curve25519.c: improve formula alignment

Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/7750)

(cherry picked from commit 425dde5d75a6ffff9a1027107711b99da39e3519)

1 files changed, 513 insertions, 523 deletions

diff --git a/crypto/ec/curve25519.c b/crypto/ec/curve25519.c
index d331784685..aa999cc591 100644
--- a/crypto/ec/curve25519.c
+++ b/crypto/ec/curve25519.c
@@ -759,6 +759,7 @@ static void x25519_scalar_mult(uint8_t out[32], const uint8_t scalar[32],
  */
 typedef int32_t fe[10];
 
+static const int64_t kBottom21Bits =  0x1fffffLL;
 static const int64_t kBottom25Bits = 0x1ffffffLL;
 static const int64_t kBottom26Bits = 0x3ffffffLL;
 static const int64_t kTop39Bits = 0xfffffffffe000000LL;
@@ -768,7 +769,7 @@ static uint64_t load_3(const uint8_t *in)
 {
     uint64_t result;
 
-    result = (uint64_t)in[0];
+    result  = ((uint64_t)in[0]);
     result |= ((uint64_t)in[1]) << 8;
     result |= ((uint64_t)in[2]) << 16;
     return result;
@@ -778,7 +779,7 @@ static uint64_t load_4(const uint8_t *in)
 {
     uint64_t result;
 
-    result = (uint64_t)in[0];
+    result  = ((uint64_t)in[0]);
     result |= ((uint64_t)in[1]) << 8;
     result |= ((uint64_t)in[2]) << 16;
     result |= ((uint64_t)in[3]) << 24;
@@ -788,16 +789,16 @@ static uint64_t load_4(const uint8_t *in)
 static void fe_frombytes(fe h, const uint8_t *s)
 {
     /* Ignores top bit of h. */
-    int64_t h0 = load_4(s);
-    int64_t h1 = load_3(s + 4) << 6;
-    int64_t h2 = load_3(s + 7) << 5;
-    int64_t h3 = load_3(s + 10) << 3;
-    int64_t h4 = load_3(s + 13) << 2;
-    int64_t h5 = load_4(s + 16);
-    int64_t h6 = load_3(s + 20) << 7;
-    int64_t h7 = load_3(s + 23) << 5;
-    int64_t h8 = load_3(s + 26) << 4;
-    int64_t h9 = (load_3(s + 29) & 8388607) << 2;
+    int64_t h0 =  load_4(s);
+    int64_t h1 =  load_3(s +  4) << 6;
+    int64_t h2 =  load_3(s +  7) << 5;
+    int64_t h3 =  load_3(s + 10) << 3;
+    int64_t h4 =  load_3(s + 13) << 2;
+    int64_t h5 =  load_4(s + 16);
+    int64_t h6 =  load_3(s + 20) << 7;
+    int64_t h7 =  load_3(s + 23) << 5;
+    int64_t h8 =  load_3(s + 26) << 4;
+    int64_t h9 = (load_3(s + 29) & 0x7fffff) << 2;
     int64_t carry0;
     int64_t carry1;
     int64_t carry2;
@@ -905,38 +906,38 @@ static void fe_tobytes(uint8_t *s, const fe h)
      * evidently 2^255 h10-2^255 q = 0.
      * Goal: Output h0+...+2^230 h9.
      */
-    s[0] = (uint8_t)(h0 >> 0);
-    s[1] = (uint8_t)(h0 >> 8);
-    s[2] = (uint8_t)(h0 >> 16);
-    s[3] = (uint8_t)((h0 >> 24) | ((uint32_t)(h1) << 2));
-    s[4] = (uint8_t)(h1 >> 6);
-    s[5] = (uint8_t)(h1 >> 14);
-    s[6] = (uint8_t)((h1 >> 22) | ((uint32_t)(h2) << 3));
-    s[7] = (uint8_t)(h2 >> 5);
-    s[8] = (uint8_t)(h2 >> 13);
-    s[9] = (uint8_t)((h2 >> 21) | ((uint32_t)(h3) << 5));
-    s[10] = (uint8_t)(h3 >> 3);
-    s[11] = (uint8_t)(h3 >> 11);
+    s[ 0] = (uint8_t) (h0 >>  0);
+    s[ 1] = (uint8_t) (h0 >>  8);
+    s[ 2] = (uint8_t) (h0 >> 16);
+    s[ 3] = (uint8_t)((h0 >> 24) | ((uint32_t)(h1) << 2));
+    s[ 4] = (uint8_t) (h1 >>  6);
+    s[ 5] = (uint8_t) (h1 >> 14);
+    s[ 6] = (uint8_t)((h1 >> 22) | ((uint32_t)(h2) << 3));
+    s[ 7] = (uint8_t) (h2 >>  5);
+    s[ 8] = (uint8_t) (h2 >> 13);
+    s[ 9] = (uint8_t)((h2 >> 21) | ((uint32_t)(h3) << 5));
+    s[10] = (uint8_t) (h3 >>  3);
+    s[11] = (uint8_t) (h3 >> 11);
     s[12] = (uint8_t)((h3 >> 19) | ((uint32_t)(h4) << 6));
-    s[13] = (uint8_t)(h4 >> 2);
-    s[14] = (uint8_t)(h4 >> 10);
-    s[15] = (uint8_t)(h4 >> 18);
-    s[16] = (uint8_t)(h5 >> 0);
-    s[17] = (uint8_t)(h5 >> 8);
-    s[18] = (uint8_t)(h5 >> 16);
+    s[13] = (uint8_t) (h4 >>  2);
+    s[14] = (uint8_t) (h4 >> 10);
+    s[15] = (uint8_t) (h4 >> 18);
+    s[16] = (uint8_t) (h5 >>  0);
+    s[17] = (uint8_t) (h5 >>  8);
+    s[18] = (uint8_t) (h5 >> 16);
     s[19] = (uint8_t)((h5 >> 24) | ((uint32_t)(h6) << 1));
-    s[20] = (uint8_t)(h6 >> 7);
-    s[21] = (uint8_t)(h6 >> 15);
+    s[20] = (uint8_t) (h6 >>  7);
+    s[21] = (uint8_t) (h6 >> 15);
     s[22] = (uint8_t)((h6 >> 23) | ((uint32_t)(h7) << 3));
-    s[23] = (uint8_t)(h7 >> 5);
-    s[24] = (uint8_t)(h7 >> 13);
+    s[23] = (uint8_t) (h7 >>  5);
+    s[24] = (uint8_t) (h7 >> 13);
     s[25] = (uint8_t)((h7 >> 21) | ((uint32_t)(h8) << 4));
-    s[26] = (uint8_t)(h8 >> 4);
-    s[27] = (uint8_t)(h8 >> 12);
+    s[26] = (uint8_t) (h8 >>  4);
+    s[27] = (uint8_t) (h8 >> 12);
     s[28] = (uint8_t)((h8 >> 20) | ((uint32_t)(h9) << 6));
-    s[29] = (uint8_t)(h9 >> 2);
-    s[30] = (uint8_t)(h9 >> 10);
-    s[31] = (uint8_t)(h9 >> 18);
+    s[29] = (uint8_t) (h9 >>  2);
+    s[30] = (uint8_t) (h9 >> 10);
+    s[31] = (uint8_t) (h9 >> 18);
 }
 
 /* h = f */
@@ -1166,16 +1167,16 @@ static void fe_mul(fe h, const fe f, const fe g)
     int64_t f9g7_38 = f9_2 * (int64_t) g7_19;
     int64_t f9g8_19 = f9   * (int64_t) g8_19;
     int64_t f9g9_38 = f9_2 * (int64_t) g9_19;
-    int64_t h0 = f0g0+f1g9_38+f2g8_19+f3g7_38+f4g6_19+f5g5_38+f6g4_19+f7g3_38+f8g2_19+f9g1_38;
-    int64_t h1 = f0g1+f1g0   +f2g9_19+f3g8_19+f4g7_19+f5g6_19+f6g5_19+f7g4_19+f8g3_19+f9g2_19;
-    int64_t h2 = f0g2+f1g1_2 +f2g0   +f3g9_38+f4g8_19+f5g7_38+f6g6_19+f7g5_38+f8g4_19+f9g3_38;
-    int64_t h3 = f0g3+f1g2   +f2g1   +f3g0   +f4g9_19+f5g8_19+f6g7_19+f7g6_19+f8g5_19+f9g4_19;
-    int64_t h4 = f0g4+f1g3_2 +f2g2   +f3g1_2 +f4g0   +f5g9_38+f6g8_19+f7g7_38+f8g6_19+f9g5_38;
-    int64_t h5 = f0g5+f1g4   +f2g3   +f3g2   +f4g1   +f5g0   +f6g9_19+f7g8_19+f8g7_19+f9g6_19;
-    int64_t h6 = f0g6+f1g5_2 +f2g4   +f3g3_2 +f4g2   +f5g1_2 +f6g0   +f7g9_38+f8g8_19+f9g7_38;
-    int64_t h7 = f0g7+f1g6   +f2g5   +f3g4   +f4g3   +f5g2   +f6g1   +f7g0   +f8g9_19+f9g8_19;
-    int64_t h8 = f0g8+f1g7_2 +f2g6   +f3g5_2 +f4g4   +f5g3_2 +f6g2   +f7g1_2 +f8g0   +f9g9_38;
-    int64_t h9 = f0g9+f1g8   +f2g7   +f3g6   +f4g5   +f5g4   +f6g3   +f7g2   +f8g1   +f9g0   ;
+    int64_t h0 = f0g0 + f1g9_38 + f2g8_19 + f3g7_38 + f4g6_19 + f5g5_38 + f6g4_19 + f7g3_38 + f8g2_19 + f9g1_38;
+    int64_t h1 = f0g1 + f1g0    + f2g9_19 + f3g8_19 + f4g7_19 + f5g6_19 + f6g5_19 + f7g4_19 + f8g3_19 + f9g2_19;
+    int64_t h2 = f0g2 + f1g1_2  + f2g0    + f3g9_38 + f4g8_19 + f5g7_38 + f6g6_19 + f7g5_38 + f8g4_19 + f9g3_38;
+    int64_t h3 = f0g3 + f1g2    + f2g1    + f3g0    + f4g9_19 + f5g8_19 + f6g7_19 + f7g6_19 + f8g5_19 + f9g4_19;
+    int64_t h4 = f0g4 + f1g3_2  + f2g2    + f3g1_2  + f4g0    + f5g9_38 + f6g8_19 + f7g7_38 + f8g6_19 + f9g5_38;
+    int64_t h5 = f0g5 + f1g4    + f2g3    + f3g2    + f4g1    + f5g0    + f6g9_19 + f7g8_19 + f8g7_19 + f9g6_19;
+    int64_t h6 = f0g6 + f1g5_2  + f2g4    + f3g3_2  + f4g2    + f5g1_2  + f6g0    + f7g9_38 + f8g8_19 + f9g7_38;
+    int64_t h7 = f0g7 + f1g6    + f2g5    + f3g4    + f4g3    + f5g2    + f6g1    + f7g0    + f8g9_19 + f9g8_19;
+    int64_t h8 = f0g8 + f1g7_2  + f2g6    + f3g5_2  + f4g4    + f5g3_2  + f6g2    + f7g1_2  + f8g0    + f9g9_38;
+    int64_t h9 = f0g9 + f1g8    + f2g7    + f3g6    + f4g5    + f5g4    + f6g3    + f7g2    + f8g1    + f9g0   ;
     int64_t carry0;
     int64_t carry1;
     int64_t carry2;
@@ -1340,16 +1341,16 @@ static void fe_sq(fe h, const fe f)
     int64_t f8f8_19 = f8   * (int64_t) f8_19;
     int64_t f8f9_38 = f8   * (int64_t) f9_38;
     int64_t f9f9_38 = f9   * (int64_t) f9_38;
-    int64_t h0 = f0f0  +f1f9_76+f2f8_38+f3f7_76+f4f6_38+f5f5_38;
-    int64_t h1 = f0f1_2+f2f9_38+f3f8_38+f4f7_38+f5f6_38;
-    int64_t h2 = f0f2_2+f1f1_2 +f3f9_76+f4f8_38+f5f7_76+f6f6_19;
-    int64_t h3 = f0f3_2+f1f2_2 +f4f9_38+f5f8_38+f6f7_38;
-    int64_t h4 = f0f4_2+f1f3_4 +f2f2   +f5f9_76+f6f8_38+f7f7_38;
-    int64_t h5 = f0f5_2+f1f4_2 +f2f3_2 +f6f9_38+f7f8_38;
-    int64_t h6 = f0f6_2+f1f5_4 +f2f4_2 +f3f3_2 +f7f9_76+f8f8_19;
-    int64_t h7 = f0f7_2+f1f6_2 +f2f5_2 +f3f4_2 +f8f9_38;
-    int64_t h8 = f0f8_2+f1f7_4 +f2f6_2 +f3f5_4 +f4f4   +f9f9_38;
-    int64_t h9 = f0f9_2+f1f8_2 +f2f7_2 +f3f6_2 +f4f5_2;
+    int64_t h0 = f0f0   + f1f9_76 + f2f8_38 + f3f7_76 + f4f6_38 + f5f5_38;
+    int64_t h1 = f0f1_2 + f2f9_38 + f3f8_38 + f4f7_38 + f5f6_38;
+    int64_t h2 = f0f2_2 + f1f1_2  + f3f9_76 + f4f8_38 + f5f7_76 + f6f6_19;
+    int64_t h3 = f0f3_2 + f1f2_2  + f4f9_38 + f5f8_38 + f6f7_38;
+    int64_t h4 = f0f4_2 + f1f3_4  + f2f2    + f5f9_76 + f6f8_38 + f7f7_38;
+    int64_t h5 = f0f5_2 + f1f4_2  + f2f3_2  + f6f9_38 + f7f8_38;
+    int64_t h6 = f0f6_2 + f1f5_4  + f2f4_2  + f3f3_2  + f7f9_76 + f8f8_19;
+    int64_t h7 = f0f7_2 + f1f6_2  + f2f5_2  + f3f4_2  + f8f9_38;
+    int64_t h8 = f0f8_2 + f1f7_4  + f2f6_2  + f3f5_4  + f4f4    + f9f9_38;
+    int64_t h9 = f0f9_2 + f1f8_2  + f2f7_2  + f3f6_2  + f4f5_2;
     int64_t carry0;
     int64_t carry1;
     int64_t carry2;
@@ -1649,16 +1650,16 @@ static void fe_sq2(fe h, const fe f)
     int64_t f8f8_19 = f8   * (int64_t) f8_19;
     int64_t f8f9_38 = f8   * (int64_t) f9_38;
     int64_t f9f9_38 = f9   * (int64_t) f9_38;
-    int64_t h0 = f0f0  +f1f9_76+f2f8_38+f3f7_76+f4f6_38+f5f5_38;
-    int64_t h1 = f0f1_2+f2f9_38+f3f8_38+f4f7_38+f5f6_38;
-    int64_t h2 = f0f2_2+f1f1_2 +f3f9_76+f4f8_38+f5f7_76+f6f6_19;
-    int64_t h3 = f0f3_2+f1f2_2 +f4f9_38+f5f8_38+f6f7_38;
-    int64_t h4 = f0f4_2+f1f3_4 +f2f2   +f5f9_76+f6f8_38+f7f7_38;
-    int64_t h5 = f0f5_2+f1f4_2 +f2f3_2 +f6f9_38+f7f8_38;
-    int64_t h6 = f0f6_2+f1f5_4 +f2f4_2 +f3f3_2 +f7f9_76+f8f8_19;
-    int64_t h7 = f0f7_2+f1f6_2 +f2f5_2 +f3f4_2 +f8f9_38;
-    int64_t h8 = f0f8_2+f1f7_4 +f2f6_2 +f3f5_4 +f4f4   +f9f9_38;
-    int64_t h9 = f0f9_2+f1f8_2 +f2f7_2 +f3f6_2 +f4f5_2;
+    int64_t h0 = f0f0   + f1f9_76 + f2f8_38 + f3f7_76 + f4f6_38 + f5f5_38;
+    int64_t h1 = f0f1_2 + f2f9_38 + f3f8_38 + f4f7_38 + f5f6_38;
+    int64_t h2 = f0f2_2 + f1f1_2  + f3f9_76 + f4f8_38 + f5f7_76 + f6f6_19;
+    int64_t h3 = f0f3_2 + f1f2_2  + f4f9_38 + f5f8_38 + f6f7_38;
+    int64_t h4 = f0f4_2 + f1f3_4  + f2f2    + f5f9_76 + f6f8_38 + f7f7_38;
+    int64_t h5 = f0f5_2 + f1f4_2  + f2f3_2  + f6f9_38 + f7f8_38;
+    int64_t h6 = f0f6_2 + f1f5_4  + f2f4_2  + f3f3_2  + f7f9_76 + f8f8_19;
+    int64_t h7 = f0f7_2 + f1f6_2  + f2f5_2  + f3f4_2  + f8f9_38;
+    int64_t h8 = f0f8_2 + f1f7_4  + f2f6_2  + f3f5_4  + f4f4    + f9f9_38;
+    int64_t h9 = f0f9_2 + f1f8_2  + f2f7_2  + f3f6_2  + f4f5_2;
     int64_t carry0;
     int64_t carry1;
     int64_t carry2;
@@ -4612,30 +4613,30 @@ static void ge_double_scalarmult_vartime(ge_p2 *r, const uint8_t *a,
 */
 static void x25519_sc_reduce(uint8_t *s)
 {
-    int64_t s0 = 2097151 & load_3(s);
-    int64_t s1 = 2097151 & (load_4(s + 2) >> 5);
-    int64_t s2 = 2097151 & (load_3(s + 5) >> 2);
-    int64_t s3 = 2097151 & (load_4(s + 7) >> 7);
-    int64_t s4 = 2097151 & (load_4(s + 10) >> 4);
-    int64_t s5 = 2097151 & (load_3(s + 13) >> 1);
-    int64_t s6 = 2097151 & (load_4(s + 15) >> 6);
-    int64_t s7 = 2097151 & (load_3(s + 18) >> 3);
-    int64_t s8 = 2097151 & load_3(s + 21);
-    int64_t s9 = 2097151 & (load_4(s + 23) >> 5);
-    int64_t s10 = 2097151 & (load_3(s + 26) >> 2);
-    int64_t s11 = 2097151 & (load_4(s + 28) >> 7);
-    int64_t s12 = 2097151 & (load_4(s + 31) >> 4);
-    int64_t s13 = 2097151 & (load_3(s + 34) >> 1);
-    int64_t s14 = 2097151 & (load_4(s + 36) >> 6);
-    int64_t s15 = 2097151 & (load_3(s + 39) >> 3);
-    int64_t s16 = 2097151 & load_3(s + 42);
-    int64_t s17 = 2097151 & (load_4(s + 44) >> 5);
-    int64_t s18 = 2097151 & (load_3(s + 47) >> 2);
-    int64_t s19 = 2097151 & (load_4(s + 49) >> 7);
-    int64_t s20 = 2097151 & (load_4(s + 52) >> 4);
-    int64_t s21 = 2097151 & (load_3(s + 55) >> 1);
-    int64_t s22 = 2097151 & (load_4(s + 57) >> 6);
-    int64_t s23 = (load_4(s + 60) >> 3);
+    int64_t s0  = kBottom21Bits &  load_3(s);
+    int64_t s1  = kBottom21Bits & (load_4(s +  2) >> 5);
+    int64_t s2  = kBottom21Bits & (load_3(s +  5) >> 2);
+    int64_t s3  = kBottom21Bits & (load_4(s +  7) >> 7);
+    int64_t s4  = kBottom21Bits & (load_4(s + 10) >> 4);
+    int64_t s5  = kBottom21Bits & (load_3(s + 13) >> 1);
+    int64_t s6  = kBottom21Bits & (load_4(s + 15) >> 6);
+    int64_t s7  = kBottom21Bits & (load_3(s + 18) >> 3);
+    int64_t s8  = kBottom21Bits &  load_3(s + 21);
+    int64_t s9  = kBottom21Bits & (load_4(s + 23) >> 5);
+    int64_t s10 = kBottom21Bits & (load_3(s + 26) >> 2);
+    int64_t s11 = kBottom21Bits & (load_4(s + 28) >> 7);
+    int64_t s12 = kBottom21Bits & (load_4(s + 31) >> 4);
+    int64_t s13 = kBottom21Bits & (load_3(s + 34) >> 1);
+    int64_t s14 = kBottom21Bits & (load_4(s + 36) >> 6);
+    int64_t s15 = kBottom21Bits & (load_3(s + 39) >> 3);
+    int64_t s16 = kBottom21Bits &  load_3(s + 42);
+    int64_t s17 = kBottom21Bits & (load_4(s + 44) >> 5);
+    int64_t s18 = kBottom21Bits & (load_3(s + 47) >> 2);
+    int64_t s19 = kBottom21Bits & (load_4(s + 49) >> 7);
+    int64_t s20 = kBottom21Bits & (load_4(s + 52) >> 4);
+    int64_t s21 = kBottom21Bits & (load_3(s + 55) >> 1);
+    int64_t s22 = kBottom21Bits & (load_4(s + 57) >> 6);
+    int64_t s23 =                 (load_4(s + 60) >> 3);
     int64_t carry0;
     int64_t carry1;
     int64_t carry2;
@@ -4660,7 +4661,7 @@ static void x25519_sc_reduce(uint8_t *s)
     s14 -= s23 * 997805;
     s15 += s23 * 136657;
     s16 -= s23 * 683901;
-    s23 = 0;
+    s23  = 0;
 
     s10 += s22 * 666643;
     s11 += s22 * 470296;
@@ -4668,46 +4669,46 @@ static void x25519_sc_reduce(uint8_t *s)
     s13 -= s22 * 997805;
     s14 += s22 * 136657;
     s15 -= s22 * 683901;
-    s22 = 0;
+    s22  = 0;
 
-    s9 += s21 * 666643;
+    s9  += s21 * 666643;
     s10 += s21 * 470296;
     s11 += s21 * 654183;
     s12 -= s21 * 997805;
     s13 += s21 * 136657;
     s14 -= s21 * 683901;
-    s21 = 0;
+    s21  = 0;
 
-    s8 += s20 * 666643;
-    s9 += s20 * 470296;
+    s8  += s20 * 666643;
+    s9  += s20 * 470296;
     s10 += s20 * 654183;
     s11 -= s20 * 997805;
     s12 += s20 * 136657;
     s13 -= s20 * 683901;
-    s20 = 0;
+    s20  = 0;
 
-    s7 += s19 * 666643;
-    s8 += s19 * 470296;
-    s9 += s19 * 654183;
+    s7  += s19 * 666643;
+    s8  += s19 * 470296;
+    s9  += s19 * 654183;
     s10 -= s19 * 997805;
     s11 += s19 * 136657;
     s12 -= s19 * 683901;
-    s19 = 0;
+    s19  = 0;
 
-    s6 += s18 * 666643;
-    s7 += s18 * 470296;
-    s8 += s18 * 654183;
-    s9 -= s18 * 997805;
+    s6  += s18 * 666643;
+    s7  += s18 * 470296;
+    s8  += s18 * 654183;
+    s9  -= s18 * 997805;
     s10 += s18 * 136657;
     s11 -= s18 * 683901;
-    s18 = 0;
+    s18  = 0;
 
     carry6 = (s6 + (1 << 20)) >> 21;
-    s7 += carry6;
-    s6 -= carry6 * (1 << 21);
+    s7  += carry6;
+    s6  -= carry6 * (1 << 21);
     carry8 = (s8 + (1 << 20)) >> 21;
-    s9 += carry8;
-    s8 -= carry8 * (1 << 21);
+    s9  += carry8;
+    s8  -= carry8 * (1 << 21);
     carry10 = (s10 + (1 << 20)) >> 21;
     s11 += carry10;
     s10 -= carry10 * (1 << 21);
@@ -4722,11 +4723,11 @@ static void x25519_sc_reduce(uint8_t *s)
     s16 -= carry16 * (1 << 21);
 
     carry7 = (s7 + (1 << 20)) >> 21;
-    s8 += carry7;
-    s7 -= carry7 * (1 << 21);
+    s8  += carry7;
+    s7  -= carry7 * (1 << 21);
     carry9 = (s9 + (1 << 20)) >> 21;
     s10 += carry9;
-    s9 -= carry9 * (1 << 21);
+    s9  -= carry9 * (1 << 21);
     carry11 = (s11 + (1 << 20)) >> 21;
     s12 += carry11;
     s11 -= carry11 * (1 << 21);
@@ -4737,130 +4738,130 @@ static void x25519_sc_reduce(uint8_t *s)
     s16 += carry15;
     s15 -= carry15 * (1 << 21);
 
-    s5 += s17 * 666643;
-    s6 += s17 * 470296;
-    s7 += s17 * 654183;
-    s8 -= s17 * 997805;
-    s9 += s17 * 136657;
+    s5  += s17 * 666643;
+    s6  += s17 * 470296;
+    s7  += s17 * 654183;
+    s8  -= s17 * 997805;
+    s9  += s17 * 136657;
     s10 -= s17 * 683901;
-    s17 = 0;
-
-    s4 += s16 * 666643;
-    s5 += s16 * 470296;
-    s6 += s16 * 654183;
-    s7 -= s16 * 997805;
-    s8 += s16 * 136657;
-    s9 -= s16 * 683901;
-    s16 = 0;
-
-    s3 += s15 * 666643;
-    s4 += s15 * 470296;
-    s5 += s15 * 654183;
-    s6 -= s15 * 997805;
-    s7 += s15 * 136657;
-    s8 -= s15 * 683901;
-    s15 = 0;
-
-    s2 += s14 * 666643;
-    s3 += s14 * 470296;
-    s4 += s14 * 654183;
-    s5 -= s14 * 997805;
-    s6 += s14 * 136657;
-    s7 -= s14 * 683901;
-    s14 = 0;
-
-    s1 += s13 * 666643;
-    s2 += s13 * 470296;
-    s3 += s13 * 654183;
-    s4 -= s13 * 997805;
-    s5 += s13 * 136657;
-    s6 -= s13 * 683901;
-    s13 = 0;
-
-    s0 += s12 * 666643;
-    s1 += s12 * 470296;
-    s2 += s12 * 654183;
-    s3 -= s12 * 997805;
-    s4 += s12 * 136657;
-    s5 -= s12 * 683901;
-    s12 = 0;
+    s17  = 0;
+
+    s4  += s16 * 666643;
+    s5  += s16 * 470296;
+    s6  += s16 * 654183;
+    s7  -= s16 * 997805;
+    s8  += s16 * 136657;
+    s9  -= s16 * 683901;
+    s16  = 0;
+
+    s3  += s15 * 666643;
+    s4  += s15 * 470296;
+    s5  += s15 * 654183;
+    s6  -= s15 * 997805;
+    s7  += s15 * 136657;
+    s8  -= s15 * 683901;
+    s15  = 0;
+
+    s2  += s14 * 666643;
+    s3  += s14 * 470296;
+    s4  += s14 * 654183;
+    s5  -= s14 * 997805;
+    s6  += s14 * 136657;
+    s7  -= s14 * 683901;
+    s14  = 0;
+
+    s1  += s13 * 666643;
+    s2  += s13 * 470296;
+    s3  += s13 * 654183;
+    s4  -= s13 * 997805;
+    s5  += s13 * 136657;
+    s6  -= s13 * 683901;
+    s13  = 0;
+
+    s0  += s12 * 666643;
+    s1  += s12 * 470296;
+    s2  += s12 * 654183;
+    s3  -= s12 * 997805;
+    s4  += s12 * 136657;
+    s5  -= s12 * 683901;
+    s12  = 0;
 
     carry0 = (s0 + (1 << 20)) >> 21;
-    s1 += carry0;
-    s0 -= carry0 * (1 << 21);
+    s1  += carry0;
+    s0  -= carry0 * (1 << 21);
     carry2 = (s2 + (1 << 20)) >> 21;
-    s3 += carry2;
-    s2 -= carry2 * (1 << 21);
+    s3  += carry2;
+    s2  -= carry2 * (1 << 21);
     carry4 = (s4 + (1 << 20)) >> 21;
-    s5 += carry4;
-    s4 -= carry4 * (1 << 21);
+    s5  += carry4;
+    s4  -= carry4 * (1 << 21);
     carry6 = (s6 + (1 << 20)) >> 21;
     s7 += carry6;
     s6 -= carry6 * (1 << 21);
     carry8 = (s8 + (1 << 20)) >> 21;
-    s9 += carry8;
-    s8 -= carry8 * (1 << 21);
+    s9  += carry8;
+    s8  -= carry8 * (1 << 21);
     carry10 = (s10 + (1 << 20)) >> 21;
     s11 += carry10;
     s10 -= carry10 * (1 << 21);
 
     carry1 = (s1 + (1 << 20)) >> 21;
-    s2 += carry1;
-    s1 -= carry1 * (1 << 21);
+    s2  += carry1;
+    s1  -= carry1 * (1 << 21);
     carry3 = (s3 + (1 << 20)) >> 21;
-    s4 += carry3;
-    s3 -= carry3 * (1 << 21);
+    s4  += carry3;
+    s3  -= carry3 * (1 << 21);
     carry5 = (s5 + (1 << 20)) >> 21;
-    s6 += carry5;
-    s5 -= carry5 * (1 << 21);
+    s6  += carry5;
+    s5  -= carry5 * (1 << 21);
     carry7 = (s7 + (1 << 20)) >> 21;
-    s8 += carry7;
-    s7 -= carry7 * (1 << 21);
+    s8  += carry7;
+    s7  -= carry7 * (1 << 21);
     carry9 = (s9 + (1 << 20)) >> 21;
     s10 += carry9;
-    s9 -= carry9 * (1 << 21);
+    s9  -= carry9 * (1 << 21);
     carry11 = (s11 + (1 << 20)) >> 21;
     s12 += carry11;
     s11 -= carry11 * (1 << 21);
 
-    s0 += s12 * 666643;
-    s1 += s12 * 470296;
-    s2 += s12 * 654183;
-    s3 -= s12 * 997805;
-    s4 += s12 * 136657;
-    s5 -= s12 * 683901;
-    s12 = 0;
+    s0  += s12 * 666643;
+    s1  += s12 * 470296;
+    s2  += s12 * 654183;
+    s3  -= s12 * 997805;
+    s4  += s12 * 136657;
+    s5  -= s12 * 683901;
+    s12  = 0;
 
     carry0 = s0 >> 21;
-    s1 += carry0;
-    s0 -= carry0 * (1 << 21);
+    s1  += carry0;
+    s0  -= carry0 * (1 << 21);
     carry1 = s1 >> 21;
-    s2 += carry1;
-    s1 -= carry1 * (1 << 21);
+    s2  += carry1;
+    s1  -= carry1 * (1 << 21);
     carry2 = s2 >> 21;
-    s3 += carry2;
-    s2 -= carry2 * (1 << 21);
+    s3  += carry2;
+    s2  -= carry2 * (1 << 21);
     carry3 = s3 >> 21;
-    s4 += carry3;
-    s3 -= carry3 * (1 << 21);
+    s4  += carry3;
+    s3  -= carry3 * (1 << 21);
     carry4 = s4 >> 21;
-    s5 += carry4;
-    s4 -= carry4 * (1 << 21);
+    s5  += carry4;
+    s4  -= carry4 * (1 << 21);
     carry5 = s5 >> 21;
-    s6 += carry5;
-    s5 -= carry5 * (1 << 21);
+    s6  += carry5;
+    s5  -= carry5 * (1 << 21);
     carry6 = s6 >> 21;
-    s7 += carry6;
-    s6 -= carry6 * (1 << 21);
+    s7  += carry6;
+    s6  -= carry6 * (1 << 21);
     carry7 = s7 >> 21;
-    s8 += carry7;
-    s7 -= carry7 * (1 << 21);
+    s8  += carry7;
+    s7  -= carry7 * (1 << 21);
     carry8 = s8 >> 21;
-    s9 += carry8;
-    s8 -= carry8 * (1 << 21);
+    s9  += carry8;
+    s8  -= carry8 * (1 << 21);
     carry9 = s9 >> 21;
     s10 += carry9;
-    s9 -= carry9 * (1 << 21);
+    s9  -= carry9 * (1 << 21);
     carry10 = s10 >> 21;
     s11 += carry10;
     s10 -= carry10 * (1 << 21);
@@ -4868,80 +4869,80 @@ static void x25519_sc_reduce(uint8_t *s)
     s12 += carry11;
     s11 -= carry11 * (1 << 21);
 
-    s0 += s12 * 666643;
-    s1 += s12 * 470296;
-    s2 += s12 * 654183;
-    s3 -= s12 * 997805;
-    s4 += s12 * 136657;
-    s5 -= s12 * 683901;
-    s12 = 0;
+    s0  += s12 * 666643;
+    s1  += s12 * 470296;
+    s2  += s12 * 654183;
+    s3  -= s12 * 997805;
+    s4  += s12 * 136657;
+    s5  -= s12 * 683901;
+    s12  = 0;
 
     carry0 = s0 >> 21;
-    s1 += carry0;
-    s0 -= carry0 * (1 << 21);
+    s1  += carry0;
+    s0  -= carry0 * (1 << 21);
     carry1 = s1 >> 21;
-    s2 += carry1;
-    s1 -= carry1 * (1 << 21);
+    s2  += carry1;
+    s1  -= carry1 * (1 << 21);
     carry2 = s2 >> 21;
-    s3 += carry2;
-    s2 -= carry2 * (1 << 21);
+    s3  += carry2;
+    s2  -= carry2 * (1 << 21);
     carry3 = s3 >> 21;
-    s4 += carry3;
-    s3 -= carry3 * (1 << 21);
+    s4  += carry3;
+    s3  -= carry3 * (1 << 21);
     carry4 = s4 >> 21;
-    s5 += carry4;
-    s4 -= carry4 * (1 << 21);
+    s5  += carry4;
+    s4  -= carry4 * (1 << 21);
     carry5 = s5 >> 21;
-    s6 += carry5;
-    s5 -= carry5 * (1 << 21);
+    s6  += carry5;
+    s5  -= carry5 * (1 << 21);
     carry6 = s6 >> 21;
-    s7 += carry6;
-    s6 -= carry6 * (1 << 21);
+    s7  += carry6;
+    s6  -= carry6 * (1 << 21);
     carry7 = s7 >> 21;
-    s8 += carry7;
-    s7 -= carry7 * (1 << 21);
+    s8  += carry7;
+    s7  -= carry7 * (1 << 21);
     carry8 = s8 >> 21;
-    s9 += carry8;
-    s8 -= carry8 * (1 << 21);
+    s9  += carry8;
+    s8  -= carry8 * (1 << 21);
     carry9 = s9 >> 21;
     s10 += carry9;
-    s9 -= carry9 * (1 << 21);
+    s9  -= carry9 * (1 << 21);
     carry10 = s10 >> 21;
     s11 += carry10;
     s10 -= carry10 * (1 << 21);
 
-    s[0] = (uint8_t)(s0 >> 0);
-    s[1] = (uint8_t)(s0 >> 8);
-    s[2] = (uint8_t)((s0 >> 16) | (s1 << 5));
-    s[3] = (uint8_t)(s1 >> 3);
-    s[4] = (uint8_t)(s1 >> 11);
-    s[5] = (uint8_t)((s1 >> 19) | (s2 << 2));
-    s[6] = (uint8_t)(s2 >> 6);
-    s[7] = (uint8_t)((s2 >> 14) | (s3 << 7));
-    s[8] = (uint8_t)(s3 >> 1);
-    s[9] = (uint8_t)(s3 >> 9);
-    s[10] = (uint8_t)((s3 >> 17) | (s4 << 4));
-    s[11] = (uint8_t)(s4 >> 4);
-    s[12] = (uint8_t)(s4 >> 12);
-    s[13] = (uint8_t)((s4 >> 20) | (s5 << 1));
-    s[14] = (uint8_t)(s5 >> 7);
-    s[15] = (uint8_t)((s5 >> 15) | (s6 << 6));
-    s[16] = (uint8_t)(s6 >> 2);
-    s[17] = (uint8_t)(s6 >> 10);
-    s[18] = (uint8_t)((s6 >> 18) | (s7 << 3));
-    s[19] = (uint8_t)(s7 >> 5);
-    s[20] = (uint8_t)(s7 >> 13);
-    s[21] = (uint8_t)(s8 >> 0);
-    s[22] = (uint8_t)(s8 >> 8);
-    s[23] = (uint8_t)((s8 >> 16) | (s9 << 5));
-    s[24] = (uint8_t)(s9 >> 3);
-    s[25] = (uint8_t)(s9 >> 11);
-    s[26] = (uint8_t)((s9 >> 19) | (s10 << 2));
-    s[27] = (uint8_t)(s10 >> 6);
-    s[28] = (uint8_t)((s10 >> 14) | (s11 << 7));
-    s[29] = (uint8_t)(s11 >> 1);
-    s[30] = (uint8_t)(s11 >> 9);
-    s[31] = (uint8_t)(s11 >> 17);
+    s[ 0] = (uint8_t) (s0  >>  0);
+    s[ 1] = (uint8_t) (s0  >>  8);
+    s[ 2] = (uint8_t)((s0  >> 16) | (s1  <<  5));
+    s[ 3] = (uint8_t) (s1  >>  3);
+    s[ 4] = (uint8_t) (s1  >> 11);
+    s[ 5] = (uint8_t)((s1  >> 19) | (s2  <<  2));
+    s[ 6] = (uint8_t) (s2  >>  6);
+    s[ 7] = (uint8_t)((s2  >> 14) | (s3  <<  7));
+    s[ 8] = (uint8_t) (s3  >>  1);
+    s[ 9] = (uint8_t) (s3  >>  9);
+    s[10] = (uint8_t)((s3  >> 17) | (s4  <<  4));
+    s[11] = (uint8_t) (s4  >>  4);
+    s[12] = (uint8_t) (s4  >> 12);
+    s[13] = (uint8_t)((s4  >> 20) | (s5  <<  1));
+    s[14] = (uint8_t) (s5  >>  7);
+    s[15] = (uint8_t)((s5  >> 15) | (s6  <<  6));
+    s[16] = (uint8_t) (s6  >>  2);
+    s[17] = (uint8_t) (s6  >> 10);
+    s[18] = (uint8_t)((s6  >> 18) | (s7  <<  3));
+    s[19] = (uint8_t) (s7  >>  5);
+    s[20] = (uint8_t) (s7  >> 13);
+    s[21] = (uint8_t) (s8  >>  0);
+    s[22] = (uint8_t) (s8  >>  8);
+    s[23] = (uint8_t)((s8  >> 16) | (s9  <<  5));
+    s[24] = (uint8_t) (s9  >>  3);
+    s[25] = (uint8_t) (s9  >> 11);
+    s[26] = (uint8_t)((s9  >> 19) | (s10 <<  2));
+    s[27] = (uint8_t) (s10 >>  6);
+    s[28] = (uint8_t)((s10 >> 14) | (s11 <<  7));
+    s[29] = (uint8_t) (s11 >>  1);
+    s[30] = (uint8_t) (s11 >>  9);
+    s[31] = (uint8_t) (s11 >> 17);
 }
 
 /*
@@ -4957,42 +4958,42 @@ static void x25519_sc_reduce(uint8_t *s)
 static void sc_muladd(uint8_t *s, const uint8_t *a, const uint8_t *b,
                       const uint8_t *c)
 {
-    int64_t a0 = 2097151 & load_3(a);
-    int64_t a1 = 2097151 & (load_4(a + 2) >> 5);
-    int64_t a2 = 2097151 & (load_3(a + 5) >> 2);
-    int64_t a3 = 2097151 & (load_4(a + 7) >> 7);
-    int64_t a4 = 2097151 & (load_4(a + 10) >> 4);
-    int64_t a5 = 2097151 & (load_3(a + 13) >> 1);
-    int64_t a6 = 2097151 & (load_4(a + 15) >> 6);
-    int64_t a7 = 2097151 & (load_3(a + 18) >> 3);
-    int64_t a8 = 2097151 & load_3(a + 21);
-    int64_t a9 = 2097151 & (load_4(a + 23) >> 5);
-    int64_t a10 = 2097151 & (load_3(a + 26) >> 2);
-    int64_t a11 = (load_4(a + 28) >> 7);
-    int64_t b0 = 2097151 & load_3(b);
-    int64_t b1 = 2097151 & (load_4(b + 2) >> 5);
-    int64_t b2 = 2097151 & (load_3(b + 5) >> 2);
-    int64_t b3 = 2097151 & (load_4(b + 7) >> 7);
-    int64_t b4 = 2097151 & (load_4(b + 10) >> 4);
-    int64_t b5 = 2097151 & (load_3(b + 13) >> 1);
-    int64_t b6 = 2097151 & (load_4(b + 15) >> 6);
-    int64_t b7 = 2097151 & (load_3(b + 18) >> 3);
-    int64_t b8 = 2097151 & load_3(b + 21);
-    int64_t b9 = 2097151 & (load_4(b + 23) >> 5);
-    int64_t b10 = 2097151 & (load_3(b + 26) >> 2);
-    int64_t b11 = (load_4(b + 28) >> 7);
-    int64_t c0 = 2097151 & load_3(c);
-    int64_t c1 = 2097151 & (load_4(c + 2) >> 5);
-    int64_t c2 = 2097151 & (load_3(c + 5) >> 2);
-    int64_t c3 = 2097151 & (load_4(c + 7) >> 7);
-    int64_t c4 = 2097151 & (load_4(c + 10) >> 4);
-    int64_t c5 = 2097151 & (load_3(c + 13) >> 1);
-    int64_t c6 = 2097151 & (load_4(c + 15) >> 6);
-    int64_t c7 = 2097151 & (load_3(c + 18) >> 3);
-    int64_t c8 = 2097151 & load_3(c + 21);
-    int64_t c9 = 2097151 & (load_4(c + 23) >> 5);
-    int64_t c10 = 2097151 & (load_3(c + 26) >> 2);
-    int64_t c11 = (load_4(c + 28) >> 7);
+    int64_t a0  = kBottom21Bits &  load_3(a);
+    int64_t a1  = kBottom21Bits & (load_4(a +  2) >> 5);
+    int64_t a2  = kBottom21Bits & (load_3(a +  5) >> 2);
+    int64_t a3  = kBottom21Bits & (load_4(a +  7) >> 7);
+    int64_t a4  = kBottom21Bits & (load_4(a + 10) >> 4);
+    int64_t a5  = kBottom21Bits & (load_3(a + 13) >> 1);
+    int64_t a6  = kBottom21Bits & (load_4(a + 15) >> 6);
+    int64_t a7  = kBottom21Bits & (load_3(a + 18) >> 3);
+    int64_t a8  = kBottom21Bits &  load_3(a + 21);
+    int64_t a9  = kBottom21Bits & (load_4(a + 23) >> 5);
+    int64_t a10 = kBottom21Bits & (load_3(a + 26) >> 2);
+    int64_t a11 =                 (load_4(a + 28) >> 7);
+    int64_t b0  = kBottom21Bits &  load_3(b);
+    int64_t b1  = kBottom21Bits & (load_4(b +  2) >> 5);
+    int64_t b2  = kBottom21Bits & (load_3(b +  5) >> 2);
+    int64_t b3  = kBottom21Bits & (load_4(b +  7) >> 7);
+    int64_t b4  = kBottom21Bits & (load_4(b + 10) >> 4);
+    int64_t b5  = kBottom21Bits & (load_3(b + 13) >> 1);
+    int64_t b6  = kBottom21Bits & (load_4(b + 15) >> 6);
+    int64_t b7  = kBottom21Bits & (load_3(b + 18) >> 3);
+    int64_t b8  = kBottom21Bits &  load_3(b + 21);
+    int64_t b9  = kBottom21Bits & (load_4(b + 23) >> 5);
+    int64_t b10 = kBottom21Bits & (load_3(b + 26) >> 2);
+    int64_t b11 =                 (load_4(b + 28) >> 7);
+    int64_t c0  = kBottom21Bits &  load_3(c);
+    int64_t c1  = kBottom21Bits & (load_4(c +  2) >> 5);
+    int64_t c2  = kBottom21Bits & (load_3(c +  5) >> 2);
+    int64_t c3  = kBottom21Bits & (load_4(c +  7) >> 7);
+    int64_t c4  = kBottom21Bits & (load_4(c + 10) >> 4);
+    int64_t c5  = kBottom21Bits & (load_3(c + 13) >> 1);
+    int64_t c6  = kBottom21Bits & (load_4(c + 15) >> 6);
+    int64_t c7  = kBottom21Bits & (load_3(c + 18) >> 3);
+    int64_t c8  = kBottom21Bits &  load_3(c + 21);
+    int64_t c9  = kBottom21Bits & (load_4(c + 23) >> 5);
+    int64_t c10 = kBottom21Bits & (load_3(c + 26) >> 2);
+    int64_t c11 =                 (load_4(c + 28) >> 7);
     int64_t s0;
     int64_t s1;
     int64_t s2;
@@ -5041,57 +5042,46 @@ static void sc_muladd(uint8_t *s, const uint8_t *a, const uint8_t *b,
     int64_t carry21;
     int64_t carry22;
 
-    s0 = c0 + a0 * b0;
-    s1 = c1 + a0 * b1 + a1 * b0;
-    s2 = c2 + a0 * b2 + a1 * b1 + a2 * b0;
-    s3 = c3 + a0 * b3 + a1 * b2 + a2 * b1 + a3 * b0;
-    s4 = c4 + a0 * b4 + a1 * b3 + a2 * b2 + a3 * b1 + a4 * b0;
-    s5 = c5 + a0 * b5 + a1 * b4 + a2 * b3 + a3 * b2 + a4 * b1 + a5 * b0;
-    s6 = c6 + a0 * b6 + a1 * b5 + a2 * b4 + a3 * b3 + a4 * b2 + a5 * b1 +
-        a6 * b0;
-    s7 = c7 + a0 * b7 + a1 * b6 + a2 * b5 + a3 * b4 + a4 * b3 + a5 * b2 +
-        a6 * b1 + a7 * b0;
-    s8 = c8 + a0 * b8 + a1 * b7 + a2 * b6 + a3 * b5 + a4 * b4 + a5 * b3 +
-        a6 * b2 + a7 * b1 + a8 * b0;
-    s9 = c9 + a0 * b9 + a1 * b8 + a2 * b7 + a3 * b6 + a4 * b5 + a5 * b4 +
-        a6 * b3 + a7 * b2 + a8 * b1 + a9 * b0;
-    s10 = c10 + a0 * b10 + a1 * b9 + a2 * b8 + a3 * b7 + a4 * b6 + a5 * b5 +
-        a6 * b4 + a7 * b3 + a8 * b2 + a9 * b1 + a10 * b0;
-    s11 = c11 + a0 * b11 + a1 * b10 + a2 * b9 + a3 * b8 + a4 * b7 + a5 * b6 +
-        a6 * b5 + a7 * b4 + a8 * b3 + a9 * b2 + a10 * b1 + a11 * b0;
-    s12 = a1 * b11 + a2 * b10 + a3 * b9 + a4 * b8 + a5 * b7 + a6 * b6 +
-        a7 * b5 + a8 * b4 + a9 * b3 + a10 * b2 + a11 * b1;
-    s13 = a2 * b11 + a3 * b10 + a4 * b9 + a5 * b8 + a6 * b7 + a7 * b6 +
-        a8 * b5 + a9 * b4 + a10 * b3 + a11 * b2;
-    s14 = a3 * b11 + a4 * b10 + a5 * b9 + a6 * b8 + a7 * b7 + a8 * b6 +
-        a9 * b5 + a10 * b4 + a11 * b3;
-    s15 = a4 * b11 + a5 * b10 + a6 * b9 + a7 * b8 + a8 * b7 + a9 * b6 +
-        a10 * b5 + a11 * b4;
-    s16 = a5 * b11 + a6 * b10 + a7 * b9 + a8 * b8 + a9 * b7 + a10 * b6 +
-        a11 * b5;
-    s17 = a6 * b11 + a7 * b10 + a8 * b9 + a9 * b8 + a10 * b7 + a11 * b6;
-    s18 = a7 * b11 + a8 * b10 + a9 * b9 + a10 * b8 + a11 * b7;
-    s19 = a8 * b11 + a9 * b10 + a10 * b9 + a11 * b8;
-    s20 = a9 * b11 + a10 * b10 + a11 * b9;
-    s21 = a10 * b11 + a11 * b10;
-    s22 = a11 * b11;
-    s23 = 0;
+    s0  = c0   +   a0 * b0;
+    s1  = c1   +   a0 * b1   +   a1 * b0;
+    s2  = c2   +   a0 * b2   +   a1 * b1   +   a2 * b0;
+    s3  = c3   +   a0 * b3   +   a1 * b2   +   a2 * b1  +   a3 * b0;
+    s4  = c4   +   a0 * b4   +   a1 * b3   +   a2 * b2  +   a3 * b1  +   a4 * b0;
+    s5  = c5   +   a0 * b5   +   a1 * b4   +   a2 * b3  +   a3 * b2  +   a4 * b1  +   a5 * b0;
+    s6  = c6   +   a0 * b6   +   a1 * b5   +   a2 * b4  +   a3 * b3  +   a4 * b2  +   a5 * b1 +   a6 * b0;
+    s7  = c7   +   a0 * b7   +   a1 * b6   +   a2 * b5  +   a3 * b4  +   a4 * b3  +   a5 * b2 +   a6 * b1   +   a7 * b0;
+    s8  = c8   +   a0 * b8   +   a1 * b7   +   a2 * b6  +   a3 * b5  +   a4 * b4  +   a5 * b3 +   a6 * b2   +   a7 * b1   +   a8 * b0;
+    s9  = c9   +   a0 * b9   +   a1 * b8   +   a2 * b7  +   a3 * b6  +   a4 * b5  +   a5 * b4 +   a6 * b3   +   a7 * b2   +   a8 * b1  +   a9 * b0;
+    s10 = c10  +   a0 * b10  +   a1 * b9   +   a2 * b8  +   a3 * b7  +   a4 * b6  +   a5 * b5 +   a6 * b4   +   a7 * b3   +   a8 * b2  +   a9 * b1  +  a10 * b0;
+    s11 = c11  +   a0 * b11  +   a1 * b10  +   a2 * b9  +   a3 * b8  +   a4 * b7  +   a5 * b6 +   a6 * b5   +   a7 * b4   +   a8 * b3  +   a9 * b2  +  a10 * b1  +  a11 * b0;
+    s12 =          a1 * b11  +   a2 * b10  +   a3 * b9  +   a4 * b8  +   a5 * b7  +   a6 * b6 +   a7 * b5   +   a8 * b4   +   a9 * b3  +  a10 * b2  +  a11 * b1;
+    s13 =          a2 * b11  +   a3 * b10  +   a4 * b9  +   a5 * b8  +   a6 * b7  +   a7 * b6 +   a8 * b5   +   a9 * b4   +  a10 * b3  +  a11 * b2;
+    s14 =          a3 * b11  +   a4 * b10  +   a5 * b9  +   a6 * b8  +   a7 * b7  +   a8 * b6 +   a9 * b5   +  a10 * b4   +  a11 * b3;
+    s15 =          a4 * b11  +   a5 * b10  +   a6 * b9  +   a7 * b8  +   a8 * b7  +   a9 * b6 +  a10 * b5   +  a11 * b4;
+    s16 =          a5 * b11  +   a6 * b10  +   a7 * b9  +   a8 * b8  +   a9 * b7  +  a10 * b6 +  a11 * b5;
+    s17 =          a6 * b11  +   a7 * b10  +   a8 * b9  +   a9 * b8  +  a10 * b7  +  a11 * b6;
+    s18 =          a7 * b11  +   a8 * b10  +   a9 * b9  +  a10 * b8  +  a11 * b7;
+    s19 =          a8 * b11  +   a9 * b10  +  a10 * b9  +  a11 * b8;
+    s20 =          a9 * b11  +  a10 * b10  +  a11 * b9;
+    s21 =         a10 * b11  +  a11 * b10;
+    s22 =         a11 * b11;
+    s23 =         0;
 
     carry0 = (s0 + (1 << 20)) >> 21;
-    s1 += carry0;
-    s0 -= carry0 * (1 << 21);
+    s1  += carry0;
+    s0  -= carry0 * (1 << 21);
     carry2 = (s2 + (1 << 20)) >> 21;
-    s3 += carry2;
-    s2 -= carry2 * (1 << 21);
+    s3  += carry2;
+    s2  -= carry2 * (1 << 21);
     carry4 = (s4 + (1 << 20)) >> 21;
-    s5 += carry4;
-    s4 -= carry4 * (1 << 21);
+    s5  += carry4;
+    s4  -= carry4 * (1 << 21);
     carry6 = (s6 + (1 << 20)) >> 21;
-    s7 += carry6;
-    s6 -= carry6 * (1 << 21);
+    s7  += carry6;
+    s6  -= carry6 * (1 << 21);
     carry8 = (s8 + (1 << 20)) >> 21;
-    s9 += carry8;
-    s8 -= carry8 * (1 << 21);
+    s9  += carry8;
+    s8  -= carry8 * (1 << 21);
     carry10 = (s10 + (1 << 20)) >> 21;
     s11 += carry10;
     s10 -= carry10 * (1 << 21);
@@ -5115,20 +5105,20 @@ static void sc_muladd(uint8_t *s, const uint8_t *a, const uint8_t *b,
     s22 -= carry22 * (1 << 21);
 
     carry1 = (s1 + (1 << 20)) >> 21;
-    s2 += carry1;
-    s1 -= carry1 * (1 << 21);
+    s2  += carry1;
+    s1  -= carry1 * (1 << 21);
     carry3 = (s3 + (1 << 20)) >> 21;
-    s4 += carry3;
-    s3 -= carry