fs/reiserfs/bitmap.c

   1 /*
   2  * Copyright 2000 by Hans Reiser, licensing governed by reiserfs/README
   3  */
   4 /* Reiserfs block (de)allocator, bitmap-based. */
   5
   6 #include <linux/time.h>
   7 #include "reiserfs.h"
   8 #include <linux/errno.h>
   9 #include <linux/buffer_head.h>
  10 #include <linux/kernel.h>
  11 #include <linux/pagemap.h>
  12 #include <linux/vmalloc.h>
  13 #include <linux/quotaops.h>
  14 #include <linux/seq_file.h>
  15
  16 #define PREALLOCATION_SIZE 9
  17
  18 /* different reiserfs block allocator options */
  19
  20 #define SB_ALLOC_OPTS(s) (REISERFS_SB(s)->s_alloc_options.bits)
  21
  22 #define  _ALLOC_concentrating_formatted_nodes 0
  23 #define  _ALLOC_displacing_large_files 1
  24 #define  _ALLOC_displacing_new_packing_localities 2
  25 #define  _ALLOC_old_hashed_relocation 3
  26 #define  _ALLOC_new_hashed_relocation 4
  27 #define  _ALLOC_skip_busy 5
  28 #define  _ALLOC_displace_based_on_dirid 6
  29 #define  _ALLOC_hashed_formatted_nodes 7
  30 #define  _ALLOC_old_way 8
  31 #define  _ALLOC_hundredth_slices 9
  32 #define  _ALLOC_dirid_groups 10
  33 #define  _ALLOC_oid_groups 11
  34 #define  _ALLOC_packing_groups 12
  35
  36 #define  concentrating_formatted_nodes(s)       test_bit(_ALLOC_concentrating_formatted_nodes, &SB_ALLOC_OPTS(s))
  37 #define  displacing_large_files(s)              test_bit(_ALLOC_displacing_large_files, &SB_ALLOC_OPTS(s))
  38 #define  displacing_new_packing_localities(s)   test_bit(_ALLOC_displacing_new_packing_localities, &SB_ALLOC_OPTS(s))
  39
  40 #define SET_OPTION(optname) \
  41    do { \
  42         reiserfs_info(s, "block allocator option \"%s\" is set", #optname); \
  43         set_bit(_ALLOC_ ## optname , &SB_ALLOC_OPTS(s)); \
  44     } while(0)
  45 #define TEST_OPTION(optname, s) \
  46     test_bit(_ALLOC_ ## optname , &SB_ALLOC_OPTS(s))
  47
  48 static inline void get_bit_address(struct super_block *s,
  49                                    b_blocknr_t block,
  50                                    unsigned int *bmap_nr,
  51                                    unsigned int *offset)
  52 {
  53         /* It is in the bitmap block number equal to the block
  54          * number divided by the number of bits in a block. */
  55         *bmap_nr = block >> (s->s_blocksize_bits + 3);
  56         /* Within that bitmap block it is located at bit offset *offset. */
  57         *offset = block & ((s->s_blocksize << 3) - 1);
  58 }
  59
  60 int is_reusable(struct super_block *s, b_blocknr_t block, int bit_value)
  61 {
  62         unsigned int bmap, offset;
  63         unsigned int bmap_count = reiserfs_bmap_count(s);
  64
  65         if (block == 0 || block >= SB_BLOCK_COUNT(s)) {
  66                 reiserfs_error(s, "vs-4010",
  67                                "block number is out of range %lu (%u)",
  68                                block, SB_BLOCK_COUNT(s));
  69                 return 0;
  70         }
  71
  72         get_bit_address(s, block, &bmap, &offset);
  73
  74         /* Old format filesystem? Unlikely, but the bitmaps are all up front so
  75          * we need to account for it. */
  76         if (unlikely(test_bit(REISERFS_OLD_FORMAT,
  77                               &(REISERFS_SB(s)->s_properties)))) {
  78                 b_blocknr_t bmap1 = REISERFS_SB(s)->s_sbh->b_blocknr + 1;
  79                 if (block >= bmap1 &&
  80                     block <= bmap1 + bmap_count) {
  81                         reiserfs_error(s, "vs-4019", "bitmap block %lu(%u) "
  82                                        "can't be freed or reused",
  83                                        block, bmap_count);
  84                         return 0;
  85                 }
  86         } else {
  87                 if (offset == 0) {
  88                         reiserfs_error(s, "vs-4020", "bitmap block %lu(%u) "
  89                                        "can't be freed or reused",
  90                                        block, bmap_count);
  91                         return 0;
  92                 }
  93         }
  94
  95         if (bmap >= bmap_count) {
  96                 reiserfs_error(s, "vs-4030", "bitmap for requested block "
  97                                "is out of range: block=%lu, bitmap_nr=%u",
  98                                block, bmap);
  99                 return 0;
 100         }
 101
 102         if (bit_value == 0 && block == SB_ROOT_BLOCK(s)) {
 103                 reiserfs_error(s, "vs-4050", "this is root block (%u), "
 104                                "it must be busy", SB_ROOT_BLOCK(s));
 105                 return 0;
 106         }
 107
 108         return 1;
 109 }
 110
 111 /* searches in journal structures for a given block number (bmap, off). If block
 112    is found in reiserfs journal it suggests next free block candidate to test. */
 113 static inline int is_block_in_journal(struct super_block *s, unsigned int bmap,
 114                                       int off, int *next)
 115 {
 116         b_blocknr_t tmp;
 117
 118         if (reiserfs_in_journal(s, bmap, off, 1, &tmp)) {
 119                 if (tmp) {      /* hint supplied */
 120                         *next = tmp;
 121                         PROC_INFO_INC(s, scan_bitmap.in_journal_hint);
 122                 } else {
 123                         (*next) = off + 1;      /* inc offset to avoid looping. */
 124                         PROC_INFO_INC(s, scan_bitmap.in_journal_nohint);
 125                 }
 126                 PROC_INFO_INC(s, scan_bitmap.retry);
 127                 return 1;
 128         }
 129         return 0;
 130 }
 131
 132 /* it searches for a window of zero bits with given minimum and maximum lengths in one bitmap
 133  * block; */
 134 static int scan_bitmap_block(struct reiserfs_transaction_handle *th,
 135                              unsigned int bmap_n, int *beg, int boundary,
 136                              int min, int max, int unfm)
 137 {
 138         struct super_block *s = th->t_super;
 139         struct reiserfs_bitmap_info *bi = &SB_AP_BITMAP(s)[bmap_n];
 140         struct buffer_head *bh;
 141         int end, next;
 142         int org = *beg;
 143
 144         BUG_ON(!th->t_trans_id);
 145
 146         RFALSE(bmap_n >= reiserfs_bmap_count(s), "Bitmap %u is out of "
 147                "range (0..%u)", bmap_n, reiserfs_bmap_count(s) - 1);
 148         PROC_INFO_INC(s, scan_bitmap.bmap);
 149 /* this is unclear and lacks comments, explain how journal bitmaps
 150    work here for the reader.  Convey a sense of the design here. What
 151    is a window? */
 152 /* - I mean `a window of zero bits' as in description of this function - Zam. */
 153
 154         if (!bi) {
 155                 reiserfs_error(s, "jdm-4055", "NULL bitmap info pointer "
 156                                "for bitmap %d", bmap_n);
 157                 return 0;
 158         }
 159
 160         bh = reiserfs_read_bitmap_block(s, bmap_n);
 161         if (bh == NULL)
 162                 return 0;
 163
 164         while (1) {
 165               cont:
 166                 if (bi->free_count < min) {
 167                         brelse(bh);
 168                         return 0;       // No free blocks in this bitmap
 169                 }
 170
 171                 /* search for a first zero bit -- beginning of a window */
 172                 *beg = reiserfs_find_next_zero_le_bit
 173                     ((unsigned long *)(bh->b_data), boundary, *beg);
 174
 175                 if (*beg + min > boundary) {    /* search for a zero bit fails or the rest of bitmap block
 176                                                  * cannot contain a zero window of minimum size */
 177                         brelse(bh);
 178                         return 0;
 179                 }
 180
 181                 if (unfm && is_block_in_journal(s, bmap_n, *beg, beg))
 182                         continue;
 183                 /* first zero bit found; we check next bits */
 184                 for (end = *beg + 1;; end++) {
 185                         if (end >= *beg + max || end >= boundary
 186                             || reiserfs_test_le_bit(end, bh->b_data)) {
 187                                 next = end;
 188                                 break;
 189                         }
 190                         /* finding the other end of zero bit window requires looking into journal structures (in
 191                          * case of searching for free blocks for unformatted nodes) */
 192                         if (unfm && is_block_in_journal(s, bmap_n, end, &next))
 193                                 break;
 194                 }
 195
 196                 /* now (*beg) points to beginning of zero bits window,
 197                  * (end) points to one bit after the window end */
 198                 if (end - *beg >= min) {        /* it seems we have found window of proper size */
 199                         int i;
 200                         reiserfs_prepare_for_journal(s, bh, 1);
 201                         /* try to set all blocks used checking are they still free */
 202                         for (i = *beg; i < end; i++) {
 203                                 /* It seems that we should not check in journal again. */
 204                                 if (reiserfs_test_and_set_le_bit
 205                                     (i, bh->b_data)) {
 206                                         /* bit was set by another process
 207                                          * while we slept in prepare_for_journal() */
 208                                         PROC_INFO_INC(s, scan_bitmap.stolen);
 209                                         if (i >= *beg + min) {  /* we can continue with smaller set of allocated blocks,
 210                                                                  * if length of this set is more or equal to `min' */
 211                                                 end = i;
 212                                                 break;
 213                                         }
 214                                         /* otherwise we clear all bit were set ... */
 215                                         while (--i >= *beg)
 216                                                 reiserfs_clear_le_bit
 217                                                     (i, bh->b_data);
 218                                         reiserfs_restore_prepared_buffer(s, bh);
 219                                         *beg = org;
 220                                         /* ... and search again in current block from beginning */
 221                                         goto cont;
 222                                 }
 223                         }
 224                         bi->free_count -= (end - *beg);
 225                         journal_mark_dirty(th, s, bh);
 226                         brelse(bh);
 227
 228                         /* free block count calculation */
 229                         reiserfs_prepare_for_journal(s, SB_BUFFER_WITH_SB(s),
 230                                                      1);
 231                         PUT_SB_FREE_BLOCKS(s, SB_FREE_BLOCKS(s) - (end - *beg));
 232                         journal_mark_dirty(th, s, SB_BUFFER_WITH_SB(s));
 233
 234                         return end - (*beg);
 235                 } else {
 236                         *beg = next;
 237                 }
 238         }
 239 }
 240
 241 static int bmap_hash_id(struct super_block *s, u32 id)
 242 {
 243         char *hash_in = NULL;
 244         unsigned long hash;
 245         unsigned bm;
 246
 247         if (id <= 2) {
 248                 bm = 1;
 249         } else {
 250                 hash_in = (char *)(&id);
 251                 hash = keyed_hash(hash_in, 4);
 252                 bm = hash % reiserfs_bmap_count(s);
 253                 if (!bm)
 254                         bm = 1;
 255         }
 256         /* this can only be true when SB_BMAP_NR = 1 */
 257         if (bm >= reiserfs_bmap_count(s))
 258                 bm = 0;
 259         return bm;
 260 }
 261
 262 /*
 263  * hashes the id and then returns > 0 if the block group for the
 264  * corresponding hash is full
 265  */
 266 static inline int block_group_used(struct super_block *s, u32 id)
 267 {
 268         int bm = bmap_hash_id(s, id);
 269         struct reiserfs_bitmap_info *info = &SB_AP_BITMAP(s)[bm];
 270
 271         /* If we don't have cached information on this bitmap block, we're
 272          * going to have to load it later anyway. Loading it here allows us
 273          * to make a better decision. This favors long-term performance gain
 274          * with a better on-disk layout vs. a short term gain of skipping the
 275          * read and potentially having a bad placement. */
 276         if (info->free_count == UINT_MAX) {
 277                 struct buffer_head *bh = reiserfs_read_bitmap_block(s, bm);
 278                 brelse(bh);
 279         }
 280
 281         if (info->free_count > ((s->s_blocksize << 3) * 60 / 100)) {
 282                 return 0;
 283         }
 284         return 1;
 285 }
 286
 287 /*
 288  * the packing is returned in disk byte order
 289  */
 290 __le32 reiserfs_choose_packing(struct inode * dir)
 291 {
 292         __le32 packing;
 293         if (TEST_OPTION(packing_groups, dir->i_sb)) {
 294                 u32 parent_dir = le32_to_cpu(INODE_PKEY(dir)->k_dir_id);
 295                 /*
 296                  * some versions of reiserfsck expect packing locality 1 to be
 297                  * special
 298                  */
 299                 if (parent_dir == 1 || block_group_used(dir->i_sb, parent_dir))
 300                         packing = INODE_PKEY(dir)->k_objectid;
 301                 else
 302                         packing = INODE_PKEY(dir)->k_dir_id;
 303         } else
 304                 packing = INODE_PKEY(dir)->k_objectid;
 305         return packing;
 306 }
 307
 308 /* Tries to find contiguous zero bit window (given size) in given region of
 309  * bitmap and place new blocks there. Returns number of allocated blocks. */
 310 static int scan_bitmap(struct reiserfs_transaction_handle *th,
 311                        b_blocknr_t * start, b_blocknr_t finish,
 312                        int min, int max, int unfm, sector_t file_block)
 313 {
 314         int nr_allocated = 0;
 315         struct super_block *s = th->t_super;
 316         /* find every bm and bmap and bmap_nr in this file, and change them all to bitmap_blocknr
 317          * - Hans, it is not a block number - Zam. */
 318
 319         unsigned int bm, off;
 320         unsigned int end_bm, end_off;
 321         unsigned int off_max = s->s_blocksize << 3;
 322
 323         BUG_ON(!th->t_trans_id);
 324
 325         PROC_INFO_INC(s, scan_bitmap.call);
 326         if (SB_FREE_BLOCKS(s) <= 0)
 327                 return 0;       // No point in looking for more free blocks
 328
 329         get_bit_address(s, *start, &bm, &off);
 330         get_bit_address(s, finish, &end_bm, &end_off);
 331         if (bm > reiserfs_bmap_count(s))
 332                 return 0;
 333         if (end_bm > reiserfs_bmap_count(s))
 334                 end_bm = reiserfs_bmap_count(s);
 335
 336         /* When the bitmap is more than 10% free, anyone can allocate.
 337          * When it's less than 10% free, only files that already use the
 338          * bitmap are allowed. Once we pass 80% full, this restriction
 339          * is lifted.
 340          *
 341          * We do this so that files that grow later still have space close to
 342          * their original allocation. This improves locality, and presumably
 343          * performance as a result.
 344          *
 345          * This is only an allocation policy and does not make up for getting a
 346          * bad hint. Decent hinting must be implemented for this to work well.
 347          */
 348         if (TEST_OPTION(skip_busy, s)
 349             && SB_FREE_BLOCKS(s) > SB_BLOCK_COUNT(s) / 20) {
 350                 for (; bm < end_bm; bm++, off = 0) {
 351                         if ((off && (!unfm || (file_block != 0)))
 352                             || SB_AP_BITMAP(s)[bm].free_count >
 353                             (s->s_blocksize << 3) / 10)
 354                                 nr_allocated =
 355                                     scan_bitmap_block(th, bm, &off, off_max,
 356                                                       min, max, unfm);
 357                         if (nr_allocated)
 358                                 goto ret;
 359                 }
 360                 /* we know from above that start is a reasonable number */
 361                 get_bit_address(s, *start, &bm, &off);
 362         }
 363
 364         for (; bm < end_bm; bm++, off = 0) {
 365                 nr_allocated =
 366                     scan_bitmap_block(th, bm, &off, off_max, min, max, unfm);
 367                 if (nr_allocated)
 368                         goto ret;
 369         }
 370
 371         nr_allocated =
 372             scan_bitmap_block(th, bm, &off, end_off + 1, min, max, unfm);
 373
 374       ret:
 375         *start = bm * off_max + off;
 376         return nr_allocated;
 377
 378 }
 379
 380 static void _reiserfs_free_block(struct reiserfs_transaction_handle *th,
 381                                  struct inode *inode, b_blocknr_t block,
 382                                  int for_unformatted)
 383 {
 384         struct super_block *s = th->t_super;
 385         struct reiserfs_super_block *rs;
 386         struct buffer_head *sbh, *bmbh;
 387         struct reiserfs_bitmap_info *apbi;
 388         unsigned int nr, offset;
 389
 390         BUG_ON(!th->t_trans_id);
 391
 392         PROC_INFO_INC(s, free_block);
 393
 394         rs = SB_DISK_SUPER_BLOCK(s);
 395         sbh = SB_BUFFER_WITH_SB(s);
 396         apbi = SB_AP_BITMAP(s);
 397
 398         get_bit_address(s, block, &nr, &offset);
 399
 400         if (nr >= reiserfs_bmap_count(s)) {
 401                 reiserfs_error(s, "vs-4075", "block %lu is out of range",
 402                                block);
 403                 return;
 404         }
 405
 406         bmbh = reiserfs_read_bitmap_block(s, nr);
 407         if (!bmbh)
 408                 return;
 409
 410         reiserfs_prepare_for_journal(s, bmbh, 1);
 411
 412         /* clear bit for the given block in bit map */
 413         if (!reiserfs_test_and_clear_le_bit(offset, bmbh->b_data)) {
 414                 reiserfs_error(s, "vs-4080",
 415                                "block %lu: bit already cleared", block);
 416         }
 417         apbi[nr].free_count++;
 418         journal_mark_dirty(th, s, bmbh);
 419         brelse(bmbh);
 420
 421         reiserfs_prepare_for_journal(s, sbh, 1);
 422         /* update super block */
 423         set_sb_free_blocks(rs, sb_free_blocks(rs) + 1);
 424
 425         journal_mark_dirty(th, s, sbh);
 426         if (for_unformatted) {
 427                 int depth = reiserfs_write_unlock_nested(s);
 428                 dquot_free_block_nodirty(inode, 1);
 429                 reiserfs_write_lock_nested(s, depth);
 430         }
 431 }
 432
 433 void reiserfs_free_block(struct reiserfs_transaction_handle *th,
 434                          struct inode *inode, b_blocknr_t block,
 435                          int for_unformatted)
 436 {
 437         struct super_block *s = th->t_super;
 438         BUG_ON(!th->t_trans_id);
 439
 440         RFALSE(!s, "vs-4061: trying to free block on nonexistent device");
 441         if (!is_reusable(s, block, 1))
 442                 return;
 443
 444         if (block > sb_block_count(REISERFS_SB(s)->s_rs)) {
 445                 reiserfs_error(th->t_super, "bitmap-4072",
 446                                "Trying to free block outside file system "
 447                                "boundaries (%lu > %lu)",
 448                                block, sb_block_count(REISERFS_SB(s)->s_rs));
 449                 return;
 450         }
 451         /* mark it before we clear it, just in case */
 452         journal_mark_freed(th, s, block);
 453         _reiserfs_free_block(th, inode, block, for_unformatted);
 454 }
 455
 456 /* preallocated blocks don't need to be run through journal_mark_freed */
 457 static void reiserfs_free_prealloc_block(struct reiserfs_transaction_handle *th,
 458                                          struct inode *inode, b_blocknr_t block)
 459 {
 460         BUG_ON(!th->t_trans_id);
 461         RFALSE(!th->t_super,
 462                "vs-4060: trying to free block on nonexistent device");
 463         if (!is_reusable(th->t_super, block, 1))
 464                 return;
 465         _reiserfs_free_block(th, inode, block, 1);
 466 }
 467
 468 static void __discard_prealloc(struct reiserfs_transaction_handle *th,
 469                                struct reiserfs_inode_info *ei)
 470 {
 471         unsigned long save = ei->i_prealloc_block;
 472         int dirty = 0;
 473         struct inode *inode = &ei->vfs_inode;
 474         BUG_ON(!th->t_trans_id);
 475 #ifdef CONFIG_REISERFS_CHECK
 476         if (ei->i_prealloc_count < 0)
 477                 reiserfs_error(th->t_super, "zam-4001",
 478                                "inode has negative prealloc blocks count.");
 479 #endif
 480         while (ei->i_prealloc_count > 0) {
 481                 reiserfs_free_prealloc_block(th, inode, ei->i_prealloc_block);
 482                 ei->i_prealloc_block++;
 483                 ei->i_prealloc_count--;
 484                 dirty = 1;
 485         }
 486         if (dirty)
 487                 reiserfs_update_sd(th, inode);
 488         ei->i_prealloc_block = save;
 489         list_del_init(&(ei->i_prealloc_list));
 490 }
 491
 492 /* FIXME: It should be inline function */
 493 void reiserfs_discard_prealloc(struct reiserfs_transaction_handle *th,
 494                                struct inode *inode)
 495 {
 496         struct reiserfs_inode_info *ei = REISERFS_I(inode);
 497         BUG_ON(!th->t_trans_id);
 498         if (ei->i_prealloc_count)
 499                 __discard_prealloc(th, ei);
 500 }
 501
 502 void reiserfs_discard_all_prealloc(struct reiserfs_transaction_handle *th)
 503 {
 504         struct list_head *plist = &SB_JOURNAL(th->t_super)->j_prealloc_list;
 505
 506         BUG_ON(!th->t_trans_id);
 507
 508         while (!list_empty(plist)) {
 509                 struct reiserfs_inode_info *ei;
 510                 ei = list_entry(plist->next, struct reiserfs_inode_info,
 511                                 i_prealloc_list);
 512 #ifdef CONFIG_REISERFS_CHECK
 513                 if (!ei->i_prealloc_count) {
 514                         reiserfs_error(th->t_super, "zam-4001",
 515                                        "inode is in prealloc list but has "
 516                                        "no preallocated blocks.");
 517                 }
 518 #endif
 519                 __discard_prealloc(th, ei);
 520         }
 521 }
 522
 523 void reiserfs_init_alloc_options(struct super_block *s)
 524 {
 525         set_bit(_ALLOC_skip_busy, &SB_ALLOC_OPTS(s));
 526         set_bit(_ALLOC_dirid_groups, &SB_ALLOC_OPTS(s));
 527         set_bit(_ALLOC_packing_groups, &SB_ALLOC_OPTS(s));
 528 }
 529
 530 /* block allocator related options are parsed here */
 531 int reiserfs_parse_alloc_options(struct super_block *s, char *options)
 532 {
 533         char *this_char, *value;
 534
 535         REISERFS_SB(s)->s_alloc_options.bits = 0;       /* clear default settings */
 536
 537         while ((this_char = strsep(&options, ":")) != NULL) {
 538                 if ((value = strchr(this_char, '=')) != NULL)
 539                         *value++ = 0;
 540
 541                 if (!strcmp(this_char, "concentrating_formatted_nodes")) {
 542                         int temp;
 543                         SET_OPTION(concentrating_formatted_nodes);
 544                         temp = (value
 545                                 && *value) ? simple_strtoul(value, &value,
 546                                                             0) : 10;
 547                         if (temp <= 0 || temp > 100) {
 548                                 REISERFS_SB(s)->s_alloc_options.border = 10;
 549                         } else {
 550                                 REISERFS_SB(s)->s_alloc_options.border =
 551                                     100 / temp;
 552                         }
 553                         continue;
 554                 }
 555                 if (!strcmp(this_char, "displacing_large_files")) {
 556                         SET_OPTION(displacing_large_files);
 557                         REISERFS_SB(s)->s_alloc_options.large_file_size =
 558                             (value
 559                              && *value) ? simple_strtoul(value, &value, 0) : 16;
 560                         continue;
 561                 }
 562                 if (!strcmp(this_char, "displacing_new_packing_localities")) {
 563                         SET_OPTION(displacing_new_packing_localities);
 564                         continue;
 565                 };
 566
 567                 if (!strcmp(this_char, "old_hashed_relocation")) {
 568                         SET_OPTION(old_hashed_relocation);
 569                         continue;
 570                 }
 571
 572                 if (!strcmp(this_char, "new_hashed_relocation")) {
 573                         SET_OPTION(new_hashed_relocation);
 574                         continue;
 575                 }
 576
 577                 if (!strcmp(this_char, "dirid_groups")) {
 578                         SET_OPTION(dirid_groups);
 579                         continue;
 580                 }
 581                 if (!strcmp(this_char, "oid_groups")) {
 582                         SET_OPTION(oid_groups);
 583                         continue;
 584                 }
 585                 if (!strcmp(this_char, "packing_groups")) {
 586                         SET_OPTION(packing_groups);
 587                         continue;
 588                 }
 589                 if (!strcmp(this_char, "hashed_formatted_nodes")) {
 590                         SET_OPTION(hashed_formatted_nodes);
 591                         continue;
 592                 }
 593
 594                 if (!strcmp(this_char, "skip_busy")) {
 595                         SET_OPTION(skip_busy);
 596                         continue;
 597                 }
 598
 599                 if (!strcmp(this_char, "hundredth_slices")) {
 600                         SET_OPTION(hundredth_slices);
 601                         continue;
 602                 }
 603
 604                 if (!strcmp(this_char, "old_way")) {
 605                         SET_OPTION(old_way);
 606                         continue;
 607                 }
 608
 609                 if (!strcmp(this_char, "displace_based_on_dirid")) {
 610                         SET_OPTION(displace_based_on_dirid);
 611                         continue;
 612                 }
 613
 614                 if (!strcmp(this_char, "preallocmin")) {
 615                         REISERFS_SB(s)->s_alloc_options.preallocmin =
 616                             (value
 617                              && *value) ? simple_strtoul(value, &value, 0) : 4;
 618                         continue;
 619                 }
 620
 621                 if (!strcmp(this_char, "preallocsize")) {
 622                         REISERFS_SB(s)->s_alloc_options.preallocsize =
 623                             (value
 624                              && *value) ? simple_strtoul(value, &value,
 625                                                          0) :
 626                             PREALLOCATION_SIZE;
 627                         continue;
 628                 }
 629
 630                 reiserfs_warning(s, "zam-4001", "unknown option - %s",
 631                                  this_char);
 632                 return 1;
 633         }
 634
 635         reiserfs_info(s, "allocator options = [%08x]\n", SB_ALLOC_OPTS(s));
 636         return 0;
 637 }
 638
 639 static void print_sep(struct seq_file *seq, int *first)
 640 {
 641         if (!*first)
 642                 seq_puts(seq, ":");
 643         else
 644                 *first = 0;
 645 }
 646
 647 void show_alloc_options(struct seq_file *seq, struct super_block *s)
 648 {
 649         int first = 1;
 650
 651         if (SB_ALLOC_OPTS(s) == ((1 << _ALLOC_skip_busy) |
 652                 (1 << _ALLOC_dirid_groups) | (1 << _ALLOC_packing_groups)))
 653                 return;
 654
 655         seq_puts(seq, ",alloc=");
 656
 657         if (TEST_OPTION(concentrating_formatted_nodes, s)) {
 658                 print_sep(seq, &first);
 659                 if (REISERFS_SB(s)->s_alloc_options.border != 10) {
 660                         seq_printf(seq, "concentrating_formatted_nodes=%d",
 661                                 100 / REISERFS_SB(s)->s_alloc_options.border);
 662                 } else
 663                         seq_puts(seq, "concentrating_formatted_nodes");
 664         }
 665         if (TEST_OPTION(displacing_large_files, s)) {
 666                 print_sep(seq, &first);
 667                 if (REISERFS_SB(s)->s_alloc_options.large_file_size != 16) {
 668                         seq_printf(seq, "displacing_large_files=%lu",
 669                             REISERFS_SB(s)->s_alloc_options.large_file_size);
 670                 } else
 671                         seq_puts(seq, "displacing_large_files");
 672         }
 673         if (TEST_OPTION(displacing_new_packing_localities, s)) {
 674                 print_sep(seq, &first);
 675                 seq_puts(seq, "displacing_new_packing_localities");
 676         }
 677         if (TEST_OPTION(old_hashed_relocation, s)) {
 678                 print_sep(seq, &first);
 679                 seq_puts(seq, "old_hashed_relocation");
 680         }
 681         if (TEST_OPTION(new_hashed_relocation, s)) {
 682                 print_sep(seq, &first);
 683                 seq_puts(seq, "new_hashed_relocation");
 684         }
 685         if (TEST_OPTION(dirid_groups, s)) {
 686                 print_sep(seq, &first);
 687                 seq_puts(seq, "dirid_groups");
 688         }
 689         if (TEST_OPTION(oid_groups, s)) {
 690                 print_sep(seq, &first);
 691                 seq_puts(seq, "oid_groups");
 692         }
 693         if (TEST_OPTION(packing_groups, s)) {
 694                 print_sep(seq, &first);
 695                 seq_puts(seq, "packing_groups");
 696         }
 697         if (TEST_OPTION(hashed_formatted_nodes, s)) {
 698                 print_sep(seq, &first);
 699                 seq_puts(seq, "hashed_formatted_nodes");
 700         }
 701         if (TEST_OPTION(skip_busy, s)) {
 702                 print_sep(seq, &first);
 703                 seq_puts(seq, "skip_busy");
 704         }
 705         if (TEST_OPTION(hundredth_slices, s)) {
 706                 print_sep(seq, &first);
 707                 seq_puts(seq, "hundredth_slices");
 708         }
 709         if (TEST_OPTION(old_way, s)) {
 710                 print_sep(seq, &first);
 711                 seq_puts(seq, "old_way");
 712         }
 713         if (TEST_OPTION(displace_based_on_dirid, s)) {
 714                 print_sep(seq, &first);
 715                 seq_puts(seq, "displace_based_on_dirid");
 716         }
 717         if (REISERFS_SB(s)->s_alloc_options.preallocmin != 0) {
 718                 print_sep(seq, &first);
 719                 seq_printf(seq, "preallocmin=%d",
 720                                 REISERFS_SB(s)->s_alloc_options.preallocmin);
 721         }
 722         if (REISERFS_SB(s)->s_alloc_options.preallocsize != 17) {
 723                 print_sep(seq, &first);
 724                 seq_printf(seq, "preallocsize=%d",
 725                                 REISERFS_SB(s)->s_alloc_options.preallocsize);
 726         }
 727 }
 728
 729 static inline void new_hashed_relocation(reiserfs_blocknr_hint_t * hint)
 730 {
 731         char *hash_in;
 732         if (hint->formatted_node) {
 733                 hash_in = (char *)&hint->key.k_dir_id;
 734         } else {
 735                 if (!hint->inode) {
 736                         //hint->search_start = hint->beg;
 737                         hash_in = (char *)&hint->key.k_dir_id;
 738                 } else
 739                     if (TEST_OPTION(displace_based_on_dirid, hint->th->t_super))
 740                         hash_in = (char *)(&INODE_PKEY(hint->inode)->k_dir_id);
 741                 else
 742                         hash_in =
 743                             (char *)(&INODE_PKEY(hint->inode)->k_objectid);
 744         }
 745
 746         hint->search_start =
 747             hint->beg + keyed_hash(hash_in, 4) % (hint->end - hint->beg);
 748 }
 749
 750 /*
 751  * Relocation based on dirid, hashing them into a given bitmap block
 752  * files. Formatted nodes are unaffected, a separate policy covers them
 753  */
 754 static void dirid_groups(reiserfs_blocknr_hint_t * hint)
 755 {
 756         unsigned long hash;
 757         __u32 dirid = 0;
 758         int bm = 0;
 759         struct super_block *sb = hint->th->t_super;
 760         if (hint->inode)
 761                 dirid = le32_to_cpu(INODE_PKEY(hint->inode)->k_dir_id);
 762         else if (hint->formatted_node)
 763                 dirid = hint->key.k_dir_id;
 764
 765         if (dirid) {
 766                 bm = bmap_hash_id(sb, dirid);
 767                 hash = bm * (sb->s_blocksize << 3);
 768                 /* give a portion of the block group to metadata */
 769                 if (hint->inode)
 770                         hash += sb->s_blocksize / 2;
 771                 hint->search_start = hash;
 772         }
 773 }
 774
 775 /*
 776  * Relocation based on oid, hashing them into a given bitmap block
 777  * files. Formatted nodes are unaffected, a separate policy covers them
 778  */
 779 static void oid_groups(reiserfs_blocknr_hint_t * hint)
 780 {
 781         if (hint->inode) {
 782                 unsigned long hash;
 783                 __u32 oid;
 784                 __u32 dirid;
 785                 int bm;
 786
 787                 dirid = le32_to_cpu(INODE_PKEY(hint->inode)->k_dir_id);
 788
 789                 /* keep the root dir and it's first set of subdirs close to
 790                  * the start of the disk
 791                  */
 792                 if (dirid <= 2)
 793                         hash = (hint->inode->i_sb->s_blocksize << 3);
 794                 else {
 795                         oid = le32_to_cpu(INODE_PKEY(hint->inode)->k_objectid);
 796                         bm = bmap_hash_id(hint->inode->i_sb, oid);
 797                         hash = bm * (hint->inode->i_sb->s_blocksize << 3);
 798                 }
 799                 hint->search_start = hash;
 800         }
 801 }
 802
 803 /* returns 1 if it finds an indirect item and gets valid hint info
 804  * from it, otherwise 0
 805  */
 806 static int get_left_neighbor(reiserfs_blocknr_hint_t * hint)
 807 {
 808         struct treepath *path;
 809         struct buffer_head *bh;
 810         struct item_head *ih;
 811         int pos_in_item;
 812         __le32 *item;
 813         int ret = 0;
 814
 815         if (!hint->path)        /* reiserfs code can call this function w/o pointer to path
 816                                  * structure supplied; then we rely on supplied search_start */
 817                 return 0;
 818
 819         path = hint->path;
 820         bh = get_last_bh(path);
 821         RFALSE(!bh, "green-4002: Illegal path specified to get_left_neighbor");
 822         ih = get_ih(path);
 823         pos_in_item = path->pos_in_item;
 824         item = get_item(path);
 825
 826         hint->search_start = bh->b_blocknr;
 827
 828         if (!hint->formatted_node && is_indirect_le_ih(ih)) {
 829                 /* for indirect item: go to left and look for the first non-hole entry
 830                    in the indirect item */
 831                 if (pos_in_item == I_UNFM_NUM(ih))
 832                         pos_in_item--;
 833 //          pos_in_item = I_UNFM_NUM (ih) - 1;
 834                 while (pos_in_item >= 0) {
 835                         int t = get_block_num(item, pos_in_item);
 836                         if (t) {
 837                                 hint->search_start = t;
 838                                 ret = 1;
 839                                 break;
 840                         }
 841                         pos_in_item--;
 842                 }
 843         }
 844
 845         /* does result value fit into specified region? */
 846         return ret;
 847 }
 848
 849 /* should be, if formatted node, then try to put on first part of the device
 850    specified as number of percent with mount option device, else try to put
 851    on last of device.  This is not to say it is good code to do so,
 852    but the effect should be measured.  */
 853 static inline void set_border_in_hint(struct super_block *s,
 854                                       reiserfs_blocknr_hint_t * hint)
 855 {
 856         b_blocknr_t border =
 857             SB_BLOCK_COUNT(s) / REISERFS_SB(s)->s_alloc_options.border;
 858
 859         if (hint->formatted_node)
 860                 hint->end = border - 1;
 861         else
 862                 hint->beg = border;
 863 }
 864
 865 static inline void displace_large_file(reiserfs_blocknr_hint_t * hint)
 866 {
 867         if (TEST_OPTION(displace_based_on_dirid, hint->th->t_super))
 868                 hint->search_start =
 869                     hint->beg +
 870                     keyed_hash((char *)(&INODE_PKEY(hint->inode)->k_dir_id),
 871                                4) % (hint->end - hint->beg);
 872         else
 873                 hint->search_start =
 874                     hint->beg +
 875                     keyed_hash((char *)(&INODE_PKEY(hint->inode)->k_objectid),
 876                                4) % (hint->end - hint->beg);
 877 }
 878
 879 static inline void hash_formatted_node(reiserfs_blocknr_hint_t * hint)
 880 {
 881         char *hash_in;
 882
 883         if (!hint->inode)
 884                 hash_in = (char *)&hint->key.k_dir_id;
 885         else if (TEST_OPTION(displace_based_on_dirid, hint->th->t_super))
 886                 hash_in = (char *)(&INODE_PKEY(hint->inode)->k_dir_id);
 887         else
 888                 hash_in = (char *)(&INODE_PKEY(hint->inode)->k_objectid);
 889
 890         hint->search_start =
 891             hint->beg + keyed_hash(hash_in, 4) % (hint->end - hint->beg);
 892 }
 893
 894 static inline int
 895 this_blocknr_allocation_would_make_it_a_large_file(reiserfs_blocknr_hint_t *
 896                                                    hint)
 897 {
 898         return hint->block ==
 899             REISERFS_SB(hint->th->t_super)->s_alloc_options.large_file_size;
 900 }
 901
 902 #ifdef DISPLACE_NEW_PACKING_LOCALITIES
 903 static inline void displace_new_packing_locality(reiserfs_blocknr_hint_t * hint)
 904 {
 905         struct in_core_key *key = &hint->key;
 906
 907         hint->th->displace_new_blocks = 0;
 908         hint->search_start =
 909             hint->beg + keyed_hash((char *)(&key->k_objectid),
 910                                    4) % (hint->end - hint->beg);
 911 }
 912 #endif
 913
 914 static inline int old_hashed_relocation(reiserfs_blocknr_hint_t * hint)
 915 {
 916         b_blocknr_t border;
 917         u32 hash_in;
 918
 919         if (hint->formatted_node || hint->inode == NULL) {
 920                 return 0;
 921         }
 922
 923         hash_in = le32_to_cpu((INODE_PKEY(hint->inode))->k_dir_id);
 924         border =
 925             hint->beg + (u32) keyed_hash(((char *)(&hash_in)),
 926                                          4) % (hint->end - hint->beg - 1);
 927         if (border > hint->search_start)
 928                 hint->search_start = border;
 929
 930         return 1;
 931 }
 932
 933 static inline int old_way(reiserfs_blocknr_hint_t * hint)
 934 {
 935         b_blocknr_t border;
 936
 937         if (hint->formatted_node || hint->inode == NULL) {
 938                 return 0;
 939         }
 940
 941         border =
 942             hint->beg +
 943             le32_to_cpu(INODE_PKEY(hint->inode)->k_dir_id) % (hint->end -
 944                                                               hint->beg);
 945         if (border > hint->search_start)
 946                 hint->search_start = border;
 947
 948         return 1;
 949 }
 950
 951 static inline void hundredth_slices(reiserfs_blocknr_hint_t * hint)
 952 {
 953         struct in_core_key *key = &hint->key;
 954         b_blocknr_t slice_start;
 955
 956         slice_start =
 957             (keyed_hash((char *)(&key->k_dir_id), 4) % 100) * (hint->end / 100);
 958         if (slice_start > hint->search_start
 959             || slice_start + (hint->end / 100) <= hint->search_start) {
 960                 hint->search_start = slice_start;
 961         }
 962 }
 963
 964 static void determine_search_start(reiserfs_blocknr_hint_t * hint,
 965                                    int amount_needed)
 966 {
 967         struct super_block *s = hint->th->t_super;
 968         int unfm_hint;
 969
 970         hint->beg = 0;
 971         hint->end = SB_BLOCK_COUNT(s) - 1;
 972
 973         /* This is former border algorithm. Now with tunable border offset */
 974         if (concentrating_formatted_nodes(s))
 975                 set_border_in_hint(s, hint);
 976
 977 #ifdef DISPLACE_NEW_PACKING_LOCALITIES
 978         /* whenever we create a new directory, we displace it.  At first we will
 979            hash for location, later we might look for a moderately empty place for
 980            it */
 981         if (displacing_new_packing_localities(s)
 982             && hint->th->displace_new_blocks) {
 983                 displace_new_packing_locality(hint);
 984
 985                 /* we do not continue determine_search_start,
 986                  * if new packing locality is being displaced */
 987                 return;
 988         }
 989 #endif
 990
 991         /* all persons should feel encouraged to add more special cases here and
 992          * test them */
 993
 994         if (displacing_large_files(s) && !hint->formatted_node
 995             && this_blocknr_allocation_would_make_it_a_large_file(hint)) {
 996                 displace_large_file(hint);
 997                 return;
 998         }
 999
1000         /* if none of our special cases is relevant, use the left neighbor in the
1001            tree order of the new node we are allocating for */
1002         if (hint->formatted_node && TEST_OPTION(hashed_formatted_nodes, s)) {
1003                 hash_formatted_node(hint);
1004                 return;
1005         }
1006
1007         unfm_hint = get_left_neighbor(hint);
1008
1009         /* Mimic old block allocator behaviour, that is if VFS allowed for preallocation,
1010            new blocks are displaced based on directory ID. Also, if suggested search_start
1011            is less than last preallocated block, we start searching from it, assuming that
1012            HDD dataflow is faster in forward direction */
1013         if (TEST_OPTION(old_way, s)) {
1014                 if (!hint->formatted_node) {
1015                         if (!reiserfs_hashed_relocation(s))
1016                                 old_way(hint);
1017                         else if (!reiserfs_no_unhashed_relocation(s))
1018                                 old_hashed_relocation(hint);
1019
1020                         if (hint->inode
1021                             && hint->search_start <
1022                             REISERFS_I(hint->inode)->i_prealloc_block)
1023                                 hint->search_start =
1024                                     REISERFS_I(hint->inode)->i_prealloc_block;
1025                 }
1026                 return;
1027         }
1028
1029         /* This is an approach proposed by Hans */
1030         if (TEST_OPTION(hundredth_slices, s)
1031             && !(displacing_large_files(s) && !hint->formatted_node)) {
1032                 hundredth_slices(hint);
1033                 return;
1034         }
1035
1036         /* old_hashed_relocation only works on unformatted */
1037         if (!unfm_hint && !hint->formatted_node &&
1038             TEST_OPTION(old_hashed_relocation, s)) {
1039                 old_hashed_relocation(hint);
1040         }
1041         /* new_hashed_relocation works with both formatted/unformatted nodes */
1042         if ((!unfm_hint || hint->formatted_node) &&
1043             TEST_OPTION(new_hashed_relocation, s)) {
1044                 new_hashed_relocation(hint);
1045         }
1046         /* dirid grouping works only on unformatted nodes */
1047         if (!unfm_hint && !hint->formatted_node && TEST_OPTION(dirid_groups, s)) {
1048                 dirid_groups(hint);
1049         }
1050 #ifdef DISPLACE_NEW_PACKING_LOCALITIES
1051         if (hint->formatted_node && TEST_OPTION(dirid_groups, s)) {
1052                 dirid_groups(hint);
1053         }
1054 #endif
1055
1056         /* oid grouping works only on unformatted nodes */
1057         if (!unfm_hint && !hint->formatted_node && TEST_OPTION(oid_groups, s)) {
1058                 oid_groups(hint);
1059         }
1060         return;
1061 }
1062
1063 static int determine_prealloc_size(reiserfs_blocknr_hint_t * hint)
1064 {
1065         /* make minimum size a mount option and benchmark both ways */
1066         /* we preallocate blocks only for regular files, specific size */
1067         /* benchmark preallocating always and see what happens */
1068
1069         hint->prealloc_size = 0;
1070
1071         if (!hint->formatted_node && hint->preallocate) {
1072                 if (S_ISREG(hint->inode->i_mode)
1073                     && hint->inode->i_size >=
1074                     REISERFS_SB(hint->th->t_super)->s_alloc_options.
1075                     preallocmin * hint->inode->i_sb->s_blocksize)
1076                         hint->prealloc_size =
1077                             REISERFS_SB(hint->th->t_super)->s_alloc_options.
1078                             preallocsize - 1;
1079         }
1080         return CARRY_ON;
1081 }
1082
1083 /* XXX I know it could be merged with upper-level function;
1084    but may be result function would be too complex. */
1085 static inline int allocate_without_wrapping_disk(reiserfs_blocknr_hint_t * hint,
1086                                                  b_blocknr_t * new_blocknrs,
1087                                                  b_blocknr_t start,
1088                                                  b_blocknr_t finish, int min,
1089                                                  int amount_needed,
1090                                                  int prealloc_size)
1091 {
1092         int rest = amount_needed;
1093         int nr_allocated;
1094
1095         while (rest > 0 && start <= finish) {
1096                 nr_allocated = scan_bitmap(hint->th, &start, finish, min,
1097                                            rest + prealloc_size,
1098                                            !hint->formatted_node, hint->block);
1099
1100                 if (nr_allocated == 0)  /* no new blocks allocated, return */
1101                         break;
1102
1103                 /* fill free_blocknrs array first */
1104                 while (rest > 0 && nr_allocated > 0) {
1105                         *new_blocknrs++ = start++;
1106                         rest--;
1107                         nr_allocated--;
1108                 }
1109
1110                 /* do we have something to fill prealloc. array also ? */
1111                 if (nr_allocated > 0) {
1112                         /* it means prealloc_size was greater that 0 and we do preallocation */
1113                         list_add(&REISERFS_I(hint->inode)->i_prealloc_list,
1114                                  &SB_JOURNAL(hint->th->t_super)->
1115                                  j_prealloc_list);
1116                         REISERFS_I(hint->inode)->i_prealloc_block = start;
1117                         REISERFS_I(hint->inode)->i_prealloc_count =
1118                             nr_allocated;
1119                         break;
1120                 }
1121         }
1122
1123         return (amount_needed - rest);
1124 }
1125
1126 static inline int blocknrs_and_prealloc_arrays_from_search_start
1127     (reiserfs_blocknr_hint_t * hint, b_blocknr_t * new_blocknrs,
1128      int amount_needed) {
1129         struct super_block *s = hint->th->t_super;
1130         b_blocknr_t start = hint->search_start;
1131         b_blocknr_t finish = SB_BLOCK_COUNT(s) - 1;
1132         int passno = 0;
1133         int nr_allocated = 0;
1134         int depth;
1135
1136         determine_prealloc_size(hint);
1137         if (!hint->formatted_node) {
1138                 int quota_ret;
1139 #ifdef REISERQUOTA_DEBUG
1140                 reiserfs_debug(s, REISERFS_DEBUG_CODE,
1141                                "reiserquota: allocating %d blocks id=%u",
1142                                amount_needed, hint->inode->i_uid);
1143 #endif
1144                 depth = reiserfs_write_unlock_nested(s);
1145                 quota_ret =
1146                     dquot_alloc_block_nodirty(hint->inode, amount_needed);
1147                 if (quota_ret) {        /* Quota exceeded? */
1148                         reiserfs_write_lock_nested(s, depth);
1149                         return QUOTA_EXCEEDED;
1150                 }
1151                 if (hint->preallocate && hint->prealloc_size) {
1152 #ifdef REISERQUOTA_DEBUG
1153                         reiserfs_debug(s, REISERFS_DEBUG_CODE,
1154                                        "reiserquota: allocating (prealloc) %d blocks id=%u",
1155                                        hint->prealloc_size, hint->inode->i_uid);
1156 #endif
1157                         quota_ret = dquot_prealloc_block_nodirty(hint->inode,
1158                                                          hint->prealloc_size);
1159                         if (quota_ret)
1160                                 hint->preallocate = hint->prealloc_size = 0;
1161                 }
1162                 /* for unformatted nodes, force large allocations */
1163                 reiserfs_write_lock_nested(s, depth);
1164         }
1165
1166         do {
1167                 switch (passno++) {
1168                 case 0: /* Search from hint->search_start to end of disk */
1169                         start = hint->search_start;
1170                         finish = SB_BLOCK_COUNT(s) - 1;
1171                         break;
1172                 case 1: /* Search from hint->beg to hint->search_start */
1173                         start = hint->beg;
1174                         finish = hint->search_start;
1175                         break;
1176                 case 2: /* Last chance: Search from 0 to hint->beg */
1177                         start = 0;
1178                         finish = hint->beg;
1179                         break;
1180                 default:        /* We've tried searching everywhere, not enough space */
1181                         /* Free the blocks */
1182                         if (!hint->formatted_node) {
1183 #ifdef REISERQUOTA_DEBUG
1184                                 reiserfs_debug(s, REISERFS_DEBUG_CODE,
1185                                                "reiserquota: freeing (nospace) %d blocks id=%u",
1186                                                amount_needed +
1187                                                hint->prealloc_size -
1188                                                nr_allocated,
1189                                                hint->inode->i_uid);
1190 #endif
1191                                 /* Free not allocated blocks */
1192                                 depth = reiserfs_write_unlock_nested(s);
1193                                 dquot_free_block_nodirty(hint->inode,
1194                                         amount_needed + hint->prealloc_size -
1195                                         nr_allocated);
1196                                 reiserfs_write_lock_nested(s, depth);
1197                         }
1198                         while (nr_allocated--)
1199                                 reiserfs_free_block(hint->th, hint->inode,
1200                                                     new_blocknrs[nr_allocated],
1201                                                     !hint->formatted_node);
1202
1203                         return NO_DISK_SPACE;
1204                 }
1205         } while ((nr_allocated += allocate_without_wrapping_disk(hint,
1206                                                                  new_blocknrs +
1207                                                                  nr_allocated,
1208                                                                  start, finish,
1209                                                                  1,
1210                                                                  amount_needed -
1211                                                                  nr_allocated,
1212                                                                  hint->
1213                                                                  prealloc_size))
1214                  < amount_needed);
1215         if (!hint->formatted_node &&
1216             amount_needed + hint->prealloc_size >
1217             nr_allocated + REISERFS_I(hint->inode)->i_prealloc_count) {
1218                 /* Some of preallocation blocks were not allocated */
1219 #ifdef REISERQUOTA_DEBUG
1220                 reiserfs_debug(s, REISERFS_DEBUG_CODE,
1221                                "reiserquota: freeing (failed prealloc) %d blocks id=%u",
1222                                amount_needed + hint->prealloc_size -
1223                                nr_allocated -
1224                                REISERFS_I(hint->inode)->i_prealloc_count,
1225                                hint->inode->i_uid);
1226 #endif
1227
1228                 depth = reiserfs_write_unlock_nested(s);
1229                 dquot_free_block_nodirty(hint->inode, amount_needed +
1230                                          hint->prealloc_size - nr_allocated -
1231                                          REISERFS_I(hint->inode)->
1232                                          i_prealloc_count);
1233                 reiserfs_write_lock_nested(s, depth);
1234         }
1235
1236         return CARRY_ON;
1237 }
1238
1239 /* grab new blocknrs from preallocated list */
1240 /* return amount still needed after using them */
1241 static int use_preallocated_list_if_available(reiserfs_blocknr_hint_t * hint,
1242                                               b_blocknr_t * new_blocknrs,
1243                                               int amount_needed)
1244 {
1245         struct inode *inode = hint->inode;
1246
1247         if (REISERFS_I(inode)->i_prealloc_count > 0) {
1248                 while (amount_needed) {
1249
1250                         *new_blocknrs++ = REISERFS_I(inode)->i_prealloc_block++;
1251                         REISERFS_I(inode)->i_prealloc_count--;
1252
1253                         amount_needed--;
1254
1255                         if (REISERFS_I(inode)->i_prealloc_count <= 0) {
1256                                 list_del(&REISERFS_I(inode)->i_prealloc_list);
1257                                 break;
1258                         }
1259                 }
1260         }
1261         /* return amount still needed after using preallocated blocks */
1262         return amount_needed;
1263 }
1264
1265 int reiserfs_allocate_blocknrs(reiserfs_blocknr_hint_t * hint, b_blocknr_t * new_blocknrs, int amount_needed, int reserved_by_us        /* Amount of blocks we have
1266                                                                                                                                            already reserved */ )
1267 {
1268         int initial_amount_needed = amount_needed;
1269         int ret;
1270         struct super_block *s = hint->th->t_super;
1271
1272         /* Check if there is enough space, taking into account reserved space */
1273         if (SB_FREE_BLOCKS(s) - REISERFS_SB(s)->reserved_blocks <
1274             amount_needed - reserved_by_us)
1275                 return NO_DISK_SPACE;
1276         /* should this be if !hint->inode &&  hint->preallocate? */
1277         /* do you mean hint->formatted_node can be removed ? - Zam */
1278         /* hint->formatted_node cannot be removed because we try to access
1279            inode information here, and there is often no inode assotiated with
1280            metadata allocations - green */
1281
1282         if (!hint->formatted_node && hint->preallocate) {
1283                 amount_needed = use_preallocated_list_if_available
1284                     (hint, new_blocknrs, amount_needed);
1285                 if (amount_needed == 0) /* all blocknrs we need we got from
1286                                            prealloc. list */
1287                         return CARRY_ON;
1288                 new_blocknrs += (initial_amount_needed - amount_needed);
1289         }
1290
1291         /* find search start and save it in hint structure */
1292         determine_search_start(hint, amount_needed);
1293         if (hint->search_start >= SB_BLOCK_COUNT(s))
1294                 hint->search_start = SB_BLOCK_COUNT(s) - 1;
1295
1296         /* allocation itself; fill new_blocknrs and preallocation arrays */
1297         ret = blocknrs_and_prealloc_arrays_from_search_start
1298             (hint, new_blocknrs, amount_needed);
1299
1300         /* we used prealloc. list to fill (partially) new_blocknrs array. If final allocation fails we
1301          * need to return blocks back to prealloc. list or just free them. -- Zam (I chose second
1302          * variant) */
1303
1304         if (ret != CARRY_ON) {
1305                 while (amount_needed++ < initial_amount_needed) {
1306                         reiserfs_free_block(hint->th, hint->inode,
1307                                             *(--new_blocknrs), 1);
1308                 }
1309         }
1310         return ret;
1311 }
1312
1313 void reiserfs_cache_bitmap_metadata(struct super_block *sb,
1314                                     struct buffer_head *bh,
1315                                     struct reiserfs_bitmap_info *info)
1316 {
1317         unsigned long *cur = (unsigned long *)(bh->b_data + bh->b_size);
1318
1319         /* The first bit must ALWAYS be 1 */
1320         if (!reiserfs_test_le_bit(0, (unsigned long *)bh->b_data))
1321                 reiserfs_error(sb, "reiserfs-2025", "bitmap block %lu is "
1322                                "corrupted: first bit must be 1", bh->b_blocknr);
1323
1324         info->free_count = 0;
1325
1326         while (--cur >= (unsigned long *)bh->b_data) {
1327                 /* 0 and ~0 are special, we can optimize for them */
1328                 if (*cur == 0)
1329                         info->free_count += BITS_PER_LONG;
1330                 else if (*cur != ~0L)   /* A mix, investigate */
1331                         info->free_count += BITS_PER_LONG - hweight_long(*cur);
1332         }
1333 }
1334
1335 struct buffer_head *reiserfs_read_bitmap_block(struct super_block *sb,
1336                                                unsigned int bitmap)
1337 {
1338         b_blocknr_t block = (sb->s_blocksize << 3) * bitmap;
1339         struct reiserfs_bitmap_info *info = SB_AP_BITMAP(sb) + bitmap;
1340         struct buffer_head *bh;
1341
1342         /* Way old format filesystems had the bitmaps packed up front.
1343          * I doubt there are any of these left, but just in case... */
1344         if (unlikely(test_bit(REISERFS_OLD_FORMAT,
1345                               &(REISERFS_SB(sb)->s_properties))))
1346                 block = REISERFS_SB(sb)->s_sbh->b_blocknr + 1 + bitmap;
1347         else if (bitmap == 0)
1348                 block = (REISERFS_DISK_OFFSET_IN_BYTES >> sb->s_blocksize_bits) + 1;
1349
1350         bh = sb_bread(sb, block);
1351         if (bh == NULL)
1352                 reiserfs_warning(sb, "sh-2029: %s: bitmap block (#%u) "
1353                                  "reading failed", __func__, block);
1354         else {
1355                 if (buffer_locked(bh)) {
1356                         int depth;
1357                         PROC_INFO_INC(sb, scan_bitmap.wait);
1358                         depth = reiserfs_write_unlock_nested(sb);
1359                         __wait_on_buffer(bh);
1360                         reiserfs_write_lock_nested(sb, depth);
1361                 }
1362                 BUG_ON(!buffer_uptodate(bh));
1363                 BUG_ON(atomic_read(&bh->b_count) == 0);
1364
1365                 if (info->free_count == UINT_MAX)
1366                         reiserfs_cache_bitmap_metadata(sb, bh, info);
1367         }
1368
1369         return bh;
1370 }
1371
1372 int reiserfs_init_bitmap_cache(struct super_block *sb)
1373 {
1374         struct reiserfs_bitmap_info *bitmap;
1375         unsigned int bmap_nr = reiserfs_bmap_count(sb);
1376
1377         bitmap = vmalloc(sizeof(*bitmap) * bmap_nr);
1378         if (bitmap == NULL)
1379                 return -ENOMEM;
1380
1381         memset(bitmap, 0xff, sizeof(*bitmap) * bmap_nr);
1382
1383         SB_AP_BITMAP(sb) = bitmap;
1384
1385         return 0;
1386 }
1387
1388 void reiserfs_free_bitmap_cache(struct super_block *sb)
1389 {
1390         if (SB_AP_BITMAP(sb)) {
1391                 vfree(SB_AP_BITMAP(sb));
1392                 SB_AP_BITMAP(sb) = NULL;
1393         }
1394 }