pack-objects.h

   1 #ifndef PACK_OBJECTS_H
   2 #define PACK_OBJECTS_H
   3
   4 #include "object-store.h"
   5 #include "thread-utils.h"
   6 #include "pack.h"
   7
   8 #define DEFAULT_DELTA_CACHE_SIZE (256 * 1024 * 1024)
   9
  10 #define OE_DFS_STATE_BITS       2
  11 #define OE_DEPTH_BITS           12
  12 #define OE_IN_PACK_BITS         10
  13 #define OE_Z_DELTA_BITS         20
  14 /*
  15  * Note that oe_set_size() becomes expensive when the given size is
  16  * above this limit. Don't lower it too much.
  17  */
  18 #define OE_SIZE_BITS            31
  19 #define OE_DELTA_SIZE_BITS      23
  20
  21 /*
  22  * State flags for depth-first search used for analyzing delta cycles.
  23  *
  24  * The depth is measured in delta-links to the base (so if A is a delta
  25  * against B, then A has a depth of 1, and B a depth of 0).
  26  */
  27 enum dfs_state {
  28         DFS_NONE = 0,
  29         DFS_ACTIVE,
  30         DFS_DONE,
  31         DFS_NUM_STATES
  32 };
  33
  34 /*
  35  * The size of struct nearly determines pack-objects's memory
  36  * consumption. This struct is packed tight for that reason. When you
  37  * add or reorder something in this struct, think a bit about this.
  38  *
  39  * basic object info
  40  * -----------------
  41  * idx.oid is filled up before delta searching starts. idx.crc32 is
  42  * only valid after the object is written out and will be used for
  43  * generating the index. idx.offset will be both gradually set and
  44  * used in writing phase (base objects get offset first, then deltas
  45  * refer to them)
  46  *
  47  * "size" is the uncompressed object size. Compressed size of the raw
  48  * data for an object in a pack is not stored anywhere but is computed
  49  * and made available when reverse .idx is made. Note that when a
  50  * delta is reused, "size" is the uncompressed _delta_ size, not the
  51  * canonical one after the delta has been applied.
  52  *
  53  * "hash" contains a path name hash which is used for sorting the
  54  * delta list and also during delta searching. Once prepare_pack()
  55  * returns it's no longer needed.
  56  *
  57  * source pack info
  58  * ----------------
  59  * The (in_pack, in_pack_offset) tuple contains the location of the
  60  * object in the source pack. in_pack_header_size allows quickly
  61  * skipping the header and going straight to the zlib stream.
  62  *
  63  * "type" and "in_pack_type" both describe object type. in_pack_type
  64  * may contain a delta type, while type is always the canonical type.
  65  *
  66  * deltas
  67  * ------
  68  * Delta links (delta, delta_child and delta_sibling) are created to
  69  * reflect that delta graph from the source pack then updated or added
  70  * during delta searching phase when we find better deltas.
  71  *
  72  * delta_child and delta_sibling are last needed in
  73  * compute_write_order(). "delta" and "delta_size" must remain valid
  74  * at object writing phase in case the delta is not cached.
  75  *
  76  * If a delta is cached in memory and is compressed, delta_data points
  77  * to the data and z_delta_size contains the compressed size. If it's
  78  * uncompressed [1], z_delta_size must be zero. delta_size is always
  79  * the uncompressed size and must be valid even if the delta is not
  80  * cached.
  81  *
  82  * [1] during try_delta phase we don't bother with compressing because
  83  * the delta could be quickly replaced with a better one.
  84  */
  85 struct object_entry {
  86         struct pack_idx_entry idx;
  87         void *delta_data;       /* cached delta (uncompressed) */
  88         off_t in_pack_offset;
  89         uint32_t hash;                  /* name hint hash */
  90         unsigned size_:OE_SIZE_BITS;
  91         unsigned size_valid:1;
  92         uint32_t delta_idx;     /* delta base object */
  93         uint32_t delta_child_idx; /* deltified objects who bases me */
  94         uint32_t delta_sibling_idx; /* other deltified objects who
  95                                      * uses the same base as me
  96                                      */
  97         unsigned delta_size_:OE_DELTA_SIZE_BITS; /* delta data size (uncompressed) */
  98         unsigned delta_size_valid:1;
  99         unsigned char in_pack_header_size;
 100         unsigned in_pack_idx:OE_IN_PACK_BITS;   /* already in pack */
 101         unsigned z_delta_size:OE_Z_DELTA_BITS;
 102         unsigned type_valid:1;
 103         unsigned no_try_delta:1;
 104         unsigned type_:TYPE_BITS;
 105         unsigned in_pack_type:TYPE_BITS; /* could be delta */
 106
 107         unsigned preferred_base:1; /*
 108                                     * we do not pack this, but is available
 109                                     * to be used as the base object to delta
 110                                     * objects against.
 111                                     */
 112         unsigned tagged:1; /* near the very tip of refs */
 113         unsigned filled:1; /* assigned write-order */
 114         unsigned dfs_state:OE_DFS_STATE_BITS;
 115         unsigned depth:OE_DEPTH_BITS;
 116         unsigned ext_base:1; /* delta_idx points outside packlist */
 117
 118         /*
 119          * pahole results on 64-bit linux (gcc and clang)
 120          *
 121          *   size: 80, bit_padding: 9 bits
 122          *
 123          * and on 32-bit (gcc)
 124          *
 125          *   size: 76, bit_padding: 9 bits
 126          */
 127 };
 128
 129 struct packing_data {
 130         struct object_entry *objects;
 131         uint32_t nr_objects, nr_alloc;
 132
 133         int32_t *index;
 134         uint32_t index_size;
 135
 136         unsigned int *in_pack_pos;
 137         unsigned long *delta_size;
 138
 139         /*
 140          * Only one of these can be non-NULL and they have different
 141          * sizes. if in_pack_by_idx is allocated, oe_in_pack() returns
 142          * the pack of an object using in_pack_idx field. If not,
 143          * in_pack[] array is used the same way as in_pack_pos[]
 144          */
 145         struct packed_git **in_pack_by_idx;
 146         struct packed_git **in_pack;
 147
 148 #ifndef NO_PTHREADS
 149         pthread_mutex_t lock;
 150 #endif
 151
 152         /*
 153          * This list contains entries for bases which we know the other side
 154          * has (e.g., via reachability bitmaps), but which aren't in our
 155          * "objects" list.
 156          */
 157         struct object_entry *ext_bases;
 158         uint32_t nr_ext, alloc_ext;
 159
 160         uintmax_t oe_size_limit;
 161         uintmax_t oe_delta_size_limit;
 162
 163         /* delta islands */
 164         unsigned int *tree_depth;
 165         unsigned char *layer;
 166 };
 167
 168 void prepare_packing_data(struct packing_data *pdata);
 169
 170 static inline void packing_data_lock(struct packing_data *pdata)
 171 {
 172 #ifndef NO_PTHREADS
 173         pthread_mutex_lock(&pdata->lock);
 174 #endif
 175 }
 176 static inline void packing_data_unlock(struct packing_data *pdata)
 177 {
 178 #ifndef NO_PTHREADS
 179         pthread_mutex_unlock(&pdata->lock);
 180 #endif
 181 }
 182
 183 struct object_entry *packlist_alloc(struct packing_data *pdata,
 184                                     const unsigned char *sha1,
 185                                     uint32_t index_pos);
 186
 187 struct object_entry *packlist_find(struct packing_data *pdata,
 188                                    const unsigned char *sha1,
 189                                    uint32_t *index_pos);
 190
 191 static inline uint32_t pack_name_hash(const char *name)
 192 {
 193         uint32_t c, hash = 0;
 194
 195         if (!name)
 196                 return 0;
 197
 198         /*
 199          * This effectively just creates a sortable number from the
 200          * last sixteen non-whitespace characters. Last characters
 201          * count "most", so things that end in ".c" sort together.
 202          */
 203         while ((c = *name++) != 0) {
 204                 if (isspace(c))
 205                         continue;
 206                 hash = (hash >> 2) + (c << 24);
 207         }
 208         return hash;
 209 }
 210
 211 static inline enum object_type oe_type(const struct object_entry *e)
 212 {
 213         return e->type_valid ? e->type_ : OBJ_BAD;
 214 }
 215
 216 static inline void oe_set_type(struct object_entry *e,
 217                                enum object_type type)
 218 {
 219         if (type >= OBJ_ANY)
 220                 BUG("OBJ_ANY cannot be set in pack-objects code");
 221
 222         e->type_valid = type >= OBJ_NONE;
 223         e->type_ = (unsigned)type;
 224 }
 225
 226 static inline unsigned int oe_in_pack_pos(const struct packing_data *pack,
 227                                           const struct object_entry *e)
 228 {
 229         return pack->in_pack_pos[e - pack->objects];
 230 }
 231
 232 static inline void oe_set_in_pack_pos(const struct packing_data *pack,
 233                                       const struct object_entry *e,
 234                                       unsigned int pos)
 235 {
 236         pack->in_pack_pos[e - pack->objects] = pos;
 237 }
 238
 239 static inline struct packed_git *oe_in_pack(const struct packing_data *pack,
 240                                             const struct object_entry *e)
 241 {
 242         if (pack->in_pack_by_idx)
 243                 return pack->in_pack_by_idx[e->in_pack_idx];
 244         else
 245                 return pack->in_pack[e - pack->objects];
 246 }
 247
 248 void oe_map_new_pack(struct packing_data *pack,
 249                      struct packed_git *p);
 250 static inline void oe_set_in_pack(struct packing_data *pack,
 251                                   struct object_entry *e,
 252                                   struct packed_git *p)
 253 {
 254         if (!p->index)
 255                 oe_map_new_pack(pack, p);
 256         if (pack->in_pack_by_idx)
 257                 e->in_pack_idx = p->index;
 258         else
 259                 pack->in_pack[e - pack->objects] = p;
 260 }
 261
 262 static inline struct object_entry *oe_delta(
 263                 const struct packing_data *pack,
 264                 const struct object_entry *e)
 265 {
 266         if (!e->delta_idx)
 267                 return NULL;
 268         if (e->ext_base)
 269                 return &pack->ext_bases[e->delta_idx - 1];
 270         else
 271                 return &pack->objects[e->delta_idx - 1];
 272 }
 273
 274 static inline void oe_set_delta(struct packing_data *pack,
 275                                 struct object_entry *e,
 276                                 struct object_entry *delta)
 277 {
 278         if (delta)
 279                 e->delta_idx = (delta - pack->objects) + 1;
 280         else
 281                 e->delta_idx = 0;
 282 }
 283
 284 void oe_set_delta_ext(struct packing_data *pack,
 285                       struct object_entry *e,
 286                       const unsigned char *sha1);
 287
 288 static inline struct object_entry *oe_delta_child(
 289                 const struct packing_data *pack,
 290                 const struct object_entry *e)
 291 {
 292         if (e->delta_child_idx)
 293                 return &pack->objects[e->delta_child_idx - 1];
 294         return NULL;
 295 }
 296
 297 static inline void oe_set_delta_child(struct packing_data *pack,
 298                                       struct object_entry *e,
 299                                       struct object_entry *delta)
 300 {
 301         if (delta)
 302                 e->delta_child_idx = (delta - pack->objects) + 1;
 303         else
 304                 e->delta_child_idx = 0;
 305 }
 306
 307 static inline struct object_entry *oe_delta_sibling(
 308                 const struct packing_data *pack,
 309                 const struct object_entry *e)
 310 {
 311         if (e->delta_sibling_idx)
 312                 return &pack->objects[e->delta_sibling_idx - 1];
 313         return NULL;
 314 }
 315
 316 static inline void oe_set_delta_sibling(struct packing_data *pack,
 317                                         struct object_entry *e,
 318                                         struct object_entry *delta)
 319 {
 320         if (delta)
 321                 e->delta_sibling_idx = (delta - pack->objects) + 1;
 322         else
 323                 e->delta_sibling_idx = 0;
 324 }
 325
 326 unsigned long oe_get_size_slow(struct packing_data *pack,
 327                                const struct object_entry *e);
 328 static inline unsigned long oe_size(struct packing_data *pack,
 329                                     const struct object_entry *e)
 330 {
 331         if (e->size_valid)
 332                 return e->size_;
 333
 334         return oe_get_size_slow(pack, e);
 335 }
 336
 337 static inline int oe_size_less_than(struct packing_data *pack,
 338                                     const struct object_entry *lhs,
 339                                     unsigned long rhs)
 340 {
 341         if (lhs->size_valid)
 342                 return lhs->size_ < rhs;
 343         if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
 344                 return 0;
 345         return oe_get_size_slow(pack, lhs) < rhs;
 346 }
 347
 348 static inline int oe_size_greater_than(struct packing_data *pack,
 349                                        const struct object_entry *lhs,
 350                                        unsigned long rhs)
 351 {
 352         if (lhs->size_valid)
 353                 return lhs->size_ > rhs;
 354         if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
 355                 return 1;
 356         return oe_get_size_slow(pack, lhs) > rhs;
 357 }
 358
 359 static inline void oe_set_size(struct packing_data *pack,
 360                                struct object_entry *e,
 361                                unsigned long size)
 362 {
 363         if (size < pack->oe_size_limit) {
 364                 e->size_ = size;
 365                 e->size_valid = 1;
 366         } else {
 367                 e->size_valid = 0;
 368                 if (oe_get_size_slow(pack, e) != size)
 369                         BUG("'size' is supposed to be the object size!");
 370         }
 371 }
 372
 373 static inline unsigned long oe_delta_size(struct packing_data *pack,
 374                                           const struct object_entry *e)
 375 {
 376         if (e->delta_size_valid)
 377                 return e->delta_size_;
 378
 379         /*
 380          * pack->delta_size[] can't be NULL because oe_set_delta_size()
 381          * must have been called when a new delta is saved with
 382          * oe_set_delta().
 383          * If oe_delta() returns NULL (i.e. default state, which means
 384          * delta_size_valid is also false), then the caller must never
 385          * call oe_delta_size().
 386          */
 387         return pack->delta_size[e - pack->objects];
 388 }
 389
 390 static inline void oe_set_delta_size(struct packing_data *pack,
 391                                      struct object_entry *e,
 392                                      unsigned long size)
 393 {
 394         if (size < pack->oe_delta_size_limit) {
 395                 e->delta_size_ = size;
 396                 e->delta_size_valid = 1;
 397         } else {
 398                 packing_data_lock(pack);
 399                 if (!pack->delta_size)
 400                         ALLOC_ARRAY(pack->delta_size, pack->nr_alloc);
 401                 packing_data_unlock(pack);
 402
 403                 pack->delta_size[e - pack->objects] = size;
 404                 e->delta_size_valid = 0;
 405         }
 406 }
 407
 408 static inline unsigned int oe_tree_depth(struct packing_data *pack,
 409                                          struct object_entry *e)
 410 {
 411         if (!pack->tree_depth)
 412                 return 0;
 413         return pack->tree_depth[e - pack->objects];
 414 }
 415
 416 static inline void oe_set_tree_depth(struct packing_data *pack,
 417                                      struct object_entry *e,
 418                                      unsigned int tree_depth)
 419 {
 420         if (!pack->tree_depth)
 421                 ALLOC_ARRAY(pack->tree_depth, pack->nr_objects);
 422         pack->tree_depth[e - pack->objects] = tree_depth;
 423 }
 424
 425 static inline unsigned char oe_layer(struct packing_data *pack,
 426                                      struct object_entry *e)
 427 {
 428         if (!pack->layer)
 429                 return 0;
 430         return pack->layer[e - pack->objects];
 431 }
 432
 433 static inline void oe_set_layer(struct packing_data *pack,
 434                                 struct object_entry *e,
 435                                 unsigned char layer)
 436 {
 437         if (!pack->layer)
 438                 ALLOC_ARRAY(pack->layer, pack->nr_objects);
 439         pack->layer[e - pack->objects] = layer;
 440 }
 441
 442 #endif