2004-01-19 18:08:53

by fire-eyes

[permalink] [raw]
Subject: 2.6.1: atkbd.c errors + mouse errors with a belkin KVM

I'm seeing some strange behavior using kernel 2.6.1 and a Belkin KVM. I
have talked to two others who have seen the same problem. This did not
happen in 2.6.0 or any 2.4 kernel.

When switching ports using the keyboard, that is hitting scroll lock
twice and then the port # I want ( 1 - 16), then coming back the same
way, I often get the following error:

kernel: atkbd.c: Unknown key released (translated set 2, code 0x7a on
isa0060/serio0).

I have also seen this same error while switching from X11 to console,
but only once. In that one case, I lost mouse control, and keyboard
control, and was unable to regain either. I had to reboot to rectify the
situation.

I'm not sure if this happens when pushing the buttons on the front of
the kvm.

KVM Information:

Belkin Omniview Matrix2 , Model F1DM216T
Maufacturer URL:
http://catalog.belkin.com/IWCatProductPage.process?Merchant_Id=1&Product_Id=122933

Is there any further information I can submit to help solve this?


2004-01-19 19:29:29

by Andries Brouwer

[permalink] [raw]
Subject: Re: 2.6.1: atkbd.c errors + mouse errors with a belkin KVM

On Mon, Jan 19, 2004 at 01:08:47PM -0500, fire-eyes wrote:

> I'm seeing some strange behavior using kernel 2.6.1 and a Belkin KVM.
> ... I often get the following error:
>
> kernel: atkbd.c: Unknown key released (translated set 2, code 0x7a on
> isa0060/serio0).

It is not really an error - the kernel is just being a bit noisy.
The 0x7a was really 0xfa, the ACK that a keyboard command succeeded.

Sooner or later the noise will go away. For now it is more interesting
to fix bugs in behaviour.

2004-01-19 20:43:55

by Vojtech Pavlik

[permalink] [raw]
Subject: Re: 2.6.1: atkbd.c errors + mouse errors with a belkin KVM

On Mon, Jan 19, 2004 at 08:29:05PM +0100, Andries Brouwer wrote:
> On Mon, Jan 19, 2004 at 01:08:47PM -0500, fire-eyes wrote:
>
> > I'm seeing some strange behavior using kernel 2.6.1 and a Belkin KVM.
> > ... I often get the following error:
> >
> > kernel: atkbd.c: Unknown key released (translated set 2, code 0x7a on
> > isa0060/serio0).
>
> It is not really an error - the kernel is just being a bit noisy.
> The 0x7a was really 0xfa, the ACK that a keyboard command succeeded.
>
> Sooner or later the noise will go away. For now it is more interesting
> to fix bugs in behaviour.

Well, the kernel is quite rightfully noisy at this point - getting
unexpected ACKs is rather suspicious.

--
Vojtech Pavlik
SuSE Labs, SuSE CR

2004-01-19 21:41:05

by Andries Brouwer

[permalink] [raw]
Subject: Re: 2.6.1: atkbd.c errors + mouse errors with a belkin KVM

On Mon, Jan 19, 2004 at 09:43:48PM +0100, Vojtech Pavlik wrote:

> > > kernel: atkbd.c: Unknown key released (translated set 2, code 0x7a on
> > > isa0060/serio0).
> >
> > It is not really an error - the kernel is just being a bit noisy.
> > The 0x7a was really 0xfa, the ACK that a keyboard command succeeded.
> >
> > Sooner or later the noise will go away. For now it is more interesting
> > to fix bugs in behaviour.
>
> Well, the kernel is quite rightfully noisy at this point - getting
> unexpected ACKs is rather suspicious.

Yes, I do not propose that you remove the noise today.
But nothing is wrong.

Several utilities access the keyboard directly. (For example X and kbdrate.)
Since setting keyboard rate and starting X are rather common actions,
most people will see these messages a few times in their syslog.

A few months from now, when we are satisfied that the keyboard code
is in perfect shape, these messages can be silenced by default.

Andries

2004-01-19 21:50:40

by Vojtech Pavlik

[permalink] [raw]
Subject: Re: 2.6.1: atkbd.c errors + mouse errors with a belkin KVM

On Mon, Jan 19, 2004 at 10:40:36PM +0100, Andries Brouwer wrote:
> On Mon, Jan 19, 2004 at 09:43:48PM +0100, Vojtech Pavlik wrote:
>
> > > > kernel: atkbd.c: Unknown key released (translated set 2, code 0x7a on
> > > > isa0060/serio0).
> > >
> > > It is not really an error - the kernel is just being a bit noisy.
> > > The 0x7a was really 0xfa, the ACK that a keyboard command succeeded.
> > >
> > > Sooner or later the noise will go away. For now it is more interesting
> > > to fix bugs in behaviour.
> >
> > Well, the kernel is quite rightfully noisy at this point - getting
> > unexpected ACKs is rather suspicious.
>
> Yes, I do not propose that you remove the noise today.
> But nothing is wrong.
>
> Several utilities access the keyboard directly. (For example X and kbdrate.)
> Since setting keyboard rate and starting X are rather common actions,
> most people will see these messages a few times in their syslog.

kbdrate is supposed to use the ioctls, as is X. Actually kbdrate does
use the ioctls on my systems.

> A few months from now, when we are satisfied that the keyboard code
> is in perfect shape, these messages can be silenced by default.

Actually, I think they could stay, because they'll not happen under
normal circumstances by then.

--
Vojtech Pavlik
SuSE Labs, SuSE CR