[SOLVED] Intel AX200 driver crashes computer

f1refly · 2024-02-23 19:50:28

I just got a wifi card because powerline isn't stable in my flat. I read online that intel cards are better supported than realtek so after some looking around I got an intel ax200 + pcie adapter because my computer doesn't have a free m.2 slot. After installing it seemeed to work for a few hours, but suddenly the network disconnected and I couldn't get it to work again without a reboot. After some more tries, each ending with a crashed iwlwifi, I also noticed that other attached devices would cease to work, sometimes my mouse and keyboard (attached to the same usb port), sometimes my screen wouldn't update, sometimes the system would freeze alltogether. Other times, it would just be the wifi again.

I tested the latest linux-lts and have updated to the newest versions of all packages. I also tested various options for iwlwifi and iwlmvm I found online in related discussions, but they didn't seem to have any effect on the problem:

options iwlwifi 11n_disable=1 power_save=0
options iwlmvm power_scheme=1

Some information about my system:

# sudo dmidecode -t 2
# dmidecode 3.5
Getting SMBIOS data from sysfs.
SMBIOS 3.5.0 present.

Handle 0x0002, DMI type 2, 15 bytes
Base Board Information
	Manufacturer: ASUSTeK COMPUTER INC.
	Product Name: PRIME X670-P
	Version: Rev 1.xx
	Serial Number: MB-1234567890
	Asset Tag: Default string
	Features:
		Board is a hosting board
		Board is replaceable
	Location In Chassis: Default string
	Chassis Handle: 0x0003
	Type: Motherboard
	Contained Object Handles: 0

# sudo lspci -v
[...]
08:00.0 Network controller: Intel Corporation Wi-Fi 6 AX200 (rev 1a)
	Subsystem: Intel Corporation Wi-Fi 6 AX200NGW
	Flags: bus master, fast devsel, latency 0, IRQ 37, IOMMU group 19
	Memory at fc600000 (64-bit, non-prefetchable) [size=16K]
	Capabilities: [c8] Power Management version 3
	Capabilities: [d0] MSI: Enable- Count=1/1 Maskable- 64bit+
	Capabilities: [40] Express Endpoint, MSI 00
	Capabilities: [80] MSI-X: Enable+ Count=16 Masked-
	Capabilities: [100] Advanced Error Reporting
	Capabilities: [14c] Latency Tolerance Reporting
	Capabilities: [154] L1 PM Substates
	Kernel driver in use: iwlwifi
	Kernel modules: iwlwifi
[...]

Here's an exemplary crash dump: dmesg. Notice how there are many crashes back to back at the end of the file, while one trace is right during system boot.

Last edited by f1refly (2024-02-26 09:13:45)

V1del · 2024-02-23 22:05:23

Try updating there have been a bunch of fixes for the wifi subsystem and iwlwifi in particular in the just released linux-lts 6.6.18

seth · 2024-02-23 22:58:39

If that doesn't fix it, please post your complete system journal for an affected boot:

sudo journalctl -b -1 | curl -F 'file=@-' 0x0.st

for the previous one.
From the symptoms "pcie_aspm=off", https://wiki.archlinux.org/title/Kernel_parameters

Edit:

powerline isn't stable in my flat

Direct wall outlet and avoid nonlinear resistors on the line (tansformators and the order in which you plug PC and PL can have a huge impact…)

Last edited by seth (2024-02-23 23:05:17)

f1refly · 2024-02-24 12:43:29

V1del wrote:

Try updating there have been a bunch of fixes for the wifi subsystem and iwlwifi in particular in the just released linux-lts 6.6.18

I did, but unfortunately it didn't seem to have had any effect

seth wrote:

From the symptoms "pcie_aspm=off"

I tried

pcie_aspm=off

and saw no effects. Here's a journalctl dump of the last boot which ended with my computer freezing and me restarting it by holding the power button: http://0x0.st/H5lX.txt

Ignore the broken borgbackup, I think at least that should be fixed now

Last edited by f1refly (2024-02-24 13:13:47)

seth · 2024-02-24 15:01:32

Feb 24 08:29:01 host kernel: Command line: BOOT_IMAGE=/vmlinuz-linux-lts [###filesystems###]

Don't do that - the UUIDs are not sensitive data but to prevent access collisions and you probably killed more kernel parameters.

Feb 24 08:29:14 host NetworkManager[920]: <info>  [1708759754.3652] manager: NetworkManager state is now CONNECTED_SITE
Feb 24 08:29:14 host NetworkManager[920]: <info>  [1708759754.4755] manager: NetworkManager state is now CONNECTED_GLOBAL
Feb 24 09:42:27 host NetworkManager[920]: <info>  [1708764147.8363] manager: NetworkManager state is now CONNECTED_SITE
Feb 24 09:42:27 host NetworkManager[920]: <info>  [1708764147.9471] manager: NetworkManager state is now CONNECTED_GLOBAL

and there's no iwlwifi crash in that journal at all?

However

restarting it by holding the power button

Avoid that at all costs: https://wiki.archlinux.org/title/Keyboa … el_(SysRq)

f1refly · 2024-02-24 16:41:28

seth wrote:

Don't do that - the UUIDs are not sensitive data but to prevent access collisions and you probably killed more kernel parameters.

I might have been a bit over ethusiastic while removing information, sorry. My full cmdline is

Feb 24 08:29:01 lina kernel: Command line: BOOT_IMAGE=/vmlinuz-linux-lts root=UUID=30954e44-bae9-4732-b294-fc9638cc44e9 rw rootflags=subvol=@ pcie_aspm=off root=/dev/mapper/cryptroot rootflags=subvol=@ cryptdevice=/dev/disk/by-uuid/53255116-9ad3-4493-9da8-14f5d65577d2:cryptroot

, so they where missing the recommended

pcie_aspm=off

and

rw

seth wrote:

and there's no iwlwifi crash in that journal at all?

I just noticed as well. I just saw the coredump in the journal and failed to see it wasn't the driver this time. Not sure why teamspeak3 crashed there, it's not the best of programs. I wonder why the log doesn't report a connectivity loss at the end, because I definitly lost connection. I'll try to get a better log. This was the last boot before the current one:

http://0x0.st/H5D6.txt

In this case, a watchdog rebooted the system without my interaction while I was away.

seth · 2024-02-24 17:14:57

Feb 24 15:33:52 host wpa_supplicant[977]: wlp8s0: WPA: Group rekeying completed with MAC [GTK=CCMP]
Feb 24 15:43:52 host wpa_supplicant[977]: wlp8s0: WPA: Group rekeying completed with MAC [GTK=CCMP]
Feb 24 15:53:52 host wpa_supplicant[977]: wlp8s0: WPA: Group rekeying completed with MAC [GTK=CCMP]
Feb 24 16:03:52 host wpa_supplicant[977]: wlp8s0: WPA: Group rekeying completed with MAC [GTK=CCMP]
Feb 24 16:13:52 host wpa_supplicant[977]: wlp8s0: WPA: Group rekeying completed with MAC [GTK=CCMP]
Feb 24 16:23:52 host wpa_supplicant[977]: wlp8s0: WPA: Group rekeying completed with MAC [GTK=CCMP]

That's rather often, but I suspect the problems start here:

Feb 24 15:53:52 host wpa_supplicant[977]: wlp8s0: WPA: Group rekeying completed with MAC [GTK=CCMP]
Feb 24 16:01:00 host CROND[6457]: (root) CMD (run-parts /etc/cron.hourly)
Feb 24 16:01:00 host CROND[6456]: (root) CMDEND (run-parts /etc/cron.hourly)
Feb 24 16:03:27 host syncthing[981]: [BAWBN] INFO: Sent usage report (version 3)
Feb 24 16:03:52 host wpa_supplicant[977]: wlp8s0: WPA: Group rekeying completed with MAC [GTK=CCMP]
Feb 24 16:03:56 host syncthing[981]: [BAWBN] INFO: Couldn't add port mapping for 172.16.0.117 (external port 43734 -> internal port 22000/TCP): UPnP Error: Not available Action (403)

What's in your hourly cron jobs?
What if you increase the rekey interval (like 1h is pretty standard)?
What if you disable syncthing?

f1refly · 2024-02-24 17:51:43

seth wrote:

What's in your hourly cron jobs?

Just the default anacron call to check if anacron already ran and if it runs on battery. I can remove that file without a problem I think.

seth wrote:

What if you increase the rekey interval (like 1h is pretty standard)?

I'm not sure I can do that - it is my understanding that the rekey interval is set by the access point, right? And the interface of my access point doesn't seem to expose that setting to me (It's an AVM Fritzbox 7530 AX)

seth wrote:

What if you disable syncthing?

I will try disabling syncthing right now and using my computer for a while.

f1refly · 2024-02-24 20:18:46

Disabling syncthing did not help, it happened again: http://0x0.st/H5d7.txt

seth · 2024-02-24 22:08:22

Nothing and then crashes on the next scheduled re-key (yes, you'd configure that in the AP - and changing that in the FritzBox apparently really requires patching the firmware…)
You're not wildly scanning around.

Can you test this against a different AP (eg. your phone hotspot)?
Maybe the wifi is only the messenger, what else is on the bus?

lspci -tvnn

f1refly · 2024-02-25 07:49:03

seth wrote:

Can you test this against a different AP (eg. your phone hotspot)?

Unfortunately I'm out and about today, I'll try this tomorrow.

seth wrote:

Maybe the wifi is only the messenger, what else is on the bus?

Output: http://0x0.st/H5hf.txt

seth · 2024-02-25 08:57:11

"Everything" - USB, nvme, sata and an RTL8125 ethernet

Do you have any external USB devices attached?
Can you use the rj45 connection for a while to test whether the system remains stable then?
https://wiki.archlinux.org/title/Solid_ … leshooting - "nvme_core.default_ps_max_latency_us=0 iommu=soft"

Last edited by seth (2024-02-25 08:57:24)

f1refly · 2024-02-25 18:51:56

seth wrote:

Do you have any external USB devices attached?

Mouse, Keyboard and a Logitech Headset dongle via a KVM switch and a steam controller wireless dongle directly.

seth wrote:

Can you use the rj45 connection for a while to test whether the system remains stable then?

This is how I used my computer for quite a while and never had any issues (apart from the abysmal speed), so it should be fine stability-wise. On the other hand, I noticed that the performance was even worse than I remembered it, so maybe there's really something wrong with the bus. When I use

seth wrote:

https://wiki.archlinux.org/title/Solid_ … leshooting - "nvme_core.default_ps_max_latency_us=0 iommu=soft"

I'll set those options for the next boot.

f1refly · 2024-02-25 19:32:06

Also, recently one of my HDDs lost all of its data. SMART data indicates it's all good, but the GPT table (and its backups?) vanished and I couldn't find any data to rescue as well. Since the sata controller is on the same pcie bus, maybe those events are correlated? I'm unfortunately unsure when the event occured precisely and if the wifi issues started only after the hdd died...

EDIT:
I confused two volume. The volume that couldn't be found anymore actually resides on two ssds which are also attached via sata. When I remove the wifi card from my system those ssdd are found again and mount without a problem! So, something has to be seriously broken here and seems to interrupt pcie functionality.

Last edited by f1refly (2024-02-25 20:15:45)

seth · 2024-02-25 21:20:36

When I remove the wifi card from my system

I got an intel ax200 + pcie adapter because my computer doesn't have a free m.2 slot.

Does just having the adapter there trash the system?
Can you plug the adapter into a different PCIe slot?

This is how I used my computer for quite a while and never had any issues (apart from the abysmal speed)

Which is because of

I just got a wifi card because powerline isn't stable in my flat.

? (You've a 2.5GBE, one would expect that to be serviceable…)

f1refly · 2024-02-25 21:44:05

seth wrote:

Does just having the adapter there trash the system?
Can you plug the adapter into a different PCIe slot?

I didn't disassemble the adapter yet, but I tried the first x1 slot (the one causing issues), the second x16 slot (card isn't recognized whatsoever, ssds are recognized again), and the third and last x16 slot (both network card and ssds are recognized).

seth wrote:

You've a 2.5GBE, one would expect that to be serviceable…

...which is then attached to a powerline module limiting speeds to ~1MiB/s because running cables in a flat I'm going to move out of in a year or two is not worth the effort

My computer has been running smoothly for the last hour, and some performance issues I've had with my graphics card didn't show this time as well. I'll use the computer for work tomorrow, hopefully the issue is solved now by physically moving the card to another slot. Seemingly either the graphics card or the wifi card emit enough EMI to cause errors.

f1refly · 2024-02-26 09:12:42

After a night idling and four hours of work the computer still works. I'll mark this as solved, thank you for helping me figure this out seth

Arch Linux

#1 2024-02-23 19:50:28

[SOLVED] Intel AX200 driver crashes computer

#2 2024-02-23 22:05:23

Re: [SOLVED] Intel AX200 driver crashes computer

#3 2024-02-23 22:58:39

Re: [SOLVED] Intel AX200 driver crashes computer

#4 2024-02-24 12:43:29

Re: [SOLVED] Intel AX200 driver crashes computer

#5 2024-02-24 15:01:32

Re: [SOLVED] Intel AX200 driver crashes computer

#6 2024-02-24 16:41:28

Re: [SOLVED] Intel AX200 driver crashes computer

#7 2024-02-24 17:14:57

Re: [SOLVED] Intel AX200 driver crashes computer

#8 2024-02-24 17:51:43

Re: [SOLVED] Intel AX200 driver crashes computer

#9 2024-02-24 20:18:46

Re: [SOLVED] Intel AX200 driver crashes computer

#10 2024-02-24 22:08:22

Re: [SOLVED] Intel AX200 driver crashes computer

#11 2024-02-25 07:49:03

Re: [SOLVED] Intel AX200 driver crashes computer

#12 2024-02-25 08:57:11

Re: [SOLVED] Intel AX200 driver crashes computer

#13 2024-02-25 18:51:56

Re: [SOLVED] Intel AX200 driver crashes computer

#14 2024-02-25 19:32:06

Re: [SOLVED] Intel AX200 driver crashes computer

#15 2024-02-25 21:20:36

Re: [SOLVED] Intel AX200 driver crashes computer

#16 2024-02-25 21:44:05

Re: [SOLVED] Intel AX200 driver crashes computer

#17 2024-02-26 09:12:42

Re: [SOLVED] Intel AX200 driver crashes computer

Board footer