Home | History | Annotate | Line # | Download | only in net
if_bridge.c revision 1.91.2.2
      1 /*	$NetBSD: if_bridge.c,v 1.91.2.2 2015/06/06 14:40:25 skrll 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 (at) 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  * 3. All advertising materials mentioning features or use of this software
     51  *    must display the following acknowledgement:
     52  *	This product includes software developed by Jason L. Wright
     53  * 4. The name of the author may not be used to endorse or promote products
     54  *    derived from this software without specific prior written permission.
     55  *
     56  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
     57  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
     58  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
     59  * DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT,
     60  * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
     61  * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
     62  * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     63  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
     64  * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
     65  * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
     66  * POSSIBILITY OF SUCH DAMAGE.
     67  *
     68  * OpenBSD: if_bridge.c,v 1.60 2001/06/15 03:38:33 itojun Exp
     69  */
     70 
     71 /*
     72  * Network interface bridge support.
     73  *
     74  * TODO:
     75  *
     76  *	- Currently only supports Ethernet-like interfaces (Ethernet,
     77  *	  802.11, VLANs on Ethernet, etc.)  Figure out a nice way
     78  *	  to bridge other types of interfaces (FDDI-FDDI, and maybe
     79  *	  consider heterogenous bridges).
     80  */
     81 
     82 #include <sys/cdefs.h>
     83 __KERNEL_RCSID(0, "$NetBSD: if_bridge.c,v 1.91.2.2 2015/06/06 14:40:25 skrll Exp $");
     84 
     85 #ifdef _KERNEL_OPT
     86 #include "opt_bridge_ipf.h"
     87 #include "opt_inet.h"
     88 #endif /* _KERNEL_OPT */
     89 
     90 #include <sys/param.h>
     91 #include <sys/kernel.h>
     92 #include <sys/mbuf.h>
     93 #include <sys/queue.h>
     94 #include <sys/socket.h>
     95 #include <sys/socketvar.h> /* for softnet_lock */
     96 #include <sys/sockio.h>
     97 #include <sys/systm.h>
     98 #include <sys/proc.h>
     99 #include <sys/pool.h>
    100 #include <sys/kauth.h>
    101 #include <sys/cpu.h>
    102 #include <sys/cprng.h>
    103 #include <sys/mutex.h>
    104 #include <sys/kmem.h>
    105 
    106 #include <net/bpf.h>
    107 #include <net/if.h>
    108 #include <net/if_dl.h>
    109 #include <net/if_types.h>
    110 #include <net/if_llc.h>
    111 #include <net/pktqueue.h>
    112 
    113 #include <net/if_ether.h>
    114 #include <net/if_bridgevar.h>
    115 
    116 #if defined(BRIDGE_IPF)
    117 /* Used for bridge_ip[6]_checkbasic */
    118 #include <netinet/in.h>
    119 #include <netinet/in_systm.h>
    120 #include <netinet/ip.h>
    121 #include <netinet/ip_var.h>
    122 #include <netinet/ip_private.h>		/* XXX */
    123 
    124 #include <netinet/ip6.h>
    125 #include <netinet6/in6_var.h>
    126 #include <netinet6/ip6_var.h>
    127 #include <netinet6/ip6_private.h>	/* XXX */
    128 #endif /* BRIDGE_IPF */
    129 
    130 /*
    131  * Size of the route hash table.  Must be a power of two.
    132  */
    133 #ifndef BRIDGE_RTHASH_SIZE
    134 #define	BRIDGE_RTHASH_SIZE		1024
    135 #endif
    136 
    137 #define	BRIDGE_RTHASH_MASK		(BRIDGE_RTHASH_SIZE - 1)
    138 
    139 #include "carp.h"
    140 #if NCARP > 0
    141 #include <netinet/in.h>
    142 #include <netinet/in_var.h>
    143 #include <netinet/ip_carp.h>
    144 #endif
    145 
    146 __CTASSERT(sizeof(struct ifbifconf) == sizeof(struct ifbaconf));
    147 __CTASSERT(offsetof(struct ifbifconf, ifbic_len) == offsetof(struct ifbaconf, ifbac_len));
    148 __CTASSERT(offsetof(struct ifbifconf, ifbic_buf) == offsetof(struct ifbaconf, ifbac_buf));
    149 
    150 /*
    151  * Maximum number of addresses to cache.
    152  */
    153 #ifndef BRIDGE_RTABLE_MAX
    154 #define	BRIDGE_RTABLE_MAX		100
    155 #endif
    156 
    157 /*
    158  * Spanning tree defaults.
    159  */
    160 #define	BSTP_DEFAULT_MAX_AGE		(20 * 256)
    161 #define	BSTP_DEFAULT_HELLO_TIME		(2 * 256)
    162 #define	BSTP_DEFAULT_FORWARD_DELAY	(15 * 256)
    163 #define	BSTP_DEFAULT_HOLD_TIME		(1 * 256)
    164 #define	BSTP_DEFAULT_BRIDGE_PRIORITY	0x8000
    165 #define	BSTP_DEFAULT_PORT_PRIORITY	0x80
    166 #define	BSTP_DEFAULT_PATH_COST		55
    167 
    168 /*
    169  * Timeout (in seconds) for entries learned dynamically.
    170  */
    171 #ifndef BRIDGE_RTABLE_TIMEOUT
    172 #define	BRIDGE_RTABLE_TIMEOUT		(20 * 60)	/* same as ARP */
    173 #endif
    174 
    175 /*
    176  * Number of seconds between walks of the route list.
    177  */
    178 #ifndef BRIDGE_RTABLE_PRUNE_PERIOD
    179 #define	BRIDGE_RTABLE_PRUNE_PERIOD	(5 * 60)
    180 #endif
    181 
    182 #define BRIDGE_RT_INTR_LOCK(_sc)	mutex_enter((_sc)->sc_rtlist_intr_lock)
    183 #define BRIDGE_RT_INTR_UNLOCK(_sc)	mutex_exit((_sc)->sc_rtlist_intr_lock)
    184 #define BRIDGE_RT_INTR_LOCKED(_sc)	mutex_owned((_sc)->sc_rtlist_intr_lock)
    185 
    186 #define BRIDGE_RT_LOCK(_sc)	if ((_sc)->sc_rtlist_lock) \
    187 					mutex_enter((_sc)->sc_rtlist_lock)
    188 #define BRIDGE_RT_UNLOCK(_sc)	if ((_sc)->sc_rtlist_lock) \
    189 					mutex_exit((_sc)->sc_rtlist_lock)
    190 #define BRIDGE_RT_LOCKED(_sc)	(!(_sc)->sc_rtlist_lock || \
    191 				 mutex_owned((_sc)->sc_rtlist_lock))
    192 
    193 #define BRIDGE_RT_PSZ_PERFORM(_sc) \
    194 				if ((_sc)->sc_rtlist_psz != NULL) \
    195 					pserialize_perform((_sc)->sc_rtlist_psz);
    196 
    197 #ifdef BRIDGE_MPSAFE
    198 #define BRIDGE_RT_RENTER(__s)	do { \
    199 					if (!cpu_intr_p()) \
    200 						__s = pserialize_read_enter(); \
    201 					else \
    202 						__s = splhigh(); \
    203 				} while (0)
    204 #define BRIDGE_RT_REXIT(__s)	do { \
    205 					if (!cpu_intr_p()) \
    206 						pserialize_read_exit(__s); \
    207 					else \
    208 						splx(__s); \
    209 				} while (0)
    210 #else /* BRIDGE_MPSAFE */
    211 #define BRIDGE_RT_RENTER(__s)	do { __s = 0; } while (0)
    212 #define BRIDGE_RT_REXIT(__s)	do { (void)__s; } while (0)
    213 #endif /* BRIDGE_MPSAFE */
    214 
    215 int	bridge_rtable_prune_period = BRIDGE_RTABLE_PRUNE_PERIOD;
    216 
    217 static struct pool bridge_rtnode_pool;
    218 static struct work bridge_rtage_wk;
    219 
    220 void	bridgeattach(int);
    221 
    222 static int	bridge_clone_create(struct if_clone *, int);
    223 static int	bridge_clone_destroy(struct ifnet *);
    224 
    225 static int	bridge_ioctl(struct ifnet *, u_long, void *);
    226 static int	bridge_init(struct ifnet *);
    227 static void	bridge_stop(struct ifnet *, int);
    228 static void	bridge_start(struct ifnet *);
    229 
    230 static void	bridge_input(struct ifnet *, struct mbuf *);
    231 static void	bridge_forward(void *);
    232 
    233 static void	bridge_timer(void *);
    234 
    235 static void	bridge_broadcast(struct bridge_softc *, struct ifnet *,
    236 				 struct mbuf *);
    237 
    238 static int	bridge_rtupdate(struct bridge_softc *, const uint8_t *,
    239 				struct ifnet *, int, uint8_t);
    240 static struct ifnet *bridge_rtlookup(struct bridge_softc *, const uint8_t *);
    241 static void	bridge_rttrim(struct bridge_softc *);
    242 static void	bridge_rtage(struct bridge_softc *);
    243 static void	bridge_rtage_work(struct work *, void *);
    244 static void	bridge_rtflush(struct bridge_softc *, int);
    245 static int	bridge_rtdaddr(struct bridge_softc *, const uint8_t *);
    246 static void	bridge_rtdelete(struct bridge_softc *, struct ifnet *ifp);
    247 
    248 static void	bridge_rtable_init(struct bridge_softc *);
    249 static void	bridge_rtable_fini(struct bridge_softc *);
    250 
    251 static struct bridge_rtnode *bridge_rtnode_lookup(struct bridge_softc *,
    252 						  const uint8_t *);
    253 static int	bridge_rtnode_insert(struct bridge_softc *,
    254 				     struct bridge_rtnode *);
    255 static void	bridge_rtnode_remove(struct bridge_softc *,
    256 				     struct bridge_rtnode *);
    257 static void	bridge_rtnode_destroy(struct bridge_rtnode *);
    258 
    259 static struct bridge_iflist *bridge_lookup_member(struct bridge_softc *,
    260 						  const char *name);
    261 static struct bridge_iflist *bridge_lookup_member_if(struct bridge_softc *,
    262 						     struct ifnet *ifp);
    263 static void	bridge_release_member(struct bridge_softc *, struct bridge_iflist *);
    264 static void	bridge_delete_member(struct bridge_softc *,
    265 				     struct bridge_iflist *);
    266 static struct bridge_iflist *bridge_try_hold_bif(struct bridge_iflist *);
    267 
    268 static int	bridge_ioctl_add(struct bridge_softc *, void *);
    269 static int	bridge_ioctl_del(struct bridge_softc *, void *);
    270 static int	bridge_ioctl_gifflags(struct bridge_softc *, void *);
    271 static int	bridge_ioctl_sifflags(struct bridge_softc *, void *);
    272 static int	bridge_ioctl_scache(struct bridge_softc *, void *);
    273 static int	bridge_ioctl_gcache(struct bridge_softc *, void *);
    274 static int	bridge_ioctl_gifs(struct bridge_softc *, void *);
    275 static int	bridge_ioctl_rts(struct bridge_softc *, void *);
    276 static int	bridge_ioctl_saddr(struct bridge_softc *, void *);
    277 static int	bridge_ioctl_sto(struct bridge_softc *, void *);
    278 static int	bridge_ioctl_gto(struct bridge_softc *, void *);
    279 static int	bridge_ioctl_daddr(struct bridge_softc *, void *);
    280 static int	bridge_ioctl_flush(struct bridge_softc *, void *);
    281 static int	bridge_ioctl_gpri(struct bridge_softc *, void *);
    282 static int	bridge_ioctl_spri(struct bridge_softc *, void *);
    283 static int	bridge_ioctl_ght(struct bridge_softc *, void *);
    284 static int	bridge_ioctl_sht(struct bridge_softc *, void *);
    285 static int	bridge_ioctl_gfd(struct bridge_softc *, void *);
    286 static int	bridge_ioctl_sfd(struct bridge_softc *, void *);
    287 static int	bridge_ioctl_gma(struct bridge_softc *, void *);
    288 static int	bridge_ioctl_sma(struct bridge_softc *, void *);
    289 static int	bridge_ioctl_sifprio(struct bridge_softc *, void *);
    290 static int	bridge_ioctl_sifcost(struct bridge_softc *, void *);
    291 #if defined(BRIDGE_IPF)
    292 static int	bridge_ioctl_gfilt(struct bridge_softc *, void *);
    293 static int	bridge_ioctl_sfilt(struct bridge_softc *, void *);
    294 static int	bridge_ipf(void *, struct mbuf **, struct ifnet *, int);
    295 static int	bridge_ip_checkbasic(struct mbuf **mp);
    296 # ifdef INET6
    297 static int	bridge_ip6_checkbasic(struct mbuf **mp);
    298 # endif /* INET6 */
    299 #endif /* BRIDGE_IPF */
    300 
    301 static void bridge_sysctl_fwdq_setup(struct sysctllog **clog,
    302     struct bridge_softc *sc);
    303 
    304 struct bridge_control {
    305 	int	(*bc_func)(struct bridge_softc *, void *);
    306 	int	bc_argsize;
    307 	int	bc_flags;
    308 };
    309 
    310 #define	BC_F_COPYIN		0x01	/* copy arguments in */
    311 #define	BC_F_COPYOUT		0x02	/* copy arguments out */
    312 #define	BC_F_SUSER		0x04	/* do super-user check */
    313 #define BC_F_XLATEIN		0x08	/* xlate arguments in */
    314 #define BC_F_XLATEOUT		0x10	/* xlate arguments out */
    315 
    316 static const struct bridge_control bridge_control_table[] = {
    317 [BRDGADD] = {bridge_ioctl_add, sizeof(struct ifbreq), BC_F_COPYIN|BC_F_SUSER},
    318 [BRDGDEL] = {bridge_ioctl_del, sizeof(struct ifbreq), BC_F_COPYIN|BC_F_SUSER},
    319 
    320 [BRDGGIFFLGS] = {bridge_ioctl_gifflags, sizeof(struct ifbreq), BC_F_COPYIN|BC_F_COPYOUT},
    321 [BRDGSIFFLGS] = {bridge_ioctl_sifflags, sizeof(struct ifbreq), BC_F_COPYIN|BC_F_SUSER},
    322 
    323 [BRDGSCACHE] = {bridge_ioctl_scache, sizeof(struct ifbrparam), BC_F_COPYIN|BC_F_SUSER},
    324 [BRDGGCACHE] = {bridge_ioctl_gcache, sizeof(struct ifbrparam), BC_F_COPYOUT},
    325 
    326 [OBRDGGIFS] = {bridge_ioctl_gifs, sizeof(struct ifbifconf), BC_F_COPYIN|BC_F_COPYOUT},
    327 [OBRDGRTS] = {bridge_ioctl_rts, sizeof(struct ifbaconf), BC_F_COPYIN|BC_F_COPYOUT},
    328 
    329 [BRDGSADDR] = {bridge_ioctl_saddr, sizeof(struct ifbareq), BC_F_COPYIN|BC_F_SUSER},
    330 
    331 [BRDGSTO] = {bridge_ioctl_sto, sizeof(struct ifbrparam), BC_F_COPYIN|BC_F_SUSER},
    332 [BRDGGTO] = {bridge_ioctl_gto, sizeof(struct ifbrparam), BC_F_COPYOUT},
    333 
    334 [BRDGDADDR] = {bridge_ioctl_daddr, sizeof(struct ifbareq), BC_F_COPYIN|BC_F_SUSER},
    335 
    336 [BRDGFLUSH] = {bridge_ioctl_flush, sizeof(struct ifbreq), BC_F_COPYIN|BC_F_SUSER},
    337 
    338 [BRDGGPRI] = {bridge_ioctl_gpri, sizeof(struct ifbrparam), BC_F_COPYOUT},
    339 [BRDGSPRI] = {bridge_ioctl_spri, sizeof(struct ifbrparam), BC_F_COPYIN|BC_F_SUSER},
    340 
    341 [BRDGGHT] = {bridge_ioctl_ght, sizeof(struct ifbrparam), BC_F_COPYOUT},
    342 [BRDGSHT] = {bridge_ioctl_sht, sizeof(struct ifbrparam), BC_F_COPYIN|BC_F_SUSER},
    343 
    344 [BRDGGFD] = {bridge_ioctl_gfd, sizeof(struct ifbrparam), BC_F_COPYOUT},
    345 [BRDGSFD] = {bridge_ioctl_sfd, sizeof(struct ifbrparam), BC_F_COPYIN|BC_F_SUSER},
    346 
    347 [BRDGGMA] = {bridge_ioctl_gma, sizeof(struct ifbrparam), BC_F_COPYOUT},
    348 [BRDGSMA] = {bridge_ioctl_sma, sizeof(struct ifbrparam), BC_F_COPYIN|BC_F_SUSER},
    349 
    350 [BRDGSIFPRIO] = {bridge_ioctl_sifprio, sizeof(struct ifbreq), BC_F_COPYIN|BC_F_SUSER},
    351 
    352 [BRDGSIFCOST] = {bridge_ioctl_sifcost, sizeof(struct ifbreq), BC_F_COPYIN|BC_F_SUSER},
    353 #if defined(BRIDGE_IPF)
    354 [BRDGGFILT] = {bridge_ioctl_gfilt, sizeof(struct ifbrparam), BC_F_COPYOUT},
    355 [BRDGSFILT] = {bridge_ioctl_sfilt, sizeof(struct ifbrparam), BC_F_COPYIN|BC_F_SUSER},
    356 #endif /* BRIDGE_IPF */
    357 [BRDGGIFS] = {bridge_ioctl_gifs, sizeof(struct ifbifconf), BC_F_XLATEIN|BC_F_XLATEOUT},
    358 [BRDGRTS] = {bridge_ioctl_rts, sizeof(struct ifbaconf), BC_F_XLATEIN|BC_F_XLATEOUT},
    359 };
    360 
    361 static const int bridge_control_table_size = __arraycount(bridge_control_table);
    362 
    363 static LIST_HEAD(, bridge_softc) bridge_list;
    364 static kmutex_t bridge_list_lock;
    365 
    366 static struct if_clone bridge_cloner =
    367     IF_CLONE_INITIALIZER("bridge", bridge_clone_create, bridge_clone_destroy);
    368 
    369 /*
    370  * bridgeattach:
    371  *
    372  *	Pseudo-device attach routine.
    373  */
    374 void
    375 bridgeattach(int n)
    376 {
    377 
    378 	pool_init(&bridge_rtnode_pool, sizeof(struct bridge_rtnode),
    379 	    0, 0, 0, "brtpl", NULL, IPL_NET);
    380 
    381 	LIST_INIT(&bridge_list);
    382 	mutex_init(&bridge_list_lock, MUTEX_DEFAULT, IPL_NET);
    383 	if_clone_attach(&bridge_cloner);
    384 }
    385 
    386 /*
    387  * bridge_clone_create:
    388  *
    389  *	Create a new bridge instance.
    390  */
    391 static int
    392 bridge_clone_create(struct if_clone *ifc, int unit)
    393 {
    394 	struct bridge_softc *sc;
    395 	struct ifnet *ifp;
    396 	int error, flags;
    397 
    398 	sc = kmem_zalloc(sizeof(*sc),  KM_SLEEP);
    399 	ifp = &sc->sc_if;
    400 
    401 	sc->sc_brtmax = BRIDGE_RTABLE_MAX;
    402 	sc->sc_brttimeout = BRIDGE_RTABLE_TIMEOUT;
    403 	sc->sc_bridge_max_age = BSTP_DEFAULT_MAX_AGE;
    404 	sc->sc_bridge_hello_time = BSTP_DEFAULT_HELLO_TIME;
    405 	sc->sc_bridge_forward_delay = BSTP_DEFAULT_FORWARD_DELAY;
    406 	sc->sc_bridge_priority = BSTP_DEFAULT_BRIDGE_PRIORITY;
    407 	sc->sc_hold_time = BSTP_DEFAULT_HOLD_TIME;
    408 	sc->sc_filter_flags = 0;
    409 
    410 	/* Initialize our routing table. */
    411 	bridge_rtable_init(sc);
    412 
    413 #ifdef BRIDGE_MPSAFE
    414 	flags = WQ_MPSAFE;
    415 #else
    416 	flags = 0;
    417 #endif
    418 	error = workqueue_create(&sc->sc_rtage_wq, "bridge_rtage",
    419 	    bridge_rtage_work, sc, PRI_SOFTNET, IPL_SOFTNET, flags);
    420 	if (error)
    421 		panic("%s: workqueue_create %d\n", __func__, error);
    422 
    423 	callout_init(&sc->sc_brcallout, 0);
    424 	callout_init(&sc->sc_bstpcallout, 0);
    425 
    426 	LIST_INIT(&sc->sc_iflist);
    427 #ifdef BRIDGE_MPSAFE
    428 	sc->sc_iflist_intr_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NET);
    429 	sc->sc_iflist_psz = pserialize_create();
    430 	sc->sc_iflist_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_SOFTNET);
    431 #else
    432 	sc->sc_iflist_intr_lock = NULL;
    433 	sc->sc_iflist_psz = NULL;
    434 	sc->sc_iflist_lock = NULL;
    435 #endif
    436 	cv_init(&sc->sc_iflist_cv, "if_bridge_cv");
    437 
    438 	if_initname(ifp, ifc->ifc_name, unit);
    439 	ifp->if_softc = sc;
    440 	ifp->if_mtu = ETHERMTU;
    441 	ifp->if_ioctl = bridge_ioctl;
    442 	ifp->if_output = bridge_output;
    443 	ifp->if_start = bridge_start;
    444 	ifp->if_stop = bridge_stop;
    445 	ifp->if_init = bridge_init;
    446 	ifp->if_type = IFT_BRIDGE;
    447 	ifp->if_addrlen = 0;
    448 	ifp->if_dlt = DLT_EN10MB;
    449 	ifp->if_hdrlen = ETHER_HDR_LEN;
    450 
    451 	sc->sc_fwd_pktq = pktq_create(IFQ_MAXLEN, bridge_forward, sc);
    452 	KASSERT(sc->sc_fwd_pktq != NULL);
    453 
    454 	bridge_sysctl_fwdq_setup(&ifp->if_sysctl_log, sc);
    455 
    456 	if_attach(ifp);
    457 
    458 	if_alloc_sadl(ifp);
    459 
    460 	mutex_enter(&bridge_list_lock);
    461 	LIST_INSERT_HEAD(&bridge_list, sc, sc_list);
    462 	mutex_exit(&bridge_list_lock);
    463 
    464 	return (0);
    465 }
    466 
    467 /*
    468  * bridge_clone_destroy:
    469  *
    470  *	Destroy a bridge instance.
    471  */
    472 static int
    473 bridge_clone_destroy(struct ifnet *ifp)
    474 {
    475 	struct bridge_softc *sc = ifp->if_softc;
    476 	struct bridge_iflist *bif;
    477 	int s;
    478 
    479 	/* Must be called during IFF_RUNNING, i.e., before bridge_stop */
    480 	pktq_barrier(sc->sc_fwd_pktq);
    481 
    482 	s = splnet();
    483 
    484 	bridge_stop(ifp, 1);
    485 
    486 	BRIDGE_LOCK(sc);
    487 	while ((bif = LIST_FIRST(&sc->sc_iflist)) != NULL)
    488 		bridge_delete_member(sc, bif);
    489 	BRIDGE_UNLOCK(sc);
    490 
    491 	mutex_enter(&bridge_list_lock);
    492 	LIST_REMOVE(sc, sc_list);
    493 	mutex_exit(&bridge_list_lock);
    494 
    495 	splx(s);
    496 
    497 	if_detach(ifp);
    498 
    499 	/* Should be called after if_detach for safe */
    500 	pktq_flush(sc->sc_fwd_pktq);
    501 	pktq_destroy(sc->sc_fwd_pktq);
    502 
    503 	/* Tear down the routing table. */
    504 	bridge_rtable_fini(sc);
    505 
    506 	cv_destroy(&sc->sc_iflist_cv);
    507 	if (sc->sc_iflist_intr_lock)
    508 		mutex_obj_free(sc->sc_iflist_intr_lock);
    509 
    510 	if (sc->sc_iflist_psz)
    511 		pserialize_destroy(sc->sc_iflist_psz);
    512 	if (sc->sc_iflist_lock)
    513 		mutex_obj_free(sc->sc_iflist_lock);
    514 
    515 	workqueue_destroy(sc->sc_rtage_wq);
    516 
    517 	kmem_free(sc, sizeof(*sc));
    518 
    519 	return (0);
    520 }
    521 
    522 static int
    523 bridge_sysctl_fwdq_maxlen(SYSCTLFN_ARGS)
    524 {
    525 	struct sysctlnode node = *rnode;
    526 	const struct bridge_softc *sc =	node.sysctl_data;
    527 	return sysctl_pktq_maxlen(SYSCTLFN_CALL(rnode), sc->sc_fwd_pktq);
    528 }
    529 
    530 #define	SYSCTL_BRIDGE_PKTQ(cn, c)					\
    531 	static int							\
    532 	bridge_sysctl_fwdq_##cn(SYSCTLFN_ARGS)				\
    533 	{								\
    534 		struct sysctlnode node = *rnode;			\
    535 		const struct bridge_softc *sc =	node.sysctl_data;	\
    536 		return sysctl_pktq_count(SYSCTLFN_CALL(rnode),		\
    537 					 sc->sc_fwd_pktq, c);		\
    538 	}
    539 
    540 SYSCTL_BRIDGE_PKTQ(items, PKTQ_NITEMS)
    541 SYSCTL_BRIDGE_PKTQ(drops, PKTQ_DROPS)
    542 
    543 static void
    544 bridge_sysctl_fwdq_setup(struct sysctllog **clog, struct bridge_softc *sc)
    545 {
    546 	const struct sysctlnode *cnode, *rnode;
    547 	sysctlfn len_func = NULL, maxlen_func = NULL, drops_func = NULL;
    548 	const char *ifname = sc->sc_if.if_xname;
    549 
    550 	len_func = bridge_sysctl_fwdq_items;
    551 	maxlen_func = bridge_sysctl_fwdq_maxlen;
    552 	drops_func = bridge_sysctl_fwdq_drops;
    553 
    554 	if (sysctl_createv(clog, 0, NULL, &rnode,
    555 			   CTLFLAG_PERMANENT,
    556 			   CTLTYPE_NODE, "interfaces",
    557 			   SYSCTL_DESCR("Per-interface controls"),
    558 			   NULL, 0, NULL, 0,
    559 			   CTL_NET, CTL_CREATE, CTL_EOL) != 0)
    560 		goto bad;
    561 
    562 	if (sysctl_createv(clog, 0, &rnode, &rnode,
    563 			   CTLFLAG_PERMANENT,
    564 			   CTLTYPE_NODE, ifname,
    565 			   SYSCTL_DESCR("Interface controls"),
    566 			   NULL, 0, NULL, 0,
    567 			   CTL_CREATE, CTL_EOL) != 0)
    568 		goto bad;
    569 
    570 	if (sysctl_createv(clog, 0, &rnode, &rnode,
    571 			   CTLFLAG_PERMANENT,
    572 			   CTLTYPE_NODE, "fwdq",
    573 			   SYSCTL_DESCR("Protocol input queue controls"),
    574 			   NULL, 0, NULL, 0,
    575 			   CTL_CREATE, CTL_EOL) != 0)
    576 		goto bad;
    577 
    578 	if (sysctl_createv(clog, 0, &rnode, &cnode,
    579 			   CTLFLAG_PERMANENT,
    580 			   CTLTYPE_INT, "len",
    581 			   SYSCTL_DESCR("Current forwarding queue length"),
    582 			   len_func, 0, (void *)sc, 0,
    583 			   CTL_CREATE, IFQCTL_LEN, CTL_EOL) != 0)
    584 		goto bad;
    585 
    586 	if (sysctl_createv(clog, 0, &rnode, &cnode,
    587 			   CTLFLAG_PERMANENT|CTLFLAG_READWRITE,
    588 			   CTLTYPE_INT, "maxlen",
    589 			   SYSCTL_DESCR("Maximum allowed forwarding queue length"),
    590 			   maxlen_func, 0, (void *)sc, 0,
    591 			   CTL_CREATE, IFQCTL_MAXLEN, CTL_EOL) != 0)
    592 		goto bad;
    593 
    594 	if (sysctl_createv(clog, 0, &rnode, &cnode,
    595 			   CTLFLAG_PERMANENT,
    596 			   CTLTYPE_INT, "drops",
    597 			   SYSCTL_DESCR("Packets dropped due to full forwarding queue"),
    598 			   drops_func, 0, (void *)sc, 0,
    599 			   CTL_CREATE, IFQCTL_DROPS, CTL_EOL) != 0)
    600 		goto bad;
    601 
    602 	return;
    603 bad:
    604 	aprint_error("%s: could not attach sysctl nodes\n", ifname);
    605 	return;
    606 }
    607 
    608 /*
    609  * bridge_ioctl:
    610  *
    611  *	Handle a control request from the operator.
    612  */
    613 static int
    614 bridge_ioctl(struct ifnet *ifp, u_long cmd, void *data)
    615 {
    616 	struct bridge_softc *sc = ifp->if_softc;
    617 	struct lwp *l = curlwp;	/* XXX */
    618 	union {
    619 		struct ifbreq ifbreq;
    620 		struct ifbifconf ifbifconf;
    621 		struct ifbareq ifbareq;
    622 		struct ifbaconf ifbaconf;
    623 		struct ifbrparam ifbrparam;
    624 	} args;
    625 	struct ifdrv *ifd = (struct ifdrv *) data;
    626 	const struct bridge_control *bc = NULL; /* XXXGCC */
    627 	int s, error = 0;
    628 
    629 	/* Authorize command before calling splnet(). */
    630 	switch (cmd) {
    631 	case SIOCGDRVSPEC:
    632 	case SIOCSDRVSPEC:
    633 		if (ifd->ifd_cmd >= bridge_control_table_size
    634 		    || (bc = &bridge_control_table[ifd->ifd_cmd]) == NULL) {
    635 			error = EINVAL;
    636 			return error;
    637 		}
    638 
    639 		/* We only care about BC_F_SUSER at this point. */
    640 		if ((bc->bc_flags & BC_F_SUSER) == 0)
    641 			break;
    642 
    643 		error = kauth_authorize_network(l->l_cred,
    644 		    KAUTH_NETWORK_INTERFACE_BRIDGE,
    645 		    cmd == SIOCGDRVSPEC ?
    646 		     KAUTH_REQ_NETWORK_INTERFACE_BRIDGE_GETPRIV :
    647 		     KAUTH_REQ_NETWORK_INTERFACE_BRIDGE_SETPRIV,
    648 		     ifd, NULL, NULL);
    649 		if (error)
    650 			return (error);
    651 
    652 		break;
    653 	}
    654 
    655 	s = splnet();
    656 
    657 	switch (cmd) {
    658 	case SIOCGDRVSPEC:
    659 	case SIOCSDRVSPEC:
    660 		KASSERT(bc != NULL);
    661 		if (cmd == SIOCGDRVSPEC &&
    662 		    (bc->bc_flags & (BC_F_COPYOUT|BC_F_XLATEOUT)) == 0) {
    663 			error = EINVAL;
    664 			break;
    665 		}
    666 		else if (cmd == SIOCSDRVSPEC &&
    667 		    (bc->bc_flags & (BC_F_COPYOUT|BC_F_XLATEOUT)) != 0) {
    668 			error = EINVAL;
    669 			break;
    670 		}
    671 
    672 		/* BC_F_SUSER is checked above, before splnet(). */
    673 
    674 		if ((bc->bc_flags & (BC_F_XLATEIN|BC_F_XLATEOUT)) == 0
    675 		    && (ifd->ifd_len != bc->bc_argsize
    676 			|| ifd->ifd_len > sizeof(args))) {
    677 			error = EINVAL;
    678 			break;
    679 		}
    680 
    681 		memset(&args, 0, sizeof(args));
    682 		if (bc->bc_flags & BC_F_COPYIN) {
    683 			error = copyin(ifd->ifd_data, &args, ifd->ifd_len);
    684 			if (error)
    685 				break;
    686 		} else if (bc->bc_flags & BC_F_XLATEIN) {
    687 			args.ifbifconf.ifbic_len = ifd->ifd_len;
    688 			args.ifbifconf.ifbic_buf = ifd->ifd_data;
    689 		}
    690 
    691 		error = (*bc->bc_func)(sc, &args);
    692 		if (error)
    693 			break;
    694 
    695 		if (bc->bc_flags & BC_F_COPYOUT) {
    696 			error = copyout(&args, ifd->ifd_data, ifd->ifd_len);
    697 		} else if (bc->bc_flags & BC_F_XLATEOUT) {
    698 			ifd->ifd_len = args.ifbifconf.ifbic_len;
    699 			ifd->ifd_data = args.ifbifconf.ifbic_buf;
    700 		}
    701 		break;
    702 
    703 	case SIOCSIFFLAGS:
    704 		if ((error = ifioctl_common(ifp, cmd, data)) != 0)
    705 			break;
    706 		switch (ifp->if_flags & (IFF_UP|IFF_RUNNING)) {
    707 		case IFF_RUNNING:
    708 			/*
    709 			 * If interface is marked down and it is running,
    710 			 * then stop and disable it.
    711 			 */
    712 			(*ifp->if_stop)(ifp, 1);
    713 			break;
    714 		case IFF_UP:
    715 			/*
    716 			 * If interface is marked up and it is stopped, then
    717 			 * start it.
    718 			 */
    719 			error = (*ifp->if_init)(ifp);
    720 			break;
    721 		default:
    722 			break;
    723 		}
    724 		break;
    725 
    726 	case SIOCSIFMTU:
    727 		if ((error = ifioctl_common(ifp, cmd, data)) == ENETRESET)
    728 			error = 0;
    729 		break;
    730 
    731 	default:
    732 		error = ifioctl_common(ifp, cmd, data);
    733 		break;
    734 	}
    735 
    736 	splx(s);
    737 
    738 	return (error);
    739 }
    740 
    741 /*
    742  * bridge_lookup_member:
    743  *
    744  *	Lookup a bridge member interface.
    745  */
    746 static struct bridge_iflist *
    747 bridge_lookup_member(struct bridge_softc *sc, const char *name)
    748 {
    749 	struct bridge_iflist *bif;
    750 	struct ifnet *ifp;
    751 	int s;
    752 
    753 	BRIDGE_PSZ_RENTER(s);
    754 
    755 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
    756 		ifp = bif->bif_ifp;
    757 		if (strcmp(ifp->if_xname, name) == 0)
    758 			break;
    759 	}
    760 	bif = bridge_try_hold_bif(bif);
    761 
    762 	BRIDGE_PSZ_REXIT(s);
    763 
    764 	return bif;
    765 }
    766 
    767 /*
    768  * bridge_lookup_member_if:
    769  *
    770  *	Lookup a bridge member interface by ifnet*.
    771  */
    772 static struct bridge_iflist *
    773 bridge_lookup_member_if(struct bridge_softc *sc, struct ifnet *member_ifp)
    774 {
    775 	struct bridge_iflist *bif;
    776 	int s;
    777 
    778 	BRIDGE_PSZ_RENTER(s);
    779 
    780 	bif = member_ifp->if_bridgeif;
    781 	bif = bridge_try_hold_bif(bif);
    782 
    783 	BRIDGE_PSZ_REXIT(s);
    784 
    785 	return bif;
    786 }
    787 
    788 static struct bridge_iflist *
    789 bridge_try_hold_bif(struct bridge_iflist *bif)
    790 {
    791 #ifdef BRIDGE_MPSAFE
    792 	if (bif != NULL) {
    793 		if (bif->bif_waiting)
    794 			bif = NULL;
    795 		else
    796 			atomic_inc_32(&bif->bif_refs);
    797 	}
    798 #endif
    799 	return bif;
    800 }
    801 
    802 /*
    803  * bridge_release_member:
    804  *
    805  *	Release the specified member interface.
    806  */
    807 static void
    808 bridge_release_member(struct bridge_softc *sc, struct bridge_iflist *bif)
    809 {
    810 #ifdef BRIDGE_MPSAFE
    811 	uint32_t refs;
    812 
    813 	refs = atomic_dec_uint_nv(&bif->bif_refs);
    814 	if (__predict_false(refs == 0 && bif->bif_waiting)) {
    815 		BRIDGE_INTR_LOCK(sc);
    816 		cv_broadcast(&sc->sc_iflist_cv);
    817 		BRIDGE_INTR_UNLOCK(sc);
    818 	}
    819 #else
    820 	(void)sc;
    821 	(void)bif;
    822 #endif
    823 }
    824 
    825 /*
    826  * bridge_delete_member:
    827  *
    828  *	Delete the specified member interface.
    829  */
    830 static void
    831 bridge_delete_member(struct bridge_softc *sc, struct bridge_iflist *bif)
    832 {
    833 	struct ifnet *ifs = bif->bif_ifp;
    834 
    835 	KASSERT(BRIDGE_LOCKED(sc));
    836 
    837 	ifs->if_input = ether_input;
    838 	ifs->if_bridge = NULL;
    839 	ifs->if_bridgeif = NULL;
    840 
    841 	LIST_REMOVE(bif, bif_next);
    842 
    843 	BRIDGE_PSZ_PERFORM(sc);
    844 
    845 	BRIDGE_UNLOCK(sc);
    846 
    847 #ifdef BRIDGE_MPSAFE
    848 	BRIDGE_INTR_LOCK(sc);
    849 	bif->bif_waiting = true;
    850 	membar_sync();
    851 	while (bif->bif_refs > 0) {
    852 		aprint_debug("%s: cv_wait on iflist\n", __func__);
    853 		cv_wait(&sc->sc_iflist_cv, sc->sc_iflist_intr_lock);
    854 	}
    855 	bif->bif_waiting = false;
    856 	BRIDGE_INTR_UNLOCK(sc);
    857 #endif
    858 
    859 	kmem_free(bif, sizeof(*bif));
    860 
    861 	BRIDGE_LOCK(sc);
    862 }
    863 
    864 static int
    865 bridge_ioctl_add(struct bridge_softc *sc, void *arg)
    866 {
    867 	struct ifbreq *req = arg;
    868 	struct bridge_iflist *bif = NULL;
    869 	struct ifnet *ifs;
    870 	int error = 0;
    871 
    872 	ifs = ifunit(req->ifbr_ifsname);
    873 	if (ifs == NULL)
    874 		return (ENOENT);
    875 
    876 	if (sc->sc_if.if_mtu != ifs->if_mtu)
    877 		return (EINVAL);
    878 
    879 	if (ifs->if_bridge == sc)
    880 		return (EEXIST);
    881 
    882 	if (ifs->if_bridge != NULL)
    883 		return (EBUSY);
    884 
    885 	if (ifs->if_input != ether_input)
    886 		return EINVAL;
    887 
    888 	/* FIXME: doesn't work with non-IFF_SIMPLEX interfaces */
    889 	if ((ifs->if_flags & IFF_SIMPLEX) == 0)
    890 		return EINVAL;
    891 
    892 	bif = kmem_alloc(sizeof(*bif), KM_SLEEP);
    893 
    894 	switch (ifs->if_type) {
    895 	case IFT_ETHER:
    896 		/*
    897 		 * Place the interface into promiscuous mode.
    898 		 */
    899 		error = ifpromisc(ifs, 1);
    900 		if (error)
    901 			goto out;
    902 		break;
    903 	default:
    904 		error = EINVAL;
    905 		goto out;
    906 	}
    907 
    908 	bif->bif_ifp = ifs;
    909 	bif->bif_flags = IFBIF_LEARNING | IFBIF_DISCOVER;
    910 	bif->bif_priority = BSTP_DEFAULT_PORT_PRIORITY;
    911 	bif->bif_path_cost = BSTP_DEFAULT_PATH_COST;
    912 	bif->bif_refs = 0;
    913 	bif->bif_waiting = false;
    914 
    915 	BRIDGE_LOCK(sc);
    916 
    917 	ifs->if_bridge = sc;
    918 	ifs->if_bridgeif = bif;
    919 	LIST_INSERT_HEAD(&sc->sc_iflist, bif, bif_next);
    920 	ifs->if_input = bridge_input;
    921 
    922 	BRIDGE_UNLOCK(sc);
    923 
    924 	if (sc->sc_if.if_flags & IFF_RUNNING)
    925 		bstp_initialization(sc);
    926 	else
    927 		bstp_stop(sc);
    928 
    929  out:
    930 	if (error) {
    931 		if (bif != NULL)
    932 			kmem_free(bif, sizeof(*bif));
    933 	}
    934 	return (error);
    935 }
    936 
    937 static int
    938 bridge_ioctl_del(struct bridge_softc *sc, void *arg)
    939 {
    940 	struct ifbreq *req = arg;
    941 	const char *name = req->ifbr_ifsname;
    942 	struct bridge_iflist *bif;
    943 	struct ifnet *ifs;
    944 
    945 	BRIDGE_LOCK(sc);
    946 
    947 	/*
    948 	 * Don't use bridge_lookup_member. We want to get a member
    949 	 * with bif_refs == 0.
    950 	 */
    951 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
    952 		ifs = bif->bif_ifp;
    953 		if (strcmp(ifs->if_xname, name) == 0)
    954 			break;
    955 	}
    956 
    957 	if (bif == NULL) {
    958 		BRIDGE_UNLOCK(sc);
    959 		return ENOENT;
    960 	}
    961 
    962 	bridge_delete_member(sc, bif);
    963 
    964 	BRIDGE_UNLOCK(sc);
    965 
    966 	switch (ifs->if_type) {
    967 	case IFT_ETHER:
    968 		/*
    969 		 * Take the interface out of promiscuous mode.
    970 		 * Don't call it with holding a spin lock.
    971 		 */
    972 		(void) ifpromisc(ifs, 0);
    973 		break;
    974 	default:
    975 #ifdef DIAGNOSTIC
    976 		panic("bridge_delete_member: impossible");
    977 #endif
    978 		break;
    979 	}
    980 
    981 	bridge_rtdelete(sc, ifs);
    982 
    983 	if (sc->sc_if.if_flags & IFF_RUNNING)
    984 		bstp_initialization(sc);
    985 
    986 	return 0;
    987 }
    988 
    989 static int
    990 bridge_ioctl_gifflags(struct bridge_softc *sc, void *arg)
    991 {
    992 	struct ifbreq *req = arg;
    993 	struct bridge_iflist *bif;
    994 
    995 	bif = bridge_lookup_member(sc, req->ifbr_ifsname);
    996 	if (bif == NULL)
    997 		return (ENOENT);
    998 
    999 	req->ifbr_ifsflags = bif->bif_flags;
   1000 	req->ifbr_state = bif->bif_state;
   1001 	req->ifbr_priority = bif->bif_priority;
   1002 	req->ifbr_path_cost = bif->bif_path_cost;
   1003 	req->ifbr_portno = bif->bif_ifp->if_index & 0xff;
   1004 
   1005 	bridge_release_member(sc, bif);
   1006 
   1007 	return (0);
   1008 }
   1009 
   1010 static int
   1011 bridge_ioctl_sifflags(struct bridge_softc *sc, void *arg)
   1012 {
   1013 	struct ifbreq *req = arg;
   1014 	struct bridge_iflist *bif;
   1015 
   1016 	bif = bridge_lookup_member(sc, req->ifbr_ifsname);
   1017 	if (bif == NULL)
   1018 		return (ENOENT);
   1019 
   1020 	if (req->ifbr_ifsflags & IFBIF_STP) {
   1021 		switch (bif->bif_ifp->if_type) {
   1022 		case IFT_ETHER:
   1023 			/* These can do spanning tree. */
   1024 			break;
   1025 
   1026 		default:
   1027 			/* Nothing else can. */
   1028 			bridge_release_member(sc, bif);
   1029 			return (EINVAL);
   1030 		}
   1031 	}
   1032 
   1033 	bif->bif_flags = req->ifbr_ifsflags;
   1034 
   1035 	bridge_release_member(sc, bif);
   1036 
   1037 	if (sc->sc_if.if_flags & IFF_RUNNING)
   1038 		bstp_initialization(sc);
   1039 
   1040 	return (0);
   1041 }
   1042 
   1043 static int
   1044 bridge_ioctl_scache(struct bridge_softc *sc, void *arg)
   1045 {
   1046 	struct ifbrparam *param = arg;
   1047 
   1048 	sc->sc_brtmax = param->ifbrp_csize;
   1049 	bridge_rttrim(sc);
   1050 
   1051 	return (0);
   1052 }
   1053 
   1054 static int
   1055 bridge_ioctl_gcache(struct bridge_softc *sc, void *arg)
   1056 {
   1057 	struct ifbrparam *param = arg;
   1058 
   1059 	param->ifbrp_csize = sc->sc_brtmax;
   1060 
   1061 	return (0);
   1062 }
   1063 
   1064 static int
   1065 bridge_ioctl_gifs(struct bridge_softc *sc, void *arg)
   1066 {
   1067 	struct ifbifconf *bifc = arg;
   1068 	struct bridge_iflist *bif;
   1069 	struct ifbreq *breqs;
   1070 	int i, count, error = 0;
   1071 
   1072 retry:
   1073 	BRIDGE_LOCK(sc);
   1074 	count = 0;
   1075 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next)
   1076 		count++;
   1077 	BRIDGE_UNLOCK(sc);
   1078 
   1079 	if (count == 0) {
   1080 		bifc->ifbic_len = 0;
   1081 		return 0;
   1082 	}
   1083 
   1084 	if (bifc->ifbic_len == 0 || bifc->ifbic_len < (sizeof(*breqs) * count)) {
   1085 		/* Tell that a larger buffer is needed */
   1086 		bifc->ifbic_len = sizeof(*breqs) * count;
   1087 		return 0;
   1088 	}
   1089 
   1090 	breqs = kmem_alloc(sizeof(*breqs) * count, KM_SLEEP);
   1091 
   1092 	BRIDGE_LOCK(sc);
   1093 
   1094 	i = 0;
   1095 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next)
   1096 		i++;
   1097 	if (i > count) {
   1098 		/*
   1099 		 * The number of members has been increased.
   1100 		 * We need more memory!
   1101 		 */
   1102 		BRIDGE_UNLOCK(sc);
   1103 		kmem_free(breqs, sizeof(*breqs) * count);
   1104 		goto retry;
   1105 	}
   1106 
   1107 	i = 0;
   1108 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
   1109 		struct ifbreq *breq = &breqs[i++];
   1110 		memset(breq, 0, sizeof(*breq));
   1111 
   1112 		strlcpy(breq->ifbr_ifsname, bif->bif_ifp->if_xname,
   1113 		    sizeof(breq->ifbr_ifsname));
   1114 		breq->ifbr_ifsflags = bif->bif_flags;
   1115 		breq->ifbr_state = bif->bif_state;
   1116 		breq->ifbr_priority = bif->bif_priority;
   1117 		breq->ifbr_path_cost = bif->bif_path_cost;
   1118 		breq->ifbr_portno = bif->bif_ifp->if_index & 0xff;
   1119 	}
   1120 
   1121 	/* Don't call copyout with holding the mutex */
   1122 	BRIDGE_UNLOCK(sc);
   1123 
   1124 	for (i = 0; i < count; i++) {
   1125 		error = copyout(&breqs[i], bifc->ifbic_req + i, sizeof(*breqs));
   1126 		if (error)
   1127 			break;
   1128 	}
   1129 	bifc->ifbic_len = sizeof(*breqs) * i;
   1130 
   1131 	kmem_free(breqs, sizeof(*breqs) * count);
   1132 
   1133 	return error;
   1134 }
   1135 
   1136 static int
   1137 bridge_ioctl_rts(struct bridge_softc *sc, void *arg)
   1138 {
   1139 	struct ifbaconf *bac = arg;
   1140 	struct bridge_rtnode *brt;
   1141 	struct ifbareq bareq;
   1142 	int count = 0, error = 0, len;
   1143 
   1144 	if (bac->ifbac_len == 0)
   1145 		return (0);
   1146 
   1147 	BRIDGE_RT_INTR_LOCK(sc);
   1148 
   1149 	len = bac->ifbac_len;
   1150 	LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) {
   1151 		if (len < sizeof(bareq))
   1152 			goto out;
   1153 		memset(&bareq, 0, sizeof(bareq));
   1154 		strlcpy(bareq.ifba_ifsname, brt->brt_ifp->if_xname,
   1155 		    sizeof(bareq.ifba_ifsname));
   1156 		memcpy(bareq.ifba_dst, brt->brt_addr, sizeof(brt->brt_addr));
   1157 		if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) {
   1158 			bareq.ifba_expire = brt->brt_expire - time_uptime;
   1159 		} else
   1160 			bareq.ifba_expire = 0;
   1161 		bareq.ifba_flags = brt->brt_flags;
   1162 
   1163 		error = copyout(&bareq, bac->ifbac_req + count, sizeof(bareq));
   1164 		if (error)
   1165 			goto out;
   1166 		count++;
   1167 		len -= sizeof(bareq);
   1168 	}
   1169  out:
   1170 	BRIDGE_RT_INTR_UNLOCK(sc);
   1171 
   1172 	bac->ifbac_len = sizeof(bareq) * count;
   1173 	return (error);
   1174 }
   1175 
   1176 static int
   1177 bridge_ioctl_saddr(struct bridge_softc *sc, void *arg)
   1178 {
   1179 	struct ifbareq *req = arg;
   1180 	struct bridge_iflist *bif;
   1181 	int error;
   1182 
   1183 	bif = bridge_lookup_member(sc, req->ifba_ifsname);
   1184 	if (bif == NULL)
   1185 		return (ENOENT);
   1186 
   1187 	error = bridge_rtupdate(sc, req->ifba_dst, bif->bif_ifp, 1,
   1188 	    req->ifba_flags);
   1189 
   1190 	bridge_release_member(sc, bif);
   1191 
   1192 	return (error);
   1193 }
   1194 
   1195 static int
   1196 bridge_ioctl_sto(struct bridge_softc *sc, void *arg)
   1197 {
   1198 	struct ifbrparam *param = arg;
   1199 
   1200 	sc->sc_brttimeout = param->ifbrp_ctime;
   1201 
   1202 	return (0);
   1203 }
   1204 
   1205 static int
   1206 bridge_ioctl_gto(struct bridge_softc *sc, void *arg)
   1207 {
   1208 	struct ifbrparam *param = arg;
   1209 
   1210 	param->ifbrp_ctime = sc->sc_brttimeout;
   1211 
   1212 	return (0);
   1213 }
   1214 
   1215 static int
   1216 bridge_ioctl_daddr(struct bridge_softc *sc, void *arg)
   1217 {
   1218 	struct ifbareq *req = arg;
   1219 
   1220 	return (bridge_rtdaddr(sc, req->ifba_dst));
   1221 }
   1222 
   1223 static int
   1224 bridge_ioctl_flush(struct bridge_softc *sc, void *arg)
   1225 {
   1226 	struct ifbreq *req = arg;
   1227 
   1228 	bridge_rtflush(sc, req->ifbr_ifsflags);
   1229 
   1230 	return (0);
   1231 }
   1232 
   1233 static int
   1234 bridge_ioctl_gpri(struct bridge_softc *sc, void *arg)
   1235 {
   1236 	struct ifbrparam *param = arg;
   1237 
   1238 	param->ifbrp_prio = sc->sc_bridge_priority;
   1239 
   1240 	return (0);
   1241 }
   1242 
   1243 static int
   1244 bridge_ioctl_spri(struct bridge_softc *sc, void *arg)
   1245 {
   1246 	struct ifbrparam *param = arg;
   1247 
   1248 	sc->sc_bridge_priority = param->ifbrp_prio;
   1249 
   1250 	if (sc->sc_if.if_flags & IFF_RUNNING)
   1251 		bstp_initialization(sc);
   1252 
   1253 	return (0);
   1254 }
   1255 
   1256 static int
   1257 bridge_ioctl_ght(struct bridge_softc *sc, void *arg)
   1258 {
   1259 	struct ifbrparam *param = arg;
   1260 
   1261 	param->ifbrp_hellotime = sc->sc_bridge_hello_time >> 8;
   1262 
   1263 	return (0);
   1264 }
   1265 
   1266 static int
   1267 bridge_ioctl_sht(struct bridge_softc *sc, void *arg)
   1268 {
   1269 	struct ifbrparam *param = arg;
   1270 
   1271 	if (param->ifbrp_hellotime == 0)
   1272 		return (EINVAL);
   1273 	sc->sc_bridge_hello_time = param->ifbrp_hellotime << 8;
   1274 
   1275 	if (sc->sc_if.if_flags & IFF_RUNNING)
   1276 		bstp_initialization(sc);
   1277 
   1278 	return (0);
   1279 }
   1280 
   1281 static int
   1282 bridge_ioctl_gfd(struct bridge_softc *sc, void *arg)
   1283 {
   1284 	struct ifbrparam *param = arg;
   1285 
   1286 	param->ifbrp_fwddelay = sc->sc_bridge_forward_delay >> 8;
   1287 
   1288 	return (0);
   1289 }
   1290 
   1291 static int
   1292 bridge_ioctl_sfd(struct bridge_softc *sc, void *arg)
   1293 {
   1294 	struct ifbrparam *param = arg;
   1295 
   1296 	if (param->ifbrp_fwddelay == 0)
   1297 		return (EINVAL);
   1298 	sc->sc_bridge_forward_delay = param->ifbrp_fwddelay << 8;
   1299 
   1300 	if (sc->sc_if.if_flags & IFF_RUNNING)
   1301 		bstp_initialization(sc);
   1302 
   1303 	return (0);
   1304 }
   1305 
   1306 static int
   1307 bridge_ioctl_gma(struct bridge_softc *sc, void *arg)
   1308 {
   1309 	struct ifbrparam *param = arg;
   1310 
   1311 	param->ifbrp_maxage = sc->sc_bridge_max_age >> 8;
   1312 
   1313 	return (0);
   1314 }
   1315 
   1316 static int
   1317 bridge_ioctl_sma(struct bridge_softc *sc, void *arg)
   1318 {
   1319 	struct ifbrparam *param = arg;
   1320 
   1321 	if (param->ifbrp_maxage == 0)
   1322 		return (EINVAL);
   1323 	sc->sc_bridge_max_age = param->ifbrp_maxage << 8;
   1324 
   1325 	if (sc->sc_if.if_flags & IFF_RUNNING)
   1326 		bstp_initialization(sc);
   1327 
   1328 	return (0);
   1329 }
   1330 
   1331 static int
   1332 bridge_ioctl_sifprio(struct bridge_softc *sc, void *arg)
   1333 {
   1334 	struct ifbreq *req = arg;
   1335 	struct bridge_iflist *bif;
   1336 
   1337 	bif = bridge_lookup_member(sc, req->ifbr_ifsname);
   1338 	if (bif == NULL)
   1339 		return (ENOENT);
   1340 
   1341 	bif->bif_priority = req->ifbr_priority;
   1342 
   1343 	if (sc->sc_if.if_flags & IFF_RUNNING)
   1344 		bstp_initialization(sc);
   1345 
   1346 	bridge_release_member(sc, bif);
   1347 
   1348 	return (0);
   1349 }
   1350 
   1351 #if defined(BRIDGE_IPF)
   1352 static int
   1353 bridge_ioctl_gfilt(struct bridge_softc *sc, void *arg)
   1354 {
   1355 	struct ifbrparam *param = arg;
   1356 
   1357 	param->ifbrp_filter = sc->sc_filter_flags;
   1358 
   1359 	return (0);
   1360 }
   1361 
   1362 static int
   1363 bridge_ioctl_sfilt(struct bridge_softc *sc, void *arg)
   1364 {
   1365 	struct ifbrparam *param = arg;
   1366 	uint32_t nflags, oflags;
   1367 
   1368 	if (param->ifbrp_filter & ~IFBF_FILT_MASK)
   1369 		return (EINVAL);
   1370 
   1371 	nflags = param->ifbrp_filter;
   1372 	oflags = sc->sc_filter_flags;
   1373 
   1374 	if ((nflags & IFBF_FILT_USEIPF) && !(oflags & IFBF_FILT_USEIPF)) {
   1375 		pfil_add_hook((void *)bridge_ipf, NULL, PFIL_IN|PFIL_OUT,
   1376 			sc->sc_if.if_pfil);
   1377 	}
   1378 	if (!(nflags & IFBF_FILT_USEIPF) && (oflags & IFBF_FILT_USEIPF)) {
   1379 		pfil_remove_hook((void *)bridge_ipf, NULL, PFIL_IN|PFIL_OUT,
   1380 			sc->sc_if.if_pfil);
   1381 	}
   1382 
   1383 	sc->sc_filter_flags = nflags;
   1384 
   1385 	return (0);
   1386 }
   1387 #endif /* BRIDGE_IPF */
   1388 
   1389 static int
   1390 bridge_ioctl_sifcost(struct bridge_softc *sc, void *arg)
   1391 {
   1392 	struct ifbreq *req = arg;
   1393 	struct bridge_iflist *bif;
   1394 
   1395 	bif = bridge_lookup_member(sc, req->ifbr_ifsname);
   1396 	if (bif == NULL)
   1397 		return (ENOENT);
   1398 
   1399 	bif->bif_path_cost = req->ifbr_path_cost;
   1400 
   1401 	if (sc->sc_if.if_flags & IFF_RUNNING)
   1402 		bstp_initialization(sc);
   1403 
   1404 	bridge_release_member(sc, bif);
   1405 
   1406 	return (0);
   1407 }
   1408 
   1409 /*
   1410  * bridge_ifdetach:
   1411  *
   1412  *	Detach an interface from a bridge.  Called when a member
   1413  *	interface is detaching.
   1414  */
   1415 void
   1416 bridge_ifdetach(struct ifnet *ifp)
   1417 {
   1418 	struct bridge_softc *sc = ifp->if_bridge;
   1419 	struct ifbreq breq;
   1420 
   1421 	/* ioctl_lock should prevent this from happening */
   1422 	KASSERT(sc != NULL);
   1423 
   1424 	memset(&breq, 0, sizeof(breq));
   1425 	strlcpy(breq.ifbr_ifsname, ifp->if_xname, sizeof(breq.ifbr_ifsname));
   1426 
   1427 	(void) bridge_ioctl_del(sc, &breq);
   1428 }
   1429 
   1430 /*
   1431  * bridge_init:
   1432  *
   1433  *	Initialize a bridge interface.
   1434  */
   1435 static int
   1436 bridge_init(struct ifnet *ifp)
   1437 {
   1438 	struct bridge_softc *sc = ifp->if_softc;
   1439 
   1440 	if (ifp->if_flags & IFF_RUNNING)
   1441 		return (0);
   1442 
   1443 	callout_reset(&sc->sc_brcallout, bridge_rtable_prune_period * hz,
   1444 	    bridge_timer, sc);
   1445 
   1446 	ifp->if_flags |= IFF_RUNNING;
   1447 	bstp_initialization(sc);
   1448 	return (0);
   1449 }
   1450 
   1451 /*
   1452  * bridge_stop:
   1453  *
   1454  *	Stop the bridge interface.
   1455  */
   1456 static void
   1457 bridge_stop(struct ifnet *ifp, int disable)
   1458 {
   1459 	struct bridge_softc *sc = ifp->if_softc;
   1460 
   1461 	if ((ifp->if_flags & IFF_RUNNING) == 0)
   1462 		return;
   1463 
   1464 	callout_stop(&sc->sc_brcallout);
   1465 	bstp_stop(sc);
   1466 
   1467 	bridge_rtflush(sc, IFBF_FLUSHDYN);
   1468 
   1469 	ifp->if_flags &= ~IFF_RUNNING;
   1470 }
   1471 
   1472 /*
   1473  * bridge_enqueue:
   1474  *
   1475  *	Enqueue a packet on a bridge member interface.
   1476  */
   1477 void
   1478 bridge_enqueue(struct bridge_softc *sc, struct ifnet *dst_ifp, struct mbuf *m,
   1479     int runfilt)
   1480 {
   1481 	ALTQ_DECL(struct altq_pktattr pktattr;)
   1482 	int len, error;
   1483 	short mflags;
   1484 
   1485 	/*
   1486 	 * Clear any in-bound checksum flags for this packet.
   1487 	 */
   1488 	m->m_pkthdr.csum_flags = 0;
   1489 
   1490 	if (runfilt) {
   1491 		if (pfil_run_hooks(sc->sc_if.if_pfil, &m,
   1492 		    dst_ifp, PFIL_OUT) != 0) {
   1493 			if (m != NULL)
   1494 				m_freem(m);
   1495 			return;
   1496 		}
   1497 		if (m == NULL)
   1498 			return;
   1499 	}
   1500 
   1501 #ifdef ALTQ
   1502 	/*
   1503 	 * If ALTQ is enabled on the member interface, do
   1504 	 * classification; the queueing discipline might
   1505 	 * not require classification, but might require
   1506 	 * the address family/header pointer in the pktattr.
   1507 	 */
   1508 	if (ALTQ_IS_ENABLED(&dst_ifp->if_snd)) {
   1509 		/* XXX IFT_ETHER */
   1510 		altq_etherclassify(&dst_ifp->if_snd, m, &pktattr);
   1511 	}
   1512 #endif /* ALTQ */
   1513 
   1514 	len = m->m_pkthdr.len;
   1515 	m->m_flags |= M_PROTO1;
   1516 	mflags = m->m_flags;
   1517 
   1518 	IFQ_ENQUEUE(&dst_ifp->if_snd, m, &pktattr, error);
   1519 
   1520 	if (error) {
   1521 		/* mbuf is already freed */
   1522 		sc->sc_if.if_oerrors++;
   1523 		return;
   1524 	}
   1525 
   1526 	sc->sc_if.if_opackets++;
   1527 	sc->sc_if.if_obytes += len;
   1528 
   1529 	dst_ifp->if_obytes += len;
   1530 
   1531 	if (mflags & M_MCAST) {
   1532 		sc->sc_if.if_omcasts++;
   1533 		dst_ifp->if_omcasts++;
   1534 	}
   1535 
   1536 	if ((dst_ifp->if_flags & IFF_OACTIVE) == 0)
   1537 		(*dst_ifp->if_start)(dst_ifp);
   1538 }
   1539 
   1540 /*
   1541  * bridge_output:
   1542  *
   1543  *	Send output from a bridge member interface.  This
   1544  *	performs the bridging function for locally originated
   1545  *	packets.
   1546  *
   1547  *	The mbuf has the Ethernet header already attached.  We must
   1548  *	enqueue or free the mbuf before returning.
   1549  */
   1550 int
   1551 bridge_output(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *sa,
   1552     struct rtentry *rt)
   1553 {
   1554 	struct ether_header *eh;
   1555 	struct ifnet *dst_if;
   1556 	struct bridge_softc *sc;
   1557 #ifndef BRIDGE_MPSAFE
   1558 	int s;
   1559 #endif
   1560 
   1561 	if (m->m_len < ETHER_HDR_LEN) {
   1562 		m = m_pullup(m, ETHER_HDR_LEN);
   1563 		if (m == NULL)
   1564 			return (0);
   1565 	}
   1566 
   1567 	eh = mtod(m, struct ether_header *);
   1568 	sc = ifp->if_bridge;
   1569 
   1570 #ifndef BRIDGE_MPSAFE
   1571 	s = splnet();
   1572 #endif
   1573 
   1574 	/*
   1575 	 * If bridge is down, but the original output interface is up,
   1576 	 * go ahead and send out that interface.  Otherwise, the packet
   1577 	 * is dropped below.
   1578 	 */
   1579 	if (__predict_false(sc == NULL) ||
   1580 	    (sc->sc_if.if_flags & IFF_RUNNING) == 0) {
   1581 		dst_if = ifp;
   1582 		goto sendunicast;
   1583 	}
   1584 
   1585 	/*
   1586 	 * If the packet is a multicast, or we don't know a better way to
   1587 	 * get there, send to all interfaces.
   1588 	 */
   1589 	if (ETHER_IS_MULTICAST(eh->ether_dhost))
   1590 		dst_if = NULL;
   1591 	else
   1592 		dst_if = bridge_rtlookup(sc, eh->ether_dhost);
   1593 	if (dst_if == NULL) {
   1594 		struct bridge_iflist *bif;
   1595 		struct mbuf *mc;
   1596 		int used = 0;
   1597 		int ss;
   1598 
   1599 		BRIDGE_PSZ_RENTER(ss);
   1600 		LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
   1601 			bif = bridge_try_hold_bif(bif);
   1602 			if (bif == NULL)
   1603 				continue;
   1604 			BRIDGE_PSZ_REXIT(ss);
   1605 
   1606 			dst_if = bif->bif_ifp;
   1607 			if ((dst_if->if_flags & IFF_RUNNING) == 0)
   1608 				goto next;
   1609 
   1610 			/*
   1611 			 * If this is not the original output interface,
   1612 			 * and the interface is participating in spanning
   1613 			 * tree, make sure the port is in a state that
   1614 			 * allows forwarding.
   1615 			 */
   1616 			if (dst_if != ifp &&
   1617 			    (bif->bif_flags & IFBIF_STP) != 0) {
   1618 				switch (bif->bif_state) {
   1619 				case BSTP_IFSTATE_BLOCKING:
   1620 				case BSTP_IFSTATE_LISTENING:
   1621 				case BSTP_IFSTATE_DISABLED:
   1622 					goto next;
   1623 				}
   1624 			}
   1625 
   1626 			if (LIST_NEXT(bif, bif_next) == NULL) {
   1627 				used = 1;
   1628 				mc = m;
   1629 			} else {
   1630 				mc = m_copym(m, 0, M_COPYALL, M_NOWAIT);
   1631 				if (mc == NULL) {
   1632 					sc->sc_if.if_oerrors++;
   1633 					goto next;
   1634 				}
   1635 			}
   1636 
   1637 			bridge_enqueue(sc, dst_if, mc, 0);
   1638 next:
   1639 			bridge_release_member(sc, bif);
   1640 			BRIDGE_PSZ_RENTER(ss);
   1641 		}
   1642 		BRIDGE_PSZ_REXIT(ss);
   1643 
   1644 		if (used == 0)
   1645 			m_freem(m);
   1646 #ifndef BRIDGE_MPSAFE
   1647 		splx(s);
   1648 #endif
   1649 		return (0);
   1650 	}
   1651 
   1652  sendunicast:
   1653 	/*
   1654 	 * XXX Spanning tree consideration here?
   1655 	 */
   1656 
   1657 	if ((dst_if->if_flags & IFF_RUNNING) == 0) {
   1658 		m_freem(m);
   1659 #ifndef BRIDGE_MPSAFE
   1660 		splx(s);
   1661 #endif
   1662 		return (0);
   1663 	}
   1664 
   1665 	bridge_enqueue(sc, dst_if, m, 0);
   1666 
   1667 #ifndef BRIDGE_MPSAFE
   1668 	splx(s);
   1669 #endif
   1670 	return (0);
   1671 }
   1672 
   1673 /*
   1674  * bridge_start:
   1675  *
   1676  *	Start output on a bridge.
   1677  *
   1678  *	NOTE: This routine should never be called in this implementation.
   1679  */
   1680 static void
   1681 bridge_start(struct ifnet *ifp)
   1682 {
   1683 
   1684 	printf("%s: bridge_start() called\n", ifp->if_xname);
   1685 }
   1686 
   1687 /*
   1688  * bridge_forward:
   1689  *
   1690  *	The forwarding function of the bridge.
   1691  */
   1692 static void
   1693 bridge_forward(void *v)
   1694 {
   1695 	struct bridge_softc *sc = v;
   1696 	struct mbuf *m;
   1697 	struct bridge_iflist *bif;
   1698 	struct ifnet *src_if, *dst_if;
   1699 	struct ether_header *eh;
   1700 #ifndef BRIDGE_MPSAFE
   1701 	int s;
   1702 
   1703 	KERNEL_LOCK(1, NULL);
   1704 	mutex_enter(softnet_lock);
   1705 #endif
   1706 
   1707 	if ((sc->sc_if.if_flags & IFF_RUNNING) == 0) {
   1708 #ifndef BRIDGE_MPSAFE
   1709 		mutex_exit(softnet_lock);
   1710 		KERNEL_UNLOCK_ONE(NULL);
   1711 #endif
   1712 		return;
   1713 	}
   1714 
   1715 #ifndef BRIDGE_MPSAFE
   1716 	s = splnet();
   1717 #endif
   1718 	while ((m = pktq_dequeue(sc->sc_fwd_pktq)) != NULL) {
   1719 		src_if = m->m_pkthdr.rcvif;
   1720 
   1721 		sc->sc_if.if_ipackets++;
   1722 		sc->sc_if.if_ibytes += m->m_pkthdr.len;
   1723 
   1724 		/*
   1725 		 * Look up the bridge_iflist.
   1726 		 */
   1727 		bif = bridge_lookup_member_if(sc, src_if);
   1728 		if (bif == NULL) {
   1729 			/* Interface is not a bridge member (anymore?) */
   1730 			m_freem(m);
   1731 			continue;
   1732 		}
   1733 
   1734 		if (bif->bif_flags & IFBIF_STP) {
   1735 			switch (bif->bif_state) {
   1736 			case BSTP_IFSTATE_BLOCKING:
   1737 			case BSTP_IFSTATE_LISTENING:
   1738 			case BSTP_IFSTATE_DISABLED:
   1739 				m_freem(m);
   1740 				bridge_release_member(sc, bif);
   1741 				continue;
   1742 			}
   1743 		}
   1744 
   1745 		eh = mtod(m, struct ether_header *);
   1746 
   1747 		/*
   1748 		 * If the interface is learning, and the source
   1749 		 * address is valid and not multicast, record
   1750 		 * the address.
   1751 		 */
   1752 		if ((bif->bif_flags & IFBIF_LEARNING) != 0 &&
   1753 		    ETHER_IS_MULTICAST(eh->ether_shost) == 0 &&
   1754 		    (eh->ether_shost[0] == 0 &&
   1755 		     eh->ether_shost[1] == 0 &&
   1756 		     eh->ether_shost[2] == 0 &&
   1757 		     eh->ether_shost[3] == 0 &&
   1758 		     eh->ether_shost[4] == 0 &&
   1759 		     eh->ether_shost[5] == 0) == 0) {
   1760 			(void) bridge_rtupdate(sc, eh->ether_shost,
   1761 			    src_if, 0, IFBAF_DYNAMIC);
   1762 		}
   1763 
   1764 		if ((bif->bif_flags & IFBIF_STP) != 0 &&
   1765 		    bif->bif_state == BSTP_IFSTATE_LEARNING) {
   1766 			m_freem(m);
   1767 			bridge_release_member(sc, bif);
   1768 			continue;
   1769 		}
   1770 
   1771 		bridge_release_member(sc, bif);
   1772 
   1773 		/*
   1774 		 * At this point, the port either doesn't participate
   1775 		 * in spanning tree or it is in the forwarding state.
   1776 		 */
   1777 
   1778 		/*
   1779 		 * If the packet is unicast, destined for someone on
   1780 		 * "this" side of the bridge, drop it.
   1781 		 */
   1782 		if ((m->m_flags & (M_BCAST|M_MCAST)) == 0) {
   1783 			dst_if = bridge_rtlookup(sc, eh->ether_dhost);
   1784 			if (src_if == dst_if) {
   1785 				m_freem(m);
   1786 				continue;
   1787 			}
   1788 		} else {
   1789 			/* ...forward it to all interfaces. */
   1790 			sc->sc_if.if_imcasts++;
   1791 			dst_if = NULL;
   1792 		}
   1793 
   1794 		if (pfil_run_hooks(sc->sc_if.if_pfil, &m,
   1795 		    m->m_pkthdr.rcvif, PFIL_IN) != 0) {
   1796 			if (m != NULL)
   1797 				m_freem(m);
   1798 			continue;
   1799 		}
   1800 		if (m == NULL)
   1801 			continue;
   1802 
   1803 		if (dst_if == NULL) {
   1804 			bridge_broadcast(sc, src_if, m);
   1805 			continue;
   1806 		}
   1807 
   1808 		/*
   1809 		 * At this point, we're dealing with a unicast frame
   1810 		 * going to a different interface.
   1811 		 */
   1812 		if ((dst_if->if_flags & IFF_RUNNING) == 0) {
   1813 			m_freem(m);
   1814 			continue;
   1815 		}
   1816 
   1817 		bif = bridge_lookup_member_if(sc, dst_if);
   1818 		if (bif == NULL) {
   1819 			/* Not a member of the bridge (anymore?) */
   1820 			m_freem(m);
   1821 			continue;
   1822 		}
   1823 
   1824 		if (bif->bif_flags & IFBIF_STP) {
   1825 			switch (bif->bif_state) {
   1826 			case BSTP_IFSTATE_DISABLED:
   1827 			case BSTP_IFSTATE_BLOCKING:
   1828 				m_freem(m);
   1829 				bridge_release_member(sc, bif);
   1830 				continue;
   1831 			}
   1832 		}
   1833 
   1834 		bridge_release_member(sc, bif);
   1835 
   1836 		bridge_enqueue(sc, dst_if, m, 1);
   1837 	}
   1838 #ifndef BRIDGE_MPSAFE
   1839 	splx(s);
   1840 	mutex_exit(softnet_lock);
   1841 	KERNEL_UNLOCK_ONE(NULL);
   1842 #endif
   1843 }
   1844 
   1845 static bool
   1846 bstp_state_before_learning(struct bridge_iflist *bif)
   1847 {
   1848 	if (bif->bif_flags & IFBIF_STP) {
   1849 		switch (bif->bif_state) {
   1850 		case BSTP_IFSTATE_BLOCKING:
   1851 		case BSTP_IFSTATE_LISTENING:
   1852 		case BSTP_IFSTATE_DISABLED:
   1853 			return true;
   1854 		}
   1855 	}
   1856 	return false;
   1857 }
   1858 
   1859 static bool
   1860 bridge_ourether(struct bridge_iflist *bif, struct ether_header *eh, int src)
   1861 {
   1862 	uint8_t *ether = src ? eh->ether_shost : eh->ether_dhost;
   1863 
   1864 	if (memcmp(CLLADDR(bif->bif_ifp->if_sadl), ether, ETHER_ADDR_LEN) == 0
   1865 #if NCARP > 0
   1866 	    || (bif->bif_ifp->if_carp &&
   1867 	        carp_ourether(bif->bif_ifp->if_carp, eh, IFT_ETHER, src) != NULL)
   1868 #endif /* NCARP > 0 */
   1869 	    )
   1870 		return true;
   1871 
   1872 	return false;
   1873 }
   1874 
   1875 /*
   1876  * bridge_input:
   1877  *
   1878  *	Receive input from a member interface.  Queue the packet for
   1879  *	bridging if it is not for us.
   1880  */
   1881 static void
   1882 bridge_input(struct ifnet *ifp, struct mbuf *m)
   1883 {
   1884 	struct bridge_softc *sc = ifp->if_bridge;
   1885 	struct bridge_iflist *bif;
   1886 	struct ether_header *eh;
   1887 
   1888 	if (__predict_false(sc == NULL) ||
   1889 	    (sc->sc_if.if_flags & IFF_RUNNING) == 0) {
   1890 		ether_input(ifp, m);
   1891 		return;
   1892 	}
   1893 
   1894 	bif = bridge_lookup_member_if(sc, ifp);
   1895 	if (bif == NULL) {
   1896 		ether_input(ifp, m);
   1897 		return;
   1898 	}
   1899 
   1900 	eh = mtod(m, struct ether_header *);
   1901 
   1902 	if (ETHER_IS_MULTICAST(eh->ether_dhost)) {
   1903 		if (memcmp(etherbroadcastaddr,
   1904 		    eh->ether_dhost, ETHER_ADDR_LEN) == 0)
   1905 			m->m_flags |= M_BCAST;
   1906 		else
   1907 			m->m_flags |= M_MCAST;
   1908 	}
   1909 
   1910 	/*
   1911 	 * A 'fast' path for packets addressed to interfaces that are
   1912 	 * part of this bridge.
   1913 	 */
   1914 	if (!(m->m_flags & (M_BCAST|M_MCAST)) &&
   1915 	    !bstp_state_before_learning(bif)) {
   1916 		struct bridge_iflist *_bif;
   1917 		struct ifnet *_ifp = NULL;
   1918 		int s;
   1919 
   1920 		BRIDGE_PSZ_RENTER(s);
   1921 		LIST_FOREACH(_bif, &sc->sc_iflist, bif_next) {
   1922 			/* It is destined for us. */
   1923 			if (bridge_ourether(_bif, eh, 0)) {
   1924 				_bif = bridge_try_hold_bif(_bif);
   1925 				BRIDGE_PSZ_REXIT(s);
   1926 				if (_bif == NULL)
   1927 					goto out;
   1928 				if (_bif->bif_flags & IFBIF_LEARNING)
   1929 					(void) bridge_rtupdate(sc,
   1930 					    eh->ether_shost, ifp, 0, IFBAF_DYNAMIC);
   1931 				_ifp = m->m_pkthdr.rcvif = _bif->bif_ifp;
   1932 				bridge_release_member(sc, _bif);
   1933 				goto out;
   1934 			}
   1935 
   1936 			/* We just received a packet that we sent out. */
   1937 			if (bridge_ourether(_bif, eh, 1))
   1938 				break;
   1939 		}
   1940 		BRIDGE_PSZ_REXIT(s);
   1941 out:
   1942 
   1943 		if (_bif != NULL) {
   1944 			bridge_release_member(sc, bif);
   1945 			if (_ifp != NULL)
   1946 				ether_input(_ifp, m);
   1947 			else
   1948 				m_freem(m);
   1949 			return;
   1950 		}
   1951 	}
   1952 
   1953 	/* Tap off 802.1D packets; they do not get forwarded. */
   1954 	if (bif->bif_flags & IFBIF_STP &&
   1955 	    memcmp(eh->ether_dhost, bstp_etheraddr, ETHER_ADDR_LEN) == 0) {
   1956 		bstp_input(sc, bif, m);
   1957 		bridge_release_member(sc, bif);
   1958 		return;
   1959 	}
   1960 
   1961 	/*
   1962 	 * A normal switch would discard the packet here, but that's not what
   1963 	 * we've done historically. This also prevents some obnoxious behaviour.
   1964 	 */
   1965 	if (bstp_state_before_learning(bif)) {
   1966 		bridge_release_member(sc, bif);
   1967 		ether_input(ifp, m);
   1968 		return;
   1969 	}
   1970 
   1971 	bridge_release_member(sc, bif);
   1972 
   1973 	/* Queue the packet for bridge forwarding. */
   1974 	if (__predict_false(!pktq_enqueue(sc->sc_fwd_pktq, m, 0)))
   1975 		m_freem(m);
   1976 }
   1977 
   1978 /*
   1979  * bridge_broadcast:
   1980  *
   1981  *	Send a frame to all interfaces that are members of
   1982  *	the bridge, except for the one on which the packet
   1983  *	arrived.
   1984  */
   1985 static void
   1986 bridge_broadcast(struct bridge_softc *sc, struct ifnet *src_if,
   1987     struct mbuf *m)
   1988 {
   1989 	struct bridge_iflist *bif;
   1990 	struct mbuf *mc;
   1991 	struct ifnet *dst_if;
   1992 	bool used, bmcast;
   1993 	int s;
   1994 
   1995 	used = bmcast = m->m_flags & (M_BCAST|M_MCAST);
   1996 
   1997 	BRIDGE_PSZ_RENTER(s);
   1998 	LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
   1999 		bif = bridge_try_hold_bif(bif);
   2000 		if (bif == NULL)
   2001 			continue;
   2002 		BRIDGE_PSZ_REXIT(s);
   2003 
   2004 		dst_if = bif->bif_ifp;
   2005 		if (dst_if == src_if)
   2006 			goto next;
   2007 
   2008 		if (bif->bif_flags & IFBIF_STP) {
   2009 			switch (bif->bif_state) {
   2010 			case BSTP_IFSTATE_BLOCKING:
   2011 			case BSTP_IFSTATE_DISABLED:
   2012 				goto next;
   2013 			}
   2014 		}
   2015 
   2016 		if ((bif->bif_flags & IFBIF_DISCOVER) == 0 && !bmcast)
   2017 			goto next;
   2018 
   2019 		if ((dst_if->if_flags & IFF_RUNNING) == 0)
   2020 			goto next;
   2021 
   2022 		if (!used && LIST_NEXT(bif, bif_next) == NULL) {
   2023 			mc = m;
   2024 			used = true;
   2025 		} else {
   2026 			mc = m_copym(m, 0, M_COPYALL, M_DONTWAIT);
   2027 			if (mc == NULL) {
   2028 				sc->sc_if.if_oerrors++;
   2029 				goto next;
   2030 			}
   2031 		}
   2032 
   2033 		bridge_enqueue(sc, dst_if, mc, 1);
   2034 next:
   2035 		bridge_release_member(sc, bif);
   2036 		BRIDGE_PSZ_RENTER(s);
   2037 	}
   2038 	BRIDGE_PSZ_REXIT(s);
   2039 
   2040 	if (bmcast)
   2041 		ether_input(src_if, m);
   2042 	else if (!used)
   2043 		m_freem(m);
   2044 }
   2045 
   2046 static int
   2047 bridge_rtalloc(struct bridge_softc *sc, const uint8_t *dst,
   2048     struct bridge_rtnode **brtp)
   2049 {
   2050 	struct bridge_rtnode *brt;
   2051 	int error;
   2052 
   2053 	if (sc->sc_brtcnt >= sc->sc_brtmax)
   2054 		return ENOSPC;
   2055 
   2056 	/*
   2057 	 * Allocate a new bridge forwarding node, and
   2058 	 * initialize the expiration time and Ethernet
   2059 	 * address.
   2060 	 */
   2061 	brt = pool_get(&bridge_rtnode_pool, PR_NOWAIT);
   2062 	if (brt == NULL)
   2063 		return ENOMEM;
   2064 
   2065 	memset(brt, 0, sizeof(*brt));
   2066 	brt->brt_expire = time_uptime + sc->sc_brttimeout;
   2067 	brt->brt_flags = IFBAF_DYNAMIC;
   2068 	memcpy(brt->brt_addr, dst, ETHER_ADDR_LEN);
   2069 
   2070 	BRIDGE_RT_INTR_LOCK(sc);
   2071 	error = bridge_rtnode_insert(sc, brt);
   2072 	BRIDGE_RT_INTR_UNLOCK(sc);
   2073 
   2074 	if (error != 0) {
   2075 		pool_put(&bridge_rtnode_pool, brt);
   2076 		return error;
   2077 	}
   2078 
   2079 	*brtp = brt;
   2080 	return 0;
   2081 }
   2082 
   2083 /*
   2084  * bridge_rtupdate:
   2085  *
   2086  *	Add a bridge routing entry.
   2087  */
   2088 static int
   2089 bridge_rtupdate(struct bridge_softc *sc, const uint8_t *dst,
   2090     struct ifnet *dst_if, int setflags, uint8_t flags)
   2091 {
   2092 	struct bridge_rtnode *brt;
   2093 	int s;
   2094 
   2095 again:
   2096 	/*
   2097 	 * A route for this destination might already exist.  If so,
   2098 	 * update it, otherwise create a new one.
   2099 	 */
   2100 	BRIDGE_RT_RENTER(s);
   2101 	brt = bridge_rtnode_lookup(sc, dst);
   2102 
   2103 	if (brt != NULL) {
   2104 		brt->brt_ifp = dst_if;
   2105 		if (setflags) {
   2106 			brt->brt_flags = flags;
   2107 			if (flags & IFBAF_STATIC)
   2108 				brt->brt_expire = 0;
   2109 			else
   2110 				brt->brt_expire = time_uptime + sc->sc_brttimeout;
   2111 		} else {
   2112 			if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
   2113 				brt->brt_expire = time_uptime + sc->sc_brttimeout;
   2114 		}
   2115 	}
   2116 	BRIDGE_RT_REXIT(s);
   2117 
   2118 	if (brt == NULL) {
   2119 		int r;
   2120 
   2121 		r = bridge_rtalloc(sc, dst, &brt);
   2122 		if (r != 0)
   2123 			return r;
   2124 		goto again;
   2125 	}
   2126 
   2127 	return 0;
   2128 }
   2129 
   2130 /*
   2131  * bridge_rtlookup:
   2132  *
   2133  *	Lookup the destination interface for an address.
   2134  */
   2135 static struct ifnet *
   2136 bridge_rtlookup(struct bridge_softc *sc, const uint8_t *addr)
   2137 {
   2138 	struct bridge_rtnode *brt;
   2139 	struct ifnet *ifs = NULL;
   2140 	int s;
   2141 
   2142 	BRIDGE_RT_RENTER(s);
   2143 	brt = bridge_rtnode_lookup(sc, addr);
   2144 	if (brt != NULL)
   2145 		ifs = brt->brt_ifp;
   2146 	BRIDGE_RT_REXIT(s);
   2147 
   2148 	return ifs;
   2149 }
   2150 
   2151 typedef bool (*bridge_iterate_cb_t)
   2152     (struct bridge_softc *, struct bridge_rtnode *, bool *, void *);
   2153 
   2154 /*
   2155  * bridge_rtlist_iterate_remove:
   2156  *
   2157  *	It iterates on sc->sc_rtlist and removes rtnodes of it which func
   2158  *	callback judges to remove. Removals of rtnodes are done in a manner
   2159  *	of pserialize. To this end, all kmem_* operations are placed out of
   2160  *	mutexes.
   2161  */
   2162 static void
   2163 bridge_rtlist_iterate_remove(struct bridge_softc *sc, bridge_iterate_cb_t func, void *arg)
   2164 {
   2165 	struct bridge_rtnode *brt, *nbrt;
   2166 	struct bridge_rtnode **brt_list;
   2167 	int i, count;
   2168 
   2169 retry:
   2170 	count = sc->sc_brtcnt;
   2171 	if (count == 0)
   2172 		return;
   2173 	brt_list = kmem_alloc(sizeof(struct bridge_rtnode *) * count, KM_SLEEP);
   2174 
   2175 	BRIDGE_RT_LOCK(sc);
   2176 	BRIDGE_RT_INTR_LOCK(sc);
   2177 	if (__predict_false(sc->sc_brtcnt > count)) {
   2178 		/* The rtnodes increased, we need more memory */
   2179 		BRIDGE_RT_INTR_UNLOCK(sc);
   2180 		BRIDGE_RT_UNLOCK(sc);
   2181 		kmem_free(brt_list, sizeof(*brt_list) * count);
   2182 		goto retry;
   2183 	}
   2184 
   2185 	i = 0;
   2186 	LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
   2187 		bool need_break = false;
   2188 		if (func(sc, brt, &need_break, arg)) {
   2189 			bridge_rtnode_remove(sc, brt);
   2190 			brt_list[i++] = brt;
   2191 		}
   2192 		if (need_break)
   2193 			break;
   2194 	}
   2195 	BRIDGE_RT_INTR_UNLOCK(sc);
   2196 
   2197 	if (i > 0)
   2198 		BRIDGE_RT_PSZ_PERFORM(sc);
   2199 	BRIDGE_RT_UNLOCK(sc);
   2200 
   2201 	while (--i >= 0)
   2202 		bridge_rtnode_destroy(brt_list[i]);
   2203 
   2204 	kmem_free(brt_list, sizeof(*brt_list) * count);
   2205 }
   2206 
   2207 static bool
   2208 bridge_rttrim0_cb(struct bridge_softc *sc, struct bridge_rtnode *brt,
   2209     bool *need_break, void *arg)
   2210 {
   2211 	if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) {
   2212 		/* Take into account of the subsequent removal */
   2213 		if ((sc->sc_brtcnt - 1) <= sc->sc_brtmax)
   2214 			*need_break = true;
   2215 		return true;
   2216 	} else
   2217 		return false;
   2218 }
   2219 
   2220 static void
   2221 bridge_rttrim0(struct bridge_softc *sc)
   2222 {
   2223 	bridge_rtlist_iterate_remove(sc, bridge_rttrim0_cb, NULL);
   2224 }
   2225 
   2226 /*
   2227  * bridge_rttrim:
   2228  *
   2229  *	Trim the routine table so that we have a number
   2230  *	of routing entries less than or equal to the
   2231  *	maximum number.
   2232  */
   2233 static void
   2234 bridge_rttrim(struct bridge_softc *sc)
   2235 {
   2236 
   2237 	/* Make sure we actually need to do this. */
   2238 	if (sc->sc_brtcnt <= sc->sc_brtmax)
   2239 		return;
   2240 
   2241 	/* Force an aging cycle; this might trim enough addresses. */
   2242 	bridge_rtage(sc);
   2243 	if (sc->sc_brtcnt <= sc->sc_brtmax)
   2244 		return;
   2245 
   2246 	bridge_rttrim0(sc);
   2247 
   2248 	return;
   2249 }
   2250 
   2251 /*
   2252  * bridge_timer:
   2253  *
   2254  *	Aging timer for the bridge.
   2255  */
   2256 static void
   2257 bridge_timer(void *arg)
   2258 {
   2259 	struct bridge_softc *sc = arg;
   2260 
   2261 	workqueue_enqueue(sc->sc_rtage_wq, &bridge_rtage_wk, NULL);
   2262 }
   2263 
   2264 static void
   2265 bridge_rtage_work(struct work *wk, void *arg)
   2266 {
   2267 	struct bridge_softc *sc = arg;
   2268 
   2269 	KASSERT(wk == &bridge_rtage_wk);
   2270 
   2271 	bridge_rtage(sc);
   2272 
   2273 	if (sc->sc_if.if_flags & IFF_RUNNING)
   2274 		callout_reset(&sc->sc_brcallout,
   2275 		    bridge_rtable_prune_period * hz, bridge_timer, sc);
   2276 }
   2277 
   2278 static bool
   2279 bridge_rtage_cb(struct bridge_softc *sc, struct bridge_rtnode *brt,
   2280     bool *need_break, void *arg)
   2281 {
   2282 	if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC &&
   2283 	    time_uptime >= brt->brt_expire)
   2284 		return true;
   2285 	else
   2286 		return false;
   2287 }
   2288 
   2289 /*
   2290  * bridge_rtage:
   2291  *
   2292  *	Perform an aging cycle.
   2293  */
   2294 static void
   2295 bridge_rtage(struct bridge_softc *sc)
   2296 {
   2297 	bridge_rtlist_iterate_remove(sc, bridge_rtage_cb, NULL);
   2298 }
   2299 
   2300 
   2301 static bool
   2302 bridge_rtflush_cb(struct bridge_softc *sc, struct bridge_rtnode *brt,
   2303     bool *need_break, void *arg)
   2304 {
   2305 	int full = *(int*)arg;
   2306 
   2307 	if (full || (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
   2308 		return true;
   2309 	else
   2310 		return false;
   2311 }
   2312 
   2313 /*
   2314  * bridge_rtflush:
   2315  *
   2316  *	Remove all dynamic addresses from the bridge.
   2317  */
   2318 static void
   2319 bridge_rtflush(struct bridge_softc *sc, int full)
   2320 {
   2321 	bridge_rtlist_iterate_remove(sc, bridge_rtflush_cb, &full);
   2322 }
   2323 
   2324 /*
   2325  * bridge_rtdaddr:
   2326  *
   2327  *	Remove an address from the table.
   2328  */
   2329 static int
   2330 bridge_rtdaddr(struct bridge_softc *sc, const uint8_t *addr)
   2331 {
   2332 	struct bridge_rtnode *brt;
   2333 
   2334 	BRIDGE_RT_LOCK(sc);
   2335 	BRIDGE_RT_INTR_LOCK(sc);
   2336 	if ((brt = bridge_rtnode_lookup(sc, addr)) == NULL) {
   2337 		BRIDGE_RT_INTR_UNLOCK(sc);
   2338 		BRIDGE_RT_UNLOCK(sc);
   2339 		return ENOENT;
   2340 	}
   2341 	bridge_rtnode_remove(sc, brt);
   2342 	BRIDGE_RT_INTR_UNLOCK(sc);
   2343 	BRIDGE_RT_PSZ_PERFORM(sc);
   2344 	BRIDGE_RT_UNLOCK(sc);
   2345 
   2346 	bridge_rtnode_destroy(brt);
   2347 
   2348 	return 0;
   2349 }
   2350 
   2351 /*
   2352  * bridge_rtdelete:
   2353  *
   2354  *	Delete routes to a speicifc member interface.
   2355  */
   2356 static void
   2357 bridge_rtdelete(struct bridge_softc *sc, struct ifnet *ifp)
   2358 {
   2359 	struct bridge_rtnode *brt, *nbrt;
   2360 
   2361 	BRIDGE_RT_LOCK(sc);
   2362 	BRIDGE_RT_INTR_LOCK(sc);
   2363 	LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
   2364 		if (brt->brt_ifp == ifp)
   2365 			break;
   2366 	}
   2367 	if (brt == NULL) {
   2368 		BRIDGE_RT_INTR_UNLOCK(sc);
   2369 		BRIDGE_RT_UNLOCK(sc);
   2370 		return;
   2371 	}
   2372 	bridge_rtnode_remove(sc, brt);
   2373 	BRIDGE_RT_INTR_UNLOCK(sc);
   2374 	BRIDGE_RT_PSZ_PERFORM(sc);
   2375 	BRIDGE_RT_UNLOCK(sc);
   2376 
   2377 	bridge_rtnode_destroy(brt);
   2378 }
   2379 
   2380 /*
   2381  * bridge_rtable_init:
   2382  *
   2383  *	Initialize the route table for this bridge.
   2384  */
   2385 static void
   2386 bridge_rtable_init(struct bridge_softc *sc)
   2387 {
   2388 	int i;
   2389 
   2390 	sc->sc_rthash = kmem_alloc(sizeof(*sc->sc_rthash) * BRIDGE_RTHASH_SIZE,
   2391 	    KM_SLEEP);
   2392 
   2393 	for (i = 0; i < BRIDGE_RTHASH_SIZE; i++)
   2394 		LIST_INIT(&sc->sc_rthash[i]);
   2395 
   2396 	sc->sc_rthash_key = cprng_fast32();
   2397 
   2398 	LIST_INIT(&sc->sc_rtlist);
   2399 
   2400 	sc->sc_rtlist_intr_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_NET);
   2401 #ifdef BRIDGE_MPSAFE
   2402 	sc->sc_rtlist_psz = pserialize_create();
   2403 	sc->sc_rtlist_lock = mutex_obj_alloc(MUTEX_DEFAULT, IPL_SOFTNET);
   2404 #else
   2405 	sc->sc_rtlist_psz = NULL;
   2406 	sc->sc_rtlist_lock = NULL;
   2407 #endif
   2408 }
   2409 
   2410 /*
   2411  * bridge_rtable_fini:
   2412  *
   2413  *	Deconstruct the route table for this bridge.
   2414  */
   2415 static void
   2416 bridge_rtable_fini(struct bridge_softc *sc)
   2417 {
   2418 
   2419 	kmem_free(sc->sc_rthash, sizeof(*sc->sc_rthash) * BRIDGE_RTHASH_SIZE);
   2420 	if (sc->sc_rtlist_intr_lock)
   2421 		mutex_obj_free(sc->sc_rtlist_intr_lock);
   2422 	if (sc->sc_rtlist_lock)
   2423 		mutex_obj_free(sc->sc_rtlist_lock);
   2424 	if (sc->sc_rtlist_psz)
   2425 		pserialize_destroy(sc->sc_rtlist_psz);
   2426 }
   2427 
   2428 /*
   2429  * The following hash function is adapted from "Hash Functions" by Bob Jenkins
   2430  * ("Algorithm Alley", Dr. Dobbs Journal, September 1997).
   2431  */
   2432 #define	mix(a, b, c)							\
   2433 do {									\
   2434 	a -= b; a -= c; a ^= (c >> 13);					\
   2435 	b -= c; b -= a; b ^= (a << 8);					\
   2436 	c -= a; c -= b; c ^= (b >> 13);					\
   2437 	a -= b; a -= c; a ^= (c >> 12);					\
   2438 	b -= c; b -= a; b ^= (a << 16);					\
   2439 	c -= a; c -= b; c ^= (b >> 5);					\
   2440 	a -= b; a -= c; a ^= (c >> 3);					\
   2441 	b -= c; b -= a; b ^= (a << 10);					\
   2442 	c -= a; c -= b; c ^= (b >> 15);					\
   2443 } while (/*CONSTCOND*/0)
   2444 
   2445 static inline uint32_t
   2446 bridge_rthash(struct bridge_softc *sc, const uint8_t *addr)
   2447 {
   2448 	uint32_t a = 0x9e3779b9, b = 0x9e3779b9, c = sc->sc_rthash_key;
   2449 
   2450 	b += addr[5] << 8;
   2451 	b += addr[4];
   2452 	a += addr[3] << 24;
   2453 	a += addr[2] << 16;
   2454 	a += addr[1] << 8;
   2455 	a += addr[0];
   2456 
   2457 	mix(a, b, c);
   2458 
   2459 	return (c & BRIDGE_RTHASH_MASK);
   2460 }
   2461 
   2462 #undef mix
   2463 
   2464 /*
   2465  * bridge_rtnode_lookup:
   2466  *
   2467  *	Look up a bridge route node for the specified destination.
   2468  */
   2469 static struct bridge_rtnode *
   2470 bridge_rtnode_lookup(struct bridge_softc *sc, const uint8_t *addr)
   2471 {
   2472 	struct bridge_rtnode *brt;
   2473 	uint32_t hash;
   2474 	int dir;
   2475 
   2476 	hash = bridge_rthash(sc, addr);
   2477 	LIST_FOREACH(brt, &sc->sc_rthash[hash], brt_hash) {
   2478 		dir = memcmp(addr, brt->brt_addr, ETHER_ADDR_LEN);
   2479 		if (dir == 0)
   2480 			return (brt);
   2481 		if (dir > 0)
   2482 			return (NULL);
   2483 	}
   2484 
   2485 	return (NULL);
   2486 }
   2487 
   2488 /*
   2489  * bridge_rtnode_insert:
   2490  *
   2491  *	Insert the specified bridge node into the route table.  We
   2492  *	assume the entry is not already in the table.
   2493  */
   2494 static int
   2495 bridge_rtnode_insert(struct bridge_softc *sc, struct bridge_rtnode *brt)
   2496 {
   2497 	struct bridge_rtnode *lbrt;
   2498 	uint32_t hash;
   2499 	int dir;
   2500 
   2501 	KASSERT(BRIDGE_RT_INTR_LOCKED(sc));
   2502 
   2503 	hash = bridge_rthash(sc, brt->brt_addr);
   2504 
   2505 	lbrt = LIST_FIRST(&sc->sc_rthash[hash]);
   2506 	if (lbrt == NULL) {
   2507 		LIST_INSERT_HEAD(&sc->sc_rthash[hash], brt, brt_hash);
   2508 		goto out;
   2509 	}
   2510 
   2511 	do {
   2512 		dir = memcmp(brt->brt_addr, lbrt->brt_addr, ETHER_ADDR_LEN);
   2513 		if (dir == 0)
   2514 			return (EEXIST);
   2515 		if (dir > 0) {
   2516 			LIST_INSERT_BEFORE(lbrt, brt, brt_hash);
   2517 			goto out;
   2518 		}
   2519 		if (LIST_NEXT(lbrt, brt_hash) == NULL) {
   2520 			LIST_INSERT_AFTER(lbrt, brt, brt_hash);
   2521 			goto out;
   2522 		}
   2523 		lbrt = LIST_NEXT(lbrt, brt_hash);
   2524 	} while (lbrt != NULL);
   2525 
   2526 #ifdef DIAGNOSTIC
   2527 	panic("bridge_rtnode_insert: impossible");
   2528 #endif
   2529 
   2530  out:
   2531 	LIST_INSERT_HEAD(&sc->sc_rtlist, brt, brt_list);
   2532 	sc->sc_brtcnt++;
   2533 
   2534 	return (0);
   2535 }
   2536 
   2537 /*
   2538  * bridge_rtnode_remove:
   2539  *
   2540  *	Remove a bridge rtnode from the rthash and the rtlist of a bridge.
   2541  */
   2542 static void
   2543 bridge_rtnode_remove(struct bridge_softc *sc, struct bridge_rtnode *brt)
   2544 {
   2545 
   2546 	KASSERT(BRIDGE_RT_INTR_LOCKED(sc));
   2547 
   2548 	LIST_REMOVE(brt, brt_hash);
   2549 	LIST_REMOVE(brt, brt_list);
   2550 	sc->sc_brtcnt--;
   2551 }
   2552 
   2553 /*
   2554  * bridge_rtnode_destroy:
   2555  *
   2556  *	Destroy a bridge rtnode.
   2557  */
   2558 static void
   2559 bridge_rtnode_destroy(struct bridge_rtnode *brt)
   2560 {
   2561 
   2562 	pool_put(&bridge_rtnode_pool, brt);
   2563 }
   2564 
   2565 #if defined(BRIDGE_IPF)
   2566 extern pfil_head_t *inet_pfil_hook;                 /* XXX */
   2567 extern pfil_head_t *inet6_pfil_hook;                /* XXX */
   2568 
   2569 /*
   2570  * Send bridge packets through IPF if they are one of the types IPF can deal
   2571  * with, or if they are ARP or REVARP.  (IPF will pass ARP and REVARP without
   2572  * question.)
   2573  */
   2574 static int
   2575 bridge_ipf(void *arg, struct mbuf **mp, struct ifnet *ifp, int dir)
   2576 {
   2577 	int snap, error;
   2578 	struct ether_header *eh1, eh2;
   2579 	struct llc llc1;
   2580 	uint16_t ether_type;
   2581 
   2582 	snap = 0;
   2583 	error = -1;	/* Default error if not error == 0 */
   2584 	eh1 = mtod(*mp, struct ether_header *);
   2585 	ether_type = ntohs(eh1->ether_type);
   2586 
   2587 	/*
   2588 	 * Check for SNAP/LLC.
   2589 	 */
   2590         if (ether_type < ETHERMTU) {
   2591                 struct llc *llc2 = (struct llc *)(eh1 + 1);
   2592 
   2593                 if ((*mp)->m_len >= ETHER_HDR_LEN + 8 &&
   2594                     llc2->llc_dsap == LLC_SNAP_LSAP &&
   2595                     llc2->llc_ssap == LLC_SNAP_LSAP &&
   2596                     llc2->llc_control == LLC_UI) {
   2597                 	ether_type = htons(llc2->llc_un.type_snap.ether_type);
   2598 			snap = 1;
   2599                 }
   2600         }
   2601 
   2602 	/*
   2603 	 * If we're trying to filter bridge traffic, don't look at anything
   2604 	 * other than IP and ARP traffic.  If the filter doesn't understand
   2605 	 * IPv6, don't allow IPv6 through the bridge either.  This is lame
   2606 	 * since if we really wanted, say, an AppleTalk filter, we are hosed,
   2607 	 * but of course we don't have an AppleTalk filter to begin with.
   2608 	 * (Note that since IPF doesn't understand ARP it will pass *ALL*
   2609 	 * ARP traffic.)
   2610 	 */
   2611 	switch (ether_type) {
   2612 		case ETHERTYPE_ARP:
   2613 		case ETHERTYPE_REVARP:
   2614 			return 0; /* Automatically pass */
   2615 		case ETHERTYPE_IP:
   2616 # ifdef INET6
   2617 		case ETHERTYPE_IPV6:
   2618 # endif /* INET6 */
   2619 			break;
   2620 		default:
   2621 			goto bad;
   2622 	}
   2623 
   2624 	/* Strip off the Ethernet header and keep a copy. */
   2625 	m_copydata(*mp, 0, ETHER_HDR_LEN, (void *) &eh2);
   2626 	m_adj(*mp, ETHER_HDR_LEN);
   2627 
   2628 	/* Strip off snap header, if present */
   2629 	if (snap) {
   2630 		m_copydata(*mp, 0, sizeof(struct llc), (void *) &llc1);
   2631 		m_adj(*mp, sizeof(struct llc));
   2632 	}
   2633 
   2634 	/*
   2635 	 * Check basic packet sanity and run IPF through pfil.
   2636 	 */
   2637 	KASSERT(!cpu_intr_p());
   2638 	switch (ether_type)
   2639 	{
   2640 	case ETHERTYPE_IP :
   2641 		error = (dir == PFIL_IN) ? bridge_ip_checkbasic(mp) : 0;
   2642 		if (error == 0)
   2643 			error = pfil_run_hooks(inet_pfil_hook, mp, ifp, dir);
   2644 		break;
   2645 # ifdef INET6
   2646 	case ETHERTYPE_IPV6 :
   2647 		error = (dir == PFIL_IN) ? bridge_ip6_checkbasic(mp) : 0;
   2648 		if (error == 0)
   2649 			error = pfil_run_hooks(inet6_pfil_hook, mp, ifp, dir);
   2650 		break;
   2651 # endif
   2652 	default :
   2653 		error = 0;
   2654 		break;
   2655 	}
   2656 
   2657 	if (*mp == NULL)
   2658 		return error;
   2659 	if (error != 0)
   2660 		goto bad;
   2661 
   2662 	error = -1;
   2663 
   2664 	/*
   2665 	 * Finally, put everything back the way it was and return
   2666 	 */
   2667 	if (snap) {
   2668 		M_PREPEND(*mp, sizeof(struct llc), M_DONTWAIT);
   2669 		if (*mp == NULL)
   2670 			return error;
   2671 		bcopy(&llc1, mtod(*mp, void *), sizeof(struct llc));
   2672 	}
   2673 
   2674 	M_PREPEND(*mp, ETHER_HDR_LEN, M_DONTWAIT);
   2675 	if (*mp == NULL)
   2676 		return error;
   2677 	bcopy(&eh2, mtod(*mp, void *), ETHER_HDR_LEN);
   2678 
   2679 	return 0;
   2680 
   2681     bad:
   2682 	m_freem(*mp);
   2683 	*mp = NULL;
   2684 	return error;
   2685 }
   2686 
   2687 /*
   2688  * Perform basic checks on header size since
   2689  * IPF assumes ip_input has already processed
   2690  * it for it.  Cut-and-pasted from ip_input.c.
   2691  * Given how simple the IPv6 version is,
   2692  * does the IPv4 version really need to be
   2693  * this complicated?
   2694  *
   2695  * XXX Should we update ipstat here, or not?
   2696  * XXX Right now we update ipstat but not
   2697  * XXX csum_counter.
   2698  */
   2699 static int
   2700 bridge_ip_checkbasic(struct mbuf **mp)
   2701 {
   2702 	struct mbuf *m = *mp;
   2703 	struct ip *ip;
   2704 	int len, hlen;
   2705 
   2706 	if (*mp == NULL)
   2707 		return -1;
   2708 
   2709 	if (IP_HDR_ALIGNED_P(mtod(m, void *)) == 0) {
   2710 		if ((m = m_copyup(m, sizeof(struct ip),
   2711 			(max_linkhdr + 3) & ~3)) == NULL) {
   2712 			/* XXXJRT new stat, please */
   2713 			ip_statinc(IP_STAT_TOOSMALL);
   2714 			goto bad;
   2715 		}
   2716 	} else if (__predict_false(m->m_len < sizeof (struct ip))) {
   2717 		if ((m = m_pullup(m, sizeof (struct ip))) == NULL) {
   2718 			ip_statinc(IP_STAT_TOOSMALL);
   2719 			goto bad;
   2720 		}
   2721 	}
   2722 	ip = mtod(m, struct ip *);
   2723 	if (ip == NULL) goto bad;
   2724 
   2725 	if (ip->ip_v != IPVERSION) {
   2726 		ip_statinc(IP_STAT_BADVERS);
   2727 		goto bad;
   2728 	}
   2729 	hlen = ip->ip_hl << 2;
   2730 	if (hlen < sizeof(struct ip)) { /* minimum header length */
   2731 		ip_statinc(IP_STAT_BADHLEN);
   2732 		goto bad;
   2733 	}
   2734 	if (hlen > m->m_len) {
   2735 		if ((m = m_pullup(m, hlen)) == 0) {
   2736 			ip_statinc(IP_STAT_BADHLEN);
   2737 			goto bad;
   2738 		}
   2739 		ip = mtod(m, struct ip *);
   2740 		if (ip == NULL) goto bad;
   2741 	}
   2742 
   2743         switch (m->m_pkthdr.csum_flags &
   2744                 ((m->m_pkthdr.rcvif->if_csum_flags_rx & M_CSUM_IPv4) |
   2745                  M_CSUM_IPv4_BAD)) {
   2746         case M_CSUM_IPv4|M_CSUM_IPv4_BAD:
   2747                 /* INET_CSUM_COUNTER_INCR(&ip_hwcsum_bad); */
   2748                 goto bad;
   2749 
   2750         case M_CSUM_IPv4:
   2751                 /* Checksum was okay. */
   2752                 /* INET_CSUM_COUNTER_INCR(&ip_hwcsum_ok); */
   2753                 break;
   2754 
   2755         default:
   2756                 /* Must compute it ourselves. */
   2757                 /* INET_CSUM_COUNTER_INCR(&ip_swcsum); */
   2758                 if (in_cksum(m, hlen) != 0)
   2759                         goto bad;
   2760                 break;
   2761         }
   2762 
   2763         /* Retrieve the packet length. */
   2764         len = ntohs(ip->ip_len);
   2765 
   2766         /*
   2767          * Check for additional length bogosity
   2768          */
   2769         if (len < hlen) {
   2770 		ip_statinc(IP_STAT_BADLEN);
   2771                 goto bad;
   2772         }
   2773 
   2774         /*
   2775          * Check that the amount of data in the buffers
   2776          * is as at least much as the IP header would have us expect.
   2777          * Drop packet if shorter than we expect.
   2778          */
   2779         if (m->m_pkthdr.len < len) {
   2780 		ip_statinc(IP_STAT_TOOSHORT);
   2781                 goto bad;
   2782         }
   2783 
   2784 	/* Checks out, proceed */
   2785 	*mp = m;
   2786 	return 0;
   2787 
   2788     bad:
   2789 	*mp = m;
   2790 	return -1;
   2791 }
   2792 
   2793 # ifdef INET6
   2794 /*
   2795  * Same as above, but for IPv6.
   2796  * Cut-and-pasted from ip6_input.c.
   2797  * XXX Should we update ip6stat, or not?
   2798  */
   2799 static int
   2800 bridge_ip6_checkbasic(struct mbuf **mp)
   2801 {
   2802 	struct mbuf *m = *mp;
   2803 	struct ip6_hdr *ip6;
   2804 
   2805         /*
   2806          * If the IPv6 header is not aligned, slurp it up into a new
   2807          * mbuf with space for link headers, in the event we forward
   2808          * it.  Otherwise, if it is aligned, make sure the entire base
   2809          * IPv6 header is in the first mbuf of the chain.
   2810          */
   2811         if (IP6_HDR_ALIGNED_P(mtod(m, void *)) == 0) {
   2812                 struct ifnet *inifp = m->m_pkthdr.rcvif;
   2813                 if ((m = m_copyup(m, sizeof(struct ip6_hdr),
   2814                                   (max_linkhdr + 3) & ~3)) == NULL) {
   2815                         /* XXXJRT new stat, please */
   2816 			ip6_statinc(IP6_STAT_TOOSMALL);
   2817                         in6_ifstat_inc(inifp, ifs6_in_hdrerr);
   2818                         goto bad;
   2819                 }
   2820         } else if (__predict_false(m->m_len < sizeof(struct ip6_hdr))) {
   2821                 struct ifnet *inifp = m->m_pkthdr.rcvif;
   2822                 if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) {
   2823 			ip6_statinc(IP6_STAT_TOOSMALL);
   2824                         in6_ifstat_inc(inifp, ifs6_in_hdrerr);
   2825                         goto bad;
   2826                 }
   2827         }
   2828 
   2829         ip6 = mtod(m, struct ip6_hdr *);
   2830 
   2831         if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) {
   2832 		ip6_statinc(IP6_STAT_BADVERS);
   2833                 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr);
   2834                 goto bad;
   2835         }
   2836 
   2837 	/* Checks out, proceed */
   2838 	*mp = m;
   2839 	return 0;
   2840 
   2841     bad:
   2842 	*mp = m;
   2843 	return -1;
   2844 }
   2845 # endif /* INET6 */
   2846 #endif /* BRIDGE_IPF */
   2847