You are not logged in.
EDIT: OP is somehow using a laptop chip on a Desktop?
there's nothing that would stop one to implant a mobile cpu/soc on a standard atx form factor
the difference is that "mobile platform" is targetted to run on battery and hence is optimized for power efficiency
re-using and hacking platforms to do what they're not desinged for is as old as 386dx or maybe even older
look up videos of those taiwan or china tech cities and thier huge warehouse-shopping malls: you'd be surprised by what weird systems exist we don't know of "in the west"
Offline
there's nothing that would stop one to implant a mobile cpu/soc on a standard atx form factor
the difference is that "mobile platform" is targetted to run on battery and hence is optimized for power efficiency
re-using and hacking platforms to do what they're not desinged for is as old as 386dx or maybe even older
look up videos of those taiwan or china tech cities and thier huge warehouse-shopping malls: you'd be surprised by what weird systems exist we don't know of "in the west"
I know that these exist. Have known for a long while. Just didn't expect that a Zen 4 chip would be that readily available for that
Offline
I know that these exist. Have known for a long while. Just didn't expect that a Zen 4 chip would be that readily available for that
Yes, these exist. You can even get them in a 16core 7945HX variant and from Amazon.
A couple months back it was just much cheaper than getting a 7700+ITX mainboard.
As for the main problem, I downgraded to the 6.10.2-arch1-2 kernel and it seems to be running stable so far. So the issue seems to be somwhere in the 6.10.3 kernel.
Offline
Downgraded to 6.10.2-zen1-2-zen too. Stable so far, but can't be 100% sure, yet.
Offline
fedro wrote:after upgrading to 6.10.3-zen1-2-zen, random reboots happen, apparently for no reason and nothing on journalctl seems to be related to the sudden reboots.
What was the pre upgrade kernel that does not have the issue?
[2024-08-08T10:07:49+0200] [ALPM] upgraded linux-zen (6.10.2.zen1-1 -> 6.10.3.zen1-2)
[2024-08-08T10:07:51+0200] [ALPM] upgraded linux-zen-headers (6.10.2.zen1-1 -> 6.10.3.zen1-2)
so 6.10.2.zen1-1 seems to be the latest stable version (at least on my machine)
Offline
99% of spontaneous reboots are due to a hardware error, especially on Ryzen, and the kernel has nothing to do with it.
A hardware error does not mean that the CPU is broken and must be replaced, just that the voltages must be adjusted manually.
Excuse my poor English.
Offline
99% of spontaneous reboots are due to a hardware error, especially on Ryzen, and the kernel has nothing to do with it.
A hardware error does not mean that the CPU is broken and must be replaced, just that the voltages must be adjusted manually.
We now have 3 people in this thread and 1 in a different thread experiencing this issue with the new 6.10.3 kernel... sorry, but that does not sound like a hardware error to me.
Offline
This is a hardware issue, undervolt, overheat, bad cpu or ram. Kernel panics don't result in spontanous reboots.
Since a downgrade can however mitigate it (though there seems disagreement about the offending version), the question is what kernel change raised the likelihood of provoking it itfp.
Edit: the other one percent would be cosmic rays, secret weapon tests or the gods are hating you.
Last edited by seth (2024-08-09 08:22:12)
Offline
Since a downgrade can however mitigate it (though there seems disagreement about the offending version), the question is what kernel change raised the likelihood of provoking it itfp.
fedro went from 6.10.2.zen1-1 to 6.10.3.zen1-2, so he skipped 6.10.2-zen1-2 and 6.10.3-zen1-1 while upgrading.
I can confirm that the issue initially appeared for me on 6.10.3-arch1-1 and persisted when i upgraded to 6.10.3-arch1-2 only a couple of hours later. They disappeared when i downgraded to 6.10.2-arch1-2. So not really a disagreement here.
Offline
just out of curiosity: I've seen zen kernel on several threads about these issues - as I use regular std kernel and don't encounter them: how about re-test with std instead of zen?
could be that changes in zen are a cause
Offline
just out of curiosity: I've seen zen kernel on several threads about these issues - as I use regular std kernel and don't encounter them: how about re-test with std instead of zen?
could be that changes in zen are a cause
I don't use zen kernel and have this issue, so it does not seem like it is caused by zen additions.
Offline
I'm experiencing the same issue while using the standard kernel.
Offline
I can confirm this started after the update to Kernel 6.10.3-arch1-2. No zen additions. Previous kernel was linux-6.9.9.arch1-1 The crashes are random and sometimes happen just from idling, but no pattern. The remind me of an overheating gpu, but that isn't the problem. I've been running this computer, GEEKOM A7, since March with absolutely no issues before this update. It dual boots into Win 11 with no problem.
.o+` -----------
`ooo/ OS: Arch Linux x86_64
`+oooo: Host: A7
`+oooooo: Kernel: 6.10.3-arch1-2
-+oooooo+: Uptime: 4 mins
`/:-:++oooo+: Packages: 1101 (pacman), 6 (flatpak)
`/++++/+++++++: Shell: bash 5.2.32
`/++++++++++++++: Resolution: 3840x2160
`/+++ooooooooooooo/` DE: GNOME 46.4
./ooosssso++osssssso+` WM: Mutter
.oossssso-````/ossssss+` WM Theme: Adwaita
-osssssso. :ssssssso. Theme: Adwaita-dark [GTK2/3]
:osssssss/ osssso+++. Icons: Adwaita [GTK2/3]
/ossssssss/ +ssssooo/- Terminal: gnome-terminal
`/ossssso+/:- -:/+osssso+- CPU: AMD Ryzen 9 7940HS w/ Radeon 780M Graphics (16) @ 5.263GHz
`+sso+:-` `.-/+oso: GPU: AMD ATI 65:00.0 Phoenix1
`++:. `-/+/ Memory: 7981MiB / 31377MiB
Last edited by dobie2564 (2024-08-09 12:28:22)
Offline
quick update: i downgraded to 6.10.2-zen1-1-zen (which was the latest stable version i could use without issues) and my pc is running smoothly since 6+ hrs (while average uptime was 2hrs at best)
i doubt it's a matter of faulty hardware since i've never had this issue before and i think i'll stick to this kernel release for now...
Offline
Repeat after me: it **very most likely** /is/ faulty hardware.
The kernel update will do something somewhere to expose that, but you do not get spontanous reboots from kernel bugs.
I guess loqs is already compiling bisection kernels but even iff and when we identify the breaking commit, be prepared to here from upstream that this is not gonna be reverted because your system is undervolted - so start looking into hardware related mitigations *now*, or you might find yourself in a position saying "I should have looked into hardware related mitigations while I still had a chance"
Offline
exact same issue, with lenovo ideapad 15AMN8 Ryzen 3 series 7000 on main kernel and lts (seems less common on lts tho). I first noticed it today, and tried to probe all hardware in order to find a faulty unit. Can you confirm it is kernel related?
It indeed happened while using an external screen (but i used that screen almost everyday from april till june without any issue like that). Never under/over volted anything. I bought this laptop on december 23 and immediately removed windows to install gnulinux.
Offline
As mentioned in your thread and by all accounts only the latest kernels are affected (but all of them, incl the LTS one)
The "undervolting" is more like an optimisitic calibration of HW vendors, the 5000 Ryzens are OTR for this, intel just pulled this w/ their raptor lakes, releasing emergency microcodes updates…
Speaking of which, https://archlinux.org/packages/core/any/amd-ucode/
https://wiki.gentoo.org/wiki/AMD_microc … ware_files
/usr/lib/firmware/amd-ucode/README.zst
Offline
$ git bisect start
status: waiting for both good and bad commits
$ git bisect bad v6.10.3
status: waiting for good commit(s), bad commit known
$ git bisect good v6.10.2
Bisecting: 404 revisions left to test after this (roughly 9 steps)
[e158d5ecd0c0676d13524275f8b7ff06f4468163] usb: typec-mux: nb7vpq904m: unregister typec switch on probe error and remove
linux-stable-6.10.2.r405.ge158d5ecd0c0-1-x86_64.pkg.tar.zst/linux-stable-headers-6.10.2.r405.ge158d5ecd0c0-1-x86_64.pkg.tar.zst
Last edited by loqs (2024-08-09 21:31:54)
Online
Repeat after me: it **very most likely** /is/ faulty hardware.
The kernel update will do something somewhere to expose that, but you do not get spontanous reboots from kernel bugs.I guess loqs is already compiling bisection kernels but even iff and when we identify the breaking commit, be prepared to here from upstream that this is not gonna be reverted because your system is undervolted - so start looking into hardware related mitigations *now*, or you might find yourself in a position saying "I should have looked into hardware related mitigations while I still had a chance"
No one ever mentioned undrvolting, but you. We're talking about LAPTOP chips here, which usually can't be undervolted in BIOS and I'm not even sure if you can reliably undervolt them, considering that they have way stricter power limits. Also, what "hardware related mitigations" are you even talking about?
Offline
The "undervolting" is more like an optimisitic calibration of HW vendors
Not you - Acer, Lenovo, whatever to brag about battery lifetime.
If the system doesn't overheat and the RAM is ok, this is some CPU issue and it's most likely power related.
The system kernel doesn't just hard-reboot the system because it feels like. Also there's a very distinct HW pattern emerging.
Also, what "hardware related mitigations" are you even talking about?
Driving the CPU at higher voltages or lower c-states. Depending on what actually causes this.
Or RMA.
Offline
4th time random reboot in 2 days. Think my update/upgrade was a day or 2 before that. Never happened before. Ryzen 7900.
I have avoided all the previous Ryzen processor/firmware bugs/fixes in all these years but looks like the combo of something on new kernels? triggered something in the ryzens? Ah well. I am not grabbing a drop box 'kernel' , does arch nnot keep even a single prior version in the sync dbs? Damn.
Offline
The issue with Kernel 6.10.3 seems to be popping up in other places too.
https://www.reddit.com/r/Fedora/comment … em_did_it/
4th time random reboot in 2 days. Think my update/upgrade was a day or 2 before that. Never happened before. Ryzen 7900.
I have avoided all the previous Ryzen processor/firmware bugs/fixes in all these years but looks like the combo of something on new kernels? triggered something in the ryzens? Ah well. I am not grabbing a drop box 'kernel' , does arch nnot keep even a single prior version in the sync dbs? Damn.
Just download 6.10.2 and downgrade.
wget https://archive.archlinux.org/packages/l/linux/linux-6.10.2.arch1-2-x86_64.pkg.tar.zst
wget https://archive.archlinux.org/packages/l/linux/linux-headers-6.10.2.arch1-2-x86_64.pkg.tar.zst
sudo pacman -U linux-6.10.2.arch1-2-x86_64.pkg.tar.zst linux-headers-6.10.2.arch1-2-x86_64.pkg.tar.zst
Offline
The issue with Kernel 6.10.3 seems to be popping up in other places too.
https://www.reddit.com/r/Fedora/comment … em_did_it/
Edit- Reinstalling it fixed the issue weirdly enough
THAT smells a LOT like user-error
I doubt that if it really is some kernel change installing from scratch would magically fix it - it just doesn't add up
Offline
Just download 6.10.2 and downgrade.
wget https://archive.archlinux.org/packages/l/linux/linux-6.10.2.arch1-2-x86_64.pkg.tar.zst wget https://archive.archlinux.org/packages/l/linux/linux-headers-6.10.2.arch1-2-x86_64.pkg.tar.zst sudo pacman -U linux-6.10.2.arch1-2-x86_64.pkg.tar.zst linux-headers-6.10.2.arch1-2-x86_64.pkg.tar.zst
To avoid skipping signature verification and add the packages to the pacman cache:
# pacman -U https://archive.archlinux.org/packages/l/linux/linux-6.10.2.arch1-2-x86_64.pkg.tar.zst https://archive.archlinux.org/packages/l/linux/linux-headers-6.10.2.arch1-2-x86_64.pkg.tar.zst
@Gnat if you are not willing to test the bisection kernel I provided, I can provide you with the PKGBUILD so you can build it yourself or possibly grommit could provide you with a signed bisection kernel.
Online
Ah sorry, didn't really get that you wanted me to test it.
I've got it running right now. No instant crash yet.
Since i dont really know how to trigger the crash/reboot (apart from using it for some time) i will report back later.
Edit: Man, its only slightly sus to download a kernel from a GDrive
Seems to be running stable after 30 minutes.
Last edited by Gnat (2024-08-10 15:52:32)
Offline