sha1_file.c 47.1 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
const unsigned char null_sha1[20];
J
Junio C Hamano 已提交
26

27 28
static unsigned int sha1_file_open_flag = O_NOATIME;

29 30 31 32 33 34 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
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 */
};
63 64 65 66 67 68 69 70 71 72 73 74 75 76

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;
}

77 78 79
int safe_create_leading_directories(char *path)
{
	char *pos = path;
80 81
	struct stat st;

82 83
	if (*pos == '/')
		pos++;
84 85 86 87 88 89

	while (pos) {
		pos = strchr(pos, '/');
		if (!pos)
			break;
		*pos = 0;
90 91 92
		if (!stat(path, &st)) {
			/* path exists */
			if (!S_ISDIR(st.st_mode)) {
93
				*pos = '/';
94
				return -3;
95
			}
96
		}
97 98 99 100
		else if (mkdir(path, 0777)) {
			*pos = '/';
			return -1;
		}
101 102 103 104
		else if (adjust_shared_perm(path)) {
			*pos = '/';
			return -2;
		}
105 106 107 108
		*pos++ = '/';
	}
	return 0;
}
109

110 111
char * sha1_to_hex(const unsigned char *sha1)
{
L
Linus Torvalds 已提交
112 113
	static int bufno;
	static char hexbuffer[4][50];
114
	static const char hex[] = "0123456789abcdef";
L
Linus Torvalds 已提交
115
	char *buffer = hexbuffer[3 & ++bufno], *buf = buffer;
116 117 118 119 120 121 122
	int i;

	for (i = 0; i < 20; i++) {
		unsigned int val = *sha1++;
		*buf++ = hex[val >> 4];
		*buf++ = hex[val & 0xf];
	}
123 124
	*buf = '\0';

125 126 127
	return buffer;
}

128 129 130 131 132 133 134 135 136 137 138 139
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];
	}
}

140 141
/*
 * NOTE! This returns a statically allocated buffer, so you have to be
142
 * careful about using it. Do a "xstrdup()" if you need to save the
143
 * filename.
144 145 146
 *
 * 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
147
 * DB_ENVIRONMENT environment variable if it is not found in
148
 * the primary object database.
149 150 151 152 153 154
 */
char *sha1_file_name(const unsigned char *sha1)
{
	static char *name, *base;

	if (!base) {
J
Junio C Hamano 已提交
155
		const char *sha1_file_directory = get_object_directory();
156
		int len = strlen(sha1_file_directory);
157
		base = xmalloc(len + 60);
158 159 160 161 162 163
		memcpy(base, sha1_file_directory, len);
		memset(base+len, 0, 60);
		base[len] = '/';
		base[len+3] = '/';
		name = base + len + 1;
	}
164
	fill_sha1_path(name, sha1);
165 166 167
	return base;
}

168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217
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;
}

218 219
struct alternate_object_database *alt_odb_list;
static struct alternate_object_database **alt_odb_tail;
220

M
Martin Waitz 已提交
221 222
static void read_info_alternates(const char * alternates, int depth);

J
Junio C Hamano 已提交
223 224
/*
 * Prepare alternate object database registry.
225 226 227 228 229
 *
 * 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,
230 231
 * whose contents is similar to that environment variable but can be
 * LF separated.  Its base points at a statically allocated buffer that
232 233 234 235 236
 * 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
237
 */
M
Martin Waitz 已提交
238
static int link_alt_odb_entry(const char * entry, int len, const char * relative_base, int depth)
239
{
M
Martin Waitz 已提交
240
	struct stat st;
241
	const char *objdir = get_object_directory();
M
Martin Waitz 已提交
242 243 244 245 246
	struct alternate_object_database *ent;
	struct alternate_object_database *alt;
	/* 43 = 40-byte + 2 '/' + terminating NUL */
	int pfxlen = len;
	int entlen = pfxlen + 43;
247
	int base_len = -1;
248

M
Martin Waitz 已提交
249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316
	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;
	}

317
	last = alt;
318 319 320 321 322 323 324 325
	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
326 327
		while (cp < ep && *cp != sep)
			cp++;
328
		if (last != cp) {
M
Martin Waitz 已提交
329 330 331 332 333 334
			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);
335
			}
336
		}
337
		while (cp < ep && *cp == sep)
338 339
			cp++;
		last = cp;
340
	}
341 342
}

M
Martin Waitz 已提交
343
static void read_info_alternates(const char * relative_base, int depth)
344
{
345
	char *map;
346
	struct stat st;
M
Martin Waitz 已提交
347 348
	char path[PATH_MAX];
	int fd;
349

M
Martin Waitz 已提交
350
	sprintf(path, "%s/info/alternates", relative_base);
351 352 353 354 355
	fd = open(path, O_RDONLY);
	if (fd < 0)
		return;
	if (fstat(fd, &st) || (st.st_size == 0)) {
		close(fd);
356
		return;
357
	}
358 359 360 361 362
	map = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
	close(fd);
	if (map == MAP_FAILED)
		return;

M
Martin Waitz 已提交
363 364
	link_alt_odb_entries(map, map + st.st_size, '\n', relative_base, depth);

365
	munmap(map, st.st_size);
366 367
}

M
Martin Waitz 已提交
368 369
void prepare_alt_odb(void)
{
T
Timo Hirvonen 已提交
370
	const char *alt;
M
Martin Waitz 已提交
371 372 373 374 375 376 377 378 379 380 381 382

	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);
}

383 384 385
static char *find_sha1_file(const unsigned char *sha1, struct stat *st)
{
	char *name = sha1_file_name(sha1);
386
	struct alternate_object_database *alt;
387 388 389

	if (!stat(name, st))
		return name;
390
	prepare_alt_odb();
391 392
	for (alt = alt_odb_list; alt; alt = alt->next) {
		name = alt->name;
393
		fill_sha1_path(name, sha1);
394 395
		if (!stat(alt->base, st))
			return alt->base;
396 397 398 399
	}
	return NULL;
}

400 401 402
static unsigned int pack_used_ctr;
static size_t pack_mapped;
static size_t page_size;
403
struct packed_git *packed_git;
404 405 406 407 408 409 410 411

static int check_packed_git_idx(const char *path, unsigned long *idx_size_,
				void **idx_map_)
{
	void *idx_map;
	unsigned int *index;
	unsigned long idx_size;
	int nr, i;
412
	int fd = open(path, O_RDONLY);
413 414 415 416 417 418 419 420 421 422 423 424 425 426
	struct stat st;
	if (fd < 0)
		return -1;
	if (fstat(fd, &st)) {
		close(fd);
		return -1;
	}
	idx_size = st.st_size;
	idx_map = mmap(NULL, idx_size, PROT_READ, MAP_PRIVATE, fd, 0);
	close(fd);
	if (idx_map == MAP_FAILED)
		return -1;

	index = idx_map;
427 428
	*idx_map_ = idx_map;
	*idx_size_ = idx_size;
429 430

	/* check index map */
431
	if (idx_size < 4*256 + 20 + 20)
432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453
		return error("index file too small");
	nr = 0;
	for (i = 0; i < 256; i++) {
		unsigned int n = ntohl(index[i]);
		if (n < nr)
			return error("non-monotonic index");
		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)
		return error("wrong index file size");

	return 0;
}

454
static int unuse_one_window(void)
455
{
456 457
	struct packed_git *p, *lru_p = NULL;
	struct pack_window *w, *w_l, *lru_w = NULL, *lru_l = NULL;
458 459

	for (p = packed_git; p; p = p->next) {
460 461 462 463 464 465 466 467 468 469
		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;
				}
			}
			w_l = w;
		}
470
	}
471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486
	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;
			if (!lru_p->windows) {
				close(lru_p->pack_fd);
				lru_p->pack_fd = -1;
			}
		}
		free(lru_w);
		return 1;
	}
	return 0;
487 488
}

489
void unuse_pack(struct pack_window **w_cursor)
490
{
491 492 493 494 495
	struct pack_window *w = *w_cursor;
	if (w) {
		w->inuse_cnt--;
		*w_cursor = NULL;
	}
496 497
}

498
static void open_packed_git(struct packed_git *p)
499
{
500 501 502 503 504 505 506 507 508 509
	struct stat st;
	struct pack_header hdr;
	unsigned char sha1[20];
	unsigned char *idx_sha1;

	p->pack_fd = open(p->pack_name, O_RDONLY);
	if (p->pack_fd < 0 || fstat(p->pack_fd, &st))
		die("packfile %s cannot be opened", p->pack_name);

	/* If we created the struct before we had the pack we lack size. */
510 511 512 513
	if (!p->pack_size) {
		if (!S_ISREG(st.st_mode))
			die("packfile %s not a regular file", p->pack_name);
		p->pack_size = st.st_size;
514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539
	} else if (p->pack_size != st.st_size)
		die("packfile %s size changed", p->pack_name);

	/* Verify we recognize this pack file format. */
	read_or_die(p->pack_fd, &hdr, sizeof(hdr));
	if (hdr.hdr_signature != htonl(PACK_SIGNATURE))
		die("file %s is not a GIT packfile", p->pack_name);
	if (!pack_version_ok(hdr.hdr_version))
		die("packfile %s is version %u and not supported"
			" (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))
		die("packfile %s claims to have %u objects"
			" 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)
		die("end of packfile %s is unavailable", p->pack_name);
	read_or_die(p->pack_fd, sha1, sizeof(sha1));
	idx_sha1 = ((unsigned char *)p->index_base) + p->index_size - 40;
	if (hashcmp(sha1, idx_sha1))
		die("packfile %s does not match index", p->pack_name);
}

540 541 542 543 544 545 546 547 548 549 550 551 552
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);
}

553 554 555 556
unsigned char* use_pack(struct packed_git *p,
		struct pack_window **w_cursor,
		unsigned long offset,
		unsigned int *left)
557
{
558
	struct pack_window *win = *w_cursor;
559

560 561
	if (p->pack_fd == -1)
		open_packed_git(p);
562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592

	/* 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) {
			if (!page_size)
				page_size = getpagesize();
			win = xcalloc(1, sizeof(*win));
			win->offset = (offset / page_size) * page_size;
			win->len = p->pack_size - win->offset;
			if (win->len > packed_git_window_size)
				win->len = packed_git_window_size;
			pack_mapped += win->len;
			while (packed_git_limit < pack_mapped && unuse_one_window())
				; /* nothing */
			win->base = mmap(NULL, win->len,
				PROT_READ, MAP_PRIVATE,
				p->pack_fd, win->offset);
			if (win->base == MAP_FAILED)
593 594 595
				die("packfile %s cannot be mapped: %s",
					p->pack_name,
					strerror(errno));
596 597 598
			win->next = p->windows;
			p->windows = win;
		}
599
	}
600 601 602 603 604
	if (win != *w_cursor) {
		win->last_used = pack_used_ctr++;
		win->inuse_cnt++;
		*w_cursor = win;
	}
605
	offset -= win->offset;
606 607 608
	if (left)
		*left = win->len - offset;
	return win->base + offset;
609 610
}

611
struct packed_git *add_packed_git(char *path, int path_len, int local)
612 613 614 615 616
{
	struct stat st;
	struct packed_git *p;
	unsigned long idx_size;
	void *idx_map;
617
	unsigned char sha1[20];
618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636

	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;
637
	p->windows = NULL;
638
	p->pack_fd = -1;
639
	p->pack_local = local;
P
Pavel Roskin 已提交
640
	if ((path_len > 44) && !get_sha1_hex(path + path_len - 44, sha1))
641
		hashcpy(p->sha1, sha1);
642 643 644
	return p;
}

645
struct packed_git *parse_pack_index(unsigned char *sha1)
646 647 648 649 650
{
	char *path = sha1_pack_index_name(sha1);
	return parse_pack_index_file(sha1, path);
}

651
struct packed_git *parse_pack_index_file(const unsigned char *sha1, char *idx_path)
652 653 654 655
{
	struct packed_git *p;
	unsigned long idx_size;
	void *idx_map;
656
	char *path;
657

658
	if (check_packed_git_idx(idx_path, &idx_size, &idx_map))
659 660 661 662 663 664 665 666 667 668
		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;
669
	p->windows = NULL;
670
	p->pack_fd = -1;
671
	hashcpy(p->sha1, sha1);
672 673 674 675 676 677 678 679 680
	return p;
}

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

681
static void prepare_packed_git_one(char *objdir, int local)
682 683 684 685 686 687 688 689 690
{
	char path[PATH_MAX];
	int len;
	DIR *dir;
	struct dirent *de;

	sprintf(path, "%s/pack", objdir);
	len = strlen(path);
	dir = opendir(path);
691
	if (!dir) {
J
Junio C Hamano 已提交
692
		if (errno != ENOENT)
693
			error("unable to open object pack directory: %s: %s",
J
Junio C Hamano 已提交
694
			      path, strerror(errno));
695
		return;
696
	}
697 698 699 700 701
	path[len++] = '/';
	while ((de = readdir(dir)) != NULL) {
		int namelen = strlen(de->d_name);
		struct packed_git *p;

702
		if (!has_extension(de->d_name, ".idx"))
703 704 705 706
			continue;

		/* we have .idx.  Is it a file we can map? */
		strcpy(path + len, de->d_name);
707 708 709 710 711 712
		for (p = packed_git; p; p = p->next) {
			if (!memcmp(path, p->pack_name, len + namelen - 4))
				break;
		}
		if (p)
			continue;
713
		p = add_packed_git(path, len + namelen, local);
714 715 716 717 718
		if (!p)
			continue;
		p->next = packed_git;
		packed_git = p;
	}
719
	closedir(dir);
720 721
}

722
static int prepare_packed_git_run_once = 0;
723
void prepare_packed_git(void)
724
{
725
	struct alternate_object_database *alt;
726

727
	if (prepare_packed_git_run_once)
728
		return;
729
	prepare_packed_git_one(get_object_directory(), 1);
730
	prepare_alt_odb();
731
	for (alt = alt_odb_list; alt; alt = alt->next) {
732
		alt->name[-1] = 0;
733
		prepare_packed_git_one(alt->base, 0);
734
		alt->name[-1] = '/';
735
	}
736 737 738
	prepare_packed_git_run_once = 1;
}

739
void reprepare_packed_git(void)
740 741 742
{
	prepare_packed_git_run_once = 0;
	prepare_packed_git();
743 744
}

745
int check_sha1_signature(const unsigned char *sha1, void *map, unsigned long size, const char *type)
746 747
{
	unsigned char real_sha1[20];
748
	hash_sha1_file(map, size, type, real_sha1);
749
	return hashcmp(sha1, real_sha1) ? -1 : 0;
750 751
}

752
void *map_sha1_file(const unsigned char *sha1, unsigned long *size)
753 754 755
{
	struct stat st;
	void *map;
756
	int fd;
757 758 759 760 761
	char *filename = find_sha1_file(sha1, &st);

	if (!filename) {
		return NULL;
	}
762

763
	fd = open(filename, O_RDONLY | sha1_file_open_flag);
764
	if (fd < 0) {
765 766 767 768 769 770 771 772 773 774 775
		/* 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;
		}

776 777 778
		/* If it failed once, it will probably fail again.
		 * Stop using O_NOATIME
		 */
779
		sha1_file_open_flag = 0;
780 781 782
	}
	map = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
	close(fd);
P
Pavel Roskin 已提交
783
	if (map == MAP_FAILED)
784 785 786 787 788
		return NULL;
	*size = st.st_size;
	return map;
}

789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804
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;
}

805
unsigned long unpack_object_header_gently(const unsigned char *buf, unsigned long len, enum object_type *type, unsigned long *sizep)
806
{
807
	unsigned shift;
808 809
	unsigned char c;
	unsigned long size;
810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835
	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 */
836
	};
837
	enum object_type type;
838

839 840 841 842 843
	/* Get the data stream */
	memset(stream, 0, sizeof(*stream));
	stream->next_in = map;
	stream->avail_in = mapsize;
	stream->next_out = buffer;
844 845
	stream->avail_out = bufsiz;

846
	if (legacy_loose_object(map)) {
847 848 849 850
		inflateInit(stream);
		return inflate(stream, 0);
	}

851 852
	used = unpack_object_header_gently(map, mapsize, &type, &size);
	if (!used || !valid_loose_object_type[type])
853
		return -1;
854 855
	map += used;
	mapsize -= used;
856 857 858 859

	/* Set up the stream for the rest.. */
	stream->next_in = map;
	stream->avail_in = mapsize;
860
	inflateInit(stream);
861 862

	/* And generate the fake traditional header */
863 864
	stream->total_out = 1 + snprintf(buffer, bufsiz, "%s %lu",
					 type_names[type], size);
865
	return 0;
866 867
}

L
Linus Torvalds 已提交
868
static void *unpack_sha1_rest(z_stream *stream, void *buffer, unsigned long size)
869 870
{
	int bytes = strlen(buffer) + 1;
871
	unsigned char *buf = xmalloc(1+size);
872
	unsigned long n;
873

874 875 876 877 878
	n = stream->total_out - bytes;
	if (n > size)
		n = size;
	memcpy(buf, (char *) buffer + bytes, n);
	bytes = n;
879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894
	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.
 */
895
static int parse_sha1_header(char *hdr, char *type, unsigned long *sizep)
896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939
{
	int i;
	unsigned long size;

	/*
	 * The type can be at most ten bytes (including the 
	 * terminating '\0' that we add), and is followed by
	 * a space. 
	 */
	i = 10;
	for (;;) {
		char c = *hdr++;
		if (c == ' ')
			break;
		if (!--i)
			return -1;
		*type++ = c;
	}
	*type = 0;

	/*
	 * 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
	 */
	return *hdr ? -1 : 0;
}

940 941
void * unpack_sha1_file(void *map, unsigned long mapsize, char *type, unsigned long *size)
{
942
	int ret;
943
	z_stream stream;
944
	char hdr[8192];
945

946 947
	ret = unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr));
	if (ret < Z_OK || parse_sha1_header(hdr, type, size) < 0)
948 949
		return NULL;

950
	return unpack_sha1_rest(&stream, hdr, *size);
951 952
}

953
static unsigned long get_delta_base(struct packed_git *p,
954
				    struct pack_window **w_curs,
955 956 957 958 959
				    unsigned long offset,
				    enum object_type kind,
				    unsigned long delta_obj_offset,
				    unsigned long *base_obj_offset)
{
960
	unsigned char *base_info = use_pack(p, w_curs, offset, NULL);
961 962
	unsigned long base_offset;

963 964 965 966 967 968
	/* 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.
	 */
969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996
	if (kind == OBJ_OFS_DELTA) {
		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");
		offset += used;
	} else if (kind == OBJ_REF_DELTA) {
		/* 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));
		offset += 20;
	} else
		die("I am totally screwed");
	*base_obj_offset = base_offset;
	return offset;
}

J
Junio C Hamano 已提交
997
/* forward declaration for a mutually recursive function */
N
Nicolas Pitre 已提交
998
static int packed_object_info(struct packed_git *p, unsigned long offset,
J
Junio C Hamano 已提交
999 1000
			      char *type, unsigned long *sizep);

N
Nicolas Pitre 已提交
1001
static int packed_delta_info(struct packed_git *p,
1002
			     struct pack_window **w_curs,
N
Nicolas Pitre 已提交
1003
			     unsigned long offset,
1004 1005
			     enum object_type kind,
			     unsigned long obj_offset,
1006
			     char *type,
N
Nicolas Pitre 已提交
1007
			     unsigned long *sizep)
1008
{
N
Nicolas Pitre 已提交
1009
	unsigned long base_offset;
J
Junio C Hamano 已提交
1010

1011 1012
	offset = get_delta_base(p, w_curs, offset, kind,
		obj_offset, &base_offset);
J
Junio C Hamano 已提交
1013

1014 1015 1016 1017 1018
	/* 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
1019
	if (packed_object_info(p, base_offset, type, NULL))
1020
		die("cannot get info for delta-pack base");
1021

1022 1023
	if (sizep) {
		const unsigned char *data;
1024
		unsigned char delta_head[20], *in;
1025 1026 1027 1028 1029 1030 1031 1032 1033
		unsigned long result_size;
		z_stream stream;
		int st;

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

		inflateInit(&stream);
1034 1035 1036 1037 1038 1039 1040
		do {
			in = use_pack(p, w_curs, offset, &stream.avail_in);
			stream.next_in = in;
			st = inflate(&stream, Z_FINISH);
			offset += stream.next_in - in;
		} while ((st == Z_OK || st == Z_BUF_ERROR)
			&& stream.total_out < sizeof(delta_head));
1041 1042 1043 1044 1045 1046 1047 1048 1049
		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;
1050 1051 1052

		/* ignore base size */
		get_delta_hdr_size(&data, delta_head+sizeof(delta_head));
1053

1054
		/* Read the result size */
1055
		result_size = get_delta_hdr_size(&data, delta_head+sizeof(delta_head));
1056 1057
		*sizep = result_size;
	}
1058 1059 1060
	return 0;
}

1061 1062 1063 1064 1065
static unsigned long unpack_object_header(struct packed_git *p,
		struct pack_window **w_curs,
		unsigned long offset,
		enum object_type *type,
		unsigned long *sizep)
1066
{
1067 1068
	unsigned char *base;
	unsigned int left;
1069
	unsigned long used;
1070

1071 1072 1073 1074 1075 1076
	/* use_pack() assures us we have [base, base + 20) available
	 * as a range that we can look at at.  (Its actually the hash
	 * size that is assurred.)  With our object header encoding
	 * the maximum deflated object size is 2^137, which is just
	 * insane, so we know won't exceed what we have been given.
	 */
1077 1078
	base = use_pack(p, w_curs, offset, &left);
	used = unpack_object_header_gently(base, left, type, sizep);
1079 1080 1081 1082
	if (!used)
		die("object offset outside of pack file");

	return offset + used;
1083 1084
}

N
Nicolas Pitre 已提交
1085 1086
void packed_object_info_detail(struct packed_git *p,
			       unsigned long offset,
1087 1088 1089
			       char *type,
			       unsigned long *size,
			       unsigned long *store_size,
J
Junio C Hamano 已提交
1090
			       unsigned int *delta_chain_length,
1091 1092
			       unsigned char *base_sha1)
{
1093
	struct pack_window *w_curs = NULL;
1094
	unsigned long obj_offset, val;
N
Nicolas Pitre 已提交
1095
	unsigned char *next_sha1;
1096 1097
	enum object_type kind;

N
Nicolas Pitre 已提交
1098
	*delta_chain_length = 0;
1099
	obj_offset = offset;
1100
	offset = unpack_object_header(p, &w_curs, offset, &kind, size);
N
Nicolas Pitre 已提交
1101 1102 1103 1104

	for (;;) {
		switch (kind) {
		default:
1105
			die("pack %s contains unknown object type %d",
N
Nicolas Pitre 已提交
1106 1107 1108 1109 1110 1111 1112
			    p->pack_name, kind);
		case OBJ_COMMIT:
		case OBJ_TREE:
		case OBJ_BLOB:
		case OBJ_TAG:
			strcpy(type, type_names[kind]);
			*store_size = 0; /* notyet */
1113
			unuse_pack(&w_curs);
N
Nicolas Pitre 已提交
1114
			return;
1115
		case OBJ_OFS_DELTA:
1116 1117
			get_delta_base(p, &w_curs, offset, kind,
				obj_offset, &offset);
1118 1119 1120 1121 1122
			if (*delta_chain_length == 0) {
				/* TODO: find base_sha1 as pointed by offset */
			}
			break;
		case OBJ_REF_DELTA:
1123
			next_sha1 = use_pack(p, &w_curs, offset, NULL);
N
Nicolas Pitre 已提交
1124 1125 1126 1127 1128
			if (*delta_chain_length == 0)
				hashcpy(base_sha1, next_sha1);
			offset = find_pack_entry_one(next_sha1, p);
			break;
		}
1129
		obj_offset = offset;
1130
		offset = unpack_object_header(p, &w_curs, offset, &kind, &val);
N
Nicolas Pitre 已提交
1131
		(*delta_chain_length)++;
1132 1133 1134
	}
}

N
Nicolas Pitre 已提交
1135
static int packed_object_info(struct packed_git *p, unsigned long offset,
1136 1137
			      char *type, unsigned long *sizep)
{
1138
	struct pack_window *w_curs = NULL;
1139
	unsigned long size, obj_offset = offset;
1140
	enum object_type kind;
1141
	int r;
1142

1143
	offset = unpack_object_header(p, &w_curs, offset, &kind, &size);
1144

1145
	switch (kind) {
1146 1147
	case OBJ_OFS_DELTA:
	case OBJ_REF_DELTA:
1148 1149 1150 1151
		r = packed_delta_info(p, &w_curs, offset, kind,
			obj_offset, type, sizep);
		unuse_pack(&w_curs);
		return r;
1152 1153 1154 1155
	case OBJ_COMMIT:
	case OBJ_TREE:
	case OBJ_BLOB:
	case OBJ_TAG:
1156
		strcpy(type, type_names[kind]);
1157
		unuse_pack(&w_curs);
1158
		break;
1159
	default:
1160
		die("pack %s contains unknown object type %d",
1161
		    p->pack_name, kind);
1162
	}
1163 1164
	if (sizep)
		*sizep = size;
1165 1166 1167
	return 0;
}

1168
static void *unpack_compressed_entry(struct packed_git *p,
1169
				    struct pack_window **w_curs,
1170 1171
				    unsigned long offset,
				    unsigned long size)
1172 1173 1174
{
	int st;
	z_stream stream;
1175
	unsigned char *buffer, *in;
1176 1177 1178 1179 1180 1181 1182 1183

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

	inflateInit(&stream);
1184 1185 1186 1187 1188 1189
	do {
		in = use_pack(p, w_curs, offset, &stream.avail_in);
		stream.next_in = in;
		st = inflate(&stream, Z_FINISH);
		offset += stream.next_in - in;
	} while (st == Z_OK || st == Z_BUF_ERROR);
1190 1191 1192 1193 1194 1195 1196 1197 1198
	inflateEnd(&stream);
	if ((st != Z_STREAM_END) || stream.total_out != size) {
		free(buffer);
		return NULL;
	}

	return buffer;
}

1199
static void *unpack_delta_entry(struct packed_git *p,
1200
				struct pack_window **w_curs,
1201
				unsigned long offset,
1202
				unsigned long delta_size,
1203 1204
				enum object_type kind,
				unsigned long obj_offset,
1205
				char *type,
1206
				unsigned long *sizep)
1207
{
1208
	void *delta_data, *result, *base;
N
Nicolas Pitre 已提交
1209 1210
	unsigned long result_size, base_size, base_offset;

1211 1212
	offset = get_delta_base(p, w_curs, offset, kind,
		obj_offset, &base_offset);
1213
	base = unpack_entry(p, base_offset, type, &base_size);
1214
	if (!base)
N
Nicolas Pitre 已提交
1215 1216
		die("failed to read delta base object at %lu from %s",
		    base_offset, p->pack_name);
1217

1218
	delta_data = unpack_compressed_entry(p, w_curs, offset, delta_size);
1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229
	result = patch_delta(base, base_size,
			     delta_data, delta_size,
			     &result_size);
	if (!result)
		die("failed to apply delta");
	free(delta_data);
	free(base);
	*sizep = result_size;
	return result;
}

1230
void *unpack_entry(struct packed_git *p, unsigned long offset,
1231 1232
			  char *type, unsigned long *sizep)
{
1233
	struct pack_window *w_curs = NULL;
1234 1235
	unsigned long size, obj_offset = offset;
	enum object_type kind;
1236
	void *retval;
1237

1238
	offset = unpack_object_header(p, &w_curs, offset, &kind, &size);
1239
	switch (kind) {
1240 1241
	case OBJ_OFS_DELTA:
	case OBJ_REF_DELTA:
1242 1243
		retval = unpack_delta_entry(p, &w_curs, offset, size,
			kind, obj_offset, type, sizep);
1244
		break;
1245 1246 1247 1248
	case OBJ_COMMIT:
	case OBJ_TREE:
	case OBJ_BLOB:
	case OBJ_TAG:
1249 1250
		strcpy(type, type_names[kind]);
		*sizep = size;
1251
		retval = unpack_compressed_entry(p, &w_curs, offset, size);
1252
		break;
1253
	default:
1254
		die("unknown object type %i in %s", kind, p->pack_name);
1255
	}
1256
	unuse_pack(&w_curs);
1257
	return retval;
1258 1259
}

1260 1261
int num_packed_objects(const struct packed_git *p)
{
1262
	/* See check_packed_git_idx() */
1263 1264 1265 1266 1267 1268 1269 1270 1271
	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;
1272
	hashcpy(sha1, (unsigned char *) index + (24 * n) + 4);
1273 1274 1275
	return 0;
}

N
Nicolas Pitre 已提交
1276 1277
unsigned long find_pack_entry_one(const unsigned char *sha1,
				  struct packed_git *p)
1278
{
1279
	unsigned int *level1_ofs = p->index_base;
1280 1281 1282 1283 1284 1285
	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;
1286
		int cmp = hashcmp((unsigned char *)index + (24 * mi) + 4, sha1);
N
Nicolas Pitre 已提交
1287 1288
		if (!cmp)
			return ntohl(*((unsigned int *) ((char *) index + (24 * mi))));
1289 1290 1291 1292 1293 1294 1295 1296
		if (cmp > 0)
			hi = mi;
		else
			lo = mi+1;
	} while (lo < hi);
	return 0;
}

1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314
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;
}

1315
static int find_pack_entry(const unsigned char *sha1, struct pack_entry *e, const char **ignore_packed)
1316 1317
{
	struct packed_git *p;
N
Nicolas Pitre 已提交
1318 1319
	unsigned long offset;

1320 1321 1322
	prepare_packed_git();

	for (p = packed_git; p; p = p->next) {
1323 1324 1325
		if (ignore_packed) {
			const char **ig;
			for (ig = ignore_packed; *ig; ig++)
1326
				if (!matches_pack_name(p, *ig))
1327 1328 1329 1330
					break;
			if (*ig)
				continue;
		}
N
Nicolas Pitre 已提交
1331 1332 1333 1334 1335
		offset = find_pack_entry_one(sha1, p);
		if (offset) {
			e->offset = offset;
			e->p = p;
			hashcpy(e->sha1, sha1);
1336
			return 1;
N
Nicolas Pitre 已提交
1337
		}
1338 1339 1340 1341
	}
	return 0;
}

1342 1343 1344 1345 1346 1347
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
1348
		if (find_pack_entry_one(sha1, p))
1349 1350 1351 1352 1353 1354
			return p;
	}
	return NULL;
	
}

1355
static int sha1_loose_object_info(const unsigned char *sha1, char *type, unsigned long *sizep)
1356
{
1357
	int status;
1358 1359 1360
	unsigned long mapsize, size;
	void *map;
	z_stream stream;
1361
	char hdr[128];
1362

1363
	map = map_sha1_file(sha1, &mapsize);
1364 1365
	if (!map)
		return error("unable to find %s", sha1_to_hex(sha1));
1366 1367 1368 1369 1370
	if (unpack_sha1_header(&stream, map, mapsize, hdr, sizeof(hdr)) < 0)
		status = error("unable to unpack %s header",
			       sha1_to_hex(sha1));
	if (parse_sha1_header(hdr, type, &size) < 0)
		status = error("unable to parse %s header", sha1_to_hex(sha1));
1371
	else {
1372
		status = 0;
1373 1374
		if (sizep)
			*sizep = size;
1375 1376 1377 1378 1379 1380
	}
	inflateEnd(&stream);
	munmap(map, mapsize);
	return status;
}

1381 1382 1383 1384 1385 1386 1387 1388 1389
int sha1_object_info(const unsigned char *sha1, char *type, unsigned long *sizep)
{
	struct pack_entry e;

	if (!find_pack_entry(sha1, &e, NULL)) {
		reprepare_packed_git();
		if (!find_pack_entry(sha1, &e, NULL))
			return sha1_loose_object_info(sha1, type, sizep);
	}
1390
	return packed_object_info(e.p, e.offset, type, sizep);
1391 1392
}

1393 1394 1395 1396
static void *read_packed_sha1(const unsigned char *sha1, char *type, unsigned long *size)
{
	struct pack_entry e;

1397
	if (!find_pack_entry(sha1, &e, NULL)) {
1398 1399 1400
		error("cannot read sha1_file for %s", sha1_to_hex(sha1));
		return NULL;
	}
1401
	return unpack_entry(e.p, e.offset, type, size);
1402 1403
}

1404 1405 1406 1407
void * read_sha1_file(const unsigned char *sha1, char *type, unsigned long *size)
{
	unsigned long mapsize;
	void *map, *buf;
1408
	struct pack_entry e;
1409

1410
	if (find_pack_entry(sha1, &e, NULL))
1411
		return read_packed_sha1(sha1, type, size);
1412
	map = map_sha1_file(sha1, &mapsize);
1413 1414 1415 1416 1417
	if (map) {
		buf = unpack_sha1_file(map, mapsize, type, size);
		munmap(map, mapsize);
		return buf;
	}
1418
	reprepare_packed_git();
1419
	if (find_pack_entry(sha1, &e, NULL))
1420
		return read_packed_sha1(sha1, type, size);
1421
	return NULL;
1422 1423
}

1424
void *read_object_with_reference(const unsigned char *sha1,
1425
				 const char *required_type,
1426 1427
				 unsigned long *size,
				 unsigned char *actual_sha1_return)
1428 1429 1430 1431
{
	char type[20];
	void *buffer;
	unsigned long isize;
1432
	unsigned char actual_sha1[20];
1433

1434
	hashcpy(actual_sha1, sha1);
1435 1436 1437
	while (1) {
		int ref_length = -1;
		const char *ref_type = NULL;
1438

1439 1440 1441 1442 1443 1444
		buffer = read_sha1_file(actual_sha1, type, &isize);
		if (!buffer)
			return NULL;
		if (!strcmp(type, required_type)) {
			*size = isize;
			if (actual_sha1_return)
1445
				hashcpy(actual_sha1_return, actual_sha1);
1446 1447 1448
			return buffer;
		}
		/* Handle references */
1449
		else if (!strcmp(type, commit_type))
1450
			ref_type = "tree ";
1451
		else if (!strcmp(type, tag_type))
1452 1453 1454 1455 1456 1457
			ref_type = "object ";
		else {
			free(buffer);
			return NULL;
		}
		ref_length = strlen(ref_type);
1458

1459
		if (memcmp(buffer, ref_type, ref_length) ||
1460
		    get_sha1_hex((char *) buffer + ref_length, actual_sha1)) {
1461 1462 1463
			free(buffer);
			return NULL;
		}
1464
		free(buffer);
1465 1466
		/* Now we have the ID of the referred-to object in
		 * actual_sha1.  Check again. */
1467 1468 1469
	}
}

R
Rene Scharfe 已提交
1470 1471 1472
static void write_sha1_file_prepare(void *buf, unsigned long len,
                                    const char *type, unsigned char *sha1,
                                    unsigned char *hdr, int *hdrlen)
1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485
{
	SHA_CTX c;

	/* Generate the header */
	*hdrlen = sprintf((char *)hdr, "%s %lu", type, len)+1;

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

1486 1487 1488 1489 1490 1491
/*
 * 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.
 */
1492
static int link_temp_to_file(const char *tmpfile, const char *filename)
1493 1494
{
	int ret;
S
Shawn Pearce 已提交
1495
	char *dir;
1496 1497 1498 1499 1500

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

	/*
S
Shawn Pearce 已提交
1501
	 * Try to mkdir the last path component if that failed.
1502 1503 1504 1505 1506 1507
	 *
	 * 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
1508 1509 1510
	dir = strrchr(filename, '/');
	if (dir) {
		*dir = 0;
1511
		if (!mkdir(filename, 0777) && adjust_shared_perm(filename)) {
1512
			*dir = '/';
S
Shawn Pearce 已提交
1513
			return -2;
1514
		}
S
Shawn Pearce 已提交
1515 1516 1517 1518
		*dir = '/';
		if (!link(tmpfile, filename))
			return 0;
		ret = errno;
1519 1520 1521 1522 1523 1524 1525
	}
	return ret;
}

/*
 * Move the just written object into its final resting place
 */
1526
int move_temp_to_file(const char *tmpfile, const char *filename)
1527 1528
{
	int ret = link_temp_to_file(tmpfile, filename);
1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542

	/*
	 * 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))
1543
			return 0;
1544
		ret = errno;
1545 1546 1547 1548
	}
	unlink(tmpfile);
	if (ret) {
		if (ret != EEXIST) {
1549
			return error("unable to write sha1 filename %s: %s\n", filename, strerror(ret));
1550 1551 1552 1553 1554 1555 1556
		}
		/* FIXME!!! Collision check here ? */
	}

	return 0;
}

L
Linus Torvalds 已提交
1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570
static int write_buffer(int fd, const void *buf, size_t len)
{
	while (len) {
		ssize_t size;

		size = write(fd, buf, len);
		if (!size)
			return error("file write: disk full");
		if (size < 0) {
			if (errno == EINTR || errno == EAGAIN)
				continue;
			return error("file write error (%s)", strerror(errno));
		}
		len -= size;
1571
		buf = (char *) buf + size;
L
Linus Torvalds 已提交
1572 1573 1574 1575
	}
	return 0;
}

1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618
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)
{
	int obj_type, hdr;

	if (use_legacy_headers) {
		while (deflate(stream, 0) == Z_OK)
			/* nothing */;
		return;
	}
	if (!strcmp(type, blob_type))
		obj_type = OBJ_BLOB;
	else if (!strcmp(type, tree_type))
		obj_type = OBJ_TREE;
	else if (!strcmp(type, commit_type))
		obj_type = OBJ_COMMIT;
	else if (!strcmp(type, tag_type))
		obj_type = OBJ_TAG;
	else
		die("trying to generate bogus object of type '%s'", type);
	hdr = write_binary_header(stream->next_out, obj_type, len);
	stream->total_out = hdr;
	stream->next_out += hdr;
	stream->avail_out -= hdr;
}

R
Rene Scharfe 已提交
1619 1620 1621 1622 1623 1624 1625 1626 1627
int hash_sha1_file(void *buf, unsigned long len, const char *type,
                   unsigned char *sha1)
{
	unsigned char hdr[50];
	int hdrlen;
	write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen);
	return 0;
}

1628
int write_sha1_file(void *buf, unsigned long len, const char *type, unsigned char *returnsha1)
1629 1630
{
	int size;
1631
	unsigned char *compressed;
1632 1633
	z_stream stream;
	unsigned char sha1[20];
1634
	char *filename;
1635
	static char tmpfile[PATH_MAX];
1636
	unsigned char hdr[50];
1637
	int fd, hdrlen;
1638

1639 1640 1641
	/* 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
1642 1643
	write_sha1_file_prepare(buf, len, type, sha1, hdr, &hdrlen);
	filename = sha1_file_name(sha1);
1644
	if (returnsha1)
1645
		hashcpy(returnsha1, sha1);
1646 1647
	if (has_sha1_file(sha1))
		return 0;
1648 1649
	fd = open(filename, O_RDONLY);
	if (fd >= 0) {
1650
		/*
1651 1652
		 * FIXME!!! We might do collision checking here, but we'd
		 * need to uncompress the old file and check it. Later.
1653
		 */
1654
		close(fd);
1655 1656 1657
		return 0;
	}

1658
	if (errno != ENOENT) {
1659
		return error("sha1 file %s: %s\n", filename, strerror(errno));
1660 1661 1662
	}

	snprintf(tmpfile, sizeof(tmpfile), "%s/obj_XXXXXX", get_object_directory());
1663

1664 1665
	fd = mkstemp(tmpfile);
	if (fd < 0) {
1666 1667 1668 1669
		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));
1670 1671
	}

1672 1673
	/* Set it up */
	memset(&stream, 0, sizeof(stream));
1674
	deflateInit(&stream, zlib_compression_level);
1675
	size = 8 + deflateBound(&stream, len+hdrlen);
1676
	compressed = xmalloc(size);
1677 1678 1679 1680

	/* Compress it */
	stream.next_out = compressed;
	stream.avail_out = size;
1681 1682 1683 1684

	/* First header.. */
	stream.next_in = hdr;
	stream.avail_in = hdrlen;
1685
	setup_object_header(&stream, type, len);
1686 1687 1688 1689

	/* Then the data itself.. */
	stream.next_in = buf;
	stream.avail_in = len;
1690 1691 1692 1693 1694
	while (deflate(&stream, Z_FINISH) == Z_OK)
		/* nothing */;
	deflateEnd(&stream);
	size = stream.total_out;

L
Linus Torvalds 已提交
1695 1696
	if (write_buffer(fd, compressed, size) < 0)
		die("unable to write sha1 file");
1697
	fchmod(fd, 0444);
1698
	close(fd);
1699
	free(compressed);
1700

1701
	return move_temp_to_file(tmpfile, filename);
1702
}
1703

L
Linus Torvalds 已提交
1704 1705 1706 1707 1708
/*
 * 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
1709
{
L
Linus Torvalds 已提交
1710
	size_t size;
D
Daniel Barkalow 已提交
1711
	z_stream stream;
L
Linus Torvalds 已提交
1712 1713 1714 1715 1716 1717
	unsigned char *unpacked;
	unsigned long len;
	char type[20];
	char hdr[50];
	int hdrlen;
	void *buf;
1718

1719
	/* need to unpack and recompress it by itself */
L
Linus Torvalds 已提交
1720
	unpacked = read_packed_sha1(sha1, type, &len);
D
Daniel Barkalow 已提交
1721

L
Linus Torvalds 已提交
1722
	hdrlen = sprintf(hdr, "%s %lu", type, len) + 1;
D
Daniel Barkalow 已提交
1723

L
Linus Torvalds 已提交
1724 1725
	/* Set it up */
	memset(&stream, 0, sizeof(stream));
1726
	deflateInit(&stream, zlib_compression_level);
L
Linus Torvalds 已提交
1727 1728
	size = deflateBound(&stream, len + hdrlen);
	buf = xmalloc(size);
D
Daniel Barkalow 已提交
1729

L
Linus Torvalds 已提交
1730 1731 1732
	/* Compress it */
	stream.next_out = buf;
	stream.avail_out = size;
D
Daniel Barkalow 已提交
1733

L
Linus Torvalds 已提交
1734 1735 1736 1737 1738
	/* First header.. */
	stream.next_in = (void *)hdr;
	stream.avail_in = hdrlen;
	while (deflate(&stream, 0) == Z_OK)
		/* nothing */;
1739

L
Linus Torvalds 已提交
1740 1741 1742 1743 1744 1745 1746
	/* Then the data itself.. */
	stream.next_in = unpacked;
	stream.avail_in = len;
	while (deflate(&stream, Z_FINISH) == Z_OK)
		/* nothing */;
	deflateEnd(&stream);
	free(unpacked);
1747

L
Linus Torvalds 已提交
1748 1749 1750 1751 1752 1753 1754 1755
	*objsize = stream.total_out;
	return buf;
}

int write_sha1_to_fd(int fd, const unsigned char *sha1)
{
	int retval;
	unsigned long objsize;
1756
	void *buf = map_sha1_file(sha1, &objsize);
L
Linus Torvalds 已提交
1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767

	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
1768 1769
}

1770 1771
int write_sha1_from_fd(const unsigned char *sha1, int fd, char *buffer,
		       size_t bufsize, size_t *bufposn)
1772
{
1773
	char tmpfile[PATH_MAX];
1774 1775 1776
	int local;
	z_stream stream;
	unsigned char real_sha1[20];
1777
	unsigned char discard[4096];
1778 1779 1780
	int ret;
	SHA_CTX c;

1781
	snprintf(tmpfile, sizeof(tmpfile), "%s/obj_XXXXXX", get_object_directory());
1782

1783
	local = mkstemp(tmpfile);
1784 1785 1786 1787 1788 1789
	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));
	}
1790 1791 1792 1793 1794 1795 1796 1797 1798

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

	inflateInit(&stream);

	SHA1_Init(&c);

	do {
		ssize_t size;
1799 1800
		if (*bufposn) {
			stream.avail_in = *bufposn;
P
Pavel Roskin 已提交
1801
			stream.next_in = (unsigned char *) buffer;
1802 1803 1804 1805 1806 1807 1808
			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
1809 1810
			if (write_buffer(local, buffer, *bufposn - stream.avail_in) < 0)
				die("unable to write sha1 file");
1811 1812 1813 1814 1815 1816 1817
			memmove(buffer, buffer + *bufposn - stream.avail_in,
				stream.avail_in);
			*bufposn = stream.avail_in;
			if (ret != Z_OK)
				break;
		}
		size = read(fd, buffer + *bufposn, bufsize - *bufposn);
1818 1819
		if (size <= 0) {
			close(local);
1820
			unlink(tmpfile);
1821 1822 1823 1824 1825
			if (!size)
				return error("Connection closed?");
			perror("Reading from connection");
			return -1;
		}
1826 1827
		*bufposn += size;
	} while (1);
1828 1829 1830 1831 1832
	inflateEnd(&stream);

	close(local);
	SHA1_Final(real_sha1, &c);
	if (ret != Z_STREAM_END) {
1833
		unlink(tmpfile);
1834 1835
		return error("File %s corrupted", sha1_to_hex(sha1));
	}
1836
	if (hashcmp(sha1, real_sha1)) {
1837
		unlink(tmpfile);
1838
		return error("File %s has bad hash", sha1_to_hex(sha1));
1839
	}
1840 1841

	return move_temp_to_file(tmpfile, sha1_file_name(sha1));
1842 1843
}

1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859
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;
}

1860
int has_sha1_pack(const unsigned char *sha1, const char **ignore_packed)
1861 1862
{
	struct pack_entry e;
1863
	return find_pack_entry(sha1, &e, ignore_packed);
1864 1865
}

1866 1867 1868
int has_sha1_file(const unsigned char *sha1)
{
	struct stat st;
1869 1870
	struct pack_entry e;

1871
	if (find_pack_entry(sha1, &e, NULL))
1872
		return 1;
1873
	return find_sha1_file(sha1, &st) ? 1 : 0;
1874
}
J
Junio C Hamano 已提交
1875

1876 1877
/*
 * reads from fd as long as possible into a supplied buffer of size bytes.
1878
 * If necessary the buffer's size is increased using realloc()
1879 1880 1881 1882 1883 1884 1885
 *
 * 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)
1886
{
1887 1888 1889
	char* buf = *return_buf;
	unsigned long size = *return_size;
	int iret;
1890
	unsigned long off = 0;
1891

1892
	do {
1893
		iret = xread(fd, buf + off, size - off);
1894 1895 1896 1897
		if (iret > 0) {
			off += iret;
			if (off == size) {
				size *= 2;
J
Jonas Fonseca 已提交
1898
				buf = xrealloc(buf, size);
1899 1900 1901
			}
		}
	} while (iret > 0);
1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913

	*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
1914
	char *buf = xmalloc(size);
1915 1916 1917
	int ret;

	if (read_pipe(fd, &buf, &size)) {
1918 1919 1920
		free(buf);
		return -1;
	}
1921

1922
	if (!type)
1923
		type = blob_type;
1924
	if (write_object)
1925
		ret = write_sha1_file(buf, size, type, sha1);
R
Rene Scharfe 已提交
1926 1927
	else
		ret = hash_sha1_file(buf, size, type, sha1);
1928 1929 1930 1931
	free(buf);
	return ret;
}

1932
int index_fd(unsigned char *sha1, int fd, struct stat *st, int write_object, const char *type)
J
Junio C Hamano 已提交
1933 1934
{
	unsigned long size = st->st_size;
1935 1936
	void *buf;
	int ret;
J
Junio C Hamano 已提交
1937

1938
	buf = "";
J
Junio C Hamano 已提交
1939
	if (size)
1940
		buf = mmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
J
Junio C Hamano 已提交
1941
	close(fd);
P
Pavel Roskin 已提交
1942
	if (buf == MAP_FAILED)
J
Junio C Hamano 已提交
1943 1944
		return -1;

1945
	if (!type)
1946
		type = blob_type;
1947 1948
	if (write_object)
		ret = write_sha1_file(buf, size, type, sha1);
R
Rene Scharfe 已提交
1949 1950
	else
		ret = hash_sha1_file(buf, size, type, sha1);
1951 1952 1953
	if (size)
		munmap(buf, size);
	return ret;
J
Junio C Hamano 已提交
1954
}
1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978

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));
		if (index_fd(sha1, fd, st, write_object, NULL) < 0)
			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
1979 1980 1981
		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))
1982 1983 1984 1985 1986 1987 1988 1989 1990
			return error("%s: failed to insert into database",
				     path);
		free(target);
		break;
	default:
		return error("%s: unsupported file type", path);
	}
	return 0;
}