KVM Switch, SuSE Linux

[spoddynerd]

I have:
- An AMD Athlon XP 1500 running SuSE Linux 9.1
- a P4 running WinXP
- A Belkin 4-port KVM switch

Both pcs are connected to the KVM switch.

Everything is fine when I boot the AMD/linux box with the KVM switch on the correct channel and don't switch channels. But if I subsequently switch to the P4/WinXP and then back again to the AMD/Linux, the AMD/Linux ceases to respond correctly - very jumpy, no response to clicks - to the mouse (Wireless Microsoft Intellimouse). There is no such problem with the P4/WinXP following switching.

Could this be an IRQ conflict problem? How can I find out for sure? (I am a newbie to Linux - please respond with this in mind).

Additional info: when the AMD athlon was used as an XP box, there were intermittent problems such as no response to the mouse (solved by rebooting), occasional failure to recognise the network card, USB ports extremely unreliable, etc.

 

[edfair]

KVM switches are working at a hardware level that is more basic than the OS and the connections are supposedly emulated when switched so there is no indication that there has been a disconnection.
I've replaced several Belkins for this type of problem and the problem has moved with them when tested on other systems. When they fail in a test environment I throw them away.
I have 2 Cybex 4 port boxes and they seem to be power sensitive. When they have the identical failure I can bring up another box to supply additional 5v and things start working again. Don't know if that solution might be an indication of the same type of problem on the Belkins. It hasn't been in my case, but might be in yours.

Ed Fair
Give the wrong symptoms, get the wrong solutions.

 

[spoddynerd]

When the AMD was used as an XP box instead of a Linux box, on the same port of the same KVM switch, the switching was o.k.
Why are the results different between Linux and XP if the the hardware is the same and the switch operates at the hardware level?

 

[edfair]

Have never traced a Linux problem like this. Can only report that SCO Unix stuff doesn't have a problem. Have about 10 systems happy with them.
Could you throw another box on the third port? Just to supply more power for long enough to test and see if more power might resolve the issue.

Ed Fair
Give the wrong symptoms, get the wrong solutions.