ebtables.8

.TH EBTABLES 8  "03 December 2002"
.\"
.\" Man page written by Bart De Schuymer <bdschuym@pandora.be>
.\" It is based on the iptables man page.
.\"
.\" Iptables page by Herve Eychenne March 2000.
.\"
.\"     This program is free software; you can redistribute it and/or modify
.\"     it under the terms of the GNU General Public License as published by
.\"     the Free Software Foundation; either version 2 of the License, or
.\"     (at your option) any later version.
.\"
.\"     This program is distributed in the hope that it will be useful,
.\"     but WITHOUT ANY WARRANTY; without even the implied warranty of
.\"     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
.\"     GNU General Public License for more details.
.\"
.\"     You should have received a copy of the GNU General Public License
.\"     along with this program; if not, write to the Free Software
.\"     Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
.\"     
.\"
.SH NAME
ebtables (v.2.0) \- Ethernet bridge frame table administration
.SH SYNOPSIS
.BR "ebtables -[ADI] " "chain rule-specification " [ options ]
.br
.BR "ebtables -P " "chain target"
.br
.BR "ebtables -[FLZ] [" "chain" "]"
.br
.BR "ebtables -[NX] " chain
.br
.BR "ebtables -E " "old-chain-name new-chain-name"
.br
.BR "ebtables --init-table"
.br
.BR "ebtables --atomic-init "
.br
.BR "ebtables --atomic-save "
.br
.BR "ebtables --atomic-commit "
.br
.SH DESCRIPTION
.B ebtables
is used to set up, maintain, and inspect the tables of Ethernet frame
rules in the Linux kernel. It works analogous as iptables, but is less
complicated. This man page is written with the man page of iptables
next to it, so don't be surprised to see copied sentences and structure.

There are three tables with built-in chains. Each chain is a list
of rules which can match frames: each rule specifies what to do with a
frame which matches. This is called a 'target'. The tables are used to
divide functionality into different sets of chains.

.SS TARGETS
A firewall rule specifies criteria for a frame, and a target. If the
frame does not match, the next rule in the chain is the examined one; if
it does match, then the next thing to do is specified by the target.
This target can be one of these values:
.IR ACCEPT ,
.IR DROP ,
.IR CONTINUE ,
.IR RETURN ,
an extention.
.PP
.I ACCEPT
means to let the frame through.
.I DROP
means the frame has to be dropped.
.I CONTINUE
means the next rule has to be checked. This can be handy to know how many
frames pass a certain point in the chain or to log those frames.
.I RETURN
means stop traversing this chain and resume at the next rule in the
previous (calling) chain.
For the
other targets see the
.B "TARGET EXTENSIONS"
section.
.SS TABLES
There are three tables.
.TP
.B "-t, --table"
This option specifies the frame matching table which the command should
operate on. If specified it should be the first option. The tables are: 
.BR filter ,
this is the default table and contains three chains: 
.B INPUT 
(for frames destined for the bridge itself), 
.B OUTPUT 
(for locally-generated frames) and
.B FORWARD 
(for frames being bridged).
.BR nat ,
this table is used to change the mac addresses and contains three chains: 
.B PREROUTING 
(for altering frames as soon as they come in), 
.B OUTPUT 
(for altering locally generated frames before they are bridged) and 
.B POSTROUTING
(for altering frames as they are about to go out). A small note on the naming
of chains POSTROUTING and PREROUTING: it would be more accurate to call them
PREFORWARDING and POSTFORWARDING, but for all those who come from the
.BR iptables " world to " ebtables
it is easier to have the same names.
.BR broute ,
this table is used to make a brouter, it has one chain:
.BR BROUTING .
The targets
.BR DROP " and " ACCEPT
have special meaning in this table.
.B DROP
actually means the frame has to be routed, while
.B ACCEPT
means the frame has to be bridged. The
.B BROUTING
chain is traversed very early. It is only traversed by frames entering on
a bridge enslaved nic that is in forwarding state. Normally those frames
would be bridged, but you can decide otherwise here. The
.B redirect
target is very handy here.
.SH OPTIONS
The options can be divided into several different groups.
.SS COMMANDS
These options specify the specific actions to perform; only one of them
can be specified on the command line (the
.B -Z
command is an exception). All these options only apply to the selected
(or default) table.
.TP
.B "-A, --append"
Append a rule to the end of the selected chain.
.TP
.B "-D, --delete"
Delete the specified rule from the selected chain. There are two ways to
use this command. The first is by specifying an interval of rule numbers
to delete, syntax: start_nr[:end_nr]. The second usage is by specifying
the complete rule as it would have been specified when it was added.
.TP
.B "-I, --insert"
Insert the specified rule into the selected chain at the specified rule number (1 meaning
the head of the chain).
.TP
.B "-L, --list"
List all rules in the selected chain. If no chain is selected, all chains
are listed.
.br
The following three options change the output:
.br
.B "--Ln"
.br
Puts rule numbers in front of every rule.
.br
.B "--Lc"
.br
Shows the counters at the end of every rule, there is a frame counter
(pcnt) and a byte counter (bcnt).
.br
.B "--Lx"
.br
The output is directly usable as executable commands in a script, to be
run f.e. at bootup. This option is incompatible with the previous two
options. When no chain name was specified for the
.B "-L"
command, all necessary commands for making the user defined chains and
renaming the standard chains will be made.
.TP
.B "-F, --flush"
Flush the selected chain. If no chain is selected, every chain will be
flushed. This does not change the policy of the chain.
.TP
.B "--init-table"
Replace the current table data by the initial table data.
.TP
.B "-Z, --zero"
Put the counters of the selected chain on zero. If no chain is selected, all the counters
are put on zero. This can be used in conjunction with the -L command (see above). 
This will cause the rule counters to be printed on the screen before they are put on zero.
.TP
.B "-P, --policy"
Set the policy for the chain to the given target. The policy is either
.B ACCEPT
, either
.BR DROP .
.TP
.B "-N, --new-chain"
Create a new user-defined chain by the given name. The number of
user-defined chains is unlimited. A chain name has max length of 31.
.TP
.B "-X, --delete-chain"
Delete the specified user-defined chain. There must be no references to the
chain,
.B ebtables
will complain if there are.
.TP
.B "-E, --rename-chain"
Rename the specified chain to the new name. This has no effect on the
structure of the table. It is also allowed to rename a base chain, f.e.
if you like PREBRIDGING more than PREROUTING. Be sure to talk about the
standard chain names when you would ask a question on a mailing list.
.TP
.B "--atomic-init"
Copy the kernel's initial data of the table to the specified
file. This can be used as the first action, after which rules are added
to the file. The file can be specified using the
.B --atomic-file
option or through the
.IR EBTABLES_ATOMIC_FILE " environment variable."
.TP
.B "--atomic-save"
Copy the kernel's current data of the table to the specified
file. This can be used as the first action, after which rules are added
to the file. The file can be specified using the
.B --atomic-file
option or through the
.IR EBTABLES_ATOMIC_FILE " environment variable."
.TP
.B "--atomic-commit"
Replace the kernel table data with the data contained in the specified
file. This is a useful command that allows you to put all your rules of a
certain table into the kernel at once, saving the kernel a lot of precious
time and allowing atomic updates of the tables. The file which contains
the table data is constructed by using either the
.B "--atomic-init"
or the
.B "--atomic-save"
command to get a starting file. After that, using the
.B "--atomic-file"
option when constructing rules or setting the
.IR EBTABLES_ATOMIC_FILE " environment variable"
allows you to extend the file and build the complete table before
commiting it to the kernel.
.SS
PARAMETERS
The following parameters make up a rule specification (as used in the add
and delete commands). A "!" argument before the specification inverts the
test for that specification. Apart from these standard parameters, there are others, see
.BR "MATCH EXTENSIONS" .
.TP
.BR "-p, --protocol " "[!] \fIprotocol\fP"
The protocol that was responsible for creating the frame. This can be a
hexadecimal number, above 
.IR 0x0600 ,
a name (e.g.
.I ARP
) or
.BR LENGTH .
The protocol field of the Ethernet frame can be used to denote the
length of the header (802.2/802.3 networks). When the value of that field is
below (or equals)
.IR 0x0600 ,
the value equals the size of the header and shouldn't be used as a
protocol number. Instead, all frames where the protocol field is used as
the length field are assumed to be of the same 'protocol'. The protocol
name used in
.B ebtables
for these frames is
.BR LENGTH .
.br
The file
.B /etc/ethertypes
can be used to show readable
characters instead of hexadecimal numbers for the protocols. For example,
.I 0x0800
will be represented by 
.IR IPV4 .
The use of this file is not case sensitive. 
See that file for more information. The flag 
.B --proto
is an alias for this option.
.TP 
.BR "-i, --in-interface " "[!] \fIname\fP"
The interface via which a frame is received (for the
.BR INPUT ,
.BR FORWARD ,
.BR PREROUTING " and " BROUTING
chains). The flag
.B --in-if
is an alias for this option.
.TP
.BR "--logical-in " "[!] \fIname\fP"
The (logical) bridge interface via which a frame is received (for the
.BR INPUT ,
.BR FORWARD ,
.BR PREROUTING " and " BROUTING
chains).
.TP
.BR "-o, --out-interface " "[!] \fIname\fP"
The interface via which a frame is going to be sent (for the
.BR OUTPUT ,
.B FORWARD
and
.B POSTROUTING
chains). The flag
.B --out-if
is an alias for this option.
.TP
.BR "--logical-out " "[!] \fIname\fP"
The (logical) bridge interface via which a frame is going to be sent (for
the
.BR OUTPUT ,
.B FORWARD
and
.B POSTROUTING
chains).
.TP
.BR "-s, --source " "[!] \fIaddress\fP[/\fImask\fP]"
The source mac address. Both mask and address are written as 6 hexadecimal
numbers seperated by colons. Alternatively one can specify Unicast,
Multicast or Broadcast.
.br
Unicast=00:00:00:00:00:00/01:00:00:00:00:00,
Multicast=01:00:00:00:00:00/01:00:00:00:00:00 and
Broadcast=ff:ff:ff:ff:ff:ff/ff:ff:ff:ff:ff:ff. Note that a broadcast
address will also match the multicast specification. The flag
.B --src
is an alias for this option.
.TP
.BR "-d, --destination " "[!] \fIaddress\fP[/\fImask\fP]"
The destination mac address. See -s (above) for more details. The flag
.B --dst
is an alias for this option.

.SS OTHER OPTIONS
.TP
.B "-V, --version"
Show the version of the userprogram.
.TP
.B "-h, --help"
Give a brief description of the command syntax. Here you can also specify
names of extensions and
.B ebtables
will try to write help about those extensions. E.g. ebtables -h snat log ip arp.
.TP
.BR "-j, --jump " "\fItarget\fP"
The target of the rule. This is one of the following values:
.BR ACCEPT ,
.BR DROP ,
.BR CONTINUE ,
.BR RETURN ,
a target extension (see
.BR "TARGET EXTENSIONS" ")"
or a user defined chain name.
.TP
.B --atomic-file file
Let the command operate on the specified file. The data of the table to
operate on will be extracted from the file and the result of the operation
will be saved back into the file. If specified, this option should come
before the command specification. An alternative that should be preferred,
is setting the
.BR EBTABLES_ATOMIC_FILE "environment variable."
.TP
.B -M, --modprobe program
When talking to the kernel, use this program to try to automatically load
missing kernel modules.
.SH MATCH EXTENSIONS
.B ebtables
extensions are precompiled into the userspace tool. So there is no need
to explicitly load them with a -m option like in iptables. However, these
extensions deal with functionality supported by supplemental kernel modules.
.SS ip
Specify ip specific fields. These will only work if the protocol equals
.BR IPv4 .
.TP
.BR "--ip-source " "[!] \fIaddress\fP[/\fImask\fP]"
The source ip address.
The flag
.B --ip-src
is an alias for this option.
.TP
.BR "--ip-destination " "[!] \fIaddress\fP[/\fImask\fP]"
The destination ip address.
The flag
.B --ip-dst
is an alias for this option.
.TP
.BR "--ip-tos " "[!] \fItos\fP"
The ip type of service, in hexadecimal numbers.
.BR IPv4 .
.TP
.BR "--ip-protocol " "[!] \fIprotocol\fP"
The ip protocol.
The flag
.B --ip-proto
is an alias for this option.
.TP
.BR "--ip-source-port " "[!] \fIport\fP[:\fIport\fP]"
The source port or port range for the ip protocols 6 (TCP) and 17
(UDP). If the first port is omitted, "0" is assumed; if the last
is omitted, "65535" is assumed. The flag
.B --ip-sport
is an alias for this option.
.TP
.BR "--ip-destination-port " "[!] \fIport\fP[:\fIport\fP]"
The destination port or port range for ip protocols 6 (TCP) and
17 (UDP). The flag
.B --ip-dport
is an alias for this option.
.SS arp
Specify arp specific fields. These will only work if the protocol equals
.BR ARP " or " RARP .
.TP
.BR "--arp-opcode " "[!] \fIopcode\fP"
The (r)arp opcode (decimal or a string, for more details see
.BR "ebtables -h arp" ).
.TP
.BR "--arp-htype " "[!] \fIhardware type\fP"
The hardware type, this can be a decimal or the string "Ethernet". This
is normally Ethernet (value 1).
.TP
.BR "--arp-ptype " "[!] \fIprotocol type\fP"
The protocol type for which the (r)arp is used (hexadecimal or the string "IPv4").
This is normally IPv4 (0x0800). 
.TP
.BR "--arp-ip-src " "[!] \fIaddress\fP[/\fImask\fP]"
The ARP IP source address specification.
.TP
.BR "--arp-ip-dst " "[!] \fIaddress\fP[/\fImask\fP]"
The ARP IP destination address specification.
.SS vlan
Specify 802.1Q Tag Control Information fields. These will only work if the protocol equals
.BR 802_1Q.
Also see extension help by 
.BR "ebtables -h vlan" .
.TP
.BR "--vlan-id " "[!] \fIid\fP"
The VLAN identifier field, VID (decimal number from 0 to 4094).
.TP
.BR "--vlan-prio " "[!] \fIprio\fP"
The user_priority field, this can be a decimal number from 0 to 7. 
Required VID to be 0 (null VID) or not specified vlan-id parameter (in this case VID deliberately be set to 0).
.TP
.BR "--vlan-encap " "[!] \fItype\fP"
The encapsulated Ethernet frame type/length, this can be a hexadecimal
number from 0x0000 to 0xFFFF.
Usually it's 0x0800 (IPv4). See also 
.B /etc/ethertypes 
file.
.SS mark_m
.TP
.BR "--mark " "[!] [\fIvalue\fP][/\fImask\fP]"
Matches frames with the given unsigned mark value. If a mark value and 
mask is specified, the logical AND of the mark value of the frame and
the user specified mask is taken before comparing with the user specified
mark value. If only a mask is specified (start with '/') the logical AND
of the mark value of the frame and the user specified mark is taken and
the result is compared with zero.

.SH WATCHER EXTENSION(S)
Watchers are things that only look at frames passing by. These watchers only
see the frame if the frame passes all the matches of the rule.
.SS log
The fact that the log module is a watcher lets us log stuff while giving a target
by choice. Note that the log module therefore is not a target.
.TP
.B "--log"
.br
Use this if you won't specify any other log options, so if you want to use the default
settings: log-prefix="", no arp logging, no ip logging, log-level=info.
.TP
.B --log-level "\fIlevel\fP"
.br
defines the logging level. For the possible values: ebtables -h log.
The default level is 
.IR info .
.TP
.BR --log-prefix " \fItext\fP"
.br
defines the prefix to be printed before the logging information.
.TP
.B --log-ip 
.br
will log the ip information when a frame made by the ip protocol matches 
the rule. The default is no ip information logging.
.TP
.B --log-arp
.br
will log the (r)arp information when a frame made by the (r)arp protocols
matches the rule. The default is no (r)arp information logging.
.SS TARGET EXTENSIONS
.TP
.B snat
The
.B snat
target can only be used in the
.BR POSTROUTING " chain of the " nat " table."
It specifies that the source mac address has to be changed.
.br
.BR "--to-source " "\fIaddress\fP"
.br
The flag
.B --to-src
is an alias for this option.
.br
.BR "--snat-target " "\fItarget\fP"
.br
Specifies the standard target. After doing the snat, the rule still has 
to give a standard target so
.B ebtables
knows what to do.
The default target is ACCEPT. Making it CONTINUE could let you use
multiple target extensions on the same frame. Making it DROP doesn't
make sense, but you could do that too. RETURN is also allowed. Note
that using RETURN in a base chain is not allowed.
.TP
.B dnat
The
.B dnat
target can only be used in the
.BR BROUTING " chain of the " broute " table and the "
.BR PREROUTING " and " OUTPUT " chains of the " nat " table."
It specifies that the destination mac address has to be changed.
.br
.BR "--to-destination " "\fIaddress\fP"
.br
The flag
.B --to-dst
is an alias for this option.
.br
.BR "--dnat-target " "\fItarget\fP"
.br
Specifies the standard target. After doing the dnat, the rule still has to
give a standard target so
.B ebtables
knows what to do.
The default target is ACCEPT. Making it CONTINUE could let you use 
multiple target extensions on the same frame. Making it DROP only makes
sense in the BROUTING chain but using the redirect target is more logical
there. RETURN is also allowed. Note
that using RETURN in a base chain is not allowed.
.TP
.B redirect
The
.B redirect
target will change the MAC target address to that of the bridge device the
frame arrived on. This target can only be used in the
.BR BROUTING " chain of the " broute " table and the "
.BR PREROUTING " chain of the " nat " table."
.br
.BR "--redirect-target " "\fItarget\fP"
.br
Specifies the standard target. After doing the MAC redirect, the rule
still has to give a standard target so
.B ebtables
knows what to do.
The default target is ACCEPT. Making it CONTINUE could let you use 
multiple target extensions on the same frame. Making it DROP in the
BROUTING chain will let the frames be routed. RETURN is also allowed. Note
that using RETURN in a base chain is not allowed.
.TP
.B mark
The mark target can be used in every chain of every table. It is possible
to use the marking of a frame/packet in both ebtables and iptables, 
if the br-nf code is compiled into the kernel. Both put the marking at the
same place. So, you can consider this fact as a feature, or as something to
watch out for.
.br
.BR "--mark-target " "\fItarget\fP"
.br
Specifies the standard target. After marking the frame, the rule
still has to give a standard target so
.B ebtables
knows what to do.
The default target is ACCEPT. Making it CONTINUE can let you do other
things with the frame in other rules of the chain.
.br
.BR "--set-mark " "\fIvalue\fP"
.br
Mark the frame with the specified unsigned value.
.br
.SH FILES
.I /etc/ethertypes
.SH ENVIRONMENT VARIABLES
.I EBTABLES_ATOMIC_FILE
.SH BUGS
This won't work on an architecture with a user32/kernel64 situation like the Sparc64.
.SH AUTHOR
.IR "" "Bart De Schuymer <" bdschuym@pandora.be >
.SH SEE ALSO
.BR iptables "(8), " brctl (8)