gvdb-builder.c

   1 /*
   2  * Copyright © 2010 Codethink Limited
   3  *
   4  * This library is free software; you can redistribute it and/or
   5  * modify it under the terms of the GNU Lesser General Public
   6  * License as published by the Free Software Foundation; either
   7  * version 2 of the licence, or (at your option) any later version.
   8  *
   9  * This library is distributed in the hope that it will be useful,
  10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  12  * Lesser General Public License for more details.
  13  *
  14  * You should have received a copy of the GNU Lesser General Public
  15  * License along with this library; if not, write to the
  16  * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  17  * Boston, MA 02111-1307, USA.
  18  *
  19  * Author: Ryan Lortie <desrt@desrt.ca>
  20  */
  21
  22 #include "gvdb-builder.h"
  23 #include "gvdb-format.h"
  24
  25 #include <glib.h>
  26 #include <fcntl.h>
  27 #include <unistd.h>
  28 #include <string.h>
  29
  30
  31 struct _GvdbItem
  32 {
  33   gchar *key;
  34   guint32 hash_value;
  35   guint32_le assigned_index;
  36   GvdbItem *parent;
  37   GvdbItem *sibling;
  38   GvdbItem *next;
  39   GVariant *options;
  40
  41   /* one of:
  42    * this:
  43    */
  44   GVariant *value;
  45
  46   /* this: */
  47   GHashTable *table;
  48
  49   /* or this: */
  50   GvdbItem *child;
  51 };
  52
  53 static void
  54 gvdb_item_free (gpointer data)
  55 {
  56   GvdbItem *item = data;
  57
  58   g_free (item->key);
  59
  60   if (item->value)
  61     g_variant_unref (item->value);
  62
  63   if (item->options)
  64     g_variant_unref (item->options);
  65
  66   if (item->table)
  67     g_hash_table_unref (item->table);
  68
  69   g_slice_free (GvdbItem, item);
  70 }
  71
  72 GHashTable *
  73 gvdb_hash_table_new (GHashTable  *parent,
  74                      const gchar *name_in_parent)
  75 {
  76   GHashTable *table;
  77
  78   table = g_hash_table_new_full (g_str_hash, g_str_equal,
  79                                  g_free, gvdb_item_free);
  80
  81   if (parent)
  82     {
  83       GvdbItem *item;
  84
  85       item = gvdb_hash_table_insert (parent, name_in_parent);
  86       gvdb_item_set_hash_table (item, table);
  87     }
  88
  89   return table;
  90 }
  91
  92 static guint32
  93 djb_hash (const gchar *key)
  94 {
  95   guint32 hash_value = 5381;
  96
  97   while (*key)
  98     hash_value = hash_value * 33 + *key++;
  99
 100   return hash_value;
 101 }
 102
 103 GvdbItem *
 104 gvdb_hash_table_insert (GHashTable  *table,
 105                         const gchar *key)
 106 {
 107   GvdbItem *item;
 108
 109   item = g_slice_new0 (GvdbItem);
 110   item->key = g_strdup (key);
 111   item->hash_value = djb_hash (key);
 112
 113   g_hash_table_insert (table, g_strdup (key), item);
 114
 115   return item;
 116 }
 117
 118 void
 119 gvdb_hash_table_insert_string (GHashTable  *table,
 120                                const gchar *key,
 121                                const gchar *value)
 122 {
 123   GvdbItem *item;
 124
 125   item = gvdb_hash_table_insert (table, key);
 126   gvdb_item_set_value (item, g_variant_new_string (value));
 127 }
 128
 129 void
 130 gvdb_item_set_value (GvdbItem *item,
 131                      GVariant *value)
 132 {
 133   g_return_if_fail (!item->value && !item->table && !item->child);
 134
 135   item->value = g_variant_ref_sink (value);
 136 }
 137
 138 void
 139 gvdb_item_set_options (GvdbItem *item,
 140                        GVariant *options)
 141 {
 142   g_return_if_fail (!item->options);
 143
 144   item->options = g_variant_ref_sink (options);
 145 }
 146
 147 void
 148 gvdb_item_set_hash_table (GvdbItem   *item,
 149                           GHashTable *table)
 150 {
 151   g_return_if_fail (!item->value && !item->table && !item->child);
 152
 153   item->table = g_hash_table_ref (table);
 154 }
 155
 156 void
 157 gvdb_item_set_parent (GvdbItem *item,
 158                       GvdbItem *parent)
 159 {
 160   GvdbItem **node;
 161
 162   g_return_if_fail (g_str_has_prefix (item->key, parent->key));
 163   g_return_if_fail (!parent->value && !parent->table);
 164   g_return_if_fail (!item->parent && !item->sibling);
 165
 166   for (node = &parent->child; *node; node = &(*node)->sibling)
 167     if (strcmp ((*node)->key, item->key) > 0)
 168       break;
 169
 170   item->parent = parent;
 171   item->sibling = *node;
 172   *node = item;
 173 }
 174
 175 typedef struct
 176 {
 177   GvdbItem **buckets;
 178   gint n_buckets;
 179 } HashTable;
 180
 181 HashTable *
 182 hash_table_new (gint n_buckets)
 183 {
 184   HashTable *table;
 185
 186   table = g_slice_new (HashTable);
 187   table->buckets = g_new0 (GvdbItem *, n_buckets);
 188   table->n_buckets = n_buckets;
 189
 190   return table;
 191 }
 192
 193 static void
 194 hash_table_insert (gpointer key,
 195                    gpointer value,
 196                    gpointer data)
 197 {
 198   guint32 hash_value, bucket;
 199   HashTable *table = data;
 200   GvdbItem *item = value;
 201
 202   hash_value = djb_hash (key);
 203   bucket = hash_value % table->n_buckets;
 204   item->next = table->buckets[bucket];
 205   table->buckets[bucket] = item;
 206 }
 207
 208 static guint32_le
 209 item_to_index (GvdbItem *item)
 210 {
 211   if (item != NULL)
 212     return item->assigned_index;
 213
 214   return guint32_to_le (-1u);
 215 }
 216
 217 typedef struct
 218 {
 219   GQueue *chunks;
 220   guint64 offset;
 221   gboolean byteswap;
 222 } FileBuilder;
 223
 224 typedef struct
 225 {
 226   gsize offset;
 227   gsize size;
 228   gpointer data;
 229 } FileChunk;
 230
 231 static gpointer
 232 file_builder_allocate (FileBuilder         *fb,
 233                        guint                alignment,
 234                        gsize                size,
 235                        struct gvdb_pointer *pointer)
 236 {
 237   FileChunk *chunk;
 238
 239   fb->offset += (-fb->offset) & (alignment - 1);
 240   chunk = g_slice_new (FileChunk);
 241   chunk->offset = fb->offset;
 242   chunk->size = size;
 243   chunk->data = g_malloc (size);
 244
 245   pointer->start = guint32_to_le (fb->offset);
 246   fb->offset += size;
 247   pointer->end = guint32_to_le (fb->offset);
 248
 249   g_queue_push_tail (fb->chunks, chunk);
 250
 251   return chunk->data;
 252 }
 253
 254 static void
 255 file_builder_add_value (FileBuilder         *fb,
 256                         GVariant            *value,
 257                         struct gvdb_pointer *pointer)
 258 {
 259   GVariant *variant, *normal;
 260   gpointer data;
 261   gsize size;
 262
 263   if (fb->byteswap)
 264     {
 265       value = g_variant_byteswap (value);
 266       variant = g_variant_new_variant (value);
 267       g_variant_unref (value);
 268     }
 269   else
 270     variant = g_variant_new_variant (value);
 271
 272   normal = g_variant_get_normal_form (variant);
 273   g_variant_unref (variant);
 274
 275   size = g_variant_get_size (normal);
 276   data = file_builder_allocate (fb, 8, size, pointer);
 277   g_variant_store (normal, data);
 278   g_variant_unref (normal);
 279 }
 280
 281 static void
 282 file_builder_add_options (FileBuilder         *fb,
 283                           GVariant            *options,
 284                           struct gvdb_pointer *pointer)
 285 {
 286   GVariant *normal;
 287   gpointer data;
 288   gsize size;
 289
 290   if (options)
 291     {
 292       if (fb->byteswap)
 293         {
 294           options = g_variant_byteswap (options);
 295           normal = g_variant_get_normal_form (options);
 296           g_variant_unref (options);
 297         }
 298       else
 299         normal = g_variant_get_normal_form (options);
 300
 301       size = g_variant_get_size (normal);
 302       data = file_builder_allocate (fb, 8, size, pointer);
 303       g_variant_store (normal, data);
 304       g_variant_unref (normal);
 305     }
 306 }
 307
 308 static void
 309 file_builder_add_string (FileBuilder *fb,
 310                          const gchar *string,
 311                          guint32_le  *start,
 312                          guint16_le  *size)
 313 {
 314   FileChunk *chunk;
 315   gsize length;
 316
 317   length = strlen (string);
 318
 319   chunk = g_slice_new (FileChunk);
 320   chunk->offset = fb->offset;
 321   chunk->size = length;
 322   chunk->data = g_malloc (length);
 323   memcpy (chunk->data, string, length);
 324
 325   *start = guint32_to_le (fb->offset);
 326   *size = guint16_to_le (length);
 327   fb->offset += length;
 328
 329   g_queue_push_tail (fb->chunks, chunk);
 330 }
 331
 332 static void
 333 file_builder_allocate_for_hash (FileBuilder            *fb,
 334                                 gsize                   n_buckets,
 335                                 gsize                   n_items,
 336                                 guint                   bloom_shift,
 337                                 gsize                   n_bloom_words,
 338                                 guint32_le            **bloom_filter,
 339                                 guint32_le            **hash_buckets,
 340                                 struct gvdb_hash_item **hash_items,
 341                                 struct gvdb_pointer    *pointer)
 342 {
 343   guint32_le bloom_hdr, table_hdr;
 344   guchar *data;
 345   gsize size;
 346
 347   g_assert (n_bloom_words < (1u << 27));
 348
 349   bloom_hdr = guint32_to_le (bloom_shift << 27 | n_bloom_words);
 350   table_hdr = guint32_to_le (n_buckets);
 351
 352   size = sizeof bloom_hdr + sizeof table_hdr +
 353          n_bloom_words * sizeof (guint32_le) +
 354          n_buckets     * sizeof (guint32_le) +
 355          n_items       * sizeof (struct gvdb_hash_item);
 356
 357   data = file_builder_allocate (fb, 4, size, pointer);
 358
 359 #define chunk(s) (size -= (s), data += (s), data - (s))
 360   memcpy (chunk (sizeof bloom_hdr), &bloom_hdr, sizeof bloom_hdr);
 361   memcpy (chunk (sizeof table_hdr), &table_hdr, sizeof table_hdr);
 362   *bloom_filter = (guint32_le *) chunk (n_bloom_words * sizeof (guint32_le));
 363   *hash_buckets = (guint32_le *) chunk (n_buckets * sizeof (guint32_le));
 364   *hash_items = (struct gvdb_hash_item *) chunk (n_items *
 365                   sizeof (struct gvdb_hash_item));
 366   g_assert (size == 0);
 367 #undef chunk
 368
 369   memset (*bloom_filter, 0, n_bloom_words * sizeof (guint32_le));
 370 }
 371
 372 static void
 373 file_builder_add_hash (FileBuilder         *fb,
 374                        GHashTable          *table,
 375                        struct gvdb_pointer *pointer)
 376 {
 377   guint32_le *buckets, *bloom_filter;
 378   struct gvdb_hash_item *items;
 379   HashTable *mytable;
 380   GvdbItem *item;
 381   guint32 index;
 382   gint bucket;
 383
 384   mytable = hash_table_new (g_hash_table_size (table));
 385   g_hash_table_foreach (table, hash_table_insert, mytable);
 386   index = 0;
 387
 388   for (bucket = 0; bucket < mytable->n_buckets; bucket++)
 389     for (item = mytable->buckets[bucket]; item; item = item->next)
 390       item->assigned_index = guint32_to_le (index++);
 391
 392   file_builder_allocate_for_hash (fb, mytable->n_buckets, index, 5, 0,
 393                                   &bloom_filter, &buckets, &items, pointer);
 394
 395   index = 0;
 396   for (bucket = 0; bucket < mytable->n_buckets; bucket++)
 397     {
 398       buckets[bucket] = guint32_to_le (index);
 399
 400       for (item = mytable->buckets[bucket]; item; item = item->next)
 401         {
 402           struct gvdb_hash_item *entry = items++;
 403           const gchar *basename;
 404
 405           g_assert (index == guint32_from_le (item->assigned_index));
 406           entry->hash_value = guint32_to_le (item->hash_value);
 407           entry->parent = item_to_index (item->parent);
 408           entry->unused = 0;
 409
 410           if (item->parent != NULL)
 411             basename = item->key + strlen (item->parent->key);
 412           else
 413             basename = item->key;
 414
 415           file_builder_add_string (fb, basename,
 416                                    &entry->key_start,
 417                                    &entry->key_size);
 418
 419           if (item->value != NULL)
 420             {
 421               g_assert (item->child == NULL && item->table == NULL);
 422
 423               file_builder_add_value (fb, item->value, &entry->value.pointer);
 424               file_builder_add_options (fb, item->options, &entry->options);
 425               entry->type = 'v';
 426             }
 427
 428           if (item->child != NULL)
 429             {
 430               guint32 children = 0;
 431               guint32_le *offsets;
 432               GvdbItem *child;
 433
 434               g_assert (item->table == NULL);
 435
 436               for (child = item->child; child; child = child->sibling)
 437                 children++;
 438
 439               offsets = file_builder_allocate (fb, 4, 4 * children,
 440                                                &entry->value.pointer);
 441               entry->type = 'L';
 442
 443               for (child = item->child; child; child = child->sibling)
 444                 offsets[--children] = child->assigned_index;
 445
 446               g_assert (children == 0);
 447             }
 448
 449           if (item->table != NULL)
 450             {
 451               entry->type = 'H';
 452               file_builder_add_hash (fb, item->table, &entry->value.pointer);
 453             }
 454
 455           index++;
 456         }
 457     }
 458 }
 459
 460 static FileBuilder *
 461 file_builder_new (gboolean byteswap)
 462 {
 463   FileBuilder *builder;
 464
 465   builder = g_slice_new (FileBuilder);
 466   builder->chunks = g_queue_new ();
 467   builder->offset = sizeof (struct gvdb_header);
 468   builder->byteswap = byteswap;
 469
 470   return builder;
 471 }
 472
 473 static GString *
 474 file_builder_serialise (FileBuilder          *fb,
 475                         struct gvdb_pointer   root)
 476 {
 477   struct gvdb_header header;
 478   GString *result;
 479
 480   if (fb->byteswap)
 481     {
 482       header.signature[0] = GVDB_SWAPPED_SIGNATURE0;
 483       header.signature[1] = GVDB_SWAPPED_SIGNATURE1;
 484     }
 485   else
 486     {
 487       header.signature[0] = GVDB_SIGNATURE0;
 488       header.signature[1] = GVDB_SIGNATURE1;
 489     }
 490
 491   result = g_string_new (NULL);
 492
 493   header.root = root;
 494   g_string_append_len (result, (gpointer) &header, sizeof header);
 495
 496   while (!g_queue_is_empty (fb->chunks))
 497     {
 498       FileChunk *chunk = g_queue_pop_head (fb->chunks);
 499
 500       if (result->len != chunk->offset)
 501         {
 502           gchar zero[8] = { 0, };
 503
 504           g_assert (chunk->offset > result->len);
 505           g_assert (chunk->offset - result->len < 8);
 506
 507           g_string_append_len (result, zero, chunk->offset - result->len);
 508           g_assert (result->len == chunk->offset);
 509         }
 510
 511       g_string_append_len (result, chunk->data, chunk->size);
 512       g_free (chunk->data);
 513     }
 514
 515   g_queue_free (fb->chunks);
 516   g_slice_free (FileBuilder, fb);
 517
 518   return result;
 519 }
 520
 521 gboolean
 522 gvdb_table_write_contents (GHashTable   *table,
 523                            const gchar  *filename,
 524                            gboolean      byteswap,
 525                            GError      **error)
 526 {
 527   struct gvdb_pointer root;
 528   gboolean status;
 529   FileBuilder *fb;
 530   GString *str;
 531
 532   fb = file_builder_new (byteswap);
 533   file_builder_add_hash (fb, table, &root);
 534   str = file_builder_serialise (fb, root);
 535
 536   status = g_file_set_contents (filename, str->str, str->len, error);
 537   g_string_free (str, TRUE);
 538
 539   return status;
 540 }