sha1_file.c 51.9 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

17 18 19 20 21 22 23 24
#ifndef O_NOATIME
#if defined(__linux__) && (defined(__i386__) || defined(__PPC__))
#define O_NOATIME 01000000
#else
#define O_NOATIME 0
#endif
#endif

25 26 27 28 29 30
#ifdef NO_C99_FORMAT
#define SZ_FMT "lu"
#else
#define SZ_FMT "zu"
#endif

31
const unsigned char null_sha1[20];
J
Junio C Hamano 已提交
32

33 34
static unsigned int sha1_file_open_flag = O_NOATIME;

35 36 37 38 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
signed char hexval_table[256] = {
	 -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 */
};
69 70 71 72 73 74 75 76 77 78 79 80 81 82

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;
}

83 84 85
int safe_create_leading_directories(char *path)
{
	char *pos = path;
86 87
	struct stat st;

88 89
	if (*pos == '/')
		pos++;
90 91 92 93 94 95

	while (pos) {
		pos = strchr(pos, '/');
		if (!pos)
			break;
		*pos = 0;
96 97 98
		if (!stat(path, &st)) {
			/* path exists */
			if (!S_ISDIR(st.st_mode)) {
99
				*pos = '/';
100
				return -3;
101
			}
102
		}
103 104 105 106
		else if (mkdir(path, 0777)) {
			*pos = '/';
			return -1;
		}
107 108 109 110
		else if (adjust_shared_perm(path)) {
			*pos = '/';
			return -2;
		}
111 112 113 114
		*pos++ = '/';
	}
	return 0;
}
115

116 117
char * sha1_to_hex(const unsigned char *sha1)
{
L
Linus Torvalds 已提交
118 119
	static int bufno;
	static char hexbuffer[4][50];
120
	static const char hex[] = "0123456789abcdef";
L
Linus Torvalds 已提交
121
	char *buffer = hexbuffer[3 & ++bufno], *buf = buffer;
122 123 124 125 126 127 128
	int i;

	for (i = 0; i < 20; i++) {
		unsigned int val = *sha1++;
		*buf++ = hex[val >> 4];
		*buf++ = hex[val & 0xf];
	}
129 130
	*buf = '\0';

131 132 133
	return buffer;
}

134 135 136 137 138 139 140 141 142 143 144 145
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];
	}
}

146 147
/*
 * NOTE! This returns a statically allocated buffer, so you have to be
148
 * careful about using it. Do a "xstrdup()" if you need to save the
149
 * filename.
150 151 152
 *
 * 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
153
 * DB_ENVIRONMENT environment variable if it is not found in
154
 * the primary object database.
155 156 157 158 159 160
 */
char *sha1_file_name(const unsigned char *sha1)
{
	static char *name, *base;

	if (!base) {
J
Junio C Hamano 已提交
161
		const char *sha1_file_directory = get_object_directory();
162
		int len = strlen(sha1_file_directory);
163
		base = xmalloc(len + 60);
164 165 166 167 168 169
		memcpy(base, sha1_file_directory, len);
		memset(base+len, 0, 60);
		base[len] = '/';
		base[len+3] = '/';
		name = base + len + 1;
	}
170
	fill_sha1_path(name, sha1);
171 172 173
	return base;
}

174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223
char *sha1_pack_name(const unsigned char *sha1)
{
	static const char hex[] = "0123456789abcdef";
	static char *name, *base, *buf;
	int i;

	if (!base) {
		const char *sha1_file_directory = get_object_directory();
		int len = strlen(sha1_file_directory);
		base = xmalloc(len + 60);
		sprintf(base, "%s/pack/pack-1234567890123456789012345678901234567890.pack", sha1_file_directory);
		name = base + len + 11;
	}

	buf = name;

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

char *sha1_pack_index_name(const unsigned char *sha1)
{
	static const char hex[] = "0123456789abcdef";
	static char *name, *base, *buf;
	int i;

	if (!base) {
		const char *sha1_file_directory = get_object_directory();
		int len = strlen(sha1_file_directory);
		base = xmalloc(len + 60);
		sprintf(base, "%s/pack/pack-1234567890123456789012345678901234567890.idx", sha1_file_directory);
		name = base + len + 11;
	}

	buf = name;

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

224 225
struct alternate_object_database *alt_odb_list;
static struct alternate_object_database **alt_odb_tail;
226

M
Martin Waitz 已提交
227 228
static void read_info_alternates(const char * alternates, int depth);

J
Junio C Hamano 已提交
229 230
/*
 * Prepare alternate object database registry.
231 232 233 234 235
 *
 * 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,
236 237
 * whose contents is similar to that environment variable but can be
 * LF separated.  Its base points at a statically allocated buffer that
238 239 240 241 242
 * 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
243
 */
M
Martin Waitz 已提交
244
static int link_alt_odb_entry(const char * entry, int len, const char * relative_base, int depth)
245
{
M
Martin Waitz 已提交
246
	struct stat st;
247
	const char *objdir = get_object_directory();
M
Martin Waitz 已提交
248 249 250 251 252
	struct alternate_object_database *ent;
	struct alternate_object_database *alt;
	/* 43 = 40-byte + 2 '/' + terminating NUL */
	int pfxlen = len;
	int entlen = pfxlen + 43;
253
	int base_len = -1;
254

M
Martin Waitz 已提交
255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322
	if (*entry != '/' && relative_base) {
		/* Relative alt-odb */
		if (base_len < 0)
			base_len = strlen(relative_base) + 1;
		entlen += base_len;
		pfxlen += base_len;
	}
	ent = xmalloc(sizeof(*ent) + entlen);

	if (*entry != '/' && relative_base) {
		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 */
	if (stat(ent->base, &st) || !S_ISDIR(st.st_mode)) {
		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;
	}

323
	last = alt;
324 325 326 327 328 329 330 331
	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
332 333
		while (cp < ep && *cp != sep)
			cp++;
334
		if (last != cp) {
M
Martin Waitz 已提交
335 336 337 338 339 340
			if ((*last != '/') && depth) {
				error("%s: ignoring relative alternate object store %s",
						relative_base, last);
			} else {
				link_alt_odb_entry(last, cp - last,
						relative_base, depth);
341
			}
342
		}
343
		while (cp < ep && *cp == sep)
344 345
			cp++;
		last = cp;
346
	}
347 348
}

M
Martin Waitz 已提交
349
static void read_info_alternates(const char * relative_base, int depth)
350
{
351
	char *map;
352
	struct stat st;
M
Martin Waitz 已提交
353 354
	char path[PATH_MAX];
	int fd;
355

M
Martin Waitz 已提交
356
	sprintf(path, "%s/info/alternates", relative_base);
357 358 359 360 361
	fd = open(path, O_RDONLY);
	if (fd < 0)
		return;
	if (fstat(fd, &st) || (st.st_size == 0)) {
		close(fd);
362
		return;
363
	}
364
	map = xmmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
365 366
	close(fd);

M
Martin Waitz 已提交
367 368
	link_alt_odb_entries(map, map + st.st_size, '\n', relative_base, depth);

369
	munmap(map, st.st_size);
370 371
}

M
Martin Waitz 已提交
372 373
void prepare_alt_odb(void)
{
T
Timo Hirvonen 已提交
374
	const char *alt;
M
Martin Waitz 已提交
375 376 377 378 379 380 381 382 383 384 385 386

	alt = getenv(ALTERNATE_DB_ENVIRONMENT);
	if (!alt) alt = "";

	if (alt_odb_tail)
		return;
	alt_odb_tail = &alt_odb_list;
	link_alt_odb_entries(alt, alt + strlen(alt), ':', NULL, 0);

	read_info_alternates(get_object_directory(), 0);
}

387 388 389
static char *find_sha1_file(const unsigned char *sha1, struct stat *st)
{
	char *name = sha1_file_name(sha1);
390
	struct alternate_object_database *alt;
391 392 393

	if (!stat(name, st))
		return name;
394
	prepare_alt_odb();
395 396
	for (alt = alt_odb_list; alt; alt = alt->next) {
		name = alt->name;
397
		fill_sha1_path(name, sha1);
398 399
		if (!stat(alt->base, st))
			return alt->base;
400 401 402 403
	}
	return NULL;
}

404
static unsigned int pack_used_ctr;
405 406 407 408
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;
409
static size_t pack_mapped;
410
struct packed_git *packed_git;
411

412 413 414
void pack_report()
{
	fprintf(stderr,
415 416 417
		"pack_report: getpagesize()            = %10" SZ_FMT "\n"
		"pack_report: core.packedGitWindowSize = %10" SZ_FMT "\n"
		"pack_report: core.packedGitLimit      = %10" SZ_FMT "\n",
418
		(size_t) getpagesize(),
419 420 421 422 423 424
		packed_git_window_size,
		packed_git_limit);
	fprintf(stderr,
		"pack_report: pack_used_ctr            = %10u\n"
		"pack_report: pack_mmap_calls          = %10u\n"
		"pack_report: pack_open_windows        = %10u / %10u\n"
425 426
		"pack_report: pack_mapped              = "
			"%10" SZ_FMT " / %10" SZ_FMT "\n",
427 428 429 430 431 432
		pack_used_ctr,
		pack_mmap_calls,
		pack_open_windows, peak_pack_open_windows,
		pack_mapped, peak_pack_mapped);
}

433 434 435 436
static int check_packed_git_idx(const char *path, unsigned long *idx_size_,
				void **idx_map_)
{
	void *idx_map;
437
	uint32_t *index;
438 439
	unsigned long idx_size;
	int nr, i;
440
	int fd = open(path, O_RDONLY);
441 442 443 444 445 446 447 448
	struct stat st;
	if (fd < 0)
		return -1;
	if (fstat(fd, &st)) {
		close(fd);
		return -1;
	}
	idx_size = st.st_size;
449
	idx_map = xmmap(NULL, idx_size, PROT_READ, MAP_PRIVATE, fd, 0);
450 451 452
	close(fd);

	index = idx_map;
453 454
	*idx_map_ = idx_map;
	*idx_size_ = idx_size;
455 456

	/* check index map */
457
	if (idx_size < 4*256 + 20 + 20)
458 459 460 461 462 463 464 465 466 467 468 469
		return error("index file %s is too small", path);

	/* a future index format would start with this, as older git
	 * binaries would fail the non-monotonic index check below.
	 * give a nicer warning to the user if we can.
	 */
	if (index[0] == htonl(PACK_IDX_SIGNATURE))
		return error("index file %s is a newer version"
			" and is not supported by this binary"
			" (try upgrading GIT to a newer version)",
			path);

470 471 472 473
	nr = 0;
	for (i = 0; i < 256; i++) {
		unsigned int n = ntohl(index[i]);
		if (n < nr)
474
			return error("non-monotonic index %s", path);
475 476 477 478 479 480 481 482 483 484 485
		nr = n;
	}

	/*
	 * 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)
486
		return error("wrong index file size in %s", path);
487 488 489 490

	return 0;
}

491 492 493 494
static void scan_windows(struct packed_git *p,
	struct packed_git **lru_p,
	struct pack_window **lru_w,
	struct pack_window **lru_l)
495
{
496 497 498 499 500 501 502 503
	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;
504 505
			}
		}
506
		w_l = w;
507
	}
508 509 510 511 512 513 514 515 516 517 518
}

static int unuse_one_window(struct packed_git *current)
{
	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);
519 520 521 522 523 524 525
	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;
526
			if (!lru_p->windows && lru_p != current) {
527 528 529 530 531
				close(lru_p->pack_fd);
				lru_p->pack_fd = -1;
			}
		}
		free(lru_w);
532
		pack_open_windows--;
533 534 535
		return 1;
	}
	return 0;
536 537
}

538 539 540 541 542 543 544
void release_pack_memory(size_t need)
{
	size_t cur = pack_mapped;
	while (need >= (cur - pack_mapped) && unuse_one_window(NULL))
		; /* nothing */
}

545
void unuse_pack(struct pack_window **w_cursor)
546
{
547 548 549 550 551
	struct pack_window *w = *w_cursor;
	if (w) {
		w->inuse_cnt--;
		*w_cursor = NULL;
	}
552 553
}

554 555 556 557 558
/*
 * 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)
559
{
560 561 562 563
	struct stat st;
	struct pack_header hdr;
	unsigned char sha1[20];
	unsigned char *idx_sha1;
564
	long fd_flag;
565 566 567

	p->pack_fd = open(p->pack_name, O_RDONLY);
	if (p->pack_fd < 0 || fstat(p->pack_fd, &st))
568
		return -1;
569 570

	/* If we created the struct before we had the pack we lack size. */
571 572
	if (!p->pack_size) {
		if (!S_ISREG(st.st_mode))
573
			return error("packfile %s not a regular file", p->pack_name);
574
		p->pack_size = st.st_size;
575
	} else if (p->pack_size != st.st_size)
576
		return error("packfile %s size changed", p->pack_name);
577

578 579 580 581 582
	/* 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)
583
		return error("cannot determine file descriptor flags");
584 585
	fd_flag |= FD_CLOEXEC;
	if (fcntl(p->pack_fd, F_SETFD, fd_flag) == -1)
586
		return error("cannot set FD_CLOEXEC");
587

588
	/* Verify we recognize this pack file format. */
589
	if (read_in_full(p->pack_fd, &hdr, sizeof(hdr)) != sizeof(hdr))
590
		return error("file %s is far too short to be a packfile", p->pack_name);
591
	if (hdr.hdr_signature != htonl(PACK_SIGNATURE))
592
		return error("file %s is not a GIT packfile", p->pack_name);
593
	if (!pack_version_ok(hdr.hdr_version))
594
		return error("packfile %s is version %u and not supported"
595 596 597 598 599
			" (try upgrading GIT to a newer version)",
			p->pack_name, ntohl(hdr.hdr_version));

	/* Verify the pack matches its index. */
	if (num_packed_objects(p) != ntohl(hdr.hdr_entries))
600
		return error("packfile %s claims to have %u objects"
601 602 603 604
			" while index size indicates %u objects",
			p->pack_name, ntohl(hdr.hdr_entries),
			num_packed_objects(p));
	if (lseek(p->pack_fd, p->pack_size - sizeof(sha1), SEEK_SET) == -1)
605
		return error("end of packfile %s is unavailable", p->pack_name);
606
	if (read_in_full(p->pack_fd, sha1, sizeof(sha1)) != sizeof(sha1))
607
		return error("packfile %s signature is unavailable", p->pack_name);
608 609
	idx_sha1 = ((unsigned char *)p->index_base) + p->index_size - 40;
	if (hashcmp(sha1, idx_sha1))
610 611
		return error("packfile %s does not match index", p->pack_name);
	return 0;
612 613
}

614 615 616 617 618 619 620 621 622 623 624
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;
}

625 626 627 628 629 630 631 632 633 634 635 636 637
static int in_window(struct pack_window *win, unsigned long offset)
{
	/* 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);
}

638 639 640 641
unsigned char* use_pack(struct packed_git *p,
		struct pack_window **w_cursor,
		unsigned long offset,
		unsigned int *left)
642
{
643
	struct pack_window *win = *w_cursor;
644

645 646
	if (p->pack_fd == -1 && open_packed_git(p))
		die("packfile %s cannot be accessed", p->pack_name);
647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663

	/* 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) {
664
			size_t window_align = packed_git_window_size / 2;
665
			win = xcalloc(1, sizeof(*win));
666
			win->offset = (offset / window_align) * window_align;
667 668 669 670
			win->len = p->pack_size - win->offset;
			if (win->len > packed_git_window_size)
				win->len = packed_git_window_size;
			pack_mapped += win->len;
671 672
			while (packed_git_limit < pack_mapped
				&& unuse_one_window(p))
673
				; /* nothing */
674
			win->base = xmmap(NULL, win->len,
675 676 677
				PROT_READ, MAP_PRIVATE,
				p->pack_fd, win->offset);
			if (win->base == MAP_FAILED)
678 679 680
				die("packfile %s cannot be mapped: %s",
					p->pack_name,
					strerror(errno));
681 682 683 684 685 686
			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;
687 688 689
			win->next = p->windows;
			p->windows = win;
		}
690
	}
691 692 693 694 695
	if (win != *w_cursor) {
		win->last_used = pack_used_ctr++;
		win->inuse_cnt++;
		*w_cursor = win;
	}
696
	offset -= win->offset;
697 698 699
	if (left)
		*left = win->len - offset;
	return win->base + offset;
700 701
}

702
struct packed_git *add_packed_git(char *path, int path_len, int local)
703 704 705 706 707
{
	struct stat st;
	struct packed_git *p;
	unsigned long idx_size;
	void *idx_map;
708
	unsigned char sha1[20];
709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727

	if (check_packed_git_idx(path, &idx_size, &idx_map))
		return NULL;

	/* do we have a corresponding .pack file? */
	strcpy(path + path_len - 4, ".pack");
	if (stat(path, &st) || !S_ISREG(st.st_mode)) {
		munmap(idx_map, idx_size);
		return NULL;
	}
	/* ok, it looks sane as far as we can check without
	 * actually mapping the pack file.
	 */
	p = xmalloc(sizeof(*p) + path_len + 2);
	strcpy(p->pack_name, path);
	p->index_size = idx_size;
	p->pack_size = st.st_size;
	p->index_base = idx_map;
	p->next = NULL;
728
	p->windows = NULL;
729
	p->pack_fd = -1;
730
	p->pack_local = local;
P
Pavel Roskin 已提交
731
	if ((path_len > 44) && !get_sha1_hex(path + path_len - 44, sha1))
732
		hashcpy(p->sha1, sha1);
733 734 735
	return p;
}

736
struct packed_git *parse_pack_index(unsigned char *sha1)
737 738 739 740 741
{
	char *path = sha1_pack_index_name(sha1);
	return parse_pack_index_file(sha1, path);
}

742
struct packed_git *parse_pack_index_file(const unsigned char *sha1, char *idx_path)
743 744 745 746
{
	struct packed_git *p;
	unsigned long idx_size;
	void *idx_map;
747
	char *path;
748

749
	if (check_packed_git_idx(idx_path, &idx_size, &idx_map))
750 751 752 753 754 755 756 757 758 759
		return NULL;

	path = sha1_pack_name(sha1);

	p = xmalloc(sizeof(*p) + strlen(path) + 2);
	strcpy(p->pack_name, path);
	p->index_size = idx_size;
	p->pack_size = 0;
	p->index_base = idx_map;
	p->next = NULL;
760
	p->windows = NULL;
761
	p->pack_fd = -1;
762
	hashcpy(p->sha1, sha1);
763 764 765 766 767 768 769 770 771
	return p;
}

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

772
static void prepare_packed_git_one(char *objdir, int local)
773 774 775 776 777 778 779 780 781
{
	char path[PATH_MAX];
	int len;
	DIR *dir;
	struct dirent *de;

	sprintf(path, "%s/pack", objdir);
	len = strlen(path);
	dir = opendir(path);
782
	if (!dir) {
J
Junio C Hamano 已提交
783
		if (errno != ENOENT)
784
			error("unable to open object pack directory: %s: %s",
J
Junio C Hamano 已提交
785
			      path, strerror(errno));
786
		return;
787
	}
788 789 790 791 792
	path[len++] = '/';
	while ((de = readdir(dir)) != NULL) {
		int namelen = strlen(de->d_name);
		struct packed_git *p;

793
		if (!has_extension(de->d_name, ".idx"))
794 795
			continue;

796
		/* Don't reopen a pack we already have. */
797
		strcpy(path + len, de->d_name);
798 799 800 801 802 803
		for (p = packed_git; p; p = p->next) {
			if (!memcmp(path, p->pack_name, len + namelen - 4))
				break;
		}
		if (p)
			continue;
804 805 806
		/* See if it really is a valid .idx file with corresponding
		 * .pack file that we can map.
		 */
807
		p = add_packed_git(path, len + namelen, local);
808 809
		if (!p)
			continue;
810
		install_packed_git(p);
811
	}
812
	closedir(dir);
813 814
}

815
static int prepare_packed_git_run_once = 0;
816
void prepare_packed_git(void)
817
{
818
	struct alternate_object_database *alt;
819

820
	if (prepare_packed_git_run_once)
821
		return;
822
	prepare_packed_git_one(get_object_directory(), 1);
823
	prepare_alt_odb();
824
	for (alt = alt_odb_list; alt; alt = alt->next) {
825
		alt->name[-1] = 0;
826
		prepare_packed_git_one(alt->base, 0);
827
		alt->name[-1] = '/';
828
	}
829 830 831
	prepare_packed_git_run_once = 1;
}

832
void reprepare_packed_git(void)
833 834 835
{
	prepare_packed_git_run_once = 0;
	prepare_packed_git();
836 837
}

838
int check_sha1_signature(const unsigned char *sha1, void *map, unsigned long size, const char *type)
839 840
{
	unsigned char real_sha1[20];
841
	hash_sha1_file(map, size, type, real_sha1);
842
	return hashcmp(sha1, real_sha1) ? -1 : 0;
843 844
}

845
void *map_sha1_file(const unsigned char *sha1, unsigned long *size)
846 847 848
{
	struct stat st;
	void *map;
849
	int fd;
850 851 852 853 854
	char *filename = find_sha1_file(sha1, &st);

	if (!filename) {
		return NULL;
	}
855

856
	fd = open(filename, O_RDONLY | sha1_file_open_flag);
857
	if (fd < 0) {
858 859 860 861 862 863 864 865 866 867 868
		/* See if it works without O_NOATIME */
		switch (sha1_file_open_flag) {
		default:
			fd = open(filename, O_RDONLY);
			if (fd >= 0)
				break;
		/* Fallthrough */
		case 0:
			return NULL;
		}

869 870 871
		/* If it failed once, it will probably fail again.
		 * Stop using O_NOATIME
		 */
872
		sha1_file_open_flag = 0;
873
	}
874
	map = xmmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
875 876 877 878 879
	close(fd);
	*size = st.st_size;
	return map;
}

880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895
int legacy_loose_object(unsigned char *map)
{
	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;
}

896
unsigned long unpack_object_header_gently(const unsigned char *buf, unsigned long len, enum object_type *type, unsigned long *sizep)
897
{
898
	unsigned shift;
899 900
	unsigned char c;
	unsigned long size;
901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926
	unsigned long used = 0;

	c = buf[used++];
	*type = (c >> 4) & 7;
	size = c & 15;
	shift = 4;
	while (c & 0x80) {
		if (len <= used)
			return 0;
		if (sizeof(long) * 8 <= shift)
			return 0;
		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 */
927
	};
928
	enum object_type type;
929

930 931 932 933 934
	/* Get the data stream */
	memset(stream, 0, sizeof(*stream));
	stream->next_in = map;
	stream->avail_in = mapsize;
	stream->next_out = buffer;
935 936
	stream->avail_out = bufsiz;

937
	if (legacy_loose_object(map)) {
938 939 940 941
		inflateInit(stream);
		return inflate(stream, 0);
	}

942 943
	used = unpack_object_header_gently(map, mapsize, &type, &size);
	if (!used || !valid_loose_object_type[type])
944
		return -1;
945 946
	map += used;
	mapsize -= used;
947 948 949 950

	/* Set up the stream for the rest.. */
	stream->next_in = map;
	stream->avail_in = mapsize;
951
	inflateInit(stream);
952 953

	/* And generate the fake traditional header */
954
	stream->total_out = 1 + snprintf(buffer, bufsiz, "%s %lu",
955
					 typename(type), size);
956
	return 0;
957 958
}

L
Linus Torvalds 已提交
959
static void *unpack_sha1_rest(z_stream *stream, void *buffer, unsigned long size)
960 961
{
	int bytes = strlen(buffer) + 1;
962
	unsigned char *buf = xmalloc(1+size);
963
	unsigned long n;
964

965 966 967 968 969
	n = stream->total_out - bytes;
	if (n > size)
		n = size;
	memcpy(buf, (char *) buffer + bytes, n);
	bytes = n;
970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985
	if (bytes < size) {
		stream->next_out = buf + bytes;
		stream->avail_out = size - bytes;
		while (inflate(stream, Z_FINISH) == Z_OK)
			/* nothing */;
	}
	buf[size] = 0;
	inflateEnd(stream);
	return buf;
}

/*
 * 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.
 */
986
static int parse_sha1_header(const char *hdr, unsigned long *sizep)
987
{
988
	char type[10];
989 990 991 992 993 994
	int i;
	unsigned long size;

	/*
	 * The type can be at most ten bytes (including the 
	 * terminating '\0' that we add), and is followed by
995
	 * a space.
996
	 */
997
	i = 0;
998 999 1000 1001
	for (;;) {
		char c = *hdr++;
		if (c == ' ')
			break;
1002 1003
		type[i++] = c;
		if (i >= sizeof(type))
1004 1005
			return -1;
	}
1006
	type[i] = 0;
1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028

	/*
	 * 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
	 */
1029
	return *hdr ? -1 : type_from_string(type);
1030 1031
}

1032
void * unpack_sha1_file(void *map, unsigned long mapsize, enum object_type *type, unsigned long *size)
1033
{
1034
	int ret;
1035
	z_stream stream;
1036
	char hdr[8192];
1037

1038
	ret = unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr));
1039
	if (ret < Z_OK || (*type = parse_sha1_header(hdr, size)) < 0)
1040 1041
		return NULL;

1042
	return unpack_sha1_rest(&stream, hdr, *size);
1043 1044
}

1045
static unsigned long get_delta_base(struct packed_git *p,
1046
				    struct pack_window **w_curs,
1047
				    unsigned long *curpos,
1048
				    enum object_type type,
1049
				    unsigned long delta_obj_offset)
1050
{
1051
	unsigned char *base_info = use_pack(p, w_curs, *curpos, NULL);
1052 1053
	unsigned long base_offset;

1054 1055 1056 1057 1058 1059
	/* 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.
	 */
1060
	if (type == OBJ_OFS_DELTA) {
1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073
		unsigned used = 0;
		unsigned char c = base_info[used++];
		base_offset = c & 127;
		while (c & 128) {
			base_offset += 1;
			if (!base_offset || base_offset & ~(~0UL >> 7))
				die("offset value overflow for delta base object");
			c = base_info[used++];
			base_offset = (base_offset << 7) + (c & 127);
		}
		base_offset = delta_obj_offset - base_offset;
		if (base_offset >= delta_obj_offset)
			die("delta base offset out of bound");
1074
		*curpos += used;
1075
	} else if (type == OBJ_REF_DELTA) {
1076 1077 1078 1079 1080
		/* The base entry _must_ be in the same pack */
		base_offset = find_pack_entry_one(base_info, p);
		if (!base_offset)
			die("failed to find delta-pack base object %s",
				sha1_to_hex(base_info));
1081
		*curpos += 20;
1082 1083
	} else
		die("I am totally screwed");
1084
	return base_offset;
1085 1086
}

J
Junio C Hamano 已提交
1087
/* forward declaration for a mutually recursive function */
N
Nicolas Pitre 已提交
1088
static int packed_object_info(struct packed_git *p, unsigned long offset,
1089
			      unsigned long *sizep);
J
Junio C Hamano 已提交
1090

N
Nicolas Pitre 已提交
1091
static int packed_delta_info(struct packed_git *p,
1092
			     struct pack_window **w_curs,
1093
			     unsigned long curpos,
1094
			     enum object_type type,
1095
			     unsigned long obj_offset,
N
Nicolas Pitre 已提交
1096
			     unsigned long *sizep)
1097
{
N
Nicolas Pitre 已提交
1098
	unsigned long base_offset;
J
Junio C Hamano 已提交
1099

1100 1101
	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
1102

1103 1104 1105 1106 1107 1108 1109
	/* 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.
	 */
	if (sizep) {
		const unsigned char *data;
1110
		unsigned char delta_head[20], *in;
1111 1112 1113 1114 1115 1116 1117 1118
		z_stream stream;
		int st;

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

		inflateInit(&stream);
1119
		do {
1120
			in = use_pack(p, w_curs, curpos, &stream.avail_in);
1121 1122
			stream.next_in = in;
			st = inflate(&stream, Z_FINISH);
1123
			curpos += stream.next_in - in;
1124 1125
		} while ((st == Z_OK || st == Z_BUF_ERROR)
			&& stream.total_out < sizeof(delta_head));
1126 1127 1128 1129 1130 1131 1132 1133 1134
		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;
1135 1136 1137

		/* ignore base size */
		get_delta_hdr_size(&data, delta_head+sizeof(delta_head));
1138

1139
		/* Read the result size */
1140
		*sizep = get_delta_hdr_size(&data, delta_head+sizeof(delta_head));
1141
	}
1142 1143

	return type;
1144 1145
}

1146 1147 1148 1149
static int unpack_object_header(struct packed_git *p,
				struct pack_window **w_curs,
				unsigned long *curpos,
				unsigned long *sizep)
1150
{
1151 1152
	unsigned char *base;
	unsigned int left;
1153
	unsigned long used;
1154
	enum object_type type;
1155

1156 1157
	/* 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
1158
	 * size that is assured.)  With our object header encoding
1159 1160 1161
	 * the maximum deflated object size is 2^137, which is just
	 * insane, so we know won't exceed what we have been given.
	 */
1162 1163
	base = use_pack(p, w_curs, *curpos, &left);
	used = unpack_object_header_gently(base, left, &type, sizep);
1164 1165
	if (!used)
		die("object offset outside of pack file");
1166
	*curpos += used;
1167

1168
	return type;
1169 1170
}

1171 1172 1173 1174 1175 1176
const char *packed_object_info_detail(struct packed_git *p,
				      unsigned long obj_offset,
				      unsigned long *size,
				      unsigned long *store_size,
				      unsigned int *delta_chain_length,
				      unsigned char *base_sha1)
1177
{
1178
	struct pack_window *w_curs = NULL;
1179
	unsigned long curpos, dummy;
N
Nicolas Pitre 已提交
1180
	unsigned char *next_sha1;
1181
	enum object_type type;
1182

N
Nicolas Pitre 已提交
1183
	*delta_chain_length = 0;
1184
	curpos = obj_offset;
1185
	type = unpack_object_header(p, &w_curs, &curpos, size);
N
Nicolas Pitre 已提交
1186 1187

	for (;;) {
1188
		switch (type) {
N
Nicolas Pitre 已提交
1189
		default:
1190
			die("pack %s contains unknown object type %d",
1191
			    p->pack_name, type);
N
Nicolas Pitre 已提交
1192 1193 1194 1195 1196
		case OBJ_COMMIT:
		case OBJ_TREE:
		case OBJ_BLOB:
		case OBJ_TAG:
			*store_size = 0; /* notyet */
1197
			unuse_pack(&w_curs);
1198
			return typename(type);
1199
		case OBJ_OFS_DELTA:
1200
			obj_offset = get_delta_base(p, &w_curs, &curpos, type, obj_offset);
1201
			if (*delta_chain_length == 0) {
1202
				/* TODO: find base_sha1 as pointed by curpos */
1203
				hashclr(base_sha1);
1204 1205 1206
			}
			break;
		case OBJ_REF_DELTA:
1207
			next_sha1 = use_pack(p, &w_curs, curpos, NULL);
N
Nicolas Pitre 已提交
1208 1209
			if (*delta_chain_length == 0)
				hashcpy(base_sha1, next_sha1);
1210
			obj_offset = find_pack_entry_one(next_sha1, p);
N
Nicolas Pitre 已提交
1211 1212 1213
			break;
		}
		(*delta_chain_length)++;
1214
		curpos = obj_offset;
1215
		type = unpack_object_header(p, &w_curs, &curpos, &dummy);
1216 1217 1218
	}
}

1219
static int packed_object_info(struct packed_git *p, unsigned long obj_offset,
1220
			      unsigned long *sizep)
1221
{
1222
	struct pack_window *w_curs = NULL;
1223
	unsigned long size, curpos = obj_offset;
1224
	enum object_type type;
1225

1226
	type = unpack_object_header(p, &w_curs, &curpos, &size);
1227

1228
	switch (type) {
1229 1230
	case OBJ_OFS_DELTA:
	case OBJ_REF_DELTA:
1231 1232
		type = packed_delta_info(p, &w_curs, curpos,
					 type, obj_offset, sizep);
1233
		break;
1234 1235 1236 1237
	case OBJ_COMMIT:
	case OBJ_TREE:
	case OBJ_BLOB:
	case OBJ_TAG:
1238 1239
		if (sizep)
			*sizep = size;
1240
		break;
1241
	default:
1242
		die("pack %s contains unknown object type %d",
1243
		    p->pack_name, type);
1244
	}
1245
	unuse_pack(&w_curs);
1246
	return type;
1247 1248
}

1249
static void *unpack_compressed_entry(struct packed_git *p,
1250
				    struct pack_window **w_curs,
1251
				    unsigned long curpos,
1252
				    unsigned long size)
1253 1254 1255
{
	int st;
	z_stream stream;
1256
	unsigned char *buffer, *in;
1257 1258 1259 1260 1261 1262 1263 1264

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

	inflateInit(&stream);
1265
	do {
1266
		in = use_pack(p, w_curs, curpos, &stream.avail_in);
1267 1268
		stream.next_in = in;
		st = inflate(&stream, Z_FINISH);
1269
		curpos += stream.next_in - in;
1270
	} while (st == Z_OK || st == Z_BUF_ERROR);
1271 1272 1273 1274 1275 1276 1277 1278 1279
	inflateEnd(&stream);
	if ((st != Z_STREAM_END) || stream.total_out != size) {
		free(buffer);
		return NULL;
	}

	return buffer;
}

1280
static void *unpack_delta_entry(struct packed_git *p,
1281
				struct pack_window **w_curs,
1282
				unsigned long curpos,
1283
				unsigned long delta_size,
1284
				unsigned long obj_offset,
1285
				enum object_type *type,
1286
				unsigned long *sizep)
1287
{
1288
	void *delta_data, *result, *base;
1289
	unsigned long base_size, base_offset;
N
Nicolas Pitre 已提交
1290

1291
	base_offset = get_delta_base(p, w_curs, &curpos, *type, obj_offset);
1292
	base = unpack_entry(p, base_offset, type, &base_size);
1293
	if (!base)
N
Nicolas Pitre 已提交
1294 1295
		die("failed to read delta base object at %lu from %s",
		    base_offset, p->pack_name);
1296

1297
	delta_data = unpack_compressed_entry(p, w_curs, curpos, delta_size);
1298 1299
	result = patch_delta(base, base_size,
			     delta_data, delta_size,
1300
			     sizep);
1301 1302 1303 1304 1305 1306 1307
	if (!result)
		die("failed to apply delta");
	free(delta_data);
	free(base);
	return result;
}

1308
void *unpack_entry(struct packed_git *p, unsigned long obj_offset,
1309
		   enum object_type *type, unsigned long *sizep)
1310
{
1311
	struct pack_window *w_curs = NULL;
1312 1313
	unsigned long curpos = obj_offset;
	void *data;
1314

1315 1316
	*type = unpack_object_header(p, &w_curs, &curpos, sizep);
	switch (*type) {
1317 1318
	case OBJ_OFS_DELTA:
	case OBJ_REF_DELTA:
1319 1320
		data = unpack_delta_entry(p, &w_curs, curpos, *sizep,
					  obj_offset, type, sizep);
1321
		break;
1322 1323 1324 1325
	case OBJ_COMMIT:
	case OBJ_TREE:
	case OBJ_BLOB:
	case OBJ_TAG:
1326
		data = unpack_compressed_entry(p, &w_curs, curpos, *sizep);
1327
		break;
1328
	default:
1329
		die("unknown object type %i in %s", *type, p->pack_name);
1330
	}
1331
	unuse_pack(&w_curs);
1332
	return data;
1333 1334
}

1335 1336
int num_packed_objects(const struct packed_git *p)
{
1337
	/* See check_packed_git_idx() */
1338 1339 1340 1341 1342 1343 1344 1345 1346
	return (p->index_size - 20 - 20 - 4*256) / 24;
}

int nth_packed_object_sha1(const struct packed_git *p, int n,
			   unsigned char* sha1)
{
	void *index = p->index_base + 256;
	if (n < 0 || num_packed_objects(p) <= n)
		return -1;
1347
	hashcpy(sha1, (unsigned char *) index + (24 * n) + 4);
1348 1349 1350
	return 0;
}

N
Nicolas Pitre 已提交
1351 1352
unsigned long find_pack_entry_one(const unsigned char *sha1,
				  struct packed_git *p)
1353
{
1354
	uint32_t *level1_ofs = p->index_base;
1355 1356 1357 1358 1359 1360
	int hi = ntohl(level1_ofs[*sha1]);
	int lo = ((*sha1 == 0x0) ? 0 : ntohl(level1_ofs[*sha1 - 1]));
	void *index = p->index_base + 256;

	do {
		int mi = (lo + hi) / 2;
1361
		int cmp = hashcmp((unsigned char *)index + (24 * mi) + 4, sha1);
N
Nicolas Pitre 已提交
1362
		if (!cmp)
1363
			return ntohl(*((uint32_t *)((char *)index + (24 * mi))));
1364 1365 1366 1367 1368 1369 1370 1371
		if (cmp > 0)
			hi = mi;
		else
			lo = mi+1;
	} while (lo < hi);
	return 0;
}

1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389
static int matches_pack_name(struct packed_git *p, const char *ig)
{
	const char *last_c, *c;

	if (!strcmp(p->pack_name, ig))
		return 0;

	for (c = p->pack_name, last_c = c; *c;)
		if (*c == '/')
			last_c = ++c;
		else
			++c;
	if (!strcmp(last_c, ig))
		return 0;

	return 1;
}

1390
static int find_pack_entry(const unsigned char *sha1, struct pack_entry *e, const char **ignore_packed)
1391 1392
{
	struct packed_git *p;
N
Nicolas Pitre 已提交
1393 1394
	unsigned long offset;

1395 1396 1397
	prepare_packed_git();

	for (p = packed_git; p; p = p->next) {
1398 1399 1400
		if (ignore_packed) {
			const char **ig;
			for (ig = ignore_packed; *ig; ig++)
1401
				if (!matches_pack_name(p, *ig))
1402 1403 1404 1405
					break;
			if (*ig)
				continue;
		}
N
Nicolas Pitre 已提交
1406 1407
		offset = find_pack_entry_one(sha1, p);
		if (offset) {
1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419
			/*
			 * 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);
				continue;
			}
N
Nicolas Pitre 已提交
1420 1421 1422
			e->offset = offset;
			e->p = p;
			hashcpy(e->sha1, sha1);
1423
			return 1;
N
Nicolas Pitre 已提交
1424
		}
1425 1426 1427 1428
	}
	return 0;
}

1429 1430 1431 1432 1433 1434
struct packed_git *find_sha1_pack(const unsigned char *sha1, 
				  struct packed_git *packs)
{
	struct packed_git *p;

	for (p = packs; p; p = p->next) {
N
Nicolas Pitre 已提交
1435
		if (find_pack_entry_one(sha1, p))
1436 1437 1438
			return p;
	}
	return NULL;
1439

1440 1441
}

1442
static int sha1_loose_object_info(const unsigned char *sha1, unsigned long *sizep)
1443
{
1444
	int status;
1445 1446 1447
	unsigned long mapsize, size;
	void *map;
	z_stream stream;
N
Nicolas Pitre 已提交
1448
	char hdr[32];
1449

1450
	map = map_sha1_file(sha1, &mapsize);
1451 1452
	if (!map)
		return error("unable to find %s", sha1_to_hex(sha1));
1453 1454 1455
	if (unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)) < 0)
		status = error("unable to unpack %s header",
			       sha1_to_hex(sha1));
1456
	else if ((status = parse_sha1_header(hdr, &size)) < 0)
1457
		status = error("unable to parse %s header", sha1_to_hex(sha1));
1458 1459
	else if (sizep)
		*sizep = size;
1460 1461 1462 1463 1464
	inflateEnd(&stream);
	munmap(map, mapsize);
	return status;
}

1465
int sha1_object_info(const unsigned char *sha1, unsigned long *sizep)
1466 1467 1468 1469 1470 1471
{
	struct pack_entry e;

	if (!find_pack_entry(sha1, &e, NULL)) {
		reprepare_packed_git();
		if (!find_pack_entry(sha1, &e, NULL))
1472
			return sha1_loose_object_info(sha1, sizep);
1473
	}
1474
	return packed_object_info(e.p, e.offset, sizep);
1475 1476
}

1477 1478
static void *read_packed_sha1(const unsigned char *sha1,
			      enum object_type *type, unsigned long *size)
1479 1480 1481
{
	struct pack_entry e;

1482
	if (!find_pack_entry(sha1, &e, NULL))
1483
		return NULL;
1484 1485
	else
		return unpack_entry(e.p, e.offset, type, size);
1486 1487
}

1488 1489 1490 1491 1492 1493 1494 1495
/*
 * 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];
1496
	enum object_type type;
1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513
	void *buf;
	unsigned long size;
} *cached_objects;
static int cached_object_nr, cached_object_alloc;

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;
	}
	return NULL;
}

1514 1515
int pretend_sha1_file(void *buf, unsigned long len, enum object_type type,
		      unsigned char *sha1)
1516 1517 1518
{
	struct cached_object *co;

1519
	hash_sha1_file(buf, len, typename(type), sha1);
1520 1521 1522 1523 1524 1525 1526 1527 1528 1529
	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;
1530
	co->type = type;
J
Junio C Hamano 已提交
1531 1532
	co->buf = xmalloc(len);
	memcpy(co->buf, buf, len);
1533 1534 1535 1536
	hashcpy(co->sha1, sha1);
	return 0;
}

1537 1538
void *read_sha1_file(const unsigned char *sha1, enum object_type *type,
		     unsigned long *size)
1539 1540 1541
{
	unsigned long mapsize;
	void *map, *buf;
1542 1543 1544 1545 1546 1547 1548
	struct cached_object *co;

	co = find_cached_object(sha1);
	if (co) {
		buf = xmalloc(co->size + 1);
		memcpy(buf, co->buf, co->size);
		((char*)buf)[co->size] = 0;
1549
		*type = co->type;
1550 1551 1552
		*size = co->size;
		return buf;
	}
1553

1554 1555 1556
	buf = read_packed_sha1(sha1, type, size);
	if (buf)
		return buf;
1557
	map = map_sha1_file(sha1, &mapsize);
1558 1559 1560 1561 1562
	if (map) {
		buf = unpack_sha1_file(map, mapsize, type, size);
		munmap(map, mapsize);
		return buf;
	}
1563
	reprepare_packed_git();
1564
	return read_packed_sha1(sha1, type, size);
1565 1566
}

1567
void *read_object_with_reference(const unsigned char *sha1,
1568
				 const char *required_type_name,
1569 1570
				 unsigned long *size,
				 unsigned char *actual_sha1_return)
1571
{
1572
	enum object_type type, required_type;
1573 1574
	void *buffer;
	unsigned long isize;
1575
	unsigned char actual_sha1[20];
1576

1577
	required_type = type_from_string(required_type_name);
1578
	hashcpy(actual_sha1, sha1);
1579 1580 1581
	while (1) {
		int ref_length = -1;
		const char *ref_type = NULL;
1582

1583
		buffer = read_sha1_file(actual_sha1, &type, &isize);
1584 1585
		if (!buffer)
			return NULL;
1586
		if (type == required_type) {
1587 1588
			*size = isize;
			if (actual_sha1_return)
1589
				hashcpy(actual_sha1_return, actual_sha1);
1590 1591 1592
			return buffer;
		}
		/* Handle references */
1593
		else if (type == OBJ_COMMIT)
1594
			ref_type = "tree ";
1595
		else if (type == OBJ_TAG)
1596 1597 1598 1599 1600 1601
			ref_type = "object ";
		else {
			free(buffer);
			return NULL;
		}
		ref_length = strlen(ref_type);
1602

1603
		if (memcmp(buffer, ref_type, ref_length) ||
1604
		    get_sha1_hex((char *) buffer + ref_length, actual_sha1)) {
1605 1606 1607
			free(buffer);
			return NULL;
		}
1608
		free(buffer);
1609 1610
		/* Now we have the ID of the referred-to object in
		 * actual_sha1.  Check again. */
1611 1612 1613
	}
}

R
Rene Scharfe 已提交
1614 1615
static void write_sha1_file_prepare(void *buf, unsigned long len,
                                    const char *type, unsigned char *sha1,
N
Nicolas Pitre 已提交
1616
                                    char *hdr, int *hdrlen)
1617 1618 1619 1620
{
	SHA_CTX c;

	/* Generate the header */
N
Nicolas Pitre 已提交
1621
	*hdrlen = sprintf(hdr, "%s %lu", type, len)+1;
1622 1623 1624 1625 1626 1627 1628 1629

	/* Sha1.. */
	SHA1_Init(&c);
	SHA1_Update(&c, hdr, *hdrlen);
	SHA1_Update(&c, buf, len);
	SHA1_Final(sha1, &c);
}

1630 1631 1632 1633 1634 1635
/*
 * Link the tempfile to the final place, possibly creating the
 * last directory level as you do so.
 *
 * Returns the errno on failure, 0 on success.
 */
1636
static int link_temp_to_file(const char *tmpfile, const char *filename)
1637 1638
{
	int ret;
S
Shawn Pearce 已提交
1639
	char *dir;
1640 1641 1642 1643 1644

	if (!link(tmpfile, filename))
		return 0;

	/*
S
Shawn Pearce 已提交
1645
	 * Try to mkdir the last path component if that failed.
1646 1647 1648 1649 1650 1651
	 *
	 * Re-try the "link()" regardless of whether the mkdir
	 * succeeds, since a race might mean that somebody
	 * else succeeded.
	 */
	ret = errno;
S
Shawn Pearce 已提交
1652 1653 1654
	dir = strrchr(filename, '/');
	if (dir) {
		*dir = 0;
1655
		if (!mkdir(filename, 0777) && adjust_shared_perm(filename)) {
1656
			*dir = '/';
S
Shawn Pearce 已提交
1657
			return -2;
1658
		}
S
Shawn Pearce 已提交
1659 1660 1661 1662
		*dir = '/';
		if (!link(tmpfile, filename))
			return 0;
		ret = errno;
1663 1664 1665 1666 1667 1668 1669
	}
	return ret;
}

/*
 * Move the just written object into its final resting place
 */
1670
int move_temp_to_file(const char *tmpfile, const char *filename)
1671 1672
{
	int ret = link_temp_to_file(tmpfile, filename);
1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686

	/*
	 * 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))
1687
			return 0;
1688
		ret = errno;
1689 1690 1691 1692
	}
	unlink(tmpfile);
	if (ret) {
		if (ret != EEXIST) {
1693
			return error("unable to write sha1 filename %s: %s\n", filename, strerror(ret));
1694 1695 1696 1697 1698 1699 1700
		}
		/* FIXME!!! Collision check here ? */
	}

	return 0;
}

L
Linus Torvalds 已提交
1701 1702
static int write_buffer(int fd, const void *buf, size_t len)
{
L
Linus Torvalds 已提交
1703
	if (write_in_full(fd, buf, len) < 0)
1704
		return error("file write error (%s)", strerror(errno));
L
Linus Torvalds 已提交
1705 1706 1707
	return 0;
}

1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727
static int write_binary_header(unsigned char *hdr, enum object_type type, unsigned long len)
{
	int hdr_len;
	unsigned char c;

	c = (type << 4) | (len & 15);
	len >>= 4;
	hdr_len = 1;
	while (len) {
		*hdr++ = c | 0x80;
		hdr_len++;
		c = (len & 0x7f);
		len >>= 7;
	}
	*hdr = c;
	return hdr_len;
}

static void setup_object_header(z_stream *stream, const char *type, unsigned long len)
{
N
Nicolas Pitre 已提交
1728
	int obj_type, hdrlen;
1729 1730 1731 1732 1733 1734

	if (use_legacy_headers) {
		while (deflate(stream, 0) == Z_OK)
			/* nothing */;
		return;
	}
1735
	obj_type = type_from_string(type);
N
Nicolas Pitre 已提交
1736 1737 1738 1739
	hdrlen = write_binary_header(stream->next_out, obj_type, len);
	stream->total_out = hdrlen;
	stream->next_out += hdrlen;
	stream->avail_out -= hdrlen;
1740 1741
}

R
Rene Scharfe 已提交
1742 1743 1744
int hash_sha1_file(void *buf, unsigned long len, const char *type,
                   unsigned char *sha1)
{
N
Nicolas Pitre 已提交
1745
	char hdr[32];
R
Rene Scharfe 已提交
1746 1747 1748 1749 1750
	int hdrlen;
	write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen);
	return 0;
}

1751
int write_sha1_file(void *buf, unsigned long len, const char *type, unsigned char *returnsha1)
1752 1753
{
	int size;
1754
	unsigned char *compressed;
1755 1756
	z_stream stream;
	unsigned char sha1[20];
1757
	char *filename;
1758
	static char tmpfile[PATH_MAX];
N
Nicolas Pitre 已提交
1759
	char hdr[32];
1760
	int fd, hdrlen;
1761

1762 1763 1764
	/* Normally if we have it in the pack then we do not bother writing
	 * it out into .git/objects/??/?{38} file.
	 */
R
Rene Scharfe 已提交
1765 1766
	write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen);
	filename = sha1_file_name(sha1);
1767
	if (returnsha1)
1768
		hashcpy(returnsha1, sha1);
1769 1770
	if (has_sha1_file(sha1))
		return 0;
1771 1772
	fd = open(filename, O_RDONLY);
	if (fd >= 0) {
1773
		/*
1774 1775
		 * FIXME!!! We might do collision checking here, but we'd
		 * need to uncompress the old file and check it. Later.
1776
		 */
1777
		close(fd);
1778 1779 1780
		return 0;
	}

1781
	if (errno != ENOENT) {
1782
		return error("sha1 file %s: %s\n", filename, strerror(errno));
1783 1784 1785
	}

	snprintf(tmpfile, sizeof(tmpfile), "%s/obj_XXXXXX", get_object_directory());
1786

1787 1788
	fd = mkstemp(tmpfile);
	if (fd < 0) {
1789 1790 1791 1792
		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));
1793 1794
	}

1795 1796
	/* Set it up */
	memset(&stream, 0, sizeof(stream));
1797
	deflateInit(&stream, zlib_compression_level);
1798
	size = 8 + deflateBound(&stream, len+hdrlen);
1799
	compressed = xmalloc(size);
1800 1801 1802 1803

	/* Compress it */
	stream.next_out = compressed;
	stream.avail_out = size;
1804 1805

	/* First header.. */
N
Nicolas Pitre 已提交
1806
	stream.next_in = (unsigned char *)hdr;
1807
	stream.avail_in = hdrlen;
1808
	setup_object_header(&stream, type, len);
1809 1810 1811 1812

	/* Then the data itself.. */
	stream.next_in = buf;
	stream.avail_in = len;
1813 1814 1815 1816 1817
	while (deflate(&stream, Z_FINISH) == Z_OK)
		/* nothing */;
	deflateEnd(&stream);
	size = stream.total_out;

L
Linus Torvalds 已提交
1818 1819
	if (write_buffer(fd, compressed, size) < 0)
		die("unable to write sha1 file");
1820
	fchmod(fd, 0444);
1821
	close(fd);
1822
	free(compressed);
1823

1824
	return move_temp_to_file(tmpfile, filename);
1825
}
1826

L
Linus Torvalds 已提交
1827 1828 1829 1830 1831
/*
 * We need to unpack and recompress the object for writing
 * it out to a different file.
 */
static void *repack_object(const unsigned char *sha1, unsigned long *objsize)
D
Daniel Barkalow 已提交
1832
{
L
Linus Torvalds 已提交
1833
	size_t size;
D
Daniel Barkalow 已提交
1834
	z_stream stream;
L
Linus Torvalds 已提交
1835 1836
	unsigned char *unpacked;
	unsigned long len;
1837
	enum object_type type;
N
Nicolas Pitre 已提交
1838
	char hdr[32];
L
Linus Torvalds 已提交
1839 1840
	int hdrlen;
	void *buf;
1841

1842
	/* need to unpack and recompress it by itself */
1843
	unpacked = read_packed_sha1(sha1, &type, &len);
1844 1845
	if (!unpacked)
		error("cannot read sha1_file for %s", sha1_to_hex(sha1));
D
Daniel Barkalow 已提交
1846

1847
	hdrlen = sprintf(hdr, "%s %lu", typename(type), len) + 1;
D
Daniel Barkalow 已提交
1848

L
Linus Torvalds 已提交
1849 1850
	/* Set it up */
	memset(&stream, 0, sizeof(stream));
1851
	deflateInit(&stream, zlib_compression_level);
L
Linus Torvalds 已提交
1852 1853
	size = deflateBound(&stream, len + hdrlen);
	buf = xmalloc(size);
D
Daniel Barkalow 已提交
1854

L
Linus Torvalds 已提交
1855 1856 1857
	/* Compress it */
	stream.next_out = buf;
	stream.avail_out = size;
D
Daniel Barkalow 已提交
1858

L
Linus Torvalds 已提交
1859 1860 1861 1862 1863
	/* First header.. */
	stream.next_in = (void *)hdr;
	stream.avail_in = hdrlen;
	while (deflate(&stream, 0) == Z_OK)
		/* nothing */;
1864

L
Linus Torvalds 已提交
1865 1866 1867 1868 1869 1870 1871
	/* Then the data itself.. */
	stream.next_in = unpacked;
	stream.avail_in = len;
	while (deflate(&stream, Z_FINISH) == Z_OK)
		/* nothing */;
	deflateEnd(&stream);
	free(unpacked);
1872

L
Linus Torvalds 已提交
1873 1874 1875 1876 1877 1878 1879 1880
	*objsize = stream.total_out;
	return buf;
}

int write_sha1_to_fd(int fd, const unsigned char *sha1)
{
	int retval;
	unsigned long objsize;
1881
	void *buf = map_sha1_file(sha1, &objsize);
L
Linus Torvalds 已提交
1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892

	if (buf) {
		retval = write_buffer(fd, buf, objsize);
		munmap(buf, objsize);
		return retval;
	}

	buf = repack_object(sha1, &objsize);
	retval = write_buffer(fd, buf, objsize);
	free(buf);
	return retval;
D
Daniel Barkalow 已提交
1893 1894
}

1895 1896
int write_sha1_from_fd(const unsigned char *sha1, int fd, char *buffer,
		       size_t bufsize, size_t *bufposn)
1897
{
1898
	char tmpfile[PATH_MAX];
1899 1900 1901
	int local;
	z_stream stream;
	unsigned char real_sha1[20];
1902
	unsigned char discard[4096];
1903 1904 1905
	int ret;
	SHA_CTX c;

1906
	snprintf(tmpfile, sizeof(tmpfile), "%s/obj_XXXXXX", get_object_directory());
1907

1908
	local = mkstemp(tmpfile);
1909 1910 1911 1912 1913 1914
	if (local < 0) {
		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));
	}
1915 1916 1917 1918 1919 1920 1921 1922 1923

	memset(&stream, 0, sizeof(stream));

	inflateInit(&stream);

	SHA1_Init(&c);

	do {
		ssize_t size;
1924 1925
		if (*bufposn) {
			stream.avail_in = *bufposn;
P
Pavel Roskin 已提交
1926
			stream.next_in = (unsigned char *) buffer;
1927 1928 1929 1930 1931 1932 1933
			do {
				stream.next_out = discard;
				stream.avail_out = sizeof(discard);
				ret = inflate(&stream, Z_SYNC_FLUSH);
				SHA1_Update(&c, discard, sizeof(discard) -
					    stream.avail_out);
			} while (stream.avail_in && ret == Z_OK);
L
Linus Torvalds 已提交
1934 1935
			if (write_buffer(local, buffer, *bufposn - stream.avail_in) < 0)
				die("unable to write sha1 file");
1936 1937 1938 1939 1940 1941
			memmove(buffer, buffer + *bufposn - stream.avail_in,
				stream.avail_in);
			*bufposn = stream.avail_in;
			if (ret != Z_OK)
				break;
		}
1942
		size = xread(fd, buffer + *bufposn, bufsize - *bufposn);
1943 1944
		if (size <= 0) {
			close(local);
1945
			unlink(tmpfile);
1946 1947 1948 1949 1950
			if (!size)
				return error("Connection closed?");
			perror("Reading from connection");
			return -1;
		}
1951 1952
		*bufposn += size;
	} while (1);
1953 1954 1955 1956 1957
	inflateEnd(&stream);

	close(local);
	SHA1_Final(real_sha1, &c);
	if (ret != Z_STREAM_END) {
1958
		unlink(tmpfile);
1959 1960
		return error("File %s corrupted", sha1_to_hex(sha1));
	}
1961
	if (hashcmp(sha1, real_sha1)) {
1962
		unlink(tmpfile);
1963
		return error("File %s has bad hash", sha1_to_hex(sha1));
1964
	}
1965 1966

	return move_temp_to_file(tmpfile, sha1_file_name(sha1));
1967 1968
}

1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984
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;
}

1985
int has_sha1_pack(const unsigned char *sha1, const char **ignore_packed)
1986 1987
{
	struct pack_entry e;
1988
	return find_pack_entry(sha1, &e, ignore_packed);
1989 1990
}

1991 1992 1993
int has_sha1_file(const unsigned char *sha1)
{
	struct stat st;
1994 1995
	struct pack_entry e;

1996
	if (find_pack_entry(sha1, &e, NULL))
1997
		return 1;
1998
	return find_sha1_file(sha1, &st) ? 1 : 0;
1999
}
J
Junio C Hamano 已提交
2000

2001 2002
/*
 * reads from fd as long as possible into a supplied buffer of size bytes.
2003
 * If necessary the buffer's size is increased using realloc()
2004 2005 2006 2007 2008 2009 2010
 *
 * returns 0 if anything went fine and -1 otherwise
 *
 * NOTE: both buf and size may change, but even when -1 is returned
 * you still have to free() it yourself.
 */
int read_pipe(int fd, char** return_buf, unsigned long* return_size)
2011
{
2012 2013 2014
	char* buf = *return_buf;
	unsigned long size = *return_size;
	int iret;
2015
	unsigned long off = 0;
2016

2017
	do {
2018
		iret = xread(fd, buf + off, size - off);
2019 2020 2021 2022
		if (iret > 0) {
			off += iret;
			if (off == size) {
				size *= 2;
J
Jonas Fonseca 已提交
2023
				buf = xrealloc(buf, size);
2024 2025 2026
			}
		}
	} while (iret > 0);
2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038

	*return_buf = buf;
	*return_size = off;

	if (iret < 0)
		return -1;
	return 0;
}

int index_pipe(unsigned char *sha1, int fd, const char *type, int write_object)
{
	unsigned long size = 4096;
J
Jonas Fonseca 已提交
2039
	char *buf = xmalloc(size);
2040 2041 2042
	int ret;

	if (read_pipe(fd, &buf, &size)) {
2043 2044 2045
		free(buf);
		return -1;
	}
2046

2047
	if (!type)
2048
		type = blob_type;
2049
	if (write_object)
2050
		ret = write_sha1_file(buf, size, type, sha1);
R
Rene Scharfe 已提交
2051 2052
	else
		ret = hash_sha1_file(buf, size, type, sha1);
2053 2054 2055 2056
	free(buf);
	return ret;
}

2057
int index_fd(unsigned char *sha1, int fd, struct stat *st, int write_object,
2058
	     enum object_type type, const char *path)
J
Junio C Hamano 已提交
2059 2060
{
	unsigned long size = st->st_size;
2061
	void *buf;
L
Linus Torvalds 已提交
2062
	int ret, re_allocated = 0;
J
Junio C Hamano 已提交
2063

2064
	buf = "";
J
Junio C Hamano 已提交
2065
	if (size)
2066
		buf = xmmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
J
Junio C Hamano 已提交
2067 2068
	close(fd);

2069
	if (!type)
2070
		type = OBJ_BLOB;
L
Linus Torvalds 已提交
2071 2072 2073 2074

	/*
	 * Convert blobs to git internal format
	 */
2075
	if ((type == OBJ_BLOB) && S_ISREG(st->st_mode)) {
L
Linus Torvalds 已提交
2076 2077
		unsigned long nsize = size;
		char *nbuf = buf;
2078
		if (convert_to_git(path, &nbuf, &nsize)) {
L
Linus Torvalds 已提交
2079 2080 2081 2082 2083 2084 2085 2086
			if (size)
				munmap(buf, size);
			size = nsize;
			buf = nbuf;
			re_allocated = 1;
		}
	}

2087
	if (write_object)
2088
		ret = write_sha1_file(buf, size, typename(type), sha1);
R
Rene Scharfe 已提交
2089
	else
2090
		ret = hash_sha1_file(buf, size, typename(type), sha1);
L
Linus Torvalds 已提交
2091 2092 2093 2094
	if (re_allocated) {
		free(buf);
		return ret;
	}
2095 2096 2097
	if (size)
		munmap(buf, size);
	return ret;
J
Junio C Hamano 已提交
2098
}
2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110

int index_path(unsigned char *sha1, const char *path, struct stat *st, int write_object)
{
	int fd;
	char *target;

	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));
2111
		if (index_fd(sha1, fd, st, write_object, OBJ_BLOB, path) < 0)
2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122
			return error("%s: failed to insert into database",
				     path);
		break;
	case S_IFLNK:
		target = xmalloc(st->st_size+1);
		if (readlink(path, target, st->st_size+1) != st->st_size) {
			char *errstr = strerror(errno);
			free(target);
			return error("readlink(\"%s\"): %s", path,
			             errstr);
		}
R
Rene Scharfe 已提交
2123 2124 2125
		if (!write_object)
			hash_sha1_file(target, st->st_size, blob_type, sha1);
		else if (write_sha1_file(target, st->st_size, blob_type, sha1))
2126 2127 2128 2129 2130 2131 2132 2133 2134
			return error("%s: failed to insert into database",
				     path);
		free(target);
		break;
	default:
		return error("%s: unsupported file type", path);
	}
	return 0;
}
2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155

int read_pack_header(int fd, struct pack_header *header)
{
	char *c = (char*)header;
	ssize_t remaining = sizeof(struct pack_header);
	do {
		ssize_t r = xread(fd, c, remaining);
		if (r <= 0)
			/* "eof before pack header was fully read" */
			return PH_ERROR_EOF;
		remaining -= r;
		c += r;
	} while (remaining > 0);
	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;
}