Re: another Pentium bug
H. Peter Anvin (hpa@transmeta.com)
11 Nov 1997 11:20:41 GMT
Followup to: <Pine.LNX.3.95.971111052324.2924D-100000@vuser.vu.union.edu>
By author: Rob Hagopian <hagopiar@vuser.vu.union.edu>
In newsgroup: linux.dev.kernel
>
> As one learns in computer theory, this is totally useless. Just as there's
> no way to test for an infinite loop.
> -Rob H.
>
> On Mon, 10 Nov 1997, Just this guy you know wrote:
>
> > The only real way fully test a program to see if the opcode is ever called
> > would be to grovel through the binary, executing it as you go and checking
> > to make sure no illegal opcodes are generated. I suppose that to
> > eloquently do this, you could write a 386 emulator that would
> > automatically trip an exception at the illegal opcodes in question.
>
Well, you *could* do it by simulating the entire program. No, you
don't want to do this, although simulating a functional x86 on a
broken x86 is probably one of the more efficient kinds of emulation,
it still stinks.
-hpa
--
PGP: 2047/2A960705 BA 03 D3 2C 14 A8 A8 BD 1E DF FE 69 EE 35 BD 74
See http://www.zytor.com/~hpa/ for web page and full PGP public key
I am Bahá'í -- ask me about it or see http://www.bahai.org/
"To love another person is to see the face of God." -- Les Misérables