Section: Maintenance Commands (8)Updated: 3 July 2000Local indexUp
pppoe - user-space PPPoE client.
pppd pty 'pppoe [pppoe_options]' [pppd_options]
pppoe -A [pppoe_options]
pppoe is a user-space client for PPPoE (Point-to-Point Protocol
over Ethernet) for Linux and other UNIX systems. pppoe works in
concert with the pppd PPP daemon to provide a PPP connection
over Ethernet, as is used by many DSL service providers.
The -I option specifies the Ethernet interface to use. Under Linux,
it is typically eth0 or eth1. The interface should be "up"
before you start pppoe, but should not be configured to have
an IP address.
The -T option causes pppoe to exit if no session traffic
is detected for timeout seconds. I recommend that you use this
option as an extra safety measure, but if you do, you should make sure
that PPP generates enough traffic so the timeout will normally not be
triggered. The best way to do this is to use the
lcp-echo-interval option to pppd. You should set the
PPPoE timeout to be about four times the LCP echo interval.
The -D option causes every packet to be dumped to the specified
file_name. This is intended for debugging only; it produces huge
amounts of output and greatly reduces performance.
The -V option causes pppoe to print its version number and
The -A option causes pppoe to send a PADI packet and then print
the names of access concentrators in each PADO packet it receives. Do not
use this option in conjunction with pppd; the -A option is
meant to be used interactively to give interesting information about the
Specifies the desired service name. pppoe will only initiate sessions
with access concentrators which can provide the specified service. In
most cases, you should not specify this option. Use it only if you
know that there are multiple access concentrators or know that you need a
specific service name.
Specifies the desired access concentrator name. pppoe will only
initiate sessions with the specified access concentrator. In
most cases, you should not specify this option. Use it only if you
know that there are multiple access concentrators. If both the
-S and -C options are specified, they must both match
for pppoe to initiate a session.
Causes pppoe to use the Host-Uniq tag in its discovery packets. This
lets you run multiple pppoe daemons without having their discovery
packets interfere with one another. You must supply this option to
allpppoe daemons if you intend to run multiple daemons
Causes pppoe to use synchronous PPP encapsulation. If you
use this option, then you must use the sync option with
pppd. You are encouraged to use this option if it works, because
it greatly reduces the CPU overhead of pppoe. However, it
MAY be unreliable on slow machines -- there is a race condition between
pppd writing data and pppoe reading it. For this reason, the default
setting is asynchronous. If you encounter bugs or crashes with Synchronous
PPP, turn it off -- don't e-mail me for support!
Causes pppoe to clamp the TCP maximum segment size at the specified
value. Because of PPPoE overhead, the maximum segment size for PPPoE is
smaller than for normal Ethernet encapsulation. This could cause problems
for machines on a LAN behind a gateway using PPPoE. If you have a LAN
behind a gateway, and the gateway connects to the Internet using PPPoE,
you are strongly recommended to use a -m 1412 option. This avoids
having to set the MTU on all the hosts on the LAN.
to use the indicated Ethernet MAC address as the source address for
must be specified in the
syntax. If this option is specified,
puts the interface into promiscuous mode.
Causes pppoe to write its process-ID to the specified file. This
can be used to locate and kill pppoe processes.
Causes pppoe to skip the discovery phase and move directly to the
session phase. The session is given by sess and the MAC address of
the peer by mac. This mode is not meant for normal use; it
is designed only for pppoe-server(8).
Causes pppoe not to open a discovery socket. This mode is
not meant for normal use; it is designed only for
Causes pppoe to terminate an existing session by sending a PADT frame,
and then exit. You must use the -e option in conjunction with this
option to specify the session to kill. This may be useful for killing
sessions when a buggy peer does not realize the session has ended.
Causes pppoe to perform discovery and then exit, after printing
session information to standard output. The session information is printed
in exactly the format expected by the -e option. This option lets
you initiate a PPPoE discovery, perform some other work, and then start
the actual PPP session. Be careful; if you use this option in a loop,
you can create many sessions, which may annoy your peer.
The -f option sets the Ethernet frame types for PPPoE discovery
and session frames. The types are specified as hexadecimal numbers
separated by a colon. Standard PPPoE uses frame types 8863:8864.
You should not use this option unless you are absolutely sure
the peer you are dealing with uses non-standard frame types. If your
ISP uses non-standard frame types, complain!
The -h option causes pppoe to print usage information and
PPPoE (Point-to-Point Protocol over Ethernet) is described in RFC 2516
and is a protocol which allows the session abstraction to be maintained
over bridged Ethernet networks.
PPPoE works by encapsulating PPP frames in Ethernet frames. The protocol
has two distinct stages: The discovery and the session stage.
In the discovery stage, the host broadcasts a special PADI (PPPoE
Active Discovery Initiation) frame to discover any access
concentrators. The access concentrators (typically, only one
access concentrator) reply with PADO (PPPoE Active Discovery Offer)
packets, announcing their presence and the services they offer. The
host picks one of the access concentrators and transmits a PADR (PPPoE
Active Discovery Request) packet, asking for a session. The access
concentrator replies with a PADS (PPPoE Active Discovery
Session-Confirmation) packet. The protocol then moves to the session stage.
In the session stage, the host and access concentrator exchange PPP frames
embedded in Ethernet frames. The normal Ethernet MTU is 1500 bytes, but
the PPPoE overhead plus two bytes of overhead for the encapsulated PPP
frame mean that the MTU of the PPP interface is at most 1492 bytes.
This causes all kinds of problems if you are using a Linux machine
as a firewall and interfaces behind the firewall have an MTU greater than
1492. In fact, to be safe, I recommend setting the MTU of machines
behind the firewall to 1412, to allow for worst-case TCP and IP options
in their respective headers.
Normally, PPP uses the Link Control Protocol (LCP) to shut down a PPP
link. However, the PPPoE specification allows the link to be shut down
with a special PADT (PPPoE Active Discovery Terminate) packet. This client
recognizes this packet and will correctly terminate if a terminate request
is received for the PPP session.
My design goals for this PPPoE client were as follows, in descending order
It must work.
It must be a user-space program and not a kernel patch.
The code must be easy to read and maintain.
It must be fully compliant with RFC 2516, the proposed PPPoE standard.
It must never hang up forever -- if the connection is broken, it must
detect this and exit, allowing a wrapper script to restart the connection.
It must be fairly efficient.
I believe I have achieved all of these goals, but (of course) am open
to suggestions, patches and ideas. See my home page,
http://www.roaringpenguin.com, for contact information.
For best results, you must give pppd an mtu option of
1492. I have observed problems with excessively-large frames
unless I set this option. Also, if pppoe is running on a firewall
machine, all machines behind the firewall should have MTU's of 1412.
If you have problems, check your system logs. pppoe logs interesting
things to syslog. You may have to turn on logging of debug-level
messages for complete diagnosis.
pppoe was written by David F. Skoll <firstname.lastname@example.org>,
with much inspiration from an earlier version by Luke Stras.