NVRM: RmInitAdapter failed!

storms · 2019-03-11 16:54:27

hey guys.

so i'm trying to get bumblebee to work, no luck so far.

-->>FRESH ARCH INSTALL<<--
installed i3,spotify, steam and a couple of other things but have yet to mess with system settings and what not.

computer name:

Acer Aspire E5-571G

i added some stuff:

 8G ram, 500GB SSD

not that it matters just to give you guys as much info as possible.

uname -a

 
Linux void 5.0.0-arch1-1-ARCH #1 SMP PREEMPT Mon Mar 4 14:11:43 UTC 2019 x86_64 GNU/Linux

lspci | egrep 'VGA|3D'

 
00:02.0 VGA compatible controller: Intel Corporation Haswell-ULT Integrated Graphics Controller (rev 0b)
03:00.0 3D controller: NVIDIA Corporation GM108M [GeForce 840M] (rev a2)

packages installed:

nvidia
lib32-nvidia-utils
nvidia-settings
nvidia-utils
bumblebee
mesa
xf86-video-intel
lib32-virtualgl
mesa-demos
xorg-server
xorg-server-common

optirun -vvv glxgears:

[ 1033.858761] [DEBUG]Reading file: /etc/bumblebee/bumblebee.conf
[ 1033.859447] [DEBUG]optirun version 3.2.1 starting...
[ 1033.859477] [DEBUG]Active configuration:
[ 1033.859486] [DEBUG] bumblebeed config file: /etc/bumblebee/bumblebee.conf
[ 1033.859494] [DEBUG] X display: :8
[ 1033.859502] [DEBUG] LD_LIBRARY_PATH: /usr/lib/nvidia:/usr/lib32/nvidia:/usr/lib:/usr/lib32
[ 1033.859576] [DEBUG] Socket path: /var/run/bumblebee.socket
[ 1033.859627] [DEBUG] Accel/display bridge: auto
[ 1033.859648] [DEBUG] VGL Compression: proxy
[ 1033.859667] [DEBUG] VGLrun extra options:
[ 1033.859691] [DEBUG] Primus LD Path: /usr/lib/primus:/usr/lib32/primus
[ 1033.859750] [DEBUG]Using auto-detected bridge virtualgl
[ 1042.063171] [INFO]Response: No - error: [XORG] (EE) NVIDIA(GPU-0): Failed to initialize the NVIDIA GPU at PCI:3:0:0.  Please

[ 1042.063186] [ERROR]Cannot access secondary GPU - error: [XORG] (EE) NVIDIA(GPU-0): Failed to initialize the NVIDIA GPU at PCI:3:0:0.  Please

[ 1042.063191] [DEBUG]Socket closed.
[ 1042.063200] [ERROR]Aborting because fallback start is disabled.
[ 1042.063204] [DEBUG]Killing all remaining processes.

/var/log/Xorg.8.log

/etc/bumblebee/bumblebee.conf

etc/bumblebee/xorg.conf.nvidia

dmesg

bootloeader options (systemd)

title	Arch Linux
linux	/vmlinuz-linux
initrd	/intel-ucode.img
initrd	/initramfs-linux.img
options	cryptdevice=/dev/sda2:luks:allow-discards resume=/dev/mapper/vg0-swap root=/dev/mapper/vg0-root rw quiet modprobe.blacklist=nouveau rcutree.rcu_idle_gp_delay=1

hope you guys can help me, let me know if you need any more info.

edit:

lsmod

i found this thread about the "RmInitAdapter failed!" , sadly i dont have the

 
Option "RegistryDwords" "PowerMizerEnable=0x1; PerfLevelSrc=0x2222; PowerMizerDefaultAC=0x1

on my xorg.conf.nvidia...
also as pointed out in the wiki and on my systemd conf, i do have

rcutree.rcu_idle_gp_delay=1

digging into the nvidia forums:
nvidia-smi -q

No devices were found

edit edit (i will lose my mind...)
honestly i'm just trying anything and everything at this point...
found another thread that said to use

pcie_aspm=off

on the kernel, that didnt work

so i found out that nvidia has a bug report tool, if someone is crazy enough to understand it....
here it is, nvidia-bug-report.sh

about the report:
aside from the obvious

RmInitAdapter failed!

i found this:

[   11.287967] bbswitch: module verification failed: signature and/or required key missing - tainting kernel
[   11.288207] bbswitch: version 0.8
[   11.288213] bbswitch: Found integrated VGA device 0000:00:02.0: \_SB_.PCI0.GFX0
[   11.288222] bbswitch: Found discrete VGA device 0000:03:00.0: \_SB_.PCI0.RP05.PEGP
[   11.288236] ACPI Warning: \_SB.PCI0.RP05.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20181213/nsarguments-66)
[   11.288361] bbswitch: detected an Optimus _DSM function
[   11.288375] pci 0000:03:00.0: enabling device (0006 -> 0007)
[   11.288435] bbswitch: Succesfully loaded. Discrete card 0000:03:00.0 is on
[   11.303338] bbswitch: disabling discrete graphics

also

[   45.003098] bbswitch: enabling discrete graphics
[   45.068449] IPMI message handler: version 39.2
[   45.070418] ipmi device interface
[   45.190433] nvidia: module license 'NVIDIA' taints kernel.
[   45.190435] Disabling lock debugging due to kernel taint
[   45.204573] nvidia-nvlink: Nvlink Core is being initialized, major device number 236
[   45.305139] NVRM: loading NVIDIA UNIX x86_64 Kernel Module  418.43  Tue Feb 19 01:12:11 CST 2019
[   53.504298] NVRM: RmInitAdapter failed! (0x26:0x65:1106)
[   53.504326] NVRM: rm_init_adapter failed for device bearing minor number 0
[   68.249397] NVRM: RmInitAdapter failed! (0x26:0x65:1106)
[   68.249430] NVRM: rm_init_adapter failed for device bearing minor number 0
[  234.642029] NVRM: RmInitAdapter failed! (0x26:0x65:1106)
[  234.642068] NVRM: rm_init_adapter failed for device bearing minor number 0

is the connection with the gpu broken or something? not sure what to make of this:

*** /proc/driver/nvidia/./gpus/0000:03:00.0/information
*** ls: -r--r--r-- 1 root root 0 2019-03-11 18:02:31.852095882 +0000 /proc/driver/nvidia/./gpus/0000:03:00.0/information
Model: 		 Unknown
IRQ:   		 48
GPU UUID: 	 GPU-????????-????-????-????-????????????
Video BIOS: 	 ??.??.??.??.??
Bus Type: 	 PCIe
DMA Size: 	 40 bits
DMA Mask: 	 0xffffffffff
Bus Location: 	 0000:03:00.0
Device Minor: 	 0
Blacklisted:	 No

i didnt changed the ???, thats how the log shows it...but then again i'm no nvidia/hardware expert, so it could be nothing related

Last edited by storms (2019-03-11 18:57:02)

loqs · 2019-03-11 17:25:15

[   24.909697] nvidia-nvlink: Nvlink Core is being initialized, major device number 236
[   25.010189] NVRM: loading NVIDIA UNIX x86_64 Kernel Module  418.43  Tue Feb 19 01:12:11 CST 2019
[   33.219403] NVRM: RmInitAdapter failed! (0x26:0x65:1106)
[   33.219445] NVRM: rm_init_adapter failed for device bearing minor number 0

seth · 2019-03-11 20:49:45

The GPU doesn't power on after the initial power down by bbswitch.
=> Try the lts kernel on whether that's a 5.0 thing.
You can also try to bring the GPU on by hand before attempting an optirun.
Ultimately, you might have to go w/o powering the chip down while unused (until this gets fixed - "somehow")

storms · 2019-03-11 21:53:21

hello!
sorry i stopped updating this thread (my bad)
so, i went ahead and made a post on the nvidia dev forums. ( https://devtalk.nvidia.com/default/topi … =7#5320483 )
things i did so far:

pacman -Rns bbswitch

tried diferent kernel options

acpi_osi=! acpi_osi="Windows 2009"
acpi_osi=! acpi_osi="Windows 2015"

and installed (thanks irc guys! love you <3) the lts kernel
still the same problem BUT this time is called for nouveau
so i removed

modprobe.blacklist=nouveau

rebooted, tried the optirun -vvv glxgears and:

[   19.295902] [DEBUG]Reading file: /etc/bumblebee/bumblebee.conf
[   19.296510] [DEBUG]optirun version 3.2.1 starting...
[   19.296543] [DEBUG]Active configuration:
[   19.296562] [DEBUG] bumblebeed config file: /etc/bumblebee/bumblebee.conf
[   19.296573] [DEBUG] X display: :8
[   19.296584] [DEBUG] LD_LIBRARY_PATH:
[   19.296594] [DEBUG] Socket path: /var/run/bumblebee.socket
[   19.296604] [DEBUG] Accel/display bridge: auto
[   19.296614] [DEBUG] VGL Compression: proxy
[   19.296627] [DEBUG] VGLrun extra options:
[   19.296658] [DEBUG] Primus LD Path: /usr/lib/primus:/usr/lib32/primus
[   19.296797] [DEBUG]Using auto-detected bridge virtualgl
[   22.775242] [INFO]Response: No - error: [XORG] (EE) Failed to load module "nouveau" (module does not exist, 0)

[   22.775261] [ERROR]Cannot access secondary GPU - error: [XORG] (EE) Failed to load module "nouveau" (module does not exist, 0)

[   22.775268] [DEBUG]Socket closed.
[   22.775279] [ERROR]Aborting because fallback start is disabled.
[   22.775284] [DEBUG]Killing all remaining processes.

nvidia-bug-report-nouveau-not-blacklisted
dmesg-nouveau-not-blacklisted

Last edited by storms (2019-03-11 21:58:43)

seth · 2019-03-11 21:56:00

the lts kernel still the same problem BUT this time is called for nouveau

You forgot to install https://www.archlinux.org/packages/extr … vidia-lts/ ;-)

storms · 2019-03-11 22:01:44

seth wrote:

the lts kernel still the same problem BUT this time is called for nouveau
You forgot to install https://www.archlinux.org/packages/extr … vidia-lts/ ;-)

oh...i'm a dumb dumb.
ok i installed it, we are back at square one.

same problem

[   87.934803] [DEBUG]Reading file: /etc/bumblebee/bumblebee.conf
[   87.935520] [DEBUG]optirun version 3.2.1 starting...
[   87.935551] [DEBUG]Active configuration:
[   87.935559] [DEBUG] bumblebeed config file: /etc/bumblebee/bumblebee.conf
[   87.935568] [DEBUG] X display: :8
[   87.935576] [DEBUG] LD_LIBRARY_PATH: /usr/lib/nvidia:/usr/lib32/nvidia:/usr/lib:/usr/lib32
[   87.935612] [DEBUG] Socket path: /var/run/bumblebee.socket
[   87.935645] [DEBUG] Accel/display bridge: auto
[   87.935667] [DEBUG] VGL Compression: proxy
[   87.935687] [DEBUG] VGLrun extra options:
[   87.935705] [DEBUG] Primus LD Path: /usr/lib/primus:/usr/lib32/primus
[   87.935769] [DEBUG]Using auto-detected bridge virtualgl
[   96.135200] [INFO]Response: No - error: [XORG] (EE) NVIDIA(GPU-0): Failed to initialize the NVIDIA GPU at PCI:3:0:0.  Please

[   96.135217] [ERROR]Cannot access secondary GPU - error: [XORG] (EE) NVIDIA(GPU-0): Failed to initialize the NVIDIA GPU at PCI:3:0:0.  Please

[   96.135221] [DEBUG]Socket closed.
[   96.135233] [ERROR]Aborting because fallback start is disabled.
[   96.135237] [DEBUG]Killing all remaining processes.

just to make sure, should i remove the nvidia package?

edit: removing it, just to make sure!

Last edited by storms (2019-03-11 22:03:59)

seth · 2019-03-11 22:06:43

"should i remove the nvidia package"
No, the kernels are independent.

Since you've removed bbswitch, can we get an updated dmesg? (there should be some change as the device should not be initially down)

storms · 2019-03-11 22:10:54

hey, thanks for the quick reply.

so i did lots of things.
removed nvidia ( not needed,now i know, will install it again)
tried optirun -vvv glxgears WITHOUT modprobe.blacklist=nouveau | same error
now i just tried optirun -vvv glxgears WITH modprobe.blacklist=nouveau | same error.

heres the dmesg (lts-kernel, with modprobe.blacklist=nouveau)
new-dmesg

Last edited by storms (2019-03-11 22:11:14)

seth · 2019-03-11 22:15:32

Sanity check: does the nvidia driver load and take control of the GPU if you remove bumblebee?
Your graphical session might fail for that config, use the multi-user.target: https://wiki.archlinux.org/index.php/Sy … _boot_into - you can still ask nvidia-smi there.

storms · 2019-03-11 22:30:47

seth wrote:

Sanity check: does the nvidia driver load and take control of the GPU if you remove bumblebee?
Your graphical session might fail for that config, use the multi-user.target: https://wiki.archlinux.org/index.php/Sy … _boot_into - you can still ask nvidia-smi there.

Ok I think you are on to something, on the multi user the nvidia-smi said the same thing
"No devices were found"

Also, how do I revert back?
I tried
Systemctl disable multi-target
Systemctl enable graphical.target

Systemctl set-default graphical.target

Systemctl get-default
Says that the default it graphical.target but X doesn't load

Edit:
https://imgur.com/a/3Q3j9QG

Last edited by storms (2019-03-11 22:37:25)

storms · 2019-03-11 22:46:53

nevermind, I'M BACK!

i'm might now nothing about nvidia but i still know how to read logs, had to change the

Section "OutputClass"
    Identifier "nvidia"
    MatchDriver "nvidia-drm"
    Driver "nvidia"
    Option "AllowEmptyInitialConfiguration"
    Option "PrimaryGPU" "no"
    ModulePath "/usr/lib/nvidia/xorg"
    ModulePath "/usr/lib/xorg/modules"
EndSection

Option "PrimaryGPU" "no"

it was "yes"

the guys on the "other side" (nvidia dev forums) said that the gpu might be broken...
also said the only way to test it was to install...windows *gasp*

Last edited by storms (2019-03-11 22:53:51)

seth · 2019-03-11 22:59:24

Change where? The bumblebee config?
Is the notebook on battery or the wall plug?

storms · 2019-03-11 23:01:13

well it does have a battery but its connected to the plug.
the change i made was in

/usr/share/X11/xorg.conf.d/10-nvidia-drm-outputclass.conf

also full conf file, just because...

Section "OutputClass"
    Identifier "intel"
    MatchDriver "i915"
    Driver "modesetting"
EndSection

Section "OutputClass"
    Identifier "nvidia"
    MatchDriver "nvidia-drm"
    Driver "nvidia"
    Option "AllowEmptyInitialConfiguration"
    Option "PrimaryGPU" "no"
    ModulePath "/usr/lib/nvidia/xorg"
    ModulePath "/usr/lib/xorg/modules"
EndSection

edit: also i dont have bumblebee anymore, by suggestion of an user on the other side i pacman -Rns it

Last edited by storms (2019-03-11 23:05:22)

loqs · 2019-03-12 00:31:06

Without bumblebee installed and without the nvidia GPU being set as PrimaryGPU is the system now just using the integrated GPU?

Ropid · 2019-03-12 09:23:03

storms wrote:

[...]
the change i made was in
/usr/share/X11/xorg.conf.d/10-nvidia-drm-outputclass.conf
[...]

Perhaps try to copy the file to /etc/X11/xorg.conf.d/ and then apply your change there on the copy. The idea about this would be that your change won't be overwritten when the package gets updated that owns the original file in /usr.

Last edited by Ropid (2019-03-12 09:23:20)

storms · 2019-03-13 09:10:49

loqs wrote:

Without bumblebee installed and without the nvidia GPU being set as PrimaryGPU is the system now just using the integrated GPU?

hello! sorry for the delay.
i assume so...
i have my DE all loaded up, i can watch videos and what not.
did some google:

glxinfo|egrep "OpenGL vendor|OpenGL renderer"

 
OpenGL vendor string: Intel Open Source Technology Center
OpenGL renderer string: Mesa DRI Intel(R) Haswell Mobile

lspci -vnnn | perl -lne 'print if /^\d+\:.+(\[\S+\:\S+\])/' | grep VGA

00:02.0 VGA compatible controller [0300]: Intel Corporation Haswell-ULT Integrated Graphics Controller [8086:0a16] (rev 0b) (prog-if 00 [VGA controller])

Ropid wrote:

storms wrote:
[...]
the change i made was in
/usr/share/X11/xorg.conf.d/10-nvidia-drm-outputclass.conf
[...]
Perhaps try to copy the file to /etc/X11/xorg.conf.d/ and then apply your change there on the copy. The idea about this would be that your change won't be overwritten when the package gets updated that owns the original file in /usr.

did it, do i need to change the name?
i just did a simple cp.

seth · 2019-03-13 21:47:00

Copying is ok.
You're on the intel chip now.

One thing you could try w/o installing windows is to install the 390xx packages and try whether nvidia-smi can talk to the GPU.
Do you maybe have GPU related settings in the BIOS?

storms · 2019-03-14 09:58:31

Good morning you beautiful people!

so i did the 390 thingy, lets get to it:

pacman -Qn | grep nvidia

lib32-nvidia-390xx-utils 390.116-2
nvidia-390xx 390.116-5
nvidia-390xx-utils 390.116-1
nvidia-settings 418.43-1

so, i went ahead reboot, loaded on the normal kernel ( i have been testing the zen kernel, looks good, no problems so far)
so the wiki says that i needed to use the dkms for customs karnels, butttttt i'm lazy so i just loaded on the normal kernel.
anyway my xorg fails to load.
loged in root, got a dmesg
dmesg
xorg log

edit:
looks like its the same problem to me...

dmesg

[   24.858112] NVRM: RmInitAdapter failed! (0x26:0x65:1123)
[   24.858139] NVRM: rm_init_adapter failed for device bearing minor number 0

Last edited by storms (2019-03-14 12:17:20)

seth · 2019-03-14 13:01:14

Unless proof to the contrary, the GPU looks dead :-(

Arch Linux

#1 2019-03-11 16:54:27

NVRM: RmInitAdapter failed!

#2 2019-03-11 17:25:15

Re: NVRM: RmInitAdapter failed!

#3 2019-03-11 20:49:45

Re: NVRM: RmInitAdapter failed!

#4 2019-03-11 21:53:21

Re: NVRM: RmInitAdapter failed!

#5 2019-03-11 21:56:00

Re: NVRM: RmInitAdapter failed!

#6 2019-03-11 22:01:44

Re: NVRM: RmInitAdapter failed!

#7 2019-03-11 22:06:43

Re: NVRM: RmInitAdapter failed!

#8 2019-03-11 22:10:54

Re: NVRM: RmInitAdapter failed!

#9 2019-03-11 22:15:32

Re: NVRM: RmInitAdapter failed!

#10 2019-03-11 22:30:47

Re: NVRM: RmInitAdapter failed!

#11 2019-03-11 22:46:53

Re: NVRM: RmInitAdapter failed!

#12 2019-03-11 22:59:24

Re: NVRM: RmInitAdapter failed!

#13 2019-03-11 23:01:13

Re: NVRM: RmInitAdapter failed!

#14 2019-03-12 00:31:06

Re: NVRM: RmInitAdapter failed!

#15 2019-03-12 09:23:03

Re: NVRM: RmInitAdapter failed!

#16 2019-03-13 09:10:49

Re: NVRM: RmInitAdapter failed!

#17 2019-03-13 21:47:00

Re: NVRM: RmInitAdapter failed!

#18 2019-03-14 09:58:31

Re: NVRM: RmInitAdapter failed!

#19 2019-03-14 13:01:14

Re: NVRM: RmInitAdapter failed!

Board footer