Hello all,

I just installed Debian and have hooked up eth0 to a Netgear wireless router (I'm using wired until I get my wireless drivers to work). Whenever I run dhclient, ifup eth0, /etc/init.d/networking restart, or anything at all to request a DHCP address, I get this:

I also have an XP box, also hooked up through eth0 to the router. DHCP on that machine works just dandy. The Windows command ipconfig /all tells me that the DHCP server is 10.0.0.1, the address of my router.

ping 10.0.0.1 gives this:

/etc/network/interfaces reads in part:

Also, for some reason, occasionally I see ifup eth0 run on boot, and it connects successfully. But then when X starts up, it somehow... disconnects? The "Network Tools" utility says only about 15KB were sent. I'm not sure what to call it, since Debian still claims my connection is active. But I'm in the same situation as if DHCP had failed. And it wouldn't make sense for a lease to expire that quickly. Perhaps it's a separate issue.

Any help would be greatly appreciated! Thanks in advance.

1. check your cable
2. check the setup of your router - it may be configured to only accept requests from certain MAC numbers, or it may be set to dispense static addresses and you have no static address assigned to your Linux box.

Thanks, I'm checking my router setup just in case, but I think both those issues are addressed by the fact that the Internet works fine when I boot into XP. Perhaps I wasn't clear; I have an XP box connected to the router, and another box that dual-boots Debian and XP. Both XPs connect to the Internet with no problems. Debian is the only one with DHCP issues. Since Debian and XP share the MAC address of my Ethernet card, I think we can eliminate that proposal. Unless I'm missing something.

Thanks in advance again!

I wonder if it doesn't like the fact that eth0 is defined as both allow-hotplug and allow-auto? Try commenting out the auto line and see if that changes it.

I would suggest removing the "allow-hotplug" line since that's only needed if your NIC is removable (like a PCMCIA type) - I can't see how it would otherwise affect your system though.

What is the NIC as determined by matching entries in 'lspci' and 'lspci -n'? It is possible that the wrong driver is loaded so it's best to track the driver from the VendorID:ProductID numbers.

And just to check one more time using 'ifconfig', try:

ifdown eth0
ifconfig eth0 up 10.0.0.6 netmask 255.255.255.0

then try that ping again.

I also forgot - when you have that trouble, can you log into the router via the other computer and check the active leases? Some routers are a bit stupid and require you to drop the lease - most however drop the lease as soon as the cable is disconnected or the computer powered down or else they hand out a new lease anyway.

I tried commenting out the auto line, the allow-hotplug line and then both just for good measure, but nothing changed.

I tried ifdown eth0 and got this:
So apparently it's aware of the router, at least; just not able to connect to it? I tried ifconfig eth0 up 10.0.0.6 netmask 255.255.255.0 as well and got this:

lspci gives me:
And lscpi -n gives me:
Is that what you're looking for? If not, I'll be happy to give any other information.

Have you tried "ifup --force eth0"?

Guess I don't have much else to offer.

The bit you posted:

03:00.0 0280: 14e4:4328 (rev 03)

That's your Wireless gizmo. The posts I have dug up all say "use ndiswrapper" which means installing ndiswrapper and setting up the WinDuhs driver.

Have another try at identifying the wired network device.

The other lines you posted like:
SIOCSIFADDR: No such device
indicate that the driver is not working for whatever reason (usually missing firmware, but in some cases the wrong driver loads as already mentioned). At least we've established that the DHCP messages are probably misleading because your NIC isn't actually working.

Can you also post the output of:
dmesg|grep eth

Thanks for your help guys, finally got ndiswrapper working (on Ubuntu) and everything works great now.