Hi everyone,
I’m encountering persistent system freezes on my Fedora 41 installation along with a series of ACPI and NVIDIA-related issues—even though my hardware works flawlessly under Windows. I’ve performed several troubleshooting steps and would appreciate any insights or suggestions.
My System Details:
- Fedora Version: 41
- Kernel Version: 6.12.11-200.fc41.x86_64
- CPU: 12 cores (6 physical, 12 threads) – 12th Gen Intel® Core™ i5-12400F
- GPU: NVIDIA GeForce RTX 3060 Ti Lite Hash Rate (GA104)
- Kernel Driver in Use: nvidia
- Modules: nouveau, nvidia_drm, nvidia
- BIOS: Fully updated (version 3601, released 10/12/2024 by American Megatrends)
Issues & Observations:
- ACPI _DSM Failures:
My system log repeatedly shows failures evaluating ACPI Device-Specific Methods (_DSM):
ACPI: : failed to evaluate _DSM bf0212f2-788f-c64d-a5b3-1f738e285ade rev:0 func:2 (0x1001)
ACPI: : failed to evaluate _DSM bf0212f2-788f-c64d-a5b3-1f738e285ade rev:0 func:1 (0x1001)
…
ACPI: : failed to evaluate _DSM bf0212f2-788f-c64d-a5b3-1f738e285ade rev:0 func:12 (0x1001)
These errors appear to be tied to the PEG1 slot (which likely affects my NVIDIA GPU) and suggest that the firmware may be struggling with device-specific object parsing.
NVIDIA Power Management Issues:
I’m also seeing the following message from the NVIDIA power management daemon:
/usr/bin/nvidia-powerd[1026]: Found unsupported configuration. Exiting…
This indicates that the current setup might not be acceptable for the driver’s power management, potentially contributing to the freezes.
Other Driver/Peripheral Messages:
intel-lpss probe failures (error -22)
i801_smbus: "SMBus base address uninitialized, upgrade BIOS" (even though my BIOS is updated)
Minor errors from udev (EVIOCSKEYCODE) and occasional Bluetooth/gdm-password warnings
Kernel Parameter Tweaks:
I’m currently booting with these GRUB parameters:
rhgb quiet rd.driver.blacklist=nouveau modprobe.blacklist=nouveau usbcore.autosuspend=-1 pci=noacpi processor.max_cstate=1 idle=nomwait
I’ve experimented with additional tweaks (such as acpi=strict, acpi_osi=Linux, acpi=off, and nomodeset) but none of these have completely prevented the spontaneous freezes—whether idle or during active use.
Attempt to Force NVIDIA’s Performance State:
To address the GPU power management issue while preserving proper visual quality, I added the following to my GRUB command line:
nvidia-drm.modeset=1 nvidia.NVreg_RegistryDwords=‘PowerMizerEnable=0x1;PowerMizerDefaultAC=0x1’
My complete last attempt was with:
GRUB_CMDLINE_LINUX=“rhgb quiet rd.driver.blacklist=nouveau modprobe.blacklist=nouveau usbcore.autosuspend=-1 pci=noacpi processor.max_cstate=1 idle=nomwait nvidia-drm.modeset=1 nvidia.NVreg_RegistryDwords=‘PowerMizerEnable=0x1;PowerMizerDefaultAC=0x1’”
After a reboot, the system managed to boot properly and the NVIDIA modules (including DRM) loaded without fatal errors. However, I still experienced a freeze after about a week of uptime. Additionally, I received a GUI notification that “GNOME Software unexpectedly terminated” (with a corresponding crash report from ABRT) which might hint at broader issues.
that’s why I switched back to the most stable configuration for me:
GRUB_CMDLINE_LINUX=“rhgb quiet rd.driver.blacklist=nouveau modprobe.blacklist=nouveau usbcore.autosuspend=-1 pci=noacpi processor.max_cstate=1 idle=nomwait”
Why I’m Not Switching to nvidia-drm.modeset=0:
I’ve tested nvidia-drm.modeset=0 in the past; unfortunately, that option degrades display quality and resolution to an unacceptable level. I must maintain the proper visual experience while resolving these freezes.
Are there additional tweaks or alternative registry options that could help improve stability without sacrificing display quality?
I appreciate any guidance and insights as I continue troubleshooting this freezing issue. Thanks in advance for your help!