[plug] Machine Check Exception

[bob]

On Sun, 26 Dec 2004 12:11 am, Bernard Blackham wrote:
> On Sat, Dec 25, 2004 at 10:44:50PM +0800, bob wrote:
> > Got a panic this evening and I was hoping someone coud have a look at
> > the following syslog dump and perhaps tell me what happened. Is this
> > RAM failing? (once again google has failed to provide an anwer by
> > flooding the response with blog entries - none of which are worth
> > anything :()
>
> Looks like bad RAM or a cosmic ray to me. It'd probably be worth
> running memtest over it. If you do find you have bad RAM and you're
> too environmentally-conscious to just replace it, you could try the
> badram patch which tells the kernel to avoid certain bits of memory.

Yes, though that might be the case. Oh well another reboot to run memtest 
and see how that goes... otherwise, cosmics as you say.

To tell the truth the box in question is probably just sulking. I'm in the 
midst of configing up a replacement and it may have decided to go on 
strike. :)

> > Dec 25 21:40:03 fluff kernel: CPU 0: Machine Check Exception:
> > 0000000000000004
> > Dec 25 21:40:03 fluff kernel: Bank 2: b600000000140000 at
> > 003446ff003446ff
>
> Problem #1
>
> > Dec 25 21:40:03 fluff kernel: Kernel panic: CPU context corrupt
>
> This and the rest would follow from problem #1. (Well rather, the
> rest is probably indeed a kernel bug, but in an odd failure case
> like this, it's not too surprising it hasn't been caught yet).

I'm prepared to forgive a lot of coding errors in the face of hardware 
failure :).

> Bernard.

-- 
... [concerning quotation marks] even if we *did* quote anybody in this
business, it probably would be gibberish.
		-- Thom McLeod