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 "cache.h"
  15 #include "parse-options.h"
  16 #include "run-command.h"
  17 #include "sigchain.h"
  18 #include "argv-array.h"
  19 #include "commit.h"
  20
  21 #define FAILED_RUN "failed to run %s"
  22
  23 static const char * const builtin_gc_usage[] = {
  24         N_("git gc [options]"),
  25         NULL
  26 };
  27
  28 static int pack_refs = 1;
  29 static int aggressive_depth = 250;
  30 static int aggressive_window = 250;
  31 static int gc_auto_threshold = 6700;
  32 static int gc_auto_pack_limit = 50;
  33 static int detach_auto = 1;
  34 static const char *prune_expire = "2.weeks.ago";
  35
  36 static struct argv_array pack_refs_cmd = ARGV_ARRAY_INIT;
  37 static struct argv_array reflog = ARGV_ARRAY_INIT;
  38 static struct argv_array repack = ARGV_ARRAY_INIT;
  39 static struct argv_array prune = ARGV_ARRAY_INIT;
  40 static struct argv_array rerere = ARGV_ARRAY_INIT;
  41
  42 static char *pidfile;
  43
  44 static void remove_pidfile(void)
  45 {
  46         if (pidfile)
  47                 unlink(pidfile);
  48 }
  49
  50 static void remove_pidfile_on_signal(int signo)
  51 {
  52         remove_pidfile();
  53         sigchain_pop(signo);
  54         raise(signo);
  55 }
  56
  57 static int gc_config(const char *var, const char *value, void *cb)
  58 {
  59         if (!strcmp(var, "gc.packrefs")) {
  60                 if (value && !strcmp(value, "notbare"))
  61                         pack_refs = -1;
  62                 else
  63                         pack_refs = git_config_bool(var, value);
  64                 return 0;
  65         }
  66         if (!strcmp(var, "gc.aggressivewindow")) {
  67                 aggressive_window = git_config_int(var, value);
  68                 return 0;
  69         }
  70         if (!strcmp(var, "gc.aggressivedepth")) {
  71                 aggressive_depth = git_config_int(var, value);
  72                 return 0;
  73         }
  74         if (!strcmp(var, "gc.auto")) {
  75                 gc_auto_threshold = git_config_int(var, value);
  76                 return 0;
  77         }
  78         if (!strcmp(var, "gc.autopacklimit")) {
  79                 gc_auto_pack_limit = git_config_int(var, value);
  80                 return 0;
  81         }
  82         if (!strcmp(var, "gc.autodetach")) {
  83                 detach_auto = git_config_bool(var, value);
  84                 return 0;
  85         }
  86         if (!strcmp(var, "gc.pruneexpire")) {
  87                 if (value && strcmp(value, "now")) {
  88                         unsigned long now = approxidate("now");
  89                         if (approxidate(value) >= now)
  90                                 return error(_("Invalid %s: '%s'"), var, value);
  91                 }
  92                 return git_config_string(&prune_expire, var, value);
  93         }
  94         return git_default_config(var, value, cb);
  95 }
  96
  97 static int too_many_loose_objects(void)
  98 {
  99         /*
 100          * Quickly check if a "gc" is needed, by estimating how
 101          * many loose objects there are.  Because SHA-1 is evenly
 102          * distributed, we can check only one and get a reasonable
 103          * estimate.
 104          */
 105         char path[PATH_MAX];
 106         const char *objdir = get_object_directory();
 107         DIR *dir;
 108         struct dirent *ent;
 109         int auto_threshold;
 110         int num_loose = 0;
 111         int needed = 0;
 112
 113         if (gc_auto_threshold <= 0)
 114                 return 0;
 115
 116         if (sizeof(path) <= snprintf(path, sizeof(path), "%s/17", objdir)) {
 117                 warning(_("insanely long object directory %.*s"), 50, objdir);
 118                 return 0;
 119         }
 120         dir = opendir(path);
 121         if (!dir)
 122                 return 0;
 123
 124         auto_threshold = (gc_auto_threshold + 255) / 256;
 125         while ((ent = readdir(dir)) != NULL) {
 126                 if (strspn(ent->d_name, "0123456789abcdef") != 38 ||
 127                     ent->d_name[38] != '\0')
 128                         continue;
 129                 if (++num_loose > auto_threshold) {
 130                         needed = 1;
 131                         break;
 132                 }
 133         }
 134         closedir(dir);
 135         return needed;
 136 }
 137
 138 static int too_many_packs(void)
 139 {
 140         struct packed_git *p;
 141         int cnt;
 142
 143         if (gc_auto_pack_limit <= 0)
 144                 return 0;
 145
 146         prepare_packed_git();
 147         for (cnt = 0, p = packed_git; p; p = p->next) {
 148                 if (!p->pack_local)
 149                         continue;
 150                 if (p->pack_keep)
 151                         continue;
 152                 /*
 153                  * Perhaps check the size of the pack and count only
 154                  * very small ones here?
 155                  */
 156                 cnt++;
 157         }
 158         return gc_auto_pack_limit <= cnt;
 159 }
 160
 161 static void add_repack_all_option(void)
 162 {
 163         if (prune_expire && !strcmp(prune_expire, "now"))
 164                 argv_array_push(&repack, "-a");
 165         else {
 166                 argv_array_push(&repack, "-A");
 167                 if (prune_expire)
 168                         argv_array_pushf(&repack, "--unpack-unreachable=%s", prune_expire);
 169         }
 170 }
 171
 172 static int need_to_gc(void)
 173 {
 174         /*
 175          * Setting gc.auto to 0 or negative can disable the
 176          * automatic gc.
 177          */
 178         if (gc_auto_threshold <= 0)
 179                 return 0;
 180
 181         /*
 182          * If there are too many loose objects, but not too many
 183          * packs, we run "repack -d -l".  If there are too many packs,
 184          * we run "repack -A -d -l".  Otherwise we tell the caller
 185          * there is no need.
 186          */
 187         if (too_many_packs())
 188                 add_repack_all_option();
 189         else if (!too_many_loose_objects())
 190                 return 0;
 191
 192         if (run_hook_le(NULL, "pre-auto-gc", NULL))
 193                 return 0;
 194         return 1;
 195 }
 196
 197 /* return NULL on success, else hostname running the gc */
 198 static const char *lock_repo_for_gc(int force, pid_t* ret_pid)
 199 {
 200         static struct lock_file lock;
 201         char my_host[128];
 202         struct strbuf sb = STRBUF_INIT;
 203         struct stat st;
 204         uintmax_t pid;
 205         FILE *fp;
 206         int fd;
 207
 208         if (pidfile)
 209                 /* already locked */
 210                 return NULL;
 211
 212         if (gethostname(my_host, sizeof(my_host)))
 213                 strcpy(my_host, "unknown");
 214
 215         fd = hold_lock_file_for_update(&lock, git_path("gc.pid"),
 216                                        LOCK_DIE_ON_ERROR);
 217         if (!force) {
 218                 static char locking_host[128];
 219                 int should_exit;
 220                 fp = fopen(git_path("gc.pid"), "r");
 221                 memset(locking_host, 0, sizeof(locking_host));
 222                 should_exit =
 223                         fp != NULL &&
 224                         !fstat(fileno(fp), &st) &&
 225                         /*
 226                          * 12 hour limit is very generous as gc should
 227                          * never take that long. On the other hand we
 228                          * don't really need a strict limit here,
 229                          * running gc --auto one day late is not a big
 230                          * problem. --force can be used in manual gc
 231                          * after the user verifies that no gc is
 232                          * running.
 233                          */
 234                         time(NULL) - st.st_mtime <= 12 * 3600 &&
 235                         fscanf(fp, "%"PRIuMAX" %127c", &pid, locking_host) == 2 &&
 236                         /* be gentle to concurrent "gc" on remote hosts */
 237                         (strcmp(locking_host, my_host) || !kill(pid, 0) || errno == EPERM);
 238                 if (fp != NULL)
 239                         fclose(fp);
 240                 if (should_exit) {
 241                         if (fd >= 0)
 242                                 rollback_lock_file(&lock);
 243                         *ret_pid = pid;
 244                         return locking_host;
 245                 }
 246         }
 247
 248         strbuf_addf(&sb, "%"PRIuMAX" %s",
 249                     (uintmax_t) getpid(), my_host);
 250         write_in_full(fd, sb.buf, sb.len);
 251         strbuf_release(&sb);
 252         commit_lock_file(&lock);
 253
 254         pidfile = git_pathdup("gc.pid");
 255         sigchain_push_common(remove_pidfile_on_signal);
 256         atexit(remove_pidfile);
 257
 258         return NULL;
 259 }
 260
 261 int cmd_gc(int argc, const char **argv, const char *prefix)
 262 {
 263         int aggressive = 0;
 264         int auto_gc = 0;
 265         int quiet = 0;
 266         int force = 0;
 267         const char *name;
 268         pid_t pid;
 269
 270         struct option builtin_gc_options[] = {
 271                 OPT__QUIET(&quiet, N_("suppress progress reporting")),
 272                 { OPTION_STRING, 0, "prune", &prune_expire, N_("date"),
 273                         N_("prune unreferenced objects"),
 274                         PARSE_OPT_OPTARG, NULL, (intptr_t)prune_expire },
 275                 OPT_BOOL(0, "aggressive", &aggressive, N_("be more thorough (increased runtime)")),
 276                 OPT_BOOL(0, "auto", &auto_gc, N_("enable auto-gc mode")),
 277                 OPT_BOOL(0, "force", &force, N_("force running gc even if there may be another gc running")),
 278                 OPT_END()
 279         };
 280
 281         if (argc == 2 && !strcmp(argv[1], "-h"))
 282                 usage_with_options(builtin_gc_usage, builtin_gc_options);
 283
 284         argv_array_pushl(&pack_refs_cmd, "pack-refs", "--all", "--prune", NULL);
 285         argv_array_pushl(&reflog, "reflog", "expire", "--all", NULL);
 286         argv_array_pushl(&repack, "repack", "-d", "-l", NULL);
 287         argv_array_pushl(&prune, "prune", "--expire", NULL );
 288         argv_array_pushl(&rerere, "rerere", "gc", NULL);
 289
 290         git_config(gc_config, NULL);
 291
 292         if (pack_refs < 0)
 293                 pack_refs = !is_bare_repository();
 294
 295         argc = parse_options(argc, argv, prefix, builtin_gc_options,
 296                              builtin_gc_usage, 0);
 297         if (argc > 0)
 298                 usage_with_options(builtin_gc_usage, builtin_gc_options);
 299
 300         if (aggressive) {
 301                 argv_array_push(&repack, "-f");
 302                 if (aggressive_depth > 0)
 303                         argv_array_pushf(&repack, "--depth=%d", aggressive_depth);
 304                 if (aggressive_window > 0)
 305                         argv_array_pushf(&repack, "--window=%d", aggressive_window);
 306         }
 307         if (quiet)
 308                 argv_array_push(&repack, "-q");
 309
 310         if (auto_gc) {
 311                 /*
 312                  * Auto-gc should be least intrusive as possible.
 313                  */
 314                 if (!need_to_gc())
 315                         return 0;
 316                 if (!quiet) {
 317                         if (detach_auto)
 318                                 fprintf(stderr, _("Auto packing the repository in background for optimum performance.\n"));
 319                         else
 320                                 fprintf(stderr, _("Auto packing the repository for optimum performance.\n"));
 321                         fprintf(stderr, _("See \"git help gc\" for manual housekeeping.\n"));
 322                 }
 323                 if (detach_auto)
 324                         /*
 325                          * failure to daemonize is ok, we'll continue
 326                          * in foreground
 327                          */
 328                         daemonize();
 329         } else
 330                 add_repack_all_option();
 331
 332         name = lock_repo_for_gc(force, &pid);
 333         if (name) {
 334                 if (auto_gc)
 335                         return 0; /* be quiet on --auto */
 336                 die(_("gc is already running on machine '%s' pid %"PRIuMAX" (use --force if not)"),
 337                     name, (uintmax_t)pid);
 338         }
 339
 340         if (pack_refs && run_command_v_opt(pack_refs_cmd.argv, RUN_GIT_CMD))
 341                 return error(FAILED_RUN, pack_refs_cmd.argv[0]);
 342
 343         if (run_command_v_opt(reflog.argv, RUN_GIT_CMD))
 344                 return error(FAILED_RUN, reflog.argv[0]);
 345
 346         if (run_command_v_opt(repack.argv, RUN_GIT_CMD))
 347                 return error(FAILED_RUN, repack.argv[0]);
 348
 349         if (prune_expire) {
 350                 argv_array_push(&prune, prune_expire);
 351                 if (quiet)
 352                         argv_array_push(&prune, "--no-progress");
 353                 if (run_command_v_opt(prune.argv, RUN_GIT_CMD))
 354                         return error(FAILED_RUN, prune.argv[0]);
 355         }
 356
 357         if (run_command_v_opt(rerere.argv, RUN_GIT_CMD))
 358                 return error(FAILED_RUN, rerere.argv[0]);
 359
 360         if (auto_gc && too_many_loose_objects())
 361                 warning(_("There are too many unreachable loose objects; "
 362                         "run 'git prune' to remove them."));
 363
 364         return 0;
 365 }