1 /* $NetBSD: if_bridge.c,v 1.31 2005/06/01 19:45:34 jdc Exp $ */ 2 3 /*- 4 * SPDX-License-Identifier: BSD-4-Clause 5 * 6 * Copyright 2001 Wasabi Systems, Inc. 7 * All rights reserved. 8 * 9 * Written by Jason R. Thorpe for Wasabi Systems, Inc. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 3. All advertising materials mentioning features or use of this software 20 * must display the following acknowledgement: 21 * This product includes software developed for the NetBSD Project by 22 * Wasabi Systems, Inc. 23 * 4. The name of Wasabi Systems, Inc. may not be used to endorse 24 * or promote products derived from this software without specific prior 25 * written permission. 26 * 27 * THIS SOFTWARE IS PROVIDED BY WASABI SYSTEMS, INC. ``AS IS'' AND 28 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 29 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 30 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL WASABI SYSTEMS, INC 31 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 32 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 33 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 34 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 35 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 36 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 37 * POSSIBILITY OF SUCH DAMAGE. 38 */ 39 40 /* 41 * Copyright (c) 1999, 2000 Jason L. Wright (jason@thought.net) 42 * All rights reserved. 43 * 44 * Redistribution and use in source and binary forms, with or without 45 * modification, are permitted provided that the following conditions 46 * are met: 47 * 1. Redistributions of source code must retain the above copyright 48 * notice, this list of conditions and the following disclaimer. 49 * 2. Redistributions in binary form must reproduce the above copyright 50 * notice, this list of conditions and the following disclaimer in the 51 * documentation and/or other materials provided with the distribution. 52 * 53 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR 54 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED 55 * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE 56 * DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, 57 * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES 58 * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR 59 * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 60 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, 61 * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN 62 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 63 * POSSIBILITY OF SUCH DAMAGE. 64 * 65 * OpenBSD: if_bridge.c,v 1.60 2001/06/15 03:38:33 itojun Exp 66 */ 67 68 /* 69 * Network interface bridge support. 70 * 71 * TODO: 72 * 73 * - Currently only supports Ethernet-like interfaces (Ethernet, 74 * 802.11, VLANs on Ethernet, etc.) Figure out a nice way 75 * to bridge other types of interfaces (maybe consider 76 * heterogeneous bridges). 77 */ 78 79 #include <sys/cdefs.h> 80 __FBSDID("$FreeBSD$"); 81 82 #include "opt_inet.h" 83 #include "opt_inet6.h" 84 85 #include <sys/param.h> 86 #include <sys/eventhandler.h> 87 #include <sys/mbuf.h> 88 #include <sys/malloc.h> 89 #include <sys/protosw.h> 90 #include <sys/systm.h> 91 #include <sys/jail.h> 92 #include <sys/time.h> 93 #include <sys/socket.h> /* for net/if.h */ 94 #include <sys/sockio.h> 95 #include <sys/ctype.h> /* string functions */ 96 #include <sys/kernel.h> 97 #include <sys/random.h> 98 #include <sys/syslog.h> 99 #include <sys/sysctl.h> 100 #include <vm/uma.h> 101 #include <sys/module.h> 102 #include <sys/priv.h> 103 #include <sys/proc.h> 104 #include <sys/lock.h> 105 #include <sys/mutex.h> 106 107 #include <net/bpf.h> 108 #include <net/if.h> 109 #include <net/if_clone.h> 110 #include <net/if_dl.h> 111 #include <net/if_types.h> 112 #include <net/if_var.h> 113 #include <net/pfil.h> 114 #include <net/vnet.h> 115 116 #include <netinet/in.h> 117 #include <netinet/in_systm.h> 118 #include <netinet/in_var.h> 119 #include <netinet/ip.h> 120 #include <netinet/ip_var.h> 121 #ifdef INET6 122 #include <netinet/ip6.h> 123 #include <netinet6/ip6_var.h> 124 #include <netinet6/in6_ifattach.h> 125 #endif 126 #if defined(INET) || defined(INET6) 127 #include <netinet/ip_carp.h> 128 #endif 129 #include <machine/in_cksum.h> 130 #include <netinet/if_ether.h> 131 #include <net/bridgestp.h> 132 #include <net/if_bridgevar.h> 133 #include <net/if_llc.h> 134 #include <net/if_vlan_var.h> 135 136 #include <net/route.h> 137 138 #ifdef INET6 139 /* 140 * XXX: declare here to avoid to include many inet6 related files.. 141 * should be more generalized? 142 */ 143 extern void nd6_setmtu(struct ifnet *); 144 #endif 145 146 /* 147 * Size of the route hash table. Must be a power of two. 148 */ 149 #ifndef BRIDGE_RTHASH_SIZE 150 #define BRIDGE_RTHASH_SIZE 1024 151 #endif 152 153 #define BRIDGE_RTHASH_MASK (BRIDGE_RTHASH_SIZE - 1) 154 155 /* 156 * Default maximum number of addresses to cache. 157 */ 158 #ifndef BRIDGE_RTABLE_MAX 159 #define BRIDGE_RTABLE_MAX 2000 160 #endif 161 162 /* 163 * Timeout (in seconds) for entries learned dynamically. 164 */ 165 #ifndef BRIDGE_RTABLE_TIMEOUT 166 #define BRIDGE_RTABLE_TIMEOUT (20 * 60) /* same as ARP */ 167 #endif 168 169 /* 170 * Number of seconds between walks of the route list. 171 */ 172 #ifndef BRIDGE_RTABLE_PRUNE_PERIOD 173 #define BRIDGE_RTABLE_PRUNE_PERIOD (5 * 60) 174 #endif 175 176 /* 177 * List of capabilities to possibly mask on the member interface. 178 */ 179 #define BRIDGE_IFCAPS_MASK (IFCAP_TOE|IFCAP_TSO|IFCAP_TXCSUM|\ 180 IFCAP_TXCSUM_IPV6) 181 182 /* 183 * List of capabilities to strip 184 */ 185 #define BRIDGE_IFCAPS_STRIP IFCAP_LRO 186 187 /* 188 * Bridge locking 189 * 190 * The bridge relies heavily on the epoch(9) system to protect its data 191 * structures. This means we can safely use CK_LISTs while in NET_EPOCH, but we 192 * must ensure there is only one writer at a time. 193 * 194 * That is: for read accesses we only need to be in NET_EPOCH, but for write 195 * accesses we must hold: 196 * 197 * - BRIDGE_RT_LOCK, for any change to bridge_rtnodes 198 * - BRIDGE_LOCK, for any other change 199 * 200 * The BRIDGE_LOCK is a sleepable lock, because it is held across ioctl() 201 * calls to bridge member interfaces and these ioctl()s can sleep. 202 * The BRIDGE_RT_LOCK is a non-sleepable mutex, because it is sometimes 203 * required while we're in NET_EPOCH and then we're not allowed to sleep. 204 */ 205 #define BRIDGE_LOCK_INIT(_sc) do { \ 206 sx_init(&(_sc)->sc_sx, "if_bridge"); \ 207 mtx_init(&(_sc)->sc_rt_mtx, "if_bridge rt", NULL, MTX_DEF); \ 208 } while (0) 209 #define BRIDGE_LOCK_DESTROY(_sc) do { \ 210 sx_destroy(&(_sc)->sc_sx); \ 211 mtx_destroy(&(_sc)->sc_rt_mtx); \ 212 } while (0) 213 #define BRIDGE_LOCK(_sc) sx_xlock(&(_sc)->sc_sx) 214 #define BRIDGE_UNLOCK(_sc) sx_xunlock(&(_sc)->sc_sx) 215 #define BRIDGE_LOCK_ASSERT(_sc) sx_assert(&(_sc)->sc_sx, SX_XLOCKED) 216 #define BRIDGE_LOCK_OR_NET_EPOCH_ASSERT(_sc) \ 217 MPASS(in_epoch(net_epoch_preempt) || sx_xlocked(&(_sc)->sc_sx)) 218 #define BRIDGE_UNLOCK_ASSERT(_sc) sx_assert(&(_sc)->sc_sx, SX_UNLOCKED) 219 #define BRIDGE_RT_LOCK(_sc) mtx_lock(&(_sc)->sc_rt_mtx) 220 #define BRIDGE_RT_UNLOCK(_sc) mtx_unlock(&(_sc)->sc_rt_mtx) 221 #define BRIDGE_RT_LOCK_ASSERT(_sc) mtx_assert(&(_sc)->sc_rt_mtx, MA_OWNED) 222 #define BRIDGE_RT_LOCK_OR_NET_EPOCH_ASSERT(_sc) \ 223 MPASS(in_epoch(net_epoch_preempt) || mtx_owned(&(_sc)->sc_rt_mtx)) 224 225 /* 226 * Bridge interface list entry. 227 */ 228 struct bridge_iflist { 229 CK_LIST_ENTRY(bridge_iflist) bif_next; 230 struct ifnet *bif_ifp; /* member if */ 231 struct bstp_port bif_stp; /* STP state */ 232 uint32_t bif_flags; /* member if flags */ 233 int bif_savedcaps; /* saved capabilities */ 234 uint32_t bif_addrmax; /* max # of addresses */ 235 uint32_t bif_addrcnt; /* cur. # of addresses */ 236 uint32_t bif_addrexceeded;/* # of address violations */ 237 struct epoch_context bif_epoch_ctx; 238 }; 239 240 /* 241 * Bridge route node. 242 */ 243 struct bridge_rtnode { 244 CK_LIST_ENTRY(bridge_rtnode) brt_hash; /* hash table linkage */ 245 CK_LIST_ENTRY(bridge_rtnode) brt_list; /* list linkage */ 246 struct bridge_iflist *brt_dst; /* destination if */ 247 unsigned long brt_expire; /* expiration time */ 248 uint8_t brt_flags; /* address flags */ 249 uint8_t brt_addr[ETHER_ADDR_LEN]; 250 uint16_t brt_vlan; /* vlan id */ 251 struct vnet *brt_vnet; 252 struct epoch_context brt_epoch_ctx; 253 }; 254 #define brt_ifp brt_dst->bif_ifp 255 256 /* 257 * Software state for each bridge. 258 */ 259 struct bridge_softc { 260 struct ifnet *sc_ifp; /* make this an interface */ 261 LIST_ENTRY(bridge_softc) sc_list; 262 struct sx sc_sx; 263 struct mtx sc_rt_mtx; 264 uint32_t sc_brtmax; /* max # of addresses */ 265 uint32_t sc_brtcnt; /* cur. # of addresses */ 266 uint32_t sc_brttimeout; /* rt timeout in seconds */ 267 struct callout sc_brcallout; /* bridge callout */ 268 CK_LIST_HEAD(, bridge_iflist) sc_iflist; /* member interface list */ 269 CK_LIST_HEAD(, bridge_rtnode) *sc_rthash; /* our forwarding table */ 270 CK_LIST_HEAD(, bridge_rtnode) sc_rtlist; /* list version of above */ 271 uint32_t sc_rthash_key; /* key for hash */ 272 CK_LIST_HEAD(, bridge_iflist) sc_spanlist; /* span ports list */ 273 struct bstp_state sc_stp; /* STP state */ 274 uint32_t sc_brtexceeded; /* # of cache drops */ 275 struct ifnet *sc_ifaddr; /* member mac copied from */ 276 struct ether_addr sc_defaddr; /* Default MAC address */ 277 struct epoch_context sc_epoch_ctx; 278 }; 279 280 VNET_DEFINE_STATIC(struct sx, bridge_list_sx); 281 #define V_bridge_list_sx VNET(bridge_list_sx) 282 static eventhandler_tag bridge_detach_cookie; 283 284 int bridge_rtable_prune_period = BRIDGE_RTABLE_PRUNE_PERIOD; 285 286 VNET_DEFINE_STATIC(uma_zone_t, bridge_rtnode_zone); 287 #define V_bridge_rtnode_zone VNET(bridge_rtnode_zone) 288 289 static int bridge_clone_create(struct if_clone *, int, caddr_t); 290 static void bridge_clone_destroy(struct ifnet *); 291 292 static int bridge_ioctl(struct ifnet *, u_long, caddr_t); 293 static void bridge_mutecaps(struct bridge_softc *); 294 static void bridge_set_ifcap(struct bridge_softc *, struct bridge_iflist *, 295 int); 296 static void bridge_ifdetach(void *arg __unused, struct ifnet *); 297 static void bridge_init(void *); 298 static void bridge_dummynet(struct mbuf *, struct ifnet *); 299 static void bridge_stop(struct ifnet *, int); 300 static int bridge_transmit(struct ifnet *, struct mbuf *); 301 #ifdef ALTQ 302 static void bridge_altq_start(if_t); 303 static int bridge_altq_transmit(if_t, struct mbuf *); 304 #endif 305 static void bridge_qflush(struct ifnet *); 306 static struct mbuf *bridge_input(struct ifnet *, struct mbuf *); 307 static int bridge_output(struct ifnet *, struct mbuf *, struct sockaddr *, 308 struct rtentry *); 309 static int bridge_enqueue(struct bridge_softc *, struct ifnet *, 310 struct mbuf *); 311 static void bridge_rtdelete(struct bridge_softc *, struct ifnet *ifp, int); 312 313 static void bridge_forward(struct bridge_softc *, struct bridge_iflist *, 314 struct mbuf *m); 315 316 static void bridge_timer(void *); 317 318 static void bridge_broadcast(struct bridge_softc *, struct ifnet *, 319 struct mbuf *, int); 320 static void bridge_span(struct bridge_softc *, struct mbuf *); 321 322 static int bridge_rtupdate(struct bridge_softc *, const uint8_t *, 323 uint16_t, struct bridge_iflist *, int, uint8_t); 324 static struct ifnet *bridge_rtlookup(struct bridge_softc *, const uint8_t *, 325 uint16_t); 326 static void bridge_rttrim(struct bridge_softc *); 327 static void bridge_rtage(struct bridge_softc *); 328 static void bridge_rtflush(struct bridge_softc *, int); 329 static int bridge_rtdaddr(struct bridge_softc *, const uint8_t *, 330 uint16_t); 331 332 static void bridge_rtable_init(struct bridge_softc *); 333 static void bridge_rtable_fini(struct bridge_softc *); 334 335 static int bridge_rtnode_addr_cmp(const uint8_t *, const uint8_t *); 336 static struct bridge_rtnode *bridge_rtnode_lookup(struct bridge_softc *, 337 const uint8_t *, uint16_t); 338 static int bridge_rtnode_insert(struct bridge_softc *, 339 struct bridge_rtnode *); 340 static void bridge_rtnode_destroy(struct bridge_softc *, 341 struct bridge_rtnode *); 342 static void bridge_rtable_expire(struct ifnet *, int); 343 static void bridge_state_change(struct ifnet *, int); 344 345 static struct bridge_iflist *bridge_lookup_member(struct bridge_softc *, 346 const char *name); 347 static struct bridge_iflist *bridge_lookup_member_if(struct bridge_softc *, 348 struct ifnet *ifp); 349 static void bridge_delete_member(struct bridge_softc *, 350 struct bridge_iflist *, int); 351 static void bridge_delete_span(struct bridge_softc *, 352 struct bridge_iflist *); 353 354 static int bridge_ioctl_add(struct bridge_softc *, void *); 355 static int bridge_ioctl_del(struct bridge_softc *, void *); 356 static int bridge_ioctl_gifflags(struct bridge_softc *, void *); 357 static int bridge_ioctl_sifflags(struct bridge_softc *, void *); 358 static int bridge_ioctl_scache(struct bridge_softc *, void *); 359 static int bridge_ioctl_gcache(struct bridge_softc *, void *); 360 static int bridge_ioctl_gifs(struct bridge_softc *, void *); 361 static int bridge_ioctl_rts(struct bridge_softc *, void *); 362 static int bridge_ioctl_saddr(struct bridge_softc *, void *); 363 static int bridge_ioctl_sto(struct bridge_softc *, void *); 364 static int bridge_ioctl_gto(struct bridge_softc *, void *); 365 static int bridge_ioctl_daddr(struct bridge_softc *, void *); 366 static int bridge_ioctl_flush(struct bridge_softc *, void *); 367 static int bridge_ioctl_gpri(struct bridge_softc *, void *); 368 static int bridge_ioctl_spri(struct bridge_softc *, void *); 369 static int bridge_ioctl_ght(struct bridge_softc *, void *); 370 static int bridge_ioctl_sht(struct bridge_softc *, void *); 371 static int bridge_ioctl_gfd(struct bridge_softc *, void *); 372 static int bridge_ioctl_sfd(struct bridge_softc *, void *); 373 static int bridge_ioctl_gma(struct bridge_softc *, void *); 374 static int bridge_ioctl_sma(struct bridge_softc *, void *); 375 static int bridge_ioctl_sifprio(struct bridge_softc *, void *); 376 static int bridge_ioctl_sifcost(struct bridge_softc *, void *); 377 static int bridge_ioctl_sifmaxaddr(struct bridge_softc *, void *); 378 static int bridge_ioctl_addspan(struct bridge_softc *, void *); 379 static int bridge_ioctl_delspan(struct bridge_softc *, void *); 380 static int bridge_ioctl_gbparam(struct bridge_softc *, void *); 381 static int bridge_ioctl_grte(struct bridge_softc *, void *); 382 static int bridge_ioctl_gifsstp(struct bridge_softc *, void *); 383 static int bridge_ioctl_sproto(struct bridge_softc *, void *); 384 static int bridge_ioctl_stxhc(struct bridge_softc *, void *); 385 static int bridge_pfil(struct mbuf **, struct ifnet *, struct ifnet *, 386 int); 387 static int bridge_ip_checkbasic(struct mbuf **mp); 388 #ifdef INET6 389 static int bridge_ip6_checkbasic(struct mbuf **mp); 390 #endif /* INET6 */ 391 static int bridge_fragment(struct ifnet *, struct mbuf **mp, 392 struct ether_header *, int, struct llc *); 393 static void bridge_linkstate(struct ifnet *ifp); 394 static void bridge_linkcheck(struct bridge_softc *sc); 395 396 /* The default bridge vlan is 1 (IEEE 802.1Q-2003 Table 9-2) */ 397 #define VLANTAGOF(_m) \ 398 (_m->m_flags & M_VLANTAG) ? EVL_VLANOFTAG(_m->m_pkthdr.ether_vtag) : 1 399 400 static struct bstp_cb_ops bridge_ops = { 401 .bcb_state = bridge_state_change, 402 .bcb_rtage = bridge_rtable_expire 403 }; 404 405 SYSCTL_DECL(_net_link); 406 static SYSCTL_NODE(_net_link, IFT_BRIDGE, bridge, CTLFLAG_RW | CTLFLAG_MPSAFE, 0, 407 "Bridge"); 408 409 /* only pass IP[46] packets when pfil is enabled */ 410 VNET_DEFINE_STATIC(int, pfil_onlyip) = 1; 411 #define V_pfil_onlyip VNET(pfil_onlyip) 412 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_onlyip, 413 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_onlyip), 0, 414 "Only pass IP packets when pfil is enabled"); 415 416 /* run pfil hooks on the bridge interface */ 417 VNET_DEFINE_STATIC(int, pfil_bridge) = 1; 418 #define V_pfil_bridge VNET(pfil_bridge) 419 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_bridge, 420 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_bridge), 0, 421 "Packet filter on the bridge interface"); 422 423 /* layer2 filter with ipfw */ 424 VNET_DEFINE_STATIC(int, pfil_ipfw); 425 #define V_pfil_ipfw VNET(pfil_ipfw) 426 427 /* layer2 ARP filter with ipfw */ 428 VNET_DEFINE_STATIC(int, pfil_ipfw_arp); 429 #define V_pfil_ipfw_arp VNET(pfil_ipfw_arp) 430 SYSCTL_INT(_net_link_bridge, OID_AUTO, ipfw_arp, 431 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_ipfw_arp), 0, 432 "Filter ARP packets through IPFW layer2"); 433 434 /* run pfil hooks on the member interface */ 435 VNET_DEFINE_STATIC(int, pfil_member) = 1; 436 #define V_pfil_member VNET(pfil_member) 437 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_member, 438 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_member), 0, 439 "Packet filter on the member interface"); 440 441 /* run pfil hooks on the physical interface for locally destined packets */ 442 VNET_DEFINE_STATIC(int, pfil_local_phys); 443 #define V_pfil_local_phys VNET(pfil_local_phys) 444 SYSCTL_INT(_net_link_bridge, OID_AUTO, pfil_local_phys, 445 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(pfil_local_phys), 0, 446 "Packet filter on the physical interface for locally destined packets"); 447 448 /* log STP state changes */ 449 VNET_DEFINE_STATIC(int, log_stp); 450 #define V_log_stp VNET(log_stp) 451 SYSCTL_INT(_net_link_bridge, OID_AUTO, log_stp, 452 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(log_stp), 0, 453 "Log STP state changes"); 454 455 /* share MAC with first bridge member */ 456 VNET_DEFINE_STATIC(int, bridge_inherit_mac); 457 #define V_bridge_inherit_mac VNET(bridge_inherit_mac) 458 SYSCTL_INT(_net_link_bridge, OID_AUTO, inherit_mac, 459 CTLFLAG_RWTUN | CTLFLAG_VNET, &VNET_NAME(bridge_inherit_mac), 0, 460 "Inherit MAC address from the first bridge member"); 461 462 VNET_DEFINE_STATIC(int, allow_llz_overlap) = 0; 463 #define V_allow_llz_overlap VNET(allow_llz_overlap) 464 SYSCTL_INT(_net_link_bridge, OID_AUTO, allow_llz_overlap, 465 CTLFLAG_RW | CTLFLAG_VNET, &VNET_NAME(allow_llz_overlap), 0, 466 "Allow overlap of link-local scope " 467 "zones of a bridge interface and the member interfaces"); 468 469 struct bridge_control { 470 int (*bc_func)(struct bridge_softc *, void *); 471 int bc_argsize; 472 int bc_flags; 473 }; 474 475 #define BC_F_COPYIN 0x01 /* copy arguments in */ 476 #define BC_F_COPYOUT 0x02 /* copy arguments out */ 477 #define BC_F_SUSER 0x04 /* do super-user check */ 478 479 const struct bridge_control bridge_control_table[] = { 480 { bridge_ioctl_add, sizeof(struct ifbreq), 481 BC_F_COPYIN|BC_F_SUSER }, 482 { bridge_ioctl_del, sizeof(struct ifbreq), 483 BC_F_COPYIN|BC_F_SUSER }, 484 485 { bridge_ioctl_gifflags, sizeof(struct ifbreq), 486 BC_F_COPYIN|BC_F_COPYOUT }, 487 { bridge_ioctl_sifflags, sizeof(struct ifbreq), 488 BC_F_COPYIN|BC_F_SUSER }, 489 490 { bridge_ioctl_scache, sizeof(struct ifbrparam), 491 BC_F_COPYIN|BC_F_SUSER }, 492 { bridge_ioctl_gcache, sizeof(struct ifbrparam), 493 BC_F_COPYOUT }, 494 495 { bridge_ioctl_gifs, sizeof(struct ifbifconf), 496 BC_F_COPYIN|BC_F_COPYOUT }, 497 { bridge_ioctl_rts, sizeof(struct ifbaconf), 498 BC_F_COPYIN|BC_F_COPYOUT }, 499 500 { bridge_ioctl_saddr, sizeof(struct ifbareq), 501 BC_F_COPYIN|BC_F_SUSER }, 502 503 { bridge_ioctl_sto, sizeof(struct ifbrparam), 504 BC_F_COPYIN|BC_F_SUSER }, 505 { bridge_ioctl_gto, sizeof(struct ifbrparam), 506 BC_F_COPYOUT }, 507 508 { bridge_ioctl_daddr, sizeof(struct ifbareq), 509 BC_F_COPYIN|BC_F_SUSER }, 510 511 { bridge_ioctl_flush, sizeof(struct ifbreq), 512 BC_F_COPYIN|BC_F_SUSER }, 513 514 { bridge_ioctl_gpri, sizeof(struct ifbrparam), 515 BC_F_COPYOUT }, 516 { bridge_ioctl_spri, sizeof(struct ifbrparam), 517 BC_F_COPYIN|BC_F_SUSER }, 518 519 { bridge_ioctl_ght, sizeof(struct ifbrparam), 520 BC_F_COPYOUT }, 521 { bridge_ioctl_sht, sizeof(struct ifbrparam), 522 BC_F_COPYIN|BC_F_SUSER }, 523 524 { bridge_ioctl_gfd, sizeof(struct ifbrparam), 525 BC_F_COPYOUT }, 526 { bridge_ioctl_sfd, sizeof(struct ifbrparam), 527 BC_F_COPYIN|BC_F_SUSER }, 528 529 { bridge_ioctl_gma, sizeof(struct ifbrparam), 530 BC_F_COPYOUT }, 531 { bridge_ioctl_sma, sizeof(struct ifbrparam), 532 BC_F_COPYIN|BC_F_SUSER }, 533 534 { bridge_ioctl_sifprio, sizeof(struct ifbreq), 535 BC_F_COPYIN|BC_F_SUSER }, 536 537 { bridge_ioctl_sifcost, sizeof(struct ifbreq), 538 BC_F_COPYIN|BC_F_SUSER }, 539 540 { bridge_ioctl_addspan, sizeof(struct ifbreq), 541 BC_F_COPYIN|BC_F_SUSER }, 542 { bridge_ioctl_delspan, sizeof(struct ifbreq), 543 BC_F_COPYIN|BC_F_SUSER }, 544 545 { bridge_ioctl_gbparam, sizeof(struct ifbropreq), 546 BC_F_COPYOUT }, 547 548 { bridge_ioctl_grte, sizeof(struct ifbrparam), 549 BC_F_COPYOUT }, 550 551 { bridge_ioctl_gifsstp, sizeof(struct ifbpstpconf), 552 BC_F_COPYIN|BC_F_COPYOUT }, 553 554 { bridge_ioctl_sproto, sizeof(struct ifbrparam), 555 BC_F_COPYIN|BC_F_SUSER }, 556 557 { bridge_ioctl_stxhc, sizeof(struct ifbrparam), 558 BC_F_COPYIN|BC_F_SUSER }, 559 560 { bridge_ioctl_sifmaxaddr, sizeof(struct ifbreq), 561 BC_F_COPYIN|BC_F_SUSER }, 562 563 }; 564 const int bridge_control_table_size = nitems(bridge_control_table); 565 566 VNET_DEFINE_STATIC(LIST_HEAD(, bridge_softc), bridge_list); 567 #define V_bridge_list VNET(bridge_list) 568 #define BRIDGE_LIST_LOCK_INIT(x) sx_init(&V_bridge_list_sx, \ 569 "if_bridge list") 570 #define BRIDGE_LIST_LOCK_DESTROY(x) sx_destroy(&V_bridge_list_sx) 571 #define BRIDGE_LIST_LOCK(x) sx_xlock(&V_bridge_list_sx) 572 #define BRIDGE_LIST_UNLOCK(x) sx_xunlock(&V_bridge_list_sx) 573 574 VNET_DEFINE_STATIC(struct if_clone *, bridge_cloner); 575 #define V_bridge_cloner VNET(bridge_cloner) 576 577 static const char bridge_name[] = "bridge"; 578 579 static void 580 vnet_bridge_init(const void *unused __unused) 581 { 582 583 V_bridge_rtnode_zone = uma_zcreate("bridge_rtnode", 584 sizeof(struct bridge_rtnode), NULL, NULL, NULL, NULL, 585 UMA_ALIGN_PTR, 0); 586 BRIDGE_LIST_LOCK_INIT(); 587 LIST_INIT(&V_bridge_list); 588 V_bridge_cloner = if_clone_simple(bridge_name, 589 bridge_clone_create, bridge_clone_destroy, 0); 590 } 591 VNET_SYSINIT(vnet_bridge_init, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY, 592 vnet_bridge_init, NULL); 593 594 static void 595 vnet_bridge_uninit(const void *unused __unused) 596 { 597 598 if_clone_detach(V_bridge_cloner); 599 V_bridge_cloner = NULL; 600 BRIDGE_LIST_LOCK_DESTROY(); 601 602 /* Callbacks may use the UMA zone. */ 603 epoch_drain_callbacks(net_epoch_preempt); 604 605 uma_zdestroy(V_bridge_rtnode_zone); 606 } 607 VNET_SYSUNINIT(vnet_bridge_uninit, SI_SUB_PSEUDO, SI_ORDER_ANY, 608 vnet_bridge_uninit, NULL); 609 610 static int 611 bridge_modevent(module_t mod, int type, void *data) 612 { 613 614 switch (type) { 615 case MOD_LOAD: 616 bridge_dn_p = bridge_dummynet; 617 bridge_detach_cookie = EVENTHANDLER_REGISTER( 618 ifnet_departure_event, bridge_ifdetach, NULL, 619 EVENTHANDLER_PRI_ANY); 620 break; 621 case MOD_UNLOAD: 622 EVENTHANDLER_DEREGISTER(ifnet_departure_event, 623 bridge_detach_cookie); 624 bridge_dn_p = NULL; 625 break; 626 default: 627 return (EOPNOTSUPP); 628 } 629 return (0); 630 } 631 632 static moduledata_t bridge_mod = { 633 "if_bridge", 634 bridge_modevent, 635 0 636 }; 637 638 DECLARE_MODULE(if_bridge, bridge_mod, SI_SUB_PSEUDO, SI_ORDER_ANY); 639 MODULE_VERSION(if_bridge, 1); 640 MODULE_DEPEND(if_bridge, bridgestp, 1, 1, 1); 641 642 /* 643 * handler for net.link.bridge.ipfw 644 */ 645 static int 646 sysctl_pfil_ipfw(SYSCTL_HANDLER_ARGS) 647 { 648 int enable = V_pfil_ipfw; 649 int error; 650 651 error = sysctl_handle_int(oidp, &enable, 0, req); 652 enable &= 1; 653 654 if (enable != V_pfil_ipfw) { 655 V_pfil_ipfw = enable; 656 657 /* 658 * Disable pfil so that ipfw doesnt run twice, if the user 659 * really wants both then they can re-enable pfil_bridge and/or 660 * pfil_member. Also allow non-ip packets as ipfw can filter by 661 * layer2 type. 662 */ 663 if (V_pfil_ipfw) { 664 V_pfil_onlyip = 0; 665 V_pfil_bridge = 0; 666 V_pfil_member = 0; 667 } 668 } 669 670 return (error); 671 } 672 SYSCTL_PROC(_net_link_bridge, OID_AUTO, ipfw, 673 CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_VNET | CTLFLAG_NEEDGIANT, 674 &VNET_NAME(pfil_ipfw), 0, &sysctl_pfil_ipfw, "I", 675 "Layer2 filter with IPFW"); 676 677 #ifdef VIMAGE 678 static void 679 bridge_reassign(struct ifnet *ifp, struct vnet *newvnet, char *arg) 680 { 681 struct bridge_softc *sc = ifp->if_softc; 682 struct bridge_iflist *bif; 683 684 BRIDGE_LOCK(sc); 685 686 while ((bif = CK_LIST_FIRST(&sc->sc_iflist)) != NULL) 687 bridge_delete_member(sc, bif, 0); 688 689 while ((bif = CK_LIST_FIRST(&sc->sc_spanlist)) != NULL) { 690 bridge_delete_span(sc, bif); 691 } 692 693 BRIDGE_UNLOCK(sc); 694 695 ether_reassign(ifp, newvnet, arg); 696 } 697 #endif 698 699 /* 700 * bridge_clone_create: 701 * 702 * Create a new bridge instance. 703 */ 704 static int 705 bridge_clone_create(struct if_clone *ifc, int unit, caddr_t params) 706 { 707 struct bridge_softc *sc; 708 struct ifnet *ifp; 709 710 sc = malloc(sizeof(*sc), M_DEVBUF, M_WAITOK|M_ZERO); 711 ifp = sc->sc_ifp = if_alloc(IFT_ETHER); 712 if (ifp == NULL) { 713 free(sc, M_DEVBUF); 714 return (ENOSPC); 715 } 716 717 BRIDGE_LOCK_INIT(sc); 718 sc->sc_brtmax = BRIDGE_RTABLE_MAX; 719 sc->sc_brttimeout = BRIDGE_RTABLE_TIMEOUT; 720 721 /* Initialize our routing table. */ 722 bridge_rtable_init(sc); 723 724 callout_init_mtx(&sc->sc_brcallout, &sc->sc_rt_mtx, 0); 725 726 CK_LIST_INIT(&sc->sc_iflist); 727 CK_LIST_INIT(&sc->sc_spanlist); 728 729 ifp->if_softc = sc; 730 if_initname(ifp, bridge_name, unit); 731 ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST; 732 ifp->if_ioctl = bridge_ioctl; 733 #ifdef ALTQ 734 ifp->if_start = bridge_altq_start; 735 ifp->if_transmit = bridge_altq_transmit; 736 IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen); 737 ifp->if_snd.ifq_drv_maxlen = 0; 738 IFQ_SET_READY(&ifp->if_snd); 739 #else 740 ifp->if_transmit = bridge_transmit; 741 #endif 742 ifp->if_qflush = bridge_qflush; 743 ifp->if_init = bridge_init; 744 ifp->if_type = IFT_BRIDGE; 745 746 ether_gen_addr(ifp, &sc->sc_defaddr); 747 748 bstp_attach(&sc->sc_stp, &bridge_ops); 749 ether_ifattach(ifp, sc->sc_defaddr.octet); 750 /* Now undo some of the damage... */ 751 ifp->if_baudrate = 0; 752 ifp->if_type = IFT_BRIDGE; 753 #ifdef VIMAGE 754 ifp->if_reassign = bridge_reassign; 755 #endif 756 757 BRIDGE_LIST_LOCK(); 758 LIST_INSERT_HEAD(&V_bridge_list, sc, sc_list); 759 BRIDGE_LIST_UNLOCK(); 760 761 return (0); 762 } 763 764 static void 765 bridge_clone_destroy_cb(struct epoch_context *ctx) 766 { 767 struct bridge_softc *sc; 768 769 sc = __containerof(ctx, struct bridge_softc, sc_epoch_ctx); 770 771 BRIDGE_LOCK_DESTROY(sc); 772 free(sc, M_DEVBUF); 773 } 774 775 /* 776 * bridge_clone_destroy: 777 * 778 * Destroy a bridge instance. 779 */ 780 static void 781 bridge_clone_destroy(struct ifnet *ifp) 782 { 783 struct bridge_softc *sc = ifp->if_softc; 784 struct bridge_iflist *bif; 785 struct epoch_tracker et; 786 787 BRIDGE_LOCK(sc); 788 789 bridge_stop(ifp, 1); 790 ifp->if_flags &= ~IFF_UP; 791 792 while ((bif = CK_LIST_FIRST(&sc->sc_iflist)) != NULL) 793 bridge_delete_member(sc, bif, 0); 794 795 while ((bif = CK_LIST_FIRST(&sc->sc_spanlist)) != NULL) { 796 bridge_delete_span(sc, bif); 797 } 798 799 /* Tear down the routing table. */ 800 bridge_rtable_fini(sc); 801 802 BRIDGE_UNLOCK(sc); 803 804 NET_EPOCH_ENTER(et); 805 806 callout_drain(&sc->sc_brcallout); 807 808 BRIDGE_LIST_LOCK(); 809 LIST_REMOVE(sc, sc_list); 810 BRIDGE_LIST_UNLOCK(); 811 812 bstp_detach(&sc->sc_stp); 813 #ifdef ALTQ 814 IFQ_PURGE(&ifp->if_snd); 815 #endif 816 NET_EPOCH_EXIT(et); 817 818 ether_ifdetach(ifp); 819 if_free(ifp); 820 821 NET_EPOCH_CALL(bridge_clone_destroy_cb, &sc->sc_epoch_ctx); 822 } 823 824 /* 825 * bridge_ioctl: 826 * 827 * Handle a control request from the operator. 828 */ 829 static int 830 bridge_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data) 831 { 832 struct bridge_softc *sc = ifp->if_softc; 833 struct ifreq *ifr = (struct ifreq *)data; 834 struct bridge_iflist *bif; 835 struct thread *td = curthread; 836 union { 837 struct ifbreq ifbreq; 838 struct ifbifconf ifbifconf; 839 struct ifbareq ifbareq; 840 struct ifbaconf ifbaconf; 841 struct ifbrparam ifbrparam; 842 struct ifbropreq ifbropreq; 843 } args; 844 struct ifdrv *ifd = (struct ifdrv *) data; 845 const struct bridge_control *bc; 846 int error = 0, oldmtu; 847 848 BRIDGE_LOCK(sc); 849 850 switch (cmd) { 851 case SIOCADDMULTI: 852 case SIOCDELMULTI: 853 break; 854 855 case SIOCGDRVSPEC: 856 case SIOCSDRVSPEC: 857 if (ifd->ifd_cmd >= bridge_control_table_size) { 858 error = EINVAL; 859 break; 860 } 861 bc = &bridge_control_table[ifd->ifd_cmd]; 862 863 if (cmd == SIOCGDRVSPEC && 864 (bc->bc_flags & BC_F_COPYOUT) == 0) { 865 error = EINVAL; 866 break; 867 } 868 else if (cmd == SIOCSDRVSPEC && 869 (bc->bc_flags & BC_F_COPYOUT) != 0) { 870 error = EINVAL; 871 break; 872 } 873 874 if (bc->bc_flags & BC_F_SUSER) { 875 error = priv_check(td, PRIV_NET_BRIDGE); 876 if (error) 877 break; 878 } 879 880 if (ifd->ifd_len != bc->bc_argsize || 881 ifd->ifd_len > sizeof(args)) { 882 error = EINVAL; 883 break; 884 } 885 886 bzero(&args, sizeof(args)); 887 if (bc->bc_flags & BC_F_COPYIN) { 888 error = copyin(ifd->ifd_data, &args, ifd->ifd_len); 889 if (error) 890 break; 891 } 892 893 oldmtu = ifp->if_mtu; 894 error = (*bc->bc_func)(sc, &args); 895 if (error) 896 break; 897 898 /* 899 * Bridge MTU may change during addition of the first port. 900 * If it did, do network layer specific procedure. 901 */ 902 if (ifp->if_mtu != oldmtu) { 903 #ifdef INET6 904 nd6_setmtu(ifp); 905 #endif 906 rt_updatemtu(ifp); 907 } 908 909 if (bc->bc_flags & BC_F_COPYOUT) 910 error = copyout(&args, ifd->ifd_data, ifd->ifd_len); 911 912 break; 913 914 case SIOCSIFFLAGS: 915 if (!(ifp->if_flags & IFF_UP) && 916 (ifp->if_drv_flags & IFF_DRV_RUNNING)) { 917 /* 918 * If interface is marked down and it is running, 919 * then stop and disable it. 920 */ 921 bridge_stop(ifp, 1); 922 } else if ((ifp->if_flags & IFF_UP) && 923 !(ifp->if_drv_flags & IFF_DRV_RUNNING)) { 924 /* 925 * If interface is marked up and it is stopped, then 926 * start it. 927 */ 928 BRIDGE_UNLOCK(sc); 929 (*ifp->if_init)(sc); 930 BRIDGE_LOCK(sc); 931 } 932 break; 933 934 case SIOCSIFMTU: 935 oldmtu = sc->sc_ifp->if_mtu; 936 937 if (ifr->ifr_mtu < 576) { 938 error = EINVAL; 939 break; 940 } 941 if (CK_LIST_EMPTY(&sc->sc_iflist)) { 942 sc->sc_ifp->if_mtu = ifr->ifr_mtu; 943 break; 944 } 945 CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 946 error = (*bif->bif_ifp->if_ioctl)(bif->bif_ifp, 947 SIOCSIFMTU, (caddr_t)ifr); 948 if (error != 0) { 949 log(LOG_NOTICE, "%s: invalid MTU: %u for" 950 " member %s\n", sc->sc_ifp->if_xname, 951 ifr->ifr_mtu, 952 bif->bif_ifp->if_xname); 953 error = EINVAL; 954 break; 955 } 956 } 957 if (error) { 958 /* Restore the previous MTU on all member interfaces. */ 959 ifr->ifr_mtu = oldmtu; 960 CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 961 (*bif->bif_ifp->if_ioctl)(bif->bif_ifp, 962 SIOCSIFMTU, (caddr_t)ifr); 963 } 964 } else { 965 sc->sc_ifp->if_mtu = ifr->ifr_mtu; 966 } 967 break; 968 default: 969 /* 970 * drop the lock as ether_ioctl() will call bridge_start() and 971 * cause the lock to be recursed. 972 */ 973 BRIDGE_UNLOCK(sc); 974 error = ether_ioctl(ifp, cmd, data); 975 BRIDGE_LOCK(sc); 976 break; 977 } 978 979 BRIDGE_UNLOCK(sc); 980 981 return (error); 982 } 983 984 /* 985 * bridge_mutecaps: 986 * 987 * Clear or restore unwanted capabilities on the member interface 988 */ 989 static void 990 bridge_mutecaps(struct bridge_softc *sc) 991 { 992 struct bridge_iflist *bif; 993 int enabled, mask; 994 995 BRIDGE_LOCK_ASSERT(sc); 996 997 /* Initial bitmask of capabilities to test */ 998 mask = BRIDGE_IFCAPS_MASK; 999 1000 CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 1001 /* Every member must support it or its disabled */ 1002 mask &= bif->bif_savedcaps; 1003 } 1004 1005 CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 1006 enabled = bif->bif_ifp->if_capenable; 1007 enabled &= ~BRIDGE_IFCAPS_STRIP; 1008 /* strip off mask bits and enable them again if allowed */ 1009 enabled &= ~BRIDGE_IFCAPS_MASK; 1010 enabled |= mask; 1011 bridge_set_ifcap(sc, bif, enabled); 1012 } 1013 } 1014 1015 static void 1016 bridge_set_ifcap(struct bridge_softc *sc, struct bridge_iflist *bif, int set) 1017 { 1018 struct ifnet *ifp = bif->bif_ifp; 1019 struct ifreq ifr; 1020 int error, mask, stuck; 1021 1022 bzero(&ifr, sizeof(ifr)); 1023 ifr.ifr_reqcap = set; 1024 1025 if (ifp->if_capenable != set) { 1026 error = (*ifp->if_ioctl)(ifp, SIOCSIFCAP, (caddr_t)&ifr); 1027 if (error) 1028 if_printf(sc->sc_ifp, 1029 "error setting capabilities on %s: %d\n", 1030 ifp->if_xname, error); 1031 mask = BRIDGE_IFCAPS_MASK | BRIDGE_IFCAPS_STRIP; 1032 stuck = ifp->if_capenable & mask & ~set; 1033 if (stuck != 0) 1034 if_printf(sc->sc_ifp, 1035 "can't disable some capabilities on %s: 0x%x\n", 1036 ifp->if_xname, stuck); 1037 } 1038 } 1039 1040 /* 1041 * bridge_lookup_member: 1042 * 1043 * Lookup a bridge member interface. 1044 */ 1045 static struct bridge_iflist * 1046 bridge_lookup_member(struct bridge_softc *sc, const char *name) 1047 { 1048 struct bridge_iflist *bif; 1049 struct ifnet *ifp; 1050 1051 BRIDGE_LOCK_OR_NET_EPOCH_ASSERT(sc); 1052 1053 CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 1054 ifp = bif->bif_ifp; 1055 if (strcmp(ifp->if_xname, name) == 0) 1056 return (bif); 1057 } 1058 1059 return (NULL); 1060 } 1061 1062 /* 1063 * bridge_lookup_member_if: 1064 * 1065 * Lookup a bridge member interface by ifnet*. 1066 */ 1067 static struct bridge_iflist * 1068 bridge_lookup_member_if(struct bridge_softc *sc, struct ifnet *member_ifp) 1069 { 1070 struct bridge_iflist *bif; 1071 1072 BRIDGE_LOCK_OR_NET_EPOCH_ASSERT(sc); 1073 1074 CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 1075 if (bif->bif_ifp == member_ifp) 1076 return (bif); 1077 } 1078 1079 return (NULL); 1080 } 1081 1082 static void 1083 bridge_delete_member_cb(struct epoch_context *ctx) 1084 { 1085 struct bridge_iflist *bif; 1086 1087 bif = __containerof(ctx, struct bridge_iflist, bif_epoch_ctx); 1088 1089 free(bif, M_DEVBUF); 1090 } 1091 1092 /* 1093 * bridge_delete_member: 1094 * 1095 * Delete the specified member interface. 1096 */ 1097 static void 1098 bridge_delete_member(struct bridge_softc *sc, struct bridge_iflist *bif, 1099 int gone) 1100 { 1101 struct ifnet *ifs = bif->bif_ifp; 1102 struct ifnet *fif = NULL; 1103 struct bridge_iflist *bifl; 1104 1105 BRIDGE_LOCK_ASSERT(sc); 1106 1107 if (bif->bif_flags & IFBIF_STP) 1108 bstp_disable(&bif->bif_stp); 1109 1110 ifs->if_bridge = NULL; 1111 CK_LIST_REMOVE(bif, bif_next); 1112 1113 /* 1114 * If removing the interface that gave the bridge its mac address, set 1115 * the mac address of the bridge to the address of the next member, or 1116 * to its default address if no members are left. 1117 */ 1118 if (V_bridge_inherit_mac && sc->sc_ifaddr == ifs) { 1119 if (CK_LIST_EMPTY(&sc->sc_iflist)) { 1120 bcopy(&sc->sc_defaddr, 1121 IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN); 1122 sc->sc_ifaddr = NULL; 1123 } else { 1124 bifl = CK_LIST_FIRST(&sc->sc_iflist); 1125 fif = bifl->bif_ifp; 1126 bcopy(IF_LLADDR(fif), 1127 IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN); 1128 sc->sc_ifaddr = fif; 1129 } 1130 EVENTHANDLER_INVOKE(iflladdr_event, sc->sc_ifp); 1131 } 1132 1133 bridge_linkcheck(sc); 1134 bridge_mutecaps(sc); /* recalcuate now this interface is removed */ 1135 BRIDGE_RT_LOCK(sc); 1136 bridge_rtdelete(sc, ifs, IFBF_FLUSHALL); 1137 BRIDGE_RT_UNLOCK(sc); 1138 KASSERT(bif->bif_addrcnt == 0, 1139 ("%s: %d bridge routes referenced", __func__, bif->bif_addrcnt)); 1140 1141 ifs->if_bridge_output = NULL; 1142 ifs->if_bridge_input = NULL; 1143 ifs->if_bridge_linkstate = NULL; 1144 if (!gone) { 1145 switch (ifs->if_type) { 1146 case IFT_ETHER: 1147 case IFT_L2VLAN: 1148 /* 1149 * Take the interface out of promiscuous mode, but only 1150 * if it was promiscuous in the first place. It might 1151 * not be if we're in the bridge_ioctl_add() error path. 1152 */ 1153 if (ifs->if_flags & IFF_PROMISC) 1154 (void) ifpromisc(ifs, 0); 1155 break; 1156 1157 case IFT_GIF: 1158 break; 1159 1160 default: 1161 #ifdef DIAGNOSTIC 1162 panic("bridge_delete_member: impossible"); 1163 #endif 1164 break; 1165 } 1166 /* reneable any interface capabilities */ 1167 bridge_set_ifcap(sc, bif, bif->bif_savedcaps); 1168 } 1169 bstp_destroy(&bif->bif_stp); /* prepare to free */ 1170 1171 NET_EPOCH_CALL(bridge_delete_member_cb, &bif->bif_epoch_ctx); 1172 } 1173 1174 /* 1175 * bridge_delete_span: 1176 * 1177 * Delete the specified span interface. 1178 */ 1179 static void 1180 bridge_delete_span(struct bridge_softc *sc, struct bridge_iflist *bif) 1181 { 1182 BRIDGE_LOCK_ASSERT(sc); 1183 1184 KASSERT(bif->bif_ifp->if_bridge == NULL, 1185 ("%s: not a span interface", __func__)); 1186 1187 CK_LIST_REMOVE(bif, bif_next); 1188 1189 NET_EPOCH_CALL(bridge_delete_member_cb, &bif->bif_epoch_ctx); 1190 } 1191 1192 static int 1193 bridge_ioctl_add(struct bridge_softc *sc, void *arg) 1194 { 1195 struct ifbreq *req = arg; 1196 struct bridge_iflist *bif = NULL; 1197 struct ifnet *ifs; 1198 int error = 0; 1199 1200 ifs = ifunit(req->ifbr_ifsname); 1201 if (ifs == NULL) 1202 return (ENOENT); 1203 if (ifs->if_ioctl == NULL) /* must be supported */ 1204 return (EINVAL); 1205 1206 /* If it's in the span list, it can't be a member. */ 1207 CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) 1208 if (ifs == bif->bif_ifp) 1209 return (EBUSY); 1210 1211 if (ifs->if_bridge == sc) 1212 return (EEXIST); 1213 1214 if (ifs->if_bridge != NULL) 1215 return (EBUSY); 1216 1217 switch (ifs->if_type) { 1218 case IFT_ETHER: 1219 case IFT_L2VLAN: 1220 case IFT_GIF: 1221 /* permitted interface types */ 1222 break; 1223 default: 1224 return (EINVAL); 1225 } 1226 1227 #ifdef INET6 1228 /* 1229 * Two valid inet6 addresses with link-local scope must not be 1230 * on the parent interface and the member interfaces at the 1231 * same time. This restriction is needed to prevent violation 1232 * of link-local scope zone. Attempts to add a member 1233 * interface which has inet6 addresses when the parent has 1234 * inet6 triggers removal of all inet6 addresses on the member 1235 * interface. 1236 */ 1237 1238 /* Check if the parent interface has a link-local scope addr. */ 1239 if (V_allow_llz_overlap == 0 && 1240 in6ifa_llaonifp(sc->sc_ifp) != NULL) { 1241 /* 1242 * If any, remove all inet6 addresses from the member 1243 * interfaces. 1244 */ 1245 CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 1246 if (in6ifa_llaonifp(bif->bif_ifp)) { 1247 in6_ifdetach(bif->bif_ifp); 1248 if_printf(sc->sc_ifp, 1249 "IPv6 addresses on %s have been removed " 1250 "before adding it as a member to prevent " 1251 "IPv6 address scope violation.\n", 1252 bif->bif_ifp->if_xname); 1253 } 1254 } 1255 if (in6ifa_llaonifp(ifs)) { 1256 in6_ifdetach(ifs); 1257 if_printf(sc->sc_ifp, 1258 "IPv6 addresses on %s have been removed " 1259 "before adding it as a member to prevent " 1260 "IPv6 address scope violation.\n", 1261 ifs->if_xname); 1262 } 1263 } 1264 #endif 1265 /* Allow the first Ethernet member to define the MTU */ 1266 if (CK_LIST_EMPTY(&sc->sc_iflist)) 1267 sc->sc_ifp->if_mtu = ifs->if_mtu; 1268 else if (sc->sc_ifp->if_mtu != ifs->if_mtu) { 1269 if_printf(sc->sc_ifp, "invalid MTU: %u(%s) != %u\n", 1270 ifs->if_mtu, ifs->if_xname, sc->sc_ifp->if_mtu); 1271 return (EINVAL); 1272 } 1273 1274 bif = malloc(sizeof(*bif), M_DEVBUF, M_NOWAIT|M_ZERO); 1275 if (bif == NULL) 1276 return (ENOMEM); 1277 1278 bif->bif_ifp = ifs; 1279 bif->bif_flags = IFBIF_LEARNING | IFBIF_DISCOVER; 1280 bif->bif_savedcaps = ifs->if_capenable; 1281 1282 /* 1283 * Assign the interface's MAC address to the bridge if it's the first 1284 * member and the MAC address of the bridge has not been changed from 1285 * the default randomly generated one. 1286 */ 1287 if (V_bridge_inherit_mac && CK_LIST_EMPTY(&sc->sc_iflist) && 1288 !memcmp(IF_LLADDR(sc->sc_ifp), sc->sc_defaddr.octet, ETHER_ADDR_LEN)) { 1289 bcopy(IF_LLADDR(ifs), IF_LLADDR(sc->sc_ifp), ETHER_ADDR_LEN); 1290 sc->sc_ifaddr = ifs; 1291 EVENTHANDLER_INVOKE(iflladdr_event, sc->sc_ifp); 1292 } 1293 1294 ifs->if_bridge = sc; 1295 ifs->if_bridge_output = bridge_output; 1296 ifs->if_bridge_input = bridge_input; 1297 ifs->if_bridge_linkstate = bridge_linkstate; 1298 bstp_create(&sc->sc_stp, &bif->bif_stp, bif->bif_ifp); 1299 /* 1300 * XXX: XLOCK HERE!?! 1301 * 1302 * NOTE: insert_***HEAD*** should be safe for the traversals. 1303 */ 1304 CK_LIST_INSERT_HEAD(&sc->sc_iflist, bif, bif_next); 1305 1306 /* Set interface capabilities to the intersection set of all members */ 1307 bridge_mutecaps(sc); 1308 bridge_linkcheck(sc); 1309 1310 /* Place the interface into promiscuous mode */ 1311 switch (ifs->if_type) { 1312 case IFT_ETHER: 1313 case IFT_L2VLAN: 1314 error = ifpromisc(ifs, 1); 1315 break; 1316 } 1317 1318 if (error) 1319 bridge_delete_member(sc, bif, 0); 1320 return (error); 1321 } 1322 1323 static int 1324 bridge_ioctl_del(struct bridge_softc *sc, void *arg) 1325 { 1326 struct ifbreq *req = arg; 1327 struct bridge_iflist *bif; 1328 1329 bif = bridge_lookup_member(sc, req->ifbr_ifsname); 1330 if (bif == NULL) 1331 return (ENOENT); 1332 1333 bridge_delete_member(sc, bif, 0); 1334 1335 return (0); 1336 } 1337 1338 static int 1339 bridge_ioctl_gifflags(struct bridge_softc *sc, void *arg) 1340 { 1341 struct ifbreq *req = arg; 1342 struct bridge_iflist *bif; 1343 struct bstp_port *bp; 1344 1345 bif = bridge_lookup_member(sc, req->ifbr_ifsname); 1346 if (bif == NULL) 1347 return (ENOENT); 1348 1349 bp = &bif->bif_stp; 1350 req->ifbr_ifsflags = bif->bif_flags; 1351 req->ifbr_state = bp->bp_state; 1352 req->ifbr_priority = bp->bp_priority; 1353 req->ifbr_path_cost = bp->bp_path_cost; 1354 req->ifbr_portno = bif->bif_ifp->if_index & 0xfff; 1355 req->ifbr_proto = bp->bp_protover; 1356 req->ifbr_role = bp->bp_role; 1357 req->ifbr_stpflags = bp->bp_flags; 1358 req->ifbr_addrcnt = bif->bif_addrcnt; 1359 req->ifbr_addrmax = bif->bif_addrmax; 1360 req->ifbr_addrexceeded = bif->bif_addrexceeded; 1361 1362 /* Copy STP state options as flags */ 1363 if (bp->bp_operedge) 1364 req->ifbr_ifsflags |= IFBIF_BSTP_EDGE; 1365 if (bp->bp_flags & BSTP_PORT_AUTOEDGE) 1366 req->ifbr_ifsflags |= IFBIF_BSTP_AUTOEDGE; 1367 if (bp->bp_ptp_link) 1368 req->ifbr_ifsflags |= IFBIF_BSTP_PTP; 1369 if (bp->bp_flags & BSTP_PORT_AUTOPTP) 1370 req->ifbr_ifsflags |= IFBIF_BSTP_AUTOPTP; 1371 if (bp->bp_flags & BSTP_PORT_ADMEDGE) 1372 req->ifbr_ifsflags |= IFBIF_BSTP_ADMEDGE; 1373 if (bp->bp_flags & BSTP_PORT_ADMCOST) 1374 req->ifbr_ifsflags |= IFBIF_BSTP_ADMCOST; 1375 return (0); 1376 } 1377 1378 static int 1379 bridge_ioctl_sifflags(struct bridge_softc *sc, void *arg) 1380 { 1381 struct epoch_tracker et; 1382 struct ifbreq *req = arg; 1383 struct bridge_iflist *bif; 1384 struct bstp_port *bp; 1385 int error; 1386 1387 bif = bridge_lookup_member(sc, req->ifbr_ifsname); 1388 if (bif == NULL) 1389 return (ENOENT); 1390 bp = &bif->bif_stp; 1391 1392 if (req->ifbr_ifsflags & IFBIF_SPAN) 1393 /* SPAN is readonly */ 1394 return (EINVAL); 1395 1396 NET_EPOCH_ENTER(et); 1397 1398 if (req->ifbr_ifsflags & IFBIF_STP) { 1399 if ((bif->bif_flags & IFBIF_STP) == 0) { 1400 error = bstp_enable(&bif->bif_stp); 1401 if (error) { 1402 NET_EPOCH_EXIT(et); 1403 return (error); 1404 } 1405 } 1406 } else { 1407 if ((bif->bif_flags & IFBIF_STP) != 0) 1408 bstp_disable(&bif->bif_stp); 1409 } 1410 1411 /* Pass on STP flags */ 1412 bstp_set_edge(bp, req->ifbr_ifsflags & IFBIF_BSTP_EDGE ? 1 : 0); 1413 bstp_set_autoedge(bp, req->ifbr_ifsflags & IFBIF_BSTP_AUTOEDGE ? 1 : 0); 1414 bstp_set_ptp(bp, req->ifbr_ifsflags & IFBIF_BSTP_PTP ? 1 : 0); 1415 bstp_set_autoptp(bp, req->ifbr_ifsflags & IFBIF_BSTP_AUTOPTP ? 1 : 0); 1416 1417 /* Save the bits relating to the bridge */ 1418 bif->bif_flags = req->ifbr_ifsflags & IFBIFMASK; 1419 1420 NET_EPOCH_EXIT(et); 1421 1422 return (0); 1423 } 1424 1425 static int 1426 bridge_ioctl_scache(struct bridge_softc *sc, void *arg) 1427 { 1428 struct ifbrparam *param = arg; 1429 1430 sc->sc_brtmax = param->ifbrp_csize; 1431 bridge_rttrim(sc); 1432 1433 return (0); 1434 } 1435 1436 static int 1437 bridge_ioctl_gcache(struct bridge_softc *sc, void *arg) 1438 { 1439 struct ifbrparam *param = arg; 1440 1441 param->ifbrp_csize = sc->sc_brtmax; 1442 1443 return (0); 1444 } 1445 1446 static int 1447 bridge_ioctl_gifs(struct bridge_softc *sc, void *arg) 1448 { 1449 struct ifbifconf *bifc = arg; 1450 struct bridge_iflist *bif; 1451 struct ifbreq breq; 1452 char *buf, *outbuf; 1453 int count, buflen, len, error = 0; 1454 1455 count = 0; 1456 CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) 1457 count++; 1458 CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) 1459 count++; 1460 1461 buflen = sizeof(breq) * count; 1462 if (bifc->ifbic_len == 0) { 1463 bifc->ifbic_len = buflen; 1464 return (0); 1465 } 1466 outbuf = malloc(buflen, M_TEMP, M_NOWAIT | M_ZERO); 1467 if (outbuf == NULL) 1468 return (ENOMEM); 1469 1470 count = 0; 1471 buf = outbuf; 1472 len = min(bifc->ifbic_len, buflen); 1473 bzero(&breq, sizeof(breq)); 1474 CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 1475 if (len < sizeof(breq)) 1476 break; 1477 1478 strlcpy(breq.ifbr_ifsname, bif->bif_ifp->if_xname, 1479 sizeof(breq.ifbr_ifsname)); 1480 /* Fill in the ifbreq structure */ 1481 error = bridge_ioctl_gifflags(sc, &breq); 1482 if (error) 1483 break; 1484 memcpy(buf, &breq, sizeof(breq)); 1485 count++; 1486 buf += sizeof(breq); 1487 len -= sizeof(breq); 1488 } 1489 CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) { 1490 if (len < sizeof(breq)) 1491 break; 1492 1493 strlcpy(breq.ifbr_ifsname, bif->bif_ifp->if_xname, 1494 sizeof(breq.ifbr_ifsname)); 1495 breq.ifbr_ifsflags = bif->bif_flags; 1496 breq.ifbr_portno = bif->bif_ifp->if_index & 0xfff; 1497 memcpy(buf, &breq, sizeof(breq)); 1498 count++; 1499 buf += sizeof(breq); 1500 len -= sizeof(breq); 1501 } 1502 1503 bifc->ifbic_len = sizeof(breq) * count; 1504 error = copyout(outbuf, bifc->ifbic_req, bifc->ifbic_len); 1505 free(outbuf, M_TEMP); 1506 return (error); 1507 } 1508 1509 static int 1510 bridge_ioctl_rts(struct bridge_softc *sc, void *arg) 1511 { 1512 struct ifbaconf *bac = arg; 1513 struct bridge_rtnode *brt; 1514 struct ifbareq bareq; 1515 char *buf, *outbuf; 1516 int count, buflen, len, error = 0; 1517 1518 if (bac->ifbac_len == 0) 1519 return (0); 1520 1521 count = 0; 1522 CK_LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) 1523 count++; 1524 buflen = sizeof(bareq) * count; 1525 1526 outbuf = malloc(buflen, M_TEMP, M_NOWAIT | M_ZERO); 1527 if (outbuf == NULL) 1528 return (ENOMEM); 1529 1530 count = 0; 1531 buf = outbuf; 1532 len = min(bac->ifbac_len, buflen); 1533 bzero(&bareq, sizeof(bareq)); 1534 CK_LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) { 1535 if (len < sizeof(bareq)) 1536 goto out; 1537 strlcpy(bareq.ifba_ifsname, brt->brt_ifp->if_xname, 1538 sizeof(bareq.ifba_ifsname)); 1539 memcpy(bareq.ifba_dst, brt->brt_addr, sizeof(brt->brt_addr)); 1540 bareq.ifba_vlan = brt->brt_vlan; 1541 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC && 1542 time_uptime < brt->brt_expire) 1543 bareq.ifba_expire = brt->brt_expire - time_uptime; 1544 else 1545 bareq.ifba_expire = 0; 1546 bareq.ifba_flags = brt->brt_flags; 1547 1548 memcpy(buf, &bareq, sizeof(bareq)); 1549 count++; 1550 buf += sizeof(bareq); 1551 len -= sizeof(bareq); 1552 } 1553 out: 1554 bac->ifbac_len = sizeof(bareq) * count; 1555 error = copyout(outbuf, bac->ifbac_req, bac->ifbac_len); 1556 free(outbuf, M_TEMP); 1557 return (error); 1558 } 1559 1560 static int 1561 bridge_ioctl_saddr(struct bridge_softc *sc, void *arg) 1562 { 1563 struct ifbareq *req = arg; 1564 struct bridge_iflist *bif; 1565 struct epoch_tracker et; 1566 int error; 1567 1568 NET_EPOCH_ENTER(et); 1569 bif = bridge_lookup_member(sc, req->ifba_ifsname); 1570 if (bif == NULL) { 1571 NET_EPOCH_EXIT(et); 1572 return (ENOENT); 1573 } 1574 1575 /* bridge_rtupdate() may acquire the lock. */ 1576 error = bridge_rtupdate(sc, req->ifba_dst, req->ifba_vlan, bif, 1, 1577 req->ifba_flags); 1578 NET_EPOCH_EXIT(et); 1579 1580 return (error); 1581 } 1582 1583 static int 1584 bridge_ioctl_sto(struct bridge_softc *sc, void *arg) 1585 { 1586 struct ifbrparam *param = arg; 1587 1588 sc->sc_brttimeout = param->ifbrp_ctime; 1589 return (0); 1590 } 1591 1592 static int 1593 bridge_ioctl_gto(struct bridge_softc *sc, void *arg) 1594 { 1595 struct ifbrparam *param = arg; 1596 1597 param->ifbrp_ctime = sc->sc_brttimeout; 1598 return (0); 1599 } 1600 1601 static int 1602 bridge_ioctl_daddr(struct bridge_softc *sc, void *arg) 1603 { 1604 struct ifbareq *req = arg; 1605 1606 return (bridge_rtdaddr(sc, req->ifba_dst, req->ifba_vlan)); 1607 } 1608 1609 static int 1610 bridge_ioctl_flush(struct bridge_softc *sc, void *arg) 1611 { 1612 struct ifbreq *req = arg; 1613 1614 BRIDGE_RT_LOCK(sc); 1615 bridge_rtflush(sc, req->ifbr_ifsflags); 1616 BRIDGE_RT_UNLOCK(sc); 1617 1618 return (0); 1619 } 1620 1621 static int 1622 bridge_ioctl_gpri(struct bridge_softc *sc, void *arg) 1623 { 1624 struct ifbrparam *param = arg; 1625 struct bstp_state *bs = &sc->sc_stp; 1626 1627 param->ifbrp_prio = bs->bs_bridge_priority; 1628 return (0); 1629 } 1630 1631 static int 1632 bridge_ioctl_spri(struct bridge_softc *sc, void *arg) 1633 { 1634 struct ifbrparam *param = arg; 1635 1636 return (bstp_set_priority(&sc->sc_stp, param->ifbrp_prio)); 1637 } 1638 1639 static int 1640 bridge_ioctl_ght(struct bridge_softc *sc, void *arg) 1641 { 1642 struct ifbrparam *param = arg; 1643 struct bstp_state *bs = &sc->sc_stp; 1644 1645 param->ifbrp_hellotime = bs->bs_bridge_htime >> 8; 1646 return (0); 1647 } 1648 1649 static int 1650 bridge_ioctl_sht(struct bridge_softc *sc, void *arg) 1651 { 1652 struct ifbrparam *param = arg; 1653 1654 return (bstp_set_htime(&sc->sc_stp, param->ifbrp_hellotime)); 1655 } 1656 1657 static int 1658 bridge_ioctl_gfd(struct bridge_softc *sc, void *arg) 1659 { 1660 struct ifbrparam *param = arg; 1661 struct bstp_state *bs = &sc->sc_stp; 1662 1663 param->ifbrp_fwddelay = bs->bs_bridge_fdelay >> 8; 1664 return (0); 1665 } 1666 1667 static int 1668 bridge_ioctl_sfd(struct bridge_softc *sc, void *arg) 1669 { 1670 struct ifbrparam *param = arg; 1671 1672 return (bstp_set_fdelay(&sc->sc_stp, param->ifbrp_fwddelay)); 1673 } 1674 1675 static int 1676 bridge_ioctl_gma(struct bridge_softc *sc, void *arg) 1677 { 1678 struct ifbrparam *param = arg; 1679 struct bstp_state *bs = &sc->sc_stp; 1680 1681 param->ifbrp_maxage = bs->bs_bridge_max_age >> 8; 1682 return (0); 1683 } 1684 1685 static int 1686 bridge_ioctl_sma(struct bridge_softc *sc, void *arg) 1687 { 1688 struct ifbrparam *param = arg; 1689 1690 return (bstp_set_maxage(&sc->sc_stp, param->ifbrp_maxage)); 1691 } 1692 1693 static int 1694 bridge_ioctl_sifprio(struct bridge_softc *sc, void *arg) 1695 { 1696 struct ifbreq *req = arg; 1697 struct bridge_iflist *bif; 1698 1699 bif = bridge_lookup_member(sc, req->ifbr_ifsname); 1700 if (bif == NULL) 1701 return (ENOENT); 1702 1703 return (bstp_set_port_priority(&bif->bif_stp, req->ifbr_priority)); 1704 } 1705 1706 static int 1707 bridge_ioctl_sifcost(struct bridge_softc *sc, void *arg) 1708 { 1709 struct ifbreq *req = arg; 1710 struct bridge_iflist *bif; 1711 1712 bif = bridge_lookup_member(sc, req->ifbr_ifsname); 1713 if (bif == NULL) 1714 return (ENOENT); 1715 1716 return (bstp_set_path_cost(&bif->bif_stp, req->ifbr_path_cost)); 1717 } 1718 1719 static int 1720 bridge_ioctl_sifmaxaddr(struct bridge_softc *sc, void *arg) 1721 { 1722 struct ifbreq *req = arg; 1723 struct bridge_iflist *bif; 1724 1725 bif = bridge_lookup_member(sc, req->ifbr_ifsname); 1726 if (bif == NULL) 1727 return (ENOENT); 1728 1729 bif->bif_addrmax = req->ifbr_addrmax; 1730 return (0); 1731 } 1732 1733 static int 1734 bridge_ioctl_addspan(struct bridge_softc *sc, void *arg) 1735 { 1736 struct ifbreq *req = arg; 1737 struct bridge_iflist *bif = NULL; 1738 struct ifnet *ifs; 1739 1740 ifs = ifunit(req->ifbr_ifsname); 1741 if (ifs == NULL) 1742 return (ENOENT); 1743 1744 CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) 1745 if (ifs == bif->bif_ifp) 1746 return (EBUSY); 1747 1748 if (ifs->if_bridge != NULL) 1749 return (EBUSY); 1750 1751 switch (ifs->if_type) { 1752 case IFT_ETHER: 1753 case IFT_GIF: 1754 case IFT_L2VLAN: 1755 break; 1756 default: 1757 return (EINVAL); 1758 } 1759 1760 bif = malloc(sizeof(*bif), M_DEVBUF, M_NOWAIT|M_ZERO); 1761 if (bif == NULL) 1762 return (ENOMEM); 1763 1764 bif->bif_ifp = ifs; 1765 bif->bif_flags = IFBIF_SPAN; 1766 1767 CK_LIST_INSERT_HEAD(&sc->sc_spanlist, bif, bif_next); 1768 1769 return (0); 1770 } 1771 1772 static int 1773 bridge_ioctl_delspan(struct bridge_softc *sc, void *arg) 1774 { 1775 struct ifbreq *req = arg; 1776 struct bridge_iflist *bif; 1777 struct ifnet *ifs; 1778 1779 ifs = ifunit(req->ifbr_ifsname); 1780 if (ifs == NULL) 1781 return (ENOENT); 1782 1783 CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) 1784 if (ifs == bif->bif_ifp) 1785 break; 1786 1787 if (bif == NULL) 1788 return (ENOENT); 1789 1790 bridge_delete_span(sc, bif); 1791 1792 return (0); 1793 } 1794 1795 static int 1796 bridge_ioctl_gbparam(struct bridge_softc *sc, void *arg) 1797 { 1798 struct ifbropreq *req = arg; 1799 struct bstp_state *bs = &sc->sc_stp; 1800 struct bstp_port *root_port; 1801 1802 req->ifbop_maxage = bs->bs_bridge_max_age >> 8; 1803 req->ifbop_hellotime = bs->bs_bridge_htime >> 8; 1804 req->ifbop_fwddelay = bs->bs_bridge_fdelay >> 8; 1805 1806 root_port = bs->bs_root_port; 1807 if (root_port == NULL) 1808 req->ifbop_root_port = 0; 1809 else 1810 req->ifbop_root_port = root_port->bp_ifp->if_index; 1811 1812 req->ifbop_holdcount = bs->bs_txholdcount; 1813 req->ifbop_priority = bs->bs_bridge_priority; 1814 req->ifbop_protocol = bs->bs_protover; 1815 req->ifbop_root_path_cost = bs->bs_root_pv.pv_cost; 1816 req->ifbop_bridgeid = bs->bs_bridge_pv.pv_dbridge_id; 1817 req->ifbop_designated_root = bs->bs_root_pv.pv_root_id; 1818 req->ifbop_designated_bridge = bs->bs_root_pv.pv_dbridge_id; 1819 req->ifbop_last_tc_time.tv_sec = bs->bs_last_tc_time.tv_sec; 1820 req->ifbop_last_tc_time.tv_usec = bs->bs_last_tc_time.tv_usec; 1821 1822 return (0); 1823 } 1824 1825 static int 1826 bridge_ioctl_grte(struct bridge_softc *sc, void *arg) 1827 { 1828 struct ifbrparam *param = arg; 1829 1830 param->ifbrp_cexceeded = sc->sc_brtexceeded; 1831 return (0); 1832 } 1833 1834 static int 1835 bridge_ioctl_gifsstp(struct bridge_softc *sc, void *arg) 1836 { 1837 struct ifbpstpconf *bifstp = arg; 1838 struct bridge_iflist *bif; 1839 struct bstp_port *bp; 1840 struct ifbpstpreq bpreq; 1841 char *buf, *outbuf; 1842 int count, buflen, len, error = 0; 1843 1844 count = 0; 1845 CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 1846 if ((bif->bif_flags & IFBIF_STP) != 0) 1847 count++; 1848 } 1849 1850 buflen = sizeof(bpreq) * count; 1851 if (bifstp->ifbpstp_len == 0) { 1852 bifstp->ifbpstp_len = buflen; 1853 return (0); 1854 } 1855 1856 outbuf = malloc(buflen, M_TEMP, M_NOWAIT | M_ZERO); 1857 if (outbuf == NULL) 1858 return (ENOMEM); 1859 1860 count = 0; 1861 buf = outbuf; 1862 len = min(bifstp->ifbpstp_len, buflen); 1863 bzero(&bpreq, sizeof(bpreq)); 1864 CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 1865 if (len < sizeof(bpreq)) 1866 break; 1867 1868 if ((bif->bif_flags & IFBIF_STP) == 0) 1869 continue; 1870 1871 bp = &bif->bif_stp; 1872 bpreq.ifbp_portno = bif->bif_ifp->if_index & 0xfff; 1873 bpreq.ifbp_fwd_trans = bp->bp_forward_transitions; 1874 bpreq.ifbp_design_cost = bp->bp_desg_pv.pv_cost; 1875 bpreq.ifbp_design_port = bp->bp_desg_pv.pv_port_id; 1876 bpreq.ifbp_design_bridge = bp->bp_desg_pv.pv_dbridge_id; 1877 bpreq.ifbp_design_root = bp->bp_desg_pv.pv_root_id; 1878 1879 memcpy(buf, &bpreq, sizeof(bpreq)); 1880 count++; 1881 buf += sizeof(bpreq); 1882 len -= sizeof(bpreq); 1883 } 1884 1885 bifstp->ifbpstp_len = sizeof(bpreq) * count; 1886 error = copyout(outbuf, bifstp->ifbpstp_req, bifstp->ifbpstp_len); 1887 free(outbuf, M_TEMP); 1888 return (error); 1889 } 1890 1891 static int 1892 bridge_ioctl_sproto(struct bridge_softc *sc, void *arg) 1893 { 1894 struct ifbrparam *param = arg; 1895 1896 return (bstp_set_protocol(&sc->sc_stp, param->ifbrp_proto)); 1897 } 1898 1899 static int 1900 bridge_ioctl_stxhc(struct bridge_softc *sc, void *arg) 1901 { 1902 struct ifbrparam *param = arg; 1903 1904 return (bstp_set_holdcount(&sc->sc_stp, param->ifbrp_txhc)); 1905 } 1906 1907 /* 1908 * bridge_ifdetach: 1909 * 1910 * Detach an interface from a bridge. Called when a member 1911 * interface is detaching. 1912 */ 1913 static void 1914 bridge_ifdetach(void *arg __unused, struct ifnet *ifp) 1915 { 1916 struct bridge_softc *sc = ifp->if_bridge; 1917 struct bridge_iflist *bif; 1918 1919 if (ifp->if_flags & IFF_RENAMING) 1920 return; 1921 if (V_bridge_cloner == NULL) { 1922 /* 1923 * This detach handler can be called after 1924 * vnet_bridge_uninit(). Just return in that case. 1925 */ 1926 return; 1927 } 1928 /* Check if the interface is a bridge member */ 1929 if (sc != NULL) { 1930 BRIDGE_LOCK(sc); 1931 1932 bif = bridge_lookup_member_if(sc, ifp); 1933 if (bif != NULL) 1934 bridge_delete_member(sc, bif, 1); 1935 1936 BRIDGE_UNLOCK(sc); 1937 return; 1938 } 1939 1940 /* Check if the interface is a span port */ 1941 BRIDGE_LIST_LOCK(); 1942 LIST_FOREACH(sc, &V_bridge_list, sc_list) { 1943 BRIDGE_LOCK(sc); 1944 CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) 1945 if (ifp == bif->bif_ifp) { 1946 bridge_delete_span(sc, bif); 1947 break; 1948 } 1949 1950 BRIDGE_UNLOCK(sc); 1951 } 1952 BRIDGE_LIST_UNLOCK(); 1953 } 1954 1955 /* 1956 * bridge_init: 1957 * 1958 * Initialize a bridge interface. 1959 */ 1960 static void 1961 bridge_init(void *xsc) 1962 { 1963 struct bridge_softc *sc = (struct bridge_softc *)xsc; 1964 struct ifnet *ifp = sc->sc_ifp; 1965 1966 if (ifp->if_drv_flags & IFF_DRV_RUNNING) 1967 return; 1968 1969 BRIDGE_LOCK(sc); 1970 callout_reset(&sc->sc_brcallout, bridge_rtable_prune_period * hz, 1971 bridge_timer, sc); 1972 1973 ifp->if_drv_flags |= IFF_DRV_RUNNING; 1974 bstp_init(&sc->sc_stp); /* Initialize Spanning Tree */ 1975 1976 BRIDGE_UNLOCK(sc); 1977 } 1978 1979 /* 1980 * bridge_stop: 1981 * 1982 * Stop the bridge interface. 1983 */ 1984 static void 1985 bridge_stop(struct ifnet *ifp, int disable) 1986 { 1987 struct bridge_softc *sc = ifp->if_softc; 1988 1989 BRIDGE_LOCK_ASSERT(sc); 1990 1991 if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) 1992 return; 1993 1994 BRIDGE_RT_LOCK(sc); 1995 callout_stop(&sc->sc_brcallout); 1996 1997 bstp_stop(&sc->sc_stp); 1998 1999 bridge_rtflush(sc, IFBF_FLUSHDYN); 2000 BRIDGE_RT_UNLOCK(sc); 2001 2002 ifp->if_drv_flags &= ~IFF_DRV_RUNNING; 2003 } 2004 2005 /* 2006 * bridge_enqueue: 2007 * 2008 * Enqueue a packet on a bridge member interface. 2009 * 2010 */ 2011 static int 2012 bridge_enqueue(struct bridge_softc *sc, struct ifnet *dst_ifp, struct mbuf *m) 2013 { 2014 int len, err = 0; 2015 short mflags; 2016 struct mbuf *m0; 2017 2018 /* We may be sending a fragment so traverse the mbuf */ 2019 for (; m; m = m0) { 2020 m0 = m->m_nextpkt; 2021 m->m_nextpkt = NULL; 2022 len = m->m_pkthdr.len; 2023 mflags = m->m_flags; 2024 2025 /* 2026 * If underlying interface can not do VLAN tag insertion itself 2027 * then attach a packet tag that holds it. 2028 */ 2029 if ((m->m_flags & M_VLANTAG) && 2030 (dst_ifp->if_capenable & IFCAP_VLAN_HWTAGGING) == 0) { 2031 m = ether_vlanencap(m, m->m_pkthdr.ether_vtag); 2032 if (m == NULL) { 2033 if_printf(dst_ifp, 2034 "unable to prepend VLAN header\n"); 2035 if_inc_counter(dst_ifp, IFCOUNTER_OERRORS, 1); 2036 continue; 2037 } 2038 m->m_flags &= ~M_VLANTAG; 2039 } 2040 2041 M_ASSERTPKTHDR(m); /* We shouldn't transmit mbuf without pkthdr */ 2042 if ((err = dst_ifp->if_transmit(dst_ifp, m))) { 2043 m_freem(m0); 2044 if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1); 2045 break; 2046 } 2047 2048 if_inc_counter(sc->sc_ifp, IFCOUNTER_OPACKETS, 1); 2049 if_inc_counter(sc->sc_ifp, IFCOUNTER_OBYTES, len); 2050 if (mflags & M_MCAST) 2051 if_inc_counter(sc->sc_ifp, IFCOUNTER_OMCASTS, 1); 2052 } 2053 2054 return (err); 2055 } 2056 2057 /* 2058 * bridge_dummynet: 2059 * 2060 * Receive a queued packet from dummynet and pass it on to the output 2061 * interface. 2062 * 2063 * The mbuf has the Ethernet header already attached. 2064 */ 2065 static void 2066 bridge_dummynet(struct mbuf *m, struct ifnet *ifp) 2067 { 2068 struct bridge_softc *sc; 2069 2070 sc = ifp->if_bridge; 2071 2072 /* 2073 * The packet didnt originate from a member interface. This should only 2074 * ever happen if a member interface is removed while packets are 2075 * queued for it. 2076 */ 2077 if (sc == NULL) { 2078 m_freem(m); 2079 return; 2080 } 2081 2082 if (PFIL_HOOKED_OUT(V_inet_pfil_head) 2083 #ifdef INET6 2084 || PFIL_HOOKED_OUT(V_inet6_pfil_head) 2085 #endif 2086 ) { 2087 if (bridge_pfil(&m, sc->sc_ifp, ifp, PFIL_OUT) != 0) 2088 return; 2089 if (m == NULL) 2090 return; 2091 } 2092 2093 bridge_enqueue(sc, ifp, m); 2094 } 2095 2096 /* 2097 * bridge_output: 2098 * 2099 * Send output from a bridge member interface. This 2100 * performs the bridging function for locally originated 2101 * packets. 2102 * 2103 * The mbuf has the Ethernet header already attached. We must 2104 * enqueue or free the mbuf before returning. 2105 */ 2106 static int 2107 bridge_output(struct ifnet *ifp, struct mbuf *m, struct sockaddr *sa, 2108 struct rtentry *rt) 2109 { 2110 struct ether_header *eh; 2111 struct ifnet *bifp, *dst_if; 2112 struct bridge_softc *sc; 2113 uint16_t vlan; 2114 2115 NET_EPOCH_ASSERT(); 2116 2117 if (m->m_len < ETHER_HDR_LEN) { 2118 m = m_pullup(m, ETHER_HDR_LEN); 2119 if (m == NULL) 2120 return (0); 2121 } 2122 2123 eh = mtod(m, struct ether_header *); 2124 sc = ifp->if_bridge; 2125 vlan = VLANTAGOF(m); 2126 2127 bifp = sc->sc_ifp; 2128 2129 /* 2130 * If bridge is down, but the original output interface is up, 2131 * go ahead and send out that interface. Otherwise, the packet 2132 * is dropped below. 2133 */ 2134 if ((bifp->if_drv_flags & IFF_DRV_RUNNING) == 0) { 2135 dst_if = ifp; 2136 goto sendunicast; 2137 } 2138 2139 /* 2140 * If the packet is a multicast, or we don't know a better way to 2141 * get there, send to all interfaces. 2142 */ 2143 if (ETHER_IS_MULTICAST(eh->ether_dhost)) 2144 dst_if = NULL; 2145 else 2146 dst_if = bridge_rtlookup(sc, eh->ether_dhost, vlan); 2147 /* Tap any traffic not passing back out the originating interface */ 2148 if (dst_if != ifp) 2149 ETHER_BPF_MTAP(bifp, m); 2150 if (dst_if == NULL) { 2151 struct bridge_iflist *bif; 2152 struct mbuf *mc; 2153 int used = 0; 2154 2155 bridge_span(sc, m); 2156 2157 CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 2158 dst_if = bif->bif_ifp; 2159 2160 if (dst_if->if_type == IFT_GIF) 2161 continue; 2162 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) 2163 continue; 2164 2165 /* 2166 * If this is not the original output interface, 2167 * and the interface is participating in spanning 2168 * tree, make sure the port is in a state that 2169 * allows forwarding. 2170 */ 2171 if (dst_if != ifp && (bif->bif_flags & IFBIF_STP) && 2172 bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) 2173 continue; 2174 2175 if (CK_LIST_NEXT(bif, bif_next) == NULL) { 2176 used = 1; 2177 mc = m; 2178 } else { 2179 mc = m_dup(m, M_NOWAIT); 2180 if (mc == NULL) { 2181 if_inc_counter(bifp, IFCOUNTER_OERRORS, 1); 2182 continue; 2183 } 2184 } 2185 2186 bridge_enqueue(sc, dst_if, mc); 2187 } 2188 if (used == 0) 2189 m_freem(m); 2190 return (0); 2191 } 2192 2193 sendunicast: 2194 /* 2195 * XXX Spanning tree consideration here? 2196 */ 2197 2198 bridge_span(sc, m); 2199 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) { 2200 m_freem(m); 2201 return (0); 2202 } 2203 2204 bridge_enqueue(sc, dst_if, m); 2205 return (0); 2206 } 2207 2208 /* 2209 * bridge_transmit: 2210 * 2211 * Do output on a bridge. 2212 * 2213 */ 2214 static int 2215 bridge_transmit(struct ifnet *ifp, struct mbuf *m) 2216 { 2217 struct bridge_softc *sc; 2218 struct ether_header *eh; 2219 struct ifnet *dst_if; 2220 int error = 0; 2221 2222 sc = ifp->if_softc; 2223 2224 ETHER_BPF_MTAP(ifp, m); 2225 2226 eh = mtod(m, struct ether_header *); 2227 2228 if (((m->m_flags & (M_BCAST|M_MCAST)) == 0) && 2229 (dst_if = bridge_rtlookup(sc, eh->ether_dhost, 1)) != NULL) { 2230 error = bridge_enqueue(sc, dst_if, m); 2231 } else 2232 bridge_broadcast(sc, ifp, m, 0); 2233 2234 return (error); 2235 } 2236 2237 #ifdef ALTQ 2238 static void 2239 bridge_altq_start(if_t ifp) 2240 { 2241 struct ifaltq *ifq = &ifp->if_snd; 2242 struct mbuf *m; 2243 2244 IFQ_LOCK(ifq); 2245 IFQ_DEQUEUE_NOLOCK(ifq, m); 2246 while (m != NULL) { 2247 bridge_transmit(ifp, m); 2248 IFQ_DEQUEUE_NOLOCK(ifq, m); 2249 } 2250 IFQ_UNLOCK(ifq); 2251 } 2252 2253 static int 2254 bridge_altq_transmit(if_t ifp, struct mbuf *m) 2255 { 2256 int err; 2257 2258 if (ALTQ_IS_ENABLED(&ifp->if_snd)) { 2259 IFQ_ENQUEUE(&ifp->if_snd, m, err); 2260 if (err == 0) 2261 bridge_altq_start(ifp); 2262 } else 2263 err = bridge_transmit(ifp, m); 2264 2265 return (err); 2266 } 2267 #endif /* ALTQ */ 2268 2269 /* 2270 * The ifp->if_qflush entry point for if_bridge(4) is no-op. 2271 */ 2272 static void 2273 bridge_qflush(struct ifnet *ifp __unused) 2274 { 2275 } 2276 2277 /* 2278 * bridge_forward: 2279 * 2280 * The forwarding function of the bridge. 2281 * 2282 * NOTE: Releases the lock on return. 2283 */ 2284 static void 2285 bridge_forward(struct bridge_softc *sc, struct bridge_iflist *sbif, 2286 struct mbuf *m) 2287 { 2288 struct bridge_iflist *dbif; 2289 struct ifnet *src_if, *dst_if, *ifp; 2290 struct ether_header *eh; 2291 uint16_t vlan; 2292 uint8_t *dst; 2293 int error; 2294 2295 NET_EPOCH_ASSERT(); 2296 2297 src_if = m->m_pkthdr.rcvif; 2298 ifp = sc->sc_ifp; 2299 2300 if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1); 2301 if_inc_counter(ifp, IFCOUNTER_IBYTES, m->m_pkthdr.len); 2302 vlan = VLANTAGOF(m); 2303 2304 if ((sbif->bif_flags & IFBIF_STP) && 2305 sbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) 2306 goto drop; 2307 2308 eh = mtod(m, struct ether_header *); 2309 dst = eh->ether_dhost; 2310 2311 /* If the interface is learning, record the address. */ 2312 if (sbif->bif_flags & IFBIF_LEARNING) { 2313 error = bridge_rtupdate(sc, eh->ether_shost, vlan, 2314 sbif, 0, IFBAF_DYNAMIC); 2315 /* 2316 * If the interface has addresses limits then deny any source 2317 * that is not in the cache. 2318 */ 2319 if (error && sbif->bif_addrmax) 2320 goto drop; 2321 } 2322 2323 if ((sbif->bif_flags & IFBIF_STP) != 0 && 2324 sbif->bif_stp.bp_state == BSTP_IFSTATE_LEARNING) 2325 goto drop; 2326 2327 /* 2328 * At this point, the port either doesn't participate 2329 * in spanning tree or it is in the forwarding state. 2330 */ 2331 2332 /* 2333 * If the packet is unicast, destined for someone on 2334 * "this" side of the bridge, drop it. 2335 */ 2336 if ((m->m_flags & (M_BCAST|M_MCAST)) == 0) { 2337 dst_if = bridge_rtlookup(sc, dst, vlan); 2338 if (src_if == dst_if) 2339 goto drop; 2340 } else { 2341 /* 2342 * Check if its a reserved multicast address, any address 2343 * listed in 802.1D section 7.12.6 may not be forwarded by the 2344 * bridge. 2345 * This is currently 01-80-C2-00-00-00 to 01-80-C2-00-00-0F 2346 */ 2347 if (dst[0] == 0x01 && dst[1] == 0x80 && 2348 dst[2] == 0xc2 && dst[3] == 0x00 && 2349 dst[4] == 0x00 && dst[5] <= 0x0f) 2350 goto drop; 2351 2352 /* ...forward it to all interfaces. */ 2353 if_inc_counter(ifp, IFCOUNTER_IMCASTS, 1); 2354 dst_if = NULL; 2355 } 2356 2357 /* 2358 * If we have a destination interface which is a member of our bridge, 2359 * OR this is a unicast packet, push it through the bpf(4) machinery. 2360 * For broadcast or multicast packets, don't bother because it will 2361 * be reinjected into ether_input. We do this before we pass the packets 2362 * through the pfil(9) framework, as it is possible that pfil(9) will 2363 * drop the packet, or possibly modify it, making it difficult to debug 2364 * firewall issues on the bridge. 2365 */ 2366 if (dst_if != NULL || (m->m_flags & (M_BCAST | M_MCAST)) == 0) 2367 ETHER_BPF_MTAP(ifp, m); 2368 2369 /* run the packet filter */ 2370 if (PFIL_HOOKED_IN(V_inet_pfil_head) 2371 #ifdef INET6 2372 || PFIL_HOOKED_IN(V_inet6_pfil_head) 2373 #endif 2374 ) { 2375 if (bridge_pfil(&m, ifp, src_if, PFIL_IN) != 0) 2376 return; 2377 if (m == NULL) 2378 return; 2379 } 2380 2381 if (dst_if == NULL) { 2382 bridge_broadcast(sc, src_if, m, 1); 2383 return; 2384 } 2385 2386 /* 2387 * At this point, we're dealing with a unicast frame 2388 * going to a different interface. 2389 */ 2390 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) 2391 goto drop; 2392 2393 dbif = bridge_lookup_member_if(sc, dst_if); 2394 if (dbif == NULL) 2395 /* Not a member of the bridge (anymore?) */ 2396 goto drop; 2397 2398 /* Private segments can not talk to each other */ 2399 if (sbif->bif_flags & dbif->bif_flags & IFBIF_PRIVATE) 2400 goto drop; 2401 2402 if ((dbif->bif_flags & IFBIF_STP) && 2403 dbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) 2404 goto drop; 2405 2406 if (PFIL_HOOKED_OUT(V_inet_pfil_head) 2407 #ifdef INET6 2408 || PFIL_HOOKED_OUT(V_inet6_pfil_head) 2409 #endif 2410 ) { 2411 if (bridge_pfil(&m, ifp, dst_if, PFIL_OUT) != 0) 2412 return; 2413 if (m == NULL) 2414 return; 2415 } 2416 2417 bridge_enqueue(sc, dst_if, m); 2418 return; 2419 2420 drop: 2421 m_freem(m); 2422 } 2423 2424 /* 2425 * bridge_input: 2426 * 2427 * Receive input from a member interface. Queue the packet for 2428 * bridging if it is not for us. 2429 */ 2430 static struct mbuf * 2431 bridge_input(struct ifnet *ifp, struct mbuf *m) 2432 { 2433 struct bridge_softc *sc = ifp->if_bridge; 2434 struct bridge_iflist *bif, *bif2; 2435 struct ifnet *bifp; 2436 struct ether_header *eh; 2437 struct mbuf *mc, *mc2; 2438 uint16_t vlan; 2439 int error; 2440 2441 NET_EPOCH_ASSERT(); 2442 2443 if ((sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) 2444 return (m); 2445 2446 bifp = sc->sc_ifp; 2447 vlan = VLANTAGOF(m); 2448 2449 /* 2450 * Implement support for bridge monitoring. If this flag has been 2451 * set on this interface, discard the packet once we push it through 2452 * the bpf(4) machinery, but before we do, increment the byte and 2453 * packet counters associated with this interface. 2454 */ 2455 if ((bifp->if_flags & IFF_MONITOR) != 0) { 2456 m->m_pkthdr.rcvif = bifp; 2457 ETHER_BPF_MTAP(bifp, m); 2458 if_inc_counter(bifp, IFCOUNTER_IPACKETS, 1); 2459 if_inc_counter(bifp, IFCOUNTER_IBYTES, m->m_pkthdr.len); 2460 m_freem(m); 2461 return (NULL); 2462 } 2463 bif = bridge_lookup_member_if(sc, ifp); 2464 if (bif == NULL) { 2465 return (m); 2466 } 2467 2468 eh = mtod(m, struct ether_header *); 2469 2470 bridge_span(sc, m); 2471 2472 if (m->m_flags & (M_BCAST|M_MCAST)) { 2473 /* Tap off 802.1D packets; they do not get forwarded. */ 2474 if (memcmp(eh->ether_dhost, bstp_etheraddr, 2475 ETHER_ADDR_LEN) == 0) { 2476 bstp_input(&bif->bif_stp, ifp, m); /* consumes mbuf */ 2477 return (NULL); 2478 } 2479 2480 if ((bif->bif_flags & IFBIF_STP) && 2481 bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) { 2482 return (m); 2483 } 2484 2485 /* 2486 * Make a deep copy of the packet and enqueue the copy 2487 * for bridge processing; return the original packet for 2488 * local processing. 2489 */ 2490 mc = m_dup(m, M_NOWAIT); 2491 if (mc == NULL) { 2492 return (m); 2493 } 2494 2495 /* Perform the bridge forwarding function with the copy. */ 2496 bridge_forward(sc, bif, mc); 2497 2498 /* 2499 * Reinject the mbuf as arriving on the bridge so we have a 2500 * chance at claiming multicast packets. We can not loop back 2501 * here from ether_input as a bridge is never a member of a 2502 * bridge. 2503 */ 2504 KASSERT(bifp->if_bridge == NULL, 2505 ("loop created in bridge_input")); 2506 mc2 = m_dup(m, M_NOWAIT); 2507 if (mc2 != NULL) { 2508 /* Keep the layer3 header aligned */ 2509 int i = min(mc2->m_pkthdr.len, max_protohdr); 2510 mc2 = m_copyup(mc2, i, ETHER_ALIGN); 2511 } 2512 if (mc2 != NULL) { 2513 mc2->m_pkthdr.rcvif = bifp; 2514 (*bifp->if_input)(bifp, mc2); 2515 } 2516 2517 /* Return the original packet for local processing. */ 2518 return (m); 2519 } 2520 2521 if ((bif->bif_flags & IFBIF_STP) && 2522 bif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) { 2523 return (m); 2524 } 2525 2526 #if (defined(INET) || defined(INET6)) 2527 # define OR_CARP_CHECK_WE_ARE_DST(iface) \ 2528 || ((iface)->if_carp \ 2529 && (*carp_forus_p)((iface), eh->ether_dhost)) 2530 # define OR_CARP_CHECK_WE_ARE_SRC(iface) \ 2531 || ((iface)->if_carp \ 2532 && (*carp_forus_p)((iface), eh->ether_shost)) 2533 #else 2534 # define OR_CARP_CHECK_WE_ARE_DST(iface) 2535 # define OR_CARP_CHECK_WE_ARE_SRC(iface) 2536 #endif 2537 2538 #ifdef INET6 2539 # define OR_PFIL_HOOKED_INET6 \ 2540 || PFIL_HOOKED_IN(V_inet6_pfil_head) 2541 #else 2542 # define OR_PFIL_HOOKED_INET6 2543 #endif 2544 2545 #define GRAB_OUR_PACKETS(iface) \ 2546 if ((iface)->if_type == IFT_GIF) \ 2547 continue; \ 2548 /* It is destined for us. */ \ 2549 if (memcmp(IF_LLADDR((iface)), eh->ether_dhost, ETHER_ADDR_LEN) == 0 \ 2550 OR_CARP_CHECK_WE_ARE_DST((iface)) \ 2551 ) { \ 2552 if (bif->bif_flags & IFBIF_LEARNING) { \ 2553 error = bridge_rtupdate(sc, eh->ether_shost, \ 2554 vlan, bif, 0, IFBAF_DYNAMIC); \ 2555 if (error && bif->bif_addrmax) { \ 2556 m_freem(m); \ 2557 return (NULL); \ 2558 } \ 2559 } \ 2560 m->m_pkthdr.rcvif = iface; \ 2561 if ((iface) == ifp) { \ 2562 /* Skip bridge processing... src == dest */ \ 2563 return (m); \ 2564 } \ 2565 /* It's passing over or to the bridge, locally. */ \ 2566 ETHER_BPF_MTAP(bifp, m); \ 2567 if_inc_counter(bifp, IFCOUNTER_IPACKETS, 1); \ 2568 if_inc_counter(bifp, IFCOUNTER_IBYTES, m->m_pkthdr.len); \ 2569 /* Filter on the physical interface. */ \ 2570 if (V_pfil_local_phys && (PFIL_HOOKED_IN(V_inet_pfil_head) \ 2571 OR_PFIL_HOOKED_INET6)) { \ 2572 if (bridge_pfil(&m, NULL, ifp, \ 2573 PFIL_IN) != 0 || m == NULL) { \ 2574 return (NULL); \ 2575 } \ 2576 } \ 2577 if ((iface) != bifp) \ 2578 ETHER_BPF_MTAP(iface, m); \ 2579 return (m); \ 2580 } \ 2581 \ 2582 /* We just received a packet that we sent out. */ \ 2583 if (memcmp(IF_LLADDR((iface)), eh->ether_shost, ETHER_ADDR_LEN) == 0 \ 2584 OR_CARP_CHECK_WE_ARE_SRC((iface)) \ 2585 ) { \ 2586 m_freem(m); \ 2587 return (NULL); \ 2588 } 2589 2590 /* 2591 * Unicast. Make sure it's not for the bridge. 2592 */ 2593 do { GRAB_OUR_PACKETS(bifp) } while (0); 2594 2595 /* 2596 * Give a chance for ifp at first priority. This will help when the 2597 * packet comes through the interface like VLAN's with the same MACs 2598 * on several interfaces from the same bridge. This also will save 2599 * some CPU cycles in case the destination interface and the input 2600 * interface (eq ifp) are the same. 2601 */ 2602 do { GRAB_OUR_PACKETS(ifp) } while (0); 2603 2604 /* Now check the all bridge members. */ 2605 CK_LIST_FOREACH(bif2, &sc->sc_iflist, bif_next) { 2606 GRAB_OUR_PACKETS(bif2->bif_ifp) 2607 } 2608 2609 #undef OR_CARP_CHECK_WE_ARE_DST 2610 #undef OR_CARP_CHECK_WE_ARE_SRC 2611 #undef OR_PFIL_HOOKED_INET6 2612 #undef GRAB_OUR_PACKETS 2613 2614 /* Perform the bridge forwarding function. */ 2615 bridge_forward(sc, bif, m); 2616 2617 return (NULL); 2618 } 2619 2620 /* 2621 * bridge_broadcast: 2622 * 2623 * Send a frame to all interfaces that are members of 2624 * the bridge, except for the one on which the packet 2625 * arrived. 2626 * 2627 * NOTE: Releases the lock on return. 2628 */ 2629 static void 2630 bridge_broadcast(struct bridge_softc *sc, struct ifnet *src_if, 2631 struct mbuf *m, int runfilt) 2632 { 2633 struct bridge_iflist *dbif, *sbif; 2634 struct mbuf *mc; 2635 struct ifnet *dst_if; 2636 int used = 0, i; 2637 2638 NET_EPOCH_ASSERT(); 2639 2640 sbif = bridge_lookup_member_if(sc, src_if); 2641 2642 /* Filter on the bridge interface before broadcasting */ 2643 if (runfilt && (PFIL_HOOKED_OUT(V_inet_pfil_head) 2644 #ifdef INET6 2645 || PFIL_HOOKED_OUT(V_inet6_pfil_head) 2646 #endif 2647 )) { 2648 if (bridge_pfil(&m, sc->sc_ifp, NULL, PFIL_OUT) != 0) 2649 return; 2650 if (m == NULL) 2651 return; 2652 } 2653 2654 CK_LIST_FOREACH(dbif, &sc->sc_iflist, bif_next) { 2655 dst_if = dbif->bif_ifp; 2656 if (dst_if == src_if) 2657 continue; 2658 2659 /* Private segments can not talk to each other */ 2660 if (sbif && (sbif->bif_flags & dbif->bif_flags & IFBIF_PRIVATE)) 2661 continue; 2662 2663 if ((dbif->bif_flags & IFBIF_STP) && 2664 dbif->bif_stp.bp_state == BSTP_IFSTATE_DISCARDING) 2665 continue; 2666 2667 if ((dbif->bif_flags & IFBIF_DISCOVER) == 0 && 2668 (m->m_flags & (M_BCAST|M_MCAST)) == 0) 2669 continue; 2670 2671 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) 2672 continue; 2673 2674 if (CK_LIST_NEXT(dbif, bif_next) == NULL) { 2675 mc = m; 2676 used = 1; 2677 } else { 2678 mc = m_dup(m, M_NOWAIT); 2679 if (mc == NULL) { 2680 if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1); 2681 continue; 2682 } 2683 } 2684 2685 /* 2686 * Filter on the output interface. Pass a NULL bridge interface 2687 * pointer so we do not redundantly filter on the bridge for 2688 * each interface we broadcast on. 2689 */ 2690 if (runfilt && (PFIL_HOOKED_OUT(V_inet_pfil_head) 2691 #ifdef INET6 2692 || PFIL_HOOKED_OUT(V_inet6_pfil_head) 2693 #endif 2694 )) { 2695 if (used == 0) { 2696 /* Keep the layer3 header aligned */ 2697 i = min(mc->m_pkthdr.len, max_protohdr); 2698 mc = m_copyup(mc, i, ETHER_ALIGN); 2699 if (mc == NULL) { 2700 if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1); 2701 continue; 2702 } 2703 } 2704 if (bridge_pfil(&mc, NULL, dst_if, PFIL_OUT) != 0) 2705 continue; 2706 if (mc == NULL) 2707 continue; 2708 } 2709 2710 bridge_enqueue(sc, dst_if, mc); 2711 } 2712 if (used == 0) 2713 m_freem(m); 2714 } 2715 2716 /* 2717 * bridge_span: 2718 * 2719 * Duplicate a packet out one or more interfaces that are in span mode, 2720 * the original mbuf is unmodified. 2721 */ 2722 static void 2723 bridge_span(struct bridge_softc *sc, struct mbuf *m) 2724 { 2725 struct bridge_iflist *bif; 2726 struct ifnet *dst_if; 2727 struct mbuf *mc; 2728 2729 NET_EPOCH_ASSERT(); 2730 2731 if (CK_LIST_EMPTY(&sc->sc_spanlist)) 2732 return; 2733 2734 CK_LIST_FOREACH(bif, &sc->sc_spanlist, bif_next) { 2735 dst_if = bif->bif_ifp; 2736 2737 if ((dst_if->if_drv_flags & IFF_DRV_RUNNING) == 0) 2738 continue; 2739 2740 mc = m_dup(m, M_NOWAIT); 2741 if (mc == NULL) { 2742 if_inc_counter(sc->sc_ifp, IFCOUNTER_OERRORS, 1); 2743 continue; 2744 } 2745 2746 bridge_enqueue(sc, dst_if, mc); 2747 } 2748 } 2749 2750 /* 2751 * bridge_rtupdate: 2752 * 2753 * Add a bridge routing entry. 2754 */ 2755 static int 2756 bridge_rtupdate(struct bridge_softc *sc, const uint8_t *dst, uint16_t vlan, 2757 struct bridge_iflist *bif, int setflags, uint8_t flags) 2758 { 2759 struct bridge_rtnode *brt; 2760 int error; 2761 2762 BRIDGE_LOCK_OR_NET_EPOCH_ASSERT(sc); 2763 2764 /* Check the source address is valid and not multicast. */ 2765 if (ETHER_IS_MULTICAST(dst) || 2766 (dst[0] == 0 && dst[1] == 0 && dst[2] == 0 && 2767 dst[3] == 0 && dst[4] == 0 && dst[5] == 0) != 0) 2768 return (EINVAL); 2769 2770 /* 802.1p frames map to vlan 1 */ 2771 if (vlan == 0) 2772 vlan = 1; 2773 2774 /* 2775 * A route for this destination might already exist. If so, 2776 * update it, otherwise create a new one. 2777 */ 2778 if ((brt = bridge_rtnode_lookup(sc, dst, vlan)) == NULL) { 2779 BRIDGE_RT_LOCK(sc); 2780 2781 /* Check again, now that we have the lock. There could have 2782 * been a race and we only want to insert this once. */ 2783 if ((brt = bridge_rtnode_lookup(sc, dst, vlan)) != NULL) { 2784 BRIDGE_RT_UNLOCK(sc); 2785 return (0); 2786 } 2787 2788 if (sc->sc_brtcnt >= sc->sc_brtmax) { 2789 sc->sc_brtexceeded++; 2790 BRIDGE_RT_UNLOCK(sc); 2791 return (ENOSPC); 2792 } 2793 /* Check per interface address limits (if enabled) */ 2794 if (bif->bif_addrmax && bif->bif_addrcnt >= bif->bif_addrmax) { 2795 bif->bif_addrexceeded++; 2796 BRIDGE_RT_UNLOCK(sc); 2797 return (ENOSPC); 2798 } 2799 2800 /* 2801 * Allocate a new bridge forwarding node, and 2802 * initialize the expiration time and Ethernet 2803 * address. 2804 */ 2805 brt = uma_zalloc(V_bridge_rtnode_zone, M_NOWAIT | M_ZERO); 2806 if (brt == NULL) { 2807 BRIDGE_RT_UNLOCK(sc); 2808 return (ENOMEM); 2809 } 2810 brt->brt_vnet = curvnet; 2811 2812 if (bif->bif_flags & IFBIF_STICKY) 2813 brt->brt_flags = IFBAF_STICKY; 2814 else 2815 brt->brt_flags = IFBAF_DYNAMIC; 2816 2817 memcpy(brt->brt_addr, dst, ETHER_ADDR_LEN); 2818 brt->brt_vlan = vlan; 2819 2820 if ((error = bridge_rtnode_insert(sc, brt)) != 0) { 2821 uma_zfree(V_bridge_rtnode_zone, brt); 2822 BRIDGE_RT_UNLOCK(sc); 2823 return (error); 2824 } 2825 brt->brt_dst = bif; 2826 bif->bif_addrcnt++; 2827 2828 BRIDGE_RT_UNLOCK(sc); 2829 } 2830 2831 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC && 2832 brt->brt_dst != bif) { 2833 BRIDGE_RT_LOCK(sc); 2834 brt->brt_dst->bif_addrcnt--; 2835 brt->brt_dst = bif; 2836 brt->brt_dst->bif_addrcnt++; 2837 BRIDGE_RT_UNLOCK(sc); 2838 } 2839 2840 if ((flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) 2841 brt->brt_expire = time_uptime + sc->sc_brttimeout; 2842 if (setflags) 2843 brt->brt_flags = flags; 2844 2845 return (0); 2846 } 2847 2848 /* 2849 * bridge_rtlookup: 2850 * 2851 * Lookup the destination interface for an address. 2852 */ 2853 static struct ifnet * 2854 bridge_rtlookup(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan) 2855 { 2856 struct bridge_rtnode *brt; 2857 2858 NET_EPOCH_ASSERT(); 2859 2860 if ((brt = bridge_rtnode_lookup(sc, addr, vlan)) == NULL) 2861 return (NULL); 2862 2863 return (brt->brt_ifp); 2864 } 2865 2866 /* 2867 * bridge_rttrim: 2868 * 2869 * Trim the routine table so that we have a number 2870 * of routing entries less than or equal to the 2871 * maximum number. 2872 */ 2873 static void 2874 bridge_rttrim(struct bridge_softc *sc) 2875 { 2876 struct bridge_rtnode *brt, *nbrt; 2877 2878 NET_EPOCH_ASSERT(); 2879 BRIDGE_RT_LOCK_ASSERT(sc); 2880 2881 /* Make sure we actually need to do this. */ 2882 if (sc->sc_brtcnt <= sc->sc_brtmax) 2883 return; 2884 2885 /* Force an aging cycle; this might trim enough addresses. */ 2886 bridge_rtage(sc); 2887 if (sc->sc_brtcnt <= sc->sc_brtmax) 2888 return; 2889 2890 CK_LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) { 2891 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) { 2892 bridge_rtnode_destroy(sc, brt); 2893 if (sc->sc_brtcnt <= sc->sc_brtmax) 2894 return; 2895 } 2896 } 2897 } 2898 2899 /* 2900 * bridge_timer: 2901 * 2902 * Aging timer for the bridge. 2903 */ 2904 static void 2905 bridge_timer(void *arg) 2906 { 2907 struct bridge_softc *sc = arg; 2908 2909 BRIDGE_RT_LOCK_ASSERT(sc); 2910 2911 /* Destruction of rtnodes requires a proper vnet context */ 2912 CURVNET_SET(sc->sc_ifp->if_vnet); 2913 bridge_rtage(sc); 2914 2915 if (sc->sc_ifp->if_drv_flags & IFF_DRV_RUNNING) 2916 callout_reset(&sc->sc_brcallout, 2917 bridge_rtable_prune_period * hz, bridge_timer, sc); 2918 CURVNET_RESTORE(); 2919 } 2920 2921 /* 2922 * bridge_rtage: 2923 * 2924 * Perform an aging cycle. 2925 */ 2926 static void 2927 bridge_rtage(struct bridge_softc *sc) 2928 { 2929 struct bridge_rtnode *brt, *nbrt; 2930 2931 BRIDGE_RT_LOCK_ASSERT(sc); 2932 2933 CK_LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) { 2934 if ((brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) { 2935 if (time_uptime >= brt->brt_expire) 2936 bridge_rtnode_destroy(sc, brt); 2937 } 2938 } 2939 } 2940 2941 /* 2942 * bridge_rtflush: 2943 * 2944 * Remove all dynamic addresses from the bridge. 2945 */ 2946 static void 2947 bridge_rtflush(struct bridge_softc *sc, int full) 2948 { 2949 struct bridge_rtnode *brt, *nbrt; 2950 2951 BRIDGE_RT_LOCK_ASSERT(sc); 2952 2953 CK_LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) { 2954 if (full || (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) 2955 bridge_rtnode_destroy(sc, brt); 2956 } 2957 } 2958 2959 /* 2960 * bridge_rtdaddr: 2961 * 2962 * Remove an address from the table. 2963 */ 2964 static int 2965 bridge_rtdaddr(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan) 2966 { 2967 struct bridge_rtnode *brt; 2968 int found = 0; 2969 2970 BRIDGE_RT_LOCK(sc); 2971 2972 /* 2973 * If vlan is zero then we want to delete for all vlans so the lookup 2974 * may return more than one. 2975 */ 2976 while ((brt = bridge_rtnode_lookup(sc, addr, vlan)) != NULL) { 2977 bridge_rtnode_destroy(sc, brt); 2978 found = 1; 2979 } 2980 2981 BRIDGE_RT_UNLOCK(sc); 2982 2983 return (found ? 0 : ENOENT); 2984 } 2985 2986 /* 2987 * bridge_rtdelete: 2988 * 2989 * Delete routes to a speicifc member interface. 2990 */ 2991 static void 2992 bridge_rtdelete(struct bridge_softc *sc, struct ifnet *ifp, int full) 2993 { 2994 struct bridge_rtnode *brt, *nbrt; 2995 2996 BRIDGE_RT_LOCK_ASSERT(sc); 2997 2998 CK_LIST_FOREACH_SAFE(brt, &sc->sc_rtlist, brt_list, nbrt) { 2999 if (brt->brt_ifp == ifp && (full || 3000 (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC)) 3001 bridge_rtnode_destroy(sc, brt); 3002 } 3003 } 3004 3005 /* 3006 * bridge_rtable_init: 3007 * 3008 * Initialize the route table for this bridge. 3009 */ 3010 static void 3011 bridge_rtable_init(struct bridge_softc *sc) 3012 { 3013 int i; 3014 3015 sc->sc_rthash = malloc(sizeof(*sc->sc_rthash) * BRIDGE_RTHASH_SIZE, 3016 M_DEVBUF, M_WAITOK); 3017 3018 for (i = 0; i < BRIDGE_RTHASH_SIZE; i++) 3019 CK_LIST_INIT(&sc->sc_rthash[i]); 3020 3021 sc->sc_rthash_key = arc4random(); 3022 CK_LIST_INIT(&sc->sc_rtlist); 3023 } 3024 3025 /* 3026 * bridge_rtable_fini: 3027 * 3028 * Deconstruct the route table for this bridge. 3029 */ 3030 static void 3031 bridge_rtable_fini(struct bridge_softc *sc) 3032 { 3033 3034 KASSERT(sc->sc_brtcnt == 0, 3035 ("%s: %d bridge routes referenced", __func__, sc->sc_brtcnt)); 3036 free(sc->sc_rthash, M_DEVBUF); 3037 } 3038 3039 /* 3040 * The following hash function is adapted from "Hash Functions" by Bob Jenkins 3041 * ("Algorithm Alley", Dr. Dobbs Journal, September 1997). 3042 */ 3043 #define mix(a, b, c) \ 3044 do { \ 3045 a -= b; a -= c; a ^= (c >> 13); \ 3046 b -= c; b -= a; b ^= (a << 8); \ 3047 c -= a; c -= b; c ^= (b >> 13); \ 3048 a -= b; a -= c; a ^= (c >> 12); \ 3049 b -= c; b -= a; b ^= (a << 16); \ 3050 c -= a; c -= b; c ^= (b >> 5); \ 3051 a -= b; a -= c; a ^= (c >> 3); \ 3052 b -= c; b -= a; b ^= (a << 10); \ 3053 c -= a; c -= b; c ^= (b >> 15); \ 3054 } while (/*CONSTCOND*/0) 3055 3056 static __inline uint32_t 3057 bridge_rthash(struct bridge_softc *sc, const uint8_t *addr) 3058 { 3059 uint32_t a = 0x9e3779b9, b = 0x9e3779b9, c = sc->sc_rthash_key; 3060 3061 b += addr[5] << 8; 3062 b += addr[4]; 3063 a += addr[3] << 24; 3064 a += addr[2] << 16; 3065 a += addr[1] << 8; 3066 a += addr[0]; 3067 3068 mix(a, b, c); 3069 3070 return (c & BRIDGE_RTHASH_MASK); 3071 } 3072 3073 #undef mix 3074 3075 static int 3076 bridge_rtnode_addr_cmp(const uint8_t *a, const uint8_t *b) 3077 { 3078 int i, d; 3079 3080 for (i = 0, d = 0; i < ETHER_ADDR_LEN && d == 0; i++) { 3081 d = ((int)a[i]) - ((int)b[i]); 3082 } 3083 3084 return (d); 3085 } 3086 3087 /* 3088 * bridge_rtnode_lookup: 3089 * 3090 * Look up a bridge route node for the specified destination. Compare the 3091 * vlan id or if zero then just return the first match. 3092 */ 3093 static struct bridge_rtnode * 3094 bridge_rtnode_lookup(struct bridge_softc *sc, const uint8_t *addr, uint16_t vlan) 3095 { 3096 struct bridge_rtnode *brt; 3097 uint32_t hash; 3098 int dir; 3099 3100 BRIDGE_RT_LOCK_OR_NET_EPOCH_ASSERT(sc); 3101 3102 hash = bridge_rthash(sc, addr); 3103 CK_LIST_FOREACH(brt, &sc->sc_rthash[hash], brt_hash) { 3104 dir = bridge_rtnode_addr_cmp(addr, brt->brt_addr); 3105 if (dir == 0 && (brt->brt_vlan == vlan || vlan == 0)) 3106 return (brt); 3107 if (dir > 0) 3108 return (NULL); 3109 } 3110 3111 return (NULL); 3112 } 3113 3114 /* 3115 * bridge_rtnode_insert: 3116 * 3117 * Insert the specified bridge node into the route table. We 3118 * assume the entry is not already in the table. 3119 */ 3120 static int 3121 bridge_rtnode_insert(struct bridge_softc *sc, struct bridge_rtnode *brt) 3122 { 3123 struct bridge_rtnode *lbrt; 3124 uint32_t hash; 3125 int dir; 3126 3127 BRIDGE_RT_LOCK_ASSERT(sc); 3128 3129 hash = bridge_rthash(sc, brt->brt_addr); 3130 3131 lbrt = CK_LIST_FIRST(&sc->sc_rthash[hash]); 3132 if (lbrt == NULL) { 3133 CK_LIST_INSERT_HEAD(&sc->sc_rthash[hash], brt, brt_hash); 3134 goto out; 3135 } 3136 3137 do { 3138 dir = bridge_rtnode_addr_cmp(brt->brt_addr, lbrt->brt_addr); 3139 if (dir == 0 && brt->brt_vlan == lbrt->brt_vlan) 3140 return (EEXIST); 3141 if (dir > 0) { 3142 CK_LIST_INSERT_BEFORE(lbrt, brt, brt_hash); 3143 goto out; 3144 } 3145 if (CK_LIST_NEXT(lbrt, brt_hash) == NULL) { 3146 CK_LIST_INSERT_AFTER(lbrt, brt, brt_hash); 3147 goto out; 3148 } 3149 lbrt = CK_LIST_NEXT(lbrt, brt_hash); 3150 } while (lbrt != NULL); 3151 3152 #ifdef DIAGNOSTIC 3153 panic("bridge_rtnode_insert: impossible"); 3154 #endif 3155 3156 out: 3157 CK_LIST_INSERT_HEAD(&sc->sc_rtlist, brt, brt_list); 3158 sc->sc_brtcnt++; 3159 3160 return (0); 3161 } 3162 3163 static void 3164 bridge_rtnode_destroy_cb(struct epoch_context *ctx) 3165 { 3166 struct bridge_rtnode *brt; 3167 3168 brt = __containerof(ctx, struct bridge_rtnode, brt_epoch_ctx); 3169 3170 CURVNET_SET(brt->brt_vnet); 3171 uma_zfree(V_bridge_rtnode_zone, brt); 3172 CURVNET_RESTORE(); 3173 } 3174 3175 /* 3176 * bridge_rtnode_destroy: 3177 * 3178 * Destroy a bridge rtnode. 3179 */ 3180 static void 3181 bridge_rtnode_destroy(struct bridge_softc *sc, struct bridge_rtnode *brt) 3182 { 3183 BRIDGE_RT_LOCK_ASSERT(sc); 3184 3185 CK_LIST_REMOVE(brt, brt_hash); 3186 3187 CK_LIST_REMOVE(brt, brt_list); 3188 sc->sc_brtcnt--; 3189 brt->brt_dst->bif_addrcnt--; 3190 3191 NET_EPOCH_CALL(bridge_rtnode_destroy_cb, &brt->brt_epoch_ctx); 3192 } 3193 3194 /* 3195 * bridge_rtable_expire: 3196 * 3197 * Set the expiry time for all routes on an interface. 3198 */ 3199 static void 3200 bridge_rtable_expire(struct ifnet *ifp, int age) 3201 { 3202 struct bridge_softc *sc = ifp->if_bridge; 3203 struct bridge_rtnode *brt; 3204 3205 CURVNET_SET(ifp->if_vnet); 3206 BRIDGE_RT_LOCK(sc); 3207 3208 /* 3209 * If the age is zero then flush, otherwise set all the expiry times to 3210 * age for the interface 3211 */ 3212 if (age == 0) 3213 bridge_rtdelete(sc, ifp, IFBF_FLUSHDYN); 3214 else { 3215 CK_LIST_FOREACH(brt, &sc->sc_rtlist, brt_list) { 3216 /* Cap the expiry time to 'age' */ 3217 if (brt->brt_ifp == ifp && 3218 brt->brt_expire > time_uptime + age && 3219 (brt->brt_flags & IFBAF_TYPEMASK) == IFBAF_DYNAMIC) 3220 brt->brt_expire = time_uptime + age; 3221 } 3222 } 3223 BRIDGE_RT_UNLOCK(sc); 3224 CURVNET_RESTORE(); 3225 } 3226 3227 /* 3228 * bridge_state_change: 3229 * 3230 * Callback from the bridgestp code when a port changes states. 3231 */ 3232 static void 3233 bridge_state_change(struct ifnet *ifp, int state) 3234 { 3235 struct bridge_softc *sc = ifp->if_bridge; 3236 static const char *stpstates[] = { 3237 "disabled", 3238 "listening", 3239 "learning", 3240 "forwarding", 3241 "blocking", 3242 "discarding" 3243 }; 3244 3245 CURVNET_SET(ifp->if_vnet); 3246 if (V_log_stp) 3247 log(LOG_NOTICE, "%s: state changed to %s on %s\n", 3248 sc->sc_ifp->if_xname, stpstates[state], ifp->if_xname); 3249 CURVNET_RESTORE(); 3250 } 3251 3252 /* 3253 * Send bridge packets through pfil if they are one of the types pfil can deal 3254 * with, or if they are ARP or REVARP. (pfil will pass ARP and REVARP without 3255 * question.) If *bifp or *ifp are NULL then packet filtering is skipped for 3256 * that interface. 3257 */ 3258 static int 3259 bridge_pfil(struct mbuf **mp, struct ifnet *bifp, struct ifnet *ifp, int dir) 3260 { 3261 int snap, error, i, hlen; 3262 struct ether_header *eh1, eh2; 3263 struct ip *ip; 3264 struct llc llc1; 3265 u_int16_t ether_type; 3266 pfil_return_t rv; 3267 3268 snap = 0; 3269 error = -1; /* Default error if not error == 0 */ 3270 3271 #if 0 3272 /* we may return with the IP fields swapped, ensure its not shared */ 3273 KASSERT(M_WRITABLE(*mp), ("%s: modifying a shared mbuf", __func__)); 3274 #endif 3275 3276 if (V_pfil_bridge == 0 && V_pfil_member == 0 && V_pfil_ipfw == 0) 3277 return (0); /* filtering is disabled */ 3278 3279 i = min((*mp)->m_pkthdr.len, max_protohdr); 3280 if ((*mp)->m_len < i) { 3281 *mp = m_pullup(*mp, i); 3282 if (*mp == NULL) { 3283 printf("%s: m_pullup failed\n", __func__); 3284 return (-1); 3285 } 3286 } 3287 3288 eh1 = mtod(*mp, struct ether_header *); 3289 ether_type = ntohs(eh1->ether_type); 3290 3291 /* 3292 * Check for SNAP/LLC. 3293 */ 3294 if (ether_type < ETHERMTU) { 3295 struct llc *llc2 = (struct llc *)(eh1 + 1); 3296 3297 if ((*mp)->m_len >= ETHER_HDR_LEN + 8 && 3298 llc2->llc_dsap == LLC_SNAP_LSAP && 3299 llc2->llc_ssap == LLC_SNAP_LSAP && 3300 llc2->llc_control == LLC_UI) { 3301 ether_type = htons(llc2->llc_un.type_snap.ether_type); 3302 snap = 1; 3303 } 3304 } 3305 3306 /* 3307 * If we're trying to filter bridge traffic, don't look at anything 3308 * other than IP and ARP traffic. If the filter doesn't understand 3309 * IPv6, don't allow IPv6 through the bridge either. This is lame 3310 * since if we really wanted, say, an AppleTalk filter, we are hosed, 3311 * but of course we don't have an AppleTalk filter to begin with. 3312 * (Note that since pfil doesn't understand ARP it will pass *ALL* 3313 * ARP traffic.) 3314 */ 3315 switch (ether_type) { 3316 case ETHERTYPE_ARP: 3317 case ETHERTYPE_REVARP: 3318 if (V_pfil_ipfw_arp == 0) 3319 return (0); /* Automatically pass */ 3320 break; 3321 3322 case ETHERTYPE_IP: 3323 #ifdef INET6 3324 case ETHERTYPE_IPV6: 3325 #endif /* INET6 */ 3326 break; 3327 default: 3328 /* 3329 * Check to see if the user wants to pass non-ip 3330 * packets, these will not be checked by pfil(9) and 3331 * passed unconditionally so the default is to drop. 3332 */ 3333 if (V_pfil_onlyip) 3334 goto bad; 3335 } 3336 3337 /* Run the packet through pfil before stripping link headers */ 3338 if (PFIL_HOOKED_OUT(V_link_pfil_head) && V_pfil_ipfw != 0 && 3339 dir == PFIL_OUT && ifp != NULL) { 3340 switch (pfil_run_hooks(V_link_pfil_head, mp, ifp, dir, NULL)) { 3341 case PFIL_DROPPED: 3342 return (EACCES); 3343 case PFIL_CONSUMED: 3344 return (0); 3345 } 3346 } 3347 3348 /* Strip off the Ethernet header and keep a copy. */ 3349 m_copydata(*mp, 0, ETHER_HDR_LEN, (caddr_t) &eh2); 3350 m_adj(*mp, ETHER_HDR_LEN); 3351 3352 /* Strip off snap header, if present */ 3353 if (snap) { 3354 m_copydata(*mp, 0, sizeof(struct llc), (caddr_t) &llc1); 3355 m_adj(*mp, sizeof(struct llc)); 3356 } 3357 3358 /* 3359 * Check the IP header for alignment and errors 3360 */ 3361 if (dir == PFIL_IN) { 3362 switch (ether_type) { 3363 case ETHERTYPE_IP: 3364 error = bridge_ip_checkbasic(mp); 3365 break; 3366 #ifdef INET6 3367 case ETHERTYPE_IPV6: 3368 error = bridge_ip6_checkbasic(mp); 3369 break; 3370 #endif /* INET6 */ 3371 default: 3372 error = 0; 3373 } 3374 if (error) 3375 goto bad; 3376 } 3377 3378 error = 0; 3379 3380 /* 3381 * Run the packet through pfil 3382 */ 3383 rv = PFIL_PASS; 3384 switch (ether_type) { 3385 case ETHERTYPE_IP: 3386 /* 3387 * Run pfil on the member interface and the bridge, both can 3388 * be skipped by clearing pfil_member or pfil_bridge. 3389 * 3390 * Keep the order: 3391 * in_if -> bridge_if -> out_if 3392 */ 3393 if (V_pfil_bridge && dir == PFIL_OUT && bifp != NULL && (rv = 3394 pfil_run_hooks(V_inet_pfil_head, mp, bifp, dir, NULL)) != 3395 PFIL_PASS) 3396 break; 3397 3398 if (V_pfil_member && ifp != NULL && (rv = 3399 pfil_run_hooks(V_inet_pfil_head, mp, ifp, dir, NULL)) != 3400 PFIL_PASS) 3401 break; 3402 3403 if (V_pfil_bridge && dir == PFIL_IN && bifp != NULL && (rv = 3404 pfil_run_hooks(V_inet_pfil_head, mp, bifp, dir, NULL)) != 3405 PFIL_PASS) 3406 break; 3407 3408 /* check if we need to fragment the packet */ 3409 /* bridge_fragment generates a mbuf chain of packets */ 3410 /* that already include eth headers */ 3411 if (V_pfil_member && ifp != NULL && dir == PFIL_OUT) { 3412 i = (*mp)->m_pkthdr.len; 3413 if (i > ifp->if_mtu) { 3414 error = bridge_fragment(ifp, mp, &eh2, snap, 3415 &llc1); 3416 return (error); 3417 } 3418 } 3419 3420 /* Recalculate the ip checksum. */ 3421 ip = mtod(*mp, struct ip *); 3422 hlen = ip->ip_hl << 2; 3423 if (hlen < sizeof(struct ip)) 3424 goto bad; 3425 if (hlen > (*mp)->m_len) { 3426 if ((*mp = m_pullup(*mp, hlen)) == NULL) 3427 goto bad; 3428 ip = mtod(*mp, struct ip *); 3429 if (ip == NULL) 3430 goto bad; 3431 } 3432 ip->ip_sum = 0; 3433 if (hlen == sizeof(struct ip)) 3434 ip->ip_sum = in_cksum_hdr(ip); 3435 else 3436 ip->ip_sum = in_cksum(*mp, hlen); 3437 3438 break; 3439 #ifdef INET6 3440 case ETHERTYPE_IPV6: 3441 if (V_pfil_bridge && dir == PFIL_OUT && bifp != NULL && (rv = 3442 pfil_run_hooks(V_inet6_pfil_head, mp, bifp, dir, NULL)) != 3443 PFIL_PASS) 3444 break; 3445 3446 if (V_pfil_member && ifp != NULL && (rv = 3447 pfil_run_hooks(V_inet6_pfil_head, mp, ifp, dir, NULL)) != 3448 PFIL_PASS) 3449 break; 3450 3451 if (V_pfil_bridge && dir == PFIL_IN && bifp != NULL && (rv = 3452 pfil_run_hooks(V_inet6_pfil_head, mp, bifp, dir, NULL)) != 3453 PFIL_PASS) 3454 break; 3455 break; 3456 #endif 3457 } 3458 3459 switch (rv) { 3460 case PFIL_CONSUMED: 3461 return (0); 3462 case PFIL_DROPPED: 3463 return (EACCES); 3464 default: 3465 break; 3466 } 3467 3468 error = -1; 3469 3470 /* 3471 * Finally, put everything back the way it was and return 3472 */ 3473 if (snap) { 3474 M_PREPEND(*mp, sizeof(struct llc), M_NOWAIT); 3475 if (*mp == NULL) 3476 return (error); 3477 bcopy(&llc1, mtod(*mp, caddr_t), sizeof(struct llc)); 3478 } 3479 3480 M_PREPEND(*mp, ETHER_HDR_LEN, M_NOWAIT); 3481 if (*mp == NULL) 3482 return (error); 3483 bcopy(&eh2, mtod(*mp, caddr_t), ETHER_HDR_LEN); 3484 3485 return (0); 3486 3487 bad: 3488 m_freem(*mp); 3489 *mp = NULL; 3490 return (error); 3491 } 3492 3493 /* 3494 * Perform basic checks on header size since 3495 * pfil assumes ip_input has already processed 3496 * it for it. Cut-and-pasted from ip_input.c. 3497 * Given how simple the IPv6 version is, 3498 * does the IPv4 version really need to be 3499 * this complicated? 3500 * 3501 * XXX Should we update ipstat here, or not? 3502 * XXX Right now we update ipstat but not 3503 * XXX csum_counter. 3504 */ 3505 static int 3506 bridge_ip_checkbasic(struct mbuf **mp) 3507 { 3508 struct mbuf *m = *mp; 3509 struct ip *ip; 3510 int len, hlen; 3511 u_short sum; 3512 3513 if (*mp == NULL) 3514 return (-1); 3515 3516 if (IP_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) { 3517 if ((m = m_copyup(m, sizeof(struct ip), 3518 (max_linkhdr + 3) & ~3)) == NULL) { 3519 /* XXXJRT new stat, please */ 3520 KMOD_IPSTAT_INC(ips_toosmall); 3521 goto bad; 3522 } 3523 } else if (__predict_false(m->m_len < sizeof (struct ip))) { 3524 if ((m = m_pullup(m, sizeof (struct ip))) == NULL) { 3525 KMOD_IPSTAT_INC(ips_toosmall); 3526 goto bad; 3527 } 3528 } 3529 ip = mtod(m, struct ip *); 3530 if (ip == NULL) goto bad; 3531 3532 if (ip->ip_v != IPVERSION) { 3533 KMOD_IPSTAT_INC(ips_badvers); 3534 goto bad; 3535 } 3536 hlen = ip->ip_hl << 2; 3537 if (hlen < sizeof(struct ip)) { /* minimum header length */ 3538 KMOD_IPSTAT_INC(ips_badhlen); 3539 goto bad; 3540 } 3541 if (hlen > m->m_len) { 3542 if ((m = m_pullup(m, hlen)) == NULL) { 3543 KMOD_IPSTAT_INC(ips_badhlen); 3544 goto bad; 3545 } 3546 ip = mtod(m, struct ip *); 3547 if (ip == NULL) goto bad; 3548 } 3549 3550 if (m->m_pkthdr.csum_flags & CSUM_IP_CHECKED) { 3551 sum = !(m->m_pkthdr.csum_flags & CSUM_IP_VALID); 3552 } else { 3553 if (hlen == sizeof(struct ip)) { 3554 sum = in_cksum_hdr(ip); 3555 } else { 3556 sum = in_cksum(m, hlen); 3557 } 3558 } 3559 if (sum) { 3560 KMOD_IPSTAT_INC(ips_badsum); 3561 goto bad; 3562 } 3563 3564 /* Retrieve the packet length. */ 3565 len = ntohs(ip->ip_len); 3566 3567 /* 3568 * Check for additional length bogosity 3569 */ 3570 if (len < hlen) { 3571 KMOD_IPSTAT_INC(ips_badlen); 3572 goto bad; 3573 } 3574 3575 /* 3576 * Check that the amount of data in the buffers 3577 * is as at least much as the IP header would have us expect. 3578 * Drop packet if shorter than we expect. 3579 */ 3580 if (m->m_pkthdr.len < len) { 3581 KMOD_IPSTAT_INC(ips_tooshort); 3582 goto bad; 3583 } 3584 3585 /* Checks out, proceed */ 3586 *mp = m; 3587 return (0); 3588 3589 bad: 3590 *mp = m; 3591 return (-1); 3592 } 3593 3594 #ifdef INET6 3595 /* 3596 * Same as above, but for IPv6. 3597 * Cut-and-pasted from ip6_input.c. 3598 * XXX Should we update ip6stat, or not? 3599 */ 3600 static int 3601 bridge_ip6_checkbasic(struct mbuf **mp) 3602 { 3603 struct mbuf *m = *mp; 3604 struct ip6_hdr *ip6; 3605 3606 /* 3607 * If the IPv6 header is not aligned, slurp it up into a new 3608 * mbuf with space for link headers, in the event we forward 3609 * it. Otherwise, if it is aligned, make sure the entire base 3610 * IPv6 header is in the first mbuf of the chain. 3611 */ 3612 if (IP6_HDR_ALIGNED_P(mtod(m, caddr_t)) == 0) { 3613 struct ifnet *inifp = m->m_pkthdr.rcvif; 3614 if ((m = m_copyup(m, sizeof(struct ip6_hdr), 3615 (max_linkhdr + 3) & ~3)) == NULL) { 3616 /* XXXJRT new stat, please */ 3617 IP6STAT_INC(ip6s_toosmall); 3618 in6_ifstat_inc(inifp, ifs6_in_hdrerr); 3619 goto bad; 3620 } 3621 } else if (__predict_false(m->m_len < sizeof(struct ip6_hdr))) { 3622 struct ifnet *inifp = m->m_pkthdr.rcvif; 3623 if ((m = m_pullup(m, sizeof(struct ip6_hdr))) == NULL) { 3624 IP6STAT_INC(ip6s_toosmall); 3625 in6_ifstat_inc(inifp, ifs6_in_hdrerr); 3626 goto bad; 3627 } 3628 } 3629 3630 ip6 = mtod(m, struct ip6_hdr *); 3631 3632 if ((ip6->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) { 3633 IP6STAT_INC(ip6s_badvers); 3634 in6_ifstat_inc(m->m_pkthdr.rcvif, ifs6_in_hdrerr); 3635 goto bad; 3636 } 3637 3638 /* Checks out, proceed */ 3639 *mp = m; 3640 return (0); 3641 3642 bad: 3643 *mp = m; 3644 return (-1); 3645 } 3646 #endif /* INET6 */ 3647 3648 /* 3649 * bridge_fragment: 3650 * 3651 * Fragment mbuf chain in multiple packets and prepend ethernet header. 3652 */ 3653 static int 3654 bridge_fragment(struct ifnet *ifp, struct mbuf **mp, struct ether_header *eh, 3655 int snap, struct llc *llc) 3656 { 3657 struct mbuf *m = *mp, *nextpkt = NULL, *mprev = NULL, *mcur = NULL; 3658 struct ip *ip; 3659 int error = -1; 3660 3661 if (m->m_len < sizeof(struct ip) && 3662 (m = m_pullup(m, sizeof(struct ip))) == NULL) 3663 goto dropit; 3664 ip = mtod(m, struct ip *); 3665 3666 m->m_pkthdr.csum_flags |= CSUM_IP; 3667 error = ip_fragment(ip, &m, ifp->if_mtu, ifp->if_hwassist); 3668 if (error) 3669 goto dropit; 3670 3671 /* 3672 * Walk the chain and re-add the Ethernet header for 3673 * each mbuf packet. 3674 */ 3675 for (mcur = m; mcur; mcur = mcur->m_nextpkt) { 3676 nextpkt = mcur->m_nextpkt; 3677 mcur->m_nextpkt = NULL; 3678 if (snap) { 3679 M_PREPEND(mcur, sizeof(struct llc), M_NOWAIT); 3680 if (mcur == NULL) { 3681 error = ENOBUFS; 3682 if (mprev != NULL) 3683 mprev->m_nextpkt = nextpkt; 3684 goto dropit; 3685 } 3686 bcopy(llc, mtod(mcur, caddr_t),sizeof(struct llc)); 3687 } 3688 3689 M_PREPEND(mcur, ETHER_HDR_LEN, M_NOWAIT); 3690 if (mcur == NULL) { 3691 error = ENOBUFS; 3692 if (mprev != NULL) 3693 mprev->m_nextpkt = nextpkt; 3694 goto dropit; 3695 } 3696 bcopy(eh, mtod(mcur, caddr_t), ETHER_HDR_LEN); 3697 3698 /* 3699 * The previous two M_PREPEND could have inserted one or two 3700 * mbufs in front so we have to update the previous packet's 3701 * m_nextpkt. 3702 */ 3703 mcur->m_nextpkt = nextpkt; 3704 if (mprev != NULL) 3705 mprev->m_nextpkt = mcur; 3706 else { 3707 /* The first mbuf in the original chain needs to be 3708 * updated. */ 3709 *mp = mcur; 3710 } 3711 mprev = mcur; 3712 } 3713 3714 KMOD_IPSTAT_INC(ips_fragmented); 3715 return (error); 3716 3717 dropit: 3718 for (mcur = *mp; mcur; mcur = m) { /* droping the full packet chain */ 3719 m = mcur->m_nextpkt; 3720 m_freem(mcur); 3721 } 3722 return (error); 3723 } 3724 3725 static void 3726 bridge_linkstate(struct ifnet *ifp) 3727 { 3728 struct bridge_softc *sc = ifp->if_bridge; 3729 struct bridge_iflist *bif; 3730 struct epoch_tracker et; 3731 3732 NET_EPOCH_ENTER(et); 3733 3734 bif = bridge_lookup_member_if(sc, ifp); 3735 if (bif == NULL) { 3736 NET_EPOCH_EXIT(et); 3737 return; 3738 } 3739 bridge_linkcheck(sc); 3740 3741 bstp_linkstate(&bif->bif_stp); 3742 3743 NET_EPOCH_EXIT(et); 3744 } 3745 3746 static void 3747 bridge_linkcheck(struct bridge_softc *sc) 3748 { 3749 struct bridge_iflist *bif; 3750 int new_link, hasls; 3751 3752 BRIDGE_LOCK_OR_NET_EPOCH_ASSERT(sc); 3753 3754 new_link = LINK_STATE_DOWN; 3755 hasls = 0; 3756 /* Our link is considered up if at least one of our ports is active */ 3757 CK_LIST_FOREACH(bif, &sc->sc_iflist, bif_next) { 3758 if (bif->bif_ifp->if_capabilities & IFCAP_LINKSTATE) 3759 hasls++; 3760 if (bif->bif_ifp->if_link_state == LINK_STATE_UP) { 3761 new_link = LINK_STATE_UP; 3762 break; 3763 } 3764 } 3765 if (!CK_LIST_EMPTY(&sc->sc_iflist) && !hasls) { 3766 /* If no interfaces support link-state then we default to up */ 3767 new_link = LINK_STATE_UP; 3768 } 3769 if_link_state_change(sc->sc_ifp, new_link); 3770 } 3771