Kernel update issues

MikeK61

Im new to Solus but love it already. My name is Mike from germany. Using Solus 4.3 since last june. Iḿ wondering why i always have issues after an kernel update.
Everything seems to work fine but i get unexpected freeze or reboots with error message about prcessor 2 after reboot.
After a few days the problems seems to be gone.
Same happened last night after i updated to kernel 6.0.10.

Any ideas ?

stalebrim

MikeK61 What does the error message say?

MikeK61

Thats a very long text in dos mode during boot i cant remember.
I have to write that down the next time it occurs.

ATM it is working fine again, have no issues since 7 hours now.
I would just like to understand why it always happens after a kernel update.

MikeK61

Just found this in journalctl.

Nov 30 18:48:13 solus kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Nov 30 18:48:13 solus kernel: ACPI: _PR.C000: Found 2 idle states
Nov 30 18:48:13 solus kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Nov 30 18:48:13 solus kernel: ACPI: _PR.C002: Found 2 idle states
Nov 30 18:48:13 solus kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Nov 30 18:48:13 solus kernel: ACPI: _PR.C004: Found 2 idle states
Nov 30 18:48:13 solus kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Nov 30 18:48:13 solus kernel: ACPI: _PR.C006: Found 2 idle states
Nov 30 18:48:13 solus kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Nov 30 18:48:13 solus kernel: ACPI: _PR.C008: Found 2 idle states
Nov 30 18:48:13 solus kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Nov 30 18:48:13 solus kernel: ACPI: _PR.C00A: Found 2 idle states
Nov 30 18:48:13 solus kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Nov 30 18:48:13 solus kernel: ACPI: _PR.C00C: Found 2 idle states
Nov 30 18:48:13 solus kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Nov 30 18:48:13 solus kernel: ACPI: _PR.C00E: Found 2 idle states
Nov 30 18:48:13 solus kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Nov 30 18:48:13 solus kernel: ACPI: _PR.C001: Found 2 idle states
Nov 30 18:48:13 solus kernel: [Firmware Bug]: ACPI MWAIT C-state 0x0 not supported by HW (0x0)
Nov 30 18:48:13 solus kernel: ACPI: _PR.C003: Found 2 idle states
Nov 30 18:48:13 solus kernel: mce: [Hardware Error]: CPU 13: Machine Check: 0 Bank 5: bea0000000000108
Nov 30 18:48:13 solus kernel: mce: [Hardware Error]: TSC 0 ADDR 1ffff8105f930 MISC d012000100000000 SYND 4>
Nov 30 18:48:13 solus kernel: mce: [Hardware Error]: PROCESSOR 2:800f11 TIME 1669830492 SOCKET 0 APIC b mi>
Nov 30

ReillyBrogan

Have you tried updating your motherbord firmware?

MikeK61

Do you mean Bios ?
Bios is already the latest version.

MikeK61

I bootet in kernel 5.15.77 and made some stress tests.
After 12 hours no issues.

MB Crosshair 6 Hero Bios 8601 (august22)
Ryzen 1800X
GPU Sapphire 6600XT
16GB Gskill
Samsung 960 Evo.

ReillyBrogan

So it looks like mce hardware error indicates some kind of CPU issue potentially. You may have a hardware issue of some kind. I believe newer kernels added support for more diagnostics around this which might be why you're only seeing there error on 6.0

MikeK61

I never had an issue before.
My system is running 24/7 for more than 5 years on win and Linux doing distributed computing.
Even after 6 hours of gaming on win 10 nothing unpredicted happens.

Is it possible to avoid kernel updates ?
I unchecked kernels in software center but they install anyways.

stalebrim

MikeK61

Is it possible to avoid kernel updates ?

ReillyBrogan

Just use the LTS kernel. It receives just as many updates as the current kernel but they're all bugfixes/security fixes and not new features.

MikeK61

Thanks for the replies.
Thats what im doing for now.
I have read that it is possible to include a working ACPI table manually to the kernel.
This fix was for Ubuntu, is this possible on Solus also ?
I have noticed my CPU is downclocking to much in idle state.
It should be 2200Mhz but it is around 1880 so i assume the vcore is to low as well.

MikeK61

Now kernel 6.0.11 is available as an update.
If i install it 5.15.77 will go.

Can i save it somehow ?
Or can i manually reinstall it afterwards.

ReillyBrogan

MikeK61 You can install linux-lts which will remain on the 5.15 branch for quite some time.

MikeK61

Thank you very much ReillyBrogan.
This is good to know as the last chance.

MikeK61

I can report that i was affected by the idle bug.
Just dont know why it didnt happen with older kernels.
After reading dozens of forum threads i deactivated c-state in the bios.

System:
Host: solus Kernel: 6.0.10-224.current arch: x86_64 bits: 64 compiler: gcc
v: 12.2.0 Desktop: KDE Plasma v: 5.26.4 Distro: Solus 4.3 fortitude
Info:
Processes: 345 Uptime: 22h 27m Memory: 15.55 GiB used: 3.52 GiB (22.7%)
Init: systemd target: graphical (5) Compilers: gcc: N/A Packages: 1127
Shell: Bash v: 5.1.16 inxi: 3.3.23

Looks fine now.
Thank you for pointing me to the right direction.