[plug] Machine Check Exception

[Bernard Blackham]

On Sun, Dec 26, 2004 at 12:16:21AM +0800, Shayne O'Neill wrote:
> I'm still wondering what the elf bin load just before crash is. Unless
> that dump aint meant to be read linear.

As linearly as most kernel panics (first to last, and bottom up :)

I'm guessing what happened is that it tried to exec an executable,
but whilst removing the pages of the previous executable (as happens
when execve() is called), a certain region of memory triggered an
MCE which printed the first "CPU 0: Machine Check Exception: ..."
message.

The call trace then occurs when the MCE triggers a panic and sync()
is called, even though some earlier function call asked not to be
interrupted, but there's not much choice in the matter at that
stage.

... I think - take this all with a pinch of salt. I've read a fair
few kernel oopses in my time but I can't claim to be an expert in
kernel scheduling and VM handling, yet :P

Bernard.

-- 
 Bernard Blackham <bernard at blackham dot com dot au>