rbnacl-libsodium 0.5.0.1 → 0.6.0

data/vendor/libsodium/src/libsodium/crypto_scalarmult/curve25519/ref10/fe_frombytes_curve25519_ref10.c

@@ -0,0 +1,73 @@
+#include "fe.h"
+#include "crypto_int64.h"
+#include "crypto_uint64.h"
+
+#ifndef HAVE_TI_MODE
+
+static crypto_uint64 load_3(const unsigned char *in)
+{
+  crypto_uint64 result;
+  result = (crypto_uint64) in[0];
+  result |= ((crypto_uint64) in[1]) << 8;
+  result |= ((crypto_uint64) in[2]) << 16;
+  return result;
+}
+
+static crypto_uint64 load_4(const unsigned char *in)
+{
+  crypto_uint64 result;
+  result = (crypto_uint64) in[0];
+  result |= ((crypto_uint64) in[1]) << 8;
+  result |= ((crypto_uint64) in[2]) << 16;
+  result |= ((crypto_uint64) in[3]) << 24;
+  return result;
+}
+
+void fe_frombytes(fe h,const unsigned char *s)
+{
+  crypto_int64 h0 = load_4(s);
+  crypto_int64 h1 = load_3(s + 4) << 6;
+  crypto_int64 h2 = load_3(s + 7) << 5;
+  crypto_int64 h3 = load_3(s + 10) << 3;
+  crypto_int64 h4 = load_3(s + 13) << 2;
+  crypto_int64 h5 = load_4(s + 16);
+  crypto_int64 h6 = load_3(s + 20) << 7;
+  crypto_int64 h7 = load_3(s + 23) << 5;
+  crypto_int64 h8 = load_3(s + 26) << 4;
+  crypto_int64 h9 = (load_3(s + 29) & 8388607) << 2;
+  crypto_int64 carry0;
+  crypto_int64 carry1;
+  crypto_int64 carry2;
+  crypto_int64 carry3;
+  crypto_int64 carry4;
+  crypto_int64 carry5;
+  crypto_int64 carry6;
+  crypto_int64 carry7;
+  crypto_int64 carry8;
+  crypto_int64 carry9;
+
+  carry9 = (h9 + (crypto_int64) (1<<24)) >> 25; h0 += carry9 * 19; h9 -= carry9 << 25;
+  carry1 = (h1 + (crypto_int64) (1<<24)) >> 25; h2 += carry1; h1 -= carry1 << 25;
+  carry3 = (h3 + (crypto_int64) (1<<24)) >> 25; h4 += carry3; h3 -= carry3 << 25;
+  carry5 = (h5 + (crypto_int64) (1<<24)) >> 25; h6 += carry5; h5 -= carry5 << 25;
+  carry7 = (h7 + (crypto_int64) (1<<24)) >> 25; h8 += carry7; h7 -= carry7 << 25;
+
+  carry0 = (h0 + (crypto_int64) (1<<25)) >> 26; h1 += carry0; h0 -= carry0 << 26;
+  carry2 = (h2 + (crypto_int64) (1<<25)) >> 26; h3 += carry2; h2 -= carry2 << 26;
+  carry4 = (h4 + (crypto_int64) (1<<25)) >> 26; h5 += carry4; h4 -= carry4 << 26;
+  carry6 = (h6 + (crypto_int64) (1<<25)) >> 26; h7 += carry6; h6 -= carry6 << 26;
+  carry8 = (h8 + (crypto_int64) (1<<25)) >> 26; h9 += carry8; h8 -= carry8 << 26;
+
+  h[0] = h0;
+  h[1] = h1;
+  h[2] = h2;
+  h[3] = h3;
+  h[4] = h4;
+  h[5] = h5;
+  h[6] = h6;
+  h[7] = h7;
+  h[8] = h8;
+  h[9] = h9;
+}
+
+#endif

data/vendor/libsodium/src/libsodium/crypto_scalarmult/curve25519/ref10/fe_invert_curve25519_ref10.c

@@ -0,0 +1,18 @@
+#include "fe.h"
+
+#ifndef HAVE_TI_MODE
+
+void fe_invert(fe out,fe z)
+{
+  fe t0;
+  fe t1;
+  fe t2;
+  fe t3;
+  int i;
+
+#include "pow225521.h"
+
+  return;
+}
+
+#endif

data/vendor/libsodium/src/libsodium/crypto_scalarmult/curve25519/ref10/fe_mul121666_curve25519_ref10.c

@@ -0,0 +1,74 @@
+#include "fe.h"
+#include "crypto_int64.h"
+
+#ifndef HAVE_TI_MODE
+
+/*
+h = f * 121666
+Can overlap h with f.
+
+Preconditions:
+   |f| bounded by 1.1*2^26,1.1*2^25,1.1*2^26,1.1*2^25,etc.
+
+Postconditions:
+   |h| bounded by 1.1*2^25,1.1*2^24,1.1*2^25,1.1*2^24,etc.
+*/
+
+void fe_mul121666(fe h,fe f)
+{
+  crypto_int32 f0 = f[0];
+  crypto_int32 f1 = f[1];
+  crypto_int32 f2 = f[2];
+  crypto_int32 f3 = f[3];
+  crypto_int32 f4 = f[4];
+  crypto_int32 f5 = f[5];
+  crypto_int32 f6 = f[6];
+  crypto_int32 f7 = f[7];
+  crypto_int32 f8 = f[8];
+  crypto_int32 f9 = f[9];
+  crypto_int64 h0 = f0 * (crypto_int64) 121666;
+  crypto_int64 h1 = f1 * (crypto_int64) 121666;
+  crypto_int64 h2 = f2 * (crypto_int64) 121666;
+  crypto_int64 h3 = f3 * (crypto_int64) 121666;
+  crypto_int64 h4 = f4 * (crypto_int64) 121666;
+  crypto_int64 h5 = f5 * (crypto_int64) 121666;
+  crypto_int64 h6 = f6 * (crypto_int64) 121666;
+  crypto_int64 h7 = f7 * (crypto_int64) 121666;
+  crypto_int64 h8 = f8 * (crypto_int64) 121666;
+  crypto_int64 h9 = f9 * (crypto_int64) 121666;
+  crypto_int64 carry0;
+  crypto_int64 carry1;
+  crypto_int64 carry2;
+  crypto_int64 carry3;
+  crypto_int64 carry4;
+  crypto_int64 carry5;
+  crypto_int64 carry6;
+  crypto_int64 carry7;
+  crypto_int64 carry8;
+  crypto_int64 carry9;
+
+  carry9 = (h9 + (crypto_int64) (1<<24)) >> 25; h0 += carry9 * 19; h9 -= carry9 << 25;
+  carry1 = (h1 + (crypto_int64) (1<<24)) >> 25; h2 += carry1; h1 -= carry1 << 25;
+  carry3 = (h3 + (crypto_int64) (1<<24)) >> 25; h4 += carry3; h3 -= carry3 << 25;
+  carry5 = (h5 + (crypto_int64) (1<<24)) >> 25; h6 += carry5; h5 -= carry5 << 25;
+  carry7 = (h7 + (crypto_int64) (1<<24)) >> 25; h8 += carry7; h7 -= carry7 << 25;
+
+  carry0 = (h0 + (crypto_int64) (1<<25)) >> 26; h1 += carry0; h0 -= carry0 << 26;
+  carry2 = (h2 + (crypto_int64) (1<<25)) >> 26; h3 += carry2; h2 -= carry2 << 26;
+  carry4 = (h4 + (crypto_int64) (1<<25)) >> 26; h5 += carry4; h4 -= carry4 << 26;
+  carry6 = (h6 + (crypto_int64) (1<<25)) >> 26; h7 += carry6; h6 -= carry6 << 26;
+  carry8 = (h8 + (crypto_int64) (1<<25)) >> 26; h9 += carry8; h8 -= carry8 << 26;
+
+  h[0] = h0;
+  h[1] = h1;
+  h[2] = h2;
+  h[3] = h3;
+  h[4] = h4;
+  h[5] = h5;
+  h[6] = h6;
+  h[7] = h7;
+  h[8] = h8;
+  h[9] = h9;
+}
+
+#endif

data/vendor/libsodium/src/libsodium/crypto_scalarmult/curve25519/ref10/fe_mul_curve25519_ref10.c

@@ -0,0 +1,257 @@
+#include "fe.h"
+#include "crypto_int64.h"
+
+#ifndef HAVE_TI_MODE
+
+/*
+h = f * g
+Can overlap h with f or g.
+
+Preconditions:
+   |f| bounded by 1.1*2^26,1.1*2^25,1.1*2^26,1.1*2^25,etc.
+   |g| bounded by 1.1*2^26,1.1*2^25,1.1*2^26,1.1*2^25,etc.
+
+Postconditions:
+   |h| bounded by 1.1*2^25,1.1*2^24,1.1*2^25,1.1*2^24,etc.
+*/
+
+/*
+Notes on implementation strategy:
+
+Using schoolbook multiplication.
+Karatsuba would save a little in some cost models.
+
+Most multiplications by 2 and 19 are 32-bit precomputations;
+cheaper than 64-bit postcomputations.
+
+There is one remaining multiplication by 19 in the carry chain;
+one *19 precomputation can be merged into this,
+but the resulting data flow is considerably less clean.
+
+There are 12 carries below.
+10 of them are 2-way parallelizable and vectorizable.
+Can get away with 11 carries, but then data flow is much deeper.
+
+With tighter constraints on inputs can squeeze carries into int32.
+*/
+
+void fe_mul(fe h,fe f,fe g)
+{
+  crypto_int32 f0 = f[0];
+  crypto_int32 f1 = f[1];
+  crypto_int32 f2 = f[2];
+  crypto_int32 f3 = f[3];
+  crypto_int32 f4 = f[4];
+  crypto_int32 f5 = f[5];
+  crypto_int32 f6 = f[6];
+  crypto_int32 f7 = f[7];
+  crypto_int32 f8 = f[8];
+  crypto_int32 f9 = f[9];
+  crypto_int32 g0 = g[0];
+  crypto_int32 g1 = g[1];
+  crypto_int32 g2 = g[2];
+  crypto_int32 g3 = g[3];
+  crypto_int32 g4 = g[4];
+  crypto_int32 g5 = g[5];
+  crypto_int32 g6 = g[6];
+  crypto_int32 g7 = g[7];
+  crypto_int32 g8 = g[8];
+  crypto_int32 g9 = g[9];
+  crypto_int32 g1_19 = 19 * g1; /* 1.4*2^29 */
+  crypto_int32 g2_19 = 19 * g2; /* 1.4*2^30; still ok */
+  crypto_int32 g3_19 = 19 * g3;
+  crypto_int32 g4_19 = 19 * g4;
+  crypto_int32 g5_19 = 19 * g5;
+  crypto_int32 g6_19 = 19 * g6;
+  crypto_int32 g7_19 = 19 * g7;
+  crypto_int32 g8_19 = 19 * g8;
+  crypto_int32 g9_19 = 19 * g9;
+  crypto_int32 f1_2 = 2 * f1;
+  crypto_int32 f3_2 = 2 * f3;
+  crypto_int32 f5_2 = 2 * f5;
+  crypto_int32 f7_2 = 2 * f7;
+  crypto_int32 f9_2 = 2 * f9;
+  crypto_int64 f0g0    = f0   * (crypto_int64) g0;
+  crypto_int64 f0g1    = f0   * (crypto_int64) g1;
+  crypto_int64 f0g2    = f0   * (crypto_int64) g2;
+  crypto_int64 f0g3    = f0   * (crypto_int64) g3;
+  crypto_int64 f0g4    = f0   * (crypto_int64) g4;
+  crypto_int64 f0g5    = f0   * (crypto_int64) g5;
+  crypto_int64 f0g6    = f0   * (crypto_int64) g6;
+  crypto_int64 f0g7    = f0   * (crypto_int64) g7;
+  crypto_int64 f0g8    = f0   * (crypto_int64) g8;
+  crypto_int64 f0g9    = f0   * (crypto_int64) g9;
+  crypto_int64 f1g0    = f1   * (crypto_int64) g0;
+  crypto_int64 f1g1_2  = f1_2 * (crypto_int64) g1;
+  crypto_int64 f1g2    = f1   * (crypto_int64) g2;
+  crypto_int64 f1g3_2  = f1_2 * (crypto_int64) g3;
+  crypto_int64 f1g4    = f1   * (crypto_int64) g4;
+  crypto_int64 f1g5_2  = f1_2 * (crypto_int64) g5;
+  crypto_int64 f1g6    = f1   * (crypto_int64) g6;
+  crypto_int64 f1g7_2  = f1_2 * (crypto_int64) g7;
+  crypto_int64 f1g8    = f1   * (crypto_int64) g8;
+  crypto_int64 f1g9_38 = f1_2 * (crypto_int64) g9_19;
+  crypto_int64 f2g0    = f2   * (crypto_int64) g0;
+  crypto_int64 f2g1    = f2   * (crypto_int64) g1;
+  crypto_int64 f2g2    = f2   * (crypto_int64) g2;
+  crypto_int64 f2g3    = f2   * (crypto_int64) g3;
+  crypto_int64 f2g4    = f2   * (crypto_int64) g4;
+  crypto_int64 f2g5    = f2   * (crypto_int64) g5;
+  crypto_int64 f2g6    = f2   * (crypto_int64) g6;
+  crypto_int64 f2g7    = f2   * (crypto_int64) g7;
+  crypto_int64 f2g8_19 = f2   * (crypto_int64) g8_19;
+  crypto_int64 f2g9_19 = f2   * (crypto_int64) g9_19;
+  crypto_int64 f3g0    = f3   * (crypto_int64) g0;
+  crypto_int64 f3g1_2  = f3_2 * (crypto_int64) g1;
+  crypto_int64 f3g2    = f3   * (crypto_int64) g2;
+  crypto_int64 f3g3_2  = f3_2 * (crypto_int64) g3;
+  crypto_int64 f3g4    = f3   * (crypto_int64) g4;
+  crypto_int64 f3g5_2  = f3_2 * (crypto_int64) g5;
+  crypto_int64 f3g6    = f3   * (crypto_int64) g6;
+  crypto_int64 f3g7_38 = f3_2 * (crypto_int64) g7_19;
+  crypto_int64 f3g8_19 = f3   * (crypto_int64) g8_19;
+  crypto_int64 f3g9_38 = f3_2 * (crypto_int64) g9_19;
+  crypto_int64 f4g0    = f4   * (crypto_int64) g0;
+  crypto_int64 f4g1    = f4   * (crypto_int64) g1;
+  crypto_int64 f4g2    = f4   * (crypto_int64) g2;
+  crypto_int64 f4g3    = f4   * (crypto_int64) g3;
+  crypto_int64 f4g4    = f4   * (crypto_int64) g4;
+  crypto_int64 f4g5    = f4   * (crypto_int64) g5;
+  crypto_int64 f4g6_19 = f4   * (crypto_int64) g6_19;
+  crypto_int64 f4g7_19 = f4   * (crypto_int64) g7_19;
+  crypto_int64 f4g8_19 = f4   * (crypto_int64) g8_19;
+  crypto_int64 f4g9_19 = f4   * (crypto_int64) g9_19;
+  crypto_int64 f5g0    = f5   * (crypto_int64) g0;
+  crypto_int64 f5g1_2  = f5_2 * (crypto_int64) g1;
+  crypto_int64 f5g2    = f5   * (crypto_int64) g2;
+  crypto_int64 f5g3_2  = f5_2 * (crypto_int64) g3;
+  crypto_int64 f5g4    = f5   * (crypto_int64) g4;
+  crypto_int64 f5g5_38 = f5_2 * (crypto_int64) g5_19;
+  crypto_int64 f5g6_19 = f5   * (crypto_int64) g6_19;
+  crypto_int64 f5g7_38 = f5_2 * (crypto_int64) g7_19;
+  crypto_int64 f5g8_19 = f5   * (crypto_int64) g8_19;
+  crypto_int64 f5g9_38 = f5_2 * (crypto_int64) g9_19;
+  crypto_int64 f6g0    = f6   * (crypto_int64) g0;
+  crypto_int64 f6g1    = f6   * (crypto_int64) g1;
+  crypto_int64 f6g2    = f6   * (crypto_int64) g2;
+  crypto_int64 f6g3    = f6   * (crypto_int64) g3;
+  crypto_int64 f6g4_19 = f6   * (crypto_int64) g4_19;
+  crypto_int64 f6g5_19 = f6   * (crypto_int64) g5_19;
+  crypto_int64 f6g6_19 = f6   * (crypto_int64) g6_19;
+  crypto_int64 f6g7_19 = f6   * (crypto_int64) g7_19;
+  crypto_int64 f6g8_19 = f6   * (crypto_int64) g8_19;
+  crypto_int64 f6g9_19 = f6   * (crypto_int64) g9_19;
+  crypto_int64 f7g0    = f7   * (crypto_int64) g0;
+  crypto_int64 f7g1_2  = f7_2 * (crypto_int64) g1;
+  crypto_int64 f7g2    = f7   * (crypto_int64) g2;
+  crypto_int64 f7g3_38 = f7_2 * (crypto_int64) g3_19;
+  crypto_int64 f7g4_19 = f7   * (crypto_int64) g4_19;
+  crypto_int64 f7g5_38 = f7_2 * (crypto_int64) g5_19;
+  crypto_int64 f7g6_19 = f7   * (crypto_int64) g6_19;
+  crypto_int64 f7g7_38 = f7_2 * (crypto_int64) g7_19;
+  crypto_int64 f7g8_19 = f7   * (crypto_int64) g8_19;
+  crypto_int64 f7g9_38 = f7_2 * (crypto_int64) g9_19;
+  crypto_int64 f8g0    = f8   * (crypto_int64) g0;
+  crypto_int64 f8g1    = f8   * (crypto_int64) g1;
+  crypto_int64 f8g2_19 = f8   * (crypto_int64) g2_19;
+  crypto_int64 f8g3_19 = f8   * (crypto_int64) g3_19;
+  crypto_int64 f8g4_19 = f8   * (crypto_int64) g4_19;
+  crypto_int64 f8g5_19 = f8   * (crypto_int64) g5_19;
+  crypto_int64 f8g6_19 = f8   * (crypto_int64) g6_19;
+  crypto_int64 f8g7_19 = f8   * (crypto_int64) g7_19;
+  crypto_int64 f8g8_19 = f8   * (crypto_int64) g8_19;
+  crypto_int64 f8g9_19 = f8   * (crypto_int64) g9_19;
+  crypto_int64 f9g0    = f9   * (crypto_int64) g0;
+  crypto_int64 f9g1_38 = f9_2 * (crypto_int64) g1_19;
+  crypto_int64 f9g2_19 = f9   * (crypto_int64) g2_19;
+  crypto_int64 f9g3_38 = f9_2 * (crypto_int64) g3_19;
+  crypto_int64 f9g4_19 = f9   * (crypto_int64) g4_19;
+  crypto_int64 f9g5_38 = f9_2 * (crypto_int64) g5_19;
+  crypto_int64 f9g6_19 = f9   * (crypto_int64) g6_19;
+  crypto_int64 f9g7_38 = f9_2 * (crypto_int64) g7_19;
+  crypto_int64 f9g8_19 = f9   * (crypto_int64) g8_19;
+  crypto_int64 f9g9_38 = f9_2 * (crypto_int64) g9_19;
+  crypto_int64 h0 = f0g0+f1g9_38+f2g8_19+f3g7_38+f4g6_19+f5g5_38+f6g4_19+f7g3_38+f8g2_19+f9g1_38;
+  crypto_int64 h1 = f0g1+f1g0   +f2g9_19+f3g8_19+f4g7_19+f5g6_19+f6g5_19+f7g4_19+f8g3_19+f9g2_19;
+  crypto_int64 h2 = f0g2+f1g1_2 +f2g0   +f3g9_38+f4g8_19+f5g7_38+f6g6_19+f7g5_38+f8g4_19+f9g3_38;
+  crypto_int64 h3 = f0g3+f1g2   +f2g1   +f3g0   +f4g9_19+f5g8_19+f6g7_19+f7g6_19+f8g5_19+f9g4_19;
+  crypto_int64 h4 = f0g4+f1g3_2 +f2g2   +f3g1_2 +f4g0   +f5g9_38+f6g8_19+f7g7_38+f8g6_19+f9g5_38;
+  crypto_int64 h5 = f0g5+f1g4   +f2g3   +f3g2   +f4g1   +f5g0   +f6g9_19+f7g8_19+f8g7_19+f9g6_19;
+  crypto_int64 h6 = f0g6+f1g5_2 +f2g4   +f3g3_2 +f4g2   +f5g1_2 +f6g0   +f7g9_38+f8g8_19+f9g7_38;
+  crypto_int64 h7 = f0g7+f1g6   +f2g5   +f3g4   +f4g3   +f5g2   +f6g1   +f7g0   +f8g9_19+f9g8_19;
+  crypto_int64 h8 = f0g8+f1g7_2 +f2g6   +f3g5_2 +f4g4   +f5g3_2 +f6g2   +f7g1_2 +f8g0   +f9g9_38;
+  crypto_int64 h9 = f0g9+f1g8   +f2g7   +f3g6   +f4g5   +f5g4   +f6g3   +f7g2   +f8g1   +f9g0   ;
+  crypto_int64 carry0;
+  crypto_int64 carry1;
+  crypto_int64 carry2;
+  crypto_int64 carry3;
+  crypto_int64 carry4;
+  crypto_int64 carry5;
+  crypto_int64 carry6;
+  crypto_int64 carry7;
+  crypto_int64 carry8;
+  crypto_int64 carry9;
+
+  /*
+  |h0| <= (1.1*1.1*2^52*(1+19+19+19+19)+1.1*1.1*2^50*(38+38+38+38+38))
+    i.e. |h0| <= 1.2*2^59; narrower ranges for h2, h4, h6, h8
+  |h1| <= (1.1*1.1*2^51*(1+1+19+19+19+19+19+19+19+19))
+    i.e. |h1| <= 1.5*2^58; narrower ranges for h3, h5, h7, h9
+  */
+
+  carry0 = (h0 + (crypto_int64) (1<<25)) >> 26; h1 += carry0; h0 -= carry0 << 26;
+  carry4 = (h4 + (crypto_int64) (1<<25)) >> 26; h5 += carry4; h4 -= carry4 << 26;
+  /* |h0| <= 2^25 */
+  /* |h4| <= 2^25 */
+  /* |h1| <= 1.51*2^58 */
+  /* |h5| <= 1.51*2^58 */
+
+  carry1 = (h1 + (crypto_int64) (1<<24)) >> 25; h2 += carry1; h1 -= carry1 << 25;
+  carry5 = (h5 + (crypto_int64) (1<<24)) >> 25; h6 += carry5; h5 -= carry5 << 25;
+  /* |h1| <= 2^24; from now on fits into int32 */
+  /* |h5| <= 2^24; from now on fits into int32 */
+  /* |h2| <= 1.21*2^59 */
+  /* |h6| <= 1.21*2^59 */
+
+  carry2 = (h2 + (crypto_int64) (1<<25)) >> 26; h3 += carry2; h2 -= carry2 << 26;
+  carry6 = (h6 + (crypto_int64) (1<<25)) >> 26; h7 += carry6; h6 -= carry6 << 26;
+  /* |h2| <= 2^25; from now on fits into int32 unchanged */
+  /* |h6| <= 2^25; from now on fits into int32 unchanged */
+  /* |h3| <= 1.51*2^58 */
+  /* |h7| <= 1.51*2^58 */
+
+  carry3 = (h3 + (crypto_int64) (1<<24)) >> 25; h4 += carry3; h3 -= carry3 << 25;
+  carry7 = (h7 + (crypto_int64) (1<<24)) >> 25; h8 += carry7; h7 -= carry7 << 25;
+  /* |h3| <= 2^24; from now on fits into int32 unchanged */
+  /* |h7| <= 2^24; from now on fits into int32 unchanged */
+  /* |h4| <= 1.52*2^33 */
+  /* |h8| <= 1.52*2^33 */
+
+  carry4 = (h4 + (crypto_int64) (1<<25)) >> 26; h5 += carry4; h4 -= carry4 << 26;
+  carry8 = (h8 + (crypto_int64) (1<<25)) >> 26; h9 += carry8; h8 -= carry8 << 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| <= 1.51*2^58 */
+
+  carry9 = (h9 + (crypto_int64) (1<<24)) >> 25; h0 += carry9 * 19; h9 -= carry9 << 25;
+  /* |h9| <= 2^24; from now on fits into int32 unchanged */
+  /* |h0| <= 1.8*2^37 */
+
+  carry0 = (h0 + (crypto_int64) (1<<25)) >> 26; h1 += carry0; h0 -= carry0 << 26;
+  /* |h0| <= 2^25; from now on fits into int32 unchanged */
+  /* |h1| <= 1.01*2^24 */
+
+  h[0] = h0;
+  h[1] = h1;
+  h[2] = h2;
+  h[3] = h3;
+  h[4] = h4;
+  h[5] = h5;
+  h[6] = h6;
+  h[7] = h7;
+  h[8] = h8;
+  h[9] = h9;
+}
+
+#endif

data/vendor/libsodium/src/libsodium/crypto_scalarmult/curve25519/ref10/fe_sq_curve25519_ref10.c

@@ -0,0 +1,153 @@
+#include "fe.h"
+#include "crypto_int64.h"
+
+#ifndef HAVE_TI_MODE
+
+/*
+h = f * f
+Can overlap h with f.
+
+Preconditions:
+   |f| bounded by 1.1*2^26,1.1*2^25,1.1*2^26,1.1*2^25,etc.
+
+Postconditions:
+   |h| bounded by 1.1*2^25,1.1*2^24,1.1*2^25,1.1*2^24,etc.
+*/
+
+/*
+See fe_mul.c for discussion of implementation strategy.
+*/
+
+void fe_sq(fe h,fe f)
+{
+  crypto_int32 f0 = f[0];
+  crypto_int32 f1 = f[1];
+  crypto_int32 f2 = f[2];
+  crypto_int32 f3 = f[3];
+  crypto_int32 f4 = f[4];
+  crypto_int32 f5 = f[5];
+  crypto_int32 f6 = f[6];
+  crypto_int32 f7 = f[7];
+  crypto_int32 f8 = f[8];
+  crypto_int32 f9 = f[9];
+  crypto_int32 f0_2 = 2 * f0;
+  crypto_int32 f1_2 = 2 * f1;
+  crypto_int32 f2_2 = 2 * f2;
+  crypto_int32 f3_2 = 2 * f3;
+  crypto_int32 f4_2 = 2 * f4;
+  crypto_int32 f5_2 = 2 * f5;
+  crypto_int32 f6_2 = 2 * f6;
+  crypto_int32 f7_2 = 2 * f7;
+  crypto_int32 f5_38 = 38 * f5; /* 1.31*2^30 */
+  crypto_int32 f6_19 = 19 * f6; /* 1.31*2^30 */
+  crypto_int32 f7_38 = 38 * f7; /* 1.31*2^30 */
+  crypto_int32 f8_19 = 19 * f8; /* 1.31*2^30 */
+  crypto_int32 f9_38 = 38 * f9; /* 1.31*2^30 */
+  crypto_int64 f0f0    = f0   * (crypto_int64) f0;
+  crypto_int64 f0f1_2  = f0_2 * (crypto_int64) f1;
+  crypto_int64 f0f2_2  = f0_2 * (crypto_int64) f2;
+  crypto_int64 f0f3_2  = f0_2 * (crypto_int64) f3;
+  crypto_int64 f0f4_2  = f0_2 * (crypto_int64) f4;
+  crypto_int64 f0f5_2  = f0_2 * (crypto_int64) f5;
+  crypto_int64 f0f6_2  = f0_2 * (crypto_int64) f6;
+  crypto_int64 f0f7_2  = f0_2 * (crypto_int64) f7;
+  crypto_int64 f0f8_2  = f0_2 * (crypto_int64) f8;
+  crypto_int64 f0f9_2  = f0_2 * (crypto_int64) f9;
+  crypto_int64 f1f1_2  = f1_2 * (crypto_int64) f1;
+  crypto_int64 f1f2_2  = f1_2 * (crypto_int64) f2;
+  crypto_int64 f1f3_4  = f1_2 * (crypto_int64) f3_2;
+  crypto_int64 f1f4_2  = f1_2 * (crypto_int64) f4;
+  crypto_int64 f1f5_4  = f1_2 * (crypto_int64) f5_2;
+  crypto_int64 f1f6_2  = f1_2 * (crypto_int64) f6;
+  crypto_int64 f1f7_4  = f1_2 * (crypto_int64) f7_2;
+  crypto_int64 f1f8_2  = f1_2 * (crypto_int64) f8;
+  crypto_int64 f1f9_76 = f1_2 * (crypto_int64) f9_38;
+  crypto_int64 f2f2    = f2   * (crypto_int64) f2;
+  crypto_int64 f2f3_2  = f2_2 * (crypto_int64) f3;
+  crypto_int64 f2f4_2  = f2_2 * (crypto_int64) f4;
+  crypto_int64 f2f5_2  = f2_2 * (crypto_int64) f5;
+  crypto_int64 f2f6_2  = f2_2 * (crypto_int64) f6;
+  crypto_int64 f2f7_2  = f2_2 * (crypto_int64) f7;
+  crypto_int64 f2f8_38 = f2_2 * (crypto_int64) f8_19;
+  crypto_int64 f2f9_38 = f2   * (crypto_int64) f9_38;
+  crypto_int64 f3f3_2  = f3_2 * (crypto_int64) f3;
+  crypto_int64 f3f4_2  = f3_2 * (crypto_int64) f4;
+  crypto_int64 f3f5_4  = f3_2 * (crypto_int64) f5_2;
+  crypto_int64 f3f6_2  = f3_2 * (crypto_int64) f6;
+  crypto_int64 f3f7_76 = f3_2 * (crypto_int64) f7_38;
+  crypto_int64 f3f8_38 = f3_2 * (crypto_int64) f8_19;
+  crypto_int64 f3f9_76 = f3_2 * (crypto_int64) f9_38;
+  crypto_int64 f4f4    = f4   * (crypto_int64) f4;
+  crypto_int64 f4f5_2  = f4_2 * (crypto_int64) f5;
+  crypto_int64 f4f6_38 = f4_2 * (crypto_int64) f6_19;
+  crypto_int64 f4f7_38 = f4   * (crypto_int64) f7_38;
+  crypto_int64 f4f8_38 = f4_2 * (crypto_int64) f8_19;
+  crypto_int64 f4f9_38 = f4   * (crypto_int64) f9_38;
+  crypto_int64 f5f5_38 = f5   * (crypto_int64) f5_38;
+  crypto_int64 f5f6_38 = f5_2 * (crypto_int64) f6_19;
+  crypto_int64 f5f7_76 = f5_2 * (crypto_int64) f7_38;
+  crypto_int64 f5f8_38 = f5_2 * (crypto_int64) f8_19;
+  crypto_int64 f5f9_76 = f5_2 * (crypto_int64) f9_38;
+  crypto_int64 f6f6_19 = f6   * (crypto_int64) f6_19;
+  crypto_int64 f6f7_38 = f6   * (crypto_int64) f7_38;
+  crypto_int64 f6f8_38 = f6_2 * (crypto_int64) f8_19;
+  crypto_int64 f6f9_38 = f6   * (crypto_int64) f9_38;
+  crypto_int64 f7f7_38 = f7   * (crypto_int64) f7_38;
+  crypto_int64 f7f8_38 = f7_2 * (crypto_int64) f8_19;
+  crypto_int64 f7f9_76 = f7_2 * (crypto_int64) f9_38;
+  crypto_int64 f8f8_19 = f8   * (crypto_int64) f8_19;
+  crypto_int64 f8f9_38 = f8   * (crypto_int64) f9_38;
+  crypto_int64 f9f9_38 = f9   * (crypto_int64) f9_38;
+  crypto_int64 h0 = f0f0  +f1f9_76+f2f8_38+f3f7_76+f4f6_38+f5f5_38;
+  crypto_int64 h1 = f0f1_2+f2f9_38+f3f8_38+f4f7_38+f5f6_38;
+  crypto_int64 h2 = f0f2_2+f1f1_2 +f3f9_76+f4f8_38+f5f7_76+f6f6_19;
+  crypto_int64 h3 = f0f3_2+f1f2_2 +f4f9_38+f5f8_38+f6f7_38;
+  crypto_int64 h4 = f0f4_2+f1f3_4 +f2f2   +f5f9_76+f6f8_38+f7f7_38;
+  crypto_int64 h5 = f0f5_2+f1f4_2 +f2f3_2 +f6f9_38+f7f8_38;
+  crypto_int64 h6 = f0f6_2+f1f5_4 +f2f4_2 +f3f3_2 +f7f9_76+f8f8_19;
+  crypto_int64 h7 = f0f7_2+f1f6_2 +f2f5_2 +f3f4_2 +f8f9_38;
+  crypto_int64 h8 = f0f8_2+f1f7_4 +f2f6_2 +f3f5_4 +f4f4   +f9f9_38;
+  crypto_int64 h9 = f0f9_2+f1f8_2 +f2f7_2 +f3f6_2 +f4f5_2;
+  crypto_int64 carry0;
+  crypto_int64 carry1;
+  crypto_int64 carry2;
+  crypto_int64 carry3;
+  crypto_int64 carry4;
+  crypto_int64 carry5;
+  crypto_int64 carry6;
+  crypto_int64 carry7;
+  crypto_int64 carry8;
+  crypto_int64 carry9;
+
+  carry0 = (h0 + (crypto_int64) (1<<25)) >> 26; h1 += carry0; h0 -= carry0 << 26;
+  carry4 = (h4 + (crypto_int64) (1<<25)) >> 26; h5 += carry4; h4 -= carry4 << 26;
+
+  carry1 = (h1 + (crypto_int64) (1<<24)) >> 25; h2 += carry1; h1 -= carry1 << 25;
+  carry5 = (h5 + (crypto_int64) (1<<24)) >> 25; h6 += carry5; h5 -= carry5 << 25;
+
+  carry2 = (h2 + (crypto_int64) (1<<25)) >> 26; h3 += carry2; h2 -= carry2 << 26;
+  carry6 = (h6 + (crypto_int64) (1<<25)) >> 26; h7 += carry6; h6 -= carry6 << 26;
+
+  carry3 = (h3 + (crypto_int64) (1<<24)) >> 25; h4 += carry3; h3 -= carry3 << 25;
+  carry7 = (h7 + (crypto_int64) (1<<24)) >> 25; h8 += carry7; h7 -= carry7 << 25;
+
+  carry4 = (h4 + (crypto_int64) (1<<25)) >> 26; h5 += carry4; h4 -= carry4 << 26;
+  carry8 = (h8 + (crypto_int64) (1<<25)) >> 26; h9 += carry8; h8 -= carry8 << 26;
+
+  carry9 = (h9 + (crypto_int64) (1<<24)) >> 25; h0 += carry9 * 19; h9 -= carry9 << 25;
+
+  carry0 = (h0 + (crypto_int64) (1<<25)) >> 26; h1 += carry0; h0 -= carry0 << 26;
+
+  h[0] = h0;
+  h[1] = h1;
+  h[2] = h2;
+  h[3] = h3;
+  h[4] = h4;
+  h[5] = h5;
+  h[6] = h6;
+  h[7] = h7;
+  h[8] = h8;
+  h[9] = h9;
+}
+
+#endif