Asus A8V-E network controller

Hi. I'm running suse 9.2 x86_64 on an amd 64 with asus A8V-E deluxe (PCI express) motherboard. Am a linux newbie!

I installed suse about 2 weeks ago using the 'mini installation' and then over ftp by telling the kernel to use sk98lin (I think that's the one) to talk to the A8V-E's built in Ethernet controllor, which is a Marvell Yukon Gigabyte NIC on the motherboard. Everything went beautifully. Full install over the internet (nice and fast from university here). Harddrives (2xSATA) recognized, drivers installed, gfx fine etc...

HOWEVER. In the last few days I have totally lost net access. If I try to access a website through firefox, it times out after about 5 mins. If I try to get email (from MS exchange server) through gnome evolution 2, similar things happen. Sometimes it appears to work for the first minute or so after booting. Then nothing.

I've tried a few things (as follows) but am stuck! Please help!

1) Is the network working? Yes - lights are on, windowsxp dual boot has not problems (I'm writing this from xp on the same machine) etc.

2) run ifconfig (I've replaced some of the numbers here with letters):

eth0 Link encap:Ethernet HWaddr x:y:z:83:36:51
inet addr:a.b.66.2 Bcast:a.b.66.255 Mask:255.255.255.0
inet6 addr: g::h:i:fe83:3651/64 Scope:Link
UP BROADCAST NOTRAILERS RUNNING MULTICAST MTU:1500 Metric:1
RX packets:416 errors:0 dropped:0 overruns:0 frame:0
TX packets:236 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:257811 (251.7 Kb) TX bytes:63338 (61.8 Kb)
Interrupt:58 Memory:dc000000-0


lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:40 errors:0 dropped:0 overruns:0 frame:0
TX packets:40 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:2608 (2.5 Kb) TX bytes:2608 (2.5 Kb)

seems all fine. ifstatus gives similar info but with DHCP etc. all nicely configured, as expected.

3) check out /var/log/messages
This is where it gets interesting. From my latest bootup:

eth0 config..
Mar 18 10:43:26 medah kernel: speed: 100
Mar 18 10:43:26 medah kernel: autonegotiation: yes
Mar 18 10:43:26 medah kernel: duplex mode: full
Mar 18 10:43:26 medah kernel: flowctrl: symmetric
Mar 18 10:43:26 medah kernel: irq moderation: disabled
Mar 18 10:43:26 medah kernel: scatter-gather: enabled
Mar 18 10:43:26 medah kernel: rx-polling: enabled

but then..

Mar 18 10:43:29 medah kernel: eth0: -- ERROR --
Mar 18 10:43:29 medah kernel: Class: Hardware failure
Mar 18 10:43:29 medah kernel: Nr: 0x271
Mar 18 10:43:29 medah kernel: Msg: Uncorrectable PCI Express
error

However, after booting, ifconfig *still* gives the correct looking info and everything looks fine in yast2. Sounds dramatic-- 'hardware failure'!

4) ping something. anything! nothing works. all hosts unreachable.

Any ideas? I suspect a driver issue but the weird thing is that this worked perfectly at first, in fact I installed suse over ftp! Perhaps evolution has screwed up some settings? Or could I inadvertently have messed the eth0 settings in yast2? Or does the PCI-E driver for the eth0 just not work properly?

Thanks very much for any help
Dave

There can be many reasons I'll try to give you the commands to test a few of them

first log in as root

try to find the routing table, the default route might be missing
#ip r l (I prefer this one but when not available use route)
#route

see if you can find a route 'default or 0/0 or 0.0.0.0 mask 0.0.0.0' anyway something
that catches all.

there will be a 'via' or 'gateway' address try to ping this one
if this one replies, your computer can be ok (doesn't have to be
but the chances are good)

it might be a problem with your DNS
take a look at
cat /etc/resolv.conf

are there any nameservers listed ?


in case none of the above helps, try to explain me where you ip adres comes from, eg you configure it, or it is given with DHCP (from where) what type of internet connection do you use? ADSL-route, pppoe to your PC, anything else?

do you have some firewall configuration?

iptables -L
iptables -t nat -L

can give you the rules that are configured

I hope any of the above helps, there is a last step you can do, using ethereal or tcpdump, that
can give lots of information but you need to understand what is going on. Let me know if you
can find anything

Tlowk

Hi. Thanks for your reponse. I've tried some of the stuff you suggested, results as follows...

1) ip r l
this gives:
a.b.66.0/24 dev eth0 proto kernel scope link src a.b.66.2
c.d.0.0/16 dev eth0 scope link
127.0.0.0/8 dev lo scope link
default via a.b.66.1 dev eth0

2) try to ping a.b.66.1
this results in "destination host unreachable"
btw, if I do the same in windowsXP from the command prompt, it's fine and I get a reply from a.b.66.1

3) take a look at cat /etc/resolv.conf - are there any nameservers listed ?
yes - a.b.63.100 and a.b.63.101 and a.b.5.12
it also says resolv.conf modified by dhcpcd, which makes sense since my IP address is assigned by DHCP. The DHCP server knows this machine's mac address too, since we had to set that up to get it working in windows (Mac address is hardware, not OS, specific, right?)

4) do you have some firewall configuration?
yes - suse firewall
iptables -L gives a lot of info which I'm not sure how to decipher
iptables -t nat -L gives:

Chain PREROUTING (policy ACCEPT)
target prot opt source destination

Chain POSTROUTING (policy ACCEPT)
target prot opt source destination

Chain OUTPUT (policy ACCEPT)
target prot opt source destination

5) other info. After a ping, TX packets in ipconfig increases.

Otherwise, I am clueless about what it could be. Am considering a reinstall of suse 9.2 (although have done that once already). Or just using windowsXP--- it's one heck of a lot easier!!!! :-)

Dave

At this point I see 3 possibilities

- drivers don't work (but they did before)
- the gateway doesn't like something that is different in linux as in windows
- the firewall is to close ?


try a ping to your gateway and in another window try
tcpdump -i eth0 icmp

this assumes that your gateway replies to a ping (normally it does but there can be
reasons to avoid that)

when you see an ICMP REQUEST and an ICMP REPLY in the tcpdump but not in the window
where your start the ping, you can assume that your firewall stops it.

does the number of transmit and receive in ifconfig increase both?

The mac adres is hardware related, but can easily be changed in linux (take a look at man ifconfig)

iptables -L INPUT
can may help, but maybe there is a command to temporary disable the firewall. I'd try
/etc/init.d/iptables stop
The other way can be

iptables -P INPUT ACCEPT
iptables -P OUTPUT ACCEPT
iptables -F
iptables -t nat -F

after this iptables -t nat -L and iptables -L should not return a lot anymore and I think that
most of your firewall is disabled until the next reboot.

when nothing works try to give the
iptables -L
iptables -t nat -L
output maybe I can understand more of it. But that will only be useful when you see the icmp reply arriving in tcpdump

Tlowk

Hi. I haven't yet tried any of the above. I've tried to delete the network card in YAST and then reinstall. Didn't achieve anything.
Then I thought, right, I've had it with screwing around, I'm not here to become a linux expert I just need it to WORK!! So I tried to reinstall Suse from the mini-installation CD that I have. That gave some weird issues. Choosing 'installation' after booting from the CD and it bailed out, saying:
RAMDISK: compressed image found at block 0
crc error
CFS: cannot open root device "<NULL>" or unknown-block(3,3)
Please append a correct "root=" boot option
Kernel panic - not syncing: VFS: unable to mount root fs on unknown-block(3,3)

WHAT THE??!? crc error? what the heck is going on now?
Anyway, I've abandoned it all for the time being. Thoroughly fed up with linux. Am currently working in Windows XP on the same machine with no issues. No CRC errors, web access is fine, etc....

Any suggestions from anyone would be much appreciated. Thanks
Dave

Further developments. Suspecting faulty RAM I have run memtest-86 and one of my (brand new!) 512Mb crucial ballastix modules has faults in 3 locations. Perhaps this has been the problem all along.... linux seems ok now!!! (I am writing this from within suse - web access is back!)

Has anyone else had RAM problems like this? I thought crucial ballistix would be good!

By the way, when I pulled the module out it was quite warm, I guess that's to be expected, has a massive heatsink on it and had just been doing an intensive memory test. Does high performance RAM run warm?

best wishes
Dave