FS#75587 - [linux] iwlwifi crashes when roaming from associated access point to a new access point

Attached to Project: Arch Linux
Opened by Tommy Zhang (T-J-M) - Friday, 12 August 2022, 22:16 GMT
Last edited by Jelle van der Waa (jelly) - Thursday, 14 September 2023, 17:53 GMT
Task Type Bug Report
Category Kernel
Status Closed
Assigned To Tobias Powalowski (tpowa)
Jan Alexander Steffens (heftig)
David Runge (dvzrv)
Architecture All
Severity High
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No


Intel AC 7260 is a standard wireless adapter employed by multiple laptops. In arch kernel 5.18.16, this device is driven by iwlwifi kernel module. Given the mobility of laptops, it's not a rare case when one needs to connect to an 802.11 network deployed with multiple access points(APs), which usually comes with wap-enterprise. Most network engineering who design and deploy the enterprise 802.11 network don't intentionally avoid the emergence of overlapped regions between APs, as a particular device named access controller(AC) would handle the client-AP association relations in the overlapped areas. Therefore, given the signal strength fluctuation and the load condition of a particular AP, the AC might disassociate a specific client and appoint a new AP to associate. Therefore, a laptop with multiple AP in reach could periodically suffer from the reassociation process without movement. However, moving from one AP to another may artificially trigger a reassociation.
When the reassociation happens, iwlwifi crashes, and the kernel captures the error log. There is a current workaround without rebooting, which removes the adapter via /sys/class/pci_bus/0000:0X/device/remove, removes the iwlwifi related modules via modprob -r and then rescan the PCI devices via /sys/bus/pci/rescan. However, iwlwifi would crash next time when reassociation happens.

Additional info:
* package version(s): linux 5.18.16.arch1-1
* config: iwlwifi.conf
* log file: iwlwifi.dmesg

Steps to reproduce:
0. Borrow a ThinkPad T440 with FRU capable Intel AC-7260 wireless adapter
1. Use NetworkManager to connect to a WPA-Enterprise 802.11 network with multiple APs
2. With the laptop in hand, gently walk from one AP to another.
3. If nothing abnormal in dmesg, keep walking between multiple APs.
This task depends upon

Closed by  Jelle van der Waa (jelly)
Thursday, 14 September 2023, 17:53 GMT
Reason for closing:  Deferred
Additional comments about closing:  Old kernel, please retry with the latest
Comment by Tommy Zhang (T-J-M) - Friday, 12 August 2022, 22:31 GMT
P.S. The author of this bug report understands that Arch might not responsible for maintaining iwlwifi. However, the kernel bug tracker suggests that non-developers report kernel-related bugs to distro-specific bug trackers first. They probably expect someone knowledgeable to polish and forward this report to the upstream.
Comment by Toolybird (Toolybird) - Sunday, 14 August 2022, 00:11 GMT
> They probably expect someone knowledgeable to polish and forward this report to the upstream

That "someone knowledgeable" is you :) i.e. the person who can reproduce the problem is best placed to write a proper bug report and send it upstream. The usual advice for kernel bugs is here [1]. If it's a regression then bisection is an option. It's also worth trying latest -rc kernel to see if the issue is fixed. Otherwise you can try the support channels (forum/IRC/etc) to try and find knowledgeable and helpful community members such as @loqs

[1] https://wiki.archlinux.org/title/Kernel#Troubleshooting