Hi Salvatore, sure I can run any test to help, but given I had fixed my
original issue, first let's agree on how to prepare the environment to try
to reproduce the original scenario and test it.
1 - Rollback fixes:
 This means block the devices previously allowed via usbguard. (I could try
to do it with the 6.11.5 and see if it lock) or do it with kernel 6.10.12
which allowed me to fix it originally.
 Another change i should do, are some new udev rules added on my machine to
avoid the mt7921e wifi be suspended. Details at (
https://github.com/fritzmatias/fritzmatias/blob/main/linux/kernel/mt7921e.md
)
2- Then try to reproduce the kernel crash running the 'usbguard
list-devices' with kernel 6.11.5 as originally. (what to do if does not
:O).
3- if step 2 is reproducible, Add mt7921e udev rules, and run 'usbguard
list-devices'. This could validate my guess.
4- remove mt7921e udev rules.
5- If step 2 is confirmed, then uninstall OOT nvidia drivers. And run
'usbguard list-devices' with kernel 6.11.5.
6- independently of step 5's output. Install nvidia open kernel 545, and
run 'usbguard list-devices' with kernel 6.11.5, to confirm/deny
similarities as step 2.

Let me know if that procedure is good enough or if you can improve it.
My speculative guess, is related to the buggy hw mt7921e (wifi &
bluetooth): when it fails because of a suspension call, the usb core locks,
unable to handle that scenario.

Open to suggestions
Thanks in advance.

Reply via email to