Re: [lamps] [EXTERNAL] Re: LAMPS Virtual Interim in Sept. 2022

[Ilari Liusvaara <[e-mail]>]

On Fri, Sep 16, 2022 at 04:58:22PM +0000, Blumenthal, Uri - 0553 - MITLL wrote:
> Mike Ounsworth wrote:
> > Could I get a slot at the LAMPS interim to discuss the hash-then-sign
> > issue for Dilithium and Falcon?
> 
> I think this issue is worth discussing.
> 
> > Issue summary:
> >
> > - Needing to stream your entire message to your crypto module is dumb
> > (think streaming an entire firmware image to your network HSM for
> > code-signing, or to your TPM for secure boot validation; yuck).
> 
> I'm not sure I agree here - if you sign a hash of something, you can't be
> sure it's a hash of what you wanted to sign.

Then there is the problem is that carelessly done pre-hashing can lead
to signature standing for two things, which is considered to break the
signature algorithm.

Pre-hashed variant of Ed25519 does very nasty hacks prevent that
attack.


Related question: How common it is that HSM itself checks format of
the data it is signing? Without such checks, it is completely open to
signing whatever.


> > - Both Dilithium and Falcon have, as their first internal step' a hash
> > of the message prepended with a nonce (the pubkey for Dilithium, and a
> > random r for Falcon), I assume in order to block pre-computed collision
> > attacks.
> > - If you, for example, do SHA256(m) before calling Dilithium.sign(),
> > then you have re-introduced that collision attack.
> 
> This assumes one can find collisions in the hash function used. For
> SHA2 and SHA3 it's a tall assumption. AFAIK, the main reason SHA3 hasn't
> superseded SHA2 at this point (despite that SHA3 is cryptographically 
> nicer and better) is that SHA2 proved to be "strong enough".

I am aware of at least one place where anti-collision countermeasures
are mandated for SHA-2.


> > - You can externalize that first hashing step of the Dilithium /
> > Falcon sign / verify algs outside of the crypto module without
> > breaking interop, but doing so will need to be mentioned in the
> > standards, and will need security review.
> 
> Offhand, I doubt this can be secure...

I think it might be safe enough for Dilithium. However, for Falcon, 
there is serious issue that one of the hash inputs must be generated by
the crypto module. And then it is not possible at all for SPHINCS+.



-Ilari