Running corefreqd causes kernel bug in Enable_ACPI_CPPC #531
Comments
|
@archiecarrot123 Hello, Can you try the branch hotfix_optimizations which attempts to fix some |
|
Using hotfix_optimizations, I get a kernel bug in CoreFreqK_mmap instead. I did try using the master branch but I didn't actually read the call trace. This is the call trace: |
|
Thank you for you return. So there is no way to check if the changes of the branch are helpful or not ? Did it happen that you ran CoreFreq successfully with Gentoo ? I'm also reviewing Perhaps your kernel is not compatible. Also it could be drivers incompatibilities; especially those accessing the SMU. Here's my boot command line. However I don't face a crash when command line is barely empty. Looking at this trace I see an X570 chipset, the same as mine (ASUS Crosshair VIII); my processor is a |
|
Meanwhile looking at both of your logs which leads to the function Do you have EDIT: Also let me known about adding |
|
I just checked and Also, my processor is a It does seem very strange that it crashes on asserting that the VMA is write protected. |
Some times a go Just to verify about a Gentoo or Kernel issue; do you mind to boot the CoreFreq Live CD ? |
|
I tried booting the live CD this morning, and CoreFreq worked fine, so this is presumably an issue with my kernel. CoreFreq worked previously, but at some point stopped working. I noticed this bug first with an older version of CoreFreq installed from an ebuild, then made the report when it persisted for a CoreFreq 2.0.0 ebuild. I am going to test setting |
Thank you for your help.
EDIT: Meanwhile I'm also trying make menuconfig ## to disable CONFIG_ANON_VMA_NAME
make -j32 pacman-pkg
pacman -U linux-upstream-6.14.0_rc4+-66-x86_64.pkg.tar.zst linux-upstream-headers-6.14.0_rc4+-66-x86_64.pkg.tar.zst
uname -a
Linux RYZEN 6.14.0-rc4+ #66 SMP PREEMPT_DYNAMIC Wed Feb 26 10:41:00 CET 2025 x86_64 GNU/Linux
zgrep ANON_VMA /proc/config.gz
# CONFIG_ANON_VMA_NAME is not set
cd src/CoreFreq
make clean
make -jinsmod build/corefreqk.ko Register_ClockSource=1 Register_Governor=1 Register_CPU_Idle=1 Register_CPU_Freq=1 Override_SubCstate="1,1,1,1,1,1,0,0" C3U_Enable=1 C2U_Enable=1 ServiceProcessor=12; \
echo "corefreq_tsc" > /sys/devices/system/clocksource/clocksource0/current_clocksource; \
./build/corefreqd -d; \
rmmod corefreqk
|
|
Setting Any other ideas, apart from using your (or some other known good) config and performing a bisection? |
No other idea yet. I had tried to build against kernel |
|
To reproduce the bug, is there a Gentoo live with similar environment? (kernel, compiler) |
Hello, |
|
Running corefreqd (CoreFreq version 2.0.0) as root causes a kernel bug and hangs.
This is the trace:
The text was updated successfully, but these errors were encountered: