[SOLVED] Asus N580VD problems with bbswitch/bumblebee (GTX 1050 GPU)

cdutsov · 2018-10-17 10:20:11

Hello everyone!

I have Asus N580VD-FY588 laptop and I have read every thread in every forum and still didn't manage to get the discrete GPU working. I am using the latest kernel version: 4.18.14-arch1-1-ARCH, and I have tried bbswitch, bbswitch-dkms, nvidia-410, nvidia-390, nvidia-lts and every possible package. My greatest concern is that sometimes (very rarely) the card does work and I am able to run "optirun glxgears" for example, but after I stop optirun, the card doesn't turn on ever again.

My current setup is:

lspci -v:
01:00.0 3D controller: NVIDIA Corporation GP107M [GeForce GTX 1050 Mobile] (rev ff) (prog-if ff)
	!!! Unknown header type 7f
	Kernel modules: nouveau, nvidia_drm, nvidia

GRUB_CMDLINE_LINUX_DEFAULT="verbose acpi_osi=! acpi_osi=\"Windows 2009\" rcutree.rcu_idle_gp_delay=1"

pacman -Q | grep nvidia:

nvidia-390xx-dkms 390.87-6
nvidia-390xx-settings 390.87-1
nvidia-390xx-utils 390.87-1

I get the following error on boot, after a 30 second delay:

[   36.080473] ACPI Error: Method parse/execution failed \_SB.PCI0.GPPR, AE_AML_LOOP_TIMEOUT (20180531/psparse-516)
[   36.082572] ACPI Error: Method parse/execution failed \_SB.PCI0.PGON, AE_AML_LOOP_TIMEOUT (20180531/psparse-516)
[   36.084681] ACPI Error: Method parse/execution failed \_SB.PCI0.PEG0.PEGP._ON, AE_AML_LOOP_TIMEOUT (20180531/psparse-516)
[   36.085988] ACPI Error: Method parse/execution failed \_SB.PCI0.PEG0.PEGP._PS0, AE_AML_LOOP_TIMEOUT (20180531/psparse-516)

After the error is shown, the computer boots up.

This is the error I get with optirun:

[user@computer][~] > optirun -vv glxgears                                                                                                                                                                         :)
[ 1375.115242] [DEBUG]Reading file: /etc/bumblebee/bumblebee.conf
[ 1375.115358] [INFO]Configured driver: nvidia
[ 1375.115634] [DEBUG]optirun version 3.2.1 starting...
[ 1375.115652] [DEBUG]Active configuration:
[ 1375.115654] [DEBUG] bumblebeed config file: /etc/bumblebee/bumblebee.conf
[ 1375.115656] [DEBUG] X display: :8
[ 1375.115658] [DEBUG] LD_LIBRARY_PATH: /usr/lib/nvidia:/usr/lib32/nvidia:/usr/lib:/usr/lib32
[ 1375.115660] [DEBUG] Socket path: /var/run/bumblebee.socket
[ 1375.115680] [DEBUG] Accel/display bridge: auto
[ 1375.115681] [DEBUG] VGL Compression: proxy
[ 1375.115684] [DEBUG] VGLrun extra options: 
[ 1375.115705] [DEBUG] Primus LD Path: /usr/lib/primus:/usr/lib32/primus
[ 1375.115731] [DEBUG]Using auto-detected bridge virtualgl
[ 1375.781731] [INFO]Response: No - error: Could not enable discrete graphics card

[ 1375.781771] [ERROR]Cannot access secondary GPU - error: Could not enable discrete graphics card

[ 1375.781782] [DEBUG]Socket closed.
[ 1375.781815] [ERROR]Aborting because fallback start is disabled.
[ 1375.781825] [DEBUG]Killing all remaining processes.

I have tried all possible kernel parameter combinations but none seem to work. I have read somewhere that if the card is turned off, sometimes it can't turn on again and a hard reset is needed. This laptop doesn't have an easy access to the battery so I can't take it out easily.

Has anyone managed to run the Nvidia GTX 1050 GPU on this Asus N580 laptop?

Thanks in advance!

EDIT:
It turned out that the GPU was fried. I got a new motherboard and everything is working out of the box now! Thanks for the replies!

Last edited by cdutsov (2019-01-15 19:05:54)

anotherAlex154 · 2018-10-20 21:39:43

try to get the status of your discrete card with:

 cat /proc/acpi/bbswitch

if it's "OFF" then turn it back on with:

 tee /proc/acpi/bbswitch <<<ON

then check the status again. if it doesn't cut on then the error will be in:

 dmesg |tail -1

perhaps some of your drivers are getting in the way. could you also run:

 xrandr --listproviders

https://github.com/Bumblebee-Project/bbswitch

Last edited by anotherAlex154 (2018-10-20 21:44:42)

tejonaco · 2018-10-20 21:51:36

Some time ago I made an script to install bumblebee with nvidia 390xx driver, maybe it works too for you.

#!/bin/bash

#Script preparado para tarjeta nvidia GeForce 820M, puede no funcionar con otras

#Instalación de los paquetes necesarios
sudo pacman -S --noconfirm --needed mesa
sudo pacman -S --noconfirm --needed nvidia-390xx nvidia-390xx-utils
sudo pacman -S --noconfirm --needed nvidia-390xx-settings
sudo pacman -S --noconfirm --needed xf86-video-intel
sudo pacman -S --noconfirm --needed bumblebee
sudo pacman -S --noconfirm --needed lib32-nvidia-390xx-utils
sudo pacman -S --noconfirm --needed primus
sudo pacman -S --noconfirm --needed lib32-primus
sudo pacman -S --noconfirm --needed bbswitch

#Activar el servicio de bumblebee y añadir al usuario al grupo bumblebee

sudo systemctl enable bumblebeed
sudo gpasswd -a $USER bumblebee

#Poner a nouveau en la lista negra
sudo touch /etc/modprobe.d/blacklist-nouveau.conf
sudo echo "blacklist nouveau" | sudo tee -a /etc/modprobe.d/blacklist-nouveau.conf
sudo echo "options nouveau modeset=0" | sudo tee -a /etc/modprobe.d/blacklist-nouveau.conf
sudo mkinitcpio -p linux

#Nota: si se quiere modificar luego la resolución el comando para abrir nvidia_settings es:
#sudo optirun nvidia-settings -c :8

Anotations are in spanish, but I trust you manage well with this and google translate, but you shoud'nt need it.

cdutsov · 2018-10-21 06:12:52

anotherAlex154 wrote:

try to get the status of your discrete card with:
 cat /proc/acpi/bbswitch 
if it's "OFF" then turn it back on with:
 tee /proc/acpi/bbswitch <<<ON 
then check the status again. if it doesn't cut on then the error will be in:
 dmesg |tail -1 
perhaps some of your drivers are getting in the way. could you also run:
 xrandr --listproviders 
https://github.com/Bumblebee-Project/bbswitch

[me@rambo][~] > cat /proc/acpi/bbswitch                                                                                                                                                                  130 :(
0000:01:00.0 OFF
[me@rambo][~] > sudo tee /proc/acpi/bbswitch <<< ON                                                                                                                                                          :)
ON
[me@rambo][~] > dmesg | tail -1                                                                                                                                                                              :)
[  272.693656] pci 0000:01:00.0: Refused to change power state, currently in D3
[me@rambo][~] > xrandr --listproviders                                                                                                                                                                       :)
Providers: number : 1
Provider 0: id: 0x48 cap: 0xb, Source Output, Sink Output, Sink Offload crtcs: 4 outputs: 5 associated providers: 0 name:Intel
[me@rambo][~] >                                                                                                                                                                                              :)

This is the result I get. I have TLP installed but I have blacklisted the Nvidia PCI bus. I have also tried with the kernel parameter pcie_port_pm=off but to no avail.

If I try to modprobe nvidia I get the following messages in dmesg:

[ 1084.767549] nvidia: module license 'NVIDIA' taints kernel.
[ 1084.767553] Disabling lock debugging due to kernel taint
[ 1084.803729] nvidia-nvlink: Nvlink Core is being initialized, major device number 237
[ 1084.804040] nvidia 0000:01:00.0: Refused to change power state, currently in D3
[ 1084.804048] NVRM: This is a 64-bit BAR mapped above 4GB by the system
               NVRM: BIOS or the Linux kernel, but the PCI bridge
               NVRM: immediately upstream of this GPU does not define
               NVRM: a matching prefetchable memory window.
[ 1084.804048] NVRM: This may be due to a known Linux kernel bug.  Please
               NVRM: see the README section on 64-bit BARs for additional
               NVRM: information.
[ 1084.804273] nvidia: probe of 0000:01:00.0 failed with error -1
[ 1084.804294] NVRM: The NVIDIA probe routine failed for 1 device(s).
[ 1084.804295] NVRM: None of the NVIDIA graphics adapters were initialized!
[ 1084.804418] nvidia-nvlink: Unregistered the Nvlink Core, major device number 237

Last edited by cdutsov (2018-10-21 06:24:31)

cdutsov · 2018-10-21 06:15:12

tejonaco wrote:

Some time ago I made an script to install bumblebee with nvidia 390xx driver, maybe it works too for you.

#!/bin/bash

#Script preparado para tarjeta nvidia GeForce 820M, puede no funcionar con otras

#Instalación de los paquetes necesarios
sudo pacman -S --noconfirm --needed mesa
sudo pacman -S --noconfirm --needed nvidia-390xx nvidia-390xx-utils
sudo pacman -S --noconfirm --needed nvidia-390xx-settings
sudo pacman -S --noconfirm --needed xf86-video-intel
sudo pacman -S --noconfirm --needed bumblebee
sudo pacman -S --noconfirm --needed lib32-nvidia-390xx-utils
sudo pacman -S --noconfirm --needed primus
sudo pacman -S --noconfirm --needed lib32-primus
sudo pacman -S --noconfirm --needed bbswitch

#Activar el servicio de bumblebee y añadir al usuario al grupo bumblebee

sudo systemctl enable bumblebeed

sudo gpasswd -a $USER bumblebee

#Poner a nouveau en la lista negra
sudo touch /etc/modprobe.d/blacklist-nouveau.conf
sudo echo "blacklist nouveau" | sudo tee -a /etc/modprobe.d/blacklist-nouveau.conf
sudo echo "options nouveau modeset=0" | sudo tee -a /etc/modprobe.d/blacklist-nouveau.conf
sudo mkinitcpio -p linux

#Nota: si se quiere modificar luego la resolución el comando para abrir nvidia_settings es:
#sudo optirun nvidia-settings -c :8

Anotations are in spanish, but I trust you manage well with this and google translate, but you shoud'nt need it.

Thank you for the reply but I have done all the steps in the script manualy as well as I have all the packages installed.

cdutsov · 2018-10-21 15:31:01

Ok, I managed to start the card again by totally disabling TLP. After it ran for a few minutes I got the following error:

[  968.443750] bbswitch: enabling discrete graphics
[  968.746950] nvidia-nvlink: Nvlink Core is being initialized, major device number 237
[  968.847775] NVRM: loading NVIDIA UNIX x86_64 Kernel Module  410.57  Tue Sep 18 23:25:09 CDT 2018 (using threaded interrupts)
[  969.625808] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms  410.57  Tue Sep 18 23:23:43 CDT 2018
[ 1003.759700] NVRM: GPU at PCI:0000:01:00: GPU-e3a9bba3-0564-5a5f-313c-cd6c7882093e
[ 1003.759706] NVRM: Xid (PCI:0000:01:00): 31, Ch 00000012, engmask 00000110, intr 10000000
[ 1003.763729] NVRM: Xid (PCI:0000:01:00): 62, 0a8a(2a88) 00000000 00000000
[ 1007.761767] NVRM: Xid (PCI:0000:01:00): 44, Ch 00000010, engmask 00000101, intr 10000000

Now it won't run after a reboot. Any ideas?

radiomike · 2018-11-05 12:17:23

With this card you should be using the latest drivers. In your case nvidia-dkms, would seem to make most sense.

The 390 version drivers are for older hardware and may not fully support newer cards. Nvidia-lts is just for the lts kernel, using the dkms version will build modules for any installed kernels.

I have similar hardware and wasn't able to get bumblebee/bbswitch to sort out power management for this card, although didn't try very hard as I manly use the machine plugged in and use nvidia-xrun which gives better performance than optirun or primusrun.

If you want to get power management running you may want to look at the dell xps 9570 thread. Also you may see different results depending on whether your machine is plugged in or not.

Last edited by radiomike (2018-11-05 12:38:09)

cdutsov · 2018-11-09 21:09:01

Thank you for your reply! I have tried the newest drivers also, but to no avail. I always get the RmInitAdapter Fail error.
I did try nvidia-xrun also but I get the same error when the nvidia module is loaded.

Arch Linux

#1 2018-10-17 10:20:11

[SOLVED] Asus N580VD problems with bbswitch/bumblebee (GTX 1050 GPU)

#2 2018-10-20 21:39:43

Re: [SOLVED] Asus N580VD problems with bbswitch/bumblebee (GTX 1050 GPU)

#3 2018-10-20 21:51:36

Re: [SOLVED] Asus N580VD problems with bbswitch/bumblebee (GTX 1050 GPU)

#4 2018-10-21 06:12:52

Re: [SOLVED] Asus N580VD problems with bbswitch/bumblebee (GTX 1050 GPU)

#5 2018-10-21 06:15:12

Re: [SOLVED] Asus N580VD problems with bbswitch/bumblebee (GTX 1050 GPU)

#6 2018-10-21 15:31:01

Re: [SOLVED] Asus N580VD problems with bbswitch/bumblebee (GTX 1050 GPU)

#7 2018-11-05 12:17:23

Re: [SOLVED] Asus N580VD problems with bbswitch/bumblebee (GTX 1050 GPU)

#8 2018-11-09 21:09:01

Re: [SOLVED] Asus N580VD problems with bbswitch/bumblebee (GTX 1050 GPU)

Board footer