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         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
 116         /*
 117          * pahole results on 64-bit linux (gcc and clang)
 118          *
 119          *   size: 80, bit_padding: 9 bits
 120          *
 121          * and on 32-bit (gcc)
 122          *
 123          *   size: 76, bit_padding: 9 bits
 124          */
 125 };
 126
 127 struct 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         unsigned long *delta_size;
 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 #ifndef NO_PTHREADS
 147         pthread_mutex_t lock;
 148 #endif
 149
 150         uintmax_t oe_size_limit;
 151         uintmax_t oe_delta_size_limit;
 152 };
 153
 154 void prepare_packing_data(struct packing_data *pdata);
 155
 156 static inline void packing_data_lock(struct packing_data *pdata)
 157 {
 158 #ifndef NO_PTHREADS
 159         pthread_mutex_lock(&pdata->lock);
 160 #endif
 161 }
 162 static inline void packing_data_unlock(struct packing_data *pdata)
 163 {
 164 #ifndef NO_PTHREADS
 165         pthread_mutex_unlock(&pdata->lock);
 166 #endif
 167 }
 168
 169 struct object_entry *packlist_alloc(struct packing_data *pdata,
 170                                     const unsigned char *sha1,
 171                                     uint32_t index_pos);
 172
 173 struct object_entry *packlist_find(struct packing_data *pdata,
 174                                    const unsigned char *sha1,
 175                                    uint32_t *index_pos);
 176
 177 static inline uint32_t pack_name_hash(const char *name)
 178 {
 179         uint32_t c, hash = 0;
 180
 181         if (!name)
 182                 return 0;
 183
 184         /*
 185          * This effectively just creates a sortable number from the
 186          * last sixteen non-whitespace characters. Last characters
 187          * count "most", so things that end in ".c" sort together.
 188          */
 189         while ((c = *name++) != 0) {
 190                 if (isspace(c))
 191                         continue;
 192                 hash = (hash >> 2) + (c << 24);
 193         }
 194         return hash;
 195 }
 196
 197 static inline enum object_type oe_type(const struct object_entry *e)
 198 {
 199         return e->type_valid ? e->type_ : OBJ_BAD;
 200 }
 201
 202 static inline void oe_set_type(struct object_entry *e,
 203                                enum object_type type)
 204 {
 205         if (type >= OBJ_ANY)
 206                 BUG("OBJ_ANY cannot be set in pack-objects code");
 207
 208         e->type_valid = type >= OBJ_NONE;
 209         e->type_ = (unsigned)type;
 210 }
 211
 212 static inline unsigned int oe_in_pack_pos(const struct packing_data *pack,
 213                                           const struct object_entry *e)
 214 {
 215         return pack->in_pack_pos[e - pack->objects];
 216 }
 217
 218 static inline void oe_set_in_pack_pos(const struct packing_data *pack,
 219                                       const struct object_entry *e,
 220                                       unsigned int pos)
 221 {
 222         pack->in_pack_pos[e - pack->objects] = pos;
 223 }
 224
 225 static inline struct packed_git *oe_in_pack(const struct packing_data *pack,
 226                                             const struct object_entry *e)
 227 {
 228         if (pack->in_pack_by_idx)
 229                 return pack->in_pack_by_idx[e->in_pack_idx];
 230         else
 231                 return pack->in_pack[e - pack->objects];
 232 }
 233
 234 void oe_map_new_pack(struct packing_data *pack,
 235                      struct packed_git *p);
 236 static inline void oe_set_in_pack(struct packing_data *pack,
 237                                   struct object_entry *e,
 238                                   struct packed_git *p)
 239 {
 240         if (!p->index)
 241                 oe_map_new_pack(pack, p);
 242         if (pack->in_pack_by_idx)
 243                 e->in_pack_idx = p->index;
 244         else
 245                 pack->in_pack[e - pack->objects] = p;
 246 }
 247
 248 static inline struct object_entry *oe_delta(
 249                 const struct packing_data *pack,
 250                 const struct object_entry *e)
 251 {
 252         if (e->delta_idx)
 253                 return &pack->objects[e->delta_idx - 1];
 254         return NULL;
 255 }
 256
 257 static inline void oe_set_delta(struct packing_data *pack,
 258                                 struct object_entry *e,
 259                                 struct object_entry *delta)
 260 {
 261         if (delta)
 262                 e->delta_idx = (delta - pack->objects) + 1;
 263         else
 264                 e->delta_idx = 0;
 265 }
 266
 267 static inline struct object_entry *oe_delta_child(
 268                 const struct packing_data *pack,
 269                 const struct object_entry *e)
 270 {
 271         if (e->delta_child_idx)
 272                 return &pack->objects[e->delta_child_idx - 1];
 273         return NULL;
 274 }
 275
 276 static inline void oe_set_delta_child(struct packing_data *pack,
 277                                       struct object_entry *e,
 278                                       struct object_entry *delta)
 279 {
 280         if (delta)
 281                 e->delta_child_idx = (delta - pack->objects) + 1;
 282         else
 283                 e->delta_child_idx = 0;
 284 }
 285
 286 static inline struct object_entry *oe_delta_sibling(
 287                 const struct packing_data *pack,
 288                 const struct object_entry *e)
 289 {
 290         if (e->delta_sibling_idx)
 291                 return &pack->objects[e->delta_sibling_idx - 1];
 292         return NULL;
 293 }
 294
 295 static inline void oe_set_delta_sibling(struct packing_data *pack,
 296                                         struct object_entry *e,
 297                                         struct object_entry *delta)
 298 {
 299         if (delta)
 300                 e->delta_sibling_idx = (delta - pack->objects) + 1;
 301         else
 302                 e->delta_sibling_idx = 0;
 303 }
 304
 305 unsigned long oe_get_size_slow(struct packing_data *pack,
 306                                const struct object_entry *e);
 307 static inline unsigned long oe_size(struct packing_data *pack,
 308                                     const struct object_entry *e)
 309 {
 310         if (e->size_valid)
 311                 return e->size_;
 312
 313         return oe_get_size_slow(pack, e);
 314 }
 315
 316 static inline int oe_size_less_than(struct packing_data *pack,
 317                                     const struct object_entry *lhs,
 318                                     unsigned long rhs)
 319 {
 320         if (lhs->size_valid)
 321                 return lhs->size_ < rhs;
 322         if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
 323                 return 0;
 324         return oe_get_size_slow(pack, lhs) < rhs;
 325 }
 326
 327 static inline int oe_size_greater_than(struct packing_data *pack,
 328                                        const struct object_entry *lhs,
 329                                        unsigned long rhs)
 330 {
 331         if (lhs->size_valid)
 332                 return lhs->size_ > rhs;
 333         if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
 334                 return 1;
 335         return oe_get_size_slow(pack, lhs) > rhs;
 336 }
 337
 338 static inline void oe_set_size(struct packing_data *pack,
 339                                struct object_entry *e,
 340                                unsigned long size)
 341 {
 342         if (size < pack->oe_size_limit) {
 343                 e->size_ = size;
 344                 e->size_valid = 1;
 345         } else {
 346                 e->size_valid = 0;
 347                 if (oe_get_size_slow(pack, e) != size)
 348                         BUG("'size' is supposed to be the object size!");
 349         }
 350 }
 351
 352 static inline unsigned long oe_delta_size(struct packing_data *pack,
 353                                           const struct object_entry *e)
 354 {
 355         if (e->delta_size_valid)
 356                 return e->delta_size_;
 357
 358         /*
 359          * pack->detla_size[] can't be NULL because oe_set_delta_size()
 360          * must have been called when a new delta is saved with
 361          * oe_set_delta().
 362          * If oe_delta() returns NULL (i.e. default state, which means
 363          * delta_size_valid is also false), then the caller must never
 364          * call oe_delta_size().
 365          */
 366         return pack->delta_size[e - pack->objects];
 367 }
 368
 369 static inline void oe_set_delta_size(struct packing_data *pack,
 370                                      struct object_entry *e,
 371                                      unsigned long size)
 372 {
 373         if (size < pack->oe_delta_size_limit) {
 374                 e->delta_size_ = size;
 375                 e->delta_size_valid = 1;
 376         } else {
 377                 packing_data_lock(pack);
 378                 if (!pack->delta_size)
 379                         ALLOC_ARRAY(pack->delta_size, pack->nr_alloc);
 380                 packing_data_unlock(pack);
 381
 382                 pack->delta_size[e - pack->objects] = size;
 383                 e->delta_size_valid = 0;
 384         }
 385 }
 386
 387 #endif