PGP source code

Ben Laurie ben at algroup.co.uk
Tue, 04 Sep 2001 19:48:10 +0100


M Taylor wrote:
> 
> On Tue, 4 Sep 2001, Richard Clayton wrote:
> 
> > In 1997 there was a great deal of interest in an announcement by Intel
> > of "programmable microcode" such that 2K chunks of microcode could be
> > replaced at POST time (when you first switch the machine on) so as to
> > correct bugs (such as the "floating point problem" that caused them so
> > much PR trouble).
> >
> > There's a little more in this October 2000 BYTE article:
> >
> >         http://www.byte.com/documents/s=479/BYT20001016S0006/
> >
> > Although BYTE talks about "checksums", the microcode is signed (I am
> > told on excellent authority [though admittedly, my memory may not be as
> > good as the authority!]) by a 2048 bit RSA key...
> >
> > There is indeed, as BYTE says, very little documented about this on
> > Intel's site (and certainly no mention of RSA or of the way that the
> > chip checks the signature), though there is occasional mention of
> > microcode upgrades for chips such as the Xeon.
> 
> <http://developer.intel.com/design/pentium4/manuals/24547204.pdf>
> Section 8.11.3 (pg 309)
> 
> Based on my reading of the details given, it is not talking about a RSA
> style of signature. I might be wrong though, it is lower level then I am
> use to dealing with.

It does not specify the signature at all, AFAICS.

Cheers,

Ben.

--
http://www.apache-ssl.org/ben.html

"There is no limit to what a man can do or how far he can go if he
doesn't mind who gets the credit." - Robert Woodruff