Merge master.kernel.org:/pub/scm/linux/kernel/git/davej/cpufreq
[linux-2.6] / 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 /*
12  *  This file contains functions dealing with S+tree
13  *
14  * B_IS_IN_TREE
15  * copy_item_head
16  * comp_short_keys
17  * comp_keys
18  * comp_short_le_keys
19  * le_key2cpu_key
20  * comp_le_keys
21  * bin_search
22  * get_lkey
23  * get_rkey
24  * key_in_buffer
25  * decrement_bcount
26  * decrement_counters_in_path
27  * reiserfs_check_path
28  * pathrelse_and_restore
29  * pathrelse
30  * search_by_key_reada
31  * search_by_key
32  * search_for_position_by_key
33  * comp_items
34  * prepare_for_direct_item
35  * prepare_for_direntry_item
36  * prepare_for_delete_or_cut
37  * calc_deleted_bytes_number
38  * init_tb_struct
39  * padd_item
40  * reiserfs_delete_item
41  * reiserfs_delete_solid_item
42  * reiserfs_delete_object
43  * maybe_indirect_to_direct
44  * indirect_to_direct_roll_back
45  * reiserfs_cut_from_item
46  * truncate_directory
47  * reiserfs_do_truncate
48  * reiserfs_paste_into_item
49  * reiserfs_insert_item
50  */
51
52 #include <linux/time.h>
53 #include <linux/string.h>
54 #include <linux/pagemap.h>
55 #include <linux/reiserfs_fs.h>
56 #include <linux/smp_lock.h>
57 #include <linux/buffer_head.h>
58 #include <linux/quotaops.h>
59
60 /* Does the buffer contain a disk block which is in the tree. */
61 inline int B_IS_IN_TREE(const struct buffer_head *p_s_bh)
62 {
63
64         RFALSE(B_LEVEL(p_s_bh) > MAX_HEIGHT,
65                "PAP-1010: block (%b) has too big level (%z)", p_s_bh, p_s_bh);
66
67         return (B_LEVEL(p_s_bh) != FREE_LEVEL);
68 }
69
70 //
71 // to gets item head in le form
72 //
73 inline void copy_item_head(struct item_head *p_v_to,
74                            const struct item_head *p_v_from)
75 {
76         memcpy(p_v_to, p_v_from, IH_SIZE);
77 }
78
79 /* k1 is pointer to on-disk structure which is stored in little-endian
80    form. k2 is pointer to cpu variable. For key of items of the same
81    object this returns 0.
82    Returns: -1 if key1 < key2 
83    0 if key1 == key2
84    1 if key1 > key2 */
85 inline int comp_short_keys(const struct reiserfs_key *le_key,
86                            const struct cpu_key *cpu_key)
87 {
88         __u32 n;
89         n = le32_to_cpu(le_key->k_dir_id);
90         if (n < cpu_key->on_disk_key.k_dir_id)
91                 return -1;
92         if (n > cpu_key->on_disk_key.k_dir_id)
93                 return 1;
94         n = le32_to_cpu(le_key->k_objectid);
95         if (n < cpu_key->on_disk_key.k_objectid)
96                 return -1;
97         if (n > cpu_key->on_disk_key.k_objectid)
98                 return 1;
99         return 0;
100 }
101
102 /* k1 is pointer to on-disk structure which is stored in little-endian
103    form. k2 is pointer to cpu variable.
104    Compare keys using all 4 key fields.
105    Returns: -1 if key1 < key2 0
106    if key1 = key2 1 if key1 > key2 */
107 static inline int comp_keys(const struct reiserfs_key *le_key,
108                             const struct cpu_key *cpu_key)
109 {
110         int retval;
111
112         retval = comp_short_keys(le_key, cpu_key);
113         if (retval)
114                 return retval;
115         if (le_key_k_offset(le_key_version(le_key), le_key) <
116             cpu_key_k_offset(cpu_key))
117                 return -1;
118         if (le_key_k_offset(le_key_version(le_key), le_key) >
119             cpu_key_k_offset(cpu_key))
120                 return 1;
121
122         if (cpu_key->key_length == 3)
123                 return 0;
124
125         /* this part is needed only when tail conversion is in progress */
126         if (le_key_k_type(le_key_version(le_key), le_key) <
127             cpu_key_k_type(cpu_key))
128                 return -1;
129
130         if (le_key_k_type(le_key_version(le_key), le_key) >
131             cpu_key_k_type(cpu_key))
132                 return 1;
133
134         return 0;
135 }
136
137 inline int comp_short_le_keys(const struct reiserfs_key *key1,
138                               const struct reiserfs_key *key2)
139 {
140         __u32 *p_s_1_u32, *p_s_2_u32;
141         int n_key_length = REISERFS_SHORT_KEY_LEN;
142
143         p_s_1_u32 = (__u32 *) key1;
144         p_s_2_u32 = (__u32 *) key2;
145         for (; n_key_length--; ++p_s_1_u32, ++p_s_2_u32) {
146                 if (le32_to_cpu(*p_s_1_u32) < le32_to_cpu(*p_s_2_u32))
147                         return -1;
148                 if (le32_to_cpu(*p_s_1_u32) > le32_to_cpu(*p_s_2_u32))
149                         return 1;
150         }
151         return 0;
152 }
153
154 inline void le_key2cpu_key(struct cpu_key *to, const struct reiserfs_key *from)
155 {
156         int version;
157         to->on_disk_key.k_dir_id = le32_to_cpu(from->k_dir_id);
158         to->on_disk_key.k_objectid = le32_to_cpu(from->k_objectid);
159
160         // find out version of the key
161         version = le_key_version(from);
162         to->version = version;
163         to->on_disk_key.k_offset = le_key_k_offset(version, from);
164         to->on_disk_key.k_type = le_key_k_type(version, from);
165 }
166
167 // this does not say which one is bigger, it only returns 1 if keys
168 // are not equal, 0 otherwise
169 inline int comp_le_keys(const struct reiserfs_key *k1,
170                         const struct reiserfs_key *k2)
171 {
172         return memcmp(k1, k2, sizeof(struct reiserfs_key));
173 }
174
175 /**************************************************************************
176  *  Binary search toolkit function                                        *
177  *  Search for an item in the array by the item key                       *
178  *  Returns:    1 if found,  0 if not found;                              *
179  *        *p_n_pos = number of the searched element if found, else the    *
180  *        number of the first element that is larger than p_v_key.        *
181  **************************************************************************/
182 /* For those not familiar with binary search: n_lbound is the leftmost item that it
183  could be, n_rbound the rightmost item that it could be.  We examine the item
184  halfway between n_lbound and n_rbound, and that tells us either that we can increase
185  n_lbound, or decrease n_rbound, or that we have found it, or if n_lbound <= n_rbound that
186  there are no possible items, and we have not found it. With each examination we
187  cut the number of possible items it could be by one more than half rounded down,
188  or we find it. */
189 static inline int bin_search(const void *p_v_key,       /* Key to search for.                   */
190                              const void *p_v_base,      /* First item in the array.             */
191                              int p_n_num,       /* Number of items in the array.        */
192                              int p_n_width,     /* Item size in the array.
193                                                    searched. Lest the reader be
194                                                    confused, note that this is crafted
195                                                    as a general function, and when it
196                                                    is applied specifically to the array
197                                                    of item headers in a node, p_n_width
198                                                    is actually the item header size not
199                                                    the item size.                      */
200                              int *p_n_pos       /* Number of the searched for element. */
201     )
202 {
203         int n_rbound, n_lbound, n_j;
204
205         for (n_j = ((n_rbound = p_n_num - 1) + (n_lbound = 0)) / 2;
206              n_lbound <= n_rbound; n_j = (n_rbound + n_lbound) / 2)
207                 switch (comp_keys
208                         ((struct reiserfs_key *)((char *)p_v_base +
209                                                  n_j * p_n_width),
210                          (struct cpu_key *)p_v_key)) {
211                 case -1:
212                         n_lbound = n_j + 1;
213                         continue;
214                 case 1:
215                         n_rbound = n_j - 1;
216                         continue;
217                 case 0:
218                         *p_n_pos = n_j;
219                         return ITEM_FOUND;      /* Key found in the array.  */
220                 }
221
222         /* bin_search did not find given key, it returns position of key,
223            that is minimal and greater than the given one. */
224         *p_n_pos = n_lbound;
225         return ITEM_NOT_FOUND;
226 }
227
228 #ifdef CONFIG_REISERFS_CHECK
229 extern struct tree_balance *cur_tb;
230 #endif
231
232 /* Minimal possible key. It is never in the tree. */
233 const struct reiserfs_key MIN_KEY = { 0, 0, {{0, 0},} };
234
235 /* Maximal possible key. It is never in the tree. */
236 static const struct reiserfs_key MAX_KEY = {
237         __constant_cpu_to_le32(0xffffffff),
238         __constant_cpu_to_le32(0xffffffff),
239         {{__constant_cpu_to_le32(0xffffffff),
240           __constant_cpu_to_le32(0xffffffff)},}
241 };
242
243 /* Get delimiting key of the buffer by looking for it in the buffers in the path, starting from the bottom
244    of the path, and going upwards.  We must check the path's validity at each step.  If the key is not in
245    the path, there is no delimiting key in the tree (buffer is first or last buffer in tree), and in this
246    case we return a special key, either MIN_KEY or MAX_KEY. */
247 static inline const struct reiserfs_key *get_lkey(const struct treepath
248                                                   *p_s_chk_path,
249                                                   const struct super_block
250                                                   *p_s_sb)
251 {
252         int n_position, n_path_offset = p_s_chk_path->path_length;
253         struct buffer_head *p_s_parent;
254
255         RFALSE(n_path_offset < FIRST_PATH_ELEMENT_OFFSET,
256                "PAP-5010: invalid offset in the path");
257
258         /* While not higher in path than first element. */
259         while (n_path_offset-- > FIRST_PATH_ELEMENT_OFFSET) {
260
261                 RFALSE(!buffer_uptodate
262                        (PATH_OFFSET_PBUFFER(p_s_chk_path, n_path_offset)),
263                        "PAP-5020: parent is not uptodate");
264
265                 /* Parent at the path is not in the tree now. */
266                 if (!B_IS_IN_TREE
267                     (p_s_parent =
268                      PATH_OFFSET_PBUFFER(p_s_chk_path, n_path_offset)))
269                         return &MAX_KEY;
270                 /* Check whether position in the parent is correct. */
271                 if ((n_position =
272                      PATH_OFFSET_POSITION(p_s_chk_path,
273                                           n_path_offset)) >
274                     B_NR_ITEMS(p_s_parent))
275                         return &MAX_KEY;
276                 /* Check whether parent at the path really points to the child. */
277                 if (B_N_CHILD_NUM(p_s_parent, n_position) !=
278                     PATH_OFFSET_PBUFFER(p_s_chk_path,
279                                         n_path_offset + 1)->b_blocknr)
280                         return &MAX_KEY;
281                 /* Return delimiting key if position in the parent is not equal to zero. */
282                 if (n_position)
283                         return B_N_PDELIM_KEY(p_s_parent, n_position - 1);
284         }
285         /* Return MIN_KEY if we are in the root of the buffer tree. */
286         if (PATH_OFFSET_PBUFFER(p_s_chk_path, FIRST_PATH_ELEMENT_OFFSET)->
287             b_blocknr == SB_ROOT_BLOCK(p_s_sb))
288                 return &MIN_KEY;
289         return &MAX_KEY;
290 }
291
292 /* Get delimiting key of the buffer at the path and its right neighbor. */
293 inline const struct reiserfs_key *get_rkey(const struct treepath *p_s_chk_path,
294                                            const struct super_block *p_s_sb)
295 {
296         int n_position, n_path_offset = p_s_chk_path->path_length;
297         struct buffer_head *p_s_parent;
298
299         RFALSE(n_path_offset < FIRST_PATH_ELEMENT_OFFSET,
300                "PAP-5030: invalid offset in the path");
301
302         while (n_path_offset-- > FIRST_PATH_ELEMENT_OFFSET) {
303
304                 RFALSE(!buffer_uptodate
305                        (PATH_OFFSET_PBUFFER(p_s_chk_path, n_path_offset)),
306                        "PAP-5040: parent is not uptodate");
307
308                 /* Parent at the path is not in the tree now. */
309                 if (!B_IS_IN_TREE
310                     (p_s_parent =
311                      PATH_OFFSET_PBUFFER(p_s_chk_path, n_path_offset)))
312                         return &MIN_KEY;
313                 /* Check whether position in the parent is correct. */
314                 if ((n_position =
315                      PATH_OFFSET_POSITION(p_s_chk_path,
316                                           n_path_offset)) >
317                     B_NR_ITEMS(p_s_parent))
318                         return &MIN_KEY;
319                 /* Check whether parent at the path really points to the child. */
320                 if (B_N_CHILD_NUM(p_s_parent, n_position) !=
321                     PATH_OFFSET_PBUFFER(p_s_chk_path,
322                                         n_path_offset + 1)->b_blocknr)
323                         return &MIN_KEY;
324                 /* Return delimiting key if position in the parent is not the last one. */
325                 if (n_position != B_NR_ITEMS(p_s_parent))
326                         return B_N_PDELIM_KEY(p_s_parent, n_position);
327         }
328         /* Return MAX_KEY if we are in the root of the buffer tree. */
329         if (PATH_OFFSET_PBUFFER(p_s_chk_path, FIRST_PATH_ELEMENT_OFFSET)->
330             b_blocknr == SB_ROOT_BLOCK(p_s_sb))
331                 return &MAX_KEY;
332         return &MIN_KEY;
333 }
334
335 /* Check whether a key is contained in the tree rooted from a buffer at a path. */
336 /* This works by looking at the left and right delimiting keys for the buffer in the last path_element in
337    the path.  These delimiting keys are stored at least one level above that buffer in the tree. If the
338    buffer is the first or last node in the tree order then one of the delimiting keys may be absent, and in
339    this case get_lkey and get_rkey return a special key which is MIN_KEY or MAX_KEY. */
340 static inline int key_in_buffer(struct treepath *p_s_chk_path,  /* Path which should be checked.  */
341                                 const struct cpu_key *p_s_key,  /* Key which should be checked.   */
342                                 struct super_block *p_s_sb      /* Super block pointer.           */
343     )
344 {
345
346         RFALSE(!p_s_key || p_s_chk_path->path_length < FIRST_PATH_ELEMENT_OFFSET
347                || p_s_chk_path->path_length > MAX_HEIGHT,
348                "PAP-5050: pointer to the key(%p) is NULL or invalid path length(%d)",
349                p_s_key, p_s_chk_path->path_length);
350         RFALSE(!PATH_PLAST_BUFFER(p_s_chk_path)->b_bdev,
351                "PAP-5060: device must not be NODEV");
352
353         if (comp_keys(get_lkey(p_s_chk_path, p_s_sb), p_s_key) == 1)
354                 /* left delimiting key is bigger, that the key we look for */
355                 return 0;
356         //  if ( comp_keys(p_s_key, get_rkey(p_s_chk_path, p_s_sb)) != -1 )
357         if (comp_keys(get_rkey(p_s_chk_path, p_s_sb), p_s_key) != 1)
358                 /* p_s_key must be less than right delimitiing key */
359                 return 0;
360         return 1;
361 }
362
363 inline void decrement_bcount(struct buffer_head *p_s_bh)
364 {
365         if (p_s_bh) {
366                 if (atomic_read(&(p_s_bh->b_count))) {
367                         put_bh(p_s_bh);
368                         return;
369                 }
370                 reiserfs_panic(NULL,
371                                "PAP-5070: decrement_bcount: trying to free free buffer %b",
372                                p_s_bh);
373         }
374 }
375
376 /* Decrement b_count field of the all buffers in the path. */
377 void decrement_counters_in_path(struct treepath *p_s_search_path)
378 {
379         int n_path_offset = p_s_search_path->path_length;
380
381         RFALSE(n_path_offset < ILLEGAL_PATH_ELEMENT_OFFSET ||
382                n_path_offset > EXTENDED_MAX_HEIGHT - 1,
383                "PAP-5080: invalid path offset of %d", n_path_offset);
384
385         while (n_path_offset > ILLEGAL_PATH_ELEMENT_OFFSET) {
386                 struct buffer_head *bh;
387
388                 bh = PATH_OFFSET_PBUFFER(p_s_search_path, n_path_offset--);
389                 decrement_bcount(bh);
390         }
391         p_s_search_path->path_length = ILLEGAL_PATH_ELEMENT_OFFSET;
392 }
393
394 int reiserfs_check_path(struct treepath *p)
395 {
396         RFALSE(p->path_length != ILLEGAL_PATH_ELEMENT_OFFSET,
397                "path not properly relsed");
398         return 0;
399 }
400
401 /* Release all buffers in the path. Restore dirty bits clean
402 ** when preparing the buffer for the log
403 **
404 ** only called from fix_nodes()
405 */
406 void pathrelse_and_restore(struct super_block *s, struct treepath *p_s_search_path)
407 {
408         int n_path_offset = p_s_search_path->path_length;
409
410         RFALSE(n_path_offset < ILLEGAL_PATH_ELEMENT_OFFSET,
411                "clm-4000: invalid path offset");
412
413         while (n_path_offset > ILLEGAL_PATH_ELEMENT_OFFSET) {
414                 reiserfs_restore_prepared_buffer(s,
415                                                  PATH_OFFSET_PBUFFER
416                                                  (p_s_search_path,
417                                                   n_path_offset));
418                 brelse(PATH_OFFSET_PBUFFER(p_s_search_path, n_path_offset--));
419         }
420         p_s_search_path->path_length = ILLEGAL_PATH_ELEMENT_OFFSET;
421 }
422
423 /* Release all buffers in the path. */
424 void pathrelse(struct treepath *p_s_search_path)
425 {
426         int n_path_offset = p_s_search_path->path_length;
427
428         RFALSE(n_path_offset < ILLEGAL_PATH_ELEMENT_OFFSET,
429                "PAP-5090: invalid path offset");
430
431         while (n_path_offset > ILLEGAL_PATH_ELEMENT_OFFSET)
432                 brelse(PATH_OFFSET_PBUFFER(p_s_search_path, n_path_offset--));
433
434         p_s_search_path->path_length = ILLEGAL_PATH_ELEMENT_OFFSET;
435 }
436
437 static int is_leaf(char *buf, int blocksize, struct buffer_head *bh)
438 {
439         struct block_head *blkh;
440         struct item_head *ih;
441         int used_space;
442         int prev_location;
443         int i;
444         int nr;
445
446         blkh = (struct block_head *)buf;
447         if (blkh_level(blkh) != DISK_LEAF_NODE_LEVEL) {
448                 reiserfs_warning(NULL,
449                                  "is_leaf: this should be caught earlier");
450                 return 0;
451         }
452
453         nr = blkh_nr_item(blkh);
454         if (nr < 1 || nr > ((blocksize - BLKH_SIZE) / (IH_SIZE + MIN_ITEM_LEN))) {
455                 /* item number is too big or too small */
456                 reiserfs_warning(NULL, "is_leaf: nr_item seems wrong: %z", bh);
457                 return 0;
458         }
459         ih = (struct item_head *)(buf + BLKH_SIZE) + nr - 1;
460         used_space = BLKH_SIZE + IH_SIZE * nr + (blocksize - ih_location(ih));
461         if (used_space != blocksize - blkh_free_space(blkh)) {
462                 /* free space does not match to calculated amount of use space */
463                 reiserfs_warning(NULL, "is_leaf: free space seems wrong: %z",
464                                  bh);
465                 return 0;
466         }
467         // FIXME: it is_leaf will hit performance too much - we may have
468         // return 1 here
469
470         /* check tables of item heads */
471         ih = (struct item_head *)(buf + BLKH_SIZE);
472         prev_location = blocksize;
473         for (i = 0; i < nr; i++, ih++) {
474                 if (le_ih_k_type(ih) == TYPE_ANY) {
475                         reiserfs_warning(NULL,
476                                          "is_leaf: wrong item type for item %h",
477                                          ih);
478                         return 0;
479                 }
480                 if (ih_location(ih) >= blocksize
481                     || ih_location(ih) < IH_SIZE * nr) {
482                         reiserfs_warning(NULL,
483                                          "is_leaf: item location seems wrong: %h",
484                                          ih);
485                         return 0;
486                 }
487                 if (ih_item_len(ih) < 1
488                     || ih_item_len(ih) > MAX_ITEM_LEN(blocksize)) {
489                         reiserfs_warning(NULL,
490                                          "is_leaf: item length seems wrong: %h",
491                                          ih);
492                         return 0;
493                 }
494                 if (prev_location - ih_location(ih) != ih_item_len(ih)) {
495                         reiserfs_warning(NULL,
496                                          "is_leaf: item location seems wrong (second one): %h",
497                                          ih);
498                         return 0;
499                 }
500                 prev_location = ih_location(ih);
501         }
502
503         // one may imagine much more checks
504         return 1;
505 }
506
507 /* returns 1 if buf looks like an internal node, 0 otherwise */
508 static int is_internal(char *buf, int blocksize, struct buffer_head *bh)
509 {
510         struct block_head *blkh;
511         int nr;
512         int used_space;
513
514         blkh = (struct block_head *)buf;
515         nr = blkh_level(blkh);
516         if (nr <= DISK_LEAF_NODE_LEVEL || nr > MAX_HEIGHT) {
517                 /* this level is not possible for internal nodes */
518                 reiserfs_warning(NULL,
519                                  "is_internal: this should be caught earlier");
520                 return 0;
521         }
522
523         nr = blkh_nr_item(blkh);
524         if (nr > (blocksize - BLKH_SIZE - DC_SIZE) / (KEY_SIZE + DC_SIZE)) {
525                 /* for internal which is not root we might check min number of keys */
526                 reiserfs_warning(NULL,
527                                  "is_internal: number of key seems wrong: %z",
528                                  bh);
529                 return 0;
530         }
531
532         used_space = BLKH_SIZE + KEY_SIZE * nr + DC_SIZE * (nr + 1);
533         if (used_space != blocksize - blkh_free_space(blkh)) {
534                 reiserfs_warning(NULL,
535                                  "is_internal: free space seems wrong: %z", bh);
536                 return 0;
537         }
538         // one may imagine much more checks
539         return 1;
540 }
541
542 // make sure that bh contains formatted node of reiserfs tree of
543 // 'level'-th level
544 static int is_tree_node(struct buffer_head *bh, int level)
545 {
546         if (B_LEVEL(bh) != level) {
547                 reiserfs_warning(NULL,
548                                  "is_tree_node: node level %d does not match to the expected one %d",
549                                  B_LEVEL(bh), level);
550                 return 0;
551         }
552         if (level == DISK_LEAF_NODE_LEVEL)
553                 return is_leaf(bh->b_data, bh->b_size, bh);
554
555         return is_internal(bh->b_data, bh->b_size, bh);
556 }
557
558 #define SEARCH_BY_KEY_READA 16
559
560 /* The function is NOT SCHEDULE-SAFE! */
561 static void search_by_key_reada(struct super_block *s,
562                                 struct buffer_head **bh,
563                                 unsigned long *b, int num)
564 {
565         int i, j;
566
567         for (i = 0; i < num; i++) {
568                 bh[i] = sb_getblk(s, b[i]);
569         }
570         for (j = 0; j < i; j++) {
571                 /*
572                  * note, this needs attention if we are getting rid of the BKL
573                  * you have to make sure the prepared bit isn't set on this buffer
574                  */
575                 if (!buffer_uptodate(bh[j]))
576                         ll_rw_block(READA, 1, bh + j);
577                 brelse(bh[j]);
578         }
579 }
580
581 /**************************************************************************
582  * Algorithm   SearchByKey                                                *
583  *             look for item in the Disk S+Tree by its key                *
584  * Input:  p_s_sb   -  super block                                        *
585  *         p_s_key  - pointer to the key to search                        *
586  * Output: ITEM_FOUND, ITEM_NOT_FOUND or IO_ERROR                         *
587  *         p_s_search_path - path from the root to the needed leaf        *
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 /* The function is NOT SCHEDULE-SAFE! */
604 int search_by_key(struct super_block *p_s_sb, const struct cpu_key *p_s_key,    /* Key to search. */
605                   struct treepath *p_s_search_path,/* This structure was
606                                                    allocated and initialized
607                                                    by the calling
608                                                    function. It is filled up
609                                                    by this function.  */
610                   int n_stop_level      /* How far down the tree to search. To
611                                            stop at leaf level - set to
612                                            DISK_LEAF_NODE_LEVEL */
613     )
614 {
615         int n_block_number;
616         int expected_level;
617         struct buffer_head *p_s_bh;
618         struct path_element *p_s_last_element;
619         int n_node_level, n_retval;
620         int right_neighbor_of_leaf_node;
621         int fs_gen;
622         struct buffer_head *reada_bh[SEARCH_BY_KEY_READA];
623         unsigned long reada_blocks[SEARCH_BY_KEY_READA];
624         int reada_count = 0;
625
626 #ifdef CONFIG_REISERFS_CHECK
627         int n_repeat_counter = 0;
628 #endif
629
630         PROC_INFO_INC(p_s_sb, search_by_key);
631
632         /* As we add each node to a path we increase its count.  This means that
633            we must be careful to release all nodes in a path before we either
634            discard the path struct or re-use the path struct, as we do here. */
635
636         decrement_counters_in_path(p_s_search_path);
637
638         right_neighbor_of_leaf_node = 0;
639
640         /* With each iteration of this loop we search through the items in the
641            current node, and calculate the next current node(next path element)
642            for the next iteration of this loop.. */
643         n_block_number = SB_ROOT_BLOCK(p_s_sb);
644         expected_level = -1;
645         while (1) {
646
647 #ifdef CONFIG_REISERFS_CHECK
648                 if (!(++n_repeat_counter % 50000))
649                         reiserfs_warning(p_s_sb, "PAP-5100: search_by_key: %s:"
650                                          "there were %d iterations of while loop "
651                                          "looking for key %K",
652                                          current->comm, n_repeat_counter,
653                                          p_s_key);
654 #endif
655
656                 /* prep path to have another element added to it. */
657                 p_s_last_element =
658                     PATH_OFFSET_PELEMENT(p_s_search_path,
659                                          ++p_s_search_path->path_length);
660                 fs_gen = get_generation(p_s_sb);
661
662                 /* Read the next tree node, and set the last element in the path to
663                    have a pointer to it. */
664                 if ((p_s_bh = p_s_last_element->pe_buffer =
665                      sb_getblk(p_s_sb, n_block_number))) {
666                         if (!buffer_uptodate(p_s_bh) && reada_count > 1) {
667                                 search_by_key_reada(p_s_sb, reada_bh,
668                                                     reada_blocks, reada_count);
669                         }
670                         ll_rw_block(READ, 1, &p_s_bh);
671                         wait_on_buffer(p_s_bh);
672                         if (!buffer_uptodate(p_s_bh))
673                                 goto io_error;
674                 } else {
675                       io_error:
676                         p_s_search_path->path_length--;
677                         pathrelse(p_s_search_path);
678                         return IO_ERROR;
679                 }
680                 reada_count = 0;
681                 if (expected_level == -1)
682                         expected_level = SB_TREE_HEIGHT(p_s_sb);
683                 expected_level--;
684
685                 /* It is possible that schedule occurred. We must check whether the key
686                    to search is still in the tree rooted from the current buffer. If
687                    not then repeat search from the root. */
688                 if (fs_changed(fs_gen, p_s_sb) &&
689                     (!B_IS_IN_TREE(p_s_bh) ||
690                      B_LEVEL(p_s_bh) != expected_level ||
691                      !key_in_buffer(p_s_search_path, p_s_key, p_s_sb))) {
692                         PROC_INFO_INC(p_s_sb, search_by_key_fs_changed);
693                         PROC_INFO_INC(p_s_sb, search_by_key_restarted);
694                         PROC_INFO_INC(p_s_sb,
695                                       sbk_restarted[expected_level - 1]);
696                         decrement_counters_in_path(p_s_search_path);
697
698                         /* Get the root block number so that we can repeat the search
699                            starting from the root. */
700                         n_block_number = SB_ROOT_BLOCK(p_s_sb);
701                         expected_level = -1;
702                         right_neighbor_of_leaf_node = 0;
703
704                         /* repeat search from the root */
705                         continue;
706                 }
707
708                 /* only check that the key is in the buffer if p_s_key is not
709                    equal to the MAX_KEY. Latter case is only possible in
710                    "finish_unfinished()" processing during mount. */
711                 RFALSE(comp_keys(&MAX_KEY, p_s_key) &&
712                        !key_in_buffer(p_s_search_path, p_s_key, p_s_sb),
713                        "PAP-5130: key is not in the buffer");
714 #ifdef CONFIG_REISERFS_CHECK
715                 if (cur_tb) {
716                         print_cur_tb("5140");
717                         reiserfs_panic(p_s_sb,
718                                        "PAP-5140: search_by_key: schedule occurred in do_balance!");
719                 }
720 #endif
721
722                 // make sure, that the node contents look like a node of
723                 // certain level
724                 if (!is_tree_node(p_s_bh, expected_level)) {
725                         reiserfs_warning(p_s_sb, "vs-5150: search_by_key: "
726                                          "invalid format found in block %ld. Fsck?",
727                                          p_s_bh->b_blocknr);
728                         pathrelse(p_s_search_path);
729                         return IO_ERROR;
730                 }
731
732                 /* ok, we have acquired next formatted node in the tree */
733                 n_node_level = B_LEVEL(p_s_bh);
734
735                 PROC_INFO_BH_STAT(p_s_sb, p_s_bh, n_node_level - 1);
736
737                 RFALSE(n_node_level < n_stop_level,
738                        "vs-5152: tree level (%d) is less than stop level (%d)",
739                        n_node_level, n_stop_level);
740
741                 n_retval = bin_search(p_s_key, B_N_PITEM_HEAD(p_s_bh, 0),
742                                       B_NR_ITEMS(p_s_bh),
743                                       (n_node_level ==
744                                        DISK_LEAF_NODE_LEVEL) ? IH_SIZE :
745                                       KEY_SIZE,
746                                       &(p_s_last_element->pe_position));
747                 if (n_node_level == n_stop_level) {
748                         return n_retval;
749                 }
750
751                 /* we are not in the stop level */
752                 if (n_retval == ITEM_FOUND)
753                         /* item has been found, so we choose the pointer which is to the right of the found one */
754                         p_s_last_element->pe_position++;
755
756                 /* if item was not found we choose the position which is to
757                    the left of the found item. This requires no code,
758                    bin_search did it already. */
759
760                 /* So we have chosen a position in the current node which is
761                    an internal node.  Now we calculate child block number by
762                    position in the node. */
763                 n_block_number =
764                     B_N_CHILD_NUM(p_s_bh, p_s_last_element->pe_position);
765
766                 /* if we are going to read leaf nodes, try for read ahead as well */
767                 if ((p_s_search_path->reada & PATH_READA) &&
768                     n_node_level == DISK_LEAF_NODE_LEVEL + 1) {
769                         int pos = p_s_last_element->pe_position;
770                         int limit = B_NR_ITEMS(p_s_bh);
771                         struct reiserfs_key *le_key;
772
773                         if (p_s_search_path->reada & PATH_READA_BACK)
774                                 limit = 0;
775                         while (reada_count < SEARCH_BY_KEY_READA) {
776                                 if (pos == limit)
777                                         break;
778                                 reada_blocks[reada_count++] =
779                                     B_N_CHILD_NUM(p_s_bh, pos);
780                                 if (p_s_search_path->reada & PATH_READA_BACK)
781                                         pos--;
782                                 else
783                                         pos++;
784
785                                 /*
786                                  * check to make sure we're in the same object
787                                  */
788                                 le_key = B_N_PDELIM_KEY(p_s_bh, pos);
789                                 if (le32_to_cpu(le_key->k_objectid) !=
790                                     p_s_key->on_disk_key.k_objectid) {
791                                         break;
792                                 }
793                         }
794                 }
795         }
796 }
797
798 /* Form the path to an item and position in this item which contains
799    file byte defined by p_s_key. If there is no such item
800    corresponding to the key, we point the path to the item with
801    maximal key less than p_s_key, and *p_n_pos_in_item is set to one
802    past the last entry/byte in the item.  If searching for entry in a
803    directory item, and it is not found, *p_n_pos_in_item is set to one
804    entry more than the entry with maximal key which is less than the
805    sought key.
806
807    Note that if there is no entry in this same node which is one more,
808    then we point to an imaginary entry.  for direct items, the
809    position is in units of bytes, for indirect items the position is
810    in units of blocknr entries, for directory items the position is in
811    units of directory entries.  */
812
813 /* The function is NOT SCHEDULE-SAFE! */
814 int search_for_position_by_key(struct super_block *p_s_sb,      /* Pointer to the super block.          */
815                                const struct cpu_key *p_cpu_key, /* Key to search (cpu variable)         */
816                                struct treepath *p_s_search_path /* Filled up by this function.          */
817     )
818 {
819         struct item_head *p_le_ih;      /* pointer to on-disk structure */
820         int n_blk_size;
821         loff_t item_offset, offset;
822         struct reiserfs_dir_entry de;
823         int retval;
824
825         /* If searching for directory entry. */
826         if (is_direntry_cpu_key(p_cpu_key))
827                 return search_by_entry_key(p_s_sb, p_cpu_key, p_s_search_path,
828                                            &de);
829
830         /* If not searching for directory entry. */
831
832         /* If item is found. */
833         retval = search_item(p_s_sb, p_cpu_key, p_s_search_path);
834         if (retval == IO_ERROR)
835                 return retval;
836         if (retval == ITEM_FOUND) {
837
838                 RFALSE(!ih_item_len
839                        (B_N_PITEM_HEAD
840                         (PATH_PLAST_BUFFER(p_s_search_path),
841                          PATH_LAST_POSITION(p_s_search_path))),
842                        "PAP-5165: item length equals zero");
843
844                 pos_in_item(p_s_search_path) = 0;
845                 return POSITION_FOUND;
846         }
847
848         RFALSE(!PATH_LAST_POSITION(p_s_search_path),
849                "PAP-5170: position equals zero");
850
851         /* Item is not found. Set path to the previous item. */
852         p_le_ih =
853             B_N_PITEM_HEAD(PATH_PLAST_BUFFER(p_s_search_path),
854                            --PATH_LAST_POSITION(p_s_search_path));
855         n_blk_size = p_s_sb->s_blocksize;
856
857         if (comp_short_keys(&(p_le_ih->ih_key), p_cpu_key)) {
858                 return FILE_NOT_FOUND;
859         }
860         // FIXME: quite ugly this far
861
862         item_offset = le_ih_k_offset(p_le_ih);
863         offset = cpu_key_k_offset(p_cpu_key);
864
865         /* Needed byte is contained in the item pointed to by the path. */
866         if (item_offset <= offset &&
867             item_offset + op_bytes_number(p_le_ih, n_blk_size) > offset) {
868                 pos_in_item(p_s_search_path) = offset - item_offset;
869                 if (is_indirect_le_ih(p_le_ih)) {
870                         pos_in_item(p_s_search_path) /= n_blk_size;
871                 }
872                 return POSITION_FOUND;
873         }
874
875         /* Needed byte is not contained in the item pointed to by the
876            path. Set pos_in_item out of the item. */
877         if (is_indirect_le_ih(p_le_ih))
878                 pos_in_item(p_s_search_path) =
879                     ih_item_len(p_le_ih) / UNFM_P_SIZE;
880         else
881                 pos_in_item(p_s_search_path) = ih_item_len(p_le_ih);
882
883         return POSITION_NOT_FOUND;
884 }
885
886 /* Compare given item and item pointed to by the path. */
887 int comp_items(const struct item_head *stored_ih, const struct treepath *p_s_path)
888 {
889         struct buffer_head *p_s_bh;
890         struct item_head *ih;
891
892         /* Last buffer at the path is not in the tree. */
893         if (!B_IS_IN_TREE(p_s_bh = PATH_PLAST_BUFFER(p_s_path)))
894                 return 1;
895
896         /* Last path position is invalid. */
897         if (PATH_LAST_POSITION(p_s_path) >= B_NR_ITEMS(p_s_bh))
898                 return 1;
899
900         /* we need only to know, whether it is the same item */
901         ih = get_ih(p_s_path);
902         return memcmp(stored_ih, ih, IH_SIZE);
903 }
904
905 /* unformatted nodes are not logged anymore, ever.  This is safe
906 ** now
907 */
908 #define held_by_others(bh) (atomic_read(&(bh)->b_count) > 1)
909
910 // block can not be forgotten as it is in I/O or held by someone
911 #define block_in_use(bh) (buffer_locked(bh) || (held_by_others(bh)))
912
913 // prepare for delete or cut of direct item
914 static inline int prepare_for_direct_item(struct treepath *path,
915                                           struct item_head *le_ih,
916                                           struct inode *inode,
917                                           loff_t new_file_length, int *cut_size)
918 {
919         loff_t round_len;
920
921         if (new_file_length == max_reiserfs_offset(inode)) {
922                 /* item has to be deleted */
923                 *cut_size = -(IH_SIZE + ih_item_len(le_ih));
924                 return M_DELETE;
925         }
926         // new file gets truncated
927         if (get_inode_item_key_version(inode) == KEY_FORMAT_3_6) {
928                 // 
929                 round_len = ROUND_UP(new_file_length);
930                 /* this was n_new_file_length < le_ih ... */
931                 if (round_len < le_ih_k_offset(le_ih)) {
932                         *cut_size = -(IH_SIZE + ih_item_len(le_ih));
933                         return M_DELETE;        /* Delete this item. */
934                 }
935                 /* Calculate first position and size for cutting from item. */
936                 pos_in_item(path) = round_len - (le_ih_k_offset(le_ih) - 1);
937                 *cut_size = -(ih_item_len(le_ih) - pos_in_item(path));
938
939                 return M_CUT;   /* Cut from this item. */
940         }
941
942         // old file: items may have any length
943
944         if (new_file_length < le_ih_k_offset(le_ih)) {
945                 *cut_size = -(IH_SIZE + ih_item_len(le_ih));
946                 return M_DELETE;        /* Delete this item. */
947         }
948         /* Calculate first position and size for cutting from item. */
949         *cut_size = -(ih_item_len(le_ih) -
950                       (pos_in_item(path) =
951                        new_file_length + 1 - le_ih_k_offset(le_ih)));
952         return M_CUT;           /* Cut from this item. */
953 }
954
955 static inline int prepare_for_direntry_item(struct treepath *path,
956                                             struct item_head *le_ih,
957                                             struct inode *inode,
958                                             loff_t new_file_length,
959                                             int *cut_size)
960 {
961         if (le_ih_k_offset(le_ih) == DOT_OFFSET &&
962             new_file_length == max_reiserfs_offset(inode)) {
963                 RFALSE(ih_entry_count(le_ih) != 2,
964                        "PAP-5220: incorrect empty directory item (%h)", le_ih);
965                 *cut_size = -(IH_SIZE + ih_item_len(le_ih));
966                 return M_DELETE;        /* Delete the directory item containing "." and ".." entry. */
967         }
968
969         if (ih_entry_count(le_ih) == 1) {
970                 /* Delete the directory item such as there is one record only
971                    in this item */
972                 *cut_size = -(IH_SIZE + ih_item_len(le_ih));
973                 return M_DELETE;
974         }
975
976         /* Cut one record from the directory item. */
977         *cut_size =
978             -(DEH_SIZE +
979               entry_length(get_last_bh(path), le_ih, pos_in_item(path)));
980         return M_CUT;
981 }
982
983 #define JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD (2 * JOURNAL_PER_BALANCE_CNT + 1)
984
985 /*  If the path points to a directory or direct item, calculate mode and the size cut, for balance.
986     If the path points to an indirect item, remove some number of its unformatted nodes.
987     In case of file truncate calculate whether this item must be deleted/truncated or last
988     unformatted node of this item will be converted to a direct item.
989     This function returns a determination of what balance mode the calling function should employ. */
990 static char prepare_for_delete_or_cut(struct reiserfs_transaction_handle *th, struct inode *inode, struct treepath *p_s_path, const struct cpu_key *p_s_item_key, int *p_n_removed,     /* Number of unformatted nodes which were removed
991                                                                                                                                                                                    from end of the file. */
992                                       int *p_n_cut_size, unsigned long long n_new_file_length   /* MAX_KEY_OFFSET in case of delete. */
993     )
994 {
995         struct super_block *p_s_sb = inode->i_sb;
996         struct item_head *p_le_ih = PATH_PITEM_HEAD(p_s_path);
997         struct buffer_head *p_s_bh = PATH_PLAST_BUFFER(p_s_path);
998
999         BUG_ON(!th->t_trans_id);
1000
1001         /* Stat_data item. */
1002         if (is_statdata_le_ih(p_le_ih)) {
1003
1004                 RFALSE(n_new_file_length != max_reiserfs_offset(inode),
1005                        "PAP-5210: mode must be M_DELETE");
1006
1007                 *p_n_cut_size = -(IH_SIZE + ih_item_len(p_le_ih));
1008                 return M_DELETE;
1009         }
1010
1011         /* Directory item. */
1012         if (is_direntry_le_ih(p_le_ih))
1013                 return prepare_for_direntry_item(p_s_path, p_le_ih, inode,
1014                                                  n_new_file_length,
1015                                                  p_n_cut_size);
1016
1017         /* Direct item. */
1018         if (is_direct_le_ih(p_le_ih))
1019                 return prepare_for_direct_item(p_s_path, p_le_ih, inode,
1020                                                n_new_file_length, p_n_cut_size);
1021
1022         /* Case of an indirect item. */
1023         {
1024             int blk_size = p_s_sb->s_blocksize;
1025             struct item_head s_ih;
1026             int need_re_search;
1027             int delete = 0;
1028             int result = M_CUT;
1029             int pos = 0;
1030
1031             if ( n_new_file_length == max_reiserfs_offset (inode) ) {
1032                 /* prepare_for_delete_or_cut() is called by
1033                  * reiserfs_delete_item() */
1034                 n_new_file_length = 0;
1035                 delete = 1;
1036             }
1037
1038             do {
1039                 need_re_search = 0;
1040                 *p_n_cut_size = 0;
1041                 p_s_bh = PATH_PLAST_BUFFER(p_s_path);
1042                 copy_item_head(&s_ih, PATH_PITEM_HEAD(p_s_path));
1043                 pos = I_UNFM_NUM(&s_ih);
1044
1045                 while (le_ih_k_offset (&s_ih) + (pos - 1) * blk_size > n_new_file_length) {
1046                     __u32 *unfm, block;
1047
1048                     /* Each unformatted block deletion may involve one additional
1049                      * bitmap block into the transaction, thereby the initial
1050                      * journal space reservation might not be enough. */
1051                     if (!delete && (*p_n_cut_size) != 0 &&
1052                         reiserfs_transaction_free_space(th) < JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD) {
1053                         break;
1054                     }
1055
1056                     unfm = (__u32 *)B_I_PITEM(p_s_bh, &s_ih) + pos - 1;
1057                     block = get_block_num(unfm, 0);
1058
1059                     if (block != 0) {
1060                         reiserfs_prepare_for_journal(p_s_sb, p_s_bh, 1);
1061                         put_block_num(unfm, 0, 0);
1062                         journal_mark_dirty (th, p_s_sb, p_s_bh);
1063                         reiserfs_free_block(th, inode, block, 1);
1064                     }
1065
1066                     cond_resched();
1067
1068                     if (item_moved (&s_ih, p_s_path))  {
1069                         need_re_search = 1;
1070                         break;
1071                     }
1072
1073                     pos --;
1074                     (*p_n_removed) ++;
1075                     (*p_n_cut_size) -= UNFM_P_SIZE;
1076
1077                     if (pos == 0) {
1078                         (*p_n_cut_size) -= IH_SIZE;
1079                         result = M_DELETE;
1080                         break;
1081                     }
1082                 }
1083                 /* a trick.  If the buffer has been logged, this will do nothing.  If
1084                 ** we've broken the loop without logging it, it will restore the
1085                 ** buffer */
1086                 reiserfs_restore_prepared_buffer(p_s_sb, p_s_bh);
1087             } while (need_re_search &&
1088                      search_for_position_by_key(p_s_sb, p_s_item_key, p_s_path) == POSITION_FOUND);
1089             pos_in_item(p_s_path) = pos * UNFM_P_SIZE;
1090
1091             if (*p_n_cut_size == 0) {
1092                 /* Nothing were cut. maybe convert last unformatted node to the
1093                  * direct item? */
1094                 result = M_CONVERT;
1095             }
1096             return result;
1097         }
1098 }
1099
1100 /* Calculate number of bytes which will be deleted or cut during balance */
1101 static int calc_deleted_bytes_number(struct tree_balance *p_s_tb, char c_mode)
1102 {
1103         int n_del_size;
1104         struct item_head *p_le_ih = PATH_PITEM_HEAD(p_s_tb->tb_path);
1105
1106         if (is_statdata_le_ih(p_le_ih))
1107                 return 0;
1108
1109         n_del_size =
1110             (c_mode ==
1111              M_DELETE) ? ih_item_len(p_le_ih) : -p_s_tb->insert_size[0];
1112         if (is_direntry_le_ih(p_le_ih)) {
1113                 // return EMPTY_DIR_SIZE; /* We delete emty directoris only. */
1114                 // we can't use EMPTY_DIR_SIZE, as old format dirs have a different
1115                 // empty size.  ick. FIXME, is this right?
1116                 //
1117                 return n_del_size;
1118         }
1119
1120         if (is_indirect_le_ih(p_le_ih))
1121                 n_del_size = (n_del_size / UNFM_P_SIZE) * (PATH_PLAST_BUFFER(p_s_tb->tb_path)->b_size); // - get_ih_free_space (p_le_ih);
1122         return n_del_size;
1123 }
1124
1125 static void init_tb_struct(struct reiserfs_transaction_handle *th,
1126                            struct tree_balance *p_s_tb,
1127                            struct super_block *p_s_sb,
1128                            struct treepath *p_s_path, int n_size)
1129 {
1130
1131         BUG_ON(!th->t_trans_id);
1132
1133         memset(p_s_tb, '\0', sizeof(struct tree_balance));
1134         p_s_tb->transaction_handle = th;
1135         p_s_tb->tb_sb = p_s_sb;
1136         p_s_tb->tb_path = p_s_path;
1137         PATH_OFFSET_PBUFFER(p_s_path, ILLEGAL_PATH_ELEMENT_OFFSET) = NULL;
1138         PATH_OFFSET_POSITION(p_s_path, ILLEGAL_PATH_ELEMENT_OFFSET) = 0;
1139         p_s_tb->insert_size[0] = n_size;
1140 }
1141
1142 void padd_item(char *item, int total_length, int length)
1143 {
1144         int i;
1145
1146         for (i = total_length; i > length;)
1147                 item[--i] = 0;
1148 }
1149
1150 #ifdef REISERQUOTA_DEBUG
1151 char key2type(struct reiserfs_key *ih)
1152 {
1153         if (is_direntry_le_key(2, ih))
1154                 return 'd';
1155         if (is_direct_le_key(2, ih))
1156                 return 'D';
1157         if (is_indirect_le_key(2, ih))
1158                 return 'i';
1159         if (is_statdata_le_key(2, ih))
1160                 return 's';
1161         return 'u';
1162 }
1163
1164 char head2type(struct item_head *ih)
1165 {
1166         if (is_direntry_le_ih(ih))
1167                 return 'd';
1168         if (is_direct_le_ih(ih))
1169                 return 'D';
1170         if (is_indirect_le_ih(ih))
1171                 return 'i';
1172         if (is_statdata_le_ih(ih))
1173                 return 's';
1174         return 'u';
1175 }
1176 #endif
1177
1178 /* Delete object item. */
1179 int reiserfs_delete_item(struct reiserfs_transaction_handle *th, struct treepath *p_s_path,     /* Path to the deleted item. */
1180                          const struct cpu_key *p_s_item_key,    /* Key to search for the deleted item.  */
1181                          struct inode *p_s_inode,       /* inode is here just to update i_blocks and quotas */
1182                          struct buffer_head *p_s_un_bh)
1183 {                               /* NULL or unformatted node pointer.    */
1184         struct super_block *p_s_sb = p_s_inode->i_sb;
1185         struct tree_balance s_del_balance;
1186         struct item_head s_ih;
1187         struct item_head *q_ih;
1188         int quota_cut_bytes;
1189         int n_ret_value, n_del_size, n_removed;
1190
1191 #ifdef CONFIG_REISERFS_CHECK
1192         char c_mode;
1193         int n_iter = 0;
1194 #endif
1195
1196         BUG_ON(!th->t_trans_id);
1197
1198         init_tb_struct(th, &s_del_balance, p_s_sb, p_s_path,
1199                        0 /*size is unknown */ );
1200
1201         while (1) {
1202                 n_removed = 0;
1203
1204 #ifdef CONFIG_REISERFS_CHECK
1205                 n_iter++;
1206                 c_mode =
1207 #endif
1208                     prepare_for_delete_or_cut(th, p_s_inode, p_s_path,
1209                                               p_s_item_key, &n_removed,
1210                                               &n_del_size,
1211                                               max_reiserfs_offset(p_s_inode));
1212
1213                 RFALSE(c_mode != M_DELETE, "PAP-5320: mode must be M_DELETE");
1214
1215                 copy_item_head(&s_ih, PATH_PITEM_HEAD(p_s_path));
1216                 s_del_balance.insert_size[0] = n_del_size;
1217
1218                 n_ret_value = fix_nodes(M_DELETE, &s_del_balance, NULL, NULL);
1219                 if (n_ret_value != REPEAT_SEARCH)
1220                         break;
1221
1222                 PROC_INFO_INC(p_s_sb, delete_item_restarted);
1223
1224                 // file system changed, repeat search
1225                 n_ret_value =
1226                     search_for_position_by_key(p_s_sb, p_s_item_key, p_s_path);
1227                 if (n_ret_value == IO_ERROR)
1228                         break;
1229                 if (n_ret_value == FILE_NOT_FOUND) {
1230                         reiserfs_warning(p_s_sb,
1231                                          "vs-5340: reiserfs_delete_item: "
1232                                          "no items of the file %K found",
1233                                          p_s_item_key);
1234                         break;
1235                 }
1236         }                       /* while (1) */
1237
1238         if (n_ret_value != CARRY_ON) {
1239                 unfix_nodes(&s_del_balance);
1240                 return 0;
1241         }
1242         // reiserfs_delete_item returns item length when success
1243         n_ret_value = calc_deleted_bytes_number(&s_del_balance, M_DELETE);
1244         q_ih = get_ih(p_s_path);
1245         quota_cut_bytes = ih_item_len(q_ih);
1246
1247         /* hack so the quota code doesn't have to guess if the file
1248          ** has a tail.  On tail insert, we allocate quota for 1 unformatted node.
1249          ** We test the offset because the tail might have been
1250          ** split into multiple items, and we only want to decrement for
1251          ** the unfm node once
1252          */
1253         if (!S_ISLNK(p_s_inode->i_mode) && is_direct_le_ih(q_ih)) {
1254                 if ((le_ih_k_offset(q_ih) & (p_s_sb->s_blocksize - 1)) == 1) {
1255                         quota_cut_bytes = p_s_sb->s_blocksize + UNFM_P_SIZE;
1256                 } else {
1257                         quota_cut_bytes = 0;
1258                 }
1259         }
1260
1261         if (p_s_un_bh) {
1262                 int off;
1263                 char *data;
1264
1265                 /* We are in direct2indirect conversion, so move tail contents
1266                    to the unformatted node */
1267                 /* note, we do the copy before preparing the buffer because we
1268                  ** don't care about the contents of the unformatted node yet.
1269                  ** the only thing we really care about is the direct item's data
1270                  ** is in the unformatted node.
1271                  **
1272                  ** Otherwise, we would have to call reiserfs_prepare_for_journal on
1273                  ** the unformatted node, which might schedule, meaning we'd have to
1274                  ** loop all the way back up to the start of the while loop.
1275                  **
1276                  ** The unformatted node must be dirtied later on.  We can't be
1277                  ** sure here if the entire tail has been deleted yet.
1278                  **
1279                  ** p_s_un_bh is from the page cache (all unformatted nodes are
1280                  ** from the page cache) and might be a highmem page.  So, we
1281                  ** can't use p_s_un_bh->b_data.
1282                  ** -clm
1283                  */
1284
1285                 data = kmap_atomic(p_s_un_bh->b_page, KM_USER0);
1286                 off = ((le_ih_k_offset(&s_ih) - 1) & (PAGE_CACHE_SIZE - 1));
1287                 memcpy(data + off,
1288                        B_I_PITEM(PATH_PLAST_BUFFER(p_s_path), &s_ih),
1289                        n_ret_value);
1290                 kunmap_atomic(data, KM_USER0);
1291         }
1292         /* Perform balancing after all resources have been collected at once. */
1293         do_balance(&s_del_balance, NULL, NULL, M_DELETE);
1294
1295 #ifdef REISERQUOTA_DEBUG
1296         reiserfs_debug(p_s_sb, REISERFS_DEBUG_CODE,
1297                        "reiserquota delete_item(): freeing %u, id=%u type=%c",
1298                        quota_cut_bytes, p_s_inode->i_uid, head2type(&s_ih));
1299 #endif
1300         DQUOT_FREE_SPACE_NODIRTY(p_s_inode, quota_cut_bytes);
1301
1302         /* Return deleted body length */
1303         return n_ret_value;
1304 }
1305
1306 /* Summary Of Mechanisms For Handling Collisions Between Processes:
1307
1308  deletion of the body of the object is performed by iput(), with the
1309  result that if multiple processes are operating on a file, the
1310  deletion of the body of the file is deferred until the last process
1311  that has an open inode performs its iput().
1312
1313  writes and truncates are protected from collisions by use of
1314  semaphores.
1315
1316  creates, linking, and mknod are protected from collisions with other
1317  processes by making the reiserfs_add_entry() the last step in the
1318  creation, and then rolling back all changes if there was a collision.
1319  - Hans
1320 */
1321
1322 /* this deletes item which never gets split */
1323 void reiserfs_delete_solid_item(struct reiserfs_transaction_handle *th,
1324                                 struct inode *inode, struct reiserfs_key *key)
1325 {
1326         struct tree_balance tb;
1327         INITIALIZE_PATH(path);
1328         int item_len = 0;
1329         int tb_init = 0;
1330         struct cpu_key cpu_key;
1331         int retval;
1332         int quota_cut_bytes = 0;
1333
1334         BUG_ON(!th->t_trans_id);
1335
1336         le_key2cpu_key(&cpu_key, key);
1337
1338         while (1) {
1339                 retval = search_item(th->t_super, &cpu_key, &path);
1340                 if (retval == IO_ERROR) {
1341                         reiserfs_warning(th->t_super,
1342                                          "vs-5350: reiserfs_delete_solid_item: "
1343                                          "i/o failure occurred trying to delete %K",
1344                                          &cpu_key);
1345                         break;
1346                 }
1347                 if (retval != ITEM_FOUND) {
1348                         pathrelse(&path);
1349                         // No need for a warning, if there is just no free space to insert '..' item into the newly-created subdir
1350                         if (!
1351                             ((unsigned long long)
1352                              GET_HASH_VALUE(le_key_k_offset
1353                                             (le_key_version(key), key)) == 0
1354                              && (unsigned long long)
1355                              GET_GENERATION_NUMBER(le_key_k_offset
1356                                                    (le_key_version(key),
1357                                                     key)) == 1))
1358                                 reiserfs_warning(th->t_super,
1359                                                  "vs-5355: reiserfs_delete_solid_item: %k not found",
1360                                                  key);
1361                         break;
1362                 }
1363                 if (!tb_init) {
1364                         tb_init = 1;
1365                         item_len = ih_item_len(PATH_PITEM_HEAD(&path));
1366                         init_tb_struct(th, &tb, th->t_super, &path,
1367                                        -(IH_SIZE + item_len));
1368                 }
1369                 quota_cut_bytes = ih_item_len(PATH_PITEM_HEAD(&path));
1370
1371                 retval = fix_nodes(M_DELETE, &tb, NULL, NULL);
1372                 if (retval == REPEAT_SEARCH) {
1373                         PROC_INFO_INC(th->t_super, delete_solid_item_restarted);
1374                         continue;
1375                 }
1376
1377                 if (retval == CARRY_ON) {
1378                         do_balance(&tb, NULL, NULL, M_DELETE);
1379                         if (inode) {    /* Should we count quota for item? (we don't count quotas for save-links) */
1380 #ifdef REISERQUOTA_DEBUG
1381                                 reiserfs_debug(th->t_super, REISERFS_DEBUG_CODE,
1382                                                "reiserquota delete_solid_item(): freeing %u id=%u type=%c",
1383                                                quota_cut_bytes, inode->i_uid,
1384                                                key2type(key));
1385 #endif
1386                                 DQUOT_FREE_SPACE_NODIRTY(inode,
1387                                                          quota_cut_bytes);
1388                         }
1389                         break;
1390                 }
1391                 // IO_ERROR, NO_DISK_SPACE, etc
1392                 reiserfs_warning(th->t_super,
1393                                  "vs-5360: reiserfs_delete_solid_item: "
1394                                  "could not delete %K due to fix_nodes failure",
1395                                  &cpu_key);
1396                 unfix_nodes(&tb);
1397                 break;
1398         }
1399
1400         reiserfs_check_path(&path);
1401 }
1402
1403 int reiserfs_delete_object(struct reiserfs_transaction_handle *th,
1404                            struct inode *inode)
1405 {
1406         int err;
1407         inode->i_size = 0;
1408         BUG_ON(!th->t_trans_id);
1409
1410         /* for directory this deletes item containing "." and ".." */
1411         err =
1412             reiserfs_do_truncate(th, inode, NULL, 0 /*no timestamp updates */ );
1413         if (err)
1414                 return err;
1415
1416 #if defined( USE_INODE_GENERATION_COUNTER )
1417         if (!old_format_only(th->t_super)) {
1418                 __le32 *inode_generation;
1419
1420                 inode_generation =
1421                     &REISERFS_SB(th->t_super)->s_rs->s_inode_generation;
1422                 *inode_generation =
1423                     cpu_to_le32(le32_to_cpu(*inode_generation) + 1);
1424         }
1425 /* USE_INODE_GENERATION_COUNTER */
1426 #endif
1427         reiserfs_delete_solid_item(th, inode, INODE_PKEY(inode));
1428
1429         return err;
1430 }
1431
1432 static void unmap_buffers(struct page *page, loff_t pos)
1433 {
1434         struct buffer_head *bh;
1435         struct buffer_head *head;
1436         struct buffer_head *next;
1437         unsigned long tail_index;
1438         unsigned long cur_index;
1439
1440         if (page) {
1441                 if (page_has_buffers(page)) {
1442                         tail_index = pos & (PAGE_CACHE_SIZE - 1);
1443                         cur_index = 0;
1444                         head = page_buffers(page);
1445                         bh = head;
1446                         do {
1447                                 next = bh->b_this_page;
1448
1449                                 /* we want to unmap the buffers that contain the tail, and
1450                                  ** all the buffers after it (since the tail must be at the
1451                                  ** end of the file).  We don't want to unmap file data
1452                                  ** before the tail, since it might be dirty and waiting to
1453                                  ** reach disk
1454                                  */
1455                                 cur_index += bh->b_size;
1456                                 if (cur_index > tail_index) {
1457                                         reiserfs_unmap_buffer(bh);
1458                                 }
1459                                 bh = next;
1460                         } while (bh != head);
1461                         if (PAGE_SIZE == bh->b_size) {
1462                                 cancel_dirty_page(page, PAGE_CACHE_SIZE);
1463                         }
1464                 }
1465         }
1466 }
1467
1468 static int maybe_indirect_to_direct(struct reiserfs_transaction_handle *th,
1469                                     struct inode *p_s_inode,
1470                                     struct page *page,
1471                                     struct treepath *p_s_path,
1472                                     const struct cpu_key *p_s_item_key,
1473                                     loff_t n_new_file_size, char *p_c_mode)
1474 {
1475         struct super_block *p_s_sb = p_s_inode->i_sb;
1476         int n_block_size = p_s_sb->s_blocksize;
1477         int cut_bytes;
1478         BUG_ON(!th->t_trans_id);
1479         BUG_ON(n_new_file_size != p_s_inode->i_size);
1480
1481         /* the page being sent in could be NULL if there was an i/o error
1482          ** reading in the last block.  The user will hit problems trying to
1483          ** read the file, but for now we just skip the indirect2direct
1484          */
1485         if (atomic_read(&p_s_inode->i_count) > 1 ||
1486             !tail_has_to_be_packed(p_s_inode) ||
1487             !page || (REISERFS_I(p_s_inode)->i_flags & i_nopack_mask)) {
1488                 // leave tail in an unformatted node    
1489                 *p_c_mode = M_SKIP_BALANCING;
1490                 cut_bytes =
1491                     n_block_size - (n_new_file_size & (n_block_size - 1));
1492                 pathrelse(p_s_path);
1493                 return cut_bytes;
1494         }
1495         /* Permorm the conversion to a direct_item. */
1496         /*return indirect_to_direct (p_s_inode, p_s_path, p_s_item_key, n_new_file_size, p_c_mode); */
1497         return indirect2direct(th, p_s_inode, page, p_s_path, p_s_item_key,
1498                                n_new_file_size, p_c_mode);
1499 }
1500
1501 /* we did indirect_to_direct conversion. And we have inserted direct
1502    item successesfully, but there were no disk space to cut unfm
1503    pointer being converted. Therefore we have to delete inserted
1504    direct item(s) */
1505 static void indirect_to_direct_roll_back(struct reiserfs_transaction_handle *th,
1506                                          struct inode *inode, struct treepath *path)
1507 {
1508         struct cpu_key tail_key;
1509         int tail_len;
1510         int removed;
1511         BUG_ON(!th->t_trans_id);
1512
1513         make_cpu_key(&tail_key, inode, inode->i_size + 1, TYPE_DIRECT, 4);      // !!!!
1514         tail_key.key_length = 4;
1515
1516         tail_len =
1517             (cpu_key_k_offset(&tail_key) & (inode->i_sb->s_blocksize - 1)) - 1;
1518         while (tail_len) {
1519                 /* look for the last byte of the tail */
1520                 if (search_for_position_by_key(inode->i_sb, &tail_key, path) ==
1521                     POSITION_NOT_FOUND)
1522                         reiserfs_panic(inode->i_sb,
1523                                        "vs-5615: indirect_to_direct_roll_back: found invalid item");
1524                 RFALSE(path->pos_in_item !=
1525                        ih_item_len(PATH_PITEM_HEAD(path)) - 1,
1526                        "vs-5616: appended bytes found");
1527                 PATH_LAST_POSITION(path)--;
1528
1529                 removed =
1530                     reiserfs_delete_item(th, path, &tail_key, inode,
1531                                          NULL /*unbh not needed */ );
1532                 RFALSE(removed <= 0
1533                        || removed > tail_len,
1534                        "vs-5617: there was tail %d bytes, removed item length %d bytes",
1535                        tail_len, removed);
1536                 tail_len -= removed;
1537                 set_cpu_key_k_offset(&tail_key,
1538                                      cpu_key_k_offset(&tail_key) - removed);
1539         }
1540         reiserfs_warning(inode->i_sb,
1541                          "indirect_to_direct_roll_back: indirect_to_direct conversion has been rolled back due to lack of disk space");
1542         //mark_file_without_tail (inode);
1543         mark_inode_dirty(inode);
1544 }
1545
1546 /* (Truncate or cut entry) or delete object item. Returns < 0 on failure */
1547 int reiserfs_cut_from_item(struct reiserfs_transaction_handle *th,
1548                            struct treepath *p_s_path,
1549                            struct cpu_key *p_s_item_key,
1550                            struct inode *p_s_inode,
1551                            struct page *page, loff_t n_new_file_size)
1552 {
1553         struct super_block *p_s_sb = p_s_inode->i_sb;
1554         /* Every function which is going to call do_balance must first
1555            create a tree_balance structure.  Then it must fill up this
1556            structure by using the init_tb_struct and fix_nodes functions.
1557            After that we can make tree balancing. */
1558         struct tree_balance s_cut_balance;
1559         struct item_head *p_le_ih;
1560         int n_cut_size = 0,     /* Amount to be cut. */
1561             n_ret_value = CARRY_ON, n_removed = 0,      /* Number of the removed unformatted nodes. */
1562             n_is_inode_locked = 0;
1563         char c_mode;            /* Mode of the balance. */
1564         int retval2 = -1;
1565         int quota_cut_bytes;
1566         loff_t tail_pos = 0;
1567
1568         BUG_ON(!th->t_trans_id);
1569
1570         init_tb_struct(th, &s_cut_balance, p_s_inode->i_sb, p_s_path,
1571                        n_cut_size);
1572
1573         /* Repeat this loop until we either cut the item without needing
1574            to balance, or we fix_nodes without schedule occurring */
1575         while (1) {
1576                 /* Determine the balance mode, position of the first byte to
1577                    be cut, and size to be cut.  In case of the indirect item
1578                    free unformatted nodes which are pointed to by the cut
1579                    pointers. */
1580
1581                 c_mode =
1582                     prepare_for_delete_or_cut(th, p_s_inode, p_s_path,
1583                                               p_s_item_key, &n_removed,
1584                                               &n_cut_size, n_new_file_size);
1585                 if (c_mode == M_CONVERT) {
1586                         /* convert last unformatted node to direct item or leave
1587                            tail in the unformatted node */
1588                         RFALSE(n_ret_value != CARRY_ON,
1589                                "PAP-5570: can not convert twice");
1590
1591                         n_ret_value =
1592                             maybe_indirect_to_direct(th, p_s_inode, page,
1593                                                      p_s_path, p_s_item_key,
1594                                                      n_new_file_size, &c_mode);
1595                         if (c_mode == M_SKIP_BALANCING)
1596                                 /* tail has been left in the unformatted node */
1597                                 return n_ret_value;
1598
1599                         n_is_inode_locked = 1;
1600
1601                         /* removing of last unformatted node will change value we
1602                            have to return to truncate. Save it */
1603                         retval2 = n_ret_value;
1604                         /*retval2 = p_s_sb->s_blocksize - (n_new_file_size & (p_s_sb->s_blocksize - 1)); */
1605
1606                         /* So, we have performed the first part of the conversion:
1607                            inserting the new direct item.  Now we are removing the
1608                            last unformatted node pointer. Set key to search for
1609                            it. */
1610                         set_cpu_key_k_type(p_s_item_key, TYPE_INDIRECT);
1611                         p_s_item_key->key_length = 4;
1612                         n_new_file_size -=
1613                             (n_new_file_size & (p_s_sb->s_blocksize - 1));
1614                         tail_pos = n_new_file_size;
1615                         set_cpu_key_k_offset(p_s_item_key, n_new_file_size + 1);
1616                         if (search_for_position_by_key
1617                             (p_s_sb, p_s_item_key,
1618                              p_s_path) == POSITION_NOT_FOUND) {
1619                                 print_block(PATH_PLAST_BUFFER(p_s_path), 3,
1620                                             PATH_LAST_POSITION(p_s_path) - 1,
1621                                             PATH_LAST_POSITION(p_s_path) + 1);
1622                                 reiserfs_panic(p_s_sb,
1623                                                "PAP-5580: reiserfs_cut_from_item: item to convert does not exist (%K)",
1624                                                p_s_item_key);
1625                         }
1626                         continue;
1627                 }
1628                 if (n_cut_size == 0) {
1629                         pathrelse(p_s_path);
1630                         return 0;
1631                 }
1632
1633                 s_cut_balance.insert_size[0] = n_cut_size;
1634
1635                 n_ret_value = fix_nodes(c_mode, &s_cut_balance, NULL, NULL);
1636                 if (n_ret_value != REPEAT_SEARCH)
1637                         break;
1638
1639                 PROC_INFO_INC(p_s_sb, cut_from_item_restarted);
1640
1641                 n_ret_value =
1642                     search_for_position_by_key(p_s_sb, p_s_item_key, p_s_path);
1643                 if (n_ret_value == POSITION_FOUND)
1644                         continue;
1645
1646                 reiserfs_warning(p_s_sb,
1647                                  "PAP-5610: reiserfs_cut_from_item: item %K not found",
1648                                  p_s_item_key);
1649                 unfix_nodes(&s_cut_balance);
1650                 return (n_ret_value == IO_ERROR) ? -EIO : -ENOENT;
1651         }                       /* while */
1652
1653         // check fix_nodes results (IO_ERROR or NO_DISK_SPACE)
1654         if (n_ret_value != CARRY_ON) {
1655                 if (n_is_inode_locked) {
1656                         // FIXME: this seems to be not needed: we are always able
1657                         // to cut item
1658                         indirect_to_direct_roll_back(th, p_s_inode, p_s_path);
1659                 }
1660                 if (n_ret_value == NO_DISK_SPACE)
1661                         reiserfs_warning(p_s_sb, "NO_DISK_SPACE");
1662                 unfix_nodes(&s_cut_balance);
1663                 return -EIO;
1664         }
1665
1666         /* go ahead and perform balancing */
1667
1668         RFALSE(c_mode == M_PASTE || c_mode == M_INSERT, "invalid mode");
1669
1670         /* Calculate number of bytes that need to be cut from the item. */
1671         quota_cut_bytes =
1672             (c_mode ==
1673              M_DELETE) ? ih_item_len(get_ih(p_s_path)) : -s_cut_balance.
1674             insert_size[0];
1675         if (retval2 == -1)
1676                 n_ret_value = calc_deleted_bytes_number(&s_cut_balance, c_mode);
1677         else
1678                 n_ret_value = retval2;
1679
1680         /* For direct items, we only change the quota when deleting the last
1681          ** item.
1682          */
1683         p_le_ih = PATH_PITEM_HEAD(s_cut_balance.tb_path);
1684         if (!S_ISLNK(p_s_inode->i_mode) && is_direct_le_ih(p_le_ih)) {
1685                 if (c_mode == M_DELETE &&
1686                     (le_ih_k_offset(p_le_ih) & (p_s_sb->s_blocksize - 1)) ==
1687                     1) {
1688                         // FIXME: this is to keep 3.5 happy
1689                         REISERFS_I(p_s_inode)->i_first_direct_byte = U32_MAX;
1690                         quota_cut_bytes = p_s_sb->s_blocksize + UNFM_P_SIZE;
1691                 } else {
1692                         quota_cut_bytes = 0;
1693                 }
1694         }
1695 #ifdef CONFIG_REISERFS_CHECK
1696         if (n_is_inode_locked) {
1697                 struct item_head *le_ih =
1698                     PATH_PITEM_HEAD(s_cut_balance.tb_path);
1699                 /* we are going to complete indirect2direct conversion. Make
1700                    sure, that we exactly remove last unformatted node pointer
1701                    of the item */
1702                 if (!is_indirect_le_ih(le_ih))
1703                         reiserfs_panic(p_s_sb,
1704                                        "vs-5652: reiserfs_cut_from_item: "
1705                                        "item must be indirect %h", le_ih);
1706
1707                 if (c_mode == M_DELETE && ih_item_len(le_ih) != UNFM_P_SIZE)
1708                         reiserfs_panic(p_s_sb,
1709                                        "vs-5653: reiserfs_cut_from_item: "
1710                                        "completing indirect2direct conversion indirect item %h "
1711                                        "being deleted must be of 4 byte long",
1712                                        le_ih);
1713
1714                 if (c_mode == M_CUT
1715                     && s_cut_balance.insert_size[0] != -UNFM_P_SIZE) {
1716                         reiserfs_panic(p_s_sb,
1717                                        "vs-5654: reiserfs_cut_from_item: "
1718                                        "can not complete indirect2direct conversion of %h (CUT, insert_size==%d)",
1719                                        le_ih, s_cut_balance.insert_size[0]);
1720                 }
1721                 /* it would be useful to make sure, that right neighboring
1722                    item is direct item of this file */
1723         }
1724 #endif
1725
1726         do_balance(&s_cut_balance, NULL, NULL, c_mode);
1727         if (n_is_inode_locked) {
1728                 /* we've done an indirect->direct conversion.  when the data block
1729                  ** was freed, it was removed from the list of blocks that must
1730                  ** be flushed before the transaction commits, make sure to
1731                  ** unmap and invalidate it
1732                  */
1733                 unmap_buffers(page, tail_pos);
1734                 REISERFS_I(p_s_inode)->i_flags &= ~i_pack_on_close_mask;
1735         }
1736 #ifdef REISERQUOTA_DEBUG
1737         reiserfs_debug(p_s_inode->i_sb, REISERFS_DEBUG_CODE,
1738                        "reiserquota cut_from_item(): freeing %u id=%u type=%c",
1739                        quota_cut_bytes, p_s_inode->i_uid, '?');
1740 #endif
1741         DQUOT_FREE_SPACE_NODIRTY(p_s_inode, quota_cut_bytes);
1742         return n_ret_value;
1743 }
1744
1745 static void truncate_directory(struct reiserfs_transaction_handle *th,
1746                                struct inode *inode)
1747 {
1748         BUG_ON(!th->t_trans_id);
1749         if (inode->i_nlink)
1750                 reiserfs_warning(inode->i_sb,
1751                                  "vs-5655: truncate_directory: link count != 0");
1752
1753         set_le_key_k_offset(KEY_FORMAT_3_5, INODE_PKEY(inode), DOT_OFFSET);
1754         set_le_key_k_type(KEY_FORMAT_3_5, INODE_PKEY(inode), TYPE_DIRENTRY);
1755         reiserfs_delete_solid_item(th, inode, INODE_PKEY(inode));
1756         reiserfs_update_sd(th, inode);
1757         set_le_key_k_offset(KEY_FORMAT_3_5, INODE_PKEY(inode), SD_OFFSET);
1758         set_le_key_k_type(KEY_FORMAT_3_5, INODE_PKEY(inode), TYPE_STAT_DATA);
1759 }
1760
1761 /* Truncate file to the new size. Note, this must be called with a transaction
1762    already started */
1763 int reiserfs_do_truncate(struct reiserfs_transaction_handle *th, struct inode *p_s_inode,       /* ->i_size contains new
1764                                                                                                    size */
1765                          struct page *page,     /* up to date for last block */
1766                          int update_timestamps  /* when it is called by
1767                                                    file_release to convert
1768                                                    the tail - no timestamps
1769                                                    should be updated */
1770     )
1771 {
1772         INITIALIZE_PATH(s_search_path); /* Path to the current object item. */
1773         struct item_head *p_le_ih;      /* Pointer to an item header. */
1774         struct cpu_key s_item_key;      /* Key to search for a previous file item. */
1775         loff_t n_file_size,     /* Old file size. */
1776          n_new_file_size;       /* New file size. */
1777         int n_deleted;          /* Number of deleted or truncated bytes. */
1778         int retval;
1779         int err = 0;
1780
1781         BUG_ON(!th->t_trans_id);
1782         if (!
1783             (S_ISREG(p_s_inode->i_mode) || S_ISDIR(p_s_inode->i_mode)
1784              || S_ISLNK(p_s_inode->i_mode)))
1785                 return 0;
1786
1787         if (S_ISDIR(p_s_inode->i_mode)) {
1788                 // deletion of directory - no need to update timestamps
1789                 truncate_directory(th, p_s_inode);
1790                 return 0;
1791         }
1792
1793         /* Get new file size. */
1794         n_new_file_size = p_s_inode->i_size;
1795
1796         // FIXME: note, that key type is unimportant here
1797         make_cpu_key(&s_item_key, p_s_inode, max_reiserfs_offset(p_s_inode),
1798                      TYPE_DIRECT, 3);
1799
1800         retval =
1801             search_for_position_by_key(p_s_inode->i_sb, &s_item_key,
1802                                        &s_search_path);
1803         if (retval == IO_ERROR) {
1804                 reiserfs_warning(p_s_inode->i_sb,
1805                                  "vs-5657: reiserfs_do_truncate: "
1806                                  "i/o failure occurred trying to truncate %K",
1807                                  &s_item_key);
1808                 err = -EIO;
1809                 goto out;
1810         }
1811         if (retval == POSITION_FOUND || retval == FILE_NOT_FOUND) {
1812                 reiserfs_warning(p_s_inode->i_sb,
1813                                  "PAP-5660: reiserfs_do_truncate: "
1814                                  "wrong result %d of search for %K", retval,
1815                                  &s_item_key);
1816
1817                 err = -EIO;
1818                 goto out;
1819         }
1820
1821         s_search_path.pos_in_item--;
1822
1823         /* Get real file size (total length of all file items) */
1824         p_le_ih = PATH_PITEM_HEAD(&s_search_path);
1825         if (is_statdata_le_ih(p_le_ih))
1826                 n_file_size = 0;
1827         else {
1828                 loff_t offset = le_ih_k_offset(p_le_ih);
1829                 int bytes =
1830                     op_bytes_number(p_le_ih, p_s_inode->i_sb->s_blocksize);
1831
1832                 /* this may mismatch with real file size: if last direct item
1833                    had no padding zeros and last unformatted node had no free
1834                    space, this file would have this file size */
1835                 n_file_size = offset + bytes - 1;
1836         }
1837         /*
1838          * are we doing a full truncate or delete, if so
1839          * kick in the reada code
1840          */
1841         if (n_new_file_size == 0)
1842                 s_search_path.reada = PATH_READA | PATH_READA_BACK;
1843
1844         if (n_file_size == 0 || n_file_size < n_new_file_size) {
1845                 goto update_and_out;
1846         }
1847
1848         /* Update key to search for the last file item. */
1849         set_cpu_key_k_offset(&s_item_key, n_file_size);
1850
1851         do {
1852                 /* Cut or delete file item. */
1853                 n_deleted =
1854                     reiserfs_cut_from_item(th, &s_search_path, &s_item_key,
1855                                            p_s_inode, page, n_new_file_size);
1856                 if (n_deleted < 0) {
1857                         reiserfs_warning(p_s_inode->i_sb,
1858                                          "vs-5665: reiserfs_do_truncate: reiserfs_cut_from_item failed");
1859                         reiserfs_check_path(&s_search_path);
1860                         return 0;
1861                 }
1862
1863                 RFALSE(n_deleted > n_file_size,
1864                        "PAP-5670: reiserfs_cut_from_item: too many bytes deleted: deleted %d, file_size %lu, item_key %K",
1865                        n_deleted, n_file_size, &s_item_key);
1866
1867                 /* Change key to search the last file item. */
1868                 n_file_size -= n_deleted;
1869
1870                 set_cpu_key_k_offset(&s_item_key, n_file_size);
1871
1872                 /* While there are bytes to truncate and previous file item is presented in the tree. */
1873
1874                 /*
1875                  ** This loop could take a really long time, and could log 
1876                  ** many more blocks than a transaction can hold.  So, we do a polite
1877                  ** journal end here, and if the transaction needs ending, we make
1878                  ** sure the file is consistent before ending the current trans
1879                  ** and starting a new one
1880                  */
1881                 if (journal_transaction_should_end(th, 0) ||
1882                     reiserfs_transaction_free_space(th) <= JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD) {
1883                         int orig_len_alloc = th->t_blocks_allocated;
1884                         decrement_counters_in_path(&s_search_path);
1885
1886                         if (update_timestamps) {
1887                                 p_s_inode->i_mtime = p_s_inode->i_ctime =
1888                                     CURRENT_TIME_SEC;
1889                         }
1890                         reiserfs_update_sd(th, p_s_inode);
1891
1892                         err = journal_end(th, p_s_inode->i_sb, orig_len_alloc);
1893                         if (err)
1894                                 goto out;
1895                         err = journal_begin(th, p_s_inode->i_sb,
1896                                             JOURNAL_FOR_FREE_BLOCK_AND_UPDATE_SD + JOURNAL_PER_BALANCE_CNT * 4) ;
1897                         if (err)
1898                                 goto out;
1899                         reiserfs_update_inode_transaction(p_s_inode);
1900                 }
1901         } while (n_file_size > ROUND_UP(n_new_file_size) &&
1902                  search_for_position_by_key(p_s_inode->i_sb, &s_item_key,
1903                                             &s_search_path) == POSITION_FOUND);
1904
1905         RFALSE(n_file_size > ROUND_UP(n_new_file_size),
1906                "PAP-5680: truncate did not finish: new_file_size %Ld, current %Ld, oid %d",
1907                n_new_file_size, n_file_size, s_item_key.on_disk_key.k_objectid);
1908
1909       update_and_out:
1910         if (update_timestamps) {
1911                 // this is truncate, not file closing
1912                 p_s_inode->i_mtime = p_s_inode->i_ctime = CURRENT_TIME_SEC;
1913         }
1914         reiserfs_update_sd(th, p_s_inode);
1915
1916       out:
1917         pathrelse(&s_search_path);
1918         return err;
1919 }
1920
1921 #ifdef CONFIG_REISERFS_CHECK
1922 // this makes sure, that we __append__, not overwrite or add holes
1923 static void check_research_for_paste(struct treepath *path,
1924                                      const struct cpu_key *p_s_key)
1925 {
1926         struct item_head *found_ih = get_ih(path);
1927
1928         if (is_direct_le_ih(found_ih)) {
1929                 if (le_ih_k_offset(found_ih) +
1930                     op_bytes_number(found_ih,
1931                                     get_last_bh(path)->b_size) !=
1932                     cpu_key_k_offset(p_s_key)
1933                     || op_bytes_number(found_ih,
1934                                        get_last_bh(path)->b_size) !=
1935                     pos_in_item(path))
1936                         reiserfs_panic(NULL,
1937                                        "PAP-5720: check_research_for_paste: "
1938                                        "found direct item %h or position (%d) does not match to key %K",
1939                                        found_ih, pos_in_item(path), p_s_key);
1940         }
1941         if (is_indirect_le_ih(found_ih)) {
1942                 if (le_ih_k_offset(found_ih) +
1943                     op_bytes_number(found_ih,
1944                                     get_last_bh(path)->b_size) !=
1945                     cpu_key_k_offset(p_s_key)
1946                     || I_UNFM_NUM(found_ih) != pos_in_item(path)
1947                     || get_ih_free_space(found_ih) != 0)
1948                         reiserfs_panic(NULL,
1949                                        "PAP-5730: check_research_for_paste: "
1950                                        "found indirect item (%h) or position (%d) does not match to key (%K)",
1951                                        found_ih, pos_in_item(path), p_s_key);
1952         }
1953 }
1954 #endif                          /* config reiserfs check */
1955
1956 /* Paste bytes to the existing item. Returns bytes number pasted into the item. */
1957 int reiserfs_paste_into_item(struct reiserfs_transaction_handle *th, struct treepath *p_s_search_path,  /* Path to the pasted item.          */
1958                              const struct cpu_key *p_s_key,     /* Key to search for the needed item. */
1959                              struct inode *inode,       /* Inode item belongs to */
1960                              const char *p_c_body,      /* Pointer to the bytes to paste.    */
1961                              int n_pasted_size)
1962 {                               /* Size of pasted bytes.             */
1963         struct tree_balance s_paste_balance;
1964         int retval;
1965         int fs_gen;
1966
1967         BUG_ON(!th->t_trans_id);
1968
1969         fs_gen = get_generation(inode->i_sb);
1970
1971 #ifdef REISERQUOTA_DEBUG
1972         reiserfs_debug(inode->i_sb, REISERFS_DEBUG_CODE,
1973                        "reiserquota paste_into_item(): allocating %u id=%u type=%c",
1974                        n_pasted_size, inode->i_uid,
1975                        key2type(&(p_s_key->on_disk_key)));
1976 #endif
1977
1978         if (DQUOT_ALLOC_SPACE_NODIRTY(inode, n_pasted_size)) {
1979                 pathrelse(p_s_search_path);
1980                 return -EDQUOT;
1981         }
1982         init_tb_struct(th, &s_paste_balance, th->t_super, p_s_search_path,
1983                        n_pasted_size);
1984 #ifdef DISPLACE_NEW_PACKING_LOCALITIES
1985         s_paste_balance.key = p_s_key->on_disk_key;
1986 #endif
1987
1988         /* DQUOT_* can schedule, must check before the fix_nodes */
1989         if (fs_changed(fs_gen, inode->i_sb)) {
1990                 goto search_again;
1991         }
1992
1993         while ((retval =
1994                 fix_nodes(M_PASTE, &s_paste_balance, NULL,
1995                           p_c_body)) == REPEAT_SEARCH) {
1996               search_again:
1997                 /* file system changed while we were in the fix_nodes */
1998                 PROC_INFO_INC(th->t_super, paste_into_item_restarted);
1999                 retval =
2000                     search_for_position_by_key(th->t_super, p_s_key,
2001                                                p_s_search_path);
2002                 if (retval == IO_ERROR) {
2003                         retval = -EIO;
2004                         goto error_out;
2005                 }
2006                 if (retval == POSITION_FOUND) {
2007                         reiserfs_warning(inode->i_sb,
2008                                          "PAP-5710: reiserfs_paste_into_item: entry or pasted byte (%K) exists",
2009                                          p_s_key);
2010                         retval = -EEXIST;
2011                         goto error_out;
2012                 }
2013 #ifdef CONFIG_REISERFS_CHECK
2014                 check_research_for_paste(p_s_search_path, p_s_key);
2015 #endif
2016         }
2017
2018         /* Perform balancing after all resources are collected by fix_nodes, and
2019            accessing them will not risk triggering schedule. */
2020         if (retval == CARRY_ON) {
2021                 do_balance(&s_paste_balance, NULL /*ih */ , p_c_body, M_PASTE);
2022                 return 0;
2023         }
2024         retval = (retval == NO_DISK_SPACE) ? -ENOSPC : -EIO;
2025       error_out:
2026         /* this also releases the path */
2027         unfix_nodes(&s_paste_balance);
2028 #ifdef REISERQUOTA_DEBUG
2029         reiserfs_debug(inode->i_sb, REISERFS_DEBUG_CODE,
2030                        "reiserquota paste_into_item(): freeing %u id=%u type=%c",
2031                        n_pasted_size, inode->i_uid,
2032                        key2type(&(p_s_key->on_disk_key)));
2033 #endif
2034         DQUOT_FREE_SPACE_NODIRTY(inode, n_pasted_size);
2035         return retval;
2036 }
2037
2038 /* Insert new item into the buffer at the path. */
2039 int reiserfs_insert_item(struct reiserfs_transaction_handle *th, struct treepath *p_s_path,     /* Path to the inserteded item.         */
2040                          const struct cpu_key *key, struct item_head *p_s_ih,   /* Pointer to the item header to insert. */
2041                          struct inode *inode, const char *p_c_body)
2042 {                               /* Pointer to the bytes to insert.      */
2043         struct tree_balance s_ins_balance;
2044         int retval;
2045         int fs_gen = 0;
2046         int quota_bytes = 0;
2047
2048         BUG_ON(!th->t_trans_id);
2049
2050         if (inode) {            /* Do we count quotas for item? */
2051                 fs_gen = get_generation(inode->i_sb);
2052                 quota_bytes = ih_item_len(p_s_ih);
2053
2054                 /* hack so the quota code doesn't have to guess if the file has
2055                  ** a tail, links are always tails, so there's no guessing needed
2056                  */
2057                 if (!S_ISLNK(inode->i_mode) && is_direct_le_ih(p_s_ih)) {
2058                         quota_bytes = inode->i_sb->s_blocksize + UNFM_P_SIZE;
2059                 }
2060 #ifdef REISERQUOTA_DEBUG
2061                 reiserfs_debug(inode->i_sb, REISERFS_DEBUG_CODE,
2062                                "reiserquota insert_item(): allocating %u id=%u type=%c",
2063                                quota_bytes, inode->i_uid, head2type(p_s_ih));
2064 #endif
2065                 /* We can't dirty inode here. It would be immediately written but
2066                  * appropriate stat item isn't inserted yet... */
2067                 if (DQUOT_ALLOC_SPACE_NODIRTY(inode, quota_bytes)) {
2068                         pathrelse(p_s_path);
2069                         return -EDQUOT;
2070                 }
2071         }
2072         init_tb_struct(th, &s_ins_balance, th->t_super, p_s_path,
2073                        IH_SIZE + ih_item_len(p_s_ih));
2074 #ifdef DISPLACE_NEW_PACKING_LOCALITIES
2075         s_ins_balance.key = key->on_disk_key;
2076 #endif
2077         /* DQUOT_* can schedule, must check to be sure calling fix_nodes is safe */
2078         if (inode && fs_changed(fs_gen, inode->i_sb)) {
2079                 goto search_again;
2080         }
2081
2082         while ((retval =
2083                 fix_nodes(M_INSERT, &s_ins_balance, p_s_ih,
2084                           p_c_body)) == REPEAT_SEARCH) {
2085               search_again:
2086                 /* file system changed while we were in the fix_nodes */
2087                 PROC_INFO_INC(th->t_super, insert_item_restarted);
2088                 retval = search_item(th->t_super, key, p_s_path);
2089                 if (retval == IO_ERROR) {
2090                         retval = -EIO;
2091                         goto error_out;
2092                 }
2093                 if (retval == ITEM_FOUND) {
2094                         reiserfs_warning(th->t_super,
2095                                          "PAP-5760: reiserfs_insert_item: "
2096                                          "key %K already exists in the tree",
2097                                          key);
2098                         retval = -EEXIST;
2099                         goto error_out;
2100                 }
2101         }
2102
2103         /* make balancing after all resources will be collected at a time */
2104         if (retval == CARRY_ON) {
2105                 do_balance(&s_ins_balance, p_s_ih, p_c_body, M_INSERT);
2106                 return 0;
2107         }
2108
2109         retval = (retval == NO_DISK_SPACE) ? -ENOSPC : -EIO;
2110       error_out:
2111         /* also releases the path */
2112         unfix_nodes(&s_ins_balance);
2113 #ifdef REISERQUOTA_DEBUG
2114         reiserfs_debug(th->t_super, REISERFS_DEBUG_CODE,
2115                        "reiserquota insert_item(): freeing %u id=%u type=%c",
2116                        quota_bytes, inode->i_uid, head2type(p_s_ih));
2117 #endif
2118         if (inode)
2119                 DQUOT_FREE_SPACE_NODIRTY(inode, quota_bytes);
2120         return retval;
2121 }