wolfSSL 3.11.1 for TLS1.3 beta
Fork of wolfSSL by
Diff: wolfcrypt/src/ge_low_mem.c
- Revision:
- 13:80fb167dafdf
- Parent:
- 11:cee25a834751
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/wolfcrypt/src/ge_low_mem.c Tue May 30 06:16:19 2017 +0000 @@ -0,0 +1,565 @@ +/* ge_low_mem.c + * + * Copyright (C) 2006-2016 wolfSSL Inc. + * + * This file is part of wolfSSL. + * + * wolfSSL is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * wolfSSL is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1335, USA + */ + + + /* Based from Daniel Beer's public domain work. */ + +#ifdef HAVE_CONFIG_H + #include <config.h> +#endif + +#include <wolfssl/wolfcrypt/settings.h> + +#if defined(CURVED25519_SMALL) /* use slower code that takes less memory */ +#if defined(HAVE_ED25519) + +#include <wolfssl/wolfcrypt/ge_operations.h> +#include <wolfssl/wolfcrypt/error-crypt.h> +#ifdef NO_INLINE + #include <wolfssl/wolfcrypt/misc.h> +#else + #define WOLFSSL_MISC_INCLUDED + #include <wolfcrypt/src/misc.c> +#endif + +void ed25519_smult(ge_p3 *r, const ge_p3 *a, const byte *e); +void ed25519_add(ge_p3 *r, const ge_p3 *a, const ge_p3 *b); +void ed25519_double(ge_p3 *r, const ge_p3 *a); + + +static const byte ed25519_order[F25519_SIZE] = { + 0xed, 0xd3, 0xf5, 0x5c, 0x1a, 0x63, 0x12, 0x58, + 0xd6, 0x9c, 0xf7, 0xa2, 0xde, 0xf9, 0xde, 0x14, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x10 +}; + +/*Arithmetic modulo the group order m = 2^252 + + 27742317777372353535851937790883648493 = + 7237005577332262213973186563042994240857116359379907606001950938285454250989 */ + +static const word32 m[32] = { + 0xED,0xD3,0xF5,0x5C,0x1A,0x63,0x12,0x58,0xD6,0x9C,0xF7,0xA2,0xDE,0xF9, + 0xDE,0x14,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, + 0x00,0x00,0x00,0x10 +}; + +static const word32 mu[33] = { + 0x1B,0x13,0x2C,0x0A,0xA3,0xE5,0x9C,0xED,0xA7,0x29,0x63,0x08,0x5D,0x21, + 0x06,0x21,0xEB,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF, + 0xFF,0xFF,0xFF,0xFF,0x0F +}; + + +int ge_compress_key(byte* out, const byte* xIn, const byte* yIn, + word32 keySz) +{ + byte tmp[F25519_SIZE]; + byte parity; + byte pt[32]; + int i; + + fe_copy(tmp, xIn); + parity = (tmp[0] & 1) << 7; + + fe_copy(pt, yIn); + pt[31] |= parity; + + for(i = 0; i < 32; i++) { + out[32-i-1] = pt[i]; + } + (void)keySz; + return 0; +} + + +static word32 lt(word32 a,word32 b) /* 16-bit inputs */ +{ + unsigned int x = a; + x -= (unsigned int) b; /* 0..65535: no; 4294901761..4294967295: yes */ + x >>= 31; /* 0: no; 1: yes */ + return x; +} + + +/* Reduce coefficients of r before calling reduce_add_sub */ +static void reduce_add_sub(word32 *r) +{ + word32 pb = 0; + word32 b; + word32 mask; + int i; + unsigned char t[32]; + + for(i=0;i<32;i++) + { + pb += m[i]; + b = lt(r[i],pb); + t[i] = r[i]-pb+(b<<8); + pb = b; + } + mask = b - 1; + for(i=0;i<32;i++) + r[i] ^= mask & (r[i] ^ t[i]); +} + + +/* Reduce coefficients of x before calling barrett_reduce */ +static void barrett_reduce(word32* r, word32 x[64]) +{ + /* See HAC, Alg. 14.42 */ + int i,j; + word32 q2[66]; + word32 *q3 = q2 + 33; + word32 r1[33]; + word32 r2[33]; + word32 carry; + word32 pb = 0; + word32 b; + + for (i = 0;i < 66;++i) q2[i] = 0; + for (i = 0;i < 33;++i) r2[i] = 0; + + for(i=0;i<33;i++) + for(j=0;j<33;j++) + if(i+j >= 31) q2[i+j] += mu[i]*x[j+31]; + carry = q2[31] >> 8; + q2[32] += carry; + carry = q2[32] >> 8; + q2[33] += carry; + + for(i=0;i<33;i++)r1[i] = x[i]; + for(i=0;i<32;i++) + for(j=0;j<33;j++) + if(i+j < 33) r2[i+j] += m[i]*q3[j]; + + for(i=0;i<32;i++) + { + carry = r2[i] >> 8; + r2[i+1] += carry; + r2[i] &= 0xff; + } + + for(i=0;i<32;i++) + { + pb += r2[i]; + b = lt(r1[i],pb); + r[i] = r1[i]-pb+(b<<8); + pb = b; + } + + /* XXX: Can it really happen that r<0?, See HAC, Alg 14.42, Step 3 + * r is an unsigned type. + * If so: Handle it here! + */ + + reduce_add_sub(r); + reduce_add_sub(r); +} + + +void sc_reduce(unsigned char x[64]) +{ + int i; + word32 t[64]; + word32 r[32]; + for(i=0;i<64;i++) t[i] = x[i]; + barrett_reduce(r, t); + for(i=0;i<32;i++) x[i] = (r[i] & 0xFF); +} + + +void sc_muladd(byte* out, const byte* a, const byte* b, const byte* c) +{ + + byte s[32]; + byte e[64]; + + XMEMSET(e, 0, sizeof(e)); + XMEMCPY(e, b, 32); + + /* Obtain e */ + sc_reduce(e); + + /* Compute s = ze + k */ + fprime_mul(s, a, e, ed25519_order); + fprime_add(s, c, ed25519_order); + + XMEMCPY(out, s, 32); +} + + +/* Base point is (numbers wrapped): + * + * x = 151122213495354007725011514095885315114 + * 54012693041857206046113283949847762202 + * y = 463168356949264781694283940034751631413 + * 07993866256225615783033603165251855960 + * + * y is derived by transforming the original Montgomery base (u=9). x + * is the corresponding positive coordinate for the new curve equation. + * t is x*y. + */ +const ge_p3 ed25519_base = { + { + 0x1a, 0xd5, 0x25, 0x8f, 0x60, 0x2d, 0x56, 0xc9, + 0xb2, 0xa7, 0x25, 0x95, 0x60, 0xc7, 0x2c, 0x69, + 0x5c, 0xdc, 0xd6, 0xfd, 0x31, 0xe2, 0xa4, 0xc0, + 0xfe, 0x53, 0x6e, 0xcd, 0xd3, 0x36, 0x69, 0x21 + }, + { + 0x58, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, + 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, + 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, + 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66 + }, + {1, 0}, + { + 0xa3, 0xdd, 0xb7, 0xa5, 0xb3, 0x8a, 0xde, 0x6d, + 0xf5, 0x52, 0x51, 0x77, 0x80, 0x9f, 0xf0, 0x20, + 0x7d, 0xe3, 0xab, 0x64, 0x8e, 0x4e, 0xea, 0x66, + 0x65, 0x76, 0x8b, 0xd7, 0x0f, 0x5f, 0x87, 0x67 + }, + +}; + + +const ge_p3 ed25519_neutral = { + {0}, + {1, 0}, + {1, 0}, + {0}, + +}; + + +static const byte ed25519_d[F25519_SIZE] = { + 0xa3, 0x78, 0x59, 0x13, 0xca, 0x4d, 0xeb, 0x75, + 0xab, 0xd8, 0x41, 0x41, 0x4d, 0x0a, 0x70, 0x00, + 0x98, 0xe8, 0x79, 0x77, 0x79, 0x40, 0xc7, 0x8c, + 0x73, 0xfe, 0x6f, 0x2b, 0xee, 0x6c, 0x03, 0x52 +}; + + +/* k = 2d */ +static const byte ed25519_k[F25519_SIZE] = { + 0x59, 0xf1, 0xb2, 0x26, 0x94, 0x9b, 0xd6, 0xeb, + 0x56, 0xb1, 0x83, 0x82, 0x9a, 0x14, 0xe0, 0x00, + 0x30, 0xd1, 0xf3, 0xee, 0xf2, 0x80, 0x8e, 0x19, + 0xe7, 0xfc, 0xdf, 0x56, 0xdc, 0xd9, 0x06, 0x24 +}; + + +void ed25519_add(ge_p3 *r, + const ge_p3 *p1, const ge_p3 *p2) +{ + /* Explicit formulas database: add-2008-hwcd-3 + * + * source 2008 Hisil--Wong--Carter--Dawson, + * http://eprint.iacr.org/2008/522, Section 3.1 + * appliesto extended-1 + * parameter k + * assume k = 2 d + * compute A = (Y1-X1)(Y2-X2) + * compute B = (Y1+X1)(Y2+X2) + * compute C = T1 k T2 + * compute D = Z1 2 Z2 + * compute E = B - A + * compute F = D - C + * compute G = D + C + * compute H = B + A + * compute X3 = E F + * compute Y3 = G H + * compute T3 = E H + * compute Z3 = F G + */ + byte a[F25519_SIZE]; + byte b[F25519_SIZE]; + byte c[F25519_SIZE]; + byte d[F25519_SIZE]; + byte e[F25519_SIZE]; + byte f[F25519_SIZE]; + byte g[F25519_SIZE]; + byte h[F25519_SIZE]; + + /* A = (Y1-X1)(Y2-X2) */ + fe_sub(c, p1->Y, p1->X); + fe_sub(d, p2->Y, p2->X); + fe_mul__distinct(a, c, d); + + /* B = (Y1+X1)(Y2+X2) */ + fe_add(c, p1->Y, p1->X); + fe_add(d, p2->Y, p2->X); + fe_mul__distinct(b, c, d); + + /* C = T1 k T2 */ + fe_mul__distinct(d, p1->T, p2->T); + fe_mul__distinct(c, d, ed25519_k); + + /* D = Z1 2 Z2 */ + fe_mul__distinct(d, p1->Z, p2->Z); + fe_add(d, d, d); + + /* E = B - A */ + fe_sub(e, b, a); + + /* F = D - C */ + fe_sub(f, d, c); + + /* G = D + C */ + fe_add(g, d, c); + + /* H = B + A */ + fe_add(h, b, a); + + /* X3 = E F */ + fe_mul__distinct(r->X, e, f); + + /* Y3 = G H */ + fe_mul__distinct(r->Y, g, h); + + /* T3 = E H */ + fe_mul__distinct(r->T, e, h); + + /* Z3 = F G */ + fe_mul__distinct(r->Z, f, g); +} + + +void ed25519_double(ge_p3 *r, const ge_p3 *p) +{ + /* Explicit formulas database: dbl-2008-hwcd + * + * source 2008 Hisil--Wong--Carter--Dawson, + * http://eprint.iacr.org/2008/522, Section 3.3 + * compute A = X1^2 + * compute B = Y1^2 + * compute C = 2 Z1^2 + * compute D = a A + * compute E = (X1+Y1)^2-A-B + * compute G = D + B + * compute F = G - C + * compute H = D - B + * compute X3 = E F + * compute Y3 = G H + * compute T3 = E H + * compute Z3 = F G + */ + byte a[F25519_SIZE]; + byte b[F25519_SIZE]; + byte c[F25519_SIZE]; + byte e[F25519_SIZE]; + byte f[F25519_SIZE]; + byte g[F25519_SIZE]; + byte h[F25519_SIZE]; + + /* A = X1^2 */ + fe_mul__distinct(a, p->X, p->X); + + /* B = Y1^2 */ + fe_mul__distinct(b, p->Y, p->Y); + + /* C = 2 Z1^2 */ + fe_mul__distinct(c, p->Z, p->Z); + fe_add(c, c, c); + + /* D = a A (alter sign) */ + /* E = (X1+Y1)^2-A-B */ + fe_add(f, p->X, p->Y); + fe_mul__distinct(e, f, f); + fe_sub(e, e, a); + fe_sub(e, e, b); + + /* G = D + B */ + fe_sub(g, b, a); + + /* F = G - C */ + fe_sub(f, g, c); + + /* H = D - B */ + fe_neg(h, b); + fe_sub(h, h, a); + + /* X3 = E F */ + fe_mul__distinct(r->X, e, f); + + /* Y3 = G H */ + fe_mul__distinct(r->Y, g, h); + + /* T3 = E H */ + fe_mul__distinct(r->T, e, h); + + /* Z3 = F G */ + fe_mul__distinct(r->Z, f, g); +} + + +void ed25519_smult(ge_p3 *r_out, const ge_p3 *p, const byte *e) +{ + ge_p3 r; + int i; + + XMEMCPY(&r, &ed25519_neutral, sizeof(r)); + + for (i = 255; i >= 0; i--) { + const byte bit = (e[i >> 3] >> (i & 7)) & 1; + ge_p3 s; + + ed25519_double(&r, &r); + ed25519_add(&s, &r, p); + + fe_select(r.X, r.X, s.X, bit); + fe_select(r.Y, r.Y, s.Y, bit); + fe_select(r.Z, r.Z, s.Z, bit); + fe_select(r.T, r.T, s.T, bit); + } + XMEMCPY(r_out, &r, sizeof(r)); +} + + +void ge_scalarmult_base(ge_p3 *R,const unsigned char *nonce) +{ + ed25519_smult(R, &ed25519_base, nonce); +} + + +/* pack the point h into array s */ +void ge_p3_tobytes(unsigned char *s,const ge_p3 *h) +{ + byte x[F25519_SIZE]; + byte y[F25519_SIZE]; + byte z1[F25519_SIZE]; + byte parity; + + fe_inv__distinct(z1, h->Z); + fe_mul__distinct(x, h->X, z1); + fe_mul__distinct(y, h->Y, z1); + + fe_normalize(x); + fe_normalize(y); + + parity = (x[0] & 1) << 7; + fe_copy(s, y); + fe_normalize(s); + s[31] |= parity; +} + + +/* pack the point h into array s */ +void ge_tobytes(unsigned char *s,const ge_p2 *h) +{ + byte x[F25519_SIZE]; + byte y[F25519_SIZE]; + byte z1[F25519_SIZE]; + byte parity; + + fe_inv__distinct(z1, h->Z); + fe_mul__distinct(x, h->X, z1); + fe_mul__distinct(y, h->Y, z1); + + fe_normalize(x); + fe_normalize(y); + + parity = (x[0] & 1) << 7; + fe_copy(s, y); + fe_normalize(s); + s[31] |= parity; +} + + +/* + Test if the public key can be uncompressed and negate it (-X,Y,Z,-T) + return 0 on success + */ +int ge_frombytes_negate_vartime(ge_p3 *p,const unsigned char *s) +{ + + byte parity; + byte x[F25519_SIZE]; + byte y[F25519_SIZE]; + byte a[F25519_SIZE]; + byte b[F25519_SIZE]; + byte c[F25519_SIZE]; + int ret = 0; + + /* unpack the key s */ + parity = s[31] >> 7; + fe_copy(y, s); + y[31] &= 127; + + fe_mul__distinct(c, y, y); + fe_mul__distinct(b, c, ed25519_d); + fe_add(a, b, f25519_one); + fe_inv__distinct(b, a); + fe_sub(a, c, f25519_one); + fe_mul__distinct(c, a, b); + fe_sqrt(a, c); + fe_neg(b, a); + fe_select(x, a, b, (a[0] ^ parity) & 1); + + /* test that x^2 is equal to c */ + fe_mul__distinct(a, x, x); + fe_normalize(a); + fe_normalize(c); + ret |= ConstantCompare(a, c, F25519_SIZE); + + /* project the key s onto p */ + fe_copy(p->X, x); + fe_copy(p->Y, y); + fe_load(p->Z, 1); + fe_mul__distinct(p->T, x, y); + + /* negate, the point becomes (-X,Y,Z,-T) */ + fe_neg(p->X,p->X); + fe_neg(p->T,p->T); + + return ret; +} + + +int ge_double_scalarmult_vartime(ge_p2* R, const unsigned char *h, + const ge_p3 *inA,const unsigned char *sig) +{ + ge_p3 p, A; + int ret = 0; + + XMEMCPY(&A, inA, sizeof(ge_p3)); + + /* find SB */ + ed25519_smult(&p, &ed25519_base, sig); + + /* find H(R,A,M) * -A */ + ed25519_smult(&A, &A, h); + + /* SB + -H(R,A,M)A */ + ed25519_add(&A, &p, &A); + + fe_copy(R->X, A.X); + fe_copy(R->Y, A.Y); + fe_copy(R->Z, A.Z); + + return ret; +} + +#endif /* HAVE_ED25519 */ +#endif /* CURVED25519_SMALL */ + +