Comment # 47 on bug 1173733 from
(In reply to Stefan Dirsch from comment #33)
> (In reply to Mister Pend from comment #30)
> Then there is something fishy. 
> 
> You must have edited manually /etc/modprobe.d/50-nvidia-default.conf in
> order to have a different
> NVreg_DeviceFileGID there. 33 is the group ID of video group. That's why we
> use it here. Probably it no longer matters since permissions are meanwhile
> set via udev/logind (ACLs). So I guess you can ignore this.

Just out of curiosity, I clean installed my test machine (formerly working),
and now seeing the same issues on that machine. And looking at
/etc/modprobe.d/50-nvidia-default.conf, the NVreg_DeviceFileGID is still 483.
Installed from NVIDIA's repository. I can promise you 100% I haven't manually
edited this file. Resulting nvidia/cuda packages:

rpm -qa | grep -i nvidia
nvidia-gfxG05-kmp-default-440.100_k5.3.18_lp152.19-lp152.26.1.x86_64
nvidia-computeG05-440.100-lp152.26.1.x86_64
x11-video-nvidiaG05-440.100-lp152.26.1.x86_64

rpm -qa | grep -i cuda
cuda-nvprune-10-1-10.1.105-1.x86_64
cuda-curand-10-1-10.1.105-1.x86_64
cuda-visual-tools-10-1-10.1.105-1.x86_64
cuda-sanitizer-api-10-1-10.1.105-1.x86_64
cuda-driver-dev-10-1-10.1.105-1.x86_64
cuda-gdb-10-1-10.1.105-1.x86_64
cuda-compiler-10-1-10.1.105-1.x86_64
cuda-nsight-systems-10-1-10.1.105-1.x86_64
cuda-nvjpeg-dev-10-1-10.1.105-1.x86_64
cuda-tools-10-1-10.1.105-1.x86_64
cuda-nvjpeg-10-1-10.1.105-1.x86_64
cuda-cudart-10-1-10.1.105-1.x86_64
cuda-command-line-tools-10-1-10.1.105-1.x86_64
cuda-nvgraph-10-1-10.1.105-1.x86_64
cuda-gpu-library-advisor-10-1-10.1.105-1.x86_64
cuda-curand-dev-10-1-10.1.105-1.x86_64
cuda-nvcc-10-1-10.1.105-1.x86_64
cuda-nvprof-10-1-10.1.105-1.x86_64
cuda-npp-10-1-10.1.105-1.x86_64
cuda-cuobjdump-10-1-10.1.105-1.x86_64
cuda-npp-dev-10-1-10.1.105-1.x86_64
cuda-libraries-dev-10-1-10.1.105-1.x86_64
cuda-toolkit-10-1-10.1.105-1.x86_64
cuda-repo-opensuse15-10-1-local-10.1.105-418.39-1.0-1.x86_64
cuda-nvml-dev-10-1-10.1.105-1.x86_64
cuda-misc-headers-10-1-10.1.105-1.x86_64
cuda-cufft-10-1-10.1.105-1.x86_64
cuda-cusparse-dev-10-1-10.1.105-1.x86_64
cuda-cupti-10-1-10.1.105-1.x86_64
cuda-nvrtc-10-1-10.1.105-1.x86_64
cuda-nsight-compute-10-1-10.1.105-1.x86_64
cuda-cusolver-10-1-10.1.105-1.x86_64
cuda-nvgraph-dev-10-1-10.1.105-1.x86_64
cuda-cudart-dev-10-1-10.1.105-1.x86_64
cuda-samples-10-1-10.1.105-1.x86_64
cuda-nsight-10-1-10.1.105-1.x86_64
cuda-nvvp-10-1-10.1.105-1.x86_64
cuda-documentation-10-1-10.1.105-1.x86_64
cuda-nvdisasm-10-1-10.1.105-1.x86_64
cuda-nvrtc-dev-10-1-10.1.105-1.x86_64
cuda-nvtx-10-1-10.1.105-1.x86_64
cuda-cusparse-10-1-10.1.105-1.x86_64
cuda-cufft-dev-10-1-10.1.105-1.x86_64
cuda-license-10-1-10.1.105-1.x86_64
cuda-memcheck-10-1-10.1.105-1.x86_64
cuda-cusolver-dev-10-1-10.1.105-1.x86_64

If there are further tests needed, I'm happy to help. My test machine I can
rebuild in under 30 minutes, so happy for potentially destructive tests too.


You are receiving this mail because: