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/queue.h>
20 #include <sys/uio.h>
21 #include <sys/socket.h>
22 #include <sys/wait.h>
23 #include <sys/syslimits.h>
24 #include <sys/resource.h>
25 #include <sys/socket.h>
26 
27 #include <errno.h>
28 #include <err.h>
29 #include <fcntl.h>
30 #include <stdio.h>
31 #include <stdlib.h>
32 #include <string.h>
33 #include <stdint.h>
34 #include <sha1.h>
35 #include <zlib.h>
36 #include <ctype.h>
37 #include <limits.h>
38 #include <imsg.h>
39 #include <time.h>
40 #include <uuid.h>
41 #include <netdb.h>
42 #include <netinet/in.h>
43 
44 #include "got_error.h"
45 #include "got_reference.h"
46 #include "got_repository.h"
47 #include "got_path.h"
48 #include "got_cancel.h"
49 #include "got_worktree.h"
50 #include "got_object.h"
51 #include "got_opentemp.h"
52 #include "got_fetch.h"
53 
54 #include "got_lib_delta.h"
55 #include "got_lib_inflate.h"
56 #include "got_lib_object.h"
57 #include "got_lib_object_parse.h"
58 #include "got_lib_object_create.h"
59 #include "got_lib_pack.h"
60 #include "got_lib_sha1.h"
61 #include "got_lib_privsep.h"
62 #include "got_lib_object_cache.h"
63 #include "got_lib_repository.h"
64 
65 #define GOT_PROTOMAX	64
66 #define GOT_HOSTMAX	256
67 #define GOT_PATHMAX	512
68 #define GOT_REPOMAX	256
69 #define GOT_PORTMAX	16
70 #define GOT_URIMAX	1024
71 
72 static int
73 hassuffix(char *base, char *suf)
74 {
75 	int nb, ns;
76 
77 	nb = strlen(base);
78 	ns = strlen(suf);
79 	if (ns <= nb && strcmp(base + (nb - ns), suf) == 0)
80 		return 1;
81 	return 0;
82 }
83 
84 static const struct got_error *
85 dial_ssh(int *fetchfd, const char *host, const char *port, const char *path,
86     const char *direction)
87 {
88 	const struct got_error *error = NULL;
89 	int pid, pfd[2];
90 	char cmd[64];
91 
92 	*fetchfd = -1;
93 
94 	if (pipe(pfd) == -1)
95 		return got_error_from_errno("pipe");
96 
97 	pid = fork();
98 	if (pid == -1) {
99 		error = got_error_from_errno("fork");
100 		close(pfd[0]);
101 		close(pfd[1]);
102 		return error;
103 	} else if (pid == 0) {
104 		int n;
105 		close(pfd[1]);
106 		dup2(pfd[0], 0);
107 		dup2(pfd[0], 1);
108 		n = snprintf(cmd, sizeof(cmd), "git-%s-pack", direction);
109 		if (n < 0 || n >= sizeof(cmd))
110 			err(1, "snprintf");
111 		if (execlp("ssh", "ssh", host, cmd, path, NULL) == -1)
112 			err(1, "execlp");
113 		abort(); /* not reached */
114 	} else {
115 		close(pfd[0]);
116 		*fetchfd = pfd[1];
117 		return NULL;
118 	}
119 }
120 
121 static const struct got_error *
122 dial_git(int *fetchfd, const char *host, const char *port, const char *path,
123     const char *direction)
124 {
125 	const struct got_error *err = NULL;
126 	struct addrinfo hints, *servinfo, *p;
127 	char *cmd = NULL, *pkt = NULL;
128 	int fd = -1, totlen, r, eaicode;
129 
130 	*fetchfd = -1;
131 
132 	memset(&hints, 0, sizeof hints);
133 	hints.ai_family = AF_UNSPEC;
134 	hints.ai_socktype = SOCK_STREAM;
135 	eaicode = getaddrinfo(host, port, &hints, &servinfo);
136 	if (eaicode)
137 		return got_error_msg(GOT_ERR_ADDRINFO, gai_strerror(eaicode));
138 
139 	for (p = servinfo; p != NULL; p = p->ai_next) {
140 		if ((fd = socket(p->ai_family, p->ai_socktype,
141 		    p->ai_protocol)) == -1)
142 			continue;
143 		if (connect(fd, p->ai_addr, p->ai_addrlen) == 0)
144 			break;
145 		err = got_error_from_errno("connect");
146 		close(fd);
147 	}
148 	if (p == NULL)
149 		goto done;
150 
151 	if (asprintf(&cmd, "git-%s-pack %s", direction, path) == -1) {
152 		err = got_error_from_errno("asprintf");
153 		goto done;
154 	}
155 	totlen = 4 + strlen(cmd) + 1 + strlen("host=") + strlen(host) + 1;
156 	if (asprintf(&pkt, "%04x%s", totlen, cmd) == -1) {
157 		err = got_error_from_errno("asprintf");
158 		goto done;
159 	}
160 	r = write(fd, pkt, strlen(pkt) + 1);
161 	if (r == -1) {
162 		err = got_error_from_errno("write");
163 		goto done;
164 	}
165 	if (asprintf(&pkt, "host=%s", host) == -1) {
166 		err = got_error_from_errno("asprintf");
167 		goto done;
168 	}
169 	r = write(fd, pkt, strlen(pkt) + 1);
170 	if (r == -1) {
171 		err = got_error_from_errno("write");
172 		goto done;
173 	}
174 done:
175 	free(cmd);
176 	free(pkt);
177 	if (err) {
178 		if (fd != -1)
179 			close(fd);
180 	} else
181 		*fetchfd = fd;
182 	return err;
183 }
184 
185 const struct got_error *
186 got_fetch_connect(int *fetchfd, const char *proto, const char *host,
187     const char *port, const char *server_path)
188 {
189 	const struct got_error *err = NULL;
190 
191 	*fetchfd = -1;
192 
193 	if (strcmp(proto, "ssh") == 0 || strcmp(proto, "git+ssh") == 0)
194 		err = dial_ssh(fetchfd, host, port, server_path, "upload");
195 	else if (strcmp(proto, "git") == 0)
196 		err = dial_git(fetchfd, host, port, server_path, "upload");
197 	else if (strcmp(proto, "http") == 0 || strcmp(proto, "git+http") == 0)
198 		err = got_error_path(proto, GOT_ERR_NOT_IMPL);
199 	else
200 		err = got_error_path(proto, GOT_ERR_BAD_PROTO);
201 	return err;
202 }
203 
204 const struct got_error *
205 got_fetch_parse_uri(char **proto, char **host, char **port,
206     char **server_path, char **repo_name, const char *uri)
207 {
208 	const struct got_error *err = NULL;
209 	char *s, *p, *q;
210 	int n, hasport;
211 
212 	*proto = *host = *port = *server_path = *repo_name = NULL;
213 
214 	p = strstr(uri, "://");
215 	if (!p) {
216 		return got_error(GOT_ERR_PARSE_URI);
217 	}
218 	*proto = strndup(uri, p - uri);
219 	if (proto == NULL) {
220 		err = got_error_from_errno("strndup");
221 		goto done;
222 	}
223 
224 	hasport = (strcmp(*proto, "git") == 0 ||
225 	    strstr(*proto, "http") == *proto);
226 	s = p + 3;
227 	p = NULL;
228 	if (!hasport) {
229 		p = strstr(s, ":");
230 		if (p != NULL)
231 			p++;
232 	}
233 	if (p == NULL)
234 		p = strstr(s, "/");
235 	if (p == NULL || strlen(p) == 1) {
236 		err = got_error(GOT_ERR_PARSE_URI);
237 		goto done;
238 	}
239 
240 	q = memchr(s, ':', p - s);
241 	if (q) {
242 		*host = strndup(s, q - s);
243 		if (*host == NULL) {
244 			err = got_error_from_errno("strndup");
245 			goto done;
246 		}
247 		*port = strndup(q + 1, p - (q + 1));
248 		if (*port == NULL) {
249 			err = got_error_from_errno("strndup");
250 			goto done;
251 		}
252 	} else {
253 		*host = strndup(s, p - s);
254 		if (*host == NULL) {
255 			err = got_error_from_errno("strndup");
256 			goto done;
257 		}
258 		if (asprintf(port, "%u", GOT_DEFAULT_GIT_PORT) == -1) {
259 			err = got_error_from_errno("asprintf");
260 			goto done;
261 		}
262 	}
263 
264 	*server_path = strdup(p);
265 	if (*server_path == NULL) {
266 		err = got_error_from_errno("strdup");
267 		goto done;
268 	}
269 
270 	p = strrchr(p, '/') + 1;
271 	if (!p || strlen(p) == 0) {
272 		//werrstr("missing repository in uri");
273 		err = got_error(GOT_ERR_PARSE_URI);
274 		goto done;
275 	}
276 	n = strlen(p);
277 	if (hassuffix(p, ".git"))
278 		n -= 4;
279 	*repo_name = strndup(p, (p + n) - p);
280 	if (*repo_name == NULL) {
281 		err = got_error_from_errno("strndup");
282 		goto done;
283 	}
284 done:
285 	if (err) {
286 		free(*proto);
287 		*proto = NULL;
288 		free(*host);
289 		*host = NULL;
290 		free(*port);
291 		*port = NULL;
292 		free(*server_path);
293 		*server_path = NULL;
294 		free(*repo_name);
295 		*repo_name = NULL;
296 	}
297 	return err;
298 }
299 
300 const struct got_error*
301 got_fetch_pack(struct got_object_id **pack_hash, struct got_pathlist_head *refs,
302     struct got_pathlist_head *symrefs, int fetchfd, struct got_repository *repo)
303 {
304 	int imsg_fetchfds[2], imsg_idxfds[2];
305 	int packfd = -1, npackfd = -1, idxfd = -1, nidxfd = -1, nfetchfd = -1;
306 	int status, done = 0;
307 	const struct got_error *err;
308 	struct imsgbuf ibuf;
309 	pid_t pid;
310 	char *tmppackpath = NULL, *tmpidxpath = NULL;
311 	char *packpath = NULL, *idxpath = NULL, *id_str = NULL;
312 	const char *repo_path = got_repo_get_path(repo);
313 	struct got_pathlist_head have_refs;
314 	struct got_pathlist_entry *pe;
315 	char *path;
316 
317 	*pack_hash = NULL;
318 
319 	TAILQ_INIT(&have_refs);
320 
321 	if (asprintf(&path, "%s/%s/fetching.pack",
322 	    repo_path, GOT_OBJECTS_PACK_DIR) == -1) {
323 		err = got_error_from_errno("asprintf");
324 		goto done;
325 	}
326 	err = got_opentemp_named_fd(&tmppackpath, &packfd, path);
327 	free(path);
328 	if (err)
329 		goto done;
330 	npackfd = dup(packfd);
331 	if (npackfd == -1) {
332 		err = got_error_from_errno("dup");
333 		goto done;
334 	}
335 	if (asprintf(&path, "%s/%s/fetching.idx",
336 	    repo_path, GOT_OBJECTS_PACK_DIR) == -1) {
337 		err = got_error_from_errno("asprintf");
338 		goto done;
339 	}
340 	err = got_opentemp_named_fd(&tmpidxpath, &idxfd, path);
341 	free(path);
342 	if (err)
343 		goto done;
344 	nidxfd = dup(idxfd);
345 	if (nidxfd == -1) {
346 		err = got_error_from_errno("dup");
347 		goto done;
348 	}
349 
350 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_fetchfds) == -1) {
351 		err = got_error_from_errno("socketpair");
352 		goto done;
353 	}
354 
355 	pid = fork();
356 	if (pid == -1) {
357 		err = got_error_from_errno("fork");
358 		goto done;
359 	} else if (pid == 0){
360 		got_privsep_exec_child(imsg_fetchfds,
361 		    GOT_PATH_PROG_FETCH_PACK, ".");
362 	}
363 
364 	if (close(imsg_fetchfds[1]) != 0) {
365 		err = got_error_from_errno("close");
366 		goto done;
367 	}
368 	imsg_init(&ibuf, imsg_fetchfds[0]);
369 	nfetchfd = dup(fetchfd);
370 	if (nfetchfd == -1) {
371 		err = got_error_from_errno("dup");
372 		goto done;
373 	}
374 	err = got_privsep_send_fetch_req(&ibuf, nfetchfd, &have_refs);
375 	if (err != NULL)
376 		goto done;
377 	nfetchfd = -1;
378 	err = got_privsep_send_tmpfd(&ibuf, npackfd);
379 	if (err != NULL)
380 		goto done;
381 	npackfd = dup(packfd);
382 	if (npackfd == -1) {
383 		err = got_error_from_errno("dup");
384 		goto done;
385 	}
386 
387 	while (!done) {
388 		struct got_object_id *id = NULL;
389 		char *refname = NULL;
390 
391 		err = got_privsep_recv_fetch_progress(&done,
392 		    &id, &refname, symrefs, &ibuf);
393 		if (err != NULL)
394 			goto done;
395 		if (done)
396 			*pack_hash = id;
397 		else if (refname && id) {
398 			err = got_pathlist_append(refs, refname, id);
399 			if (err)
400 				goto done;
401 		}
402 		/* TODO remote status / download progress callback */
403 	}
404 	if (waitpid(pid, &status, 0) == -1) {
405 		err = got_error_from_errno("waitpid");
406 		goto done;
407 	}
408 
409 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_idxfds) == -1) {
410 		err = got_error_from_errno("socketpair");
411 		goto done;
412 	}
413 	pid = fork();
414 	if (pid == -1) {
415 		err= got_error_from_errno("fork");
416 		goto done;
417 	} else if (pid == 0)
418 		got_privsep_exec_child(imsg_idxfds,
419 		    GOT_PATH_PROG_INDEX_PACK, ".");
420 	if (close(imsg_idxfds[1]) != 0) {
421 		err = got_error_from_errno("close");
422 		goto done;
423 	}
424 	imsg_init(&ibuf, imsg_idxfds[0]);
425 
426 	err = got_privsep_send_index_pack_req(&ibuf, npackfd, *pack_hash);
427 	if (err != NULL)
428 		goto done;
429 	npackfd = -1;
430 	err = got_privsep_send_tmpfd(&ibuf, nidxfd);
431 	if (err != NULL)
432 		goto done;
433 	nidxfd = -1;
434 	err = got_privsep_wait_index_pack_done(&ibuf);
435 	if (err != NULL)
436 		goto done;
437 	imsg_clear(&ibuf);
438 	if (close(imsg_idxfds[0]) == -1) {
439 		err = got_error_from_errno("close");
440 		goto done;
441 	}
442 	if (waitpid(pid, &status, 0) == -1) {
443 		err = got_error_from_errno("waitpid");
444 		goto done;
445 	}
446 
447 	err = got_object_id_str(&id_str, *pack_hash);
448 	if (err)
449 		goto done;
450 	if (asprintf(&packpath, "%s/%s/pack-%s.pack",
451 	    repo_path, GOT_OBJECTS_PACK_DIR, id_str) == -1) {
452 		err = got_error_from_errno("asprintf");
453 		goto done;
454 	}
455 
456 	if (asprintf(&idxpath, "%s/%s/pack-%s.idx",
457 	    repo_path, GOT_OBJECTS_PACK_DIR, id_str) == -1) {
458 		err = got_error_from_errno("asprintf");
459 		goto done;
460 	}
461 
462 	if (rename(tmppackpath, packpath) == -1) {
463 		err = got_error_from_errno3("rename", tmppackpath, packpath);
464 		goto done;
465 	}
466 	if (rename(tmpidxpath, idxpath) == -1) {
467 		err = got_error_from_errno3("rename", tmpidxpath, idxpath);
468 		goto done;
469 	}
470 
471 done:
472 	if (nfetchfd != -1 && close(nfetchfd) == -1 && err == NULL)
473 		err = got_error_from_errno("close");
474 	if (npackfd != -1 && close(npackfd) == -1 && err == NULL)
475 		err = got_error_from_errno("close");
476 	if (packfd != -1 && close(packfd) == -1 && err == NULL)
477 		err = got_error_from_errno("close");
478 	if (idxfd != -1 && close(idxfd) == -1 && err == NULL)
479 		err = got_error_from_errno("close");
480 	free(tmppackpath);
481 	free(tmpidxpath);
482 	free(idxpath);
483 	free(packpath);
484 
485 	if (err) {
486 		free(*pack_hash);
487 		*pack_hash = NULL;
488 		TAILQ_FOREACH(pe, refs, entry) {
489 			free((void *)pe->path);
490 			free(pe->data);
491 		}
492 		got_pathlist_free(refs);
493 		TAILQ_FOREACH(pe, symrefs, entry) {
494 			free((void *)pe->path);
495 			free(pe->data);
496 		}
497 		got_pathlist_free(symrefs);
498 	}
499 	return err;
500 }