[Bug] AC Evo WHEEL CRASH SPOTTED

[fenixadam]

Category

Other

Description

I finally caught the crash (apparently it does not show on journalctl).

I disabled the iGPU and i was able to see with the dedicated GPU the crash.

Steps to Reproduce

• Open AC or AC Evo
• At some point restarting session multiple times will trigger crash

Kernel Version

7.2.8-2-cachyos

Driver version

v0.42.4

Distribution

Cachy OS

Installation Method

DKMS

dmesg Output

Current sysfs Values (if relevant)

Game/Application

AC EVO or AC

USB capture (recommended for FFB / wheel-state bugs)

No response

Checklist

• I installed the driver via setup.sh (or dkms-update.sh)
• I verified the driver loaded (lsmod | grep hid_logitech_dd)
• I checked dmesg for errors

Additional Context

No response

[mescon]

That photo is the whole answer, thank you; it is the freeze #90 was chasing, and it never reached the journal because it dies inside the USB interrupt. The fault is in this driver: it matched every incoming report against the previous command's reply buffer, which lives on the stack of whoever sent that command, whenever the send lock was held. Four of our own writers (the rev-light worker among them, on your wheel) hold that lock without a reply buffer, so once the process that sent the last command had exited, the next report during their hold read a stack that no longer existed. Fix in #129; it ships in 0.42.5 once CI is green.

Update: 0.42.5 is out on every channel with the fix. Please run it for a while in your usual sessions; a week without a freeze and this closes.

[fenixadam]

Thank you very much for your response. :)