[PlanetCCRMA] Nvidia problem

izzy atorima@vfemail.net
Tue Nov 7 00:19:00 2006


izzy wrote:
> Ah, I just noticed this:
>
> kernel-2.6.16-1.2080.16.rdt.rhfc5.ccrma
> kernel-devel-2.6.16-1.2080.16.rrt.rhfc5.ccrma
>
> That just might be the problem.  Perhaps, I should use my glasses!
> Downloading kernel-devel-2.6.16-1.2080.16.rdt.rhfc5.ccrma.i686.rpm 
> <http://ccrma.stanford.edu/planetccrma/mirror/fedora/linux/planetcore/5/i386/kernel-devel-2.6.16-1.2080.16.rdt.rhfc5.ccrma.i686.rpm> 
>
>
> I'll try that and see if it works a little better for me. and maybe, 
> I'll give the edge kernel a try (since I seem to have the sources?)
>
> Thanks!
Ha! Well,  with the correct sources the nvidia installer did indeed 
build the module and install it

So, I rebooted, but...the module failed to load and then... just after 
the login splash the window manager froze. uh oh...

Then I tried to boot into the base kernel. The display was looking 
something like a painting by Pollock or Klee...but not what i wanted.

So I rebooted into run level 3 and had a look a xorg.conf.  The nvidia 
installer had re-written it. Like some bad prose. I tried to edit it.  
But, X wasn't buying it.

I tried to remove 'nvidia' with yum. Bad move. I tried to re-install it 
with nvidia installer. Stupid move.

xorg-system-display-config tried to fix this mess and I got back to the 
login spash and the frozen window manger.  I had had enough. I shut it 
down down. Enough 'work' for tonight.

Did I mention I'm writing this from the ubuntu box in the back room? its 
very...brown. I wonder what kind of latency I could get with this puppy 
and the soundblaster?  Did I mention that the nvidia installer sucks?

I wonder - if I ordered a fedora core 6 cd will planet core 6 will be 
built by the time it arrives?

A Happy thought!  Well, I must sleep now so that I may destroy a new 
tommorrow.

-izzy
> -
>> Brady Bellinger wrote:
>>> It sounds like you don't have a matching -devel kernel rpm installed.
>>> In your second paragraph below, you refer to two different kernels.
>>> To make sure do this:
>>>
>>> $ uname -r
>>>
>>> This will display your current running kernel.  Now see if you have
>>> the corresponding kernel-devel package installed:
>>>
>>> $ rpm -qa | grep kernel
>>>
>>> This will list the currently installed kernel packages.  More than
>>> likely you're a very slight version off.
>>>
>>> Hope it helps,
>>>
>>> Brady
>>
>> Hi,  Brady.
>>
>> # rpm -qa | grep kernel
>> kernel-2.6.17-1.2187_FC5
>> kernel-2.6.16-1.2080.16.rdt.rhfc5.ccrma
>> kernel-2.6.18-1.2200.fc5
>> kernel-devel-2.6.16-1.2080.16.rrt.rhfc5.ccrma
>> kernel-module-alsa-2.6.16-1.2080.16.rdt.rhfc5.ccrma-1.0.11-1.rhfc5.ccrma
>> yum-kernel-module-0.6-3.fc5
>>
>> and I did boot into kernel-2.6.16-1.2080.16.rdt.rhfc5.ccrma when I 
>> ran the installer so, I believe it should have installed for that 
>> kernel.
>>
>> I do have the nvidia -kmod from livna working for 
>> kernel-2.6.18-1.2200.fc5. I've tried un-installing it and then 
>> running the installer for kernel-2.6.16-1.2080.16.rdt.rhfc5.ccrma 
>> just on the off-chance there was a conflict. It made no difference. 
>> You should be able to install for differnet  kernels, yes? (On my 
>> 'old' RH8 you could not) Can't quite figure out why this isn't 
>> working. I keep messing with it.
>>
>> Thanks!
>> -izzy
>>>
>>> On 11/6/06, izzy <atorima@vfemail.net> wrote:
>>>> Fernando,
>>>>
>>>> I ran this again- I tried a couple of different times before and might
>>>> not have had the correct source on the previous attempt-
>>>>
>>>> Here I have kernel-devel-2.6.16-1.2080.16.rrt.rhfc5.ccrma.i686.rpm
>>>> installed and am installing for 2.6.16-1.2080.16.rdt.rhfc5.ccrma
>>>>
>>>> Booting into run level 3, 2.6.16-1.2080.16.rdt.rhfc5.ccrma and running
>>>> "sh NVIDIA-Linux-x86-1.0-8774-pkg1.run"
>>>>
>>>> I get-
>>>>
>>>> ERROR: Unable to find the kernel source tree for the currently running
>>>> kernel.
>>>> Please make sure you have installed the kernel source files for your
>>>> kernel and that they are properly configured; on Red Hat Linux 
>>>> systems,
>>>> for example, be sure you have the 'kernel-source' RPM installed. If 
>>>> you
>>>> know the correct kernel source files are installed, you may specify 
>>>> the
>>>> kernel source path with the '--kernel-source-path' command line 
>>>> option.
>>>>
>>>> If I use -
>>>>
>>>> "sh NVIDIA-Linux-x86-1.0-8774-pkg1.run
>>>> -ľkernel-source-path=/usr/src/kernels/2.6.16-1.2080.16.rrt.rhfc5.ccrma-i686" 
>>>>
>>>>
>>>> (which is where the kernel source is installed, I believe?) I get -
>>>>
>>>> ERROR: Unable to load the kernel module 'nvidia.ko'. This happens most
>>>> frequently when this kernel module was built against the wrong or
>>>> improperly configured kernel sources, with a version of gcc that 
>>>> differs
>>>> from the one used to build the target kernel, or if a driver such as
>>>> rivafb/nvidiafb is present and prevents the NVIDIA kernel module from
>>>> obtaining ownership of the NVIDIA graphics device(s).
>>>>
>>>> I will post the installer log output below.
>>>>
>>>> Your help is always appreciated,
>>>> Thanks!
>>>> - izzy
>>>> > I doubt that the problem is non-matching versions of gcc. I don't 
>>>> think
>>>> > there was any new version updated recently.
>>>> >
>>>> > Do you have the "kernel-devel" package for the kernel you want to 
>>>> build
>>>> > for installed?
>>>> On 11/6/06, izzy <atorima@vfemail.net> wrote:
>>>>> Fernando,
>>>>>
>>>>> I ran this again- I tried a couple of different times before and 
>>>>> might
>>>>> not have had the correct source on the previous attempt-
>>>>>
>>>>> Here I have kernel-devel-2.6.16-1.2080.16.rrt.rhfc5.ccrma.i686.rpm
>>>>> installed and am installing for 2.6.16-1.2080.16.rdt.rhfc5.ccrma
>>>>>
>>>>> Booting into run level 3, 2.6.16-1.2080.16.rdt.rhfc5.ccrma and 
>>>>> running
>>>>> "sh NVIDIA-Linux-x86-1.0-8774-pkg1.run"
>>>>>
>>>>> I get-
>>>>>
>>>>> ERROR: Unable to find the kernel source tree for the currently 
>>>>> running
>>>>> kernel.
>>>>> Please make sure you have installed the kernel source files for your
>>>>> kernel and that they are properly configured; on Red Hat Linux 
>>>>> systems,
>>>>> for example, be sure you have the 'kernel-source' RPM installed. 
>>>>> If you
>>>>> know the correct kernel source files are installed, you may 
>>>>> specify the
>>>>> kernel source path with the '--kernel-source-path' command line 
>>>>> option.
>>>>>
>>>>> If I use -
>>>>>
>>>>> "sh NVIDIA-Linux-x86-1.0-8774-pkg1.run
>>>>> -ľkernel-source-path=/usr/src/kernels/2.6.16-1.2080.16.rrt.rhfc5.ccrma-i686" 
>>>>>
>>>>>
>>>>> (which is where the kernel source is installed, I believe?) I get -
>>>>>
>>>>> ERROR: Unable to load the kernel module 'nvidia.ko'. This happens 
>>>>> most
>>>>> frequently when this kernel module was built against the wrong or
>>>>> improperly configured kernel sources, with a version of gcc that 
>>>>> differs
>>>>> from the one used to build the target kernel, or if a driver such as
>>>>> rivafb/nvidiafb is present and prevents the NVIDIA kernel module from
>>>>> obtaining ownership of the NVIDIA graphics device(s).
>>>>>
>>>>> I will post the installer log output below.
>>>>>
>>>>> Your help is always appreciated,
>>>>> Thanks!
>>>>> - izzy
>>>>> > I doubt that the problem is non-matching versions of gcc. I 
>>>>> don't think
>>>>> > there was any new version updated recently.
>>>>> >
>>>>> > Do you have the "kernel-devel" package for the kernel you want 
>>>>> to build
>>>>> > for installed?
>>>>> >
>>>>> > Looks like the nvidia module is unhappy about one (only one?) 
>>>>> symbol in
>>>>> > the kernel, that is, it would seem the nvidia module that was 
>>>>> built was
>>>>> > not built for the kernel you are booting to.
>>>>> >
>>>>> > -- Fernando
>>>>> >
>>>>> >
>>>>> >