1#ifndef PACK_OBJECTS_H
2#define PACK_OBJECTS_H
3#include "object-store.h"
5#include "thread-utils.h"
6#include "pack.h"
7struct repository;
9#define DEFAULT_DELTA_CACHE_SIZE (256 * 1024 * 1024)
11#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/*
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 */
29enum dfs_state {
30 DFS_NONE = 0,
31 DFS_ACTIVE,
32 DFS_DONE,
33 DFS_NUM_STATES
34};
35/*
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 */
87struct 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 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 /*
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};
130struct packing_data {
132 struct repository *repo;
133 struct object_entry *objects;
134 uint32_t nr_objects, nr_alloc;
135 int32_t *index;
137 uint32_t index_size;
138 unsigned int *in_pack_pos;
140 unsigned long *delta_size;
141 /*
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#ifndef NO_PTHREADS
152 pthread_mutex_t lock;
153#endif
154 /*
156 * This list contains entries for bases which we know the other side
157 * has (e.g., via reachability bitmaps), but which aren't in our
158 * "objects" list.
159 */
160 struct object_entry *ext_bases;
161 uint32_t nr_ext, alloc_ext;
162 uintmax_t oe_size_limit;
164 uintmax_t oe_delta_size_limit;
165 /* delta islands */
167 unsigned int *tree_depth;
168 unsigned char *layer;
169};
170void prepare_packing_data(struct repository *r, struct packing_data *pdata);
172static inline void packing_data_lock(struct packing_data *pdata)
174{
175#ifndef NO_PTHREADS
176 pthread_mutex_lock(&pdata->lock);
177#endif
178}
179static inline void packing_data_unlock(struct packing_data *pdata)
180{
181#ifndef NO_PTHREADS
182 pthread_mutex_unlock(&pdata->lock);
183#endif
184}
185struct object_entry *packlist_alloc(struct packing_data *pdata,
187 const unsigned char *sha1,
188 uint32_t index_pos);
189struct object_entry *packlist_find(struct packing_data *pdata,
191 const unsigned char *sha1,
192 uint32_t *index_pos);
193static inline uint32_t pack_name_hash(const char *name)
195{
196 uint32_t c, hash = 0;
197 if (!name)
199 return 0;
200 /*
202 * This effectively just creates a sortable number from the
203 * last sixteen non-whitespace characters. Last characters
204 * count "most", so things that end in ".c" sort together.
205 */
206 while ((c = *name++) != 0) {
207 if (isspace(c))
208 continue;
209 hash = (hash >> 2) + (c << 24);
210 }
211 return hash;
212}
213static inline enum object_type oe_type(const struct object_entry *e)
215{
216 return e->type_valid ? e->type_ : OBJ_BAD;
217}
218static inline void oe_set_type(struct object_entry *e,
220 enum object_type type)
221{
222 if (type >= OBJ_ANY)
223 BUG("OBJ_ANY cannot be set in pack-objects code");
224 e->type_valid = type >= OBJ_NONE;
226 e->type_ = (unsigned)type;
227}
228static inline unsigned int oe_in_pack_pos(const struct packing_data *pack,
230 const struct object_entry *e)
231{
232 return pack->in_pack_pos[e - pack->objects];
233}
234static inline void oe_set_in_pack_pos(const struct packing_data *pack,
236 const struct object_entry *e,
237 unsigned int pos)
238{
239 pack->in_pack_pos[e - pack->objects] = pos;
240}
241static inline struct packed_git *oe_in_pack(const struct packing_data *pack,
243 const struct object_entry *e)
244{
245 if (pack->in_pack_by_idx)
246 return pack->in_pack_by_idx[e->in_pack_idx];
247 else
248 return pack->in_pack[e - pack->objects];
249}
250void oe_map_new_pack(struct packing_data *pack,
252 struct packed_git *p);
253static inline void oe_set_in_pack(struct packing_data *pack,
254 struct object_entry *e,
255 struct packed_git *p)
256{
257 if (!p->index)
258 oe_map_new_pack(pack, p);
259 if (pack->in_pack_by_idx)
260 e->in_pack_idx = p->index;
261 else
262 pack->in_pack[e - pack->objects] = p;
263}
264static inline struct object_entry *oe_delta(
266 const struct packing_data *pack,
267 const struct object_entry *e)
268{
269 if (!e->delta_idx)
270 return NULL;
271 if (e->ext_base)
272 return &pack->ext_bases[e->delta_idx - 1];
273 else
274 return &pack->objects[e->delta_idx - 1];
275}
276static inline void oe_set_delta(struct packing_data *pack,
278 struct object_entry *e,
279 struct object_entry *delta)
280{
281 if (delta)
282 e->delta_idx = (delta - pack->objects) + 1;
283 else
284 e->delta_idx = 0;
285}
286void oe_set_delta_ext(struct packing_data *pack,
288 struct object_entry *e,
289 const unsigned char *sha1);
290static inline struct object_entry *oe_delta_child(
292 const struct packing_data *pack,
293 const struct object_entry *e)
294{
295 if (e->delta_child_idx)
296 return &pack->objects[e->delta_child_idx - 1];
297 return NULL;
298}
299static inline void oe_set_delta_child(struct packing_data *pack,
301 struct object_entry *e,
302 struct object_entry *delta)
303{
304 if (delta)
305 e->delta_child_idx = (delta - pack->objects) + 1;
306 else
307 e->delta_child_idx = 0;
308}
309static inline struct object_entry *oe_delta_sibling(
311 const struct packing_data *pack,
312 const struct object_entry *e)
313{
314 if (e->delta_sibling_idx)
315 return &pack->objects[e->delta_sibling_idx - 1];
316 return NULL;
317}
318static inline void oe_set_delta_sibling(struct packing_data *pack,
320 struct object_entry *e,
321 struct object_entry *delta)
322{
323 if (delta)
324 e->delta_sibling_idx = (delta - pack->objects) + 1;
325 else
326 e->delta_sibling_idx = 0;
327}
328unsigned long oe_get_size_slow(struct packing_data *pack,
330 const struct object_entry *e);
331static inline unsigned long oe_size(struct packing_data *pack,
332 const struct object_entry *e)
333{
334 if (e->size_valid)
335 return e->size_;
336 return oe_get_size_slow(pack, e);
338}
339static inline int oe_size_less_than(struct packing_data *pack,
341 const struct object_entry *lhs,
342 unsigned long rhs)
343{
344 if (lhs->size_valid)
345 return lhs->size_ < rhs;
346 if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
347 return 0;
348 return oe_get_size_slow(pack, lhs) < rhs;
349}
350static inline int oe_size_greater_than(struct packing_data *pack,
352 const struct object_entry *lhs,
353 unsigned long rhs)
354{
355 if (lhs->size_valid)
356 return lhs->size_ > rhs;
357 if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
358 return 1;
359 return oe_get_size_slow(pack, lhs) > rhs;
360}
361static inline void oe_set_size(struct packing_data *pack,
363 struct object_entry *e,
364 unsigned long size)
365{
366 if (size < pack->oe_size_limit) {
367 e->size_ = size;
368 e->size_valid = 1;
369 } else {
370 e->size_valid = 0;
371 if (oe_get_size_slow(pack, e) != size)
372 BUG("'size' is supposed to be the object size!");
373 }
374}
375static inline unsigned long oe_delta_size(struct packing_data *pack,
377 const struct object_entry *e)
378{
379 if (e->delta_size_valid)
380 return e->delta_size_;
381 /*
383 * pack->delta_size[] can't be NULL because oe_set_delta_size()
384 * must have been called when a new delta is saved with
385 * oe_set_delta().
386 * If oe_delta() returns NULL (i.e. default state, which means
387 * delta_size_valid is also false), then the caller must never
388 * call oe_delta_size().
389 */
390 return pack->delta_size[e - pack->objects];
391}
392static inline void oe_set_delta_size(struct packing_data *pack,
394 struct object_entry *e,
395 unsigned long size)
396{
397 if (size < pack->oe_delta_size_limit) {
398 e->delta_size_ = size;
399 e->delta_size_valid = 1;
400 } else {
401 packing_data_lock(pack);
402 if (!pack->delta_size)
403 ALLOC_ARRAY(pack->delta_size, pack->nr_alloc);
404 packing_data_unlock(pack);
405 pack->delta_size[e - pack->objects] = size;
407 e->delta_size_valid = 0;
408 }
409}
410static inline unsigned int oe_tree_depth(struct packing_data *pack,
412 struct object_entry *e)
413{
414 if (!pack->tree_depth)
415 return 0;
416 return pack->tree_depth[e - pack->objects];
417}
418static inline void oe_set_tree_depth(struct packing_data *pack,
420 struct object_entry *e,
421 unsigned int tree_depth)
422{
423 if (!pack->tree_depth)
424 ALLOC_ARRAY(pack->tree_depth, pack->nr_objects);
425 pack->tree_depth[e - pack->objects] = tree_depth;
426}
427static inline unsigned char oe_layer(struct packing_data *pack,
429 struct object_entry *e)
430{
431 if (!pack->layer)
432 return 0;
433 return pack->layer[e - pack->objects];
434}
435static inline void oe_set_layer(struct packing_data *pack,
437 struct object_entry *e,
438 unsigned char layer)
439{
440 if (!pack->layer)
441 ALLOC_ARRAY(pack->layer, pack->nr_objects);
442 pack->layer[e - pack->objects] = layer;
443}
444#endif