[thirdparty/git.git] / pack-objects.h

#ifndef PACK_OBJECTS_H
#define PACK_OBJECTS_H

#include "object-store-ll.h"
#include "thread-utils.h"
#include "pack.h"

struct repository;

#define DEFAULT_DELTA_CACHE_SIZE (256 * 1024 * 1024)

#define OE_DFS_STATE_BITS	2
#define OE_DEPTH_BITS		12
#define OE_IN_PACK_BITS		10
#define OE_Z_DELTA_BITS		20
/*
 * Note that oe_set_size() becomes expensive when the given size is
 * above this limit. Don't lower it too much.
 */
#define OE_SIZE_BITS		31
#define OE_DELTA_SIZE_BITS	23

/*
 * State flags for depth-first search used for analyzing delta cycles.
 *
 * The depth is measured in delta-links to the base (so if A is a delta
 * against B, then A has a depth of 1, and B a depth of 0).
 */
enum dfs_state {
	DFS_NONE = 0,
	DFS_ACTIVE,
	DFS_DONE,
	DFS_NUM_STATES
};

/*
 * The size of struct nearly determines pack-objects's memory
 * consumption. This struct is packed tight for that reason. When you
 * add or reorder something in this struct, think a bit about this.
 *
 * basic object info
 * -----------------
 * idx.oid is filled up before delta searching starts. idx.crc32 is
 * only valid after the object is written out and will be used for
 * generating the index. idx.offset will be both gradually set and
 * used in writing phase (base objects get offset first, then deltas
 * refer to them)
 *
 * "size" is the uncompressed object size. Compressed size of the raw
 * data for an object in a pack is not stored anywhere but is computed
 * and made available when reverse .idx is made. Note that when a
 * delta is reused, "size" is the uncompressed _delta_ size, not the
 * canonical one after the delta has been applied.
 *
 * "hash" contains a path name hash which is used for sorting the
 * delta list and also during delta searching. Once prepare_pack()
 * returns it's no longer needed.
 *
 * source pack info
 * ----------------
 * The (in_pack, in_pack_offset) tuple contains the location of the
 * object in the source pack. in_pack_header_size allows quickly
 * skipping the header and going straight to the zlib stream.
 *
 * "type" and "in_pack_type" both describe object type. in_pack_type
 * may contain a delta type, while type is always the canonical type.
 *
 * deltas
 * ------
 * Delta links (delta, delta_child and delta_sibling) are created to
 * reflect that delta graph from the source pack then updated or added
 * during delta searching phase when we find better deltas.
 *
 * delta_child and delta_sibling are last needed in
 * compute_write_order(). "delta" and "delta_size" must remain valid
 * at object writing phase in case the delta is not cached.
 *
 * If a delta is cached in memory and is compressed, delta_data points
 * to the data and z_delta_size contains the compressed size. If it's
 * uncompressed [1], z_delta_size must be zero. delta_size is always
 * the uncompressed size and must be valid even if the delta is not
 * cached.
 *
 * [1] during try_delta phase we don't bother with compressing because
 * the delta could be quickly replaced with a better one.
 */
struct object_entry {
	struct pack_idx_entry idx;
	void *delta_data;	/* cached delta (uncompressed) */
	off_t in_pack_offset;
	uint32_t hash;			/* name hint hash */
	unsigned size_:OE_SIZE_BITS;
	unsigned size_valid:1;
	uint32_t delta_idx;	/* delta base object */
	uint32_t delta_child_idx; /* deltified objects who bases me */
	uint32_t delta_sibling_idx; /* other deltified objects who
				     * uses the same base as me
				     */
	unsigned delta_size_:OE_DELTA_SIZE_BITS; /* delta data size (uncompressed) */
	unsigned delta_size_valid:1;
	unsigned char in_pack_header_size;
	unsigned in_pack_idx:OE_IN_PACK_BITS;	/* already in pack */
	unsigned z_delta_size:OE_Z_DELTA_BITS;
	unsigned type_valid:1;
	unsigned no_try_delta:1;
	unsigned type_:TYPE_BITS;
	unsigned in_pack_type:TYPE_BITS; /* could be delta */

	unsigned preferred_base:1; /*
				    * we do not pack this, but is available
				    * to be used as the base object to delta
				    * objects against.
				    */
	unsigned tagged:1; /* near the very tip of refs */
	unsigned filled:1; /* assigned write-order */
	unsigned dfs_state:OE_DFS_STATE_BITS;
	unsigned depth:OE_DEPTH_BITS;
	unsigned ext_base:1; /* delta_idx points outside packlist */
};

struct packing_data {
	struct repository *repo;
	struct object_entry *objects;
	uint32_t nr_objects, nr_alloc;

	int32_t *index;
	uint32_t index_size;

	unsigned int *in_pack_pos;
	unsigned long *delta_size;

	/*
	 * Only one of these can be non-NULL and they have different
	 * sizes. if in_pack_by_idx is allocated, oe_in_pack() returns
	 * the pack of an object using in_pack_idx field. If not,
	 * in_pack[] array is used the same way as in_pack_pos[]
	 */
	struct packed_git **in_pack_by_idx;
	struct packed_git **in_pack;

	/*
	 * During packing with multiple threads, protect the in-core
	 * object database from concurrent accesses.
	 */
	pthread_mutex_t odb_lock;

	/*
	 * This list contains entries for bases which we know the other side
	 * has (e.g., via reachability bitmaps), but which aren't in our
	 * "objects" list.
	 */
	struct object_entry *ext_bases;
	uint32_t nr_ext, alloc_ext;

	uintmax_t oe_size_limit;
	uintmax_t oe_delta_size_limit;

	/* delta islands */
	unsigned int *tree_depth;
	unsigned char *layer;

	/*
	 * Used when writing cruft packs.
	 *
	 * Object mtimes are stored in pack order when writing, but
	 * written out in lexicographic (index) order.
	 */
	uint32_t *cruft_mtime;
};

void prepare_packing_data(struct repository *r, struct packing_data *pdata);

/* Protect access to object database */
static inline void packing_data_lock(struct packing_data *pdata)
{
	pthread_mutex_lock(&pdata->odb_lock);
}
static inline void packing_data_unlock(struct packing_data *pdata)
{
	pthread_mutex_unlock(&pdata->odb_lock);
}

struct object_entry *packlist_alloc(struct packing_data *pdata,
				    const struct object_id *oid);

struct object_entry *packlist_find(struct packing_data *pdata,
				   const struct object_id *oid);

static inline uint32_t pack_name_hash(const char *name)
{
	uint32_t c, hash = 0;

	if (!name)
		return 0;

	/*
	 * This effectively just creates a sortable number from the
	 * last sixteen non-whitespace characters. Last characters
	 * count "most", so things that end in ".c" sort together.
	 */
	while ((c = *name++) != 0) {
		if (isspace(c))
			continue;
		hash = (hash >> 2) + (c << 24);
	}
	return hash;
}

static inline enum object_type oe_type(const struct object_entry *e)
{
	return e->type_valid ? e->type_ : OBJ_BAD;
}

static inline void oe_set_type(struct object_entry *e,
			       enum object_type type)
{
	if (type >= OBJ_ANY)
		BUG("OBJ_ANY cannot be set in pack-objects code");

	e->type_valid = type >= OBJ_NONE;
	e->type_ = (unsigned)type;
}

static inline unsigned int oe_in_pack_pos(const struct packing_data *pack,
					  const struct object_entry *e)
{
	return pack->in_pack_pos[e - pack->objects];
}

static inline void oe_set_in_pack_pos(const struct packing_data *pack,
				      const struct object_entry *e,
				      unsigned int pos)
{
	pack->in_pack_pos[e - pack->objects] = pos;
}

static inline struct packed_git *oe_in_pack(const struct packing_data *pack,
					    const struct object_entry *e)
{
	if (pack->in_pack_by_idx)
		return pack->in_pack_by_idx[e->in_pack_idx];
	else
		return pack->in_pack[e - pack->objects];
}

void oe_map_new_pack(struct packing_data *pack);

static inline void oe_set_in_pack(struct packing_data *pack,
				  struct object_entry *e,
				  struct packed_git *p)
{
	if (pack->in_pack_by_idx) {
		if (p->index) {
			e->in_pack_idx = p->index;
			return;
		}
		/*
		 * We're accessing packs by index, but this pack doesn't have
		 * an index (e.g., because it was added since we created the
		 * in_pack_by_idx array). Bail to oe_map_new_pack(), which
		 * will convert us to using the full in_pack array, and then
		 * fall through to our in_pack handling.
		 */
		oe_map_new_pack(pack);
	}
	pack->in_pack[e - pack->objects] = p;
}

void oe_set_delta_ext(struct packing_data *pack,
		      struct object_entry *e,
		      const struct object_id *oid);

static inline unsigned int oe_tree_depth(struct packing_data *pack,
					 struct object_entry *e)
{
	if (!pack->tree_depth)
		return 0;
	return pack->tree_depth[e - pack->objects];
}

static inline void oe_set_layer(struct packing_data *pack,
				struct object_entry *e,
				unsigned char layer)
{
	if (!pack->layer)
		CALLOC_ARRAY(pack->layer, pack->nr_alloc);
	pack->layer[e - pack->objects] = layer;
}

static inline uint32_t oe_cruft_mtime(struct packing_data *pack,
				      struct object_entry *e)
{
	if (!pack->cruft_mtime)
		return 0;
	return pack->cruft_mtime[e - pack->objects];
}

static inline void oe_set_cruft_mtime(struct packing_data *pack,
				      struct object_entry *e,
				      uint32_t mtime)
{
	if (!pack->cruft_mtime)
		CALLOC_ARRAY(pack->cruft_mtime, pack->nr_alloc);
	pack->cruft_mtime[e - pack->objects] = mtime;
}

#endif
Commit	Line	Data
2834bc27 VM	1	#ifndef PACK_OBJECTS_H
	2	#define PACK_OBJECTS_H
	3
a034e910	4	#include "object-store-ll.h"
9ac3f0e5	5	#include "thread-utils.h"
ef3ca954	6	#include "pack.h"
43fa44fa	7
7c141127 NTND	8	struct repository;
7c141127 NTND	9
9806f5a7 NTND	10	#define DEFAULT_DELTA_CACHE_SIZE (256 * 1024 * 1024)
9806f5a7 NTND	11
0c6804ab	12	#define OE_DFS_STATE_BITS 2
b5c0cbd8	13	#define OE_DEPTH_BITS 12
43fa44fa	14	#define OE_IN_PACK_BITS 10
0cb3c142	15	#define OE_Z_DELTA_BITS 20
ac77d0c3 NTND	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
9ac3f0e5	21	#define OE_DELTA_SIZE_BITS 23
0c6804ab NTND	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
8d6ccce1	36	/*
3b13a5f2 NTND	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	*
8d6ccce1 NTND	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
27a7d067 NTND	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.
8d6ccce1 NTND	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	*/
2834bc27 VM	87	struct object_entry {
2834bc27 VM	88	struct pack_idx_entry idx;
2834bc27	89	void delta_data; / cached delta (uncompressed) */
3b13a5f2	90	off_t in_pack_offset;
2834bc27	91	uint32_t hash; /* name hint hash */
ac77d0c3 NTND	92	unsigned size_:OE_SIZE_BITS;
ac77d0c3 NTND	93	unsigned size_valid:1;
898eba5e NTND	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	*/
0aca34e8 NTND	99	unsigned delta_size_:OE_DELTA_SIZE_BITS; /* delta data size (uncompressed) */
0aca34e8 NTND	100	unsigned delta_size_valid:1;
9ac3f0e5	101	unsigned char in_pack_header_size;
3b13a5f2	102	unsigned in_pack_idx:OE_IN_PACK_BITS; /* already in pack */
0cb3c142	103	unsigned z_delta_size:OE_Z_DELTA_BITS;
3b13a5f2	104	unsigned type_valid:1;
3b13a5f2	105	unsigned no_try_delta:1;
9ac3f0e5	106	unsigned type_:TYPE_BITS;
fd9b1bae	107	unsigned in_pack_type:TYPE_BITS; /* could be delta */
c8d521fa	108
2834bc27 VM	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	*/
2834bc27 VM	114	unsigned tagged:1; /* near the very tip of refs */
2834bc27 VM	115	unsigned filled:1; /* assigned write-order */
0c6804ab	116	unsigned dfs_state:OE_DFS_STATE_BITS;
b5c0cbd8	117	unsigned depth:OE_DEPTH_BITS;
6a1e32d5	118	unsigned ext_base:1; /* delta_idx points outside packlist */
2834bc27 VM	119	};
	120
	121	struct packing_data {
7c141127	122	struct repository *repo;
2834bc27 VM	123	struct object_entry *objects;
	124	uint32_t nr_objects, nr_alloc;
	125
	126	int32_t *index;
	127	uint32_t index_size;
06af3bba NTND	128
06af3bba NTND	129	unsigned int *in_pack_pos;
9ac3f0e5	130	unsigned long *delta_size;
43fa44fa NTND	131
	132	/*
	133	* Only one of these can be non-NULL and they have different
	134	* sizes. if in_pack_by_idx is allocated, oe_in_pack() returns
	135	* the pack of an object using in_pack_idx field. If not,
	136	* in_pack[] array is used the same way as in_pack_pos[]
	137	*/
	138	struct packed_git **in_pack_by_idx;
	139	struct packed_git **in_pack;
ac77d0c3	140
edb673cf PH	141	/*
	142	* During packing with multiple threads, protect the in-core
	143	* object database from concurrent accesses.
	144	*/
	145	pthread_mutex_t odb_lock;
9ac3f0e5	146
6a1e32d5 JK	147	/*
	148	* This list contains entries for bases which we know the other side
	149	* has (e.g., via reachability bitmaps), but which aren't in our
	150	* "objects" list.
	151	*/
	152	struct object_entry *ext_bases;
	153	uint32_t nr_ext, alloc_ext;
	154
ac77d0c3	155	uintmax_t oe_size_limit;
9ac3f0e5	156	uintmax_t oe_delta_size_limit;
108f5303 CC	157
	158	/* delta islands */
	159	unsigned int *tree_depth;
fe0ac2fb	160	unsigned char *layer;
5dfaf49a TB	161
	162	/*
	163	* Used when writing cruft packs.
	164	*
	165	* Object mtimes are stored in pack order when writing, but
	166	* written out in lexicographic (index) order.
	167	*/
	168	uint32_t *cruft_mtime;
2834bc27 VM	169	};
2834bc27 VM	170
7c141127	171	void prepare_packing_data(struct repository r, struct packing_data pdata);
9ac3f0e5	172
edb673cf	173	/* Protect access to object database */
9ac3f0e5 NTND	174	static inline void packing_data_lock(struct packing_data *pdata)
9ac3f0e5 NTND	175	{
edb673cf	176	pthread_mutex_lock(&pdata->odb_lock);
9ac3f0e5 NTND	177	}
	178	static inline void packing_data_unlock(struct packing_data *pdata)
	179	{
edb673cf	180	pthread_mutex_unlock(&pdata->odb_lock);
9ac3f0e5 NTND	181	}
9ac3f0e5 NTND	182
2834bc27	183	struct object_entry packlist_alloc(struct packing_data pdata,
3a37876b	184	const struct object_id *oid);
2834bc27 VM	185
2834bc27 VM	186	struct object_entry packlist_find(struct packing_data pdata,
3a37876b	187	const struct object_id *oid);
2834bc27	188
68fb36eb VM	189	static inline uint32_t pack_name_hash(const char *name)
	190	{
	191	uint32_t c, hash = 0;
	192
	193	if (!name)
	194	return 0;
	195
	196	/*
	197	* This effectively just creates a sortable number from the
	198	* last sixteen non-whitespace characters. Last characters
	199	* count "most", so things that end in ".c" sort together.
	200	*/
	201	while ((c = *name++) != 0) {
	202	if (isspace(c))
	203	continue;
	204	hash = (hash >> 2) + (c << 24);
	205	}
	206	return hash;
	207	}
	208
fd9b1bae NTND	209	static inline enum object_type oe_type(const struct object_entry *e)
	210	{
	211	return e->type_valid ? e->type_ : OBJ_BAD;
	212	}
	213
	214	static inline void oe_set_type(struct object_entry *e,
	215	enum object_type type)
	216	{
	217	if (type >= OBJ_ANY)
	218	BUG("OBJ_ANY cannot be set in pack-objects code");
	219
	220	e->type_valid = type >= OBJ_NONE;
	221	e->type_ = (unsigned)type;
	222	}
	223
06af3bba NTND	224	static inline unsigned int oe_in_pack_pos(const struct packing_data *pack,
	225	const struct object_entry *e)
	226	{
	227	return pack->in_pack_pos[e - pack->objects];
	228	}
	229
	230	static inline void oe_set_in_pack_pos(const struct packing_data *pack,
	231	const struct object_entry *e,
	232	unsigned int pos)
	233	{
	234	pack->in_pack_pos[e - pack->objects] = pos;
	235	}
	236
43fa44fa NTND	237	static inline struct packed_git oe_in_pack(const struct packing_data pack,
	238	const struct object_entry *e)
	239	{
	240	if (pack->in_pack_by_idx)
	241	return pack->in_pack_by_idx[e->in_pack_idx];
	242	else
	243	return pack->in_pack[e - pack->objects];
	244	}
	245
c409d108 JK	246	void oe_map_new_pack(struct packing_data *pack);
c409d108 JK	247
43fa44fa NTND	248	static inline void oe_set_in_pack(struct packing_data *pack,
	249	struct object_entry *e,
	250	struct packed_git *p)
	251	{
f66e0401 JK	252	if (pack->in_pack_by_idx) {
	253	if (p->index) {
	254	e->in_pack_idx = p->index;
	255	return;
	256	}
	257	/*
	258	* We're accessing packs by index, but this pack doesn't have
	259	* an index (e.g., because it was added since we created the
	260	* in_pack_by_idx array). Bail to oe_map_new_pack(), which
	261	* will convert us to using the full in_pack array, and then
	262	* fall through to our in_pack handling.
	263	*/
c409d108	264	oe_map_new_pack(pack);
f66e0401 JK	265	}
f66e0401 JK	266	pack->in_pack[e - pack->objects] = p;
43fa44fa NTND	267	}
43fa44fa NTND	268
6a1e32d5 JK	269	void oe_set_delta_ext(struct packing_data *pack,
6a1e32d5 JK	270	struct object_entry *e,
a93c141d	271	const struct object_id *oid);
6a1e32d5	272
108f5303 CC	273	static inline unsigned int oe_tree_depth(struct packing_data *pack,
	274	struct object_entry *e)
	275	{
	276	if (!pack->tree_depth)
	277	return 0;
	278	return pack->tree_depth[e - pack->objects];
	279	}
	280
fe0ac2fb CC	281	static inline void oe_set_layer(struct packing_data *pack,
	282	struct object_entry *e,
	283	unsigned char layer)
	284	{
	285	if (!pack->layer)
e159b810	286	CALLOC_ARRAY(pack->layer, pack->nr_alloc);
fe0ac2fb CC	287	pack->layer[e - pack->objects] = layer;
	288	}
	289
5dfaf49a TB	290	static inline uint32_t oe_cruft_mtime(struct packing_data *pack,
	291	struct object_entry *e)
	292	{
	293	if (!pack->cruft_mtime)
	294	return 0;
	295	return pack->cruft_mtime[e - pack->objects];
	296	}
	297
	298	static inline void oe_set_cruft_mtime(struct packing_data *pack,
	299	struct object_entry *e,
	300	uint32_t mtime)
	301	{
	302	if (!pack->cruft_mtime)
	303	CALLOC_ARRAY(pack->cruft_mtime, pack->nr_alloc);
	304	pack->cruft_mtime[e - pack->objects] = mtime;
	305	}
	306
2834bc27	307	#endif