2 * Copyright (c) 2018, 2019 Ori Bernstein <ori@openbsd.org>
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.
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.
16 #include "got_compat.h"
18 #include <sys/types.h>
20 #include <sys/queue.h>
22 #include <sys/socket.h>
24 #include <sys/resource.h>
25 #include <sys/socket.h>
40 #include "got_error.h"
41 #include "got_reference.h"
42 #include "got_repository.h"
44 #include "got_cancel.h"
45 #include "got_worktree.h"
46 #include "got_object.h"
47 #include "got_opentemp.h"
48 #include "got_fetch.h"
50 #include "got_lib_delta.h"
51 #include "got_lib_inflate.h"
52 #include "got_lib_object.h"
53 #include "got_lib_object_parse.h"
54 #include "got_lib_object_create.h"
55 #include "got_lib_pack.h"
56 #include "got_lib_hash.h"
57 #include "got_lib_privsep.h"
58 #include "got_lib_object_cache.h"
59 #include "got_lib_repository.h"
60 #include "got_lib_dial.h"
61 #include "got_lib_pkt.h"
64 #define nitems(_a) (sizeof((_a)) / sizeof((_a)[0]))
68 #define ssizeof(_x) ((ssize_t)(sizeof(_x)))
72 #define MIN(_a,_b) ((_a) < (_b) ? (_a) : (_b))
75 const struct got_error *
76 got_fetch_connect(pid_t *fetchpid, int *fetchfd, const char *proto,
77 const char *host, const char *port, const char *server_path, int verbosity)
79 const struct got_error *err = NULL;
84 if (strcmp(proto, "ssh") == 0 || strcmp(proto, "git+ssh") == 0)
85 err = got_dial_ssh(fetchpid, fetchfd, host, port,
86 server_path, GOT_DIAL_CMD_FETCH, verbosity);
87 else if (strcmp(proto, "git") == 0)
88 err = got_dial_git(fetchfd, host, port, server_path,
90 else if (strcmp(proto, "http") == 0 || strcmp(proto, "git+http") == 0)
91 err = got_error_path(proto, GOT_ERR_NOT_IMPL);
93 err = got_error_path(proto, GOT_ERR_BAD_PROTO);
97 const struct got_error *
98 got_fetch_pack(struct got_object_id **pack_hash, struct got_pathlist_head *refs,
99 struct got_pathlist_head *symrefs, const char *remote_name,
100 int mirror_references, int fetch_all_branches,
101 struct got_pathlist_head *wanted_branches,
102 struct got_pathlist_head *wanted_refs, int list_refs_only, int verbosity,
103 int fetchfd, struct got_repository *repo, const char *worktree_refname,
104 const char *remote_head, int no_head, got_fetch_progress_cb progress_cb,
108 int imsg_fetchfds[2], imsg_idxfds[2];
109 int packfd = -1, npackfd = -1, idxfd = -1, nidxfd = -1, nfetchfd = -1;
111 int fetchstatus, idxstatus, done = 0;
112 const struct got_error *err;
113 struct imsgbuf fetchibuf, idxibuf;
114 pid_t fetchpid, idxpid;
115 char *tmppackpath = NULL, *tmpidxpath = NULL;
116 char *packpath = NULL, *idxpath = NULL, *id_str = NULL;
117 const char *repo_path = NULL;
118 struct got_pathlist_head have_refs;
119 struct got_pathlist_entry *pe;
120 struct got_reflist_head my_refs;
121 struct got_reflist_entry *re;
122 off_t packfile_size = 0;
123 struct got_packfile_hdr pack_hdr;
126 char *progress = NULL;
131 * Prevent fetching of references that won't make any
132 * sense outside of the remote repository's context.
134 TAILQ_FOREACH(pe, wanted_refs, entry) {
135 const char *refname = pe->path;
136 if (strncmp(refname, "refs/got/", 9) == 0 ||
137 strncmp(refname, "got/", 4) == 0 ||
138 strncmp(refname, "refs/remotes/", 13) == 0 ||
139 strncmp(refname, "remotes/", 8) == 0)
140 return got_error_path(refname, GOT_ERR_FETCH_BAD_REF);
144 repo_path = got_repo_get_path_git_dir(repo);
146 for (i = 0; i < nitems(tmpfds); i++)
149 TAILQ_INIT(&have_refs);
150 TAILQ_INIT(&my_refs);
152 if (!list_refs_only) {
153 err = got_ref_list(&my_refs, repo, NULL,
154 got_ref_cmp_by_name, NULL);
159 TAILQ_FOREACH(re, &my_refs, entry) {
160 struct got_object_id *id;
163 if (got_ref_is_symbolic(re->ref))
166 err = got_ref_resolve(&id, repo, re->ref);
169 refname = strdup(got_ref_get_name(re->ref));
170 if (refname == NULL) {
171 err = got_error_from_errno("strdup");
174 err = got_pathlist_append(&have_refs, refname, id);
179 if (list_refs_only) {
180 packfd = got_opentempfd();
182 err = got_error_from_errno("got_opentempfd");
186 if (asprintf(&path, "%s/%s/fetching.pack",
187 repo_path, GOT_OBJECTS_PACK_DIR) == -1) {
188 err = got_error_from_errno("asprintf");
191 err = got_opentemp_named_fd(&tmppackpath, &packfd, path, "");
195 if (fchmod(packfd, GOT_DEFAULT_FILE_MODE) != 0) {
196 err = got_error_from_errno2("fchmod", tmppackpath);
200 if (list_refs_only) {
201 idxfd = got_opentempfd();
203 err = got_error_from_errno("got_opentempfd");
207 if (asprintf(&path, "%s/%s/fetching.idx",
208 repo_path, GOT_OBJECTS_PACK_DIR) == -1) {
209 err = got_error_from_errno("asprintf");
212 err = got_opentemp_named_fd(&tmpidxpath, &idxfd, path, "");
216 if (fchmod(idxfd, GOT_DEFAULT_FILE_MODE) != 0) {
217 err = got_error_from_errno2("fchmod", tmpidxpath);
223 err = got_error_from_errno("dup");
227 for (i = 0; i < nitems(tmpfds); i++) {
228 tmpfds[i] = got_opentempfd();
229 if (tmpfds[i] == -1) {
230 err = got_error_from_errno("got_opentempfd");
235 if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_fetchfds) == -1) {
236 err = got_error_from_errno("socketpair");
241 if (fetchpid == -1) {
242 err = got_error_from_errno("fork");
244 } else if (fetchpid == 0){
245 got_privsep_exec_child(imsg_fetchfds,
246 GOT_PATH_PROG_FETCH_PACK, tmppackpath);
249 if (close(imsg_fetchfds[1]) == -1) {
250 err = got_error_from_errno("close");
253 imsg_init(&fetchibuf, imsg_fetchfds[0]);
254 nfetchfd = dup(fetchfd);
255 if (nfetchfd == -1) {
256 err = got_error_from_errno("dup");
259 err = got_privsep_send_fetch_req(&fetchibuf, nfetchfd, &have_refs,
260 fetch_all_branches, wanted_branches, wanted_refs,
261 list_refs_only, worktree_refname, remote_head, no_head, verbosity);
265 npackfd = dup(packfd);
267 err = got_error_from_errno("dup");
270 err = got_privsep_send_fetch_outfd(&fetchibuf, npackfd);
276 progress = calloc(GOT_PKT_MAX, 1);
277 if (progress == NULL) {
278 err = got_error_from_errno("calloc");
282 *pack_hash = calloc(1, sizeof(**pack_hash));
283 if (*pack_hash == NULL) {
284 err = got_error_from_errno("calloc");
289 struct got_object_id *id = NULL;
290 char *refname = NULL;
291 char *server_progress = NULL;
292 off_t packfile_size_cur = 0;
294 err = got_privsep_recv_fetch_progress(&done,
295 &id, &refname, symrefs, &server_progress,
296 &packfile_size_cur, (*pack_hash)->sha1, &fetchibuf);
299 /* Don't report size progress for an empty pack file. */
300 if (packfile_size_cur <= ssizeof(pack_hdr) + SHA1_DIGEST_LENGTH)
301 packfile_size_cur = 0;
302 if (!done && refname && id) {
303 err = got_pathlist_insert(NULL, refs, refname, id);
306 } else if (!done && server_progress) {
309 * XXX git-daemon tends to send batched output with
310 * lines spanning separate packets. Buffer progress
311 * output until we see a CR or LF to avoid giving
312 * partial lines of progress output to the callback.
314 if (strlcat(progress, server_progress,
315 GOT_PKT_MAX) >= GOT_PKT_MAX) {
316 progress[0] = '\0'; /* discard */
319 while ((p = strchr(progress, '\r')) != NULL ||
320 (p = strchr(progress, '\n')) != NULL) {
325 if (asprintf(&s, "%s%s", progress,
326 c == '\n' ? "\n" : "") == -1) {
327 err = got_error_from_errno("asprintf");
330 err = progress_cb(progress_arg, s,
331 packfile_size_cur, 0, 0, 0, 0);
335 n = strlen(progress);
336 if (n < GOT_PKT_MAX - 1) {
337 memmove(progress, &progress[n + 1],
338 GOT_PKT_MAX - n - 1);
342 free(server_progress);
345 } else if (!done && packfile_size_cur != packfile_size) {
346 err = progress_cb(progress_arg, NULL,
347 packfile_size_cur, 0, 0, 0, 0);
350 packfile_size = packfile_size_cur;
353 if (close(imsg_fetchfds[0]) == -1) {
354 err = got_error_from_errno("close");
357 if (waitpid(fetchpid, &fetchstatus, 0) == -1) {
358 err = got_error_from_errno("waitpid");
362 if (lseek(packfd, 0, SEEK_SET) == -1) {
363 err = got_error_from_errno("lseek");
367 /* If zero data was fetched without error we are already up-to-date. */
368 if (packfile_size == 0) {
372 } else if (packfile_size < ssizeof(pack_hdr) + SHA1_DIGEST_LENGTH) {
373 err = got_error_msg(GOT_ERR_BAD_PACKFILE, "short pack file");
378 n = read(packfd, &pack_hdr, ssizeof(pack_hdr));
380 err = got_error_from_errno("read");
383 if (n != ssizeof(pack_hdr)) {
384 err = got_error(GOT_ERR_IO);
387 if (pack_hdr.signature != htobe32(GOT_PACKFILE_SIGNATURE)) {
388 err = got_error_msg(GOT_ERR_BAD_PACKFILE,
389 "bad pack file signature");
392 if (pack_hdr.version != htobe32(GOT_PACKFILE_VERSION)) {
393 err = got_error_msg(GOT_ERR_BAD_PACKFILE,
394 "bad pack file version");
397 nobj = be32toh(pack_hdr.nobjects);
399 packfile_size > ssizeof(pack_hdr) + SHA1_DIGEST_LENGTH)
400 return got_error_msg(GOT_ERR_BAD_PACKFILE,
401 "bad pack file with zero objects");
403 packfile_size <= ssizeof(pack_hdr) + SHA1_DIGEST_LENGTH)
404 return got_error_msg(GOT_ERR_BAD_PACKFILE,
405 "empty pack file with non-zero object count");
409 * If the pack file contains no objects, we may only need to update
410 * references in our repository. The caller will take care of that.
418 if (lseek(packfd, 0, SEEK_SET) == -1) {
419 err = got_error_from_errno("lseek");
423 if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_idxfds) == -1) {
424 err = got_error_from_errno("socketpair");
429 err= got_error_from_errno("fork");
431 } else if (idxpid == 0)
432 got_privsep_exec_child(imsg_idxfds,
433 GOT_PATH_PROG_INDEX_PACK, tmppackpath);
434 if (close(imsg_idxfds[1]) == -1) {
435 err = got_error_from_errno("close");
438 imsg_init(&idxibuf, imsg_idxfds[0]);
440 npackfd = dup(packfd);
442 err = got_error_from_errno("dup");
445 err = got_privsep_send_index_pack_req(&idxibuf, (*pack_hash)->sha1,
450 err = got_privsep_send_index_pack_outfd(&idxibuf, nidxfd);
454 for (i = 0; i < nitems(tmpfds); i++) {
455 err = got_privsep_send_tmpfd(&idxibuf, tmpfds[i]);
462 int nobj_total, nobj_indexed, nobj_loose, nobj_resolved;
464 err = got_privsep_recv_index_progress(&done, &nobj_total,
465 &nobj_indexed, &nobj_loose, &nobj_resolved,
469 if (nobj_indexed != 0) {
470 err = progress_cb(progress_arg, NULL,
471 packfile_size, nobj_total,
472 nobj_indexed, nobj_loose, nobj_resolved);
477 if (close(imsg_idxfds[0]) == -1) {
478 err = got_error_from_errno("close");
481 if (waitpid(idxpid, &idxstatus, 0) == -1) {
482 err = got_error_from_errno("waitpid");
486 err = got_object_id_str(&id_str, *pack_hash);
489 if (asprintf(&packpath, "%s/%s/pack-%s.pack",
490 repo_path, GOT_OBJECTS_PACK_DIR, id_str) == -1) {
491 err = got_error_from_errno("asprintf");
495 if (asprintf(&idxpath, "%s/%s/pack-%s.idx",
496 repo_path, GOT_OBJECTS_PACK_DIR, id_str) == -1) {
497 err = got_error_from_errno("asprintf");
503 if (rename(tmppackpath, packpath) == -1) {
504 err = got_error_from_errno3("rename", tmppackpath, packpath);
509 if (rename(tmpidxpath, idxpath) == -1) {
510 err = got_error_from_errno3("rename", tmpidxpath, idxpath);
517 if (tmppackpath && unlink(tmppackpath) == -1 && err == NULL)
518 err = got_error_from_errno2("unlink", tmppackpath);
519 if (tmpidxpath && unlink(tmpidxpath) == -1 && err == NULL)
520 err = got_error_from_errno2("unlink", tmpidxpath);
521 if (nfetchfd != -1 && close(nfetchfd) == -1 && err == NULL)
522 err = got_error_from_errno("close");
523 if (npackfd != -1 && close(npackfd) == -1 && err == NULL)
524 err = got_error_from_errno("close");
525 if (packfd != -1 && close(packfd) == -1 && err == NULL)
526 err = got_error_from_errno("close");
527 if (idxfd != -1 && close(idxfd) == -1 && err == NULL)
528 err = got_error_from_errno("close");
529 for (i = 0; i < nitems(tmpfds); i++) {
530 if (tmpfds[i] != -1 && close(tmpfds[i]) == -1 && err == NULL)
531 err = got_error_from_errno("close");
540 got_pathlist_free(&have_refs, GOT_PATHLIST_FREE_ALL);
541 got_ref_list_free(&my_refs);
545 got_pathlist_free(refs, GOT_PATHLIST_FREE_ALL);
546 got_pathlist_free(symrefs, GOT_PATHLIST_FREE_ALL);