Btrfs: Extent based page cache code. This uses an rbtree of extents and tests

instead of buffer heads.

Signed-off-by: Chris Mason <chris.mason@oracle.com>

9 files changed, 2203 insertions, 669 deletions

diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile
index 9321438c5ee..ea7a2287035 100644
--- a/fs/btrfs/Makefile
+++ b/fs/btrfs/Makefile
@@ -4,7 +4,8 @@ ifneq ($(KERNELRELEASE),)
 obj-m  := btrfs.o
 btrfs-y := super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
 	   hash.o file-item.o inode-item.o inode-map.o disk-io.o \
-	   transaction.o bit-radix.o inode.o file.o tree-defrag.o
+	   transaction.o bit-radix.o inode.o file.o tree-defrag.o \
+	   extent_map.o
 
 #btrfs-y := ctree.o disk-io.o radix-tree.o extent-tree.o print-tree.o \
 #	  root-tree.o dir-item.o hash.o file-item.o inode-item.o \
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index f446f161f73..d1d5af471c3 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -19,11 +19,14 @@
 #ifndef __BTRFS_I__
 #define __BTRFS_I__
 
+#include "extent_map.h"
+
 /* in memory btrfs inode */
 struct btrfs_inode {
 	struct btrfs_root *root;
 	struct btrfs_block_group_cache *block_group;
 	struct btrfs_key location;
+	struct extent_map_tree extent_tree;
 	struct inode vfs_inode;
 
 	/*
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index c7f5161271d..947f061ed11 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -1176,7 +1176,14 @@ int btrfs_commit_write(struct file *file, struct page *page,
 		       unsigned from, unsigned to);
 int btrfs_get_block(struct inode *inode, sector_t iblock,
 		    struct buffer_head *result, int create);
+struct extent_map *btrfs_get_extent(struct inode *inode, struct page *page,
+				    size_t page_offset, u64 start, u64 end,
+				    int create);
+int btrfs_update_inode(struct btrfs_trans_handle *trans,
+			      struct btrfs_root *root,
+			      struct inode *inode);
 /* file.c */
+int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end);
 extern struct file_operations btrfs_file_operations;
 int btrfs_drop_extents(struct btrfs_trans_handle *trans,
 		       struct btrfs_root *root, struct inode *inode,
diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h
index 81fd18cbd82..5261733b873 100644
--- a/fs/btrfs/disk-io.h
+++ b/fs/btrfs/disk-io.h
@@ -77,7 +77,6 @@ int btrfs_insert_dev_radix(struct btrfs_root *root,
 			   u64 num_blocks);
 int btrfs_map_bh_to_logical(struct btrfs_root *root, struct buffer_head *bh,
 			     u64 logical);
-int btrfs_releasepage(struct page *page, gfp_t flags);
 void btrfs_btree_balance_dirty(struct btrfs_root *root);
 int btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root);
 void btrfs_mark_buffer_dirty(struct buffer_head *bh);
diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c
new file mode 100644
index 00000000000..d378edf0964
--- /dev/null
+++ b/fs/btrfs/extent_map.c
@@ -0,0 +1,1656 @@
+#include <linux/bitops.h>
+#include <linux/slab.h>
+#include <linux/bio.h>
+#include <linux/mm.h>
+#include <linux/gfp.h>
+#include <linux/pagemap.h>
+#include <linux/page-flags.h>
+#include <linux/module.h>
+#include <linux/spinlock.h>
+#include <linux/blkdev.h>
+#include "extent_map.h"
+
+static struct kmem_cache *extent_map_cache;
+static struct kmem_cache *extent_state_cache;
+
+struct tree_entry {
+	u64 start;
+	u64 end;
+	int in_tree;
+	struct rb_node rb_node;
+};
+
+/* bits for the extent state */
+#define EXTENT_DIRTY 1
+#define EXTENT_WRITEBACK (1 << 1)
+#define EXTENT_UPTODATE (1 << 2)
+#define EXTENT_LOCKED (1 << 3)
+#define EXTENT_NEW (1 << 4)
+#define EXTENT_DELALLOC (1 << 5)
+
+#define EXTENT_IOBITS (EXTENT_LOCKED | EXTENT_WRITEBACK)
+
+static LIST_HEAD(all_states);
+spinlock_t state_lock = SPIN_LOCK_UNLOCKED;
+
+void __init extent_map_init(void)
+{
+	extent_map_cache = kmem_cache_create("extent_map",
+					    sizeof(struct extent_map), 0,
+					    SLAB_RECLAIM_ACCOUNT |
+					    SLAB_DESTROY_BY_RCU,
+					    NULL);
+	extent_state_cache = kmem_cache_create("extent_state",
+					    sizeof(struct extent_state), 0,
+					    SLAB_RECLAIM_ACCOUNT |
+					    SLAB_DESTROY_BY_RCU,
+					    NULL);
+}
+
+void __exit extent_map_exit(void)
+{
+	while(!list_empty(&all_states)) {
+		struct extent_state *state;
+		struct list_head *cur = all_states.next;
+		state = list_entry(cur, struct extent_state, list);
+		printk("found leaked state %Lu %Lu state %d in_tree %d\n",
+		       state->start, state->end, state->state, state->in_tree);
+		list_del(&state->list);
+		kfree(state);
+	}
+	if (extent_map_cache)
+		kmem_cache_destroy(extent_map_cache);
+	if (extent_state_cache)
+		kmem_cache_destroy(extent_state_cache);
+}
+
+void extent_map_tree_init(struct extent_map_tree *tree,
+			  struct address_space *mapping, gfp_t mask)
+{
+	tree->map.rb_node = NULL;
+	tree->state.rb_node = NULL;
+	rwlock_init(&tree->lock);
+	tree->mapping = mapping;
+}
+EXPORT_SYMBOL(extent_map_tree_init);
+
+struct extent_map *alloc_extent_map(gfp_t mask)
+{
+	struct extent_map *em;
+	em = kmem_cache_alloc(extent_map_cache, mask);
+	if (!em || IS_ERR(em))
+		return em;
+	em->in_tree = 0;
+	atomic_set(&em->refs, 1);
+	return em;
+}
+EXPORT_SYMBOL(alloc_extent_map);
+
+void free_extent_map(struct extent_map *em)
+{
+	if (atomic_dec_and_test(&em->refs)) {
+		WARN_ON(em->in_tree);
+		kmem_cache_free(extent_map_cache, em);
+	}
+}
+EXPORT_SYMBOL(free_extent_map);
+
+
+struct extent_state *alloc_extent_state(gfp_t mask)
+{
+	struct extent_state *state;
+	state = kmem_cache_alloc(extent_state_cache, mask);
+	if (!state || IS_ERR(state))
+		return state;
+	state->state = 0;
+	state->in_tree = 0;
+	atomic_set(&state->refs, 1);
+	init_waitqueue_head(&state->wq);
+	spin_lock_irq(&state_lock);
+	list_add(&state->list, &all_states);
+	spin_unlock_irq(&state_lock);
+	return state;
+}
+EXPORT_SYMBOL(alloc_extent_state);
+
+void free_extent_state(struct extent_state *state)
+{
+	if (atomic_dec_and_test(&state->refs)) {
+		WARN_ON(state->in_tree);
+		spin_lock_irq(&state_lock);
+		list_del_init(&state->list);
+		spin_unlock_irq(&state_lock);
+		kmem_cache_free(extent_state_cache, state);
+	}
+}
+EXPORT_SYMBOL(free_extent_state);
+
+static struct rb_node *tree_insert(struct rb_root *root, u64 offset,
+				   struct rb_node *node)
+{
+	struct rb_node ** p = &root->rb_node;
+	struct rb_node * parent = NULL;
+	struct tree_entry *entry;
+
+	while(*p) {
+		parent = *p;
+		entry = rb_entry(parent, struct tree_entry, rb_node);
+
+		if (offset < entry->start)
+			p = &(*p)->rb_left;
+		else if (offset > entry->end)
+			p = &(*p)->rb_right;
+		else
+			return parent;
+	}
+
+	entry = rb_entry(node, struct tree_entry, rb_node);
+	entry->in_tree = 1;
+	rb_link_node(node, parent, p);
+	rb_insert_color(node, root);
+	return NULL;
+}
+
+static struct rb_node *__tree_search(struct rb_root *root, u64 offset,
+				   struct rb_node **prev_ret)
+{
+	struct rb_node * n = root->rb_node;
+	struct rb_node *prev = NULL;
+	struct tree_entry *entry;
+	struct tree_entry *prev_entry = NULL;
+
+	while(n) {
+		entry = rb_entry(n, struct tree_entry, rb_node);
+		prev = n;
+		prev_entry = entry;
+
+		if (offset < entry->start)
+			n = n->rb_left;
+		else if (offset > entry->end)
+			n = n->rb_right;
+		else
+			return n;
+	}
+	if (!prev_ret)
+		return NULL;
+	while(prev && offset > prev_entry->end) {
+		prev = rb_next(prev);
+		prev_entry = rb_entry(prev, struct tree_entry, rb_node);
+	}
+	*prev_ret = prev;
+	return NULL;
+}
+
+static inline struct rb_node *tree_search(struct rb_root *root, u64 offset)
+{
+	struct rb_node *prev;
+	struct rb_node *ret;
+	ret = __tree_search(root, offset, &prev);
+	if (!ret)
+		return prev;
+	return ret;
+}
+
+static int tree_delete(struct rb_root *root, u64 offset)
+{
+	struct rb_node *node;
+	struct tree_entry *entry;
+
+	node = __tree_search(root, offset, NULL);
+	if (!node)
+		return -ENOENT;
+	entry = rb_entry(node, struct tree_entry, rb_node);
+	entry->in_tree = 0;
+	rb_erase(node, root);
+	return 0;
+}
+
+/*
+ * add_extent_mapping tries a simple backward merge with existing
+ * mappings.  The extent_map struct passed in will be inserted into
+ * the tree directly (no copies made, just a reference taken).
+ */
+int add_extent_mapping(struct extent_map_tree *tree,
+		       struct extent_map *em)
+{
+	int ret = 0;
+	struct extent_map *prev = NULL;
+	struct rb_node *rb;
+
+	write_lock_irq(&tree->lock);
+	rb = tree_insert(&tree->map, em->end, &em->rb_node);
+	if (rb) {
+		prev = rb_entry(rb, struct extent_map, rb_node);
+		printk("found extent map %Lu %Lu on insert of %Lu %Lu\n", prev->start, prev->end, em->start, em->end);
+		ret = -EEXIST;
+		goto out;
+	}
+	atomic_inc(&em->refs);
+	if (em->start != 0) {
+		rb = rb_prev(&em->rb_node);
+		if (rb)
+			prev = rb_entry(rb, struct extent_map, rb_node);
+		if (prev && prev->end + 1 == em->start &&
+		    ((em->block_start == 0 && prev->block_start == 0) ||
+			     (em->block_start == prev->block_end + 1))) {
+			em->start = prev->start;
+			em->block_start = prev->block_start;
+			rb_erase(&prev->rb_node, &tree->map);
+			prev->in_tree = 0;
+			free_extent_map(prev);
+		}
+	 }
+out:
+	write_unlock_irq(&tree->lock);
+	return ret;
+}
+EXPORT_SYMBOL(add_extent_mapping);
+
+/*
+ * lookup_extent_mapping returns the first extent_map struct in the
+ * tree that intersects the [start, end] (inclusive) range.  There may
+ * be additional objects in the tree that intersect, so check the object
+ * returned carefully to make sure you don't need additional lookups.
+ */
+struct extent_map *lookup_extent_mapping(struct extent_map_tree *tree,
+					 u64 start, u64 end)
+{
+	struct extent_map *em;
+	struct rb_node *rb_node;
+
+	read_lock_irq(&tree->lock);
+	rb_node = tree_search(&tree->map, start);
+	if (!rb_node) {
+		em = NULL;
+		goto out;
+	}
+	if (IS_ERR(rb_node)) {
+		em = ERR_PTR(PTR_ERR(rb_node));
+		goto out;
+	}
+	em = rb_entry(rb_node, struct extent_map, rb_node);
+	if (em->end < start || em->start > end) {
+		em = NULL;
+		goto out;
+	}
+	atomic_inc(&em->refs);
+out:
+	read_unlock_irq(&tree->lock);
+	return em;
+}
+EXPORT_SYMBOL(lookup_extent_mapping);
+
+/*
+ * removes an extent_map struct from the tree.  No reference counts are
+ * dropped, and no checks are done to  see if the range is in use
+ */
+int remove_extent_mapping(struct extent_map_tree *tree, struct extent_map *em)
+{
+	int ret;
+
+	write_lock_irq(&tree->lock);
+	ret = tree_delete(&tree->map, em->end);
+	write_unlock_irq(&tree->lock);
+	return ret;
+}
+EXPORT_SYMBOL(remove_extent_mapping);
+
+/*
+ * utility function to look for merge candidates inside a given range.
+ * Any extents with matching state are merged together into a single
+ * extent in the tree.  Extents with EXTENT_IO in their state field
+ * are not merged because the end_io handlers need to be able to do
+ * operations on them without sleeping (or doing allocations/splits).
+ *
+ * This should be called with the tree lock held.
+ */
+static int merge_state(struct extent_map_tree *tree,
+		       struct extent_state *state)
+{
+	struct extent_state *other;
+	struct rb_node *other_node;
+
+	if (state->state & EXTENT_IOBITS)
+		return 0;
+
+	other_node = rb_prev(&state->rb_node);
+	if (other_node) {
+		other = rb_entry(other_node, struct extent_state, rb_node);
+		if (other->end == state->start - 1 &&
+		    other->state == state->state) {
+			state->start = other->start;
+			other->in_tree = 0;
+			rb_erase(&other->rb_node, &tree->state);
+			free_extent_state(other);
+		}
+	}
+	other_node = rb_next(&state->rb_node);
+	if (other_node) {
+		other = rb_entry(other_node, struct extent_state, rb_node);
+		if (other->start == state->end + 1 &&
+		    other->state == state->state) {
+			other->start = state->start;
+			state->in_tree = 0;
+			rb_erase(&state->rb_node, &tree->state);
+			free_extent_state(state);
+		}
+	}
+	return 0;
+}
+
+/*
+ * insert an extent_state struct into the tree.  'bits' are set on the
+ * struct before it is inserted.
+ *
+ * This may return -EEXIST if the extent is already there, in which case the
+ * state struct is freed.
+ *
+ * The tree lock is not taken internally.  This is a utility function and
+ * probably isn't what you want to call (see set/clear_extent_bit).
+ */
+static int insert_state(struct extent_map_tree *tree,
+			struct extent_state *state, u64 start, u64 end,
+			int bits)
+{
+	struct rb_node *node;
+
+	if (end < start) {
+		printk("end < start %Lu %Lu\n", end, start);
+		WARN_ON(1);
+	}
+	state->state |= bits;
+	state->start = start;
+	state->end = end;
+	if ((end & 4095) == 0) {
+		printk("insert state %Lu %Lu strange end\n", start, end);
+		WARN_ON(1);
+	}
+	node = tree_insert(&tree->state, end, &state->rb_node);
+	if (node) {
+		struct extent_state *found;
+		found = rb_entry(node, struct extent_state, rb_node);
+printk("found node %Lu %Lu on insert of %Lu %Lu\n", found->start, found->end, start, end);
+		free_extent_state(state);
+		return -EEXIST;
+	}
+	merge_state(tree, state);
+	return 0;
+}
+
+/*
+ * split a given extent state struct in two, inserting the preallocated
+ * struct 'prealloc' as the newly created second half.  'split' indicates an
+ * offset inside 'orig' where it should be split.
+ *
+ * Before calling,
+ * the tree has 'orig' at [orig->start, orig->end].  After calling, there
+ * are two extent state structs in the tree:
+ * prealloc: [orig->start, split - 1]
+ * orig: [ split, orig->end ]
+ *
+ * The tree locks are not taken by this function. They need to be held
+ * by the caller.
+ */
+static int split_state(struct extent_map_tree *tree, struct extent_state *orig,
+		       struct extent_state *prealloc, u64 split)
+{
+	struct rb_node *node;
+	prealloc->start = orig->start;
+	prealloc->end = split - 1;
+	prealloc->state = orig->state;
+	orig->start = split;
+	if ((prealloc->end & 4095) == 0) {
+		printk("insert state %Lu %Lu strange end\n", prealloc->start,
+		       prealloc->end);
+		WARN_ON(1);
+	}
+	node = tree_insert(&tree->state, prealloc->end, &prealloc->rb_node);
+	if (node) {
+		struct extent_state *found;
+		found = rb_entry(node, struct extent_state, rb_node);
+printk("found node %Lu %Lu on insert of %Lu %Lu\n", found->start, found->end, prealloc->start, prealloc->end);
+		free_extent_state(prealloc);
+		return -EEXIST;
+	}
+	return 0;
+}
+
+/*
+ * utility function to clear some bits in an extent state struct.
+ * it will optionally wake up any one waiting on this state (wake == 1), or
+ * forcibly remove the state from the tree (delete == 1).
+ *
+ * If no bits are set on the state struct after clearing things, the
+ * struct is freed and removed from the tree
+ */
+static int clear_state_bit(struct extent_map_tree *tree,
+			    struct extent_state *state, int bits, int wake,
+			    int delete)
+{
+	int ret = state->state & bits;
+	state->state &= ~bits;
+	if (wake)
+		wake_up(&state->wq);
+	if (delete || state->state == 0) {
+		if (state->in_tree) {
+			rb_erase(&state->rb_node, &tree->state);
+			state->in_tree = 0;
+			free_extent_state(state);
+		} else {
+			WARN_ON(1);
+		}
+	} else {
+		merge_state(tree, state);
+	}
+	return ret;
+}
+
+/*
+ * clear some bits on a range in the tree.  This may require splitting
+ * or inserting elements in the tree, so the gfp mask is used to
+ * indicate which allocations or sleeping are allowed.
+ *
+ * pass 'wake' == 1 to kick any sleepers, and 'delete' == 1 to remove
+ * the given range from the tree regardless of state (ie for truncate).
+ *
+ * the range [start, end] is inclusive.
+ *
+ * This takes the tree lock, and returns < 0 on error, > 0 if any of the
+ * bits were already set, or zero if none of the bits were already set.
+ */
+int clear_extent_bit(struct extent_map_tree *tree, u64 start, u64 end,
+		     int bits, int wake, int delete, gfp_t mask)
+{
+	struct extent_state *state;
+	struct extent_state *prealloc = NULL;
+	struct rb_node *node;
+	int err;
+	int set = 0;
+
+again:
+	if (!prealloc && (mask & __GFP_WAIT)) {
+		prealloc = alloc_extent_state(mask);
+		if (!prealloc)
+			return -ENOMEM;
+	}
+
+	write_lock_irq(&tree->lock);
+	/*
+	 * this search will find the extents that end after
+	 * our range starts
+	 */
+	node = tree_search(&tree->state, start);
+	if (!node)
+		goto out;
+	state = rb_entry(node, struct extent_state, rb_node);
+	if (state->start > end)
+		goto out;
+	WARN_ON(state->end < start);
+
+	/*
+	 *     | ---- desired range ---- |
+	 *  | state | or
+	 *  | ------------- state -------------- |
+	 *
+	 * We need to split the extent we found, and may flip
+	 * bits on second half.
+	 *
+	 * If the extent we found extends past our range, we
+	 * just split and search again.  It'll get split again
+	 * the next time though.
+	 *
+	 * If the extent we found is inside our range, we clear
+	 * the desired bit on it.
+	 */
+
+	if (state->start < start) {
+		err = split_state(tree, state, prealloc, start);
+		BUG_ON(err == -EEXIST);
+		prealloc = NULL;
+		if (err)
+			goto out;
+		if (state->end <= end) {
+			start = state->end + 1;
+			set |= clear_state_bit(tree, state, bits,
+					wake, delete);
+		} else {
+			start = state->start;
+		}
+		goto search_again;
+	}
+	/*
+	 * | ---- desired range ---- |
+	 *                        | state |
+	 * We need to split the extent, and clear the bit
+	 * on the first half
+	 */
+	if (state->start <= end && state->end > end) {
+		err = split_state(tree, state, prealloc, end + 1);
+		BUG_ON(err == -EEXIST);
+
+		if (wake)
+			wake_up(&state->wq);
+		set |= clear_state_bit(tree, prealloc, bits,
+				       wake, delete);
+		prealloc = NULL;
+		goto out;
+	}
+
+	start = state->end + 1;
+	set |= clear_state_bit(tree, state, bits, wake, delete);
+	goto search_again;
+
+out:
+	write_unlock_irq(&tree->lock);
+	if (prealloc)
+		free_extent_state(prealloc);
+
+	return set;
+
+search_again:
+	if (start >= end)
+		goto out;
+	write_unlock_irq(&tree->lock);
+	if (mask & __GFP_WAIT)
+		cond_resched();
+	goto again;
+}
+EXPORT_SYMBOL(clear_extent_bit);
+
+static int wait_on_state(struct extent_map_tree *tree,
+			 struct extent_state *state)
+{
+	DEFINE_WAIT(wait);
+	prepare_to_wait(&state->wq, &wait, TASK_UNINTERRUPTIBLE);
+	read_unlock_irq(&tree->lock);
+	schedule();
+	read_lock_irq(&tree->lock);
+	finish_wait(&state->wq, &wait);
+	return 0;
+}
+
+/*
+ * waits for one or more bits to clear on a range in the state tree.
+ * The range [start, end] is inclusive.
+ * The tree lock is taken by this function
+ */
+int wait_extent_bit(struct extent_map_tree *tree, u64 start, u64 end, int bits)
+{
+	struct extent_state *state;
+	struct rb_node *node;
+
+	read_lock_irq(&tree->lock);
+again:
+	while (1) {
+		/*
+		 * this search will find all the extents that end after
+		 * our range starts
+		 */
+		node = tree_search(&tree->state, start);
+		if (!node)
+			break;
+
+		state = rb_entry(node, struct extent_state, rb_node);
+
+		if (state->start > end)
+			goto out;
+
+		if (state->state & bits) {
+			start = state->start;
+			atomic_inc(&state->refs);
+			wait_on_state(tree, state);
+			free_extent_state(state);
+			goto again;
+		}
+		start = state->end + 1;
+
+		if (start > end)
+			break;
+
+		if (need_resched()) {
+			read_unlock_irq(&tree->lock);
+			cond_resched();
+			read_lock_irq(&tree->lock);
+		}
+	}
+out:
+	read_unlock_irq(&tree->lock);
+	return 0;
+}
+EXPORT_SYMBOL(wait_extent_bit);
+
+/*
+ * set some bits on a range in the tree.  This may require allocations
+ * or sleeping, so the gfp mask is used to indicate what is allowed.
+ *
+ * If 'exclusive' == 1, this will fail with -EEXIST if some part of the
+ * range already has the desired bits set.  The start of the existing
+ * range is returned in failed_start in this case.
+ *
+ * [start, end] is inclusive
+ * This takes the tree lock.
+ */
+int set_extent_bit(struct extent_map_tree *tree, u64 start, u64 end, int bits,
+		   int exclusive, u64 *failed_start, gfp_t mask)
+{
+	struct extent_state *state;
+	struct extent_state *prealloc = NULL;
+	struct rb_node *node;
+	int err = 0;
+	int set;
+	u64 last_start;
+	u64 last_end;
+again:
+	if (!prealloc && (mask & __GFP_WAIT)) {
+		prealloc = alloc_extent_state(mask);
+		if (!prealloc)
+			return -ENOMEM;
+	}
+
+	write_lock_irq(&tree->lock);
+	/*
+	 * this search will find all the extents that end after
+	 * our range starts.
+	 */
+	node = tree_search(&tree->state, start);
+	if (!node) {
+		err = insert_state(tree, prealloc, start, end, bits);
+		prealloc = NULL;
+		BUG_ON(err == -EEXIST);
+		goto out;
+	}
+
+	state = rb_entry(node, struct extent_state, rb_node);
+	last_start = state->start;
+	last_end = state->end;
+
+	/*
+	 * | ---- desired range ---- |
+	 * | state |
+	 *
+	 * Just lock what we found and keep going
+	 */
+	if (state->start == start && state->end <= end) {
+		set = state->state & bits;
+		if (set && exclusive) {
+			*failed_start = state->start;
+			err = -EEXIST;
+			goto out;
+		}
+		state->state |= bits;
+		start = state->end + 1;
+		merge_state(tree, state);
+		goto search_again;
+	}
+
+	/*
+	 *     | ---- desired range ---- |
+	 * | state |
+	 *   or
+	 * | ------------- state -------------- |
+	 *
+	 * We need to split the extent we found, and may flip bits on
+	 * second half.
+	 *
+	 * If the extent we found extends past our
+	 * range, we just split and search again.  It'll get split
+	 * again the next time though.
+	 *
+	 * If the extent we found is inside our range, we set the
+	 * desired bit on it.
+	 */
+	if (state->start < start) {
+		set = state->state & bits;
+		if (exclusive && set) {
+			*failed_start = start;
+			err = -EEXIST;
+			goto out;
+		}
+		err = split_state(tree, state, prealloc, start);
+		BUG_ON(err == -EEXIST);
+		prealloc = NULL;
+		if (err)
+			goto out;
+		if (state->end <= end) {
+			state->state |= bits;
+			start = state->end + 1;
+			merge_state(tree, state);
+		} else {
+			start = state->start;
+		}
+		goto search_again;
+	}
+	/*
+	 * | ---- desired range ---- |
+	 *                        | state |
+	 * We need to split the extent, and set the bit
+	 * on the first half
+	 */
+	if (state->start <= end && state->end > end) {
+		set = state->state & bits;
+		if (exclusive && set) {
+			*failed_start = start;
+			err = -EEXIST;
+			goto out;
+		}
+		err = split_state(tree, state, prealloc, end + 1);
+		BUG_ON(err == -EEXIST);
+
+		prealloc->state |= bits;
+		merge_state(tree, prealloc);
+		prealloc = NULL;
+		goto out;
+	}
+
+	/*
+	 * | ---- desired range ---- |
+	 *     | state | or               | state |
+	 *
+	 * There's a hole, we need to insert something in it and
+	 * ignore the extent we found.
+	 */
+	if (state->start > start) {
+		u64 this_end;
+		if (end < last_start)
+			this_end = end;
+		else
+			this_end = last_start -1;
+		err = insert_state(tree, prealloc, start, this_end,
+				   bits);
+		prealloc = NULL;
+		BUG_ON(err == -EEXIST);
+		if (err)
+			goto out;
+		start = this_end + 1;
+		goto search_again;
+	}
+	goto search_again;
+
+out:
+	write_unlock_irq(&tree->lock);
+	if (prealloc)
+		free_extent_state(prealloc);
+
+	return err;
+
+search_again:
+	if (start > end)
+		goto out;
+	write_unlock_irq(&tree->lock);
+	if (mask & __GFP_WAIT)
+		cond_resched();
+	goto again;
+}
+EXPORT_SYMBOL(set_extent_bit);
+
+/* wrappers around set/clear extent bit */
+int set_extent_dirty(struct extent_map_tree *tree, u64 start, u64 end,
+		     gfp_t mask)
+{
+	return set_extent_bit(tree, start, end, EXTENT_DIRTY, 0, NULL,
+			      mask);
+}
+EXPORT_SYMBOL(set_extent_dirty);
+
+int clear_extent_dirty(struct extent_map_tree *tree, u64 start, u64 end,
+		       gfp_t mask)
+{
+	return clear_extent_bit(tree, start, end, EXTENT_DIRTY, 0, 0, mask);
+}
+EXPORT_SYMBOL(clear_extent_dirty);
+
+int set_extent_new(struct extent_map_tree *tree, u64 start, u64 end,
+		     gfp_t mask)
+{
+	return set_extent_bit(tree, start, end, EXTENT_NEW, 0, NULL,
+			      mask);
+}
+EXPORT_SYMBOL(set_extent_new);
+
+int clear_extent_new(struct extent_map_tree *tree, u64 start, u64 end,
+		       gfp_t mask)
+{
+	return clear_extent_bit(tree, start, end, EXTENT_NEW, 0, 0, mask);
+}
+EXPORT_SYMBOL(clear_extent_new);
+
+int set_extent_uptodate(struct extent_map_tree *tree, u64 start, u64 end,
+			gfp_t mask)
+{
+	return set_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, NULL,
+			      mask);
+}
+EXPORT_SYMBOL(set_extent_uptodate);
+
+int clear_extent_uptodate(struct extent_map_tree *tree, u64 start, u64 end,
+			  gfp_t mask)
+{
+	return clear_extent_bit(tree, start, end, EXTENT_UPTODATE, 0, 0, mask);
+}
+EXPORT_SYMBOL(clear_extent_uptodate);
+
+int set_extent_writeback(struct extent_map_tree *tree, u64 start, u64 end,
+			 gfp_t mask)
+{
+	return set_extent_bit(tree, start, end, EXTENT_WRITEBACK,
+			      0, NULL, mask);
+}
+EXPORT_SYMBOL(set_extent_writeback);
+
+int clear_extent_writeback(struct extent_map_tree *tree, u64 start, u64 end,
+			   gfp_t mask)
+{
+	return clear_extent_bit(tree, start, end, EXTENT_WRITEBACK, 1, 0, mask);
+}
+EXPORT_SYMBOL(clear_extent_writeback);
+
+int wait_on_extent_writeback(struct extent_map_tree *tree, u64 start, u64 end)
+{
+	return wait_extent_bit(tree, start, end, EXTENT_WRITEBACK);
+}
+EXPORT_SYMBOL(wait_on_extent_writeback);
+
+/*
+ * locks a range in ascending order, waiting for any locked regions
+ * it hits on the way.  [start,end] are inclusive, and this will sleep.
+ */
+int lock_extent(struct extent_map_tree *tree, u64 start, u64 end, gfp_t mask)
+{
+	int err;
+	u64 failed_start;
+	while (1) {
+		err = set_extent_bit(tree, start, end, EXTENT_LOCKED, 1,
+				     &failed_start, mask);
+		if (err == -EEXIST && (mask & __GFP_WAIT)) {
+			wait_extent_bit(tree, failed_start, end, EXTENT_LOCKED);
+			start = failed_start;
+		} else {
+			break;
+		}
+		WARN_ON(start > end);
+	}
+	return err;
+}
+EXPORT_SYMBOL(lock_extent);
+
+int unlock_extent(struct extent_map_tree *tree, u64 start, u64 end,
+		  gfp_t mask)
+{
+	return clear_extent_bit(tree, start, end, EXTENT_LOCKED, 1, 0, mask);
+}
+EXPORT_SYMBOL(unlock_extent);
+
+/*
+ * helper function to set pages and extents in the tree dirty
+ */
+int set_range_dirty(struct extent_map_tree *tree, u64 start, u64 end)
+{
+	unsigned long index = start >> PAGE_CACHE_SHIFT;
+	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
+	struct page *page;
+
+	while (index <= end_index) {
+		page = find_get_page(tree->mapping, index);
+		BUG_ON(!page);
+		__set_page_dirty_nobuffers(page);
+		page_cache_release(page);
+		index++;
+	}
+	set_extent_dirty(tree, start, end, GFP_NOFS);
+	return 0;
+}
+EXPORT_SYMBOL(set_range_dirty);
+
+/*
+ * helper function to set both pages and extents in the tree writeback
+ */
+int set_range_writeback(struct extent_map_tree *tree, u64 start, u64 end)
+{
+	unsigned long index = start >> PAGE_CACHE_SHIFT;
+	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
+	struct page *page;
+
+	while (index <= end_index) {
+		page = find_get_page(tree->mapping, index);
+		BUG_ON(!page);
+		set_page_writeback(page);
+		page_cache_release(page);
+		index++;
+	}
+	set_extent_writeback(tree, start, end, GFP_NOFS);
+	return 0;
+}
+EXPORT_SYMBOL(set_range_writeback);
+
+/*
+ * helper function to lock both pages and extents in the tree.
+ * pages must be locked first.
+ */
+int lock_range(struct extent_map_tree *tree, u64 start, u64 end)
+{
+	unsigned long index = start >> PAGE_CACHE_SHIFT;
+	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
+	struct page *page;
+	int err;
+
+	while (index <= end_index) {
+		page = grab_cache_page(tree->mapping, index);
+		if (!page) {
+			err = -ENOMEM;
+			goto failed;
+		}
+		if (IS_ERR(page)) {
+			err = PTR_ERR(page);
+			goto failed;
+		}
+		index++;
+	}
+	lock_extent(tree, start, end, GFP_NOFS);
+	return 0;
+
+failed:
+	/*
+	 * we failed above in getting the page at 'index', so we undo here
+	 * up to but not including the page at 'index'
+	 */
+	end_index = index;
+	index = start >> PAGE_CACHE_SHIFT;
+	while (index < end_index) {
+		page = find_get_page(tree->mapping, index);
+		unlock_page(page);
+		page_cache_release(page);
+		index++;
+	}
+	return err;
+}
+EXPORT_SYMBOL(lock_range);
+
+/*
+ * helper function to unlock both pages and extents in the tree.
+ */
+int unlock_range(struct extent_map_tree *tree, u64 start, u64 end)
+{
+	unsigned long index = start >> PAGE_CACHE_SHIFT;
+	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
+	struct page *page;
+
+	while (index <= end_index) {
+		page = find_get_page(tree->mapping, index);
+		unlock_page(page);
+		page_cache_release(page);
+		index++;
+	}
+	unlock_extent(tree, start, end, GFP_NOFS);
+	return 0;
+}
+EXPORT_SYMBOL(unlock_range);
+
+/*
+ * searches a range in the state tree for a given mask.
+ * If 'filled' == 1, this returns 1 only if ever extent in the tree
+ * has the bits set.  Otherwise, 1 is returned if any bit in the
+ * range is found set.
+ */
+static int test_range_bit(struct extent_map_tree *tree, u64 start, u64 end,
+			  int bits, int filled)
+{
+	struct extent_state *state = NULL;
+	struct rb_node *node;
+	int bitset = 0;
+
+	read_lock_irq(&tree->lock);
+	node = tree_search(&tree->state, start);
+	while (node && start <= end) {
+		state = rb_entry(node, struct extent_state, rb_node);
+		if (state->start > end)
+			break;
+
+		if (filled && state->start > start) {
+			bitset = 0;
+			break;
+		}
+		if (state->state & bits) {
+			bitset = 1;
+			if (!filled)
+				break;
+		} else if (filled) {
+			bitset = 0;
+			break;
+		}
+		start = state->end + 1;
+		if (start > end)
+			break;
+		node = rb_next(node);
+	}
+	read_unlock_irq(&tree->lock);
+	return bitset;
+}
+
+/*
+ * helper function to set a given page up to date if all the
+ * extents in the tree for that page are up to date
+ */
+static int check_page_uptodate(struct extent_map_tree *tree,
+			       struct page *page)
+{
+	u64 start = page->index << PAGE_CACHE_SHIFT;
+	u64 end = start + PAGE_CACHE_SIZE - 1;
+