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         /*
 149          * During packing with multiple threads, protect the in-core
 150          * object database from concurrent accesses.
 151          */
 152         pthread_mutex_t odb_lock;
 153
 154         /*
 155          * This list contains entries for bases which we know the other side
 156          * has (e.g., via reachability bitmaps), but which aren't in our
 157          * "objects" list.
 158          */
 159         struct object_entry *ext_bases;
 160         uint32_t nr_ext, alloc_ext;
 161
 162         uintmax_t oe_size_limit;
 163         uintmax_t oe_delta_size_limit;
 164
 165         /* delta islands */
 166         unsigned int *tree_depth;
 167         unsigned char *layer;
 168 };
 169
 170 void prepare_packing_data(struct packing_data *pdata);
 171
 172 /* Protect access to object database */
 173 static inline void packing_data_lock(struct packing_data *pdata)
 174 {
 175         pthread_mutex_lock(&pdata->odb_lock);
 176 }
 177 static inline void packing_data_unlock(struct packing_data *pdata)
 178 {
 179         pthread_mutex_unlock(&pdata->odb_lock);
 180 }
 181
 182 struct object_entry *packlist_alloc(struct packing_data *pdata,
 183                                     const unsigned char *sha1,
 184                                     uint32_t index_pos);
 185
 186 struct object_entry *packlist_find(struct packing_data *pdata,
 187                                    const unsigned char *sha1,
 188                                    uint32_t *index_pos);
 189
 190 static inline uint32_t pack_name_hash(const char *name)
 191 {
 192         uint32_t c, hash = 0;
 193
 194         if (!name)
 195                 return 0;
 196
 197         /*
 198          * This effectively just creates a sortable number from the
 199          * last sixteen non-whitespace characters. Last characters
 200          * count "most", so things that end in ".c" sort together.
 201          */
 202         while ((c = *name++) != 0) {
 203                 if (isspace(c))
 204                         continue;
 205                 hash = (hash >> 2) + (c << 24);
 206         }
 207         return hash;
 208 }
 209
 210 static inline enum object_type oe_type(const struct object_entry *e)
 211 {
 212         return e->type_valid ? e->type_ : OBJ_BAD;
 213 }
 214
 215 static inline void oe_set_type(struct object_entry *e,
 216                                enum object_type type)
 217 {
 218         if (type >= OBJ_ANY)
 219                 BUG("OBJ_ANY cannot be set in pack-objects code");
 220
 221         e->type_valid = type >= OBJ_NONE;
 222         e->type_ = (unsigned)type;
 223 }
 224
 225 static inline unsigned int oe_in_pack_pos(const struct packing_data *pack,
 226                                           const struct object_entry *e)
 227 {
 228         return pack->in_pack_pos[e - pack->objects];
 229 }
 230
 231 static inline void oe_set_in_pack_pos(const struct packing_data *pack,
 232                                       const struct object_entry *e,
 233                                       unsigned int pos)
 234 {
 235         pack->in_pack_pos[e - pack->objects] = pos;
 236 }
 237
 238 static inline struct packed_git *oe_in_pack(const struct packing_data *pack,
 239                                             const struct object_entry *e)
 240 {
 241         if (pack->in_pack_by_idx)
 242                 return pack->in_pack_by_idx[e->in_pack_idx];
 243         else
 244                 return pack->in_pack[e - pack->objects];
 245 }
 246
 247 void oe_map_new_pack(struct packing_data *pack,
 248                      struct packed_git *p);
 249 static inline void oe_set_in_pack(struct packing_data *pack,
 250                                   struct object_entry *e,
 251                                   struct packed_git *p)
 252 {
 253         if (!p->index)
 254                 oe_map_new_pack(pack, p);
 255         if (pack->in_pack_by_idx)
 256                 e->in_pack_idx = p->index;
 257         else
 258                 pack->in_pack[e - pack->objects] = p;
 259 }
 260
 261 static inline struct object_entry *oe_delta(
 262                 const struct packing_data *pack,
 263                 const struct object_entry *e)
 264 {
 265         if (!e->delta_idx)
 266                 return NULL;
 267         if (e->ext_base)
 268                 return &pack->ext_bases[e->delta_idx - 1];
 269         else
 270                 return &pack->objects[e->delta_idx - 1];
 271 }
 272
 273 static inline void oe_set_delta(struct packing_data *pack,
 274                                 struct object_entry *e,
 275                                 struct object_entry *delta)
 276 {
 277         if (delta)
 278                 e->delta_idx = (delta - pack->objects) + 1;
 279         else
 280                 e->delta_idx = 0;
 281 }
 282
 283 void oe_set_delta_ext(struct packing_data *pack,
 284                       struct object_entry *e,
 285                       const unsigned char *sha1);
 286
 287 static inline struct object_entry *oe_delta_child(
 288                 const struct packing_data *pack,
 289                 const struct object_entry *e)
 290 {
 291         if (e->delta_child_idx)
 292                 return &pack->objects[e->delta_child_idx - 1];
 293         return NULL;
 294 }
 295
 296 static inline void oe_set_delta_child(struct packing_data *pack,
 297                                       struct object_entry *e,
 298                                       struct object_entry *delta)
 299 {
 300         if (delta)
 301                 e->delta_child_idx = (delta - pack->objects) + 1;
 302         else
 303                 e->delta_child_idx = 0;
 304 }
 305
 306 static inline struct object_entry *oe_delta_sibling(
 307                 const struct packing_data *pack,
 308                 const struct object_entry *e)
 309 {
 310         if (e->delta_sibling_idx)
 311                 return &pack->objects[e->delta_sibling_idx - 1];
 312         return NULL;
 313 }
 314
 315 static inline void oe_set_delta_sibling(struct packing_data *pack,
 316                                         struct object_entry *e,
 317                                         struct object_entry *delta)
 318 {
 319         if (delta)
 320                 e->delta_sibling_idx = (delta - pack->objects) + 1;
 321         else
 322                 e->delta_sibling_idx = 0;
 323 }
 324
 325 unsigned long oe_get_size_slow(struct packing_data *pack,
 326                                const struct object_entry *e);
 327 static inline unsigned long oe_size(struct packing_data *pack,
 328                                     const struct object_entry *e)
 329 {
 330         if (e->size_valid)
 331                 return e->size_;
 332
 333         return oe_get_size_slow(pack, e);
 334 }
 335
 336 static inline int oe_size_less_than(struct packing_data *pack,
 337                                     const struct object_entry *lhs,
 338                                     unsigned long rhs)
 339 {
 340         if (lhs->size_valid)
 341                 return lhs->size_ < rhs;
 342         if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
 343                 return 0;
 344         return oe_get_size_slow(pack, lhs) < rhs;
 345 }
 346
 347 static inline int oe_size_greater_than(struct packing_data *pack,
 348                                        const struct object_entry *lhs,
 349                                        unsigned long rhs)
 350 {
 351         if (lhs->size_valid)
 352                 return lhs->size_ > rhs;
 353         if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
 354                 return 1;
 355         return oe_get_size_slow(pack, lhs) > rhs;
 356 }
 357
 358 static inline void oe_set_size(struct packing_data *pack,
 359                                struct object_entry *e,
 360                                unsigned long size)
 361 {
 362         if (size < pack->oe_size_limit) {
 363                 e->size_ = size;
 364                 e->size_valid = 1;
 365         } else {
 366                 e->size_valid = 0;
 367                 if (oe_get_size_slow(pack, e) != size)
 368                         BUG("'size' is supposed to be the object size!");
 369         }
 370 }
 371
 372 static inline unsigned long oe_delta_size(struct packing_data *pack,
 373                                           const struct object_entry *e)
 374 {
 375         if (e->delta_size_valid)
 376                 return e->delta_size_;
 377
 378         /*
 379          * pack->delta_size[] can't be NULL because oe_set_delta_size()
 380          * must have been called when a new delta is saved with
 381          * oe_set_delta().
 382          * If oe_delta() returns NULL (i.e. default state, which means
 383          * delta_size_valid is also false), then the caller must never
 384          * call oe_delta_size().
 385          */
 386         return pack->delta_size[e - pack->objects];
 387 }
 388
 389 static inline void oe_set_delta_size(struct packing_data *pack,
 390                                      struct object_entry *e,
 391                                      unsigned long size)
 392 {
 393         if (size < pack->oe_delta_size_limit) {
 394                 e->delta_size_ = size;
 395                 e->delta_size_valid = 1;
 396         } else {
 397                 packing_data_lock(pack);
 398                 if (!pack->delta_size)
 399                         ALLOC_ARRAY(pack->delta_size, pack->nr_alloc);
 400                 packing_data_unlock(pack);
 401
 402                 pack->delta_size[e - pack->objects] = size;
 403                 e->delta_size_valid = 0;
 404         }
 405 }
 406
 407 static inline unsigned int oe_tree_depth(struct packing_data *pack,
 408                                          struct object_entry *e)
 409 {
 410         if (!pack->tree_depth)
 411                 return 0;
 412         return pack->tree_depth[e - pack->objects];
 413 }
 414
 415 static inline void oe_set_tree_depth(struct packing_data *pack,
 416                                      struct object_entry *e,
 417                                      unsigned int tree_depth)
 418 {
 419         if (!pack->tree_depth)
 420                 CALLOC_ARRAY(pack->tree_depth, pack->nr_alloc);
 421         pack->tree_depth[e - pack->objects] = tree_depth;
 422 }
 423
 424 static inline unsigned char oe_layer(struct packing_data *pack,
 425                                      struct object_entry *e)
 426 {
 427         if (!pack->layer)
 428                 return 0;
 429         return pack->layer[e - pack->objects];
 430 }
 431
 432 static inline void oe_set_layer(struct packing_data *pack,
 433                                 struct object_entry *e,
 434                                 unsigned char layer)
 435 {
 436         if (!pack->layer)
 437                 CALLOC_ARRAY(pack->layer, pack->nr_alloc);
 438         pack->layer[e - pack->objects] = layer;
 439 }
 440
 441 #endif