[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#662895: i8042: keys get stuck and keyboard stops responding



Hi,

GSR wrote:

> Newly installed machine and every few hours the PS/2 keyboard dies
> leaving a key stuck, but the USB mouse still works. Focusing different
> windows shows something being typed or a modifier key taking effect
> (depends, each time is a different key, sometimes they do not conflict
> with the wm and you can also move windows with mouse, etc). Alt-Sysrq
> or C-A-Fx stop working too when it happens.
>
> Accesing the machine via ssh shows the system still works, allows
> changing VT with chvt (just to see what they have, no input allowed),
> killing programs or rebooting, but no obvious way to recover the kbd
> other than restart. Echoing things to /proc/sysrq-trigger is allowed
> but the naK or Raw ones do not manage to get the kbd back in line,
[...]
> After searching some more and finding I am not alone with respect to
> stuck keyboard, I also found that someone suggests (in
> https://bugzilla.redhat.com/show_bug.cgi?id=676446 ) to invoke
>
> echo -n reconnect > /sys/devices/platform/i8042/serio0/drvctl
>
> and so far it seems to recover the keyboard

Thanks for a clear report, and sorry for the slow response.
Quick ideas:

 1) Please try a squeeze kernel (it should work fine on a
    wheezy/sid system) to see if this is a regression.

 2) Please attach /proc/interrupts

 3) Output from acpidump would also be interesting.

Hope that helps,
Jonathan

[...]
> In the recent days I have seen lines in the logs like:
>
> ---8<---
> atkbd serio0: Unknown key pressed (translated set 2, code 0x5b on isa0060/serio0).
> atkbd serio0: Use 'setkeycodes 5b <keycode>' to make it known.
> atkbd serio0: Unknown key pressed (translated set 2, code 0x0 on isa0060/serio0).
> atkbd serio0: Use 'setkeycodes 00 <keycode>' to make it known.
> atkbd serio0: Unknown key pressed (translated set 2, code 0xc0 on isa0060/serio0).
> atkbd serio0: Use 'setkeycodes e040 <keycode>' to make it known.
> hda-intel: IRQ timing workaround is activated for card #0. Suggest a bigger bdl_pos_adj.
> ACPI: EC: GPE storm detected, transactions will use polling mode
> atkbd serio0: Unknown key pressed (translated set 2, code 0x8b on isa0060/serio0).
> atkbd serio0: Use 'setkeycodes e00b <keycode>' to make it known.
> --->8--
>
> but they are not immediate to the kbd hang. Just adding in case they
> are related or can provide extra hints about what is going on (IRQ
> mess? ACPI gone mad?). The serio0 ones show different codes each time.
[...]
> [    0.000000] DMI: System manufacturer System Product Name/P7P55D, BIOS 1807    07/22/2010
[...]
> bios_vendor: American Megatrends Inc.
> bios_version: 1807   
> board_vendor: ASUSTeK Computer INC.
> board_name: P7P55D
> board_version: Rev 1.xx



Reply to: