fs/reiserfs/stree.c

   1 /*
   2  *  Copyright 2000 by Hans Reiser, licensing governed by reiserfs/README
   3  */
   4
   5 /*
   6  *  Written by Anatoly P. Pinchuk pap@namesys.botik.ru
   7  *  Programm System Institute
   8  *  Pereslavl-Zalessky Russia
   9  */
  10
  11 #include <linux/time.h>
  12 #include <linux/string.h>
  13 #include <linux/pagemap.h>
  14 #include <linux/bio.h>
  15 #include "reiserfs.h"
  16 #include <linux/buffer_head.h>
  17 #include <linux/quotaops.h>
  18
  19 /* Does the buffer contain a disk block which is in the tree. */
  20 inline int B_IS_IN_TREE(const struct buffer_head *bh)
  21 {
  22
  23         RFALSE(B_LEVEL(bh) > MAX_HEIGHT,
  24                "PAP-1010: block (%b) has too big level (%z)", bh, bh);
  25
  26         return (B_LEVEL(bh) != FREE_LEVEL);
  27 }
  28
  29 /* to get item head in le form */
  30 inline void copy_item_head(struct item_head *to,
  31                            const struct item_head *from)
  32 {
  33         memcpy(to, from, IH_SIZE);
  34 }
  35
  36 /*
  37  * k1 is pointer to on-disk structure which is stored in little-endian
  38  * form. k2 is pointer to cpu variable. For key of items of the same
  39  * object this returns 0.
  40  * Returns: -1 if key1 < key2
  41  * 0 if key1 == key2
  42  * 1 if key1 > key2
  43  */
  44 inline int comp_short_keys(const struct reiserfs_key *le_key,
  45                            const struct cpu_key *cpu_key)
  46 {
  47         __u32 n;
  48         n = le32_to_cpu(le_key->k_dir_id);
  49         if (n < cpu_key->on_disk_key.k_dir_id)
  50                 return -1;
  51         if (n > cpu_key->on_disk_key.k_dir_id)
  52                 return 1;
  53         n = le32_to_cpu(le_key->k_objectid);
  54         if (n < cpu_key->on_disk_key.k_objectid)
  55                 return -1;
  56         if (n > cpu_key->on_disk_key.k_objectid)
  57                 return 1;
  58         return 0;
  59 }
  60
  61 /*
  62  * k1 is pointer to on-disk structure which is stored in little-endian
  63  * form. k2 is pointer to cpu variable.
  64  * Compare keys using all 4 key fields.
  65  * Returns: -1 if key1 < key2 0
  66  * if key1 = key2 1 if key1 > key2
  67  */
  68 static inline int comp_keys(const struct reiserfs_key *le_key,
  69                             const struct cpu_key *cpu_key)
  70 {
  71         int retval;
  72
  73         retval = comp_short_keys(le_key, cpu_key);
  74         if (retval)
  75                 return retval;
  76         if (le_key_k_offset(le_key_version(le_key), le_key) <
  77             cpu_key_k_offset(cpu_key))
  78                 return -1;
  79         if (le_key_k_offset(le_key_version(le_key), le_key) >
  80             cpu_key_k_offset(cpu_key))
  81                 return 1;
  82
  83         if (cpu_key->key_length == 3)
  84                 return 0;
  85
  86         /* this part is needed only when tail conversion is in progress */
  87         if (le_key_k_type(le_key_version(le_key), le_key) <
  88             cpu_key_k_type(cpu_key))
  89                 return -1;
  90
  91         if (le_key_k_type(le_key_version(le_key), le_key) >
  92             cpu_key_k_type(cpu_key))
  93                 return 1;
  94
  95         return 0;
  96 }
  97
  98 inline int comp_short_le_keys(const struct reiserfs_key *key1,
  99                               const struct reiserfs_key *key2)
 100 {
 101         __u32 *k1_u32, *k2_u32;
 102         int key_length = REISERFS_SHORT_KEY_LEN;
 103
 104         k1_u32 = (__u32 *) key1;
 105         k2_u32 = (__u32 *) key2;
 106         for (; key_length--; ++k1_u32, ++k2_u32) {
 107                 if (le32_to_cpu(*k1_u32) < le32_to_cpu(*k2_u32))
 108                         return -1;
 109                 if (le32_to_cpu(*k1_u32) > le32_to_cpu(*k2_u32))
 110                         return 1;
 111         }
 112         return 0;
 113 }
 114
 115 inline void le_key2cpu_key(struct cpu_key *to, const struct reiserfs_key *from)
 116 {
 117         int version;
 118         to->on_disk_key.k_dir_id = le32_to_cpu(from->k_dir_id);
 119         to->on_disk_key.k_objectid = le32_to_cpu(from->k_objectid);
 120
 121         /* find out version of the key */
 122         version = le_key_version(from);
 123         to->version = version;
 124         to->on_disk_key.k_offset = le_key_k_offset(version, from);
 125         to->on_disk_key.k_type = le_key_k_type(version, from);
 126 }
 127
 128 /*
 129  * this does not say which one is bigger, it only returns 1 if keys
 130  * are not equal, 0 otherwise
 131  */
 132 inline int comp_le_keys(const struct reiserfs_key *k1,
 133                         const struct reiserfs_key *k2)
 134 {
 135         return memcmp(k1, k2, sizeof(struct reiserfs_key));
 136 }
 137
 138 /**************************************************************************
 139  *  Binary search toolkit function                                        *
 140  *  Search for an item in the array by the item key                       *
 141  *  Returns:    1 if found,  0 if not found;                              *
 142  *        *pos = number of the searched element if found, else the        *
 143  *        number of the first element that is larger than key.            *
 144  **************************************************************************/
 145 /*
 146  * For those not familiar with binary search: lbound is the leftmost item
 147  * that it could be, rbound the rightmost item that it could be.  We examine
 148  * the item halfway between lbound and rbound, and that tells us either
 149  * that we can increase lbound, or decrease rbound, or that we have found it,
 150  * or if lbound <= rbound that there are no possible items, and we have not
 151  * found it. With each examination we cut the number of possible items it
 152  * could be by one more than half rounded down, or we find it.
 153  */
 154 static inline int bin_search(const void *key,   /* Key to search for. */
 155                              const void *base,  /* First item in the array. */
 156                              int num,   /* Number of items in the array. */
 157                              /*
 158                               * Item size in the array.  searched. Lest the
 159                               * reader be confused, note that this is crafted
 160                               * as a general function, and when it is applied
 161                               * specifically to the array of item headers in a
 162                               * node, width is actually the item header size
 163                               * not the item size.
 164                               */
 165                              int width,
 166                              int *pos /* Number of the searched for element. */
 167     )
 168 {
 169         int rbound, lbound, j;
 170
 171         for (j = ((rbound = num - 1) + (lbound = 0)) / 2;
 172              lbound <= rbound; j = (rbound + lbound) / 2)
 173                 switch (comp_keys
 174                         ((struct reiserfs_key *)((char *)base + j * width),
 175                          (struct cpu_key *)key)) {
 176                 case -1:
 177                         lbound = j + 1;
 178                         continue;
 179                 case 1:
 180                         rbound = j - 1;
 181                         continue;
 182                 case 0:
 183                         *pos = j;
 184                         return ITEM_FOUND;      /* Key found in the array.  */
 185                 }
 186
 187         /*
 188          * bin_search did not find given key, it returns position of key,
 189          * that is minimal and greater than the given one.
 190          */
 191         *pos = lbound;
 192         return ITEM_NOT_FOUND;
 193 }
 194
 195
 196 /* Minimal possible key. It is never in the tree. */
 197 const struct reiserfs_key MIN_KEY = { 0, 0, {{0, 0},} };
 198
 199 /* Maximal possible key. It is never in the tree. */
 200 static const struct reiserfs_key MAX_KEY = {
 201         cpu_to_le32(0xffffffff),
 202         cpu_to_le32(0xffffffff),
 203         {{cpu_to_le32(0xffffffff),
 204           cpu_to_le32(0xffffffff)},}
 205 };
 206
 207 /*
 208  * Get delimiting key of the buffer by looking for it in the buffers in the
 209  * path, starting from the bottom of the path, and going upwards.  We must
 210  * check the path's validity at each step.  If the key is not in the path,
 211  * there is no delimiting key in the tree (buffer is first or last buffer
 212  * in tree), and in this case we return a special key, either MIN_KEY or
 213  * MAX_KEY.
 214  */
 215 static inline const struct reiserfs_key *get_lkey(const struct treepath *chk_path,
 216                                                   const struct super_block *sb)
 217 {
 218         int position, path_offset = chk_path->path_length;
 219         struct buffer_head *parent;
 220
 221         RFALSE(path_offset < FIRST_PATH_ELEMENT_OFFSET,
 222                "PAP-5010: invalid offset in the path");
 223
 224         /* While not higher in path than first element. */
 225         while (path_offset-- > FIRST_PATH_ELEMENT_OFFSET) {
 226
 227                 RFALSE(!buffer_uptodate
 228                        (PATH_OFFSET_PBUFFER(chk_path, path_offset)),
 229                        "PAP-5020: parent is not uptodate");
 230
 231                 /* Parent at the path is not in the tree now. */
 232                 if (!B_IS_IN_TREE
 233                     (parent =
 234                      PATH_OFFSET_PBUFFER(chk_path, path_offset)))
 235                         return &MAX_KEY;
 236                 /* Check whether position in the parent is correct. */
 237                 if ((position =
 238                      PATH_OFFSET_POSITION(chk_path,
 239                                           path_offset)) >
 240                     B_NR_ITEMS(parent))
 241                         return &MAX_KEY;
 242                 /* Check whether parent at the path really points to the child. */
 243                 if (B_N_CHILD_NUM(parent, position) !=
 244                     PATH_OFFSET_PBUFFER(chk_path,
 245                                         path_offset + 1)->b_blocknr)
 246                         return &MAX_KEY;
 247                 /*
 248                  * Return delimiting key if position in the parent
 249                  * is not equal to zero.
 250                  */
 251                 if (position)
 252                         return internal_key(parent, position - 1);
 253         }
 254         /* Return MIN_KEY if we are in the root of the buffer tree. */
 255         if (PATH_OFFSET_PBUFFER(chk_path, FIRST_PATH_ELEMENT_OFFSET)->
 256             b_blocknr == SB_ROOT_BLOCK(sb))
 257                 return &MIN_KEY;
 258         return &MAX_KEY;
 259 }
 260
 261 /* Get delimiting key of the buffer at the path and its right neighbor. */
 262 inline const struct reiserfs_key *get_rkey(const struct treepath *chk_path,
 263                                            const struct super_block *sb)
 264 {
 265         int position, path_offset = chk_path->path_length;
 266         struct buffer_head *parent;
 267
 268         RFALSE(path_offset < FIRST_PATH_ELEMENT_OFFSET,
 269                "PAP-5030: invalid offset in the path");
 270
 271         while (path_offset-- > FIRST_PATH_ELEMENT_OFFSET) {
 272
 273                 RFALSE(!buffer_uptodate
 274                        (PATH_OFFSET_PBUFFER(chk_path, path_offset)),
 275                        "PAP-5040: parent is not uptodate");
 276
 277                 /* Parent at the path is not in the tree now. */
 278                 if (!B_IS_IN_TREE
 279                     (parent =
 280                      PATH_OFFSET_PBUFFER(chk_path, path_offset)))
 281                         return &MIN_KEY;
 282                 /* Check whether position in the parent is correct. */
 283                 if ((position =
 284                      PATH_OFFSET_POSITION(chk_path,
 285                                           path_offset)) >
 286                     B_NR_ITEMS(parent))
 287                         return &MIN_KEY;
 288                 /*
 289                  * Check whether parent at the path really points
 290                  * to the child.
 291                  */
 292                 if (B_N_CHILD_NUM(parent, position) !=
 293                     PATH_OFFSET_PBUFFER(chk_path,
 294                                         path_offset + 1)->b_blocknr)
 295                         return &MIN_KEY;
 296
 297                 /*
 298                  * Return delimiting key if position in the parent
 299                  * is not the last one.
 300                  */
 301                 if (position != B_NR_ITEMS(parent))
 302                         return internal_key(parent, position);
 303         }
 304
 305         /* Return MAX_KEY if we are in the root of the buffer tree. */
 306         if (PATH_OFFSET_PBUFFER(chk_path, FIRST_PATH_ELEMENT_OFFSET)->
 307             b_blocknr == SB_ROOT_BLOCK(sb))
 308                 return &MAX_KEY;
 309         return &MIN_KEY;
 310 }
 311
 312 /*
 313  * Check whether a key is contained in the tree rooted from a buffer at a path.
 314  * This works by looking at the left and right delimiting keys for the buffer
 315  * in the last path_element in the path.  These delimiting keys are stored
 316  * at least one level above that buffer in the tree. If the buffer is the
 317  * first or last node in the tree order then one of the delimiting keys may
 318  * be absent, and in this case get_lkey and get_rkey return a special key
 319  * which is MIN_KEY or MAX_KEY.
 320  */
 321 static inline int key_in_buffer(
 322                                 /* Path which should be checked. */
 323                                 struct treepath *chk_path,
 324                                 /* Key which should be checked. */
 325                                 const struct cpu_key *key,
 326                                 struct super_block *sb
 327     )
 328 {
 329
 330         RFALSE(!key || chk_path->path_length < FIRST_PATH_ELEMENT_OFFSET
 331                || chk_path->path_length > MAX_HEIGHT,
 332                "PAP-5050: pointer to the key(%p) is NULL or invalid path length(%d)",
 333                key, chk_path->path_length);
 334         RFALSE(!PATH_PLAST_BUFFER(chk_path)->b_bdev,
 335                "PAP-5060: device must not be NODEV");
 336
 337         if (comp_keys(get_lkey(chk_path, sb), key) == 1)
 338                 /* left delimiting key is bigger, that the key we look for */
 339                 return 0;
 340         /*  if ( comp_keys(key, get_rkey(chk_path, sb)) != -1 ) */
 341         if (comp_keys(get_rkey(chk_path, sb), key) != 1)
 342                 /* key must be less than right delimitiing key */
 343                 return 0;
 344         return 1;
 345 }
 346
 347 int reiserfs_check_path(struct treepath *p)
 348 {
 349         RFALSE(p->path_length != ILLEGAL_PATH_ELEMENT_OFFSET,
 350                "path not properly relsed");
 351         return 0;
 352 }
 353
 354 /*
 355  * Drop the reference to each buffer in a path and restore
 356  * dirty bits clean when preparing the buffer for the log.
 357  * This version should only be called from fix_nodes()
 358  */
 359 void pathrelse_and_restore(struct super_block *sb,
 360                            struct treepath *search_path)
 361 {
 362         int path_offset = search_path->path_length;
 363
 364         RFALSE(path_offset < ILLEGAL_PATH_ELEMENT_OFFSET,
 365                "clm-4000: invalid path offset");
 366
 367         while (path_offset > ILLEGAL_PATH_ELEMENT_OFFSET) {
 368                 struct buffer_head *bh;
 369                 bh = PATH_OFFSET_PBUFFER(search_path, path_offset--);
 370                 reiserfs_restore_prepared_buffer(sb, bh);
 371                 brelse(bh);
 372         }
 373         search_path->path_length = ILLEGAL_PATH_ELEMENT_OFFSET;
 374 }
 375
 376 /* Drop the reference to each buffer in a path */
 377 void pathrelse(struct treepath *search_path)
 378 {
 379         int path_offset = search_path->path_length;
 380
 381         RFALSE(path_offset < ILLEGAL_PATH_ELEMENT_OFFSET,
 382                "PAP-5090: invalid path offset");
 383
 384         while (path_offset > ILLEGAL_PATH_ELEMENT_OFFSET)
 385                 brelse(PATH_OFFSET_PBUFFER(search_path, path_offset--));
 386
 387         search_path->path_length = ILLEGAL_PATH_ELEMENT_OFFSET;
 388 }
 389
 390 static int has_valid_deh_location(struct buffer_head *bh, struct item_head *ih)
 391 {
 392         struct reiserfs_de_head *deh;
 393         int i;
 394
 395         deh = B_I_DEH(bh, ih);
 396         for (i = 0; i < ih_entry_count(ih); i++) {
 397                 if (deh_location(&deh[i]) > ih_item_len(ih)) {
 398                         reiserfs_warning(NULL, "reiserfs-5094",
 399                                          "directory entry location seems wrong %h",
 400                                          &deh[i]);
 401                         return 0;
 402                 }
 403         }
 404
 405         return 1;
 406 }
 407
 408 static int is_leaf(char *buf, int blocksize, struct buffer_head *bh)
 409 {
 410         struct block_head *blkh;
 411         struct item_head *ih;
 412         int used_space;
 413         int prev_location;
 414         int i;
 415         int nr;
 416
 417         blkh = (struct block_head *)buf;
 418         if (blkh_level(blkh) != DISK_LEAF_NODE_LEVEL) {
 419                 reiserfs_warning(NULL, "reiserfs-5080",
 420                                  "this should be caught earlier");
 421                 return 0;
 422         }
 423
 424         nr = blkh_nr_item(blkh);
 425         if (nr < 1 || nr > ((blocksize - BLKH_SIZE) / (IH_SIZE + MIN_ITEM_LEN))) {
 426                 /* item number is too big or too small */
 427                 reiserfs_warning(NULL, "reiserfs-5081",
 428                                  "nr_item seems wrong: %z", bh);
 429                 return 0;
 430         }
 431         ih = (struct item_head *)(buf + BLKH_SIZE) + nr - 1;
 432         used_space = BLKH_SIZE + IH_SIZE * nr + (blocksize - ih_location(ih));
 433
 434         /* free space does not match to calculated amount of use space */
 435         if (used_space != blocksize - blkh_free_space(blkh)) {
 436                 reiserfs_warning(NULL, "reiserfs-5082",
 437                                  "free space seems wrong: %z", bh);
 438                 return 0;
 439         }
 440         /*
 441          * FIXME: it is_leaf will hit performance too much - we may have
 442          * return 1 here
 443          */
 444
 445         /* check tables of item heads */
 446         ih = (struct item_head *)(buf + BLKH_SIZE);
 447         prev_location = blocksize;
 448         for (i = 0; i < nr; i++, ih++) {
 449                 if (le_ih_k_type(ih) == TYPE_ANY) {
 450                         reiserfs_warning(NULL, "reiserfs-5083",
 451                                          "wrong item type for item %h",
 452                                          ih);
 453                         return 0;
 454                 }
 455                 if (ih_location(ih) >= blocksize
 456                     || ih_location(ih) < IH_SIZE * nr) {
 457                         reiserfs_warning(NULL, "reiserfs-5084",
 458                                          "item location seems wrong: %h",
 459                                          ih);
 460                         return 0;
 461                 }
 462                 if (ih_item_len(ih) < 1
 463                     || ih_item_len(ih) > MAX_ITEM_LEN(blocksize)) {
 464                         reiserfs_warning(NULL, "reiserfs-5085",
 465                                          "item length seems wrong: %h",
 466                                          ih);
 467                         return 0;
 468                 }
 469                 if (prev_location - ih_location(ih) != ih_item_len(ih)) {
 470                         reiserfs_warning(NULL, "reiserfs-5086",
 471                                          "item location seems wrong "
 472                                          "(second one): %h", ih);
 473                         return 0;
 474                 }
 475                 if (is_direntry_le_ih(ih)) {
 476                         if (ih_item_len(ih) < (ih_entry_count(ih) * IH_SIZE)) {
 477                                 reiserfs_warning(NULL, "reiserfs-5093",
 478                                                  "item entry count seems wrong %h",
 479                                                  ih);
 480                                 return 0;
 481                         }
 482                         return has_valid_deh_location(bh, ih);
 483                 }
 484                 prev_location = ih_location(ih);
 485         }
 486
 487         /* one may imagine many more checks */
 488         return 1;
 489 }
 490
 491 /* returns 1 if buf looks like an internal node, 0 otherwise */
 492 static int is_internal(char *buf, int blocksize, struct buffer_head *bh)
 493 {
 494         struct block_head *blkh;
 495         int nr;
 496         int used_space;
 497
 498         blkh = (struct block_head *)buf;
 499         nr = blkh_level(blkh);
 500         if (nr <= DISK_LEAF_NODE_LEVEL || nr > MAX_HEIGHT) {
 501                 /* this level is not possible for internal nodes */
 502                 reiserfs_warning(NULL, "reiserfs-5087",
 503                                  "this should be caught earlier");
 504                 return 0;
 505         }
 506
 507         nr = blkh_nr_item(blkh);
 508         /* for internal which is not root we might check min number of keys */
 509         if (nr > (blocksize - BLKH_SIZE - DC_SIZE) / (KEY_SIZE + DC_SIZE)) {
 510                 reiserfs_warning(NULL, "reiserfs-5088",
 511                                  "number of key seems wrong: %z", bh);
 512                 return 0;
 513         }
 514
 515         used_space = BLKH_SIZE + KEY_SIZE * nr + DC_SIZE * (nr + 1);
 516         if (used_space != blocksize - blkh_free_space(blkh)) {
 517                 reiserfs_warning(NULL, "reiserfs-5089",
 518                                  "free space seems wrong: %z", bh);
 519                 return 0;
 520         }
 521
 522         /* one may imagine many more checks */
 523         return 1;
 524 }
 525
 526 /*
 527  * make sure that bh contains formatted node of reiserfs tree of
 528  * 'level'-th level
 529  */
 530 static int is_tree_node(struct buffer_head *bh, int level)
 531 {
 532         if (B_LEVEL(bh) != level) {
 533                 reiserfs_warning(NULL, "reiserfs-5090", "node level %d does "
 534                                  "not match to the expected one %d",
 535                                  B_LEVEL(bh), level);
 536                 return 0;
 537         }
 538         if (level == DISK_LEAF_NODE_LEVEL)
 539                 return is_leaf(bh->b_data, bh->b_size, bh);
 540
 541         return is_internal(bh->b_data, bh->b_size, bh);
 542 }
 543
 544 #define SEARCH_BY_KEY_READA 16
 545
 546 /*
 547  * The function is NOT SCHEDULE-SAFE!
 548  * It might unlock the write lock if we needed to wait for a block
 549  * to be read. Note that in this case it won't recover the lock to avoid
 550  * high contention resulting from too much lock requests, especially
 551  * the caller (search_by_key) will perform other schedule-unsafe
 552  * operations just after calling this function.
 553  *
 554  * @return depth of lock to be restored after read completes
 555  */
 556 static int search_by_key_reada(struct super_block *s,
 557                                 struct buffer_head **bh,
 558                                 b_blocknr_t *b, int num)
 559 {
 560         int i, j;
 561         int depth = -1;
 562
 563         for (i = 0; i < num; i++) {
 564                 bh[i] = sb_getblk(s, b[i]);
 565         }
 566         /*
 567          * We are going to read some blocks on which we
 568          * have a reference. It's safe, though we might be
 569          * reading blocks concurrently changed if we release
 570          * the lock. But it's still fine because we check later
 571          * if the tree changed
 572          */
 573         for (j = 0; j < i; j++) {
 574                 /*
 575                  * note, this needs attention if we are getting rid of the BKL
 576                  * you have to make sure the prepared bit isn't set on this
 577                  * buffer
 578                  */
 579                 if (!buffer_uptodate(bh[j])) {
 580                         if (depth == -1)
 581                                 depth = reiserfs_write_unlock_nested(s);
 582                         ll_rw_block(REQ_OP_READ, REQ_RAHEAD, 1, bh + j);
 583                 }
 584                 brelse(bh[j]);
 585         }
 586         return depth;
 587 }
 588
 589 /*
 590  * This function fills up the path from the root to the leaf as it
 591  * descends the tree looking for the key.  It uses reiserfs_bread to
 592  * try to find buffers in the cache given their block number.  If it
 593  * does not find them in the cache it reads them from disk.  For each
 594  * node search_by_key finds using reiserfs_bread it then uses
 595  * bin_search to look through that node.  bin_search will find the
 596  * position of the block_number of the next node if it is looking
 597  * through an internal node.  If it is looking through a leaf node
 598  * bin_search will find the position of the item which has key either
 599  * equal to given key, or which is the maximal key less than the given
 600  * key.  search_by_key returns a path that must be checked for the
 601  * correctness of the top of the path but need not be checked for the
 602  * correctness of the bottom of the path
 603  */
 604 /*
 605  * search_by_key - search for key (and item) in stree
 606  * @sb: superblock
 607  * @key: pointer to key to search for
 608  * @search_path: Allocated and initialized struct treepath; Returned filled
 609  *               on success.
 610  * @stop_level: How far down the tree to search, Use DISK_LEAF_NODE_LEVEL to
 611  *              stop at leaf level.
 612  *
 613  * The function is NOT SCHEDULE-SAFE!
 614  */
 615 int search_by_key(struct super_block *sb, const struct cpu_key *key,
 616                   struct treepath *search_path, int stop_level)
 617 {
 618         b_blocknr_t block_number;
 619         int expected_level;
 620         struct buffer_head *bh;
 621         struct path_element *last_element;
 622         int node_level, retval;
 623         int fs_gen;
 624         struct buffer_head *reada_bh[SEARCH_BY_KEY_READA];
 625         b_blocknr_t reada_blocks[SEARCH_BY_KEY_READA];
 626         int reada_count = 0;
 627
 628 #ifdef CONFIG_REISERFS_CHECK
 629         int repeat_counter = 0;
 630 #endif
 631
 632         PROC_INFO_INC(sb, search_by_key);
 633
 634         /*
 635          * As we add each node to a path we increase its count.  This means
 636          * that we must be careful to release all nodes in a path before we
 637          * either discard the path struct or re-use the path struct, as we
 638          * do here.
 639          */
 640
 641         pathrelse(search_path);
 642
 643         /*
 644          * With each iteration of this loop we search through the items in the
 645          * current node, and calculate the next current node(next path element)
 646          * for the next iteration of this loop..
 647          */
 648         block_number = SB_ROOT_BLOCK(sb);
 649         expected_level = -1;
 650         while (1) {
 651
 652 #ifdef CONFIG_REISERFS_CHECK
 653                 if (!(++repeat_counter % 50000))
 654                         reiserfs_warning(sb, "PAP-5100",
 655                                          "%s: there were %d iterations of "
 656                                          "while loop looking for key %K",
 657                                          current->comm, repeat_counter,
 658                                          key);
 659 #endif
 660
 661                 /* prep path to have another element added to it. */
 662                 last_element =
 663                     PATH_OFFSET_PELEMENT(search_path,
 664                                          ++search_path->path_length);
 665                 fs_gen = get_generation(sb);
 666
 667                 /*
 668                  * Read the next tree node, and set the last element
 669                  * in the path to have a pointer to it.
 670                  */
 671                 if ((bh = last_element->pe_buffer =
 672                      sb_getblk(sb, block_number))) {
 673
 674                         /*
 675                          * We'll need to drop the lock if we encounter any
 676                          * buffers that need to be read. If all of them are
 677                          * already up to date, we don't need to drop the lock.
 678                          */
 679                         int depth = -1;
 680
 681                         if (!buffer_uptodate(bh) && reada_count > 1)
 682                                 depth = search_by_key_reada(sb, reada_bh,
 683                                                     reada_blocks, reada_count);
 684
 685                         if (!buffer_uptodate(bh) && depth == -1)
 686                                 depth = reiserfs_write_unlock_nested(sb);
 687
 688                         ll_rw_block(REQ_OP_READ, 0, 1, &bh);
 689                         wait_on_buffer(bh);
 690
 691                         if (depth != -1)
 692                                 reiserfs_write_lock_nested(sb, depth);
 693                         if (!buffer_uptodate(bh))
 694                                 goto io_error;
 695                 } else {
 696 io_error:
 697                         search_path->path_length--;
 698                         pathrelse(search_path);
 699                         return IO_ERROR;
 700                 }
 701                 reada_count = 0;
 702                 if (expected_level == -1)
 703                         expected_level = SB_TREE_HEIGHT(sb);
 704                 expected_level--;
 705
 706                 /*
 707                  * It is possible that schedule occurred. We must check
 708                  * whether the key to search is still in the tree rooted
 709                  * from the current buffer. If not then repeat search
 710                  * from the root.
 711                  */
 712                 if (fs_changed(fs_gen, sb) &&
 713                     (!B_IS_IN_TREE(bh) ||
 714                      B_LEVEL(bh) != expected_level ||
 715                      !key_in_buffer(search_path, key, sb))) {
 716                         PROC_INFO_INC(sb, search_by_key_fs_changed);
 717                         PROC_INFO_INC(sb, search_by_key_restarted);
 718                         PROC_INFO_INC(sb,
 719                                       sbk_restarted[expected_level - 1]);
 720                         pathrelse(search_path);
 721
 722                         /*
 723                          * Get the root block number so that we can
 724                          * repeat the search starting from the root.
 725                          */
 726                         block_number = SB_ROOT_BLOCK(sb);
 727                         expected_level = -1;
 728
 729                         /* repeat search from the root */
 730                         continue;
 731                 }
 732
 733                 /*
 734                  * only check that the key is in the buffer if key is not
 735                  * equal to the MAX_KEY. Latter case is only possible in
 736                  * "finish_unfinished()" processing during mount.
 737                  */
 738                 RFALSE(comp_keys(&MAX_KEY, key) &&
 739                        !key_in_buffer(search_path, key, sb),
 740                        "PAP-5130: key is not in the buffer");
 741 #ifdef CONFIG_REISERFS_CHECK
 742                 if (REISERFS_SB(sb)->cur_tb) {
 743                         print_cur_tb("5140");
 744                         reiserfs_panic(sb, "PAP-5140",
 745                                        "schedule occurred in do_balance!");
 746                 }
 747 #endif
 748
 749                 /*
 750                  * make sure, that the node contents look like a node of
 751                  * certain level
 752                  */
 753                 if (!is_tree_node(bh, expected_level)) {
 754                         reiserfs_error(sb, "vs-5150",
 755                                        "invalid format found in block %ld. "
 756                                        "Fsck?", bh->b_blocknr);
 757                         pathrelse(search_path);
 758                         return IO_ERROR;
 759                 }
 760
 761                 /* ok, we have acquired next formatted node in the tree */
 762                 node_level = B_LEVEL(bh);
 763
 764                 PROC_INFO_BH_STAT(sb, bh, node_level - 1);
 765
 766                 RFALSE(node_level < stop_level,
 767                        "vs-5152: tree level (%d) is less than stop level (%d)",
 768                        node_level, stop_level);
 769
 770                 retval = bin_search(key, item_head(bh, 0),
 771                                       B_NR_ITEMS(bh),
 772                                       (node_level ==
 773                                        DISK_LEAF_NODE_LEVEL) ? IH_SIZE :
 774                                       KEY_SIZE,
 775                                       &last_element->pe_position);
 776                 if (node_level == stop_level) {
 777                         return retval;
 778                 }
 779
 780                 /* we are not in the stop level */
 781                 /*
 782                  * item has been found, so we choose the pointer which
 783                  * is to the right of the found one
 784                  */
 785                 if (retval == ITEM_FOUND)
 786                         last_element->pe_position++;
 787
 788                 /*
 789                  * if item was not found we choose the position which is to
 790                  * the left of the found item. This requires no code,
 791                  * bin_search did it already.
 792                  */
 793
 794                 /*
 795                  * So we have chosen a position in the current node which is
 796                  * an internal node.  Now we calculate child block number by
 797                  * position in the node.
 798                  */
 799                 block_number =
 800                     B_N_CHILD_NUM(bh, last_element->pe_position);
 801
 802                 /*
 803                  * if we are going to read leaf nodes, try for read
 804                  * ahead as well
 805                  */
 806                 if ((search_path->reada & PATH_READA) &&
 807                     node_level == DISK_LEAF_NODE_LEVEL + 1) {
 808                         int pos = last_element->pe_position;
 809                         int limit = B_NR_ITEMS(bh);
 810                         struct reiserfs_key *le_key;
 811
 812                         if (search_path->reada & PATH_READA_BACK)
 813                                 limit = 0;
 814                         while (reada_count < SEARCH_BY_KEY_READA) {
 815                                 if (pos == limit)
 816                                         break;
 817                                 reada_blocks[reada_count++] =
 818                                     B_N_CHILD_NUM(bh, pos);
 819                                 if (search_path->reada & PATH_READA_BACK)
 820                                         pos--;
 821                                 else
 822                                         pos++;
 823
 824                                 /*
 825                                  * check to make sure we're in the same object
 826                                  */
 827                                 le_key = internal_key(bh, pos);
 828                                 if (le32_to_cpu(le_key->k_objectid) !=
 829                                     key->on_disk_key.k_objectid) {
 830                                         break;
 831                                 }
 832                         }
 833                 }
 834         }
 835 }
 836
 837 /*
 838  * Form the path to an item and position in this item which contains
 839  * file byte defined by key. If there is no such item
 840  * corresponding to the key, we point the path to the item with
 841  * maximal key less than key, and *pos_in_item is set to one
 842  * past the last entry/byte in the item.  If searching for entry in a
 843  * directory item, and it is not found, *pos_in_item is set to one
 844  * entry more than the entry with maximal key which is less than the
 845  * sought key.
 846  *
 847  * Note that if there is no entry in this same node which is one more,
 848  * then we point to an imaginary entry.  for direct items, the
 849  * position is in units of bytes, for indirect items the position is
 850  * in units of blocknr entries, for directory items the position is in
 851  * units of directory entries.
 852  */
 853 /* The function is NOT SCHEDULE-SAFE! */
 854 int search_for_position_by_key(struct super_block *sb,
 855                                /* Key to search (cpu variable) */
 856                                const struct cpu_key *p_cpu_key,
 857                                /* Filled up by this function. */
 858                                struct treepath *search_path)
 859 {
 860         struct item_head *p_le_ih;      /* pointer to on-disk structure */
 861         int blk_size;
 862         loff_t item_offset, offset;
 863         struct reiserfs_dir_entry de;
 864         int retval;
 865
 866         /* If searching for directory entry. */
 867         if (is_direntry_cpu_key(p_cpu_key))
 868                 return search_by_entry_key(sb, p_cpu_key, search_path,
 869                                            &de);
 870
 871         /* If not searching for directory entry. */
 872
 873         /* If item is found. */
 874         retval = search_item(sb, p_cpu_key, search_path);
 875         if (retval == IO_ERROR)
 876                 return retval;
 877         if (retval == ITEM_FOUND) {
 878
 879                 RFALSE(!ih_item_len
 880                        (item_head
 881                         (PATH_PLAST_BUFFER(search_path),
 882                          PATH_LAST_POSITION(search_path))),
 883                        "PAP-5165: item length equals zero");
 884
 885                 pos_in_item(search_path) = 0;
 886                 return POSITION_FOUND;
 887         }
 888
 889         RFALSE(!PATH_LAST_POSITION(search_path),
 890                "PAP-5170: position equals zero");
 891
 892         /* Item is not found. Set path to the previous item. */
 893         p_le_ih =
 894             item_head(PATH_PLAST_BUFFER(search_path),
 895                            --PATH_LAST_POSITION(search_path));
 896         blk_size = sb->s_blocksize;
 897
 898         if (comp_short_keys(&p_le_ih->ih_key, p_cpu_key))
 899                 return FILE_NOT_FOUND;
 900
 901         /* FIXME: quite ugly this far */
 902
 903         item_offset = le_ih_k_offset(p_le_ih);
 904         offset = cpu_key_k_offset(p_cpu_key);
 905
 906         /* Needed byte is contained in the item pointed to by the path. */
 907         if (item_offset <= offset &&
 908             item_offset + op_bytes_number(p_le_ih, blk_size) > offset) {
 909                 pos_in_item(search_path) = offset - item_offset;
 910                 if (is_indirect_le_ih(p_le_ih)) {
 911                         pos_in_item(search_path) /= blk_size;
 912                 }
 913                 return POSITION_FOUND;
 914         }
 915
 916         /*
 917          * Needed byte is not contained in the item pointed to by the
 918          * path. Set pos_in_item out of the item.
 919          */
 920         if (is_indirect_le_ih(p_le_ih))
 921                 pos_in_item(search_path) =
 922                     ih_item_len(p_le_ih) / UNFM_P_SIZE;
 923         else
 924                 pos_in_item(search_path) = ih_item_len(p_le_ih);
 925
 926         return POSITION_NOT_FOUND;
 927 }
 928
 929 /* Compare given item and item pointed to by the path. */
 930 int comp_items(const struct item_head *stored_ih, const struct treepath *path)
 931 {
 932         struct buffer_head *bh = PATH_PLAST_BUFFER(path);
 933         struct item_head *ih;
 934
 935         /* Last buffer at the path is not in the tree. */
 936         if (!B_IS_IN_TREE(bh))
 937                 return 1;
 938
 939         /* Last path position is invalid. */
 940         if (PATH_LAST_POSITION(path) >= B_NR_ITEMS(bh))
 941                 return 1;
 942
 943         /* we need only to know, whether it is the same item */
 944         ih = tp_item_head(path);
 945         return memcmp(stored_ih, ih, IH_SIZE);
 946 }
 947
 948 /* prepare for delete or cut of direct item */
 949 static inline int prepare_for_direct_item(struct treepath *path,
 950                                           struct item_head *le_ih,
 951                                           struct inode *inode,
 952                                           loff_t new_file_length, int *cut_size)
 953 {
 954         loff_t round_len;
 955
 956         if (new_file_length == max_reiserfs_offset(inode)) {
 957                 /* item has to be deleted */
 958                 *cut_size = -(IH_SIZE + ih_item_len(le_ih));
 959                 return M_DELETE;
 960         }
 961         /* new file gets truncated */
 962         if (get_inode_item_key_version(inode) == KEY_FORMAT_3_6) {
 963                 round_len = ROUND_UP(new_file_length);
 964                 /* this was new_file_length < le_ih ... */
 965                 if (round_len < le_ih_k_offset(le_ih)) {
 966                         *cut_size = -(IH_SIZE + ih_item_len(le_ih));
 967                         return M_DELETE;        /* Delete this item. */
 968                 }
 969                 /* Calculate first position and size for cutting from item. */
 970                 pos_in_item(path) = round_len - (le_ih_k_offset(le_ih) - 1);
 971                 *cut_size = -(ih_item_len(le_ih) - pos_in_item(path));
 972
 973                 return M_CUT;   /* Cut from this item. */
 974         }
 975
 976         /* old file: items may have any length */
 977
 978         if (new_file_length < le_ih_k_offset(le_ih)) {
 979                 *cut_size = -(IH_SIZE + ih_item_len(le_ih));
 980                 return M_DELETE;        /* Delete this item. */
 981         }
 982
 983         /* Calculate first position and size for cutting from item. */
 984         *cut_size = -(ih_item_len(le_ih) -
 985                       (pos_in_item(path) =
 986                        new_file_length + 1 - le_ih_k_offset(le_ih)));
 987         return M_CUT;           /* Cut from this item. */
 988 }
 989
 990 static inline int prepare_for_direntry_item(struct treepath *path,
 991                                             struct item_head *le_ih,
 992                                             struct inode *inode,
 993                                             loff_t new_file_length,
 994                                             int *cut_size)
 995 {
 996         if (le_ih_k_offset(le_ih) == DOT_OFFSET &&
 997             new_file_length == max_reiserfs_offset(inode)) {
 998                 RFALSE(ih_entry_count(le_ih) != 2,
 999                        "PAP-5220: incorrect empty directory item (%h)", le_ih);
1000                 *cut_size = -(IH_SIZE + ih_item_len(le_ih));
1001                 /* Delete the directory item containing "." and ".." entry. */
1002                 return M_DELETE;
1003         }
1004
1005         if (ih_entry_count(le_ih) == 1) {
1006                 /*
1007                  * Delete the directory item such as there is one record only
1008                  * in this item
1009                  */
1010                 *cut_size = -(IH_SIZE + ih_item_len(le_ih));
1011                 return M_DELETE;
1012         }
1013
1014         /* Cut one record from the directory item. */
1015         *cut_size =
1016             -(DEH_SIZE +
1017               entry_length(get_last_bh(path), le_ih, pos_in_item(path)));
1018         return M_CUT;
1019 }
1020
1021 #define JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD (2 * JOURNAL_PER_BALANCE_CNT + 1)
1022
1023 /*
1024  * If the path points to a directory or direct item, calculate mode
1025  * and the size cut, for balance.
1026  * If the path points to an indirect item, remove some number of its
1027  * unformatted nodes.
1028  * In case of file truncate calculate whether this item must be
1029  * deleted/truncated or last unformatted node of this item will be
1030  * converted to a direct item.
1031  * This function returns a determination of what balance mode the
1032  * calling function should employ.
1033  */
1034 static char prepare_for_delete_or_cut(struct reiserfs_transaction_handle *th,
1035                                       struct inode *inode,
1036                                       struct treepath *path,
1037                                       const struct cpu_key *item_key,
1038                                       /*
1039                                        * Number of unformatted nodes
1040                                        * which were removed from end
1041                                        * of the file.
1042                                        */
1043                                       int *removed,
1044                                       int *cut_size,
1045                                       /* MAX_KEY_OFFSET in case of delete. */
1046                                       unsigned long long new_file_length
1047     )
1048 {
1049         struct super_block *sb = inode->i_sb;
1050         struct item_head *p_le_ih = tp_item_head(path);
1051         struct buffer_head *bh = PATH_PLAST_BUFFER(path);
1052
1053         BUG_ON(!th->t_trans_id);
1054
1055         /* Stat_data item. */
1056         if (is_statdata_le_ih(p_le_ih)) {
1057
1058                 RFALSE(new_file_length != max_reiserfs_offset(inode),
1059                        "PAP-5210: mode must be M_DELETE");
1060
1061                 *cut_size = -(IH_SIZE + ih_item_len(p_le_ih));
1062                 return M_DELETE;
1063         }
1064
1065         /* Directory item. */
1066         if (is_direntry_le_ih(p_le_ih))
1067                 return prepare_for_direntry_item(path, p_le_ih, inode,
1068                                                  new_file_length,
1069                                                  cut_size);
1070
1071         /* Direct item. */
1072         if (is_direct_le_ih(p_le_ih))
1073                 return prepare_for_direct_item(path, p_le_ih, inode,
1074                                                new_file_length, cut_size);
1075
1076         /* Case of an indirect item. */
1077         {
1078             int blk_size = sb->s_blocksize;
1079             struct item_head s_ih;
1080             int need_re_search;
1081             int delete = 0;
1082             int result = M_CUT;
1083             int pos = 0;
1084
1085             if ( new_file_length == max_reiserfs_offset (inode) ) {
1086                 /*
1087                  * prepare_for_delete_or_cut() is called by
1088                  * reiserfs_delete_item()
1089                  */
1090                 new_file_length = 0;
1091                 delete = 1;
1092             }
1093
1094             do {
1095                 need_re_search = 0;
1096                 *cut_size = 0;
1097                 bh = PATH_PLAST_BUFFER(path);
1098                 copy_item_head(&s_ih, tp_item_head(path));
1099                 pos = I_UNFM_NUM(&s_ih);
1100
1101                 while (le_ih_k_offset (&s_ih) + (pos - 1) * blk_size > new_file_length) {
1102                     __le32 *unfm;
1103                     __u32 block;
1104
1105                     /*
1106                      * Each unformatted block deletion may involve
1107                      * one additional bitmap block into the transaction,
1108                      * thereby the initial journal space reservation
1109                      * might not be enough.
1110                      */
1111                     if (!delete && (*cut_size) != 0 &&
1112                         reiserfs_transaction_free_space(th) < JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD)
1113                         break;
1114
1115                     unfm = (__le32 *)ih_item_body(bh, &s_ih) + pos - 1;
1116                     block = get_block_num(unfm, 0);
1117
1118                     if (block != 0) {
1119                         reiserfs_prepare_for_journal(sb, bh, 1);
1120                         put_block_num(unfm, 0, 0);
1121                         journal_mark_dirty(th, bh);
1122                         reiserfs_free_block(th, inode, block, 1);
1123                     }
1124
1125                     reiserfs_cond_resched(sb);
1126
1127                     if (item_moved (&s_ih, path))  {
1128                         need_re_search = 1;
1129                         break;
1130                     }
1131
1132                     pos --;
1133                     (*removed)++;
1134                     (*cut_size) -= UNFM_P_SIZE;
1135
1136                     if (pos == 0) {
1137                         (*cut_size) -= IH_SIZE;
1138                         result = M_DELETE;
1139                         break;
1140                     }
1141                 }
1142                 /*
1143                  * a trick.  If the buffer has been logged, this will
1144                  * do nothing.  If we've broken the loop without logging
1145                  * it, it will restore the buffer
1146                  */
1147                 reiserfs_restore_prepared_buffer(sb, bh);
1148             } while (need_re_search &&
1149                      search_for_position_by_key(sb, item_key, path) == POSITION_FOUND);
1150             pos_in_item(path) = pos * UNFM_P_SIZE;
1151
1152             if (*cut_size == 0) {
1153                 /*
1154                  * Nothing was cut. maybe convert last unformatted node to the
1155                  * direct item?
1156                  */
1157                 result = M_CONVERT;
1158             }
1159             return result;
1160         }
1161 }
1162
1163 /* Calculate number of bytes which will be deleted or cut during balance */
1164 static int calc_deleted_bytes_number(struct tree_balance *tb, char mode)
1165 {
1166         int del_size;
1167         struct item_head *p_le_ih = tp_item_head(tb->tb_path);
1168
1169         if (is_statdata_le_ih(p_le_ih))
1170                 return 0;
1171
1172         del_size =
1173             (mode ==
1174              M_DELETE) ? ih_item_len(p_le_ih) : -tb->insert_size[0];
1175         if (is_direntry_le_ih(p_le_ih)) {
1176                 /*
1177                  * return EMPTY_DIR_SIZE; We delete emty directories only.
1178                  * we can't use EMPTY_DIR_SIZE, as old format dirs have a
1179                  * different empty size.  ick. FIXME, is this right?
1180                  */
1181                 return del_size;
1182         }
1183
1184         if (is_indirect_le_ih(p_le_ih))
1185                 del_size = (del_size / UNFM_P_SIZE) *
1186                                 (PATH_PLAST_BUFFER(tb->tb_path)->b_size);
1187         return del_size;
1188 }
1189
1190 static void init_tb_struct(struct reiserfs_transaction_handle *th,
1191                            struct tree_balance *tb,
1192                            struct super_block *sb,
1193                            struct treepath *path, int size)
1194 {
1195
1196         BUG_ON(!th->t_trans_id);
1197
1198         memset(tb, '\0', sizeof(struct tree_balance));
1199         tb->transaction_handle = th;
1200         tb->tb_sb = sb;
1201         tb->tb_path = path;
1202         PATH_OFFSET_PBUFFER(path, ILLEGAL_PATH_ELEMENT_OFFSET) = NULL;
1203         PATH_OFFSET_POSITION(path, ILLEGAL_PATH_ELEMENT_OFFSET) = 0;
1204         tb->insert_size[0] = size;
1205 }
1206
1207 void padd_item(char *item, int total_length, int length)
1208 {
1209         int i;
1210
1211         for (i = total_length; i > length;)
1212                 item[--i] = 0;
1213 }
1214
1215 #ifdef REISERQUOTA_DEBUG
1216 char key2type(struct reiserfs_key *ih)
1217 {
1218         if (is_direntry_le_key(2, ih))
1219                 return 'd';
1220         if (is_direct_le_key(2, ih))
1221                 return 'D';
1222         if (is_indirect_le_key(2, ih))
1223                 return 'i';
1224         if (is_statdata_le_key(2, ih))
1225                 return 's';
1226         return 'u';
1227 }
1228
1229 char head2type(struct item_head *ih)
1230 {
1231         if (is_direntry_le_ih(ih))
1232                 return 'd';
1233         if (is_direct_le_ih(ih))
1234                 return 'D';
1235         if (is_indirect_le_ih(ih))
1236                 return 'i';
1237         if (is_statdata_le_ih(ih))
1238                 return 's';
1239         return 'u';
1240 }
1241 #endif
1242
1243 /*
1244  * Delete object item.
1245  * th       - active transaction handle
1246  * path     - path to the deleted item
1247  * item_key - key to search for the deleted item
1248  * indode   - used for updating i_blocks and quotas
1249  * un_bh    - NULL or unformatted node pointer
1250  */
1251 int reiserfs_delete_item(struct reiserfs_transaction_handle *th,
1252                          struct treepath *path, const struct cpu_key *item_key,
1253                          struct inode *inode, struct buffer_head *un_bh)
1254 {
1255         struct super_block *sb = inode->i_sb;
1256         struct tree_balance s_del_balance;
1257         struct item_head s_ih;
1258         struct item_head *q_ih;
1259         int quota_cut_bytes;
1260         int ret_value, del_size, removed;
1261         int depth;
1262
1263 #ifdef CONFIG_REISERFS_CHECK
1264         char mode;
1265         int iter = 0;
1266 #endif
1267
1268         BUG_ON(!th->t_trans_id);
1269
1270         init_tb_struct(th, &s_del_balance, sb, path,
1271                        0 /*size is unknown */ );
1272
1273         while (1) {
1274                 removed = 0;
1275
1276 #ifdef CONFIG_REISERFS_CHECK
1277                 iter++;
1278                 mode =
1279 #endif
1280                     prepare_for_delete_or_cut(th, inode, path,
1281                                               item_key, &removed,
1282                                               &del_size,
1283                                               max_reiserfs_offset(inode));
1284
1285                 RFALSE(mode != M_DELETE, "PAP-5320: mode must be M_DELETE");
1286
1287                 copy_item_head(&s_ih, tp_item_head(path));
1288                 s_del_balance.insert_size[0] = del_size;
1289
1290                 ret_value = fix_nodes(M_DELETE, &s_del_balance, NULL, NULL);
1291                 if (ret_value != REPEAT_SEARCH)
1292                         break;
1293
1294                 PROC_INFO_INC(sb, delete_item_restarted);
1295
1296                 /* file system changed, repeat search */
1297                 ret_value =
1298                     search_for_position_by_key(sb, item_key, path);
1299                 if (ret_value == IO_ERROR)
1300                         break;
1301                 if (ret_value == FILE_NOT_FOUND) {
1302                         reiserfs_warning(sb, "vs-5340",
1303                                          "no items of the file %K found",
1304                                          item_key);
1305                         break;
1306                 }
1307         }                       /* while (1) */
1308
1309         if (ret_value != CARRY_ON) {
1310                 unfix_nodes(&s_del_balance);
1311                 return 0;
1312         }
1313
1314         /* reiserfs_delete_item returns item length when success */
1315         ret_value = calc_deleted_bytes_number(&s_del_balance, M_DELETE);
1316         q_ih = tp_item_head(path);
1317         quota_cut_bytes = ih_item_len(q_ih);
1318
1319         /*
1320          * hack so the quota code doesn't have to guess if the file has a
1321          * tail.  On tail insert, we allocate quota for 1 unformatted node.
1322          * We test the offset because the tail might have been
1323          * split into multiple items, and we only want to decrement for
1324          * the unfm node once
1325          */
1326         if (!S_ISLNK(inode->i_mode) && is_direct_le_ih(q_ih)) {
1327                 if ((le_ih_k_offset(q_ih) & (sb->s_blocksize - 1)) == 1) {
1328                         quota_cut_bytes = sb->s_blocksize + UNFM_P_SIZE;
1329                 } else {
1330                         quota_cut_bytes = 0;
1331                 }
1332         }
1333
1334         if (un_bh) {
1335                 int off;
1336                 char *data;
1337
1338                 /*
1339                  * We are in direct2indirect conversion, so move tail contents
1340                  * to the unformatted node
1341                  */
1342                 /*
1343                  * note, we do the copy before preparing the buffer because we
1344                  * don't care about the contents of the unformatted node yet.
1345                  * the only thing we really care about is the direct item's
1346                  * data is in the unformatted node.
1347                  *
1348                  * Otherwise, we would have to call
1349                  * reiserfs_prepare_for_journal on the unformatted node,
1350                  * which might schedule, meaning we'd have to loop all the
1351                  * way back up to the start of the while loop.
1352                  *
1353                  * The unformatted node must be dirtied later on.  We can't be
1354                  * sure here if the entire tail has been deleted yet.
1355                  *
1356                  * un_bh is from the page cache (all unformatted nodes are
1357                  * from the page cache) and might be a highmem page.  So, we
1358                  * can't use un_bh->b_data.
1359                  * -clm
1360                  */
1361
1362                 data = kmap_atomic(un_bh->b_page);
1363                 off = ((le_ih_k_offset(&s_ih) - 1) & (PAGE_SIZE - 1));
1364                 memcpy(data + off,
1365                        ih_item_body(PATH_PLAST_BUFFER(path), &s_ih),
1366                        ret_value);
1367                 kunmap_atomic(data);
1368         }
1369
1370         /* Perform balancing after all resources have been collected at once. */
1371         do_balance(&s_del_balance, NULL, NULL, M_DELETE);
1372
1373 #ifdef REISERQUOTA_DEBUG
1374         reiserfs_debug(sb, REISERFS_DEBUG_CODE,
1375                        "reiserquota delete_item(): freeing %u, id=%u type=%c",
1376                        quota_cut_bytes, inode->i_uid, head2type(&s_ih));
1377 #endif
1378         depth = reiserfs_write_unlock_nested(inode->i_sb);
1379         dquot_free_space_nodirty(inode, quota_cut_bytes);
1380         reiserfs_write_lock_nested(inode->i_sb, depth);
1381
1382         /* Return deleted body length */
1383         return ret_value;
1384 }
1385
1386 /*
1387  * Summary Of Mechanisms For Handling Collisions Between Processes:
1388  *
1389  *  deletion of the body of the object is performed by iput(), with the
1390  *  result that if multiple processes are operating on a file, the
1391  *  deletion of the body of the file is deferred until the last process
1392  *  that has an open inode performs its iput().
1393  *
1394  *  writes and truncates are protected from collisions by use of
1395  *  semaphores.
1396  *
1397  *  creates, linking, and mknod are protected from collisions with other
1398  *  processes by making the reiserfs_add_entry() the last step in the
1399  *  creation, and then rolling back all changes if there was a collision.
1400  *  - Hans
1401 */
1402
1403 /* this deletes item which never gets split */
1404 void reiserfs_delete_solid_item(struct reiserfs_transaction_handle *th,
1405                                 struct inode *inode, struct reiserfs_key *key)
1406 {
1407         struct super_block *sb = th->t_super;
1408         struct tree_balance tb;
1409         INITIALIZE_PATH(path);
1410         int item_len = 0;
1411         int tb_init = 0;
1412         struct cpu_key cpu_key;
1413         int retval;
1414         int quota_cut_bytes = 0;
1415
1416         BUG_ON(!th->t_trans_id);
1417
1418         le_key2cpu_key(&cpu_key, key);
1419
1420         while (1) {
1421                 retval = search_item(th->t_super, &cpu_key, &path);
1422                 if (retval == IO_ERROR) {
1423                         reiserfs_error(th->t_super, "vs-5350",
1424                                        "i/o failure occurred trying "
1425                                        "to delete %K", &cpu_key);
1426                         break;
1427                 }
1428                 if (retval != ITEM_FOUND) {
1429                         pathrelse(&path);
1430                         /*
1431                          * No need for a warning, if there is just no free
1432                          * space to insert '..' item into the
1433                          * newly-created subdir
1434                          */
1435                         if (!
1436                             ((unsigned long long)
1437                              GET_HASH_VALUE(le_key_k_offset
1438                                             (le_key_version(key), key)) == 0
1439                              && (unsigned long long)
1440                              GET_GENERATION_NUMBER(le_key_k_offset
1441                                                    (le_key_version(key),
1442                                                     key)) == 1))
1443                                 reiserfs_warning(th->t_super, "vs-5355",
1444                                                  "%k not found", key);
1445                         break;
1446                 }
1447                 if (!tb_init) {
1448                         tb_init = 1;
1449                         item_len = ih_item_len(tp_item_head(&path));
1450                         init_tb_struct(th, &tb, th->t_super, &path,
1451                                        -(IH_SIZE + item_len));
1452                 }
1453                 quota_cut_bytes = ih_item_len(tp_item_head(&path));
1454
1455                 retval = fix_nodes(M_DELETE, &tb, NULL, NULL);
1456                 if (retval == REPEAT_SEARCH) {
1457                         PROC_INFO_INC(th->t_super, delete_solid_item_restarted);
1458                         continue;
1459                 }
1460
1461                 if (retval == CARRY_ON) {
1462                         do_balance(&tb, NULL, NULL, M_DELETE);
1463                         /*
1464                          * Should we count quota for item? (we don't
1465                          * count quotas for save-links)
1466                          */
1467                         if (inode) {
1468                                 int depth;
1469 #ifdef REISERQUOTA_DEBUG
1470                                 reiserfs_debug(th->t_super, REISERFS_DEBUG_CODE,
1471                                                "reiserquota delete_solid_item(): freeing %u id=%u type=%c",
1472                                                quota_cut_bytes, inode->i_uid,
1473                                                key2type(key));
1474 #endif
1475                                 depth = reiserfs_write_unlock_nested(sb);
1476                                 dquot_free_space_nodirty(inode,
1477                                                          quota_cut_bytes);
1478                                 reiserfs_write_lock_nested(sb, depth);
1479                         }
1480                         break;
1481                 }
1482
1483                 /* IO_ERROR, NO_DISK_SPACE, etc */
1484                 reiserfs_warning(th->t_super, "vs-5360",
1485                                  "could not delete %K due to fix_nodes failure",
1486                                  &cpu_key);
1487                 unfix_nodes(&tb);
1488                 break;
1489         }
1490
1491         reiserfs_check_path(&path);
1492 }
1493
1494 int reiserfs_delete_object(struct reiserfs_transaction_handle *th,
1495                            struct inode *inode)
1496 {
1497         int err;
1498         inode->i_size = 0;
1499         BUG_ON(!th->t_trans_id);
1500
1501         /* for directory this deletes item containing "." and ".." */
1502         err =
1503             reiserfs_do_truncate(th, inode, NULL, 0 /*no timestamp updates */ );
1504         if (err)
1505                 return err;
1506
1507 #if defined( USE_INODE_GENERATION_COUNTER )
1508         if (!old_format_only(th->t_super)) {
1509                 __le32 *inode_generation;
1510
1511                 inode_generation =
1512                     &REISERFS_SB(th->t_super)->s_rs->s_inode_generation;
1513                 le32_add_cpu(inode_generation, 1);
1514         }
1515 /* USE_INODE_GENERATION_COUNTER */
1516 #endif
1517         reiserfs_delete_solid_item(th, inode, INODE_PKEY(inode));
1518
1519         return err;
1520 }
1521
1522 static void unmap_buffers(struct page *page, loff_t pos)
1523 {
1524         struct buffer_head *bh;
1525         struct buffer_head *head;
1526         struct buffer_head *next;
1527         unsigned long tail_index;
1528         unsigned long cur_index;
1529
1530         if (page) {
1531                 if (page_has_buffers(page)) {
1532                         tail_index = pos & (PAGE_SIZE - 1);
1533                         cur_index = 0;
1534                         head = page_buffers(page);
1535                         bh = head;
1536                         do {
1537                                 next = bh->b_this_page;
1538
1539                                 /*
1540                                  * we want to unmap the buffers that contain
1541                                  * the tail, and all the buffers after it
1542                                  * (since the tail must be at the end of the
1543                                  * file).  We don't want to unmap file data
1544                                  * before the tail, since it might be dirty
1545                                  * and waiting to reach disk
1546                                  */
1547                                 cur_index += bh->b_size;
1548                                 if (cur_index > tail_index) {
1549                                         reiserfs_unmap_buffer(bh);
1550                                 }
1551                                 bh = next;
1552                         } while (bh != head);
1553                 }
1554         }
1555 }
1556
1557 static int maybe_indirect_to_direct(struct reiserfs_transaction_handle *th,
1558                                     struct inode *inode,
1559                                     struct page *page,
1560                                     struct treepath *path,
1561                                     const struct cpu_key *item_key,
1562                                     loff_t new_file_size, char *mode)
1563 {
1564         struct super_block *sb = inode->i_sb;
1565         int block_size = sb->s_blocksize;
1566         int cut_bytes;
1567         BUG_ON(!th->t_trans_id);
1568         BUG_ON(new_file_size != inode->i_size);
1569
1570         /*
1571          * the page being sent in could be NULL if there was an i/o error
1572          * reading in the last block.  The user will hit problems trying to
1573          * read the file, but for now we just skip the indirect2direct
1574          */
1575         if (atomic_read(&inode->i_count) > 1 ||
1576             !tail_has_to_be_packed(inode) ||
1577             !page || (REISERFS_I(inode)->i_flags & i_nopack_mask)) {
1578                 /* leave tail in an unformatted node */
1579                 *mode = M_SKIP_BALANCING;
1580                 cut_bytes =
1581                     block_size - (new_file_size & (block_size - 1));
1582                 pathrelse(path);
1583                 return cut_bytes;
1584         }
1585
1586         /* Perform the conversion to a direct_item. */
1587         return indirect2direct(th, inode, page, path, item_key,
1588                                new_file_size, mode);
1589 }
1590
1591 /*
1592  * we did indirect_to_direct conversion. And we have inserted direct
1593  * item successesfully, but there were no disk space to cut unfm
1594  * pointer being converted. Therefore we have to delete inserted
1595  * direct item(s)
1596  */
1597 static void indirect_to_direct_roll_back(struct reiserfs_transaction_handle *th,
1598                                          struct inode *inode, struct treepath *path)
1599 {
1600         struct cpu_key tail_key;
1601         int tail_len;
1602         int removed;
1603         BUG_ON(!th->t_trans_id);
1604
1605         make_cpu_key(&tail_key, inode, inode->i_size + 1, TYPE_DIRECT, 4);
1606         tail_key.key_length = 4;
1607
1608         tail_len =
1609             (cpu_key_k_offset(&tail_key) & (inode->i_sb->s_blocksize - 1)) - 1;
1610         while (tail_len) {
1611                 /* look for the last byte of the tail */
1612                 if (search_for_position_by_key(inode->i_sb, &tail_key, path) ==
1613                     POSITION_NOT_FOUND)
1614                         reiserfs_panic(inode->i_sb, "vs-5615",
1615                                        "found invalid item");
1616                 RFALSE(path->pos_in_item !=
1617                        ih_item_len(tp_item_head(path)) - 1,
1618                        "vs-5616: appended bytes found");
1619                 PATH_LAST_POSITION(path)--;
1620
1621                 removed =
1622                     reiserfs_delete_item(th, path, &tail_key, inode,
1623                                          NULL /*unbh not needed */ );
1624                 RFALSE(removed <= 0
1625                        || removed > tail_len,
1626                        "vs-5617: there was tail %d bytes, removed item length %d bytes",
1627                        tail_len, removed);
1628                 tail_len -= removed;
1629                 set_cpu_key_k_offset(&tail_key,
1630                                      cpu_key_k_offset(&tail_key) - removed);
1631         }
1632         reiserfs_warning(inode->i_sb, "reiserfs-5091", "indirect_to_direct "
1633                          "conversion has been rolled back due to "
1634                          "lack of disk space");
1635         mark_inode_dirty(inode);
1636 }
1637
1638 /* (Truncate or cut entry) or delete object item. Returns < 0 on failure */
1639 int reiserfs_cut_from_item(struct reiserfs_transaction_handle *th,
1640                            struct treepath *path,
1641                            struct cpu_key *item_key,
1642                            struct inode *inode,
1643                            struct page *page, loff_t new_file_size)
1644 {
1645         struct super_block *sb = inode->i_sb;
1646         /*
1647          * Every function which is going to call do_balance must first
1648          * create a tree_balance structure.  Then it must fill up this
1649          * structure by using the init_tb_struct and fix_nodes functions.
1650          * After that we can make tree balancing.
1651          */
1652         struct tree_balance s_cut_balance;
1653         struct item_head *p_le_ih;
1654         int cut_size = 0;       /* Amount to be cut. */
1655         int ret_value = CARRY_ON;
1656         int removed = 0;        /* Number of the removed unformatted nodes. */
1657         int is_inode_locked = 0;
1658         char mode;              /* Mode of the balance. */
1659         int retval2 = -1;
1660         int quota_cut_bytes;
1661         loff_t tail_pos = 0;
1662         int depth;
1663
1664         BUG_ON(!th->t_trans_id);
1665
1666         init_tb_struct(th, &s_cut_balance, inode->i_sb, path,
1667                        cut_size);
1668
1669         /*
1670          * Repeat this loop until we either cut the item without needing
1671          * to balance, or we fix_nodes without schedule occurring
1672          */
1673         while (1) {
1674                 /*
1675                  * Determine the balance mode, position of the first byte to
1676                  * be cut, and size to be cut.  In case of the indirect item
1677                  * free unformatted nodes which are pointed to by the cut
1678                  * pointers.
1679                  */
1680
1681                 mode =
1682                     prepare_for_delete_or_cut(th, inode, path,
1683                                               item_key, &removed,
1684                                               &cut_size, new_file_size);
1685                 if (mode == M_CONVERT) {
1686                         /*
1687                          * convert last unformatted node to direct item or
1688                          * leave tail in the unformatted node
1689                          */
1690                         RFALSE(ret_value != CARRY_ON,
1691                                "PAP-5570: can not convert twice");
1692
1693                         ret_value =
1694                             maybe_indirect_to_direct(th, inode, page,
1695                                                      path, item_key,
1696                                                      new_file_size, &mode);
1697                         if (mode == M_SKIP_BALANCING)
1698                                 /* tail has been left in the unformatted node */
1699                                 return ret_value;
1700
1701                         is_inode_locked = 1;
1702
1703                         /*
1704                          * removing of last unformatted node will
1705                          * change value we have to return to truncate.
1706                          * Save it
1707                          */
1708                         retval2 = ret_value;
1709
1710                         /*
1711                          * So, we have performed the first part of the
1712                          * conversion:
1713                          * inserting the new direct item.  Now we are
1714                          * removing the last unformatted node pointer.
1715                          * Set key to search for it.
1716                          */
1717                         set_cpu_key_k_type(item_key, TYPE_INDIRECT);
1718                         item_key->key_length = 4;
1719                         new_file_size -=
1720                             (new_file_size & (sb->s_blocksize - 1));
1721                         tail_pos = new_file_size;
1722                         set_cpu_key_k_offset(item_key, new_file_size + 1);
1723                         if (search_for_position_by_key
1724                             (sb, item_key,
1725                              path) == POSITION_NOT_FOUND) {
1726                                 print_block(PATH_PLAST_BUFFER(path), 3,
1727                                             PATH_LAST_POSITION(path) - 1,
1728                                             PATH_LAST_POSITION(path) + 1);
1729                                 reiserfs_panic(sb, "PAP-5580", "item to "
1730                                                "convert does not exist (%K)",
1731                                                item_key);
1732                         }
1733                         continue;
1734                 }
1735                 if (cut_size == 0) {
1736                         pathrelse(path);
1737                         return 0;
1738                 }
1739
1740                 s_cut_balance.insert_size[0] = cut_size;
1741
1742                 ret_value = fix_nodes(mode, &s_cut_balance, NULL, NULL);
1743                 if (ret_value != REPEAT_SEARCH)
1744                         break;
1745
1746                 PROC_INFO_INC(sb, cut_from_item_restarted);
1747
1748                 ret_value =
1749                     search_for_position_by_key(sb, item_key, path);
1750                 if (ret_value == POSITION_FOUND)
1751                         continue;
1752
1753                 reiserfs_warning(sb, "PAP-5610", "item %K not found",
1754                                  item_key);
1755                 unfix_nodes(&s_cut_balance);
1756                 return (ret_value == IO_ERROR) ? -EIO : -ENOENT;
1757         }                       /* while */
1758
1759         /* check fix_nodes results (IO_ERROR or NO_DISK_SPACE) */
1760         if (ret_value != CARRY_ON) {
1761                 if (is_inode_locked) {
1762                         /*
1763                          * FIXME: this seems to be not needed: we are always
1764                          * able to cut item
1765                          */
1766                         indirect_to_direct_roll_back(th, inode, path);
1767                 }
1768                 if (ret_value == NO_DISK_SPACE)
1769                         reiserfs_warning(sb, "reiserfs-5092",
1770                                          "NO_DISK_SPACE");
1771                 unfix_nodes(&s_cut_balance);
1772                 return -EIO;
1773         }
1774
1775         /* go ahead and perform balancing */
1776
1777         RFALSE(mode == M_PASTE || mode == M_INSERT, "invalid mode");
1778
1779         /* Calculate number of bytes that need to be cut from the item. */
1780         quota_cut_bytes =
1781             (mode ==
1782              M_DELETE) ? ih_item_len(tp_item_head(path)) : -s_cut_balance.
1783             insert_size[0];
1784         if (retval2 == -1)
1785                 ret_value = calc_deleted_bytes_number(&s_cut_balance, mode);
1786         else
1787                 ret_value = retval2;
1788
1789         /*
1790          * For direct items, we only change the quota when deleting the last
1791          * item.
1792          */
1793         p_le_ih = tp_item_head(s_cut_balance.tb_path);
1794         if (!S_ISLNK(inode->i_mode) && is_direct_le_ih(p_le_ih)) {
1795                 if (mode == M_DELETE &&
1796                     (le_ih_k_offset(p_le_ih) & (sb->s_blocksize - 1)) ==
1797                     1) {
1798                         /* FIXME: this is to keep 3.5 happy */
1799                         REISERFS_I(inode)->i_first_direct_byte = U32_MAX;
1800                         quota_cut_bytes = sb->s_blocksize + UNFM_P_SIZE;
1801                 } else {
1802                         quota_cut_bytes = 0;
1803                 }
1804         }
1805 #ifdef CONFIG_REISERFS_CHECK
1806         if (is_inode_locked) {
1807                 struct item_head *le_ih =
1808                     tp_item_head(s_cut_balance.tb_path);
1809                 /*
1810                  * we are going to complete indirect2direct conversion. Make
1811                  * sure, that we exactly remove last unformatted node pointer
1812                  * of the item
1813                  */
1814                 if (!is_indirect_le_ih(le_ih))
1815                         reiserfs_panic(sb, "vs-5652",
1816                                        "item must be indirect %h", le_ih);
1817
1818                 if (mode == M_DELETE && ih_item_len(le_ih) != UNFM_P_SIZE)
1819                         reiserfs_panic(sb, "vs-5653", "completing "
1820                                        "indirect2direct conversion indirect "
1821                                        "item %h being deleted must be of "
1822                                        "4 byte long", le_ih);
1823
1824                 if (mode == M_CUT
1825                     && s_cut_balance.insert_size[0] != -UNFM_P_SIZE) {
1826                         reiserfs_panic(sb, "vs-5654", "can not complete "
1827                                        "indirect2direct conversion of %h "
1828                                        "(CUT, insert_size==%d)",
1829                                        le_ih, s_cut_balance.insert_size[0]);
1830                 }
1831                 /*
1832                  * it would be useful to make sure, that right neighboring
1833                  * item is direct item of this file
1834                  */
1835         }
1836 #endif
1837
1838         do_balance(&s_cut_balance, NULL, NULL, mode);
1839         if (is_inode_locked) {
1840                 /*
1841                  * we've done an indirect->direct conversion.  when the
1842                  * data block was freed, it was removed from the list of
1843                  * blocks that must be flushed before the transaction
1844                  * commits, make sure to unmap and invalidate it
1845                  */
1846                 unmap_buffers(page, tail_pos);
1847                 REISERFS_I(inode)->i_flags &= ~i_pack_on_close_mask;
1848         }
1849 #ifdef REISERQUOTA_DEBUG
1850         reiserfs_debug(inode->i_sb, REISERFS_DEBUG_CODE,
1851                        "reiserquota cut_from_item(): freeing %u id=%u type=%c",
1852                        quota_cut_bytes, inode->i_uid, '?');
1853 #endif
1854         depth = reiserfs_write_unlock_nested(sb);
1855         dquot_free_space_nodirty(inode, quota_cut_bytes);
1856         reiserfs_write_lock_nested(sb, depth);
1857         return ret_value;
1858 }
1859
1860 static void truncate_directory(struct reiserfs_transaction_handle *th,
1861                                struct inode *inode)
1862 {
1863         BUG_ON(!th->t_trans_id);
1864         if (inode->i_nlink)
1865                 reiserfs_error(inode->i_sb, "vs-5655", "link count != 0");
1866
1867         set_le_key_k_offset(KEY_FORMAT_3_5, INODE_PKEY(inode), DOT_OFFSET);
1868         set_le_key_k_type(KEY_FORMAT_3_5, INODE_PKEY(inode), TYPE_DIRENTRY);
1869         reiserfs_delete_solid_item(th, inode, INODE_PKEY(inode));
1870         reiserfs_update_sd(th, inode);
1871         set_le_key_k_offset(KEY_FORMAT_3_5, INODE_PKEY(inode), SD_OFFSET);
1872         set_le_key_k_type(KEY_FORMAT_3_5, INODE_PKEY(inode), TYPE_STAT_DATA);
1873 }
1874
1875 /*
1876  * Truncate file to the new size. Note, this must be called with a
1877  * transaction already started
1878  */
1879 int reiserfs_do_truncate(struct reiserfs_transaction_handle *th,
1880                          struct inode *inode,   /* ->i_size contains new size */
1881                          struct page *page,     /* up to date for last block */
1882                          /*
1883                           * when it is called by file_release to convert
1884                           * the tail - no timestamps should be updated
1885                           */
1886                          int update_timestamps
1887     )
1888 {
1889         INITIALIZE_PATH(s_search_path); /* Path to the current object item. */
1890         struct item_head *p_le_ih;      /* Pointer to an item header. */
1891
1892         /* Key to search for a previous file item. */
1893         struct cpu_key s_item_key;
1894         loff_t file_size,       /* Old file size. */
1895          new_file_size; /* New file size. */
1896         int deleted;            /* Number of deleted or truncated bytes. */
1897         int retval;
1898         int err = 0;
1899
1900         BUG_ON(!th->t_trans_id);
1901         if (!
1902             (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode)
1903              || S_ISLNK(inode->i_mode)))
1904                 return 0;
1905
1906         /* deletion of directory - no need to update timestamps */
1907         if (S_ISDIR(inode->i_mode)) {
1908                 truncate_directory(th, inode);
1909                 return 0;
1910         }
1911
1912         /* Get new file size. */
1913         new_file_size = inode->i_size;
1914
1915         /* FIXME: note, that key type is unimportant here */
1916         make_cpu_key(&s_item_key, inode, max_reiserfs_offset(inode),
1917                      TYPE_DIRECT, 3);
1918
1919         retval =
1920             search_for_position_by_key(inode->i_sb, &s_item_key,
1921                                        &s_search_path);
1922         if (retval == IO_ERROR) {
1923                 reiserfs_error(inode->i_sb, "vs-5657",
1924                                "i/o failure occurred trying to truncate %K",
1925                                &s_item_key);
1926                 err = -EIO;
1927                 goto out;
1928         }
1929         if (retval == POSITION_FOUND || retval == FILE_NOT_FOUND) {
1930                 reiserfs_error(inode->i_sb, "PAP-5660",
1931                                "wrong result %d of search for %K", retval,
1932                                &s_item_key);
1933
1934                 err = -EIO;
1935                 goto out;
1936         }
1937
1938         s_search_path.pos_in_item--;
1939
1940         /* Get real file size (total length of all file items) */
1941         p_le_ih = tp_item_head(&s_search_path);
1942         if (is_statdata_le_ih(p_le_ih))
1943                 file_size = 0;
1944         else {
1945                 loff_t offset = le_ih_k_offset(p_le_ih);
1946                 int bytes =
1947                     op_bytes_number(p_le_ih, inode->i_sb->s_blocksize);
1948
1949                 /*
1950                  * this may mismatch with real file size: if last direct item
1951                  * had no padding zeros and last unformatted node had no free
1952                  * space, this file would have this file size
1953                  */
1954                 file_size = offset + bytes - 1;
1955         }
1956         /*
1957          * are we doing a full truncate or delete, if so
1958          * kick in the reada code
1959          */
1960         if (new_file_size == 0)
1961                 s_search_path.reada = PATH_READA | PATH_READA_BACK;
1962
1963         if (file_size == 0 || file_size < new_file_size) {
1964                 goto update_and_out;
1965         }
1966
1967         /* Update key to search for the last file item. */
1968         set_cpu_key_k_offset(&s_item_key, file_size);
1969
1970         do {
1971                 /* Cut or delete file item. */
1972                 deleted =
1973                     reiserfs_cut_from_item(th, &s_search_path, &s_item_key,
1974                                            inode, page, new_file_size);
1975                 if (deleted < 0) {
1976                         reiserfs_warning(inode->i_sb, "vs-5665",
1977                                          "reiserfs_cut_from_item failed");
1978                         reiserfs_check_path(&s_search_path);
1979                         return 0;
1980                 }
1981
1982                 RFALSE(deleted > file_size,
1983                        "PAP-5670: reiserfs_cut_from_item: too many bytes deleted: deleted %d, file_size %lu, item_key %K",
1984                        deleted, file_size, &s_item_key);
1985
1986                 /* Change key to search the last file item. */
1987                 file_size -= deleted;
1988
1989                 set_cpu_key_k_offset(&s_item_key, file_size);
1990
1991                 /*
1992                  * While there are bytes to truncate and previous
1993                  * file item is presented in the tree.
1994                  */
1995
1996                 /*
1997                  * This loop could take a really long time, and could log
1998                  * many more blocks than a transaction can hold.  So, we do
1999                  * a polite journal end here, and if the transaction needs
2000                  * ending, we make sure the file is consistent before ending
2001                  * the current trans and starting a new one
2002                  */
2003                 if (journal_transaction_should_end(th, 0) ||
2004                     reiserfs_transaction_free_space(th) <= JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD) {
2005                         pathrelse(&s_search_path);
2006
2007                         if (update_timestamps) {
2008                                 inode->i_mtime = current_time(inode);
2009                                 inode->i_ctime = current_time(inode);
2010                         }
2011                         reiserfs_update_sd(th, inode);
2012
2013                         err = journal_end(th);
2014                         if (err)
2015                                 goto out;
2016                         err = journal_begin(th, inode->i_sb,
2017                                             JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD + JOURNAL_PER_BALANCE_CNT * 4) ;
2018                         if (err)
2019                                 goto out;
2020                         reiserfs_update_inode_transaction(inode);
2021                 }
2022         } while (file_size > ROUND_UP(new_file_size) &&
2023                  search_for_position_by_key(inode->i_sb, &s_item_key,
2024                                             &s_search_path) == POSITION_FOUND);
2025
2026         RFALSE(file_size > ROUND_UP(new_file_size),
2027                "PAP-5680: truncate did not finish: new_file_size %lld, current %lld, oid %d",
2028                new_file_size, file_size, s_item_key.on_disk_key.k_objectid);
2029
2030 update_and_out:
2031         if (update_timestamps) {
2032                 /* this is truncate, not file closing */
2033                 inode->i_mtime = current_time(inode);
2034                 inode->i_ctime = current_time(inode);
2035         }
2036         reiserfs_update_sd(th, inode);
2037
2038 out:
2039         pathrelse(&s_search_path);
2040         return err;
2041 }
2042
2043 #ifdef CONFIG_REISERFS_CHECK
2044 /* this makes sure, that we __append__, not overwrite or add holes */
2045 static void check_research_for_paste(struct treepath *path,
2046                                      const struct cpu_key *key)
2047 {
2048         struct item_head *found_ih = tp_item_head(path);
2049
2050         if (is_direct_le_ih(found_ih)) {
2051                 if (le_ih_k_offset(found_ih) +
2052                     op_bytes_number(found_ih,
2053                                     get_last_bh(path)->b_size) !=
2054                     cpu_key_k_offset(key)
2055                     || op_bytes_number(found_ih,
2056                                        get_last_bh(path)->b_size) !=
2057                     pos_in_item(path))
2058                         reiserfs_panic(NULL, "PAP-5720", "found direct item "
2059                                        "%h or position (%d) does not match "
2060                                        "to key %K", found_ih,
2061                                        pos_in_item(path), key);
2062         }
2063         if (is_indirect_le_ih(found_ih)) {
2064                 if (le_ih_k_offset(found_ih) +
2065                     op_bytes_number(found_ih,
2066                                     get_last_bh(path)->b_size) !=
2067                     cpu_key_k_offset(key)
2068                     || I_UNFM_NUM(found_ih) != pos_in_item(path)
2069                     || get_ih_free_space(found_ih) != 0)
2070                         reiserfs_panic(NULL, "PAP-5730", "found indirect "
2071                                        "item (%h) or position (%d) does not "
2072                                        "match to key (%K)",
2073                                        found_ih, pos_in_item(path), key);
2074         }
2075 }
2076 #endif                          /* config reiserfs check */
2077
2078 /*
2079  * Paste bytes to the existing item.
2080  * Returns bytes number pasted into the item.
2081  */
2082 int reiserfs_paste_into_item(struct reiserfs_transaction_handle *th,
2083                              /* Path to the pasted item. */
2084                              struct treepath *search_path,
2085                              /* Key to search for the needed item. */
2086                              const struct cpu_key *key,
2087                              /* Inode item belongs to */
2088                              struct inode *inode,
2089                              /* Pointer to the bytes to paste. */
2090                              const char *body,
2091                              /* Size of pasted bytes. */
2092                              int pasted_size)
2093 {
2094         struct super_block *sb = inode->i_sb;
2095         struct tree_balance s_paste_balance;
2096         int retval;
2097         int fs_gen;
2098         int depth;
2099
2100         BUG_ON(!th->t_trans_id);
2101
2102         fs_gen = get_generation(inode->i_sb);
2103
2104 #ifdef REISERQUOTA_DEBUG
2105         reiserfs_debug(inode->i_sb, REISERFS_DEBUG_CODE,
2106                        "reiserquota paste_into_item(): allocating %u id=%u type=%c",
2107                        pasted_size, inode->i_uid,
2108                        key2type(&key->on_disk_key));
2109 #endif
2110
2111         depth = reiserfs_write_unlock_nested(sb);
2112         retval = dquot_alloc_space_nodirty(inode, pasted_size);
2113         reiserfs_write_lock_nested(sb, depth);
2114         if (retval) {
2115                 pathrelse(search_path);
2116                 return retval;
2117         }
2118         init_tb_struct(th, &s_paste_balance, th->t_super, search_path,
2119                        pasted_size);
2120 #ifdef DISPLACE_NEW_PACKING_LOCALITIES
2121         s_paste_balance.key = key->on_disk_key;
2122 #endif
2123
2124         /* DQUOT_* can schedule, must check before the fix_nodes */
2125         if (fs_changed(fs_gen, inode->i_sb)) {
2126                 goto search_again;
2127         }
2128
2129         while ((retval =
2130                 fix_nodes(M_PASTE, &s_paste_balance, NULL,
2131                           body)) == REPEAT_SEARCH) {
2132 search_again:
2133                 /* file system changed while we were in the fix_nodes */
2134                 PROC_INFO_INC(th->t_super, paste_into_item_restarted);
2135                 retval =
2136                     search_for_position_by_key(th->t_super, key,
2137                                                search_path);
2138                 if (retval == IO_ERROR) {
2139                         retval = -EIO;
2140                         goto error_out;
2141                 }
2142                 if (retval == POSITION_FOUND) {
2143                         reiserfs_warning(inode->i_sb, "PAP-5710",
2144                                          "entry or pasted byte (%K) exists",
2145                                          key);
2146                         retval = -EEXIST;
2147                         goto error_out;
2148                 }
2149 #ifdef CONFIG_REISERFS_CHECK
2150                 check_research_for_paste(search_path, key);
2151 #endif
2152         }
2153
2154         /*
2155          * Perform balancing after all resources are collected by fix_nodes,
2156          * and accessing them will not risk triggering schedule.
2157          */
2158         if (retval == CARRY_ON) {
2159                 do_balance(&s_paste_balance, NULL /*ih */ , body, M_PASTE);
2160                 return 0;
2161         }
2162         retval = (retval == NO_DISK_SPACE) ? -ENOSPC : -EIO;
2163 error_out:
2164         /* this also releases the path */
2165         unfix_nodes(&s_paste_balance);
2166 #ifdef REISERQUOTA_DEBUG
2167         reiserfs_debug(inode->i_sb, REISERFS_DEBUG_CODE,
2168                        "reiserquota paste_into_item(): freeing %u id=%u type=%c",
2169                        pasted_size, inode->i_uid,
2170                        key2type(&key->on_disk_key));
2171 #endif
2172         depth = reiserfs_write_unlock_nested(sb);
2173         dquot_free_space_nodirty(inode, pasted_size);
2174         reiserfs_write_lock_nested(sb, depth);
2175         return retval;
2176 }
2177
2178 /*
2179  * Insert new item into the buffer at the path.
2180  * th   - active transaction handle
2181  * path - path to the inserted item
2182  * ih   - pointer to the item header to insert
2183  * body - pointer to the bytes to insert
2184  */
2185 int reiserfs_insert_item(struct reiserfs_transaction_handle *th,
2186                          struct treepath *path, const struct cpu_key *key,
2187                          struct item_head *ih, struct inode *inode,
2188                          const char *body)
2189 {
2190         struct tree_balance s_ins_balance;
2191         int retval;
2192         int fs_gen = 0;
2193         int quota_bytes = 0;
2194
2195         BUG_ON(!th->t_trans_id);
2196
2197         if (inode) {            /* Do we count quotas for item? */
2198                 int depth;
2199                 fs_gen = get_generation(inode->i_sb);
2200                 quota_bytes = ih_item_len(ih);
2201
2202                 /*
2203                  * hack so the quota code doesn't have to guess
2204                  * if the file has a tail, links are always tails,
2205                  * so there's no guessing needed
2206                  */
2207                 if (!S_ISLNK(inode->i_mode) && is_direct_le_ih(ih))
2208                         quota_bytes = inode->i_sb->s_blocksize + UNFM_P_SIZE;
2209 #ifdef REISERQUOTA_DEBUG
2210                 reiserfs_debug(inode->i_sb, REISERFS_DEBUG_CODE,
2211                                "reiserquota insert_item(): allocating %u id=%u type=%c",
2212                                quota_bytes, inode->i_uid, head2type(ih));
2213 #endif
2214                 /*
2215                  * We can't dirty inode here. It would be immediately
2216                  * written but appropriate stat item isn't inserted yet...
2217                  */
2218                 depth = reiserfs_write_unlock_nested(inode->i_sb);
2219                 retval = dquot_alloc_space_nodirty(inode, quota_bytes);
2220                 reiserfs_write_lock_nested(inode->i_sb, depth);
2221                 if (retval) {
2222                         pathrelse(path);
2223                         return retval;
2224                 }
2225         }
2226         init_tb_struct(th, &s_ins_balance, th->t_super, path,
2227                        IH_SIZE + ih_item_len(ih));
2228 #ifdef DISPLACE_NEW_PACKING_LOCALITIES
2229         s_ins_balance.key = key->on_disk_key;
2230 #endif
2231         /*
2232          * DQUOT_* can schedule, must check to be sure calling
2233          * fix_nodes is safe
2234          */
2235         if (inode && fs_changed(fs_gen, inode->i_sb)) {
2236                 goto search_again;
2237         }
2238
2239         while ((retval =
2240                 fix_nodes(M_INSERT, &s_ins_balance, ih,
2241                           body)) == REPEAT_SEARCH) {
2242 search_again:
2243                 /* file system changed while we were in the fix_nodes */
2244                 PROC_INFO_INC(th->t_super, insert_item_restarted);
2245                 retval = search_item(th->t_super, key, path);
2246                 if (retval == IO_ERROR) {
2247                         retval = -EIO;
2248                         goto error_out;
2249                 }
2250                 if (retval == ITEM_FOUND) {
2251                         reiserfs_warning(th->t_super, "PAP-5760",
2252                                          "key %K already exists in the tree",
2253                                          key);
2254                         retval = -EEXIST;
2255                         goto error_out;
2256                 }
2257         }
2258
2259         /* make balancing after all resources will be collected at a time */
2260         if (retval == CARRY_ON) {
2261                 do_balance(&s_ins_balance, ih, body, M_INSERT);
2262                 return 0;
2263         }
2264
2265         retval = (retval == NO_DISK_SPACE) ? -ENOSPC : -EIO;
2266 error_out:
2267         /* also releases the path */
2268         unfix_nodes(&s_ins_balance);
2269 #ifdef REISERQUOTA_DEBUG
2270         if (inode)
2271                 reiserfs_debug(th->t_super, REISERFS_DEBUG_CODE,
2272                        "reiserquota insert_item(): freeing %u id=%u type=%c",
2273                        quota_bytes, inode->i_uid, head2type(ih));
2274 #endif
2275         if (inode) {
2276                 int depth = reiserfs_write_unlock_nested(inode->i_sb);
2277                 dquot_free_space_nodirty(inode, quota_bytes);
2278                 reiserfs_write_lock_nested(inode->i_sb, depth);
2279         }
2280         return retval;
2281 }