changeset 2:86e0b50a9b58 libtommath-orig ltm-0.30-orig

ltm 0.30 orig import
author Matt Johnston <matt@ucc.asn.au>
date Mon, 31 May 2004 18:25:22 +0000
parents
children e1037a1e12e7
files LICENSE bn_error.c bn_fast_mp_invmod.c bn_fast_mp_montgomery_reduce.c bn_fast_s_mp_mul_digs.c bn_fast_s_mp_mul_high_digs.c bn_fast_s_mp_sqr.c bn_mp_2expt.c bn_mp_abs.c bn_mp_add.c bn_mp_add_d.c bn_mp_addmod.c bn_mp_and.c bn_mp_clamp.c bn_mp_clear.c bn_mp_clear_multi.c bn_mp_cmp.c bn_mp_cmp_d.c bn_mp_cmp_mag.c bn_mp_cnt_lsb.c bn_mp_copy.c bn_mp_count_bits.c bn_mp_div.c bn_mp_div_2.c bn_mp_div_2d.c bn_mp_div_3.c bn_mp_div_d.c bn_mp_dr_is_modulus.c bn_mp_dr_reduce.c bn_mp_dr_setup.c bn_mp_exch.c bn_mp_expt_d.c bn_mp_exptmod.c bn_mp_exptmod_fast.c bn_mp_exteuclid.c bn_mp_fread.c bn_mp_fwrite.c bn_mp_gcd.c bn_mp_get_int.c bn_mp_grow.c bn_mp_init.c bn_mp_init_copy.c bn_mp_init_multi.c bn_mp_init_set.c bn_mp_init_set_int.c bn_mp_init_size.c bn_mp_invmod.c bn_mp_is_square.c bn_mp_jacobi.c bn_mp_karatsuba_mul.c bn_mp_karatsuba_sqr.c bn_mp_lcm.c bn_mp_lshd.c bn_mp_mod.c bn_mp_mod_2d.c bn_mp_mod_d.c bn_mp_montgomery_calc_normalization.c bn_mp_montgomery_reduce.c bn_mp_montgomery_setup.c bn_mp_mul.c bn_mp_mul_2.c bn_mp_mul_2d.c bn_mp_mul_d.c bn_mp_mulmod.c bn_mp_n_root.c bn_mp_neg.c bn_mp_or.c bn_mp_prime_fermat.c bn_mp_prime_is_divisible.c bn_mp_prime_is_prime.c bn_mp_prime_miller_rabin.c bn_mp_prime_next_prime.c bn_mp_prime_random_ex.c bn_mp_radix_size.c bn_mp_radix_smap.c bn_mp_rand.c bn_mp_read_radix.c bn_mp_read_signed_bin.c bn_mp_read_unsigned_bin.c bn_mp_reduce.c bn_mp_reduce_2k.c bn_mp_reduce_2k_setup.c bn_mp_reduce_is_2k.c bn_mp_reduce_setup.c bn_mp_rshd.c bn_mp_set.c bn_mp_set_int.c bn_mp_shrink.c bn_mp_signed_bin_size.c bn_mp_sqr.c bn_mp_sqrmod.c bn_mp_sqrt.c bn_mp_sub.c bn_mp_sub_d.c bn_mp_submod.c bn_mp_to_signed_bin.c bn_mp_to_unsigned_bin.c bn_mp_toom_mul.c bn_mp_toom_sqr.c bn_mp_toradix.c bn_mp_toradix_n.c bn_mp_unsigned_bin_size.c bn_mp_xor.c bn_mp_zero.c bn_prime_sizes_tab.c bn_prime_tab.c bn_reverse.c bn_s_mp_add.c bn_s_mp_exptmod.c bn_s_mp_mul_digs.c bn_s_mp_mul_high_digs.c bn_s_mp_sqr.c bn_s_mp_sub.c bncore.c booker.pl changes.txt demo/demo.c etc/2kprime.1 etc/2kprime.c etc/drprime.c etc/drprimes.28 etc/drprimes.txt etc/makefile etc/makefile.msvc etc/mersenne.c etc/mont.c etc/pprime.c etc/prime.1024 etc/prime.512 etc/timer.asm etc/tune.c gen.pl logs/README logs/add.log logs/addsub.png logs/expt.log logs/expt.png logs/expt_2k.log logs/expt_dr.log logs/graphs.dem logs/index.html logs/invmod.log logs/invmod.png logs/k7/README logs/k7/add.log logs/k7/addsub.png logs/k7/expt.log logs/k7/expt.png logs/k7/expt_dr.log logs/k7/graphs.dem logs/k7/index.html logs/k7/invmod.log logs/k7/invmod.png logs/k7/mult.log logs/k7/mult.png logs/k7/mult_kara.log logs/k7/sqr.log logs/k7/sqr_kara.log logs/k7/sub.log logs/mult.log logs/mult.png logs/mult_kara.log logs/p4/README logs/p4/add.log logs/p4/addsub.png logs/p4/expt.log logs/p4/expt.png logs/p4/expt_dr.log logs/p4/graphs.dem logs/p4/index.html logs/p4/invmod.log logs/p4/invmod.png logs/p4/mult.log logs/p4/mult.png logs/p4/mult_kara.log logs/p4/sqr.log logs/p4/sqr_kara.log logs/p4/sub.log logs/sqr.log logs/sqr.old logs/sqr_kara.log logs/sub.log makefile makefile.bcc makefile.cygwin_dll makefile.msvc mtest/logtab.h mtest/mpi-config.h mtest/mpi-types.h mtest/mpi.c mtest/mpi.h mtest/mtest.c pics/design_process.sxd pics/design_process.tif pics/expt_state.sxd pics/expt_state.tif pics/makefile pics/primality.tif pics/radix.sxd pics/sliding_window.sxd pics/sliding_window.tif pre_gen/mpi.c tommath.h
diffstat 203 files changed, 25560 insertions(+), 0 deletions(-) [+]
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/LICENSE	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,4 @@
+LibTomMath is hereby released into the Public Domain.  
+
+-- Tom St Denis
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_error.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,41 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+static const struct {
+     int code;
+     char *msg;
+} msgs[] = {
+     { MP_OKAY, "Successful" },
+     { MP_MEM,  "Out of heap" },
+     { MP_VAL,  "Value out of range" }
+};
+
+/* return a char * string for a given code */
+char *mp_error_to_string(int code)
+{
+   int x;
+
+   /* scan the lookup table for the given message */
+   for (x = 0; x < (int)(sizeof(msgs) / sizeof(msgs[0])); x++) {
+       if (msgs[x].code == code) {
+          return msgs[x].msg;
+       }
+   }
+
+   /* generic reply for invalid code */
+   return "Invalid error code";
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_fast_mp_invmod.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,143 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* computes the modular inverse via binary extended euclidean algorithm, 
+ * that is c = 1/a mod b 
+ *
+ * Based on mp_invmod except this is optimized for the case where b is 
+ * odd as per HAC Note 14.64 on pp. 610
+ */
+int
+fast_mp_invmod (mp_int * a, mp_int * b, mp_int * c)
+{
+  mp_int  x, y, u, v, B, D;
+  int     res, neg;
+
+  /* 2. [modified] b must be odd   */
+  if (mp_iseven (b) == 1) {
+    return MP_VAL;
+  }
+
+  /* init all our temps */
+  if ((res = mp_init_multi(&x, &y, &u, &v, &B, &D, NULL)) != MP_OKAY) {
+     return res;
+  }
+
+  /* x == modulus, y == value to invert */
+  if ((res = mp_copy (b, &x)) != MP_OKAY) {
+    goto __ERR;
+  }
+
+  /* we need y = |a| */
+  if ((res = mp_abs (a, &y)) != MP_OKAY) {
+    goto __ERR;
+  }
+
+  /* 3. u=x, v=y, A=1, B=0, C=0,D=1 */
+  if ((res = mp_copy (&x, &u)) != MP_OKAY) {
+    goto __ERR;
+  }
+  if ((res = mp_copy (&y, &v)) != MP_OKAY) {
+    goto __ERR;
+  }
+  mp_set (&D, 1);
+
+top:
+  /* 4.  while u is even do */
+  while (mp_iseven (&u) == 1) {
+    /* 4.1 u = u/2 */
+    if ((res = mp_div_2 (&u, &u)) != MP_OKAY) {
+      goto __ERR;
+    }
+    /* 4.2 if B is odd then */
+    if (mp_isodd (&B) == 1) {
+      if ((res = mp_sub (&B, &x, &B)) != MP_OKAY) {
+        goto __ERR;
+      }
+    }
+    /* B = B/2 */
+    if ((res = mp_div_2 (&B, &B)) != MP_OKAY) {
+      goto __ERR;
+    }
+  }
+
+  /* 5.  while v is even do */
+  while (mp_iseven (&v) == 1) {
+    /* 5.1 v = v/2 */
+    if ((res = mp_div_2 (&v, &v)) != MP_OKAY) {
+      goto __ERR;
+    }
+    /* 5.2 if D is odd then */
+    if (mp_isodd (&D) == 1) {
+      /* D = (D-x)/2 */
+      if ((res = mp_sub (&D, &x, &D)) != MP_OKAY) {
+        goto __ERR;
+      }
+    }
+    /* D = D/2 */
+    if ((res = mp_div_2 (&D, &D)) != MP_OKAY) {
+      goto __ERR;
+    }
+  }
+
+  /* 6.  if u >= v then */
+  if (mp_cmp (&u, &v) != MP_LT) {
+    /* u = u - v, B = B - D */
+    if ((res = mp_sub (&u, &v, &u)) != MP_OKAY) {
+      goto __ERR;
+    }
+
+    if ((res = mp_sub (&B, &D, &B)) != MP_OKAY) {
+      goto __ERR;
+    }
+  } else {
+    /* v - v - u, D = D - B */
+    if ((res = mp_sub (&v, &u, &v)) != MP_OKAY) {
+      goto __ERR;
+    }
+
+    if ((res = mp_sub (&D, &B, &D)) != MP_OKAY) {
+      goto __ERR;
+    }
+  }
+
+  /* if not zero goto step 4 */
+  if (mp_iszero (&u) == 0) {
+    goto top;
+  }
+
+  /* now a = C, b = D, gcd == g*v */
+
+  /* if v != 1 then there is no inverse */
+  if (mp_cmp_d (&v, 1) != MP_EQ) {
+    res = MP_VAL;
+    goto __ERR;
+  }
+
+  /* b is now the inverse */
+  neg = a->sign;
+  while (D.sign == MP_NEG) {
+    if ((res = mp_add (&D, b, &D)) != MP_OKAY) {
+      goto __ERR;
+    }
+  }
+  mp_exch (&D, c);
+  c->sign = neg;
+  res = MP_OKAY;
+
+__ERR:mp_clear_multi (&x, &y, &u, &v, &B, &D, NULL);
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_fast_mp_montgomery_reduce.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,167 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* computes xR**-1 == x (mod N) via Montgomery Reduction
+ *
+ * This is an optimized implementation of mp_montgomery_reduce
+ * which uses the comba method to quickly calculate the columns of the
+ * reduction.
+ *
+ * Based on Algorithm 14.32 on pp.601 of HAC.
+*/
+int
+fast_mp_montgomery_reduce (mp_int * x, mp_int * n, mp_digit rho)
+{
+  int     ix, res, olduse;
+  mp_word W[MP_WARRAY];
+
+  /* get old used count */
+  olduse = x->used;
+
+  /* grow a as required */
+  if (x->alloc < n->used + 1) {
+    if ((res = mp_grow (x, n->used + 1)) != MP_OKAY) {
+      return res;
+    }
+  }
+
+  /* first we have to get the digits of the input into
+   * an array of double precision words W[...]
+   */
+  {
+    register mp_word *_W;
+    register mp_digit *tmpx;
+
+    /* alias for the W[] array */
+    _W   = W;
+
+    /* alias for the digits of  x*/
+    tmpx = x->dp;
+
+    /* copy the digits of a into W[0..a->used-1] */
+    for (ix = 0; ix < x->used; ix++) {
+      *_W++ = *tmpx++;
+    }
+
+    /* zero the high words of W[a->used..m->used*2] */
+    for (; ix < n->used * 2 + 1; ix++) {
+      *_W++ = 0;
+    }
+  }
+
+  /* now we proceed to zero successive digits
+   * from the least significant upwards
+   */
+  for (ix = 0; ix < n->used; ix++) {
+    /* mu = ai * m' mod b
+     *
+     * We avoid a double precision multiplication (which isn't required)
+     * by casting the value down to a mp_digit.  Note this requires
+     * that W[ix-1] have  the carry cleared (see after the inner loop)
+     */
+    register mp_digit mu;
+    mu = (mp_digit) (((W[ix] & MP_MASK) * rho) & MP_MASK);
+
+    /* a = a + mu * m * b**i
+     *
+     * This is computed in place and on the fly.  The multiplication
+     * by b**i is handled by offseting which columns the results
+     * are added to.
+     *
+     * Note the comba method normally doesn't handle carries in the
+     * inner loop In this case we fix the carry from the previous
+     * column since the Montgomery reduction requires digits of the
+     * result (so far) [see above] to work.  This is
+     * handled by fixing up one carry after the inner loop.  The
+     * carry fixups are done in order so after these loops the
+     * first m->used words of W[] have the carries fixed
+     */
+    {
+      register int iy;
+      register mp_digit *tmpn;
+      register mp_word *_W;
+
+      /* alias for the digits of the modulus */
+      tmpn = n->dp;
+
+      /* Alias for the columns set by an offset of ix */
+      _W = W + ix;
+
+      /* inner loop */
+      for (iy = 0; iy < n->used; iy++) {
+          *_W++ += ((mp_word)mu) * ((mp_word)*tmpn++);
+      }
+    }
+
+    /* now fix carry for next digit, W[ix+1] */
+    W[ix + 1] += W[ix] >> ((mp_word) DIGIT_BIT);
+  }
+
+  /* now we have to propagate the carries and
+   * shift the words downward [all those least
+   * significant digits we zeroed].
+   */
+  {
+    register mp_digit *tmpx;
+    register mp_word *_W, *_W1;
+
+    /* nox fix rest of carries */
+
+    /* alias for current word */
+    _W1 = W + ix;
+
+    /* alias for next word, where the carry goes */
+    _W = W + ++ix;
+
+    for (; ix <= n->used * 2 + 1; ix++) {
+      *_W++ += *_W1++ >> ((mp_word) DIGIT_BIT);
+    }
+
+    /* copy out, A = A/b**n
+     *
+     * The result is A/b**n but instead of converting from an
+     * array of mp_word to mp_digit than calling mp_rshd
+     * we just copy them in the right order
+     */
+
+    /* alias for destination word */
+    tmpx = x->dp;
+
+    /* alias for shifted double precision result */
+    _W = W + n->used;
+
+    for (ix = 0; ix < n->used + 1; ix++) {
+      *tmpx++ = (mp_digit)(*_W++ & ((mp_word) MP_MASK));
+    }
+
+    /* zero oldused digits, if the input a was larger than
+     * m->used+1 we'll have to clear the digits
+     */
+    for (; ix < olduse; ix++) {
+      *tmpx++ = 0;
+    }
+  }
+
+  /* set the max used and clamp */
+  x->used = n->used + 1;
+  mp_clamp (x);
+
+  /* if A >= m then A = A - m */
+  if (mp_cmp_mag (x, n) != MP_LT) {
+    return s_mp_sub (x, n, x);
+  }
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_fast_s_mp_mul_digs.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,130 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* Fast (comba) multiplier
+ *
+ * This is the fast column-array [comba] multiplier.  It is 
+ * designed to compute the columns of the product first 
+ * then handle the carries afterwards.  This has the effect 
+ * of making the nested loops that compute the columns very
+ * simple and schedulable on super-scalar processors.
+ *
+ * This has been modified to produce a variable number of 
+ * digits of output so if say only a half-product is required 
+ * you don't have to compute the upper half (a feature 
+ * required for fast Barrett reduction).
+ *
+ * Based on Algorithm 14.12 on pp.595 of HAC.
+ *
+ */
+int
+fast_s_mp_mul_digs (mp_int * a, mp_int * b, mp_int * c, int digs)
+{
+  int     olduse, res, pa, ix;
+  mp_word W[MP_WARRAY];
+
+  /* grow the destination as required */
+  if (c->alloc < digs) {
+    if ((res = mp_grow (c, digs)) != MP_OKAY) {
+      return res;
+    }
+  }
+
+  /* clear temp buf (the columns) */
+  memset (W, 0, sizeof (mp_word) * digs);
+
+  /* calculate the columns */
+  pa = a->used;
+  for (ix = 0; ix < pa; ix++) {
+    /* this multiplier has been modified to allow you to 
+     * control how many digits of output are produced.  
+     * So at most we want to make upto "digs" digits of output.
+     *
+     * this adds products to distinct columns (at ix+iy) of W
+     * note that each step through the loop is not dependent on
+     * the previous which means the compiler can easily unroll
+     * the loop without scheduling problems
+     */
+    {
+      register mp_digit tmpx, *tmpy;
+      register mp_word *_W;
+      register int iy, pb;
+
+      /* alias for the the word on the left e.g. A[ix] * A[iy] */
+      tmpx = a->dp[ix];
+
+      /* alias for the right side */
+      tmpy = b->dp;
+
+      /* alias for the columns, each step through the loop adds a new
+         term to each column
+       */
+      _W = W + ix;
+
+      /* the number of digits is limited by their placement.  E.g.
+         we avoid multiplying digits that will end up above the # of
+         digits of precision requested
+       */
+      pb = MIN (b->used, digs - ix);
+
+      for (iy = 0; iy < pb; iy++) {
+        *_W++ += ((mp_word)tmpx) * ((mp_word)*tmpy++);
+      }
+    }
+
+  }
+
+  /* setup dest */
+  olduse = c->used;
+  c->used = digs;
+
+  {
+    register mp_digit *tmpc;
+
+    /* At this point W[] contains the sums of each column.  To get the
+     * correct result we must take the extra bits from each column and
+     * carry them down
+     *
+     * Note that while this adds extra code to the multiplier it 
+     * saves time since the carry propagation is removed from the 
+     * above nested loop.This has the effect of reducing the work 
+     * from N*(N+N*c)==N**2 + c*N**2 to N**2 + N*c where c is the 
+     * cost of the shifting.  On very small numbers this is slower 
+     * but on most cryptographic size numbers it is faster.
+     *
+     * In this particular implementation we feed the carries from
+     * behind which means when the loop terminates we still have one
+     * last digit to copy
+     */
+    tmpc = c->dp;
+    for (ix = 1; ix < digs; ix++) {
+      /* forward the carry from the previous temp */
+      W[ix] += (W[ix - 1] >> ((mp_word) DIGIT_BIT));
+
+      /* now extract the previous digit [below the carry] */
+      *tmpc++ = (mp_digit) (W[ix - 1] & ((mp_word) MP_MASK));
+    }
+    /* fetch the last digit */
+    *tmpc++ = (mp_digit) (W[digs - 1] & ((mp_word) MP_MASK));
+
+    /* clear unused digits [that existed in the old copy of c] */
+    for (; ix < olduse; ix++) {
+      *tmpc++ = 0;
+    }
+  }
+  mp_clamp (c);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_fast_s_mp_mul_high_digs.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,98 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+ #include <tommath.h>
+
+/* this is a modified version of fast_s_mp_mul_digs that only produces
+ * output digits *above* digs.  See the comments for fast_s_mp_mul_digs
+ * to see how it works.
+ *
+ * This is used in the Barrett reduction since for one of the multiplications
+ * only the higher digits were needed.  This essentially halves the work.
+ *
+ * Based on Algorithm 14.12 on pp.595 of HAC.
+ */
+int
+fast_s_mp_mul_high_digs (mp_int * a, mp_int * b, mp_int * c, int digs)
+{
+  int     oldused, newused, res, pa, pb, ix;
+  mp_word W[MP_WARRAY];
+
+  /* calculate size of product and allocate more space if required */
+  newused = a->used + b->used + 1;
+  if (c->alloc < newused) {
+    if ((res = mp_grow (c, newused)) != MP_OKAY) {
+      return res;
+    }
+  }
+
+  /* like the other comba method we compute the columns first */
+  pa = a->used;
+  pb = b->used;
+  memset (W + digs, 0, (pa + pb + 1 - digs) * sizeof (mp_word));
+  for (ix = 0; ix < pa; ix++) {
+    {
+      register mp_digit tmpx, *tmpy;
+      register int iy;
+      register mp_word *_W;
+
+      /* work todo, that is we only calculate digits that are at "digs" or above  */
+      iy = digs - ix;
+
+      /* copy of word on the left of A[ix] * B[iy] */
+      tmpx = a->dp[ix];
+
+      /* alias for right side */
+      tmpy = b->dp + iy;
+     
+      /* alias for the columns of output.  Offset to be equal to or above the 
+       * smallest digit place requested 
+       */
+      _W = W + digs;     
+      
+      /* skip cases below zero where ix > digs */
+      if (iy < 0) {
+         iy    = abs(iy);
+         tmpy += iy;
+         _W   += iy;
+         iy    = 0;
+      }
+
+      /* compute column products for digits above the minimum */
+      for (; iy < pb; iy++) {
+         *_W++ += ((mp_word) tmpx) * ((mp_word)*tmpy++);
+      }
+    }
+  }
+
+  /* setup dest */
+  oldused = c->used;
+  c->used = newused;
+
+  /* now convert the array W downto what we need
+   *
+   * See comments in bn_fast_s_mp_mul_digs.c
+   */
+  for (ix = digs + 1; ix < newused; ix++) {
+    W[ix] += (W[ix - 1] >> ((mp_word) DIGIT_BIT));
+    c->dp[ix - 1] = (mp_digit) (W[ix - 1] & ((mp_word) MP_MASK));
+  }
+  c->dp[newused - 1] = (mp_digit) (W[newused - 1] & ((mp_word) MP_MASK));
+
+  for (; ix < oldused; ix++) {
+    c->dp[ix] = 0;
+  }
+  mp_clamp (c);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_fast_s_mp_sqr.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,139 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* fast squaring
+ *
+ * This is the comba method where the columns of the product
+ * are computed first then the carries are computed.  This
+ * has the effect of making a very simple inner loop that
+ * is executed the most
+ *
+ * W2 represents the outer products and W the inner.
+ *
+ * A further optimizations is made because the inner
+ * products are of the form "A * B * 2".  The *2 part does
+ * not need to be computed until the end which is good
+ * because 64-bit shifts are slow!
+ *
+ * Based on Algorithm 14.16 on pp.597 of HAC.
+ *
+ */
+int fast_s_mp_sqr (mp_int * a, mp_int * b)
+{
+  int     olduse, newused, res, ix, pa;
+  mp_word W2[MP_WARRAY], W[MP_WARRAY];
+
+  /* calculate size of product and allocate as required */
+  pa = a->used;
+  newused = pa + pa + 1;
+  if (b->alloc < newused) {
+    if ((res = mp_grow (b, newused)) != MP_OKAY) {
+      return res;
+    }
+  }
+
+  /* zero temp buffer (columns)
+   * Note that there are two buffers.  Since squaring requires
+   * a outer and inner product and the inner product requires
+   * computing a product and doubling it (a relatively expensive
+   * op to perform n**2 times if you don't have to) the inner and
+   * outer products are computed in different buffers.  This way
+   * the inner product can be doubled using n doublings instead of
+   * n**2
+   */
+  memset (W,  0, newused * sizeof (mp_word));
+  memset (W2, 0, newused * sizeof (mp_word));
+
+  /* This computes the inner product.  To simplify the inner N**2 loop
+   * the multiplication by two is done afterwards in the N loop.
+   */
+  for (ix = 0; ix < pa; ix++) {
+    /* compute the outer product
+     *
+     * Note that every outer product is computed
+     * for a particular column only once which means that
+     * there is no need todo a double precision addition
+     * into the W2[] array.
+     */
+    W2[ix + ix] = ((mp_word)a->dp[ix]) * ((mp_word)a->dp[ix]);
+
+    {
+      register mp_digit tmpx, *tmpy;
+      register mp_word *_W;
+      register int iy;
+
+      /* copy of left side */
+      tmpx = a->dp[ix];
+
+      /* alias for right side */
+      tmpy = a->dp + (ix + 1);
+
+      /* the column to store the result in */
+      _W = W + (ix + ix + 1);
+
+      /* inner products */
+      for (iy = ix + 1; iy < pa; iy++) {
+          *_W++ += ((mp_word)tmpx) * ((mp_word)*tmpy++);
+      }
+    }
+  }
+
+  /* setup dest */
+  olduse  = b->used;
+  b->used = newused;
+
+  /* now compute digits
+   *
+   * We have to double the inner product sums, add in the
+   * outer product sums, propagate carries and convert
+   * to single precision.
+   */
+  {
+    register mp_digit *tmpb;
+
+    /* double first value, since the inner products are
+     * half of what they should be
+     */
+    W[0] += W[0] + W2[0];
+
+    tmpb = b->dp;
+    for (ix = 1; ix < newused; ix++) {
+      /* double/add next digit */
+      W[ix] += W[ix] + W2[ix];
+
+      /* propagate carry forwards [from the previous digit] */
+      W[ix] = W[ix] + (W[ix - 1] >> ((mp_word) DIGIT_BIT));
+
+      /* store the current digit now that the carry isn't
+       * needed
+       */
+      *tmpb++ = (mp_digit) (W[ix - 1] & ((mp_word) MP_MASK));
+    }
+    /* set the last value.  Note even if the carry is zero
+     * this is required since the next step will not zero
+     * it if b originally had a value at b->dp[2*a.used]
+     */
+    *tmpb++ = (mp_digit) (W[(newused) - 1] & ((mp_word) MP_MASK));
+
+    /* clear high digits of b if there were any originally */
+    for (; ix < olduse; ix++) {
+      *tmpb++ = 0;
+    }
+  }
+
+  mp_clamp (b);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_2expt.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,42 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* computes a = 2**b 
+ *
+ * Simple algorithm which zeroes the int, grows it then just sets one bit
+ * as required.
+ */
+int
+mp_2expt (mp_int * a, int b)
+{
+  int     res;
+
+  /* zero a as per default */
+  mp_zero (a);
+
+  /* grow a to accomodate the single bit */
+  if ((res = mp_grow (a, b / DIGIT_BIT + 1)) != MP_OKAY) {
+    return res;
+  }
+
+  /* set the used count of where the bit will go */
+  a->used = b / DIGIT_BIT + 1;
+
+  /* put the single bit in its place */
+  a->dp[b / DIGIT_BIT] = 1 << (b % DIGIT_BIT);
+
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_abs.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,37 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* b = |a| 
+ *
+ * Simple function copies the input and fixes the sign to positive
+ */
+int
+mp_abs (mp_int * a, mp_int * b)
+{
+  int     res;
+
+  /* copy a to b */
+  if (a != b) {
+     if ((res = mp_copy (a, b)) != MP_OKAY) {
+       return res;
+     }
+  }
+
+  /* force the sign of b to positive */
+  b->sign = MP_ZPOS;
+
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_add.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,47 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* high level addition (handles signs) */
+int mp_add (mp_int * a, mp_int * b, mp_int * c)
+{
+  int     sa, sb, res;
+
+  /* get sign of both inputs */
+  sa = a->sign;
+  sb = b->sign;
+
+  /* handle two cases, not four */
+  if (sa == sb) {
+    /* both positive or both negative */
+    /* add their magnitudes, copy the sign */
+    c->sign = sa;
+    res = s_mp_add (a, b, c);
+  } else {
+    /* one positive, the other negative */
+    /* subtract the one with the greater magnitude from */
+    /* the one of the lesser magnitude.  The result gets */
+    /* the sign of the one with the greater magnitude. */
+    if (mp_cmp_mag (a, b) == MP_LT) {
+      c->sign = sb;
+      res = s_mp_sub (b, a, c);
+    } else {
+      c->sign = sa;
+      res = s_mp_sub (a, b, c);
+    }
+  }
+  return res;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_add_d.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,103 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* single digit addition */
+int
+mp_add_d (mp_int * a, mp_digit b, mp_int * c)
+{
+  int     res, ix, oldused;
+  mp_digit *tmpa, *tmpc, mu;
+
+  /* grow c as required */
+  if (c->alloc < a->used + 1) {
+     if ((res = mp_grow(c, a->used + 1)) != MP_OKAY) {
+        return res;
+     }
+  }
+
+  /* if a is negative and |a| >= b, call c = |a| - b */
+  if (a->sign == MP_NEG && (a->used > 1 || a->dp[0] >= b)) {
+     /* temporarily fix sign of a */
+     a->sign = MP_ZPOS;
+
+     /* c = |a| - b */
+     res = mp_sub_d(a, b, c);
+
+     /* fix sign  */
+     a->sign = c->sign = MP_NEG;
+
+     return res;
+  }
+
+  /* old number of used digits in c */
+  oldused = c->used;
+
+  /* sign always positive */
+  c->sign = MP_ZPOS;
+
+  /* source alias */
+  tmpa    = a->dp;
+
+  /* destination alias */
+  tmpc    = c->dp;
+
+  /* if a is positive */
+  if (a->sign == MP_ZPOS) {
+     /* add digit, after this we're propagating
+      * the carry.
+      */
+     *tmpc   = *tmpa++ + b;
+     mu      = *tmpc >> DIGIT_BIT;
+     *tmpc++ &= MP_MASK;
+
+     /* now handle rest of the digits */
+     for (ix = 1; ix < a->used; ix++) {
+        *tmpc   = *tmpa++ + mu;
+        mu      = *tmpc >> DIGIT_BIT;
+        *tmpc++ &= MP_MASK;
+     }
+     /* set final carry */
+     ix++;
+     *tmpc++  = mu;
+
+     /* setup size */
+     c->used = a->used + 1;
+  } else {
+     /* a was negative and |a| < b */
+     c->used  = 1;
+
+     /* the result is a single digit */
+     if (a->used == 1) {
+        *tmpc++  =  b - a->dp[0];
+     } else {
+        *tmpc++  =  b;
+     }
+
+     /* setup count so the clearing of oldused
+      * can fall through correctly
+      */
+     ix       = 1;
+  }
+
+  /* now zero to oldused */
+  while (ix++ < oldused) {
+     *tmpc++ = 0;
+  }
+  mp_clamp(c);
+
+  return MP_OKAY;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_addmod.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,35 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* d = a + b (mod c) */
+int
+mp_addmod (mp_int * a, mp_int * b, mp_int * c, mp_int * d)
+{
+  int     res;
+  mp_int  t;
+
+  if ((res = mp_init (&t)) != MP_OKAY) {
+    return res;
+  }
+
+  if ((res = mp_add (a, b, &t)) != MP_OKAY) {
+    mp_clear (&t);
+    return res;
+  }
+  res = mp_mod (&t, c, d);
+  mp_clear (&t);
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_and.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,51 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* AND two ints together */
+int
+mp_and (mp_int * a, mp_int * b, mp_int * c)
+{
+  int     res, ix, px;
+  mp_int  t, *x;
+
+  if (a->used > b->used) {
+    if ((res = mp_init_copy (&t, a)) != MP_OKAY) {
+      return res;
+    }
+    px = b->used;
+    x = b;
+  } else {
+    if ((res = mp_init_copy (&t, b)) != MP_OKAY) {
+      return res;
+    }
+    px = a->used;
+    x = a;
+  }
+
+  for (ix = 0; ix < px; ix++) {
+    t.dp[ix] &= x->dp[ix];
+  }
+
+  /* zero digits above the last from the smallest mp_int */
+  for (; ix < t.used; ix++) {
+    t.dp[ix] = 0;
+  }
+
+  mp_clamp (&t);
+  mp_exch (c, &t);
+  mp_clear (&t);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_clamp.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,38 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* trim unused digits 
+ *
+ * This is used to ensure that leading zero digits are
+ * trimed and the leading "used" digit will be non-zero
+ * Typically very fast.  Also fixes the sign if there
+ * are no more leading digits
+ */
+void
+mp_clamp (mp_int * a)
+{
+  /* decrease used while the most significant digit is
+   * zero.
+   */
+  while (a->used > 0 && a->dp[a->used - 1] == 0) {
+    --(a->used);
+  }
+
+  /* reset the sign flag if used == 0 */
+  if (a->used == 0) {
+    a->sign = MP_ZPOS;
+  }
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_clear.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,34 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* clear one (frees)  */
+void
+mp_clear (mp_int * a)
+{
+  /* only do anything if a hasn't been freed previously */
+  if (a->dp != NULL) {
+    /* first zero the digits */
+    memset (a->dp, 0, sizeof (mp_digit) * a->used);
+
+    /* free ram */
+    XFREE(a->dp);
+
+    /* reset members to make debugging easier */
+    a->dp    = NULL;
+    a->alloc = a->used = 0;
+    a->sign  = MP_ZPOS;
+  }
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_clear_multi.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,28 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+#include <stdarg.h>
+
+void mp_clear_multi(mp_int *mp, ...) 
+{
+    mp_int* next_mp = mp;
+    va_list args;
+    va_start(args, mp);
+    while (next_mp != NULL) {
+        mp_clear(next_mp);
+        next_mp = va_arg(args, mp_int*);
+    }
+    va_end(args);
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_cmp.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,37 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* compare two ints (signed)*/
+int
+mp_cmp (mp_int * a, mp_int * b)
+{
+  /* compare based on sign */
+  if (a->sign != b->sign) {
+     if (a->sign == MP_NEG) {
+        return MP_LT;
+     } else {
+        return MP_GT;
+     }
+  }
+  
+  /* compare digits */
+  if (a->sign == MP_NEG) {
+     /* if negative compare opposite direction */
+     return mp_cmp_mag(b, a);
+  } else {
+     return mp_cmp_mag(a, b);
+  }
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_cmp_d.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,38 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* compare a digit */
+int mp_cmp_d(mp_int * a, mp_digit b)
+{
+  /* compare based on sign */
+  if (a->sign == MP_NEG) {
+    return MP_LT;
+  }
+
+  /* compare based on magnitude */
+  if (a->used > 1) {
+    return MP_GT;
+  }
+
+  /* compare the only digit of a to b */
+  if (a->dp[0] > b) {
+    return MP_GT;
+  } else if (a->dp[0] < b) {
+    return MP_LT;
+  } else {
+    return MP_EQ;
+  }
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_cmp_mag.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,49 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* compare maginitude of two ints (unsigned) */
+int mp_cmp_mag (mp_int * a, mp_int * b)
+{
+  int     n;
+  mp_digit *tmpa, *tmpb;
+
+  /* compare based on # of non-zero digits */
+  if (a->used > b->used) {
+    return MP_GT;
+  }
+  
+  if (a->used < b->used) {
+    return MP_LT;
+  }
+
+  /* alias for a */
+  tmpa = a->dp + (a->used - 1);
+
+  /* alias for b */
+  tmpb = b->dp + (a->used - 1);
+
+  /* compare based on digits  */
+  for (n = 0; n < a->used; ++n, --tmpa, --tmpb) {
+    if (*tmpa > *tmpb) {
+      return MP_GT;
+    }
+
+    if (*tmpa < *tmpb) {
+      return MP_LT;
+    }
+  }
+  return MP_EQ;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_cnt_lsb.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,47 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+static const int lnz[16] = { 
+   4, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0
+};
+
+/* Counts the number of lsbs which are zero before the first zero bit */
+int mp_cnt_lsb(mp_int *a)
+{
+   int x;
+   mp_digit q, qq;
+
+   /* easy out */
+   if (mp_iszero(a) == 1) {
+      return 0;
+   }
+
+   /* scan lower digits until non-zero */
+   for (x = 0; x < a->used && a->dp[x] == 0; x++);
+   q = a->dp[x];
+   x *= DIGIT_BIT;
+
+   /* now scan this digit until a 1 is found */
+   if ((q & 1) == 0) {
+      do {
+         qq  = q & 15;
+         x  += lnz[qq];
+         q >>= 4;
+      } while (qq == 0);
+   }
+   return x;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_copy.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,62 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* copy, b = a */
+int
+mp_copy (mp_int * a, mp_int * b)
+{
+  int     res, n;
+
+  /* if dst == src do nothing */
+  if (a == b) {
+    return MP_OKAY;
+  }
+
+  /* grow dest */
+  if (b->alloc < a->used) {
+     if ((res = mp_grow (b, a->used)) != MP_OKAY) {
+        return res;
+     }
+  }
+
+  /* zero b and copy the parameters over */
+  {
+    register mp_digit *tmpa, *tmpb;
+
+    /* pointer aliases */
+
+    /* source */
+    tmpa = a->dp;
+
+    /* destination */
+    tmpb = b->dp;
+
+    /* copy all the digits */
+    for (n = 0; n < a->used; n++) {
+      *tmpb++ = *tmpa++;
+    }
+
+    /* clear high digits */
+    for (; n < b->used; n++) {
+      *tmpb++ = 0;
+    }
+  }
+
+  /* copy used count and sign */
+  b->used = a->used;
+  b->sign = a->sign;
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_count_bits.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,39 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* returns the number of bits in an int */
+int
+mp_count_bits (mp_int * a)
+{
+  int     r;
+  mp_digit q;
+
+  /* shortcut */
+  if (a->used == 0) {
+    return 0;
+  }
+
+  /* get number of digits and add that */
+  r = (a->used - 1) * DIGIT_BIT;
+  
+  /* take the last digit and count the bits in it */
+  q = a->dp[a->used - 1];
+  while (q > ((mp_digit) 0)) {
+    ++r;
+    q >>= ((mp_digit) 1);
+  }
+  return r;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_div.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,211 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* integer signed division. 
+ * c*b + d == a [e.g. a/b, c=quotient, d=remainder]
+ * HAC pp.598 Algorithm 14.20
+ *
+ * Note that the description in HAC is horribly 
+ * incomplete.  For example, it doesn't consider 
+ * the case where digits are removed from 'x' in 
+ * the inner loop.  It also doesn't consider the 
+ * case that y has fewer than three digits, etc..
+ *
+ * The overall algorithm is as described as 
+ * 14.20 from HAC but fixed to treat these cases.
+*/
+int mp_div (mp_int * a, mp_int * b, mp_int * c, mp_int * d)
+{
+  mp_int  q, x, y, t1, t2;
+  int     res, n, t, i, norm, neg;
+
+  /* is divisor zero ? */
+  if (mp_iszero (b) == 1) {
+    return MP_VAL;
+  }
+
+  /* if a < b then q=0, r = a */
+  if (mp_cmp_mag (a, b) == MP_LT) {
+    if (d != NULL) {
+      res = mp_copy (a, d);
+    } else {
+      res = MP_OKAY;
+    }
+    if (c != NULL) {
+      mp_zero (c);
+    }
+    return res;
+  }
+
+  if ((res = mp_init_size (&q, a->used + 2)) != MP_OKAY) {
+    return res;
+  }
+  q.used = a->used + 2;
+
+  if ((res = mp_init (&t1)) != MP_OKAY) {
+    goto __Q;
+  }
+
+  if ((res = mp_init (&t2)) != MP_OKAY) {
+    goto __T1;
+  }
+
+  if ((res = mp_init_copy (&x, a)) != MP_OKAY) {
+    goto __T2;
+  }
+
+  if ((res = mp_init_copy (&y, b)) != MP_OKAY) {
+    goto __X;
+  }
+
+  /* fix the sign */
+  neg = (a->sign == b->sign) ? MP_ZPOS : MP_NEG;
+  x.sign = y.sign = MP_ZPOS;
+
+  /* normalize both x and y, ensure that y >= b/2, [b == 2**DIGIT_BIT] */
+  norm = mp_count_bits(&y) % DIGIT_BIT;
+  if (norm < (int)(DIGIT_BIT-1)) {
+     norm = (DIGIT_BIT-1) - norm;
+     if ((res = mp_mul_2d (&x, norm, &x)) != MP_OKAY) {
+       goto __Y;
+     }
+     if ((res = mp_mul_2d (&y, norm, &y)) != MP_OKAY) {
+       goto __Y;
+     }
+  } else {
+     norm = 0;
+  }
+
+  /* note hac does 0 based, so if used==5 then its 0,1,2,3,4, e.g. use 4 */
+  n = x.used - 1;
+  t = y.used - 1;
+
+  /* while (x >= y*b**n-t) do { q[n-t] += 1; x -= y*b**{n-t} } */
+  if ((res = mp_lshd (&y, n - t)) != MP_OKAY) { /* y = y*b**{n-t} */
+    goto __Y;
+  }
+
+  while (mp_cmp (&x, &y) != MP_LT) {
+    ++(q.dp[n - t]);
+    if ((res = mp_sub (&x, &y, &x)) != MP_OKAY) {
+      goto __Y;
+    }
+  }
+
+  /* reset y by shifting it back down */
+  mp_rshd (&y, n - t);
+
+  /* step 3. for i from n down to (t + 1) */
+  for (i = n; i >= (t + 1); i--) {
+    if (i > x.used) {
+      continue;
+    }
+
+    /* step 3.1 if xi == yt then set q{i-t-1} to b-1, 
+     * otherwise set q{i-t-1} to (xi*b + x{i-1})/yt */
+    if (x.dp[i] == y.dp[t]) {
+      q.dp[i - t - 1] = ((((mp_digit)1) << DIGIT_BIT) - 1);
+    } else {
+      mp_word tmp;
+      tmp = ((mp_word) x.dp[i]) << ((mp_word) DIGIT_BIT);
+      tmp |= ((mp_word) x.dp[i - 1]);
+      tmp /= ((mp_word) y.dp[t]);
+      if (tmp > (mp_word) MP_MASK)
+        tmp = MP_MASK;
+      q.dp[i - t - 1] = (mp_digit) (tmp & (mp_word) (MP_MASK));
+    }
+
+    /* while (q{i-t-1} * (yt * b + y{t-1})) > 
+             xi * b**2 + xi-1 * b + xi-2 
+     
+       do q{i-t-1} -= 1; 
+    */
+    q.dp[i - t - 1] = (q.dp[i - t - 1] + 1) & MP_MASK;
+    do {
+      q.dp[i - t - 1] = (q.dp[i - t - 1] - 1) & MP_MASK;
+
+      /* find left hand */
+      mp_zero (&t1);
+      t1.dp[0] = (t - 1 < 0) ? 0 : y.dp[t - 1];
+      t1.dp[1] = y.dp[t];
+      t1.used = 2;
+      if ((res = mp_mul_d (&t1, q.dp[i - t - 1], &t1)) != MP_OKAY) {
+        goto __Y;
+      }
+
+      /* find right hand */
+      t2.dp[0] = (i - 2 < 0) ? 0 : x.dp[i - 2];
+      t2.dp[1] = (i - 1 < 0) ? 0 : x.dp[i - 1];
+      t2.dp[2] = x.dp[i];
+      t2.used = 3;
+    } while (mp_cmp_mag(&t1, &t2) == MP_GT);
+
+    /* step 3.3 x = x - q{i-t-1} * y * b**{i-t-1} */
+    if ((res = mp_mul_d (&y, q.dp[i - t - 1], &t1)) != MP_OKAY) {
+      goto __Y;
+    }
+
+    if ((res = mp_lshd (&t1, i - t - 1)) != MP_OKAY) {
+      goto __Y;
+    }
+
+    if ((res = mp_sub (&x, &t1, &x)) != MP_OKAY) {
+      goto __Y;
+    }
+
+    /* if x < 0 then { x = x + y*b**{i-t-1}; q{i-t-1} -= 1; } */
+    if (x.sign == MP_NEG) {
+      if ((res = mp_copy (&y, &t1)) != MP_OKAY) {
+        goto __Y;
+      }
+      if ((res = mp_lshd (&t1, i - t - 1)) != MP_OKAY) {
+        goto __Y;
+      }
+      if ((res = mp_add (&x, &t1, &x)) != MP_OKAY) {
+        goto __Y;
+      }
+
+      q.dp[i - t - 1] = (q.dp[i - t - 1] - 1UL) & MP_MASK;
+    }
+  }
+
+  /* now q is the quotient and x is the remainder 
+   * [which we have to normalize] 
+   */
+  
+  /* get sign before writing to c */
+  x.sign = a->sign;
+
+  if (c != NULL) {
+    mp_clamp (&q);
+    mp_exch (&q, c);
+    c->sign = neg;
+  }
+
+  if (d != NULL) {
+    mp_div_2d (&x, norm, &x, NULL);
+    mp_exch (&x, d);
+  }
+
+  res = MP_OKAY;
+
+__Y:mp_clear (&y);
+__X:mp_clear (&x);
+__T2:mp_clear (&t2);
+__T1:mp_clear (&t1);
+__Q:mp_clear (&q);
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_div_2.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,62 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* b = a/2 */
+int mp_div_2(mp_int * a, mp_int * b)
+{
+  int     x, res, oldused;
+
+  /* copy */
+  if (b->alloc < a->used) {
+    if ((res = mp_grow (b, a->used)) != MP_OKAY) {
+      return res;
+    }
+  }
+
+  oldused = b->used;
+  b->used = a->used;
+  {
+    register mp_digit r, rr, *tmpa, *tmpb;
+
+    /* source alias */
+    tmpa = a->dp + b->used - 1;
+
+    /* dest alias */
+    tmpb = b->dp + b->used - 1;
+
+    /* carry */
+    r = 0;
+    for (x = b->used - 1; x >= 0; x--) {
+      /* get the carry for the next iteration */
+      rr = *tmpa & 1;
+
+      /* shift the current digit, add in carry and store */
+      *tmpb-- = (*tmpa-- >> 1) | (r << (DIGIT_BIT - 1));
+
+      /* forward carry to next iteration */
+      r = rr;
+    }
+
+    /* zero excess digits */
+    tmpb = b->dp + b->used;
+    for (x = b->used; x < oldused; x++) {
+      *tmpb++ = 0;
+    }
+  }
+  b->sign = a->sign;
+  mp_clamp (b);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_div_2d.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,91 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* shift right by a certain bit count (store quotient in c, optional remainder in d) */
+int mp_div_2d (mp_int * a, int b, mp_int * c, mp_int * d)
+{
+  mp_digit D, r, rr;
+  int     x, res;
+  mp_int  t;
+
+
+  /* if the shift count is <= 0 then we do no work */
+  if (b <= 0) {
+    res = mp_copy (a, c);
+    if (d != NULL) {
+      mp_zero (d);
+    }
+    return res;
+  }
+
+  if ((res = mp_init (&t)) != MP_OKAY) {
+    return res;
+  }
+
+  /* get the remainder */
+  if (d != NULL) {
+    if ((res = mp_mod_2d (a, b, &t)) != MP_OKAY) {
+      mp_clear (&t);
+      return res;
+    }
+  }
+
+  /* copy */
+  if ((res = mp_copy (a, c)) != MP_OKAY) {
+    mp_clear (&t);
+    return res;
+  }
+
+  /* shift by as many digits in the bit count */
+  if (b >= (int)DIGIT_BIT) {
+    mp_rshd (c, b / DIGIT_BIT);
+  }
+
+  /* shift any bit count < DIGIT_BIT */
+  D = (mp_digit) (b % DIGIT_BIT);
+  if (D != 0) {
+    register mp_digit *tmpc, mask, shift;
+
+    /* mask */
+    mask = (((mp_digit)1) << D) - 1;
+
+    /* shift for lsb */
+    shift = DIGIT_BIT - D;
+
+    /* alias */
+    tmpc = c->dp + (c->used - 1);
+
+    /* carry */
+    r = 0;
+    for (x = c->used - 1; x >= 0; x--) {
+      /* get the lower  bits of this word in a temp */
+      rr = *tmpc & mask;
+
+      /* shift the current word and mix in the carry bits from the previous word */
+      *tmpc = (*tmpc >> D) | (r << shift);
+      --tmpc;
+
+      /* set the carry to the carry bits of the current word found above */
+      r = rr;
+    }
+  }
+  mp_clamp (c);
+  if (d != NULL) {
+    mp_exch (&t, d);
+  }
+  mp_clear (&t);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_div_3.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,73 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* divide by three (based on routine from MPI and the GMP manual) */
+int
+mp_div_3 (mp_int * a, mp_int *c, mp_digit * d)
+{
+  mp_int   q;
+  mp_word  w, t;
+  mp_digit b;
+  int      res, ix;
+  
+  /* b = 2**DIGIT_BIT / 3 */
+  b = (((mp_word)1) << ((mp_word)DIGIT_BIT)) / ((mp_word)3);
+
+  if ((res = mp_init_size(&q, a->used)) != MP_OKAY) {
+     return res;
+  }
+  
+  q.used = a->used;
+  q.sign = a->sign;
+  w = 0;
+  for (ix = a->used - 1; ix >= 0; ix--) {
+     w = (w << ((mp_word)DIGIT_BIT)) | ((mp_word)a->dp[ix]);
+
+     if (w >= 3) {
+        /* multiply w by [1/3] */
+        t = (w * ((mp_word)b)) >> ((mp_word)DIGIT_BIT);
+
+        /* now subtract 3 * [w/3] from w, to get the remainder */
+        w -= t+t+t;
+
+        /* fixup the remainder as required since
+         * the optimization is not exact.
+         */
+        while (w >= 3) {
+           t += 1;
+           w -= 3;
+        }
+      } else {
+        t = 0;
+      }
+      q.dp[ix] = (mp_digit)t;
+  }
+
+  /* [optional] store the remainder */
+  if (d != NULL) {
+     *d = (mp_digit)w;
+  }
+
+  /* [optional] store the quotient */
+  if (c != NULL) {
+     mp_clamp(&q);
+     mp_exch(&q, c);
+  }
+  mp_clear(&q);
+  
+  return res;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_div_d.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,102 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+static int s_is_power_of_two(mp_digit b, int *p)
+{
+   int x;
+
+   for (x = 1; x < DIGIT_BIT; x++) {
+      if (b == (((mp_digit)1)<<x)) {
+         *p = x;
+         return 1;
+      }
+   }
+   return 0;
+}
+
+/* single digit division (based on routine from MPI) */
+int mp_div_d (mp_int * a, mp_digit b, mp_int * c, mp_digit * d)
+{
+  mp_int  q;
+  mp_word w;
+  mp_digit t;
+  int     res, ix;
+
+  /* cannot divide by zero */
+  if (b == 0) {
+     return MP_VAL;
+  }
+
+  /* quick outs */
+  if (b == 1 || mp_iszero(a) == 1) {
+     if (d != NULL) {
+        *d = 0;
+     }
+     if (c != NULL) {
+        return mp_copy(a, c);
+     }
+     return MP_OKAY;
+  }
+
+  /* power of two ? */
+  if (s_is_power_of_two(b, &ix) == 1) {
+     if (d != NULL) {
+        *d = a->dp[0] & ((1<<ix) - 1);
+     }
+     if (c != NULL) {
+        return mp_div_2d(a, ix, c, NULL);
+     }
+     return MP_OKAY;
+  }
+
+  /* three? */
+  if (b == 3) {
+     return mp_div_3(a, c, d);
+  }
+
+  /* no easy answer [c'est la vie].  Just division */
+  if ((res = mp_init_size(&q, a->used)) != MP_OKAY) {
+     return res;
+  }
+  
+  q.used = a->used;
+  q.sign = a->sign;
+  w = 0;
+  for (ix = a->used - 1; ix >= 0; ix--) {
+     w = (w << ((mp_word)DIGIT_BIT)) | ((mp_word)a->dp[ix]);
+     
+     if (w >= b) {
+        t = (mp_digit)(w / b);
+        w -= ((mp_word)t) * ((mp_word)b);
+      } else {
+        t = 0;
+      }
+      q.dp[ix] = (mp_digit)t;
+  }
+  
+  if (d != NULL) {
+     *d = (mp_digit)w;
+  }
+  
+  if (c != NULL) {
+     mp_clamp(&q);
+     mp_exch(&q, c);
+  }
+  mp_clear(&q);
+  
+  return res;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_dr_is_modulus.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,37 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* determines if a number is a valid DR modulus */
+int mp_dr_is_modulus(mp_int *a)
+{
+   int ix;
+
+   /* must be at least two digits */
+   if (a->used < 2) {
+      return 0;
+   }
+
+   /* must be of the form b**k - a [a <= b] so all
+    * but the first digit must be equal to -1 (mod b).
+    */
+   for (ix = 1; ix < a->used; ix++) {
+       if (a->dp[ix] != MP_MASK) {
+          return 0;
+       }
+   }
+   return 1;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_dr_reduce.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,88 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* reduce "x" in place modulo "n" using the Diminished Radix algorithm.
+ *
+ * Based on algorithm from the paper
+ *
+ * "Generating Efficient Primes for Discrete Log Cryptosystems"
+ *                 Chae Hoon Lim, Pil Loong Lee,
+ *          POSTECH Information Research Laboratories
+ *
+ * The modulus must be of a special format [see manual]
+ *
+ * Has been modified to use algorithm 7.10 from the LTM book instead
+ *
+ * Input x must be in the range 0 <= x <= (n-1)**2
+ */
+int
+mp_dr_reduce (mp_int * x, mp_int * n, mp_digit k)
+{
+  int      err, i, m;
+  mp_word  r;
+  mp_digit mu, *tmpx1, *tmpx2;
+
+  /* m = digits in modulus */
+  m = n->used;
+
+  /* ensure that "x" has at least 2m digits */
+  if (x->alloc < m + m) {
+    if ((err = mp_grow (x, m + m)) != MP_OKAY) {
+      return err;
+    }
+  }
+
+/* top of loop, this is where the code resumes if
+ * another reduction pass is required.
+ */
+top:
+  /* aliases for digits */
+  /* alias for lower half of x */
+  tmpx1 = x->dp;
+
+  /* alias for upper half of x, or x/B**m */
+  tmpx2 = x->dp + m;
+
+  /* set carry to zero */
+  mu = 0;
+
+  /* compute (x mod B**m) + k * [x/B**m] inline and inplace */
+  for (i = 0; i < m; i++) {
+      r         = ((mp_word)*tmpx2++) * ((mp_word)k) + *tmpx1 + mu;
+      *tmpx1++  = (mp_digit)(r & MP_MASK);
+      mu        = (mp_digit)(r >> ((mp_word)DIGIT_BIT));
+  }
+
+  /* set final carry */
+  *tmpx1++ = mu;
+
+  /* zero words above m */
+  for (i = m + 1; i < x->used; i++) {
+      *tmpx1++ = 0;
+  }
+
+  /* clamp, sub and return */
+  mp_clamp (x);
+
+  /* if x >= n then subtract and reduce again
+   * Each successive "recursion" makes the input smaller and smaller.
+   */
+  if (mp_cmp_mag (x, n) != MP_LT) {
+    s_mp_sub(x, n, x);
+    goto top;
+  }
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_dr_setup.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,26 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* determines the setup value */
+void mp_dr_setup(mp_int *a, mp_digit *d)
+{
+   /* the casts are required if DIGIT_BIT is one less than
+    * the number of bits in a mp_digit [e.g. DIGIT_BIT==31]
+    */
+   *d = (mp_digit)((((mp_word)1) << ((mp_word)DIGIT_BIT)) - 
+        ((mp_word)a->dp[0]));
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_exch.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,28 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* swap the elements of two integers, for cases where you can't simply swap the 
+ * mp_int pointers around
+ */
+void
+mp_exch (mp_int * a, mp_int * b)
+{
+  mp_int  t;
+
+  t  = *a;
+  *a = *b;
+  *b = t;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_expt_d.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,51 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* calculate c = a**b  using a square-multiply algorithm */
+int mp_expt_d (mp_int * a, mp_digit b, mp_int * c)
+{
+  int     res, x;
+  mp_int  g;
+
+  if ((res = mp_init_copy (&g, a)) != MP_OKAY) {
+    return res;
+  }
+
+  /* set initial result */
+  mp_set (c, 1);
+
+  for (x = 0; x < (int) DIGIT_BIT; x++) {
+    /* square */
+    if ((res = mp_sqr (c, c)) != MP_OKAY) {
+      mp_clear (&g);
+      return res;
+    }
+
+    /* if the bit is set multiply */
+    if ((b & (mp_digit) (((mp_digit)1) << (DIGIT_BIT - 1))) != 0) {
+      if ((res = mp_mul (c, &g, c)) != MP_OKAY) {
+         mp_clear (&g);
+         return res;
+      }
+    }
+
+    /* shift to next bit */
+    b <<= 1;
+  }
+
+  mp_clear (&g);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_exptmod.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,78 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+
+/* this is a shell function that calls either the normal or Montgomery
+ * exptmod functions.  Originally the call to the montgomery code was
+ * embedded in the normal function but that wasted alot of stack space
+ * for nothing (since 99% of the time the Montgomery code would be called)
+ */
+int mp_exptmod (mp_int * G, mp_int * X, mp_int * P, mp_int * Y)
+{
+  int dr;
+
+  /* modulus P must be positive */
+  if (P->sign == MP_NEG) {
+     return MP_VAL;
+  }
+
+  /* if exponent X is negative we have to recurse */
+  if (X->sign == MP_NEG) {
+     mp_int tmpG, tmpX;
+     int err;
+
+     /* first compute 1/G mod P */
+     if ((err = mp_init(&tmpG)) != MP_OKAY) {
+        return err;
+     }
+     if ((err = mp_invmod(G, P, &tmpG)) != MP_OKAY) {
+        mp_clear(&tmpG);
+        return err;
+     }
+
+     /* now get |X| */
+     if ((err = mp_init(&tmpX)) != MP_OKAY) {
+        mp_clear(&tmpG);
+        return err;
+     }
+     if ((err = mp_abs(X, &tmpX)) != MP_OKAY) {
+        mp_clear_multi(&tmpG, &tmpX, NULL);
+        return err;
+     }
+
+     /* and now compute (1/G)**|X| instead of G**X [X < 0] */
+     err = mp_exptmod(&tmpG, &tmpX, P, Y);
+     mp_clear_multi(&tmpG, &tmpX, NULL);
+     return err;
+  }
+
+  /* is it a DR modulus? */
+  dr = mp_dr_is_modulus(P);
+
+  /* if not, is it a uDR modulus? */
+  if (dr == 0) {
+     dr = mp_reduce_is_2k(P) << 1;
+  }
+    
+  /* if the modulus is odd or dr != 0 use the fast method */
+  if (mp_isodd (P) == 1 || dr !=  0) {
+    return mp_exptmod_fast (G, X, P, Y, dr);
+  } else {
+    /* otherwise use the generic Barrett reduction technique */
+    return s_mp_exptmod (G, X, P, Y);
+  }
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_exptmod_fast.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,287 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* computes Y == G**X mod P, HAC pp.616, Algorithm 14.85
+ *
+ * Uses a left-to-right k-ary sliding window to compute the modular exponentiation.
+ * The value of k changes based on the size of the exponent.
+ *
+ * Uses Montgomery or Diminished Radix reduction [whichever appropriate]
+ */
+
+#ifdef MP_LOW_MEM
+   #define TAB_SIZE 32
+#else
+   #define TAB_SIZE 256
+#endif
+
+int
+mp_exptmod_fast (mp_int * G, mp_int * X, mp_int * P, mp_int * Y, int redmode)
+{
+  mp_int  M[TAB_SIZE], res;
+  mp_digit buf, mp;
+  int     err, bitbuf, bitcpy, bitcnt, mode, digidx, x, y, winsize;
+
+  /* use a pointer to the reduction algorithm.  This allows us to use
+   * one of many reduction algorithms without modding the guts of
+   * the code with if statements everywhere.
+   */
+  int     (*redux)(mp_int*,mp_int*,mp_digit);
+
+  /* find window size */
+  x = mp_count_bits (X);
+  if (x <= 7) {
+    winsize = 2;
+  } else if (x <= 36) {
+    winsize = 3;
+  } else if (x <= 140) {
+    winsize = 4;
+  } else if (x <= 450) {
+    winsize = 5;
+  } else if (x <= 1303) {
+    winsize = 6;
+  } else if (x <= 3529) {
+    winsize = 7;
+  } else {
+    winsize = 8;
+  }
+
+#ifdef MP_LOW_MEM
+  if (winsize > 5) {
+     winsize = 5;
+  }
+#endif
+
+  /* init M array */
+  /* init first cell */
+  if ((err = mp_init(&M[1])) != MP_OKAY) {
+     return err;
+  }
+
+  /* now init the second half of the array */
+  for (x = 1<<(winsize-1); x < (1 << winsize); x++) {
+    if ((err = mp_init(&M[x])) != MP_OKAY) {
+      for (y = 1<<(winsize-1); y < x; y++) {
+        mp_clear (&M[y]);
+      }
+      mp_clear(&M[1]);
+      return err;
+    }
+  }
+
+  /* determine and setup reduction code */
+  if (redmode == 0) {
+     /* now setup montgomery  */
+     if ((err = mp_montgomery_setup (P, &mp)) != MP_OKAY) {
+        goto __M;
+     }
+
+     /* automatically pick the comba one if available (saves quite a few calls/ifs) */
+     if (((P->used * 2 + 1) < MP_WARRAY) &&
+          P->used < (1 << ((CHAR_BIT * sizeof (mp_word)) - (2 * DIGIT_BIT)))) {
+        redux = fast_mp_montgomery_reduce;
+     } else {
+        /* use slower baseline Montgomery method */
+        redux = mp_montgomery_reduce;
+     }
+  } else if (redmode == 1) {
+     /* setup DR reduction for moduli of the form B**k - b */
+     mp_dr_setup(P, &mp);
+     redux = mp_dr_reduce;
+  } else {
+     /* setup DR reduction for moduli of the form 2**k - b */
+     if ((err = mp_reduce_2k_setup(P, &mp)) != MP_OKAY) {
+        goto __M;
+     }
+     redux = mp_reduce_2k;
+  }
+
+  /* setup result */
+  if ((err = mp_init (&res)) != MP_OKAY) {
+    goto __M;
+  }
+
+  /* create M table
+   *
+   * The M table contains powers of the input base, e.g. M[x] = G^x mod P
+   *
+   * The first half of the table is not computed though accept for M[0] and M[1]
+   */
+
+  if (redmode == 0) {
+     /* now we need R mod m */
+     if ((err = mp_montgomery_calc_normalization (&res, P)) != MP_OKAY) {
+       goto __RES;
+     }
+
+     /* now set M[1] to G * R mod m */
+     if ((err = mp_mulmod (G, &res, P, &M[1])) != MP_OKAY) {
+       goto __RES;
+     }
+  } else {
+     mp_set(&res, 1);
+     if ((err = mp_mod(G, P, &M[1])) != MP_OKAY) {
+        goto __RES;
+     }
+  }
+
+  /* compute the value at M[1<<(winsize-1)] by squaring M[1] (winsize-1) times */
+  if ((err = mp_copy (&M[1], &M[1 << (winsize - 1)])) != MP_OKAY) {
+    goto __RES;
+  }
+
+  for (x = 0; x < (winsize - 1); x++) {
+    if ((err = mp_sqr (&M[1 << (winsize - 1)], &M[1 << (winsize - 1)])) != MP_OKAY) {
+      goto __RES;
+    }
+    if ((err = redux (&M[1 << (winsize - 1)], P, mp)) != MP_OKAY) {
+      goto __RES;
+    }
+  }
+
+  /* create upper table */
+  for (x = (1 << (winsize - 1)) + 1; x < (1 << winsize); x++) {
+    if ((err = mp_mul (&M[x - 1], &M[1], &M[x])) != MP_OKAY) {
+      goto __RES;
+    }
+    if ((err = redux (&M[x], P, mp)) != MP_OKAY) {
+      goto __RES;
+    }
+  }
+
+  /* set initial mode and bit cnt */
+  mode   = 0;
+  bitcnt = 1;
+  buf    = 0;
+  digidx = X->used - 1;
+  bitcpy = 0;
+  bitbuf = 0;
+
+  for (;;) {
+    /* grab next digit as required */
+    if (--bitcnt == 0) {
+      /* if digidx == -1 we are out of digits so break */
+      if (digidx == -1) {
+        break;
+      }
+      /* read next digit and reset bitcnt */
+      buf    = X->dp[digidx--];
+      bitcnt = (int)DIGIT_BIT;
+    }
+
+    /* grab the next msb from the exponent */
+    y     = (mp_digit)(buf >> (DIGIT_BIT - 1)) & 1;
+    buf <<= (mp_digit)1;
+
+    /* if the bit is zero and mode == 0 then we ignore it
+     * These represent the leading zero bits before the first 1 bit
+     * in the exponent.  Technically this opt is not required but it
+     * does lower the # of trivial squaring/reductions used
+     */
+    if (mode == 0 && y == 0) {
+      continue;
+    }
+
+    /* if the bit is zero and mode == 1 then we square */
+    if (mode == 1 && y == 0) {
+      if ((err = mp_sqr (&res, &res)) != MP_OKAY) {
+        goto __RES;
+      }
+      if ((err = redux (&res, P, mp)) != MP_OKAY) {
+        goto __RES;
+      }
+      continue;
+    }
+
+    /* else we add it to the window */
+    bitbuf |= (y << (winsize - ++bitcpy));
+    mode    = 2;
+
+    if (bitcpy == winsize) {
+      /* ok window is filled so square as required and multiply  */
+      /* square first */
+      for (x = 0; x < winsize; x++) {
+        if ((err = mp_sqr (&res, &res)) != MP_OKAY) {
+          goto __RES;
+        }
+        if ((err = redux (&res, P, mp)) != MP_OKAY) {
+          goto __RES;
+        }
+      }
+
+      /* then multiply */
+      if ((err = mp_mul (&res, &M[bitbuf], &res)) != MP_OKAY) {
+        goto __RES;
+      }
+      if ((err = redux (&res, P, mp)) != MP_OKAY) {
+        goto __RES;
+      }
+
+      /* empty window and reset */
+      bitcpy = 0;
+      bitbuf = 0;
+      mode   = 1;
+    }
+  }
+
+  /* if bits remain then square/multiply */
+  if (mode == 2 && bitcpy > 0) {
+    /* square then multiply if the bit is set */
+    for (x = 0; x < bitcpy; x++) {
+      if ((err = mp_sqr (&res, &res)) != MP_OKAY) {
+        goto __RES;
+      }
+      if ((err = redux (&res, P, mp)) != MP_OKAY) {
+        goto __RES;
+      }
+
+      /* get next bit of the window */
+      bitbuf <<= 1;
+      if ((bitbuf & (1 << winsize)) != 0) {
+        /* then multiply */
+        if ((err = mp_mul (&res, &M[1], &res)) != MP_OKAY) {
+          goto __RES;
+        }
+        if ((err = redux (&res, P, mp)) != MP_OKAY) {
+          goto __RES;
+        }
+      }
+    }
+  }
+
+  if (redmode == 0) {
+     /* fixup result if Montgomery reduction is used
+      * recall that any value in a Montgomery system is
+      * actually multiplied by R mod n.  So we have
+      * to reduce one more time to cancel out the factor
+      * of R.
+      */
+     if ((err = mp_montgomery_reduce (&res, P, mp)) != MP_OKAY) {
+       goto __RES;
+     }
+  }
+
+  /* swap res with Y */
+  mp_exch (&res, Y);
+  err = MP_OKAY;
+__RES:mp_clear (&res);
+__M:
+  mp_clear(&M[1]);
+  for (x = 1<<(winsize-1); x < (1 << winsize); x++) {
+    mp_clear (&M[x]);
+  }
+  return err;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_exteuclid.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,69 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* Extended euclidean algorithm of (a, b) produces 
+   a*u1 + b*u2 = u3
+ */
+int mp_exteuclid(mp_int *a, mp_int *b, mp_int *U1, mp_int *U2, mp_int *U3)
+{
+   mp_int u1,u2,u3,v1,v2,v3,t1,t2,t3,q,tmp;
+   int err;
+
+   if ((err = mp_init_multi(&u1, &u2, &u3, &v1, &v2, &v3, &t1, &t2, &t3, &q, &tmp, NULL)) != MP_OKAY) {
+      return err;
+   }
+
+   /* initialize, (u1,u2,u3) = (1,0,a) */
+   mp_set(&u1, 1);
+   if ((err = mp_copy(a, &u3)) != MP_OKAY)                                        { goto _ERR; }
+
+   /* initialize, (v1,v2,v3) = (0,1,b) */
+   mp_set(&v2, 1);
+   if ((err = mp_copy(b, &v3)) != MP_OKAY)                                        { goto _ERR; }
+
+   /* loop while v3 != 0 */
+   while (mp_iszero(&v3) == MP_NO) {
+       /* q = u3/v3 */
+       if ((err = mp_div(&u3, &v3, &q, NULL)) != MP_OKAY)                         { goto _ERR; }
+
+       /* (t1,t2,t3) = (u1,u2,u3) - (v1,v2,v3)q */
+       if ((err = mp_mul(&v1, &q, &tmp)) != MP_OKAY)                              { goto _ERR; }
+       if ((err = mp_sub(&u1, &tmp, &t1)) != MP_OKAY)                             { goto _ERR; }
+       if ((err = mp_mul(&v2, &q, &tmp)) != MP_OKAY)                              { goto _ERR; }
+       if ((err = mp_sub(&u2, &tmp, &t2)) != MP_OKAY)                             { goto _ERR; }
+       if ((err = mp_mul(&v3, &q, &tmp)) != MP_OKAY)                              { goto _ERR; }
+       if ((err = mp_sub(&u3, &tmp, &t3)) != MP_OKAY)                             { goto _ERR; }
+
+       /* (u1,u2,u3) = (v1,v2,v3) */
+       if ((err = mp_copy(&v1, &u1)) != MP_OKAY)                                  { goto _ERR; }
+       if ((err = mp_copy(&v2, &u2)) != MP_OKAY)                                  { goto _ERR; }
+       if ((err = mp_copy(&v3, &u3)) != MP_OKAY)                                  { goto _ERR; }
+
+       /* (v1,v2,v3) = (t1,t2,t3) */
+       if ((err = mp_copy(&t1, &v1)) != MP_OKAY)                                  { goto _ERR; }
+       if ((err = mp_copy(&t2, &v2)) != MP_OKAY)                                  { goto _ERR; }
+       if ((err = mp_copy(&t3, &v3)) != MP_OKAY)                                  { goto _ERR; }
+   }
+
+   /* copy result out */
+   if (U1 != NULL) { mp_exch(U1, &u1); }
+   if (U2 != NULL) { mp_exch(U2, &u2); }
+   if (U3 != NULL) { mp_exch(U3, &u3); }
+
+   err = MP_OKAY;
+_ERR: mp_clear_multi(&u1, &u2, &u3, &v1, &v2, &v3, &t1, &t2, &t3, &q, &tmp, NULL);
+   return err;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_fread.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,61 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* read a bigint from a file stream in ASCII */
+int mp_fread(mp_int *a, int radix, FILE *stream)
+{
+   int err, ch, neg, y;
+   
+   /* clear a */
+   mp_zero(a);
+   
+   /* if first digit is - then set negative */
+   ch = fgetc(stream);
+   if (ch == '-') {
+      neg = MP_NEG;
+      ch = fgetc(stream);
+   } else {
+      neg = MP_ZPOS;
+   }
+   
+   for (;;) {
+      /* find y in the radix map */
+      for (y = 0; y < radix; y++) {
+          if (mp_s_rmap[y] == ch) {
+             break;
+          }
+      }
+      if (y == radix) {
+         break;
+      }
+      
+      /* shift up and add */
+      if ((err = mp_mul_d(a, radix, a)) != MP_OKAY) {
+         return err;
+      }
+      if ((err = mp_add_d(a, y, a)) != MP_OKAY) {
+         return err;
+      }
+      
+      ch = fgetc(stream);
+   }
+   if (mp_cmp_d(a, 0) != MP_EQ) {
+      a->sign = neg;
+   }
+   
+   return MP_OKAY;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_fwrite.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,46 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+int mp_fwrite(mp_int *a, int radix, FILE *stream)
+{
+   char *buf;
+   int err, len, x;
+   
+   if ((err = mp_radix_size(a, radix, &len)) != MP_OKAY) {
+      return err;
+   }
+
+   buf = OPT_CAST(char) XMALLOC (len);
+   if (buf == NULL) {
+      return MP_MEM;
+   }
+   
+   if ((err = mp_toradix(a, buf, radix)) != MP_OKAY) {
+      XFREE (buf);
+      return err;
+   }
+   
+   for (x = 0; x < len; x++) {
+       if (fputc(buf[x], stream) == EOF) {
+          XFREE (buf);
+          return MP_VAL;
+       }
+   }
+   
+   XFREE (buf);
+   return MP_OKAY;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_gcd.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,107 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* Greatest Common Divisor using the binary method */
+int mp_gcd (mp_int * a, mp_int * b, mp_int * c)
+{
+  mp_int  u, v;
+  int     k, u_lsb, v_lsb, res;
+
+  /* either zero than gcd is the largest */
+  if (mp_iszero (a) == 1 && mp_iszero (b) == 0) {
+    return mp_abs (b, c);
+  }
+  if (mp_iszero (a) == 0 && mp_iszero (b) == 1) {
+    return mp_abs (a, c);
+  }
+
+  /* optimized.  At this point if a == 0 then
+   * b must equal zero too
+   */
+  if (mp_iszero (a) == 1) {
+    mp_zero(c);
+    return MP_OKAY;
+  }
+
+  /* get copies of a and b we can modify */
+  if ((res = mp_init_copy (&u, a)) != MP_OKAY) {
+    return res;
+  }
+
+  if ((res = mp_init_copy (&v, b)) != MP_OKAY) {
+    goto __U;
+  }
+
+  /* must be positive for the remainder of the algorithm */
+  u.sign = v.sign = MP_ZPOS;
+
+  /* B1.  Find the common power of two for u and v */
+  u_lsb = mp_cnt_lsb(&u);
+  v_lsb = mp_cnt_lsb(&v);
+  k     = MIN(u_lsb, v_lsb);
+
+  if (k > 0) {
+     /* divide the power of two out */
+     if ((res = mp_div_2d(&u, k, &u, NULL)) != MP_OKAY) {
+        goto __V;
+     }
+
+     if ((res = mp_div_2d(&v, k, &v, NULL)) != MP_OKAY) {
+        goto __V;
+     }
+  }
+
+  /* divide any remaining factors of two out */
+  if (u_lsb != k) {
+     if ((res = mp_div_2d(&u, u_lsb - k, &u, NULL)) != MP_OKAY) {
+        goto __V;
+     }
+  }
+
+  if (v_lsb != k) {
+     if ((res = mp_div_2d(&v, v_lsb - k, &v, NULL)) != MP_OKAY) {
+        goto __V;
+     }
+  }
+
+  while (mp_iszero(&v) == 0) {
+     /* make sure v is the largest */
+     if (mp_cmp_mag(&u, &v) == MP_GT) {
+        /* swap u and v to make sure v is >= u */
+        mp_exch(&u, &v);
+     }
+     
+     /* subtract smallest from largest */
+     if ((res = s_mp_sub(&v, &u, &v)) != MP_OKAY) {
+        goto __V;
+     }
+     
+     /* Divide out all factors of two */
+     if ((res = mp_div_2d(&v, mp_cnt_lsb(&v), &v, NULL)) != MP_OKAY) {
+        goto __V;
+     } 
+  } 
+
+  /* multiply by 2**k which we divided out at the beginning */
+  if ((res = mp_mul_2d (&u, k, c)) != MP_OKAY) {
+     goto __V;
+  }
+  c->sign = MP_ZPOS;
+  res = MP_OKAY;
+__V:mp_clear (&u);
+__U:mp_clear (&v);
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_get_int.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,39 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* get the lower 32-bits of an mp_int */
+unsigned long mp_get_int(mp_int * a) 
+{
+  int i;
+  unsigned long res;
+
+  if (a->used == 0) {
+     return 0;
+  }
+
+  /* get number of digits of the lsb we have to read */
+  i = MIN(a->used,(int)((sizeof(unsigned long)*CHAR_BIT+DIGIT_BIT-1)/DIGIT_BIT))-1;
+
+  /* get most significant digit of result */
+  res = DIGIT(a,i);
+   
+  while (--i >= 0) {
+    res = (res << DIGIT_BIT) | DIGIT(a,i);
+  }
+
+  /* force result to 32-bits always so it is consistent on non 32-bit platforms */
+  return res & 0xFFFFFFFFUL;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_grow.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,51 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* grow as required */
+int mp_grow (mp_int * a, int size)
+{
+  int     i;
+  mp_digit *tmp;
+
+  /* if the alloc size is smaller alloc more ram */
+  if (a->alloc < size) {
+    /* ensure there are always at least MP_PREC digits extra on top */
+    size += (MP_PREC * 2) - (size % MP_PREC);
+
+    /* reallocate the array a->dp
+     *
+     * We store the return in a temporary variable
+     * in case the operation failed we don't want
+     * to overwrite the dp member of a.
+     */
+    tmp = OPT_CAST(mp_digit) XREALLOC (a->dp, sizeof (mp_digit) * size);
+    if (tmp == NULL) {
+      /* reallocation failed but "a" is still valid [can be freed] */
+      return MP_MEM;
+    }
+
+    /* reallocation succeeded so set a->dp */
+    a->dp = tmp;
+
+    /* zero excess digits */
+    i        = a->alloc;
+    a->alloc = size;
+    for (; i < a->alloc; i++) {
+      a->dp[i] = 0;
+    }
+  }
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_init.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,33 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* init a new bigint */
+int mp_init (mp_int * a)
+{
+  /* allocate memory required and clear it */
+  a->dp = OPT_CAST(mp_digit) XCALLOC (sizeof (mp_digit), MP_PREC);
+  if (a->dp == NULL) {
+    return MP_MEM;
+  }
+
+  /* set the used to zero, allocated digits to the default precision
+   * and sign to positive */
+  a->used  = 0;
+  a->alloc = MP_PREC;
+  a->sign  = MP_ZPOS;
+
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_init_copy.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,26 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* creates "a" then copies b into it */
+int mp_init_copy (mp_int * a, mp_int * b)
+{
+  int     res;
+
+  if ((res = mp_init (a)) != MP_OKAY) {
+    return res;
+  }
+  return mp_copy (b, a);
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_init_multi.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,53 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+#include <stdarg.h>
+
+int mp_init_multi(mp_int *mp, ...) 
+{
+    mp_err res = MP_OKAY;      /* Assume ok until proven otherwise */
+    int n = 0;                 /* Number of ok inits */
+    mp_int* cur_arg = mp;
+    va_list args;
+
+    va_start(args, mp);        /* init args to next argument from caller */
+    while (cur_arg != NULL) {
+        if (mp_init(cur_arg) != MP_OKAY) {
+            /* Oops - error! Back-track and mp_clear what we already
+               succeeded in init-ing, then return error.
+            */
+            va_list clean_args;
+            
+            /* end the current list */
+            va_end(args);
+            
+            /* now start cleaning up */            
+            cur_arg = mp;
+            va_start(clean_args, mp);
+            while (n--) {
+                mp_clear(cur_arg);
+                cur_arg = va_arg(clean_args, mp_int*);
+            }
+            va_end(clean_args);
+            res = MP_MEM;
+            break;
+        }
+        n++;
+        cur_arg = va_arg(args, mp_int*);
+    }
+    va_end(args);
+    return res;                /* Assumed ok, if error flagged above. */
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_init_set.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,26 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* initialize and set a digit */
+int mp_init_set (mp_int * a, mp_digit b)
+{
+  int err;
+  if ((err = mp_init(a)) != MP_OKAY) {
+     return err;
+  }
+  mp_set(a, b);
+  return err;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_init_set_int.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,25 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* initialize and set a digit */
+int mp_init_set_int (mp_int * a, unsigned long b)
+{
+  int err;
+  if ((err = mp_init(a)) != MP_OKAY) {
+     return err;
+  }
+  return mp_set_int(a, b);
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_init_size.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,33 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* init an mp_init for a given size */
+int mp_init_size (mp_int * a, int size)
+{
+  /* pad size so there are always extra digits */
+  size += (MP_PREC * 2) - (size % MP_PREC);	
+  
+  /* alloc mem */
+  a->dp = OPT_CAST(mp_digit) XCALLOC (sizeof (mp_digit), size);
+  if (a->dp == NULL) {
+    return MP_MEM;
+  }
+  a->used  = 0;
+  a->alloc = size;
+  a->sign  = MP_ZPOS;
+
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_invmod.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,174 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* hac 14.61, pp608 */
+int mp_invmod (mp_int * a, mp_int * b, mp_int * c)
+{
+  mp_int  x, y, u, v, A, B, C, D;
+  int     res;
+
+  /* b cannot be negative */
+  if (b->sign == MP_NEG || mp_iszero(b) == 1) {
+    return MP_VAL;
+  }
+
+  /* if the modulus is odd we can use a faster routine instead */
+  if (mp_isodd (b) == 1) {
+    return fast_mp_invmod (a, b, c);
+  }
+  
+  /* init temps */
+  if ((res = mp_init_multi(&x, &y, &u, &v, 
+                           &A, &B, &C, &D, NULL)) != MP_OKAY) {
+     return res;
+  }
+
+  /* x = a, y = b */
+  if ((res = mp_copy (a, &x)) != MP_OKAY) {
+    goto __ERR;
+  }
+  if ((res = mp_copy (b, &y)) != MP_OKAY) {
+    goto __ERR;
+  }
+
+  /* 2. [modified] if x,y are both even then return an error! */
+  if (mp_iseven (&x) == 1 && mp_iseven (&y) == 1) {
+    res = MP_VAL;
+    goto __ERR;
+  }
+
+  /* 3. u=x, v=y, A=1, B=0, C=0,D=1 */
+  if ((res = mp_copy (&x, &u)) != MP_OKAY) {
+    goto __ERR;
+  }
+  if ((res = mp_copy (&y, &v)) != MP_OKAY) {
+    goto __ERR;
+  }
+  mp_set (&A, 1);
+  mp_set (&D, 1);
+
+top:
+  /* 4.  while u is even do */
+  while (mp_iseven (&u) == 1) {
+    /* 4.1 u = u/2 */
+    if ((res = mp_div_2 (&u, &u)) != MP_OKAY) {
+      goto __ERR;
+    }
+    /* 4.2 if A or B is odd then */
+    if (mp_isodd (&A) == 1 || mp_isodd (&B) == 1) {
+      /* A = (A+y)/2, B = (B-x)/2 */
+      if ((res = mp_add (&A, &y, &A)) != MP_OKAY) {
+         goto __ERR;
+      }
+      if ((res = mp_sub (&B, &x, &B)) != MP_OKAY) {
+         goto __ERR;
+      }
+    }
+    /* A = A/2, B = B/2 */
+    if ((res = mp_div_2 (&A, &A)) != MP_OKAY) {
+      goto __ERR;
+    }
+    if ((res = mp_div_2 (&B, &B)) != MP_OKAY) {
+      goto __ERR;
+    }
+  }
+
+  /* 5.  while v is even do */
+  while (mp_iseven (&v) == 1) {
+    /* 5.1 v = v/2 */
+    if ((res = mp_div_2 (&v, &v)) != MP_OKAY) {
+      goto __ERR;
+    }
+    /* 5.2 if C or D is odd then */
+    if (mp_isodd (&C) == 1 || mp_isodd (&D) == 1) {
+      /* C = (C+y)/2, D = (D-x)/2 */
+      if ((res = mp_add (&C, &y, &C)) != MP_OKAY) {
+         goto __ERR;
+      }
+      if ((res = mp_sub (&D, &x, &D)) != MP_OKAY) {
+         goto __ERR;
+      }
+    }
+    /* C = C/2, D = D/2 */
+    if ((res = mp_div_2 (&C, &C)) != MP_OKAY) {
+      goto __ERR;
+    }
+    if ((res = mp_div_2 (&D, &D)) != MP_OKAY) {
+      goto __ERR;
+    }
+  }
+
+  /* 6.  if u >= v then */
+  if (mp_cmp (&u, &v) != MP_LT) {
+    /* u = u - v, A = A - C, B = B - D */
+    if ((res = mp_sub (&u, &v, &u)) != MP_OKAY) {
+      goto __ERR;
+    }
+
+    if ((res = mp_sub (&A, &C, &A)) != MP_OKAY) {
+      goto __ERR;
+    }
+
+    if ((res = mp_sub (&B, &D, &B)) != MP_OKAY) {
+      goto __ERR;
+    }
+  } else {
+    /* v - v - u, C = C - A, D = D - B */
+    if ((res = mp_sub (&v, &u, &v)) != MP_OKAY) {
+      goto __ERR;
+    }
+
+    if ((res = mp_sub (&C, &A, &C)) != MP_OKAY) {
+      goto __ERR;
+    }
+
+    if ((res = mp_sub (&D, &B, &D)) != MP_OKAY) {
+      goto __ERR;
+    }
+  }
+
+  /* if not zero goto step 4 */
+  if (mp_iszero (&u) == 0)
+    goto top;
+
+  /* now a = C, b = D, gcd == g*v */
+
+  /* if v != 1 then there is no inverse */
+  if (mp_cmp_d (&v, 1) != MP_EQ) {
+    res = MP_VAL;
+    goto __ERR;
+  }
+
+  /* if its too low */
+  while (mp_cmp_d(&C, 0) == MP_LT) {
+      if ((res = mp_add(&C, b, &C)) != MP_OKAY) {
+         goto __ERR;
+      }
+  }
+  
+  /* too big */
+  while (mp_cmp_mag(&C, b) != MP_LT) {
+      if ((res = mp_sub(&C, b, &C)) != MP_OKAY) {
+         goto __ERR;
+      }
+  }
+  
+  /* C is now the inverse */
+  mp_exch (&C, c);
+  res = MP_OKAY;
+__ERR:mp_clear_multi (&x, &y, &u, &v, &A, &B, &C, &D, NULL);
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_is_square.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,103 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* Check if remainders are possible squares - fast exclude non-squares */
+static const char rem_128[128] = {
+ 0, 0, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1,
+ 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1,
+ 1, 0, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1,
+ 1, 0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1,
+ 0, 0, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1,
+ 1, 0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1,
+ 1, 0, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1,
+ 1, 0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1
+};
+
+static const char rem_105[105] = {
+ 0, 0, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1,
+ 0, 0, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1,
+ 0, 1, 1, 1, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1,
+ 1, 0, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1,
+ 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1,
+ 1, 1, 1, 1, 0, 1, 0, 1, 1, 0, 0, 1, 1, 1, 1,
+ 1, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 1, 1
+};
+
+/* Store non-zero to ret if arg is square, and zero if not */
+int mp_is_square(mp_int *arg,int *ret) 
+{
+  int           res;
+  mp_digit      c;
+  mp_int        t;
+  unsigned long r;
+
+  /* Default to Non-square :) */
+  *ret = MP_NO; 
+
+  if (arg->sign == MP_NEG) {
+    return MP_VAL;
+  }
+
+  /* digits used?  (TSD) */
+  if (arg->used == 0) {
+     return MP_OKAY;
+  }
+
+  /* First check mod 128 (suppose that DIGIT_BIT is at least 7) */
+  if (rem_128[127 & DIGIT(arg,0)] == 1) {
+     return MP_OKAY;
+  }
+
+  /* Next check mod 105 (3*5*7) */
+  if ((res = mp_mod_d(arg,105,&c)) != MP_OKAY) {
+     return res;
+  }
+  if (rem_105[c] == 1) {
+     return MP_OKAY;
+  }
+
+  /* product of primes less than 2^31 */
+  if ((res = mp_init_set_int(&t,11L*13L*17L*19L*23L*29L*31L)) != MP_OKAY) {
+     return res;
+  }
+  if ((res = mp_mod(arg,&t,&t)) != MP_OKAY) {
+     goto ERR;
+  }
+  r = mp_get_int(&t);
+  /* Check for other prime modules, note it's not an ERROR but we must
+   * free "t" so the easiest way is to goto ERR.  We know that res
+   * is already equal to MP_OKAY from the mp_mod call 
+   */ 
+  if ( (1L<<(r%11)) & 0x5C4L )             goto ERR;
+  if ( (1L<<(r%13)) & 0x9E4L )             goto ERR;
+  if ( (1L<<(r%17)) & 0x5CE8L )            goto ERR;
+  if ( (1L<<(r%19)) & 0x4F50CL )           goto ERR;
+  if ( (1L<<(r%23)) & 0x7ACCA0L )          goto ERR;
+  if ( (1L<<(r%29)) & 0xC2EDD0CL )         goto ERR;
+  if ( (1L<<(r%31)) & 0x6DE2B848L )        goto ERR;
+
+  /* Final check - is sqr(sqrt(arg)) == arg ? */
+  if ((res = mp_sqrt(arg,&t)) != MP_OKAY) {
+     goto ERR;
+  }
+  if ((res = mp_sqr(&t,&t)) != MP_OKAY) {
+     goto ERR;
+  }
+
+  *ret = (mp_cmp_mag(&t,arg) == MP_EQ) ? MP_YES : MP_NO;
+ERR:mp_clear(&t);
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_jacobi.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,99 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* computes the jacobi c = (a | n) (or Legendre if n is prime)
+ * HAC pp. 73 Algorithm 2.149
+ */
+int mp_jacobi (mp_int * a, mp_int * p, int *c)
+{
+  mp_int  a1, p1;
+  int     k, s, r, res;
+  mp_digit residue;
+
+  /* if p <= 0 return MP_VAL */
+  if (mp_cmp_d(p, 0) != MP_GT) {
+     return MP_VAL;
+  }
+
+  /* step 1.  if a == 0, return 0 */
+  if (mp_iszero (a) == 1) {
+    *c = 0;
+    return MP_OKAY;
+  }
+
+  /* step 2.  if a == 1, return 1 */
+  if (mp_cmp_d (a, 1) == MP_EQ) {
+    *c = 1;
+    return MP_OKAY;
+  }
+
+  /* default */
+  s = 0;
+
+  /* step 3.  write a = a1 * 2**k  */
+  if ((res = mp_init_copy (&a1, a)) != MP_OKAY) {
+    return res;
+  }
+
+  if ((res = mp_init (&p1)) != MP_OKAY) {
+    goto __A1;
+  }
+
+  /* divide out larger power of two */
+  k = mp_cnt_lsb(&a1);
+  if ((res = mp_div_2d(&a1, k, &a1, NULL)) != MP_OKAY) {
+     goto __P1;
+  }
+
+  /* step 4.  if e is even set s=1 */
+  if ((k & 1) == 0) {
+    s = 1;
+  } else {
+    /* else set s=1 if p = 1/7 (mod 8) or s=-1 if p = 3/5 (mod 8) */
+    residue = p->dp[0] & 7;
+
+    if (residue == 1 || residue == 7) {
+      s = 1;
+    } else if (residue == 3 || residue == 5) {
+      s = -1;
+    }
+  }
+
+  /* step 5.  if p == 3 (mod 4) *and* a1 == 3 (mod 4) then s = -s */
+  if ( ((p->dp[0] & 3) == 3) && ((a1.dp[0] & 3) == 3)) {
+    s = -s;
+  }
+
+  /* if a1 == 1 we're done */
+  if (mp_cmp_d (&a1, 1) == MP_EQ) {
+    *c = s;
+  } else {
+    /* n1 = n mod a1 */
+    if ((res = mp_mod (p, &a1, &p1)) != MP_OKAY) {
+      goto __P1;
+    }
+    if ((res = mp_jacobi (&p1, &a1, &r)) != MP_OKAY) {
+      goto __P1;
+    }
+    *c = s * r;
+  }
+
+  /* done */
+  res = MP_OKAY;
+__P1:mp_clear (&p1);
+__A1:mp_clear (&a1);
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_karatsuba_mul.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,164 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* c = |a| * |b| using Karatsuba Multiplication using 
+ * three half size multiplications
+ *
+ * Let B represent the radix [e.g. 2**DIGIT_BIT] and 
+ * let n represent half of the number of digits in 
+ * the min(a,b)
+ *
+ * a = a1 * B**n + a0
+ * b = b1 * B**n + b0
+ *
+ * Then, a * b => 
+   a1b1 * B**2n + ((a1 - a0)(b1 - b0) + a0b0 + a1b1) * B + a0b0
+ *
+ * Note that a1b1 and a0b0 are used twice and only need to be 
+ * computed once.  So in total three half size (half # of 
+ * digit) multiplications are performed, a0b0, a1b1 and 
+ * (a1-b1)(a0-b0)
+ *
+ * Note that a multiplication of half the digits requires
+ * 1/4th the number of single precision multiplications so in 
+ * total after one call 25% of the single precision multiplications 
+ * are saved.  Note also that the call to mp_mul can end up back 
+ * in this function if the a0, a1, b0, or b1 are above the threshold.  
+ * This is known as divide-and-conquer and leads to the famous 
+ * O(N**lg(3)) or O(N**1.584) work which is asymptopically lower than 
+ * the standard O(N**2) that the baseline/comba methods use.  
+ * Generally though the overhead of this method doesn't pay off 
+ * until a certain size (N ~ 80) is reached.
+ */
+int mp_karatsuba_mul (mp_int * a, mp_int * b, mp_int * c)
+{
+  mp_int  x0, x1, y0, y1, t1, x0y0, x1y1;
+  int     B, err;
+
+  /* default the return code to an error */
+  err = MP_MEM;
+
+  /* min # of digits */
+  B = MIN (a->used, b->used);
+
+  /* now divide in two */
+  B = B >> 1;
+
+  /* init copy all the temps */
+  if (mp_init_size (&x0, B) != MP_OKAY)
+    goto ERR;
+  if (mp_init_size (&x1, a->used - B) != MP_OKAY)
+    goto X0;
+  if (mp_init_size (&y0, B) != MP_OKAY)
+    goto X1;
+  if (mp_init_size (&y1, b->used - B) != MP_OKAY)
+    goto Y0;
+
+  /* init temps */
+  if (mp_init_size (&t1, B * 2) != MP_OKAY)
+    goto Y1;
+  if (mp_init_size (&x0y0, B * 2) != MP_OKAY)
+    goto T1;
+  if (mp_init_size (&x1y1, B * 2) != MP_OKAY)
+    goto X0Y0;
+
+  /* now shift the digits */
+  x0.sign = x1.sign = a->sign;
+  y0.sign = y1.sign = b->sign;
+
+  x0.used = y0.used = B;
+  x1.used = a->used - B;
+  y1.used = b->used - B;
+
+  {
+    register int x;
+    register mp_digit *tmpa, *tmpb, *tmpx, *tmpy;
+
+    /* we copy the digits directly instead of using higher level functions
+     * since we also need to shift the digits
+     */
+    tmpa = a->dp;
+    tmpb = b->dp;
+
+    tmpx = x0.dp;
+    tmpy = y0.dp;
+    for (x = 0; x < B; x++) {
+      *tmpx++ = *tmpa++;
+      *tmpy++ = *tmpb++;
+    }
+
+    tmpx = x1.dp;
+    for (x = B; x < a->used; x++) {
+      *tmpx++ = *tmpa++;
+    }
+
+    tmpy = y1.dp;
+    for (x = B; x < b->used; x++) {
+      *tmpy++ = *tmpb++;
+    }
+  }
+
+  /* only need to clamp the lower words since by definition the 
+   * upper words x1/y1 must have a known number of digits
+   */
+  mp_clamp (&x0);
+  mp_clamp (&y0);
+
+  /* now calc the products x0y0 and x1y1 */
+  /* after this x0 is no longer required, free temp [x0==t2]! */
+  if (mp_mul (&x0, &y0, &x0y0) != MP_OKAY)  
+    goto X1Y1;          /* x0y0 = x0*y0 */
+  if (mp_mul (&x1, &y1, &x1y1) != MP_OKAY)
+    goto X1Y1;          /* x1y1 = x1*y1 */
+
+  /* now calc x1-x0 and y1-y0 */
+  if (mp_sub (&x1, &x0, &t1) != MP_OKAY)
+    goto X1Y1;          /* t1 = x1 - x0 */
+  if (mp_sub (&y1, &y0, &x0) != MP_OKAY)
+    goto X1Y1;          /* t2 = y1 - y0 */
+  if (mp_mul (&t1, &x0, &t1) != MP_OKAY)
+    goto X1Y1;          /* t1 = (x1 - x0) * (y1 - y0) */
+
+  /* add x0y0 */
+  if (mp_add (&x0y0, &x1y1, &x0) != MP_OKAY)
+    goto X1Y1;          /* t2 = x0y0 + x1y1 */
+  if (mp_sub (&x0, &t1, &t1) != MP_OKAY)
+    goto X1Y1;          /* t1 = x0y0 + x1y1 - (x1-x0)*(y1-y0) */
+
+  /* shift by B */
+  if (mp_lshd (&t1, B) != MP_OKAY)
+    goto X1Y1;          /* t1 = (x0y0 + x1y1 - (x1-x0)*(y1-y0))<<B */
+  if (mp_lshd (&x1y1, B * 2) != MP_OKAY)
+    goto X1Y1;          /* x1y1 = x1y1 << 2*B */
+
+  if (mp_add (&x0y0, &t1, &t1) != MP_OKAY)
+    goto X1Y1;          /* t1 = x0y0 + t1 */
+  if (mp_add (&t1, &x1y1, c) != MP_OKAY)
+    goto X1Y1;          /* t1 = x0y0 + t1 + x1y1 */
+
+  /* Algorithm succeeded set the return code to MP_OKAY */
+  err = MP_OKAY;
+
+X1Y1:mp_clear (&x1y1);
+X0Y0:mp_clear (&x0y0);
+T1:mp_clear (&t1);
+Y1:mp_clear (&y1);
+Y0:mp_clear (&y0);
+X1:mp_clear (&x1);
+X0:mp_clear (&x0);
+ERR:
+  return err;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_karatsuba_sqr.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,115 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* Karatsuba squaring, computes b = a*a using three 
+ * half size squarings
+ *
+ * See comments of mp_karatsuba_mul for details.  It 
+ * is essentially the same algorithm but merely 
+ * tuned to perform recursive squarings.
+ */
+int mp_karatsuba_sqr (mp_int * a, mp_int * b)
+{
+  mp_int  x0, x1, t1, t2, x0x0, x1x1;
+  int     B, err;
+
+  err = MP_MEM;
+
+  /* min # of digits */
+  B = a->used;
+
+  /* now divide in two */
+  B = B >> 1;
+
+  /* init copy all the temps */
+  if (mp_init_size (&x0, B) != MP_OKAY)
+    goto ERR;
+  if (mp_init_size (&x1, a->used - B) != MP_OKAY)
+    goto X0;
+
+  /* init temps */
+  if (mp_init_size (&t1, a->used * 2) != MP_OKAY)
+    goto X1;
+  if (mp_init_size (&t2, a->used * 2) != MP_OKAY)
+    goto T1;
+  if (mp_init_size (&x0x0, B * 2) != MP_OKAY)
+    goto T2;
+  if (mp_init_size (&x1x1, (a->used - B) * 2) != MP_OKAY)
+    goto X0X0;
+
+  {
+    register int x;
+    register mp_digit *dst, *src;
+
+    src = a->dp;
+
+    /* now shift the digits */
+    dst = x0.dp;
+    for (x = 0; x < B; x++) {
+      *dst++ = *src++;
+    }
+
+    dst = x1.dp;
+    for (x = B; x < a->used; x++) {
+      *dst++ = *src++;
+    }
+  }
+
+  x0.used = B;
+  x1.used = a->used - B;
+
+  mp_clamp (&x0);
+
+  /* now calc the products x0*x0 and x1*x1 */
+  if (mp_sqr (&x0, &x0x0) != MP_OKAY)
+    goto X1X1;           /* x0x0 = x0*x0 */
+  if (mp_sqr (&x1, &x1x1) != MP_OKAY)
+    goto X1X1;           /* x1x1 = x1*x1 */
+
+  /* now calc (x1-x0)**2 */
+  if (mp_sub (&x1, &x0, &t1) != MP_OKAY)
+    goto X1X1;           /* t1 = x1 - x0 */
+  if (mp_sqr (&t1, &t1) != MP_OKAY)
+    goto X1X1;           /* t1 = (x1 - x0) * (x1 - x0) */
+
+  /* add x0y0 */
+  if (s_mp_add (&x0x0, &x1x1, &t2) != MP_OKAY)
+    goto X1X1;           /* t2 = x0x0 + x1x1 */
+  if (mp_sub (&t2, &t1, &t1) != MP_OKAY)
+    goto X1X1;           /* t1 = x0x0 + x1x1 - (x1-x0)*(x1-x0) */
+
+  /* shift by B */
+  if (mp_lshd (&t1, B) != MP_OKAY)
+    goto X1X1;           /* t1 = (x0x0 + x1x1 - (x1-x0)*(x1-x0))<<B */
+  if (mp_lshd (&x1x1, B * 2) != MP_OKAY)
+    goto X1X1;           /* x1x1 = x1x1 << 2*B */
+
+  if (mp_add (&x0x0, &t1, &t1) != MP_OKAY)
+    goto X1X1;           /* t1 = x0x0 + t1 */
+  if (mp_add (&t1, &x1x1, b) != MP_OKAY)
+    goto X1X1;           /* t1 = x0x0 + t1 + x1x1 */
+
+  err = MP_OKAY;
+
+X1X1:mp_clear (&x1x1);
+X0X0:mp_clear (&x0x0);
+T2:mp_clear (&t2);
+T1:mp_clear (&t1);
+X1:mp_clear (&x1);
+X0:mp_clear (&x0);
+ERR:
+  return err;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_lcm.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,54 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* computes least common multiple as |a*b|/(a, b) */
+int mp_lcm (mp_int * a, mp_int * b, mp_int * c)
+{
+  int     res;
+  mp_int  t1, t2;
+
+
+  if ((res = mp_init_multi (&t1, &t2, NULL)) != MP_OKAY) {
+    return res;
+  }
+
+  /* t1 = get the GCD of the two inputs */
+  if ((res = mp_gcd (a, b, &t1)) != MP_OKAY) {
+    goto __T;
+  }
+
+  /* divide the smallest by the GCD */
+  if (mp_cmp_mag(a, b) == MP_LT) {
+     /* store quotient in t2 such that t2 * b is the LCM */
+     if ((res = mp_div(a, &t1, &t2, NULL)) != MP_OKAY) {
+        goto __T;
+     }
+     res = mp_mul(b, &t2, c);
+  } else {
+     /* store quotient in t2 such that t2 * a is the LCM */
+     if ((res = mp_div(b, &t1, &t2, NULL)) != MP_OKAY) {
+        goto __T;
+     }
+     res = mp_mul(a, &t2, c);
+  }
+
+  /* fix the sign to positive */
+  c->sign = MP_ZPOS;
+
+__T:
+  mp_clear_multi (&t1, &t2, NULL);
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_lshd.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,61 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* shift left a certain amount of digits */
+int mp_lshd (mp_int * a, int b)
+{
+  int     x, res;
+
+  /* if its less than zero return */
+  if (b <= 0) {
+    return MP_OKAY;
+  }
+
+  /* grow to fit the new digits */
+  if (a->alloc < a->used + b) {
+     if ((res = mp_grow (a, a->used + b)) != MP_OKAY) {
+       return res;
+     }
+  }
+
+  {
+    register mp_digit *top, *bottom;
+
+    /* increment the used by the shift amount then copy upwards */
+    a->used += b;
+
+    /* top */
+    top = a->dp + a->used - 1;
+
+    /* base */
+    bottom = a->dp + a->used - 1 - b;
+
+    /* much like mp_rshd this is implemented using a sliding window
+     * except the window goes the otherway around.  Copying from
+     * the bottom to the top.  see bn_mp_rshd.c for more info.
+     */
+    for (x = a->used - 1; x >= b; x--) {
+      *top-- = *bottom--;
+    }
+
+    /* zero the lower digits */
+    top = a->dp;
+    for (x = 0; x < b; x++) {
+      *top++ = 0;
+    }
+  }
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_mod.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,42 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* c = a mod b, 0 <= c < b */
+int
+mp_mod (mp_int * a, mp_int * b, mp_int * c)
+{
+  mp_int  t;
+  int     res;
+
+  if ((res = mp_init (&t)) != MP_OKAY) {
+    return res;
+  }
+
+  if ((res = mp_div (a, b, NULL, &t)) != MP_OKAY) {
+    mp_clear (&t);
+    return res;
+  }
+
+  if (t.sign != b->sign) {
+    res = mp_add (b, &t, c);
+  } else {
+    res = MP_OKAY;
+    mp_exch (&t, c);
+  }
+
+  mp_clear (&t);
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_mod_2d.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,49 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* calc a value mod 2**b */
+int
+mp_mod_2d (mp_int * a, int b, mp_int * c)
+{
+  int     x, res;
+
+  /* if b is <= 0 then zero the int */
+  if (b <= 0) {
+    mp_zero (c);
+    return MP_OKAY;
+  }
+
+  /* if the modulus is larger than the value than return */
+  if (b > (int) (a->used * DIGIT_BIT)) {
+    res = mp_copy (a, c);
+    return res;
+  }
+
+  /* copy */
+  if ((res = mp_copy (a, c)) != MP_OKAY) {
+    return res;
+  }
+
+  /* zero digits above the last digit of the modulus */
+  for (x = (b / DIGIT_BIT) + ((b % DIGIT_BIT) == 0 ? 0 : 1); x < c->used; x++) {
+    c->dp[x] = 0;
+  }
+  /* clear the digit that is not completely outside/inside the modulus */
+  c->dp[b / DIGIT_BIT] &=
+    (mp_digit) ((((mp_digit) 1) << (((mp_digit) b) % DIGIT_BIT)) - ((mp_digit) 1));
+  mp_clamp (c);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_mod_d.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,21 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+int
+mp_mod_d (mp_int * a, mp_digit b, mp_digit * c)
+{
+  return mp_div_d(a, b, NULL, c);
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_montgomery_calc_normalization.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,53 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* calculates a = B^n mod b for Montgomery reduction
+ * Where B is the base [e.g. 2^DIGIT_BIT].
+ * B^n mod b is computed by first computing
+ * A = B^(n-1) which doesn't require a reduction but a simple OR.
+ * then C = A * B = B^n is computed by performing upto DIGIT_BIT
+ * shifts with subtractions when the result is greater than b.
+ *
+ * The method is slightly modified to shift B unconditionally upto just under
+ * the leading bit of b.  This saves alot of multiple precision shifting.
+ */
+int
+mp_montgomery_calc_normalization (mp_int * a, mp_int * b)
+{
+  int     x, bits, res;
+
+  /* how many bits of last digit does b use */
+  bits = mp_count_bits (b) % DIGIT_BIT;
+
+  /* compute A = B^(n-1) * 2^(bits-1) */
+  if ((res = mp_2expt (a, (b->used - 1) * DIGIT_BIT + bits - 1)) != MP_OKAY) {
+    return res;
+  }
+
+  /* now compute C = A * B mod b */
+  for (x = bits - 1; x < (int)DIGIT_BIT; x++) {
+    if ((res = mp_mul_2 (a, a)) != MP_OKAY) {
+      return res;
+    }
+    if (mp_cmp_mag (a, b) != MP_LT) {
+      if ((res = s_mp_sub (a, b, a)) != MP_OKAY) {
+        return res;
+      }
+    }
+  }
+
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_montgomery_reduce.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,112 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* computes xR**-1 == x (mod N) via Montgomery Reduction */
+int
+mp_montgomery_reduce (mp_int * x, mp_int * n, mp_digit rho)
+{
+  int     ix, res, digs;
+  mp_digit mu;
+
+  /* can the fast reduction [comba] method be used?
+   *
+   * Note that unlike in mp_mul you're safely allowed *less*
+   * than the available columns [255 per default] since carries
+   * are fixed up in the inner loop.
+   */
+  digs = n->used * 2 + 1;
+  if ((digs < MP_WARRAY) &&
+      n->used <
+      (1 << ((CHAR_BIT * sizeof (mp_word)) - (2 * DIGIT_BIT)))) {
+    return fast_mp_montgomery_reduce (x, n, rho);
+  }
+
+  /* grow the input as required */
+  if (x->alloc < digs) {
+    if ((res = mp_grow (x, digs)) != MP_OKAY) {
+      return res;
+    }
+  }
+  x->used = digs;
+
+  for (ix = 0; ix < n->used; ix++) {
+    /* mu = ai * rho mod b
+     *
+     * The value of rho must be precalculated via
+     * bn_mp_montgomery_setup() such that
+     * it equals -1/n0 mod b this allows the
+     * following inner loop to reduce the
+     * input one digit at a time
+     */
+    mu = (mp_digit) (((mp_word)x->dp[ix]) * ((mp_word)rho) & MP_MASK);
+
+    /* a = a + mu * m * b**i */
+    {
+      register int iy;
+      register mp_digit *tmpn, *tmpx, u;
+      register mp_word r;
+
+      /* alias for digits of the modulus */
+      tmpn = n->dp;
+
+      /* alias for the digits of x [the input] */
+      tmpx = x->dp + ix;
+
+      /* set the carry to zero */
+      u = 0;
+
+      /* Multiply and add in place */
+      for (iy = 0; iy < n->used; iy++) {
+        /* compute product and sum */
+        r       = ((mp_word)mu) * ((mp_word)*tmpn++) +
+                  ((mp_word) u) + ((mp_word) * tmpx);
+
+        /* get carry */
+        u       = (mp_digit)(r >> ((mp_word) DIGIT_BIT));
+
+        /* fix digit */
+        *tmpx++ = (mp_digit)(r & ((mp_word) MP_MASK));
+      }
+      /* At this point the ix'th digit of x should be zero */
+
+
+      /* propagate carries upwards as required*/
+      while (u) {
+        *tmpx   += u;
+        u        = *tmpx >> DIGIT_BIT;
+        *tmpx++ &= MP_MASK;
+      }
+    }
+  }
+
+  /* at this point the n.used'th least
+   * significant digits of x are all zero
+   * which means we can shift x to the
+   * right by n.used digits and the
+   * residue is unchanged.
+   */
+
+  /* x = x/b**n.used */
+  mp_clamp(x);
+  mp_rshd (x, n->used);
+
+  /* if x >= n then x = x - n */
+  if (mp_cmp_mag (x, n) != MP_LT) {
+    return s_mp_sub (x, n, x);
+  }
+
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_montgomery_setup.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,53 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* setups the montgomery reduction stuff */
+int
+mp_montgomery_setup (mp_int * n, mp_digit * rho)
+{
+  mp_digit x, b;
+
+/* fast inversion mod 2**k
+ *
+ * Based on the fact that
+ *
+ * XA = 1 (mod 2**n)  =>  (X(2-XA)) A = 1 (mod 2**2n)
+ *                    =>  2*X*A - X*X*A*A = 1
+ *                    =>  2*(1) - (1)     = 1
+ */
+  b = n->dp[0];
+
+  if ((b & 1) == 0) {
+    return MP_VAL;
+  }
+
+  x = (((b + 2) & 4) << 1) + b; /* here x*a==1 mod 2**4 */
+  x *= 2 - b * x;               /* here x*a==1 mod 2**8 */
+#if !defined(MP_8BIT)
+  x *= 2 - b * x;               /* here x*a==1 mod 2**16 */
+#endif
+#if defined(MP_64BIT) || !(defined(MP_8BIT) || defined(MP_16BIT))
+  x *= 2 - b * x;               /* here x*a==1 mod 2**32 */
+#endif
+#ifdef MP_64BIT
+  x *= 2 - b * x;               /* here x*a==1 mod 2**64 */
+#endif
+
+  /* rho = -1/m mod b */
+  *rho = (((mp_digit) 1 << ((mp_digit) DIGIT_BIT)) - x) & MP_MASK;
+
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_mul.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,48 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* high level multiplication (handles sign) */
+int mp_mul (mp_int * a, mp_int * b, mp_int * c)
+{
+  int     res, neg;
+  neg = (a->sign == b->sign) ? MP_ZPOS : MP_NEG;
+
+  /* use Toom-Cook? */
+  if (MIN (a->used, b->used) >= TOOM_MUL_CUTOFF) {
+    res = mp_toom_mul(a, b, c);
+  /* use Karatsuba? */
+  } else if (MIN (a->used, b->used) >= KARATSUBA_MUL_CUTOFF) {
+    res = mp_karatsuba_mul (a, b, c);
+  } else {
+    /* can we use the fast multiplier?
+     *
+     * The fast multiplier can be used if the output will 
+     * have less than MP_WARRAY digits and the number of 
+     * digits won't affect carry propagation
+     */
+    int     digs = a->used + b->used + 1;
+
+    if ((digs < MP_WARRAY) &&
+        MIN(a->used, b->used) <= 
+        (1 << ((CHAR_BIT * sizeof (mp_word)) - (2 * DIGIT_BIT)))) {
+      res = fast_s_mp_mul_digs (a, b, c, digs);
+    } else {
+      res = s_mp_mul (a, b, c);
+    }
+  }
+  c->sign = neg;
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_mul_2.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,76 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* b = a*2 */
+int mp_mul_2(mp_int * a, mp_int * b)
+{
+  int     x, res, oldused;
+
+  /* grow to accomodate result */
+  if (b->alloc < a->used + 1) {
+    if ((res = mp_grow (b, a->used + 1)) != MP_OKAY) {
+      return res;
+    }
+  }
+
+  oldused = b->used;
+  b->used = a->used;
+
+  {
+    register mp_digit r, rr, *tmpa, *tmpb;
+
+    /* alias for source */
+    tmpa = a->dp;
+    
+    /* alias for dest */
+    tmpb = b->dp;
+
+    /* carry */
+    r = 0;
+    for (x = 0; x < a->used; x++) {
+    
+      /* get what will be the *next* carry bit from the 
+       * MSB of the current digit 
+       */
+      rr = *tmpa >> ((mp_digit)(DIGIT_BIT - 1));
+      
+      /* now shift up this digit, add in the carry [from the previous] */
+      *tmpb++ = ((*tmpa++ << ((mp_digit)1)) | r) & MP_MASK;
+      
+      /* copy the carry that would be from the source 
+       * digit into the next iteration 
+       */
+      r = rr;
+    }
+
+    /* new leading digit? */
+    if (r != 0) {
+      /* add a MSB which is always 1 at this point */
+      *tmpb = 1;
+      ++(b->used);
+    }
+
+    /* now zero any excess digits on the destination 
+     * that we didn't write to 
+     */
+    tmpb = b->dp + b->used;
+    for (x = b->used; x < oldused; x++) {
+      *tmpb++ = 0;
+    }
+  }
+  b->sign = a->sign;
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_mul_2d.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,79 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* shift left by a certain bit count */
+int mp_mul_2d (mp_int * a, int b, mp_int * c)
+{
+  mp_digit d;
+  int      res;
+
+  /* copy */
+  if (a != c) {
+     if ((res = mp_copy (a, c)) != MP_OKAY) {
+       return res;
+     }
+  }
+
+  if (c->alloc < (int)(c->used + b/DIGIT_BIT + 1)) {
+     if ((res = mp_grow (c, c->used + b / DIGIT_BIT + 1)) != MP_OKAY) {
+       return res;
+     }
+  }
+
+  /* shift by as many digits in the bit count */
+  if (b >= (int)DIGIT_BIT) {
+    if ((res = mp_lshd (c, b / DIGIT_BIT)) != MP_OKAY) {
+      return res;
+    }
+  }
+
+  /* shift any bit count < DIGIT_BIT */
+  d = (mp_digit) (b % DIGIT_BIT);
+  if (d != 0) {
+    register mp_digit *tmpc, shift, mask, r, rr;
+    register int x;
+
+    /* bitmask for carries */
+    mask = (((mp_digit)1) << d) - 1;
+
+    /* shift for msbs */
+    shift = DIGIT_BIT - d;
+
+    /* alias */
+    tmpc = c->dp;
+
+    /* carry */
+    r    = 0;
+    for (x = 0; x < c->used; x++) {
+      /* get the higher bits of the current word */
+      rr = (*tmpc >> shift) & mask;
+
+      /* shift the current word and OR in the carry */
+      *tmpc = ((*tmpc << d) | r) & MP_MASK;
+      ++tmpc;
+
+      /* set the carry to the carry bits of the current word */
+      r = rr;
+    }
+    
+    /* set final carry */
+    if (r != 0) {
+       c->dp[(c->used)++] = r;
+    }
+  }
+  mp_clamp (c);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_mul_d.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,72 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* multiply by a digit */
+int
+mp_mul_d (mp_int * a, mp_digit b, mp_int * c)
+{
+  mp_digit u, *tmpa, *tmpc;
+  mp_word  r;
+  int      ix, res, olduse;
+
+  /* make sure c is big enough to hold a*b */
+  if (c->alloc < a->used + 1) {
+    if ((res = mp_grow (c, a->used + 1)) != MP_OKAY) {
+      return res;
+    }
+  }
+
+  /* get the original destinations used count */
+  olduse = c->used;
+
+  /* set the sign */
+  c->sign = a->sign;
+
+  /* alias for a->dp [source] */
+  tmpa = a->dp;
+
+  /* alias for c->dp [dest] */
+  tmpc = c->dp;
+
+  /* zero carry */
+  u = 0;
+
+  /* compute columns */
+  for (ix = 0; ix < a->used; ix++) {
+    /* compute product and carry sum for this term */
+    r       = ((mp_word) u) + ((mp_word)*tmpa++) * ((mp_word)b);
+
+    /* mask off higher bits to get a single digit */
+    *tmpc++ = (mp_digit) (r & ((mp_word) MP_MASK));
+
+    /* send carry into next iteration */
+    u       = (mp_digit) (r >> ((mp_word) DIGIT_BIT));
+  }
+
+  /* store final carry [if any] */
+  *tmpc++ = u;
+
+  /* now zero digits above the top */
+  while (ix++ < olduse) {
+     *tmpc++ = 0;
+  }
+
+  /* set used count */
+  c->used = a->used + 1;
+  mp_clamp(c);
+
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_mulmod.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,35 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* d = a * b (mod c) */
+int
+mp_mulmod (mp_int * a, mp_int * b, mp_int * c, mp_int * d)
+{
+  int     res;
+  mp_int  t;
+
+  if ((res = mp_init (&t)) != MP_OKAY) {
+    return res;
+  }
+
+  if ((res = mp_mul (a, b, &t)) != MP_OKAY) {
+    mp_clear (&t);
+    return res;
+  }
+  res = mp_mod (&t, c, d);
+  mp_clear (&t);
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_n_root.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,126 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* find the n'th root of an integer 
+ *
+ * Result found such that (c)**b <= a and (c+1)**b > a 
+ *
+ * This algorithm uses Newton's approximation 
+ * x[i+1] = x[i] - f(x[i])/f'(x[i]) 
+ * which will find the root in log(N) time where 
+ * each step involves a fair bit.  This is not meant to 
+ * find huge roots [square and cube, etc].
+ */
+int mp_n_root (mp_int * a, mp_digit b, mp_int * c)
+{
+  mp_int  t1, t2, t3;
+  int     res, neg;
+
+  /* input must be positive if b is even */
+  if ((b & 1) == 0 && a->sign == MP_NEG) {
+    return MP_VAL;
+  }
+
+  if ((res = mp_init (&t1)) != MP_OKAY) {
+    return res;
+  }
+
+  if ((res = mp_init (&t2)) != MP_OKAY) {
+    goto __T1;
+  }
+
+  if ((res = mp_init (&t3)) != MP_OKAY) {
+    goto __T2;
+  }
+
+  /* if a is negative fudge the sign but keep track */
+  neg     = a->sign;
+  a->sign = MP_ZPOS;
+
+  /* t2 = 2 */
+  mp_set (&t2, 2);
+
+  do {
+    /* t1 = t2 */
+    if ((res = mp_copy (&t2, &t1)) != MP_OKAY) {
+      goto __T3;
+    }
+
+    /* t2 = t1 - ((t1**b - a) / (b * t1**(b-1))) */
+    
+    /* t3 = t1**(b-1) */
+    if ((res = mp_expt_d (&t1, b - 1, &t3)) != MP_OKAY) {   
+      goto __T3;
+    }
+
+    /* numerator */
+    /* t2 = t1**b */
+    if ((res = mp_mul (&t3, &t1, &t2)) != MP_OKAY) {    
+      goto __T3;
+    }
+
+    /* t2 = t1**b - a */
+    if ((res = mp_sub (&t2, a, &t2)) != MP_OKAY) {  
+      goto __T3;
+    }
+
+    /* denominator */
+    /* t3 = t1**(b-1) * b  */
+    if ((res = mp_mul_d (&t3, b, &t3)) != MP_OKAY) {    
+      goto __T3;
+    }
+
+    /* t3 = (t1**b - a)/(b * t1**(b-1)) */
+    if ((res = mp_div (&t2, &t3, &t3, NULL)) != MP_OKAY) {  
+      goto __T3;
+    }
+
+    if ((res = mp_sub (&t1, &t3, &t2)) != MP_OKAY) {
+      goto __T3;
+    }
+  }  while (mp_cmp (&t1, &t2) != MP_EQ);
+
+  /* result can be off by a few so check */
+  for (;;) {
+    if ((res = mp_expt_d (&t1, b, &t2)) != MP_OKAY) {
+      goto __T3;
+    }
+
+    if (mp_cmp (&t2, a) == MP_GT) {
+      if ((res = mp_sub_d (&t1, 1, &t1)) != MP_OKAY) {
+         goto __T3;
+      }
+    } else {
+      break;
+    }
+  }
+
+  /* reset the sign of a first */
+  a->sign = neg;
+
+  /* set the result */
+  mp_exch (&t1, c);
+
+  /* set the sign of the result */
+  c->sign = neg;
+
+  res = MP_OKAY;
+
+__T3:mp_clear (&t3);
+__T2:mp_clear (&t2);
+__T1:mp_clear (&t1);
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_neg.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,28 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* b = -a */
+int mp_neg (mp_int * a, mp_int * b)
+{
+  int     res;
+  if ((res = mp_copy (a, b)) != MP_OKAY) {
+    return res;
+  }
+  if (mp_iszero(b) != MP_YES) {
+     b->sign = (a->sign == MP_ZPOS) ? MP_NEG : MP_ZPOS;
+  }
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_or.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,44 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* OR two ints together */
+int mp_or (mp_int * a, mp_int * b, mp_int * c)
+{
+  int     res, ix, px;
+  mp_int  t, *x;
+
+  if (a->used > b->used) {
+    if ((res = mp_init_copy (&t, a)) != MP_OKAY) {
+      return res;
+    }
+    px = b->used;
+    x = b;
+  } else {
+    if ((res = mp_init_copy (&t, b)) != MP_OKAY) {
+      return res;
+    }
+    px = a->used;
+    x = a;
+  }
+
+  for (ix = 0; ix < px; ix++) {
+    t.dp[ix] |= x->dp[ix];
+  }
+  mp_clamp (&t);
+  mp_exch (c, &t);
+  mp_clear (&t);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_prime_fermat.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,56 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* performs one Fermat test.
+ * 
+ * If "a" were prime then b**a == b (mod a) since the order of
+ * the multiplicative sub-group would be phi(a) = a-1.  That means
+ * it would be the same as b**(a mod (a-1)) == b**1 == b (mod a).
+ *
+ * Sets result to 1 if the congruence holds, or zero otherwise.
+ */
+int mp_prime_fermat (mp_int * a, mp_int * b, int *result)
+{
+  mp_int  t;
+  int     err;
+
+  /* default to composite  */
+  *result = MP_NO;
+
+  /* ensure b > 1 */
+  if (mp_cmp_d(b, 1) != MP_GT) {
+     return MP_VAL;
+  }
+
+  /* init t */
+  if ((err = mp_init (&t)) != MP_OKAY) {
+    return err;
+  }
+
+  /* compute t = b**a mod a */
+  if ((err = mp_exptmod (b, a, a, &t)) != MP_OKAY) {
+    goto __T;
+  }
+
+  /* is it equal to b? */
+  if (mp_cmp (&t, b) == MP_EQ) {
+    *result = MP_YES;
+  }
+
+  err = MP_OKAY;
+__T:mp_clear (&t);
+  return err;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_prime_is_divisible.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,44 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* determines if an integers is divisible by one 
+ * of the first PRIME_SIZE primes or not
+ *
+ * sets result to 0 if not, 1 if yes
+ */
+int mp_prime_is_divisible (mp_int * a, int *result)
+{
+  int     err, ix;
+  mp_digit res;
+
+  /* default to not */
+  *result = MP_NO;
+
+  for (ix = 0; ix < PRIME_SIZE; ix++) {
+    /* what is a mod __prime_tab[ix] */
+    if ((err = mp_mod_d (a, __prime_tab[ix], &res)) != MP_OKAY) {
+      return err;
+    }
+
+    /* is the residue zero? */
+    if (res == 0) {
+      *result = MP_YES;
+      return MP_OKAY;
+    }
+  }
+
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_prime_is_prime.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,77 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* performs a variable number of rounds of Miller-Rabin
+ *
+ * Probability of error after t rounds is no more than
+ * (1/4)^t when 1 <= t <= PRIME_SIZE
+ *
+ * Sets result to 1 if probably prime, 0 otherwise
+ */
+int mp_prime_is_prime (mp_int * a, int t, int *result)
+{
+  mp_int  b;
+  int     ix, err, res;
+
+  /* default to no */
+  *result = MP_NO;
+
+  /* valid value of t? */
+  if (t <= 0 || t > PRIME_SIZE) {
+    return MP_VAL;
+  }
+
+  /* is the input equal to one of the primes in the table? */
+  for (ix = 0; ix < PRIME_SIZE; ix++) {
+      if (mp_cmp_d(a, __prime_tab[ix]) == MP_EQ) {
+         *result = 1;
+         return MP_OKAY;
+      }
+  }
+
+  /* first perform trial division */
+  if ((err = mp_prime_is_divisible (a, &res)) != MP_OKAY) {
+    return err;
+  }
+
+  /* return if it was trivially divisible */
+  if (res == MP_YES) {
+    return MP_OKAY;
+  }
+
+  /* now perform the miller-rabin rounds */
+  if ((err = mp_init (&b)) != MP_OKAY) {
+    return err;
+  }
+
+  for (ix = 0; ix < t; ix++) {
+    /* set the prime */
+    mp_set (&b, __prime_tab[ix]);
+
+    if ((err = mp_prime_miller_rabin (a, &b, &res)) != MP_OKAY) {
+      goto __B;
+    }
+
+    if (res == MP_NO) {
+      goto __B;
+    }
+  }
+
+  /* passed the test */
+  *result = MP_YES;
+__B:mp_clear (&b);
+  return err;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_prime_miller_rabin.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,97 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* Miller-Rabin test of "a" to the base of "b" as described in 
+ * HAC pp. 139 Algorithm 4.24
+ *
+ * Sets result to 0 if definitely composite or 1 if probably prime.
+ * Randomly the chance of error is no more than 1/4 and often 
+ * very much lower.
+ */
+int mp_prime_miller_rabin (mp_int * a, mp_int * b, int *result)
+{
+  mp_int  n1, y, r;
+  int     s, j, err;
+
+  /* default */
+  *result = MP_NO;
+
+  /* ensure b > 1 */
+  if (mp_cmp_d(b, 1) != MP_GT) {
+     return MP_VAL;
+  }     
+
+  /* get n1 = a - 1 */
+  if ((err = mp_init_copy (&n1, a)) != MP_OKAY) {
+    return err;
+  }
+  if ((err = mp_sub_d (&n1, 1, &n1)) != MP_OKAY) {
+    goto __N1;
+  }
+
+  /* set 2**s * r = n1 */
+  if ((err = mp_init_copy (&r, &n1)) != MP_OKAY) {
+    goto __N1;
+  }
+
+  /* count the number of least significant bits
+   * which are zero
+   */
+  s = mp_cnt_lsb(&r);
+
+  /* now divide n - 1 by 2**s */
+  if ((err = mp_div_2d (&r, s, &r, NULL)) != MP_OKAY) {
+    goto __R;
+  }
+
+  /* compute y = b**r mod a */
+  if ((err = mp_init (&y)) != MP_OKAY) {
+    goto __R;
+  }
+  if ((err = mp_exptmod (b, &r, a, &y)) != MP_OKAY) {
+    goto __Y;
+  }
+
+  /* if y != 1 and y != n1 do */
+  if (mp_cmp_d (&y, 1) != MP_EQ && mp_cmp (&y, &n1) != MP_EQ) {
+    j = 1;
+    /* while j <= s-1 and y != n1 */
+    while ((j <= (s - 1)) && mp_cmp (&y, &n1) != MP_EQ) {
+      if ((err = mp_sqrmod (&y, a, &y)) != MP_OKAY) {
+         goto __Y;
+      }
+
+      /* if y == 1 then composite */
+      if (mp_cmp_d (&y, 1) == MP_EQ) {
+         goto __Y;
+      }
+
+      ++j;
+    }
+
+    /* if y != n1 then composite */
+    if (mp_cmp (&y, &n1) != MP_EQ) {
+      goto __Y;
+    }
+  }
+
+  /* probably prime now */
+  *result = MP_YES;
+__Y:mp_clear (&y);
+__R:mp_clear (&r);
+__N1:mp_clear (&n1);
+  return err;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_prime_next_prime.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,164 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* finds the next prime after the number "a" using "t" trials
+ * of Miller-Rabin.
+ *
+ * bbs_style = 1 means the prime must be congruent to 3 mod 4
+ */
+int mp_prime_next_prime(mp_int *a, int t, int bbs_style)
+{
+   int      err, res, x, y;
+   mp_digit res_tab[PRIME_SIZE], step, kstep;
+   mp_int   b;
+
+   /* ensure t is valid */
+   if (t <= 0 || t > PRIME_SIZE) {
+      return MP_VAL;
+   }
+
+   /* force positive */
+   a->sign = MP_ZPOS;
+
+   /* simple algo if a is less than the largest prime in the table */
+   if (mp_cmp_d(a, __prime_tab[PRIME_SIZE-1]) == MP_LT) {
+      /* find which prime it is bigger than */
+      for (x = PRIME_SIZE - 2; x >= 0; x--) {
+          if (mp_cmp_d(a, __prime_tab[x]) != MP_LT) {
+             if (bbs_style == 1) {
+                /* ok we found a prime smaller or
+                 * equal [so the next is larger]
+                 *
+                 * however, the prime must be
+                 * congruent to 3 mod 4
+                 */
+                if ((__prime_tab[x + 1] & 3) != 3) {
+                   /* scan upwards for a prime congruent to 3 mod 4 */
+                   for (y = x + 1; y < PRIME_SIZE; y++) {
+                       if ((__prime_tab[y] & 3) == 3) {
+                          mp_set(a, __prime_tab[y]);
+                          return MP_OKAY;
+                       }
+                   }
+                }
+             } else {
+                mp_set(a, __prime_tab[x + 1]);
+                return MP_OKAY;
+             }
+          }
+      }
+      /* at this point a maybe 1 */
+      if (mp_cmp_d(a, 1) == MP_EQ) {
+         mp_set(a, 2);
+         return MP_OKAY;
+      }
+      /* fall through to the sieve */
+   }
+
+   /* generate a prime congruent to 3 mod 4 or 1/3 mod 4? */
+   if (bbs_style == 1) {
+      kstep   = 4;
+   } else {
+      kstep   = 2;
+   }
+
+   /* at this point we will use a combination of a sieve and Miller-Rabin */
+
+   if (bbs_style == 1) {
+      /* if a mod 4 != 3 subtract the correct value to make it so */
+      if ((a->dp[0] & 3) != 3) {
+         if ((err = mp_sub_d(a, (a->dp[0] & 3) + 1, a)) != MP_OKAY) { return err; };
+      }
+   } else {
+      if (mp_iseven(a) == 1) {
+         /* force odd */
+         if ((err = mp_sub_d(a, 1, a)) != MP_OKAY) {
+            return err;
+         }
+      }
+   }
+
+   /* generate the restable */
+   for (x = 1; x < PRIME_SIZE; x++) {
+      if ((err = mp_mod_d(a, __prime_tab[x], res_tab + x)) != MP_OKAY) {
+         return err;
+      }
+   }
+
+   /* init temp used for Miller-Rabin Testing */
+   if ((err = mp_init(&b)) != MP_OKAY) {
+      return err;
+   }
+
+   for (;;) {
+      /* skip to the next non-trivially divisible candidate */
+      step = 0;
+      do {
+         /* y == 1 if any residue was zero [e.g. cannot be prime] */
+         y     =  0;
+
+         /* increase step to next candidate */
+         step += kstep;
+
+         /* compute the new residue without using division */
+         for (x = 1; x < PRIME_SIZE; x++) {
+             /* add the step to each residue */
+             res_tab[x] += kstep;
+
+             /* subtract the modulus [instead of using division] */
+             if (res_tab[x] >= __prime_tab[x]) {
+                res_tab[x]  -= __prime_tab[x];
+             }
+
+             /* set flag if zero */
+             if (res_tab[x] == 0) {
+                y = 1;
+             }
+         }
+      } while (y == 1 && step < ((((mp_digit)1)<<DIGIT_BIT) - kstep));
+
+      /* add the step */
+      if ((err = mp_add_d(a, step, a)) != MP_OKAY) {
+         goto __ERR;
+      }
+
+      /* if didn't pass sieve and step == MAX then skip test */
+      if (y == 1 && step >= ((((mp_digit)1)<<DIGIT_BIT) - kstep)) {
+         continue;
+      }
+
+      /* is this prime? */
+      for (x = 0; x < t; x++) {
+          mp_set(&b, __prime_tab[t]);
+          if ((err = mp_prime_miller_rabin(a, &b, &res)) != MP_OKAY) {
+             goto __ERR;
+          }
+          if (res == MP_NO) {
+             break;
+          }
+      }
+
+      if (res == MP_YES) {
+         break;
+      }
+   }
+
+   err = MP_OKAY;
+__ERR:
+   mp_clear(&b);
+   return err;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_prime_random_ex.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,118 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* makes a truly random prime of a given size (bits),
+ *
+ * Flags are as follows:
+ * 
+ *   LTM_PRIME_BBS      - make prime congruent to 3 mod 4
+ *   LTM_PRIME_SAFE     - make sure (p-1)/2 is prime as well (implies LTM_PRIME_BBS)
+ *   LTM_PRIME_2MSB_OFF - make the 2nd highest bit zero
+ *   LTM_PRIME_2MSB_ON  - make the 2nd highest bit one
+ *
+ * You have to supply a callback which fills in a buffer with random bytes.  "dat" is a parameter you can
+ * have passed to the callback (e.g. a state or something).  This function doesn't use "dat" itself
+ * so it can be NULL
+ *
+ */
+
+/* This is possibly the mother of all prime generation functions, muahahahahaha! */
+int mp_prime_random_ex(mp_int *a, int t, int size, int flags, ltm_prime_callback cb, void *dat)
+{
+   unsigned char *tmp, maskAND, maskOR_msb, maskOR_lsb;
+   int res, err, bsize, maskOR_msb_offset;
+
+   /* sanity check the input */
+   if (size <= 1 || t <= 0) {
+      return MP_VAL;
+   }
+
+   /* LTM_PRIME_SAFE implies LTM_PRIME_BBS */
+   if (flags & LTM_PRIME_SAFE) {
+      flags |= LTM_PRIME_BBS;
+   }
+
+   /* calc the byte size */
+   bsize = (size>>3)+(size&7?1:0);
+
+   /* we need a buffer of bsize bytes */
+   tmp = OPT_CAST(unsigned char) XMALLOC(bsize);
+   if (tmp == NULL) {
+      return MP_MEM;
+   }
+
+   /* calc the maskAND value for the MSbyte*/
+   maskAND = 0xFF >> (8 - (size & 7));
+
+   /* calc the maskOR_msb */
+   maskOR_msb        = 0;
+   maskOR_msb_offset = (size - 2) >> 3;
+   if (flags & LTM_PRIME_2MSB_ON) {
+      maskOR_msb     |= 1 << ((size - 2) & 7);
+   } else if (flags & LTM_PRIME_2MSB_OFF) {
+      maskAND        &= ~(1 << ((size - 2) & 7));
+   }
+
+   /* get the maskOR_lsb */
+   maskOR_lsb         = 0;
+   if (flags & LTM_PRIME_BBS) {
+      maskOR_lsb     |= 3;
+   }
+
+   do {
+      /* read the bytes */
+      if (cb(tmp, bsize, dat) != bsize) {
+         err = MP_VAL;
+         goto error;
+      }
+ 
+      /* work over the MSbyte */
+      tmp[0]    &= maskAND;
+      tmp[0]    |= 1 << ((size - 1) & 7);
+
+      /* mix in the maskORs */
+      tmp[maskOR_msb_offset]   |= maskOR_msb;
+      tmp[bsize-1]             |= maskOR_lsb;
+
+      /* read it in */
+      if ((err = mp_read_unsigned_bin(a, tmp, bsize)) != MP_OKAY)     { goto error; }
+
+      /* is it prime? */
+      if ((err = mp_prime_is_prime(a, t, &res)) != MP_OKAY)           { goto error; }
+
+      if (flags & LTM_PRIME_SAFE) {
+         /* see if (a-1)/2 is prime */
+         if ((err = mp_sub_d(a, 1, a)) != MP_OKAY)                    { goto error; }
+         if ((err = mp_div_2(a, a)) != MP_OKAY)                       { goto error; }
+ 
+         /* is it prime? */
+         if ((err = mp_prime_is_prime(a, t, &res)) != MP_OKAY)        { goto error; }
+      }
+   } while (res == MP_NO);
+
+   if (flags & LTM_PRIME_SAFE) {
+      /* restore a to the original value */
+      if ((err = mp_mul_2(a, a)) != MP_OKAY)                          { goto error; }
+      if ((err = mp_add_d(a, 1, a)) != MP_OKAY)                       { goto error; }
+   }
+
+   err = MP_OKAY;
+error:
+   XFREE(tmp);
+   return err;
+}
+
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_radix_size.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,65 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* returns size of ASCII reprensentation */
+int mp_radix_size (mp_int * a, int radix, int *size)
+{
+  int     res, digs;
+  mp_int  t;
+  mp_digit d;
+
+  *size = 0;
+
+  /* special case for binary */
+  if (radix == 2) {
+    *size = mp_count_bits (a) + (a->sign == MP_NEG ? 1 : 0) + 1;
+    return MP_OKAY;
+  }
+
+  /* make sure the radix is in range */
+  if (radix < 2 || radix > 64) {
+    return MP_VAL;
+  }
+
+  /* init a copy of the input */
+  if ((res = mp_init_copy (&t, a)) != MP_OKAY) {
+    return res;
+  }
+
+  /* digs is the digit count */
+  digs = 0;
+
+  /* if it's negative add one for the sign */
+  if (t.sign == MP_NEG) {
+    ++digs;
+    t.sign = MP_ZPOS;
+  }
+
+  /* fetch out all of the digits */
+  while (mp_iszero (&t) == 0) {
+    if ((res = mp_div_d (&t, (mp_digit) radix, &t, &d)) != MP_OKAY) {
+      mp_clear (&t);
+      return res;
+    }
+    ++digs;
+  }
+  mp_clear (&t);
+
+  /* return digs + 1, the 1 is for the NULL byte that would be required. */
+  *size = digs + 1;
+  return MP_OKAY;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_radix_smap.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,18 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* chars used in radix conversions */
+const char *mp_s_rmap = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz+/";
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_rand.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,49 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* makes a pseudo-random int of a given size */
+int
+mp_rand (mp_int * a, int digits)
+{
+  int     res;
+  mp_digit d;
+
+  mp_zero (a);
+  if (digits <= 0) {
+    return MP_OKAY;
+  }
+
+  /* first place a random non-zero digit */
+  do {
+    d = ((mp_digit) abs (rand ()));
+  } while (d == 0);
+
+  if ((res = mp_add_d (a, d, a)) != MP_OKAY) {
+    return res;
+  }
+
+  while (digits-- > 0) {
+    if ((res = mp_lshd (a, 1)) != MP_OKAY) {
+      return res;
+    }
+
+    if ((res = mp_add_d (a, ((mp_digit) abs (rand ())), a)) != MP_OKAY) {
+      return res;
+    }
+  }
+
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_read_radix.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,76 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* read a string [ASCII] in a given radix */
+int mp_read_radix (mp_int * a, char *str, int radix)
+{
+  int     y, res, neg;
+  char    ch;
+
+  /* make sure the radix is ok */
+  if (radix < 2 || radix > 64) {
+    return MP_VAL;
+  }
+
+  /* if the leading digit is a 
+   * minus set the sign to negative. 
+   */
+  if (*str == '-') {
+    ++str;
+    neg = MP_NEG;
+  } else {
+    neg = MP_ZPOS;
+  }
+
+  /* set the integer to the default of zero */
+  mp_zero (a);
+  
+  /* process each digit of the string */
+  while (*str) {
+    /* if the radix < 36 the conversion is case insensitive
+     * this allows numbers like 1AB and 1ab to represent the same  value
+     * [e.g. in hex]
+     */
+    ch = (char) ((radix < 36) ? toupper (*str) : *str);
+    for (y = 0; y < 64; y++) {
+      if (ch == mp_s_rmap[y]) {
+         break;
+      }
+    }
+
+    /* if the char was found in the map 
+     * and is less than the given radix add it
+     * to the number, otherwise exit the loop. 
+     */
+    if (y < radix) {
+      if ((res = mp_mul_d (a, (mp_digit) radix, a)) != MP_OKAY) {
+         return res;
+      }
+      if ((res = mp_add_d (a, (mp_digit) y, a)) != MP_OKAY) {
+         return res;
+      }
+    } else {
+      break;
+    }
+    ++str;
+  }
+  
+  /* set the sign only if a != 0 */
+  if (mp_iszero(a) != 1) {
+     a->sign = neg;
+  }
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_read_signed_bin.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,36 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* read signed bin, big endian, first byte is 0==positive or 1==negative */
+int
+mp_read_signed_bin (mp_int * a, unsigned char *b, int c)
+{
+  int     res;
+
+  /* read magnitude */
+  if ((res = mp_read_unsigned_bin (a, b + 1, c - 1)) != MP_OKAY) {
+    return res;
+  }
+
+  /* first byte is 0 for positive, non-zero for negative */
+  if (b[0] == 0) {
+     a->sign = MP_ZPOS;
+  } else {
+     a->sign = MP_NEG;
+  }
+
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_read_unsigned_bin.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,50 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* reads a unsigned char array, assumes the msb is stored first [big endian] */
+int
+mp_read_unsigned_bin (mp_int * a, unsigned char *b, int c)
+{
+  int     res;
+
+  /* make sure there are at least two digits */
+  if (a->alloc < 2) {
+     if ((res = mp_grow(a, 2)) != MP_OKAY) {
+        return res;
+     }
+  }
+
+  /* zero the int */
+  mp_zero (a);
+
+  /* read the bytes in */
+  while (c-- > 0) {
+    if ((res = mp_mul_2d (a, 8, a)) != MP_OKAY) {
+      return res;
+    }
+
+#ifndef MP_8BIT
+      a->dp[0] |= *b++;
+      a->used += 1;
+#else
+      a->dp[0] = (*b & MP_MASK);
+      a->dp[1] |= ((*b++ >> 7U) & 1);
+      a->used += 2;
+#endif
+  }
+  mp_clamp (a);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_reduce.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,84 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* reduces x mod m, assumes 0 < x < m**2, mu is 
+ * precomputed via mp_reduce_setup.
+ * From HAC pp.604 Algorithm 14.42
+ */
+int
+mp_reduce (mp_int * x, mp_int * m, mp_int * mu)
+{
+  mp_int  q;
+  int     res, um = m->used;
+
+  /* q = x */
+  if ((res = mp_init_copy (&q, x)) != MP_OKAY) {
+    return res;
+  }
+
+  /* q1 = x / b**(k-1)  */
+  mp_rshd (&q, um - 1);         
+
+  /* according to HAC this optimization is ok */
+  if (((unsigned long) um) > (((mp_digit)1) << (DIGIT_BIT - 1))) {
+    if ((res = mp_mul (&q, mu, &q)) != MP_OKAY) {
+      goto CLEANUP;
+    }
+  } else {
+    if ((res = s_mp_mul_high_digs (&q, mu, &q, um - 1)) != MP_OKAY) {
+      goto CLEANUP;
+    }
+  }
+
+  /* q3 = q2 / b**(k+1) */
+  mp_rshd (&q, um + 1);         
+
+  /* x = x mod b**(k+1), quick (no division) */
+  if ((res = mp_mod_2d (x, DIGIT_BIT * (um + 1), x)) != MP_OKAY) {
+    goto CLEANUP;
+  }
+
+  /* q = q * m mod b**(k+1), quick (no division) */
+  if ((res = s_mp_mul_digs (&q, m, &q, um + 1)) != MP_OKAY) {
+    goto CLEANUP;
+  }
+
+  /* x = x - q */
+  if ((res = mp_sub (x, &q, x)) != MP_OKAY) {
+    goto CLEANUP;
+  }
+
+  /* If x < 0, add b**(k+1) to it */
+  if (mp_cmp_d (x, 0) == MP_LT) {
+    mp_set (&q, 1);
+    if ((res = mp_lshd (&q, um + 1)) != MP_OKAY)
+      goto CLEANUP;
+    if ((res = mp_add (x, &q, x)) != MP_OKAY)
+      goto CLEANUP;
+  }
+
+  /* Back off if it's too big */
+  while (mp_cmp (x, m) != MP_LT) {
+    if ((res = s_mp_sub (x, m, x)) != MP_OKAY) {
+      goto CLEANUP;
+    }
+  }
+  
+CLEANUP:
+  mp_clear (&q);
+
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_reduce_2k.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,56 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* reduces a modulo n where n is of the form 2**p - d */
+int
+mp_reduce_2k(mp_int *a, mp_int *n, mp_digit d)
+{
+   mp_int q;
+   int    p, res;
+   
+   if ((res = mp_init(&q)) != MP_OKAY) {
+      return res;
+   }
+   
+   p = mp_count_bits(n);    
+top:
+   /* q = a/2**p, a = a mod 2**p */
+   if ((res = mp_div_2d(a, p, &q, a)) != MP_OKAY) {
+      goto ERR;
+   }
+   
+   if (d != 1) {
+      /* q = q * d */
+      if ((res = mp_mul_d(&q, d, &q)) != MP_OKAY) { 
+         goto ERR;
+      }
+   }
+   
+   /* a = a + q */
+   if ((res = s_mp_add(a, &q, a)) != MP_OKAY) {
+      goto ERR;
+   }
+   
+   if (mp_cmp_mag(a, n) != MP_LT) {
+      s_mp_sub(a, n, a);
+      goto top;
+   }
+   
+ERR:
+   mp_clear(&q);
+   return res;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_reduce_2k_setup.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,42 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* determines the setup value */
+int 
+mp_reduce_2k_setup(mp_int *a, mp_digit *d)
+{
+   int res, p;
+   mp_int tmp;
+   
+   if ((res = mp_init(&tmp)) != MP_OKAY) {
+      return res;
+   }
+   
+   p = mp_count_bits(a);
+   if ((res = mp_2expt(&tmp, p)) != MP_OKAY) {
+      mp_clear(&tmp);
+      return res;
+   }
+   
+   if ((res = s_mp_sub(&tmp, a, &tmp)) != MP_OKAY) {
+      mp_clear(&tmp);
+      return res;
+   }
+   
+   *d = tmp.dp[0];
+   mp_clear(&tmp);
+   return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_reduce_is_2k.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,45 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* determines if mp_reduce_2k can be used */
+int mp_reduce_is_2k(mp_int *a)
+{
+   int ix, iy, iz, iw;
+   
+   if (a->used == 0) {
+      return 0;
+   } else if (a->used == 1) {
+      return 1;
+   } else if (a->used > 1) {
+      iy = mp_count_bits(a);
+      iz = 1;
+      iw = 1;
+    
+      /* Test every bit from the second digit up, must be 1 */
+      for (ix = DIGIT_BIT; ix < iy; ix++) {
+          if ((a->dp[iw] & iz) == 0) {
+             return 0;
+          }
+          iz <<= 1;
+          if (iz > (int)MP_MASK) {
+             ++iw;
+             iz = 1;
+          }
+      }
+   }
+   return 1;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_reduce_setup.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,29 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* pre-calculate the value required for Barrett reduction
+ * For a given modulus "b" it calulates the value required in "a"
+ */
+int
+mp_reduce_setup (mp_int * a, mp_int * b)
+{
+  int     res;
+  
+  if ((res = mp_2expt (a, b->used * 2 * DIGIT_BIT)) != MP_OKAY) {
+    return res;
+  }
+  return mp_div (a, b, a, NULL);
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_rshd.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,66 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* shift right a certain amount of digits */
+void mp_rshd (mp_int * a, int b)
+{
+  int     x;
+
+  /* if b <= 0 then ignore it */
+  if (b <= 0) {
+    return;
+  }
+
+  /* if b > used then simply zero it and return */
+  if (a->used <= b) {
+    mp_zero (a);
+    return;
+  }
+
+  {
+    register mp_digit *bottom, *top;
+
+    /* shift the digits down */
+
+    /* bottom */
+    bottom = a->dp;
+
+    /* top [offset into digits] */
+    top = a->dp + b;
+
+    /* this is implemented as a sliding window where 
+     * the window is b-digits long and digits from 
+     * the top of the window are copied to the bottom
+     *
+     * e.g.
+
+     b-2 | b-1 | b0 | b1 | b2 | ... | bb |   ---->
+                 /\                   |      ---->
+                  \-------------------/      ---->
+     */
+    for (x = 0; x < (a->used - b); x++) {
+      *bottom++ = *top++;
+    }
+
+    /* zero the top digits */
+    for (; x < a->used; x++) {
+      *bottom++ = 0;
+    }
+  }
+  
+  /* remove excess digits */
+  a->used -= b;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_set.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,23 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* set to a digit */
+void mp_set (mp_int * a, mp_digit b)
+{
+  mp_zero (a);
+  a->dp[0] = b & MP_MASK;
+  a->used  = (a->dp[0] != 0) ? 1 : 0;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_set_int.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,42 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* set a 32-bit const */
+int mp_set_int (mp_int * a, unsigned long b)
+{
+  int     x, res;
+
+  mp_zero (a);
+  
+  /* set four bits at a time */
+  for (x = 0; x < 8; x++) {
+    /* shift the number up four bits */
+    if ((res = mp_mul_2d (a, 4, a)) != MP_OKAY) {
+      return res;
+    }
+
+    /* OR in the top four bits of the source */
+    a->dp[0] |= (b >> 28) & 15;
+
+    /* shift the source up to the next four bits */
+    b <<= 4;
+
+    /* ensure that digits are not clamped off */
+    a->used += 1;
+  }
+  mp_clamp (a);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_shrink.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,29 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* shrink a bignum */
+int mp_shrink (mp_int * a)
+{
+  mp_digit *tmp;
+  if (a->alloc != a->used && a->used > 0) {
+    if ((tmp = OPT_CAST(mp_digit) XREALLOC (a->dp, sizeof (mp_digit) * a->used)) == NULL) {
+      return MP_MEM;
+    }
+    a->dp    = tmp;
+    a->alloc = a->used;
+  }
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_signed_bin_size.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,21 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* get the size for an signed equivalent */
+int mp_signed_bin_size (mp_int * a)
+{
+  return 1 + mp_unsigned_bin_size (a);
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_sqr.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,41 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* computes b = a*a */
+int
+mp_sqr (mp_int * a, mp_int * b)
+{
+  int     res;
+
+  /* use Toom-Cook? */
+  if (a->used >= TOOM_SQR_CUTOFF) {
+    res = mp_toom_sqr(a, b);
+  /* Karatsuba? */
+  } else if (a->used >= KARATSUBA_SQR_CUTOFF) {
+    res = mp_karatsuba_sqr (a, b);
+  } else {
+    /* can we use the fast comba multiplier? */
+    if ((a->used * 2 + 1) < MP_WARRAY && 
+         a->used < 
+         (1 << (sizeof(mp_word) * CHAR_BIT - 2*DIGIT_BIT - 1))) {
+      res = fast_s_mp_sqr (a, b);
+    } else {
+      res = s_mp_sqr (a, b);
+    }
+  }
+  b->sign = MP_ZPOS;
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_sqrmod.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,35 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* c = a * a (mod b) */
+int
+mp_sqrmod (mp_int * a, mp_int * b, mp_int * c)
+{
+  int     res;
+  mp_int  t;
+
+  if ((res = mp_init (&t)) != MP_OKAY) {
+    return res;
+  }
+
+  if ((res = mp_sqr (a, &t)) != MP_OKAY) {
+    mp_clear (&t);
+    return res;
+  }
+  res = mp_mod (&t, b, c);
+  mp_clear (&t);
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_sqrt.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,75 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* this function is less generic than mp_n_root, simpler and faster */
+int mp_sqrt(mp_int *arg, mp_int *ret) 
+{
+  int res;
+  mp_int t1,t2;
+
+  /* must be positive */
+  if (arg->sign == MP_NEG) {
+    return MP_VAL;
+  }
+
+  /* easy out */
+  if (mp_iszero(arg) == MP_YES) {
+    mp_zero(ret);
+    return MP_OKAY;
+  }
+
+  if ((res = mp_init_copy(&t1, arg)) != MP_OKAY) {
+    return res;
+  }
+
+  if ((res = mp_init(&t2)) != MP_OKAY) {
+    goto E2;
+  }
+
+  /* First approx. (not very bad for large arg) */
+  mp_rshd (&t1,t1.used/2);
+
+  /* t1 > 0  */ 
+  if ((res = mp_div(arg,&t1,&t2,NULL)) != MP_OKAY) {
+    goto E1;
+  }
+  if ((res = mp_add(&t1,&t2,&t1)) != MP_OKAY) {
+    goto E1;
+  }
+  if ((res = mp_div_2(&t1,&t1)) != MP_OKAY) {
+    goto E1;
+  }
+  /* And now t1 > sqrt(arg) */
+  do { 
+    if ((res = mp_div(arg,&t1,&t2,NULL)) != MP_OKAY) {
+      goto E1;
+    }
+    if ((res = mp_add(&t1,&t2,&t1)) != MP_OKAY) {
+      goto E1;
+    }
+    if ((res = mp_div_2(&t1,&t1)) != MP_OKAY) {
+      goto E1;
+    }
+    /* t1 >= sqrt(arg) >= t2 at this point */
+  } while (mp_cmp_mag(&t1,&t2) == MP_GT);
+
+  mp_exch(&t1,ret);
+
+E1: mp_clear(&t2);
+E2: mp_clear(&t1);
+  return res;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_sub.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,53 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* high level subtraction (handles signs) */
+int
+mp_sub (mp_int * a, mp_int * b, mp_int * c)
+{
+  int     sa, sb, res;
+
+  sa = a->sign;
+  sb = b->sign;
+
+  if (sa != sb) {
+    /* subtract a negative from a positive, OR */
+    /* subtract a positive from a negative. */
+    /* In either case, ADD their magnitudes, */
+    /* and use the sign of the first number. */
+    c->sign = sa;
+    res = s_mp_add (a, b, c);
+  } else {
+    /* subtract a positive from a positive, OR */
+    /* subtract a negative from a negative. */
+    /* First, take the difference between their */
+    /* magnitudes, then... */
+    if (mp_cmp_mag (a, b) != MP_LT) {
+      /* Copy the sign from the first */
+      c->sign = sa;
+      /* The first has a larger or equal magnitude */
+      res = s_mp_sub (a, b, c);
+    } else {
+      /* The result has the *opposite* sign from */
+      /* the first number. */
+      c->sign = (sa == MP_ZPOS) ? MP_NEG : MP_ZPOS;
+      /* The second has a larger magnitude */
+      res = s_mp_sub (b, a, c);
+    }
+  }
+  return res;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_sub_d.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,83 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* single digit subtraction */
+int
+mp_sub_d (mp_int * a, mp_digit b, mp_int * c)
+{
+  mp_digit *tmpa, *tmpc, mu;
+  int       res, ix, oldused;
+
+  /* grow c as required */
+  if (c->alloc < a->used + 1) {
+     if ((res = mp_grow(c, a->used + 1)) != MP_OKAY) {
+        return res;
+     }
+  }
+
+  /* if a is negative just do an unsigned
+   * addition [with fudged signs]
+   */
+  if (a->sign == MP_NEG) {
+     a->sign = MP_ZPOS;
+     res     = mp_add_d(a, b, c);
+     a->sign = c->sign = MP_NEG;
+     return res;
+  }
+
+  /* setup regs */
+  oldused = c->used;
+  tmpa    = a->dp;
+  tmpc    = c->dp;
+
+  /* if a <= b simply fix the single digit */
+  if ((a->used == 1 && a->dp[0] <= b) || a->used == 0) {
+     if (a->used == 1) {
+        *tmpc++ = b - *tmpa;
+     } else {
+        *tmpc++ = b;
+     }
+     ix      = 1;
+
+     /* negative/1digit */
+     c->sign = MP_NEG;
+     c->used = 1;
+  } else {
+     /* positive/size */
+     c->sign = MP_ZPOS;
+     c->used = a->used;
+
+     /* subtract first digit */
+     *tmpc    = *tmpa++ - b;
+     mu       = *tmpc >> (sizeof(mp_digit) * CHAR_BIT - 1);
+     *tmpc++ &= MP_MASK;
+
+     /* handle rest of the digits */
+     for (ix = 1; ix < a->used; ix++) {
+        *tmpc    = *tmpa++ - mu;
+        mu       = *tmpc >> (sizeof(mp_digit) * CHAR_BIT - 1);
+        *tmpc++ &= MP_MASK;
+     }
+  }
+
+  /* zero excess digits */
+  while (ix++ < oldused) {
+     *tmpc++ = 0;
+  }
+  mp_clamp(c);
+  return MP_OKAY;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_submod.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,36 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* d = a - b (mod c) */
+int
+mp_submod (mp_int * a, mp_int * b, mp_int * c, mp_int * d)
+{
+  int     res;
+  mp_int  t;
+
+
+  if ((res = mp_init (&t)) != MP_OKAY) {
+    return res;
+  }
+
+  if ((res = mp_sub (a, b, &t)) != MP_OKAY) {
+    mp_clear (&t);
+    return res;
+  }
+  res = mp_mod (&t, c, d);
+  mp_clear (&t);
+  return res;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_to_signed_bin.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,28 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* store in signed [big endian] format */
+int
+mp_to_signed_bin (mp_int * a, unsigned char *b)
+{
+  int     res;
+
+  if ((res = mp_to_unsigned_bin (a, b + 1)) != MP_OKAY) {
+    return res;
+  }
+  b[0] = (unsigned char) ((a->sign == MP_ZPOS) ? 0 : 1);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_to_unsigned_bin.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,43 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* store in unsigned [big endian] format */
+int
+mp_to_unsigned_bin (mp_int * a, unsigned char *b)
+{
+  int     x, res;
+  mp_int  t;
+
+  if ((res = mp_init_copy (&t, a)) != MP_OKAY) {
+    return res;
+  }
+
+  x = 0;
+  while (mp_iszero (&t) == 0) {
+#ifndef MP_8BIT
+      b[x++] = (unsigned char) (t.dp[0] & 255);
+#else
+      b[x++] = (unsigned char) (t.dp[0] | ((t.dp[1] & 0x01) << 7));
+#endif
+    if ((res = mp_div_2d (&t, 8, &t, NULL)) != MP_OKAY) {
+      mp_clear (&t);
+      return res;
+    }
+  }
+  bn_reverse (b, x);
+  mp_clear (&t);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_toom_mul.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,272 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* multiplication using the Toom-Cook 3-way algorithm */
+int mp_toom_mul(mp_int *a, mp_int *b, mp_int *c)
+{
+    mp_int w0, w1, w2, w3, w4, tmp1, tmp2, a0, a1, a2, b0, b1, b2;
+    int res, B;
+        
+    /* init temps */
+    if ((res = mp_init_multi(&w0, &w1, &w2, &w3, &w4, 
+                             &a0, &a1, &a2, &b0, &b1, 
+                             &b2, &tmp1, &tmp2, NULL)) != MP_OKAY) {
+       return res;
+    }
+    
+    /* B */
+    B = MIN(a->used, b->used) / 3;
+    
+    /* a = a2 * B**2 + a1 * B + a0 */
+    if ((res = mp_mod_2d(a, DIGIT_BIT * B, &a0)) != MP_OKAY) {
+       goto ERR;
+    }
+
+    if ((res = mp_copy(a, &a1)) != MP_OKAY) {
+       goto ERR;
+    }
+    mp_rshd(&a1, B);
+    mp_mod_2d(&a1, DIGIT_BIT * B, &a1);
+
+    if ((res = mp_copy(a, &a2)) != MP_OKAY) {
+       goto ERR;
+    }
+    mp_rshd(&a2, B*2);
+    
+    /* b = b2 * B**2 + b1 * B + b0 */
+    if ((res = mp_mod_2d(b, DIGIT_BIT * B, &b0)) != MP_OKAY) {
+       goto ERR;
+    }
+
+    if ((res = mp_copy(b, &b1)) != MP_OKAY) {
+       goto ERR;
+    }
+    mp_rshd(&b1, B);
+    mp_mod_2d(&b1, DIGIT_BIT * B, &b1);
+
+    if ((res = mp_copy(b, &b2)) != MP_OKAY) {
+       goto ERR;
+    }
+    mp_rshd(&b2, B*2);
+    
+    /* w0 = a0*b0 */
+    if ((res = mp_mul(&a0, &b0, &w0)) != MP_OKAY) {
+       goto ERR;
+    }
+    
+    /* w4 = a2 * b2 */
+    if ((res = mp_mul(&a2, &b2, &w4)) != MP_OKAY) {
+       goto ERR;
+    }
+    
+    /* w1 = (a2 + 2(a1 + 2a0))(b2 + 2(b1 + 2b0)) */
+    if ((res = mp_mul_2(&a0, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp1, &a1, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_mul_2(&tmp1, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp1, &a2, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    
+    if ((res = mp_mul_2(&b0, &tmp2)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp2, &b1, &tmp2)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_mul_2(&tmp2, &tmp2)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp2, &b2, &tmp2)) != MP_OKAY) {
+       goto ERR;
+    }
+    
+    if ((res = mp_mul(&tmp1, &tmp2, &w1)) != MP_OKAY) {
+       goto ERR;
+    }
+    
+    /* w3 = (a0 + 2(a1 + 2a2))(b0 + 2(b1 + 2b2)) */
+    if ((res = mp_mul_2(&a2, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp1, &a1, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_mul_2(&tmp1, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp1, &a0, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    
+    if ((res = mp_mul_2(&b2, &tmp2)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp2, &b1, &tmp2)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_mul_2(&tmp2, &tmp2)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp2, &b0, &tmp2)) != MP_OKAY) {
+       goto ERR;
+    }
+    
+    if ((res = mp_mul(&tmp1, &tmp2, &w3)) != MP_OKAY) {
+       goto ERR;
+    }
+    
+
+    /* w2 = (a2 + a1 + a0)(b2 + b1 + b0) */
+    if ((res = mp_add(&a2, &a1, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp1, &a0, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&b2, &b1, &tmp2)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp2, &b0, &tmp2)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_mul(&tmp1, &tmp2, &w2)) != MP_OKAY) {
+       goto ERR;
+    }
+    
+    /* now solve the matrix 
+    
+       0  0  0  0  1
+       1  2  4  8  16
+       1  1  1  1  1
+       16 8  4  2  1
+       1  0  0  0  0
+       
+       using 12 subtractions, 4 shifts, 
+              2 small divisions and 1 small multiplication 
+     */
+     
+     /* r1 - r4 */
+     if ((res = mp_sub(&w1, &w4, &w1)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r3 - r0 */
+     if ((res = mp_sub(&w3, &w0, &w3)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r1/2 */
+     if ((res = mp_div_2(&w1, &w1)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r3/2 */
+     if ((res = mp_div_2(&w3, &w3)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r2 - r0 - r4 */
+     if ((res = mp_sub(&w2, &w0, &w2)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_sub(&w2, &w4, &w2)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r1 - r2 */
+     if ((res = mp_sub(&w1, &w2, &w1)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r3 - r2 */
+     if ((res = mp_sub(&w3, &w2, &w3)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r1 - 8r0 */
+     if ((res = mp_mul_2d(&w0, 3, &tmp1)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_sub(&w1, &tmp1, &w1)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r3 - 8r4 */
+     if ((res = mp_mul_2d(&w4, 3, &tmp1)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_sub(&w3, &tmp1, &w3)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* 3r2 - r1 - r3 */
+     if ((res = mp_mul_d(&w2, 3, &w2)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_sub(&w2, &w1, &w2)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_sub(&w2, &w3, &w2)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r1 - r2 */
+     if ((res = mp_sub(&w1, &w2, &w1)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r3 - r2 */
+     if ((res = mp_sub(&w3, &w2, &w3)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r1/3 */
+     if ((res = mp_div_3(&w1, &w1, NULL)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r3/3 */
+     if ((res = mp_div_3(&w3, &w3, NULL)) != MP_OKAY) {
+        goto ERR;
+     }
+     
+     /* at this point shift W[n] by B*n */
+     if ((res = mp_lshd(&w1, 1*B)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_lshd(&w2, 2*B)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_lshd(&w3, 3*B)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_lshd(&w4, 4*B)) != MP_OKAY) {
+        goto ERR;
+     }     
+     
+     if ((res = mp_add(&w0, &w1, c)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_add(&w2, &w3, &tmp1)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_add(&w4, &tmp1, &tmp1)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_add(&tmp1, c, c)) != MP_OKAY) {
+        goto ERR;
+     }     
+     
+ERR:
+     mp_clear_multi(&w0, &w1, &w2, &w3, &w4, 
+                    &a0, &a1, &a2, &b0, &b1, 
+                    &b2, &tmp1, &tmp2, NULL);
+     return res;
+}     
+     
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_toom_sqr.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,220 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* squaring using Toom-Cook 3-way algorithm */
+int
+mp_toom_sqr(mp_int *a, mp_int *b)
+{
+    mp_int w0, w1, w2, w3, w4, tmp1, a0, a1, a2;
+    int res, B;
+
+    /* init temps */
+    if ((res = mp_init_multi(&w0, &w1, &w2, &w3, &w4, &a0, &a1, &a2, &tmp1, NULL)) != MP_OKAY) {
+       return res;
+    }
+
+    /* B */
+    B = a->used / 3;
+
+    /* a = a2 * B**2 + a1 * B + a0 */
+    if ((res = mp_mod_2d(a, DIGIT_BIT * B, &a0)) != MP_OKAY) {
+       goto ERR;
+    }
+
+    if ((res = mp_copy(a, &a1)) != MP_OKAY) {
+       goto ERR;
+    }
+    mp_rshd(&a1, B);
+    mp_mod_2d(&a1, DIGIT_BIT * B, &a1);
+
+    if ((res = mp_copy(a, &a2)) != MP_OKAY) {
+       goto ERR;
+    }
+    mp_rshd(&a2, B*2);
+
+    /* w0 = a0*a0 */
+    if ((res = mp_sqr(&a0, &w0)) != MP_OKAY) {
+       goto ERR;
+    }
+
+    /* w4 = a2 * a2 */
+    if ((res = mp_sqr(&a2, &w4)) != MP_OKAY) {
+       goto ERR;
+    }
+
+    /* w1 = (a2 + 2(a1 + 2a0))**2 */
+    if ((res = mp_mul_2(&a0, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp1, &a1, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_mul_2(&tmp1, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp1, &a2, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+
+    if ((res = mp_sqr(&tmp1, &w1)) != MP_OKAY) {
+       goto ERR;
+    }
+
+    /* w3 = (a0 + 2(a1 + 2a2))**2 */
+    if ((res = mp_mul_2(&a2, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp1, &a1, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_mul_2(&tmp1, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp1, &a0, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+
+    if ((res = mp_sqr(&tmp1, &w3)) != MP_OKAY) {
+       goto ERR;
+    }
+
+
+    /* w2 = (a2 + a1 + a0)**2 */
+    if ((res = mp_add(&a2, &a1, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_add(&tmp1, &a0, &tmp1)) != MP_OKAY) {
+       goto ERR;
+    }
+    if ((res = mp_sqr(&tmp1, &w2)) != MP_OKAY) {
+       goto ERR;
+    }
+
+    /* now solve the matrix
+
+       0  0  0  0  1
+       1  2  4  8  16
+       1  1  1  1  1
+       16 8  4  2  1
+       1  0  0  0  0
+
+       using 12 subtractions, 4 shifts, 2 small divisions and 1 small multiplication.
+     */
+
+     /* r1 - r4 */
+     if ((res = mp_sub(&w1, &w4, &w1)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r3 - r0 */
+     if ((res = mp_sub(&w3, &w0, &w3)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r1/2 */
+     if ((res = mp_div_2(&w1, &w1)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r3/2 */
+     if ((res = mp_div_2(&w3, &w3)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r2 - r0 - r4 */
+     if ((res = mp_sub(&w2, &w0, &w2)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_sub(&w2, &w4, &w2)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r1 - r2 */
+     if ((res = mp_sub(&w1, &w2, &w1)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r3 - r2 */
+     if ((res = mp_sub(&w3, &w2, &w3)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r1 - 8r0 */
+     if ((res = mp_mul_2d(&w0, 3, &tmp1)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_sub(&w1, &tmp1, &w1)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r3 - 8r4 */
+     if ((res = mp_mul_2d(&w4, 3, &tmp1)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_sub(&w3, &tmp1, &w3)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* 3r2 - r1 - r3 */
+     if ((res = mp_mul_d(&w2, 3, &w2)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_sub(&w2, &w1, &w2)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_sub(&w2, &w3, &w2)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r1 - r2 */
+     if ((res = mp_sub(&w1, &w2, &w1)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r3 - r2 */
+     if ((res = mp_sub(&w3, &w2, &w3)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r1/3 */
+     if ((res = mp_div_3(&w1, &w1, NULL)) != MP_OKAY) {
+        goto ERR;
+     }
+     /* r3/3 */
+     if ((res = mp_div_3(&w3, &w3, NULL)) != MP_OKAY) {
+        goto ERR;
+     }
+
+     /* at this point shift W[n] by B*n */
+     if ((res = mp_lshd(&w1, 1*B)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_lshd(&w2, 2*B)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_lshd(&w3, 3*B)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_lshd(&w4, 4*B)) != MP_OKAY) {
+        goto ERR;
+     }
+
+     if ((res = mp_add(&w0, &w1, b)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_add(&w2, &w3, &tmp1)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_add(&w4, &tmp1, &tmp1)) != MP_OKAY) {
+        goto ERR;
+     }
+     if ((res = mp_add(&tmp1, b, b)) != MP_OKAY) {
+        goto ERR;
+     }
+
+ERR:
+     mp_clear_multi(&w0, &w1, &w2, &w3, &w4, &a0, &a1, &a2, &tmp1, NULL);
+     return res;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_toradix.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,69 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* stores a bignum as a ASCII string in a given radix (2..64) */
+int mp_toradix (mp_int * a, char *str, int radix)
+{
+  int     res, digs;
+  mp_int  t;
+  mp_digit d;
+  char   *_s = str;
+
+  /* check range of the radix */
+  if (radix < 2 || radix > 64) {
+    return MP_VAL;
+  }
+
+  /* quick out if its zero */
+  if (mp_iszero(a) == 1) {
+     *str++ = '0';
+     *str = '\0';
+     return MP_OKAY;
+  }
+
+  if ((res = mp_init_copy (&t, a)) != MP_OKAY) {
+    return res;
+  }
+
+  /* if it is negative output a - */
+  if (t.sign == MP_NEG) {
+    ++_s;
+    *str++ = '-';
+    t.sign = MP_ZPOS;
+  }
+
+  digs = 0;
+  while (mp_iszero (&t) == 0) {
+    if ((res = mp_div_d (&t, (mp_digit) radix, &t, &d)) != MP_OKAY) {
+      mp_clear (&t);
+      return res;
+    }
+    *str++ = mp_s_rmap[d];
+    ++digs;
+  }
+
+  /* reverse the digits of the string.  In this case _s points
+   * to the first digit [exluding the sign] of the number]
+   */
+  bn_reverse ((unsigned char *)_s, digs);
+
+  /* append a NULL so the string is properly terminated */
+  *str = '\0';
+
+  mp_clear (&t);
+  return MP_OKAY;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_toradix_n.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,83 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* stores a bignum as a ASCII string in a given radix (2..64) 
+ *
+ * Stores upto maxlen-1 chars and always a NULL byte 
+ */
+int mp_toradix_n(mp_int * a, char *str, int radix, int maxlen)
+{
+  int     res, digs;
+  mp_int  t;
+  mp_digit d;
+  char   *_s = str;
+
+  /* check range of the maxlen, radix */
+  if (maxlen < 3 || radix < 2 || radix > 64) {
+    return MP_VAL;
+  }
+
+  /* quick out if its zero */
+  if (mp_iszero(a) == 1) {
+     *str++ = '0';
+     *str = '\0';
+     return MP_OKAY;
+  }
+
+  if ((res = mp_init_copy (&t, a)) != MP_OKAY) {
+    return res;
+  }
+
+  /* if it is negative output a - */
+  if (t.sign == MP_NEG) {
+    /* we have to reverse our digits later... but not the - sign!! */
+    ++_s;
+
+    /* store the flag and mark the number as positive */
+    *str++ = '-';
+    t.sign = MP_ZPOS;
+ 
+    /* subtract a char */
+    --maxlen;
+  }
+
+  digs = 0;
+  while (mp_iszero (&t) == 0) {
+    if ((res = mp_div_d (&t, (mp_digit) radix, &t, &d)) != MP_OKAY) {
+      mp_clear (&t);
+      return res;
+    }
+    *str++ = mp_s_rmap[d];
+    ++digs;
+
+    if (--maxlen == 1) {
+       /* no more room */
+       break;
+    }
+  }
+
+  /* reverse the digits of the string.  In this case _s points
+   * to the first digit [exluding the sign] of the number]
+   */
+  bn_reverse ((unsigned char *)_s, digs);
+
+  /* append a NULL so the string is properly terminated */
+  *str = '\0';
+
+  mp_clear (&t);
+  return MP_OKAY;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_unsigned_bin_size.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,23 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* get the size for an unsigned equivalent */
+int
+mp_unsigned_bin_size (mp_int * a)
+{
+  int     size = mp_count_bits (a);
+  return (size / 8 + ((size & 7) != 0 ? 1 : 0));
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_xor.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,45 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* XOR two ints together */
+int
+mp_xor (mp_int * a, mp_int * b, mp_int * c)
+{
+  int     res, ix, px;
+  mp_int  t, *x;
+
+  if (a->used > b->used) {
+    if ((res = mp_init_copy (&t, a)) != MP_OKAY) {
+      return res;
+    }
+    px = b->used;
+    x = b;
+  } else {
+    if ((res = mp_init_copy (&t, b)) != MP_OKAY) {
+      return res;
+    }
+    px = a->used;
+    x = a;
+  }
+
+  for (ix = 0; ix < px; ix++) {
+    t.dp[ix] ^= x->dp[ix];
+  }
+  mp_clamp (&t);
+  mp_exch (c, &t);
+  mp_clear (&t);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_mp_zero.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,24 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* set to zero */
+void
+mp_zero (mp_int * a)
+{
+  a->sign = MP_ZPOS;
+  a->used = 0;
+  memset (a->dp, 0, sizeof (mp_digit) * a->alloc);
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_prime_sizes_tab.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,51 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* this table gives the # of rabin miller trials for a prob of failure lower than 2^-96 */
+static const struct {
+   int k, t;
+} sizes[] = {
+{   128,    28 },
+{   256,    16 },
+{   384,    10 },
+{   512,     7 },
+{   640,     6 },
+{   768,     5 },
+{   896,     4 },
+{  1024,     4 },
+{  1152,     3 },
+{  1280,     3 },
+{  1408,     3 },
+{  1536,     3 },
+{  1664,     3 },
+{  1792,     2 } };
+
+/* returns # of RM trials required for a given bit size */
+int mp_prime_rabin_miller_trials(int size)
+{
+   int x;
+
+   for (x = 0; x < (int)(sizeof(sizes)/(sizeof(sizes[0]))); x++) {
+       if (sizes[x].k == size) {
+          return sizes[x].t;
+       } else if (sizes[x].k > size) {
+          return (x == 0) ? sizes[0].t : sizes[x - 1].t;
+       }
+   }
+   return 1;
+}
+
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_prime_tab.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,55 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+const mp_digit __prime_tab[] = {
+  0x0002, 0x0003, 0x0005, 0x0007, 0x000B, 0x000D, 0x0011, 0x0013,
+  0x0017, 0x001D, 0x001F, 0x0025, 0x0029, 0x002B, 0x002F, 0x0035,
+  0x003B, 0x003D, 0x0043, 0x0047, 0x0049, 0x004F, 0x0053, 0x0059,
+  0x0061, 0x0065, 0x0067, 0x006B, 0x006D, 0x0071, 0x007F,
+#ifndef MP_8BIT
+  0x0083,
+  0x0089, 0x008B, 0x0095, 0x0097, 0x009D, 0x00A3, 0x00A7, 0x00AD,
+  0x00B3, 0x00B5, 0x00BF, 0x00C1, 0x00C5, 0x00C7, 0x00D3, 0x00DF,
+  0x00E3, 0x00E5, 0x00E9, 0x00EF, 0x00F1, 0x00FB, 0x0101, 0x0107,
+  0x010D, 0x010F, 0x0115, 0x0119, 0x011B, 0x0125, 0x0133, 0x0137,
+
+  0x0139, 0x013D, 0x014B, 0x0151, 0x015B, 0x015D, 0x0161, 0x0167,
+  0x016F, 0x0175, 0x017B, 0x017F, 0x0185, 0x018D, 0x0191, 0x0199,
+  0x01A3, 0x01A5, 0x01AF, 0x01B1, 0x01B7, 0x01BB, 0x01C1, 0x01C9,
+  0x01CD, 0x01CF, 0x01D3, 0x01DF, 0x01E7, 0x01EB, 0x01F3, 0x01F7,
+  0x01FD, 0x0209, 0x020B, 0x021D, 0x0223, 0x022D, 0x0233, 0x0239,
+  0x023B, 0x0241, 0x024B, 0x0251, 0x0257, 0x0259, 0x025F, 0x0265,
+  0x0269, 0x026B, 0x0277, 0x0281, 0x0283, 0x0287, 0x028D, 0x0293,
+  0x0295, 0x02A1, 0x02A5, 0x02AB, 0x02B3, 0x02BD, 0x02C5, 0x02CF,
+
+  0x02D7, 0x02DD, 0x02E3, 0x02E7, 0x02EF, 0x02F5, 0x02F9, 0x0301,
+  0x0305, 0x0313, 0x031D, 0x0329, 0x032B, 0x0335, 0x0337, 0x033B,
+  0x033D, 0x0347, 0x0355, 0x0359, 0x035B, 0x035F, 0x036D, 0x0371,
+  0x0373, 0x0377, 0x038B, 0x038F, 0x0397, 0x03A1, 0x03A9, 0x03AD,
+  0x03B3, 0x03B9, 0x03C7, 0x03CB, 0x03D1, 0x03D7, 0x03DF, 0x03E5,
+  0x03F1, 0x03F5, 0x03FB, 0x03FD, 0x0407, 0x0409, 0x040F, 0x0419,
+  0x041B, 0x0425, 0x0427, 0x042D, 0x043F, 0x0443, 0x0445, 0x0449,
+  0x044F, 0x0455, 0x045D, 0x0463, 0x0469, 0x047F, 0x0481, 0x048B,
+
+  0x0493, 0x049D, 0x04A3, 0x04A9, 0x04B1, 0x04BD, 0x04C1, 0x04C7,
+  0x04CD, 0x04CF, 0x04D5, 0x04E1, 0x04EB, 0x04FD, 0x04FF, 0x0503,
+  0x0509, 0x050B, 0x0511, 0x0515, 0x0517, 0x051B, 0x0527, 0x0529,
+  0x052F, 0x0551, 0x0557, 0x055D, 0x0565, 0x0577, 0x0581, 0x058F,
+  0x0593, 0x0595, 0x0599, 0x059F, 0x05A7, 0x05AB, 0x05AD, 0x05B3,
+  0x05BF, 0x05C9, 0x05CB, 0x05CF, 0x05D1, 0x05D5, 0x05DB, 0x05E7,
+  0x05F3, 0x05FB, 0x0607, 0x060D, 0x0611, 0x0617, 0x061F, 0x0623,
+  0x062B, 0x062F, 0x063D, 0x0641, 0x0647, 0x0649, 0x064D, 0x0653
+#endif
+};
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_reverse.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,33 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* reverse an array, used for radix code */
+void
+bn_reverse (unsigned char *s, int len)
+{
+  int     ix, iy;
+  unsigned char t;
+
+  ix = 0;
+  iy = len - 1;
+  while (ix < iy) {
+    t     = s[ix];
+    s[ix] = s[iy];
+    s[iy] = t;
+    ++ix;
+    --iy;
+  }
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_s_mp_add.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,103 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* low level addition, based on HAC pp.594, Algorithm 14.7 */
+int
+s_mp_add (mp_int * a, mp_int * b, mp_int * c)
+{
+  mp_int *x;
+  int     olduse, res, min, max;
+
+  /* find sizes, we let |a| <= |b| which means we have to sort
+   * them.  "x" will point to the input with the most digits
+   */
+  if (a->used > b->used) {
+    min = b->used;
+    max = a->used;
+    x = a;
+  } else {
+    min = a->used;
+    max = b->used;
+    x = b;
+  }
+
+  /* init result */
+  if (c->alloc < max + 1) {
+    if ((res = mp_grow (c, max + 1)) != MP_OKAY) {
+      return res;
+    }
+  }
+
+  /* get old used digit count and set new one */
+  olduse = c->used;
+  c->used = max + 1;
+
+  {
+    register mp_digit u, *tmpa, *tmpb, *tmpc;
+    register int i;
+
+    /* alias for digit pointers */
+
+    /* first input */
+    tmpa = a->dp;
+
+    /* second input */
+    tmpb = b->dp;
+
+    /* destination */
+    tmpc = c->dp;
+
+    /* zero the carry */
+    u = 0;
+    for (i = 0; i < min; i++) {
+      /* Compute the sum at one digit, T[i] = A[i] + B[i] + U */
+      *tmpc = *tmpa++ + *tmpb++ + u;
+
+      /* U = carry bit of T[i] */
+      u = *tmpc >> ((mp_digit)DIGIT_BIT);
+
+      /* take away carry bit from T[i] */
+      *tmpc++ &= MP_MASK;
+    }
+
+    /* now copy higher words if any, that is in A+B 
+     * if A or B has more digits add those in 
+     */
+    if (min != max) {
+      for (; i < max; i++) {
+        /* T[i] = X[i] + U */
+        *tmpc = x->dp[i] + u;
+
+        /* U = carry bit of T[i] */
+        u = *tmpc >> ((mp_digit)DIGIT_BIT);
+
+        /* take away carry bit from T[i] */
+        *tmpc++ &= MP_MASK;
+      }
+    }
+
+    /* add carry */
+    *tmpc++ = u;
+
+    /* clear digits above oldused */
+    for (i = c->used; i < olduse; i++) {
+      *tmpc++ = 0;
+    }
+  }
+
+  mp_clamp (c);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_s_mp_exptmod.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,234 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+#ifdef MP_LOW_MEM
+   #define TAB_SIZE 32
+#else
+   #define TAB_SIZE 256
+#endif
+
+int s_mp_exptmod (mp_int * G, mp_int * X, mp_int * P, mp_int * Y)
+{
+  mp_int  M[TAB_SIZE], res, mu;
+  mp_digit buf;
+  int     err, bitbuf, bitcpy, bitcnt, mode, digidx, x, y, winsize;
+
+  /* find window size */
+  x = mp_count_bits (X);
+  if (x <= 7) {
+    winsize = 2;
+  } else if (x <= 36) {
+    winsize = 3;
+  } else if (x <= 140) {
+    winsize = 4;
+  } else if (x <= 450) {
+    winsize = 5;
+  } else if (x <= 1303) {
+    winsize = 6;
+  } else if (x <= 3529) {
+    winsize = 7;
+  } else {
+    winsize = 8;
+  }
+
+#ifdef MP_LOW_MEM
+    if (winsize > 5) {
+       winsize = 5;
+    }
+#endif
+
+  /* init M array */
+  /* init first cell */
+  if ((err = mp_init(&M[1])) != MP_OKAY) {
+     return err; 
+  }
+
+  /* now init the second half of the array */
+  for (x = 1<<(winsize-1); x < (1 << winsize); x++) {
+    if ((err = mp_init(&M[x])) != MP_OKAY) {
+      for (y = 1<<(winsize-1); y < x; y++) {
+        mp_clear (&M[y]);
+      }
+      mp_clear(&M[1]);
+      return err;
+    }
+  }
+
+  /* create mu, used for Barrett reduction */
+  if ((err = mp_init (&mu)) != MP_OKAY) {
+    goto __M;
+  }
+  if ((err = mp_reduce_setup (&mu, P)) != MP_OKAY) {
+    goto __MU;
+  }
+
+  /* create M table
+   *
+   * The M table contains powers of the base, 
+   * e.g. M[x] = G**x mod P
+   *
+   * The first half of the table is not 
+   * computed though accept for M[0] and M[1]
+   */
+  if ((err = mp_mod (G, P, &M[1])) != MP_OKAY) {
+    goto __MU;
+  }
+
+  /* compute the value at M[1<<(winsize-1)] by squaring 
+   * M[1] (winsize-1) times 
+   */
+  if ((err = mp_copy (&M[1], &M[1 << (winsize - 1)])) != MP_OKAY) {
+    goto __MU;
+  }
+
+  for (x = 0; x < (winsize - 1); x++) {
+    if ((err = mp_sqr (&M[1 << (winsize - 1)], 
+                       &M[1 << (winsize - 1)])) != MP_OKAY) {
+      goto __MU;
+    }
+    if ((err = mp_reduce (&M[1 << (winsize - 1)], P, &mu)) != MP_OKAY) {
+      goto __MU;
+    }
+  }
+
+  /* create upper table, that is M[x] = M[x-1] * M[1] (mod P)
+   * for x = (2**(winsize - 1) + 1) to (2**winsize - 1)
+   */
+  for (x = (1 << (winsize - 1)) + 1; x < (1 << winsize); x++) {
+    if ((err = mp_mul (&M[x - 1], &M[1], &M[x])) != MP_OKAY) {
+      goto __MU;
+    }
+    if ((err = mp_reduce (&M[x], P, &mu)) != MP_OKAY) {
+      goto __MU;
+    }
+  }
+
+  /* setup result */
+  if ((err = mp_init (&res)) != MP_OKAY) {
+    goto __MU;
+  }
+  mp_set (&res, 1);
+
+  /* set initial mode and bit cnt */
+  mode   = 0;
+  bitcnt = 1;
+  buf    = 0;
+  digidx = X->used - 1;
+  bitcpy = 0;
+  bitbuf = 0;
+
+  for (;;) {
+    /* grab next digit as required */
+    if (--bitcnt == 0) {
+      /* if digidx == -1 we are out of digits */
+      if (digidx == -1) {
+        break;
+      }
+      /* read next digit and reset the bitcnt */
+      buf    = X->dp[digidx--];
+      bitcnt = (int) DIGIT_BIT;
+    }
+
+    /* grab the next msb from the exponent */
+    y     = (buf >> (mp_digit)(DIGIT_BIT - 1)) & 1;
+    buf <<= (mp_digit)1;
+
+    /* if the bit is zero and mode == 0 then we ignore it
+     * These represent the leading zero bits before the first 1 bit
+     * in the exponent.  Technically this opt is not required but it
+     * does lower the # of trivial squaring/reductions used
+     */
+    if (mode == 0 && y == 0) {
+      continue;
+    }
+
+    /* if the bit is zero and mode == 1 then we square */
+    if (mode == 1 && y == 0) {
+      if ((err = mp_sqr (&res, &res)) != MP_OKAY) {
+        goto __RES;
+      }
+      if ((err = mp_reduce (&res, P, &mu)) != MP_OKAY) {
+        goto __RES;
+      }
+      continue;
+    }
+
+    /* else we add it to the window */
+    bitbuf |= (y << (winsize - ++bitcpy));
+    mode    = 2;
+
+    if (bitcpy == winsize) {
+      /* ok window is filled so square as required and multiply  */
+      /* square first */
+      for (x = 0; x < winsize; x++) {
+        if ((err = mp_sqr (&res, &res)) != MP_OKAY) {
+          goto __RES;
+        }
+        if ((err = mp_reduce (&res, P, &mu)) != MP_OKAY) {
+          goto __RES;
+        }
+      }
+
+      /* then multiply */
+      if ((err = mp_mul (&res, &M[bitbuf], &res)) != MP_OKAY) {
+        goto __RES;
+      }
+      if ((err = mp_reduce (&res, P, &mu)) != MP_OKAY) {
+        goto __RES;
+      }
+
+      /* empty window and reset */
+      bitcpy = 0;
+      bitbuf = 0;
+      mode   = 1;
+    }
+  }
+
+  /* if bits remain then square/multiply */
+  if (mode == 2 && bitcpy > 0) {
+    /* square then multiply if the bit is set */
+    for (x = 0; x < bitcpy; x++) {
+      if ((err = mp_sqr (&res, &res)) != MP_OKAY) {
+        goto __RES;
+      }
+      if ((err = mp_reduce (&res, P, &mu)) != MP_OKAY) {
+        goto __RES;
+      }
+
+      bitbuf <<= 1;
+      if ((bitbuf & (1 << winsize)) != 0) {
+        /* then multiply */
+        if ((err = mp_mul (&res, &M[1], &res)) != MP_OKAY) {
+          goto __RES;
+        }
+        if ((err = mp_reduce (&res, P, &mu)) != MP_OKAY) {
+          goto __RES;
+        }
+      }
+    }
+  }
+
+  mp_exch (&res, Y);
+  err = MP_OKAY;
+__RES:mp_clear (&res);
+__MU:mp_clear (&mu);
+__M:
+  mp_clear(&M[1]);
+  for (x = 1<<(winsize-1); x < (1 << winsize); x++) {
+    mp_clear (&M[x]);
+  }
+  return err;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_s_mp_mul_digs.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,85 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* multiplies |a| * |b| and only computes upto digs digits of result
+ * HAC pp. 595, Algorithm 14.12  Modified so you can control how 
+ * many digits of output are created.
+ */
+int
+s_mp_mul_digs (mp_int * a, mp_int * b, mp_int * c, int digs)
+{
+  mp_int  t;
+  int     res, pa, pb, ix, iy;
+  mp_digit u;
+  mp_word r;
+  mp_digit tmpx, *tmpt, *tmpy;
+
+  /* can we use the fast multiplier? */
+  if (((digs) < MP_WARRAY) &&
+      MIN (a->used, b->used) < 
+          (1 << ((CHAR_BIT * sizeof (mp_word)) - (2 * DIGIT_BIT)))) {
+    return fast_s_mp_mul_digs (a, b, c, digs);
+  }
+
+  if ((res = mp_init_size (&t, digs)) != MP_OKAY) {
+    return res;
+  }
+  t.used = digs;
+
+  /* compute the digits of the product directly */
+  pa = a->used;
+  for (ix = 0; ix < pa; ix++) {
+    /* set the carry to zero */
+    u = 0;
+
+    /* limit ourselves to making digs digits of output */
+    pb = MIN (b->used, digs - ix);
+
+    /* setup some aliases */
+    /* copy of the digit from a used within the nested loop */
+    tmpx = a->dp[ix];
+    
+    /* an alias for the destination shifted ix places */
+    tmpt = t.dp + ix;
+    
+    /* an alias for the digits of b */
+    tmpy = b->dp;
+
+    /* compute the columns of the output and propagate the carry */
+    for (iy = 0; iy < pb; iy++) {
+      /* compute the column as a mp_word */
+      r       = ((mp_word)*tmpt) +
+                ((mp_word)tmpx) * ((mp_word)*tmpy++) +
+                ((mp_word) u);
+
+      /* the new column is the lower part of the result */
+      *tmpt++ = (mp_digit) (r & ((mp_word) MP_MASK));
+
+      /* get the carry word from the result */
+      u       = (mp_digit) (r >> ((mp_word) DIGIT_BIT));
+    }
+    /* set carry if it is placed below digs */
+    if (ix + iy < digs) {
+      *tmpt = u;
+    }
+  }
+
+  mp_clamp (&t);
+  mp_exch (&t, c);
+
+  mp_clear (&t);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_s_mp_mul_high_digs.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,73 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* multiplies |a| * |b| and does not compute the lower digs digits
+ * [meant to get the higher part of the product]
+ */
+int
+s_mp_mul_high_digs (mp_int * a, mp_int * b, mp_int * c, int digs)
+{
+  mp_int  t;
+  int     res, pa, pb, ix, iy;
+  mp_digit u;
+  mp_word r;
+  mp_digit tmpx, *tmpt, *tmpy;
+
+  /* can we use the fast multiplier? */
+  if (((a->used + b->used + 1) < MP_WARRAY)
+      && MIN (a->used, b->used) < (1 << ((CHAR_BIT * sizeof (mp_word)) - (2 * DIGIT_BIT)))) {
+    return fast_s_mp_mul_high_digs (a, b, c, digs);
+  }
+
+  if ((res = mp_init_size (&t, a->used + b->used + 1)) != MP_OKAY) {
+    return res;
+  }
+  t.used = a->used + b->used + 1;
+
+  pa = a->used;
+  pb = b->used;
+  for (ix = 0; ix < pa; ix++) {
+    /* clear the carry */
+    u = 0;
+
+    /* left hand side of A[ix] * B[iy] */
+    tmpx = a->dp[ix];
+
+    /* alias to the address of where the digits will be stored */
+    tmpt = &(t.dp[digs]);
+
+    /* alias for where to read the right hand side from */
+    tmpy = b->dp + (digs - ix);
+
+    for (iy = digs - ix; iy < pb; iy++) {
+      /* calculate the double precision result */
+      r       = ((mp_word)*tmpt) +
+                ((mp_word)tmpx) * ((mp_word)*tmpy++) +
+                ((mp_word) u);
+
+      /* get the lower part */
+      *tmpt++ = (mp_digit) (r & ((mp_word) MP_MASK));
+
+      /* carry the carry */
+      u       = (mp_digit) (r >> ((mp_word) DIGIT_BIT));
+    }
+    *tmpt = u;
+  }
+  mp_clamp (&t);
+  mp_exch (&t, c);
+  mp_clear (&t);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_s_mp_sqr.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,79 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* low level squaring, b = a*a, HAC pp.596-597, Algorithm 14.16 */
+int
+s_mp_sqr (mp_int * a, mp_int * b)
+{
+  mp_int  t;
+  int     res, ix, iy, pa;
+  mp_word r;
+  mp_digit u, tmpx, *tmpt;
+
+  pa = a->used;
+  if ((res = mp_init_size (&t, 2*pa + 1)) != MP_OKAY) {
+    return res;
+  }
+
+  /* default used is maximum possible size */
+  t.used = 2*pa + 1;
+
+  for (ix = 0; ix < pa; ix++) {
+    /* first calculate the digit at 2*ix */
+    /* calculate double precision result */
+    r = ((mp_word) t.dp[2*ix]) +
+        ((mp_word)a->dp[ix])*((mp_word)a->dp[ix]);
+
+    /* store lower part in result */
+    t.dp[ix+ix] = (mp_digit) (r & ((mp_word) MP_MASK));
+
+    /* get the carry */
+    u           = (mp_digit)(r >> ((mp_word) DIGIT_BIT));
+
+    /* left hand side of A[ix] * A[iy] */
+    tmpx        = a->dp[ix];
+
+    /* alias for where to store the results */
+    tmpt        = t.dp + (2*ix + 1);
+    
+    for (iy = ix + 1; iy < pa; iy++) {
+      /* first calculate the product */
+      r       = ((mp_word)tmpx) * ((mp_word)a->dp[iy]);
+
+      /* now calculate the double precision result, note we use
+       * addition instead of *2 since it's easier to optimize
+       */
+      r       = ((mp_word) *tmpt) + r + r + ((mp_word) u);
+
+      /* store lower part */
+      *tmpt++ = (mp_digit) (r & ((mp_word) MP_MASK));
+
+      /* get carry */
+      u       = (mp_digit)(r >> ((mp_word) DIGIT_BIT));
+    }
+    /* propagate upwards */
+    while (u != ((mp_digit) 0)) {
+      r       = ((mp_word) *tmpt) + ((mp_word) u);
+      *tmpt++ = (mp_digit) (r & ((mp_word) MP_MASK));
+      u       = (mp_digit)(r >> ((mp_word) DIGIT_BIT));
+    }
+  }
+
+  mp_clamp (&t);
+  mp_exch (&t, b);
+  mp_clear (&t);
+  return MP_OKAY;
+}
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bn_s_mp_sub.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,83 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* low level subtraction (assumes |a| > |b|), HAC pp.595 Algorithm 14.9 */
+int
+s_mp_sub (mp_int * a, mp_int * b, mp_int * c)
+{
+  int     olduse, res, min, max;
+
+  /* find sizes */
+  min = b->used;
+  max = a->used;
+
+  /* init result */
+  if (c->alloc < max) {
+    if ((res = mp_grow (c, max)) != MP_OKAY) {
+      return res;
+    }
+  }
+  olduse = c->used;
+  c->used = max;
+
+  {
+    register mp_digit u, *tmpa, *tmpb, *tmpc;
+    register int i;
+
+    /* alias for digit pointers */
+    tmpa = a->dp;
+    tmpb = b->dp;
+    tmpc = c->dp;
+
+    /* set carry to zero */
+    u = 0;
+    for (i = 0; i < min; i++) {
+      /* T[i] = A[i] - B[i] - U */
+      *tmpc = *tmpa++ - *tmpb++ - u;
+
+      /* U = carry bit of T[i]
+       * Note this saves performing an AND operation since
+       * if a carry does occur it will propagate all the way to the
+       * MSB.  As a result a single shift is enough to get the carry
+       */
+      u = *tmpc >> ((mp_digit)(CHAR_BIT * sizeof (mp_digit) - 1));
+
+      /* Clear carry from T[i] */
+      *tmpc++ &= MP_MASK;
+    }
+
+    /* now copy higher words if any, e.g. if A has more digits than B  */
+    for (; i < max; i++) {
+      /* T[i] = A[i] - U */
+      *tmpc = *tmpa++ - u;
+
+      /* U = carry bit of T[i] */
+      u = *tmpc >> ((mp_digit)(CHAR_BIT * sizeof (mp_digit) - 1));
+
+      /* Clear carry from T[i] */
+      *tmpc++ &= MP_MASK;
+    }
+
+    /* clear digits above used (since we may not have grown result above) */
+    for (i = c->used; i < olduse; i++) {
+      *tmpc++ = 0;
+    }
+  }
+
+  mp_clamp (c);
+  return MP_OKAY;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/bncore.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,31 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis
+ *
+ * LibTomMath is a library that provides multiple-precision
+ * integer arithmetic as well as number theoretic functionality.
+ *
+ * The library was designed directly after the MPI library by
+ * Michael Fromberger but has been written from scratch with
+ * additional optimizations in place.
+ *
+ * The library is free for all purposes without any express
+ * guarantee it works.
+ *
+ * Tom St Denis, [email protected], http://math.libtomcrypt.org
+ */
+#include <tommath.h>
+
+/* Known optimal configurations
+
+ CPU                    /Compiler     /MUL CUTOFF/SQR CUTOFF
+-------------------------------------------------------------
+ Intel P4               /GCC v3.2     /        70/       108
+ AMD Athlon XP          /GCC v3.2     /       109/       127
+
+*/
+
+/* configured for a AMD XP Thoroughbred core with etc/tune.c */
+int     KARATSUBA_MUL_CUTOFF = 109,      /* Min. number of digits before Karatsuba multiplication is used. */
+        KARATSUBA_SQR_CUTOFF = 127,      /* Min. number of digits before Karatsuba squaring is used. */
+        
+        TOOM_MUL_CUTOFF      = 350,      /* no optimal values of these are known yet so set em high */
+        TOOM_SQR_CUTOFF      = 400; 
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/booker.pl	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,261 @@
+#!/bin/perl
+#
+#Used to prepare the book "tommath.src" for LaTeX by pre-processing it into a .tex file
+#
+#Essentially you write the "tommath.src" as normal LaTex except where you want code snippets you put
+#
+#EXAM,file
+#
+#This preprocessor will then open "file" and insert it as a verbatim copy.
+#
+#Tom St Denis
+
+#get graphics type
+if (shift =~ /PDF/) {
+   $graph = "";
+} else {
+   $graph = ".ps";
+}   
+
+open(IN,"<tommath.src") or die "Can't open source file";
+open(OUT,">tommath.tex") or die "Can't open destination file";
+
+print "Scanning for sections\n";
+$chapter = $section = $subsection = 0;
+$x = 0;
+while (<IN>) {
+   print ".";
+   if (!(++$x % 80)) { print "\n"; }
+   #update the headings 
+   if (~($_ =~ /\*/)) {
+      if ($_ =~ /\\chapter{.+}/) {
+          ++$chapter;
+          $section = $subsection = 0;
+      } elsif ($_ =~ /\\section{.+}/) {
+          ++$section;
+          $subsection = 0;
+      } elsif ($_ =~ /\\subsection{.+}/) {
+          ++$subsection;
+      }
+   }      
+
+   if ($_ =~ m/MARK/) {
+      @m = split(",",$_);
+      chomp(@m[1]);
+      $index1{@m[1]} = $chapter;
+      $index2{@m[1]} = $section;
+      $index3{@m[1]} = $subsection;
+   }
+}
+close(IN);
+
+open(IN,"<tommath.src") or die "Can't open source file";
+$readline = $wroteline = 0;
+$srcline = 0;
+
+while (<IN>) {
+   ++$readline;
+   ++$srcline;
+   
+   if ($_ =~ m/MARK/) {
+   } elsif ($_ =~ m/EXAM/ || $_ =~ m/LIST/) {
+      if ($_ =~ m/EXAM/) {
+         $skipheader = 1;
+      } else {
+         $skipheader = 0;
+      }
+      
+      # EXAM,file
+      chomp($_);
+      @m = split(",",$_);
+      open(SRC,"<$m[1]") or die "Error:$srcline:Can't open source file $m[1]";
+      
+      print "$srcline:Inserting $m[1]:";
+      
+      $line = 0;
+      $tmp = $m[1];
+      $tmp =~ s/_/"\\_"/ge;
+      print OUT "\\vspace{+3mm}\\begin{small}\n\\hspace{-5.1mm}{\\bf File}: $tmp\n\\vspace{-3mm}\n\\begin{alltt}\n";
+      $wroteline += 5;
+      
+      if ($skipheader == 1) {
+         # scan till next end of comment, e.g. skip license 
+         while (<SRC>) {
+            $text[$line++] = $_;
+            last if ($_ =~ /tommath\.h/);
+         }
+      }
+      
+      $inline = 0;
+      while (<SRC>) {
+         $text[$line++] = $_;
+         ++$inline;
+         chomp($_);
+         $_ =~ s/\t/"    "/ge;
+         $_ =~ s/{/"^{"/ge;
+         $_ =~ s/}/"^}"/ge;
+         $_ =~ s/\\/'\symbol{92}'/ge;
+         $_ =~ s/\^/"\\"/ge;
+           
+         printf OUT ("%03d   ", $line);
+         for ($x = 0; $x < length($_); $x++) {
+             print OUT chr(vec($_, $x, 8));
+             if ($x == 75) { 
+                 print OUT "\n      ";
+                 ++$wroteline;
+             }
+         }
+         print OUT "\n";
+         ++$wroteline;
+      }
+      $totlines = $line;
+      print OUT "\\end{alltt}\n\\end{small}\n";
+      close(SRC);
+      print "$inline lines\n";
+      $wroteline += 2;
+   } elsif ($_ =~ m/@\d+,.+@/) {
+     # line contains [number,text]
+     # e.g. @14,for (ix = 0)@
+     $txt = $_;
+     while ($txt =~ m/@\d+,.+@/) {
+        @m = split("@",$txt);      # splits into text, one, two
+        @parms = split(",",$m[1]);  # splits one,two into two elements 
+                
+        # now search from $parms[0] down for $parms[1] 
+        $found1 = 0;
+        $found2 = 0;
+        for ($i = $parms[0]; $i < $totlines && $found1 == 0; $i++) {
+           if ($text[$i] =~ m/\Q$parms[1]\E/) {
+              $foundline1 = $i + 1;
+              $found1 = 1;
+           }
+        }
+        
+        # now search backwards
+        for ($i = $parms[0] - 1; $i >= 0 && $found2 == 0; $i--) {
+           if ($text[$i] =~ m/\Q$parms[1]\E/) {
+              $foundline2 = $i + 1;
+              $found2 = 1;
+           }
+        }
+        
+        # now use the closest match or the first if tied
+        if ($found1 == 1 && $found2 == 0) {
+           $found = 1;
+           $foundline = $foundline1;
+        } elsif ($found1 == 0 && $found2 == 1) {
+           $found = 1;
+           $foundline = $foundline2;
+        } elsif ($found1 == 1 && $found2 == 1) {
+           $found = 1;
+           if (($foundline1 - $parms[0]) <= ($parms[0] - $foundline2)) {
+              $foundline = $foundline1;
+           } else {
+              $foundline = $foundline2;
+           }
+        } else {
+           $found = 0;
+        }
+                      
+        # if found replace 
+        if ($found == 1) {
+           $delta = $parms[0] - $foundline;
+           print "Found replacement tag for \"$parms[1]\" on line $srcline which refers to line $foundline (delta $delta)\n";
+           $_ =~ s/@\Q$m[1]\E@/$foundline/;
+        } else {
+           print "ERROR:  The tag \"$parms[1]\" on line $srcline was not found in the most recently parsed source!\n";
+        }
+        
+        # remake the rest of the line 
+        $cnt = @m;
+        $txt = "";
+        for ($i = 2; $i < $cnt; $i++) {
+            $txt = $txt . $m[$i] . "@";
+        }
+     }
+     print OUT $_;
+     ++$wroteline;
+   } elsif ($_ =~ /~.+~/) {
+      # line contains a ~text~ pair used to refer to indexing :-)
+      $txt = $_;
+      while ($txt =~ /~.+~/) {
+         @m = split("~", $txt);
+         
+         # word is the second position
+         $word = @m[1];
+         $a = $index1{$word};
+         $b = $index2{$word};
+         $c = $index3{$word};
+         
+         # if chapter (a) is zero it wasn't found
+         if ($a == 0) {
+            print "ERROR: the tag \"$word\" on line $srcline was not found previously marked.\n";
+         } else {
+            # format the tag as x, x.y or x.y.z depending on the values
+            $str = $a;
+            $str = $str . ".$b" if ($b != 0);
+            $str = $str . ".$c" if ($c != 0);
+            
+            if ($b == 0 && $c == 0) {
+               # its a chapter
+               if ($a <= 10) {
+                  if ($a == 1) {
+                     $str = "chapter one";
+                  } elsif ($a == 2) {
+                     $str = "chapter two";
+                  } elsif ($a == 3) {
+                     $str = "chapter three";
+                  } elsif ($a == 4) {
+                     $str = "chapter four";
+                  } elsif ($a == 5) {
+                     $str = "chapter five";
+                  } elsif ($a == 6) {
+                     $str = "chapter six";
+                  } elsif ($a == 7) {
+                     $str = "chapter seven";
+                  } elsif ($a == 8) {
+                     $str = "chapter eight";
+                  } elsif ($a == 9) {
+                     $str = "chapter nine";
+                  } elsif ($a == 2) {
+                     $str = "chapter ten";
+                  }
+               } else {
+                  $str = "chapter " . $str;
+               }
+            } else {
+               $str = "section " . $str     if ($b != 0 && $c == 0);            
+               $str = "sub-section " . $str if ($b != 0 && $c != 0);
+            }
+            
+            #substitute
+            $_ =~ s/~\Q$word\E~/$str/;
+            
+            print "Found replacement tag for marker \"$word\" on line $srcline which refers to $str\n";
+         }
+         
+         # remake rest of the line
+         $cnt = @m;
+         $txt = "";
+         for ($i = 2; $i < $cnt; $i++) {
+             $txt = $txt . $m[$i] . "~";
+         }
+      }
+      print OUT $_;
+      ++$wroteline;
+   } elsif ($_ =~ m/FIGU/) {
+      # FIGU,file,caption
+      chomp($_);
+      @m = split(",", $_);
+      print OUT "\\begin{center}\n\\begin{figure}[here]\n\\includegraphics{pics/$m[1]$graph}\n";
+      print OUT "\\caption{$m[2]}\n\\label{pic:$m[1]}\n\\end{figure}\n\\end{center}\n";
+      $wroteline += 4;
+   } else {
+      print OUT $_;
+      ++$wroteline;
+   }
+}
+print "Read $readline lines, wrote $wroteline lines\n";
+
+close (OUT);
+close (IN);
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/changes.txt	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,304 @@
+April 11th, 2004
+v0.30  -- Added "mp_toradix_n" which stores upto "n-1" least significant digits of an mp_int
+       -- Johan Lindh sent a patch so MSVC wouldn't whine about redefining malloc [in weird dll modes]
+       -- Henrik Goldman spotted a missing OPT_CAST in mp_fwrite()
+       -- Tuned tommath.h so that when MP_LOW_MEM is defined MP_PREC shall be reduced.
+          [I also allow MP_PREC to be externally defined now]
+       -- Sped up mp_cnt_lsb() by using a 4x4 table [e.g. 4x speedup]
+       -- Added mp_prime_random_ex() which is a more versatile prime generator accurate to
+          exact bit lengths (unlike the deprecated but still available mp_prime_random() which
+          is only accurate to byte lengths).  See the new LTM_PRIME_* flags ;-)
+       -- Alex Polushin contributed an optimized mp_sqrt() as well as mp_get_int() and mp_is_square().
+          I've cleaned them all up to be a little more consistent [along with one bug fix] for this release.
+       -- Added mp_init_set and mp_init_set_int to initialize and set small constants with one function
+          call.
+       -- Removed /etclib directory [um LibTomPoly deprecates this].
+       -- Fixed mp_mod() so the sign of the result agrees with the sign of the modulus.
+       ++ N.B.  My semester is almost up so expect updates to the textbook to be posted to the libtomcrypt.org 
+          website.  
+
+Jan 25th, 2004
+v0.29  ++ Note: "Henrik" from the v0.28 changelog refers to Henrik Goldman ;-)
+       -- Added fix to mp_shrink to prevent a realloc when used == 0 [e.g. realloc zero bytes???]
+       -- Made the mp_prime_rabin_miller_trials() function internal table smaller and also
+          set the minimum number of tests to two (sounds a bit safer).
+       -- Added a mp_exteuclid() which computes the extended euclidean algorithm.
+       -- Fixed a memory leak in s_mp_exptmod() [called when Barrett reduction is to be used] which would arise
+          if a multiplication or subsequent reduction failed [would not free the temp result].
+       -- Made an API change to mp_radix_size().  It now returns an error code and stores the required size
+          through an "int star" passed to it.
+
+Dec 24th, 2003
+v0.28  -- Henrik Goldman suggested I add casts to the montomgery code [stores into mu...] so compilers wouldn't
+          spew [erroneous] diagnostics... fixed.
+       -- Henrik Goldman also spotted two typos.  One in mp_radix_size() and another in mp_toradix().
+       -- Added fix to mp_shrink() to avoid a memory leak.
+       -- Added mp_prime_random() which requires a callback to make truly random primes of a given nature
+          (idea from chat with Niels Ferguson at Crypto'03)
+       -- Picked up a second wind.  I'm filled with Gooo.  Mission Gooo!
+       -- Removed divisions from mp_reduce_is_2k()
+       -- Sped up mp_div_d() [general case] to use only one division per digit instead of two.
+       -- Added the heap macros from LTC to LTM.  Now you can easily [by editing four lines of tommath.h]
+          change the name of the heap functions used in LTM [also compatible with LTC via MPI mode]
+       -- Added bn_prime_rabin_miller_trials() which gives the number of Rabin-Miller trials to achieve
+          a failure rate of less than 2^-96
+       -- fixed bug in fast_mp_invmod().  The initial testing logic was wrong.  An invalid input is not when
+          "a" and "b" are even it's when "b" is even [the algo is for odd moduli only].
+       -- Started a new manual [finally].  It is incomplete and will be finished as time goes on.  I had to stop
+          adding full demos around half way in chapter three so I could at least get a good portion of the
+          manual done.   If you really need help using the library you can always email me!
+       -- My Textbook is now included as part of the package [all Public Domain]
+
+Sept 19th, 2003
+v0.27  -- Removed changes.txt~ which was made by accident since "kate" decided it was
+          a good time to re-enable backups... [kde is fun!]
+       -- In mp_grow() "a->dp" is not overwritten by realloc call [re: memory leak]
+          Now if mp_grow() fails the mp_int is still valid and can be cleared via
+          mp_clear() to reclaim the memory.
+       -- Henrik Goldman found a buffer overflow bug in mp_add_d().  Fixed.
+       -- Cleaned up mp_mul_d() to be much easier to read and follow.
+
+Aug 29th, 2003
+v0.26  -- Fixed typo that caused warning with GCC 3.2
+       -- Martin Marcel noticed a bug in mp_neg() that allowed negative zeroes.
+          Also, Martin is the fellow who noted the bugs in mp_gcd() of 0.24/0.25.
+       -- Martin Marcel noticed an optimization [and slight bug] in mp_lcm().
+       -- Added fix to mp_read_unsigned_bin to prevent a buffer overflow.
+       -- Beefed up the comments in the baseline multipliers [and montgomery]
+       -- Added "mont" demo to the makefile.msvc in etc/
+       -- Optimized sign compares in mp_cmp from 4 to 2 cases.
+
+Aug 4th, 2003
+v0.25  -- Fix to mp_gcd again... oops (0,-a) == (-a, 0) == a
+       -- Fix to mp_clear which didn't reset the sign  [Greg Rose]
+       -- Added mp_error_to_string() to convert return codes to strings.  [Greg Rose]
+       -- Optimized fast_mp_invmod() to do the test for invalid inputs [both even]
+          first so temps don't have to be initialized if it's going to fail.
+       -- Optimized mp_gcd() by removing mp_div_2d calls for when one of the inputs
+          is odd.
+       -- Tons of new comments, some indentation fixups, etc.
+       -- mp_jacobi() returns MP_VAL if the modulus is less than or equal to zero.
+       -- fixed two typos in the header of each file :-)
+       -- LibTomMath is officially Public Domain [see LICENSE]
+
+July 15th, 2003
+v0.24  -- Optimized mp_add_d and mp_sub_d to not allocate temporary variables
+       -- Fixed mp_gcd() so the gcd of 0,0 is 0.  Allows the gcd operation to be chained
+          e.g. (0,0,a) == a [instead of 1]
+       -- Should be one of the last release for a while.  Working on LibTomMath book now.
+       -- optimized the pprime demo [/etc/pprime.c] to first make a huge table of single
+          digit primes then it reads them randomly instead of randomly choosing/testing single
+          digit primes.
+
+July 12th, 2003
+v0.23  -- Optimized mp_prime_next_prime() to not use mp_mod [via is_divisible()] in each
+          iteration.  Instead now a smaller table is kept of the residues which can be updated
+          without division.
+       -- Fixed a bug in next_prime() where an input of zero would be treated as odd and
+          have two added to it [to move to the next odd].
+       -- fixed a bug in prime_fermat() and prime_miller_rabin() which allowed the base
+          to be negative, zero or one.  Normally the test is only valid if the base is
+          greater than one.
+       -- changed the next_prime() prototype to accept a new parameter "bbs_style" which
+          will find the next prime congruent to 3 mod 4.  The default [bbs_style==0] will
+          make primes which are either congruent to 1 or 3 mod 4.
+       -- fixed mp_read_unsigned_bin() so that it doesn't include both code for
+          the case DIGIT_BIT < 8 and >= 8
+       -- optimized div_d() to easy out on division by 1 [or if a == 0] and use
+          logical shifts if the divisor is a power of two.
+       -- the default DIGIT_BIT type was not int for non-default builds.  Fixed.
+
+July 2nd, 2003
+v0.22  -- Fixed up mp_invmod so the result is properly in range now [was always congruent to the inverse...]
+       -- Fixed up s_mp_exptmod and mp_exptmod_fast so the lower half of the pre-computed table isn't allocated
+          which makes the algorithm use half as much ram.
+       -- Fixed the install script not to make the book :-) [which isn't included anyways]
+       -- added mp_cnt_lsb() which counts how many of the lsbs are zero
+       -- optimized mp_gcd() to use the new mp_cnt_lsb() to replace multiple divisions by two by a single division.
+       -- applied similar optimization to mp_prime_miller_rabin().
+       -- Fixed a bug in both mp_invmod() and fast_mp_invmod() which tested for odd
+          via "mp_iseven() == 0" which is not valid [since zero is not even either].
+
+June 19th, 2003
+v0.21  -- Fixed bug in mp_mul_d which would not handle sign correctly [would not always forward it]
+       -- Removed the #line lines from gen.pl [was in violation of ISO C]
+
+June 8th, 2003
+v0.20  -- Removed the book from the package.  Added the TDCAL license document.
+       -- This release is officially pure-bred TDCAL again [last officially TDCAL based release was v0.16]
+
+June 6th, 2003
+v0.19  -- Fixed a bug in mp_montgomery_reduce() which was introduced when I tweaked mp_rshd() in the previous release.
+          Essentially the digits were not trimmed before the compare which cause a subtraction to occur all the time.
+       -- Fixed up etc/tune.c a bit to stop testing new cutoffs after 16 failures [to find more optimal points].
+          Brute force ho!
+
+
+May 29th, 2003
+v0.18  -- Fixed a bug in s_mp_sqr which would handle carries properly just not very elegantly.
+          (e.g. correct result, just bad looking code)
+       -- Fixed bug in mp_sqr which still had a 512 constant instead of MP_WARRAY
+       -- Added Toom-Cook multipliers [needs tuning!]
+       -- Added efficient divide by 3 algorithm mp_div_3
+       -- Re-wrote mp_div_d to be faster than calling mp_div
+       -- Added in a donated BCC makefile and a single page LTM poster ([email protected])
+       -- Added mp_reduce_2k which reduces an input modulo n = 2**p - k for any single digit k
+       -- Made the exptmod system be aware of the 2k reduction algorithms.
+       -- Rewrote mp_dr_reduce to be smaller, simpler and easier to understand.
+
+May 17th, 2003
+v0.17  -- Benjamin Goldberg submitted optimized mp_add and mp_sub routines.  A new gen.pl as well
+          as several smaller suggestions.  Thanks!
+       -- removed call to mp_cmp in inner loop of mp_div and put mp_cmp_mag in its place :-)
+       -- Fixed bug in mp_exptmod that would cause it to fail for odd moduli when DIGIT_BIT != 28
+       -- mp_exptmod now also returns errors if the modulus is negative and will handle negative exponents
+       -- mp_prime_is_prime will now return true if the input is one of the primes in the prime table
+       -- Damian M Gryski ([email protected]) found a index out of bounds error in the
+          mp_fast_s_mp_mul_high_digs function which didn't come up before.  (fixed)
+       -- Refactored the DR reduction code so there is only one function per file.
+       -- Fixed bug in the mp_mul() which would erroneously avoid the faster multiplier [comba] when it was
+          allowed.  The bug would not cause the incorrect value to be produced just less efficient (fixed)
+       -- Fixed similar bug in the Montgomery reduction code.
+       -- Added tons of (mp_digit) casts so the 7/15/28/31 bit digit code will work flawlessly out of the box.
+          Also added limited support for 64-bit machines with a 60-bit digit.  Both thanks to Tom Wu ([email protected])
+       -- Added new comments here and there, cleaned up some code [style stuff]
+       -- Fixed a lingering typo in mp_exptmod* that would set bitcnt to zero then one.  Very silly stuff :-)
+       -- Fixed up mp_exptmod_fast so it would set "redux" to the comba Montgomery reduction if allowed.  This
+          saves quite a few calls and if statements.
+       -- Added etc/mont.c a test of the Montgomery reduction [assuming all else works :-| ]
+       -- Fixed up etc/tune.c to use a wider test range [more appropriate] also added a x86 based addition which
+          uses RDTSC for high precision timing.
+       -- Updated demo/demo.c to remove MPI stuff [won't work anyways], made the tests run for 2 seconds each so its
+          not so insanely slow.  Also made the output space delimited [and fixed up various errors]
+       -- Added logs directory, logs/graph.dem which will use gnuplot to make a series of PNG files
+          that go with the pre-made index.html.  You have to build [via make timing] and run ltmtest first in the
+          root of the package.
+       -- Fixed a bug in mp_sub and mp_add where "-a - -a" or "-a + a" would produce -0 as the result [obviously invalid].
+       -- Fixed a bug in mp_rshd.  If the count == a.used it should zero/return [instead of shifting]
+       -- Fixed a "off-by-one" bug in mp_mul2d.  The initial size check on alloc would be off by one if the residue
+          shifting caused a carry.
+       -- Fixed a bug where s_mp_mul_digs() would not call the Comba based routine if allowed.  This made Barrett reduction
+          slower than it had to be.
+
+Mar 29th, 2003
+v0.16  -- Sped up mp_div by making normalization one shift call
+       -- Sped up mp_mul_2d/mp_div_2d by aliasing pointers :-)
+       -- Cleaned up mp_gcd to use the macros for odd/even detection
+       -- Added comments here and there, mostly there but occasionally here too.
+
+Mar 22nd, 2003
+v0.15  -- Added series of prime testing routines to lib
+       -- Fixed up etc/tune.c
+       -- Added DR reduction algorithm
+       -- Beefed up the manual more.
+       -- Fixed up demo/demo.c so it doesn't have so many warnings and it does the full series of
+          tests
+       -- Added "pre-gen" directory which will hold a "gen.pl"'ed copy of the entire lib [done at
+          zipup time so its always the latest]
+       -- Added conditional casts for C++ users [boo!]
+
+Mar 15th, 2003
+v0.14  -- Tons of manual updates
+       -- cleaned up the directory
+       -- added MSVC makefiles
+       -- source changes [that I don't recall]
+       -- Fixed up the lshd/rshd code to use pointer aliasing
+       -- Fixed up the mul_2d and div_2d to not call rshd/lshd unless needed
+       -- Fixed up etc/tune.c a tad
+       -- fixed up demo/demo.c to output comma-delimited results of timing
+          also fixed up timing demo to use a finer granularity for various functions
+       -- fixed up demo/demo.c testing to pause during testing so my Duron won't catch on fire
+          [stays around 31-35C during testing :-)]
+
+Feb 13th, 2003
+v0.13  -- tons of minor speed-ups in low level add, sub, mul_2 and div_2 which propagate
+          to other functions like mp_invmod, mp_div, etc...
+       -- Sped up mp_exptmod_fast by using new code to find R mod m [e.g. B^n mod m]
+       -- minor fixes
+
+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
+          file packagage any more.  Instead of just adding "bn.c" you have to add
+          libtommath.a
+       -- Renamed "bn.h" to "tommath.h"
+       -- Changes to the manual to reflect all of this
+       -- Used GNU Indent to clean up the source
+
+Jan 15th, 2003
+v0.11  -- More subtle fixes
+       -- Moved to gentoo linux [hurrah!] so made *nix specific fixes to the make process
+       -- Sped up the montgomery reduction code quite a bit
+       -- fixed up demo so when building timing for the x86 it assumes ELF format now
+
+Jan 9th, 2003
+v0.10  -- Pekka Riikonen suggested fixes to the radix conversion code.
+       -- Added baseline montgomery and comba montgomery reductions, sped up exptmods
+          [to a point, see bn.h for MONTGOMERY_EXPT_CUTOFF]
+
+Jan 6th, 2003
+v0.09  -- Updated the manual to reflect recent changes.  :-)
+       -- Added Jacobi function (mp_jacobi) to supplement the number theory side of the lib
+       -- Added a Mersenne prime finder demo in ./etc/mersenne.c
+
+Jan 2nd, 2003
+v0.08  -- Sped up the multipliers by moving the inner loop variables into a smaller scope
+       -- Corrected a bunch of small "warnings"
+       -- Added more comments
+       -- Made "mtest" be able to use /dev/random, /dev/urandom or stdin for RNG data
+       -- Corrected some bugs where error messages were potentially ignored
+       -- add etc/pprime.c program which makes numbers which are provably prime.
+
+Jan 1st, 2003
+v0.07  -- Removed alot of heap operations from core functions to speed them up
+       -- Added a root finding function [and mp_sqrt macro like from MPI]
+       -- Added more to manual
+
+Dec 31st, 2002
+v0.06  -- Sped up the s_mp_add, s_mp_sub which inturn sped up mp_invmod, mp_exptmod, etc...
+       -- Cleaned up the header a bit more
+
+Dec 30th, 2002
+v0.05  -- Builds with MSVC out of the box
+       -- Fixed a bug in mp_invmod w.r.t. even moduli
+       -- Made mp_toradix and mp_read_radix use char instead of unsigned char arrays
+       -- Fixed up exptmod to use fewer multiplications
+       -- Fixed up mp_init_size to use only one heap operation
+          -- Note there is a slight "off-by-one" bug in the library somewhere
+             without the padding (see the source for comment) the library
+             crashes in libtomcrypt.  Anyways a reasonable workaround is to pad the
+             numbers which will always correct it since as the numbers grow the padding
+             will still be beyond the end of the number
+       -- Added more to the manual
+
+Dec 29th, 2002
+v0.04  -- Fixed a memory leak in mp_to_unsigned_bin
+       -- optimized invmod code
+       -- Fixed bug in mp_div
+       -- use exchange instead of copy for results
+       -- added a bit more to the manual
+
+Dec 27th, 2002
+v0.03  -- Sped up s_mp_mul_high_digs by not computing the carries of the lower digits
+       -- Fixed a bug where mp_set_int wouldn't zero the value first and set the used member.
+       -- fixed a bug in s_mp_mul_high_digs where the limit placed on the result digits was not calculated properly
+       -- fixed bugs in add/sub/mul/sqr_mod functions where if the modulus and dest were the same it wouldn't work
+       -- fixed a bug in mp_mod and mp_mod_d concerning negative inputs
+       -- mp_mul_d didn't preserve sign
+       -- Many many many many fixes
+       -- Works in LibTomCrypt now :-)
+       -- Added iterations to the timing demos... more accurate.
+       -- Tom needs a job.
+
+Dec 26th, 2002
+v0.02  -- Fixed a few "slips" in the manual.  This is "LibTomMath" afterall :-)
+       -- Added mp_cmp_mag, mp_neg, mp_abs and mp_radix_size that were missing.
+       -- Sped up the fast [comba] multipliers more [yahoo!]
+
+Dec 25th,2002
+v0.01  -- Initial release.  Gimme a break.
+       -- Todo list,
+           add details to manual [e.g. algorithms]
+           more comments in code
+           example programs
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/demo/demo.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,785 @@
+#include <time.h>
+
+#define TESTING
+
+#ifdef IOWNANATHLON
+#include <unistd.h>
+#define SLEEP sleep(4)
+#else
+#define SLEEP
+#endif
+
+#include "tommath.h"
+
+#ifdef TIMER
+ulong64 _tt;
+
+#if defined(__i386__) || defined(_M_IX86) || defined(_M_AMD64)
+/* RDTSC from Scott Duplichan */
+static ulong64 TIMFUNC (void)
+   {
+   #if defined __GNUC__
+      #ifdef __i386__
+         ulong64 a;
+         __asm__ __volatile__ ("rdtsc ":"=A" (a));
+         return a;
+      #else /* gcc-IA64 version */
+         unsigned long result;
+         __asm__ __volatile__("mov %0=ar.itc" : "=r"(result) :: "memory");
+         while (__builtin_expect ((int) result == -1, 0))
+         __asm__ __volatile__("mov %0=ar.itc" : "=r"(result) :: "memory");
+         return result;
+      #endif
+
+   // Microsoft and Intel Windows compilers
+   #elif defined _M_IX86
+     __asm rdtsc
+   #elif defined _M_AMD64
+     return __rdtsc ();
+   #elif defined _M_IA64
+     #if defined __INTEL_COMPILER
+       #include <ia64intrin.h>
+     #endif
+      return __getReg (3116);
+   #else
+     #error need rdtsc function for this build
+   #endif
+   }
+#else
+#define TIMFUNC clock
+#endif
+
+ulong64 rdtsc(void) { return TIMFUNC() - _tt; }
+void reset(void) { _tt = TIMFUNC(); }
+
+#endif
+
+void ndraw(mp_int *a, char *name)
+{
+   char buf[4096];
+   printf("%s: ", name);
+   mp_toradix(a, buf, 64);
+   printf("%s\n", buf);
+}
+
+static void draw(mp_int *a)
+{
+   ndraw(a, "");
+}
+
+
+unsigned long lfsr = 0xAAAAAAAAUL;
+
+int lbit(void)
+{
+   if (lfsr & 0x80000000UL) {
+      lfsr = ((lfsr << 1) ^ 0x8000001BUL) & 0xFFFFFFFFUL;
+      return 1;
+   } else {
+      lfsr <<= 1;
+      return 0;
+   }
+}
+
+int myrng(unsigned char *dst, int len, void *dat)
+{
+   int x;
+   for (x = 0; x < len; x++) dst[x] = rand() & 0xFF;
+   return len;
+}
+
+
+#define DO2(x) x; x;
+#define DO4(x) DO2(x); DO2(x);
+#define DO8(x) DO4(x); DO4(x);
+#define DO(x)  DO8(x); DO8(x);
+
+   char cmd[4096], buf[4096];
+int main(void)
+{
+   mp_int a, b, c, d, e, f;
+   unsigned long expt_n, add_n, sub_n, mul_n, div_n, sqr_n, mul2d_n, div2d_n, gcd_n, lcm_n, inv_n,
+                 div2_n, mul2_n, add_d_n, sub_d_n, t;
+   unsigned rr;
+   int i, n, err, cnt, ix, old_kara_m, old_kara_s;
+
+#ifdef TIMER
+   ulong64 tt, CLK_PER_SEC;
+   FILE *log, *logb, *logc;
+#endif
+
+   mp_init(&a);
+   mp_init(&b);
+   mp_init(&c);
+   mp_init(&d);
+   mp_init(&e);
+   mp_init(&f);
+
+   srand(time(NULL));
+
+#ifdef TESTING
+  // test mp_get_int
+  printf("Testing: mp_get_int\n");
+  for(i=0;i<1000;++i) {
+    t = (unsigned long)rand()*rand()+1;
+    mp_set_int(&a,t);
+    if (t!=mp_get_int(&a)) { 
+      printf("mp_get_int() bad result!\n");
+      return 1;
+    }
+  }
+  mp_set_int(&a,0);
+  if (mp_get_int(&a)!=0)
+  { printf("mp_get_int() bad result!\n");
+    return 1;
+  }
+  mp_set_int(&a,0xffffffff);
+  if (mp_get_int(&a)!=0xffffffff)
+  { printf("mp_get_int() bad result!\n");
+    return 1;
+  }
+
+  // test mp_sqrt
+  printf("Testing: mp_sqrt\n");
+  for (i=0;i<10000;++i) { 
+    printf("%6d\r", i); fflush(stdout);
+    n = (rand()&15)+1;
+    mp_rand(&a,n);
+    if (mp_sqrt(&a,&b) != MP_OKAY)
+    { printf("mp_sqrt() error!\n");
+      return 1;
+    }
+    mp_n_root(&a,2,&a);
+    if (mp_cmp_mag(&b,&a) != MP_EQ)
+    { printf("mp_sqrt() bad result!\n");
+      return 1;
+    }
+  }
+
+  printf("\nTesting: mp_is_square\n");
+  for (i=0;i<100000;++i) {
+    printf("%6d\r", i); fflush(stdout);
+
+    /* test mp_is_square false negatives */
+    n = (rand()&7)+1;
+    mp_rand(&a,n);
+    mp_sqr(&a,&a);
+    if (mp_is_square(&a,&n)!=MP_OKAY) { 
+      printf("fn:mp_is_square() error!\n");
+      return 1;
+    }
+    if (n==0) { 
+      printf("fn:mp_is_square() bad result!\n");
+      return 1;
+    }
+
+    /* test for false positives */
+    mp_add_d(&a, 1, &a);
+    if (mp_is_square(&a,&n)!=MP_OKAY) { 
+      printf("fp:mp_is_square() error!\n");
+      return 1;
+    }
+    if (n==1) { 
+      printf("fp:mp_is_square() bad result!\n");
+      return 1;
+    }
+
+  }
+  printf("\n\n");
+#endif
+
+#ifdef TESTING 
+   /* test for size */
+   for (ix = 16; ix < 512; ix++) {
+       printf("Testing (not safe-prime): %9d bits    \r", ix); fflush(stdout);
+       err = mp_prime_random_ex(&a, 8, ix, (rand()&1)?LTM_PRIME_2MSB_OFF:LTM_PRIME_2MSB_ON, myrng, NULL);
+       if (err != MP_OKAY) {
+          printf("failed with err code %d\n", err);
+          return EXIT_FAILURE;
+       }
+       if (mp_count_bits(&a) != ix) {
+          printf("Prime is %d not %d bits!!!\n", mp_count_bits(&a), ix);
+          return EXIT_FAILURE;
+       }
+   }
+
+   for (ix = 16; ix < 512; ix++) {
+       printf("Testing (   safe-prime): %9d bits    \r", ix); fflush(stdout);
+       err = mp_prime_random_ex(&a, 8, ix, ((rand()&1)?LTM_PRIME_2MSB_OFF:LTM_PRIME_2MSB_ON)|LTM_PRIME_SAFE, myrng, NULL);
+       if (err != MP_OKAY) {
+          printf("failed with err code %d\n", err);
+          return EXIT_FAILURE;
+       }
+       if (mp_count_bits(&a) != ix) {
+          printf("Prime is %d not %d bits!!!\n", mp_count_bits(&a), ix);
+          return EXIT_FAILURE;
+       }
+       /* let's see if it's really a safe prime */
+       mp_sub_d(&a, 1, &a);
+       mp_div_2(&a, &a);
+       mp_prime_is_prime(&a, 8, &cnt);
+       if (cnt != MP_YES) {
+          printf("sub is not prime!\n");
+          return EXIT_FAILURE;
+       }
+   }
+
+   printf("\n\n");
+#endif
+
+#ifdef TESTING
+   mp_read_radix(&a, "123456", 10);
+   mp_toradix_n(&a, buf, 10, 3);
+   printf("a == %s\n", buf);
+   mp_toradix_n(&a, buf, 10, 4);
+   printf("a == %s\n", buf);
+   mp_toradix_n(&a, buf, 10, 30);
+   printf("a == %s\n", buf);
+#endif
+
+
+#if 0
+   for (;;) {
+      fgets(buf, sizeof(buf), stdin);
+      mp_read_radix(&a, buf, 10);
+      mp_prime_next_prime(&a, 5, 1);
+      mp_toradix(&a, buf, 10);
+      printf("%s, %lu\n", buf, a.dp[0] & 3);
+   }
+#endif
+
+#if 0
+{
+   mp_word aa, bb;
+
+   for (;;) {
+       aa = abs(rand()) & MP_MASK;
+       bb = abs(rand()) & MP_MASK;
+      if (MULT(aa,bb) != (aa*bb)) {
+             printf("%llu * %llu == %llu or %llu?\n", aa, bb, (ulong64)MULT(aa,bb), (ulong64)(aa*bb));
+             return 0;
+          }
+   }
+}
+#endif
+
+#ifdef TESTING
+   /* test mp_cnt_lsb */
+   printf("testing mp_cnt_lsb...\n");
+   mp_set(&a, 1);
+   for (ix = 0; ix < 1024; ix++) {
+       if (mp_cnt_lsb(&a) != ix) {
+          printf("Failed at %d, %d\n", ix, mp_cnt_lsb(&a));
+          return 0;
+       }
+       mp_mul_2(&a, &a);
+   }
+#endif
+
+/* test mp_reduce_2k */
+#ifdef TESTING
+   printf("Testing mp_reduce_2k...\n");
+   for (cnt = 3; cnt <= 384; ++cnt) {
+       mp_digit tmp;
+       mp_2expt(&a, cnt);
+       mp_sub_d(&a, 2, &a);  /* a = 2**cnt - 2 */
+
+
+       printf("\nTesting %4d bits", cnt);
+       printf("(%d)", mp_reduce_is_2k(&a));
+       mp_reduce_2k_setup(&a, &tmp);
+       printf("(%d)", tmp);
+       for (ix = 0; ix < 10000; ix++) {
+           if (!(ix & 127)) {printf("."); fflush(stdout); }
+           mp_rand(&b, (cnt/DIGIT_BIT  + 1) * 2);
+           mp_copy(&c, &b);
+           mp_mod(&c, &a, &c);
+           mp_reduce_2k(&b, &a, 1);
+           if (mp_cmp(&c, &b)) {
+              printf("FAILED\n");
+              exit(0);
+           }
+        }
+    }
+#endif
+
+
+/* test mp_div_3  */
+#ifdef TESTING
+   printf("Testing mp_div_3...\n");
+   mp_set(&d, 3);
+   for (cnt = 0; cnt < 1000000; ) {
+      mp_digit r1, r2;
+
+      if (!(++cnt & 127)) printf("%9d\r", cnt);
+      mp_rand(&a, abs(rand()) % 128 + 1);
+      mp_div(&a, &d, &b, &e);
+      mp_div_3(&a, &c, &r2);
+
+      if (mp_cmp(&b, &c) || mp_cmp_d(&e, r2)) {
+         printf("\n\nmp_div_3 => Failure\n");
+      }
+   }
+   printf("\n\nPassed div_3 testing\n");
+#endif
+
+/* test the DR reduction */
+#ifdef TESTING
+   printf("testing mp_dr_reduce...\n");
+   for (cnt = 2; cnt < 128; cnt++) {
+       printf("%d digit modulus\n", cnt);
+       mp_grow(&a, cnt);
+       mp_zero(&a);
+       for (ix = 1; ix < cnt; ix++) {
+           a.dp[ix] = MP_MASK;
+       }
+       a.used = cnt;
+       mp_prime_next_prime(&a, 3, 0);
+
+       mp_rand(&b, cnt - 1);
+       mp_copy(&b, &c);
+
+      rr = 0;
+      do {
+         if (!(rr & 127)) { printf("%9lu\r", rr); fflush(stdout); }
+         mp_sqr(&b, &b); mp_add_d(&b, 1, &b);
+         mp_copy(&b, &c);
+
+         mp_mod(&b, &a, &b);
+         mp_dr_reduce(&c, &a, (1<<DIGIT_BIT)-a.dp[0]);
+
+         if (mp_cmp(&b, &c) != MP_EQ) {
+            printf("Failed on trial %lu\n", rr); exit(-1);
+
+         }
+      } while (++rr < 100000);
+      printf("Passed DR test for %d digits\n", cnt);
+   }
+#endif
+
+#ifdef TIMER
+      /* temp. turn off TOOM */
+      TOOM_MUL_CUTOFF = TOOM_SQR_CUTOFF = 100000;
+
+      reset();
+      sleep(1);
+      CLK_PER_SEC = rdtsc();
+
+      printf("CLK_PER_SEC == %lu\n", CLK_PER_SEC);
+      
+
+      log = fopen("logs/add.log", "w");
+      for (cnt = 8; cnt <= 128; cnt += 8) {
+         SLEEP;
+         mp_rand(&a, cnt);
+         mp_rand(&b, cnt);
+         reset();
+         rr = 0;
+         do {
+            DO(mp_add(&a,&b,&c));
+            rr += 16;
+         } while (rdtsc() < (CLK_PER_SEC * 2));
+         tt = rdtsc();
+         printf("Adding\t\t%4d-bit => %9llu/sec, %9llu ticks\n", mp_count_bits(&a), (((ulong64)rr)*CLK_PER_SEC)/tt, tt);
+         fprintf(log, "%d %9llu\n", cnt*DIGIT_BIT, (((ulong64)rr)*CLK_PER_SEC)/tt); fflush(log);
+      }
+      fclose(log);
+
+      log = fopen("logs/sub.log", "w");
+      for (cnt = 8; cnt <= 128; cnt += 8) {
+         SLEEP;
+         mp_rand(&a, cnt);
+         mp_rand(&b, cnt);
+         reset();
+         rr = 0;
+         do {
+            DO(mp_sub(&a,&b,&c));
+            rr += 16;
+         } while (rdtsc() < (CLK_PER_SEC * 2));
+         tt = rdtsc();
+         printf("Subtracting\t\t%4d-bit => %9llu/sec, %9llu ticks\n", mp_count_bits(&a), (((ulong64)rr)*CLK_PER_SEC)/tt, tt);
+         fprintf(log, "%d %9llu\n", cnt*DIGIT_BIT, (((ulong64)rr)*CLK_PER_SEC)/tt);  fflush(log);
+      }
+      fclose(log);
+
+   /* do mult/square twice, first without karatsuba and second with */
+mult_test:   
+   old_kara_m = KARATSUBA_MUL_CUTOFF;
+   old_kara_s = KARATSUBA_SQR_CUTOFF;
+   for (ix = 0; ix < 2; ix++) {
+      printf("With%s Karatsuba\n", (ix==0)?"out":"");
+
+      KARATSUBA_MUL_CUTOFF = (ix==0)?9999:old_kara_m;
+      KARATSUBA_SQR_CUTOFF = (ix==0)?9999:old_kara_s;
+
+      log = fopen((ix==0)?"logs/mult.log":"logs/mult_kara.log", "w");
+      for (cnt = 32; cnt <= 288; cnt += 8) {
+         SLEEP;
+         mp_rand(&a, cnt);
+         mp_rand(&b, cnt);
+         reset();
+         rr = 0;
+         do {
+            DO(mp_mul(&a, &b, &c));
+            rr += 16;
+         } while (rdtsc() < (CLK_PER_SEC * 2));
+         tt = rdtsc();
+         printf("Multiplying\t%4d-bit => %9llu/sec, %9llu ticks\n", mp_count_bits(&a), (((ulong64)rr)*CLK_PER_SEC)/tt, tt);
+         fprintf(log, "%d %9llu\n", mp_count_bits(&a), (((ulong64)rr)*CLK_PER_SEC)/tt);  fflush(log);
+      }
+      fclose(log);
+
+      log = fopen((ix==0)?"logs/sqr.log":"logs/sqr_kara.log", "w");
+      for (cnt = 32; cnt <= 288; cnt += 8) {
+         SLEEP;
+         mp_rand(&a, cnt);
+         reset();
+         rr = 0;
+         do {
+            DO(mp_sqr(&a, &b));
+            rr += 16;
+         } while (rdtsc() < (CLK_PER_SEC * 2));
+         tt = rdtsc();
+         printf("Squaring\t%4d-bit => %9llu/sec, %9llu ticks\n", mp_count_bits(&a), (((ulong64)rr)*CLK_PER_SEC)/tt, tt);
+         fprintf(log, "%d %9llu\n", mp_count_bits(&a), (((ulong64)rr)*CLK_PER_SEC)/tt);  fflush(log);
+      }
+      fclose(log);
+
+   }
+expt_test:
+  {
+      char *primes[] = {
+         /* 2K moduli mersenne primes */
+         "6864797660130609714981900799081393217269435300143305409394463459185543183397656052122559640661454554977296311391480858037121987999716643812574028291115057151",
+         "531137992816767098689588206552468627329593117727031923199444138200403559860852242739162502265229285668889329486246501015346579337652707239409519978766587351943831270835393219031728127",
+         "10407932194664399081925240327364085538615262247266704805319112350403608059673360298012239441732324184842421613954281007791383566248323464908139906605677320762924129509389220345773183349661583550472959420547689811211693677147548478866962501384438260291732348885311160828538416585028255604666224831890918801847068222203140521026698435488732958028878050869736186900714720710555703168729087",
+         "1475979915214180235084898622737381736312066145333169775147771216478570297878078949377407337049389289382748507531496480477281264838760259191814463365330269540496961201113430156902396093989090226259326935025281409614983499388222831448598601834318536230923772641390209490231836446899608210795482963763094236630945410832793769905399982457186322944729636418890623372171723742105636440368218459649632948538696905872650486914434637457507280441823676813517852099348660847172579408422316678097670224011990280170474894487426924742108823536808485072502240519452587542875349976558572670229633962575212637477897785501552646522609988869914013540483809865681250419497686697771007",
+         "259117086013202627776246767922441530941818887553125427303974923161874019266586362086201209516800483406550695241733194177441689509238807017410377709597512042313066624082916353517952311186154862265604547691127595848775610568757931191017711408826252153849035830401185072116424747461823031471398340229288074545677907941037288235820705892351068433882986888616658650280927692080339605869308790500409503709875902119018371991620994002568935113136548829739112656797303241986517250116412703509705427773477972349821676443446668383119322540099648994051790241624056519054483690809616061625743042361721863339415852426431208737266591962061753535748892894599629195183082621860853400937932839420261866586142503251450773096274235376822938649407127700846077124211823080804139298087057504713825264571448379371125032081826126566649084251699453951887789613650248405739378594599444335231188280123660406262468609212150349937584782292237144339628858485938215738821232393687046160677362909315071",
+         "190797007524439073807468042969529173669356994749940177394741882673528979787005053706368049835514900244303495954950709725762186311224148828811920216904542206960744666169364221195289538436845390250168663932838805192055137154390912666527533007309292687539092257043362517857366624699975402375462954490293259233303137330643531556539739921926201438606439020075174723029056838272505051571967594608350063404495977660656269020823960825567012344189908927956646011998057988548630107637380993519826582389781888135705408653045219655801758081251164080554609057468028203308718724654081055323215860189611391296030471108443146745671967766308925858547271507311563765171008318248647110097614890313562856541784154881743146033909602737947385055355960331855614540900081456378659068370317267696980001187750995491090350108417050917991562167972281070161305972518044872048331306383715094854938415738549894606070722584737978176686422134354526989443028353644037187375385397838259511833166416134323695660367676897722287918773420968982326089026150031515424165462111337527431154890666327374921446276833564519776797633875503548665093914556482031482248883127023777039667707976559857333357013727342079099064400455741830654320379350833236245819348824064783585692924881021978332974949906122664421376034687815350484991",
+
+         /* DR moduli */
+         "14059105607947488696282932836518693308967803494693489478439861164411992439598399594747002144074658928593502845729752797260025831423419686528151609940203368612079",
+         "101745825697019260773923519755878567461315282017759829107608914364075275235254395622580447400994175578963163918967182013639660669771108475957692810857098847138903161308502419410142185759152435680068435915159402496058513611411688900243039",
+         "736335108039604595805923406147184530889923370574768772191969612422073040099331944991573923112581267542507986451953227192970402893063850485730703075899286013451337291468249027691733891486704001513279827771740183629161065194874727962517148100775228363421083691764065477590823919364012917984605619526140821797602431",
+         "38564998830736521417281865696453025806593491967131023221754800625044118265468851210705360385717536794615180260494208076605798671660719333199513807806252394423283413430106003596332513246682903994829528690198205120921557533726473585751382193953592127439965050261476810842071573684505878854588706623484573925925903505747545471088867712185004135201289273405614415899438276535626346098904241020877974002916168099951885406379295536200413493190419727789712076165162175783",
+         "542189391331696172661670440619180536749994166415993334151601745392193484590296600979602378676624808129613777993466242203025054573692562689251250471628358318743978285860720148446448885701001277560572526947619392551574490839286458454994488665744991822837769918095117129546414124448777033941223565831420390846864429504774477949153794689948747680362212954278693335653935890352619041936727463717926744868338358149568368643403037768649616778526013610493696186055899318268339432671541328195724261329606699831016666359440874843103020666106568222401047720269951530296879490444224546654729111504346660859907296364097126834834235287147",
+         "1487259134814709264092032648525971038895865645148901180585340454985524155135260217788758027400478312256339496385275012465661575576202252063145698732079880294664220579764848767704076761853197216563262660046602703973050798218246170835962005598561669706844469447435461092542265792444947706769615695252256130901271870341005768912974433684521436211263358097522726462083917939091760026658925757076733484173202927141441492573799914240222628795405623953109131594523623353044898339481494120112723445689647986475279242446083151413667587008191682564376412347964146113898565886683139407005941383669325997475076910488086663256335689181157957571445067490187939553165903773554290260531009121879044170766615232300936675369451260747671432073394867530820527479172464106442450727640226503746586340279816318821395210726268291535648506190714616083163403189943334431056876038286530365757187367147446004855912033137386225053275419626102417236133948503",
+         "1095121115716677802856811290392395128588168592409109494900178008967955253005183831872715423151551999734857184538199864469605657805519106717529655044054833197687459782636297255219742994736751541815269727940751860670268774903340296040006114013971309257028332849679096824800250742691718610670812374272414086863715763724622797509437062518082383056050144624962776302147890521249477060215148275163688301275847155316042279405557632639366066847442861422164832655874655824221577849928863023018366835675399949740429332468186340518172487073360822220449055340582568461568645259954873303616953776393853174845132081121976327462740354930744487429617202585015510744298530101547706821590188733515880733527449780963163909830077616357506845523215289297624086914545378511082534229620116563260168494523906566709418166011112754529766183554579321224940951177394088465596712620076240067370589036924024728375076210477267488679008016579588696191194060127319035195370137160936882402244399699172017835144537488486396906144217720028992863941288217185353914991583400421682751000603596655790990815525126154394344641336397793791497068253936771017031980867706707490224041075826337383538651825493679503771934836094655802776331664261631740148281763487765852746577808019633679",
+
+         /* generic unrestricted moduli */
+         "17933601194860113372237070562165128350027320072176844226673287945873370751245439587792371960615073855669274087805055507977323024886880985062002853331424203",
+         "2893527720709661239493896562339544088620375736490408468011883030469939904368086092336458298221245707898933583190713188177399401852627749210994595974791782790253946539043962213027074922559572312141181787434278708783207966459019479487",
+         "347743159439876626079252796797422223177535447388206607607181663903045907591201940478223621722118173270898487582987137708656414344685816179420855160986340457973820182883508387588163122354089264395604796675278966117567294812714812796820596564876450716066283126720010859041484786529056457896367683122960411136319",
+         "47266428956356393164697365098120418976400602706072312735924071745438532218237979333351774907308168340693326687317443721193266215155735814510792148768576498491199122744351399489453533553203833318691678263241941706256996197460424029012419012634671862283532342656309677173602509498417976091509154360039893165037637034737020327399910409885798185771003505320583967737293415979917317338985837385734747478364242020380416892056650841470869294527543597349250299539682430605173321029026555546832473048600327036845781970289288898317888427517364945316709081173840186150794397479045034008257793436817683392375274635794835245695887",
+         "436463808505957768574894870394349739623346440601945961161254440072143298152040105676491048248110146278752857839930515766167441407021501229924721335644557342265864606569000117714935185566842453630868849121480179691838399545644365571106757731317371758557990781880691336695584799313313687287468894148823761785582982549586183756806449017542622267874275103877481475534991201849912222670102069951687572917937634467778042874315463238062009202992087620963771759666448266532858079402669920025224220613419441069718482837399612644978839925207109870840278194042158748845445131729137117098529028886770063736487420613144045836803985635654192482395882603511950547826439092832800532152534003936926017612446606135655146445620623395788978726744728503058670046885876251527122350275750995227",
+         "11424167473351836398078306042624362277956429440521137061889702611766348760692206243140413411077394583180726863277012016602279290144126785129569474909173584789822341986742719230331946072730319555984484911716797058875905400999504305877245849119687509023232790273637466821052576859232452982061831009770786031785669030271542286603956118755585683996118896215213488875253101894663403069677745948305893849505434201763745232895780711972432011344857521691017896316861403206449421332243658855453435784006517202894181640562433575390821384210960117518650374602256601091379644034244332285065935413233557998331562749140202965844219336298970011513882564935538704289446968322281451907487362046511461221329799897350993370560697505809686438782036235372137015731304779072430260986460269894522159103008260495503005267165927542949439526272736586626709581721032189532726389643625590680105784844246152702670169304203783072275089194754889511973916207",
+         "1214855636816562637502584060163403830270705000634713483015101384881871978446801224798536155406895823305035467591632531067547890948695117172076954220727075688048751022421198712032848890056357845974246560748347918630050853933697792254955890439720297560693579400297062396904306270145886830719309296352765295712183040773146419022875165382778007040109957609739589875590885701126197906063620133954893216612678838507540777138437797705602453719559017633986486649523611975865005712371194067612263330335590526176087004421363598470302731349138773205901447704682181517904064735636518462452242791676541725292378925568296858010151852326316777511935037531017413910506921922450666933202278489024521263798482237150056835746454842662048692127173834433089016107854491097456725016327709663199738238442164843147132789153725513257167915555162094970853584447993125488607696008169807374736711297007473812256272245489405898470297178738029484459690836250560495461579533254473316340608217876781986188705928270735695752830825527963838355419762516246028680280988020401914551825487349990306976304093109384451438813251211051597392127491464898797406789175453067960072008590614886532333015881171367104445044718144312416815712216611576221546455968770801413440778423979",
+         NULL
+      };
+   log = fopen("logs/expt.log", "w");
+   logb = fopen("logs/expt_dr.log", "w");
+   logc = fopen("logs/expt_2k.log", "w");
+   for (n = 0; primes[n]; n++) {
+      SLEEP;
+      mp_read_radix(&a, primes[n], 10);
+      mp_zero(&b);
+      for (rr = 0; rr < mp_count_bits(&a); rr++) {
+         mp_mul_2(&b, &b);
+         b.dp[0] |= lbit();
+         b.used  += 1;
+      }
+      mp_sub_d(&a, 1, &c);
+      mp_mod(&b, &c, &b);
+      mp_set(&c, 3);
+      reset();
+      rr = 0;
+      do {
+         DO(mp_exptmod(&c, &b, &a, &d));
+         rr += 16;
+      } while (rdtsc() < (CLK_PER_SEC * 2));
+      tt = rdtsc();
+      mp_sub_d(&a, 1, &e);
+      mp_sub(&e, &b, &b);
+      mp_exptmod(&c, &b, &a, &e);  /* c^(p-1-b) mod a */
+      mp_mulmod(&e, &d, &a, &d);   /* c^b * c^(p-1-b) == c^p-1 == 1 */
+      if (mp_cmp_d(&d, 1)) {
+         printf("Different (%d)!!!\n", mp_count_bits(&a));
+         draw(&d);
+         exit(0);
+      }
+      printf("Exponentiating\t%4d-bit => %9llu/sec, %9llu ticks\n", mp_count_bits(&a), (((ulong64)rr)*CLK_PER_SEC)/tt, tt);
+      fprintf((n < 6) ? logc : (n < 13) ? logb : log, "%d %9llu\n", mp_count_bits(&a), (((ulong64)rr)*CLK_PER_SEC)/tt);
+   }
+   }
+   fclose(log);
+   fclose(logb);
+   fclose(logc);
+
+   log = fopen("logs/invmod.log", "w");
+   for (cnt = 4; cnt <= 128; cnt += 4) {
+      SLEEP;
+      mp_rand(&a, cnt);
+      mp_rand(&b, cnt);
+
+      do {
+         mp_add_d(&b, 1, &b);
+         mp_gcd(&a, &b, &c);
+      } while (mp_cmp_d(&c, 1) != MP_EQ);
+
+      reset();
+      rr = 0;
+      do {
+         DO(mp_invmod(&b, &a, &c));
+         rr += 16;
+      } while (rdtsc() < (CLK_PER_SEC * 2));
+      tt = rdtsc();
+      mp_mulmod(&b, &c, &a, &d);
+      if (mp_cmp_d(&d, 1) != MP_EQ) {
+         printf("Failed to invert\n");
+         return 0;
+      }
+      printf("Inverting mod\t%4d-bit => %9llu/sec, %9llu ticks\n", mp_count_bits(&a), (((ulong64)rr)*CLK_PER_SEC)/tt, tt);
+      fprintf(log, "%d %9llu\n", cnt*DIGIT_BIT, (((ulong64)rr)*CLK_PER_SEC)/tt);
+   }
+   fclose(log);
+
+   return 0;
+
+#endif
+
+   div2_n = mul2_n = inv_n = expt_n = lcm_n = gcd_n = add_n =
+   sub_n = mul_n = div_n = sqr_n = mul2d_n = div2d_n = cnt = add_d_n = sub_d_n= 0;
+
+   /* force KARA and TOOM to enable despite cutoffs */
+   KARATSUBA_SQR_CUTOFF = KARATSUBA_MUL_CUTOFF = 110;
+   TOOM_SQR_CUTOFF      = TOOM_MUL_CUTOFF      = 150;
+
+   for (;;) {
+       /* randomly clear and re-init one variable, this has the affect of triming the alloc space */
+       switch (abs(rand()) % 7) {
+           case 0:  mp_clear(&a); mp_init(&a); break;
+           case 1:  mp_clear(&b); mp_init(&b); break;
+           case 2:  mp_clear(&c); mp_init(&c); break;
+           case 3:  mp_clear(&d); mp_init(&d); break;
+           case 4:  mp_clear(&e); mp_init(&e); break;
+           case 5:  mp_clear(&f); mp_init(&f); break;
+           case 6:  break; /* don't clear any */
+       }
+
+
+       printf("%4lu/%4lu/%4lu/%4lu/%4lu/%4lu/%4lu/%4lu/%4lu/%4lu/%4lu/%4lu/%4lu/%4lu/%4lu ", add_n, sub_n, mul_n, div_n, sqr_n, mul2d_n, div2d_n, gcd_n, lcm_n, expt_n, inv_n, div2_n, mul2_n, add_d_n, sub_d_n);
+       fgets(cmd, 4095, stdin);
+       cmd[strlen(cmd)-1] = 0;
+       printf("%s  ]\r",cmd); fflush(stdout);
+       if (!strcmp(cmd, "mul2d")) { ++mul2d_n;
+          fgets(buf, 4095, stdin); mp_read_radix(&a, buf, 64);
+          fgets(buf, 4095, stdin); sscanf(buf, "%d", &rr);
+          fgets(buf, 4095, stdin); mp_read_radix(&b, buf, 64);
+
+          mp_mul_2d(&a, rr, &a);
+          a.sign = b.sign;
+          if (mp_cmp(&a, &b) != MP_EQ) {
+             printf("mul2d failed, rr == %d\n",rr);
+             draw(&a);
+             draw(&b);
+             return 0;
+          }
+       } else if (!strcmp(cmd, "div2d")) { ++div2d_n;
+          fgets(buf, 4095, stdin); mp_read_radix(&a, buf, 64);
+          fgets(buf, 4095, stdin); sscanf(buf, "%d", &rr);
+          fgets(buf, 4095, stdin); mp_read_radix(&b, buf, 64);
+
+          mp_div_2d(&a, rr, &a, &e);
+          a.sign = b.sign;
+          if (a.used == b.used && a.used == 0) { a.sign = b.sign = MP_ZPOS; }
+          if (mp_cmp(&a, &b) != MP_EQ) {
+             printf("div2d failed, rr == %d\n",rr);
+             draw(&a);
+             draw(&b);
+             return 0;
+          }
+       } else if (!strcmp(cmd, "add")) { ++add_n;
+          fgets(buf, 4095, stdin);  mp_read_radix(&a, buf, 64);
+          fgets(buf, 4095, stdin);  mp_read_radix(&b, buf, 64);
+          fgets(buf, 4095, stdin);  mp_read_radix(&c, buf, 64);
+          mp_copy(&a, &d);
+          mp_add(&d, &b, &d);
+          if (mp_cmp(&c, &d) != MP_EQ) {
+             printf("add %lu failure!\n", add_n);
+draw(&a);draw(&b);draw(&c);draw(&d);
+             return 0;
+          }
+
+          /* test the sign/unsigned storage functions */
+
+          rr = mp_signed_bin_size(&c);
+          mp_to_signed_bin(&c, (unsigned char *)cmd);
+          memset(cmd+rr, rand()&255, sizeof(cmd)-rr);
+          mp_read_signed_bin(&d, (unsigned char *)cmd, rr);
+          if (mp_cmp(&c, &d) != MP_EQ) {
+             printf("mp_signed_bin failure!\n");
+             draw(&c);
+             draw(&d);
+             return 0;
+          }
+
+
+          rr = mp_unsigned_bin_size(&c);
+          mp_to_unsigned_bin(&c, (unsigned char *)cmd);
+          memset(cmd+rr, rand()&255, sizeof(cmd)-rr);
+          mp_read_unsigned_bin(&d, (unsigned char *)cmd, rr);
+          if (mp_cmp_mag(&c, &d) != MP_EQ) {
+             printf("mp_unsigned_bin failure!\n");
+             draw(&c);
+             draw(&d);
+             return 0;
+          }
+
+       } else if (!strcmp(cmd, "sub")) { ++sub_n;
+          fgets(buf, 4095, stdin);  mp_read_radix(&a, buf, 64);
+          fgets(buf, 4095, stdin);  mp_read_radix(&b, buf, 64);
+          fgets(buf, 4095, stdin);  mp_read_radix(&c, buf, 64);
+          mp_copy(&a, &d);
+          mp_sub(&d, &b, &d);
+          if (mp_cmp(&c, &d) != MP_EQ) {
+             printf("sub %lu failure!\n", sub_n);
+draw(&a);draw(&b);draw(&c);draw(&d);
+             return 0;
+          }
+       } else if (!strcmp(cmd, "mul")) { ++mul_n;
+          fgets(buf, 4095, stdin);  mp_read_radix(&a, buf, 64);
+          fgets(buf, 4095, stdin);  mp_read_radix(&b, buf, 64);
+          fgets(buf, 4095, stdin);  mp_read_radix(&c, buf, 64);
+          mp_copy(&a, &d);
+          mp_mul(&d, &b, &d);
+          if (mp_cmp(&c, &d) != MP_EQ) {
+             printf("mul %lu failure!\n", mul_n);
+draw(&a);draw(&b);draw(&c);draw(&d);
+             return 0;
+          }
+       } else if (!strcmp(cmd, "div")) { ++div_n;
+          fgets(buf, 4095, stdin); mp_read_radix(&a, buf, 64);
+          fgets(buf, 4095, stdin); mp_read_radix(&b, buf, 64);
+          fgets(buf, 4095, stdin); mp_read_radix(&c, buf, 64);
+          fgets(buf, 4095, stdin); mp_read_radix(&d, buf, 64);
+
+          mp_div(&a, &b, &e, &f);
+          if (mp_cmp(&c, &e) != MP_EQ || mp_cmp(&d, &f) != MP_EQ) {
+             printf("div %lu failure!\n", div_n);
+draw(&a);draw(&b);draw(&c);draw(&d); draw(&e); draw(&f);
+             return 0;
+          }
+
+       } else if (!strcmp(cmd, "sqr")) { ++sqr_n;
+          fgets(buf, 4095, stdin);  mp_read_radix(&a, buf, 64);
+          fgets(buf, 4095, stdin);  mp_read_radix(&b, buf, 64);
+          mp_copy(&a, &c);
+          mp_sqr(&c, &c);
+          if (mp_cmp(&b, &c) != MP_EQ) {
+             printf("sqr %lu failure!\n", sqr_n);
+draw(&a);draw(&b);draw(&c);
+             return 0;
+          }
+       } else if (!strcmp(cmd, "gcd")) { ++gcd_n;
+          fgets(buf, 4095, stdin);  mp_read_radix(&a, buf, 64);
+          fgets(buf, 4095, stdin);  mp_read_radix(&b, buf, 64);
+          fgets(buf, 4095, stdin);  mp_read_radix(&c, buf, 64);
+          mp_copy(&a, &d);
+          mp_gcd(&d, &b, &d);
+          d.sign = c.sign;
+          if (mp_cmp(&c, &d) != MP_EQ) {
+             printf("gcd %lu failure!\n", gcd_n);
+draw(&a);draw(&b);draw(&c);draw(&d);
+             return 0;
+          }
+       } else if (!strcmp(cmd, "lcm")) { ++lcm_n;
+             fgets(buf, 4095, stdin);  mp_read_radix(&a, buf, 64);
+             fgets(buf, 4095, stdin);  mp_read_radix(&b, buf, 64);
+             fgets(buf, 4095, stdin);  mp_read_radix(&c, buf, 64);
+             mp_copy(&a, &d);
+             mp_lcm(&d, &b, &d);
+             d.sign = c.sign;
+             if (mp_cmp(&c, &d) != MP_EQ) {
+                printf("lcm %lu failure!\n", lcm_n);
+   draw(&a);draw(&b);draw(&c);draw(&d);
+                return 0;
+             }
+       } else if (!strcmp(cmd, "expt")) {  ++expt_n;
+             fgets(buf, 4095, stdin);  mp_read_radix(&a, buf, 64);
+             fgets(buf, 4095, stdin);  mp_read_radix(&b, buf, 64);
+             fgets(buf, 4095, stdin);  mp_read_radix(&c, buf, 64);
+             fgets(buf, 4095, stdin);  mp_read_radix(&d, buf, 64);
+             mp_copy(&a, &e);
+             mp_exptmod(&e, &b, &c, &e);
+             if (mp_cmp(&d, &e) != MP_EQ) {
+                printf("expt %lu failure!\n", expt_n);
+   draw(&a);draw(&b);draw(&c);draw(&d); draw(&e);
+                return 0;
+             }
+       } else if (!strcmp(cmd, "invmod")) {  ++inv_n;
+             fgets(buf, 4095, stdin);  mp_read_radix(&a, buf, 64);
+             fgets(buf, 4095, stdin);  mp_read_radix(&b, buf, 64);
+             fgets(buf, 4095, stdin);  mp_read_radix(&c, buf, 64);
+             mp_invmod(&a, &b, &d);
+             mp_mulmod(&d,&a,&b,&e);
+             if (mp_cmp_d(&e, 1) != MP_EQ) {
+                printf("inv [wrong value from MPI?!] failure\n");
+                draw(&a);draw(&b);draw(&c);draw(&d);
+                mp_gcd(&a, &b, &e);
+                draw(&e);
+                return 0;
+             }
+
+       } else if (!strcmp(cmd, "div2")) { ++div2_n;
+             fgets(buf, 4095, stdin);  mp_read_radix(&a, buf, 64);
+             fgets(buf, 4095, stdin);  mp_read_radix(&b, buf, 64);
+             mp_div_2(&a, &c);
+             if (mp_cmp(&c, &b) != MP_EQ) {
+                 printf("div_2 %lu failure\n", div2_n);
+                 draw(&a);
+                 draw(&b);
+                 draw(&c);
+                 return 0;
+             }
+       } else if (!strcmp(cmd, "mul2")) { ++mul2_n;
+             fgets(buf, 4095, stdin);  mp_read_radix(&a, buf, 64);
+             fgets(buf, 4095, stdin);  mp_read_radix(&b, buf, 64);
+             mp_mul_2(&a, &c);
+             if (mp_cmp(&c, &b) != MP_EQ) {
+                 printf("mul_2 %lu failure\n", mul2_n);
+                 draw(&a);
+                 draw(&b);
+                 draw(&c);
+                 return 0;
+             }
+       } else if (!strcmp(cmd, "add_d")) { ++add_d_n;
+              fgets(buf, 4095, stdin); mp_read_radix(&a, buf, 64);
+              fgets(buf, 4095, stdin); sscanf(buf, "%d", &ix);
+              fgets(buf, 4095, stdin); mp_read_radix(&b, buf, 64);
+              mp_add_d(&a, ix, &c);
+              if (mp_cmp(&b, &c) != MP_EQ) {
+                 printf("add_d %lu failure\n", add_d_n);
+                 draw(&a);
+                 draw(&b);
+                 draw(&c);
+                 printf("d == %d\n", ix);
+                 return 0;
+              }
+       } else if (!strcmp(cmd, "sub_d")) { ++sub_d_n;
+              fgets(buf, 4095, stdin); mp_read_radix(&a, buf, 64);
+              fgets(buf, 4095, stdin); sscanf(buf, "%d", &ix);
+              fgets(buf, 4095, stdin); mp_read_radix(&b, buf, 64);
+              mp_sub_d(&a, ix, &c);
+              if (mp_cmp(&b, &c) != MP_EQ) {
+                 printf("sub_d %lu failure\n", sub_d_n);
+                 draw(&a);
+                 draw(&b);
+                 draw(&c);
+                 printf("d == %d\n", ix);
+                 return 0;
+              }
+       }
+   }
+   return 0;
+}
+
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/etc/2kprime.1	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,2 @@
+256-bits (k = 36113) = 115792089237316195423570985008687907853269984665640564039457584007913129603823
+512-bits (k = 38117) = 13407807929942597099574024998205846127479365820592393377723561443721764030073546976801874298166903427690031858186486050853753882811946569946433649006045979
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/etc/2kprime.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,80 @@
+/* Makes safe primes of a 2k nature */
+#include <tommath.h>
+#include <time.h>
+
+int sizes[] = {256, 512, 768, 1024, 1536, 2048, 3072, 4096};
+
+int main(void)
+{
+   char buf[2000];
+   int x, y;
+   mp_int q, p;
+   FILE *out;
+   clock_t t1;
+   mp_digit z;
+   
+   mp_init_multi(&q, &p, NULL);
+   
+   out = fopen("2kprime.1", "w");
+   for (x = 0; x < (int)(sizeof(sizes) / sizeof(sizes[0])); x++) {
+   top:
+       mp_2expt(&q, sizes[x]);
+       mp_add_d(&q, 3, &q);
+       z = -3;
+       
+       t1 = clock();
+       for(;;) {
+         mp_sub_d(&q, 4, &q);
+         z += 4;
+
+         if (z > MP_MASK) {
+            printf("No primes of size %d found\n", sizes[x]);
+            break;
+         }
+         
+         if (clock() - t1 > CLOCKS_PER_SEC) { 
+            printf("."); fflush(stdout);
+//            sleep((clock() - t1 + CLOCKS_PER_SEC/2)/CLOCKS_PER_SEC);
+            t1 = clock();
+         }
+         
+         /* quick test on q */
+         mp_prime_is_prime(&q, 1, &y);
+         if (y == 0) {
+            continue;
+         }
+
+         /* find (q-1)/2 */
+         mp_sub_d(&q, 1, &p);
+         mp_div_2(&p, &p);
+         mp_prime_is_prime(&p, 3, &y);
+         if (y == 0) {
+            continue;
+         }
+
+         /* test on q */
+         mp_prime_is_prime(&q, 3, &y);
+         if (y == 0) {
+            continue;
+         }
+
+         break;
+       }
+       
+       if (y == 0) {
+          ++sizes[x];
+          goto top;
+       }
+       
+       mp_toradix(&q, buf, 10);
+       printf("\n\n%d-bits (k = %lu) = %s\n", sizes[x], z, buf);
+       fprintf(out, "%d-bits (k = %lu) = %s\n", sizes[x], z, buf); fflush(out);
+   }
+   
+   return 0;
+}   
+       
+         
+            
+            
+          
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/etc/drprime.c	Mon May 31 18:25:22 2004 +0000
@@ -0,0 +1,60 @@
+/* Makes safe primes of a DR nature */
+#include <tommath.h>
+
+int sizes[] = { 1+256/DIGIT_BIT, 1+512/DIGIT_BIT, 1+768/DIGIT_BIT, 1+1024/DIGIT_BIT, 1+2048/DIGIT_BIT, 1+4096/DIGIT_BIT };
+int main(void)
+{
+   int res, x, y;
+   char buf[4096];
+   FILE *out;
+   mp_int a, b;
+   
+   mp_init(&a);
+   mp_init(&b);
+   
+   out = fopen("drprimes.txt", "w");
+   for (x = 0; x < (int)(sizeof(sizes)/sizeof(sizes[0])); x++) {
+   top:
+       printf("Seeking a %d-bit safe prime\n", sizes[x] * DIGIT_BIT);
+       mp_grow(&a, sizes[x]);
+       mp_zero(&a);
+       for (y = 1; y < sizes[x]; y++) {
+           a.dp[y] = MP_MASK;
+       }
+       
+       /* make a DR modulus */
+       a.dp[0] = -1;
+       a.used = sizes[x];
+       
+       /* now loop */
+       res = 0;
+       for (;;) { 
+          a.dp[0] += 4;
+          if (a.dp[0] >= MP_MASK) break;
+          mp_prime_is_prime(&a, 1, &res);
+          if (res == 0) continue;
+          printf("."); fflush(stdout);
+          mp_sub_d(&a, 1, &b);
+          mp_div_2(&b, &b);