gdb/bcache.c

   1 /* Implement a cached obstack.
   2    Written by Fred Fish <fnf@cygnus.com>
   3    Rewritten by Jim Blandy <jimb@cygnus.com>
   4
   5    Copyright (C) 1999, 2000, 2002, 2003, 2007, 2008, 2009, 2010
   6    Free Software Foundation, Inc.
   7
   8    This file is part of GDB.
   9
  10    This program is free software; you can redistribute it and/or modify
  11    it under the terms of the GNU General Public License as published by
  12    the Free Software Foundation; either version 3 of the License, or
  13    (at your option) any later version.
  14
  15    This program is distributed in the hope that it will be useful,
  16    but WITHOUT ANY WARRANTY; without even the implied warranty of
  17    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18    GNU General Public License for more details.
  19
  20    You should have received a copy of the GNU General Public License
  21    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
  22
  23 #include "defs.h"
  24 #include "gdb_obstack.h"
  25 #include "bcache.h"
  26 #include "gdb_string.h"         /* For memcpy declaration */
  27 #include "gdb_assert.h"
  28
  29 #include <stddef.h>
  30 #include <stdlib.h>
  31
  32 /* The type used to hold a single bcache string.  The user data is
  33    stored in d.data.  Since it can be any type, it needs to have the
  34    same alignment as the most strict alignment of any type on the host
  35    machine.  I don't know of any really correct way to do this in
  36    stock ANSI C, so just do it the same way obstack.h does.  */
  37
  38 struct bstring
  39 {
  40   /* Hash chain.  */
  41   struct bstring *next;
  42   /* Assume the data length is no more than 64k.  */
  43   unsigned short length;
  44   /* The half hash hack.  This contains the upper 16 bits of the hash
  45      value and is used as a pre-check when comparing two strings and
  46      avoids the need to do length or memcmp calls.  It proves to be
  47      roughly 100% effective.  */
  48   unsigned short half_hash;
  49
  50   union
  51   {
  52     char data[1];
  53     double dummy;
  54   }
  55   d;
  56 };
  57
  58
  59 /* The structure for a bcache itself.  The bcache is initialized, in
  60    bcache_xmalloc(), by filling it with zeros and then setting the
  61    corresponding obstack's malloc() and free() methods.  */
  62
  63 struct bcache
  64 {
  65   /* All the bstrings are allocated here.  */
  66   struct obstack cache;
  67
  68   /* How many hash buckets we're using.  */
  69   unsigned int num_buckets;
  70
  71   /* Hash buckets.  This table is allocated using malloc, so when we
  72      grow the table we can return the old table to the system.  */
  73   struct bstring **bucket;
  74
  75   /* Statistics.  */
  76   unsigned long unique_count;   /* number of unique strings */
  77   long total_count;     /* total number of strings cached, including dups */
  78   long unique_size;     /* size of unique strings, in bytes */
  79   long total_size;      /* total number of bytes cached, including dups */
  80   long structure_size;  /* total size of bcache, including infrastructure */
  81   /* Number of times that the hash table is expanded and hence
  82      re-built, and the corresponding number of times that a string is
  83      [re]hashed as part of entering it into the expanded table.  The
  84      total number of hashes can be computed by adding TOTAL_COUNT to
  85      expand_hash_count.  */
  86   unsigned long expand_count;
  87   unsigned long expand_hash_count;
  88   /* Number of times that the half-hash compare hit (compare the upper
  89      16 bits of hash values) hit, but the corresponding combined
  90      length/data compare missed.  */
  91   unsigned long half_hash_miss_count;
  92 };
  93
  94 /* The old hash function was stolen from SDBM. This is what DB 3.0 uses now,
  95  * and is better than the old one.
  96  */
  97 \f
  98 unsigned long
  99 hash(const void *addr, int length)
 100 {
 101   const unsigned char *k, *e;
 102   unsigned long h;
 103
 104   k = (const unsigned char *)addr;
 105   e = k+length;
 106   for (h=0; k< e;++k)
 107     {
 108       h *=16777619;
 109       h ^= *k;
 110     }
 111   return (h);
 112 }
 113 \f
 114 /* Growing the bcache's hash table.  */
 115
 116 /* If the average chain length grows beyond this, then we want to
 117    resize our hash table.  */
 118 #define CHAIN_LENGTH_THRESHOLD (5)
 119
 120 static void
 121 expand_hash_table (struct bcache *bcache)
 122 {
 123   /* A table of good hash table sizes.  Whenever we grow, we pick the
 124      next larger size from this table.  sizes[i] is close to 1 << (i+10),
 125      so we roughly double the table size each time.  After we fall off
 126      the end of this table, we just double.  Don't laugh --- there have
 127      been executables sighted with a gigabyte of debug info.  */
 128   static unsigned long sizes[] = {
 129     1021, 2053, 4099, 8191, 16381, 32771,
 130     65537, 131071, 262144, 524287, 1048573, 2097143,
 131     4194301, 8388617, 16777213, 33554467, 67108859, 134217757,
 132     268435459, 536870923, 1073741827, 2147483659UL
 133   };
 134   unsigned int new_num_buckets;
 135   struct bstring **new_buckets;
 136   unsigned int i;
 137
 138   /* Count the stats.  Every unique item needs to be re-hashed and
 139      re-entered.  */
 140   bcache->expand_count++;
 141   bcache->expand_hash_count += bcache->unique_count;
 142
 143   /* Find the next size.  */
 144   new_num_buckets = bcache->num_buckets * 2;
 145   for (i = 0; i < (sizeof (sizes) / sizeof (sizes[0])); i++)
 146     if (sizes[i] > bcache->num_buckets)
 147       {
 148         new_num_buckets = sizes[i];
 149         break;
 150       }
 151
 152   /* Allocate the new table.  */
 153   {
 154     size_t new_size = new_num_buckets * sizeof (new_buckets[0]);
 155
 156     new_buckets = (struct bstring **) xmalloc (new_size);
 157     memset (new_buckets, 0, new_size);
 158
 159     bcache->structure_size -= (bcache->num_buckets
 160                                * sizeof (bcache->bucket[0]));
 161     bcache->structure_size += new_size;
 162   }
 163
 164   /* Rehash all existing strings.  */
 165   for (i = 0; i < bcache->num_buckets; i++)
 166     {
 167       struct bstring *s, *next;
 168
 169       for (s = bcache->bucket[i]; s; s = next)
 170         {
 171           struct bstring **new_bucket;
 172           next = s->next;
 173
 174           new_bucket = &new_buckets[(hash (&s->d.data, s->length)
 175                                      % new_num_buckets)];
 176           s->next = *new_bucket;
 177           *new_bucket = s;
 178         }
 179     }
 180
 181   /* Plug in the new table.  */
 182   if (bcache->bucket)
 183     xfree (bcache->bucket);
 184   bcache->bucket = new_buckets;
 185   bcache->num_buckets = new_num_buckets;
 186 }
 187
 188 \f
 189 /* Looking up things in the bcache.  */
 190
 191 /* The number of bytes needed to allocate a struct bstring whose data
 192    is N bytes long.  */
 193 #define BSTRING_SIZE(n) (offsetof (struct bstring, d.data) + (n))
 194
 195 /* Find a copy of the LENGTH bytes at ADDR in BCACHE.  If BCACHE has
 196    never seen those bytes before, add a copy of them to BCACHE.  In
 197    either case, return a pointer to BCACHE's copy of that string.  */
 198 const void *
 199 bcache (const void *addr, int length, struct bcache *bcache)
 200 {
 201   return bcache_full (addr, length, bcache, NULL);
 202 }
 203
 204 /* Find a copy of the LENGTH bytes at ADDR in BCACHE.  If BCACHE has
 205    never seen those bytes before, add a copy of them to BCACHE.  In
 206    either case, return a pointer to BCACHE's copy of that string.  If
 207    optional ADDED is not NULL, return 1 in case of new entry or 0 if
 208    returning an old entry.  */
 209
 210 const void *
 211 bcache_full (const void *addr, int length, struct bcache *bcache, int *added)
 212 {
 213   unsigned long full_hash;
 214   unsigned short half_hash;
 215   int hash_index;
 216   struct bstring *s;
 217
 218   if (added)
 219     *added = 0;
 220
 221   /* Lazily initialize the obstack.  This can save quite a bit of
 222      memory in some cases.  */
 223   if (bcache->total_count == 0)
 224     {
 225       /* We could use obstack_specify_allocation here instead, but
 226          gdb_obstack.h specifies the allocation/deallocation
 227          functions.  */
 228       obstack_init (&bcache->cache);
 229     }
 230
 231   /* If our average chain length is too high, expand the hash table.  */
 232   if (bcache->unique_count >= bcache->num_buckets * CHAIN_LENGTH_THRESHOLD)
 233     expand_hash_table (bcache);
 234
 235   bcache->total_count++;
 236   bcache->total_size += length;
 237
 238   full_hash = hash (addr, length);
 239   half_hash = (full_hash >> 16);
 240   hash_index = full_hash % bcache->num_buckets;
 241
 242   /* Search the hash bucket for a string identical to the caller's.
 243      As a short-circuit first compare the upper part of each hash
 244      values.  */
 245   for (s = bcache->bucket[hash_index]; s; s = s->next)
 246     {
 247       if (s->half_hash == half_hash)
 248         {
 249           if (s->length == length
 250               && ! memcmp (&s->d.data, addr, length))
 251             return &s->d.data;
 252           else
 253             bcache->half_hash_miss_count++;
 254         }
 255     }
 256
 257   /* The user's string isn't in the list.  Insert it after *ps.  */
 258   {
 259     struct bstring *new
 260       = obstack_alloc (&bcache->cache, BSTRING_SIZE (length));
 261
 262     memcpy (&new->d.data, addr, length);
 263     new->length = length;
 264     new->next = bcache->bucket[hash_index];
 265     new->half_hash = half_hash;
 266     bcache->bucket[hash_index] = new;
 267
 268     bcache->unique_count++;
 269     bcache->unique_size += length;
 270     bcache->structure_size += BSTRING_SIZE (length);
 271
 272     if (added)
 273       *added = 1;
 274
 275     return &new->d.data;
 276   }
 277 }
 278 \f
 279 /* Allocating and freeing bcaches.  */
 280
 281 struct bcache *
 282 bcache_xmalloc (void)
 283 {
 284   /* Allocate the bcache pre-zeroed.  */
 285   struct bcache *b = XCALLOC (1, struct bcache);
 286
 287   return b;
 288 }
 289
 290 /* Free all the storage associated with BCACHE.  */
 291 void
 292 bcache_xfree (struct bcache *bcache)
 293 {
 294   if (bcache == NULL)
 295     return;
 296   /* Only free the obstack if we actually initialized it.  */
 297   if (bcache->total_count > 0)
 298     obstack_free (&bcache->cache, 0);
 299   xfree (bcache->bucket);
 300   xfree (bcache);
 301 }
 302
 303
 304 \f
 305 /* Printing statistics.  */
 306
 307 static void
 308 print_percentage (int portion, int total)
 309 {
 310   if (total == 0)
 311     /* i18n: Like "Percentage of duplicates, by count: (not applicable)" */
 312     printf_filtered (_("(not applicable)\n"));
 313   else
 314     printf_filtered ("%3d%%\n", (int) (portion * 100.0 / total));
 315 }
 316
 317
 318 /* Print statistics on BCACHE's memory usage and efficacity at
 319    eliminating duplication.  NAME should describe the kind of data
 320    BCACHE holds.  Statistics are printed using `printf_filtered' and
 321    its ilk.  */
 322 void
 323 print_bcache_statistics (struct bcache *c, char *type)
 324 {
 325   int occupied_buckets;
 326   int max_chain_length;
 327   int median_chain_length;
 328   int max_entry_size;
 329   int median_entry_size;
 330
 331   /* Count the number of occupied buckets, tally the various string
 332      lengths, and measure chain lengths.  */
 333   {
 334     unsigned int b;
 335     int *chain_length = XCALLOC (c->num_buckets + 1, int);
 336     int *entry_size = XCALLOC (c->unique_count + 1, int);
 337     int stringi = 0;
 338
 339     occupied_buckets = 0;
 340
 341     for (b = 0; b < c->num_buckets; b++)
 342       {
 343         struct bstring *s = c->bucket[b];
 344
 345         chain_length[b] = 0;
 346
 347         if (s)
 348           {
 349             occupied_buckets++;
 350
 351             while (s)
 352               {
 353                 gdb_assert (b < c->num_buckets);
 354                 chain_length[b]++;
 355                 gdb_assert (stringi < c->unique_count);
 356                 entry_size[stringi++] = s->length;
 357                 s = s->next;
 358               }
 359           }
 360       }
 361
 362     /* To compute the median, we need the set of chain lengths sorted.  */
 363     qsort (chain_length, c->num_buckets, sizeof (chain_length[0]),
 364            compare_positive_ints);
 365     qsort (entry_size, c->unique_count, sizeof (entry_size[0]),
 366            compare_positive_ints);
 367
 368     if (c->num_buckets > 0)
 369       {
 370         max_chain_length = chain_length[c->num_buckets - 1];
 371         median_chain_length = chain_length[c->num_buckets / 2];
 372       }
 373     else
 374       {
 375         max_chain_length = 0;
 376         median_chain_length = 0;
 377       }
 378     if (c->unique_count > 0)
 379       {
 380         max_entry_size = entry_size[c->unique_count - 1];
 381         median_entry_size = entry_size[c->unique_count / 2];
 382       }
 383     else
 384       {
 385         max_entry_size = 0;
 386         median_entry_size = 0;
 387       }
 388
 389     xfree (chain_length);
 390     xfree (entry_size);
 391   }
 392
 393   printf_filtered (_("  Cached '%s' statistics:\n"), type);
 394   printf_filtered (_("    Total object count:  %ld\n"), c->total_count);
 395   printf_filtered (_("    Unique object count: %lu\n"), c->unique_count);
 396   printf_filtered (_("    Percentage of duplicates, by count: "));
 397   print_percentage (c->total_count - c->unique_count, c->total_count);
 398   printf_filtered ("\n");
 399
 400   printf_filtered (_("    Total object size:   %ld\n"), c->total_size);
 401   printf_filtered (_("    Unique object size:  %ld\n"), c->unique_size);
 402   printf_filtered (_("    Percentage of duplicates, by size:  "));
 403   print_percentage (c->total_size - c->unique_size, c->total_size);
 404   printf_filtered ("\n");
 405
 406   printf_filtered (_("    Max entry size:     %d\n"), max_entry_size);
 407   printf_filtered (_("    Average entry size: "));
 408   if (c->unique_count > 0)
 409     printf_filtered ("%ld\n", c->unique_size / c->unique_count);
 410   else
 411     /* i18n: "Average entry size: (not applicable)" */
 412     printf_filtered (_("(not applicable)\n"));
 413   printf_filtered (_("    Median entry size:  %d\n"), median_entry_size);
 414   printf_filtered ("\n");
 415
 416   printf_filtered (_("    Total memory used by bcache, including overhead: %ld\n"),
 417                    c->structure_size);
 418   printf_filtered (_("    Percentage memory overhead: "));
 419   print_percentage (c->structure_size - c->unique_size, c->unique_size);
 420   printf_filtered (_("    Net memory savings:         "));
 421   print_percentage (c->total_size - c->structure_size, c->total_size);
 422   printf_filtered ("\n");
 423
 424   printf_filtered (_("    Hash table size:           %3d\n"), c->num_buckets);
 425   printf_filtered (_("    Hash table expands:        %lu\n"),
 426                    c->expand_count);
 427   printf_filtered (_("    Hash table hashes:         %lu\n"),
 428                    c->total_count + c->expand_hash_count);
 429   printf_filtered (_("    Half hash misses:          %lu\n"),
 430                    c->half_hash_miss_count);
 431   printf_filtered (_("    Hash table population:     "));
 432   print_percentage (occupied_buckets, c->num_buckets);
 433   printf_filtered (_("    Median hash chain length:  %3d\n"),
 434                    median_chain_length);
 435   printf_filtered (_("    Average hash chain length: "));
 436   if (c->num_buckets > 0)
 437     printf_filtered ("%3lu\n", c->unique_count / c->num_buckets);
 438   else
 439     /* i18n: "Average hash chain length: (not applicable)" */
 440     printf_filtered (_("(not applicable)\n"));
 441   printf_filtered (_("    Maximum hash chain length: %3d\n"), max_chain_length);
 442   printf_filtered ("\n");
 443 }
 444
 445 int
 446 bcache_memory_used (struct bcache *bcache)
 447 {
 448   if (bcache->total_count == 0)
 449     return 0;
 450   return obstack_memory_used (&bcache->cache);
 451 }