1 .\" $NetBSD: bridge.4,v 1.5 2004/01/31 20:14:11 jdc Exp $
3 .\" Copyright 2001 Wasabi Systems, Inc.
4 .\" All rights reserved.
6 .\" Written by Jason R. Thorpe for Wasabi Systems, Inc.
8 .\" Redistribution and use in source and binary forms, with or without
9 .\" modification, are permitted provided that the following conditions
11 .\" 1. Redistributions of source code must retain the above copyright
12 .\" notice, this list of conditions and the following disclaimer.
13 .\" 2. Redistributions in binary form must reproduce the above copyright
14 .\" notice, this list of conditions and the following disclaimer in the
15 .\" documentation and/or other materials provided with the distribution.
16 .\" 3. All advertising materials mentioning features or use of this software
17 .\" must display the following acknowledgement:
18 .\" This product includes software developed for the NetBSD Project by
19 .\" Wasabi Systems, Inc.
20 .\" 4. The name of Wasabi Systems, Inc. may not be used to endorse
21 .\" or promote products derived from this software without specific prior
22 .\" written permission.
24 .\" THIS SOFTWARE IS PROVIDED BY WASABI SYSTEMS, INC. ``AS IS'' AND
25 .\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
26 .\" TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
27 .\" PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL WASABI SYSTEMS, INC
28 .\" BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
29 .\" CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
30 .\" SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
31 .\" INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
32 .\" CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
33 .\" ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
34 .\" POSSIBILITY OF SUCH DAMAGE.
43 .Nd network bridge device
45 To compile this driver into the kernel,
46 place the following line in your
47 kernel configuration file:
48 .Bd -ragged -offset indent
49 .Cd "device if_bridge"
52 Alternatively, to load the driver as a
53 module at boot time, place the following lines in
55 .Bd -literal -offset indent
62 driver creates a logical link between two or more IEEE 802 networks
64 .Dq "similar enough" )
66 For example, it is possible to bridge Ethernet and 802.11 networks together,
67 but it is not possible to bridge Ethernet and Token Ring together.
71 interface is created at runtime using interface cloning.
73 most easily done with the
83 interface randomly chooses a link (MAC) address in the range reserved for
84 locally administered addresses when it is created.
85 This address is guaranteed to be unique
89 interfaces on the local machine.
90 Thus you can theoretically have two bridges on the different machines with
91 the same link addresses.
92 The address can be changed by assigning the desired link address using
98 .Va net.link.bridge.inherit_mac
99 has non-zero value, newly created bridge will inherit MAC address
100 from its first member instead of choosing random link-level address.
101 This will provide more predictable bridge MAC without any
102 additional configuration, but currently this feature is known
103 to break some L2 protocols, for example PPPoE that is provided
108 Now this feature is considered as experimental and is turned off
111 A bridge can be used to provide several services, such as a simple
112 802.11-to-Ethernet bridge for wireless hosts, and traffic isolation.
114 A bridge works like a switch, forwarding traffic from one interface
116 Multicast and broadcast packets are always forwarded to all
117 interfaces that are part of the bridge.
118 For unicast traffic, the bridge learns which MAC addresses are associated
119 with which interfaces and will forward the traffic selectively.
121 All the bridged member interfaces need to be up in order to pass network traffic.
122 These can be enabled using
125 .Va ifconfig_ Ns Ao Ar interface Ac Ns Li ="up"
129 The MTU of the first member interface to be added is used as the bridge MTU.
130 All additional members are required to have exactly the same value.
132 The TXCSUM capability is disabled for any interface added to the bridge, and it
133 is restored when the interface is removed again.
137 where the packets are discarded after
139 processing, and are not processed or forwarded further.
140 This can be used to multiplex the input of two or more interfaces into a single
143 This is useful for reconstructing the traffic for network taps
144 that transmit the RX/TX signals out through two separate interfaces.
149 address family on bridge interfaces.
152 variable configures an IPv6 link-local address on
155 .Bd -literal -offset indent
156 ifconfig_bridge0_ipv6="up"
159 or in a more explicit manner:
160 .Bd -literal -offset indent
161 ifconfig_bridge0_ipv6="inet6 auto_linklocal"
166 address family has a concept of scope zone.
167 Bridging multiple interfaces change the zone configuration because
168 multiple links are merged to each other and form a new single link
169 while the member interfaces still work individually.
170 This means each member interface still has a separate link-local scope
173 interface has another single,
174 aggregated link-local scope zone at the same time.
175 This situation is clearly against the description
176 .Qq zones of the same scope cannot overlap
179 Although it works in most cases,
180 it can cause some conterintuitive or undesirable behavior in some
181 edge cases when both of the
183 interface and one of the member interface have an IPv6 address
184 and applications use both of them.
186 To prevent this situation,
188 checks whether a link-local scoped IPv6 address is configured on
189 a member interface to be added and the
194 interface has IPv6 addresses,
195 IPv6 addresses on the member interface will be automatically removed
196 before the interface is added.
198 This behavior can be disabled by setting
201 .Va net.link.bridge.allow_llz_overlap
209 interface flag are not enabled by default on
212 .Va net.inet6.ip6.accept_rtadv
214 .Va net.inet6.ip6.auto_linklocal
220 driver implements the Rapid Spanning Tree Protocol (RSTP or 802.1w) with
221 backwards compatibility with the legacy Spanning Tree Protocol (STP).
222 Spanning Tree is used to detect and remove loops in a network topology.
224 RSTP provides faster spanning tree convergence than legacy STP, the protocol
225 will exchange information with neighbouring switches to quickly transition to
226 forwarding without creating loops.
228 The code will default to RSTP mode but will downgrade any port connected to a
229 legacy STP network so is fully backward compatible.
230 A bridge can be forced to operate in STP mode without rapid state transitions
236 The bridge can log STP port changes to
239 .Va net.link.bridge.log_stp
244 Packet filtering can be used with any firewall package that hooks in via the
247 When filtering is enabled, bridged packets will pass through the filter
248 inbound on the originating interface, on the bridge interface and outbound on
249 the appropriate interfaces.
250 Either stage can be disabled.
251 The filtering behaviour can be controlled using
253 .Bl -tag -width ".Va net.link.bridge.pfil_onlyip"
254 .It Va net.link.bridge.pfil_onlyip
255 Controls the handling of non-IP packets which are not passed to
259 to only allow IP packets to pass (subject to firewall rules), set to
261 to unconditionally pass all non-IP Ethernet frames.
262 .It Va net.link.bridge.pfil_member
265 to enable filtering on the incoming and outgoing member interfaces, set
269 .It Va net.link.bridge.pfil_bridge
272 to enable filtering on the bridge interface, set
276 .It Va net.link.bridge.pfil_local_phys
279 to additionally filter on the physical interface for locally destined packets.
282 to disable this feature.
283 .It Va net.link.bridge.ipfw
286 to enable layer2 filtering with
291 This needs to be enabled for
300 will be disabled so that IPFW
301 is not run twice; these can be re-enabled if desired.
302 .It Va net.link.bridge.ipfw_arp
305 to enable layer2 ARP filtering with
315 ARP and REVARP packets are forwarded without being filtered and others
316 that are not IP nor IPv6 packets are not forwarded when
319 IPFW can filter Ethernet types using
321 so all packets are passed to
322 the filter for processing.
324 The packets originating from the bridging host will be seen by
325 the filter on the interface that is looked up in the routing
328 The packets destined to the bridging host will be seen by the filter
329 on the interface with the MAC address equal to the packet's destination
331 There are situations when some of the bridge members are sharing
332 the same MAC address (for example the
334 interfaces: they are currently sharing the
335 MAC address of the parent physical interface).
336 It is not possible to distinguish between these interfaces using
337 their MAC address, excluding the case when the packet's destination
338 MAC address is equal to the MAC address of the interface on which
339 the packet was entered to the system.
340 In this case the filter will see the incoming packet on this
342 In all other cases the interface seen by the packet filter is chosen
343 from the list of bridge members with the same MAC address and the
344 result strongly depends on the member addition sequence and the
345 actual implementation of
347 It is not recommended to rely on the order chosen by the current
349 implementation: it can be changed in the future.
351 The previous paragraph is best illustrated with the following
356 the MAC address of the incoming packet's destination is
357 .Nm nn:nn:nn:nn:nn:nn ,
359 the interface on which packet entered the system is
364 .Nm xx:xx:xx:xx:xx:xx ,
366 there are possibly other bridge members with the same MAC address
367 .Nm xx:xx:xx:xx:xx:xx ,
369 the bridge has more than one interface that are sharing the
371 .Nm yy:yy:yy:yy:yy:yy ;
378 Then if the MAC address
379 .Nm nn:nn:nn:nn:nn:nn
381 .Nm xx:xx:xx:xx:xx:xx
382 then the filter will see the packet on the interface
384 no matter if there are any other bridge members carrying the same
386 But if the MAC address
387 .Nm nn:nn:nn:nn:nn:nn
389 .Nm yy:yy:yy:yy:yy:yy
390 then the interface that will be seen by the filter is one of the
392 It is not possible to predict the name of the actual interface
393 without the knowledge of the system state and the
395 implementation details.
397 This problem arises for any bridge members that are sharing the same
398 MAC address, not only to the
400 ones: they we taken just as the example of such situation.
401 So if one wants the filter the locally destined packets based on
402 their interface name, one should be aware of this implication.
403 The described situation will appear at least on the filtering bridges
404 that are doing IP-forwarding; in some of such cases it is better
405 to assign the IP address only to the
407 interface and not to the bridge members.
409 .Va net.link.bridge.pfil_local_phys
410 will let you do the additional filtering on the physical interface.
412 The following when placed in the file
414 will cause a bridge called
416 to be created, and will add the interfaces
420 to the bridge, and then enable packet forwarding.
421 Such a configuration could be used to implement a simple
422 802.11-to-Ethernet bridge (assuming the 802.11 interface is
424 .Bd -literal -offset indent
425 cloned_interfaces="bridge0"
426 ifconfig_bridge0="addm wlan0 addm fxp0 up"
429 For the bridge to forward packets all member interfaces and the bridge need
431 The above example would also require:
432 .Bd -literal -offset indent
433 create_args_wlan0="wlanmode hostap"
434 ifconfig_wlan0="up ssid my_ap mode 11g"
438 Consider a system with two 4-port Ethernet boards.
439 The following will cause a bridge consisting of all 8 ports with Rapid Spanning
440 Tree enabled to be created:
441 .Bd -literal -offset indent
442 ifconfig bridge0 create
444 addm fxp0 stp fxp0 \e
445 addm fxp1 stp fxp1 \e
446 addm fxp2 stp fxp2 \e
447 addm fxp3 stp fxp3 \e
448 addm fxp4 stp fxp4 \e
449 addm fxp5 stp fxp5 \e
450 addm fxp6 stp fxp6 \e
451 addm fxp7 stp fxp7 \e
455 The bridge can be used as a regular host interface at the same time as bridging
456 between its member ports.
457 In this example, the bridge connects em0 and em1, and will receive its IP
458 address through DHCP:
459 .Bd -literal -offset indent
460 cloned_interfaces="bridge0"
461 ifconfig_bridge0="addm em0 addm em1 DHCP"
466 The bridge can tunnel Ethernet across an IP internet using the EtherIP
468 This can be combined with
470 to provide an encrypted connection.
473 interface and set the local and remote IP addresses for the
474 tunnel, these are reversed on the remote bridge.
475 .Bd -literal -offset indent
477 ifconfig gif0 tunnel 1.2.3.4 5.6.7.8 up
478 ifconfig bridge0 create
479 ifconfig bridge0 addm fxp0 addm gif0 up
484 6.1, 6.2, 6.3, 7.0, 7.1, and 7.2 have a bug in the EtherIP protocol.
485 For more details and workaround, see
497 driver first appeared in
503 driver was originally written by
505 .Aq jason@thought.net
506 as part of an undergraduate independent study at the University of
507 North Carolina at Greensboro.
511 driver has been heavily modified from the original version by
513 .Aq thorpej@wasabisystems.com .
515 Rapid Spanning Tree Protocol (RSTP) support was added by
517 .Aq thompsa@FreeBSD.org .
521 driver currently supports only Ethernet and Ethernet-like (e.g., 802.11)
522 network devices, with exactly the same interface MTU size as the bridge device.