1 /*
2  * Copyright (c) 2024 Tobias Heider <me@tobhe.de>
3  * Copyright (c) 2022 Omar Polo <op@openbsd.org>
4  *
5  * Permission to use, copy, modify, and distribute this software for any
6  * purpose with or without fee is hereby granted, provided that the above
7  * copyright notice and this permission notice appear in all copies.
8  *
9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
16  */
17 
18 #include <sys/types.h>
19 #include <sys/queue.h>
20 #include <sys/socket.h>
21 
22 #include <err.h>
23 #include <errno.h>
24 #include <limits.h>
25 #include <netdb.h>
26 #include <poll.h>
27 #include <stdio.h>
28 #include <stdlib.h>
29 #include <string.h>
30 #include <tls.h>
31 #include <unistd.h>
32 
33 #include "got_error.h"
34 #include "got_path.h"
35 #include "got_version.h"
36 
37 #include "got_lib_pkt.h"
38 
39 #include "bufio.h"
40 
41 #define UPLOAD_PACK_ADV "application/x-git-upload-pack-advertisement"
42 #define UPLOAD_PACK_REQ "application/x-git-upload-pack-request"
43 #define UPLOAD_PACK_RES "application/x-git-upload-pack-result"
44 
45 #define	GOT_USERAGENT	"got/" GOT_VERSION_STR
46 #define MINIMUM(a, b)	((a) < (b) ? (a) : (b))
47 #define hasprfx(str, p)	(strncasecmp(str, p, strlen(p)) == 0)
48 
49 FILE *tmp;
50 
51 static int	verbose;
52 
53 static char *
54 bufio_getdelim_sync(struct bufio *bio, const char *nl, size_t *len)
55 {
56 	int	r;
57 
58 	do {
59 		r = bufio_read(bio);
60 		if (r == -1 && errno != EAGAIN)
61 			errx(1, "bufio_read: %s", bufio_io_err(bio));
62 	} while (r == -1 && errno == EAGAIN);
63 	return buf_getdelim(&bio->rbuf, nl, len);
64 }
65 
66 static size_t
67 bufio_drain_sync(struct bufio *bio, void *d, size_t len)
68 {
69 	int	r;
70 
71 	do {
72 		r = bufio_read(bio);
73 		if (r == -1 && errno != EAGAIN)
74 			errx(1, "bufio_read: %s", bufio_io_err(bio));
75 	} while (r == -1 && errno == EAGAIN);
76 	return bufio_drain(bio, d, len);
77 }
78 
79 static void
80 bufio_close_sync(struct bufio *bio)
81 {
82 	int	 r;
83 
84 	do {
85 		r = bufio_close(bio);
86 		if (r == -1 && errno == EAGAIN)
87 			errx(1, "bufio_read: %s", bufio_io_err(bio));
88 	} while (r == -1 && errno == EAGAIN);
89 }
90 
91 static long long
92 hexstrtonum(const char *str, long long min, long long max, const char **errstr)
93 {
94 	long long	 lval;
95 	char		*cp;
96 
97 	errno = 0;
98 	lval = strtoll(str, &cp, 16);
99 	if (*str == '\0' || *cp != '\0') {
100 		*errstr = "not a number";
101 		return 0;
102 	}
103 	if ((errno == ERANGE && (lval == LONG_MAX || lval == LONG_MIN)) ||
104 	    lval < min || lval > max) {
105 		*errstr = "out of range";
106 		return 0;
107 	}
108 
109 	*errstr = NULL;
110 	return lval;
111 }
112 
113 static int
114 dial(int https, const char *host, const char *port)
115 {
116 	struct addrinfo		 hints, *res, *res0;
117 	int			 error, saved_errno, fd = -1;
118 	const char		*cause = NULL;
119 
120 	memset(&hints, 0, sizeof(hints));
121 	hints.ai_family = AF_UNSPEC;
122 	hints.ai_socktype = SOCK_STREAM;
123 	error = getaddrinfo(host, port, &hints, &res0);
124 	if (error) {
125 		warnx("%s", gai_strerror(error));
126 		return -1;
127 	}
128 
129 	for (res = res0; res; res = res->ai_next) {
130 		fd = socket(res->ai_family, res->ai_socktype,
131 		    res->ai_protocol);
132 		if (fd == -1) {
133 			cause = "socket";
134 			continue;
135 		}
136 
137 		if (connect(fd, res->ai_addr, res->ai_addrlen) == 0)
138 			break;
139 
140 		cause = "connect";
141 		saved_errno = errno;
142 		close(fd);
143 		fd = -1;
144 		errno = saved_errno;
145 	}
146 	freeaddrinfo(res0);
147 
148 	if (fd == -1) {
149 		warn("%s", cause);
150 		return -1;
151 	}
152 
153 	return fd;
154 }
155 
156 static int
157 http_open(struct bufio *bio, int https, const char *method, const char *host, const char *port,
158     const char *path, const char *path_sufx, const char *query, const char *ctype)
159 {
160 	const char	*chdr = NULL, *te = "";
161 	char		*p, *req;
162 	int		 r;
163 	
164 	if (strcmp(method, "POST") == 0)
165 		te = "\r\nTransfer-Encoding: chunked\r\n";
166 
167 	if (ctype)
168 		chdr = "Content-Type: ";
169 
170 	r = asprintf(&p, "%s%s/%s%s%s", got_path_is_absolute(path) ? "" :"/",
171 	    path, path_sufx, query ? "?" : "", query ? query : "");
172 	if (r == -1)
173 		err(1, "asprintf");
174 
175 	r = asprintf(&req, "%s %s HTTP/1.1\r\n"
176 	    "Host: %s\r\n"
177 	    "Connection: close\r\n"
178 	    "User-agent: %s\r\n"
179 	    "%s%s%s\r\n",
180 	    method, p, host, GOT_USERAGENT,
181 	    chdr ? chdr : "", ctype ? ctype : "", te);
182 	if (r == -1)
183 		err(1, "asprintf");
184 	free(p);
185 
186 	if (verbose > 0)
187 		fprintf(stderr, "%s: request: %s\n", getprogname(), req);
188 	
189 
190 	r = bufio_compose(bio, req, r);
191 	if (r == -1)
192 		err(1, "bufio_compose_fmt");
193 	free(req);
194 
195 	do {
196 		r = bufio_write(bio);
197 		if (r == -1 && errno != EAGAIN)
198 			errx(1, "bufio_read: %s", bufio_io_err(bio));
199 	} while (bio->wbuf.len != 0);
200 
201 	return 0;
202 }
203 
204 static int
205 http_parse_reply(struct bufio *bio, int *chunked, const char *expected_ctype)
206 {
207 	char		*cp, *line;
208 	size_t		 linelen;
209 
210 	*chunked = 0;
211 
212 	line = bufio_getdelim_sync(bio, "\r\n", &linelen);
213 	if (line == NULL) {
214 		warnx("%s: bufio_getdelim_sync()", __func__);
215 		return -1;
216 	}
217 
218 	if (verbose > 0)
219 		fprintf(stderr, "%s: response: %s\n", getprogname(), line);
220 
221 	if ((cp = strchr(line, ' ')) == NULL) {
222 		warnx("malformed HTTP response");
223 		return -1;
224 	}
225 	cp++;
226 
227 	if (strncmp(cp, "200 ", 4) != 0) {
228 		warnx("malformed HTTP response");
229 		return -1;
230 	}
231 	buf_drain(&bio->rbuf, linelen);
232 
233 	while(1) {
234 		line = bufio_getdelim_sync(bio, "\r\n", &linelen);
235 		if (line == NULL) {
236 			warnx("%s: bufio_getdelim_sync()", __func__);
237 			return -1;
238 		}
239 		if (*line == '\0') {
240 			buf_drain(&bio->rbuf, linelen);
241 			break;
242 		}
243 
244 		if (hasprfx(line, "content-type:")) {
245 			cp = strchr(line, ':') + 1;
246 			cp += strspn(cp, " \t");
247 			cp[strcspn(cp, " \t")] = '\0';
248 			if (strcmp(cp, expected_ctype) != 0) {
249 				warnx("server not using the \"smart\" "
250 				    "HTTP protocol.");
251 				return -1;
252 			}
253 		}
254 		if (hasprfx(line, "transfer-encoding:")) {
255 			cp = strchr(line, ':') + 1;
256 			cp += strspn(cp, " \t");
257 			cp[strcspn(cp, " \t")] = '\0';
258 			if (strcmp(cp, "chunked") != 0) {
259 				warnx("unknown transfer-encoding");
260 				return -1;
261 			}
262 			*chunked = 1;
263 		}
264 		buf_drain(&bio->rbuf, linelen);
265 	}
266 
267 	return 0;
268 }
269 
270 static ssize_t
271 http_read(struct bufio *bio, int chunked, size_t *chunksz, char *buf, size_t bufsz)
272 {
273 	const char	*errstr;
274 	char		*line = NULL;
275 	size_t		 r;
276 	ssize_t		 ret = 0, linelen;
277 
278 	if (!chunked)
279 		return bufio_drain_sync(bio, buf, bufsz);
280 
281 	while (bufsz > 0) {
282 		if (*chunksz == 0) {
283 		again:
284 			line = bufio_getdelim_sync(bio, "\r\n", &linelen);
285 			if (line == NULL) {
286 				buf_drain(&bio->rbuf, linelen);
287 				break;
288 			}
289 			if (*line == '\0') {
290 				buf_drain(&bio->rbuf, linelen);
291 				goto again; /* was the CRLF after the chunk */
292 			}
293 
294 			*chunksz = hexstrtonum(line, 0, INT_MAX, &errstr);
295 			if (errstr != NULL) {
296 				warnx("invalid HTTP chunk: size is %s (%s)",
297 				    errstr, line);
298 				ret = -1;
299 				break;
300 			}
301 
302 			if (*chunksz == 0) {
303 				buf_drain(&bio->rbuf, linelen);
304 				break;
305 			}
306 			buf_drain(&bio->rbuf, linelen);
307 		}
308 
309 		r = bufio_drain_sync(bio, buf, MINIMUM(*chunksz, bufsz));
310 		if (r == 0) {
311 			break;
312 		}
313 
314 		ret += r;
315 		buf += r;
316 		bufsz -= r;
317 		*chunksz -= r;
318 	}
319 
320 	return ret;
321 }
322 
323 static int
324 http_chunk(struct bufio *bio, const void *buf, size_t len)
325 {
326 	int r;
327 
328 	if (bufio_compose_fmt(bio, "%zx\r\n", len) ||
329 	    bufio_compose(bio, buf, len) ||
330 	    bufio_compose(bio, "\r\n", 2))
331 		return 1;
332 
333 	do {
334 		r = bufio_write(bio);
335 		if (r == -1 && errno != EAGAIN)
336 			errx(1, "bufio_read: %s", bufio_io_err(bio));
337 	} while (bio->wbuf.len != 0);
338 
339 	return 0;
340 }
341 
342 static int
343 get_refs(int https, const char *host, const char *port, const char *path)
344 {
345 	struct bufio		 bio;
346 	char			 buf[GOT_PKT_MAX];
347 	const struct got_error	*e;
348 	size_t			 chunksz = 0;
349 	ssize_t			 r;
350 	int			 skip;
351 	int			 chunked;
352 	int			 sock;
353 	int			 ret = -1;
354 
355 	if ((sock = dial(https, host, port)) == -1)
356 		return -1;
357 
358 	if (bufio_init(&bio)) {
359 		warnx("bufio_init");
360 		goto err;
361 	}
362 	bufio_set_fd(&bio, sock);
363 	if (https && bufio_starttls(&bio, host, 0, NULL, 0, NULL, 0) == -1) {
364 		warnx("bufio_starttls");
365 		goto err;
366 	}
367 
368 	if (http_open(&bio, https, "GET", host, port, path, "info/refs",
369 	    "service=git-upload-pack", NULL) == -1)
370 		goto err;
371 
372 	/* Fetch the initial reference announcement from the server. */
373 	if (http_parse_reply(&bio, &chunked, UPLOAD_PACK_ADV) == -1)
374 		goto err;
375 
376 	/* skip first pack; why git over http is like this? */
377 	r = http_read(&bio, chunked, &chunksz, buf, 4);
378 	if (r <= 0)
379 		goto err;
380 
381 	e = got_pkt_readlen(&skip, buf, verbose);
382 	if (e) {
383 		warnx("%s", e->msg);
384 		goto err;
385 	}
386 
387 	/* TODO: validate it's # service=git-upload-pack\n */
388 	while (skip > 0) {
389 		r = http_read(&bio, chunked, &chunksz, buf,
390 		    MINIMUM(skip, sizeof(buf)));
391 		if (r <= 0)
392 			goto err;
393 		skip -= r;
394 	}
395 
396 	for (;;) {
397 		r = http_read(&bio, chunked, &chunksz, buf, sizeof(buf));
398 		if (r == -1)
399 			goto err;
400 
401 		if (r == 0)
402 			break;
403 
404 		fwrite(buf, 1, r, stdout);
405 	}
406 
407 	fflush(stdout);
408 	ret = 0;
409 err:
410 	bufio_close_sync(&bio);
411 	bufio_free(&bio);
412 	return ret;
413 }
414 
415 static int
416 upload_request(int https, const char *host, const char *port, const char *path,
417     FILE *in)
418 {
419 	struct bufio		 bio;
420 	char			 buf[GOT_PKT_MAX];
421 	const struct got_error	*e;
422 	ssize_t			 r;
423 	size_t			 chunksz = 0;
424 	int			 t;
425 	int			 chunked;
426 	int			 sock;
427 	int			 ret = -1;
428 
429 	if ((sock = dial(https, host, port)) == -1)
430 		return -1;
431 
432 	if (bufio_init(&bio)) {
433 		warnx("bufio_init");
434 		goto err;
435 	}
436 	bufio_set_fd(&bio, sock);
437 	if (https && bufio_starttls(&bio, host, 0, NULL, 0, NULL, 0) == -1) {
438 		warnx("bufio_starttls");
439 		goto err;
440 	}
441 #ifndef PROFILE
442 	/* TODO: can we push this upwards such that get_refs() is covered? */
443 	if (pledge("stdio", NULL) == -1)
444 		err(1, "pledge");
445 #endif
446 	if (http_open(&bio, https, "POST", host, port, path, "git-upload-pack",
447 	    NULL, UPLOAD_PACK_REQ) == -1)
448 		goto err;
449 
450 	/*
451 	 * Read have/want lines generated by got-fetch-pack and forward
452 	 * them to the server in the POST request body.
453 	 */
454 	for (;;) {
455 		r = fread(buf, 1, 4, in);
456 		if (r != 4)
457 			goto err;
458 
459 		e = got_pkt_readlen(&t, buf, verbose);
460 		if (e) {
461 			warnx("%s", e->msg);
462 			goto err;
463 		}
464 
465 		if (t == 0) {
466 			const char *flushpkt = "0000";
467 			if (http_chunk(&bio, flushpkt, strlen(flushpkt)))
468 				goto err;
469 			continue; /* got-fetch-pack will send "done" */
470 		}
471 
472 		if (t < 6) {
473 			warnx("pktline len is too small");
474 			goto err;
475 		}
476 
477 		r = fread(buf + 4, 1, t - 4, in);
478 		if (r != t - 4)
479 			goto err;
480 
481 		if (http_chunk(&bio, buf, t))
482 			goto err;
483 
484 		/*
485 		 * Once got-fetch-pack is done the server will
486 		 * send pack file data.
487 		 */
488 		if (t == 9 && strncmp(buf + 4, "done\n", 5) == 0) {
489 			if (http_chunk(&bio, NULL, 0))
490 				goto err;
491 			break;
492 		}
493 	}
494 
495 	if (http_parse_reply(&bio, &chunked, UPLOAD_PACK_RES) == -1)
496 		goto err;
497 
498 	/* Fetch pack file data from server. */
499 	for (;;) {
500 		r = http_read(&bio, chunked, &chunksz, buf, sizeof(buf));
501 		if (r == -1)
502 			goto err;
503 
504 		if (r == 0)
505 			break;
506 
507 		fwrite(buf, 1, r, stdout);
508 	}
509 
510 	ret = 0;
511 err:
512 	bufio_close_sync(&bio);
513 	bufio_free(&bio);
514 	return ret;
515 }
516 
517 static __dead void
518 usage(void)
519 {
520 	fprintf(stderr, "usage: %s [-qv] proto host port path\n",
521 	     getprogname());
522 	exit(1);
523 }
524 
525 int
526 main(int argc, char **argv)
527 {
528 	struct pollfd	 pfd;
529 	const char	*host, *port;
530 	char		*path;
531 	int		 https = 0;
532 	int		 ch;
533 
534 #ifndef PROFILE
535 	if (pledge("stdio rpath inet dns unveil", NULL) == -1)
536 		err(1, "pledge");
537 #endif
538 
539 	while ((ch = getopt(argc, argv, "qv")) != -1) {
540 		switch (ch) {
541 		case 'q':
542 			verbose = -1;
543 			break;
544 		case 'v':
545 			verbose++;
546 			break;
547 		default:
548 			usage();
549 		}
550 	}
551 	argc -= optind;
552 	argv += optind;
553 
554 	if (argc != 4)
555 		usage();
556 
557 	https = strcmp(argv[0], "https") == 0;
558 #ifndef PROFILE
559 	if (https) {
560 		if (unveil("/etc/ssl/cert.pem", "r") == -1)
561 			err(1, "unveil /etc/ssl/cert.pem");
562 	} else {
563 		/* drop "rpath" */
564 		if (pledge("stdio inet dns unveil", NULL) == -1)
565 			err(1, "pledge");
566 	}
567 #else
568 	if (unveil("gmon.out", "rwc") != 0)
569 		err(1, "unveil gmon.out");
570 #endif
571 	if (unveil(NULL, NULL) == -1)
572 		err(1, "unveil NULL");
573 
574 	host = argv[1];
575 	port = argv[2];
576 	path = argv[3];
577 	got_path_strip_trailing_slashes(path);
578 
579 	if (get_refs(https, host, port, path) == -1)
580 		errx(1, "failed to get refs");
581 
582 	pfd.fd = 0;
583 	pfd.events = POLLIN;
584 	if (poll(&pfd, 1, INFTIM) == -1)
585 		err(1, "poll");
586 
587 	if ((ch = fgetc(stdin)) == EOF)
588 		return 0;
589 
590 	ungetc(ch, stdin);
591 	if (upload_request(https, host, port, path, stdin) == -1) {
592 		fflush(tmp);
593 		errx(1, "failed to upload request");
594 	}
595 
596 	return 0;
597 }