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 heterogeneous 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 IFCAP_TXCSUM_IPV6) 171 172 /* 173 * List of capabilities to strip 174 */ 175 #define BRIDGE_IFCAPS_STRIP IFCAP_LRO 176 177 /* 178 * Bridge interface list entry. 179 */ 180 struct bridge_iflist { 181 LIST_ENTRY(bridge_iflist) bif_next; 182 struct ifnet *bif_ifp; /* member if */ 183 struct bstp_port bif_stp; /* STP state */ 184 uint32_t bif_flags; /* member if flags */ 185 int bif_savedcaps; /* saved capabilities */ 186 uint32_t bif_addrmax; /* max # of addresses */ 187 uint32_t bif_addrcnt; /* cur. # of addresses */ 188 uint32_t bif_addrexceeded;/* # of address violations */ 189 }; 190 191 /* 192 * Bridge route node. 193 */ 194 struct bridge_rtnode { 195 LIST_ENTRY(bridge_rtnode) brt_hash; /* hash table linkage */ 196 LIST_ENTRY(bridge_rtnode) brt_list; /* list linkage */ 197 struct bridge_iflist *brt_dst; /* destination if */ 198 unsigned long brt_expire; /* expiration time */ 199 uint8_t brt_flags; /* address flags */ 200 uint8_t brt_addr[ETHER_ADDR_LEN]; 201 uint16_t brt_vlan; /* vlan id */ 202 }; 203 #define brt_ifp brt_dst->bif_ifp 204 205 /* 206 * Software state for each bridge. 207 */ 208 struct bridge_softc { 209 struct ifnet *sc_ifp; /* make this an interface */ 210 LIST_ENTRY(bridge_softc) sc_list; 211 struct mtx sc_mtx; 212 struct cv sc_cv; 213 uint32_t sc_brtmax; /* max # of addresses */ 214 uint32_t sc_brtcnt; /* cur. # of addresses */ 215 uint32_t sc_brttimeout; /* rt timeout in seconds */ 216 struct callout sc_brcallout; /* bridge callout */ 217 uint32_t sc_iflist_ref; /* refcount for sc_iflist */ 218 uint32_t sc_iflist_xcnt; /* refcount for sc_iflist */ 219 LIST_HEAD(, bridge_iflist) sc_iflist; /* member interface list */ 220 LIST_HEAD(, bridge_rtnode) *sc_rthash; /* our forwarding table */ 221 LIST_HEAD(, bridge_rtnode) sc_rtlist; /* list version of above */ 222 uint32_t sc_rthash_key; /* key for hash */ 223 LIST_HEAD(, bridge_iflist) sc_spanlist; /* span ports list */ 224 struct bstp_state sc_stp; /* STP state */ 225 uint32_t sc_brtexceeded; /* # of cache drops */ 226 struct ifnet *sc_ifaddr; /* member mac copied from */ 227 u_char sc_defaddr[6]; /* Default MAC address */ 228 }; 229 230 static VNET_DEFINE(struct mtx, bridge_list_mtx); 231 #define V_bridge_list_mtx VNET(bridge_list_mtx) 232 static eventhandler_tag bridge_detach_cookie; 233 234 int bridge_rtable_prune_period = BRIDGE_RTABLE_PRUNE_PERIOD; 235 236 uma_zone_t bridge_rtnode_zone; 237 238 static int bridge_clone_create(struct if_clone *, int, caddr_t); 239 static void bridge_clone_destroy(struct ifnet *); 240 241 static int bridge_ioctl(struct ifnet *, u_long, caddr_t); 242 static void bridge_mutecaps(struct bridge_softc *); 243 static void bridge_set_ifcap(struct bridge_softc *, struct bridge_iflist *, 244 int); 245 static void bridge_ifdetach(void *arg __unused, struct ifnet *); 246 static void bridge_init(void *); 247 static void bridge_dummynet(struct mbuf *, struct ifnet *); 248 static void bridge_stop(struct ifnet *, int); 249 static int bridge_transmit(struct ifnet *, struct mbuf *); 250 static void bridge_qflush(struct ifnet *); 251 static struct mbuf *bridge_input(struct ifnet *, struct mbuf *); 252 static int bridge_output(struct ifnet *, struct mbuf *, struct sockaddr *, 253 struct rtentry *); 254 static int bridge_enqueue(struct bridge_softc *, struct ifnet *, 255 struct mbuf *); 256 static void bridge_rtdelete(struct bridge_softc *, struct ifnet *ifp, int); 257 258 static void bridge_forward(struct bridge_softc *, struct bridge_iflist *, 259 struct mbuf *m); 260 261 static void bridge_timer(void *); 262 263 static void bridge_broadcast(struct bridge_softc *, struct ifnet *, 264 struct mbuf *, int); 265 static void bridge_span(struct bridge_softc *, struct mbuf *); 266 267 static int bridge_rtupdate(struct bridge_softc *, const uint8_t *, 268 uint16_t, struct bridge_iflist *, int, uint8_t); 269 static struct ifnet *bridge_rtlookup(struct bridge_softc *, const uint8_t *, 270 uint16_t); 271 static void bridge_rttrim(struct bridge_softc *); 272 static void bridge_rtage(struct bridge_softc *); 273 static void bridge_rtflush(struct bridge_softc *, int); 274 static int bridge_rtdaddr(struct bridge_softc *, const uint8_t *, 275 uint16_t); 276 277 static void bridge_rtable_init(struct bridge_softc *); 278 static void bridge_rtable_fini(struct bridge_softc *); 279 280 static int bridge_rtnode_addr_cmp(const uint8_t *, const uint8_t *); 281 static struct bridge_rtnode *bridge_rtnode_lookup(struct bridge_softc *, 282 const uint8_t *, uint16_t); 283 static int bridge_rtnode_insert(struct bridge_softc *, 284 struct bridge_rtnode *); 285 static void bridge_rtnode_destroy(struct bridge_softc *, 286 struct bridge_rtnode *); 287 static void bridge_rtable_expire(struct ifnet *, int); 288 static void bridge_state_change(struct ifnet *, int); 289 290 static struct bridge_iflist *bridge_lookup_member(struct bridge_softc *, 291 const char *name); 292 static struct bridge_iflist *bridge_lookup_member_if(struct bridge_softc *, 293 struct ifnet *ifp); 294 static void bridge_delete_member(struct bridge_softc *, 295 struct bridge_iflist *, int); 296 static void bridge_delete_span(struct bridge_softc *, 297 struct bridge_iflist *); 298 299 static int bridge_ioctl_add(struct bridge_softc *, void *); 300 static int bridge_ioctl_del(struct bridge_softc *, void *); 301 static int bridge_ioctl_gifflags(struct bridge_softc *, void *); 302 static int bridge_ioctl_sifflags(struct bridge_softc *, void *); 303 static int bridge_ioctl_scache(struct bridge_softc *, void *); 304 static int bridge_ioctl_gcache(struct bridge_softc *, void *); 305 static int bridge_ioctl_gifs(struct bridge_softc *, void *); 306 static int bridge_ioctl_rts(struct bridge_softc *, void *); 307 static int bridge_ioctl_saddr(struct bridge_softc *, void *); 308 static int bridge_ioctl_sto(struct bridge_softc *, void *); 309 static int bridge_ioctl_gto(struct bridge_softc *, void *); 310 static int bridge_ioctl_daddr(struct bridge_softc *, void *); 311 static int bridge_ioctl_flush(struct bridge_softc *, void *); 312 static int bridge_ioctl_gpri(struct bridge_softc *, void *); 313 static int bridge_ioctl_spri(struct bridge_softc *, void *); 314 static int bridge_ioctl_ght(struct bridge_softc *, void *); 315 static int bridge_ioctl_sht(struct bridge_softc *, void *); 316 static int bridge_ioctl_gfd(struct bridge_softc *, void *); 317 static int bridge_ioctl_sfd(struct bridge_softc *, void *); 318 static int bridge_ioctl_gma(struct bridge_softc *, void *); 319 static int bridge_ioctl_sma(struct bridge_softc *, void *); 320 static int bridge_ioctl_sifprio(struct bridge_softc *, void *); 321 static int bridge_ioctl_sifcost(struct bridge_softc *, void *); 322 static int bridge_ioctl_sifmaxaddr(struct bridge_softc *, void *); 323 static int bridge_ioctl_addspan(struct bridge_softc *, void *); 324 static int bridge_ioctl_delspan(struct bridge_softc *, void *); 325 static int bridge_ioctl_gbparam(struct bridge_softc *, void *); 326 static int bridge_ioctl_grte(struct bridge_softc *, void *); 327 static int bridge_ioctl_gifsstp(struct bridge_softc *, void *); 328 static int bridge_ioctl_sproto(struct bridge_softc *, void *); 329 static int bridge_ioctl_stxhc(struct bridge_softc *, void *); 330 static int bridge_pfil(struct mbuf **, struct ifnet *, struct ifnet *, 331 int); 332 static int bridge_ip_checkbasic(struct mbuf **mp); 333 #ifdef INET6 334 static int bridge_ip6_checkbasic(struct mbuf **mp); 335 #endif /* INET6 */ 336 static int bridge_fragment(struct ifnet *, struct mbuf **mp, 337 struct ether_header *, int, struct llc *); 338 static void bridge_linkstate(struct ifnet *ifp); 339 static void bridge_linkcheck(struct bridge_softc *sc); 340 341 extern void (*bridge_linkstate_p)(struct ifnet *ifp); 342 343 /* The default bridge vlan is 1 (IEEE 802.1Q-2003 Table 9-2) */ 344 #define VLANTAGOF(_m) \ 345 (_m->m_flags & M_VLANTAG) ? EVL_VLANOFTAG(_m->m_pkthdr.ether_vtag) : 1 346 347 static struct bstp_cb_ops bridge_ops = { 348 .bcb_state = bridge_state_change, 349 .bcb_rtage = bridge_rtable_expire 350 }; 351 352 SYSCTL_DECL(_net_link); 353 static SYSCTL_NODE(_net_link, IFT_BRIDGE, bridge, CTLFLAG_RW, 0, "Bridge"); 354 355 /* only pass IP[46] packets when pfil is enabled */ 356 static VNET_DEFINE(int, pfil_onlyip) = 1; 357 #define V_pfil_onlyip VNET(pfil_onlyip) 358 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_onlyip, 359 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_onlyip), 0, 360 "Only pass IP packets when pfil is enabled"); 361 362 /* run pfil hooks on the bridge interface */ 363 static VNET_DEFINE(int, pfil_bridge) = 1; 364 #define V_pfil_bridge VNET(pfil_bridge) 365 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_bridge, 366 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_bridge), 0, 367 "Packet filter on the bridge interface"); 368 369 /* layer2 filter with ipfw */ 370 static VNET_DEFINE(int, pfil_ipfw); 371 #define V_pfil_ipfw VNET(pfil_ipfw) 372 373 /* layer2 ARP filter with ipfw */ 374 static VNET_DEFINE(int, pfil_ipfw_arp); 375 #define V_pfil_ipfw_arp VNET(pfil_ipfw_arp) 376 SYSCTL_INT(_net_link_bridge, OID_AUTO, ipfw_arp, 377 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_ipfw_arp), 0, 378 "Filter ARP packets through IPFW layer2"); 379 380 /* run pfil hooks on the member interface */ 381 static VNET_DEFINE(int, pfil_member) = 1; 382 #define V_pfil_member VNET(pfil_member) 383 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_member, 384 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_member), 0, 385 "Packet filter on the member interface"); 386 387 /* run pfil hooks on the physical interface for locally destined packets */ 388 static VNET_DEFINE(int, pfil_local_phys); 389 #define V_pfil_local_phys VNET(pfil_local_phys) 390 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_local_phys, 391 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_local_phys), 0, 392 "Packet filter on the physical interface for locally destined packets"); 393 394 /* log STP state changes */ 395 static VNET_DEFINE(int, log_stp); 396 #define V_log_stp VNET(log_stp) 397 SYSCTL_INT(_net_link_bridge, OID_AUTO, log_stp, 398 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(log_stp), 0, 399 "Log STP state changes"); 400 401 /* share MAC with first bridge member */ 402 static VNET_DEFINE(int, bridge_inherit_mac); 403 #define V_bridge_inherit_mac VNET(bridge_inherit_mac) 404 SYSCTL_INT(_net_link_bridge, OID_AUTO, inherit_mac, 405 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(bridge_inherit_mac), 0, 406 "Inherit MAC address from the first bridge member"); 407 408 static VNET_DEFINE(int, allow_llz_overlap) = 0; 409 #define V_allow_llz_overlap VNET(allow_llz_overlap) 410 SYSCTL_INT(_net_link_bridge, OID_AUTO, allow_llz_overlap, 411 CTLFLAG_RW | CTLFLAG_VNET, &VNET_NAME(allow_llz_overlap), 0, 412 "Allow overlap of link-local scope " 413 "zones of a bridge interface and the member interfaces"); 414 415 struct bridge_control { 416 int (*bc_func)(struct bridge_softc *, void *); 417 int bc_argsize; 418 int bc_flags; 419 }; 420 421 #define BC_F_COPYIN 0x01 /* copy arguments in */ 422 #define BC_F_COPYOUT 0x02 /* copy arguments out */ 423 #define BC_F_SUSER 0x04 /* do super-user check */ 424 425 const struct bridge_control bridge_control_table[] = { 426 { bridge_ioctl_add, sizeof(struct ifbreq), 427 BC_F_COPYIN|BC_F_SUSER }, 428 { bridge_ioctl_del, sizeof(struct ifbreq), 429 BC_F_COPYIN|BC_F_SUSER }, 430 431 { bridge_ioctl_gifflags, sizeof(struct ifbreq), 432 BC_F_COPYIN|BC_F_COPYOUT }, 433 { bridge_ioctl_sifflags, sizeof(struct ifbreq), 434 BC_F_COPYIN|BC_F_SUSER }, 435 436 { bridge_ioctl_scache, sizeof(struct ifbrparam), 437 BC_F_COPYIN|BC_F_SUSER }, 438 { bridge_ioctl_gcache, sizeof(struct ifbrparam), 439 BC_F_COPYOUT }, 440 441 { bridge_ioctl_gifs, sizeof(struct ifbifconf), 442 BC_F_COPYIN|BC_F_COPYOUT }, 443 { bridge_ioctl_rts, sizeof(struct ifbaconf), 444 BC_F_COPYIN|BC_F_COPYOUT }, 445 446 { bridge_ioctl_saddr, sizeof(struct ifbareq), 447 BC_F_COPYIN|BC_F_SUSER }, 448 449 { bridge_ioctl_sto, sizeof(struct ifbrparam), 450 BC_F_COPYIN|BC_F_SUSER }, 451 { bridge_ioctl_gto, sizeof(struct ifbrparam), 452 BC_F_COPYOUT }, 453 454 { bridge_ioctl_daddr, sizeof(struct ifbareq), 455 BC_F_COPYIN|BC_F_SUSER }, 456 457 { bridge_ioctl_flush, sizeof(struct ifbreq), 458 BC_F_COPYIN|BC_F_SUSER }, 459 460 { bridge_ioctl_gpri, sizeof(struct ifbrparam), 461 BC_F_COPYOUT }, 462 { bridge_ioctl_spri, sizeof(struct ifbrparam), 463 BC_F_COPYIN|BC_F_SUSER }, 464 465 { bridge_ioctl_ght, sizeof(struct ifbrparam), 466 BC_F_COPYOUT }, 467 { bridge_ioctl_sht, sizeof(struct ifbrparam), 468 BC_F_COPYIN|BC_F_SUSER }, 469 470 { bridge_ioctl_gfd, sizeof(struct ifbrparam), 471 BC_F_COPYOUT }, 472 { bridge_ioctl_sfd, sizeof(struct ifbrparam), 473 BC_F_COPYIN|BC_F_SUSER }, 474 475 { bridge_ioctl_gma, sizeof(struct ifbrparam), 476 BC_F_COPYOUT }, 477 { bridge_ioctl_sma, sizeof(struct ifbrparam), 478 BC_F_COPYIN|BC_F_SUSER }, 479 480 { bridge_ioctl_sifprio, sizeof(struct ifbreq), 481 BC_F_COPYIN|BC_F_SUSER }, 482 483 { bridge_ioctl_sifcost, sizeof(struct ifbreq), 484 BC_F_COPYIN|BC_F_SUSER }, 485 486 { bridge_ioctl_addspan, sizeof(struct ifbreq), 487 BC_F_COPYIN|BC_F_SUSER }, 488 { bridge_ioctl_delspan, sizeof(struct ifbreq), 489 BC_F_COPYIN|BC_F_SUSER }, 490 491 { bridge_ioctl_gbparam, sizeof(struct ifbropreq), 492 BC_F_COPYOUT }, 493 494 { bridge_ioctl_grte, sizeof(struct ifbrparam), 495 BC_F_COPYOUT }, 496 497 { bridge_ioctl_gifsstp, sizeof(struct ifbpstpconf), 498 BC_F_COPYIN|BC_F_COPYOUT }, 499 500 { bridge_ioctl_sproto, sizeof(struct ifbrparam), 501 BC_F_COPYIN|BC_F_SUSER }, 502 503 { bridge_ioctl_stxhc, sizeof(struct ifbrparam), 504 BC_F_COPYIN|BC_F_SUSER }, 505 506 { bridge_ioctl_sifmaxaddr, sizeof(struct ifbreq), 507 BC_F_COPYIN|BC_F_SUSER }, 508 509 }; 510 const int bridge_control_table_size = nitems(bridge_control_table); 511 512 static VNET_DEFINE(LIST_HEAD(, bridge_softc), bridge_list); 513 #define V_bridge_list VNET(bridge_list) 514 #define BRIDGE_LIST_LOCK_INIT(x) mtx_init(&V_bridge_list_mtx, \ 515 "if_bridge list", NULL, MTX_DEF) 516 #define BRIDGE_LIST_LOCK_DESTROY(x) mtx_destroy(&V_bridge_list_mtx) 517 #define BRIDGE_LIST_LOCK(x) mtx_lock(&V_bridge_list_mtx) 518 #define BRIDGE_LIST_UNLOCK(x) mtx_unlock(&V_bridge_list_mtx) 519 520 static VNET_DEFINE(struct if_clone *, bridge_cloner); 521 #define V_bridge_cloner VNET(bridge_cloner) 522 523 static const char bridge_name[] = "bridge"; 524 525 static void 526 vnet_bridge_init(const void *unused __unused) 527 { 528 529 BRIDGE_LIST_LOCK_INIT(); 530 LIST_INIT(&V_bridge_list); 531 V_bridge_cloner = if_clone_simple(bridge_name, 532 bridge_clone_create, bridge_clone_destroy, 0); 533 } 534 VNET_SYSINIT(vnet_bridge_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY, 535 vnet_bridge_init, NULL); 536 537 static void 538 vnet_bridge_uninit(const void *unused __unused) 539 { 540 541 if_clone_detach(V_bridge_cloner); 542 V_bridge_cloner = NULL; 543 BRIDGE_LIST_LOCK_DESTROY(); 544 } 545 VNET_SYSUNINIT(vnet_bridge_uninit, SI_SUB_PSEUDO, SI_ORDER_ANY, 546 vnet_bridge_uninit, NULL); 547 548 static int 549 bridge_modevent(module_t mod, int type, void *data) 550 { 551 552 switch (type) { 553 case MOD_LOAD: 554 bridge_rtnode_zone = uma_zcreate("bridge_rtnode", 555 sizeof(struct bridge_rtnode), NULL, NULL, NULL, NULL, 556 UMA_ALIGN_PTR, 0); 557 bridge_input_p = bridge_input; 558 bridge_output_p = bridge_output; 559 bridge_dn_p = bridge_dummynet; 560 bridge_linkstate_p = bridge_linkstate; 561 bridge_detach_cookie = EVENTHANDLER_REGISTER( 562 ifnet_departure_event, bridge_ifdetach, NULL, 563 EVENTHANDLER_PRI_ANY); 564 break; 565 case MOD_UNLOAD: 566 EVENTHANDLER_DEREGISTER(ifnet_departure_event, 567 bridge_detach_cookie); 568 uma_zdestroy(bridge_rtnode_zone); 569 bridge_input_p = NULL; 570 bridge_output_p = NULL; 571 bridge_dn_p = NULL; 572 bridge_linkstate_p = NULL; 573 break; 574 default: 575 return (EOPNOTSUPP); 576 } 577 return (0); 578 } 579 580 static moduledata_t bridge_mod = { 581 "if_bridge", 582 bridge_modevent, 583 0 584 }; 585 586 DECLARE_MODULE(if_bridge, bridge_mod, SI_SUB_PSEUDO, SI_ORDER_ANY); 587 MODULE_DEPEND(if_bridge, bridgestp, 1, 1, 1); 588 589 /* 590 * handler for net.link.bridge.ipfw 591 */ 592 static int 593 sysctl_pfil_ipfw(SYSCTL_HANDLER_ARGS) 594 { 595 int enable = V_pfil_ipfw; 596 int error; 597 598 error = sysctl_handle_int(oidp, &enable, 0, req); 599 enable &= 1; 600 601 if (enable != V_pfil_ipfw) { 602 V_pfil_ipfw = enable; 603 604 /* 605 * Disable pfil so that ipfw doesnt run twice, if the user 606 * really wants both then they can re-enable pfil_bridge and/or 607 * pfil_member. Also allow non-ip packets as ipfw can filter by 608 * layer2 type. 609 */ 610 if (V_pfil_ipfw) { 611 V_pfil_onlyip = 0; 612 V_pfil_bridge = 0; 613 V_pfil_member = 0; 614 } 615 } 616 617 return (error); 618 } 619 SYSCTL_PROC(_net_link_bridge, OID_AUTO, ipfw, 620 CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_VNET, 621 &VNET_NAME(pfil_ipfw), 0, &sysctl_pfil_ipfw, "I", 622 "Layer2 filter with IPFW"); 623 624 /* 625 * bridge_clone_create: 626 * 627 * Create a new bridge instance. 628 */ 629 static int 630 bridge_clone_create(struct if_clone *ifc, int unit, caddr_t params) 631 { 632 struct bridge_softc *sc, *sc2; 633 struct ifnet *bifp, *ifp; 634 int fb, retry; 635 unsigned long hostid; 636 637 sc = malloc(sizeof(*sc), M_DEVBUF, M_WAITOK|M_ZERO); 638 ifp = sc->sc_ifp = if_alloc(IFT_ETHER); 639 if (ifp == NULL) { 640 free(sc, M_DEVBUF); 641 return (ENOSPC); 642 } 643 644 BRIDGE_LOCK_INIT(sc); 645 sc->sc_brtmax = BRIDGE_RTABLE_MAX; 646 sc->sc_brttimeout = BRIDGE_RTABLE_TIMEOUT; 647 648 /* Initialize our routing table. */ 649 bridge_rtable_init(sc); 650 651 callout_init_mtx(&sc->sc_brcallout, &sc->sc_mtx, 0); 652 653 LIST_INIT(&sc->sc_iflist); 654 LIST_INIT(&sc->sc_spanlist); 655 656 ifp->if_softc = sc; 657 if_initname(ifp, bridge_name, unit); 658 ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST; 659 ifp->if_ioctl = bridge_ioctl; 660 ifp->if_transmit = bridge_transmit; 661 ifp->if_qflush = bridge_qflush; 662 ifp->if_init = bridge_init; 663 ifp->if_type = IFT_BRIDGE; 664 665 /* 666 * Generate an ethernet address with a locally administered address. 667 * 668 * Since we are using random ethernet addresses for the bridge, it is 669 * possible that we might have address collisions, so make sure that 670 * this hardware address isn't already in use on another bridge. 671 * The first try uses the hostid and falls back to arc4rand(). 672 */ 673 fb = 0; 674 getcredhostid(curthread->td_ucred, &hostid); 675 do { 676 if (fb || hostid == 0) { 677 arc4rand(sc->sc_defaddr, ETHER_ADDR_LEN, 1); 678 sc->sc_defaddr[0] &= ~1;/* clear multicast bit */ 679 sc->sc_defaddr[0] |= 2; /* set the LAA bit */ 680 } else { 681 sc->sc_defaddr[0] = 0x2; 682 sc->sc_defaddr[1] = (hostid >> 24) & 0xff; 683 sc->sc_defaddr[2] = (hostid >> 16) & 0xff; 684 sc->sc_defaddr[3] = (hostid >> 8 ) & 0xff; 685 sc->sc_defaddr[4] = hostid & 0xff; 686 sc->sc_defaddr[5] = ifp->if_dunit & 0xff; 687 } 688 689 fb = 1; 690 retry = 0; 691 BRIDGE_LIST_LOCK(); 692 LIST_FOREACH(sc2, &V_bridge_list, sc_list) { 693 bifp = sc2->sc_ifp; 694 if (memcmp(sc->sc_defaddr, 695 IF_LLADDR(bifp), ETHER_ADDR_LEN) == 0) { 696 retry = 1; 697 break; 698 } 699 } 700 BRIDGE_LIST_UNLOCK(); 701 } while (retry == 1); 702 703 bstp_attach(&sc->sc_stp, &bridge_ops); 704 ether_ifattach(ifp, sc->sc_defaddr); 705 /* Now undo some of the damage... */ 706 ifp->if_baudrate = 0; 707 ifp->if_type = IFT_BRIDGE; 708 709 BRIDGE_LIST_LOCK(); 710 LIST_INSERT_HEAD(&V_bridge_list, sc, sc_list); 711 BRIDGE_LIST_UNLOCK(); 712 713 return (0); 714 } 715 716 /* 717 * bridge_clone_destroy: 718 * 719 * Destroy a bridge instance. 720 */ 721 static void 722 bridge_clone_destroy(struct ifnet *ifp) 723 { 724 struct bridge_softc *sc = ifp->if_softc; 725 struct bridge_iflist *bif; 726 727 BRIDGE_LOCK(sc); 728 729 bridge_stop(ifp, 1); 730 ifp->if_flags &= ~IFF_UP; 731 732 while ((bif = LIST_FIRST(&sc->sc_iflist)) != NULL) 733 bridge_delete_member(sc, bif, 0); 734 735 while ((bif = LIST_FIRST(&sc->sc_spanlist)) != NULL) { 736 bridge_delete_span(sc, bif); 737 } 738 739 BRIDGE_UNLOCK(sc); 740 741 callout_drain(&sc->sc_brcallout); 742 743 BRIDGE_LIST_LOCK(); 744 LIST_REMOVE(sc, sc_list); 745 BRIDGE_LIST_UNLOCK(); 746 747 bstp_detach(&sc->sc_stp); 748 ether_ifdetach(ifp); 749 if_free(ifp); 750 751 /* Tear down the routing table. */ 752 bridge_rtable_fini(sc); 753 754 BRIDGE_LOCK_DESTROY(sc); 755 free(sc, M_DEVBUF); 756 } 757 758 /* 759 * bridge_ioctl: 760 * 761 * Handle a control request from the operator. 762 */ 763 static int 764 bridge_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data) 765 { 766 struct bridge_softc *sc = ifp->if_softc; 767 struct ifreq *ifr = (struct ifreq *)data; 768 struct bridge_iflist *bif; 769 struct thread *td = curthread; 770 union { 771 struct ifbreq ifbreq; 772 struct ifbifconf ifbifconf; 773 struct ifbareq ifbareq; 774 struct ifbaconf ifbaconf; 775 struct ifbrparam ifbrparam; 776 struct ifbropreq ifbropreq; 777 } args; 778 struct ifdrv *ifd = (struct ifdrv *) data; 779 const struct bridge_control *bc; 780 int error = 0; 781 782 switch (cmd) { 783 784 case SIOCADDMULTI: 785 case SIOCDELMULTI: 786 break; 787 788 case SIOCGDRVSPEC: 789 case SIOCSDRVSPEC: 790 if (ifd->ifd_cmd >= bridge_control_table_size) { 791 error = EINVAL; 792 break; 793 } 794 bc = &bridge_control_table[ifd->ifd_cmd]; 795 796 if (cmd == SIOCGDRVSPEC && 797 (bc->bc_flags & BC_F_COPYOUT) == 0) { 798 error = EINVAL; 799 break; 800 } 801 else if (cmd == SIOCSDRVSPEC && 802 (bc->bc_flags & BC_F_COPYOUT) != 0) { 803 error = EINVAL; 804 break; 805 } 806 807 if (bc->bc_flags & BC_F_SUSER) { 808 error = priv_check(td, PRIV_NET_BRIDGE); 809 if (error) 810 break; 811 } 812 813 if (ifd->ifd_len != bc->bc_argsize || 814 ifd->ifd_len > sizeof(args)) { 815 error = EINVAL; 816 break; 817 } 818 819 bzero(&args, sizeof(args)); 820 if (bc->bc_flags & BC_F_COPYIN) { 821 error = copyin(ifd->ifd_data, &args, ifd->ifd_len); 822 if (error) 823 break; 824 } 825 826 BRIDGE_LOCK(sc); 827 error = (*bc->bc_func)(sc, &args); 828 BRIDGE_UNLOCK(sc); 829 if (error) 830 break; 831 832 if (bc->bc_flags & BC_F_COPYOUT) 833 error = copyout(&args, ifd->ifd_data, ifd->ifd_len); 834 835 break; 836 837 case SIOCSIFFLAGS: 838 if (!(ifp->if_flags & IFF_UP) && 839 (ifp->if_drv_flags & IFF_DRV_RUNNING)) { 840 /* 841 * If interface is marked down and it is running, 842 * then stop and disable it. 843 */ 844 BRIDGE_LOCK(sc); 845 bridge_stop(ifp, 1); 846 BRIDGE_UNLOCK(sc); 847 } else if ((ifp->if_flags & IFF_UP) && 848 !(ifp->if_drv_flags & IFF_DRV_RUNNING)) { 849 /* 850 * If interface is marked up and it is stopped, then 851 * start it. 852 */ 853 (*ifp->if_init)(sc); 854 } 855 break; 856 857 case SIOCSIFMTU: 858 if (ifr->ifr_mtu < 576) { 859 error = EINVAL; 860 break; 861 } 862 if (LIST_EMPTY(&sc->sc_iflist)) { 863 sc->sc_ifp->if_mtu = ifr->ifr_mtu; 864 break; 865 } 866 BRIDGE_LOCK(sc); 867 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 868 if (bif->bif_ifp->if_mtu != ifr->ifr_mtu) { 869 log(LOG_NOTICE, "%s: invalid MTU: %u(%s)" 870 " != %d\n", sc->sc_ifp->if_xname, 871 bif->bif_ifp->if_mtu, 872 bif->bif_ifp->if_xname, ifr->ifr_mtu); 873 error = EINVAL; 874 break; 875 } 876 } 877 if (!error) 878 sc->sc_ifp->if_mtu = ifr->ifr_mtu; 879 BRIDGE_UNLOCK(sc); 880 break; 881 default: 882 /* 883 * drop the lock as ether_ioctl() will call bridge_start() and 884 * cause the lock to be recursed. 885 */ 886 error = ether_ioctl(ifp, cmd, data); 887 break; 888 } 889 890 return (error); 891 } 892 893 /* 894 * bridge_mutecaps: 895 * 896 * Clear or restore unwanted capabilities on the member interface 897 */ 898 static void 899 bridge_mutecaps(struct bridge_softc *sc) 900 { 901 struct bridge_iflist *bif; 902 int enabled, mask; 903 904 /* Initial bitmask of capabilities to test */ 905 mask = BRIDGE_IFCAPS_MASK; 906 907 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 908 /* Every member must support it or its disabled */ 909 mask &= bif->bif_savedcaps; 910 } 911 912 BRIDGE_XLOCK(sc); 913 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 914 enabled = bif->bif_ifp->if_capenable; 915 enabled &= ~BRIDGE_IFCAPS_STRIP; 916 /* strip off mask bits and enable them again if allowed */ 917 enabled &= ~BRIDGE_IFCAPS_MASK; 918 enabled |= mask; 919 BRIDGE_UNLOCK(sc); 920 bridge_set_ifcap(sc, bif, enabled); 921 BRIDGE_LOCK(sc); 922 } 923 BRIDGE_XDROP(sc); 924 925 } 926 927 static void 928 bridge_set_ifcap(struct bridge_softc *sc, struct bridge_iflist *bif, int set) 929 { 930 struct ifnet *ifp = bif->bif_ifp; 931 struct ifreq ifr; 932 int error; 933 934 BRIDGE_UNLOCK_ASSERT(sc); 935 936 bzero(&ifr, sizeof(ifr)); 937 ifr.ifr_reqcap = set; 938 939 if (ifp->if_capenable != set) { 940 error = (*ifp->if_ioctl)(ifp, SIOCSIFCAP, (caddr_t)&ifr); 941 if (error) 942 if_printf(sc->sc_ifp, 943 "error setting capabilities on %s: %d\n", 944 ifp->if_xname, error); 945 if ((ifp->if_capenable & ~set) != 0) 946 if_printf(sc->sc_ifp, 947 "can't disable some capabilities on %s: 0x%x\n", 948 ifp->if_xname, ifp->if_capenable & ~set); 949 } 950 } 951 952 /* 953 * bridge_lookup_member: 954 * 955 * Lookup a bridge member interface. 956 */ 957 static struct bridge_iflist * 958 bridge_lookup_member(struct bridge_softc *sc, const char *name) 959 { 960 struct bridge_iflist *bif; 961 struct ifnet *ifp; 962 963 BRIDGE_LOCK_ASSERT(sc); 964 965 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 966 ifp = bif->bif_ifp; 967 if (strcmp(ifp->if_xname, name) == 0) 968 return (bif); 969 } 970 971 return (NULL); 972 } 973 974 /* 975 * bridge_lookup_member_if: 976 * 977 * Lookup a bridge member interface by ifnet*. 978 */ 979 static struct bridge_iflist * 980 bridge_lookup_member_if(struct bridge_softc *sc, struct ifnet *member_ifp) 981 { 982 struct bridge_iflist *bif; 983 984 BRIDGE_LOCK_ASSERT(sc); 985 986 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 987 if (bif->bif_ifp == member_ifp) 988 return (bif); 989 } 990 991 return (NULL); 992 } 993 994 /* 995 * bridge_delete_member: 996 * 997 * Delete the specified member interface. 998 */ 999 static void 1000 bridge_delete_member(struct bridge_softc *sc, struct bridge_iflist *bif, 1001 int gone) 1002 { 1003 struct ifnet *ifs = bif->bif_ifp; 1004 struct ifnet *fif = NULL; 1005 1006 BRIDGE_LOCK_ASSERT(sc); 1007 1008 if (bif->bif_flags & IFBIF_STP) 1009 bstp_disable(&bif->bif_stp); 1010 1011 ifs->if_bridge = NULL; 1012 BRIDGE_XLOCK(sc); 1013 LIST_REMOVE(bif, bif_next); 1014 BRIDGE_XDROP(sc); 1015 1016 /* 1017 * If removing the interface that gave the bridge its mac address, set 1018 * the mac address of the bridge to the address of the next member, or 1019 * to its default address if no members are left. 1020 */ 1021 if (V_bridge_inherit_mac && sc->sc_ifaddr == ifs) { 1022 if (LIST_EMPTY(&sc->sc_iflist)) { 1023 bcopy(sc->sc_defaddr, 1024 IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN); 1025 sc->sc_ifaddr = NULL; 1026 } else { 1027 fif = LIST_FIRST(&sc->sc_iflist)->bif_ifp; 1028 bcopy(IF_LLADDR(fif), 1029 IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN); 1030 sc->sc_ifaddr = fif; 1031 } 1032 EVENTHANDLER_INVOKE(iflladdr_event, sc->sc_ifp); 1033 } 1034 1035 bridge_linkcheck(sc); 1036 bridge_mutecaps(sc); /* recalcuate now this interface is removed */ 1037 bridge_rtdelete(sc, ifs, IFBF_FLUSHALL); 1038 KASSERT(bif->bif_addrcnt == 0, 1039 ("%s: %d bridge routes referenced", __func__, bif->bif_addrcnt)); 1040 1041 BRIDGE_UNLOCK(sc); 1042 if (!gone) { 1043 switch (ifs->if_type) { 1044 case IFT_ETHER: 1045 case IFT_L2VLAN: 1046 /* 1047 * Take the interface out of promiscuous mode, but only 1048 * if it was promiscuous in the first place. It might 1049 * not be if we're in the bridge_ioctl_add() error path. 1050 */ 1051 if (ifs->if_flags & IFF_PROMISC) 1052 (void) ifpromisc(ifs, 0); 1053 break; 1054 1055 case IFT_GIF: 1056 break; 1057 1058 default: 1059 #ifdef DIAGNOSTIC 1060 panic("bridge_delete_member: impossible"); 1061 #endif 1062 break; 1063 } 1064 /* reneable any interface capabilities */ 1065 bridge_set_ifcap(sc, bif, bif->bif_savedcaps); 1066 } 1067 bstp_destroy(&bif->bif_stp); /* prepare to free */ 1068 BRIDGE_LOCK(sc); 1069 free(bif, M_DEVBUF); 1070 } 1071 1072 /* 1073 * bridge_delete_span: 1074 * 1075 * Delete the specified span interface. 1076 */ 1077 static void 1078 bridge_delete_span(struct bridge_softc *sc, struct bridge_iflist *bif) 1079 { 1080 BRIDGE_LOCK_ASSERT(sc); 1081 1082 KASSERT(bif->bif_ifp->if_bridge == NULL, 1083 ("%s: not a span interface", __func__)); 1084 1085 LIST_REMOVE(bif, bif_next); 1086 free(bif, M_DEVBUF); 1087 } 1088 1089 static int 1090 bridge_ioctl_add(struct bridge_softc *sc, void *arg) 1091 { 1092 struct ifbreq *req = arg; 1093 struct bridge_iflist *bif = NULL; 1094 struct ifnet *ifs; 1095 int error = 0; 1096 1097 ifs = ifunit(req->ifbr_ifsname); 1098 if (ifs == NULL) 1099 return (ENOENT); 1100 if (ifs->if_ioctl == NULL) /* must be supported */ 1101 return (EINVAL); 1102 1103 /* If it's in the span list, it can't be a member. */ 1104 LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) 1105 if (ifs == bif->bif_ifp) 1106 return (EBUSY); 1107 1108 if (ifs->if_bridge == sc) 1109 return (EEXIST); 1110 1111 if (ifs->if_bridge != NULL) 1112 return (EBUSY); 1113 1114 switch (ifs->if_type) { 1115 case IFT_ETHER: 1116 case IFT_L2VLAN: 1117 case IFT_GIF: 1118 /* permitted interface types */ 1119 break; 1120 default: 1121 return (EINVAL); 1122 } 1123 1124 #ifdef INET6 1125 /* 1126 * Two valid inet6 addresses with link-local scope must not be 1127 * on the parent interface and the member interfaces at the 1128 * same time. This restriction is needed to prevent violation 1129 * of link-local scope zone. Attempts to add a member 1130 * interface which has inet6 addresses when the parent has 1131 * inet6 triggers removal of all inet6 addresses on the member 1132 * interface. 1133 */ 1134 1135 /* Check if the parent interface has a link-local scope addr. */ 1136 if (V_allow_llz_overlap == 0 && 1137 in6ifa_llaonifp(sc->sc_ifp) != NULL) { 1138 /* 1139 * If any, remove all inet6 addresses from the member 1140 * interfaces. 1141 */ 1142 BRIDGE_XLOCK(sc); 1143 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 1144 if (in6ifa_llaonifp(bif->bif_ifp)) { 1145 BRIDGE_UNLOCK(sc); 1146 in6_ifdetach(bif->bif_ifp); 1147 BRIDGE_LOCK(sc); 1148 if_printf(sc->sc_ifp, 1149 "IPv6 addresses on %s have been removed " 1150 "before adding it as a member to prevent " 1151 "IPv6 address scope violation.\n", 1152 bif->bif_ifp->if_xname); 1153 } 1154 } 1155 BRIDGE_XDROP(sc); 1156 if (in6ifa_llaonifp(ifs)) { 1157 BRIDGE_UNLOCK(sc); 1158 in6_ifdetach(ifs); 1159 BRIDGE_LOCK(sc); 1160 if_printf(sc->sc_ifp, 1161 "IPv6 addresses on %s have been removed " 1162 "before adding it as a member to prevent " 1163 "IPv6 address scope violation.\n", 1164 ifs->if_xname); 1165 } 1166 } 1167 #endif 1168 /* Allow the first Ethernet member to define the MTU */ 1169 if (LIST_EMPTY(&sc->sc_iflist)) 1170 sc->sc_ifp->if_mtu = ifs->if_mtu; 1171 else if (sc->sc_ifp->if_mtu != ifs->if_mtu) { 1172 if_printf(sc->sc_ifp, "invalid MTU: %u(%s) != %u\n", 1173 ifs->if_mtu, ifs->if_xname, sc->sc_ifp->if_mtu); 1174 return (EINVAL); 1175 } 1176 1177 bif = malloc(sizeof(*bif), M_DEVBUF, M_NOWAIT|M_ZERO); 1178 if (bif == NULL) 1179 return (ENOMEM); 1180 1181 bif->bif_ifp = ifs; 1182 bif->bif_flags = IFBIF_LEARNING | IFBIF_DISCOVER; 1183 bif->bif_savedcaps = ifs->if_capenable; 1184 1185 /* 1186 * Assign the interface's MAC address to the bridge if it's the first 1187 * member and the MAC address of the bridge has not been changed from 1188 * the default randomly generated one. 1189 */ 1190 if (V_bridge_inherit_mac && LIST_EMPTY(&sc->sc_iflist) && 1191 !memcmp(IF_LLADDR(sc->sc_ifp), sc->sc_defaddr, ETHER_ADDR_LEN)) { 1192 bcopy(IF_LLADDR(ifs), IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN); 1193 sc->sc_ifaddr = ifs; 1194 EVENTHANDLER_INVOKE(iflladdr_event, sc->sc_ifp); 1195 } 1196 1197 ifs->if_bridge = sc; 1198 bstp_create(&sc->sc_stp, &bif->bif_stp, bif->bif_ifp); 1199 /* 1200 * XXX: XLOCK HERE!?! 1201 * 1202 * NOTE: insert_***HEAD*** should be safe for the traversals. 1203 */ 1204 LIST_INSERT_HEAD(&sc->sc_iflist, bif, bif_next); 1205 1206 /* Set interface capabilities to the intersection set of all members */ 1207 bridge_mutecaps(sc); 1208 bridge_linkcheck(sc); 1209 1210 /* Place the interface into promiscuous mode */ 1211 switch (ifs->if_type) { 1212 case IFT_ETHER: 1213 case IFT_L2VLAN: 1214 BRIDGE_UNLOCK(sc); 1215 error = ifpromisc(ifs, 1); 1216 BRIDGE_LOCK(sc); 1217 break; 1218 } 1219 1220 if (error) 1221 bridge_delete_member(sc, bif, 0); 1222 return (error); 1223 } 1224 1225 static int 1226 bridge_ioctl_del(struct bridge_softc *sc, void *arg) 1227 { 1228 struct ifbreq *req = arg; 1229 struct bridge_iflist *bif; 1230 1231 bif = bridge_lookup_member(sc, req->ifbr_ifsname); 1232 if (bif == NULL) 1233 return (ENOENT); 1234 1235 bridge_delete_member(sc, bif, 0); 1236 1237 return (0); 1238 } 1239 1240 static int 1241 bridge_ioctl_gifflags(struct bridge_softc *sc, void *arg) 1242 { 1243 struct ifbreq *req = arg; 1244 struct bridge_iflist *bif; 1245 struct bstp_port *bp; 1246 1247 bif = bridge_lookup_member(sc, req->ifbr_ifsname); 1248 if (bif == NULL) 1249 return (ENOENT); 1250 1251 bp = &bif->bif_stp; 1252 req->ifbr_ifsflags = bif->bif_flags; 1253 req->ifbr_state = bp->bp_state; 1254 req->ifbr_priority = bp->bp_priority; 1255 req->ifbr_path_cost = bp->bp_path_cost; 1256 req->ifbr_portno = bif->bif_ifp->if_index & 0xfff; 1257 req->ifbr_proto = bp->bp_protover; 1258 req->ifbr_role = bp->bp_role; 1259 req->ifbr_stpflags = bp->bp_flags; 1260 req->ifbr_addrcnt = bif->bif_addrcnt; 1261 req->ifbr_addrmax = bif->bif_addrmax; 1262 req->ifbr_addrexceeded = bif->bif_addrexceeded; 1263 1264 /* Copy STP state options as flags */ 1265 if (bp->bp_operedge) 1266 req->ifbr_ifsflags |= IFBIF_BSTP_EDGE; 1267 if (bp->bp_flags & BSTP_PORT_AUTOEDGE) 1268 req->ifbr_ifsflags |= IFBIF_BSTP_AUTOEDGE; 1269 if (bp->bp_ptp_link) 1270 req->ifbr_ifsflags |= IFBIF_BSTP_PTP; 1271 if (bp->bp_flags & BSTP_PORT_AUTOPTP) 1272 req->ifbr_ifsflags |= IFBIF_BSTP_AUTOPTP; 1273 if (bp->bp_flags & BSTP_PORT_ADMEDGE) 1274 req->ifbr_ifsflags |= IFBIF_BSTP_ADMEDGE; 1275 if (bp->bp_flags & BSTP_PORT_ADMCOST) 1276 req->ifbr_ifsflags |= IFBIF_BSTP_ADMCOST; 1277 return (0); 1278 } 1279 1280 static int 1281 bridge_ioctl_sifflags(struct bridge_softc *sc, void *arg) 1282 { 1283 struct ifbreq *req = arg; 1284 struct bridge_iflist *bif; 1285 struct bstp_port *bp; 1286 int error; 1287 1288 bif = bridge_lookup_member(sc, req->ifbr_ifsname); 1289 if (bif == NULL) 1290 return (ENOENT); 1291 bp = &bif->bif_stp; 1292 1293 if (req->ifbr_ifsflags & IFBIF_SPAN) 1294 /* SPAN is readonly */ 1295 return (EINVAL); 1296 1297 if (req->ifbr_ifsflags & IFBIF_STP) { 1298 if ((bif->bif_flags & IFBIF_STP) == 0) { 1299 error = bstp_enable(&bif->bif_stp); 1300 if (error) 1301 return (error); 1302 } 1303 } else { 1304 if ((bif->bif_flags & IFBIF_STP) != 0) 1305 bstp_disable(&bif->bif_stp); 1306 } 1307 1308 /* Pass on STP flags */ 1309 bstp_set_edge(bp, req->ifbr_ifsflags & IFBIF_BSTP_EDGE ? 1 : 0); 1310 bstp_set_autoedge(bp, req->ifbr_ifsflags & IFBIF_BSTP_AUTOEDGE ? 1 : 0); 1311 bstp_set_ptp(bp, req->ifbr_ifsflags & IFBIF_BSTP_PTP ? 1 : 0); 1312 bstp_set_autoptp(bp, req->ifbr_ifsflags & IFBIF_BSTP_AUTOPTP ? 1 : 0); 1313 1314 /* Save the bits relating to the bridge */ 1315 bif->bif_flags = req->ifbr_ifsflags & IFBIFMASK; 1316 1317 return (0); 1318 } 1319 1320 static int 1321 bridge_ioctl_scache(struct bridge_softc *sc, void *arg) 1322 { 1323 struct ifbrparam *param = arg; 1324 1325 sc->sc_brtmax = param->ifbrp_csize; 1326 bridge_rttrim(sc); 1327 1328 return (0); 1329 } 1330 1331 static int 1332 bridge_ioctl_gcache(struct bridge_softc *sc, void *arg) 1333 { 1334 struct ifbrparam *param = arg; 1335 1336 param->ifbrp_csize = sc->sc_brtmax; 1337 1338 return (0); 1339 } 1340 1341 static int 1342 bridge_ioctl_gifs(struct bridge_softc *sc, void *arg) 1343 { 1344 struct ifbifconf *bifc = arg; 1345 struct bridge_iflist *bif; 1346 struct ifbreq breq; 1347 char *buf, *outbuf; 1348 int count, buflen, len, error = 0; 1349 1350 count = 0; 1351 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) 1352 count++; 1353 LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) 1354 count++; 1355 1356 buflen = sizeof(breq) * count; 1357 if (bifc->ifbic_len == 0) { 1358 bifc->ifbic_len = buflen; 1359 return (0); 1360 } 1361 BRIDGE_UNLOCK(sc); 1362 outbuf = malloc(buflen, M_TEMP, M_WAITOK | M_ZERO); 1363 BRIDGE_LOCK(sc); 1364 1365 count = 0; 1366 buf = outbuf; 1367 len = min(bifc->ifbic_len, buflen); 1368 bzero(&breq, sizeof(breq)); 1369 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 1370 if (len < sizeof(breq)) 1371 break; 1372 1373 strlcpy(breq.ifbr_ifsname, bif->bif_ifp->if_xname, 1374 sizeof(breq.ifbr_ifsname)); 1375 /* Fill in the ifbreq structure */ 1376 error = bridge_ioctl_gifflags(sc, &breq); 1377 if (error) 1378 break; 1379 memcpy(buf, &breq, sizeof(breq)); 1380 count++; 1381 buf += sizeof(breq); 1382 len -= sizeof(breq); 1383 } 1384 LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) { 1385 if (len < sizeof(breq)) 1386 break; 1387 1388 strlcpy(breq.ifbr_ifsname, bif->bif_ifp->if_xname, 1389 sizeof(breq.ifbr_ifsname)); 1390 breq.ifbr_ifsflags = bif->bif_flags; 1391 breq.ifbr_portno = bif->bif_ifp->if_index & 0xfff; 1392 memcpy(buf, &breq, sizeof(breq)); 1393 count++; 1394 buf += sizeof(breq); 1395 len -= sizeof(breq); 1396 } 1397 1398 BRIDGE_UNLOCK(sc); 1399 bifc->ifbic_len = sizeof(breq) * count; 1400 error = copyout(outbuf, bifc->ifbic_req, bifc->ifbic_len); 1401 BRIDGE_LOCK(sc); 1402 free(outbuf, M_TEMP); 1403 return (error); 1404 } 1405 1406 static int 1407 bridge_ioctl_rts(struct bridge_softc *sc, void *arg) 1408 { 1409 struct ifbaconf *bac = arg; 1410 struct bridge_rtnode *brt; 1411 struct ifbareq bareq; 1412 char *buf, *outbuf; 1413 int count, buflen, len, error = 0; 1414 1415 if (bac->ifbac_len == 0) 1416 return (0); 1417 1418 count = 0; 1419 LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) 1420 count++; 1421 buflen = sizeof(bareq) * count; 1422 1423 BRIDGE_UNLOCK(sc); 1424 outbuf = malloc(buflen, M_TEMP, M_WAITOK | M_ZERO); 1425 BRIDGE_LOCK(sc); 1426 1427 count = 0; 1428 buf = outbuf; 1429 len = min(bac->ifbac_len, buflen); 1430 bzero(&bareq, sizeof(bareq)); 1431 LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) { 1432 if (len < sizeof(bareq)) 1433 goto out; 1434 strlcpy(bareq.ifba_ifsname, brt->brt_ifp->if_xname, 1435 sizeof(bareq.ifba_ifsname)); 1436 memcpy(bareq.ifba_dst, brt->brt_addr, sizeof(brt->brt_addr)); 1437 bareq.ifba_vlan = brt->brt_vlan; 1438 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC && 1439 time_uptime < brt->brt_expire) 1440 bareq.ifba_expire = brt->brt_expire - time_uptime; 1441 else 1442 bareq.ifba_expire = 0; 1443 bareq.ifba_flags = brt->brt_flags; 1444 1445 memcpy(buf, &bareq, sizeof(bareq)); 1446 count++; 1447 buf += sizeof(bareq); 1448 len -= sizeof(bareq); 1449 } 1450 out: 1451 BRIDGE_UNLOCK(sc); 1452 bac->ifbac_len = sizeof(bareq) * count; 1453 error = copyout(outbuf, bac->ifbac_req, bac->ifbac_len); 1454 BRIDGE_LOCK(sc); 1455 free(outbuf, M_TEMP); 1456 return (error); 1457 } 1458 1459 static int 1460 bridge_ioctl_saddr(struct bridge_softc *sc, void *arg) 1461 { 1462 struct ifbareq *req = arg; 1463 struct bridge_iflist *bif; 1464 int error; 1465 1466 bif = bridge_lookup_member(sc, req->ifba_ifsname); 1467 if (bif == NULL) 1468 return (ENOENT); 1469 1470 error = bridge_rtupdate(sc, req->ifba_dst, req->ifba_vlan, bif, 1, 1471 req->ifba_flags); 1472 1473 return (error); 1474 } 1475 1476 static int 1477 bridge_ioctl_sto(struct bridge_softc *sc, void *arg) 1478 { 1479 struct ifbrparam *param = arg; 1480 1481 sc->sc_brttimeout = param->ifbrp_ctime; 1482 return (0); 1483 } 1484 1485 static int 1486 bridge_ioctl_gto(struct bridge_softc *sc, void *arg) 1487 { 1488 struct ifbrparam *param = arg; 1489 1490 param->ifbrp_ctime = sc->sc_brttimeout; 1491 return (0); 1492 } 1493 1494 static int 1495 bridge_ioctl_daddr(struct bridge_softc *sc, void *arg) 1496 { 1497 struct ifbareq *req = arg; 1498 1499 return (bridge_rtdaddr(sc, req->ifba_dst, req->ifba_vlan)); 1500 } 1501 1502 static int 1503 bridge_ioctl_flush(struct bridge_softc *sc, void *arg) 1504 { 1505 struct ifbreq *req = arg; 1506 1507 bridge_rtflush(sc, req->ifbr_ifsflags); 1508 return (0); 1509 } 1510 1511 static int 1512 bridge_ioctl_gpri(struct bridge_softc *sc, void *arg) 1513 { 1514 struct ifbrparam *param = arg; 1515 struct bstp_state *bs = &sc->sc_stp; 1516 1517 param->ifbrp_prio = bs->bs_bridge_priority; 1518 return (0); 1519 } 1520 1521 static int 1522 bridge_ioctl_spri(struct bridge_softc *sc, void *arg) 1523 { 1524 struct ifbrparam *param = arg; 1525 1526 return (bstp_set_priority(&sc->sc_stp, param->ifbrp_prio)); 1527 } 1528 1529 static int 1530 bridge_ioctl_ght(struct bridge_softc *sc, void *arg) 1531 { 1532 struct ifbrparam *param = arg; 1533 struct bstp_state *bs = &sc->sc_stp; 1534 1535 param->ifbrp_hellotime = bs->bs_bridge_htime >> 8; 1536 return (0); 1537 } 1538 1539 static int 1540 bridge_ioctl_sht(struct bridge_softc *sc, void *arg) 1541 { 1542 struct ifbrparam *param = arg; 1543 1544 return (bstp_set_htime(&sc->sc_stp, param->ifbrp_hellotime)); 1545 } 1546 1547 static int 1548 bridge_ioctl_gfd(struct bridge_softc *sc, void *arg) 1549 { 1550 struct ifbrparam *param = arg; 1551 struct bstp_state *bs = &sc->sc_stp; 1552 1553 param->ifbrp_fwddelay = bs->bs_bridge_fdelay >> 8; 1554 return (0); 1555 } 1556 1557 static int 1558 bridge_ioctl_sfd(struct bridge_softc *sc, void *arg) 1559 { 1560 struct ifbrparam *param = arg; 1561 1562 return (bstp_set_fdelay(&sc->sc_stp, param->ifbrp_fwddelay)); 1563 } 1564 1565 static int 1566 bridge_ioctl_gma(struct bridge_softc *sc, void *arg) 1567 { 1568 struct ifbrparam *param = arg; 1569 struct bstp_state *bs = &sc->sc_stp; 1570 1571 param->ifbrp_maxage = bs->bs_bridge_max_age >> 8; 1572 return (0); 1573 } 1574 1575 static int 1576 bridge_ioctl_sma(struct bridge_softc *sc, void *arg) 1577 { 1578 struct ifbrparam *param = arg; 1579 1580 return (bstp_set_maxage(&sc->sc_stp, param->ifbrp_maxage)); 1581 } 1582 1583 static int 1584 bridge_ioctl_sifprio(struct bridge_softc *sc, void *arg) 1585 { 1586 struct ifbreq *req = arg; 1587 struct bridge_iflist *bif; 1588 1589 bif = bridge_lookup_member(sc, req->ifbr_ifsname); 1590 if (bif == NULL) 1591 return (ENOENT); 1592 1593 return (bstp_set_port_priority(&bif->bif_stp, req->ifbr_priority)); 1594 } 1595 1596 static int 1597 bridge_ioctl_sifcost(struct bridge_softc *sc, void *arg) 1598 { 1599 struct ifbreq *req = arg; 1600 struct bridge_iflist *bif; 1601 1602 bif = bridge_lookup_member(sc, req->ifbr_ifsname); 1603 if (bif == NULL) 1604 return (ENOENT); 1605 1606 return (bstp_set_path_cost(&bif->bif_stp, req->ifbr_path_cost)); 1607 } 1608 1609 static int 1610 bridge_ioctl_sifmaxaddr(struct bridge_softc *sc, void *arg) 1611 { 1612 struct ifbreq *req = arg; 1613 struct bridge_iflist *bif; 1614 1615 bif = bridge_lookup_member(sc, req->ifbr_ifsname); 1616 if (bif == NULL) 1617 return (ENOENT); 1618 1619 bif->bif_addrmax = req->ifbr_addrmax; 1620 return (0); 1621 } 1622 1623 static int 1624 bridge_ioctl_addspan(struct bridge_softc *sc, void *arg) 1625 { 1626 struct ifbreq *req = arg; 1627 struct bridge_iflist *bif = NULL; 1628 struct ifnet *ifs; 1629 1630 ifs = ifunit(req->ifbr_ifsname); 1631 if (ifs == NULL) 1632 return (ENOENT); 1633 1634 LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) 1635 if (ifs == bif->bif_ifp) 1636 return (EBUSY); 1637 1638 if (ifs->if_bridge != NULL) 1639 return (EBUSY); 1640 1641 switch (ifs->if_type) { 1642 case IFT_ETHER: 1643 case IFT_GIF: 1644 case IFT_L2VLAN: 1645 break; 1646 default: 1647 return (EINVAL); 1648 } 1649 1650 bif = malloc(sizeof(*bif), M_DEVBUF, M_NOWAIT|M_ZERO); 1651 if (bif == NULL) 1652 return (ENOMEM); 1653 1654 bif->bif_ifp = ifs; 1655 bif->bif_flags = IFBIF_SPAN; 1656 1657 LIST_INSERT_HEAD(&sc->sc_spanlist, bif, bif_next); 1658 1659 return (0); 1660 } 1661 1662 static int 1663 bridge_ioctl_delspan(struct bridge_softc *sc, void *arg) 1664 { 1665 struct ifbreq *req = arg; 1666 struct bridge_iflist *bif; 1667 struct ifnet *ifs; 1668 1669 ifs = ifunit(req->ifbr_ifsname); 1670 if (ifs == NULL) 1671 return (ENOENT); 1672 1673 LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) 1674 if (ifs == bif->bif_ifp) 1675 break; 1676 1677 if (bif == NULL) 1678 return (ENOENT); 1679 1680 bridge_delete_span(sc, bif); 1681 1682 return (0); 1683 } 1684 1685 static int 1686 bridge_ioctl_gbparam(struct bridge_softc *sc, void *arg) 1687 { 1688 struct ifbropreq *req = arg; 1689 struct bstp_state *bs = &sc->sc_stp; 1690 struct bstp_port *root_port; 1691 1692 req->ifbop_maxage = bs->bs_bridge_max_age >> 8; 1693 req->ifbop_hellotime = bs->bs_bridge_htime >> 8; 1694 req->ifbop_fwddelay = bs->bs_bridge_fdelay >> 8; 1695 1696 root_port = bs->bs_root_port; 1697 if (root_port == NULL) 1698 req->ifbop_root_port = 0; 1699 else 1700 req->ifbop_root_port = root_port->bp_ifp->if_index; 1701 1702 req->ifbop_holdcount = bs->bs_txholdcount; 1703 req->ifbop_priority = bs->bs_bridge_priority; 1704 req->ifbop_protocol = bs->bs_protover; 1705 req->ifbop_root_path_cost = bs->bs_root_pv.pv_cost; 1706 req->ifbop_bridgeid = bs->bs_bridge_pv.pv_dbridge_id; 1707 req->ifbop_designated_root = bs->bs_root_pv.pv_root_id; 1708 req->ifbop_designated_bridge = bs->bs_root_pv.pv_dbridge_id; 1709 req->ifbop_last_tc_time.tv_sec = bs->bs_last_tc_time.tv_sec; 1710 req->ifbop_last_tc_time.tv_usec = bs->bs_last_tc_time.tv_usec; 1711 1712 return (0); 1713 } 1714 1715 static int 1716 bridge_ioctl_grte(struct bridge_softc *sc, void *arg) 1717 { 1718 struct ifbrparam *param = arg; 1719 1720 param->ifbrp_cexceeded = sc->sc_brtexceeded; 1721 return (0); 1722 } 1723 1724 static int 1725 bridge_ioctl_gifsstp(struct bridge_softc *sc, void *arg) 1726 { 1727 struct ifbpstpconf *bifstp = arg; 1728 struct bridge_iflist *bif; 1729 struct bstp_port *bp; 1730 struct ifbpstpreq bpreq; 1731 char *buf, *outbuf; 1732 int count, buflen, len, error = 0; 1733 1734 count = 0; 1735 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 1736 if ((bif->bif_flags & IFBIF_STP) != 0) 1737 count++; 1738 } 1739 1740 buflen = sizeof(bpreq) * count; 1741 if (bifstp->ifbpstp_len == 0) { 1742 bifstp->ifbpstp_len = buflen; 1743 return (0); 1744 } 1745 1746 BRIDGE_UNLOCK(sc); 1747 outbuf = malloc(buflen, M_TEMP, M_WAITOK | M_ZERO); 1748 BRIDGE_LOCK(sc); 1749 1750 count = 0; 1751 buf = outbuf; 1752 len = min(bifstp->ifbpstp_len, buflen); 1753 bzero(&bpreq, sizeof(bpreq)); 1754 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 1755 if (len < sizeof(bpreq)) 1756 break; 1757 1758 if ((bif->bif_flags & IFBIF_STP) == 0) 1759 continue; 1760 1761 bp = &bif->bif_stp; 1762 bpreq.ifbp_portno = bif->bif_ifp->if_index & 0xfff; 1763 bpreq.ifbp_fwd_trans = bp->bp_forward_transitions; 1764 bpreq.ifbp_design_cost = bp->bp_desg_pv.pv_cost; 1765 bpreq.ifbp_design_port = bp->bp_desg_pv.pv_port_id; 1766 bpreq.ifbp_design_bridge = bp->bp_desg_pv.pv_dbridge_id; 1767 bpreq.ifbp_design_root = bp->bp_desg_pv.pv_root_id; 1768 1769 memcpy(buf, &bpreq, sizeof(bpreq)); 1770 count++; 1771 buf += sizeof(bpreq); 1772 len -= sizeof(bpreq); 1773 } 1774 1775 BRIDGE_UNLOCK(sc); 1776 bifstp->ifbpstp_len = sizeof(bpreq) * count; 1777 error = copyout(outbuf, bifstp->ifbpstp_req, bifstp->ifbpstp_len); 1778 BRIDGE_LOCK(sc); 1779 free(outbuf, M_TEMP); 1780 return (error); 1781 } 1782 1783 static int 1784 bridge_ioctl_sproto(struct bridge_softc *sc, void *arg) 1785 { 1786 struct ifbrparam *param = arg; 1787 1788 return (bstp_set_protocol(&sc->sc_stp, param->ifbrp_proto)); 1789 } 1790 1791 static int 1792 bridge_ioctl_stxhc(struct bridge_softc *sc, void *arg) 1793 { 1794 struct ifbrparam *param = arg; 1795 1796 return (bstp_set_holdcount(&sc->sc_stp, param->ifbrp_txhc)); 1797 } 1798 1799 /* 1800 * bridge_ifdetach: 1801 * 1802 * Detach an interface from a bridge. Called when a member 1803 * interface is detaching. 1804 */ 1805 static void 1806 bridge_ifdetach(void *arg __unused, struct ifnet *ifp) 1807 { 1808 struct bridge_softc *sc = ifp->if_bridge; 1809 struct bridge_iflist *bif; 1810 1811 if (ifp->if_flags & IFF_RENAMING) 1812 return; 1813 if (V_bridge_cloner == NULL) { 1814 /* 1815 * This detach handler can be called after 1816 * vnet_bridge_uninit(). Just return in that case. 1817 */ 1818 return; 1819 } 1820 /* Check if the interface is a bridge member */ 1821 if (sc != NULL) { 1822 BRIDGE_LOCK(sc); 1823 1824 bif = bridge_lookup_member_if(sc, ifp); 1825 if (bif != NULL) 1826 bridge_delete_member(sc, bif, 1); 1827 1828 BRIDGE_UNLOCK(sc); 1829 return; 1830 } 1831 1832 /* Check if the interface is a span port */ 1833 BRIDGE_LIST_LOCK(); 1834 LIST_FOREACH(sc, &V_bridge_list, sc_list) { 1835 BRIDGE_LOCK(sc); 1836 LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) 1837 if (ifp == bif->bif_ifp) { 1838 bridge_delete_span(sc, bif); 1839 break; 1840 } 1841 1842 BRIDGE_UNLOCK(sc); 1843 } 1844 BRIDGE_LIST_UNLOCK(); 1845 } 1846 1847 /* 1848 * bridge_init: 1849 * 1850 * Initialize a bridge interface. 1851 */ 1852 static void 1853 bridge_init(void *xsc) 1854 { 1855 struct bridge_softc *sc = (struct bridge_softc *)xsc; 1856 struct ifnet *ifp = sc->sc_ifp; 1857 1858 if (ifp->if_drv_flags & IFF_DRV_RUNNING) 1859 return; 1860 1861 BRIDGE_LOCK(sc); 1862 callout_reset(&sc->sc_brcallout, bridge_rtable_prune_period * hz, 1863 bridge_timer, sc); 1864 1865 ifp->if_drv_flags |= IFF_DRV_RUNNING; 1866 bstp_init(&sc->sc_stp); /* Initialize Spanning Tree */ 1867 1868 BRIDGE_UNLOCK(sc); 1869 } 1870 1871 /* 1872 * bridge_stop: 1873 * 1874 * Stop the bridge interface. 1875 */ 1876 static void 1877 bridge_stop(struct ifnet *ifp, int disable) 1878 { 1879 struct bridge_softc *sc = ifp->if_softc; 1880 1881 BRIDGE_LOCK_ASSERT(sc); 1882 1883 if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) 1884 return; 1885 1886 callout_stop(&sc->sc_brcallout); 1887 bstp_stop(&sc->sc_stp); 1888 1889 bridge_rtflush(sc, IFBF_FLUSHDYN); 1890 1891 ifp->if_drv_flags &= ~IFF_DRV_RUNNING; 1892 } 1893 1894 /* 1895 * bridge_enqueue: 1896 * 1897 * Enqueue a packet on a bridge member interface. 1898 * 1899 */ 1900 static int 1901 bridge_enqueue(struct bridge_softc *sc, struct ifnet *dst_ifp, struct mbuf *m) 1902 { 1903 int len, err = 0; 1904 short mflags; 1905 struct mbuf *m0; 1906 1907 /* We may be sending a fragment so traverse the mbuf */ 1908 for (; m; m = m0) { 1909 m0 = m->m_nextpkt; 1910 m->m_nextpkt = NULL; 1911 len = m->m_pkthdr.len; 1912 mflags = m->m_flags; 1913 1914 /* 1915 * If underlying interface can not do VLAN tag insertion itself 1916 * then attach a packet tag that holds it. 1917 */ 1918 if ((m->m_flags & M_VLANTAG) && 1919 (dst_ifp->if_capenable & IFCAP_VLAN_HWTAGGING) == 0) { 1920 m = ether_vlanencap(m, m->m_pkthdr.ether_vtag); 1921 if (m == NULL) { 1922 if_printf(dst_ifp, 1923 "unable to prepend VLAN header\n"); 1924 if_inc_counter(dst_ifp, IFCOUNTER_OERRORS, 1); 1925 continue; 1926 } 1927 m->m_flags &= ~M_VLANTAG; 1928 } 1929 1930 M_ASSERTPKTHDR(m); /* We shouldn't transmit mbuf without pkthdr */ 1931 if ((err = dst_ifp->if_transmit(dst_ifp, m))) { 1932 m_freem(m0); 1933 if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1); 1934 break; 1935 } 1936 1937 if_inc_counter(sc->sc_ifp, IFCOUNTER_OPACKETS, 1); 1938 if_inc_counter(sc->sc_ifp, IFCOUNTER_OBYTES, len); 1939 if (mflags & M_MCAST) 1940 if_inc_counter(sc->sc_ifp, IFCOUNTER_OMCASTS, 1); 1941 } 1942 1943 return (err); 1944 } 1945 1946 /* 1947 * bridge_dummynet: 1948 * 1949 * Receive a queued packet from dummynet and pass it on to the output 1950 * interface. 1951 * 1952 * The mbuf has the Ethernet header already attached. 1953 */ 1954 static void 1955 bridge_dummynet(struct mbuf *m, struct ifnet *ifp) 1956 { 1957 struct bridge_softc *sc; 1958 1959 sc = ifp->if_bridge; 1960 1961 /* 1962 * The packet didnt originate from a member interface. This should only 1963 * ever happen if a member interface is removed while packets are 1964 * queued for it. 1965 */ 1966 if (sc == NULL) { 1967 m_freem(m); 1968 return; 1969 } 1970 1971 if (PFIL_HOOKED(&V_inet_pfil_hook) 1972 #ifdef INET6 1973 || PFIL_HOOKED(&V_inet6_pfil_hook) 1974 #endif 1975 ) { 1976 if (bridge_pfil(&m, sc->sc_ifp, ifp, PFIL_OUT) != 0) 1977 return; 1978 if (m == NULL) 1979 return; 1980 } 1981 1982 bridge_enqueue(sc, ifp, m); 1983 } 1984 1985 /* 1986 * bridge_output: 1987 * 1988 * Send output from a bridge member interface. This 1989 * performs the bridging function for locally originated 1990 * packets. 1991 * 1992 * The mbuf has the Ethernet header already attached. We must 1993 * enqueue or free the mbuf before returning. 1994 */ 1995 static int 1996 bridge_output(struct ifnet *ifp, struct mbuf *m, struct sockaddr *sa, 1997 struct rtentry *rt) 1998 { 1999 struct ether_header *eh; 2000 struct ifnet *dst_if; 2001 struct bridge_softc *sc; 2002 uint16_t vlan; 2003 2004 if (m->m_len < ETHER_HDR_LEN) { 2005 m = m_pullup(m, ETHER_HDR_LEN); 2006 if (m == NULL) 2007 return (0); 2008 } 2009 2010 eh = mtod(m, struct ether_header *); 2011 sc = ifp->if_bridge; 2012 vlan = VLANTAGOF(m); 2013 2014 BRIDGE_LOCK(sc); 2015 2016 /* 2017 * If bridge is down, but the original output interface is up, 2018 * go ahead and send out that interface. Otherwise, the packet 2019 * is dropped below. 2020 */ 2021 if ((sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) { 2022 dst_if = ifp; 2023 goto sendunicast; 2024 } 2025 2026 /* 2027 * If the packet is a multicast, or we don't know a better way to 2028 * get there, send to all interfaces. 2029 */ 2030 if (ETHER_IS_MULTICAST(eh->ether_dhost)) 2031 dst_if = NULL; 2032 else 2033 dst_if = bridge_rtlookup(sc, eh->ether_dhost, vlan); 2034 if (dst_if == NULL) { 2035 struct bridge_iflist *bif; 2036 struct mbuf *mc; 2037 int error = 0, used = 0; 2038 2039 bridge_span(sc, m); 2040 2041 BRIDGE_LOCK2REF(sc, error); 2042 if (error) { 2043 m_freem(m); 2044 return (0); 2045 } 2046 2047 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 2048 dst_if = bif->bif_ifp; 2049 2050 if (dst_if->if_type == IFT_GIF) 2051 continue; 2052 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) 2053 continue; 2054 2055 /* 2056 * If this is not the original output interface, 2057 * and the interface is participating in spanning 2058 * tree, make sure the port is in a state that 2059 * allows forwarding. 2060 */ 2061 if (dst_if != ifp && (bif->bif_flags & IFBIF_STP) && 2062 bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) 2063 continue; 2064 2065 if (LIST_NEXT(bif, bif_next) == NULL) { 2066 used = 1; 2067 mc = m; 2068 } else { 2069 mc = m_copypacket(m, M_NOWAIT); 2070 if (mc == NULL) { 2071 if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1); 2072 continue; 2073 } 2074 } 2075 2076 bridge_enqueue(sc, dst_if, mc); 2077 } 2078 if (used == 0) 2079 m_freem(m); 2080 BRIDGE_UNREF(sc); 2081 return (0); 2082 } 2083 2084 sendunicast: 2085 /* 2086 * XXX Spanning tree consideration here? 2087 */ 2088 2089 bridge_span(sc, m); 2090 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) { 2091 m_freem(m); 2092 BRIDGE_UNLOCK(sc); 2093 return (0); 2094 } 2095 2096 BRIDGE_UNLOCK(sc); 2097 bridge_enqueue(sc, dst_if, m); 2098 return (0); 2099 } 2100 2101 /* 2102 * bridge_transmit: 2103 * 2104 * Do output on a bridge. 2105 * 2106 */ 2107 static int 2108 bridge_transmit(struct ifnet *ifp, struct mbuf *m) 2109 { 2110 struct bridge_softc *sc; 2111 struct ether_header *eh; 2112 struct ifnet *dst_if; 2113 int error = 0; 2114 2115 sc = ifp->if_softc; 2116 2117 ETHER_BPF_MTAP(ifp, m); 2118 2119 eh = mtod(m, struct ether_header *); 2120 2121 BRIDGE_LOCK(sc); 2122 if (((m->m_flags & (M_BCAST|M_MCAST)) == 0) && 2123 (dst_if = bridge_rtlookup(sc, eh->ether_dhost, 1)) != NULL) { 2124 BRIDGE_UNLOCK(sc); 2125 error = bridge_enqueue(sc, dst_if, m); 2126 } else 2127 bridge_broadcast(sc, ifp, m, 0); 2128 2129 return (error); 2130 } 2131 2132 /* 2133 * The ifp->if_qflush entry point for if_bridge(4) is no-op. 2134 */ 2135 static void 2136 bridge_qflush(struct ifnet *ifp __unused) 2137 { 2138 } 2139 2140 /* 2141 * bridge_forward: 2142 * 2143 * The forwarding function of the bridge. 2144 * 2145 * NOTE: Releases the lock on return. 2146 */ 2147 static void 2148 bridge_forward(struct bridge_softc *sc, struct bridge_iflist *sbif, 2149 struct mbuf *m) 2150 { 2151 struct bridge_iflist *dbif; 2152 struct ifnet *src_if, *dst_if, *ifp; 2153 struct ether_header *eh; 2154 uint16_t vlan; 2155 uint8_t *dst; 2156 int error; 2157 2158 src_if = m->m_pkthdr.rcvif; 2159 ifp = sc->sc_ifp; 2160 2161 if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1); 2162 if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len); 2163 vlan = VLANTAGOF(m); 2164 2165 if ((sbif->bif_flags & IFBIF_STP) && 2166 sbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) 2167 goto drop; 2168 2169 eh = mtod(m, struct ether_header *); 2170 dst = eh->ether_dhost; 2171 2172 /* If the interface is learning, record the address. */ 2173 if (sbif->bif_flags & IFBIF_LEARNING) { 2174 error = bridge_rtupdate(sc, eh->ether_shost, vlan, 2175 sbif, 0, IFBAF_DYNAMIC); 2176 /* 2177 * If the interface has addresses limits then deny any source 2178 * that is not in the cache. 2179 */ 2180 if (error && sbif->bif_addrmax) 2181 goto drop; 2182 } 2183 2184 if ((sbif->bif_flags & IFBIF_STP) != 0 && 2185 sbif->bif_stp.bp_state == BSTP_IFSTATE_LEARNING) 2186 goto drop; 2187 2188 /* 2189 * At this point, the port either doesn't participate 2190 * in spanning tree or it is in the forwarding state. 2191 */ 2192 2193 /* 2194 * If the packet is unicast, destined for someone on 2195 * "this" side of the bridge, drop it. 2196 */ 2197 if ((m->m_flags & (M_BCAST|M_MCAST)) == 0) { 2198 dst_if = bridge_rtlookup(sc, dst, vlan); 2199 if (src_if == dst_if) 2200 goto drop; 2201 } else { 2202 /* 2203 * Check if its a reserved multicast address, any address 2204 * listed in 802.1D section 7.12.6 may not be forwarded by the 2205 * bridge. 2206 * This is currently 01-80-C2-00-00-00 to 01-80-C2-00-00-0F 2207 */ 2208 if (dst[0] == 0x01 && dst[1] == 0x80 && 2209 dst[2] == 0xc2 && dst[3] == 0x00 && 2210 dst[4] == 0x00 && dst[5] <= 0x0f) 2211 goto drop; 2212 2213 /* ...forward it to all interfaces. */ 2214 if_inc_counter(ifp, IFCOUNTER_IMCASTS, 1); 2215 dst_if = NULL; 2216 } 2217 2218 /* 2219 * If we have a destination interface which is a member of our bridge, 2220 * OR this is a unicast packet, push it through the bpf(4) machinery. 2221 * For broadcast or multicast packets, don't bother because it will 2222 * be reinjected into ether_input. We do this before we pass the packets 2223 * through the pfil(9) framework, as it is possible that pfil(9) will 2224 * drop the packet, or possibly modify it, making it difficult to debug 2225 * firewall issues on the bridge. 2226 */ 2227 if (dst_if != NULL || (m->m_flags & (M_BCAST | M_MCAST)) == 0) 2228 ETHER_BPF_MTAP(ifp, m); 2229 2230 /* run the packet filter */ 2231 if (PFIL_HOOKED(&V_inet_pfil_hook) 2232 #ifdef INET6 2233 || PFIL_HOOKED(&V_inet6_pfil_hook) 2234 #endif 2235 ) { 2236 BRIDGE_UNLOCK(sc); 2237 if (bridge_pfil(&m, ifp, src_if, PFIL_IN) != 0) 2238 return; 2239 if (m == NULL) 2240 return; 2241 BRIDGE_LOCK(sc); 2242 } 2243 2244 if (dst_if == NULL) { 2245 bridge_broadcast(sc, src_if, m, 1); 2246 return; 2247 } 2248 2249 /* 2250 * At this point, we're dealing with a unicast frame 2251 * going to a different interface. 2252 */ 2253 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) 2254 goto drop; 2255 2256 dbif = bridge_lookup_member_if(sc, dst_if); 2257 if (dbif == NULL) 2258 /* Not a member of the bridge (anymore?) */ 2259 goto drop; 2260 2261 /* Private segments can not talk to each other */ 2262 if (sbif->bif_flags & dbif->bif_flags & IFBIF_PRIVATE) 2263 goto drop; 2264 2265 if ((dbif->bif_flags & IFBIF_STP) && 2266 dbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) 2267 goto drop; 2268 2269 BRIDGE_UNLOCK(sc); 2270 2271 if (PFIL_HOOKED(&V_inet_pfil_hook) 2272 #ifdef INET6 2273 || PFIL_HOOKED(&V_inet6_pfil_hook) 2274 #endif 2275 ) { 2276 if (bridge_pfil(&m, ifp, dst_if, PFIL_OUT) != 0) 2277 return; 2278 if (m == NULL) 2279 return; 2280 } 2281 2282 bridge_enqueue(sc, dst_if, m); 2283 return; 2284 2285 drop: 2286 BRIDGE_UNLOCK(sc); 2287 m_freem(m); 2288 } 2289 2290 /* 2291 * bridge_input: 2292 * 2293 * Receive input from a member interface. Queue the packet for 2294 * bridging if it is not for us. 2295 */ 2296 static struct mbuf * 2297 bridge_input(struct ifnet *ifp, struct mbuf *m) 2298 { 2299 struct bridge_softc *sc = ifp->if_bridge; 2300 struct bridge_iflist *bif, *bif2; 2301 struct ifnet *bifp; 2302 struct ether_header *eh; 2303 struct mbuf *mc, *mc2; 2304 uint16_t vlan; 2305 int error; 2306 2307 if ((sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) 2308 return (m); 2309 2310 bifp = sc->sc_ifp; 2311 vlan = VLANTAGOF(m); 2312 2313 /* 2314 * Implement support for bridge monitoring. If this flag has been 2315 * set on this interface, discard the packet once we push it through 2316 * the bpf(4) machinery, but before we do, increment the byte and 2317 * packet counters associated with this interface. 2318 */ 2319 if ((bifp->if_flags & IFF_MONITOR) != 0) { 2320 m->m_pkthdr.rcvif = bifp; 2321 ETHER_BPF_MTAP(bifp, m); 2322 if_inc_counter(bifp, IFCOUNTER_IPACKETS, 1); 2323 if_inc_counter(bifp, IFCOUNTER_IBYTES, m->m_pkthdr.len); 2324 m_freem(m); 2325 return (NULL); 2326 } 2327 BRIDGE_LOCK(sc); 2328 bif = bridge_lookup_member_if(sc, ifp); 2329 if (bif == NULL) { 2330 BRIDGE_UNLOCK(sc); 2331 return (m); 2332 } 2333 2334 eh = mtod(m, struct ether_header *); 2335 2336 bridge_span(sc, m); 2337 2338 if (m->m_flags & (M_BCAST|M_MCAST)) { 2339 /* Tap off 802.1D packets; they do not get forwarded. */ 2340 if (memcmp(eh->ether_dhost, bstp_etheraddr, 2341 ETHER_ADDR_LEN) == 0) { 2342 bstp_input(&bif->bif_stp, ifp, m); /* consumes mbuf */ 2343 BRIDGE_UNLOCK(sc); 2344 return (NULL); 2345 } 2346 2347 if ((bif->bif_flags & IFBIF_STP) && 2348 bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) { 2349 BRIDGE_UNLOCK(sc); 2350 return (m); 2351 } 2352 2353 /* 2354 * Make a deep copy of the packet and enqueue the copy 2355 * for bridge processing; return the original packet for 2356 * local processing. 2357 */ 2358 mc = m_dup(m, M_NOWAIT); 2359 if (mc == NULL) { 2360 BRIDGE_UNLOCK(sc); 2361 return (m); 2362 } 2363 2364 /* Perform the bridge forwarding function with the copy. */ 2365 bridge_forward(sc, bif, mc); 2366 2367 /* 2368 * Reinject the mbuf as arriving on the bridge so we have a 2369 * chance at claiming multicast packets. We can not loop back 2370 * here from ether_input as a bridge is never a member of a 2371 * bridge. 2372 */ 2373 KASSERT(bifp->if_bridge == NULL, 2374 ("loop created in bridge_input")); 2375 mc2 = m_dup(m, M_NOWAIT); 2376 if (mc2 != NULL) { 2377 /* Keep the layer3 header aligned */ 2378 int i = min(mc2->m_pkthdr.len, max_protohdr); 2379 mc2 = m_copyup(mc2, i, ETHER_ALIGN); 2380 } 2381 if (mc2 != NULL) { 2382 mc2->m_pkthdr.rcvif = bifp; 2383 (*bifp->if_input)(bifp, mc2); 2384 } 2385 2386 /* Return the original packet for local processing. */ 2387 return (m); 2388 } 2389 2390 if ((bif->bif_flags & IFBIF_STP) && 2391 bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) { 2392 BRIDGE_UNLOCK(sc); 2393 return (m); 2394 } 2395 2396 #if (defined(INET) || defined(INET6)) 2397 # define OR_CARP_CHECK_WE_ARE_DST(iface) \ 2398 || ((iface)->if_carp \ 2399 && (*carp_forus_p)((iface), eh->ether_dhost)) 2400 # define OR_CARP_CHECK_WE_ARE_SRC(iface) \ 2401 || ((iface)->if_carp \ 2402 && (*carp_forus_p)((iface), eh->ether_shost)) 2403 #else 2404 # define OR_CARP_CHECK_WE_ARE_DST(iface) 2405 # define OR_CARP_CHECK_WE_ARE_SRC(iface) 2406 #endif 2407 2408 #ifdef INET6 2409 # define OR_PFIL_HOOKED_INET6 \ 2410 || PFIL_HOOKED(&V_inet6_pfil_hook) 2411 #else 2412 # define OR_PFIL_HOOKED_INET6 2413 #endif 2414 2415 #define GRAB_OUR_PACKETS(iface) \ 2416 if ((iface)->if_type == IFT_GIF) \ 2417 continue; \ 2418 /* It is destined for us. */ \ 2419 if (memcmp(IF_LLADDR((iface)), eh->ether_dhost, ETHER_ADDR_LEN) == 0 \ 2420 OR_CARP_CHECK_WE_ARE_DST((iface)) \ 2421 ) { \ 2422 if ((iface)->if_type == IFT_BRIDGE) { \ 2423 ETHER_BPF_MTAP(iface, m); \ 2424 if_inc_counter(iface, IFCOUNTER_IPACKETS, 1); \ 2425 if_inc_counter(iface, IFCOUNTER_IBYTES, m->m_pkthdr.len); \ 2426 /* Filter on the physical interface. */ \ 2427 if (V_pfil_local_phys && \ 2428 (PFIL_HOOKED(&V_inet_pfil_hook) \ 2429 OR_PFIL_HOOKED_INET6)) { \ 2430 if (bridge_pfil(&m, NULL, ifp, \ 2431 PFIL_IN) != 0 || m == NULL) { \ 2432 BRIDGE_UNLOCK(sc); \ 2433 return (NULL); \ 2434 } \ 2435 eh = mtod(m, struct ether_header *); \ 2436 } \ 2437 } \ 2438 if (bif->bif_flags & IFBIF_LEARNING) { \ 2439 error = bridge_rtupdate(sc, eh->ether_shost, \ 2440 vlan, bif, 0, IFBAF_DYNAMIC); \ 2441 if (error && bif->bif_addrmax) { \ 2442 BRIDGE_UNLOCK(sc); \ 2443 m_freem(m); \ 2444 return (NULL); \ 2445 } \ 2446 } \ 2447 m->m_pkthdr.rcvif = iface; \ 2448 BRIDGE_UNLOCK(sc); \ 2449 return (m); \ 2450 } \ 2451 \ 2452 /* We just received a packet that we sent out. */ \ 2453 if (memcmp(IF_LLADDR((iface)), eh->ether_shost, ETHER_ADDR_LEN) == 0 \ 2454 OR_CARP_CHECK_WE_ARE_SRC((iface)) \ 2455 ) { \ 2456 BRIDGE_UNLOCK(sc); \ 2457 m_freem(m); \ 2458 return (NULL); \ 2459 } 2460 2461 /* 2462 * Unicast. Make sure it's not for the bridge. 2463 */ 2464 do { GRAB_OUR_PACKETS(bifp) } while (0); 2465 2466 /* 2467 * Give a chance for ifp at first priority. This will help when the 2468 * packet comes through the interface like VLAN's with the same MACs 2469 * on several interfaces from the same bridge. This also will save 2470 * some CPU cycles in case the destination interface and the input 2471 * interface (eq ifp) are the same. 2472 */ 2473 do { GRAB_OUR_PACKETS(ifp) } while (0); 2474 2475 /* Now check the all bridge members. */ 2476 LIST_FOREACH(bif2, &sc->sc_iflist, bif_next) { 2477 GRAB_OUR_PACKETS(bif2->bif_ifp) 2478 } 2479 2480 #undef OR_CARP_CHECK_WE_ARE_DST 2481 #undef OR_CARP_CHECK_WE_ARE_SRC 2482 #undef OR_PFIL_HOOKED_INET6 2483 #undef GRAB_OUR_PACKETS 2484 2485 /* Perform the bridge forwarding function. */ 2486 bridge_forward(sc, bif, m); 2487 2488 return (NULL); 2489 } 2490 2491 /* 2492 * bridge_broadcast: 2493 * 2494 * Send a frame to all interfaces that are members of 2495 * the bridge, except for the one on which the packet 2496 * arrived. 2497 * 2498 * NOTE: Releases the lock on return. 2499 */ 2500 static void 2501 bridge_broadcast(struct bridge_softc *sc, struct ifnet *src_if, 2502 struct mbuf *m, int runfilt) 2503 { 2504 struct bridge_iflist *dbif, *sbif; 2505 struct mbuf *mc; 2506 struct ifnet *dst_if; 2507 int error = 0, used = 0, i; 2508 2509 sbif = bridge_lookup_member_if(sc, src_if); 2510 2511 BRIDGE_LOCK2REF(sc, error); 2512 if (error) { 2513 m_freem(m); 2514 return; 2515 } 2516 2517 /* Filter on the bridge interface before broadcasting */ 2518 if (runfilt && (PFIL_HOOKED(&V_inet_pfil_hook) 2519 #ifdef INET6 2520 || PFIL_HOOKED(&V_inet6_pfil_hook) 2521 #endif 2522 )) { 2523 if (bridge_pfil(&m, sc->sc_ifp, NULL, PFIL_OUT) != 0) 2524 goto out; 2525 if (m == NULL) 2526 goto out; 2527 } 2528 2529 LIST_FOREACH(dbif, &sc->sc_iflist, bif_next) { 2530 dst_if = dbif->bif_ifp; 2531 if (dst_if == src_if) 2532 continue; 2533 2534 /* Private segments can not talk to each other */ 2535 if (sbif && (sbif->bif_flags & dbif->bif_flags & IFBIF_PRIVATE)) 2536 continue; 2537 2538 if ((dbif->bif_flags & IFBIF_STP) && 2539 dbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) 2540 continue; 2541 2542 if ((dbif->bif_flags & IFBIF_DISCOVER) == 0 && 2543 (m->m_flags & (M_BCAST|M_MCAST)) == 0) 2544 continue; 2545 2546 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) 2547 continue; 2548 2549 if (LIST_NEXT(dbif, bif_next) == NULL) { 2550 mc = m; 2551 used = 1; 2552 } else { 2553 mc = m_dup(m, M_NOWAIT); 2554 if (mc == NULL) { 2555 if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1); 2556 continue; 2557 } 2558 } 2559 2560 /* 2561 * Filter on the output interface. Pass a NULL bridge interface 2562 * pointer so we do not redundantly filter on the bridge for 2563 * each interface we broadcast on. 2564 */ 2565 if (runfilt && (PFIL_HOOKED(&V_inet_pfil_hook) 2566 #ifdef INET6 2567 || PFIL_HOOKED(&V_inet6_pfil_hook) 2568 #endif 2569 )) { 2570 if (used == 0) { 2571 /* Keep the layer3 header aligned */ 2572 i = min(mc->m_pkthdr.len, max_protohdr); 2573 mc = m_copyup(mc, i, ETHER_ALIGN); 2574 if (mc == NULL) { 2575 if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1); 2576 continue; 2577 } 2578 } 2579 if (bridge_pfil(&mc, NULL, dst_if, PFIL_OUT) != 0) 2580 continue; 2581 if (mc == NULL) 2582 continue; 2583 } 2584 2585 bridge_enqueue(sc, dst_if, mc); 2586 } 2587 if (used == 0) 2588 m_freem(m); 2589 2590 out: 2591 BRIDGE_UNREF(sc); 2592 } 2593 2594 /* 2595 * bridge_span: 2596 * 2597 * Duplicate a packet out one or more interfaces that are in span mode, 2598 * the original mbuf is unmodified. 2599 */ 2600 static void 2601 bridge_span(struct bridge_softc *sc, struct mbuf *m) 2602 { 2603 struct bridge_iflist *bif; 2604 struct ifnet *dst_if; 2605 struct mbuf *mc; 2606 2607 if (LIST_EMPTY(&sc->sc_spanlist)) 2608 return; 2609 2610 LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) { 2611 dst_if = bif->bif_ifp; 2612 2613 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) 2614 continue; 2615 2616 mc = m_copypacket(m, M_NOWAIT); 2617 if (mc == NULL) { 2618 if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1); 2619 continue; 2620 } 2621 2622 bridge_enqueue(sc, dst_if, mc); 2623 } 2624 } 2625 2626 /* 2627 * bridge_rtupdate: 2628 * 2629 * Add a bridge routing entry. 2630 */ 2631 static int 2632 bridge_rtupdate(struct bridge_softc *sc, const uint8_t *dst, uint16_t vlan, 2633 struct bridge_iflist *bif, int setflags, uint8_t flags) 2634 { 2635 struct bridge_rtnode *brt; 2636 int error; 2637 2638 BRIDGE_LOCK_ASSERT(sc); 2639 2640 /* Check the source address is valid and not multicast. */ 2641 if (ETHER_IS_MULTICAST(dst) || 2642 (dst[0] == 0 && dst[1] == 0 && dst[2] == 0 && 2643 dst[3] == 0 && dst[4] == 0 && dst[5] == 0) != 0) 2644 return (EINVAL); 2645 2646 /* 802.1p frames map to vlan 1 */ 2647 if (vlan == 0) 2648 vlan = 1; 2649 2650 /* 2651 * A route for this destination might already exist. If so, 2652 * update it, otherwise create a new one. 2653 */ 2654 if ((brt = bridge_rtnode_lookup(sc, dst, vlan)) == NULL) { 2655 if (sc->sc_brtcnt >= sc->sc_brtmax) { 2656 sc->sc_brtexceeded++; 2657 return (ENOSPC); 2658 } 2659 /* Check per interface address limits (if enabled) */ 2660 if (bif->bif_addrmax && bif->bif_addrcnt >= bif->bif_addrmax) { 2661 bif->bif_addrexceeded++; 2662 return (ENOSPC); 2663 } 2664 2665 /* 2666 * Allocate a new bridge forwarding node, and 2667 * initialize the expiration time and Ethernet 2668 * address. 2669 */ 2670 brt = uma_zalloc(bridge_rtnode_zone, M_NOWAIT | M_ZERO); 2671 if (brt == NULL) 2672 return (ENOMEM); 2673 2674 if (bif->bif_flags & IFBIF_STICKY) 2675 brt->brt_flags = IFBAF_STICKY; 2676 else 2677 brt->brt_flags = IFBAF_DYNAMIC; 2678 2679 memcpy(brt->brt_addr, dst, ETHER_ADDR_LEN); 2680 brt->brt_vlan = vlan; 2681 2682 if ((error = bridge_rtnode_insert(sc, brt)) != 0) { 2683 uma_zfree(bridge_rtnode_zone, brt); 2684 return (error); 2685 } 2686 brt->brt_dst = bif; 2687 bif->bif_addrcnt++; 2688 } 2689 2690 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC && 2691 brt->brt_dst != bif) { 2692 brt->brt_dst->bif_addrcnt--; 2693 brt->brt_dst = bif; 2694 brt->brt_dst->bif_addrcnt++; 2695 } 2696 2697 if ((flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) 2698 brt->brt_expire = time_uptime + sc->sc_brttimeout; 2699 if (setflags) 2700 brt->brt_flags = flags; 2701 2702 return (0); 2703 } 2704 2705 /* 2706 * bridge_rtlookup: 2707 * 2708 * Lookup the destination interface for an address. 2709 */ 2710 static struct ifnet * 2711 bridge_rtlookup(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan) 2712 { 2713 struct bridge_rtnode *brt; 2714 2715 BRIDGE_LOCK_ASSERT(sc); 2716 2717 if ((brt = bridge_rtnode_lookup(sc, addr, vlan)) == NULL) 2718 return (NULL); 2719 2720 return (brt->brt_ifp); 2721 } 2722 2723 /* 2724 * bridge_rttrim: 2725 * 2726 * Trim the routine table so that we have a number 2727 * of routing entries less than or equal to the 2728 * maximum number. 2729 */ 2730 static void 2731 bridge_rttrim(struct bridge_softc *sc) 2732 { 2733 struct bridge_rtnode *brt, *nbrt; 2734 2735 BRIDGE_LOCK_ASSERT(sc); 2736 2737 /* Make sure we actually need to do this. */ 2738 if (sc->sc_brtcnt <= sc->sc_brtmax) 2739 return; 2740 2741 /* Force an aging cycle; this might trim enough addresses. */ 2742 bridge_rtage(sc); 2743 if (sc->sc_brtcnt <= sc->sc_brtmax) 2744 return; 2745 2746 LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) { 2747 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) { 2748 bridge_rtnode_destroy(sc, brt); 2749 if (sc->sc_brtcnt <= sc->sc_brtmax) 2750 return; 2751 } 2752 } 2753 } 2754 2755 /* 2756 * bridge_timer: 2757 * 2758 * Aging timer for the bridge. 2759 */ 2760 static void 2761 bridge_timer(void *arg) 2762 { 2763 struct bridge_softc *sc = arg; 2764 2765 BRIDGE_LOCK_ASSERT(sc); 2766 2767 bridge_rtage(sc); 2768 2769 if (sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING) 2770 callout_reset(&sc->sc_brcallout, 2771 bridge_rtable_prune_period * hz, bridge_timer, sc); 2772 } 2773 2774 /* 2775 * bridge_rtage: 2776 * 2777 * Perform an aging cycle. 2778 */ 2779 static void 2780 bridge_rtage(struct bridge_softc *sc) 2781 { 2782 struct bridge_rtnode *brt, *nbrt; 2783 2784 BRIDGE_LOCK_ASSERT(sc); 2785 2786 LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) { 2787 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) { 2788 if (time_uptime >= brt->brt_expire) 2789 bridge_rtnode_destroy(sc, brt); 2790 } 2791 } 2792 } 2793 2794 /* 2795 * bridge_rtflush: 2796 * 2797 * Remove all dynamic addresses from the bridge. 2798 */ 2799 static void 2800 bridge_rtflush(struct bridge_softc *sc, int full) 2801 { 2802 struct bridge_rtnode *brt, *nbrt; 2803 2804 BRIDGE_LOCK_ASSERT(sc); 2805 2806 LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) { 2807 if (full || (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) 2808 bridge_rtnode_destroy(sc, brt); 2809 } 2810 } 2811 2812 /* 2813 * bridge_rtdaddr: 2814 * 2815 * Remove an address from the table. 2816 */ 2817 static int 2818 bridge_rtdaddr(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan) 2819 { 2820 struct bridge_rtnode *brt; 2821 int found = 0; 2822 2823 BRIDGE_LOCK_ASSERT(sc); 2824 2825 /* 2826 * If vlan is zero then we want to delete for all vlans so the lookup 2827 * may return more than one. 2828 */ 2829 while ((brt = bridge_rtnode_lookup(sc, addr, vlan)) != NULL) { 2830 bridge_rtnode_destroy(sc, brt); 2831 found = 1; 2832 } 2833 2834 return (found ? 0 : ENOENT); 2835 } 2836 2837 /* 2838 * bridge_rtdelete: 2839 * 2840 * Delete routes to a speicifc member interface. 2841 */ 2842 static void 2843 bridge_rtdelete(struct bridge_softc *sc, struct ifnet *ifp, int full) 2844 { 2845 struct bridge_rtnode *brt, *nbrt; 2846 2847 BRIDGE_LOCK_ASSERT(sc); 2848 2849 LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) { 2850 if (brt->brt_ifp == ifp && (full || 2851 (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)) 2852 bridge_rtnode_destroy(sc, brt); 2853 } 2854 } 2855 2856 /* 2857 * bridge_rtable_init: 2858 * 2859 * Initialize the route table for this bridge. 2860 */ 2861 static void 2862 bridge_rtable_init(struct bridge_softc *sc) 2863 { 2864 int i; 2865 2866 sc->sc_rthash = malloc(sizeof(*sc->sc_rthash) * BRIDGE_RTHASH_SIZE, 2867 M_DEVBUF, M_WAITOK); 2868 2869 for (i = 0; i < BRIDGE_RTHASH_SIZE; i++) 2870 LIST_INIT(&sc->sc_rthash[i]); 2871 2872 sc->sc_rthash_key = arc4random(); 2873 LIST_INIT(&sc->sc_rtlist); 2874 } 2875 2876 /* 2877 * bridge_rtable_fini: 2878 * 2879 * Deconstruct the route table for this bridge. 2880 */ 2881 static void 2882 bridge_rtable_fini(struct bridge_softc *sc) 2883 { 2884 2885 KASSERT(sc->sc_brtcnt == 0, 2886 ("%s: %d bridge routes referenced", __func__, sc->sc_brtcnt)); 2887 free(sc->sc_rthash, M_DEVBUF); 2888 } 2889 2890 /* 2891 * The following hash function is adapted from "Hash Functions" by Bob Jenkins 2892 * ("Algorithm Alley", Dr. Dobbs Journal, September 1997). 2893 */ 2894 #define mix(a, b, c) \ 2895 do { \ 2896 a -= b; a -= c; a ^= (c >> 13); \ 2897 b -= c; b -= a; b ^= (a << 8); \ 2898 c -= a; c -= b; c ^= (b >> 13); \ 2899 a -= b; a -= c; a ^= (c >> 12); \ 2900 b -= c; b -= a; b ^= (a << 16); \ 2901 c -= a; c -= b; c ^= (b >> 5); \ 2902 a -= b; a -= c; a ^= (c >> 3); \ 2903 b -= c; b -= a; b ^= (a << 10); \ 2904 c -= a; c -= b; c ^= (b >> 15); \ 2905 } while (/*CONSTCOND*/0) 2906 2907 static __inline uint32_t 2908 bridge_rthash(struct bridge_softc *sc, const uint8_t *addr) 2909 { 2910 uint32_t a = 0x9e3779b9, b = 0x9e3779b9, c = sc->sc_rthash_key; 2911 2912 b += addr[5] << 8; 2913 b += addr[4]; 2914 a += addr[3] << 24; 2915 a += addr[2] << 16; 2916 a += addr[1] << 8; 2917 a += addr[0]; 2918 2919 mix(a, b, c); 2920 2921 return (c & BRIDGE_RTHASH_MASK); 2922 } 2923 2924 #undef mix 2925 2926 static int 2927 bridge_rtnode_addr_cmp(const uint8_t *a, const uint8_t *b) 2928 { 2929 int i, d; 2930 2931 for (i = 0, d = 0; i < ETHER_ADDR_LEN && d == 0; i++) { 2932 d = ((int)a[i]) - ((int)b[i]); 2933 } 2934 2935 return (d); 2936 } 2937 2938 /* 2939 * bridge_rtnode_lookup: 2940 * 2941 * Look up a bridge route node for the specified destination. Compare the 2942 * vlan id or if zero then just return the first match. 2943 */ 2944 static struct bridge_rtnode * 2945 bridge_rtnode_lookup(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan) 2946 { 2947 struct bridge_rtnode *brt; 2948 uint32_t hash; 2949 int dir; 2950 2951 BRIDGE_LOCK_ASSERT(sc); 2952 2953 hash = bridge_rthash(sc, addr); 2954 LIST_FOREACH(brt, &sc->sc_rthash[hash], brt_hash) { 2955 dir = bridge_rtnode_addr_cmp(addr, brt->brt_addr); 2956 if (dir == 0 && (brt->brt_vlan == vlan || vlan == 0)) 2957 return (brt); 2958 if (dir > 0) 2959 return (NULL); 2960 } 2961 2962 return (NULL); 2963 } 2964 2965 /* 2966 * bridge_rtnode_insert: 2967 * 2968 * Insert the specified bridge node into the route table. We 2969 * assume the entry is not already in the table. 2970 */ 2971 static int 2972 bridge_rtnode_insert(struct bridge_softc *sc, struct bridge_rtnode *brt) 2973 { 2974 struct bridge_rtnode *lbrt; 2975 uint32_t hash; 2976 int dir; 2977 2978 BRIDGE_LOCK_ASSERT(sc); 2979 2980 hash = bridge_rthash(sc, brt->brt_addr); 2981 2982 lbrt = LIST_FIRST(&sc->sc_rthash[hash]); 2983 if (lbrt == NULL) { 2984 LIST_INSERT_HEAD(&sc->sc_rthash[hash], brt, brt_hash); 2985 goto out; 2986 } 2987 2988 do { 2989 dir = bridge_rtnode_addr_cmp(brt->brt_addr, lbrt->brt_addr); 2990 if (dir == 0 && brt->brt_vlan == lbrt->brt_vlan) 2991 return (EEXIST); 2992 if (dir > 0) { 2993 LIST_INSERT_BEFORE(lbrt, brt, brt_hash); 2994 goto out; 2995 } 2996 if (LIST_NEXT(lbrt, brt_hash) == NULL) { 2997 LIST_INSERT_AFTER(lbrt, brt, brt_hash); 2998 goto out; 2999 } 3000 lbrt = LIST_NEXT(lbrt, brt_hash); 3001 } while (lbrt != NULL); 3002 3003 #ifdef DIAGNOSTIC 3004 panic("bridge_rtnode_insert: impossible"); 3005 #endif 3006 3007 out: 3008 LIST_INSERT_HEAD(&sc->sc_rtlist, brt, brt_list); 3009 sc->sc_brtcnt++; 3010 3011 return (0); 3012 } 3013 3014 /* 3015 * bridge_rtnode_destroy: 3016 * 3017 * Destroy a bridge rtnode. 3018 */ 3019 static void 3020 bridge_rtnode_destroy(struct bridge_softc *sc, struct bridge_rtnode *brt) 3021 { 3022 BRIDGE_LOCK_ASSERT(sc); 3023 3024 LIST_REMOVE(brt, brt_hash); 3025 3026 LIST_REMOVE(brt, brt_list); 3027 sc->sc_brtcnt--; 3028 brt->brt_dst->bif_addrcnt--; 3029 uma_zfree(bridge_rtnode_zone, brt); 3030 } 3031 3032 /* 3033 * bridge_rtable_expire: 3034 * 3035 * Set the expiry time for all routes on an interface. 3036 */ 3037 static void 3038 bridge_rtable_expire(struct ifnet *ifp, int age) 3039 { 3040 struct bridge_softc *sc = ifp->if_bridge; 3041 struct bridge_rtnode *brt; 3042 3043 BRIDGE_LOCK(sc); 3044 3045 /* 3046 * If the age is zero then flush, otherwise set all the expiry times to 3047 * age for the interface 3048 */ 3049 if (age == 0) 3050 bridge_rtdelete(sc, ifp, IFBF_FLUSHDYN); 3051 else { 3052 LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) { 3053 /* Cap the expiry time to 'age' */ 3054 if (brt->brt_ifp == ifp && 3055 brt->brt_expire > time_uptime + age && 3056 (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) 3057 brt->brt_expire = time_uptime + age; 3058 } 3059 } 3060 BRIDGE_UNLOCK(sc); 3061 } 3062 3063 /* 3064 * bridge_state_change: 3065 * 3066 * Callback from the bridgestp code when a port changes states. 3067 */ 3068 static void 3069 bridge_state_change(struct ifnet *ifp, int state) 3070 { 3071 struct bridge_softc *sc = ifp->if_bridge; 3072 static const char *stpstates[] = { 3073 "disabled", 3074 "listening", 3075 "learning", 3076 "forwarding", 3077 "blocking", 3078 "discarding" 3079 }; 3080 3081 CURVNET_SET(ifp->if_vnet); 3082 if (V_log_stp) 3083 log(LOG_NOTICE, "%s: state changed to %s on %s\n", 3084 sc->sc_ifp->if_xname, stpstates[state], ifp->if_xname); 3085 CURVNET_RESTORE(); 3086 } 3087 3088 /* 3089 * Send bridge packets through pfil if they are one of the types pfil can deal 3090 * with, or if they are ARP or REVARP. (pfil will pass ARP and REVARP without 3091 * question.) If *bifp or *ifp are NULL then packet filtering is skipped for 3092 * that interface. 3093 */ 3094 static int 3095 bridge_pfil(struct mbuf **mp, struct ifnet *bifp, struct ifnet *ifp, int dir) 3096 { 3097 int snap, error, i, hlen; 3098 struct ether_header *eh1, eh2; 3099 struct ip *ip; 3100 struct llc llc1; 3101 u_int16_t ether_type; 3102 3103 snap = 0; 3104 error = -1; /* Default error if not error == 0 */ 3105 3106 #if 0 3107 /* we may return with the IP fields swapped, ensure its not shared */ 3108 KASSERT(M_WRITABLE(*mp), ("%s: modifying a shared mbuf", __func__)); 3109 #endif 3110 3111 if (V_pfil_bridge == 0 && V_pfil_member == 0 && V_pfil_ipfw == 0) 3112 return (0); /* filtering is disabled */ 3113 3114 i = min((*mp)->m_pkthdr.len, max_protohdr); 3115 if ((*mp)->m_len < i) { 3116 *mp = m_pullup(*mp, i); 3117 if (*mp == NULL) { 3118 printf("%s: m_pullup failed\n", __func__); 3119 return (-1); 3120 } 3121 } 3122 3123 eh1 = mtod(*mp, struct ether_header *); 3124 ether_type = ntohs(eh1->ether_type); 3125 3126 /* 3127 * Check for SNAP/LLC. 3128 */ 3129 if (ether_type < ETHERMTU) { 3130 struct llc *llc2 = (struct llc *)(eh1 + 1); 3131 3132 if ((*mp)->m_len >= ETHER_HDR_LEN + 8 && 3133 llc2->llc_dsap == LLC_SNAP_LSAP && 3134 llc2->llc_ssap == LLC_SNAP_LSAP && 3135 llc2->llc_control == LLC_UI) { 3136 ether_type = htons(llc2->llc_un.type_snap.ether_type); 3137 snap = 1; 3138 } 3139 } 3140 3141 /* 3142 * If we're trying to filter bridge traffic, don't look at anything 3143 * other than IP and ARP traffic. If the filter doesn't understand 3144 * IPv6, don't allow IPv6 through the bridge either. This is lame 3145 * since if we really wanted, say, an AppleTalk filter, we are hosed, 3146 * but of course we don't have an AppleTalk filter to begin with. 3147 * (Note that since pfil doesn't understand ARP it will pass *ALL* 3148 * ARP traffic.) 3149 */ 3150 switch (ether_type) { 3151 case ETHERTYPE_ARP: 3152 case ETHERTYPE_REVARP: 3153 if (V_pfil_ipfw_arp == 0) 3154 return (0); /* Automatically pass */ 3155 break; 3156 3157 case ETHERTYPE_IP: 3158 #ifdef INET6 3159 case ETHERTYPE_IPV6: 3160 #endif /* INET6 */ 3161 break; 3162 default: 3163 /* 3164 * Check to see if the user wants to pass non-ip 3165 * packets, these will not be checked by pfil(9) and 3166 * passed unconditionally so the default is to drop. 3167 */ 3168 if (V_pfil_onlyip) 3169 goto bad; 3170 } 3171 3172 /* Run the packet through pfil before stripping link headers */ 3173 if (PFIL_HOOKED(&V_link_pfil_hook) && V_pfil_ipfw != 0 && 3174 dir == PFIL_OUT && ifp != NULL) { 3175 3176 error = pfil_run_hooks(&V_link_pfil_hook, mp, ifp, dir, NULL); 3177 3178 if (*mp == NULL || error != 0) /* packet consumed by filter */ 3179 return (error); 3180 } 3181 3182 /* Strip off the Ethernet header and keep a copy. */ 3183 m_copydata(*mp, 0, ETHER_HDR_LEN, (caddr_t) &eh2); 3184 m_adj(*mp, ETHER_HDR_LEN); 3185 3186 /* Strip off snap header, if present */ 3187 if (snap) { 3188 m_copydata(*mp, 0, sizeof(struct llc), (caddr_t) &llc1); 3189 m_adj(*mp, sizeof(struct llc)); 3190 } 3191 3192 /* 3193 * Check the IP header for alignment and errors 3194 */ 3195 if (dir == PFIL_IN) { 3196 switch (ether_type) { 3197 case ETHERTYPE_IP: 3198 error = bridge_ip_checkbasic(mp); 3199 break; 3200 #ifdef INET6 3201 case ETHERTYPE_IPV6: 3202 error = bridge_ip6_checkbasic(mp); 3203 break; 3204 #endif /* INET6 */ 3205 default: 3206 error = 0; 3207 } 3208 if (error) 3209 goto bad; 3210 } 3211 3212 error = 0; 3213 3214 /* 3215 * Run the packet through pfil 3216 */ 3217 switch (ether_type) { 3218 case ETHERTYPE_IP: 3219 /* 3220 * Run pfil on the member interface and the bridge, both can 3221 * be skipped by clearing pfil_member or pfil_bridge. 3222 * 3223 * Keep the order: 3224 * in_if -> bridge_if -> out_if 3225 */ 3226 if (V_pfil_bridge && dir == PFIL_OUT && bifp != NULL) 3227 error = pfil_run_hooks(&V_inet_pfil_hook, mp, bifp, 3228 dir, NULL); 3229 3230 if (*mp == NULL || error != 0) /* filter may consume */ 3231 break; 3232 3233 if (V_pfil_member && ifp != NULL) 3234 error = pfil_run_hooks(&V_inet_pfil_hook, mp, ifp, 3235 dir, NULL); 3236 3237 if (*mp == NULL || error != 0) /* filter may consume */ 3238 break; 3239 3240 if (V_pfil_bridge && dir == PFIL_IN && bifp != NULL) 3241 error = pfil_run_hooks(&V_inet_pfil_hook, mp, bifp, 3242 dir, NULL); 3243 3244 if (*mp == NULL || error != 0) /* filter may consume */ 3245 break; 3246 3247 /* check if we need to fragment the packet */ 3248 /* bridge_fragment generates a mbuf chain of packets */ 3249 /* that already include eth headers */ 3250 if (V_pfil_member && ifp != NULL && dir == PFIL_OUT) { 3251 i = (*mp)->m_pkthdr.len; 3252 if (i > ifp->if_mtu) { 3253 error = bridge_fragment(ifp, mp, &eh2, snap, 3254 &llc1); 3255 return (error); 3256 } 3257 } 3258 3259 /* Recalculate the ip checksum. */ 3260 ip = mtod(*mp, struct ip *); 3261 hlen = ip->ip_hl << 2; 3262 if (hlen < sizeof(struct ip)) 3263 goto bad; 3264 if (hlen > (*mp)->m_len) { 3265 if ((*mp = m_pullup(*mp, hlen)) == NULL) 3266 goto bad; 3267 ip = mtod(*mp, struct ip *); 3268 if (ip == NULL) 3269 goto bad; 3270 } 3271 ip->ip_sum = 0; 3272 if (hlen == sizeof(struct ip)) 3273 ip->ip_sum = in_cksum_hdr(ip); 3274 else 3275 ip->ip_sum = in_cksum(*mp, hlen); 3276 3277 break; 3278 #ifdef INET6 3279 case ETHERTYPE_IPV6: 3280 if (V_pfil_bridge && dir == PFIL_OUT && bifp != NULL) 3281 error = pfil_run_hooks(&V_inet6_pfil_hook, mp, bifp, 3282 dir, NULL); 3283 3284 if (*mp == NULL || error != 0) /* filter may consume */ 3285 break; 3286 3287 if (V_pfil_member && ifp != NULL) 3288 error = pfil_run_hooks(&V_inet6_pfil_hook, mp, ifp, 3289 dir, NULL); 3290 3291 if (*mp == NULL || error != 0) /* filter may consume */ 3292 break; 3293 3294 if (V_pfil_bridge && dir == PFIL_IN && bifp != NULL) 3295 error = pfil_run_hooks(&V_inet6_pfil_hook, mp, bifp, 3296 dir, NULL); 3297 break; 3298 #endif 3299 default: 3300 error = 0; 3301 break; 3302 } 3303 3304 if (*mp == NULL) 3305 return (error); 3306 if (error != 0) 3307 goto bad; 3308 3309 error = -1; 3310 3311 /* 3312 * Finally, put everything back the way it was and return 3313 */ 3314 if (snap) { 3315 M_PREPEND(*mp, sizeof(struct llc), M_NOWAIT); 3316 if (*mp == NULL) 3317 return (error); 3318 bcopy(&llc1, mtod(*mp, caddr_t), sizeof(struct llc)); 3319 } 3320 3321 M_PREPEND(*mp, ETHER_HDR_LEN, M_NOWAIT); 3322 if (*mp == NULL) 3323 return (error); 3324 bcopy(&eh2, mtod(*mp, caddr_t), ETHER_HDR_LEN); 3325 3326 return (0); 3327 3328 bad: 3329 m_freem(*mp); 3330 *mp = NULL; 3331 return (error); 3332 } 3333 3334 /* 3335 * Perform basic checks on header size since 3336 * pfil assumes ip_input has already processed 3337 * it for it. Cut-and-pasted from ip_input.c. 3338 * Given how simple the IPv6 version is, 3339 * does the IPv4 version really need to be 3340 * this complicated? 3341 * 3342 * XXX Should we update ipstat here, or not? 3343 * XXX Right now we update ipstat but not 3344 * XXX csum_counter. 3345 */ 3346 static int 3347 bridge_ip_checkbasic(struct mbuf **mp) 3348 { 3349 struct mbuf *m = *mp; 3350 struct ip *ip; 3351 int len, hlen; 3352 u_short sum; 3353 3354 if (*mp == NULL) 3355 return (-1); 3356 3357 if (IP_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) { 3358 if ((m = m_copyup(m, sizeof(struct ip), 3359 (max_linkhdr + 3) & ~3)) == NULL) { 3360 /* XXXJRT new stat, please */ 3361 KMOD_IPSTAT_INC(ips_toosmall); 3362 goto bad; 3363 } 3364 } else if (__predict_false(m->m_len < sizeof (struct ip))) { 3365 if ((m = m_pullup(m, sizeof (struct ip))) == NULL) { 3366 KMOD_IPSTAT_INC(ips_toosmall); 3367 goto bad; 3368 } 3369 } 3370 ip = mtod(m, struct ip *); 3371 if (ip == NULL) goto bad; 3372 3373 if (ip->ip_v != IPVERSION) { 3374 KMOD_IPSTAT_INC(ips_badvers); 3375 goto bad; 3376 } 3377 hlen = ip->ip_hl << 2; 3378 if (hlen < sizeof(struct ip)) { /* minimum header length */ 3379 KMOD_IPSTAT_INC(ips_badhlen); 3380 goto bad; 3381 } 3382 if (hlen > m->m_len) { 3383 if ((m = m_pullup(m, hlen)) == NULL) { 3384 KMOD_IPSTAT_INC(ips_badhlen); 3385 goto bad; 3386 } 3387 ip = mtod(m, struct ip *); 3388 if (ip == NULL) goto bad; 3389 } 3390 3391 if (m->m_pkthdr.csum_flags & CSUM_IP_CHECKED) { 3392 sum = !(m->m_pkthdr.csum_flags & CSUM_IP_VALID); 3393 } else { 3394 if (hlen == sizeof(struct ip)) { 3395 sum = in_cksum_hdr(ip); 3396 } else { 3397 sum = in_cksum(m, hlen); 3398 } 3399 } 3400 if (sum) { 3401 KMOD_IPSTAT_INC(ips_badsum); 3402 goto bad; 3403 } 3404 3405 /* Retrieve the packet length. */ 3406 len = ntohs(ip->ip_len); 3407 3408 /* 3409 * Check for additional length bogosity 3410 */ 3411 if (len < hlen) { 3412 KMOD_IPSTAT_INC(ips_badlen); 3413 goto bad; 3414 } 3415 3416 /* 3417 * Check that the amount of data in the buffers 3418 * is as at least much as the IP header would have us expect. 3419 * Drop packet if shorter than we expect. 3420 */ 3421 if (m->m_pkthdr.len < len) { 3422 KMOD_IPSTAT_INC(ips_tooshort); 3423 goto bad; 3424 } 3425 3426 /* Checks out, proceed */ 3427 *mp = m; 3428 return (0); 3429 3430 bad: 3431 *mp = m; 3432 return (-1); 3433 } 3434 3435 #ifdef INET6 3436 /* 3437 * Same as above, but for IPv6. 3438 * Cut-and-pasted from ip6_input.c. 3439 * XXX Should we update ip6stat, or not? 3440 */ 3441 static int 3442 bridge_ip6_checkbasic(struct mbuf **mp) 3443 { 3444 struct mbuf *m = *mp; 3445 struct ip6_hdr *ip6; 3446 3447 /* 3448 * If the IPv6 header is not aligned, slurp it up into a new 3449 * mbuf with space for link headers, in the event we forward 3450 * it. Otherwise, if it is aligned, make sure the entire base 3451 * IPv6 header is in the first mbuf of the chain. 3452 */ 3453 if (IP6_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) { 3454 struct ifnet *inifp = m->m_pkthdr.rcvif; 3455 if ((m = m_copyup(m, sizeof(struct ip6_hdr), 3456 (max_linkhdr + 3) & ~3)) == NULL) { 3457 /* XXXJRT new stat, please */ 3458 IP6STAT_INC(ip6s_toosmall); 3459 in6_ifstat_inc(inifp, ifs6_in_hdrerr); 3460 goto bad; 3461 } 3462 } else if (__predict_false(m->m_len < sizeof(struct ip6_hdr))) { 3463 struct ifnet *inifp = m->m_pkthdr.rcvif; 3464 if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) { 3465 IP6STAT_INC(ip6s_toosmall); 3466 in6_ifstat_inc(inifp, ifs6_in_hdrerr); 3467 goto bad; 3468 } 3469 } 3470 3471 ip6 = mtod(m, struct ip6_hdr *); 3472 3473 if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) { 3474 IP6STAT_INC(ip6s_badvers); 3475 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr); 3476 goto bad; 3477 } 3478 3479 /* Checks out, proceed */ 3480 *mp = m; 3481 return (0); 3482 3483 bad: 3484 *mp = m; 3485 return (-1); 3486 } 3487 #endif /* INET6 */ 3488 3489 /* 3490 * bridge_fragment: 3491 * 3492 * Fragment mbuf chain in multiple packets and prepend ethernet header. 3493 */ 3494 static int 3495 bridge_fragment(struct ifnet *ifp, struct mbuf **mp, struct ether_header *eh, 3496 int snap, struct llc *llc) 3497 { 3498 struct mbuf *m = *mp, *nextpkt = NULL, *mprev = NULL, *mcur = NULL; 3499 struct ip *ip; 3500 int error = -1; 3501 3502 if (m->m_len < sizeof(struct ip) && 3503 (m = m_pullup(m, sizeof(struct ip))) == NULL) 3504 goto dropit; 3505 ip = mtod(m, struct ip *); 3506 3507 m->m_pkthdr.csum_flags |= CSUM_IP; 3508 error = ip_fragment(ip, &m, ifp->if_mtu, ifp->if_hwassist); 3509 if (error) 3510 goto dropit; 3511 3512 /* 3513 * Walk the chain and re-add the Ethernet header for 3514 * each mbuf packet. 3515 */ 3516 for (mcur = m; mcur; mcur = mcur->m_nextpkt) { 3517 nextpkt = mcur->m_nextpkt; 3518 mcur->m_nextpkt = NULL; 3519 if (snap) { 3520 M_PREPEND(mcur, sizeof(struct llc), M_NOWAIT); 3521 if (mcur == NULL) { 3522 error = ENOBUFS; 3523 if (mprev != NULL) 3524 mprev->m_nextpkt = nextpkt; 3525 goto dropit; 3526 } 3527 bcopy(llc, mtod(mcur, caddr_t),sizeof(struct llc)); 3528 } 3529 3530 M_PREPEND(mcur, ETHER_HDR_LEN, M_NOWAIT); 3531 if (mcur == NULL) { 3532 error = ENOBUFS; 3533 if (mprev != NULL) 3534 mprev->m_nextpkt = nextpkt; 3535 goto dropit; 3536 } 3537 bcopy(eh, mtod(mcur, caddr_t), ETHER_HDR_LEN); 3538 3539 /* 3540 * The previous two M_PREPEND could have inserted one or two 3541 * mbufs in front so we have to update the previous packet's 3542 * m_nextpkt. 3543 */ 3544 mcur->m_nextpkt = nextpkt; 3545 if (mprev != NULL) 3546 mprev->m_nextpkt = mcur; 3547 else { 3548 /* The first mbuf in the original chain needs to be 3549 * updated. */ 3550 *mp = mcur; 3551 } 3552 mprev = mcur; 3553 } 3554 3555 KMOD_IPSTAT_INC(ips_fragmented); 3556 return (error); 3557 3558 dropit: 3559 for (mcur = *mp; mcur; mcur = m) { /* droping the full packet chain */ 3560 m = mcur->m_nextpkt; 3561 m_freem(mcur); 3562 } 3563 return (error); 3564 } 3565 3566 static void 3567 bridge_linkstate(struct ifnet *ifp) 3568 { 3569 struct bridge_softc *sc = ifp->if_bridge; 3570 struct bridge_iflist *bif; 3571 3572 BRIDGE_LOCK(sc); 3573 bif = bridge_lookup_member_if(sc, ifp); 3574 if (bif == NULL) { 3575 BRIDGE_UNLOCK(sc); 3576 return; 3577 } 3578 bridge_linkcheck(sc); 3579 BRIDGE_UNLOCK(sc); 3580 3581 bstp_linkstate(&bif->bif_stp); 3582 } 3583 3584 static void 3585 bridge_linkcheck(struct bridge_softc *sc) 3586 { 3587 struct bridge_iflist *bif; 3588 int new_link, hasls; 3589 3590 BRIDGE_LOCK_ASSERT(sc); 3591 new_link = LINK_STATE_DOWN; 3592 hasls = 0; 3593 /* Our link is considered up if at least one of our ports is active */ 3594 LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 3595 if (bif->bif_ifp->if_capabilities & IFCAP_LINKSTATE) 3596 hasls++; 3597 if (bif->bif_ifp->if_link_state == LINK_STATE_UP) { 3598 new_link = LINK_STATE_UP; 3599 break; 3600 } 3601 } 3602 if (!LIST_EMPTY(&sc->sc_iflist) && !hasls) { 3603 /* If no interfaces support link-state then we default to up */ 3604 new_link = LINK_STATE_UP; 3605 } 3606 if_link_state_change(sc->sc_ifp, new_link); 3607 } 3608