[Solved° System 'soft' crashing often

Trevor_B · 2025-02-09 14:34:22

After spending many years on a old PC setup running arch : Nvidia and Intel perfectly

I've just built a new PC
ASRock B650 Pro RS, AMD Ryzen 7 7700, WD Black SN850X, Kingston FURY Beast 32 Go (2 x 16 Go) DDR5 6000 MHz CL36 (KF560C36BBE2K2-32)
Installion using Archinstaller with systemd, choosing All open-sources as drivers, no swap, running the latest kernel 6.13.1-arch2-1

I'm having 'soft' crashes often, soft in the sense, the system logs me out, but does not do a full reboot

I've just dropped the ram speeds again as this seems to help, spacing out the crashes, now trying DDR4800 rather than DDR6000 but as of yet I do not have a stable system

For info I'm running the IGPU as don’t have the GPU until next week

here is the log after the last crash
https://pastebin.com/qJHRdy7g

But sadly this does not mean anything to me, I would be grateful for some heads up

Thanks

Last edited by Trevor_B (2025-03-03 00:57:16)

BS86 · 2025-02-09 14:44:13

Trevor_B wrote:

I've just dropped the ram speeds again as this seems to help, spacing out the crashes, now trying DDR4800 rather than DDR6000 but as of yet I do not have a stable system

If that has an influence on the issue but is still not stable, maybe try reseating your RAM and CPU and do a full memtest run. Either you haven't installed them properly or you have a hardware defect, most likely in RAM or some other component.

BenjB83 · 2025-02-10 07:24:34

From the pastebin it looks like the amdgpu driver is causing issues. It is only working with fairly new AMD cards and from what I remember, it has issues with integrated ones. I would try to disable / remove it and use noveau or something for now, until you get the GPU.

You could also try to use a different Kernel, like the LTS one, but I doubt, that this is really the issue.

Trevor_B · 2025-02-10 07:52:00

BenjB83 wrote:

From the pastebin it looks like the amdgpu driver is causing issues. It is only working with fairly new AMD cards and from what I remember, it has issues with integrated ones. I would try to disable / remove it and use noveau or something for now, until you get the GPU.
You could also try to use a different Kernel, like the LTS one, but I doubt, that this is really the issue.

Thanks, I'm awaiting the delivery of a card, as I believe this may be the problem as well, a week before it arrives ?

For info ran a memtest all perfect

Lone_Wolf · 2025-02-10 11:53:52

Feb 09 13:48:27 archlinux kernel: amdgpu 0000:0c:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00101030

That is a nasty bug that affects some chipsets on low level .
It happens less often noawadays, but is near impossible to troubleshoot unless you find a reliable way to cause the crash.

Check if you are running latest firmware for that motherboard AND have configured microcode loading .

Trevor_B · 2025-02-10 13:07:19

Lone_Wolf wrote:

Feb 09 13:48:27 archlinux kernel: amdgpu 0000:0c:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00101030
That is a nasty bug that affects some chipsets on low level .
It happens less often noawadays, but is near impossible to troubleshoot unless you find a reliable way to cause the crash.
Check if you are running latest firmware for that motherboard AND have configured microcode loading .

Hi and thanks

Yes I have amd-ucode installed (I ran the latest archinstaller, so it was done)

I double checked running
udo pacman -S amd-ucode
[sudo] password for trevor:
warning: amd-ucode-20250109.7673dffd-1 is up to date -- reinstalling
said N

also, I have in /etc/mkinitcpio.conf
HOOKS=(base udev autodetect microcode modconf kms keyboard keymap consolefont block filesystems fsck)

ran as well
[trevor@archlinux ~]$ lsinitcpio /boot/initramfs-linux.img | grep microcode
kernel/x86/microcode/
kernel/x86/microcode/AuthenticAMD.bin

My MB board is running the latest Bios, quite a few reports about it running DDR5 6000 ram speeds that’s why I'm running that at 4800 with a slightly lower voltage, it seems more stable

Trevor_B · 2025-02-10 13:13:00

I checked my /boot/loader/entries/ 2025-02-01_16-40-38_linux.conf

# Created by: archinstall
# Created on: 2025-02-01_16-40-38
title Arch Linux (linux)
linux /vmlinuz-linux
initrd /initramfs-linux.img
#options root=PARTUUID=f7f73d3f-1c1b-48a2-8216-777388a30e25 zswap.enabled=0 rw rootfstype=ext4
options root=PARTUUID=f7f73d3f-1c1b-48a2-8216-777388a30e25 zswap.enabled=0 rw rootfstype=ext4

On the wiki its suggests
2.1.3.2
systemd-boot
Use the initrd option to load the microcode, before the initial ramdisk, as follows:

/boot/loader/entries/entry.conf

title Arch Linux
linux /vmlinuz-linux
initrd /cpu_manufacturer-ucode.img (this I do not have in my conf file)
initrd /initramfs-linux.img
...

so not quite the same as what I have, do I need , should I modify anything ? like adding initrd /amd-ucode.img for instance ?

Thanks

Last edited by Trevor_B (2025-02-10 14:21:00)

Lone_Wolf · 2025-02-10 14:59:33

There are ways to have mkinitcpio include the microcode into the initramfs, see https://wiki.archlinux.org/title/Mkinit … mmon_hooks microcode hook.

If you use that there's no need for a separate amd-ucode.img anymore.

Please post your /etc/mkinitcpio.conf if you're unsure.

Trevor_B · 2025-02-10 15:40:47

Thanks again getting back to me

I am not advanced enough to understand "here are ways to have mkinitcpio include the microcode into the initramfs,.........."

here are my mkinitcpio.conf details
https://pastebin.com/qsV7Bdcv

also screen shot of my /boot folder if any use

xerxes_ · 2025-02-10 20:38:03

Sometimes it is not enough to have microcode installed, it have to be checked to be sure it is loaded - post output of command:

sudo journalctl -b -g microcode

Trevor_B · 2025-02-11 07:26:59

Here you go

Feb 10 15:16:19 archlinux kernel: microcode: Current revision: 0x0a601209
Feb 10 15:16:19 archlinux kernel: microcode: Updated early from: 0x0a601209
Feb 11 08:22:37 archlinux sudo[58376]: trevor : TTY=pts/0 ; PWD=/home/trevor ; USER=root ; COMMAND=/usr/bin/journalctl -b -g microcode

running [trevor@archlinux ~]$ journalctl -p 3 -xb this morning give me the usual results

Feb 10 15:16:19 archlinux kernel: hub 8-0:1.0: config failed, hub doesn't have any ports! (err -19)
Feb 10 15:16:19 archlinux kernel: amdgpu 0000:0c:00.0: [drm] *ERROR* lttpr_caps phy_repeater_cnt is 0x0, forcing it to 0x80.
Feb 10 15:16:29 archlinux lightdm[842]: gkr-pam: unable to locate daemon control file

For info since the last ram settings, the system has been stable for almost one day (touch wood)

Lone_Wolf · 2025-02-11 10:50:01

That indicates microcode is present and used .
You can move amd-ucode.img away from the /boot folder to verify if it's still needed.
(make sure to have a recent arch install iso on usb available just in case).

Trevor_B · 2025-02-11 14:06:37

Lone_Wolf wrote:

That indicates microcode is present and used .
You can move amd-ucode.img away from the /boot folder to verify if it's still needed.
(make sure to have a recent arch install iso on usb available just in case).

Thank you

I'll try that. I have a second SSD with a full system, so I can always boot into that and replace the amd-ucode.img if need be

I have just received a RX6600 XT GPU card so will first see how that goes, then I dis-activate the iGpu as I wont need that

Many thanks for all your help

I'll mark solved in the next couple of days after more testing

Arch Linux

#1 2025-02-09 14:34:22

[Solved° System 'soft' crashing often

#2 2025-02-09 14:44:13

Re: [Solved° System 'soft' crashing often

#3 2025-02-10 07:24:34

Re: [Solved° System 'soft' crashing often

#4 2025-02-10 07:52:00

Re: [Solved° System 'soft' crashing often

#5 2025-02-10 11:53:52

Re: [Solved° System 'soft' crashing often

#6 2025-02-10 13:07:19

Re: [Solved° System 'soft' crashing often

#7 2025-02-10 13:13:00

Re: [Solved° System 'soft' crashing often

#8 2025-02-10 14:59:33

Re: [Solved° System 'soft' crashing often

#9 2025-02-10 15:40:47

Re: [Solved° System 'soft' crashing often

#10 2025-02-10 20:38:03

Re: [Solved° System 'soft' crashing often

#11 2025-02-11 07:26:59

Re: [Solved° System 'soft' crashing often

#12 2025-02-11 10:50:01

Re: [Solved° System 'soft' crashing often

#13 2025-02-11 14:06:37

Re: [Solved° System 'soft' crashing often

Board footer