1  /* mpn_gcd_22 -- double limb greatest common divisor.
       2  
       3  Copyright 1994, 1996, 2000, 2001, 2009, 2012, 2019 Free Software Foundation, Inc.
       4  
       5  This file is part of the GNU MP Library.
       6  
       7  The GNU MP Library is free software; you can redistribute it and/or modify
       8  it under the terms of either:
       9  
      10    * the GNU Lesser General Public License as published by the Free
      11      Software Foundation; either version 3 of the License, or (at your
      12      option) any later version.
      13  
      14  or
      15  
      16    * the GNU General Public License as published by the Free Software
      17      Foundation; either version 2 of the License, or (at your option) any
      18      later version.
      19  
      20  or both in parallel, as here.
      21  
      22  The GNU MP Library is distributed in the hope that it will be useful, but
      23  WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
      24  or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
      25  for more details.
      26  
      27  You should have received copies of the GNU General Public License and the
      28  GNU Lesser General Public License along with the GNU MP Library.  If not,
      29  see https://www.gnu.org/licenses/.  */
      30  
      31  #include "gmp-impl.h"
      32  #include "longlong.h"
      33  
      34  #if GMP_NAIL_BITS > 0
      35  #error Nails not supported.
      36  #endif
      37  
      38  mp_double_limb_t
      39  mpn_gcd_22 (mp_limb_t u1, mp_limb_t u0, mp_limb_t v1, mp_limb_t v0)
      40  {
      41    mp_double_limb_t g;
      42    ASSERT (u0 & v0 & 1);
      43  
      44    /* Implicit least significant bit */
      45    u0 = (u0 >> 1) | (u1 << (GMP_LIMB_BITS - 1));
      46    u1 >>= 1;
      47  
      48    v0 = (v0 >> 1) | (v1 << (GMP_LIMB_BITS - 1));
      49    v1 >>= 1;
      50  
      51    while (u1 || v1) /* u1 == 0 can happen at most twice per call */
      52      {
      53        mp_limb_t vgtu, t1, t0;
      54        sub_ddmmss (t1, t0, u1, u0, v1, v0);
      55        vgtu = LIMB_HIGHBIT_TO_MASK(t1);
      56  
      57        if (UNLIKELY (t0 == 0))
      58  	{
      59  	  if (t1 == 0)
      60  	    {
      61  	      g.d1 = (u1 << 1) | (u0 >> (GMP_LIMB_BITS - 1));
      62  	      g.d0 = (u0 << 1) | 1;
      63  	      return g;
      64  	    }
      65  	  int c;
      66  	  count_trailing_zeros (c, t1);
      67  
      68  	  /* v1 = min (u1, v1) */
      69  	  v1 += (vgtu & t1);
      70  	  /* u0 = |u1 - v1| */
      71  	  u0 = (t1 ^ vgtu) - vgtu;
      72  	  ASSERT (c < GMP_LIMB_BITS - 1);
      73  	  u0 >>= c + 1;
      74  	  u1 = 0;
      75  	}
      76        else
      77  	{
      78  	  int c;
      79  	  count_trailing_zeros (c, t0);
      80  	  c++;
      81  	  /* V <-- min (U, V).
      82  
      83  	     Assembly version should use cmov. Another alternative,
      84  	     avoiding carry propagation, would be
      85  
      86  	     v0 += vgtu & t0; v1 += vtgu & (u1 - v1);
      87  	  */
      88  	  add_ssaaaa (v1, v0, v1, v0, vgtu & t1, vgtu & t0);
      89  	  /* U  <--  |U - V|
      90  	     No carry handling needed in this conditional negation,
      91  	     since t0 != 0. */
      92  	  u0 = (t0 ^ vgtu) - vgtu;
      93  	  u1 = t1 ^ vgtu;
      94  	  if (UNLIKELY (c == GMP_LIMB_BITS))
      95  	    {
      96  	      u0 = u1;
      97  	      u1 = 0;
      98  	    }
      99  	  else
     100  	    {
     101  	      u0 = (u0 >> c) | (u1 << (GMP_LIMB_BITS - c));
     102  	      u1 >>= c;
     103  	    }
     104  	}
     105      }
     106    while ((v0 | u0) & GMP_LIMB_HIGHBIT)
     107      { /* At most two iterations */
     108        mp_limb_t vgtu, t0;
     109        int c;
     110        sub_ddmmss (vgtu, t0, 0, u0, 0, v0);
     111        if (UNLIKELY (t0 == 0))
     112  	{
     113  	  g.d1 = u0 >> (GMP_LIMB_BITS - 1);
     114  	  g.d0 = (u0 << 1) | 1;
     115  	  return g;
     116  	}
     117  
     118        /* v <-- min (u, v) */
     119        v0 += (vgtu & t0);
     120  
     121        /* u <-- |u - v| */
     122        u0 = (t0 ^ vgtu) - vgtu;
     123  
     124        count_trailing_zeros (c, t0);
     125        u0 = (u0 >> 1) >> c;
     126      }
     127  
     128    g.d0 = mpn_gcd_11 ((u0 << 1) + 1, (v0 << 1) + 1);
     129    g.d1 = 0;
     130    return g;
     131  }