Fix overflowing multiplication in FrodoKEM AES

このコミットが含まれているのは:
Thom Wiggers 2020-11-17 12:25:34 +01:00
コミット 3efcd2b186
この署名に対応する既知のキーがデータベースに存在しません
GPGキーID: 001BB0A7CE26E363
3個のファイルの変更13行の追加13行の削除

ファイルの表示

@ -110,10 +110,10 @@ int PQCLEAN_FRODOKEM1344AES_OPT_mul_add_sa_plus_e(uint16_t *out, const uint16_t
uint16_t sum[PARAMS_PARALLEL] = {0};
for (j = 0; j < PARAMS_N; j++) { // Matrix-vector multiplication
uint16_t sp = s[i * PARAMS_N + j];
sum[0] += sp * a_cols_t[(k + 0) * PARAMS_N + j];
sum[1] += sp * a_cols_t[(k + 1) * PARAMS_N + j];
sum[2] += sp * a_cols_t[(k + 2) * PARAMS_N + j];
sum[3] += sp * a_cols_t[(k + 3) * PARAMS_N + j];
sum[0] += (uint16_t)(sp * (uint32_t)a_cols_t[(k + 0) * PARAMS_N + j]);
sum[1] += (uint16_t)(sp * (uint32_t)a_cols_t[(k + 1) * PARAMS_N + j]);
sum[2] += (uint16_t)(sp * (uint32_t)a_cols_t[(k + 2) * PARAMS_N + j]);
sum[3] += (uint16_t)(sp * (uint32_t)a_cols_t[(k + 3) * PARAMS_N + j]);
}
out[i * PARAMS_N + kk + k + 0] += sum[0];
out[i * PARAMS_N + kk + k + 2] += sum[2];

ファイルの表示

@ -80,7 +80,7 @@ int PQCLEAN_FRODOKEM640AES_OPT_mul_add_sa_plus_e(uint16_t *out, const uint16_t *
*((uint32_t *)&out[i]) = *((uint32_t *)&e[i]);
}
int k;
size_t k;
uint16_t a_cols[PARAMS_N * PARAMS_STRIPE_STEP] = {0};
uint16_t a_cols_t[PARAMS_N * PARAMS_STRIPE_STEP];
uint16_t a_cols_temp[PARAMS_N * PARAMS_STRIPE_STEP] = {0};
@ -110,10 +110,10 @@ int PQCLEAN_FRODOKEM640AES_OPT_mul_add_sa_plus_e(uint16_t *out, const uint16_t *
uint16_t sum[PARAMS_PARALLEL] = {0};
for (j = 0; j < PARAMS_N; j++) { // Matrix-vector multiplication
uint16_t sp = s[i * PARAMS_N + j];
sum[0] += sp * a_cols_t[(k + 0) * PARAMS_N + j];
sum[1] += sp * a_cols_t[(k + 1) * PARAMS_N + j];
sum[2] += sp * a_cols_t[(k + 2) * PARAMS_N + j];
sum[3] += sp * a_cols_t[(k + 3) * PARAMS_N + j];
sum[0] += (uint16_t)(sp * (uint32_t)a_cols_t[(k + 0) * PARAMS_N + j]);
sum[1] += (uint16_t)(sp * (uint32_t)a_cols_t[(k + 1) * PARAMS_N + j]);
sum[2] += (uint16_t)(sp * (uint32_t)a_cols_t[(k + 2) * PARAMS_N + j]);
sum[3] += (uint16_t)(sp * (uint32_t)a_cols_t[(k + 3) * PARAMS_N + j]);
}
out[i * PARAMS_N + kk + k + 0] += sum[0];
out[i * PARAMS_N + kk + k + 2] += sum[2];

ファイルの表示

@ -110,10 +110,10 @@ int PQCLEAN_FRODOKEM976AES_OPT_mul_add_sa_plus_e(uint16_t *out, const uint16_t *
uint16_t sum[PARAMS_PARALLEL] = {0};
for (j = 0; j < PARAMS_N; j++) { // Matrix-vector multiplication
uint16_t sp = s[i * PARAMS_N + j];
sum[0] += sp * a_cols_t[(k + 0) * PARAMS_N + j];
sum[1] += sp * a_cols_t[(k + 1) * PARAMS_N + j];
sum[2] += sp * a_cols_t[(k + 2) * PARAMS_N + j];
sum[3] += sp * a_cols_t[(k + 3) * PARAMS_N + j];
sum[0] += (uint16_t)(sp * (uint32_t)a_cols_t[(k + 0) * PARAMS_N + j]);
sum[1] += (uint16_t)(sp * (uint32_t)a_cols_t[(k + 1) * PARAMS_N + j]);
sum[2] += (uint16_t)(sp * (uint32_t)a_cols_t[(k + 2) * PARAMS_N + j]);
sum[3] += (uint16_t)(sp * (uint32_t)a_cols_t[(k + 3) * PARAMS_N + j]);
}
out[i * PARAMS_N + kk + k + 0] += sum[0];
out[i * PARAMS_N + kk + k + 2] += sum[2];