[Solved] Can't get nvidia driver to load on fresh install

dchw · 2017-11-26 00:18:27

As noted in the title, I have a machine that has a fresh Arch installation on it, which seems to have misbehaving drivers. My system is a dual-GPU system, where one is using vfio-pci for some VM gaming, and the other is intended for Linux. I have installed the nvidia driver, but it doesn't act like it was installed. Last time I used Arch (almost a year ago) I don't remember it being this hard?

When I try to start X, I get errors saying that there were no screens found. It looks like I haven't finished configuring it yet here?

[  1740.725] 
X.Org X Server 1.19.5
Release Date: 2017-10-12
[  1740.727] X Protocol Version 11, Revision 0
[  1740.727] Build Operating System: Linux 4.9.54-1-lts x86_64 
[  1740.728] Current Operating System: Linux host 4.13.12-1-ARCH #1 SMP PREEMPT Wed Nov 8 11:54:06 CET 2017 x86_64
[  1740.728] Kernel command line: initrd=\intel-ucode.img initrd=\initramfs-linux.img root=/dev/sda3 rw intel_iommu=on
[  1740.729] Build Date: 12 October 2017  09:59:53PM
[  1740.729]  
[  1740.730] Current version of pixman: 0.34.0
[  1740.731] 	Before reporting problems, check http://wiki.x.org
	to make sure that you have the latest version.
[  1740.731] Markers: (--) probed, (**) from config file, (==) default setting,
	(++) from command line, (!!) notice, (II) informational,
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[  1740.733] (==) Log file: "/var/log/Xorg.0.log", Time: Sat Nov 25 16:41:08 2017
[  1740.735] (==) Using system config directory "/usr/share/X11/xorg.conf.d"
[  1740.735] (==) No Layout section.  Using the first Screen section.
[  1740.735] (==) No screen section available. Using defaults.
[  1740.735] (**) |-->Screen "Default Screen Section" (0)
[  1740.735] (**) |   |-->Monitor "<default monitor>"
[  1740.736] (==) No monitor specified for screen "Default Screen Section".
	Using a default monitor configuration.
[  1740.736] (==) Automatically adding devices
[  1740.736] (==) Automatically enabling devices
[  1740.736] (==) Automatically adding GPU devices
[  1740.736] (==) Automatically binding GPU devices
[  1740.736] (==) Max clients allowed: 256, resource mask: 0x1fffff
[  1740.737] (WW) The directory "/usr/share/fonts/Type1/" does not exist.
[  1740.737] 	Entry deleted from font path.
[  1740.737] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/100dpi/".
[  1740.737] 	Entry deleted from font path.
[  1740.737] 	(Run 'mkfontdir' on "/usr/share/fonts/100dpi/").
[  1740.737] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/75dpi/".
[  1740.737] 	Entry deleted from font path.
[  1740.737] 	(Run 'mkfontdir' on "/usr/share/fonts/75dpi/").
[  1740.737] (==) FontPath set to:
	/usr/share/fonts/misc/,
	/usr/share/fonts/TTF/,
	/usr/share/fonts/OTF/
[  1740.737] (==) ModulePath set to "/usr/lib/xorg/modules"
[  1740.737] (II) The server relies on udev to provide the list of input devices.
	If no devices become available, reconfigure udev or disable AutoAddDevices.
[  1740.737] (II) Loader magic: 0x55e385429d60
[  1740.737] (II) Module ABI versions:
[  1740.737] 	X.Org ANSI C Emulation: 0.4
[  1740.737] 	X.Org Video Driver: 23.0
[  1740.737] 	X.Org XInput driver : 24.1
[  1740.737] 	X.Org Server Extension : 10.0
[  1740.738] (++) using VT number 1

[  1740.738] (--) controlling tty is VT number 1, auto-enabling KeepTty
[  1740.739] (II) systemd-logind: took control of session /org/freedesktop/login1/session/c1
[  1740.743] (--) PCI: (0:1:0:0) 10de:1b81:1458:3701 rev 161, Mem @ 0xfa000000/16777216, 0xe0000000/268435456, 0xf0000000/33554432, I/O @ 0x0000e000/128, BIOS @ 0x????????/524288
[  1740.743] (--) PCI:*(0:2:0:0) 10de:1401:3842:3967 rev 161, Mem @ 0xf8000000/16777216, 0xc0000000/268435456, 0xd0000000/33554432, I/O @ 0x0000d000/128, BIOS @ 0x????????/131072
[  1740.743] (WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)
[  1740.743] (II) LoadModule: "glx"
[  1740.744] (II) Loading /usr/lib/xorg/modules/extensions/libglx.so
[  1740.750] (II) Module glx: vendor="X.Org Foundation"
[  1740.750] 	compiled for 1.19.5, module version = 1.0.0
[  1740.750] 	ABI class: X.Org Server Extension, version 10.0
[  1740.750] (==) Matched nouveau as autoconfigured driver 0
[  1740.750] (==) Matched nv as autoconfigured driver 1
[  1740.750] (==) Matched modesetting as autoconfigured driver 2
[  1740.750] (==) Matched fbdev as autoconfigured driver 3
[  1740.750] (==) Matched vesa as autoconfigured driver 4
[  1740.750] (==) Assigned the driver to the xf86ConfigLayout
[  1740.750] (II) LoadModule: "nouveau"
[  1740.750] (WW) Warning, couldn't open module nouveau
[  1740.750] (II) UnloadModule: "nouveau"
[  1740.750] (II) Unloading nouveau
[  1740.750] (EE) Failed to load module "nouveau" (module does not exist, 0)
[  1740.750] (II) LoadModule: "nv"
[  1740.750] (WW) Warning, couldn't open module nv
[  1740.750] (II) UnloadModule: "nv"
[  1740.750] (II) Unloading nv
[  1740.750] (EE) Failed to load module "nv" (module does not exist, 0)
[  1740.750] (II) LoadModule: "modesetting"
[  1740.750] (II) Loading /usr/lib/xorg/modules/drivers/modesetting_drv.so
[  1740.751] (II) Module modesetting: vendor="X.Org Foundation"
[  1740.751] 	compiled for 1.19.5, module version = 1.19.5
[  1740.751] 	Module class: X.Org Video Driver
[  1740.751] 	ABI class: X.Org Video Driver, version 23.0
[  1740.751] (II) LoadModule: "fbdev"
[  1740.751] (WW) Warning, couldn't open module fbdev
[  1740.751] (II) UnloadModule: "fbdev"
[  1740.751] (II) Unloading fbdev
[  1740.751] (EE) Failed to load module "fbdev" (module does not exist, 0)
[  1740.751] (II) LoadModule: "vesa"
[  1740.751] (WW) Warning, couldn't open module vesa
[  1740.751] (II) UnloadModule: "vesa"
[  1740.751] (II) Unloading vesa
[  1740.751] (EE) Failed to load module "vesa" (module does not exist, 0)
[  1740.751] (II) modesetting: Driver for Modesetting Kernel Drivers: kms
[  1740.751] (EE) open /dev/dri/card0: No such file or directory
[  1740.751] (WW) Falling back to old probe method for modesetting
[  1740.751] (EE) open /dev/dri/card0: No such file or directory
[  1740.751] (EE) Screen 0 deleted because of no matching config section.
[  1740.751] (II) UnloadModule: "modesetting"
[  1740.751] (EE) Device(s) detected, but none match those in the config file.
[  1740.751] (EE) 
Fatal server error:
[  1740.751] (EE) no screens found(EE) 
[  1740.751] (EE) 
Please consult the The X.Org Foundation support 
	 at http://wiki.x.org
 for help. 
[  1740.751] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[  1740.751] (EE) 
[  1740.753] (EE) Server terminated with error (1). Closing log file.

If I provide a minimal configuration:

[  3744.467] 
X.Org X Server 1.19.5
Release Date: 2017-10-12
[  3744.467] X Protocol Version 11, Revision 0
[  3744.468] Build Operating System: Linux 4.9.54-1-lts x86_64 
[  3744.468] Current Operating System: Linux host 4.13.12-1-ARCH #1 SMP PREEMPT Wed Nov 8 11:54:06 CET 2017 x86_64
[  3744.468] Kernel command line: initrd=\intel-ucode.img initrd=\initramfs-linux.img root=/dev/sda3 rw intel_iommu=on
[  3744.468] Build Date: 12 October 2017  09:59:53PM
[  3744.468]  
[  3744.469] Current version of pixman: 0.34.0
[  3744.469] 	Before reporting problems, check http://wiki.x.org
	to make sure that you have the latest version.
[  3744.469] Markers: (--) probed, (**) from config file, (==) default setting,
	(++) from command line, (!!) notice, (II) informational,
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[  3744.470] (==) Log file: "/var/log/Xorg.0.log", Time: Sat Nov 25 17:14:31 2017
[  3744.470] (==) Using config directory: "/etc/X11/xorg.conf.d"
[  3744.471] (==) Using system config directory "/usr/share/X11/xorg.conf.d"
[  3744.471] (==) No Layout section.  Using the first Screen section.
[  3744.471] (==) No screen section available. Using defaults.
[  3744.471] (**) |-->Screen "Default Screen Section" (0)
[  3744.471] (**) |   |-->Monitor "<default monitor>"
[  3744.471] (==) No device specified for screen "Default Screen Section".
	Using the first device section listed.
[  3744.471] (**) |   |-->Device "Nvidia Card"
[  3744.471] (==) No monitor specified for screen "Default Screen Section".
	Using a default monitor configuration.
[  3744.471] (==) Automatically adding devices
[  3744.471] (==) Automatically enabling devices
[  3744.471] (==) Automatically adding GPU devices
[  3744.471] (==) Automatically binding GPU devices
[  3744.471] (==) Max clients allowed: 256, resource mask: 0x1fffff
[  3744.471] (WW) The directory "/usr/share/fonts/Type1/" does not exist.
[  3744.471] 	Entry deleted from font path.
[  3744.471] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/100dpi/".
[  3744.471] 	Entry deleted from font path.
[  3744.471] 	(Run 'mkfontdir' on "/usr/share/fonts/100dpi/").
[  3744.471] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/75dpi/".
[  3744.471] 	Entry deleted from font path.
[  3744.471] 	(Run 'mkfontdir' on "/usr/share/fonts/75dpi/").
[  3744.471] (==) FontPath set to:
	/usr/share/fonts/misc/,
	/usr/share/fonts/TTF/,
	/usr/share/fonts/OTF/
[  3744.471] (==) ModulePath set to "/usr/lib/xorg/modules"
[  3744.471] (II) The server relies on udev to provide the list of input devices.
	If no devices become available, reconfigure udev or disable AutoAddDevices.
[  3744.471] (II) Loader magic: 0x563b1b19fd60
[  3744.471] (II) Module ABI versions:
[  3744.471] 	X.Org ANSI C Emulation: 0.4
[  3744.471] 	X.Org Video Driver: 23.0
[  3744.471] 	X.Org XInput driver : 24.1
[  3744.471] 	X.Org Server Extension : 10.0
[  3744.471] (++) using VT number 1

[  3744.471] (--) controlling tty is VT number 1, auto-enabling KeepTty
[  3744.473] (II) systemd-logind: took control of session /org/freedesktop/login1/session/c1
[  3744.477] (--) PCI: (0:1:0:0) 10de:1b81:1458:3701 rev 161, Mem @ 0xfa000000/16777216, 0xe0000000/268435456, 0xf0000000/33554432, I/O @ 0x0000e000/128, BIOS @ 0x????????/524288
[  3744.477] (--) PCI:*(0:2:0:0) 10de:1401:3842:3967 rev 161, Mem @ 0xf8000000/16777216, 0xc0000000/268435456, 0xd0000000/33554432, I/O @ 0x0000d000/128, BIOS @ 0x????????/131072
[  3744.477] (WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)
[  3744.477] (II) LoadModule: "glx"
[  3744.477] (II) Loading /usr/lib/xorg/modules/extensions/libglx.so
[  3744.478] (II) Module glx: vendor="X.Org Foundation"
[  3744.478] 	compiled for 1.19.5, module version = 1.0.0
[  3744.478] 	ABI class: X.Org Server Extension, version 10.0
[  3744.478] (II) LoadModule: "nvidia"
[  3744.478] (II) Loading /usr/lib/xorg/modules/drivers/nvidia_drv.so
[  3744.483] (II) Module nvidia: vendor="NVIDIA Corporation"
[  3744.483] 	compiled for 4.0.2, module version = 1.0.0
[  3744.483] 	Module class: X.Org Video Driver
[  3744.483] (II) NVIDIA dlloader X Driver  387.34  Tue Nov 21 01:38:22 PST 2017
[  3744.483] (II) NVIDIA Unified Driver for all Supported NVIDIA GPUs
[  3744.484] (II) Loading sub module "fb"
[  3744.484] (II) LoadModule: "fb"
[  3744.485] (II) Loading /usr/lib/xorg/modules/libfb.so
[  3744.486] (II) Module fb: vendor="X.Org Foundation"
[  3744.486] 	compiled for 1.19.5, module version = 1.0.0
[  3744.486] 	ABI class: X.Org ANSI C Emulation, version 0.4
[  3744.486] (II) Loading sub module "wfb"
[  3744.486] (II) LoadModule: "wfb"
[  3744.486] (II) Loading /usr/lib/xorg/modules/libwfb.so
[  3744.487] (II) Module wfb: vendor="X.Org Foundation"
[  3744.487] 	compiled for 1.19.5, module version = 1.0.0
[  3744.487] 	ABI class: X.Org ANSI C Emulation, version 0.4
[  3744.487] (II) Loading sub module "ramdac"
[  3744.487] (II) LoadModule: "ramdac"
[  3744.487] (II) Module "ramdac" already built-in
[  3744.490] (EE) NVIDIA: Failed to initialize the NVIDIA kernel module. Please see the
[  3744.490] (EE) NVIDIA:     system's kernel log for additional error messages and
[  3744.490] (EE) NVIDIA:     consult the NVIDIA README for details.
[  3744.490] (EE) No devices detected.
[  3744.490] (EE) 
Fatal server error:
[  3744.490] (EE) no screens found(EE) 
[  3744.490] (EE) 
Please consult the The X.Org Foundation support 
	 at http://wiki.x.org
 for help. 
[  3744.490] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[  3744.490] (EE) 
[  3744.492] (EE) Server terminated with error (1). Closing log file.

I see that it mentions checking kernel logs for additional errors. I cant seem to find anything with "nvidia", "NVIDIA". I do see some messages regarding sound cards when I grep for "NVidia", though.

Driver is installed:

host% pacman -Q | grep nvidia

nvidia 387.34-3
nvidia-utils 387.34-1

Modprobe for nvidia module:

host% modprobe -vv nvidia

modprobe: INFO: custom logging function 0x40a2a0 registered
modprobe: FATAL: Module nvidia not found in directory /lib/modules/4.13.12-1-ARCH

My card configuration:

host% lspci -nnk  

...
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GP104 [GeForce GTX 1070] [10de:10f0] (rev a1)
        Subsystem: Gigabyte Technology Co., Ltd GP104 [GeForce GTX 1070] [1458:3701]
        Kernel driver in use: vfio-pci
        Kernel modules: noveau
01:00.1 Audio device [0403]: NVIDIA Corporation GP104 High Definition Audio Controller [10de:10f0] (rev a1)
        Subsystem: Gigabyte Technology Co., Ltd GP104 High Definition Audio Controller [1458:3701]
        Kernel driver in use: vfio-pci
        Kernel modules: snd_hda_intel
02:00.0 VGA compatible controller [0300]: NVIDIA Corporation GM206 [GeForce GTX 960] [10de:1401] (rev a1)
        Subsystem: eVga.com. Corp. Device [3842:3967]
        Kernel modules: noveau
02:00.1 Audio device [0403]: NVIDIA Corporation Device [10de:0fba] (rev a1)
        Subsystem: eVga.com. Corp. Device [3842:3967]
        Kernel driver in use: snd_hda_intel
        Kernel modules: snd_hda_intel     
...

I do not have the persistence systemd unit enabled; I have never had to use that in the past. In addition the NVidia wiki page mentions that it blacklists noveau; I am not sure where to look to confirm that is the case.

If you need any other logs from me, let me know and I can get them uploaded quickly. I appreciate any help!

Last edited by dchw (2017-11-27 04:07:53)

loqs · 2017-11-26 00:27:49

Can you run the following please

# depmod 4.13.12-1-ARCH
$ modprobe -vv nvidia

dchw · 2017-11-26 00:42:03

host% sudo depmod 4.13.12-1-ARCH
host% echo $?
0

host% sudo modprobe -vv nvidia

modprobe: INFO: custom logging function 0x40a2a0 registered
insmod /lib/modules/4.13.12-1-ARCH/extramodules/nvidia.ko.gz
modprobe: INFO: context 0x1b023f0 released

Looks like that did it:

host% lspci -nnk  

...
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GP104 [GeForce GTX 1070] [10de:10f0] (rev a1)
        Subsystem: Gigabyte Technology Co., Ltd GP104 [GeForce GTX 1070] [1458:3701]
        Kernel driver in use: vfio-pci
        Kernel modules: noveau, nvidia_drm, nvidia
01:00.1 Audio device [0403]: NVIDIA Corporation GP104 High Definition Audio Controller [10de:10f0] (rev a1)
        Subsystem: Gigabyte Technology Co., Ltd GP104 High Definition Audio Controller [1458:3701]
        Kernel driver in use: vfio-pci
        Kernel modules: snd_hda_intel
02:00.0 VGA compatible controller [0300]: NVIDIA Corporation GM206 [GeForce GTX 960] [10de:1401] (rev a1)
        Subsystem: eVga.com. Corp. Device [3842:3967]
        Kernel driver in use: nvidia
        Kernel modules: noveau, nvidia_drm, nvidia
02:00.1 Audio device [0403]: NVIDIA Corporation Device [10de:0fba] (rev a1)
        Subsystem: eVga.com. Corp. Device [3842:3967]
        Kernel driver in use: snd_hda_intel
        Kernel modules: snd_hda_intel     
...

I want to be able to do this myself in the future. Can you help me understand what led you to depmod, and why this was an issue for me?

Scimmia · 2017-11-26 00:44:03

With what nvidia package?

Edit: nevermind, I see it now. Missed it the first time.

Last edited by Scimmia (2017-11-26 00:45:02)

loqs · 2017-11-26 00:48:29

https://bugs.archlinux.org/task/56449 which was unfortunately closed. What I noticed was with the move the the 4.14 kernel in testing the linux package gained a new hook that runs depmods when external modules such as nvidia are added.
So the nvidia package which used to run depmod in that packages install script had that script removed and relied upon the hook instead. nvidia 387.34 is released and versions are built for linux in testing and linux in core.
However the version built for core nvidia 387.34-3 did not have the install file put back so when nvidia 387.34-3 is installed nothing runs depmod.
Edit:
@Scimmia that explanation sound plausible?

Last edited by loqs (2017-11-26 00:52:00)

Scimmia · 2017-11-26 00:53:43

Yes, that sounds like what happened here, but the linked ticket is totally different since the person filing it specifically said the files were in the 4.14 dir.

loqs · 2017-11-26 00:56:00

My mistake on the ticket should dchw open a new ticket or you being aware of the issue enough?
Edit:
Opening new ticket as requested https://bugs.archlinux.org/task/56454
Edit2:
https://bugs.archlinux.org/task/56455 yes they said "Modules are placed in the extramodules-4.14-ARCH instead of extramodules-4.13-ARCH."
but now have amended their position "I was just somehow lost in svntogit and thought that wrong version number in PKGBUILD was the cause."

Last edited by loqs (2017-11-26 01:49:23)

jeckhack · 2017-11-26 01:57:05

Hi. That was me who opened https://bugs.archlinux.org/task/56449. The reason of this confusion about wrong folder is that I started to mess with the testing repository to understand the nature of this bug. That was a bad idea to do at 3 a.m.

dchw · 2017-11-27 04:05:59

Verified the fix mentioned in the bug. Marking thread as resolved.

Thanks for all your help! Every time I come here I find the best help!

Last edited by dchw (2017-11-27 04:06:45)

Arch Linux

#1 2017-11-26 00:18:27

[Solved] Can't get nvidia driver to load on fresh install

#2 2017-11-26 00:27:49

Re: [Solved] Can't get nvidia driver to load on fresh install

#3 2017-11-26 00:42:03

Re: [Solved] Can't get nvidia driver to load on fresh install

#4 2017-11-26 00:44:03

Re: [Solved] Can't get nvidia driver to load on fresh install

#5 2017-11-26 00:48:29

Re: [Solved] Can't get nvidia driver to load on fresh install

#6 2017-11-26 00:53:43

Re: [Solved] Can't get nvidia driver to load on fresh install

#7 2017-11-26 00:56:00

Re: [Solved] Can't get nvidia driver to load on fresh install

#8 2017-11-26 01:57:05

Re: [Solved] Can't get nvidia driver to load on fresh install

#9 2017-11-27 04:05:59

Re: [Solved] Can't get nvidia driver to load on fresh install

Board footer