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 "utf8.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                                is_hfs_dotgit(name) ||
 176                                is_ntfs_dotgit(name));
 177                 has_zero_pad |= *(char *)desc.buffer == '0';
 178                 update_tree_entry(&desc);
 179
 180                 switch (mode) {
 181                 /*
 182                  * Standard modes..
 183                  */
 184                 case S_IFREG | 0755:
 185                 case S_IFREG | 0644:
 186                 case S_IFLNK:
 187                 case S_IFDIR:
 188                 case S_IFGITLINK:
 189                         break;
 190                 /*
 191                  * This is nonstandard, but we had a few of these
 192                  * early on when we honored the full set of mode
 193                  * bits..
 194                  */
 195                 case S_IFREG | 0664:
 196                         if (!strict)
 197                                 break;
 198                 default:
 199                         has_bad_modes = 1;
 200                 }
 201
 202                 if (o_name) {
 203                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 204                         case TREE_UNORDERED:
 205                                 not_properly_sorted = 1;
 206                                 break;
 207                         case TREE_HAS_DUPS:
 208                                 has_dup_entries = 1;
 209                                 break;
 210                         default:
 211                                 break;
 212                         }
 213                 }
 214
 215                 o_mode = mode;
 216                 o_name = name;
 217         }
 218
 219         retval = 0;
 220         if (has_null_sha1)
 221                 retval += error_func(&item->object, FSCK_WARN, "contains entries pointing to null sha1");
 222         if (has_full_path)
 223                 retval += error_func(&item->object, FSCK_WARN, "contains full pathnames");
 224         if (has_empty_name)
 225                 retval += error_func(&item->object, FSCK_WARN, "contains empty pathname");
 226         if (has_dot)
 227                 retval += error_func(&item->object, FSCK_WARN, "contains '.'");
 228         if (has_dotdot)
 229                 retval += error_func(&item->object, FSCK_WARN, "contains '..'");
 230         if (has_dotgit)
 231                 retval += error_func(&item->object, FSCK_WARN, "contains '.git'");
 232         if (has_zero_pad)
 233                 retval += error_func(&item->object, FSCK_WARN, "contains zero-padded file modes");
 234         if (has_bad_modes)
 235                 retval += error_func(&item->object, FSCK_WARN, "contains bad file modes");
 236         if (has_dup_entries)
 237                 retval += error_func(&item->object, FSCK_ERROR, "contains duplicate file entries");
 238         if (not_properly_sorted)
 239                 retval += error_func(&item->object, FSCK_ERROR, "not properly sorted");
 240         return retval;
 241 }
 242
 243 static int fsck_ident(const char **ident, struct object *obj, fsck_error error_func)
 244 {
 245         char *end;
 246
 247         if (**ident == '<')
 248                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - missing space before email");
 249         *ident += strcspn(*ident, "<>\n");
 250         if (**ident == '>')
 251                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - bad name");
 252         if (**ident != '<')
 253                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - missing email");
 254         if ((*ident)[-1] != ' ')
 255                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - missing space before email");
 256         (*ident)++;
 257         *ident += strcspn(*ident, "<>\n");
 258         if (**ident != '>')
 259                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - bad email");
 260         (*ident)++;
 261         if (**ident != ' ')
 262                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - missing space before date");
 263         (*ident)++;
 264         if (**ident == '0' && (*ident)[1] != ' ')
 265                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - zero-padded date");
 266         if (date_overflows(strtoul(*ident, &end, 10)))
 267                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - date causes integer overflow");
 268         if (end == *ident || *end != ' ')
 269                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - bad date");
 270         *ident = end + 1;
 271         if ((**ident != '+' && **ident != '-') ||
 272             !isdigit((*ident)[1]) ||
 273             !isdigit((*ident)[2]) ||
 274             !isdigit((*ident)[3]) ||
 275             !isdigit((*ident)[4]) ||
 276             ((*ident)[5] != '\n'))
 277                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - bad time zone");
 278         (*ident) += 6;
 279         return 0;
 280 }
 281
 282 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 283                               fsck_error error_func)
 284 {
 285         const char *tmp;
 286         unsigned char tree_sha1[20], sha1[20];
 287         struct commit_graft *graft;
 288         int parents = 0;
 289         int err;
 290
 291         buffer = skip_prefix(buffer, "tree ");
 292         if (!buffer)
 293                 return error_func(&commit->object, FSCK_ERROR, "invalid format - expected 'tree' line");
 294         if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n')
 295                 return error_func(&commit->object, FSCK_ERROR, "invalid 'tree' line format - bad sha1");
 296         buffer += 41;
 297         while ((tmp = skip_prefix(buffer, "parent "))) {
 298                 buffer = tmp;
 299                 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n')
 300                         return error_func(&commit->object, FSCK_ERROR, "invalid 'parent' line format - bad sha1");
 301                 buffer += 41;
 302                 parents++;
 303         }
 304         graft = lookup_commit_graft(commit->object.sha1);
 305         if (graft) {
 306                 struct commit_list *p = commit->parents;
 307                 parents = 0;
 308                 while (p) {
 309                         p = p->next;
 310                         parents++;
 311                 }
 312                 if (graft->nr_parent == -1 && !parents)
 313                         ; /* shallow commit */
 314                 else if (graft->nr_parent != parents)
 315                         return error_func(&commit->object, FSCK_ERROR, "graft objects missing");
 316         } else {
 317                 struct commit_list *p = commit->parents;
 318                 while (p && parents) {
 319                         p = p->next;
 320                         parents--;
 321                 }
 322                 if (p || parents)
 323                         return error_func(&commit->object, FSCK_ERROR, "parent objects missing");
 324         }
 325         buffer = skip_prefix(buffer, "author ");
 326         if (!buffer)
 327                 return error_func(&commit->object, FSCK_ERROR, "invalid format - expected 'author' line");
 328         err = fsck_ident(&buffer, &commit->object, error_func);
 329         if (err)
 330                 return err;
 331         buffer = skip_prefix(buffer, "committer ");
 332         if (!buffer)
 333                 return error_func(&commit->object, FSCK_ERROR, "invalid format - expected 'committer' line");
 334         err = fsck_ident(&buffer, &commit->object, error_func);
 335         if (err)
 336                 return err;
 337         if (!commit->tree)
 338                 return error_func(&commit->object, FSCK_ERROR, "could not load commit's tree %s", sha1_to_hex(tree_sha1));
 339
 340         return 0;
 341 }
 342
 343 static int fsck_commit(struct commit *commit, fsck_error error_func)
 344 {
 345         const char *buffer = get_commit_buffer(commit, NULL);
 346         int ret = fsck_commit_buffer(commit, buffer, error_func);
 347         unuse_commit_buffer(commit, buffer);
 348         return ret;
 349 }
 350
 351 static int fsck_tag(struct tag *tag, fsck_error error_func)
 352 {
 353         struct object *tagged = tag->tagged;
 354
 355         if (!tagged)
 356                 return error_func(&tag->object, FSCK_ERROR, "could not load tagged object");
 357         return 0;
 358 }
 359
 360 int fsck_object(struct object *obj, int strict, fsck_error error_func)
 361 {
 362         if (!obj)
 363                 return error_func(obj, FSCK_ERROR, "no valid object to fsck");
 364
 365         if (obj->type == OBJ_BLOB)
 366                 return 0;
 367         if (obj->type == OBJ_TREE)
 368                 return fsck_tree((struct tree *) obj, strict, error_func);
 369         if (obj->type == OBJ_COMMIT)
 370                 return fsck_commit((struct commit *) obj, error_func);
 371         if (obj->type == OBJ_TAG)
 372                 return fsck_tag((struct tag *) obj, error_func);
 373
 374         return error_func(obj, FSCK_ERROR, "unknown type '%d' (internal fsck error)",
 375                           obj->type);
 376 }
 377
 378 int fsck_error_function(struct object *obj, int type, const char *fmt, ...)
 379 {
 380         va_list ap;
 381         struct strbuf sb = STRBUF_INIT;
 382
 383         strbuf_addf(&sb, "object %s:", sha1_to_hex(obj->sha1));
 384
 385         va_start(ap, fmt);
 386         strbuf_vaddf(&sb, fmt, ap);
 387         va_end(ap);
 388
 389         error("%s", sb.buf);
 390         strbuf_release(&sb);
 391         return 1;
 392 }