Non-Page Memory Error At Seemingly Random Intervals

virtualras · 2026-01-02 03:54:36

I have been battling with these errors for a while now, and have ran out of options in my own understanding of Arch.

Background: I have been receiving an error that crashes my system. Sometimes it comes in a kernel panic (usually when playing a game), sometimes it just freezes. Nothing resolves it but a hard reset.
Hardware wise; I have tried a new everything. Ship of Theseus style, the entire PC is brand new but still experiencing the same issue. So hardware can, with relative safety, be ruled out. I have tried memtest for extended periods with no results, the CPU is brand new with fresh thermal paste, the cooler, RAM, GPU, Motherboard, even the case is new. I am relatively confident it is some software or configuration issue at hand.

As for software;
I am running the latest software and have noticed no difference. pacman -Syu has not yielded results.

My kernel is 6.18.2-arch2-1
My desktop environment is Hyprland
I am running the dot-4 hyprland dotfiles (not sure if it matters, but I see no issue stating it)
Systemd everything (where possible), which will come into the matter later

As for hardware:
CPU: Ryzen 7 9800X3D
GPU: Nvidia 4070Ti (I have also attempted this with an AMD 6900 XT, same results)
RAM: 16GB X 2 T-Force Delta RGB DDR5
MOBO: Gigabyte X870 Aorus Elite WIFI7 ICE
PSU: Lian Li EDGE 100W

all drivers are up to date, and have been regularly updated with no change to performance.

As I said before, much of this I believe is meaningless as I've tested so many different things that I have an entire second PC running Bazzite, I see no issues there so I can only find that I've messed up a configuration somewhere.

Journald does not record the errors when they arrive, just recording up to the errors. I have, however, taken a picture of one of the errors. I had

journalctl -f

running when I ran into the issue. I am looking for a way to upload images the follows the rules, please bear with me while I work through this, but the error states:

non-paged memory
list_del corruption. prev->next should be [index, i presume], but was [index, i also presume]

It then lists a series of modules that were linked, a CPU, UID, PID, as well as some other data. Please let me know the correct image hosting page to use so I can be as precise as possible.

This has plagued me across GPUs, CPUs, MOBOs, RAM, PSUs; I have no other options. I hope it is as simple as "do X" and makes me feel dumb.

Thanks

seth · 2026-01-02 11:14:58

Please let me know the correct image hosting page

imgur is a popular choice.

Also

Sometimes it comes in a kernel panic (usually when playing a game)

Do you get one of those fancy QR-coded bluescreens and therefore have a link for such panic?

virtualras · 2026-01-02 18:48:49

Thanks Seth, here's a link to the error. The kernel panic does at times give me a QR code, unfortunately I don't have the link saved, but it mentions "Fatal exception in interrupt" in the picture I have (the image is too poor for my phone to process the QR code). I'll try to save it in the future.

Potomac · 2026-01-02 19:54:05

Temperatures of components are Ok ?
Voltages ok ?

Try with one memory module installed to see if the problem disapears.

You can also test with LTS Linux kernel.

Kernel 6.18.2 has bugs with AMD graphic cards that can led to random crash, some users downgrade kernel to 6.17.9 version.

Last edited by Potomac (2026-01-02 19:58:33)

seth · 2026-01-02 20:35:18

Also, since it crashes from eviction, do you run a zram swap device AND zswap? (zram is loaded)
Try neither.

I didn't see amdgpu, but nvidia *might* be a candidate (just as every OOT driver)

virtualras · 2026-01-02 23:01:32

Temperatures appear to be normal. I had an AIO pump fail in the past, that failure caused an immediate shutdown, this is just straight freezing. I'll try with one memory stick and let you know if that changes anything.

I currently have an Nvidia card in, but I'll try downgrading the kernel if the single memory module still doesn't change anything.

I believe i only have zram. running lsblk brings back:

zram0       253:0    0    4G  0 disk [SWAP]

Running some games does cause zram to completely fill and crash the system (it's currently sitting at 98% full, spooky), not sure if that is a symptom of this problem or a different problem. Is it possible to run without swap? From what I understand that's not advisable, but I'm wondering if shutting it off stops the crashes, possibly pointing to a zram misconfiguration.

I'll try running with one memory module in the meantime and report back if that changes anything.

seth · 2026-01-02 23:12:54

I believe i…

Don't. Check:
https://wiki.archlinux.org/title/Zswap# … parameters

Is it possible to run without swap? From what I understand that's not advisable

Yes (if you don't have too high memory demands) and no, it's not.
=> https://wiki.archlinux.org/title/Swap#Swap_file

virtualras · 2026-01-03 00:16:25

> grep -r . /sys/module/zswap/parameters/
/sys/module/zswap/parameters/enabled:N
/sys/module/zswap/parameters/shrinker_enabled:Y
/sys/module/zswap/parameters/max_pool_percent:20
/sys/module/zswap/parameters/compressor:zstd
/sys/module/zswap/parameters/accept_threshold_percent:90

Here is the output.

seth · 2026-01-03 06:22:33

So - no.
However, and w/

Running some games does cause zram to completely fill and crash the system (it's currently sitting at 98% full, spooky)

your game experience, the compressed swap devices in RAM can completely blow up and into your face, so I'd disable the zram swap, create a swap file and see whether
a) you're swapping a lot there
b) you might benefit from lowering vm.swappiness (depending on the access patterns)
c) you still run into kernel panics or those were just results of the memory pressure.

Ceterum censeo: when you're constantly finding yourself swapping out a lot, you ideally need more RAM

virtualras · 2026-01-03 06:52:44

I'll try setting up a swap file and dropping zram. The curious thing is that it starts swapping a little too early and a little too liberally. I've had swap fill well before RAM is even halfway full.

Man, I hope I don't need more RAM. My wallet isn't ready for that reality

So far I haven't ran into any crashes today after Potomac's suggestion. Hopefully with yours, it'll behave with both sticks. If it does, I'll mark the issue as solved and hopefully that'll be that. Thanks everyone

Arch Linux

#1 2026-01-02 03:54:36

Non-Page Memory Error At Seemingly Random Intervals

#2 2026-01-02 11:14:58

Re: Non-Page Memory Error At Seemingly Random Intervals

#3 2026-01-02 18:48:49

Re: Non-Page Memory Error At Seemingly Random Intervals

#4 2026-01-02 19:54:05

Re: Non-Page Memory Error At Seemingly Random Intervals

#5 2026-01-02 20:35:18

Re: Non-Page Memory Error At Seemingly Random Intervals

#6 2026-01-02 23:01:32

Re: Non-Page Memory Error At Seemingly Random Intervals

#7 2026-01-02 23:12:54

Re: Non-Page Memory Error At Seemingly Random Intervals

#8 2026-01-03 00:16:25

Re: Non-Page Memory Error At Seemingly Random Intervals

#9 2026-01-03 06:22:33

Re: Non-Page Memory Error At Seemingly Random Intervals

#10 2026-01-03 06:52:44

Re: Non-Page Memory Error At Seemingly Random Intervals

Board footer