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 struct repository;
   9
  10 #define DEFAULT_DELTA_CACHE_SIZE (256 * 1024 * 1024)
  11
  12 #define OE_DFS_STATE_BITS       2
  13 #define OE_DEPTH_BITS           12
  14 #define OE_IN_PACK_BITS         10
  15 #define OE_Z_DELTA_BITS         20
  16 /*
  17  * Note that oe_set_size() becomes expensive when the given size is
  18  * above this limit. Don't lower it too much.
  19  */
  20 #define OE_SIZE_BITS            31
  21 #define OE_DELTA_SIZE_BITS      23
  22
  23 /*
  24  * State flags for depth-first search used for analyzing delta cycles.
  25  *
  26  * The depth is measured in delta-links to the base (so if A is a delta
  27  * against B, then A has a depth of 1, and B a depth of 0).
  28  */
  29 enum dfs_state {
  30         DFS_NONE = 0,
  31         DFS_ACTIVE,
  32         DFS_DONE,
  33         DFS_NUM_STATES
  34 };
  35
  36 /*
  37  * The size of struct nearly determines pack-objects's memory
  38  * consumption. This struct is packed tight for that reason. When you
  39  * add or reorder something in this struct, think a bit about this.
  40  *
  41  * basic object info
  42  * -----------------
  43  * idx.oid is filled up before delta searching starts. idx.crc32 is
  44  * only valid after the object is written out and will be used for
  45  * generating the index. idx.offset will be both gradually set and
  46  * used in writing phase (base objects get offset first, then deltas
  47  * refer to them)
  48  *
  49  * "size" is the uncompressed object size. Compressed size of the raw
  50  * data for an object in a pack is not stored anywhere but is computed
  51  * and made available when reverse .idx is made. Note that when a
  52  * delta is reused, "size" is the uncompressed _delta_ size, not the
  53  * canonical one after the delta has been applied.
  54  *
  55  * "hash" contains a path name hash which is used for sorting the
  56  * delta list and also during delta searching. Once prepare_pack()
  57  * returns it's no longer needed.
  58  *
  59  * source pack info
  60  * ----------------
  61  * The (in_pack, in_pack_offset) tuple contains the location of the
  62  * object in the source pack. in_pack_header_size allows quickly
  63  * skipping the header and going straight to the zlib stream.
  64  *
  65  * "type" and "in_pack_type" both describe object type. in_pack_type
  66  * may contain a delta type, while type is always the canonical type.
  67  *
  68  * deltas
  69  * ------
  70  * Delta links (delta, delta_child and delta_sibling) are created to
  71  * reflect that delta graph from the source pack then updated or added
  72  * during delta searching phase when we find better deltas.
  73  *
  74  * delta_child and delta_sibling are last needed in
  75  * compute_write_order(). "delta" and "delta_size" must remain valid
  76  * at object writing phase in case the delta is not cached.
  77  *
  78  * If a delta is cached in memory and is compressed, delta_data points
  79  * to the data and z_delta_size contains the compressed size. If it's
  80  * uncompressed [1], z_delta_size must be zero. delta_size is always
  81  * the uncompressed size and must be valid even if the delta is not
  82  * cached.
  83  *
  84  * [1] during try_delta phase we don't bother with compressing because
  85  * the delta could be quickly replaced with a better one.
  86  */
  87 struct object_entry {
  88         struct pack_idx_entry idx;
  89         void *delta_data;       /* cached delta (uncompressed) */
  90         off_t in_pack_offset;
  91         uint32_t hash;                  /* name hint hash */
  92         unsigned size_:OE_SIZE_BITS;
  93         unsigned size_valid:1;
  94         uint32_t delta_idx;     /* delta base object */
  95         uint32_t delta_child_idx; /* deltified objects who bases me */
  96         uint32_t delta_sibling_idx; /* other deltified objects who
  97                                      * uses the same base as me
  98                                      */
  99         unsigned delta_size_:OE_DELTA_SIZE_BITS; /* delta data size (uncompressed) */
 100         unsigned delta_size_valid:1;
 101         unsigned char in_pack_header_size;
 102         unsigned in_pack_idx:OE_IN_PACK_BITS;   /* already in pack */
 103         unsigned z_delta_size:OE_Z_DELTA_BITS;
 104         unsigned type_valid:1;
 105         unsigned no_try_delta:1;
 106         unsigned type_:TYPE_BITS;
 107         unsigned in_pack_type:TYPE_BITS; /* could be delta */
 108
 109         unsigned preferred_base:1; /*
 110                                     * we do not pack this, but is available
 111                                     * to be used as the base object to delta
 112                                     * objects against.
 113                                     */
 114         unsigned tagged:1; /* near the very tip of refs */
 115         unsigned filled:1; /* assigned write-order */
 116         unsigned dfs_state:OE_DFS_STATE_BITS;
 117         unsigned depth:OE_DEPTH_BITS;
 118         unsigned ext_base:1; /* delta_idx points outside packlist */
 119
 120         /*
 121          * pahole results on 64-bit linux (gcc and clang)
 122          *
 123          *   size: 80, bit_padding: 9 bits
 124          *
 125          * and on 32-bit (gcc)
 126          *
 127          *   size: 76, bit_padding: 9 bits
 128          */
 129 };
 130
 131 struct packing_data {
 132         struct repository *repo;
 133         struct object_entry *objects;
 134         uint32_t nr_objects, nr_alloc;
 135
 136         int32_t *index;
 137         uint32_t index_size;
 138
 139         unsigned int *in_pack_pos;
 140         unsigned long *delta_size;
 141
 142         /*
 143          * Only one of these can be non-NULL and they have different
 144          * sizes. if in_pack_by_idx is allocated, oe_in_pack() returns
 145          * the pack of an object using in_pack_idx field. If not,
 146          * in_pack[] array is used the same way as in_pack_pos[]
 147          */
 148         struct packed_git **in_pack_by_idx;
 149         struct packed_git **in_pack;
 150
 151         /*
 152          * During packing with multiple threads, protect the in-core
 153          * object database from concurrent accesses.
 154          */
 155         pthread_mutex_t odb_lock;
 156
 157         /*
 158          * This list contains entries for bases which we know the other side
 159          * has (e.g., via reachability bitmaps), but which aren't in our
 160          * "objects" list.
 161          */
 162         struct object_entry *ext_bases;
 163         uint32_t nr_ext, alloc_ext;
 164
 165         uintmax_t oe_size_limit;
 166         uintmax_t oe_delta_size_limit;
 167
 168         /* delta islands */
 169         unsigned int *tree_depth;
 170         unsigned char *layer;
 171 };
 172
 173 void prepare_packing_data(struct repository *r, struct packing_data *pdata);
 174
 175 /* Protect access to object database */
 176 static inline void packing_data_lock(struct packing_data *pdata)
 177 {
 178         pthread_mutex_lock(&pdata->odb_lock);
 179 }
 180 static inline void packing_data_unlock(struct packing_data *pdata)
 181 {
 182         pthread_mutex_unlock(&pdata->odb_lock);
 183 }
 184
 185 struct object_entry *packlist_alloc(struct packing_data *pdata,
 186                                     const unsigned char *sha1,
 187                                     uint32_t index_pos);
 188
 189 struct object_entry *packlist_find(struct packing_data *pdata,
 190                                    const unsigned char *sha1,
 191                                    uint32_t *index_pos);
 192
 193 static inline uint32_t pack_name_hash(const char *name)
 194 {
 195         uint32_t c, hash = 0;
 196
 197         if (!name)
 198                 return 0;
 199
 200         /*
 201          * This effectively just creates a sortable number from the
 202          * last sixteen non-whitespace characters. Last characters
 203          * count "most", so things that end in ".c" sort together.
 204          */
 205         while ((c = *name++) != 0) {
 206                 if (isspace(c))
 207                         continue;
 208                 hash = (hash >> 2) + (c << 24);
 209         }
 210         return hash;
 211 }
 212
 213 static inline enum object_type oe_type(const struct object_entry *e)
 214 {
 215         return e->type_valid ? e->type_ : OBJ_BAD;
 216 }
 217
 218 static inline void oe_set_type(struct object_entry *e,
 219                                enum object_type type)
 220 {
 221         if (type >= OBJ_ANY)
 222                 BUG("OBJ_ANY cannot be set in pack-objects code");
 223
 224         e->type_valid = type >= OBJ_NONE;
 225         e->type_ = (unsigned)type;
 226 }
 227
 228 static inline unsigned int oe_in_pack_pos(const struct packing_data *pack,
 229                                           const struct object_entry *e)
 230 {
 231         return pack->in_pack_pos[e - pack->objects];
 232 }
 233
 234 static inline void oe_set_in_pack_pos(const struct packing_data *pack,
 235                                       const struct object_entry *e,
 236                                       unsigned int pos)
 237 {
 238         pack->in_pack_pos[e - pack->objects] = pos;
 239 }
 240
 241 static inline struct packed_git *oe_in_pack(const struct packing_data *pack,
 242                                             const struct object_entry *e)
 243 {
 244         if (pack->in_pack_by_idx)
 245                 return pack->in_pack_by_idx[e->in_pack_idx];
 246         else
 247                 return pack->in_pack[e - pack->objects];
 248 }
 249
 250 void oe_map_new_pack(struct packing_data *pack,
 251                      struct packed_git *p);
 252 static inline void oe_set_in_pack(struct packing_data *pack,
 253                                   struct object_entry *e,
 254                                   struct packed_git *p)
 255 {
 256         if (!p->index)
 257                 oe_map_new_pack(pack, p);
 258         if (pack->in_pack_by_idx)
 259                 e->in_pack_idx = p->index;
 260         else
 261                 pack->in_pack[e - pack->objects] = p;
 262 }
 263
 264 static inline struct object_entry *oe_delta(
 265                 const struct packing_data *pack,
 266                 const struct object_entry *e)
 267 {
 268         if (!e->delta_idx)
 269                 return NULL;
 270         if (e->ext_base)
 271                 return &pack->ext_bases[e->delta_idx - 1];
 272         else
 273                 return &pack->objects[e->delta_idx - 1];
 274 }
 275
 276 static inline void oe_set_delta(struct packing_data *pack,
 277                                 struct object_entry *e,
 278                                 struct object_entry *delta)
 279 {
 280         if (delta)
 281                 e->delta_idx = (delta - pack->objects) + 1;
 282         else
 283                 e->delta_idx = 0;
 284 }
 285
 286 void oe_set_delta_ext(struct packing_data *pack,
 287                       struct object_entry *e,
 288                       const unsigned char *sha1);
 289
 290 static inline struct object_entry *oe_delta_child(
 291                 const struct packing_data *pack,
 292                 const struct object_entry *e)
 293 {
 294         if (e->delta_child_idx)
 295                 return &pack->objects[e->delta_child_idx - 1];
 296         return NULL;
 297 }
 298
 299 static inline void oe_set_delta_child(struct packing_data *pack,
 300                                       struct object_entry *e,
 301                                       struct object_entry *delta)
 302 {
 303         if (delta)
 304                 e->delta_child_idx = (delta - pack->objects) + 1;
 305         else
 306                 e->delta_child_idx = 0;
 307 }
 308
 309 static inline struct object_entry *oe_delta_sibling(
 310                 const struct packing_data *pack,
 311                 const struct object_entry *e)
 312 {
 313         if (e->delta_sibling_idx)
 314                 return &pack->objects[e->delta_sibling_idx - 1];
 315         return NULL;
 316 }
 317
 318 static inline void oe_set_delta_sibling(struct packing_data *pack,
 319                                         struct object_entry *e,
 320                                         struct object_entry *delta)
 321 {
 322         if (delta)
 323                 e->delta_sibling_idx = (delta - pack->objects) + 1;
 324         else
 325                 e->delta_sibling_idx = 0;
 326 }
 327
 328 unsigned long oe_get_size_slow(struct packing_data *pack,
 329                                const struct object_entry *e);
 330 static inline unsigned long oe_size(struct packing_data *pack,
 331                                     const struct object_entry *e)
 332 {
 333         if (e->size_valid)
 334                 return e->size_;
 335
 336         return oe_get_size_slow(pack, e);
 337 }
 338
 339 static inline int oe_size_less_than(struct packing_data *pack,
 340                                     const struct object_entry *lhs,
 341                                     unsigned long rhs)
 342 {
 343         if (lhs->size_valid)
 344                 return lhs->size_ < rhs;
 345         if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
 346                 return 0;
 347         return oe_get_size_slow(pack, lhs) < rhs;
 348 }
 349
 350 static inline int oe_size_greater_than(struct packing_data *pack,
 351                                        const struct object_entry *lhs,
 352                                        unsigned long rhs)
 353 {
 354         if (lhs->size_valid)
 355                 return lhs->size_ > rhs;
 356         if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
 357                 return 1;
 358         return oe_get_size_slow(pack, lhs) > rhs;
 359 }
 360
 361 static inline void oe_set_size(struct packing_data *pack,
 362                                struct object_entry *e,
 363                                unsigned long size)
 364 {
 365         if (size < pack->oe_size_limit) {
 366                 e->size_ = size;
 367                 e->size_valid = 1;
 368         } else {
 369                 e->size_valid = 0;
 370                 if (oe_get_size_slow(pack, e) != size)
 371                         BUG("'size' is supposed to be the object size!");
 372         }
 373 }
 374
 375 static inline unsigned long oe_delta_size(struct packing_data *pack,
 376                                           const struct object_entry *e)
 377 {
 378         if (e->delta_size_valid)
 379                 return e->delta_size_;
 380
 381         /*
 382          * pack->delta_size[] can't be NULL because oe_set_delta_size()
 383          * must have been called when a new delta is saved with
 384          * oe_set_delta().
 385          * If oe_delta() returns NULL (i.e. default state, which means
 386          * delta_size_valid is also false), then the caller must never
 387          * call oe_delta_size().
 388          */
 389         return pack->delta_size[e - pack->objects];
 390 }
 391
 392 static inline void oe_set_delta_size(struct packing_data *pack,
 393                                      struct object_entry *e,
 394                                      unsigned long size)
 395 {
 396         if (size < pack->oe_delta_size_limit) {
 397                 e->delta_size_ = size;
 398                 e->delta_size_valid = 1;
 399         } else {
 400                 packing_data_lock(pack);
 401                 if (!pack->delta_size)
 402                         ALLOC_ARRAY(pack->delta_size, pack->nr_alloc);
 403                 packing_data_unlock(pack);
 404
 405                 pack->delta_size[e - pack->objects] = size;
 406                 e->delta_size_valid = 0;
 407         }
 408 }
 409
 410 static inline unsigned int oe_tree_depth(struct packing_data *pack,
 411                                          struct object_entry *e)
 412 {
 413         if (!pack->tree_depth)
 414                 return 0;
 415         return pack->tree_depth[e - pack->objects];
 416 }
 417
 418 static inline void oe_set_tree_depth(struct packing_data *pack,
 419                                      struct object_entry *e,
 420                                      unsigned int tree_depth)
 421 {
 422         if (!pack->tree_depth)
 423                 CALLOC_ARRAY(pack->tree_depth, pack->nr_alloc);
 424         pack->tree_depth[e - pack->objects] = tree_depth;
 425 }
 426
 427 static inline unsigned char oe_layer(struct packing_data *pack,
 428                                      struct object_entry *e)
 429 {
 430         if (!pack->layer)
 431                 return 0;
 432         return pack->layer[e - pack->objects];
 433 }
 434
 435 static inline void oe_set_layer(struct packing_data *pack,
 436                                 struct object_entry *e,
 437                                 unsigned char layer)
 438 {
 439         if (!pack->layer)
 440                 CALLOC_ARRAY(pack->layer, pack->nr_alloc);
 441         pack->layer[e - pack->objects] = layer;
 442 }
 443
 444 #endif