LinuxQuestions.org
Welcome to the most active Linux Forum on the web.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Debian
User Name
Password
Debian This forum is for the discussion of Debian Linux.

Notices


Reply
  Search this Thread
Old 11-02-2014, 09:37 AM   #1
f4s1ss
LQ Newbie
 
Registered: Nov 2014
Posts: 2

Rep: Reputation: Disabled
nvidia drivers 340.32 installation using debirf


Hi everyone,

I'm building a custom debian live distribution which is running from RAM only (also without X). It has some access to the network share and some special packages. All this i did using the debirf package. Computers which running this live distro are diskless machines with the Tesla K20m GPU. Using the debirf which uses chroot (i guess) to access the created system I'm able to install the required packages i.e linux-headers, gcc-4.6, make, build-essential even also create the symbolic link for gcc-4.6 which is required for NVIDIA-Linux-x86_64-340.32.run. During the build i get this error

Code:
    DKMS make.log for nvidia-current-340.32 for kernel 3.2.0-4-amd64 (x86_64)
    Sat Nov 1 17:38:33 CET 2014
    make: Entering directory `/var/lib/dkms/nvidia-current/340.32/build'
    make KBUILD_VERBOSE=1 -C /lib/modules/3.2.0-4-amd64/build M=/var/lib/dkms/nvidi$
    make[1]: Entering directory `/usr/src/linux-headers-3.2.0-4-amd64'
    make -C /usr/src/linux-headers-3.2.0-4-amd64 \
    KBUILD_SRC=/usr/src/linux-headers-3.2.0-4-common \
    KBUILD_EXTMOD="/var/lib/dkms/nvidia-current/340.32/build" -f /usr/src/l$
    modules
    make: file.c:145: enter_file: Assertion `*name != '\0'' failed.
    Aborted
    make[2]: *** [sub-make] Error 134
    make[1]: *** [all] Error 2
    make[1]: Leaving directory `/usr/src/linux-headers-3.2.0-4-amd64'
    make: *** [modules] Error 2
    make: Leaving directory `/var/lib/dkms/nvidia-current/340.32/build'
^^^Any ideas about this?^^^


The second approach to build the drivers manually inside a running machine was successful. I'm also able to --add-this-kernel and to have already prepared builded drivers. (when i put these pre-build drivers to debirf I got the same error as above appears). After the installation on the headless machines I'm able to run nvidia-detect, nvidia-smi or nvidia-modprobe. the outputs are following:
Code:
    nvidia-detect
    No NVIDIA GPU detected.
Code:
    nvidia-smi
    Unable to determine the device handle for GPU 0000:83:00.0: Unknown Error
Code:
    nvidia-debugdump -l
    Found 1 NVIDIA devices
    Error: nvmlDeviceGetHandleByIndex(): Unknown Error
    FAILED to get details on GPU (0x0): Unknown Error
Code:
    dmesg | grep NVRM
    [ 2400.286037] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 340.32 Tue Aug 5 20:58:26 PDT 2014
    [ 2405.568003] NVRM: RmInitAdapter failed! (0x26:0xffffffff:1218)
    [ 2405.573892] NVRM: rm_init_adapter failed for device bearing minor number 0
    [ 2405.580824] NVRM: nvidia_frontend_open: minor 0, module->open() failed, error -5
Code:
    cat /proc/driver/nvidia/gpus/0000\:83\:00.0/information
    Model: Tesla K20m
    IRQ: 121
    GPU UUID: GPU-????????-????-????-????-????????????
    Video BIOS: ??.??.??.??.??
    Bus Type: PCIe
    DMA Size: 40 bits
    DMA Mask: 0xffffffffff
    Bus Location: 0000:83:00.0
Ideas about this? (already third forum where i posted this issue and still no answer...)
 
Old 11-11-2014, 10:29 PM   #2
ordealbyfire83
Member
 
Registered: Oct 2006
Location: Leiden, Netherlands
Distribution: LFS, Ubuntu Hardy
Posts: 302

Rep: Reputation: 89
Building things like modules via chroot can be tricky, since the kernel you are building against isn't necessarily the same kernel that is running (uname -r in the chroot will give you that of the host). Also you should mount/bind your /dev and /proc from the host before entering the chroot. Perhaps this is a factor as you said that it worked when building in a running machine.
 
Old 11-12-2014, 01:18 AM   #3
f4s1ss
LQ Newbie
 
Registered: Nov 2014
Posts: 2

Original Poster
Rep: Reputation: Disabled
Quote:
Originally Posted by ordealbyfire83 View Post
Building things like modules via chroot can be tricky, since the kernel you are building against isn't necessarily the same kernel that is running (uname -r in the chroot will give you that of the host). Also you should mount/bind your /dev and /proc from the host before entering the chroot. Perhaps this is a factor as you said that it worked when building in a running machine.
Hi ordealbyfire83,
aure the kernels have to be the same and they are. I updated the machine to the newest version. I also configured debirf to use the same sources as the building environment. Mounting of /dev and /proc is done automatically by debirf i guess. Anyway thank you for your reply!
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
[SOLVED] browser and X crash with pipelight, nvidia 340.17, kernel 3.15.5 mostlyharmless Slackware 1 09-08-2014 09:48 AM
startx-KDE hangs - Kernel 3.16.1 with NVIDIA-Linux-x86-340.32 firedrake Linux - Laptop and Netbook 7 09-02-2014 04:18 AM
[SOLVED] 24 bit depth on some apps with NVIDIA 340.24 Refractor Slackware 3 08-20-2014 12:53 AM
nvidia drivers installation perfectpol7 Linux - Hardware 1 03-02-2009 02:04 AM
Installation of NVidia drivers... Grim Reaper Linux - Software 2 01-30-2003 03:38 PM

LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Debian

All times are GMT -5. The time now is 07:34 AM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Open Source Consulting | Domain Registration