path: root/security/nss/lib/freebl/verified/Hacl_Curve25519_64.c

/* MIT License
 *
 * Copyright (c) 2016-2022 INRIA, CMU and Microsoft Corporation
 * Copyright (c) 2022-2023 HACL* Contributors
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to deal
 * in the Software without restriction, including without limitation the rights
 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 * copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in all
 * copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 */

#include "Hacl_Curve25519_64.h"

#include "internal/Vale.h"
#include "internal/Hacl_Krmllib.h"
#include "config.h"
#include "curve25519-inline.h"

static inline void
add_scalar0(uint64_t *out, uint64_t *f1, uint64_t f2)
{
#if HACL_CAN_COMPILE_INLINE_ASM
    add_scalar(out, f1, f2);
#else
    uint64_t uu____0 = add_scalar_e(out, f1, f2);
#endif
}

static inline void
fadd0(uint64_t *out, uint64_t *f1, uint64_t *f2)
{
#if HACL_CAN_COMPILE_INLINE_ASM
    fadd(out, f1, f2);
#else
    uint64_t uu____0 = fadd_e(out, f1, f2);
#endif
}

static inline void
fsub0(uint64_t *out, uint64_t *f1, uint64_t *f2)
{
#if HACL_CAN_COMPILE_INLINE_ASM
    fsub(out, f1, f2);
#else
    uint64_t uu____0 = fsub_e(out, f1, f2);
#endif
}

static inline void
fmul0(uint64_t *out, uint64_t *f1, uint64_t *f2, uint64_t *tmp)
{
#if HACL_CAN_COMPILE_INLINE_ASM
    fmul(out, f1, f2, tmp);
#else
    uint64_t uu____0 = fmul_e(tmp, f1, out, f2);
#endif
}

static inline void
fmul20(uint64_t *out, uint64_t *f1, uint64_t *f2, uint64_t *tmp)
{
#if HACL_CAN_COMPILE_INLINE_ASM
    fmul2(out, f1, f2, tmp);
#else
    uint64_t uu____0 = fmul2_e(tmp, f1, out, f2);
#endif
}

static inline void
fmul_scalar0(uint64_t *out, uint64_t *f1, uint64_t f2)
{
#if HACL_CAN_COMPILE_INLINE_ASM
    fmul_scalar(out, f1, f2);
#else
    uint64_t uu____0 = fmul_scalar_e(out, f1, f2);
#endif
}

static inline void
fsqr0(uint64_t *out, uint64_t *f1, uint64_t *tmp)
{
#if HACL_CAN_COMPILE_INLINE_ASM
    fsqr(out, f1, tmp);
#else
    uint64_t uu____0 = fsqr_e(tmp, f1, out);
#endif
}

static inline void
fsqr20(uint64_t *out, uint64_t *f, uint64_t *tmp)
{
#if HACL_CAN_COMPILE_INLINE_ASM
    fsqr2(out, f, tmp);
#else
    uint64_t uu____0 = fsqr2_e(tmp, f, out);
#endif
}

static inline void
cswap20(uint64_t bit, uint64_t *p1, uint64_t *p2)
{
#if HACL_CAN_COMPILE_INLINE_ASM
    cswap2(bit, p1, p2);
#else
    uint64_t uu____0 = cswap2_e(bit, p1, p2);
#endif
}

static const uint8_t g25519[32U] = { (uint8_t)9U };

static void
point_add_and_double(uint64_t *q, uint64_t *p01_tmp1, uint64_t *tmp2)
{
    uint64_t *nq = p01_tmp1;
    uint64_t *nq_p1 = p01_tmp1 + (uint32_t)8U;
    uint64_t *tmp1 = p01_tmp1 + (uint32_t)16U;
    uint64_t *x1 = q;
    uint64_t *x2 = nq;
    uint64_t *z2 = nq + (uint32_t)4U;
    uint64_t *z3 = nq_p1 + (uint32_t)4U;
    uint64_t *a = tmp1;
    uint64_t *b = tmp1 + (uint32_t)4U;
    uint64_t *ab = tmp1;
    uint64_t *dc = tmp1 + (uint32_t)8U;
    fadd0(a, x2, z2);
    fsub0(b, x2, z2);
    uint64_t *x3 = nq_p1;
    uint64_t *z31 = nq_p1 + (uint32_t)4U;
    uint64_t *d0 = dc;
    uint64_t *c0 = dc + (uint32_t)4U;
    fadd0(c0, x3, z31);
    fsub0(d0, x3, z31);
    fmul20(dc, dc, ab, tmp2);
    fadd0(x3, d0, c0);
    fsub0(z31, d0, c0);
    uint64_t *a1 = tmp1;
    uint64_t *b1 = tmp1 + (uint32_t)4U;
    uint64_t *d = tmp1 + (uint32_t)8U;
    uint64_t *c = tmp1 + (uint32_t)12U;
    uint64_t *ab1 = tmp1;
    uint64_t *dc1 = tmp1 + (uint32_t)8U;
    fsqr20(dc1, ab1, tmp2);
    fsqr20(nq_p1, nq_p1, tmp2);
    a1[0U] = c[0U];
    a1[1U] = c[1U];
    a1[2U] = c[2U];
    a1[3U] = c[3U];
    fsub0(c, d, c);
    fmul_scalar0(b1, c, (uint64_t)121665U);
    fadd0(b1, b1, d);
    fmul20(nq, dc1, ab1, tmp2);
    fmul0(z3, z3, x1, tmp2);
}

static void
point_double(uint64_t *nq, uint64_t *tmp1, uint64_t *tmp2)
{
    uint64_t *x2 = nq;
    uint64_t *z2 = nq + (uint32_t)4U;
    uint64_t *a = tmp1;
    uint64_t *b = tmp1 + (uint32_t)4U;
    uint64_t *d = tmp1 + (uint32_t)8U;
    uint64_t *c = tmp1 + (uint32_t)12U;
    uint64_t *ab = tmp1;
    uint64_t *dc = tmp1 + (uint32_t)8U;
    fadd0(a, x2, z2);
    fsub0(b, x2, z2);
    fsqr20(dc, ab, tmp2);
    a[0U] = c[0U];
    a[1U] = c[1U];
    a[2U] = c[2U];
    a[3U] = c[3U];
    fsub0(c, d, c);
    fmul_scalar0(b, c, (uint64_t)121665U);
    fadd0(b, b, d);
    fmul20(nq, dc, ab, tmp2);
}

static void
montgomery_ladder(uint64_t *out, uint8_t *key, uint64_t *init)
{
    uint64_t tmp2[16U] = { 0U };
    uint64_t p01_tmp1_swap[33U] = { 0U };
    uint64_t *p0 = p01_tmp1_swap;
    uint64_t *p01 = p01_tmp1_swap;
    uint64_t *p03 = p01;
    uint64_t *p11 = p01 + (uint32_t)8U;
    memcpy(p11, init, (uint32_t)8U * sizeof(uint64_t));
    uint64_t *x0 = p03;
    uint64_t *z0 = p03 + (uint32_t)4U;
    x0[0U] = (uint64_t)1U;
    x0[1U] = (uint64_t)0U;
    x0[2U] = (uint64_t)0U;
    x0[3U] = (uint64_t)0U;
    z0[0U] = (uint64_t)0U;
    z0[1U] = (uint64_t)0U;
    z0[2U] = (uint64_t)0U;
    z0[3U] = (uint64_t)0U;
    uint64_t *p01_tmp1 = p01_tmp1_swap;
    uint64_t *p01_tmp11 = p01_tmp1_swap;
    uint64_t *nq1 = p01_tmp1_swap;
    uint64_t *nq_p11 = p01_tmp1_swap + (uint32_t)8U;
    uint64_t *swap = p01_tmp1_swap + (uint32_t)32U;
    cswap20((uint64_t)1U, nq1, nq_p11);
    point_add_and_double(init, p01_tmp11, tmp2);
    swap[0U] = (uint64_t)1U;
    for (uint32_t i = (uint32_t)0U; i < (uint32_t)251U; i++) {
        uint64_t *p01_tmp12 = p01_tmp1_swap;
        uint64_t *swap1 = p01_tmp1_swap + (uint32_t)32U;
        uint64_t *nq2 = p01_tmp12;
        uint64_t *nq_p12 = p01_tmp12 + (uint32_t)8U;
        uint64_t
            bit =
                (uint64_t)(key[((uint32_t)253U - i) / (uint32_t)8U] >> ((uint32_t)253U - i) % (uint32_t)8U & (uint8_t)1U);
        uint64_t sw = swap1[0U] ^ bit;
        cswap20(sw, nq2, nq_p12);
        point_add_and_double(init, p01_tmp12, tmp2);
        swap1[0U] = bit;
    }
    uint64_t sw = swap[0U];
    cswap20(sw, nq1, nq_p11);
    uint64_t *nq10 = p01_tmp1;
    uint64_t *tmp1 = p01_tmp1 + (uint32_t)16U;
    point_double(nq10, tmp1, tmp2);
    point_double(nq10, tmp1, tmp2);
    point_double(nq10, tmp1, tmp2);
    memcpy(out, p0, (uint32_t)8U * sizeof(uint64_t));
}

static void
fsquare_times(uint64_t *o, uint64_t *inp, uint64_t *tmp, uint32_t n)
{
    fsqr0(o, inp, tmp);
    for (uint32_t i = (uint32_t)0U; i < n - (uint32_t)1U; i++) {
        fsqr0(o, o, tmp);
    }
}

static void
finv(uint64_t *o, uint64_t *i, uint64_t *tmp)
{
    uint64_t t1[16U] = { 0U };
    uint64_t *a1 = t1;
    uint64_t *b1 = t1 + (uint32_t)4U;
    uint64_t *t010 = t1 + (uint32_t)12U;
    uint64_t *tmp10 = tmp;
    fsquare_times(a1, i, tmp10, (uint32_t)1U);
    fsquare_times(t010, a1, tmp10, (uint32_t)2U);
    fmul0(b1, t010, i, tmp);
    fmul0(a1, b1, a1, tmp);
    fsquare_times(t010, a1, tmp10, (uint32_t)1U);
    fmul0(b1, t010, b1, tmp);
    fsquare_times(t010, b1, tmp10, (uint32_t)5U);
    fmul0(b1, t010, b1, tmp);
    uint64_t *b10 = t1 + (uint32_t)4U;
    uint64_t *c10 = t1 + (uint32_t)8U;
    uint64_t *t011 = t1 + (uint32_t)12U;
    uint64_t *tmp11 = tmp;
    fsquare_times(t011, b10, tmp11, (uint32_t)10U);
    fmul0(c10, t011, b10, tmp);
    fsquare_times(t011, c10, tmp11, (uint32_t)20U);
    fmul0(t011, t011, c10, tmp);
    fsquare_times(t011, t011, tmp11, (uint32_t)10U);
    fmul0(b10, t011, b10, tmp);
    fsquare_times(t011, b10, tmp11, (uint32_t)50U);
    fmul0(c10, t011, b10, tmp);
    uint64_t *b11 = t1 + (uint32_t)4U;
    uint64_t *c1 = t1 + (uint32_t)8U;
    uint64_t *t01 = t1 + (uint32_t)12U;
    uint64_t *tmp1 = tmp;
    fsquare_times(t01, c1, tmp1, (uint32_t)100U);
    fmul0(t01, t01, c1, tmp);
    fsquare_times(t01, t01, tmp1, (uint32_t)50U);
    fmul0(t01, t01, b11, tmp);
    fsquare_times(t01, t01, tmp1, (uint32_t)5U);
    uint64_t *a = t1;
    uint64_t *t0 = t1 + (uint32_t)12U;
    fmul0(o, t0, a, tmp);
}

static void
store_felem(uint64_t *b, uint64_t *f)
{
    uint64_t f30 = f[3U];
    uint64_t top_bit0 = f30 >> (uint32_t)63U;
    f[3U] = f30 & (uint64_t)0x7fffffffffffffffU;
    add_scalar0(f, f, (uint64_t)19U * top_bit0);
    uint64_t f31 = f[3U];
    uint64_t top_bit = f31 >> (uint32_t)63U;
    f[3U] = f31 & (uint64_t)0x7fffffffffffffffU;
    add_scalar0(f, f, (uint64_t)19U * top_bit);
    uint64_t f0 = f[0U];
    uint64_t f1 = f[1U];
    uint64_t f2 = f[2U];
    uint64_t f3 = f[3U];
    uint64_t m0 = FStar_UInt64_gte_mask(f0, (uint64_t)0xffffffffffffffedU);
    uint64_t m1 = FStar_UInt64_eq_mask(f1, (uint64_t)0xffffffffffffffffU);
    uint64_t m2 = FStar_UInt64_eq_mask(f2, (uint64_t)0xffffffffffffffffU);
    uint64_t m3 = FStar_UInt64_eq_mask(f3, (uint64_t)0x7fffffffffffffffU);
    uint64_t mask = ((m0 & m1) & m2) & m3;
    uint64_t f0_ = f0 - (mask & (uint64_t)0xffffffffffffffedU);
    uint64_t f1_ = f1 - (mask & (uint64_t)0xffffffffffffffffU);
    uint64_t f2_ = f2 - (mask & (uint64_t)0xffffffffffffffffU);
    uint64_t f3_ = f3 - (mask & (uint64_t)0x7fffffffffffffffU);
    uint64_t o0 = f0_;
    uint64_t o1 = f1_;
    uint64_t o2 = f2_;
    uint64_t o3 = f3_;
    b[0U] = o0;
    b[1U] = o1;
    b[2U] = o2;
    b[3U] = o3;
}

static void
encode_point(uint8_t *o, uint64_t *i)
{
    uint64_t *x = i;
    uint64_t *z = i + (uint32_t)4U;
    uint64_t tmp[4U] = { 0U };
    uint64_t u64s[4U] = { 0U };
    uint64_t tmp_w[16U] = { 0U };
    finv(tmp, z, tmp_w);
    fmul0(tmp, tmp, x, tmp_w);
    store_felem(u64s, tmp);
    KRML_MAYBE_FOR4(i0,
                    (uint32_t)0U,
                    (uint32_t)4U,
                    (uint32_t)1U,
                    store64_le(o + i0 * (uint32_t)8U, u64s[i0]););
}

/**
Compute the scalar multiple of a point.

@param out Pointer to 32 bytes of memory, allocated by the caller, where the resulting point is written to.
@param priv Pointer to 32 bytes of memory where the secret/private key is read from.
@param pub Pointer to 32 bytes of memory where the public point is read from.
*/
void
Hacl_Curve25519_64_scalarmult(uint8_t *out, uint8_t *priv, uint8_t *pub)
{
    uint64_t init[8U] = { 0U };
    uint64_t tmp[4U] = { 0U };
    KRML_MAYBE_FOR4(i,
                    (uint32_t)0U,
                    (uint32_t)4U,
                    (uint32_t)1U,
                    uint64_t *os = tmp;
                    uint8_t *bj = pub + i * (uint32_t)8U;
                    uint64_t u = load64_le(bj);
                    uint64_t r = u;
                    uint64_t x = r;
                    os[i] = x;);
    uint64_t tmp3 = tmp[3U];
    tmp[3U] = tmp3 & (uint64_t)0x7fffffffffffffffU;
    uint64_t *x = init;
    uint64_t *z = init + (uint32_t)4U;
    z[0U] = (uint64_t)1U;
    z[1U] = (uint64_t)0U;
    z[2U] = (uint64_t)0U;
    z[3U] = (uint64_t)0U;
    x[0U] = tmp[0U];
    x[1U] = tmp[1U];
    x[2U] = tmp[2U];
    x[3U] = tmp[3U];
    montgomery_ladder(init, priv, init);
    encode_point(out, init);
}

/**
Calculate a public point from a secret/private key.

This computes a scalar multiplication of the secret/private key with the curve's basepoint.

@param pub Pointer to 32 bytes of memory, allocated by the caller, where the resulting point is written to.
@param priv Pointer to 32 bytes of memory where the secret/private key is read from.
*/
void
Hacl_Curve25519_64_secret_to_public(uint8_t *pub, uint8_t *priv)
{
    uint8_t basepoint[32U] = { 0U };
    for (uint32_t i = (uint32_t)0U; i < (uint32_t)32U; i++) {
        uint8_t *os = basepoint;
        uint8_t x = g25519[i];
        os[i] = x;
    }
    Hacl_Curve25519_64_scalarmult(pub, priv, basepoint);
}

/**
Execute the diffie-hellmann key exchange.

@param out Pointer to 32 bytes of memory, allocated by the caller, where the resulting point is written to.
@param priv Pointer to 32 bytes of memory where **our** secret/private key is read from.
@param pub Pointer to 32 bytes of memory where **their** public point is read from.
*/
bool
Hacl_Curve25519_64_ecdh(uint8_t *out, uint8_t *priv, uint8_t *pub)
{
    uint8_t zeros[32U] = { 0U };
    Hacl_Curve25519_64_scalarmult(out, priv, pub);
    uint8_t res = (uint8_t)255U;
    for (uint32_t i = (uint32_t)0U; i < (uint32_t)32U; i++) {
        uint8_t uu____0 = FStar_UInt8_eq_mask(out[i], zeros[i]);
        res = uu____0 & res;
    }
    uint8_t z = res;
    bool r = z == (uint8_t)255U;
    return !r;
}