sha1_file.c 62.4 KB
Newer Older
1 2 3 4 5 6 7 8 9
/*
 * GIT - The information manager from hell
 *
 * Copyright (C) Linus Torvalds, 2005
 *
 * This handles basic git sha1 object files - packing, unpacking,
 * creation etc.
 */
#include "cache.h"
10
#include "delta.h"
11
#include "pack.h"
12 13 14 15
#include "blob.h"
#include "commit.h"
#include "tag.h"
#include "tree.h"
16
#include "refs.h"
17
#include "pack-revindex.h"
18
#include "sha1-lookup.h"
19

20 21 22 23 24 25 26 27
#ifndef O_NOATIME
#if defined(__linux__) && (defined(__i386__) || defined(__PPC__))
#define O_NOATIME 01000000
#else
#define O_NOATIME 0
#endif
#endif

28 29
#ifdef NO_C99_FORMAT
#define SZ_FMT "lu"
30
static unsigned long sz_fmt(size_t s) { return (unsigned long)s; }
31 32
#else
#define SZ_FMT "zu"
33
static size_t sz_fmt(size_t s) { return s; }
34 35
#endif

36
const unsigned char null_sha1[20];
J
Junio C Hamano 已提交
37

38
const signed char hexval_table[256] = {
39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 00-07 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 08-0f */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 10-17 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 18-1f */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 20-27 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 28-2f */
	  0,  1,  2,  3,  4,  5,  6,  7,		/* 30-37 */
	  8,  9, -1, -1, -1, -1, -1, -1,		/* 38-3f */
	 -1, 10, 11, 12, 13, 14, 15, -1,		/* 40-47 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 48-4f */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 50-57 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 58-5f */
	 -1, 10, 11, 12, 13, 14, 15, -1,		/* 60-67 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 68-67 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 70-77 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 78-7f */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 80-87 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 88-8f */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 90-97 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* 98-9f */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* a0-a7 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* a8-af */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* b0-b7 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* b8-bf */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* c0-c7 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* c8-cf */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* d0-d7 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* d8-df */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* e0-e7 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* e8-ef */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* f0-f7 */
	 -1, -1, -1, -1, -1, -1, -1, -1,		/* f8-ff */
};
72 73 74 75 76 77 78 79 80 81 82 83 84 85

int get_sha1_hex(const char *hex, unsigned char *sha1)
{
	int i;
	for (i = 0; i < 20; i++) {
		unsigned int val = (hexval(hex[0]) << 4) | hexval(hex[1]);
		if (val & ~0xff)
			return -1;
		*sha1++ = val;
		hex += 2;
	}
	return 0;
}

86 87 88 89 90 91 92
static inline int offset_1st_component(const char *path)
{
	if (has_dos_drive_prefix(path))
		return 2 + (path[2] == '/');
	return *path == '/';
}

93 94
int safe_create_leading_directories(char *path)
{
95
	char *pos = path + offset_1st_component(path);
96 97
	struct stat st;

98 99 100 101
	while (pos) {
		pos = strchr(pos, '/');
		if (!pos)
			break;
102 103 104 105 106
		while (*++pos == '/')
			;
		if (!*pos)
			break;
		*--pos = '\0';
107 108 109
		if (!stat(path, &st)) {
			/* path exists */
			if (!S_ISDIR(st.st_mode)) {
110
				*pos = '/';
111
				return -3;
112
			}
113
		}
114 115 116 117
		else if (mkdir(path, 0777)) {
			*pos = '/';
			return -1;
		}
118 119 120 121
		else if (adjust_shared_perm(path)) {
			*pos = '/';
			return -2;
		}
122 123 124 125
		*pos++ = '/';
	}
	return 0;
}
126

127 128 129 130 131 132 133 134 135
int safe_create_leading_directories_const(const char *path)
{
	/* path points to cache entries, so xstrdup before messing with it */
	char *buf = xstrdup(path);
	int result = safe_create_leading_directories(buf);
	free(buf);
	return result;
}

136
char *sha1_to_hex(const unsigned char *sha1)
137
{
L
Linus Torvalds 已提交
138 139
	static int bufno;
	static char hexbuffer[4][50];
140
	static const char hex[] = "0123456789abcdef";
L
Linus Torvalds 已提交
141
	char *buffer = hexbuffer[3 & ++bufno], *buf = buffer;
142 143 144 145 146 147 148
	int i;

	for (i = 0; i < 20; i++) {
		unsigned int val = *sha1++;
		*buf++ = hex[val >> 4];
		*buf++ = hex[val & 0xf];
	}
149 150
	*buf = '\0';

151 152 153
	return buffer;
}

154 155 156 157 158 159 160 161 162 163 164 165
static void fill_sha1_path(char *pathbuf, const unsigned char *sha1)
{
	int i;
	for (i = 0; i < 20; i++) {
		static char hex[] = "0123456789abcdef";
		unsigned int val = sha1[i];
		char *pos = pathbuf + i*2 + (i > 0);
		*pos++ = hex[val >> 4];
		*pos = hex[val & 0xf];
	}
}

166 167
/*
 * NOTE! This returns a statically allocated buffer, so you have to be
168
 * careful about using it. Do an "xstrdup()" if you need to save the
169
 * filename.
170 171 172
 *
 * Also note that this returns the location for creating.  Reading
 * SHA1 file can happen from any alternate directory listed in the
J
Junio C Hamano 已提交
173
 * DB_ENVIRONMENT environment variable if it is not found in
174
 * the primary object database.
175 176 177 178 179 180
 */
char *sha1_file_name(const unsigned char *sha1)
{
	static char *name, *base;

	if (!base) {
J
Junio C Hamano 已提交
181
		const char *sha1_file_directory = get_object_directory();
182
		int len = strlen(sha1_file_directory);
183
		base = xmalloc(len + 60);
184 185 186 187 188 189
		memcpy(base, sha1_file_directory, len);
		memset(base+len, 0, 60);
		base[len] = '/';
		base[len+3] = '/';
		name = base + len + 1;
	}
190
	fill_sha1_path(name, sha1);
191 192 193
	return base;
}

194
static char *sha1_get_pack_name(const unsigned char *sha1,
J
Junio C Hamano 已提交
195
				char **name, char **base, const char *which)
196 197
{
	static const char hex[] = "0123456789abcdef";
198
	char *buf;
199 200
	int i;

201
	if (!*base) {
202 203
		const char *sha1_file_directory = get_object_directory();
		int len = strlen(sha1_file_directory);
204
		*base = xmalloc(len + 60);
J
Junio C Hamano 已提交
205 206
		sprintf(*base, "%s/pack/pack-1234567890123456789012345678901234567890.%s",
			sha1_file_directory, which);
207
		*name = *base + len + 11;
208 209
	}

210
	buf = *name;
211 212 213 214 215 216

	for (i = 0; i < 20; i++) {
		unsigned int val = *sha1++;
		*buf++ = hex[val >> 4];
		*buf++ = hex[val & 0xf];
	}
J
Junio C Hamano 已提交
217

218
	return *base;
219 220
}

221
char *sha1_pack_name(const unsigned char *sha1)
222
{
223
	static char *name, *base;
224

J
Junio C Hamano 已提交
225
	return sha1_get_pack_name(sha1, &name, &base, "pack");
226
}
227

228 229 230
char *sha1_pack_index_name(const unsigned char *sha1)
{
	static char *name, *base;
J
Junio C Hamano 已提交
231

J
Junio C Hamano 已提交
232
	return sha1_get_pack_name(sha1, &name, &base, "idx");
233 234
}

235 236
struct alternate_object_database *alt_odb_list;
static struct alternate_object_database **alt_odb_tail;
237

M
Martin Waitz 已提交
238 239
static void read_info_alternates(const char * alternates, int depth);

J
Junio C Hamano 已提交
240 241
/*
 * Prepare alternate object database registry.
242 243 244 245 246
 *
 * The variable alt_odb_list points at the list of struct
 * alternate_object_database.  The elements on this list come from
 * non-empty elements from colon separated ALTERNATE_DB_ENVIRONMENT
 * environment variable, and $GIT_OBJECT_DIRECTORY/info/alternates,
247 248
 * whose contents is similar to that environment variable but can be
 * LF separated.  Its base points at a statically allocated buffer that
249 250 251 252 253
 * contains "/the/directory/corresponding/to/.git/objects/...", while
 * its name points just after the slash at the end of ".git/objects/"
 * in the example above, and has enough space to hold 40-byte hex
 * SHA1, an extra slash for the first level indirection, and the
 * terminating NUL.
J
Junio C Hamano 已提交
254
 */
M
Martin Waitz 已提交
255
static int link_alt_odb_entry(const char * entry, int len, const char * relative_base, int depth)
256
{
257
	const char *objdir = get_object_directory();
M
Martin Waitz 已提交
258 259 260 261 262
	struct alternate_object_database *ent;
	struct alternate_object_database *alt;
	/* 43 = 40-byte + 2 '/' + terminating NUL */
	int pfxlen = len;
	int entlen = pfxlen + 43;
263
	int base_len = -1;
264

265
	if (!is_absolute_path(entry) && relative_base) {
M
Martin Waitz 已提交
266 267 268 269 270 271 272 273
		/* Relative alt-odb */
		if (base_len < 0)
			base_len = strlen(relative_base) + 1;
		entlen += base_len;
		pfxlen += base_len;
	}
	ent = xmalloc(sizeof(*ent) + entlen);

274
	if (!is_absolute_path(entry) && relative_base) {
M
Martin Waitz 已提交
275 276 277 278 279 280 281 282 283 284 285 286
		memcpy(ent->base, relative_base, base_len - 1);
		ent->base[base_len - 1] = '/';
		memcpy(ent->base + base_len, entry, len);
	}
	else
		memcpy(ent->base, entry, pfxlen);

	ent->name = ent->base + pfxlen + 1;
	ent->base[pfxlen + 3] = '/';
	ent->base[pfxlen] = ent->base[entlen-1] = 0;

	/* Detect cases where alternate disappeared */
287
	if (!is_directory(ent->base)) {
M
Martin Waitz 已提交
288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332
		error("object directory %s does not exist; "
		      "check .git/objects/info/alternates.",
		      ent->base);
		free(ent);
		return -1;
	}

	/* Prevent the common mistake of listing the same
	 * thing twice, or object directory itself.
	 */
	for (alt = alt_odb_list; alt; alt = alt->next) {
		if (!memcmp(ent->base, alt->base, pfxlen)) {
			free(ent);
			return -1;
		}
	}
	if (!memcmp(ent->base, objdir, pfxlen)) {
		free(ent);
		return -1;
	}

	/* add the alternate entry */
	*alt_odb_tail = ent;
	alt_odb_tail = &(ent->next);
	ent->next = NULL;

	/* recursively add alternates */
	read_info_alternates(ent->base, depth + 1);

	ent->base[pfxlen] = '/';

	return 0;
}

static void link_alt_odb_entries(const char *alt, const char *ep, int sep,
				 const char *relative_base, int depth)
{
	const char *cp, *last;

	if (depth > 5) {
		error("%s: ignoring alternate object stores, nesting too deep.",
				relative_base);
		return;
	}

333
	last = alt;
334 335 336 337 338 339 340 341
	while (last < ep) {
		cp = last;
		if (cp < ep && *cp == '#') {
			while (cp < ep && *cp != sep)
				cp++;
			last = cp + 1;
			continue;
		}
M
Martin Waitz 已提交
342 343
		while (cp < ep && *cp != sep)
			cp++;
344
		if (last != cp) {
345
			if (!is_absolute_path(last) && depth) {
M
Martin Waitz 已提交
346 347 348 349 350
				error("%s: ignoring relative alternate object store %s",
						relative_base, last);
			} else {
				link_alt_odb_entry(last, cp - last,
						relative_base, depth);
351
			}
352
		}
353
		while (cp < ep && *cp == sep)
354 355
			cp++;
		last = cp;
356
	}
357 358
}

M
Martin Waitz 已提交
359
static void read_info_alternates(const char * relative_base, int depth)
360
{
361
	char *map;
362
	size_t mapsz;
363
	struct stat st;
364 365 366 367 368
	const char alt_file_name[] = "info/alternates";
	/* Given that relative_base is no longer than PATH_MAX,
	   ensure that "path" has enough space to append "/", the
	   file name, "info/alternates", and a trailing NUL.  */
	char path[PATH_MAX + 1 + sizeof alt_file_name];
M
Martin Waitz 已提交
369
	int fd;
370

371
	sprintf(path, "%s/%s", relative_base, alt_file_name);
372 373 374 375 376
	fd = open(path, O_RDONLY);
	if (fd < 0)
		return;
	if (fstat(fd, &st) || (st.st_size == 0)) {
		close(fd);
377
		return;
378
	}
379 380
	mapsz = xsize_t(st.st_size);
	map = xmmap(NULL, mapsz, PROT_READ, MAP_PRIVATE, fd, 0);
381 382
	close(fd);

383
	link_alt_odb_entries(map, map + mapsz, '\n', relative_base, depth);
M
Martin Waitz 已提交
384

385
	munmap(map, mapsz);
386 387
}

388 389 390
void add_to_alternates_file(const char *reference)
{
	struct lock_file *lock = xcalloc(1, sizeof(struct lock_file));
391
	int fd = hold_lock_file_for_append(lock, git_path("objects/info/alternates"), LOCK_DIE_ON_ERROR);
392 393 394 395 396 397 398 399
	char *alt = mkpath("%s/objects\n", reference);
	write_or_die(fd, alt, strlen(alt));
	if (commit_lock_file(lock))
		die("could not close alternates file");
	if (alt_odb_tail)
		link_alt_odb_entries(alt, alt + strlen(alt), '\n', NULL, 0);
}

400 401 402 403 404 405 406 407 408 409
void foreach_alt_odb(alt_odb_fn fn, void *cb)
{
	struct alternate_object_database *ent;

	prepare_alt_odb();
	for (ent = alt_odb_list; ent; ent = ent->next)
		if (fn(ent, cb))
			return;
}

M
Martin Waitz 已提交
410 411
void prepare_alt_odb(void)
{
T
Timo Hirvonen 已提交
412
	const char *alt;
M
Martin Waitz 已提交
413

S
Shawn O. Pearce 已提交
414 415 416
	if (alt_odb_tail)
		return;

M
Martin Waitz 已提交
417 418 419 420
	alt = getenv(ALTERNATE_DB_ENVIRONMENT);
	if (!alt) alt = "";

	alt_odb_tail = &alt_odb_list;
421
	link_alt_odb_entries(alt, alt + strlen(alt), PATH_SEP, NULL, 0);
M
Martin Waitz 已提交
422 423 424 425

	read_info_alternates(get_object_directory(), 0);
}

426
static int has_loose_object(const unsigned char *sha1)
427 428
{
	char *name = sha1_file_name(sha1);
429
	struct alternate_object_database *alt;
430

431 432
	if (!access(name, F_OK))
		return 1;
433
	prepare_alt_odb();
434 435
	for (alt = alt_odb_list; alt; alt = alt->next) {
		name = alt->name;
436
		fill_sha1_path(name, sha1);
437 438
		if (!access(alt->base, F_OK))
			return 1;
439
	}
440
	return 0;
441 442
}

443
static unsigned int pack_used_ctr;
444 445 446 447
static unsigned int pack_mmap_calls;
static unsigned int peak_pack_open_windows;
static unsigned int pack_open_windows;
static size_t peak_pack_mapped;
448
static size_t pack_mapped;
449
struct packed_git *packed_git;
450

J
Junio C Hamano 已提交
451
void pack_report(void)
452 453
{
	fprintf(stderr,
454 455 456
		"pack_report: getpagesize()            = %10" SZ_FMT "\n"
		"pack_report: core.packedGitWindowSize = %10" SZ_FMT "\n"
		"pack_report: core.packedGitLimit      = %10" SZ_FMT "\n",
457 458 459
		sz_fmt(getpagesize()),
		sz_fmt(packed_git_window_size),
		sz_fmt(packed_git_limit));
460 461 462 463
	fprintf(stderr,
		"pack_report: pack_used_ctr            = %10u\n"
		"pack_report: pack_mmap_calls          = %10u\n"
		"pack_report: pack_open_windows        = %10u / %10u\n"
464 465
		"pack_report: pack_mapped              = "
			"%10" SZ_FMT " / %10" SZ_FMT "\n",
466 467 468
		pack_used_ctr,
		pack_mmap_calls,
		pack_open_windows, peak_pack_open_windows,
469
		sz_fmt(pack_mapped), sz_fmt(peak_pack_mapped));
470 471
}

472
static int check_packed_git_idx(const char *path,  struct packed_git *p)
473 474
{
	void *idx_map;
475
	struct pack_idx_header *hdr;
476
	size_t idx_size;
477
	uint32_t version, nr, i, *index;
478
	int fd = open(path, O_RDONLY);
479
	struct stat st;
480

481 482 483 484 485 486
	if (fd < 0)
		return -1;
	if (fstat(fd, &st)) {
		close(fd);
		return -1;
	}
487
	idx_size = xsize_t(st.st_size);
488 489 490 491
	if (idx_size < 4 * 256 + 20 + 20) {
		close(fd);
		return error("index file %s is too small", path);
	}
492
	idx_map = xmmap(NULL, idx_size, PROT_READ, MAP_PRIVATE, fd, 0);
493 494
	close(fd);

495 496
	hdr = idx_map;
	if (hdr->idx_signature == htonl(PACK_IDX_SIGNATURE)) {
497 498 499
		version = ntohl(hdr->idx_version);
		if (version < 2 || version > 2) {
			munmap(idx_map, idx_size);
500
			return error("index file %s is version %"PRIu32
501 502 503 504 505 506
				     " and is not supported by this binary"
				     " (try upgrading GIT to a newer version)",
				     path, version);
		}
	} else
		version = 1;
507

508
	nr = 0;
509
	index = idx_map;
510 511
	if (version > 1)
		index += 2;  /* skip index header */
512
	for (i = 0; i < 256; i++) {
513
		uint32_t n = ntohl(index[i]);
514 515
		if (n < nr) {
			munmap(idx_map, idx_size);
516
			return error("non-monotonic index %s", path);
517
		}
518 519 520
		nr = n;
	}

521 522 523 524 525 526 527 528 529 530
	if (version == 1) {
		/*
		 * Total size:
		 *  - 256 index entries 4 bytes each
		 *  - 24-byte entries * nr (20-byte sha1 + 4-byte offset)
		 *  - 20-byte SHA1 of the packfile
		 *  - 20-byte SHA1 file checksum
		 */
		if (idx_size != 4*256 + nr * 24 + 20 + 20) {
			munmap(idx_map, idx_size);
531
			return error("wrong index v1 file size in %s", path);
532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547
		}
	} else if (version == 2) {
		/*
		 * Minimum size:
		 *  - 8 bytes of header
		 *  - 256 index entries 4 bytes each
		 *  - 20-byte sha1 entry * nr
		 *  - 4-byte crc entry * nr
		 *  - 4-byte offset entry * nr
		 *  - 20-byte SHA1 of the packfile
		 *  - 20-byte SHA1 file checksum
		 * And after the 4-byte offset table might be a
		 * variable sized table containing 8-byte entries
		 * for offsets larger than 2^31.
		 */
		unsigned long min_size = 8 + 4*256 + nr*(20 + 4 + 4) + 20 + 20;
L
Linus Torvalds 已提交
548 549 550 551
		unsigned long max_size = min_size;
		if (nr)
			max_size += (nr - 1)*8;
		if (idx_size < min_size || idx_size > max_size) {
552
			munmap(idx_map, idx_size);
553
			return error("wrong index v2 file size in %s", path);
554
		}
555 556 557 558 559 560 561 562 563
		if (idx_size != min_size &&
		    /*
		     * make sure we can deal with large pack offsets.
		     * 31-bit signed offset won't be enough, neither
		     * 32-bit unsigned one will be.
		     */
		    (sizeof(off_t) <= 4)) {
			munmap(idx_map, idx_size);
			return error("pack too large for current definition of off_t in %s", path);
564
		}
565
	}
566

567
	p->index_version = version;
568 569
	p->index_data = idx_map;
	p->index_size = idx_size;
N
Nicolas Pitre 已提交
570
	p->num_objects = nr;
571 572 573
	return 0;
}

574
int open_pack_index(struct packed_git *p)
575 576 577 578 579 580 581 582 583 584 585 586 587 588
{
	char *idx_name;
	int ret;

	if (p->index_data)
		return 0;

	idx_name = xstrdup(p->pack_name);
	strcpy(idx_name + strlen(idx_name) - strlen(".pack"), ".idx");
	ret = check_packed_git_idx(idx_name, p);
	free(idx_name);
	return ret;
}

589 590 591 592
static void scan_windows(struct packed_git *p,
	struct packed_git **lru_p,
	struct pack_window **lru_w,
	struct pack_window **lru_l)
593
{
594 595 596 597 598 599 600 601
	struct pack_window *w, *w_l;

	for (w_l = NULL, w = p->windows; w; w = w->next) {
		if (!w->inuse_cnt) {
			if (!*lru_w || w->last_used < (*lru_w)->last_used) {
				*lru_p = p;
				*lru_w = w;
				*lru_l = w_l;
602 603
			}
		}
604
		w_l = w;
605
	}
606 607
}

608
static int unuse_one_window(struct packed_git *current, int keep_fd)
609 610 611 612 613 614 615 616
{
	struct packed_git *p, *lru_p = NULL;
	struct pack_window *lru_w = NULL, *lru_l = NULL;

	if (current)
		scan_windows(current, &lru_p, &lru_w, &lru_l);
	for (p = packed_git; p; p = p->next)
		scan_windows(p, &lru_p, &lru_w, &lru_l);
617 618 619 620 621 622 623
	if (lru_p) {
		munmap(lru_w->base, lru_w->len);
		pack_mapped -= lru_w->len;
		if (lru_l)
			lru_l->next = lru_w->next;
		else {
			lru_p->windows = lru_w->next;
624
			if (!lru_p->windows && lru_p->pack_fd != keep_fd) {
625 626 627 628 629
				close(lru_p->pack_fd);
				lru_p->pack_fd = -1;
			}
		}
		free(lru_w);
630
		pack_open_windows--;
631 632 633
		return 1;
	}
	return 0;
634 635
}

636
void release_pack_memory(size_t need, int fd)
637 638
{
	size_t cur = pack_mapped;
639
	while (need >= (cur - pack_mapped) && unuse_one_window(NULL, fd))
640 641 642
		; /* nothing */
}

643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658
void close_pack_windows(struct packed_git *p)
{
	while (p->windows) {
		struct pack_window *w = p->windows;

		if (w->inuse_cnt)
			die("pack '%s' still has open windows to it",
			    p->pack_name);
		munmap(w->base, w->len);
		pack_mapped -= w->len;
		pack_open_windows--;
		p->windows = w->next;
		free(w);
	}
}

659
void unuse_pack(struct pack_window **w_cursor)
660
{
661 662 663 664 665
	struct pack_window *w = *w_cursor;
	if (w) {
		w->inuse_cnt--;
		*w_cursor = NULL;
	}
666 667
}

668 669 670 671 672
/*
 * Do not call this directly as this leaks p->pack_fd on error return;
 * call open_packed_git() instead.
 */
static int open_packed_git_1(struct packed_git *p)
673
{
674 675 676 677
	struct stat st;
	struct pack_header hdr;
	unsigned char sha1[20];
	unsigned char *idx_sha1;
678
	long fd_flag;
679

680 681 682
	if (!p->index_data && open_pack_index(p))
		return error("packfile %s index unavailable", p->pack_name);

683 684
	p->pack_fd = open(p->pack_name, O_RDONLY);
	if (p->pack_fd < 0 || fstat(p->pack_fd, &st))
685
		return -1;
686 687

	/* If we created the struct before we had the pack we lack size. */
688 689
	if (!p->pack_size) {
		if (!S_ISREG(st.st_mode))
690
			return error("packfile %s not a regular file", p->pack_name);
691
		p->pack_size = st.st_size;
692
	} else if (p->pack_size != st.st_size)
693
		return error("packfile %s size changed", p->pack_name);
694

695 696 697 698 699
	/* We leave these file descriptors open with sliding mmap;
	 * there is no point keeping them open across exec(), though.
	 */
	fd_flag = fcntl(p->pack_fd, F_GETFD, 0);
	if (fd_flag < 0)
700
		return error("cannot determine file descriptor flags");
701 702
	fd_flag |= FD_CLOEXEC;
	if (fcntl(p->pack_fd, F_SETFD, fd_flag) == -1)
703
		return error("cannot set FD_CLOEXEC");
704

705
	/* Verify we recognize this pack file format. */
706
	if (read_in_full(p->pack_fd, &hdr, sizeof(hdr)) != sizeof(hdr))
707
		return error("file %s is far too short to be a packfile", p->pack_name);
708
	if (hdr.hdr_signature != htonl(PACK_SIGNATURE))
709
		return error("file %s is not a GIT packfile", p->pack_name);
710
	if (!pack_version_ok(hdr.hdr_version))
711 712
		return error("packfile %s is version %"PRIu32" and not"
			" supported (try upgrading GIT to a newer version)",
713 714 715
			p->pack_name, ntohl(hdr.hdr_version));

	/* Verify the pack matches its index. */
N
Nicolas Pitre 已提交
716
	if (p->num_objects != ntohl(hdr.hdr_entries))
717 718
		return error("packfile %s claims to have %"PRIu32" objects"
			     " while index indicates %"PRIu32" objects",
N
Nicolas Pitre 已提交
719 720
			     p->pack_name, ntohl(hdr.hdr_entries),
			     p->num_objects);
721
	if (lseek(p->pack_fd, p->pack_size - sizeof(sha1), SEEK_SET) == -1)
722
		return error("end of packfile %s is unavailable", p->pack_name);
723
	if (read_in_full(p->pack_fd, sha1, sizeof(sha1)) != sizeof(sha1))
724
		return error("packfile %s signature is unavailable", p->pack_name);
725
	idx_sha1 = ((unsigned char *)p->index_data) + p->index_size - 40;
726
	if (hashcmp(sha1, idx_sha1))
727 728
		return error("packfile %s does not match index", p->pack_name);
	return 0;
729 730
}

731 732 733 734 735 736 737 738 739 740 741
static int open_packed_git(struct packed_git *p)
{
	if (!open_packed_git_1(p))
		return 0;
	if (p->pack_fd != -1) {
		close(p->pack_fd);
		p->pack_fd = -1;
	}
	return -1;
}

742
static int in_window(struct pack_window *win, off_t offset)
743 744 745 746 747 748 749 750 751 752 753 754
{
	/* We must promise at least 20 bytes (one hash) after the
	 * offset is available from this window, otherwise the offset
	 * is not actually in this window and a different window (which
	 * has that one hash excess) must be used.  This is to support
	 * the object header and delta base parsing routines below.
	 */
	off_t win_off = win->offset;
	return win_off <= offset
		&& (offset + 20) <= (win_off + win->len);
}

755 756
unsigned char* use_pack(struct packed_git *p,
		struct pack_window **w_cursor,
757
		off_t offset,
758
		unsigned int *left)
759
{
760
	struct pack_window *win = *w_cursor;
761

762 763
	if (p->pack_fd == -1 && open_packed_git(p))
		die("packfile %s cannot be accessed", p->pack_name);
764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780

	/* Since packfiles end in a hash of their content and its
	 * pointless to ask for an offset into the middle of that
	 * hash, and the in_window function above wouldn't match
	 * don't allow an offset too close to the end of the file.
	 */
	if (offset > (p->pack_size - 20))
		die("offset beyond end of packfile (truncated pack?)");

	if (!win || !in_window(win, offset)) {
		if (win)
			win->inuse_cnt--;
		for (win = p->windows; win; win = win->next) {
			if (in_window(win, offset))
				break;
		}
		if (!win) {
781
			size_t window_align = packed_git_window_size / 2;
782
			off_t len;
783
			win = xcalloc(1, sizeof(*win));
784
			win->offset = (offset / window_align) * window_align;
785 786 787 788
			len = p->pack_size - win->offset;
			if (len > packed_git_window_size)
				len = packed_git_window_size;
			win->len = (size_t)len;
789
			pack_mapped += win->len;
790
			while (packed_git_limit < pack_mapped
791
				&& unuse_one_window(p, p->pack_fd))
792
				; /* nothing */
793
			win->base = xmmap(NULL, win->len,
794 795 796
				PROT_READ, MAP_PRIVATE,
				p->pack_fd, win->offset);
			if (win->base == MAP_FAILED)
797 798 799
				die("packfile %s cannot be mapped: %s",
					p->pack_name,
					strerror(errno));
800 801 802 803 804 805
			pack_mmap_calls++;
			pack_open_windows++;
			if (pack_mapped > peak_pack_mapped)
				peak_pack_mapped = pack_mapped;
			if (pack_open_windows > peak_pack_open_windows)
				peak_pack_open_windows = pack_open_windows;
806 807 808
			win->next = p->windows;
			p->windows = win;
		}
809
	}
810 811 812 813 814
	if (win != *w_cursor) {
		win->last_used = pack_used_ctr++;
		win->inuse_cnt++;
		*w_cursor = win;
	}
815
	offset -= win->offset;
816
	if (left)
817
		*left = win->len - xsize_t(offset);
818
	return win->base + offset;
819 820
}

821 822 823 824 825 826 827 828
static struct packed_git *alloc_packed_git(int extra)
{
	struct packed_git *p = xmalloc(sizeof(*p) + extra);
	memset(p, 0, sizeof(*p));
	p->pack_fd = -1;
	return p;
}

829
struct packed_git *add_packed_git(const char *path, int path_len, int local)
830 831
{
	struct stat st;
832
	struct packed_git *p = alloc_packed_git(path_len + 2);
833

834 835 836 837 838
	/*
	 * Make sure a corresponding .pack file exists and that
	 * the index looks sane.
	 */
	path_len -= strlen(".idx");
839 840
	if (path_len < 1) {
		free(p);
841
		return NULL;
842
	}
843 844
	memcpy(p->pack_name, path, path_len);
	strcpy(p->pack_name + path_len, ".pack");
845
	if (stat(p->pack_name, &st) || !S_ISREG(st.st_mode)) {
846
		free(p);
847 848
		return NULL;
	}
849

850 851 852 853
	/* ok, it looks sane as far as we can check without
	 * actually mapping the pack file.
	 */
	p->pack_size = st.st_size;
854
	p->pack_local = local;
855
	p->mtime = st.st_mtime;
856 857
	if (path_len < 40 || get_sha1_hex(path + path_len - 40, p->sha1))
		hashclr(p->sha1);
858 859 860
	return p;
}

861
struct packed_git *parse_pack_index(unsigned char *sha1)
862
{
863
	const char *idx_path = sha1_pack_index_name(sha1);
864
	const char *path = sha1_pack_name(sha1);
865
	struct packed_git *p = alloc_packed_git(strlen(path) + 1);
866

867 868
	strcpy(p->pack_name, path);
	hashcpy(p->sha1, sha1);
869 870
	if (check_packed_git_idx(idx_path, p)) {
		free(p);
871
		return NULL;
872
	}
873 874 875 876 877 878 879 880 881 882

	return p;
}

void install_packed_git(struct packed_git *pack)
{
	pack->next = packed_git;
	packed_git = pack;
}

883
static void prepare_packed_git_one(char *objdir, int local)
884
{
885 886 887 888
	/* Ensure that this buffer is large enough so that we can
	   append "/pack/" without clobbering the stack even if
	   strlen(objdir) were PATH_MAX.  */
	char path[PATH_MAX + 1 + 4 + 1 + 1];
889 890 891 892 893 894 895
	int len;
	DIR *dir;
	struct dirent *de;

	sprintf(path, "%s/pack", objdir);
	len = strlen(path);
	dir = opendir(path);
896
	if (!dir) {
J
Junio C Hamano 已提交
897
		if (errno != ENOENT)
898
			error("unable to open object pack directory: %s: %s",
J
Junio C Hamano 已提交
899
			      path, strerror(errno));
900
		return;
901
	}
902 903 904 905 906
	path[len++] = '/';
	while ((de = readdir(dir)) != NULL) {
		int namelen = strlen(de->d_name);
		struct packed_git *p;

907
		if (!has_extension(de->d_name, ".idx"))
908 909
			continue;

910 911 912
		if (len + namelen + 1 > sizeof(path))
			continue;

913
		/* Don't reopen a pack we already have. */
914
		strcpy(path + len, de->d_name);
915 916 917 918 919 920
		for (p = packed_git; p; p = p->next) {
			if (!memcmp(path, p->pack_name, len + namelen - 4))
				break;
		}
		if (p)
			continue;
921 922 923
		/* See if it really is a valid .idx file with corresponding
		 * .pack file that we can map.
		 */
924
		p = add_packed_git(path, len + namelen, local);
925 926
		if (!p)
			continue;
927
		install_packed_git(p);
928
	}
929
	closedir(dir);
930 931
}

932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985
static int sort_pack(const void *a_, const void *b_)
{
	struct packed_git *a = *((struct packed_git **)a_);
	struct packed_git *b = *((struct packed_git **)b_);
	int st;

	/*
	 * Local packs tend to contain objects specific to our
	 * variant of the project than remote ones.  In addition,
	 * remote ones could be on a network mounted filesystem.
	 * Favor local ones for these reasons.
	 */
	st = a->pack_local - b->pack_local;
	if (st)
		return -st;

	/*
	 * Younger packs tend to contain more recent objects,
	 * and more recent objects tend to get accessed more
	 * often.
	 */
	if (a->mtime < b->mtime)
		return 1;
	else if (a->mtime == b->mtime)
		return 0;
	return -1;
}

static void rearrange_packed_git(void)
{
	struct packed_git **ary, *p;
	int i, n;

	for (n = 0, p = packed_git; p; p = p->next)
		n++;
	if (n < 2)
		return;

	/* prepare an array of packed_git for easier sorting */
	ary = xcalloc(n, sizeof(struct packed_git *));
	for (n = 0, p = packed_git; p; p = p->next)
		ary[n++] = p;

	qsort(ary, n, sizeof(struct packed_git *), sort_pack);

	/* link them back again */
	for (i = 0; i < n - 1; i++)
		ary[i]->next = ary[i + 1];
	ary[n - 1]->next = NULL;
	packed_git = ary[0];

	free(ary);
}

986
static int prepare_packed_git_run_once = 0;
987
void prepare_packed_git(void)
988
{
989
	struct alternate_object_database *alt;
990

991
	if (prepare_packed_git_run_once)
992
		return;
993
	prepare_packed_git_one(get_object_directory(), 1);
994
	prepare_alt_odb();
995
	for (alt = alt_odb_list; alt; alt = alt->next) {
996
		alt->name[-1] = 0;
997
		prepare_packed_git_one(alt->base, 0);
998
		alt->name[-1] = '/';
999
	}
1000
	rearrange_packed_git();
1001 1002 1003
	prepare_packed_git_run_once = 1;
}

1004
void reprepare_packed_git(void)
1005
{
1006
	discard_revindex();
1007 1008
	prepare_packed_git_run_once = 0;
	prepare_packed_git();
1009 1010
}

1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022
static void mark_bad_packed_object(struct packed_git *p,
				   const unsigned char *sha1)
{
	unsigned i;
	for (i = 0; i < p->num_bad_objects; i++)
		if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i))
			return;
	p->bad_object_sha1 = xrealloc(p->bad_object_sha1, 20 * (p->num_bad_objects + 1));
	hashcpy(p->bad_object_sha1 + 20 * p->num_bad_objects, sha1);
	p->num_bad_objects++;
}

1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034
static int has_packed_and_bad(const unsigned char *sha1)
{
	struct packed_git *p;
	unsigned i;

	for (p = packed_git; p; p = p->next)
		for (i = 0; i < p->num_bad_objects; i++)
			if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i))
				return 1;
	return 0;
}

1035
int check_sha1_signature(const unsigned char *sha1, void *map, unsigned long size, const char *type)
1036 1037
{
	unsigned char real_sha1[20];
1038
	hash_sha1_file(map, size, type, real_sha1);
1039
	return hashcmp(sha1, real_sha1) ? -1 : 0;
1040 1041
}

1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077
static int git_open_noatime(const char *name)
{
	static int sha1_file_open_flag = O_NOATIME;
	int fd = open(name, O_RDONLY | sha1_file_open_flag);

	/* Might the failure be due to O_NOATIME? */
	if (fd < 0 && errno != ENOENT && sha1_file_open_flag) {
		fd = open(name, O_RDONLY);
		if (fd >= 0)
			sha1_file_open_flag = 0;
	}
	return fd;
}

static int open_sha1_file(const unsigned char *sha1)
{
	int fd;
	char *name = sha1_file_name(sha1);
	struct alternate_object_database *alt;

	fd = git_open_noatime(name);
	if (fd >= 0)
		return fd;

	prepare_alt_odb();
	errno = ENOENT;
	for (alt = alt_odb_list; alt; alt = alt->next) {
		name = alt->name;
		fill_sha1_path(name, sha1);
		fd = git_open_noatime(alt->base);
		if (fd >= 0)
			return fd;
	}
	return -1;
}

J
Junio C Hamano 已提交
1078
static void *map_sha1_file(const unsigned char *sha1, unsigned long *size)
1079 1080
{
	void *map;
1081
	int fd;
1082

1083 1084 1085 1086
	fd = open_sha1_file(sha1);
	map = NULL;
	if (fd >= 0) {
		struct stat st;
1087

1088 1089 1090
		if (!fstat(fd, &st)) {
			*size = xsize_t(st.st_size);
			map = xmmap(NULL, *size, PROT_READ, MAP_PRIVATE, fd, 0);
1091
		}
1092
		close(fd);
1093 1094 1095 1096
	}
	return map;
}

1097
static int legacy_loose_object(unsigned char *map)
1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112
{
	unsigned int word;

	/*
	 * Is it a zlib-compressed buffer? If so, the first byte
	 * must be 0x78 (15-bit window size, deflated), and the
	 * first 16-bit word is evenly divisible by 31
	 */
	word = (map[0] << 8) + map[1];
	if (map[0] == 0x78 && !(word % 31))
		return 1;
	else
		return 0;
}

1113 1114
unsigned long unpack_object_header_buffer(const unsigned char *buf,
		unsigned long len, enum object_type *type, unsigned long *sizep)
1115
{
1116
	unsigned shift;
1117 1118
	unsigned char c;
	unsigned long size;
1119 1120 1121 1122 1123 1124 1125
	unsigned long used = 0;

	c = buf[used++];
	*type = (c >> 4) & 7;
	size = c & 15;
	shift = 4;
	while (c & 0x80) {
1126 1127
		if (len <= used || sizeof(long) * 8 <= shift) {
			error("bad object header");
1128
			return 0;
1129
		}
1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144
		c = buf[used++];
		size += (c & 0x7f) << shift;
		shift += 7;
	}
	*sizep = size;
	return used;
}

static int unpack_sha1_header(z_stream *stream, unsigned char *map, unsigned long mapsize, void *buffer, unsigned long bufsiz)
{
	unsigned long size, used;
	static const char valid_loose_object_type[8] = {
		0, /* OBJ_EXT */
		1, 1, 1, 1, /* "commit", "tree", "blob", "tag" */
		0, /* "delta" and others are invalid in a loose object */
1145
	};
1146
	enum object_type type;
1147

1148 1149 1150 1151 1152
	/* Get the data stream */
	memset(stream, 0, sizeof(*stream));
	stream->next_in = map;
	stream->avail_in = mapsize;
	stream->next_out = buffer;
1153 1154
	stream->avail_out = bufsiz;

1155
	if (legacy_loose_object(map)) {
1156 1157 1158 1159
		inflateInit(stream);
		return inflate(stream, 0);
	}

1160 1161 1162 1163 1164 1165 1166 1167

	/*
	 * There used to be a second loose object header format which
	 * was meant to mimic the in-pack format, allowing for direct
	 * copy of the object data.  This format turned up not to be
	 * really worth it and we don't write it any longer.  But we
	 * can still read it.
	 */
1168
	used = unpack_object_header_buffer(map, mapsize, &type, &size);
1169
	if (!used || !valid_loose_object_type[type])
1170
		return -1;
1171 1172
	map += used;
	mapsize -= used;
1173 1174 1175 1176

	/* Set up the stream for the rest.. */
	stream->next_in = map;
	stream->avail_in = mapsize;
1177
	inflateInit(stream);
1178 1179

	/* And generate the fake traditional header */
1180
	stream->total_out = 1 + snprintf(buffer, bufsiz, "%s %lu",
1181
					 typename(type), size);
1182
	return 0;
1183 1184
}

1185
static void *unpack_sha1_rest(z_stream *stream, void *buffer, unsigned long size, const unsigned char *sha1)
1186 1187
{
	int bytes = strlen(buffer) + 1;
1188
	unsigned char *buf = xmalloc(1+size);
1189
	unsigned long n;
1190
	int status = Z_OK;
1191

1192 1193 1194 1195 1196
	n = stream->total_out - bytes;
	if (n > size)
		n = size;
	memcpy(buf, (char *) buffer + bytes, n);
	bytes = n;
1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210
	if (bytes <= size) {
		/*
		 * The above condition must be (bytes <= size), not
		 * (bytes < size).  In other words, even though we
		 * expect no more output and set avail_out to zer0,
		 * the input zlib stream may have bytes that express
		 * "this concludes the stream", and we *do* want to
		 * eat that input.
		 *
		 * Otherwise we would not be able to test that we
		 * consumed all the input to reach the expected size;
		 * we also want to check that zlib tells us that all
		 * went well with status == Z_STREAM_END at the end.
		 */
1211 1212
		stream->next_out = buf + bytes;
		stream->avail_out = size - bytes;
1213 1214
		while (status == Z_OK)
			status = inflate(stream, Z_FINISH);
1215 1216
	}
	buf[size] = 0;
1217
	if (status == Z_STREAM_END && !stream->avail_in) {
1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228
		inflateEnd(stream);
		return buf;
	}

	if (status < 0)
		error("corrupt loose object '%s'", sha1_to_hex(sha1));
	else if (stream->avail_in)
		error("garbage at end of loose object '%s'",
		      sha1_to_hex(sha1));
	free(buf);
	return NULL;
1229 1230 1231 1232 1233 1234 1235
}

/*
 * We used to just use "sscanf()", but that's actually way
 * too permissive for what we want to check. So do an anal
 * object header parse by hand.
 */
1236
static int parse_sha1_header(const char *hdr, unsigned long *sizep)
1237
{
1238
	char type[10];
1239 1240 1241 1242
	int i;
	unsigned long size;

	/*
J
Junio C Hamano 已提交
1243
	 * The type can be at most ten bytes (including the
1244
	 * terminating '\0' that we add), and is followed by
1245
	 * a space.
1246
	 */
1247
	i = 0;
1248 1249 1250 1251
	for (;;) {
		char c = *hdr++;
		if (c == ' ')
			break;
1252 1253
		type[i++] = c;
		if (i >= sizeof(type))
1254 1255
			return -1;
	}
1256
	type[i] = 0;
1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278

	/*
	 * The length must follow immediately, and be in canonical
	 * decimal format (ie "010" is not valid).
	 */
	size = *hdr++ - '0';
	if (size > 9)
		return -1;
	if (size) {
		for (;;) {
			unsigned long c = *hdr - '0';
			if (c > 9)
				break;
			hdr++;
			size = size * 10 + c;
		}
	}
	*sizep = size;

	/*
	 * The length must be followed by a zero byte
	 */
1279
	return *hdr ? -1 : type_from_string(type);
1280 1281
}

1282
static void *unpack_sha1_file(void *map, unsigned long mapsize, enum object_type *type, unsigned long *size, const unsigned char *sha1)
1283
{
1284
	int ret;
1285
	z_stream stream;
1286
	char hdr[8192];
1287

1288
	ret = unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr));
1289
	if (ret < Z_OK || (*type = parse_sha1_header(hdr, size)) < 0)
1290 1291
		return NULL;

1292
	return unpack_sha1_rest(&stream, hdr, *size, sha1);
1293 1294
}

N
Nicolas Pitre 已提交
1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331
unsigned long get_size_from_delta(struct packed_git *p,
				  struct pack_window **w_curs,
			          off_t curpos)
{
	const unsigned char *data;
	unsigned char delta_head[20], *in;
	z_stream stream;
	int st;

	memset(&stream, 0, sizeof(stream));
	stream.next_out = delta_head;
	stream.avail_out = sizeof(delta_head);

	inflateInit(&stream);
	do {
		in = use_pack(p, w_curs, curpos, &stream.avail_in);
		stream.next_in = in;
		st = inflate(&stream, Z_FINISH);
		curpos += stream.next_in - in;
	} while ((st == Z_OK || st == Z_BUF_ERROR) &&
		 stream.total_out < sizeof(delta_head));
	inflateEnd(&stream);
	if ((st != Z_STREAM_END) && stream.total_out != sizeof(delta_head))
		die("delta data unpack-initial failed");

	/* Examine the initial part of the delta to figure out
	 * the result size.
	 */
	data = delta_head;

	/* ignore base size */
	get_delta_hdr_size(&data, delta_head+sizeof(delta_head));

	/* Read the result size */
	return get_delta_hdr_size(&data, delta_head+sizeof(delta_head));
}

1332
static off_t get_delta_base(struct packed_git *p,
1333
				    struct pack_window **w_curs,
1334
				    off_t *curpos,
1335
				    enum object_type type,
1336
				    off_t delta_obj_offset)
1337
{
1338
	unsigned char *base_info = use_pack(p, w_curs, *curpos, NULL);
1339
	off_t base_offset;
1340

1341 1342 1343 1344 1345 1346
	/* use_pack() assured us we have [base_info, base_info + 20)
	 * as a range that we can look at without walking off the
	 * end of the mapped window.  Its actually the hash size
	 * that is assured.  An OFS_DELTA longer than the hash size
	 * is stupid, as then a REF_DELTA would be smaller to store.
	 */
1347
	if (type == OBJ_OFS_DELTA) {
1348 1349 1350 1351 1352
		unsigned used = 0;
		unsigned char c = base_info[used++];
		base_offset = c & 127;
		while (c & 128) {
			base_offset += 1;
1353
			if (!base_offset || MSB(base_offset, 7))
1354
				return 0;  /* overflow */
1355 1356 1357 1358
			c = base_info[used++];
			base_offset = (base_offset << 7) + (c & 127);
		}
		base_offset = delta_obj_offset - base_offset;
1359
		if (base_offset <= 0 || base_offset >= delta_obj_offset)
1360
			return 0;  /* out of bound */
1361
		*curpos += used;
1362
	} else if (type == OBJ_REF_DELTA) {
1363 1364
		/* The base entry _must_ be in the same pack */
		base_offset = find_pack_entry_one(base_info, p);
1365
		*curpos += 20;
1366 1367
	} else
		die("I am totally screwed");
1368
	return base_offset;
1369 1370
}

J
Junio C Hamano 已提交
1371
/* forward declaration for a mutually recursive function */
1372
static int packed_object_info(struct packed_git *p, off_t offset,
1373
			      unsigned long *sizep);
J
Junio C Hamano 已提交
1374

N
Nicolas Pitre 已提交
1375
static int packed_delta_info(struct packed_git *p,
1376
			     struct pack_window **w_curs,
1377
			     off_t curpos,
1378
			     enum object_type type,
1379
			     off_t obj_offset,
N
Nicolas Pitre 已提交
1380
			     unsigned long *sizep)
1381
{
1382
	off_t base_offset;
J
Junio C Hamano 已提交
1383

1384 1385
	base_offset = get_delta_base(p, w_curs, &curpos, type, obj_offset);
	type = packed_object_info(p, base_offset, NULL);
J
Junio C Hamano 已提交
1386

1387 1388 1389 1390 1391
	/* We choose to only get the type of the base object and
	 * ignore potentially corrupt pack file that expects the delta
	 * based on a base with a wrong size.  This saves tons of
	 * inflate() calls.
	 */
N
Nicolas Pitre 已提交
1392 1393
	if (sizep)
		*sizep = get_size_from_delta(p, w_curs, curpos);
1394 1395

	return type;
1396 1397
}

1398 1399
static int unpack_object_header(struct packed_git *p,
				struct pack_window **w_curs,
1400
				off_t *curpos,
1401
				unsigned long *sizep)
1402
{
1403 1404
	unsigned char *base;
	unsigned int left;
1405
	unsigned long used;
1406
	enum object_type type;
1407

1408 1409
	/* use_pack() assures us we have [base, base + 20) available
	 * as a range that we can look at at.  (Its actually the hash
P
Pavel Roskin 已提交
1410
	 * size that is assured.)  With our object header encoding
1411 1412 1413
	 * the maximum deflated object size is 2^137, which is just
	 * insane, so we know won't exceed what we have been given.
	 */
1414
	base = use_pack(p, w_curs, *curpos, &left);
1415 1416 1417 1418 1419
	used = unpack_object_header_buffer(base, left, &type, sizep);
	if (!used) {
		type = OBJ_BAD;
	} else
		*curpos += used;
1420

1421
	return type;
1422 1423
}

1424
const char *packed_object_info_detail(struct packed_git *p,
1425
				      off_t obj_offset,
1426 1427 1428 1429
				      unsigned long *size,
				      unsigned long *store_size,
				      unsigned int *delta_chain_length,
				      unsigned char *base_sha1)
1430
{
1431
	struct pack_window *w_curs = NULL;
1432 1433
	off_t curpos;
	unsigned long dummy;
N
Nicolas Pitre 已提交
1434
	unsigned char *next_sha1;
1435
	enum object_type type;
1436
	struct revindex_entry *revidx;
1437

N
Nicolas Pitre 已提交
1438
	*delta_chain_length = 0;
1439
	curpos = obj_offset;
1440
	type = unpack_object_header(p, &w_curs, &curpos, size);
N
Nicolas Pitre 已提交
1441

1442 1443 1444
	revidx = find_pack_revindex(p, obj_offset);
	*store_size = revidx[1].offset - obj_offset;

N
Nicolas Pitre 已提交
1445
	for (;;) {
1446
		switch (type) {
N
Nicolas Pitre 已提交
1447
		default:
1448
			die("pack %s contains unknown object type %d",
1449
			    p->pack_name, type);
N
Nicolas Pitre 已提交
1450 1451 1452 1453
		case OBJ_COMMIT:
		case OBJ_TREE:
		case OBJ_BLOB:
		case OBJ_TAG:
1454
			unuse_pack(&w_curs);
1455
			return typename(type);
1456
		case OBJ_OFS_DELTA:
1457
			obj_offset = get_delta_base(p, &w_curs, &curpos, type, obj_offset);
1458 1459 1460
			if (!obj_offset)
				die("pack %s contains bad delta base reference of type %s",
				    p->pack_name, typename(type));
1461
			if (*delta_chain_length == 0) {
1462 1463
				revidx = find_pack_revindex(p, obj_offset);
				hashcpy(base_sha1, nth_packed_object_sha1(p, revidx->nr));
1464 1465 1466
			}
			break;
		case OBJ_REF_DELTA:
1467
			next_sha1 = use_pack(p, &w_curs, curpos, NULL);
N
Nicolas Pitre 已提交
1468 1469
			if (*delta_chain_length == 0)
				hashcpy(base_sha1, next_sha1);
1470
			obj_offset = find_pack_entry_one(next_sha1, p);
N
Nicolas Pitre 已提交
1471 1472 1473
			break;
		}
		(*delta_chain_length)++;
1474
		curpos = obj_offset;
1475
		type = unpack_object_header(p, &w_curs, &curpos, &dummy);
1476 1477 1478
	}
}

1479
static int packed_object_info(struct packed_git *p, off_t obj_offset,
1480
			      unsigned long *sizep)
1481
{
1482
	struct pack_window *w_curs = NULL;
1483 1484
	unsigned long size;
	off_t curpos = obj_offset;
1485
	enum object_type type;
1486

1487
	type = unpack_object_header(p, &w_curs, &curpos, &size);
1488

1489
	switch (type) {
1490 1491
	case OBJ_OFS_DELTA:
	case OBJ_REF_DELTA:
1492 1493
		type = packed_delta_info(p, &w_curs, curpos,
					 type, obj_offset, sizep);
1494
		break;
1495 1496 1497 1498
	case OBJ_COMMIT:
	case OBJ_TREE:
	case OBJ_BLOB:
	case OBJ_TAG:
1499 1500
		if (sizep)
			*sizep = size;
1501
		break;
1502
	default:
1503
		die("pack %s contains unknown object type %d",
1504
		    p->pack_name, type);
1505
	}
1506
	unuse_pack(&w_curs);
1507
	return type;
1508 1509
}

1510
static void *unpack_compressed_entry(struct packed_git *p,
1511
				    struct pack_window **w_curs,
1512
				    off_t curpos,
1513
				    unsigned long size)
1514 1515 1516
{
	int st;
	z_stream stream;
1517
	unsigned char *buffer, *in;
1518 1519 1520 1521 1522 1523 1524 1525

	buffer = xmalloc(size + 1);
	buffer[size] = 0;
	memset(&stream, 0, sizeof(stream));
	stream.next_out = buffer;
	stream.avail_out = size;

	inflateInit(&stream);
1526
	do {
1527
		in = use_pack(p, w_curs, curpos, &stream.avail_in);
1528 1529
		stream.next_in = in;
		st = inflate(&stream, Z_FINISH);
1530
		curpos += stream.next_in - in;
1531
	} while (st == Z_OK || st == Z_BUF_ERROR);
1532 1533 1534 1535 1536 1537 1538 1539 1540
	inflateEnd(&stream);
	if ((st != Z_STREAM_END) || stream.total_out != size) {
		free(buffer);
		return NULL;
	}

	return buffer;
}

1541 1542
#define MAX_DELTA_CACHE (256)

1543
static size_t delta_base_cached;
1544 1545 1546 1547 1548 1549

static struct delta_base_cache_lru_list {
	struct delta_base_cache_lru_list *prev;
	struct delta_base_cache_lru_list *next;
} delta_base_cache_lru = { &delta_base_cache_lru, &delta_base_cache_lru };

1550
static struct delta_base_cache_entry {
1551 1552
	struct delta_base_cache_lru_list lru;
	void *data;
1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564
	struct packed_git *p;
	off_t base_offset;
	unsigned long size;
	enum object_type type;
} delta_base_cache[MAX_DELTA_CACHE];

static unsigned long pack_entry_hash(struct packed_git *p, off_t base_offset)
{
	unsigned long hash;

	hash = (unsigned long)p + (unsigned long)base_offset;
	hash += (hash >> 8) + (hash >> 16);
1565
	return hash % MAX_DELTA_CACHE;
1566 1567
}

1568
static void *cache_or_unpack_entry(struct packed_git *p, off_t base_offset,
1569
	unsigned long *base_size, enum object_type *type, int keep_cache)
1570
{
1571 1572 1573 1574 1575
	void *ret;
	unsigned long hash = pack_entry_hash(p, base_offset);
	struct delta_base_cache_entry *ent = delta_base_cache + hash;

	ret = ent->data;
1576 1577
	if (!ret || ent->p != p || ent->base_offset != base_offset)
		return unpack_entry(p, base_offset, type, base_size);
1578

1579
	if (!keep_cache) {
1580
		ent->data = NULL;
1581 1582
		ent->lru.next->prev = ent->lru.prev;
		ent->lru.prev->next = ent->lru.next;
1583
		delta_base_cached -= ent->size;
P
Pierre Habouzit 已提交
1584 1585
	} else {
		ret = xmemdupz(ent->data, ent->size);
1586
	}
1587 1588 1589
	*type = ent->type;
	*base_size = ent->size;
	return ret;
1590 1591
}

1592 1593 1594 1595 1596
static inline void release_delta_base_cache(struct delta_base_cache_entry *ent)
{
	if (ent->data) {
		free(ent->data);
		ent->data = NULL;
1597 1598
		ent->lru.next->prev = ent->lru.prev;
		ent->lru.prev->next = ent->lru.next;
1599 1600 1601 1602
		delta_base_cached -= ent->size;
	}
}

1603 1604 1605
static void add_delta_base_cache(struct packed_git *p, off_t base_offset,
	void *base, unsigned long base_size, enum object_type type)
{
1606
	unsigned long hash = pack_entry_hash(p, base_offset);
1607
	struct delta_base_cache_entry *ent = delta_base_cache + hash;
1608
	struct delta_base_cache_lru_list *lru;
1609

1610 1611
	release_delta_base_cache(ent);
	delta_base_cached += base_size;
1612 1613 1614 1615 1616 1617

	for (lru = delta_base_cache_lru.next;
	     delta_base_cached > delta_base_cache_limit
	     && lru != &delta_base_cache_lru;
	     lru = lru->next) {
		struct delta_base_cache_entry *f = (void *)lru;
1618 1619 1620
		if (f->type == OBJ_BLOB)
			release_delta_base_cache(f);
	}
1621 1622 1623 1624 1625 1626 1627
	for (lru = delta_base_cache_lru.next;
	     delta_base_cached > delta_base_cache_limit
	     && lru != &delta_base_cache_lru;
	     lru = lru->next) {
		struct delta_base_cache_entry *f = (void *)lru;
		release_delta_base_cache(f);
	}
1628

1629 1630 1631 1632 1633
	ent->p = p;
	ent->base_offset = base_offset;
	ent->type = type;
	ent->data = base;
	ent->size = base_size;
1634 1635 1636 1637
	ent->lru.next = &delta_base_cache_lru;
	ent->lru.prev = delta_base_cache_lru.prev;
	delta_base_cache_lru.prev->next = &ent->lru;
	delta_base_cache_lru.prev = &ent->lru;
1638 1639
}

1640
static void *unpack_delta_entry(struct packed_git *p,
1641
				struct pack_window **w_curs,
1642
				off_t curpos,
1643
				unsigned long delta_size,
1644
				off_t obj_offset,
1645
				enum object_type *type,
1646
				unsigned long *sizep)
1647
{
1648
	void *delta_data, *result, *base;
1649 1650
	unsigned long base_size;
	off_t base_offset;
N
Nicolas Pitre 已提交
1651

1652
	base_offset = get_delta_base(p, w_curs, &curpos, *type, obj_offset);
1653 1654 1655 1656 1657 1658
	if (!base_offset) {
		error("failed to validate delta base reference "
		      "at offset %"PRIuMAX" from %s",
		      (uintmax_t)curpos, p->pack_name);
		return NULL;
	}
1659
	unuse_pack(w_curs);
1660
	base = cache_or_unpack_entry(p, base_offset, &base_size, type, 0);
1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675
	if (!base) {
		/*
		 * We're probably in deep shit, but let's try to fetch
		 * the required base anyway from another pack or loose.
		 * This is costly but should happen only in the presence
		 * of a corrupted pack, and is better than failing outright.
		 */
		struct revindex_entry *revidx = find_pack_revindex(p, base_offset);
		const unsigned char *base_sha1 =
					nth_packed_object_sha1(p, revidx->nr);
		error("failed to read delta base object %s"
		      " at offset %"PRIuMAX" from %s",
		      sha1_to_hex(base_sha1), (uintmax_t)base_offset,
		      p->pack_name);
		mark_bad_packed_object(p, base_sha1);
1676
		base = read_object(base_sha1, type, &base_size);
1677 1678 1679
		if (!base)
			return NULL;
	}
1680

1681
	delta_data = unpack_compressed_entry(p, w_curs, curpos, delta_size);
1682 1683 1684 1685 1686 1687 1688
	if (!delta_data) {
		error("failed to unpack compressed delta "
		      "at offset %"PRIuMAX" from %s",
		      (uintmax_t)curpos, p->pack_name);
		free(base);
		return NULL;
	}
1689 1690
	result = patch_delta(base, base_size,
			     delta_data, delta_size,
1691
			     sizep);
1692 1693 1694
	if (!result)
		die("failed to apply delta");
	free(delta_data);
1695
	add_delta_base_cache(p, base_offset, base, base_size, *type);
1696 1697 1698
	return result;
}

1699 1700
int do_check_packed_object_crc;

1701
void *unpack_entry(struct packed_git *p, off_t obj_offset,
1702
		   enum object_type *type, unsigned long *sizep)
1703
{
1704
	struct pack_window *w_curs = NULL;
1705
	off_t curpos = obj_offset;
1706
	void *data;
1707

1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720
	if (do_check_packed_object_crc && p->index_version > 1) {
		struct revindex_entry *revidx = find_pack_revindex(p, obj_offset);
		unsigned long len = revidx[1].offset - obj_offset;
		if (check_pack_crc(p, &w_curs, obj_offset, len, revidx->nr)) {
			const unsigned char *sha1 =
				nth_packed_object_sha1(p, revidx->nr);
			error("bad packed object CRC for %s",
			      sha1_to_hex(sha1));
			mark_bad_packed_object(p, sha1);
			return NULL;
		}
	}

1721 1722
	*type = unpack_object_header(p, &w_curs, &curpos, sizep);
	switch (*type) {
1723 1724
	case OBJ_OFS_DELTA:
	case OBJ_REF_DELTA:
1725 1726
		data = unpack_delta_entry(p, &w_curs, curpos, *sizep,
					  obj_offset, type, sizep);
1727
		break;
1728 1729 1730 1731
	case OBJ_COMMIT:
	case OBJ_TREE:
	case OBJ_BLOB:
	case OBJ_TAG:
1732
		data = unpack_compressed_entry(p, &w_curs, curpos, *sizep);
1733
		break;
1734
	default:
1735 1736 1737
		data = NULL;
		error("unknown object type %i at offset %"PRIuMAX" in %s",
		      *type, (uintmax_t)obj_offset, p->pack_name);
1738
	}
1739
	unuse_pack(&w_curs);
1740
	return data;
1741 1742
}

1743
const unsigned char *nth_packed_object_sha1(struct packed_git *p,
1744
					    uint32_t n)
1745
{
1746
	const unsigned char *index = p->index_data;
1747 1748 1749 1750 1751
	if (!index) {
		if (open_pack_index(p))
			return NULL;
		index = p->index_data;
	}
N
Nicolas Pitre 已提交
1752
	if (n >= p->num_objects)
1753
		return NULL;
1754 1755 1756 1757 1758 1759 1760 1761 1762
	index += 4 * 256;
	if (p->index_version == 1) {
		return index + 24 * n + 4;
	} else {
		index += 8;
		return index + 20 * n;
	}
}

N
Nicolas Pitre 已提交
1763
off_t nth_packed_object_offset(const struct packed_git *p, uint32_t n)
1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778
{
	const unsigned char *index = p->index_data;
	index += 4 * 256;
	if (p->index_version == 1) {
		return ntohl(*((uint32_t *)(index + 24 * n)));
	} else {
		uint32_t off;
		index += 8 + p->num_objects * (20 + 4);
		off = ntohl(*((uint32_t *)(index + 4 * n)));
		if (!(off & 0x80000000))
			return off;
		index += p->num_objects * 4 + (off & 0x7fffffff) * 8;
		return (((uint64_t)ntohl(*((uint32_t *)(index + 0)))) << 32) |
				   ntohl(*((uint32_t *)(index + 4)));
	}
1779 1780
}

1781
off_t find_pack_entry_one(const unsigned char *sha1,
N
Nicolas Pitre 已提交
1782
				  struct packed_git *p)
1783
{
1784 1785
	const uint32_t *level1_ofs = p->index_data;
	const unsigned char *index = p->index_data;
1786 1787 1788 1789 1790 1791
	unsigned hi, lo, stride;
	static int use_lookup = -1;
	static int debug_lookup = -1;

	if (debug_lookup < 0)
		debug_lookup = !!getenv("GIT_DEBUG_LOOKUP");
1792

1793 1794 1795 1796 1797 1798
	if (!index) {
		if (open_pack_index(p))
			return 0;
		level1_ofs = p->index_data;
		index = p->index_data;
	}
1799 1800 1801 1802
	if (p->index_version > 1) {
		level1_ofs += 2;
		index += 8;
	}
1803
	index += 4 * 256;
1804 1805
	hi = ntohl(level1_ofs[*sha1]);
	lo = ((*sha1 == 0x0) ? 0 : ntohl(level1_ofs[*sha1 - 1]));
1806 1807 1808 1809 1810 1811 1812 1813
	if (p->index_version > 1) {
		stride = 20;
	} else {
		stride = 24;
		index += 4;
	}

	if (debug_lookup)
1814
		printf("%02x%02x%02x... lo %u hi %u nr %"PRIu32"\n",
1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825
		       sha1[0], sha1[1], sha1[2], lo, hi, p->num_objects);

	if (use_lookup < 0)
		use_lookup = !!getenv("GIT_USE_LOOKUP");
	if (use_lookup) {
		int pos = sha1_entry_pos(index, stride, 0,
					 lo, hi, p->num_objects, sha1);
		if (pos < 0)
			return 0;
		return nth_packed_object_offset(p, pos);
	}
1826 1827

	do {
1828
		unsigned mi = (lo + hi) / 2;
1829 1830 1831 1832 1833
		int cmp = hashcmp(index + mi * stride, sha1);

		if (debug_lookup)
			printf("lo %u hi %u rg %u mi %u\n",
			       lo, hi, hi - lo, mi);
N
Nicolas Pitre 已提交
1834
		if (!cmp)
1835
			return nth_packed_object_offset(p, mi);
1836 1837 1838 1839 1840 1841 1842 1843
		if (cmp > 0)
			hi = mi;
		else
			lo = mi+1;
	} while (lo < hi);
	return 0;
}

1844
int matches_pack_name(struct packed_git *p, const char *name)
1845 1846 1847
{
	const char *last_c, *c;

1848 1849
	if (!strcmp(p->pack_name, name))
		return 1;
1850 1851 1852 1853 1854 1855

	for (c = p->pack_name, last_c = c; *c;)
		if (*c == '/')
			last_c = ++c;
		else
			++c;
1856 1857
	if (!strcmp(last_c, name))
		return 1;
1858

1859
	return 0;
1860 1861
}

1862
static int find_pack_entry(const unsigned char *sha1, struct pack_entry *e, const char **ignore_packed)
1863
{
1864
	static struct packed_git *last_found = (void *)1;
1865
	struct packed_git *p;
1866
	off_t offset;
N
Nicolas Pitre 已提交
1867

1868
	prepare_packed_git();
1869 1870 1871
	if (!packed_git)
		return 0;
	p = (last_found == (void *)1) ? packed_git : last_found;
1872

1873
	do {
1874 1875 1876
		if (ignore_packed) {
			const char **ig;
			for (ig = ignore_packed; *ig; ig++)
1877
				if (matches_pack_name(p, *ig))
1878 1879
					break;
			if (*ig)
1880
				goto next;
1881
		}
1882

1883 1884 1885 1886 1887 1888 1889
		if (p->num_bad_objects) {
			unsigned i;
			for (i = 0; i < p->num_bad_objects; i++)
				if (!hashcmp(sha1, p->bad_object_sha1 + 20 * i))
					goto next;
		}

N
Nicolas Pitre 已提交
1890 1891
		offset = find_pack_entry_one(sha1, p);
		if (offset) {
1892 1893 1894 1895 1896 1897 1898 1899 1900 1901
			/*
			 * We are about to tell the caller where they can
			 * locate the requested object.  We better make
			 * sure the packfile is still here and can be
			 * accessed before supplying that answer, as
			 * it may have been deleted since the index
			 * was loaded!
			 */
			if (p->pack_fd == -1 && open_packed_git(p)) {
				error("packfile %s cannot be accessed", p->pack_name);
1902
				goto next;
1903
			}
N
Nicolas Pitre 已提交
1904 1905 1906
			e->offset = offset;
			e->p = p;
			hashcpy(e->sha1, sha1);
1907
			last_found = p;
1908
			return 1;
N
Nicolas Pitre 已提交
1909
		}
1910 1911 1912 1913 1914 1915 1916 1917 1918

		next:
		if (p == last_found)
			p = packed_git;
		else
			p = p->next;
		if (p == last_found)
			p = p->next;
	} while (p);
1919 1920 1921
	return 0;
}

J
Junio C Hamano 已提交
1922
struct packed_git *find_sha1_pack(const unsigned char *sha1,
1923 1924 1925 1926 1927
				  struct packed_git *packs)
{
	struct packed_git *p;

	for (p = packs; p; p = p->next) {
N
Nicolas Pitre 已提交
1928
		if (find_pack_entry_one(sha1, p))
1929 1930 1931
			return p;
	}
	return NULL;
1932

1933 1934
}

1935
static int sha1_loose_object_info(const unsigned char *sha1, unsigned long *sizep)
1936
{
1937
	int status;
1938 1939 1940
	unsigned long mapsize, size;
	void *map;
	z_stream stream;
N
Nicolas Pitre 已提交
1941
	char hdr[32];
1942

1943
	map = map_sha1_file(sha1, &mapsize);
1944 1945
	if (!map)
		return error("unable to find %s", sha1_to_hex(sha1));
1946 1947 1948
	if (unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)) < 0)
		status = error("unable to unpack %s header",
			       sha1_to_hex(sha1));
1949
	else if ((status = parse_sha1_header(hdr, &size)) < 0)
1950
		status = error("unable to parse %s header", sha1_to_hex(sha1));
1951 1952
	else if (sizep)
		*sizep = size;
1953 1954 1955 1956 1957
	inflateEnd(&stream);
	munmap(map, mapsize);
	return status;
}

1958
int sha1_object_info(const unsigned char *sha1, unsigned long *sizep)
1959 1960
{
	struct pack_entry e;
1961
	int status;
1962 1963

	if (!find_pack_entry(sha1, &e, NULL)) {
1964 1965 1966 1967 1968 1969
		/* Most likely it's a loose object. */
		status = sha1_loose_object_info(sha1, sizep);
		if (status >= 0)
			return status;

		/* Not a loose object; someone else may have just packed it. */
1970 1971
		reprepare_packed_git();
		if (!find_pack_entry(sha1, &e, NULL))
1972
			return status;
1973
	}
1974
	return packed_object_info(e.p, e.offset, sizep);
1975 1976
}

1977 1978
static void *read_packed_sha1(const unsigned char *sha1,
			      enum object_type *type, unsigned long *size)
1979 1980
{
	struct pack_entry e;
1981
	void *data;
1982

1983
	if (!find_pack_entry(sha1, &e, NULL))
1984
		return NULL;
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995
	data = cache_or_unpack_entry(e.p, e.offset, size, type, 1);
	if (!data) {
		/*
		 * We're probably in deep shit, but let's try to fetch
		 * the required object anyway from another pack or loose.
		 * This should happen only in the presence of a corrupted
		 * pack, and is better than failing outright.
		 */
		error("failed to read object %s at offset %"PRIuMAX" from %s",
		      sha1_to_hex(sha1), (uintmax_t)e.offset, e.p->pack_name);
		mark_bad_packed_object(e.p, sha1);
1996
		data = read_object(sha1, type, size);
1997 1998
	}
	return data;
1999 2000
}

2001 2002 2003 2004 2005 2006 2007 2008
/*
 * This is meant to hold a *small* number of objects that you would
 * want read_sha1_file() to be able to return, but yet you do not want
 * to write them into the object store (e.g. a browse-only
 * application).
 */
static struct cached_object {
	unsigned char sha1[20];
2009
	enum object_type type;
2010 2011 2012 2013 2014
	void *buf;
	unsigned long size;
} *cached_objects;
static int cached_object_nr, cached_object_alloc;

J
Jeff King 已提交
2015 2016 2017 2018 2019 2020 2021 2022 2023
static struct cached_object empty_tree = {
	/* empty tree sha1: 4b825dc642cb6eb9a060e54bf8d69288fbee4904 */
	"\x4b\x82\x5d\xc6\x42\xcb\x6e\xb9\xa0\x60"
	"\xe5\x4b\xf8\xd6\x92\x88\xfb\xee\x49\x04",
	OBJ_TREE,
	"",
	0
};

2024 2025 2026 2027 2028 2029 2030 2031 2032
static struct cached_object *find_cached_object(const unsigned char *sha1)
{
	int i;
	struct cached_object *co = cached_objects;

	for (i = 0; i < cached_object_nr; i++, co++) {
		if (!hashcmp(co->sha1, sha1))
			return co;
	}
J
Jeff King 已提交
2033 2034
	if (!hashcmp(sha1, empty_tree.sha1))
		return &empty_tree;
2035 2036 2037
	return NULL;
}

2038 2039
int pretend_sha1_file(void *buf, unsigned long len, enum object_type type,
		      unsigned char *sha1)
2040 2041 2042
{
	struct cached_object *co;

2043
	hash_sha1_file(buf, len, typename(type), sha1);
2044 2045 2046 2047 2048 2049 2050 2051 2052 2053
	if (has_sha1_file(sha1) || find_cached_object(sha1))
		return 0;
	if (cached_object_alloc <= cached_object_nr) {
		cached_object_alloc = alloc_nr(cached_object_alloc);
		cached_objects = xrealloc(cached_objects,
					  sizeof(*cached_objects) *
					  cached_object_alloc);
	}
	co = &cached_objects[cached_object_nr++];
	co->size = len;
2054
	co->type = type;
J
Junio C Hamano 已提交
2055 2056
	co->buf = xmalloc(len);
	memcpy(co->buf, buf, len);
2057 2058 2059 2060
	hashcpy(co->sha1, sha1);
	return 0;
}

2061 2062
void *read_object(const unsigned char *sha1, enum object_type *type,
		  unsigned long *size)
2063 2064 2065
{
	unsigned long mapsize;
	void *map, *buf;
2066 2067 2068 2069
	struct cached_object *co;

	co = find_cached_object(sha1);
	if (co) {
2070
		*type = co->type;
2071
		*size = co->size;
P
Pierre Habouzit 已提交
2072
		return xmemdupz(co->buf, co->size);
2073
	}
2074

2075 2076 2077
	buf = read_packed_sha1(sha1, type, size);
	if (buf)
		return buf;
2078
	map = map_sha1_file(sha1, &mapsize);
2079
	if (map) {
2080
		buf = unpack_sha1_file(map, mapsize, type, size, sha1);
2081 2082 2083
		munmap(map, mapsize);
		return buf;
	}
2084
	reprepare_packed_git();
2085
	return read_packed_sha1(sha1, type, size);
2086 2087
}

2088 2089 2090 2091 2092 2093 2094 2095 2096 2097
void *read_sha1_file(const unsigned char *sha1, enum object_type *type,
		     unsigned long *size)
{
	void *data = read_object(sha1, type, size);
	/* legacy behavior is to die on corrupted objects */
	if (!data && (has_loose_object(sha1) || has_packed_and_bad(sha1)))
		die("object %s is corrupted", sha1_to_hex(sha1));
	return data;
}

2098
void *read_object_with_reference(const unsigned char *sha1,
2099
				 const char *required_type_name,
2100 2101
				 unsigned long *size,
				 unsigned char *actual_sha1_return)
2102
{
2103
	enum object_type type, required_type;
2104 2105
	void *buffer;
	unsigned long isize;
2106
	unsigned char actual_sha1[20];
2107

2108
	required_type = type_from_string(required_type_name);
2109
	hashcpy(actual_sha1, sha1);
2110 2111 2112
	while (1) {
		int ref_length = -1;
		const char *ref_type = NULL;
2113

2114
		buffer = read_sha1_file(actual_sha1, &type, &isize);
2115 2116
		if (!buffer)
			return NULL;
2117
		if (type == required_type) {
2118 2119
			*size = isize;
			if (actual_sha1_return)
2120
				hashcpy(actual_sha1_return, actual_sha1);
2121 2122 2123
			return buffer;
		}
		/* Handle references */
2124
		else if (type == OBJ_COMMIT)
2125
			ref_type = "tree ";
2126
		else if (type == OBJ_TAG)
2127 2128 2129 2130 2131 2132
			ref_type = "object ";
		else {
			free(buffer);
			return NULL;
		}
		ref_length = strlen(ref_type);
2133

2134 2135
		if (ref_length + 40 > isize ||
		    memcmp(buffer, ref_type, ref_length) ||
2136
		    get_sha1_hex((char *) buffer + ref_length, actual_sha1)) {
2137 2138 2139
			free(buffer);
			return NULL;
		}
2140
		free(buffer);
2141 2142
		/* Now we have the ID of the referred-to object in
		 * actual_sha1.  Check again. */
2143 2144 2145
	}
}

N
Nicolas Pitre 已提交
2146
static void write_sha1_file_prepare(const void *buf, unsigned long len,
R
Rene Scharfe 已提交
2147
                                    const char *type, unsigned char *sha1,
N
Nicolas Pitre 已提交
2148
                                    char *hdr, int *hdrlen)
2149
{
2150
	git_SHA_CTX c;
2151 2152

	/* Generate the header */
N
Nicolas Pitre 已提交
2153
	*hdrlen = sprintf(hdr, "%s %lu", type, len)+1;
2154 2155

	/* Sha1.. */
2156 2157 2158 2159
	git_SHA1_Init(&c);
	git_SHA1_Update(&c, hdr, *hdrlen);
	git_SHA1_Update(&c, buf, len);
	git_SHA1_Final(sha1, &c);
2160 2161
}

2162 2163 2164
/*
 * Move the just written object into its final resting place
 */
2165
int move_temp_to_file(const char *tmpfile, const char *filename)
2166
{
2167 2168 2169
	int ret = 0;
	if (link(tmpfile, filename))
		ret = errno;
2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183

	/*
	 * Coda hack - coda doesn't like cross-directory links,
	 * so we fall back to a rename, which will mean that it
	 * won't be able to check collisions, but that's not a
	 * big deal.
	 *
	 * The same holds for FAT formatted media.
	 *
	 * When this succeeds, we just return 0. We have nothing
	 * left to unlink.
	 */
	if (ret && ret != EEXIST) {
		if (!rename(tmpfile, filename))
2184
			return 0;
2185
		ret = errno;
2186 2187 2188 2189
	}
	unlink(tmpfile);
	if (ret) {
		if (ret != EEXIST) {
2190
			return error("unable to write sha1 filename %s: %s\n", filename, strerror(ret));
2191 2192 2193 2194 2195 2196 2197
		}
		/* FIXME!!! Collision check here ? */
	}

	return 0;
}

L
Linus Torvalds 已提交
2198 2199
static int write_buffer(int fd, const void *buf, size_t len)
{
L
Linus Torvalds 已提交
2200
	if (write_in_full(fd, buf, len) < 0)
2201
		return error("file write error (%s)", strerror(errno));
L
Linus Torvalds 已提交
2202 2203 2204
	return 0;
}

N
Nicolas Pitre 已提交
2205
int hash_sha1_file(const void *buf, unsigned long len, const char *type,
R
Rene Scharfe 已提交
2206 2207
                   unsigned char *sha1)
{
N
Nicolas Pitre 已提交
2208
	char hdr[32];
R
Rene Scharfe 已提交
2209 2210 2211 2212 2213
	int hdrlen;
	write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen);
	return 0;
}

2214 2215 2216
/* Finalize a file on disk, and close it. */
static void close_sha1_file(int fd)
{
2217 2218
	if (fsync_object_files)
		fsync_or_die(fd, "sha1 file");
2219 2220 2221 2222 2223
	fchmod(fd, 0444);
	if (close(fd) != 0)
		die("unable to write sha1 file");
}

2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252
/* Size of directory component, including the ending '/' */
static inline int directory_size(const char *filename)
{
	const char *s = strrchr(filename, '/');
	if (!s)
		return 0;
	return s - filename + 1;
}

/*
 * This creates a temporary file in the same directory as the final
 * 'filename'
 *
 * We want to avoid cross-directory filename renames, because those
 * can have problems on various filesystems (FAT, NFS, Coda).
 */
static int create_tmpfile(char *buffer, size_t bufsiz, const char *filename)
{
	int fd, dirlen = directory_size(filename);

	if (dirlen + 20 > bufsiz) {
		errno = ENAMETOOLONG;
		return -1;
	}
	memcpy(buffer, filename, dirlen);
	strcpy(buffer + dirlen, "tmp_obj_XXXXXX");
	fd = mkstemp(buffer);
	if (fd < 0 && dirlen) {
		/* Make sure the directory exists */
P
Patrick Higgins 已提交
2253
		memcpy(buffer, filename, dirlen);
2254
		buffer[dirlen-1] = 0;
2255
		if (mkdir(buffer, 0777) || adjust_shared_perm(buffer))
2256 2257 2258 2259 2260 2261 2262 2263 2264
			return -1;

		/* Try again */
		strcpy(buffer + dirlen - 1, "/tmp_obj_XXXXXX");
		fd = mkstemp(buffer);
	}
	return fd;
}

2265 2266
static int write_loose_object(const unsigned char *sha1, char *hdr, int hdrlen,
			      void *buf, unsigned long len, time_t mtime)
2267
{
2268
	int fd, size, ret;
2269
	unsigned char *compressed;
2270
	z_stream stream;
2271
	char *filename;
2272
	static char tmpfile[PATH_MAX];
2273

R
Rene Scharfe 已提交
2274
	filename = sha1_file_name(sha1);
2275
	fd = create_tmpfile(tmpfile, sizeof(tmpfile), filename);
2276
	if (fd < 0) {
2277 2278 2279 2280
		if (errno == EPERM)
			return error("insufficient permission for adding an object to repository database %s\n", get_object_directory());
		else
			return error("unable to create temporary sha1 filename %s: %s\n", tmpfile, strerror(errno));
2281 2282
	}

2283 2284
	/* Set it up */
	memset(&stream, 0, sizeof(stream));
2285
	deflateInit(&stream, zlib_compression_level);
2286
	size = 8 + deflateBound(&stream, len+hdrlen);
2287
	compressed = xmalloc(size);
2288 2289 2290 2291

	/* Compress it */
	stream.next_out = compressed;
	stream.avail_out = size;
2292 2293

	/* First header.. */
N
Nicolas Pitre 已提交
2294
	stream.next_in = (unsigned char *)hdr;
2295
	stream.avail_in = hdrlen;
2296 2297
	while (deflate(&stream, 0) == Z_OK)
		/* nothing */;
2298 2299 2300 2301

	/* Then the data itself.. */
	stream.next_in = buf;
	stream.avail_in = len;
2302 2303 2304 2305 2306 2307 2308 2309
	ret = deflate(&stream, Z_FINISH);
	if (ret != Z_STREAM_END)
		die("unable to deflate new object %s (%d)", sha1_to_hex(sha1), ret);

	ret = deflateEnd(&stream);
	if (ret != Z_OK)
		die("deflateEnd on object %s failed (%d)", sha1_to_hex(sha1), ret);

2310 2311
	size = stream.total_out;

L
Linus Torvalds 已提交
2312 2313
	if (write_buffer(fd, compressed, size) < 0)
		die("unable to write sha1 file");
2314
	close_sha1_file(fd);
2315
	free(compressed);
2316

2317 2318 2319 2320 2321 2322 2323 2324 2325
	if (mtime) {
		struct utimbuf utb;
		utb.actime = mtime;
		utb.modtime = mtime;
		if (utime(tmpfile, &utb) < 0)
			warning("failed utime() on %s: %s",
				tmpfile, strerror(errno));
	}

2326
	return move_temp_to_file(tmpfile, filename);
2327
}
2328

2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352
int write_sha1_file(void *buf, unsigned long len, const char *type, unsigned char *returnsha1)
{
	unsigned char sha1[20];
	char hdr[32];
	int hdrlen;

	/* Normally if we have it in the pack then we do not bother writing
	 * it out into .git/objects/??/?{38} file.
	 */
	write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen);
	if (returnsha1)
		hashcpy(returnsha1, sha1);
	if (has_sha1_file(sha1))
		return 0;
	return write_loose_object(sha1, hdr, hdrlen, buf, len, 0);
}

int force_object_loose(const unsigned char *sha1, time_t mtime)
{
	void *buf;
	unsigned long len;
	enum object_type type;
	char hdr[32];
	int hdrlen;
2353
	int ret;
2354

2355
	if (has_loose_object(sha1))
2356 2357 2358 2359 2360
		return 0;
	buf = read_packed_sha1(sha1, &type, &len);
	if (!buf)
		return error("cannot read sha1_file for %s", sha1_to_hex(sha1));
	hdrlen = sprintf(hdr, "%s %lu", typename(type), len) + 1;
2361 2362 2363 2364
	ret = write_loose_object(sha1, hdr, hdrlen, buf, len, mtime);
	free(buf);

	return ret;
2365 2366
}

2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382
int has_pack_index(const unsigned char *sha1)
{
	struct stat st;
	if (stat(sha1_pack_index_name(sha1), &st))
		return 0;
	return 1;
}

int has_pack_file(const unsigned char *sha1)
{
	struct stat st;
	if (stat(sha1_pack_name(sha1), &st))
		return 0;
	return 1;
}

2383
int has_sha1_pack(const unsigned char *sha1, const char **ignore_packed)
2384 2385
{
	struct pack_entry e;
2386
	return find_pack_entry(sha1, &e, ignore_packed);
2387 2388
}

2389 2390
int has_sha1_file(const unsigned char *sha1)
{
2391 2392
	struct pack_entry e;

2393
	if (find_pack_entry(sha1, &e, NULL))
2394
		return 1;
2395
	return has_loose_object(sha1);
2396
}
J
Junio C Hamano 已提交
2397

2398 2399
static int index_mem(unsigned char *sha1, void *buf, size_t size,
		     int write_object, enum object_type type, const char *path)
2400
{
L
Linus Torvalds 已提交
2401
	int ret, re_allocated = 0;
J
Junio C Hamano 已提交
2402

2403
	if (!type)
2404
		type = OBJ_BLOB;
L
Linus Torvalds 已提交
2405 2406 2407 2408

	/*
	 * Convert blobs to git internal format
	 */
2409
	if ((type == OBJ_BLOB) && path) {
2410
		struct strbuf nbuf = STRBUF_INIT;
2411 2412
		if (convert_to_git(path, buf, size, &nbuf,
		                   write_object ? safe_crlf : 0)) {
2413
			buf = strbuf_detach(&nbuf, &size);
L
Linus Torvalds 已提交
2414 2415 2416 2417
			re_allocated = 1;
		}
	}

2418
	if (write_object)
2419
		ret = write_sha1_file(buf, size, typename(type), sha1);
R
Rene Scharfe 已提交
2420
	else
2421
		ret = hash_sha1_file(buf, size, typename(type), sha1);
2422
	if (re_allocated)
L
Linus Torvalds 已提交
2423
		free(buf);
2424 2425 2426 2427 2428 2429 2430 2431 2432 2433
	return ret;
}

int index_fd(unsigned char *sha1, int fd, struct stat *st, int write_object,
	     enum object_type type, const char *path)
{
	int ret;
	size_t size = xsize_t(st->st_size);

	if (!S_ISREG(st->st_mode)) {
2434
		struct strbuf sbuf = STRBUF_INIT;
2435 2436 2437 2438 2439 2440 2441 2442 2443
		if (strbuf_read(&sbuf, fd, 4096) >= 0)
			ret = index_mem(sha1, sbuf.buf, sbuf.len, write_object,
					type, path);
		else
			ret = -1;
		strbuf_release(&sbuf);
	} else if (size) {
		void *buf = xmmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
		ret = index_mem(sha1, buf, size, write_object, type, path);
2444
		munmap(buf, size);
2445 2446 2447
	} else
		ret = index_mem(sha1, NULL, size, write_object, type, path);
	close(fd);
2448
	return ret;
J
Junio C Hamano 已提交
2449
}
2450 2451 2452 2453 2454

int index_path(unsigned char *sha1, const char *path, struct stat *st, int write_object)
{
	int fd;
	char *target;
2455
	size_t len;
2456 2457 2458 2459 2460 2461 2462

	switch (st->st_mode & S_IFMT) {
	case S_IFREG:
		fd = open(path, O_RDONLY);
		if (fd < 0)
			return error("open(\"%s\"): %s", path,
				     strerror(errno));
2463
		if (index_fd(sha1, fd, st, write_object, OBJ_BLOB, path) < 0)
2464 2465 2466 2467
			return error("%s: failed to insert into database",
				     path);
		break;
	case S_IFLNK:
2468 2469 2470
		len = xsize_t(st->st_size);
		target = xmalloc(len + 1);
		if (readlink(path, target, len + 1) != st->st_size) {
2471 2472 2473 2474 2475
			char *errstr = strerror(errno);
			free(target);
			return error("readlink(\"%s\"): %s", path,
			             errstr);
		}
R
Rene Scharfe 已提交
2476
		if (!write_object)
2477 2478
			hash_sha1_file(target, len, blob_type, sha1);
		else if (write_sha1_file(target, len, blob_type, sha1))
2479 2480 2481 2482
			return error("%s: failed to insert into database",
				     path);
		free(target);
		break;
2483 2484
	case S_IFDIR:
		return resolve_gitlink_ref(path, "HEAD", sha1);
2485 2486 2487 2488 2489
	default:
		return error("%s: unsupported file type", path);
	}
	return 0;
}
2490 2491 2492

int read_pack_header(int fd, struct pack_header *header)
{
2493 2494 2495 2496
	if (read_in_full(fd, header, sizeof(*header)) < sizeof(*header))
		/* "eof before pack header was fully read" */
		return PH_ERROR_EOF;

2497 2498 2499 2500 2501 2502 2503 2504
	if (header->hdr_signature != htonl(PACK_SIGNATURE))
		/* "protocol error (pack signature mismatch detected)" */
		return PH_ERROR_PACK_SIGNATURE;
	if (!pack_version_ok(header->hdr_version))
		/* "protocol error (pack version unsupported)" */
		return PH_ERROR_PROTOCOL;
	return 0;
}