| View previous topic :: View next topic |
| Author |
Message |
Cadex
n00b
Joined: 16 Mar 2005
Posts: 14
|
 Posted: Wed Mar 16, 2005 4:49 pm Post subject: e100 NIC doesn't receive or handle ARP responses (I guess) |
 |
|
Hello,
I am just about to finish my Stage3-Gentoo-installation on my Compaq Armada E700 notebook, but for some strange reason I just can't get the integrated Intel Etherexpress PRO/100 Mini-PCI card running. I've seen quite a number of threads on that topic, but though I've read many of them my problem seems to be different.
<EDIT>
I ran ethereal on my other PC and tried to ping that computer. The following two packets appeared over and over again:
| Code: | No. Time Source Destination Protocol Info
3 14.555090 192.168.0.11 Broadcast ARP Who has 192.168.0.1? Tell 192.168.0.11
Frame 3 (60 bytes on wire, 60 bytes captured)
Ethernet II, Src: 00:d0:59:36:6a:30, Dst: ff:ff:ff:ff:ff:ff
Address Resolution Protocol (request)
No. Time Source Destination Protocol Info
4 14.555172 192.168.0.1 192.168.0.11 ARP 192.168.0.1 is at 00:07:95:ad:4c:66
Frame 4 (42 bytes on wire, 42 bytes captured)
Ethernet II, Src: 00:07:95:ad:4c:66, Dst: 00:d0:59:36:6a:30
Address Resolution Protocol (reply) |
I'm not really familiar with ARP and other network protocols, but I guess this means that the e100 NIC successfully transmits the ARP package to query for the MAC of 192.168.0.1, my PC returns the correct response, but for some reason this ARP response doesn't arrive or isn't handled for some reason...
</EDIT>
I use the kernel sources linux-2.6.10-gentoo-r6 , and tried both eepro100 and e100. I compiled them directly into the kernel, not as modules. I configured the card as 192.168.0.11
Now when the system boots, I see no error messages, it looks to me like the card is correctly recognized. But: I can't even ping my other PC at 192.168.0.1 . But looking at my ethernet switch, I can see lights flickering when I try to ping the other computer, so there seems to be some kind of communication...
I also have a RT8139 PCMCIA-card which is eth1, configured to 192.168.0.12 & compiled as a module, and when I stop eth0, modprobe 8139too, plug the network cable into the 8139 adapter and start eth1, I can successfully ping my other computer without any problems!
I don't think the e100 network device is actually broken, as for example I could successfully PXE boot from the e100 card.
I can't actually tell you if the e100 card is working when booting the Gentoo LiveCD, as I don't have a CDROM drive - I was installing Gentoo from an existing SuSE 9.2 installation.
The problem does not appear to be Gentoo-specific: When I installed SuSE using the SuSE 9.2 installation floppy disks, I had the same problem: When I used the builtin Mini-PCI e100 card I could not reach my FTP server, using the RTL8139 PCMCIA card it worked fine. Same problem with the finished SuSE 9.2 installation, which is using a precompiled "2.6.8-24" SuSE kernel.
BUT: Prior to installing Gentoo, I tried "Linux from scratch" 6.0-1 (?), and there the e100 NIC worked fine after installation! (Kernel 2.6.8.1, driver eepro100)
Anyone got any idea? If you need further information, please ask, I still only have basic Linux skills and don't actually know which information is relevant in order to fix my problem... |
|
| Back to top |
|
 |
egberts
Guru
Joined: 04 Nov 2003
Posts: 359
Location: Dimmed Cathode Ray Tube
|
| Posted: Thu Mar 17, 2005 4:03 am Post subject: |
|
|
Two tools in your arsenal:
1. mii-tool (emerge sys-apps/ethtool)
2. tcpdump (emerge net-analyzer/tcpdump)
# tcpdump -i eth0 -vv -s 0
and you'll be able to study how the packet go back and forth.
# mii-tool eth0
and you'll be able to ascertain that you have a true UP link. |
|
| Back to top |
|
|
egberts
Guru
Joined: 04 Nov 2003
Posts: 359
Location: Dimmed Cathode Ray Tube
|
| Posted: Thu Mar 17, 2005 4:05 am Post subject: |
|
|
Whoa!
You cannot share the same subnet between eth0 (Intel EtherExpressPRO100) and eth1 (RealTek).
Put a way different IP address on eth1, like 192.168.1.x on eth1.
# vi /etc/conf.d/net
< change your eth1 address>
Make sure that your subnet is 255.255.255.0 or /24.
# /etc/init.d/net.eth1 restart
# /etc/init.d/net.eth0 restart |
|
| Back to top |
|
|
egberts
Guru
Joined: 04 Nov 2003
Posts: 359
Location: Dimmed Cathode Ray Tube
|
| Posted: Thu Mar 17, 2005 4:10 am Post subject: |
|
|
I get the sinking feeling that you're attempting to use both eth0 and eth1 at the same hub/switch.
Well, unless you're a skillful network engineer, a guru with route tables, or can do masterful load-balancing setup, you can't expect to utilize both eth ports AND have them both within the same subnet.
These two eth ports CAN however reside on the same hub/switch IF they have different subnets; otherwise you'll only be able to use one of them.
If you wish to attain gainful uses of eth0, then you must disable eth1 or change eth1 subnet to something else. |
|
| Back to top |
|
|
Cadex
n00b
Joined: 16 Mar 2005
Posts: 14
|
| Posted: Thu Mar 17, 2005 7:54 am Post subject: |
|
|
Hi egberts, thank you for your tips.
Unfortunately the two NICs in the same subnet shouldn't be the problem, for the following reasons:
- I always only use one at a time (by stopping one interface and starting the other - I don't actually want to use them both at a time, I just configured the second card so I still have network access)
- the problem already existed before I even configured the secondary PCMCIA card with the 8139 chipset
I tried "mii-tool eth0", but the result seems ok:
| Code: | | eth0: negotiated 100baseTx-FD flow-control, link ok |
I also tried forcing different speeds (100/10Mbps) and both full and half duplex, but that didn't help either.
I can't currently run tcpdump, as I started emerging kde about 10 hours ago on my plain gentoo install, and it still hasn't finished compiling yet... (400MHz only) I'll try tcpdump afterwards.
But I more and more get the impression that this isn't a simple misconfiguration bug. 
I now remember some other problem I encountered a few weeks ago, when I was using SuSE:
When I inserted my PCMCIA NIC in my lower PCMCIA slot, the card was properly recognized but it seemed like the card could send out packets but not receive any, as ifconfig showed lots of TX packets but RX packets always stayed 0. I tried some testing, unplugged the card several times - and by accident I plugged the card into the upper PCMCIA slot - and all of a sudden the card worked without a problem! |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Networking & Security
