Re: NMI

Linus Torvalds (torvalds@transmeta.com)
Fri, 14 Nov 1997 08:56:59 -0800 (PST)

On Fri, 14 Nov 1997, Richard B. Johnson wrote:
> Execution of:
> unsigned char main[]={0xCD,0x02,0x00};
> (INT 2)
> ...produces a race condition where the trap executes thousands of times!

..but it should still be interruptible, no?

Essentially, what happens is that it ends up doing the NMI over and over
again until you kill it.

Right now I can't think of a good way to avoid this, except to actually
disassemble the instruction that we hit on when taking the f00f bug. We
probably need to do that anyway to handle the "int3" (one byte) and the
"int 0x3" (two bytes) instructions correctly to make debuggers happier..

Anybody want to tackle these issues?

Linus