read-cache.c 12 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 7 8
#include "cache.h"

struct cache_entry **active_cache = NULL;
9
unsigned int active_nr = 0, active_alloc = 0, active_cache_changed = 0;
10

11 12 13 14 15 16 17 18 19
/*
 * 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);
20
#ifdef USE_NSEC
21 22 23 24 25 26 27 28 29 30
	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);
}

31
int ce_match_stat(struct cache_entry *ce, struct stat *st)
32 33 34
{
	unsigned int changed = 0;

35 36 37
	switch (ntohl(ce->ce_mode) & S_IFMT) {
	case S_IFREG:
		changed |= !S_ISREG(st->st_mode) ? TYPE_CHANGED : 0;
38 39 40
		/* We consider only the owner x bit to be relevant for "mode changes" */
		if (0100 & (ntohl(ce->ce_mode) ^ st->st_mode))
			changed |= MODE_CHANGED;
41 42 43 44 45 46 47
		break;
	case S_IFLNK:
		changed |= !S_ISLNK(st->st_mode) ? TYPE_CHANGED : 0;
		break;
	default:
		die("internal error: ce_mode is %o", ntohl(ce->ce_mode));
	}
48
	if (ce->ce_mtime.sec != htonl(st->st_mtime))
49
		changed |= MTIME_CHANGED;
50 51 52
	if (ce->ce_ctime.sec != htonl(st->st_ctime))
		changed |= CTIME_CHANGED;

53
#ifdef USE_NSEC
54 55 56 57 58
	/*
	 * 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.
	 */
59
	if (ce->ce_mtime.nsec != htonl(st->st_mtim.tv_nsec))
60
		changed |= MTIME_CHANGED;
61
	if (ce->ce_ctime.nsec != htonl(st->st_ctim.tv_nsec))
62
		changed |= CTIME_CHANGED;
63 64 65 66
#endif	

	if (ce->ce_uid != htonl(st->st_uid) ||
	    ce->ce_gid != htonl(st->st_gid))
67
		changed |= OWNER_CHANGED;
68
	if (ce->ce_ino != htonl(st->st_ino))
69
		changed |= INODE_CHANGED;
70 71 72 73 74 75 76 77 78 79 80

#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

81
	if (ce->ce_size != htonl(st->st_size))
82 83 84 85
		changed |= DATA_CHANGED;
	return changed;
}

86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104
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;
}

105
int cache_name_compare(const char *name1, int flags1, const char *name2, int flags2)
106
{
107 108
	int len1 = flags1 & CE_NAMEMASK;
	int len2 = flags2 & CE_NAMEMASK;
109 110 111 112 113 114 115 116 117 118
	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;
119 120 121 122
	if (flags1 < flags2)
		return -1;
	if (flags1 > flags2)
		return 1;
123 124 125 126 127 128 129 130 131 132 133 134
	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];
135
		int cmp = cache_name_compare(name, namelen, ce->name, ntohs(ce->ce_flags));
136
		if (!cmp)
137
			return next;
138 139 140 141 142 143
		if (cmp < 0) {
			last = next;
			continue;
		}
		first = next+1;
	}
144
	return -first-1;
145 146
}

147
/* Remove entry, return true if there are more entries to go.. */
148
int remove_cache_entry_at(int pos)
149
{
150
	active_cache_changed = 1;
151 152 153 154 155 156 157
	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;
}

158 159 160
int remove_file_from_cache(char *path)
{
	int pos = cache_name_pos(path, strlen(path));
161 162 163
	if (pos < 0)
		pos = -pos-1;
	while (pos < active_nr && !strcmp(active_cache[pos]->name, path))
164
		remove_cache_entry_at(pos);
165 166 167
	return 0;
}

168
int ce_same_name(struct cache_entry *a, struct cache_entry *b)
169 170 171 172 173
{
	int len = ce_namelen(a);
	return ce_namelen(b) == len && !memcmp(a->name, b->name, len);
}

174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193
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;
194 195
		if (!matchlen)
			return 1;
196 197 198 199
	}
	return 0;
}

200 201 202
/*
 * Do we have another file that has the beginning components being a
 * proper superset of the name we're trying to add?
203
 */
204
static int has_file_name(const struct cache_entry *ce, int pos, int ok_to_replace)
205
{
206 207
	int retval = 0;
	int len = ce_namelen(ce);
208
	int stage = ce_stage(ce);
209
	const char *name = ce->name;
210

211 212
	while (pos < active_nr) {
		struct cache_entry *p = active_cache[pos++];
213

214
		if (len >= ce_namelen(p))
215
			break;
216 217
		if (memcmp(name, p->name, len))
			break;
218 219
		if (ce_stage(p) != stage)
			continue;
220 221 222 223 224 225
		if (p->name[len] != '/')
			continue;
		retval = -1;
		if (!ok_to_replace)
			break;
		remove_cache_entry_at(--pos);
226
	}
227 228
	return retval;
}
229

230 231 232 233 234 235 236
/*
 * 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;
237
	int stage = ce_stage(ce);
238 239
	const char *name = ce->name;
	const char *slash = name + ce_namelen(ce);
240

241 242
	for (;;) {
		int len;
243

244 245 246 247 248 249 250
		for (;;) {
			if (*--slash == '/')
				break;
			if (slash <= ce->name)
				return retval;
		}
		len = slash - name;
251

252
		pos = cache_name_pos(name, ntohs(create_ce_flags(len, stage)));
253 254 255 256
		if (pos >= 0) {
			retval = -1;
			if (ok_to_replace)
				break;
257
			remove_cache_entry_at(pos);
258 259 260 261 262 263
			continue;
		}

		/*
		 * Trivial optimization: if we find an entry that
		 * already matches the sub-directory, then we know
264
		 * we're ok, and we can exit.
265 266
		 */
		pos = -pos-1;
267
		while (pos < active_nr) {
268
			struct cache_entry *p = active_cache[pos];
269 270 271 272 273 274 275 276 277 278 279 280
			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++;
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
	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);
308 309
}

310
int add_cache_entry(struct cache_entry *ce, int option)
311 312
{
	int pos;
313 314
	int ok_to_add = option & ADD_CACHE_OK_TO_ADD;
	int ok_to_replace = option & ADD_CACHE_OK_TO_REPLACE;
315
	int skip_df_check = option & ADD_CACHE_SKIP_DFCHECK;
316
	pos = cache_name_pos(ce->name, ntohs(ce->ce_flags));
317 318

	/* existing match? Just replace it */
319
	if (pos >= 0) {
320
		active_cache_changed = 1;
321
		active_cache[pos] = ce;
322 323
		return 0;
	}
324
	pos = -pos-1;
325

326 327 328 329 330
	/*
	 * Inserting a merged entry ("stage 0") into the index
	 * will always replace all non-merged entries..
	 */
	if (pos < active_nr && ce_stage(ce) == 0) {
331
		while (ce_same_name(active_cache[pos], ce)) {
332
			ok_to_add = 1;
333
			if (!remove_cache_entry_at(pos))
334 335 336 337
				break;
		}
	}

338 339 340
	if (!ok_to_add)
		return -1;

341
	if (!skip_df_check && check_file_directory_conflict(ce, pos, ok_to_replace)) {
342 343
		if (!ok_to_replace)
			return -1;
344
		pos = cache_name_pos(ce->name, ntohs(ce->ce_flags));
345 346
		pos = -pos-1;
	}
347

348 349 350
	/* Make sure the array is big enough .. */
	if (active_nr == active_alloc) {
		active_alloc = alloc_nr(active_alloc);
351
		active_cache = xrealloc(active_cache, active_alloc * sizeof(struct cache_entry *));
352 353 354 355 356 357 358
	}

	/* 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;
359
	active_cache_changed = 1;
360 361 362
	return 0;
}

363 364 365 366 367
static int verify_hdr(struct cache_header *hdr, unsigned long size)
{
	SHA_CTX c;
	unsigned char sha1[20];

368
	if (hdr->hdr_signature != htonl(CACHE_SIGNATURE))
369
		return error("bad signature");
370 371
	if (hdr->hdr_version != htonl(2))
		return error("bad index version");
372
	SHA1_Init(&c);
373
	SHA1_Update(&c, hdr, size - 20);
374
	SHA1_Final(sha1, &c);
375 376
	if (memcmp(sha1, (void *)hdr + size - 20, 20))
		return error("bad index file sha1 signature");
377 378 379 380 381 382 383 384 385 386 387 388 389 390 391
	return 0;
}

int read_cache(void)
{
	int fd, i;
	struct stat st;
	unsigned long size, offset;
	void *map;
	struct cache_header *hdr;

	errno = EBUSY;
	if (active_cache)
		return error("more than one cachefile");
	errno = ENOENT;
392
	fd = open(get_index_file(), O_RDONLY);
393 394 395
	if (fd < 0)
		return (errno == ENOENT) ? 0 : error("open failed");

396
	size = 0; // avoid gcc warning
Pavel Roskin's avatar
Pavel Roskin committed
397
	map = MAP_FAILED;
398 399 400
	if (!fstat(fd, &st)) {
		size = st.st_size;
		errno = EINVAL;
401
		if (size >= sizeof(struct cache_header) + 20)
402
			map = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0);
403 404
	}
	close(fd);
Pavel Roskin's avatar
Pavel Roskin committed
405
	if (map == MAP_FAILED)
406 407 408 409 410 411
		return error("mmap failed");

	hdr = map;
	if (verify_hdr(hdr, size) < 0)
		goto unmap;

412
	active_nr = ntohl(hdr->hdr_entries);
413 414 415 416
	active_alloc = alloc_nr(active_nr);
	active_cache = calloc(active_alloc, sizeof(struct cache_entry *));

	offset = sizeof(*hdr);
417
	for (i = 0; i < active_nr; i++) {
418 419 420 421 422 423 424 425 426 427 428 429
		struct cache_entry *ce = map + offset;
		offset = offset + ce_size(ce);
		active_cache[i] = ce;
	}
	return active_nr;

unmap:
	munmap(map, size);
	errno = EINVAL;
	return error("verify header failed");
}

430
#define WRITE_BUFFER_SIZE 8192
431
static unsigned char write_buffer[WRITE_BUFFER_SIZE];
432 433
static unsigned long write_buffer_len;

434
static int ce_write(SHA_CTX *context, int fd, void *data, unsigned int len)
435 436 437 438 439 440 441 442 443
{
	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) {
444
			SHA1_Update(context, write_buffer, WRITE_BUFFER_SIZE);
445 446 447 448 449 450 451 452 453 454 455
			if (write(fd, write_buffer, WRITE_BUFFER_SIZE) != WRITE_BUFFER_SIZE)
				return -1;
			buffered = 0;
		}
		write_buffer_len = buffered;
		len -= partial;
		data += partial;
 	}
 	return 0;
}

456
static int ce_flush(SHA_CTX *context, int fd)
457 458
{
	unsigned int left = write_buffer_len;
459

460 461
	if (left) {
		write_buffer_len = 0;
462
		SHA1_Update(context, write_buffer, left);
463
	}
464

465 466 467 468 469 470 471
	/* Flush first if not enough space for SHA1 signature */
	if (left + 20 > WRITE_BUFFER_SIZE) {
		if (write(fd, write_buffer, left) != left)
			return -1;
		left = 0;
	}

472 473 474 475 476
	/* Append the SHA1 signature at the end */
	SHA1_Final(write_buffer + left, context);
	left += 20;
	if (write(fd, write_buffer, left) != left)
		return -1;
477 478 479
	return 0;
}

480 481 482 483
int write_cache(int newfd, struct cache_entry **cache, int entries)
{
	SHA_CTX c;
	struct cache_header hdr;
484 485 486 487 488
	int i, removed;

	for (i = removed = 0; i < entries; i++)
		if (!cache[i]->ce_mode)
			removed++;
489

490
	hdr.hdr_signature = htonl(CACHE_SIGNATURE);
491
	hdr.hdr_version = htonl(2);
492
	hdr.hdr_entries = htonl(entries - removed);
493 494

	SHA1_Init(&c);
495
	if (ce_write(&c, newfd, &hdr, sizeof(hdr)) < 0)
496 497 498 499
		return -1;

	for (i = 0; i < entries; i++) {
		struct cache_entry *ce = cache[i];
500 501
		if (!ce->ce_mode)
			continue;
502
		if (ce_write(&c, newfd, ce, ce_size(ce)) < 0)
503 504
			return -1;
	}
505
	return ce_flush(&c, newfd);
506
}