APIC error 80(80) and 00(80)

APIC error 80(80) and 00(80)

[Timur Alperovich]

Hi,

I apologize if others have asked this question before, but I'm a bit
confused by the APIC error messages, i.e. I couldn't find what they
mean. After inserting a kernel module I wrote and trying to read the
device it created, I see the following in dmesg:

Losing some ticks... checking if CPU frequency changed.
warning: many lost ticks.
Your time source seems to be instable or some driver is hogging interupts
rip tg3_read32+0x9/0xa [tg3]
Falling back to HPET
APIC error on CPU3: 00(80)
APIC error on CPU3: 80(80)

where the last message was repeated 1600 times. Could anyone tell me
what it means or of a better place to ask about it?

Thank you,
Timur

Re: APIC error 80(80) and 00(80)

[Maciej W. Rozycki]

On Wed, 30 Apr 2008, Timur Alperovich wrote:

> I apologize if others have asked this question before, but I'm a bit
> confused by the APIC error messages, i.e. I couldn't find what they

 It sounds strange -- the errors are explained in the source next to the 
message you are seeing.

> mean. After inserting a kernel module I wrote and trying to read the
> device it created, I see the following in dmesg:
> 
> Losing some ticks... checking if CPU frequency changed.
> warning: many lost ticks.
> Your time source seems to be instable or some driver is hogging interupts
> rip tg3_read32+0x9/0xa [tg3]
> Falling back to HPET
> APIC error on CPU3: 00(80)
> APIC error on CPU3: 80(80)
> 
> where the last message was repeated 1600 times. Could anyone tell me
> what it means or of a better place to ask about it?

 This is "Illegal register address", which means some piece of code tried
to access an inexistent APIC register.  If you want to find out where the
code is, then trigger an oops from where the message is printed.  Of
course it looks like the tip of an iceberg anyway...

  Maciej

Re: APIC error 80(80) and 00(80)

[Robert Hancock]

Timur Alperovich wrote:
> Hi,
> 
> I apologize if others have asked this question before, but I'm a bit
> confused by the APIC error messages, i.e. I couldn't find what they
> mean. After inserting a kernel module I wrote and trying to read the
> device it created, I see the following in dmesg:
> 
> Losing some ticks... checking if CPU frequency changed.
> warning: many lost ticks.
> Your time source seems to be instable or some driver is hogging interupts
> rip tg3_read32+0x9/0xa [tg3]
> Falling back to HPET
> APIC error on CPU3: 00(80)
> APIC error on CPU3: 80(80)
> 
> where the last message was repeated 1600 times. Could anyone tell me
> what it means or of a better place to ask about it?

Not sure exactly what that means, but sounds like something really wonky 
happened.. I'd suspect your code somehow got stuck in a loop with 
interrupts off on one CPU, or corrupted memory, or something.