read-cache.c 25 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
int cache_errno;
28

29 30
static void *cache_mmap;
static size_t cache_mmap_size;
31

32 33 34 35 36 37 38 39 40
/*
 * 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);
41
#ifdef USE_NSEC
42 43 44 45 46 47 48 49
	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
50 51 52

	if (assume_unchanged)
		ce->ce_flags |= htons(CE_VALID);
53 54
}

Junio C Hamano's avatar
Junio C Hamano committed
55 56 57 58 59 60 61
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];
62
		if (!index_fd(sha1, fd, st, 0, OBJ_BLOB, ce->name))
63
			match = hashcmp(sha1, ce->sha1);
64
		/* index_fd() closed the file descriptor already */
Junio C Hamano's avatar
Junio C Hamano committed
65 66 67 68
	}
	return match;
}

69
static int ce_compare_link(struct cache_entry *ce, size_t expected_size)
Junio C Hamano's avatar
Junio C Hamano committed
70 71 72 73 74
{
	int match = -1;
	char *target;
	void *buffer;
	unsigned long size;
75
	enum object_type type;
Junio C Hamano's avatar
Junio C Hamano committed
76 77 78 79 80 81 82 83
	int len;

	target = xmalloc(expected_size);
	len = readlink(ce->name, target, expected_size);
	if (len != expected_size) {
		free(target);
		return -1;
	}
84
	buffer = read_sha1_file(ce->sha1, &type, &size);
Junio C Hamano's avatar
Junio C Hamano committed
85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103
	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:
104
		if (ce_compare_link(ce, xsize_t(st->st_size)))
Junio C Hamano's avatar
Junio C Hamano committed
105 106 107 108 109 110 111 112
			return DATA_CHANGED;
		break;
	default:
		return TYPE_CHANGED;
	}
	return 0;
}

Junio C Hamano's avatar
Junio C Hamano committed
113
static int ce_match_stat_basic(struct cache_entry *ce, struct stat *st)
114 115 116
{
	unsigned int changed = 0;

117 118 119
	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
120 121 122 123 124
		/* 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)))
125
			changed |= MODE_CHANGED;
126 127
		break;
	case S_IFLNK:
128 129 130
		if (!S_ISLNK(st->st_mode) &&
		    (has_symlinks || !S_ISREG(st->st_mode)))
			changed |= TYPE_CHANGED;
131 132 133 134
		break;
	default:
		die("internal error: ce_mode is %o", ntohl(ce->ce_mode));
	}
135
	if (ce->ce_mtime.sec != htonl(st->st_mtime))
136
		changed |= MTIME_CHANGED;
137 138 139
	if (ce->ce_ctime.sec != htonl(st->st_ctime))
		changed |= CTIME_CHANGED;

140
#ifdef USE_NSEC
141 142 143 144 145
	/*
	 * 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.
	 */
146
	if (ce->ce_mtime.nsec != htonl(st->st_mtim.tv_nsec))
147
		changed |= MTIME_CHANGED;
148
	if (ce->ce_ctime.nsec != htonl(st->st_ctim.tv_nsec))
149
		changed |= CTIME_CHANGED;
150 151 152 153
#endif	

	if (ce->ce_uid != htonl(st->st_uid) ||
	    ce->ce_gid != htonl(st->st_gid))
154
		changed |= OWNER_CHANGED;
155
	if (ce->ce_ino != htonl(st->st_ino))
156
		changed |= INODE_CHANGED;
157 158 159 160 161 162 163 164 165 166 167

#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

168
	if (ce->ce_size != htonl(st->st_size))
169
		changed |= DATA_CHANGED;
170

Junio C Hamano's avatar
Junio C Hamano committed
171 172 173
	return changed;
}

174
int ce_match_stat(struct cache_entry *ce, struct stat *st, int options)
Junio C Hamano's avatar
Junio C Hamano committed
175
{
Junio C Hamano's avatar
Junio C Hamano committed
176
	unsigned int changed;
177 178
	int ignore_valid = options & 01;
	int assume_racy_is_modified = options & 02;
Junio C Hamano's avatar
Junio C Hamano committed
179 180 181 182 183 184 185 186 187

	/*
	 * 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
188

Junio C Hamano's avatar
Junio C Hamano committed
189 190 191 192 193 194 195 196 197 198 199 200 201 202
	/*
	 * 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
203
	 * carefully than others.
Junio C Hamano's avatar
Junio C Hamano committed
204 205 206
	 */
	if (!changed &&
	    index_file_timestamp &&
207 208 209 210 211 212
	    index_file_timestamp <= ntohl(ce->ce_mtime.sec)) {
		if (assume_racy_is_modified)
			changed |= DATA_CHANGED;
		else
			changed |= ce_modified_check_fs(ce, st);
	}
213

Junio C Hamano's avatar
Junio C Hamano committed
214
	return changed;
215 216
}

Junio C Hamano's avatar
Junio C Hamano committed
217
int ce_modified(struct cache_entry *ce, struct stat *st, int really)
218
{
Junio C Hamano's avatar
Junio C Hamano committed
219
	int changed, changed_fs;
Junio C Hamano's avatar
Junio C Hamano committed
220
	changed = ce_match_stat(ce, st, really);
221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236
	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
237 238 239
	changed_fs = ce_modified_check_fs(ce, st);
	if (changed_fs)
		return changed | changed_fs;
240 241 242
	return 0;
}

243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261
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;
}

262
int cache_name_compare(const char *name1, int flags1, const char *name2, int flags2)
263
{
264 265
	int len1 = flags1 & CE_NAMEMASK;
	int len2 = flags2 & CE_NAMEMASK;
266 267 268 269 270 271 272 273 274 275
	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
276

277 278 279
	/* Compare stages  */
	flags1 &= CE_STAGEMASK;
	flags2 &= CE_STAGEMASK;
Junio C Hamano's avatar
Junio C Hamano committed
280

281 282 283 284
	if (flags1 < flags2)
		return -1;
	if (flags1 > flags2)
		return 1;
285 286 287 288 289 290 291 292 293 294 295 296
	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];
297
		int cmp = cache_name_compare(name, namelen, ce->name, ntohs(ce->ce_flags));
298
		if (!cmp)
299
			return next;
300 301 302 303 304 305
		if (cmp < 0) {
			last = next;
			continue;
		}
		first = next+1;
	}
306
	return -first-1;
307 308
}

309
/* Remove entry, return true if there are more entries to go.. */
310
int remove_cache_entry_at(int pos)
311
{
312
	active_cache_changed = 1;
313 314 315 316 317 318 319
	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
320
int remove_file_from_cache(const char *path)
321 322
{
	int pos = cache_name_pos(path, strlen(path));
323 324 325
	if (pos < 0)
		pos = -pos-1;
	while (pos < active_nr && !strcmp(active_cache[pos]->name, path))
326
		remove_cache_entry_at(pos);
327 328 329
	return 0;
}

Johannes Schindelin's avatar
Johannes Schindelin committed
330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348
int add_file_to_index(const char *path, int verbose)
{
	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);

349
	if (trust_executable_bit && has_symlinks)
350 351
		ce->ce_mode = create_ce_mode(st.st_mode);
	else {
352 353
		/* 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
354
		 */
355
		struct cache_entry *ent;
Johannes Schindelin's avatar
Johannes Schindelin committed
356
		int pos = cache_name_pos(path, namelen);
357 358 359

		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
360 361 362 363
	}

	if (index_path(ce->sha1, path, &st, 1))
		die("unable to index file %s", path);
364
	if (add_cache_entry(ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE))
Johannes Schindelin's avatar
Johannes Schindelin committed
365 366 367 368 369 370 371
		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;
}

372
int ce_same_name(struct cache_entry *a, struct cache_entry *b)
373 374 375 376 377
{
	int len = ce_namelen(a);
	return ce_namelen(b) == len && !memcmp(a->name, b->name, len);
}

378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397
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;
398 399
		if (!matchlen)
			return 1;
400 401 402 403
	}
	return 0;
}

404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467
/*
 * 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++;
	}
}

468 469 470
/*
 * Do we have another file that has the beginning components being a
 * proper superset of the name we're trying to add?
471
 */
472
static int has_file_name(const struct cache_entry *ce, int pos, int ok_to_replace)
473
{
474 475
	int retval = 0;
	int len = ce_namelen(ce);
476
	int stage = ce_stage(ce);
477
	const char *name = ce->name;
478

479 480
	while (pos < active_nr) {
		struct cache_entry *p = active_cache[pos++];
481

482
		if (len >= ce_namelen(p))
483
			break;
484 485
		if (memcmp(name, p->name, len))
			break;
486 487
		if (ce_stage(p) != stage)
			continue;
488 489 490 491 492 493
		if (p->name[len] != '/')
			continue;
		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)));
521 522
		if (pos >= 0) {
			retval = -1;
523
			if (!ok_to_replace)
524
				break;
525
			remove_cache_entry_at(pos);
526 527 528 529 530 531
			continue;
		}

		/*
		 * Trivial optimization: if we find an entry that
		 * already matches the sub-directory, then we know
532
		 * we're ok, and we can exit.
533 534
		 */
		pos = -pos-1;
535
		while (pos < active_nr) {
536
			struct cache_entry *p = active_cache[pos];
537 538 539 540 541 542 543 544 545 546 547 548
			if ((ce_namelen(p) <= len) ||
			    (p->name[len] != '/') ||
			    memcmp(p->name, name, len))
				break; /* not our subdirectory */
			if (ce_stage(p) == stage)
				/* 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++;
549
		}
550
	}
551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575
	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)
{
	/*
	 * We check if the path is a sub-path of a subsequent pathname
	 * first, since removing those will not change the position
	 * in the array
	 */
	int retval = has_file_name(ce, pos, ok_to_replace);
	/*
	 * Then check if the path might have a clashing sub-directory
	 * before it.
	 */
	return retval + has_dir_name(ce, pos, ok_to_replace);
576 577
}

578
int add_cache_entry(struct cache_entry *ce, int option)
579 580
{
	int pos;
581 582
	int ok_to_add = option & ADD_CACHE_OK_TO_ADD;
	int ok_to_replace = option & ADD_CACHE_OK_TO_REPLACE;
583
	int skip_df_check = option & ADD_CACHE_SKIP_DFCHECK;
Junio C Hamano's avatar
Junio C Hamano committed
584

585
	pos = cache_name_pos(ce->name, ntohs(ce->ce_flags));
586

Junio C Hamano's avatar
Junio C Hamano committed
587
	/* existing match? Just replace it. */
588
	if (pos >= 0) {
589
		active_cache_changed = 1;
590
		active_cache[pos] = ce;
591 592
		return 0;
	}
593
	pos = -pos-1;
594

595 596 597 598 599
	/*
	 * Inserting a merged entry ("stage 0") into the index
	 * will always replace all non-merged entries..
	 */
	if (pos < active_nr && ce_stage(ce) == 0) {
600
		while (ce_same_name(active_cache[pos], ce)) {
601
			ok_to_add = 1;
602
			if (!remove_cache_entry_at(pos))
603 604 605 606
				break;
		}
	}

607 608
	if (!ok_to_add)
		return -1;
609 610
	if (!verify_path(ce->name))
		return -1;
611

Junio C Hamano's avatar
Junio C Hamano committed
612 613
	if (!skip_df_check &&
	    check_file_directory_conflict(ce, pos, ok_to_replace)) {
614
		if (!ok_to_replace)
615
			return error("'%s' appears as both a file and as a directory", ce->name);
616
		pos = cache_name_pos(ce->name, ntohs(ce->ce_flags));
617 618
		pos = -pos-1;
	}
619

620 621 622
	/* Make sure the array is big enough .. */
	if (active_nr == active_alloc) {
		active_alloc = alloc_nr(active_alloc);
623
		active_cache = xrealloc(active_cache, active_alloc * sizeof(struct cache_entry *));
624 625 626 627 628 629 630
	}

	/* 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;
631
	active_cache_changed = 1;
632 633 634
	return 0;
}

635 636 637 638 639 640 641 642 643 644 645
/*
 * "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.
 */
646
struct cache_entry *refresh_cache_entry(struct cache_entry *ce, int really)
647 648 649 650 651
{
	struct stat st;
	struct cache_entry *updated;
	int changed, size;

652 653 654 655
	if (lstat(ce->name, &st) < 0) {
		cache_errno = errno;
		return NULL;
	}
656 657 658 659 660 661 662

	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
663
			return ce;
664 665
	}

666 667 668 669
	if (ce_modified(ce, &st, really)) {
		cache_errno = EINVAL;
		return NULL;
	}
670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711

	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;
		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;
		}

712 713
		new = refresh_cache_entry(ce, really);
		if (new == ce)
714
			continue;
715 716
		if (!new) {
			if (not_new && cache_errno == ENOENT)
717
				continue;
718
			if (really && cache_errno == EINVAL) {
719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739
				/* 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;
}

740
static int verify_hdr(struct cache_header *hdr, unsigned long size)
741 742
{
	SHA_CTX c;
743
	unsigned char sha1[20];
744

745
	if (hdr->hdr_signature != htonl(CACHE_SIGNATURE))
746
		return error("bad signature");
747 748
	if (hdr->hdr_version != htonl(2))
		return error("bad index version");
749
	SHA1_Init(&c);
750
	SHA1_Update(&c, hdr, size - 20);
751
	SHA1_Final(sha1, &c);
752
	if (hashcmp(sha1, (unsigned char *)hdr + size - 20))
753
		return error("bad index file sha1 signature");
754 755 756
	return 0;
}

757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773
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)
774 775 776 777 778 779
{
	return read_cache_from(get_index_file());
}

/* remember to discard_cache() before reading a different cache! */
int read_cache_from(const char *path)
780 781 782
{
	int fd, i;
	struct stat st;
783
	unsigned long offset;
784 785 786
	struct cache_header *hdr;

	errno = EBUSY;
787
	if (cache_mmap)
788 789
		return active_nr;

790
	errno = ENOENT;
Junio C Hamano's avatar
Junio C Hamano committed
791
	index_file_timestamp = 0;
792
	fd = open(path, O_RDONLY);
793 794 795 796 797
	if (fd < 0) {
		if (errno == ENOENT)
			return 0;
		die("index file open failed (%s)", strerror(errno));
	}
798 799

	if (!fstat(fd, &st)) {
800
		cache_mmap_size = xsize_t(st.st_size);
801
		errno = EINVAL;
802
		if (cache_mmap_size >= sizeof(struct cache_header) + 20)
803
			cache_mmap = xmmap(NULL, cache_mmap_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
804 805 806 807
		else
			die("index file smaller than expected");
	} else
		die("cannot stat the open index (%s)", strerror(errno));
808 809
	close(fd);

810 811
	hdr = cache_mmap;
	if (verify_hdr(hdr, cache_mmap_size) < 0)
812 813
		goto unmap;

814
	active_nr = ntohl(hdr->hdr_entries);
815
	active_alloc = alloc_nr(active_nr);
816
	active_cache = xcalloc(active_alloc, sizeof(struct cache_entry *));
817 818

	offset = sizeof(*hdr);
819
	for (i = 0; i < active_nr; i++) {
820
		struct cache_entry *ce = (struct cache_entry *) ((char *) cache_mmap + offset);
821 822 823
		offset = offset + ce_size(ce);
		active_cache[i] = ce;
	}
Junio C Hamano's avatar
Junio C Hamano committed
824
	index_file_timestamp = st.st_mtime;
825
	while (offset <= cache_mmap_size - 20 - 8) {
826 827 828 829 830 831 832
		/* 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;
833
		memcpy(&extsize, (char *) cache_mmap + offset + 4, 4);
834
		extsize = ntohl(extsize);
835 836
		if (read_index_extension(((const char *) cache_mmap) + offset,
					 (char *) cache_mmap + offset + 8,
837
					 extsize) < 0)
838 839 840 841
			goto unmap;
		offset += 8;
		offset += extsize;
	}
842 843 844
	return active_nr;

unmap:
845
	munmap(cache_mmap, cache_mmap_size);
846
	errno = EINVAL;
847
	die("index file corrupt");
848 849
}

850
int discard_cache(void)
851 852 853
{
	int ret;

854 855 856
	active_nr = active_cache_changed = 0;
	index_file_timestamp = 0;
	cache_tree_free(&active_cache_tree);
857 858 859 860 861 862 863 864 865 866
	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;
}

867
#define WRITE_BUFFER_SIZE 8192
868
static unsigned char write_buffer[WRITE_BUFFER_SIZE];
869 870
static unsigned long write_buffer_len;

871 872 873 874 875
static int ce_write_flush(SHA_CTX *context, int fd)
{
	unsigned int buffered = write_buffer_len;
	if (buffered) {
		SHA1_Update(context, write_buffer, buffered);
876
		if (write_in_full(fd, write_buffer, buffered) != buffered)
877 878 879 880 881 882
			return -1;
		write_buffer_len = 0;
	}
	return 0;
}

883
static int ce_write(SHA_CTX *context, int fd, void *data, unsigned int len)
884 885 886 887 888 889 890 891 892
{
	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) {
893 894
			write_buffer_len = buffered;
			if (ce_write_flush(context, fd))
895 896 897 898 899
				return -1;
			buffered = 0;
		}
		write_buffer_len = buffered;
		len -= partial;
900
		data = (char *) data + partial;
901 902 903 904
 	}
 	return 0;
}

905
static int write_index_ext_header(SHA_CTX *context, int fd,
906
				  unsigned int ext, unsigned int sz)
907 908 909
{
	ext = htonl(ext);
	sz = htonl(sz);
David Rientjes's avatar
David Rientjes committed
910 911
	return ((ce_write(context, fd, &ext, 4) < 0) ||
		(ce_write(context, fd, &sz, 4) < 0)) ? -1 : 0;
912 913 914
}

static int ce_flush(SHA_CTX *context, int fd)
915 916
{
	unsigned int left = write_buffer_len;
917

918 919
	if (left) {
		write_buffer_len = 0;
920
		SHA1_Update(context, write_buffer, left);
921
	}
922

923 924
	/* Flush first if not enough space for SHA1 signature */
	if (left + 20 > WRITE_BUFFER_SIZE) {
925
		if (write_in_full(fd, write_buffer, left) != left)
926 927 928 929
			return -1;
		left = 0;
	}

930
	/* Append the SHA1 signature at the end */
931
	SHA1_Final(write_buffer + left, context);
932
	left += 20;
933
	return (write_in_full(fd, write_buffer, left) != left) ? -1 : 0;
934 935
}

Junio C Hamano's avatar
Junio C Hamano committed
936 937 938 939 940 941 942 943 944 945 946 947 948 949 950
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)) {
951 952 953 954 955 956 957 958 959 960 961
		/* 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
		 *
962
		 * but it does not.  When the second update-index runs,
963 964 965 966 967 968 969 970 971 972 973 974 975
		 * 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
976 977 978 979
		ce->ce_size = htonl(0);
	}
}

980
int write_cache(int newfd, struct cache_entry **cache, int entries)
981 982 983
{
	SHA_CTX c;
	struct cache_header hdr;
984
	int i, removed;
985 986 987 988

	for (i = removed = 0; i < entries; i++)
		if (!cache[i]->ce_mode)
			removed++;
989

990
	hdr.hdr_signature = htonl(CACHE_SIGNATURE);
991
	hdr.hdr_version = htonl(2);
992
	hdr.hdr_entries = htonl(entries - removed);
993 994

	SHA1_Init(&c);
995
	if (ce_write(&c, newfd, &hdr, sizeof(hdr)) < 0)
996 997 998 999
		return -1;

	for (i = 0; i < entries; i++) {
		struct cache_entry *ce = cache[i];
1000 1001
		if (!ce->ce_mode)
			continue;
1002 1003
		if (index_file_timestamp &&
		    index_file_timestamp <= ntohl(ce->ce_mtime.sec))
Junio C Hamano's avatar
Junio C Hamano committed
1004
			ce_smudge_racily_clean_entry(ce);
1005
		if (ce_write(&c, newfd, ce, ce_size(ce)) < 0)
1006 1007
			return -1;
	}
1008

1009 1010 1011 1012 1013 1014 1015
	/* 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))
1016
			free(data);
1017 1018 1019 1020 1021 1022
		else {
			free(data);
			return -1;
		}
	}
	return ce_flush(&c, newfd);
1023
}