fs/btrfs/extent_io.c

   1 #include <linux/bitops.h>
   2 #include <linux/slab.h>
   3 #include <linux/bio.h>
   4 #include <linux/mm.h>
   5 #include <linux/gfp.h>
   6 #include <linux/pagemap.h>
   7 #include <linux/page-flags.h>
   8 #include <linux/module.h>
   9 #include <linux/spinlock.h>
  10 #include <linux/blkdev.h>
  11 #include <linux/swap.h>
  12 #include <linux/writeback.h>
  13 #include <linux/pagevec.h>
  14 #include "extent_io.h"
  15 #include "extent_map.h"
  16 #include "compat.h"
  17 #include "ctree.h"
  18 #include "btrfs_inode.h"
  19
  20 static struct kmem_cache *extent_state_cache;
  21 static struct kmem_cache *extent_buffer_cache;
  22
  23 static LIST_HEAD(buffers);
  24 static LIST_HEAD(states);
  25
  26 #define LEAK_DEBUG 0
  27 #if LEAK_DEBUG
  28 static DEFINE_SPINLOCK(leak_lock);
  29 #endif
  30
  31 #define BUFFER_LRU_MAX 64
  32
  33 struct tree_entry {
  34         u64 start;
  35         u64 end;
  36         struct rb_node rb_node;
  37 };
  38
  39 struct extent_page_data {
  40         struct bio *bio;
  41         struct extent_io_tree *tree;
  42         get_extent_t *get_extent;
  43
  44         /* tells writepage not to lock the state bits for this range
  45          * it still does the unlocking
  46          */
  47         unsigned int extent_locked:1;
  48
  49         /* tells the submit_bio code to use a WRITE_SYNC */
  50         unsigned int sync_io:1;
  51 };
  52
  53 int __init extent_io_init(void)
  54 {
  55         extent_state_cache = kmem_cache_create("extent_state",
  56                         sizeof(struct extent_state), 0,
  57                         SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
  58         if (!extent_state_cache)
  59                 return -ENOMEM;
  60
  61         extent_buffer_cache = kmem_cache_create("extent_buffers",
  62                         sizeof(struct extent_buffer), 0,
  63                         SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
  64         if (!extent_buffer_cache)
  65                 goto free_state_cache;
  66         return 0;
  67
  68 free_state_cache:
  69         kmem_cache_destroy(extent_state_cache);
  70         return -ENOMEM;
  71 }
  72
  73 void extent_io_exit(void)
  74 {
  75         struct extent_state *state;
  76         struct extent_buffer *eb;
  77
  78         while (!list_empty(&states)) {
  79                 state = list_entry(states.next, struct extent_state, leak_list);
  80                 printk(KERN_ERR "btrfs state leak: start %llu end %llu "
  81                        "state %lu in tree %p refs %d\n",
  82                        (unsigned long long)state->start,
  83                        (unsigned long long)state->end,
  84                        state->state, state->tree, atomic_read(&state->refs));
  85                 list_del(&state->leak_list);
  86                 kmem_cache_free(extent_state_cache, state);
  87
  88         }
  89
  90         while (!list_empty(&buffers)) {
  91                 eb = list_entry(buffers.next, struct extent_buffer, leak_list);
  92                 printk(KERN_ERR "btrfs buffer leak start %llu len %lu "
  93                        "refs %d\n", (unsigned long long)eb->start,
  94                        eb->len, atomic_read(&eb->refs));
  95                 list_del(&eb->leak_list);
  96                 kmem_cache_free(extent_buffer_cache, eb);
  97         }
  98         if (extent_state_cache)
  99                 kmem_cache_destroy(extent_state_cache);
 100         if (extent_buffer_cache)
 101                 kmem_cache_destroy(extent_buffer_cache);
 102 }
 103
 104 void extent_io_tree_init(struct extent_io_tree *tree,
 105                           struct address_space *mapping, gfp_t mask)
 106 {
 107         tree->state.rb_node = NULL;
 108         tree->buffer.rb_node = NULL;
 109         tree->ops = NULL;
 110         tree->dirty_bytes = 0;
 111         spin_lock_init(&tree->lock);
 112         spin_lock_init(&tree->buffer_lock);
 113         tree->mapping = mapping;
 114 }
 115
 116 static struct extent_state *alloc_extent_state(gfp_t mask)
 117 {
 118         struct extent_state *state;
 119 #if LEAK_DEBUG
 120         unsigned long flags;
 121 #endif
 122
 123         state = kmem_cache_alloc(extent_state_cache, mask);
 124         if (!state)
 125                 return state;
 126         state->state = 0;
 127         state->private = 0;
 128         state->tree = NULL;
 129 #if LEAK_DEBUG
 130         spin_lock_irqsave(&leak_lock, flags);
 131         list_add(&state->leak_list, &states);
 132         spin_unlock_irqrestore(&leak_lock, flags);
 133 #endif
 134         atomic_set(&state->refs, 1);
 135         init_waitqueue_head(&state->wq);
 136         return state;
 137 }
 138
 139 static void free_extent_state(struct extent_state *state)
 140 {
 141         if (!state)
 142                 return;
 143         if (atomic_dec_and_test(&state->refs)) {
 144 #if LEAK_DEBUG
 145                 unsigned long flags;
 146 #endif
 147                 WARN_ON(state->tree);
 148 #if LEAK_DEBUG
 149                 spin_lock_irqsave(&leak_lock, flags);
 150                 list_del(&state->leak_list);
 151                 spin_unlock_irqrestore(&leak_lock, flags);
 152 #endif
 153                 kmem_cache_free(extent_state_cache, state);
 154         }
 155 }
 156
 157 static struct rb_node *tree_insert(struct rb_root *root, u64 offset,
 158                                    struct rb_node *node)
 159 {
 160         struct rb_node **p = &root->rb_node;
 161         struct rb_node *parent = NULL;
 162         struct tree_entry *entry;
 163
 164         while (*p) {
 165                 parent = *p;
 166                 entry = rb_entry(parent, struct tree_entry, rb_node);
 167
 168                 if (offset < entry->start)
 169                         p = &(*p)->rb_left;
 170                 else if (offset > entry->end)
 171                         p = &(*p)->rb_right;
 172                 else
 173                         return parent;
 174         }
 175
 176         entry = rb_entry(node, struct tree_entry, rb_node);
 177         rb_link_node(node, parent, p);
 178         rb_insert_color(node, root);
 179         return NULL;
 180 }
 181
 182 static struct rb_node *__etree_search(struct extent_io_tree *tree, u64 offset,
 183                                      struct rb_node **prev_ret,
 184                                      struct rb_node **next_ret)
 185 {
 186         struct rb_root *root = &tree->state;
 187         struct rb_node *n = root->rb_node;
 188         struct rb_node *prev = NULL;
 189         struct rb_node *orig_prev = NULL;
 190         struct tree_entry *entry;
 191         struct tree_entry *prev_entry = NULL;
 192
 193         while (n) {
 194                 entry = rb_entry(n, struct tree_entry, rb_node);
 195                 prev = n;
 196                 prev_entry = entry;
 197
 198                 if (offset < entry->start)
 199                         n = n->rb_left;
 200                 else if (offset > entry->end)
 201                         n = n->rb_right;
 202                 else
 203                         return n;
 204         }
 205
 206         if (prev_ret) {
 207                 orig_prev = prev;
 208                 while (prev && offset > prev_entry->end) {
 209                         prev = rb_next(prev);
 210                         prev_entry = rb_entry(prev, struct tree_entry, rb_node);
 211                 }
 212                 *prev_ret = prev;
 213                 prev = orig_prev;
 214         }
 215
 216         if (next_ret) {
 217                 prev_entry = rb_entry(prev, struct tree_entry, rb_node);
 218                 while (prev && offset < prev_entry->start) {
 219                         prev = rb_prev(prev);
 220                         prev_entry = rb_entry(prev, struct tree_entry, rb_node);
 221                 }
 222                 *next_ret = prev;
 223         }
 224         return NULL;
 225 }
 226
 227 static inline struct rb_node *tree_search(struct extent_io_tree *tree,
 228                                           u64 offset)
 229 {
 230         struct rb_node *prev = NULL;
 231         struct rb_node *ret;
 232
 233         ret = __etree_search(tree, offset, &prev, NULL);
 234         if (!ret)
 235                 return prev;
 236         return ret;
 237 }
 238
 239 static struct extent_buffer *buffer_tree_insert(struct extent_io_tree *tree,
 240                                           u64 offset, struct rb_node *node)
 241 {
 242         struct rb_root *root = &tree->buffer;
 243         struct rb_node **p = &root->rb_node;
 244         struct rb_node *parent = NULL;
 245         struct extent_buffer *eb;
 246
 247         while (*p) {
 248                 parent = *p;
 249                 eb = rb_entry(parent, struct extent_buffer, rb_node);
 250
 251                 if (offset < eb->start)
 252                         p = &(*p)->rb_left;
 253                 else if (offset > eb->start)
 254                         p = &(*p)->rb_right;
 255                 else
 256                         return eb;
 257         }
 258
 259         rb_link_node(node, parent, p);
 260         rb_insert_color(node, root);
 261         return NULL;
 262 }
 263
 264 static struct extent_buffer *buffer_search(struct extent_io_tree *tree,
 265                                            u64 offset)
 266 {
 267         struct rb_root *root = &tree->buffer;
 268         struct rb_node *n = root->rb_node;
 269         struct extent_buffer *eb;
 270
 271         while (n) {
 272                 eb = rb_entry(n, struct extent_buffer, rb_node);
 273                 if (offset < eb->start)
 274                         n = n->rb_left;
 275                 else if (offset > eb->start)
 276                         n = n->rb_right;
 277                 else
 278                         return eb;
 279         }
 280         return NULL;
 281 }
 282
 283 /*
 284  * utility function to look for merge candidates inside a given range.
 285  * Any extents with matching state are merged together into a single
 286  * extent in the tree.  Extents with EXTENT_IO in their state field
 287  * are not merged because the end_io handlers need to be able to do
 288  * operations on them without sleeping (or doing allocations/splits).
 289  *
 290  * This should be called with the tree lock held.
 291  */
 292 static int merge_state(struct extent_io_tree *tree,
 293                        struct extent_state *state)
 294 {
 295         struct extent_state *other;
 296         struct rb_node *other_node;
 297
 298         if (state->state & (EXTENT_IOBITS | EXTENT_BOUNDARY))
 299                 return 0;
 300
 301         other_node = rb_prev(&state->rb_node);
 302         if (other_node) {
 303                 other = rb_entry(other_node, struct extent_state, rb_node);
 304                 if (other->end == state->start - 1 &&
 305                     other->state == state->state) {
 306                         state->start = other->start;
 307                         other->tree = NULL;
 308                         rb_erase(&other->rb_node, &tree->state);
 309                         free_extent_state(other);
 310                 }
 311         }
 312         other_node = rb_next(&state->rb_node);
 313         if (other_node) {
 314                 other = rb_entry(other_node, struct extent_state, rb_node);
 315                 if (other->start == state->end + 1 &&
 316                     other->state == state->state) {
 317                         other->start = state->start;
 318                         state->tree = NULL;
 319                         rb_erase(&state->rb_node, &tree->state);
 320                         free_extent_state(state);
 321                 }
 322         }
 323         return 0;
 324 }
 325
 326 static void set_state_cb(struct extent_io_tree *tree,
 327                          struct extent_state *state,
 328                          unsigned long bits)
 329 {
 330         if (tree->ops && tree->ops->set_bit_hook) {
 331                 tree->ops->set_bit_hook(tree->mapping->host, state->start,
 332                                         state->end, state->state, bits);
 333         }
 334 }
 335
 336 static void clear_state_cb(struct extent_io_tree *tree,
 337                            struct extent_state *state,
 338                            unsigned long bits)
 339 {
 340         if (tree->ops && tree->ops->clear_bit_hook) {
 341                 tree->ops->clear_bit_hook(tree->mapping->host, state->start,
 342                                           state->end, state->state, bits);
 343         }
 344 }
 345
 346 /*
 347  * insert an extent_state struct into the tree.  'bits' are set on the
 348  * struct before it is inserted.
 349  *
 350  * This may return -EEXIST if the extent is already there, in which case the
 351  * state struct is freed.
 352  *
 353  * The tree lock is not taken internally.  This is a utility function and
 354  * probably isn't what you want to call (see set/clear_extent_bit).
 355  */
 356 static int insert_state(struct extent_io_tree *tree,
 357                         struct extent_state *state, u64 start, u64 end,
 358                         int bits)
 359 {
 360         struct rb_node *node;
 361
 362         if (end < start) {
 363                 printk(KERN_ERR "btrfs end < start %llu %llu\n",
 364                        (unsigned long long)end,
 365                        (unsigned long long)start);
 366                 WARN_ON(1);
 367         }
 368         if (bits & EXTENT_DIRTY)
 369                 tree->dirty_bytes += end - start + 1;
 370         set_state_cb(tree, state, bits);
 371         state->state |= bits;
 372         state->start = start;
 373         state->end = end;
 374         node = tree_insert(&tree->state, end, &state->rb_node);
 375         if (node) {
 376                 struct extent_state *found;
 377                 found = rb_entry(node, struct extent_state, rb_node);
 378                 printk(KERN_ERR "btrfs found node %llu %llu on insert of "
 379                        "%llu %llu\n", (unsigned long long)found->start,
 380                        (unsigned long long)found->end,
 381                        (unsigned long long)start, (unsigned long long)end);
 382                 free_extent_state(state);
 383                 return -EEXIST;
 384         }
 385         state->tree = tree;
 386         merge_state(tree, state);
 387         return 0;
 388 }
 389
 390 /*
 391  * split a given extent state struct in two, inserting the preallocated
 392  * struct 'prealloc' as the newly created second half.  'split' indicates an
 393  * offset inside 'orig' where it should be split.
 394  *
 395  * Before calling,
 396  * the tree has 'orig' at [orig->start, orig->end].  After calling, there
 397  * are two extent state structs in the tree:
 398  * prealloc: [orig->start, split - 1]
 399  * orig: [ split, orig->end ]
 400  *
 401  * The tree locks are not taken by this function. They need to be held
 402  * by the caller.
 403  */
 404 static int split_state(struct extent_io_tree *tree, struct extent_state *orig,
 405                        struct extent_state *prealloc, u64 split)
 406 {
 407         struct rb_node *node;
 408         prealloc->start = orig->start;
 409         prealloc->end = split - 1;
 410         prealloc->state = orig->state;
 411         orig->start = split;
 412
 413         node = tree_insert(&tree->state, prealloc->end, &prealloc->rb_node);
 414         if (node) {
 415                 free_extent_state(prealloc);
 416                 return -EEXIST;
 417         }
 418         prealloc->tree = tree;
 419         return 0;
 420 }
 421
 422 /*
 423  * utility function to clear some bits in an extent state struct.
 424  * it will optionally wake up any one waiting on this state (wake == 1), or
 425  * forcibly remove the state from the tree (delete == 1).
 426  *
 427  * If no bits are set on the state struct after clearing things, the
 428  * struct is freed and removed from the tree
 429  */
 430 static int clear_state_bit(struct extent_io_tree *tree,
 431                             struct extent_state *state, int bits, int wake,
 432                             int delete)
 433 {
 434         int ret = state->state & bits;
 435
 436         if ((bits & EXTENT_DIRTY) && (state->state & EXTENT_DIRTY)) {
 437                 u64 range = state->end - state->start + 1;
 438                 WARN_ON(range > tree->dirty_bytes);
 439                 tree->dirty_bytes -= range;
 440         }
 441         clear_state_cb(tree, state, bits);
 442         state->state &= ~bits;
 443         if (wake)
 444                 wake_up(&state->wq);
 445         if (delete || state->state == 0) {
 446                 if (state->tree) {
 447                         clear_state_cb(tree, state, state->state);
 448                         rb_erase(&state->rb_node, &tree->state);
 449                         state->tree = NULL;
 450                         free_extent_state(state);
 451                 } else {
 452                         WARN_ON(1);
 453                 }
 454         } else {
 455                 merge_state(tree, state);
 456         }
 457         return ret;
 458 }
 459
 460 /*
 461  * clear some bits on a range in the tree.  This may require splitting
 462  * or inserting elements in the tree, so the gfp mask is used to
 463  * indicate which allocations or sleeping are allowed.
 464  *
 465  * pass 'wake' == 1 to kick any sleepers, and 'delete' == 1 to remove
 466  * the given range from the tree regardless of state (ie for truncate).
 467  *
 468  * the range [start, end] is inclusive.
 469  *
 470  * This takes the tree lock, and returns < 0 on error, > 0 if any of the
 471  * bits were already set, or zero if none of the bits were already set.
 472  */
 473 int clear_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 474                      int bits, int wake, int delete, gfp_t mask)
 475 {
 476         struct extent_state *state;
 477         struct extent_state *prealloc = NULL;
 478         struct rb_node *node;
 479         int err;
 480         int set = 0;
 481
 482 again:
 483         if (!prealloc && (mask & __GFP_WAIT)) {
 484                 prealloc = alloc_extent_state(mask);
 485                 if (!prealloc)
 486                         return -ENOMEM;
 487         }
 488
 489         spin_lock(&tree->lock);
 490         /*
 491          * this search will find the extents that end after
 492          * our range starts
 493          */
 494         node = tree_search(tree, start);
 495         if (!node)
 496                 goto out;
 497         state = rb_entry(node, struct extent_state, rb_node);
 498         if (state->start > end)
 499                 goto out;
 500         WARN_ON(state->end < start);
 501
 502         /*
 503          *     | ---- desired range ---- |
 504          *  | state | or
 505          *  | ------------- state -------------- |
 506          *
 507          * We need to split the extent we found, and may flip
 508          * bits on second half.
 509          *
 510          * If the extent we found extends past our range, we
 511          * just split and search again.  It'll get split again
 512          * the next time though.
 513          *
 514          * If the extent we found is inside our range, we clear
 515          * the desired bit on it.
 516          */
 517
 518         if (state->start < start) {
 519                 if (!prealloc)
 520                         prealloc = alloc_extent_state(GFP_ATOMIC);
 521                 err = split_state(tree, state, prealloc, start);
 522                 BUG_ON(err == -EEXIST);
 523                 prealloc = NULL;
 524                 if (err)
 525                         goto out;
 526                 if (state->end <= end) {
 527                         start = state->end + 1;
 528                         set |= clear_state_bit(tree, state, bits,
 529                                         wake, delete);
 530                 } else {
 531                         start = state->start;
 532                 }
 533                 goto search_again;
 534         }
 535         /*
 536          * | ---- desired range ---- |
 537          *                        | state |
 538          * We need to split the extent, and clear the bit
 539          * on the first half
 540          */
 541         if (state->start <= end && state->end > end) {
 542                 if (!prealloc)
 543                         prealloc = alloc_extent_state(GFP_ATOMIC);
 544                 err = split_state(tree, state, prealloc, end + 1);
 545                 BUG_ON(err == -EEXIST);
 546
 547                 if (wake)
 548                         wake_up(&state->wq);
 549                 set |= clear_state_bit(tree, prealloc, bits,
 550                                        wake, delete);
 551                 prealloc = NULL;
 552                 goto out;
 553         }
 554
 555         start = state->end + 1;
 556         set |= clear_state_bit(tree, state, bits, wake, delete);
 557         goto search_again;
 558
 559 out:
 560         spin_unlock(&tree->lock);
 561         if (prealloc)
 562                 free_extent_state(prealloc);
 563
 564         return set;
 565
 566 search_again:
 567         if (start > end)
 568                 goto out;
 569         spin_unlock(&tree->lock);
 570         if (mask & __GFP_WAIT)
 571                 cond_resched();
 572         goto again;
 573 }
 574
 575 static int wait_on_state(struct extent_io_tree *tree,
 576                          struct extent_state *state)
 577                 __releases(tree->lock)
 578                 __acquires(tree->lock)
 579 {
 580         DEFINE_WAIT(wait);
 581         prepare_to_wait(&state->wq, &wait, TASK_UNINTERRUPTIBLE);
 582         spin_unlock(&tree->lock);
 583         schedule();
 584         spin_lock(&tree->lock);
 585         finish_wait(&state->wq, &wait);
 586         return 0;
 587 }
 588
 589 /*
 590  * waits for one or more bits to clear on a range in the state tree.
 591  * The range [start, end] is inclusive.
 592  * The tree lock is taken by this function
 593  */
 594 int wait_extent_bit(struct extent_io_tree *tree, u64 start, u64 end, int bits)
 595 {
 596         struct extent_state *state;
 597         struct rb_node *node;
 598
 599         spin_lock(&tree->lock);
 600 again:
 601         while (1) {
 602                 /*
 603                  * this search will find all the extents that end after
 604                  * our range starts
 605                  */
 606                 node = tree_search(tree, start);
 607                 if (!node)
 608                         break;
 609
 610                 state = rb_entry(node, struct extent_state, rb_node);
 611
 612                 if (state->start > end)
 613                         goto out;
 614
 615                 if (state->state & bits) {
 616                         start = state->start;
 617                         atomic_inc(&state->refs);
 618                         wait_on_state(tree, state);
 619                         free_extent_state(state);
 620                         goto again;
 621                 }
 622                 start = state->end + 1;
 623
 624                 if (start > end)
 625                         break;
 626
 627                 if (need_resched()) {
 628                         spin_unlock(&tree->lock);
 629                         cond_resched();
 630                         spin_lock(&tree->lock);
 631                 }
 632         }
 633 out:
 634         spin_unlock(&tree->lock);
 635         return 0;
 636 }
 637
 638 static void set_state_bits(struct extent_io_tree *tree,
 639                            struct extent_state *state,
 640                            int bits)
 641 {
 642         if ((bits & EXTENT_DIRTY) && !(state->state & EXTENT_DIRTY)) {
 643                 u64 range = state->end - state->start + 1;
 644                 tree->dirty_bytes += range;
 645         }
 646         set_state_cb(tree, state, bits);
 647         state->state |= bits;
 648 }
 649
 650 /*
 651  * set some bits on a range in the tree.  This may require allocations
 652  * or sleeping, so the gfp mask is used to indicate what is allowed.
 653  *
 654  * If 'exclusive' == 1, this will fail with -EEXIST if some part of the
 655  * range already has the desired bits set.  The start of the existing
 656  * range is returned in failed_start in this case.
 657  *
 658  * [start, end] is inclusive
 659  * This takes the tree lock.
 660  */
 661 static int set_extent_bit(struct extent_io_tree *tree, u64 start, u64 end,
 662                           int bits, int exclusive, u64 *failed_start,
 663                           gfp_t mask)
 664 {
 665         struct extent_state *state;
 666         struct extent_state *prealloc = NULL;
 667         struct rb_node *node;
 668         int err = 0;
 669         int set;
 670         u64 last_start;
 671         u64 last_end;
 672 again:
 673         if (!prealloc && (mask & __GFP_WAIT)) {
 674                 prealloc = alloc_extent_state(mask);
 675                 if (!prealloc)
 676                         return -ENOMEM;
 677         }
 678
 679         spin_lock(&tree->lock);
 680         /*
 681          * this search will find all the extents that end after
 682          * our range starts.
 683          */
 684         node = tree_search(tree, start);
 685         if (!node) {
 686                 err = insert_state(tree, prealloc, start, end, bits);
 687                 prealloc = NULL;
 688                 BUG_ON(err == -EEXIST);
 689                 goto out;
 690         }
 691
 692         state = rb_entry(node, struct extent_state, rb_node);
 693         last_start = state->start;
 694         last_end = state->end;
 695
 696         /*
 697          * | ---- desired range ---- |
 698          * | state |
 699          *
 700          * Just lock what we found and keep going
 701          */
 702         if (state->start == start && state->end <= end) {
 703                 set = state->state & bits;
 704                 if (set && exclusive) {
 705                         *failed_start = state->start;
 706                         err = -EEXIST;
 707                         goto out;
 708                 }
 709                 set_state_bits(tree, state, bits);
 710                 start = state->end + 1;
 711                 merge_state(tree, state);
 712                 goto search_again;
 713         }
 714
 715         /*
 716          *     | ---- desired range ---- |
 717          * | state |
 718          *   or
 719          * | ------------- state -------------- |
 720          *
 721          * We need to split the extent we found, and may flip bits on
 722          * second half.
 723          *
 724          * If the extent we found extends past our
 725          * range, we just split and search again.  It'll get split
 726          * again the next time though.
 727          *
 728          * If the extent we found is inside our range, we set the
 729          * desired bit on it.
 730          */
 731         if (state->start < start) {
 732                 set = state->state & bits;
 733                 if (exclusive && set) {
 734                         *failed_start = start;
 735                         err = -EEXIST;
 736                         goto out;
 737                 }
 738                 err = split_state(tree, state, prealloc, start);
 739                 BUG_ON(err == -EEXIST);
 740                 prealloc = NULL;
 741                 if (err)
 742                         goto out;
 743                 if (state->end <= end) {
 744                         set_state_bits(tree, state, bits);
 745                         start = state->end + 1;
 746                         merge_state(tree, state);
 747                 } else {
 748                         start = state->start;
 749                 }
 750                 goto search_again;
 751         }
 752         /*
 753          * | ---- desired range ---- |
 754          *     | state | or               | state |
 755          *
 756          * There's a hole, we need to insert something in it and
 757          * ignore the extent we found.
 758          */
 759         if (state->start > start) {
 760                 u64 this_end;
 761                 if (end < last_start)
 762                         this_end = end;
 763                 else
 764                         this_end = last_start - 1;
 765                 err = insert_state(tree, prealloc, start, this_end,
 766                                    bits);
 767                 prealloc = NULL;
 768                 BUG_ON(err == -EEXIST);
 769                 if (err)
 770                         goto out;
 771                 start = this_end + 1;
 772                 goto search_again;
 773         }
 774         /*
 775          * | ---- desired range ---- |
 776          *                        | state |
 777          * We need to split the extent, and set the bit
 778          * on the first half
 779          */
 780         if (state->start <= end && state->end > end) {
 781                 set = state->state & bits;
 782                 if (exclusive && set) {
 783                         *failed_start = start;
 784                         err = -EEXIST;
 785                         goto out;
 786                 }
 787                 err = split_state(tree, state, prealloc, end + 1);
 788                 BUG_ON(err == -EEXIST);
 789
 790                 set_state_bits(tree, prealloc, bits);
 791                 merge_state(tree, prealloc);
 792                 prealloc = NULL;
 793                 goto out;
 794         }
 795
 796         goto search_again;
 797
 798 out:
 799         spin_unlock(&tree->lock);
 800         if (prealloc)
 801                 free_extent_state(prealloc);
 802
 803         return err;
 804
 805 search_again:
 806         if (start > end)
 807                 goto out;
 808         spin_unlock(&tree->lock);
 809         if (mask & __GFP_WAIT)
 810                 cond_resched();
 811         goto again;
 812 }
 813
 814 /* wrappers around set/clear extent bit */
 815 int set_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
 816                      gfp_t mask)
 817 {
 818         return set_extent_bit(tree, start, end, EXTENT_DIRTY, 0, NULL,
 819                               mask);
 820 }
 821
 822 int set_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end,
 823                        gfp_t mask)
 824 {
 825         return set_extent_bit(tree, start, end, EXTENT_ORDERED, 0, NULL, mask);
 826 }
 827
 828 int set_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
 829                     int bits, gfp_t mask)
 830 {
 831         return set_extent_bit(tree, start, end, bits, 0, NULL,
 832                               mask);
 833 }
 834
 835 int clear_extent_bits(struct extent_io_tree *tree, u64 start, u64 end,
 836                       int bits, gfp_t mask)
 837 {
 838         return clear_extent_bit(tree, start, end, bits, 0, 0, mask);
 839 }
 840
 841 int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end,
 842                      gfp_t mask)
 843 {
 844         return set_extent_bit(tree, start, end,
 845                               EXTENT_DELALLOC | EXTENT_DIRTY,
 846                               0, NULL, mask);
 847 }
 848
 849 int clear_extent_dirty(struct extent_io_tree *tree, u64 start, u64 end,
 850                        gfp_t mask)
 851 {
 852         return clear_extent_bit(tree, start, end,
 853                                 EXTENT_DIRTY | EXTENT_DELALLOC, 0, 0, mask);
 854 }
 855
 856 int clear_extent_ordered(struct extent_io_tree *tree, u64 start, u64 end,
 857                          gfp_t mask)
 858 {
 859         return clear_extent_bit(tree, start, end, EXTENT_ORDERED, 1, 0, mask);
 860 }
 861
 862 int set_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
 863                      gfp_t mask)
 864 {
 865         return set_extent_bit(tree, start, end, EXTENT_NEW, 0, NULL,
 866                               mask);
 867 }
 868
 869 static int clear_extent_new(struct extent_io_tree *tree, u64 start, u64 end,
 870                        gfp_t mask)
 871 {
 872         return clear_extent_bit(tree, start, end, EXTENT_NEW, 0, 0, mask);
 873 }
 874
 875 int set_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
 876                         gfp_t mask)
 877 {
 878         return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, NULL,
 879                               mask);
 880 }
 881
 882 static int clear_extent_uptodate(struct extent_io_tree *tree, u64 start,
 883                                  u64 end, gfp_t mask)
 884 {
 885         return clear_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, 0, mask);
 886 }
 887
 888 static int set_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end,
 889                          gfp_t mask)
 890 {
 891         return set_extent_bit(tree, start, end, EXTENT_WRITEBACK,
 892                               0, NULL, mask);
 893 }
 894
 895 static int clear_extent_writeback(struct extent_io_tree *tree, u64 start,
 896                                   u64 end, gfp_t mask)
 897 {
 898         return clear_extent_bit(tree, start, end, EXTENT_WRITEBACK, 1, 0, mask);
 899 }
 900
 901 int wait_on_extent_writeback(struct extent_io_tree *tree, u64 start, u64 end)
 902 {
 903         return wait_extent_bit(tree, start, end, EXTENT_WRITEBACK);
 904 }
 905
 906 /*
 907  * either insert or lock state struct between start and end use mask to tell
 908  * us if waiting is desired.
 909  */
 910 int lock_extent(struct extent_io_tree *tree, u64 start, u64 end, gfp_t mask)
 911 {
 912         int err;
 913         u64 failed_start;
 914         while (1) {
 915                 err = set_extent_bit(tree, start, end, EXTENT_LOCKED, 1,
 916                                      &failed_start, mask);
 917                 if (err == -EEXIST && (mask & __GFP_WAIT)) {
 918                         wait_extent_bit(tree, failed_start, end, EXTENT_LOCKED);
 919                         start = failed_start;
 920                 } else {
 921                         break;
 922                 }
 923                 WARN_ON(start > end);
 924         }
 925         return err;
 926 }
 927
 928 int try_lock_extent(struct extent_io_tree *tree, u64 start, u64 end,
 929                     gfp_t mask)
 930 {
 931         int err;
 932         u64 failed_start;
 933
 934         err = set_extent_bit(tree, start, end, EXTENT_LOCKED, 1,
 935                              &failed_start, mask);
 936         if (err == -EEXIST) {
 937                 if (failed_start > start)
 938                         clear_extent_bit(tree, start, failed_start - 1,
 939                                          EXTENT_LOCKED, 1, 0, mask);
 940                 return 0;
 941         }
 942         return 1;
 943 }
 944
 945 int unlock_extent(struct extent_io_tree *tree, u64 start, u64 end,
 946                   gfp_t mask)
 947 {
 948         return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, mask);
 949 }
 950
 951 /*
 952  * helper function to set pages and extents in the tree dirty
 953  */
 954 int set_range_dirty(struct extent_io_tree *tree, u64 start, u64 end)
 955 {
 956         unsigned long index = start >> PAGE_CACHE_SHIFT;
 957         unsigned long end_index = end >> PAGE_CACHE_SHIFT;
 958         struct page *page;
 959
 960         while (index <= end_index) {
 961                 page = find_get_page(tree->mapping, index);
 962                 BUG_ON(!page);
 963                 __set_page_dirty_nobuffers(page);
 964                 page_cache_release(page);
 965                 index++;
 966         }
 967         set_extent_dirty(tree, start, end, GFP_NOFS);
 968         return 0;
 969 }
 970
 971 /*
 972  * helper function to set both pages and extents in the tree writeback
 973  */
 974 static int set_range_writeback(struct extent_io_tree *tree, u64 start, u64 end)
 975 {
 976         unsigned long index = start >> PAGE_CACHE_SHIFT;
 977         unsigned long end_index = end >> PAGE_CACHE_SHIFT;
 978         struct page *page;
 979
 980         while (index <= end_index) {
 981                 page = find_get_page(tree->mapping, index);
 982                 BUG_ON(!page);
 983                 set_page_writeback(page);
 984                 page_cache_release(page);
 985                 index++;
 986         }
 987         set_extent_writeback(tree, start, end, GFP_NOFS);
 988         return 0;
 989 }
 990
 991 /*
 992  * find the first offset in the io tree with 'bits' set. zero is
 993  * returned if we find something, and *start_ret and *end_ret are
 994  * set to reflect the state struct that was found.
 995  *
 996  * If nothing was found, 1 is returned, < 0 on error
 997  */
 998 int find_first_extent_bit(struct extent_io_tree *tree, u64 start,
 999                           u64 *start_ret, u64 *end_ret, int bits)
1000 {
1001         struct rb_node *node;
1002         struct extent_state *state;
1003         int ret = 1;
1004
1005         spin_lock(&tree->lock);
1006         /*
1007          * this search will find all the extents that end after
1008          * our range starts.
1009          */
1010         node = tree_search(tree, start);
1011         if (!node)
1012                 goto out;
1013
1014         while (1) {
1015                 state = rb_entry(node, struct extent_state, rb_node);
1016                 if (state->end >= start && (state->state & bits)) {
1017                         *start_ret = state->start;
1018                         *end_ret = state->end;
1019                         ret = 0;
1020                         break;
1021                 }
1022                 node = rb_next(node);
1023                 if (!node)
1024                         break;
1025         }
1026 out:
1027         spin_unlock(&tree->lock);
1028         return ret;
1029 }
1030
1031 /* find the first state struct with 'bits' set after 'start', and
1032  * return it.  tree->lock must be held.  NULL will returned if
1033  * nothing was found after 'start'
1034  */
1035 struct extent_state *find_first_extent_bit_state(struct extent_io_tree *tree,
1036                                                  u64 start, int bits)
1037 {
1038         struct rb_node *node;
1039         struct extent_state *state;
1040
1041         /*
1042          * this search will find all the extents that end after
1043          * our range starts.
1044          */
1045         node = tree_search(tree, start);
1046         if (!node)
1047                 goto out;
1048
1049         while (1) {
1050                 state = rb_entry(node, struct extent_state, rb_node);
1051                 if (state->end >= start && (state->state & bits))
1052                         return state;
1053
1054                 node = rb_next(node);
1055                 if (!node)
1056                         break;
1057         }
1058 out:
1059         return NULL;
1060 }
1061
1062 /*
1063  * find a contiguous range of bytes in the file marked as delalloc, not
1064  * more than 'max_bytes'.  start and end are used to return the range,
1065  *
1066  * 1 is returned if we find something, 0 if nothing was in the tree
1067  */
1068 static noinline u64 find_delalloc_range(struct extent_io_tree *tree,
1069                                         u64 *start, u64 *end, u64 max_bytes)
1070 {
1071         struct rb_node *node;
1072         struct extent_state *state;
1073         u64 cur_start = *start;
1074         u64 found = 0;
1075         u64 total_bytes = 0;
1076
1077         spin_lock(&tree->lock);
1078
1079         /*
1080          * this search will find all the extents that end after
1081          * our range starts.
1082          */
1083         node = tree_search(tree, cur_start);
1084         if (!node) {
1085                 if (!found)
1086                         *end = (u64)-1;
1087                 goto out;
1088         }
1089
1090         while (1) {
1091                 state = rb_entry(node, struct extent_state, rb_node);
1092                 if (found && (state->start != cur_start ||
1093                               (state->state & EXTENT_BOUNDARY))) {
1094                         goto out;
1095                 }
1096                 if (!(state->state & EXTENT_DELALLOC)) {
1097                         if (!found)
1098                                 *end = state->end;
1099                         goto out;
1100                 }
1101                 if (!found)
1102                         *start = state->start;
1103                 found++;
1104                 *end = state->end;
1105                 cur_start = state->end + 1;
1106                 node = rb_next(node);
1107                 if (!node)
1108                         break;
1109                 total_bytes += state->end - state->start + 1;
1110                 if (total_bytes >= max_bytes)
1111                         break;
1112         }
1113 out:
1114         spin_unlock(&tree->lock);
1115         return found;
1116 }
1117
1118 static noinline int __unlock_for_delalloc(struct inode *inode,
1119                                           struct page *locked_page,
1120                                           u64 start, u64 end)
1121 {
1122         int ret;
1123         struct page *pages[16];
1124         unsigned long index = start >> PAGE_CACHE_SHIFT;
1125         unsigned long end_index = end >> PAGE_CACHE_SHIFT;
1126         unsigned long nr_pages = end_index - index + 1;
1127         int i;
1128
1129         if (index == locked_page->index && end_index == index)
1130                 return 0;
1131
1132         while (nr_pages > 0) {
1133                 ret = find_get_pages_contig(inode->i_mapping, index,
1134                                      min_t(unsigned long, nr_pages,
1135                                      ARRAY_SIZE(pages)), pages);
1136                 for (i = 0; i < ret; i++) {
1137                         if (pages[i] != locked_page)
1138                                 unlock_page(pages[i]);
1139                         page_cache_release(pages[i]);
1140                 }
1141                 nr_pages -= ret;
1142                 index += ret;
1143                 cond_resched();
1144         }
1145         return 0;
1146 }
1147
1148 static noinline int lock_delalloc_pages(struct inode *inode,
1149                                         struct page *locked_page,
1150                                         u64 delalloc_start,
1151                                         u64 delalloc_end)
1152 {
1153         unsigned long index = delalloc_start >> PAGE_CACHE_SHIFT;
1154         unsigned long start_index = index;
1155         unsigned long end_index = delalloc_end >> PAGE_CACHE_SHIFT;
1156         unsigned long pages_locked = 0;
1157         struct page *pages[16];
1158         unsigned long nrpages;
1159         int ret;
1160         int i;
1161
1162         /* the caller is responsible for locking the start index */
1163         if (index == locked_page->index && index == end_index)
1164                 return 0;
1165
1166         /* skip the page at the start index */
1167         nrpages = end_index - index + 1;
1168         while (nrpages > 0) {
1169                 ret = find_get_pages_contig(inode->i_mapping, index,
1170                                      min_t(unsigned long,
1171                                      nrpages, ARRAY_SIZE(pages)), pages);
1172                 if (ret == 0) {
1173                         ret = -EAGAIN;
1174                         goto done;
1175                 }
1176                 /* now we have an array of pages, lock them all */
1177                 for (i = 0; i < ret; i++) {
1178                         /*
1179                          * the caller is taking responsibility for
1180                          * locked_page
1181                          */
1182                         if (pages[i] != locked_page) {
1183                                 lock_page(pages[i]);
1184                                 if (!PageDirty(pages[i]) ||
1185                                     pages[i]->mapping != inode->i_mapping) {
1186                                         ret = -EAGAIN;
1187                                         unlock_page(pages[i]);
1188                                         page_cache_release(pages[i]);
1189                                         goto done;
1190                                 }
1191                         }
1192                         page_cache_release(pages[i]);
1193                         pages_locked++;
1194                 }
1195                 nrpages -= ret;
1196                 index += ret;
1197                 cond_resched();
1198         }
1199         ret = 0;
1200 done:
1201         if (ret && pages_locked) {
1202                 __unlock_for_delalloc(inode, locked_page,
1203                               delalloc_start,
1204                               ((u64)(start_index + pages_locked - 1)) <<
1205                               PAGE_CACHE_SHIFT);
1206         }
1207         return ret;
1208 }
1209
1210 /*
1211  * find a contiguous range of bytes in the file marked as delalloc, not
1212  * more than 'max_bytes'.  start and end are used to return the range,
1213  *
1214  * 1 is returned if we find something, 0 if nothing was in the tree
1215  */
1216 static noinline u64 find_lock_delalloc_range(struct inode *inode,
1217                                              struct extent_io_tree *tree,
1218                                              struct page *locked_page,
1219                                              u64 *start, u64 *end,
1220                                              u64 max_bytes)
1221 {
1222         u64 delalloc_start;
1223         u64 delalloc_end;
1224         u64 found;
1225         int ret;
1226         int loops = 0;
1227
1228 again:
1229         /* step one, find a bunch of delalloc bytes starting at start */
1230         delalloc_start = *start;
1231         delalloc_end = 0;
1232         found = find_delalloc_range(tree, &delalloc_start, &delalloc_end,
1233                                     max_bytes);
1234         if (!found || delalloc_end <= *start) {
1235                 *start = delalloc_start;
1236                 *end = delalloc_end;
1237                 return found;
1238         }
1239
1240         /*
1241          * start comes from the offset of locked_page.  We have to lock
1242          * pages in order, so we can't process delalloc bytes before
1243          * locked_page
1244          */
1245         if (delalloc_start < *start)
1246                 delalloc_start = *start;
1247
1248         /*
1249          * make sure to limit the number of pages we try to lock down
1250          * if we're looping.
1251          */
1252         if (delalloc_end + 1 - delalloc_start > max_bytes && loops)
1253                 delalloc_end = delalloc_start + PAGE_CACHE_SIZE - 1;
1254
1255         /* step two, lock all the pages after the page that has start */
1256         ret = lock_delalloc_pages(inode, locked_page,
1257                                   delalloc_start, delalloc_end);
1258         if (ret == -EAGAIN) {
1259                 /* some of the pages are gone, lets avoid looping by
1260                  * shortening the size of the delalloc range we're searching
1261                  */
1262                 if (!loops) {
1263                         unsigned long offset = (*start) & (PAGE_CACHE_SIZE - 1);
1264                         max_bytes = PAGE_CACHE_SIZE - offset;
1265                         loops = 1;
1266                         goto again;
1267                 } else {
1268                         found = 0;
1269                         goto out_failed;
1270                 }
1271         }
1272         BUG_ON(ret);
1273
1274         /* step three, lock the state bits for the whole range */
1275         lock_extent(tree, delalloc_start, delalloc_end, GFP_NOFS);
1276
1277         /* then test to make sure it is all still delalloc */
1278         ret = test_range_bit(tree, delalloc_start, delalloc_end,
1279                              EXTENT_DELALLOC, 1);
1280         if (!ret) {
1281                 unlock_extent(tree, delalloc_start, delalloc_end, GFP_NOFS);
1282                 __unlock_for_delalloc(inode, locked_page,
1283                               delalloc_start, delalloc_end);
1284                 cond_resched();
1285                 goto again;
1286         }
1287         *start = delalloc_start;
1288         *end = delalloc_end;
1289 out_failed:
1290         return found;
1291 }
1292
1293 int extent_clear_unlock_delalloc(struct inode *inode,
1294                                 struct extent_io_tree *tree,
1295                                 u64 start, u64 end, struct page *locked_page,
1296                                 int unlock_pages,
1297                                 int clear_unlock,
1298                                 int clear_delalloc, int clear_dirty,
1299                                 int set_writeback,
1300                                 int end_writeback)
1301 {
1302         int ret;
1303         struct page *pages[16];
1304         unsigned long index = start >> PAGE_CACHE_SHIFT;
1305         unsigned long end_index = end >> PAGE_CACHE_SHIFT;
1306         unsigned long nr_pages = end_index - index + 1;
1307         int i;
1308         int clear_bits = 0;
1309
1310         if (clear_unlock)
1311                 clear_bits |= EXTENT_LOCKED;
1312         if (clear_dirty)
1313                 clear_bits |= EXTENT_DIRTY;
1314
1315         if (clear_delalloc)
1316                 clear_bits |= EXTENT_DELALLOC;
1317
1318         clear_extent_bit(tree, start, end, clear_bits, 1, 0, GFP_NOFS);
1319         if (!(unlock_pages || clear_dirty || set_writeback || end_writeback))
1320                 return 0;
1321
1322         while (nr_pages > 0) {
1323                 ret = find_get_pages_contig(inode->i_mapping, index,
1324                                      min_t(unsigned long,
1325                                      nr_pages, ARRAY_SIZE(pages)), pages);
1326                 for (i = 0; i < ret; i++) {
1327                         if (pages[i] == locked_page) {
1328                                 page_cache_release(pages[i]);
1329                                 continue;
1330                         }
1331                         if (clear_dirty)
1332                                 clear_page_dirty_for_io(pages[i]);
1333                         if (set_writeback)
1334                                 set_page_writeback(pages[i]);
1335                         if (end_writeback)
1336                                 end_page_writeback(pages[i]);
1337                         if (unlock_pages)
1338                                 unlock_page(pages[i]);
1339                         page_cache_release(pages[i]);
1340                 }
1341                 nr_pages -= ret;
1342                 index += ret;
1343                 cond_resched();
1344         }
1345         return 0;
1346 }
1347
1348 /*
1349  * count the number of bytes in the tree that have a given bit(s)
1350  * set.  This can be fairly slow, except for EXTENT_DIRTY which is
1351  * cached.  The total number found is returned.
1352  */
1353 u64 count_range_bits(struct extent_io_tree *tree,
1354                      u64 *start, u64 search_end, u64 max_bytes,
1355                      unsigned long bits)
1356 {
1357         struct rb_node *node;
1358         struct extent_state *state;
1359         u64 cur_start = *start;
1360         u64 total_bytes = 0;
1361         int found = 0;
1362
1363         if (search_end <= cur_start) {
1364                 WARN_ON(1);
1365                 return 0;
1366         }
1367
1368         spin_lock(&tree->lock);
1369         if (cur_start == 0 && bits == EXTENT_DIRTY) {
1370                 total_bytes = tree->dirty_bytes;
1371                 goto out;
1372         }
1373         /*
1374          * this search will find all the extents that end after
1375          * our range starts.
1376          */
1377         node = tree_search(tree, cur_start);
1378         if (!node)
1379                 goto out;
1380
1381         while (1) {
1382                 state = rb_entry(node, struct extent_state, rb_node);
1383                 if (state->start > search_end)
1384                         break;
1385                 if (state->end >= cur_start && (state->state & bits)) {
1386                         total_bytes += min(search_end, state->end) + 1 -
1387                                        max(cur_start, state->start);
1388                         if (total_bytes >= max_bytes)
1389                                 break;
1390                         if (!found) {
1391                                 *start = state->start;
1392                                 found = 1;
1393                         }
1394                 }
1395                 node = rb_next(node);
1396                 if (!node)
1397                         break;
1398         }
1399 out:
1400         spin_unlock(&tree->lock);
1401         return total_bytes;
1402 }
1403
1404 #if 0
1405 /*
1406  * helper function to lock both pages and extents in the tree.
1407  * pages must be locked first.
1408  */
1409 static int lock_range(struct extent_io_tree *tree, u64 start, u64 end)
1410 {
1411         unsigned long index = start >> PAGE_CACHE_SHIFT;
1412         unsigned long end_index = end >> PAGE_CACHE_SHIFT;
1413         struct page *page;
1414         int err;
1415
1416         while (index <= end_index) {
1417                 page = grab_cache_page(tree->mapping, index);
1418                 if (!page) {
1419                         err = -ENOMEM;
1420                         goto failed;
1421                 }
1422                 if (IS_ERR(page)) {
1423                         err = PTR_ERR(page);
1424                         goto failed;
1425                 }
1426                 index++;
1427         }
1428         lock_extent(tree, start, end, GFP_NOFS);
1429         return 0;
1430
1431 failed:
1432         /*
1433          * we failed above in getting the page at 'index', so we undo here
1434          * up to but not including the page at 'index'
1435          */
1436         end_index = index;
1437         index = start >> PAGE_CACHE_SHIFT;
1438         while (index < end_index) {
1439                 page = find_get_page(tree->mapping, index);
1440                 unlock_page(page);
1441                 page_cache_release(page);
1442                 index++;
1443         }
1444         return err;
1445 }
1446
1447 /*
1448  * helper function to unlock both pages and extents in the tree.
1449  */
1450 static int unlock_range(struct extent_io_tree *tree, u64 start, u64 end)
1451 {
1452         unsigned long index = start >> PAGE_CACHE_SHIFT;
1453         unsigned long end_index = end >> PAGE_CACHE_SHIFT;
1454         struct page *page;
1455
1456         while (index <= end_index) {
1457                 page = find_get_page(tree->mapping, index);
1458                 unlock_page(page);
1459                 page_cache_release(page);
1460                 index++;
1461         }
1462         unlock_extent(tree, start, end, GFP_NOFS);
1463         return 0;
1464 }
1465 #endif
1466
1467 /*
1468  * set the private field for a given byte offset in the tree.  If there isn't
1469  * an extent_state there already, this does nothing.
1470  */
1471 int set_state_private(struct extent_io_tree *tree, u64 start, u64 private)
1472 {
1473         struct rb_node *node;
1474         struct extent_state *state;
1475         int ret = 0;
1476
1477         spin_lock(&tree->lock);
1478         /*
1479          * this search will find all the extents that end after
1480          * our range starts.
1481          */
1482         node = tree_search(tree, start);
1483         if (!node) {
1484                 ret = -ENOENT;
1485                 goto out;
1486         }
1487         state = rb_entry(node, struct extent_state, rb_node);
1488         if (state->start != start) {
1489                 ret = -ENOENT;
1490                 goto out;
1491         }
1492         state->private = private;
1493 out:
1494         spin_unlock(&tree->lock);
1495         return ret;
1496 }
1497
1498 int get_state_private(struct extent_io_tree *tree, u64 start, u64 *private)
1499 {
1500         struct rb_node *node;
1501         struct extent_state *state;
1502         int ret = 0;
1503
1504         spin_lock(&tree->lock);
1505         /*
1506          * this search will find all the extents that end after
1507          * our range starts.
1508          */
1509         node = tree_search(tree, start);
1510         if (!node) {
1511                 ret = -ENOENT;
1512                 goto out;
1513         }
1514         state = rb_entry(node, struct extent_state, rb_node);
1515         if (state->start != start) {
1516                 ret = -ENOENT;
1517                 goto out;
1518         }
1519         *private = state->private;
1520 out:
1521         spin_unlock(&tree->lock);
1522         return ret;
1523 }
1524
1525 /*
1526  * searches a range in the state tree for a given mask.
1527  * If 'filled' == 1, this returns 1 only if every extent in the tree
1528  * has the bits set.  Otherwise, 1 is returned if any bit in the
1529  * range is found set.
1530  */
1531 int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end,
1532                    int bits, int filled)
1533 {
1534         struct extent_state *state = NULL;
1535         struct rb_node *node;
1536         int bitset = 0;
1537
1538         spin_lock(&tree->lock);
1539         node = tree_search(tree, start);
1540         while (node && start <= end) {
1541                 state = rb_entry(node, struct extent_state, rb_node);
1542
1543                 if (filled && state->start > start) {
1544                         bitset = 0;
1545                         break;
1546                 }
1547
1548                 if (state->start > end)
1549                         break;
1550
1551                 if (state->state & bits) {
1552                         bitset = 1;
1553                         if (!filled)
1554                                 break;
1555                 } else if (filled) {
1556                         bitset = 0;
1557                         break;
1558                 }
1559                 start = state->end + 1;
1560                 if (start > end)
1561                         break;
1562                 node = rb_next(node);
1563                 if (!node) {
1564                         if (filled)
1565                                 bitset = 0;
1566                         break;
1567                 }
1568         }
1569         spin_unlock(&tree->lock);
1570         return bitset;
1571 }
1572
1573 /*
1574  * helper function to set a given page up to date if all the
1575  * extents in the tree for that page are up to date
1576  */
1577 static int check_page_uptodate(struct extent_io_tree *tree,
1578                                struct page *page)
1579 {
1580         u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
1581         u64 end = start + PAGE_CACHE_SIZE - 1;
1582         if (test_range_bit(tree, start, end, EXTENT_UPTODATE, 1))
1583                 SetPageUptodate(page);
1584         return 0;
1585 }
1586
1587 /*
1588  * helper function to unlock a page if all the extents in the tree
1589  * for that page are unlocked
1590  */
1591 static int check_page_locked(struct extent_io_tree *tree,
1592                              struct page *page)
1593 {
1594         u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
1595         u64 end = start + PAGE_CACHE_SIZE - 1;
1596         if (!test_range_bit(tree, start, end, EXTENT_LOCKED, 0))
1597                 unlock_page(page);
1598         return 0;
1599 }
1600
1601 /*
1602  * helper function to end page writeback if all the extents
1603  * in the tree for that page are done with writeback
1604  */
1605 static int check_page_writeback(struct extent_io_tree *tree,
1606                              struct page *page)
1607 {
1608         u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
1609         u64 end = start + PAGE_CACHE_SIZE - 1;
1610         if (!test_range_bit(tree, start, end, EXTENT_WRITEBACK, 0))
1611                 end_page_writeback(page);
1612         return 0;
1613 }
1614
1615 /* lots and lots of room for performance fixes in the end_bio funcs */
1616
1617 /*
1618  * after a writepage IO is done, we need to:
1619  * clear the uptodate bits on error
1620  * clear the writeback bits in the extent tree for this IO
1621  * end_page_writeback if the page has no more pending IO
1622  *
1623  * Scheduling is not allowed, so the extent state tree is expected
1624  * to have one and only one object corresponding to this IO.
1625  */
1626 static void end_bio_extent_writepage(struct bio *bio, int err)
1627 {
1628         int uptodate = err == 0;
1629         struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
1630         struct extent_io_tree *tree;
1631         u64 start;
1632         u64 end;
1633         int whole_page;
1634         int ret;
1635
1636         do {
1637                 struct page *page = bvec->bv_page;
1638                 tree = &BTRFS_I(page->mapping->host)->io_tree;
1639
1640                 start = ((u64)page->index << PAGE_CACHE_SHIFT) +
1641                          bvec->bv_offset;
1642                 end = start + bvec->bv_len - 1;
1643
1644                 if (bvec->bv_offset == 0 && bvec->bv_len == PAGE_CACHE_SIZE)
1645                         whole_page = 1;
1646                 else
1647                         whole_page = 0;
1648
1649                 if (--bvec >= bio->bi_io_vec)
1650                         prefetchw(&bvec->bv_page->flags);
1651                 if (tree->ops && tree->ops->writepage_end_io_hook) {
1652                         ret = tree->ops->writepage_end_io_hook(page, start,
1653                                                        end, NULL, uptodate);
1654                         if (ret)
1655                                 uptodate = 0;
1656                 }
1657
1658                 if (!uptodate && tree->ops &&
1659                     tree->ops->writepage_io_failed_hook) {
1660                         ret = tree->ops->writepage_io_failed_hook(bio, page,
1661                                                          start, end, NULL);
1662                         if (ret == 0) {
1663                                 uptodate = (err == 0);
1664                                 continue;
1665                         }
1666                 }
1667
1668                 if (!uptodate) {
1669                         clear_extent_uptodate(tree, start, end, GFP_ATOMIC);
1670                         ClearPageUptodate(page);
1671                         SetPageError(page);
1672                 }
1673
1674                 clear_extent_writeback(tree, start, end, GFP_ATOMIC);
1675
1676                 if (whole_page)
1677                         end_page_writeback(page);
1678                 else
1679                         check_page_writeback(tree, page);
1680         } while (bvec >= bio->bi_io_vec);
1681
1682         bio_put(bio);
1683 }
1684
1685 /*
1686  * after a readpage IO is done, we need to:
1687  * clear the uptodate bits on error
1688  * set the uptodate bits if things worked
1689  * set the page up to date if all extents in the tree are uptodate
1690  * clear the lock bit in the extent tree
1691  * unlock the page if there are no other extents locked for it
1692  *
1693  * Scheduling is not allowed, so the extent state tree is expected
1694  * to have one and only one object corresponding to this IO.
1695  */
1696 static void end_bio_extent_readpage(struct bio *bio, int err)
1697 {
1698         int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
1699         struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
1700         struct extent_io_tree *tree;
1701         u64 start;
1702         u64 end;
1703         int whole_page;
1704         int ret;
1705
1706         if (err)
1707                 uptodate = 0;
1708
1709         do {
1710                 struct page *page = bvec->bv_page;
1711                 tree = &BTRFS_I(page->mapping->host)->io_tree;
1712
1713                 start = ((u64)page->index << PAGE_CACHE_SHIFT) +
1714                         bvec->bv_offset;
1715                 end = start + bvec->bv_len - 1;
1716
1717                 if (bvec->bv_offset == 0 && bvec->bv_len == PAGE_CACHE_SIZE)
1718                         whole_page = 1;
1719                 else
1720                         whole_page = 0;
1721
1722                 if (--bvec >= bio->bi_io_vec)
1723                         prefetchw(&bvec->bv_page->flags);
1724
1725                 if (uptodate && tree->ops && tree->ops->readpage_end_io_hook) {
1726                         ret = tree->ops->readpage_end_io_hook(page, start, end,
1727                                                               NULL);
1728                         if (ret)
1729                                 uptodate = 0;
1730                 }
1731                 if (!uptodate && tree->ops &&
1732                     tree->ops->readpage_io_failed_hook) {
1733                         ret = tree->ops->readpage_io_failed_hook(bio, page,
1734                                                          start, end, NULL);
1735                         if (ret == 0) {
1736                                 uptodate =
1737                                         test_bit(BIO_UPTODATE, &bio->bi_flags);
1738                                 if (err)
1739                                         uptodate = 0;
1740                                 continue;
1741                         }
1742                 }
1743
1744                 if (uptodate) {
1745                         set_extent_uptodate(tree, start, end,
1746                                             GFP_ATOMIC);
1747                 }
1748                 unlock_extent(tree, start, end, GFP_ATOMIC);
1749
1750                 if (whole_page) {
1751                         if (uptodate) {
1752                                 SetPageUptodate(page);
1753                         } else {
1754                                 ClearPageUptodate(page);
1755                                 SetPageError(page);
1756                         }
1757                         unlock_page(page);
1758                 } else {
1759                         if (uptodate) {
1760                                 check_page_uptodate(tree, page);
1761                         } else {
1762                                 ClearPageUptodate(page);
1763                                 SetPageError(page);
1764                         }
1765                         check_page_locked(tree, page);
1766                 }
1767         } while (bvec >= bio->bi_io_vec);
1768
1769         bio_put(bio);
1770 }
1771
1772 /*
1773  * IO done from prepare_write is pretty simple, we just unlock
1774  * the structs in the extent tree when done, and set the uptodate bits
1775  * as appropriate.
1776  */
1777 static void end_bio_extent_preparewrite(struct bio *bio, int err)
1778 {
1779         const int uptodate = test_bit(BIO_UPTODATE, &bio->bi_flags);
1780         struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
1781         struct extent_io_tree *tree;
1782         u64 start;
1783         u64 end;
1784
1785         do {
1786                 struct page *page = bvec->bv_page;
1787                 tree = &BTRFS_I(page->mapping->host)->io_tree;
1788
1789                 start = ((u64)page->index << PAGE_CACHE_SHIFT) +
1790                         bvec->bv_offset;
1791                 end = start + bvec->bv_len - 1;
1792
1793                 if (--bvec >= bio->bi_io_vec)
1794                         prefetchw(&bvec->bv_page->flags);
1795
1796                 if (uptodate) {
1797                         set_extent_uptodate(tree, start, end, GFP_ATOMIC);
1798                 } else {
1799                         ClearPageUptodate(page);
1800                         SetPageError(page);
1801                 }
1802
1803                 unlock_extent(tree, start, end, GFP_ATOMIC);
1804
1805         } while (bvec >= bio->bi_io_vec);
1806
1807         bio_put(bio);
1808 }
1809
1810 static struct bio *
1811 extent_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs,
1812                  gfp_t gfp_flags)
1813 {
1814         struct bio *bio;
1815
1816         bio = bio_alloc(gfp_flags, nr_vecs);
1817
1818         if (bio == NULL && (current->flags & PF_MEMALLOC)) {
1819                 while (!bio && (nr_vecs /= 2))
1820                         bio = bio_alloc(gfp_flags, nr_vecs);
1821         }
1822
1823         if (bio) {
1824                 bio->bi_size = 0;
1825                 bio->bi_bdev = bdev;
1826                 bio->bi_sector = first_sector;
1827         }
1828         return bio;
1829 }
1830
1831 static int submit_one_bio(int rw, struct bio *bio, int mirror_num,
1832                           unsigned long bio_flags)
1833 {
1834         int ret = 0;
1835         struct bio_vec *bvec = bio->bi_io_vec + bio->bi_vcnt - 1;
1836         struct page *page = bvec->bv_page;
1837         struct extent_io_tree *tree = bio->bi_private;
1838         u64 start;
1839         u64 end;
1840
1841         start = ((u64)page->index << PAGE_CACHE_SHIFT) + bvec->bv_offset;
1842         end = start + bvec->bv_len - 1;
1843
1844         bio->bi_private = NULL;
1845
1846         bio_get(bio);
1847
1848         if (tree->ops && tree->ops->submit_bio_hook)
1849                 tree->ops->submit_bio_hook(page->mapping->host, rw, bio,
1850                                            mirror_num, bio_flags);
1851         else
1852                 submit_bio(rw, bio);
1853         if (bio_flagged(bio, BIO_EOPNOTSUPP))
1854                 ret = -EOPNOTSUPP;
1855         bio_put(bio);
1856         return ret;
1857 }
1858
1859 static int submit_extent_page(int rw, struct extent_io_tree *tree,
1860                               struct page *page, sector_t sector,
1861                               size_t size, unsigned long offset,
1862                               struct block_device *bdev,
1863                               struct bio **bio_ret,
1864                               unsigned long max_pages,
1865                               bio_end_io_t end_io_func,
1866                               int mirror_num,
1867                               unsigned long prev_bio_flags,
1868                               unsigned long bio_flags)
1869 {
1870         int ret = 0;
1871         struct bio *bio;
1872         int nr;
1873         int contig = 0;
1874         int this_compressed = bio_flags & EXTENT_BIO_COMPRESSED;
1875         int old_compressed = prev_bio_flags & EXTENT_BIO_COMPRESSED;
1876         size_t page_size = min_t(size_t, size, PAGE_CACHE_SIZE);
1877
1878         if (bio_ret && *bio_ret) {
1879                 bio = *bio_ret;
1880                 if (old_compressed)
1881                         contig = bio->bi_sector == sector;
1882                 else
1883                         contig = bio->bi_sector + (bio->bi_size >> 9) ==
1884                                 sector;
1885
1886                 if (prev_bio_flags != bio_flags || !contig ||
1887                     (tree->ops && tree->ops->merge_bio_hook &&
1888                      tree->ops->merge_bio_hook(page, offset, page_size, bio,
1889                                                bio_flags)) ||
1890                     bio_add_page(bio, page, page_size, offset) < page_size) {
1891                         ret = submit_one_bio(rw, bio, mirror_num,
1892                                              prev_bio_flags);
1893                         bio = NULL;
1894                 } else {
1895                         return 0;
1896                 }
1897         }
1898         if (this_compressed)
1899                 nr = BIO_MAX_PAGES;
1900         else
1901                 nr = bio_get_nr_vecs(bdev);
1902
1903         bio = extent_bio_alloc(bdev, sector, nr, GFP_NOFS | __GFP_HIGH);
1904
1905         bio_add_page(bio, page, page_size, offset);
1906         bio->bi_end_io = end_io_func;
1907         bio->bi_private = tree;
1908
1909         if (bio_ret)
1910                 *bio_ret = bio;
1911         else
1912                 ret = submit_one_bio(rw, bio, mirror_num, bio_flags);
1913
1914         return ret;
1915 }
1916
1917 void set_page_extent_mapped(struct page *page)
1918 {
1919         if (!PagePrivate(page)) {
1920                 SetPagePrivate(page);
1921                 page_cache_get(page);
1922                 set_page_private(page, EXTENT_PAGE_PRIVATE);
1923         }
1924 }
1925
1926 static void set_page_extent_head(struct page *page, unsigned long len)
1927 {
1928         set_page_private(page, EXTENT_PAGE_PRIVATE_FIRST_PAGE | len << 2);
1929 }
1930
1931 /*
1932  * basic readpage implementation.  Locked extent state structs are inserted
1933  * into the tree that are removed when the IO is done (by the end_io
1934  * handlers)
1935  */
1936 static int __extent_read_full_page(struct extent_io_tree *tree,
1937                                    struct page *page,
1938                                    get_extent_t *get_extent,
1939                                    struct bio **bio, int mirror_num,
1940                                    unsigned long *bio_flags)
1941 {
1942         struct inode *inode = page->mapping->host;
1943         u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
1944         u64 page_end = start + PAGE_CACHE_SIZE - 1;
1945         u64 end;
1946         u64 cur = start;
1947         u64 extent_offset;
1948         u64 last_byte = i_size_read(inode);
1949         u64 block_start;
1950         u64 cur_end;
1951         sector_t sector;
1952         struct extent_map *em;
1953         struct block_device *bdev;
1954         int ret;
1955         int nr = 0;
1956         size_t page_offset = 0;
1957         size_t iosize;
1958         size_t disk_io_size;
1959         size_t blocksize = inode->i_sb->s_blocksize;
1960         unsigned long this_bio_flag = 0;
1961
1962         set_page_extent_mapped(page);
1963
1964         end = page_end;
1965         lock_extent(tree, start, end, GFP_NOFS);
1966
1967         if (page->index == last_byte >> PAGE_CACHE_SHIFT) {
1968                 char *userpage;
1969                 size_t zero_offset = last_byte & (PAGE_CACHE_SIZE - 1);
1970
1971                 if (zero_offset) {
1972                         iosize = PAGE_CACHE_SIZE - zero_offset;
1973                         userpage = kmap_atomic(page, KM_USER0);
1974                         memset(userpage + zero_offset, 0, iosize);
1975                         flush_dcache_page(page);
1976                         kunmap_atomic(userpage, KM_USER0);
1977                 }
1978         }
1979         while (cur <= end) {
1980                 if (cur >= last_byte) {
1981                         char *userpage;
1982                         iosize = PAGE_CACHE_SIZE - page_offset;
1983                         userpage = kmap_atomic(page, KM_USER0);
1984                         memset(userpage + page_offset, 0, iosize);
1985                         flush_dcache_page(page);
1986                         kunmap_atomic(userpage, KM_USER0);
1987                         set_extent_uptodate(tree, cur, cur + iosize - 1,
1988                                             GFP_NOFS);
1989                         unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
1990                         break;
1991                 }
1992                 em = get_extent(inode, page, page_offset, cur,
1993                                 end - cur + 1, 0);
1994                 if (IS_ERR(em) || !em) {
1995                         SetPageError(page);
1996                         unlock_extent(tree, cur, end, GFP_NOFS);
1997                         break;
1998                 }
1999                 extent_offset = cur - em->start;
2000                 BUG_ON(extent_map_end(em) <= cur);
2001                 BUG_ON(end < cur);
2002
2003                 if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags))
2004                         this_bio_flag = EXTENT_BIO_COMPRESSED;
2005
2006                 iosize = min(extent_map_end(em) - cur, end - cur + 1);
2007                 cur_end = min(extent_map_end(em) - 1, end);
2008                 iosize = (iosize + blocksize - 1) & ~((u64)blocksize - 1);
2009                 if (this_bio_flag & EXTENT_BIO_COMPRESSED) {
2010                         disk_io_size = em->block_len;
2011                         sector = em->block_start >> 9;
2012                 } else {
2013                         sector = (em->block_start + extent_offset) >> 9;
2014                         disk_io_size = iosize;
2015                 }
2016                 bdev = em->bdev;
2017                 block_start = em->block_start;
2018                 if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags))
2019                         block_start = EXTENT_MAP_HOLE;
2020                 free_extent_map(em);
2021                 em = NULL;
2022
2023                 /* we've found a hole, just zero and go on */
2024                 if (block_start == EXTENT_MAP_HOLE) {
2025                         char *userpage;
2026                         userpage = kmap_atomic(page, KM_USER0);
2027                         memset(userpage + page_offset, 0, iosize);
2028                         flush_dcache_page(page);
2029                         kunmap_atomic(userpage, KM_USER0);
2030
2031                         set_extent_uptodate(tree, cur, cur + iosize - 1,
2032                                             GFP_NOFS);
2033                         unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
2034                         cur = cur + iosize;
2035                         page_offset += iosize;
2036                         continue;
2037                 }
2038                 /* the get_extent function already copied into the page */
2039                 if (test_range_bit(tree, cur, cur_end, EXTENT_UPTODATE, 1)) {
2040                         check_page_uptodate(tree, page);
2041                         unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
2042                         cur = cur + iosize;
2043                         page_offset += iosize;
2044                         continue;
2045                 }
2046                 /* we have an inline extent but it didn't get marked up
2047                  * to date.  Error out
2048                  */
2049                 if (block_start == EXTENT_MAP_INLINE) {
2050                         SetPageError(page);
2051                         unlock_extent(tree, cur, cur + iosize - 1, GFP_NOFS);
2052                         cur = cur + iosize;
2053                         page_offset += iosize;
2054                         continue;
2055                 }
2056
2057                 ret = 0;
2058                 if (tree->ops && tree->ops->readpage_io_hook) {
2059                         ret = tree->ops->readpage_io_hook(page, cur,
2060                                                           cur + iosize - 1);
2061                 }
2062                 if (!ret) {
2063                         unsigned long pnr = (last_byte >> PAGE_CACHE_SHIFT) + 1;
2064                         pnr -= page->index;
2065                         ret = submit_extent_page(READ, tree, page,
2066                                          sector, disk_io_size, page_offset,
2067                                          bdev, bio, pnr,
2068                                          end_bio_extent_readpage, mirror_num,
2069                                          *bio_flags,
2070                                          this_bio_flag);
2071                         nr++;
2072                         *bio_flags = this_bio_flag;
2073                 }
2074                 if (ret)
2075                         SetPageError(page);
2076                 cur = cur + iosize;
2077                 page_offset += iosize;
2078         }
2079         if (!nr) {
2080                 if (!PageError(page))
2081                         SetPageUptodate(page);
2082                 unlock_page(page);
2083         }
2084         return 0;
2085 }
2086
2087 int extent_read_full_page(struct extent_io_tree *tree, struct page *page,
2088                             get_extent_t *get_extent)
2089 {
2090         struct bio *bio = NULL;
2091         unsigned long bio_flags = 0;
2092         int ret;
2093
2094         ret = __extent_read_full_page(tree, page, get_extent, &bio, 0,
2095                                       &bio_flags);
2096         if (bio)
2097                 submit_one_bio(READ, bio, 0, bio_flags);
2098         return ret;
2099 }
2100
2101 static noinline void update_nr_written(struct page *page,
2102                                       struct writeback_control *wbc,
2103                                       unsigned long nr_written)
2104 {
2105         wbc->nr_to_write -= nr_written;
2106         if (wbc->range_cyclic || (wbc->nr_to_write > 0 &&
2107             wbc->range_start == 0 && wbc->range_end == LLONG_MAX))
2108                 page->mapping->writeback_index = page->index + nr_written;
2109 }
2110
2111 /*
2112  * the writepage semantics are similar to regular writepage.  extent
2113  * records are inserted to lock ranges in the tree, and as dirty areas
2114  * are found, they are marked writeback.  Then the lock bits are removed
2115  * and the end_io handler clears the writeback ranges
2116  */
2117 static int __extent_writepage(struct page *page, struct writeback_control *wbc,
2118                               void *data)
2119 {
2120         struct inode *inode = page->mapping->host;
2121         struct extent_page_data *epd = data;
2122         struct extent_io_tree *tree = epd->tree;
2123         u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
2124         u64 delalloc_start;
2125         u64 page_end = start + PAGE_CACHE_SIZE - 1;
2126         u64 end;
2127         u64 cur = start;
2128         u64 extent_offset;
2129         u64 last_byte = i_size_read(inode);
2130         u64 block_start;
2131         u64 iosize;
2132         u64 unlock_start;
2133         sector_t sector;
2134         struct extent_map *em;
2135         struct block_device *bdev;
2136         int ret;
2137         int nr = 0;
2138         size_t pg_offset = 0;
2139         size_t blocksize;
2140         loff_t i_size = i_size_read(inode);
2141         unsigned long end_index = i_size >> PAGE_CACHE_SHIFT;
2142         u64 nr_delalloc;
2143         u64 delalloc_end;
2144         int page_started;
2145         int compressed;
2146         int write_flags;
2147         unsigned long nr_written = 0;
2148
2149         if (wbc->sync_mode == WB_SYNC_ALL)
2150                 write_flags = WRITE_SYNC_PLUG;
2151         else
2152                 write_flags = WRITE;
2153
2154         WARN_ON(!PageLocked(page));
2155         pg_offset = i_size & (PAGE_CACHE_SIZE - 1);
2156         if (page->index > end_index ||
2157            (page->index == end_index && !pg_offset)) {
2158                 page->mapping->a_ops->invalidatepage(page, 0);
2159                 unlock_page(page);
2160                 return 0;
2161         }
2162
2163         if (page->index == end_index) {
2164                 char *userpage;
2165
2166                 userpage = kmap_atomic(page, KM_USER0);
2167                 memset(userpage + pg_offset, 0,
2168                        PAGE_CACHE_SIZE - pg_offset);
2169                 kunmap_atomic(userpage, KM_USER0);
2170                 flush_dcache_page(page);
2171         }
2172         pg_offset = 0;
2173
2174         set_page_extent_mapped(page);
2175
2176         delalloc_start = start;
2177         delalloc_end = 0;
2178         page_started = 0;
2179         if (!epd->extent_locked) {
2180                 /*
2181                  * make sure the wbc mapping index is at least updated
2182                  * to this page.
2183                  */
2184                 update_nr_written(page, wbc, 0);
2185
2186                 while (delalloc_end < page_end) {
2187                         nr_delalloc = find_lock_delalloc_range(inode, tree,
2188                                                        page,
2189                                                        &delalloc_start,
2190                                                        &delalloc_end,
2191                                                        128 * 1024 * 1024);
2192                         if (nr_delalloc == 0) {
2193                                 delalloc_start = delalloc_end + 1;
2194                                 continue;
2195                         }
2196                         tree->ops->fill_delalloc(inode, page, delalloc_start,
2197                                                  delalloc_end, &page_started,
2198                                                  &nr_written);
2199                         delalloc_start = delalloc_end + 1;
2200                 }
2201
2202                 /* did the fill delalloc function already unlock and start
2203                  * the IO?
2204                  */
2205                 if (page_started) {
2206                         ret = 0;
2207                         /*
2208                          * we've unlocked the page, so we can't update
2209                          * the mapping's writeback index, just update
2210                          * nr_to_write.
2211                          */
2212                         wbc->nr_to_write -= nr_written;
2213                         goto done_unlocked;
2214                 }
2215         }
2216         lock_extent(tree, start, page_end, GFP_NOFS);
2217
2218         unlock_start = start;
2219
2220         if (tree->ops && tree->ops->writepage_start_hook) {
2221                 ret = tree->ops->writepage_start_hook(page, start,
2222                                                       page_end);
2223                 if (ret == -EAGAIN) {
2224                         unlock_extent(tree, start, page_end, GFP_NOFS);
2225                         redirty_page_for_writepage(wbc, page);
2226                         update_nr_written(page, wbc, nr_written);
2227                         unlock_page(page);
2228                         ret = 0;
2229                         goto done_unlocked;
2230                 }
2231         }
2232
2233         /*
2234          * we don't want to touch the inode after unlocking the page,
2235          * so we update the mapping writeback index now
2236          */
2237         update_nr_written(page, wbc, nr_written + 1);
2238
2239         end = page_end;
2240         if (test_range_bit(tree, start, page_end, EXTENT_DELALLOC, 0))
2241                 printk(KERN_ERR "btrfs delalloc bits after lock_extent\n");
2242
2243         if (last_byte <= start) {
2244                 clear_extent_dirty(tree, start, page_end, GFP_NOFS);
2245                 unlock_extent(tree, start, page_end, GFP_NOFS);
2246                 if (tree->ops && tree->ops->writepage_end_io_hook)
2247                         tree->ops->writepage_end_io_hook(page, start,
2248                                                          page_end, NULL, 1);
2249                 unlock_start = page_end + 1;
2250                 goto done;
2251         }
2252
2253         set_extent_uptodate(tree, start, page_end, GFP_NOFS);
2254         blocksize = inode->i_sb->s_blocksize;
2255
2256         while (cur <= end) {
2257                 if (cur >= last_byte) {
2258                         clear_extent_dirty(tree, cur, page_end, GFP_NOFS);
2259                         unlock_extent(tree, unlock_start, page_end, GFP_NOFS);
2260                         if (tree->ops && tree->ops->writepage_end_io_hook)
2261                                 tree->ops->writepage_end_io_hook(page, cur,
2262                                                          page_end, NULL, 1);
2263                         unlock_start = page_end + 1;
2264                         break;
2265                 }
2266                 em = epd->get_extent(inode, page, pg_offset, cur,
2267                                      end - cur + 1, 1);
2268                 if (IS_ERR(em) || !em) {
2269                         SetPageError(page);
2270                         break;
2271                 }
2272
2273                 extent_offset = cur - em->start;
2274                 BUG_ON(extent_map_end(em) <= cur);
2275                 BUG_ON(end < cur);
2276                 iosize = min(extent_map_end(em) - cur, end - cur + 1);
2277                 iosize = (iosize + blocksize - 1) & ~((u64)blocksize - 1);
2278                 sector = (em->block_start + extent_offset) >> 9;
2279                 bdev = em->bdev;
2280                 block_start = em->block_start;
2281                 compressed = test_bit(EXTENT_FLAG_COMPRESSED, &em->flags);
2282                 free_extent_map(em);
2283                 em = NULL;
2284
2285                 /*
2286                  * compressed and inline extents are written through other
2287                  * paths in the FS
2288                  */
2289                 if (compressed || block_start == EXTENT_MAP_HOLE ||
2290                     block_start == EXTENT_MAP_INLINE) {
2291                         clear_extent_dirty(tree, cur,
2292                                            cur + iosize - 1, GFP_NOFS);
2293
2294                         unlock_extent(tree, unlock_start, cur + iosize - 1,
2295                                       GFP_NOFS);
2296
2297                         /*
2298                          * end_io notification does not happen here for
2299                          * compressed extents
2300                          */
2301                         if (!compressed && tree->ops &&
2302                             tree->ops->writepage_end_io_hook)
2303                                 tree->ops->writepage_end_io_hook(page, cur,
2304                                                          cur + iosize - 1,
2305                                                          NULL, 1);
2306                         else if (compressed) {
2307                                 /* we don't want to end_page_writeback on
2308                                  * a compressed extent.  this happens
2309                                  * elsewhere
2310                                  */
2311                                 nr++;
2312                         }
2313
2314                         cur += iosize;
2315                         pg_offset += iosize;
2316                         unlock_start = cur;
2317                         continue;
2318                 }
2319                 /* leave this out until we have a page_mkwrite call */
2320                 if (0 && !test_range_bit(tree, cur, cur + iosize - 1,
2321                                    EXTENT_DIRTY, 0)) {
2322                         cur = cur + iosize;
2323                         pg_offset += iosize;
2324                         continue;
2325                 }
2326
2327                 clear_extent_dirty(tree, cur, cur + iosize - 1, GFP_NOFS);
2328                 if (tree->ops && tree->ops->writepage_io_hook) {
2329                         ret = tree->ops->writepage_io_hook(page, cur,
2330                                                 cur + iosize - 1);
2331                 } else {
2332                         ret = 0;
2333                 }
2334                 if (ret) {
2335                         SetPageError(page);
2336                 } else {
2337                         unsigned long max_nr = end_index + 1;
2338
2339                         set_range_writeback(tree, cur, cur + iosize - 1);
2340                         if (!PageWriteback(page)) {
2341                                 printk(KERN_ERR "btrfs warning page %lu not "
2342                                        "writeback, cur %llu end %llu\n",
2343                                        page->index, (unsigned long long)cur,
2344                                        (unsigned long long)end);
2345                         }
2346
2347                         ret = submit_extent_page(write_flags, tree, page,
2348                                                  sector, iosize, pg_offset,
2349                                                  bdev, &epd->bio, max_nr,
2350                                                  end_bio_extent_writepage,
2351                                                  0, 0, 0);
2352                         if (ret)
2353                                 SetPageError(page);
2354                 }
2355                 cur = cur + iosize;
2356                 pg_offset += iosize;
2357                 nr++;
2358         }
2359 done:
2360         if (nr == 0) {
2361                 /* make sure the mapping tag for page dirty gets cleared */
2362                 set_page_writeback(page);
2363                 end_page_writeback(page);
2364         }
2365         if (unlock_start <= page_end)
2366                 unlock_extent(tree, unlock_start, page_end, GFP_NOFS);
2367         unlock_page(page);
2368
2369 done_unlocked:
2370
2371         return 0;
2372 }
2373
2374 /**
2375  * write_cache_pages - walk the list of dirty pages of the given address space and write all of them.
2376  * @mapping: address space structure to write
2377  * @wbc: subtract the number of written pages from *@wbc->nr_to_write
2378  * @writepage: function called for each page
2379  * @data: data passed to writepage function
2380  *
2381  * If a page is already under I/O, write_cache_pages() skips it, even
2382  * if it's dirty.  This is desirable behaviour for memory-cleaning writeback,
2383  * but it is INCORRECT for data-integrity system calls such as fsync().  fsync()
2384  * and msync() need to guarantee that all the data which was dirty at the time
2385  * the call was made get new I/O started against them.  If wbc->sync_mode is
2386  * WB_SYNC_ALL then we were called for data integrity and we must wait for
2387  * existing IO to complete.
2388  */
2389 static int extent_write_cache_pages(struct extent_io_tree *tree,
2390                              struct address_space *mapping,
2391                              struct writeback_control *wbc,
2392                              writepage_t writepage, void *data,
2393                              void (*flush_fn)(void *))
2394 {
2395         struct backing_dev_info *bdi = mapping->backing_dev_info;
2396         int ret = 0;
2397         int done = 0;
2398         struct pagevec pvec;
2399         int nr_pages;
2400         pgoff_t index;
2401         pgoff_t end;            /* Inclusive */
2402         int scanned = 0;
2403         int range_whole = 0;
2404
2405         pagevec_init(&pvec, 0);
2406         if (wbc->range_cyclic) {
2407                 index = mapping->writeback_index; /* Start from prev offset */
2408                 end = -1;
2409         } else {
2410                 index = wbc->range_start >> PAGE_CACHE_SHIFT;
2411                 end = wbc->range_end >> PAGE_CACHE_SHIFT;
2412                 if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
2413                         range_whole = 1;
2414                 scanned = 1;
2415         }
2416 retry:
2417         while (!done && (index <= end) &&
2418                (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
2419                               PAGECACHE_TAG_DIRTY, min(end - index,
2420                                   (pgoff_t)PAGEVEC_SIZE-1) + 1))) {
2421                 unsigned i;
2422
2423                 scanned = 1;
2424                 for (i = 0; i < nr_pages; i++) {
2425                         struct page *page = pvec.pages[i];
2426
2427                         /*
2428                          * At this point we hold neither mapping->tree_lock nor
2429                          * lock on the page itself: the page may be truncated or
2430                          * invalidated (changing page->mapping to NULL), or even
2431                          * swizzled back from swapper_space to tmpfs file
2432                          * mapping
2433                          */
2434                         if (tree->ops && tree->ops->write_cache_pages_lock_hook)
2435                                 tree->ops->write_cache_pages_lock_hook(page);
2436                         else
2437                                 lock_page(page);
2438
2439                         if (unlikely(page->mapping != mapping)) {
2440                                 unlock_page(page);
2441                                 continue;
2442                         }
2443
2444                         if (!wbc->range_cyclic && page->index > end) {
2445                                 done = 1;
2446                                 unlock_page(page);
2447                                 continue;
2448                         }
2449
2450                         if (wbc->sync_mode != WB_SYNC_NONE) {
2451                                 if (PageWriteback(page))
2452                                         flush_fn(data);
2453                                 wait_on_page_writeback(page);
2454                         }
2455
2456                         if (PageWriteback(page) ||
2457                             !clear_page_dirty_for_io(page)) {
2458                                 unlock_page(page);
2459                                 continue;
2460                         }
2461
2462                         ret = (*writepage)(page, wbc, data);
2463
2464                         if (unlikely(ret == AOP_WRITEPAGE_ACTIVATE)) {
2465                                 unlock_page(page);
2466                                 ret = 0;
2467                         }
2468                         if (ret || wbc->nr_to_write <= 0)
2469                                 done = 1;
2470                         if (wbc->nonblocking && bdi_write_congested(bdi)) {
2471                                 wbc->encountered_congestion = 1;
2472                                 done = 1;
2473                         }
2474                 }
2475                 pagevec_release(&pvec);
2476                 cond_resched();
2477         }
2478         if (!scanned && !done) {
2479                 /*
2480                  * We hit the last page and there is more work to be done: wrap
2481                  * back to the start of the file
2482                  */
2483                 scanned = 1;
2484                 index = 0;
2485                 goto retry;
2486         }
2487         return ret;
2488 }
2489
2490 static void flush_epd_write_bio(struct extent_page_data *epd)
2491 {
2492         if (epd->bio) {
2493                 if (epd->sync_io)
2494                         submit_one_bio(WRITE_SYNC, epd->bio, 0, 0);
2495                 else
2496                         submit_one_bio(WRITE, epd->bio, 0, 0);
2497                 epd->bio = NULL;
2498         }
2499 }
2500
2501 static noinline void flush_write_bio(void *data)
2502 {
2503         struct extent_page_data *epd = data;
2504         flush_epd_write_bio(epd);
2505 }
2506
2507 int extent_write_full_page(struct extent_io_tree *tree, struct page *page,
2508                           get_extent_t *get_extent,
2509                           struct writeback_control *wbc)
2510 {
2511         int ret;
2512         struct address_space *mapping = page->mapping;
2513         struct extent_page_data epd = {
2514                 .bio = NULL,
2515                 .tree = tree,
2516                 .get_extent = get_extent,
2517                 .extent_locked = 0,
2518                 .sync_io = wbc->sync_mode == WB_SYNC_ALL,
2519         };
2520         struct writeback_control wbc_writepages = {
2521                 .bdi            = wbc->bdi,
2522                 .sync_mode      = wbc->sync_mode,
2523                 .older_than_this = NULL,
2524                 .nr_to_write    = 64,
2525                 .range_start    = page_offset(page) + PAGE_CACHE_SIZE,
2526                 .range_end      = (loff_t)-1,
2527         };
2528
2529         ret = __extent_writepage(page, wbc, &epd);
2530
2531         extent_write_cache_pages(tree, mapping, &wbc_writepages,
2532                                  __extent_writepage, &epd, flush_write_bio);
2533         flush_epd_write_bio(&epd);
2534         return ret;
2535 }
2536
2537 int extent_write_locked_range(struct extent_io_tree *tree, struct inode *inode,
2538                               u64 start, u64 end, get_extent_t *get_extent,
2539                               int mode)
2540 {
2541         int ret = 0;
2542         struct address_space *mapping = inode->i_mapping;
2543         struct page *page;
2544         unsigned long nr_pages = (end - start + PAGE_CACHE_SIZE) >>
2545                 PAGE_CACHE_SHIFT;
2546
2547         struct extent_page_data epd = {
2548                 .bio = NULL,
2549                 .tree = tree,
2550                 .get_extent = get_extent,
2551                 .extent_locked = 1,
2552                 .sync_io = mode == WB_SYNC_ALL,
2553         };
2554         struct writeback_control wbc_writepages = {
2555                 .bdi            = inode->i_mapping->backing_dev_info,
2556                 .sync_mode      = mode,
2557                 .older_than_this = NULL,
2558                 .nr_to_write    = nr_pages * 2,
2559                 .range_start    = start,
2560                 .range_end      = end + 1,
2561         };
2562
2563         while (start <= end) {
2564                 page = find_get_page(mapping, start >> PAGE_CACHE_SHIFT);
2565                 if (clear_page_dirty_for_io(page))
2566                         ret = __extent_writepage(page, &wbc_writepages, &epd);
2567                 else {
2568                         if (tree->ops && tree->ops->writepage_end_io_hook)
2569                                 tree->ops->writepage_end_io_hook(page, start,
2570                                                  start + PAGE_CACHE_SIZE - 1,
2571                                                  NULL, 1);
2572                         unlock_page(page);
2573                 }
2574                 page_cache_release(page);
2575                 start += PAGE_CACHE_SIZE;
2576         }
2577
2578         flush_epd_write_bio(&epd);
2579         return ret;
2580 }
2581
2582 int extent_writepages(struct extent_io_tree *tree,
2583                       struct address_space *mapping,
2584                       get_extent_t *get_extent,
2585                       struct writeback_control *wbc)
2586 {
2587         int ret = 0;
2588         struct extent_page_data epd = {
2589                 .bio = NULL,
2590                 .tree = tree,
2591                 .get_extent = get_extent,
2592                 .extent_locked = 0,
2593                 .sync_io = wbc->sync_mode == WB_SYNC_ALL,
2594         };
2595
2596         ret = extent_write_cache_pages(tree, mapping, wbc,
2597                                        __extent_writepage, &epd,
2598                                        flush_write_bio);
2599         flush_epd_write_bio(&epd);
2600         return ret;
2601 }
2602
2603 int extent_readpages(struct extent_io_tree *tree,
2604                      struct address_space *mapping,
2605                      struct list_head *pages, unsigned nr_pages,
2606                      get_extent_t get_extent)
2607 {
2608         struct bio *bio = NULL;
2609         unsigned page_idx;
2610         struct pagevec pvec;
2611         unsigned long bio_flags = 0;
2612
2613         pagevec_init(&pvec, 0);
2614         for (page_idx = 0; page_idx < nr_pages; page_idx++) {
2615                 struct page *page = list_entry(pages->prev, struct page, lru);
2616
2617                 prefetchw(&page->flags);
2618                 list_del(&page->lru);
2619                 /*
2620                  * what we want to do here is call add_to_page_cache_lru,
2621                  * but that isn't exported, so we reproduce it here
2622                  */
2623                 if (!add_to_page_cache(page, mapping,
2624                                         page->index, GFP_KERNEL)) {
2625
2626                         /* open coding of lru_cache_add, also not exported */
2627                         page_cache_get(page);
2628                         if (!pagevec_add(&pvec, page))
2629                                 __pagevec_lru_add_file(&pvec);
2630                         __extent_read_full_page(tree, page, get_extent,
2631                                                 &bio, 0, &bio_flags);
2632                 }
2633                 page_cache_release(page);
2634         }
2635         if (pagevec_count(&pvec))
2636                 __pagevec_lru_add_file(&pvec);
2637         BUG_ON(!list_empty(pages));
2638         if (bio)
2639                 submit_one_bio(READ, bio, 0, bio_flags);
2640         return 0;
2641 }
2642
2643 /*
2644  * basic invalidatepage code, this waits on any locked or writeback
2645  * ranges corresponding to the page, and then deletes any extent state
2646  * records from the tree
2647  */
2648 int extent_invalidatepage(struct extent_io_tree *tree,
2649                           struct page *page, unsigned long offset)
2650 {
2651         u64 start = ((u64)page->index << PAGE_CACHE_SHIFT);
2652         u64 end = start + PAGE_CACHE_SIZE - 1;
2653         size_t blocksize = page->mapping->host->i_sb->s_blocksize;
2654
2655         start += (offset + blocksize - 1) & ~(blocksize - 1);
2656         if (start > end)
2657                 return 0;
2658
2659         lock_extent(tree, start, end, GFP_NOFS);
2660         wait_on_extent_writeback(tree, start, end);
2661         clear_extent_bit(tree, start, end,
2662                          EXTENT_LOCKED | EXTENT_DIRTY | EXTENT_DELALLOC,
2663                          1, 1, GFP_NOFS);
2664         return 0;
2665 }
2666
2667 /*
2668  * simple commit_write call, set_range_dirty is used to mark both
2669  * the pages and the extent records as dirty
2670  */
2671 int extent_commit_write(struct extent_io_tree *tree,
2672                         struct inode *inode, struct page *page,
2673                         unsigned from, unsigned to)
2674 {
2675         loff_t pos = ((loff_t)page->index << PAGE_CACHE_SHIFT) + to;
2676
2677         set_page_extent_mapped(page);
2678         set_page_dirty(page);
2679
2680         if (pos > inode->i_size) {
2681                 i_size_write(inode, pos);
2682                 mark_inode_dirty(inode);
2683         }
2684         return 0;
2685 }
2686
2687 int extent_prepare_write(struct extent_io_tree *tree,
2688                          struct inode *inode, struct page *page,
2689                          unsigned from, unsigned to, get_extent_t *get_extent)
2690 {
2691         u64 page_start = (u64)page->index << PAGE_CACHE_SHIFT;
2692         u64 page_end = page_start + PAGE_CACHE_SIZE - 1;
2693         u64 block_start;
2694         u64 orig_block_start;
2695         u64 block_end;
2696         u64 cur_end;
2697         struct extent_map *em;
2698         unsigned blocksize = 1 << inode->i_blkbits;
2699         size_t page_offset = 0;
2700         size_t block_off_start;
2701         size_t block_off_end;
2702         int err = 0;
2703         int iocount = 0;
2704         int ret = 0;
2705         int isnew;
2706
2707         set_page_extent_mapped(page);
2708
2709         block_start = (page_start + from) & ~((u64)blocksize - 1);
2710         block_end = (page_start + to - 1) | (blocksize - 1);
2711         orig_block_start = block_start;
2712
2713         lock_extent(tree, page_start, page_end, GFP_NOFS);
2714         while (block_start <= block_end) {
2715                 em = get_extent(inode, page, page_offset, block_start,
2716                                 block_end - block_start + 1, 1);
2717                 if (IS_ERR(em) || !em)
2718                         goto err;
2719
2720                 cur_end = min(block_end, extent_map_end(em) - 1);
2721                 block_off_start = block_start & (PAGE_CACHE_SIZE - 1);
2722                 block_off_end = block_off_start + blocksize;
2723                 isnew = clear_extent_new(tree, block_start, cur_end, GFP_NOFS);
2724
2725                 if (!PageUptodate(page) && isnew &&
2726                     (block_off_end > to || block_off_start < from)) {
2727                         void *kaddr;
2728
2729                         kaddr = kmap_atomic(page, KM_USER0);
2730                         if (block_off_end > to)
2731                                 memset(kaddr + to, 0, block_off_end - to);
2732                         if (block_off_start < from)
2733                                 memset(kaddr + block_off_start, 0,
2734                                        from - block_off_start);
2735                         flush_dcache_page(page);
2736                         kunmap_atomic(kaddr, KM_USER0);
2737                 }
2738                 if ((em->block_start != EXTENT_MAP_HOLE &&
2739                      em->block_start != EXTENT_MAP_INLINE) &&
2740                     !isnew && !PageUptodate(page) &&
2741                     (block_off_end > to || block_off_start < from) &&
2742                     !test_range_bit(tree, block_start, cur_end,
2743                                     EXTENT_UPTODATE, 1)) {
2744                         u64 sector;
2745                         u64 extent_offset = block_start - em->start;
2746                         size_t iosize;
2747                         sector = (em->block_start + extent_offset) >> 9;
2748                         iosize = (cur_end - block_start + blocksize) &
2749                                 ~((u64)blocksize - 1);
2750                         /*
2751                          * we've already got the extent locked, but we
2752                          * need to split the state such that our end_bio
2753                          * handler can clear the lock.
2754                          */
2755                         set_extent_bit(tree, block_start,
2756                                        block_start + iosize - 1,
2757                                        EXTENT_LOCKED, 0, NULL, GFP_NOFS);
2758                         ret = submit_extent_page(READ, tree, page,
2759                                          sector, iosize, page_offset, em->bdev,
2760                                          NULL, 1,
2761                                          end_bio_extent_preparewrite, 0,
2762                                          0, 0);
2763                         iocount++;
2764                         block_start = block_start + iosize;
2765                 } else {
2766                         set_extent_uptodate(tree, block_start, cur_end,
2767                                             GFP_NOFS);
2768                         unlock_extent(tree, block_start, cur_end, GFP_NOFS);
2769                         block_start = cur_end + 1;
2770                 }
2771                 page_offset = block_start & (PAGE_CACHE_SIZE - 1);
2772                 free_extent_map(em);
2773         }
2774         if (iocount) {
2775                 wait_extent_bit(tree, orig_block_start,
2776                                 block_end, EXTENT_LOCKED);
2777         }
2778         check_page_uptodate(tree, page);
2779 err:
2780         /* FIXME, zero out newly allocated blocks on error */
2781         return err;
2782 }
2783
2784 /*
2785  * a helper for releasepage, this tests for areas of the page that
2786  * are locked or under IO and drops the related state bits if it is safe
2787  * to drop the page.
2788  */
2789 int try_release_extent_state(struct extent_map_tree *map,
2790                              struct extent_io_tree *tree, struct page *page,
2791                              gfp_t mask)
2792 {
2793         u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
2794         u64 end = start + PAGE_CACHE_SIZE - 1;
2795         int ret = 1;
2796
2797         if (test_range_bit(tree, start, end,
2798                            EXTENT_IOBITS | EXTENT_ORDERED, 0))
2799                 ret = 0;
2800         else {
2801                 if ((mask & GFP_NOFS) == GFP_NOFS)
2802                         mask = GFP_NOFS;
2803                 clear_extent_bit(tree, start, end, EXTENT_UPTODATE,
2804                                  1, 1, mask);
2805         }
2806         return ret;
2807 }
2808
2809 /*
2810  * a helper for releasepage.  As long as there are no locked extents
2811  * in the range corresponding to the page, both state records and extent
2812  * map records are removed
2813  */
2814 int try_release_extent_mapping(struct extent_map_tree *map,
2815                                struct extent_io_tree *tree, struct page *page,
2816                                gfp_t mask)
2817 {
2818         struct extent_map *em;
2819         u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
2820         u64 end = start + PAGE_CACHE_SIZE - 1;
2821
2822         if ((mask & __GFP_WAIT) &&
2823             page->mapping->host->i_size > 16 * 1024 * 1024) {
2824                 u64 len;
2825                 while (start <= end) {
2826                         len = end - start + 1;
2827                         spin_lock(&map->lock);
2828                         em = lookup_extent_mapping(map, start, len);
2829                         if (!em || IS_ERR(em)) {
2830                                 spin_unlock(&map->lock);
2831                                 break;
2832                         }
2833                         if (test_bit(EXTENT_FLAG_PINNED, &em->flags) ||
2834                             em->start != start) {
2835                                 spin_unlock(&map->lock);
2836                                 free_extent_map(em);
2837                                 break;
2838                         }
2839                         if (!test_range_bit(tree, em->start,
2840                                             extent_map_end(em) - 1,
2841                                             EXTENT_LOCKED | EXTENT_WRITEBACK |
2842                                             EXTENT_ORDERED,
2843                                             0)) {
2844                                 remove_extent_mapping(map, em);
2845                                 /* once for the rb tree */
2846                                 free_extent_map(em);
2847                         }
2848                         start = extent_map_end(em);
2849                         spin_unlock(&map->lock);
2850
2851                         /* once for us */
2852                         free_extent_map(em);
2853                 }
2854         }
2855         return try_release_extent_state(map, tree, page, mask);
2856 }
2857
2858 sector_t extent_bmap(struct address_space *mapping, sector_t iblock,
2859                 get_extent_t *get_extent)
2860 {
2861         struct inode *inode = mapping->host;
2862         u64 start = iblock << inode->i_blkbits;
2863         sector_t sector = 0;
2864         size_t blksize = (1 << inode->i_blkbits);
2865         struct extent_map *em;
2866
2867         lock_extent(&BTRFS_I(inode)->io_tree, start, start + blksize - 1,
2868                     GFP_NOFS);
2869         em = get_extent(inode, NULL, 0, start, blksize, 0);
2870         unlock_extent(&BTRFS_I(inode)->io_tree, start, start + blksize - 1,
2871                       GFP_NOFS);
2872         if (!em || IS_ERR(em))
2873                 return 0;
2874
2875         if (em->block_start > EXTENT_MAP_LAST_BYTE)
2876                 goto out;
2877
2878         sector = (em->block_start + start - em->start) >> inode->i_blkbits;
2879 out:
2880         free_extent_map(em);
2881         return sector;
2882 }
2883
2884 int extent_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
2885                 __u64 start, __u64 len, get_extent_t *get_extent)
2886 {
2887         int ret;
2888         u64 off = start;
2889         u64 max = start + len;
2890         u32 flags = 0;
2891         u64 disko = 0;
2892         struct extent_map *em = NULL;
2893         int end = 0;
2894         u64 em_start = 0, em_len = 0;
2895         unsigned long emflags;
2896         ret = 0;
2897
2898         if (len == 0)
2899                 return -EINVAL;
2900
2901         lock_extent(&BTRFS_I(inode)->io_tree, start, start + len,
2902                 GFP_NOFS);
2903         em = get_extent(inode, NULL, 0, off, max - off, 0);
2904         if (!em)
2905                 goto out;
2906         if (IS_ERR(em)) {
2907                 ret = PTR_ERR(em);
2908                 goto out;
2909         }
2910         while (!end) {
2911                 off = em->start + em->len;
2912                 if (off >= max)
2913                         end = 1;
2914
2915                 em_start = em->start;
2916                 em_len = em->len;
2917
2918                 disko = 0;
2919                 flags = 0;
2920
2921                 if (em->block_start == EXTENT_MAP_LAST_BYTE) {
2922                         end = 1;
2923                         flags |= FIEMAP_EXTENT_LAST;
2924                 } else if (em->block_start == EXTENT_MAP_HOLE) {
2925                         flags |= FIEMAP_EXTENT_UNWRITTEN;
2926                 } else if (em->block_start == EXTENT_MAP_INLINE) {
2927                         flags |= (FIEMAP_EXTENT_DATA_INLINE |
2928                                   FIEMAP_EXTENT_NOT_ALIGNED);
2929                 } else if (em->block_start == EXTENT_MAP_DELALLOC) {
2930                         flags |= (FIEMAP_EXTENT_DELALLOC |
2931                                   FIEMAP_EXTENT_UNKNOWN);
2932                 } else {
2933                         disko = em->block_start;
2934                 }
2935                 if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags))
2936                         flags |= FIEMAP_EXTENT_ENCODED;
2937
2938                 emflags = em->flags;
2939                 free_extent_map(em);
2940                 em = NULL;
2941
2942                 if (!end) {
2943                         em = get_extent(inode, NULL, 0, off, max - off, 0);
2944                         if (!em)
2945                                 goto out;
2946                         if (IS_ERR(em)) {
2947                                 ret = PTR_ERR(em);
2948                                 goto out;
2949                         }
2950                         emflags = em->flags;
2951                 }
2952                 if (test_bit(EXTENT_FLAG_VACANCY, &emflags)) {
2953                         flags |= FIEMAP_EXTENT_LAST;
2954                         end = 1;
2955                 }
2956
2957                 ret = fiemap_fill_next_extent(fieinfo, em_start, disko,
2958                                         em_len, flags);
2959                 if (ret)
2960                         goto out_free;
2961         }
2962 out_free:
2963         free_extent_map(em);
2964 out:
2965         unlock_extent(&BTRFS_I(inode)->io_tree, start, start + len,
2966                         GFP_NOFS);
2967         return ret;
2968 }
2969
2970 static inline struct page *extent_buffer_page(struct extent_buffer *eb,
2971                                               unsigned long i)
2972 {
2973         struct page *p;
2974         struct address_space *mapping;
2975
2976         if (i == 0)
2977                 return eb->first_page;
2978         i += eb->start >> PAGE_CACHE_SHIFT;
2979         mapping = eb->first_page->mapping;
2980         if (!mapping)
2981                 return NULL;
2982
2983         /*
2984          * extent_buffer_page is only called after pinning the page
2985          * by increasing the reference count.  So we know the page must
2986          * be in the radix tree.
2987          */
2988         rcu_read_lock();
2989         p = radix_tree_lookup(&mapping->page_tree, i);
2990         rcu_read_unlock();
2991
2992         return p;
2993 }
2994
2995 static inline unsigned long num_extent_pages(u64 start, u64 len)
2996 {
2997         return ((start + len + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT) -
2998                 (start >> PAGE_CACHE_SHIFT);
2999 }
3000
3001 static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree,
3002                                                    u64 start,
3003                                                    unsigned long len,
3004                                                    gfp_t mask)
3005 {
3006         struct extent_buffer *eb = NULL;
3007 #if LEAK_DEBUG
3008         unsigned long flags;
3009 #endif
3010
3011         eb = kmem_cache_zalloc(extent_buffer_cache, mask);
3012         eb->start = start;
3013         eb->len = len;
3014         spin_lock_init(&eb->lock);
3015         init_waitqueue_head(&eb->lock_wq);
3016
3017 #if LEAK_DEBUG
3018         spin_lock_irqsave(&leak_lock, flags);
3019         list_add(&eb->leak_list, &buffers);
3020         spin_unlock_irqrestore(&leak_lock, flags);
3021 #endif
3022         atomic_set(&eb->refs, 1);
3023
3024         return eb;
3025 }
3026
3027 static void __free_extent_buffer(struct extent_buffer *eb)
3028 {
3029 #if LEAK_DEBUG
3030         unsigned long flags;
3031         spin_lock_irqsave(&leak_lock, flags);
3032         list_del(&eb->leak_list);
3033         spin_unlock_irqrestore(&leak_lock, flags);
3034 #endif
3035         kmem_cache_free(extent_buffer_cache, eb);
3036 }
3037
3038 struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
3039                                           u64 start, unsigned long len,
3040                                           struct page *page0,
3041                                           gfp_t mask)
3042 {
3043         unsigned long num_pages = num_extent_pages(start, len);
3044         unsigned long i;
3045         unsigned long index = start >> PAGE_CACHE_SHIFT;
3046         struct extent_buffer *eb;
3047         struct extent_buffer *exists = NULL;
3048         struct page *p;
3049         struct address_space *mapping = tree->mapping;
3050         int uptodate = 1;
3051
3052         spin_lock(&tree->buffer_lock);
3053         eb = buffer_search(tree, start);
3054         if (eb) {
3055                 atomic_inc(&eb->refs);
3056                 spin_unlock(&tree->buffer_lock);
3057                 mark_page_accessed(eb->first_page);
3058                 return eb;
3059         }
3060         spin_unlock(&tree->buffer_lock);
3061
3062         eb = __alloc_extent_buffer(tree, start, len, mask);
3063         if (!eb)
3064                 return NULL;
3065
3066         if (page0) {
3067                 eb->first_page = page0;
3068                 i = 1;
3069                 index++;
3070                 page_cache_get(page0);
3071                 mark_page_accessed(page0);
3072                 set_page_extent_mapped(page0);
3073                 set_page_extent_head(page0, len);
3074                 uptodate = PageUptodate(page0);
3075         } else {
3076                 i = 0;
3077         }
3078         for (; i < num_pages; i++, index++) {
3079                 p = find_or_create_page(mapping, index, mask | __GFP_HIGHMEM);
3080                 if (!p) {
3081                         WARN_ON(1);
3082                         goto free_eb;
3083                 }
3084                 set_page_extent_mapped(p);
3085                 mark_page_accessed(p);
3086                 if (i == 0) {
3087                         eb->first_page = p;
3088                         set_page_extent_head(p, len);
3089                 } else {
3090                         set_page_private(p, EXTENT_PAGE_PRIVATE);
3091                 }
3092                 if (!PageUptodate(p))
3093                         uptodate = 0;
3094                 unlock_page(p);
3095         }
3096         if (uptodate)
3097                 set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
3098
3099         spin_lock(&tree->buffer_lock);
3100         exists = buffer_tree_insert(tree, start, &eb->rb_node);
3101         if (exists) {
3102                 /* add one reference for the caller */
3103                 atomic_inc(&exists->refs);
3104                 spin_unlock(&tree->buffer_lock);
3105                 goto free_eb;
3106         }
3107         spin_unlock(&tree->buffer_lock);
3108
3109         /* add one reference for the tree */
3110         atomic_inc(&eb->refs);
3111         return eb;
3112
3113 free_eb:
3114         if (!atomic_dec_and_test(&eb->refs))
3115                 return exists;
3116         for (index = 1; index < i; index++)
3117                 page_cache_release(extent_buffer_page(eb, index));
3118         page_cache_release(extent_buffer_page(eb, 0));
3119         __free_extent_buffer(eb);
3120         return exists;
3121 }
3122
3123 struct extent_buffer *find_extent_buffer(struct extent_io_tree *tree,
3124                                          u64 start, unsigned long len,
3125                                           gfp_t mask)
3126 {
3127         struct extent_buffer *eb;
3128
3129         spin_lock(&tree->buffer_lock);
3130         eb = buffer_search(tree, start);
3131         if (eb)
3132                 atomic_inc(&eb->refs);
3133         spin_unlock(&tree->buffer_lock);
3134
3135         if (eb)
3136                 mark_page_accessed(eb->first_page);
3137
3138         return eb;
3139 }
3140
3141 void free_extent_buffer(struct extent_buffer *eb)
3142 {
3143         if (!eb)
3144                 return;
3145
3146         if (!atomic_dec_and_test(&eb->refs))
3147                 return;
3148
3149         WARN_ON(1);
3150 }
3151
3152 int clear_extent_buffer_dirty(struct extent_io_tree *tree,
3153                               struct extent_buffer *eb)
3154 {
3155         unsigned long i;
3156         unsigned long num_pages;
3157         struct page *page;
3158
3159         num_pages = num_extent_pages(eb->start, eb->len);
3160
3161         for (i = 0; i < num_pages; i++) {
3162                 page = extent_buffer_page(eb, i);
3163                 if (!PageDirty(page))
3164                         continue;
3165
3166                 lock_page(page);
3167                 if (i == 0)
3168                         set_page_extent_head(page, eb->len);
3169                 else
3170                         set_page_private(page, EXTENT_PAGE_PRIVATE);
3171
3172                 clear_page_dirty_for_io(page);
3173                 spin_lock_irq(&page->mapping->tree_lock);
3174                 if (!PageDirty(page)) {
3175                         radix_tree_tag_clear(&page->mapping->page_tree,
3176                                                 page_index(page),
3177                                                 PAGECACHE_TAG_DIRTY);
3178                 }
3179                 spin_unlock_irq(&page->mapping->tree_lock);
3180                 unlock_page(page);
3181         }
3182         return 0;
3183 }
3184
3185 int wait_on_extent_buffer_writeback(struct extent_io_tree *tree,
3186                                     struct extent_buffer *eb)
3187 {
3188         return wait_on_extent_writeback(tree, eb->start,
3189                                         eb->start + eb->len - 1);
3190 }
3191
3192 int set_extent_buffer_dirty(struct extent_io_tree *tree,
3193                              struct extent_buffer *eb)
3194 {
3195         unsigned long i;
3196         unsigned long num_pages;
3197         int was_dirty = 0;
3198
3199         was_dirty = test_and_set_bit(EXTENT_BUFFER_DIRTY, &eb->bflags);
3200         num_pages = num_extent_pages(eb->start, eb->len);
3201         for (i = 0; i < num_pages; i++)
3202                 __set_page_dirty_nobuffers(extent_buffer_page(eb, i));
3203         return was_dirty;
3204 }
3205
3206 int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
3207                                 struct extent_buffer *eb)
3208 {
3209         unsigned long i;
3210         struct page *page;
3211         unsigned long num_pages;
3212
3213         num_pages = num_extent_pages(eb->start, eb->len);
3214         clear_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
3215
3216         clear_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
3217                               GFP_NOFS);
3218         for (i = 0; i < num_pages; i++) {
3219                 page = extent_buffer_page(eb, i);
3220                 if (page)
3221                         ClearPageUptodate(page);
3222         }
3223         return 0;
3224 }
3225
3226 int set_extent_buffer_uptodate(struct extent_io_tree *tree,
3227                                 struct extent_buffer *eb)
3228 {
3229         unsigned long i;
3230         struct page *page;
3231         unsigned long num_pages;
3232
3233         num_pages = num_extent_pages(eb->start, eb->len);
3234
3235         set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
3236                             GFP_NOFS);
3237         for (i = 0; i < num_pages; i++) {
3238                 page = extent_buffer_page(eb, i);
3239                 if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) ||
3240                     ((i == num_pages - 1) &&
3241                      ((eb->start + eb->len) & (PAGE_CACHE_SIZE - 1)))) {
3242                         check_page_uptodate(tree, page);
3243                         continue;
3244                 }
3245                 SetPageUptodate(page);
3246         }
3247         return 0;
3248 }
3249
3250 int extent_range_uptodate(struct extent_io_tree *tree,
3251                           u64 start, u64 end)
3252 {
3253         struct page *page;
3254         int ret;
3255         int pg_uptodate = 1;
3256         int uptodate;
3257         unsigned long index;
3258
3259         ret = test_range_bit(tree, start, end, EXTENT_UPTODATE, 1);
3260         if (ret)
3261                 return 1;
3262         while (start <= end) {
3263                 index = start >> PAGE_CACHE_SHIFT;
3264                 page = find_get_page(tree->mapping, index);
3265                 uptodate = PageUptodate(page);
3266                 page_cache_release(page);
3267                 if (!uptodate) {
3268                         pg_uptodate = 0;
3269                         break;
3270                 }
3271                 start += PAGE_CACHE_SIZE;
3272         }
3273         return pg_uptodate;
3274 }
3275
3276 int extent_buffer_uptodate(struct extent_io_tree *tree,
3277                            struct extent_buffer *eb)
3278 {
3279         int ret = 0;
3280         unsigned long num_pages;
3281         unsigned long i;
3282         struct page *page;
3283         int pg_uptodate = 1;
3284
3285         if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))
3286                 return 1;
3287
3288         ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1,
3289                            EXTENT_UPTODATE, 1);
3290         if (ret)
3291                 return ret;
3292
3293         num_pages = num_extent_pages(eb->start, eb->len);
3294         for (i = 0; i < num_pages; i++) {
3295                 page = extent_buffer_page(eb, i);
3296                 if (!PageUptodate(page)) {
3297                         pg_uptodate = 0;
3298                         break;
3299                 }
3300         }
3301         return pg_uptodate;
3302 }
3303
3304 int read_extent_buffer_pages(struct extent_io_tree *tree,
3305                              struct extent_buffer *eb,
3306                              u64 start, int wait,
3307                              get_extent_t *get_extent, int mirror_num)
3308 {
3309         unsigned long i;
3310         unsigned long start_i;
3311         struct page *page;
3312         int err;
3313         int ret = 0;
3314         int locked_pages = 0;
3315         int all_uptodate = 1;
3316         int inc_all_pages = 0;
3317         unsigned long num_pages;
3318         struct bio *bio = NULL;
3319         unsigned long bio_flags = 0;
3320
3321         if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))
3322                 return 0;
3323
3324         if (test_range_bit(tree, eb->start, eb->start + eb->len - 1,
3325                            EXTENT_UPTODATE, 1)) {
3326                 return 0;
3327         }
3328
3329         if (start) {
3330                 WARN_ON(start < eb->start);
3331                 start_i = (start >> PAGE_CACHE_SHIFT) -
3332                         (eb->start >> PAGE_CACHE_SHIFT);
3333         } else {
3334                 start_i = 0;
3335         }
3336
3337         num_pages = num_extent_pages(eb->start, eb->len);
3338         for (i = start_i; i < num_pages; i++) {
3339                 page = extent_buffer_page(eb, i);
3340                 if (!wait) {
3341                         if (!trylock_page(page))
3342                                 goto unlock_exit;
3343                 } else {
3344                         lock_page(page);
3345                 }
3346                 locked_pages++;
3347                 if (!PageUptodate(page))
3348                         all_uptodate = 0;
3349         }
3350         if (all_uptodate) {
3351                 if (start_i == 0)
3352                         set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
3353                 goto unlock_exit;
3354         }
3355
3356         for (i = start_i; i < num_pages; i++) {
3357                 page = extent_buffer_page(eb, i);
3358                 if (inc_all_pages)
3359                         page_cache_get(page);
3360                 if (!PageUptodate(page)) {
3361                         if (start_i == 0)
3362                                 inc_all_pages = 1;
3363                         ClearPageError(page);
3364                         err = __extent_read_full_page(tree, page,
3365                                                       get_extent, &bio,
3366                                                       mirror_num, &bio_flags);
3367                         if (err)
3368                                 ret = err;
3369                 } else {
3370                         unlock_page(page);
3371                 }
3372         }
3373
3374         if (bio)
3375                 submit_one_bio(READ, bio, mirror_num, bio_flags);
3376
3377         if (ret || !wait)
3378                 return ret;
3379
3380         for (i = start_i; i < num_pages; i++) {
3381                 page = extent_buffer_page(eb, i);
3382                 wait_on_page_locked(page);
3383                 if (!PageUptodate(page))
3384                         ret = -EIO;
3385         }
3386
3387         if (!ret)
3388                 set_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
3389         return ret;
3390
3391 unlock_exit:
3392         i = start_i;
3393         while (locked_pages > 0) {
3394                 page = extent_buffer_page(eb, i);
3395                 i++;
3396                 unlock_page(page);
3397                 locked_pages--;
3398         }
3399         return ret;
3400 }
3401
3402 void read_extent_buffer(struct extent_buffer *eb, void *dstv,
3403                         unsigned long start,
3404                         unsigned long len)
3405 {
3406         size_t cur;
3407         size_t offset;
3408         struct page *page;
3409         char *kaddr;
3410         char *dst = (char *)dstv;
3411         size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
3412         unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
3413
3414         WARN_ON(start > eb->len);
3415         WARN_ON(start + len > eb->start + eb->len);
3416
3417         offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
3418
3419         while (len > 0) {
3420                 page = extent_buffer_page(eb, i);
3421
3422                 cur = min(len, (PAGE_CACHE_SIZE - offset));
3423                 kaddr = kmap_atomic(page, KM_USER1);
3424                 memcpy(dst, kaddr + offset, cur);
3425                 kunmap_atomic(kaddr, KM_USER1);
3426
3427                 dst += cur;
3428                 len -= cur;
3429                 offset = 0;
3430                 i++;
3431         }
3432 }
3433
3434 int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start,
3435                                unsigned long min_len, char **token, char **map,
3436                                unsigned long *map_start,
3437                                unsigned long *map_len, int km)
3438 {
3439         size_t offset = start & (PAGE_CACHE_SIZE - 1);
3440         char *kaddr;
3441         struct page *p;
3442         size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
3443         unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
3444         unsigned long end_i = (start_offset + start + min_len - 1) >>
3445                 PAGE_CACHE_SHIFT;
3446
3447         if (i != end_i)
3448                 return -EINVAL;
3449
3450         if (i == 0) {
3451                 offset = start_offset;
3452                 *map_start = 0;
3453         } else {
3454                 offset = 0;
3455                 *map_start = ((u64)i << PAGE_CACHE_SHIFT) - start_offset;
3456         }
3457
3458         if (start + min_len > eb->len) {
3459                 printk(KERN_ERR "btrfs bad mapping eb start %llu len %lu, "
3460                        "wanted %lu %lu\n", (unsigned long long)eb->start,
3461                        eb->len, start, min_len);
3462                 WARN_ON(1);
3463         }
3464
3465         p = extent_buffer_page(eb, i);
3466         kaddr = kmap_atomic(p, km);
3467         *token = kaddr;
3468         *map = kaddr + offset;
3469         *map_len = PAGE_CACHE_SIZE - offset;
3470         return 0;
3471 }
3472
3473 int map_extent_buffer(struct extent_buffer *eb, unsigned long start,
3474                       unsigned long min_len,
3475                       char **token, char **map,
3476                       unsigned long *map_start,
3477                       unsigned long *map_len, int km)
3478 {
3479         int err;
3480         int save = 0;
3481         if (eb->map_token) {
3482                 unmap_extent_buffer(eb, eb->map_token, km);
3483                 eb->map_token = NULL;
3484                 save = 1;
3485         }
3486         err = map_private_extent_buffer(eb, start, min_len, token, map,
3487                                        map_start, map_len, km);
3488         if (!err && save) {
3489                 eb->map_token = *token;
3490                 eb->kaddr = *map;
3491                 eb->map_start = *map_start;
3492                 eb->map_len = *map_len;
3493         }
3494         return err;
3495 }
3496
3497 void unmap_extent_buffer(struct extent_buffer *eb, char *token, int km)
3498 {
3499         kunmap_atomic(token, km);
3500 }
3501
3502 int memcmp_extent_buffer(struct extent_buffer *eb, const void *ptrv,
3503                           unsigned long start,
3504                           unsigned long len)
3505 {
3506         size_t cur;
3507         size_t offset;
3508         struct page *page;
3509         char *kaddr;
3510         char *ptr = (char *)ptrv;
3511         size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
3512         unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
3513         int ret = 0;
3514
3515         WARN_ON(start > eb->len);
3516         WARN_ON(start + len > eb->start + eb->len);
3517
3518         offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
3519
3520         while (len > 0) {
3521                 page = extent_buffer_page(eb, i);
3522
3523                 cur = min(len, (PAGE_CACHE_SIZE - offset));
3524
3525                 kaddr = kmap_atomic(page, KM_USER0);
3526                 ret = memcmp(ptr, kaddr + offset, cur);
3527                 kunmap_atomic(kaddr, KM_USER0);
3528                 if (ret)
3529                         break;
3530
3531                 ptr += cur;
3532                 len -= cur;
3533                 offset = 0;
3534                 i++;
3535         }
3536         return ret;
3537 }
3538
3539 void write_extent_buffer(struct extent_buffer *eb, const void *srcv,
3540                          unsigned long start, unsigned long len)
3541 {
3542         size_t cur;
3543         size_t offset;
3544         struct page *page;
3545         char *kaddr;
3546         char *src = (char *)srcv;
3547         size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
3548         unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
3549
3550         WARN_ON(start > eb->len);
3551         WARN_ON(start + len > eb->start + eb->len);
3552
3553         offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
3554
3555         while (len > 0) {
3556                 page = extent_buffer_page(eb, i);
3557                 WARN_ON(!PageUptodate(page));
3558
3559                 cur = min(len, PAGE_CACHE_SIZE - offset);
3560                 kaddr = kmap_atomic(page, KM_USER1);
3561                 memcpy(kaddr + offset, src, cur);
3562                 kunmap_atomic(kaddr, KM_USER1);
3563
3564                 src += cur;
3565                 len -= cur;
3566                 offset = 0;
3567                 i++;
3568         }
3569 }
3570
3571 void memset_extent_buffer(struct extent_buffer *eb, char c,
3572                           unsigned long start, unsigned long len)
3573 {
3574         size_t cur;
3575         size_t offset;
3576         struct page *page;
3577         char *kaddr;
3578         size_t start_offset = eb->start & ((u64)PAGE_CACHE_SIZE - 1);
3579         unsigned long i = (start_offset + start) >> PAGE_CACHE_SHIFT;
3580
3581         WARN_ON(start > eb->len);
3582         WARN_ON(start + len > eb->start + eb->len);
3583
3584         offset = (start_offset + start) & ((unsigned long)PAGE_CACHE_SIZE - 1);
3585
3586         while (len > 0) {
3587                 page = extent_buffer_page(eb, i);
3588                 WARN_ON(!PageUptodate(page));
3589
3590                 cur = min(len, PAGE_CACHE_SIZE - offset);
3591                 kaddr = kmap_atomic(page, KM_USER0);
3592                 memset(kaddr + offset, c, cur);
3593                 kunmap_atomic(kaddr, KM_USER0);
3594
3595                 len -= cur;
3596                 offset = 0;
3597                 i++;
3598         }
3599 }
3600
3601 void copy_extent_buffer(struct extent_buffer *dst, struct extent_buffer *src,
3602                         unsigned long dst_offset, unsigned long src_offset,
3603                         unsigned long len)
3604 {
3605         u64 dst_len = dst->len;
3606         size_t cur;
3607         size_t offset;
3608         struct page *page;
3609         char *kaddr;
3610         size_t start_offset = dst->start & ((u64)PAGE_CACHE_SIZE - 1);
3611         unsigned long i = (start_offset + dst_offset) >> PAGE_CACHE_SHIFT;
3612
3613         WARN_ON(src->len != dst_len);
3614
3615         offset = (start_offset + dst_offset) &
3616                 ((unsigned long)PAGE_CACHE_SIZE - 1);
3617
3618         while (len > 0) {
3619                 page = extent_buffer_page(dst, i);
3620                 WARN_ON(!PageUptodate(page));
3621
3622                 cur = min(len, (unsigned long)(PAGE_CACHE_SIZE - offset));
3623
3624                 kaddr = kmap_atomic(page, KM_USER0);
3625                 read_extent_buffer(src, kaddr + offset, src_offset, cur);
3626                 kunmap_atomic(kaddr, KM_USER0);
3627
3628                 src_offset += cur;
3629                 len -= cur;
3630                 offset = 0;
3631                 i++;
3632         }
3633 }
3634
3635 static void move_pages(struct page *dst_page, struct page *src_page,
3636                        unsigned long dst_off, unsigned long src_off,
3637                        unsigned long len)
3638 {
3639         char *dst_kaddr = kmap_atomic(dst_page, KM_USER0);
3640         if (dst_page == src_page) {
3641                 memmove(dst_kaddr + dst_off, dst_kaddr + src_off, len);
3642         } else {
3643                 char *src_kaddr = kmap_atomic(src_page, KM_USER1);
3644                 char *p = dst_kaddr + dst_off + len;
3645                 char *s = src_kaddr + src_off + len;
3646
3647                 while (len--)
3648                         *--p = *--s;
3649
3650                 kunmap_atomic(src_kaddr, KM_USER1);
3651         }
3652         kunmap_atomic(dst_kaddr, KM_USER0);
3653 }
3654
3655 static void copy_pages(struct page *dst_page, struct page *src_page,
3656                        unsigned long dst_off, unsigned long src_off,
3657                        unsigned long len)
3658 {
3659         char *dst_kaddr = kmap_atomic(dst_page, KM_USER0);
3660         char *src_kaddr;
3661
3662         if (dst_page != src_page)
3663                 src_kaddr = kmap_atomic(src_page, KM_USER1);
3664         else
3665                 src_kaddr = dst_kaddr;
3666
3667         memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len);
3668         kunmap_atomic(dst_kaddr, KM_USER0);
3669         if (dst_page != src_page)
3670                 kunmap_atomic(src_kaddr, KM_USER1);
3671 }
3672
3673 void memcpy_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
3674                            unsigned long src_offset, unsigned long len)
3675 {
3676         size_t cur;
3677         size_t dst_off_in_page;
3678         size_t src_off_in_page;
3679         size_t start_offset = dst->start & ((u64)PAGE_CACHE_SIZE - 1);
3680         unsigned long dst_i;
3681         unsigned long src_i;
3682
3683         if (src_offset + len > dst->len) {
3684                 printk(KERN_ERR "btrfs memmove bogus src_offset %lu move "
3685                        "len %lu dst len %lu\n", src_offset, len, dst->len);
3686                 BUG_ON(1);
3687         }
3688         if (dst_offset + len > dst->len) {
3689                 printk(KERN_ERR "btrfs memmove bogus dst_offset %lu move "
3690                        "len %lu dst len %lu\n", dst_offset, len, dst->len);
3691                 BUG_ON(1);
3692         }
3693
3694         while (len > 0) {
3695                 dst_off_in_page = (start_offset + dst_offset) &
3696                         ((unsigned long)PAGE_CACHE_SIZE - 1);
3697                 src_off_in_page = (start_offset + src_offset) &
3698                         ((unsigned long)PAGE_CACHE_SIZE - 1);
3699
3700                 dst_i = (start_offset + dst_offset) >> PAGE_CACHE_SHIFT;
3701                 src_i = (start_offset + src_offset) >> PAGE_CACHE_SHIFT;
3702
3703                 cur = min(len, (unsigned long)(PAGE_CACHE_SIZE -
3704                                                src_off_in_page));
3705                 cur = min_t(unsigned long, cur,
3706                         (unsigned long)(PAGE_CACHE_SIZE - dst_off_in_page));
3707
3708                 copy_pages(extent_buffer_page(dst, dst_i),
3709                            extent_buffer_page(dst, src_i),
3710                            dst_off_in_page, src_off_in_page, cur);
3711
3712                 src_offset += cur;
3713                 dst_offset += cur;
3714                 len -= cur;
3715         }
3716 }
3717
3718 void memmove_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,
3719                            unsigned long src_offset, unsigned long len)
3720 {
3721         size_t cur;
3722         size_t dst_off_in_page;
3723         size_t src_off_in_page;
3724         unsigned long dst_end = dst_offset + len - 1;
3725         unsigned long src_end = src_offset + len - 1;
3726         size_t start_offset = dst->start & ((u64)PAGE_CACHE_SIZE - 1);
3727         unsigned long dst_i;
3728         unsigned long src_i;
3729
3730         if (src_offset + len > dst->len) {
3731                 printk(KERN_ERR "btrfs memmove bogus src_offset %lu move "
3732                        "len %lu len %lu\n", src_offset, len, dst->len);
3733                 BUG_ON(1);
3734         }
3735         if (dst_offset + len > dst->len) {
3736                 printk(KERN_ERR "btrfs memmove bogus dst_offset %lu move "
3737                        "len %lu len %lu\n", dst_offset, len, dst->len);
3738                 BUG_ON(1);
3739         }
3740         if (dst_offset < src_offset) {
3741                 memcpy_extent_buffer(dst, dst_offset, src_offset, len);
3742                 return;
3743         }
3744         while (len > 0) {
3745                 dst_i = (start_offset + dst_end) >> PAGE_CACHE_SHIFT;
3746                 src_i = (start_offset + src_end) >> PAGE_CACHE_SHIFT;
3747
3748                 dst_off_in_page = (start_offset + dst_end) &
3749                         ((unsigned long)PAGE_CACHE_SIZE - 1);
3750                 src_off_in_page = (start_offset + src_end) &
3751                         ((unsigned long)PAGE_CACHE_SIZE - 1);
3752
3753                 cur = min_t(unsigned long, len, src_off_in_page + 1);
3754                 cur = min(cur, dst_off_in_page + 1);
3755                 move_pages(extent_buffer_page(dst, dst_i),
3756                            extent_buffer_page(dst, src_i),
3757                            dst_off_in_page - cur + 1,
3758                            src_off_in_page - cur + 1, cur);
3759
3760                 dst_end -= cur;
3761                 src_end -= cur;
3762                 len -= cur;
3763         }
3764 }
3765
3766 int try_release_extent_buffer(struct extent_io_tree *tree, struct page *page)
3767 {
3768         u64 start = page_offset(page);
3769         struct extent_buffer *eb;
3770         int ret = 1;
3771         unsigned long i;
3772         unsigned long num_pages;
3773
3774         spin_lock(&tree->buffer_lock);
3775         eb = buffer_search(tree, start);
3776         if (!eb)
3777                 goto out;
3778
3779         if (atomic_read(&eb->refs) > 1) {
3780                 ret = 0;
3781                 goto out;
3782         }
3783         if (test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)) {
3784                 ret = 0;
3785                 goto out;
3786         }
3787         /* at this point we can safely release the extent buffer */
3788         num_pages = num_extent_pages(eb->start, eb->len);
3789         for (i = 0; i < num_pages; i++)
3790                 page_cache_release(extent_buffer_page(eb, i));
3791         rb_erase(&eb->rb_node, &tree->buffer);
3792         __free_extent_buffer(eb);
3793 out:
3794         spin_unlock(&tree->buffer_lock);
3795         return ret;
3796 }