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#include "builtin.h"
14#include "repository.h"
15#include "config.h"
16#include "tempfile.h"
17#include "lockfile.h"
18#include "parse-options.h"
19#include "run-command.h"
20#include "sigchain.h"
21#include "argv-array.h"
22#include "commit.h"
23#include "commit-graph.h"
24#include "packfile.h"
25#include "object-store.h"
26#include "pack.h"
27#include "pack-objects.h"
28#include "blob.h"
29#include "tree.h"
30#define FAILED_RUN "failed to run %s"
32static const char * const builtin_gc_usage[] = {
34 N_("git gc [<options>]"),
35 NULL
36};
37static int pack_refs = 1;
39static int prune_reflogs = 1;
40static int aggressive_depth = 50;
41static int aggressive_window = 250;
42static int gc_auto_threshold = 6700;
43static int gc_auto_pack_limit = 50;
44static int detach_auto = 1;
45static timestamp_t gc_log_expire_time;
46static const char *gc_log_expire = "1.day.ago";
47static const char *prune_expire = "2.weeks.ago";
48static const char *prune_worktrees_expire = "3.months.ago";
49static unsigned long big_pack_threshold;
50static unsigned long max_delta_cache_size = DEFAULT_DELTA_CACHE_SIZE;
51static struct argv_array pack_refs_cmd = ARGV_ARRAY_INIT;
53static struct argv_array reflog = ARGV_ARRAY_INIT;
54static struct argv_array repack = ARGV_ARRAY_INIT;
55static struct argv_array prune = ARGV_ARRAY_INIT;
56static struct argv_array prune_worktrees = ARGV_ARRAY_INIT;
57static struct argv_array rerere = ARGV_ARRAY_INIT;
58static struct tempfile *pidfile;
60static struct lock_file log_lock;
61static struct string_list pack_garbage = STRING_LIST_INIT_DUP;
63static void clean_pack_garbage(void)
65{
66 int i;
67 for (i = 0; i < pack_garbage.nr; i++)
68 unlink_or_warn(pack_garbage.items[i].string);
69 string_list_clear(&pack_garbage, 0);
70}
71static void report_pack_garbage(unsigned seen_bits, const char *path)
73{
74 if (seen_bits == PACKDIR_FILE_IDX)
75 string_list_append(&pack_garbage, path);
76}
77static void process_log_file(void)
79{
80 struct stat st;
81 if (fstat(get_lock_file_fd(&log_lock), &st)) {
82 /*
83 * Perhaps there was an i/o error or another
84 * unlikely situation. Try to make a note of
85 * this in gc.log along with any existing
86 * messages.
87 */
88 int saved_errno = errno;
89 fprintf(stderr, _("Failed to fstat %s: %s"),
90 get_tempfile_path(log_lock.tempfile),
91 strerror(saved_errno));
92 fflush(stderr);
93 commit_lock_file(&log_lock);
94 errno = saved_errno;
95 } else if (st.st_size) {
96 /* There was some error recorded in the lock file */
97 commit_lock_file(&log_lock);
98 } else {
99 /* No error, clean up any old gc.log */
100 unlink(git_path("gc.log"));
101 rollback_lock_file(&log_lock);
102 }
103}
104static void process_log_file_at_exit(void)
106{
107 fflush(stderr);
108 process_log_file();
109}
110static void process_log_file_on_signal(int signo)
112{
113 process_log_file();
114 sigchain_pop(signo);
115 raise(signo);
116}
117static int gc_config_is_timestamp_never(const char *var)
119{
120 const char *value;
121 timestamp_t expire;
122 if (!git_config_get_value(var, &value) && value) {
124 if (parse_expiry_date(value, &expire))
125 die(_("failed to parse '%s' value '%s'"), var, value);
126 return expire == 0;
127 }
128 return 0;
129}
130static void gc_config(void)
132{
133 const char *value;
134 if (!git_config_get_value("gc.packrefs", &value)) {
136 if (value && !strcmp(value, "notbare"))
137 pack_refs = -1;
138 else
139 pack_refs = git_config_bool("gc.packrefs", value);
140 }
141 if (gc_config_is_timestamp_never("gc.reflogexpire") &&
143 gc_config_is_timestamp_never("gc.reflogexpireunreachable"))
144 prune_reflogs = 0;
145 git_config_get_int("gc.aggressivewindow", &aggressive_window);
147 git_config_get_int("gc.aggressivedepth", &aggressive_depth);
148 git_config_get_int("gc.auto", &gc_auto_threshold);
149 git_config_get_int("gc.autopacklimit", &gc_auto_pack_limit);
150 git_config_get_bool("gc.autodetach", &detach_auto);
151 git_config_get_expiry("gc.pruneexpire", &prune_expire);
152 git_config_get_expiry("gc.worktreepruneexpire", &prune_worktrees_expire);
153 git_config_get_expiry("gc.logexpiry", &gc_log_expire);
154 git_config_get_ulong("gc.bigpackthreshold", &big_pack_threshold);
156 git_config_get_ulong("pack.deltacachesize", &max_delta_cache_size);
157 git_config(git_default_config, NULL);
159}
160static int too_many_loose_objects(void)
162{
163 /*
164 * Quickly check if a "gc" is needed, by estimating how
165 * many loose objects there are. Because SHA-1 is evenly
166 * distributed, we can check only one and get a reasonable
167 * estimate.
168 */
169 DIR *dir;
170 struct dirent *ent;
171 int auto_threshold;
172 int num_loose = 0;
173 int needed = 0;
174 const unsigned hexsz_loose = the_hash_algo->hexsz - 2;
175 dir = opendir(git_path("objects/17"));
177 if (!dir)
178 return 0;
179 auto_threshold = DIV_ROUND_UP(gc_auto_threshold, 256);
181 while ((ent = readdir(dir)) != NULL) {
182 if (strspn(ent->d_name, "0123456789abcdef") != hexsz_loose ||
183 ent->d_name[hexsz_loose] != '\0')
184 continue;
185 if (++num_loose > auto_threshold) {
186 needed = 1;
187 break;
188 }
189 }
190 closedir(dir);
191 return needed;
192}
193static struct packed_git *find_base_packs(struct string_list *packs,
195 unsigned long limit)
196{
197 struct packed_git *p, *base = NULL;
198 for (p = get_all_packs(the_repository); p; p = p->next) {
200 if (!p->pack_local)
201 continue;
202 if (limit) {
203 if (p->pack_size >= limit)
204 string_list_append(packs, p->pack_name);
205 } else if (!base || base->pack_size < p->pack_size) {
206 base = p;
207 }
208 }
209 if (base)
211 string_list_append(packs, base->pack_name);
212 return base;
214}
215static int too_many_packs(void)
217{
218 struct packed_git *p;
219 int cnt;
220 if (gc_auto_pack_limit <= 0)
222 return 0;
223 for (cnt = 0, p = get_all_packs(the_repository); p; p = p->next) {
225 if (!p->pack_local)
226 continue;
227 if (p->pack_keep)
228 continue;
229 /*
230 * Perhaps check the size of the pack and count only
231 * very small ones here?
232 */
233 cnt++;
234 }
235 return gc_auto_pack_limit < cnt;
236}
237static uint64_t total_ram(void)
239{
240#if defined(HAVE_SYSINFO)
241 struct sysinfo si;
242 if (!sysinfo(&si))
244 return si.totalram;
245#elif defined(HAVE_BSD_SYSCTL) && (defined(HW_MEMSIZE) || defined(HW_PHYSMEM))
246 int64_t physical_memory;
247 int mib[2];
248 size_t length;
249 mib[0] = CTL_HW;
251# if defined(HW_MEMSIZE)
252 mib[1] = HW_MEMSIZE;
253# else
254 mib[1] = HW_PHYSMEM;
255# endif
256 length = sizeof(int64_t);
257 if (!sysctl(mib, 2, &physical_memory, &length, NULL, 0))
258 return physical_memory;
259#elif defined(GIT_WINDOWS_NATIVE)
260 MEMORYSTATUSEX memInfo;
261 memInfo.dwLength = sizeof(MEMORYSTATUSEX);
263 if (GlobalMemoryStatusEx(&memInfo))
264 return memInfo.ullTotalPhys;
265#endif
266 return 0;
267}
268static uint64_t estimate_repack_memory(struct packed_git *pack)
270{
271 unsigned long nr_objects = approximate_object_count();
272 size_t os_cache, heap;
273 if (!pack || !nr_objects)
275 return 0;
276 /*
278 * First we have to scan through at least one pack.
279 * Assume enough room in OS file cache to keep the entire pack
280 * or we may accidentally evict data of other processes from
281 * the cache.
282 */
283 os_cache = pack->pack_size + pack->index_size;
284 /* then pack-objects needs lots more for book keeping */
285 heap = sizeof(struct object_entry) * nr_objects;
286 /*
287 * internal rev-list --all --objects takes up some memory too,
288 * let's say half of it is for blobs
289 */
290 heap += sizeof(struct blob) * nr_objects / 2;
291 /*
292 * and the other half is for trees (commits and tags are
293 * usually insignificant)
294 */
295 heap += sizeof(struct tree) * nr_objects / 2;
296 /* and then obj_hash[], underestimated in fact */
297 heap += sizeof(struct object *) * nr_objects;
298 /* revindex is used also */
299 heap += sizeof(struct revindex_entry) * nr_objects;
300 /*
301 * read_sha1_file() (either at delta calculation phase, or
302 * writing phase) also fills up the delta base cache
303 */
304 heap += delta_base_cache_limit;
305 /* and of course pack-objects has its own delta cache */
306 heap += max_delta_cache_size;
307 return os_cache + heap;
309}
310static int keep_one_pack(struct string_list_item *item, void *data)
312{
313 argv_array_pushf(&repack, "--keep-pack=%s", basename(item->string));
314 return 0;
315}
316static void add_repack_all_option(struct string_list *keep_pack)
318{
319 if (prune_expire && !strcmp(prune_expire, "now"))
320 argv_array_push(&repack, "-a");
321 else {
322 argv_array_push(&repack, "-A");
323 if (prune_expire)
324 argv_array_pushf(&repack, "--unpack-unreachable=%s", prune_expire);
325 }
326 if (keep_pack)
328 for_each_string_list(keep_pack, keep_one_pack, NULL);
329}
330static void add_repack_incremental_option(void)
332{
333 argv_array_push(&repack, "--no-write-bitmap-index");
334}
335static int need_to_gc(void)
337{
338 /*
339 * Setting gc.auto to 0 or negative can disable the
340 * automatic gc.
341 */
342 if (gc_auto_threshold <= 0)
343 return 0;
344 /*
346 * If there are too many loose objects, but not too many
347 * packs, we run "repack -d -l". If there are too many packs,
348 * we run "repack -A -d -l". Otherwise we tell the caller
349 * there is no need.
350 */
351 if (too_many_packs()) {
352 struct string_list keep_pack = STRING_LIST_INIT_NODUP;
353 if (big_pack_threshold) {
355 find_base_packs(&keep_pack, big_pack_threshold);
356 if (keep_pack.nr >= gc_auto_pack_limit) {
357 big_pack_threshold = 0;
358 string_list_clear(&keep_pack, 0);
359 find_base_packs(&keep_pack, 0);
360 }
361 } else {
362 struct packed_git *p = find_base_packs(&keep_pack, 0);
363 uint64_t mem_have, mem_want;
364 mem_have = total_ram();
366 mem_want = estimate_repack_memory(p);
367 /*
369 * Only allow 1/2 of memory for pack-objects, leave
370 * the rest for the OS and other processes in the
371 * system.
372 */
373 if (!mem_have || mem_want < mem_have / 2)
374 string_list_clear(&keep_pack, 0);
375 }
376 add_repack_all_option(&keep_pack);
378 string_list_clear(&keep_pack, 0);
379 } else if (too_many_loose_objects())
380 add_repack_incremental_option();
381 else
382 return 0;
383 if (run_hook_le(NULL, "pre-auto-gc", NULL))
385 return 0;
386 return 1;
387}
388/* return NULL on success, else hostname running the gc */
390static const char *lock_repo_for_gc(int force, pid_t* ret_pid)
391{
392 struct lock_file lock = LOCK_INIT;
393 char my_host[HOST_NAME_MAX + 1];
394 struct strbuf sb = STRBUF_INIT;
395 struct stat st;
396 uintmax_t pid;
397 FILE *fp;
398 int fd;
399 char *pidfile_path;
400 if (is_tempfile_active(pidfile))
402 /* already locked */
403 return NULL;
404 if (xgethostname(my_host, sizeof(my_host)))
406 xsnprintf(my_host, sizeof(my_host), "unknown");
407 pidfile_path = git_pathdup("gc.pid");
409 fd = hold_lock_file_for_update(&lock, pidfile_path,
410 LOCK_DIE_ON_ERROR);
411 if (!force) {
412 static char locking_host[HOST_NAME_MAX + 1];
413 static char *scan_fmt;
414 int should_exit;
415 if (!scan_fmt)
417 scan_fmt = xstrfmt("%s %%%ds", "%"SCNuMAX, HOST_NAME_MAX);
418 fp = fopen(pidfile_path, "r");
419 memset(locking_host, 0, sizeof(locking_host));
420 should_exit =
421 fp != NULL &&
422 !fstat(fileno(fp), &st) &&
423 /*
424 * 12 hour limit is very generous as gc should
425 * never take that long. On the other hand we
426 * don't really need a strict limit here,
427 * running gc --auto one day late is not a big
428 * problem. --force can be used in manual gc
429 * after the user verifies that no gc is
430 * running.
431 */
432 time(NULL) - st.st_mtime <= 12 * 3600 &&
433 fscanf(fp, scan_fmt, &pid, locking_host) == 2 &&
434 /* be gentle to concurrent "gc" on remote hosts */
435 (strcmp(locking_host, my_host) || !kill(pid, 0) || errno == EPERM);
436 if (fp != NULL)
437 fclose(fp);
438 if (should_exit) {
439 if (fd >= 0)
440 rollback_lock_file(&lock);
441 *ret_pid = pid;
442 free(pidfile_path);
443 return locking_host;
444 }
445 }
446 strbuf_addf(&sb, "%"PRIuMAX" %s",
448 (uintmax_t) getpid(), my_host);
449 write_in_full(fd, sb.buf, sb.len);
450 strbuf_release(&sb);
451 commit_lock_file(&lock);
452 pidfile = register_tempfile(pidfile_path);
453 free(pidfile_path);
454 return NULL;
455}
456/*
458 * Returns 0 if there was no previous error and gc can proceed, 1 if
459 * gc should not proceed due to an error in the last run. Prints a
460 * message and returns -1 if an error occured while reading gc.log
461 */
462static int report_last_gc_error(void)
463{
464 struct strbuf sb = STRBUF_INIT;
465 int ret = 0;
466 ssize_t len;
467 struct stat st;
468 char *gc_log_path = git_pathdup("gc.log");
469 if (stat(gc_log_path, &st)) {
471 if (errno == ENOENT)
472 goto done;
473 ret = error_errno(_("cannot stat '%s'"), gc_log_path);
475 goto done;
476 }
477 if (st.st_mtime < gc_log_expire_time)
479 goto done;
480 len = strbuf_read_file(&sb, gc_log_path, 0);
482 if (len < 0)
483 ret = error_errno(_("cannot read '%s'"), gc_log_path);
484 else if (len > 0) {
485 /*
486 * A previous gc failed. Report the error, and don't
487 * bother with an automatic gc run since it is likely
488 * to fail in the same way.
489 */
490 warning(_("The last gc run reported the following. "
491 "Please correct the root cause\n"
492 "and remove %s.\n"
493 "Automatic cleanup will not be performed "
494 "until the file is removed.\n\n"
495 "%s"),
496 gc_log_path, sb.buf);
497 ret = 1;
498 }
499 strbuf_release(&sb);
500done:
501 free(gc_log_path);
502 return ret;
503}
504static void gc_before_repack(void)
506{
507 /*
508 * We may be called twice, as both the pre- and
509 * post-daemonized phases will call us, but running these
510 * commands more than once is pointless and wasteful.
511 */
512 static int done = 0;
513 if (done++)
514 return;
515 if (pack_refs && run_command_v_opt(pack_refs_cmd.argv, RUN_GIT_CMD))
517 die(FAILED_RUN, pack_refs_cmd.argv[0]);
518 if (prune_reflogs && run_command_v_opt(reflog.argv, RUN_GIT_CMD))
520 die(FAILED_RUN, reflog.argv[0]);
521}
522int cmd_gc(int argc, const char **argv, const char *prefix)
524{
525 int aggressive = 0;
526 int auto_gc = 0;
527 int quiet = 0;
528 int force = 0;
529 const char *name;
530 pid_t pid;
531 int daemonized = 0;
532 int keep_base_pack = -1;
533 timestamp_t dummy;
534 struct option builtin_gc_options[] = {
536 OPT__QUIET(&quiet, N_("suppress progress reporting")),
537 { OPTION_STRING, 0, "prune", &prune_expire, N_("date"),
538 N_("prune unreferenced objects"),
539 PARSE_OPT_OPTARG, NULL, (intptr_t)prune_expire },
540 OPT_BOOL(0, "aggressive", &aggressive, N_("be more thorough (increased runtime)")),
541 OPT_BOOL_F(0, "auto", &auto_gc, N_("enable auto-gc mode"),
542 PARSE_OPT_NOCOMPLETE),
543 OPT_BOOL_F(0, "force", &force,
544 N_("force running gc even if there may be another gc running"),
545 PARSE_OPT_NOCOMPLETE),
546 OPT_BOOL(0, "keep-largest-pack", &keep_base_pack,
547 N_("repack all other packs except the largest pack")),
548 OPT_END()
549 };
550 if (argc == 2 && !strcmp(argv[1], "-h"))
552 usage_with_options(builtin_gc_usage, builtin_gc_options);
553 argv_array_pushl(&pack_refs_cmd, "pack-refs", "--all", "--prune", NULL);
555 argv_array_pushl(&reflog, "reflog", "expire", "--all", NULL);
556 argv_array_pushl(&repack, "repack", "-d", "-l", NULL);
557 argv_array_pushl(&prune, "prune", "--expire", NULL);
558 argv_array_pushl(&prune_worktrees, "worktree", "prune", "--expire", NULL);
559 argv_array_pushl(&rerere, "rerere", "gc", NULL);
560 /* default expiry time, overwritten in gc_config */
562 gc_config();
563 if (parse_expiry_date(gc_log_expire, &gc_log_expire_time))
564 die(_("failed to parse gc.logexpiry value %s"), gc_log_expire);
565 if (pack_refs < 0)
567 pack_refs = !is_bare_repository();
568 argc = parse_options(argc, argv, prefix, builtin_gc_options,
570 builtin_gc_usage, 0);
571 if (argc > 0)
572 usage_with_options(builtin_gc_usage, builtin_gc_options);
573 if (prune_expire && parse_expiry_date(prune_expire, &dummy))
575 die(_("failed to parse prune expiry value %s"), prune_expire);
576 if (aggressive) {
578 argv_array_push(&repack, "-f");
579 if (aggressive_depth > 0)
580 argv_array_pushf(&repack, "--depth=%d", aggressive_depth);
581 if (aggressive_window > 0)
582 argv_array_pushf(&repack, "--window=%d", aggressive_window);
583 }
584 if (quiet)
585 argv_array_push(&repack, "-q");
586 if (auto_gc) {
588 /*
589 * Auto-gc should be least intrusive as possible.
590 */
591 if (!need_to_gc())
592 return 0;
593 if (!quiet) {
594 if (detach_auto)
595 fprintf(stderr, _("Auto packing the repository in background for optimum performance.\n"));
596 else
597 fprintf(stderr, _("Auto packing the repository for optimum performance.\n"));
598 fprintf(stderr, _("See \"git help gc\" for manual housekeeping.\n"));
599 }
600 if (detach_auto) {
601 int ret = report_last_gc_error();
602 if (ret < 0)
603 /* an I/O error occured, already reported */
604 exit(128);
605 if (ret == 1)
606 /* Last gc --auto failed. Skip this one. */
607 return 0;
608 if (lock_repo_for_gc(force, &pid))
610 return 0;
611 gc_before_repack(); /* dies on failure */
612 delete_tempfile(&pidfile);
613 /*
615 * failure to daemonize is ok, we'll continue
616 * in foreground
617 */
618 daemonized = !daemonize();
619 }
620 } else {
621 struct string_list keep_pack = STRING_LIST_INIT_NODUP;
622 if (keep_base_pack != -1) {
624 if (keep_base_pack)
625 find_base_packs(&keep_pack, 0);
626 } else if (big_pack_threshold) {
627 find_base_packs(&keep_pack, big_pack_threshold);
628 }
629 add_repack_all_option(&keep_pack);
631 string_list_clear(&keep_pack, 0);
632 }
633 name = lock_repo_for_gc(force, &pid);
635 if (name) {
636 if (auto_gc)
637 return 0; /* be quiet on --auto */
638 die(_("gc is already running on machine '%s' pid %"PRIuMAX" (use --force if not)"),
639 name, (uintmax_t)pid);
640 }
641 if (daemonized) {
643 hold_lock_file_for_update(&log_lock,
644 git_path("gc.log"),
645 LOCK_DIE_ON_ERROR);
646 dup2(get_lock_file_fd(&log_lock), 2);
647 sigchain_push_common(process_log_file_on_signal);
648 atexit(process_log_file_at_exit);
649 }
650 gc_before_repack();
652 if (!repository_format_precious_objects) {
654 close_object_store(the_repository->objects);
655 if (run_command_v_opt(repack.argv, RUN_GIT_CMD))
656 die(FAILED_RUN, repack.argv[0]);
657 if (prune_expire) {
659 argv_array_push(&prune, prune_expire);
660 if (quiet)
661 argv_array_push(&prune, "--no-progress");
662 if (repository_format_partial_clone)
663 argv_array_push(&prune,
664 "--exclude-promisor-objects");
665 if (run_command_v_opt(prune.argv, RUN_GIT_CMD))
666 die(FAILED_RUN, prune.argv[0]);
667 }
668 }
669 if (prune_worktrees_expire) {
671 argv_array_push(&prune_worktrees, prune_worktrees_expire);
672 if (run_command_v_opt(prune_worktrees.argv, RUN_GIT_CMD))
673 die(FAILED_RUN, prune_worktrees.argv[0]);
674 }
675 if (run_command_v_opt(rerere.argv, RUN_GIT_CMD))
677 die(FAILED_RUN, rerere.argv[0]);
678 report_garbage = report_pack_garbage;
680 reprepare_packed_git(the_repository);
681 if (pack_garbage.nr > 0) {
682 close_object_store(the_repository->objects);
683 clean_pack_garbage();
684 }
685 prepare_repo_settings(the_repository);
687 if (the_repository->settings.gc_write_commit_graph == 1)
688 write_commit_graph_reachable(get_object_directory(),
689 !quiet && !daemonized ? COMMIT_GRAPH_WRITE_PROGRESS : 0,
690 NULL);
691 if (auto_gc && too_many_loose_objects())
693 warning(_("There are too many unreachable loose objects; "
694 "run 'git prune' to remove them."));
695 if (!daemonized)
697 unlink(git_path("gc.log"));
698 return 0;
700}