Linksys NC100 ethernet problems (tulip driver) [SOLVED]

Message

meson_ray · Post by **meson_ray** » Fri Sep 14, 2007 4:40 am

This ethernet card works perfectly fine with the livecd, and gives me an ip. When I configure the kernel, I add the support for the kernel driver, and compile the kernel, and reboot, and it doesn't work. I have had both gentoo and fedora installed on this computer before (just uninstalled fedora) and it has worked fine in both, so I know the card works, I'm just having some problems getting it working this time... I may have misconfigured the kernel, but I'm not sure what it would be...

Here's some outputs, edited for context:

Code: Select all

# dmesg | grep -C 3 tulip
Linux Tulip driver version 1.1.15-NAPI (Feb 27, 2007)
PCI: Enabling device 0000:00:0a.0 (0014 -> 0017)
PCI: Found IRQ 9 for device 0000:00:0a.0
tulip0:  MII transceiver #1 config 1000 status 786d advertising 05e1.
tulip0:  MII transceiver #2 config 1000 status 786d advertising 05e1.
tulip0:  MII transceiver #3 config 1000 status 786d advertising 05e1.
tulip0:  MII transceiver #4 config 1000 status 786d advertising 05e1.
eth0: ADMtek Comet rev 17 at MMIO 0xd6800000, MACID, IRQ 9.
netconsole: not configured, aborting

later on in boot process:

0000:00:0a.0: tulip_stop_rxtx() failed (CSR5 0xfc07c056 CSR6 0xff970111)
eth0: Setting full-duplex based on MII#1 link partner capability of 45e1.
eth0: no IPv6 routers present
0000:00:0a.0: tulip_stop_rxtx() failed (CSR5 0xfc07c056 CSR6 0xff970111)

ifconfig -a shows output for eth0 - so it knows the card is there, but there is no ip address.

On startup, dhcpcd eth0 times out and will not give an IP. When manually setting an ip and gateway through ifconfig and route, it refuses to connect to the internet.

Any suggestions that anyone has would be greatly appreciated...

Post by **John R. Graham** » Fri Sep 14, 2007 5:12 am

What's the exact Tulip variant you have enabled? For the LinkSys NC100, I use "DECchip Tulip (dc2114x) PCI support" (symbol TULIP) and build it as a module. UDEV detects and loads it just fine.

- John

meson_ray · Post by **meson_ray** » Fri Sep 14, 2007 2:45 pm

I'm using the dc2114x driver, here's the exact kernel configure options:

Code: Select all

[*] "Tulip" family network device support
<M> Early DECchip Tulip (dc2104x) PCI support (EXPERIMENTAL)
<*> DECchip Tulip (dc2114x) PCI support
[*] New bus configuration (experimental)
[*] Use PCI shared mem for NIC registers
[*] Use NAPI RX polling
[*]    Use Interrupt Mitigation
<M> Generic DECchip & Digital EtherWORKS PCI/EISA
<M> Winbond W89c840 Ethernet support
<M> Davicom DM910x/DM980x support
<M> ULi M526x controller support

I'll try recompiling the kernel with the tulip drivers as a module, and see if that works.

*EDIT*

Didn't work, but I noticed that the version of the driver is different on the livecd:

2.6.19 LIVECD kernel: Linux Tulip driver version 1.1.14-NAPI (May 11, 2002)

Installed 2.6.22 kernel: Linux Tulip driver version 1.1.15-NAPI (Feb 27. 2007)

I guess I'll try re-configuring the kernel from scratch for now and see if I accidentally turned something off that needed to be on.

Post by **John R. Graham** » Fri Sep 14, 2007 3:00 pm

For what it's worth, the other difference between my (working) and your (broken) configuration is that all four options for the driver are turned off. The one marked "EXPERIMENTAL" has comments that state that it hasn't been tested on very many cards.

- John

meson_ray · Post by **meson_ray** » Fri Sep 14, 2007 3:11 pm

Thanks for your help

I had those off at one point and when it didn't work I turned them on, but I'll try turning those off before going through the whole config process again.

Post by **John R. Graham** » Fri Sep 14, 2007 11:20 pm

Just for the record, now that I am home, I can report:

Code: Select all

ceres ~ # cd /var/log
ceres log # zgrep Tulip messages.*.gz | head -n1
messages.1.gz:Sep  9 12:21:51 ceres Linux Tulip driver version 1.1.15 (Feb 27, 2007)
ceres log # uname -r
2.6.22-reiser4-r5
ceres log #

It's been stable as a rock here.

- John

meson_ray · Post by **meson_ray** » Sat Sep 15, 2007 1:50 am

Found the problem, it had to do with ACPI and IRQ conflicts. So it's working now.

Thanks for the help!

calculator · Post by **calculator** » Sun Sep 16, 2007 8:23 pm

meson_ray
I have someone problem. How u solved this in detail?

xbmodder · Post by **xbmodder** » Sun Sep 16, 2007 9:15 pm

In general, let's try to support the companies that support us. Buy Intel.

meson_ray · Post by **meson_ray** » Thu Sep 20, 2007 3:12 am

calculator:

For me, it was an IRQ conflict, so I enabled ACPI in the kernel. You can follow the guide in the handbook and look through the options for ACPI support. Hopefully that works for you.

Xbmodder:

It's just a really old card that I had lying around... I can't afford to spend $40 on a new card (when I already have one) just so I can support companies that support linux.