[SOLVED] nvidia: Unable to change power state from D3cold to D0

clumsyelefant · 2025-09-03 13:29:25

Hi everyone!

This is my first time posting something in this forum. I never thought I could end up begging for help for such a problem, but its really almost my last shot.

The problem is I've unsuccessfully been trying to make my NVIDIA MX250 GPU work on Arch Linux. After following the official Arch Linux installation guide almost religiously I always end up getting the same error:

Sep 03 14:51:20 banana kernel: nvidia: loading out-of-tree module taints kernel.
Sep 03 14:51:20 banana kernel: nvidia: module license 'NVIDIA' taints kernel.
Sep 03 14:51:20 banana kernel: nvidia: module verification failed: signature and/or required key missing - tainting kernel
Sep 03 14:51:20 banana kernel: nvidia: module license taints kernel.
Sep 03 14:51:20 banana kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 238
Sep 03 14:51:20 banana kernel: nvidia 0000:02:00.0: Unable to change power state from D3cold to D0, device inaccessible
Sep 03 14:51:20 banana kernel: nvidia 0000:02:00.0: probe with driver nvidia failed with error -1
Sep 03 14:51:20 banana kernel: nvidia-nvlink: Unregistered Nvlink Core, major device number 238
Sep 03 14:51:21 banana systemd-modules-load[313]: Failed to insert module 'nvidia_uvm': No such device
Sep 03 14:51:22 banana kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 237
Sep 03 14:51:22 banana kernel: nvidia 0000:02:00.0: Unable to change power state from D3cold to D0, device inaccessible
Sep 03 14:51:22 banana kernel: nvidia 0000:02:00.0: probe with driver nvidia failed with error -1
Sep 03 14:51:22 banana kernel: nvidia-nvlink: Unregistered Nvlink Core, major device number 237
Sep 03 14:51:22 banana kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 511
Sep 03 14:51:22 banana kernel: nvidia 0000:02:00.0: Unable to change power state from D3cold to D0, device inaccessible
Sep 03 14:51:22 banana kernel: nvidia 0000:02:00.0: probe with driver nvidia failed with error -1
Sep 03 14:51:22 banana kernel: nvidia-nvlink: Unregistered Nvlink Core, major device number 511

This is the laptop's information:

OS: Arch Linux x86_64
Host: HP Pavilion x360 Convertible 14-dh0xxx
Kernel: Linux 6.16.4-arch1-1
Uptime: 16 mins
Packages: 206 (pacman)
Shell: bash 5.3.3
Display (BOE082C): 1920x1080 @ 60 Hz in 14" [Built-in]
Terminal: /dev/pts/0
CPU: Intel(R) Core(TM) i7-8565U (8) @ 4.60 GHz
GPU 1: NVIDIA GeForce MX250 [Discrete]
GPU 2: Intel UHD Graphics 620 @ 1.15 GHz [Integrated]
Memory: 541.15 MiB / 15.35 GiB (3%)
Swap: 0 B / 8.00 GiB (0%)
Disk (/): 3.79 GiB / 224.81 GiB (2%) - ext4
Battery (TF03041XL): 100% [AC Connected]
Locale: en_US.UTF-8

Similarly, I’ve included outputs from diagnostic commands to help identify whether the issue is related to power management, driver loading, or hardware detection:

$ lspci -k | grep -A 3 -i "nvidia\|3d"
02:00.0 3D controller: NVIDIA Corporation GP108M [GeForce MX250] (rev a1)
        Subsystem: Hewlett-Packard Company Device 85c6
        Kernel modules: nouveau, nvidia_drm, nvidia

No results for: $ lspci -s 02:00.0 -vvv | grep -i power

$ mokutil --sb-state
SecureBoot disabled
Platform is in Setup Mode

$ pacman -Q | grep -i nvidia
linux-firmware-nvidia 20250808-1
nvidia 580.76.05-7
nvidia-utils 580.76.05-4

$ cat /proc/cmdline
initrd=\intel-ucode.img initrd=\initramfs-linux.img root=/dev/nvme0n1p3 rw

$ sudo modprobe nvidia
modprobe: ERROR: could not insert 'nvidia': No such device
$ sudo dmesg | tail -20
[    5.981573] input: HDA Intel PCH Headphone as /devices/pci0000:00/0000:00:1f.3/sound/card0/input32
[    5.981757] input: HDA Intel PCH HDMI/DP,pcm=3 as /devices/pci0000:00/0000:00:1f.3/sound/card0/input33
[    5.981990] input: HDA Intel PCH HDMI/DP,pcm=7 as /devices/pci0000:00/0000:00:1f.3/sound/card0/input34
[    5.982179] input: HDA Intel PCH HDMI/DP,pcm=8 as /devices/pci0000:00/0000:00:1f.3/sound/card0/input35
[ 1333.351296] nvidia-nvlink: Nvlink Core is being initialized, major device number 511

[ 1333.360564] nvidia 0000:02:00.0: Unable to change power state from D3cold to D0, device inaccessible
[ 1333.360731] NVRM: The NVIDIA GPU 0000:02:00.0
               NVRM: (PCI ID: 10de:1d13) installed in this system has
               NVRM: fallen off the bus and is not responding to commands.
[ 1333.360784] nvidia 0000:02:00.0: probe with driver nvidia failed with error -1
[ 1333.360823] NVRM: The NVIDIA probe routine failed for 1 device(s).
[ 1333.360824] NVRM: None of the NVIDIA devices were initialized.
[ 1333.361682] nvidia-nvlink: Unregistered Nvlink Core, major device number 511

$ sudo dmesg | grep -i "nvidia\|pci\|acpi\|power\|D3cold" | tail -50
[    3.363831] NVRM: The NVIDIA probe routine failed for 1 device(s).
[    3.363833] NVRM: None of the NVIDIA devices were initialized.
[    3.370653] nvidia-nvlink: Unregistered Nvlink Core, major device number 238
[    3.465155] i2c_hid_acpi i2c-ELAN2514:00: i2c_hid_get_input: IRQ triggered but there's no data
[    3.465282] input: ELAN2514:00 04F3:23E3 Touchscreen as /devices/pci0000:00/0000:00:15.1/i2c_designware.1/i2c-2/i2c-ELAN2514:00/0018:04F3:23E3.0002/input/input9
[    3.465403] input: ELAN2514:00 04F3:23E3 as /devices/pci0000:00/0000:00:15.1/i2c_designware.1/i2c-2/i2c-ELAN2514:00/0018:04F3:23E3.0002/input/input10
[    3.465477] input: ELAN2514:00 04F3:23E3 as /devices/pci0000:00/0000:00:15.1/i2c_designware.1/i2c-2/i2c-ELAN2514:00/0018:04F3:23E3.0002/input/input11
[    3.465549] input: ELAN2514:00 04F3:23E3 Stylus as /devices/pci0000:00/0000:00:15.1/i2c_designware.1/i2c-2/i2c-ELAN2514:00/0018:04F3:23E3.0002/input/input12
[    3.516650] input: SYNA328B:00 06CB:CD50 Mouse as /devices/pci0000:00/0000:00:15.0/i2c_designware.0/i2c-1/i2c-SYNA328B:00/0018:06CB:CD50.0003/input/input15
[    3.516789] input: SYNA328B:00 06CB:CD50 Touchpad as /devices/pci0000:00/0000:00:15.0/i2c_designware.0/i2c-1/i2c-SYNA328B:00/0018:06CB:CD50.0003/input/input16
[    3.533577] acer_wmi: Acer Laptop ACPI-WMI Extras
[    3.704726] input: ELAN2514:00 04F3:23E3 as /devices/pci0000:00/0000:00:15.1/i2c_designware.1/i2c-2/i2c-ELAN2514:00/0018:04F3:23E3.0002/input/input18
[    3.704886] input: ELAN2514:00 04F3:23E3 UNKNOWN as /devices/pci0000:00/0000:00:15.1/i2c_designware.1/i2c-2/i2c-ELAN2514:00/0018:04F3:23E3.0002/input/input19
[    3.704938] input: ELAN2514:00 04F3:23E3 UNKNOWN as /devices/pci0000:00/0000:00:15.1/i2c_designware.1/i2c-2/i2c-ELAN2514:00/0018:04F3:23E3.0002/input/input20
[    3.704987] input: ELAN2514:00 04F3:23E3 Stylus as /devices/pci0000:00/0000:00:15.1/i2c_designware.1/i2c-2/i2c-ELAN2514:00/0018:04F3:23E3.0002/input/input21
[    3.897706] input: SYNA328B:00 06CB:CD50 Mouse as /devices/pci0000:00/0000:00:15.0/i2c_designware.0/i2c-1/i2c-SYNA328B:00/0018:06CB:CD50.0003/input/input25
[    3.897920] input: SYNA328B:00 06CB:CD50 Touchpad as /devices/pci0000:00/0000:00:15.0/i2c_designware.0/i2c-1/i2c-SYNA328B:00/0018:06CB:CD50.0003/input/input26
[    4.181897] ACPI: video: Video Device [GFX0] (multi-head: yes  rom: no  post: no)
[    4.182164] ACPI: video: [Firmware Bug]: ACPI(PXSX) defines _DOD but not _DOS
[    4.182176] ACPI: video: Video Device [PXSX] (multi-head: yes  rom: yes  post: no)
[    4.559250] nvidia-nvlink: Nvlink Core is being initialized, major device number 237
[    4.568059] nvidia 0000:02:00.0: Unable to change power state from D3cold to D0, device inaccessible
[    4.568234] NVRM: The NVIDIA GPU 0000:02:00.0
               NVRM: (PCI ID: 10de:1d13) installed in this system has
[    4.568325] nvidia 0000:02:00.0: probe with driver nvidia failed with error -1
[    4.568345] NVRM: The NVIDIA probe routine failed for 1 device(s).
[    4.568347] NVRM: None of the NVIDIA devices were initialized.
[    4.569170] nvidia-nvlink: Unregistered Nvlink Core, major device number 237
[    5.253981] nvidia-nvlink: Nvlink Core is being initialized, major device number 511
[    5.263966] nvidia 0000:02:00.0: Unable to change power state from D3cold to D0, device inaccessible
[    5.264107] NVRM: The NVIDIA GPU 0000:02:00.0
               NVRM: (PCI ID: 10de:1d13) installed in this system has
[    5.264164] nvidia 0000:02:00.0: probe with driver nvidia failed with error -1
[    5.264177] NVRM: The NVIDIA probe routine failed for 1 device(s).
[    5.264177] NVRM: None of the NVIDIA devices were initialized.
[    5.264872] nvidia-nvlink: Unregistered Nvlink Core, major device number 511
[    5.308735] snd_hda_codec_realtek hdaudioC0D0: ALC295: picked fixup  for PCI SSID 103c:85c6
[    5.981368] input: HDA Intel PCH Mic as /devices/pci0000:00/0000:00:1f.3/sound/card0/input31
[    5.981573] input: HDA Intel PCH Headphone as /devices/pci0000:00/0000:00:1f.3/sound/card0/input32
[    5.981757] input: HDA Intel PCH HDMI/DP,pcm=3 as /devices/pci0000:00/0000:00:1f.3/sound/card0/input33
[    5.981990] input: HDA Intel PCH HDMI/DP,pcm=7 as /devices/pci0000:00/0000:00:1f.3/sound/card0/input34
[    5.982179] input: HDA Intel PCH HDMI/DP,pcm=8 as /devices/pci0000:00/0000:00:1f.3/sound/card0/input35
[ 1333.351296] nvidia-nvlink: Nvlink Core is being initialized, major device number 511
[ 1333.360564] nvidia 0000:02:00.0: Unable to change power state from D3cold to D0, device inaccessible
[ 1333.360731] NVRM: The NVIDIA GPU 0000:02:00.0
               NVRM: (PCI ID: 10de:1d13) installed in this system has
[ 1333.360784] nvidia 0000:02:00.0: probe with driver nvidia failed with error -1
[ 1333.360823] NVRM: The NVIDIA probe routine failed for 1 device(s).
[ 1333.360824] NVRM: None of the NVIDIA devices were initialized.
[ 1333.361682] nvidia-nvlink: Unregistered Nvlink Core, major device number 511


$ cat /etc/mkinitcpio.conf | grep MODULES
# MODULES
#     MODULES=(usbhid xhci_hcd)
MODULES=()
# 'filesystems' is _required_ unless you specify your fs modules in MODULES
##   This setup specifies all modules in the MODULES setting above.
# MODULES_DECOMPRESS
#MODULES_DECOMPRESS="no"

sudo cat /sys/bus/pci/devices/0000:02:00.0/power_state
D3cold

I would also like to mention that I unsuccessfully tried to make this graphics card work in both Ubuntu and Fedora (always following official guidelines, and getting the same error). And I already went through the following posts (however I might be missing something):
- https://bbs.archlinux.org/viewtopic.php?id=306509
- https://bbs.archlinux.org/viewtopic.php?id=286976
- https://forums.developer.nvidia.com/t/b … oot/112912
- https://www.reddit.com/r/eGPU/comments/ … tate_from/
- https://bbs.archlinux.org/viewtopic.php?id=299927
- https://bbs.archlinux.org/viewtopic.php?id=301985

Last edited by clumsyelefant (2025-09-04 13:26:30)

seth · 2025-09-03 14:36:36

Don't grep, please post your complete system journal for the boot:

sudo journalctl -b | curl -F 'file=@-' 0x0.st

The GPU is "only" a 3D accelerator, ie. supposed to be available via eg. prime-run.
Did you attempt "rcutree.gp_init_delay=1" or a later pci bus rescan?

tried to make this graphics card work in both Ubuntu and Fedora

Is there any parallel OS installed?
Is there any reason to believe the HW works *at all*?
Can you drive it w/ nouveau?

clumsyelefant · 2025-09-03 15:56:46

Hi!

Thanks for your fast reply.

Of course, here you have the complete system journal: http://0x0.st/KXMJ.txt

I see, but even if it is only a 3D accelerator, shouldn't "nvidia-smi" command work properly?
I tried (separately) both, adding "rcutree.gp_init_delay=1" and trying a later pci bus rescan using "echo 1 | sudo tee /sys/bus/pci/rescan" but none of them worked.

The answers to the other questions:
- Is there any parallel OS installed? No
- Is there any reason to believe the HW works *at all*? Yes, it works perfectly on Windows 11
- Can you drive it w/ nouveau? No, nouveau has the same problem

Thank you again for your time and effort!

seth · 2025-09-03 18:52:22

shouldn't "nvidia-smi" command work properly?

Yes.

How do we square

Is there any parallel OS installed?
No

with

Yes, it works perfectly on Windows 11

?
(Though there's no indication of a parallel OS in the journal)

Here's your problem

Sep 03 17:23:14 potato kernel: pci 0000:02:00.0: ROM [mem 0xfff80000-0xffffffff pref]: can't claim; no compatible bridge window
Sep 03 17:23:14 potato kernel: pci 0000:02:00.0: ROM [mem size 0x00080000 pref]: can't assign; no space
Sep 03 17:23:14 potato kernel: pci 0000:02:00.0: ROM [mem size 0x00080000 pref]: failed to assign
Sep 03 17:23:14 potato kernel: pci 0000:02:00.0: CLS mismatch (64 != 1020), using 64 bytes

=> https://bbs.archlinux.org/viewtopic.php … 0#p2169960

Sep 03 17:23:14 potato kernel: pci_bus 0000:00: Some PCI device resources are unassigned, try booting with pci=realloc

Didn't help there, still worth a shot - the resolution in that thread is disappointing (it suddenly worked, or maybe not) but the user also struggled to configure the kernel parameters, so god knows what happened back then.
Also check your UEFI settings on whether you can toggle the resizable BAR support.

clumsyelefant · 2025-09-04 07:36:33

Hello again,

Sorry, I should have specified more. I know the hardware works because I’ve been using Windows 11 (with no dual-boot setup) and the GPU drivers worked perfectly. The problem only happens when I install Linux; right now, the only OS installed is Arch Linux.

I will follow your error analysis and try to resolve this. Of course, if I manage to access the GPU, I’ll document the solution as clearly as possible so others with the same issue can fix it too.

Thanks again for your time and patience!

seth · 2025-09-04 07:42:04

You might also want to retry to reset the cmos and/or completely power down (remove the battery) the system - if you "shut down" (and really just hibernated) windows to replace it w/ linux, it might have left the hardware in an inconsistent state.

clumsyelefant · 2025-09-04 08:32:18

Hello again,

I followed the thread you proposed (https://bbs.archlinux.org/viewtopic.php … 0#p2169960) and I got to an interesting workaround.

As you initially proposed I unsuccessfully rescanned the pci bus using

echo 1 | sudo tee /sys/bus/pci/rescan

. However, in the other thread you also recommend to remove the device before the rescan using

echo 1 | sudo tee /sys/bus/pci/devices/0000\:02\:00.0/remove

and it worked!

The commands I used are:

$ echo 1 | sudo tee /sys/bus/pci/devices/0000\:02\:00.0/remove
$ echo 1 | sudo tee /sys/bus/pci/rescan

Since I am new to this """""low level""""" linux configurations I should have known that its recommended to remove the specific device before rescanning the pci bus.

I created a simple systemd service to perform this rescan during the boot and works perfectly now. Nevertheless, the other solution you propose might also work:

seth wrote:

You might also want to retry to reset the cmos and/or completely power down (remove the battery) the system - if you "shut down" (and really just hibernated) windows to replace it w/ linux, it might have left the hardware in an inconsistent state.

Should I now tag this post as SOLVED?

You really cannot imagine how thankful I am, you really are a lifesaver.

Thanks!

seth · 2025-09-04 12:54:14

I'd check whether you can avoid the late rescan and get away w/ the realloc or disabling rebar, but if the rescan reliably gets you around the allocation issues and you're fine w/ that then yes, pleasemark resolved threads by editing your initial posts subject - so others will know that there's no task left, but maybe a solution to find.
Thanks.

Arch Linux

#1 2025-09-03 13:29:25

[SOLVED] nvidia: Unable to change power state from D3cold to D0

#2 2025-09-03 14:36:36

Re: [SOLVED] nvidia: Unable to change power state from D3cold to D0

#3 2025-09-03 15:56:46

Re: [SOLVED] nvidia: Unable to change power state from D3cold to D0

#4 2025-09-03 18:52:22

Re: [SOLVED] nvidia: Unable to change power state from D3cold to D0

#5 2025-09-04 07:36:33

Re: [SOLVED] nvidia: Unable to change power state from D3cold to D0

#6 2025-09-04 07:42:04

Re: [SOLVED] nvidia: Unable to change power state from D3cold to D0

#7 2025-09-04 08:32:18

Re: [SOLVED] nvidia: Unable to change power state from D3cold to D0

#8 2025-09-04 12:54:14

Re: [SOLVED] nvidia: Unable to change power state from D3cold to D0

Board footer