SEO

Hello,

I just installed Debian 4.0 for an AMD-64 processor. Unfortunately, I
cannot set the screen resolution to something else than 640x480, on 60Hz.
My video card is a GeForce 7100 GS; capable of much more than only basic
vga....

I tried to download a driver from nvidia.com, but then I need to
recompile a lot of things. libc header files are missing, ect. Just a
little bit too complicated for a starting linux user

Does anybody have a suggestion how to set the higher screen resolutions?

Thanx in advance for any hint!

/Eilko Nijboer.

On Wed, 30 Apr 2008 17:59:23 +0200, Eilko wrote:

> I just installed Debian 4.0 for an AMD-64 processor. Unfortunately, I
> cannot set the screen resolution to something else than 640x480, on
> 60Hz. My video card is a GeForce 7100 GS; capable of much more than only
> basic vga....

Very well supported by nVidia's drivers.

> I tried to download a driver from nvidia.com, but then I need to
> recompile a lot of things. libc header files are missing, ect. Just a
> little bit too complicated for a starting linux user

Why in hell would you ever think that Debian is suitable for a complete
novice?

apt-get install build-essential
apt-get install linux-headers
apt-get install linux-kbuild

> Does anybody have a suggestion how to set the higher screen resolutions?

Install the nVidia drivers.

On Wed, 30 Apr 2008, Eilko wrote:

> Hello,
>
> I just installed Debian 4.0 for an AMD-64 processor. Unfortunately, I cannot
> set the screen resolution to something else than 640x480, on 60Hz.
> My video card is a GeForce 7100 GS; capable of much more than only basic
> vga....
>
> I tried to download a driver from nvidia.com, but then I need to recompile a
> lot of things. libc header files are missing, ect. Just a little bit too
> complicated for a starting linux user
>
Well no, your problem is that you decided you knew what you needed, and
then did a selective install.

Unless someone has an actually good reason, there is no reason to do a
selective install nowadays. Hard drives are so big that the space saved
from a limited install will be next to nothing.

But, a selective install requires that you know what you are doing. Guess
wrong and you end up going through these iterations.

I did a partial install the first time, in late 2000, but then it was a
240meg hard drive. I realized immediately the futility of it all. Ever
since, I've had a large enough hard drive to do a full install, whether
it was that 2gig that I got in mid-2001 specifically to run Linux, or
the 20gig or later 160gig I've upgraded to. Always more space than I
needed.

I've been using only Linux since mid-2001. When I moved to the latest
version of my distribution a couple of months ago, I did a selective
install, and I left out things I thought were irrelevant, we're talking
relatively minor things. Yet it did cause problems, and it took effort
to figure out where the problem lay. If I'd just done a full install,
which requires no effort on the user's part, that problem would not
arise.

When I've installed the nvidia driver, I've not ever had a problem,
it just compiles and there it is. It actually is simple, so long
as you did a full install and everything expected is there.

Michael

On Wed, 30 Apr 2008 14:27:32 -0400, Michael Black wrote:

> Well no, your problem is that you decided you knew what you needed, and
> then did a selective install.

Not really. It is the Debian developers who have decided what the
install is to be and there is little choice for the end user. The end
user gets to choose which *servers* are to be installed but not the
headers required to build the nVidia driver nor a compiler.

Dave Uhring wrote:
> On Wed, 30 Apr 2008 14:27:32 -0400, Michael Black wrote:
>
>> Well no, your problem is that you decided you knew what you needed, and
>> then did a selective install.
>
> Not really. It is the Debian developers who have decided what the
> install is to be and there is little choice for the end user. The end
> user gets to choose which *servers* are to be installed but not the
> headers required to build the nVidia driver nor a compiler.

I think you're right.

In the meantime I executed the build-essential, linux-headers and
linux-kbuild commands; and that works. That is, there are many linux
headers; I choose the amd-x64 variant.

But I still fail to install the drivers.. When I installed Debian,
I entered my ip-address, the subnet and the gateway, and a number of dns
servers, and it al works when I login under my user account. But
installing nVidia is blocked because X is running.
When I login as a single-user (via the Grub menu), X is not started, but
the network doesn't work either! So I can start installing the drivers
but during that process it wants to download additional headers from
nvidia.com and that fails...

My Debian learning curve isn't as steep as it should be, I need another
hint... help?

Thanx in advance..
/Eilko.

Eilko <nijboer@xs4all.nl> wrote:
> Dave Uhring wrote:
>> On Wed, 30 Apr 2008 14:27:32 -0400, Michael Black wrote:
>>
>>> Well no, your problem is that you decided you knew what you needed, and
>>> then did a selective install.
>>
>> Not really. It is the Debian developers who have decided what the
>> install is to be and there is little choice for the end user. The end
>> user gets to choose which *servers* are to be installed but not the
>> headers required to build the nVidia driver nor a compiler.
>
> I think you're right.
>
> In the meantime I executed the build-essential, linux-headers and
> linux-kbuild commands; and that works. That is, there are many linux
> headers; I choose the amd-x64 variant.
>
> But I still fail to install the drivers.. When I installed Debian,
> I entered my ip-address, the subnet and the gateway, and a number of dns
> servers, and it al works when I login under my user account. But
> installing nVidia is blocked because X is running.
> When I login as a single-user (via the Grub menu), X is not started, but
> the network doesn't work either! So I can start installing the drivers
> but during that process it wants to download additional headers from
> nvidia.com and that fails...
>
> My Debian learning curve isn't as steep as it should be, I need another
> hint... help?

boot into single user as you were doing..
And start the network.
/etc/init.d/network start should do the trick.
THEN run the nvidia-installer

Just because you booted into a runlevel without a facility, doesn't mean you
can't start that facility yourself.

One thing that always bugged me about debian was their stupid idiotic init.d
scheme. In other more sensible distros, the runlevels have meaning, 1 is
single user, 2 is multiuser, 3 is multiuser with network, 5 is multiuser
with network and display manager.

But debian fecked that idea up something rotten and put the display manager
in every single runlevel apart from the single user mode.

You can correct this fatal and moronic mistake by going into /etc/rc3.d and
deleting the links to kdm, gdm and xdm...

then you could start up in runlevel 3 to do maintenance tasks that require
the use of basic system functions like networking without all that faffing
about in single user mode.
--
| spike1@freenet.co.uk | Windows95 (noun): 32 bit extensions and a |
| | graphical shell for a 16 bit patch to an 8 bit |
| Andrew Halliwell BSc | operating system originally coded for a 4 bit |
| in |microprocessor, written by a 2 bit company, that|
| Computer Science | can't stand 1 bit of competition. |

On Thu, 01 May 2008 08:59:45 +0200, Eilko wrote:

> But I still fail to install the drivers.. When I installed Debian,
> I entered my ip-address, the subnet and the gateway, and a number of dns
> servers, and it al works when I login under my user account. But
> installing nVidia is blocked because X is running.

You can kill X while keeping the network up. See below.

> When I login as a single-user (via the Grub menu), X is not started, but
> the network doesn't work either! So I can start installing the drivers
> but during that process it wants to download additional headers from
> nvidia.com and that fails...

None of that makes any sense. As the unprivileged user you download the
driver from nVidia. The file name will be NVIDIA-Linux-x86_64-169.12-pkg2.run
Then press three keys simultaneously, ALT-CTRL-F2. That will take you to
a console session and you log in as root. Alternatively, you can skip
that step and su to root and...

Execute '/etc/init.d/gdm stop'. Now change directory (cd) to where you
downloaded the nVidia driver and execute

sh NVIDIA-Linux-x86_64-169.12-pkg2.run -a

You will need to answer a few Yes or No questions; read the screens and
use a little common sense.

After the driver and libraries are installed, '/etc/init.d/gdm start'
will bring up the login screen again following the nVidia splash screen.

I demand that Andrew Halliwell may or may not have written...

[snip]
> One thing that always bugged me about debian was their stupid idiotic
> init.d scheme. In other more sensible distros, the runlevels have meaning,
> 1 is single user, 2 is multiuser, 3 is multiuser with network, 5 is
> multiuser with network and display manager.

> But debian fecked that idea up something rotten and put the display manager
> in every single runlevel apart from the single user mode.

Alternatively, everybody else messed it up by not including networking for
runlevels 1 and 2.

> You can correct this fatal and moronic mistake by going into /etc/rc3.d and
> deleting the links to kdm, gdm and xdm...

What mistake?

(BTW, I use none of the above, but still have a display manager.)

> then you could start up in runlevel 3 to do maintenance tasks that require
> the use of basic system functions like networking without all that faffing
> about in single user mode.

OTOH, start up in single-user mode, faff around a bit, press ^D to resume
booting. Works nicely, and doesn't have things like apache running (where
runlevel 3 probably will). :-)

(My laptop's in runlevel 1 as I write this, providing routing for another box
whose wireless networking basically isn't. All that it took was to bring up
eth0, set up a few firewall rules and enable IP forwarding; wlan0 was brought
up in runlevel S.)

--
| Darren Salt | linux or ds at | nr. Ashington, | Toon
| RISC OS, Linux | youmustbejoking,demon,co,uk | Northumberland | Army
| Kill all extremists!

The two most common elements in the universe are hydrogen and stupidity.

Dave Uhring wrote:
> On Thu, 01 May 2008 08:59:45 +0200, Eilko wrote:
>
> Execute '/etc/init.d/gdm stop'. Now change directory (cd) to where you
> downloaded the nVidia driver and execute
>
> sh NVIDIA-Linux-x86_64-169.12-pkg2.run -a
>
> You will need to answer a few Yes or No questions; read the screens and
> use a little common sense.

Thanks, I made a lot of progress. I think I'm almost finished

I found out that I needed to install the kernel sources first.
I found out that they had to be unpacked manually.
I found out that I need to tell the nvidia script where to find the
sources. I know that this works, because previously the kernel.h could
not be found, but now the script doesn't complain about kernel.h
anymore. So most likely it found that headerfile.

But the latest error report of nvidia seems to be more tricky.
It claims that it cannot find
/usr/src/linux-source-2.6.18/include/linux/version.h which is true
because that file doesn't exist. Similar files do exist on different
places in the source tree, but they have a different meaning..

On nvidia.com it was advised to use the script parameter
--kernel-include-dir but that parameter is not recognized by the script.

Am I making progress in the right direction? Or am I making a stupid
fundamental error? How do I correct the version.h error?

/Eilko.

p.s.
One thing is for sure, Debian is not suitable for a complete novice.
Yesterday I was a complete novice, today I'm just a novice.. But I like
the learning process And thanks for your help!

On Thu, 01 May 2008 19:37:36 +0200, Eilko wrote:

> I found out that I needed to install the kernel sources first.

Not really necessary.

> I found out that they had to be unpacked manually.

Yup.

> I found out that I need to tell the nvidia script where to find the
> sources. I know that this works, because previously the kernel.h could
> not be found, but now the script doesn't complain about kernel.h
> anymore. So most likely it found that headerfile.
>
> But the latest error report of nvidia seems to be more tricky.
> It claims that it cannot find
> /usr/src/linux-source-2.6.18/include/linux/version.h which is true
> because that file doesn't exist. Similar files do exist on different
> places in the source tree, but they have a different meaning..

I'm running Solaris ATM and can't tell you exactly where that file is but
I think that you will find it in under /usr/src/linux-headers-2.6.18.x

Create a symbolic link in /usr/src

ln -s linux-headers-2.6.18.x linux

Then the nVidia installer script should be able to find it.

The other option is to copy /boot/config-2.6.18-5-686 to
/usr/src/linux-source-2.6.18/.config, then

cd /usr/src/linux-source-2.6.18
make oldconfig
make bzImage

and hit Control-C after a page of output scrolls by.

Then change your symlink

cd ../
ln -sf linux-source-2.6.18 linux

> On nvidia.com it was advised to use the script parameter
> --kernel-include-dir but that parameter is not recognized by the script.

It should be; more likely you invoked it incorrectly.

> Am I making progress in the right direction? Or am I making a stupid
> fundamental error? How do I correct the version.h error?

You're doing OK. The header files which the nVidia script failed to find
are probably present where I said. Take a look and see.

> One thing is for sure, Debian is not suitable for a complete novice.
> Yesterday I was a complete novice, today I'm just a novice.. But I like
> the learning process And thanks for your help!

AFAIK the only OS which comes with the nVidia drivers installed is
Solaris, which is not "free" in the Stallman sense but is available for
free download and use.

http://www.opensolaris.org/os/downloads/

I'm currently using the Solaris Express Community Edition.

Now it may be necessary for you to create a config file for X if one does
not already exist or modify the one which is there at /etc/X11/xorg.conf.
Get your driver installed and post again if you still can't get things
working correctly.