fsck.c

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