I did it just like that. First put the driver to the proper dir. and then the depmod -a. No luck

nimmer

hi,

i have the same problem with SuSE 8.2 Pro on a P4P800 deluxe (same network chip and same asus driver) ;-)

i haven' find any solution for this problem. Setting mtu value don't change anything (i have tested for all recommended value like 1492, 1472, 1452 ..) and for a lot of other value ... nothing change.

i try to use the syskonnect patch et driver .... nothing change :-(
i send an e-mail to asus tech center ... they don't reply ... :?

sam

On a suse system you have to copy 2 files that are located in the /boot directory to your kernel source dir. Have you done that? It's in the readme of the driver.

J.H.

Yes, i have done that :-(

My first idea when i see that the network doesn't work was to look in this readme file to verify that i done all in the good way.

In order to be sure, i have clean all the system, re-install suse linux, and re-make the network driver ... and the result was the same.

Network access seems to be ok except for some internet sites (google, ... ) :-(

sam

Do you have the same problem with driver version 044?

JH

yes, I have the same problem with driver version 042 and version 044. :-(
Is there any other linux distribution that i can try to test my hardware?? or any other solution?

thanks for your help
sam

Easier solution for now (I am hoping that the user community and ASUS develop a good solution) is to put another NIC in the MoBo, with a driver known to work.

I dual boot my systems, and even under Windows, there are some strange things that happen with the LOM Gigabit NIC. Google is one of the strange things there as well!

I am trying to decode the makefile, and either change the makefile to fit the layout of the file system in my distribution, or change the file system by creating symbolic links to the correct files for the "make load to work"

Here is makefile:

#***************************************************************************
#
# Makefile for the Linux driver for the 3Com Gigabit NIC (3C2000)
# Copyright (C) 2002 3Com Corporation
#
#***************************************************************************

#
# Target file name
#
TARGET = 3c2000.o

#
# Object files
#
OBJECTS = skge.o skaddr.o skgehwt.o skgeinit.o skgepnmi.o skgesirq.o \
ski2c.o sklm80.o skqueue.o skrlmt.o sktimer.o skvpd.o skxmac2.o \
skproc.o skcsum.o

#
# Compiler
#
CC = gcc

#
# Compiler flags
#

CFLAGS = -D__KERNEL__
#CFLAGS += -I/lib/modules/$(shell uname -r)/build/include
CFLAGS += -Wall
CFLAGS += -Wstrict-prototypes
#CFLAGS += -Wno-trigraphs
CFLAGS += -O2
#CFLAGS += -fno-strict-aliasing
#CFLAGS += -fno-common
#CFLAGS += -fomit-frame-pointer
#CFLAGS += -pipe
#CFLAGS += -mpreferred-stack-boundary=2
#CFLAGS += -march=i686
CFLAGS += -DMODULE
#CFLAGS += -DMODVERSIONS
CFLAGS += -I.
CFLAGS += -DSK_USE_CSUM
#CFLAGS += -nostdinc

#
# Include directory
#
ifneq (,$(wildcard /usr/src/linux))
INCLUDE_DIR = /usr/src/linux/include
else
INCLUDE_DIR = /usr/src/linux-2.4/include
endif

CFLAGS += -I $(INCLUDE_DIR)

#
# MODVersions setting
#
ifneq (,$(wildcard $(INCLUDE_DIR)/linux/modversions.h))
CFLAGS += -include $(INCLUDE_DIR)/linux/modversions.h
endif
CFLAGS += $(shell [ -f $(INCLUDE_DIR)/linux/modversions.h ] && echo -DMODVERSIONS)

#
# Version file
#
VERSION_FILE = $(INCLUDE_DIR)/linux/version.h

# Get the kernel version from version.h
KERNEL_VER = $(shell gcc -E -dM $(VERSION_FILE) | grep UTS | awk '{ print $$3; }' | sed -e 's/\"//g')

ifneq (,$(wildcard /lib/modules/$(KERNEL_VER)/kernel))
INSTALL_DIR = /lib/modules/$(KERNEL_VER)/kernel/drivers/net
else
INSTALL_DIR = /lib/modules/$(KERNEL_VER)/net
endif

#
# SMP Configuration
#
CONFIG_FILE = $(INCLUDE_DIR)/linux/autoconf.h
SMP_FLAG = $(shell gcc -E -dM $(CONFIG_FILE) | grep CONFIG_SMP | awk '{print $$3; }')

ifeq ($(SMP_FLAG), 1)
CFLAGS += -D__SMP__
endif

#
# Target generation
#
$(TARGET) : $(OBJECTS)
ld -r -o $(TARGET) $(OBJECTS)

skge.o : skge.c h/skversion.h h/skdrv1st.h h/skdrv2nd.h

skaddr.o : skaddr.c h/skdrv1st.h h/skdrv2nd.h

skgehwt.o : skgehwt.c h/skdrv1st.h h/skdrv2nd.h

skgeinit.o : skgeinit.c h/skdrv1st.h h/skdrv2nd.h

skgepnmi.o : skgepnmi.c h/skdrv1st.h h/sktypes.h h/xmac_ii.h h/skdebug.h \
h/skqueue.h h/skgepnmi.h h/skgesirq.h h/skcsum.h h/skvpd.h \
h/skgehw.h h/skgeinit.h h/skdrv2nd.h h/skgepnm2.h skgemib.c

skgesirq.o : skgesirq.c h/skdrv1st.h h/skgepnmi.h h/skrlmt.h h/skdrv2nd.h

ski2c.o : ski2c.c h/skdrv1st.h h/lm80.h h/skdrv2nd.h

sklm80.o : sklm80.c h/skdrv1st.h h/lm80.h h/skdrv2nd.h

skqueue.o : skqueue.c h/skdrv1st.h h/skqueue.h h/skdrv2nd.h

skrlmt.o : skrlmt.c h/skdrv1st.h h/skdrv2nd.h

sktimer.o : sktimer.c h/skdrv1st.h h/skdrv2nd.h

skvpd.o : skvpd.c h/skdrv1st.h h/sktypes.h h/skdebug.h h/skdrv2nd.h

skxmac2.o : skxmac2.c h/skdrv1st.h h/skdrv2nd.h

skproc.o : skproc.c h/skdrv1st.h h/skdrv2nd.h

skcsum.o : skcsum.c h/skdrv1st.h h/skcsum.h h/skdrv2nd.h h/sktypes.h \
h/skqueue.h

#
# Cleanup
#
.PHONY : clean
clean :
-rm -f $(TARGET) $(OBJECTS)

#
# Module installation
#
.PHONY : install
install: $(TARGET)
@echo "Found kernel version $(KERNEL_VER)"
@echo "Install directory is $(INSTALL_DIR)"
mkdir -p $(INSTALL_DIR)
install -m 644 $(TARGET) $(INSTALL_DIR)
depmod -a

#
# Module removal
#
.PHONY : remove
remove:
@echo "Removing Module from $(INSTALL_DIR)"
rm -f $(INSTALL_DIR)/$(TARGET)
depmod -a

#
# Module loading
#
.PHONY : load
load: $(TARGET)
insmod $(TARGET)

and the source is located in /usr/src, there is no Linux or Linux 2.4 directory.

Any suggestions (Libranet Distro)

i know i know, ... i have a peabird based on realtek 8139 100Mbps chip; and with 5 free pci slot i think it was not a problem to use it! lol

But i doesn't like this solution for 2 reasons:

1) i have payed for this gigabit chip ... so i want to use it :-p
2) i can't know if the chip is working fine or if it has a big default :-(

I thought that Asus or 3com was a little more professionnals ... :-/

any other solutions ? ;-)
thanks for help
sam

hi,

sam_00;
i don't think there is a solution for now, the only thing i think we can do is to test every distro or modules we can, when a wide selection will done we can said if this is a module bug or a hardware issue.

for now i heavely think it's a module bug because of the problem is not present under win$ and we all have different mobo (p4p or p4c, different revisions ...)

By the way, talking about it is a good thing, the problem is not really obvious to discover and as finegan i think we search a lot before thinking it was the network chip more than us .

for now i tested under :
distro; gentoo 1.4rc4, redhat9, knoppix 3.1
with modules; 42,44 and the kernel 2.4.20 module
with or without the patch explained on my first post (if necessary)

with a p4p800DLX and a p4c800DLX (no rev for now)

This afternoon i'll test with a mandrake (yeak :/) 9.1 because the x86-secret.com said it work with... i want to verify it, if rh9, suse and gentoo don't work i don't understand why it could under a crappy mandrake.

And the funny thing is that I have not yet had any problem with a site like google or anyone for that matter. The only thing I had problems with was a system inside my local lan. But that could be because the P4C800 Deluxe system is inside the lan and doesn't have a public IPv4 address. And I don't know if the problem was with my system or with that other system because that other system is down now, and I don't know the reason yet.

JH

roscop>

i understand ...

i have tested this under 3 linux distributions:

- Debian 3.0r1 (classic boot or bf2.4)
- Redhat 9.0
- Suse 8.2 Pro

i think i will try with mandrake 9.1 this night ... but like you i don't understand how it could work fine with mandrake and not with suse or redhat ....

Sam

Hi,

As i fear, it don't work more with a mandrake 9.1 and asus module 044...
I took the mandrake kernel sources and the modules and only build the module.

It don't solve our problem, the module is well maked and loader, network configured , but always no responses from google or others.

Hi,

A possible answer on x86-secret is to have a kernel vanilla (the original one from www.kernel.org) 2.4.21 OR higher , 2.4.22pre4 was also succefully tested for them.

WARNING i remember it always work for them ! on mandrake or redhat... So don't consider this information like a solution until other people confirm it, anyway you can test it to report result

I've tested on my gentoo with the vanilla 2.4.21 (used the 3c2000 driver from asus, the sk98linux from kernel refuse to insert ) and it don't work anymore :/ (same problem with www.google.Fr and other) even if i try to tweak the mtu.
Someone also tested a 2.4.21 kernel or newer ?

I'd like somes responses from other people to be sure it's not me or my configuration.

Thank by advance.

hi all,

finally it work !
After failed with vanilla 2.4.21 i remembered the 2 links given by jhaprins on up for kernel 2.4.20 and 2.4.21. And with the patch for 2.4.21 on vanilla it work !!! (you must use the sk98lin module in the kernel, not the 3c2000 from asus)

For the other maybe, i don't tested. For gentoo-sources 2.4.20-r5 the patch fail (many hunk couldn't be applied). I added successfully the grsecurity patch on vanilla, use iptables and all of what i need in kernel.
No problem so far.
BTW the mtu still at default value.
Tested on p4p800 and p4c800, thank jhaprins, i'm really irritated to don't tried soon the vanilla with this patch :/

Nice to hear that this patch really worked for you. I allready send a mail to redhat if they could insert this patch into there next kernel.

JH