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 VNET_DEFINE_STATIC(uma_zone_t, bridge_rtnode_zone);
 239 #define V_bridge_rtnode_zone    VNET(bridge_rtnode_zone)
 240
 241 static int      bridge_clone_create(struct if_clone *, int, caddr_t);
 242 static void     bridge_clone_destroy(struct ifnet *);
 243
 244 static int      bridge_ioctl(struct ifnet *, u_long, caddr_t);
 245 static void     bridge_mutecaps(struct bridge_softc *);
 246 static void     bridge_set_ifcap(struct bridge_softc *, struct bridge_iflist *,
 247                     int);
 248 static void     bridge_ifdetach(void *arg __unused, struct ifnet *);
 249 static void     bridge_init(void *);
 250 static void     bridge_dummynet(struct mbuf *, struct ifnet *);
 251 static void     bridge_stop(struct ifnet *, int);
 252 static int      bridge_transmit(struct ifnet *, struct mbuf *);
 253 static void     bridge_qflush(struct ifnet *);
 254 static struct mbuf *bridge_input(struct ifnet *, struct mbuf *);
 255 static int      bridge_output(struct ifnet *, struct mbuf *, struct sockaddr *,
 256                     struct rtentry *);
 257 static int      bridge_enqueue(struct bridge_softc *, struct ifnet *,
 258                     struct mbuf *);
 259 static void     bridge_rtdelete(struct bridge_softc *, struct ifnet *ifp, int);
 260
 261 static void     bridge_forward(struct bridge_softc *, struct bridge_iflist *,
 262                     struct mbuf *m);
 263
 264 static void     bridge_timer(void *);
 265
 266 static void     bridge_broadcast(struct bridge_softc *, struct ifnet *,
 267                     struct mbuf *, int);
 268 static void     bridge_span(struct bridge_softc *, struct mbuf *);
 269
 270 static int      bridge_rtupdate(struct bridge_softc *, const uint8_t *,
 271                     uint16_t, struct bridge_iflist *, int, uint8_t);
 272 static struct ifnet *bridge_rtlookup(struct bridge_softc *, const uint8_t *,
 273                     uint16_t);
 274 static void     bridge_rttrim(struct bridge_softc *);
 275 static void     bridge_rtage(struct bridge_softc *);
 276 static void     bridge_rtflush(struct bridge_softc *, int);
 277 static int      bridge_rtdaddr(struct bridge_softc *, const uint8_t *,
 278                     uint16_t);
 279
 280 static void     bridge_rtable_init(struct bridge_softc *);
 281 static void     bridge_rtable_fini(struct bridge_softc *);
 282
 283 static int      bridge_rtnode_addr_cmp(const uint8_t *, const uint8_t *);
 284 static struct bridge_rtnode *bridge_rtnode_lookup(struct bridge_softc *,
 285                     const uint8_t *, uint16_t);
 286 static int      bridge_rtnode_insert(struct bridge_softc *,
 287                     struct bridge_rtnode *);
 288 static void     bridge_rtnode_destroy(struct bridge_softc *,
 289                     struct bridge_rtnode *);
 290 static void     bridge_rtable_expire(struct ifnet *, int);
 291 static void     bridge_state_change(struct ifnet *, int);
 292
 293 static struct bridge_iflist *bridge_lookup_member(struct bridge_softc *,
 294                     const char *name);
 295 static struct bridge_iflist *bridge_lookup_member_if(struct bridge_softc *,
 296                     struct ifnet *ifp);
 297 static void     bridge_delete_member(struct bridge_softc *,
 298                     struct bridge_iflist *, int);
 299 static void     bridge_delete_span(struct bridge_softc *,
 300                     struct bridge_iflist *);
 301
 302 static int      bridge_ioctl_add(struct bridge_softc *, void *);
 303 static int      bridge_ioctl_del(struct bridge_softc *, void *);
 304 static int      bridge_ioctl_gifflags(struct bridge_softc *, void *);
 305 static int      bridge_ioctl_sifflags(struct bridge_softc *, void *);
 306 static int      bridge_ioctl_scache(struct bridge_softc *, void *);
 307 static int      bridge_ioctl_gcache(struct bridge_softc *, void *);
 308 static int      bridge_ioctl_gifs(struct bridge_softc *, void *);
 309 static int      bridge_ioctl_rts(struct bridge_softc *, void *);
 310 static int      bridge_ioctl_saddr(struct bridge_softc *, void *);
 311 static int      bridge_ioctl_sto(struct bridge_softc *, void *);
 312 static int      bridge_ioctl_gto(struct bridge_softc *, void *);
 313 static int      bridge_ioctl_daddr(struct bridge_softc *, void *);
 314 static int      bridge_ioctl_flush(struct bridge_softc *, void *);
 315 static int      bridge_ioctl_gpri(struct bridge_softc *, void *);
 316 static int      bridge_ioctl_spri(struct bridge_softc *, void *);
 317 static int      bridge_ioctl_ght(struct bridge_softc *, void *);
 318 static int      bridge_ioctl_sht(struct bridge_softc *, void *);
 319 static int      bridge_ioctl_gfd(struct bridge_softc *, void *);
 320 static int      bridge_ioctl_sfd(struct bridge_softc *, void *);
 321 static int      bridge_ioctl_gma(struct bridge_softc *, void *);
 322 static int      bridge_ioctl_sma(struct bridge_softc *, void *);
 323 static int      bridge_ioctl_sifprio(struct bridge_softc *, void *);
 324 static int      bridge_ioctl_sifcost(struct bridge_softc *, void *);
 325 static int      bridge_ioctl_sifmaxaddr(struct bridge_softc *, void *);
 326 static int      bridge_ioctl_addspan(struct bridge_softc *, void *);
 327 static int      bridge_ioctl_delspan(struct bridge_softc *, void *);
 328 static int      bridge_ioctl_gbparam(struct bridge_softc *, void *);
 329 static int      bridge_ioctl_grte(struct bridge_softc *, void *);
 330 static int      bridge_ioctl_gifsstp(struct bridge_softc *, void *);
 331 static int      bridge_ioctl_sproto(struct bridge_softc *, void *);
 332 static int      bridge_ioctl_stxhc(struct bridge_softc *, void *);
 333 static int      bridge_pfil(struct mbuf **, struct ifnet *, struct ifnet *,
 334                     int);
 335 static int      bridge_ip_checkbasic(struct mbuf **mp);
 336 #ifdef INET6
 337 static int      bridge_ip6_checkbasic(struct mbuf **mp);
 338 #endif /* INET6 */
 339 static int      bridge_fragment(struct ifnet *, struct mbuf **mp,
 340                     struct ether_header *, int, struct llc *);
 341 static void     bridge_linkstate(struct ifnet *ifp);
 342 static void     bridge_linkcheck(struct bridge_softc *sc);
 343
 344
 345 /* The default bridge vlan is 1 (IEEE 802.1Q-2003 Table 9-2) */
 346 #define VLANTAGOF(_m)   \
 347     (_m->m_flags & M_VLANTAG) ? EVL_VLANOFTAG(_m->m_pkthdr.ether_vtag) : 1
 348
 349 static struct bstp_cb_ops bridge_ops = {
 350         .bcb_state = bridge_state_change,
 351         .bcb_rtage = bridge_rtable_expire
 352 };
 353
 354 SYSCTL_DECL(_net_link);
 355 static SYSCTL_NODE(_net_link, IFT_BRIDGE, bridge, CTLFLAG_RW, 0, "Bridge");
 356
 357 /* only pass IP[46] packets when pfil is enabled */
 358 VNET_DEFINE_STATIC(int, pfil_onlyip) = 1;
 359 #define V_pfil_onlyip   VNET(pfil_onlyip)
 360 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_onlyip,
 361     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_onlyip), 0,
 362     "Only pass IP packets when pfil is enabled");
 363
 364 /* run pfil hooks on the bridge interface */
 365 VNET_DEFINE_STATIC(int, pfil_bridge) = 1;
 366 #define V_pfil_bridge   VNET(pfil_bridge)
 367 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_bridge,
 368     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_bridge), 0,
 369     "Packet filter on the bridge interface");
 370
 371 /* layer2 filter with ipfw */
 372 VNET_DEFINE_STATIC(int, pfil_ipfw);
 373 #define V_pfil_ipfw     VNET(pfil_ipfw)
 374
 375 /* layer2 ARP filter with ipfw */
 376 VNET_DEFINE_STATIC(int, pfil_ipfw_arp);
 377 #define V_pfil_ipfw_arp VNET(pfil_ipfw_arp)
 378 SYSCTL_INT(_net_link_bridge, OID_AUTO, ipfw_arp,
 379     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_ipfw_arp), 0,
 380     "Filter ARP packets through IPFW layer2");
 381
 382 /* run pfil hooks on the member interface */
 383 VNET_DEFINE_STATIC(int, pfil_member) = 1;
 384 #define V_pfil_member   VNET(pfil_member)
 385 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_member,
 386     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_member), 0,
 387     "Packet filter on the member interface");
 388
 389 /* run pfil hooks on the physical interface for locally destined packets */
 390 VNET_DEFINE_STATIC(int, pfil_local_phys);
 391 #define V_pfil_local_phys       VNET(pfil_local_phys)
 392 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_local_phys,
 393     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_local_phys), 0,
 394     "Packet filter on the physical interface for locally destined packets");
 395
 396 /* log STP state changes */
 397 VNET_DEFINE_STATIC(int, log_stp);
 398 #define V_log_stp       VNET(log_stp)
 399 SYSCTL_INT(_net_link_bridge, OID_AUTO, log_stp,
 400     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(log_stp), 0,
 401     "Log STP state changes");
 402
 403 /* share MAC with first bridge member */
 404 VNET_DEFINE_STATIC(int, bridge_inherit_mac);
 405 #define V_bridge_inherit_mac    VNET(bridge_inherit_mac)
 406 SYSCTL_INT(_net_link_bridge, OID_AUTO, inherit_mac,
 407     CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(bridge_inherit_mac), 0,
 408     "Inherit MAC address from the first bridge member");
 409
 410 VNET_DEFINE_STATIC(int, allow_llz_overlap) = 0;
 411 #define V_allow_llz_overlap     VNET(allow_llz_overlap)
 412 SYSCTL_INT(_net_link_bridge, OID_AUTO, allow_llz_overlap,
 413     CTLFLAG_RW | CTLFLAG_VNET, &VNET_NAME(allow_llz_overlap), 0,
 414     "Allow overlap of link-local scope "
 415     "zones of a bridge interface and the member interfaces");
 416
 417 struct bridge_control {
 418         int     (*bc_func)(struct bridge_softc *, void *);
 419         int     bc_argsize;
 420         int     bc_flags;
 421 };
 422
 423 #define BC_F_COPYIN             0x01    /* copy arguments in */
 424 #define BC_F_COPYOUT            0x02    /* copy arguments out */
 425 #define BC_F_SUSER              0x04    /* do super-user check */
 426
 427 const struct bridge_control bridge_control_table[] = {
 428         { bridge_ioctl_add,             sizeof(struct ifbreq),
 429           BC_F_COPYIN|BC_F_SUSER },
 430         { bridge_ioctl_del,             sizeof(struct ifbreq),
 431           BC_F_COPYIN|BC_F_SUSER },
 432
 433         { bridge_ioctl_gifflags,        sizeof(struct ifbreq),
 434           BC_F_COPYIN|BC_F_COPYOUT },
 435         { bridge_ioctl_sifflags,        sizeof(struct ifbreq),
 436           BC_F_COPYIN|BC_F_SUSER },
 437
 438         { bridge_ioctl_scache,          sizeof(struct ifbrparam),
 439           BC_F_COPYIN|BC_F_SUSER },
 440         { bridge_ioctl_gcache,          sizeof(struct ifbrparam),
 441           BC_F_COPYOUT },
 442
 443         { bridge_ioctl_gifs,            sizeof(struct ifbifconf),
 444           BC_F_COPYIN|BC_F_COPYOUT },
 445         { bridge_ioctl_rts,             sizeof(struct ifbaconf),
 446           BC_F_COPYIN|BC_F_COPYOUT },
 447
 448         { bridge_ioctl_saddr,           sizeof(struct ifbareq),
 449           BC_F_COPYIN|BC_F_SUSER },
 450
 451         { bridge_ioctl_sto,             sizeof(struct ifbrparam),
 452           BC_F_COPYIN|BC_F_SUSER },
 453         { bridge_ioctl_gto,             sizeof(struct ifbrparam),
 454           BC_F_COPYOUT },
 455
 456         { bridge_ioctl_daddr,           sizeof(struct ifbareq),
 457           BC_F_COPYIN|BC_F_SUSER },
 458
 459         { bridge_ioctl_flush,           sizeof(struct ifbreq),
 460           BC_F_COPYIN|BC_F_SUSER },
 461
 462         { bridge_ioctl_gpri,            sizeof(struct ifbrparam),
 463           BC_F_COPYOUT },
 464         { bridge_ioctl_spri,            sizeof(struct ifbrparam),
 465           BC_F_COPYIN|BC_F_SUSER },
 466
 467         { bridge_ioctl_ght,             sizeof(struct ifbrparam),
 468           BC_F_COPYOUT },
 469         { bridge_ioctl_sht,             sizeof(struct ifbrparam),
 470           BC_F_COPYIN|BC_F_SUSER },
 471
 472         { bridge_ioctl_gfd,             sizeof(struct ifbrparam),
 473           BC_F_COPYOUT },
 474         { bridge_ioctl_sfd,             sizeof(struct ifbrparam),
 475           BC_F_COPYIN|BC_F_SUSER },
 476
 477         { bridge_ioctl_gma,             sizeof(struct ifbrparam),
 478           BC_F_COPYOUT },
 479         { bridge_ioctl_sma,             sizeof(struct ifbrparam),
 480           BC_F_COPYIN|BC_F_SUSER },
 481
 482         { bridge_ioctl_sifprio,         sizeof(struct ifbreq),
 483           BC_F_COPYIN|BC_F_SUSER },
 484
 485         { bridge_ioctl_sifcost,         sizeof(struct ifbreq),
 486           BC_F_COPYIN|BC_F_SUSER },
 487
 488         { bridge_ioctl_addspan,         sizeof(struct ifbreq),
 489           BC_F_COPYIN|BC_F_SUSER },
 490         { bridge_ioctl_delspan,         sizeof(struct ifbreq),
 491           BC_F_COPYIN|BC_F_SUSER },
 492
 493         { bridge_ioctl_gbparam,         sizeof(struct ifbropreq),
 494           BC_F_COPYOUT },
 495
 496         { bridge_ioctl_grte,            sizeof(struct ifbrparam),
 497           BC_F_COPYOUT },
 498
 499         { bridge_ioctl_gifsstp,         sizeof(struct ifbpstpconf),
 500           BC_F_COPYIN|BC_F_COPYOUT },
 501
 502         { bridge_ioctl_sproto,          sizeof(struct ifbrparam),
 503           BC_F_COPYIN|BC_F_SUSER },
 504
 505         { bridge_ioctl_stxhc,           sizeof(struct ifbrparam),
 506           BC_F_COPYIN|BC_F_SUSER },
 507
 508         { bridge_ioctl_sifmaxaddr,      sizeof(struct ifbreq),
 509           BC_F_COPYIN|BC_F_SUSER },
 510
 511 };
 512 const int bridge_control_table_size = nitems(bridge_control_table);
 513
 514 VNET_DEFINE_STATIC(LIST_HEAD(, bridge_softc), bridge_list);
 515 #define V_bridge_list   VNET(bridge_list)
 516 #define BRIDGE_LIST_LOCK_INIT(x)        mtx_init(&V_bridge_list_mtx,    \
 517                                             "if_bridge list", NULL, MTX_DEF)
 518 #define BRIDGE_LIST_LOCK_DESTROY(x)     mtx_destroy(&V_bridge_list_mtx)
 519 #define BRIDGE_LIST_LOCK(x)             mtx_lock(&V_bridge_list_mtx)
 520 #define BRIDGE_LIST_UNLOCK(x)           mtx_unlock(&V_bridge_list_mtx)
 521
 522 VNET_DEFINE_STATIC(struct if_clone *, bridge_cloner);
 523 #define V_bridge_cloner VNET(bridge_cloner)
 524
 525 static const char bridge_name[] = "bridge";
 526
 527 static void
 528 vnet_bridge_init(const void *unused __unused)
 529 {
 530
 531         V_bridge_rtnode_zone = uma_zcreate("bridge_rtnode",
 532             sizeof(struct bridge_rtnode), NULL, NULL, NULL, NULL,
 533             UMA_ALIGN_PTR, 0);
 534         BRIDGE_LIST_LOCK_INIT();
 535         LIST_INIT(&V_bridge_list);
 536         V_bridge_cloner = if_clone_simple(bridge_name,
 537             bridge_clone_create, bridge_clone_destroy, 0);
 538 }
 539 VNET_SYSINIT(vnet_bridge_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
 540     vnet_bridge_init, NULL);
 541
 542 static void
 543 vnet_bridge_uninit(const void *unused __unused)
 544 {
 545
 546         if_clone_detach(V_bridge_cloner);
 547         V_bridge_cloner = NULL;
 548         BRIDGE_LIST_LOCK_DESTROY();
 549         uma_zdestroy(V_bridge_rtnode_zone);
 550 }
 551 VNET_SYSUNINIT(vnet_bridge_uninit, SI_SUB_PSEUDO, SI_ORDER_ANY,
 552     vnet_bridge_uninit, NULL);
 553
 554 static int
 555 bridge_modevent(module_t mod, int type, void *data)
 556 {
 557
 558         switch (type) {
 559         case MOD_LOAD:
 560                 bridge_dn_p = bridge_dummynet;
 561                 bridge_detach_cookie = EVENTHANDLER_REGISTER(
 562                     ifnet_departure_event, bridge_ifdetach, NULL,
 563                     EVENTHANDLER_PRI_ANY);
 564                 break;
 565         case MOD_UNLOAD:
 566                 EVENTHANDLER_DEREGISTER(ifnet_departure_event,
 567                     bridge_detach_cookie);
 568                 bridge_dn_p = NULL;
 569                 break;
 570         default:
 571                 return (EOPNOTSUPP);
 572         }
 573         return (0);
 574 }
 575
 576 static moduledata_t bridge_mod = {
 577         "if_bridge",
 578         bridge_modevent,
 579         0
 580 };
 581
 582 DECLARE_MODULE(if_bridge, bridge_mod, SI_SUB_PSEUDO, SI_ORDER_ANY);
 583 MODULE_VERSION(if_bridge, 1);
 584 MODULE_DEPEND(if_bridge, bridgestp, 1, 1, 1);
 585
 586 /*
 587  * handler for net.link.bridge.ipfw
 588  */
 589 static int
 590 sysctl_pfil_ipfw(SYSCTL_HANDLER_ARGS)
 591 {
 592         int enable = V_pfil_ipfw;
 593         int error;
 594
 595         error = sysctl_handle_int(oidp, &enable, 0, req);
 596         enable &= 1;
 597
 598         if (enable != V_pfil_ipfw) {
 599                 V_pfil_ipfw = enable;
 600
 601                 /*
 602                  * Disable pfil so that ipfw doesnt run twice, if the user
 603                  * really wants both then they can re-enable pfil_bridge and/or
 604                  * pfil_member. Also allow non-ip packets as ipfw can filter by
 605                  * layer2 type.
 606                  */
 607                 if (V_pfil_ipfw) {
 608                         V_pfil_onlyip = 0;
 609                         V_pfil_bridge = 0;
 610                         V_pfil_member = 0;
 611                 }
 612         }
 613
 614         return (error);
 615 }
 616 SYSCTL_PROC(_net_link_bridge, OID_AUTO, ipfw,
 617     CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_VNET,
 618     &VNET_NAME(pfil_ipfw), 0, &sysctl_pfil_ipfw, "I",
 619     "Layer2 filter with IPFW");
 620
 621 /*
 622  * bridge_clone_create:
 623  *
 624  *      Create a new bridge instance.
 625  */
 626 static int
 627 bridge_clone_create(struct if_clone *ifc, int unit, caddr_t params)
 628 {
 629         struct bridge_softc *sc, *sc2;
 630         struct ifnet *bifp, *ifp;
 631         int fb, retry;
 632         unsigned long hostid;
 633
 634         sc = malloc(sizeof(*sc), M_DEVBUF, M_WAITOK|M_ZERO);
 635         ifp = sc->sc_ifp = if_alloc(IFT_ETHER);
 636         if (ifp == NULL) {
 637                 free(sc, M_DEVBUF);
 638                 return (ENOSPC);
 639         }
 640
 641         BRIDGE_LOCK_INIT(sc);
 642         sc->sc_brtmax = BRIDGE_RTABLE_MAX;
 643         sc->sc_brttimeout = BRIDGE_RTABLE_TIMEOUT;
 644
 645         /* Initialize our routing table. */
 646         bridge_rtable_init(sc);
 647
 648         callout_init_mtx(&sc->sc_brcallout, &sc->sc_mtx, 0);
 649
 650         LIST_INIT(&sc->sc_iflist);
 651         LIST_INIT(&sc->sc_spanlist);
 652
 653         ifp->if_softc = sc;
 654         if_initname(ifp, bridge_name, unit);
 655         ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
 656         ifp->if_ioctl = bridge_ioctl;
 657         ifp->if_transmit = bridge_transmit;
 658         ifp->if_qflush = bridge_qflush;
 659         ifp->if_init = bridge_init;
 660         ifp->if_type = IFT_BRIDGE;
 661
 662         /*
 663          * Generate an ethernet address with a locally administered address.
 664          *
 665          * Since we are using random ethernet addresses for the bridge, it is
 666          * possible that we might have address collisions, so make sure that
 667          * this hardware address isn't already in use on another bridge.
 668          * The first try uses the hostid and falls back to arc4rand().
 669          */
 670         fb = 0;
 671         getcredhostid(curthread->td_ucred, &hostid);
 672         do {
 673                 if (fb || hostid == 0) {
 674                         ether_gen_addr(ifp, &sc->sc_defaddr);
 675                 } else {
 676                         sc->sc_defaddr.octet[0] = 0x2;
 677                         sc->sc_defaddr.octet[1] = (hostid >> 24) & 0xff;
 678                         sc->sc_defaddr.octet[2] = (hostid >> 16) & 0xff;
 679                         sc->sc_defaddr.octet[3] = (hostid >> 8 ) & 0xff;
 680                         sc->sc_defaddr.octet[4] =  hostid        & 0xff;
 681                         sc->sc_defaddr.octet[5] = ifp->if_dunit & 0xff;
 682                 }
 683
 684                 fb = 1;
 685                 retry = 0;
 686                 BRIDGE_LIST_LOCK();
 687                 LIST_FOREACH(sc2, &V_bridge_list, sc_list) {
 688                         bifp = sc2->sc_ifp;
 689                         if (memcmp(sc->sc_defaddr.octet,
 690                             IF_LLADDR(bifp), ETHER_ADDR_LEN) == 0) {
 691                                 retry = 1;
 692                                 break;
 693                         }
 694                 }
 695                 BRIDGE_LIST_UNLOCK();
 696         } while (retry == 1);
 697
 698         bstp_attach(&sc->sc_stp, &bridge_ops);
 699         ether_ifattach(ifp, sc->sc_defaddr.octet);
 700         /* Now undo some of the damage... */
 701         ifp->if_baudrate = 0;
 702         ifp->if_type = IFT_BRIDGE;
 703
 704         BRIDGE_LIST_LOCK();
 705         LIST_INSERT_HEAD(&V_bridge_list, sc, sc_list);
 706         BRIDGE_LIST_UNLOCK();
 707
 708         return (0);
 709 }
 710
 711 /*
 712  * bridge_clone_destroy:
 713  *
 714  *      Destroy a bridge instance.
 715  */
 716 static void
 717 bridge_clone_destroy(struct ifnet *ifp)
 718 {
 719         struct bridge_softc *sc = ifp->if_softc;
 720         struct bridge_iflist *bif;
 721
 722         BRIDGE_LOCK(sc);
 723
 724         bridge_stop(ifp, 1);
 725         ifp->if_flags &= ~IFF_UP;
 726
 727         while ((bif = LIST_FIRST(&sc->sc_iflist)) != NULL)
 728                 bridge_delete_member(sc, bif, 0);
 729
 730         while ((bif = LIST_FIRST(&sc->sc_spanlist)) != NULL) {
 731                 bridge_delete_span(sc, bif);
 732         }
 733
 734         /* Tear down the routing table. */
 735         bridge_rtable_fini(sc);
 736
 737         BRIDGE_UNLOCK(sc);
 738
 739         callout_drain(&sc->sc_brcallout);
 740
 741         BRIDGE_LIST_LOCK();
 742         LIST_REMOVE(sc, sc_list);
 743         BRIDGE_LIST_UNLOCK();
 744
 745         bstp_detach(&sc->sc_stp);
 746         ether_ifdetach(ifp);
 747         if_free(ifp);
 748
 749         BRIDGE_LOCK_DESTROY(sc);
 750         free(sc, M_DEVBUF);
 751 }
 752
 753 /*
 754  * bridge_ioctl:
 755  *
 756  *      Handle a control request from the operator.
 757  */
 758 static int
 759 bridge_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
 760 {
 761         struct bridge_softc *sc = ifp->if_softc;
 762         struct ifreq *ifr = (struct ifreq *)data;
 763         struct bridge_iflist *bif;
 764         struct thread *td = curthread;
 765         union {
 766                 struct ifbreq ifbreq;
 767                 struct ifbifconf ifbifconf;
 768                 struct ifbareq ifbareq;
 769                 struct ifbaconf ifbaconf;
 770                 struct ifbrparam ifbrparam;
 771                 struct ifbropreq ifbropreq;
 772         } args;
 773         struct ifdrv *ifd = (struct ifdrv *) data;
 774         const struct bridge_control *bc;
 775         int error = 0;
 776
 777         switch (cmd) {
 778
 779         case SIOCADDMULTI:
 780         case SIOCDELMULTI:
 781                 break;
 782
 783         case SIOCGDRVSPEC:
 784         case SIOCSDRVSPEC:
 785                 if (ifd->ifd_cmd >= bridge_control_table_size) {
 786                         error = EINVAL;
 787                         break;
 788                 }
 789                 bc = &bridge_control_table[ifd->ifd_cmd];
 790
 791                 if (cmd == SIOCGDRVSPEC &&
 792                     (bc->bc_flags & BC_F_COPYOUT) == 0) {
 793                         error = EINVAL;
 794                         break;
 795                 }
 796                 else if (cmd == SIOCSDRVSPEC &&
 797                     (bc->bc_flags & BC_F_COPYOUT) != 0) {
 798                         error = EINVAL;
 799                         break;
 800                 }
 801
 802                 if (bc->bc_flags & BC_F_SUSER) {
 803                         error = priv_check(td, PRIV_NET_BRIDGE);
 804                         if (error)
 805                                 break;
 806                 }
 807
 808                 if (ifd->ifd_len != bc->bc_argsize ||
 809                     ifd->ifd_len > sizeof(args)) {
 810                         error = EINVAL;
 811                         break;
 812                 }
 813
 814                 bzero(&args, sizeof(args));
 815                 if (bc->bc_flags & BC_F_COPYIN) {
 816                         error = copyin(ifd->ifd_data, &args, ifd->ifd_len);
 817                         if (error)
 818                                 break;
 819                 }
 820
 821                 BRIDGE_LOCK(sc);
 822                 error = (*bc->bc_func)(sc, &args);
 823                 BRIDGE_UNLOCK(sc);
 824                 if (error)
 825                         break;
 826
 827                 if (bc->bc_flags & BC_F_COPYOUT)
 828                         error = copyout(&args, ifd->ifd_data, ifd->ifd_len);
 829
 830                 break;
 831
 832         case SIOCSIFFLAGS:
 833                 if (!(ifp->if_flags & IFF_UP) &&
 834                     (ifp->if_drv_flags & IFF_DRV_RUNNING)) {
 835                         /*
 836                          * If interface is marked down and it is running,
 837                          * then stop and disable it.
 838                          */
 839                         BRIDGE_LOCK(sc);
 840                         bridge_stop(ifp, 1);
 841                         BRIDGE_UNLOCK(sc);
 842                 } else if ((ifp->if_flags & IFF_UP) &&
 843                     !(ifp->if_drv_flags & IFF_DRV_RUNNING)) {
 844                         /*
 845                          * If interface is marked up and it is stopped, then
 846                          * start it.
 847                          */
 848                         (*ifp->if_init)(sc);
 849                 }
 850                 break;
 851
 852         case SIOCSIFMTU:
 853                 if (ifr->ifr_mtu < 576) {
 854                         error = EINVAL;
 855                         break;
 856                 }
 857                 if (LIST_EMPTY(&sc->sc_iflist)) {
 858                         sc->sc_ifp->if_mtu = ifr->ifr_mtu;
 859                         break;
 860                 }
 861                 BRIDGE_LOCK(sc);
 862                 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
 863                         if (bif->bif_ifp->if_mtu != ifr->ifr_mtu) {
 864                                 log(LOG_NOTICE, "%s: invalid MTU: %u(%s)"
 865                                     " != %d\n", sc->sc_ifp->if_xname,
 866                                     bif->bif_ifp->if_mtu,
 867                                     bif->bif_ifp->if_xname, ifr->ifr_mtu);
 868                                 error = EINVAL;
 869                                 break;
 870                         }
 871                 }
 872                 if (!error)
 873                         sc->sc_ifp->if_mtu = ifr->ifr_mtu;
 874                 BRIDGE_UNLOCK(sc);
 875                 break;
 876         default:
 877                 /*
 878                  * drop the lock as ether_ioctl() will call bridge_start() and
 879                  * cause the lock to be recursed.
 880                  */
 881                 error = ether_ioctl(ifp, cmd, data);
 882                 break;
 883         }
 884
 885         return (error);
 886 }
 887
 888 /*
 889  * bridge_mutecaps:
 890  *
 891  *      Clear or restore unwanted capabilities on the member interface
 892  */
 893 static void
 894 bridge_mutecaps(struct bridge_softc *sc)
 895 {
 896         struct bridge_iflist *bif;
 897         int enabled, mask;
 898
 899         /* Initial bitmask of capabilities to test */
 900         mask = BRIDGE_IFCAPS_MASK;
 901
 902         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
 903                 /* Every member must support it or its disabled */
 904                 mask &= bif->bif_savedcaps;
 905         }
 906
 907         BRIDGE_XLOCK(sc);
 908         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
 909                 enabled = bif->bif_ifp->if_capenable;
 910                 enabled &= ~BRIDGE_IFCAPS_STRIP;
 911                 /* strip off mask bits and enable them again if allowed */
 912                 enabled &= ~BRIDGE_IFCAPS_MASK;
 913                 enabled |= mask;
 914                 BRIDGE_UNLOCK(sc);
 915                 bridge_set_ifcap(sc, bif, enabled);
 916                 BRIDGE_LOCK(sc);
 917         }
 918         BRIDGE_XDROP(sc);
 919
 920 }
 921
 922 static void
 923 bridge_set_ifcap(struct bridge_softc *sc, struct bridge_iflist *bif, int set)
 924 {
 925         struct ifnet *ifp = bif->bif_ifp;
 926         struct ifreq ifr;
 927         int error, mask, stuck;
 928
 929         BRIDGE_UNLOCK_ASSERT(sc);
 930
 931         bzero(&ifr, sizeof(ifr));
 932         ifr.ifr_reqcap = set;
 933
 934         if (ifp->if_capenable != set) {
 935                 error = (*ifp->if_ioctl)(ifp, SIOCSIFCAP, (caddr_t)&ifr);
 936                 if (error)
 937                         if_printf(sc->sc_ifp,
 938                             "error setting capabilities on %s: %d\n",
 939                             ifp->if_xname, error);
 940                 mask = BRIDGE_IFCAPS_MASK | BRIDGE_IFCAPS_STRIP;
 941                 stuck = ifp->if_capenable & mask & ~set;
 942                 if (stuck != 0)
 943                         if_printf(sc->sc_ifp,
 944                             "can't disable some capabilities on %s: 0x%x\n",
 945                             ifp->if_xname, stuck);
 946         }
 947 }
 948
 949 /*
 950  * bridge_lookup_member:
 951  *
 952  *      Lookup a bridge member interface.
 953  */
 954 static struct bridge_iflist *
 955 bridge_lookup_member(struct bridge_softc *sc, const char *name)
 956 {
 957         struct bridge_iflist *bif;
 958         struct ifnet *ifp;
 959
 960         BRIDGE_LOCK_ASSERT(sc);
 961
 962         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
 963                 ifp = bif->bif_ifp;
 964                 if (strcmp(ifp->if_xname, name) == 0)
 965                         return (bif);
 966         }
 967
 968         return (NULL);
 969 }
 970
 971 /*
 972  * bridge_lookup_member_if:
 973  *
 974  *      Lookup a bridge member interface by ifnet*.
 975  */
 976 static struct bridge_iflist *
 977 bridge_lookup_member_if(struct bridge_softc *sc, struct ifnet *member_ifp)
 978 {
 979         struct bridge_iflist *bif;
 980
 981         BRIDGE_LOCK_ASSERT(sc);
 982
 983         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
 984                 if (bif->bif_ifp == member_ifp)
 985                         return (bif);
 986         }
 987
 988         return (NULL);
 989 }
 990
 991 /*
 992  * bridge_delete_member:
 993  *
 994  *      Delete the specified member interface.
 995  */
 996 static void
 997 bridge_delete_member(struct bridge_softc *sc, struct bridge_iflist *bif,
 998     int gone)
 999 {
1000         struct ifnet *ifs = bif->bif_ifp;
1001         struct ifnet *fif = NULL;
1002
1003         BRIDGE_LOCK_ASSERT(sc);
1004
1005         if (bif->bif_flags & IFBIF_STP)
1006                 bstp_disable(&bif->bif_stp);
1007
1008         ifs->if_bridge = NULL;
1009         BRIDGE_XLOCK(sc);
1010         LIST_REMOVE(bif, bif_next);
1011         BRIDGE_XDROP(sc);
1012
1013         /*
1014          * If removing the interface that gave the bridge its mac address, set
1015          * the mac address of the bridge to the address of the next member, or
1016          * to its default address if no members are left.
1017          */
1018         if (V_bridge_inherit_mac && sc->sc_ifaddr == ifs) {
1019                 if (LIST_EMPTY(&sc->sc_iflist)) {
1020                         bcopy(&sc->sc_defaddr,
1021                             IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN);
1022                         sc->sc_ifaddr = NULL;
1023                 } else {
1024                         fif = LIST_FIRST(&sc->sc_iflist)->bif_ifp;
1025                         bcopy(IF_LLADDR(fif),
1026                             IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN);
1027                         sc->sc_ifaddr = fif;
1028                 }
1029                 EVENTHANDLER_INVOKE(iflladdr_event, sc->sc_ifp);
1030         }
1031
1032         bridge_linkcheck(sc);
1033         bridge_mutecaps(sc);    /* recalcuate now this interface is removed */
1034         bridge_rtdelete(sc, ifs, IFBF_FLUSHALL);
1035         KASSERT(bif->bif_addrcnt == 0,
1036             ("%s: %d bridge routes referenced", __func__, bif->bif_addrcnt));
1037
1038         ifs->if_bridge_output = NULL;
1039         ifs->if_bridge_input = NULL;
1040         ifs->if_bridge_linkstate = NULL;
1041         BRIDGE_UNLOCK(sc);
1042         if (!gone) {
1043                 switch (ifs->if_type) {
1044                 case IFT_ETHER:
1045                 case IFT_L2VLAN:
1046                         /*
1047                          * Take the interface out of promiscuous mode, but only
1048                          * if it was promiscuous in the first place. It might
1049                          * not be if we're in the bridge_ioctl_add() error path.
1050                          */
1051                         if (ifs->if_flags & IFF_PROMISC)
1052                                 (void) ifpromisc(ifs, 0);
1053                         break;
1054
1055                 case IFT_GIF:
1056                         break;
1057
1058                 default:
1059 #ifdef DIAGNOSTIC
1060                         panic("bridge_delete_member: impossible");
1061 #endif
1062                         break;
1063                 }
1064                 /* reneable any interface capabilities */
1065                 bridge_set_ifcap(sc, bif, bif->bif_savedcaps);
1066         }
1067         bstp_destroy(&bif->bif_stp);    /* prepare to free */
1068         BRIDGE_LOCK(sc);
1069         free(bif, M_DEVBUF);
1070 }
1071
1072 /*
1073  * bridge_delete_span:
1074  *
1075  *      Delete the specified span interface.
1076  */
1077 static void
1078 bridge_delete_span(struct bridge_softc *sc, struct bridge_iflist *bif)
1079 {
1080         BRIDGE_LOCK_ASSERT(sc);
1081
1082         KASSERT(bif->bif_ifp->if_bridge == NULL,
1083             ("%s: not a span interface", __func__));
1084
1085         LIST_REMOVE(bif, bif_next);
1086         free(bif, M_DEVBUF);
1087 }
1088
1089 static int
1090 bridge_ioctl_add(struct bridge_softc *sc, void *arg)
1091 {
1092         struct ifbreq *req = arg;
1093         struct bridge_iflist *bif = NULL;
1094         struct ifnet *ifs;
1095         int error = 0;
1096
1097         ifs = ifunit(req->ifbr_ifsname);
1098         if (ifs == NULL)
1099                 return (ENOENT);
1100         if (ifs->if_ioctl == NULL)      /* must be supported */
1101                 return (EINVAL);
1102
1103         /* If it's in the span list, it can't be a member. */
1104         LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1105                 if (ifs == bif->bif_ifp)
1106                         return (EBUSY);
1107
1108         if (ifs->if_bridge == sc)
1109                 return (EEXIST);
1110
1111         if (ifs->if_bridge != NULL)
1112                 return (EBUSY);
1113
1114         switch (ifs->if_type) {
1115         case IFT_ETHER:
1116         case IFT_L2VLAN:
1117         case IFT_GIF:
1118                 /* permitted interface types */
1119                 break;
1120         default:
1121                 return (EINVAL);
1122         }
1123
1124 #ifdef INET6
1125         /*
1126          * Two valid inet6 addresses with link-local scope must not be
1127          * on the parent interface and the member interfaces at the
1128          * same time.  This restriction is needed to prevent violation
1129          * of link-local scope zone.  Attempts to add a member
1130          * interface which has inet6 addresses when the parent has
1131          * inet6 triggers removal of all inet6 addresses on the member
1132          * interface.
1133          */
1134
1135         /* Check if the parent interface has a link-local scope addr. */
1136         if (V_allow_llz_overlap == 0 &&
1137             in6ifa_llaonifp(sc->sc_ifp) != NULL) {
1138                 /*
1139                  * If any, remove all inet6 addresses from the member
1140                  * interfaces.
1141                  */
1142                 BRIDGE_XLOCK(sc);
1143                 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
1144                         if (in6ifa_llaonifp(bif->bif_ifp)) {
1145                                 BRIDGE_UNLOCK(sc);
1146                                 in6_ifdetach(bif->bif_ifp);
1147                                 BRIDGE_LOCK(sc);
1148                                 if_printf(sc->sc_ifp,
1149                                     "IPv6 addresses on %s have been removed "
1150                                     "before adding it as a member to prevent "
1151                                     "IPv6 address scope violation.\n",
1152                                     bif->bif_ifp->if_xname);
1153                         }
1154                 }
1155                 BRIDGE_XDROP(sc);
1156                 if (in6ifa_llaonifp(ifs)) {
1157                         BRIDGE_UNLOCK(sc);
1158                         in6_ifdetach(ifs);
1159                         BRIDGE_LOCK(sc);
1160                         if_printf(sc->sc_ifp,
1161                             "IPv6 addresses on %s have been removed "
1162                             "before adding it as a member to prevent "
1163                             "IPv6 address scope violation.\n",
1164                             ifs->if_xname);
1165                 }
1166         }
1167 #endif
1168         /* Allow the first Ethernet member to define the MTU */
1169         if (LIST_EMPTY(&sc->sc_iflist))
1170                 sc->sc_ifp->if_mtu = ifs->if_mtu;
1171         else if (sc->sc_ifp->if_mtu != ifs->if_mtu) {
1172                 if_printf(sc->sc_ifp, "invalid MTU: %u(%s) != %u\n",
1173                     ifs->if_mtu, ifs->if_xname, sc->sc_ifp->if_mtu);
1174                 return (EINVAL);
1175         }
1176
1177         bif = malloc(sizeof(*bif), M_DEVBUF, M_NOWAIT|M_ZERO);
1178         if (bif == NULL)
1179                 return (ENOMEM);
1180
1181         bif->bif_ifp = ifs;
1182         bif->bif_flags = IFBIF_LEARNING | IFBIF_DISCOVER;
1183         bif->bif_savedcaps = ifs->if_capenable;
1184
1185         /*
1186          * Assign the interface's MAC address to the bridge if it's the first
1187          * member and the MAC address of the bridge has not been changed from
1188          * the default randomly generated one.
1189          */
1190         if (V_bridge_inherit_mac && LIST_EMPTY(&sc->sc_iflist) &&
1191             !memcmp(IF_LLADDR(sc->sc_ifp), sc->sc_defaddr.octet, ETHER_ADDR_LEN)) {
1192                 bcopy(IF_LLADDR(ifs), IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN);
1193                 sc->sc_ifaddr = ifs;
1194                 EVENTHANDLER_INVOKE(iflladdr_event, sc->sc_ifp);
1195         }
1196
1197         ifs->if_bridge = sc;
1198         ifs->if_bridge_output = bridge_output;
1199         ifs->if_bridge_input = bridge_input;
1200         ifs->if_bridge_linkstate = bridge_linkstate;
1201         bstp_create(&sc->sc_stp, &bif->bif_stp, bif->bif_ifp);
1202         /*
1203          * XXX: XLOCK HERE!?!
1204          *
1205          * NOTE: insert_***HEAD*** should be safe for the traversals.
1206          */
1207         LIST_INSERT_HEAD(&sc->sc_iflist, bif, bif_next);
1208
1209         /* Set interface capabilities to the intersection set of all members */
1210         bridge_mutecaps(sc);
1211         bridge_linkcheck(sc);
1212
1213         /* Place the interface into promiscuous mode */
1214         switch (ifs->if_type) {
1215                 case IFT_ETHER:
1216                 case IFT_L2VLAN:
1217                         BRIDGE_UNLOCK(sc);
1218                         error = ifpromisc(ifs, 1);
1219                         BRIDGE_LOCK(sc);
1220                         break;
1221         }
1222
1223         if (error)
1224                 bridge_delete_member(sc, bif, 0);
1225         return (error);
1226 }
1227
1228 static int
1229 bridge_ioctl_del(struct bridge_softc *sc, void *arg)
1230 {
1231         struct ifbreq *req = arg;
1232         struct bridge_iflist *bif;
1233
1234         bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1235         if (bif == NULL)
1236                 return (ENOENT);
1237
1238         bridge_delete_member(sc, bif, 0);
1239
1240         return (0);
1241 }
1242
1243 static int
1244 bridge_ioctl_gifflags(struct bridge_softc *sc, void *arg)
1245 {
1246         struct ifbreq *req = arg;
1247         struct bridge_iflist *bif;
1248         struct bstp_port *bp;
1249
1250         bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1251         if (bif == NULL)
1252                 return (ENOENT);
1253
1254         bp = &bif->bif_stp;
1255         req->ifbr_ifsflags = bif->bif_flags;
1256         req->ifbr_state = bp->bp_state;
1257         req->ifbr_priority = bp->bp_priority;
1258         req->ifbr_path_cost = bp->bp_path_cost;
1259         req->ifbr_portno = bif->bif_ifp->if_index & 0xfff;
1260         req->ifbr_proto = bp->bp_protover;
1261         req->ifbr_role = bp->bp_role;
1262         req->ifbr_stpflags = bp->bp_flags;
1263         req->ifbr_addrcnt = bif->bif_addrcnt;
1264         req->ifbr_addrmax = bif->bif_addrmax;
1265         req->ifbr_addrexceeded = bif->bif_addrexceeded;
1266
1267         /* Copy STP state options as flags */
1268         if (bp->bp_operedge)
1269                 req->ifbr_ifsflags |= IFBIF_BSTP_EDGE;
1270         if (bp->bp_flags & BSTP_PORT_AUTOEDGE)
1271                 req->ifbr_ifsflags |= IFBIF_BSTP_AUTOEDGE;
1272         if (bp->bp_ptp_link)
1273                 req->ifbr_ifsflags |= IFBIF_BSTP_PTP;
1274         if (bp->bp_flags & BSTP_PORT_AUTOPTP)
1275                 req->ifbr_ifsflags |= IFBIF_BSTP_AUTOPTP;
1276         if (bp->bp_flags & BSTP_PORT_ADMEDGE)
1277                 req->ifbr_ifsflags |= IFBIF_BSTP_ADMEDGE;
1278         if (bp->bp_flags & BSTP_PORT_ADMCOST)
1279                 req->ifbr_ifsflags |= IFBIF_BSTP_ADMCOST;
1280         return (0);
1281 }
1282
1283 static int
1284 bridge_ioctl_sifflags(struct bridge_softc *sc, void *arg)
1285 {
1286         struct ifbreq *req = arg;
1287         struct bridge_iflist *bif;
1288         struct bstp_port *bp;
1289         int error;
1290
1291         bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1292         if (bif == NULL)
1293                 return (ENOENT);
1294         bp = &bif->bif_stp;
1295
1296         if (req->ifbr_ifsflags & IFBIF_SPAN)
1297                 /* SPAN is readonly */
1298                 return (EINVAL);
1299
1300         if (req->ifbr_ifsflags & IFBIF_STP) {
1301                 if ((bif->bif_flags & IFBIF_STP) == 0) {
1302                         error = bstp_enable(&bif->bif_stp);
1303                         if (error)
1304                                 return (error);
1305                 }
1306         } else {
1307                 if ((bif->bif_flags & IFBIF_STP) != 0)
1308                         bstp_disable(&bif->bif_stp);
1309         }
1310
1311         /* Pass on STP flags */
1312         bstp_set_edge(bp, req->ifbr_ifsflags & IFBIF_BSTP_EDGE ? 1 : 0);
1313         bstp_set_autoedge(bp, req->ifbr_ifsflags & IFBIF_BSTP_AUTOEDGE ? 1 : 0);
1314         bstp_set_ptp(bp, req->ifbr_ifsflags & IFBIF_BSTP_PTP ? 1 : 0);
1315         bstp_set_autoptp(bp, req->ifbr_ifsflags & IFBIF_BSTP_AUTOPTP ? 1 : 0);
1316
1317         /* Save the bits relating to the bridge */
1318         bif->bif_flags = req->ifbr_ifsflags & IFBIFMASK;
1319
1320         return (0);
1321 }
1322
1323 static int
1324 bridge_ioctl_scache(struct bridge_softc *sc, void *arg)
1325 {
1326         struct ifbrparam *param = arg;
1327
1328         sc->sc_brtmax = param->ifbrp_csize;
1329         bridge_rttrim(sc);
1330
1331         return (0);
1332 }
1333
1334 static int
1335 bridge_ioctl_gcache(struct bridge_softc *sc, void *arg)
1336 {
1337         struct ifbrparam *param = arg;
1338
1339         param->ifbrp_csize = sc->sc_brtmax;
1340
1341         return (0);
1342 }
1343
1344 static int
1345 bridge_ioctl_gifs(struct bridge_softc *sc, void *arg)
1346 {
1347         struct ifbifconf *bifc = arg;
1348         struct bridge_iflist *bif;
1349         struct ifbreq breq;
1350         char *buf, *outbuf;
1351         int count, buflen, len, error = 0;
1352
1353         count = 0;
1354         LIST_FOREACH(bif, &sc->sc_iflist, bif_next)
1355                 count++;
1356         LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1357                 count++;
1358
1359         buflen = sizeof(breq) * count;
1360         if (bifc->ifbic_len == 0) {
1361                 bifc->ifbic_len = buflen;
1362                 return (0);
1363         }
1364         BRIDGE_UNLOCK(sc);
1365         outbuf = malloc(buflen, M_TEMP, M_WAITOK | M_ZERO);
1366         BRIDGE_LOCK(sc);
1367
1368         count = 0;
1369         buf = outbuf;
1370         len = min(bifc->ifbic_len, buflen);
1371         bzero(&breq, sizeof(breq));
1372         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
1373                 if (len < sizeof(breq))
1374                         break;
1375
1376                 strlcpy(breq.ifbr_ifsname, bif->bif_ifp->if_xname,
1377                     sizeof(breq.ifbr_ifsname));
1378                 /* Fill in the ifbreq structure */
1379                 error = bridge_ioctl_gifflags(sc, &breq);
1380                 if (error)
1381                         break;
1382                 memcpy(buf, &breq, sizeof(breq));
1383                 count++;
1384                 buf += sizeof(breq);
1385                 len -= sizeof(breq);
1386         }
1387         LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) {
1388                 if (len < sizeof(breq))
1389                         break;
1390
1391                 strlcpy(breq.ifbr_ifsname, bif->bif_ifp->if_xname,
1392                     sizeof(breq.ifbr_ifsname));
1393                 breq.ifbr_ifsflags = bif->bif_flags;
1394                 breq.ifbr_portno = bif->bif_ifp->if_index & 0xfff;
1395                 memcpy(buf, &breq, sizeof(breq));
1396                 count++;
1397                 buf += sizeof(breq);
1398                 len -= sizeof(breq);
1399         }
1400
1401         BRIDGE_UNLOCK(sc);
1402         bifc->ifbic_len = sizeof(breq) * count;
1403         error = copyout(outbuf, bifc->ifbic_req, bifc->ifbic_len);
1404         BRIDGE_LOCK(sc);
1405         free(outbuf, M_TEMP);
1406         return (error);
1407 }
1408
1409 static int
1410 bridge_ioctl_rts(struct bridge_softc *sc, void *arg)
1411 {
1412         struct ifbaconf *bac = arg;
1413         struct bridge_rtnode *brt;
1414         struct ifbareq bareq;
1415         char *buf, *outbuf;
1416         int count, buflen, len, error = 0;
1417
1418         if (bac->ifbac_len == 0)
1419                 return (0);
1420
1421         count = 0;
1422         LIST_FOREACH(brt, &sc->sc_rtlist, brt_list)
1423                 count++;
1424         buflen = sizeof(bareq) * count;
1425
1426         BRIDGE_UNLOCK(sc);
1427         outbuf = malloc(buflen, M_TEMP, M_WAITOK | M_ZERO);
1428         BRIDGE_LOCK(sc);
1429
1430         count = 0;
1431         buf = outbuf;
1432         len = min(bac->ifbac_len, buflen);
1433         bzero(&bareq, sizeof(bareq));
1434         LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) {
1435                 if (len < sizeof(bareq))
1436                         goto out;
1437                 strlcpy(bareq.ifba_ifsname, brt->brt_ifp->if_xname,
1438                     sizeof(bareq.ifba_ifsname));
1439                 memcpy(bareq.ifba_dst, brt->brt_addr, sizeof(brt->brt_addr));
1440                 bareq.ifba_vlan = brt->brt_vlan;
1441                 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC &&
1442                                 time_uptime < brt->brt_expire)
1443                         bareq.ifba_expire = brt->brt_expire - time_uptime;
1444                 else
1445                         bareq.ifba_expire = 0;
1446                 bareq.ifba_flags = brt->brt_flags;
1447
1448                 memcpy(buf, &bareq, sizeof(bareq));
1449                 count++;
1450                 buf += sizeof(bareq);
1451                 len -= sizeof(bareq);
1452         }
1453 out:
1454         BRIDGE_UNLOCK(sc);
1455         bac->ifbac_len = sizeof(bareq) * count;
1456         error = copyout(outbuf, bac->ifbac_req, bac->ifbac_len);
1457         BRIDGE_LOCK(sc);
1458         free(outbuf, M_TEMP);
1459         return (error);
1460 }
1461
1462 static int
1463 bridge_ioctl_saddr(struct bridge_softc *sc, void *arg)
1464 {
1465         struct ifbareq *req = arg;
1466         struct bridge_iflist *bif;
1467         int error;
1468
1469         bif = bridge_lookup_member(sc, req->ifba_ifsname);
1470         if (bif == NULL)
1471                 return (ENOENT);
1472
1473         error = bridge_rtupdate(sc, req->ifba_dst, req->ifba_vlan, bif, 1,
1474             req->ifba_flags);
1475
1476         return (error);
1477 }
1478
1479 static int
1480 bridge_ioctl_sto(struct bridge_softc *sc, void *arg)
1481 {
1482         struct ifbrparam *param = arg;
1483
1484         sc->sc_brttimeout = param->ifbrp_ctime;
1485         return (0);
1486 }
1487
1488 static int
1489 bridge_ioctl_gto(struct bridge_softc *sc, void *arg)
1490 {
1491         struct ifbrparam *param = arg;
1492
1493         param->ifbrp_ctime = sc->sc_brttimeout;
1494         return (0);
1495 }
1496
1497 static int
1498 bridge_ioctl_daddr(struct bridge_softc *sc, void *arg)
1499 {
1500         struct ifbareq *req = arg;
1501
1502         return (bridge_rtdaddr(sc, req->ifba_dst, req->ifba_vlan));
1503 }
1504
1505 static int
1506 bridge_ioctl_flush(struct bridge_softc *sc, void *arg)
1507 {
1508         struct ifbreq *req = arg;
1509
1510         bridge_rtflush(sc, req->ifbr_ifsflags);
1511         return (0);
1512 }
1513
1514 static int
1515 bridge_ioctl_gpri(struct bridge_softc *sc, void *arg)
1516 {
1517         struct ifbrparam *param = arg;
1518         struct bstp_state *bs = &sc->sc_stp;
1519
1520         param->ifbrp_prio = bs->bs_bridge_priority;
1521         return (0);
1522 }
1523
1524 static int
1525 bridge_ioctl_spri(struct bridge_softc *sc, void *arg)
1526 {
1527         struct ifbrparam *param = arg;
1528
1529         return (bstp_set_priority(&sc->sc_stp, param->ifbrp_prio));
1530 }
1531
1532 static int
1533 bridge_ioctl_ght(struct bridge_softc *sc, void *arg)
1534 {
1535         struct ifbrparam *param = arg;
1536         struct bstp_state *bs = &sc->sc_stp;
1537
1538         param->ifbrp_hellotime = bs->bs_bridge_htime >> 8;
1539         return (0);
1540 }
1541
1542 static int
1543 bridge_ioctl_sht(struct bridge_softc *sc, void *arg)
1544 {
1545         struct ifbrparam *param = arg;
1546
1547         return (bstp_set_htime(&sc->sc_stp, param->ifbrp_hellotime));
1548 }
1549
1550 static int
1551 bridge_ioctl_gfd(struct bridge_softc *sc, void *arg)
1552 {
1553         struct ifbrparam *param = arg;
1554         struct bstp_state *bs = &sc->sc_stp;
1555
1556         param->ifbrp_fwddelay = bs->bs_bridge_fdelay >> 8;
1557         return (0);
1558 }
1559
1560 static int
1561 bridge_ioctl_sfd(struct bridge_softc *sc, void *arg)
1562 {
1563         struct ifbrparam *param = arg;
1564
1565         return (bstp_set_fdelay(&sc->sc_stp, param->ifbrp_fwddelay));
1566 }
1567
1568 static int
1569 bridge_ioctl_gma(struct bridge_softc *sc, void *arg)
1570 {
1571         struct ifbrparam *param = arg;
1572         struct bstp_state *bs = &sc->sc_stp;
1573
1574         param->ifbrp_maxage = bs->bs_bridge_max_age >> 8;
1575         return (0);
1576 }
1577
1578 static int
1579 bridge_ioctl_sma(struct bridge_softc *sc, void *arg)
1580 {
1581         struct ifbrparam *param = arg;
1582
1583         return (bstp_set_maxage(&sc->sc_stp, param->ifbrp_maxage));
1584 }
1585
1586 static int
1587 bridge_ioctl_sifprio(struct bridge_softc *sc, void *arg)
1588 {
1589         struct ifbreq *req = arg;
1590         struct bridge_iflist *bif;
1591
1592         bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1593         if (bif == NULL)
1594                 return (ENOENT);
1595
1596         return (bstp_set_port_priority(&bif->bif_stp, req->ifbr_priority));
1597 }
1598
1599 static int
1600 bridge_ioctl_sifcost(struct bridge_softc *sc, void *arg)
1601 {
1602         struct ifbreq *req = arg;
1603         struct bridge_iflist *bif;
1604
1605         bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1606         if (bif == NULL)
1607                 return (ENOENT);
1608
1609         return (bstp_set_path_cost(&bif->bif_stp, req->ifbr_path_cost));
1610 }
1611
1612 static int
1613 bridge_ioctl_sifmaxaddr(struct bridge_softc *sc, void *arg)
1614 {
1615         struct ifbreq *req = arg;
1616         struct bridge_iflist *bif;
1617
1618         bif = bridge_lookup_member(sc, req->ifbr_ifsname);
1619         if (bif == NULL)
1620                 return (ENOENT);
1621
1622         bif->bif_addrmax = req->ifbr_addrmax;
1623         return (0);
1624 }
1625
1626 static int
1627 bridge_ioctl_addspan(struct bridge_softc *sc, void *arg)
1628 {
1629         struct ifbreq *req = arg;
1630         struct bridge_iflist *bif = NULL;
1631         struct ifnet *ifs;
1632
1633         ifs = ifunit(req->ifbr_ifsname);
1634         if (ifs == NULL)
1635                 return (ENOENT);
1636
1637         LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1638                 if (ifs == bif->bif_ifp)
1639                         return (EBUSY);
1640
1641         if (ifs->if_bridge != NULL)
1642                 return (EBUSY);
1643
1644         switch (ifs->if_type) {
1645                 case IFT_ETHER:
1646                 case IFT_GIF:
1647                 case IFT_L2VLAN:
1648                         break;
1649                 default:
1650                         return (EINVAL);
1651         }
1652
1653         bif = malloc(sizeof(*bif), M_DEVBUF, M_NOWAIT|M_ZERO);
1654         if (bif == NULL)
1655                 return (ENOMEM);
1656
1657         bif->bif_ifp = ifs;
1658         bif->bif_flags = IFBIF_SPAN;
1659
1660         LIST_INSERT_HEAD(&sc->sc_spanlist, bif, bif_next);
1661
1662         return (0);
1663 }
1664
1665 static int
1666 bridge_ioctl_delspan(struct bridge_softc *sc, void *arg)
1667 {
1668         struct ifbreq *req = arg;
1669         struct bridge_iflist *bif;
1670         struct ifnet *ifs;
1671
1672         ifs = ifunit(req->ifbr_ifsname);
1673         if (ifs == NULL)
1674                 return (ENOENT);
1675
1676         LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1677                 if (ifs == bif->bif_ifp)
1678                         break;
1679
1680         if (bif == NULL)
1681                 return (ENOENT);
1682
1683         bridge_delete_span(sc, bif);
1684
1685         return (0);
1686 }
1687
1688 static int
1689 bridge_ioctl_gbparam(struct bridge_softc *sc, void *arg)
1690 {
1691         struct ifbropreq *req = arg;
1692         struct bstp_state *bs = &sc->sc_stp;
1693         struct bstp_port *root_port;
1694
1695         req->ifbop_maxage = bs->bs_bridge_max_age >> 8;
1696         req->ifbop_hellotime = bs->bs_bridge_htime >> 8;
1697         req->ifbop_fwddelay = bs->bs_bridge_fdelay >> 8;
1698
1699         root_port = bs->bs_root_port;
1700         if (root_port == NULL)
1701                 req->ifbop_root_port = 0;
1702         else
1703                 req->ifbop_root_port = root_port->bp_ifp->if_index;
1704
1705         req->ifbop_holdcount = bs->bs_txholdcount;
1706         req->ifbop_priority = bs->bs_bridge_priority;
1707         req->ifbop_protocol = bs->bs_protover;
1708         req->ifbop_root_path_cost = bs->bs_root_pv.pv_cost;
1709         req->ifbop_bridgeid = bs->bs_bridge_pv.pv_dbridge_id;
1710         req->ifbop_designated_root = bs->bs_root_pv.pv_root_id;
1711         req->ifbop_designated_bridge = bs->bs_root_pv.pv_dbridge_id;
1712         req->ifbop_last_tc_time.tv_sec = bs->bs_last_tc_time.tv_sec;
1713         req->ifbop_last_tc_time.tv_usec = bs->bs_last_tc_time.tv_usec;
1714
1715         return (0);
1716 }
1717
1718 static int
1719 bridge_ioctl_grte(struct bridge_softc *sc, void *arg)
1720 {
1721         struct ifbrparam *param = arg;
1722
1723         param->ifbrp_cexceeded = sc->sc_brtexceeded;
1724         return (0);
1725 }
1726
1727 static int
1728 bridge_ioctl_gifsstp(struct bridge_softc *sc, void *arg)
1729 {
1730         struct ifbpstpconf *bifstp = arg;
1731         struct bridge_iflist *bif;
1732         struct bstp_port *bp;
1733         struct ifbpstpreq bpreq;
1734         char *buf, *outbuf;
1735         int count, buflen, len, error = 0;
1736
1737         count = 0;
1738         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
1739                 if ((bif->bif_flags & IFBIF_STP) != 0)
1740                         count++;
1741         }
1742
1743         buflen = sizeof(bpreq) * count;
1744         if (bifstp->ifbpstp_len == 0) {
1745                 bifstp->ifbpstp_len = buflen;
1746                 return (0);
1747         }
1748
1749         BRIDGE_UNLOCK(sc);
1750         outbuf = malloc(buflen, M_TEMP, M_WAITOK | M_ZERO);
1751         BRIDGE_LOCK(sc);
1752
1753         count = 0;
1754         buf = outbuf;
1755         len = min(bifstp->ifbpstp_len, buflen);
1756         bzero(&bpreq, sizeof(bpreq));
1757         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
1758                 if (len < sizeof(bpreq))
1759                         break;
1760
1761                 if ((bif->bif_flags & IFBIF_STP) == 0)
1762                         continue;
1763
1764                 bp = &bif->bif_stp;
1765                 bpreq.ifbp_portno = bif->bif_ifp->if_index & 0xfff;
1766                 bpreq.ifbp_fwd_trans = bp->bp_forward_transitions;
1767                 bpreq.ifbp_design_cost = bp->bp_desg_pv.pv_cost;
1768                 bpreq.ifbp_design_port = bp->bp_desg_pv.pv_port_id;
1769                 bpreq.ifbp_design_bridge = bp->bp_desg_pv.pv_dbridge_id;
1770                 bpreq.ifbp_design_root = bp->bp_desg_pv.pv_root_id;
1771
1772                 memcpy(buf, &bpreq, sizeof(bpreq));
1773                 count++;
1774                 buf += sizeof(bpreq);
1775                 len -= sizeof(bpreq);
1776         }
1777
1778         BRIDGE_UNLOCK(sc);
1779         bifstp->ifbpstp_len = sizeof(bpreq) * count;
1780         error = copyout(outbuf, bifstp->ifbpstp_req, bifstp->ifbpstp_len);
1781         BRIDGE_LOCK(sc);
1782         free(outbuf, M_TEMP);
1783         return (error);
1784 }
1785
1786 static int
1787 bridge_ioctl_sproto(struct bridge_softc *sc, void *arg)
1788 {
1789         struct ifbrparam *param = arg;
1790
1791         return (bstp_set_protocol(&sc->sc_stp, param->ifbrp_proto));
1792 }
1793
1794 static int
1795 bridge_ioctl_stxhc(struct bridge_softc *sc, void *arg)
1796 {
1797         struct ifbrparam *param = arg;
1798
1799         return (bstp_set_holdcount(&sc->sc_stp, param->ifbrp_txhc));
1800 }
1801
1802 /*
1803  * bridge_ifdetach:
1804  *
1805  *      Detach an interface from a bridge.  Called when a member
1806  *      interface is detaching.
1807  */
1808 static void
1809 bridge_ifdetach(void *arg __unused, struct ifnet *ifp)
1810 {
1811         struct bridge_softc *sc = ifp->if_bridge;
1812         struct bridge_iflist *bif;
1813
1814         if (ifp->if_flags & IFF_RENAMING)
1815                 return;
1816         if (V_bridge_cloner == NULL) {
1817                 /*
1818                  * This detach handler can be called after
1819                  * vnet_bridge_uninit().  Just return in that case.
1820                  */
1821                 return;
1822         }
1823         /* Check if the interface is a bridge member */
1824         if (sc != NULL) {
1825                 BRIDGE_LOCK(sc);
1826
1827                 bif = bridge_lookup_member_if(sc, ifp);
1828                 if (bif != NULL)
1829                         bridge_delete_member(sc, bif, 1);
1830
1831                 BRIDGE_UNLOCK(sc);
1832                 return;
1833         }
1834
1835         /* Check if the interface is a span port */
1836         BRIDGE_LIST_LOCK();
1837         LIST_FOREACH(sc, &V_bridge_list, sc_list) {
1838                 BRIDGE_LOCK(sc);
1839                 LIST_FOREACH(bif, &sc->sc_spanlist, bif_next)
1840                         if (ifp == bif->bif_ifp) {
1841                                 bridge_delete_span(sc, bif);
1842                                 break;
1843                         }
1844
1845                 BRIDGE_UNLOCK(sc);
1846         }
1847         BRIDGE_LIST_UNLOCK();
1848 }
1849
1850 /*
1851  * bridge_init:
1852  *
1853  *      Initialize a bridge interface.
1854  */
1855 static void
1856 bridge_init(void *xsc)
1857 {
1858         struct bridge_softc *sc = (struct bridge_softc *)xsc;
1859         struct ifnet *ifp = sc->sc_ifp;
1860
1861         if (ifp->if_drv_flags & IFF_DRV_RUNNING)
1862                 return;
1863
1864         BRIDGE_LOCK(sc);
1865         callout_reset(&sc->sc_brcallout, bridge_rtable_prune_period * hz,
1866             bridge_timer, sc);
1867
1868         ifp->if_drv_flags |= IFF_DRV_RUNNING;
1869         bstp_init(&sc->sc_stp);         /* Initialize Spanning Tree */
1870
1871         BRIDGE_UNLOCK(sc);
1872 }
1873
1874 /*
1875  * bridge_stop:
1876  *
1877  *      Stop the bridge interface.
1878  */
1879 static void
1880 bridge_stop(struct ifnet *ifp, int disable)
1881 {
1882         struct bridge_softc *sc = ifp->if_softc;
1883
1884         BRIDGE_LOCK_ASSERT(sc);
1885
1886         if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0)
1887                 return;
1888
1889         callout_stop(&sc->sc_brcallout);
1890         bstp_stop(&sc->sc_stp);
1891
1892         bridge_rtflush(sc, IFBF_FLUSHDYN);
1893
1894         ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
1895 }
1896
1897 /*
1898  * bridge_enqueue:
1899  *
1900  *      Enqueue a packet on a bridge member interface.
1901  *
1902  */
1903 static int
1904 bridge_enqueue(struct bridge_softc *sc, struct ifnet *dst_ifp, struct mbuf *m)
1905 {
1906         int len, err = 0;
1907         short mflags;
1908         struct mbuf *m0;
1909
1910         /* We may be sending a fragment so traverse the mbuf */
1911         for (; m; m = m0) {
1912                 m0 = m->m_nextpkt;
1913                 m->m_nextpkt = NULL;
1914                 len = m->m_pkthdr.len;
1915                 mflags = m->m_flags;
1916
1917                 /*
1918                  * If underlying interface can not do VLAN tag insertion itself
1919                  * then attach a packet tag that holds it.
1920                  */
1921                 if ((m->m_flags & M_VLANTAG) &&
1922                     (dst_ifp->if_capenable & IFCAP_VLAN_HWTAGGING) == 0) {
1923                         m = ether_vlanencap(m, m->m_pkthdr.ether_vtag);
1924                         if (m == NULL) {
1925                                 if_printf(dst_ifp,
1926                                     "unable to prepend VLAN header\n");
1927                                 if_inc_counter(dst_ifp, IFCOUNTER_OERRORS, 1);
1928                                 continue;
1929                         }
1930                         m->m_flags &= ~M_VLANTAG;
1931                 }
1932
1933                 M_ASSERTPKTHDR(m); /* We shouldn't transmit mbuf without pkthdr */
1934                 if ((err = dst_ifp->if_transmit(dst_ifp, m))) {
1935                         m_freem(m0);
1936                         if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
1937                         break;
1938                 }
1939
1940                 if_inc_counter(sc->sc_ifp, IFCOUNTER_OPACKETS, 1);
1941                 if_inc_counter(sc->sc_ifp, IFCOUNTER_OBYTES, len);
1942                 if (mflags & M_MCAST)
1943                         if_inc_counter(sc->sc_ifp, IFCOUNTER_OMCASTS, 1);
1944         }
1945
1946         return (err);
1947 }
1948
1949 /*
1950  * bridge_dummynet:
1951  *
1952  *      Receive a queued packet from dummynet and pass it on to the output
1953  *      interface.
1954  *
1955  *      The mbuf has the Ethernet header already attached.
1956  */
1957 static void
1958 bridge_dummynet(struct mbuf *m, struct ifnet *ifp)
1959 {
1960         struct bridge_softc *sc;
1961
1962         sc = ifp->if_bridge;
1963
1964         /*
1965          * The packet didnt originate from a member interface. This should only
1966          * ever happen if a member interface is removed while packets are
1967          * queued for it.
1968          */
1969         if (sc == NULL) {
1970                 m_freem(m);
1971                 return;
1972         }
1973
1974         if (PFIL_HOOKED_OUT(V_inet_pfil_head)
1975 #ifdef INET6
1976             || PFIL_HOOKED_OUT(V_inet6_pfil_head)
1977 #endif
1978             ) {
1979                 if (bridge_pfil(&m, sc->sc_ifp, ifp, PFIL_OUT) != 0)
1980                         return;
1981                 if (m == NULL)
1982                         return;
1983         }
1984
1985         bridge_enqueue(sc, ifp, m);
1986 }
1987
1988 /*
1989  * bridge_output:
1990  *
1991  *      Send output from a bridge member interface.  This
1992  *      performs the bridging function for locally originated
1993  *      packets.
1994  *
1995  *      The mbuf has the Ethernet header already attached.  We must
1996  *      enqueue or free the mbuf before returning.
1997  */
1998 static int
1999 bridge_output(struct ifnet *ifp, struct mbuf *m, struct sockaddr *sa,
2000     struct rtentry *rt)
2001 {
2002         struct ether_header *eh;
2003         struct ifnet *dst_if;
2004         struct bridge_softc *sc;
2005         uint16_t vlan;
2006
2007         if (m->m_len < ETHER_HDR_LEN) {
2008                 m = m_pullup(m, ETHER_HDR_LEN);
2009                 if (m == NULL)
2010                         return (0);
2011         }
2012
2013         eh = mtod(m, struct ether_header *);
2014         sc = ifp->if_bridge;
2015         vlan = VLANTAGOF(m);
2016
2017         BRIDGE_LOCK(sc);
2018
2019         /*
2020          * If bridge is down, but the original output interface is up,
2021          * go ahead and send out that interface.  Otherwise, the packet
2022          * is dropped below.
2023          */
2024         if ((sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) {
2025                 dst_if = ifp;
2026                 goto sendunicast;
2027         }
2028
2029         /*
2030          * If the packet is a multicast, or we don't know a better way to
2031          * get there, send to all interfaces.
2032          */
2033         if (ETHER_IS_MULTICAST(eh->ether_dhost))
2034                 dst_if = NULL;
2035         else
2036                 dst_if = bridge_rtlookup(sc, eh->ether_dhost, vlan);
2037         if (dst_if == NULL) {
2038                 struct bridge_iflist *bif;
2039                 struct mbuf *mc;
2040                 int error = 0, used = 0;
2041
2042                 bridge_span(sc, m);
2043
2044                 BRIDGE_LOCK2REF(sc, error);
2045                 if (error) {
2046                         m_freem(m);
2047                         return (0);
2048                 }
2049
2050                 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
2051                         dst_if = bif->bif_ifp;
2052
2053                         if (dst_if->if_type == IFT_GIF)
2054                                 continue;
2055                         if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
2056                                 continue;
2057
2058                         /*
2059                          * If this is not the original output interface,
2060                          * and the interface is participating in spanning
2061                          * tree, make sure the port is in a state that
2062                          * allows forwarding.
2063                          */
2064                         if (dst_if != ifp && (bif->bif_flags & IFBIF_STP) &&
2065                             bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
2066                                 continue;
2067
2068                         if (LIST_NEXT(bif, bif_next) == NULL) {
2069                                 used = 1;
2070                                 mc = m;
2071                         } else {
2072                                 mc = m_copypacket(m, M_NOWAIT);
2073                                 if (mc == NULL) {
2074                                         if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
2075                                         continue;
2076                                 }
2077                         }
2078
2079                         bridge_enqueue(sc, dst_if, mc);
2080                 }
2081                 if (used == 0)
2082                         m_freem(m);
2083                 BRIDGE_UNREF(sc);
2084                 return (0);
2085         }
2086
2087 sendunicast:
2088         /*
2089          * XXX Spanning tree consideration here?
2090          */
2091
2092         bridge_span(sc, m);
2093         if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) {
2094                 m_freem(m);
2095                 BRIDGE_UNLOCK(sc);
2096                 return (0);
2097         }
2098
2099         BRIDGE_UNLOCK(sc);
2100         bridge_enqueue(sc, dst_if, m);
2101         return (0);
2102 }
2103
2104 /*
2105  * bridge_transmit:
2106  *
2107  *      Do output on a bridge.
2108  *
2109  */
2110 static int
2111 bridge_transmit(struct ifnet *ifp, struct mbuf *m)
2112 {
2113         struct bridge_softc *sc;
2114         struct ether_header *eh;
2115         struct ifnet *dst_if;
2116         int error = 0;
2117
2118         sc = ifp->if_softc;
2119
2120         ETHER_BPF_MTAP(ifp, m);
2121
2122         eh = mtod(m, struct ether_header *);
2123
2124         BRIDGE_LOCK(sc);
2125         if (((m->m_flags & (M_BCAST|M_MCAST)) == 0) &&
2126             (dst_if = bridge_rtlookup(sc, eh->ether_dhost, 1)) != NULL) {
2127                 BRIDGE_UNLOCK(sc);
2128                 error = bridge_enqueue(sc, dst_if, m);
2129         } else
2130                 bridge_broadcast(sc, ifp, m, 0);
2131
2132         return (error);
2133 }
2134
2135 /*
2136  * The ifp->if_qflush entry point for if_bridge(4) is no-op.
2137  */
2138 static void
2139 bridge_qflush(struct ifnet *ifp __unused)
2140 {
2141 }
2142
2143 /*
2144  * bridge_forward:
2145  *
2146  *      The forwarding function of the bridge.
2147  *
2148  *      NOTE: Releases the lock on return.
2149  */
2150 static void
2151 bridge_forward(struct bridge_softc *sc, struct bridge_iflist *sbif,
2152     struct mbuf *m)
2153 {
2154         struct bridge_iflist *dbif;
2155         struct ifnet *src_if, *dst_if, *ifp;
2156         struct ether_header *eh;
2157         uint16_t vlan;
2158         uint8_t *dst;
2159         int error;
2160
2161         src_if = m->m_pkthdr.rcvif;
2162         ifp = sc->sc_ifp;
2163
2164         if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1);
2165         if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len);
2166         vlan = VLANTAGOF(m);
2167
2168         if ((sbif->bif_flags & IFBIF_STP) &&
2169             sbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
2170                 goto drop;
2171
2172         eh = mtod(m, struct ether_header *);
2173         dst = eh->ether_dhost;
2174
2175         /* If the interface is learning, record the address. */
2176         if (sbif->bif_flags & IFBIF_LEARNING) {
2177                 error = bridge_rtupdate(sc, eh->ether_shost, vlan,
2178                     sbif, 0, IFBAF_DYNAMIC);
2179                 /*
2180                  * If the interface has addresses limits then deny any source
2181                  * that is not in the cache.
2182                  */
2183                 if (error && sbif->bif_addrmax)
2184                         goto drop;
2185         }
2186
2187         if ((sbif->bif_flags & IFBIF_STP) != 0 &&
2188             sbif->bif_stp.bp_state == BSTP_IFSTATE_LEARNING)
2189                 goto drop;
2190
2191         /*
2192          * At this point, the port either doesn't participate
2193          * in spanning tree or it is in the forwarding state.
2194          */
2195
2196         /*
2197          * If the packet is unicast, destined for someone on
2198          * "this" side of the bridge, drop it.
2199          */
2200         if ((m->m_flags & (M_BCAST|M_MCAST)) == 0) {
2201                 dst_if = bridge_rtlookup(sc, dst, vlan);
2202                 if (src_if == dst_if)
2203                         goto drop;
2204         } else {
2205                 /*
2206                  * Check if its a reserved multicast address, any address
2207                  * listed in 802.1D section 7.12.6 may not be forwarded by the
2208                  * bridge.
2209                  * This is currently 01-80-C2-00-00-00 to 01-80-C2-00-00-0F
2210                  */
2211                 if (dst[0] == 0x01 && dst[1] == 0x80 &&
2212                     dst[2] == 0xc2 && dst[3] == 0x00 &&
2213                     dst[4] == 0x00 && dst[5] <= 0x0f)
2214                         goto drop;
2215
2216                 /* ...forward it to all interfaces. */
2217                 if_inc_counter(ifp, IFCOUNTER_IMCASTS, 1);
2218                 dst_if = NULL;
2219         }
2220
2221         /*
2222          * If we have a destination interface which is a member of our bridge,
2223          * OR this is a unicast packet, push it through the bpf(4) machinery.
2224          * For broadcast or multicast packets, don't bother because it will
2225          * be reinjected into ether_input. We do this before we pass the packets
2226          * through the pfil(9) framework, as it is possible that pfil(9) will
2227          * drop the packet, or possibly modify it, making it difficult to debug
2228          * firewall issues on the bridge.
2229          */
2230         if (dst_if != NULL || (m->m_flags & (M_BCAST | M_MCAST)) == 0)
2231                 ETHER_BPF_MTAP(ifp, m);
2232
2233         /* run the packet filter */
2234         if (PFIL_HOOKED_IN(V_inet_pfil_head)
2235 #ifdef INET6
2236             || PFIL_HOOKED_IN(V_inet6_pfil_head)
2237 #endif
2238             ) {
2239                 BRIDGE_UNLOCK(sc);
2240                 if (bridge_pfil(&m, ifp, src_if, PFIL_IN) != 0)
2241                         return;
2242                 if (m == NULL)
2243                         return;
2244                 BRIDGE_LOCK(sc);
2245         }
2246
2247         if (dst_if == NULL) {
2248                 bridge_broadcast(sc, src_if, m, 1);
2249                 return;
2250         }
2251
2252         /*
2253          * At this point, we're dealing with a unicast frame
2254          * going to a different interface.
2255          */
2256         if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
2257                 goto drop;
2258
2259         dbif = bridge_lookup_member_if(sc, dst_if);
2260         if (dbif == NULL)
2261                 /* Not a member of the bridge (anymore?) */
2262                 goto drop;
2263
2264         /* Private segments can not talk to each other */
2265         if (sbif->bif_flags & dbif->bif_flags & IFBIF_PRIVATE)
2266                 goto drop;
2267
2268         if ((dbif->bif_flags & IFBIF_STP) &&
2269             dbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
2270                 goto drop;
2271
2272         BRIDGE_UNLOCK(sc);
2273
2274         if (PFIL_HOOKED_OUT(V_inet_pfil_head)
2275 #ifdef INET6
2276             || PFIL_HOOKED_OUT(V_inet6_pfil_head)
2277 #endif
2278             ) {
2279                 if (bridge_pfil(&m, ifp, dst_if, PFIL_OUT) != 0)
2280                         return;
2281                 if (m == NULL)
2282                         return;
2283         }
2284
2285         bridge_enqueue(sc, dst_if, m);
2286         return;
2287
2288 drop:
2289         BRIDGE_UNLOCK(sc);
2290         m_freem(m);
2291 }
2292
2293 /*
2294  * bridge_input:
2295  *
2296  *      Receive input from a member interface.  Queue the packet for
2297  *      bridging if it is not for us.
2298  */
2299 static struct mbuf *
2300 bridge_input(struct ifnet *ifp, struct mbuf *m)
2301 {
2302         struct bridge_softc *sc = ifp->if_bridge;
2303         struct bridge_iflist *bif, *bif2;
2304         struct ifnet *bifp;
2305         struct ether_header *eh;
2306         struct mbuf *mc, *mc2;
2307         uint16_t vlan;
2308         int error;
2309
2310         if ((sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING) == 0)
2311                 return (m);
2312
2313         bifp = sc->sc_ifp;
2314         vlan = VLANTAGOF(m);
2315
2316         /*
2317          * Implement support for bridge monitoring. If this flag has been
2318          * set on this interface, discard the packet once we push it through
2319          * the bpf(4) machinery, but before we do, increment the byte and
2320          * packet counters associated with this interface.
2321          */
2322         if ((bifp->if_flags & IFF_MONITOR) != 0) {
2323                 m->m_pkthdr.rcvif  = bifp;
2324                 ETHER_BPF_MTAP(bifp, m);
2325                 if_inc_counter(bifp, IFCOUNTER_IPACKETS, 1);
2326                 if_inc_counter(bifp, IFCOUNTER_IBYTES, m->m_pkthdr.len);
2327                 m_freem(m);
2328                 return (NULL);
2329         }
2330         BRIDGE_LOCK(sc);
2331         bif = bridge_lookup_member_if(sc, ifp);
2332         if (bif == NULL) {
2333                 BRIDGE_UNLOCK(sc);
2334                 return (m);
2335         }
2336
2337         eh = mtod(m, struct ether_header *);
2338
2339         bridge_span(sc, m);
2340
2341         if (m->m_flags & (M_BCAST|M_MCAST)) {
2342                 /* Tap off 802.1D packets; they do not get forwarded. */
2343                 if (memcmp(eh->ether_dhost, bstp_etheraddr,
2344                     ETHER_ADDR_LEN) == 0) {
2345                         bstp_input(&bif->bif_stp, ifp, m); /* consumes mbuf */
2346                         BRIDGE_UNLOCK(sc);
2347                         return (NULL);
2348                 }
2349
2350                 if ((bif->bif_flags & IFBIF_STP) &&
2351                     bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) {
2352                         BRIDGE_UNLOCK(sc);
2353                         return (m);
2354                 }
2355
2356                 /*
2357                  * Make a deep copy of the packet and enqueue the copy
2358                  * for bridge processing; return the original packet for
2359                  * local processing.
2360                  */
2361                 mc = m_dup(m, M_NOWAIT);
2362                 if (mc == NULL) {
2363                         BRIDGE_UNLOCK(sc);
2364                         return (m);
2365                 }
2366
2367                 /* Perform the bridge forwarding function with the copy. */
2368                 bridge_forward(sc, bif, mc);
2369
2370                 /*
2371                  * Reinject the mbuf as arriving on the bridge so we have a
2372                  * chance at claiming multicast packets. We can not loop back
2373                  * here from ether_input as a bridge is never a member of a
2374                  * bridge.
2375                  */
2376                 KASSERT(bifp->if_bridge == NULL,
2377                     ("loop created in bridge_input"));
2378                 mc2 = m_dup(m, M_NOWAIT);
2379                 if (mc2 != NULL) {
2380                         /* Keep the layer3 header aligned */
2381                         int i = min(mc2->m_pkthdr.len, max_protohdr);
2382                         mc2 = m_copyup(mc2, i, ETHER_ALIGN);
2383                 }
2384                 if (mc2 != NULL) {
2385                         mc2->m_pkthdr.rcvif = bifp;
2386                         (*bifp->if_input)(bifp, mc2);
2387                 }
2388
2389                 /* Return the original packet for local processing. */
2390                 return (m);
2391         }
2392
2393         if ((bif->bif_flags & IFBIF_STP) &&
2394             bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) {
2395                 BRIDGE_UNLOCK(sc);
2396                 return (m);
2397         }
2398
2399 #if (defined(INET) || defined(INET6))
2400 #   define OR_CARP_CHECK_WE_ARE_DST(iface) \
2401         || ((iface)->if_carp \
2402             && (*carp_forus_p)((iface), eh->ether_dhost))
2403 #   define OR_CARP_CHECK_WE_ARE_SRC(iface) \
2404         || ((iface)->if_carp \
2405             && (*carp_forus_p)((iface), eh->ether_shost))
2406 #else
2407 #   define OR_CARP_CHECK_WE_ARE_DST(iface)
2408 #   define OR_CARP_CHECK_WE_ARE_SRC(iface)
2409 #endif
2410
2411 #ifdef INET6
2412 #   define OR_PFIL_HOOKED_INET6 \
2413         || PFIL_HOOKED_IN(V_inet6_pfil_head)
2414 #else
2415 #   define OR_PFIL_HOOKED_INET6
2416 #endif
2417
2418 #define GRAB_OUR_PACKETS(iface) \
2419         if ((iface)->if_type == IFT_GIF) \
2420                 continue; \
2421         /* It is destined for us. */ \
2422         if (memcmp(IF_LLADDR((iface)), eh->ether_dhost,  ETHER_ADDR_LEN) == 0 \
2423             OR_CARP_CHECK_WE_ARE_DST((iface))                           \
2424             ) {                                                         \
2425                 if (bif->bif_flags & IFBIF_LEARNING) {                  \
2426                         error = bridge_rtupdate(sc, eh->ether_shost,    \
2427                             vlan, bif, 0, IFBAF_DYNAMIC);               \
2428                         if (error && bif->bif_addrmax) {                \
2429                                 BRIDGE_UNLOCK(sc);                      \
2430                                 m_freem(m);                             \
2431                                 return (NULL);                          \
2432                         }                                               \
2433                 }                                                       \
2434                 m->m_pkthdr.rcvif = iface;                              \
2435                 if ((iface) == ifp) {                                   \
2436                         /* Skip bridge processing... src == dest */     \
2437                         BRIDGE_UNLOCK(sc);                              \
2438                         return (m);                                     \
2439                 }                                                       \
2440                 /* It's passing over or to the bridge, locally. */      \
2441                 ETHER_BPF_MTAP(bifp, m);                                \
2442                 if_inc_counter(bifp, IFCOUNTER_IPACKETS, 1);            \
2443                 if_inc_counter(bifp, IFCOUNTER_IBYTES, m->m_pkthdr.len); \
2444                 /* Filter on the physical interface. */                 \
2445                 if (V_pfil_local_phys && (PFIL_HOOKED_IN(V_inet_pfil_head) \
2446                      OR_PFIL_HOOKED_INET6)) {                           \
2447                         if (bridge_pfil(&m, NULL, ifp,                  \
2448                             PFIL_IN) != 0 || m == NULL) {               \
2449                                 BRIDGE_UNLOCK(sc);                      \
2450                                 return (NULL);                          \
2451                         }                                               \
2452                 }                                                       \
2453                 BRIDGE_UNLOCK(sc);                                      \
2454                 return (m);                                             \
2455         }                                                               \
2456                                                                         \
2457         /* We just received a packet that we sent out. */               \
2458         if (memcmp(IF_LLADDR((iface)), eh->ether_shost, ETHER_ADDR_LEN) == 0 \
2459             OR_CARP_CHECK_WE_ARE_SRC((iface))                   \
2460             ) {                                                         \
2461                 BRIDGE_UNLOCK(sc);                                      \
2462                 m_freem(m);                                             \
2463                 return (NULL);                                          \
2464         }
2465
2466         /*
2467          * Unicast.  Make sure it's not for the bridge.
2468          */
2469         do { GRAB_OUR_PACKETS(bifp) } while (0);
2470
2471         /*
2472          * Give a chance for ifp at first priority. This will help when the
2473          * packet comes through the interface like VLAN's with the same MACs
2474          * on several interfaces from the same bridge. This also will save
2475          * some CPU cycles in case the destination interface and the input
2476          * interface (eq ifp) are the same.
2477          */
2478         do { GRAB_OUR_PACKETS(ifp) } while (0);
2479
2480         /* Now check the all bridge members. */
2481         LIST_FOREACH(bif2, &sc->sc_iflist, bif_next) {
2482                 GRAB_OUR_PACKETS(bif2->bif_ifp)
2483         }
2484
2485 #undef OR_CARP_CHECK_WE_ARE_DST
2486 #undef OR_CARP_CHECK_WE_ARE_SRC
2487 #undef OR_PFIL_HOOKED_INET6
2488 #undef GRAB_OUR_PACKETS
2489
2490         /* Perform the bridge forwarding function. */
2491         bridge_forward(sc, bif, m);
2492
2493         return (NULL);
2494 }
2495
2496 /*
2497  * bridge_broadcast:
2498  *
2499  *      Send a frame to all interfaces that are members of
2500  *      the bridge, except for the one on which the packet
2501  *      arrived.
2502  *
2503  *      NOTE: Releases the lock on return.
2504  */
2505 static void
2506 bridge_broadcast(struct bridge_softc *sc, struct ifnet *src_if,
2507     struct mbuf *m, int runfilt)
2508 {
2509         struct bridge_iflist *dbif, *sbif;
2510         struct mbuf *mc;
2511         struct ifnet *dst_if;
2512         int error = 0, used = 0, i;
2513
2514         sbif = bridge_lookup_member_if(sc, src_if);
2515
2516         BRIDGE_LOCK2REF(sc, error);
2517         if (error) {
2518                 m_freem(m);
2519                 return;
2520         }
2521
2522         /* Filter on the bridge interface before broadcasting */
2523         if (runfilt && (PFIL_HOOKED_OUT(V_inet_pfil_head)
2524 #ifdef INET6
2525             || PFIL_HOOKED_OUT(V_inet6_pfil_head)
2526 #endif
2527             )) {
2528                 if (bridge_pfil(&m, sc->sc_ifp, NULL, PFIL_OUT) != 0)
2529                         goto out;
2530                 if (m == NULL)
2531                         goto out;
2532         }
2533
2534         LIST_FOREACH(dbif, &sc->sc_iflist, bif_next) {
2535                 dst_if = dbif->bif_ifp;
2536                 if (dst_if == src_if)
2537                         continue;
2538
2539                 /* Private segments can not talk to each other */
2540                 if (sbif && (sbif->bif_flags & dbif->bif_flags & IFBIF_PRIVATE))
2541                         continue;
2542
2543                 if ((dbif->bif_flags & IFBIF_STP) &&
2544                     dbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING)
2545                         continue;
2546
2547                 if ((dbif->bif_flags & IFBIF_DISCOVER) == 0 &&
2548                     (m->m_flags & (M_BCAST|M_MCAST)) == 0)
2549                         continue;
2550
2551                 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
2552                         continue;
2553
2554                 if (LIST_NEXT(dbif, bif_next) == NULL) {
2555                         mc = m;
2556                         used = 1;
2557                 } else {
2558                         mc = m_dup(m, M_NOWAIT);
2559                         if (mc == NULL) {
2560                                 if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
2561                                 continue;
2562                         }
2563                 }
2564
2565                 /*
2566                  * Filter on the output interface. Pass a NULL bridge interface
2567                  * pointer so we do not redundantly filter on the bridge for
2568                  * each interface we broadcast on.
2569                  */
2570                 if (runfilt && (PFIL_HOOKED_OUT(V_inet_pfil_head)
2571 #ifdef INET6
2572                     || PFIL_HOOKED_OUT(V_inet6_pfil_head)
2573 #endif
2574                     )) {
2575                         if (used == 0) {
2576                                 /* Keep the layer3 header aligned */
2577                                 i = min(mc->m_pkthdr.len, max_protohdr);
2578                                 mc = m_copyup(mc, i, ETHER_ALIGN);
2579                                 if (mc == NULL) {
2580                                         if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
2581                                         continue;
2582                                 }
2583                         }
2584                         if (bridge_pfil(&mc, NULL, dst_if, PFIL_OUT) != 0)
2585                                 continue;
2586                         if (mc == NULL)
2587                                 continue;
2588                 }
2589
2590                 bridge_enqueue(sc, dst_if, mc);
2591         }
2592         if (used == 0)
2593                 m_freem(m);
2594
2595 out:
2596         BRIDGE_UNREF(sc);
2597 }
2598
2599 /*
2600  * bridge_span:
2601  *
2602  *      Duplicate a packet out one or more interfaces that are in span mode,
2603  *      the original mbuf is unmodified.
2604  */
2605 static void
2606 bridge_span(struct bridge_softc *sc, struct mbuf *m)
2607 {
2608         struct bridge_iflist *bif;
2609         struct ifnet *dst_if;
2610         struct mbuf *mc;
2611
2612         if (LIST_EMPTY(&sc->sc_spanlist))
2613                 return;
2614
2615         LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) {
2616                 dst_if = bif->bif_ifp;
2617
2618                 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0)
2619                         continue;
2620
2621                 mc = m_copypacket(m, M_NOWAIT);
2622                 if (mc == NULL) {
2623                         if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1);
2624                         continue;
2625                 }
2626
2627                 bridge_enqueue(sc, dst_if, mc);
2628         }
2629 }
2630
2631 /*
2632  * bridge_rtupdate:
2633  *
2634  *      Add a bridge routing entry.
2635  */
2636 static int
2637 bridge_rtupdate(struct bridge_softc *sc, const uint8_t *dst, uint16_t vlan,
2638     struct bridge_iflist *bif, int setflags, uint8_t flags)
2639 {
2640         struct bridge_rtnode *brt;
2641         int error;
2642
2643         BRIDGE_LOCK_ASSERT(sc);
2644
2645         /* Check the source address is valid and not multicast. */
2646         if (ETHER_IS_MULTICAST(dst) ||
2647             (dst[0] == 0 && dst[1] == 0 && dst[2] == 0 &&
2648              dst[3] == 0 && dst[4] == 0 && dst[5] == 0) != 0)
2649                 return (EINVAL);
2650
2651         /* 802.1p frames map to vlan 1 */
2652         if (vlan == 0)
2653                 vlan = 1;
2654
2655         /*
2656          * A route for this destination might already exist.  If so,
2657          * update it, otherwise create a new one.
2658          */
2659         if ((brt = bridge_rtnode_lookup(sc, dst, vlan)) == NULL) {
2660                 if (sc->sc_brtcnt >= sc->sc_brtmax) {
2661                         sc->sc_brtexceeded++;
2662                         return (ENOSPC);
2663                 }
2664                 /* Check per interface address limits (if enabled) */
2665                 if (bif->bif_addrmax && bif->bif_addrcnt >= bif->bif_addrmax) {
2666                         bif->bif_addrexceeded++;
2667                         return (ENOSPC);
2668                 }
2669
2670                 /*
2671                  * Allocate a new bridge forwarding node, and
2672                  * initialize the expiration time and Ethernet
2673                  * address.
2674                  */
2675                 brt = uma_zalloc(V_bridge_rtnode_zone, M_NOWAIT | M_ZERO);
2676                 if (brt == NULL)
2677                         return (ENOMEM);
2678
2679                 if (bif->bif_flags & IFBIF_STICKY)
2680                         brt->brt_flags = IFBAF_STICKY;
2681                 else
2682                         brt->brt_flags = IFBAF_DYNAMIC;
2683
2684                 memcpy(brt->brt_addr, dst, ETHER_ADDR_LEN);
2685                 brt->brt_vlan = vlan;
2686
2687                 if ((error = bridge_rtnode_insert(sc, brt)) != 0) {
2688                         uma_zfree(V_bridge_rtnode_zone, brt);
2689                         return (error);
2690                 }
2691                 brt->brt_dst = bif;
2692                 bif->bif_addrcnt++;
2693         }
2694
2695         if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC &&
2696             brt->brt_dst != bif) {
2697                 brt->brt_dst->bif_addrcnt--;
2698                 brt->brt_dst = bif;
2699                 brt->brt_dst->bif_addrcnt++;
2700         }
2701
2702         if ((flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
2703                 brt->brt_expire = time_uptime + sc->sc_brttimeout;
2704         if (setflags)
2705                 brt->brt_flags = flags;
2706
2707         return (0);
2708 }
2709
2710 /*
2711  * bridge_rtlookup:
2712  *
2713  *      Lookup the destination interface for an address.
2714  */
2715 static struct ifnet *
2716 bridge_rtlookup(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan)
2717 {
2718         struct bridge_rtnode *brt;
2719
2720         BRIDGE_LOCK_ASSERT(sc);
2721
2722         if ((brt = bridge_rtnode_lookup(sc, addr, vlan)) == NULL)
2723                 return (NULL);
2724
2725         return (brt->brt_ifp);
2726 }
2727
2728 /*
2729  * bridge_rttrim:
2730  *
2731  *      Trim the routine table so that we have a number
2732  *      of routing entries less than or equal to the
2733  *      maximum number.
2734  */
2735 static void
2736 bridge_rttrim(struct bridge_softc *sc)
2737 {
2738         struct bridge_rtnode *brt, *nbrt;
2739
2740         BRIDGE_LOCK_ASSERT(sc);
2741
2742         /* Make sure we actually need to do this. */
2743         if (sc->sc_brtcnt <= sc->sc_brtmax)
2744                 return;
2745
2746         /* Force an aging cycle; this might trim enough addresses. */
2747         bridge_rtage(sc);
2748         if (sc->sc_brtcnt <= sc->sc_brtmax)
2749                 return;
2750
2751         LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
2752                 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) {
2753                         bridge_rtnode_destroy(sc, brt);
2754                         if (sc->sc_brtcnt <= sc->sc_brtmax)
2755                                 return;
2756                 }
2757         }
2758 }
2759
2760 /*
2761  * bridge_timer:
2762  *
2763  *      Aging timer for the bridge.
2764  */
2765 static void
2766 bridge_timer(void *arg)
2767 {
2768         struct bridge_softc *sc = arg;
2769
2770         BRIDGE_LOCK_ASSERT(sc);
2771
2772         /* Destruction of rtnodes requires a proper vnet context */
2773         CURVNET_SET(sc->sc_ifp->if_vnet);
2774         bridge_rtage(sc);
2775
2776         if (sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING)
2777                 callout_reset(&sc->sc_brcallout,
2778                     bridge_rtable_prune_period * hz, bridge_timer, sc);
2779         CURVNET_RESTORE();
2780 }
2781
2782 /*
2783  * bridge_rtage:
2784  *
2785  *      Perform an aging cycle.
2786  */
2787 static void
2788 bridge_rtage(struct bridge_softc *sc)
2789 {
2790         struct bridge_rtnode *brt, *nbrt;
2791
2792         BRIDGE_LOCK_ASSERT(sc);
2793
2794         LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
2795                 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) {
2796                         if (time_uptime >= brt->brt_expire)
2797                                 bridge_rtnode_destroy(sc, brt);
2798                 }
2799         }
2800 }
2801
2802 /*
2803  * bridge_rtflush:
2804  *
2805  *      Remove all dynamic addresses from the bridge.
2806  */
2807 static void
2808 bridge_rtflush(struct bridge_softc *sc, int full)
2809 {
2810         struct bridge_rtnode *brt, *nbrt;
2811
2812         BRIDGE_LOCK_ASSERT(sc);
2813
2814         LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
2815                 if (full || (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
2816                         bridge_rtnode_destroy(sc, brt);
2817         }
2818 }
2819
2820 /*
2821  * bridge_rtdaddr:
2822  *
2823  *      Remove an address from the table.
2824  */
2825 static int
2826 bridge_rtdaddr(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan)
2827 {
2828         struct bridge_rtnode *brt;
2829         int found = 0;
2830
2831         BRIDGE_LOCK_ASSERT(sc);
2832
2833         /*
2834          * If vlan is zero then we want to delete for all vlans so the lookup
2835          * may return more than one.
2836          */
2837         while ((brt = bridge_rtnode_lookup(sc, addr, vlan)) != NULL) {
2838                 bridge_rtnode_destroy(sc, brt);
2839                 found = 1;
2840         }
2841
2842         return (found ? 0 : ENOENT);
2843 }
2844
2845 /*
2846  * bridge_rtdelete:
2847  *
2848  *      Delete routes to a speicifc member interface.
2849  */
2850 static void
2851 bridge_rtdelete(struct bridge_softc *sc, struct ifnet *ifp, int full)
2852 {
2853         struct bridge_rtnode *brt, *nbrt;
2854
2855         BRIDGE_LOCK_ASSERT(sc);
2856
2857         LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) {
2858                 if (brt->brt_ifp == ifp && (full ||
2859                             (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC))
2860                         bridge_rtnode_destroy(sc, brt);
2861         }
2862 }
2863
2864 /*
2865  * bridge_rtable_init:
2866  *
2867  *      Initialize the route table for this bridge.
2868  */
2869 static void
2870 bridge_rtable_init(struct bridge_softc *sc)
2871 {
2872         int i;
2873
2874         sc->sc_rthash = malloc(sizeof(*sc->sc_rthash) * BRIDGE_RTHASH_SIZE,
2875             M_DEVBUF, M_WAITOK);
2876
2877         for (i = 0; i < BRIDGE_RTHASH_SIZE; i++)
2878                 LIST_INIT(&sc->sc_rthash[i]);
2879
2880         sc->sc_rthash_key = arc4random();
2881         LIST_INIT(&sc->sc_rtlist);
2882 }
2883
2884 /*
2885  * bridge_rtable_fini:
2886  *
2887  *      Deconstruct the route table for this bridge.
2888  */
2889 static void
2890 bridge_rtable_fini(struct bridge_softc *sc)
2891 {
2892
2893         KASSERT(sc->sc_brtcnt == 0,
2894             ("%s: %d bridge routes referenced", __func__, sc->sc_brtcnt));
2895         free(sc->sc_rthash, M_DEVBUF);
2896 }
2897
2898 /*
2899  * The following hash function is adapted from "Hash Functions" by Bob Jenkins
2900  * ("Algorithm Alley", Dr. Dobbs Journal, September 1997).
2901  */
2902 #define mix(a, b, c)                                                    \
2903 do {                                                                    \
2904         a -= b; a -= c; a ^= (c >> 13);                                 \
2905         b -= c; b -= a; b ^= (a << 8);                                  \
2906         c -= a; c -= b; c ^= (b >> 13);                                 \
2907         a -= b; a -= c; a ^= (c >> 12);                                 \
2908         b -= c; b -= a; b ^= (a << 16);                                 \
2909         c -= a; c -= b; c ^= (b >> 5);                                  \
2910         a -= b; a -= c; a ^= (c >> 3);                                  \
2911         b -= c; b -= a; b ^= (a << 10);                                 \
2912         c -= a; c -= b; c ^= (b >> 15);                                 \
2913 } while (/*CONSTCOND*/0)
2914
2915 static __inline uint32_t
2916 bridge_rthash(struct bridge_softc *sc, const uint8_t *addr)
2917 {
2918         uint32_t a = 0x9e3779b9, b = 0x9e3779b9, c = sc->sc_rthash_key;
2919
2920         b += addr[5] << 8;
2921         b += addr[4];
2922         a += addr[3] << 24;
2923         a += addr[2] << 16;
2924         a += addr[1] << 8;
2925         a += addr[0];
2926
2927         mix(a, b, c);
2928
2929         return (c & BRIDGE_RTHASH_MASK);
2930 }
2931
2932 #undef mix
2933
2934 static int
2935 bridge_rtnode_addr_cmp(const uint8_t *a, const uint8_t *b)
2936 {
2937         int i, d;
2938
2939         for (i = 0, d = 0; i < ETHER_ADDR_LEN && d == 0; i++) {
2940                 d = ((int)a[i]) - ((int)b[i]);
2941         }
2942
2943         return (d);
2944 }
2945
2946 /*
2947  * bridge_rtnode_lookup:
2948  *
2949  *      Look up a bridge route node for the specified destination. Compare the
2950  *      vlan id or if zero then just return the first match.
2951  */
2952 static struct bridge_rtnode *
2953 bridge_rtnode_lookup(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan)
2954 {
2955         struct bridge_rtnode *brt;
2956         uint32_t hash;
2957         int dir;
2958
2959         BRIDGE_LOCK_ASSERT(sc);
2960
2961         hash = bridge_rthash(sc, addr);
2962         LIST_FOREACH(brt, &sc->sc_rthash[hash], brt_hash) {
2963                 dir = bridge_rtnode_addr_cmp(addr, brt->brt_addr);
2964                 if (dir == 0 && (brt->brt_vlan == vlan || vlan == 0))
2965                         return (brt);
2966                 if (dir > 0)
2967                         return (NULL);
2968         }
2969
2970         return (NULL);
2971 }
2972
2973 /*
2974  * bridge_rtnode_insert:
2975  *
2976  *      Insert the specified bridge node into the route table.  We
2977  *      assume the entry is not already in the table.
2978  */
2979 static int
2980 bridge_rtnode_insert(struct bridge_softc *sc, struct bridge_rtnode *brt)
2981 {
2982         struct bridge_rtnode *lbrt;
2983         uint32_t hash;
2984         int dir;
2985
2986         BRIDGE_LOCK_ASSERT(sc);
2987
2988         hash = bridge_rthash(sc, brt->brt_addr);
2989
2990         lbrt = LIST_FIRST(&sc->sc_rthash[hash]);
2991         if (lbrt == NULL) {
2992                 LIST_INSERT_HEAD(&sc->sc_rthash[hash], brt, brt_hash);
2993                 goto out;
2994         }
2995
2996         do {
2997                 dir = bridge_rtnode_addr_cmp(brt->brt_addr, lbrt->brt_addr);
2998                 if (dir == 0 && brt->brt_vlan == lbrt->brt_vlan)
2999                         return (EEXIST);
3000                 if (dir > 0) {
3001                         LIST_INSERT_BEFORE(lbrt, brt, brt_hash);
3002                         goto out;
3003                 }
3004                 if (LIST_NEXT(lbrt, brt_hash) == NULL) {
3005                         LIST_INSERT_AFTER(lbrt, brt, brt_hash);
3006                         goto out;
3007                 }
3008                 lbrt = LIST_NEXT(lbrt, brt_hash);
3009         } while (lbrt != NULL);
3010
3011 #ifdef DIAGNOSTIC
3012         panic("bridge_rtnode_insert: impossible");
3013 #endif
3014
3015 out:
3016         LIST_INSERT_HEAD(&sc->sc_rtlist, brt, brt_list);
3017         sc->sc_brtcnt++;
3018
3019         return (0);
3020 }
3021
3022 /*
3023  * bridge_rtnode_destroy:
3024  *
3025  *      Destroy a bridge rtnode.
3026  */
3027 static void
3028 bridge_rtnode_destroy(struct bridge_softc *sc, struct bridge_rtnode *brt)
3029 {
3030         BRIDGE_LOCK_ASSERT(sc);
3031
3032         LIST_REMOVE(brt, brt_hash);
3033
3034         LIST_REMOVE(brt, brt_list);
3035         sc->sc_brtcnt--;
3036         brt->brt_dst->bif_addrcnt--;
3037         uma_zfree(V_bridge_rtnode_zone, brt);
3038 }
3039
3040 /*
3041  * bridge_rtable_expire:
3042  *
3043  *      Set the expiry time for all routes on an interface.
3044  */
3045 static void
3046 bridge_rtable_expire(struct ifnet *ifp, int age)
3047 {
3048         struct bridge_softc *sc = ifp->if_bridge;
3049         struct bridge_rtnode *brt;
3050
3051         CURVNET_SET(ifp->if_vnet);
3052         BRIDGE_LOCK(sc);
3053
3054         /*
3055          * If the age is zero then flush, otherwise set all the expiry times to
3056          * age for the interface
3057          */
3058         if (age == 0)
3059                 bridge_rtdelete(sc, ifp, IFBF_FLUSHDYN);
3060         else {
3061                 LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) {
3062                         /* Cap the expiry time to 'age' */
3063                         if (brt->brt_ifp == ifp &&
3064                             brt->brt_expire > time_uptime + age &&
3065                             (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)
3066                                 brt->brt_expire = time_uptime + age;
3067                 }
3068         }
3069         BRIDGE_UNLOCK(sc);
3070         CURVNET_RESTORE();
3071 }
3072
3073 /*
3074  * bridge_state_change:
3075  *
3076  *      Callback from the bridgestp code when a port changes states.
3077  */
3078 static void
3079 bridge_state_change(struct ifnet *ifp, int state)
3080 {
3081         struct bridge_softc *sc = ifp->if_bridge;
3082         static const char *stpstates[] = {
3083                 "disabled",
3084                 "listening",
3085                 "learning",
3086                 "forwarding",
3087                 "blocking",
3088                 "discarding"
3089         };
3090
3091         CURVNET_SET(ifp->if_vnet);
3092         if (V_log_stp)
3093                 log(LOG_NOTICE, "%s: state changed to %s on %s\n",
3094                     sc->sc_ifp->if_xname, stpstates[state], ifp->if_xname);
3095         CURVNET_RESTORE();
3096 }
3097
3098 /*
3099  * Send bridge packets through pfil if they are one of the types pfil can deal
3100  * with, or if they are ARP or REVARP.  (pfil will pass ARP and REVARP without
3101  * question.) If *bifp or *ifp are NULL then packet filtering is skipped for
3102  * that interface.
3103  */
3104 static int
3105 bridge_pfil(struct mbuf **mp, struct ifnet *bifp, struct ifnet *ifp, int dir)
3106 {
3107         int snap, error, i, hlen;
3108         struct ether_header *eh1, eh2;
3109         struct ip *ip;
3110         struct llc llc1;
3111         u_int16_t ether_type;
3112         pfil_return_t rv;
3113
3114         snap = 0;
3115         error = -1;     /* Default error if not error == 0 */
3116
3117 #if 0
3118         /* we may return with the IP fields swapped, ensure its not shared */
3119         KASSERT(M_WRITABLE(*mp), ("%s: modifying a shared mbuf", __func__));
3120 #endif
3121
3122         if (V_pfil_bridge == 0 && V_pfil_member == 0 && V_pfil_ipfw == 0)
3123                 return (0); /* filtering is disabled */
3124
3125         i = min((*mp)->m_pkthdr.len, max_protohdr);
3126         if ((*mp)->m_len < i) {
3127             *mp = m_pullup(*mp, i);
3128             if (*mp == NULL) {
3129                 printf("%s: m_pullup failed\n", __func__);
3130                 return (-1);
3131             }
3132         }
3133
3134         eh1 = mtod(*mp, struct ether_header *);
3135         ether_type = ntohs(eh1->ether_type);
3136
3137         /*
3138          * Check for SNAP/LLC.
3139          */
3140         if (ether_type < ETHERMTU) {
3141                 struct llc *llc2 = (struct llc *)(eh1 + 1);
3142
3143                 if ((*mp)->m_len >= ETHER_HDR_LEN + 8 &&
3144                     llc2->llc_dsap == LLC_SNAP_LSAP &&
3145                     llc2->llc_ssap == LLC_SNAP_LSAP &&
3146                     llc2->llc_control == LLC_UI) {
3147                         ether_type = htons(llc2->llc_un.type_snap.ether_type);
3148                         snap = 1;
3149                 }
3150         }
3151
3152         /*
3153          * If we're trying to filter bridge traffic, don't look at anything
3154          * other than IP and ARP traffic.  If the filter doesn't understand
3155          * IPv6, don't allow IPv6 through the bridge either.  This is lame
3156          * since if we really wanted, say, an AppleTalk filter, we are hosed,
3157          * but of course we don't have an AppleTalk filter to begin with.
3158          * (Note that since pfil doesn't understand ARP it will pass *ALL*
3159          * ARP traffic.)
3160          */
3161         switch (ether_type) {
3162                 case ETHERTYPE_ARP:
3163                 case ETHERTYPE_REVARP:
3164                         if (V_pfil_ipfw_arp == 0)
3165                                 return (0); /* Automatically pass */
3166                         break;
3167
3168                 case ETHERTYPE_IP:
3169 #ifdef INET6
3170                 case ETHERTYPE_IPV6:
3171 #endif /* INET6 */
3172                         break;
3173                 default:
3174                         /*
3175                          * Check to see if the user wants to pass non-ip
3176                          * packets, these will not be checked by pfil(9) and
3177                          * passed unconditionally so the default is to drop.
3178                          */
3179                         if (V_pfil_onlyip)
3180                                 goto bad;
3181         }
3182
3183         /* Run the packet through pfil before stripping link headers */
3184         if (PFIL_HOOKED_OUT(V_link_pfil_head) && V_pfil_ipfw != 0 &&
3185             dir == PFIL_OUT && ifp != NULL) {
3186                 switch (pfil_run_hooks(V_link_pfil_head, mp, ifp, dir, NULL)) {
3187                 case PFIL_DROPPED:
3188                         return (EPERM);
3189                 case PFIL_CONSUMED:
3190                         return (0);
3191                 }
3192         }
3193
3194         /* Strip off the Ethernet header and keep a copy. */
3195         m_copydata(*mp, 0, ETHER_HDR_LEN, (caddr_t) &eh2);
3196         m_adj(*mp, ETHER_HDR_LEN);
3197
3198         /* Strip off snap header, if present */
3199         if (snap) {
3200                 m_copydata(*mp, 0, sizeof(struct llc), (caddr_t) &llc1);
3201                 m_adj(*mp, sizeof(struct llc));
3202         }
3203
3204         /*
3205          * Check the IP header for alignment and errors
3206          */
3207         if (dir == PFIL_IN) {
3208                 switch (ether_type) {
3209                         case ETHERTYPE_IP:
3210                                 error = bridge_ip_checkbasic(mp);
3211                                 break;
3212 #ifdef INET6
3213                         case ETHERTYPE_IPV6:
3214                                 error = bridge_ip6_checkbasic(mp);
3215                                 break;
3216 #endif /* INET6 */
3217                         default:
3218                                 error = 0;
3219                 }
3220                 if (error)
3221                         goto bad;
3222         }
3223
3224         error = 0;
3225
3226         /*
3227          * Run the packet through pfil
3228          */
3229         rv = PFIL_PASS;
3230         switch (ether_type) {
3231         case ETHERTYPE_IP:
3232                 /*
3233                  * Run pfil on the member interface and the bridge, both can
3234                  * be skipped by clearing pfil_member or pfil_bridge.
3235                  *
3236                  * Keep the order:
3237                  *   in_if -> bridge_if -> out_if
3238                  */
3239                 if (V_pfil_bridge && dir == PFIL_OUT && bifp != NULL && (rv =
3240                     pfil_run_hooks(V_inet_pfil_head, mp, bifp, dir, NULL)) !=
3241                     PFIL_PASS)
3242                         break;
3243
3244                 if (V_pfil_member && ifp != NULL && (rv =
3245                     pfil_run_hooks(V_inet_pfil_head, mp, ifp, dir, NULL)) !=
3246                     PFIL_PASS)
3247                         break;
3248
3249                 if (V_pfil_bridge && dir == PFIL_IN && bifp != NULL && (rv =
3250                     pfil_run_hooks(V_inet_pfil_head, mp, bifp, dir, NULL)) !=
3251                     PFIL_PASS)
3252                         break;
3253
3254                 /* check if we need to fragment the packet */
3255                 /* bridge_fragment generates a mbuf chain of packets */
3256                 /* that already include eth headers */
3257                 if (V_pfil_member && ifp != NULL && dir == PFIL_OUT) {
3258                         i = (*mp)->m_pkthdr.len;
3259                         if (i > ifp->if_mtu) {
3260                                 error = bridge_fragment(ifp, mp, &eh2, snap,
3261                                             &llc1);
3262                                 return (error);
3263                         }
3264                 }
3265
3266                 /* Recalculate the ip checksum. */
3267                 ip = mtod(*mp, struct ip *);
3268                 hlen = ip->ip_hl << 2;
3269                 if (hlen < sizeof(struct ip))
3270                         goto bad;
3271                 if (hlen > (*mp)->m_len) {
3272                         if ((*mp = m_pullup(*mp, hlen)) == NULL)
3273                                 goto bad;
3274                         ip = mtod(*mp, struct ip *);
3275                         if (ip == NULL)
3276                                 goto bad;
3277                 }
3278                 ip->ip_sum = 0;
3279                 if (hlen == sizeof(struct ip))
3280                         ip->ip_sum = in_cksum_hdr(ip);
3281                 else
3282                         ip->ip_sum = in_cksum(*mp, hlen);
3283
3284                 break;
3285 #ifdef INET6
3286         case ETHERTYPE_IPV6:
3287                 if (V_pfil_bridge && dir == PFIL_OUT && bifp != NULL && (rv =
3288                     pfil_run_hooks(V_inet6_pfil_head, mp, bifp, dir, NULL)) !=
3289                     PFIL_PASS)
3290                         break;
3291
3292                 if (V_pfil_member && ifp != NULL && (rv =
3293                     pfil_run_hooks(V_inet6_pfil_head, mp, ifp, dir, NULL)) !=
3294                     PFIL_PASS)
3295                         break;
3296
3297                 if (V_pfil_bridge && dir == PFIL_IN && bifp != NULL && (rv =
3298                     pfil_run_hooks(V_inet6_pfil_head, mp, bifp, dir, NULL)) !=
3299                     PFIL_PASS)
3300                         break;
3301                 break;
3302 #endif
3303         }
3304
3305         switch (rv) {
3306         case PFIL_CONSUMED:
3307                 return (0);
3308         case PFIL_DROPPED:
3309                 return (EPERM);
3310         default:
3311                 break;
3312         }
3313
3314         error = -1;
3315
3316         /*
3317          * Finally, put everything back the way it was and return
3318          */
3319         if (snap) {
3320                 M_PREPEND(*mp, sizeof(struct llc), M_NOWAIT);
3321                 if (*mp == NULL)
3322                         return (error);
3323                 bcopy(&llc1, mtod(*mp, caddr_t), sizeof(struct llc));
3324         }
3325
3326         M_PREPEND(*mp, ETHER_HDR_LEN, M_NOWAIT);
3327         if (*mp == NULL)
3328                 return (error);
3329         bcopy(&eh2, mtod(*mp, caddr_t), ETHER_HDR_LEN);
3330
3331         return (0);
3332
3333 bad:
3334         m_freem(*mp);
3335         *mp = NULL;
3336         return (error);
3337 }
3338
3339 /*
3340  * Perform basic checks on header size since
3341  * pfil assumes ip_input has already processed
3342  * it for it.  Cut-and-pasted from ip_input.c.
3343  * Given how simple the IPv6 version is,
3344  * does the IPv4 version really need to be
3345  * this complicated?
3346  *
3347  * XXX Should we update ipstat here, or not?
3348  * XXX Right now we update ipstat but not
3349  * XXX csum_counter.
3350  */
3351 static int
3352 bridge_ip_checkbasic(struct mbuf **mp)
3353 {
3354         struct mbuf *m = *mp;
3355         struct ip *ip;
3356         int len, hlen;
3357         u_short sum;
3358
3359         if (*mp == NULL)
3360                 return (-1);
3361
3362         if (IP_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) {
3363                 if ((m = m_copyup(m, sizeof(struct ip),
3364                         (max_linkhdr + 3) & ~3)) == NULL) {
3365                         /* XXXJRT new stat, please */
3366                         KMOD_IPSTAT_INC(ips_toosmall);
3367                         goto bad;
3368                 }
3369         } else if (__predict_false(m->m_len < sizeof (struct ip))) {
3370                 if ((m = m_pullup(m, sizeof (struct ip))) == NULL) {
3371                         KMOD_IPSTAT_INC(ips_toosmall);
3372                         goto bad;
3373                 }
3374         }
3375         ip = mtod(m, struct ip *);
3376         if (ip == NULL) goto bad;
3377
3378         if (ip->ip_v != IPVERSION) {
3379                 KMOD_IPSTAT_INC(ips_badvers);
3380                 goto bad;
3381         }
3382         hlen = ip->ip_hl << 2;
3383         if (hlen < sizeof(struct ip)) { /* minimum header length */
3384                 KMOD_IPSTAT_INC(ips_badhlen);
3385                 goto bad;
3386         }
3387         if (hlen > m->m_len) {
3388                 if ((m = m_pullup(m, hlen)) == NULL) {
3389                         KMOD_IPSTAT_INC(ips_badhlen);
3390                         goto bad;
3391                 }
3392                 ip = mtod(m, struct ip *);
3393                 if (ip == NULL) goto bad;
3394         }
3395
3396         if (m->m_pkthdr.csum_flags & CSUM_IP_CHECKED) {
3397                 sum = !(m->m_pkthdr.csum_flags & CSUM_IP_VALID);
3398         } else {
3399                 if (hlen == sizeof(struct ip)) {
3400                         sum = in_cksum_hdr(ip);
3401                 } else {
3402                         sum = in_cksum(m, hlen);
3403                 }
3404         }
3405         if (sum) {
3406                 KMOD_IPSTAT_INC(ips_badsum);
3407                 goto bad;
3408         }
3409
3410         /* Retrieve the packet length. */
3411         len = ntohs(ip->ip_len);
3412
3413         /*
3414          * Check for additional length bogosity
3415          */
3416         if (len < hlen) {
3417                 KMOD_IPSTAT_INC(ips_badlen);
3418                 goto bad;
3419         }
3420
3421         /*
3422          * Check that the amount of data in the buffers
3423          * is as at least much as the IP header would have us expect.
3424          * Drop packet if shorter than we expect.
3425          */
3426         if (m->m_pkthdr.len < len) {
3427                 KMOD_IPSTAT_INC(ips_tooshort);
3428                 goto bad;
3429         }
3430
3431         /* Checks out, proceed */
3432         *mp = m;
3433         return (0);
3434
3435 bad:
3436         *mp = m;
3437         return (-1);
3438 }
3439
3440 #ifdef INET6
3441 /*
3442  * Same as above, but for IPv6.
3443  * Cut-and-pasted from ip6_input.c.
3444  * XXX Should we update ip6stat, or not?
3445  */
3446 static int
3447 bridge_ip6_checkbasic(struct mbuf **mp)
3448 {
3449         struct mbuf *m = *mp;
3450         struct ip6_hdr *ip6;
3451
3452         /*
3453          * If the IPv6 header is not aligned, slurp it up into a new
3454          * mbuf with space for link headers, in the event we forward
3455          * it.  Otherwise, if it is aligned, make sure the entire base
3456          * IPv6 header is in the first mbuf of the chain.
3457          */
3458         if (IP6_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) {
3459                 struct ifnet *inifp = m->m_pkthdr.rcvif;
3460                 if ((m = m_copyup(m, sizeof(struct ip6_hdr),
3461                             (max_linkhdr + 3) & ~3)) == NULL) {
3462                         /* XXXJRT new stat, please */
3463                         IP6STAT_INC(ip6s_toosmall);
3464                         in6_ifstat_inc(inifp, ifs6_in_hdrerr);
3465                         goto bad;
3466                 }
3467         } else if (__predict_false(m->m_len < sizeof(struct ip6_hdr))) {
3468                 struct ifnet *inifp = m->m_pkthdr.rcvif;
3469                 if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) {
3470                         IP6STAT_INC(ip6s_toosmall);
3471                         in6_ifstat_inc(inifp, ifs6_in_hdrerr);
3472                         goto bad;
3473                 }
3474         }
3475
3476         ip6 = mtod(m, struct ip6_hdr *);
3477
3478         if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) {
3479                 IP6STAT_INC(ip6s_badvers);
3480                 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr);
3481                 goto bad;
3482         }
3483
3484         /* Checks out, proceed */
3485         *mp = m;
3486         return (0);
3487
3488 bad:
3489         *mp = m;
3490         return (-1);
3491 }
3492 #endif /* INET6 */
3493
3494 /*
3495  * bridge_fragment:
3496  *
3497  *      Fragment mbuf chain in multiple packets and prepend ethernet header.
3498  */
3499 static int
3500 bridge_fragment(struct ifnet *ifp, struct mbuf **mp, struct ether_header *eh,
3501     int snap, struct llc *llc)
3502 {
3503         struct mbuf *m = *mp, *nextpkt = NULL, *mprev = NULL, *mcur = NULL;
3504         struct ip *ip;
3505         int error = -1;
3506
3507         if (m->m_len < sizeof(struct ip) &&
3508             (m = m_pullup(m, sizeof(struct ip))) == NULL)
3509                 goto dropit;
3510         ip = mtod(m, struct ip *);
3511
3512         m->m_pkthdr.csum_flags |= CSUM_IP;
3513         error = ip_fragment(ip, &m, ifp->if_mtu, ifp->if_hwassist);
3514         if (error)
3515                 goto dropit;
3516
3517         /*
3518          * Walk the chain and re-add the Ethernet header for
3519          * each mbuf packet.
3520          */
3521         for (mcur = m; mcur; mcur = mcur->m_nextpkt) {
3522                 nextpkt = mcur->m_nextpkt;
3523                 mcur->m_nextpkt = NULL;
3524                 if (snap) {
3525                         M_PREPEND(mcur, sizeof(struct llc), M_NOWAIT);
3526                         if (mcur == NULL) {
3527                                 error = ENOBUFS;
3528                                 if (mprev != NULL)
3529                                         mprev->m_nextpkt = nextpkt;
3530                                 goto dropit;
3531                         }
3532                         bcopy(llc, mtod(mcur, caddr_t),sizeof(struct llc));
3533                 }
3534
3535                 M_PREPEND(mcur, ETHER_HDR_LEN, M_NOWAIT);
3536                 if (mcur == NULL) {
3537                         error = ENOBUFS;
3538                         if (mprev != NULL)
3539                                 mprev->m_nextpkt = nextpkt;
3540                         goto dropit;
3541                 }
3542                 bcopy(eh, mtod(mcur, caddr_t), ETHER_HDR_LEN);
3543
3544                 /*
3545                  * The previous two M_PREPEND could have inserted one or two
3546                  * mbufs in front so we have to update the previous packet's
3547                  * m_nextpkt.
3548                  */
3549                 mcur->m_nextpkt = nextpkt;
3550                 if (mprev != NULL)
3551                         mprev->m_nextpkt = mcur;
3552                 else {
3553                         /* The first mbuf in the original chain needs to be
3554                          * updated. */
3555                         *mp = mcur;
3556                 }
3557                 mprev = mcur;
3558         }
3559
3560         KMOD_IPSTAT_INC(ips_fragmented);
3561         return (error);
3562
3563 dropit:
3564         for (mcur = *mp; mcur; mcur = m) { /* droping the full packet chain */
3565                 m = mcur->m_nextpkt;
3566                 m_freem(mcur);
3567         }
3568         return (error);
3569 }
3570
3571 static void
3572 bridge_linkstate(struct ifnet *ifp)
3573 {
3574         struct bridge_softc *sc = ifp->if_bridge;
3575         struct bridge_iflist *bif;
3576
3577         BRIDGE_LOCK(sc);
3578         bif = bridge_lookup_member_if(sc, ifp);
3579         if (bif == NULL) {
3580                 BRIDGE_UNLOCK(sc);
3581                 return;
3582         }
3583         bridge_linkcheck(sc);
3584         BRIDGE_UNLOCK(sc);
3585
3586         bstp_linkstate(&bif->bif_stp);
3587 }
3588
3589 static void
3590 bridge_linkcheck(struct bridge_softc *sc)
3591 {
3592         struct bridge_iflist *bif;
3593         int new_link, hasls;
3594
3595         BRIDGE_LOCK_ASSERT(sc);
3596         new_link = LINK_STATE_DOWN;
3597         hasls = 0;
3598         /* Our link is considered up if at least one of our ports is active */
3599         LIST_FOREACH(bif, &sc->sc_iflist, bif_next) {
3600                 if (bif->bif_ifp->if_capabilities & IFCAP_LINKSTATE)
3601                         hasls++;
3602                 if (bif->bif_ifp->if_link_state == LINK_STATE_UP) {
3603                         new_link = LINK_STATE_UP;
3604                         break;
3605                 }
3606         }
3607         if (!LIST_EMPTY(&sc->sc_iflist) && !hasls) {
3608                 /* If no interfaces support link-state then we default to up */
3609                 new_link = LINK_STATE_UP;
3610         }
3611         if_link_state_change(sc->sc_ifp, new_link);
3612 }