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 pthread_mutex_t lock;
152 /*
154 * This list contains entries for bases which we know the other side
155 * has (e.g., via reachability bitmaps), but which aren't in our
156 * "objects" list.
157 */
158 struct object_entry *ext_bases;
159 uint32_t nr_ext, alloc_ext;
160 uintmax_t oe_size_limit;
162 uintmax_t oe_delta_size_limit;
163 /* delta islands */
165 unsigned int *tree_depth;
166 unsigned char *layer;
167};
168void prepare_packing_data(struct repository *r, struct packing_data *pdata);
170static inline void packing_data_lock(struct packing_data *pdata)
172{
173 pthread_mutex_lock(&pdata->lock);
174}
175static inline void packing_data_unlock(struct packing_data *pdata)
176{
177 pthread_mutex_unlock(&pdata->lock);
178}
179struct object_entry *packlist_alloc(struct packing_data *pdata,
181 const unsigned char *sha1,
182 uint32_t index_pos);
183struct object_entry *packlist_find(struct packing_data *pdata,
185 const unsigned char *sha1,
186 uint32_t *index_pos);
187static inline uint32_t pack_name_hash(const char *name)
189{
190 uint32_t c, hash = 0;
191 if (!name)
193 return 0;
194 /*
196 * This effectively just creates a sortable number from the
197 * last sixteen non-whitespace characters. Last characters
198 * count "most", so things that end in ".c" sort together.
199 */
200 while ((c = *name++) != 0) {
201 if (isspace(c))
202 continue;
203 hash = (hash >> 2) + (c << 24);
204 }
205 return hash;
206}
207static inline enum object_type oe_type(const struct object_entry *e)
209{
210 return e->type_valid ? e->type_ : OBJ_BAD;
211}
212static inline void oe_set_type(struct object_entry *e,
214 enum object_type type)
215{
216 if (type >= OBJ_ANY)
217 BUG("OBJ_ANY cannot be set in pack-objects code");
218 e->type_valid = type >= OBJ_NONE;
220 e->type_ = (unsigned)type;
221}
222static inline unsigned int oe_in_pack_pos(const struct packing_data *pack,
224 const struct object_entry *e)
225{
226 return pack->in_pack_pos[e - pack->objects];
227}
228static inline void oe_set_in_pack_pos(const struct packing_data *pack,
230 const struct object_entry *e,
231 unsigned int pos)
232{
233 pack->in_pack_pos[e - pack->objects] = pos;
234}
235static inline struct packed_git *oe_in_pack(const struct packing_data *pack,
237 const struct object_entry *e)
238{
239 if (pack->in_pack_by_idx)
240 return pack->in_pack_by_idx[e->in_pack_idx];
241 else
242 return pack->in_pack[e - pack->objects];
243}
244void oe_map_new_pack(struct packing_data *pack,
246 struct packed_git *p);
247static inline void oe_set_in_pack(struct packing_data *pack,
248 struct object_entry *e,
249 struct packed_git *p)
250{
251 if (!p->index)
252 oe_map_new_pack(pack, p);
253 if (pack->in_pack_by_idx)
254 e->in_pack_idx = p->index;
255 else
256 pack->in_pack[e - pack->objects] = p;
257}
258static inline struct object_entry *oe_delta(
260 const struct packing_data *pack,
261 const struct object_entry *e)
262{
263 if (!e->delta_idx)
264 return NULL;
265 if (e->ext_base)
266 return &pack->ext_bases[e->delta_idx - 1];
267 else
268 return &pack->objects[e->delta_idx - 1];
269}
270static inline void oe_set_delta(struct packing_data *pack,
272 struct object_entry *e,
273 struct object_entry *delta)
274{
275 if (delta)
276 e->delta_idx = (delta - pack->objects) + 1;
277 else
278 e->delta_idx = 0;
279}
280void oe_set_delta_ext(struct packing_data *pack,
282 struct object_entry *e,
283 const unsigned char *sha1);
284static inline struct object_entry *oe_delta_child(
286 const struct packing_data *pack,
287 const struct object_entry *e)
288{
289 if (e->delta_child_idx)
290 return &pack->objects[e->delta_child_idx - 1];
291 return NULL;
292}
293static inline void oe_set_delta_child(struct packing_data *pack,
295 struct object_entry *e,
296 struct object_entry *delta)
297{
298 if (delta)
299 e->delta_child_idx = (delta - pack->objects) + 1;
300 else
301 e->delta_child_idx = 0;
302}
303static inline struct object_entry *oe_delta_sibling(
305 const struct packing_data *pack,
306 const struct object_entry *e)
307{
308 if (e->delta_sibling_idx)
309 return &pack->objects[e->delta_sibling_idx - 1];
310 return NULL;
311}
312static inline void oe_set_delta_sibling(struct packing_data *pack,
314 struct object_entry *e,
315 struct object_entry *delta)
316{
317 if (delta)
318 e->delta_sibling_idx = (delta - pack->objects) + 1;
319 else
320 e->delta_sibling_idx = 0;
321}
322unsigned long oe_get_size_slow(struct packing_data *pack,
324 const struct object_entry *e);
325static inline unsigned long oe_size(struct packing_data *pack,
326 const struct object_entry *e)
327{
328 if (e->size_valid)
329 return e->size_;
330 return oe_get_size_slow(pack, e);
332}
333static inline int oe_size_less_than(struct packing_data *pack,
335 const struct object_entry *lhs,
336 unsigned long rhs)
337{
338 if (lhs->size_valid)
339 return lhs->size_ < rhs;
340 if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
341 return 0;
342 return oe_get_size_slow(pack, lhs) < rhs;
343}
344static inline int oe_size_greater_than(struct packing_data *pack,
346 const struct object_entry *lhs,
347 unsigned long rhs)
348{
349 if (lhs->size_valid)
350 return lhs->size_ > rhs;
351 if (rhs < pack->oe_size_limit) /* rhs < 2^x <= lhs ? */
352 return 1;
353 return oe_get_size_slow(pack, lhs) > rhs;
354}
355static inline void oe_set_size(struct packing_data *pack,
357 struct object_entry *e,
358 unsigned long size)
359{
360 if (size < pack->oe_size_limit) {
361 e->size_ = size;
362 e->size_valid = 1;
363 } else {
364 e->size_valid = 0;
365 if (oe_get_size_slow(pack, e) != size)
366 BUG("'size' is supposed to be the object size!");
367 }
368}
369static inline unsigned long oe_delta_size(struct packing_data *pack,
371 const struct object_entry *e)
372{
373 if (e->delta_size_valid)
374 return e->delta_size_;
375 /*
377 * pack->delta_size[] can't be NULL because oe_set_delta_size()
378 * must have been called when a new delta is saved with
379 * oe_set_delta().
380 * If oe_delta() returns NULL (i.e. default state, which means
381 * delta_size_valid is also false), then the caller must never
382 * call oe_delta_size().
383 */
384 return pack->delta_size[e - pack->objects];
385}
386static inline void oe_set_delta_size(struct packing_data *pack,
388 struct object_entry *e,
389 unsigned long size)
390{
391 if (size < pack->oe_delta_size_limit) {
392 e->delta_size_ = size;
393 e->delta_size_valid = 1;
394 } else {
395 packing_data_lock(pack);
396 if (!pack->delta_size)
397 ALLOC_ARRAY(pack->delta_size, pack->nr_alloc);
398 packing_data_unlock(pack);
399 pack->delta_size[e - pack->objects] = size;
401 e->delta_size_valid = 0;
402 }
403}
404static inline unsigned int oe_tree_depth(struct packing_data *pack,
406 struct object_entry *e)
407{
408 if (!pack->tree_depth)
409 return 0;
410 return pack->tree_depth[e - pack->objects];
411}
412static inline void oe_set_tree_depth(struct packing_data *pack,
414 struct object_entry *e,
415 unsigned int tree_depth)
416{
417 if (!pack->tree_depth)
418 CALLOC_ARRAY(pack->tree_depth, pack->nr_alloc);
419 pack->tree_depth[e - pack->objects] = tree_depth;
420}
421static inline unsigned char oe_layer(struct packing_data *pack,
423 struct object_entry *e)
424{
425 if (!pack->layer)
426 return 0;
427 return pack->layer[e - pack->objects];
428}
429static inline void oe_set_layer(struct packing_data *pack,
431 struct object_entry *e,
432 unsigned char layer)
433{
434 if (!pack->layer)
435 CALLOC_ARRAY(pack->layer, pack->nr_alloc);
436 pack->layer[e - pack->objects] = layer;
437}
438#endif