stormlib 0.1.0 → 1.0.1

data/ext/stormlib/StormLib/src/libtommath/bn_fast_mp_montgomery_reduce.c

@@ -0,0 +1,172 @@
+#include "tommath.h"
+#ifdef BN_FAST_MP_MONTGOMERY_REDUCE_C
+/* 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, tomstdenis@gmail.com, http://libtom.org
+ */
+
+/* computes xR**-1 == x (mod N) via Montgomery Reduction
+ *
+ * This is an optimized implementation of 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;
+}
+#endif
+
+/* $Source: /cvs/libtom/libtommath/bn_fast_mp_montgomery_reduce.c,v $ */
+/* $Revision: 1.4 $ */
+/* $Date: 2006/12/28 01:25:13 $ */

data/ext/stormlib/StormLib/src/libtommath/bn_fast_s_mp_mul_digs.c

@@ -0,0 +1,107 @@
+#include "tommath.h"
+#ifdef BN_FAST_S_MP_MUL_DIGS_C
+/* 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, tomstdenis@gmail.com, http://libtom.org
+ */
+
+/* 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, iz;
+  mp_digit W[MP_WARRAY];
+  register mp_word  _W;
+
+  /* grow the destination as required */
+  if (c->alloc < digs) {
+    if ((res = mp_grow (c, digs)) != MP_OKAY) {
+      return res;
+    }
+  }
+
+  /* number of output digits to produce */
+  pa = MIN(digs, a->used + b->used);
+
+  /* clear the carry */
+  _W = 0;
+  for (ix = 0; ix < pa; ix++) { 
+      int      tx, ty;
+      int      iy;
+      mp_digit *tmpx, *tmpy;
+
+      /* get offsets into the two bignums */
+      ty = MIN(b->used-1, ix);
+      tx = ix - ty;
+
+      /* setup temp aliases */
+      tmpx = a->dp + tx;
+      tmpy = b->dp + ty;
+
+      /* this is the number of times the loop will iterrate, essentially 
+         while (tx++ < a->used && ty-- >= 0) { ... }
+       */
+      iy = MIN(a->used-tx, ty+1);
+
+      /* execute loop */
+      for (iz = 0; iz < iy; ++iz) {
+         _W += ((mp_word)*tmpx++)*((mp_word)*tmpy--);
+
+      }
+
+      /* store term */
+      W[ix] = ((mp_digit)_W) & MP_MASK;
+
+      /* make next carry */
+      _W = _W >> ((mp_word)DIGIT_BIT);
+ }
+
+  /* setup dest */
+  olduse  = c->used;
+  c->used = pa;
+
+  {
+    register mp_digit *tmpc;
+    tmpc = c->dp;
+    for (ix = 0; ix < pa+1; ix++) {
+      /* now extract the previous digit [below the carry] */
+      *tmpc++ = W[ix];
+    }
+
+    /* clear unused digits [that existed in the old copy of c] */
+    for (; ix < olduse; ix++) {
+      *tmpc++ = 0;
+    }
+  }
+  mp_clamp (c);
+  return MP_OKAY;
+}
+#endif
+
+/* $Source: /cvs/libtom/libtommath/bn_fast_s_mp_mul_digs.c,v $ */
+/* $Revision: 1.8 $ */
+/* $Date: 2006/12/28 01:25:13 $ */

data/ext/stormlib/StormLib/src/libtommath/bn_fast_s_mp_mul_high_digs.c

@@ -0,0 +1,98 @@
+#include "tommath.h"
+#ifdef BN_FAST_S_MP_MUL_HIGH_DIGS_C
+/* 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, tomstdenis@gmail.com, http://libtom.org
+ */
+
+/* this is a modified version of fast_s_mul_digs that only produces
+ * output digits *above* digs.  See the comments for fast_s_mul_digs
+ * 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     olduse, res, pa, ix, iz;
+  mp_digit W[MP_WARRAY];
+  mp_word  _W;
+
+  /* grow the destination as required */
+  pa = a->used + b->used;
+  if (c->alloc < pa) {
+    if ((res = mp_grow (c, pa)) != MP_OKAY) {
+      return res;
+    }
+  }
+
+  /* number of output digits to produce */
+  pa = a->used + b->used;
+  _W = 0;
+  for (ix = digs; ix < pa; ix++) { 
+      int      tx, ty, iy;
+      mp_digit *tmpx, *tmpy;
+
+      /* get offsets into the two bignums */
+      ty = MIN(b->used-1, ix);
+      tx = ix - ty;
+
+      /* setup temp aliases */
+      tmpx = a->dp + tx;
+      tmpy = b->dp + ty;
+
+      /* this is the number of times the loop will iterrate, essentially its 
+         while (tx++ < a->used && ty-- >= 0) { ... }
+       */
+      iy = MIN(a->used-tx, ty+1);
+
+      /* execute loop */
+      for (iz = 0; iz < iy; iz++) {
+         _W += ((mp_word)*tmpx++)*((mp_word)*tmpy--);
+      }
+
+      /* store term */
+      W[ix] = ((mp_digit)_W) & MP_MASK;
+
+      /* make next carry */
+      _W = _W >> ((mp_word)DIGIT_BIT);
+  }
+  
+  /* setup dest */
+  olduse  = c->used;
+  c->used = pa;
+
+  {
+    register mp_digit *tmpc;
+
+    tmpc = c->dp + digs;
+    for (ix = digs; ix < pa; ix++) {
+      /* now extract the previous digit [below the carry] */
+      *tmpc++ = W[ix];
+    }
+
+    /* clear unused digits [that existed in the old copy of c] */
+    for (; ix < olduse; ix++) {
+      *tmpc++ = 0;
+    }
+  }
+  mp_clamp (c);
+  return MP_OKAY;
+}
+#endif
+
+/* $Source: /cvs/libtom/libtommath/bn_fast_s_mp_mul_high_digs.c,v $ */
+/* $Revision: 1.6 $ */
+/* $Date: 2006/12/28 01:25:13 $ */

data/ext/stormlib/StormLib/src/libtommath/bn_fast_s_mp_sqr.c

@@ -0,0 +1,114 @@
+#include "tommath.h"
+#ifdef BN_FAST_S_MP_SQR_C
+/* 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, tomstdenis@gmail.com, http://libtom.org
+ */
+
+/* the jist of squaring...
+ * you do like mult except the offset of the tmpx [one that 
+ * starts closer to zero] can't equal the offset of tmpy.  
+ * So basically you set up iy like before then you min it with
+ * (ty-tx) so that it never happens.  You double all those 
+ * you add in the inner loop
+
+After that loop you do the squares and add them in.
+*/
+
+int fast_s_mp_sqr (mp_int * a, mp_int * b)
+{
+  int       olduse, res, pa, ix, iz;
+  mp_digit   W[MP_WARRAY], *tmpx;
+  mp_word   W1;
+
+  /* grow the destination as required */
+  pa = a->used + a->used;
+  if (b->alloc < pa) {
+    if ((res = mp_grow (b, pa)) != MP_OKAY) {
+      return res;
+    }
+  }
+
+  /* number of output digits to produce */
+  W1 = 0;
+  for (ix = 0; ix < pa; ix++) { 
+      int      tx, ty, iy;
+      mp_word  _W;
+      mp_digit *tmpy;
+
+      /* clear counter */
+      _W = 0;
+
+      /* get offsets into the two bignums */
+      ty = MIN(a->used-1, ix);
+      tx = ix - ty;
+
+      /* setup temp aliases */
+      tmpx = a->dp + tx;
+      tmpy = a->dp + ty;
+
+      /* this is the number of times the loop will iterrate, essentially
+         while (tx++ < a->used && ty-- >= 0) { ... }
+       */
+      iy = MIN(a->used-tx, ty+1);
+
+      /* now for squaring tx can never equal ty 
+       * we halve the distance since they approach at a rate of 2x
+       * and we have to round because odd cases need to be executed
+       */
+      iy = MIN(iy, (ty-tx+1)>>1);
+
+      /* execute loop */
+      for (iz = 0; iz < iy; iz++) {
+         _W += ((mp_word)*tmpx++)*((mp_word)*tmpy--);
+      }
+
+      /* double the inner product and add carry */
+      _W = _W + _W + W1;
+
+      /* even columns have the square term in them */
+      if ((ix&1) == 0) {
+         _W += ((mp_word)a->dp[ix>>1])*((mp_word)a->dp[ix>>1]);
+      }
+
+      /* store it */
+      W[ix] = (mp_digit)(_W & MP_MASK);
+
+      /* make next carry */
+      W1 = _W >> ((mp_word)DIGIT_BIT);
+  }
+
+  /* setup dest */
+  olduse  = b->used;
+  b->used = a->used+a->used;
+
+  {
+    mp_digit *tmpb;
+    tmpb = b->dp;
+    for (ix = 0; ix < pa; ix++) {
+      *tmpb++ = W[ix] & MP_MASK;
+    }
+
+    /* clear unused digits [that existed in the old copy of c] */
+    for (; ix < olduse; ix++) {
+      *tmpb++ = 0;
+    }
+  }
+  mp_clamp (b);
+  return MP_OKAY;
+}
+#endif
+
+/* $Source: /cvs/libtom/libtommath/bn_fast_s_mp_sqr.c,v $ */
+/* $Revision: 1.4 $ */
+/* $Date: 2006/12/28 01:25:13 $ */

data/ext/stormlib/StormLib/src/libtommath/bn_mp_2expt.c

@@ -0,0 +1,48 @@
+#include "tommath.h"
+#ifdef BN_MP_2EXPT_C
+/* 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, tomstdenis@gmail.com, http://libtom.org
+ */
+
+/* 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] = ((mp_digit)1) << (b % DIGIT_BIT);
+
+  return MP_OKAY;
+}
+#endif
+
+/* $Source: /cvs/libtom/libtommath/bn_mp_2expt.c,v $ */
+/* $Revision: 1.4 $ */
+/* $Date: 2006/12/28 01:25:13 $ */

data/ext/stormlib/StormLib/src/libtommath/bn_mp_abs.c

@@ -0,0 +1,43 @@
+#include "tommath.h"
+#ifdef BN_MP_ABS_C
+/* 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, tomstdenis@gmail.com, http://libtom.org
+ */
+
+/* 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;
+}
+#endif
+
+/* $Source: /cvs/libtom/libtommath/bn_mp_abs.c,v $ */
+/* $Revision: 1.4 $ */
+/* $Date: 2006/12/28 01:25:13 $ */

data/ext/stormlib/StormLib/src/libtommath/bn_mp_add.c

@@ -0,0 +1,53 @@
+#include "tommath.h"
+#ifdef BN_MP_ADD_C
+/* 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, tomstdenis@gmail.com, http://libtom.org
+ */
+
+/* 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;
+}
+
+#endif
+
+/* $Source: /cvs/libtom/libtommath/bn_mp_add.c,v $ */
+/* $Revision: 1.4 $ */
+/* $Date: 2006/12/28 01:25:13 $ */