I am using the red hat first beta of red hat 10(soon going bad to 9) but I am trying to figure out why the nvidia drivers won''t install. I ran the setup program out of x and it goes all the way to the point where it is going to build the kernel. But i am thinking it won''t do it since I have a beta and it does''t support it on the driver..although that would seem odd that people with a newer version wouldn''t be able to use nvidia drivers... Here''s what i did: logged in to the command line with telinit 3 under root. Then I did: sh (filename).run Next it put me in their little setup program, i chose accept, tried both find and not to find latest kernel things. Can''t remember excatly what it said, but it I know the second error said : Error Could not build kernel. am i doing something wrong? i didnt even get to the point to change my xfree settings... I am not sure how to get the log to see what happened...anyone know how to do that? Thanks

Is the source (or at least the headers) for your kernel installed?

not sure, how do you check?

^bump^

Anyone know an easy way to install them? I tried to install them by: Downloading the newest drivers, then going into the terminal and logging out of x. Next, I typed: sh (filename) .run . It put me in the driver program and I cilcked accept. Then it tries to build the kernel but comes up with errors.

Thanks

[edited by - HTML on October 5, 2003 12:01:59 PM]

Since you said Red Hat: check if you have a package beginning with "kernel-source" installed (a full example being: kernel-source-2.4.22-1.2087.nptl.i386.rpm). If it is not yet installed and the GUI installation tool doesn''t have an option to do so, check the CDs for a kernel-source package corresponding to the version of the kernel you''re running (you can find this at a terminal with uname -r).

The NVidia drivers only really need the headers, but few distributions (Debian being an exception) offer packages of the kernel headers seperate from the source.

Yup, the kernel-source is installed. It still can''t compile though...

You''re going to kill me, but I saw this same problem on my machine.

I remember it being simple to solve...but I can''t recall (sorry, I deal with so many different machines and issue all day in my work I loose details sometimes). It was on my Severn system.

I remember I ended up having to run the steps manually, if you read the INSTALL or README with the package. I also recall that it was something documented there. I think it might have had something to do with the fact I was running a newer kernel than it expected, so I might have used some force option to bypass the exit error.

Hrms, I found it in my history.

Looks like I ran:

sh NVIDIA-Linux-x86-1.0-4496-pkg2.run --extract-only
cd NVIDIA* # rest of file name
export IGNORE_CC_MISMATCH=yes
export CC=gcc
make
make install

Then I could modprobe nvidia.

I see I vi''d into the Makefile, which isn''t helpful. Maybe if you can post up your error output it''ll fully jog my memory?

Sorry not more helpful. ( I have no idea what exactly fixed the conflict *grin*).

Interim

Assuming it isn''t some really fundamental issue (missing compiler or something), I still think that a kernel header problem would be causing that. Maybe the installer simply can''t find the headers because Red Hat''s package puts them some place weird? From the README:

quote:

---

Q:
I can''t build the NVIDIA kernel module, or I can build the NVIDIA kernel module, but modprobe/insmod fails to load the module into my kernel. What''s wrong?

---

quote:

---

A:
These problems are generally caused by the build using the wrong kernel header files (ie header files for a different kernel version than the one you are running). The convention used to be that kernel header files should be stored in "/usr/include/linux/", but that is deprecated in favor of "/lib/modules/`uname -r`/build/include". The nvidia-installer should be able to determine the location on your system; however, if you encounter a problem you can force the build to use certain header files by using the --kernel-include-dir option. Obviously, for this to work, you need the appropriate kernel header files installed on your system. Consult the documentation that came with your distribution; some distributions don''t install the kernel header files by default, or they install headers that don''t coincide properly with the kernel you are running.

---

Maybe the bolded bit will help?

Actually that may have been the issue, I did dive into the Makefile, odds are I might have altered the include paths. Though I would like to think I read the entire readme. I did see in my history I was checking for the kernel sources.

I used to be in the habit of creating a quick note for each app I custom built, but having switched off regular Unix admin to work on Windows the past few years I''ve gotten lazy. *grin*

Maybe between Null''s and my ideas, you might have enough to tough it out yourself. G''luck.

Interim

quote:

---

Original post by Interim
Looks like I ran:

sh NVIDIA-Linux-x86-1.0-4496-pkg2.run --extract-only
cd NVIDIA* # rest of file name
export IGNORE_CC_MISMATCH=yes
export CC=gcc
make
make install

---

I had to do something like this in SuSE, because the kernel has to be updated with the same version of gcc as it was originally, unless you export IGNORE_CC_MISMATH=yes. What error message exactly are you getting?