2004-10-06 05:36:44

by Anthony DiSante

[permalink] [raw]
Subject: KVM -> jumping mouse... still no solution?

Hello,

I first got a KVM switch around the time of kernel 2.2.something, and when
using it to switch to a Linux system, the mouse "freaks out." It's fine if
you don't move it, but if you move it N/E/NE it's really slow and jerky, and
if you move it S/W/SW even a hair, it slams down to the SW corner of the
screen and acts like you hit all the mouse's buttons 50 times simultaneously.

When switching to an MS Windows system (any version from 98 on up; haven't
tried anything earlier) the mouse works fine, it just pauses for maybe a
second at first, during which I assume it's doing some kind of PS/2 reset.

It used to be that switching out of X-windows with Ctrl-Alt-F[1-6] and then
back to VT7 would reset the mouse, but that hasn't worked in about a year
for me. I was also able to run a little script to send a few specific chars
to the mouse device that seemed to reset it... that too no longer works.
The only thing that works now is unplugging the mouse from the KVM and then
back in.

The other day I came across this (kerneltrap.org/node/view/2199): "Use
psmouse.proto=bare on the kernel command line, or proto=bare on the
psmouse module command line." But that makes the mouse's scroll-wheel not
work. (And this problem doesn't exist with some of the mouse drivers, but
it does with IMPS/2, which is the only one I've ever been able to get the
scroll wheel working with.)

Is there really no solution to this problem? If Microsoft can figure it
out, I'm sure someone in the Linux community can... not that I'm
volunteering, of course...

-Anthony DiSante
http://nodivisions.com/


2004-10-06 05:50:14

by Jesper Juhl

[permalink] [raw]
Subject: Re: KVM -> jumping mouse... still no solution?

On Wed, 6 Oct 2004, Anthony DiSante wrote:

> you don't move it, but if you move it N/E/NE it's really slow and jerky, and
> if you move it S/W/SW even a hair, it slams down to the SW corner of the
> screen and acts like you hit all the mouse's buttons 50 times simultaneously.
>
I've had similar problems with my mouse and KVM switch.

> The other day I came across this (kerneltrap.org/node/view/2199): "Use
> psmouse.proto=bare on the kernel command line, or proto=bare on the
> psmouse module command line." But that makes the mouse's scroll-wheel not
> work. (And this problem doesn't exist with some of the mouse drivers, but it
> does with IMPS/2, which is the only one I've ever been able to get the scroll
> wheel working with.)
>
psmouse.proto=imps solves the problem for me (wheel works as well).
The funny thing is that I don't need to do anything like this when I boot
a 2.4 kernel, only 2.6 kernels show this behaviour on my system.???

--
Jesper Juhl

2004-10-06 06:15:11

by Anthony DiSante

[permalink] [raw]
Subject: Re: KVM -> jumping mouse... still no solution?

Jesper Juhl wrote:
>>you don't move it, but if you move it N/E/NE it's really slow and jerky, and
>>if you move it S/W/SW even a hair, it slams down to the SW corner of the
>>screen and acts like you hit all the mouse's buttons 50 times simultaneously.
>>
>
> I've had similar problems with my mouse and KVM switch.
>
>
>>The other day I came across this (kerneltrap.org/node/view/2199): "Use
>>psmouse.proto=bare on the kernel command line, or proto=bare on the
>>psmouse module command line." But that makes the mouse's scroll-wheel not
>>work. (And this problem doesn't exist with some of the mouse drivers, but it
>>does with IMPS/2, which is the only one I've ever been able to get the scroll
>>wheel working with.)
>>
>
> psmouse.proto=imps solves the problem for me (wheel works as well).
> The funny thing is that I don't need to do anything like this when I boot
> a 2.4 kernel, only 2.6 kernels show this behaviour on my system.???

That doesn't make any difference on my system. Mouse freakout is just the same.

-Anthony DiSante
http://nodivisions.com/

2004-10-06 07:24:24

by Rajendra

[permalink] [raw]
Subject: Re: KVM -> jumping mouse... still no solution?

One quick solution I know of is to restart the gpm daemon,
(/etc/init.d/gpm restart) that resets the mouse settings.
But this is not the correct way, there should be some way
where the driver automatically detects and resets the mouse.

bye,
-rpm
On Wed, 2004-10-06 at 11:06, Anthony DiSante wrote:
> Hello,
>
> I first got a KVM switch around the time of kernel 2.2.something, and when
> using it to switch to a Linux system, the mouse "freaks out." It's fine if
> you don't move it, but if you move it N/E/NE it's really slow and jerky, and
> if you move it S/W/SW even a hair, it slams down to the SW corner of the
> screen and acts like you hit all the mouse's buttons 50 times simultaneously.
>
> When switching to an MS Windows system (any version from 98 on up; haven't
> tried anything earlier) the mouse works fine, it just pauses for maybe a
> second at first, during which I assume it's doing some kind of PS/2 reset.
>
> It used to be that switching out of X-windows with Ctrl-Alt-F[1-6] and then
> back to VT7 would reset the mouse, but that hasn't worked in about a year
> for me. I was also able to run a little script to send a few specific chars
> to the mouse device that seemed to reset it... that too no longer works.
> The only thing that works now is unplugging the mouse from the KVM and then
> back in.
>
> The other day I came across this (kerneltrap.org/node/view/2199): "Use
> psmouse.proto=bare on the kernel command line, or proto=bare on the
> psmouse module command line." But that makes the mouse's scroll-wheel not
> work. (And this problem doesn't exist with some of the mouse drivers, but
> it does with IMPS/2, which is the only one I've ever been able to get the
> scroll wheel working with.)
>
> Is there really no solution to this problem? If Microsoft can figure it
> out, I'm sure someone in the Linux community can... not that I'm
> volunteering, of course...
>
> -Anthony DiSante
> http://nodivisions.com/
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/

2004-10-06 10:10:12

by Anthony DiSante

[permalink] [raw]
Subject: Re: KVM -> jumping mouse... still no solution?

Rajendra P Mishra wrote:
>>I first got a KVM switch around the time of kernel 2.2.something, and when
>>using it to switch to a Linux system, the mouse "freaks out." It's fine if
>>you don't move it, but if you move it N/E/NE it's really slow and jerky, and
>>if you move it S/W/SW even a hair, it slams down to the SW corner of the
>>screen and acts like you hit all the mouse's buttons 50 times simultaneously.
>>
>>When switching to an MS Windows system (any version from 98 on up; haven't
>>tried anything earlier) the mouse works fine, it just pauses for maybe a
>>second at first, during which I assume it's doing some kind of PS/2 reset.
>>
>>It used to be that switching out of X-windows with Ctrl-Alt-F[1-6] and then
>>back to VT7 would reset the mouse, but that hasn't worked in about a year
>>for me. I was also able to run a little script to send a few specific chars
>>to the mouse device that seemed to reset it... that too no longer works.
>>The only thing that works now is unplugging the mouse from the KVM and then
>>back in.
>>
>>The other day I came across this (kerneltrap.org/node/view/2199): "Use
>>psmouse.proto=bare on the kernel command line, or proto=bare on the
>>psmouse module command line." But that makes the mouse's scroll-wheel not
>>work. (And this problem doesn't exist with some of the mouse drivers, but
>>it does with IMPS/2, which is the only one I've ever been able to get the
>>scroll wheel working with.)
>>
>>Is there really no solution to this problem? If Microsoft can figure it
>>out, I'm sure someone in the Linux community can... not that I'm
>>volunteering, of course...
>
> One quick solution I know of is to restart the gpm daemon,
> (/etc/init.d/gpm restart) that resets the mouse settings.
> But this is not the correct way, there should be some way
> where the driver automatically detects and resets the mouse.
>

That doesn't work for me either. Hmm... maybe it's because I'm using a
Microsoft mouse... only plays nice with Windows systems?

-Anthony DiSante
http://nodivisions.com/

2004-10-06 16:26:15

by Dmitry Torokhov

[permalink] [raw]
Subject: Re: KVM -> jumping mouse... still no solution?

On Wednesday 06 October 2004 05:09 am, Anthony DiSante wrote:
> ?> One quick solution I know of is to restart the gpm daemon,
> ?> (/etc/init.d/gpm ?restart) that resets the mouse settings.
> ?> But this is not the correct way, there should be some way
> ?> where the driver automatically detects and resets the mouse.
> ?>
>
> That doesn't work for me either. ?Hmm... maybe it's because I'm using a
> Microsoft mouse... only plays nice with Windows systems?
>

In 2.6 GPM does not have access to the hardware and therefore cannot reset
it. A temporary fix can be downloaded from here:

http://bugme.osdl.org/attachment.cgi?id=3244&action=view

It will for only on pre 2.6.9-rc2 kernels. When using kernels 2.6.9-rc3+
there is no automatic re-synchronization [yet] but one can restore a mouse
by issuing the following command:

echo -n "reconnect" > /sys/bus/serio/devices/serioX/driver

To find out which 'serioX' your mouse connected to just look into their
respective "driver" attributes - the one with "psmouse" is the one you
need.

--
Dmitry

2004-10-06 22:11:40

by Peter Williams

[permalink] [raw]
Subject: Re: KVM -> jumping mouse... still no solution?

Anthony DiSante wrote:
> Rajendra P Mishra wrote:
>
>>> I first got a KVM switch around the time of kernel 2.2.something, and
>>> when using it to switch to a Linux system, the mouse "freaks out."
>>> It's fine if you don't move it, but if you move it N/E/NE it's really
>>> slow and jerky, and if you move it S/W/SW even a hair, it slams down
>>> to the SW corner of the screen and acts like you hit all the mouse's
>>> buttons 50 times simultaneously.
>>>
>>> When switching to an MS Windows system (any version from 98 on up;
>>> haven't tried anything earlier) the mouse works fine, it just pauses
>>> for maybe a second at first, during which I assume it's doing some
>>> kind of PS/2 reset.
>>>
>>> It used to be that switching out of X-windows with Ctrl-Alt-F[1-6]
>>> and then back to VT7 would reset the mouse, but that hasn't worked in
>>> about a year for me. I was also able to run a little script to send
>>> a few specific chars to the mouse device that seemed to reset it...
>>> that too no longer works. The only thing that works now is unplugging
>>> the mouse from the KVM and then back in.
>>>
>>> The other day I came across this (kerneltrap.org/node/view/2199):
>>> "Use psmouse.proto=bare on the kernel command line, or proto=bare on the
>>> psmouse module command line." But that makes the mouse's
>>> scroll-wheel not work. (And this problem doesn't exist with some of
>>> the mouse drivers, but it does with IMPS/2, which is the only one
>>> I've ever been able to get the scroll wheel working with.)
>>>
>>> Is there really no solution to this problem? If Microsoft can figure
>>> it out, I'm sure someone in the Linux community can... not that I'm
>>> volunteering, of course...
>
> >
> > One quick solution I know of is to restart the gpm daemon,
> > (/etc/init.d/gpm restart) that resets the mouse settings.
> > But this is not the correct way, there should be some way
> > where the driver automatically detects and resets the mouse.
> >
>
> That doesn't work for me either. Hmm... maybe it's because I'm using a
> Microsoft mouse... only plays nice with Windows systems?

I'm having similar problems to this only with the keyboard not the
mouse. It started happening in 2.6.8 and 2.6.9. I've recently noticed
that the problems only seem to occur after I've used the KVM switch to
connect to a Windows XP. After I switch back to Linux I get spurious
characters inserted into the input stream and random deletions.
Rebooting cures the problem.

Now that I've detected the association with using the KVM to connect to
a Windows box I've started using a terminal server client to connect to
the Windows box when possible.

Peter
--
Peter Williams [email protected]

"Learning, n. The kind of ignorance distinguishing the studious."
-- Ambrose Bierce