Re: getting around Ken Thompson's compiler Trojan

From: Alan J Rosenthal (flaps@dgp.toronto.edu)
Date: 01/28/03

  • Next message: Jim: "A "send email" utility?" 
    From: flaps@dgp.toronto.edu (Alan J Rosenthal)
Date: 28 Jan 2003 17:11:42 GMT

    Richard Caley <MYFIRSTNAME@MYLASTNAME.org.uk> writes:
    >Barry Margolin writes:
    >> BTW, not only does the compiler have to recognize itself, but it also has
    >> to recognize the login program, so it can reinsert the backdoor.
    >
    >The little nerd who lives in my head is jumping up and down and making
    >me post the fact that recognising the compiler is almost certainly an
    >undecidable problem.

    Oh, of course it is. Ken Thompson never claimed to have solved it.
    The claimed trojan horse in the compiler was highly that-compiler-specific.

    Actually, I'm not sure I should agree that "recognizing the compiler is an
    undecidable problem", because I'm not sure it's a well-defined problem,
    because "the compiler" is not well-defined. So it may be even worse than the
    little nerd who lives in your head noticed. 

    -- 
very frequently asked questions at
ftp://rtfm.mit.edu/pub/faqs/computer-security/most-common-qs