DHCP problem with Knoppix 3.3, 3Com 3C905B card

[Li Gwai Lo]

I am trying to use Knoppix 3.3 EN (dated 11-03) on my Asus P5A / K6-2/400 / 3Com 3C905B.

The network that I am on uses DHCP. Knoppix seems to set the ethernet address to be the address of the gateway -- I get x.x.0.1 as my machine's address, when I should be getting x.x.0.5 instead.

Running the "netcardsetup" script and rejecting DHCP -- giving it all of the same values that it should get from DHCP if its auto-config were working -- doesn't work either. I can't even ping my gateway.

I know the machine and ethernet card function (I'm currently running Win98 on the box in order to post this).

Does anyone know what the problem might be? As an aside, I have only managed to get one Linux distro to network properly with this machine, a Redhat 6.2 (or maybe 6.1) edition that was designated by Oracle as the officially-supported distro to work with their database. All of the other distros that I have tried have resulted in weeks of frustrating trial-and-error without ever getting a network connection to work, even though they seem to recognize everything fine. (Similarly, I tried replacing the card with an Intel card, whatever was supposedly Linus' favorite that month, and I couldn't get anything to work with it either!)

[Li Gwai Lo]

The saga continues. . . .

I gave it one more shot after failing five times to get a working network connection either by DHCP or by manually setting the values.

On the sixth reboot, which I did solely to get a snapshot of what the automated DHCP request would get, it magically got a working set of values -- setting it to the same address that I'd been trying to get it to accept for hours.

Does anyone know why DHCP would work only occasionally? This does NOT appear to be a problem with the rest of the network; every bootup with a Windows-based machine (98SE, 2K, and XP in use) has worked fine.

The only clue that appears in the data is that the HWAddr is "FF:FF:FF:FF:FF:FF" on the times it failed, and a valid card number on the one time it worked:

<pre>
eth0 Link encap:Ethernet HWaddr 00:10:5B:10:6C:54
inet addr:182.143.0.5 Bcast:182.143.0.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:11 errors:0 dropped:0 overruns:0 frame:0
TX packets:3 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100
RX bytes:1883 (1.8 KiB) TX bytes:1522 (1.4 KiB)
Interrupt:10 Base address:0xb800
</pre>

instead of
<pre>
eth0 Link encap:Ethernet HWaddr FF:FF:FF:FF:FF:FF
inet addr:182.143.0.5 Bcast:182.143.0.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100
RX bytes:0 (0.0 b) TX bytes:0 (0.0 b)
Interrupt:10 Base address:0xb800
</pre>

What might cause this??

[Li Gwai Lo]

I found a new clue. When I booted this time, I saw an error message come up which looked something like this:

eth0: command didn't complete! 0x3002 status 0xffff

This is a paraphrase; the message was substantially similar but may have been slightly different. I couldn't write it down before X11 started and the message disappeared, but I memorized the failed command (definitely 0x3002) at least.

This came up on an attempt to boot without specifying a config, and the card didn't initialize properly. After rebooting, it came up ok and didn't show this message.

Does anyone know what this command (0x3002) is, or what to do to prevent it from erroring out?

BTW, Knoppix now *usually* comes up without a problem on the SECOND try. It seems like it always fails on the first bootup, then a shutdown and reboot will allow it to initialize the network card correctly so that it can talk to the world.

[windos_no_thanks]

BTW, Knoppix now *usually* comes up without a problem on the SECOND try. It seems like it always fails on the first bootup, then a shutdown and reboot will allow it to initialize the network card correctly so that it can talk to the world.
Have you always been running windos in between ?
Some people over here suspect that a service pack issued by M$ adds code that intentionally messes up the EEPROM of some network cards while shutting down:
http://www.knoppix.net/forum/viewtopic.php?t=2164