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
  10 static int fsck_walk_tree(struct tree *tree, fsck_walk_func walk, void *data)
  11 {
  12         struct tree_desc desc;
  13         struct name_entry entry;
  14         int res = 0;
  15
  16         if (parse_tree(tree))
  17                 return -1;
  18
  19         init_tree_desc(&desc, tree->buffer, tree->size);
  20         while (tree_entry(&desc, &entry)) {
  21                 int result;
  22
  23                 if (S_ISGITLINK(entry.mode))
  24                         continue;
  25                 if (S_ISDIR(entry.mode))
  26                         result = walk(&lookup_tree(entry.sha1)->object, OBJ_TREE, data);
  27                 else if (S_ISREG(entry.mode) || S_ISLNK(entry.mode))
  28                         result = walk(&lookup_blob(entry.sha1)->object, OBJ_BLOB, data);
  29                 else {
  30                         result = error("in tree %s: entry %s has bad mode %.6o",
  31                                         sha1_to_hex(tree->object.sha1), entry.path, entry.mode);
  32                 }
  33                 if (result < 0)
  34                         return result;
  35                 if (!res)
  36                         res = result;
  37         }
  38         return res;
  39 }
  40
  41 static int fsck_walk_commit(struct commit *commit, fsck_walk_func walk, void *data)
  42 {
  43         struct commit_list *parents;
  44         int res;
  45         int result;
  46
  47         if (parse_commit(commit))
  48                 return -1;
  49
  50         result = walk((struct object *)commit->tree, OBJ_TREE, data);
  51         if (result < 0)
  52                 return result;
  53         res = result;
  54
  55         parents = commit->parents;
  56         while (parents) {
  57                 result = walk((struct object *)parents->item, OBJ_COMMIT, data);
  58                 if (result < 0)
  59                         return result;
  60                 if (!res)
  61                         res = result;
  62                 parents = parents->next;
  63         }
  64         return res;
  65 }
  66
  67 static int fsck_walk_tag(struct tag *tag, fsck_walk_func walk, void *data)
  68 {
  69         if (parse_tag(tag))
  70                 return -1;
  71         return walk(tag->tagged, OBJ_ANY, data);
  72 }
  73
  74 int fsck_walk(struct object *obj, fsck_walk_func walk, void *data)
  75 {
  76         if (!obj)
  77                 return -1;
  78         switch (obj->type) {
  79         case OBJ_BLOB:
  80                 return 0;
  81         case OBJ_TREE:
  82                 return fsck_walk_tree((struct tree *)obj, walk, data);
  83         case OBJ_COMMIT:
  84                 return fsck_walk_commit((struct commit *)obj, walk, data);
  85         case OBJ_TAG:
  86                 return fsck_walk_tag((struct tag *)obj, walk, data);
  87         default:
  88                 error("Unknown object type for %s", sha1_to_hex(obj->sha1));
  89                 return -1;
  90         }
  91 }
  92
  93 /*
  94  * The entries in a tree are ordered in the _path_ order,
  95  * which means that a directory entry is ordered by adding
  96  * a slash to the end of it.
  97  *
  98  * So a directory called "a" is ordered _after_ a file
  99  * called "a.c", because "a/" sorts after "a.c".
 100  */
 101 #define TREE_UNORDERED (-1)
 102 #define TREE_HAS_DUPS  (-2)
 103
 104 static int verify_ordered(unsigned mode1, const char *name1, unsigned mode2, const char *name2)
 105 {
 106         int len1 = strlen(name1);
 107         int len2 = strlen(name2);
 108         int len = len1 < len2 ? len1 : len2;
 109         unsigned char c1, c2;
 110         int cmp;
 111
 112         cmp = memcmp(name1, name2, len);
 113         if (cmp < 0)
 114                 return 0;
 115         if (cmp > 0)
 116                 return TREE_UNORDERED;
 117
 118         /*
 119          * Ok, the first <len> characters are the same.
 120          * Now we need to order the next one, but turn
 121          * a '\0' into a '/' for a directory entry.
 122          */
 123         c1 = name1[len];
 124         c2 = name2[len];
 125         if (!c1 && !c2)
 126                 /*
 127                  * git-write-tree used to write out a nonsense tree that has
 128                  * entries with the same name, one blob and one tree.  Make
 129                  * sure we do not have duplicate entries.
 130                  */
 131                 return TREE_HAS_DUPS;
 132         if (!c1 && S_ISDIR(mode1))
 133                 c1 = '/';
 134         if (!c2 && S_ISDIR(mode2))
 135                 c2 = '/';
 136         return c1 < c2 ? 0 : TREE_UNORDERED;
 137 }
 138
 139 static int fsck_tree(struct tree *item, int strict, fsck_error error_func)
 140 {
 141         int retval;
 142         int has_null_sha1 = 0;
 143         int has_full_path = 0;
 144         int has_empty_name = 0;
 145         int has_dot = 0;
 146         int has_dotdot = 0;
 147         int has_dotgit = 0;
 148         int has_zero_pad = 0;
 149         int has_bad_modes = 0;
 150         int has_dup_entries = 0;
 151         int not_properly_sorted = 0;
 152         struct tree_desc desc;
 153         unsigned o_mode;
 154         const char *o_name;
 155
 156         init_tree_desc(&desc, item->buffer, item->size);
 157
 158         o_mode = 0;
 159         o_name = NULL;
 160
 161         while (desc.size) {
 162                 unsigned mode;
 163                 const char *name;
 164                 const unsigned char *sha1;
 165
 166                 sha1 = tree_entry_extract(&desc, &name, &mode);
 167
 168                 has_null_sha1 |= is_null_sha1(sha1);
 169                 has_full_path |= !!strchr(name, '/');
 170                 has_empty_name |= !*name;
 171                 has_dot |= !strcmp(name, ".");
 172                 has_dotdot |= !strcmp(name, "..");
 173                 has_dotgit |= !strcmp(name, ".git");
 174                 has_zero_pad |= *(char *)desc.buffer == '0';
 175                 update_tree_entry(&desc);
 176
 177                 switch (mode) {
 178                 /*
 179                  * Standard modes..
 180                  */
 181                 case S_IFREG | 0755:
 182                 case S_IFREG | 0644:
 183                 case S_IFLNK:
 184                 case S_IFDIR:
 185                 case S_IFGITLINK:
 186                         break;
 187                 /*
 188                  * This is nonstandard, but we had a few of these
 189                  * early on when we honored the full set of mode
 190                  * bits..
 191                  */
 192                 case S_IFREG | 0664:
 193                         if (!strict)
 194                                 break;
 195                 default:
 196                         has_bad_modes = 1;
 197                 }
 198
 199                 if (o_name) {
 200                         switch (verify_ordered(o_mode, o_name, mode, name)) {
 201                         case TREE_UNORDERED:
 202                                 not_properly_sorted = 1;
 203                                 break;
 204                         case TREE_HAS_DUPS:
 205                                 has_dup_entries = 1;
 206                                 break;
 207                         default:
 208                                 break;
 209                         }
 210                 }
 211
 212                 o_mode = mode;
 213                 o_name = name;
 214         }
 215
 216         retval = 0;
 217         if (has_null_sha1)
 218                 retval += error_func(&item->object, FSCK_WARN, "contains entries pointing to null sha1");
 219         if (has_full_path)
 220                 retval += error_func(&item->object, FSCK_WARN, "contains full pathnames");
 221         if (has_empty_name)
 222                 retval += error_func(&item->object, FSCK_WARN, "contains empty pathname");
 223         if (has_dot)
 224                 retval += error_func(&item->object, FSCK_WARN, "contains '.'");
 225         if (has_dotdot)
 226                 retval += error_func(&item->object, FSCK_WARN, "contains '..'");
 227         if (has_dotgit)
 228                 retval += error_func(&item->object, FSCK_WARN, "contains '.git'");
 229         if (has_zero_pad)
 230                 retval += error_func(&item->object, FSCK_WARN, "contains zero-padded file modes");
 231         if (has_bad_modes)
 232                 retval += error_func(&item->object, FSCK_WARN, "contains bad file modes");
 233         if (has_dup_entries)
 234                 retval += error_func(&item->object, FSCK_ERROR, "contains duplicate file entries");
 235         if (not_properly_sorted)
 236                 retval += error_func(&item->object, FSCK_ERROR, "not properly sorted");
 237         return retval;
 238 }
 239
 240 static int fsck_ident(const char **ident, struct object *obj, fsck_error error_func)
 241 {
 242         char *end;
 243
 244         if (**ident == '<')
 245                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - missing space before email");
 246         *ident += strcspn(*ident, "<>\n");
 247         if (**ident == '>')
 248                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - bad name");
 249         if (**ident != '<')
 250                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - missing email");
 251         if ((*ident)[-1] != ' ')
 252                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - missing space before email");
 253         (*ident)++;
 254         *ident += strcspn(*ident, "<>\n");
 255         if (**ident != '>')
 256                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - bad email");
 257         (*ident)++;
 258         if (**ident != ' ')
 259                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - missing space before date");
 260         (*ident)++;
 261         if (**ident == '0' && (*ident)[1] != ' ')
 262                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - zero-padded date");
 263         if (date_overflows(strtoul(*ident, &end, 10)))
 264                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - date causes integer overflow");
 265         if (end == *ident || *end != ' ')
 266                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - bad date");
 267         *ident = end + 1;
 268         if ((**ident != '+' && **ident != '-') ||
 269             !isdigit((*ident)[1]) ||
 270             !isdigit((*ident)[2]) ||
 271             !isdigit((*ident)[3]) ||
 272             !isdigit((*ident)[4]) ||
 273             ((*ident)[5] != '\n'))
 274                 return error_func(obj, FSCK_ERROR, "invalid author/committer line - bad time zone");
 275         (*ident) += 6;
 276         return 0;
 277 }
 278
 279 static int fsck_commit_buffer(struct commit *commit, const char *buffer,
 280                               fsck_error error_func)
 281 {
 282         unsigned char tree_sha1[20], sha1[20];
 283         struct commit_graft *graft;
 284         unsigned parent_count, parent_line_count = 0;
 285         int err;
 286
 287         if (!skip_prefix(buffer, "tree ", &buffer))
 288                 return error_func(&commit->object, FSCK_ERROR, "invalid format - expected 'tree' line");
 289         if (get_sha1_hex(buffer, tree_sha1) || buffer[40] != '\n')
 290                 return error_func(&commit->object, FSCK_ERROR, "invalid 'tree' line format - bad sha1");
 291         buffer += 41;
 292         while (skip_prefix(buffer, "parent ", &buffer)) {
 293                 if (get_sha1_hex(buffer, sha1) || buffer[40] != '\n')
 294                         return error_func(&commit->object, FSCK_ERROR, "invalid 'parent' line format - bad sha1");
 295                 buffer += 41;
 296                 parent_line_count++;
 297         }
 298         graft = lookup_commit_graft(commit->object.sha1);
 299         parent_count = commit_list_count(commit->parents);
 300         if (graft) {
 301                 if (graft->nr_parent == -1 && !parent_count)
 302                         ; /* shallow commit */
 303                 else if (graft->nr_parent != parent_count)
 304                         return error_func(&commit->object, FSCK_ERROR, "graft objects missing");
 305         } else {
 306                 if (parent_count != parent_line_count)
 307                         return error_func(&commit->object, FSCK_ERROR, "parent objects missing");
 308         }
 309         if (!skip_prefix(buffer, "author ", &buffer))
 310                 return error_func(&commit->object, FSCK_ERROR, "invalid format - expected 'author' line");
 311         err = fsck_ident(&buffer, &commit->object, error_func);
 312         if (err)
 313                 return err;
 314         if (!skip_prefix(buffer, "committer ", &buffer))
 315                 return error_func(&commit->object, FSCK_ERROR, "invalid format - expected 'committer' line");
 316         err = fsck_ident(&buffer, &commit->object, error_func);
 317         if (err)
 318                 return err;
 319         if (!commit->tree)
 320                 return error_func(&commit->object, FSCK_ERROR, "could not load commit's tree %s", sha1_to_hex(tree_sha1));
 321
 322         return 0;
 323 }
 324
 325 static int fsck_commit(struct commit *commit, fsck_error error_func)
 326 {
 327         const char *buffer = get_commit_buffer(commit, NULL);
 328         int ret = fsck_commit_buffer(commit, buffer, error_func);
 329         unuse_commit_buffer(commit, buffer);
 330         return ret;
 331 }
 332
 333 static int fsck_tag(struct tag *tag, fsck_error error_func)
 334 {
 335         struct object *tagged = tag->tagged;
 336
 337         if (!tagged)
 338                 return error_func(&tag->object, FSCK_ERROR, "could not load tagged object");
 339         return 0;
 340 }
 341
 342 int fsck_object(struct object *obj, int strict, fsck_error error_func)
 343 {
 344         if (!obj)
 345                 return error_func(obj, FSCK_ERROR, "no valid object to fsck");
 346
 347         if (obj->type == OBJ_BLOB)
 348                 return 0;
 349         if (obj->type == OBJ_TREE)
 350                 return fsck_tree((struct tree *) obj, strict, error_func);
 351         if (obj->type == OBJ_COMMIT)
 352                 return fsck_commit((struct commit *) obj, error_func);
 353         if (obj->type == OBJ_TAG)
 354                 return fsck_tag((struct tag *) obj, error_func);
 355
 356         return error_func(obj, FSCK_ERROR, "unknown type '%d' (internal fsck error)",
 357                           obj->type);
 358 }
 359
 360 int fsck_error_function(struct object *obj, int type, const char *fmt, ...)
 361 {
 362         va_list ap;
 363         struct strbuf sb = STRBUF_INIT;
 364
 365         strbuf_addf(&sb, "object %s:", sha1_to_hex(obj->sha1));
 366
 367         va_start(ap, fmt);
 368         strbuf_vaddf(&sb, fmt, ap);
 369         va_end(ap);
 370
 371         error("%s", sb.buf);
 372         strbuf_release(&sb);
 373         return 1;
 374 }