RTX 3060 "RmInitAdapter failed! (0x24:0x72:1743)" + ACPI AE_ALREADY_EXISTS on Oracle Linux 9

Graphics / Linux Linux
1

System Specifications:

  • GPU: NVIDIA GeForce RTX 3060 12GB (GA106, dual BIOS switch)

  • OS: Oracle Linux 9.7 (UEK kernel: 6.12.0-200.74.27.1.el9uek.x86_64 and 5.14.0-611.45.1.el9_7.x86_64)

  • Driver: NVIDIA 595.58.03 (installed via runfile with DKMS)

  • Motherboard: ASUS [insert model if known]

  • CPU: Intel 12th Gen Alder Lake-S

Issue Summary:
After my RTX 3060 returned from repair, the GPU is not initializing in Linux. The system detects the GPU (lspci shows NVIDIA GA106 [GeForce RTX 3060]), but nvidia-smi returns “No devices found”. The kernel logs show:

  1. Firmware initialization failure:

    text

    Copy

    NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x24:0x72:1743)

  2. ACPI conflicts:

    text

    Copy

    ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS

  3. Missing Video BIOS:

    text

    Copy

    Video BIOS: ??.??.??.??.??

Troubleshooting Steps Attempted:

  1. Driver Reinstallation:

    • Removed all NVIDIA packages (dnf remove nvidia* + dkms remove).

    • Reinstalled driver 595.58.03 with DKMS and dependencies (libglvnd-devel, pkg-config).

    • Tried both silent and interactive installation.

  2. Kernel Testing:

    • Tested on multiple kernels (UEK 6.12 and 5.14), but the issue persists.

  3. BIOS Settings:

    • Enabled Above 4G Decoding and Resizable BAR.

    • Set PCIe to Gen 3 (not Auto/Gen 4).

    • Disabled CSM (UEFI mode only).

  4. Hardware Checks:

    • Reseated the GPU and tested in the primary PCIe slot.

    • Performed cold power cycles (unplugged PSU for 5+ minutes).

  5. Firmware Workarounds:

    • Added kernel parameters:

      text

      Copy

      nvidia-drm.modeset=1 nvidia.NVreg_EnableGpuFirmware=0

    • No change in behavior.

Additional Observations:

  • The GPU’s Video BIOS is not detected (cat /proc/driver/nvidia/gpus/*/information).

  • The ACPI error (AE_ALREADY_EXISTS) suggests a conflict in the GPU’s power management tables.

  • The GSP firmware (used by newer drivers) may be failing to initialize due to corrupted state or BIOS incompatibility.

Questions for the Community:

  1. Has anyone encountered RmInitAdapter failed! (0x24:0x72:1743) on an RTX 3060 with driver 595.xx?

    • If so, were you able to resolve it? How?

  2. Could the ACPI AE_ALREADY_EXISTS error be preventing the GPU from initializing?

    • Is this a known issue with ASUS motherboards or Oracle Linux 9?

  3. Is there a way to force-reload the GPU’s Video BIOS or reset its firmware state?

    • (I’ve already tried cold power cycles and reseating.)

  4. Should I try downgrading the driver to a pre-GSP version (e.g., 535.xx)?

    • If yes, which version is most stable for the RTX 3060 on Oracle Linux 9?

  5. Are there additional kernel parameters or BIOS settings I should try?

Logs for Reference:

bash

Copy

# lspci (GPU is detected)
01:00.0 VGA compatible controller: NVIDIA Corporation GA106 [GeForce RTX 3060] (rev a1)

# dmesg (errors)
[    4.786540] ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS
[    4.841003] NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x24:0x72:1743)

# nvidia-smi
No devices were found

# /proc/driver/nvidia/gpus/*/information
Video BIOS:      ??.??.??.??.??

Final Notes:

  • I suspect this is either a firmware state issue (from improper shutdowns before repair) or a BIOS/ACPI incompatibility.

  • Any guidance or similar experiences would be greatly appreciated!

2

I am seeing the same error message with an RTX A4500. I’ve tested 535, 560, and 595 driver versions on 6.2, 6.8, and 6.17 kernels.

The GPU itself is accessible with the NVFlash utility.

Build GUID : 849F47DED3A842E596028E2AEEEEDA26
Build Number : 30525064
IFR Subsystem ID : 10DE-163C
Subsystem Vendor ID : 0x10DE
Subsystem ID : 0x163C
Version : 94.02.88.00.01
Image Hash : N/A
Hierarchy ID : Normal Board
Build Date : 10/12/21
Modification Date : 10/12/21
UEFI Version : 0x60010 ( x64 )
UEFI Variant ID : 0x000000000000000A ( GA1xx )
UEFI Signer(s) : Microsoft Corporation UEFI CA 2011
XUSB-FW Version ID : N/A
XUSB-FW Build Time : N/A
InfoROM Version : G132.0510.00.01
InfoROM Backup : Present
License Placeholder : Present
GPU Mode : Graphics
CEC OTA-signed Blob : Not Present

The dmesg log:

[ 13.408465] nvidia: loading out-of-tree module taints kernel.
[ 13.408476] nvidia: module license ‘NVIDIA’ taints kernel.
[ 13.408478] Disabling lock debugging due to kernel taint
[ 13.408482] nvidia: module verification failed: signature and/or required key missing - tainting kernel
[ 13.408484] nvidia: module license taints kernel.
[ 13.515086] nvidia-nvlink: Nvlink Core is being initialized, major device number 237

[ 13.523410] nvidia 0000:29:00.0: vgaarb: VGA decodes changed: olddecodes=io+mem,decodes=none:owns=io+mem
[ 13.569293] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 595.58.03 Tue Mar 17 19:54:55 UTC 2026
[ 13.676907] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 595.58.03 Tue Mar 17 19:37:50 UTC 2026
[ 13.680490] [drm] [nvidia-drm] [GPU ID 0x00002900] Loading driver
[ 14.135556] NVRM: GPU 0000:29:00.0: RmInitAdapter failed! (0x24:0x72:1743)
[ 14.135564] NVRM: GPU 0000:29:00.0: rm_init_adapter failed, device minor number 0
[ 14.135693] [drm:nv_drm_dev_load [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00002900] Failed to allocate NvKmsKapiDevice
[ 14.440412] NVRM: GPU 0000:29:00.0: RmInitAdapter failed! (0x24:0x72:1743)
[ 14.440420] NVRM: GPU 0000:29:00.0: rm_init_adapter failed, device minor number 0