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