You are not logged in.
Okay, I cannot find any information about this problem anywhere and I haven't had any luck on the Bumblebee or Archlinux IRC rooms. I'd really be thankful if someone can help with this.
I have Optimus set up and working with my Intel CPU and my Nvidia 860M on my laptop. Optirun and Primusrun appear to work as expected and I can run different applications on my nvidia video card.
However, many 3D applications completely stop being responsive after a short period of time when ran using optirun or primusrun. They lock up and cease updating their display buffer. If I terminate the application in any way, either by terminating the application's process or the optirun/primusrun process, my entire system freezes up and I need to do a hard reset.
I am not sure how best to debug this. If you need me to post anything here I will gladly. Also, I had just installed Arch Linux on my laptop after using OpenSuSE. Bumblebee seemed to work fine on OpenSuSE so I doubt it is a hardware problem.
Thank you a lot for taking your time reading this, I hope you all have happy holidays and such.
Offline
Still can't figure out a way around it. This seems to happen with every program I run through primusrun/optirun eventually. The only thing I haven't noticed trigger it is unigine-heaven benchmark. But I figure it might happen eventually there too. I've tried to use linux-lts kernel. Same results. Pograms just freeze and when their process is killed the entire system freezes. It's not just xorg; killing the process from a terminal also renders the system unresponsive. Hard reboot is required. It happens across a wide range of different types of apps; blender, pymol, KSP, TF2. If anyone could help me even begin to figure out what is going on it'd be great. Thanks again for taking the time to read this!
Offline
Same problem here. Looks like anytime the application looses focus the whole system just freezes for me..
Though some applications just freeze and system is still responsive and system freezes only after killing application..
Offline
I have the same problem. Running minecraft. After some time (around 1 to 3 minutes) the game freezes. When I try to kill java the rest of the system freezes. Also `sudo reboot` won't work (just hangs). Only a hard reset using the power button works.
$ uname -r
4.3.3-2-ARCH$ pacman -Qs nvidia
local/bumblebee 3.2.1-10
NVIDIA Optimus support for Linux through VirtualGL
local/cuda 7.5.18-1
NVIDIA's GPU programming toolkit
local/lib32-nvidia-utils 358.16-1
NVIDIA drivers utilities (32-bit)
local/libvdpau 1.1.1-2
Nvidia VDPAU library
local/nvidia 358.16-3
NVIDIA drivers for linux
local/nvidia-utils 358.16-1
NVIDIA drivers utilities
local/opencl-nvidia 358.16-1
OpenCL implemention for NVIDIAThe The interesting lines in journalctl from starting Minecraft using primusrun until the reboot
Jän 09 19:26:41 burner-red kernel: bbswitch: enabling discrete graphics
Jän 09 19:26:42 burner-red kernel: nvidia: module license 'NVIDIA' taints kernel.
Jän 09 19:26:42 burner-red kernel: Disabling lock debugging due to kernel taint
Jän 09 19:26:42 burner-red kernel: ACPI Error: [AR02] Namespace lookup failure, AE_NOT_FOUND (20150818/psargs-359)
Jän 09 19:26:42 burner-red kernel: ACPI Error: Method parse/execution failed [\_SB.PCI0.PEG0._PRT] (Node ffff8802568d79d8), AE_NOT_FOUND (20150818/psparse-542)
Jän 09 19:26:42 burner-red kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 247
Jän 09 19:26:42 burner-red kernel: [drm] Initialized nvidia-drm 0.0.0 20150116 for 0000:01:00.0 on minor 1
Jän 09 19:26:42 burner-red kernel: NVRM: loading NVIDIA UNIX x86_64 Kernel Module 358.16 Mon Nov 16 19:25:55 PST 2015
Jän 09 19:26:42 burner-red kernel: vgaarb: this pci device is not a vga device
Jän 09 19:26:42 burner-red kernel: ACPI Error: [AR02] Namespace lookup failure, AE_NOT_FOUND (20150818/psargs-359)
Jän 09 19:26:42 burner-red kernel: ACPI Error: Method parse/execution failed [\_SB.PCI0.PEG0._PRT] (Node ffff8802568d79d8), AE_NOT_FOUND (20150818/psparse-542)
Jän 09 19:26:42 burner-red kernel: ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150818/nsarguments-95)
Jän 09 19:26:42 burner-red kernel: ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150818/nsarguments-95)
Jän 09 19:26:42 burner-red kernel: ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150818/nsarguments-95)
Jän 09 19:26:42 burner-red kernel: ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150818/nsarguments-95)
Jän 09 19:26:42 burner-red kernel: ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150818/nsarguments-95)
Jän 09 19:26:42 burner-red kernel: ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150818/nsarguments-95)
Jän 09 19:26:42 burner-red kernel: ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150818/nsarguments-95)
Jän 09 19:26:42 burner-red kernel: ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150818/nsarguments-95)
Jän 09 19:26:43 burner-red kernel: vgaarb: this pci device is not a vga device
Jän 09 19:26:43 burner-red kernel: nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 358.16 Mon Nov 16 18:58:09 PST 2015
Jän 09 19:26:43 burner-red bumblebeed[391]: [ 3436.010305] [WARN][XORG] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/100dpi/".
Jän 09 19:26:43 burner-red bumblebeed[391]: [ 3436.010323] [WARN][XORG] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/75dpi/".
Jän 09 19:26:43 burner-red bumblebeed[391]: [ 3436.010339] [ERROR][XORG] (EE) /dev/dri/card0: failed to set DRM interface version 1.4: Permission denied
Jän 09 19:26:43 burner-red bumblebeed[391]: [ 3436.010342] [WARN][XORG] (WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)
Jän 09 19:26:43 burner-red bumblebeed[391]: [ 3436.010345] [WARN][XORG] (WW) Warning, couldn't open module mouse
Jän 09 19:26:43 burner-red bumblebeed[391]: [ 3436.010346] [ERROR][XORG] (EE) Failed to load module "mouse" (module does not exist, 0)
Jän 09 19:26:43 burner-red bumblebeed[391]: [ 3436.010352] [WARN][XORG] (WW) NVIDIA(0): Unable to get display device for DPI computation.
Jän 09 19:26:43 burner-red bumblebeed[391]: [ 3436.010356] [WARN][XORG] (WW) Warning, couldn't open module mouse
Jän 09 19:26:43 burner-red bumblebeed[391]: [ 3436.010358] [ERROR][XORG] (EE) Failed to load module "mouse" (module does not exist, 0)
Jän 09 19:26:43 burner-red bumblebeed[391]: [ 3436.010362] [ERROR][XORG] (EE) /dev/dri/card0: failed to set DRM interface version 1.4: Permission denied
-- Reboot --
Jän 09 19:31:10 burner-red systemd-journald[163]: Runtime journal (/run/log/journal/) is currently using 8.0M.
Maximum allowed usage is set to 395.2M.
Leaving at least 592.8M free (of currently available 3.8G of space).
Enforced usage limit is thus 395.2M, of which 387.2M are still available.Offline
Does anyone know where to report a bug or how to investigate this issue further?
To see if the driver is the issue I tried to install nvidia-352 from AUR, but I must have done something wrong because bumblebee wasn't able to load the NVIDIA-driver after the downgrade.
Last edited by NeroBurner (2016-01-10 21:22:29)
Offline
The 358 drivers has been know to crash windows machines with the 860M. I downgraded to 355 and it doesn't crash. Hopefully the issue is fixed in 361.
I tried to use 361 from testing, but it is currently broken with bumblebee.
**Edit
I had to use nvidia 355.11-4 with linux 4.2.5-1 to get bumblebee to work. I can list all the packages I downgraded if needed.
Last edited by Extremeadin (2016-01-11 01:12:43)
Offline
Thanks for the tipp, But I ran into some problems.
I tried to downgrade the nvidia-driver and its dependencies with the following command.
sudo pacman -U nvidia-355.11-4-x86_64.pkg.tar.xz nvidia-utils-355.11-1-x86_64.pkg.tar.xz lib32-nvidia-utils-355.11-1-x86_64.pkg.tar.xz linux-4.2.5-1-x86_64.pkg.tar.xz bbswitch-0.8-38-x86_64.pkg.tar.xz virtualbox-host-modules-5.0.10-2.1-x86_64.pkg.tar.xzAfter a restart I tried to run glxgears on the dedicated card and got the following error:
$ primusrun glxgears
primus: fatal: Bumblebee daemon reported: error: [XORG] (EE) /dev/dri/card0: failed to set DRM interface version 1.4: Permission denied
$ optirun glxgears
[ 267.080695] [ERROR]Cannot access secondary GPU - error: [XORG] (EE) /dev/dri/card0: failed to set DRM interface version 1.4: Permission denied
[ 267.080722] [ERROR]Aborting because fallback start is disabled.Offline
I maybehave found a workaround
When running 'primusrun glxspheres' after about half a minute glxpheres freezes, but I was still able to open a shell. In the output of dmesg there was the following:
NVRM: GPU at 0000:01:00.0 has fallen off the bus.A bit searching got me an old workaround from 2012
http://www.cyberciti.biz/faq/debian-ubu … n-off-bus/
And the proposed solution is to set the NVIDIA-Card into persistence mode. Because I'm using an optimus-system the nvidia-module is not loaded at startup. My solution is to start a program with primusrun and (quickly!!!) run
sudo primusrun /usr/bin/nvidia-smi -pm 1in another terminal. Just loading the nvidia module and then set the persistence mode did not work for me
TL;DR
start your program. For example
primusrun glxspheres64and then in a seperate terminal (before the system freezes:
sudo primusrun /usr/bin/nvidia-smi -pm 1Offline
Apparently I'm also using xorg 1.17 with the old drivers. The complete list of downgraded packages I'm using is
freetype2-2.6.1-1
lib32-nvidia-utils-355.11-1
nvidia-355.11-4
nvidia-utils-355.11-1
xf86-input-evdev-2.10.0-1
xf86-input-libinput-0.15.0-1
xf86-video-intel-1_2.99.917+478+gdf72bc5-2
xorg-server-1.17.4-2
xorg-server-common-1.17.4-2
xorg-server-xwayland-1.17.4-2
bbswitch-0.8-38
linux-4.2.5-1
linux-headers-4.2.5-1
You probably don't need to downgrade freetype2, however. My fonts just all looked weird after updating it.
Additionally, the bug with primus appears to be fixed, so you could also try 361.16 from testing.
Offline
thanks for the list!
My workaround with setting persistence mode also does not work, still falling off the bus
Offline
I also have a GTX860M and 358.16-3 drivers:
$ pacman -Ss nvidia | grep -i instal
extra/libvdpau 1.1.1-2 [installed]
extra/nvidia 358.16-3 [installed]
extra/nvidia-utils 358.16-1 [installed]
extra/opencl-nvidia 358.16-1 [installed]
community/bumblebee 3.2.1-10 [installed]
community/cuda 7.5.18-1 [installed]I can run glxspheres64 with primusrun and optirun (and also in cpu).
I am not running any 3D stuff for the last weeks, only cuda stuff and is running ok.
I only discovered I have some kind of problems because i was mounting a USB flash drive and saw my dmesg shouting like crazy:
[25350.680162] ACPI Error: [AR02] Namespace lookup failure, AE_NOT_FOUND (20150818/psargs-359)
[25350.680182] ACPI Error: Method parse/execution failed [\_SB.PCI0.PEG0._PRT] (Node ffff8802568d7a00), AE_NOT_FOUND (20150818/psparse-542)
[25350.686214] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150818/nsarguments-95)These lines are repeated over and over.
Anyone knows when the 361 drivers will leave testing?
The bug with bumblebee (https://bugs.archlinux.org/task/47688) was fixed some days ago.
Offline
My guess is that they are waiting until the Civilization 5 crash is fixed before it gets released.
Offline
The new nvidia 361.18-3 driver came out and all works fine again
Offline