Re: [Cfrg] CFRG Curves document fixes

I dunno how I managed to typo this, since I was doing it in SAGE. Do you 
get

07fff4181ac6cc95ec1c16a94a0f74d12da232ce40a77552281d282b
b60c0b56fd2464c335543936521c24403085d59a449a5037514a879d

Cheers,
-- Mike

On 05/11/2015 10:14 PM, Ilari Liusvaara wrote:
> On Mon, May 11, 2015 at 05:37:01PM -0700, Thomas DuBuisson wrote:
>> All,
>>
>> I have produced an executable (Cryptol) specification of the Turner
>> draft and now the cfrg-curves document.  In doing so I believe I have
>> found a bug; the division optimization is only valid for 25519, if I
>> understand correctly.  The offending text:
>>
>> ```
>> Return x_2 * (z_2^(p - 2))
>> ```
>>
>> appears as field division in my reference implementation.
> I think the formula above is correct for all GF(p) (except for producing
> 0 for 0, which is actually desirable to handle z=0 right).
>
>> As a matter of full disclosure: my specification "passes the tests" in
>> that the Curve25519 passes 7.1 and 8.X while 448 passes section 7.1
>> vectors but not 8.X (ECDH).  I will investigate the failure when time
>> allows, though if someone wishes to confirm this KAT is correct I
>> certainly would not object.
> I tried running through KAT with my own implementation:
> - [PASS] KAT Curve25519 DHF #1
> - [PASS] KAT Curve25519 DHF #2
> - [PASS] KAT Curve25519 ECDH Alice public
> - [PASS] KAT Curve25519 ECDH Bob public
> - [PASS]     Curve25519 ECDH consistancy[1]
> - [PASS] KAT Curve25519 ECDH Shared secret
> - [PASS] KAT Curve448 DHF #1
> - [PASS] KAT Curve448 DHF #2
> - [PASS] KAT Curve448 ECDH Alice public
> - [PASS] KAT Curve448 ECDH Bob public
> - [PASS]     Curve448 ECDH consistancy[1]
> - [FAIL] KAT Curve448 ECDH Shared secret
>
> Was that KAT Curve448 ECDH Shared secret the KAT you saw failing?
> Any others failed for you?
>
>
> [1] Where ECDH consistency means the consistency conditions:
> - DHF(a, G) = DHF_base(a)
> - DHF(b, G) = DHF_base(b)
> - DHF(a, B) = DHF(b, A)
>
> (My implementation does have different faster algorithm for standard
> basepoint).
>
>
> -Ilari
>
> _______________________________________________
> Cfrg mailing list
> Cfrg@irtf.org
> http://www.irtf.org/mailman/listinfo/cfrg