C++ improvements/typo's

18 files changed, 105 insertions, 32 deletions

diff --git a/libtommath/bn_mp_div.c b/libtommath/bn_mp_div.c
index 71de55b..bca227d 100644
--- a/libtommath/bn_mp_div.c
+++ b/libtommath/bn_mp_div.c
@@ -31,7 +31,7 @@ mp_err mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d)
    }
 
    /* init our temps */
-   if ((err = mp_init_multi(&ta, &tb, &tq, &q, NULL)) != MP_OKAY) {
+   if ((err = mp_init_multi(&ta, &tb, &tq, &q, (void *)NULL)) != MP_OKAY) {
       return err;
    }
 
@@ -64,7 +64,7 @@ mp_err mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d)
       d->sign = MP_IS_ZERO(d) ? MP_ZPOS : n;
    }
 LBL_ERR:
-   mp_clear_multi(&ta, &tb, &tq, &q, NULL);
+   mp_clear_multi(&ta, &tb, &tq, &q, (void *)NULL);
    return err;
 }
 
diff --git a/libtommath/bn_mp_exptmod.c b/libtommath/bn_mp_exptmod.c
index 5f811eb..1ca2e93 100644
--- a/libtommath/bn_mp_exptmod.c
+++ b/libtommath/bn_mp_exptmod.c
@@ -26,7 +26,7 @@ mp_err mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y)
          return MP_VAL;
       }
 
-      if ((err = mp_init_multi(&tmpG, &tmpX, NULL)) != MP_OKAY) {
+      if ((err = mp_init_multi(&tmpG, &tmpX, (void *)NULL)) != MP_OKAY) {
          return err;
       }
 
@@ -43,7 +43,7 @@ mp_err mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y)
       /* and now compute (1/G)**|X| instead of G**X [X < 0] */
       err = mp_exptmod(&tmpG, &tmpX, P, Y);
 LBL_ERR:
-      mp_clear_multi(&tmpG, &tmpX, NULL);
+      mp_clear_multi(&tmpG, &tmpX, (void *)NULL);
       return err;
    }
 
diff --git a/libtommath/bn_mp_exteuclid.c b/libtommath/bn_mp_exteuclid.c
index faf47ba..fbac454 100644
--- a/libtommath/bn_mp_exteuclid.c
+++ b/libtommath/bn_mp_exteuclid.c
@@ -11,7 +11,7 @@ mp_err mp_exteuclid(const mp_int *a, const mp_int *b, mp_int *U1, mp_int *U2, mp
    mp_int u1, u2, u3, v1, v2, v3, t1, t2, t3, q, tmp;
    mp_err err;
 
-   if ((err = mp_init_multi(&u1, &u2, &u3, &v1, &v2, &v3, &t1, &t2, &t3, &q, &tmp, NULL)) != MP_OKAY) {
+   if ((err = mp_init_multi(&u1, &u2, &u3, &v1, &v2, &v3, &t1, &t2, &t3, &q, &tmp, (void *)NULL)) != MP_OKAY) {
       return err;
    }
 
@@ -67,7 +67,7 @@ mp_err mp_exteuclid(const mp_int *a, const mp_int *b, mp_int *U1, mp_int *U2, mp
 
    err = MP_OKAY;
 LBL_ERR:
-   mp_clear_multi(&u1, &u2, &u3, &v1, &v2, &v3, &t1, &t2, &t3, &q, &tmp, NULL);
+   mp_clear_multi(&u1, &u2, &u3, &v1, &v2, &v3, &t1, &t2, &t3, &q, &tmp, (void *)NULL);
    return err;
 }
 #endif
diff --git a/libtommath/bn_mp_lcm.c b/libtommath/bn_mp_lcm.c
index c32b269..5be3fc6 100644
--- a/libtommath/bn_mp_lcm.c
+++ b/libtommath/bn_mp_lcm.c
@@ -10,7 +10,7 @@ mp_err mp_lcm(const mp_int *a, const mp_int *b, mp_int *c)
    mp_int  t1, t2;
 
 
-   if ((err = mp_init_multi(&t1, &t2, NULL)) != MP_OKAY) {
+   if ((err = mp_init_multi(&t1, &t2, (void *)NULL)) != MP_OKAY) {
       return err;
    }
 
@@ -38,7 +38,7 @@ mp_err mp_lcm(const mp_int *a, const mp_int *b, mp_int *c)
    c->sign = MP_ZPOS;
 
 LBL_T:
-   mp_clear_multi(&t1, &t2, NULL);
+   mp_clear_multi(&t1, &t2, (void *)NULL);
    return err;
 }
 #endif
diff --git a/libtommath/bn_mp_mul.c b/libtommath/bn_mp_mul.c
index 561913a..91707cd 100644
--- a/libtommath/bn_mp_mul.c
+++ b/libtommath/bn_mp_mul.c
@@ -17,7 +17,7 @@ mp_err mp_mul(const mp_int *a, const mp_int *b, mp_int *c)
         * The bigger one needs to be at least about one MP_KARATSUBA_MUL_CUTOFF bigger
         * to make some sense, but it depends on architecture, OS, position of the
         * stars... so YMMV.
-        * Using it to cut the input into slices small enough for fast_s_mp_mul_digs
+        * Using it to cut the input into slices small enough for s_mp_mul_digs_fast
         * was actually slower on the author's machine, but YMMV.
         */
        (min_len >= MP_KARATSUBA_MUL_CUTOFF) &&
diff --git a/libtommath/bn_mp_prime_frobenius_underwood.c b/libtommath/bn_mp_prime_frobenius_underwood.c
index 253e8d5..be4bd08 100644
--- a/libtommath/bn_mp_prime_frobenius_underwood.c
+++ b/libtommath/bn_mp_prime_frobenius_underwood.c
@@ -32,7 +32,7 @@ mp_err mp_prime_frobenius_underwood(const mp_int *N, mp_bool *result)
 
    *result = MP_NO;
 
-   if ((err = mp_init_multi(&T1z, &T2z, &Np1z, &sz, &tz, NULL)) != MP_OKAY) {
+   if ((err = mp_init_multi(&T1z, &T2z, &Np1z, &sz, &tz, (void *)NULL)) != MP_OKAY) {
       return err;
    }
 
@@ -124,7 +124,7 @@ mp_err mp_prime_frobenius_underwood(const mp_int *N, mp_bool *result)
    }
 
 LBL_FU_ERR:
-   mp_clear_multi(&tz, &sz, &Np1z, &T2z, &T1z, NULL);
+   mp_clear_multi(&tz, &sz, &Np1z, &T2z, &T1z, (void *)NULL);
    return err;
 }
 
diff --git a/libtommath/bn_mp_prime_strong_lucas_selfridge.c b/libtommath/bn_mp_prime_strong_lucas_selfridge.c
index b50bbcd..617648c 100644
--- a/libtommath/bn_mp_prime_strong_lucas_selfridge.c
+++ b/libtommath/bn_mp_prime_strong_lucas_selfridge.c
@@ -73,7 +73,7 @@ mp_err mp_prime_strong_lucas_selfridge(const mp_int *a, mp_bool *result)
    */
 
    if ((err = mp_init_multi(&Dz, &gcd, &Np1, &Uz, &Vz, &U2mz, &V2mz, &Qmz, &Q2mz, &Qkdz, &T1z, &T2z, &T3z, &T4z, &Q2kdz,
-                            NULL)) != MP_OKAY) {
+                            (void *)NULL)) != MP_OKAY) {
       return err;
    }
 
@@ -281,7 +281,7 @@ mp_err mp_prime_strong_lucas_selfridge(const mp_int *a, mp_bool *result)
       }
    }
 LBL_LS_ERR:
-   mp_clear_multi(&Q2kdz, &T4z, &T3z, &T2z, &T1z, &Qkdz, &Q2mz, &Qmz, &V2mz, &U2mz, &Vz, &Uz, &Np1, &gcd, &Dz, NULL);
+   mp_clear_multi(&Q2kdz, &T4z, &T3z, &T2z, &T1z, &Qkdz, &Q2mz, &Qmz, &V2mz, &U2mz, &Vz, &Uz, &Np1, &gcd, &Dz, (void *)NULL);
    return err;
 }
 #endif
diff --git a/libtommath/bn_mp_root_n.c b/libtommath/bn_mp_root_n.c
index 5b92ff5..51ad759 100644
--- a/libtommath/bn_mp_root_n.c
+++ b/libtommath/bn_mp_root_n.c
@@ -27,7 +27,7 @@ mp_err mp_root_n(const mp_int *a, int b, mp_int *c)
       return MP_VAL;
    }
 
-   if ((err = mp_init_multi(&t1, &t2, &t3, NULL)) != MP_OKAY) {
+   if ((err = mp_init_multi(&t1, &t2, &t3, (void *)NULL)) != MP_OKAY) {
       return err;
    }
 
@@ -134,7 +134,7 @@ mp_err mp_root_n(const mp_int *a, int b, mp_int *c)
    c->sign = a->sign;
 
 LBL_ERR:
-   mp_clear_multi(&t1, &t2, &t3, NULL);
+   mp_clear_multi(&t1, &t2, &t3, (void *)NULL);
    return err;
 }
 
diff --git a/libtommath/bn_mp_sqrt.c b/libtommath/bn_mp_sqrt.c
index 82d6824..dcf28fd 100644
--- a/libtommath/bn_mp_sqrt.c
+++ b/libtommath/bn_mp_sqrt.c
@@ -3,11 +3,24 @@
 /* LibTomMath, multiple-precision integer library -- Tom St Denis */
 /* SPDX-License-Identifier: Unlicense */
 
+#ifndef NO_FLOATING_POINT
+#include <float.h>
+#include <math.h>
+#if (MP_DIGIT_BIT != 28) || (FLT_RADIX != 2) || (DBL_MANT_DIG != 53) || (DBL_MAX_EXP != 1024)
+#define NO_FLOATING_POINT
+#endif
+#endif
+
 /* this function is less generic than mp_n_root, simpler and faster */
 mp_err mp_sqrt(const mp_int *arg, mp_int *ret)
 {
    mp_err err;
    mp_int t1, t2;
+#ifndef NO_FLOATING_POINT
+   int i, j, k;
+   volatile double d;
+   mp_digit dig;
+#endif
 
    /* must be positive */
    if (arg->sign == MP_NEG) {
@@ -20,6 +33,64 @@ mp_err mp_sqrt(const mp_int *arg, mp_int *ret)
       return MP_OKAY;
    }
 
+#ifndef NO_FLOATING_POINT
+
+   i = (arg->used / 2) - 1;
+   j = 2 * i;
+   if ((err = mp_init_size(&t1, i+2)) != MP_OKAY) {
+      return err;
+   }
+
+   if ((err = mp_init(&t2)) != MP_OKAY) {
+      goto E2;
+   }
+
+   for (k = 0; k < i; ++k) {
+      t1.dp[k] = (mp_digit) 0;
+   }
+
+   /* Estimate the square root using the hardware floating point unit. */
+
+   d = 0.0;
+   for (k = arg->used-1; k >= j; --k) {
+      d = ldexp(d, MP_DIGIT_BIT) + (double)(arg->dp[k]);
+   }
+
+   /*
+    * At this point, d is the nearest floating point number to the most
+    * significant 1 or 2 mp_digits of arg. Extract its square root.
+    */
+
+   d = sqrt(d);
+
+   /* dig is the most significant mp_digit of the square root */
+
+   dig = (mp_digit) ldexp(d, -MP_DIGIT_BIT);
+
+   /*
+    * If the most significant digit is nonzero, find the next digit down
+    * by subtracting MP_DIGIT_BIT times thie most significant digit.
+    * Subtract one from the result so that our initial estimate is always
+    * low.
+    */
+
+   if (dig) {
+      t1.used = i+2;
+      d -= ldexp((double) dig, MP_DIGIT_BIT);
+      if (d >= 1.0) {
+         t1.dp[i+1] = dig;
+         t1.dp[i] = ((mp_digit) d) - 1;
+      } else {
+         t1.dp[i+1] = dig-1;
+         t1.dp[i] = MP_DIGIT_MAX;
+      }
+   } else {
+      t1.used = i+1;
+      t1.dp[i] = ((mp_digit) d) - 1;
+   }
+
+#else
+
    if ((err = mp_init_copy(&t1, arg)) != MP_OKAY) {
       return err;
    }
@@ -31,6 +102,8 @@ mp_err mp_sqrt(const mp_int *arg, mp_int *ret)
    /* First approx. (not very bad for large arg) */
    mp_rshd(&t1, t1.used/2);
 
+#endif
+
    /* t1 > 0  */
    if ((err = mp_div(arg, &t1, &t2, NULL)) != MP_OKAY) {
       goto E1;
diff --git a/libtommath/bn_mp_sqrtmod_prime.c b/libtommath/bn_mp_sqrtmod_prime.c
index a833ed7..425ae17 100644
--- a/libtommath/bn_mp_sqrtmod_prime.c
+++ b/libtommath/bn_mp_sqrtmod_prime.c
@@ -25,7 +25,7 @@ mp_err mp_sqrtmod_prime(const mp_int *n, const mp_int *prime, mp_int *ret)
    if ((err = mp_kronecker(n, prime, &legendre)) != MP_OKAY)        return err;
    if (legendre == -1)                                           return MP_VAL; /* quadratic non-residue mod prime */
 
-   if ((err = mp_init_multi(&t1, &C, &Q, &S, &Z, &M, &T, &R, &two, NULL)) != MP_OKAY) {
+   if ((err = mp_init_multi(&t1, &C, &Q, &S, &Z, &M, &T, &R, &two, (void *)NULL)) != MP_OKAY) {
       return err;
    }
 
@@ -111,7 +111,7 @@ mp_err mp_sqrtmod_prime(const mp_int *n, const mp_int *prime, mp_int *ret)
    }
 
 cleanup:
-   mp_clear_multi(&t1, &C, &Q, &S, &Z, &M, &T, &R, &two, NULL);
+   mp_clear_multi(&t1, &C, &Q, &S, &Z, &M, &T, &R, &two, (void *)NULL);
    return err;
 }
 
diff --git a/libtommath/bn_s_mp_balance_mul.c b/libtommath/bn_s_mp_balance_mul.c
index 7ece5d7..557cc1d 100644
--- a/libtommath/bn_s_mp_balance_mul.c
+++ b/libtommath/bn_s_mp_balance_mul.c
@@ -19,7 +19,7 @@ mp_err s_mp_balance_mul(const mp_int *a, const mp_int *b, mp_int *c)
    if ((err = mp_init_size(&a0, bsize + 2)) != MP_OKAY) {
       return err;
    }
-   if ((err = mp_init_multi(&tmp, &r, NULL)) != MP_OKAY) {
+   if ((err = mp_init_multi(&tmp, &r, (void *)NULL)) != MP_OKAY) {
       mp_clear(&a0);
       return err;
    }
@@ -75,7 +75,7 @@ mp_err s_mp_balance_mul(const mp_int *a, const mp_int *b, mp_int *c)
 
    mp_exch(&r,c);
 LBL_ERR:
-   mp_clear_multi(&a0, &tmp, &r,NULL);
+   mp_clear_multi(&a0, &tmp, &r, (void *)NULL);
    return err;
 }
 #endif
diff --git a/libtommath/bn_s_mp_invmod_fast.c b/libtommath/bn_s_mp_invmod_fast.c
index 677d7ab..e76c604 100644
--- a/libtommath/bn_s_mp_invmod_fast.c
+++ b/libtommath/bn_s_mp_invmod_fast.c
@@ -21,7 +21,7 @@ mp_err s_mp_invmod_fast(const mp_int *a, const mp_int *b, mp_int *c)
    }
 
    /* init all our temps */
-   if ((err = mp_init_multi(&x, &y, &u, &v, &B, &D, NULL)) != MP_OKAY) {
+   if ((err = mp_init_multi(&x, &y, &u, &v, &B, &D, (void *)NULL)) != MP_OKAY) {
       return err;
    }
 
@@ -112,7 +112,7 @@ top:
    err = MP_OKAY;
 
 LBL_ERR:
-   mp_clear_multi(&x, &y, &u, &v, &B, &D, NULL);
+   mp_clear_multi(&x, &y, &u, &v, &B, &D, (void *)NULL);
    return err;
 }
 #endif
diff --git a/libtommath/bn_s_mp_invmod_slow.c b/libtommath/bn_s_mp_invmod_slow.c
index 4c5db33..6079d4b 100644
--- a/libtommath/bn_s_mp_invmod_slow.c
+++ b/libtommath/bn_s_mp_invmod_slow.c
@@ -16,7 +16,7 @@ mp_err s_mp_invmod_slow(const mp_int *a, const mp_int *b, mp_int *c)
 
    /* init temps */
    if ((err = mp_init_multi(&x, &y, &u, &v,
-                            &A, &B, &C, &D, NULL)) != MP_OKAY) {
+                            &A, &B, &C, &D, (void *)NULL)) != MP_OKAY) {
       return err;
    }
 
@@ -113,7 +113,7 @@ top:
    mp_exch(&C, c);
    err = MP_OKAY;
 LBL_ERR:
-   mp_clear_multi(&x, &y, &u, &v, &A, &B, &C, &D, NULL);
+   mp_clear_multi(&x, &y, &u, &v, &A, &B, &C, &D, (void *)NULL);
    return err;
 }
 #endif
diff --git a/libtommath/bn_s_mp_log.c b/libtommath/bn_s_mp_log.c
index a75212a..6ead7d9 100644
--- a/libtommath/bn_s_mp_log.c
+++ b/libtommath/bn_s_mp_log.c
@@ -17,7 +17,7 @@ mp_err s_mp_log(const mp_int *a, mp_digit base, int *c)
 
    if ((err =
            mp_init_multi(&bracket_low, &bracket_high,
-                         &bracket_mid, &t, &bi_base, NULL)) != MP_OKAY) {
+                         &bracket_mid, &t, &bi_base, (void *)NULL)) != MP_OKAY) {
       return err;
    }
 
@@ -73,7 +73,7 @@ mp_err s_mp_log(const mp_int *a, mp_digit base, int *c)
 
 LBL_END:
    mp_clear_multi(&bracket_low, &bracket_high, &bracket_mid,
-                  &t, &bi_base, NULL);
+                  &t, &bi_base, (void *)NULL);
    return err;
 }
 
diff --git a/libtommath/bn_s_mp_mul_high_digs_fast.c b/libtommath/bn_s_mp_mul_high_digs_fast.c
index 4ce7f59..04c74a8 100644
--- a/libtommath/bn_s_mp_mul_high_digs_fast.c
+++ b/libtommath/bn_s_mp_mul_high_digs_fast.c
@@ -3,8 +3,8 @@
 /* LibTomMath, multiple-precision integer library -- Tom St Denis */
 /* SPDX-License-Identifier: Unlicense */
 
-/* this is a modified version of fast_s_mul_digs that only produces
- * output digits *above* digs.  See the comments for fast_s_mul_digs
+/* this is a modified version of s_mp_mul_digs_fast  that only produces
+ * output digits *above* digs.  See the comments for s_mp_mul_digs_fast
  * to see how it works.
  *
  * This is used in the Barrett reduction since for one of the multiplications
diff --git a/libtommath/bn_s_mp_rand_jenkins.c b/libtommath/bn_s_mp_rand_jenkins.c
index da0771c..c64afac 100644
--- a/libtommath/bn_s_mp_rand_jenkins.c
+++ b/libtommath/bn_s_mp_rand_jenkins.c
@@ -27,10 +27,10 @@ static uint64_t s_rand_jenkins_val(void)
 
 void s_mp_rand_jenkins_init(uint64_t seed)
 {
-   uint64_t i;
+   int i;
    jenkins_x.a = 0xf1ea5eedULL;
    jenkins_x.b = jenkins_x.c = jenkins_x.d = seed;
-   for (i = 0uLL; i < 20uLL; ++i) {
+   for (i = 0; i < 20; ++i) {
       (void)s_rand_jenkins_val();
    }
 }
diff --git a/libtommath/bn_s_mp_toom_mul.c b/libtommath/bn_s_mp_toom_mul.c
index eefce6c..fd574a2 100644
--- a/libtommath/bn_s_mp_toom_mul.c
+++ b/libtommath/bn_s_mp_toom_mul.c
@@ -36,7 +36,7 @@ mp_err s_mp_toom_mul(const mp_int *a, const mp_int *b, mp_int *c)
    mp_err err;
 
    /* init temps */
-   if ((err = mp_init_multi(&S1, &S2, &T1, NULL)) != MP_OKAY) {
+   if ((err = mp_init_multi(&S1, &S2, &T1, (void *)NULL)) != MP_OKAY) {
       return err;
    }
 
@@ -208,7 +208,7 @@ LBL_ERRa2:
 LBL_ERRa1:
    mp_clear(&a0);
 LBL_ERRa0:
-   mp_clear_multi(&S1, &S2, &T1, NULL);
+   mp_clear_multi(&S1, &S2, &T1, (void *)NULL);
    return err;
 }
 
diff --git a/libtommath/changes.txt b/libtommath/changes.txt
index 80ff7dd..ac9e49a 100644
--- a/libtommath/changes.txt
+++ b/libtommath/changes.txt
@@ -423,7 +423,7 @@ v0.13  -- tons of minor speed-ups in low level add, sub, mul_2 and div_2 which p
 Jan 17th, 2003
 v0.12  -- re-wrote the majority of the makefile so its more portable and will
           install via "make install" on most *nix platforms
-       -- Re-packaged all the source as seperate files.  Means the library a single
+       -- Re-packaged all the source as separate files.  Means the library a single
           file packagage any more.  Instead of just adding "bn.c" you have to add
           libtommath.a
        -- Renamed "bn.h" to "tommath.h"