read-cache.c 25.7 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2 3 4 5
/*
 * GIT - The information manager from hell
 *
 * Copyright (C) Linus Torvalds, 2005
 */
6
#include "cache.h"
7 8 9 10 11 12 13 14 15 16 17 18 19
#include "cache-tree.h"

/* Index extensions.
 *
 * The first letter should be 'A'..'Z' for extensions that are not
 * necessary for a correct operation (i.e. optimization data).
 * When new extensions are added that _needs_ to be understood in
 * order to correctly interpret the index file, pick character that
 * is outside the range, to cause the reader to abort.
 */

#define CACHE_EXT(s) ( (s[0]<<24)|(s[1]<<16)|(s[2]<<8)|(s[3]) )
#define CACHE_EXT_TREE 0x54524545	/* "TREE" */
20

21
struct cache_entry **active_cache;
Junio C Hamano's avatar
Junio C Hamano committed
22
static time_t index_file_timestamp;
23
unsigned int active_nr, active_alloc, active_cache_changed;
24

25
struct cache_tree *active_cache_tree;
26

27 28
static void *cache_mmap;
static size_t cache_mmap_size;
29

30 31 32 33 34 35 36 37 38
/*
 * This only updates the "non-critical" parts of the directory
 * cache, ie the parts that aren't tracked by GIT, and only used
 * to validate the cache.
 */
void fill_stat_cache_info(struct cache_entry *ce, struct stat *st)
{
	ce->ce_ctime.sec = htonl(st->st_ctime);
	ce->ce_mtime.sec = htonl(st->st_mtime);
39
#ifdef USE_NSEC
40 41 42 43 44 45 46 47
	ce->ce_ctime.nsec = htonl(st->st_ctim.tv_nsec);
	ce->ce_mtime.nsec = htonl(st->st_mtim.tv_nsec);
#endif
	ce->ce_dev = htonl(st->st_dev);
	ce->ce_ino = htonl(st->st_ino);
	ce->ce_uid = htonl(st->st_uid);
	ce->ce_gid = htonl(st->st_gid);
	ce->ce_size = htonl(st->st_size);
Junio C Hamano's avatar
Junio C Hamano committed
48 49 50

	if (assume_unchanged)
		ce->ce_flags |= htons(CE_VALID);
51 52
}

Junio C Hamano's avatar
Junio C Hamano committed
53 54 55 56 57 58 59
static int ce_compare_data(struct cache_entry *ce, struct stat *st)
{
	int match = -1;
	int fd = open(ce->name, O_RDONLY);

	if (fd >= 0) {
		unsigned char sha1[20];
60
		if (!index_fd(sha1, fd, st, 0, OBJ_BLOB, ce->name))
61
			match = hashcmp(sha1, ce->sha1);
62
		/* index_fd() closed the file descriptor already */
Junio C Hamano's avatar
Junio C Hamano committed
63 64 65 66
	}
	return match;
}

67
static int ce_compare_link(struct cache_entry *ce, size_t expected_size)
Junio C Hamano's avatar
Junio C Hamano committed
68 69 70 71 72
{
	int match = -1;
	char *target;
	void *buffer;
	unsigned long size;
73
	enum object_type type;
Junio C Hamano's avatar
Junio C Hamano committed
74 75 76 77 78 79 80 81
	int len;

	target = xmalloc(expected_size);
	len = readlink(ce->name, target, expected_size);
	if (len != expected_size) {
		free(target);
		return -1;
	}
82
	buffer = read_sha1_file(ce->sha1, &type, &size);
Junio C Hamano's avatar
Junio C Hamano committed
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101
	if (!buffer) {
		free(target);
		return -1;
	}
	if (size == expected_size)
		match = memcmp(buffer, target, size);
	free(buffer);
	free(target);
	return match;
}

static int ce_modified_check_fs(struct cache_entry *ce, struct stat *st)
{
	switch (st->st_mode & S_IFMT) {
	case S_IFREG:
		if (ce_compare_data(ce, st))
			return DATA_CHANGED;
		break;
	case S_IFLNK:
102
		if (ce_compare_link(ce, xsize_t(st->st_size)))
Junio C Hamano's avatar
Junio C Hamano committed
103 104 105 106 107 108 109 110
			return DATA_CHANGED;
		break;
	default:
		return TYPE_CHANGED;
	}
	return 0;
}

Junio C Hamano's avatar
Junio C Hamano committed
111
static int ce_match_stat_basic(struct cache_entry *ce, struct stat *st)
112 113 114
{
	unsigned int changed = 0;

115 116 117
	switch (ntohl(ce->ce_mode) & S_IFMT) {
	case S_IFREG:
		changed |= !S_ISREG(st->st_mode) ? TYPE_CHANGED : 0;
Junio C Hamano's avatar
Junio C Hamano committed
118 119 120 121 122
		/* We consider only the owner x bit to be relevant for
		 * "mode changes"
		 */
		if (trust_executable_bit &&
		    (0100 & (ntohl(ce->ce_mode) ^ st->st_mode)))
123
			changed |= MODE_CHANGED;
124 125
		break;
	case S_IFLNK:
126 127 128
		if (!S_ISLNK(st->st_mode) &&
		    (has_symlinks || !S_ISREG(st->st_mode)))
			changed |= TYPE_CHANGED;
129 130 131 132
		break;
	default:
		die("internal error: ce_mode is %o", ntohl(ce->ce_mode));
	}
133
	if (ce->ce_mtime.sec != htonl(st->st_mtime))
134
		changed |= MTIME_CHANGED;
135 136 137
	if (ce->ce_ctime.sec != htonl(st->st_ctime))
		changed |= CTIME_CHANGED;

138
#ifdef USE_NSEC
139 140 141 142 143
	/*
	 * nsec seems unreliable - not all filesystems support it, so
	 * as long as it is in the inode cache you get right nsec
	 * but after it gets flushed, you get zero nsec.
	 */
144
	if (ce->ce_mtime.nsec != htonl(st->st_mtim.tv_nsec))
145
		changed |= MTIME_CHANGED;
146
	if (ce->ce_ctime.nsec != htonl(st->st_ctim.tv_nsec))
147
		changed |= CTIME_CHANGED;
148 149 150 151
#endif	

	if (ce->ce_uid != htonl(st->st_uid) ||
	    ce->ce_gid != htonl(st->st_gid))
152
		changed |= OWNER_CHANGED;
153
	if (ce->ce_ino != htonl(st->st_ino))
154
		changed |= INODE_CHANGED;
155 156 157 158 159 160 161 162 163 164 165

#ifdef USE_STDEV
	/*
	 * st_dev breaks on network filesystems where different
	 * clients will have different views of what "device"
	 * the filesystem is on
	 */
	if (ce->ce_dev != htonl(st->st_dev))
		changed |= INODE_CHANGED;
#endif

166
	if (ce->ce_size != htonl(st->st_size))
167
		changed |= DATA_CHANGED;
168

Junio C Hamano's avatar
Junio C Hamano committed
169 170 171
	return changed;
}

172
int ce_match_stat(struct cache_entry *ce, struct stat *st, int options)
Junio C Hamano's avatar
Junio C Hamano committed
173
{
Junio C Hamano's avatar
Junio C Hamano committed
174
	unsigned int changed;
175 176
	int ignore_valid = options & 01;
	int assume_racy_is_modified = options & 02;
Junio C Hamano's avatar
Junio C Hamano committed
177 178 179 180 181 182 183 184 185

	/*
	 * If it's marked as always valid in the index, it's
	 * valid whatever the checked-out copy says.
	 */
	if (!ignore_valid && (ce->ce_flags & htons(CE_VALID)))
		return 0;

	changed = ce_match_stat_basic(ce, st);
Junio C Hamano's avatar
Junio C Hamano committed
186

Junio C Hamano's avatar
Junio C Hamano committed
187 188 189 190 191 192 193 194 195 196 197 198 199 200
	/*
	 * Within 1 second of this sequence:
	 * 	echo xyzzy >file && git-update-index --add file
	 * running this command:
	 * 	echo frotz >file
	 * would give a falsely clean cache entry.  The mtime and
	 * length match the cache, and other stat fields do not change.
	 *
	 * We could detect this at update-index time (the cache entry
	 * being registered/updated records the same time as "now")
	 * and delay the return from git-update-index, but that would
	 * effectively mean we can make at most one commit per second,
	 * which is not acceptable.  Instead, we check cache entries
	 * whose mtime are the same as the index file timestamp more
Junio C Hamano's avatar
Junio C Hamano committed
201
	 * carefully than others.
Junio C Hamano's avatar
Junio C Hamano committed
202 203 204
	 */
	if (!changed &&
	    index_file_timestamp &&
205 206 207 208 209 210
	    index_file_timestamp <= ntohl(ce->ce_mtime.sec)) {
		if (assume_racy_is_modified)
			changed |= DATA_CHANGED;
		else
			changed |= ce_modified_check_fs(ce, st);
	}
211

Junio C Hamano's avatar
Junio C Hamano committed
212
	return changed;
213 214
}

Junio C Hamano's avatar
Junio C Hamano committed
215
int ce_modified(struct cache_entry *ce, struct stat *st, int really)
216
{
Junio C Hamano's avatar
Junio C Hamano committed
217
	int changed, changed_fs;
Junio C Hamano's avatar
Junio C Hamano committed
218
	changed = ce_match_stat(ce, st, really);
219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234
	if (!changed)
		return 0;
	/*
	 * If the mode or type has changed, there's no point in trying
	 * to refresh the entry - it's not going to match
	 */
	if (changed & (MODE_CHANGED | TYPE_CHANGED))
		return changed;

	/* Immediately after read-tree or update-index --cacheinfo,
	 * the length field is zero.  For other cases the ce_size
	 * should match the SHA1 recorded in the index entry.
	 */
	if ((changed & DATA_CHANGED) && ce->ce_size != htonl(0))
		return changed;

Junio C Hamano's avatar
Junio C Hamano committed
235 236 237
	changed_fs = ce_modified_check_fs(ce, st);
	if (changed_fs)
		return changed | changed_fs;
238 239 240
	return 0;
}

241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259
int base_name_compare(const char *name1, int len1, int mode1,
		      const char *name2, int len2, int mode2)
{
	unsigned char c1, c2;
	int len = len1 < len2 ? len1 : len2;
	int cmp;

	cmp = memcmp(name1, name2, len);
	if (cmp)
		return cmp;
	c1 = name1[len];
	c2 = name2[len];
	if (!c1 && S_ISDIR(mode1))
		c1 = '/';
	if (!c2 && S_ISDIR(mode2))
		c2 = '/';
	return (c1 < c2) ? -1 : (c1 > c2) ? 1 : 0;
}

260
int cache_name_compare(const char *name1, int flags1, const char *name2, int flags2)
261
{
262 263
	int len1 = flags1 & CE_NAMEMASK;
	int len2 = flags2 & CE_NAMEMASK;
264 265 266 267 268 269 270 271 272 273
	int len = len1 < len2 ? len1 : len2;
	int cmp;

	cmp = memcmp(name1, name2, len);
	if (cmp)
		return cmp;
	if (len1 < len2)
		return -1;
	if (len1 > len2)
		return 1;
Junio C Hamano's avatar
Junio C Hamano committed
274

275 276 277
	/* Compare stages  */
	flags1 &= CE_STAGEMASK;
	flags2 &= CE_STAGEMASK;
Junio C Hamano's avatar
Junio C Hamano committed
278

279 280 281 282
	if (flags1 < flags2)
		return -1;
	if (flags1 > flags2)
		return 1;
283 284 285 286 287 288 289 290 291 292 293 294
	return 0;
}

int cache_name_pos(const char *name, int namelen)
{
	int first, last;

	first = 0;
	last = active_nr;
	while (last > first) {
		int next = (last + first) >> 1;
		struct cache_entry *ce = active_cache[next];
295
		int cmp = cache_name_compare(name, namelen, ce->name, ntohs(ce->ce_flags));
296
		if (!cmp)
297
			return next;
298 299 300 301 302 303
		if (cmp < 0) {
			last = next;
			continue;
		}
		first = next+1;
	}
304
	return -first-1;
305 306
}

307
/* Remove entry, return true if there are more entries to go.. */
308
int remove_cache_entry_at(int pos)
309
{
310
	active_cache_changed = 1;
311 312 313 314 315 316 317
	active_nr--;
	if (pos >= active_nr)
		return 0;
	memmove(active_cache + pos, active_cache + pos + 1, (active_nr - pos) * sizeof(struct cache_entry *));
	return 1;
}

Junio C Hamano's avatar
Junio C Hamano committed
318
int remove_file_from_cache(const char *path)
319 320
{
	int pos = cache_name_pos(path, strlen(path));
321 322 323
	if (pos < 0)
		pos = -pos-1;
	while (pos < active_nr && !strcmp(active_cache[pos]->name, path))
324
		remove_cache_entry_at(pos);
325 326 327
	return 0;
}

328
int add_file_to_cache(const char *path, int verbose)
Johannes Schindelin's avatar
Johannes Schindelin committed
329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346
{
	int size, namelen;
	struct stat st;
	struct cache_entry *ce;

	if (lstat(path, &st))
		die("%s: unable to stat (%s)", path, strerror(errno));

	if (!S_ISREG(st.st_mode) && !S_ISLNK(st.st_mode))
		die("%s: can only add regular files or symbolic links", path);

	namelen = strlen(path);
	size = cache_entry_size(namelen);
	ce = xcalloc(1, size);
	memcpy(ce->name, path, namelen);
	ce->ce_flags = htons(namelen);
	fill_stat_cache_info(ce, &st);

347
	if (trust_executable_bit && has_symlinks)
348 349
		ce->ce_mode = create_ce_mode(st.st_mode);
	else {
350 351
		/* If there is an existing entry, pick the mode bits and type
		 * from it, otherwise assume unexecutable regular file.
Johannes Schindelin's avatar
Johannes Schindelin committed
352
		 */
353
		struct cache_entry *ent;
Johannes Schindelin's avatar
Johannes Schindelin committed
354
		int pos = cache_name_pos(path, namelen);
355 356 357

		ent = (0 <= pos) ? active_cache[pos] : NULL;
		ce->ce_mode = ce_mode_from_stat(ent, st.st_mode);
Johannes Schindelin's avatar
Johannes Schindelin committed
358 359 360 361
	}

	if (index_path(ce->sha1, path, &st, 1))
		die("unable to index file %s", path);
362
	if (add_cache_entry(ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE))
Johannes Schindelin's avatar
Johannes Schindelin committed
363 364 365 366 367 368 369
		die("unable to add %s to index",path);
	if (verbose)
		printf("add '%s'\n", path);
	cache_tree_invalidate_path(active_cache_tree, path);
	return 0;
}

370
int ce_same_name(struct cache_entry *a, struct cache_entry *b)
371 372 373 374 375
{
	int len = ce_namelen(a);
	return ce_namelen(b) == len && !memcmp(a->name, b->name, len);
}

376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395
int ce_path_match(const struct cache_entry *ce, const char **pathspec)
{
	const char *match, *name;
	int len;

	if (!pathspec)
		return 1;

	len = ce_namelen(ce);
	name = ce->name;
	while ((match = *pathspec++) != NULL) {
		int matchlen = strlen(match);
		if (matchlen > len)
			continue;
		if (memcmp(name, match, matchlen))
			continue;
		if (matchlen && name[matchlen-1] == '/')
			return 1;
		if (name[matchlen] == '/' || !name[matchlen])
			return 1;
396 397
		if (!matchlen)
			return 1;
398 399 400 401
	}
	return 0;
}

402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465
/*
 * We fundamentally don't like some paths: we don't want
 * dot or dot-dot anywhere, and for obvious reasons don't
 * want to recurse into ".git" either.
 *
 * Also, we don't want double slashes or slashes at the
 * end that can make pathnames ambiguous.
 */
static int verify_dotfile(const char *rest)
{
	/*
	 * The first character was '.', but that
	 * has already been discarded, we now test
	 * the rest.
	 */
	switch (*rest) {
	/* "." is not allowed */
	case '\0': case '/':
		return 0;

	/*
	 * ".git" followed by  NUL or slash is bad. This
	 * shares the path end test with the ".." case.
	 */
	case 'g':
		if (rest[1] != 'i')
			break;
		if (rest[2] != 't')
			break;
		rest += 2;
	/* fallthrough */
	case '.':
		if (rest[1] == '\0' || rest[1] == '/')
			return 0;
	}
	return 1;
}

int verify_path(const char *path)
{
	char c;

	goto inside;
	for (;;) {
		if (!c)
			return 1;
		if (c == '/') {
inside:
			c = *path++;
			switch (c) {
			default:
				continue;
			case '/': case '\0':
				break;
			case '.':
				if (verify_dotfile(path))
					continue;
			}
			return 0;
		}
		c = *path++;
	}
}

466 467 468
/*
 * Do we have another file that has the beginning components being a
 * proper superset of the name we're trying to add?
469
 */
470
static int has_file_name(const struct cache_entry *ce, int pos, int ok_to_replace)
471
{
472 473
	int retval = 0;
	int len = ce_namelen(ce);
474
	int stage = ce_stage(ce);
475
	const char *name = ce->name;
476

477 478
	while (pos < active_nr) {
		struct cache_entry *p = active_cache[pos++];
479

480
		if (len >= ce_namelen(p))
481
			break;
482 483
		if (memcmp(name, p->name, len))
			break;
484 485
		if (ce_stage(p) != stage)
			continue;
486 487
		if (p->name[len] != '/')
			continue;
488 489
		if (!ce_stage(p) && !p->ce_mode)
			continue;
490 491 492 493
		retval = -1;
		if (!ok_to_replace)
			break;
		remove_cache_entry_at(--pos);
494
	}
495 496
	return retval;
}
497

498 499 500 501 502 503 504
/*
 * Do we have another file with a pathname that is a proper
 * subset of the name we're trying to add?
 */
static int has_dir_name(const struct cache_entry *ce, int pos, int ok_to_replace)
{
	int retval = 0;
505
	int stage = ce_stage(ce);
506 507
	const char *name = ce->name;
	const char *slash = name + ce_namelen(ce);
508

509 510
	for (;;) {
		int len;
511

512 513 514 515 516 517 518
		for (;;) {
			if (*--slash == '/')
				break;
			if (slash <= ce->name)
				return retval;
		}
		len = slash - name;
519

520
		pos = cache_name_pos(name, ntohs(create_ce_flags(len, stage)));
521
		if (pos >= 0) {
522 523 524 525 526 527 528 529 530 531 532 533 534 535 536
			/*
			 * Found one, but not so fast.  This could
			 * be a marker that says "I was here, but
			 * I am being removed".  Such an entry is
			 * not a part of the resulting tree, and
			 * it is Ok to have a directory at the same
			 * path.
			 */
			if (stage || active_cache[pos]->ce_mode) {
				retval = -1;
				if (!ok_to_replace)
					break;
				remove_cache_entry_at(pos);
				continue;
			}
537
		}
538 539
		else
			pos = -pos-1;
540 541 542 543

		/*
		 * Trivial optimization: if we find an entry that
		 * already matches the sub-directory, then we know
544
		 * we're ok, and we can exit.
545
		 */
546
		while (pos < active_nr) {
547
			struct cache_entry *p = active_cache[pos];
548 549 550 551
			if ((ce_namelen(p) <= len) ||
			    (p->name[len] != '/') ||
			    memcmp(p->name, name, len))
				break; /* not our subdirectory */
552
			if (ce_stage(p) == stage && (stage || p->ce_mode))
553 554 555 556 557 558 559
				/* p is at the same stage as our entry, and
				 * is a subdirectory of what we are looking
				 * at, so we cannot have conflicts at our
				 * level or anything shorter.
				 */
				return retval;
			pos++;
560
		}
561
	}
562 563 564 565 566 567 568 569 570 571 572 573 574 575
	return retval;
}

/* We may be in a situation where we already have path/file and path
 * is being added, or we already have path and path/file is being
 * added.  Either one would result in a nonsense tree that has path
 * twice when git-write-tree tries to write it out.  Prevent it.
 * 
 * If ok-to-replace is specified, we remove the conflicting entries
 * from the cache so the caller should recompute the insert position.
 * When this happens, we return non-zero.
 */
static int check_file_directory_conflict(const struct cache_entry *ce, int pos, int ok_to_replace)
{
576 577 578 579 580 581 582 583
	int retval;

	/*
	 * When ce is an "I am going away" entry, we allow it to be added
	 */
	if (!ce_stage(ce) && !ce->ce_mode)
		return 0;

584 585 586
	/*
	 * We check if the path is a sub-path of a subsequent pathname
	 * first, since removing those will not change the position
587
	 * in the array.
588
	 */
589 590
	retval = has_file_name(ce, pos, ok_to_replace);

591 592 593 594 595
	/*
	 * Then check if the path might have a clashing sub-directory
	 * before it.
	 */
	return retval + has_dir_name(ce, pos, ok_to_replace);
596 597
}

598
int add_cache_entry(struct cache_entry *ce, int option)
599 600
{
	int pos;
601 602
	int ok_to_add = option & ADD_CACHE_OK_TO_ADD;
	int ok_to_replace = option & ADD_CACHE_OK_TO_REPLACE;
603
	int skip_df_check = option & ADD_CACHE_SKIP_DFCHECK;
Junio C Hamano's avatar
Junio C Hamano committed
604

605
	pos = cache_name_pos(ce->name, ntohs(ce->ce_flags));
606

Junio C Hamano's avatar
Junio C Hamano committed
607
	/* existing match? Just replace it. */
608
	if (pos >= 0) {
609
		active_cache_changed = 1;
610
		active_cache[pos] = ce;
611 612
		return 0;
	}
613
	pos = -pos-1;
614

615 616 617 618 619
	/*
	 * Inserting a merged entry ("stage 0") into the index
	 * will always replace all non-merged entries..
	 */
	if (pos < active_nr && ce_stage(ce) == 0) {
620
		while (ce_same_name(active_cache[pos], ce)) {
621
			ok_to_add = 1;
622
			if (!remove_cache_entry_at(pos))
623 624 625 626
				break;
		}
	}

627 628
	if (!ok_to_add)
		return -1;
629 630
	if (!verify_path(ce->name))
		return -1;
631

Junio C Hamano's avatar
Junio C Hamano committed
632 633
	if (!skip_df_check &&
	    check_file_directory_conflict(ce, pos, ok_to_replace)) {
634
		if (!ok_to_replace)
635
			return error("'%s' appears as both a file and as a directory", ce->name);
636
		pos = cache_name_pos(ce->name, ntohs(ce->ce_flags));
637 638
		pos = -pos-1;
	}
639

640 641 642
	/* Make sure the array is big enough .. */
	if (active_nr == active_alloc) {
		active_alloc = alloc_nr(active_alloc);
643
		active_cache = xrealloc(active_cache, active_alloc * sizeof(struct cache_entry *));
644 645 646 647 648 649 650
	}

	/* Add it in.. */
	active_nr++;
	if (active_nr > pos)
		memmove(active_cache + pos + 1, active_cache + pos, (active_nr - pos - 1) * sizeof(ce));
	active_cache[pos] = ce;
651
	active_cache_changed = 1;
652 653 654
	return 0;
}

655 656 657 658 659 660 661 662 663 664 665
/*
 * "refresh" does not calculate a new sha1 file or bring the
 * cache up-to-date for mode/content changes. But what it
 * _does_ do is to "re-match" the stat information of a file
 * with the cache, so that you can refresh the cache for a
 * file that hasn't been changed but where the stat entry is
 * out of date.
 *
 * For example, you'd want to do this after doing a "git-read-tree",
 * to link up the stat cache details with the proper files.
 */
666
static struct cache_entry *refresh_cache_ent(struct cache_entry *ce, int really, int *err)
667 668 669 670 671
{
	struct stat st;
	struct cache_entry *updated;
	int changed, size;

672
	if (lstat(ce->name, &st) < 0) {
673 674
		if (err)
			*err = errno;
675 676
		return NULL;
	}
677 678 679 680 681 682 683

	changed = ce_match_stat(ce, &st, really);
	if (!changed) {
		if (really && assume_unchanged &&
		    !(ce->ce_flags & htons(CE_VALID)))
			; /* mark this one VALID again */
		else
684
			return ce;
685 686
	}

687
	if (ce_modified(ce, &st, really)) {
688 689
		if (err)
			*err = EINVAL;
690 691
		return NULL;
	}
692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720

	size = ce_size(ce);
	updated = xmalloc(size);
	memcpy(updated, ce, size);
	fill_stat_cache_info(updated, &st);

	/* In this case, if really is not set, we should leave
	 * CE_VALID bit alone.  Otherwise, paths marked with
	 * --no-assume-unchanged (i.e. things to be edited) will
	 * reacquire CE_VALID bit automatically, which is not
	 * really what we want.
	 */
	if (!really && assume_unchanged && !(ce->ce_flags & htons(CE_VALID)))
		updated->ce_flags &= ~htons(CE_VALID);

	return updated;
}

int refresh_cache(unsigned int flags)
{
	int i;
	int has_errors = 0;
	int really = (flags & REFRESH_REALLY) != 0;
	int allow_unmerged = (flags & REFRESH_UNMERGED) != 0;
	int quiet = (flags & REFRESH_QUIET) != 0;
	int not_new = (flags & REFRESH_IGNORE_MISSING) != 0;

	for (i = 0; i < active_nr; i++) {
		struct cache_entry *ce, *new;
721 722
		int cache_errno = 0;

723 724 725 726 727 728 729 730 731 732 733 734 735
		ce = active_cache[i];
		if (ce_stage(ce)) {
			while ((i < active_nr) &&
			       ! strcmp(active_cache[i]->name, ce->name))
				i++;
			i--;
			if (allow_unmerged)
				continue;
			printf("%s: needs merge\n", ce->name);
			has_errors = 1;
			continue;
		}

736
		new = refresh_cache_ent(ce, really, &cache_errno);
737
		if (new == ce)
738
			continue;
739 740
		if (!new) {
			if (not_new && cache_errno == ENOENT)
741
				continue;
742
			if (really && cache_errno == EINVAL) {
743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763
				/* If we are doing --really-refresh that
				 * means the index is not valid anymore.
				 */
				ce->ce_flags &= ~htons(CE_VALID);
				active_cache_changed = 1;
			}
			if (quiet)
				continue;
			printf("%s: needs update\n", ce->name);
			has_errors = 1;
			continue;
		}
		active_cache_changed = 1;
		/* You can NOT just free active_cache[i] here, since it
		 * might not be necessarily malloc()ed but can also come
		 * from mmap(). */
		active_cache[i] = new;
	}
	return has_errors;
}

764 765 766 767 768
struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really)
{
	return refresh_cache_ent(ce, really, NULL);
}

769
static int verify_hdr(struct cache_header *hdr, unsigned long size)
770 771
{
	SHA_CTX c;
772
	unsigned char sha1[20];
773

774
	if (hdr->hdr_signature != htonl(CACHE_SIGNATURE))
775
		return error("bad signature");
776 777
	if (hdr->hdr_version != htonl(2))
		return error("bad index version");
778
	SHA1_Init(&c);
779
	SHA1_Update(&c, hdr, size - 20);
780
	SHA1_Final(sha1, &c);
781
	if (hashcmp(sha1, (unsigned char *)hdr + size - 20))
782
		return error("bad index file sha1 signature");
783 784 785
	return 0;
}

786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802
static int read_index_extension(const char *ext, void *data, unsigned long sz)
{
	switch (CACHE_EXT(ext)) {
	case CACHE_EXT_TREE:
		active_cache_tree = cache_tree_read(data, sz);
		break;
	default:
		if (*ext < 'A' || 'Z' < *ext)
			return error("index uses %.4s extension, which we do not understand",
				     ext);
		fprintf(stderr, "ignoring %.4s extension\n", ext);
		break;
	}
	return 0;
}

int read_cache(void)
803 804 805 806 807 808
{
	return read_cache_from(get_index_file());
}

/* remember to discard_cache() before reading a different cache! */
int read_cache_from(const char *path)
809 810 811
{
	int fd, i;
	struct stat st;
812
	unsigned long offset;
813 814 815
	struct cache_header *hdr;

	errno = EBUSY;
816
	if (cache_mmap)
817 818
		return active_nr;

819
	errno = ENOENT;
Junio C Hamano's avatar
Junio C Hamano committed
820
	index_file_timestamp = 0;
821
	fd = open(path, O_RDONLY);
822 823 824 825 826
	if (fd < 0) {
		if (errno == ENOENT)
			return 0;
		die("index file open failed (%s)", strerror(errno));
	}
827 828

	if (!fstat(fd, &st)) {
829
		cache_mmap_size = xsize_t(st.st_size);
830
		errno = EINVAL;
831
		if (cache_mmap_size >= sizeof(struct cache_header) + 20)
832
			cache_mmap = xmmap(NULL, cache_mmap_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
833 834 835 836
		else
			die("index file smaller than expected");
	} else
		die("cannot stat the open index (%s)", strerror(errno));
837 838
	close(fd);

839 840
	hdr = cache_mmap;
	if (verify_hdr(hdr, cache_mmap_size) < 0)
841 842
		goto unmap;

843
	active_nr = ntohl(hdr->hdr_entries);
844
	active_alloc = alloc_nr(active_nr);
845
	active_cache = xcalloc(active_alloc, sizeof(struct cache_entry *));
846 847

	offset = sizeof(*hdr);
848
	for (i = 0; i < active_nr; i++) {
849
		struct cache_entry *ce = (struct cache_entry *) ((char *) cache_mmap + offset);
850 851 852
		offset = offset + ce_size(ce);
		active_cache[i] = ce;
	}
Junio C Hamano's avatar
Junio C Hamano committed
853
	index_file_timestamp = st.st_mtime;
854
	while (offset <= cache_mmap_size - 20 - 8) {
855 856 857 858 859 860 861
		/* After an array of active_nr index entries,
		 * there can be arbitrary number of extended
		 * sections, each of which is prefixed with
		 * extension name (4-byte) and section length
		 * in 4-byte network byte order.
		 */
		unsigned long extsize;
862
		memcpy(&extsize, (char *) cache_mmap + offset + 4, 4);
863
		extsize = ntohl(extsize);
864 865
		if (read_index_extension(((const char *) cache_mmap) + offset,
					 (char *) cache_mmap + offset + 8,
866
					 extsize) < 0)
867 868 869 870
			goto unmap;
		offset += 8;
		offset += extsize;
	}
871 872 873
	return active_nr;

unmap:
874
	munmap(cache_mmap, cache_mmap_size);
875
	errno = EINVAL;
876
	die("index file corrupt");
877 878
}

879
int discard_cache(void)
880 881 882
{
	int ret;

883 884 885
	active_nr = active_cache_changed = 0;
	index_file_timestamp = 0;
	cache_tree_free(&active_cache_tree);
886 887 888 889 890 891 892 893 894 895
	if (cache_mmap == NULL)
		return 0;
	ret = munmap(cache_mmap, cache_mmap_size);
	cache_mmap = NULL;
	cache_mmap_size = 0;

	/* no need to throw away allocated active_cache */
	return ret;
}

896
#define WRITE_BUFFER_SIZE 8192
897
static unsigned char write_buffer[WRITE_BUFFER_SIZE];
898 899
static unsigned long write_buffer_len;

900 901 902 903 904
static int ce_write_flush(SHA_CTX *context, int fd)
{
	unsigned int buffered = write_buffer_len;
	if (buffered) {
		SHA1_Update(context, write_buffer, buffered);
905
		if (write_in_full(fd, write_buffer, buffered) != buffered)
906 907 908 909 910 911
			return -1;
		write_buffer_len = 0;
	}
	return 0;
}

912
static int ce_write(SHA_CTX *context, int fd, void *data, unsigned int len)
913 914 915 916 917 918 919 920 921
{
	while (len) {
		unsigned int buffered = write_buffer_len;
		unsigned int partial = WRITE_BUFFER_SIZE - buffered;
		if (partial > len)
			partial = len;
		memcpy(write_buffer + buffered, data, partial);
		buffered += partial;
		if (buffered == WRITE_BUFFER_SIZE) {
922 923
			write_buffer_len = buffered;
			if (ce_write_flush(context, fd))
924 925 926 927 928
				return -1;
			buffered = 0;
		}
		write_buffer_len = buffered;
		len -= partial;
929
		data = (char *) data + partial;
930 931 932 933
 	}
 	return 0;
}

934
static int write_index_ext_header(SHA_CTX *context, int fd,
935
				  unsigned int ext, unsigned int sz)
936 937 938
{
	ext = htonl(ext);
	sz = htonl(sz);
David Rientjes's avatar
David Rientjes committed
939 940
	return ((ce_write(context, fd, &ext, 4) < 0) ||
		(ce_write(context, fd, &sz, 4) < 0)) ? -1 : 0;
941 942 943
}

static int ce_flush(SHA_CTX *context, int fd)
944 945
{
	unsigned int left = write_buffer_len;
946

947 948
	if (left) {
		write_buffer_len = 0;
949
		SHA1_Update(context, write_buffer, left);
950
	}
951

952 953
	/* Flush first if not enough space for SHA1 signature */
	if (left + 20 > WRITE_BUFFER_SIZE) {
954
		if (write_in_full(fd, write_buffer, left) != left)
955 956 957 958
			return -1;
		left = 0;
	}

959
	/* Append the SHA1 signature at the end */
960
	SHA1_Final(write_buffer + left, context);
961
	left += 20;
962
	return (write_in_full(fd, write_buffer, left) != left) ? -1 : 0;
963 964
}

Junio C Hamano's avatar
Junio C Hamano committed
965 966 967 968 969 970 971 972 973 974 975 976 977 978 979
static void ce_smudge_racily_clean_entry(struct cache_entry *ce)
{
	/*
	 * The only thing we care about in this function is to smudge the
	 * falsely clean entry due to touch-update-touch race, so we leave
	 * everything else as they are.  We are called for entries whose
	 * ce_mtime match the index file mtime.
	 */
	struct stat st;

	if (lstat(ce->name, &st) < 0)
		return;
	if (ce_match_stat_basic(ce, &st))
		return;
	if (ce_modified_check_fs(ce, &st)) {
980 981 982 983 984 985 986 987 988 989 990
		/* This is "racily clean"; smudge it.  Note that this
		 * is a tricky code.  At first glance, it may appear
		 * that it can break with this sequence:
		 *
		 * $ echo xyzzy >frotz
		 * $ git-update-index --add frotz
		 * $ : >frotz
		 * $ sleep 3
		 * $ echo filfre >nitfol
		 * $ git-update-index --add nitfol
		 *
991
		 * but it does not.  When the second update-index runs,
992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004
		 * it notices that the entry "frotz" has the same timestamp
		 * as index, and if we were to smudge it by resetting its
		 * size to zero here, then the object name recorded
		 * in index is the 6-byte file but the cached stat information
		 * becomes zero --- which would then match what we would
		 * obtain from the filesystem next time we stat("frotz"). 
		 *
		 * However, the second update-index, before calling
		 * this function, notices that the cached size is 6
		 * bytes and what is on the filesystem is an empty
		 * file, and never calls us, so the cached size information
		 * for "frotz" stays 6 which does not match the filesystem.
		 */
Junio C Hamano's avatar
Junio C Hamano committed
1005 1006 1007 1008
		ce->ce_size = htonl(0);
	}
}

1009
int write_cache(int newfd, struct cache_entry **cache, int entries)
1010 1011 1012
{
	SHA_CTX c;
	struct cache_header hdr;
1013
	int i, removed;
1014 1015 1016 1017

	for (i = removed = 0; i < entries; i++)
		if (!cache[i]->ce_mode)
			removed++;
1018

1019
	hdr.hdr_signature = htonl(CACHE_SIGNATURE);
1020
	hdr.hdr_version = htonl(2);
1021
	hdr.hdr_entries = htonl(entries - removed);
1022 1023

	SHA1_Init(&c);
1024
	if (ce_write(&c, newfd, &hdr, sizeof(hdr)) < 0)
1025 1026 1027 1028
		return -1;

	for (i = 0; i < entries; i++) {
		struct cache_entry *ce = cache[i];
1029 1030
		if (!ce->ce_mode)
			continue;
1031 1032
		if (index_file_timestamp &&
		    index_file_timestamp <= ntohl(ce->ce_mtime.sec))
Junio C Hamano's avatar
Junio C Hamano committed
1033
			ce_smudge_racily_clean_entry(ce);
1034
		if (ce_write(&c, newfd, ce, ce_size(ce)) < 0)
1035 1036
			return -1;
	}
1037

1038 1039 1040 1041 1042 1043 1044
	/* Write extension data here */
	if (active_cache_tree) {
		unsigned long sz;
		void *data = cache_tree_write(active_cache_tree, &sz);
		if (data &&
		    !write_index_ext_header(&c, newfd, CACHE_EXT_TREE, sz) &&
		    !ce_write(&c, newfd, data, sz))
1045
			free(data);
1046 1047 1048 1049 1050 1051
		else {
			free(data);
			return -1;
		}
	}
	return ce_flush(&c, newfd);
1052
}