]> git.ipfire.org Git - thirdparty/git.git/blob - builtin/gc.c
The seventh batch
[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 = refs_for_each_ref(get_main_ref_store(the_repository),
911 dfs_on_ref, &data);
912
913 repo_clear_commit_marks(the_repository, SEEN);
914
915 return result;
916 }
917
918 static int run_write_commit_graph(struct maintenance_run_opts *opts)
919 {
920 struct child_process child = CHILD_PROCESS_INIT;
921
922 child.git_cmd = child.close_object_store = 1;
923 strvec_pushl(&child.args, "commit-graph", "write",
924 "--split", "--reachable", NULL);
925
926 if (opts->quiet)
927 strvec_push(&child.args, "--no-progress");
928
929 return !!run_command(&child);
930 }
931
932 static int maintenance_task_commit_graph(struct maintenance_run_opts *opts)
933 {
934 prepare_repo_settings(the_repository);
935 if (!the_repository->settings.core_commit_graph)
936 return 0;
937
938 if (run_write_commit_graph(opts)) {
939 error(_("failed to write commit-graph"));
940 return 1;
941 }
942
943 return 0;
944 }
945
946 static int fetch_remote(struct remote *remote, void *cbdata)
947 {
948 struct maintenance_run_opts *opts = cbdata;
949 struct child_process child = CHILD_PROCESS_INIT;
950
951 if (remote->skip_default_update)
952 return 0;
953
954 child.git_cmd = 1;
955 strvec_pushl(&child.args, "fetch", remote->name,
956 "--prefetch", "--prune", "--no-tags",
957 "--no-write-fetch-head", "--recurse-submodules=no",
958 NULL);
959
960 if (opts->quiet)
961 strvec_push(&child.args, "--quiet");
962
963 return !!run_command(&child);
964 }
965
966 static int maintenance_task_prefetch(struct maintenance_run_opts *opts)
967 {
968 if (for_each_remote(fetch_remote, opts)) {
969 error(_("failed to prefetch remotes"));
970 return 1;
971 }
972
973 return 0;
974 }
975
976 static int maintenance_task_gc(struct maintenance_run_opts *opts)
977 {
978 struct child_process child = CHILD_PROCESS_INIT;
979
980 child.git_cmd = child.close_object_store = 1;
981 strvec_push(&child.args, "gc");
982
983 if (opts->auto_flag)
984 strvec_push(&child.args, "--auto");
985 if (opts->quiet)
986 strvec_push(&child.args, "--quiet");
987 else
988 strvec_push(&child.args, "--no-quiet");
989
990 return run_command(&child);
991 }
992
993 static int prune_packed(struct maintenance_run_opts *opts)
994 {
995 struct child_process child = CHILD_PROCESS_INIT;
996
997 child.git_cmd = 1;
998 strvec_push(&child.args, "prune-packed");
999
1000 if (opts->quiet)
1001 strvec_push(&child.args, "--quiet");
1002
1003 return !!run_command(&child);
1004 }
1005
1006 struct write_loose_object_data {
1007 FILE *in;
1008 int count;
1009 int batch_size;
1010 };
1011
1012 static int loose_object_auto_limit = 100;
1013
1014 static int loose_object_count(const struct object_id *oid UNUSED,
1015 const char *path UNUSED,
1016 void *data)
1017 {
1018 int *count = (int*)data;
1019 if (++(*count) >= loose_object_auto_limit)
1020 return 1;
1021 return 0;
1022 }
1023
1024 static int loose_object_auto_condition(void)
1025 {
1026 int count = 0;
1027
1028 git_config_get_int("maintenance.loose-objects.auto",
1029 &loose_object_auto_limit);
1030
1031 if (!loose_object_auto_limit)
1032 return 0;
1033 if (loose_object_auto_limit < 0)
1034 return 1;
1035
1036 return for_each_loose_file_in_objdir(the_repository->objects->odb->path,
1037 loose_object_count,
1038 NULL, NULL, &count);
1039 }
1040
1041 static int bail_on_loose(const struct object_id *oid UNUSED,
1042 const char *path UNUSED,
1043 void *data UNUSED)
1044 {
1045 return 1;
1046 }
1047
1048 static int write_loose_object_to_stdin(const struct object_id *oid,
1049 const char *path UNUSED,
1050 void *data)
1051 {
1052 struct write_loose_object_data *d = (struct write_loose_object_data *)data;
1053
1054 fprintf(d->in, "%s\n", oid_to_hex(oid));
1055
1056 return ++(d->count) > d->batch_size;
1057 }
1058
1059 static int pack_loose(struct maintenance_run_opts *opts)
1060 {
1061 struct repository *r = the_repository;
1062 int result = 0;
1063 struct write_loose_object_data data;
1064 struct child_process pack_proc = CHILD_PROCESS_INIT;
1065
1066 /*
1067 * Do not start pack-objects process
1068 * if there are no loose objects.
1069 */
1070 if (!for_each_loose_file_in_objdir(r->objects->odb->path,
1071 bail_on_loose,
1072 NULL, NULL, NULL))
1073 return 0;
1074
1075 pack_proc.git_cmd = 1;
1076
1077 strvec_push(&pack_proc.args, "pack-objects");
1078 if (opts->quiet)
1079 strvec_push(&pack_proc.args, "--quiet");
1080 strvec_pushf(&pack_proc.args, "%s/pack/loose", r->objects->odb->path);
1081
1082 pack_proc.in = -1;
1083
1084 if (start_command(&pack_proc)) {
1085 error(_("failed to start 'git pack-objects' process"));
1086 return 1;
1087 }
1088
1089 data.in = xfdopen(pack_proc.in, "w");
1090 data.count = 0;
1091 data.batch_size = 50000;
1092
1093 for_each_loose_file_in_objdir(r->objects->odb->path,
1094 write_loose_object_to_stdin,
1095 NULL,
1096 NULL,
1097 &data);
1098
1099 fclose(data.in);
1100
1101 if (finish_command(&pack_proc)) {
1102 error(_("failed to finish 'git pack-objects' process"));
1103 result = 1;
1104 }
1105
1106 return result;
1107 }
1108
1109 static int maintenance_task_loose_objects(struct maintenance_run_opts *opts)
1110 {
1111 return prune_packed(opts) || pack_loose(opts);
1112 }
1113
1114 static int incremental_repack_auto_condition(void)
1115 {
1116 struct packed_git *p;
1117 int incremental_repack_auto_limit = 10;
1118 int count = 0;
1119
1120 prepare_repo_settings(the_repository);
1121 if (!the_repository->settings.core_multi_pack_index)
1122 return 0;
1123
1124 git_config_get_int("maintenance.incremental-repack.auto",
1125 &incremental_repack_auto_limit);
1126
1127 if (!incremental_repack_auto_limit)
1128 return 0;
1129 if (incremental_repack_auto_limit < 0)
1130 return 1;
1131
1132 for (p = get_packed_git(the_repository);
1133 count < incremental_repack_auto_limit && p;
1134 p = p->next) {
1135 if (!p->multi_pack_index)
1136 count++;
1137 }
1138
1139 return count >= incremental_repack_auto_limit;
1140 }
1141
1142 static int multi_pack_index_write(struct maintenance_run_opts *opts)
1143 {
1144 struct child_process child = CHILD_PROCESS_INIT;
1145
1146 child.git_cmd = 1;
1147 strvec_pushl(&child.args, "multi-pack-index", "write", NULL);
1148
1149 if (opts->quiet)
1150 strvec_push(&child.args, "--no-progress");
1151
1152 if (run_command(&child))
1153 return error(_("failed to write multi-pack-index"));
1154
1155 return 0;
1156 }
1157
1158 static int multi_pack_index_expire(struct maintenance_run_opts *opts)
1159 {
1160 struct child_process child = CHILD_PROCESS_INIT;
1161
1162 child.git_cmd = child.close_object_store = 1;
1163 strvec_pushl(&child.args, "multi-pack-index", "expire", NULL);
1164
1165 if (opts->quiet)
1166 strvec_push(&child.args, "--no-progress");
1167
1168 if (run_command(&child))
1169 return error(_("'git multi-pack-index expire' failed"));
1170
1171 return 0;
1172 }
1173
1174 #define TWO_GIGABYTES (INT32_MAX)
1175
1176 static off_t get_auto_pack_size(void)
1177 {
1178 /*
1179 * The "auto" value is special: we optimize for
1180 * one large pack-file (i.e. from a clone) and
1181 * expect the rest to be small and they can be
1182 * repacked quickly.
1183 *
1184 * The strategy we select here is to select a
1185 * size that is one more than the second largest
1186 * pack-file. This ensures that we will repack
1187 * at least two packs if there are three or more
1188 * packs.
1189 */
1190 off_t max_size = 0;
1191 off_t second_largest_size = 0;
1192 off_t result_size;
1193 struct packed_git *p;
1194 struct repository *r = the_repository;
1195
1196 reprepare_packed_git(r);
1197 for (p = get_all_packs(r); p; p = p->next) {
1198 if (p->pack_size > max_size) {
1199 second_largest_size = max_size;
1200 max_size = p->pack_size;
1201 } else if (p->pack_size > second_largest_size)
1202 second_largest_size = p->pack_size;
1203 }
1204
1205 result_size = second_largest_size + 1;
1206
1207 /* But limit ourselves to a batch size of 2g */
1208 if (result_size > TWO_GIGABYTES)
1209 result_size = TWO_GIGABYTES;
1210
1211 return result_size;
1212 }
1213
1214 static int multi_pack_index_repack(struct maintenance_run_opts *opts)
1215 {
1216 struct child_process child = CHILD_PROCESS_INIT;
1217
1218 child.git_cmd = child.close_object_store = 1;
1219 strvec_pushl(&child.args, "multi-pack-index", "repack", NULL);
1220
1221 if (opts->quiet)
1222 strvec_push(&child.args, "--no-progress");
1223
1224 strvec_pushf(&child.args, "--batch-size=%"PRIuMAX,
1225 (uintmax_t)get_auto_pack_size());
1226
1227 if (run_command(&child))
1228 return error(_("'git multi-pack-index repack' failed"));
1229
1230 return 0;
1231 }
1232
1233 static int maintenance_task_incremental_repack(struct maintenance_run_opts *opts)
1234 {
1235 prepare_repo_settings(the_repository);
1236 if (!the_repository->settings.core_multi_pack_index) {
1237 warning(_("skipping incremental-repack task because core.multiPackIndex is disabled"));
1238 return 0;
1239 }
1240
1241 if (multi_pack_index_write(opts))
1242 return 1;
1243 if (multi_pack_index_expire(opts))
1244 return 1;
1245 if (multi_pack_index_repack(opts))
1246 return 1;
1247 return 0;
1248 }
1249
1250 typedef int maintenance_task_fn(struct maintenance_run_opts *opts);
1251
1252 /*
1253 * An auto condition function returns 1 if the task should run
1254 * and 0 if the task should NOT run. See needs_to_gc() for an
1255 * example.
1256 */
1257 typedef int maintenance_auto_fn(void);
1258
1259 struct maintenance_task {
1260 const char *name;
1261 maintenance_task_fn *fn;
1262 maintenance_auto_fn *auto_condition;
1263 unsigned enabled:1;
1264
1265 enum schedule_priority schedule;
1266
1267 /* -1 if not selected. */
1268 int selected_order;
1269 };
1270
1271 enum maintenance_task_label {
1272 TASK_PREFETCH,
1273 TASK_LOOSE_OBJECTS,
1274 TASK_INCREMENTAL_REPACK,
1275 TASK_GC,
1276 TASK_COMMIT_GRAPH,
1277 TASK_PACK_REFS,
1278
1279 /* Leave as final value */
1280 TASK__COUNT
1281 };
1282
1283 static struct maintenance_task tasks[] = {
1284 [TASK_PREFETCH] = {
1285 "prefetch",
1286 maintenance_task_prefetch,
1287 },
1288 [TASK_LOOSE_OBJECTS] = {
1289 "loose-objects",
1290 maintenance_task_loose_objects,
1291 loose_object_auto_condition,
1292 },
1293 [TASK_INCREMENTAL_REPACK] = {
1294 "incremental-repack",
1295 maintenance_task_incremental_repack,
1296 incremental_repack_auto_condition,
1297 },
1298 [TASK_GC] = {
1299 "gc",
1300 maintenance_task_gc,
1301 need_to_gc,
1302 1,
1303 },
1304 [TASK_COMMIT_GRAPH] = {
1305 "commit-graph",
1306 maintenance_task_commit_graph,
1307 should_write_commit_graph,
1308 },
1309 [TASK_PACK_REFS] = {
1310 "pack-refs",
1311 maintenance_task_pack_refs,
1312 pack_refs_condition,
1313 },
1314 };
1315
1316 static int compare_tasks_by_selection(const void *a_, const void *b_)
1317 {
1318 const struct maintenance_task *a = a_;
1319 const struct maintenance_task *b = b_;
1320
1321 return b->selected_order - a->selected_order;
1322 }
1323
1324 static int maintenance_run_tasks(struct maintenance_run_opts *opts)
1325 {
1326 int i, found_selected = 0;
1327 int result = 0;
1328 struct lock_file lk;
1329 struct repository *r = the_repository;
1330 char *lock_path = xstrfmt("%s/maintenance", r->objects->odb->path);
1331
1332 if (hold_lock_file_for_update(&lk, lock_path, LOCK_NO_DEREF) < 0) {
1333 /*
1334 * Another maintenance command is running.
1335 *
1336 * If --auto was provided, then it is likely due to a
1337 * recursive process stack. Do not report an error in
1338 * that case.
1339 */
1340 if (!opts->auto_flag && !opts->quiet)
1341 warning(_("lock file '%s' exists, skipping maintenance"),
1342 lock_path);
1343 free(lock_path);
1344 return 0;
1345 }
1346 free(lock_path);
1347
1348 for (i = 0; !found_selected && i < TASK__COUNT; i++)
1349 found_selected = tasks[i].selected_order >= 0;
1350
1351 if (found_selected)
1352 QSORT(tasks, TASK__COUNT, compare_tasks_by_selection);
1353
1354 for (i = 0; i < TASK__COUNT; i++) {
1355 if (found_selected && tasks[i].selected_order < 0)
1356 continue;
1357
1358 if (!found_selected && !tasks[i].enabled)
1359 continue;
1360
1361 if (opts->auto_flag &&
1362 (!tasks[i].auto_condition ||
1363 !tasks[i].auto_condition()))
1364 continue;
1365
1366 if (opts->schedule && tasks[i].schedule < opts->schedule)
1367 continue;
1368
1369 trace2_region_enter("maintenance", tasks[i].name, r);
1370 if (tasks[i].fn(opts)) {
1371 error(_("task '%s' failed"), tasks[i].name);
1372 result = 1;
1373 }
1374 trace2_region_leave("maintenance", tasks[i].name, r);
1375 }
1376
1377 rollback_lock_file(&lk);
1378 return result;
1379 }
1380
1381 static void initialize_maintenance_strategy(void)
1382 {
1383 char *config_str;
1384
1385 if (git_config_get_string("maintenance.strategy", &config_str))
1386 return;
1387
1388 if (!strcasecmp(config_str, "incremental")) {
1389 tasks[TASK_GC].schedule = SCHEDULE_NONE;
1390 tasks[TASK_COMMIT_GRAPH].enabled = 1;
1391 tasks[TASK_COMMIT_GRAPH].schedule = SCHEDULE_HOURLY;
1392 tasks[TASK_PREFETCH].enabled = 1;
1393 tasks[TASK_PREFETCH].schedule = SCHEDULE_HOURLY;
1394 tasks[TASK_INCREMENTAL_REPACK].enabled = 1;
1395 tasks[TASK_INCREMENTAL_REPACK].schedule = SCHEDULE_DAILY;
1396 tasks[TASK_LOOSE_OBJECTS].enabled = 1;
1397 tasks[TASK_LOOSE_OBJECTS].schedule = SCHEDULE_DAILY;
1398 tasks[TASK_PACK_REFS].enabled = 1;
1399 tasks[TASK_PACK_REFS].schedule = SCHEDULE_WEEKLY;
1400 }
1401 }
1402
1403 static void initialize_task_config(int schedule)
1404 {
1405 int i;
1406 struct strbuf config_name = STRBUF_INIT;
1407 gc_config();
1408
1409 if (schedule)
1410 initialize_maintenance_strategy();
1411
1412 for (i = 0; i < TASK__COUNT; i++) {
1413 int config_value;
1414 char *config_str;
1415
1416 strbuf_reset(&config_name);
1417 strbuf_addf(&config_name, "maintenance.%s.enabled",
1418 tasks[i].name);
1419
1420 if (!git_config_get_bool(config_name.buf, &config_value))
1421 tasks[i].enabled = config_value;
1422
1423 strbuf_reset(&config_name);
1424 strbuf_addf(&config_name, "maintenance.%s.schedule",
1425 tasks[i].name);
1426
1427 if (!git_config_get_string(config_name.buf, &config_str)) {
1428 tasks[i].schedule = parse_schedule(config_str);
1429 free(config_str);
1430 }
1431 }
1432
1433 strbuf_release(&config_name);
1434 }
1435
1436 static int task_option_parse(const struct option *opt UNUSED,
1437 const char *arg, int unset)
1438 {
1439 int i, num_selected = 0;
1440 struct maintenance_task *task = NULL;
1441
1442 BUG_ON_OPT_NEG(unset);
1443
1444 for (i = 0; i < TASK__COUNT; i++) {
1445 if (tasks[i].selected_order >= 0)
1446 num_selected++;
1447 if (!strcasecmp(tasks[i].name, arg)) {
1448 task = &tasks[i];
1449 }
1450 }
1451
1452 if (!task) {
1453 error(_("'%s' is not a valid task"), arg);
1454 return 1;
1455 }
1456
1457 if (task->selected_order >= 0) {
1458 error(_("task '%s' cannot be selected multiple times"), arg);
1459 return 1;
1460 }
1461
1462 task->selected_order = num_selected + 1;
1463
1464 return 0;
1465 }
1466
1467 static int maintenance_run(int argc, const char **argv, const char *prefix)
1468 {
1469 int i;
1470 struct maintenance_run_opts opts;
1471 struct option builtin_maintenance_run_options[] = {
1472 OPT_BOOL(0, "auto", &opts.auto_flag,
1473 N_("run tasks based on the state of the repository")),
1474 OPT_CALLBACK(0, "schedule", &opts.schedule, N_("frequency"),
1475 N_("run tasks based on frequency"),
1476 maintenance_opt_schedule),
1477 OPT_BOOL(0, "quiet", &opts.quiet,
1478 N_("do not report progress or other information over stderr")),
1479 OPT_CALLBACK_F(0, "task", NULL, N_("task"),
1480 N_("run a specific task"),
1481 PARSE_OPT_NONEG, task_option_parse),
1482 OPT_END()
1483 };
1484 memset(&opts, 0, sizeof(opts));
1485
1486 opts.quiet = !isatty(2);
1487
1488 for (i = 0; i < TASK__COUNT; i++)
1489 tasks[i].selected_order = -1;
1490
1491 argc = parse_options(argc, argv, prefix,
1492 builtin_maintenance_run_options,
1493 builtin_maintenance_run_usage,
1494 PARSE_OPT_STOP_AT_NON_OPTION);
1495
1496 if (opts.auto_flag && opts.schedule)
1497 die(_("use at most one of --auto and --schedule=<frequency>"));
1498
1499 initialize_task_config(opts.schedule);
1500
1501 if (argc != 0)
1502 usage_with_options(builtin_maintenance_run_usage,
1503 builtin_maintenance_run_options);
1504 return maintenance_run_tasks(&opts);
1505 }
1506
1507 static char *get_maintpath(void)
1508 {
1509 struct strbuf sb = STRBUF_INIT;
1510 const char *p = the_repository->worktree ?
1511 the_repository->worktree : the_repository->gitdir;
1512
1513 strbuf_realpath(&sb, p, 1);
1514 return strbuf_detach(&sb, NULL);
1515 }
1516
1517 static char const * const builtin_maintenance_register_usage[] = {
1518 "git maintenance register [--config-file <path>]",
1519 NULL
1520 };
1521
1522 static int maintenance_register(int argc, const char **argv, const char *prefix)
1523 {
1524 char *config_file = NULL;
1525 struct option options[] = {
1526 OPT_STRING(0, "config-file", &config_file, N_("file"), N_("use given config file")),
1527 OPT_END(),
1528 };
1529 int found = 0;
1530 const char *key = "maintenance.repo";
1531 char *maintpath = get_maintpath();
1532 struct string_list_item *item;
1533 const struct string_list *list;
1534
1535 argc = parse_options(argc, argv, prefix, options,
1536 builtin_maintenance_register_usage, 0);
1537 if (argc)
1538 usage_with_options(builtin_maintenance_register_usage,
1539 options);
1540
1541 /* Disable foreground maintenance */
1542 git_config_set("maintenance.auto", "false");
1543
1544 /* Set maintenance strategy, if unset */
1545 if (git_config_get("maintenance.strategy"))
1546 git_config_set("maintenance.strategy", "incremental");
1547
1548 if (!git_config_get_string_multi(key, &list)) {
1549 for_each_string_list_item(item, list) {
1550 if (!strcmp(maintpath, item->string)) {
1551 found = 1;
1552 break;
1553 }
1554 }
1555 }
1556
1557 if (!found) {
1558 int rc;
1559 char *global_config_file = NULL;
1560
1561 if (!config_file) {
1562 global_config_file = git_global_config();
1563 config_file = global_config_file;
1564 }
1565 if (!config_file)
1566 die(_("$HOME not set"));
1567 rc = git_config_set_multivar_in_file_gently(
1568 config_file, "maintenance.repo", maintpath,
1569 CONFIG_REGEX_NONE, NULL, 0);
1570 free(global_config_file);
1571
1572 if (rc)
1573 die(_("unable to add '%s' value of '%s'"),
1574 key, maintpath);
1575 }
1576
1577 free(maintpath);
1578 return 0;
1579 }
1580
1581 static char const * const builtin_maintenance_unregister_usage[] = {
1582 "git maintenance unregister [--config-file <path>] [--force]",
1583 NULL
1584 };
1585
1586 static int maintenance_unregister(int argc, const char **argv, const char *prefix)
1587 {
1588 int force = 0;
1589 char *config_file = NULL;
1590 struct option options[] = {
1591 OPT_STRING(0, "config-file", &config_file, N_("file"), N_("use given config file")),
1592 OPT__FORCE(&force,
1593 N_("return success even if repository was not registered"),
1594 PARSE_OPT_NOCOMPLETE),
1595 OPT_END(),
1596 };
1597 const char *key = "maintenance.repo";
1598 char *maintpath = get_maintpath();
1599 int found = 0;
1600 struct string_list_item *item;
1601 const struct string_list *list;
1602 struct config_set cs = { { 0 } };
1603
1604 argc = parse_options(argc, argv, prefix, options,
1605 builtin_maintenance_unregister_usage, 0);
1606 if (argc)
1607 usage_with_options(builtin_maintenance_unregister_usage,
1608 options);
1609
1610 if (config_file) {
1611 git_configset_init(&cs);
1612 git_configset_add_file(&cs, config_file);
1613 }
1614 if (!(config_file
1615 ? git_configset_get_string_multi(&cs, key, &list)
1616 : git_config_get_string_multi(key, &list))) {
1617 for_each_string_list_item(item, list) {
1618 if (!strcmp(maintpath, item->string)) {
1619 found = 1;
1620 break;
1621 }
1622 }
1623 }
1624
1625 if (found) {
1626 int rc;
1627 char *global_config_file = NULL;
1628
1629 if (!config_file) {
1630 global_config_file = git_global_config();
1631 config_file = global_config_file;
1632 }
1633 if (!config_file)
1634 die(_("$HOME not set"));
1635 rc = git_config_set_multivar_in_file_gently(
1636 config_file, key, NULL, maintpath, NULL,
1637 CONFIG_FLAGS_MULTI_REPLACE | CONFIG_FLAGS_FIXED_VALUE);
1638 free(global_config_file);
1639
1640 if (rc &&
1641 (!force || rc == CONFIG_NOTHING_SET))
1642 die(_("unable to unset '%s' value of '%s'"),
1643 key, maintpath);
1644 } else if (!force) {
1645 die(_("repository '%s' is not registered"), maintpath);
1646 }
1647
1648 git_configset_clear(&cs);
1649 free(maintpath);
1650 return 0;
1651 }
1652
1653 static const char *get_frequency(enum schedule_priority schedule)
1654 {
1655 switch (schedule) {
1656 case SCHEDULE_HOURLY:
1657 return "hourly";
1658 case SCHEDULE_DAILY:
1659 return "daily";
1660 case SCHEDULE_WEEKLY:
1661 return "weekly";
1662 default:
1663 BUG("invalid schedule %d", schedule);
1664 }
1665 }
1666
1667 /*
1668 * get_schedule_cmd` reads the GIT_TEST_MAINT_SCHEDULER environment variable
1669 * to mock the schedulers that `git maintenance start` rely on.
1670 *
1671 * For test purpose, GIT_TEST_MAINT_SCHEDULER can be set to a comma-separated
1672 * list of colon-separated key/value pairs where each pair contains a scheduler
1673 * and its corresponding mock.
1674 *
1675 * * If $GIT_TEST_MAINT_SCHEDULER is not set, return false and leave the
1676 * arguments unmodified.
1677 *
1678 * * If $GIT_TEST_MAINT_SCHEDULER is set, return true.
1679 * In this case, the *cmd value is read as input.
1680 *
1681 * * if the input value *cmd is the key of one of the comma-separated list
1682 * item, then *is_available is set to true and *cmd is modified and becomes
1683 * the mock command.
1684 *
1685 * * if the input value *cmd isn’t the key of any of the comma-separated list
1686 * item, then *is_available is set to false.
1687 *
1688 * Ex.:
1689 * GIT_TEST_MAINT_SCHEDULER not set
1690 * +-------+-------------------------------------------------+
1691 * | Input | Output |
1692 * | *cmd | return code | *cmd | *is_available |
1693 * +-------+-------------+-------------------+---------------+
1694 * | "foo" | false | "foo" (unchanged) | (unchanged) |
1695 * +-------+-------------+-------------------+---------------+
1696 *
1697 * GIT_TEST_MAINT_SCHEDULER set to “foo:./mock_foo.sh,bar:./mock_bar.sh”
1698 * +-------+-------------------------------------------------+
1699 * | Input | Output |
1700 * | *cmd | return code | *cmd | *is_available |
1701 * +-------+-------------+-------------------+---------------+
1702 * | "foo" | true | "./mock.foo.sh" | true |
1703 * | "qux" | true | "qux" (unchanged) | false |
1704 * +-------+-------------+-------------------+---------------+
1705 */
1706 static int get_schedule_cmd(const char **cmd, int *is_available)
1707 {
1708 char *testing = xstrdup_or_null(getenv("GIT_TEST_MAINT_SCHEDULER"));
1709 struct string_list_item *item;
1710 struct string_list list = STRING_LIST_INIT_NODUP;
1711
1712 if (!testing)
1713 return 0;
1714
1715 if (is_available)
1716 *is_available = 0;
1717
1718 string_list_split_in_place(&list, testing, ",", -1);
1719 for_each_string_list_item(item, &list) {
1720 struct string_list pair = STRING_LIST_INIT_NODUP;
1721
1722 if (string_list_split_in_place(&pair, item->string, ":", 2) != 2)
1723 continue;
1724
1725 if (!strcmp(*cmd, pair.items[0].string)) {
1726 *cmd = pair.items[1].string;
1727 if (is_available)
1728 *is_available = 1;
1729 string_list_clear(&list, 0);
1730 UNLEAK(testing);
1731 return 1;
1732 }
1733 }
1734
1735 string_list_clear(&list, 0);
1736 free(testing);
1737 return 1;
1738 }
1739
1740 static int get_random_minute(void)
1741 {
1742 /* Use a static value when under tests. */
1743 if (getenv("GIT_TEST_MAINT_SCHEDULER"))
1744 return 13;
1745
1746 return git_rand() % 60;
1747 }
1748
1749 static int is_launchctl_available(void)
1750 {
1751 const char *cmd = "launchctl";
1752 int is_available;
1753 if (get_schedule_cmd(&cmd, &is_available))
1754 return is_available;
1755
1756 #ifdef __APPLE__
1757 return 1;
1758 #else
1759 return 0;
1760 #endif
1761 }
1762
1763 static char *launchctl_service_name(const char *frequency)
1764 {
1765 struct strbuf label = STRBUF_INIT;
1766 strbuf_addf(&label, "org.git-scm.git.%s", frequency);
1767 return strbuf_detach(&label, NULL);
1768 }
1769
1770 static char *launchctl_service_filename(const char *name)
1771 {
1772 char *expanded;
1773 struct strbuf filename = STRBUF_INIT;
1774 strbuf_addf(&filename, "~/Library/LaunchAgents/%s.plist", name);
1775
1776 expanded = interpolate_path(filename.buf, 1);
1777 if (!expanded)
1778 die(_("failed to expand path '%s'"), filename.buf);
1779
1780 strbuf_release(&filename);
1781 return expanded;
1782 }
1783
1784 static char *launchctl_get_uid(void)
1785 {
1786 return xstrfmt("gui/%d", getuid());
1787 }
1788
1789 static int launchctl_boot_plist(int enable, const char *filename)
1790 {
1791 const char *cmd = "launchctl";
1792 int result;
1793 struct child_process child = CHILD_PROCESS_INIT;
1794 char *uid = launchctl_get_uid();
1795
1796 get_schedule_cmd(&cmd, NULL);
1797 strvec_split(&child.args, cmd);
1798 strvec_pushl(&child.args, enable ? "bootstrap" : "bootout", uid,
1799 filename, NULL);
1800
1801 child.no_stderr = 1;
1802 child.no_stdout = 1;
1803
1804 if (start_command(&child))
1805 die(_("failed to start launchctl"));
1806
1807 result = finish_command(&child);
1808
1809 free(uid);
1810 return result;
1811 }
1812
1813 static int launchctl_remove_plist(enum schedule_priority schedule)
1814 {
1815 const char *frequency = get_frequency(schedule);
1816 char *name = launchctl_service_name(frequency);
1817 char *filename = launchctl_service_filename(name);
1818 int result = launchctl_boot_plist(0, filename);
1819 unlink(filename);
1820 free(filename);
1821 free(name);
1822 return result;
1823 }
1824
1825 static int launchctl_remove_plists(void)
1826 {
1827 return launchctl_remove_plist(SCHEDULE_HOURLY) ||
1828 launchctl_remove_plist(SCHEDULE_DAILY) ||
1829 launchctl_remove_plist(SCHEDULE_WEEKLY);
1830 }
1831
1832 static int launchctl_list_contains_plist(const char *name, const char *cmd)
1833 {
1834 struct child_process child = CHILD_PROCESS_INIT;
1835
1836 strvec_split(&child.args, cmd);
1837 strvec_pushl(&child.args, "list", name, NULL);
1838
1839 child.no_stderr = 1;
1840 child.no_stdout = 1;
1841
1842 if (start_command(&child))
1843 die(_("failed to start launchctl"));
1844
1845 /* Returns failure if 'name' doesn't exist. */
1846 return !finish_command(&child);
1847 }
1848
1849 static int launchctl_schedule_plist(const char *exec_path, enum schedule_priority schedule)
1850 {
1851 int i, fd;
1852 const char *preamble, *repeat;
1853 const char *frequency = get_frequency(schedule);
1854 char *name = launchctl_service_name(frequency);
1855 char *filename = launchctl_service_filename(name);
1856 struct lock_file lk = LOCK_INIT;
1857 static unsigned long lock_file_timeout_ms = ULONG_MAX;
1858 struct strbuf plist = STRBUF_INIT, plist2 = STRBUF_INIT;
1859 struct stat st;
1860 const char *cmd = "launchctl";
1861 int minute = get_random_minute();
1862
1863 get_schedule_cmd(&cmd, NULL);
1864 preamble = "<?xml version=\"1.0\"?>\n"
1865 "<!DOCTYPE plist PUBLIC \"-//Apple//DTD PLIST 1.0//EN\" \"http://www.apple.com/DTDs/PropertyList-1.0.dtd\">\n"
1866 "<plist version=\"1.0\">"
1867 "<dict>\n"
1868 "<key>Label</key><string>%s</string>\n"
1869 "<key>ProgramArguments</key>\n"
1870 "<array>\n"
1871 "<string>%s/git</string>\n"
1872 "<string>--exec-path=%s</string>\n"
1873 "<string>for-each-repo</string>\n"
1874 "<string>--keep-going</string>\n"
1875 "<string>--config=maintenance.repo</string>\n"
1876 "<string>maintenance</string>\n"
1877 "<string>run</string>\n"
1878 "<string>--schedule=%s</string>\n"
1879 "</array>\n"
1880 "<key>StartCalendarInterval</key>\n"
1881 "<array>\n";
1882 strbuf_addf(&plist, preamble, name, exec_path, exec_path, frequency);
1883
1884 switch (schedule) {
1885 case SCHEDULE_HOURLY:
1886 repeat = "<dict>\n"
1887 "<key>Hour</key><integer>%d</integer>\n"
1888 "<key>Minute</key><integer>%d</integer>\n"
1889 "</dict>\n";
1890 for (i = 1; i <= 23; i++)
1891 strbuf_addf(&plist, repeat, i, minute);
1892 break;
1893
1894 case SCHEDULE_DAILY:
1895 repeat = "<dict>\n"
1896 "<key>Day</key><integer>%d</integer>\n"
1897 "<key>Hour</key><integer>0</integer>\n"
1898 "<key>Minute</key><integer>%d</integer>\n"
1899 "</dict>\n";
1900 for (i = 1; i <= 6; i++)
1901 strbuf_addf(&plist, repeat, i, minute);
1902 break;
1903
1904 case SCHEDULE_WEEKLY:
1905 strbuf_addf(&plist,
1906 "<dict>\n"
1907 "<key>Day</key><integer>0</integer>\n"
1908 "<key>Hour</key><integer>0</integer>\n"
1909 "<key>Minute</key><integer>%d</integer>\n"
1910 "</dict>\n",
1911 minute);
1912 break;
1913
1914 default:
1915 /* unreachable */
1916 break;
1917 }
1918 strbuf_addstr(&plist, "</array>\n</dict>\n</plist>\n");
1919
1920 if (safe_create_leading_directories(filename))
1921 die(_("failed to create directories for '%s'"), filename);
1922
1923 if ((long)lock_file_timeout_ms < 0 &&
1924 git_config_get_ulong("gc.launchctlplistlocktimeoutms",
1925 &lock_file_timeout_ms))
1926 lock_file_timeout_ms = 150;
1927
1928 fd = hold_lock_file_for_update_timeout(&lk, filename, LOCK_DIE_ON_ERROR,
1929 lock_file_timeout_ms);
1930
1931 /*
1932 * Does this file already exist? With the intended contents? Is it
1933 * registered already? Then it does not need to be re-registered.
1934 */
1935 if (!stat(filename, &st) && st.st_size == plist.len &&
1936 strbuf_read_file(&plist2, filename, plist.len) == plist.len &&
1937 !strbuf_cmp(&plist, &plist2) &&
1938 launchctl_list_contains_plist(name, cmd))
1939 rollback_lock_file(&lk);
1940 else {
1941 if (write_in_full(fd, plist.buf, plist.len) < 0 ||
1942 commit_lock_file(&lk))
1943 die_errno(_("could not write '%s'"), filename);
1944
1945 /* bootout might fail if not already running, so ignore */
1946 launchctl_boot_plist(0, filename);
1947 if (launchctl_boot_plist(1, filename))
1948 die(_("failed to bootstrap service %s"), filename);
1949 }
1950
1951 free(filename);
1952 free(name);
1953 strbuf_release(&plist);
1954 strbuf_release(&plist2);
1955 return 0;
1956 }
1957
1958 static int launchctl_add_plists(void)
1959 {
1960 const char *exec_path = git_exec_path();
1961
1962 return launchctl_schedule_plist(exec_path, SCHEDULE_HOURLY) ||
1963 launchctl_schedule_plist(exec_path, SCHEDULE_DAILY) ||
1964 launchctl_schedule_plist(exec_path, SCHEDULE_WEEKLY);
1965 }
1966
1967 static int launchctl_update_schedule(int run_maintenance, int fd UNUSED)
1968 {
1969 if (run_maintenance)
1970 return launchctl_add_plists();
1971 else
1972 return launchctl_remove_plists();
1973 }
1974
1975 static int is_schtasks_available(void)
1976 {
1977 const char *cmd = "schtasks";
1978 int is_available;
1979 if (get_schedule_cmd(&cmd, &is_available))
1980 return is_available;
1981
1982 #ifdef GIT_WINDOWS_NATIVE
1983 return 1;
1984 #else
1985 return 0;
1986 #endif
1987 }
1988
1989 static char *schtasks_task_name(const char *frequency)
1990 {
1991 struct strbuf label = STRBUF_INIT;
1992 strbuf_addf(&label, "Git Maintenance (%s)", frequency);
1993 return strbuf_detach(&label, NULL);
1994 }
1995
1996 static int schtasks_remove_task(enum schedule_priority schedule)
1997 {
1998 const char *cmd = "schtasks";
1999 struct child_process child = CHILD_PROCESS_INIT;
2000 const char *frequency = get_frequency(schedule);
2001 char *name = schtasks_task_name(frequency);
2002
2003 get_schedule_cmd(&cmd, NULL);
2004 strvec_split(&child.args, cmd);
2005 strvec_pushl(&child.args, "/delete", "/tn", name, "/f", NULL);
2006 free(name);
2007
2008 return run_command(&child);
2009 }
2010
2011 static int schtasks_remove_tasks(void)
2012 {
2013 return schtasks_remove_task(SCHEDULE_HOURLY) ||
2014 schtasks_remove_task(SCHEDULE_DAILY) ||
2015 schtasks_remove_task(SCHEDULE_WEEKLY);
2016 }
2017
2018 static int schtasks_schedule_task(const char *exec_path, enum schedule_priority schedule)
2019 {
2020 const char *cmd = "schtasks";
2021 int result;
2022 struct child_process child = CHILD_PROCESS_INIT;
2023 const char *xml;
2024 struct tempfile *tfile;
2025 const char *frequency = get_frequency(schedule);
2026 char *name = schtasks_task_name(frequency);
2027 struct strbuf tfilename = STRBUF_INIT;
2028 int minute = get_random_minute();
2029
2030 get_schedule_cmd(&cmd, NULL);
2031
2032 strbuf_addf(&tfilename, "%s/schedule_%s_XXXXXX",
2033 get_git_common_dir(), frequency);
2034 tfile = xmks_tempfile(tfilename.buf);
2035 strbuf_release(&tfilename);
2036
2037 if (!fdopen_tempfile(tfile, "w"))
2038 die(_("failed to create temp xml file"));
2039
2040 xml = "<?xml version=\"1.0\" ?>\n"
2041 "<Task version=\"1.4\" xmlns=\"http://schemas.microsoft.com/windows/2004/02/mit/task\">\n"
2042 "<Triggers>\n"
2043 "<CalendarTrigger>\n";
2044 fputs(xml, tfile->fp);
2045
2046 switch (schedule) {
2047 case SCHEDULE_HOURLY:
2048 fprintf(tfile->fp,
2049 "<StartBoundary>2020-01-01T01:%02d:00</StartBoundary>\n"
2050 "<Enabled>true</Enabled>\n"
2051 "<ScheduleByDay>\n"
2052 "<DaysInterval>1</DaysInterval>\n"
2053 "</ScheduleByDay>\n"
2054 "<Repetition>\n"
2055 "<Interval>PT1H</Interval>\n"
2056 "<Duration>PT23H</Duration>\n"
2057 "<StopAtDurationEnd>false</StopAtDurationEnd>\n"
2058 "</Repetition>\n",
2059 minute);
2060 break;
2061
2062 case SCHEDULE_DAILY:
2063 fprintf(tfile->fp,
2064 "<StartBoundary>2020-01-01T00:%02d:00</StartBoundary>\n"
2065 "<Enabled>true</Enabled>\n"
2066 "<ScheduleByWeek>\n"
2067 "<DaysOfWeek>\n"
2068 "<Monday />\n"
2069 "<Tuesday />\n"
2070 "<Wednesday />\n"
2071 "<Thursday />\n"
2072 "<Friday />\n"
2073 "<Saturday />\n"
2074 "</DaysOfWeek>\n"
2075 "<WeeksInterval>1</WeeksInterval>\n"
2076 "</ScheduleByWeek>\n",
2077 minute);
2078 break;
2079
2080 case SCHEDULE_WEEKLY:
2081 fprintf(tfile->fp,
2082 "<StartBoundary>2020-01-01T00:%02d:00</StartBoundary>\n"
2083 "<Enabled>true</Enabled>\n"
2084 "<ScheduleByWeek>\n"
2085 "<DaysOfWeek>\n"
2086 "<Sunday />\n"
2087 "</DaysOfWeek>\n"
2088 "<WeeksInterval>1</WeeksInterval>\n"
2089 "</ScheduleByWeek>\n",
2090 minute);
2091 break;
2092
2093 default:
2094 break;
2095 }
2096
2097 xml = "</CalendarTrigger>\n"
2098 "</Triggers>\n"
2099 "<Principals>\n"
2100 "<Principal id=\"Author\">\n"
2101 "<LogonType>InteractiveToken</LogonType>\n"
2102 "<RunLevel>LeastPrivilege</RunLevel>\n"
2103 "</Principal>\n"
2104 "</Principals>\n"
2105 "<Settings>\n"
2106 "<MultipleInstancesPolicy>IgnoreNew</MultipleInstancesPolicy>\n"
2107 "<Enabled>true</Enabled>\n"
2108 "<Hidden>true</Hidden>\n"
2109 "<UseUnifiedSchedulingEngine>true</UseUnifiedSchedulingEngine>\n"
2110 "<WakeToRun>false</WakeToRun>\n"
2111 "<ExecutionTimeLimit>PT72H</ExecutionTimeLimit>\n"
2112 "<Priority>7</Priority>\n"
2113 "</Settings>\n"
2114 "<Actions Context=\"Author\">\n"
2115 "<Exec>\n"
2116 "<Command>\"%s\\headless-git.exe\"</Command>\n"
2117 "<Arguments>--exec-path=\"%s\" for-each-repo --keep-going --config=maintenance.repo maintenance run --schedule=%s</Arguments>\n"
2118 "</Exec>\n"
2119 "</Actions>\n"
2120 "</Task>\n";
2121 fprintf(tfile->fp, xml, exec_path, exec_path, frequency);
2122 strvec_split(&child.args, cmd);
2123 strvec_pushl(&child.args, "/create", "/tn", name, "/f", "/xml",
2124 get_tempfile_path(tfile), NULL);
2125 close_tempfile_gently(tfile);
2126
2127 child.no_stdout = 1;
2128 child.no_stderr = 1;
2129
2130 if (start_command(&child))
2131 die(_("failed to start schtasks"));
2132 result = finish_command(&child);
2133
2134 delete_tempfile(&tfile);
2135 free(name);
2136 return result;
2137 }
2138
2139 static int schtasks_schedule_tasks(void)
2140 {
2141 const char *exec_path = git_exec_path();
2142
2143 return schtasks_schedule_task(exec_path, SCHEDULE_HOURLY) ||
2144 schtasks_schedule_task(exec_path, SCHEDULE_DAILY) ||
2145 schtasks_schedule_task(exec_path, SCHEDULE_WEEKLY);
2146 }
2147
2148 static int schtasks_update_schedule(int run_maintenance, int fd UNUSED)
2149 {
2150 if (run_maintenance)
2151 return schtasks_schedule_tasks();
2152 else
2153 return schtasks_remove_tasks();
2154 }
2155
2156 MAYBE_UNUSED
2157 static int check_crontab_process(const char *cmd)
2158 {
2159 struct child_process child = CHILD_PROCESS_INIT;
2160
2161 strvec_split(&child.args, cmd);
2162 strvec_push(&child.args, "-l");
2163 child.no_stdin = 1;
2164 child.no_stdout = 1;
2165 child.no_stderr = 1;
2166 child.silent_exec_failure = 1;
2167
2168 if (start_command(&child))
2169 return 0;
2170 /* Ignore exit code, as an empty crontab will return error. */
2171 finish_command(&child);
2172 return 1;
2173 }
2174
2175 static int is_crontab_available(void)
2176 {
2177 const char *cmd = "crontab";
2178 int is_available;
2179
2180 if (get_schedule_cmd(&cmd, &is_available))
2181 return is_available;
2182
2183 #ifdef __APPLE__
2184 /*
2185 * macOS has cron, but it requires special permissions and will
2186 * create a UI alert when attempting to run this command.
2187 */
2188 return 0;
2189 #else
2190 return check_crontab_process(cmd);
2191 #endif
2192 }
2193
2194 #define BEGIN_LINE "# BEGIN GIT MAINTENANCE SCHEDULE"
2195 #define END_LINE "# END GIT MAINTENANCE SCHEDULE"
2196
2197 static int crontab_update_schedule(int run_maintenance, int fd)
2198 {
2199 const char *cmd = "crontab";
2200 int result = 0;
2201 int in_old_region = 0;
2202 struct child_process crontab_list = CHILD_PROCESS_INIT;
2203 struct child_process crontab_edit = CHILD_PROCESS_INIT;
2204 FILE *cron_list, *cron_in;
2205 struct strbuf line = STRBUF_INIT;
2206 struct tempfile *tmpedit = NULL;
2207 int minute = get_random_minute();
2208
2209 get_schedule_cmd(&cmd, NULL);
2210 strvec_split(&crontab_list.args, cmd);
2211 strvec_push(&crontab_list.args, "-l");
2212 crontab_list.in = -1;
2213 crontab_list.out = dup(fd);
2214 crontab_list.git_cmd = 0;
2215
2216 if (start_command(&crontab_list))
2217 return error(_("failed to run 'crontab -l'; your system might not support 'cron'"));
2218
2219 /* Ignore exit code, as an empty crontab will return error. */
2220 finish_command(&crontab_list);
2221
2222 tmpedit = mks_tempfile_t(".git_cron_edit_tmpXXXXXX");
2223 if (!tmpedit) {
2224 result = error(_("failed to create crontab temporary file"));
2225 goto out;
2226 }
2227 cron_in = fdopen_tempfile(tmpedit, "w");
2228 if (!cron_in) {
2229 result = error(_("failed to open temporary file"));
2230 goto out;
2231 }
2232
2233 /*
2234 * Read from the .lock file, filtering out the old
2235 * schedule while appending the new schedule.
2236 */
2237 cron_list = fdopen(fd, "r");
2238 rewind(cron_list);
2239
2240 while (!strbuf_getline_lf(&line, cron_list)) {
2241 if (!in_old_region && !strcmp(line.buf, BEGIN_LINE))
2242 in_old_region = 1;
2243 else if (in_old_region && !strcmp(line.buf, END_LINE))
2244 in_old_region = 0;
2245 else if (!in_old_region)
2246 fprintf(cron_in, "%s\n", line.buf);
2247 }
2248 strbuf_release(&line);
2249
2250 if (run_maintenance) {
2251 struct strbuf line_format = STRBUF_INIT;
2252 const char *exec_path = git_exec_path();
2253
2254 fprintf(cron_in, "%s\n", BEGIN_LINE);
2255 fprintf(cron_in,
2256 "# The following schedule was created by Git\n");
2257 fprintf(cron_in, "# Any edits made in this region might be\n");
2258 fprintf(cron_in,
2259 "# replaced in the future by a Git command.\n\n");
2260
2261 strbuf_addf(&line_format,
2262 "%%d %%s * * %%s \"%s/git\" --exec-path=\"%s\" for-each-repo --keep-going --config=maintenance.repo maintenance run --schedule=%%s\n",
2263 exec_path, exec_path);
2264 fprintf(cron_in, line_format.buf, minute, "1-23", "*", "hourly");
2265 fprintf(cron_in, line_format.buf, minute, "0", "1-6", "daily");
2266 fprintf(cron_in, line_format.buf, minute, "0", "0", "weekly");
2267 strbuf_release(&line_format);
2268
2269 fprintf(cron_in, "\n%s\n", END_LINE);
2270 }
2271
2272 fflush(cron_in);
2273
2274 strvec_split(&crontab_edit.args, cmd);
2275 strvec_push(&crontab_edit.args, get_tempfile_path(tmpedit));
2276 crontab_edit.git_cmd = 0;
2277
2278 if (start_command(&crontab_edit)) {
2279 result = error(_("failed to run 'crontab'; your system might not support 'cron'"));
2280 goto out;
2281 }
2282
2283 if (finish_command(&crontab_edit))
2284 result = error(_("'crontab' died"));
2285 else
2286 fclose(cron_list);
2287 out:
2288 delete_tempfile(&tmpedit);
2289 return result;
2290 }
2291
2292 static int real_is_systemd_timer_available(void)
2293 {
2294 struct child_process child = CHILD_PROCESS_INIT;
2295
2296 strvec_pushl(&child.args, "systemctl", "--user", "list-timers", NULL);
2297 child.no_stdin = 1;
2298 child.no_stdout = 1;
2299 child.no_stderr = 1;
2300 child.silent_exec_failure = 1;
2301
2302 if (start_command(&child))
2303 return 0;
2304 if (finish_command(&child))
2305 return 0;
2306 return 1;
2307 }
2308
2309 static int is_systemd_timer_available(void)
2310 {
2311 const char *cmd = "systemctl";
2312 int is_available;
2313
2314 if (get_schedule_cmd(&cmd, &is_available))
2315 return is_available;
2316
2317 return real_is_systemd_timer_available();
2318 }
2319
2320 static char *xdg_config_home_systemd(const char *filename)
2321 {
2322 return xdg_config_home_for("systemd/user", filename);
2323 }
2324
2325 #define SYSTEMD_UNIT_FORMAT "git-maintenance@%s.%s"
2326
2327 static int systemd_timer_delete_timer_file(enum schedule_priority priority)
2328 {
2329 int ret = 0;
2330 const char *frequency = get_frequency(priority);
2331 char *local_timer_name = xstrfmt(SYSTEMD_UNIT_FORMAT, frequency, "timer");
2332 char *filename = xdg_config_home_systemd(local_timer_name);
2333
2334 if (unlink(filename) && !is_missing_file_error(errno))
2335 ret = error_errno(_("failed to delete '%s'"), filename);
2336
2337 free(filename);
2338 free(local_timer_name);
2339 return ret;
2340 }
2341
2342 static int systemd_timer_delete_service_template(void)
2343 {
2344 int ret = 0;
2345 char *local_service_name = xstrfmt(SYSTEMD_UNIT_FORMAT, "", "service");
2346 char *filename = xdg_config_home_systemd(local_service_name);
2347 if (unlink(filename) && !is_missing_file_error(errno))
2348 ret = error_errno(_("failed to delete '%s'"), filename);
2349
2350 free(filename);
2351 free(local_service_name);
2352 return ret;
2353 }
2354
2355 /*
2356 * Write the schedule information into a git-maintenance@<schedule>.timer
2357 * file using a custom minute. This timer file cannot use the templating
2358 * system, so we generate a specific file for each.
2359 */
2360 static int systemd_timer_write_timer_file(enum schedule_priority schedule,
2361 int minute)
2362 {
2363 int res = -1;
2364 char *filename;
2365 FILE *file;
2366 const char *unit;
2367 char *schedule_pattern = NULL;
2368 const char *frequency = get_frequency(schedule);
2369 char *local_timer_name = xstrfmt(SYSTEMD_UNIT_FORMAT, frequency, "timer");
2370
2371 filename = xdg_config_home_systemd(local_timer_name);
2372
2373 if (safe_create_leading_directories(filename)) {
2374 error(_("failed to create directories for '%s'"), filename);
2375 goto error;
2376 }
2377 file = fopen_or_warn(filename, "w");
2378 if (!file)
2379 goto error;
2380
2381 switch (schedule) {
2382 case SCHEDULE_HOURLY:
2383 schedule_pattern = xstrfmt("*-*-* 1..23:%02d:00", minute);
2384 break;
2385
2386 case SCHEDULE_DAILY:
2387 schedule_pattern = xstrfmt("Tue..Sun *-*-* 0:%02d:00", minute);
2388 break;
2389
2390 case SCHEDULE_WEEKLY:
2391 schedule_pattern = xstrfmt("Mon 0:%02d:00", minute);
2392 break;
2393
2394 default:
2395 BUG("Unhandled schedule_priority");
2396 }
2397
2398 unit = "# This file was created and is maintained by Git.\n"
2399 "# Any edits made in this file might be replaced in the future\n"
2400 "# by a Git command.\n"
2401 "\n"
2402 "[Unit]\n"
2403 "Description=Optimize Git repositories data\n"
2404 "\n"
2405 "[Timer]\n"
2406 "OnCalendar=%s\n"
2407 "Persistent=true\n"
2408 "\n"
2409 "[Install]\n"
2410 "WantedBy=timers.target\n";
2411 if (fprintf(file, unit, schedule_pattern) < 0) {
2412 error(_("failed to write to '%s'"), filename);
2413 fclose(file);
2414 goto error;
2415 }
2416 if (fclose(file) == EOF) {
2417 error_errno(_("failed to flush '%s'"), filename);
2418 goto error;
2419 }
2420
2421 res = 0;
2422
2423 error:
2424 free(schedule_pattern);
2425 free(local_timer_name);
2426 free(filename);
2427 return res;
2428 }
2429
2430 /*
2431 * No matter the schedule, we use the same service and can make use of the
2432 * templating system. When installing git-maintenance@<schedule>.timer,
2433 * systemd will notice that git-maintenance@.service exists as a template
2434 * and will use this file and insert the <schedule> into the template at
2435 * the position of "%i".
2436 */
2437 static int systemd_timer_write_service_template(const char *exec_path)
2438 {
2439 int res = -1;
2440 char *filename;
2441 FILE *file;
2442 const char *unit;
2443 char *local_service_name = xstrfmt(SYSTEMD_UNIT_FORMAT, "", "service");
2444
2445 filename = xdg_config_home_systemd(local_service_name);
2446 if (safe_create_leading_directories(filename)) {
2447 error(_("failed to create directories for '%s'"), filename);
2448 goto error;
2449 }
2450 file = fopen_or_warn(filename, "w");
2451 if (!file)
2452 goto error;
2453
2454 unit = "# This file was created and is maintained by Git.\n"
2455 "# Any edits made in this file might be replaced in the future\n"
2456 "# by a Git command.\n"
2457 "\n"
2458 "[Unit]\n"
2459 "Description=Optimize Git repositories data\n"
2460 "\n"
2461 "[Service]\n"
2462 "Type=oneshot\n"
2463 "ExecStart=\"%s/git\" --exec-path=\"%s\" for-each-repo --keep-going --config=maintenance.repo maintenance run --schedule=%%i\n"
2464 "LockPersonality=yes\n"
2465 "MemoryDenyWriteExecute=yes\n"
2466 "NoNewPrivileges=yes\n"
2467 "RestrictAddressFamilies=AF_UNIX AF_INET AF_INET6 AF_VSOCK\n"
2468 "RestrictNamespaces=yes\n"
2469 "RestrictRealtime=yes\n"
2470 "RestrictSUIDSGID=yes\n"
2471 "SystemCallArchitectures=native\n"
2472 "SystemCallFilter=@system-service\n";
2473 if (fprintf(file, unit, exec_path, exec_path) < 0) {
2474 error(_("failed to write to '%s'"), filename);
2475 fclose(file);
2476 goto error;
2477 }
2478 if (fclose(file) == EOF) {
2479 error_errno(_("failed to flush '%s'"), filename);
2480 goto error;
2481 }
2482
2483 res = 0;
2484
2485 error:
2486 free(local_service_name);
2487 free(filename);
2488 return res;
2489 }
2490
2491 static int systemd_timer_enable_unit(int enable,
2492 enum schedule_priority schedule,
2493 int minute)
2494 {
2495 const char *cmd = "systemctl";
2496 struct child_process child = CHILD_PROCESS_INIT;
2497 const char *frequency = get_frequency(schedule);
2498
2499 /*
2500 * Disabling the systemd unit while it is already disabled makes
2501 * systemctl print an error.
2502 * Let's ignore it since it means we already are in the expected state:
2503 * the unit is disabled.
2504 *
2505 * On the other hand, enabling a systemd unit which is already enabled
2506 * produces no error.
2507 */
2508 if (!enable)
2509 child.no_stderr = 1;
2510 else if (systemd_timer_write_timer_file(schedule, minute))
2511 return -1;
2512
2513 get_schedule_cmd(&cmd, NULL);
2514 strvec_split(&child.args, cmd);
2515 strvec_pushl(&child.args, "--user", enable ? "enable" : "disable",
2516 "--now", NULL);
2517 strvec_pushf(&child.args, SYSTEMD_UNIT_FORMAT, frequency, "timer");
2518
2519 if (start_command(&child))
2520 return error(_("failed to start systemctl"));
2521 if (finish_command(&child))
2522 /*
2523 * Disabling an already disabled systemd unit makes
2524 * systemctl fail.
2525 * Let's ignore this failure.
2526 *
2527 * Enabling an enabled systemd unit doesn't fail.
2528 */
2529 if (enable)
2530 return error(_("failed to run systemctl"));
2531 return 0;
2532 }
2533
2534 /*
2535 * A previous version of Git wrote the timer units as template files.
2536 * Clean these up, if they exist.
2537 */
2538 static void systemd_timer_delete_stale_timer_templates(void)
2539 {
2540 char *timer_template_name = xstrfmt(SYSTEMD_UNIT_FORMAT, "", "timer");
2541 char *filename = xdg_config_home_systemd(timer_template_name);
2542
2543 if (unlink(filename) && !is_missing_file_error(errno))
2544 warning(_("failed to delete '%s'"), filename);
2545
2546 free(filename);
2547 free(timer_template_name);
2548 }
2549
2550 static int systemd_timer_delete_unit_files(void)
2551 {
2552 systemd_timer_delete_stale_timer_templates();
2553
2554 /* Purposefully not short-circuited to make sure all are called. */
2555 return systemd_timer_delete_timer_file(SCHEDULE_HOURLY) |
2556 systemd_timer_delete_timer_file(SCHEDULE_DAILY) |
2557 systemd_timer_delete_timer_file(SCHEDULE_WEEKLY) |
2558 systemd_timer_delete_service_template();
2559 }
2560
2561 static int systemd_timer_delete_units(void)
2562 {
2563 int minute = get_random_minute();
2564 /* Purposefully not short-circuited to make sure all are called. */
2565 return systemd_timer_enable_unit(0, SCHEDULE_HOURLY, minute) |
2566 systemd_timer_enable_unit(0, SCHEDULE_DAILY, minute) |
2567 systemd_timer_enable_unit(0, SCHEDULE_WEEKLY, minute) |
2568 systemd_timer_delete_unit_files();
2569 }
2570
2571 static int systemd_timer_setup_units(void)
2572 {
2573 int minute = get_random_minute();
2574 const char *exec_path = git_exec_path();
2575
2576 int ret = systemd_timer_write_service_template(exec_path) ||
2577 systemd_timer_enable_unit(1, SCHEDULE_HOURLY, minute) ||
2578 systemd_timer_enable_unit(1, SCHEDULE_DAILY, minute) ||
2579 systemd_timer_enable_unit(1, SCHEDULE_WEEKLY, minute);
2580
2581 if (ret)
2582 systemd_timer_delete_units();
2583 else
2584 systemd_timer_delete_stale_timer_templates();
2585
2586 return ret;
2587 }
2588
2589 static int systemd_timer_update_schedule(int run_maintenance, int fd UNUSED)
2590 {
2591 if (run_maintenance)
2592 return systemd_timer_setup_units();
2593 else
2594 return systemd_timer_delete_units();
2595 }
2596
2597 enum scheduler {
2598 SCHEDULER_INVALID = -1,
2599 SCHEDULER_AUTO,
2600 SCHEDULER_CRON,
2601 SCHEDULER_SYSTEMD,
2602 SCHEDULER_LAUNCHCTL,
2603 SCHEDULER_SCHTASKS,
2604 };
2605
2606 static const struct {
2607 const char *name;
2608 int (*is_available)(void);
2609 int (*update_schedule)(int run_maintenance, int fd);
2610 } scheduler_fn[] = {
2611 [SCHEDULER_CRON] = {
2612 .name = "crontab",
2613 .is_available = is_crontab_available,
2614 .update_schedule = crontab_update_schedule,
2615 },
2616 [SCHEDULER_SYSTEMD] = {
2617 .name = "systemctl",
2618 .is_available = is_systemd_timer_available,
2619 .update_schedule = systemd_timer_update_schedule,
2620 },
2621 [SCHEDULER_LAUNCHCTL] = {
2622 .name = "launchctl",
2623 .is_available = is_launchctl_available,
2624 .update_schedule = launchctl_update_schedule,
2625 },
2626 [SCHEDULER_SCHTASKS] = {
2627 .name = "schtasks",
2628 .is_available = is_schtasks_available,
2629 .update_schedule = schtasks_update_schedule,
2630 },
2631 };
2632
2633 static enum scheduler parse_scheduler(const char *value)
2634 {
2635 if (!value)
2636 return SCHEDULER_INVALID;
2637 else if (!strcasecmp(value, "auto"))
2638 return SCHEDULER_AUTO;
2639 else if (!strcasecmp(value, "cron") || !strcasecmp(value, "crontab"))
2640 return SCHEDULER_CRON;
2641 else if (!strcasecmp(value, "systemd") ||
2642 !strcasecmp(value, "systemd-timer"))
2643 return SCHEDULER_SYSTEMD;
2644 else if (!strcasecmp(value, "launchctl"))
2645 return SCHEDULER_LAUNCHCTL;
2646 else if (!strcasecmp(value, "schtasks"))
2647 return SCHEDULER_SCHTASKS;
2648 else
2649 return SCHEDULER_INVALID;
2650 }
2651
2652 static int maintenance_opt_scheduler(const struct option *opt, const char *arg,
2653 int unset)
2654 {
2655 enum scheduler *scheduler = opt->value;
2656
2657 BUG_ON_OPT_NEG(unset);
2658
2659 *scheduler = parse_scheduler(arg);
2660 if (*scheduler == SCHEDULER_INVALID)
2661 return error(_("unrecognized --scheduler argument '%s'"), arg);
2662 return 0;
2663 }
2664
2665 struct maintenance_start_opts {
2666 enum scheduler scheduler;
2667 };
2668
2669 static enum scheduler resolve_scheduler(enum scheduler scheduler)
2670 {
2671 if (scheduler != SCHEDULER_AUTO)
2672 return scheduler;
2673
2674 #if defined(__APPLE__)
2675 return SCHEDULER_LAUNCHCTL;
2676
2677 #elif defined(GIT_WINDOWS_NATIVE)
2678 return SCHEDULER_SCHTASKS;
2679
2680 #elif defined(__linux__)
2681 if (is_systemd_timer_available())
2682 return SCHEDULER_SYSTEMD;
2683 else if (is_crontab_available())
2684 return SCHEDULER_CRON;
2685 else
2686 die(_("neither systemd timers nor crontab are available"));
2687
2688 #else
2689 return SCHEDULER_CRON;
2690 #endif
2691 }
2692
2693 static void validate_scheduler(enum scheduler scheduler)
2694 {
2695 if (scheduler == SCHEDULER_INVALID)
2696 BUG("invalid scheduler");
2697 if (scheduler == SCHEDULER_AUTO)
2698 BUG("resolve_scheduler should have been called before");
2699
2700 if (!scheduler_fn[scheduler].is_available())
2701 die(_("%s scheduler is not available"),
2702 scheduler_fn[scheduler].name);
2703 }
2704
2705 static int update_background_schedule(const struct maintenance_start_opts *opts,
2706 int enable)
2707 {
2708 unsigned int i;
2709 int result = 0;
2710 struct lock_file lk;
2711 char *lock_path = xstrfmt("%s/schedule", the_repository->objects->odb->path);
2712
2713 if (hold_lock_file_for_update(&lk, lock_path, LOCK_NO_DEREF) < 0) {
2714 free(lock_path);
2715 return error(_("another process is scheduling background maintenance"));
2716 }
2717
2718 for (i = 1; i < ARRAY_SIZE(scheduler_fn); i++) {
2719 if (enable && opts->scheduler == i)
2720 continue;
2721 if (!scheduler_fn[i].is_available())
2722 continue;
2723 scheduler_fn[i].update_schedule(0, get_lock_file_fd(&lk));
2724 }
2725
2726 if (enable)
2727 result = scheduler_fn[opts->scheduler].update_schedule(
2728 1, get_lock_file_fd(&lk));
2729
2730 rollback_lock_file(&lk);
2731
2732 free(lock_path);
2733 return result;
2734 }
2735
2736 static const char *const builtin_maintenance_start_usage[] = {
2737 N_("git maintenance start [--scheduler=<scheduler>]"),
2738 NULL
2739 };
2740
2741 static int maintenance_start(int argc, const char **argv, const char *prefix)
2742 {
2743 struct maintenance_start_opts opts = { 0 };
2744 struct option options[] = {
2745 OPT_CALLBACK_F(
2746 0, "scheduler", &opts.scheduler, N_("scheduler"),
2747 N_("scheduler to trigger git maintenance run"),
2748 PARSE_OPT_NONEG, maintenance_opt_scheduler),
2749 OPT_END()
2750 };
2751 const char *register_args[] = { "register", NULL };
2752
2753 argc = parse_options(argc, argv, prefix, options,
2754 builtin_maintenance_start_usage, 0);
2755 if (argc)
2756 usage_with_options(builtin_maintenance_start_usage, options);
2757
2758 opts.scheduler = resolve_scheduler(opts.scheduler);
2759 validate_scheduler(opts.scheduler);
2760
2761 if (update_background_schedule(&opts, 1))
2762 die(_("failed to set up maintenance schedule"));
2763
2764 if (maintenance_register(ARRAY_SIZE(register_args)-1, register_args, NULL))
2765 warning(_("failed to add repo to global config"));
2766 return 0;
2767 }
2768
2769 static const char *const builtin_maintenance_stop_usage[] = {
2770 "git maintenance stop",
2771 NULL
2772 };
2773
2774 static int maintenance_stop(int argc, const char **argv, const char *prefix)
2775 {
2776 struct option options[] = {
2777 OPT_END()
2778 };
2779 argc = parse_options(argc, argv, prefix, options,
2780 builtin_maintenance_stop_usage, 0);
2781 if (argc)
2782 usage_with_options(builtin_maintenance_stop_usage, options);
2783 return update_background_schedule(NULL, 0);
2784 }
2785
2786 static const char * const builtin_maintenance_usage[] = {
2787 N_("git maintenance <subcommand> [<options>]"),
2788 NULL,
2789 };
2790
2791 int cmd_maintenance(int argc, const char **argv, const char *prefix)
2792 {
2793 parse_opt_subcommand_fn *fn = NULL;
2794 struct option builtin_maintenance_options[] = {
2795 OPT_SUBCOMMAND("run", &fn, maintenance_run),
2796 OPT_SUBCOMMAND("start", &fn, maintenance_start),
2797 OPT_SUBCOMMAND("stop", &fn, maintenance_stop),
2798 OPT_SUBCOMMAND("register", &fn, maintenance_register),
2799 OPT_SUBCOMMAND("unregister", &fn, maintenance_unregister),
2800 OPT_END(),
2801 };
2802
2803 argc = parse_options(argc, argv, prefix, builtin_maintenance_options,
2804 builtin_maintenance_usage, 0);
2805 return fn(argc, argv, prefix);
2806 }