You are not logged in.

#1 2017-02-09 17:29:28

nesquil
Member
Registered: 2017-02-09
Posts: 2

[Solved] Can't turn off nvidia card

Hello arch forums,
I have been going through posts and git commit comments, and I believe I found my issue with its fix as well. However, it is an old commit and I imagine the bumblebee-git source from the aur has it already included.
This is a similar issue but with an older driver. I am on NVidia 375.26-6.
https://github.com/Bumblebee-Project/Bu … issues/732
This is the commit to bumblebee/src/module that should(?) fix my driver unloading issue.
https://github.com/Bumblebee-Project/Bu … 454de8a04a

I believe this is because in order to execute

startx

without freezing when loading the config file, I have to execute

sudo modprobe nvidia

. I have to do this after switching to another virtual terminal to avoid xorg autorunning. I don't get any error logs from Xorg when it hangs at the config file.

Please help me figure out how to get my nvidia card off. I have

optirun glxspheres64 --info

working, which is good because it used to just do a segmentation fault. 
bbswitch reports my card on and

sudo tee /proc/acpi/bbswitch <<< OFF

does not help.

                   -`                    nesquil@hermes 
                  .o+`                   -------------- 
                 `ooo/                   OS: Arch Linux x86_64 
                `+oooo:                  Model: P65_67RSRP 
               `+oooooo:                 Kernel: 4.9.7-1-ARCH 
               -+oooooo+:                Uptime: 12 hours, 24 mins 
             `/:-:++oooo+:               Packages: 781 
            `/++++/+++++++:              Shell: bash 4.4.11 
           `/++++++++++++++:             Resolution: 1920x1080 
          `/+++ooooooooooooo/`           WM: i3 
         ./ooosssso++osssssso+`          Theme: Adwaita-dark [GTK2/3] 
        .oossssso-````/ossssss+`         Icons: Adwaita [GTK2/3] 
       -osssssso.      :ssssssso.        Terminal: urxvt 
      :osssssss/        osssso+++.       CPU: Intel i7-6700HQ (8) @ 3.5GHz 
     /ossssssss/        +ssssooo/-       GPU: Intel Integrated Graphics 
   `/ossssso+/:-        -:/+osssso+-     GPU: NVIDIA GeForce GTX 1060 
  `+sso+:-`                 `.-/+oso:    Memory: 2415MiB / 15935MiB 

edit: 1 application using bumblebeed? is it bumblebeed itself or is this my problem

[nesquil@hermes .scripts]$ optirun --status
Bumblebee status: Ready (3.2.1-50-gd469a32). X is PID 8873, 1 applications using bumblebeed.

Bits of dmesg --human

[  +0.242876] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms  375.26  Thu Dec  8 18:04:14 PST 2016
[  +0.000473] nvidia-modeset: Allocated GPU:0 (GPU-b1d865e2-e5bb-468d-4bc0-1198b00c8681) @ PCI:0000:01:00.0
[Feb 9 00:52] bbswitch: device 0000:01:00.0 is in use by driver 'nvidia', refusing OFF
[Feb 9 00:53] bbswitch: device 0000:01:00.0 is in use by driver 'nvidia', refusing OFF
[Feb 9 00:57] glxspheres64[8932]: segfault at 0 ip 00007f8103c913d2 sp 00007ffdc4fff540 error 4 in libGL.so.1[7f8103c6c000+41000]
[Feb 9 00:58] glxspheres64[8952]: segfault at 0 ip 00007f6f49c3a3d2 sp 00007ffe33487870 error 4 in libGL.so.1[7f6f49c15000+41000]
...
...
[Feb 9 10:03] nvidia-modeset: Unloading
[  +2.735763] nvidia-nvlink: Unregistered the Nvlink Core, major device number 244
[  +0.000024] NVRM: Attempting to remove minor device 0 with non-zero usage count!
[  +0.000001] ------------[ cut here ]------------
[  +0.000098] WARNING: CPU: 6 PID: 16914 at /build/nvidia/src/NVIDIA-Linux-x86_64-375.26-no-compat32/kernel/nvidia/nv.c:3498 nvidia_remove+0x2d3/0x360 [nv
[  +0.000001] Modules linked in: ath9k_htc ath9k_common ath9k_hw ath uinput hid_generic usbhid hid tun ax25 rfcomm nvidia(PO-) ctr ccm bnep btusb uvcvideo
[  +0.000032]  psmouse shpchp i2c_smbus intel_rapl_perf led_class pcspkr thermal fjes battery wmi evdev tpm_tis tpm_tis_core tpm acpi_pad ac mac_hid sch_f
[  +0.000023] CPU: 6 PID: 16914 Comm: rmmod Tainted: P  R  U     O    4.9.7-1-ARCH #1
[  +0.000001] Hardware name: Notebook                        P65_67RSRP/P65_67RSRP, BIOS 1.05.02LS1 07/23/2016
[  +0.000001]  ffffc900065abd38 ffffffff813055b0 0000000000000000 0000000000000000
[  +0.000002]  ffffc900065abd78 ffffffff8107eb0b 00000daaba4cb520 ffff88040ac13000
[  +0.000002]  ffff8804a01fb800 ffff8804ab958000 dead000000000100 000000000101a010
[  +0.000002] Call Trace:
[  +0.000004]  [<ffffffff813055b0>] dump_stack+0x63/0x83
[  +0.000001]  [<ffffffff8107eb0b>] __warn+0xcb/0xf0
[  +0.000002]  [<ffffffff8107ec3d>] warn_slowpath_null+0x1d/0x20
[  +0.000066]  [<ffffffffa0a87453>] nvidia_remove+0x2d3/0x360 [nvidia]
[  +0.000002]  [<ffffffff8135c839>] pci_device_remove+0x39/0xc0
[  +0.000002]  [<ffffffff81451d31>] __device_release_driver+0xa1/0x160
[  +0.000002]  [<ffffffff81452964>] driver_detach+0xb4/0xc0
[  +0.000001]  [<ffffffff81451725>] bus_remove_driver+0x55/0xd0
[  +0.000002]  [<ffffffff814531bc>] driver_unregister+0x2c/0x50
[  +0.000001]  [<ffffffff8135afa0>] pci_unregister_driver+0x40/0xa0
[  +0.000051]  [<ffffffffa0a84a5a>] nvidia_exit_module+0x3a/0xb0 [nvidia]
[  +0.000076]  [<ffffffffa105b905>] nvidia_frontend_exit_module+0x2a/0x2c [nvidia]
[  +0.000003]  [<ffffffff8110ae02>] SyS_delete_module+0x192/0x260
[  +0.000001]  [<ffffffff810035f8>] ? exit_to_usermode_loop+0x48/0xb0
[  +0.000002]  [<ffffffff8160a737>] entry_SYSCALL_64_fastpath+0x1a/0xa9
[  +0.000015] ---[ end trace c4f929d5bb8419f8 ]---
[Feb 9 10:05] bbswitch: device 0000:01:00.0 is in use by driver 'nvidia', refusing OFF
[Feb 9 10:06] bbswitch: device 0000:01:00.0 is in use by driver 'nvidia', refusing OFF

edit:
after adding pcie_port_pm=0 to my boot commands, and made only one xorg.conf file which had a section for my "card0" (shows up in org log as /dri/card0, those who have this issue will know) device to use the intel driver. I kept having a no devices found after getting only my intel card on. Now primusrun works wonderfully. smile

Last edited by nesquil (2017-02-09 23:47:49)

Offline

Board footer

Powered by FluxBB