rbnacl-libsodium 1.0.11 → 1.0.13

data/vendor/libsodium/src/libsodium/crypto_box/curve25519xsalsa20poly1305/box_curve25519xsalsa20poly1305.c

@@ -0,0 +1,150 @@
+#include <string.h>
+
+#include "crypto_box_curve25519xsalsa20poly1305.h"
+#include "crypto_core_hsalsa20.h"
+#include "crypto_hash_sha512.h"
+#include "crypto_scalarmult_curve25519.h"
+#include "crypto_secretbox_xsalsa20poly1305.h"
+#include "randombytes.h"
+#include "utils.h"
+
+int
+crypto_box_curve25519xsalsa20poly1305_seed_keypair(unsigned char *pk,
+                                                   unsigned char *sk,
+                                                   const unsigned char *seed)
+{
+    unsigned char hash[64];
+
+    crypto_hash_sha512(hash, seed, 32);
+    memcpy(sk, hash, 32);
+    sodium_memzero(hash, sizeof hash);
+
+    return crypto_scalarmult_curve25519_base(pk, sk);
+}
+
+int
+crypto_box_curve25519xsalsa20poly1305_keypair(unsigned char *pk,
+                                              unsigned char *sk)
+{
+    randombytes_buf(sk, 32);
+
+    return crypto_scalarmult_curve25519_base(pk, sk);
+}
+
+int
+crypto_box_curve25519xsalsa20poly1305_beforenm(unsigned char *k,
+                                               const unsigned char *pk,
+                                               const unsigned char *sk)
+{
+    static const unsigned char zero[16] = { 0 };
+    unsigned char s[32];
+
+    if (crypto_scalarmult_curve25519(s, sk, pk) != 0) {
+        return -1;
+    }
+    return crypto_core_hsalsa20(k, zero, s, NULL);
+}
+
+int
+crypto_box_curve25519xsalsa20poly1305_afternm(unsigned char *c,
+                                              const unsigned char *m,
+                                              unsigned long long mlen,
+                                              const unsigned char *n,
+                                              const unsigned char *k)
+{
+    return crypto_secretbox_xsalsa20poly1305(c, m, mlen, n, k);
+}
+
+int
+crypto_box_curve25519xsalsa20poly1305_open_afternm(unsigned char *m,
+                                                   const unsigned char *c,
+                                                   unsigned long long clen,
+                                                   const unsigned char *n,
+                                                   const unsigned char *k)
+{
+    return crypto_secretbox_xsalsa20poly1305_open(m, c, clen, n, k);
+}
+
+int
+crypto_box_curve25519xsalsa20poly1305(unsigned char *c, const unsigned char *m,
+                                      unsigned long long   mlen,
+                                      const unsigned char *n,
+                                      const unsigned char *pk,
+                                      const unsigned char *sk)
+{
+    unsigned char k[crypto_box_curve25519xsalsa20poly1305_BEFORENMBYTES];
+    int           ret;
+
+    if (crypto_box_curve25519xsalsa20poly1305_beforenm(k, pk, sk) != 0) {
+        return -1;
+    }
+    ret = crypto_box_curve25519xsalsa20poly1305_afternm(c, m, mlen, n, k);
+    sodium_memzero(k, sizeof k);
+
+    return ret;
+}
+
+int
+crypto_box_curve25519xsalsa20poly1305_open(
+    unsigned char *m, const unsigned char *c, unsigned long long clen,
+    const unsigned char *n, const unsigned char *pk, const unsigned char *sk)
+{
+    unsigned char k[crypto_box_curve25519xsalsa20poly1305_BEFORENMBYTES];
+    int           ret;
+
+    if (crypto_box_curve25519xsalsa20poly1305_beforenm(k, pk, sk) != 0) {
+        return -1;
+    }
+    ret = crypto_box_curve25519xsalsa20poly1305_open_afternm(m, c, clen, n, k);
+    sodium_memzero(k, sizeof k);
+
+    return ret;
+}
+
+size_t
+crypto_box_curve25519xsalsa20poly1305_seedbytes(void)
+{
+    return crypto_box_curve25519xsalsa20poly1305_SEEDBYTES;
+}
+
+size_t
+crypto_box_curve25519xsalsa20poly1305_publickeybytes(void)
+{
+    return crypto_box_curve25519xsalsa20poly1305_PUBLICKEYBYTES;
+}
+
+size_t
+crypto_box_curve25519xsalsa20poly1305_secretkeybytes(void)
+{
+    return crypto_box_curve25519xsalsa20poly1305_SECRETKEYBYTES;
+}
+
+size_t
+crypto_box_curve25519xsalsa20poly1305_beforenmbytes(void)
+{
+    return crypto_box_curve25519xsalsa20poly1305_BEFORENMBYTES;
+}
+
+size_t
+crypto_box_curve25519xsalsa20poly1305_noncebytes(void)
+{
+    return crypto_box_curve25519xsalsa20poly1305_NONCEBYTES;
+}
+
+size_t
+crypto_box_curve25519xsalsa20poly1305_zerobytes(void)
+{
+    return crypto_box_curve25519xsalsa20poly1305_ZEROBYTES;
+}
+
+size_t
+crypto_box_curve25519xsalsa20poly1305_boxzerobytes(void)
+{
+    return crypto_box_curve25519xsalsa20poly1305_BOXZEROBYTES;
+}
+
+size_t
+crypto_box_curve25519xsalsa20poly1305_macbytes(void)
+{
+    return crypto_box_curve25519xsalsa20poly1305_MACBYTES;
+}

data/vendor/libsodium/src/libsodium/crypto_core/curve25519/ref10/curve25519_ref10.c

@@ -1,12 +1,15 @@
 #include <stddef.h>
 #include <stdint.h>
 #include <string.h>
+
 #include "crypto_verify_32.h"
 #include "private/curve25519_ref10.h"
 
-static uint64_t load_3(const unsigned char *in)
+static inline uint64_t
+load_3(const unsigned char *in)
 {
     uint64_t result;
+
     result = (uint64_t) in[0];
     result |= ((uint64_t) in[1]) << 8;
     result |= ((uint64_t) in[2]) << 16;
@@ -14,9 +17,11 @@ static uint64_t load_3(const unsigned char *in)
     return result;
 }
 
-static uint64_t load_4(const unsigned char *in)
+static inline uint64_t
+load_4(const unsigned char *in)
 {
     uint64_t result;
+
     result = (uint64_t) in[0];
     result |= ((uint64_t) in[1]) << 8;
     result |= ((uint64_t) in[2]) << 16;
@@ -29,7 +34,8 @@ static uint64_t load_4(const unsigned char *in)
  h = 0
  */
 
-void fe_0(fe h)
+void
+fe_0(fe h)
 {
     memset(&h[0], 0, 10 * sizeof h[0]);
 }
@@ -38,7 +44,8 @@ void fe_0(fe h)
  h = 1
  */
 
-void fe_1(fe h)
+void
+fe_1(fe h)
 {
     h[0] = 1;
     h[1] = 0;
@@ -57,7 +64,8 @@ void fe_1(fe h)
  |h| bounded by 1.1*2^26,1.1*2^25,1.1*2^26,1.1*2^25,etc.
  */
 
-void fe_add(fe h,const fe f,const fe g)
+void
+fe_add(fe h, const fe f, const fe g)
 {
     int32_t f0 = f[0];
     int32_t f1 = f[1];
@@ -69,6 +77,7 @@ void fe_add(fe h,const fe f,const fe g)
     int32_t f7 = f[7];
     int32_t f8 = f[8];
     int32_t f9 = f[9];
+
     int32_t g0 = g[0];
     int32_t g1 = g[1];
     int32_t g2 = g[2];
@@ -79,6 +88,7 @@ void fe_add(fe h,const fe f,const fe g)
     int32_t g7 = g[7];
     int32_t g8 = g[8];
     int32_t g9 = g[9];
+
     int32_t h0 = f0 + g0;
     int32_t h1 = f1 + g1;
     int32_t h2 = f2 + g2;
@@ -89,6 +99,7 @@ void fe_add(fe h,const fe f,const fe g)
     int32_t h7 = f7 + g7;
     int32_t h8 = f8 + g8;
     int32_t h9 = f9 + g9;
+
     h[0] = h0;
     h[1] = h1;
     h[2] = h2;
@@ -108,7 +119,8 @@ void fe_add(fe h,const fe f,const fe g)
  Preconditions: b in {0,1}.
  */
 
-void fe_cmov(fe f,const fe g,unsigned int b)
+void
+fe_cmov(fe f, const fe g, unsigned int b)
 {
     int32_t f0 = f[0];
     int32_t f1 = f[1];
@@ -120,6 +132,7 @@ void fe_cmov(fe f,const fe g,unsigned int b)
     int32_t f7 = f[7];
     int32_t f8 = f[8];
     int32_t f9 = f[9];
+
     int32_t g0 = g[0];
     int32_t g1 = g[1];
     int32_t g2 = g[2];
@@ -130,6 +143,7 @@ void fe_cmov(fe f,const fe g,unsigned int b)
     int32_t g7 = g[7];
     int32_t g8 = g[8];
     int32_t g9 = g[9];
+
     int32_t x0 = f0 ^ g0;
     int32_t x1 = f1 ^ g1;
     int32_t x2 = f2 ^ g2;
@@ -140,7 +154,8 @@ void fe_cmov(fe f,const fe g,unsigned int b)
     int32_t x7 = f7 ^ g7;
     int32_t x8 = f8 ^ g8;
     int32_t x9 = f9 ^ g9;
-    b = (unsigned int) (- (int) b);
+
+    b = (unsigned int) (-(int) b);
     x0 &= b;
     x1 &= b;
     x2 &= b;
@@ -167,7 +182,8 @@ void fe_cmov(fe f,const fe g,unsigned int b)
  h = f
  */
 
-void fe_copy(fe h,const fe f)
+void
+fe_copy(fe h, const fe f)
 {
     int32_t f0 = f[0];
     int32_t f1 = f[1];
@@ -179,6 +195,7 @@ void fe_copy(fe h,const fe f)
     int32_t f7 = f[7];
     int32_t f8 = f[8];
     int32_t f9 = f[9];
+
     h[0] = f0;
     h[1] = f1;
     h[2] = f2;
@@ -195,7 +212,8 @@ void fe_copy(fe h,const fe f)
  Ignores top bit of h.
  */
 
-void fe_frombytes(fe h,const unsigned char *s)
+void
+fe_frombytes(fe h, const unsigned char *s)
 {
     int64_t h0 = load_4(s);
     int64_t h1 = load_3(s + 4) << 6;
@@ -207,6 +225,7 @@ void fe_frombytes(fe h,const unsigned char *s)
     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 carry0;
     int64_t carry1;
     int64_t carry2;
@@ -218,17 +237,37 @@ void fe_frombytes(fe h,const unsigned char *s)
     int64_t carry8;
     int64_t carry9;
 
-    carry9 = (h9 + (int64_t) (1L << 24)) >> 25; h0 += carry9 * 19; h9 -= carry9 * ((uint64_t) 1L << 25);
-    carry1 = (h1 + (int64_t) (1L << 24)) >> 25; h2 += carry1; h1 -= carry1 * ((uint64_t) 1L << 25);
-    carry3 = (h3 + (int64_t) (1L << 24)) >> 25; h4 += carry3; h3 -= carry3 * ((uint64_t) 1L << 25);
-    carry5 = (h5 + (int64_t) (1L << 24)) >> 25; h6 += carry5; h5 -= carry5 * ((uint64_t) 1L << 25);
-    carry7 = (h7 + (int64_t) (1L << 24)) >> 25; h8 += carry7; h7 -= carry7 * ((uint64_t) 1L << 25);
-
-    carry0 = (h0 + (int64_t) (1L << 25)) >> 26; h1 += carry0; h0 -= carry0 * ((uint64_t) 1L << 26);
-    carry2 = (h2 + (int64_t) (1L << 25)) >> 26; h3 += carry2; h2 -= carry2 * ((uint64_t) 1L << 26);
-    carry4 = (h4 + (int64_t) (1L << 25)) >> 26; h5 += carry4; h4 -= carry4 * ((uint64_t) 1L << 26);
-    carry6 = (h6 + (int64_t) (1L << 25)) >> 26; h7 += carry6; h6 -= carry6 * ((uint64_t) 1L << 26);
-    carry8 = (h8 + (int64_t) (1L << 25)) >> 26; h9 += carry8; h8 -= carry8 * ((uint64_t) 1L << 26);
+    carry9 = (h9 + (int64_t)(1L << 24)) >> 25;
+    h0 += carry9 * 19;
+    h9 -= carry9 * ((uint64_t) 1L << 25);
+    carry1 = (h1 + (int64_t)(1L << 24)) >> 25;
+    h2 += carry1;
+    h1 -= carry1 * ((uint64_t) 1L << 25);
+    carry3 = (h3 + (int64_t)(1L << 24)) >> 25;
+    h4 += carry3;
+    h3 -= carry3 * ((uint64_t) 1L << 25);
+    carry5 = (h5 + (int64_t)(1L << 24)) >> 25;
+    h6 += carry5;
+    h5 -= carry5 * ((uint64_t) 1L << 25);
+    carry7 = (h7 + (int64_t)(1L << 24)) >> 25;
+    h8 += carry7;
+    h7 -= carry7 * ((uint64_t) 1L << 25);
+
+    carry0 = (h0 + (int64_t)(1L << 25)) >> 26;
+    h1 += carry0;
+    h0 -= carry0 * ((uint64_t) 1L << 26);
+    carry2 = (h2 + (int64_t)(1L << 25)) >> 26;
+    h3 += carry2;
+    h2 -= carry2 * ((uint64_t) 1L << 26);
+    carry4 = (h4 + (int64_t)(1L << 25)) >> 26;
+    h5 += carry4;
+    h4 -= carry4 * ((uint64_t) 1L << 26);
+    carry6 = (h6 + (int64_t)(1L << 25)) >> 26;
+    h7 += carry6;
+    h6 -= carry6 * ((uint64_t) 1L << 26);
+    carry8 = (h8 + (int64_t)(1L << 25)) >> 26;
+    h9 += carry8;
+    h8 -= carry8 * ((uint64_t) 1L << 26);
 
     h[0] = (int32_t) h0;
     h[1] = (int32_t) h1;
@@ -267,7 +306,8 @@ void fe_frombytes(fe h,const unsigned char *s)
  so floor(2^(-255)(h + 19 2^(-25) h9 + 2^(-1))) = q.
  */
 
-void fe_tobytes(unsigned char *s,const fe h)
+void
+fe_tobytes(unsigned char *s, const fe h)
 {
     int32_t h0 = h[0];
     int32_t h1 = h[1];
@@ -279,6 +319,7 @@ void fe_tobytes(unsigned char *s,const fe h)
     int32_t h7 = h[7];
     int32_t h8 = h[8];
     int32_t h9 = h[9];
+
     int32_t q;
     int32_t carry0;
     int32_t carry1;
@@ -307,16 +348,35 @@ void fe_tobytes(unsigned char *s,const fe h)
     h0 += 19 * q;
     /* Goal: Output h-2^255 q, which is between 0 and 2^255-20. */
 
-    carry0 = h0 >> 26; h1 += carry0; h0 -= carry0 * ((uint32_t) 1L << 26);
-    carry1 = h1 >> 25; h2 += carry1; h1 -= carry1 * ((uint32_t) 1L << 25);
-    carry2 = h2 >> 26; h3 += carry2; h2 -= carry2 * ((uint32_t) 1L << 26);
-    carry3 = h3 >> 25; h4 += carry3; h3 -= carry3 * ((uint32_t) 1L << 25);
-    carry4 = h4 >> 26; h5 += carry4; h4 -= carry4 * ((uint32_t) 1L << 26);
-    carry5 = h5 >> 25; h6 += carry5; h5 -= carry5 * ((uint32_t) 1L << 25);
-    carry6 = h6 >> 26; h7 += carry6; h6 -= carry6 * ((uint32_t) 1L << 26);
-    carry7 = h7 >> 25; h8 += carry7; h7 -= carry7 * ((uint32_t) 1L << 25);
-    carry8 = h8 >> 26; h9 += carry8; h8 -= carry8 * ((uint32_t) 1L << 26);
-    carry9 = h9 >> 25;               h9 -= carry9 * ((uint32_t) 1L << 25);
+    carry0 = h0 >> 26;
+    h1 += carry0;
+    h0 -= carry0 * ((uint32_t) 1L << 26);
+    carry1 = h1 >> 25;
+    h2 += carry1;
+    h1 -= carry1 * ((uint32_t) 1L << 25);
+    carry2 = h2 >> 26;
+    h3 += carry2;
+    h2 -= carry2 * ((uint32_t) 1L << 26);
+    carry3 = h3 >> 25;
+    h4 += carry3;
+    h3 -= carry3 * ((uint32_t) 1L << 25);
+    carry4 = h4 >> 26;
+    h5 += carry4;
+    h4 -= carry4 * ((uint32_t) 1L << 26);
+    carry5 = h5 >> 25;
+    h6 += carry5;
+    h5 -= carry5 * ((uint32_t) 1L << 25);
+    carry6 = h6 >> 26;
+    h7 += carry6;
+    h6 -= carry6 * ((uint32_t) 1L << 26);
+    carry7 = h7 >> 25;
+    h8 += carry7;
+    h7 -= carry7 * ((uint32_t) 1L << 25);
+    carry8 = h8 >> 26;
+    h9 += carry8;
+    h8 -= carry8 * ((uint32_t) 1L << 26);
+    carry9 = h9 >> 25;
+    h9 -= carry9 * ((uint32_t) 1L << 25);
     /* h10 = carry9 */
 
     /*
@@ -326,16 +386,16 @@ void fe_tobytes(unsigned char *s,const fe h)
      Goal: Output h0+...+2^230 h9.
      */
 
-    s[0] = h0 >> 0;
-    s[1] = h0 >> 8;
-    s[2] = h0 >> 16;
-    s[3] = (h0 >> 24) | (h1 * ((uint32_t) 1 << 2));
-    s[4] = h1 >> 6;
-    s[5] = h1 >> 14;
-    s[6] = (h1 >> 22) | (h2 * ((uint32_t) 1 << 3));
-    s[7] = h2 >> 5;
-    s[8] = h2 >> 13;
-    s[9] = (h2 >> 21) | (h3 * ((uint32_t) 1 << 5));
+    s[0]  = h0 >> 0;
+    s[1]  = h0 >> 8;
+    s[2]  = h0 >> 16;
+    s[3]  = (h0 >> 24) | (h1 * ((uint32_t) 1 << 2));
+    s[4]  = h1 >> 6;
+    s[5]  = h1 >> 14;
+    s[6]  = (h1 >> 22) | (h2 * ((uint32_t) 1 << 3));
+    s[7]  = h2 >> 5;
+    s[8]  = h2 >> 13;
+    s[9]  = (h2 >> 21) | (h3 * ((uint32_t) 1 << 5));
     s[10] = h3 >> 3;
     s[11] = h3 >> 11;
     s[12] = (h3 >> 19) | (h4 * ((uint32_t) 1 << 6));
@@ -368,10 +428,12 @@ void fe_tobytes(unsigned char *s,const fe h)
  |f| bounded by 1.1*2^26,1.1*2^25,1.1*2^26,1.1*2^25,etc.
  */
 
-int fe_isnegative(const fe f)
+int
+fe_isnegative(const fe f)
 {
     unsigned char s[32];
-    fe_tobytes(s,f);
+
+    fe_tobytes(s, f);
 
     return s[0] & 1;
 }
@@ -386,12 +448,14 @@ int fe_isnegative(const fe f)
 
 static unsigned char zero[32];
 
-int fe_isnonzero(const fe f)
+int
+fe_isnonzero(const fe f)
 {
     unsigned char s[32];
-    fe_tobytes(s,f);
 
-    return crypto_verify_32(s,zero);
+    fe_tobytes(s, f);
+
+    return crypto_verify_32(s, zero);
 }
 
 /*
@@ -426,7 +490,8 @@ int fe_isnonzero(const fe f)
  With tighter constraints on inputs can squeeze carries into int32.
  */
 
-void fe_mul(fe h,const fe f,const fe g)
+void
+fe_mul(fe h, const fe f, const fe g)
 {
     int32_t f0 = f[0];
     int32_t f1 = f[1];
@@ -438,6 +503,7 @@ void fe_mul(fe h,const fe f,const fe g)
     int32_t f7 = f[7];
     int32_t f8 = f[8];
     int32_t f9 = f[9];
+
     int32_t g0 = g[0];
     int32_t g1 = g[1];
     int32_t g2 = g[2];
@@ -448,6 +514,7 @@ void fe_mul(fe h,const fe f,const fe g)
     int32_t g7 = g[7];
     int32_t g8 = g[8];
     int32_t g9 = g[9];
+
     int32_t g1_19 = 19 * g1; /* 1.959375*2^29 */
     int32_t g2_19 = 19 * g2; /* 1.959375*2^30; still ok */
     int32_t g3_19 = 19 * g3;
@@ -457,121 +524,134 @@ void fe_mul(fe h,const fe f,const fe g)
     int32_t g7_19 = 19 * g7;
     int32_t g8_19 = 19 * g8;
     int32_t g9_19 = 19 * g9;
-    int32_t f1_2 = 2 * f1;
-    int32_t f3_2 = 2 * f3;
-    int32_t f5_2 = 2 * f5;
-    int32_t f7_2 = 2 * f7;
-    int32_t f9_2 = 2 * f9;
-    int64_t f0g0    = f0   * (int64_t) g0;
-    int64_t f0g1    = f0   * (int64_t) g1;
-    int64_t f0g2    = f0   * (int64_t) g2;
-    int64_t f0g3    = f0   * (int64_t) g3;
-    int64_t f0g4    = f0   * (int64_t) g4;
-    int64_t f0g5    = f0   * (int64_t) g5;
-    int64_t f0g6    = f0   * (int64_t) g6;
-    int64_t f0g7    = f0   * (int64_t) g7;
-    int64_t f0g8    = f0   * (int64_t) g8;
-    int64_t f0g9    = f0   * (int64_t) g9;
-    int64_t f1g0    = f1   * (int64_t) g0;
+    int32_t f1_2  = 2 * f1;
+    int32_t f3_2  = 2 * f3;
+    int32_t f5_2  = 2 * f5;
+    int32_t f7_2  = 2 * f7;
+    int32_t f9_2  = 2 * f9;
+
+    int64_t f0g0    = f0 * (int64_t) g0;
+    int64_t f0g1    = f0 * (int64_t) g1;
+    int64_t f0g2    = f0 * (int64_t) g2;
+    int64_t f0g3    = f0 * (int64_t) g3;
+    int64_t f0g4    = f0 * (int64_t) g4;
+    int64_t f0g5    = f0 * (int64_t) g5;
+    int64_t f0g6    = f0 * (int64_t) g6;
+    int64_t f0g7    = f0 * (int64_t) g7;
+    int64_t f0g8    = f0 * (int64_t) g8;
+    int64_t f0g9    = f0 * (int64_t) g9;
+    int64_t f1g0    = f1 * (int64_t) g0;
     int64_t f1g1_2  = f1_2 * (int64_t) g1;
-    int64_t f1g2    = f1   * (int64_t) g2;
+    int64_t f1g2    = f1 * (int64_t) g2;
     int64_t f1g3_2  = f1_2 * (int64_t) g3;
-    int64_t f1g4    = f1   * (int64_t) g4;
+    int64_t f1g4    = f1 * (int64_t) g4;
     int64_t f1g5_2  = f1_2 * (int64_t) g5;
-    int64_t f1g6    = f1   * (int64_t) g6;
+    int64_t f1g6    = f1 * (int64_t) g6;
     int64_t f1g7_2  = f1_2 * (int64_t) g7;
-    int64_t f1g8    = f1   * (int64_t) g8;
+    int64_t f1g8    = f1 * (int64_t) g8;
     int64_t f1g9_38 = f1_2 * (int64_t) g9_19;
-    int64_t f2g0    = f2   * (int64_t) g0;
-    int64_t f2g1    = f2   * (int64_t) g1;
-    int64_t f2g2    = f2   * (int64_t) g2;
-    int64_t f2g3    = f2   * (int64_t) g3;
-    int64_t f2g4    = f2   * (int64_t) g4;
-    int64_t f2g5    = f2   * (int64_t) g5;
-    int64_t f2g6    = f2   * (int64_t) g6;
-    int64_t f2g7    = f2   * (int64_t) g7;
-    int64_t f2g8_19 = f2   * (int64_t) g8_19;
-    int64_t f2g9_19 = f2   * (int64_t) g9_19;
-    int64_t f3g0    = f3   * (int64_t) g0;
+    int64_t f2g0    = f2 * (int64_t) g0;
+    int64_t f2g1    = f2 * (int64_t) g1;
+    int64_t f2g2    = f2 * (int64_t) g2;
+    int64_t f2g3    = f2 * (int64_t) g3;
+    int64_t f2g4    = f2 * (int64_t) g4;
+    int64_t f2g5    = f2 * (int64_t) g5;
+    int64_t f2g6    = f2 * (int64_t) g6;
+    int64_t f2g7    = f2 * (int64_t) g7;
+    int64_t f2g8_19 = f2 * (int64_t) g8_19;
+    int64_t f2g9_19 = f2 * (int64_t) g9_19;
+    int64_t f3g0    = f3 * (int64_t) g0;
     int64_t f3g1_2  = f3_2 * (int64_t) g1;
-    int64_t f3g2    = f3   * (int64_t) g2;
+    int64_t f3g2    = f3 * (int64_t) g2;
     int64_t f3g3_2  = f3_2 * (int64_t) g3;
-    int64_t f3g4    = f3   * (int64_t) g4;
+    int64_t f3g4    = f3 * (int64_t) g4;
     int64_t f3g5_2  = f3_2 * (int64_t) g5;
-    int64_t f3g6    = f3   * (int64_t) g6;
+    int64_t f3g6    = f3 * (int64_t) g6;
     int64_t f3g7_38 = f3_2 * (int64_t) g7_19;
-    int64_t f3g8_19 = f3   * (int64_t) g8_19;
+    int64_t f3g8_19 = f3 * (int64_t) g8_19;
     int64_t f3g9_38 = f3_2 * (int64_t) g9_19;
-    int64_t f4g0    = f4   * (int64_t) g0;
-    int64_t f4g1    = f4   * (int64_t) g1;
-    int64_t f4g2    = f4   * (int64_t) g2;
-    int64_t f4g3    = f4   * (int64_t) g3;
-    int64_t f4g4    = f4   * (int64_t) g4;
-    int64_t f4g5    = f4   * (int64_t) g5;
-    int64_t f4g6_19 = f4   * (int64_t) g6_19;
-    int64_t f4g7_19 = f4   * (int64_t) g7_19;
-    int64_t f4g8_19 = f4   * (int64_t) g8_19;
-    int64_t f4g9_19 = f4   * (int64_t) g9_19;
-    int64_t f5g0    = f5   * (int64_t) g0;
+    int64_t f4g0    = f4 * (int64_t) g0;
+    int64_t f4g1    = f4 * (int64_t) g1;
+    int64_t f4g2    = f4 * (int64_t) g2;
+    int64_t f4g3    = f4 * (int64_t) g3;
+    int64_t f4g4    = f4 * (int64_t) g4;
+    int64_t f4g5    = f4 * (int64_t) g5;
+    int64_t f4g6_19 = f4 * (int64_t) g6_19;
+    int64_t f4g7_19 = f4 * (int64_t) g7_19;
+    int64_t f4g8_19 = f4 * (int64_t) g8_19;
+    int64_t f4g9_19 = f4 * (int64_t) g9_19;
+    int64_t f5g0    = f5 * (int64_t) g0;
     int64_t f5g1_2  = f5_2 * (int64_t) g1;
-    int64_t f5g2    = f5   * (int64_t) g2;
+    int64_t f5g2    = f5 * (int64_t) g2;
     int64_t f5g3_2  = f5_2 * (int64_t) g3;
-    int64_t f5g4    = f5   * (int64_t) g4;
+    int64_t f5g4    = f5 * (int64_t) g4;
     int64_t f5g5_38 = f5_2 * (int64_t) g5_19;
-    int64_t f5g6_19 = f5   * (int64_t) g6_19;
+    int64_t f5g6_19 = f5 * (int64_t) g6_19;
     int64_t f5g7_38 = f5_2 * (int64_t) g7_19;
-    int64_t f5g8_19 = f5   * (int64_t) g8_19;
+    int64_t f5g8_19 = f5 * (int64_t) g8_19;
     int64_t f5g9_38 = f5_2 * (int64_t) g9_19;
-    int64_t f6g0    = f6   * (int64_t) g0;
-    int64_t f6g1    = f6   * (int64_t) g1;
-    int64_t f6g2    = f6   * (int64_t) g2;
-    int64_t f6g3    = f6   * (int64_t) g3;
-    int64_t f6g4_19 = f6   * (int64_t) g4_19;
-    int64_t f6g5_19 = f6   * (int64_t) g5_19;
-    int64_t f6g6_19 = f6   * (int64_t) g6_19;
-    int64_t f6g7_19 = f6   * (int64_t) g7_19;
-    int64_t f6g8_19 = f6   * (int64_t) g8_19;
-    int64_t f6g9_19 = f6   * (int64_t) g9_19;
-    int64_t f7g0    = f7   * (int64_t) g0;
+    int64_t f6g0    = f6 * (int64_t) g0;
+    int64_t f6g1    = f6 * (int64_t) g1;
+    int64_t f6g2    = f6 * (int64_t) g2;
+    int64_t f6g3    = f6 * (int64_t) g3;
+    int64_t f6g4_19 = f6 * (int64_t) g4_19;
+    int64_t f6g5_19 = f6 * (int64_t) g5_19;
+    int64_t f6g6_19 = f6 * (int64_t) g6_19;
+    int64_t f6g7_19 = f6 * (int64_t) g7_19;
+    int64_t f6g8_19 = f6 * (int64_t) g8_19;
+    int64_t f6g9_19 = f6 * (int64_t) g9_19;
+    int64_t f7g0    = f7 * (int64_t) g0;
     int64_t f7g1_2  = f7_2 * (int64_t) g1;
-    int64_t f7g2    = f7   * (int64_t) g2;
+    int64_t f7g2    = f7 * (int64_t) g2;
     int64_t f7g3_38 = f7_2 * (int64_t) g3_19;
-    int64_t f7g4_19 = f7   * (int64_t) g4_19;
+    int64_t f7g4_19 = f7 * (int64_t) g4_19;
     int64_t f7g5_38 = f7_2 * (int64_t) g5_19;
-    int64_t f7g6_19 = f7   * (int64_t) g6_19;
+    int64_t f7g6_19 = f7 * (int64_t) g6_19;
     int64_t f7g7_38 = f7_2 * (int64_t) g7_19;
-    int64_t f7g8_19 = f7   * (int64_t) g8_19;
+    int64_t f7g8_19 = f7 * (int64_t) g8_19;
     int64_t f7g9_38 = f7_2 * (int64_t) g9_19;
-    int64_t f8g0    = f8   * (int64_t) g0;
-    int64_t f8g1    = f8   * (int64_t) g1;
-    int64_t f8g2_19 = f8   * (int64_t) g2_19;
-    int64_t f8g3_19 = f8   * (int64_t) g3_19;
-    int64_t f8g4_19 = f8   * (int64_t) g4_19;
-    int64_t f8g5_19 = f8   * (int64_t) g5_19;
-    int64_t f8g6_19 = f8   * (int64_t) g6_19;
-    int64_t f8g7_19 = f8   * (int64_t) g7_19;
-    int64_t f8g8_19 = f8   * (int64_t) g8_19;
-    int64_t f8g9_19 = f8   * (int64_t) g9_19;
-    int64_t f9g0    = f9   * (int64_t) g0;
+    int64_t f8g0    = f8 * (int64_t) g0;
+    int64_t f8g1    = f8 * (int64_t) g1;
+    int64_t f8g2_19 = f8 * (int64_t) g2_19;
+    int64_t f8g3_19 = f8 * (int64_t) g3_19;
+    int64_t f8g4_19 = f8 * (int64_t) g4_19;
+    int64_t f8g5_19 = f8 * (int64_t) g5_19;
+    int64_t f8g6_19 = f8 * (int64_t) g6_19;
+    int64_t f8g7_19 = f8 * (int64_t) g7_19;
+    int64_t f8g8_19 = f8 * (int64_t) g8_19;
+    int64_t f8g9_19 = f8 * (int64_t) g9_19;
+    int64_t f9g0    = f9 * (int64_t) g0;
     int64_t f9g1_38 = f9_2 * (int64_t) g1_19;
-    int64_t f9g2_19 = f9   * (int64_t) g2_19;
+    int64_t f9g2_19 = f9 * (int64_t) g2_19;
     int64_t f9g3_38 = f9_2 * (int64_t) g3_19;
-    int64_t f9g4_19 = f9   * (int64_t) g4_19;
+    int64_t f9g4_19 = f9 * (int64_t) g4_19;
     int64_t f9g5_38 = f9_2 * (int64_t) g5_19;
-    int64_t f9g6_19 = f9   * (int64_t) g6_19;
+    int64_t f9g6_19 = f9 * (int64_t) g6_19;
     int64_t f9g7_38 = f9_2 * (int64_t) g7_19;
-    int64_t f9g8_19 = f9   * (int64_t) g8_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;
@@ -590,46 +670,70 @@ void fe_mul(fe h,const fe f,const fe g)
      i.e. |h1| <= 1.7*2^59; narrower ranges for h3, h5, h7, h9
      */
 
-    carry0 = (h0 + (int64_t) (1L << 25)) >> 26; h1 += carry0; h0 -= carry0 * ((uint64_t) 1L << 26);
-    carry4 = (h4 + (int64_t) (1L << 25)) >> 26; h5 += carry4; h4 -= carry4 * ((uint64_t) 1L << 26);
+    carry0 = (h0 + (int64_t)(1L << 25)) >> 26;
+    h1 += carry0;
+    h0 -= carry0 * ((uint64_t) 1L << 26);
+    carry4 = (h4 + (int64_t)(1L << 25)) >> 26;
+    h5 += carry4;
+    h4 -= carry4 * ((uint64_t) 1L << 26);
     /* |h0| <= 2^25 */
     /* |h4| <= 2^25 */
     /* |h1| <= 1.71*2^59 */
     /* |h5| <= 1.71*2^59 */
 
-    carry1 = (h1 + (int64_t) (1L << 24)) >> 25; h2 += carry1; h1 -= carry1 * ((uint64_t) 1L << 25);
-    carry5 = (h5 + (int64_t) (1L << 24)) >> 25; h6 += carry5; h5 -= carry5 * ((uint64_t) 1L << 25);
+    carry1 = (h1 + (int64_t)(1L << 24)) >> 25;
+    h2 += carry1;
+    h1 -= carry1 * ((uint64_t) 1L << 25);
+    carry5 = (h5 + (int64_t)(1L << 24)) >> 25;
+    h6 += carry5;
+    h5 -= carry5 * ((uint64_t) 1L << 25);
     /* |h1| <= 2^24; from now on fits into int32 */
     /* |h5| <= 2^24; from now on fits into int32 */
     /* |h2| <= 1.41*2^60 */
     /* |h6| <= 1.41*2^60 */
 
-    carry2 = (h2 + (int64_t) (1L << 25)) >> 26; h3 += carry2; h2 -= carry2 * ((uint64_t) 1L << 26);
-    carry6 = (h6 + (int64_t) (1L << 25)) >> 26; h7 += carry6; h6 -= carry6 * ((uint64_t) 1L << 26);
+    carry2 = (h2 + (int64_t)(1L << 25)) >> 26;
+    h3 += carry2;
+    h2 -= carry2 * ((uint64_t) 1L << 26);
+    carry6 = (h6 + (int64_t)(1L << 25)) >> 26;
+    h7 += carry6;
+    h6 -= carry6 * ((uint64_t) 1L << 26);
     /* |h2| <= 2^25; from now on fits into int32 unchanged */
     /* |h6| <= 2^25; from now on fits into int32 unchanged */
     /* |h3| <= 1.71*2^59 */
     /* |h7| <= 1.71*2^59 */
 
-    carry3 = (h3 + (int64_t) (1L << 24)) >> 25; h4 += carry3; h3 -= carry3 * ((uint64_t) 1L << 25);
-    carry7 = (h7 + (int64_t) (1L << 24)) >> 25; h8 += carry7; h7 -= carry7 * ((uint64_t) 1L << 25);
+    carry3 = (h3 + (int64_t)(1L << 24)) >> 25;
+    h4 += carry3;
+    h3 -= carry3 * ((uint64_t) 1L << 25);
+    carry7 = (h7 + (int64_t)(1L << 24)) >> 25;
+    h8 += carry7;
+    h7 -= carry7 * ((uint64_t) 1L << 25);
     /* |h3| <= 2^24; from now on fits into int32 unchanged */
     /* |h7| <= 2^24; from now on fits into int32 unchanged */
     /* |h4| <= 1.72*2^34 */
     /* |h8| <= 1.41*2^60 */
 
-    carry4 = (h4 + (int64_t) (1L << 25)) >> 26; h5 += carry4; h4 -= carry4 * ((uint64_t) 1L << 26);
-    carry8 = (h8 + (int64_t) (1L << 25)) >> 26; h9 += carry8; h8 -= carry8 * ((uint64_t) 1L << 26);
+    carry4 = (h4 + (int64_t)(1L << 25)) >> 26;
+    h5 += carry4;
+    h4 -= carry4 * ((uint64_t) 1L << 26);
+    carry8 = (h8 + (int64_t)(1L << 25)) >> 26;
+    h9 += carry8;
+    h8 -= carry8 * ((uint64_t) 1L << 26);
     /* |h4| <= 2^25; from now on fits into int32 unchanged */
     /* |h8| <= 2^25; from now on fits into int32 unchanged */
     /* |h5| <= 1.01*2^24 */
     /* |h9| <= 1.71*2^59 */
 
-    carry9 = (h9 + (int64_t) (1L << 24)) >> 25; h0 += carry9 * 19; h9 -= carry9 * ((uint64_t) 1L << 25);
+    carry9 = (h9 + (int64_t)(1L << 24)) >> 25;
+    h0 += carry9 * 19;
+    h9 -= carry9 * ((uint64_t) 1L << 25);
     /* |h9| <= 2^24; from now on fits into int32 unchanged */
     /* |h0| <= 1.1*2^39 */
 
-    carry0 = (h0 + (int64_t) (1L << 25)) >> 26; h1 += carry0; h0 -= carry0 * ((uint64_t) 1L << 26);
+    carry0 = (h0 + (int64_t)(1L << 25)) >> 26;
+    h1 += carry0;
+    h0 -= carry0 * ((uint64_t) 1L << 26);
     /* |h0| <= 2^25; from now on fits into int32 unchanged */
     /* |h1| <= 1.01*2^24 */
 
@@ -655,7 +759,8 @@ void fe_mul(fe h,const fe f,const fe g)
  |h| bounded by 1.1*2^25,1.1*2^24,1.1*2^25,1.1*2^24,etc.
  */
 
-void fe_neg(fe h,const fe f)
+void
+fe_neg(fe h, const fe f)
 {
     int32_t f0 = f[0];
     int32_t f1 = f[1];
@@ -667,6 +772,7 @@ void fe_neg(fe h,const fe f)
     int32_t f7 = f[7];
     int32_t f8 = f[8];
     int32_t f9 = f[9];
+
     int32_t h0 = -f0;
     int32_t h1 = -f1;
     int32_t h2 = -f2;
@@ -677,6 +783,7 @@ void fe_neg(fe h,const fe f)
     int32_t h7 = -f7;
     int32_t h8 = -f8;
     int32_t h9 = -f9;
+
     h[0] = h0;
     h[1] = h1;
     h[2] = h2;
@@ -704,7 +811,8 @@ void fe_neg(fe h,const fe f)
  See fe_mul.c for discussion of implementation strategy.
  */
 
-void fe_sq(fe h,const fe f)
+void
+fe_sq(fe h, const fe f)
 {
     int32_t f0 = f[0];
     int32_t f1 = f[1];
@@ -716,20 +824,22 @@ void fe_sq(fe h,const fe f)
     int32_t f7 = f[7];
     int32_t f8 = f[8];
     int32_t f9 = f[9];
-    int32_t f0_2 = 2 * f0;
-    int32_t f1_2 = 2 * f1;
-    int32_t f2_2 = 2 * f2;
-    int32_t f3_2 = 2 * f3;
-    int32_t f4_2 = 2 * f4;
-    int32_t f5_2 = 2 * f5;
-    int32_t f6_2 = 2 * f6;
-    int32_t f7_2 = 2 * f7;
+
+    int32_t f0_2  = 2 * f0;
+    int32_t f1_2  = 2 * f1;
+    int32_t f2_2  = 2 * f2;
+    int32_t f3_2  = 2 * f3;
+    int32_t f4_2  = 2 * f4;
+    int32_t f5_2  = 2 * f5;
+    int32_t f6_2  = 2 * f6;
+    int32_t f7_2  = 2 * f7;
     int32_t f5_38 = 38 * f5; /* 1.959375*2^30 */
     int32_t f6_19 = 19 * f6; /* 1.959375*2^30 */
     int32_t f7_38 = 38 * f7; /* 1.959375*2^30 */
     int32_t f8_19 = 19 * f8; /* 1.959375*2^30 */
     int32_t f9_38 = 38 * f9; /* 1.959375*2^30 */
-    int64_t f0f0    = f0   * (int64_t) f0;
+
+    int64_t f0f0    = f0 * (int64_t) f0;
     int64_t f0f1_2  = f0_2 * (int64_t) f1;
     int64_t f0f2_2  = f0_2 * (int64_t) f2;
     int64_t f0f3_2  = f0_2 * (int64_t) f3;
@@ -748,14 +858,14 @@ void fe_sq(fe h,const fe f)
     int64_t f1f7_4  = f1_2 * (int64_t) f7_2;
     int64_t f1f8_2  = f1_2 * (int64_t) f8;
     int64_t f1f9_76 = f1_2 * (int64_t) f9_38;
-    int64_t f2f2    = f2   * (int64_t) f2;
+    int64_t f2f2    = f2 * (int64_t) f2;
     int64_t f2f3_2  = f2_2 * (int64_t) f3;
     int64_t f2f4_2  = f2_2 * (int64_t) f4;
     int64_t f2f5_2  = f2_2 * (int64_t) f5;
     int64_t f2f6_2  = f2_2 * (int64_t) f6;
     int64_t f2f7_2  = f2_2 * (int64_t) f7;
     int64_t f2f8_38 = f2_2 * (int64_t) f8_19;
-    int64_t f2f9_38 = f2   * (int64_t) f9_38;
+    int64_t f2f9_38 = f2 * (int64_t) f9_38;
     int64_t f3f3_2  = f3_2 * (int64_t) f3;
     int64_t f3f4_2  = f3_2 * (int64_t) f4;
     int64_t f3f5_4  = f3_2 * (int64_t) f5_2;
@@ -763,37 +873,39 @@ void fe_sq(fe h,const fe f)
     int64_t f3f7_76 = f3_2 * (int64_t) f7_38;
     int64_t f3f8_38 = f3_2 * (int64_t) f8_19;
     int64_t f3f9_76 = f3_2 * (int64_t) f9_38;
-    int64_t f4f4    = f4   * (int64_t) f4;
+    int64_t f4f4    = f4 * (int64_t) f4;
     int64_t f4f5_2  = f4_2 * (int64_t) f5;
     int64_t f4f6_38 = f4_2 * (int64_t) f6_19;
-    int64_t f4f7_38 = f4   * (int64_t) f7_38;
+    int64_t f4f7_38 = f4 * (int64_t) f7_38;
     int64_t f4f8_38 = f4_2 * (int64_t) f8_19;
-    int64_t f4f9_38 = f4   * (int64_t) f9_38;
-    int64_t f5f5_38 = f5   * (int64_t) f5_38;
+    int64_t f4f9_38 = f4 * (int64_t) f9_38;
+    int64_t f5f5_38 = f5 * (int64_t) f5_38;
     int64_t f5f6_38 = f5_2 * (int64_t) f6_19;
     int64_t f5f7_76 = f5_2 * (int64_t) f7_38;
     int64_t f5f8_38 = f5_2 * (int64_t) f8_19;
     int64_t f5f9_76 = f5_2 * (int64_t) f9_38;
-    int64_t f6f6_19 = f6   * (int64_t) f6_19;
-    int64_t f6f7_38 = f6   * (int64_t) f7_38;
+    int64_t f6f6_19 = f6 * (int64_t) f6_19;
+    int64_t f6f7_38 = f6 * (int64_t) f7_38;
     int64_t f6f8_38 = f6_2 * (int64_t) f8_19;
-    int64_t f6f9_38 = f6   * (int64_t) f9_38;
-    int64_t f7f7_38 = f7   * (int64_t) f7_38;
+    int64_t f6f9_38 = f6 * (int64_t) f9_38;
+    int64_t f7f7_38 = f7 * (int64_t) f7_38;
     int64_t f7f8_38 = f7_2 * (int64_t) f8_19;
     int64_t f7f9_76 = f7_2 * (int64_t) f9_38;
-    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 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 carry0;
     int64_t carry1;
     int64_t carry2;
@@ -805,24 +917,48 @@ void fe_sq(fe h,const fe f)
     int64_t carry8;
     int64_t carry9;
 
-    carry0 = (h0 + (int64_t) (1L << 25)) >> 26; h1 += carry0; h0 -= carry0 * ((uint64_t) 1L << 26);
-    carry4 = (h4 + (int64_t) (1L << 25)) >> 26; h5 += carry4; h4 -= carry4 * ((uint64_t) 1L << 26);
-
-    carry1 = (h1 + (int64_t) (1L << 24)) >> 25; h2 += carry1; h1 -= carry1 * ((uint64_t) 1L << 25);
-    carry5 = (h5 + (int64_t) (1L << 24)) >> 25; h6 += carry5; h5 -= carry5 * ((uint64_t) 1L << 25);
-
-    carry2 = (h2 + (int64_t) (1L << 25)) >> 26; h3 += carry2; h2 -= carry2 * ((uint64_t) 1L << 26);
-    carry6 = (h6 + (int64_t) (1L << 25)) >> 26; h7 += carry6; h6 -= carry6 * ((uint64_t) 1L << 26);
-
-    carry3 = (h3 + (int64_t) (1L << 24)) >> 25; h4 += carry3; h3 -= carry3 * ((uint64_t) 1L << 25);
-    carry7 = (h7 + (int64_t) (1L << 24)) >> 25; h8 += carry7; h7 -= carry7 * ((uint64_t) 1L << 25);
-
-    carry4 = (h4 + (int64_t) (1L << 25)) >> 26; h5 += carry4; h4 -= carry4 * ((uint64_t) 1L << 26);
-    carry8 = (h8 + (int64_t) (1L << 25)) >> 26; h9 += carry8; h8 -= carry8 * ((uint64_t) 1L << 26);
-
-    carry9 = (h9 + (int64_t) (1L << 24)) >> 25; h0 += carry9 * 19; h9 -= carry9 * ((uint64_t) 1L << 25);
-
-    carry0 = (h0 + (int64_t) (1L << 25)) >> 26; h1 += carry0; h0 -= carry0 * ((uint64_t) 1L << 26);
+    carry0 = (h0 + (int64_t)(1L << 25)) >> 26;
+    h1 += carry0;
+    h0 -= carry0 * ((uint64_t) 1L << 26);
+    carry4 = (h4 + (int64_t)(1L << 25)) >> 26;
+    h5 += carry4;
+    h4 -= carry4 * ((uint64_t) 1L << 26);
+
+    carry1 = (h1 + (int64_t)(1L << 24)) >> 25;
+    h2 += carry1;
+    h1 -= carry1 * ((uint64_t) 1L << 25);
+    carry5 = (h5 + (int64_t)(1L << 24)) >> 25;
+    h6 += carry5;
+    h5 -= carry5 * ((uint64_t) 1L << 25);
+
+    carry2 = (h2 + (int64_t)(1L << 25)) >> 26;
+    h3 += carry2;
+    h2 -= carry2 * ((uint64_t) 1L << 26);
+    carry6 = (h6 + (int64_t)(1L << 25)) >> 26;
+    h7 += carry6;
+    h6 -= carry6 * ((uint64_t) 1L << 26);
+
+    carry3 = (h3 + (int64_t)(1L << 24)) >> 25;
+    h4 += carry3;
+    h3 -= carry3 * ((uint64_t) 1L << 25);
+    carry7 = (h7 + (int64_t)(1L << 24)) >> 25;
+    h8 += carry7;
+    h7 -= carry7 * ((uint64_t) 1L << 25);
+
+    carry4 = (h4 + (int64_t)(1L << 25)) >> 26;
+    h5 += carry4;
+    h4 -= carry4 * ((uint64_t) 1L << 26);
+    carry8 = (h8 + (int64_t)(1L << 25)) >> 26;
+    h9 += carry8;
+    h8 -= carry8 * ((uint64_t) 1L << 26);
+
+    carry9 = (h9 + (int64_t)(1L << 24)) >> 25;
+    h0 += carry9 * 19;
+    h9 -= carry9 * ((uint64_t) 1L << 25);
+
+    carry0 = (h0 + (int64_t)(1L << 25)) >> 26;
+    h1 += carry0;
+    h0 -= carry0 * ((uint64_t) 1L << 26);
 
     h[0] = (int32_t) h0;
     h[1] = (int32_t) h1;
@@ -851,7 +987,8 @@ void fe_sq(fe h,const fe f)
  See fe_mul.c for discussion of implementation strategy.
  */
 
-void fe_sq2(fe h,const fe f)
+void
+fe_sq2(fe h, const fe f)
 {
     int32_t f0 = f[0];
     int32_t f1 = f[1];
@@ -863,20 +1000,22 @@ void fe_sq2(fe h,const fe f)
     int32_t f7 = f[7];
     int32_t f8 = f[8];
     int32_t f9 = f[9];
-    int32_t f0_2 = 2 * f0;
-    int32_t f1_2 = 2 * f1;
-    int32_t f2_2 = 2 * f2;
-    int32_t f3_2 = 2 * f3;
-    int32_t f4_2 = 2 * f4;
-    int32_t f5_2 = 2 * f5;
-    int32_t f6_2 = 2 * f6;
-    int32_t f7_2 = 2 * f7;
+
+    int32_t f0_2  = 2 * f0;
+    int32_t f1_2  = 2 * f1;
+    int32_t f2_2  = 2 * f2;
+    int32_t f3_2  = 2 * f3;
+    int32_t f4_2  = 2 * f4;
+    int32_t f5_2  = 2 * f5;
+    int32_t f6_2  = 2 * f6;
+    int32_t f7_2  = 2 * f7;
     int32_t f5_38 = 38 * f5; /* 1.959375*2^30 */
     int32_t f6_19 = 19 * f6; /* 1.959375*2^30 */
     int32_t f7_38 = 38 * f7; /* 1.959375*2^30 */
     int32_t f8_19 = 19 * f8; /* 1.959375*2^30 */
     int32_t f9_38 = 38 * f9; /* 1.959375*2^30 */
-    int64_t f0f0    = f0   * (int64_t) f0;
+
+    int64_t f0f0    = f0 * (int64_t) f0;
     int64_t f0f1_2  = f0_2 * (int64_t) f1;
     int64_t f0f2_2  = f0_2 * (int64_t) f2;
     int64_t f0f3_2  = f0_2 * (int64_t) f3;
@@ -895,14 +1034,14 @@ void fe_sq2(fe h,const fe f)
     int64_t f1f7_4  = f1_2 * (int64_t) f7_2;
     int64_t f1f8_2  = f1_2 * (int64_t) f8;
     int64_t f1f9_76 = f1_2 * (int64_t) f9_38;
-    int64_t f2f2    = f2   * (int64_t) f2;
+    int64_t f2f2    = f2 * (int64_t) f2;
     int64_t f2f3_2  = f2_2 * (int64_t) f3;
     int64_t f2f4_2  = f2_2 * (int64_t) f4;
     int64_t f2f5_2  = f2_2 * (int64_t) f5;
     int64_t f2f6_2  = f2_2 * (int64_t) f6;
     int64_t f2f7_2  = f2_2 * (int64_t) f7;
     int64_t f2f8_38 = f2_2 * (int64_t) f8_19;
-    int64_t f2f9_38 = f2   * (int64_t) f9_38;
+    int64_t f2f9_38 = f2 * (int64_t) f9_38;
     int64_t f3f3_2  = f3_2 * (int64_t) f3;
     int64_t f3f4_2  = f3_2 * (int64_t) f4;
     int64_t f3f5_4  = f3_2 * (int64_t) f5_2;
@@ -910,37 +1049,39 @@ void fe_sq2(fe h,const fe f)
     int64_t f3f7_76 = f3_2 * (int64_t) f7_38;
     int64_t f3f8_38 = f3_2 * (int64_t) f8_19;
     int64_t f3f9_76 = f3_2 * (int64_t) f9_38;
-    int64_t f4f4    = f4   * (int64_t) f4;
+    int64_t f4f4    = f4 * (int64_t) f4;
     int64_t f4f5_2  = f4_2 * (int64_t) f5;
     int64_t f4f6_38 = f4_2 * (int64_t) f6_19;
-    int64_t f4f7_38 = f4   * (int64_t) f7_38;
+    int64_t f4f7_38 = f4 * (int64_t) f7_38;
     int64_t f4f8_38 = f4_2 * (int64_t) f8_19;
-    int64_t f4f9_38 = f4   * (int64_t) f9_38;
-    int64_t f5f5_38 = f5   * (int64_t) f5_38;
+    int64_t f4f9_38 = f4 * (int64_t) f9_38;
+    int64_t f5f5_38 = f5 * (int64_t) f5_38;
     int64_t f5f6_38 = f5_2 * (int64_t) f6_19;
     int64_t f5f7_76 = f5_2 * (int64_t) f7_38;
     int64_t f5f8_38 = f5_2 * (int64_t) f8_19;
     int64_t f5f9_76 = f5_2 * (int64_t) f9_38;
-    int64_t f6f6_19 = f6   * (int64_t) f6_19;
-    int64_t f6f7_38 = f6   * (int64_t) f7_38;
+    int64_t f6f6_19 = f6 * (int64_t) f6_19;
+    int64_t f6f7_38 = f6 * (int64_t) f7_38;
     int64_t f6f8_38 = f6_2 * (int64_t) f8_19;
-    int64_t f6f9_38 = f6   * (int64_t) f9_38;
-    int64_t f7f7_38 = f7   * (int64_t) f7_38;
+    int64_t f6f9_38 = f6 * (int64_t) f9_38;
+    int64_t f7f7_38 = f7 * (int64_t) f7_38;
     int64_t f7f8_38 = f7_2 * (int64_t) f8_19;
     int64_t f7f9_76 = f7_2 * (int64_t) f9_38;
-    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 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 carry0;
     int64_t carry1;
     int64_t carry2;
@@ -963,24 +1104,48 @@ void fe_sq2(fe h,const fe f)
     h8 += h8;
     h9 += h9;
 
-    carry0 = (h0 + (int64_t) (1L << 25)) >> 26; h1 += carry0; h0 -= carry0 * ((uint64_t) 1L << 26);
-    carry4 = (h4 + (int64_t) (1L << 25)) >> 26; h5 += carry4; h4 -= carry4 * ((uint64_t) 1L << 26);
-
-    carry1 = (h1 + (int64_t) (1L << 24)) >> 25; h2 += carry1; h1 -= carry1 * ((uint64_t) 1L << 25);
-    carry5 = (h5 + (int64_t) (1L << 24)) >> 25; h6 += carry5; h5 -= carry5 * ((uint64_t) 1L << 25);
-
-    carry2 = (h2 + (int64_t) (1L << 25)) >> 26; h3 += carry2; h2 -= carry2 * ((uint64_t) 1L << 26);
-    carry6 = (h6 + (int64_t) (1L << 25)) >> 26; h7 += carry6; h6 -= carry6 * ((uint64_t) 1L << 26);
-
-    carry3 = (h3 + (int64_t) (1L << 24)) >> 25; h4 += carry3; h3 -= carry3 * ((uint64_t) 1L << 25);
-    carry7 = (h7 + (int64_t) (1L << 24)) >> 25; h8 += carry7; h7 -= carry7 * ((uint64_t) 1L << 25);
-
-    carry4 = (h4 + (int64_t) (1L << 25)) >> 26; h5 += carry4; h4 -= carry4 * ((uint64_t) 1L << 26);
-    carry8 = (h8 + (int64_t) (1L << 25)) >> 26; h9 += carry8; h8 -= carry8 * ((uint64_t) 1L << 26);
-
-    carry9 = (h9 + (int64_t) (1L << 24)) >> 25; h0 += carry9 * 19; h9 -= carry9 * ((uint64_t) 1L << 25);
-
-    carry0 = (h0 + (int64_t) (1L << 25)) >> 26; h1 += carry0; h0 -= carry0 * ((uint64_t) 1L << 26);
+    carry0 = (h0 + (int64_t)(1L << 25)) >> 26;
+    h1 += carry0;
+    h0 -= carry0 * ((uint64_t) 1L << 26);
+    carry4 = (h4 + (int64_t)(1L << 25)) >> 26;
+    h5 += carry4;
+    h4 -= carry4 * ((uint64_t) 1L << 26);
+
+    carry1 = (h1 + (int64_t)(1L << 24)) >> 25;
+    h2 += carry1;
+    h1 -= carry1 * ((uint64_t) 1L << 25);
+    carry5 = (h5 + (int64_t)(1L << 24)) >> 25;
+    h6 += carry5;
+    h5 -= carry5 * ((uint64_t) 1L << 25);
+
+    carry2 = (h2 + (int64_t)(1L << 25)) >> 26;
+    h3 += carry2;
+    h2 -= carry2 * ((uint64_t) 1L << 26);
+    carry6 = (h6 + (int64_t)(1L << 25)) >> 26;
+    h7 += carry6;
+    h6 -= carry6 * ((uint64_t) 1L << 26);
+
+    carry3 = (h3 + (int64_t)(1L << 24)) >> 25;
+    h4 += carry3;
+    h3 -= carry3 * ((uint64_t) 1L << 25);
+    carry7 = (h7 + (int64_t)(1L << 24)) >> 25;
+    h8 += carry7;
+    h7 -= carry7 * ((uint64_t) 1L << 25);
+
+    carry4 = (h4 + (int64_t)(1L << 25)) >> 26;
+    h5 += carry4;
+    h4 -= carry4 * ((uint64_t) 1L << 26);
+    carry8 = (h8 + (int64_t)(1L << 25)) >> 26;
+    h9 += carry8;
+    h8 -= carry8 * ((uint64_t) 1L << 26);
+
+    carry9 = (h9 + (int64_t)(1L << 24)) >> 25;
+    h0 += carry9 * 19;
+    h9 -= carry9 * ((uint64_t) 1L << 25);
+
+    carry0 = (h0 + (int64_t)(1L << 25)) >> 26;
+    h1 += carry0;
+    h0 -= carry0 * ((uint64_t) 1L << 26);
 
     h[0] = (int32_t) h0;
     h[1] = (int32_t) h1;
@@ -994,12 +1159,13 @@ void fe_sq2(fe h,const fe f)
     h[9] = (int32_t) h9;
 }
 
-void fe_invert(fe out,const fe z)
+void
+fe_invert(fe out, const fe z)
 {
-    fe t0;
-    fe t1;
-    fe t2;
-    fe t3;
+    fe  t0;
+    fe  t1;
+    fe  t2;
+    fe  t3;
     int i;
 
     fe_sq(t0, z);
@@ -1051,11 +1217,12 @@ void fe_invert(fe out,const fe z)
     fe_mul(out, t1, t0);
 }
 
-void fe_pow22523(fe out,const fe z)
+void
+fe_pow22523(fe out, const fe z)
 {
-    fe t0;
-    fe t1;
-    fe t2;
+    fe  t0;
+    fe  t1;
+    fe  t2;
     int i;
 
     fe_sq(t0, z);
@@ -1117,7 +1284,8 @@ void fe_pow22523(fe out,const fe z)
  |h| bounded by 1.1*2^26,1.1*2^25,1.1*2^26,1.1*2^25,etc.
  */
 
-void fe_sub(fe h,const fe f,const fe g)
+void
+fe_sub(fe h, const fe f, const fe g)
 {
     int32_t f0 = f[0];
     int32_t f1 = f[1];
@@ -1139,6 +1307,7 @@ void fe_sub(fe h,const fe f,const fe g)
     int32_t g7 = g[7];
     int32_t g8 = g[8];
     int32_t g9 = g[9];
+
     int32_t h0 = f0 - g0;
     int32_t h1 = f1 - g1;
     int32_t h2 = f2 - g2;
@@ -1149,6 +1318,7 @@ void fe_sub(fe h,const fe f,const fe g)
     int32_t h7 = f7 - g7;
     int32_t h8 = f8 - g8;
     int32_t h9 = f9 - g9;
+
     h[0] = h0;
     h[1] = h1;
     h[2] = h2;
@@ -1165,7 +1335,8 @@ void fe_sub(fe h,const fe f,const fe g)
  r = p + q
  */
 
-void ge_add(ge_p1p1 *r,const ge_p3 *p,const ge_cached *q)
+void
+ge_add(ge_p1p1 *r, const ge_p3 *p, const ge_cached *q)
 {
     fe t0;
 
@@ -1182,53 +1353,63 @@ void ge_add(ge_p1p1 *r,const ge_p3 *p,const ge_cached *q)
     fe_sub(r->T, t0, r->T);
 }
 
-static void slide(signed char *r,const unsigned char *a)
+static void
+slide(signed char *r, const unsigned char *a)
 {
     int i;
     int b;
     int k;
+    int ribs;
+    int cmp;
 
-    for (i = 0;i < 256;++i)
+    for (i = 0; i < 256; ++i) {
         r[i] = 1 & (a[i >> 3] >> (i & 7));
-
-    for (i = 0;i < 256;++i)
+    }
+    for (i = 0; i < 256; ++i) {
         if (r[i]) {
-            for (b = 1;b <= 6 && i + b < 256;++b) {
+            for (b = 1; b <= 6 && i + b < 256; ++b) {
                 if (r[i + b]) {
-                    if (r[i] + (r[i + b] << b) <= 15) {
-                        r[i] += r[i + b] << b; r[i + b] = 0;
-                    } else if (r[i] - (r[i + b] << b) >= -15) {
-                        r[i] -= r[i + b] << b;
-                        for (k = i + b;k < 256;++k) {
-                            if (!r[k]) {
-                                r[k] = 1;
-                                break;
+                    ribs = r[i + b] << b;
+                    cmp = r[i] + ribs;
+                    if (cmp <= 15) {
+                        r[i] = cmp;
+                        r[i + b] = 0;
+                    } else {
+                        cmp = r[i] - ribs;
+                        if (cmp >= -15) {
+                            r[i] = cmp;
+                            for (k = i + b; k < 256; ++k) {
+                                if (!r[k]) {
+                                    r[k] = 1;
+                                    break;
+                                }
+                                r[k] = 0;
                             }
-                            r[k] = 0;
+                        } else {
+                            break;
                         }
-                    } else
-                        break;
+                    }
                 }
             }
         }
-
+    }
 }
 
 static const ge_precomp Bi[8] = {
 #include "base2.h"
 };
 
-/* 37095705934669439343138083508754565189542113879843219016388785533085940283555 */
-static const fe d = {
-    -10913610,13857413,-15372611,6949391,114729,-8787816,-6275908,-3247719,-18696448,-12055116
-};
+/* 37095705934669439343138083508754565189542113879843219016388785533085940283555
+ */
+static const fe d = { -10913610, 13857413, -15372611, 6949391,   114729,
+                      -8787816,  -6275908, -3247719,  -18696448, -12055116 };
 
 /* sqrt(-1) */
-static const fe sqrtm1 = {
-    -32595792,-7943725,9377950,3500415,12389472,-272473,-25146209,-2005654,326686,11406482
-};
+static const fe sqrtm1 = { -32595792, -7943725,  9377950,  3500415, 12389472,
+                           -272473,   -25146209, -2005654, 326686,  11406482 };
 
-int ge_frombytes_negate_vartime(ge_p3 *h,const unsigned char *s)
+int
+ge_frombytes_negate_vartime(ge_p3 *h, const unsigned char *s)
 {
     fe u;
     fe v;
@@ -1236,38 +1417,38 @@ int ge_frombytes_negate_vartime(ge_p3 *h,const unsigned char *s)
     fe vxx;
     fe check;
 
-    fe_frombytes(h->Y,s);
+    fe_frombytes(h->Y, s);
     fe_1(h->Z);
-    fe_sq(u,h->Y);
-    fe_mul(v,u,d);
-    fe_sub(u,u,h->Z);       /* u = y^2-1 */
-    fe_add(v,v,h->Z);       /* v = dy^2+1 */
-
-    fe_sq(v3,v);
-    fe_mul(v3,v3,v);        /* v3 = v^3 */
-    fe_sq(h->X,v3);
-    fe_mul(h->X,h->X,v);
-    fe_mul(h->X,h->X,u);    /* x = uv^7 */
-
-    fe_pow22523(h->X,h->X); /* x = (uv^7)^((q-5)/8) */
-    fe_mul(h->X,h->X,v3);
-    fe_mul(h->X,h->X,u);    /* x = uv^3(uv^7)^((q-5)/8) */
-
-    fe_sq(vxx,h->X);
-    fe_mul(vxx,vxx,v);
-    fe_sub(check,vxx,u);    /* vx^2-u */
+    fe_sq(u, h->Y);
+    fe_mul(v, u, d);
+    fe_sub(u, u, h->Z); /* u = y^2-1 */
+    fe_add(v, v, h->Z); /* v = dy^2+1 */
+
+    fe_sq(v3, v);
+    fe_mul(v3, v3, v); /* v3 = v^3 */
+    fe_sq(h->X, v3);
+    fe_mul(h->X, h->X, v);
+    fe_mul(h->X, h->X, u); /* x = uv^7 */
+
+    fe_pow22523(h->X, h->X); /* x = (uv^7)^((q-5)/8) */
+    fe_mul(h->X, h->X, v3);
+    fe_mul(h->X, h->X, u); /* x = uv^3(uv^7)^((q-5)/8) */
+
+    fe_sq(vxx, h->X);
+    fe_mul(vxx, vxx, v);
+    fe_sub(check, vxx, u); /* vx^2-u */
     if (fe_isnonzero(check)) {
-        fe_add(check,vxx,u);  /* vx^2+u */
+        fe_add(check, vxx, u); /* vx^2+u */
         if (fe_isnonzero(check)) {
             return -1;
         }
-        fe_mul(h->X,h->X,sqrtm1);
+        fe_mul(h->X, h->X, sqrtm1);
     }
 
     if (fe_isnegative(h->X) == (s[31] >> 7)) {
-        fe_neg(h->X,h->X);
+        fe_neg(h->X, h->X);
     }
-    fe_mul(h->T,h->X,h->Y);
+    fe_mul(h->T, h->X, h->Y);
 
     return 0;
 }
@@ -1276,7 +1457,8 @@ int ge_frombytes_negate_vartime(ge_p3 *h,const unsigned char *s)
  r = p + q
  */
 
-void ge_madd(ge_p1p1 *r,const ge_p3 *p,const ge_precomp *q)
+void
+ge_madd(ge_p1p1 *r, const ge_p3 *p, const ge_precomp *q)
 {
     fe t0;
 
@@ -1296,7 +1478,8 @@ void ge_madd(ge_p1p1 *r,const ge_p3 *p,const ge_precomp *q)
  r = p - q
  */
 
-void ge_msub(ge_p1p1 *r,const ge_p3 *p,const ge_precomp *q)
+void
+ge_msub(ge_p1p1 *r, const ge_p3 *p, const ge_precomp *q)
 {
     fe t0;
 
@@ -1316,26 +1499,29 @@ void ge_msub(ge_p1p1 *r,const ge_p3 *p,const ge_precomp *q)
  r = p
  */
 
-extern void ge_p1p1_to_p2(ge_p2 *r,const ge_p1p1 *p)
+extern void
+ge_p1p1_to_p2(ge_p2 *r, const ge_p1p1 *p)
 {
-    fe_mul(r->X,p->X,p->T);
-    fe_mul(r->Y,p->Y,p->Z);
-    fe_mul(r->Z,p->Z,p->T);
+    fe_mul(r->X, p->X, p->T);
+    fe_mul(r->Y, p->Y, p->Z);
+    fe_mul(r->Z, p->Z, p->T);
 }
 
 /*
  r = p
  */
 
-extern void ge_p1p1_to_p3(ge_p3 *r,const ge_p1p1 *p)
+extern void
+ge_p1p1_to_p3(ge_p3 *r, const ge_p1p1 *p)
 {
-    fe_mul(r->X,p->X,p->T);
-    fe_mul(r->Y,p->Y,p->Z);
-    fe_mul(r->Z,p->Z,p->T);
-    fe_mul(r->T,p->X,p->Y);
+    fe_mul(r->X, p->X, p->T);
+    fe_mul(r->Y, p->Y, p->Z);
+    fe_mul(r->Z, p->Z, p->T);
+    fe_mul(r->T, p->X, p->Y);
 }
 
-void ge_p2_0(ge_p2 *h)
+void
+ge_p2_0(ge_p2 *h)
 {
     fe_0(h->X);
     fe_1(h->Y);
@@ -1346,7 +1532,8 @@ void ge_p2_0(ge_p2 *h)
  r = 2 * p
  */
 
-void ge_p2_dbl(ge_p1p1 *r,const ge_p2 *p)
+void
+ge_p2_dbl(ge_p1p1 *r, const ge_p2 *p)
 {
     fe t0;
 
@@ -1361,7 +1548,8 @@ void ge_p2_dbl(ge_p1p1 *r,const ge_p2 *p)
     fe_sub(r->T, r->T, r->Z);
 }
 
-void ge_p3_0(ge_p3 *h)
+void
+ge_p3_0(ge_p3 *h)
 {
     fe_0(h->X);
     fe_1(h->Y);
@@ -1373,40 +1561,44 @@ void ge_p3_0(ge_p3 *h)
  r = p
  */
 
-/* 2 * d = 16295367250680780974490674513165176452449235426866156013048779062215315747161 */
-static const fe d2 = {
-    -21827239,-5839606,-30745221,13898782,229458,15978800,-12551817,-6495438,29715968,9444199
-};
+/* 2 * d =
+ * 16295367250680780974490674513165176452449235426866156013048779062215315747161
+ */
+static const fe d2 = { -21827239, -5839606,  -30745221, 13898782, 229458,
+                       15978800,  -12551817, -6495438,  29715968, 9444199 };
 
-extern void ge_p3_to_cached(ge_cached *r,const ge_p3 *p)
+extern void
+ge_p3_to_cached(ge_cached *r, const ge_p3 *p)
 {
-    fe_add(r->YplusX,p->Y,p->X);
-    fe_sub(r->YminusX,p->Y,p->X);
-    fe_copy(r->Z,p->Z);
-    fe_mul(r->T2d,p->T,d2);
+    fe_add(r->YplusX, p->Y, p->X);
+    fe_sub(r->YminusX, p->Y, p->X);
+    fe_copy(r->Z, p->Z);
+    fe_mul(r->T2d, p->T, d2);
 }
 
 /*
  r = p
  */
 
-extern void ge_p3_to_p2(ge_p2 *r,const ge_p3 *p)
+extern void
+ge_p3_to_p2(ge_p2 *r, const ge_p3 *p)
 {
-    fe_copy(r->X,p->X);
-    fe_copy(r->Y,p->Y);
-    fe_copy(r->Z,p->Z);
+    fe_copy(r->X, p->X);
+    fe_copy(r->Y, p->Y);
+    fe_copy(r->Z, p->Z);
 }
 
-void ge_p3_tobytes(unsigned char *s,const ge_p3 *h)
+void
+ge_p3_tobytes(unsigned char *s, const ge_p3 *h)
 {
     fe recip;
     fe x;
     fe y;
 
-    fe_invert(recip,h->Z);
-    fe_mul(x,h->X,recip);
-    fe_mul(y,h->Y,recip);
-    fe_tobytes(s,y);
+    fe_invert(recip, h->Z);
+    fe_mul(x, h->X, recip);
+    fe_mul(y, h->Y, recip);
+    fe_tobytes(s, y);
     s[31] ^= fe_isnegative(x) << 7;
 }
 
@@ -1414,45 +1606,53 @@ void ge_p3_tobytes(unsigned char *s,const ge_p3 *h)
  r = 2 * p
  */
 
-void ge_p3_dbl(ge_p1p1 *r,const ge_p3 *p)
+void
+ge_p3_dbl(ge_p1p1 *r, const ge_p3 *p)
 {
     ge_p2 q;
-    ge_p3_to_p2(&q,p);
-    ge_p2_dbl(r,&q);
+    ge_p3_to_p2(&q, p);
+    ge_p2_dbl(r, &q);
 }
 
-void ge_precomp_0(ge_precomp *h)
+void
+ge_precomp_0(ge_precomp *h)
 {
     fe_1(h->yplusx);
     fe_1(h->yminusx);
     fe_0(h->xy2d);
 }
 
-static unsigned char equal(signed char b,signed char c)
+static unsigned char
+equal(signed char b, signed char c)
 {
     unsigned char ub = b;
     unsigned char uc = c;
-    unsigned char x = ub ^ uc; /* 0: yes; 1..255: no */
-    uint32_t y = x; /* 0: yes; 1..255: no */
-    y -= 1; /* 4294967295: yes; 0..254: no */
+    unsigned char x  = ub ^ uc; /* 0: yes; 1..255: no */
+    uint32_t      y  = x;       /* 0: yes; 1..255: no */
+
+    y -= 1;   /* 4294967295: yes; 0..254: no */
     y >>= 31; /* 1: yes; 0: no */
 
     return y;
 }
 
-static unsigned char negative(signed char b)
+static unsigned char
+negative(signed char b)
 {
-    uint64_t x = b; /* 18446744073709551361..18446744073709551615: yes; 0..255: no */
+    uint64_t x =
+        b; /* 18446744073709551361..18446744073709551615: yes; 0..255: no */
+
     x >>= 63; /* 1: yes; 0: no */
 
     return x;
 }
 
-static void cmov(ge_precomp *t,const ge_precomp *u,unsigned char b)
+static void
+cmov(ge_precomp *t, const ge_precomp *u, unsigned char b)
 {
-    fe_cmov(t->yplusx,u->yplusx,b);
-    fe_cmov(t->yminusx,u->yminusx,b);
-    fe_cmov(t->xy2d,u->xy2d,b);
+    fe_cmov(t->yplusx, u->yplusx, b);
+    fe_cmov(t->yminusx, u->yminusx, b);
+    fe_cmov(t->xy2d, u->xy2d, b);
 }
 
 /* base[i][j] = (j+1)*256^i*B */
@@ -1460,32 +1660,34 @@ static const ge_precomp base[32][8] = {
 #include "base.h"
 };
 
-static void ge_select(ge_precomp *t,int pos,signed char b)
+static void
+ge_select(ge_precomp *t, int pos, signed char b)
 {
-    ge_precomp minust;
+    ge_precomp    minust;
     unsigned char bnegative = negative(b);
-    unsigned char babs = b - (((-bnegative) & b) * ((signed char) 1 << 1));
+    unsigned char babs      = b - (((-bnegative) & b) * ((signed char) 1 << 1));
 
     ge_precomp_0(t);
-    cmov(t,&base[pos][0],equal(babs,1));
-    cmov(t,&base[pos][1],equal(babs,2));
-    cmov(t,&base[pos][2],equal(babs,3));
-    cmov(t,&base[pos][3],equal(babs,4));
-    cmov(t,&base[pos][4],equal(babs,5));
-    cmov(t,&base[pos][5],equal(babs,6));
-    cmov(t,&base[pos][6],equal(babs,7));
-    cmov(t,&base[pos][7],equal(babs,8));
-    fe_copy(minust.yplusx,t->yminusx);
-    fe_copy(minust.yminusx,t->yplusx);
-    fe_neg(minust.xy2d,t->xy2d);
-    cmov(t,&minust,bnegative);
+    cmov(t, &base[pos][0], equal(babs, 1));
+    cmov(t, &base[pos][1], equal(babs, 2));
+    cmov(t, &base[pos][2], equal(babs, 3));
+    cmov(t, &base[pos][3], equal(babs, 4));
+    cmov(t, &base[pos][4], equal(babs, 5));
+    cmov(t, &base[pos][5], equal(babs, 6));
+    cmov(t, &base[pos][6], equal(babs, 7));
+    cmov(t, &base[pos][7], equal(babs, 8));
+    fe_copy(minust.yplusx, t->yminusx);
+    fe_copy(minust.yminusx, t->yplusx);
+    fe_neg(minust.xy2d, t->xy2d);
+    cmov(t, &minust, bnegative);
 }
 
 /*
  r = p - q
  */
 
-void ge_sub(ge_p1p1 *r,const ge_p3 *p,const ge_cached *q)
+void
+ge_sub(ge_p1p1 *r, const ge_p3 *p, const ge_cached *q)
 {
     fe t0;
 
@@ -1502,16 +1704,17 @@ void ge_sub(ge_p1p1 *r,const ge_p3 *p,const ge_cached *q)
     fe_add(r->T, t0, r->T);
 }
 
-void ge_tobytes(unsigned char *s,const ge_p2 *h)
+void
+ge_tobytes(unsigned char *s, const ge_p2 *h)
 {
     fe recip;
     fe x;
     fe y;
 
-    fe_invert(recip,h->Z);
-    fe_mul(x,h->X,recip);
-    fe_mul(y,h->Y,recip);
-    fe_tobytes(s,y);
+    fe_invert(recip, h->Z);
+    fe_mul(x, h->X, recip);
+    fe_mul(y, h->Y, recip);
+    fe_tobytes(s, y);
     s[31] ^= fe_isnegative(x) << 7;
 }
 
@@ -1531,110 +1734,146 @@ void ge_tobytes(unsigned char *s,const ge_p2 *h)
  B is the Ed25519 base point (x,4/5) with x positive.
  */
 
-void ge_double_scalarmult_vartime(ge_p2 *r,const unsigned char *a,const ge_p3 *A,const unsigned char *b)
+void
+ge_double_scalarmult_vartime(ge_p2 *r, const unsigned char *a, const ge_p3 *A,
+                             const unsigned char *b)
 {
     signed char aslide[256];
     signed char bslide[256];
-    ge_cached Ai[8]; /* A,3A,5A,7A,9A,11A,13A,15A */
-    ge_p1p1 t;
-    ge_p3 u;
-    ge_p3 A2;
-    int i;
-
-    slide(aslide,a);
-    slide(bslide,b);
-
-    ge_p3_to_cached(&Ai[0],A);
-    ge_p3_dbl(&t,A); ge_p1p1_to_p3(&A2,&t);
-    ge_add(&t,&A2,&Ai[0]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[1],&u);
-    ge_add(&t,&A2,&Ai[1]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[2],&u);
-    ge_add(&t,&A2,&Ai[2]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[3],&u);
-    ge_add(&t,&A2,&Ai[3]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[4],&u);
-    ge_add(&t,&A2,&Ai[4]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[5],&u);
-    ge_add(&t,&A2,&Ai[5]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[6],&u);
-    ge_add(&t,&A2,&Ai[6]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[7],&u);
+    ge_cached   Ai[8]; /* A,3A,5A,7A,9A,11A,13A,15A */
+    ge_p1p1     t;
+    ge_p3       u;
+    ge_p3       A2;
+    int         i;
+
+    slide(aslide, a);
+    slide(bslide, b);
+
+    ge_p3_to_cached(&Ai[0], A);
+    ge_p3_dbl(&t, A);
+    ge_p1p1_to_p3(&A2, &t);
+    ge_add(&t, &A2, &Ai[0]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[1], &u);
+    ge_add(&t, &A2, &Ai[1]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[2], &u);
+    ge_add(&t, &A2, &Ai[2]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[3], &u);
+    ge_add(&t, &A2, &Ai[3]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[4], &u);
+    ge_add(&t, &A2, &Ai[4]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[5], &u);
+    ge_add(&t, &A2, &Ai[5]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[6], &u);
+    ge_add(&t, &A2, &Ai[6]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[7], &u);
 
     ge_p2_0(r);
 
-    for (i = 255;i >= 0;--i) {
-        if (aslide[i] || bslide[i]) break;
+    for (i = 255; i >= 0; --i) {
+        if (aslide[i] || bslide[i])
+            break;
     }
 
-    for (;i >= 0;--i) {
-        ge_p2_dbl(&t,r);
+    for (; i >= 0; --i) {
+        ge_p2_dbl(&t, r);
 
         if (aslide[i] > 0) {
-            ge_p1p1_to_p3(&u,&t);
-            ge_add(&t,&u,&Ai[aslide[i]/2]);
+            ge_p1p1_to_p3(&u, &t);
+            ge_add(&t, &u, &Ai[aslide[i] / 2]);
         } else if (aslide[i] < 0) {
-            ge_p1p1_to_p3(&u,&t);
-            ge_sub(&t,&u,&Ai[(-aslide[i])/2]);
+            ge_p1p1_to_p3(&u, &t);
+            ge_sub(&t, &u, &Ai[(-aslide[i]) / 2]);
         }
 
         if (bslide[i] > 0) {
-            ge_p1p1_to_p3(&u,&t);
-            ge_madd(&t,&u,&Bi[bslide[i]/2]);
+            ge_p1p1_to_p3(&u, &t);
+            ge_madd(&t, &u, &Bi[bslide[i] / 2]);
         } else if (bslide[i] < 0) {
-            ge_p1p1_to_p3(&u,&t);
-            ge_msub(&t,&u,&Bi[(-bslide[i])/2]);
+            ge_p1p1_to_p3(&u, &t);
+            ge_msub(&t, &u, &Bi[(-bslide[i]) / 2]);
         }
 
-        ge_p1p1_to_p2(r,&t);
+        ge_p1p1_to_p2(r, &t);
     }
 }
 
-void ge_scalarmult_vartime(ge_p3 *r,const unsigned char *a,const ge_p3 *A)
+void
+ge_scalarmult_vartime(ge_p3 *r, const unsigned char *a, const ge_p3 *A)
 {
     signed char aslide[256];
-    ge_cached Ai[8];
-    ge_p1p1 t;
-    ge_p3 u;
-    ge_p3 A2;
-    int i;
-
-    slide(aslide,a);
-
-    ge_p3_to_cached(&Ai[0],A);
-    ge_p3_dbl(&t,A); ge_p1p1_to_p3(&A2,&t);
-    ge_add(&t,&A2,&Ai[0]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[1],&u);
-    ge_add(&t,&A2,&Ai[1]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[2],&u);
-    ge_add(&t,&A2,&Ai[2]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[3],&u);
-    ge_add(&t,&A2,&Ai[3]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[4],&u);
-    ge_add(&t,&A2,&Ai[4]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[5],&u);
-    ge_add(&t,&A2,&Ai[5]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[6],&u);
-    ge_add(&t,&A2,&Ai[6]); ge_p1p1_to_p3(&u,&t); ge_p3_to_cached(&Ai[7],&u);
+    ge_cached   Ai[8];
+    ge_p1p1     t;
+    ge_p3       u;
+    ge_p3       A2;
+    int         i;
+
+    slide(aslide, a);
+
+    ge_p3_to_cached(&Ai[0], A);
+    ge_p3_dbl(&t, A);
+    ge_p1p1_to_p3(&A2, &t);
+    ge_add(&t, &A2, &Ai[0]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[1], &u);
+    ge_add(&t, &A2, &Ai[1]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[2], &u);
+    ge_add(&t, &A2, &Ai[2]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[3], &u);
+    ge_add(&t, &A2, &Ai[3]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[4], &u);
+    ge_add(&t, &A2, &Ai[4]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[5], &u);
+    ge_add(&t, &A2, &Ai[5]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[6], &u);
+    ge_add(&t, &A2, &Ai[6]);
+    ge_p1p1_to_p3(&u, &t);
+    ge_p3_to_cached(&Ai[7], &u);
 
     ge_p3_0(r);
 
-    for (i = 255;i >= 0;--i) {
-        if (aslide[i]) break;
+    for (i = 255; i >= 0; --i) {
+        if (aslide[i])
+            break;
     }
 
-    for (;i >= 0;--i) {
-        ge_p3_dbl(&t,r);
+    for (; i >= 0; --i) {
+        ge_p3_dbl(&t, r);
 
         if (aslide[i] > 0) {
-            ge_p1p1_to_p3(&u,&t);
-            ge_add(&t,&u,&Ai[aslide[i]/2]);
+            ge_p1p1_to_p3(&u, &t);
+            ge_add(&t, &u, &Ai[aslide[i] / 2]);
         } else if (aslide[i] < 0) {
-            ge_p1p1_to_p3(&u,&t);
-            ge_sub(&t,&u,&Ai[(-aslide[i])/2]);
+            ge_p1p1_to_p3(&u, &t);
+            ge_sub(&t, &u, &Ai[(-aslide[i]) / 2]);
         }
 
-        ge_p1p1_to_p3(r,&t);
+        ge_p1p1_to_p3(r, &t);
     }
 }
 
-void ge_scalarmult_base(ge_p3 *h,const unsigned char *a)
+void
+ge_scalarmult_base(ge_p3 *h, const unsigned char *a)
 {
     signed char e[64];
     signed char carry;
-    ge_p1p1 r;
-    ge_p2 s;
-    ge_precomp t;
-    int i;
+    ge_p1p1     r;
+    ge_p2       s;
+    ge_precomp  t;
+    int         i;
 
-    for (i = 0;i < 32;++i) {
+    for (i = 0; i < 32; ++i) {
         e[2 * i + 0] = (a[i] >> 0) & 15;
         e[2 * i + 1] = (a[i] >> 4) & 15;
     }
@@ -1642,7 +1881,7 @@ void ge_scalarmult_base(ge_p3 *h,const unsigned char *a)
     /* e[63] is between 0 and 7 */
 
     carry = 0;
-    for (i = 0;i < 63;++i) {
+    for (i = 0; i < 63; ++i) {
         e[i] += carry;
         carry = e[i] + 8;
         carry >>= 4;
@@ -1652,19 +1891,25 @@ void ge_scalarmult_base(ge_p3 *h,const unsigned char *a)
     /* each e[i] is between -8 and 8 */
 
     ge_p3_0(h);
-    for (i = 1;i < 64;i += 2) {
-        ge_select(&t,i / 2,e[i]);
-        ge_madd(&r,h,&t); ge_p1p1_to_p3(h,&r);
+    for (i = 1; i < 64; i += 2) {
+        ge_select(&t, i / 2, e[i]);
+        ge_madd(&r, h, &t);
+        ge_p1p1_to_p3(h, &r);
     }
 
-    ge_p3_dbl(&r,h);  ge_p1p1_to_p2(&s,&r);
-    ge_p2_dbl(&r,&s); ge_p1p1_to_p2(&s,&r);
-    ge_p2_dbl(&r,&s); ge_p1p1_to_p2(&s,&r);
-    ge_p2_dbl(&r,&s); ge_p1p1_to_p3(h,&r);
-
-    for (i = 0;i < 64;i += 2) {
-        ge_select(&t,i / 2,e[i]);
-        ge_madd(&r,h,&t); ge_p1p1_to_p3(h,&r);
+    ge_p3_dbl(&r, h);
+    ge_p1p1_to_p2(&s, &r);
+    ge_p2_dbl(&r, &s);
+    ge_p1p1_to_p2(&s, &r);
+    ge_p2_dbl(&r, &s);
+    ge_p1p1_to_p2(&s, &r);
+    ge_p2_dbl(&r, &s);
+    ge_p1p1_to_p3(h, &r);
+
+    for (i = 0; i < 64; i += 2) {
+        ge_select(&t, i / 2, e[i]);
+        ge_madd(&r, h, &t);
+        ge_p1p1_to_p3(h, &r);
     }
 }
 
@@ -1679,44 +1924,49 @@ void ge_scalarmult_base(ge_p3 *h,const unsigned char *a)
  where l = 2^252 + 27742317777372353535851937790883648493.
  */
 
-void sc_muladd(unsigned char *s,const unsigned char *a,const unsigned char *b,const unsigned char *c)
+void
+sc_muladd(unsigned char *s, const unsigned char *a, const unsigned char *b,
+          const unsigned char *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 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 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 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 s0;
     int64_t s1;
     int64_t s2;
@@ -1741,6 +1991,7 @@ void sc_muladd(unsigned char *s,const unsigned char *a,const unsigned char *b,co
     int64_t s21;
     int64_t s22;
     int64_t s23;
+
     int64_t carry0;
     int64_t carry1;
     int64_t carry2;
@@ -1765,55 +2016,112 @@ void sc_muladd(unsigned char *s,const unsigned char *a,const unsigned char *b,co
     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;
+    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 + (int64_t) (1L << 20)) >> 21; s1 += carry0; s0 -= carry0 * ((uint64_t) 1L << 21);
-    carry2 = (s2 + (int64_t) (1L << 20)) >> 21; s3 += carry2; s2 -= carry2 * ((uint64_t) 1L << 21);
-    carry4 = (s4 + (int64_t) (1L << 20)) >> 21; s5 += carry4; s4 -= carry4 * ((uint64_t) 1L << 21);
-    carry6 = (s6 + (int64_t) (1L << 20)) >> 21; s7 += carry6; s6 -= carry6 * ((uint64_t) 1L << 21);
-    carry8 = (s8 + (int64_t) (1L << 20)) >> 21; s9 += carry8; s8 -= carry8 * ((uint64_t) 1L << 21);
-    carry10 = (s10 + (int64_t) (1L << 20)) >> 21; s11 += carry10; s10 -= carry10 * ((uint64_t) 1L << 21);
-    carry12 = (s12 + (int64_t) (1L << 20)) >> 21; s13 += carry12; s12 -= carry12 * ((uint64_t) 1L << 21);
-    carry14 = (s14 + (int64_t) (1L << 20)) >> 21; s15 += carry14; s14 -= carry14 * ((uint64_t) 1L << 21);
-    carry16 = (s16 + (int64_t) (1L << 20)) >> 21; s17 += carry16; s16 -= carry16 * ((uint64_t) 1L << 21);
-    carry18 = (s18 + (int64_t) (1L << 20)) >> 21; s19 += carry18; s18 -= carry18 * ((uint64_t) 1L << 21);
-    carry20 = (s20 + (int64_t) (1L << 20)) >> 21; s21 += carry20; s20 -= carry20 * ((uint64_t) 1L << 21);
-    carry22 = (s22 + (int64_t) (1L << 20)) >> 21; s23 += carry22; s22 -= carry22 * ((uint64_t) 1L << 21);
-
-    carry1 = (s1 + (int64_t) (1L << 20)) >> 21; s2 += carry1; s1 -= carry1 * ((uint64_t) 1L << 21);
-    carry3 = (s3 + (int64_t) (1L << 20)) >> 21; s4 += carry3; s3 -= carry3 * ((uint64_t) 1L << 21);
-    carry5 = (s5 + (int64_t) (1L << 20)) >> 21; s6 += carry5; s5 -= carry5 * ((uint64_t) 1L << 21);
-    carry7 = (s7 + (int64_t) (1L << 20)) >> 21; s8 += carry7; s7 -= carry7 * ((uint64_t) 1L << 21);
-    carry9 = (s9 + (int64_t) (1L << 20)) >> 21; s10 += carry9; s9 -= carry9 * ((uint64_t) 1L << 21);
-    carry11 = (s11 + (int64_t) (1L << 20)) >> 21; s12 += carry11; s11 -= carry11 * ((uint64_t) 1L << 21);
-    carry13 = (s13 + (int64_t) (1L << 20)) >> 21; s14 += carry13; s13 -= carry13 * ((uint64_t) 1L << 21);
-    carry15 = (s15 + (int64_t) (1L << 20)) >> 21; s16 += carry15; s15 -= carry15 * ((uint64_t) 1L << 21);
-    carry17 = (s17 + (int64_t) (1L << 20)) >> 21; s18 += carry17; s17 -= carry17 * ((uint64_t) 1L << 21);
-    carry19 = (s19 + (int64_t) (1L << 20)) >> 21; s20 += carry19; s19 -= carry19 * ((uint64_t) 1L << 21);
-    carry21 = (s21 + (int64_t) (1L << 20)) >> 21; s22 += carry21; s21 -= carry21 * ((uint64_t) 1L << 21);
+    carry0 = (s0 + (int64_t)(1L << 20)) >> 21;
+    s1 += carry0;
+    s0 -= carry0 * ((uint64_t) 1L << 21);
+    carry2 = (s2 + (int64_t)(1L << 20)) >> 21;
+    s3 += carry2;
+    s2 -= carry2 * ((uint64_t) 1L << 21);
+    carry4 = (s4 + (int64_t)(1L << 20)) >> 21;
+    s5 += carry4;
+    s4 -= carry4 * ((uint64_t) 1L << 21);
+    carry6 = (s6 + (int64_t)(1L << 20)) >> 21;
+    s7 += carry6;
+    s6 -= carry6 * ((uint64_t) 1L << 21);
+    carry8 = (s8 + (int64_t)(1L << 20)) >> 21;
+    s9 += carry8;
+    s8 -= carry8 * ((uint64_t) 1L << 21);
+    carry10 = (s10 + (int64_t)(1L << 20)) >> 21;
+    s11 += carry10;
+    s10 -= carry10 * ((uint64_t) 1L << 21);
+    carry12 = (s12 + (int64_t)(1L << 20)) >> 21;
+    s13 += carry12;
+    s12 -= carry12 * ((uint64_t) 1L << 21);
+    carry14 = (s14 + (int64_t)(1L << 20)) >> 21;
+    s15 += carry14;
+    s14 -= carry14 * ((uint64_t) 1L << 21);
+    carry16 = (s16 + (int64_t)(1L << 20)) >> 21;
+    s17 += carry16;
+    s16 -= carry16 * ((uint64_t) 1L << 21);
+    carry18 = (s18 + (int64_t)(1L << 20)) >> 21;
+    s19 += carry18;
+    s18 -= carry18 * ((uint64_t) 1L << 21);
+    carry20 = (s20 + (int64_t)(1L << 20)) >> 21;
+    s21 += carry20;
+    s20 -= carry20 * ((uint64_t) 1L << 21);
+    carry22 = (s22 + (int64_t)(1L << 20)) >> 21;
+    s23 += carry22;
+    s22 -= carry22 * ((uint64_t) 1L << 21);
+
+    carry1 = (s1 + (int64_t)(1L << 20)) >> 21;
+    s2 += carry1;
+    s1 -= carry1 * ((uint64_t) 1L << 21);
+    carry3 = (s3 + (int64_t)(1L << 20)) >> 21;
+    s4 += carry3;
+    s3 -= carry3 * ((uint64_t) 1L << 21);
+    carry5 = (s5 + (int64_t)(1L << 20)) >> 21;
+    s6 += carry5;
+    s5 -= carry5 * ((uint64_t) 1L << 21);
+    carry7 = (s7 + (int64_t)(1L << 20)) >> 21;
+    s8 += carry7;
+    s7 -= carry7 * ((uint64_t) 1L << 21);
+    carry9 = (s9 + (int64_t)(1L << 20)) >> 21;
+    s10 += carry9;
+    s9 -= carry9 * ((uint64_t) 1L << 21);
+    carry11 = (s11 + (int64_t)(1L << 20)) >> 21;
+    s12 += carry11;
+    s11 -= carry11 * ((uint64_t) 1L << 21);
+    carry13 = (s13 + (int64_t)(1L << 20)) >> 21;
+    s14 += carry13;
+    s13 -= carry13 * ((uint64_t) 1L << 21);
+    carry15 = (s15 + (int64_t)(1L << 20)) >> 21;
+    s16 += carry15;
+    s15 -= carry15 * ((uint64_t) 1L << 21);
+    carry17 = (s17 + (int64_t)(1L << 20)) >> 21;
+    s18 += carry17;
+    s17 -= carry17 * ((uint64_t) 1L << 21);
+    carry19 = (s19 + (int64_t)(1L << 20)) >> 21;
+    s20 += carry19;
+    s19 -= carry19 * ((uint64_t) 1L << 21);
+    carry21 = (s21 + (int64_t)(1L << 20)) >> 21;
+    s22 += carry21;
+    s21 -= carry21 * ((uint64_t) 1L << 21);
 
     s11 += s23 * 666643;
     s12 += s23 * 470296;
@@ -1857,18 +2165,40 @@ void sc_muladd(unsigned char *s,const unsigned char *a,const unsigned char *b,co
     s10 += s18 * 136657;
     s11 -= s18 * 683901;
 
-    carry6 = (s6 + (int64_t) (1L << 20)) >> 21; s7 += carry6; s6 -= carry6 * ((uint64_t) 1L << 21);
-    carry8 = (s8 + (int64_t) (1L << 20)) >> 21; s9 += carry8; s8 -= carry8 * ((uint64_t) 1L << 21);
-    carry10 = (s10 + (int64_t) (1L << 20)) >> 21; s11 += carry10; s10 -= carry10 * ((uint64_t) 1L << 21);
-    carry12 = (s12 + (int64_t) (1L << 20)) >> 21; s13 += carry12; s12 -= carry12 * ((uint64_t) 1L << 21);
-    carry14 = (s14 + (int64_t) (1L << 20)) >> 21; s15 += carry14; s14 -= carry14 * ((uint64_t) 1L << 21);
-    carry16 = (s16 + (int64_t) (1L << 20)) >> 21; s17 += carry16; s16 -= carry16 * ((uint64_t) 1L << 21);
-
-    carry7 = (s7 + (int64_t) (1L << 20)) >> 21; s8 += carry7; s7 -= carry7 * ((uint64_t) 1L << 21);
-    carry9 = (s9 + (int64_t) (1L << 20)) >> 21; s10 += carry9; s9 -= carry9 * ((uint64_t) 1L << 21);
-    carry11 = (s11 + (int64_t) (1L << 20)) >> 21; s12 += carry11; s11 -= carry11 * ((uint64_t) 1L << 21);
-    carry13 = (s13 + (int64_t) (1L << 20)) >> 21; s14 += carry13; s13 -= carry13 * ((uint64_t) 1L << 21);
-    carry15 = (s15 + (int64_t) (1L << 20)) >> 21; s16 += carry15; s15 -= carry15 * ((uint64_t) 1L << 21);
+    carry6 = (s6 + (int64_t)(1L << 20)) >> 21;
+    s7 += carry6;
+    s6 -= carry6 * ((uint64_t) 1L << 21);
+    carry8 = (s8 + (int64_t)(1L << 20)) >> 21;
+    s9 += carry8;
+    s8 -= carry8 * ((uint64_t) 1L << 21);
+    carry10 = (s10 + (int64_t)(1L << 20)) >> 21;
+    s11 += carry10;
+    s10 -= carry10 * ((uint64_t) 1L << 21);
+    carry12 = (s12 + (int64_t)(1L << 20)) >> 21;
+    s13 += carry12;
+    s12 -= carry12 * ((uint64_t) 1L << 21);
+    carry14 = (s14 + (int64_t)(1L << 20)) >> 21;
+    s15 += carry14;
+    s14 -= carry14 * ((uint64_t) 1L << 21);
+    carry16 = (s16 + (int64_t)(1L << 20)) >> 21;
+    s17 += carry16;
+    s16 -= carry16 * ((uint64_t) 1L << 21);
+
+    carry7 = (s7 + (int64_t)(1L << 20)) >> 21;
+    s8 += carry7;
+    s7 -= carry7 * ((uint64_t) 1L << 21);
+    carry9 = (s9 + (int64_t)(1L << 20)) >> 21;
+    s10 += carry9;
+    s9 -= carry9 * ((uint64_t) 1L << 21);
+    carry11 = (s11 + (int64_t)(1L << 20)) >> 21;
+    s12 += carry11;
+    s11 -= carry11 * ((uint64_t) 1L << 21);
+    carry13 = (s13 + (int64_t)(1L << 20)) >> 21;
+    s14 += carry13;
+    s13 -= carry13 * ((uint64_t) 1L << 21);
+    carry15 = (s15 + (int64_t)(1L << 20)) >> 21;
+    s16 += carry15;
+    s15 -= carry15 * ((uint64_t) 1L << 21);
 
     s5 += s17 * 666643;
     s6 += s17 * 470296;
@@ -1913,19 +2243,43 @@ void sc_muladd(unsigned char *s,const unsigned char *a,const unsigned char *b,co
     s5 -= s12 * 683901;
     s12 = 0;
 
-    carry0 = (s0 + (int64_t) (1L << 20)) >> 21; s1 += carry0; s0 -= carry0 * ((uint64_t) 1L << 21);
-    carry2 = (s2 + (int64_t) (1L << 20)) >> 21; s3 += carry2; s2 -= carry2 * ((uint64_t) 1L << 21);
-    carry4 = (s4 + (int64_t) (1L << 20)) >> 21; s5 += carry4; s4 -= carry4 * ((uint64_t) 1L << 21);
-    carry6 = (s6 + (int64_t) (1L << 20)) >> 21; s7 += carry6; s6 -= carry6 * ((uint64_t) 1L << 21);
-    carry8 = (s8 + (int64_t) (1L << 20)) >> 21; s9 += carry8; s8 -= carry8 * ((uint64_t) 1L << 21);
-    carry10 = (s10 + (int64_t) (1L << 20)) >> 21; s11 += carry10; s10 -= carry10 * ((uint64_t) 1L << 21);
-
-    carry1 = (s1 + (int64_t) (1L << 20)) >> 21; s2 += carry1; s1 -= carry1 * ((uint64_t) 1L << 21);
-    carry3 = (s3 + (int64_t) (1L << 20)) >> 21; s4 += carry3; s3 -= carry3 * ((uint64_t) 1L << 21);
-    carry5 = (s5 + (int64_t) (1L << 20)) >> 21; s6 += carry5; s5 -= carry5 * ((uint64_t) 1L << 21);
-    carry7 = (s7 + (int64_t) (1L << 20)) >> 21; s8 += carry7; s7 -= carry7 * ((uint64_t) 1L << 21);
-    carry9 = (s9 + (int64_t) (1L << 20)) >> 21; s10 += carry9; s9 -= carry9 * ((uint64_t) 1L << 21);
-    carry11 = (s11 + (int64_t) (1L << 20)) >> 21; s12 += carry11; s11 -= carry11 * ((uint64_t) 1L << 21);
+    carry0 = (s0 + (int64_t)(1L << 20)) >> 21;
+    s1 += carry0;
+    s0 -= carry0 * ((uint64_t) 1L << 21);
+    carry2 = (s2 + (int64_t)(1L << 20)) >> 21;
+    s3 += carry2;
+    s2 -= carry2 * ((uint64_t) 1L << 21);
+    carry4 = (s4 + (int64_t)(1L << 20)) >> 21;
+    s5 += carry4;
+    s4 -= carry4 * ((uint64_t) 1L << 21);
+    carry6 = (s6 + (int64_t)(1L << 20)) >> 21;
+    s7 += carry6;
+    s6 -= carry6 * ((uint64_t) 1L << 21);
+    carry8 = (s8 + (int64_t)(1L << 20)) >> 21;
+    s9 += carry8;
+    s8 -= carry8 * ((uint64_t) 1L << 21);
+    carry10 = (s10 + (int64_t)(1L << 20)) >> 21;
+    s11 += carry10;
+    s10 -= carry10 * ((uint64_t) 1L << 21);
+
+    carry1 = (s1 + (int64_t)(1L << 20)) >> 21;
+    s2 += carry1;
+    s1 -= carry1 * ((uint64_t) 1L << 21);
+    carry3 = (s3 + (int64_t)(1L << 20)) >> 21;
+    s4 += carry3;
+    s3 -= carry3 * ((uint64_t) 1L << 21);
+    carry5 = (s5 + (int64_t)(1L << 20)) >> 21;
+    s6 += carry5;
+    s5 -= carry5 * ((uint64_t) 1L << 21);
+    carry7 = (s7 + (int64_t)(1L << 20)) >> 21;
+    s8 += carry7;
+    s7 -= carry7 * ((uint64_t) 1L << 21);
+    carry9 = (s9 + (int64_t)(1L << 20)) >> 21;
+    s10 += carry9;
+    s9 -= carry9 * ((uint64_t) 1L << 21);
+    carry11 = (s11 + (int64_t)(1L << 20)) >> 21;
+    s12 += carry11;
+    s11 -= carry11 * ((uint64_t) 1L << 21);
 
     s0 += s12 * 666643;
     s1 += s12 * 470296;
@@ -1935,18 +2289,42 @@ void sc_muladd(unsigned char *s,const unsigned char *a,const unsigned char *b,co
     s5 -= s12 * 683901;
     s12 = 0;
 
-    carry0 = s0 >> 21; s1 += carry0; s0 -= carry0 * ((uint64_t) 1L << 21);
-    carry1 = s1 >> 21; s2 += carry1; s1 -= carry1 * ((uint64_t) 1L << 21);
-    carry2 = s2 >> 21; s3 += carry2; s2 -= carry2 * ((uint64_t) 1L << 21);
-    carry3 = s3 >> 21; s4 += carry3; s3 -= carry3 * ((uint64_t) 1L << 21);
-    carry4 = s4 >> 21; s5 += carry4; s4 -= carry4 * ((uint64_t) 1L << 21);
-    carry5 = s5 >> 21; s6 += carry5; s5 -= carry5 * ((uint64_t) 1L << 21);
-    carry6 = s6 >> 21; s7 += carry6; s6 -= carry6 * ((uint64_t) 1L << 21);
-    carry7 = s7 >> 21; s8 += carry7; s7 -= carry7 * ((uint64_t) 1L << 21);
-    carry8 = s8 >> 21; s9 += carry8; s8 -= carry8 * ((uint64_t) 1L << 21);
-    carry9 = s9 >> 21; s10 += carry9; s9 -= carry9 * ((uint64_t) 1L << 21);
-    carry10 = s10 >> 21; s11 += carry10; s10 -= carry10 * ((uint64_t) 1L << 21);
-    carry11 = s11 >> 21; s12 += carry11; s11 -= carry11 * ((uint64_t) 1L << 21);
+    carry0 = s0 >> 21;
+    s1 += carry0;
+    s0 -= carry0 * ((uint64_t) 1L << 21);
+    carry1 = s1 >> 21;
+    s2 += carry1;
+    s1 -= carry1 * ((uint64_t) 1L << 21);
+    carry2 = s2 >> 21;
+    s3 += carry2;
+    s2 -= carry2 * ((uint64_t) 1L << 21);
+    carry3 = s3 >> 21;
+    s4 += carry3;
+    s3 -= carry3 * ((uint64_t) 1L << 21);
+    carry4 = s4 >> 21;
+    s5 += carry4;
+    s4 -= carry4 * ((uint64_t) 1L << 21);
+    carry5 = s5 >> 21;
+    s6 += carry5;
+    s5 -= carry5 * ((uint64_t) 1L << 21);
+    carry6 = s6 >> 21;
+    s7 += carry6;
+    s6 -= carry6 * ((uint64_t) 1L << 21);
+    carry7 = s7 >> 21;
+    s8 += carry7;
+    s7 -= carry7 * ((uint64_t) 1L << 21);
+    carry8 = s8 >> 21;
+    s9 += carry8;
+    s8 -= carry8 * ((uint64_t) 1L << 21);
+    carry9 = s9 >> 21;
+    s10 += carry9;
+    s9 -= carry9 * ((uint64_t) 1L << 21);
+    carry10 = s10 >> 21;
+    s11 += carry10;
+    s10 -= carry10 * ((uint64_t) 1L << 21);
+    carry11 = s11 >> 21;
+    s12 += carry11;
+    s11 -= carry11 * ((uint64_t) 1L << 21);
 
     s0 += s12 * 666643;
     s1 += s12 * 470296;
@@ -1955,28 +2333,50 @@ void sc_muladd(unsigned char *s,const unsigned char *a,const unsigned char *b,co
     s4 += s12 * 136657;
     s5 -= s12 * 683901;
 
-    carry0 = s0 >> 21; s1 += carry0; s0 -= carry0 * ((uint64_t) 1L << 21);
-    carry1 = s1 >> 21; s2 += carry1; s1 -= carry1 * ((uint64_t) 1L << 21);
-    carry2 = s2 >> 21; s3 += carry2; s2 -= carry2 * ((uint64_t) 1L << 21);
-    carry3 = s3 >> 21; s4 += carry3; s3 -= carry3 * ((uint64_t) 1L << 21);
-    carry4 = s4 >> 21; s5 += carry4; s4 -= carry4 * ((uint64_t) 1L << 21);
-    carry5 = s5 >> 21; s6 += carry5; s5 -= carry5 * ((uint64_t) 1L << 21);
-    carry6 = s6 >> 21; s7 += carry6; s6 -= carry6 * ((uint64_t) 1L << 21);
-    carry7 = s7 >> 21; s8 += carry7; s7 -= carry7 * ((uint64_t) 1L << 21);
-    carry8 = s8 >> 21; s9 += carry8; s8 -= carry8 * ((uint64_t) 1L << 21);
-    carry9 = s9 >> 21; s10 += carry9; s9 -= carry9 * ((uint64_t) 1L << 21);
-    carry10 = s10 >> 21; s11 += carry10; s10 -= carry10 * ((uint64_t) 1L << 21);
-
-    s[0] = s0 >> 0;
-    s[1] = s0 >> 8;
-    s[2] = (s0 >> 16) | (s1 * ((uint64_t) 1 << 5));
-    s[3] = s1 >> 3;
-    s[4] = s1 >> 11;
-    s[5] = (s1 >> 19) | (s2 * ((uint64_t) 1 << 2));
-    s[6] = s2 >> 6;
-    s[7] = (s2 >> 14) | (s3 * ((uint64_t) 1 << 7));
-    s[8] = s3 >> 1;
-    s[9] = s3 >> 9;
+    carry0 = s0 >> 21;
+    s1 += carry0;
+    s0 -= carry0 * ((uint64_t) 1L << 21);
+    carry1 = s1 >> 21;
+    s2 += carry1;
+    s1 -= carry1 * ((uint64_t) 1L << 21);
+    carry2 = s2 >> 21;
+    s3 += carry2;
+    s2 -= carry2 * ((uint64_t) 1L << 21);
+    carry3 = s3 >> 21;
+    s4 += carry3;
+    s3 -= carry3 * ((uint64_t) 1L << 21);
+    carry4 = s4 >> 21;
+    s5 += carry4;
+    s4 -= carry4 * ((uint64_t) 1L << 21);
+    carry5 = s5 >> 21;
+    s6 += carry5;
+    s5 -= carry5 * ((uint64_t) 1L << 21);
+    carry6 = s6 >> 21;
+    s7 += carry6;
+    s6 -= carry6 * ((uint64_t) 1L << 21);
+    carry7 = s7 >> 21;
+    s8 += carry7;
+    s7 -= carry7 * ((uint64_t) 1L << 21);
+    carry8 = s8 >> 21;
+    s9 += carry8;
+    s8 -= carry8 * ((uint64_t) 1L << 21);
+    carry9 = s9 >> 21;
+    s10 += carry9;
+    s9 -= carry9 * ((uint64_t) 1L << 21);
+    carry10 = s10 >> 21;
+    s11 += carry10;
+    s10 -= carry10 * ((uint64_t) 1L << 21);
+
+    s[0]  = s0 >> 0;
+    s[1]  = s0 >> 8;
+    s[2]  = (s0 >> 16) | (s1 * ((uint64_t) 1 << 5));
+    s[3]  = s1 >> 3;
+    s[4]  = s1 >> 11;
+    s[5]  = (s1 >> 19) | (s2 * ((uint64_t) 1 << 2));
+    s[6]  = s2 >> 6;
+    s[7]  = (s2 >> 14) | (s3 * ((uint64_t) 1 << 7));
+    s[8]  = s3 >> 1;
+    s[9]  = s3 >> 9;
     s[10] = (s3 >> 17) | (s4 * ((uint64_t) 1 << 4));
     s[11] = s4 >> 4;
     s[12] = s4 >> 12;
@@ -2011,18 +2411,19 @@ void sc_muladd(unsigned char *s,const unsigned char *a,const unsigned char *b,co
  Overwrites s in place.
  */
 
-void sc_reduce(unsigned char *s)
+void
+sc_reduce(unsigned char *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 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);
@@ -2037,6 +2438,7 @@ void sc_reduce(unsigned char *s)
     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 carry0;
     int64_t carry1;
     int64_t carry2;
@@ -2097,18 +2499,40 @@ void sc_reduce(unsigned char *s)
     s10 += s18 * 136657;
     s11 -= s18 * 683901;
 
-    carry6 = (s6 + (int64_t) (1L << 20)) >> 21; s7 += carry6; s6 -= carry6 * ((uint64_t) 1L << 21);
-    carry8 = (s8 + (int64_t) (1L << 20)) >> 21; s9 += carry8; s8 -= carry8 * ((uint64_t) 1L << 21);
-    carry10 = (s10 + (int64_t) (1L << 20)) >> 21; s11 += carry10; s10 -= carry10 * ((uint64_t) 1L << 21);
-    carry12 = (s12 + (int64_t) (1L << 20)) >> 21; s13 += carry12; s12 -= carry12 * ((uint64_t) 1L << 21);
-    carry14 = (s14 + (int64_t) (1L << 20)) >> 21; s15 += carry14; s14 -= carry14 * ((uint64_t) 1L << 21);
-    carry16 = (s16 + (int64_t) (1L << 20)) >> 21; s17 += carry16; s16 -= carry16 * ((uint64_t) 1L << 21);
-
-    carry7 = (s7 + (int64_t) (1L << 20)) >> 21; s8 += carry7; s7 -= carry7 * ((uint64_t) 1L << 21);
-    carry9 = (s9 + (int64_t) (1L << 20)) >> 21; s10 += carry9; s9 -= carry9 * ((uint64_t) 1L << 21);
-    carry11 = (s11 + (int64_t) (1L << 20)) >> 21; s12 += carry11; s11 -= carry11 * ((uint64_t) 1L << 21);
-    carry13 = (s13 + (int64_t) (1L << 20)) >> 21; s14 += carry13; s13 -= carry13 * ((uint64_t) 1L << 21);
-    carry15 = (s15 + (int64_t) (1L << 20)) >> 21; s16 += carry15; s15 -= carry15 * ((uint64_t) 1L << 21);
+    carry6 = (s6 + (int64_t)(1L << 20)) >> 21;
+    s7 += carry6;
+    s6 -= carry6 * ((uint64_t) 1L << 21);
+    carry8 = (s8 + (int64_t)(1L << 20)) >> 21;
+    s9 += carry8;
+    s8 -= carry8 * ((uint64_t) 1L << 21);
+    carry10 = (s10 + (int64_t)(1L << 20)) >> 21;
+    s11 += carry10;
+    s10 -= carry10 * ((uint64_t) 1L << 21);
+    carry12 = (s12 + (int64_t)(1L << 20)) >> 21;
+    s13 += carry12;
+    s12 -= carry12 * ((uint64_t) 1L << 21);
+    carry14 = (s14 + (int64_t)(1L << 20)) >> 21;
+    s15 += carry14;
+    s14 -= carry14 * ((uint64_t) 1L << 21);
+    carry16 = (s16 + (int64_t)(1L << 20)) >> 21;
+    s17 += carry16;
+    s16 -= carry16 * ((uint64_t) 1L << 21);
+
+    carry7 = (s7 + (int64_t)(1L << 20)) >> 21;
+    s8 += carry7;
+    s7 -= carry7 * ((uint64_t) 1L << 21);
+    carry9 = (s9 + (int64_t)(1L << 20)) >> 21;
+    s10 += carry9;
+    s9 -= carry9 * ((uint64_t) 1L << 21);
+    carry11 = (s11 + (int64_t)(1L << 20)) >> 21;
+    s12 += carry11;
+    s11 -= carry11 * ((uint64_t) 1L << 21);
+    carry13 = (s13 + (int64_t)(1L << 20)) >> 21;
+    s14 += carry13;
+    s13 -= carry13 * ((uint64_t) 1L << 21);
+    carry15 = (s15 + (int64_t)(1L << 20)) >> 21;
+    s16 += carry15;
+    s15 -= carry15 * ((uint64_t) 1L << 21);
 
     s5 += s17 * 666643;
     s6 += s17 * 470296;
@@ -2153,19 +2577,43 @@ void sc_reduce(unsigned char *s)
     s5 -= s12 * 683901;
     s12 = 0;
 
-    carry0 = (s0 + (int64_t) (1L << 20)) >> 21; s1 += carry0; s0 -= carry0 * ((uint64_t) 1L << 21);
-    carry2 = (s2 + (int64_t) (1L << 20)) >> 21; s3 += carry2; s2 -= carry2 * ((uint64_t) 1L << 21);
-    carry4 = (s4 + (int64_t) (1L << 20)) >> 21; s5 += carry4; s4 -= carry4 * ((uint64_t) 1L << 21);
-    carry6 = (s6 + (int64_t) (1L << 20)) >> 21; s7 += carry6; s6 -= carry6 * ((uint64_t) 1L << 21);
-    carry8 = (s8 + (int64_t) (1L << 20)) >> 21; s9 += carry8; s8 -= carry8 * ((uint64_t) 1L << 21);
-    carry10 = (s10 + (int64_t) (1L << 20)) >> 21; s11 += carry10; s10 -= carry10 * ((uint64_t) 1L << 21);
-
-    carry1 = (s1 + (int64_t) (1L << 20)) >> 21; s2 += carry1; s1 -= carry1 * ((uint64_t) 1L << 21);
-    carry3 = (s3 + (int64_t) (1L << 20)) >> 21; s4 += carry3; s3 -= carry3 * ((uint64_t) 1L << 21);
-    carry5 = (s5 + (int64_t) (1L << 20)) >> 21; s6 += carry5; s5 -= carry5 * ((uint64_t) 1L << 21);
-    carry7 = (s7 + (int64_t) (1L << 20)) >> 21; s8 += carry7; s7 -= carry7 * ((uint64_t) 1L << 21);
-    carry9 = (s9 + (int64_t) (1L << 20)) >> 21; s10 += carry9; s9 -= carry9 * ((uint64_t) 1L << 21);
-    carry11 = (s11 + (int64_t) (1L << 20)) >> 21; s12 += carry11; s11 -= carry11 * ((uint64_t) 1L << 21);
+    carry0 = (s0 + (int64_t)(1L << 20)) >> 21;
+    s1 += carry0;
+    s0 -= carry0 * ((uint64_t) 1L << 21);
+    carry2 = (s2 + (int64_t)(1L << 20)) >> 21;
+    s3 += carry2;
+    s2 -= carry2 * ((uint64_t) 1L << 21);
+    carry4 = (s4 + (int64_t)(1L << 20)) >> 21;
+    s5 += carry4;
+    s4 -= carry4 * ((uint64_t) 1L << 21);
+    carry6 = (s6 + (int64_t)(1L << 20)) >> 21;
+    s7 += carry6;
+    s6 -= carry6 * ((uint64_t) 1L << 21);
+    carry8 = (s8 + (int64_t)(1L << 20)) >> 21;
+    s9 += carry8;
+    s8 -= carry8 * ((uint64_t) 1L << 21);
+    carry10 = (s10 + (int64_t)(1L << 20)) >> 21;
+    s11 += carry10;
+    s10 -= carry10 * ((uint64_t) 1L << 21);
+
+    carry1 = (s1 + (int64_t)(1L << 20)) >> 21;
+    s2 += carry1;
+    s1 -= carry1 * ((uint64_t) 1L << 21);
+    carry3 = (s3 + (int64_t)(1L << 20)) >> 21;
+    s4 += carry3;
+    s3 -= carry3 * ((uint64_t) 1L << 21);
+    carry5 = (s5 + (int64_t)(1L << 20)) >> 21;
+    s6 += carry5;
+    s5 -= carry5 * ((uint64_t) 1L << 21);
+    carry7 = (s7 + (int64_t)(1L << 20)) >> 21;
+    s8 += carry7;
+    s7 -= carry7 * ((uint64_t) 1L << 21);
+    carry9 = (s9 + (int64_t)(1L << 20)) >> 21;
+    s10 += carry9;
+    s9 -= carry9 * ((uint64_t) 1L << 21);
+    carry11 = (s11 + (int64_t)(1L << 20)) >> 21;
+    s12 += carry11;
+    s11 -= carry11 * ((uint64_t) 1L << 21);
 
     s0 += s12 * 666643;
     s1 += s12 * 470296;
@@ -2175,18 +2623,42 @@ void sc_reduce(unsigned char *s)
     s5 -= s12 * 683901;
     s12 = 0;
 
-    carry0 = s0 >> 21; s1 += carry0; s0 -= carry0 * ((uint64_t) 1L << 21);
-    carry1 = s1 >> 21; s2 += carry1; s1 -= carry1 * ((uint64_t) 1L << 21);
-    carry2 = s2 >> 21; s3 += carry2; s2 -= carry2 * ((uint64_t) 1L << 21);
-    carry3 = s3 >> 21; s4 += carry3; s3 -= carry3 * ((uint64_t) 1L << 21);
-    carry4 = s4 >> 21; s5 += carry4; s4 -= carry4 * ((uint64_t) 1L << 21);
-    carry5 = s5 >> 21; s6 += carry5; s5 -= carry5 * ((uint64_t) 1L << 21);
-    carry6 = s6 >> 21; s7 += carry6; s6 -= carry6 * ((uint64_t) 1L << 21);
-    carry7 = s7 >> 21; s8 += carry7; s7 -= carry7 * ((uint64_t) 1L << 21);
-    carry8 = s8 >> 21; s9 += carry8; s8 -= carry8 * ((uint64_t) 1L << 21);
-    carry9 = s9 >> 21; s10 += carry9; s9 -= carry9 * ((uint64_t) 1L << 21);
-    carry10 = s10 >> 21; s11 += carry10; s10 -= carry10 * ((uint64_t) 1L << 21);
-    carry11 = s11 >> 21; s12 += carry11; s11 -= carry11 * ((uint64_t) 1L << 21);
+    carry0 = s0 >> 21;
+    s1 += carry0;
+    s0 -= carry0 * ((uint64_t) 1L << 21);
+    carry1 = s1 >> 21;
+    s2 += carry1;
+    s1 -= carry1 * ((uint64_t) 1L << 21);
+    carry2 = s2 >> 21;
+    s3 += carry2;
+    s2 -= carry2 * ((uint64_t) 1L << 21);
+    carry3 = s3 >> 21;
+    s4 += carry3;
+    s3 -= carry3 * ((uint64_t) 1L << 21);
+    carry4 = s4 >> 21;
+    s5 += carry4;
+    s4 -= carry4 * ((uint64_t) 1L << 21);
+    carry5 = s5 >> 21;
+    s6 += carry5;
+    s5 -= carry5 * ((uint64_t) 1L << 21);
+    carry6 = s6 >> 21;
+    s7 += carry6;
+    s6 -= carry6 * ((uint64_t) 1L << 21);
+    carry7 = s7 >> 21;
+    s8 += carry7;
+    s7 -= carry7 * ((uint64_t) 1L << 21);
+    carry8 = s8 >> 21;
+    s9 += carry8;
+    s8 -= carry8 * ((uint64_t) 1L << 21);
+    carry9 = s9 >> 21;
+    s10 += carry9;
+    s9 -= carry9 * ((uint64_t) 1L << 21);
+    carry10 = s10 >> 21;
+    s11 += carry10;
+    s10 -= carry10 * ((uint64_t) 1L << 21);
+    carry11 = s11 >> 21;
+    s12 += carry11;
+    s11 -= carry11 * ((uint64_t) 1L << 21);
 
     s0 += s12 * 666643;
     s1 += s12 * 470296;
@@ -2195,28 +2667,50 @@ void sc_reduce(unsigned char *s)
     s4 += s12 * 136657;
     s5 -= s12 * 683901;
 
-    carry0 = s0 >> 21; s1 += carry0; s0 -= carry0 * ((uint64_t) 1L << 21);
-    carry1 = s1 >> 21; s2 += carry1; s1 -= carry1 * ((uint64_t) 1L << 21);
-    carry2 = s2 >> 21; s3 += carry2; s2 -= carry2 * ((uint64_t) 1L << 21);
-    carry3 = s3 >> 21; s4 += carry3; s3 -= carry3 * ((uint64_t) 1L << 21);
-    carry4 = s4 >> 21; s5 += carry4; s4 -= carry4 * ((uint64_t) 1L << 21);
-    carry5 = s5 >> 21; s6 += carry5; s5 -= carry5 * ((uint64_t) 1L << 21);
-    carry6 = s6 >> 21; s7 += carry6; s6 -= carry6 * ((uint64_t) 1L << 21);
-    carry7 = s7 >> 21; s8 += carry7; s7 -= carry7 * ((uint64_t) 1L << 21);
-    carry8 = s8 >> 21; s9 += carry8; s8 -= carry8 * ((uint64_t) 1L << 21);
-    carry9 = s9 >> 21; s10 += carry9; s9 -= carry9 * ((uint64_t) 1L << 21);
-    carry10 = s10 >> 21; s11 += carry10; s10 -= carry10 * ((uint64_t) 1L << 21);
-
-    s[0] = s0 >> 0;
-    s[1] = s0 >> 8;
-    s[2] = (s0 >> 16) | (s1 * ((uint64_t) 1 << 5));
-    s[3] = s1 >> 3;
-    s[4] = s1 >> 11;
-    s[5] = (s1 >> 19) | (s2 * ((uint64_t) 1 << 2));
-    s[6] = s2 >> 6;
-    s[7] = (s2 >> 14) | (s3 * ((uint64_t) 1 << 7));
-    s[8] = s3 >> 1;
-    s[9] = s3 >> 9;
+    carry0 = s0 >> 21;
+    s1 += carry0;
+    s0 -= carry0 * ((uint64_t) 1L << 21);
+    carry1 = s1 >> 21;
+    s2 += carry1;
+    s1 -= carry1 * ((uint64_t) 1L << 21);
+    carry2 = s2 >> 21;
+    s3 += carry2;
+    s2 -= carry2 * ((uint64_t) 1L << 21);
+    carry3 = s3 >> 21;
+    s4 += carry3;
+    s3 -= carry3 * ((uint64_t) 1L << 21);
+    carry4 = s4 >> 21;
+    s5 += carry4;
+    s4 -= carry4 * ((uint64_t) 1L << 21);
+    carry5 = s5 >> 21;
+    s6 += carry5;
+    s5 -= carry5 * ((uint64_t) 1L << 21);
+    carry6 = s6 >> 21;
+    s7 += carry6;
+    s6 -= carry6 * ((uint64_t) 1L << 21);
+    carry7 = s7 >> 21;
+    s8 += carry7;
+    s7 -= carry7 * ((uint64_t) 1L << 21);
+    carry8 = s8 >> 21;
+    s9 += carry8;
+    s8 -= carry8 * ((uint64_t) 1L << 21);
+    carry9 = s9 >> 21;
+    s10 += carry9;
+    s9 -= carry9 * ((uint64_t) 1L << 21);
+    carry10 = s10 >> 21;
+    s11 += carry10;
+    s10 -= carry10 * ((uint64_t) 1L << 21);
+
+    s[0]  = s0 >> 0;
+    s[1]  = s0 >> 8;
+    s[2]  = (s0 >> 16) | (s1 * ((uint64_t) 1 << 5));
+    s[3]  = s1 >> 3;
+    s[4]  = s1 >> 11;
+    s[5]  = (s1 >> 19) | (s2 * ((uint64_t) 1 << 2));
+    s[6]  = s2 >> 6;
+    s[7]  = (s2 >> 14) | (s3 * ((uint64_t) 1 << 7));
+    s[8]  = s3 >> 1;
+    s[9]  = s3 >> 9;
     s[10] = (s3 >> 17) | (s4 * ((uint64_t) 1 << 4));
     s[11] = s4 >> 4;
     s[12] = s4 >> 12;