sha1-lookup.c

   1 #include "cache.h"
   2 #include "sha1-lookup.h"
   3
   4 /*
   5  * Conventional binary search loop looks like this:
   6  *
   7  *      unsigned lo, hi;
   8  *      do {
   9  *              unsigned mi = (lo + hi) / 2;
  10  *              int cmp = "entry pointed at by mi" minus "target";
  11  *              if (!cmp)
  12  *                      return (mi is the wanted one)
  13  *              if (cmp > 0)
  14  *                      hi = mi; "mi is larger than target"
  15  *              else
  16  *                      lo = mi+1; "mi is smaller than target"
  17  *      } while (lo < hi);
  18  *
  19  * The invariants are:
  20  *
  21  * - When entering the loop, lo points at a slot that is never
  22  *   above the target (it could be at the target), hi points at a
  23  *   slot that is guaranteed to be above the target (it can never
  24  *   be at the target).
  25  *
  26  * - We find a point 'mi' between lo and hi (mi could be the same
  27  *   as lo, but never can be as same as hi), and check if it hits
  28  *   the target.  There are three cases:
  29  *
  30  *    - if it is a hit, we are happy.
  31  *
  32  *    - if it is strictly higher than the target, we set it to hi,
  33  *      and repeat the search.
  34  *
  35  *    - if it is strictly lower than the target, we update lo to
  36  *      one slot after it, because we allow lo to be at the target.
  37  *
  38  *   If the loop exits, there is no matching entry.
  39  *
  40  * When choosing 'mi', we do not have to take the "middle" but
  41  * anywhere in between lo and hi, as long as lo <= mi < hi is
  42  * satisfied.  When we somehow know that the distance between the
  43  * target and lo is much shorter than the target and hi, we could
  44  * pick mi that is much closer to lo than the midway.
  45  *
  46  * Now, we can take advantage of the fact that SHA-1 is a good hash
  47  * function, and as long as there are enough entries in the table, we
  48  * can expect uniform distribution.  An entry that begins with for
  49  * example "deadbeef..." is much likely to appear much later than in
  50  * the midway of the table.  It can reasonably be expected to be near
  51  * 87% (222/256) from the top of the table.
  52  *
  53  * However, we do not want to pick "mi" too precisely.  If the entry at
  54  * the 87% in the above example turns out to be higher than the target
  55  * we are looking for, we would end up narrowing the search space down
  56  * only by 13%, instead of 50% we would get if we did a simple binary
  57  * search.  So we would want to hedge our bets by being less aggressive.
  58  *
  59  * The table at "table" holds at least "nr" entries of "elem_size"
  60  * bytes each.  Each entry has the SHA-1 key at "key_offset".  The
  61  * table is sorted by the SHA-1 key of the entries.  The caller wants
  62  * to find the entry with "key", and knows that the entry at "lo" is
  63  * not higher than the entry it is looking for, and that the entry at
  64  * "hi" is higher than the entry it is looking for.
  65  */
  66 int sha1_entry_pos(const void *table,
  67                    size_t elem_size,
  68                    size_t key_offset,
  69                    unsigned lo, unsigned hi, unsigned nr,
  70                    const unsigned char *key)
  71 {
  72         const unsigned char *base = table;
  73         const unsigned char *hi_key, *lo_key;
  74         unsigned ofs_0;
  75         static int debug_lookup = -1;
  76
  77         if (debug_lookup < 0)
  78                 debug_lookup = !!getenv("GIT_DEBUG_LOOKUP");
  79
  80         if (!nr || lo >= hi)
  81                 return -1;
  82
  83         if (nr == hi)
  84                 hi_key = NULL;
  85         else
  86                 hi_key = base + elem_size * hi + key_offset;
  87         lo_key = base + elem_size * lo + key_offset;
  88
  89         ofs_0 = 0;
  90         do {
  91                 int cmp;
  92                 unsigned ofs, mi, range;
  93                 unsigned lov, hiv, kyv;
  94                 const unsigned char *mi_key;
  95
  96                 range = hi - lo;
  97                 if (hi_key) {
  98                         for (ofs = ofs_0; ofs < 20; ofs++)
  99                                 if (lo_key[ofs] != hi_key[ofs])
 100                                         break;
 101                         ofs_0 = ofs;
 102                         /*
 103                          * byte 0 thru (ofs-1) are the same between
 104                          * lo and hi; ofs is the first byte that is
 105                          * different.
 106                          */
 107                         hiv = hi_key[ofs_0];
 108                         if (ofs_0 < 19)
 109                                 hiv = (hiv << 8) | hi_key[ofs_0+1];
 110                 } else {
 111                         hiv = 256;
 112                         if (ofs_0 < 19)
 113                                 hiv <<= 8;
 114                 }
 115                 lov = lo_key[ofs_0];
 116                 kyv = key[ofs_0];
 117                 if (ofs_0 < 19) {
 118                         lov = (lov << 8) | lo_key[ofs_0+1];
 119                         kyv = (kyv << 8) | key[ofs_0+1];
 120                 }
 121                 assert(lov < hiv);
 122
 123                 if (kyv < lov)
 124                         return -1 - lo;
 125                 if (hiv < kyv)
 126                         return -1 - hi;
 127
 128                 /*
 129                  * Even if we know the target is much closer to 'hi'
 130                  * than 'lo', if we pick too precisely and overshoot
 131                  * (e.g. when we know 'mi' is closer to 'hi' than to
 132                  * 'lo', pick 'mi' that is higher than the target), we
 133                  * end up narrowing the search space by a smaller
 134                  * amount (i.e. the distance between 'mi' and 'hi')
 135                  * than what we would have (i.e. about half of 'lo'
 136                  * and 'hi').  Hedge our bets to pick 'mi' less
 137                  * aggressively, i.e. make 'mi' a bit closer to the
 138                  * middle than we would otherwise pick.
 139                  */
 140                 kyv = (kyv * 6 + lov + hiv) / 8;
 141                 if (lov < hiv - 1) {
 142                         if (kyv == lov)
 143                                 kyv++;
 144                         else if (kyv == hiv)
 145                                 kyv--;
 146                 }
 147                 mi = (range - 1) * (kyv - lov) / (hiv - lov) + lo;
 148
 149                 if (debug_lookup) {
 150                         printf("lo %u hi %u rg %u mi %u ", lo, hi, range, mi);
 151                         printf("ofs %u lov %x, hiv %x, kyv %x\n",
 152                                ofs_0, lov, hiv, kyv);
 153                 }
 154                 if (!(lo <= mi && mi < hi))
 155                         die("assertion failure lo %u mi %u hi %u %s",
 156                             lo, mi, hi, sha1_to_hex(key));
 157
 158                 mi_key = base + elem_size * mi + key_offset;
 159                 cmp = memcmp(mi_key + ofs_0, key + ofs_0, 20 - ofs_0);
 160                 if (!cmp)
 161                         return mi;
 162                 if (cmp > 0) {
 163                         hi = mi;
 164                         hi_key = mi_key;
 165                 } else {
 166                         lo = mi + 1;
 167                         lo_key = mi_key + elem_size;
 168                 }
 169         } while (lo < hi);
 170         return -lo-1;
 171 }