1 /*
2  * Copyright (c) 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/queue.h>
19 #include <sys/uio.h>
20 #include <sys/time.h>
21 #include <sys/stat.h>
22 #include <sys/syslimits.h>
23 
24 #include <stdint.h>
25 #include <errno.h>
26 #include <imsg.h>
27 #include <limits.h>
28 #include <signal.h>
29 #include <stdio.h>
30 #include <stdlib.h>
31 #include <string.h>
32 #include <ctype.h>
33 #include <sha1.h>
34 #include <fcntl.h>
35 #include <zlib.h>
36 #include <err.h>
37 
38 #include "got_error.h"
39 #include "got_object.h"
40 
41 #include "got_lib_sha1.h"
42 #include "got_lib_delta.h"
43 #include "got_lib_inflate.h"
44 #include "got_lib_object.h"
45 #include "got_lib_object_parse.h"
46 #include "got_lib_privsep.h"
47 
48 #define GOT_PKTMAX	65536
49 
50 struct got_object *indexed;
51 static int chattygit;
52 static char *fetchbranch;
53 static char *upstream = "origin";
54 static char *packtmp = ".git/objects/pack/fetch.tmp";
55 static struct got_object_id zhash = {.sha1={0}};
56 
57 static char*
58 strip(char *ref)
59 {
60 	return ref;
61 }
62 
63 int
64 readn(int fd, void *buf, size_t n)
65 {
66 	ssize_t r, off;
67 
68 	off = 0;
69 	while (off != n) {
70 		r = read(fd, buf + off, n - off);
71 		if (r < 0)
72 			return -1;
73 		if (r == 0)
74 			return off;
75 		off += r;
76 	}
77 	return off;
78 }
79 
80 int
81 flushpkt(int fd)
82 {
83 	if(chattygit)
84 		fprintf(stderr, "writepkt: 0000\n");
85 	return write(fd, "0000", 4);
86 }
87 
88 
89 int
90 readpkt(int fd, char *buf, int nbuf)
91 {
92 	char len[5];
93 	char *e;
94 	int n, r;
95 
96 	if(readn(fd, len, 4) == -1){
97 		return -1;
98 	}
99 	len[4] = 0;
100 	n = strtol(len, &e, 16);
101 	if(n == 0){
102 		if(chattygit)
103 			fprintf(stderr, "readpkt: 0000\n");
104 		return 0;
105 	}
106 	if(e != len + 4 || n <= 4)
107 		err(1, "invalid packet line length");
108 	n  -= 4;
109 	if(n >= nbuf)
110 		err(1, "buffer too small");
111 	if((r = readn(fd, buf, n)) != n)
112 		return -1;
113 	buf[n] = 0;
114 	if(chattygit)
115 		fprintf(stderr, "readpkt: %s:\t%.*s\n", len, nbuf, buf);
116 	return n;
117 }
118 
119 int
120 writepkt(int fd, char *buf, int nbuf)
121 {
122 	char len[5];
123 	int i;
124 
125 	if (snprintf(len, sizeof(len), "%04x", nbuf + 4) >= sizeof(len))
126 		return -1;
127 	if(write(fd, len, 4) != 4)
128 		return -1;
129 	if(write(fd, buf, nbuf) != nbuf)
130 		return -1;
131 	if(chattygit){
132 		fprintf(stderr, "writepkt: %s:\t", len);
133 		fwrite(buf, 1, nbuf, stderr);
134 		for(i = 0; i < nbuf; i++){
135 			if(isprint(buf[i]))
136 				fputc(buf[i], stderr);
137 		}
138 	}
139 	return 0;
140 }
141 
142 
143 int
144 got_resolve_remote_ref(struct got_object_id *id, char *ref)
145 {
146 	char buf[128], *s;
147 	int r, f;
148 
149 	ref = strip(ref);
150 	if(!got_parse_sha1_digest(id->sha1, ref))
151 		return 0;
152 
153 	/* Slightly special handling: translate remote refs to local ones. */
154 	if (strcmp(ref, "HEAD") == 0) {
155 		if(snprintf(buf, sizeof(buf), ".git/HEAD") >= sizeof(buf))
156 			return -1;
157 	} else if(strstr(ref, "refs/heads") == ref) {
158 		ref += strlen("refs/heads");
159 		if(snprintf(buf, sizeof(buf),
160 		    ".git/refs/remotes/%s/%s", upstream, ref) >= sizeof(buf))
161 			return -1;
162 	} else if(strstr(ref, "refs/tags") == ref) {
163 		ref += strlen("refs/tags");
164 		if(snprintf(buf, sizeof(buf),
165 		    ".git/refs/tags/%s/%s", upstream, ref) >= sizeof(buf))
166 			return -1;
167 	} else {
168 		return -1;
169 	}
170 
171 	r = -1;
172 	s = strip(buf);
173 	if((f = open(s, O_RDONLY)) == -1)
174 		goto err;
175 	if(readn(f, buf, sizeof(buf)) < 40)
176 		goto err;
177 	if(!got_parse_sha1_digest(id->sha1, buf))
178 		goto err;
179 err:
180 	close(f);
181 	if(r == -1 && strstr(buf, "ref:") == buf)
182 		return got_resolve_remote_ref(id, buf + strlen("ref:"));
183 	return r;
184 }
185 
186 static int
187 got_check_pack_hash(int fd, size_t sz, uint8_t *hcomp)
188 {
189 	SHA1_CTX ctx;
190 	uint8_t hexpect[SHA1_DIGEST_LENGTH];
191 	char s1[SHA1_DIGEST_STRING_LENGTH + 1];
192 	char s2[SHA1_DIGEST_STRING_LENGTH + 1];
193 	uint8_t buf[32*1024];
194 	ssize_t n, r, nr;
195 
196 	if(sz < 28)
197 		return -1;
198 
199 	n = 0;
200 	SHA1Init(&ctx);
201 	while(n < sz - 20){
202 		nr = sizeof(buf);
203 		if(sz - n - 20 < sizeof(buf))
204 			nr = sz - n - 20;
205 		r = readn(fd, buf, nr);
206 		if(r != nr)
207 			return -1;
208 		SHA1Update(&ctx, buf, nr);
209 		n += r;
210 	}
211 	SHA1Final(hcomp, &ctx);
212 
213 	if(readn(fd, hexpect, sizeof(hexpect)) != sizeof(hexpect))
214 		errx(1, "truncated packfile");
215 	if(memcmp(hcomp, hexpect, SHA1_DIGEST_LENGTH) != 0){
216 		got_sha1_digest_to_str(hcomp, s1, sizeof(s1));
217 		got_sha1_digest_to_str(hexpect, s2, sizeof(s2));
218 		printf("hash mismatch %s != %s\n", s1, s2);
219 		return -1;
220 	}
221 	return 0;
222 }
223 
224 int
225 got_has_object(struct got_object_id *obj)
226 {
227 	return 0;
228 }
229 
230 /*static */int
231 got_make_pack_dir(char *path)
232 {
233 	char s[128];
234 	char *p;
235 
236 	if(snprintf(s, sizeof(s), "%s", path) >= sizeof(s))
237 		return -1;
238 	for(p=strchr(s+1, '/'); p; p=strchr(p+1, '/')){
239 		*p = 0;
240 		if (mkdir(s, 0755) == -1)
241 			if(errno != EEXIST)
242 				return -1;
243 		*p = '/';
244 	}
245 	return 0;
246 }
247 
248 static int
249 got_match_branch(char *br, char *pat)
250 {
251 	char name[128];
252 
253 	if(strstr(pat, "refs/heads") == pat) {
254 		if (snprintf(name, sizeof(name), "%s", pat) >= sizeof(name))
255 			return -1;
256 	} else if(strstr(pat, "heads")) {
257 		if (snprintf(name, sizeof(name), "refs/%s", pat) >= sizeof(name))
258 			return -1;
259 	} else {
260 		if (snprintf(name, sizeof(name), "refs/heads/%s", pat) >= sizeof(name))
261 			return -1;
262 	}
263 	return strcmp(br, name) == 0;
264 }
265 
266 static int
267 got_tokenize_refline(char *line, char **sp, size_t nsp)
268 {
269 	char *p;
270 	size_t i, j;
271 
272 	for (i = 0; i < nsp; i++) {
273 		while (isspace(*line))
274 			line++;
275 		p = line;
276 		while (*line != '\0' && (!isspace(*line) || i == nsp - 1))
277 			line++;
278 		sp[i] = strndup(p, line - p);
279 	}
280 	for (j = i; j < nsp; j++)
281 		sp[j] = NULL;
282 	return i;
283 }
284 
285 static int
286 got_fetch_pack(int fd, int packfd, char *packtmp, struct got_object_id *packid)
287 {
288 	char buf[GOT_PKTMAX], idxtmp[256], *sp[3];
289 	char hashstr[SHA1_DIGEST_STRING_LENGTH];
290 	struct got_object_id *have, *want;
291 	int nref, refsz;
292 	int i, n, req;
293 	off_t packsz;
294 
295 	nref = 0;
296 	refsz = 16;
297 	have = malloc(refsz * sizeof(have[0]));
298 	if(have == NULL)
299 		err(1, "malloc packs");
300 	want = malloc(refsz * sizeof(want[0]));
301 	if(want == NULL)
302 		err(1, "malloc packs");
303 	if(chattygit)
304 		fprintf(stderr, "starting fetch\n");
305 	while(1){
306 		n = readpkt(fd, buf, sizeof(buf));
307 		if(n == -1){
308 			err(1, "readpkt:");
309 			return -1;
310 		}
311 		if(n == 0)
312 			break;
313 		if(strncmp(buf, "ERR ", 4) == 0)
314 			err(1, "%s", buf + 4);
315 		if(got_tokenize_refline(buf, sp, 3) <= 2)
316 			err(1, "malformed ref line");
317 		if(strstr(sp[1], "^{}"))
318 			continue;
319 		if(fetchbranch && !got_match_branch(sp[1], fetchbranch))
320 			continue;
321 		if(refsz == nref + 1){
322 			refsz *= 2;
323 			have = realloc(have, refsz * sizeof(have[0]));
324 			if (have == NULL)
325 				err(1, "realloc have");
326 			want = realloc(want, refsz * sizeof(want[0]));
327 			if (want == NULL)
328 				err(1, "realloc want");
329 		}
330 		if(!got_parse_sha1_digest(want[nref].sha1, sp[0]))
331 			errx(1, "invalid hash %s", sp[0]);
332 		if (got_resolve_remote_ref(&have[nref], sp[1]) == -1)
333 			memset(&have[nref], 0, sizeof(have[nref]));
334 		// TODO: send IMSG with progress.
335 		if (chattygit)
336 			fprintf(stderr, "remote %s\n", sp[1]);
337 		nref++;
338 	}
339 
340 	req = 0;
341 	for(i = 0; i < nref; i++){
342 		if (got_object_id_cmp(&have[i], &want[i]) == 0)
343 			continue;
344 		if (got_has_object(&want[i]))
345 			continue;
346 		got_sha1_digest_to_str(want[i].sha1, hashstr, sizeof(hashstr));
347 		n = snprintf(buf, sizeof(buf), "want %s\n", hashstr);
348 		if (n >= sizeof(buf))
349 			errx(1, "undersized buffer");
350 		if(writepkt(fd, buf, n) == -1)
351 			errx(1, "could not send want");
352 		req = 1;
353 	}
354 	flushpkt(fd);
355 	for(i = 0; i < nref; i++){
356 		if (got_object_id_cmp(&have[i], &zhash) == 0)
357 			continue;
358 		got_sha1_digest_to_str(want[i].sha1, hashstr, sizeof(hashstr));
359 		n = snprintf(buf, sizeof(buf), "have %s\n", hashstr);
360 		if (n >= sizeof(buf))
361 			errx(1, "undersized buffer");
362 		if (writepkt(fd, buf, n + 1) == -1)
363 			errx(1, "could not send have");
364 	}
365 	if(!req){
366 		fprintf(stderr, "up to date\n");
367 		flushpkt(fd);
368 	}
369 	n = snprintf(buf, sizeof(buf), "done\n");
370 	if(writepkt(fd, buf, n) == -1)
371 		errx(1, "lost connection write");
372 	if(!req)
373 		return 0;
374 
375 	if((n = readpkt(fd, buf, sizeof(buf))) == -1)
376 		errx(1, "lost connection read");
377 	buf[n] = 0;
378 
379 	fprintf(stderr, "fetching...\n");
380 	packsz = 0;
381 	while(1){
382 		n = readn(fd, buf, sizeof buf);
383 		if(n == 0)
384 			break;
385 		if(n == -1 || write(packfd, buf, n) != n)
386 			err(1, "could not fetch packfile:");
387 		packsz += n;
388 	}
389 	if(lseek(packfd, 0, SEEK_SET) == -1)
390 		err(1, "packfile seek:");
391 	if(got_check_pack_hash(packfd, packsz, packid->sha1) == -1)
392 		errx(1, "corrupt packfile");
393 	close(packfd);
394 	n = strlen(packtmp) - strlen(".tmp");
395 	memcpy(idxtmp, packtmp, n);
396 	memcpy(idxtmp + n, ".idx", strlen(".idx") + 1);
397 	return 0;
398 }
399 
400 
401 int
402 main(int argc, char **argv)
403 {
404 	const struct got_error *err = NULL;
405 	int fetchfd, packfd;
406 	struct got_object_id packid;
407 	struct imsgbuf ibuf;
408 	struct imsg imsg;
409 
410 	if(getenv("GOT_DEBUG") != NULL){
411 		fprintf(stderr, "fetch-pack being chatty!\n");
412 		chattygit = 1;
413 	}
414 
415 	imsg_init(&ibuf, GOT_IMSG_FD_CHILD);
416 	if((err = got_privsep_recv_imsg(&imsg, &ibuf, 0)) != 0) {
417 		if (err->code == GOT_ERR_PRIVSEP_PIPE)
418 			err = NULL;
419 		goto done;
420 	}
421 	if (imsg.hdr.type == GOT_IMSG_STOP)
422 		goto done;
423 	if (imsg.hdr.type != GOT_IMSG_FETCH_REQUEST) {
424 		err = got_error(GOT_ERR_PRIVSEP_MSG);
425 		goto done;
426 	}
427 	if (imsg.hdr.len - IMSG_HEADER_SIZE != 0) {
428 		err = got_error(GOT_ERR_PRIVSEP_LEN);
429 		goto done;
430 	}
431 	fetchfd = imsg.fd;
432 
433 	if((err = got_privsep_recv_imsg(&imsg, &ibuf, 0)) != 0) {
434 		if (err->code == GOT_ERR_PRIVSEP_PIPE)
435 			err = NULL;
436 		goto done;
437 	}
438 	if (imsg.hdr.type == GOT_IMSG_STOP)
439 		goto done;
440 	if (imsg.hdr.type != GOT_IMSG_TMPFD) {
441 		err = got_error(GOT_ERR_PRIVSEP_MSG);
442 		goto done;
443 	}
444 	if (imsg.hdr.len - IMSG_HEADER_SIZE != 0) {
445 		err = got_error(GOT_ERR_PRIVSEP_LEN);
446 		goto done;
447 	}
448 	packfd = imsg.fd;
449 
450 	if(got_fetch_pack(fetchfd, packfd, packtmp, &packid) == -1) {
451 		err = got_error(GOT_ERR_FETCH_FAILED);
452 		goto done;
453 	}
454 done:
455 	if(err != NULL)
456 		got_privsep_send_error(&ibuf, err);
457 	else
458 		err = got_privsep_send_fetch_done(&ibuf, packid);
459 	if(err != NULL) {
460 		fprintf(stderr, "%s: %s\n", getprogname(), err->msg);
461 		got_privsep_send_error(&ibuf, err);
462 	}
463 
464 	exit(0);
465 }