]> git.ipfire.org Git - thirdparty/git.git/blob - builtin/gc.c
Merge branch 'ps/pack-refs-auto'
[thirdparty/git.git] / builtin / gc.c
1 /*
2 * git gc builtin command
3 *
4 * Cleanup unreachable files and optimize the repository.
5 *
6 * Copyright (c) 2007 James Bowes
7 *
8 * Based on git-gc.sh, which is
9 *
10 * Copyright (c) 2006 Shawn O. Pearce
11 */
12
13 #include "builtin.h"
14 #include "abspath.h"
15 #include "date.h"
16 #include "environment.h"
17 #include "hex.h"
18 #include "repository.h"
19 #include "config.h"
20 #include "tempfile.h"
21 #include "lockfile.h"
22 #include "parse-options.h"
23 #include "run-command.h"
24 #include "sigchain.h"
25 #include "strvec.h"
26 #include "commit.h"
27 #include "commit-graph.h"
28 #include "packfile.h"
29 #include "object-file.h"
30 #include "object-store-ll.h"
31 #include "pack.h"
32 #include "pack-objects.h"
33 #include "path.h"
34 #include "blob.h"
35 #include "tree.h"
36 #include "promisor-remote.h"
37 #include "refs.h"
38 #include "remote.h"
39 #include "exec-cmd.h"
40 #include "gettext.h"
41 #include "hook.h"
42 #include "setup.h"
43 #include "trace2.h"
44
45 #define FAILED_RUN "failed to run %s"
46
47 static const char * const builtin_gc_usage[] = {
48 N_("git gc [<options>]"),
49 NULL
50 };
51
52 static int pack_refs = 1;
53 static int prune_reflogs = 1;
54 static int cruft_packs = 1;
55 static unsigned long max_cruft_size;
56 static int aggressive_depth = 50;
57 static int aggressive_window = 250;
58 static int gc_auto_threshold = 6700;
59 static int gc_auto_pack_limit = 50;
60 static int detach_auto = 1;
61 static timestamp_t gc_log_expire_time;
62 static const char *gc_log_expire = "1.day.ago";
63 static const char *prune_expire = "2.weeks.ago";
64 static const char *prune_worktrees_expire = "3.months.ago";
65 static char *repack_filter;
66 static char *repack_filter_to;
67 static unsigned long big_pack_threshold;
68 static unsigned long max_delta_cache_size = DEFAULT_DELTA_CACHE_SIZE;
69
70 static struct strvec reflog = STRVEC_INIT;
71 static struct strvec repack = STRVEC_INIT;
72 static struct strvec prune = STRVEC_INIT;
73 static struct strvec prune_worktrees = STRVEC_INIT;
74 static struct strvec rerere = STRVEC_INIT;
75
76 static struct tempfile *pidfile;
77 static struct lock_file log_lock;
78
79 static struct string_list pack_garbage = STRING_LIST_INIT_DUP;
80
81 static void clean_pack_garbage(void)
82 {
83 int i;
84 for (i = 0; i < pack_garbage.nr; i++)
85 unlink_or_warn(pack_garbage.items[i].string);
86 string_list_clear(&pack_garbage, 0);
87 }
88
89 static void report_pack_garbage(unsigned seen_bits, const char *path)
90 {
91 if (seen_bits == PACKDIR_FILE_IDX)
92 string_list_append(&pack_garbage, path);
93 }
94
95 static void process_log_file(void)
96 {
97 struct stat st;
98 if (fstat(get_lock_file_fd(&log_lock), &st)) {
99 /*
100 * Perhaps there was an i/o error or another
101 * unlikely situation. Try to make a note of
102 * this in gc.log along with any existing
103 * messages.
104 */
105 int saved_errno = errno;
106 fprintf(stderr, _("Failed to fstat %s: %s"),
107 get_lock_file_path(&log_lock),
108 strerror(saved_errno));
109 fflush(stderr);
110 commit_lock_file(&log_lock);
111 errno = saved_errno;
112 } else if (st.st_size) {
113 /* There was some error recorded in the lock file */
114 commit_lock_file(&log_lock);
115 } else {
116 /* No error, clean up any old gc.log */
117 unlink(git_path("gc.log"));
118 rollback_lock_file(&log_lock);
119 }
120 }
121
122 static void process_log_file_at_exit(void)
123 {
124 fflush(stderr);
125 process_log_file();
126 }
127
128 static void process_log_file_on_signal(int signo)
129 {
130 process_log_file();
131 sigchain_pop(signo);
132 raise(signo);
133 }
134
135 static int gc_config_is_timestamp_never(const char *var)
136 {
137 const char *value;
138 timestamp_t expire;
139
140 if (!git_config_get_value(var, &value) && value) {
141 if (parse_expiry_date(value, &expire))
142 die(_("failed to parse '%s' value '%s'"), var, value);
143 return expire == 0;
144 }
145 return 0;
146 }
147
148 static void gc_config(void)
149 {
150 const char *value;
151
152 if (!git_config_get_value("gc.packrefs", &value)) {
153 if (value && !strcmp(value, "notbare"))
154 pack_refs = -1;
155 else
156 pack_refs = git_config_bool("gc.packrefs", value);
157 }
158
159 if (gc_config_is_timestamp_never("gc.reflogexpire") &&
160 gc_config_is_timestamp_never("gc.reflogexpireunreachable"))
161 prune_reflogs = 0;
162
163 git_config_get_int("gc.aggressivewindow", &aggressive_window);
164 git_config_get_int("gc.aggressivedepth", &aggressive_depth);
165 git_config_get_int("gc.auto", &gc_auto_threshold);
166 git_config_get_int("gc.autopacklimit", &gc_auto_pack_limit);
167 git_config_get_bool("gc.autodetach", &detach_auto);
168 git_config_get_bool("gc.cruftpacks", &cruft_packs);
169 git_config_get_ulong("gc.maxcruftsize", &max_cruft_size);
170 git_config_get_expiry("gc.pruneexpire", &prune_expire);
171 git_config_get_expiry("gc.worktreepruneexpire", &prune_worktrees_expire);
172 git_config_get_expiry("gc.logexpiry", &gc_log_expire);
173
174 git_config_get_ulong("gc.bigpackthreshold", &big_pack_threshold);
175 git_config_get_ulong("pack.deltacachesize", &max_delta_cache_size);
176
177 git_config_get_string("gc.repackfilter", &repack_filter);
178 git_config_get_string("gc.repackfilterto", &repack_filter_to);
179
180 git_config(git_default_config, NULL);
181 }
182
183 enum schedule_priority {
184 SCHEDULE_NONE = 0,
185 SCHEDULE_WEEKLY = 1,
186 SCHEDULE_DAILY = 2,
187 SCHEDULE_HOURLY = 3,
188 };
189
190 static enum schedule_priority parse_schedule(const char *value)
191 {
192 if (!value)
193 return SCHEDULE_NONE;
194 if (!strcasecmp(value, "hourly"))
195 return SCHEDULE_HOURLY;
196 if (!strcasecmp(value, "daily"))
197 return SCHEDULE_DAILY;
198 if (!strcasecmp(value, "weekly"))
199 return SCHEDULE_WEEKLY;
200 return SCHEDULE_NONE;
201 }
202
203 struct maintenance_run_opts {
204 int auto_flag;
205 int quiet;
206 enum schedule_priority schedule;
207 };
208
209 static int pack_refs_condition(void)
210 {
211 /*
212 * The auto-repacking logic for refs is handled by the ref backends and
213 * exposed via `git pack-refs --auto`. We thus always return truish
214 * here and let the backend decide for us.
215 */
216 return 1;
217 }
218
219 static int maintenance_task_pack_refs(MAYBE_UNUSED struct maintenance_run_opts *opts)
220 {
221 struct child_process cmd = CHILD_PROCESS_INIT;
222
223 cmd.git_cmd = 1;
224 strvec_pushl(&cmd.args, "pack-refs", "--all", "--prune", NULL);
225 if (opts->auto_flag)
226 strvec_push(&cmd.args, "--auto");
227
228 return run_command(&cmd);
229 }
230
231 static int too_many_loose_objects(void)
232 {
233 /*
234 * Quickly check if a "gc" is needed, by estimating how
235 * many loose objects there are. Because SHA-1 is evenly
236 * distributed, we can check only one and get a reasonable
237 * estimate.
238 */
239 DIR *dir;
240 struct dirent *ent;
241 int auto_threshold;
242 int num_loose = 0;
243 int needed = 0;
244 const unsigned hexsz_loose = the_hash_algo->hexsz - 2;
245
246 dir = opendir(git_path("objects/17"));
247 if (!dir)
248 return 0;
249
250 auto_threshold = DIV_ROUND_UP(gc_auto_threshold, 256);
251 while ((ent = readdir(dir)) != NULL) {
252 if (strspn(ent->d_name, "0123456789abcdef") != hexsz_loose ||
253 ent->d_name[hexsz_loose] != '\0')
254 continue;
255 if (++num_loose > auto_threshold) {
256 needed = 1;
257 break;
258 }
259 }
260 closedir(dir);
261 return needed;
262 }
263
264 static struct packed_git *find_base_packs(struct string_list *packs,
265 unsigned long limit)
266 {
267 struct packed_git *p, *base = NULL;
268
269 for (p = get_all_packs(the_repository); p; p = p->next) {
270 if (!p->pack_local || p->is_cruft)
271 continue;
272 if (limit) {
273 if (p->pack_size >= limit)
274 string_list_append(packs, p->pack_name);
275 } else if (!base || base->pack_size < p->pack_size) {
276 base = p;
277 }
278 }
279
280 if (base)
281 string_list_append(packs, base->pack_name);
282
283 return base;
284 }
285
286 static int too_many_packs(void)
287 {
288 struct packed_git *p;
289 int cnt;
290
291 if (gc_auto_pack_limit <= 0)
292 return 0;
293
294 for (cnt = 0, p = get_all_packs(the_repository); p; p = p->next) {
295 if (!p->pack_local)
296 continue;
297 if (p->pack_keep)
298 continue;
299 /*
300 * Perhaps check the size of the pack and count only
301 * very small ones here?
302 */
303 cnt++;
304 }
305 return gc_auto_pack_limit < cnt;
306 }
307
308 static uint64_t total_ram(void)
309 {
310 #if defined(HAVE_SYSINFO)
311 struct sysinfo si;
312
313 if (!sysinfo(&si))
314 return si.totalram;
315 #elif defined(HAVE_BSD_SYSCTL) && (defined(HW_MEMSIZE) || defined(HW_PHYSMEM))
316 int64_t physical_memory;
317 int mib[2];
318 size_t length;
319
320 mib[0] = CTL_HW;
321 # if defined(HW_MEMSIZE)
322 mib[1] = HW_MEMSIZE;
323 # else
324 mib[1] = HW_PHYSMEM;
325 # endif
326 length = sizeof(int64_t);
327 if (!sysctl(mib, 2, &physical_memory, &length, NULL, 0))
328 return physical_memory;
329 #elif defined(GIT_WINDOWS_NATIVE)
330 MEMORYSTATUSEX memInfo;
331
332 memInfo.dwLength = sizeof(MEMORYSTATUSEX);
333 if (GlobalMemoryStatusEx(&memInfo))
334 return memInfo.ullTotalPhys;
335 #endif
336 return 0;
337 }
338
339 static uint64_t estimate_repack_memory(struct packed_git *pack)
340 {
341 unsigned long nr_objects = repo_approximate_object_count(the_repository);
342 size_t os_cache, heap;
343
344 if (!pack || !nr_objects)
345 return 0;
346
347 /*
348 * First we have to scan through at least one pack.
349 * Assume enough room in OS file cache to keep the entire pack
350 * or we may accidentally evict data of other processes from
351 * the cache.
352 */
353 os_cache = pack->pack_size + pack->index_size;
354 /* then pack-objects needs lots more for book keeping */
355 heap = sizeof(struct object_entry) * nr_objects;
356 /*
357 * internal rev-list --all --objects takes up some memory too,
358 * let's say half of it is for blobs
359 */
360 heap += sizeof(struct blob) * nr_objects / 2;
361 /*
362 * and the other half is for trees (commits and tags are
363 * usually insignificant)
364 */
365 heap += sizeof(struct tree) * nr_objects / 2;
366 /* and then obj_hash[], underestimated in fact */
367 heap += sizeof(struct object *) * nr_objects;
368 /* revindex is used also */
369 heap += (sizeof(off_t) + sizeof(uint32_t)) * nr_objects;
370 /*
371 * read_sha1_file() (either at delta calculation phase, or
372 * writing phase) also fills up the delta base cache
373 */
374 heap += delta_base_cache_limit;
375 /* and of course pack-objects has its own delta cache */
376 heap += max_delta_cache_size;
377
378 return os_cache + heap;
379 }
380
381 static int keep_one_pack(struct string_list_item *item, void *data UNUSED)
382 {
383 strvec_pushf(&repack, "--keep-pack=%s", basename(item->string));
384 return 0;
385 }
386
387 static void add_repack_all_option(struct string_list *keep_pack)
388 {
389 if (prune_expire && !strcmp(prune_expire, "now"))
390 strvec_push(&repack, "-a");
391 else if (cruft_packs) {
392 strvec_push(&repack, "--cruft");
393 if (prune_expire)
394 strvec_pushf(&repack, "--cruft-expiration=%s", prune_expire);
395 if (max_cruft_size)
396 strvec_pushf(&repack, "--max-cruft-size=%lu",
397 max_cruft_size);
398 } else {
399 strvec_push(&repack, "-A");
400 if (prune_expire)
401 strvec_pushf(&repack, "--unpack-unreachable=%s", prune_expire);
402 }
403
404 if (keep_pack)
405 for_each_string_list(keep_pack, keep_one_pack, NULL);
406
407 if (repack_filter && *repack_filter)
408 strvec_pushf(&repack, "--filter=%s", repack_filter);
409 if (repack_filter_to && *repack_filter_to)
410 strvec_pushf(&repack, "--filter-to=%s", repack_filter_to);
411 }
412
413 static void add_repack_incremental_option(void)
414 {
415 strvec_push(&repack, "--no-write-bitmap-index");
416 }
417
418 static int need_to_gc(void)
419 {
420 /*
421 * Setting gc.auto to 0 or negative can disable the
422 * automatic gc.
423 */
424 if (gc_auto_threshold <= 0)
425 return 0;
426
427 /*
428 * If there are too many loose objects, but not too many
429 * packs, we run "repack -d -l". If there are too many packs,
430 * we run "repack -A -d -l". Otherwise we tell the caller
431 * there is no need.
432 */
433 if (too_many_packs()) {
434 struct string_list keep_pack = STRING_LIST_INIT_NODUP;
435
436 if (big_pack_threshold) {
437 find_base_packs(&keep_pack, big_pack_threshold);
438 if (keep_pack.nr >= gc_auto_pack_limit) {
439 big_pack_threshold = 0;
440 string_list_clear(&keep_pack, 0);
441 find_base_packs(&keep_pack, 0);
442 }
443 } else {
444 struct packed_git *p = find_base_packs(&keep_pack, 0);
445 uint64_t mem_have, mem_want;
446
447 mem_have = total_ram();
448 mem_want = estimate_repack_memory(p);
449
450 /*
451 * Only allow 1/2 of memory for pack-objects, leave
452 * the rest for the OS and other processes in the
453 * system.
454 */
455 if (!mem_have || mem_want < mem_have / 2)
456 string_list_clear(&keep_pack, 0);
457 }
458
459 add_repack_all_option(&keep_pack);
460 string_list_clear(&keep_pack, 0);
461 } else if (too_many_loose_objects())
462 add_repack_incremental_option();
463 else
464 return 0;
465
466 if (run_hooks("pre-auto-gc"))
467 return 0;
468 return 1;
469 }
470
471 /* return NULL on success, else hostname running the gc */
472 static const char *lock_repo_for_gc(int force, pid_t* ret_pid)
473 {
474 struct lock_file lock = LOCK_INIT;
475 char my_host[HOST_NAME_MAX + 1];
476 struct strbuf sb = STRBUF_INIT;
477 struct stat st;
478 uintmax_t pid;
479 FILE *fp;
480 int fd;
481 char *pidfile_path;
482
483 if (is_tempfile_active(pidfile))
484 /* already locked */
485 return NULL;
486
487 if (xgethostname(my_host, sizeof(my_host)))
488 xsnprintf(my_host, sizeof(my_host), "unknown");
489
490 pidfile_path = git_pathdup("gc.pid");
491 fd = hold_lock_file_for_update(&lock, pidfile_path,
492 LOCK_DIE_ON_ERROR);
493 if (!force) {
494 static char locking_host[HOST_NAME_MAX + 1];
495 static char *scan_fmt;
496 int should_exit;
497
498 if (!scan_fmt)
499 scan_fmt = xstrfmt("%s %%%ds", "%"SCNuMAX, HOST_NAME_MAX);
500 fp = fopen(pidfile_path, "r");
501 memset(locking_host, 0, sizeof(locking_host));
502 should_exit =
503 fp != NULL &&
504 !fstat(fileno(fp), &st) &&
505 /*
506 * 12 hour limit is very generous as gc should
507 * never take that long. On the other hand we
508 * don't really need a strict limit here,
509 * running gc --auto one day late is not a big
510 * problem. --force can be used in manual gc
511 * after the user verifies that no gc is
512 * running.
513 */
514 time(NULL) - st.st_mtime <= 12 * 3600 &&
515 fscanf(fp, scan_fmt, &pid, locking_host) == 2 &&
516 /* be gentle to concurrent "gc" on remote hosts */
517 (strcmp(locking_host, my_host) || !kill(pid, 0) || errno == EPERM);
518 if (fp)
519 fclose(fp);
520 if (should_exit) {
521 if (fd >= 0)
522 rollback_lock_file(&lock);
523 *ret_pid = pid;
524 free(pidfile_path);
525 return locking_host;
526 }
527 }
528
529 strbuf_addf(&sb, "%"PRIuMAX" %s",
530 (uintmax_t) getpid(), my_host);
531 write_in_full(fd, sb.buf, sb.len);
532 strbuf_release(&sb);
533 commit_lock_file(&lock);
534 pidfile = register_tempfile(pidfile_path);
535 free(pidfile_path);
536 return NULL;
537 }
538
539 /*
540 * Returns 0 if there was no previous error and gc can proceed, 1 if
541 * gc should not proceed due to an error in the last run. Prints a
542 * message and returns with a non-[01] status code if an error occurred
543 * while reading gc.log
544 */
545 static int report_last_gc_error(void)
546 {
547 struct strbuf sb = STRBUF_INIT;
548 int ret = 0;
549 ssize_t len;
550 struct stat st;
551 char *gc_log_path = git_pathdup("gc.log");
552
553 if (stat(gc_log_path, &st)) {
554 if (errno == ENOENT)
555 goto done;
556
557 ret = die_message_errno(_("cannot stat '%s'"), gc_log_path);
558 goto done;
559 }
560
561 if (st.st_mtime < gc_log_expire_time)
562 goto done;
563
564 len = strbuf_read_file(&sb, gc_log_path, 0);
565 if (len < 0)
566 ret = die_message_errno(_("cannot read '%s'"), gc_log_path);
567 else if (len > 0) {
568 /*
569 * A previous gc failed. Report the error, and don't
570 * bother with an automatic gc run since it is likely
571 * to fail in the same way.
572 */
573 warning(_("The last gc run reported the following. "
574 "Please correct the root cause\n"
575 "and remove %s\n"
576 "Automatic cleanup will not be performed "
577 "until the file is removed.\n\n"
578 "%s"),
579 gc_log_path, sb.buf);
580 ret = 1;
581 }
582 strbuf_release(&sb);
583 done:
584 free(gc_log_path);
585 return ret;
586 }
587
588 static void gc_before_repack(struct maintenance_run_opts *opts)
589 {
590 /*
591 * We may be called twice, as both the pre- and
592 * post-daemonized phases will call us, but running these
593 * commands more than once is pointless and wasteful.
594 */
595 static int done = 0;
596 if (done++)
597 return;
598
599 if (pack_refs && maintenance_task_pack_refs(opts))
600 die(FAILED_RUN, "pack-refs");
601
602 if (prune_reflogs) {
603 struct child_process cmd = CHILD_PROCESS_INIT;
604
605 cmd.git_cmd = 1;
606 strvec_pushv(&cmd.args, reflog.v);
607 if (run_command(&cmd))
608 die(FAILED_RUN, reflog.v[0]);
609 }
610 }
611
612 int cmd_gc(int argc, const char **argv, const char *prefix)
613 {
614 int aggressive = 0;
615 int quiet = 0;
616 int force = 0;
617 const char *name;
618 pid_t pid;
619 int daemonized = 0;
620 int keep_largest_pack = -1;
621 timestamp_t dummy;
622 struct child_process rerere_cmd = CHILD_PROCESS_INIT;
623 struct maintenance_run_opts opts = {0};
624
625 struct option builtin_gc_options[] = {
626 OPT__QUIET(&quiet, N_("suppress progress reporting")),
627 { OPTION_STRING, 0, "prune", &prune_expire, N_("date"),
628 N_("prune unreferenced objects"),
629 PARSE_OPT_OPTARG, NULL, (intptr_t)prune_expire },
630 OPT_BOOL(0, "cruft", &cruft_packs, N_("pack unreferenced objects separately")),
631 OPT_MAGNITUDE(0, "max-cruft-size", &max_cruft_size,
632 N_("with --cruft, limit the size of new cruft packs")),
633 OPT_BOOL(0, "aggressive", &aggressive, N_("be more thorough (increased runtime)")),
634 OPT_BOOL_F(0, "auto", &opts.auto_flag, N_("enable auto-gc mode"),
635 PARSE_OPT_NOCOMPLETE),
636 OPT_BOOL_F(0, "force", &force,
637 N_("force running gc even if there may be another gc running"),
638 PARSE_OPT_NOCOMPLETE),
639 OPT_BOOL(0, "keep-largest-pack", &keep_largest_pack,
640 N_("repack all other packs except the largest pack")),
641 OPT_END()
642 };
643
644 if (argc == 2 && !strcmp(argv[1], "-h"))
645 usage_with_options(builtin_gc_usage, builtin_gc_options);
646
647 strvec_pushl(&reflog, "reflog", "expire", "--all", NULL);
648 strvec_pushl(&repack, "repack", "-d", "-l", NULL);
649 strvec_pushl(&prune, "prune", "--expire", NULL);
650 strvec_pushl(&prune_worktrees, "worktree", "prune", "--expire", NULL);
651 strvec_pushl(&rerere, "rerere", "gc", NULL);
652
653 /* default expiry time, overwritten in gc_config */
654 gc_config();
655 if (parse_expiry_date(gc_log_expire, &gc_log_expire_time))
656 die(_("failed to parse gc.logExpiry value %s"), gc_log_expire);
657
658 if (pack_refs < 0)
659 pack_refs = !is_bare_repository();
660
661 argc = parse_options(argc, argv, prefix, builtin_gc_options,
662 builtin_gc_usage, 0);
663 if (argc > 0)
664 usage_with_options(builtin_gc_usage, builtin_gc_options);
665
666 if (prune_expire && parse_expiry_date(prune_expire, &dummy))
667 die(_("failed to parse prune expiry value %s"), prune_expire);
668
669 if (aggressive) {
670 strvec_push(&repack, "-f");
671 if (aggressive_depth > 0)
672 strvec_pushf(&repack, "--depth=%d", aggressive_depth);
673 if (aggressive_window > 0)
674 strvec_pushf(&repack, "--window=%d", aggressive_window);
675 }
676 if (quiet)
677 strvec_push(&repack, "-q");
678
679 if (opts.auto_flag) {
680 /*
681 * Auto-gc should be least intrusive as possible.
682 */
683 if (!need_to_gc())
684 return 0;
685 if (!quiet) {
686 if (detach_auto)
687 fprintf(stderr, _("Auto packing the repository in background for optimum performance.\n"));
688 else
689 fprintf(stderr, _("Auto packing the repository for optimum performance.\n"));
690 fprintf(stderr, _("See \"git help gc\" for manual housekeeping.\n"));
691 }
692 if (detach_auto) {
693 int ret = report_last_gc_error();
694
695 if (ret == 1)
696 /* Last gc --auto failed. Skip this one. */
697 return 0;
698 else if (ret)
699 /* an I/O error occurred, already reported */
700 return ret;
701
702 if (lock_repo_for_gc(force, &pid))
703 return 0;
704 gc_before_repack(&opts); /* dies on failure */
705 delete_tempfile(&pidfile);
706
707 /*
708 * failure to daemonize is ok, we'll continue
709 * in foreground
710 */
711 daemonized = !daemonize();
712 }
713 } else {
714 struct string_list keep_pack = STRING_LIST_INIT_NODUP;
715
716 if (keep_largest_pack != -1) {
717 if (keep_largest_pack)
718 find_base_packs(&keep_pack, 0);
719 } else if (big_pack_threshold) {
720 find_base_packs(&keep_pack, big_pack_threshold);
721 }
722
723 add_repack_all_option(&keep_pack);
724 string_list_clear(&keep_pack, 0);
725 }
726
727 name = lock_repo_for_gc(force, &pid);
728 if (name) {
729 if (opts.auto_flag)
730 return 0; /* be quiet on --auto */
731 die(_("gc is already running on machine '%s' pid %"PRIuMAX" (use --force if not)"),
732 name, (uintmax_t)pid);
733 }
734
735 if (daemonized) {
736 hold_lock_file_for_update(&log_lock,
737 git_path("gc.log"),
738 LOCK_DIE_ON_ERROR);
739 dup2(get_lock_file_fd(&log_lock), 2);
740 sigchain_push_common(process_log_file_on_signal);
741 atexit(process_log_file_at_exit);
742 }
743
744 gc_before_repack(&opts);
745
746 if (!repository_format_precious_objects) {
747 struct child_process repack_cmd = CHILD_PROCESS_INIT;
748
749 repack_cmd.git_cmd = 1;
750 repack_cmd.close_object_store = 1;
751 strvec_pushv(&repack_cmd.args, repack.v);
752 if (run_command(&repack_cmd))
753 die(FAILED_RUN, repack.v[0]);
754
755 if (prune_expire) {
756 struct child_process prune_cmd = CHILD_PROCESS_INIT;
757
758 /* run `git prune` even if using cruft packs */
759 strvec_push(&prune, prune_expire);
760 if (quiet)
761 strvec_push(&prune, "--no-progress");
762 if (repo_has_promisor_remote(the_repository))
763 strvec_push(&prune,
764 "--exclude-promisor-objects");
765 prune_cmd.git_cmd = 1;
766 strvec_pushv(&prune_cmd.args, prune.v);
767 if (run_command(&prune_cmd))
768 die(FAILED_RUN, prune.v[0]);
769 }
770 }
771
772 if (prune_worktrees_expire) {
773 struct child_process prune_worktrees_cmd = CHILD_PROCESS_INIT;
774
775 strvec_push(&prune_worktrees, prune_worktrees_expire);
776 prune_worktrees_cmd.git_cmd = 1;
777 strvec_pushv(&prune_worktrees_cmd.args, prune_worktrees.v);
778 if (run_command(&prune_worktrees_cmd))
779 die(FAILED_RUN, prune_worktrees.v[0]);
780 }
781
782 rerere_cmd.git_cmd = 1;
783 strvec_pushv(&rerere_cmd.args, rerere.v);
784 if (run_command(&rerere_cmd))
785 die(FAILED_RUN, rerere.v[0]);
786
787 report_garbage = report_pack_garbage;
788 reprepare_packed_git(the_repository);
789 if (pack_garbage.nr > 0) {
790 close_object_store(the_repository->objects);
791 clean_pack_garbage();
792 }
793
794 if (the_repository->settings.gc_write_commit_graph == 1)
795 write_commit_graph_reachable(the_repository->objects->odb,
796 !quiet && !daemonized ? COMMIT_GRAPH_WRITE_PROGRESS : 0,
797 NULL);
798
799 if (opts.auto_flag && too_many_loose_objects())
800 warning(_("There are too many unreachable loose objects; "
801 "run 'git prune' to remove them."));
802
803 if (!daemonized)
804 unlink(git_path("gc.log"));
805
806 return 0;
807 }
808
809 static const char *const builtin_maintenance_run_usage[] = {
810 N_("git maintenance run [--auto] [--[no-]quiet] [--task=<task>] [--schedule]"),
811 NULL
812 };
813
814 static int maintenance_opt_schedule(const struct option *opt, const char *arg,
815 int unset)
816 {
817 enum schedule_priority *priority = opt->value;
818
819 if (unset)
820 die(_("--no-schedule is not allowed"));
821
822 *priority = parse_schedule(arg);
823
824 if (!*priority)
825 die(_("unrecognized --schedule argument '%s'"), arg);
826
827 return 0;
828 }
829
830 /* Remember to update object flag allocation in object.h */
831 #define SEEN (1u<<0)
832
833 struct cg_auto_data {
834 int num_not_in_graph;
835 int limit;
836 };
837
838 static int dfs_on_ref(const char *refname UNUSED,
839 const struct object_id *oid,
840 int flags UNUSED,
841 void *cb_data)
842 {
843 struct cg_auto_data *data = (struct cg_auto_data *)cb_data;
844 int result = 0;
845 struct object_id peeled;
846 struct commit_list *stack = NULL;
847 struct commit *commit;
848
849 if (!peel_iterated_oid(oid, &peeled))
850 oid = &peeled;
851 if (oid_object_info(the_repository, oid, NULL) != OBJ_COMMIT)
852 return 0;
853
854 commit = lookup_commit(the_repository, oid);
855 if (!commit)
856 return 0;
857 if (repo_parse_commit(the_repository, commit) ||
858 commit_graph_position(commit) != COMMIT_NOT_FROM_GRAPH)
859 return 0;
860
861 data->num_not_in_graph++;
862
863 if (data->num_not_in_graph >= data->limit)
864 return 1;
865
866 commit_list_append(commit, &stack);
867
868 while (!result && stack) {
869 struct commit_list *parent;
870
871 commit = pop_commit(&stack);
872
873 for (parent = commit->parents; parent; parent = parent->next) {
874 if (repo_parse_commit(the_repository, parent->item) ||
875 commit_graph_position(parent->item) != COMMIT_NOT_FROM_GRAPH ||
876 parent->item->object.flags & SEEN)
877 continue;
878
879 parent->item->object.flags |= SEEN;
880 data->num_not_in_graph++;
881
882 if (data->num_not_in_graph >= data->limit) {
883 result = 1;
884 break;
885 }
886
887 commit_list_append(parent->item, &stack);
888 }
889 }
890
891 free_commit_list(stack);
892 return result;
893 }
894
895 static int should_write_commit_graph(void)
896 {
897 int result;
898 struct cg_auto_data data;
899
900 data.num_not_in_graph = 0;
901 data.limit = 100;
902 git_config_get_int("maintenance.commit-graph.auto",
903 &data.limit);
904
905 if (!data.limit)
906 return 0;
907 if (data.limit < 0)
908 return 1;
909
910 result = for_each_ref(dfs_on_ref, &data);
911
912 repo_clear_commit_marks(the_repository, SEEN);
913
914 return result;
915 }
916
917 static int run_write_commit_graph(struct maintenance_run_opts *opts)
918 {
919 struct child_process child = CHILD_PROCESS_INIT;
920
921 child.git_cmd = child.close_object_store = 1;
922 strvec_pushl(&child.args, "commit-graph", "write",
923 "--split", "--reachable", NULL);
924
925 if (opts->quiet)
926 strvec_push(&child.args, "--no-progress");
927
928 return !!run_command(&child);
929 }
930
931 static int maintenance_task_commit_graph(struct maintenance_run_opts *opts)
932 {
933 prepare_repo_settings(the_repository);
934 if (!the_repository->settings.core_commit_graph)
935 return 0;
936
937 if (run_write_commit_graph(opts)) {
938 error(_("failed to write commit-graph"));
939 return 1;
940 }
941
942 return 0;
943 }
944
945 static int fetch_remote(struct remote *remote, void *cbdata)
946 {
947 struct maintenance_run_opts *opts = cbdata;
948 struct child_process child = CHILD_PROCESS_INIT;
949
950 if (remote->skip_default_update)
951 return 0;
952
953 child.git_cmd = 1;
954 strvec_pushl(&child.args, "fetch", remote->name,
955 "--prefetch", "--prune", "--no-tags",
956 "--no-write-fetch-head", "--recurse-submodules=no",
957 NULL);
958
959 if (opts->quiet)
960 strvec_push(&child.args, "--quiet");
961
962 return !!run_command(&child);
963 }
964
965 static int maintenance_task_prefetch(struct maintenance_run_opts *opts)
966 {
967 if (for_each_remote(fetch_remote, opts)) {
968 error(_("failed to prefetch remotes"));
969 return 1;
970 }
971
972 return 0;
973 }
974
975 static int maintenance_task_gc(struct maintenance_run_opts *opts)
976 {
977 struct child_process child = CHILD_PROCESS_INIT;
978
979 child.git_cmd = child.close_object_store = 1;
980 strvec_push(&child.args, "gc");
981
982 if (opts->auto_flag)
983 strvec_push(&child.args, "--auto");
984 if (opts->quiet)
985 strvec_push(&child.args, "--quiet");
986 else
987 strvec_push(&child.args, "--no-quiet");
988
989 return run_command(&child);
990 }
991
992 static int prune_packed(struct maintenance_run_opts *opts)
993 {
994 struct child_process child = CHILD_PROCESS_INIT;
995
996 child.git_cmd = 1;
997 strvec_push(&child.args, "prune-packed");
998
999 if (opts->quiet)
1000 strvec_push(&child.args, "--quiet");
1001
1002 return !!run_command(&child);
1003 }
1004
1005 struct write_loose_object_data {
1006 FILE *in;
1007 int count;
1008 int batch_size;
1009 };
1010
1011 static int loose_object_auto_limit = 100;
1012
1013 static int loose_object_count(const struct object_id *oid UNUSED,
1014 const char *path UNUSED,
1015 void *data)
1016 {
1017 int *count = (int*)data;
1018 if (++(*count) >= loose_object_auto_limit)
1019 return 1;
1020 return 0;
1021 }
1022
1023 static int loose_object_auto_condition(void)
1024 {
1025 int count = 0;
1026
1027 git_config_get_int("maintenance.loose-objects.auto",
1028 &loose_object_auto_limit);
1029
1030 if (!loose_object_auto_limit)
1031 return 0;
1032 if (loose_object_auto_limit < 0)
1033 return 1;
1034
1035 return for_each_loose_file_in_objdir(the_repository->objects->odb->path,
1036 loose_object_count,
1037 NULL, NULL, &count);
1038 }
1039
1040 static int bail_on_loose(const struct object_id *oid UNUSED,
1041 const char *path UNUSED,
1042 void *data UNUSED)
1043 {
1044 return 1;
1045 }
1046
1047 static int write_loose_object_to_stdin(const struct object_id *oid,
1048 const char *path UNUSED,
1049 void *data)
1050 {
1051 struct write_loose_object_data *d = (struct write_loose_object_data *)data;
1052
1053 fprintf(d->in, "%s\n", oid_to_hex(oid));
1054
1055 return ++(d->count) > d->batch_size;
1056 }
1057
1058 static int pack_loose(struct maintenance_run_opts *opts)
1059 {
1060 struct repository *r = the_repository;
1061 int result = 0;
1062 struct write_loose_object_data data;
1063 struct child_process pack_proc = CHILD_PROCESS_INIT;
1064
1065 /*
1066 * Do not start pack-objects process
1067 * if there are no loose objects.
1068 */
1069 if (!for_each_loose_file_in_objdir(r->objects->odb->path,
1070 bail_on_loose,
1071 NULL, NULL, NULL))
1072 return 0;
1073
1074 pack_proc.git_cmd = 1;
1075
1076 strvec_push(&pack_proc.args, "pack-objects");
1077 if (opts->quiet)
1078 strvec_push(&pack_proc.args, "--quiet");
1079 strvec_pushf(&pack_proc.args, "%s/pack/loose", r->objects->odb->path);
1080
1081 pack_proc.in = -1;
1082
1083 if (start_command(&pack_proc)) {
1084 error(_("failed to start 'git pack-objects' process"));
1085 return 1;
1086 }
1087
1088 data.in = xfdopen(pack_proc.in, "w");
1089 data.count = 0;
1090 data.batch_size = 50000;
1091
1092 for_each_loose_file_in_objdir(r->objects->odb->path,
1093 write_loose_object_to_stdin,
1094 NULL,
1095 NULL,
1096 &data);
1097
1098 fclose(data.in);
1099
1100 if (finish_command(&pack_proc)) {
1101 error(_("failed to finish 'git pack-objects' process"));
1102 result = 1;
1103 }
1104
1105 return result;
1106 }
1107
1108 static int maintenance_task_loose_objects(struct maintenance_run_opts *opts)
1109 {
1110 return prune_packed(opts) || pack_loose(opts);
1111 }
1112
1113 static int incremental_repack_auto_condition(void)
1114 {
1115 struct packed_git *p;
1116 int incremental_repack_auto_limit = 10;
1117 int count = 0;
1118
1119 prepare_repo_settings(the_repository);
1120 if (!the_repository->settings.core_multi_pack_index)
1121 return 0;
1122
1123 git_config_get_int("maintenance.incremental-repack.auto",
1124 &incremental_repack_auto_limit);
1125
1126 if (!incremental_repack_auto_limit)
1127 return 0;
1128 if (incremental_repack_auto_limit < 0)
1129 return 1;
1130
1131 for (p = get_packed_git(the_repository);
1132 count < incremental_repack_auto_limit && p;
1133 p = p->next) {
1134 if (!p->multi_pack_index)
1135 count++;
1136 }
1137
1138 return count >= incremental_repack_auto_limit;
1139 }
1140
1141 static int multi_pack_index_write(struct maintenance_run_opts *opts)
1142 {
1143 struct child_process child = CHILD_PROCESS_INIT;
1144
1145 child.git_cmd = 1;
1146 strvec_pushl(&child.args, "multi-pack-index", "write", NULL);
1147
1148 if (opts->quiet)
1149 strvec_push(&child.args, "--no-progress");
1150
1151 if (run_command(&child))
1152 return error(_("failed to write multi-pack-index"));
1153
1154 return 0;
1155 }
1156
1157 static int multi_pack_index_expire(struct maintenance_run_opts *opts)
1158 {
1159 struct child_process child = CHILD_PROCESS_INIT;
1160
1161 child.git_cmd = child.close_object_store = 1;
1162 strvec_pushl(&child.args, "multi-pack-index", "expire", NULL);
1163
1164 if (opts->quiet)
1165 strvec_push(&child.args, "--no-progress");
1166
1167 if (run_command(&child))
1168 return error(_("'git multi-pack-index expire' failed"));
1169
1170 return 0;
1171 }
1172
1173 #define TWO_GIGABYTES (INT32_MAX)
1174
1175 static off_t get_auto_pack_size(void)
1176 {
1177 /*
1178 * The "auto" value is special: we optimize for
1179 * one large pack-file (i.e. from a clone) and
1180 * expect the rest to be small and they can be
1181 * repacked quickly.
1182 *
1183 * The strategy we select here is to select a
1184 * size that is one more than the second largest
1185 * pack-file. This ensures that we will repack
1186 * at least two packs if there are three or more
1187 * packs.
1188 */
1189 off_t max_size = 0;
1190 off_t second_largest_size = 0;
1191 off_t result_size;
1192 struct packed_git *p;
1193 struct repository *r = the_repository;
1194
1195 reprepare_packed_git(r);
1196 for (p = get_all_packs(r); p; p = p->next) {
1197 if (p->pack_size > max_size) {
1198 second_largest_size = max_size;
1199 max_size = p->pack_size;
1200 } else if (p->pack_size > second_largest_size)
1201 second_largest_size = p->pack_size;
1202 }
1203
1204 result_size = second_largest_size + 1;
1205
1206 /* But limit ourselves to a batch size of 2g */
1207 if (result_size > TWO_GIGABYTES)
1208 result_size = TWO_GIGABYTES;
1209
1210 return result_size;
1211 }
1212
1213 static int multi_pack_index_repack(struct maintenance_run_opts *opts)
1214 {
1215 struct child_process child = CHILD_PROCESS_INIT;
1216
1217 child.git_cmd = child.close_object_store = 1;
1218 strvec_pushl(&child.args, "multi-pack-index", "repack", NULL);
1219
1220 if (opts->quiet)
1221 strvec_push(&child.args, "--no-progress");
1222
1223 strvec_pushf(&child.args, "--batch-size=%"PRIuMAX,
1224 (uintmax_t)get_auto_pack_size());
1225
1226 if (run_command(&child))
1227 return error(_("'git multi-pack-index repack' failed"));
1228
1229 return 0;
1230 }
1231
1232 static int maintenance_task_incremental_repack(struct maintenance_run_opts *opts)
1233 {
1234 prepare_repo_settings(the_repository);
1235 if (!the_repository->settings.core_multi_pack_index) {
1236 warning(_("skipping incremental-repack task because core.multiPackIndex is disabled"));
1237 return 0;
1238 }
1239
1240 if (multi_pack_index_write(opts))
1241 return 1;
1242 if (multi_pack_index_expire(opts))
1243 return 1;
1244 if (multi_pack_index_repack(opts))
1245 return 1;
1246 return 0;
1247 }
1248
1249 typedef int maintenance_task_fn(struct maintenance_run_opts *opts);
1250
1251 /*
1252 * An auto condition function returns 1 if the task should run
1253 * and 0 if the task should NOT run. See needs_to_gc() for an
1254 * example.
1255 */
1256 typedef int maintenance_auto_fn(void);
1257
1258 struct maintenance_task {
1259 const char *name;
1260 maintenance_task_fn *fn;
1261 maintenance_auto_fn *auto_condition;
1262 unsigned enabled:1;
1263
1264 enum schedule_priority schedule;
1265
1266 /* -1 if not selected. */
1267 int selected_order;
1268 };
1269
1270 enum maintenance_task_label {
1271 TASK_PREFETCH,
1272 TASK_LOOSE_OBJECTS,
1273 TASK_INCREMENTAL_REPACK,
1274 TASK_GC,
1275 TASK_COMMIT_GRAPH,
1276 TASK_PACK_REFS,
1277
1278 /* Leave as final value */
1279 TASK__COUNT
1280 };
1281
1282 static struct maintenance_task tasks[] = {
1283 [TASK_PREFETCH] = {
1284 "prefetch",
1285 maintenance_task_prefetch,
1286 },
1287 [TASK_LOOSE_OBJECTS] = {
1288 "loose-objects",
1289 maintenance_task_loose_objects,
1290 loose_object_auto_condition,
1291 },
1292 [TASK_INCREMENTAL_REPACK] = {
1293 "incremental-repack",
1294 maintenance_task_incremental_repack,
1295 incremental_repack_auto_condition,
1296 },
1297 [TASK_GC] = {
1298 "gc",
1299 maintenance_task_gc,
1300 need_to_gc,
1301 1,
1302 },
1303 [TASK_COMMIT_GRAPH] = {
1304 "commit-graph",
1305 maintenance_task_commit_graph,
1306 should_write_commit_graph,
1307 },
1308 [TASK_PACK_REFS] = {
1309 "pack-refs",
1310 maintenance_task_pack_refs,
1311 pack_refs_condition,
1312 },
1313 };
1314
1315 static int compare_tasks_by_selection(const void *a_, const void *b_)
1316 {
1317 const struct maintenance_task *a = a_;
1318 const struct maintenance_task *b = b_;
1319
1320 return b->selected_order - a->selected_order;
1321 }
1322
1323 static int maintenance_run_tasks(struct maintenance_run_opts *opts)
1324 {
1325 int i, found_selected = 0;
1326 int result = 0;
1327 struct lock_file lk;
1328 struct repository *r = the_repository;
1329 char *lock_path = xstrfmt("%s/maintenance", r->objects->odb->path);
1330
1331 if (hold_lock_file_for_update(&lk, lock_path, LOCK_NO_DEREF) < 0) {
1332 /*
1333 * Another maintenance command is running.
1334 *
1335 * If --auto was provided, then it is likely due to a
1336 * recursive process stack. Do not report an error in
1337 * that case.
1338 */
1339 if (!opts->auto_flag && !opts->quiet)
1340 warning(_("lock file '%s' exists, skipping maintenance"),
1341 lock_path);
1342 free(lock_path);
1343 return 0;
1344 }
1345 free(lock_path);
1346
1347 for (i = 0; !found_selected && i < TASK__COUNT; i++)
1348 found_selected = tasks[i].selected_order >= 0;
1349
1350 if (found_selected)
1351 QSORT(tasks, TASK__COUNT, compare_tasks_by_selection);
1352
1353 for (i = 0; i < TASK__COUNT; i++) {
1354 if (found_selected && tasks[i].selected_order < 0)
1355 continue;
1356
1357 if (!found_selected && !tasks[i].enabled)
1358 continue;
1359
1360 if (opts->auto_flag &&
1361 (!tasks[i].auto_condition ||
1362 !tasks[i].auto_condition()))
1363 continue;
1364
1365 if (opts->schedule && tasks[i].schedule < opts->schedule)
1366 continue;
1367
1368 trace2_region_enter("maintenance", tasks[i].name, r);
1369 if (tasks[i].fn(opts)) {
1370 error(_("task '%s' failed"), tasks[i].name);
1371 result = 1;
1372 }
1373 trace2_region_leave("maintenance", tasks[i].name, r);
1374 }
1375
1376 rollback_lock_file(&lk);
1377 return result;
1378 }
1379
1380 static void initialize_maintenance_strategy(void)
1381 {
1382 char *config_str;
1383
1384 if (git_config_get_string("maintenance.strategy", &config_str))
1385 return;
1386
1387 if (!strcasecmp(config_str, "incremental")) {
1388 tasks[TASK_GC].schedule = SCHEDULE_NONE;
1389 tasks[TASK_COMMIT_GRAPH].enabled = 1;
1390 tasks[TASK_COMMIT_GRAPH].schedule = SCHEDULE_HOURLY;
1391 tasks[TASK_PREFETCH].enabled = 1;
1392 tasks[TASK_PREFETCH].schedule = SCHEDULE_HOURLY;
1393 tasks[TASK_INCREMENTAL_REPACK].enabled = 1;
1394 tasks[TASK_INCREMENTAL_REPACK].schedule = SCHEDULE_DAILY;
1395 tasks[TASK_LOOSE_OBJECTS].enabled = 1;
1396 tasks[TASK_LOOSE_OBJECTS].schedule = SCHEDULE_DAILY;
1397 tasks[TASK_PACK_REFS].enabled = 1;
1398 tasks[TASK_PACK_REFS].schedule = SCHEDULE_WEEKLY;
1399 }
1400 }
1401
1402 static void initialize_task_config(int schedule)
1403 {
1404 int i;
1405 struct strbuf config_name = STRBUF_INIT;
1406 gc_config();
1407
1408 if (schedule)
1409 initialize_maintenance_strategy();
1410
1411 for (i = 0; i < TASK__COUNT; i++) {
1412 int config_value;
1413 char *config_str;
1414
1415 strbuf_reset(&config_name);
1416 strbuf_addf(&config_name, "maintenance.%s.enabled",
1417 tasks[i].name);
1418
1419 if (!git_config_get_bool(config_name.buf, &config_value))
1420 tasks[i].enabled = config_value;
1421
1422 strbuf_reset(&config_name);
1423 strbuf_addf(&config_name, "maintenance.%s.schedule",
1424 tasks[i].name);
1425
1426 if (!git_config_get_string(config_name.buf, &config_str)) {
1427 tasks[i].schedule = parse_schedule(config_str);
1428 free(config_str);
1429 }
1430 }
1431
1432 strbuf_release(&config_name);
1433 }
1434
1435 static int task_option_parse(const struct option *opt UNUSED,
1436 const char *arg, int unset)
1437 {
1438 int i, num_selected = 0;
1439 struct maintenance_task *task = NULL;
1440
1441 BUG_ON_OPT_NEG(unset);
1442
1443 for (i = 0; i < TASK__COUNT; i++) {
1444 if (tasks[i].selected_order >= 0)
1445 num_selected++;
1446 if (!strcasecmp(tasks[i].name, arg)) {
1447 task = &tasks[i];
1448 }
1449 }
1450
1451 if (!task) {
1452 error(_("'%s' is not a valid task"), arg);
1453 return 1;
1454 }
1455
1456 if (task->selected_order >= 0) {
1457 error(_("task '%s' cannot be selected multiple times"), arg);
1458 return 1;
1459 }
1460
1461 task->selected_order = num_selected + 1;
1462
1463 return 0;
1464 }
1465
1466 static int maintenance_run(int argc, const char **argv, const char *prefix)
1467 {
1468 int i;
1469 struct maintenance_run_opts opts;
1470 struct option builtin_maintenance_run_options[] = {
1471 OPT_BOOL(0, "auto", &opts.auto_flag,
1472 N_("run tasks based on the state of the repository")),
1473 OPT_CALLBACK(0, "schedule", &opts.schedule, N_("frequency"),
1474 N_("run tasks based on frequency"),
1475 maintenance_opt_schedule),
1476 OPT_BOOL(0, "quiet", &opts.quiet,
1477 N_("do not report progress or other information over stderr")),
1478 OPT_CALLBACK_F(0, "task", NULL, N_("task"),
1479 N_("run a specific task"),
1480 PARSE_OPT_NONEG, task_option_parse),
1481 OPT_END()
1482 };
1483 memset(&opts, 0, sizeof(opts));
1484
1485 opts.quiet = !isatty(2);
1486
1487 for (i = 0; i < TASK__COUNT; i++)
1488 tasks[i].selected_order = -1;
1489
1490 argc = parse_options(argc, argv, prefix,
1491 builtin_maintenance_run_options,
1492 builtin_maintenance_run_usage,
1493 PARSE_OPT_STOP_AT_NON_OPTION);
1494
1495 if (opts.auto_flag && opts.schedule)
1496 die(_("use at most one of --auto and --schedule=<frequency>"));
1497
1498 initialize_task_config(opts.schedule);
1499
1500 if (argc != 0)
1501 usage_with_options(builtin_maintenance_run_usage,
1502 builtin_maintenance_run_options);
1503 return maintenance_run_tasks(&opts);
1504 }
1505
1506 static char *get_maintpath(void)
1507 {
1508 struct strbuf sb = STRBUF_INIT;
1509 const char *p = the_repository->worktree ?
1510 the_repository->worktree : the_repository->gitdir;
1511
1512 strbuf_realpath(&sb, p, 1);
1513 return strbuf_detach(&sb, NULL);
1514 }
1515
1516 static char const * const builtin_maintenance_register_usage[] = {
1517 "git maintenance register [--config-file <path>]",
1518 NULL
1519 };
1520
1521 static int maintenance_register(int argc, const char **argv, const char *prefix)
1522 {
1523 char *config_file = NULL;
1524 struct option options[] = {
1525 OPT_STRING(0, "config-file", &config_file, N_("file"), N_("use given config file")),
1526 OPT_END(),
1527 };
1528 int found = 0;
1529 const char *key = "maintenance.repo";
1530 char *maintpath = get_maintpath();
1531 struct string_list_item *item;
1532 const struct string_list *list;
1533
1534 argc = parse_options(argc, argv, prefix, options,
1535 builtin_maintenance_register_usage, 0);
1536 if (argc)
1537 usage_with_options(builtin_maintenance_register_usage,
1538 options);
1539
1540 /* Disable foreground maintenance */
1541 git_config_set("maintenance.auto", "false");
1542
1543 /* Set maintenance strategy, if unset */
1544 if (git_config_get("maintenance.strategy"))
1545 git_config_set("maintenance.strategy", "incremental");
1546
1547 if (!git_config_get_string_multi(key, &list)) {
1548 for_each_string_list_item(item, list) {
1549 if (!strcmp(maintpath, item->string)) {
1550 found = 1;
1551 break;
1552 }
1553 }
1554 }
1555
1556 if (!found) {
1557 int rc;
1558 char *global_config_file = NULL;
1559
1560 if (!config_file) {
1561 global_config_file = git_global_config();
1562 config_file = global_config_file;
1563 }
1564 if (!config_file)
1565 die(_("$HOME not set"));
1566 rc = git_config_set_multivar_in_file_gently(
1567 config_file, "maintenance.repo", maintpath,
1568 CONFIG_REGEX_NONE, NULL, 0);
1569 free(global_config_file);
1570
1571 if (rc)
1572 die(_("unable to add '%s' value of '%s'"),
1573 key, maintpath);
1574 }
1575
1576 free(maintpath);
1577 return 0;
1578 }
1579
1580 static char const * const builtin_maintenance_unregister_usage[] = {
1581 "git maintenance unregister [--config-file <path>] [--force]",
1582 NULL
1583 };
1584
1585 static int maintenance_unregister(int argc, const char **argv, const char *prefix)
1586 {
1587 int force = 0;
1588 char *config_file = NULL;
1589 struct option options[] = {
1590 OPT_STRING(0, "config-file", &config_file, N_("file"), N_("use given config file")),
1591 OPT__FORCE(&force,
1592 N_("return success even if repository was not registered"),
1593 PARSE_OPT_NOCOMPLETE),
1594 OPT_END(),
1595 };
1596 const char *key = "maintenance.repo";
1597 char *maintpath = get_maintpath();
1598 int found = 0;
1599 struct string_list_item *item;
1600 const struct string_list *list;
1601 struct config_set cs = { { 0 } };
1602
1603 argc = parse_options(argc, argv, prefix, options,
1604 builtin_maintenance_unregister_usage, 0);
1605 if (argc)
1606 usage_with_options(builtin_maintenance_unregister_usage,
1607 options);
1608
1609 if (config_file) {
1610 git_configset_init(&cs);
1611 git_configset_add_file(&cs, config_file);
1612 }
1613 if (!(config_file
1614 ? git_configset_get_string_multi(&cs, key, &list)
1615 : git_config_get_string_multi(key, &list))) {
1616 for_each_string_list_item(item, list) {
1617 if (!strcmp(maintpath, item->string)) {
1618 found = 1;
1619 break;
1620 }
1621 }
1622 }
1623
1624 if (found) {
1625 int rc;
1626 char *global_config_file = NULL;
1627
1628 if (!config_file) {
1629 global_config_file = git_global_config();
1630 config_file = global_config_file;
1631 }
1632 if (!config_file)
1633 die(_("$HOME not set"));
1634 rc = git_config_set_multivar_in_file_gently(
1635 config_file, key, NULL, maintpath, NULL,
1636 CONFIG_FLAGS_MULTI_REPLACE | CONFIG_FLAGS_FIXED_VALUE);
1637 free(global_config_file);
1638
1639 if (rc &&
1640 (!force || rc == CONFIG_NOTHING_SET))
1641 die(_("unable to unset '%s' value of '%s'"),
1642 key, maintpath);
1643 } else if (!force) {
1644 die(_("repository '%s' is not registered"), maintpath);
1645 }
1646
1647 git_configset_clear(&cs);
1648 free(maintpath);
1649 return 0;
1650 }
1651
1652 static const char *get_frequency(enum schedule_priority schedule)
1653 {
1654 switch (schedule) {
1655 case SCHEDULE_HOURLY:
1656 return "hourly";
1657 case SCHEDULE_DAILY:
1658 return "daily";
1659 case SCHEDULE_WEEKLY:
1660 return "weekly";
1661 default:
1662 BUG("invalid schedule %d", schedule);
1663 }
1664 }
1665
1666 /*
1667 * get_schedule_cmd` reads the GIT_TEST_MAINT_SCHEDULER environment variable
1668 * to mock the schedulers that `git maintenance start` rely on.
1669 *
1670 * For test purpose, GIT_TEST_MAINT_SCHEDULER can be set to a comma-separated
1671 * list of colon-separated key/value pairs where each pair contains a scheduler
1672 * and its corresponding mock.
1673 *
1674 * * If $GIT_TEST_MAINT_SCHEDULER is not set, return false and leave the
1675 * arguments unmodified.
1676 *
1677 * * If $GIT_TEST_MAINT_SCHEDULER is set, return true.
1678 * In this case, the *cmd value is read as input.
1679 *
1680 * * if the input value *cmd is the key of one of the comma-separated list
1681 * item, then *is_available is set to true and *cmd is modified and becomes
1682 * the mock command.
1683 *
1684 * * if the input value *cmd isn’t the key of any of the comma-separated list
1685 * item, then *is_available is set to false.
1686 *
1687 * Ex.:
1688 * GIT_TEST_MAINT_SCHEDULER not set
1689 * +-------+-------------------------------------------------+
1690 * | Input | Output |
1691 * | *cmd | return code | *cmd | *is_available |
1692 * +-------+-------------+-------------------+---------------+
1693 * | "foo" | false | "foo" (unchanged) | (unchanged) |
1694 * +-------+-------------+-------------------+---------------+
1695 *
1696 * GIT_TEST_MAINT_SCHEDULER set to “foo:./mock_foo.sh,bar:./mock_bar.sh”
1697 * +-------+-------------------------------------------------+
1698 * | Input | Output |
1699 * | *cmd | return code | *cmd | *is_available |
1700 * +-------+-------------+-------------------+---------------+
1701 * | "foo" | true | "./mock.foo.sh" | true |
1702 * | "qux" | true | "qux" (unchanged) | false |
1703 * +-------+-------------+-------------------+---------------+
1704 */
1705 static int get_schedule_cmd(const char **cmd, int *is_available)
1706 {
1707 char *testing = xstrdup_or_null(getenv("GIT_TEST_MAINT_SCHEDULER"));
1708 struct string_list_item *item;
1709 struct string_list list = STRING_LIST_INIT_NODUP;
1710
1711 if (!testing)
1712 return 0;
1713
1714 if (is_available)
1715 *is_available = 0;
1716
1717 string_list_split_in_place(&list, testing, ",", -1);
1718 for_each_string_list_item(item, &list) {
1719 struct string_list pair = STRING_LIST_INIT_NODUP;
1720
1721 if (string_list_split_in_place(&pair, item->string, ":", 2) != 2)
1722 continue;
1723
1724 if (!strcmp(*cmd, pair.items[0].string)) {
1725 *cmd = pair.items[1].string;
1726 if (is_available)
1727 *is_available = 1;
1728 string_list_clear(&list, 0);
1729 UNLEAK(testing);
1730 return 1;
1731 }
1732 }
1733
1734 string_list_clear(&list, 0);
1735 free(testing);
1736 return 1;
1737 }
1738
1739 static int get_random_minute(void)
1740 {
1741 /* Use a static value when under tests. */
1742 if (getenv("GIT_TEST_MAINT_SCHEDULER"))
1743 return 13;
1744
1745 return git_rand() % 60;
1746 }
1747
1748 static int is_launchctl_available(void)
1749 {
1750 const char *cmd = "launchctl";
1751 int is_available;
1752 if (get_schedule_cmd(&cmd, &is_available))
1753 return is_available;
1754
1755 #ifdef __APPLE__
1756 return 1;
1757 #else
1758 return 0;
1759 #endif
1760 }
1761
1762 static char *launchctl_service_name(const char *frequency)
1763 {
1764 struct strbuf label = STRBUF_INIT;
1765 strbuf_addf(&label, "org.git-scm.git.%s", frequency);
1766 return strbuf_detach(&label, NULL);
1767 }
1768
1769 static char *launchctl_service_filename(const char *name)
1770 {
1771 char *expanded;
1772 struct strbuf filename = STRBUF_INIT;
1773 strbuf_addf(&filename, "~/Library/LaunchAgents/%s.plist", name);
1774
1775 expanded = interpolate_path(filename.buf, 1);
1776 if (!expanded)
1777 die(_("failed to expand path '%s'"), filename.buf);
1778
1779 strbuf_release(&filename);
1780 return expanded;
1781 }
1782
1783 static char *launchctl_get_uid(void)
1784 {
1785 return xstrfmt("gui/%d", getuid());
1786 }
1787
1788 static int launchctl_boot_plist(int enable, const char *filename)
1789 {
1790 const char *cmd = "launchctl";
1791 int result;
1792 struct child_process child = CHILD_PROCESS_INIT;
1793 char *uid = launchctl_get_uid();
1794
1795 get_schedule_cmd(&cmd, NULL);
1796 strvec_split(&child.args, cmd);
1797 strvec_pushl(&child.args, enable ? "bootstrap" : "bootout", uid,
1798 filename, NULL);
1799
1800 child.no_stderr = 1;
1801 child.no_stdout = 1;
1802
1803 if (start_command(&child))
1804 die(_("failed to start launchctl"));
1805
1806 result = finish_command(&child);
1807
1808 free(uid);
1809 return result;
1810 }
1811
1812 static int launchctl_remove_plist(enum schedule_priority schedule)
1813 {
1814 const char *frequency = get_frequency(schedule);
1815 char *name = launchctl_service_name(frequency);
1816 char *filename = launchctl_service_filename(name);
1817 int result = launchctl_boot_plist(0, filename);
1818 unlink(filename);
1819 free(filename);
1820 free(name);
1821 return result;
1822 }
1823
1824 static int launchctl_remove_plists(void)
1825 {
1826 return launchctl_remove_plist(SCHEDULE_HOURLY) ||
1827 launchctl_remove_plist(SCHEDULE_DAILY) ||
1828 launchctl_remove_plist(SCHEDULE_WEEKLY);
1829 }
1830
1831 static int launchctl_list_contains_plist(const char *name, const char *cmd)
1832 {
1833 struct child_process child = CHILD_PROCESS_INIT;
1834
1835 strvec_split(&child.args, cmd);
1836 strvec_pushl(&child.args, "list", name, NULL);
1837
1838 child.no_stderr = 1;
1839 child.no_stdout = 1;
1840
1841 if (start_command(&child))
1842 die(_("failed to start launchctl"));
1843
1844 /* Returns failure if 'name' doesn't exist. */
1845 return !finish_command(&child);
1846 }
1847
1848 static int launchctl_schedule_plist(const char *exec_path, enum schedule_priority schedule)
1849 {
1850 int i, fd;
1851 const char *preamble, *repeat;
1852 const char *frequency = get_frequency(schedule);
1853 char *name = launchctl_service_name(frequency);
1854 char *filename = launchctl_service_filename(name);
1855 struct lock_file lk = LOCK_INIT;
1856 static unsigned long lock_file_timeout_ms = ULONG_MAX;
1857 struct strbuf plist = STRBUF_INIT, plist2 = STRBUF_INIT;
1858 struct stat st;
1859 const char *cmd = "launchctl";
1860 int minute = get_random_minute();
1861
1862 get_schedule_cmd(&cmd, NULL);
1863 preamble = "<?xml version=\"1.0\"?>\n"
1864 "<!DOCTYPE plist PUBLIC \"-//Apple//DTD PLIST 1.0//EN\" \"http://www.apple.com/DTDs/PropertyList-1.0.dtd\">\n"
1865 "<plist version=\"1.0\">"
1866 "<dict>\n"
1867 "<key>Label</key><string>%s</string>\n"
1868 "<key>ProgramArguments</key>\n"
1869 "<array>\n"
1870 "<string>%s/git</string>\n"
1871 "<string>--exec-path=%s</string>\n"
1872 "<string>for-each-repo</string>\n"
1873 "<string>--config=maintenance.repo</string>\n"
1874 "<string>maintenance</string>\n"
1875 "<string>run</string>\n"
1876 "<string>--schedule=%s</string>\n"
1877 "</array>\n"
1878 "<key>StartCalendarInterval</key>\n"
1879 "<array>\n";
1880 strbuf_addf(&plist, preamble, name, exec_path, exec_path, frequency);
1881
1882 switch (schedule) {
1883 case SCHEDULE_HOURLY:
1884 repeat = "<dict>\n"
1885 "<key>Hour</key><integer>%d</integer>\n"
1886 "<key>Minute</key><integer>%d</integer>\n"
1887 "</dict>\n";
1888 for (i = 1; i <= 23; i++)
1889 strbuf_addf(&plist, repeat, i, minute);
1890 break;
1891
1892 case SCHEDULE_DAILY:
1893 repeat = "<dict>\n"
1894 "<key>Day</key><integer>%d</integer>\n"
1895 "<key>Hour</key><integer>0</integer>\n"
1896 "<key>Minute</key><integer>%d</integer>\n"
1897 "</dict>\n";
1898 for (i = 1; i <= 6; i++)
1899 strbuf_addf(&plist, repeat, i, minute);
1900 break;
1901
1902 case SCHEDULE_WEEKLY:
1903 strbuf_addf(&plist,
1904 "<dict>\n"
1905 "<key>Day</key><integer>0</integer>\n"
1906 "<key>Hour</key><integer>0</integer>\n"
1907 "<key>Minute</key><integer>%d</integer>\n"
1908 "</dict>\n",
1909 minute);
1910 break;
1911
1912 default:
1913 /* unreachable */
1914 break;
1915 }
1916 strbuf_addstr(&plist, "</array>\n</dict>\n</plist>\n");
1917
1918 if (safe_create_leading_directories(filename))
1919 die(_("failed to create directories for '%s'"), filename);
1920
1921 if ((long)lock_file_timeout_ms < 0 &&
1922 git_config_get_ulong("gc.launchctlplistlocktimeoutms",
1923 &lock_file_timeout_ms))
1924 lock_file_timeout_ms = 150;
1925
1926 fd = hold_lock_file_for_update_timeout(&lk, filename, LOCK_DIE_ON_ERROR,
1927 lock_file_timeout_ms);
1928
1929 /*
1930 * Does this file already exist? With the intended contents? Is it
1931 * registered already? Then it does not need to be re-registered.
1932 */
1933 if (!stat(filename, &st) && st.st_size == plist.len &&
1934 strbuf_read_file(&plist2, filename, plist.len) == plist.len &&
1935 !strbuf_cmp(&plist, &plist2) &&
1936 launchctl_list_contains_plist(name, cmd))
1937 rollback_lock_file(&lk);
1938 else {
1939 if (write_in_full(fd, plist.buf, plist.len) < 0 ||
1940 commit_lock_file(&lk))
1941 die_errno(_("could not write '%s'"), filename);
1942
1943 /* bootout might fail if not already running, so ignore */
1944 launchctl_boot_plist(0, filename);
1945 if (launchctl_boot_plist(1, filename))
1946 die(_("failed to bootstrap service %s"), filename);
1947 }
1948
1949 free(filename);
1950 free(name);
1951 strbuf_release(&plist);
1952 strbuf_release(&plist2);
1953 return 0;
1954 }
1955
1956 static int launchctl_add_plists(void)
1957 {
1958 const char *exec_path = git_exec_path();
1959
1960 return launchctl_schedule_plist(exec_path, SCHEDULE_HOURLY) ||
1961 launchctl_schedule_plist(exec_path, SCHEDULE_DAILY) ||
1962 launchctl_schedule_plist(exec_path, SCHEDULE_WEEKLY);
1963 }
1964
1965 static int launchctl_update_schedule(int run_maintenance, int fd UNUSED)
1966 {
1967 if (run_maintenance)
1968 return launchctl_add_plists();
1969 else
1970 return launchctl_remove_plists();
1971 }
1972
1973 static int is_schtasks_available(void)
1974 {
1975 const char *cmd = "schtasks";
1976 int is_available;
1977 if (get_schedule_cmd(&cmd, &is_available))
1978 return is_available;
1979
1980 #ifdef GIT_WINDOWS_NATIVE
1981 return 1;
1982 #else
1983 return 0;
1984 #endif
1985 }
1986
1987 static char *schtasks_task_name(const char *frequency)
1988 {
1989 struct strbuf label = STRBUF_INIT;
1990 strbuf_addf(&label, "Git Maintenance (%s)", frequency);
1991 return strbuf_detach(&label, NULL);
1992 }
1993
1994 static int schtasks_remove_task(enum schedule_priority schedule)
1995 {
1996 const char *cmd = "schtasks";
1997 struct child_process child = CHILD_PROCESS_INIT;
1998 const char *frequency = get_frequency(schedule);
1999 char *name = schtasks_task_name(frequency);
2000
2001 get_schedule_cmd(&cmd, NULL);
2002 strvec_split(&child.args, cmd);
2003 strvec_pushl(&child.args, "/delete", "/tn", name, "/f", NULL);
2004 free(name);
2005
2006 return run_command(&child);
2007 }
2008
2009 static int schtasks_remove_tasks(void)
2010 {
2011 return schtasks_remove_task(SCHEDULE_HOURLY) ||
2012 schtasks_remove_task(SCHEDULE_DAILY) ||
2013 schtasks_remove_task(SCHEDULE_WEEKLY);
2014 }
2015
2016 static int schtasks_schedule_task(const char *exec_path, enum schedule_priority schedule)
2017 {
2018 const char *cmd = "schtasks";
2019 int result;
2020 struct child_process child = CHILD_PROCESS_INIT;
2021 const char *xml;
2022 struct tempfile *tfile;
2023 const char *frequency = get_frequency(schedule);
2024 char *name = schtasks_task_name(frequency);
2025 struct strbuf tfilename = STRBUF_INIT;
2026 int minute = get_random_minute();
2027
2028 get_schedule_cmd(&cmd, NULL);
2029
2030 strbuf_addf(&tfilename, "%s/schedule_%s_XXXXXX",
2031 get_git_common_dir(), frequency);
2032 tfile = xmks_tempfile(tfilename.buf);
2033 strbuf_release(&tfilename);
2034
2035 if (!fdopen_tempfile(tfile, "w"))
2036 die(_("failed to create temp xml file"));
2037
2038 xml = "<?xml version=\"1.0\" ?>\n"
2039 "<Task version=\"1.4\" xmlns=\"http://schemas.microsoft.com/windows/2004/02/mit/task\">\n"
2040 "<Triggers>\n"
2041 "<CalendarTrigger>\n";
2042 fputs(xml, tfile->fp);
2043
2044 switch (schedule) {
2045 case SCHEDULE_HOURLY:
2046 fprintf(tfile->fp,
2047 "<StartBoundary>2020-01-01T01:%02d:00</StartBoundary>\n"
2048 "<Enabled>true</Enabled>\n"
2049 "<ScheduleByDay>\n"
2050 "<DaysInterval>1</DaysInterval>\n"
2051 "</ScheduleByDay>\n"
2052 "<Repetition>\n"
2053 "<Interval>PT1H</Interval>\n"
2054 "<Duration>PT23H</Duration>\n"
2055 "<StopAtDurationEnd>false</StopAtDurationEnd>\n"
2056 "</Repetition>\n",
2057 minute);
2058 break;
2059
2060 case SCHEDULE_DAILY:
2061 fprintf(tfile->fp,
2062 "<StartBoundary>2020-01-01T00:%02d:00</StartBoundary>\n"
2063 "<Enabled>true</Enabled>\n"
2064 "<ScheduleByWeek>\n"
2065 "<DaysOfWeek>\n"
2066 "<Monday />\n"
2067 "<Tuesday />\n"
2068 "<Wednesday />\n"
2069 "<Thursday />\n"
2070 "<Friday />\n"
2071 "<Saturday />\n"
2072 "</DaysOfWeek>\n"
2073 "<WeeksInterval>1</WeeksInterval>\n"
2074 "</ScheduleByWeek>\n",
2075 minute);
2076 break;
2077
2078 case SCHEDULE_WEEKLY:
2079 fprintf(tfile->fp,
2080 "<StartBoundary>2020-01-01T00:%02d:00</StartBoundary>\n"
2081 "<Enabled>true</Enabled>\n"
2082 "<ScheduleByWeek>\n"
2083 "<DaysOfWeek>\n"
2084 "<Sunday />\n"
2085 "</DaysOfWeek>\n"
2086 "<WeeksInterval>1</WeeksInterval>\n"
2087 "</ScheduleByWeek>\n",
2088 minute);
2089 break;
2090
2091 default:
2092 break;
2093 }
2094
2095 xml = "</CalendarTrigger>\n"
2096 "</Triggers>\n"
2097 "<Principals>\n"
2098 "<Principal id=\"Author\">\n"
2099 "<LogonType>InteractiveToken</LogonType>\n"
2100 "<RunLevel>LeastPrivilege</RunLevel>\n"
2101 "</Principal>\n"
2102 "</Principals>\n"
2103 "<Settings>\n"
2104 "<MultipleInstancesPolicy>IgnoreNew</MultipleInstancesPolicy>\n"
2105 "<Enabled>true</Enabled>\n"
2106 "<Hidden>true</Hidden>\n"
2107 "<UseUnifiedSchedulingEngine>true</UseUnifiedSchedulingEngine>\n"
2108 "<WakeToRun>false</WakeToRun>\n"
2109 "<ExecutionTimeLimit>PT72H</ExecutionTimeLimit>\n"
2110 "<Priority>7</Priority>\n"
2111 "</Settings>\n"
2112 "<Actions Context=\"Author\">\n"
2113 "<Exec>\n"
2114 "<Command>\"%s\\headless-git.exe\"</Command>\n"
2115 "<Arguments>--exec-path=\"%s\" for-each-repo --config=maintenance.repo maintenance run --schedule=%s</Arguments>\n"
2116 "</Exec>\n"
2117 "</Actions>\n"
2118 "</Task>\n";
2119 fprintf(tfile->fp, xml, exec_path, exec_path, frequency);
2120 strvec_split(&child.args, cmd);
2121 strvec_pushl(&child.args, "/create", "/tn", name, "/f", "/xml",
2122 get_tempfile_path(tfile), NULL);
2123 close_tempfile_gently(tfile);
2124
2125 child.no_stdout = 1;
2126 child.no_stderr = 1;
2127
2128 if (start_command(&child))
2129 die(_("failed to start schtasks"));
2130 result = finish_command(&child);
2131
2132 delete_tempfile(&tfile);
2133 free(name);
2134 return result;
2135 }
2136
2137 static int schtasks_schedule_tasks(void)
2138 {
2139 const char *exec_path = git_exec_path();
2140
2141 return schtasks_schedule_task(exec_path, SCHEDULE_HOURLY) ||
2142 schtasks_schedule_task(exec_path, SCHEDULE_DAILY) ||
2143 schtasks_schedule_task(exec_path, SCHEDULE_WEEKLY);
2144 }
2145
2146 static int schtasks_update_schedule(int run_maintenance, int fd UNUSED)
2147 {
2148 if (run_maintenance)
2149 return schtasks_schedule_tasks();
2150 else
2151 return schtasks_remove_tasks();
2152 }
2153
2154 MAYBE_UNUSED
2155 static int check_crontab_process(const char *cmd)
2156 {
2157 struct child_process child = CHILD_PROCESS_INIT;
2158
2159 strvec_split(&child.args, cmd);
2160 strvec_push(&child.args, "-l");
2161 child.no_stdin = 1;
2162 child.no_stdout = 1;
2163 child.no_stderr = 1;
2164 child.silent_exec_failure = 1;
2165
2166 if (start_command(&child))
2167 return 0;
2168 /* Ignore exit code, as an empty crontab will return error. */
2169 finish_command(&child);
2170 return 1;
2171 }
2172
2173 static int is_crontab_available(void)
2174 {
2175 const char *cmd = "crontab";
2176 int is_available;
2177
2178 if (get_schedule_cmd(&cmd, &is_available))
2179 return is_available;
2180
2181 #ifdef __APPLE__
2182 /*
2183 * macOS has cron, but it requires special permissions and will
2184 * create a UI alert when attempting to run this command.
2185 */
2186 return 0;
2187 #else
2188 return check_crontab_process(cmd);
2189 #endif
2190 }
2191
2192 #define BEGIN_LINE "# BEGIN GIT MAINTENANCE SCHEDULE"
2193 #define END_LINE "# END GIT MAINTENANCE SCHEDULE"
2194
2195 static int crontab_update_schedule(int run_maintenance, int fd)
2196 {
2197 const char *cmd = "crontab";
2198 int result = 0;
2199 int in_old_region = 0;
2200 struct child_process crontab_list = CHILD_PROCESS_INIT;
2201 struct child_process crontab_edit = CHILD_PROCESS_INIT;
2202 FILE *cron_list, *cron_in;
2203 struct strbuf line = STRBUF_INIT;
2204 struct tempfile *tmpedit = NULL;
2205 int minute = get_random_minute();
2206
2207 get_schedule_cmd(&cmd, NULL);
2208 strvec_split(&crontab_list.args, cmd);
2209 strvec_push(&crontab_list.args, "-l");
2210 crontab_list.in = -1;
2211 crontab_list.out = dup(fd);
2212 crontab_list.git_cmd = 0;
2213
2214 if (start_command(&crontab_list))
2215 return error(_("failed to run 'crontab -l'; your system might not support 'cron'"));
2216
2217 /* Ignore exit code, as an empty crontab will return error. */
2218 finish_command(&crontab_list);
2219
2220 tmpedit = mks_tempfile_t(".git_cron_edit_tmpXXXXXX");
2221 if (!tmpedit) {
2222 result = error(_("failed to create crontab temporary file"));
2223 goto out;
2224 }
2225 cron_in = fdopen_tempfile(tmpedit, "w");
2226 if (!cron_in) {
2227 result = error(_("failed to open temporary file"));
2228 goto out;
2229 }
2230
2231 /*
2232 * Read from the .lock file, filtering out the old
2233 * schedule while appending the new schedule.
2234 */
2235 cron_list = fdopen(fd, "r");
2236 rewind(cron_list);
2237
2238 while (!strbuf_getline_lf(&line, cron_list)) {
2239 if (!in_old_region && !strcmp(line.buf, BEGIN_LINE))
2240 in_old_region = 1;
2241 else if (in_old_region && !strcmp(line.buf, END_LINE))
2242 in_old_region = 0;
2243 else if (!in_old_region)
2244 fprintf(cron_in, "%s\n", line.buf);
2245 }
2246 strbuf_release(&line);
2247
2248 if (run_maintenance) {
2249 struct strbuf line_format = STRBUF_INIT;
2250 const char *exec_path = git_exec_path();
2251
2252 fprintf(cron_in, "%s\n", BEGIN_LINE);
2253 fprintf(cron_in,
2254 "# The following schedule was created by Git\n");
2255 fprintf(cron_in, "# Any edits made in this region might be\n");
2256 fprintf(cron_in,
2257 "# replaced in the future by a Git command.\n\n");
2258
2259 strbuf_addf(&line_format,
2260 "%%d %%s * * %%s \"%s/git\" --exec-path=\"%s\" for-each-repo --config=maintenance.repo maintenance run --schedule=%%s\n",
2261 exec_path, exec_path);
2262 fprintf(cron_in, line_format.buf, minute, "1-23", "*", "hourly");
2263 fprintf(cron_in, line_format.buf, minute, "0", "1-6", "daily");
2264 fprintf(cron_in, line_format.buf, minute, "0", "0", "weekly");
2265 strbuf_release(&line_format);
2266
2267 fprintf(cron_in, "\n%s\n", END_LINE);
2268 }
2269
2270 fflush(cron_in);
2271
2272 strvec_split(&crontab_edit.args, cmd);
2273 strvec_push(&crontab_edit.args, get_tempfile_path(tmpedit));
2274 crontab_edit.git_cmd = 0;
2275
2276 if (start_command(&crontab_edit)) {
2277 result = error(_("failed to run 'crontab'; your system might not support 'cron'"));
2278 goto out;
2279 }
2280
2281 if (finish_command(&crontab_edit))
2282 result = error(_("'crontab' died"));
2283 else
2284 fclose(cron_list);
2285 out:
2286 delete_tempfile(&tmpedit);
2287 return result;
2288 }
2289
2290 static int real_is_systemd_timer_available(void)
2291 {
2292 struct child_process child = CHILD_PROCESS_INIT;
2293
2294 strvec_pushl(&child.args, "systemctl", "--user", "list-timers", NULL);
2295 child.no_stdin = 1;
2296 child.no_stdout = 1;
2297 child.no_stderr = 1;
2298 child.silent_exec_failure = 1;
2299
2300 if (start_command(&child))
2301 return 0;
2302 if (finish_command(&child))
2303 return 0;
2304 return 1;
2305 }
2306
2307 static int is_systemd_timer_available(void)
2308 {
2309 const char *cmd = "systemctl";
2310 int is_available;
2311
2312 if (get_schedule_cmd(&cmd, &is_available))
2313 return is_available;
2314
2315 return real_is_systemd_timer_available();
2316 }
2317
2318 static char *xdg_config_home_systemd(const char *filename)
2319 {
2320 return xdg_config_home_for("systemd/user", filename);
2321 }
2322
2323 #define SYSTEMD_UNIT_FORMAT "git-maintenance@%s.%s"
2324
2325 static int systemd_timer_delete_timer_file(enum schedule_priority priority)
2326 {
2327 int ret = 0;
2328 const char *frequency = get_frequency(priority);
2329 char *local_timer_name = xstrfmt(SYSTEMD_UNIT_FORMAT, frequency, "timer");
2330 char *filename = xdg_config_home_systemd(local_timer_name);
2331
2332 if (unlink(filename) && !is_missing_file_error(errno))
2333 ret = error_errno(_("failed to delete '%s'"), filename);
2334
2335 free(filename);
2336 free(local_timer_name);
2337 return ret;
2338 }
2339
2340 static int systemd_timer_delete_service_template(void)
2341 {
2342 int ret = 0;
2343 char *local_service_name = xstrfmt(SYSTEMD_UNIT_FORMAT, "", "service");
2344 char *filename = xdg_config_home_systemd(local_service_name);
2345 if (unlink(filename) && !is_missing_file_error(errno))
2346 ret = error_errno(_("failed to delete '%s'"), filename);
2347
2348 free(filename);
2349 free(local_service_name);
2350 return ret;
2351 }
2352
2353 /*
2354 * Write the schedule information into a git-maintenance@<schedule>.timer
2355 * file using a custom minute. This timer file cannot use the templating
2356 * system, so we generate a specific file for each.
2357 */
2358 static int systemd_timer_write_timer_file(enum schedule_priority schedule,
2359 int minute)
2360 {
2361 int res = -1;
2362 char *filename;
2363 FILE *file;
2364 const char *unit;
2365 char *schedule_pattern = NULL;
2366 const char *frequency = get_frequency(schedule);
2367 char *local_timer_name = xstrfmt(SYSTEMD_UNIT_FORMAT, frequency, "timer");
2368
2369 filename = xdg_config_home_systemd(local_timer_name);
2370
2371 if (safe_create_leading_directories(filename)) {
2372 error(_("failed to create directories for '%s'"), filename);
2373 goto error;
2374 }
2375 file = fopen_or_warn(filename, "w");
2376 if (!file)
2377 goto error;
2378
2379 switch (schedule) {
2380 case SCHEDULE_HOURLY:
2381 schedule_pattern = xstrfmt("*-*-* 1..23:%02d:00", minute);
2382 break;
2383
2384 case SCHEDULE_DAILY:
2385 schedule_pattern = xstrfmt("Tue..Sun *-*-* 0:%02d:00", minute);
2386 break;
2387
2388 case SCHEDULE_WEEKLY:
2389 schedule_pattern = xstrfmt("Mon 0:%02d:00", minute);
2390 break;
2391
2392 default:
2393 BUG("Unhandled schedule_priority");
2394 }
2395
2396 unit = "# This file was created and is maintained by Git.\n"
2397 "# Any edits made in this file might be replaced in the future\n"
2398 "# by a Git command.\n"
2399 "\n"
2400 "[Unit]\n"
2401 "Description=Optimize Git repositories data\n"
2402 "\n"
2403 "[Timer]\n"
2404 "OnCalendar=%s\n"
2405 "Persistent=true\n"
2406 "\n"
2407 "[Install]\n"
2408 "WantedBy=timers.target\n";
2409 if (fprintf(file, unit, schedule_pattern) < 0) {
2410 error(_("failed to write to '%s'"), filename);
2411 fclose(file);
2412 goto error;
2413 }
2414 if (fclose(file) == EOF) {
2415 error_errno(_("failed to flush '%s'"), filename);
2416 goto error;
2417 }
2418
2419 res = 0;
2420
2421 error:
2422 free(schedule_pattern);
2423 free(local_timer_name);
2424 free(filename);
2425 return res;
2426 }
2427
2428 /*
2429 * No matter the schedule, we use the same service and can make use of the
2430 * templating system. When installing git-maintenance@<schedule>.timer,
2431 * systemd will notice that git-maintenance@.service exists as a template
2432 * and will use this file and insert the <schedule> into the template at
2433 * the position of "%i".
2434 */
2435 static int systemd_timer_write_service_template(const char *exec_path)
2436 {
2437 int res = -1;
2438 char *filename;
2439 FILE *file;
2440 const char *unit;
2441 char *local_service_name = xstrfmt(SYSTEMD_UNIT_FORMAT, "", "service");
2442
2443 filename = xdg_config_home_systemd(local_service_name);
2444 if (safe_create_leading_directories(filename)) {
2445 error(_("failed to create directories for '%s'"), filename);
2446 goto error;
2447 }
2448 file = fopen_or_warn(filename, "w");
2449 if (!file)
2450 goto error;
2451
2452 unit = "# This file was created and is maintained by Git.\n"
2453 "# Any edits made in this file might be replaced in the future\n"
2454 "# by a Git command.\n"
2455 "\n"
2456 "[Unit]\n"
2457 "Description=Optimize Git repositories data\n"
2458 "\n"
2459 "[Service]\n"
2460 "Type=oneshot\n"
2461 "ExecStart=\"%s/git\" --exec-path=\"%s\" for-each-repo --config=maintenance.repo maintenance run --schedule=%%i\n"
2462 "LockPersonality=yes\n"
2463 "MemoryDenyWriteExecute=yes\n"
2464 "NoNewPrivileges=yes\n"
2465 "RestrictAddressFamilies=AF_UNIX AF_INET AF_INET6 AF_VSOCK\n"
2466 "RestrictNamespaces=yes\n"
2467 "RestrictRealtime=yes\n"
2468 "RestrictSUIDSGID=yes\n"
2469 "SystemCallArchitectures=native\n"
2470 "SystemCallFilter=@system-service\n";
2471 if (fprintf(file, unit, exec_path, exec_path) < 0) {
2472 error(_("failed to write to '%s'"), filename);
2473 fclose(file);
2474 goto error;
2475 }
2476 if (fclose(file) == EOF) {
2477 error_errno(_("failed to flush '%s'"), filename);
2478 goto error;
2479 }
2480
2481 res = 0;
2482
2483 error:
2484 free(local_service_name);
2485 free(filename);
2486 return res;
2487 }
2488
2489 static int systemd_timer_enable_unit(int enable,
2490 enum schedule_priority schedule,
2491 int minute)
2492 {
2493 const char *cmd = "systemctl";
2494 struct child_process child = CHILD_PROCESS_INIT;
2495 const char *frequency = get_frequency(schedule);
2496
2497 /*
2498 * Disabling the systemd unit while it is already disabled makes
2499 * systemctl print an error.
2500 * Let's ignore it since it means we already are in the expected state:
2501 * the unit is disabled.
2502 *
2503 * On the other hand, enabling a systemd unit which is already enabled
2504 * produces no error.
2505 */
2506 if (!enable)
2507 child.no_stderr = 1;
2508 else if (systemd_timer_write_timer_file(schedule, minute))
2509 return -1;
2510
2511 get_schedule_cmd(&cmd, NULL);
2512 strvec_split(&child.args, cmd);
2513 strvec_pushl(&child.args, "--user", enable ? "enable" : "disable",
2514 "--now", NULL);
2515 strvec_pushf(&child.args, SYSTEMD_UNIT_FORMAT, frequency, "timer");
2516
2517 if (start_command(&child))
2518 return error(_("failed to start systemctl"));
2519 if (finish_command(&child))
2520 /*
2521 * Disabling an already disabled systemd unit makes
2522 * systemctl fail.
2523 * Let's ignore this failure.
2524 *
2525 * Enabling an enabled systemd unit doesn't fail.
2526 */
2527 if (enable)
2528 return error(_("failed to run systemctl"));
2529 return 0;
2530 }
2531
2532 /*
2533 * A previous version of Git wrote the timer units as template files.
2534 * Clean these up, if they exist.
2535 */
2536 static void systemd_timer_delete_stale_timer_templates(void)
2537 {
2538 char *timer_template_name = xstrfmt(SYSTEMD_UNIT_FORMAT, "", "timer");
2539 char *filename = xdg_config_home_systemd(timer_template_name);
2540
2541 if (unlink(filename) && !is_missing_file_error(errno))
2542 warning(_("failed to delete '%s'"), filename);
2543
2544 free(filename);
2545 free(timer_template_name);
2546 }
2547
2548 static int systemd_timer_delete_unit_files(void)
2549 {
2550 systemd_timer_delete_stale_timer_templates();
2551
2552 /* Purposefully not short-circuited to make sure all are called. */
2553 return systemd_timer_delete_timer_file(SCHEDULE_HOURLY) |
2554 systemd_timer_delete_timer_file(SCHEDULE_DAILY) |
2555 systemd_timer_delete_timer_file(SCHEDULE_WEEKLY) |
2556 systemd_timer_delete_service_template();
2557 }
2558
2559 static int systemd_timer_delete_units(void)
2560 {
2561 int minute = get_random_minute();
2562 /* Purposefully not short-circuited to make sure all are called. */
2563 return systemd_timer_enable_unit(0, SCHEDULE_HOURLY, minute) |
2564 systemd_timer_enable_unit(0, SCHEDULE_DAILY, minute) |
2565 systemd_timer_enable_unit(0, SCHEDULE_WEEKLY, minute) |
2566 systemd_timer_delete_unit_files();
2567 }
2568
2569 static int systemd_timer_setup_units(void)
2570 {
2571 int minute = get_random_minute();
2572 const char *exec_path = git_exec_path();
2573
2574 int ret = systemd_timer_write_service_template(exec_path) ||
2575 systemd_timer_enable_unit(1, SCHEDULE_HOURLY, minute) ||
2576 systemd_timer_enable_unit(1, SCHEDULE_DAILY, minute) ||
2577 systemd_timer_enable_unit(1, SCHEDULE_WEEKLY, minute);
2578
2579 if (ret)
2580 systemd_timer_delete_units();
2581 else
2582 systemd_timer_delete_stale_timer_templates();
2583
2584 return ret;
2585 }
2586
2587 static int systemd_timer_update_schedule(int run_maintenance, int fd UNUSED)
2588 {
2589 if (run_maintenance)
2590 return systemd_timer_setup_units();
2591 else
2592 return systemd_timer_delete_units();
2593 }
2594
2595 enum scheduler {
2596 SCHEDULER_INVALID = -1,
2597 SCHEDULER_AUTO,
2598 SCHEDULER_CRON,
2599 SCHEDULER_SYSTEMD,
2600 SCHEDULER_LAUNCHCTL,
2601 SCHEDULER_SCHTASKS,
2602 };
2603
2604 static const struct {
2605 const char *name;
2606 int (*is_available)(void);
2607 int (*update_schedule)(int run_maintenance, int fd);
2608 } scheduler_fn[] = {
2609 [SCHEDULER_CRON] = {
2610 .name = "crontab",
2611 .is_available = is_crontab_available,
2612 .update_schedule = crontab_update_schedule,
2613 },
2614 [SCHEDULER_SYSTEMD] = {
2615 .name = "systemctl",
2616 .is_available = is_systemd_timer_available,
2617 .update_schedule = systemd_timer_update_schedule,
2618 },
2619 [SCHEDULER_LAUNCHCTL] = {
2620 .name = "launchctl",
2621 .is_available = is_launchctl_available,
2622 .update_schedule = launchctl_update_schedule,
2623 },
2624 [SCHEDULER_SCHTASKS] = {
2625 .name = "schtasks",
2626 .is_available = is_schtasks_available,
2627 .update_schedule = schtasks_update_schedule,
2628 },
2629 };
2630
2631 static enum scheduler parse_scheduler(const char *value)
2632 {
2633 if (!value)
2634 return SCHEDULER_INVALID;
2635 else if (!strcasecmp(value, "auto"))
2636 return SCHEDULER_AUTO;
2637 else if (!strcasecmp(value, "cron") || !strcasecmp(value, "crontab"))
2638 return SCHEDULER_CRON;
2639 else if (!strcasecmp(value, "systemd") ||
2640 !strcasecmp(value, "systemd-timer"))
2641 return SCHEDULER_SYSTEMD;
2642 else if (!strcasecmp(value, "launchctl"))
2643 return SCHEDULER_LAUNCHCTL;
2644 else if (!strcasecmp(value, "schtasks"))
2645 return SCHEDULER_SCHTASKS;
2646 else
2647 return SCHEDULER_INVALID;
2648 }
2649
2650 static int maintenance_opt_scheduler(const struct option *opt, const char *arg,
2651 int unset)
2652 {
2653 enum scheduler *scheduler = opt->value;
2654
2655 BUG_ON_OPT_NEG(unset);
2656
2657 *scheduler = parse_scheduler(arg);
2658 if (*scheduler == SCHEDULER_INVALID)
2659 return error(_("unrecognized --scheduler argument '%s'"), arg);
2660 return 0;
2661 }
2662
2663 struct maintenance_start_opts {
2664 enum scheduler scheduler;
2665 };
2666
2667 static enum scheduler resolve_scheduler(enum scheduler scheduler)
2668 {
2669 if (scheduler != SCHEDULER_AUTO)
2670 return scheduler;
2671
2672 #if defined(__APPLE__)
2673 return SCHEDULER_LAUNCHCTL;
2674
2675 #elif defined(GIT_WINDOWS_NATIVE)
2676 return SCHEDULER_SCHTASKS;
2677
2678 #elif defined(__linux__)
2679 if (is_systemd_timer_available())
2680 return SCHEDULER_SYSTEMD;
2681 else if (is_crontab_available())
2682 return SCHEDULER_CRON;
2683 else
2684 die(_("neither systemd timers nor crontab are available"));
2685
2686 #else
2687 return SCHEDULER_CRON;
2688 #endif
2689 }
2690
2691 static void validate_scheduler(enum scheduler scheduler)
2692 {
2693 if (scheduler == SCHEDULER_INVALID)
2694 BUG("invalid scheduler");
2695 if (scheduler == SCHEDULER_AUTO)
2696 BUG("resolve_scheduler should have been called before");
2697
2698 if (!scheduler_fn[scheduler].is_available())
2699 die(_("%s scheduler is not available"),
2700 scheduler_fn[scheduler].name);
2701 }
2702
2703 static int update_background_schedule(const struct maintenance_start_opts *opts,
2704 int enable)
2705 {
2706 unsigned int i;
2707 int result = 0;
2708 struct lock_file lk;
2709 char *lock_path = xstrfmt("%s/schedule", the_repository->objects->odb->path);
2710
2711 if (hold_lock_file_for_update(&lk, lock_path, LOCK_NO_DEREF) < 0) {
2712 free(lock_path);
2713 return error(_("another process is scheduling background maintenance"));
2714 }
2715
2716 for (i = 1; i < ARRAY_SIZE(scheduler_fn); i++) {
2717 if (enable && opts->scheduler == i)
2718 continue;
2719 if (!scheduler_fn[i].is_available())
2720 continue;
2721 scheduler_fn[i].update_schedule(0, get_lock_file_fd(&lk));
2722 }
2723
2724 if (enable)
2725 result = scheduler_fn[opts->scheduler].update_schedule(
2726 1, get_lock_file_fd(&lk));
2727
2728 rollback_lock_file(&lk);
2729
2730 free(lock_path);
2731 return result;
2732 }
2733
2734 static const char *const builtin_maintenance_start_usage[] = {
2735 N_("git maintenance start [--scheduler=<scheduler>]"),
2736 NULL
2737 };
2738
2739 static int maintenance_start(int argc, const char **argv, const char *prefix)
2740 {
2741 struct maintenance_start_opts opts = { 0 };
2742 struct option options[] = {
2743 OPT_CALLBACK_F(
2744 0, "scheduler", &opts.scheduler, N_("scheduler"),
2745 N_("scheduler to trigger git maintenance run"),
2746 PARSE_OPT_NONEG, maintenance_opt_scheduler),
2747 OPT_END()
2748 };
2749 const char *register_args[] = { "register", NULL };
2750
2751 argc = parse_options(argc, argv, prefix, options,
2752 builtin_maintenance_start_usage, 0);
2753 if (argc)
2754 usage_with_options(builtin_maintenance_start_usage, options);
2755
2756 opts.scheduler = resolve_scheduler(opts.scheduler);
2757 validate_scheduler(opts.scheduler);
2758
2759 if (update_background_schedule(&opts, 1))
2760 die(_("failed to set up maintenance schedule"));
2761
2762 if (maintenance_register(ARRAY_SIZE(register_args)-1, register_args, NULL))
2763 warning(_("failed to add repo to global config"));
2764 return 0;
2765 }
2766
2767 static const char *const builtin_maintenance_stop_usage[] = {
2768 "git maintenance stop",
2769 NULL
2770 };
2771
2772 static int maintenance_stop(int argc, const char **argv, const char *prefix)
2773 {
2774 struct option options[] = {
2775 OPT_END()
2776 };
2777 argc = parse_options(argc, argv, prefix, options,
2778 builtin_maintenance_stop_usage, 0);
2779 if (argc)
2780 usage_with_options(builtin_maintenance_stop_usage, options);
2781 return update_background_schedule(NULL, 0);
2782 }
2783
2784 static const char * const builtin_maintenance_usage[] = {
2785 N_("git maintenance <subcommand> [<options>]"),
2786 NULL,
2787 };
2788
2789 int cmd_maintenance(int argc, const char **argv, const char *prefix)
2790 {
2791 parse_opt_subcommand_fn *fn = NULL;
2792 struct option builtin_maintenance_options[] = {
2793 OPT_SUBCOMMAND("run", &fn, maintenance_run),
2794 OPT_SUBCOMMAND("start", &fn, maintenance_start),
2795 OPT_SUBCOMMAND("stop", &fn, maintenance_stop),
2796 OPT_SUBCOMMAND("register", &fn, maintenance_register),
2797 OPT_SUBCOMMAND("unregister", &fn, maintenance_unregister),
2798 OPT_END(),
2799 };
2800
2801 argc = parse_options(argc, argv, prefix, builtin_maintenance_options,
2802 builtin_maintenance_usage, 0);
2803 return fn(argc, argv, prefix);
2804 }