[clug] Anti-Virus Software

[steve jenkin]

Sam Couter wrote on 22/06/10 8:50 PM:
> Ivan Miljenovic <ivan.miljenovic at gmail.com> wrote:
>> Also, people at NICTA have proven an L4 kernel correct and secure (not
>> used anywhere yet though).
> 
> I think they also assumed the compiler is generating correct code. And
> the compiler is software, which as we know will always contain bugs...

Ken Thompson gave his 1984 Turing Award address on exactly this topic.
STW "reflections on trusting trust".

The key point is, if you didn't write the compiler (and linker/loader?)
yourself, using it is an act of faith.

Thompson had built a version of the C compiler that recognised and
inserted code into 2 programs: login and itself.

'login' patch was a secret user (ken) that was ID 0.
The patch to the compiler was the login patch plus its own patch.

He compiled the source with the C-code in there.
Then removed the patches from the source and recompiled with the patched
compiler.

Voila!
Clean source, but a permanent backdoor in descendants of that executable.

Inspect etc as much as you like, but you'll never find it.

Scary if you think about the complexity of chips, CPU's and hidden
electronics in todays' compute devices.

Apple iPhone could be fundamentally broken, and not even Apple would know.

-- 
Steve Jenkin, Info Tech, Systems and Design Specialist.
0412 786 915 (+61 412 786 915)
PO Box 48, Kippax ACT 2615, AUSTRALIA

sjenkin at canb.auug.org.au http://members.tip.net.au/~sjenkin