Bug 217077

Summary:	iwlwifi 6235: Wifi hardware does not recover from a ucode error. Kernel panic when reloading wifi-related modules
Product:	Drivers	Reporter:	Leo (muttoni)
Component:	network-wireless-intel	Assignee:	Default virtual assignee for network-wireless-intel (drivers_network-wireless-intel)
Status:	NEW ---
Severity:	normal	CC:	muttoni
Priority:	P1
Hardware:	Intel
OS:	Linux
Kernel Version:	6.1.12	Subsystem:
Regression:	No	Bisected commit-id:
Attachments:	dmesg after wifi stopped

Description Leo 2023-02-23 18:45:05 UTC

Created attachment 303774 [details]
dmesg after wifi stopped

Abstract: While using my wifi card "Intel 6235" sometimes the wifi stops working, eventually leading to a kernel panic (for example when I reload the wifi related kernel modules).

Long history:

Just after connecting to a network sometimes the wifi stops working. When the problem occour, the dmesg messages that I think are prominent are:
	"Error sending REPLY_SCAN_ABORT_CMD", 
	"Failed to load firmware chunk!"
	"Hardware became unavailable during restart"

One option is to restart and hope the wifi works. Other is to reload all the wifi related kernel modules:

    rmmod iwlwifi iwldvm mac80211 cfg80211 # I repeat to ensure all modules are unloaded
    #   (wait some time)
    modprobe iwlwifi

After this, sometimes the wifi works normally again, but most of the time the system starts behaving strangely, and a few seconds later the crash occurs, with the caps-lock led blinking (kernel panic).

The problem occurs intermittently. Most of the time the problem occurs a few seconds after login, when KDE/networkmanager tries to connect with my AP on a cold boot. But sometimes weeks pass with the system normally (even with dozens of suspend-resume cycles). I tried a few things but couldn't come to any conclusion as to what causes the problem:

1) Different options for iwlwifi (current one is options iwlwifi bt_coex_active=0 11n_disable=8).

2) Set some kernel parameters, like "pcie_aspm=force enable_mtrr_cleanup mtrr_spare_reg_nr=1 mtrr_gran_size=64M mtrr_chunk_size=256M"

3) Activate or deactivate my "external" GPU ([AMD/ATI] Mars [Radeon HD 8670A/8670M/8750M / R7 M370]) (currently deactivated)

4) Connect to my AP's 2.4 GHz channel, instead of the 5 GHz one.

I've enabled the iwlwifi debug with command below and collected the dmesg after the wifi stopped working:
    echo 0x7ffff > /sys/module/iwlwifi/parameters/debug

Note that my report is very similar to bug 190281, but in this one I attach the dmesg of the driver with debug enabled, and in my case there is a kernel panic sometimes.

I always compile my own kernel, and as far as I can remember this problem has been going on at least since kernel version 3.8.x.

When the wifi works, it works great (with download and upload speeds in the 100mbps range).

I know the device is old and the firmware is no longer supported. But I'm not asking for a firmware correction: being a firmware crash, I just wanted that, in case of a firmware error, it would be successfully reloaded and the wifi hardware would work again, without kernel panic. 

Please let me know if you need additional information.

Firmware-version: 18.168.6.1 6000g2b-6.ucode.