[gentoo-user] eth0 link down

[gentoo-user] eth0 link down

[James]

hello,

I have a portable that has been working great. Now, the eth0 link
does not stay up. If I watch the system boot, all looks fine.
Looking at 'dmesg' the relevant entry are:

eth0: RealTek RTL8139 at 0x2000, 00:90:f5:2b:cb:18, IRQ 11
eth0: Identified 8139 chip type 'RTL-8100B/1839D'
<snip>
[drm] Setting GART locaatoin based on new memory map
[drm] Loading R200 Microcode
[drm] writeback test succeded in 1 usecs
'eth0: link down'

I do not think that 'drm has anything to do with this, but I do not
remember it  in my dmesg before on this system. Perhaps updateding
something on the windows side (corrupted the bios?)......

I've rebuild the kernel and the realtek driver is built into the kernel.
etho, the built in realtek chips has  always worked before. I keep 
several older kernels around and 
ready to use in grub.conf. They all suffer the same problem. 
ifconfig -a and netstat -nr look normal (similar to other workstations
on the same ethernet segment).

I've double check /etc/conf.d/ and /etc/init.d/  and all of the files
and links are correct.

I have notice that /etc/init.d/net.lo got upgraded during an routine
'emerge -uDNvp world' (no I did not save the older version) 
but, I'm hoping that's not the problem.

I know the ethernet chip is good because the portable works fine
booted into windows2K. I have switched ethernet cables with a known good
cable and that is not the problem either.

Any ideas or suggestions are most welcome.


James

-- 
gentoo-user@gentoo.org mailing list

[gentoo-user] update: Re: eth0 link down

[James]

James <wireless <at> tampabay.rr.com> writes:



> eth0: RealTek RTL8139 at 0x2000, 00:90:f5:2b:cb:18, IRQ 11
> eth0: Identified 8139 chip type 'RTL-8100B/1839D'
> <snip>

> 'eth0: link down'

Well forget all of that other conjecture. I installed a pcmcia
ethernet card manufacture = dynex.
Since it uses a realtek chip, I have no problems getting the dynex to
work as eth1.

Relevant portions of dmesg:
8139too Fast Ethernet driver 0.9.27
eth0: RealTek RTL8139 at 0x2000, 00:90:f5:2b:cb:13, IRQ 11
eth0:  Identified 8139 chip type 'RTL-8100B/8139D'

eth1: RealTek RTL8139 at 0x2400, 00:13:46:37:17:c0, IRQ 5
eth1:  Identified 8139 chip type 'RTL-8139C'
[drm] Setting GART location based on new memory map
[drm] Loading R200 Microcode
[drm] writeback test succeeded in 1 usecs
eth0: link down
eth1: link up, 100Mbps, full-duplex, lpa 0x45E1
eth1: link down
eth1: link up, 100Mbps, full-duplex, lpa 0x45E1


What's going on?  We had another clevo portable running gentoo
and this happened about (3) weeks ago. I just assumed the
ethernet chip went bad. (2) indentical realtek chips on
(2) different clevo portable going bad in a (3) week period?

Very suspicious, statistically.

this onboard realtek chips works when I pull the pccard ethernet
chip and reboot and use the onboard ethernet chip.

Ideas?

james



-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] eth0 link down

[Richard Fish]

On 9/20/06, James <wireless@tampabay.rr.com> wrote:
> 'eth0: link down'

This line in dmesg means that your hardware saw the cable disconnect.
You should be able to see this effect by running "ethtool eth0 | grep
Link".  There is nothing that Gentoo can do to cause this...it is
purely between the kernel driver and the hardware.  Even "ifconfig
eth0 down" doesn't cause this.

This could be the fault of the chip in the laptop, the cable, or the
switch that you plug into.  Since you already tried a different cable,
and the NIC chip isn't easily replaced, I would try to prove that the
switch is good.  Assuming you have control over the switch, try
different port.  Or if you don't, insert a small desktop switch
between you and the upstream one.

-Richard
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] update: Re: eth0 link down

[Richard Fish]

On 9/20/06, James <wireless@tampabay.rr.com> wrote:
> [drm] Setting GART location based on new memory map

BTW, this drm stuff is completely unrelated...it has to do with your
graphics card.

> What's going on?  We had another clevo portable running gentoo
> and this happened about (3) weeks ago. I just assumed the
> ethernet chip went bad. (2) indentical realtek chips on
> (2) different clevo portable going bad in a (3) week period?

Starting to look more and more like a flakey switch to me.

-Richard
-- 
gentoo-user@gentoo.org mailing list

[gentoo-user] Re: eth0 link down

[James]

Richard Fish <bigfish <at> asmallpond.org> writes:

>  Assuming you have control over the switch, try
> different port.  Or if you don't, insert a small desktop switch
> between you and the upstream one.

Hello Richard,

Well I changed a switch in the middle and it's all fine now.
The switch must have had a bad port. I had other traffic running
just fine through the switch, so I did not think that was the problem.

Nice to know about ethtool....

thanks,

James




-- 
gentoo-user@gentoo.org mailing list