Bug#576918: Keyboard disabled: atkbd.c: Spurious ACK on isa0060/serio0 ...
On Thu, Feb 09, 2012 at 09:44:57PM -0600, Jonathan Nieder wrote:
> ael wrote:
> > With Linux 2.6.32-trunk-686 & Linux 2.6.32-3-686, at least, on this Aspire One
> > Netbook, the system frequently boots into a state in which the keyboard does
> > not respond. The mouse (Synaptics touchpad) is still active, and as far as I can
> > tell everything else is working. My only recourse is to hit the reset button.
> > So this is nondeterministic, and so probably some sort of race? I *think* that
> > it happens more frequently if I happen to touch the ps/2 touchpad during boot, but
> > it definitely does still happen when I keep my hands well away during boot.
> Hey, neat. What's the newest kernel you've tried? Please attach full
> "dmesg" output from a good boot and a bad boot, even though as you
> mentioned there might not be any significant difference between them.
> Sorry for the long quiet.
Well, it really is a long time. I am running testing so I have been
through many kernels since this report. With recent kernels - I am
running 3.1.0-1-686-pae today - this no longer happens.
*However*, there is still some nondeterministic misbehaviour with
1) On maybe 10% of boots, the right button is not recognised. I test
with mev, or if I have started X, then I use xev. There is simply
no response to the button. Reloading the psmouse module (and
presumably its dependants) does not solve the problem. Rebooting seems
to be the only recourse.
2) In X, the VerticalEdgeScroll & HorizEdgeScroll sometimes do not
work. That is cured by "rmmod psmouse; modprobe psmouse". Not sure
that this is same problem, but again nondeterministic.
I realize that this information is unlikely to be of much help in
trying to localize the problem. My *impression* as I have seen various
kernels behave in slightly different ways is that there are races
happening, but the trigering and effects vary in frequency and
importance as the code context changes. Just as one might expect :-)
If you have any suggestions for how to diagnose the problem, let me
I will try to collect a dmesg.diff for boots with working and failing
right button recognition, but IIRC I tried that sort of thing
earlier and could see no useful information.
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact firstname.lastname@example.org