fsck.c

   1 #include "cache.h"
   2 #include "object-store.h"
   3 #include "repository.h"
   4 #include "object.h"
   5 #include "attr.h"
   6 #include "blob.h"
   7 #include "tree.h"
   8 #include "tree-walk.h"
   9 #include "commit.h"
  10 #include "tag.h"
  11 #include "fsck.h"
  12 #include "refs.h"
  13 #include "url.h"
  14 #include "utf8.h"
  15 #include "decorate.h"
  16 #include "oidset.h"
  17 #include "packfile.h"
  18 #include "submodule-config.h"
  19 #include "config.h"
  20 #include "credential.h"
  21 #include "help.h"
  22
  23 #define STR(x) #x
  24 #define MSG_ID(id, msg_type) { STR(id), NULL, NULL, FSCK_##msg_type },
  25 static struct {
  26         const char *id_string;
  27         const char *downcased;
  28         const char *camelcased;
  29         enum fsck_msg_type msg_type;
  30 } msg_id_info[FSCK_MSG_MAX + 1] = {
  31         FOREACH_FSCK_MSG_ID(MSG_ID)
  32         { NULL, NULL, NULL, -1 }
  33 };
  34 #undef MSG_ID
  35 #undef STR
  36
  37 static void prepare_msg_ids(void)
  38 {
  39         int i;
  40
  41         if (msg_id_info[0].downcased)
  42                 return;
  43
  44         /* convert id_string to lower case, without underscores. */
  45         for (i = 0; i < FSCK_MSG_MAX; i++) {
  46                 const char *p = msg_id_info[i].id_string;
  47                 int len = strlen(p);
  48                 char *q = xmalloc(len);
  49
  50                 msg_id_info[i].downcased = q;
  51                 while (*p)
  52                         if (*p == '_')
  53                                 p++;
  54                         else
  55                                 *(q)++ = tolower(*(p)++);
  56                 *q = '\0';
  57
  58                 p = msg_id_info[i].id_string;
  59                 q = xmalloc(len);
  60                 msg_id_info[i].camelcased = q;
  61                 while (*p) {
  62                         if (*p == '_') {
  63                                 p++;
  64                                 if (*p)
  65                                         *q++ = *p++;
  66                         } else {
  67                                 *q++ = tolower(*p++);
  68                         }
  69                 }
  70                 *q = '\0';
  71         }
  72 }
  73
  74 static int parse_msg_id(const char *text)
  75 {
  76         int i;
  77
  78         prepare_msg_ids();
  79
  80         for (i = 0; i < FSCK_MSG_MAX; i++)
  81                 if (!strcmp(text, msg_id_info[i].downcased))
  82                         return i;
  83
  84         return -1;
  85 }
  86
  87 void list_config_fsck_msg_ids(struct string_list *list, const char *prefix)
  88 {
  89         int i;
  90
  91         prepare_msg_ids();
  92
  93         for (i = 0; i < FSCK_MSG_MAX; i++)
  94                 list_config_item(list, prefix, msg_id_info[i].camelcased);
  95 }
  96
  97 static enum fsck_msg_type fsck_msg_type(enum fsck_msg_id msg_id,
  98         struct fsck_options *options)
  99 {
 100         assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
 101
 102         if (!options->msg_type) {
 103                 enum fsck_msg_type msg_type = msg_id_info[msg_id].msg_type;
 104
 105                 if (options->strict && msg_type == FSCK_WARN)
 106                         msg_type = FSCK_ERROR;
 107                 return msg_type;
 108         }
 109
 110         return options->msg_type[msg_id];
 111 }
 112
 113 static enum fsck_msg_type parse_msg_type(const char *str)
 114 {
 115         if (!strcmp(str, "error"))
 116                 return FSCK_ERROR;
 117         else if (!strcmp(str, "warn"))
 118                 return FSCK_WARN;
 119         else if (!strcmp(str, "ignore"))
 120                 return FSCK_IGNORE;
 121         else
 122                 die("Unknown fsck message type: '%s'", str);
 123 }
 124
 125 int is_valid_msg_type(const char *msg_id, const char *msg_type)
 126 {
 127         if (parse_msg_id(msg_id) < 0)
 128                 return 0;
 129         parse_msg_type(msg_type);
 130         return 1;
 131 }
 132
 133 void fsck_set_msg_type_from_ids(struct fsck_options *options,
 134                                 enum fsck_msg_id msg_id,
 135                                 enum fsck_msg_type msg_type)
 136 {
 137         if (!options->msg_type) {
 138                 int i;
 139                 enum fsck_msg_type *severity;
 140                 ALLOC_ARRAY(severity, FSCK_MSG_MAX);
 141                 for (i = 0; i < FSCK_MSG_MAX; i++)
 142                         severity[i] = fsck_msg_type(i, options);
 143                 options->msg_type = severity;
 144         }
 145
 146         options->msg_type[msg_id] = msg_type;
 147 }
 148
 149 void fsck_set_msg_type(struct fsck_options *options,
 150                        const char *msg_id_str, const char *msg_type_str)
 151 {
 152         int msg_id = parse_msg_id(msg_id_str);
 153         enum fsck_msg_type msg_type = parse_msg_type(msg_type_str);
 154
 155         if (msg_id < 0)
 156                 die("Unhandled message id: %s", msg_id_str);
 157
 158         if (msg_type != FSCK_ERROR && msg_id_info[msg_id].msg_type == FSCK_FATAL)
 159                 die("Cannot demote %s to %s", msg_id_str, msg_type_str);
 160
 161         fsck_set_msg_type_from_ids(options, msg_id, msg_type);
 162 }
 163
 164 void fsck_set_msg_types(struct fsck_options *options, const char *values)
 165 {
 166         char *buf = xstrdup(values), *to_free = buf;
 167         int done = 0;
 168
 169         while (!done) {
 170                 int len = strcspn(buf, " ,|"), equal;
 171
 172                 done = !buf[len];
 173                 if (!len) {
 174                         buf++;
 175                         continue;
 176                 }
 177                 buf[len] = '\0';
 178
 179                 for (equal = 0;
 180                      equal < len && buf[equal] != '=' && buf[equal] != ':';
 181                      equal++)
 182                         buf[equal] = tolower(buf[equal]);
 183                 buf[equal] = '\0';
 184
 185                 if (!strcmp(buf, "skiplist")) {
 186                         if (equal == len)
 187                                 die("skiplist requires a path");
 188                         oidset_parse_file(&options->skiplist, buf + equal + 1);
 189                         buf += len + 1;
 190                         continue;
 191                 }
 192
 193                 if (equal == len)
 194                         die("Missing '=': '%s'", buf);
 195
 196                 fsck_set_msg_type(options, buf, buf + equal + 1);
 197                 buf += len + 1;
 198         }
 199         free(to_free);
 200 }
 201
 202 static int object_on_skiplist(struct fsck_options *opts,
 203                               const struct object_id *oid)
 204 {
 205         return opts && oid && oidset_contains(&opts->skiplist, oid);
 206 }
 207
 208 __attribute__((format (printf, 5, 6)))
 209 static int report(struct fsck_options *options,
 210                   const struct object_id *oid, enum object_type object_type,
 211                   enum fsck_msg_id msg_id, const char *fmt, ...)
 212 {
 213         va_list ap;
 214         struct strbuf sb = STRBUF_INIT;
 215         enum fsck_msg_type msg_type = fsck_msg_type(msg_id, options);
 216         int result;
 217
 218         if (msg_type == FSCK_IGNORE)
 219                 return 0;
 220
 221         if (object_on_skiplist(options, oid))
 222                 return 0;
 223
 224         if (msg_type == FSCK_FATAL)
 225                 msg_type = FSCK_ERROR;
 226         else if (msg_type == FSCK_INFO)
 227                 msg_type = FSCK_WARN;
 228
 229         prepare_msg_ids();
 230         strbuf_addf(&sb, "%s: ", msg_id_info[msg_id].camelcased);
 231
 232         va_start(ap, fmt);
 233         strbuf_vaddf(&sb, fmt, ap);
 234         result = options->error_func(options, oid, object_type,
 235                                      msg_type, msg_id, sb.buf);
 236         strbuf_release(&sb);
 237         va_end(ap);
 238
 239         return result;
 240 }
 241
 242 void fsck_enable_object_names(struct fsck_options *options)
 243 {
 244         if (!options->object_names)
 245                 options->object_names = kh_init_oid_map();
 246 }
 247
 248 const char *fsck_get_object_name(struct fsck_options *options,
 249                                  const struct object_id *oid)
 250 {
 251         khiter_t pos;
 252         if (!options->object_names)
 253                 return NULL;
 254         pos = kh_get_oid_map(options->object_names, *oid);
 255         if (pos >= kh_end(options->object_names))
 256                 return NULL;
 257         return kh_value(options->object_names, pos);
 258 }
 259
 260 void fsck_put_object_name(struct fsck_options *options,
 261                           const struct object_id *oid,
 262                           const char *fmt, ...)
 263 {
 264         va_list ap;
 265         struct strbuf buf = STRBUF_INIT;
 266         khiter_t pos;
 267         int hashret;
 268
 269         if (!options->object_names)
 270                 return;
 271
 272         pos = kh_put_oid_map(options->object_names, *oid, &hashret);
 273         if (!hashret)
 274                 return;
 275         va_start(ap, fmt);
 276         strbuf_vaddf(&buf, fmt, ap);
 277         kh_value(options->object_names, pos) = strbuf_detach(&buf, NULL);
 278         va_end(ap);
 279 }
 280
 281 const char *fsck_describe_object(struct fsck_options *options,
 282                                  const struct object_id *oid)
 283 {
 284         static struct strbuf bufs[] = {
 285                 STRBUF_INIT, STRBUF_INIT, STRBUF_INIT, STRBUF_INIT
 286         };
 287         static int b = 0;
 288         struct strbuf *buf;
 289         const char *name = fsck_get_object_name(options, oid);
 290
 291         buf = bufs + b;
 292         b = (b + 1) % ARRAY_SIZE(bufs);
 293         strbuf_reset(buf);
 294         strbuf_addstr(buf, oid_to_hex(oid));
 295         if (name)
 296                 strbuf_addf(buf, " (%s)", name);
 297
 298         return buf->buf;
 299 }
 300
 301 static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 302 {
 303         struct tree_desc desc;
 304         struct name_entry entry;
 305         int res = 0;
 306         const char *name;
 307
 308         if (parse_tree(tree))
 309                 return -1;
 310
 311         name = fsck_get_object_name(options, &tree->object.oid);
 312         if (init_tree_desc_gently(&desc, tree->buffer, tree->size, 0))
 313                 return -1;
 314         while (tree_entry_gently(&desc, &entry)) {
 315                 struct object *obj;
 316                 int result;
 317
 318                 if (S_ISGITLINK(entry.mode))
 319                         continue;
 320
 321                 if (S_ISDIR(entry.mode)) {
 322                         obj = (struct object *)lookup_tree(the_repository, &entry.oid);
 323                         if (name && obj)
 324                                 fsck_put_object_name(options, &entry.oid, "%s%s/",
 325                                                      name, entry.path);
 326                         result = options->walk(obj, OBJ_TREE, data, options);
 327                 }
 328                 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 329                         obj = (struct object *)lookup_blob(the_repository, &entry.oid);
 330                         if (name && obj)
 331                                 fsck_put_object_name(options, &entry.oid, "%s%s",
 332                                                      name, entry.path);
 333                         result = options->walk(obj, OBJ_BLOB, data, options);
 334                 }
 335                 else {
 336                         result = error("in tree %s: entry %s has bad mode %.6o",
 337                                        fsck_describe_object(options, &tree->object.oid),
 338                                        entry.path, entry.mode);
 339                 }
 340                 if (result < 0)
 341                         return result;
 342                 if (!res)
 343                         res = result;
 344         }
 345         return res;
 346 }
 347
 348 static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 349 {
 350         int counter = 0, generation = 0, name_prefix_len = 0;
 351         struct commit_list *parents;
 352         int res;
 353         int result;
 354         const char *name;
 355
 356         if (parse_commit(commit))
 357                 return -1;
 358
 359         name = fsck_get_object_name(options, &commit->object.oid);
 360         if (name)
 361                 fsck_put_object_name(options, get_commit_tree_oid(commit),
 362                                      "%s:", name);
 363
 364         result = options->walk((struct object *)get_commit_tree(commit),
 365                                OBJ_TREE, data, options);
 366         if (result < 0)
 367                 return result;
 368         res = result;
 369
 370         parents = commit->parents;
 371         if (name && parents) {
 372                 int len = strlen(name), power;
 373
 374                 if (len && name[len - 1] == '^') {
 375                         generation = 1;
 376                         name_prefix_len = len - 1;
 377                 }
 378                 else { /* parse ~<generation> suffix */
 379                         for (generation = 0, power = 1;
 380                              len && isdigit(name[len - 1]);
 381                              power *= 10)
 382                                 generation += power * (name[--len] - '0');
 383                         if (power > 1 && len && name[len - 1] == '~')
 384                                 name_prefix_len = len - 1;
 385                         else {
 386                                 /* Maybe a non-first parent, e.g. HEAD^2 */
 387                                 generation = 0;
 388                                 name_prefix_len = len;
 389                         }
 390                 }
 391         }
 392
 393         while (parents) {
 394                 if (name) {
 395                         struct object_id *oid = &parents->item->object.oid;
 396
 397                         if (counter++)
 398                                 fsck_put_object_name(options, oid, "%s^%d",
 399                                                      name, counter);
 400                         else if (generation > 0)
 401                                 fsck_put_object_name(options, oid, "%.*s~%d",
 402                                                      name_prefix_len, name,
 403                                                      generation + 1);
 404                         else
 405                                 fsck_put_object_name(options, oid, "%s^", name);
 406                 }
 407                 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 408                 if (result < 0)
 409                         return result;
 410                 if (!res)
 411                         res = result;
 412                 parents = parents->next;
 413         }
 414         return res;
 415 }
 416
 417 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 418 {
 419         const char *name = fsck_get_object_name(options, &tag->object.oid);
 420
 421         if (parse_tag(tag))
 422                 return -1;
 423         if (name)
 424                 fsck_put_object_name(options, &tag->tagged->oid, "%s", name);
 425         return options->walk(tag->tagged, OBJ_ANY, data, options);
 426 }
 427
 428 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 429 {
 430         if (!obj)
 431                 return -1;
 432
 433         if (obj->type == OBJ_NONE)
 434                 parse_object(the_repository, &obj->oid);
 435
 436         switch (obj->type) {
 437         case OBJ_BLOB:
 438                 return 0;
 439         case OBJ_TREE:
 440                 return fsck_walk_tree((struct tree *)obj, data, options);
 441         case OBJ_COMMIT:
 442                 return fsck_walk_commit((struct commit *)obj, data, options);
 443         case OBJ_TAG:
 444                 return fsck_walk_tag((struct tag *)obj, data, options);
 445         default:
 446                 error("Unknown object type for %s",
 447                       fsck_describe_object(options, &obj->oid));
 448                 return -1;
 449         }
 450 }
 451
 452 struct name_stack {
 453         const char **names;
 454         size_t nr, alloc;
 455 };
 456
 457 static void name_stack_push(struct name_stack *stack, const char *name)
 458 {
 459         ALLOC_GROW(stack->names, stack->nr + 1, stack->alloc);
 460         stack->names[stack->nr++] = name;
 461 }
 462
 463 static const char *name_stack_pop(struct name_stack *stack)
 464 {
 465         return stack->nr ? stack->names[--stack->nr] : NULL;
 466 }
 467
 468 static void name_stack_clear(struct name_stack *stack)
 469 {
 470         FREE_AND_NULL(stack->names);
 471         stack->nr = stack->alloc = 0;
 472 }
 473
 474 /*
 475  * The entries in a tree are ordered in the _path_ order,
 476  * which means that a directory entry is ordered by adding
 477  * a slash to the end of it.
 478  *
 479  * So a directory called "a" is ordered _after_ a file
 480  * called "a.c", because "a/" sorts after "a.c".
 481  */
 482 #define TREE_UNORDERED (-1)
 483 #define TREE_HAS_DUPS  (-2)
 484
 485 static int is_less_than_slash(unsigned char c)
 486 {
 487         return '\0' < c && c < '/';
 488 }
 489
 490 static int verify_ordered(unsigned mode1, const char *name1,
 491                           unsigned mode2, const char *name2,
 492                           struct name_stack *candidates)
 493 {
 494         int len1 = strlen(name1);
 495         int len2 = strlen(name2);
 496         int len = len1 < len2 ? len1 : len2;
 497         unsigned char c1, c2;
 498         int cmp;
 499
 500         cmp = memcmp(name1, name2, len);
 501         if (cmp < 0)
 502                 return 0;
 503         if (cmp > 0)
 504                 return TREE_UNORDERED;
 505
 506         /*
 507          * Ok, the first <len> characters are the same.
 508          * Now we need to order the next one, but turn
 509          * a '\0' into a '/' for a directory entry.
 510          */
 511         c1 = name1[len];
 512         c2 = name2[len];
 513         if (!c1 && !c2)
 514                 /*
 515                  * git-write-tree used to write out a nonsense tree that has
 516                  * entries with the same name, one blob and one tree.  Make
 517                  * sure we do not have duplicate entries.
 518                  */
 519                 return TREE_HAS_DUPS;
 520         if (!c1 && S_ISDIR(mode1))
 521                 c1 = '/';
 522         if (!c2 && S_ISDIR(mode2))
 523                 c2 = '/';
 524
 525         /*
 526          * There can be non-consecutive duplicates due to the implicitly
 527          * added slash, e.g.:
 528          *
 529          *   foo
 530          *   foo.bar
 531          *   foo.bar.baz
 532          *   foo.bar/
 533          *   foo/
 534          *
 535          * Record non-directory candidates (like "foo" and "foo.bar" in
 536          * the example) on a stack and check directory candidates (like
 537          * foo/" and "foo.bar/") against that stack.
 538          */
 539         if (!c1 && is_less_than_slash(c2)) {
 540                 name_stack_push(candidates, name1);
 541         } else if (c2 == '/' && is_less_than_slash(c1)) {
 542                 for (;;) {
 543                         const char *p;
 544                         const char *f_name = name_stack_pop(candidates);
 545
 546                         if (!f_name)
 547                                 break;
 548                         if (!skip_prefix(name2, f_name, &p))
 549                                 continue;
 550                         if (!*p)
 551                                 return TREE_HAS_DUPS;
 552                         if (is_less_than_slash(*p)) {
 553                                 name_stack_push(candidates, f_name);
 554                                 break;
 555                         }
 556                 }
 557         }
 558
 559         return c1 < c2 ? 0 : TREE_UNORDERED;
 560 }
 561
 562 static int fsck_tree(const struct object_id *tree_oid,
 563                      const char *buffer, unsigned long size,
 564                      struct fsck_options *options)
 565 {
 566         int retval = 0;
 567         int has_null_sha1 = 0;
 568         int has_full_path = 0;
 569         int has_empty_name = 0;
 570         int has_dot = 0;
 571         int has_dotdot = 0;
 572         int has_dotgit = 0;
 573         int has_zero_pad = 0;
 574         int has_bad_modes = 0;
 575         int has_dup_entries = 0;
 576         int not_properly_sorted = 0;
 577         struct tree_desc desc;
 578         unsigned o_mode;
 579         const char *o_name;
 580         struct name_stack df_dup_candidates = { NULL };
 581
 582         if (init_tree_desc_gently(&desc, buffer, size, TREE_DESC_RAW_MODES)) {
 583                 retval += report(options, tree_oid, OBJ_TREE,
 584                                  FSCK_MSG_BAD_TREE,
 585                                  "cannot be parsed as a tree");
 586                 return retval;
 587         }
 588
 589         o_mode = 0;
 590         o_name = NULL;
 591
 592         while (desc.size) {
 593                 unsigned short mode;
 594                 const char *name, *backslash;
 595                 const struct object_id *entry_oid;
 596
 597                 entry_oid = tree_entry_extract(&desc, &name, &mode);
 598
 599                 has_null_sha1 |= is_null_oid(entry_oid);
 600                 has_full_path |= !!strchr(name, '/');
 601                 has_empty_name |= !*name;
 602                 has_dot |= !strcmp(name, ".");
 603                 has_dotdot |= !strcmp(name, "..");
 604                 has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
 605                 has_zero_pad |= *(char *)desc.buffer == '0';
 606
 607                 if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
 608                         if (!S_ISLNK(mode))
 609                                 oidset_insert(&options->gitmodules_found,
 610                                               entry_oid);
 611                         else
 612                                 retval += report(options,
 613                                                  tree_oid, OBJ_TREE,
 614                                                  FSCK_MSG_GITMODULES_SYMLINK,
 615                                                  ".gitmodules is a symbolic link");
 616                 }
 617
 618                 if (is_hfs_dotgitattributes(name) || is_ntfs_dotgitattributes(name)) {
 619                         if (!S_ISLNK(mode))
 620                                 oidset_insert(&options->gitattributes_found,
 621                                               entry_oid);
 622                         else
 623                                 retval += report(options, tree_oid, OBJ_TREE,
 624                                                  FSCK_MSG_GITATTRIBUTES_SYMLINK,
 625                                                  ".gitattributes is a symlink");
 626                 }
 627
 628                 if (S_ISLNK(mode)) {
 629                         if (is_hfs_dotgitignore(name) ||
 630                             is_ntfs_dotgitignore(name))
 631                                 retval += report(options, tree_oid, OBJ_TREE,
 632                                                  FSCK_MSG_GITIGNORE_SYMLINK,
 633                                                  ".gitignore is a symlink");
 634                         if (is_hfs_dotmailmap(name) ||
 635                             is_ntfs_dotmailmap(name))
 636                                 retval += report(options, tree_oid, OBJ_TREE,
 637                                                  FSCK_MSG_MAILMAP_SYMLINK,
 638                                                  ".mailmap is a symlink");
 639                         oidset_insert(&options->symlink_targets_found,
 640                                       entry_oid);
 641                 }
 642
 643                 if ((backslash = strchr(name, '\\'))) {
 644                         while (backslash) {
 645                                 backslash++;
 646                                 has_dotgit |= is_ntfs_dotgit(backslash);
 647                                 if (is_ntfs_dotgitmodules(backslash)) {
 648                                         if (!S_ISLNK(mode))
 649                                                 oidset_insert(&options->gitmodules_found,
 650                                                               entry_oid);
 651                                         else
 652                                                 retval += report(options, tree_oid, OBJ_TREE,
 653                                                                  FSCK_MSG_GITMODULES_SYMLINK,
 654                                                                  ".gitmodules is a symbolic link");
 655                                 }
 656                                 backslash = strchr(backslash, '\\');
 657                         }
 658                 }
 659
 660                 if (update_tree_entry_gently(&desc)) {
 661                         retval += report(options, tree_oid, OBJ_TREE,
 662                                          FSCK_MSG_BAD_TREE,
 663                                          "cannot be parsed as a tree");
 664                         break;
 665                 }
 666
 667                 switch (mode) {
 668                 /*
 669                  * Standard modes..
 670                  */
 671                 case S_IFREG | 0755:
 672                 case S_IFREG | 0644:
 673                 case S_IFLNK:
 674                 case S_IFDIR:
 675                 case S_IFGITLINK:
 676                         break;
 677                 /*
 678                  * This is nonstandard, but we had a few of these
 679                  * early on when we honored the full set of mode
 680                  * bits..
 681                  */
 682                 case S_IFREG | 0664:
 683                         if (!options->strict)
 684                                 break;
 685                         /* fallthrough */
 686                 default:
 687                         has_bad_modes = 1;
 688                 }
 689
 690                 if (o_name) {
 691                         switch (verify_ordered(o_mode, o_name, mode, name,
 692                                                &df_dup_candidates)) {
 693                         case TREE_UNORDERED:
 694                                 not_properly_sorted = 1;
 695                                 break;
 696                         case TREE_HAS_DUPS:
 697                                 has_dup_entries = 1;
 698                                 break;
 699                         default:
 700                                 break;
 701                         }
 702                 }
 703
 704                 o_mode = mode;
 705                 o_name = name;
 706         }
 707
 708         name_stack_clear(&df_dup_candidates);
 709
 710         if (has_null_sha1)
 711                 retval += report(options, tree_oid, OBJ_TREE,
 712                                  FSCK_MSG_NULL_SHA1,
 713                                  "contains entries pointing to null sha1");
 714         if (has_full_path)
 715                 retval += report(options, tree_oid, OBJ_TREE,
 716                                  FSCK_MSG_FULL_PATHNAME,
 717                                  "contains full pathnames");
 718         if (has_empty_name)
 719                 retval += report(options, tree_oid, OBJ_TREE,
 720                                  FSCK_MSG_EMPTY_NAME,
 721                                  "contains empty pathname");
 722         if (has_dot)
 723                 retval += report(options, tree_oid, OBJ_TREE,
 724                                  FSCK_MSG_HAS_DOT,
 725                                  "contains '.'");
 726         if (has_dotdot)
 727                 retval += report(options, tree_oid, OBJ_TREE,
 728                                  FSCK_MSG_HAS_DOTDOT,
 729                                  "contains '..'");
 730         if (has_dotgit)
 731                 retval += report(options, tree_oid, OBJ_TREE,
 732                                  FSCK_MSG_HAS_DOTGIT,
 733                                  "contains '.git'");
 734         if (has_zero_pad)
 735                 retval += report(options, tree_oid, OBJ_TREE,
 736                                  FSCK_MSG_ZERO_PADDED_FILEMODE,
 737                                  "contains zero-padded file modes");
 738         if (has_bad_modes)
 739                 retval += report(options, tree_oid, OBJ_TREE,
 740                                  FSCK_MSG_BAD_FILEMODE,
 741                                  "contains bad file modes");
 742         if (has_dup_entries)
 743                 retval += report(options, tree_oid, OBJ_TREE,
 744                                  FSCK_MSG_DUPLICATE_ENTRIES,
 745                                  "contains duplicate file entries");
 746         if (not_properly_sorted)
 747                 retval += report(options, tree_oid, OBJ_TREE,
 748                                  FSCK_MSG_TREE_NOT_SORTED,
 749                                  "not properly sorted");
 750         return retval;
 751 }
 752
 753 static int verify_headers(const void *data, unsigned long size,
 754                           const struct object_id *oid, enum object_type type,
 755                           struct fsck_options *options)
 756 {
 757         const char *buffer = (const char *)data;
 758         unsigned long i;
 759
 760         for (i = 0; i < size; i++) {
 761                 switch (buffer[i]) {
 762                 case '\0':
 763                         return report(options, oid, type,
 764                                 FSCK_MSG_NUL_IN_HEADER,
 765                                 "unterminated header: NUL at offset %ld", i);
 766                 case '\n':
 767                         if (i + 1 < size && buffer[i + 1] == '\n')
 768                                 return 0;
 769                 }
 770         }
 771
 772         /*
 773          * We did not find double-LF that separates the header
 774          * and the body.  Not having a body is not a crime but
 775          * we do want to see the terminating LF for the last header
 776          * line.
 777          */
 778         if (size && buffer[size - 1] == '\n')
 779                 return 0;
 780
 781         return report(options, oid, type,
 782                 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 783 }
 784
 785 static int fsck_ident(const char **ident,
 786                       const struct object_id *oid, enum object_type type,
 787                       struct fsck_options *options)
 788 {
 789         const char *p = *ident;
 790         char *end;
 791
 792         *ident = strchrnul(*ident, '\n');
 793         if (**ident == '\n')
 794                 (*ident)++;
 795
 796         if (*p == '<')
 797                 return report(options, oid, type, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 798         p += strcspn(p, "<>\n");
 799         if (*p == '>')
 800                 return report(options, oid, type, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 801         if (*p != '<')
 802                 return report(options, oid, type, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 803         if (p[-1] != ' ')
 804                 return report(options, oid, type, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 805         p++;
 806         p += strcspn(p, "<>\n");
 807         if (*p != '>')
 808                 return report(options, oid, type, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 809         p++;
 810         if (*p != ' ')
 811                 return report(options, oid, type, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 812         p++;
 813         if (*p == '0' && p[1] != ' ')
 814                 return report(options, oid, type, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 815         if (date_overflows(parse_timestamp(p, &end, 10)))
 816                 return report(options, oid, type, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 817         if ((end == p || *end != ' '))
 818                 return report(options, oid, type, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 819         p = end + 1;
 820         if ((*p != '+' && *p != '-') ||
 821             !isdigit(p[1]) ||
 822             !isdigit(p[2]) ||
 823             !isdigit(p[3]) ||
 824             !isdigit(p[4]) ||
 825             (p[5] != '\n'))
 826                 return report(options, oid, type, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 827         p += 6;
 828         return 0;
 829 }
 830
 831 static int fsck_commit(const struct object_id *oid,
 832                        const char *buffer, unsigned long size,
 833                        struct fsck_options *options)
 834 {
 835         struct object_id tree_oid, parent_oid;
 836         unsigned author_count;
 837         int err;
 838         const char *buffer_begin = buffer;
 839         const char *p;
 840
 841         if (verify_headers(buffer, size, oid, OBJ_COMMIT, options))
 842                 return -1;
 843
 844         if (!skip_prefix(buffer, "tree ", &buffer))
 845                 return report(options, oid, OBJ_COMMIT, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 846         if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') {
 847                 err = report(options, oid, OBJ_COMMIT, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 848                 if (err)
 849                         return err;
 850         }
 851         buffer = p + 1;
 852         while (skip_prefix(buffer, "parent ", &buffer)) {
 853                 if (parse_oid_hex(buffer, &parent_oid, &p) || *p != '\n') {
 854                         err = report(options, oid, OBJ_COMMIT, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 855                         if (err)
 856                                 return err;
 857                 }
 858                 buffer = p + 1;
 859         }
 860         author_count = 0;
 861         while (skip_prefix(buffer, "author ", &buffer)) {
 862                 author_count++;
 863                 err = fsck_ident(&buffer, oid, OBJ_COMMIT, options);
 864                 if (err)
 865                         return err;
 866         }
 867         if (author_count < 1)
 868                 err = report(options, oid, OBJ_COMMIT, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 869         else if (author_count > 1)
 870                 err = report(options, oid, OBJ_COMMIT, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 871         if (err)
 872                 return err;
 873         if (!skip_prefix(buffer, "committer ", &buffer))
 874                 return report(options, oid, OBJ_COMMIT, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 875         err = fsck_ident(&buffer, oid, OBJ_COMMIT, options);
 876         if (err)
 877                 return err;
 878         if (memchr(buffer_begin, '\0', size)) {
 879                 err = report(options, oid, OBJ_COMMIT, FSCK_MSG_NUL_IN_COMMIT,
 880                              "NUL byte in the commit object body");
 881                 if (err)
 882                         return err;
 883         }
 884         return 0;
 885 }
 886
 887 static int fsck_tag(const struct object_id *oid, const char *buffer,
 888                     unsigned long size, struct fsck_options *options)
 889 {
 890         struct object_id tagged_oid;
 891         int tagged_type;
 892         return fsck_tag_standalone(oid, buffer, size, options, &tagged_oid,
 893                                    &tagged_type);
 894 }
 895
 896 int fsck_tag_standalone(const struct object_id *oid, const char *buffer,
 897                         unsigned long size, struct fsck_options *options,
 898                         struct object_id *tagged_oid,
 899                         int *tagged_type)
 900 {
 901         int ret = 0;
 902         char *eol;
 903         struct strbuf sb = STRBUF_INIT;
 904         const char *p;
 905
 906         ret = verify_headers(buffer, size, oid, OBJ_TAG, options);
 907         if (ret)
 908                 goto done;
 909
 910         if (!skip_prefix(buffer, "object ", &buffer)) {
 911                 ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 912                 goto done;
 913         }
 914         if (parse_oid_hex(buffer, tagged_oid, &p) || *p != '\n') {
 915                 ret = report(options, oid, OBJ_TAG, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 916                 if (ret)
 917                         goto done;
 918         }
 919         buffer = p + 1;
 920
 921         if (!skip_prefix(buffer, "type ", &buffer)) {
 922                 ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 923                 goto done;
 924         }
 925         eol = strchr(buffer, '\n');
 926         if (!eol) {
 927                 ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 928                 goto done;
 929         }
 930         *tagged_type = type_from_string_gently(buffer, eol - buffer, 1);
 931         if (*tagged_type < 0)
 932                 ret = report(options, oid, OBJ_TAG, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 933         if (ret)
 934                 goto done;
 935         buffer = eol + 1;
 936
 937         if (!skip_prefix(buffer, "tag ", &buffer)) {
 938                 ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 939                 goto done;
 940         }
 941         eol = strchr(buffer, '\n');
 942         if (!eol) {
 943                 ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 944                 goto done;
 945         }
 946         strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 947         if (check_refname_format(sb.buf, 0)) {
 948                 ret = report(options, oid, OBJ_TAG,
 949                              FSCK_MSG_BAD_TAG_NAME,
 950                              "invalid 'tag' name: %.*s",
 951                              (int)(eol - buffer), buffer);
 952                 if (ret)
 953                         goto done;
 954         }
 955         buffer = eol + 1;
 956
 957         if (!skip_prefix(buffer, "tagger ", &buffer)) {
 958                 /* early tags do not contain 'tagger' lines; warn only */
 959                 ret = report(options, oid, OBJ_TAG, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 960                 if (ret)
 961                         goto done;
 962         }
 963         else
 964                 ret = fsck_ident(&buffer, oid, OBJ_TAG, options);
 965         if (!*buffer)
 966                 goto done;
 967
 968         if (!starts_with(buffer, "\n")) {
 969                 /*
 970                  * The verify_headers() check will allow
 971                  * e.g. "[...]tagger <tagger>\nsome
 972                  * garbage\n\nmessage" to pass, thinking "some
 973                  * garbage" could be a custom header. E.g. "mktag"
 974                  * doesn't want any unknown headers.
 975                  */
 976                 ret = report(options, oid, OBJ_TAG, FSCK_MSG_EXTRA_HEADER_ENTRY, "invalid format - extra header(s) after 'tagger'");
 977                 if (ret)
 978                         goto done;
 979         }
 980
 981 done:
 982         strbuf_release(&sb);
 983         return ret;
 984 }
 985
 986 static int starts_with_dot_slash(const char *const path)
 987 {
 988         return path_match_flags(path, PATH_MATCH_STARTS_WITH_DOT_SLASH |
 989                                 PATH_MATCH_XPLATFORM);
 990 }
 991
 992 static int starts_with_dot_dot_slash(const char *const path)
 993 {
 994         return path_match_flags(path, PATH_MATCH_STARTS_WITH_DOT_DOT_SLASH |
 995                                 PATH_MATCH_XPLATFORM);
 996 }
 997
 998 static int submodule_url_is_relative(const char *url)
 999 {
1000         return starts_with_dot_slash(url) || starts_with_dot_dot_slash(url);
1001 }
1002
1003 /*
1004  * Count directory components that a relative submodule URL should chop
1005  * from the remote_url it is to be resolved against.
1006  *
1007  * In other words, this counts "../" components at the start of a
1008  * submodule URL.
1009  *
1010  * Returns the number of directory components to chop and writes a
1011  * pointer to the next character of url after all leading "./" and
1012  * "../" components to out.
1013  */
1014 static int count_leading_dotdots(const char *url, const char **out)
1015 {
1016         int result = 0;
1017         while (1) {
1018                 if (starts_with_dot_dot_slash(url)) {
1019                         result++;
1020                         url += strlen("../");
1021                         continue;
1022                 }
1023                 if (starts_with_dot_slash(url)) {
1024                         url += strlen("./");
1025                         continue;
1026                 }
1027                 *out = url;
1028                 return result;
1029         }
1030 }
1031 /*
1032  * Check whether a transport is implemented by git-remote-curl.
1033  *
1034  * If it is, returns 1 and writes the URL that would be passed to
1035  * git-remote-curl to the "out" parameter.
1036  *
1037  * Otherwise, returns 0 and leaves "out" untouched.
1038  *
1039  * Examples:
1040  *   http::https://example.com/repo.git -> 1, https://example.com/repo.git
1041  *   https://example.com/repo.git -> 1, https://example.com/repo.git
1042  *   git://example.com/repo.git -> 0
1043  *
1044  * This is for use in checking for previously exploitable bugs that
1045  * required a submodule URL to be passed to git-remote-curl.
1046  */
1047 static int url_to_curl_url(const char *url, const char **out)
1048 {
1049         /*
1050          * We don't need to check for case-aliases, "http.exe", and so
1051          * on because in the default configuration, is_transport_allowed
1052          * prevents URLs with those schemes from being cloned
1053          * automatically.
1054          */
1055         if (skip_prefix(url, "http::", out) ||
1056             skip_prefix(url, "https::", out) ||
1057             skip_prefix(url, "ftp::", out) ||
1058             skip_prefix(url, "ftps::", out))
1059                 return 1;
1060         if (starts_with(url, "http://") ||
1061             starts_with(url, "https://") ||
1062             starts_with(url, "ftp://") ||
1063             starts_with(url, "ftps://")) {
1064                 *out = url;
1065                 return 1;
1066         }
1067         return 0;
1068 }
1069
1070 static int check_submodule_url(const char *url)
1071 {
1072         const char *curl_url;
1073
1074         if (looks_like_command_line_option(url))
1075                 return -1;
1076
1077         if (submodule_url_is_relative(url) || starts_with(url, "git://")) {
1078                 char *decoded;
1079                 const char *next;
1080                 int has_nl;
1081
1082                 /*
1083                  * This could be appended to an http URL and url-decoded;
1084                  * check for malicious characters.
1085                  */
1086                 decoded = url_decode(url);
1087                 has_nl = !!strchr(decoded, '\n');
1088
1089                 free(decoded);
1090                 if (has_nl)
1091                         return -1;
1092
1093                 /*
1094                  * URLs which escape their root via "../" can overwrite
1095                  * the host field and previous components, resolving to
1096                  * URLs like https::example.com/submodule.git and
1097                  * https:///example.com/submodule.git that were
1098                  * susceptible to CVE-2020-11008.
1099                  */
1100                 if (count_leading_dotdots(url, &next) > 0 &&
1101                     (*next == ':' || *next == '/'))
1102                         return -1;
1103         }
1104
1105         else if (url_to_curl_url(url, &curl_url)) {
1106                 struct credential c = CREDENTIAL_INIT;
1107                 int ret = 0;
1108                 if (credential_from_url_gently(&c, curl_url, 1) ||
1109                     !*c.host)
1110                         ret = -1;
1111                 credential_clear(&c);
1112                 return ret;
1113         }
1114
1115         return 0;
1116 }
1117
1118 struct fsck_gitmodules_data {
1119         const struct object_id *oid;
1120         struct fsck_options *options;
1121         int ret;
1122 };
1123
1124 static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
1125 {
1126         struct fsck_gitmodules_data *data = vdata;
1127         const char *subsection, *key;
1128         size_t subsection_len;
1129         char *name;
1130
1131         if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
1132             !subsection)
1133                 return 0;
1134
1135         name = xmemdupz(subsection, subsection_len);
1136         if (check_submodule_name(name) < 0)
1137                 data->ret |= report(data->options,
1138                                     data->oid, OBJ_BLOB,
1139                                     FSCK_MSG_GITMODULES_NAME,
1140                                     "disallowed submodule name: %s",
1141                                     name);
1142         if (!strcmp(key, "url") && value &&
1143             check_submodule_url(value) < 0)
1144                 data->ret |= report(data->options,
1145                                     data->oid, OBJ_BLOB,
1146                                     FSCK_MSG_GITMODULES_URL,
1147                                     "disallowed submodule url: %s",
1148                                     value);
1149         if (!strcmp(key, "path") && value &&
1150             looks_like_command_line_option(value))
1151                 data->ret |= report(data->options,
1152                                     data->oid, OBJ_BLOB,
1153                                     FSCK_MSG_GITMODULES_PATH,
1154                                     "disallowed submodule path: %s",
1155                                     value);
1156         if (!strcmp(key, "update") && value &&
1157             parse_submodule_update_type(value) == SM_UPDATE_COMMAND)
1158                 data->ret |= report(data->options, data->oid, OBJ_BLOB,
1159                                     FSCK_MSG_GITMODULES_UPDATE,
1160                                     "disallowed submodule update setting: %s",
1161                                     value);
1162         free(name);
1163
1164         return 0;
1165 }
1166
1167 static int fsck_blob(const struct object_id *oid, const char *buf,
1168                      unsigned long size, struct fsck_options *options)
1169 {
1170         int ret = 0;
1171
1172         if (object_on_skiplist(options, oid))
1173                 return 0;
1174
1175         if (oidset_contains(&options->gitmodules_found, oid)) {
1176                 struct config_options config_opts = { 0 };
1177                 struct fsck_gitmodules_data data;
1178
1179                 oidset_insert(&options->gitmodules_done, oid);
1180
1181                 if (!buf) {
1182                         /*
1183                          * A missing buffer here is a sign that the caller found the
1184                          * blob too gigantic to load into memory. Let's just consider
1185                          * that an error.
1186                          */
1187                         return report(options, oid, OBJ_BLOB,
1188                                         FSCK_MSG_GITMODULES_LARGE,
1189                                         ".gitmodules too large to parse");
1190                 }
1191
1192                 data.oid = oid;
1193                 data.options = options;
1194                 data.ret = 0;
1195                 config_opts.error_action = CONFIG_ERROR_SILENT;
1196                 if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
1197                                         ".gitmodules", buf, size, &data, &config_opts))
1198                         data.ret |= report(options, oid, OBJ_BLOB,
1199                                         FSCK_MSG_GITMODULES_PARSE,
1200                                         "could not parse gitmodules blob");
1201                 ret |= data.ret;
1202         }
1203
1204         if (oidset_contains(&options->gitattributes_found, oid)) {
1205                 const char *ptr;
1206
1207                 oidset_insert(&options->gitattributes_done, oid);
1208
1209                 if (!buf || size > ATTR_MAX_FILE_SIZE) {
1210                         /*
1211                          * A missing buffer here is a sign that the caller found the
1212                          * blob too gigantic to load into memory. Let's just consider
1213                          * that an error.
1214                          */
1215                         return report(options, oid, OBJ_BLOB,
1216                                         FSCK_MSG_GITATTRIBUTES_LARGE,
1217                                         ".gitattributes too large to parse");
1218                 }
1219
1220                 for (ptr = buf; *ptr; ) {
1221                         const char *eol = strchrnul(ptr, '\n');
1222                         if (eol - ptr >= ATTR_MAX_LINE_LENGTH) {
1223                                 ret |= report(options, oid, OBJ_BLOB,
1224                                               FSCK_MSG_GITATTRIBUTES_LINE_LENGTH,
1225                                               ".gitattributes has too long lines to parse");
1226                                 break;
1227                         }
1228
1229                         ptr = *eol ? eol + 1 : eol;
1230                 }
1231         }
1232
1233         if (oidset_contains(&options->symlink_targets_found, oid)) {
1234                 const char *ptr = buf;
1235                 const struct object_id *reported = NULL;
1236
1237                 oidset_insert(&options->symlink_targets_done, oid);
1238
1239                 if (!buf || size > PATH_MAX) {
1240                         /*
1241                          * A missing buffer here is a sign that the caller found the
1242                          * blob too gigantic to load into memory. Let's just consider
1243                          * that an error.
1244                          */
1245                         return report(options, oid, OBJ_BLOB,
1246                                         FSCK_MSG_SYMLINK_TARGET_LENGTH,
1247                                         "symlink target too long");
1248                 }
1249
1250                 while (!reported && ptr) {
1251                         const char *p = ptr;
1252                         char c, *slash = strchrnul(ptr, '/');
1253                         char *backslash = memchr(ptr, '\\', slash - ptr);
1254
1255                         c = *slash;
1256                         *slash = '\0';
1257
1258                         while (!reported && backslash) {
1259                                 *backslash = '\0';
1260                                 if (is_ntfs_dotgit(p))
1261                                         ret |= report(options, reported = oid, OBJ_BLOB,
1262                                                       FSCK_MSG_SYMLINK_POINTS_TO_GIT_DIR,
1263                                                       "symlink target points to git dir");
1264                                 *backslash = '\\';
1265                                 p = backslash + 1;
1266                                 backslash = memchr(p, '\\', slash - p);
1267                         }
1268                         if (!reported && is_ntfs_dotgit(p))
1269                                 ret |= report(options, reported = oid, OBJ_BLOB,
1270                                               FSCK_MSG_SYMLINK_POINTS_TO_GIT_DIR,
1271                                               "symlink target points to git dir");
1272
1273                         if (!reported && is_hfs_dotgit(ptr))
1274                                 ret |= report(options, reported = oid, OBJ_BLOB,
1275                                               FSCK_MSG_SYMLINK_POINTS_TO_GIT_DIR,
1276                                               "symlink target points to git dir");
1277
1278                         *slash = c;
1279                         ptr = c ? slash + 1 : NULL;
1280                 }
1281         }
1282
1283         return ret;
1284 }
1285
1286 int fsck_object(struct object *obj, void *data, unsigned long size,
1287         struct fsck_options *options)
1288 {
1289         if (!obj)
1290                 return report(options, NULL, OBJ_NONE, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
1291
1292         if (obj->type == OBJ_BLOB)
1293                 return fsck_blob(&obj->oid, data, size, options);
1294         if (obj->type == OBJ_TREE)
1295                 return fsck_tree(&obj->oid, data, size, options);
1296         if (obj->type == OBJ_COMMIT)
1297                 return fsck_commit(&obj->oid, data, size, options);
1298         if (obj->type == OBJ_TAG)
1299                 return fsck_tag(&obj->oid, data, size, options);
1300
1301         return report(options, &obj->oid, obj->type,
1302                       FSCK_MSG_UNKNOWN_TYPE,
1303                       "unknown type '%d' (internal fsck error)",
1304                       obj->type);
1305 }
1306
1307 int fsck_error_function(struct fsck_options *o,
1308                         const struct object_id *oid,
1309                         enum object_type object_type,
1310                         enum fsck_msg_type msg_type,
1311                         enum fsck_msg_id msg_id,
1312                         const char *message)
1313 {
1314         if (msg_type == FSCK_WARN) {
1315                 warning("object %s: %s", fsck_describe_object(o, oid), message);
1316                 return 0;
1317         }
1318         error("object %s: %s", fsck_describe_object(o, oid), message);
1319         return 1;
1320 }
1321
1322 static int fsck_blobs(struct oidset *blobs_found, struct oidset *blobs_done,
1323                       enum fsck_msg_id msg_missing, enum fsck_msg_id msg_type,
1324                       struct fsck_options *options, const char *blob_type)
1325 {
1326         int ret = 0;
1327         struct oidset_iter iter;
1328         const struct object_id *oid;
1329
1330         oidset_iter_init(blobs_found, &iter);
1331         while ((oid = oidset_iter_next(&iter))) {
1332                 enum object_type type;
1333                 unsigned long size;
1334                 char *buf;
1335
1336                 if (oidset_contains(blobs_done, oid))
1337                         continue;
1338
1339                 buf = read_object_file(oid, &type, &size);
1340                 if (!buf) {
1341                         if (is_promisor_object(oid))
1342                                 continue;
1343                         ret |= report(options,
1344                                       oid, OBJ_BLOB, msg_missing,
1345                                       "unable to read %s blob", blob_type);
1346                         continue;
1347                 }
1348
1349                 if (type == OBJ_BLOB)
1350                         ret |= fsck_blob(oid, buf, size, options);
1351                 else
1352                         ret |= report(options, oid, type, msg_type,
1353                                       "non-blob found at %s", blob_type);
1354                 free(buf);
1355         }
1356
1357         oidset_clear(blobs_found);
1358         oidset_clear(blobs_done);
1359
1360         return ret;
1361 }
1362
1363 int fsck_finish(struct fsck_options *options)
1364 {
1365         int ret = 0;
1366
1367         ret |= fsck_blobs(&options->gitmodules_found, &options->gitmodules_done,
1368                           FSCK_MSG_GITMODULES_MISSING, FSCK_MSG_GITMODULES_BLOB,
1369                           options, ".gitmodules");
1370         ret |= fsck_blobs(&options->gitattributes_found, &options->gitattributes_done,
1371                           FSCK_MSG_GITATTRIBUTES_MISSING, FSCK_MSG_GITATTRIBUTES_BLOB,
1372                           options, ".gitattributes");
1373
1374         ret |= fsck_blobs(&options->symlink_targets_found, &options->symlink_targets_done,
1375                           FSCK_MSG_SYMLINK_TARGET_MISSING, FSCK_MSG_SYMLINK_TARGET_BLOB,
1376                           options, "<symlink-target>");
1377
1378         return ret;
1379 }
1380
1381 int git_fsck_config(const char *var, const char *value, void *cb)
1382 {
1383         struct fsck_options *options = cb;
1384         if (strcmp(var, "fsck.skiplist") == 0) {
1385                 const char *path;
1386                 struct strbuf sb = STRBUF_INIT;
1387
1388                 if (git_config_pathname(&path, var, value))
1389                         return 1;
1390                 strbuf_addf(&sb, "skiplist=%s", path);
1391                 free((char *)path);
1392                 fsck_set_msg_types(options, sb.buf);
1393                 strbuf_release(&sb);
1394                 return 0;
1395         }
1396
1397         if (skip_prefix(var, "fsck.", &var)) {
1398                 fsck_set_msg_type(options, var, value);
1399                 return 0;
1400         }
1401
1402         return git_default_config(var, value, cb);
1403 }
1404
1405 /*
1406  * Custom error callbacks that are used in more than one place.
1407  */
1408
1409 int fsck_error_cb_print_missing_gitmodules(struct fsck_options *o,
1410                                            const struct object_id *oid,
1411                                            enum object_type object_type,
1412                                            enum fsck_msg_type msg_type,
1413                                            enum fsck_msg_id msg_id,
1414                                            const char *message)
1415 {
1416         if (msg_id == FSCK_MSG_GITMODULES_MISSING) {
1417                 puts(oid_to_hex(oid));
1418                 return 0;
1419         }
1420         return fsck_error_function(o, oid, object_type, msg_type, msg_id, message);
1421 }