lib/libc/stdlib/heapsort.c

   1 /*-
   2  * Copyright (c) 1991, 1993
   3  *      The Regents of the University of California.  All rights reserved.
   4  *
   5  * This code is derived from software contributed to Berkeley by
   6  * Ronnie Kon at Mindcraft Inc., Kevin Lew and Elmer Yglesias.
   7  *
   8  * Redistribution and use in source and binary forms, with or without
   9  * modification, are permitted provided that the following conditions
  10  * are met:
  11  * 1. Redistributions of source code must retain the above copyright
  12  *    notice, this list of conditions and the following disclaimer.
  13  * 2. Redistributions in binary form must reproduce the above copyright
  14  *    notice, this list of conditions and the following disclaimer in the
  15  *    documentation and/or other materials provided with the distribution.
  16  * 3. All advertising materials mentioning features or use of this software
  17  *    must display the following acknowledgement:
  18  *      This product includes software developed by the University of
  19  *      California, Berkeley and its contributors.
  20  * 4. Neither the name of the University nor the names of its contributors
  21  *    may be used to endorse or promote products derived from this software
  22  *    without specific prior written permission.
  23  *
  24  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  25  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  26  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  27  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  28  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  29  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  30  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  31  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  32  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  33  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  34  * SUCH DAMAGE.
  35  *
  36  * $DragonFly: src/lib/libc/stdlib/heapsort.c,v 1.5 2005/11/20 12:37:48 swildner Exp $
  37  *
  38  * @(#)heapsort.c       8.1 (Berkeley) 6/4/93
  39  */
  40
  41 #include <errno.h>
  42 #include <stddef.h>
  43 #include <stdlib.h>
  44
  45 /*
  46  * Swap two areas of size number of bytes.  Although qsort(3) permits random
  47  * blocks of memory to be sorted, sorting pointers is almost certainly the
  48  * common case (and, were it not, could easily be made so).  Regardless, it
  49  * isn't worth optimizing; the SWAP's get sped up by the cache, and pointer
  50  * arithmetic gets lost in the time required for comparison function calls.
  51  */
  52 #define SWAP(a, b, count, size, tmp) { \
  53         count = size; \
  54         do { \
  55                 tmp = *a; \
  56                 *a++ = *b; \
  57                 *b++ = tmp; \
  58         } while (--count); \
  59 }
  60
  61 /* Copy one block of size size to another. */
  62 #define COPY(a, b, count, size, tmp1, tmp2) { \
  63         count = size; \
  64         tmp1 = a; \
  65         tmp2 = b; \
  66         do { \
  67                 *tmp1++ = *tmp2++; \
  68         } while (--count); \
  69 }
  70
  71 /*
  72  * Build the list into a heap, where a heap is defined such that for
  73  * the records K1 ... KN, Kj/2 >= Kj for 1 <= j/2 <= j <= N.
  74  *
  75  * There two cases.  If j == nmemb, select largest of Ki and Kj.  If
  76  * j < nmemb, select largest of Ki, Kj and Kj+1.
  77  */
  78 #define CREATE(initval, nmemb, par_i, child_i, par, child, size, count, tmp) { \
  79         for (par_i = initval; (child_i = par_i * 2) <= nmemb; \
  80             par_i = child_i) { \
  81                 child = base + child_i * size; \
  82                 if (child_i < nmemb && compar(child, child + size) < 0) { \
  83                         child += size; \
  84                         ++child_i; \
  85                 } \
  86                 par = base + par_i * size; \
  87                 if (compar(child, par) <= 0) \
  88                         break; \
  89                 SWAP(par, child, count, size, tmp); \
  90         } \
  91 }
  92
  93 /*
  94  * Select the top of the heap and 'heapify'.  Since by far the most expensive
  95  * action is the call to the compar function, a considerable optimization
  96  * in the average case can be achieved due to the fact that k, the displaced
  97  * elememt, is ususally quite small, so it would be preferable to first
  98  * heapify, always maintaining the invariant that the larger child is copied
  99  * over its parent's record.
 100  *
 101  * Then, starting from the *bottom* of the heap, finding k's correct place,
 102  * again maintianing the invariant.  As a result of the invariant no element
 103  * is 'lost' when k is assigned its correct place in the heap.
 104  *
 105  * The time savings from this optimization are on the order of 15-20% for the
 106  * average case. See Knuth, Vol. 3, page 158, problem 18.
 107  *
 108  * XXX Don't break the #define SELECT line, below.  Reiser cpp gets upset.
 109  */
 110 #define SELECT(par_i, child_i, nmemb, par, child, size, k, count, tmp1, tmp2) { \
 111         for (par_i = 1; (child_i = par_i * 2) <= nmemb; par_i = child_i) { \
 112                 child = base + child_i * size; \
 113                 if (child_i < nmemb && compar(child, child + size) < 0) { \
 114                         child += size; \
 115                         ++child_i; \
 116                 } \
 117                 par = base + par_i * size; \
 118                 COPY(par, child, count, size, tmp1, tmp2); \
 119         } \
 120         for (;;) { \
 121                 child_i = par_i; \
 122                 par_i = child_i / 2; \
 123                 child = base + child_i * size; \
 124                 par = base + par_i * size; \
 125                 if (child_i == 1 || compar(k, par) < 0) { \
 126                         COPY(child, k, count, size, tmp1, tmp2); \
 127                         break; \
 128                 } \
 129                 COPY(child, par, count, size, tmp1, tmp2); \
 130         } \
 131 }
 132
 133 /*
 134  * Heapsort -- Knuth, Vol. 3, page 145.  Runs in O (N lg N), both average
 135  * and worst.  While heapsort is faster than the worst case of quicksort,
 136  * the BSD quicksort does median selection so that the chance of finding
 137  * a data set that will trigger the worst case is nonexistent.  Heapsort's
 138  * only advantage over quicksort is that it requires little additional memory.
 139  */
 140 int
 141 heapsort(void *vbase, size_t nmemb, size_t size,
 142          int (*compar)(const void *, const void *))
 143 {
 144         int cnt, i, j, l;
 145         char tmp, *tmp1, *tmp2;
 146         char *base, *k, *p, *t;
 147
 148         if (nmemb <= 1)
 149                 return (0);
 150
 151         if (!size) {
 152                 errno = EINVAL;
 153                 return (-1);
 154         }
 155
 156         if ((k = malloc(size)) == NULL)
 157                 return (-1);
 158
 159         /*
 160          * Items are numbered from 1 to nmemb, so offset from size bytes
 161          * below the starting address.
 162          */
 163         base = (char *)vbase - size;
 164
 165         for (l = nmemb / 2 + 1; --l;)
 166                 CREATE(l, nmemb, i, j, t, p, size, cnt, tmp);
 167
 168         /*
 169          * For each element of the heap, save the largest element into its
 170          * final slot, save the displaced element (k), then recreate the
 171          * heap.
 172          */
 173         while (nmemb > 1) {
 174                 COPY(k, base + nmemb * size, cnt, size, tmp1, tmp2);
 175                 COPY(base + nmemb * size, base + size, cnt, size, tmp1, tmp2);
 176                 --nmemb;
 177                 SELECT(i, j, nmemb, t, p, size, k, cnt, tmp1, tmp2);
 178         }
 179         free(k);
 180         return (0);
 181 }