pack-objects.hon commit Merge branch 'sb/range-diff-colors' (30035d1)
   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 */
  27enum 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 */
  85struct 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        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 depth:OE_DEPTH_BITS;
 115        unsigned ext_base:1; /* delta_idx points outside packlist */
 116
 117        /*
 118         * pahole results on 64-bit linux (gcc and clang)
 119         *
 120         *   size: 80, bit_padding: 9 bits
 121         *
 122         * and on 32-bit (gcc)
 123         *
 124         *   size: 76, bit_padding: 9 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        unsigned long *delta_size;
 137
 138        /*
 139         * Only one of these can be non-NULL and they have different
 140         * sizes. if in_pack_by_idx is allocated, oe_in_pack() returns
 141         * the pack of an object using in_pack_idx field. If not,
 142         * in_pack[] array is used the same way as in_pack_pos[]
 143         */
 144        struct packed_git **in_pack_by_idx;
 145        struct packed_git **in_pack;
 146
 147#ifndef NO_PTHREADS
 148        pthread_mutex_t lock;
 149#endif
 150
 151        /*
 152         * This list contains entries for bases which we know the other side
 153         * has (e.g., via reachability bitmaps), but which aren't in our
 154         * "objects" list.
 155         */
 156        struct object_entry *ext_bases;
 157        uint32_t nr_ext, alloc_ext;
 158
 159        uintmax_t oe_size_limit;
 160        uintmax_t oe_delta_size_limit;
 161};
 162
 163void prepare_packing_data(struct packing_data *pdata);
 164
 165static inline void packing_data_lock(struct packing_data *pdata)
 166{
 167#ifndef NO_PTHREADS
 168        pthread_mutex_lock(&pdata->lock);
 169#endif
 170}
 171static inline void packing_data_unlock(struct packing_data *pdata)
 172{
 173#ifndef NO_PTHREADS
 174        pthread_mutex_unlock(&pdata->lock);
 175#endif
 176}
 177
 178struct object_entry *packlist_alloc(struct packing_data *pdata,
 179                                    const unsigned char *sha1,
 180                                    uint32_t index_pos);
 181
 182struct object_entry *packlist_find(struct packing_data *pdata,
 183                                   const unsigned char *sha1,
 184                                   uint32_t *index_pos);
 185
 186static inline uint32_t pack_name_hash(const char *name)
 187{
 188        uint32_t c, hash = 0;
 189
 190        if (!name)
 191                return 0;
 192
 193        /*
 194         * This effectively just creates a sortable number from the
 195         * last sixteen non-whitespace characters. Last characters
 196         * count "most", so things that end in ".c" sort together.
 197         */
 198        while ((c = *name++) != 0) {
 199                if (isspace(c))
 200                        continue;
 201                hash = (hash >> 2) + (c << 24);
 202        }
 203        return hash;
 204}
 205
 206static inline enum object_type oe_type(const struct object_entry *e)
 207{
 208        return e->type_valid ? e->type_ : OBJ_BAD;
 209}
 210
 211static inline void oe_set_type(struct object_entry *e,
 212                               enum object_type type)
 213{
 214        if (type >= OBJ_ANY)
 215                BUG("OBJ_ANY cannot be set in pack-objects code");
 216
 217        e->type_valid = type >= OBJ_NONE;
 218        e->type_ = (unsigned)type;
 219}
 220
 221static inline unsigned int oe_in_pack_pos(const struct packing_data *pack,
 222                                          const struct object_entry *e)
 223{
 224        return pack->in_pack_pos[e - pack->objects];
 225}
 226
 227static inline void oe_set_in_pack_pos(const struct packing_data *pack,
 228                                      const struct object_entry *e,
 229                                      unsigned int pos)
 230{
 231        pack->in_pack_pos[e - pack->objects] = pos;
 232}
 233
 234static inline struct packed_git *oe_in_pack(const struct packing_data *pack,
 235                                            const struct object_entry *e)
 236{
 237        if (pack->in_pack_by_idx)
 238                return pack->in_pack_by_idx[e->in_pack_idx];
 239        else
 240                return pack->in_pack[e - pack->objects];
 241}
 242
 243void oe_map_new_pack(struct packing_data *pack,
 244                     struct packed_git *p);
 245static inline void oe_set_in_pack(struct packing_data *pack,
 246                                  struct object_entry *e,
 247                                  struct packed_git *p)
 248{
 249        if (!p->index)
 250                oe_map_new_pack(pack, p);
 251        if (pack->in_pack_by_idx)
 252                e->in_pack_idx = p->index;
 253        else
 254                pack->in_pack[e - pack->objects] = p;
 255}
 256
 257static inline struct object_entry *oe_delta(
 258                const struct packing_data *pack,
 259                const struct object_entry *e)
 260{
 261        if (!e->delta_idx)
 262                return NULL;
 263        if (e->ext_base)
 264                return &pack->ext_bases[e->delta_idx - 1];
 265        else
 266                return &pack->objects[e->delta_idx - 1];
 267}
 268
 269static inline void oe_set_delta(struct packing_data *pack,
 270                                struct object_entry *e,
 271                                struct object_entry *delta)
 272{
 273        if (delta)
 274                e->delta_idx = (delta - pack->objects) + 1;
 275        else
 276                e->delta_idx = 0;
 277}
 278
 279void oe_set_delta_ext(struct packing_data *pack,
 280                      struct object_entry *e,
 281                      const unsigned char *sha1);
 282
 283static inline struct object_entry *oe_delta_child(
 284                const struct packing_data *pack,
 285                const struct object_entry *e)
 286{
 287        if (e->delta_child_idx)
 288                return &pack->objects[e->delta_child_idx - 1];
 289        return NULL;
 290}
 291
 292static inline void oe_set_delta_child(struct packing_data *pack,
 293                                      struct object_entry *e,
 294                                      struct object_entry *delta)
 295{
 296        if (delta)
 297                e->delta_child_idx = (delta - pack->objects) + 1;
 298        else
 299                e->delta_child_idx = 0;
 300}
 301
 302static inline struct object_entry *oe_delta_sibling(
 303                const struct packing_data *pack,
 304                const struct object_entry *e)
 305{
 306        if (e->delta_sibling_idx)
 307                return &pack->objects[e->delta_sibling_idx - 1];
 308        return NULL;
 309}
 310
 311static inline void oe_set_delta_sibling(struct packing_data *pack,
 312                                        struct object_entry *e,
 313                                        struct object_entry *delta)
 314{
 315        if (delta)
 316                e->delta_sibling_idx = (delta - pack->objects) + 1;
 317        else
 318                e->delta_sibling_idx = 0;
 319}
 320
 321unsigned long oe_get_size_slow(struct packing_data *pack,
 322                               const struct object_entry *e);
 323static inline unsigned long oe_size(struct packing_data *pack,
 324                                    const struct object_entry *e)
 325{
 326        if (e->size_valid)
 327                return e->size_;
 328
 329        return oe_get_size_slow(pack, e);
 330}
 331
 332static inline int oe_size_less_than(struct packing_data *pack,
 333                                    const struct object_entry *lhs,
 334                                    unsigned long rhs)
 335{
 336        if (lhs->size_valid)
 337                return lhs->size_ < rhs;
 338        if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
 339                return 0;
 340        return oe_get_size_slow(pack, lhs) < rhs;
 341}
 342
 343static inline int oe_size_greater_than(struct packing_data *pack,
 344                                       const struct object_entry *lhs,
 345                                       unsigned long rhs)
 346{
 347        if (lhs->size_valid)
 348                return lhs->size_ > rhs;
 349        if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
 350                return 1;
 351        return oe_get_size_slow(pack, lhs) > rhs;
 352}
 353
 354static inline void oe_set_size(struct packing_data *pack,
 355                               struct object_entry *e,
 356                               unsigned long size)
 357{
 358        if (size < pack->oe_size_limit) {
 359                e->size_ = size;
 360                e->size_valid = 1;
 361        } else {
 362                e->size_valid = 0;
 363                if (oe_get_size_slow(pack, e) != size)
 364                        BUG("'size' is supposed to be the object size!");
 365        }
 366}
 367
 368static inline unsigned long oe_delta_size(struct packing_data *pack,
 369                                          const struct object_entry *e)
 370{
 371        if (e->delta_size_valid)
 372                return e->delta_size_;
 373
 374        /*
 375         * pack->detla_size[] can't be NULL because oe_set_delta_size()
 376         * must have been called when a new delta is saved with
 377         * oe_set_delta().
 378         * If oe_delta() returns NULL (i.e. default state, which means
 379         * delta_size_valid is also false), then the caller must never
 380         * call oe_delta_size().
 381         */
 382        return pack->delta_size[e - pack->objects];
 383}
 384
 385static inline void oe_set_delta_size(struct packing_data *pack,
 386                                     struct object_entry *e,
 387                                     unsigned long size)
 388{
 389        if (size < pack->oe_delta_size_limit) {
 390                e->delta_size_ = size;
 391                e->delta_size_valid = 1;
 392        } else {
 393                packing_data_lock(pack);
 394                if (!pack->delta_size)
 395                        ALLOC_ARRAY(pack->delta_size, pack->nr_alloc);
 396                packing_data_unlock(pack);
 397
 398                pack->delta_size[e - pack->objects] = size;
 399                e->delta_size_valid = 0;
 400        }
 401}
 402
 403#endif