Timeline for Questions about the Curve25519-donna implementation
Current License: CC BY-SA 4.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 11, 2018 at 10:04 | comment | added | Vega4 | zmone is multiplied by x in a call to fmul(z, x, zmone) which results in the 'real' X coordinate to be put into the 'z' variable.(right?) with call to fcontract we convert the 'polynomial form' back to little-endian. Thank you dave_thompson_085. | |
| Aug 11, 2018 at 10:04 | comment | added | Vega4 | in the code above; cmult returns x and z as separate variables. isn't x already the public key? I understand that we use projective coordinates instead of affine coordinates to cut on the number of divisions. is that correct? what does crecip do? (it takes z and what's the result in zmone?) - the most misterious thing to me right now. The fmul function just seems to multiply two number which are in the 'polynomial' form. Actually from what I see seems like the 'real' Z-coordinate is in zmone; but then what's the format of z returned by cmult? how do these two differ? | |
| Aug 11, 2018 at 4:06 | comment | added | dave_thompson_085 | @Vega4: some crypto, particularly asymmetric crypto like RSA DSA DH and as here ECC, uses numbers larger than a computer can handle in one unit, which are usually called bignums; crypto is an important use of bignums but not the only one. 'Limb' is used for bignums as an analogy to 'digit'; see crypto.stackexchange.com/questions/59839/… . | |
| Aug 10, 2018 at 13:20 | comment | added | Vega4 | also is the notion of a 'limb' widespread in cryptography? where does it come from? | |
| Aug 10, 2018 at 12:49 | vote | accept | Vega4 | ||
| Aug 8, 2018 at 19:18 | history | answered | Lery | CC BY-SA 4.0 |