$NetBSD: bridge.4,v 1.5 2004/01/31 20:14:11 jdc Exp $ Copyright 2001 Wasabi Systems, Inc. All rights reserved. Written by Jason R. Thorpe for Wasabi Systems, Inc. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: 1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. 2. Redistributions in binary form must reproduc...
NAMEif_bridge - network bridge device
SYNOPSISTo compile this driver into the kernel, place the following line in your kernel configuration file:
Alternatively, to load the driver as a module at boot time, place the following lines in loader.conf5:
DESCRIPTIONThe sysctl Cm net.inet.tcp.syncookies driver creates a logical link between two or more IEEE 802 networks that use the same (or ``similar enough'' framing format. For example, it is possible to bridge Ethernet and 802.11 networks together, but it is not possible to bridge Ethernet and Token Ring together.
Each sysctl Cm net.inet.tcp.syncookies interface is created at runtime using interface cloning. This is most easily done with the ifconfig(8) create command or using the cloned_interfaces variable in rc.conf5.
The sysctl Cm net.inet.tcp.syncookies interface randomly chooses a link (MAC) address in the range reserved for locally administered addresses when it is created. This address is guaranteed to be unique only across all sysctl Cm net.inet.tcp.syncookies interfaces on the local machine. Thus you can theoretically have two bridges on the different machines with the same link addresses. The address can be changed by assigning the desired link address using ifconfig(8).
If sysctl(8) node net.link.bridge.inherit_mac has non-zero value, newly created bridge will inherit MAC address from its first member instead of choosing random link-level address. This will provide more predictable bridge MAC without any additional configuration, but currently this feature is known to break some L2 protocols, for example PPPoE that is provided by ng_pppoe4 and ppp(8). Now this feature is considered as experimental and is turned off by-default.
A bridge can be used to provide several services, such as a simple 802.11-to-Ethernet bridge for wireless hosts, and traffic isolation.
A bridge works like a switch, forwarding traffic from one interface to another. Multicast and broadcast packets are always forwarded to all interfaces that are part of the bridge. For unicast traffic, the bridge learns which MAC addresses are associated with which interfaces and will forward the traffic selectively.
All the bridged member interfaces need to be up in order to pass network traffic. These can be enabled using ifconfig(8) or ifconfig_ Ao interface Ac =up in rc.conf5.
The MTU of the first member interface to be added is used as the bridge MTU. All additional members are required to have exactly the same value.
The TXCSUM capability is disabled for any interface added to the bridge, and it is restored when the interface is removed again.
The bridge supports ``monitor mode'' where the packets are discarded after bpf(4) processing, and are not processed or forwarded further. This can be used to multiplex the input of two or more interfaces into a single bpf(4) stream. This is useful for reconstructing the traffic for network taps that transmit the RX/TX signals out through two separate interfaces.
IPV6 SUPPORTsysctl Cm net.inet.tcp.syncookies supports the AF_INET6 address family on bridge interfaces. The following rc.conf5 variable configures an IPv6 link-local address on bridge0 interface:
or in a more explicit manner:
However, the AF_INET6 address family has a concept of scope zone. Bridging multiple interfaces change the zone configuration because multiple links are merged to each other and form a new single link while the member interfaces still work individually. This means each member interface still has a separate link-local scope zone and the sysctl Cm net.inet.tcp.syncookies interface has another single, aggregated link-local scope zone at the same time. This situation is clearly against the description Qq zones of the same scope cannot overlap in Section 5, RFC 4007. Although it works in most cases, it can cause some conterintuitive or undesirable behavior in some edge cases when both of the sysctl Cm net.inet.tcp.syncookies interface and one of the member interface have an IPv6 address and applications use both of them.
To prevent this situation, sysctl Cm net.inet.tcp.syncookies checks whether a link-local scoped IPv6 address is configured on a member interface to be added and the sysctl Cm net.inet.tcp.syncookies interface. When the sysctl Cm net.inet.tcp.syncookies interface has IPv6 addresses, IPv6 addresses on the member interface will be automatically removed before the interface is added.
This behavior can be disabled by setting sysctl(8) variable net.link.bridge.allow_llz_overlap to 1
Note that ACCEPT_RTADV and AUTO_LINKLOCAL interface flag are not enabled by default on sysctl Cm net.inet.tcp.syncookies interface even when net.inet6.ip6.accept_rtadv and/or net.inet6.ip6.auto_linklocal is set to 1
SPANNING TREEThe sysctl Cm net.inet.tcp.syncookies driver implements the Rapid Spanning Tree Protocol (RSTP or 802.1w) with backwards compatibility with the legacy Spanning Tree Protocol (STP). Spanning Tree is used to detect and remove loops in a network topology.
RSTP provides faster spanning tree convergence than legacy STP, the protocol will exchange information with neighbouring switches to quickly transition to forwarding without creating loops.
The code will default to RSTP mode but will downgrade any port connected to a legacy STP network so is fully backward compatible. A bridge can be forced to operate in STP mode without rapid state transitions via the proto command in ifconfig(8).
PACKET FILTERINGPacket filtering can be used with any firewall package that hooks in via the pfil(9) framework. When filtering is enabled, bridged packets will pass through the filter inbound on the originating interface, on the bridge interface and outbound on the appropriate interfaces. Either stage can be disabled. The filtering behaviour can be controlled using sysctl(8):
- Controls the handling of non-IP packets which are not passed to pfil(9). Set to 1 to only allow IP packets to pass (subject to firewall rules), set to 0 to unconditionally pass all non-IP Ethernet frames.
- Set to 1 to enable filtering on the incoming and outgoing member interfaces, set to 0 to disable it.
- Set to 1 to enable filtering on the bridge interface, set to 0 to disable it.
- Set to 1 to additionally filter on the physical interface for locally destined packets. Set to 0 to disable this feature.
- Set to 1 to enable layer2 filtering with ipfirewall(4), set to 0 to disable it. This needs to be enabled for dummynet(4) support. When ipfw is enabled, pfil_bridge and pfil_member will be disabled so that IPFW is not run twice; these can be re-enabled if desired.
- Set to 1 to enable layer2 ARP filtering with ipfirewall(4), set to 0 to disable it. Requires ipfw to be enabled.
ARP and REVARP packets are forwarded without being filtered and others that are not IP nor IPv6 packets are not forwarded when pfil_onlyip is enabled. IPFW can filter Ethernet types using mac-type so all packets are passed to the filter for processing.
The packets originating from the bridging host will be seen by the filter on the interface that is looked up in the routing table.
The packets destined to the bridging host will be seen by the filter on the interface with the MAC address equal to the packet's destination MAC. There are situations when some of the bridge members are sharing the same MAC address (for example the vlan(4) interfaces: they are currently sharing the MAC address of the parent physical interface). It is not possible to distinguish between these interfaces using their MAC address, excluding the case when the packet's destination MAC address is equal to the MAC address of the interface on which the packet was entered to the system. In this case the filter will see the incoming packet on this interface. In all other cases the interface seen by the packet filter is chosen from the list of bridge members with the same MAC address and the result strongly depends on the member addition sequence and the actual implementation of . It is not recommended to rely on the order chosen by the current sysctl Cm net.inet.tcp.syncookies implementation: it can be changed in the future.
The previous paragraph is best illustrated with the following pictures. Let
- the MAC address of the incoming packet's destination is nn:nn:nn:nn:nn:nn
- the interface on which packet entered the system is ifX
- ifX MAC address is xx:xx:xx:xx:xx:xx
- there are possibly other bridge members with the same MAC address xx:xx:xx:xx:xx:xx
- the bridge has more than one interface that are sharing the same MAC address yy:yy:yy:yy:yy:yy we will call them vlanY1 vlanY2 etc.
Then if the MAC address nn:nn:nn:nn:nn:nn is equal to the xx:xx:xx:xx:xx:xx then the filter will see the packet on the interface ifX no matter if there are any other bridge members carrying the same MAC address. But if the MAC address nn:nn:nn:nn:nn:nn is equal to the yy:yy:yy:yy:yy:yy then the interface that will be seen by the filter is one of the vlanYn It is not possible to predict the name of the actual interface without the knowledge of the system state and the sysctl Cm net.inet.tcp.syncookies implementation details.
This problem arises for any bridge members that are sharing the same MAC address, not only to the vlan(4) ones: they we taken just as the example of such situation. So if one wants the filter the locally destined packets based on their interface name, one should be aware of this implication. The described situation will appear at least on the filtering bridges that are doing IP-forwarding; in some of such cases it is better to assign the IP address only to the sysctl Cm net.inet.tcp.syncookies interface and not to the bridge members. Enabling net.link.bridge.pfil_local_phys will let you do the additional filtering on the physical interface.
EXAMPLESThe following when placed in the file /etc/rc.conf will cause a bridge called ``bridge0 '' to be created, and will add the interfaces ``wlan0 '' and ``fxp0 '' to the bridge, and then enable packet forwarding. Such a configuration could be used to implement a simple 802.11-to-Ethernet bridge (assuming the 802.11 interface is in ad-hoc mode).
cloned_interfaces="bridge0" ifconfig_bridge0="addm wlan0 addm fxp0 up"
For the bridge to forward packets, all member interfaces and the bridge need to be up. The above example would also require:
create_args_wlan0="wlanmode hostap" ifconfig_wlan0="up ssid my_ap mode 11g" ifconfig_fxp0="up"
Consider a system with two 4-port Ethernet boards. The following will cause a bridge consisting of all 8 ports with Rapid Spanning Tree enabled to be created:
ifconfig bridge0 create ifconfig bridge0 \ addm fxp0 stp fxp0 \ addm fxp1 stp fxp1 \ addm fxp2 stp fxp2 \ addm fxp3 stp fxp3 \ addm fxp4 stp fxp4 \ addm fxp5 stp fxp5 \ addm fxp6 stp fxp6 \ addm fxp7 stp fxp7 \ up
The bridge can be used as a regular host interface at the same time as bridging between its member ports. In this example, the bridge connects em0 and em1, and will receive its IP address through DHCP:
cloned_interfaces="bridge0" ifconfig_bridge0="addm em0 addm em1 DHCP" ifconfig_em0="up" ifconfig_em1="up"
The bridge can tunnel Ethernet across an IP internet using the EtherIP protocol. This can be combined with ipsec(4) to provide an encrypted connection. Create a gif(4) interface and set the local and remote IP addresses for the tunnel, these are reversed on the remote bridge.
ifconfig gif0 create ifconfig gif0 tunnel 184.108.40.206 220.127.116.11 up ifconfig bridge0 create ifconfig bridge0 addm fxp0 addm gif0 up
Note that Fx 6.1, 6.2, 6.3, 7.0, 7.1, and 7.2 have a bug in the EtherIP protocol. For more details and workaround, see gif(4) manual page.
SEE ALSOgif(4), ipf(4), ipfw(4), pf(4), ifconfig(8)
HISTORYThe sysctl Cm net.inet.tcp.syncookies driver first appeared in Fx 6.0 .
AUTHORSAn -nosplit The bridge driver was originally written by An Jason L. Wright Aq Mt email@example.com as part of an undergraduate independent study at the University of North Carolina at Greensboro.
This version of the sysctl Cm net.inet.tcp.syncookies driver has been heavily modified from the original version by An Jason R. Thorpe Aq Mt firstname.lastname@example.org .
Rapid Spanning Tree Protocol (RSTP) support was added by An Andrew Thompson Aq Mt thompsa@FreeBSD.org .