Re: Pentium bug

linux kernel account (linker@nightshade.z.ml.org)
Sun, 9 Nov 1997 00:45:24 -0500 (EST)


I thought of this too..

First youd have to toy with the apic so that somehow the working cpu would
eventually get a interupt (RTC trick maby?). (MABY NOT POSSIBLE)

Then it would have to figure out that the other cpu isn't working. (LIKELY
EASY)

Then it would have to reboot and reinitlize the other cpu. (VERY HARD)

It's also entirely possible that killing a cpu might require a harder
reboot then is possible at this level.. It's also possible that a CPU
disapearing might confuse the hell out of the MB..

Further more, an attacker could fork off two 'evil processes'.. Enough
tries and he could get both cpus at once.. (Unless you also patched the
kernel to make all user apps run on one cpu, so that the attacker couldn't
crash both)

This whole hack if possible wouldn't really be worth the effort IMHO (but
if someone wants to try, be my guest)... If you want the problem fixed
then join a mailing list commited to harassing Intel.

On Sun, 9 Nov 1997, Adam Wiggins wrote:

> Hey ppl i have two systems effected by this bug, but it occured to
> me that since both are dual's (one has yet to get its 2nd cpu) only one
> cpu can be effected at once.
> First question, is it possible for the 2nd cpu to detect the other
> has been shutdown?
> Second question can someone write a patch to detect and handle
> this? Maybe setting some kernel flag or something so a system program can
> take mesures like ejecting all users and logging the event?
> Hows this sound? I'd be interested in feedback on this issue
>
> Cheers Adam
>