pack-objects.hon commit Merge branch 'sb/config-write-fix' (2a2c18f)
   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
 115        /*
 116         * pahole results on 64-bit linux (gcc and clang)
 117         *
 118         *   size: 80, bit_padding: 20 bits, holes: 8 bits
 119         *
 120         * and on 32-bit (gcc)
 121         *
 122         *   size: 76, bit_padding: 20 bits, holes: 8 bits
 123         */
 124};
 125
 126struct packing_data {
 127        struct object_entry *objects;
 128        uint32_t nr_objects, nr_alloc;
 129
 130        int32_t *index;
 131        uint32_t index_size;
 132
 133        unsigned int *in_pack_pos;
 134
 135        /*
 136         * Only one of these can be non-NULL and they have different
 137         * sizes. if in_pack_by_idx is allocated, oe_in_pack() returns
 138         * the pack of an object using in_pack_idx field. If not,
 139         * in_pack[] array is used the same way as in_pack_pos[]
 140         */
 141        struct packed_git **in_pack_by_idx;
 142        struct packed_git **in_pack;
 143
 144        uintmax_t oe_size_limit;
 145};
 146
 147void prepare_packing_data(struct packing_data *pdata);
 148struct object_entry *packlist_alloc(struct packing_data *pdata,
 149                                    const unsigned char *sha1,
 150                                    uint32_t index_pos);
 151
 152struct object_entry *packlist_find(struct packing_data *pdata,
 153                                   const unsigned char *sha1,
 154                                   uint32_t *index_pos);
 155
 156static inline uint32_t pack_name_hash(const char *name)
 157{
 158        uint32_t c, hash = 0;
 159
 160        if (!name)
 161                return 0;
 162
 163        /*
 164         * This effectively just creates a sortable number from the
 165         * last sixteen non-whitespace characters. Last characters
 166         * count "most", so things that end in ".c" sort together.
 167         */
 168        while ((c = *name++) != 0) {
 169                if (isspace(c))
 170                        continue;
 171                hash = (hash >> 2) + (c << 24);
 172        }
 173        return hash;
 174}
 175
 176static inline enum object_type oe_type(const struct object_entry *e)
 177{
 178        return e->type_valid ? e->type_ : OBJ_BAD;
 179}
 180
 181static inline void oe_set_type(struct object_entry *e,
 182                               enum object_type type)
 183{
 184        if (type >= OBJ_ANY)
 185                BUG("OBJ_ANY cannot be set in pack-objects code");
 186
 187        e->type_valid = type >= OBJ_NONE;
 188        e->type_ = (unsigned)type;
 189}
 190
 191static inline unsigned int oe_in_pack_pos(const struct packing_data *pack,
 192                                          const struct object_entry *e)
 193{
 194        return pack->in_pack_pos[e - pack->objects];
 195}
 196
 197static inline void oe_set_in_pack_pos(const struct packing_data *pack,
 198                                      const struct object_entry *e,
 199                                      unsigned int pos)
 200{
 201        pack->in_pack_pos[e - pack->objects] = pos;
 202}
 203
 204static inline struct packed_git *oe_in_pack(const struct packing_data *pack,
 205                                            const struct object_entry *e)
 206{
 207        if (pack->in_pack_by_idx)
 208                return pack->in_pack_by_idx[e->in_pack_idx];
 209        else
 210                return pack->in_pack[e - pack->objects];
 211}
 212
 213void oe_map_new_pack(struct packing_data *pack,
 214                     struct packed_git *p);
 215static inline void oe_set_in_pack(struct packing_data *pack,
 216                                  struct object_entry *e,
 217                                  struct packed_git *p)
 218{
 219        if (!p->index)
 220                oe_map_new_pack(pack, p);
 221        if (pack->in_pack_by_idx)
 222                e->in_pack_idx = p->index;
 223        else
 224                pack->in_pack[e - pack->objects] = p;
 225}
 226
 227static inline struct object_entry *oe_delta(
 228                const struct packing_data *pack,
 229                const struct object_entry *e)
 230{
 231        if (e->delta_idx)
 232                return &pack->objects[e->delta_idx - 1];
 233        return NULL;
 234}
 235
 236static inline void oe_set_delta(struct packing_data *pack,
 237                                struct object_entry *e,
 238                                struct object_entry *delta)
 239{
 240        if (delta)
 241                e->delta_idx = (delta - pack->objects) + 1;
 242        else
 243                e->delta_idx = 0;
 244}
 245
 246static inline struct object_entry *oe_delta_child(
 247                const struct packing_data *pack,
 248                const struct object_entry *e)
 249{
 250        if (e->delta_child_idx)
 251                return &pack->objects[e->delta_child_idx - 1];
 252        return NULL;
 253}
 254
 255static inline void oe_set_delta_child(struct packing_data *pack,
 256                                      struct object_entry *e,
 257                                      struct object_entry *delta)
 258{
 259        if (delta)
 260                e->delta_child_idx = (delta - pack->objects) + 1;
 261        else
 262                e->delta_child_idx = 0;
 263}
 264
 265static inline struct object_entry *oe_delta_sibling(
 266                const struct packing_data *pack,
 267                const struct object_entry *e)
 268{
 269        if (e->delta_sibling_idx)
 270                return &pack->objects[e->delta_sibling_idx - 1];
 271        return NULL;
 272}
 273
 274static inline void oe_set_delta_sibling(struct packing_data *pack,
 275                                        struct object_entry *e,
 276                                        struct object_entry *delta)
 277{
 278        if (delta)
 279                e->delta_sibling_idx = (delta - pack->objects) + 1;
 280        else
 281                e->delta_sibling_idx = 0;
 282}
 283
 284unsigned long oe_get_size_slow(struct packing_data *pack,
 285                               const struct object_entry *e);
 286static inline unsigned long oe_size(struct packing_data *pack,
 287                                    const struct object_entry *e)
 288{
 289        if (e->size_valid)
 290                return e->size_;
 291
 292        return oe_get_size_slow(pack, e);
 293}
 294
 295static inline int oe_size_less_than(struct packing_data *pack,
 296                                    const struct object_entry *lhs,
 297                                    unsigned long rhs)
 298{
 299        if (lhs->size_valid)
 300                return lhs->size_ < rhs;
 301        if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
 302                return 0;
 303        return oe_get_size_slow(pack, lhs) < rhs;
 304}
 305
 306static inline int oe_size_greater_than(struct packing_data *pack,
 307                                       const struct object_entry *lhs,
 308                                       unsigned long rhs)
 309{
 310        if (lhs->size_valid)
 311                return lhs->size_ > rhs;
 312        if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
 313                return 1;
 314        return oe_get_size_slow(pack, lhs) > rhs;
 315}
 316
 317static inline void oe_set_size(struct packing_data *pack,
 318                               struct object_entry *e,
 319                               unsigned long size)
 320{
 321        if (size < pack->oe_size_limit) {
 322                e->size_ = size;
 323                e->size_valid = 1;
 324        } else {
 325                e->size_valid = 0;
 326                if (oe_get_size_slow(pack, e) != size)
 327                        BUG("'size' is supposed to be the object size!");
 328        }
 329}
 330
 331static inline unsigned long oe_delta_size(struct packing_data *pack,
 332                                          const struct object_entry *e)
 333{
 334        if (e->delta_size_valid)
 335                return e->delta_size_;
 336        return oe_size(pack, e);
 337}
 338
 339static inline void oe_set_delta_size(struct packing_data *pack,
 340                                     struct object_entry *e,
 341                                     unsigned long size)
 342{
 343        e->delta_size_ = size;
 344        e->delta_size_valid = e->delta_size_ == size;
 345        if (!e->delta_size_valid && size != oe_size(pack, e))
 346                BUG("this can only happen in check_object() "
 347                    "where delta size is the same as entry size");
 348}
 349
 350#endif