philip at turmel
Feb 15, 2012, 5:28 PM
Post #4 of 9
(2269 views)
Permalink
|
Hi Frank,
On 02/15/2012 07:06 PM, Frank Peters wrote:
> After updating nvidia-drivers from 290.10 to 295.20-r1, I cannot start X
> and the log files contain this message:
>
> (EE) NVIDIA(0): Failed to initialize the NVIDIA kernel module. Please see the
> [ 3482.609] (EE) NVIDIA(0): system's kernel log for additional error messages and
> [ 3482.609] (EE) NVIDIA(0): consult the NVIDIA README for details.
> [ 3482.609] (EE) NVIDIA(0): *** Aborting ***
> [ 3482.609]
>
> NVRM: API mismatch: the client has the version 295.20, but
> NVRM: this kernel module has the version 290.10. Please
> NVRM: make sure that this kernel module and all NVIDIA driver
> NVRM: components have the same version.
> NVRM: API mismatch: the client has the version 295.20, but
> NVRM: this kernel module has the version 290.10. Please
> NVRM: make sure that this kernel module and all NVIDIA driver
> NVRM: components have the same version.
>
> What could be the problem? I've seen no other reports of
> this issue.
The nvidia-drivers ebuild follows the /usr/src/linux symlink to
find the kernel source to match the module to. It then builds
the Nvidia module and installs it in that kernel version's
part of the /lib/modules/ tree. If you aren't running that
specific kernel, you won't get the new module.
Run "find /lib/modules/ -name nvidia.ko -ls" to see what I mean.
The simplest solution is to replace the symlink, pointing it
at the kernel source tree for the version you are using, then
emerging nvidia-drivers again.
If you want the nvidia driver available for multiple kernel
versions, you will need to emerge once for each version, with
the symlink updated each time.
If you are building from git, or otherwise discarded the kernel
tree, you will have to rebuild the kernel and fix the symlink
before emerging nvidia-drivers again.
HTH,
Phil
|
