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 #include <sys/uio.h>
21 #include <sys/socket.h>
22 #include <sys/syslimits.h>
23 #include <sys/wait.h>
24 
25 #include <errno.h>
26 #include <stdio.h>
27 #include <stdlib.h>
28 #include <string.h>
29 #include <stdint.h>
30 #include <sha1.h>
31 #include <zlib.h>
32 #include <ctype.h>
33 #include <limits.h>
34 #include <imsg.h>
35 #include <time.h>
36 #include <unistd.h>
37 
38 #include "got_error.h"
39 #include "got_object.h"
40 #include "got_repository.h"
41 #include "got_opentemp.h"
42 
43 #include "got_lib_sha1.h"
44 #include "got_lib_delta.h"
45 #include "got_lib_privsep.h"
46 #include "got_lib_pack.h"
47 #include "got_lib_inflate.h"
48 #include "got_lib_object.h"
49 #include "got_lib_object_cache.h"
50 #include "got_lib_repository.h"
51 
52 #ifndef nitems
53 #define nitems(_a) (sizeof(_a) / sizeof((_a)[0]))
54 #endif
55 
56 #define GOT_OBJ_TAG_COMMIT	"commit"
57 #define GOT_OBJ_TAG_TREE	"tree"
58 #define GOT_OBJ_TAG_BLOB	"blob"
59 
60 #define GOT_COMMIT_TAG_TREE		"tree "
61 #define GOT_COMMIT_TAG_PARENT		"parent "
62 #define GOT_COMMIT_TAG_AUTHOR		"author "
63 #define GOT_COMMIT_TAG_COMMITTER	"committer "
64 
65 const struct got_error *
66 got_object_qid_alloc_partial(struct got_object_qid **qid)
67 {
68 	const struct got_error *err = NULL;
69 
70 	*qid = malloc(sizeof(**qid));
71 	if (*qid == NULL)
72 		return got_error_from_errno();
73 
74 	(*qid)->id = malloc(sizeof(*((*qid)->id)));
75 	if ((*qid)->id == NULL) {
76 		err = got_error_from_errno();
77 		got_object_qid_free(*qid);
78 		*qid = NULL;
79 		return err;
80 	}
81 
82 	return NULL;
83 }
84 
85 const struct got_error *
86 got_object_id_str(char **outbuf, struct got_object_id *id)
87 {
88 	static const size_t len = SHA1_DIGEST_STRING_LENGTH;
89 
90 	*outbuf = malloc(len);
91 	if (*outbuf == NULL)
92 		return got_error_from_errno();
93 
94 	if (got_sha1_digest_to_str(id->sha1, *outbuf, len) == NULL) {
95 		free(*outbuf);
96 		*outbuf = NULL;
97 		return got_error(GOT_ERR_BAD_OBJ_ID_STR);
98 	}
99 
100 	return NULL;
101 }
102 
103 void
104 got_object_close(struct got_object *obj)
105 {
106 	if (obj->refcnt > 0) {
107 		obj->refcnt--;
108 		if (obj->refcnt > 0)
109 			return;
110 	}
111 
112 	if (obj->flags & GOT_OBJ_FLAG_DELTIFIED) {
113 		struct got_delta *delta;
114 		while (!SIMPLEQ_EMPTY(&obj->deltas.entries)) {
115 			delta = SIMPLEQ_FIRST(&obj->deltas.entries);
116 			SIMPLEQ_REMOVE_HEAD(&obj->deltas.entries, entry);
117 			got_delta_close(delta);
118 		}
119 	}
120 	if (obj->flags & GOT_OBJ_FLAG_PACKED)
121 		free(obj->path_packfile);
122 	free(obj);
123 }
124 
125 void
126 got_object_qid_free(struct got_object_qid *qid)
127 {
128 	free(qid->id);
129 	free(qid);
130 }
131 
132 struct got_commit_object *
133 got_object_commit_alloc_partial(void)
134 {
135 	struct got_commit_object *commit;
136 
137 	commit = calloc(1, sizeof(*commit));
138 	if (commit == NULL)
139 		return NULL;
140 	commit->tree_id = malloc(sizeof(*commit->tree_id));
141 	if (commit->tree_id == NULL) {
142 		free(commit);
143 		return NULL;
144 	}
145 
146 	SIMPLEQ_INIT(&commit->parent_ids);
147 
148 	return commit;
149 }
150 
151 const struct got_error *
152 got_object_commit_add_parent(struct got_commit_object *commit,
153     const char *id_str)
154 {
155 	const struct got_error *err = NULL;
156 	struct got_object_qid *qid;
157 
158 	err = got_object_qid_alloc_partial(&qid);
159 	if (err)
160 		return err;
161 
162 	if (!got_parse_sha1_digest(qid->id->sha1, id_str)) {
163 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
164 		free(qid->id);
165 		free(qid);
166 		return err;
167 	}
168 
169 	SIMPLEQ_INSERT_TAIL(&commit->parent_ids, qid, entry);
170 	commit->nparents++;
171 
172 	return NULL;
173 }
174 
175 static const struct got_error *
176 parse_gmtoff(time_t *gmtoff, const char *tzstr)
177 {
178 	int sign = 1;
179 	const char *p = tzstr;
180 	time_t h, m;
181 
182 	*gmtoff = 0;
183 
184 	if (*p == '-')
185 		sign = -1;
186 	else if (*p != '+')
187 		return got_error(GOT_ERR_BAD_OBJ_DATA);
188 	p++;
189 	if (!isdigit(*p) && !isdigit(*(p + 1)))
190 		return got_error(GOT_ERR_BAD_OBJ_DATA);
191 	h = (((*p - '0') * 10) + (*(p + 1) - '0'));
192 
193 	p += 2;
194 	if (!isdigit(*p) && !isdigit(*(p + 1)))
195 		return got_error(GOT_ERR_BAD_OBJ_DATA);
196 	m = ((*p - '0') * 10) + (*(p + 1) - '0');
197 
198 	*gmtoff = (h * 60 * 60 + m * 60) * sign;
199 	return NULL;
200 }
201 
202 static const struct got_error *
203 parse_commit_time(time_t *time, time_t *gmtoff, char *committer)
204 {
205 	const struct got_error *err = NULL;
206 	const char *errstr;
207 	char *space, *tzstr;
208 
209 	/* Parse and strip off trailing timezone indicator string. */
210 	space = strrchr(committer, ' ');
211 	if (space == NULL)
212 		return got_error(GOT_ERR_BAD_OBJ_DATA);
213 	tzstr = strdup(space + 1);
214 	if (tzstr == NULL)
215 		return got_error_from_errno();
216 	err = parse_gmtoff(gmtoff, tzstr);
217 	free(tzstr);
218 	if (err)
219 		return err;
220 	*space = '\0';
221 
222 	/* Timestamp is separated from committer name + email by space. */
223 	space = strrchr(committer, ' ');
224 	if (space == NULL)
225 		return got_error(GOT_ERR_BAD_OBJ_DATA);
226 
227 	/* Timestamp parsed here is expressed in comitter's local time. */
228 	*time = strtonum(space + 1, 0, INT64_MAX, &errstr);
229 	if (errstr)
230 		return got_error(GOT_ERR_BAD_OBJ_DATA);
231 
232 	/* Express the time stamp in UTC. */
233 	*time -= *gmtoff;
234 
235 	/* Strip off parsed time information, leaving just author and email. */
236 	*space = '\0';
237 
238 	return NULL;
239 }
240 
241 void
242 got_object_commit_close(struct got_commit_object *commit)
243 {
244 	struct got_object_qid *qid;
245 
246 	if (commit->refcnt > 0) {
247 		commit->refcnt--;
248 		if (commit->refcnt > 0)
249 			return;
250 	}
251 
252 	while (!SIMPLEQ_EMPTY(&commit->parent_ids)) {
253 		qid = SIMPLEQ_FIRST(&commit->parent_ids);
254 		SIMPLEQ_REMOVE_HEAD(&commit->parent_ids, entry);
255 		got_object_qid_free(qid);
256 	}
257 
258 	free(commit->tree_id);
259 	free(commit->author);
260 	free(commit->committer);
261 	free(commit->logmsg);
262 	free(commit);
263 }
264 
265 const struct got_error *
266 got_object_parse_commit(struct got_commit_object **commit, char *buf, size_t len)
267 {
268 	const struct got_error *err = NULL;
269 	char *s = buf;
270 	size_t tlen;
271 	ssize_t remain = (ssize_t)len;
272  
273 	*commit = got_object_commit_alloc_partial();
274 	if (*commit == NULL)
275 		return got_error_from_errno();
276 
277 	tlen = strlen(GOT_COMMIT_TAG_TREE);
278 	if (strncmp(s, GOT_COMMIT_TAG_TREE, tlen) == 0) {
279 		remain -= tlen;
280 		if (remain < SHA1_DIGEST_STRING_LENGTH) {
281 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
282 			goto done;
283 		}
284 		s += tlen;
285 		if (!got_parse_sha1_digest((*commit)->tree_id->sha1, s)) {
286 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
287 			goto done;
288 		}
289 		remain -= SHA1_DIGEST_STRING_LENGTH;
290 		s += SHA1_DIGEST_STRING_LENGTH;
291 	} else {
292 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
293 		goto done;
294 	}
295 
296 	tlen = strlen(GOT_COMMIT_TAG_PARENT);
297 	while (strncmp(s, GOT_COMMIT_TAG_PARENT, tlen) == 0) {
298 		remain -= tlen;
299 		if (remain < SHA1_DIGEST_STRING_LENGTH) {
300 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
301 			goto done;
302 		}
303 		s += tlen;
304 		err = got_object_commit_add_parent(*commit, s);
305 		if (err)
306 			goto done;
307 
308 		remain -= SHA1_DIGEST_STRING_LENGTH;
309 		s += SHA1_DIGEST_STRING_LENGTH;
310 	}
311 
312 	tlen = strlen(GOT_COMMIT_TAG_AUTHOR);
313 	if (strncmp(s, GOT_COMMIT_TAG_AUTHOR, tlen) == 0) {
314 		char *p;
315 		size_t slen;
316 
317 		remain -= tlen;
318 		if (remain <= 0) {
319 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
320 			goto done;
321 		}
322 		s += tlen;
323 		p = strchr(s, '\n');
324 		if (p == NULL) {
325 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
326 			goto done;
327 		}
328 		*p = '\0';
329 		slen = strlen(s);
330 		err = parse_commit_time(&(*commit)->author_time,
331 		    &(*commit)->author_gmtoff, s);
332 		if (err)
333 			goto done;
334 		(*commit)->author = strdup(s);
335 		if ((*commit)->author == NULL) {
336 			err = got_error_from_errno();
337 			goto done;
338 		}
339 		s += slen + 1;
340 		remain -= slen + 1;
341 	}
342 
343 	tlen = strlen(GOT_COMMIT_TAG_COMMITTER);
344 	if (strncmp(s, GOT_COMMIT_TAG_COMMITTER, tlen) == 0) {
345 		char *p;
346 		size_t slen;
347 
348 		remain -= tlen;
349 		if (remain <= 0) {
350 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
351 			goto done;
352 		}
353 		s += tlen;
354 		p = strchr(s, '\n');
355 		if (p == NULL) {
356 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
357 			goto done;
358 		}
359 		*p = '\0';
360 		slen = strlen(s);
361 		err = parse_commit_time(&(*commit)->committer_time,
362 		    &(*commit)->committer_gmtoff, s);
363 		if (err)
364 			goto done;
365 		(*commit)->committer = strdup(s);
366 		if ((*commit)->committer == NULL) {
367 			err = got_error_from_errno();
368 			goto done;
369 		}
370 		s += slen + 1;
371 		remain -= slen + 1;
372 	}
373 
374 	(*commit)->logmsg = strndup(s, remain);
375 	if ((*commit)->logmsg == NULL) {
376 		err = got_error_from_errno();
377 		goto done;
378 	}
379 done:
380 	if (err) {
381 		got_object_commit_close(*commit);
382 		*commit = NULL;
383 	}
384 	return err;
385 }
386 
387 void
388 got_object_tree_entry_close(struct got_tree_entry *te)
389 {
390 	free(te->id);
391 	free(te->name);
392 	free(te);
393 }
394 
395 void
396 got_object_tree_close(struct got_tree_object *tree)
397 {
398 	struct got_tree_entry *te;
399 
400 	if (tree->refcnt > 0) {
401 		tree->refcnt--;
402 		if (tree->refcnt > 0)
403 			return;
404 	}
405 
406 	while (!SIMPLEQ_EMPTY(&tree->entries.head)) {
407 		te = SIMPLEQ_FIRST(&tree->entries.head);
408 		SIMPLEQ_REMOVE_HEAD(&tree->entries.head, entry);
409 		got_object_tree_entry_close(te);
410 	}
411 
412 	free(tree);
413 }
414 
415 struct got_tree_entry *
416 got_alloc_tree_entry_partial(void)
417 {
418 	struct got_tree_entry *te;
419 
420 	te = malloc(sizeof(*te));
421 	if (te == NULL)
422 		return NULL;
423 
424 	te->id = malloc(sizeof(*te->id));
425 	if (te->id == NULL) {
426 		free(te);
427 		te = NULL;
428 	}
429 	return te;
430 }
431 
432 static const struct got_error *
433 parse_tree_entry(struct got_tree_entry **te, size_t *elen, char *buf,
434     size_t maxlen)
435 {
436 	char *p = buf, *space;
437 	const struct got_error *err = NULL;
438 
439 	*te = got_alloc_tree_entry_partial();
440 	if (*te == NULL)
441 		return got_error_from_errno();
442 
443 	*elen = strlen(buf) + 1;
444 	if (*elen > maxlen) {
445 		free(*te);
446 		*te = NULL;
447 		return got_error(GOT_ERR_BAD_OBJ_DATA);
448 	}
449 
450 	space = strchr(buf, ' ');
451 	if (space == NULL) {
452 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
453 		free(*te);
454 		*te = NULL;
455 		return err;
456 	}
457 	while (*p != ' ') {
458 		if (*p < '0' && *p > '7') {
459 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
460 			goto done;
461 		}
462 		(*te)->mode <<= 3;
463 		(*te)->mode |= *p - '0';
464 		p++;
465 	}
466 
467 	(*te)->name = strdup(space + 1);
468 	if (*elen > maxlen || maxlen - *elen < SHA1_DIGEST_LENGTH) {
469 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
470 		goto done;
471 	}
472 	buf += strlen(buf) + 1;
473 	memcpy((*te)->id->sha1, buf, SHA1_DIGEST_LENGTH);
474 	*elen += SHA1_DIGEST_LENGTH;
475 done:
476 	if (err) {
477 		got_object_tree_entry_close(*te);
478 		*te = NULL;
479 	}
480 	return err;
481 }
482 
483 const struct got_error *
484 got_object_parse_tree(struct got_tree_object **tree, uint8_t *buf, size_t len)
485 {
486 	const struct got_error *err;
487 	size_t remain = len;
488 
489 	*tree = calloc(1, sizeof(**tree));
490 	if (*tree == NULL)
491 		return got_error_from_errno();
492 
493 	SIMPLEQ_INIT(&(*tree)->entries.head);
494 
495 	while (remain > 0) {
496 		struct got_tree_entry *te;
497 		size_t elen;
498 
499 		err = parse_tree_entry(&te, &elen, buf, remain);
500 		if (err)
501 			return err;
502 		(*tree)->entries.nentries++;
503 		SIMPLEQ_INSERT_TAIL(&(*tree)->entries.head, te, entry);
504 		buf += elen;
505 		remain -= elen;
506 	}
507 
508 	if (remain != 0) {
509 		got_object_tree_close(*tree);
510 		return got_error(GOT_ERR_BAD_OBJ_DATA);
511 	}
512 
513 	return NULL;
514 }
515 
516 const struct got_error *
517 got_read_file_to_mem(uint8_t **outbuf, size_t *outlen, FILE *f)
518 {
519 	const struct got_error *err = NULL;
520 	static const size_t blocksize = 512;
521 	size_t n, total, remain;
522 	uint8_t *buf;
523 
524 	*outbuf = NULL;
525 	*outlen = 0;
526 
527 	buf = malloc(blocksize);
528 	if (buf == NULL)
529 		return got_error_from_errno();
530 
531 	remain = blocksize;
532 	total = 0;
533 	while (1) {
534 		if (remain == 0) {
535 			uint8_t *newbuf;
536 			newbuf = reallocarray(buf, 1, total + blocksize);
537 			if (newbuf == NULL) {
538 				err = got_error_from_errno();
539 				goto done;
540 			}
541 			buf = newbuf;
542 			remain += blocksize;
543 		}
544 		n = fread(buf + total, 1, remain, f);
545 		if (n == 0) {
546 			if (ferror(f)) {
547 				err = got_ferror(f, GOT_ERR_IO);
548 				goto done;
549 			}
550 			break; /* EOF */
551 		}
552 		remain -= n;
553 		total += n;
554 	};
555 
556 done:
557 	if (err == NULL) {
558 		*outbuf = buf;
559 		*outlen = total;
560 	} else
561 		free(buf);
562 	return err;
563 }