The first batch post 2.24 cycle
[git.git] / pack-objects.h
blob6fe6ae5ee86ccaa6ffe619ab5c4280a76841f340
1 #ifndef PACK_OBJECTS_H
2 #define PACK_OBJECTS_H
4 #include "object-store.h"
5 #include "thread-utils.h"
6 #include "pack.h"
8 struct repository;
10 #define DEFAULT_DELTA_CACHE_SIZE (256 * 1024 * 1024)
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
17 * Note that oe_set_size() becomes expensive when the given size is
18 * above this limit. Don't lower it too much.
20 #define OE_SIZE_BITS 31
21 #define OE_DELTA_SIZE_BITS 23
24 * State flags for depth-first search used for analyzing delta cycles.
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).
29 enum dfs_state {
30 DFS_NONE = 0,
31 DFS_ACTIVE,
32 DFS_DONE,
33 DFS_NUM_STATES
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.
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)
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.
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.
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.
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.
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.
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.
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.
84 * [1] during try_delta phase we don't bother with compressing because
85 * the delta could be quickly replaced with a better one.
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
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 */
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.
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 */
121 * pahole results on 64-bit linux (gcc and clang)
123 * size: 80, bit_padding: 9 bits
125 * and on 32-bit (gcc)
127 * size: 76, bit_padding: 9 bits
131 struct packing_data {
132 struct repository *repo;
133 struct object_entry *objects;
134 uint32_t nr_objects, nr_alloc;
136 int32_t *index;
137 uint32_t index_size;
139 unsigned int *in_pack_pos;
140 unsigned long *delta_size;
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[]
148 struct packed_git **in_pack_by_idx;
149 struct packed_git **in_pack;
152 * During packing with multiple threads, protect the in-core
153 * object database from concurrent accesses.
155 pthread_mutex_t odb_lock;
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.
162 struct object_entry *ext_bases;
163 uint32_t nr_ext, alloc_ext;
165 uintmax_t oe_size_limit;
166 uintmax_t oe_delta_size_limit;
168 /* delta islands */
169 unsigned int *tree_depth;
170 unsigned char *layer;
173 void prepare_packing_data(struct repository *r, struct packing_data *pdata);
175 /* Protect access to object database */
176 static inline void packing_data_lock(struct packing_data *pdata)
178 pthread_mutex_lock(&pdata->odb_lock);
180 static inline void packing_data_unlock(struct packing_data *pdata)
182 pthread_mutex_unlock(&pdata->odb_lock);
185 struct object_entry *packlist_alloc(struct packing_data *pdata,
186 const struct object_id *oid);
188 struct object_entry *packlist_find(struct packing_data *pdata,
189 const struct object_id *oid);
191 static inline uint32_t pack_name_hash(const char *name)
193 uint32_t c, hash = 0;
195 if (!name)
196 return 0;
199 * This effectively just creates a sortable number from the
200 * last sixteen non-whitespace characters. Last characters
201 * count "most", so things that end in ".c" sort together.
203 while ((c = *name++) != 0) {
204 if (isspace(c))
205 continue;
206 hash = (hash >> 2) + (c << 24);
208 return hash;
211 static inline enum object_type oe_type(const struct object_entry *e)
213 return e->type_valid ? e->type_ : OBJ_BAD;
216 static inline void oe_set_type(struct object_entry *e,
217 enum object_type type)
219 if (type >= OBJ_ANY)
220 BUG("OBJ_ANY cannot be set in pack-objects code");
222 e->type_valid = type >= OBJ_NONE;
223 e->type_ = (unsigned)type;
226 static inline unsigned int oe_in_pack_pos(const struct packing_data *pack,
227 const struct object_entry *e)
229 return pack->in_pack_pos[e - pack->objects];
232 static inline void oe_set_in_pack_pos(const struct packing_data *pack,
233 const struct object_entry *e,
234 unsigned int pos)
236 pack->in_pack_pos[e - pack->objects] = pos;
239 static inline struct packed_git *oe_in_pack(const struct packing_data *pack,
240 const struct object_entry *e)
242 if (pack->in_pack_by_idx)
243 return pack->in_pack_by_idx[e->in_pack_idx];
244 else
245 return pack->in_pack[e - pack->objects];
248 void oe_map_new_pack(struct packing_data *pack);
250 static inline void oe_set_in_pack(struct packing_data *pack,
251 struct object_entry *e,
252 struct packed_git *p)
254 if (!p->index)
255 oe_map_new_pack(pack);
256 if (pack->in_pack_by_idx)
257 e->in_pack_idx = p->index;
258 else
259 pack->in_pack[e - pack->objects] = p;
262 static inline struct object_entry *oe_delta(
263 const struct packing_data *pack,
264 const struct object_entry *e)
266 if (!e->delta_idx)
267 return NULL;
268 if (e->ext_base)
269 return &pack->ext_bases[e->delta_idx - 1];
270 else
271 return &pack->objects[e->delta_idx - 1];
274 static inline void oe_set_delta(struct packing_data *pack,
275 struct object_entry *e,
276 struct object_entry *delta)
278 if (delta)
279 e->delta_idx = (delta - pack->objects) + 1;
280 else
281 e->delta_idx = 0;
284 void oe_set_delta_ext(struct packing_data *pack,
285 struct object_entry *e,
286 const unsigned char *sha1);
288 static inline struct object_entry *oe_delta_child(
289 const struct packing_data *pack,
290 const struct object_entry *e)
292 if (e->delta_child_idx)
293 return &pack->objects[e->delta_child_idx - 1];
294 return NULL;
297 static inline void oe_set_delta_child(struct packing_data *pack,
298 struct object_entry *e,
299 struct object_entry *delta)
301 if (delta)
302 e->delta_child_idx = (delta - pack->objects) + 1;
303 else
304 e->delta_child_idx = 0;
307 static inline struct object_entry *oe_delta_sibling(
308 const struct packing_data *pack,
309 const struct object_entry *e)
311 if (e->delta_sibling_idx)
312 return &pack->objects[e->delta_sibling_idx - 1];
313 return NULL;
316 static inline void oe_set_delta_sibling(struct packing_data *pack,
317 struct object_entry *e,
318 struct object_entry *delta)
320 if (delta)
321 e->delta_sibling_idx = (delta - pack->objects) + 1;
322 else
323 e->delta_sibling_idx = 0;
326 unsigned long oe_get_size_slow(struct packing_data *pack,
327 const struct object_entry *e);
328 static inline unsigned long oe_size(struct packing_data *pack,
329 const struct object_entry *e)
331 if (e->size_valid)
332 return e->size_;
334 return oe_get_size_slow(pack, e);
337 static inline int oe_size_less_than(struct packing_data *pack,
338 const struct object_entry *lhs,
339 unsigned long rhs)
341 if (lhs->size_valid)
342 return lhs->size_ < rhs;
343 if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
344 return 0;
345 return oe_get_size_slow(pack, lhs) < rhs;
348 static inline int oe_size_greater_than(struct packing_data *pack,
349 const struct object_entry *lhs,
350 unsigned long rhs)
352 if (lhs->size_valid)
353 return lhs->size_ > rhs;
354 if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
355 return 1;
356 return oe_get_size_slow(pack, lhs) > rhs;
359 static inline void oe_set_size(struct packing_data *pack,
360 struct object_entry *e,
361 unsigned long size)
363 if (size < pack->oe_size_limit) {
364 e->size_ = size;
365 e->size_valid = 1;
366 } else {
367 e->size_valid = 0;
368 if (oe_get_size_slow(pack, e) != size)
369 BUG("'size' is supposed to be the object size!");
373 static inline unsigned long oe_delta_size(struct packing_data *pack,
374 const struct object_entry *e)
376 if (e->delta_size_valid)
377 return e->delta_size_;
380 * pack->delta_size[] can't be NULL because oe_set_delta_size()
381 * must have been called when a new delta is saved with
382 * oe_set_delta().
383 * If oe_delta() returns NULL (i.e. default state, which means
384 * delta_size_valid is also false), then the caller must never
385 * call oe_delta_size().
387 return pack->delta_size[e - pack->objects];
390 static inline void oe_set_delta_size(struct packing_data *pack,
391 struct object_entry *e,
392 unsigned long size)
394 if (size < pack->oe_delta_size_limit) {
395 e->delta_size_ = size;
396 e->delta_size_valid = 1;
397 } else {
398 packing_data_lock(pack);
399 if (!pack->delta_size)
400 ALLOC_ARRAY(pack->delta_size, pack->nr_alloc);
401 packing_data_unlock(pack);
403 pack->delta_size[e - pack->objects] = size;
404 e->delta_size_valid = 0;
408 static inline unsigned int oe_tree_depth(struct packing_data *pack,
409 struct object_entry *e)
411 if (!pack->tree_depth)
412 return 0;
413 return pack->tree_depth[e - pack->objects];
416 static inline void oe_set_tree_depth(struct packing_data *pack,
417 struct object_entry *e,
418 unsigned int tree_depth)
420 if (!pack->tree_depth)
421 CALLOC_ARRAY(pack->tree_depth, pack->nr_alloc);
422 pack->tree_depth[e - pack->objects] = tree_depth;
425 static inline unsigned char oe_layer(struct packing_data *pack,
426 struct object_entry *e)
428 if (!pack->layer)
429 return 0;
430 return pack->layer[e - pack->objects];
433 static inline void oe_set_layer(struct packing_data *pack,
434 struct object_entry *e,
435 unsigned char layer)
437 if (!pack->layer)
438 CALLOC_ARRAY(pack->layer, pack->nr_alloc);
439 pack->layer[e - pack->objects] = layer;
442 #endif