[gradient]

Since ungrading to Leopard, I've now had three kernel panics related to switching to or from my G5 using a Belkin Flip USB kvm switch.....

Unresolved kernel trap(cpu 0): 0x400 - Inst access DAR=0x0000000052286564 PC=0x0000000000000000
Latest crash info for cpu 0:
Exception state (sv=0x60c37000)
PC=0x00000000; MSR=0x40009030; DAR=0x52286564; DSISR=0x40000000; LR=0x0091F948; R1=0x60B5BC20; XCP=0x00000010 (0x400 - Inst access)
Backtrace:
0xDF725C0C
backtrace terminated - unaligned frame address: 0xC44B13E3

Kernel loadable modules in backtrace (with dependencies):
com.apple.iokit.IOHIDFamily(1.5)@0x8f1000->0x93afff
Proceeding back via exception chain:
Exception state (sv=0x60c37000)
previously dumped as "Latest" state. skipping...
Exception state (sv=0x69f95500)
PC=0x00000000; MSR=0x0000D030; DAR=0x00000000; DSISR=0x00000000; LR=0x00000000; R1=0x00000000; XCP=0x00000000 (Unknown)

BSD process name corresponding to current thread: Unknown

Mac OS version:
9B18

Kernel version:
Darwin Kernel Version 9.1.0: Wed Oct 31 17:48:21 PDT 2007; root:xnu-1228.0.2~1/RELEASE_PPC
panic(cpu 0 caller 0xFFFF0004): 0x400 - Inst access
Latest stack backtrace for cpu 0:
Backtrace:
0x63094F73
backtrace terminated - frame not mapped or invalid: 0xD8ECEB10

Proceeding back via exception chain:
Exception state (sv=0x60c37000)
PC=0x00000000; MSR=0x40009030; DAR=0x52286564; DSISR=0x40000000; LR=0x0091F948; R1=0x60B5BC20; XCP=0x00000010 (0x400 - Inst access)
Backtrace:
0xDF725C0C
backtrace terminated - unaligned frame address: 0xC44B13E3

Kernel loadable modules in backtrace (with dependencies):
com.apple.iokit.IOHIDFamily(1.5)@0x8f1000->0x93afff
Exception state (sv=0x69f95500)
PC=0x00000000; MSR=0x0000D030; DAR=0x00000000; DSISR=0x00000000; LR=0x00000000; R1=0x00000000; XCP=0x00000000 (Unknown)

BSD process name corresponding to current thread: Unknown
KZQUU

I did post about the first one in this thread. The kp reoccurred yesterday under 10.5.1.

I'm considering just going to buy a different KVM switch (seeing that it isn't fully compatible with my XB360 either, which is the other unit attached to it) but have no idea if others are or have experienced that same issue with switches by other manufacturers. I'd hate to end up with the same issue after spending the $$.

Thanks.

[pgl]

After installing Leopard I have had at least 6 similar random kernel panic events on switching back and forth between a Mini g4 and a Windows pc. Under Tiger never had any problem.

[MallyMal]

Hmm, I seem to be having a similar problem under 10.4.11. At least that is what I think because I haven't had a KP since I disconnected from my Belkin KVM. My Thread I'll keep testing to be sure it's the KVM in my case.

[rdenton]

I also had the kernel panic problem with Leopard when using a KVM switch between Mac and XP boxes. Tried the suggestion I had seen in this or another forum about putting a powered hub between the KVM and the Mac to act as a buffer to prevent USB spikes. Anyway, it is now four days and no more panics. I'm actually using the cheap Zonet KVM switch and a generic no-name powered USB hub. This is great as now I can continue to use my Kensington Wireless desktop for MAc on both Mac and Win boxes.

Equipment: PowerMac Dual G4 Quicksilver, 10.5.1, USB 2.0 PCI card (already in use and working fine with Tiger), Zonet KVM 3122, Kensington Wireless desktop for Mac KB, Logitech optical wired mouse (Kensington wireless mouse lags too much). Generic powered 7-port usb hub ($19.99 at local PC shop).
This may not apply with the Belkin, but it seemed to me like a good $20 invested to get back into full production without having to juggle two KB/mice. YMMV.

[Chongo]

Meet Colonel Panic

[ghporter]

That's Martha Rae and you know it! Quite the panic in her day, but I digress...

I would be very interested in determining conclusively whether it is the KVM causing the problem, the keyboard/mouse combo (wireless stuff connected to a KVM sounds like asking for some sort of "stacked tolerance" problem to me), or something completely separate. I have heard of this sort of thing, but it had more to do with a very off-brand KVM that was on its last legs and didn't lock the signals at the "idle" computer end so that computer got "odd" pretty fast.