Skip to content

[BUG] nvidia GPU stuck at d0 state on 7.3.0-rc3-1 - Runtime D3 status: Not supported #1053

Closed
#1096
CachyOS/kernel-patches
#191
Closed
[BUG] nvidia GPU stuck at d0 state on 7.3.0-rc3-1 - Runtime D3 status: Not supported#1053
#1096
CachyOS/kernel-patches
#191
Assignees
ptr1337ventureoo1Naim
Labels
bugSomething isn't working

Description

@xantrk
xantrk
opened on Sep 17, 2026

Pre-flight checklist

  • I have searched existing issues and this is not a duplicate.
  • I have read the Contributing Guidelines.
  • I have verified the issue is reproducible with the latest available CachyOS kernel.
  • I have tried to reproduce the issue on Arch Linux's linux kernel.

Upstream / vanilla kernel check

I tested with a vanilla/upstream kernel and the issue does NOT reproduce there (CachyOS-specific bug)

Kernel variant

linux-cachyos-rc

System information (cachyos-bugreport.sh)

https://paste.cachyos.org/p/495c4ec.log

Manual system information (if cachyos-bugreport.sh is unavailable)

Bug description

On hybrid laptop with dGPU and iGPU, dGPU is stuck in D0 and constantly draws power while lsof is empty.

cat /proc/driver/nvidia/gpus/0000:02:00.0/power
Runtime D3 status:          Not supported

on RC kernel while supported on non-rc.

7.2.6-1-cachyos kernel returns:

cat /proc/driver/nvidia/gpus/0000:02:00.0/power
Runtime D3 status:          Enabled (fine-grained)

Steps to reproduce

  1. cat /proc/driver/nvidia/gpus/0000:02:00.0/power
  2. Runtime D3 status: Not supported (while it works on non rc kernel)

Expected behavior

D3 status should be supported.

Actual behavior

D3 status not supported.

Logs / stack traces

Additional system information

No response

Activity

  1. xantrk
    assigned
    1Naim
    and
    ptr1337
    on Sep 17, 2026
  2. xantrk
    added
    bugSomething isn't working
    on Sep 17, 2026

  3. good-news-everyone commented on Oct 22, 2026

    @good-news-everyone
    good-news-everyone
    on Oct 22, 2026

    Same here on the stable 7.3 packages, not only rc.

    Razer Blade 14 (RZ09-0508), Ryzen 9 8945HS + Radeon 780M, RTX 4070 Laptop GPU (AD106M, 10de:2820), hybrid mode, open module 615.78.08 from linux-cachyos-nvidia-open.

    /proc/driver/nvidia/gpus/0000:01:00.0/power, same driver version and module options on every kernel:

    kernel NVreg_DynamicPowerManagement Runtime D3 status 01:00.0 after boot
    7.3.0-2-cachyos 0x03 (default) Disabled by default D0, runtime_suspended_time = 0
    7.3.0-2-cachyos 0x02 Not supported D0, runtime_suspended_time = 0
    7.3.0-3-cachyos 0x02 Not supported D0, runtime_suspended_time = 0
    7.2.10-1-cachyos 0x02 Enabled (fine-grained) D3cold
    6.18.55-2-cachyos-lts 0x02 Enabled (fine-grained) D3cold

    power/control is auto on the GPU, its HDA function and the root port, and the port has _PR3, so nothing on the kernel PM side is blocking it. The driver turns RTD3 off by itself and logs nothing about it.

    This fits the explanation in CachyOS/kernel-patches#191: since 7.3 kthreads start in nullfs (32750c77e811), so filp_open() on /sys/bus/pci/devices/<id>/config in nv_get_pci_sysfs_config() fails in the nv-probe kthread. I haven't tested the patch itself.

    About the vanilla check in the report: that change is upstream, so stock 7.3 with the 615 module loaded should behave the same. A stock Arch kernel without an nvidia module for it (no nvidia-open-dkms) leaves the GPU unbound, and with nouveau blacklisted by nvidia-utils the kernel suspends it on its own. I hit exactly that on linux 7.2.11-arch1-1, so the vanilla result may not exercise the driver at all.

    For now I'm back on 7.2.10-1 until #191 lands.

  4. Acters commented on Oct 22, 2026

    @Acters
    Acters
    on Oct 22, 2026

    Since this is patching NVIDIA open gpu kernel module, can you report this on their github issue tracker too?

  5. good-news-everyone commented on Oct 22, 2026

    @good-news-everyone
    good-news-everyone
    on Oct 22, 2026

    I'll try, this is quite a challenge to create issue in their repo 🫠

  6. ventureoo
    closed this as completedin kernel-patches#191on Oct 22, 2026

  7. xantrk commented on Oct 23, 2026

    @xantrk
    xantrk
    on Oct 23, 2026
    Author

    I think the problem is sadly still happening with a different flavour.

    cat /proc/driver/nvidia/gpus/0000:02:00.0/power
    Runtime D3 status: Disabled by default
    Tegra iGPU Rail-Gating: Disabled
    Video Memory: Active

    GPU Hardware Support:
    Video Memory Self Refresh: Supported
    Video Memory Off: Supported

    S0ix Power Management:
    Platform Support: Not Supported
    Status: Disabled

    nvidia-smi
Fri Oct 23 17:28:15 2026
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 615.78.08              KMD Version: 615.78.08     CUDA UMD Version: 13.4     |
+-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 5070 ...    On  |   00000000:02:00.0 Off |                  N/A |
| N/A   45C    P8              5W /   90W |      15MiB /  12227MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI              PID   Type   Process name                        GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|    0   N/A  N/A            2683      G   niri                                      3MiB |
+-----------------------------------------------------------------------------------------+

    cat /sys/bus/pci/devices/0000:02:00.0/power/control
    cat /sys/bus/pci/devices/0000:02:00.0/power/runtime_status
    cat /sys/bus/pci/devices/0000:02:00.0/power/runtime_suspended_time
    auto
    active
    0

     furkan  cat /etc/modprobe.d/nvidia-power.conf
    options nvidia NVreg_DynamicPowerManagement=0x03
    options nvidia NVreg_DynamicPowerManagementVideoMemoryThreshold=200
    options nvidia "NVreg_PreserveVideoMemoryAllocations=1"
     furkan  cat /sys/module/nvidia/parameters/NVreg_DynamicPowerManagement
    cat /proc/driver/nvidia/gpus/0000:02:00.0/power
    journalctl -b -k | grep -Ei 'NVRM|nvidia.*(power|D3|ACPI)'
    3
    Runtime D3 status: Disabled by default
    Tegra iGPU Rail-Gating: Disabled
    Video Memory: Active

    GPU Hardware Support:
    Video Memory Self Refresh: Supported
    Video Memory Off: Supported

    S0ix Power Management:
    Platform Support: Not Supported
    Status: Disabled

    Notebook Dynamic Boost: Supported
    Oct 23 18:24:07 cachyos kernel: NVRM: GPU0 nvAssertOkFailedNoLog: Assertion failed: Invalid data passed [NV_ERR_INVALID_DATA] (0x00000025) returned from PlatformRequestHandler failed to get target temp from SBIOS @ platform_request_handler_ctrl.c:2175
    Oct 23 18:24:07 cachyos kernel: NVRM: GPU0 nvAssertOkFailedNoLog: Assertion failed: Invalid data passed [NV_ERR_INVALID_DATA] (0x00000025) returned from PlatformRequestHandler failed to get platform power mode from SBIOS @ platform_request_handler_ctrl.c:2118
    Oct 23 18:24:07 cachyos kernel: NVRM: loading NVIDIA UNIX Open Kernel Module for x86_64 615.78.08 Release Build (notroot@) Fri Oct 23 15:27:21 UTC 2026

  8. eltunas420 commented on Oct 23, 2026

    @eltunas420
    eltunas420
    on Oct 23, 2026

    Running cat /proc/driver/nvidia/gpus/0000:01:00.0/power throws the following for me:

    Runtime D3 status: Not supported
    Tegra iGPU Rail-Gating: Disabled
    Video Memory: Active

    GPU Hardware Support:
    Video Memory Self Refresh: Supported
    Video Memory Off: Supported

    S0ix Power Management:
    Platform Support: Supported
    Status: Enabled

    Notebook Dynamic Boost: Supported

  9. ventureoo
    reopened this on Oct 23, 2026

  10. ventureoo commented on Oct 23, 2026

    @ventureoo
    ventureoo
    on Oct 23, 2026
    Member

    Seems to be issue still persist, was able to reproduce it on my laptop.

    With options nvidia NVreg_DynamicPowerManagement=0x02

    [N] ~ $ cat /proc/driver/nvidia/gpus/0000:01:00.0/power
Runtime D3 status:          Not supported
Tegra iGPU Rail-Gating:     Disabled
Video Memory:               Active

GPU Hardware Support:
 Video Memory Self Refresh: Supported
 Video Memory Off:          Supported

S0ix Power Management:
 Platform Support:          Supported
 Status:                    Disabled

Notebook Dynamic Boost:     Supported

    Without options nvidia NVreg_DynamicPowerManagement=0x02

     Runtime D3 status: Disabled by default
Tegra iGPU Rail-Gating: Disabled
Video Memory: Active

GPU Hardware Support:
Video Memory Self Refresh: Supported
Video Memory Off: Supported

S0ix Power Management:
Platform Support: Not Supported
Status: Disabled

Notebook Dynamic Boost: Supported
  11. ventureoo
    self-assigned this
    on Oct 23, 2026

  12. ventureoo commented on Oct 23, 2026

    @ventureoo
    ventureoo
    on Oct 23, 2026
    Member

    Nevermind, seems to be I just got confused, because we still had pkgrel=3 in PKGBUILD, but -4 in repo actually doesn't contain fix for this issue, because it was pushed for another tag release (https://github.com/CachyOS/linux/releases/tag/cachyos-7.3.0-4), not for NVIDIA D3 fix. Please wait for linux-cachyos-nvidia-open 7.3.0-5.

  13. ventureoo
    closed this as completedon Oct 23, 2026

  14. eltunas420 commented on Oct 23, 2026

    @eltunas420
    eltunas420
    on Oct 23, 2026

    I've updated linux-cachyos-nvidia-open to 7.3.0-5 (which is the most recent version), and Runtime D3 Status now reports as "Disabled by Default"

  15. ventureoo commented on Oct 23, 2026

    @ventureoo
    ventureoo
    on Oct 23, 2026
    Member

    16. I've updated linux-cachyos-nvidia-open to 7.3.0-5 (which is the most recent version), and Runtime D3 Status now reports as "Disabled by Default"

    Are you using znver4 repository? Because packages in znver4 always have one more pkgrel value than in v3/v4.

  16. eltunas420 commented on Oct 23, 2026

    @eltunas420
    eltunas420
    on Oct 23, 2026

    17. Are you using znver4 repository?

    Yes, i do. Does this mean i should wait for 7.3.0-6?

  17. ventureoo commented on Oct 23, 2026

    @ventureoo
    ventureoo
    on Oct 23, 2026
    Member

    18. > 17. Are you using znver4 repository?

         Yes, i do. Does this mean i should wait for 7.3.0-6?

    Yes, you should wait for it.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

bugSomething isn't working

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions