Tried Mandrake, Suse and RH(

[jimscafe]

I am a real newbie to Linux, trying, like many people, to move away from Windows.

My PC is the following

Samsung Syncmaster 151BM LCD monitor

I have the Epox 8RDA motherboard with onboard network card, sound, USB and Fire Wire (it also has graphics but I have also installed a separate gforce graphics adapter).

ASUS CRW-4824A CD
GForce MX400
Logitech optical USB mouse (changed to using standard port as USB option could not be made to work)
Logitech keyboard
80GB hard drive in two partitions (before Linux) Win 98 and Win XP)

First I tried Mandrake 8.1 (I live in the Philippines so getting the latest version is not easy) but I did get it to work at 600x800 reslotion but then it hung when I quit - sent en email to Mandrake support and it was as if the guy hadn't even read my email - simply said why aren't you using the latest version. Decided on the quality of response to try Suse - so had that shipped from the US (version 8.2)
Suse also could not do anything with my LCD screen. Also the dual boot did not work, I seemed to have lost my windows.
Someone at work got windows working again and installed RedHat 9 for me, and this detected a generic LCD at the correct resolution - but having plenty of problems trying to install the drivers for the netwrok coard etc.

The question - I would like to install Suse - from what I have read it seems the system I would really like to make work. The safest thing for me to do would be to create another Linux partition and install it there - but then I might lose the information in the Linux dual boot system for RedHat and not be able to work out how to get Windows back (without a network card operating windows in the only means I have of getting onto the internet).

There must be a config file somewhere for RedHat which contains the information I would need to install into the Suse boot loader - maybe manually.
Then I need to get Suse to work with my monitor at 1024x768 (it sucks at 600x800) - so maybe I need to set this up manually and again RedHat presumably has the correct infromation somewhere.

My other option is to overwrite RedHat (which is at least working so I can start to look at using Python programming on Linux - which is my immediate aim).

The guys at work seem to use a brute force approach to solving problems (i.e. rather than figure out what was wrong with Suse they just installed Red Hat) so I would like to solve these problems myself to better understand what I am doing.

In addition, one option on the nvidia website suggests rebuilding an rpm with the command rpm --rebuild, even though this command is mentioned in my book (Linux in a Nutshell) the response is --rebuild not a recognised command.

I have heard it takes 6 months to get Linux working correctly and to understand enough to get the best out of the system - so my clock has started and I have 5 months to go!!
:rolleyes:

[ehawk]

Knoppix is famous for autodetection of hardware. Maybe it will make your life easier....I think it should take much less than 6 months....6 days maybe.

[ehawk]

Winlinux is also an option. It is installed using a loopback filesystem within Windows....wizard and everything, just like any other Windows application. You wouldn't have to worry at all about hardware, since it is already recognized by windows. Winlinux is supposed to be 100% compatible with Red Hat.

[JusKickNit]

Sorry I couldn't be more help. In order to rebuild rpms the command is rpm-build. rpm --rebuild was used a in previous versions.

[DMR]

Originally posted by jimscafe
but having plenty of problems trying to install the drivers for the netwrok coard etc. What network driver are you trying to install? From what I've read, you should be using the sis900 driver (as weird as this seems considering that the motherboard specs list a Realtek chipset). However, I've also read that there are some problems with certain versions of the drive. With that in mind, try to manually load the driver with the following command and see what happens:

modprobe sis900

Originally posted by jimscafe
There must be a config file somewhere for RedHat which contains the information I would need to install into the Suse boot loader In redhat, the bootloader configuration file are:

if you use Grub: /boot/grub/grub.conf
If you use lilo: /etc/lilo.conf

[jimscafe]

OK got some results.

modeprobe sis900
No such device

rpm-build
Command not found

rpmbuild NVIDIA_nforce etc....
File does not appear to be a spec file

rpmbuild --rebuild NVIDIA_nforce.......src.rpm

Works but produces a lot of errors e.g.
warning user buildmeiser does not exist - using root

and ends with

and exit status from /var/tmp/rpm-tmp.64324(%build)

I haven't tried knoppix yet!!

:(

[mdwatts]

Originally posted by jimscafe
OK got some results.

modeprobe sis900
No such device

rpm-build
Command not found

rpmbuild NVIDIA_nforce etc....
File does not appear to be a spec file

rpmbuild --rebuild NVIDIA_nforce.......src.rpm

Works but produces a lot of errors e.g.
warning user buildmeiser does not exist - using root

and ends with

and exit status from /var/tmp/rpm-tmp.64324(%build)

I haven't tried knoppix yet!!

:(

Is that modeprobe or modprobe and are you using root to load the module. If no such device, ensure you have PnP OS disabled in the bios.

Are you root when using rpmbuild as you need root authority as rpmbuild uses the directory structure in /usr/src/<distro name>/... to build the binary rpm package.

[DMR]

Originally posted by jimscafe
OK got some results.

modeprobe sis900
No such device Ok, here's another option:

Your motherboard seems to use NVidia's NForce2 chipset, which is comparatively new and not fully supported in Linux yet. Different kernel versions will have varying degrees of NForce2 support; some of which is only available via kernel patches. NVidia has released their own bundle of Linux network, sound, etc. drivers for the NForce2 which you might want to try. You can read the documentation and download the drivers here:

http://www.nvidia.com/object/linux_nforce_1.0-0248.html

[jimscafe]

Sorry I typed modprobe not modeprobe

Yes I logged on as root.

I have tried the nvidia website drivers.

Tried to use the rpm - said I had the wrong kernel, they seem to support 2.4.20.6 and 2.4.20.9 - I have 2.4.20.8 !!!

Then tried to use the src.rpm - which I have mentioend in this thread.

I also tried the source tar.gz file -which I unzipped OK but when I typed make came up with all kinds of errors (I posted my results on the nvidia forum but so far no replies. (At least getting some help here - thank you)

As for the PNP OS, the bios settings seem more complicated and I think I have disabled it but not sure. In simpler motherboards the setting is an easy one but for this motherboard it seems there are many more options. Maybe I should try setting different things in the PNP section. But there just seems so many options right now it is most confusing.


Seriously thinking of downloading the kernel 2.4.20.9 and reinstalling Red Hat just to get the correct rpm from nvidia, but I really felt that I should learn how to do the other stuff also (i.e. from source etc.)

More suggestions, comments please. (I even considered throwing my motherboard away and buying another easier one - what use is the latest if it doesn't work) What I don't understand is why all netowrk cards are not also NE2000 compatible - so that they can be made to work in a basic way if all else fails - ditto video cards, sound, etc. - there must be a reason I guess. I can also diasble the onboard network card and buy another, they are inexpensive enough. (But then presumably my sound and firewire would not work also!!)

[DMR]

Originally posted by jimscafe
I have tried the nvidia website drivers.

Tried to use the rpm - said I had the wrong kernel, they seem to support 2.4.20.6 and 2.4.20.9 - I have 2.4.20.8 !!!

Then tried to use the src.rpm - which I have mentioend in this thread. Sorry about that- I missed the fact that we were refering to the same file.

Originally posted by jimscafe
I also tried the source tar.gz file -which I unzipped OK but when I typed make came up with all kinds of errorsWhat errors exactly? You might not have all of the required source/header files installed.

Originally posted by jimscafe
Maybe I should try setting different things in the PNP section. But there just seems so many options right now it is most confusing.It would be the PNP option which refers to operating systems. In some BIOSes it's similar to "PNP OS? Y/N", in others there will be an "operating system" choice where you should choose "other".


Originally posted by jimscafe
I don't understand is why all netowrk cards are not also NE2000 compatible - so that they can be made to work in a basic way if all else fails - ditto video cards, sound, etc. - there must be a reason I guess. It's all about the chipsets. Drivers are so closely tied to their specific chipsets that chip manufacturers would have to agree on some common baseline circuit structure which would support a common driver. This would be a Bad Thing however, because having to maintain tight integration with lowest-common-denominator circuitry would almost certainly hinder the implementation of design improvements or the addition of new features at some point down the road.

With sound and video cards you have a better chance of finding a "generic" driver than with, say, network cards. Since SoundBlaster became the de-facto standard for sound cards, there are a wide range of cards which are "SoundBlaster compatible". Because there are vendor-independent video standards, you can almost always get a video card to work at some level by using a plain-vanilla VGA or SVGA driver.

[jimscafe]

The errors were this and many more :-

[root@localhost nforce]# make
make -C nvnet
make[1]: Entering directory `/home/test/nforce/nvnet'
cc -c -Wall -DLINUX -DMODULE -DEXPORT_SYMTAB -D__KERNEL__ -O -Wstrict-prototypes
-DCONFIG_PM -fno-strict-aliasing -mpreferred-stack-boundary=2 -march=i686 -falign-functions=4 -DMODULE -I/lib/modules/2.4.20-8/build/include nvnet.c
In file included from /usr/include/linux/module.h:20,
from nvnet.h:20,
from nvnet.c:21:
/usr/include/linux/modversions.h:1:2: #error Modules should never use kernel-headers system headers,
/usr/include/linux/modversions.h:2:2: #error but rather headers from an appropriate kernel-source package.

It does look like some header files are missing.

I will check on the exact options available in the BIOS

[jimscafe]

My BIOS settings in the PNP section are :-

Reset Configuration Data - Disabled
Resources Controlled By [Auto(ESCD)]
PCI/VGA Palette Snoop [Disabled]
Assign IRQ for VGA [Enabled]
PCI Latency Time [32]
Int Pint 1 Assignment [Auto]
Int Pint 2 Assignment [Auto]
Int Pint 3 Assignment [Auto]
Int Pint 4 Assignment [Auto]

Should I change any of these ?

[DMR]

Originally posted by jimscafe
Should I change any of these ? Nope, none of those are it. It looks like your BIOS doesn't have the option.

[mdwatts]

Originally posted by jimscafe

/usr/include/linux/modversions.h:1:2: #error Modules should never use kernel-headers system headers,
/usr/include/linux/modversions.h:2:2: #error but rather headers from an appropriate kernel-source package.


You do not have the kernel-source package installed for your current running kernel version.

Search the JL forums for 'redhat kernel-source' and you will find lots of previous threads that explain all you need to know about finding and installing the correct kernel-source package.

Redhat even has documentation on their site on the same subject.

[swiftnet]

Why don't you just use the NVidia proprietary drivers instead of the rpms?

[jimscafe]

Why don't you just use the NVidia proprietary drivers instead of the rpms

I don't understand the above comment - I thought I was trying to install the drivers from Nvidia - they come as an rpm, src.rpm and tar - hree versions - is there another way to install Nvdidia drivers.


A guy at work gave me an rpm file that will upgrade the kernel to 2.4.20-9 for which Nvidia have an rpm driver file. So I installed that (I now get two Linux options when I boot - 20-8 and 20-9), and the Nvidia driver rpm installed without a problem. My network card was detected and I can now try to get my DSL hooked up - then I have access to the internet and do not have to rely upon Win XP for that.

I appreciated the comment about not having the kernel source package - I may well need that for a future installation.

Thank for all the support guys.

[swiftnet]

NVidia has created a script that checks your system, goes to their website, downloads and installs the driver.

Here is the link. You'll have to be in a text environment to install it and you will need to be root.

http://download.nvidia.com/XFree86/Linux-x86/1.0-4496/NVIDIA-Linux-x86-1.0-4496-pkg2.run

After the download, open up a console, su into root, type telinit 3, at the login prompt login as root and run the script : . /script-name

[jimscafe]

Errr.. until I get the netowrk card working I cannot connect to the internet!!

Currently to access this forum I have to boot into XP.

[phlipant]

why don`t you post the file /proc/ioports to see if the nic is recognized. it is possible that, the nic is recognized but not properly configured, or that the module is not found or loaded.

[phlipant]

since, i need to go, i though i might expand a bit. if i type cat /proc/ioports, i will find 4 interesting lines.

a800-a81f : Realtek Semiconductor Co., Ltd. RTL-8029(AS)
a800-a81f : ne2k-pci
ac00-acff : Linksys Network Everywhere Fast Ethernet 10/100 model NC100
ac00-acff : tulip

the first two lines tell me i have a realtek 8029 with the ne2k-pci driver and the second two lines tell me i have a linksys nc100 card with the tulip driver. if i then examine /etc/modules.conf, i find two lines

alias eth0 ne2k-pci
alias eth1 tulip

telling me that the realtek(ne2k-pci) is the first card, and the linksys(tulip) is the second card.

if you find this condition, the kernel has installed your card but it is not configured. configuring your card may involve other issues(i.e. dhcp client vs static ip).

another interesting inquiry would be a command like grep eth /var/log/console. where i get the response.

Jul 29 09:59:17 philipmarsh kernel: divert: allocating divert_blk for eth0
Jul 29 09:59:17 philipmarsh kernel: eth0: RealTek RTL-8029 found at 0xa800, IRQ 11, 00:E0:29:53:BE:5C.
Jul 29 09:59:17 philipmarsh kernel: divert: allocating divert_blk for eth1
Jul 29 09:59:17 philipmarsh kernel: eth1: ADMtek Comet rev 17 at 0xf0936000, 00:04:5A:6B:A8:17, IRQ 5.

note: for the second card, i get a different response, that being the card manufacturer vs chip designer.


this information tells me that, the kernel recognizes the card and the hardware is functional, this does not tell me if the card is properly configured to talk to my isp.

[swiftnet]

Did you get all the drivers for the nforce2 chipset from NVidia?

[jimscafe]

I am up and running, nvidia drivers installed on my Suse installation.

My next step was to connect to the internet using my DSL line - I was apprehensive and delayed attempting it for some days - but it was a breeze, about 30 seconds and it was working - this is my first transmission from a Linux OS on the internet!! It feels great - I feel like it is Christmas!!

Just wanted to say thankyou to all the help I received and that I could easily have given up and gone back to W!!! if it were not for the support I got (and I know I will get in the future) on this forum.

To other newbies - these guys will help you get there if you stick it out.
:)

(I hope you can handle all this praise!! I know you are all modest people ;) )