You are not logged in.
Ok, that's useless
We need that line after the system crash-rebooted.
But you're now using the proper kernel for that.
Offline
Ahh alright, well i had it reboot by itself after waiting a bit
Offline
May 24 19:42:53 archlinux kernel: x86/amd: Previous system reset reason [0x00200800]: ACPI power state transition occurred
https://en.wikipedia.org/wiki/ACPI#Power_states
Did we not try to limit the processor c-states?
https://wiki.archlinux.org/title/Ryzen# … nd_suspend
There was a parallel window, right?
Revisit the 3rd link below. Double, triple and quadruple-check that it's disabled. Reboot windows (and linux) thrice (seriously)
Offline
There was a parallel window, right?
I have no clue what you are referring to
Did we not try to limit the processor c-states?
We did, and for that i did
GRUB_CMDLINE_LINUX="quiet splash processor.max_cstate=1"
Not sure if that was correct
Double, triple and quadruple-check that it's disabled. Reboot windows (and linux) thrice (seriously)
100% disabled. But remember my BIOS settings reset, so if i need to change anything again, youll need to let me know, i changed too much to remember.
Offline
I have no clue what you are referring to
Sorry, typo - "windows", ie this
100% disabled
But
But remember my BIOS settings reset
3rd link below.
it's NOT the BIOS setting! That's called "fast start" and completely unrelated to the problem.
I also assume that windows doesn't exhibit this behavior (ie. it's not spuriously triggered this way by a bad firmware or broken hardware)?
Not sure if that was correct
You've to run grub-mkconfig afterwards, but "processor.max_cstate=1" is actually in the journal you posted. So yeah.
Offline
So everything is correct?
And yes, i belive fast boot by default was enabled in bios. I have NOT changed that. But in power options in control panel, i HAVE disabled hibernating AND fast boot
Offline
Unless you transiently reactive c-states after the boot (hardly true if this has affected the install iso and you'd need cpupower-idle-set or maybe some powermanagement daemon does that, though unlikely) c-states are unfortunately limited. Still something changes the power state and
I also assume that windows doesn't exhibit this behavior (ie. it's not spuriously triggered this way by a bad firmware or broken hardware)?
You could look for/try a BIOS update/reset, on an intel system I'd try to disable intel rapid resume or however the S3.4 combo is labeled, but id even k whether AMD has anything like that
Offline
I've updated the bios to the latest version, i also cant find anything "intel rapid resume" related for AMD CPUs
It did not fix the linux issue again! :'(
Offline
Add "acpi=off" to the kernel parameters. This will break a lot of things but is worth testing to see whether the resets stop.
Offline
Welp that completely broke my booting into linux and windows.
Linux was crashing on start forever, tried windows, it would start to load then reboot.
Trying to reinstall grub
Offline
This cannot "break windows" and I somewhat doubt that it would lead to "linux crashing forever" either.
With "breaking a lot of things" I rather meant the inability to shutdown the system or reboot etc.
There's also no point in re-installing grub, just remove that parameter and the bootloaders commandline editor.
But I suspect you're just witnessing degrading hardware, windows now showing the same symptoms
Offline
What are the odds ill experience this on any other OS? is there anything similar to arch that i can install? im genuinely sick of this
Offline
You can just try any live distro but as mentioned: this is very most and only increasingly likely a HW issue and then you'll face this everywhere.
Offline
Alright, well ill see. If it is HW issues, is there anything to solve it? or practically a whole new pc is the solution
Well i think it is, i honestly dont even care atp.
If in the future i want to build a pc, how tf do i prevent this? is it simply, dont get an AMD chip? cuz i am NOT getting a intel cpu
Last edited by Capzay (2025-05-25 10:32:50)
Offline
dont get an AMD chip? cuz i am NOT getting a intel cpu
That's gonna severely limit your options
Is the system self-assembled?
Strip it to bare bones (remove everything but the board, CPU and RAM, - that includes any storage devices and every connection like eg. a speaker that's not really required)
Make sure all connections are properly attached (in doubt reconnect them) and also physically inspect the system for busted capacitors, loose fits, bends/undue tension.
Ideally switch the GPU or use onboard graphics/the APU (if available) - otherwise at least make absolutely sure the GPU is in the PEG slot and the dedicated power supply is correctly attached.
Also use only one RAM DIMM if there're currently two (rep. also switch among them)
Then try to boot a live system from usb and keep a very close eye on the temperatures.
Building a PC has frankly more to do with selecting fitting components and proper assembly than intel vs. AMD (though ryzens are notorious for the PBO situation so you want to be able to control that)
Offline
Wait i have one question. How is my hardware the issue? if windows can run fine, surely its the OS's problem right? so that surely can be fixed by them and not me?
Offline
tried windows, it would start to load then reboot
I guess one has to curb their expectations with windows, but would you actually call that "fine"?
The rest is triggered by the acpi, if this isn't a hibernating OS (windows) nor the running OS (you disabled c-states) it's spurious and likely power supply related.
AMD Ryzen is on the record for crashing on PBO but idk whether that would cause this specific reset reason in the CPU.
Did you run further attempts? Do you get the same reset reason all the time?
Also
Is the system self-assembled?
Offline
Did you run further attempts? Do you get the same reset reason all the time?
Nah, i gave up. Completely removed arch.
Is the system self-assembled?
Nope, bought off of ebuyer almost 5 years ago.
How likely is it, its power supply? could it still be a driver issue?
Offline
Does windows still crash?
Do other live distros run uninterrupted?
It's certainly possible that this is some off-shoot one in a million bug in the linux kernel that only affects your system, but not likely.
Whether it's the power supply or just the general ryzen problem - or such bug - is impossible to tell w/ further tests.
If windows keeps crashing and various random live distros likewise crash, you can pretty much rule out a software issue and it's gonna be the firmware (UEFI) or the hardware.
For the hardware cause you'd have to test the components individually.
Offline
Windows doesnt crash after removing grub bootloader
endeavouros crashed on install
im trying ubuntu now
if i was to get a new PSU what would i have to look for? i dont get how it could be the issue
Ubuntu install seems to fail
Last edited by Capzay (2025-05-27 14:10:22)
Offline
If you had been booting windows from grub it probably didn't boot at all because you borked the entry?
Or did you re-install windows after removing arch/grub?
Ubuntu fails like arch, system reboots?
The most likely cause would be the known ryzen issues, the wiki points out that
out of the box, Windows seems to run the CPUs at higher voltage and lower peak frequencies, compared to the stock linux kernel
but other than adjusting PBO (apparently not possible on your system), limiting the c-states (didn't work) or *maybe* overall limiting the maximum cpu frequencies, https://wiki.archlinux.org/title/CPU_fr … requencies I wouldn't know what to do about that.
There's https://bugzilla.kernel.org/show_bug.cgi?id=212087#c43 which refers to graphics, but you don't have an APU.
I'd maybe throw a curveball and hint at https://wiki.archlinux.org/title/Power_ … Management but I doubt this would affect a live distro only running from usb…
Offline
Power supply was once mentioned at the issue. I found out that my PSU is the Corsair RM-x 2021 1000W (And calculated my system to need 700-850 watts) using SPL's PSU Tier List it states the psu is very good. So i highly doubt that could be the issue. How likely is it that it is?
Offline
You main PSU theoretical capacity really doesn't matter here - the problem is what arrives at the devices at any given point.
Offline
How can that be tested? According to the tier list, the psu should be lasting me and doing its job efficiently
Offline
Forget about the PSU, the problem is the voltage/current at the chip.
The only way is to adjust settings in the BIOS (though that's not possible on your system), disabling c-states (to avoid low power conditions, apparently didn't work) and - if it's not actually the CPU but maybe some other hardware, disable power management as much as possible.
Other than that you'd need HW equipment to measure the voltage and current arriving at the CPU core (which isn't a very realistic option)
Ideally switch the GPU or use onboard graphics/the APU (if available) - otherwise at least make absolutely sure the GPU is in the PEG slot and the dedicated power supply is correctly attached.
Also use only one RAM DIMM if there're currently two (rep. also switch among them)
This is on the theory that immediate surges in power demand on those devices starve the CPU.
Did you attempt to limit the RCU callback candidates as in https://wiki.archlinux.org/title/Ryzen# … k_freezing ?
Offline