Re: Pentium ALIVE patch testing

Linus Torvalds (torvalds@transmeta.com)
14 Nov 1997 05:16:18 GMT


In article <Pine.LNX.3.95.971113173745.2302A-200000@chaos.analogic.com>,
Richard B. Johnson <root@chaos.analogic.com> wrote:
>
>
>Testing with the included program sometime gives:
>
>Nov 13 17:32:16 chaos kernel: Dazed and confused, but trying to continue
>Nov 13 17:32:16 chaos kernel: Do you have a strange power saving mode enabled?
>Nov 13 17:32:16 chaos kernel: Uhhuh. NMI received for unknown reason 30.
> [ .... ]

This is one of the buglets in the current code - this is what happens
when you do a "int $0x2" - the code thinks it's a valid (external) NMI.
Without the fix it would have generated a #GP fault (which linux turns
into a segmentation fault), because the IDT entry for #2 would have been
marked as privileged - while the new page fault handler has no
reasonable way of finding out what's up...

Perfecting the patch is going to take some time - the current one
essentially just fixes the lock-up.

Linus