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
243 Packet filtering can be used with any firewall package that hooks in via the
246 When filtering is enabled, bridged packets will pass through the filter
247 inbound on the originating interface, on the bridge interface and outbound on
248 the appropriate interfaces.
249 Either stage can be disabled.
250 The filtering behaviour can be controlled using
252 .Bl -tag -width ".Va net.link.bridge.pfil_onlyip"
253 .It Va net.link.bridge.pfil_onlyip
254 Controls the handling of non-IP packets which are not passed to
258 to only allow IP packets to pass (subject to firewall rules), set to
260 to unconditionally pass all non-IP Ethernet frames.
261 .It Va net.link.bridge.pfil_member
264 to enable filtering on the incoming and outgoing member interfaces, set
268 .It Va net.link.bridge.pfil_bridge
271 to enable filtering on the bridge interface, set
275 .It Va net.link.bridge.pfil_local_phys
278 to additionally filter on the physical interface for locally destined packets.
281 to disable this feature.
282 .It Va net.link.bridge.ipfw
285 to enable layer2 filtering with
290 This needs to be enabled for
299 will be disabled so that IPFW
300 is not run twice; these can be re-enabled if desired.
301 .It Va net.link.bridge.ipfw_arp
304 to enable layer2 ARP filtering with
314 ARP and REVARP packets are forwarded without being filtered and others
315 that are not IP nor IPv6 packets are not forwarded when
318 IPFW can filter Ethernet types using
320 so all packets are passed to
321 the filter for processing.
323 The packets originating from the bridging host will be seen by
324 the filter on the interface that is looked up in the routing
327 The packets destined to the bridging host will be seen by the filter
328 on the interface with the MAC address equal to the packet's destination
330 There are situations when some of the bridge members are sharing
331 the same MAC address (for example the
333 interfaces: they are currently sharing the
334 MAC address of the parent physical interface).
335 It is not possible to distinguish between these interfaces using
336 their MAC address, excluding the case when the packet's destination
337 MAC address is equal to the MAC address of the interface on which
338 the packet was entered to the system.
339 In this case the filter will see the incoming packet on this
341 In all other cases the interface seen by the packet filter is chosen
342 from the list of bridge members with the same MAC address and the
343 result strongly depends on the member addition sequence and the
344 actual implementation of
346 It is not recommended to rely on the order chosen by the current
348 implementation: it can be changed in the future.
350 The previous paragraph is best illustrated with the following
355 the MAC address of the incoming packet's destination is
356 .Nm nn:nn:nn:nn:nn:nn ,
358 the interface on which packet entered the system is
363 .Nm xx:xx:xx:xx:xx:xx ,
365 there are possibly other bridge members with the same MAC address
366 .Nm xx:xx:xx:xx:xx:xx ,
368 the bridge has more than one interface that are sharing the
370 .Nm yy:yy:yy:yy:yy:yy ;
377 Then if the MAC address
378 .Nm nn:nn:nn:nn:nn:nn
380 .Nm xx:xx:xx:xx:xx:xx
381 then the filter will see the packet on the interface
383 no matter if there are any other bridge members carrying the same
385 But if the MAC address
386 .Nm nn:nn:nn:nn:nn:nn
388 .Nm yy:yy:yy:yy:yy:yy
389 then the interface that will be seen by the filter is one of the
391 It is not possible to predict the name of the actual interface
392 without the knowledge of the system state and the
394 implementation details.
396 This problem arises for any bridge members that are sharing the same
397 MAC address, not only to the
399 ones: they we taken just as the example of such situation.
400 So if one wants the filter the locally destined packets based on
401 their interface name, one should be aware of this implication.
402 The described situation will appear at least on the filtering bridges
403 that are doing IP-forwarding; in some of such cases it is better
404 to assign the IP address only to the
406 interface and not to the bridge members.
408 .Va net.link.bridge.pfil_local_phys
409 will let you do the additional filtering on the physical interface.
411 The following when placed in the file
413 will cause a bridge called
415 to be created, and will add the interfaces
419 to the bridge, and then enable packet forwarding.
420 Such a configuration could be used to implement a simple
421 802.11-to-Ethernet bridge (assuming the 802.11 interface is
423 .Bd -literal -offset indent
424 cloned_interfaces="bridge0"
425 ifconfig_bridge0="addm wlan0 addm fxp0 up"
428 For the bridge to forward packets all member interfaces and the bridge need
430 The above example would also require:
431 .Bd -literal -offset indent
432 create_args_wlan0="wlanmode hostap"
433 ifconfig_wlan0="up ssid my_ap mode 11g"
437 Consider a system with two 4-port Ethernet boards.
438 The following will cause a bridge consisting of all 8 ports with Rapid Spanning
439 Tree enabled to be created:
440 .Bd -literal -offset indent
441 ifconfig bridge0 create
443 addm fxp0 stp fxp0 \e
444 addm fxp1 stp fxp1 \e
445 addm fxp2 stp fxp2 \e
446 addm fxp3 stp fxp3 \e
447 addm fxp4 stp fxp4 \e
448 addm fxp5 stp fxp5 \e
449 addm fxp6 stp fxp6 \e
450 addm fxp7 stp fxp7 \e
454 The bridge can be used as a regular host interface at the same time as bridging
455 between its member ports.
456 In this example, the bridge connects em0 and em1, and will receive its IP
457 address through DHCP:
458 .Bd -literal -offset indent
459 cloned_interfaces="bridge0"
460 ifconfig_bridge0="addm em0 addm em1 DHCP"
465 The bridge can tunnel Ethernet across an IP internet using the EtherIP
467 This can be combined with
469 to provide an encrypted connection.
472 interface and set the local and remote IP addresses for the
473 tunnel, these are reversed on the remote bridge.
474 .Bd -literal -offset indent
476 ifconfig gif0 tunnel 1.2.3.4 5.6.7.8 up
477 ifconfig bridge0 create
478 ifconfig bridge0 addm fxp0 addm gif0 up
483 6.1, 6.2, 6.3, 7.0, 7.1, and 7.2 have a bug in the EtherIP protocol.
484 For more details and workaround, see
496 driver first appeared in
502 driver was originally written by
504 .Aq jason@thought.net
505 as part of an undergraduate independent study at the University of
506 North Carolina at Greensboro.
510 driver has been heavily modified from the original version by
512 .Aq thorpej@wasabisystems.com .
514 Rapid Spanning Tree Protocol (RSTP) support was added by
516 .Aq thompsa@FreeBSD.org .
520 driver currently supports only Ethernet and Ethernet-like (e.g., 802.11)
521 network devices, with exactly the same interface MTU size as the bridge device.