tls: reorder P256 functions to make more sense

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

1 files changed, 179 insertions, 179 deletions

diff --git a/networking/tls_sp_c32.c b/networking/tls_sp_c32.c
index 1f14031..8059f6e 100644
--- a/networking/tls_sp_c32.c
+++ b/networking/tls_sp_c32.c
@@ -220,106 +220,6 @@ static void sp_256_rshift1_10(sp_digit* r, sp_digit* a)
 	r[9] = a[9] >> 1;
 }
 
-/* Multiply a number by Montogmery normalizer mod modulus (prime).
- *
- * r  The resulting Montgomery form number.
- * a  The number to convert.
- */
-static void sp_256_mod_mul_norm_10(sp_digit* r, const sp_digit* a)
-{
-	int64_t t[8];
-	int64_t a32[8];
-	int64_t o;
-
-	a32[0] = a[0];
-	a32[0] |= a[1] << 26;
-	a32[0] &= 0xffffffff;
-	a32[1] = (sp_digit)(a[1] >> 6);
-	a32[1] |= a[2] << 20;
-	a32[1] &= 0xffffffff;
-	a32[2] = (sp_digit)(a[2] >> 12);
-	a32[2] |= a[3] << 14;
-	a32[2] &= 0xffffffff;
-	a32[3] = (sp_digit)(a[3] >> 18);
-	a32[3] |= a[4] << 8;
-	a32[3] &= 0xffffffff;
-	a32[4] = (sp_digit)(a[4] >> 24);
-	a32[4] |= a[5] << 2;
-	a32[4] |= a[6] << 28;
-	a32[4] &= 0xffffffff;
-	a32[5] = (sp_digit)(a[6] >> 4);
-	a32[5] |= a[7] << 22;
-	a32[5] &= 0xffffffff;
-	a32[6] = (sp_digit)(a[7] >> 10);
-	a32[6] |= a[8] << 16;
-	a32[6] &= 0xffffffff;
-	a32[7] = (sp_digit)(a[8] >> 16);
-	a32[7] |= a[9] << 10;
-	a32[7] &= 0xffffffff;
-
-	/*  1  1  0 -1 -1 -1 -1  0 */
-	t[0] = 0 + a32[0] + a32[1] - a32[3] - a32[4] - a32[5] - a32[6];
-	/*  0  1  1  0 -1 -1 -1 -1 */
-	t[1] = 0 + a32[1] + a32[2] - a32[4] - a32[5] - a32[6] - a32[7];
-	/*  0  0  1  1  0 -1 -1 -1 */
-	t[2] = 0 + a32[2] + a32[3] - a32[5] - a32[6] - a32[7];
-	/* -1 -1  0  2  2  1  0 -1 */
-	t[3] = 0 - a32[0] - a32[1] + 2 * a32[3] + 2 * a32[4] + a32[5] - a32[7];
-	/*  0 -1 -1  0  2  2  1  0 */
-	t[4] = 0 - a32[1] - a32[2] + 2 * a32[4] + 2 * a32[5] + a32[6];
-	/*  0  0 -1 -1  0  2  2  1 */
-	t[5] = 0 - a32[2] - a32[3] + 2 * a32[5] + 2 * a32[6] + a32[7];
-	/* -1 -1  0  0  0  1  3  2 */
-	t[6] = 0 - a32[0] - a32[1] + a32[5] + 3 * a32[6] + 2 * a32[7];
-	/*  1  0 -1 -1 -1 -1  0  3 */
-	t[7] = 0 + a32[0] - a32[2] - a32[3] - a32[4] - a32[5] + 3 * a32[7];
-
-	t[1] += t[0] >> 32; t[0] &= 0xffffffff;
-	t[2] += t[1] >> 32; t[1] &= 0xffffffff;
-	t[3] += t[2] >> 32; t[2] &= 0xffffffff;
-	t[4] += t[3] >> 32; t[3] &= 0xffffffff;
-	t[5] += t[4] >> 32; t[4] &= 0xffffffff;
-	t[6] += t[5] >> 32; t[5] &= 0xffffffff;
-	t[7] += t[6] >> 32; t[6] &= 0xffffffff;
-	o     = t[7] >> 32; t[7] &= 0xffffffff;
-	t[0] += o;
-	t[3] -= o;
-	t[6] -= o;
-	t[7] += o;
-	t[1] += t[0] >> 32; t[0] &= 0xffffffff;
-	t[2] += t[1] >> 32; t[1] &= 0xffffffff;
-	t[3] += t[2] >> 32; t[2] &= 0xffffffff;
-	t[4] += t[3] >> 32; t[3] &= 0xffffffff;
-	t[5] += t[4] >> 32; t[4] &= 0xffffffff;
-	t[6] += t[5] >> 32; t[5] &= 0xffffffff;
-	t[7] += t[6] >> 32; t[6] &= 0xffffffff;
-
-	r[0] = (sp_digit)(t[0]) & 0x3ffffff;
-	r[1] = (sp_digit)(t[0] >> 26);
-	r[1] |= t[1] << 6;
-	r[1] &= 0x3ffffff;
-	r[2] = (sp_digit)(t[1] >> 20);
-	r[2] |= t[2] << 12;
-	r[2] &= 0x3ffffff;
-	r[3] = (sp_digit)(t[2] >> 14);
-	r[3] |= t[3] << 18;
-	r[3] &= 0x3ffffff;
-	r[4] = (sp_digit)(t[3] >> 8);
-	r[4] |= t[4] << 24;
-	r[4] &= 0x3ffffff;
-	r[5] = (sp_digit)(t[4] >> 2) & 0x3ffffff;
-	r[6] = (sp_digit)(t[4] >> 28);
-	r[6] |= t[5] << 4;
-	r[6] &= 0x3ffffff;
-	r[7] = (sp_digit)(t[5] >> 22);
-	r[7] |= t[6] << 10;
-	r[7] &= 0x3ffffff;
-	r[8] = (sp_digit)(t[6] >> 16);
-	r[8] |= t[7] << 16;
-	r[8] &= 0x3ffffff;
-	r[9] = (sp_digit)(t[7] >> 10);
-}
-
 /* Mul a by scalar b and add into r. (r += a * b) */
 static void sp_256_mul_add_10(sp_digit* r, const sp_digit* a, sp_digit b)
 {
@@ -335,6 +235,58 @@ static void sp_256_mul_add_10(sp_digit* r, const sp_digit* a, sp_digit b)
 	r[10] += t;
 }
 
+/* Multiply a and b into r. (r = a * b) */
+static void sp_256_mul_10(sp_digit* r, const sp_digit* a, const sp_digit* b)
+{
+	int i, j, k;
+	int64_t c;
+
+	c = ((int64_t)a[9]) * b[9];
+	r[19] = (sp_digit)(c >> 26);
+	c = (c & 0x3ffffff) << 26;
+	for (k = 17; k >= 0; k--) {
+		for (i = 9; i >= 0; i--) {
+			j = k - i;
+			if (j >= 10)
+				break;
+			if (j < 0)
+				continue;
+			c += ((int64_t)a[i]) * b[j];
+		}
+		r[k + 2] += c >> 52;
+		r[k + 1] = (c >> 26) & 0x3ffffff;
+		c = (c & 0x3ffffff) << 26;
+	}
+	r[0] = (sp_digit)(c >> 26);
+}
+
+/* Square a and put result in r. (r = a * a) */
+static void sp_256_sqr_10(sp_digit* r, const sp_digit* a)
+{
+	int i, j, k;
+	int64_t c;
+
+	c = ((int64_t)a[9]) * a[9];
+	r[19] = (sp_digit)(c >> 26);
+	c = (c & 0x3ffffff) << 26;
+	for (k = 17; k >= 0; k--) {
+		for (i = 9; i >= 0; i--) {
+			j = k - i;
+			if (j >= 10 || i <= j)
+				break;
+			if (j < 0)
+				continue;
+			c += ((int64_t)a[i]) * a[j] * 2;
+		}
+		if (i == j)
+			c += ((int64_t)a[i]) * a[i];
+		r[k + 2] += c >> 52;
+		r[k + 1] = (c >> 26) & 0x3ffffff;
+		c = (c & 0x3ffffff) << 26;
+	}
+	r[0] = (sp_digit)(c >> 26);
+}
+
 /* Divide the number by 2 mod the modulus (prime). (r = a / 2 % m) */
 static void sp_256_div2_10(sp_digit* r, const sp_digit* a, const sp_digit* m)
 {
@@ -344,25 +296,6 @@ static void sp_256_div2_10(sp_digit* r, const sp_digit* a, const sp_digit* m)
 	sp_256_rshift1_10(r, r);
 }
 
-/* Shift the result in the high 256 bits down to the bottom. */
-static void sp_256_mont_shift_10(sp_digit* r, const sp_digit* a)
-{
-	int i;
-	sp_digit n, s;
-
-	s = a[10];
-	n = a[9] >> 22;
-	for (i = 0; i < 9; i++) {
-		n += (s & 0x3ffffff) << 4;
-		r[i] = n & 0x3ffffff;
-		n >>= 26;
-		s = a[11 + i] + (s >> 26);
-	}
-	n += s << 4;
-	r[9] = n;
-	memset(&r[10], 0, sizeof(*r) * 10);
-}
-
 /* Add two Montgomery form numbers (r = a + b % m) */
 static void sp_256_mont_add_10(sp_digit* r, const sp_digit* a, const sp_digit* b,
 		const sp_digit* m)
@@ -374,6 +307,16 @@ static void sp_256_mont_add_10(sp_digit* r, const sp_digit* a, const sp_digit* b
 	sp_256_norm_10(r);
 }
 
+/* Subtract two Montgomery form numbers (r = a - b % m) */
+static void sp_256_mont_sub_10(sp_digit* r, const sp_digit* a, const sp_digit* b,
+		const sp_digit* m)
+{
+	sp_256_sub_10(r, a, b);
+	if (r[9] >> 22)
+		sp_256_add_10(r, r, m);
+	sp_256_norm_10(r);
+}
+
 /* Double a Montgomery form number (r = a + a % m) */
 static void sp_256_mont_dbl_10(sp_digit* r, const sp_digit* a, const sp_digit* m)
 {
@@ -399,14 +342,23 @@ static void sp_256_mont_tpl_10(sp_digit* r, const sp_digit* a, const sp_digit* m
 	sp_256_norm_10(r);
 }
 
-/* Subtract two Montgomery form numbers (r = a - b % m) */
-static void sp_256_mont_sub_10(sp_digit* r, const sp_digit* a, const sp_digit* b,
-		const sp_digit* m)
+/* Shift the result in the high 256 bits down to the bottom. */
+static void sp_256_mont_shift_10(sp_digit* r, const sp_digit* a)
 {
-	sp_256_sub_10(r, a, b);
-	if (r[9] >> 22)
-		sp_256_add_10(r, r, m);
-	sp_256_norm_10(r);
+	int i;
+	sp_digit n, s;
+
+	s = a[10];
+	n = a[9] >> 22;
+	for (i = 0; i < 9; i++) {
+		n += (s & 0x3ffffff) << 4;
+		r[i] = n & 0x3ffffff;
+		n >>= 26;
+		s = a[11 + i] + (s >> 26);
+	}
+	n += s << 4;
+	r[9] = n;
+	memset(&r[10], 0, sizeof(*r) * 10);
 }
 
 /* Reduce the number back to 256 bits using Montgomery reduction.
@@ -449,31 +401,6 @@ static void sp_256_mont_reduce_10(sp_digit* a, const sp_digit* m, sp_digit mp)
 	sp_256_norm_10(a);
 }
 
-/* Multiply a and b into r. (r = a * b) */
-static void sp_256_mul_10(sp_digit* r, const sp_digit* a, const sp_digit* b)
-{
-	int i, j, k;
-	int64_t c;
-
-	c = ((int64_t)a[9]) * b[9];
-	r[19] = (sp_digit)(c >> 26);
-	c = (c & 0x3ffffff) << 26;
-	for (k = 17; k >= 0; k--) {
-		for (i = 9; i >= 0; i--) {
-			j = k - i;
-			if (j >= 10)
-				break;
-			if (j < 0)
-				continue;
-			c += ((int64_t)a[i]) * b[j];
-		}
-		r[k + 2] += c >> 52;
-		r[k + 1] = (c >> 26) & 0x3ffffff;
-		c = (c & 0x3ffffff) << 26;
-	}
-	r[0] = (sp_digit)(c >> 26);
-}
-
 /* Multiply two Montogmery form numbers mod the modulus (prime).
  * (r = a * b mod m)
  *
@@ -490,33 +417,6 @@ static void sp_256_mont_mul_10(sp_digit* r, const sp_digit* a, const sp_digit* b
 	sp_256_mont_reduce_10(r, m, mp);
 }
 
-/* Square a and put result in r. (r = a * a) */
-static void sp_256_sqr_10(sp_digit* r, const sp_digit* a)
-{
-	int i, j, k;
-	int64_t c;
-
-	c = ((int64_t)a[9]) * a[9];
-	r[19] = (sp_digit)(c >> 26);
-	c = (c & 0x3ffffff) << 26;
-	for (k = 17; k >= 0; k--) {
-		for (i = 9; i >= 0; i--) {
-			j = k - i;
-			if (j >= 10 || i <= j)
-				break;
-			if (j < 0)
-				continue;
-			c += ((int64_t)a[i]) * a[j] * 2;
-		}
-		if (i == j)
-			c += ((int64_t)a[i]) * a[i];
-		r[k + 2] += c >> 52;
-		r[k + 1] = (c >> 26) & 0x3ffffff;
-		c = (c & 0x3ffffff) << 26;
-	}
-	r[0] = (sp_digit)(c >> 26);
-}
-
 /* Square the Montgomery form number. (r = a * a mod m)
  *
  * r   Result of squaring.
@@ -564,6 +464,106 @@ static void sp_256_mont_inv_10(sp_digit* r, sp_digit* a)
 	memcpy(r, t, sizeof(sp_digit) * 10);
 }
 
+/* Multiply a number by Montogmery normalizer mod modulus (prime).
+ *
+ * r  The resulting Montgomery form number.
+ * a  The number to convert.
+ */
+static void sp_256_mod_mul_norm_10(sp_digit* r, const sp_digit* a)
+{
+	int64_t t[8];
+	int64_t a32[8];
+	int64_t o;
+
+	a32[0] = a[0];
+	a32[0] |= a[1] << 26;
+	a32[0] &= 0xffffffff;
+	a32[1] = (sp_digit)(a[1] >> 6);
+	a32[1] |= a[2] << 20;
+	a32[1] &= 0xffffffff;
+	a32[2] = (sp_digit)(a[2] >> 12);
+	a32[2] |= a[3] << 14;
+	a32[2] &= 0xffffffff;
+	a32[3] = (sp_digit)(a[3] >> 18);
+	a32[3] |= a[4] << 8;
+	a32[3] &= 0xffffffff;
+	a32[4] = (sp_digit)(a[4] >> 24);
+	a32[4] |= a[5] << 2;
+	a32[4] |= a[6] << 28;
+	a32[4] &= 0xffffffff;
+	a32[5] = (sp_digit)(a[6] >> 4);
+	a32[5] |= a[7] << 22;
+	a32[5] &= 0xffffffff;
+	a32[6] = (sp_digit)(a[7] >> 10);
+	a32[6] |= a[8] << 16;
+	a32[6] &= 0xffffffff;
+	a32[7] = (sp_digit)(a[8] >> 16);
+	a32[7] |= a[9] << 10;
+	a32[7] &= 0xffffffff;
+
+	/*  1  1  0 -1 -1 -1 -1  0 */
+	t[0] = 0 + a32[0] + a32[1] - a32[3] - a32[4] - a32[5] - a32[6];
+	/*  0  1  1  0 -1 -1 -1 -1 */
+	t[1] = 0 + a32[1] + a32[2] - a32[4] - a32[5] - a32[6] - a32[7];
+	/*  0  0  1  1  0 -1 -1 -1 */
+	t[2] = 0 + a32[2] + a32[3] - a32[5] - a32[6] - a32[7];
+	/* -1 -1  0  2  2  1  0 -1 */
+	t[3] = 0 - a32[0] - a32[1] + 2 * a32[3] + 2 * a32[4] + a32[5] - a32[7];
+	/*  0 -1 -1  0  2  2  1  0 */
+	t[4] = 0 - a32[1] - a32[2] + 2 * a32[4] + 2 * a32[5] + a32[6];
+	/*  0  0 -1 -1  0  2  2  1 */
+	t[5] = 0 - a32[2] - a32[3] + 2 * a32[5] + 2 * a32[6] + a32[7];
+	/* -1 -1  0  0  0  1  3  2 */
+	t[6] = 0 - a32[0] - a32[1] + a32[5] + 3 * a32[6] + 2 * a32[7];
+	/*  1  0 -1 -1 -1 -1  0  3 */
+	t[7] = 0 + a32[0] - a32[2] - a32[3] - a32[4] - a32[5] + 3 * a32[7];
+
+	t[1] += t[0] >> 32; t[0] &= 0xffffffff;
+	t[2] += t[1] >> 32; t[1] &= 0xffffffff;
+	t[3] += t[2] >> 32; t[2] &= 0xffffffff;
+	t[4] += t[3] >> 32; t[3] &= 0xffffffff;
+	t[5] += t[4] >> 32; t[4] &= 0xffffffff;
+	t[6] += t[5] >> 32; t[5] &= 0xffffffff;
+	t[7] += t[6] >> 32; t[6] &= 0xffffffff;
+	o     = t[7] >> 32; t[7] &= 0xffffffff;
+	t[0] += o;
+	t[3] -= o;
+	t[6] -= o;
+	t[7] += o;
+	t[1] += t[0] >> 32; t[0] &= 0xffffffff;
+	t[2] += t[1] >> 32; t[1] &= 0xffffffff;
+	t[3] += t[2] >> 32; t[2] &= 0xffffffff;
+	t[4] += t[3] >> 32; t[3] &= 0xffffffff;
+	t[5] += t[4] >> 32; t[4] &= 0xffffffff;
+	t[6] += t[5] >> 32; t[5] &= 0xffffffff;
+	t[7] += t[6] >> 32; t[6] &= 0xffffffff;
+
+	r[0] = (sp_digit)(t[0]) & 0x3ffffff;
+	r[1] = (sp_digit)(t[0] >> 26);
+	r[1] |= t[1] << 6;
+	r[1] &= 0x3ffffff;
+	r[2] = (sp_digit)(t[1] >> 20);
+	r[2] |= t[2] << 12;
+	r[2] &= 0x3ffffff;
+	r[3] = (sp_digit)(t[2] >> 14);
+	r[3] |= t[3] << 18;
+	r[3] &= 0x3ffffff;
+	r[4] = (sp_digit)(t[3] >> 8);
+	r[4] |= t[4] << 24;
+	r[4] &= 0x3ffffff;
+	r[5] = (sp_digit)(t[4] >> 2) & 0x3ffffff;
+	r[6] = (sp_digit)(t[4] >> 28);
+	r[6] |= t[5] << 4;
+	r[6] &= 0x3ffffff;
+	r[7] = (sp_digit)(t[5] >> 22);
+	r[7] |= t[6] << 10;
+	r[7] &= 0x3ffffff;
+	r[8] = (sp_digit)(t[6] >> 16);
+	r[8] |= t[7] << 16;
+	r[8] &= 0x3ffffff;
+	r[9] = (sp_digit)(t[7] >> 10);
+}
+
 /* Map the Montgomery form projective co-ordinate point to an affine point.
  *
  * r  Resulting affine co-ordinate point.
@@ -808,7 +808,7 @@ static void sp_256_ecc_mulmod_base_10(sp_point* r, sp_digit* k /*, int map*/)
  * pub2x32 Point to multiply.
  * out32   Buffer to hold X ordinate.
  */
-static void sp_ecc_secret_gen_256(sp_digit priv[10], const uint8_t *pub2x32, uint8_t* out32)
+static void sp_ecc_secret_gen_256(const sp_digit priv[10], const uint8_t *pub2x32, uint8_t* out32)
 {
 	sp_point point[1];