path: root/usr.sbin/bgpd/bgpd.conf.5

.\" $OpenBSD: bgpd.conf.5,v 1.94 2009/06/07 00:31:22 claudio Exp $
.\"
.\" Copyright (c) 2004 Claudio Jeker <claudio@openbsd.org>
.\" Copyright (c) 2003, 2004 Henning Brauer <henning@openbsd.org>
.\" Copyright (c) 2002 Daniel Hartmeier <dhartmei@openbsd.org>
.\"
.\" Permission to use, copy, modify, and distribute this software for any
.\" purpose with or without fee is hereby granted, provided that the above
.\" copyright notice and this permission notice appear in all copies.
.\"
.\" THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
.\" WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
.\" MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
.\" ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
.\" WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
.\" ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
.\" OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
.\"
.Dd $Mdocdate: June 7 2009 $
.Dt BGPD.CONF 5
.Os
.Sh NAME
.Nm bgpd.conf
.Nd Border Gateway Protocol daemon configuration file
.Sh DESCRIPTION
The
.Xr bgpd 8
daemon implements the Border Gateway Protocol version 4 as described
in RFC 1771.
.Sh SECTIONS
The
.Nm
config file is divided into four main sections.
.Bl -tag -width xxxx
.It Sy Macros
User-defined variables may be defined and used later, simplifying the
configuration file.
.It Sy Global Configuration
Global settings for
.Xr bgpd 8 .
.It Sy Neighbors and Groups
.Xr bgpd 8
establishes sessions with
.Em neighbors .
The neighbor definition and properties are set in this section, as well as
grouping neighbors for the ease of configuration.
.It Sy Filter
Filter rules for incoming and outgoing
.Em UPDATES .
.El
.Pp
With the exception of macros,
the sections should be grouped and appear in
.Nm
in the order shown above.
.Pp
Comments can be put anywhere in the file using a hash mark
.Pq Sq # ,
and extend to the end of the current line.
.Pp
Additional configuration files can be included with the
.Ic include
keyword, for example:
.Bd -literal -offset indent
include "/etc/bgpd/bgpd-10.0.0.1.filter"
.Ed
.Sh MACROS
Macros can be defined that will later be expanded in context.
Macro names must start with a letter, and may contain letters, digits
and underscores.
Macro names may not be reserved words (for example,
.Ic AS ,
.Ic neighbor ,
or
.Ic group ) .
Macros are not expanded inside quotes.
.Pp
For example:
.Bd -literal -offset indent
peer1="1.2.3.4"
neighbor $peer1 {
	remote-as 65001
}
.Ed
.Sh GLOBAL CONFIGURATION
There are quite a few settings that affect the operation of the
.Xr bgpd 8
daemon globally.
.Pp
.Bl -tag -width Ds -compact
.It Ic AS Ar as-number Op Ar as-number
Set the local
.Em autonomous system
number to
.Ar as-number .
If the first AS number is a 4-byte AS it is possible to specifiy a secondary
2-byte AS number which is used for neighbors which do not support 4-byte AS
numbers.
The default for the secondary AS is 23456.
.Pp
The AS numbers are assigned by local RIRs, such as:
.Pp
.Bl -tag -width xxxxxxxx -compact
.It AfriNIC
for Africa
.It APNIC
for Asia Pacific
.It ARIN
for North America and parts of the Caribbean
.It LACNIC
for Latin America and the Caribbean
.It RIPE NCC
for Europe, the Middle East, and parts of Asia
.El
.Pp
For example:
.Bd -literal -offset indent
AS 65001
.Ed
.Pp
sets the local AS to 65001.
.Pp
The AS numbers 64512 \(en 65534 are designated for private use.
The AS number 23456 is a specially designated Autonomous System Number and
should not be used.
4-byte AS numbers are specified as two numbers separated by a dot
(ASDOT format),
for example:
.Bd -literal -offset indent
AS 3.10
.Ed
.Pp
or as a large number (ASPLAIN format), for example:
.Bd -literal -offset indent
AS 196618
.Ed
.Pp
.It Ic connect-retry Ar seconds
Set the number of seconds before retrying to open a connection.
This timer should be sufficiently large in EBGP configurations.
The default is 120 seconds.
.Pp
.It Xo
.Ic dump
.Op Ic rib Ar name
.Pq Ic table Ns \&| Ns Ic table-mp
.Ar file Op Ar timeout
.Xc
.It Xo
.Ic dump
.Pq Ic all Ns \&| Ns Ic updates
.Pq Ic in Ns \&| Ns Ic out
.Ar file Op Ar timeout
.Xc
Dump the RIB, a.k.a. the
.Em routing information base ,
and all BGP messages in Multi-threaded Routing Toolkit (MRT) format.
Dumping the RIB is normally an expensive operation,
but it should not influence the session handling.
It is possible to dump alternate RIB with the use of
.Ar name .
.Pp
For example, the following will dump the entire table to the
.Xr strftime 3 Ns -expanded
filename.
The
.Ic table-mp
format is multi-protocol capable but often not supported by 3rd-party tools.
The timeout is optional:
.Bd -literal -offset indent
dump table "/tmp/rib-dump-%H%M" 300
.Ed
.Pp
Similar to the table dump, but this time all
BGP messages and
.Em state transitions
will be dumped to the specified file:
.Bd -literal -offset indent
dump all in "/tmp/all-in-%H%M" 300
.Ed
.Pp
As before, but only the
.Em UPDATE
messages will be dumped to the file:
.Bd -literal -offset indent
dump updates in "/tmp/updates-in-%H%M" 300
.Ed
.Pp
It is also possible to dump outgoing messages:
.Bd -literal -offset indent
dump all out "/tmp/all-out-%H%M" 300
# or
dump updates out "/tmp/updates-out-%H%M" 300
.Ed
.Pp
.It Xo
.Ic fib-update
.Pq Ic yes Ns \&| Ns Ic no
.Xc
If set to
.Ic no ,
do not update the Forwarding Information Base, a.k.a. the kernel
routing table.
The default is
.Ic yes .
.Pp
.It Ic holdtime Ar seconds
Set the holdtime in seconds.
The holdtime is reset to its initial value every time either a
.Em KEEPALIVE
or an
.Em UPDATE
message is received from the neighbor.
If the holdtime expires the session is dropped.
The default is 90 seconds.
Neighboring systems negotiate the holdtime used when the connection is
established in the
.Em OPEN
messages.
Each neighbor announces its configured holdtime; the smaller one is
then agreed upon.
.Pp
.It Ic holdtime min Ar seconds
The minimal accepted holdtime in seconds.
This value must be greater than or equal to 3.
.Pp
.It Ic listen on Ar address
Specify the local IP address
.Xr bgpd 8
should listen on.
.Bd -literal -offset indent
listen on 127.0.0.1
.Ed
.Pp
.It Ic log updates
Log received and sent updates.
.Pp
.It Xo
.Ic network
.Ar address Ns Li / Ns Ar prefix
.Op Ic set ...\&
.Xc
.It Xo
.Ic network
.Pq Ic inet Ns \&| Ns Ic inet6
.Ic static Op Ic set ...\&
.Xc
.It Xo
.Ic network
.Pq Ic inet Ns \&| Ns Ic inet6
.Ic connected Op Ic set ...\&
.Xc
Announce the specified network as belonging to our AS.
If set to
.Ic connected ,
routes to directly attached networks will be announced.
If set to
.Ic static ,
all static routes will be announced.
.Bd -literal -offset indent
network 192.168.7.0/24
.Ed
.Pp
It is possible to set default
.Em AS path attributes
per
.Ic network
statement:
.Bd -literal -offset indent
network 192.168.7.0/24 set localpref 220
.Ed
.Pp
See also the
.Sx ATTRIBUTE SET
section.
.Pp
.It Xo
.Ic nexthop
.Ic qualify
.Ic via
.Pq Ic bgp Ns \&| Ns Ic default
.Xc
If set to
.Ic bgp ,
.Xr bgpd 8
may use BGP routes to verify nexthops.
If set to
.Ic default ,
bgpd may use the default route to verify nexthops.
By default bgpd will only use static routes or routes added by other routing
daemons like
.Xr ospfd 8 .
.Pp
.It Xo
.Ic rde
.Ic med
.Ic compare
.Pq Ic always Ns \&| Ns Ic strict
.Xc
If set to
.Ic always ,
the
.Em MED
attributes will always be compared.
The default is
.Ic strict ,
where the
.Em MED
is only compared between peers belonging to the same AS.
.Pp
.It Xo
.Ic rde
.Ic rib Ar name
.Op Ic no Ic evaluate
.Xc
Creat an additional RIB named
.Ar name .
It is possible to disable the decision process per RIB with the
.Ic no Ic evaluate
flag.
.Ic Adj-RIB-In
and
.Ic Loc-RIB
are created automaticaly and used as default.
.Pp
.It Xo
.Ic rde
.Ic route-age
.Pq Ic ignore Ns \&| Ns Ic evaluate
.Xc
If set to
.Ic evaluate ,
the best path selection will not only be based on the path attributes but
also on the age of the route, giving preference to the older, typically
more stable, route.
In this case the decision process is no longer deterministic.
The default is
.Ic ignore .
.Pp
.It Xo
.Ic route-collector
.Pq Ic yes Ns \&| Ns Ic no
.Xc
If set to
.Ic yes ,
the route selection process is turned off.
The default is
.Ic no .
.Pp
.It Ic router-id Ar address
Set the router ID to the given IP address, which must be local to the
machine.
.Bd -literal -offset indent
router-id 10.0.0.1
.Ed
.Pp
If not given, the BGP ID is determined as the biggest IP address assigned
to the local machine.
.Pp
.It Ic rtable Ar number
Work with the given kernel routing table
instead of the default table,
.Ar 0 .
Note that this table is used for nexthop verification as well.
Directly connected networks are always taken into account, even though
their routes live in table 0.
.Pp
.It Xo
.Ic transparent-as
.Pq Ic yes Ns \&| Ns Ic no
.Xc
If set to
.Ic yes ,
.Em AS paths
to EBGP neighbors are not prepended with their own AS.
The default is
.Ic no .
.El
.Sh NEIGHBORS AND GROUPS
.Xr bgpd 8
establishes TCP connections to other BGP speakers called
.Em neighbors .
Each neighbor is specified by a
.Ic neighbor
section, which allows properties to be set specifically for that neighbor:
.Bd -literal -offset indent
neighbor 10.0.0.2 {
	remote-as 65002
	descr "a neighbor"
}
.Ed
.Pp
Multiple neighbors can be grouped together by a
.Ic group
section.
Each
.Ic neighbor
section within the
.Ic group
section inherits all properties from its group:
.Bd -literal -offset indent
group "peering AS65002" {
	remote-as 65002
	neighbor 10.0.0.2 {
		descr "AS65002-p1"
	}
	neighbor 10.0.0.3 {
		descr "AS65002-p2"
	}
}
.Ed
.Pp
Instead of the neighbor's IP address, an address/netmask pair may be given:
.Bd -literal -offset indent
neighbor 10.0.0.0/8
.Ed
.Pp
In this case, the neighbor specification becomes a
.Em template ,
and if a neighbor connects from an IP address within the given network,
the template is
.Em cloned ,
inheriting everything from the template but the remote address, which is
replaced by the connecting neighbor's address.
With a template specification it is valid to omit
.Ic remote-as ;
.Xr bgpd 8
will then accept any AS the neighbor presents in the
.Em OPEN
message.
.Pp
There are several neighbor properties:
.Pp
.Bl -tag -width Ds -compact
.It Xo
.Ic announce
.Sm off
.Po Ic all \*(Ba none \*(Ba
.Ic self \*(Ba default-route Pc
.Sm on
.Xc
If set to
.Ic none ,
no
.Em UPDATE
messages will be sent to the neighbor.
If set to
.Ic default-route ,
only the default route will be announced to the neighbor.
If set to
.Ic all ,
all generated
.Em UPDATE
messages will be sent to the neighbor.
This is usually used for
.Em transit AS's
and
.Em IBGP
peers.
The default value
for
.Em EBGP
peers is
.Ic self ,
which limits the sent
.Em UPDATE
messages to announcements of the local AS.
The default for IBGP peers is
.Ic all .
.Pp
.It Xo
.Ic announce
.Pq Ic IPv4 Ns \&| Ns Ic IPv6
.Pq Ic none Ns \&| Ns Ic unicast
.Xc
For the given address family, control which subsequent address families
(at the moment, only
.Em none ,
which disables the announcement of that address family, and
.Em unicast
are supported) are announced during the capabilities negotiation.
Only routes for that address family and subsequent address family will be
announced and processed.
.Pp
.It Xo
.Ic announce capabilities
.Pq Ic yes Ns \&| Ns Ic no
.Xc
If set to
.Ic no ,
capability negotiation is disabled during the establishment of the session.
This can be helpful to connect to old or broken BGP implementations.
The default is
.Ic yes .
.Pp
.It Ic demote Ar group
Increase the
.Xr carp 4
demotion counter on the given interface group, usually
.Ar carp ,
when the session is not in state
.Em ESTABLISHED .
The demotion counter will be increased as soon as
.Xr bgpd 8
starts and decreased
60 seconds after the session went to state
.Em ESTABLISHED.
For neighbors added at runtime, the demotion counter is only increased after
the session has been
.Em ESTABLISHED
at least once before dropping.
.Pp
For more information on interface groups,
see the
.Ic group
keyword in
.Xr ifconfig 8 .
.Pp
.It Ic depend on Ar interface
The neighbor session will be kept in state
.Em IDLE
as long as
.Ar interface
reports no link.
For
.Xr carp 4
interfaces, no link means that the interface is currently
.Em backup .
This is primarily intended to be used with
.Xr carp 4
to reduce failover times.
.Pp
The state of the network interfaces on the system can be viewed
using the
.Cm show interfaces
command to
.Xr bgpctl 8 .
.Pp
.It Ic descr Ar description
Add a description.
The description is used when logging neighbor events, in status
reports, for specifying neighbors, etc., but has no further meaning to
.Xr bgpd 8 .
.Pp
.It Ic down
Do not start the session when bgpd comes up but stay in
.Em IDLE .
.Pp
.It Xo
.Ic dump
.Pq Ic all Ns \&| Ns Ic updates
.Pq Ic in Ns \&| Ns Ic out
.Ar file Op Ar timeout
.Xc
Do a peer specific MRT dump.
Peer specific dumps are limited to
.Ic all
and
.Ic updates .
See also the
.Ic dump
section in
.Sx GLOBAL CONFIGURATION .
.Pp
.It Xo
.Ic enforce neighbor-as
.Pq Ic yes Ns \&| Ns Ic no
.Xc
If set to
.Ic yes ,
.Em AS paths
whose
.Em leftmost AS
is not equal to the
.Em remote AS
of the neighbor are rejected and a
.Em NOTIFICATION
is sent back.
The default value for IBGP peers is
.Ic no
otherwise the default is
.Ic yes .
.Pp
.It Ic holdtime Ar seconds
Set the holdtime in seconds.
Inherited from the global configuration if not given.
.Pp
.It Ic holdtime min Ar seconds
Set the minimal acceptable holdtime.
Inherited from the global configuration if not given.
.Pp
.It Xo
.Ic ipsec
.Pq Ic ah Ns \&| Ns Ic esp
.Pq Ic in Ns \&| Ns Ic out
.Ic spi Ar spi-number authspec Op Ar encspec
.Xc
Enable IPsec with static keying.
There must be at least two
.Ic ipsec
statements per peer with manual keying, one per direction.
.Ar authspec
specifies the authentication algorithm and key.
It can be
.Bd -literal -offset indent
sha1 <key>
md5 <key>
.Ed
.Pp
.Ar encspec
specifies the encryption algorithm and key.
.Ic ah
does not support encryption.
With
.Ic esp ,
encryption is optional.
.Ar encspec
can be
.Bd -literal -offset indent
3des <key>
3des-cbc <key>
aes <key>
aes-128-cbc <key>
.Ed
.Pp
Keys must be given in hexadecimal format.
.Pp
.It Xo
.Ic ipsec
.Pq Ic ah Ns \&| Ns Ic esp
.Ic ike
.Xc
Enable IPsec with dynamic keying.
In this mode,
.Xr bgpd 8
sets up the flows, and a key management daemon such as
.Xr isakmpd 8
is responsible for managing the session keys.
With
.Xr isakmpd 8 ,
it is sufficient to copy the peer's public key, found in
.Pa /etc/isakmpd/local.pub ,
to the local machine.
It must be stored in a file
named after the peer's IP address and must be stored in
.Pa /etc/isakmpd/pubkeys/ipv4/ .
The local public key must be copied to the peer in the same way.
As
.Xr bgpd 8
manages the flows on its own, it is sufficient to restrict
.Xr isakmpd 8
to only take care of keying by specifying the flags
.Fl Ka .
This can be done in
.Xr rc.conf.local 8 .
After starting the
.Xr isakmpd 8
and
.Xr bgpd 8
daemons on both sides, the session should be established.
.Pp
.It Ic local-address Ar address
When
.Xr bgpd 8
initiates the TCP connection to the neighbor system, it normally does not
bind to a specific IP address.
If a
.Ic local-address
is given,
.Xr bgpd 8
binds to this address first.
.Pp
.It Xo
.Ic max-prefix Ar number
.Op Ic restart Ar number
.Xc
Terminate the session after
.Ar number
prefixes have been received
(no such limit is imposed by default).
If
.Ic restart
is specified, the session will be restarted after
.Ar number
minutes.
.Pp
.It Ic multihop Ar hops
Neighbors not in the same AS as the local
.Xr bgpd 8
normally have to be directly connected to the local machine.
If this is not the case, the
.Ic multihop
statement defines the maximum hops the neighbor may be away.
.Pp
.It Ic passive
Do not attempt to actively open a TCP connection to the neighbor system.
.Pp
.It Ic remote-as Ar as-number
Set the AS number of the remote system.
.Pp
.It rib .Ar name
Bind the neighbor to the specified RIB.
.Pp
.It Ic route-reflector Op Ar address
Act as an RFC 2796
.Em route-reflector
for this neighbor.
An optional cluster ID can be specified; otherwise the BGP ID will be used.
.Pp
.It Ic set Ar attribute ...
Set the
.Em AS path attributes
to some default per
.Ic neighbor
or
.Ic group
block:
.Bd -literal -offset indent
set localpref 300
.Ed
.Pp
See also the
.Sx ATTRIBUTE SET
section.
Set parameters are applied to the received prefixes; the only exceptions are
.Ic prepend-self ,
.Ic nexthop no-modify
and
.Ic nexthop self .
These sets are rewritten into filter rules and can be viewed with
.Dq bgpd -nv .
.Pp
.It Xo
.Ic softreconfig
.Pq Ic in Ns \&| Ns Ic out
.Pq Ic yes Ns \&| Ns Ic no
.Xc
Turn soft reconfiguration on or off for the specified direction.
If soft reconfiguration is turned on, filter changes will be applied on
configuration reloads.
If turned off, a BGP session needs to be cleared to apply the filter changes.
Enabling
.Ic softreconfig in
will raise the memory requirements of
.Xr bgpd 8
because the unmodified
.Em AS path attributes
need to be stored as well.
The default is
.Ic yes .
.Pp
.It Ic tcp md5sig password Ar secret
.It Ic tcp md5sig key Ar secret
Enable TCP MD5 signatures per RFC 2385.
The shared secret can either be given as a password or hexadecimal key.
.Bd -literal -offset indent
tcp md5sig password mekmidasdigoat
tcp md5sig key deadbeef
.Ed
.Pp
.It Xo
.Ic transparent-as
.Pq Ic yes Ns \&| Ns Ic no
.Xc
If set to
.Ic yes ,
.Em AS paths
to EBGP neighbors are not prepended with their own AS.
The default is inherited from the global
.Ic transparent-as
setting.
.Pp
.It Xo
.Ic ttl-security
.Pq Ic yes Ns \&| Ns Ic no
.Xc
Enable or disable ttl-security.
When enabled,
outgoing packets are sent using a TTL of 255
and a check is made against an incoming packet's TTL.
For directly connected peers,
incoming packets are required to have a TTL of 255,
ensuring they have not been routed.
For multihop peers,
incoming packets are required to have a TTL of 256 minus multihop distance,
ensuring they have not passed through more than the expected number of hops.
The default is
.Ic no .
.El
.Sh FILTER
.Xr bgpd 8
has the ability to
.Ic allow
and
.Ic deny
.Em UPDATES
based on
.Em prefix
or
.Em AS path attributes .
In addition,
.Em UPDATES
may also be modified by filter rules.
.Pp
For each
.Em UPDATE
processed by the filter, the filter rules are evaluated in sequential order,
from first to last.
The last matching
.Ic allow
or
.Ic deny
rule decides what action is taken.
.Pp
The following actions can be used in the filter:
.Bl -tag -width xxxxxxxx
.It Ic allow
The
.Em UPDATE
is passed.
.It Ic deny
The
.Em UPDATE
is blocked.
.It Ic match
Apply the filter attribute set without influencing the filter decision.
.El
.Sh PARAMETERS
The rule parameters specify the
.Em UPDATES
to which a rule applies.
An
.Em UPDATE
always comes from, or goes to, one neighbor.
Most parameters are optional, but each can appear at most once per rule.
If a parameter is specified, the rule only applies to packets with
matching attributes.
.Pp
.Bl -tag -width Ds -compact
.It Ar as-type as-number
This rule applies only to
.Em UPDATES
where the
.Em AS path
matches.
The
.Ar as-number
is matched against a part of the
.Em AS path
specified by the
.Ar as-type .
.Ar as-type
is one of the following operators:
.Pp
.Bl -tag -width transmit-as -compact
.It Ic AS
(any part)
.It Ic peer-as
(leftmost AS number)
.It Ic source-as
(rightmost AS number)
.It Ic transit-as
(all but the rightmost AS number)
.El
.Pp
Multiple
.Ar as-number
entries for a given type or
.Ar as-type as-number
entries may also be specified,
separated by commas or whitespace,
if enclosed in curly brackets:
.Bd -literal -offset indent
deny from any AS { 1, 2, 3 }
deny from any { AS 1, source-as 2, transit-as 3 }
deny from any { AS { 1, 2, 3 }, source-as 4, transit-as 5 }
.Ed
.Pp
.It Xo
.Ic community
.Ar as-number Ns Li \&: Ns Ar local
.Xc
.It Ic community Ar name
This rule applies only to
.Em UPDATES
where the
.Ic community
path attribute is present and matches.
Communities are specified as
.Ar as-number Ns Li : Ns Ar local ,
where
.Ar as-number
is an AS number and
.Ar local
is a locally significant number between zero and
.Li 65535 .
Both
.Ar as-number
and
.Ar local
may be set to
.Sq *
to do wildcard matching.
Alternatively, well-known communities may be given by name instead and
include
.Ic NO_EXPORT ,
.Ic NO_ADVERTISE ,
.Ic NO_EXPORT_SUBCONFED ,
and
.Ic NO_PEER .
Both
.Ar as-number
and
.Ar local
may be set to
.Ic neighbor-as ,
which is expanded to the current neighbor remote AS number.
.Pp
.It Xo
.Pq Ic from Ns \&| Ns Ic to
.Ar peer
.Xc
This rule applies only to
.Em UPDATES
coming from, or going to, this particular neighbor.
This parameter must be specified.
.Ar peer
is one of the following:
.Pp
.Bl -tag -width "group descr" -compact
.It Ic any
Any neighbor will be matched.
.It Ar address
Neighbors with this address will be matched.
.It Ic group Ar descr
Neighbors in this group will be matched.
.El
.Pp
Multiple
.Ar peer
entries may also be specified,
separated by commas or whitespace,
if enclosed in curly brackets:
.Bd -literal -offset indent
deny from { 128.251.16.1, 251.128.16.2, group hojo }
.Ed
.Pp
.It Pq Ic inet Ns \&| Ns Ic inet6
This rule applies only to routes matching the stated address family.
The address family needs to be set only in rules that use
.Ic prefixlen
without specifying a
.Ic prefix
beforehand.
.Pp
.It Xo
.Ic prefix
.Ar address Ns Li / Ns Ar len
.Xc
This rule applies only to
.Em UPDATES
for the specified prefix.
.Pp
Multiple
.Ar address Ns Li / Ns Ar len
entries may be specified,
separated by commas or whitespace,
if enclosed in curly brackets:
.Bd -literal -offset indent
deny from any prefix { 192.168.0.0/16, 10.0.0.0/8 }
.Ed
.Pp
Multiple lists can also be specified, which is useful for
macro expansion:
.Bd -literal -offset indent
good="{ 192.168.0.0/16, 172.16.0.0/12, 10.0.0.0/8 }"
bad="{ 224.0.0.0/4, 240.0.0.0/4 }"
ugly="{ 127.0.0.1/8, 169.254.0.0/16 }"

deny from any prefix { $good $bad $ugly }
.Ed
.Pp
.It Ic prefixlen Ar range
This rule applies only to
.Em UPDATES
for prefixes where the prefixlen matches.
Prefix length ranges are specified by using these operators:
.Bd -literal -offset indent
=	(equal)
!=	(unequal)
<	(less than)
<=	(less than or equal)
>	(greater than)
>=	(greater than or equal)
-	(range including boundaries)
><	(except range)
.Ed
.Pp
>< and -
are binary operators (they take two arguments).
For instance, to match all prefix lengths >= 8 and <= 12, and hence the
CIDR netmasks 8, 9, 10, 11 and 12:
.Bd -literal -offset indent
prefixlen 8-12
.Ed
.Pp
Or, to match all prefix lengths < 8 or > 12, and hence the CIDR netmasks
0\(en7 and 13\(en32:
.Bd -literal -offset indent
prefixlen 8><12
.Ed
.Pp
.Ic prefixlen
can be used together with
.Ic prefix .
.Pp
This will match all prefixes in the 10.0.0.0/8 netblock with netmasks longer
than 16:
.Bd -literal -offset indent
prefix 10.0.0.0/8 prefixlen > 16
.Ed
.Pp
.It Ic quick
If an
.Em UPDATE
matches a rule which has the
.Ic quick
option set, this rule is considered the last matching rule, and evaluation
of subsequent rules is skipped.
.Pp
.It Ic set Ar attribute ...
All matching rules can set the
.Em AS path attributes
to some default.
The set of every matching rule is applied, not only the last matching one.
See also the following section.
.El
.Sh ATTRIBUTE SET
.Em AS path attributes
can be modified with
.Ic set .
.Pp
.Ic set
can be used on
.Ic network
statements, in
.Ic neighbor
or
.Ic group
blocks, and on filter rules.
Attribute sets can be expressed as lists.
.Pp
The following attributes can be modified:
.Pp
.Bl -tag -width Ds -compact
.It Xo
.Ic community Op Ar delete
.Ar as-number Ns Li : Ns Ar local
.Xc
.It Xo
.Ic community Op Ar delete
.Ar name
.Xc
Set or delete the
.Em COMMUNITIES
AS path attribute.
Communities are specified as
.Ar as-number Ns Li : Ns Ar local ,
where
.Ar as-number
is an AS number and
.Ar local
is a locally-significant number between zero and
.Li 65535 .
Alternately, well-known communities may be specified by name:
.Ic NO_EXPORT ,
.Ic NO_ADVERTISE ,
.Ic NO_EXPORT_SUBCONFED ,
or
.Ic NO_PEER .
.Pp
.It Ic localpref Ar number
Set the
.Em LOCAL_PREF
AS path attribute.
If
.Ar number
starts with a plus or minus sign,
.Em LOCAL_PREF
will be adjusted by adding or subtracting
.Ar number ;
otherwise it will be set to
.Ar number .
The default is 100.
.Pp
.It Ic med Ar number
.It Ic metric Ar number
Set the
.Em MULTI_EXIT_DISC
AS path attribute.
If
.Ar number
starts with a plus or minus sign,
.Em MULTI_EXIT_DISC
will be adjusted by adding or subtracting
.Ar number ;
otherwise it will be set to
.Ar number .
.Pp
.It Xo
.Ic nexthop
.Sm off
.Po Ar address \*(Ba
.Ic blackhole \*(Ba
.Ic reject \*(Ba
.Ic self \*(Ba
.Ic no-modify Pc
.Sm on
.Xc
Set the
.Em NEXTHOP
AS path attribute
to a different nexthop address or use blackhole or reject routes.
If set to
.Em no-modify ,
the nexthop attribute is not modified.
Unless set to
.Em self ,
the nexthop is left unmodified for IBGP
sessions.
.Em self
forces the nexthop to be set to the local interface address.
.Bd -literal -offset indent
set nexthop 192.168.0.1
set nexthop blackhole
set nexthop reject
set nexthop no-modify
set nexthop self
.Ed
.Pp
.It Ic pftable Ar table
Add the prefix in the update to the specified
.Xr pf 4
table, regardless of whether or not the path was selected for routing.
This option may be useful in building realtime blacklists.
.Pp
.It Ic prepend-neighbor Ar number
Prepend the neighbor's AS
.Ar number
times to the
.Em AS path .
.Pp
.It Ic prepend-self Ar number
Prepend the local AS
.Ar number
times to the
.Em AS path .
.Pp
.It Ic rtlabel Ar label
Add the prefix with the specified
.Ar label
to the kernel routing table.
.Pp
.It Ic weight Ar number
The
.Em weight
is used to tip prefixes with equally long AS paths in one or
the other direction.
A prefix is weighed at a very late stage in the decision process.
If
.Ar number
starts with a plus or minus sign, the
.Em weight
will be adjusted by adding or subtracting
.Ar number ;
otherwise it will be set to
.Ar number .
.Em Weight
is a local non-transitive attribute and a bgpd-specific extension.
For prefixes with equally long paths, the prefix with the larger weight
is selected.
.El
.Sh FILES
.Bl -tag -width "/etc/bgpd.conf" -compact
.It Pa /etc/bgpd.conf
.Xr bgpd 8
configuration file
.El
.Sh SEE ALSO
.Xr strftime 3 ,
.Xr ipsec 4 ,
.Xr pf 4 ,
.Xr tcp 4 ,
.Xr bgpctl 8 ,
.Xr bgpd 8 ,
.Xr ipsecctl 8 ,
.Xr isakmpd 8 ,
.Xr rc.conf.local 8
.Sh HISTORY
The
.Nm
file format first appeared in
.Ox 3.5 .