Annotation of embedaddon/ntp/libparse/mfp_mul.c, revision 1.1.1.1

1.1       misho       1: /*
                      2:  * /src/NTP/ntp4-dev/libparse/mfp_mul.c,v 4.9 2005/07/17 20:34:40 kardel RELEASE_20050717_A
                      3:  *
                      4:  * mfp_mul.c,v 4.9 2005/07/17 20:34:40 kardel RELEASE_20050717_A
                      5:  *
                      6:  * $Created: Sat Aug 16 20:35:08 1997 $
                      7:  *
                      8:  * Copyright (c) 1997-2005 by Frank Kardel <kardel <AT> ntp.org>
                      9:  *
                     10:  * Redistribution and use in source and binary forms, with or without
                     11:  * modification, are permitted provided that the following conditions
                     12:  * are met:
                     13:  * 1. Redistributions of source code must retain the above copyright
                     14:  *    notice, this list of conditions and the following disclaimer.
                     15:  * 2. Redistributions in binary form must reproduce the above copyright
                     16:  *    notice, this list of conditions and the following disclaimer in the
                     17:  *    documentation and/or other materials provided with the distribution.
                     18:  * 3. Neither the name of the author nor the names of its contributors
                     19:  *    may be used to endorse or promote products derived from this software
                     20:  *    without specific prior written permission.
                     21:  *
                     22:  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
                     23:  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
                     24:  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
                     25:  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
                     26:  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
                     27:  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
                     28:  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
                     29:  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
                     30:  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
                     31:  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
                     32:  * SUCH DAMAGE.
                     33:  *
                     34:  */
                     35: #include <stdio.h>
                     36: #include "ntp_stdlib.h"
                     37: #include "ntp_types.h"
                     38: #include "ntp_fp.h"
                     39: 
                     40: #define LOW_MASK  (u_int32)((1<<(FRACTION_PREC/2))-1)
                     41: #define HIGH_MASK (u_int32)(LOW_MASK << (FRACTION_PREC/2))
                     42: 
                     43: /*
                     44:  * for those who worry about overflows (possibly triggered by static analysis tools):
                     45:  *
                     46:  * Largest value of a 2^n bit number is 2^n-1.
                     47:  * Thus the result is: (2^n-1)*(2^n-1) = 2^2n - 2^n - 2^n + 1 < 2^2n
                     48:  * Here overflow can not happen for 2 reasons:
                     49:  * 1) the code actually multiplies the absolute values of two signed
                     50:  *    64bit quantities.thus effectively multiplying 2 63bit quantities.
                     51:  * 2) Carry propagation is from low to high, building principle is
                     52:  *    addition, so no storage for the 2^2n term from above is needed.
                     53:  */
                     54: 
                     55: void
                     56: mfp_mul(
                     57:        int32   *o_i,
                     58:        u_int32 *o_f,
                     59:        int32    a_i,
                     60:        u_int32  a_f,
                     61:        int32    b_i,
                     62:        u_int32  b_f
                     63:        )
                     64: {
                     65:   int32 i, j;
                     66:   u_int32  f;
                     67:   u_long a[4];                 /* operand a */
                     68:   u_long b[4];                 /* operand b */
                     69:   u_long c[5];                 /* result c - 5 items for performance - see below */
                     70:   u_long carry;
                     71:   
                     72:   int neg = 0;
                     73: 
                     74:   if (a_i < 0)                 /* examine sign situation */
                     75:     {
                     76:       neg = 1;
                     77:       M_NEG(a_i, a_f);
                     78:     }
                     79: 
                     80:   if (b_i < 0)                 /* examine sign situation */
                     81:     {
                     82:       neg = !neg;
                     83:       M_NEG(b_i, b_f);
                     84:     }
                     85:   
                     86:   a[0] = a_f & LOW_MASK;       /* prepare a operand */
                     87:   a[1] = (a_f & HIGH_MASK) >> (FRACTION_PREC/2);
                     88:   a[2] = a_i & LOW_MASK;
                     89:   a[3] = (a_i & HIGH_MASK) >> (FRACTION_PREC/2);
                     90:   
                     91:   b[0] = b_f & LOW_MASK;       /* prepare b operand */
                     92:   b[1] = (b_f & HIGH_MASK) >> (FRACTION_PREC/2);
                     93:   b[2] = b_i & LOW_MASK;
                     94:   b[3] = (b_i & HIGH_MASK) >> (FRACTION_PREC/2);
                     95: 
                     96:   c[0] = c[1] = c[2] = c[3] = c[4] = 0;
                     97: 
                     98:   for (i = 0; i < 4; i++)      /* we do assume 32 * 32 = 64 bit multiplication */
                     99:     for (j = 0; j < 4; j++)
                    100:       {
                    101:        u_long result_low, result_high;
                    102:        int low_index = (i+j)/2;      /* formal [0..3]  - index for low long word */
                    103:        int mid_index = 1+low_index;  /* formal [1..4]! - index for high long word
                    104:                                         will generate unecessary add of 0 to c[4]
                    105:                                         but save 15 'if (result_high) expressions' */
                    106:        int high_index = 1+mid_index; /* formal [2..5]! - index for high word overflow
                    107:                                         - only assigned on overflow (limits range to 2..3) */
                    108: 
                    109:        result_low = (u_long)a[i] * (u_long)b[j];       /* partial product */
                    110: 
                    111:        if ((i+j) & 1)          /* splits across two result registers */
                    112:          {
                    113:            result_high   = result_low >> (FRACTION_PREC/2);
                    114:            result_low  <<= FRACTION_PREC/2;
                    115:            carry         = (unsigned)1<<(FRACTION_PREC/2);
                    116:          }
                    117:        else
                    118:          {                     /* stays in a result register - except for overflows */
                    119:            result_high = 0;
                    120:            carry       = 1;
                    121:          }
                    122: 
                    123:        if (((c[low_index] >> 1) + (result_low >> 1) + ((c[low_index] & result_low & carry) != 0)) &
                    124:            (u_int32)((unsigned)1<<(FRACTION_PREC - 1))) {
                    125:          result_high++;        /* propagate overflows */
                    126:         }
                    127: 
                    128:        c[low_index]   += result_low; /* add up partial products */
                    129: 
                    130:        if (((c[mid_index] >> 1) + (result_high >> 1) + ((c[mid_index] & result_high & 1) != 0)) &
                    131:            (u_int32)((unsigned)1<<(FRACTION_PREC - 1))) {
                    132:          c[high_index]++;              /* propagate overflows of high word sum */
                    133:         }
                    134: 
                    135:        c[mid_index] += result_high;  /* will add a 0 to c[4] once but saves 15 if conditions */
                    136:       }
                    137: 
                    138: #ifdef DEBUG
                    139:   if (debug > 6)
                    140:     printf("mfp_mul: 0x%04lx%04lx%04lx%04lx * 0x%04lx%04lx%04lx%04lx = 0x%08lx%08lx%08lx%08lx\n",
                    141:         a[3], a[2], a[1], a[0], b[3], b[2], b[1], b[0], c[3], c[2], c[1], c[0]);
                    142: #endif
                    143: 
                    144:   if (c[3])                    /* overflow */
                    145:     {
                    146:       i = ((unsigned)1 << (FRACTION_PREC-1)) - 1;
                    147:       f = ~(unsigned)0;
                    148:     }
                    149:   else
                    150:     {                          /* take produkt - discarding extra precision */
                    151:       i = c[2];
                    152:       f = c[1];
                    153:     }
                    154:   
                    155:   if (neg)                     /* recover sign */
                    156:     {
                    157:       M_NEG(i, f);
                    158:     }
                    159: 
                    160:   *o_i = i;
                    161:   *o_f = f;
                    162: 
                    163: #ifdef DEBUG
                    164:   if (debug > 6)
                    165:     printf("mfp_mul: %s * %s => %s\n",
                    166:           mfptoa((u_long)a_i, a_f, 6),
                    167:           mfptoa((u_long)b_i, b_f, 6),
                    168:           mfptoa((u_long)i, f, 6));
                    169: #endif
                    170: }
                    171: 
                    172: /*
                    173:  * History:
                    174:  *
                    175:  * mfp_mul.c,v
                    176:  * Revision 4.9  2005/07/17 20:34:40  kardel
                    177:  * correct carry propagation implementation
                    178:  *
                    179:  * Revision 4.8  2005/07/12 16:17:26  kardel
                    180:  * add explanation why we do not write into c[4]
                    181:  *
                    182:  * Revision 4.7  2005/04/16 17:32:10  kardel
                    183:  * update copyright
                    184:  *
                    185:  * Revision 4.6  2004/11/14 15:29:41  kardel
                    186:  * support PPSAPI, upgrade Copyright to Berkeley style
                    187:  *
                    188:  * Revision 4.3  1999/02/21 12:17:37  kardel
                    189:  * 4.91f reconcilation
                    190:  *
                    191:  * Revision 4.2  1998/12/20 23:45:28  kardel
                    192:  * fix types and warnings
                    193:  *
                    194:  * Revision 4.1  1998/05/24 07:59:57  kardel
                    195:  * conditional debug support
                    196:  *
                    197:  * Revision 4.0  1998/04/10 19:46:38  kardel
                    198:  * Start 4.0 release version numbering
                    199:  *
                    200:  * Revision 1.1  1998/04/10 19:27:47  kardel
                    201:  * initial NTP VERSION 4 integration of PARSE with GPS166 binary support
                    202:  *
                    203:  * Revision 1.1  1997/10/06 21:05:46  kardel
                    204:  * new parse structure
                    205:  *
                    206:  */

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>