nx_ed25519_scalar_test.nx

nx_ed25519_scalar_test.nx -- KAT for scalar reduction + mul. Verifies: sc_reduce side: A. sc_reduce(0^64) == 0^32 B. sc_reduce(L padded to 64) == 0^32 C. sc_reduce(L+1 padded to 64) == 1 (01 00 ... 00) D. sc_reduce(L-1 padded to 64) == L-1 (byte-exact) E. sc_reduce(2*L padded to 64) == 0 scalar mul side: F. [0]B == identity G. [1]B == B (byte-exact compressed) H. [2]B == double(B) I. [3]B == B + B + B Note: skipping [L]B == identity for this turn -- it requires 252 doublings + ~125 adds and exercises bit patterns most prone to off-by-one. It will land as part of Piece 4's verify KAT (which actually computes [s]B for non-trivial s). expect_exit: 0 license_tier: ORIGINAL