Commit Graph

7 Commits

Author SHA1 Message Date
Jack Grigg 97c27e3d5a Use complete addition in SinsemillaCommit
This is necessary because the blinding factor r can be zero with greater
than negligible probability in an adversarial case, which with incomplete
addition would cause the circuit to compute a commitment that is not on
the curve.
2021-09-28 13:13:25 -06:00
str4d 4e33fe7aec Use correct symbol for incomplete addition
Co-authored-by: Daira Hopwood <daira@jacaranda.org>
2021-08-12 21:34:35 +01:00
str4d 459e68b71e
Fix clippy lint
Co-authored-by: Daira Hopwood <daira@jacaranda.org>
2021-08-12 21:32:14 +01:00
Jack Grigg 9f3c9a7e60 Use mixed addition for Sinsemilla bases
Performance improvements:
- MerkleCRH:  ~5%
- Commit^ivk: ~1%
- NoteCommit: ~3%
2021-08-12 15:45:00 +01:00
Jack Grigg bdaf9d06cc clippy: Allow binary operators in IncompletePoint addition
It's not suspicious, it's constant time! :D
2021-04-22 12:09:32 +12:00
Jack Grigg 09e70cb6e3 Improve performance of IncompletePoint addition
We only need to track the occurrence of any edge cases, and we can do so
without expensive inversions at every addition step, by instead
performing the checks on the projective form directly.
2021-04-22 12:01:59 +12:00
Jack Grigg 907ff46078 Simulate incomplete addition
Sinsemilla will use incomplete addition inside the circuit for
efficiency, but the pasta_curves crate uses complete addition.
2021-04-20 10:04:44 +12:00