sys/net/if_bridge.c

   1 /*      $NetBSD: if_bridge.c,v 1.31 2005/06/01 19:45:34 jdc Exp $       */
   2
   3 /*-
   4  * SPDX-License-Identifier: BSD-4-Clause
   5  *
   6  * Copyright 2001 Wasabi Systems, Inc.
   7  * All rights reserved.
   8  *
   9  * Written by Jason R. Thorpe for Wasabi Systems, Inc.
  10  *
  11  * Redistribution and use in source and binary forms, with or without
  12  * modification, are permitted provided that the following conditions
  13  * are met:
  14  * 1. Redistributions of source code must retain the above copyright
  15  *    notice, this list of conditions and the following disclaimer.
  16  * 2. Redistributions in binary form must reproduce the above copyright
  17  *    notice, this list of conditions and the following disclaimer in the
  18  *    documentation and/or other materials provided with the distribution.
  19  * 3. All advertising materials mentioning features or use of this software
  20  *    must display the following acknowledgement:
  21  *      This product includes software developed for the NetBSD Project by
  22  *      Wasabi Systems, Inc.
  23  * 4. The name of Wasabi Systems, Inc. may not be used to endorse
  24  *    or promote products derived from this software without specific prior
  25  *    written permission.
  26  *
  27  * THIS SOFTWARE IS PROVIDED BY WASABI SYSTEMS, INC. ``AS IS'' AND
  28  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
  29  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
  30  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL WASABI SYSTEMS, INC
  31  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
  32  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
  33  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
  34  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
  35  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
  36  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
  37  * POSSIBILITY OF SUCH DAMAGE.
  38  */
  39
  40 /*
  41  * Copyright (c) 1999, 2000 Jason L. Wright (jason@thought.net)
  42  * All rights reserved.
  43  *
  44  * Redistribution and use in source and binary forms, with or without
  45  * modification, are permitted provided that the following conditions
  46  * are met:
  47  * 1. Redistributions of source code must retain the above copyright
  48  *    notice, this list of conditions and the following disclaimer.
  49  * 2. Redistributions in binary form must reproduce the above copyright
  50  *    notice, this list of conditions and the following disclaimer in the
  51  *    documentation and/or other materials provided with the distribution.
  52  *
  53  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
  54  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
  55  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
  56  * DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT,
  57  * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
  58  * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
  59  * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  60  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
  61  * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
  62  * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
  63  * POSSIBILITY OF SUCH DAMAGE.
  64  *
  65  * OpenBSD: if_bridge.c,v 1.60 2001/06/15 03:38:33 itojun Exp
  66  */
  67
  68 /*
  69  * Network interface bridge support.
  70  *
  71  * TODO:
  72  *
  73  *      - Currently only supports Ethernet-like interfaces (Ethernet,
  74  *        802.11, VLANs on Ethernet, etc.)  Figure out a nice way
  75  *        to bridge other types of interfaces (maybe consider
  76  *        heterogeneous bridges).
  77  */
  78
  79 #include <sys/cdefs.h>
  80 __FBSDID("$FreeBSD$");
  81
  82 #include "opt_inet.h"
  83 #include "opt_inet6.h"
  84
  85 #include <sys/param.h>
  86 #include <sys/eventhandler.h>
  87 #include <sys/mbuf.h>
  88 #include <sys/malloc.h>
  89 #include <sys/protosw.h>
  90 #include <sys/systm.h>
  91 #include <sys/jail.h>
  92 #include <sys/time.h>
  93 #include <sys/socket.h> /* for net/if.h */
  94 #include <sys/sockio.h>
  95 #include <sys/ctype.h>  /* string functions */
  96 #include <sys/kernel.h>
  97 #include <sys/random.h>
  98 #include <sys/syslog.h>
  99 #include <sys/sysctl.h>
 100 #include <vm/uma.h>
 101 #include <sys/module.h>
 102 #include <sys/priv.h>
 103 #include <sys/proc.h>
 104 #include <sys/lock.h>
 105 #include <sys/mutex.h>
 106
 107 #include <net/bpf.h>
 108 #include <net/if.h>
 109 #include <net/if_clone.h>
 110 #include <net/if_dl.h>
 111 #include <net/if_types.h>
 112 #include <net/if_var.h>
 113 #include <net/pfil.h>
 114 #include <net/vnet.h>
 115
 116 #include <netinet/in.h>
 117 #include <netinet/in_systm.h>
 118 #include <netinet/in_var.h>
 119 #include <netinet/ip.h>
 120 #include <netinet/ip_var.h>
 121 #ifdef INET6
 122 #include <netinet/ip6.h>
 123 #include <netinet6/ip6_var.h>
 124 #include <netinet6/in6_ifattach.h>
 125 #endif
 126 #if defined(INET) || defined(INET6)
 127 #include <netinet/ip_carp.h>
 128 #endif
 129 #include <machine/in_cksum.h>
 130 #include <netinet/if_ether.h>
 131 #include <net/bridgestp.h>
 132 #include <net/if_bridgevar.h>
 133 #include <net/if_llc.h>
 134 #include <net/if_vlan_var.h>
 135
 136 #include <net/route.h>
 137
 138 /*
 139  * Size of the route hash table.  Must be a power of two.
 140  */
 141 #ifndef BRIDGE_RTHASH_SIZE
 142 #define BRIDGE_RTHASH_SIZE              1024
 143 #endif
 144
 145 #define BRIDGE_RTHASH_MASK              (BRIDGE_RTHASH_SIZE - 1)
 146
 147 /*
 148  * Default maximum number of addresses to cache.
 149  */
 150 #ifndef BRIDGE_RTABLE_MAX
 151 #define BRIDGE_RTABLE_MAX               2000
 152 #endif
 153
 154 /*
 155  * Timeout (in seconds) for entries learned dynamically.
 156  */
 157 #ifndef BRIDGE_RTABLE_TIMEOUT
 158 #define BRIDGE_RTABLE_TIMEOUT           (20 * 60)       /* same as ARP */
 159 #endif
 160
 161 /*
 162  * Number of seconds between walks of the route list.
 163  */
 164 #ifndef BRIDGE_RTABLE_PRUNE_PERIOD
 165 #define BRIDGE_RTABLE_PRUNE_PERIOD      (5 * 60)
 166 #endif
 167
 168 /*
 169  * List of capabilities to possibly mask on the member interface.
 170  */
 171 #define BRIDGE_IFCAPS_MASK              (IFCAP_TOE|IFCAP_TSO|IFCAP_TXCSUM|\
 172                                          IFCAP_TXCSUM_IPV6)
 173
 174 /*
 175  * List of capabilities to strip
 176  */
 177 #define BRIDGE_IFCAPS_STRIP             IFCAP_LRO
 178
 179 /*
 180  * Bridge interface list entry.
 181  */
 182 struct bridge_iflist {
 183         LIST_ENTRY(bridge_iflist) bif_next;
 184         struct ifnet            *bif_ifp;       /* member if */
 185         struct bstp_port        bif_stp;        /* STP state */
 186         uint32_t                bif_flags;      /* member if flags */
 187         int                     bif_savedcaps;  /* saved capabilities */
 188         uint32_t                bif_addrmax;    /* max # of addresses */
 189         uint32_t                bif_addrcnt;    /* cur. # of addresses */
 190         uint32_t                bif_addrexceeded;/* # of address violations */
 191 };
 192
 193 /*
 194  * Bridge route node.
 195  */
 196 struct bridge_rtnode {
 197         LIST_ENTRY(bridge_rtnode) brt_hash;     /* hash table linkage */
 198         LIST_ENTRY(bridge_rtnode) brt_list;     /* list linkage */
 199         struct bridge_iflist    *brt_dst;       /* destination if */
 200         unsigned long           brt_expire;     /* expiration time */
 201         uint8_t                 brt_flags;      /* address flags */
 202         uint8_t                 brt_addr[ETHER_ADDR_LEN];
 203         uint16_t                brt_vlan;       /* vlan id */
 204 };
 205 #define brt_ifp                 brt_dst->bif_ifp
 206
 207 /*
 208  * Software state for each bridge.
 209  */
 210 struct bridge_softc {
 211         struct ifnet            *sc_ifp;        /* make this an interface */
 212         LIST_ENTRY(bridge_softc) sc_list;
 213         struct mtx              sc_mtx;
 214         struct cv               sc_cv;
 215         uint32_t                sc_brtmax;      /* max # of addresses */
 216         uint32_t                sc_brtcnt;      /* cur. # of addresses */
 217         uint32_t                sc_brttimeout;  /* rt timeout in seconds */
 218         struct callout          sc_brcallout;   /* bridge callout */
 219         uint32_t                sc_iflist_ref;  /* refcount for sc_iflist */
 220         uint32_t                sc_iflist_xcnt; /* refcount for sc_iflist */
 221         LIST_HEAD(, bridge_iflist) sc_iflist;   /* member interface list */
 222         LIST_HEAD(, bridge_rtnode) *sc_rthash;  /* our forwarding table */
 223         LIST_HEAD(, bridge_rtnode) sc_rtlist;   /* list version of above */
 224         uint32_t                sc_rthash_key;  /* key for hash */
 225         LIST_HEAD(, bridge_iflist) sc_spanlist; /* span ports list */
 226         struct bstp_state       sc_stp;         /* STP state */
 227         uint32_t                sc_brtexceeded; /* # of cache drops */
 228         struct ifnet            *sc_ifaddr;     /* member mac copied from */
 229         struct ether_addr       sc_defaddr;     /* Default MAC address */
 230 };
 231
 232 VNET_DEFINE_STATIC(struct mtx, bridge_list_mtx);
 233 #define V_bridge_list_mtx       VNET(bridge_list_mtx)
 234 static eventhandler_tag bridge_detach_cookie;
 235
 236 int     bridge_rtable_prune_period = BRIDGE_RTABLE_PRUNE_PERIOD;
 237
 238 uma_zone_t bridge_rtnode_zone;
 239
 240 static int      bridge_clone_create(struct if_clone *, int, caddr_t);
 241 static void     bridge_clone_destroy(struct ifnet *);
 242
 243 static int      bridge_ioctl(struct ifnet *, u_long, caddr_t);
 244 static void     bridge_mutecaps(struct bridge_softc *);
 245 static void     bridge_set_ifcap(struct bridge_softc *, struct bridge_iflist *,
 246                     int);
 247 static void     bridge_ifdetach(void *arg __unused, struct ifnet *);
 248 static void     bridge_init(void *);
 249 static void     bridge_dummynet(struct mbuf *, struct ifnet *);
 250 static void     bridge_stop(struct ifnet *, int);
 251 static int      bridge_transmit(struct ifnet *, struct mbuf *);
 252 static void     bridge_qflush(struct ifnet *);
 253 static struct mbuf *bridge_input(struct ifnet *, struct mbuf *);
 254 static int      bridge_output(struct ifnet *, struct mbuf *, struct sockaddr *,
 255                     struct rtentry *);
 256 static int      bridge_enqueue(struct bridge_softc *, struct ifnet *,
 257                     struct mbuf *);
 258 static void     bridge_rtdelete(struct bridge_softc *, struct ifnet *ifp, int);
 259
 260 static void     bridge_forward(struct bridge_softc *, struct bridge_iflist *,
 261                     struct mbuf *m);
 262
 263 static void     bridge_timer(void *);
 264
 265 static void     bridge_broadcast(struct bridge_softc *, struct ifnet *,
 266                     struct mbuf *, int);
 267 static void     bridge_span(struct bridge_softc *, struct mbuf *);
 268
 269 static int      bridge_rtupdate(struct bridge_softc *, const uint8_t *,
 270                     uint16_t, struct bridge_iflist *, int, uint8_t);
 271 static struct ifnet *bridge_rtlookup(struct bridge_softc *, const uint8_t *,
 272                     uint16_t);
 273 static void     bridge_rttrim(struct bridge_softc *);
 274 static void     bridge_rtage(struct bridge_softc *);
 275 static void     bridge_rtflush(struct bridge_softc *, int);
 276 static int      bridge_rtdaddr(struct bridge_softc *, const uint8_t *,
 277                     uint16_t);
 278
 279 static void     bridge_rtable_init(struct bridge_softc *);
 280 static void     bridge_rtable_fini(struct bridge_softc *);
 281
 282 static int      bridge_rtnode_addr_cmp(const uint8_t *, const uint8_t *);
 283 static struct bridge_rtnode *bridge_rtnode_lookup(struct bridge_softc *,
 284                     const uint8_t *, uint16_t);
 285 static int      bridge_rtnode_insert(struct bridge_softc *,
 286                     struct bridge_rtnode *);
 287 static void     bridge_rtnode_destroy(struct bridge_softc *,
 288                     struct bridge_rtnode *);
 289 static void     bridge_rtable_expire(struct ifnet *, int);
 290 static void     bridge_state_change(struct ifnet *, int);
 291
 292 static struct bridge_iflist *bridge_lookup_member(struct bridge_softc *,
 293                     const char *name);
 294 static struct bridge_iflist *bridge_lookup_member_if(struct bridge_softc *,
 295                     struct ifnet *ifp);
 296 static void     bridge_delete_member(struct bridge_softc *,
 297                     struct bridge_iflist *, int);
 298 static void     bridge_delete_span(struct bridge_softc *,
 299                     struct bridge_iflist *);
 300
 301 static int      bridge_ioctl_add(struct bridge_softc *, void *);
 302 static int      bridge_ioctl_del(struct bridge_softc *, void *);
 303 static int      bridge_ioctl_gifflags(struct bridge_softc *, void *);
 304 static int      bridge_ioctl_sifflags(struct bridge_softc *, void *);
 305 static int      bridge_ioctl_scache(struct bridge_softc *, void *);
 306 static int      bridge_ioctl_gcache(struct bridge_softc *, void *);
 307 static int      bridge_ioctl_gifs(struct bridge_softc *, void *);
 308 static int      bridge_ioctl_rts(struct bridge_softc *, void *);
 309 static int      bridge_ioctl_saddr(struct bridge_softc *, void *);
 310 static int      bridge_ioctl_sto(struct bridge_softc *, void *);
 311 static int      bridge_ioctl_gto(struct bridge_softc *, void *);
 312 static int      bridge_ioctl_daddr(struct bridge_softc *, void *);
 313 static int      bridge_ioctl_flush(struct bridge_softc *, void *);
 314 static int      bridge_ioctl_gpri(struct bridge_softc *, void *);
 315 static int      bridge_ioctl_spri(struct bridge_softc *, void *);
 316 static int      bridge_ioctl_ght(struct bridge_softc *, void *);
 317 static int      bridge_ioctl_sht(struct bridge_softc *, void *);
 318 static int      bridge_ioctl_gfd(struct bridge_softc *, void *);
 319 static int      bridge_ioctl_sfd(struct bridge_softc *, void *);
 320 static int      bridge_ioctl_gma(struct bridge_softc *, void *);
 321 static int      bridge_ioctl_sma(struct bridge_softc *, void *);
 322 static int      bridge_ioctl_sifprio(struct bridge_softc *, void *);
 323 static int      bridge_ioctl_sifcost(struct bridge_softc *, void *);
 324 static int      bridge_ioctl_sifmaxaddr(struct bridge_softc *, void *);
 325 static int      bridge_ioctl_addspan(struct bridge_softc *, void *);
 326 static int      bridge_ioctl_delspan(struct bridge_softc *, void *);
 327 static int      bridge_ioctl_gbparam(struct bridge_softc *, void *);
 328 static int      bridge_ioctl_grte(struct bridge_softc *, void *);
 329 static int      bridge_ioctl_gifsstp(struct bridge_softc *, void *);
 330 static int      bridge_ioctl_sproto(struct bridge_softc *, void *);
 331 static int      bridge_ioctl_stxhc(struct bridge_softc *, void *);
 332 static int      bridge_pfil(struct mbuf **, struct ifnet *, struct ifnet *,
 333                     int);
 334 static int      bridge_ip_checkbasic(struct mbuf **mp);
 335 #ifdef INET6
 336 static int      bridge_ip6_checkbasic(struct mbuf **mp);
 337 #endif /* INET6 */
 338 static int      bridge_fragment(struct ifnet *, struct mbuf **mp,
 339                     struct ether_header *, int, struct llc *);
 340 static void     bridge_linkstate(struct ifnet *ifp);
 341 static void     bridge_linkcheck(struct bridge_softc *sc);
 342
 343
 344 /* The default bridge vlan is 1 (IEEE 802.1Q-2003 Table 9-2) */
 345 #define VLANTAGOF(_m)   \
 346     (_m->m_flags & M_VLANTAG) ? EVL_VLANOFTAG(_m->m_pkthdr.ether_vtag) : 1
 347
 348 static struct bstp_cb_ops bridge_ops = {
 349         .bcb_state = bridge_state_change,
 350         .bcb_rtage = bridge_rtable_expire
 351 };
 352
 353 SYSCTL_DECL(_net_link);
 354 static SYSCTL_NODE(_net_link, IFT_BRIDGE, bridge, CTLFLAG_RW, 0, "Bridge");
 355
 356 /* only pass IP[46] packets when pfil is enabled */
 357 VNET_DEFINE_STATIC(int, pfil_onlyip) = 1;
 358 #define V_pfil_onlyip   VNET(pfil_onlyip)
 359 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_onlyip,
 360     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_onlyip), 0,
 361     "Only pass IP packets when pfil is enabled");
 362
 363 /* run pfil hooks on the bridge interface */
 364 VNET_DEFINE_STATIC(int, pfil_bridge) = 1;
 365 #define V_pfil_bridge   VNET(pfil_bridge)
 366 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_bridge,
 367     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_bridge), 0,
 368     "Packet filter on the bridge interface");
 369
 370 /* layer2 filter with ipfw */
 371 VNET_DEFINE_STATIC(int, pfil_ipfw);
 372 #define V_pfil_ipfw     VNET(pfil_ipfw)
 373
 374 /* layer2 ARP filter with ipfw */
 375 VNET_DEFINE_STATIC(int, pfil_ipfw_arp);
 376 #define V_pfil_ipfw_arp VNET(pfil_ipfw_arp)
 377 SYSCTL_INT(_net_link_bridge, OID_AUTO, ipfw_arp,
 378     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_ipfw_arp), 0,
 379     "Filter ARP packets through IPFW layer2");
 380
 381 /* run pfil hooks on the member interface */
 382 VNET_DEFINE_STATIC(int, pfil_member) = 1;
 383 #define V_pfil_member   VNET(pfil_member)
 384 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_member,
 385     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_member), 0,
 386     "Packet filter on the member interface");
 387
 388 /* run pfil hooks on the physical interface for locally destined packets */
 389 VNET_DEFINE_STATIC(int, pfil_local_phys);
 390 #define V_pfil_local_phys       VNET(pfil_local_phys)
 391 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_local_phys,
 392     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_local_phys), 0,
 393     "Packet filter on the physical interface for locally destined packets");
 394
 395 /* log STP state changes */
 396 VNET_DEFINE_STATIC(int, log_stp);
 397 #define V_log_stp       VNET(log_stp)
 398 SYSCTL_INT(_net_link_bridge, OID_AUTO, log_stp,
 399     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(log_stp), 0,
 400     "Log STP state changes");
 401
 402 /* share MAC with first bridge member */
 403 VNET_DEFINE_STATIC(int, bridge_inherit_mac);
 404 #define V_bridge_inherit_mac    VNET(bridge_inherit_mac)
 405 SYSCTL_INT(_net_link_bridge, OID_AUTO, inherit_mac,
 406     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(bridge_inherit_mac), 0,
 407     "Inherit MAC address from the first bridge member");
 408
 409 VNET_DEFINE_STATIC(int, allow_llz_overlap) = 0;
 410 #define V_allow_llz_overlap     VNET(allow_llz_overlap)
 411 SYSCTL_INT(_net_link_bridge, OID_AUTO, allow_llz_overlap,
 412     CTLFLAG_RW | CTLFLAG_VNET, &VNET_NAME(allow_llz_overlap), 0,
 413     "Allow overlap of link-local scope "
 414     "zones of a bridge interface and the member interfaces");
 415
 416 struct bridge_control {
 417         int     (*bc_func)(struct bridge_softc *, void *);
 418         int     bc_argsize;
 419         int     bc_flags;
 420 };
 421
 422 #define BC_F_COPYIN             0x01    /* copy arguments in */
 423 #define BC_F_COPYOUT            0x02    /* copy arguments out */
 424 #define BC_F_SUSER              0x04    /* do super-user check */
 425
 426 const struct bridge_control bridge_control_table[] = {
 427         { bridge_ioctl_add,             sizeof(struct ifbreq),
 428           BC_F_COPYIN|BC_F_SUSER },
 429         { bridge_ioctl_del,             sizeof(struct ifbreq),
 430           BC_F_COPYIN|BC_F_SUSER },
 431
 432         { bridge_ioctl_gifflags,        sizeof(struct ifbreq),
 433           BC_F_COPYIN|BC_F_COPYOUT },
 434         { bridge_ioctl_sifflags,        sizeof(struct ifbreq),
 435           BC_F_COPYIN|BC_F_SUSER },
 436
 437         { bridge_ioctl_scache,          sizeof(struct ifbrparam),
 438           BC_F_COPYIN|BC_F_SUSER },
 439         { bridge_ioctl_gcache,          sizeof(struct ifbrparam),
 440           BC_F_COPYOUT },
 441
 442         { bridge_ioctl_gifs,            sizeof(struct ifbifconf),
 443           BC_F_COPYIN|BC_F_COPYOUT },
 444         { bridge_ioctl_rts,             sizeof(struct ifbaconf),
 445           BC_F_COPYIN|BC_F_COPYOUT },
 446
 447         { bridge_ioctl_saddr,           sizeof(struct ifbareq),
 448           BC_F_COPYIN|BC_F_SUSER },
 449
 450         { bridge_ioctl_sto,             sizeof(struct ifbrparam),
 451           BC_F_COPYIN|BC_F_SUSER },
 452         { bridge_ioctl_gto,             sizeof(struct ifbrparam),
 453           BC_F_COPYOUT },
 454
 455         { bridge_ioctl_daddr,           sizeof(struct ifbareq),
 456           BC_F_COPYIN|BC_F_SUSER },
 457
 458         { bridge_ioctl_flush,           sizeof(struct ifbreq),
 459           BC_F_COPYIN|BC_F_SUSER },
 460
 461         { bridge_ioctl_gpri,            sizeof(struct ifbrparam),
 462           BC_F_COPYOUT },
 463         { bridge_ioctl_spri,            sizeof(struct ifbrparam),
 464           BC_F_COPYIN|BC_F_SUSER },
 465
 466         { bridge_ioctl_ght,             sizeof(struct ifbrparam),
 467           BC_F_COPYOUT },
 468         { bridge_ioctl_sht,             sizeof(struct ifbrparam),
 469           BC_F_COPYIN|BC_F_SUSER },
 470
 471         { bridge_ioctl_gfd,             sizeof(struct ifbrparam),
 472           BC_F_COPYOUT },
 473         { bridge_ioctl_sfd,             sizeof(struct ifbrparam),
 474           BC_F_COPYIN|BC_F_SUSER },
 475
 476         { bridge_ioctl_gma,             sizeof(struct ifbrparam),
 477           BC_F_COPYOUT },
 478         { bridge_ioctl_sma,             sizeof(struct ifbrparam),
 479           BC_F_COPYIN|BC_F_SUSER },
 480
 481         { bridge_ioctl_sifprio,         sizeof(struct ifbreq),
 482           BC_F_COPYIN|BC_F_SUSER },
 483
 484         { bridge_ioctl_sifcost,         sizeof(struct ifbreq),
 485           BC_F_COPYIN|BC_F_SUSER },
 486
 487         { bridge_ioctl_addspan,         sizeof(struct ifbreq),
 488           BC_F_COPYIN|BC_F_SUSER },
 489         { bridge_ioctl_delspan,         sizeof(struct ifbreq),
 490           BC_F_COPYIN|BC_F_SUSER },
 491
 492         { bridge_ioctl_gbparam,         sizeof(struct ifbropreq),
 493           BC_F_COPYOUT },
 494
 495         { bridge_ioctl_grte,            sizeof(struct ifbrparam),
 496           BC_F_COPYOUT },
 497
 498         { bridge_ioctl_gifsstp,         sizeof(struct ifbpstpconf),
 499           BC_F_COPYIN|BC_F_COPYOUT },
 500
 501         { bridge_ioctl_sproto,          sizeof(struct ifbrparam),
 502           BC_F_COPYIN|BC_F_SUSER },
 503
 504         { bridge_ioctl_stxhc,           sizeof(struct ifbrparam),
 505           BC_F_COPYIN|BC_F_SUSER },
 506
 507         { bridge_ioctl_sifmaxaddr,      sizeof(struct ifbreq),
 508           BC_F_COPYIN|BC_F_SUSER },
 509
 510 };
 511 const int bridge_control_table_size = nitems(bridge_control_table);
 512
 513 VNET_DEFINE_STATIC(LIST_HEAD(, bridge_softc), bridge_list);
 514 #define V_bridge_list   VNET(bridge_list)
 515 #define BRIDGE_LIST_LOCK_INIT(x)        mtx_init(&V_bridge_list_mtx,    \
 516                                             "if_bridge list", NULL, MTX_DEF)
 517 #define BRIDGE_LIST_LOCK_DESTROY(x)     mtx_destroy(&V_bridge_list_mtx)
 518 #define BRIDGE_LIST_LOCK(x)             mtx_lock(&V_bridge_list_mtx)
 519 #define BRIDGE_LIST_UNLOCK(x)           mtx_unlock(&V_bridge_list_mtx)
 520
 521 VNET_DEFINE_STATIC(struct if_clone *, bridge_cloner);
 522 #define V_bridge_cloner VNET(bridge_cloner)
 523
 524 static const char bridge_name[] = "bridge";
 525
 526 static void
 527 vnet_bridge_init(const void *unused __unused)
 528 {
 529
 530         BRIDGE_LIST_LOCK_INIT();
 531         LIST_INIT(&V_bridge_list);
 532         V_bridge_cloner = if_clone_simple(bridge_name,
 533             bridge_clone_create, bridge_clone_destroy, 0);
 534 }
 535 VNET_SYSINIT(vnet_bridge_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
 536     vnet_bridge_init, NULL);
 537
 538 static void
 539 vnet_bridge_uninit(const void *unused __unused)
 540 {
 541
 542         if_clone_detach(V_bridge_cloner);
 543         V_bridge_cloner = NULL;
 544         BRIDGE_LIST_LOCK_DESTROY();
 545 }
 546 VNET_SYSUNINIT(vnet_bridge_uninit, SI_SUB_PSEUDO, SI_ORDER_ANY,
 547     vnet_bridge_uninit, NULL);
 548
 549 static int
 550 bridge_modevent(module_t mod, int type, void *data)
 551 {
 552
 553         switch (type) {
 554         case MOD_LOAD:
 555                 bridge_rtnode_zone = uma_zcreate("bridge_rtnode",
 556                     sizeof(struct bridge_rtnode), NULL, NULL, NULL, NULL,
 557                     UMA_ALIGN_PTR, 0);
 558                 bridge_dn_p = bridge_dummynet;
 559                 bridge_detach_cookie = EVENTHANDLER_REGISTER(
 560                     ifnet_departure_event, bridge_ifdetach, NULL,
 561                     EVENTHANDLER_PRI_ANY);
 562                 break;
 563         case MOD_UNLOAD:
 564                 EVENTHANDLER_DEREGISTER(ifnet_departure_event,
 565                     bridge_detach_cookie);
 566                 uma_zdestroy(bridge_rtnode_zone);
 567                 bridge_dn_p = NULL;
 568                 break;
 569         default:
 570                 return (EOPNOTSUPP);
 571         }
 572         return (0);
 573 }
 574
 575 static moduledata_t bridge_mod = {
 576         "if_bridge",
 577         bridge_modevent,
 578         0
 579 };
 580
 581 DECLARE_MODULE(if_bridge, bridge_mod, SI_SUB_PSEUDO, SI_ORDER_ANY);
 582 MODULE_VERSION(if_bridge, 1);
 583 MODULE_DEPEND(if_bridge, bridgestp, 1, 1, 1);
 584
 585 /*
 586  * handler for net.link.bridge.ipfw
 587  */
 588 static int
 589 sysctl_pfil_ipfw(SYSCTL_HANDLER_ARGS)
 590 {
 591         int enable = V_pfil_ipfw;
 592         int error;
 593
 594         error = sysctl_handle_int(oidp, &enable, 0, req);
 595         enable &= 1;
 596
 597         if (enable != V_pfil_ipfw) {
 598                 V_pfil_ipfw = enable;
 599
 600                 /*
 601                  * Disable pfil so that ipfw doesnt run twice, if the user
 602                  * really wants both then they can re-enable pfil_bridge and/or
 603                  * pfil_member. Also allow non-ip packets as ipfw can filter by
 604                  * layer2 type.
 605                  */
 606                 if (V_pfil_ipfw) {
 607                         V_pfil_onlyip = 0;
 608                         V_pfil_bridge = 0;
 609                         V_pfil_member = 0;
 610                 }
 611         }
 612
 613         return (error);
 614 }
 615 SYSCTL_PROC(_net_link_bridge, OID_AUTO, ipfw,
 616     CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_VNET,
 617     &VNET_NAME(pfil_ipfw), 0, &sysctl_pfil_ipfw, "I",
 618     "Layer2 filter with IPFW");
 619
 620 /*
 621  * bridge_clone_create:
 622  *
 623  *      Create a new bridge instance.
 624  */
 625 static int
 626 bridge_clone_create(struct if_clone *ifc, int unit, caddr_t params)
 627 {
 628         struct bridge_softc *sc, *sc2;
 629         struct ifnet *bifp, *ifp;
 630         int fb, retry;
 631         unsigned long hostid;
 632
 633         sc = malloc(sizeof(*sc), M_DEVBUF, M_WAITOK|M_ZERO);
 634         ifp = sc->sc_ifp = if_alloc(IFT_ETHER);
 635         if (ifp == NULL) {
 636                 free(sc, M_DEVBUF);
 637                 return (ENOSPC);
 638         }
 639
 640         BRIDGE_LOCK_INIT(sc);
 641         sc->sc_brtmax = BRIDGE_RTABLE_MAX;
 642         sc->sc_brttimeout = BRIDGE_RTABLE_TIMEOUT;
 643
 644         /* Initialize our routing table. */
 645         bridge_rtable_init(sc);
 646
 647         callout_init_mtx(&sc->sc_brcallout, &sc->sc_mtx, 0);
 648
 649         LIST_INIT(&sc->sc_iflist);
 650         LIST_INIT(&sc->sc_spanlist);
 651
 652         ifp->if_softc = sc;
 653         if_initname(ifp, bridge_name, unit);
 654         ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
 655         ifp->if_ioctl = bridge_ioctl;
 656         ifp->if_transmit = bridge_transmit;
 657         ifp->if_qflush = bridge_qflush;
 658         ifp->if_init = bridge_init;
 659         ifp->if_type = IFT_BRIDGE;
 660
 661         /*
 662          * Generate an ethernet address with a locally administered address.
 663          *
 664          * Since we are using random ethernet addresses for the bridge, it is
 665          * possible that we might have address collisions, so make sure that
 666          * this hardware address isn't already in use on another bridge.
 667          * The first try uses the hostid and falls back to arc4rand().
 668          */
 669         fb = 0;
 670         getcredhostid(curthread->td_ucred, &hostid);
 671         do {
 672                 if (fb || hostid == 0) {
 673                         ether_fakeaddr(&sc->sc_defaddr);
 674                 } else {
 675                         sc->sc_defaddr.octet[0] = 0x2;
 676                         sc->sc_defaddr.octet[1] = (hostid >> 24) & 0xff;
 677                         sc->sc_defaddr.octet[2] = (hostid >> 16) & 0xff;
 678                         sc->sc_defaddr.octet[3] = (hostid >> 8 ) & 0xff;
 679                         sc->sc_defaddr.octet[4] =  hostid        & 0xff;
 680                         sc->sc_defaddr.octet[5] = ifp->if_dunit & 0xff;
 681                 }
 682
 683                 fb = 1;
 684                 retry = 0;
 685                 BRIDGE_LIST_LOCK();
 686                 LIST_FOREACH(sc2, &V_bridge_list, sc_list) {
 687                         bifp = sc2->sc_ifp;
 688                         if (memcmp(sc->sc_defaddr.octet,
 689                             IF_LLADDR(bifp), ETHER_ADDR_LEN) == 0) {
 690                                 retry = 1;
 691                                 break;
 692                         }
 693                 }
 694                 BRIDGE_LIST_UNLOCK();
 695         } while (retry == 1);
 696
 697         bstp_attach(&sc->sc_stp, &bridge_ops);
 698         ether_ifattach(ifp, sc->sc_defaddr.octet);
 699         /* Now undo some of the damage... */
 700         ifp->if_baudrate = 0;
 701         ifp->if_type = IFT_BRIDGE;
 702
 703         BRIDGE_LIST_LOCK();
 704         LIST_INSERT_HEAD(&V_bridge_list, sc, sc_list);
 705         BRIDGE_LIST_UNLOCK();
 706
 707         return (0);
 708 }
 709
 710 /*
 711  * bridge_clone_destroy:
 712  *
 713  *      Destroy a bridge instance.
 714  */
 715 static void
 716 bridge_clone_destroy(struct ifnet *ifp)
 717 {
 718         struct bridge_softc *sc = ifp->if_softc;
 719         struct bridge_iflist *bif;
 720
 721         BRIDGE_LOCK(sc);
 722
 723         bridge_stop(ifp, 1);
 724         ifp->if_flags &= ~IFF_UP;
 725
 726         while ((bif = LIST_FIRST(&sc->sc_iflist)) != NULL)
 727                 bridge_delete_member(sc, bif, 0);
 728
 729         while ((bif = LIST_FIRST(&sc->sc_spanlist)) != NULL) {
 730                 bridge_delete_span(sc, bif);
 731         }
 732
 733         BRIDGE_UNLOCK(sc);
 734
 735         callout_drain(&sc->sc_brcallout);
 736
 737         BRIDGE_LIST_LOCK();
 738         LIST_REMOVE(sc, sc_list);
 739         BRIDGE_LIST_UNLOCK();
 740
 741         bstp_detach(&sc->sc_stp);
 742         ether_ifdetach(ifp);
 743         if_free(ifp);
 744
 745         /* Tear down the routing table. */
 746         bridge_rtable_fini(sc);
 747
 748         BRIDGE_LOCK_DESTROY(sc);
 749         free(sc, M_DEVBUF);
 750 }
 751
 752 /*
 753  * bridge_ioctl:
 754  *
 755  *      Handle a control request from the operator.
 756  */
 757 static int
 758 bridge_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
 759 {
 760         struct bridge_softc *sc = ifp->if_softc;
 761         struct ifreq *ifr = (struct ifreq *)data;
 762         struct bridge_iflist *bif;
 763         struct thread *td = curthread;
 764         union {
 765                 struct ifbreq ifbreq;
 766                 struct ifbifconf ifbifconf;
 767                 struct ifbareq ifbareq;
 768                 struct ifbaconf ifbaconf;
 769                 struct ifbrparam ifbrparam;
 770                 struct ifbropreq ifbropreq;
 771         } args;
 772         struct ifdrv *ifd = (struct ifdrv *) data;
 773         const struct bridge_control *bc;
 774         int error = 0;
 775
 776         switch (cmd) {
 777
 778         case SIOCADDMULTI:
 779         case SIOCDELMULTI:
 780                 break;
 781
 782         case SIOCGDRVSPEC:
 783         case SIOCSDRVSPEC:
 784                 if (ifd->ifd_cmd >= bridge_control_table_size) {
 785                         error = EINVAL;
 786                         break;
 787                 }
 788                 bc = &bridge_control_table[ifd->ifd_cmd];
 789
 790                 if (cmd == SIOCGDRVSPEC &&
 791                     (bc->bc_flags & BC_F_COPYOUT) == 0) {
 792                         error = EINVAL;
 793                         break;
 794                 }
 795                 else if (cmd == SIOCSDRVSPEC &&
 796                     (bc->bc_flags & BC_F_COPYOUT) != 0) {
 797                         error = EINVAL;
 798                         break;
 799                 }
 800
 801                 if (bc->bc_flags & BC_F_SUSER) {
 802                         error = priv_check(td, PRIV_NET_BRIDGE);
 803                         if (error)
 804                                 break;
 805                 }
 806
 807                 if (ifd->ifd_len != bc->bc_argsize ||
 808                     ifd->ifd_len > sizeof(args)) {
 809                         error = EINVAL;
 810                         break;
 811                 }
 812
 813                 bzero(&args, sizeof(args));
 814                 if (bc->bc_flags & BC_F_COPYIN) {
 815                         error = copyin(ifd->ifd_data, &args, ifd->ifd_len);
 816                         if (error)
 817                                 break;
 818                 }
 819
 820                 BRIDGE_LOCK(sc);
 821                 error = (*bc->bc_func)(sc, &args);
 822                 BRIDGE_UNLOCK(sc);
 823                 if (error)
 824                         break;
 825
 826                 if (bc->bc_flags & BC_F_COPYOUT)
 827                         error = copyout(&args, ifd->ifd_data, ifd->ifd_len);
 828
 829                 break;
 830
 831         case SIOCSIFFLAGS:
 832                 if (!(ifp->if_flags & IFF_UP) &&
 833                     (ifp->if_drv_flags & IFF_DRV_RUNNING)) {
 834                         /*
 835                          * If interface is marked down and it is running,
 836                          * then stop and disable it.
 837                          */
 838                         BRIDGE_LOCK(sc);
 839                         bridge_stop(ifp, 1);
 840                         BRIDGE_UNLOCK(sc);
 841                 } else if ((ifp->if_flags & IFF_UP) &&
 842                     !(ifp->if_drv_flags & IFF_DRV_RUNNING)) {
 843                         /*
 844                          * If interface is marked up and it is stopped, then
 845                          * start it.
 846                          */
 847                         (*ifp->if_init)(sc);
 848                 }
 849                 break;
 850
 851         case SIOCSIFMTU:
 852                 if (ifr->ifr_mtu < 576) {
 853                         error = EINVAL;
 854                         break;
 855                 }
 856                 if (LIST_EMPTY(&sc->sc_iflist)) {
 857                         sc->sc_ifp->if_mtu = ifr->ifr_mtu;
 858                         break;
 859                 }
 860                 BRIDGE_LOCK(sc);
 861                 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
 862                         if (bif->bif_ifp->if_mtu != ifr->ifr_mtu) {
 863                                 log(LOG_NOTICE, "%s: invalid MTU: %u(%s)"
 864                                     " != %d\n", sc->sc_ifp->if_xname,
 865                                     bif->bif_ifp->if_mtu,
 866                                     bif->bif_ifp->if_xname, ifr->ifr_mtu);
 867                                 error = EINVAL;
 868                                 break;
 869                         }
 870                 }
 871                 if (!error)
 872                         sc->sc_ifp->if_mtu = ifr->ifr_mtu;
 873                 BRIDGE_UNLOCK(sc);
 874                 break;
 875         default:
 876                 /*
 877                  * drop the lock as ether_ioctl() will call bridge_start() and
 878                  * cause the lock to be recursed.
 879                  */
 880                 error = ether_ioctl(ifp, cmd, data);
 881                 break;
 882         }
 883
 884         return (error);
 885 }
 886
 887 /*
 888  * bridge_mutecaps:
 889  *
 890  *      Clear or restore unwanted capabilities on the member interface
 891  */
 892 static void
 893 bridge_mutecaps(struct bridge_softc *sc)
 894 {
 895         struct bridge_iflist *bif;
 896         int enabled, mask;
 897
 898         /* Initial bitmask of capabilities to test */
 899         mask = BRIDGE_IFCAPS_MASK;
 900
 901         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
 902                 /* Every member must support it or its disabled */
 903                 mask &= bif->bif_savedcaps;
 904         }
 905
 906         BRIDGE_XLOCK(sc);
 907         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
 908                 enabled = bif->bif_ifp->if_capenable;
 909                 enabled &= ~BRIDGE_IFCAPS_STRIP;
 910                 /* strip off mask bits and enable them again if allowed */
 911                 enabled &= ~BRIDGE_IFCAPS_MASK;
 912                 enabled |= mask;
 913                 BRIDGE_UNLOCK(sc);
 914                 bridge_set_ifcap(sc, bif, enabled);
 915                 BRIDGE_LOCK(sc);
 916         }
 917         BRIDGE_XDROP(sc);
 918
 919 }
 920
 921 static void
 922 bridge_set_ifcap(struct bridge_softc *sc, struct bridge_iflist *bif, int set)
 923 {
 924         struct ifnet *ifp = bif->bif_ifp;
 925         struct ifreq ifr;
 926         int error, mask, stuck;
 927
 928         BRIDGE_UNLOCK_ASSERT(sc);
 929
 930         bzero(&ifr, sizeof(ifr));
 931         ifr.ifr_reqcap = set;
 932
 933         if (ifp->if_capenable != set) {
 934                 error = (*ifp->if_ioctl)(ifp, SIOCSIFCAP, (caddr_t)&ifr);
 935                 if (error)
 936                         if_printf(sc->sc_ifp,
 937                             "error setting capabilities on %s: %d\n",
 938                             ifp->if_xname, error);
 939                 mask = BRIDGE_IFCAPS_MASK | BRIDGE_IFCAPS_STRIP;
 940                 stuck = ifp->if_capenable & mask & ~set;
 941                 if (stuck != 0)
 942                         if_printf(sc->sc_ifp,
 943                             "can't disable some capabilities on %s: 0x%x\n",
 944                             ifp->if_xname, stuck);
 945         }
 946 }
 947
 948 /*
 949  * bridge_lookup_member:
 950  *
 951  *      Lookup a bridge member interface.
 952  */
 953 static struct bridge_iflist *
 954 bridge_lookup_member(struct bridge_softc *sc, const char *name)
 955 {
 956         struct bridge_iflist *bif;
 957         struct ifnet *ifp;
 958
 959         BRIDGE_LOCK_ASSERT(sc);
 960
 961         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
 962                 ifp = bif->bif_ifp;
 963                 if (strcmp(ifp->if_xname, name) == 0)
 964                         return (bif);
 965         }
 966
 967         return (NULL);
 968 }
 969
 970 /*
 971  * bridge_lookup_member_if:
 972  *
 973  *      Lookup a bridge member interface by ifnet*.
 974  */
 975 static struct bridge_iflist *
 976 bridge_lookup_member_if(struct bridge_softc *sc, struct ifnet *member_ifp)
 977 {
 978         struct bridge_iflist *bif;
 979
 980         BRIDGE_LOCK_ASSERT(sc);
 981
 982         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
 983                 if (bif->bif_ifp == member_ifp)
 984                         return (bif);
 985         }
 986
 987         return (NULL);
 988 }
 989
 990 /*
 991  * bridge_delete_member:
 992  *
 993  *      Delete the specified member interface.
 994  */
 995 static void
 996 bridge_delete_member(struct bridge_softc *sc, struct bridge_iflist *bif,
 997     int gone)
 998 {
 999         struct ifnet *ifs = bif->bif_ifp;
1000         struct ifnet *fif = NULL;
1001
1002         BRIDGE_LOCK_ASSERT(sc);
1003
1004         if (bif->bif_flags & IFBIF_STP)
1005                 bstp_disable(&bif->bif_stp);
1006
1007         ifs->if_bridge = NULL;
1008         BRIDGE_XLOCK(sc);
1009         LIST_REMOVE(bif, bif_next);
1010         BRIDGE_XDROP(sc);
1011
1012         /*
1013          * If removing the interface that gave the bridge its mac address, set
1014          * the mac address of the bridge to the address of the next member, or
1015          * to its default address if no members are left.
1016          */
1017         if (V_bridge_inherit_mac && sc->sc_ifaddr == ifs) {
1018                 if (LIST_EMPTY(&sc->sc_iflist)) {
1019                         bcopy(&sc->sc_defaddr,
1020                             IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN);
1021                         sc->sc_ifaddr = NULL;
1022                 } else {
1023                         fif = LIST_FIRST(&sc->sc_iflist)->bif_ifp;
1024                         bcopy(IF_LLADDR(fif),
1025                             IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN);
1026                         sc->sc_ifaddr = fif;
1027                 }
1028                 EVENTHANDLER_INVOKE(iflladdr_event, sc->sc_ifp);
1029         }
1030
1031         bridge_linkcheck(sc);
1032         bridge_mutecaps(sc);    /* recalcuate now this interface is removed */
1033         bridge_rtdelete(sc, ifs, IFBF_FLUSHALL);
1034         KASSERT(bif->bif_addrcnt == 0,
1035             ("%s: %d bridge routes referenced", __func__, bif->bif_addrcnt));
1036
1037         ifs->if_bridge_output = NULL;
1038         ifs->if_bridge_input = NULL;
1039         ifs->if_bridge_linkstate = NULL;
1040         BRIDGE_UNLOCK(sc);
1041         if (!gone) {
1042                 switch (ifs->if_type) {
1043                 case IFT_ETHER:
1044                 case IFT_L2VLAN:
1045                         /*
1046                          * Take the interface out of promiscuous mode, but only
1047                          * if it was promiscuous in the first place. It might
1048                          * not be if we're in the bridge_ioctl_add() error path.
1049                          */
1050                         if (ifs->if_flags & IFF_PROMISC)
1051                                 (void) ifpromisc(ifs, 0);
1052                         break;
1053
1054                 case IFT_GIF:
1055                         break;
1056
1057                 default:
1058 #ifdef DIAGNOSTIC
1059                         panic("bridge_delete_member: impossible");
1060 #endif
1061                         break;
1062                 }
1063                 /* reneable any interface capabilities */
1064                 bridge_set_ifcap(sc, bif, bif->bif_savedcaps);
1065         }
1066         bstp_destroy(&bif->bif_stp);    /* prepare to free */
1067         BRIDGE_LOCK(sc);
1068         free(bif, M_DEVBUF);
1069 }
1070
1071 /*
1072  * bridge_delete_span:
1073  *
1074  *      Delete the specified span interface.
1075  */
1076 static void
1077 bridge_delete_span(struct bridge_softc *sc, struct bridge_iflist *bif)
1078 {
1079         BRIDGE_LOCK_ASSERT(sc);
1080
1081         KASSERT(bif->bif_ifp->if_bridge == NULL,
1082             ("%s: not a span interface", __func__));
1083
1084         LIST_REMOVE(bif, bif_next);
1085         free(bif, M_DEVBUF);
1086 }
1087
1088 static int
1089 bridge_ioctl_add(struct bridge_softc *sc, void *arg)
1090 {
1091         struct ifbreq *req = arg;
1092         struct bridge_iflist *bif = NULL;
1093         struct ifnet *ifs;
1094         int error = 0;
1095
1096         ifs = ifunit(req->ifbr_ifsname);
1097         if (ifs == NULL)
1098                 return (ENOENT);
1099         if (ifs->if_ioctl == NULL)      /* must be supported */
1100                 return (EINVAL);
1101
1102         /* If it's in the span list, it can't be a member. */
1103         LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1104                 if (ifs == bif->bif_ifp)
1105                         return (EBUSY);
1106
1107         if (ifs->if_bridge == sc)
1108                 return (EEXIST);
1109
1110         if (ifs->if_bridge != NULL)
1111                 return (EBUSY);
1112
1113         switch (ifs->if_type) {
1114         case IFT_ETHER:
1115         case IFT_L2VLAN:
1116         case IFT_GIF:
1117                 /* permitted interface types */
1118                 break;
1119         default:
1120                 return (EINVAL);
1121         }
1122
1123 #ifdef INET6
1124         /*
1125          * Two valid inet6 addresses with link-local scope must not be
1126          * on the parent interface and the member interfaces at the
1127          * same time.  This restriction is needed to prevent violation
1128          * of link-local scope zone.  Attempts to add a member
1129          * interface which has inet6 addresses when the parent has
1130          * inet6 triggers removal of all inet6 addresses on the member
1131          * interface.
1132          */
1133
1134         /* Check if the parent interface has a link-local scope addr. */
1135         if (V_allow_llz_overlap == 0 &&
1136             in6ifa_llaonifp(sc->sc_ifp) != NULL) {
1137                 /*
1138                  * If any, remove all inet6 addresses from the member
1139                  * interfaces.
1140                  */
1141                 BRIDGE_XLOCK(sc);
1142                 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
1143                         if (in6ifa_llaonifp(bif->bif_ifp)) {
1144                                 BRIDGE_UNLOCK(sc);
1145                                 in6_ifdetach(bif->bif_ifp);
1146                                 BRIDGE_LOCK(sc);
1147                                 if_printf(sc->sc_ifp,
1148                                     "IPv6 addresses on %s have been removed "
1149                                     "before adding it as a member to prevent "
1150                                     "IPv6 address scope violation.\n",
1151                                     bif->bif_ifp->if_xname);
1152                         }
1153                 }
1154                 BRIDGE_XDROP(sc);
1155                 if (in6ifa_llaonifp(ifs)) {
1156                         BRIDGE_UNLOCK(sc);
1157                         in6_ifdetach(ifs);
1158                         BRIDGE_LOCK(sc);
1159                         if_printf(sc->sc_ifp,
1160                             "IPv6 addresses on %s have been removed "
1161                             "before adding it as a member to prevent "
1162                             "IPv6 address scope violation.\n",
1163                             ifs->if_xname);
1164                 }
1165         }
1166 #endif
1167         /* Allow the first Ethernet member to define the MTU */
1168         if (LIST_EMPTY(&sc->sc_iflist))
1169                 sc->sc_ifp->if_mtu = ifs->if_mtu;
1170         else if (sc->sc_ifp->if_mtu != ifs->if_mtu) {
1171                 if_printf(sc->sc_ifp, "invalid MTU: %u(%s) != %u\n",
1172                     ifs->if_mtu, ifs->if_xname, sc->sc_ifp->if_mtu);
1173                 return (EINVAL);
1174         }
1175
1176         bif = malloc(sizeof(*bif), M_DEVBUF, M_NOWAIT|M_ZERO);
1177         if (bif == NULL)
1178                 return (ENOMEM);
1179
1180         bif->bif_ifp = ifs;
1181         bif->bif_flags = IFBIF_LEARNING | IFBIF_DISCOVER;
1182         bif->bif_savedcaps = ifs->if_capenable;
1183
1184         /*
1185          * Assign the interface's MAC address to the bridge if it's the first
1186          * member and the MAC address of the bridge has not been changed from
1187          * the default randomly generated one.
1188          */
1189         if (V_bridge_inherit_mac && LIST_EMPTY(&sc->sc_iflist) &&
1190             !memcmp(IF_LLADDR(sc->sc_ifp), sc->sc_defaddr.octet, ETHER_ADDR_LEN)) {
1191                 bcopy(IF_LLADDR(ifs), IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN);
1192                 sc->sc_ifaddr = ifs;
1193                 EVENTHANDLER_INVOKE(iflladdr_event, sc->sc_ifp);
1194         }
1195
1196         ifs->if_bridge = sc;
1197         ifs->if_bridge_output = bridge_output;
1198         ifs->if_bridge_input = bridge_input;
1199         ifs->if_bridge_linkstate = bridge_linkstate;
1200         bstp_create(&sc->sc_stp, &bif->bif_stp, bif->bif_ifp);
1201         /*
1202          * XXX: XLOCK HERE!?!
1203          *
1204          * NOTE: insert_***HEAD*** should be safe for the traversals.
1205          */
1206         LIST_INSERT_HEAD(&sc->sc_iflist, bif, bif_next);
1207
1208         /* Set interface capabilities to the intersection set of all members */
1209         bridge_mutecaps(sc);
1210         bridge_linkcheck(sc);
1211
1212         /* Place the interface into promiscuous mode */
1213         switch (ifs->if_type) {
1214                 case IFT_ETHER:
1215                 case IFT_L2VLAN:
1216                         BRIDGE_UNLOCK(sc);
1217                         error = ifpromisc(ifs, 1);
1218                         BRIDGE_LOCK(sc);
1219                         break;
1220         }
1221
1222         if (error)
1223                 bridge_delete_member(sc, bif, 0);
1224         return (error);
1225 }
1226
1227 static int
1228 bridge_ioctl_del(struct bridge_softc *sc, void *arg)
1229 {
1230         struct ifbreq *req = arg;
1231         struct bridge_iflist *bif;
1232
1233         bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1234         if (bif == NULL)
1235                 return (ENOENT);
1236
1237         bridge_delete_member(sc, bif, 0);
1238
1239         return (0);
1240 }
1241
1242 static int
1243 bridge_ioctl_gifflags(struct bridge_softc *sc, void *arg)
1244 {
1245         struct ifbreq *req = arg;
1246         struct bridge_iflist *bif;
1247         struct bstp_port *bp;
1248
1249         bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1250         if (bif == NULL)
1251                 return (ENOENT);
1252
1253         bp = &bif->bif_stp;
1254         req->ifbr_ifsflags = bif->bif_flags;
1255         req->ifbr_state = bp->bp_state;
1256         req->ifbr_priority = bp->bp_priority;
1257         req->ifbr_path_cost = bp->bp_path_cost;
1258         req->ifbr_portno = bif->bif_ifp->if_index & 0xfff;
1259         req->ifbr_proto = bp->bp_protover;
1260         req->ifbr_role = bp->bp_role;
1261         req->ifbr_stpflags = bp->bp_flags;
1262         req->ifbr_addrcnt = bif->bif_addrcnt;
1263         req->ifbr_addrmax = bif->bif_addrmax;
1264         req->ifbr_addrexceeded = bif->bif_addrexceeded;
1265
1266         /* Copy STP state options as flags */
1267         if (bp->bp_operedge)
1268                 req->ifbr_ifsflags |= IFBIF_BSTP_EDGE;
1269         if (bp->bp_flags & BSTP_PORT_AUTOEDGE)
1270                 req->ifbr_ifsflags |= IFBIF_BSTP_AUTOEDGE;
1271         if (bp->bp_ptp_link)
1272                 req->ifbr_ifsflags |= IFBIF_BSTP_PTP;
1273         if (bp->bp_flags & BSTP_PORT_AUTOPTP)
1274                 req->ifbr_ifsflags |= IFBIF_BSTP_AUTOPTP;
1275         if (bp->bp_flags & BSTP_PORT_ADMEDGE)
1276                 req->ifbr_ifsflags |= IFBIF_BSTP_ADMEDGE;
1277         if (bp->bp_flags & BSTP_PORT_ADMCOST)
1278                 req->ifbr_ifsflags |= IFBIF_BSTP_ADMCOST;
1279         return (0);
1280 }
1281
1282 static int
1283 bridge_ioctl_sifflags(struct bridge_softc *sc, void *arg)
1284 {
1285         struct ifbreq *req = arg;
1286         struct bridge_iflist *bif;
1287         struct bstp_port *bp;
1288         int error;
1289
1290         bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1291         if (bif == NULL)
1292                 return (ENOENT);
1293         bp = &bif->bif_stp;
1294
1295         if (req->ifbr_ifsflags & IFBIF_SPAN)
1296                 /* SPAN is readonly */
1297                 return (EINVAL);
1298
1299         if (req->ifbr_ifsflags & IFBIF_STP) {
1300                 if ((bif->bif_flags & IFBIF_STP) == 0) {
1301                         error = bstp_enable(&bif->bif_stp);
1302                         if (error)
1303                                 return (error);
1304                 }
1305         } else {
1306                 if ((bif->bif_flags & IFBIF_STP) != 0)
1307                         bstp_disable(&bif->bif_stp);
1308         }
1309
1310         /* Pass on STP flags */
1311         bstp_set_edge(bp, req->ifbr_ifsflags & IFBIF_BSTP_EDGE ? 1 : 0);
1312         bstp_set_autoedge(bp, req->ifbr_ifsflags & IFBIF_BSTP_AUTOEDGE ? 1 : 0);
1313         bstp_set_ptp(bp, req->ifbr_ifsflags & IFBIF_BSTP_PTP ? 1 : 0);
1314         bstp_set_autoptp(bp, req->ifbr_ifsflags & IFBIF_BSTP_AUTOPTP ? 1 : 0);
1315
1316         /* Save the bits relating to the bridge */
1317         bif->bif_flags = req->ifbr_ifsflags & IFBIFMASK;
1318
1319         return (0);
1320 }
1321
1322 static int
1323 bridge_ioctl_scache(struct bridge_softc *sc, void *arg)
1324 {
1325         struct ifbrparam *param = arg;
1326
1327         sc->sc_brtmax = param->ifbrp_csize;
1328         bridge_rttrim(sc);
1329
1330         return (0);
1331 }
1332
1333 static int
1334 bridge_ioctl_gcache(struct bridge_softc *sc, void *arg)
1335 {
1336         struct ifbrparam *param = arg;
1337
1338         param->ifbrp_csize = sc->sc_brtmax;
1339
1340         return (0);
1341 }
1342
1343 static int
1344 bridge_ioctl_gifs(struct bridge_softc *sc, void *arg)
1345 {
1346         struct ifbifconf *bifc = arg;
1347         struct bridge_iflist *bif;
1348         struct ifbreq breq;
1349         char *buf, *outbuf;
1350         int count, buflen, len, error = 0;
1351
1352         count = 0;
1353         LIST_FOREACH(bif, &sc->sc_iflist, bif_next)
1354                 count++;
1355         LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1356                 count++;
1357
1358         buflen = sizeof(breq) * count;
1359         if (bifc->ifbic_len == 0) {
1360                 bifc->ifbic_len = buflen;
1361                 return (0);
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(bifc->ifbic_len, buflen);
1370         bzero(&breq, sizeof(breq));
1371         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
1372                 if (len < sizeof(breq))
1373                         break;
1374
1375                 strlcpy(breq.ifbr_ifsname, bif->bif_ifp->if_xname,
1376                     sizeof(breq.ifbr_ifsname));
1377                 /* Fill in the ifbreq structure */
1378                 error = bridge_ioctl_gifflags(sc, &breq);
1379                 if (error)
1380                         break;
1381                 memcpy(buf, &breq, sizeof(breq));
1382                 count++;
1383                 buf += sizeof(breq);
1384                 len -= sizeof(breq);
1385         }
1386         LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) {
1387                 if (len < sizeof(breq))
1388                         break;
1389
1390                 strlcpy(breq.ifbr_ifsname, bif->bif_ifp->if_xname,
1391                     sizeof(breq.ifbr_ifsname));
1392                 breq.ifbr_ifsflags = bif->bif_flags;
1393                 breq.ifbr_portno = bif->bif_ifp->if_index & 0xfff;
1394                 memcpy(buf, &breq, sizeof(breq));
1395                 count++;
1396                 buf += sizeof(breq);
1397                 len -= sizeof(breq);
1398         }
1399
1400         BRIDGE_UNLOCK(sc);
1401         bifc->ifbic_len = sizeof(breq) * count;
1402         error = copyout(outbuf, bifc->ifbic_req, bifc->ifbic_len);
1403         BRIDGE_LOCK(sc);
1404         free(outbuf, M_TEMP);
1405         return (error);
1406 }
1407
1408 static int
1409 bridge_ioctl_rts(struct bridge_softc *sc, void *arg)
1410 {
1411         struct ifbaconf *bac = arg;
1412         struct bridge_rtnode *brt;
1413         struct ifbareq bareq;
1414         char *buf, *outbuf;
1415         int count, buflen, len, error = 0;
1416
1417         if (bac->ifbac_len == 0)
1418                 return (0);
1419
1420         count = 0;
1421         LIST_FOREACH(brt, &sc->sc_rtlist, brt_list)
1422                 count++;
1423         buflen = sizeof(bareq) * count;
1424
1425         BRIDGE_UNLOCK(sc);
1426         outbuf = malloc(buflen, M_TEMP, M_WAITOK | M_ZERO);
1427         BRIDGE_LOCK(sc);
1428
1429         count = 0;
1430         buf = outbuf;
1431         len = min(bac->ifbac_len, buflen);
1432         bzero(&bareq, sizeof(bareq));
1433         LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) {
1434                 if (len < sizeof(bareq))
1435                         goto out;
1436                 strlcpy(bareq.ifba_ifsname, brt->brt_ifp->if_xname,
1437                     sizeof(bareq.ifba_ifsname));
1438                 memcpy(bareq.ifba_dst, brt->brt_addr, sizeof(brt->brt_addr));
1439                 bareq.ifba_vlan = brt->brt_vlan;
1440                 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC &&
1441                                 time_uptime < brt->brt_expire)
1442                         bareq.ifba_expire = brt->brt_expire - time_uptime;
1443                 else
1444                         bareq.ifba_expire = 0;
1445                 bareq.ifba_flags = brt->brt_flags;
1446
1447                 memcpy(buf, &bareq, sizeof(bareq));
1448                 count++;
1449                 buf += sizeof(bareq);
1450                 len -= sizeof(bareq);
1451         }
1452 out:
1453         BRIDGE_UNLOCK(sc);
1454         bac->ifbac_len = sizeof(bareq) * count;
1455         error = copyout(outbuf, bac->ifbac_req, bac->ifbac_len);
1456         BRIDGE_LOCK(sc);
1457         free(outbuf, M_TEMP);
1458         return (error);
1459 }
1460
1461 static int
1462 bridge_ioctl_saddr(struct bridge_softc *sc, void *arg)
1463 {
1464         struct ifbareq *req = arg;
1465         struct bridge_iflist *bif;
1466         int error;
1467
1468         bif = bridge_lookup_member(sc, req->ifba_ifsname);
1469         if (bif == NULL)
1470                 return (ENOENT);
1471
1472         error = bridge_rtupdate(sc, req->ifba_dst, req->ifba_vlan, bif, 1,
1473             req->ifba_flags);
1474
1475         return (error);
1476 }
1477
1478 static int
1479 bridge_ioctl_sto(struct bridge_softc *sc, void *arg)
1480 {
1481         struct ifbrparam *param = arg;
1482
1483         sc->sc_brttimeout = param->ifbrp_ctime;
1484         return (0);
1485 }
1486
1487 static int
1488 bridge_ioctl_gto(struct bridge_softc *sc, void *arg)
1489 {
1490         struct ifbrparam *param = arg;
1491
1492         param->ifbrp_ctime = sc->sc_brttimeout;
1493         return (0);
1494 }
1495
1496 static int
1497 bridge_ioctl_daddr(struct bridge_softc *sc, void *arg)
1498 {
1499         struct ifbareq *req = arg;
1500
1501         return (bridge_rtdaddr(sc, req->ifba_dst, req->ifba_vlan));
1502 }
1503
1504 static int
1505 bridge_ioctl_flush(struct bridge_softc *sc, void *arg)
1506 {
1507         struct ifbreq *req = arg;
1508
1509         bridge_rtflush(sc, req->ifbr_ifsflags);
1510         return (0);
1511 }
1512
1513 static int
1514 bridge_ioctl_gpri(struct bridge_softc *sc, void *arg)
1515 {
1516         struct ifbrparam *param = arg;
1517         struct bstp_state *bs = &sc->sc_stp;
1518
1519         param->ifbrp_prio = bs->bs_bridge_priority;
1520         return (0);
1521 }
1522
1523 static int
1524 bridge_ioctl_spri(struct bridge_softc *sc, void *arg)
1525 {
1526         struct ifbrparam *param = arg;
1527
1528         return (bstp_set_priority(&sc->sc_stp, param->ifbrp_prio));
1529 }
1530
1531 static int
1532 bridge_ioctl_ght(struct bridge_softc *sc, void *arg)
1533 {
1534         struct ifbrparam *param = arg;
1535         struct bstp_state *bs = &sc->sc_stp;
1536
1537         param->ifbrp_hellotime = bs->bs_bridge_htime >> 8;
1538         return (0);
1539 }
1540
1541 static int
1542 bridge_ioctl_sht(struct bridge_softc *sc, void *arg)
1543 {
1544         struct ifbrparam *param = arg;
1545
1546         return (bstp_set_htime(&sc->sc_stp, param->ifbrp_hellotime));
1547 }
1548
1549 static int
1550 bridge_ioctl_gfd(struct bridge_softc *sc, void *arg)
1551 {
1552         struct ifbrparam *param = arg;
1553         struct bstp_state *bs = &sc->sc_stp;
1554
1555         param->ifbrp_fwddelay = bs->bs_bridge_fdelay >> 8;
1556         return (0);
1557 }
1558
1559 static int
1560 bridge_ioctl_sfd(struct bridge_softc *sc, void *arg)
1561 {
1562         struct ifbrparam *param = arg;
1563
1564         return (bstp_set_fdelay(&sc->sc_stp, param->ifbrp_fwddelay));
1565 }
1566
1567 static int
1568 bridge_ioctl_gma(struct bridge_softc *sc, void *arg)
1569 {
1570         struct ifbrparam *param = arg;
1571         struct bstp_state *bs = &sc->sc_stp;
1572
1573         param->ifbrp_maxage = bs->bs_bridge_max_age >> 8;
1574         return (0);
1575 }
1576
1577 static int
1578 bridge_ioctl_sma(struct bridge_softc *sc, void *arg)
1579 {
1580         struct ifbrparam *param = arg;
1581
1582         return (bstp_set_maxage(&sc->sc_stp, param->ifbrp_maxage));
1583 }
1584
1585 static int
1586 bridge_ioctl_sifprio(struct bridge_softc *sc, void *arg)
1587 {
1588         struct ifbreq *req = arg;
1589         struct bridge_iflist *bif;
1590
1591         bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1592         if (bif == NULL)
1593                 return (ENOENT);
1594
1595         return (bstp_set_port_priority(&bif->bif_stp, req->ifbr_priority));
1596 }
1597
1598 static int
1599 bridge_ioctl_sifcost(struct bridge_softc *sc, void *arg)
1600 {
1601         struct ifbreq *req = arg;
1602         struct bridge_iflist *bif;
1603
1604         bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1605         if (bif == NULL)
1606                 return (ENOENT);
1607
1608         return (bstp_set_path_cost(&bif->bif_stp, req->ifbr_path_cost));
1609 }
1610
1611 static int
1612 bridge_ioctl_sifmaxaddr(struct bridge_softc *sc, void *arg)
1613 {
1614         struct ifbreq *req = arg;
1615         struct bridge_iflist *bif;
1616
1617         bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1618         if (bif == NULL)
1619                 return (ENOENT);
1620
1621         bif->bif_addrmax = req->ifbr_addrmax;
1622         return (0);
1623 }
1624
1625 static int
1626 bridge_ioctl_addspan(struct bridge_softc *sc, void *arg)
1627 {
1628         struct ifbreq *req = arg;
1629         struct bridge_iflist *bif = NULL;
1630         struct ifnet *ifs;
1631
1632         ifs = ifunit(req->ifbr_ifsname);
1633         if (ifs == NULL)
1634                 return (ENOENT);
1635
1636         LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1637                 if (ifs == bif->bif_ifp)
1638                         return (EBUSY);
1639
1640         if (ifs->if_bridge != NULL)
1641                 return (EBUSY);
1642
1643         switch (ifs->if_type) {
1644                 case IFT_ETHER:
1645                 case IFT_GIF:
1646                 case IFT_L2VLAN:
1647                         break;
1648                 default:
1649                         return (EINVAL);
1650         }
1651
1652         bif = malloc(sizeof(*bif), M_DEVBUF, M_NOWAIT|M_ZERO);
1653         if (bif == NULL)
1654                 return (ENOMEM);
1655
1656         bif->bif_ifp = ifs;
1657         bif->bif_flags = IFBIF_SPAN;
1658
1659         LIST_INSERT_HEAD(&sc->sc_spanlist, bif, bif_next);
1660
1661         return (0);
1662 }
1663
1664 static int
1665 bridge_ioctl_delspan(struct bridge_softc *sc, void *arg)
1666 {
1667         struct ifbreq *req = arg;
1668         struct bridge_iflist *bif;
1669         struct ifnet *ifs;
1670
1671         ifs = ifunit(req->ifbr_ifsname);
1672         if (ifs == NULL)
1673                 return (ENOENT);
1674
1675         LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1676                 if (ifs == bif->bif_ifp)
1677                         break;
1678
1679         if (bif == NULL)
1680                 return (ENOENT);
1681
1682         bridge_delete_span(sc, bif);
1683
1684         return (0);
1685 }
1686
1687 static int
1688 bridge_ioctl_gbparam(struct bridge_softc *sc, void *arg)
1689 {
1690         struct ifbropreq *req = arg;
1691         struct bstp_state *bs = &sc->sc_stp;
1692         struct bstp_port *root_port;
1693
1694         req->ifbop_maxage = bs->bs_bridge_max_age >> 8;
1695         req->ifbop_hellotime = bs->bs_bridge_htime >> 8;
1696         req->ifbop_fwddelay = bs->bs_bridge_fdelay >> 8;
1697
1698         root_port = bs->bs_root_port;
1699         if (root_port == NULL)
1700                 req->ifbop_root_port = 0;
1701         else
1702                 req->ifbop_root_port = root_port->bp_ifp->if_index;
1703
1704         req->ifbop_holdcount = bs->bs_txholdcount;
1705         req->ifbop_priority = bs->bs_bridge_priority;
1706         req->ifbop_protocol = bs->bs_protover;
1707         req->ifbop_root_path_cost = bs->bs_root_pv.pv_cost;
1708         req->ifbop_bridgeid = bs->bs_bridge_pv.pv_dbridge_id;
1709         req->ifbop_designated_root = bs->bs_root_pv.pv_root_id;
1710         req->ifbop_designated_bridge = bs->bs_root_pv.pv_dbridge_id;
1711         req->ifbop_last_tc_time.tv_sec = bs->bs_last_tc_time.tv_sec;
1712         req->ifbop_last_tc_time.tv_usec = bs->bs_last_tc_time.tv_usec;
1713
1714         return (0);
1715 }
1716
1717 static int
1718 bridge_ioctl_grte(struct bridge_softc *sc, void *arg)
1719 {
1720         struct ifbrparam *param = arg;
1721
1722         param->ifbrp_cexceeded = sc->sc_brtexceeded;
1723         return (0);
1724 }
1725
1726 static int
1727 bridge_ioctl_gifsstp(struct bridge_softc *sc, void *arg)
1728 {
1729         struct ifbpstpconf *bifstp = arg;
1730         struct bridge_iflist *bif;
1731         struct bstp_port *bp;
1732         struct ifbpstpreq bpreq;
1733         char *buf, *outbuf;
1734         int count, buflen, len, error = 0;
1735
1736         count = 0;
1737         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
1738                 if ((bif->bif_flags & IFBIF_STP) != 0)
1739                         count++;
1740         }
1741
1742         buflen = sizeof(bpreq) * count;
1743         if (bifstp->ifbpstp_len == 0) {
1744                 bifstp->ifbpstp_len = buflen;
1745                 return (0);
1746         }
1747
1748         BRIDGE_UNLOCK(sc);
1749         outbuf = malloc(buflen, M_TEMP, M_WAITOK | M_ZERO);
1750         BRIDGE_LOCK(sc);
1751
1752         count = 0;
1753         buf = outbuf;
1754         len = min(bifstp->ifbpstp_len, buflen);
1755         bzero(&bpreq, sizeof(bpreq));
1756         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
1757                 if (len < sizeof(bpreq))
1758                         break;
1759
1760                 if ((bif->bif_flags & IFBIF_STP) == 0)
1761                         continue;
1762
1763                 bp = &bif->bif_stp;
1764                 bpreq.ifbp_portno = bif->bif_ifp->if_index & 0xfff;
1765                 bpreq.ifbp_fwd_trans = bp->bp_forward_transitions;
1766                 bpreq.ifbp_design_cost = bp->bp_desg_pv.pv_cost;
1767                 bpreq.ifbp_design_port = bp->bp_desg_pv.pv_port_id;
1768                 bpreq.ifbp_design_bridge = bp->bp_desg_pv.pv_dbridge_id;
1769                 bpreq.ifbp_design_root = bp->bp_desg_pv.pv_root_id;
1770
1771                 memcpy(buf, &bpreq, sizeof(bpreq));
1772                 count++;
1773                 buf += sizeof(bpreq);
1774                 len -= sizeof(bpreq);
1775         }
1776
1777         BRIDGE_UNLOCK(sc);
1778         bifstp->ifbpstp_len = sizeof(bpreq) * count;
1779         error = copyout(outbuf, bifstp->ifbpstp_req, bifstp->ifbpstp_len);
1780         BRIDGE_LOCK(sc);
1781         free(outbuf, M_TEMP);
1782         return (error);
1783 }
1784
1785 static int
1786 bridge_ioctl_sproto(struct bridge_softc *sc, void *arg)
1787 {
1788         struct ifbrparam *param = arg;
1789
1790         return (bstp_set_protocol(&sc->sc_stp, param->ifbrp_proto));
1791 }
1792
1793 static int
1794 bridge_ioctl_stxhc(struct bridge_softc *sc, void *arg)
1795 {
1796         struct ifbrparam *param = arg;
1797
1798         return (bstp_set_holdcount(&sc->sc_stp, param->ifbrp_txhc));
1799 }
1800
1801 /*
1802  * bridge_ifdetach:
1803  *
1804  *      Detach an interface from a bridge.  Called when a member
1805  *      interface is detaching.
1806  */
1807 static void
1808 bridge_ifdetach(void *arg __unused, struct ifnet *ifp)
1809 {
1810         struct bridge_softc *sc = ifp->if_bridge;
1811         struct bridge_iflist *bif;
1812
1813         if (ifp->if_flags & IFF_RENAMING)
1814                 return;
1815         if (V_bridge_cloner == NULL) {
1816                 /*
1817                  * This detach handler can be called after
1818                  * vnet_bridge_uninit().  Just return in that case.
1819                  */
1820                 return;
1821         }
1822         /* Check if the interface is a bridge member */
1823         if (sc != NULL) {
1824                 BRIDGE_LOCK(sc);
1825
1826                 bif = bridge_lookup_member_if(sc, ifp);
1827                 if (bif != NULL)
1828                         bridge_delete_member(sc, bif, 1);
1829
1830                 BRIDGE_UNLOCK(sc);
1831                 return;
1832         }
1833
1834         /* Check if the interface is a span port */
1835         BRIDGE_LIST_LOCK();
1836         LIST_FOREACH(sc, &V_bridge_list, sc_list) {
1837                 BRIDGE_LOCK(sc);
1838                 LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1839                         if (ifp == bif->bif_ifp) {
1840                                 bridge_delete_span(sc, bif);
1841                                 break;
1842                         }
1843
1844                 BRIDGE_UNLOCK(sc);
1845         }
1846         BRIDGE_LIST_UNLOCK();
1847 }
1848
1849 /*
1850  * bridge_init:
1851  *
1852  *      Initialize a bridge interface.
1853  */
1854 static void
1855 bridge_init(void *xsc)
1856 {
1857         struct bridge_softc *sc = (struct bridge_softc *)xsc;
1858         struct ifnet *ifp = sc->sc_ifp;
1859
1860         if (ifp->if_drv_flags & IFF_DRV_RUNNING)
1861                 return;
1862
1863         BRIDGE_LOCK(sc);
1864         callout_reset(&sc->sc_brcallout, bridge_rtable_prune_period * hz,
1865             bridge_timer, sc);
1866
1867         ifp->if_drv_flags |= IFF_DRV_RUNNING;
1868         bstp_init(&sc->sc_stp);         /* Initialize Spanning Tree */
1869
1870         BRIDGE_UNLOCK(sc);
1871 }
1872
1873 /*
1874  * bridge_stop:
1875  *
1876  *      Stop the bridge interface.
1877  */
1878 static void
1879 bridge_stop(struct ifnet *ifp, int disable)
1880 {
1881         struct bridge_softc *sc = ifp->if_softc;
1882
1883         BRIDGE_LOCK_ASSERT(sc);
1884
1885         if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0)
1886                 return;
1887
1888         callout_stop(&sc->sc_brcallout);
1889         bstp_stop(&sc->sc_stp);
1890
1891         bridge_rtflush(sc, IFBF_FLUSHDYN);
1892
1893         ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
1894 }
1895
1896 /*
1897  * bridge_enqueue:
1898  *
1899  *      Enqueue a packet on a bridge member interface.
1900  *
1901  */
1902 static int
1903 bridge_enqueue(struct bridge_softc *sc, struct ifnet *dst_ifp, struct mbuf *m)
1904 {
1905         int len, err = 0;
1906         short mflags;
1907         struct mbuf *m0;
1908
1909         /* We may be sending a fragment so traverse the mbuf */
1910         for (; m; m = m0) {
1911                 m0 = m->m_nextpkt;
1912                 m->m_nextpkt = NULL;
1913                 len = m->m_pkthdr.len;
1914                 mflags = m->m_flags;
1915
1916                 /*
1917                  * If underlying interface can not do VLAN tag insertion itself
1918                  * then attach a packet tag that holds it.
1919                  */
1920                 if ((m->m_flags & M_VLANTAG) &&
1921                     (dst_ifp->if_capenable & IFCAP_VLAN_HWTAGGING) == 0) {
1922                         m = ether_vlanencap(m, m->m_pkthdr.ether_vtag);
1923                         if (m == NULL) {
1924                                 if_printf(dst_ifp,
1925                                     "unable to prepend VLAN header\n");
1926                                 if_inc_counter(dst_ifp, IFCOUNTER_OERRORS, 1);
1927                                 continue;
1928                         }
1929                         m->m_flags &= ~M_VLANTAG;
1930                 }
1931
1932                 M_ASSERTPKTHDR(m); /* We shouldn't transmit mbuf without pkthdr */
1933                 if ((err = dst_ifp->if_transmit(dst_ifp, m))) {
1934                         m_freem(m0);
1935                         if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
1936                         break;
1937                 }
1938
1939                 if_inc_counter(sc->sc_ifp, IFCOUNTER_OPACKETS, 1);
1940                 if_inc_counter(sc->sc_ifp, IFCOUNTER_OBYTES, len);
1941                 if (mflags & M_MCAST)
1942                         if_inc_counter(sc->sc_ifp, IFCOUNTER_OMCASTS, 1);
1943         }
1944
1945         return (err);
1946 }
1947
1948 /*
1949  * bridge_dummynet:
1950  *
1951  *      Receive a queued packet from dummynet and pass it on to the output
1952  *      interface.
1953  *
1954  *      The mbuf has the Ethernet header already attached.
1955  */
1956 static void
1957 bridge_dummynet(struct mbuf *m, struct ifnet *ifp)
1958 {
1959         struct bridge_softc *sc;
1960
1961         sc = ifp->if_bridge;
1962
1963         /*
1964          * The packet didnt originate from a member interface. This should only
1965          * ever happen if a member interface is removed while packets are
1966          * queued for it.
1967          */
1968         if (sc == NULL) {
1969                 m_freem(m);
1970                 return;
1971         }
1972
1973         if (PFIL_HOOKED_OUT(V_inet_pfil_head)
1974 #ifdef INET6
1975             || PFIL_HOOKED_OUT(V_inet6_pfil_head)
1976 #endif
1977             ) {
1978                 if (bridge_pfil(&m, sc->sc_ifp, ifp, PFIL_OUT) != 0)
1979                         return;
1980                 if (m == NULL)
1981                         return;
1982         }
1983
1984         bridge_enqueue(sc, ifp, m);
1985 }
1986
1987 /*
1988  * bridge_output:
1989  *
1990  *      Send output from a bridge member interface.  This
1991  *      performs the bridging function for locally originated
1992  *      packets.
1993  *
1994  *      The mbuf has the Ethernet header already attached.  We must
1995  *      enqueue or free the mbuf before returning.
1996  */
1997 static int
1998 bridge_output(struct ifnet *ifp, struct mbuf *m, struct sockaddr *sa,
1999     struct rtentry *rt)
2000 {
2001         struct ether_header *eh;
2002         struct ifnet *dst_if;
2003         struct bridge_softc *sc;
2004         uint16_t vlan;
2005
2006         if (m->m_len < ETHER_HDR_LEN) {
2007                 m = m_pullup(m, ETHER_HDR_LEN);
2008                 if (m == NULL)
2009                         return (0);
2010         }
2011
2012         eh = mtod(m, struct ether_header *);
2013         sc = ifp->if_bridge;
2014         vlan = VLANTAGOF(m);
2015
2016         BRIDGE_LOCK(sc);
2017
2018         /*
2019          * If bridge is down, but the original output interface is up,
2020          * go ahead and send out that interface.  Otherwise, the packet
2021          * is dropped below.
2022          */
2023         if ((sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) {
2024                 dst_if = ifp;
2025                 goto sendunicast;
2026         }
2027
2028         /*
2029          * If the packet is a multicast, or we don't know a better way to
2030          * get there, send to all interfaces.
2031          */
2032         if (ETHER_IS_MULTICAST(eh->ether_dhost))
2033                 dst_if = NULL;
2034         else
2035                 dst_if = bridge_rtlookup(sc, eh->ether_dhost, vlan);
2036         if (dst_if == NULL) {
2037                 struct bridge_iflist *bif;
2038                 struct mbuf *mc;
2039                 int error = 0, used = 0;
2040
2041                 bridge_span(sc, m);
2042
2043                 BRIDGE_LOCK2REF(sc, error);
2044                 if (error) {
2045                         m_freem(m);
2046                         return (0);
2047                 }
2048
2049                 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
2050                         dst_if = bif->bif_ifp;
2051
2052                         if (dst_if->if_type == IFT_GIF)
2053                                 continue;
2054                         if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
2055                                 continue;
2056
2057                         /*
2058                          * If this is not the original output interface,
2059                          * and the interface is participating in spanning
2060                          * tree, make sure the port is in a state that
2061                          * allows forwarding.
2062                          */
2063                         if (dst_if != ifp && (bif->bif_flags & IFBIF_STP) &&
2064                             bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
2065                                 continue;
2066
2067                         if (LIST_NEXT(bif, bif_next) == NULL) {
2068                                 used = 1;
2069                                 mc = m;
2070                         } else {
2071                                 mc = m_copypacket(m, M_NOWAIT);
2072                                 if (mc == NULL) {
2073                                         if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
2074                                         continue;
2075                                 }
2076                         }
2077
2078                         bridge_enqueue(sc, dst_if, mc);
2079                 }
2080                 if (used == 0)
2081                         m_freem(m);
2082                 BRIDGE_UNREF(sc);
2083                 return (0);
2084         }
2085
2086 sendunicast:
2087         /*
2088          * XXX Spanning tree consideration here?
2089          */
2090
2091         bridge_span(sc, m);
2092         if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) {
2093                 m_freem(m);
2094                 BRIDGE_UNLOCK(sc);
2095                 return (0);
2096         }
2097
2098         BRIDGE_UNLOCK(sc);
2099         bridge_enqueue(sc, dst_if, m);
2100         return (0);
2101 }
2102
2103 /*
2104  * bridge_transmit:
2105  *
2106  *      Do output on a bridge.
2107  *
2108  */
2109 static int
2110 bridge_transmit(struct ifnet *ifp, struct mbuf *m)
2111 {
2112         struct bridge_softc *sc;
2113         struct ether_header *eh;
2114         struct ifnet *dst_if;
2115         int error = 0;
2116
2117         sc = ifp->if_softc;
2118
2119         ETHER_BPF_MTAP(ifp, m);
2120
2121         eh = mtod(m, struct ether_header *);
2122
2123         BRIDGE_LOCK(sc);
2124         if (((m->m_flags & (M_BCAST|M_MCAST)) == 0) &&
2125             (dst_if = bridge_rtlookup(sc, eh->ether_dhost, 1)) != NULL) {
2126                 BRIDGE_UNLOCK(sc);
2127                 error = bridge_enqueue(sc, dst_if, m);
2128         } else
2129                 bridge_broadcast(sc, ifp, m, 0);
2130
2131         return (error);
2132 }
2133
2134 /*
2135  * The ifp->if_qflush entry point for if_bridge(4) is no-op.
2136  */
2137 static void
2138 bridge_qflush(struct ifnet *ifp __unused)
2139 {
2140 }
2141
2142 /*
2143  * bridge_forward:
2144  *
2145  *      The forwarding function of the bridge.
2146  *
2147  *      NOTE: Releases the lock on return.
2148  */
2149 static void
2150 bridge_forward(struct bridge_softc *sc, struct bridge_iflist *sbif,
2151     struct mbuf *m)
2152 {
2153         struct bridge_iflist *dbif;
2154         struct ifnet *src_if, *dst_if, *ifp;
2155         struct ether_header *eh;
2156         uint16_t vlan;
2157         uint8_t *dst;
2158         int error;
2159
2160         src_if = m->m_pkthdr.rcvif;
2161         ifp = sc->sc_ifp;
2162
2163         if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1);
2164         if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len);
2165         vlan = VLANTAGOF(m);
2166
2167         if ((sbif->bif_flags & IFBIF_STP) &&
2168             sbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
2169                 goto drop;
2170
2171         eh = mtod(m, struct ether_header *);
2172         dst = eh->ether_dhost;
2173
2174         /* If the interface is learning, record the address. */
2175         if (sbif->bif_flags & IFBIF_LEARNING) {
2176                 error = bridge_rtupdate(sc, eh->ether_shost, vlan,
2177                     sbif, 0, IFBAF_DYNAMIC);
2178                 /*
2179                  * If the interface has addresses limits then deny any source
2180                  * that is not in the cache.
2181                  */
2182                 if (error && sbif->bif_addrmax)
2183                         goto drop;
2184         }
2185
2186         if ((sbif->bif_flags & IFBIF_STP) != 0 &&
2187             sbif->bif_stp.bp_state == BSTP_IFSTATE_LEARNING)
2188                 goto drop;
2189
2190         /*
2191          * At this point, the port either doesn't participate
2192          * in spanning tree or it is in the forwarding state.
2193          */
2194
2195         /*
2196          * If the packet is unicast, destined for someone on
2197          * "this" side of the bridge, drop it.
2198          */
2199         if ((m->m_flags & (M_BCAST|M_MCAST)) == 0) {
2200                 dst_if = bridge_rtlookup(sc, dst, vlan);
2201                 if (src_if == dst_if)
2202                         goto drop;
2203         } else {
2204                 /*
2205                  * Check if its a reserved multicast address, any address
2206                  * listed in 802.1D section 7.12.6 may not be forwarded by the
2207                  * bridge.
2208                  * This is currently 01-80-C2-00-00-00 to 01-80-C2-00-00-0F
2209                  */
2210                 if (dst[0] == 0x01 && dst[1] == 0x80 &&
2211                     dst[2] == 0xc2 && dst[3] == 0x00 &&
2212                     dst[4] == 0x00 && dst[5] <= 0x0f)
2213                         goto drop;
2214
2215                 /* ...forward it to all interfaces. */
2216                 if_inc_counter(ifp, IFCOUNTER_IMCASTS, 1);
2217                 dst_if = NULL;
2218         }
2219
2220         /*
2221          * If we have a destination interface which is a member of our bridge,
2222          * OR this is a unicast packet, push it through the bpf(4) machinery.
2223          * For broadcast or multicast packets, don't bother because it will
2224          * be reinjected into ether_input. We do this before we pass the packets
2225          * through the pfil(9) framework, as it is possible that pfil(9) will
2226          * drop the packet, or possibly modify it, making it difficult to debug
2227          * firewall issues on the bridge.
2228          */
2229         if (dst_if != NULL || (m->m_flags & (M_BCAST | M_MCAST)) == 0)
2230                 ETHER_BPF_MTAP(ifp, m);
2231
2232         /* run the packet filter */
2233         if (PFIL_HOOKED_IN(V_inet_pfil_head)
2234 #ifdef INET6
2235             || PFIL_HOOKED_IN(V_inet6_pfil_head)
2236 #endif
2237             ) {
2238                 BRIDGE_UNLOCK(sc);
2239                 if (bridge_pfil(&m, ifp, src_if, PFIL_IN) != 0)
2240                         return;
2241                 if (m == NULL)
2242                         return;
2243                 BRIDGE_LOCK(sc);
2244         }
2245
2246         if (dst_if == NULL) {
2247                 bridge_broadcast(sc, src_if, m, 1);
2248                 return;
2249         }
2250
2251         /*
2252          * At this point, we're dealing with a unicast frame
2253          * going to a different interface.
2254          */
2255         if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
2256                 goto drop;
2257
2258         dbif = bridge_lookup_member_if(sc, dst_if);
2259         if (dbif == NULL)
2260                 /* Not a member of the bridge (anymore?) */
2261                 goto drop;
2262
2263         /* Private segments can not talk to each other */
2264         if (sbif->bif_flags & dbif->bif_flags & IFBIF_PRIVATE)
2265                 goto drop;
2266
2267         if ((dbif->bif_flags & IFBIF_STP) &&
2268             dbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
2269                 goto drop;
2270
2271         BRIDGE_UNLOCK(sc);
2272
2273         if (PFIL_HOOKED_OUT(V_inet_pfil_head)
2274 #ifdef INET6
2275             || PFIL_HOOKED_OUT(V_inet6_pfil_head)
2276 #endif
2277             ) {
2278                 if (bridge_pfil(&m, ifp, dst_if, PFIL_OUT) != 0)
2279                         return;
2280                 if (m == NULL)
2281                         return;
2282         }
2283
2284         bridge_enqueue(sc, dst_if, m);
2285         return;
2286
2287 drop:
2288         BRIDGE_UNLOCK(sc);
2289         m_freem(m);
2290 }
2291
2292 /*
2293  * bridge_input:
2294  *
2295  *      Receive input from a member interface.  Queue the packet for
2296  *      bridging if it is not for us.
2297  */
2298 static struct mbuf *
2299 bridge_input(struct ifnet *ifp, struct mbuf *m)
2300 {
2301         struct bridge_softc *sc = ifp->if_bridge;
2302         struct bridge_iflist *bif, *bif2;
2303         struct ifnet *bifp;
2304         struct ether_header *eh;
2305         struct mbuf *mc, *mc2;
2306         uint16_t vlan;
2307         int error;
2308
2309         if ((sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING) == 0)
2310                 return (m);
2311
2312         bifp = sc->sc_ifp;
2313         vlan = VLANTAGOF(m);
2314
2315         /*
2316          * Implement support for bridge monitoring. If this flag has been
2317          * set on this interface, discard the packet once we push it through
2318          * the bpf(4) machinery, but before we do, increment the byte and
2319          * packet counters associated with this interface.
2320          */
2321         if ((bifp->if_flags & IFF_MONITOR) != 0) {
2322                 m->m_pkthdr.rcvif  = bifp;
2323                 ETHER_BPF_MTAP(bifp, m);
2324                 if_inc_counter(bifp, IFCOUNTER_IPACKETS, 1);
2325                 if_inc_counter(bifp, IFCOUNTER_IBYTES, m->m_pkthdr.len);
2326                 m_freem(m);
2327                 return (NULL);
2328         }
2329         BRIDGE_LOCK(sc);
2330         bif = bridge_lookup_member_if(sc, ifp);
2331         if (bif == NULL) {
2332                 BRIDGE_UNLOCK(sc);
2333                 return (m);
2334         }
2335
2336         eh = mtod(m, struct ether_header *);
2337
2338         bridge_span(sc, m);
2339
2340         if (m->m_flags & (M_BCAST|M_MCAST)) {
2341                 /* Tap off 802.1D packets; they do not get forwarded. */
2342                 if (memcmp(eh->ether_dhost, bstp_etheraddr,
2343                     ETHER_ADDR_LEN) == 0) {
2344                         bstp_input(&bif->bif_stp, ifp, m); /* consumes mbuf */
2345                         BRIDGE_UNLOCK(sc);
2346                         return (NULL);
2347                 }
2348
2349                 if ((bif->bif_flags & IFBIF_STP) &&
2350                     bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) {
2351                         BRIDGE_UNLOCK(sc);
2352                         return (m);
2353                 }
2354
2355                 /*
2356                  * Make a deep copy of the packet and enqueue the copy
2357                  * for bridge processing; return the original packet for
2358                  * local processing.
2359                  */
2360                 mc = m_dup(m, M_NOWAIT);
2361                 if (mc == NULL) {
2362                         BRIDGE_UNLOCK(sc);
2363                         return (m);
2364                 }
2365
2366                 /* Perform the bridge forwarding function with the copy. */
2367                 bridge_forward(sc, bif, mc);
2368
2369                 /*
2370                  * Reinject the mbuf as arriving on the bridge so we have a
2371                  * chance at claiming multicast packets. We can not loop back
2372                  * here from ether_input as a bridge is never a member of a
2373                  * bridge.
2374                  */
2375                 KASSERT(bifp->if_bridge == NULL,
2376                     ("loop created in bridge_input"));
2377                 mc2 = m_dup(m, M_NOWAIT);
2378                 if (mc2 != NULL) {
2379                         /* Keep the layer3 header aligned */
2380                         int i = min(mc2->m_pkthdr.len, max_protohdr);
2381                         mc2 = m_copyup(mc2, i, ETHER_ALIGN);
2382                 }
2383                 if (mc2 != NULL) {
2384                         mc2->m_pkthdr.rcvif = bifp;
2385                         (*bifp->if_input)(bifp, mc2);
2386                 }
2387
2388                 /* Return the original packet for local processing. */
2389                 return (m);
2390         }
2391
2392         if ((bif->bif_flags & IFBIF_STP) &&
2393             bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) {
2394                 BRIDGE_UNLOCK(sc);
2395                 return (m);
2396         }
2397
2398 #if (defined(INET) || defined(INET6))
2399 #   define OR_CARP_CHECK_WE_ARE_DST(iface) \
2400         || ((iface)->if_carp \
2401             && (*carp_forus_p)((iface), eh->ether_dhost))
2402 #   define OR_CARP_CHECK_WE_ARE_SRC(iface) \
2403         || ((iface)->if_carp \
2404             && (*carp_forus_p)((iface), eh->ether_shost))
2405 #else
2406 #   define OR_CARP_CHECK_WE_ARE_DST(iface)
2407 #   define OR_CARP_CHECK_WE_ARE_SRC(iface)
2408 #endif
2409
2410 #ifdef INET6
2411 #   define OR_PFIL_HOOKED_INET6 \
2412         || PFIL_HOOKED_IN(V_inet6_pfil_head)
2413 #else
2414 #   define OR_PFIL_HOOKED_INET6
2415 #endif
2416
2417 #define GRAB_OUR_PACKETS(iface) \
2418         if ((iface)->if_type == IFT_GIF) \
2419                 continue; \
2420         /* It is destined for us. */ \
2421         if (memcmp(IF_LLADDR((iface)), eh->ether_dhost,  ETHER_ADDR_LEN) == 0 \
2422             OR_CARP_CHECK_WE_ARE_DST((iface))                           \
2423             ) {                                                         \
2424                 if ((iface)->if_type == IFT_BRIDGE) {                   \
2425                         ETHER_BPF_MTAP(iface, m);                       \
2426                         if_inc_counter(iface, IFCOUNTER_IPACKETS, 1);                           \
2427                         if_inc_counter(iface, IFCOUNTER_IBYTES, m->m_pkthdr.len);               \
2428                         /* Filter on the physical interface. */         \
2429                         if (V_pfil_local_phys &&                        \
2430                             (PFIL_HOOKED_IN(V_inet_pfil_head)           \
2431                              OR_PFIL_HOOKED_INET6)) {                   \
2432                                 if (bridge_pfil(&m, NULL, ifp,          \
2433                                     PFIL_IN) != 0 || m == NULL) {       \
2434                                         BRIDGE_UNLOCK(sc);              \
2435                                         return (NULL);                  \
2436                                 }                                       \
2437                                 eh = mtod(m, struct ether_header *);    \
2438                         }                                               \
2439                 }                                                       \
2440                 if (bif->bif_flags & IFBIF_LEARNING) {                  \
2441                         error = bridge_rtupdate(sc, eh->ether_shost,    \
2442                             vlan, bif, 0, IFBAF_DYNAMIC);               \
2443                         if (error && bif->bif_addrmax) {                \
2444                                 BRIDGE_UNLOCK(sc);                      \
2445                                 m_freem(m);                             \
2446                                 return (NULL);                          \
2447                         }                                               \
2448                 }                                                       \
2449                 m->m_pkthdr.rcvif = iface;                              \
2450                 BRIDGE_UNLOCK(sc);                                      \
2451                 return (m);                                             \
2452         }                                                               \
2453                                                                         \
2454         /* We just received a packet that we sent out. */               \
2455         if (memcmp(IF_LLADDR((iface)), eh->ether_shost, ETHER_ADDR_LEN) == 0 \
2456             OR_CARP_CHECK_WE_ARE_SRC((iface))                   \
2457             ) {                                                         \
2458                 BRIDGE_UNLOCK(sc);                                      \
2459                 m_freem(m);                                             \
2460                 return (NULL);                                          \
2461         }
2462
2463         /*
2464          * Unicast.  Make sure it's not for the bridge.
2465          */
2466         do { GRAB_OUR_PACKETS(bifp) } while (0);
2467
2468         /*
2469          * Give a chance for ifp at first priority. This will help when the
2470          * packet comes through the interface like VLAN's with the same MACs
2471          * on several interfaces from the same bridge. This also will save
2472          * some CPU cycles in case the destination interface and the input
2473          * interface (eq ifp) are the same.
2474          */
2475         do { GRAB_OUR_PACKETS(ifp) } while (0);
2476
2477         /* Now check the all bridge members. */
2478         LIST_FOREACH(bif2, &sc->sc_iflist, bif_next) {
2479                 GRAB_OUR_PACKETS(bif2->bif_ifp)
2480         }
2481
2482 #undef OR_CARP_CHECK_WE_ARE_DST
2483 #undef OR_CARP_CHECK_WE_ARE_SRC
2484 #undef OR_PFIL_HOOKED_INET6
2485 #undef GRAB_OUR_PACKETS
2486
2487         /* Perform the bridge forwarding function. */
2488         bridge_forward(sc, bif, m);
2489
2490         return (NULL);
2491 }
2492
2493 /*
2494  * bridge_broadcast:
2495  *
2496  *      Send a frame to all interfaces that are members of
2497  *      the bridge, except for the one on which the packet
2498  *      arrived.
2499  *
2500  *      NOTE: Releases the lock on return.
2501  */
2502 static void
2503 bridge_broadcast(struct bridge_softc *sc, struct ifnet *src_if,
2504     struct mbuf *m, int runfilt)
2505 {
2506         struct bridge_iflist *dbif, *sbif;
2507         struct mbuf *mc;
2508         struct ifnet *dst_if;
2509         int error = 0, used = 0, i;
2510
2511         sbif = bridge_lookup_member_if(sc, src_if);
2512
2513         BRIDGE_LOCK2REF(sc, error);
2514         if (error) {
2515                 m_freem(m);
2516                 return;
2517         }
2518
2519         /* Filter on the bridge interface before broadcasting */
2520         if (runfilt && (PFIL_HOOKED_OUT(V_inet_pfil_head)
2521 #ifdef INET6
2522             || PFIL_HOOKED_OUT(V_inet6_pfil_head)
2523 #endif
2524             )) {
2525                 if (bridge_pfil(&m, sc->sc_ifp, NULL, PFIL_OUT) != 0)
2526                         goto out;
2527                 if (m == NULL)
2528                         goto out;
2529         }
2530
2531         LIST_FOREACH(dbif, &sc->sc_iflist, bif_next) {
2532                 dst_if = dbif->bif_ifp;
2533                 if (dst_if == src_if)
2534                         continue;
2535
2536                 /* Private segments can not talk to each other */
2537                 if (sbif && (sbif->bif_flags & dbif->bif_flags & IFBIF_PRIVATE))
2538                         continue;
2539
2540                 if ((dbif->bif_flags & IFBIF_STP) &&
2541                     dbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
2542                         continue;
2543
2544                 if ((dbif->bif_flags & IFBIF_DISCOVER) == 0 &&
2545                     (m->m_flags & (M_BCAST|M_MCAST)) == 0)
2546                         continue;
2547
2548                 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
2549                         continue;
2550
2551                 if (LIST_NEXT(dbif, bif_next) == NULL) {
2552                         mc = m;
2553                         used = 1;
2554                 } else {
2555                         mc = m_dup(m, M_NOWAIT);
2556                         if (mc == NULL) {
2557                                 if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
2558                                 continue;
2559                         }
2560                 }
2561
2562                 /*
2563                  * Filter on the output interface. Pass a NULL bridge interface
2564                  * pointer so we do not redundantly filter on the bridge for
2565                  * each interface we broadcast on.
2566                  */
2567                 if (runfilt && (PFIL_HOOKED_OUT(V_inet_pfil_head)
2568 #ifdef INET6
2569                     || PFIL_HOOKED_OUT(V_inet6_pfil_head)
2570 #endif
2571                     )) {
2572                         if (used == 0) {
2573                                 /* Keep the layer3 header aligned */
2574                                 i = min(mc->m_pkthdr.len, max_protohdr);
2575                                 mc = m_copyup(mc, i, ETHER_ALIGN);
2576                                 if (mc == NULL) {
2577                                         if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
2578                                         continue;
2579                                 }
2580                         }
2581                         if (bridge_pfil(&mc, NULL, dst_if, PFIL_OUT) != 0)
2582                                 continue;
2583                         if (mc == NULL)
2584                                 continue;
2585                 }
2586
2587                 bridge_enqueue(sc, dst_if, mc);
2588         }
2589         if (used == 0)
2590                 m_freem(m);
2591
2592 out:
2593         BRIDGE_UNREF(sc);
2594 }
2595
2596 /*
2597  * bridge_span:
2598  *
2599  *      Duplicate a packet out one or more interfaces that are in span mode,
2600  *      the original mbuf is unmodified.
2601  */
2602 static void
2603 bridge_span(struct bridge_softc *sc, struct mbuf *m)
2604 {
2605         struct bridge_iflist *bif;
2606         struct ifnet *dst_if;
2607         struct mbuf *mc;
2608
2609         if (LIST_EMPTY(&sc->sc_spanlist))
2610                 return;
2611
2612         LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) {
2613                 dst_if = bif->bif_ifp;
2614
2615                 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
2616                         continue;
2617
2618                 mc = m_copypacket(m, M_NOWAIT);
2619                 if (mc == NULL) {
2620                         if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
2621                         continue;
2622                 }
2623
2624                 bridge_enqueue(sc, dst_if, mc);
2625         }
2626 }
2627
2628 /*
2629  * bridge_rtupdate:
2630  *
2631  *      Add a bridge routing entry.
2632  */
2633 static int
2634 bridge_rtupdate(struct bridge_softc *sc, const uint8_t *dst, uint16_t vlan,
2635     struct bridge_iflist *bif, int setflags, uint8_t flags)
2636 {
2637         struct bridge_rtnode *brt;
2638         int error;
2639
2640         BRIDGE_LOCK_ASSERT(sc);
2641
2642         /* Check the source address is valid and not multicast. */
2643         if (ETHER_IS_MULTICAST(dst) ||
2644             (dst[0] == 0 && dst[1] == 0 && dst[2] == 0 &&
2645              dst[3] == 0 && dst[4] == 0 && dst[5] == 0) != 0)
2646                 return (EINVAL);
2647
2648         /* 802.1p frames map to vlan 1 */
2649         if (vlan == 0)
2650                 vlan = 1;
2651
2652         /*
2653          * A route for this destination might already exist.  If so,
2654          * update it, otherwise create a new one.
2655          */
2656         if ((brt = bridge_rtnode_lookup(sc, dst, vlan)) == NULL) {
2657                 if (sc->sc_brtcnt >= sc->sc_brtmax) {
2658                         sc->sc_brtexceeded++;
2659                         return (ENOSPC);
2660                 }
2661                 /* Check per interface address limits (if enabled) */
2662                 if (bif->bif_addrmax && bif->bif_addrcnt >= bif->bif_addrmax) {
2663                         bif->bif_addrexceeded++;
2664                         return (ENOSPC);
2665                 }
2666
2667                 /*
2668                  * Allocate a new bridge forwarding node, and
2669                  * initialize the expiration time and Ethernet
2670                  * address.
2671                  */
2672                 brt = uma_zalloc(bridge_rtnode_zone, M_NOWAIT | M_ZERO);
2673                 if (brt == NULL)
2674                         return (ENOMEM);
2675
2676                 if (bif->bif_flags & IFBIF_STICKY)
2677                         brt->brt_flags = IFBAF_STICKY;
2678                 else
2679                         brt->brt_flags = IFBAF_DYNAMIC;
2680
2681                 memcpy(brt->brt_addr, dst, ETHER_ADDR_LEN);
2682                 brt->brt_vlan = vlan;
2683
2684                 if ((error = bridge_rtnode_insert(sc, brt)) != 0) {
2685                         uma_zfree(bridge_rtnode_zone, brt);
2686                         return (error);
2687                 }
2688                 brt->brt_dst = bif;
2689                 bif->bif_addrcnt++;
2690         }
2691
2692         if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC &&
2693             brt->brt_dst != bif) {
2694                 brt->brt_dst->bif_addrcnt--;
2695                 brt->brt_dst = bif;
2696                 brt->brt_dst->bif_addrcnt++;
2697         }
2698
2699         if ((flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
2700                 brt->brt_expire = time_uptime + sc->sc_brttimeout;
2701         if (setflags)
2702                 brt->brt_flags = flags;
2703
2704         return (0);
2705 }
2706
2707 /*
2708  * bridge_rtlookup:
2709  *
2710  *      Lookup the destination interface for an address.
2711  */
2712 static struct ifnet *
2713 bridge_rtlookup(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan)
2714 {
2715         struct bridge_rtnode *brt;
2716
2717         BRIDGE_LOCK_ASSERT(sc);
2718
2719         if ((brt = bridge_rtnode_lookup(sc, addr, vlan)) == NULL)
2720                 return (NULL);
2721
2722         return (brt->brt_ifp);
2723 }
2724
2725 /*
2726  * bridge_rttrim:
2727  *
2728  *      Trim the routine table so that we have a number
2729  *      of routing entries less than or equal to the
2730  *      maximum number.
2731  */
2732 static void
2733 bridge_rttrim(struct bridge_softc *sc)
2734 {
2735         struct bridge_rtnode *brt, *nbrt;
2736
2737         BRIDGE_LOCK_ASSERT(sc);
2738
2739         /* Make sure we actually need to do this. */
2740         if (sc->sc_brtcnt <= sc->sc_brtmax)
2741                 return;
2742
2743         /* Force an aging cycle; this might trim enough addresses. */
2744         bridge_rtage(sc);
2745         if (sc->sc_brtcnt <= sc->sc_brtmax)
2746                 return;
2747
2748         LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
2749                 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) {
2750                         bridge_rtnode_destroy(sc, brt);
2751                         if (sc->sc_brtcnt <= sc->sc_brtmax)
2752                                 return;
2753                 }
2754         }
2755 }
2756
2757 /*
2758  * bridge_timer:
2759  *
2760  *      Aging timer for the bridge.
2761  */
2762 static void
2763 bridge_timer(void *arg)
2764 {
2765         struct bridge_softc *sc = arg;
2766
2767         BRIDGE_LOCK_ASSERT(sc);
2768
2769         bridge_rtage(sc);
2770
2771         if (sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING)
2772                 callout_reset(&sc->sc_brcallout,
2773                     bridge_rtable_prune_period * hz, bridge_timer, sc);
2774 }
2775
2776 /*
2777  * bridge_rtage:
2778  *
2779  *      Perform an aging cycle.
2780  */
2781 static void
2782 bridge_rtage(struct bridge_softc *sc)
2783 {
2784         struct bridge_rtnode *brt, *nbrt;
2785
2786         BRIDGE_LOCK_ASSERT(sc);
2787
2788         LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
2789                 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) {
2790                         if (time_uptime >= brt->brt_expire)
2791                                 bridge_rtnode_destroy(sc, brt);
2792                 }
2793         }
2794 }
2795
2796 /*
2797  * bridge_rtflush:
2798  *
2799  *      Remove all dynamic addresses from the bridge.
2800  */
2801 static void
2802 bridge_rtflush(struct bridge_softc *sc, int full)
2803 {
2804         struct bridge_rtnode *brt, *nbrt;
2805
2806         BRIDGE_LOCK_ASSERT(sc);
2807
2808         LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
2809                 if (full || (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
2810                         bridge_rtnode_destroy(sc, brt);
2811         }
2812 }
2813
2814 /*
2815  * bridge_rtdaddr:
2816  *
2817  *      Remove an address from the table.
2818  */
2819 static int
2820 bridge_rtdaddr(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan)
2821 {
2822         struct bridge_rtnode *brt;
2823         int found = 0;
2824
2825         BRIDGE_LOCK_ASSERT(sc);
2826
2827         /*
2828          * If vlan is zero then we want to delete for all vlans so the lookup
2829          * may return more than one.
2830          */
2831         while ((brt = bridge_rtnode_lookup(sc, addr, vlan)) != NULL) {
2832                 bridge_rtnode_destroy(sc, brt);
2833                 found = 1;
2834         }
2835
2836         return (found ? 0 : ENOENT);
2837 }
2838
2839 /*
2840  * bridge_rtdelete:
2841  *
2842  *      Delete routes to a speicifc member interface.
2843  */
2844 static void
2845 bridge_rtdelete(struct bridge_softc *sc, struct ifnet *ifp, int full)
2846 {
2847         struct bridge_rtnode *brt, *nbrt;
2848
2849         BRIDGE_LOCK_ASSERT(sc);
2850
2851         LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
2852                 if (brt->brt_ifp == ifp && (full ||
2853                             (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC))
2854                         bridge_rtnode_destroy(sc, brt);
2855         }
2856 }
2857
2858 /*
2859  * bridge_rtable_init:
2860  *
2861  *      Initialize the route table for this bridge.
2862  */
2863 static void
2864 bridge_rtable_init(struct bridge_softc *sc)
2865 {
2866         int i;
2867
2868         sc->sc_rthash = malloc(sizeof(*sc->sc_rthash) * BRIDGE_RTHASH_SIZE,
2869             M_DEVBUF, M_WAITOK);
2870
2871         for (i = 0; i < BRIDGE_RTHASH_SIZE; i++)
2872                 LIST_INIT(&sc->sc_rthash[i]);
2873
2874         sc->sc_rthash_key = arc4random();
2875         LIST_INIT(&sc->sc_rtlist);
2876 }
2877
2878 /*
2879  * bridge_rtable_fini:
2880  *
2881  *      Deconstruct the route table for this bridge.
2882  */
2883 static void
2884 bridge_rtable_fini(struct bridge_softc *sc)
2885 {
2886
2887         KASSERT(sc->sc_brtcnt == 0,
2888             ("%s: %d bridge routes referenced", __func__, sc->sc_brtcnt));
2889         free(sc->sc_rthash, M_DEVBUF);
2890 }
2891
2892 /*
2893  * The following hash function is adapted from "Hash Functions" by Bob Jenkins
2894  * ("Algorithm Alley", Dr. Dobbs Journal, September 1997).
2895  */
2896 #define mix(a, b, c)                                                    \
2897 do {                                                                    \
2898         a -= b; a -= c; a ^= (c >> 13);                                 \
2899         b -= c; b -= a; b ^= (a << 8);                                  \
2900         c -= a; c -= b; c ^= (b >> 13);                                 \
2901         a -= b; a -= c; a ^= (c >> 12);                                 \
2902         b -= c; b -= a; b ^= (a << 16);                                 \
2903         c -= a; c -= b; c ^= (b >> 5);                                  \
2904         a -= b; a -= c; a ^= (c >> 3);                                  \
2905         b -= c; b -= a; b ^= (a << 10);                                 \
2906         c -= a; c -= b; c ^= (b >> 15);                                 \
2907 } while (/*CONSTCOND*/0)
2908
2909 static __inline uint32_t
2910 bridge_rthash(struct bridge_softc *sc, const uint8_t *addr)
2911 {
2912         uint32_t a = 0x9e3779b9, b = 0x9e3779b9, c = sc->sc_rthash_key;
2913
2914         b += addr[5] << 8;
2915         b += addr[4];
2916         a += addr[3] << 24;
2917         a += addr[2] << 16;
2918         a += addr[1] << 8;
2919         a += addr[0];
2920
2921         mix(a, b, c);
2922
2923         return (c & BRIDGE_RTHASH_MASK);
2924 }
2925
2926 #undef mix
2927
2928 static int
2929 bridge_rtnode_addr_cmp(const uint8_t *a, const uint8_t *b)
2930 {
2931         int i, d;
2932
2933         for (i = 0, d = 0; i < ETHER_ADDR_LEN && d == 0; i++) {
2934                 d = ((int)a[i]) - ((int)b[i]);
2935         }
2936
2937         return (d);
2938 }
2939
2940 /*
2941  * bridge_rtnode_lookup:
2942  *
2943  *      Look up a bridge route node for the specified destination. Compare the
2944  *      vlan id or if zero then just return the first match.
2945  */
2946 static struct bridge_rtnode *
2947 bridge_rtnode_lookup(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan)
2948 {
2949         struct bridge_rtnode *brt;
2950         uint32_t hash;
2951         int dir;
2952
2953         BRIDGE_LOCK_ASSERT(sc);
2954
2955         hash = bridge_rthash(sc, addr);
2956         LIST_FOREACH(brt, &sc->sc_rthash[hash], brt_hash) {
2957                 dir = bridge_rtnode_addr_cmp(addr, brt->brt_addr);
2958                 if (dir == 0 && (brt->brt_vlan == vlan || vlan == 0))
2959                         return (brt);
2960                 if (dir > 0)
2961                         return (NULL);
2962         }
2963
2964         return (NULL);
2965 }
2966
2967 /*
2968  * bridge_rtnode_insert:
2969  *
2970  *      Insert the specified bridge node into the route table.  We
2971  *      assume the entry is not already in the table.
2972  */
2973 static int
2974 bridge_rtnode_insert(struct bridge_softc *sc, struct bridge_rtnode *brt)
2975 {
2976         struct bridge_rtnode *lbrt;
2977         uint32_t hash;
2978         int dir;
2979
2980         BRIDGE_LOCK_ASSERT(sc);
2981
2982         hash = bridge_rthash(sc, brt->brt_addr);
2983
2984         lbrt = LIST_FIRST(&sc->sc_rthash[hash]);
2985         if (lbrt == NULL) {
2986                 LIST_INSERT_HEAD(&sc->sc_rthash[hash], brt, brt_hash);
2987                 goto out;
2988         }
2989
2990         do {
2991                 dir = bridge_rtnode_addr_cmp(brt->brt_addr, lbrt->brt_addr);
2992                 if (dir == 0 && brt->brt_vlan == lbrt->brt_vlan)
2993                         return (EEXIST);
2994                 if (dir > 0) {
2995                         LIST_INSERT_BEFORE(lbrt, brt, brt_hash);
2996                         goto out;
2997                 }
2998                 if (LIST_NEXT(lbrt, brt_hash) == NULL) {
2999                         LIST_INSERT_AFTER(lbrt, brt, brt_hash);
3000                         goto out;
3001                 }
3002                 lbrt = LIST_NEXT(lbrt, brt_hash);
3003         } while (lbrt != NULL);
3004
3005 #ifdef DIAGNOSTIC
3006         panic("bridge_rtnode_insert: impossible");
3007 #endif
3008
3009 out:
3010         LIST_INSERT_HEAD(&sc->sc_rtlist, brt, brt_list);
3011         sc->sc_brtcnt++;
3012
3013         return (0);
3014 }
3015
3016 /*
3017  * bridge_rtnode_destroy:
3018  *
3019  *      Destroy a bridge rtnode.
3020  */
3021 static void
3022 bridge_rtnode_destroy(struct bridge_softc *sc, struct bridge_rtnode *brt)
3023 {
3024         BRIDGE_LOCK_ASSERT(sc);
3025
3026         LIST_REMOVE(brt, brt_hash);
3027
3028         LIST_REMOVE(brt, brt_list);
3029         sc->sc_brtcnt--;
3030         brt->brt_dst->bif_addrcnt--;
3031         uma_zfree(bridge_rtnode_zone, brt);
3032 }
3033
3034 /*
3035  * bridge_rtable_expire:
3036  *
3037  *      Set the expiry time for all routes on an interface.
3038  */
3039 static void
3040 bridge_rtable_expire(struct ifnet *ifp, int age)
3041 {
3042         struct bridge_softc *sc = ifp->if_bridge;
3043         struct bridge_rtnode *brt;
3044
3045         BRIDGE_LOCK(sc);
3046
3047         /*
3048          * If the age is zero then flush, otherwise set all the expiry times to
3049          * age for the interface
3050          */
3051         if (age == 0)
3052                 bridge_rtdelete(sc, ifp, IFBF_FLUSHDYN);
3053         else {
3054                 LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) {
3055                         /* Cap the expiry time to 'age' */
3056                         if (brt->brt_ifp == ifp &&
3057                             brt->brt_expire > time_uptime + age &&
3058                             (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
3059                                 brt->brt_expire = time_uptime + age;
3060                 }
3061         }
3062         BRIDGE_UNLOCK(sc);
3063 }
3064
3065 /*
3066  * bridge_state_change:
3067  *
3068  *      Callback from the bridgestp code when a port changes states.
3069  */
3070 static void
3071 bridge_state_change(struct ifnet *ifp, int state)
3072 {
3073         struct bridge_softc *sc = ifp->if_bridge;
3074         static const char *stpstates[] = {
3075                 "disabled",
3076                 "listening",
3077                 "learning",
3078                 "forwarding",
3079                 "blocking",
3080                 "discarding"
3081         };
3082
3083         CURVNET_SET(ifp->if_vnet);
3084         if (V_log_stp)
3085                 log(LOG_NOTICE, "%s: state changed to %s on %s\n",
3086                     sc->sc_ifp->if_xname, stpstates[state], ifp->if_xname);
3087         CURVNET_RESTORE();
3088 }
3089
3090 /*
3091  * Send bridge packets through pfil if they are one of the types pfil can deal
3092  * with, or if they are ARP or REVARP.  (pfil will pass ARP and REVARP without
3093  * question.) If *bifp or *ifp are NULL then packet filtering is skipped for
3094  * that interface.
3095  */
3096 static int
3097 bridge_pfil(struct mbuf **mp, struct ifnet *bifp, struct ifnet *ifp, int dir)
3098 {
3099         int snap, error, i, hlen;
3100         struct ether_header *eh1, eh2;
3101         struct ip *ip;
3102         struct llc llc1;
3103         u_int16_t ether_type;
3104         pfil_return_t rv;
3105
3106         snap = 0;
3107         error = -1;     /* Default error if not error == 0 */
3108
3109 #if 0
3110         /* we may return with the IP fields swapped, ensure its not shared */
3111         KASSERT(M_WRITABLE(*mp), ("%s: modifying a shared mbuf", __func__));
3112 #endif
3113
3114         if (V_pfil_bridge == 0 && V_pfil_member == 0 && V_pfil_ipfw == 0)
3115                 return (0); /* filtering is disabled */
3116
3117         i = min((*mp)->m_pkthdr.len, max_protohdr);
3118         if ((*mp)->m_len < i) {
3119             *mp = m_pullup(*mp, i);
3120             if (*mp == NULL) {
3121                 printf("%s: m_pullup failed\n", __func__);
3122                 return (-1);
3123             }
3124         }
3125
3126         eh1 = mtod(*mp, struct ether_header *);
3127         ether_type = ntohs(eh1->ether_type);
3128
3129         /*
3130          * Check for SNAP/LLC.
3131          */
3132         if (ether_type < ETHERMTU) {
3133                 struct llc *llc2 = (struct llc *)(eh1 + 1);
3134
3135                 if ((*mp)->m_len >= ETHER_HDR_LEN + 8 &&
3136                     llc2->llc_dsap == LLC_SNAP_LSAP &&
3137                     llc2->llc_ssap == LLC_SNAP_LSAP &&
3138                     llc2->llc_control == LLC_UI) {
3139                         ether_type = htons(llc2->llc_un.type_snap.ether_type);
3140                         snap = 1;
3141                 }
3142         }
3143
3144         /*
3145          * If we're trying to filter bridge traffic, don't look at anything
3146          * other than IP and ARP traffic.  If the filter doesn't understand
3147          * IPv6, don't allow IPv6 through the bridge either.  This is lame
3148          * since if we really wanted, say, an AppleTalk filter, we are hosed,
3149          * but of course we don't have an AppleTalk filter to begin with.
3150          * (Note that since pfil doesn't understand ARP it will pass *ALL*
3151          * ARP traffic.)
3152          */
3153         switch (ether_type) {
3154                 case ETHERTYPE_ARP:
3155                 case ETHERTYPE_REVARP:
3156                         if (V_pfil_ipfw_arp == 0)
3157                                 return (0); /* Automatically pass */
3158                         break;
3159
3160                 case ETHERTYPE_IP:
3161 #ifdef INET6
3162                 case ETHERTYPE_IPV6:
3163 #endif /* INET6 */
3164                         break;
3165                 default:
3166                         /*
3167                          * Check to see if the user wants to pass non-ip
3168                          * packets, these will not be checked by pfil(9) and
3169                          * passed unconditionally so the default is to drop.
3170                          */
3171                         if (V_pfil_onlyip)
3172                                 goto bad;
3173         }
3174
3175         /* Run the packet through pfil before stripping link headers */
3176         if (PFIL_HOOKED_OUT(V_link_pfil_head) && V_pfil_ipfw != 0 &&
3177             dir == PFIL_OUT && ifp != NULL) {
3178                 switch (pfil_run_hooks(V_link_pfil_head, mp, ifp, dir, NULL)) {
3179                 case PFIL_DROPPED:
3180                         return (EPERM);
3181                 case PFIL_CONSUMED:
3182                         return (0);
3183                 }
3184         }
3185
3186         /* Strip off the Ethernet header and keep a copy. */
3187         m_copydata(*mp, 0, ETHER_HDR_LEN, (caddr_t) &eh2);
3188         m_adj(*mp, ETHER_HDR_LEN);
3189
3190         /* Strip off snap header, if present */
3191         if (snap) {
3192                 m_copydata(*mp, 0, sizeof(struct llc), (caddr_t) &llc1);
3193                 m_adj(*mp, sizeof(struct llc));
3194         }
3195
3196         /*
3197          * Check the IP header for alignment and errors
3198          */
3199         if (dir == PFIL_IN) {
3200                 switch (ether_type) {
3201                         case ETHERTYPE_IP:
3202                                 error = bridge_ip_checkbasic(mp);
3203                                 break;
3204 #ifdef INET6
3205                         case ETHERTYPE_IPV6:
3206                                 error = bridge_ip6_checkbasic(mp);
3207                                 break;
3208 #endif /* INET6 */
3209                         default:
3210                                 error = 0;
3211                 }
3212                 if (error)
3213                         goto bad;
3214         }
3215
3216         error = 0;
3217
3218         /*
3219          * Run the packet through pfil
3220          */
3221         rv = PFIL_PASS;
3222         switch (ether_type) {
3223         case ETHERTYPE_IP:
3224                 /*
3225                  * Run pfil on the member interface and the bridge, both can
3226                  * be skipped by clearing pfil_member or pfil_bridge.
3227                  *
3228                  * Keep the order:
3229                  *   in_if -> bridge_if -> out_if
3230                  */
3231                 if (V_pfil_bridge && dir == PFIL_OUT && bifp != NULL && (rv =
3232                     pfil_run_hooks(V_inet_pfil_head, mp, bifp, dir, NULL)) !=
3233                     PFIL_PASS)
3234                         break;
3235
3236                 if (V_pfil_member && ifp != NULL && (rv =
3237                     pfil_run_hooks(V_inet_pfil_head, mp, ifp, dir, NULL)) !=
3238                     PFIL_PASS)
3239                         break;
3240
3241                 if (V_pfil_bridge && dir == PFIL_IN && bifp != NULL && (rv =
3242                     pfil_run_hooks(V_inet_pfil_head, mp, bifp, dir, NULL)) !=
3243                     PFIL_PASS)
3244                         break;
3245
3246                 /* check if we need to fragment the packet */
3247                 /* bridge_fragment generates a mbuf chain of packets */
3248                 /* that already include eth headers */
3249                 if (V_pfil_member && ifp != NULL && dir == PFIL_OUT) {
3250                         i = (*mp)->m_pkthdr.len;
3251                         if (i > ifp->if_mtu) {
3252                                 error = bridge_fragment(ifp, mp, &eh2, snap,
3253                                             &llc1);
3254                                 return (error);
3255                         }
3256                 }
3257
3258                 /* Recalculate the ip checksum. */
3259                 ip = mtod(*mp, struct ip *);
3260                 hlen = ip->ip_hl << 2;
3261                 if (hlen < sizeof(struct ip))
3262                         goto bad;
3263                 if (hlen > (*mp)->m_len) {
3264                         if ((*mp = m_pullup(*mp, hlen)) == NULL)
3265                                 goto bad;
3266                         ip = mtod(*mp, struct ip *);
3267                         if (ip == NULL)
3268                                 goto bad;
3269                 }
3270                 ip->ip_sum = 0;
3271                 if (hlen == sizeof(struct ip))
3272                         ip->ip_sum = in_cksum_hdr(ip);
3273                 else
3274                         ip->ip_sum = in_cksum(*mp, hlen);
3275
3276                 break;
3277 #ifdef INET6
3278         case ETHERTYPE_IPV6:
3279                 if (V_pfil_bridge && dir == PFIL_OUT && bifp != NULL && (rv =
3280                     pfil_run_hooks(V_inet6_pfil_head, mp, bifp, dir, NULL)) !=
3281                     PFIL_PASS)
3282                         break;
3283
3284                 if (V_pfil_member && ifp != NULL && (rv =
3285                     pfil_run_hooks(V_inet6_pfil_head, mp, ifp, dir, NULL)) !=
3286                     PFIL_PASS)
3287                         break;
3288
3289                 if (V_pfil_bridge && dir == PFIL_IN && bifp != NULL && (rv =
3290                     pfil_run_hooks(V_inet6_pfil_head, mp, bifp, dir, NULL)) !=
3291                     PFIL_PASS)
3292                         break;
3293                 break;
3294 #endif
3295         }
3296
3297         switch (rv) {
3298         case PFIL_CONSUMED:
3299                 return (0);
3300         case PFIL_DROPPED:
3301                 return (EPERM);
3302         default:
3303                 break;
3304         }
3305
3306         error = -1;
3307
3308         /*
3309          * Finally, put everything back the way it was and return
3310          */
3311         if (snap) {
3312                 M_PREPEND(*mp, sizeof(struct llc), M_NOWAIT);
3313                 if (*mp == NULL)
3314                         return (error);
3315                 bcopy(&llc1, mtod(*mp, caddr_t), sizeof(struct llc));
3316         }
3317
3318         M_PREPEND(*mp, ETHER_HDR_LEN, M_NOWAIT);
3319         if (*mp == NULL)
3320                 return (error);
3321         bcopy(&eh2, mtod(*mp, caddr_t), ETHER_HDR_LEN);
3322
3323         return (0);
3324
3325 bad:
3326         m_freem(*mp);
3327         *mp = NULL;
3328         return (error);
3329 }
3330
3331 /*
3332  * Perform basic checks on header size since
3333  * pfil assumes ip_input has already processed
3334  * it for it.  Cut-and-pasted from ip_input.c.
3335  * Given how simple the IPv6 version is,
3336  * does the IPv4 version really need to be
3337  * this complicated?
3338  *
3339  * XXX Should we update ipstat here, or not?
3340  * XXX Right now we update ipstat but not
3341  * XXX csum_counter.
3342  */
3343 static int
3344 bridge_ip_checkbasic(struct mbuf **mp)
3345 {
3346         struct mbuf *m = *mp;
3347         struct ip *ip;
3348         int len, hlen;
3349         u_short sum;
3350
3351         if (*mp == NULL)
3352                 return (-1);
3353
3354         if (IP_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) {
3355                 if ((m = m_copyup(m, sizeof(struct ip),
3356                         (max_linkhdr + 3) & ~3)) == NULL) {
3357                         /* XXXJRT new stat, please */
3358                         KMOD_IPSTAT_INC(ips_toosmall);
3359                         goto bad;
3360                 }
3361         } else if (__predict_false(m->m_len < sizeof (struct ip))) {
3362                 if ((m = m_pullup(m, sizeof (struct ip))) == NULL) {
3363                         KMOD_IPSTAT_INC(ips_toosmall);
3364                         goto bad;
3365                 }
3366         }
3367         ip = mtod(m, struct ip *);
3368         if (ip == NULL) goto bad;
3369
3370         if (ip->ip_v != IPVERSION) {
3371                 KMOD_IPSTAT_INC(ips_badvers);
3372                 goto bad;
3373         }
3374         hlen = ip->ip_hl << 2;
3375         if (hlen < sizeof(struct ip)) { /* minimum header length */
3376                 KMOD_IPSTAT_INC(ips_badhlen);
3377                 goto bad;
3378         }
3379         if (hlen > m->m_len) {
3380                 if ((m = m_pullup(m, hlen)) == NULL) {
3381                         KMOD_IPSTAT_INC(ips_badhlen);
3382                         goto bad;
3383                 }
3384                 ip = mtod(m, struct ip *);
3385                 if (ip == NULL) goto bad;
3386         }
3387
3388         if (m->m_pkthdr.csum_flags & CSUM_IP_CHECKED) {
3389                 sum = !(m->m_pkthdr.csum_flags & CSUM_IP_VALID);
3390         } else {
3391                 if (hlen == sizeof(struct ip)) {
3392                         sum = in_cksum_hdr(ip);
3393                 } else {
3394                         sum = in_cksum(m, hlen);
3395                 }
3396         }
3397         if (sum) {
3398                 KMOD_IPSTAT_INC(ips_badsum);
3399                 goto bad;
3400         }
3401
3402         /* Retrieve the packet length. */
3403         len = ntohs(ip->ip_len);
3404
3405         /*
3406          * Check for additional length bogosity
3407          */
3408         if (len < hlen) {
3409                 KMOD_IPSTAT_INC(ips_badlen);
3410                 goto bad;
3411         }
3412
3413         /*
3414          * Check that the amount of data in the buffers
3415          * is as at least much as the IP header would have us expect.
3416          * Drop packet if shorter than we expect.
3417          */
3418         if (m->m_pkthdr.len < len) {
3419                 KMOD_IPSTAT_INC(ips_tooshort);
3420                 goto bad;
3421         }
3422
3423         /* Checks out, proceed */
3424         *mp = m;
3425         return (0);
3426
3427 bad:
3428         *mp = m;
3429         return (-1);
3430 }
3431
3432 #ifdef INET6
3433 /*
3434  * Same as above, but for IPv6.
3435  * Cut-and-pasted from ip6_input.c.
3436  * XXX Should we update ip6stat, or not?
3437  */
3438 static int
3439 bridge_ip6_checkbasic(struct mbuf **mp)
3440 {
3441         struct mbuf *m = *mp;
3442         struct ip6_hdr *ip6;
3443
3444         /*
3445          * If the IPv6 header is not aligned, slurp it up into a new
3446          * mbuf with space for link headers, in the event we forward
3447          * it.  Otherwise, if it is aligned, make sure the entire base
3448          * IPv6 header is in the first mbuf of the chain.
3449          */
3450         if (IP6_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) {
3451                 struct ifnet *inifp = m->m_pkthdr.rcvif;
3452                 if ((m = m_copyup(m, sizeof(struct ip6_hdr),
3453                             (max_linkhdr + 3) & ~3)) == NULL) {
3454                         /* XXXJRT new stat, please */
3455                         IP6STAT_INC(ip6s_toosmall);
3456                         in6_ifstat_inc(inifp, ifs6_in_hdrerr);
3457                         goto bad;
3458                 }
3459         } else if (__predict_false(m->m_len < sizeof(struct ip6_hdr))) {
3460                 struct ifnet *inifp = m->m_pkthdr.rcvif;
3461                 if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) {
3462                         IP6STAT_INC(ip6s_toosmall);
3463                         in6_ifstat_inc(inifp, ifs6_in_hdrerr);
3464                         goto bad;
3465                 }
3466         }
3467
3468         ip6 = mtod(m, struct ip6_hdr *);
3469
3470         if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) {
3471                 IP6STAT_INC(ip6s_badvers);
3472                 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr);
3473                 goto bad;
3474         }
3475
3476         /* Checks out, proceed */
3477         *mp = m;
3478         return (0);
3479
3480 bad:
3481         *mp = m;
3482         return (-1);
3483 }
3484 #endif /* INET6 */
3485
3486 /*
3487  * bridge_fragment:
3488  *
3489  *      Fragment mbuf chain in multiple packets and prepend ethernet header.
3490  */
3491 static int
3492 bridge_fragment(struct ifnet *ifp, struct mbuf **mp, struct ether_header *eh,
3493     int snap, struct llc *llc)
3494 {
3495         struct mbuf *m = *mp, *nextpkt = NULL, *mprev = NULL, *mcur = NULL;
3496         struct ip *ip;
3497         int error = -1;
3498
3499         if (m->m_len < sizeof(struct ip) &&
3500             (m = m_pullup(m, sizeof(struct ip))) == NULL)
3501                 goto dropit;
3502         ip = mtod(m, struct ip *);
3503
3504         m->m_pkthdr.csum_flags |= CSUM_IP;
3505         error = ip_fragment(ip, &m, ifp->if_mtu, ifp->if_hwassist);
3506         if (error)
3507                 goto dropit;
3508
3509         /*
3510          * Walk the chain and re-add the Ethernet header for
3511          * each mbuf packet.
3512          */
3513         for (mcur = m; mcur; mcur = mcur->m_nextpkt) {
3514                 nextpkt = mcur->m_nextpkt;
3515                 mcur->m_nextpkt = NULL;
3516                 if (snap) {
3517                         M_PREPEND(mcur, sizeof(struct llc), M_NOWAIT);
3518                         if (mcur == NULL) {
3519                                 error = ENOBUFS;
3520                                 if (mprev != NULL)
3521                                         mprev->m_nextpkt = nextpkt;
3522                                 goto dropit;
3523                         }
3524                         bcopy(llc, mtod(mcur, caddr_t),sizeof(struct llc));
3525                 }
3526
3527                 M_PREPEND(mcur, ETHER_HDR_LEN, M_NOWAIT);
3528                 if (mcur == NULL) {
3529                         error = ENOBUFS;
3530                         if (mprev != NULL)
3531                                 mprev->m_nextpkt = nextpkt;
3532                         goto dropit;
3533                 }
3534                 bcopy(eh, mtod(mcur, caddr_t), ETHER_HDR_LEN);
3535
3536                 /*
3537                  * The previous two M_PREPEND could have inserted one or two
3538                  * mbufs in front so we have to update the previous packet's
3539                  * m_nextpkt.
3540                  */
3541                 mcur->m_nextpkt = nextpkt;
3542                 if (mprev != NULL)
3543                         mprev->m_nextpkt = mcur;
3544                 else {
3545                         /* The first mbuf in the original chain needs to be
3546                          * updated. */
3547                         *mp = mcur;
3548                 }
3549                 mprev = mcur;
3550         }
3551
3552         KMOD_IPSTAT_INC(ips_fragmented);
3553         return (error);
3554
3555 dropit:
3556         for (mcur = *mp; mcur; mcur = m) { /* droping the full packet chain */
3557                 m = mcur->m_nextpkt;
3558                 m_freem(mcur);
3559         }
3560         return (error);
3561 }
3562
3563 static void
3564 bridge_linkstate(struct ifnet *ifp)
3565 {
3566         struct bridge_softc *sc = ifp->if_bridge;
3567         struct bridge_iflist *bif;
3568
3569         BRIDGE_LOCK(sc);
3570         bif = bridge_lookup_member_if(sc, ifp);
3571         if (bif == NULL) {
3572                 BRIDGE_UNLOCK(sc);
3573                 return;
3574         }
3575         bridge_linkcheck(sc);
3576         BRIDGE_UNLOCK(sc);
3577
3578         bstp_linkstate(&bif->bif_stp);
3579 }
3580
3581 static void
3582 bridge_linkcheck(struct bridge_softc *sc)
3583 {
3584         struct bridge_iflist *bif;
3585         int new_link, hasls;
3586
3587         BRIDGE_LOCK_ASSERT(sc);
3588         new_link = LINK_STATE_DOWN;
3589         hasls = 0;
3590         /* Our link is considered up if at least one of our ports is active */
3591         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
3592                 if (bif->bif_ifp->if_capabilities & IFCAP_LINKSTATE)
3593                         hasls++;
3594                 if (bif->bif_ifp->if_link_state == LINK_STATE_UP) {
3595                         new_link = LINK_STATE_UP;
3596                         break;
3597                 }
3598         }
3599         if (!LIST_EMPTY(&sc->sc_iflist) && !hasls) {
3600                 /* If no interfaces support link-state then we default to up */
3601                 new_link = LINK_STATE_UP;
3602         }
3603         if_link_state_change(sc->sc_ifp, new_link);
3604 }