Hardware error causing WiFi to crash and PC to hang

indigoblue · 2026-01-25 15:26:50

Hi everyone! This is my first post here so apologies if it isn't quite formatted well or as informative as it needs to be.

I'm dealing with an issue where during heavy load (e.g. watching HD video, downloading Steam games) my WiFi eventually stops working. All the information is below. I'm not sure what else to do at this point I hope this is an issue with how I have configured NetworkManager. Any guidance would be greatly appreciated. Thank you very much for taking the time to look at my issue

Full journalctl and dmesg

These are from the same session where the crash was observed

journalctl: http://0x0.st/PZYW.txt

dmesg: http://0x0.st/PZYV.txt

Background to Problem

The output of

lspci -nn | grep -i network

is

07:00.0 Network controller [0280]: Intel Corporation Wi-Fi 6E(802.11ax) AX210/AX1675* 2x2 [Typhoon Peak] [8086:2725] (rev 1a)

Specifically it is this WiFi card

As I said, this issue only shows up on heavy load. If I don't download anything large, it won't happen.
But, if I watch HD videos on YouTube, download Steam games, after around 10-30 minutes the WiFi will suddenly cut out.
Usually, this is associated with the whole machine freezing up too, sometimes this doesn't happen though.
If the machine isn't totally frozen,

systemctl restart NetworkManager

doesn't help - the command just hangs forever.

I've been able to identify the issue in dmesg, the full output is in the paste but I think the key lines are

[ 1803.931925] iwlwifi 0000:07:00.0: Hardware error detected. Restarting.
[ 1934.781501] iwlwifi 0000:07:00.0: Error sending SYSTEM_STATISTICS_CMD: time out after 2000ms.

Since the iwlwifi 0000:07:00.0 matches the 07:00.0 Network controller, I'm sure the issue is with the WiFi hardware. Additionally, this doesn't happen on ethernet (though, it's sadly impractical for me to be on ethernet all the time).

Also, this machine has previously had Windows installed, and I didn't encounter this issue with Windows. Therefore, I suspect it is a configuration issue or Linux bug, as I think the hardware is okay. Currently I only have Linux installed and am not dual booting.

Things I've Tried So Far

Different Kernels

I've tried with 6.18.6.arch1-1 and 6.12.67-1 LTS, but both have the exact same behaviour.

Disabling Powersaving

To

/etc/modprobe.d/iwlwifi.conf

I added

options iwlwifi power_save=0 uapsd_disable=1 disable_11ax=1

To

/etc/NetworkManager/conf.d/wifi-powersave.conf

I added

[connection]
wifi.powersave = 2

And rebooted, but did not help

PCIE ASPM Off

Tried setting

pcie_aspm=off

in the kernel params, and rebooted, but did not help

Disable D3 Cold

To

/etc/udev/rules.d/99-disable-d3cold-ax210.rules

I added

ACTION=="add", SUBSYSTEM=="pci", ATTR{vendor}=="0x8086", ATTR{device}=="0x2725", ATTR{d3cold_allowed}="0"

And rebooted, but did not help

Links

I've been able to find some bug reports with the same dmesg on bugzilla, going back a few years to Linux 5

https://bugzilla.kernel.org/show_bug.cgi?id=196411
https://bugzilla.kernel.org/show_bug.cgi?id=211925
https://www.mail-archive.com/ubuntu-bug … 13400.html

It could be a Kernel bug? None of those ever seemed to get resolved conclusively

This blog is where I got a lot of ideas from the fixes I've tried so far from, he seems to have the exact same problems as me, but the fixes haven't worked for me

https://edu4rdshl.dev/posts/fixing-inte … -on-linux/

Last edited by indigoblue (2026-01-25 15:27:25)

seth · 2026-01-25 16:00:20

Neither

[ 1803.931925] iwlwifi 0000:07:00.0: Hardware error detected. Restarting.

nor any other traces of wifi failure are in http://0x0.st/PZYW.txt but there's a firmware crash and subsequent

[ 1934.799047] WARNING: CPU: 12 PID: 1062 at drivers/net/wireless/intel/iwlwifi/pcie/gen1_2/trans.c:2387 __iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]

in the http://0x0.st/PZYV.txt dmesg.

Try to add "iwlmvm.power_scheme=1" to the https://wiki.archlinux.org/title/Kernel_parameters but if this is exclusively *during* high network pressure it might not be powersaving.
You could also try to "iwlwifi.swcrypto=1"

indigoblue · 2026-01-25 19:10:04

Thanks for your reply!

No luck on "iwlmvm.power_scheme=1" or "iwlwifi.swcrypto=1" - I tried them alone and together but still got the same behaviour.

I was also thinking along similar lines in terms of powersaving. A LOT of errors like this seem associated with powersaving, but why would powersaving kick in during high load?

At the moment I'm out of ideas, I'll have to do some more searching in the next couple days when I get a chance

seth · 2026-01-25 20:18:32

Do you have a system journal covering the firmware crash for pot. userspace context?

You could try to downgrade to an older firmware, https://wiki.archlinux.org/title/Networ … are_issues

Another thing is to grow the tcp buffer but I'm not sure whether that's related - do you also get this when downloading stuff into a tmpfs?
https://www.cyberciti.biz/faq/linux-tcp-tuning/

Jan 25 12:56:02 indigo-homepc kernel: DMI: ASUS System Product Name/ROG STRIX B650-A GAMING WIFI, BIOS 1811 10/07/2023

How's the wifi signal overall?
Expectable throughput? (You didn't forget to attach the little stub antenna you likely got w/ the board?)

indigoblue · 2026-01-26 21:36:39

I've never seen anything in journalctl which looks relevant to the crash. I know the following gets spammed out in dmesg at the exact moment of the crash though

[ 1803.931925] iwlwifi 0000:07:00.0: Hardware error detected. Restarting.
// Dump out registries here
[ 1814.202153] wlan0: Limiting TX power to 19 (20 - 1) dBm as advertised by c4:e5:32:53:97:bd
[ 1934.781501] iwlwifi 0000:07:00.0: Error sending SYSTEM_STATISTICS_CMD: time out after 2000ms.
[ 1934.781506] iwlwifi 0000:07:00.0: Current CMD queue read_ptr 415 write_ptr 416
[ 1934.799034] ------------[ cut here ]------------
[ 1934.799036] Timeout waiting for hardware access (CSR_GP_CNTRL 0xffffffff)
[ 1934.799047] WARNING: CPU: 12 PID: 1062 at drivers/net/wireless/intel/iwlwifi/pcie/gen1_2/trans.c:2387 __iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]
// Dump out registries here

I attempted a downgrade to linux-firmware-intel-20250627 but I still saw the crash.

I also tried downloading a bunch of stuff into tmpfs. I wasn't familiar with tmpfs until now, I followed the steps here https://linuxvox.com/blog/linux-tmpfs/ to mount one, I downloaded about 20Gb of stuff to there before I got the crash again

Until the crash, the WiFi signal is great. Speed is also exactly what I expect. Antenna is definitely attached.

One thing I have noticed recently which might be worth mentioning is my CPU usage goes through the roof when it happens - sometimes just one core hits 100%, but occasionally almost every core hits 100% usage. Really unclear why that is, or what it's even processing

At this point, I'm sure I've configured everything right. I think the problem has got to be in the linux-firmware-intel package, I'll start opening bug report tickets further upstream

seth · 2026-01-26 21:45:02

I know the following gets spammed out in dmesg at the exact moment of the crash though

Since the journal includes the ringbuffer that should™ absolutely also be there - next to the userspace logs.

CPU usage goes through the roof when it happens

top

but if it's not a userspace process it'll be the kernel handling the crash.

I think the problem has got to be in the linux-firmware-intel package, I'll start opening bug report tickets further upstream

The firmware crashes but the immediate problem would be the aftermath of that (which also maybe the cause of the fw crash)

The symptoms still look eerily like the known powersaving problems. Have you checked

systool -vm iwlwifi

to make sure the desired parameters actually apply?

indigoblue · 2026-02-19 19:08:52

Since the journal includes the ringbuffer that should™ absolutely also be there - next to the userspace logs

I just can't see anything that looks relevant in journalctl! Which I agree is weird but everything about this is looking weird to me

top shows occasional spikes in kwin_wayland. The problem is sometimes the crash doesn't cause too much issues other than just no internet, and sometimes totally locking up the PC. When it totally locks up, top stops updating so I'm not sure in that case what's using all the CPU.

systool -vm iwlwifi output

Module = "iwlwifi"

  Attributes:
    coresize            = "610304"
    initsize            = "0"
    initstate           = "live"
    refcnt              = "1"
    srcversion          = "4D55A30077AE93B8845155A"
    taint               = ""
    uevent              = <store method only>

  Parameters:
    11n_disable         = "0"
    amsdu_size          = "0"
    bt_coex_active      = "Y"
    debug               = "0"
    disable_11ac        = "N"
    disable_11ax        = "N"
    disable_11be        = "N"
    enable_ini          = "16"
    fw_restart          = "Y"
    led_mode            = "0"
    nvm_file            = "(null)"
    power_level         = "0"
    power_save          = "N"
    remove_when_gone    = "N"
    swcrypto            = "1"
    uapsd_disable       = "3"

  Sections:

`power_save = "N"` I believe indicates that power saving is turned off as I expect

seth · 2026-02-19 21:28:44

I just can't see anything that looks relevant in journalctl!

Can you upload a journal that covers such crash?

indigoblue · 2026-02-19 22:15:08

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

I uploaded this one before but I think the link expired - I can try again to make it crash tomorrow and get a new printout if necessary

seth · 2026-02-19 22:49:58

seth wrote:

Neither
[ 1803.931925] iwlwifi 0000:07:00.0: Hardware error detected. Restarting.
nor any other traces of wifi failure are in http://0x0.st/PZYW.txt

Nor anything like https://bbs.archlinux.org/viewtopic.php … 9#p2284829

Are you *sure* that anything like this happend during the 30 minutes covered by the journal from January 25th?

sudo journalctl | grep __iwl_trans_pcie_grab_nic_access

indigoblue · 2026-02-20 00:25:01

Ah! Bingo

Feb 10 18:42:13 indigo-homepc kernel: RIP: 0010:__iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]
Feb 11 20:26:42 indigo-homepc kernel: WARNING: CPU: 15 PID: 1132 at drivers/net/wireless/intel/iwlwifi/pcie/gen1_2/trans.c:2387 __iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]
Feb 11 20:26:42 indigo-homepc kernel: RIP: 0010:__iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]
Feb 12 20:41:27 indigo-homepc kernel: WARNING: CPU: 4 PID: 1090 at drivers/net/wireless/intel/iwlwifi/pcie/gen1_2/trans.c:2387 __iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]
Feb 12 20:41:27 indigo-homepc kernel: RIP: 0010:__iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]
Feb 19 18:16:13 indigo-homepc kernel: WARNING: CPU: 1 PID: 1075 at drivers/net/wireless/intel/iwlwifi/pcie/gen1_2/trans.c:2387 __iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]
Feb 19 18:16:13 indigo-homepc kernel: RIP: 0010:__iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]
Feb 19 18:43:42 indigo-homepc kernel: WARNING: CPU: 13 PID: 250 at drivers/net/wireless/intel/iwlwifi/pcie/gen1_2/trans.c:2387 __iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]
Feb 19 18:43:42 indigo-homepc kernel: RIP: 0010:__iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]
Feb 19 19:12:48 indigo-homepc kernel: WARNING: CPU: 14 PID: 1074 at drivers/net/wireless/intel/iwlwifi/pcie/gen1_2/trans.c:2387 __iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]
Feb 19 19:12:48 indigo-homepc kernel: RIP: 0010:__iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]
Feb 19 21:01:36 indigo-homepc kernel: WARNING: CPU: 18 PID: 1078 at drivers/net/wireless/intel/iwlwifi/pcie/gen1_2/trans.c:2387 __iwl_trans_pcie_grab_nic_access+0x163/0x170 [iwlwifi]

I haven't been keeping a concrete log of every crash but I believe they all correspond to one/around the time of one

Last edited by indigoblue (2026-02-20 00:25:54)

seth · 2026-02-20 09:01:04

journalctl --list-boots
sudo journalctl -b [i]IDX[/i]

Arch Linux

#1 2026-01-25 15:26:50