Andrew's git - gitweb.git/blob

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