGE_LOCK_ASSERT(sc);
2754 
2755 	bridge_rtage(sc);
2756 
2757 	if (sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING)
2758 		callout_reset(&sc->sc_brcallout,
2759 		    bridge_rtable_prune_period * hz, bridge_timer, sc);
2760 }
2761 
2762 /*
2763  * bridge_rtage:
2764  *
2765  *	Perform an aging cycle.
2766  */
2767 static void
2768 bridge_rtage(struct bridge_softc *sc)
2769 {
2770 	struct bridge_rtnode *brt, *nbrt;
2771 
2772 	BRIDGE_LOCK_ASSERT(sc);
2773 
2774 	LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
2775 		if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) {
2776 			if (time_uptime >= brt->brt_expire)
2777 				bridge_rtnode_destroy(sc, brt);
2778 		}
2779 	}
2780 }
2781 
2782 /*
2783  * bridge_rtflush:
2784  *
2785  *	Remove all dynamic addresses from the bridge.
2786  */
2787 static void
2788 bridge_rtflush(struct bridge_softc *sc, int full)
2789 {
2790 	struct bridge_rtnode *brt, *nbrt;
2791 
2792 	BRIDGE_LOCK_ASSERT(sc);
2793 
2794 	LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
2795 		if (full || (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
2796 			bridge_rtnode_destroy(sc, brt);
2797 	}
2798 }
2799 
2800 /*
2801  * bridge_rtdaddr:
2802  *
2803  *	Remove an address from the table.
2804  */
2805 static int
2806 bridge_rtdaddr(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan)
2807 {
2808 	struct bridge_rtnode *brt;
2809 	int found = 0;
2810 
2811 	BRIDGE_LOCK_ASSERT(sc);
2812 
2813 	/*
2814 	 * If vlan is zero then we want to delete for all vlans so the lookup
2815 	 * may return more than one.
2816 	 */
2817 	while ((brt = bridge_rtnode_lookup(sc, addr, vlan)) != NULL) {
2818 		bridge_rtnode_destroy(sc, brt);
2819 		found = 1;
2820 	}
2821 
2822 	return (found ? 0 : ENOENT);
2823 }
2824 
2825 /*
2826  * bridge_rtdelete:
2827  *
2828  *	Delete routes to a speicifc member interface.
2829  */
2830 static void
2831 bridge_rtdelete(struct bridge_softc *sc, struct ifnet *ifp, int full)
2832 {
2833 	struct bridge_rtnode *brt, *nbrt;
2834 
2835 	BRIDGE_LOCK_ASSERT(sc);
2836 
2837 	LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
2838 		if (brt->brt_ifp == ifp && (full ||
2839 			    (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC))
2840 			bridge_rtnode_destroy(sc, brt);
2841 	}
2842 }
2843 
2844 /*
2845  * bridge_rtable_init:
2846  *
2847  *	Initialize the route table for this bridge.
2848  */
2849 static void
2850 bridge_rtable_init(struct bridge_softc *sc)
2851 {
2852 	int i;
2853 
2854 	sc->sc_rthash = malloc(sizeof(*sc->sc_rthash) * BRIDGE_RTHASH_SIZE,
2855 	    M_DEVBUF, M_WAITOK);
2856 
2857 	for (i = 0; i < BRIDGE_RTHASH_SIZE; i++)
2858 		LIST_INIT(&sc->sc_rthash[i]);
2859 
2860 	sc->sc_rthash_key = arc4random();
2861 	LIST_INIT(&sc->sc_rtlist);
2862 }
2863 
2864 /*
2865  * bridge_rtable_fini:
2866  *
2867  *	Deconstruct the route table for this bridge.
2868  */
2869 static void
2870 bridge_rtable_fini(struct bridge_softc *sc)
2871 {
2872 
2873 	KASSERT(sc->sc_brtcnt == 0,
2874 	    ("%s: %d bridge routes referenced", __func__, sc->sc_brtcnt));
2875 	free(sc->sc_rthash, M_DEVBUF);
2876 }
2877 
2878 /*
2879  * The following hash function is adapted from "Hash Functions" by Bob Jenkins
2880  * ("Algorithm Alley", Dr. Dobbs Journal, September 1997).
2881  */
2882 #define	mix(a, b, c)							\
2883 do {									\
2884 	a -= b; a -= c; a ^= (c >> 13);					\
2885 	b -= c; b -= a; b ^= (a << 8);					\
2886 	c -= a; c -= b; c ^= (b >> 13);					\
2887 	a -= b; a -= c; a ^= (c >> 12);					\
2888 	b -= c; b -= a; b ^= (a << 16);					\
2889 	c -= a; c -= b; c ^= (b >> 5);					\
2890 	a -= b; a -= c; a ^= (c >> 3);					\
2891 	b -= c; b -= a; b ^= (a << 10);					\
2892 	c -= a; c -= b; c ^= (b >> 15);					\
2893 } while (/*CONSTCOND*/0)
2894 
2895 static __inline uint32_t
2896 bridge_rthash(struct bridge_softc *sc, const uint8_t *addr)
2897 {
2898 	uint32_t a = 0x9e3779b9, b = 0x9e3779b9, c = sc->sc_rthash_key;
2899 
2900 	b += addr[5] << 8;
2901 	b += addr[4];
2902 	a += addr[3] << 24;
2903 	a += addr[2] << 16;
2904 	a += addr[1] << 8;
2905 	a += addr[0];
2906 
2907 	mix(a, b, c);
2908 
2909 	return (c & BRIDGE_RTHASH_MASK);
2910 }
2911 
2912 #undef mix
2913 
2914 static int
2915 bridge_rtnode_addr_cmp(const uint8_t *a, const uint8_t *b)
2916 {
2917 	int i, d;
2918 
2919 	for (i = 0, d = 0; i < ETHER_ADDR_LEN && d == 0; i++) {
2920 		d = ((int)a[i]) - ((int)b[i]);
2921 	}
2922 
2923 	return (d);
2924 }
2925 
2926 /*
2927  * bridge_rtnode_lookup:
2928  *
2929  *	Look up a bridge route node for the specified destination. Compare the
2930  *	vlan id or if zero then just return the first match.
2931  */
2932 static struct bridge_rtnode *
2933 bridge_rtnode_lookup(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan)
2934 {
2935 	struct bridge_rtnode *brt;
2936 	uint32_t hash;
2937 	int dir;
2938 
2939 	BRIDGE_LOCK_ASSERT(sc);
2940 
2941 	hash = bridge_rthash(sc, addr);
2942 	LIST_FOREACH(brt, &sc->sc_rthash[hash], brt_hash) {
2943 		dir = bridge_rtnode_addr_cmp(addr, brt->brt_addr);
2944 		if (dir == 0 && (brt->brt_vlan == vlan || vlan == 0))
2945 			return (brt);
2946 		if (dir > 0)
2947 			return (NULL);
2948 	}
2949 
2950 	return (NULL);
2951 }
2952 
2953 /*
2954  * bridge_rtnode_insert:
2955  *
2956  *	Insert the specified bridge node into the route table.  We
2957  *	assume the entry is not already in the table.
2958  */
2959 static int
2960 bridge_rtnode_insert(struct bridge_softc *sc, struct bridge_rtnode *brt)
2961 {
2962 	struct bridge_rtnode *lbrt;
2963 	uint32_t hash;
2964 	int dir;
2965 
2966 	BRIDGE_LOCK_ASSERT(sc);
2967 
2968 	hash = bridge_rthash(sc, brt->brt_addr);
2969 
2970 	lbrt = LIST_FIRST(&sc->sc_rthash[hash]);
2971 	if (lbrt == NULL) {
2972 		LIST_INSERT_HEAD(&sc->sc_rthash[hash], brt, brt_hash);
2973 		goto out;
2974 	}
2975 
2976 	do {
2977 		dir = bridge_rtnode_addr_cmp(brt->brt_addr, lbrt->brt_addr);
2978 		if (dir == 0 && brt->brt_vlan == lbrt->brt_vlan)
2979 			return (EEXIST);
2980 		if (dir > 0) {
2981 			LIST_INSERT_BEFORE(lbrt, brt, brt_hash);
2982 			goto out;
2983 		}
2984 		if (LIST_NEXT(lbrt, brt_hash) == NULL) {
2985 			LIST_INSERT_AFTER(lbrt, brt, brt_hash);
2986 			goto out;
2987 		}
2988 		lbrt = LIST_NEXT(lbrt, brt_hash);
2989 	} while (lbrt != NULL);
2990 
2991 #ifdef DIAGNOSTIC
2992 	panic("bridge_rtnode_insert: impossible");
2993 #endif
2994 
2995 out:
2996 	LIST_INSERT_HEAD(&sc->sc_rtlist, brt, brt_list);
2997 	sc->sc_brtcnt++;
2998 
2999 	return (0);
3000 }
3001 
3002 /*
3003  * bridge_rtnode_destroy:
3004  *
3005  *	Destroy a bridge rtnode.
3006  */
3007 static void
3008 bridge_rtnode_destroy(struct bridge_softc *sc, struct bridge_rtnode *brt)
3009 {
3010 	BRIDGE_LOCK_ASSERT(sc);
3011 
3012 	LIST_REMOVE(brt, brt_hash);
3013 
3014 	LIST_REMOVE(brt, brt_list);
3015 	sc->sc_brtcnt--;
3016 	brt->brt_dst->bif_addrcnt--;
3017 	uma_zfree(bridge_rtnode_zone, brt);
3018 }
3019 
3020 /*
3021  * bridge_rtable_expire:
3022  *
3023  *	Set the expiry time for all routes on an interface.
3024  */
3025 static void
3026 bridge_rtable_expire(struct ifnet *ifp, int age)
3027 {
3028 	struct bridge_softc *sc = ifp->if_bridge;
3029 	struct bridge_rtnode *brt;
3030 
3031 	BRIDGE_LOCK(sc);
3032 
3033 	/*
3034 	 * If the age is zero then flush, otherwise set all the expiry times to
3035 	 * age for the interface
3036 	 */
3037 	if (age == 0)
3038 		bridge_rtdelete(sc, ifp, IFBF_FLUSHDYN);
3039 	else {
3040 		LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) {
3041 			/* Cap the expiry time to 'age' */
3042 			if (brt->brt_ifp == ifp &&
3043 			    brt->brt_expire > time_uptime + age &&
3044 			    (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
3045 				brt->brt_expire = time_uptime + age;
3046 		}
3047 	}
3048 	BRIDGE_UNLOCK(sc);
3049 }
3050 
3051 /*
3052  * bridge_state_change:
3053  *
3054  *	Callback from the bridgestp code when a port changes states.
3055  */
3056 static void
3057 bridge_state_change(struct ifnet *ifp, int state)
3058 {
3059 	struct bridge_softc *sc = ifp->if_bridge;
3060 	static const char *stpstates[] = {
3061 		"disabled",
3062 		"listening",
3063 		"learning",
3064 		"forwarding",
3065 		"blocking",
3066 		"discarding"
3067 	};
3068 
3069 	CURVNET_SET(ifp->if_vnet);
3070 	if (V_log_stp)
3071 		log(LOG_NOTICE, "%s: state changed to %s on %s\n",
3072 		    sc->sc_ifp->if_xname, stpstates[state], ifp->if_xname);
3073 	CURVNET_RESTORE();
3074 }
3075 
3076 /*
3077  * Send bridge packets through pfil if they are one of the types pfil can deal
3078  * with, or if they are ARP or REVARP.  (pfil will pass ARP and REVARP without
3079  * question.) If *bifp or *ifp are NULL then packet filtering is skipped for
3080  * that interface.
3081  */
3082 static int
3083 bridge_pfil(struct mbuf **mp, struct ifnet *bifp, struct ifnet *ifp, int dir)
3084 {
3085 	int snap, error, i, hlen;
3086 	struct ether_header *eh1, eh2;
3087 	struct ip *ip;
3088 	struct llc llc1;
3089 	u_int16_t ether_type;
3090 
3091 	snap = 0;
3092 	error = -1;	/* Default error if not error == 0 */
3093 
3094 #if 0
3095 	/* we may return with the IP fields swapped, ensure its not shared */
3096 	KASSERT(M_WRITABLE(*mp), ("%s: modifying a shared mbuf", __func__));
3097 #endif
3098 
3099 	if (V_pfil_bridge == 0 && V_pfil_member == 0 && V_pfil_ipfw == 0)
3100 		return (0); /* filtering is disabled */
3101 
3102 	i = min((*mp)->m_pkthdr.len, max_protohdr);
3103 	if ((*mp)->m_len < i) {
3104 	    *mp = m_pullup(*mp, i);
3105 	    if (*mp == NULL) {
3106 		printf("%s: m_pullup failed\n", __func__);
3107 		return (-1);
3108 	    }
3109 	}
3110 
3111 	eh1 = mtod(*mp, struct ether_header *);
3112 	ether_type = ntohs(eh1->ether_type);
3113 
3114 	/*
3115 	 * Check for SNAP/LLC.
3116 	 */
3117 	if (ether_type < ETHERMTU) {
3118 		struct llc *llc2 = (struct llc *)(eh1 + 1);
3119 
3120 		if ((*mp)->m_len >= ETHER_HDR_LEN + 8 &&
3121 		    llc2->llc_dsap == LLC_SNAP_LSAP &&
3122 		    llc2->llc_ssap == LLC_SNAP_LSAP &&
3123 		    llc2->llc_control == LLC_UI) {
3124 			ether_type = htons(llc2->llc_un.type_snap.ether_type);
3125 			snap = 1;
3126 		}
3127 	}
3128 
3129 	/*
3130 	 * If we're trying to filter bridge traffic, don't look at anything
3131 	 * other than IP and ARP traffic.  If the filter doesn't understand
3132 	 * IPv6, don't allow IPv6 through the bridge either.  This is lame
3133 	 * since if we really wanted, say, an AppleTalk filter, we are hosed,
3134 	 * but of course we don't have an AppleTalk filter to begin with.
3135 	 * (Note that since pfil doesn't understand ARP it will pass *ALL*
3136 	 * ARP traffic.)
3137 	 */
3138 	switch (ether_type) {
3139 		case ETHERTYPE_ARP:
3140 		case ETHERTYPE_REVARP:
3141 			if (V_pfil_ipfw_arp == 0)
3142 				return (0); /* Automatically pass */
3143 			break;
3144 
3145 		case ETHERTYPE_IP:
3146 #ifdef INET6
3147 		case ETHERTYPE_IPV6:
3148 #endif /* INET6 */
3149 			break;
3150 		default:
3151 			/*
3152 			 * Check to see if the user wants to pass non-ip
3153 			 * packets, these will not be checked by pfil(9) and
3154 			 * passed unconditionally so the default is to drop.
3155 			 */
3156 			if (V_pfil_onlyip)
3157 				goto bad;
3158 	}
3159 
3160 	/* Run the packet through pfil before stripping link headers */
3161 	if (PFIL_HOOKED(&V_link_pfil_hook) && V_pfil_ipfw != 0 &&
3162 			dir == PFIL_OUT && ifp != NULL) {
3163 
3164 		error = pfil_run_hooks(&V_link_pfil_hook, mp, ifp, dir, NULL);
3165 
3166 		if (*mp == NULL || error != 0) /* packet consumed by filter */
3167 			return (error);
3168 	}
3169 
3170 	/* Strip off the Ethernet header and keep a copy. */
3171 	m_copydata(*mp, 0, ETHER_HDR_LEN, (caddr_t) &eh2);
3172 	m_adj(*mp, ETHER_HDR_LEN);
3173 
3174 	/* Strip off snap header, if present */
3175 	if (snap) {
3176 		m_copydata(*mp, 0, sizeof(struct llc), (caddr_t) &llc1);
3177 		m_adj(*mp, sizeof(struct llc));
3178 	}
3179 
3180 	/*
3181 	 * Check the IP header for alignment and errors
3182 	 */
3183 	if (dir == PFIL_IN) {
3184 		switch (ether_type) {
3185 			case ETHERTYPE_IP:
3186 				error = bridge_ip_checkbasic(mp);
3187 				break;
3188 #ifdef INET6
3189 			case ETHERTYPE_IPV6:
3190 				error = bridge_ip6_checkbasic(mp);
3191 				break;
3192 #endif /* INET6 */
3193 			default:
3194 				error = 0;
3195 		}
3196 		if (error)
3197 			goto bad;
3198 	}
3199 
3200 	error = 0;
3201 
3202 	/*
3203 	 * Run the packet through pfil
3204 	 */
3205 	switch (ether_type) {
3206 	case ETHERTYPE_IP:
3207 		/*
3208 		 * Run pfil on the member interface and the bridge, both can
3209 		 * be skipped by clearing pfil_member or pfil_bridge.
3210 		 *
3211 		 * Keep the order:
3212 		 *   in_if -> bridge_if -> out_if
3213 		 */
3214 		if (V_pfil_bridge && dir == PFIL_OUT && bifp != NULL)
3215 			error = pfil_run_hooks(&V_inet_pfil_hook, mp, bifp,
3216 					dir, NULL);
3217 
3218 		if (*mp == NULL || error != 0) /* filter may consume */
3219 			break;
3220 
3221 		if (V_pfil_member && ifp != NULL)
3222 			error = pfil_run_hooks(&V_inet_pfil_hook, mp, ifp,
3223 					dir, NULL);
3224 
3225 		if (*mp == NULL || error != 0) /* filter may consume */
3226 			break;
3227 
3228 		if (V_pfil_bridge && dir == PFIL_IN && bifp != NULL)
3229 			error = pfil_run_hooks(&V_inet_pfil_hook, mp, bifp,
3230 					dir, NULL);
3231 
3232 		if (*mp == NULL || error != 0) /* filter may consume */
3233 			break;
3234 
3235 		/* check if we need to fragment the packet */
3236 		if (V_pfil_member && ifp != NULL && dir == PFIL_OUT) {
3237 			i = (*mp)->m_pkthdr.len;
3238 			if (i > ifp->if_mtu) {
3239 				error = bridge_fragment(ifp, *mp, &eh2, snap,
3240 					    &llc1);
3241 				return (error);
3242 			}
3243 		}
3244 
3245 		/* Recalculate the ip checksum. */
3246 		ip = mtod(*mp, struct ip *);
3247 		hlen = ip->ip_hl << 2;
3248 		if (hlen < sizeof(struct ip))
3249 			goto bad;
3250 		if (hlen > (*mp)->m_len) {
3251 			if ((*mp = m_pullup(*mp, hlen)) == 0)
3252 				goto bad;
3253 			ip = mtod(*mp, struct ip *);
3254 			if (ip == NULL)
3255 				goto bad;
3256 		}
3257 		ip->ip_sum = 0;
3258 		if (hlen == sizeof(struct ip))
3259 			ip->ip_sum = in_cksum_hdr(ip);
3260 		else
3261 			ip->ip_sum = in_cksum(*mp, hlen);
3262 
3263 		break;
3264 #ifdef INET6
3265 	case ETHERTYPE_IPV6:
3266 		if (V_pfil_bridge && dir == PFIL_OUT && bifp != NULL)
3267 			error = pfil_run_hooks(&V_inet6_pfil_hook, mp, bifp,
3268 					dir, NULL);
3269 
3270 		if (*mp == NULL || error != 0) /* filter may consume */
3271 			break;
3272 
3273 		if (V_pfil_member && ifp != NULL)
3274 			error = pfil_run_hooks(&V_inet6_pfil_hook, mp, ifp,
3275 					dir, NULL);
3276 
3277 		if (*mp == NULL || error != 0) /* filter may consume */
3278 			break;
3279 
3280 		if (V_pfil_bridge && dir == PFIL_IN && bifp != NULL)
3281 			error = pfil_run_hooks(&V_inet6_pfil_hook, mp, bifp,
3282 					dir, NULL);
3283 		break;
3284 #endif
3285 	default:
3286 		error = 0;
3287 		break;
3288 	}
3289 
3290 	if (*mp == NULL)
3291 		return (error);
3292 	if (error != 0)
3293 		goto bad;
3294 
3295 	error = -1;
3296 
3297 	/*
3298 	 * Finally, put everything back the way it was and return
3299 	 */
3300 	if (snap) {
3301 		M_PREPEND(*mp, sizeof(struct llc), M_NOWAIT);
3302 		if (*mp == NULL)
3303 			return (error);
3304 		bcopy(&llc1, mtod(*mp, caddr_t), sizeof(struct llc));
3305 	}
3306 
3307 	M_PREPEND(*mp, ETHER_HDR_LEN, M_NOWAIT);
3308 	if (*mp == NULL)
3309 		return (error);
3310 	bcopy(&eh2, mtod(*mp, caddr_t), ETHER_HDR_LEN);
3311 
3312 	return (0);
3313 
3314 bad:
3315 	m_freem(*mp);
3316 	*mp = NULL;
3317 	return (error);
3318 }
3319 
3320 /*
3321  * Perform basic checks on header size since
3322  * pfil assumes ip_input has already processed
3323  * it for it.  Cut-and-pasted from ip_input.c.
3324  * Given how simple the IPv6 version is,
3325  * does the IPv4 version really need to be
3326  * this complicated?
3327  *
3328  * XXX Should we update ipstat here, or not?
3329  * XXX Right now we update ipstat but not
3330  * XXX csum_counter.
3331  */
3332 static int
3333 bridge_ip_checkbasic(struct mbuf **mp)
3334 {
3335 	struct mbuf *m = *mp;
3336 	struct ip *ip;
3337 	int len, hlen;
3338 	u_short sum;
3339 
3340 	if (*mp == NULL)
3341 		return (-1);
3342 
3343 	if (IP_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) {
3344 		if ((m = m_copyup(m, sizeof(struct ip),
3345 			(max_linkhdr + 3) & ~3)) == NULL) {
3346 			/* XXXJRT new stat, please */
3347 			KMOD_IPSTAT_INC(ips_toosmall);
3348 			goto bad;
3349 		}
3350 	} else if (__predict_false(m->m_len < sizeof (struct ip))) {
3351 		if ((m = m_pullup(m, sizeof (struct ip))) == NULL) {
3352 			KMOD_IPSTAT_INC(ips_toosmall);
3353 			goto bad;
3354 		}
3355 	}
3356 	ip = mtod(m, struct ip *);
3357 	if (ip == NULL) goto bad;
3358 
3359 	if (ip->ip_v != IPVERSION) {
3360 		KMOD_IPSTAT_INC(ips_badvers);
3361 		goto bad;
3362 	}
3363 	hlen = ip->ip_hl << 2;
3364 	if (hlen < sizeof(struct ip)) { /* minimum header length */
3365 		KMOD_IPSTAT_INC(ips_badhlen);
3366 		goto bad;
3367 	}
3368 	if (hlen > m->m_len) {
3369 		if ((m = m_pullup(m, hlen)) == 0) {
3370 			KMOD_IPSTAT_INC(ips_badhlen);
3371 			goto bad;
3372 		}
3373 		ip = mtod(m, struct ip *);
3374 		if (ip == NULL) goto bad;
3375 	}
3376 
3377 	if (m->m_pkthdr.csum_flags & CSUM_IP_CHECKED) {
3378 		sum = !(m->m_pkthdr.csum_flags & CSUM_IP_VALID);
3379 	} else {
3380 		if (hlen == sizeof(struct ip)) {
3381 			sum = in_cksum_hdr(ip);
3382 		} else {
3383 			sum = in_cksum(m, hlen);
3384 		}
3385 	}
3386 	if (sum) {
3387 		KMOD_IPSTAT_INC(ips_badsum);
3388 		goto bad;
3389 	}
3390 
3391 	/* Retrieve the packet length. */
3392 	len = ntohs(ip->ip_len);
3393 
3394 	/*
3395 	 * Check for additional length bogosity
3396 	 */
3397 	if (len < hlen) {
3398 		KMOD_IPSTAT_INC(ips_badlen);
3399 		goto bad;
3400 	}
3401 
3402 	/*
3403 	 * Check that the amount of data in the buffers
3404 	 * is as at least much as the IP header would have us expect.
3405 	 * Drop packet if shorter than we expect.
3406 	 */
3407 	if (m->m_pkthdr.len < len) {
3408 		KMOD_IPSTAT_INC(ips_tooshort);
3409 		goto bad;
3410 	}
3411 
3412 	/* Checks out, proceed */
3413 	*mp = m;
3414 	return (0);
3415 
3416 bad:
3417 	*mp = m;
3418 	return (-1);
3419 }
3420 
3421 #ifdef INET6
3422 /*
3423  * Same as above, but for IPv6.
3424  * Cut-and-pasted from ip6_input.c.
3425  * XXX Should we update ip6stat, or not?
3426  */
3427 static int
3428 bridge_ip6_checkbasic(struct mbuf **mp)
3429 {
3430 	struct mbuf *m = *mp;
3431 	struct ip6_hdr *ip6;
3432 
3433 	/*
3434 	 * If the IPv6 header is not aligned, slurp it up into a new
3435 	 * mbuf with space for link headers, in the event we forward
3436 	 * it.  Otherwise, if it is aligned, make sure the entire base
3437 	 * IPv6 header is in the first mbuf of the chain.
3438 	 */
3439 	if (IP6_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) {
3440 		struct ifnet *inifp = m->m_pkthdr.rcvif;
3441 		if ((m = m_copyup(m, sizeof(struct ip6_hdr),
3442 			    (max_linkhdr + 3) & ~3)) == NULL) {
3443 			/* XXXJRT new stat, please */
3444 			IP6STAT_INC(ip6s_toosmall);
3445 			in6_ifstat_inc(inifp, ifs6_in_hdrerr);
3446 			goto bad;
3447 		}
3448 	} else if (__predict_false(m->m_len < sizeof(struct ip6_hdr))) {
3449 		struct ifnet *inifp = m->m_pkthdr.rcvif;
3450 		if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) {
3451 			IP6STAT_INC(ip6s_toosmall);
3452 			in6_ifstat_inc(inifp, ifs6_in_hdrerr);
3453 			goto bad;
3454 		}
3455 	}
3456 
3457 	ip6 = mtod(m, struct ip6_hdr *);
3458 
3459 	if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) {
3460 		IP6STAT_INC(ip6s_badvers);
3461 		in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr);
3462 		goto bad;
3463 	}
3464 
3465 	/* Checks out, proceed */
3466 	*mp = m;
3467 	return (0);
3468 
3469 bad:
3470 	*mp = m;
3471 	return (-1);
3472 }
3473 #endif /* INET6 */
3474 
3475 /*
3476  * bridge_fragment:
3477  *
3478  *	Return a fragmented mbuf chain.
3479  */
3480 static int
3481 bridge_fragment(struct ifnet *ifp, struct mbuf *m, struct ether_header *eh,
3482     int snap, struct llc *llc)
3483 {
3484 	struct mbuf *m0;
3485 	struct ip *ip;
3486 	int error = -1;
3487 
3488 	if (m->m_len < sizeof(struct ip) &&
3489 	    (m = m_pullup(m, sizeof(struct ip))) == NULL)
3490 		goto out;
3491 	ip = mtod(m, struct ip *);
3492 
3493 	m->m_pkthdr.csum_flags |= CSUM_IP;
3494 	error = ip_fragment(ip, &m, ifp->if_mtu, ifp->if_hwassist);
3495 	if (error)
3496 		goto out;
3497 
3498 	/* walk the chain and re-add the Ethernet header */
3499 	for (m0 = m; m0; m0 = m0->m_nextpkt) {
3500 		if (error == 0) {
3501 			if (snap) {
3502 				M_PREPEND(m0, sizeof(struct llc), M_NOWAIT);
3503 				if (m0 == NULL) {
3504 					error = ENOBUFS;
3505 					continue;
3506 				}
3507 				bcopy(llc, mtod(m0, caddr_t),
3508 				    sizeof(struct llc));
3509 			}
3510 			M_PREPEND(m0, ETHER_HDR_LEN, M_NOWAIT);
3511 			if (m0 == NULL) {
3512 				error = ENOBUFS;
3513 				continue;
3514 			}
3515 			bcopy(eh, mtod(m0, caddr_t), ETHER_HDR_LEN);
3516 		} else
3517 			m_freem(m);
3518 	}
3519 
3520 	if (error == 0)
3521 		KMOD_IPSTAT_INC(ips_fragmented);
3522 
3523 	return (error);
3524 
3525 out:
3526 	if (m != NULL)
3527 		m_freem(m);
3528 	return (error);
3529 }
3530 
3531 static void
3532 bridge_linkstate(struct ifnet *ifp)
3533 {
3534 	struct bridge_softc *sc = ifp->if_bridge;
3535 	struct bridge_iflist *bif;
3536 
3537 	BRIDGE_LOCK(sc);
3538 	bif = bridge_lookup_member_if(sc, ifp);
3539 	if (bif == NULL) {
3540 		BRIDGE_UNLOCK(sc);
3541 		return;
3542 	}
3543 	bridge_linkcheck(sc);
3544 	BRIDGE_UNLOCK(sc);
3545 
3546 	bstp_linkstate(&bif->bif_stp);
3547 }
3548 
3549 static void
3550 bridge_linkcheck(struct bridge_softc *sc)
3551 {
3552 	struct bridge_iflist *bif;
3553 	int new_link, hasls;
3554 
3555 	BRIDGE_LOCK_ASSERT(sc);
3556 	new_link = LINK_STATE_DOWN;
3557 	hasls = 0;
3558 	/* Our link is considered up if at least one of our ports is active */
3559 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
3560 		if (bif->bif_ifp->if_capabilities & IFCAP_LINKSTATE)
3561 			hasls++;
3562 		if (bif->bif_ifp->if_link_state == LINK_STATE_UP) {
3563 			new_link = LINK_STATE_UP;
3564 			break;
3565 		}
3566 	}
3567 	if (!LIST_EMPTY(&sc->sc_iflist) && !hasls) {
3568 		/* If no interfaces support link-state then we default to up */
3569 		new_link = LINK_STATE_UP;
3570 	}
3571 	if_link_state_change(sc->sc_ifp, new_link);
3572 }
3573