References for recovering the y coordinate in a montgomery ladder
Phillip Hallam-Baker <[email protected]> Fri, 8 Nov 2019 23:38:19 -0500
| Newsgroups | gmane.comp.encryption.general |
|---|---|
| Message-ID | <CAMm+LwjhXyFOxh4jeUmVasUe2K3_h5AqKccscCngzDZx7bhg9g@mail.gmail.com> |
I have a problem, I am trying to add two RFC7748 aka CurveX25518 points and it is only working half the time. The problem here is that the montgomery ladder only uses the X point, the y point is ignored. And that means that the sign on the y point is also lost. Since the curve is symmetric about the x axis (y^2 = ...), guessing y is positive (or even) and drawing a line between the two points is going to give the correct x if both guesses are right or both are wrong. If one is right and the other is wrong... oops. I have found some discussion on the net on recovering the y coordinate from the Montgomery ladder. But it is kinda icky to refer to random blog posts in a spec. And not being a number theorist, it is going to take me quite a while to get familiar with the notation in the original papers. Can anyone point me to a crib? Preferably with like working code in Perl or C# :-) One of the accumulators is sU. The other seems to be (s+1)U. Since all I need is the sign bit, I could try suck it and see.... but seems icky. _______________________________________________ The cryptography mailing list [email protected] https://www.metzdowd.com/mailman/listinfo/cryptography