> Has anyone been able to get Linux working with the Cox Cable Modem
> I've tried the following:
> 1.) dhcpcd (errors out saying no DHCP server) although this works fine in
> Windows NT/95.
> 2.) Manually config ifconfig and routes. When I do this, I can't see
> There's nothing wrong with the network card I'm using since I had (just
> previously) been using it as a LAN connection.
> I'm at a loss as to why this isn't working. Anyone have any ideas?
Let me give more details. This message amounts to "I configured it and
it doesn't work."
My linux machine is dual-homed. It has two SMC Ultra Elite cards: eth0
eth0 is used for LAN and eth1 will (hopefully) be used for WAN via cable
Please note that I have been able to get the cable modem service to work
properly using Windows NT Workstation. There are no problems with DHCP
server or the modem itself. Also note that eth1 has worked when
configured for a LAN IP. I have been able to connect to the machine
through eth1. I can therefore rule out hardware issues associated with
the modem and the card. I can also rule out problems on the remote end
I have tried two approaches:
1.) dhcp using "dhcpcd eth1"
allocate them. DHCP, they said, is used to ease administration on their
end. It is probably also in place to allow further growth of their
So, lets say my IP address is: 184.108.40.206. My netmask would be
255.255.255.0 and my gateway should be 220.127.116.11.
Now I'm not sure how they assign the IP address using DHCP if it is
fixed. So I have to assume that it goes by the hostname. I assume that
DHCP has some facility for transferring the hostname to the DHCP server
so that the DHCP server knows which address to return.
I will note that I was not able to get DHCP working properly on Windows
machines until Microsoft Networking was installed, WINS server addresses
were correctly configured, and the Microsoft Network Node name was set to
the DNS name. Perhaps the method they use for assign DHCP goes through
Microsoft Networking. I am rather ignorant of how DHCP gets the job
done, so any clarification would be greatly appreciated.
When running "dhcpcd eth1" with my hostname correctly set, ifconfig
reveals the IP address is 0.0.0.0 and it remains there until dhcpcd times
out. I get an error in /var/adm/syslog saying that it was not able to
receive a DHCP command (the name which I forget) which appears to be
related to replying to the DHCP request.
Accordingly, I gave up with DHCP since there doesn't appear to be a lot
you can do with it on the client side.
2.) I decided to configure eth1 directly using the assigned network
hostname# ifconfig eth1 18.104.22.168 -nopointopoint
doing this will report back that the interface is configured for that
address with the correct subnet (255.255.255.0) and an acceptable
broadcast address. However, it appears to have a ton of options on which
I did not specify, things like "MASTER" and "SLAVE" and "PROMISC"
I'm not sure why these items are being turned on. Anyone who can fill me
in on those details would be greatly appreciated. Also note that I must
specifically disable pointtopoint or else the interface trys to set
itself up as a point-to-point interface. Perhaps someone can enlighten
me on this as well.
hostname#route add -net 22.214.171.124 eth1
hostname#route add default gateway 126.96.36.199
doing this will report back the correct routing table configuration.
This should be all there is to it, unless I'm missing something glaring.
At this point pings to 188.8.131.52 (my IP) will respond. However,
pings to ANY other address including the router never respond.
I would appreciate any insight into what may be my problem with either
the dhcpcd or the manual configuration. If you have been able to get
Also note that I have tried configuring all of this on eth0 (substituting
the interface above) to no avail.
I will monitor this newsgroup as well as email for replies.