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
  11 static int fsck_walk_tree(struct tree *tree, fsck_walk_func walk, void *data)
  12 {
  13         struct tree_desc desc;
  14         struct name_entry entry;
  15         int res = 0;
  16
  17         if (parse_tree(tree))
  18                 return -1;
  19
  20         init_tree_desc(&desc, tree->buffer, tree->size);
  21         while (tree_entry(&desc, &entry)) {
  22                 int result;
  23
  24                 if (S_ISGITLINK(entry.mode))
  25                         continue;
  26                 if (S_ISDIR(entry.mode))
  27                         result = walk(&lookup_tree(entry.sha1)->object, OBJ_TREE, data);
  28                 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode))
  29                         result = walk(&lookup_blob(entry.sha1)->object, OBJ_BLOB, data);
  30                 else {
  31                         result = error("in tree %s: entry %s has bad mode %.6o",
  32                                         sha1_to_hex(tree->object.sha1), entry.path, entry.mode);
  33                 }
  34                 if (result < 0)
  35                         return result;
  36                 if (!res)
  37                         res = result;
  38         }
  39         return res;
  40 }
  41
  42 static int fsck_walk_commit(struct commit *commit, fsck_walk_func walk, void *data)
  43 {
  44         struct commit_list *parents;
  45         int res;
  46         int result;
  47
  48         if (parse_commit(commit))
  49                 return -1;
  50
  51         result = walk((struct object *)commit->tree, OBJ_TREE, data);
  52         if (result < 0)
  53                 return result;
  54         res = result;
  55
  56         parents = commit->parents;
  57         while (parents) {
  58                 result = walk((struct object *)parents->item, OBJ_COMMIT, data);
  59                 if (result < 0)
  60                         return result;
  61                 if (!res)
  62                         res = result;
  63                 parents = parents->next;
  64         }
  65         return res;
  66 }
  67
  68 static int fsck_walk_tag(struct tag *tag, fsck_walk_func walk, void *data)
  69 {
  70         if (parse_tag(tag))
  71                 return -1;
  72         return walk(tag->tagged, OBJ_ANY, data);
  73 }
  74
  75 int fsck_walk(struct object *obj, fsck_walk_func walk, void *data)
  76 {
  77         if (!obj)
  78                 return -1;
  79         switch (obj->type) {
  80         case OBJ_BLOB:
  81                 return 0;
  82         case OBJ_TREE:
  83                 return fsck_walk_tree((struct tree *)obj, walk, data);
  84         case OBJ_COMMIT:
  85                 return fsck_walk_commit((struct commit *)obj, walk, data);
  86         case OBJ_TAG:
  87                 return fsck_walk_tag((struct tag *)obj, walk, data);
  88         default:
  89                 error("Unknown object type for %s", sha1_to_hex(obj->sha1));
  90                 return -1;
  91         }
  92 }
  93
  94 /*
  95  * The entries in a tree are ordered in the _path_ order,
  96  * which means that a directory entry is ordered by adding
  97  * a slash to the end of it.
  98  *
  99  * So a directory called "a" is ordered _after_ a file
 100  * called "a.c", because "a/" sorts after "a.c".
 101  */
 102 #define TREE_UNORDERED (-1)
 103 #define TREE_HAS_DUPS  (-2)
 104
 105 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 106 {
 107         int len1 = strlen(name1);
 108         int len2 = strlen(name2);
 109         int len = len1 < len2 ? len1 : len2;
 110         unsigned char c1, c2;
 111         int cmp;
 112
 113         cmp = memcmp(name1, name2, len);
 114         if (cmp < 0)
 115                 return 0;
 116         if (cmp > 0)
 117                 return TREE_UNORDERED;
 118
 119         /*
 120          * Ok, the first <len> characters are the same.
 121          * Now we need to order the next one, but turn
 122          * a '\0' into a '/' for a directory entry.
 123          */
 124         c1 = name1[len];
 125         c2 = name2[len];
 126         if (!c1 && !c2)
 127                 /*
 128                  * git-write-tree used to write out a nonsense tree that has
 129                  * entries with the same name, one blob and one tree.  Make
 130                  * sure we do not have duplicate entries.
 131                  */
 132                 return TREE_HAS_DUPS;
 133         if (!c1 && S_ISDIR(mode1))
 134                 c1 = '/';
 135         if (!c2 && S_ISDIR(mode2))
 136                 c2 = '/';
 137         return c1 < c2 ? 0 : TREE_UNORDERED;
 138 }
 139
 140 static int fsck_tree(struct tree *item, int strict, fsck_error error_func)
 141 {
 142         int retval;
 143         int has_null_sha1 = 0;
 144         int has_full_path = 0;
 145         int has_empty_name = 0;
 146         int has_dot = 0;
 147         int has_dotdot = 0;
 148         int has_dotgit = 0;
 149         int has_zero_pad = 0;
 150         int has_bad_modes = 0;
 151         int has_dup_entries = 0;
 152         int not_properly_sorted = 0;
 153         struct tree_desc desc;
 154         unsigned o_mode;
 155         const char *o_name;
 156
 157         init_tree_desc(&desc, item->buffer, item->size);
 158
 159         o_mode = 0;
 160         o_name = NULL;
 161
 162         while (desc.size) {
 163                 unsigned mode;
 164                 const char *name;
 165                 const unsigned char *sha1;
 166
 167                 sha1 = tree_entry_extract(&desc, &name, &mode);
 168
 169                 has_null_sha1 |= is_null_sha1(sha1);
 170                 has_full_path |= !!strchr(name, '/');
 171                 has_empty_name |= !*name;
 172                 has_dot |= !strcmp(name, ".");
 173                 has_dotdot |= !strcmp(name, "..");
 174                 has_dotgit |= !strcmp(name, ".git");
 175                 has_zero_pad |= *(char *)desc.buffer == '0';
 176                 update_tree_entry(&desc);
 177
 178                 switch (mode) {
 179                 /*
 180                  * Standard modes..
 181                  */
 182                 case S_IFREG | 0755:
 183                 case S_IFREG | 0644:
 184                 case S_IFLNK:
 185                 case S_IFDIR:
 186                 case S_IFGITLINK:
 187                         break;
 188                 /*
 189                  * This is nonstandard, but we had a few of these
 190                  * early on when we honored the full set of mode
 191                  * bits..
 192                  */
 193                 case S_IFREG | 0664:
 194                         if (!strict)
 195                                 break;
 196                 default:
 197                         has_bad_modes = 1;
 198                 }
 199
 200                 if (o_name) {
 201                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 202                         case TREE_UNORDERED:
 203                                 not_properly_sorted = 1;
 204                                 break;
 205                         case TREE_HAS_DUPS:
 206                                 has_dup_entries = 1;
 207                                 break;
 208                         default:
 209                                 break;
 210                         }
 211                 }
 212
 213                 o_mode = mode;
 214                 o_name = name;
 215         }
 216
 217         retval = 0;
 218         if (has_null_sha1)
 219                 retval += error_func(&item->object, FSCK_WARN, "contains entries pointing to null sha1");
 220         if (has_full_path)
 221                 retval += error_func(&item->object, FSCK_WARN, "contains full pathnames");
 222         if (has_empty_name)
 223                 retval += error_func(&item->object, FSCK_WARN, "contains empty pathname");
 224         if (has_dot)
 225                 retval += error_func(&item->object, FSCK_WARN, "contains '.'");
 226         if (has_dotdot)
 227                 retval += error_func(&item->object, FSCK_WARN, "contains '..'");
 228         if (has_dotgit)
 229                 retval += error_func(&item->object, FSCK_WARN, "contains '.git'");
 230         if (has_zero_pad)
 231                 retval += error_func(&item->object, FSCK_WARN, "contains zero-padded file modes");
 232         if (has_bad_modes)
 233                 retval += error_func(&item->object, FSCK_WARN, "contains bad file modes");
 234         if (has_dup_entries)
 235                 retval += error_func(&item->object, FSCK_ERROR, "contains duplicate file entries");
 236         if (not_properly_sorted)
 237                 retval += error_func(&item->object, FSCK_ERROR, "not properly sorted");
 238         return retval;
 239 }
 240
 241 static int verify_headers(const void *data, unsigned long size,
 242                           struct object *obj, fsck_error error_func)
 243 {
 244         const char *buffer = (const char *)data;
 245         unsigned long i;
 246
 247         for (i = 0; i < size; i++) {
 248                 switch (buffer[i]) {
 249                 case '\0':
 250                         return error_func(obj, FSCK_ERROR,
 251                                 "unterminated header: NUL at offset %d", i);
 252                 case '\n':
 253                         if (i + 1 < size && buffer[i + 1] == '\n')
 254                                 return 0;
 255                 }
 256         }
 257
 258         /*
 259          * We did not find double-LF that separates the header
 260          * and the body.  Not having a body is not a crime but
 261          * we do want to see the terminating LF for the last header
 262          * line.
 263          */
 264         if (size && buffer[size - 1] == '\n')
 265                 return 0;
 266
 267         return error_func(obj, FSCK_ERROR, "unterminated header");
 268 }
 269
 270 static int fsck_ident(const char **ident, struct object *obj, fsck_error error_func)
 271 {
 272         char *end;
 273
 274         if (**ident == '<')
 275                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - missing space before email");
 276         *ident += strcspn(*ident, "<>\n");
 277         if (**ident == '>')
 278                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - bad name");
 279         if (**ident != '<')
 280                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - missing email");
 281         if ((*ident)[-1] != ' ')
 282                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - missing space before email");
 283         (*ident)++;
 284         *ident += strcspn(*ident, "<>\n");
 285         if (**ident != '>')
 286                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - bad email");
 287         (*ident)++;
 288         if (**ident != ' ')
 289                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - missing space before date");
 290         (*ident)++;
 291         if (**ident == '0' && (*ident)[1] != ' ')
 292                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - zero-padded date");
 293         if (date_overflows(strtoul(*ident, &end, 10)))
 294                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - date causes integer overflow");
 295         if (end == *ident || *end != ' ')
 296                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - bad date");
 297         *ident = end + 1;
 298         if ((**ident != '+' && **ident != '-') ||
 299             !isdigit((*ident)[1]) ||
 300             !isdigit((*ident)[2]) ||
 301             !isdigit((*ident)[3]) ||
 302             !isdigit((*ident)[4]) ||
 303             ((*ident)[5] != '\n'))
 304                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - bad time zone");
 305         (*ident) += 6;
 306         return 0;
 307 }
 308
 309 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 310         unsigned long size, fsck_error error_func)
 311 {
 312         unsigned char tree_sha1[20], sha1[20];
 313         struct commit_graft *graft;
 314         unsigned parent_count, parent_line_count = 0;
 315         int err;
 316
 317         if (verify_headers(buffer, size, &commit->object, error_func))
 318                 return -1;
 319
 320         if (!skip_prefix(buffer, "tree ", &buffer))
 321                 return error_func(&commit->object, FSCK_ERROR, "invalid format - expected 'tree' line");
 322         if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n')
 323                 return error_func(&commit->object, FSCK_ERROR, "invalid 'tree' line format - bad sha1");
 324         buffer += 41;
 325         while (skip_prefix(buffer, "parent ", &buffer)) {
 326                 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n')
 327                         return error_func(&commit->object, FSCK_ERROR, "invalid 'parent' line format - bad sha1");
 328                 buffer += 41;
 329                 parent_line_count++;
 330         }
 331         graft = lookup_commit_graft(commit->object.sha1);
 332         parent_count = commit_list_count(commit->parents);
 333         if (graft) {
 334                 if (graft->nr_parent == -1 && !parent_count)
 335                         ; /* shallow commit */
 336                 else if (graft->nr_parent != parent_count)
 337                         return error_func(&commit->object, FSCK_ERROR, "graft objects missing");
 338         } else {
 339                 if (parent_count != parent_line_count)
 340                         return error_func(&commit->object, FSCK_ERROR, "parent objects missing");
 341         }
 342         if (!skip_prefix(buffer, "author ", &buffer))
 343                 return error_func(&commit->object, FSCK_ERROR, "invalid format - expected 'author' line");
 344         err = fsck_ident(&buffer, &commit->object, error_func);
 345         if (err)
 346                 return err;
 347         if (!skip_prefix(buffer, "committer ", &buffer))
 348                 return error_func(&commit->object, FSCK_ERROR, "invalid format - expected 'committer' line");
 349         err = fsck_ident(&buffer, &commit->object, error_func);
 350         if (err)
 351                 return err;
 352         if (!commit->tree)
 353                 return error_func(&commit->object, FSCK_ERROR, "could not load commit's tree %s", sha1_to_hex(tree_sha1));
 354
 355         return 0;
 356 }
 357
 358 static int fsck_commit(struct commit *commit, const char *data,
 359         unsigned long size, fsck_error error_func)
 360 {
 361         const char *buffer = data ?  data : get_commit_buffer(commit, &size);
 362         int ret = fsck_commit_buffer(commit, buffer, size, error_func);
 363         if (!data)
 364                 unuse_commit_buffer(commit, buffer);
 365         return ret;
 366 }
 367
 368 static int fsck_tag_buffer(struct tag *tag, const char *data,
 369         unsigned long size, fsck_error error_func)
 370 {
 371         unsigned char sha1[20];
 372         int ret = 0;
 373         const char *buffer;
 374         char *to_free = NULL, *eol;
 375         struct strbuf sb = STRBUF_INIT;
 376
 377         if (data)
 378                 buffer = data;
 379         else {
 380                 enum object_type type;
 381
 382                 buffer = to_free =
 383                         read_sha1_file(tag->object.sha1, &type, &size);
 384                 if (!buffer)
 385                         return error_func(&tag->object, FSCK_ERROR,
 386                                 "cannot read tag object");
 387
 388                 if (type != OBJ_TAG) {
 389                         ret = error_func(&tag->object, FSCK_ERROR,
 390                                 "expected tag got %s",
 391                             typename(type));
 392                         goto done;
 393                 }
 394         }
 395
 396         if (verify_headers(buffer, size, &tag->object, error_func))
 397                 goto done;
 398
 399         if (!skip_prefix(buffer, "object ", &buffer)) {
 400                 ret = error_func(&tag->object, FSCK_ERROR, "invalid format - expected 'object' line");
 401                 goto done;
 402         }
 403         if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n') {
 404                 ret = error_func(&tag->object, FSCK_ERROR, "invalid 'object' line format - bad sha1");
 405                 goto done;
 406         }
 407         buffer += 41;
 408
 409         if (!skip_prefix(buffer, "type ", &buffer)) {
 410                 ret = error_func(&tag->object, FSCK_ERROR, "invalid format - expected 'type' line");
 411                 goto done;
 412         }
 413         eol = strchr(buffer, '\n');
 414         if (!eol) {
 415                 ret = error_func(&tag->object, FSCK_ERROR, "invalid format - unexpected end after 'type' line");
 416                 goto done;
 417         }
 418         if (type_from_string_gently(buffer, eol - buffer, 1) < 0)
 419                 ret = error_func(&tag->object, FSCK_ERROR, "invalid 'type' value");
 420         if (ret)
 421                 goto done;
 422         buffer = eol + 1;
 423
 424         if (!skip_prefix(buffer, "tag ", &buffer)) {
 425                 ret = error_func(&tag->object, FSCK_ERROR, "invalid format - expected 'tag' line");
 426                 goto done;
 427         }
 428         eol = strchr(buffer, '\n');
 429         if (!eol) {
 430                 ret = error_func(&tag->object, FSCK_ERROR, "invalid format - unexpected end after 'type' line");
 431                 goto done;
 432         }
 433         strbuf_addf(&sb, "refs/tags/%.*s", (int)(eol - buffer), buffer);
 434         if (check_refname_format(sb.buf, 0))
 435                 error_func(&tag->object, FSCK_WARN, "invalid 'tag' name: %.*s",
 436                            (int)(eol - buffer), buffer);
 437         buffer = eol + 1;
 438
 439         if (!skip_prefix(buffer, "tagger ", &buffer))
 440                 /* early tags do not contain 'tagger' lines; warn only */
 441                 error_func(&tag->object, FSCK_WARN, "invalid format - expected 'tagger' line");
 442         else
 443                 ret = fsck_ident(&buffer, &tag->object, error_func);
 444
 445 done:
 446         strbuf_release(&sb);
 447         free(to_free);
 448         return ret;
 449 }
 450
 451 static int fsck_tag(struct tag *tag, const char *data,
 452         unsigned long size, fsck_error error_func)
 453 {
 454         struct object *tagged = tag->tagged;
 455
 456         if (!tagged)
 457                 return error_func(&tag->object, FSCK_ERROR, "could not load tagged object");
 458
 459         return fsck_tag_buffer(tag, data, size, error_func);
 460 }
 461
 462 int fsck_object(struct object *obj, void *data, unsigned long size,
 463         int strict, fsck_error error_func)
 464 {
 465         if (!obj)
 466                 return error_func(obj, FSCK_ERROR, "no valid object to fsck");
 467
 468         if (obj->type == OBJ_BLOB)
 469                 return 0;
 470         if (obj->type == OBJ_TREE)
 471                 return fsck_tree((struct tree *) obj, strict, error_func);
 472         if (obj->type == OBJ_COMMIT)
 473                 return fsck_commit((struct commit *) obj, (const char *) data,
 474                         size, error_func);
 475         if (obj->type == OBJ_TAG)
 476                 return fsck_tag((struct tag *) obj, (const char *) data,
 477                         size, error_func);
 478
 479         return error_func(obj, FSCK_ERROR, "unknown type '%d' (internal fsck error)",
 480                           obj->type);
 481 }
 482
 483 int fsck_error_function(struct object *obj, int type, const char *fmt, ...)
 484 {
 485         va_list ap;
 486         struct strbuf sb = STRBUF_INIT;
 487
 488         strbuf_addf(&sb, "object %s:", sha1_to_hex(obj->sha1));
 489
 490         va_start(ap, fmt);
 491         strbuf_vaddf(&sb, fmt, ap);
 492         va_end(ap);
 493
 494         error("%s", sb.buf);
 495         strbuf_release(&sb);
 496         return 1;
 497 }