1 /*
2  * Copyright (c) 2018, 2019 Ori Bernstein <ori@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/uio.h>
20 #include <sys/socket.h>
21 #include <sys/wait.h>
22 #include <sys/resource.h>
23 #include <sys/socket.h>
24 
25 #include <errno.h>
26 #include <err.h>
27 #include <fcntl.h>
28 #include <stdio.h>
29 #include <stdlib.h>
30 #include <string.h>
31 #include <stdint.h>
32 #include <unistd.h>
33 #include <zlib.h>
34 #include <ctype.h>
35 #include <limits.h>
36 #include <time.h>
37 
38 #include "got_error.h"
39 #include "got_reference.h"
40 #include "got_repository.h"
41 #include "got_path.h"
42 #include "got_cancel.h"
43 #include "got_worktree.h"
44 #include "got_object.h"
45 #include "got_opentemp.h"
46 #include "got_fetch.h"
47 
48 #include "got_lib_delta.h"
49 #include "got_lib_inflate.h"
50 #include "got_lib_object.h"
51 #include "got_lib_object_parse.h"
52 #include "got_lib_object_create.h"
53 #include "got_lib_pack.h"
54 #include "got_lib_sha1.h"
55 #include "got_lib_privsep.h"
56 #include "got_lib_object_cache.h"
57 #include "got_lib_repository.h"
58 #include "got_lib_dial.h"
59 #include "got_lib_pkt.h"
60 
61 #ifndef nitems
62 #define nitems(_a)	(sizeof((_a)) / sizeof((_a)[0]))
63 #endif
64 
65 #ifndef ssizeof
66 #define ssizeof(_x) ((ssize_t)(sizeof(_x)))
67 #endif
68 
69 #ifndef MIN
70 #define	MIN(_a,_b) ((_a) < (_b) ? (_a) : (_b))
71 #endif
72 
73 const struct got_error *
74 got_fetch_connect(pid_t *fetchpid, int *fetchfd, const char *proto,
75     const char *host, const char *port, const char *server_path, int verbosity)
76 {
77 	const struct got_error *err = NULL;
78 
79 	*fetchpid = -1;
80 	*fetchfd = -1;
81 
82 	if (strcmp(proto, "ssh") == 0 || strcmp(proto, "git+ssh") == 0)
83 		err = got_dial_ssh(fetchpid, fetchfd, host, port,
84 		    server_path, GOT_DIAL_DIRECTION_FETCH, verbosity);
85 	else if (strcmp(proto, "git") == 0)
86 		err = got_dial_git(fetchfd, host, port, server_path,
87 		    GOT_DIAL_DIRECTION_FETCH);
88 	else if (strcmp(proto, "http") == 0 || strcmp(proto, "git+http") == 0)
89 		err = got_error_path(proto, GOT_ERR_NOT_IMPL);
90 	else
91 		err = got_error_path(proto, GOT_ERR_BAD_PROTO);
92 	return err;
93 }
94 
95 const struct got_error *
96 got_fetch_pack(struct got_object_id **pack_hash, struct got_pathlist_head *refs,
97     struct got_pathlist_head *symrefs, const char *remote_name,
98     int mirror_references, int fetch_all_branches,
99     struct got_pathlist_head *wanted_branches,
100     struct got_pathlist_head *wanted_refs, int list_refs_only, int verbosity,
101     int fetchfd, struct got_repository *repo,
102     got_fetch_progress_cb progress_cb, void *progress_arg)
103 {
104 	size_t i;
105 	int imsg_fetchfds[2], imsg_idxfds[2];
106 	int packfd = -1, npackfd = -1, idxfd = -1, nidxfd = -1, nfetchfd = -1;
107 	int tmpfds[3];
108 	int fetchstatus, idxstatus, done = 0;
109 	const struct got_error *err;
110 	struct imsgbuf fetchibuf, idxibuf;
111 	pid_t fetchpid, idxpid;
112 	char *tmppackpath = NULL, *tmpidxpath = NULL;
113 	char *packpath = NULL, *idxpath = NULL, *id_str = NULL;
114 	const char *repo_path = NULL;
115 	struct got_pathlist_head have_refs;
116 	struct got_pathlist_entry *pe;
117 	struct got_reflist_head my_refs;
118 	struct got_reflist_entry *re;
119 	off_t packfile_size = 0;
120 	struct got_packfile_hdr pack_hdr;
121 	uint32_t nobj = 0;
122 	char *path;
123 	char *progress = NULL;
124 
125 	*pack_hash = NULL;
126 
127 	/*
128 	 * Prevent fetching of references that won't make any
129 	 * sense outside of the remote repository's context.
130 	 */
131 	TAILQ_FOREACH(pe, wanted_refs, entry) {
132 		const char *refname = pe->path;
133 		if (strncmp(refname, "refs/got/", 9) == 0 ||
134 		    strncmp(refname, "got/", 4) == 0 ||
135 		    strncmp(refname, "refs/remotes/", 13) == 0 ||
136 		    strncmp(refname, "remotes/", 8) == 0)
137 			return got_error_path(refname, GOT_ERR_FETCH_BAD_REF);
138 	}
139 
140 	if (!list_refs_only)
141 		repo_path = got_repo_get_path_git_dir(repo);
142 
143 	for (i = 0; i < nitems(tmpfds); i++)
144 		tmpfds[i] = -1;
145 
146 	TAILQ_INIT(&have_refs);
147 	TAILQ_INIT(&my_refs);
148 
149 	if (!list_refs_only) {
150 		err = got_ref_list(&my_refs, repo, NULL,
151 		    got_ref_cmp_by_name, NULL);
152 		if (err)
153 			goto done;
154 	}
155 
156 	TAILQ_FOREACH(re, &my_refs, entry) {
157 		struct got_object_id *id;
158 		const char *refname;
159 
160 		if (got_ref_is_symbolic(re->ref))
161 			continue;
162 
163 		err = got_ref_resolve(&id, repo, re->ref);
164 		if (err)
165 			goto done;
166 		refname = strdup(got_ref_get_name(re->ref));
167 		if (refname == NULL) {
168 			err = got_error_from_errno("strdup");
169 			goto done;
170 		}
171 		err = got_pathlist_append(&have_refs, refname, id);
172 		if (err)
173 			goto done;
174 	}
175 
176 	if (list_refs_only) {
177 		packfd = got_opentempfd();
178 		if (packfd == -1) {
179 			err = got_error_from_errno("got_opentempfd");
180 			goto done;
181 		}
182 	} else {
183 		if (asprintf(&path, "%s/%s/fetching.pack",
184 		    repo_path, GOT_OBJECTS_PACK_DIR) == -1) {
185 			err = got_error_from_errno("asprintf");
186 			goto done;
187 		}
188 		err = got_opentemp_named_fd(&tmppackpath, &packfd, path);
189 		free(path);
190 		if (err)
191 			goto done;
192 		if (fchmod(packfd, GOT_DEFAULT_FILE_MODE) != 0) {
193 			err = got_error_from_errno2("fchmod", tmppackpath);
194 			goto done;
195 		}
196 	}
197 	if (list_refs_only) {
198 		idxfd = got_opentempfd();
199 		if (idxfd == -1) {
200 			err = got_error_from_errno("got_opentempfd");
201 			goto done;
202 		}
203 	} else {
204 		if (asprintf(&path, "%s/%s/fetching.idx",
205 		    repo_path, GOT_OBJECTS_PACK_DIR) == -1) {
206 			err = got_error_from_errno("asprintf");
207 			goto done;
208 		}
209 		err = got_opentemp_named_fd(&tmpidxpath, &idxfd, path);
210 		free(path);
211 		if (err)
212 			goto done;
213 		if (fchmod(idxfd, GOT_DEFAULT_FILE_MODE) != 0) {
214 			err = got_error_from_errno2("fchmod", tmpidxpath);
215 			goto done;
216 		}
217 	}
218 	nidxfd = dup(idxfd);
219 	if (nidxfd == -1) {
220 		err = got_error_from_errno("dup");
221 		goto done;
222 	}
223 
224 	for (i = 0; i < nitems(tmpfds); i++) {
225 		tmpfds[i] = got_opentempfd();
226 		if (tmpfds[i] == -1) {
227 			err = got_error_from_errno("got_opentempfd");
228 			goto done;
229 		}
230 	}
231 
232 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_fetchfds) == -1) {
233 		err = got_error_from_errno("socketpair");
234 		goto done;
235 	}
236 
237 	fetchpid = fork();
238 	if (fetchpid == -1) {
239 		err = got_error_from_errno("fork");
240 		goto done;
241 	} else if (fetchpid == 0){
242 		got_privsep_exec_child(imsg_fetchfds,
243 		    GOT_PATH_PROG_FETCH_PACK, tmppackpath);
244 	}
245 
246 	if (close(imsg_fetchfds[1]) == -1) {
247 		err = got_error_from_errno("close");
248 		goto done;
249 	}
250 	imsg_init(&fetchibuf, imsg_fetchfds[0]);
251 	nfetchfd = dup(fetchfd);
252 	if (nfetchfd == -1) {
253 		err = got_error_from_errno("dup");
254 		goto done;
255 	}
256 	err = got_privsep_send_fetch_req(&fetchibuf, nfetchfd, &have_refs,
257 	    fetch_all_branches, wanted_branches, wanted_refs,
258 	    list_refs_only, verbosity);
259 	if (err != NULL)
260 		goto done;
261 	nfetchfd = -1;
262 	npackfd = dup(packfd);
263 	if (npackfd == -1) {
264 		err = got_error_from_errno("dup");
265 		goto done;
266 	}
267 	err = got_privsep_send_fetch_outfd(&fetchibuf, npackfd);
268 	if (err != NULL)
269 		goto done;
270 	npackfd = -1;
271 
272 	packfile_size = 0;
273 	progress = calloc(GOT_PKT_MAX, 1);
274 	if (progress == NULL) {
275 		err = got_error_from_errno("calloc");
276 		goto done;
277 	}
278 
279 	*pack_hash = calloc(1, sizeof(**pack_hash));
280 	if (*pack_hash == NULL) {
281 		err = got_error_from_errno("calloc");
282 		goto done;
283 	}
284 
285 	while (!done) {
286 		struct got_object_id *id = NULL;
287 		char *refname = NULL;
288 		char *server_progress = NULL;
289 		off_t packfile_size_cur = 0;
290 
291 		err = got_privsep_recv_fetch_progress(&done,
292 		    &id, &refname, symrefs, &server_progress,
293 		    &packfile_size_cur, (*pack_hash)->sha1, &fetchibuf);
294 		if (err != NULL)
295 			goto done;
296 		/* Don't report size progress for an empty pack file. */
297 		if (packfile_size_cur <= ssizeof(pack_hdr) + SHA1_DIGEST_LENGTH)
298 			packfile_size_cur = 0;
299 		if (!done && refname && id) {
300 			err = got_pathlist_insert(NULL, refs, refname, id);
301 			if (err)
302 				goto done;
303 		} else if (!done && server_progress) {
304 			char *p;
305 			/*
306 			 * XXX git-daemon tends to send batched output with
307 			 * lines spanning separate packets. Buffer progress
308 			 * output until we see a CR or LF to avoid giving
309 			 * partial lines of progress output to the callback.
310 			 */
311 			if (strlcat(progress, server_progress,
312 			    GOT_PKT_MAX) >= GOT_PKT_MAX) {
313 				progress[0] = '\0'; /* discard */
314 				continue;
315 			}
316 			while ((p = strchr(progress, '\r')) != NULL ||
317 			    (p = strchr(progress, '\n')) != NULL) {
318 				char *s;
319 				size_t n;
320 				char c = *p;
321 				*p = '\0';
322 				if (asprintf(&s, "%s%s", progress,
323 				    c == '\n' ? "\n" : "") == -1) {
324 					err = got_error_from_errno("asprintf");
325 					goto done;
326 				}
327 				err = progress_cb(progress_arg, s,
328 				    packfile_size_cur, 0, 0, 0, 0);
329 				free(s);
330 				if (err)
331 					break;
332 				n = strlen(progress);
333 				if (n < GOT_PKT_MAX - 1) {
334 					memmove(progress, &progress[n + 1],
335 					    GOT_PKT_MAX - n - 1);
336 				} else
337 					progress[0] = '\0';
338 			}
339 			free(server_progress);
340 			if (err)
341 				goto done;
342 		} else if (!done && packfile_size_cur != packfile_size) {
343 			err = progress_cb(progress_arg, NULL,
344 			    packfile_size_cur, 0, 0, 0, 0);
345 			if (err)
346 				break;
347 			packfile_size = packfile_size_cur;
348 		}
349 	}
350 	if (close(imsg_fetchfds[0]) == -1) {
351 		err = got_error_from_errno("close");
352 		goto done;
353 	}
354 	if (waitpid(fetchpid, &fetchstatus, 0) == -1) {
355 		err = got_error_from_errno("waitpid");
356 		goto done;
357 	}
358 
359 	if (lseek(packfd, 0, SEEK_SET) == -1) {
360 		err = got_error_from_errno("lseek");
361 		goto done;
362 	}
363 
364 	/* If zero data was fetched without error we are already up-to-date. */
365 	if (packfile_size == 0) {
366 		free(*pack_hash);
367 		*pack_hash = NULL;
368 		goto done;
369 	} else if (packfile_size < ssizeof(pack_hdr) + SHA1_DIGEST_LENGTH) {
370 		err = got_error_msg(GOT_ERR_BAD_PACKFILE, "short pack file");
371 		goto done;
372 	} else {
373 		ssize_t n;
374 
375 		n = read(packfd, &pack_hdr, ssizeof(pack_hdr));
376 		if (n == -1) {
377 			err = got_error_from_errno("read");
378 			goto done;
379 		}
380 		if (n != ssizeof(pack_hdr)) {
381 			err = got_error(GOT_ERR_IO);
382 			goto done;
383 		}
384 		if (pack_hdr.signature != htobe32(GOT_PACKFILE_SIGNATURE)) {
385 			err = got_error_msg(GOT_ERR_BAD_PACKFILE,
386 			    "bad pack file signature");
387 			goto done;
388 		}
389 		if (pack_hdr.version != htobe32(GOT_PACKFILE_VERSION)) {
390 			err = got_error_msg(GOT_ERR_BAD_PACKFILE,
391 			    "bad pack file version");
392 			goto done;
393 		}
394 		nobj = be32toh(pack_hdr.nobjects);
395 		if (nobj == 0 &&
396 		    packfile_size > ssizeof(pack_hdr) + SHA1_DIGEST_LENGTH)
397 			return got_error_msg(GOT_ERR_BAD_PACKFILE,
398 			    "bad pack file with zero objects");
399 		if (nobj != 0 &&
400 		    packfile_size <= ssizeof(pack_hdr) + SHA1_DIGEST_LENGTH)
401 			return got_error_msg(GOT_ERR_BAD_PACKFILE,
402 			    "empty pack file with non-zero object count");
403 	}
404 
405 	/*
406 	 * If the pack file contains no objects, we may only need to update
407 	 * references in our repository. The caller will take care of that.
408 	 */
409 	if (nobj == 0) {
410 		free(*pack_hash);
411 		*pack_hash = NULL;
412 		goto done;
413 	}
414 
415 	if (lseek(packfd, 0, SEEK_SET) == -1) {
416 		err = got_error_from_errno("lseek");
417 		goto done;
418 	}
419  
420 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_idxfds) == -1) {
421 		err = got_error_from_errno("socketpair");
422 		goto done;
423 	}
424 	idxpid = fork();
425 	if (idxpid == -1) {
426 		err= got_error_from_errno("fork");
427 		goto done;
428 	} else if (idxpid == 0)
429 		got_privsep_exec_child(imsg_idxfds,
430 		    GOT_PATH_PROG_INDEX_PACK, tmppackpath);
431 	if (close(imsg_idxfds[1]) == -1) {
432 		err = got_error_from_errno("close");
433 		goto done;
434 	}
435 	imsg_init(&idxibuf, imsg_idxfds[0]);
436 
437 	npackfd = dup(packfd);
438 	if (npackfd == -1) {
439 		err = got_error_from_errno("dup");
440 		goto done;
441 	}
442 	err = got_privsep_send_index_pack_req(&idxibuf, (*pack_hash)->sha1,
443 	    npackfd);
444 	if (err != NULL)
445 		goto done;
446 	npackfd = -1;
447 	err = got_privsep_send_index_pack_outfd(&idxibuf, nidxfd);
448 	if (err != NULL)
449 		goto done;
450 	nidxfd = -1;
451 	for (i = 0; i < nitems(tmpfds); i++) {
452 		err = got_privsep_send_tmpfd(&idxibuf, tmpfds[i]);
453 		if (err != NULL)
454 			goto done;
455 		tmpfds[i] = -1;
456 	}
457 	done = 0;
458 	while (!done) {
459 		int nobj_total, nobj_indexed, nobj_loose, nobj_resolved;
460 
461 		err = got_privsep_recv_index_progress(&done, &nobj_total,
462 		    &nobj_indexed, &nobj_loose, &nobj_resolved,
463 		    &idxibuf);
464 		if (err != NULL)
465 			goto done;
466 		if (nobj_indexed != 0) {
467 			err = progress_cb(progress_arg, NULL,
468 			    packfile_size, nobj_total,
469 			    nobj_indexed, nobj_loose, nobj_resolved);
470 			if (err)
471 				break;
472 		}
473 	}
474 	if (close(imsg_idxfds[0]) == -1) {
475 		err = got_error_from_errno("close");
476 		goto done;
477 	}
478 	if (waitpid(idxpid, &idxstatus, 0) == -1) {
479 		err = got_error_from_errno("waitpid");
480 		goto done;
481 	}
482 
483 	err = got_object_id_str(&id_str, *pack_hash);
484 	if (err)
485 		goto done;
486 	if (asprintf(&packpath, "%s/%s/pack-%s.pack",
487 	    repo_path, GOT_OBJECTS_PACK_DIR, id_str) == -1) {
488 		err = got_error_from_errno("asprintf");
489 		goto done;
490 	}
491 
492 	if (asprintf(&idxpath, "%s/%s/pack-%s.idx",
493 	    repo_path, GOT_OBJECTS_PACK_DIR, id_str) == -1) {
494 		err = got_error_from_errno("asprintf");
495 		goto done;
496 	}
497 
498 	if (rename(tmppackpath, packpath) == -1) {
499 		err = got_error_from_errno3("rename", tmppackpath, packpath);
500 		goto done;
501 	}
502 	free(tmppackpath);
503 	tmppackpath = NULL;
504 	if (rename(tmpidxpath, idxpath) == -1) {
505 		err = got_error_from_errno3("rename", tmpidxpath, idxpath);
506 		goto done;
507 	}
508 	free(tmpidxpath);
509 	tmpidxpath = NULL;
510 
511 done:
512 	if (tmppackpath && unlink(tmppackpath) == -1 && err == NULL)
513 		err = got_error_from_errno2("unlink", tmppackpath);
514 	if (tmpidxpath && unlink(tmpidxpath) == -1 && err == NULL)
515 		err = got_error_from_errno2("unlink", tmpidxpath);
516 	if (nfetchfd != -1 && close(nfetchfd) == -1 && err == NULL)
517 		err = got_error_from_errno("close");
518 	if (npackfd != -1 && close(npackfd) == -1 && err == NULL)
519 		err = got_error_from_errno("close");
520 	if (packfd != -1 && close(packfd) == -1 && err == NULL)
521 		err = got_error_from_errno("close");
522 	if (idxfd != -1 && close(idxfd) == -1 && err == NULL)
523 		err = got_error_from_errno("close");
524 	for (i = 0; i < nitems(tmpfds); i++) {
525 		if (tmpfds[i] != -1 && close(tmpfds[i]) == -1 && err == NULL)
526 			err = got_error_from_errno("close");
527 	}
528 	free(tmppackpath);
529 	free(tmpidxpath);
530 	free(idxpath);
531 	free(packpath);
532 	free(progress);
533 
534 	TAILQ_FOREACH(pe, &have_refs, entry) {
535 		free((char *)pe->path);
536 		free(pe->data);
537 	}
538 	got_pathlist_free(&have_refs);
539 	got_ref_list_free(&my_refs);
540 	if (err) {
541 		free(*pack_hash);
542 		*pack_hash = NULL;
543 		TAILQ_FOREACH(pe, refs, entry) {
544 			free((void *)pe->path);
545 			free(pe->data);
546 		}
547 		got_pathlist_free(refs);
548 		TAILQ_FOREACH(pe, symrefs, entry) {
549 			free((void *)pe->path);
550 			free(pe->data);
551 		}
552 		got_pathlist_free(symrefs);
553 	}
554 	return err;
555 }