Gotweb

Blob

Date:: Sun Jan 14 16:49:25 2018 UTC
Message:: refactor dump_packed_object() a bit; no functional change
Actions:: History | Blame | Raw File
1 /*
2  * Copyright (c) 2018 Stefan Sperling <stsp@openbsd.org>
3  *
4  * Permission to use, copy, modify, and distribute this software for any
5  * purpose with or without fee is hereby granted, provided that the above
6  * copyright notice and this permission notice appear in all copies.
7  *
8  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
15  */
16 
17 #include <sys/types.h>
18 #include <sys/stat.h>
19 #include <sys/queue.h>
20 
21 #include <dirent.h>
22 #include <errno.h>
23 #include <stdio.h>
24 #include <stdint.h>
25 #include <stdlib.h>
26 #include <string.h>
27 #include <limits.h>
28 #include <sha1.h>
29 #include <endian.h>
30 #include <zlib.h>
31 
32 #include "got_error.h"
33 #include "got_object.h"
34 #include "got_repository.h"
35 #include "got_sha1.h"
36 #include "pack.h"
37 #include "path.h"
38 
39 #define GOT_PACK_PREFIX		"pack-"
40 #define GOT_PACKFILE_SUFFIX	".pack"
41 #define GOT_PACKIDX_SUFFIX		".idx"
42 #define GOT_PACKFILE_NAMELEN	(strlen(GOT_PACK_PREFIX) + \
43 				SHA1_DIGEST_STRING_LENGTH - 1 + \
44 				strlen(GOT_PACKFILE_SUFFIX))
45 #define GOT_PACKIDX_NAMELEN	(strlen(GOT_PACK_PREFIX) + \
46 				SHA1_DIGEST_STRING_LENGTH - 1 + \
47 				strlen(GOT_PACKIDX_SUFFIX))
48 
49 #ifndef MIN
50 #define	MIN(_a,_b) ((_a) < (_b) ? (_a) : (_b))
51 #endif
52 
53 static const struct got_error *
54 verify_fanout_table(uint32_t *fanout_table)
55 {
56 	int i;
57 
58 	for (i = 0; i < 0xff - 1; i++) {
59 		if (be32toh(fanout_table[i]) > be32toh(fanout_table[i + 1]))
60 			return got_error(GOT_ERR_BAD_PACKIDX);
61 	}
62 
63 	return NULL;
64 }
65 
66 static const struct got_error *
67 get_packfile_size(size_t *size, const char *path_idx)
68 {
69 	struct stat sb;
70 	char *path_pack;
71 	char base_path[PATH_MAX];
72 	char *dot;
73 
74 	if (strlcpy(base_path, path_idx, PATH_MAX) > PATH_MAX)
75 		return got_error(GOT_ERR_NO_SPACE);
76 
77 	dot = strrchr(base_path, '.');
78 	if (dot == NULL)
79 		return got_error(GOT_ERR_BAD_PATH);
80 	*dot = '\0';
81 	if (asprintf(&path_pack, "%s.pack", base_path) == -1)
82 		return got_error(GOT_ERR_NO_MEM);
83 
84 	if (stat(path_pack, &sb) != 0) {
85 		free(path_pack);
86 		return got_error_from_errno();
87 	}
88 
89 	free(path_pack);
90 	*size = sb.st_size;
91 	return 0;
92 }
93 
94 const struct got_error *
95 got_packidx_open(struct got_packidx_v2_hdr **packidx, const char *path)
96 {
97 	struct got_packidx_v2_hdr *p;
98 	FILE *f;
99 	const struct got_error *err = NULL;
100 	size_t n, nobj, packfile_size;
101 	SHA1_CTX ctx;
102 	uint8_t sha1[SHA1_DIGEST_LENGTH];
103 
104 	SHA1Init(&ctx);
105 
106 	f = fopen(path, "rb");
107 	if (f == NULL)
108 		return got_error(GOT_ERR_BAD_PATH);
109 
110 	err = get_packfile_size(&packfile_size, path);
111 	if (err)
112 		return err;
113 
114 	p = calloc(1, sizeof(*p));
115 	if (p == NULL) {
116 		err = got_error(GOT_ERR_NO_MEM);
117 		goto done;
118 	}
119 
120 	n = fread(&p->magic, sizeof(p->magic), 1, f);
121 	if (n != 1) {
122 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
123 		goto done;
124 	}
125 
126 	if (betoh32(p->magic) != GOT_PACKIDX_V2_MAGIC) {
127 		err = got_error(GOT_ERR_BAD_PACKIDX);
128 		goto done;
129 	}
130 
131 	SHA1Update(&ctx, (uint8_t *)&p->magic, sizeof(p->magic));
132 
133 	n = fread(&p->version, sizeof(p->version), 1, f);
134 	if (n != 1) {
135 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
136 		goto done;
137 	}
138 
139 	if (betoh32(p->version) != GOT_PACKIDX_VERSION) {
140 		err = got_error(GOT_ERR_BAD_PACKIDX);
141 		goto done;
142 	}
143 
144 	SHA1Update(&ctx, (uint8_t *)&p->version, sizeof(p->version));
145 
146 	n = fread(&p->fanout_table, sizeof(p->fanout_table), 1, f);
147 	if (n != 1) {
148 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
149 		goto done;
150 	}
151 
152 	err = verify_fanout_table(p->fanout_table);
153 	if (err)
154 		goto done;
155 
156 	SHA1Update(&ctx, (uint8_t *)p->fanout_table, sizeof(p->fanout_table));
157 
158 	nobj = betoh32(p->fanout_table[0xff]);
159 
160 	p->sorted_ids = calloc(nobj, sizeof(*p->sorted_ids));
161 	if (p->sorted_ids == NULL) {
162 		err = got_error(GOT_ERR_NO_MEM);
163 		goto done;
164 	}
165 
166 	n = fread(p->sorted_ids, sizeof(*p->sorted_ids), nobj, f);
167 	if (n != nobj) {
168 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
169 		goto done;
170 	}
171 
172 	SHA1Update(&ctx, (uint8_t *)p->sorted_ids,
173 	    nobj * sizeof(*p->sorted_ids));
174 
175 	p->crc32 = calloc(nobj, sizeof(*p->crc32));
176 	if (p->crc32 == NULL) {
177 		err = got_error(GOT_ERR_NO_MEM);
178 		goto done;
179 	}
180 
181 	n = fread(p->crc32, sizeof(*p->crc32), nobj, f);
182 	if (n != nobj) {
183 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
184 		goto done;
185 	}
186 
187 	SHA1Update(&ctx, (uint8_t *)p->crc32, nobj * sizeof(*p->crc32));
188 
189 	p->offsets = calloc(nobj, sizeof(*p->offsets));
190 	if (p->offsets == NULL) {
191 		err = got_error(GOT_ERR_NO_MEM);
192 		goto done;
193 	}
194 
195 	n = fread(p->offsets, sizeof(*p->offsets), nobj, f);
196 	if (n != nobj) {
197 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
198 		goto done;
199 	}
200 
201 	SHA1Update(&ctx, (uint8_t *)p->offsets, nobj * sizeof(*p->offsets));
202 
203 	/* Large file offsets are contained only in files > 2GB. */
204 	if (packfile_size <= 0x80000000)
205 		goto checksum;
206 
207 	p->large_offsets = calloc(nobj, sizeof(*p->large_offsets));
208 	if (p->large_offsets == NULL) {
209 		err = got_error(GOT_ERR_NO_MEM);
210 		goto done;
211 	}
212 
213 	n = fread(p->large_offsets, sizeof(*p->large_offsets), nobj, f);
214 	if (n != nobj) {
215 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
216 		goto done;
217 	}
218 
219 	SHA1Update(&ctx, (uint8_t*)p->large_offsets,
220 	    nobj * sizeof(*p->large_offsets));
221 
222 checksum:
223 	n = fread(&p->trailer, sizeof(p->trailer), 1, f);
224 	if (n != 1) {
225 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
226 		goto done;
227 	}
228 
229 	SHA1Update(&ctx, p->trailer.packfile_sha1, SHA1_DIGEST_LENGTH);
230 	SHA1Final(sha1, &ctx);
231 	if (memcmp(p->trailer.packidx_sha1, sha1, SHA1_DIGEST_LENGTH) != 0)
232 		err = got_error(GOT_ERR_PACKIDX_CSUM);
233 done:
234 	fclose(f);
235 	if (err)
236 		got_packidx_close(p);
237 	else
238 		*packidx = p;
239 	return err;
240 }
241 
242 void
243 got_packidx_close(struct got_packidx_v2_hdr *packidx)
244 {
245 	free(packidx->sorted_ids);
246 	free(packidx->offsets);
247 	free(packidx->crc32);
248 	free(packidx->large_offsets);
249 	free(packidx);
250 }
251 
252 static int
253 is_packidx_filename(const char *name, size_t len)
254 {
255 	if (len != GOT_PACKIDX_NAMELEN)
256 		return 0;
257 
258 	if (strncmp(name, GOT_PACK_PREFIX, strlen(GOT_PACK_PREFIX)) != 0)
259 		return 0;
260 
261 	if (strcmp(name + strlen(GOT_PACK_PREFIX) +
262 	    SHA1_DIGEST_STRING_LENGTH - 1, GOT_PACKIDX_SUFFIX) != 0)
263 		return 0;
264 
265 	return 1;
266 }
267 
268 static off_t
269 get_object_offset(struct got_packidx_v2_hdr *packidx, int idx)
270 {
271 	uint32_t totobj = betoh32(packidx->fanout_table[0xff]);
272 	uint32_t offset = betoh32(packidx->offsets[idx]);
273 	if (offset & GOT_PACKIDX_OFFSET_VAL_IS_LARGE_IDX) {
274 		uint64_t loffset;
275 		idx = offset & GOT_PACKIDX_OFFSET_VAL_MASK;
276 		if (idx < 0 || idx > totobj || packidx->large_offsets == NULL)
277 			return -1;
278 		loffset = betoh64(packidx->large_offsets[idx]);
279 		return (loffset > INT64_MAX ? -1 : (off_t)loffset);
280 	}
281 	return (off_t)(offset & GOT_PACKIDX_OFFSET_VAL_MASK);
282 }
283 
284 static int
285 get_object_idx(struct got_packidx_v2_hdr *packidx, struct got_object_id *id)
286 {
287 	u_int8_t id0 = id->sha1[0];
288 	uint32_t totobj = betoh32(packidx->fanout_table[0xff]);
289 	int i = 0;
290 
291 	if (id0 > 0)
292 		i = betoh32(packidx->fanout_table[id0 - 1]);
293 
294 	while (i < totobj) {
295 		struct got_object_id *oid = &packidx->sorted_ids[i++];
296 		uint32_t offset;
297 		int cmp = got_object_id_cmp(id, oid);
298 
299 		if (cmp < 0)
300 			continue;
301 		if (cmp > 0)
302 			break;
303 
304 		return i;
305 	}
306 
307 	return -1;
308 }
309 
310 const struct got_error *
311 read_packfile_hdr(FILE *f, struct got_packidx_v2_hdr *packidx)
312 {
313 	const struct got_error *err = NULL;
314 	uint32_t totobj = betoh32(packidx->fanout_table[0xff]);
315 	struct got_packfile_hdr hdr;
316 	size_t n;
317 
318 	n = fread(&hdr, sizeof(hdr), 1, f);
319 	if (n != 1)
320 		return got_ferror(f, GOT_ERR_BAD_PACKIDX);
321 
322 	if (betoh32(hdr.signature) != GOT_PACKFILE_SIGNATURE ||
323 	    betoh32(hdr.version) != GOT_PACKFILE_VERSION ||
324 	    betoh32(hdr.nobjects) != totobj)
325 		err = got_error(GOT_ERR_BAD_PACKFILE);
326 
327 	return err;
328 }
329 
330 static const struct got_error *
331 dump_plain_object(FILE *infile, uint8_t type, uint64_t size, FILE *outfile)
332 {
333 	const char *type_tag = got_object_get_type_tag(type);
334 	size_t n;
335 
336 	if (type_tag == NULL)
337 		return got_error(GOT_ERR_OBJ_TYPE);
338 
339 	fprintf(outfile, "%s %llu", type_tag, size);
340 	fputc('\0', outfile);
341 
342 	while (size > 0) {
343 		uint8_t data[2048];
344 		size_t len = MIN(size, sizeof(data));
345 
346 		n = fread(data, len, 1, infile);
347 		if (n != 1)
348 			return got_ferror(infile, GOT_ERR_BAD_PACKIDX);
349 
350 		n = fwrite(data, len, 1, outfile);
351 		if (n != 1)
352 			return got_ferror(outfile, GOT_ERR_BAD_PACKIDX);
353 
354 		size -= len;
355 	}
356 
357 	return NULL;
358 }
359 
360 static const struct got_error *
361 decode_type_and_size(uint8_t *type, uint64_t *size, FILE *packfile)
362 {
363 	uint8_t t = 0;
364 	uint64_t s = 0;
365 	uint8_t sizeN;
366 	size_t n;
367 	int i = 0;
368 
369 	do {
370 		/* We do not support size values which don't fit in 64 bit. */
371 		if (i > 9)
372 			return got_error(GOT_ERR_NO_SPACE);
373 
374 		n = fread(&sizeN, sizeof(sizeN), 1, packfile);
375 		if (n != 1)
376 			return got_ferror(packfile, GOT_ERR_BAD_PACKIDX);
377 
378 		if (i == 0) {
379 			t = (sizeN & GOT_PACK_OBJ_SIZE0_TYPE_MASK) >>
380 			    GOT_PACK_OBJ_SIZE0_TYPE_MASK_SHIFT;
381 			s = (sizeN & GOT_PACK_OBJ_SIZE0_VAL_MASK);
382 		} else {
383 			size_t shift = 4 + 7 * (i - 1);
384 			s |= ((sizeN & GOT_PACK_OBJ_SIZE_VAL_MASK) << shift);
385 		}
386 		i++;
387 	} while (sizeN & GOT_PACK_OBJ_SIZE_MORE);
388 
389 	*type = t;
390 	*size = s;
391 	return NULL;
392 }
393 
394 static const struct got_error *
395 dump_packed_object(FILE **f, FILE *packfile, off_t offset)
396 {
397 	const struct got_error *err = NULL;
398 	const char *template = "/tmp/got.XXXXXXXXXX";
399 	uint8_t type;
400 	uint64_t size;
401 	FILE *outfile = NULL;
402 
403 	*f = got_opentemp();
404 	if (*f == NULL) {
405 		err = got_error(GOT_ERR_FILE_OPEN);
406 		goto done;
407 	}
408 
409 	if (fseeko(packfile, offset, SEEK_SET) != 0) {
410 		err = got_error_from_errno();
411 		goto done;
412 	}
413 
414 	err = decode_type_and_size(&type, &size, packfile);
415 	if (err)
416 		goto done;
417 
418 	switch (type) {
419 	case GOT_OBJ_TYPE_COMMIT:
420 	case GOT_OBJ_TYPE_TREE:
421 	case GOT_OBJ_TYPE_BLOB:
422 		err = dump_plain_object(packfile, type, size, *f);
423 		break;
424 	case GOT_OBJ_TYPE_REF_DELTA:
425 	case GOT_OBJ_TYPE_TAG:
426 	case GOT_OBJ_TYPE_OFFSET_DELTA:
427 	default:
428 		err = got_error(GOT_ERR_NOT_IMPL);
429 		goto done;
430 	}
431 
432 	rewind(*f);
433 done:
434 	if (err && *f)
435 		fclose(*f);
436 	return err;
437 }
438 
439 static const struct got_error *
440 extract_object(FILE **f, const char *path_packdir,
441     struct got_packidx_v2_hdr *packidx, struct got_object_id *id)
442 {
443 	const struct got_error *err = NULL;
444 	int idx = get_object_idx(packidx, id);
445 	off_t offset;
446 	char *path_packfile;
447 	FILE *packfile;
448 	char hex[SHA1_DIGEST_STRING_LENGTH];
449 	char *sha1str;
450 
451 	*f = NULL;
452 	if (idx == -1) /* object not found in pack index */
453 		return NULL;
454 
455 	offset = get_object_offset(packidx, idx);
456 	if (offset == (uint64_t)-1)
457 		return got_error(GOT_ERR_BAD_PACKIDX);
458 
459 	sha1str = got_sha1_digest_to_str(packidx->trailer.packfile_sha1,
460 	    hex, sizeof(hex));
461 	if (sha1str == NULL)
462 		return got_error(GOT_ERR_PACKIDX_CSUM);
463 
464 	if (asprintf(&path_packfile, "%s/%s%s%s", path_packdir,
465 	    GOT_PACK_PREFIX, sha1str, GOT_PACKFILE_SUFFIX) == -1)
466 		return got_error(GOT_ERR_NO_MEM);
467 
468 	packfile = fopen(path_packfile, "rb");
469 	if (packfile == NULL) {
470 		err = got_error_from_errno();
471 		goto done;
472 	}
473 
474 	err = read_packfile_hdr(packfile, packidx);
475 	if (err)
476 		goto done;
477 
478 	printf("Dumping object at offset %llu\n", offset);
479 	err = dump_packed_object(f, packfile, offset);
480 	if (err)
481 		goto done;
482 
483 done:
484 	free(path_packfile);
485 	if (packfile && fclose(packfile) == -1 && err == 0)
486 		err = got_error_from_errno();
487 	return err;
488 }
489 
490 const struct got_error *
491 got_packfile_extract_object(FILE **f, struct got_object_id *id,
492     struct got_repository *repo)
493 {
494 	const struct got_error *err = NULL;
495 	DIR *packdir = NULL;
496 	struct dirent *dent;
497 	char *path_packdir = got_repo_get_path_objects_pack(repo);
498 
499 	if (path_packdir == NULL) {
500 		err = got_error(GOT_ERR_NO_MEM);
501 		goto done;
502 	}
503 
504 	packdir = opendir(path_packdir);
505 	if (packdir == NULL) {
506 		err = got_error_from_errno();
507 		goto done;
508 	}
509 
510 	while ((dent = readdir(packdir)) != NULL) {
511 		struct got_packidx_v2_hdr *packidx;
512 		char *path_packidx, *path_object;
513 
514 		if (!is_packidx_filename(dent->d_name, dent->d_namlen))
515 			continue;
516 
517 		if (asprintf(&path_packidx, "%s/%s", path_packdir,
518 		    dent->d_name) == -1) {
519 			err = got_error(GOT_ERR_NO_MEM);
520 			goto done;
521 		}
522 
523 		err = got_packidx_open(&packidx, path_packidx);
524 		free(path_packidx);
525 		if (err)
526 			goto done;
527 
528 		err = extract_object(f, path_packdir, packidx, id);
529 		if (err)
530 			goto done;
531 		if (*f != NULL)
532 			break;
533 	}
534 
535 done:
536 	free(path_packdir);
537 	if (packdir && closedir(packdir) != 0 && err == 0)
538 		err = got_error_from_errno();
539 	return err;
540 }