sys/net/if_bridge.c

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