builtin/fsck.c

   1 #include "builtin.h"
   2 #include "cache.h"
   3 #include "commit.h"
   4 #include "tree.h"
   5 #include "blob.h"
   6 #include "tag.h"
   7 #include "refs.h"
   8 #include "pack.h"
   9 #include "cache-tree.h"
  10 #include "tree-walk.h"
  11 #include "fsck.h"
  12 #include "parse-options.h"
  13 #include "dir.h"
  14 #include "progress.h"
  15 #include "streaming.h"
  16
  17 #define REACHABLE 0x0001
  18 #define SEEN      0x0002
  19 #define HAS_OBJ   0x0004
  20
  21 static int show_root;
  22 static int show_tags;
  23 static int show_unreachable;
  24 static int include_reflogs = 1;
  25 static int check_full = 1;
  26 static int connectivity_only;
  27 static int check_strict;
  28 static int keep_cache_objects;
  29 static struct fsck_options fsck_walk_options = FSCK_OPTIONS_DEFAULT;
  30 static struct fsck_options fsck_obj_options = FSCK_OPTIONS_DEFAULT;
  31 static struct object_id head_oid;
  32 static const char *head_points_at;
  33 static int errors_found;
  34 static int write_lost_and_found;
  35 static int verbose;
  36 static int show_progress = -1;
  37 static int show_dangling = 1;
  38 #define ERROR_OBJECT 01
  39 #define ERROR_REACHABLE 02
  40 #define ERROR_PACK 04
  41 #define ERROR_REFS 010
  42
  43 static const char *describe_object(struct object *obj)
  44 {
  45         return oid_to_hex(&obj->oid);
  46 }
  47
  48 static int fsck_config(const char *var, const char *value, void *cb)
  49 {
  50         if (strcmp(var, "fsck.skiplist") == 0) {
  51                 const char *path;
  52                 struct strbuf sb = STRBUF_INIT;
  53
  54                 if (git_config_pathname(&path, var, value))
  55                         return 1;
  56                 strbuf_addf(&sb, "skiplist=%s", path);
  57                 free((char *)path);
  58                 fsck_set_msg_types(&fsck_obj_options, sb.buf);
  59                 strbuf_release(&sb);
  60                 return 0;
  61         }
  62
  63         if (skip_prefix(var, "fsck.", &var)) {
  64                 fsck_set_msg_type(&fsck_obj_options, var, value);
  65                 return 0;
  66         }
  67
  68         return git_default_config(var, value, cb);
  69 }
  70
  71 static void objreport(struct object *obj, const char *msg_type,
  72                         const char *err)
  73 {
  74         fprintf(stderr, "%s in %s %s: %s\n",
  75                 msg_type, typename(obj->type), describe_object(obj), err);
  76 }
  77
  78 static int objerror(struct object *obj, const char *err)
  79 {
  80         errors_found |= ERROR_OBJECT;
  81         objreport(obj, "error", err);
  82         return -1;
  83 }
  84
  85 static int fsck_error_func(struct fsck_options *o,
  86         struct object *obj, int type, const char *message)
  87 {
  88         objreport(obj, (type == FSCK_WARN) ? "warning" : "error", message);
  89         return (type == FSCK_WARN) ? 0 : 1;
  90 }
  91
  92 static struct object_array pending;
  93
  94 static int mark_object(struct object *obj, int type, void *data, struct fsck_options *options)
  95 {
  96         struct object *parent = data;
  97
  98         /*
  99          * The only case data is NULL or type is OBJ_ANY is when
 100          * mark_object_reachable() calls us.  All the callers of
 101          * that function has non-NULL obj hence ...
 102          */
 103         if (!obj) {
 104                 /* ... these references to parent->fld are safe here */
 105                 printf("broken link from %7s %s\n",
 106                            typename(parent->type), describe_object(parent));
 107                 printf("broken link from %7s %s\n",
 108                            (type == OBJ_ANY ? "unknown" : typename(type)), "unknown");
 109                 errors_found |= ERROR_REACHABLE;
 110                 return 1;
 111         }
 112
 113         if (type != OBJ_ANY && obj->type != type)
 114                 /* ... and the reference to parent is safe here */
 115                 objerror(parent, "wrong object type in link");
 116
 117         if (obj->flags & REACHABLE)
 118                 return 0;
 119         obj->flags |= REACHABLE;
 120         if (!(obj->flags & HAS_OBJ)) {
 121                 if (parent && !has_object_file(&obj->oid)) {
 122                         printf("broken link from %7s %s\n",
 123                                  typename(parent->type), describe_object(parent));
 124                         printf("              to %7s %s\n",
 125                                  typename(obj->type), describe_object(obj));
 126                         errors_found |= ERROR_REACHABLE;
 127                 }
 128                 return 1;
 129         }
 130
 131         add_object_array(obj, NULL, &pending);
 132         return 0;
 133 }
 134
 135 static void mark_object_reachable(struct object *obj)
 136 {
 137         mark_object(obj, OBJ_ANY, NULL, NULL);
 138 }
 139
 140 static int traverse_one_object(struct object *obj)
 141 {
 142         int result;
 143         struct tree *tree = NULL;
 144
 145         if (obj->type == OBJ_TREE) {
 146                 tree = (struct tree *)obj;
 147                 if (parse_tree(tree) < 0)
 148                         return 1; /* error already displayed */
 149         }
 150         result = fsck_walk(obj, obj, &fsck_walk_options);
 151         if (tree)
 152                 free_tree_buffer(tree);
 153         return result;
 154 }
 155
 156 static int traverse_reachable(void)
 157 {
 158         struct progress *progress = NULL;
 159         unsigned int nr = 0;
 160         int result = 0;
 161         if (show_progress)
 162                 progress = start_progress_delay(_("Checking connectivity"), 0, 0, 2);
 163         while (pending.nr) {
 164                 struct object_array_entry *entry;
 165                 struct object *obj;
 166
 167                 entry = pending.objects + --pending.nr;
 168                 obj = entry->item;
 169                 result |= traverse_one_object(obj);
 170                 display_progress(progress, ++nr);
 171         }
 172         stop_progress(&progress);
 173         return !!result;
 174 }
 175
 176 static int mark_used(struct object *obj, int type, void *data, struct fsck_options *options)
 177 {
 178         if (!obj)
 179                 return 1;
 180         obj->used = 1;
 181         return 0;
 182 }
 183
 184 /*
 185  * Check a single reachable object
 186  */
 187 static void check_reachable_object(struct object *obj)
 188 {
 189         /*
 190          * We obviously want the object to be parsed,
 191          * except if it was in a pack-file and we didn't
 192          * do a full fsck
 193          */
 194         if (!(obj->flags & HAS_OBJ)) {
 195                 if (has_sha1_pack(obj->oid.hash))
 196                         return; /* it is in pack - forget about it */
 197                 if (connectivity_only && has_object_file(&obj->oid))
 198                         return;
 199                 printf("missing %s %s\n", typename(obj->type),
 200                         describe_object(obj));
 201                 errors_found |= ERROR_REACHABLE;
 202                 return;
 203         }
 204 }
 205
 206 /*
 207  * Check a single unreachable object
 208  */
 209 static void check_unreachable_object(struct object *obj)
 210 {
 211         /*
 212          * Missing unreachable object? Ignore it. It's not like
 213          * we miss it (since it can't be reached), nor do we want
 214          * to complain about it being unreachable (since it does
 215          * not exist).
 216          */
 217         if (!obj->parsed)
 218                 return;
 219
 220         /*
 221          * Unreachable object that exists? Show it if asked to,
 222          * since this is something that is prunable.
 223          */
 224         if (show_unreachable) {
 225                 printf("unreachable %s %s\n", typename(obj->type),
 226                         describe_object(obj));
 227                 return;
 228         }
 229
 230         /*
 231          * "!used" means that nothing at all points to it, including
 232          * other unreachable objects. In other words, it's the "tip"
 233          * of some set of unreachable objects, usually a commit that
 234          * got dropped.
 235          *
 236          * Such starting points are more interesting than some random
 237          * set of unreachable objects, so we show them even if the user
 238          * hasn't asked for _all_ unreachable objects. If you have
 239          * deleted a branch by mistake, this is a prime candidate to
 240          * start looking at, for example.
 241          */
 242         if (!obj->used) {
 243                 if (show_dangling)
 244                         printf("dangling %s %s\n", typename(obj->type),
 245                                describe_object(obj));
 246                 if (write_lost_and_found) {
 247                         char *filename = git_pathdup("lost-found/%s/%s",
 248                                 obj->type == OBJ_COMMIT ? "commit" : "other",
 249                                 describe_object(obj));
 250                         FILE *f;
 251
 252                         if (safe_create_leading_directories_const(filename)) {
 253                                 error("Could not create lost-found");
 254                                 free(filename);
 255                                 return;
 256                         }
 257                         if (!(f = fopen(filename, "w")))
 258                                 die_errno("Could not open '%s'", filename);
 259                         if (obj->type == OBJ_BLOB) {
 260                                 if (stream_blob_to_fd(fileno(f), obj->oid.hash, NULL, 1))
 261                                         die_errno("Could not write '%s'", filename);
 262                         } else
 263                                 fprintf(f, "%s\n", describe_object(obj));
 264                         if (fclose(f))
 265                                 die_errno("Could not finish '%s'",
 266                                           filename);
 267                         free(filename);
 268                 }
 269                 return;
 270         }
 271
 272         /*
 273          * Otherwise? It's there, it's unreachable, and some other unreachable
 274          * object points to it. Ignore it - it's not interesting, and we showed
 275          * all the interesting cases above.
 276          */
 277 }
 278
 279 static void check_object(struct object *obj)
 280 {
 281         if (verbose)
 282                 fprintf(stderr, "Checking %s\n", describe_object(obj));
 283
 284         if (obj->flags & REACHABLE)
 285                 check_reachable_object(obj);
 286         else
 287                 check_unreachable_object(obj);
 288 }
 289
 290 static void check_connectivity(void)
 291 {
 292         int i, max;
 293
 294         /* Traverse the pending reachable objects */
 295         traverse_reachable();
 296
 297         /* Look up all the requirements, warn about missing objects.. */
 298         max = get_max_object_index();
 299         if (verbose)
 300                 fprintf(stderr, "Checking connectivity (%d objects)\n", max);
 301
 302         for (i = 0; i < max; i++) {
 303                 struct object *obj = get_indexed_object(i);
 304
 305                 if (obj)
 306                         check_object(obj);
 307         }
 308 }
 309
 310 static int fsck_obj(struct object *obj)
 311 {
 312         if (obj->flags & SEEN)
 313                 return 0;
 314         obj->flags |= SEEN;
 315
 316         if (verbose)
 317                 fprintf(stderr, "Checking %s %s\n",
 318                         typename(obj->type), describe_object(obj));
 319
 320         if (fsck_walk(obj, NULL, &fsck_obj_options))
 321                 objerror(obj, "broken links");
 322         if (fsck_object(obj, NULL, 0, &fsck_obj_options))
 323                 return -1;
 324
 325         if (obj->type == OBJ_TREE) {
 326                 struct tree *item = (struct tree *) obj;
 327
 328                 free_tree_buffer(item);
 329         }
 330
 331         if (obj->type == OBJ_COMMIT) {
 332                 struct commit *commit = (struct commit *) obj;
 333
 334                 free_commit_buffer(commit);
 335
 336                 if (!commit->parents && show_root)
 337                         printf("root %s\n", describe_object(&commit->object));
 338         }
 339
 340         if (obj->type == OBJ_TAG) {
 341                 struct tag *tag = (struct tag *) obj;
 342
 343                 if (show_tags && tag->tagged) {
 344                         printf("tagged %s %s", typename(tag->tagged->type),
 345                                 describe_object(tag->tagged));
 346                         printf(" (%s) in %s\n", tag->tag,
 347                                 describe_object(&tag->object));
 348                 }
 349         }
 350
 351         return 0;
 352 }
 353
 354 static int fsck_sha1(const unsigned char *sha1)
 355 {
 356         struct object *obj = parse_object(sha1);
 357         if (!obj) {
 358                 errors_found |= ERROR_OBJECT;
 359                 return error("%s: object corrupt or missing",
 360                              sha1_to_hex(sha1));
 361         }
 362         obj->flags |= HAS_OBJ;
 363         return fsck_obj(obj);
 364 }
 365
 366 static int fsck_obj_buffer(const unsigned char *sha1, enum object_type type,
 367                            unsigned long size, void *buffer, int *eaten)
 368 {
 369         struct object *obj;
 370         obj = parse_object_buffer(sha1, type, size, buffer, eaten);
 371         if (!obj) {
 372                 errors_found |= ERROR_OBJECT;
 373                 return error("%s: object corrupt or missing", sha1_to_hex(sha1));
 374         }
 375         obj->flags = HAS_OBJ;
 376         return fsck_obj(obj);
 377 }
 378
 379 static int default_refs;
 380
 381 static void fsck_handle_reflog_sha1(const char *refname, unsigned char *sha1)
 382 {
 383         struct object *obj;
 384
 385         if (!is_null_sha1(sha1)) {
 386                 obj = lookup_object(sha1);
 387                 if (obj) {
 388                         obj->used = 1;
 389                         mark_object_reachable(obj);
 390                 } else {
 391                         error("%s: invalid reflog entry %s", refname, sha1_to_hex(sha1));
 392                         errors_found |= ERROR_REACHABLE;
 393                 }
 394         }
 395 }
 396
 397 static int fsck_handle_reflog_ent(unsigned char *osha1, unsigned char *nsha1,
 398                 const char *email, unsigned long timestamp, int tz,
 399                 const char *message, void *cb_data)
 400 {
 401         const char *refname = cb_data;
 402
 403         if (verbose)
 404                 fprintf(stderr, "Checking reflog %s->%s\n",
 405                         sha1_to_hex(osha1), sha1_to_hex(nsha1));
 406
 407         fsck_handle_reflog_sha1(refname, osha1);
 408         fsck_handle_reflog_sha1(refname, nsha1);
 409         return 0;
 410 }
 411
 412 static int fsck_handle_reflog(const char *logname, const struct object_id *oid,
 413                               int flag, void *cb_data)
 414 {
 415         for_each_reflog_ent(logname, fsck_handle_reflog_ent, (void *)logname);
 416         return 0;
 417 }
 418
 419 static int fsck_handle_ref(const char *refname, const struct object_id *oid,
 420                            int flag, void *cb_data)
 421 {
 422         struct object *obj;
 423
 424         obj = parse_object(oid->hash);
 425         if (!obj) {
 426                 error("%s: invalid sha1 pointer %s", refname, oid_to_hex(oid));
 427                 errors_found |= ERROR_REACHABLE;
 428                 /* We'll continue with the rest despite the error.. */
 429                 return 0;
 430         }
 431         if (obj->type != OBJ_COMMIT && is_branch(refname)) {
 432                 error("%s: not a commit", refname);
 433                 errors_found |= ERROR_REFS;
 434         }
 435         default_refs++;
 436         obj->used = 1;
 437         mark_object_reachable(obj);
 438
 439         return 0;
 440 }
 441
 442 static void get_default_heads(void)
 443 {
 444         if (head_points_at && !is_null_oid(&head_oid))
 445                 fsck_handle_ref("HEAD", &head_oid, 0, NULL);
 446         for_each_rawref(fsck_handle_ref, NULL);
 447         if (include_reflogs)
 448                 for_each_reflog(fsck_handle_reflog, NULL);
 449
 450         /*
 451          * Not having any default heads isn't really fatal, but
 452          * it does mean that "--unreachable" no longer makes any
 453          * sense (since in this case everything will obviously
 454          * be unreachable by definition.
 455          *
 456          * Showing dangling objects is valid, though (as those
 457          * dangling objects are likely lost heads).
 458          *
 459          * So we just print a warning about it, and clear the
 460          * "show_unreachable" flag.
 461          */
 462         if (!default_refs) {
 463                 fprintf(stderr, "notice: No default references\n");
 464                 show_unreachable = 0;
 465         }
 466 }
 467
 468 static int fsck_loose(const unsigned char *sha1, const char *path, void *data)
 469 {
 470         if (fsck_sha1(sha1))
 471                 errors_found |= ERROR_OBJECT;
 472         return 0;
 473 }
 474
 475 static int fsck_cruft(const char *basename, const char *path, void *data)
 476 {
 477         if (!starts_with(basename, "tmp_obj_"))
 478                 fprintf(stderr, "bad sha1 file: %s\n", path);
 479         return 0;
 480 }
 481
 482 static int fsck_subdir(int nr, const char *path, void *progress)
 483 {
 484         display_progress(progress, nr + 1);
 485         return 0;
 486 }
 487
 488 static void fsck_object_dir(const char *path)
 489 {
 490         struct progress *progress = NULL;
 491
 492         if (verbose)
 493                 fprintf(stderr, "Checking object directory\n");
 494
 495         if (show_progress)
 496                 progress = start_progress(_("Checking object directories"), 256);
 497
 498         for_each_loose_file_in_objdir(path, fsck_loose, fsck_cruft, fsck_subdir,
 499                                       progress);
 500         display_progress(progress, 256);
 501         stop_progress(&progress);
 502 }
 503
 504 static int fsck_head_link(void)
 505 {
 506         int null_is_error = 0;
 507
 508         if (verbose)
 509                 fprintf(stderr, "Checking HEAD link\n");
 510
 511         head_points_at = resolve_ref_unsafe("HEAD", 0, head_oid.hash, NULL);
 512         if (!head_points_at) {
 513                 errors_found |= ERROR_REFS;
 514                 return error("Invalid HEAD");
 515         }
 516         if (!strcmp(head_points_at, "HEAD"))
 517                 /* detached HEAD */
 518                 null_is_error = 1;
 519         else if (!starts_with(head_points_at, "refs/heads/")) {
 520                 errors_found |= ERROR_REFS;
 521                 return error("HEAD points to something strange (%s)",
 522                              head_points_at);
 523         }
 524         if (is_null_oid(&head_oid)) {
 525                 if (null_is_error) {
 526                         errors_found |= ERROR_REFS;
 527                         return error("HEAD: detached HEAD points at nothing");
 528                 }
 529                 fprintf(stderr, "notice: HEAD points to an unborn branch (%s)\n",
 530                         head_points_at + 11);
 531         }
 532         return 0;
 533 }
 534
 535 static int fsck_cache_tree(struct cache_tree *it)
 536 {
 537         int i;
 538         int err = 0;
 539
 540         if (verbose)
 541                 fprintf(stderr, "Checking cache tree\n");
 542
 543         if (0 <= it->entry_count) {
 544                 struct object *obj = parse_object(it->sha1);
 545                 if (!obj) {
 546                         error("%s: invalid sha1 pointer in cache-tree",
 547                               sha1_to_hex(it->sha1));
 548                         errors_found |= ERROR_REFS;
 549                         return 1;
 550                 }
 551                 obj->used = 1;
 552                 mark_object_reachable(obj);
 553                 if (obj->type != OBJ_TREE)
 554                         err |= objerror(obj, "non-tree in cache-tree");
 555         }
 556         for (i = 0; i < it->subtree_nr; i++)
 557                 err |= fsck_cache_tree(it->down[i]->cache_tree);
 558         return err;
 559 }
 560
 561 static char const * const fsck_usage[] = {
 562         N_("git fsck [<options>] [<object>...]"),
 563         NULL
 564 };
 565
 566 static struct option fsck_opts[] = {
 567         OPT__VERBOSE(&verbose, N_("be verbose")),
 568         OPT_BOOL(0, "unreachable", &show_unreachable, N_("show unreachable objects")),
 569         OPT_BOOL(0, "dangling", &show_dangling, N_("show dangling objects")),
 570         OPT_BOOL(0, "tags", &show_tags, N_("report tags")),
 571         OPT_BOOL(0, "root", &show_root, N_("report root nodes")),
 572         OPT_BOOL(0, "cache", &keep_cache_objects, N_("make index objects head nodes")),
 573         OPT_BOOL(0, "reflogs", &include_reflogs, N_("make reflogs head nodes (default)")),
 574         OPT_BOOL(0, "full", &check_full, N_("also consider packs and alternate objects")),
 575         OPT_BOOL(0, "connectivity-only", &connectivity_only, N_("check only connectivity")),
 576         OPT_BOOL(0, "strict", &check_strict, N_("enable more strict checking")),
 577         OPT_BOOL(0, "lost-found", &write_lost_and_found,
 578                                 N_("write dangling objects in .git/lost-found")),
 579         OPT_BOOL(0, "progress", &show_progress, N_("show progress")),
 580         OPT_END(),
 581 };
 582
 583 int cmd_fsck(int argc, const char **argv, const char *prefix)
 584 {
 585         int i, heads;
 586         struct alternate_object_database *alt;
 587
 588         errors_found = 0;
 589         check_replace_refs = 0;
 590
 591         argc = parse_options(argc, argv, prefix, fsck_opts, fsck_usage, 0);
 592
 593         fsck_walk_options.walk = mark_object;
 594         fsck_obj_options.walk = mark_used;
 595         fsck_obj_options.error_func = fsck_error_func;
 596         if (check_strict)
 597                 fsck_obj_options.strict = 1;
 598
 599         if (show_progress == -1)
 600                 show_progress = isatty(2);
 601         if (verbose)
 602                 show_progress = 0;
 603
 604         if (write_lost_and_found) {
 605                 check_full = 1;
 606                 include_reflogs = 0;
 607         }
 608
 609         git_config(fsck_config, NULL);
 610
 611         fsck_head_link();
 612         if (!connectivity_only) {
 613                 fsck_object_dir(get_object_directory());
 614
 615                 prepare_alt_odb();
 616                 for (alt = alt_odb_list; alt; alt = alt->next) {
 617                         /* directory name, minus trailing slash */
 618                         size_t namelen = alt->name - alt->base - 1;
 619                         struct strbuf name = STRBUF_INIT;
 620                         strbuf_add(&name, alt->base, namelen);
 621                         fsck_object_dir(name.buf);
 622                         strbuf_release(&name);
 623                 }
 624         }
 625
 626         if (check_full) {
 627                 struct packed_git *p;
 628                 uint32_t total = 0, count = 0;
 629                 struct progress *progress = NULL;
 630
 631                 prepare_packed_git();
 632
 633                 if (show_progress) {
 634                         for (p = packed_git; p; p = p->next) {
 635                                 if (open_pack_index(p))
 636                                         continue;
 637                                 total += p->num_objects;
 638                         }
 639
 640                         progress = start_progress(_("Checking objects"), total);
 641                 }
 642                 for (p = packed_git; p; p = p->next) {
 643                         /* verify gives error messages itself */
 644                         if (verify_pack(p, fsck_obj_buffer,
 645                                         progress, count))
 646                                 errors_found |= ERROR_PACK;
 647                         count += p->num_objects;
 648                 }
 649                 stop_progress(&progress);
 650         }
 651
 652         heads = 0;
 653         for (i = 0; i < argc; i++) {
 654                 const char *arg = argv[i];
 655                 unsigned char sha1[20];
 656                 if (!get_sha1(arg, sha1)) {
 657                         struct object *obj = lookup_object(sha1);
 658
 659                         /* Error is printed by lookup_object(). */
 660                         if (!obj)
 661                                 continue;
 662
 663                         obj->used = 1;
 664                         mark_object_reachable(obj);
 665                         heads++;
 666                         continue;
 667                 }
 668                 error("invalid parameter: expected sha1, got '%s'", arg);
 669         }
 670
 671         /*
 672          * If we've not been given any explicit head information, do the
 673          * default ones from .git/refs. We also consider the index file
 674          * in this case (ie this implies --cache).
 675          */
 676         if (!heads) {
 677                 get_default_heads();
 678                 keep_cache_objects = 1;
 679         }
 680
 681         if (keep_cache_objects) {
 682                 read_cache();
 683                 for (i = 0; i < active_nr; i++) {
 684                         unsigned int mode;
 685                         struct blob *blob;
 686                         struct object *obj;
 687
 688                         mode = active_cache[i]->ce_mode;
 689                         if (S_ISGITLINK(mode))
 690                                 continue;
 691                         blob = lookup_blob(active_cache[i]->sha1);
 692                         if (!blob)
 693                                 continue;
 694                         obj = &blob->object;
 695                         obj->used = 1;
 696                         mark_object_reachable(obj);
 697                 }
 698                 if (active_cache_tree)
 699                         fsck_cache_tree(active_cache_tree);
 700         }
 701
 702         check_connectivity();
 703         return errors_found;
 704 }