| Message ID | 20181111174002.30929-1-vt@altlinux.org (mailing list archive) |
|---|---|
| State | Accepted |
| Delegated to: | Herbert Xu |
| Headers | show |
| Series | [v2] crypto: ecc - regularize scalar for scalar multiplication | expand |
On Sun, Nov 11, 2018 at 08:40:02PM +0300, Vitaly Chikunov wrote: > ecc_point_mult is supposed to be used with a regularized scalar, > otherwise, it's possible to deduce the position of the top bit of the > scalar with timing attack. This is important when the scalar is a > private key. > > ecc_point_mult is already using a regular algorithm (i.e. having an > operation flow independent of the input scalar) but regularization step > is not implemented. > > Arrange scalar to always have fixed top bit by adding a multiple of the > curve order (n). > > References: > The constant time regularization step is based on micro-ecc by Kenneth > MacKay and also referenced in the literature (Bernstein, D. J., & Lange, > T. (2017). Montgomery curves and the Montgomery ladder. (Cryptology > ePrint Archive; Vol. 2017/293). s.l.: IACR. Chapter 4.6.2.) > > Signed-off-by: Vitaly Chikunov <vt@altlinux.org> > Cc: kernel-hardening@lists.openwall.com > --- > > Changes from v1: > - No code changes, only description updates to be more informative. > > crypto/ecc.c | 16 ++++++++++++---- > 1 file changed, 12 insertions(+), 4 deletions(-) Patch applied. Thanks.
diff --git a/crypto/ecc.c b/crypto/ecc.c index 8facafd67802..adcce310f646 100644 --- a/crypto/ecc.c +++ b/crypto/ecc.c @@ -842,15 +842,23 @@ static void xycz_add_c(u64 *x1, u64 *y1, u64 *x2, u64 *y2, u64 *curve_prime, static void ecc_point_mult(struct ecc_point *result, const struct ecc_point *point, const u64 *scalar, - u64 *initial_z, u64 *curve_prime, + u64 *initial_z, const struct ecc_curve *curve, unsigned int ndigits) { /* R0 and R1 */ u64 rx[2][ECC_MAX_DIGITS]; u64 ry[2][ECC_MAX_DIGITS]; u64 z[ECC_MAX_DIGITS]; + u64 sk[2][ECC_MAX_DIGITS]; + u64 *curve_prime = curve->p; int i, nb; - int num_bits = vli_num_bits(scalar, ndigits); + int num_bits; + int carry; + + carry = vli_add(sk[0], scalar, curve->n, ndigits); + vli_add(sk[1], sk[0], curve->n, ndigits); + scalar = sk[!carry]; + num_bits = sizeof(u64) * ndigits * 8 + 1; vli_set(rx[1], point->x, ndigits); vli_set(ry[1], point->y, ndigits); @@ -1004,7 +1012,7 @@ int ecc_make_pub_key(unsigned int curve_id, unsigned int ndigits, goto out; } - ecc_point_mult(pk, &curve->g, priv, NULL, curve->p, ndigits); + ecc_point_mult(pk, &curve->g, priv, NULL, curve, ndigits); if (ecc_point_is_zero(pk)) { ret = -EAGAIN; goto err_free_point; @@ -1090,7 +1098,7 @@ int crypto_ecdh_shared_secret(unsigned int curve_id, unsigned int ndigits, goto err_alloc_product; } - ecc_point_mult(product, pk, priv, rand_z, curve->p, ndigits); + ecc_point_mult(product, pk, priv, rand_z, curve, ndigits); ecc_swap_digits(product->x, secret, ndigits);
ecc_point_mult is supposed to be used with a regularized scalar, otherwise, it's possible to deduce the position of the top bit of the scalar with timing attack. This is important when the scalar is a private key. ecc_point_mult is already using a regular algorithm (i.e. having an operation flow independent of the input scalar) but regularization step is not implemented. Arrange scalar to always have fixed top bit by adding a multiple of the curve order (n). References: The constant time regularization step is based on micro-ecc by Kenneth MacKay and also referenced in the literature (Bernstein, D. J., & Lange, T. (2017). Montgomery curves and the Montgomery ladder. (Cryptology ePrint Archive; Vol. 2017/293). s.l.: IACR. Chapter 4.6.2.) Signed-off-by: Vitaly Chikunov <vt@altlinux.org> Cc: kernel-hardening@lists.openwall.com --- Changes from v1: - No code changes, only description updates to be more informative. crypto/ecc.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-)