pack-objects.hon commit pack-objects: reuse on-disk deltas for thin "have" objects (6a1e32d)
   1#ifndef PACK_OBJECTS_H
   2#define PACK_OBJECTS_H
   3
   4#include "object-store.h"
   5#include "pack.h"
   6
   7#define DEFAULT_DELTA_CACHE_SIZE (256 * 1024 * 1024)
   8
   9#define OE_DFS_STATE_BITS       2
  10#define OE_DEPTH_BITS           12
  11#define OE_IN_PACK_BITS         10
  12#define OE_Z_DELTA_BITS         20
  13/*
  14 * Note that oe_set_size() becomes expensive when the given size is
  15 * above this limit. Don't lower it too much.
  16 */
  17#define OE_SIZE_BITS            31
  18#define OE_DELTA_SIZE_BITS      20
  19
  20/*
  21 * State flags for depth-first search used for analyzing delta cycles.
  22 *
  23 * The depth is measured in delta-links to the base (so if A is a delta
  24 * against B, then A has a depth of 1, and B a depth of 0).
  25 */
  26enum dfs_state {
  27        DFS_NONE = 0,
  28        DFS_ACTIVE,
  29        DFS_DONE,
  30        DFS_NUM_STATES
  31};
  32
  33/*
  34 * The size of struct nearly determines pack-objects's memory
  35 * consumption. This struct is packed tight for that reason. When you
  36 * add or reorder something in this struct, think a bit about this.
  37 *
  38 * basic object info
  39 * -----------------
  40 * idx.oid is filled up before delta searching starts. idx.crc32 is
  41 * only valid after the object is written out and will be used for
  42 * generating the index. idx.offset will be both gradually set and
  43 * used in writing phase (base objects get offset first, then deltas
  44 * refer to them)
  45 *
  46 * "size" is the uncompressed object size. Compressed size of the raw
  47 * data for an object in a pack is not stored anywhere but is computed
  48 * and made available when reverse .idx is made. Note that when a
  49 * delta is reused, "size" is the uncompressed _delta_ size, not the
  50 * canonical one after the delta has been applied.
  51 *
  52 * "hash" contains a path name hash which is used for sorting the
  53 * delta list and also during delta searching. Once prepare_pack()
  54 * returns it's no longer needed.
  55 *
  56 * source pack info
  57 * ----------------
  58 * The (in_pack, in_pack_offset) tuple contains the location of the
  59 * object in the source pack. in_pack_header_size allows quickly
  60 * skipping the header and going straight to the zlib stream.
  61 *
  62 * "type" and "in_pack_type" both describe object type. in_pack_type
  63 * may contain a delta type, while type is always the canonical type.
  64 *
  65 * deltas
  66 * ------
  67 * Delta links (delta, delta_child and delta_sibling) are created to
  68 * reflect that delta graph from the source pack then updated or added
  69 * during delta searching phase when we find better deltas.
  70 *
  71 * delta_child and delta_sibling are last needed in
  72 * compute_write_order(). "delta" and "delta_size" must remain valid
  73 * at object writing phase in case the delta is not cached.
  74 *
  75 * If a delta is cached in memory and is compressed, delta_data points
  76 * to the data and z_delta_size contains the compressed size. If it's
  77 * uncompressed [1], z_delta_size must be zero. delta_size is always
  78 * the uncompressed size and must be valid even if the delta is not
  79 * cached.
  80 *
  81 * [1] during try_delta phase we don't bother with compressing because
  82 * the delta could be quickly replaced with a better one.
  83 */
  84struct object_entry {
  85        struct pack_idx_entry idx;
  86        void *delta_data;       /* cached delta (uncompressed) */
  87        off_t in_pack_offset;
  88        uint32_t hash;                  /* name hint hash */
  89        unsigned size_:OE_SIZE_BITS;
  90        unsigned size_valid:1;
  91        uint32_t delta_idx;     /* delta base object */
  92        uint32_t delta_child_idx; /* deltified objects who bases me */
  93        uint32_t delta_sibling_idx; /* other deltified objects who
  94                                     * uses the same base as me
  95                                     */
  96        unsigned delta_size_:OE_DELTA_SIZE_BITS; /* delta data size (uncompressed) */
  97        unsigned delta_size_valid:1;
  98        unsigned in_pack_idx:OE_IN_PACK_BITS;   /* already in pack */
  99        unsigned z_delta_size:OE_Z_DELTA_BITS;
 100        unsigned type_valid:1;
 101        unsigned type_:TYPE_BITS;
 102        unsigned no_try_delta:1;
 103        unsigned in_pack_type:TYPE_BITS; /* could be delta */
 104        unsigned preferred_base:1; /*
 105                                    * we do not pack this, but is available
 106                                    * to be used as the base object to delta
 107                                    * objects against.
 108                                    */
 109        unsigned tagged:1; /* near the very tip of refs */
 110        unsigned filled:1; /* assigned write-order */
 111        unsigned dfs_state:OE_DFS_STATE_BITS;
 112        unsigned char in_pack_header_size;
 113        unsigned depth:OE_DEPTH_BITS;
 114        unsigned ext_base:1; /* delta_idx points outside packlist */
 115
 116        /*
 117         * pahole results on 64-bit linux (gcc and clang)
 118         *
 119         *   size: 80, bit_padding: 20 bits, holes: 8 bits
 120         *
 121         * and on 32-bit (gcc)
 122         *
 123         *   size: 76, bit_padding: 20 bits, holes: 8 bits
 124         */
 125};
 126
 127struct packing_data {
 128        struct object_entry *objects;
 129        uint32_t nr_objects, nr_alloc;
 130
 131        int32_t *index;
 132        uint32_t index_size;
 133
 134        unsigned int *in_pack_pos;
 135
 136        /*
 137         * Only one of these can be non-NULL and they have different
 138         * sizes. if in_pack_by_idx is allocated, oe_in_pack() returns
 139         * the pack of an object using in_pack_idx field. If not,
 140         * in_pack[] array is used the same way as in_pack_pos[]
 141         */
 142        struct packed_git **in_pack_by_idx;
 143        struct packed_git **in_pack;
 144
 145        /*
 146         * This list contains entries for bases which we know the other side
 147         * has (e.g., via reachability bitmaps), but which aren't in our
 148         * "objects" list.
 149         */
 150        struct object_entry *ext_bases;
 151        uint32_t nr_ext, alloc_ext;
 152
 153        uintmax_t oe_size_limit;
 154};
 155
 156void prepare_packing_data(struct packing_data *pdata);
 157struct object_entry *packlist_alloc(struct packing_data *pdata,
 158                                    const unsigned char *sha1,
 159                                    uint32_t index_pos);
 160
 161struct object_entry *packlist_find(struct packing_data *pdata,
 162                                   const unsigned char *sha1,
 163                                   uint32_t *index_pos);
 164
 165static inline uint32_t pack_name_hash(const char *name)
 166{
 167        uint32_t c, hash = 0;
 168
 169        if (!name)
 170                return 0;
 171
 172        /*
 173         * This effectively just creates a sortable number from the
 174         * last sixteen non-whitespace characters. Last characters
 175         * count "most", so things that end in ".c" sort together.
 176         */
 177        while ((c = *name++) != 0) {
 178                if (isspace(c))
 179                        continue;
 180                hash = (hash >> 2) + (c << 24);
 181        }
 182        return hash;
 183}
 184
 185static inline enum object_type oe_type(const struct object_entry *e)
 186{
 187        return e->type_valid ? e->type_ : OBJ_BAD;
 188}
 189
 190static inline void oe_set_type(struct object_entry *e,
 191                               enum object_type type)
 192{
 193        if (type >= OBJ_ANY)
 194                BUG("OBJ_ANY cannot be set in pack-objects code");
 195
 196        e->type_valid = type >= OBJ_NONE;
 197        e->type_ = (unsigned)type;
 198}
 199
 200static inline unsigned int oe_in_pack_pos(const struct packing_data *pack,
 201                                          const struct object_entry *e)
 202{
 203        return pack->in_pack_pos[e - pack->objects];
 204}
 205
 206static inline void oe_set_in_pack_pos(const struct packing_data *pack,
 207                                      const struct object_entry *e,
 208                                      unsigned int pos)
 209{
 210        pack->in_pack_pos[e - pack->objects] = pos;
 211}
 212
 213static inline struct packed_git *oe_in_pack(const struct packing_data *pack,
 214                                            const struct object_entry *e)
 215{
 216        if (pack->in_pack_by_idx)
 217                return pack->in_pack_by_idx[e->in_pack_idx];
 218        else
 219                return pack->in_pack[e - pack->objects];
 220}
 221
 222void oe_map_new_pack(struct packing_data *pack,
 223                     struct packed_git *p);
 224static inline void oe_set_in_pack(struct packing_data *pack,
 225                                  struct object_entry *e,
 226                                  struct packed_git *p)
 227{
 228        if (!p->index)
 229                oe_map_new_pack(pack, p);
 230        if (pack->in_pack_by_idx)
 231                e->in_pack_idx = p->index;
 232        else
 233                pack->in_pack[e - pack->objects] = p;
 234}
 235
 236static inline struct object_entry *oe_delta(
 237                const struct packing_data *pack,
 238                const struct object_entry *e)
 239{
 240        if (!e->delta_idx)
 241                return NULL;
 242        if (e->ext_base)
 243                return &pack->ext_bases[e->delta_idx - 1];
 244        else
 245                return &pack->objects[e->delta_idx - 1];
 246}
 247
 248static inline void oe_set_delta(struct packing_data *pack,
 249                                struct object_entry *e,
 250                                struct object_entry *delta)
 251{
 252        if (delta)
 253                e->delta_idx = (delta - pack->objects) + 1;
 254        else
 255                e->delta_idx = 0;
 256}
 257
 258void oe_set_delta_ext(struct packing_data *pack,
 259                      struct object_entry *e,
 260                      const unsigned char *sha1);
 261
 262static inline struct object_entry *oe_delta_child(
 263                const struct packing_data *pack,
 264                const struct object_entry *e)
 265{
 266        if (e->delta_child_idx)
 267                return &pack->objects[e->delta_child_idx - 1];
 268        return NULL;
 269}
 270
 271static inline void oe_set_delta_child(struct packing_data *pack,
 272                                      struct object_entry *e,
 273                                      struct object_entry *delta)
 274{
 275        if (delta)
 276                e->delta_child_idx = (delta - pack->objects) + 1;
 277        else
 278                e->delta_child_idx = 0;
 279}
 280
 281static inline struct object_entry *oe_delta_sibling(
 282                const struct packing_data *pack,
 283                const struct object_entry *e)
 284{
 285        if (e->delta_sibling_idx)
 286                return &pack->objects[e->delta_sibling_idx - 1];
 287        return NULL;
 288}
 289
 290static inline void oe_set_delta_sibling(struct packing_data *pack,
 291                                        struct object_entry *e,
 292                                        struct object_entry *delta)
 293{
 294        if (delta)
 295                e->delta_sibling_idx = (delta - pack->objects) + 1;
 296        else
 297                e->delta_sibling_idx = 0;
 298}
 299
 300unsigned long oe_get_size_slow(struct packing_data *pack,
 301                               const struct object_entry *e);
 302static inline unsigned long oe_size(struct packing_data *pack,
 303                                    const struct object_entry *e)
 304{
 305        if (e->size_valid)
 306                return e->size_;
 307
 308        return oe_get_size_slow(pack, e);
 309}
 310
 311static inline int oe_size_less_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 0;
 319        return oe_get_size_slow(pack, lhs) < rhs;
 320}
 321
 322static inline int oe_size_greater_than(struct packing_data *pack,
 323                                       const struct object_entry *lhs,
 324                                       unsigned long rhs)
 325{
 326        if (lhs->size_valid)
 327                return lhs->size_ > rhs;
 328        if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
 329                return 1;
 330        return oe_get_size_slow(pack, lhs) > rhs;
 331}
 332
 333static inline void oe_set_size(struct packing_data *pack,
 334                               struct object_entry *e,
 335                               unsigned long size)
 336{
 337        if (size < pack->oe_size_limit) {
 338                e->size_ = size;
 339                e->size_valid = 1;
 340        } else {
 341                e->size_valid = 0;
 342                if (oe_get_size_slow(pack, e) != size)
 343                        BUG("'size' is supposed to be the object size!");
 344        }
 345}
 346
 347static inline unsigned long oe_delta_size(struct packing_data *pack,
 348                                          const struct object_entry *e)
 349{
 350        if (e->delta_size_valid)
 351                return e->delta_size_;
 352        return oe_size(pack, e);
 353}
 354
 355static inline void oe_set_delta_size(struct packing_data *pack,
 356                                     struct object_entry *e,
 357                                     unsigned long size)
 358{
 359        e->delta_size_ = size;
 360        e->delta_size_valid = e->delta_size_ == size;
 361        if (!e->delta_size_valid && size != oe_size(pack, e))
 362                BUG("this can only happen in check_object() "
 363                    "where delta size is the same as entry size");
 364}
 365
 366#endif