mirror of
https://gitlab.winehq.org/wine/wine-gecko.git
synced 2024-09-13 09:24:08 -07:00
430 lines
13 KiB
C
430 lines
13 KiB
C
/*
|
|
* ***** BEGIN LICENSE BLOCK *****
|
|
* Version: MPL 1.1/GPL 2.0/LGPL 2.1
|
|
*
|
|
* The contents of this file are subject to the Mozilla Public License Version
|
|
* 1.1 (the "License"); you may not use this file except in compliance with
|
|
* the License. You may obtain a copy of the License at
|
|
* http://www.mozilla.org/MPL/
|
|
*
|
|
* Software distributed under the License is distributed on an "AS IS" basis,
|
|
* WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
|
|
* for the specific language governing rights and limitations under the
|
|
* License.
|
|
*
|
|
* The Original Code is the elliptic curve math library for prime field curves.
|
|
*
|
|
* The Initial Developer of the Original Code is
|
|
* Sun Microsystems, Inc.
|
|
* Portions created by the Initial Developer are Copyright (C) 2003
|
|
* the Initial Developer. All Rights Reserved.
|
|
*
|
|
* Contributor(s):
|
|
* Douglas Stebila <douglas@stebila.ca>
|
|
*
|
|
* Alternatively, the contents of this file may be used under the terms of
|
|
* either the GNU General Public License Version 2 or later (the "GPL"), or
|
|
* the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
|
|
* in which case the provisions of the GPL or the LGPL are applicable instead
|
|
* of those above. If you wish to allow use of your version of this file only
|
|
* under the terms of either the GPL or the LGPL, and not to allow others to
|
|
* use your version of this file under the terms of the MPL, indicate your
|
|
* decision by deleting the provisions above and replace them with the notice
|
|
* and other provisions required by the GPL or the LGPL. If you do not delete
|
|
* the provisions above, a recipient may use your version of this file under
|
|
* the terms of any one of the MPL, the GPL or the LGPL.
|
|
*
|
|
* ***** END LICENSE BLOCK ***** */
|
|
|
|
#include "ecp.h"
|
|
#include "mpi.h"
|
|
#include "mplogic.h"
|
|
#include "mpi-priv.h"
|
|
#include <stdlib.h>
|
|
|
|
/* Fast modular reduction for p256 = 2^256 - 2^224 + 2^192+ 2^96 - 1. a can be r.
|
|
* Uses algorithm 2.29 from Hankerson, Menezes, Vanstone. Guide to
|
|
* Elliptic Curve Cryptography. */
|
|
mp_err
|
|
ec_GFp_nistp256_mod(const mp_int *a, mp_int *r, const GFMethod *meth)
|
|
{
|
|
mp_err res = MP_OKAY;
|
|
mp_size a_used = MP_USED(a);
|
|
int a_bits = mpl_significant_bits(a);
|
|
mp_digit carry;
|
|
|
|
#ifdef ECL_THIRTY_TWO_BIT
|
|
mp_digit a8=0, a9=0, a10=0, a11=0, a12=0, a13=0, a14=0, a15=0;
|
|
mp_digit r0, r1, r2, r3, r4, r5, r6, r7;
|
|
int r8; /* must be a signed value ! */
|
|
#else
|
|
mp_digit a4=0, a5=0, a6=0, a7=0;
|
|
mp_digit a4h, a4l, a5h, a5l, a6h, a6l, a7h, a7l;
|
|
mp_digit r0, r1, r2, r3;
|
|
int r4; /* must be a signed value ! */
|
|
#endif
|
|
/* for polynomials larger than twice the field size
|
|
* use regular reduction */
|
|
if (a_bits < 256) {
|
|
if (a == r) return MP_OKAY;
|
|
return mp_copy(a,r);
|
|
}
|
|
if (a_bits > 512) {
|
|
MP_CHECKOK(mp_mod(a, &meth->irr, r));
|
|
} else {
|
|
|
|
#ifdef ECL_THIRTY_TWO_BIT
|
|
switch (a_used) {
|
|
case 16:
|
|
a15 = MP_DIGIT(a,15);
|
|
case 15:
|
|
a14 = MP_DIGIT(a,14);
|
|
case 14:
|
|
a13 = MP_DIGIT(a,13);
|
|
case 13:
|
|
a12 = MP_DIGIT(a,12);
|
|
case 12:
|
|
a11 = MP_DIGIT(a,11);
|
|
case 11:
|
|
a10 = MP_DIGIT(a,10);
|
|
case 10:
|
|
a9 = MP_DIGIT(a,9);
|
|
case 9:
|
|
a8 = MP_DIGIT(a,8);
|
|
}
|
|
|
|
r0 = MP_DIGIT(a,0);
|
|
r1 = MP_DIGIT(a,1);
|
|
r2 = MP_DIGIT(a,2);
|
|
r3 = MP_DIGIT(a,3);
|
|
r4 = MP_DIGIT(a,4);
|
|
r5 = MP_DIGIT(a,5);
|
|
r6 = MP_DIGIT(a,6);
|
|
r7 = MP_DIGIT(a,7);
|
|
|
|
/* sum 1 */
|
|
MP_ADD_CARRY(r3, a11, r3, 0, carry);
|
|
MP_ADD_CARRY(r4, a12, r4, carry, carry);
|
|
MP_ADD_CARRY(r5, a13, r5, carry, carry);
|
|
MP_ADD_CARRY(r6, a14, r6, carry, carry);
|
|
MP_ADD_CARRY(r7, a15, r7, carry, carry);
|
|
r8 = carry;
|
|
MP_ADD_CARRY(r3, a11, r3, 0, carry);
|
|
MP_ADD_CARRY(r4, a12, r4, carry, carry);
|
|
MP_ADD_CARRY(r5, a13, r5, carry, carry);
|
|
MP_ADD_CARRY(r6, a14, r6, carry, carry);
|
|
MP_ADD_CARRY(r7, a15, r7, carry, carry);
|
|
r8 += carry;
|
|
/* sum 2 */
|
|
MP_ADD_CARRY(r3, a12, r3, 0, carry);
|
|
MP_ADD_CARRY(r4, a13, r4, carry, carry);
|
|
MP_ADD_CARRY(r5, a14, r5, carry, carry);
|
|
MP_ADD_CARRY(r6, a15, r6, carry, carry);
|
|
MP_ADD_CARRY(r7, 0, r7, carry, carry);
|
|
r8 += carry;
|
|
/* combine last bottom of sum 3 with second sum 2 */
|
|
MP_ADD_CARRY(r0, a8, r0, 0, carry);
|
|
MP_ADD_CARRY(r1, a9, r1, carry, carry);
|
|
MP_ADD_CARRY(r2, a10, r2, carry, carry);
|
|
MP_ADD_CARRY(r3, a12, r3, carry, carry);
|
|
MP_ADD_CARRY(r4, a13, r4, carry, carry);
|
|
MP_ADD_CARRY(r5, a14, r5, carry, carry);
|
|
MP_ADD_CARRY(r6, a15, r6, carry, carry);
|
|
MP_ADD_CARRY(r7, a15, r7, carry, carry); /* from sum 3 */
|
|
r8 += carry;
|
|
/* sum 3 (rest of it)*/
|
|
MP_ADD_CARRY(r6, a14, r6, 0, carry);
|
|
MP_ADD_CARRY(r7, 0, r7, carry, carry);
|
|
r8 += carry;
|
|
/* sum 4 (rest of it)*/
|
|
MP_ADD_CARRY(r0, a9, r0, 0, carry);
|
|
MP_ADD_CARRY(r1, a10, r1, carry, carry);
|
|
MP_ADD_CARRY(r2, a11, r2, carry, carry);
|
|
MP_ADD_CARRY(r3, a13, r3, carry, carry);
|
|
MP_ADD_CARRY(r4, a14, r4, carry, carry);
|
|
MP_ADD_CARRY(r5, a15, r5, carry, carry);
|
|
MP_ADD_CARRY(r6, a13, r6, carry, carry);
|
|
MP_ADD_CARRY(r7, a8, r7, carry, carry);
|
|
r8 += carry;
|
|
/* diff 5 */
|
|
MP_SUB_BORROW(r0, a11, r0, 0, carry);
|
|
MP_SUB_BORROW(r1, a12, r1, carry, carry);
|
|
MP_SUB_BORROW(r2, a13, r2, carry, carry);
|
|
MP_SUB_BORROW(r3, 0, r3, carry, carry);
|
|
MP_SUB_BORROW(r4, 0, r4, carry, carry);
|
|
MP_SUB_BORROW(r5, 0, r5, carry, carry);
|
|
MP_SUB_BORROW(r6, a8, r6, carry, carry);
|
|
MP_SUB_BORROW(r7, a10, r7, carry, carry);
|
|
r8 -= carry;
|
|
/* diff 6 */
|
|
MP_SUB_BORROW(r0, a12, r0, 0, carry);
|
|
MP_SUB_BORROW(r1, a13, r1, carry, carry);
|
|
MP_SUB_BORROW(r2, a14, r2, carry, carry);
|
|
MP_SUB_BORROW(r3, a15, r3, carry, carry);
|
|
MP_SUB_BORROW(r4, 0, r4, carry, carry);
|
|
MP_SUB_BORROW(r5, 0, r5, carry, carry);
|
|
MP_SUB_BORROW(r6, a9, r6, carry, carry);
|
|
MP_SUB_BORROW(r7, a11, r7, carry, carry);
|
|
r8 -= carry;
|
|
/* diff 7 */
|
|
MP_SUB_BORROW(r0, a13, r0, 0, carry);
|
|
MP_SUB_BORROW(r1, a14, r1, carry, carry);
|
|
MP_SUB_BORROW(r2, a15, r2, carry, carry);
|
|
MP_SUB_BORROW(r3, a8, r3, carry, carry);
|
|
MP_SUB_BORROW(r4, a9, r4, carry, carry);
|
|
MP_SUB_BORROW(r5, a10, r5, carry, carry);
|
|
MP_SUB_BORROW(r6, 0, r6, carry, carry);
|
|
MP_SUB_BORROW(r7, a12, r7, carry, carry);
|
|
r8 -= carry;
|
|
/* diff 8 */
|
|
MP_SUB_BORROW(r0, a14, r0, 0, carry);
|
|
MP_SUB_BORROW(r1, a15, r1, carry, carry);
|
|
MP_SUB_BORROW(r2, 0, r2, carry, carry);
|
|
MP_SUB_BORROW(r3, a9, r3, carry, carry);
|
|
MP_SUB_BORROW(r4, a10, r4, carry, carry);
|
|
MP_SUB_BORROW(r5, a11, r5, carry, carry);
|
|
MP_SUB_BORROW(r6, 0, r6, carry, carry);
|
|
MP_SUB_BORROW(r7, a13, r7, carry, carry);
|
|
r8 -= carry;
|
|
|
|
/* reduce the overflows */
|
|
while (r8 > 0) {
|
|
mp_digit r8_d = r8;
|
|
MP_ADD_CARRY(r0, r8_d, r0, 0, carry);
|
|
MP_ADD_CARRY(r1, 0, r1, carry, carry);
|
|
MP_ADD_CARRY(r2, 0, r2, carry, carry);
|
|
MP_ADD_CARRY(r3, -r8_d, r3, carry, carry);
|
|
MP_ADD_CARRY(r4, MP_DIGIT_MAX, r4, carry, carry);
|
|
MP_ADD_CARRY(r5, MP_DIGIT_MAX, r5, carry, carry);
|
|
MP_ADD_CARRY(r6, -(r8_d+1), r6, carry, carry);
|
|
MP_ADD_CARRY(r7, (r8_d-1), r7, carry, carry);
|
|
r8 = carry;
|
|
}
|
|
|
|
/* reduce the underflows */
|
|
while (r8 < 0) {
|
|
mp_digit r8_d = -r8;
|
|
MP_SUB_BORROW(r0, r8_d, r0, 0, carry);
|
|
MP_SUB_BORROW(r1, 0, r1, carry, carry);
|
|
MP_SUB_BORROW(r2, 0, r2, carry, carry);
|
|
MP_SUB_BORROW(r3, -r8_d, r3, carry, carry);
|
|
MP_SUB_BORROW(r4, MP_DIGIT_MAX, r4, carry, carry);
|
|
MP_SUB_BORROW(r5, MP_DIGIT_MAX, r5, carry, carry);
|
|
MP_SUB_BORROW(r6, -(r8_d+1), r6, carry, carry);
|
|
MP_SUB_BORROW(r7, (r8_d-1), r7, carry, carry);
|
|
r8 = -carry;
|
|
}
|
|
if (a != r) {
|
|
MP_CHECKOK(s_mp_pad(r,8));
|
|
}
|
|
MP_SIGN(r) = MP_ZPOS;
|
|
MP_USED(r) = 8;
|
|
|
|
MP_DIGIT(r,7) = r7;
|
|
MP_DIGIT(r,6) = r6;
|
|
MP_DIGIT(r,5) = r5;
|
|
MP_DIGIT(r,4) = r4;
|
|
MP_DIGIT(r,3) = r3;
|
|
MP_DIGIT(r,2) = r2;
|
|
MP_DIGIT(r,1) = r1;
|
|
MP_DIGIT(r,0) = r0;
|
|
|
|
/* final reduction if necessary */
|
|
if ((r7 == MP_DIGIT_MAX) &&
|
|
((r6 > 1) || ((r6 == 1) &&
|
|
(r5 || r4 || r3 ||
|
|
((r2 == MP_DIGIT_MAX) && (r1 == MP_DIGIT_MAX)
|
|
&& (r0 == MP_DIGIT_MAX)))))) {
|
|
MP_CHECKOK(mp_sub(r, &meth->irr, r));
|
|
}
|
|
#ifdef notdef
|
|
|
|
|
|
/* smooth the negatives */
|
|
while (MP_SIGN(r) != MP_ZPOS) {
|
|
MP_CHECKOK(mp_add(r, &meth->irr, r));
|
|
}
|
|
while (MP_USED(r) > 8) {
|
|
MP_CHECKOK(mp_sub(r, &meth->irr, r));
|
|
}
|
|
|
|
/* final reduction if necessary */
|
|
if (MP_DIGIT(r,7) >= MP_DIGIT(&meth->irr,7)) {
|
|
if (mp_cmp(r,&meth->irr) != MP_LT) {
|
|
MP_CHECKOK(mp_sub(r, &meth->irr, r));
|
|
}
|
|
}
|
|
#endif
|
|
s_mp_clamp(r);
|
|
#else
|
|
switch (a_used) {
|
|
case 8:
|
|
a7 = MP_DIGIT(a,7);
|
|
case 7:
|
|
a6 = MP_DIGIT(a,6);
|
|
case 6:
|
|
a5 = MP_DIGIT(a,5);
|
|
case 5:
|
|
a4 = MP_DIGIT(a,4);
|
|
}
|
|
a7l = a7 << 32;
|
|
a7h = a7 >> 32;
|
|
a6l = a6 << 32;
|
|
a6h = a6 >> 32;
|
|
a5l = a5 << 32;
|
|
a5h = a5 >> 32;
|
|
a4l = a4 << 32;
|
|
a4h = a4 >> 32;
|
|
r3 = MP_DIGIT(a,3);
|
|
r2 = MP_DIGIT(a,2);
|
|
r1 = MP_DIGIT(a,1);
|
|
r0 = MP_DIGIT(a,0);
|
|
|
|
/* sum 1 */
|
|
MP_ADD_CARRY(r1, a5h << 32, r1, 0, carry);
|
|
MP_ADD_CARRY(r2, a6, r2, carry, carry);
|
|
MP_ADD_CARRY(r3, a7, r3, carry, carry);
|
|
r4 = carry;
|
|
MP_ADD_CARRY(r1, a5h << 32, r1, 0, carry);
|
|
MP_ADD_CARRY(r2, a6, r2, carry, carry);
|
|
MP_ADD_CARRY(r3, a7, r3, carry, carry);
|
|
r4 += carry;
|
|
/* sum 2 */
|
|
MP_ADD_CARRY(r1, a6l, r1, 0, carry);
|
|
MP_ADD_CARRY(r2, a6h | a7l, r2, carry, carry);
|
|
MP_ADD_CARRY(r3, a7h, r3, carry, carry);
|
|
r4 += carry;
|
|
MP_ADD_CARRY(r1, a6l, r1, 0, carry);
|
|
MP_ADD_CARRY(r2, a6h | a7l, r2, carry, carry);
|
|
MP_ADD_CARRY(r3, a7h, r3, carry, carry);
|
|
r4 += carry;
|
|
|
|
/* sum 3 */
|
|
MP_ADD_CARRY(r0, a4, r0, 0, carry);
|
|
MP_ADD_CARRY(r1, a5l >> 32, r1, carry, carry);
|
|
MP_ADD_CARRY(r2, 0, r2, carry, carry);
|
|
MP_ADD_CARRY(r3, a7, r3, carry, carry);
|
|
r4 += carry;
|
|
/* sum 4 */
|
|
MP_ADD_CARRY(r0, a4h | a5l, r0, 0, carry);
|
|
MP_ADD_CARRY(r1, a5h|(a6h<<32), r1, carry, carry);
|
|
MP_ADD_CARRY(r2, a7, r2, carry, carry);
|
|
MP_ADD_CARRY(r3, a6h | a4l, r3, carry, carry);
|
|
r4 += carry;
|
|
/* diff 5 */
|
|
MP_SUB_BORROW(r0, a5h | a6l, r0, 0, carry);
|
|
MP_SUB_BORROW(r1, a6h, r1, carry, carry);
|
|
MP_SUB_BORROW(r2, 0, r2, carry, carry);
|
|
MP_SUB_BORROW(r3, (a4l>>32)|a5l,r3, carry, carry);
|
|
r4 -= carry;
|
|
/* diff 6 */
|
|
MP_SUB_BORROW(r0, a6, r0, 0, carry);
|
|
MP_SUB_BORROW(r1, a7, r1, carry, carry);
|
|
MP_SUB_BORROW(r2, 0, r2, carry, carry);
|
|
MP_SUB_BORROW(r3, a4h|(a5h<<32),r3, carry, carry);
|
|
r4 -= carry;
|
|
/* diff 7 */
|
|
MP_SUB_BORROW(r0, a6h|a7l, r0, 0, carry);
|
|
MP_SUB_BORROW(r1, a7h|a4l, r1, carry, carry);
|
|
MP_SUB_BORROW(r2, a4h|a5l, r2, carry, carry);
|
|
MP_SUB_BORROW(r3, a6l, r3, carry, carry);
|
|
r4 -= carry;
|
|
/* diff 8 */
|
|
MP_SUB_BORROW(r0, a7, r0, 0, carry);
|
|
MP_SUB_BORROW(r1, a4h<<32, r1, carry, carry);
|
|
MP_SUB_BORROW(r2, a5, r2, carry, carry);
|
|
MP_SUB_BORROW(r3, a6h<<32, r3, carry, carry);
|
|
r4 -= carry;
|
|
|
|
/* reduce the overflows */
|
|
while (r4 > 0) {
|
|
mp_digit r4_long = r4;
|
|
mp_digit r4l = (r4_long << 32);
|
|
MP_ADD_CARRY(r0, r4_long, r0, 0, carry);
|
|
MP_ADD_CARRY(r1, -r4l, r1, carry, carry);
|
|
MP_ADD_CARRY(r2, MP_DIGIT_MAX, r2, carry, carry);
|
|
MP_ADD_CARRY(r3, r4l-r4_long-1,r3, carry, carry);
|
|
r4 = carry;
|
|
}
|
|
|
|
/* reduce the underflows */
|
|
while (r4 < 0) {
|
|
mp_digit r4_long = -r4;
|
|
mp_digit r4l = (r4_long << 32);
|
|
MP_SUB_BORROW(r0, r4_long, r0, 0, carry);
|
|
MP_SUB_BORROW(r1, -r4l, r1, carry, carry);
|
|
MP_SUB_BORROW(r2, MP_DIGIT_MAX, r2, carry, carry);
|
|
MP_SUB_BORROW(r3, r4l-r4_long-1,r3, carry, carry);
|
|
r4 = -carry;
|
|
}
|
|
|
|
if (a != r) {
|
|
MP_CHECKOK(s_mp_pad(r,4));
|
|
}
|
|
MP_SIGN(r) = MP_ZPOS;
|
|
MP_USED(r) = 4;
|
|
|
|
MP_DIGIT(r,3) = r3;
|
|
MP_DIGIT(r,2) = r2;
|
|
MP_DIGIT(r,1) = r1;
|
|
MP_DIGIT(r,0) = r0;
|
|
|
|
/* final reduction if necessary */
|
|
if ((r3 > 0xFFFFFFFF00000001ULL) ||
|
|
((r3 == 0xFFFFFFFF00000001ULL) &&
|
|
(r2 || (r1 >> 32)||
|
|
(r1 == 0xFFFFFFFFULL && r0 == MP_DIGIT_MAX)))) {
|
|
/* very rare, just use mp_sub */
|
|
MP_CHECKOK(mp_sub(r, &meth->irr, r));
|
|
}
|
|
|
|
s_mp_clamp(r);
|
|
#endif
|
|
}
|
|
|
|
CLEANUP:
|
|
return res;
|
|
}
|
|
|
|
/* Compute the square of polynomial a, reduce modulo p256. Store the
|
|
* result in r. r could be a. Uses optimized modular reduction for p256.
|
|
*/
|
|
mp_err
|
|
ec_GFp_nistp256_sqr(const mp_int *a, mp_int *r, const GFMethod *meth)
|
|
{
|
|
mp_err res = MP_OKAY;
|
|
|
|
MP_CHECKOK(mp_sqr(a, r));
|
|
MP_CHECKOK(ec_GFp_nistp256_mod(r, r, meth));
|
|
CLEANUP:
|
|
return res;
|
|
}
|
|
|
|
/* Compute the product of two polynomials a and b, reduce modulo p256.
|
|
* Store the result in r. r could be a or b; a could be b. Uses
|
|
* optimized modular reduction for p256. */
|
|
mp_err
|
|
ec_GFp_nistp256_mul(const mp_int *a, const mp_int *b, mp_int *r,
|
|
const GFMethod *meth)
|
|
{
|
|
mp_err res = MP_OKAY;
|
|
|
|
MP_CHECKOK(mp_mul(a, b, r));
|
|
MP_CHECKOK(ec_GFp_nistp256_mod(r, r, meth));
|
|
CLEANUP:
|
|
return res;
|
|
}
|
|
|
|
/* Wire in fast field arithmetic and precomputation of base point for
|
|
* named curves. */
|
|
mp_err
|
|
ec_group_set_gfp256(ECGroup *group, ECCurveName name)
|
|
{
|
|
if (name == ECCurve_NIST_P256) {
|
|
group->meth->field_mod = &ec_GFp_nistp256_mod;
|
|
group->meth->field_mul = &ec_GFp_nistp256_mul;
|
|
group->meth->field_sqr = &ec_GFp_nistp256_sqr;
|
|
}
|
|
return MP_OKAY;
|
|
}
|