Aarch64: failed to register with the ACPI subsystem!

Graphics / Linux Linux
1

Hi,
i’m running the drivers from

https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/sbsa/

installed according to this page

however, the seem to require acpi. We’re booting the lx2160a through uboot so i dont think acpi will work at all.

[  111.138305] [drm] [nvidia-drm] [GPU ID 0x00010100] Loading driver
[  111.138311] [drm] Initialized nvidia-drm 0.0.0 20160202 for 0001:01:00.0 on minor 0
[  134.472580] NVRM: failed to register with the ACPI subsystem!
[  134.704571] NVRM: GPU 0001:01:00.0: RmInitAdapter failed! (0x31:0xffff:2268)
[  134.704713] NVRM: GPU 0001:01:00.0: rm_init_adapter failed, device minor number 0
[  134.827918] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000001

is there a different driver that doesnt require a bios?

drivers installed via NVIDIA-Linux-aarch64-470.57.02.run behave the same

vidia-bug-report.sh will not work btw, because nvidia-smi hangs after kernel crashes

2

Same here. Just got my SolidRun Honeycomb LX2 with LX2160A as well, plugged in a 2060 to go with it. Trying to fire up the Nvidia Aarch64 drivers for my 2060.

dmesg shows the same error

[    8.576481] nvidia: loading out-of-tree module taints kernel.
[    8.576500] nvidia: module license 'NVIDIA' taints kernel.
[    8.576501] Disabling lock debugging due to kernel taint
[    8.797160] nvidia-nvlink: Nvlink Core is being initialized, major device number 506
[    8.799387] nvidia 0001:01:00.0: Adding to iommu group 4
[    8.799602] nvidia 0001:01:00.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=none
[    8.850131] NVRM: loading NVIDIA UNIX aarch64 Kernel Module  495.29.05  Thu Sep 30 16:29:07 UTC 2021
[    8.876168] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms  495.29.05  Thu Sep 30 15:49:43 UTC 2021
[    8.882428] [drm] [nvidia-drm] [GPU ID 0x00010100] Loading driver
[    8.882432] [drm] Initialized nvidia-drm 0.0.0 20160202 for 0001:01:00.0 on minor 0
[    9.075328] random: crng init done
[    9.075334] random: 7 urandom warning(s) missed due to ratelimiting
[   10.265811] NVRM: failed to register with the ACPI subsystem!
[   10.532954] NVRM: GPU 0001:01:00.0: RmInitAdapter failed! (0x24:0x54:1433)
[   10.533108] NVRM: GPU 0001:01:00.0: rm_init_adapter failed, device minor number 0
[   10.564299] NVRM: GPU 0001:01:00.0: RmInitAdapter failed! (0x24:0x54:1433)
[   10.564412] NVRM: GPU 0001:01:00.0: rm_init_adapter failed, device minor number 0
[   10.574064] NVRM: failed to unregister from the ACPI subsystem!

/var/log/Xorg.0.log shows

[    10.049] (II) Loading sub module "glxserver_nvidia"
[    10.049] (II) LoadModule: "glxserver_nvidia"
[    10.049] (II) Loading /usr/lib/xorg/modules/extensions/libglxserver_nvidia.so
[    10.072] (II) Module glxserver_nvidia: vendor="NVIDIA Corporation"
[    10.072]    compiled for 1.6.99.901, module version = 1.0.0
[    10.072]    Module class: X.Org Server Extension
[    10.072] (II) NVIDIA GLX Module  495.29.05  Thu Sep 30 15:52:00 UTC 2021
[    10.072] (II) NVIDIA: The X server supports PRIME Render Offload.
[    10.347] (EE) NVIDIA(GPU-0): Failed to initialize the NVIDIA GPU at PCI:1@1:0:0.  Please
[    10.347] (EE) NVIDIA(GPU-0):     check your system's kernel log for additional error
[    10.347] (EE) NVIDIA(GPU-0):     messages and refer to Chapter 8: Common Problems in the
[    10.347] (EE) NVIDIA(GPU-0):     README for additional information.
[    10.347] (EE) NVIDIA(GPU-0): Failed to initialize the NVIDIA graphics device!
[    10.347] (EE) NVIDIA(0): Failing initialization of X screen
[    10.347] (II) UnloadModule: "nvidia"
[    10.347] (II) UnloadSubModule: "glxserver_nvidia"
[    10.347] (II) Unloading glxserver_nvidia

Tried pci=noacpi and acpi=off and that didn’t change anything.

3

Chapter 2. Minimum Requirements seems to indicate only a limited list of CPUs are supported, but no clear explanation as to why. I also couldn’t find any of this information when I looked some 5+ months ago and ordered all the parts.

4

We are working nvidia on some issues regarding using their binary Aarch64 Linux drivers. from 2 months ago.