fsck.c

   1 #include "cache.h"
   2 #include "object.h"
   3 #include "blob.h"
   4 #include "tree.h"
   5 #include "tree-walk.h"
   6 #include "commit.h"
   7 #include "tag.h"
   8 #include "fsck.h"
   9 #include "refs.h"
  10 #include "url.h"
  11 #include "utf8.h"
  12 #include "sha1-array.h"
  13 #include "decorate.h"
  14 #include "oidset.h"
  15 #include "packfile.h"
  16 #include "submodule-config.h"
  17 #include "config.h"
  18 #include "credential.h"
  19
  20 static struct oidset gitmodules_found = OIDSET_INIT;
  21 static struct oidset gitmodules_done = OIDSET_INIT;
  22
  23 #define FSCK_FATAL -1
  24 #define FSCK_INFO -2
  25
  26 #define FOREACH_MSG_ID(FUNC) \
  27         /* fatal errors */ \
  28         FUNC(NUL_IN_HEADER, FATAL) \
  29         FUNC(UNTERMINATED_HEADER, FATAL) \
  30         /* errors */ \
  31         FUNC(BAD_DATE, ERROR) \
  32         FUNC(BAD_DATE_OVERFLOW, ERROR) \
  33         FUNC(BAD_EMAIL, ERROR) \
  34         FUNC(BAD_NAME, ERROR) \
  35         FUNC(BAD_OBJECT_SHA1, ERROR) \
  36         FUNC(BAD_PARENT_SHA1, ERROR) \
  37         FUNC(BAD_TAG_OBJECT, ERROR) \
  38         FUNC(BAD_TIMEZONE, ERROR) \
  39         FUNC(BAD_TREE, ERROR) \
  40         FUNC(BAD_TREE_SHA1, ERROR) \
  41         FUNC(BAD_TYPE, ERROR) \
  42         FUNC(DUPLICATE_ENTRIES, ERROR) \
  43         FUNC(MISSING_AUTHOR, ERROR) \
  44         FUNC(MISSING_COMMITTER, ERROR) \
  45         FUNC(MISSING_EMAIL, ERROR) \
  46         FUNC(MISSING_GRAFT, ERROR) \
  47         FUNC(MISSING_NAME_BEFORE_EMAIL, ERROR) \
  48         FUNC(MISSING_OBJECT, ERROR) \
  49         FUNC(MISSING_PARENT, ERROR) \
  50         FUNC(MISSING_SPACE_BEFORE_DATE, ERROR) \
  51         FUNC(MISSING_SPACE_BEFORE_EMAIL, ERROR) \
  52         FUNC(MISSING_TAG, ERROR) \
  53         FUNC(MISSING_TAG_ENTRY, ERROR) \
  54         FUNC(MISSING_TAG_OBJECT, ERROR) \
  55         FUNC(MISSING_TREE, ERROR) \
  56         FUNC(MISSING_TREE_OBJECT, ERROR) \
  57         FUNC(MISSING_TYPE, ERROR) \
  58         FUNC(MISSING_TYPE_ENTRY, ERROR) \
  59         FUNC(MULTIPLE_AUTHORS, ERROR) \
  60         FUNC(TAG_OBJECT_NOT_TAG, ERROR) \
  61         FUNC(TREE_NOT_SORTED, ERROR) \
  62         FUNC(UNKNOWN_TYPE, ERROR) \
  63         FUNC(ZERO_PADDED_DATE, ERROR) \
  64         FUNC(GITMODULES_MISSING, ERROR) \
  65         FUNC(GITMODULES_BLOB, ERROR) \
  66         FUNC(GITMODULES_PARSE, ERROR) \
  67         FUNC(GITMODULES_NAME, ERROR) \
  68         FUNC(GITMODULES_SYMLINK, ERROR) \
  69         FUNC(GITMODULES_URL, ERROR) \
  70         FUNC(GITMODULES_PATH, ERROR) \
  71         FUNC(GITMODULES_UPDATE, ERROR) \
  72         /* warnings */ \
  73         FUNC(BAD_FILEMODE, WARN) \
  74         FUNC(EMPTY_NAME, WARN) \
  75         FUNC(FULL_PATHNAME, WARN) \
  76         FUNC(HAS_DOT, WARN) \
  77         FUNC(HAS_DOTDOT, WARN) \
  78         FUNC(HAS_DOTGIT, WARN) \
  79         FUNC(NULL_SHA1, WARN) \
  80         FUNC(ZERO_PADDED_FILEMODE, WARN) \
  81         FUNC(NUL_IN_COMMIT, WARN) \
  82         /* infos (reported as warnings, but ignored by default) */ \
  83         FUNC(BAD_TAG_NAME, INFO) \
  84         FUNC(MISSING_TAGGER_ENTRY, INFO)
  85
  86 #define MSG_ID(id, msg_type) FSCK_MSG_##id,
  87 enum fsck_msg_id {
  88         FOREACH_MSG_ID(MSG_ID)
  89         FSCK_MSG_MAX
  90 };
  91 #undef MSG_ID
  92
  93 #define STR(x) #x
  94 #define MSG_ID(id, msg_type) { STR(id), NULL, FSCK_##msg_type },
  95 static struct {
  96         const char *id_string;
  97         const char *downcased;
  98         int msg_type;
  99 } msg_id_info[FSCK_MSG_MAX + 1] = {
 100         FOREACH_MSG_ID(MSG_ID)
 101         { NULL, NULL, -1 }
 102 };
 103 #undef MSG_ID
 104
 105 static int parse_msg_id(const char *text)
 106 {
 107         int i;
 108
 109         if (!msg_id_info[0].downcased) {
 110                 /* convert id_string to lower case, without underscores. */
 111                 for (i = 0; i < FSCK_MSG_MAX; i++) {
 112                         const char *p = msg_id_info[i].id_string;
 113                         int len = strlen(p);
 114                         char *q = xmalloc(len);
 115
 116                         msg_id_info[i].downcased = q;
 117                         while (*p)
 118                                 if (*p == '_')
 119                                         p++;
 120                                 else
 121                                         *(q)++ = tolower(*(p)++);
 122                         *q = '\0';
 123                 }
 124         }
 125
 126         for (i = 0; i < FSCK_MSG_MAX; i++)
 127                 if (!strcmp(text, msg_id_info[i].downcased))
 128                         return i;
 129
 130         return -1;
 131 }
 132
 133 static int fsck_msg_type(enum fsck_msg_id msg_id,
 134         struct fsck_options *options)
 135 {
 136         int msg_type;
 137
 138         assert(msg_id >= 0 && msg_id < FSCK_MSG_MAX);
 139
 140         if (options->msg_type)
 141                 msg_type = options->msg_type[msg_id];
 142         else {
 143                 msg_type = msg_id_info[msg_id].msg_type;
 144                 if (options->strict && msg_type == FSCK_WARN)
 145                         msg_type = FSCK_ERROR;
 146         }
 147
 148         return msg_type;
 149 }
 150
 151 static void init_skiplist(struct fsck_options *options, const char *path)
 152 {
 153         static struct oid_array skiplist = OID_ARRAY_INIT;
 154         int sorted, fd;
 155         char buffer[GIT_MAX_HEXSZ + 1];
 156         struct object_id oid;
 157
 158         if (options->skiplist)
 159                 sorted = options->skiplist->sorted;
 160         else {
 161                 sorted = 1;
 162                 options->skiplist = &skiplist;
 163         }
 164
 165         fd = open(path, O_RDONLY);
 166         if (fd < 0)
 167                 die("Could not open skip list: %s", path);
 168         for (;;) {
 169                 const char *p;
 170                 int result = read_in_full(fd, buffer, sizeof(buffer));
 171                 if (result < 0)
 172                         die_errno("Could not read '%s'", path);
 173                 if (!result)
 174                         break;
 175                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n')
 176                         die("Invalid SHA-1: %s", buffer);
 177                 oid_array_append(&skiplist, &oid);
 178                 if (sorted && skiplist.nr > 1 &&
 179                                 oidcmp(&skiplist.oid[skiplist.nr - 2],
 180                                        &oid) > 0)
 181                         sorted = 0;
 182         }
 183         close(fd);
 184
 185         if (sorted)
 186                 skiplist.sorted = 1;
 187 }
 188
 189 static int parse_msg_type(const char *str)
 190 {
 191         if (!strcmp(str, "error"))
 192                 return FSCK_ERROR;
 193         else if (!strcmp(str, "warn"))
 194                 return FSCK_WARN;
 195         else if (!strcmp(str, "ignore"))
 196                 return FSCK_IGNORE;
 197         else
 198                 die("Unknown fsck message type: '%s'", str);
 199 }
 200
 201 int is_valid_msg_type(const char *msg_id, const char *msg_type)
 202 {
 203         if (parse_msg_id(msg_id) < 0)
 204                 return 0;
 205         parse_msg_type(msg_type);
 206         return 1;
 207 }
 208
 209 void fsck_set_msg_type(struct fsck_options *options,
 210                 const char *msg_id, const char *msg_type)
 211 {
 212         int id = parse_msg_id(msg_id), type;
 213
 214         if (id < 0)
 215                 die("Unhandled message id: %s", msg_id);
 216         type = parse_msg_type(msg_type);
 217
 218         if (type != FSCK_ERROR && msg_id_info[id].msg_type == FSCK_FATAL)
 219                 die("Cannot demote %s to %s", msg_id, msg_type);
 220
 221         if (!options->msg_type) {
 222                 int i;
 223                 int *msg_type;
 224                 ALLOC_ARRAY(msg_type, FSCK_MSG_MAX);
 225                 for (i = 0; i < FSCK_MSG_MAX; i++)
 226                         msg_type[i] = fsck_msg_type(i, options);
 227                 options->msg_type = msg_type;
 228         }
 229
 230         options->msg_type[id] = type;
 231 }
 232
 233 void fsck_set_msg_types(struct fsck_options *options, const char *values)
 234 {
 235         char *buf = xstrdup(values), *to_free = buf;
 236         int done = 0;
 237
 238         while (!done) {
 239                 int len = strcspn(buf, " ,|"), equal;
 240
 241                 done = !buf[len];
 242                 if (!len) {
 243                         buf++;
 244                         continue;
 245                 }
 246                 buf[len] = '\0';
 247
 248                 for (equal = 0;
 249                      equal < len && buf[equal] != '=' && buf[equal] != ':';
 250                      equal++)
 251                         buf[equal] = tolower(buf[equal]);
 252                 buf[equal] = '\0';
 253
 254                 if (!strcmp(buf, "skiplist")) {
 255                         if (equal == len)
 256                                 die("skiplist requires a path");
 257                         init_skiplist(options, buf + equal + 1);
 258                         buf += len + 1;
 259                         continue;
 260                 }
 261
 262                 if (equal == len)
 263                         die("Missing '=': '%s'", buf);
 264
 265                 fsck_set_msg_type(options, buf, buf + equal + 1);
 266                 buf += len + 1;
 267         }
 268         free(to_free);
 269 }
 270
 271 static void append_msg_id(struct strbuf *sb, const char *msg_id)
 272 {
 273         for (;;) {
 274                 char c = *(msg_id)++;
 275
 276                 if (!c)
 277                         break;
 278                 if (c != '_')
 279                         strbuf_addch(sb, tolower(c));
 280                 else {
 281                         assert(*msg_id);
 282                         strbuf_addch(sb, *(msg_id)++);
 283                 }
 284         }
 285
 286         strbuf_addstr(sb, ": ");
 287 }
 288
 289 __attribute__((format (printf, 4, 5)))
 290 static int report(struct fsck_options *options, struct object *object,
 291         enum fsck_msg_id id, const char *fmt, ...)
 292 {
 293         va_list ap;
 294         struct strbuf sb = STRBUF_INIT;
 295         int msg_type = fsck_msg_type(id, options), result;
 296
 297         if (msg_type == FSCK_IGNORE)
 298                 return 0;
 299
 300         if (options->skiplist && object &&
 301                         oid_array_lookup(options->skiplist, &object->oid) >= 0)
 302                 return 0;
 303
 304         if (msg_type == FSCK_FATAL)
 305                 msg_type = FSCK_ERROR;
 306         else if (msg_type == FSCK_INFO)
 307                 msg_type = FSCK_WARN;
 308
 309         append_msg_id(&sb, msg_id_info[id].id_string);
 310
 311         va_start(ap, fmt);
 312         strbuf_vaddf(&sb, fmt, ap);
 313         result = options->error_func(options, object, msg_type, sb.buf);
 314         strbuf_release(&sb);
 315         va_end(ap);
 316
 317         return result;
 318 }
 319
 320 static char *get_object_name(struct fsck_options *options, struct object *obj)
 321 {
 322         if (!options->object_names)
 323                 return NULL;
 324         return lookup_decoration(options->object_names, obj);
 325 }
 326
 327 static void put_object_name(struct fsck_options *options, struct object *obj,
 328         const char *fmt, ...)
 329 {
 330         va_list ap;
 331         struct strbuf buf = STRBUF_INIT;
 332         char *existing;
 333
 334         if (!options->object_names)
 335                 return;
 336         existing = lookup_decoration(options->object_names, obj);
 337         if (existing)
 338                 return;
 339         va_start(ap, fmt);
 340         strbuf_vaddf(&buf, fmt, ap);
 341         add_decoration(options->object_names, obj, strbuf_detach(&buf, NULL));
 342         va_end(ap);
 343 }
 344
 345 static const char *describe_object(struct fsck_options *o, struct object *obj)
 346 {
 347         static struct strbuf buf = STRBUF_INIT;
 348         char *name;
 349
 350         strbuf_reset(&buf);
 351         strbuf_addstr(&buf, oid_to_hex(&obj->oid));
 352         if (o->object_names && (name = lookup_decoration(o->object_names, obj)))
 353                 strbuf_addf(&buf, " (%s)", name);
 354
 355         return buf.buf;
 356 }
 357
 358 static int fsck_walk_tree(struct tree *tree, void *data, struct fsck_options *options)
 359 {
 360         struct tree_desc desc;
 361         struct name_entry entry;
 362         int res = 0;
 363         const char *name;
 364
 365         if (parse_tree(tree))
 366                 return -1;
 367
 368         name = get_object_name(options, &tree->object);
 369         if (init_tree_desc_gently(&desc, tree->buffer, tree->size))
 370                 return -1;
 371         while (tree_entry_gently(&desc, &entry)) {
 372                 struct object *obj;
 373                 int result;
 374
 375                 if (S_ISGITLINK(entry.mode))
 376                         continue;
 377
 378                 if (S_ISDIR(entry.mode)) {
 379                         obj = (struct object *)lookup_tree(entry.oid);
 380                         if (name && obj)
 381                                 put_object_name(options, obj, "%s%s/", name,
 382                                         entry.path);
 383                         result = options->walk(obj, OBJ_TREE, data, options);
 384                 }
 385                 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode)) {
 386                         obj = (struct object *)lookup_blob(entry.oid);
 387                         if (name && obj)
 388                                 put_object_name(options, obj, "%s%s", name,
 389                                         entry.path);
 390                         result = options->walk(obj, OBJ_BLOB, data, options);
 391                 }
 392                 else {
 393                         result = error("in tree %s: entry %s has bad mode %.6o",
 394                                         describe_object(options, &tree->object), entry.path, entry.mode);
 395                 }
 396                 if (result < 0)
 397                         return result;
 398                 if (!res)
 399                         res = result;
 400         }
 401         return res;
 402 }
 403
 404 static int fsck_walk_commit(struct commit *commit, void *data, struct fsck_options *options)
 405 {
 406         int counter = 0, generation = 0, name_prefix_len = 0;
 407         struct commit_list *parents;
 408         int res;
 409         int result;
 410         const char *name;
 411
 412         if (parse_commit(commit))
 413                 return -1;
 414
 415         name = get_object_name(options, &commit->object);
 416         if (name)
 417                 put_object_name(options, &get_commit_tree(commit)->object,
 418                                 "%s:", name);
 419
 420         result = options->walk((struct object *)get_commit_tree(commit),
 421                                OBJ_TREE, data, options);
 422         if (result < 0)
 423                 return result;
 424         res = result;
 425
 426         parents = commit->parents;
 427         if (name && parents) {
 428                 int len = strlen(name), power;
 429
 430                 if (len && name[len - 1] == '^') {
 431                         generation = 1;
 432                         name_prefix_len = len - 1;
 433                 }
 434                 else { /* parse ~<generation> suffix */
 435                         for (generation = 0, power = 1;
 436                              len && isdigit(name[len - 1]);
 437                              power *= 10)
 438                                 generation += power * (name[--len] - '0');
 439                         if (power > 1 && len && name[len - 1] == '~')
 440                                 name_prefix_len = len - 1;
 441                 }
 442         }
 443
 444         while (parents) {
 445                 if (name) {
 446                         struct object *obj = &parents->item->object;
 447
 448                         if (++counter > 1)
 449                                 put_object_name(options, obj, "%s^%d",
 450                                         name, counter);
 451                         else if (generation > 0)
 452                                 put_object_name(options, obj, "%.*s~%d",
 453                                         name_prefix_len, name, generation + 1);
 454                         else
 455                                 put_object_name(options, obj, "%s^", name);
 456                 }
 457                 result = options->walk((struct object *)parents->item, OBJ_COMMIT, data, options);
 458                 if (result < 0)
 459                         return result;
 460                 if (!res)
 461                         res = result;
 462                 parents = parents->next;
 463         }
 464         return res;
 465 }
 466
 467 static int fsck_walk_tag(struct tag *tag, void *data, struct fsck_options *options)
 468 {
 469         char *name = get_object_name(options, &tag->object);
 470
 471         if (parse_tag(tag))
 472                 return -1;
 473         if (name)
 474                 put_object_name(options, tag->tagged, "%s", name);
 475         return options->walk(tag->tagged, OBJ_ANY, data, options);
 476 }
 477
 478 int fsck_walk(struct object *obj, void *data, struct fsck_options *options)
 479 {
 480         if (!obj)
 481                 return -1;
 482
 483         if (obj->type == OBJ_NONE)
 484                 parse_object(&obj->oid);
 485
 486         switch (obj->type) {
 487         case OBJ_BLOB:
 488                 return 0;
 489         case OBJ_TREE:
 490                 return fsck_walk_tree((struct tree *)obj, data, options);
 491         case OBJ_COMMIT:
 492                 return fsck_walk_commit((struct commit *)obj, data, options);
 493         case OBJ_TAG:
 494                 return fsck_walk_tag((struct tag *)obj, data, options);
 495         default:
 496                 error("Unknown object type for %s", describe_object(options, obj));
 497                 return -1;
 498         }
 499 }
 500
 501 /*
 502  * The entries in a tree are ordered in the _path_ order,
 503  * which means that a directory entry is ordered by adding
 504  * a slash to the end of it.
 505  *
 506  * So a directory called "a" is ordered _after_ a file
 507  * called "a.c", because "a/" sorts after "a.c".
 508  */
 509 #define TREE_UNORDERED (-1)
 510 #define TREE_HAS_DUPS  (-2)
 511
 512 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 513 {
 514         int len1 = strlen(name1);
 515         int len2 = strlen(name2);
 516         int len = len1 < len2 ? len1 : len2;
 517         unsigned char c1, c2;
 518         int cmp;
 519
 520         cmp = memcmp(name1, name2, len);
 521         if (cmp < 0)
 522                 return 0;
 523         if (cmp > 0)
 524                 return TREE_UNORDERED;
 525
 526         /*
 527          * Ok, the first <len> characters are the same.
 528          * Now we need to order the next one, but turn
 529          * a '\0' into a '/' for a directory entry.
 530          */
 531         c1 = name1[len];
 532         c2 = name2[len];
 533         if (!c1 && !c2)
 534                 /*
 535                  * git-write-tree used to write out a nonsense tree that has
 536                  * entries with the same name, one blob and one tree.  Make
 537                  * sure we do not have duplicate entries.
 538                  */
 539                 return TREE_HAS_DUPS;
 540         if (!c1 && S_ISDIR(mode1))
 541                 c1 = '/';
 542         if (!c2 && S_ISDIR(mode2))
 543                 c2 = '/';
 544         return c1 < c2 ? 0 : TREE_UNORDERED;
 545 }
 546
 547 static int fsck_tree(struct tree *item, struct fsck_options *options)
 548 {
 549         int retval = 0;
 550         int has_null_sha1 = 0;
 551         int has_full_path = 0;
 552         int has_empty_name = 0;
 553         int has_dot = 0;
 554         int has_dotdot = 0;
 555         int has_dotgit = 0;
 556         int has_zero_pad = 0;
 557         int has_bad_modes = 0;
 558         int has_dup_entries = 0;
 559         int not_properly_sorted = 0;
 560         struct tree_desc desc;
 561         unsigned o_mode;
 562         const char *o_name;
 563
 564         if (init_tree_desc_gently(&desc, item->buffer, item->size)) {
 565                 retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 566                 return retval;
 567         }
 568
 569         o_mode = 0;
 570         o_name = NULL;
 571
 572         while (desc.size) {
 573                 unsigned mode;
 574                 const char *name, *backslash;
 575                 const struct object_id *oid;
 576
 577                 oid = tree_entry_extract(&desc, &name, &mode);
 578
 579                 has_null_sha1 |= is_null_oid(oid);
 580                 has_full_path |= !!strchr(name, '/');
 581                 has_empty_name |= !*name;
 582                 has_dot |= !strcmp(name, ".");
 583                 has_dotdot |= !strcmp(name, "..");
 584                 has_dotgit |= is_hfs_dotgit(name) || is_ntfs_dotgit(name);
 585                 has_zero_pad |= *(char *)desc.buffer == '0';
 586
 587                 if (is_hfs_dotgitmodules(name) || is_ntfs_dotgitmodules(name)) {
 588                         if (!S_ISLNK(mode))
 589                                 oidset_insert(&gitmodules_found, oid);
 590                         else
 591                                 retval += report(options, &item->object,
 592                                                  FSCK_MSG_GITMODULES_SYMLINK,
 593                                                  ".gitmodules is a symbolic link");
 594                 }
 595
 596                 if ((backslash = strchr(name, '\\'))) {
 597                         while (backslash) {
 598                                 backslash++;
 599                                 has_dotgit |= is_ntfs_dotgit(backslash);
 600                                 if (is_ntfs_dotgitmodules(backslash)) {
 601                                         if (!S_ISLNK(mode))
 602                                                 oidset_insert(&gitmodules_found, oid);
 603                                         else
 604                                                 retval += report(options, &item->object,
 605                                                                  FSCK_MSG_GITMODULES_SYMLINK,
 606                                                                  ".gitmodules is a symbolic link");
 607                                 }
 608                                 backslash = strchr(backslash, '\\');
 609                         }
 610                 }
 611
 612                 if (update_tree_entry_gently(&desc)) {
 613                         retval += report(options, &item->object, FSCK_MSG_BAD_TREE, "cannot be parsed as a tree");
 614                         break;
 615                 }
 616
 617                 switch (mode) {
 618                 /*
 619                  * Standard modes..
 620                  */
 621                 case S_IFREG | 0755:
 622                 case S_IFREG | 0644:
 623                 case S_IFLNK:
 624                 case S_IFDIR:
 625                 case S_IFGITLINK:
 626                         break;
 627                 /*
 628                  * This is nonstandard, but we had a few of these
 629                  * early on when we honored the full set of mode
 630                  * bits..
 631                  */
 632                 case S_IFREG | 0664:
 633                         if (!options->strict)
 634                                 break;
 635                         /* fallthrough */
 636                 default:
 637                         has_bad_modes = 1;
 638                 }
 639
 640                 if (o_name) {
 641                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 642                         case TREE_UNORDERED:
 643                                 not_properly_sorted = 1;
 644                                 break;
 645                         case TREE_HAS_DUPS:
 646                                 has_dup_entries = 1;
 647                                 break;
 648                         default:
 649                                 break;
 650                         }
 651                 }
 652
 653                 o_mode = mode;
 654                 o_name = name;
 655         }
 656
 657         if (has_null_sha1)
 658                 retval += report(options, &item->object, FSCK_MSG_NULL_SHA1, "contains entries pointing to null sha1");
 659         if (has_full_path)
 660                 retval += report(options, &item->object, FSCK_MSG_FULL_PATHNAME, "contains full pathnames");
 661         if (has_empty_name)
 662                 retval += report(options, &item->object, FSCK_MSG_EMPTY_NAME, "contains empty pathname");
 663         if (has_dot)
 664                 retval += report(options, &item->object, FSCK_MSG_HAS_DOT, "contains '.'");
 665         if (has_dotdot)
 666                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTDOT, "contains '..'");
 667         if (has_dotgit)
 668                 retval += report(options, &item->object, FSCK_MSG_HAS_DOTGIT, "contains '.git'");
 669         if (has_zero_pad)
 670                 retval += report(options, &item->object, FSCK_MSG_ZERO_PADDED_FILEMODE, "contains zero-padded file modes");
 671         if (has_bad_modes)
 672                 retval += report(options, &item->object, FSCK_MSG_BAD_FILEMODE, "contains bad file modes");
 673         if (has_dup_entries)
 674                 retval += report(options, &item->object, FSCK_MSG_DUPLICATE_ENTRIES, "contains duplicate file entries");
 675         if (not_properly_sorted)
 676                 retval += report(options, &item->object, FSCK_MSG_TREE_NOT_SORTED, "not properly sorted");
 677         return retval;
 678 }
 679
 680 static int verify_headers(const void *data, unsigned long size,
 681                           struct object *obj, struct fsck_options *options)
 682 {
 683         const char *buffer = (const char *)data;
 684         unsigned long i;
 685
 686         for (i = 0; i < size; i++) {
 687                 switch (buffer[i]) {
 688                 case '\0':
 689                         return report(options, obj,
 690                                 FSCK_MSG_NUL_IN_HEADER,
 691                                 "unterminated header: NUL at offset %ld", i);
 692                 case '\n':
 693                         if (i + 1 < size && buffer[i + 1] == '\n')
 694                                 return 0;
 695                 }
 696         }
 697
 698         /*
 699          * We did not find double-LF that separates the header
 700          * and the body.  Not having a body is not a crime but
 701          * we do want to see the terminating LF for the last header
 702          * line.
 703          */
 704         if (size && buffer[size - 1] == '\n')
 705                 return 0;
 706
 707         return report(options, obj,
 708                 FSCK_MSG_UNTERMINATED_HEADER, "unterminated header");
 709 }
 710
 711 static int fsck_ident(const char **ident, struct object *obj, struct fsck_options *options)
 712 {
 713         const char *p = *ident;
 714         char *end;
 715
 716         *ident = strchrnul(*ident, '\n');
 717         if (**ident == '\n')
 718                 (*ident)++;
 719
 720         if (*p == '<')
 721                 return report(options, obj, FSCK_MSG_MISSING_NAME_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 722         p += strcspn(p, "<>\n");
 723         if (*p == '>')
 724                 return report(options, obj, FSCK_MSG_BAD_NAME, "invalid author/committer line - bad name");
 725         if (*p != '<')
 726                 return report(options, obj, FSCK_MSG_MISSING_EMAIL, "invalid author/committer line - missing email");
 727         if (p[-1] != ' ')
 728                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_EMAIL, "invalid author/committer line - missing space before email");
 729         p++;
 730         p += strcspn(p, "<>\n");
 731         if (*p != '>')
 732                 return report(options, obj, FSCK_MSG_BAD_EMAIL, "invalid author/committer line - bad email");
 733         p++;
 734         if (*p != ' ')
 735                 return report(options, obj, FSCK_MSG_MISSING_SPACE_BEFORE_DATE, "invalid author/committer line - missing space before date");
 736         p++;
 737         if (*p == '0' && p[1] != ' ')
 738                 return report(options, obj, FSCK_MSG_ZERO_PADDED_DATE, "invalid author/committer line - zero-padded date");
 739         if (date_overflows(parse_timestamp(p, &end, 10)))
 740                 return report(options, obj, FSCK_MSG_BAD_DATE_OVERFLOW, "invalid author/committer line - date causes integer overflow");
 741         if ((end == p || *end != ' '))
 742                 return report(options, obj, FSCK_MSG_BAD_DATE, "invalid author/committer line - bad date");
 743         p = end + 1;
 744         if ((*p != '+' && *p != '-') ||
 745             !isdigit(p[1]) ||
 746             !isdigit(p[2]) ||
 747             !isdigit(p[3]) ||
 748             !isdigit(p[4]) ||
 749             (p[5] != '\n'))
 750                 return report(options, obj, FSCK_MSG_BAD_TIMEZONE, "invalid author/committer line - bad time zone");
 751         p += 6;
 752         return 0;
 753 }
 754
 755 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 756         unsigned long size, struct fsck_options *options)
 757 {
 758         struct object_id tree_oid, oid;
 759         struct commit_graft *graft;
 760         unsigned parent_count, parent_line_count = 0, author_count;
 761         int err;
 762         const char *buffer_begin = buffer;
 763         const char *p;
 764
 765         if (verify_headers(buffer, size, &commit->object, options))
 766                 return -1;
 767
 768         if (!skip_prefix(buffer, "tree ", &buffer))
 769                 return report(options, &commit->object, FSCK_MSG_MISSING_TREE, "invalid format - expected 'tree' line");
 770         if (parse_oid_hex(buffer, &tree_oid, &p) || *p != '\n') {
 771                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE_SHA1, "invalid 'tree' line format - bad sha1");
 772                 if (err)
 773                         return err;
 774         }
 775         buffer = p + 1;
 776         while (skip_prefix(buffer, "parent ", &buffer)) {
 777                 if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 778                         err = report(options, &commit->object, FSCK_MSG_BAD_PARENT_SHA1, "invalid 'parent' line format - bad sha1");
 779                         if (err)
 780                                 return err;
 781                 }
 782                 buffer = p + 1;
 783                 parent_line_count++;
 784         }
 785         graft = lookup_commit_graft(&commit->object.oid);
 786         parent_count = commit_list_count(commit->parents);
 787         if (graft) {
 788                 if (graft->nr_parent == -1 && !parent_count)
 789                         ; /* shallow commit */
 790                 else if (graft->nr_parent != parent_count) {
 791                         err = report(options, &commit->object, FSCK_MSG_MISSING_GRAFT, "graft objects missing");
 792                         if (err)
 793                                 return err;
 794                 }
 795         } else {
 796                 if (parent_count != parent_line_count) {
 797                         err = report(options, &commit->object, FSCK_MSG_MISSING_PARENT, "parent objects missing");
 798                         if (err)
 799                                 return err;
 800                 }
 801         }
 802         author_count = 0;
 803         while (skip_prefix(buffer, "author ", &buffer)) {
 804                 author_count++;
 805                 err = fsck_ident(&buffer, &commit->object, options);
 806                 if (err)
 807                         return err;
 808         }
 809         if (author_count < 1)
 810                 err = report(options, &commit->object, FSCK_MSG_MISSING_AUTHOR, "invalid format - expected 'author' line");
 811         else if (author_count > 1)
 812                 err = report(options, &commit->object, FSCK_MSG_MULTIPLE_AUTHORS, "invalid format - multiple 'author' lines");
 813         if (err)
 814                 return err;
 815         if (!skip_prefix(buffer, "committer ", &buffer))
 816                 return report(options, &commit->object, FSCK_MSG_MISSING_COMMITTER, "invalid format - expected 'committer' line");
 817         err = fsck_ident(&buffer, &commit->object, options);
 818         if (err)
 819                 return err;
 820         if (!get_commit_tree(commit)) {
 821                 err = report(options, &commit->object, FSCK_MSG_BAD_TREE, "could not load commit's tree %s", oid_to_hex(&tree_oid));
 822                 if (err)
 823                         return err;
 824         }
 825         if (memchr(buffer_begin, '\0', size)) {
 826                 err = report(options, &commit->object, FSCK_MSG_NUL_IN_COMMIT,
 827                              "NUL byte in the commit object body");
 828                 if (err)
 829                         return err;
 830         }
 831         return 0;
 832 }
 833
 834 static int fsck_commit(struct commit *commit, const char *data,
 835         unsigned long size, struct fsck_options *options)
 836 {
 837         const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 838         int ret = fsck_commit_buffer(commit, buffer, size, options);
 839         if (!data)
 840                 unuse_commit_buffer(commit, buffer);
 841         return ret;
 842 }
 843
 844 static int fsck_tag_buffer(struct tag *tag, const char *data,
 845         unsigned long size, struct fsck_options *options)
 846 {
 847         struct object_id oid;
 848         int ret = 0;
 849         const char *buffer;
 850         char *to_free = NULL, *eol;
 851         struct strbuf sb = STRBUF_INIT;
 852         const char *p;
 853
 854         if (data)
 855                 buffer = data;
 856         else {
 857                 enum object_type type;
 858
 859                 buffer = to_free =
 860                         read_object_file(&tag->object.oid, &type, &size);
 861                 if (!buffer)
 862                         return report(options, &tag->object,
 863                                 FSCK_MSG_MISSING_TAG_OBJECT,
 864                                 "cannot read tag object");
 865
 866                 if (type != OBJ_TAG) {
 867                         ret = report(options, &tag->object,
 868                                 FSCK_MSG_TAG_OBJECT_NOT_TAG,
 869                                 "expected tag got %s",
 870                             type_name(type));
 871                         goto done;
 872                 }
 873         }
 874
 875         ret = verify_headers(buffer, size, &tag->object, options);
 876         if (ret)
 877                 goto done;
 878
 879         if (!skip_prefix(buffer, "object ", &buffer)) {
 880                 ret = report(options, &tag->object, FSCK_MSG_MISSING_OBJECT, "invalid format - expected 'object' line");
 881                 goto done;
 882         }
 883         if (parse_oid_hex(buffer, &oid, &p) || *p != '\n') {
 884                 ret = report(options, &tag->object, FSCK_MSG_BAD_OBJECT_SHA1, "invalid 'object' line format - bad sha1");
 885                 if (ret)
 886                         goto done;
 887         }
 888         buffer = p + 1;
 889
 890         if (!skip_prefix(buffer, "type ", &buffer)) {
 891                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE_ENTRY, "invalid format - expected 'type' line");
 892                 goto done;
 893         }
 894         eol = strchr(buffer, '\n');
 895         if (!eol) {
 896                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TYPE, "invalid format - unexpected end after 'type' line");
 897                 goto done;
 898         }
 899         if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 900                 ret = report(options, &tag->object, FSCK_MSG_BAD_TYPE, "invalid 'type' value");
 901         if (ret)
 902                 goto done;
 903         buffer = eol + 1;
 904
 905         if (!skip_prefix(buffer, "tag ", &buffer)) {
 906                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG_ENTRY, "invalid format - expected 'tag' line");
 907                 goto done;
 908         }
 909         eol = strchr(buffer, '\n');
 910         if (!eol) {
 911                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAG, "invalid format - unexpected end after 'type' line");
 912                 goto done;
 913         }
 914         strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 915         if (check_refname_format(sb.buf, 0)) {
 916                 ret = report(options, &tag->object, FSCK_MSG_BAD_TAG_NAME,
 917                            "invalid 'tag' name: %.*s",
 918                            (int)(eol - buffer), buffer);
 919                 if (ret)
 920                         goto done;
 921         }
 922         buffer = eol + 1;
 923
 924         if (!skip_prefix(buffer, "tagger ", &buffer)) {
 925                 /* early tags do not contain 'tagger' lines; warn only */
 926                 ret = report(options, &tag->object, FSCK_MSG_MISSING_TAGGER_ENTRY, "invalid format - expected 'tagger' line");
 927                 if (ret)
 928                         goto done;
 929         }
 930         else
 931                 ret = fsck_ident(&buffer, &tag->object, options);
 932
 933 done:
 934         strbuf_release(&sb);
 935         free(to_free);
 936         return ret;
 937 }
 938
 939 static int fsck_tag(struct tag *tag, const char *data,
 940         unsigned long size, struct fsck_options *options)
 941 {
 942         struct object *tagged = tag->tagged;
 943
 944         if (!tagged)
 945                 return report(options, &tag->object, FSCK_MSG_BAD_TAG_OBJECT, "could not load tagged object");
 946
 947         return fsck_tag_buffer(tag, data, size, options);
 948 }
 949
 950 /*
 951  * Like builtin/submodule--helper.c's starts_with_dot_slash, but without
 952  * relying on the platform-dependent is_dir_sep helper.
 953  *
 954  * This is for use in checking whether a submodule URL is interpreted as
 955  * relative to the current directory on any platform, since \ is a
 956  * directory separator on Windows but not on other platforms.
 957  */
 958 static int starts_with_dot_slash(const char *str)
 959 {
 960         return str[0] == '.' && (str[1] == '/' || str[1] == '\\');
 961 }
 962
 963 /*
 964  * Like starts_with_dot_slash, this is a variant of submodule--helper's
 965  * helper of the same name with the twist that it accepts backslash as a
 966  * directory separator even on non-Windows platforms.
 967  */
 968 static int starts_with_dot_dot_slash(const char *str)
 969 {
 970         return str[0] == '.' && starts_with_dot_slash(str + 1);
 971 }
 972
 973 static int submodule_url_is_relative(const char *url)
 974 {
 975         return starts_with_dot_slash(url) || starts_with_dot_dot_slash(url);
 976 }
 977
 978 /*
 979  * Count directory components that a relative submodule URL should chop
 980  * from the remote_url it is to be resolved against.
 981  *
 982  * In other words, this counts "../" components at the start of a
 983  * submodule URL.
 984  *
 985  * Returns the number of directory components to chop and writes a
 986  * pointer to the next character of url after all leading "./" and
 987  * "../" components to out.
 988  */
 989 static int count_leading_dotdots(const char *url, const char **out)
 990 {
 991         int result = 0;
 992         while (1) {
 993                 if (starts_with_dot_dot_slash(url)) {
 994                         result++;
 995                         url += strlen("../");
 996                         continue;
 997                 }
 998                 if (starts_with_dot_slash(url)) {
 999                         url += strlen("./");
1000                         continue;
1001                 }
1002                 *out = url;
1003                 return result;
1004         }
1005 }
1006 /*
1007  * Check whether a transport is implemented by git-remote-curl.
1008  *
1009  * If it is, returns 1 and writes the URL that would be passed to
1010  * git-remote-curl to the "out" parameter.
1011  *
1012  * Otherwise, returns 0 and leaves "out" untouched.
1013  *
1014  * Examples:
1015  *   http::https://example.com/repo.git -> 1, https://example.com/repo.git
1016  *   https://example.com/repo.git -> 1, https://example.com/repo.git
1017  *   git://example.com/repo.git -> 0
1018  *
1019  * This is for use in checking for previously exploitable bugs that
1020  * required a submodule URL to be passed to git-remote-curl.
1021  */
1022 static int url_to_curl_url(const char *url, const char **out)
1023 {
1024         /*
1025          * We don't need to check for case-aliases, "http.exe", and so
1026          * on because in the default configuration, is_transport_allowed
1027          * prevents URLs with those schemes from being cloned
1028          * automatically.
1029          */
1030         if (skip_prefix(url, "http::", out) ||
1031             skip_prefix(url, "https::", out) ||
1032             skip_prefix(url, "ftp::", out) ||
1033             skip_prefix(url, "ftps::", out))
1034                 return 1;
1035         if (starts_with(url, "http://") ||
1036             starts_with(url, "https://") ||
1037             starts_with(url, "ftp://") ||
1038             starts_with(url, "ftps://")) {
1039                 *out = url;
1040                 return 1;
1041         }
1042         return 0;
1043 }
1044
1045 static int check_submodule_url(const char *url)
1046 {
1047         const char *curl_url;
1048
1049         if (looks_like_command_line_option(url))
1050                 return -1;
1051
1052         if (submodule_url_is_relative(url)) {
1053                 char *decoded;
1054                 const char *next;
1055                 int has_nl;
1056
1057                 /*
1058                  * This could be appended to an http URL and url-decoded;
1059                  * check for malicious characters.
1060                  */
1061                 decoded = url_decode(url);
1062                 has_nl = !!strchr(decoded, '\n');
1063
1064                 free(decoded);
1065                 if (has_nl)
1066                         return -1;
1067
1068                 /*
1069                  * URLs which escape their root via "../" can overwrite
1070                  * the host field and previous components, resolving to
1071                  * URLs like https::example.com/submodule.git and
1072                  * https:///example.com/submodule.git that were
1073                  * susceptible to CVE-2020-11008.
1074                  */
1075                 if (count_leading_dotdots(url, &next) > 0 &&
1076                     (*next == ':' || *next == '/'))
1077                         return -1;
1078         }
1079
1080         else if (url_to_curl_url(url, &curl_url)) {
1081                 struct credential c = CREDENTIAL_INIT;
1082                 int ret = 0;
1083                 if (credential_from_url_gently(&c, curl_url, 1) ||
1084                     !*c.host)
1085                         ret = -1;
1086                 credential_clear(&c);
1087                 return ret;
1088         }
1089
1090         return 0;
1091 }
1092
1093 struct fsck_gitmodules_data {
1094         struct object *obj;
1095         struct fsck_options *options;
1096         int ret;
1097 };
1098
1099 static int fsck_gitmodules_fn(const char *var, const char *value, void *vdata)
1100 {
1101         struct fsck_gitmodules_data *data = vdata;
1102         const char *subsection, *key;
1103         int subsection_len;
1104         char *name;
1105
1106         if (parse_config_key(var, "submodule", &subsection, &subsection_len, &key) < 0 ||
1107             !subsection)
1108                 return 0;
1109
1110         name = xmemdupz(subsection, subsection_len);
1111         if (check_submodule_name(name) < 0)
1112                 data->ret |= report(data->options, data->obj,
1113                                     FSCK_MSG_GITMODULES_NAME,
1114                                     "disallowed submodule name: %s",
1115                                     name);
1116         if (!strcmp(key, "url") && value &&
1117             check_submodule_url(value) < 0)
1118                 data->ret |= report(data->options, data->obj,
1119                                     FSCK_MSG_GITMODULES_URL,
1120                                     "disallowed submodule url: %s",
1121                                     value);
1122         if (!strcmp(key, "path") && value &&
1123             looks_like_command_line_option(value))
1124                 data->ret |= report(data->options, data->obj,
1125                                     FSCK_MSG_GITMODULES_PATH,
1126                                     "disallowed submodule path: %s",
1127                                     value);
1128         if (!strcmp(key, "update") && value &&
1129             parse_submodule_update_type(value) == SM_UPDATE_COMMAND)
1130                 data->ret |= report(data->options, data->obj,
1131                                     FSCK_MSG_GITMODULES_UPDATE,
1132                                     "disallowed submodule update setting: %s",
1133                                     value);
1134         free(name);
1135
1136         return 0;
1137 }
1138
1139 static int fsck_blob(struct blob *blob, const char *buf,
1140                      unsigned long size, struct fsck_options *options)
1141 {
1142         struct fsck_gitmodules_data data;
1143
1144         if (!oidset_contains(&gitmodules_found, &blob->object.oid))
1145                 return 0;
1146         oidset_insert(&gitmodules_done, &blob->object.oid);
1147
1148         if (!buf) {
1149                 /*
1150                  * A missing buffer here is a sign that the caller found the
1151                  * blob too gigantic to load into memory. Let's just consider
1152                  * that an error.
1153                  */
1154                 return report(options, &blob->object,
1155                               FSCK_MSG_GITMODULES_PARSE,
1156                               ".gitmodules too large to parse");
1157         }
1158
1159         data.obj = &blob->object;
1160         data.options = options;
1161         data.ret = 0;
1162         if (git_config_from_mem(fsck_gitmodules_fn, CONFIG_ORIGIN_BLOB,
1163                                 ".gitmodules", buf, size, &data))
1164                 data.ret |= report(options, &blob->object,
1165                                    FSCK_MSG_GITMODULES_PARSE,
1166                                    "could not parse gitmodules blob");
1167
1168         return data.ret;
1169 }
1170
1171 int fsck_object(struct object *obj, void *data, unsigned long size,
1172         struct fsck_options *options)
1173 {
1174         if (!obj)
1175                 return report(options, obj, FSCK_MSG_BAD_OBJECT_SHA1, "no valid object to fsck");
1176
1177         if (obj->type == OBJ_BLOB)
1178                 return fsck_blob((struct blob *)obj, data, size, options);
1179         if (obj->type == OBJ_TREE)
1180                 return fsck_tree((struct tree *) obj, options);
1181         if (obj->type == OBJ_COMMIT)
1182                 return fsck_commit((struct commit *) obj, (const char *) data,
1183                         size, options);
1184         if (obj->type == OBJ_TAG)
1185                 return fsck_tag((struct tag *) obj, (const char *) data,
1186                         size, options);
1187
1188         return report(options, obj, FSCK_MSG_UNKNOWN_TYPE, "unknown type '%d' (internal fsck error)",
1189                           obj->type);
1190 }
1191
1192 int fsck_error_function(struct fsck_options *o,
1193         struct object *obj, int msg_type, const char *message)
1194 {
1195         if (msg_type == FSCK_WARN) {
1196                 warning("object %s: %s", describe_object(o, obj), message);
1197                 return 0;
1198         }
1199         error("object %s: %s", describe_object(o, obj), message);
1200         return 1;
1201 }
1202
1203 int fsck_finish(struct fsck_options *options)
1204 {
1205         int ret = 0;
1206         struct oidset_iter iter;
1207         const struct object_id *oid;
1208
1209         oidset_iter_init(&gitmodules_found, &iter);
1210         while ((oid = oidset_iter_next(&iter))) {
1211                 struct blob *blob;
1212                 enum object_type type;
1213                 unsigned long size;
1214                 char *buf;
1215
1216                 if (oidset_contains(&gitmodules_done, oid))
1217                         continue;
1218
1219                 blob = lookup_blob(oid);
1220                 if (!blob) {
1221                         struct object *obj = lookup_unknown_object(oid->hash);
1222                         ret |= report(options, obj,
1223                                       FSCK_MSG_GITMODULES_BLOB,
1224                                       "non-blob found at .gitmodules");
1225                         continue;
1226                 }
1227
1228                 buf = read_object_file(oid, &type, &size);
1229                 if (!buf) {
1230                         if (is_promisor_object(&blob->object.oid))
1231                                 continue;
1232                         ret |= report(options, &blob->object,
1233                                       FSCK_MSG_GITMODULES_MISSING,
1234                                       "unable to read .gitmodules blob");
1235                         continue;
1236                 }
1237
1238                 if (type == OBJ_BLOB)
1239                         ret |= fsck_blob(blob, buf, size, options);
1240                 else
1241                         ret |= report(options, &blob->object,
1242                                       FSCK_MSG_GITMODULES_BLOB,
1243                                       "non-blob found at .gitmodules");
1244                 free(buf);
1245         }
1246
1247
1248         oidset_clear(&gitmodules_found);
1249         oidset_clear(&gitmodules_done);
1250         return ret;
1251 }