You are not logged in.
Dear Arch community,
I have problems in getting my nvidia card to work using the prop drivers and bumblebee.
I followed the steps in bumblee wiki entry: https://wiki.archlinux.org/index.php/Bu … l.2FNVIDIA
pacman -S intel-dri xf86-video-intel bumblebee nvidia
added user to bimblebee-group, but unfortunately I get the following error:
If a run the command for the 1. time:
sudo optirun glxgears -info
[ 1168.770796] [ERROR]Cannot access secondary GPU - error: [XORG] (EE) Failed to load module "mouse" (module does not exist, 0)[ 1168.770834] [ERROR]Aborting because fallback start is disabled.
If I round it a 2. time:
sudo optirun glxgears -info
[ 2322.784161] [ERROR]Cannot access secondary GPU - error: [XORG] (EE) Server terminated successfully (0). Closing log file.[ 2322.784196] [ERROR]Aborting because fallback start is disabled.
The help section in the wiki regarding this problem (https://wiki.archlinux.org/index.php/Bu … ondary_GPU) doesnt help me. Additionally I do not have a 20-intel.conf file as is mentioned in one of the solutions there (why??)
cd /etc/X11/xorg.conf.d/
ls
00-keyboard.conf 10-evdev.conf 10-quirks.conf 50-synaptics.conf
some further information:
sudo lspci
00:02.0 VGA compatible controller: Intel Corporation Haswell-ULT Integrated Graphics Controller (rev 09)
01:00.0 3D controller: NVIDIA Corporation GK107M [GeForce GT 750M] (rev a1)lspci -v | perl -ne '/VGA/../^$/ and /VGA|Kern|3D|Display/ and print'
00:02.0 VGA compatible controller: Intel Corporation Haswell-ULT Integrated Graphics Controller (rev 09) (prog-if 00 [VGA controller])
Kernel driver in use: i915
Kernel modules: i915sudo lspci -d10de: -vvv01:00.0 3D controller: NVIDIA Corporation GK107M [GeForce GT 750M] (rev ff) (prog-if ff)
!!! Unknown header type 7f
Kernel modules: nouveau, nvidia
The wiki states, the nvidia runs out of the box:
https://wiki.archlinux.org/index.php/Ac … re_V5-573G
Everything pretty much works out of the box, follow standard documentation for details.
I found a problem which seems maybe related:
https://devtalk.nvidia.com/default/topi … r-crash/1/
But my knowledge is to limited (1 year in linux) to solve this problem on my own.
Hoping for help,
bkk
Last edited by bkk (2014-02-26 15:02:18)
Offline
Ah I replied to you in the screenshot forum. It seems that the mouse error can be solved by installing xf86-input-mouse. which is what I did to solve that error. You may also try installing linux-headers and rebuilding the driver against the current kernal. I have done all of these things, and have solved one error at a time only to be met with the error, which says that the secondary GPU cannot be initilized at its appropriate PCI address. All attempts to trouble shoot and fix this like editing configs and such results in bumblebee crashing and no longer running at boot. I have attempted to define the Nvidia card via the instructions on the wiki from both bumblebee and the nvidia page. I have tried changing the pci number in both bumblebee.conf and the xorg.conf.nvidia and in 20-nvidia.conf. I attempted to make the information between those files match. I also have removed and reinstalled the packages several times to attempt to reset them and start over. The official nvidia driver in community is flagged out of date, so I even tried nvidia-all from the aur and to no avail. I have read many accounts of various people having these problems, but they have all been several months to a year or two ago, and any fixes either do not work or fix something that is not broken on my setup. I would post configs or output, but the entire bumblebee service is broken and any output is from behavior that has been regressed further to the point where bumblebee will no longer run. My card is no longer detected as even existing on this computer. After a couple of reboots, it will show up again. Upon reinstalling bumblebee and all associated packages with the proprietary driver. The service will work again and the power saving effects will be achieved, but any attempt to start the card will result in the same thing that I started with, because I have repeated this several times. What I am trying to say is that if your problem resembles mine at all, you may be running in circles.
Last edited by agahnim (2014-02-27 03:13:27)
Offline
[This issue has been solved for me. Scroll down to "Edit3" for the solution]
I'm facing the same problems on a Dell M3800 Notebook, including a Quadro K1100M and HD4600 iGPU. There could be something wrong with my configuration, but I also think, there is some upstream bug, as I'm facing the same issue:
[ERROR]Cannot access secondary GPU - error: [XORG] (EE) Failed to load module "mouse" (module does not exist, 0)Installing xf86-input-mouse and rebooting results in the same error as agahnim already stated: "Cannot initialize GPU at PCI:2:0:0", eventhough the PCI address is correct (lscpi tells me 02:00.0). So far, I've changed BusID to PCI:2:0:0 in bumblebee configs - there is no xorg.conf file and I also don't have any intel or nvidia specific conf files in xorg.conf.d. Everything else is on stock.
Of course, I've tried numerous xorg.conf files, too, but nothing successful. I'm able to disable the Nvidia graphics and save power, but there is no way to run optirun. As stated in the Docs, I've installed (only relevant packages):
xf86-video-intel + deps
nvidia + deps
mesa
bbswitch
bumblebee
Installing and running primus didn't help; it resulted in the same errors as optirun. So, do you think, that this is a configuration related error or really something upstream, probably affecting GK107+ GPUs and / or Haswell systems? There were some documentations on how to enable bumblebee for Linux on my specific machine, but they didn't help aswell. They were made for Ubuntu 13.10 (in contrast, I'm running latest software, including Kernel 3.13 and latest Nvidia driver).
//Edit: This is my dmesg. If I interpret the error messages right, it seems to be an ACPI bug, related to my Laptop.
[ 276.082268] bbswitch: enabling discrete graphics
[ 276.717738] nvidia: module license 'NVIDIA' taints kernel.
[ 276.717742] Disabling lock debugging due to kernel taint
[ 276.724937] [drm] Initialized nvidia-drm 0.0.0 20130102 for 0000:02:00.0 on minor 1
[ 276.724941] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 331.49 Wed Feb 12 20:42:50 PST 2014
[ 276.851917] vgaarb: this pci device is not a vga device
[ 276.853336] nvidia 0000:02:00.0: irq 54 for MSI/MSI-X
[ 276.859491] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20131115/nsarguments-95)
[ 276.859534] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20131115/nsarguments-95)
[ 276.859556] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20131115/nsarguments-95)
[ 276.859575] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20131115/nsarguments-95)
[ 276.859593] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20131115/nsarguments-95)
[ 276.859612] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20131115/nsarguments-95)
[ 276.859642] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20131115/nsarguments-95)
[ 276.859661] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20131115/nsarguments-95)
[ 282.653640] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20131115/nsarguments-95)
[ 282.679204] NVRM: GPU at 0000:02:00.0 has fallen off the bus.
[ 282.679212] NVRM: os_pci_init_handle: invalid context!
[ 282.679214] NVRM: os_pci_init_handle: invalid context!
[ 282.679218] NVRM: GPU at 0000:02:00.0 has fallen off the bus.
[ 282.679220] NVRM: os_pci_init_handle: invalid context!
[ 282.679221] NVRM: os_pci_init_handle: invalid context!
[ 282.696387] NVRM: RmInitAdapter failed! (0x25:0x28:1156)
[ 282.696395] NVRM: rm_init_adapter failed for device bearing minor number 0
[ 282.696410] NVRM: nvidia_frontend_open: minor 0, module->open() failed, error -5This could be related, too (appeared upon booting - I've cut out obviously unrelated logs, like Touchpad driver loaded etc.):
[ 11.739221] bbswitch: version 0.8
[ 11.739224] bbswitch: Found integrated VGA device 0000:00:02.0: \_SB_.PCI0.GFX0
[ 11.739227] bbswitch: Found discrete VGA device 0000:02:00.0: \_SB_.PCI0.PEG0.PEGP
[ 11.739238] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20131115/nsarguments-95)
[ 11.739304] bbswitch: detected an Optimus _DSM function
[ 11.739315] pci 0000:02:00.0: enabling device (0006 -> 0007)
[ 11.739355] bbswitch: Succesfully loaded. Discrete card 0000:02:00.0 is on
[ 11.740352] bbswitch: disabling discrete graphics
[ 11.740358] ACPI Warning: \_SB_.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20131115/nsarguments-95)
[ 13.355100] [drm] Enabling RC6 states: RC6 on, RC6p off, RC6pp off
[ 13.405069] Console: switching to colour frame buffer device 240x67
[ 13.412109] i915 0000:00:02.0: fb0: inteldrmfb frame buffer device
[ 13.412110] i915 0000:00:02.0: registered panic notifier
[ 13.412221] [Firmware Bug]: ACPI(PEGP) defines _DOD but not _DOS
[ 13.432539] ACPI: Video Device [PEGP] (multi-head: yes rom: yes post: no)
[ 13.432615] input: Video Bus as /devices/LNXSYSTM:00/device:00/PNP0A08:00/device:44/LNXVIDEO:00/input/input12
[ 13.433102] ACPI: Video Device [GFX0] (multi-head: yes rom: no post: no)
[ 13.433516] acpi device:4b: registered as cooling_device11
[ 13.433647] input: Video Bus as /devices/LNXSYSTM:00/device:00/PNP0A08:00/LNXVIDEO:01/input/input13
[ 13.434589] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
[ 13.434673] ACPI Warning: 0x000000000000f040-0x000000000000f05f SystemIO conflicts with Region \_SB_.PCI0.SBUS.SMBI 1 (20131115/utaddress-251)
[ 13.434677] ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver//Edit2: This issue seems related: https://bbs.archlinux.org/viewtopic.php?pid=1384782. Okay, so it sounds like a Linux upstream bug.
//Edit3: Possible workaround for now, is adding "rcutree.rcu_idle_gp_delay=1" to Kernel line. This fixed my problem for now. Don't forget to install xf86-input-mouse and changing bumblebee configs according to your BusID, which can be found with "lspci". Good luck!
Last edited by tolga9009 (2014-03-01 19:14:21)
Offline
@tolga9009 Exactly the same behavior I had. When I tried to define the hardware with the relavent configs, the card would play a vanishing act. It reappears in the device tree once I reboot it the next day. I then have to do a clean install of the Bumblebee set up and get the powersaving effect, but still any attempts to rectify the issue have been naught.
Offline
@agahnim: My solution is consitently working now, even between reboots. I'm using almost stock configs, only BusID configured according to my card and the kernel parameter "rcutree.rcu_idle_gp_delay=1" added to my bootloader. No xorg.conf and no other GPU specific xorg.conf.d file. It sounds like bumblebee doesn't enable your GPU upon shutting down, which can lead to such problems. Usual behaviour should be that your GPU get disabled when bumblebeed starts and enabled when bumblebeed stops. Verify this - easily done with reading the power consumption via powerTop for example.
Also, please provide your full system specs, including laptop model, iGPU and dGPU. Sometimes config files aren't deleted via "pacman -Rsnc" (for whatever reason, I don't know why) command - you need to manually delete them. Delete /etc/bumblebee and reinstall bumblebee.
Offline
@agahnim: My solution is consitently working now, even between reboots. I'm using almost stock configs, only BusID configured according to my card and the kernel parameter "rcutree.rcu_idle_gp_delay=1" added to my bootloader. No xorg.conf and no other GPU specific xorg.conf.d file. It sounds like bumblebee doesn't enable your GPU upon shutting down, which can lead to such problems. Usual behaviour should be that your GPU get disabled when bumblebeed starts and enabled when bumblebeed stops. Verify this - easily done with reading the power consumption via powerTop for example.
Also, please provide your full system specs, including laptop model, iGPU and dGPU. Sometimes config files aren't deleted via "pacman -Rsnc" (for whatever reason, I don't know why) command - you need to manually delete them. Delete /etc/bumblebee and reinstall bumblebee.
Thank you for the reply. I will be attempting your solution with adding the display command along with this once I finish a college application that I am on. I will post back with details. Thank you very much. It looks like this may work. The reason that I did not post info before, was that when I posted my GPU had already vanished temporarily and the results of any commands would have been different than what they actually are after a vanilla install. I will be getting back to you shortly to verify if I am successful.
*EDIT* Adding that line to the kernel fixed my issue. Thanks a whole lot!!
Last edited by agahnim (2014-03-01 23:11:40)
Offline
Glad that I could help you! Hmm, this is probably the fix for the thread author too. I've also seen some other threads here, which are related to this issue. Next step would be identifying the upstream change causing this bug (probably kernel, because this is happening to nouveau and nvidia) - this occured around 2 - 4 weeks ago, so something late 3.12 related and still present in 3.13. I don't know, if this even could be ArchLinux related or not - haven't tried out other distros yet.
Offline
Yea, I think it is *possibly* arch specific, but I cannot say for sure. The reason that I say that is because the only documented issues that I have seen are from arch users. I definatly agree that it is a kernel issue though. They did take a long time releasing 3.13, so it may be some patching that they had to do...
Offline
I'll take a look into ArchLinux bug reports, remove the kernel parameter, produce the error messages and file a new bug report. The developers will probably know better, if there was any Arch-related patch, which might have caused this issue. Afaik, this bug was present in earlier versions of the Linux kernel and was fixed in either 3.8 or 3.10. Ubuntu 13.10 users don't suffer from this issue, as there are reports about my specific machine running totally well under Ubuntu 13.10.
//Edit: Okay, there is already a bug report, which is basically the same error: https://bugs.archlinux.org/task/39092
Last edited by tolga9009 (2014-03-02 12:16:27)
Offline
Yes. I have seen the errors on other distros anywhere from 6 months to 2 years ago, but they stopped after that, and any of the solutions that worked for those people did not work anymore. Any of the recent reports ,which have been very sparse, seem to come for arch users. Thank you for all of the help though. I stopped looking when I figured it was going to be a kernel problem and was going to wait for a fix since I did not really need my graphics card at the moment and there were very few reports that I could find of this. The original poster of this thread was the first person who I had seen encounter the problem.
Last edited by agahnim (2014-03-02 17:38:03)
Offline
Hi there!
I've experienced the exact same problem on my Acer Aspire V3 with NVIDIA 650M and bumblebee.
As tolga9009 pointed out, adding the following to /etc/grub/grub.cfg (or to /etc/default/grub) to the kernel line will fix this problem:
rcutree.rcu_idle_gp_delay=1I've had this problem since linux 3.10.
There are 10 kinds of people in this world - Those who know binary and those who don't.
Offline