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