Properly namespace everything in third_party/fiat/p256.c. This file gets #included into other files, so we shouldn't use generic names like 'fe'. This will let us import other fiat-crypto curves in the future, if we want them. Change-Id: Ie4e222729bde7e4ccd368b86fb9048a2ea4a58ac Reviewed-on: https://boringssl-review.googlesource.com/c/boringssl/+/40824 Reviewed-by: David Benjamin <davidben@google.com> Reviewed-by: Steven Valdez <svaldez@google.com> Commit-Queue: David Benjamin <davidben@google.com>
diff --git a/third_party/fiat/p256.c b/third_party/fiat/p256.c index 4a63ef5..3f6cff7 100644 --- a/third_party/fiat/p256.c +++ b/third_party/fiat/p256.c
@@ -54,129 +54,118 @@ // utility functions, handwritten -#define NBYTES 32 - #if defined(BORINGSSL_NISTP256_64BIT) -#define NLIMBS 4 -typedef uint64_t limb_t; -typedef uint64_t fe[NLIMBS]; -#else // 64BIT; else 32BIT +#define FIAT_P256_NLIMBS 4 +typedef uint64_t fiat_p256_limb_t; +typedef uint64_t fiat_p256_felem[FIAT_P256_NLIMBS]; +#else // 64BIT; else 32BIT -#define NLIMBS 8 -typedef uint32_t limb_t; -typedef uint32_t fe[NLIMBS]; +#define FIAT_P256_NLIMBS 8 +typedef uint32_t fiat_p256_limb_t; +typedef uint32_t fiat_p256_felem[FIAT_P256_NLIMBS]; -#endif // 64BIT +#endif // 64BIT -#define fe_add fiat_p256_add -#define fe_sub fiat_p256_sub -#define fe_opp fiat_p256_opp -#define fe_mul fiat_p256_mul -#define fe_sqr fiat_p256_square - -#define fe_tobytes fiat_p256_to_bytes -#define fe_frombytes fiat_p256_from_bytes - -static limb_t fe_nz(const limb_t in1[NLIMBS]) { - limb_t ret; +static fiat_p256_limb_t fiat_p256_nz( + const fiat_p256_limb_t in1[FIAT_P256_NLIMBS]) { + fiat_p256_limb_t ret; fiat_p256_nonzero(&ret, in1); return ret; } -static void fe_copy(limb_t out[NLIMBS], const limb_t in1[NLIMBS]) { - for (int i = 0; i < NLIMBS; i++) { +static void fiat_p256_copy(fiat_p256_limb_t out[FIAT_P256_NLIMBS], + const fiat_p256_limb_t in1[FIAT_P256_NLIMBS]) { + for (int i = 0; i < FIAT_P256_NLIMBS; i++) { out[i] = in1[i]; } } -static void fe_cmovznz(limb_t out[NLIMBS], limb_t t, const limb_t z[NLIMBS], - const limb_t nz[NLIMBS]) { +static void fiat_p256_cmovznz(fiat_p256_limb_t out[FIAT_P256_NLIMBS], + fiat_p256_limb_t t, + const fiat_p256_limb_t z[FIAT_P256_NLIMBS], + const fiat_p256_limb_t nz[FIAT_P256_NLIMBS]) { fiat_p256_selectznz(out, !!t, z, nz); } -static void fe_from_montgomery(fe x) { - fiat_p256_from_montgomery(x, x); +static void fiat_p256_from_generic(fiat_p256_felem out, const EC_FELEM *in) { + fiat_p256_from_bytes(out, in->bytes); } -static void fe_from_generic(fe out, const EC_FELEM *in) { - fe_frombytes(out, in->bytes); -} - -static void fe_to_generic(EC_FELEM *out, const fe in) { +static void fiat_p256_to_generic(EC_FELEM *out, const fiat_p256_felem in) { // This works because 256 is a multiple of 64, so there are no excess bytes to // zero when rounding up to |BN_ULONG|s. OPENSSL_STATIC_ASSERT( 256 / 8 == sizeof(BN_ULONG) * ((256 + BN_BITS2 - 1) / BN_BITS2), - "fe_tobytes leaves bytes uninitialized"); - fe_tobytes(out->bytes, in); + "fiat_p256_to_bytes leaves bytes uninitialized"); + fiat_p256_to_bytes(out->bytes, in); } -// fe_inv calculates |out| = |in|^{-1} +// fiat_p256_inv calculates |out| = |in|^{-1} // // Based on Fermat's Little Theorem: // a^p = a (mod p) // a^{p-1} = 1 (mod p) // a^{p-2} = a^{-1} (mod p) -static void fe_inv(fe out, const fe in) { - fe ftmp, ftmp2; +static void fiat_p256_inv(fiat_p256_felem out, const fiat_p256_felem in) { + fiat_p256_felem ftmp, ftmp2; // each e_I will hold |in|^{2^I - 1} - fe e2, e4, e8, e16, e32, e64; + fiat_p256_felem e2, e4, e8, e16, e32, e64; - fe_sqr(ftmp, in); // 2^1 - fe_mul(ftmp, in, ftmp); // 2^2 - 2^0 - fe_copy(e2, ftmp); - fe_sqr(ftmp, ftmp); // 2^3 - 2^1 - fe_sqr(ftmp, ftmp); // 2^4 - 2^2 - fe_mul(ftmp, ftmp, e2); // 2^4 - 2^0 - fe_copy(e4, ftmp); - fe_sqr(ftmp, ftmp); // 2^5 - 2^1 - fe_sqr(ftmp, ftmp); // 2^6 - 2^2 - fe_sqr(ftmp, ftmp); // 2^7 - 2^3 - fe_sqr(ftmp, ftmp); // 2^8 - 2^4 - fe_mul(ftmp, ftmp, e4); // 2^8 - 2^0 - fe_copy(e8, ftmp); + fiat_p256_square(ftmp, in); // 2^1 + fiat_p256_mul(ftmp, in, ftmp); // 2^2 - 2^0 + fiat_p256_copy(e2, ftmp); + fiat_p256_square(ftmp, ftmp); // 2^3 - 2^1 + fiat_p256_square(ftmp, ftmp); // 2^4 - 2^2 + fiat_p256_mul(ftmp, ftmp, e2); // 2^4 - 2^0 + fiat_p256_copy(e4, ftmp); + fiat_p256_square(ftmp, ftmp); // 2^5 - 2^1 + fiat_p256_square(ftmp, ftmp); // 2^6 - 2^2 + fiat_p256_square(ftmp, ftmp); // 2^7 - 2^3 + fiat_p256_square(ftmp, ftmp); // 2^8 - 2^4 + fiat_p256_mul(ftmp, ftmp, e4); // 2^8 - 2^0 + fiat_p256_copy(e8, ftmp); for (size_t i = 0; i < 8; i++) { - fe_sqr(ftmp, ftmp); - } // 2^16 - 2^8 - fe_mul(ftmp, ftmp, e8); // 2^16 - 2^0 - fe_copy(e16, ftmp); + fiat_p256_square(ftmp, ftmp); + } // 2^16 - 2^8 + fiat_p256_mul(ftmp, ftmp, e8); // 2^16 - 2^0 + fiat_p256_copy(e16, ftmp); for (size_t i = 0; i < 16; i++) { - fe_sqr(ftmp, ftmp); - } // 2^32 - 2^16 - fe_mul(ftmp, ftmp, e16); // 2^32 - 2^0 - fe_copy(e32, ftmp); + fiat_p256_square(ftmp, ftmp); + } // 2^32 - 2^16 + fiat_p256_mul(ftmp, ftmp, e16); // 2^32 - 2^0 + fiat_p256_copy(e32, ftmp); for (size_t i = 0; i < 32; i++) { - fe_sqr(ftmp, ftmp); + fiat_p256_square(ftmp, ftmp); } // 2^64 - 2^32 - fe_copy(e64, ftmp); - fe_mul(ftmp, ftmp, in); // 2^64 - 2^32 + 2^0 + fiat_p256_copy(e64, ftmp); + fiat_p256_mul(ftmp, ftmp, in); // 2^64 - 2^32 + 2^0 for (size_t i = 0; i < 192; i++) { - fe_sqr(ftmp, ftmp); + fiat_p256_square(ftmp, ftmp); } // 2^256 - 2^224 + 2^192 - fe_mul(ftmp2, e64, e32); // 2^64 - 2^0 + fiat_p256_mul(ftmp2, e64, e32); // 2^64 - 2^0 for (size_t i = 0; i < 16; i++) { - fe_sqr(ftmp2, ftmp2); - } // 2^80 - 2^16 - fe_mul(ftmp2, ftmp2, e16); // 2^80 - 2^0 + fiat_p256_square(ftmp2, ftmp2); + } // 2^80 - 2^16 + fiat_p256_mul(ftmp2, ftmp2, e16); // 2^80 - 2^0 for (size_t i = 0; i < 8; i++) { - fe_sqr(ftmp2, ftmp2); - } // 2^88 - 2^8 - fe_mul(ftmp2, ftmp2, e8); // 2^88 - 2^0 + fiat_p256_square(ftmp2, ftmp2); + } // 2^88 - 2^8 + fiat_p256_mul(ftmp2, ftmp2, e8); // 2^88 - 2^0 for (size_t i = 0; i < 4; i++) { - fe_sqr(ftmp2, ftmp2); - } // 2^92 - 2^4 - fe_mul(ftmp2, ftmp2, e4); // 2^92 - 2^0 - fe_sqr(ftmp2, ftmp2); // 2^93 - 2^1 - fe_sqr(ftmp2, ftmp2); // 2^94 - 2^2 - fe_mul(ftmp2, ftmp2, e2); // 2^94 - 2^0 - fe_sqr(ftmp2, ftmp2); // 2^95 - 2^1 - fe_sqr(ftmp2, ftmp2); // 2^96 - 2^2 - fe_mul(ftmp2, ftmp2, in); // 2^96 - 3 + fiat_p256_square(ftmp2, ftmp2); + } // 2^92 - 2^4 + fiat_p256_mul(ftmp2, ftmp2, e4); // 2^92 - 2^0 + fiat_p256_square(ftmp2, ftmp2); // 2^93 - 2^1 + fiat_p256_square(ftmp2, ftmp2); // 2^94 - 2^2 + fiat_p256_mul(ftmp2, ftmp2, e2); // 2^94 - 2^0 + fiat_p256_square(ftmp2, ftmp2); // 2^95 - 2^1 + fiat_p256_square(ftmp2, ftmp2); // 2^96 - 2^2 + fiat_p256_mul(ftmp2, ftmp2, in); // 2^96 - 3 - fe_mul(out, ftmp2, ftmp); // 2^256 - 2^224 + 2^192 + 2^96 - 3 + fiat_p256_mul(out, ftmp2, ftmp); // 2^256 - 2^224 + 2^192 + 2^96 - 3 } // Group operations @@ -194,7 +183,7 @@ // As a sanity check, a proof that these points form a commutative group: // <https://github.com/mit-plv/fiat-crypto/blob/79f8b5f39ed609339f0233098dee1a3c4e6b3080/src/Curves/Weierstrass/AffineProofs.v#L33> -// point_double calculates 2*(x_in, y_in, z_in) +// fiat_p256_point_double calculates 2*(x_in, y_in, z_in) // // The method is taken from: // http://hyperelliptic.org/EFD/g1p/auto-shortw-jacobian-3.html#doubling-dbl-2001-b @@ -205,47 +194,50 @@ // // Outputs can equal corresponding inputs, i.e., x_out == x_in is allowed. // while x_out == y_in is not (maybe this works, but it's not tested). -static void point_double(fe x_out, fe y_out, fe z_out, - const fe x_in, const fe y_in, const fe z_in) { - fe delta, gamma, beta, ftmp, ftmp2, tmptmp, alpha, fourbeta; +static void fiat_p256_point_double(fiat_p256_felem x_out, fiat_p256_felem y_out, + fiat_p256_felem z_out, + const fiat_p256_felem x_in, + const fiat_p256_felem y_in, + const fiat_p256_felem z_in) { + fiat_p256_felem delta, gamma, beta, ftmp, ftmp2, tmptmp, alpha, fourbeta; // delta = z^2 - fe_sqr(delta, z_in); + fiat_p256_square(delta, z_in); // gamma = y^2 - fe_sqr(gamma, y_in); + fiat_p256_square(gamma, y_in); // beta = x*gamma - fe_mul(beta, x_in, gamma); + fiat_p256_mul(beta, x_in, gamma); // alpha = 3*(x-delta)*(x+delta) - fe_sub(ftmp, x_in, delta); - fe_add(ftmp2, x_in, delta); + fiat_p256_sub(ftmp, x_in, delta); + fiat_p256_add(ftmp2, x_in, delta); - fe_add(tmptmp, ftmp2, ftmp2); - fe_add(ftmp2, ftmp2, tmptmp); - fe_mul(alpha, ftmp, ftmp2); + fiat_p256_add(tmptmp, ftmp2, ftmp2); + fiat_p256_add(ftmp2, ftmp2, tmptmp); + fiat_p256_mul(alpha, ftmp, ftmp2); // x' = alpha^2 - 8*beta - fe_sqr(x_out, alpha); - fe_add(fourbeta, beta, beta); - fe_add(fourbeta, fourbeta, fourbeta); - fe_add(tmptmp, fourbeta, fourbeta); - fe_sub(x_out, x_out, tmptmp); + fiat_p256_square(x_out, alpha); + fiat_p256_add(fourbeta, beta, beta); + fiat_p256_add(fourbeta, fourbeta, fourbeta); + fiat_p256_add(tmptmp, fourbeta, fourbeta); + fiat_p256_sub(x_out, x_out, tmptmp); // z' = (y + z)^2 - gamma - delta - fe_add(delta, gamma, delta); - fe_add(ftmp, y_in, z_in); - fe_sqr(z_out, ftmp); - fe_sub(z_out, z_out, delta); + fiat_p256_add(delta, gamma, delta); + fiat_p256_add(ftmp, y_in, z_in); + fiat_p256_square(z_out, ftmp); + fiat_p256_sub(z_out, z_out, delta); // y' = alpha*(4*beta - x') - 8*gamma^2 - fe_sub(y_out, fourbeta, x_out); - fe_add(gamma, gamma, gamma); - fe_sqr(gamma, gamma); - fe_mul(y_out, alpha, y_out); - fe_add(gamma, gamma, gamma); - fe_sub(y_out, y_out, gamma); + fiat_p256_sub(y_out, fourbeta, x_out); + fiat_p256_add(gamma, gamma, gamma); + fiat_p256_square(gamma, gamma); + fiat_p256_mul(y_out, alpha, y_out); + fiat_p256_add(gamma, gamma, gamma); + fiat_p256_sub(y_out, y_out, gamma); } -// point_add calcuates (x1, y1, z1) + (x2, y2, z2) +// fiat_p256_point_add calculates (x1, y1, z1) + (x2, y2, z2) // // The method is taken from: // http://hyperelliptic.org/EFD/g1p/auto-shortw-jacobian-3.html#addition-add-2007-bl, @@ -259,107 +251,119 @@ // are equal, (while not equal to the point at infinity). This case never // happens during single point multiplication, so there is no timing leak for // ECDH or ECDSA signing. -static void point_add(fe x3, fe y3, fe z3, const fe x1, - const fe y1, const fe z1, const int mixed, - const fe x2, const fe y2, const fe z2) { - fe x_out, y_out, z_out; - limb_t z1nz = fe_nz(z1); - limb_t z2nz = fe_nz(z2); +static void fiat_p256_point_add(fiat_p256_felem x3, fiat_p256_felem y3, + fiat_p256_felem z3, const fiat_p256_felem x1, + const fiat_p256_felem y1, + const fiat_p256_felem z1, const int mixed, + const fiat_p256_felem x2, + const fiat_p256_felem y2, + const fiat_p256_felem z2) { + fiat_p256_felem x_out, y_out, z_out; + fiat_p256_limb_t z1nz = fiat_p256_nz(z1); + fiat_p256_limb_t z2nz = fiat_p256_nz(z2); // z1z1 = z1z1 = z1**2 - fe z1z1; fe_sqr(z1z1, z1); + fiat_p256_felem z1z1; + fiat_p256_square(z1z1, z1); - fe u1, s1, two_z1z2; + fiat_p256_felem u1, s1, two_z1z2; if (!mixed) { // z2z2 = z2**2 - fe z2z2; fe_sqr(z2z2, z2); + fiat_p256_felem z2z2; + fiat_p256_square(z2z2, z2); // u1 = x1*z2z2 - fe_mul(u1, x1, z2z2); + fiat_p256_mul(u1, x1, z2z2); // two_z1z2 = (z1 + z2)**2 - (z1z1 + z2z2) = 2z1z2 - fe_add(two_z1z2, z1, z2); - fe_sqr(two_z1z2, two_z1z2); - fe_sub(two_z1z2, two_z1z2, z1z1); - fe_sub(two_z1z2, two_z1z2, z2z2); + fiat_p256_add(two_z1z2, z1, z2); + fiat_p256_square(two_z1z2, two_z1z2); + fiat_p256_sub(two_z1z2, two_z1z2, z1z1); + fiat_p256_sub(two_z1z2, two_z1z2, z2z2); // s1 = y1 * z2**3 - fe_mul(s1, z2, z2z2); - fe_mul(s1, s1, y1); + fiat_p256_mul(s1, z2, z2z2); + fiat_p256_mul(s1, s1, y1); } else { // We'll assume z2 = 1 (special case z2 = 0 is handled later). // u1 = x1*z2z2 - fe_copy(u1, x1); + fiat_p256_copy(u1, x1); // two_z1z2 = 2z1z2 - fe_add(two_z1z2, z1, z1); + fiat_p256_add(two_z1z2, z1, z1); // s1 = y1 * z2**3 - fe_copy(s1, y1); + fiat_p256_copy(s1, y1); } // u2 = x2*z1z1 - fe u2; fe_mul(u2, x2, z1z1); + fiat_p256_felem u2; + fiat_p256_mul(u2, x2, z1z1); // h = u2 - u1 - fe h; fe_sub(h, u2, u1); + fiat_p256_felem h; + fiat_p256_sub(h, u2, u1); - limb_t xneq = fe_nz(h); + fiat_p256_limb_t xneq = fiat_p256_nz(h); // z_out = two_z1z2 * h - fe_mul(z_out, h, two_z1z2); + fiat_p256_mul(z_out, h, two_z1z2); // z1z1z1 = z1 * z1z1 - fe z1z1z1; fe_mul(z1z1z1, z1, z1z1); + fiat_p256_felem z1z1z1; + fiat_p256_mul(z1z1z1, z1, z1z1); // s2 = y2 * z1**3 - fe s2; fe_mul(s2, y2, z1z1z1); + fiat_p256_felem s2; + fiat_p256_mul(s2, y2, z1z1z1); // r = (s2 - s1)*2 - fe r; - fe_sub(r, s2, s1); - fe_add(r, r, r); + fiat_p256_felem r; + fiat_p256_sub(r, s2, s1); + fiat_p256_add(r, r, r); - limb_t yneq = fe_nz(r); + fiat_p256_limb_t yneq = fiat_p256_nz(r); - limb_t is_nontrivial_double = constant_time_is_zero_w(xneq | yneq) & - ~constant_time_is_zero_w(z1nz) & - ~constant_time_is_zero_w(z2nz); + fiat_p256_limb_t is_nontrivial_double = constant_time_is_zero_w(xneq | yneq) & + ~constant_time_is_zero_w(z1nz) & + ~constant_time_is_zero_w(z2nz); if (is_nontrivial_double) { - point_double(x3, y3, z3, x1, y1, z1); + fiat_p256_point_double(x3, y3, z3, x1, y1, z1); return; } // I = (2h)**2 - fe i; - fe_add(i, h, h); - fe_sqr(i, i); + fiat_p256_felem i; + fiat_p256_add(i, h, h); + fiat_p256_square(i, i); // J = h * I - fe j; fe_mul(j, h, i); + fiat_p256_felem j; + fiat_p256_mul(j, h, i); // V = U1 * I - fe v; fe_mul(v, u1, i); + fiat_p256_felem v; + fiat_p256_mul(v, u1, i); // x_out = r**2 - J - 2V - fe_sqr(x_out, r); - fe_sub(x_out, x_out, j); - fe_sub(x_out, x_out, v); - fe_sub(x_out, x_out, v); + fiat_p256_square(x_out, r); + fiat_p256_sub(x_out, x_out, j); + fiat_p256_sub(x_out, x_out, v); + fiat_p256_sub(x_out, x_out, v); // y_out = r(V-x_out) - 2 * s1 * J - fe_sub(y_out, v, x_out); - fe_mul(y_out, y_out, r); - fe s1j; - fe_mul(s1j, s1, j); - fe_sub(y_out, y_out, s1j); - fe_sub(y_out, y_out, s1j); + fiat_p256_sub(y_out, v, x_out); + fiat_p256_mul(y_out, y_out, r); + fiat_p256_felem s1j; + fiat_p256_mul(s1j, s1, j); + fiat_p256_sub(y_out, y_out, s1j); + fiat_p256_sub(y_out, y_out, s1j); - fe_cmovznz(x_out, z1nz, x2, x_out); - fe_cmovznz(x3, z2nz, x1, x_out); - fe_cmovznz(y_out, z1nz, y2, y_out); - fe_cmovznz(y3, z2nz, y1, y_out); - fe_cmovznz(z_out, z1nz, z2, z_out); - fe_cmovznz(z3, z2nz, z1, z_out); + fiat_p256_cmovznz(x_out, z1nz, x2, x_out); + fiat_p256_cmovznz(x3, z2nz, x1, x_out); + fiat_p256_cmovznz(y_out, z1nz, y2, y_out); + fiat_p256_cmovznz(y3, z2nz, y1, y_out); + fiat_p256_cmovznz(z_out, z1nz, z2, z_out); + fiat_p256_cmovznz(z3, z2nz, z1, z_out); } // Base point pre computation @@ -399,7 +403,7 @@ // g_pre_comp is the table of precomputed base points #if defined(BORINGSSL_NISTP256_64BIT) -static const fe g_pre_comp[2][16][3] = { +static const fiat_p256_felem g_pre_comp[2][16][3] = { {{{0x0, 0x0, 0x0, 0x0}, {0x0, 0x0, 0x0, 0x0}, {0x0, 0x0, 0x0, 0x0}}, {{0x79e730d418a9143c, 0x75ba95fc5fedb601, 0x79fb732b77622510, 0x18905f76a53755c6}, @@ -553,181 +557,181 @@ 0x8e24d3c46860bbd0}, {0x1, 0xffffffff00000000, 0xffffffffffffffff, 0xfffffffe}}}}; #else -static const fe g_pre_comp[2][16][3] = { - {{{0x0,0x0, 0x0,0x0, 0x0,0x0, 0x0,0x0}, - {0x0,0x0, 0x0,0x0, 0x0,0x0, 0x0,0x0}, - {0x0,0x0, 0x0,0x0, 0x0,0x0, 0x0,0x0}}, - {{0x18a9143c,0x79e730d4, 0x5fedb601,0x75ba95fc, 0x77622510,0x79fb732b, - 0xa53755c6,0x18905f76}, - {0xce95560a,0xddf25357, 0xba19e45c,0x8b4ab8e4, 0xdd21f325,0xd2e88688, - 0x25885d85,0x8571ff18}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x16a0d2bb,0x4f922fc5, 0x1a623499,0xd5cc16c, 0x57c62c8b,0x9241cf3a, - 0xfd1b667f,0x2f5e6961}, - {0xf5a01797,0x5c15c70b, 0x60956192,0x3d20b44d, 0x71fdb52,0x4911b37, - 0x8d6f0f7b,0xf648f916}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0xe137bbbc,0x9e566847, 0x8a6a0bec,0xe434469e, 0x79d73463,0xb1c42761, - 0x133d0015,0x5abe0285}, - {0xc04c7dab,0x92aa837c, 0x43260c07,0x573d9f4c, 0x78e6cc37,0xc931562, - 0x6b6f7383,0x94bb725b}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0xbfe20925,0x62a8c244, 0x8fdce867,0x91c19ac3, 0xdd387063,0x5a96a5d5, - 0x21d324f6,0x61d587d4}, - {0xa37173ea,0xe87673a2, 0x53778b65,0x23848008, 0x5bab43e,0x10f8441e, - 0x4621efbe,0xfa11fe12}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x2cb19ffd,0x1c891f2b, 0xb1923c23,0x1ba8d5b, 0x8ac5ca8e,0xb6d03d67, - 0x1f13bedc,0x586eb04c}, - {0x27e8ed09,0xc35c6e5, 0x1819ede2,0x1e81a33c, 0x56c652fa,0x278fd6c0, - 0x70864f11,0x19d5ac08}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0xd2b533d5,0x62577734, 0xa1bdddc0,0x673b8af6, 0xa79ec293,0x577e7c9a, - 0xc3b266b1,0xbb6de651}, - {0xb65259b3,0xe7e9303a, 0xd03a7480,0xd6a0afd3, 0x9b3cfc27,0xc5ac83d1, - 0x5d18b99b,0x60b4619a}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x1ae5aa1c,0xbd6a38e1, 0x49e73658,0xb8b7652b, 0xee5f87ed,0xb130014, - 0xaeebffcd,0x9d0f27b2}, - {0x7a730a55,0xca924631, 0xddbbc83a,0x9c955b2f, 0xac019a71,0x7c1dfe0, - 0x356ec48d,0x244a566d}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0xf4f8b16a,0x56f8410e, 0xc47b266a,0x97241afe, 0x6d9c87c1,0xa406b8e, - 0xcd42ab1b,0x803f3e02}, - {0x4dbec69,0x7f0309a8, 0x3bbad05f,0xa83b85f7, 0xad8e197f,0xc6097273, - 0x5067adc1,0xc097440e}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0xc379ab34,0x846a56f2, 0x841df8d1,0xa8ee068b, 0x176c68ef,0x20314459, - 0x915f1f30,0xf1af32d5}, - {0x5d75bd50,0x99c37531, 0xf72f67bc,0x837cffba, 0x48d7723f,0x613a418, - 0xe2d41c8b,0x23d0f130}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0xd5be5a2b,0xed93e225, 0x5934f3c6,0x6fe79983, 0x22626ffc,0x43140926, - 0x7990216a,0x50bbb4d9}, - {0xe57ec63e,0x378191c6, 0x181dcdb2,0x65422c40, 0x236e0f6,0x41a8099b, - 0x1fe49c3,0x2b100118}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x9b391593,0xfc68b5c5, 0x598270fc,0xc385f5a2, 0xd19adcbb,0x7144f3aa, - 0x83fbae0c,0xdd558999}, - {0x74b82ff4,0x93b88b8e, 0x71e734c9,0xd2e03c40, 0x43c0322a,0x9a7a9eaf, - 0x149d6041,0xe6e4c551}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x80ec21fe,0x5fe14bfe, 0xc255be82,0xf6ce116a, 0x2f4a5d67,0x98bc5a07, - 0xdb7e63af,0xfad27148}, - {0x29ab05b3,0x90c0b6ac, 0x4e251ae6,0x37a9a83c, 0xc2aade7d,0xa7dc875, - 0x9f0e1a84,0x77387de3}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0xa56c0dd7,0x1e9ecc49, 0x46086c74,0xa5cffcd8, 0xf505aece,0x8f7a1408, - 0xbef0c47e,0xb37b85c0}, - {0xcc0e6a8f,0x3596b6e4, 0x6b388f23,0xfd6d4bbf, 0xc39cef4e,0xaba453fa, - 0xf9f628d5,0x9c135ac8}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x95c8f8be,0xa1c7294, 0x3bf362bf,0x2961c480, 0xdf63d4ac,0x9e418403, - 0x91ece900,0xc109f9cb}, - {0x58945705,0xc2d095d0, 0xddeb85c0,0xb9083d96, 0x7a40449b,0x84692b8d, - 0x2eee1ee1,0x9bc3344f}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x42913074,0xd5ae356, 0x48a542b1,0x55491b27, 0xb310732a,0x469ca665, - 0x5f1a4cc1,0x29591d52}, - {0xb84f983f,0xe76f5b6b, 0x9f5f84e1,0xbe7eef41, 0x80baa189,0x1200d496, - 0x18ef332c,0x6376551f}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}}, - {{{0x0,0x0, 0x0,0x0, 0x0,0x0, 0x0,0x0}, - {0x0,0x0, 0x0,0x0, 0x0,0x0, 0x0,0x0}, - {0x0,0x0, 0x0,0x0, 0x0,0x0, 0x0,0x0}}, - {{0x4147519a,0x20288602, 0x26b372f0,0xd0981eac, 0xa785ebc8,0xa9d4a7ca, - 0xdbdf58e9,0xd953c50d}, - {0xfd590f8f,0x9d6361cc, 0x44e6c917,0x72e9626b, 0x22eb64cf,0x7fd96110, - 0x9eb288f3,0x863ebb7e}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0xb0e63d34,0x4fe7ee31, 0xa9e54fab,0xf4600572, 0xd5e7b5a4,0xc0493334, - 0x6d54831,0x8589fb92}, - {0x6583553a,0xaa70f5cc, 0xe25649e5,0x879094a, 0x10044652,0xcc904507, - 0x2541c4f,0xebb0696d}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x3b89da99,0xabbaa0c0, 0xb8284022,0xa6f2d79e, 0xb81c05e8,0x27847862, - 0x5e54d63,0x337a4b59}, - {0x21f7794a,0x3c67500d, 0x7d6d7f61,0x207005b7, 0x4cfd6e8,0xa5a3781, - 0xf4c2fbd6,0xd65e0d5}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x6d3549cf,0xd433e50f, 0xfacd665e,0x6f33696f, 0xce11fcb4,0x695bfdac, - 0xaf7c9860,0x810ee252}, - {0x7159bb2c,0x65450fe1, 0x758b357b,0xf7dfbebe, 0xd69fea72,0x2b057e74, - 0x92731745,0xd485717a}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0xe83f7669,0xce1f69bb, 0x72877d6b,0x9f8ae82, 0x3244278d,0x9548ae54, - 0xe3c2c19c,0x207755de}, - {0x6fef1945,0x87bd61d9, 0xb12d28c3,0x18813cef, 0x72df64aa,0x9fbcd1d6, - 0x7154b00d,0x48dc5ee5}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0xf49a3154,0xef0f469e, 0x6e2b2e9a,0x3e85a595, 0xaa924a9c,0x45aaec1e, - 0xa09e4719,0xaa12dfc8}, - {0x4df69f1d,0x26f27227, 0xa2ff5e73,0xe0e4c82c, 0xb7a9dd44,0xb9d8ce73, - 0xe48ca901,0x6c036e73}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0xa47153f0,0xe1e421e1, 0x920418c9,0xb86c3b79, 0x705d7672,0x93bdce87, - 0xcab79a77,0xf25ae793}, - {0x6d869d0c,0x1f3194a3, 0x4986c264,0x9d55c882, 0x96e945e,0x49fb5ea3, - 0x13db0a3e,0x39b8e653}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x35d0b34a,0xe3417bc0, 0x8327c0a7,0x440b386b, 0xac0362d1,0x8fb7262d, - 0xe0cdf943,0x2c41114c}, - {0xad95a0b1,0x2ba5cef1, 0x67d54362,0xc09b37a8, 0x1e486c9,0x26d6cdd2, - 0x42ff9297,0x20477abf}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0xbc0a67d2,0xf121b41, 0x444d248a,0x62d4760a, 0x659b4737,0xe044f1d, - 0x250bb4a8,0x8fde365}, - {0x848bf287,0xaceec3da, 0xd3369d6e,0xc2a62182, 0x92449482,0x3582dfdc, - 0x565d6cd7,0x2f7e2fd2}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x178a876b,0xa0122b5, 0x85104b4,0x51ff96ff, 0x14f29f76,0x50b31ab, - 0x5f87d4e6,0x84abb28b}, - {0x8270790a,0xd5ed439f, 0x85e3f46b,0x2d6cb59d, 0x6c1e2212,0x75f55c1b, - 0x17655640,0xe5436f67}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x9aeb596d,0xc2965ecc, 0x23c92b4,0x1ea03e7, 0x2e013961,0x4704b4b6, - 0x905ea367,0xca8fd3f}, - {0x551b2b61,0x92523a42, 0x390fcd06,0x1eb7a89c, 0x392a63e,0xe7f1d2be, - 0x4ddb0c33,0x96dca264}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x15339848,0x231c210e, 0x70778c8d,0xe87a28e8, 0x6956e170,0x9d1de661, - 0x2bb09c0b,0x4ac3c938}, - {0x6998987d,0x19be0551, 0xae09f4d6,0x8b2376c4, 0x1a3f933d,0x1de0b765, - 0xe39705f4,0x380d94c7}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x8c31c31d,0x3685954b, 0x5bf21a0c,0x68533d00, 0x75c79ec9,0xbd7626e, - 0x42c69d54,0xca177547}, - {0xf6d2dbb2,0xcc6edaff, 0x174a9d18,0xfd0d8cbd, 0xaa4578e8,0x875e8793, - 0x9cab2ce6,0xa976a713}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0xb43ea1db,0xce37ab11, 0x5259d292,0xa7ff1a9, 0x8f84f186,0x851b0221, - 0xdefaad13,0xa7222bea}, - {0x2b0a9144,0xa2ac78ec, 0xf2fa59c5,0x5a024051, 0x6147ce38,0x91d1eca5, - 0xbc2ac690,0xbe94d523}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}, - {{0x79ec1a0f,0x2d8daefd, 0xceb39c97,0x3bbcd6fd, 0x58f61a95,0xf5575ffc, - 0xadf7b420,0xdbd986c4}, - {0x15f39eb7,0x81aa8814, 0xb98d976c,0x6ee2fcf5, 0xcf2f717d,0x5465475d, - 0x6860bbd0,0x8e24d3c4}, - {0x1,0x0, 0x0,0xffffffff, 0xffffffff,0xffffffff, 0xfffffffe,0x0}}}}; +static const fiat_p256_felem g_pre_comp[2][16][3] = { + {{{0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, + {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, + {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}}, + {{0x18a9143c, 0x79e730d4, 0x5fedb601, 0x75ba95fc, 0x77622510, 0x79fb732b, + 0xa53755c6, 0x18905f76}, + {0xce95560a, 0xddf25357, 0xba19e45c, 0x8b4ab8e4, 0xdd21f325, 0xd2e88688, + 0x25885d85, 0x8571ff18}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x16a0d2bb, 0x4f922fc5, 0x1a623499, 0xd5cc16c, 0x57c62c8b, 0x9241cf3a, + 0xfd1b667f, 0x2f5e6961}, + {0xf5a01797, 0x5c15c70b, 0x60956192, 0x3d20b44d, 0x71fdb52, 0x4911b37, + 0x8d6f0f7b, 0xf648f916}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0xe137bbbc, 0x9e566847, 0x8a6a0bec, 0xe434469e, 0x79d73463, 0xb1c42761, + 0x133d0015, 0x5abe0285}, + {0xc04c7dab, 0x92aa837c, 0x43260c07, 0x573d9f4c, 0x78e6cc37, 0xc931562, + 0x6b6f7383, 0x94bb725b}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0xbfe20925, 0x62a8c244, 0x8fdce867, 0x91c19ac3, 0xdd387063, 0x5a96a5d5, + 0x21d324f6, 0x61d587d4}, + {0xa37173ea, 0xe87673a2, 0x53778b65, 0x23848008, 0x5bab43e, 0x10f8441e, + 0x4621efbe, 0xfa11fe12}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x2cb19ffd, 0x1c891f2b, 0xb1923c23, 0x1ba8d5b, 0x8ac5ca8e, 0xb6d03d67, + 0x1f13bedc, 0x586eb04c}, + {0x27e8ed09, 0xc35c6e5, 0x1819ede2, 0x1e81a33c, 0x56c652fa, 0x278fd6c0, + 0x70864f11, 0x19d5ac08}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0xd2b533d5, 0x62577734, 0xa1bdddc0, 0x673b8af6, 0xa79ec293, 0x577e7c9a, + 0xc3b266b1, 0xbb6de651}, + {0xb65259b3, 0xe7e9303a, 0xd03a7480, 0xd6a0afd3, 0x9b3cfc27, 0xc5ac83d1, + 0x5d18b99b, 0x60b4619a}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x1ae5aa1c, 0xbd6a38e1, 0x49e73658, 0xb8b7652b, 0xee5f87ed, 0xb130014, + 0xaeebffcd, 0x9d0f27b2}, + {0x7a730a55, 0xca924631, 0xddbbc83a, 0x9c955b2f, 0xac019a71, 0x7c1dfe0, + 0x356ec48d, 0x244a566d}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0xf4f8b16a, 0x56f8410e, 0xc47b266a, 0x97241afe, 0x6d9c87c1, 0xa406b8e, + 0xcd42ab1b, 0x803f3e02}, + {0x4dbec69, 0x7f0309a8, 0x3bbad05f, 0xa83b85f7, 0xad8e197f, 0xc6097273, + 0x5067adc1, 0xc097440e}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0xc379ab34, 0x846a56f2, 0x841df8d1, 0xa8ee068b, 0x176c68ef, 0x20314459, + 0x915f1f30, 0xf1af32d5}, + {0x5d75bd50, 0x99c37531, 0xf72f67bc, 0x837cffba, 0x48d7723f, 0x613a418, + 0xe2d41c8b, 0x23d0f130}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0xd5be5a2b, 0xed93e225, 0x5934f3c6, 0x6fe79983, 0x22626ffc, 0x43140926, + 0x7990216a, 0x50bbb4d9}, + {0xe57ec63e, 0x378191c6, 0x181dcdb2, 0x65422c40, 0x236e0f6, 0x41a8099b, + 0x1fe49c3, 0x2b100118}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x9b391593, 0xfc68b5c5, 0x598270fc, 0xc385f5a2, 0xd19adcbb, 0x7144f3aa, + 0x83fbae0c, 0xdd558999}, + {0x74b82ff4, 0x93b88b8e, 0x71e734c9, 0xd2e03c40, 0x43c0322a, 0x9a7a9eaf, + 0x149d6041, 0xe6e4c551}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x80ec21fe, 0x5fe14bfe, 0xc255be82, 0xf6ce116a, 0x2f4a5d67, 0x98bc5a07, + 0xdb7e63af, 0xfad27148}, + {0x29ab05b3, 0x90c0b6ac, 0x4e251ae6, 0x37a9a83c, 0xc2aade7d, 0xa7dc875, + 0x9f0e1a84, 0x77387de3}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0xa56c0dd7, 0x1e9ecc49, 0x46086c74, 0xa5cffcd8, 0xf505aece, 0x8f7a1408, + 0xbef0c47e, 0xb37b85c0}, + {0xcc0e6a8f, 0x3596b6e4, 0x6b388f23, 0xfd6d4bbf, 0xc39cef4e, 0xaba453fa, + 0xf9f628d5, 0x9c135ac8}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x95c8f8be, 0xa1c7294, 0x3bf362bf, 0x2961c480, 0xdf63d4ac, 0x9e418403, + 0x91ece900, 0xc109f9cb}, + {0x58945705, 0xc2d095d0, 0xddeb85c0, 0xb9083d96, 0x7a40449b, 0x84692b8d, + 0x2eee1ee1, 0x9bc3344f}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x42913074, 0xd5ae356, 0x48a542b1, 0x55491b27, 0xb310732a, 0x469ca665, + 0x5f1a4cc1, 0x29591d52}, + {0xb84f983f, 0xe76f5b6b, 0x9f5f84e1, 0xbe7eef41, 0x80baa189, 0x1200d496, + 0x18ef332c, 0x6376551f}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}}, + {{{0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, + {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, + {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}}, + {{0x4147519a, 0x20288602, 0x26b372f0, 0xd0981eac, 0xa785ebc8, 0xa9d4a7ca, + 0xdbdf58e9, 0xd953c50d}, + {0xfd590f8f, 0x9d6361cc, 0x44e6c917, 0x72e9626b, 0x22eb64cf, 0x7fd96110, + 0x9eb288f3, 0x863ebb7e}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0xb0e63d34, 0x4fe7ee31, 0xa9e54fab, 0xf4600572, 0xd5e7b5a4, 0xc0493334, + 0x6d54831, 0x8589fb92}, + {0x6583553a, 0xaa70f5cc, 0xe25649e5, 0x879094a, 0x10044652, 0xcc904507, + 0x2541c4f, 0xebb0696d}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x3b89da99, 0xabbaa0c0, 0xb8284022, 0xa6f2d79e, 0xb81c05e8, 0x27847862, + 0x5e54d63, 0x337a4b59}, + {0x21f7794a, 0x3c67500d, 0x7d6d7f61, 0x207005b7, 0x4cfd6e8, 0xa5a3781, + 0xf4c2fbd6, 0xd65e0d5}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x6d3549cf, 0xd433e50f, 0xfacd665e, 0x6f33696f, 0xce11fcb4, 0x695bfdac, + 0xaf7c9860, 0x810ee252}, + {0x7159bb2c, 0x65450fe1, 0x758b357b, 0xf7dfbebe, 0xd69fea72, 0x2b057e74, + 0x92731745, 0xd485717a}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0xe83f7669, 0xce1f69bb, 0x72877d6b, 0x9f8ae82, 0x3244278d, 0x9548ae54, + 0xe3c2c19c, 0x207755de}, + {0x6fef1945, 0x87bd61d9, 0xb12d28c3, 0x18813cef, 0x72df64aa, 0x9fbcd1d6, + 0x7154b00d, 0x48dc5ee5}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0xf49a3154, 0xef0f469e, 0x6e2b2e9a, 0x3e85a595, 0xaa924a9c, 0x45aaec1e, + 0xa09e4719, 0xaa12dfc8}, + {0x4df69f1d, 0x26f27227, 0xa2ff5e73, 0xe0e4c82c, 0xb7a9dd44, 0xb9d8ce73, + 0xe48ca901, 0x6c036e73}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0xa47153f0, 0xe1e421e1, 0x920418c9, 0xb86c3b79, 0x705d7672, 0x93bdce87, + 0xcab79a77, 0xf25ae793}, + {0x6d869d0c, 0x1f3194a3, 0x4986c264, 0x9d55c882, 0x96e945e, 0x49fb5ea3, + 0x13db0a3e, 0x39b8e653}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x35d0b34a, 0xe3417bc0, 0x8327c0a7, 0x440b386b, 0xac0362d1, 0x8fb7262d, + 0xe0cdf943, 0x2c41114c}, + {0xad95a0b1, 0x2ba5cef1, 0x67d54362, 0xc09b37a8, 0x1e486c9, 0x26d6cdd2, + 0x42ff9297, 0x20477abf}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0xbc0a67d2, 0xf121b41, 0x444d248a, 0x62d4760a, 0x659b4737, 0xe044f1d, + 0x250bb4a8, 0x8fde365}, + {0x848bf287, 0xaceec3da, 0xd3369d6e, 0xc2a62182, 0x92449482, 0x3582dfdc, + 0x565d6cd7, 0x2f7e2fd2}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x178a876b, 0xa0122b5, 0x85104b4, 0x51ff96ff, 0x14f29f76, 0x50b31ab, + 0x5f87d4e6, 0x84abb28b}, + {0x8270790a, 0xd5ed439f, 0x85e3f46b, 0x2d6cb59d, 0x6c1e2212, 0x75f55c1b, + 0x17655640, 0xe5436f67}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x9aeb596d, 0xc2965ecc, 0x23c92b4, 0x1ea03e7, 0x2e013961, 0x4704b4b6, + 0x905ea367, 0xca8fd3f}, + {0x551b2b61, 0x92523a42, 0x390fcd06, 0x1eb7a89c, 0x392a63e, 0xe7f1d2be, + 0x4ddb0c33, 0x96dca264}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x15339848, 0x231c210e, 0x70778c8d, 0xe87a28e8, 0x6956e170, 0x9d1de661, + 0x2bb09c0b, 0x4ac3c938}, + {0x6998987d, 0x19be0551, 0xae09f4d6, 0x8b2376c4, 0x1a3f933d, 0x1de0b765, + 0xe39705f4, 0x380d94c7}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x8c31c31d, 0x3685954b, 0x5bf21a0c, 0x68533d00, 0x75c79ec9, 0xbd7626e, + 0x42c69d54, 0xca177547}, + {0xf6d2dbb2, 0xcc6edaff, 0x174a9d18, 0xfd0d8cbd, 0xaa4578e8, 0x875e8793, + 0x9cab2ce6, 0xa976a713}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0xb43ea1db, 0xce37ab11, 0x5259d292, 0xa7ff1a9, 0x8f84f186, 0x851b0221, + 0xdefaad13, 0xa7222bea}, + {0x2b0a9144, 0xa2ac78ec, 0xf2fa59c5, 0x5a024051, 0x6147ce38, 0x91d1eca5, + 0xbc2ac690, 0xbe94d523}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}, + {{0x79ec1a0f, 0x2d8daefd, 0xceb39c97, 0x3bbcd6fd, 0x58f61a95, 0xf5575ffc, + 0xadf7b420, 0xdbd986c4}, + {0x15f39eb7, 0x81aa8814, 0xb98d976c, 0x6ee2fcf5, 0xcf2f717d, 0x5465475d, + 0x6860bbd0, 0x8e24d3c4}, + {0x1, 0x0, 0x0, 0xffffffff, 0xffffffff, 0xffffffff, 0xfffffffe, 0x0}}}}; #endif -// select_point selects the |idx|th point from a precomputation table and -// copies it to out. -static void select_point(const limb_t idx, size_t size, - const fe pre_comp[/*size*/][3], - fe out[3]) { - OPENSSL_memset(out, 0, sizeof(fe) * 3); +// fiat_p256_select_point selects the |idx|th point from a precomputation table +// and copies it to out. +static void fiat_p256_select_point(const fiat_p256_limb_t idx, size_t size, + const fiat_p256_felem pre_comp[/*size*/][3], + fiat_p256_felem out[3]) { + OPENSSL_memset(out, 0, sizeof(fiat_p256_felem) * 3); for (size_t i = 0; i < size; i++) { - limb_t mismatch = i ^ idx; - fe_cmovznz(out[0], mismatch, pre_comp[i][0], out[0]); - fe_cmovznz(out[1], mismatch, pre_comp[i][1], out[1]); - fe_cmovznz(out[2], mismatch, pre_comp[i][2], out[2]); + fiat_p256_limb_t mismatch = i ^ idx; + fiat_p256_cmovznz(out[0], mismatch, pre_comp[i][0], out[0]); + fiat_p256_cmovznz(out[1], mismatch, pre_comp[i][1], out[1]); + fiat_p256_cmovznz(out[2], mismatch, pre_comp[i][2], out[2]); } } -// get_bit returns the |i|th bit in |in| -static char get_bit(const uint8_t *in, int i) { +// fiat_p256_get_bit returns the |i|th bit in |in| +static char fiat_p256_get_bit(const uint8_t *in, int i) { if (i < 0 || i >= 256) { return 0; } @@ -746,29 +750,30 @@ return 0; } - fe z1, z2; - fe_from_generic(z1, &point->Z); - fe_inv(z2, z1); - fe_sqr(z1, z2); + fiat_p256_felem z1, z2; + fiat_p256_from_generic(z1, &point->Z); + fiat_p256_inv(z2, z1); + fiat_p256_square(z1, z2); - // Instead of using |fe_from_montgomery| to convert the |x| coordinate and - // then calling |fe_from_montgomery| again to convert the |y| coordinate - // below, convert the common factor |z1| once now, saving one reduction. - fe_from_montgomery(z1); + // Instead of using |fiat_p256_from_montgomery| to convert the |x| coordinate + // and then calling |fiat_p256_from_montgomery| again to convert the |y| + // coordinate below, convert the common factor |z1| once now, saving one + // reduction. + fiat_p256_from_montgomery(z1, z1); if (x_out != NULL) { - fe x; - fe_from_generic(x, &point->X); - fe_mul(x, x, z1); - fe_to_generic(x_out, x); + fiat_p256_felem x; + fiat_p256_from_generic(x, &point->X); + fiat_p256_mul(x, x, z1); + fiat_p256_to_generic(x_out, x); } if (y_out != NULL) { - fe y; - fe_from_generic(y, &point->Y); - fe_mul(z1, z1, z2); - fe_mul(y, y, z1); - fe_to_generic(y_out, y); + fiat_p256_felem y; + fiat_p256_from_generic(y, &point->Y); + fiat_p256_mul(z1, z1, z2); + fiat_p256_mul(y, y, z1); + fiat_p256_to_generic(y_out, y); } return 1; @@ -776,141 +781,143 @@ static void ec_GFp_nistp256_add(const EC_GROUP *group, EC_RAW_POINT *r, const EC_RAW_POINT *a, const EC_RAW_POINT *b) { - fe x1, y1, z1, x2, y2, z2; - fe_from_generic(x1, &a->X); - fe_from_generic(y1, &a->Y); - fe_from_generic(z1, &a->Z); - fe_from_generic(x2, &b->X); - fe_from_generic(y2, &b->Y); - fe_from_generic(z2, &b->Z); - point_add(x1, y1, z1, x1, y1, z1, 0 /* both Jacobian */, x2, y2, z2); - fe_to_generic(&r->X, x1); - fe_to_generic(&r->Y, y1); - fe_to_generic(&r->Z, z1); + fiat_p256_felem x1, y1, z1, x2, y2, z2; + fiat_p256_from_generic(x1, &a->X); + fiat_p256_from_generic(y1, &a->Y); + fiat_p256_from_generic(z1, &a->Z); + fiat_p256_from_generic(x2, &b->X); + fiat_p256_from_generic(y2, &b->Y); + fiat_p256_from_generic(z2, &b->Z); + fiat_p256_point_add(x1, y1, z1, x1, y1, z1, 0 /* both Jacobian */, x2, y2, + z2); + fiat_p256_to_generic(&r->X, x1); + fiat_p256_to_generic(&r->Y, y1); + fiat_p256_to_generic(&r->Z, z1); } static void ec_GFp_nistp256_dbl(const EC_GROUP *group, EC_RAW_POINT *r, const EC_RAW_POINT *a) { - fe x, y, z; - fe_from_generic(x, &a->X); - fe_from_generic(y, &a->Y); - fe_from_generic(z, &a->Z); - point_double(x, y, z, x, y, z); - fe_to_generic(&r->X, x); - fe_to_generic(&r->Y, y); - fe_to_generic(&r->Z, z); + fiat_p256_felem x, y, z; + fiat_p256_from_generic(x, &a->X); + fiat_p256_from_generic(y, &a->Y); + fiat_p256_from_generic(z, &a->Z); + fiat_p256_point_double(x, y, z, x, y, z); + fiat_p256_to_generic(&r->X, x); + fiat_p256_to_generic(&r->Y, y); + fiat_p256_to_generic(&r->Z, z); } static void ec_GFp_nistp256_point_mul(const EC_GROUP *group, EC_RAW_POINT *r, const EC_RAW_POINT *p, const EC_SCALAR *scalar) { - fe p_pre_comp[17][3]; + fiat_p256_felem p_pre_comp[17][3]; OPENSSL_memset(&p_pre_comp, 0, sizeof(p_pre_comp)); // Precompute multiples. - fe_from_generic(p_pre_comp[1][0], &p->X); - fe_from_generic(p_pre_comp[1][1], &p->Y); - fe_from_generic(p_pre_comp[1][2], &p->Z); + fiat_p256_from_generic(p_pre_comp[1][0], &p->X); + fiat_p256_from_generic(p_pre_comp[1][1], &p->Y); + fiat_p256_from_generic(p_pre_comp[1][2], &p->Z); for (size_t j = 2; j <= 16; ++j) { if (j & 1) { - point_add(p_pre_comp[j][0], p_pre_comp[j][1], p_pre_comp[j][2], - p_pre_comp[1][0], p_pre_comp[1][1], p_pre_comp[1][2], 0, - p_pre_comp[j - 1][0], p_pre_comp[j - 1][1], - p_pre_comp[j - 1][2]); + fiat_p256_point_add(p_pre_comp[j][0], p_pre_comp[j][1], p_pre_comp[j][2], + p_pre_comp[1][0], p_pre_comp[1][1], p_pre_comp[1][2], + 0, p_pre_comp[j - 1][0], p_pre_comp[j - 1][1], + p_pre_comp[j - 1][2]); } else { - point_double(p_pre_comp[j][0], p_pre_comp[j][1], p_pre_comp[j][2], - p_pre_comp[j / 2][0], p_pre_comp[j / 2][1], - p_pre_comp[j / 2][2]); + fiat_p256_point_double(p_pre_comp[j][0], p_pre_comp[j][1], + p_pre_comp[j][2], p_pre_comp[j / 2][0], + p_pre_comp[j / 2][1], p_pre_comp[j / 2][2]); } } // Set nq to the point at infinity. - fe nq[3] = {{0}, {0}, {0}}, ftmp, tmp[3]; + fiat_p256_felem nq[3] = {{0}, {0}, {0}}, ftmp, tmp[3]; // Loop over |scalar| msb-to-lsb, incorporating |p_pre_comp| every 5th round. int skip = 1; // Save two point operations in the first round. for (size_t i = 255; i < 256; i--) { // double if (!skip) { - point_double(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2]); + fiat_p256_point_double(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2]); } // do other additions every 5 doublings if (i % 5 == 0) { - uint64_t bits = get_bit(scalar->bytes, i + 4) << 5; - bits |= get_bit(scalar->bytes, i + 3) << 4; - bits |= get_bit(scalar->bytes, i + 2) << 3; - bits |= get_bit(scalar->bytes, i + 1) << 2; - bits |= get_bit(scalar->bytes, i) << 1; - bits |= get_bit(scalar->bytes, i - 1); + uint64_t bits = fiat_p256_get_bit(scalar->bytes, i + 4) << 5; + bits |= fiat_p256_get_bit(scalar->bytes, i + 3) << 4; + bits |= fiat_p256_get_bit(scalar->bytes, i + 2) << 3; + bits |= fiat_p256_get_bit(scalar->bytes, i + 1) << 2; + bits |= fiat_p256_get_bit(scalar->bytes, i) << 1; + bits |= fiat_p256_get_bit(scalar->bytes, i - 1); uint8_t sign, digit; ec_GFp_nistp_recode_scalar_bits(&sign, &digit, bits); // select the point to add or subtract, in constant time. - select_point(digit, 17, (const fe(*)[3])p_pre_comp, tmp); - fe_opp(ftmp, tmp[1]); // (X, -Y, Z) is the negative point. - fe_cmovznz(tmp[1], sign, tmp[1], ftmp); + fiat_p256_select_point(digit, 17, (const fiat_p256_felem(*)[3])p_pre_comp, + tmp); + fiat_p256_opp(ftmp, tmp[1]); // (X, -Y, Z) is the negative point. + fiat_p256_cmovznz(tmp[1], sign, tmp[1], ftmp); if (!skip) { - point_add(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2], 0 /* mixed */, - tmp[0], tmp[1], tmp[2]); + fiat_p256_point_add(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2], + 0 /* mixed */, tmp[0], tmp[1], tmp[2]); } else { - fe_copy(nq[0], tmp[0]); - fe_copy(nq[1], tmp[1]); - fe_copy(nq[2], tmp[2]); + fiat_p256_copy(nq[0], tmp[0]); + fiat_p256_copy(nq[1], tmp[1]); + fiat_p256_copy(nq[2], tmp[2]); skip = 0; } } } - fe_to_generic(&r->X, nq[0]); - fe_to_generic(&r->Y, nq[1]); - fe_to_generic(&r->Z, nq[2]); + fiat_p256_to_generic(&r->X, nq[0]); + fiat_p256_to_generic(&r->Y, nq[1]); + fiat_p256_to_generic(&r->Z, nq[2]); } static void ec_GFp_nistp256_point_mul_base(const EC_GROUP *group, EC_RAW_POINT *r, const EC_SCALAR *scalar) { // Set nq to the point at infinity. - fe nq[3] = {{0}, {0}, {0}}, tmp[3]; + fiat_p256_felem nq[3] = {{0}, {0}, {0}}, tmp[3]; int skip = 1; // Save two point operations in the first round. for (size_t i = 31; i < 32; i--) { if (!skip) { - point_double(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2]); + fiat_p256_point_double(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2]); } // First, look 32 bits upwards. - uint64_t bits = get_bit(scalar->bytes, i + 224) << 3; - bits |= get_bit(scalar->bytes, i + 160) << 2; - bits |= get_bit(scalar->bytes, i + 96) << 1; - bits |= get_bit(scalar->bytes, i + 32); + uint64_t bits = fiat_p256_get_bit(scalar->bytes, i + 224) << 3; + bits |= fiat_p256_get_bit(scalar->bytes, i + 160) << 2; + bits |= fiat_p256_get_bit(scalar->bytes, i + 96) << 1; + bits |= fiat_p256_get_bit(scalar->bytes, i + 32); // Select the point to add, in constant time. - select_point(bits, 16, g_pre_comp[1], tmp); + fiat_p256_select_point(bits, 16, g_pre_comp[1], tmp); if (!skip) { - point_add(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2], 1 /* mixed */, tmp[0], - tmp[1], tmp[2]); + fiat_p256_point_add(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2], + 1 /* mixed */, tmp[0], tmp[1], tmp[2]); } else { - fe_copy(nq[0], tmp[0]); - fe_copy(nq[1], tmp[1]); - fe_copy(nq[2], tmp[2]); + fiat_p256_copy(nq[0], tmp[0]); + fiat_p256_copy(nq[1], tmp[1]); + fiat_p256_copy(nq[2], tmp[2]); skip = 0; } // Second, look at the current position. - bits = get_bit(scalar->bytes, i + 192) << 3; - bits |= get_bit(scalar->bytes, i + 128) << 2; - bits |= get_bit(scalar->bytes, i + 64) << 1; - bits |= get_bit(scalar->bytes, i); + bits = fiat_p256_get_bit(scalar->bytes, i + 192) << 3; + bits |= fiat_p256_get_bit(scalar->bytes, i + 128) << 2; + bits |= fiat_p256_get_bit(scalar->bytes, i + 64) << 1; + bits |= fiat_p256_get_bit(scalar->bytes, i); // Select the point to add, in constant time. - select_point(bits, 16, g_pre_comp[0], tmp); - point_add(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2], 1 /* mixed */, tmp[0], - tmp[1], tmp[2]); + fiat_p256_select_point(bits, 16, g_pre_comp[0], tmp); + fiat_p256_point_add(nq[0], nq[1], nq[2], nq[0], nq[1], nq[2], 1 /* mixed */, + tmp[0], tmp[1], tmp[2]); } - fe_to_generic(&r->X, nq[0]); - fe_to_generic(&r->Y, nq[1]); - fe_to_generic(&r->Z, nq[2]); + fiat_p256_to_generic(&r->X, nq[0]); + fiat_p256_to_generic(&r->Y, nq[1]); + fiat_p256_to_generic(&r->Z, nq[2]); } static void ec_GFp_nistp256_point_mul_public(const EC_GROUP *group, @@ -920,17 +927,18 @@ const EC_SCALAR *p_scalar) { #define P256_WSIZE_PUBLIC 4 // Precompute multiples of |p|. p_pre_comp[i] is (2*i+1) * |p|. - fe p_pre_comp[1 << (P256_WSIZE_PUBLIC-1)][3]; - fe_from_generic(p_pre_comp[0][0], &p->X); - fe_from_generic(p_pre_comp[0][1], &p->Y); - fe_from_generic(p_pre_comp[0][2], &p->Z); - fe p2[3]; - point_double(p2[0], p2[1], p2[2], p_pre_comp[0][0], p_pre_comp[0][1], - p_pre_comp[0][2]); + fiat_p256_felem p_pre_comp[1 << (P256_WSIZE_PUBLIC - 1)][3]; + fiat_p256_from_generic(p_pre_comp[0][0], &p->X); + fiat_p256_from_generic(p_pre_comp[0][1], &p->Y); + fiat_p256_from_generic(p_pre_comp[0][2], &p->Z); + fiat_p256_felem p2[3]; + fiat_p256_point_double(p2[0], p2[1], p2[2], p_pre_comp[0][0], + p_pre_comp[0][1], p_pre_comp[0][2]); for (size_t i = 1; i < OPENSSL_ARRAY_SIZE(p_pre_comp); i++) { - point_add(p_pre_comp[i][0], p_pre_comp[i][1], p_pre_comp[i][2], - p_pre_comp[i - 1][0], p_pre_comp[i - 1][1], p_pre_comp[i - 1][2], - 0 /* not mixed */, p2[0], p2[1], p2[2]); + fiat_p256_point_add(p_pre_comp[i][0], p_pre_comp[i][1], p_pre_comp[i][2], + p_pre_comp[i - 1][0], p_pre_comp[i - 1][1], + p_pre_comp[i - 1][2], 0 /* not mixed */, p2[0], p2[1], + p2[2]); } // Set up the coefficients for |p_scalar|. @@ -939,59 +947,60 @@ // Set |ret| to the point at infinity. int skip = 1; // Save some point operations. - fe ret[3] = {{0},{0},{0}}; + fiat_p256_felem ret[3] = {{0}, {0}, {0}}; for (int i = 256; i >= 0; i--) { if (!skip) { - point_double(ret[0], ret[1], ret[2], ret[0], ret[1], ret[2]); + fiat_p256_point_double(ret[0], ret[1], ret[2], ret[0], ret[1], ret[2]); } // For the |g_scalar|, we use the precomputed table without the // constant-time lookup. if (i <= 31) { // First, look 32 bits upwards. - uint64_t bits = get_bit(g_scalar->bytes, i + 224) << 3; - bits |= get_bit(g_scalar->bytes, i + 160) << 2; - bits |= get_bit(g_scalar->bytes, i + 96) << 1; - bits |= get_bit(g_scalar->bytes, i + 32); - point_add(ret[0], ret[1], ret[2], ret[0], ret[1], ret[2], 1 /* mixed */, - g_pre_comp[1][bits][0], g_pre_comp[1][bits][1], - g_pre_comp[1][bits][2]); + uint64_t bits = fiat_p256_get_bit(g_scalar->bytes, i + 224) << 3; + bits |= fiat_p256_get_bit(g_scalar->bytes, i + 160) << 2; + bits |= fiat_p256_get_bit(g_scalar->bytes, i + 96) << 1; + bits |= fiat_p256_get_bit(g_scalar->bytes, i + 32); + fiat_p256_point_add(ret[0], ret[1], ret[2], ret[0], ret[1], ret[2], + 1 /* mixed */, g_pre_comp[1][bits][0], + g_pre_comp[1][bits][1], g_pre_comp[1][bits][2]); skip = 0; // Second, look at the current position. - bits = get_bit(g_scalar->bytes, i + 192) << 3; - bits |= get_bit(g_scalar->bytes, i + 128) << 2; - bits |= get_bit(g_scalar->bytes, i + 64) << 1; - bits |= get_bit(g_scalar->bytes, i); - point_add(ret[0], ret[1], ret[2], ret[0], ret[1], ret[2], 1 /* mixed */, - g_pre_comp[0][bits][0], g_pre_comp[0][bits][1], - g_pre_comp[0][bits][2]); + bits = fiat_p256_get_bit(g_scalar->bytes, i + 192) << 3; + bits |= fiat_p256_get_bit(g_scalar->bytes, i + 128) << 2; + bits |= fiat_p256_get_bit(g_scalar->bytes, i + 64) << 1; + bits |= fiat_p256_get_bit(g_scalar->bytes, i); + fiat_p256_point_add(ret[0], ret[1], ret[2], ret[0], ret[1], ret[2], + 1 /* mixed */, g_pre_comp[0][bits][0], + g_pre_comp[0][bits][1], g_pre_comp[0][bits][2]); } int digit = p_wNAF[i]; if (digit != 0) { assert(digit & 1); int idx = digit < 0 ? (-digit) >> 1 : digit >> 1; - fe *y = &p_pre_comp[idx][1], tmp; + fiat_p256_felem *y = &p_pre_comp[idx][1], tmp; if (digit < 0) { - fe_opp(tmp, p_pre_comp[idx][1]); + fiat_p256_opp(tmp, p_pre_comp[idx][1]); y = &tmp; } if (!skip) { - point_add(ret[0], ret[1], ret[2], ret[0], ret[1], ret[2], - 0 /* not mixed */, p_pre_comp[idx][0], *y, p_pre_comp[idx][2]); + fiat_p256_point_add(ret[0], ret[1], ret[2], ret[0], ret[1], ret[2], + 0 /* not mixed */, p_pre_comp[idx][0], *y, + p_pre_comp[idx][2]); } else { - fe_copy(ret[0], p_pre_comp[idx][0]); - fe_copy(ret[1], *y); - fe_copy(ret[2], p_pre_comp[idx][2]); + fiat_p256_copy(ret[0], p_pre_comp[idx][0]); + fiat_p256_copy(ret[1], *y); + fiat_p256_copy(ret[2], p_pre_comp[idx][2]); skip = 0; } } } - fe_to_generic(&r->X, ret[0]); - fe_to_generic(&r->Y, ret[1]); - fe_to_generic(&r->Z, ret[2]); + fiat_p256_to_generic(&r->X, ret[0]); + fiat_p256_to_generic(&r->Y, ret[1]); + fiat_p256_to_generic(&r->Z, ret[2]); } static int ec_GFp_nistp256_cmp_x_coordinate(const EC_GROUP *group, @@ -1004,17 +1013,17 @@ // We wish to compare X/Z^2 with r. This is equivalent to comparing X with // r*Z^2. Note that X and Z are represented in Montgomery form, while r is // not. - fe Z2_mont; - fe_from_generic(Z2_mont, &p->Z); - fe_mul(Z2_mont, Z2_mont, Z2_mont); + fiat_p256_felem Z2_mont; + fiat_p256_from_generic(Z2_mont, &p->Z); + fiat_p256_mul(Z2_mont, Z2_mont, Z2_mont); - fe r_Z2; - fe_frombytes(r_Z2, r->bytes); // r < order < p, so this is valid. - fe_mul(r_Z2, r_Z2, Z2_mont); + fiat_p256_felem r_Z2; + fiat_p256_from_bytes(r_Z2, r->bytes); // r < order < p, so this is valid. + fiat_p256_mul(r_Z2, r_Z2, Z2_mont); - fe X; - fe_from_generic(X, &p->X); - fe_from_montgomery(X); + fiat_p256_felem X; + fiat_p256_from_generic(X, &p->X); + fiat_p256_from_montgomery(X, X); if (OPENSSL_memcmp(&r_Z2, &X, sizeof(r_Z2)) == 0) { return 1; @@ -1030,8 +1039,8 @@ // We can ignore the carry because: r + group_order < p < 2^256. EC_FELEM tmp; bn_add_words(tmp.words, r->words, group->order.d, group->order.width); - fe_from_generic(r_Z2, &tmp); - fe_mul(r_Z2, r_Z2, Z2_mont); + fiat_p256_from_generic(r_Z2, &tmp); + fiat_p256_mul(r_Z2, r_Z2, Z2_mont); if (OPENSSL_memcmp(&r_Z2, &X, sizeof(r_Z2)) == 0) { return 1; } @@ -1045,7 +1054,7 @@ out->group_finish = ec_GFp_mont_group_finish; out->group_set_curve = ec_GFp_mont_group_set_curve; out->point_get_affine_coordinates = - ec_GFp_nistp256_point_get_affine_coordinates; + ec_GFp_nistp256_point_get_affine_coordinates; out->add = ec_GFp_nistp256_add; out->dbl = ec_GFp_nistp256_dbl; out->mul = ec_GFp_nistp256_point_mul;