Btrfs: Split the extent_map code into two parts

There is now extent_map for mapping offsets in the file to disk and
extent_io for state tracking, IO submission and extent_bufers.

The new extent_map code shifts from [start,end] pairs to [start,len], and
pushes the locking out into the caller.  This allows a few performance
optimizations and is easier to use.

A number of extent_map usage bugs were fixed, mostly with failing
to remove extent_map entries when changing the file.

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

14 files changed, 3612 insertions, 3413 deletions

diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile
index 72cc3e16741..9c61609f371 100644
--- a/fs/btrfs/Makefile
+++ b/fs/btrfs/Makefile
@@ -5,7 +5,8 @@ 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 \
-	   extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o
+	   extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o \
+	   extent_io.o
 
 ifeq ($(CONFIG_FS_POSIX_ACL),y)
 btrfs-y += acl.o
diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index 9fc99883c70..9a0647dc5a0 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -20,6 +20,7 @@
 #define __BTRFS_I__
 
 #include "extent_map.h"
+#include "extent_io.h"
 
 /* in memory btrfs inode */
 struct btrfs_inode {
@@ -27,6 +28,7 @@ struct btrfs_inode {
 	struct btrfs_block_group_cache *block_group;
 	struct btrfs_key location;
 	struct extent_map_tree extent_tree;
+	struct extent_io_tree io_tree;
 	struct inode vfs_inode;
 
 	u64 ordered_trans;
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index b9f2975b55c..6c65473e0fe 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -27,6 +27,7 @@
 #include <linux/completion.h>
 #include <asm/kmap_types.h>
 #include "bit-radix.h"
+#include "extent_io.h"
 #include "extent_map.h"
 
 struct btrfs_trans_handle;
@@ -314,11 +315,11 @@ struct btrfs_fs_info {
 	struct btrfs_root *tree_root;
 	struct radix_tree_root fs_roots_radix;
 
-	struct extent_map_tree free_space_cache;
-	struct extent_map_tree block_group_cache;
-	struct extent_map_tree pinned_extents;
-	struct extent_map_tree pending_del;
-	struct extent_map_tree extent_ins;
+	struct extent_io_tree free_space_cache;
+	struct extent_io_tree block_group_cache;
+	struct extent_io_tree pinned_extents;
+	struct extent_io_tree pending_del;
+	struct extent_io_tree extent_ins;
 
 	u64 generation;
 	u64 last_trans_committed;
@@ -956,7 +957,7 @@ u32 btrfs_count_snapshots_in_path(struct btrfs_root *root,
 				  u64 first_extent);
 int btrfs_extent_post_op(struct btrfs_trans_handle *trans,
 			 struct btrfs_root *root);
-int btrfs_copy_pinned(struct btrfs_root *root, struct extent_map_tree *copy);
+int btrfs_copy_pinned(struct btrfs_root *root, struct extent_io_tree *copy);
 struct btrfs_block_group_cache *btrfs_lookup_block_group(struct
 							 btrfs_fs_info *info,
 							 u64 bytenr);
@@ -1001,7 +1002,7 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root
 		      u64 owner_objectid, u64 owner_offset, int pin);
 int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 			       struct btrfs_root *root,
-			       struct extent_map_tree *unpin);
+			       struct extent_io_tree *unpin);
 int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
 				struct btrfs_root *root,
 				u64 bytenr, u64 num_bytes,
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 5d1f9bca271..4c4ebea0b2a 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -43,14 +43,14 @@ static int check_tree_block(struct btrfs_root *root, struct extent_buffer *buf)
 }
 #endif
 
-static struct extent_map_ops btree_extent_map_ops;
+static struct extent_io_ops btree_extent_io_ops;
 
 struct extent_buffer *btrfs_find_tree_block(struct btrfs_root *root,
 					    u64 bytenr, u32 blocksize)
 {
 	struct inode *btree_inode = root->fs_info->btree_inode;
 	struct extent_buffer *eb;
-	eb = find_extent_buffer(&BTRFS_I(btree_inode)->extent_tree,
+	eb = find_extent_buffer(&BTRFS_I(btree_inode)->io_tree,
 				bytenr, blocksize, GFP_NOFS);
 	return eb;
 }
@@ -61,13 +61,13 @@ struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root,
 	struct inode *btree_inode = root->fs_info->btree_inode;
 	struct extent_buffer *eb;
 
-	eb = alloc_extent_buffer(&BTRFS_I(btree_inode)->extent_tree,
+	eb = alloc_extent_buffer(&BTRFS_I(btree_inode)->io_tree,
 				 bytenr, blocksize, NULL, GFP_NOFS);
 	return eb;
 }
 
 struct extent_map *btree_get_extent(struct inode *inode, struct page *page,
-				    size_t page_offset, u64 start, u64 end,
+				    size_t page_offset, u64 start, u64 len,
 				    int create)
 {
 	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
@@ -75,7 +75,9 @@ struct extent_map *btree_get_extent(struct inode *inode, struct page *page,
 	int ret;
 
 again:
-	em = lookup_extent_mapping(em_tree, start, end);
+	spin_lock(&em_tree->lock);
+	em = lookup_extent_mapping(em_tree, start, len);
+	spin_unlock(&em_tree->lock);
 	if (em) {
 		goto out;
 	}
@@ -85,11 +87,14 @@ again:
 		goto out;
 	}
 	em->start = 0;
-	em->end = (i_size_read(inode) & ~((u64)PAGE_CACHE_SIZE -1)) - 1;
+	em->len = i_size_read(inode);
 	em->block_start = 0;
-	em->block_end = em->end;
 	em->bdev = inode->i_sb->s_bdev;
+
+	spin_lock(&em_tree->lock);
 	ret = add_extent_mapping(em_tree, em);
+	spin_unlock(&em_tree->lock);
+
 	if (ret == -EEXIST) {
 		free_extent_map(em);
 		em = NULL;
@@ -175,13 +180,13 @@ static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf,
 
 int csum_dirty_buffer(struct btrfs_root *root, struct page *page)
 {
-	struct extent_map_tree *tree;
+	struct extent_io_tree *tree;
 	u64 start = (u64)page->index << PAGE_CACHE_SHIFT;
 	u64 found_start;
 	int found_level;
 	unsigned long len;
 	struct extent_buffer *eb;
-	tree = &BTRFS_I(page->mapping->host)->extent_tree;
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
 
 	if (page->private == EXTENT_PAGE_PRIVATE)
 		goto out;
@@ -230,16 +235,16 @@ static int btree_writepage_io_hook(struct page *page, u64 start, u64 end)
 
 static int btree_writepage(struct page *page, struct writeback_control *wbc)
 {
-	struct extent_map_tree *tree;
-	tree = &BTRFS_I(page->mapping->host)->extent_tree;
+	struct extent_io_tree *tree;
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
 	return extent_write_full_page(tree, page, btree_get_extent, wbc);
 }
 
 static int btree_writepages(struct address_space *mapping,
 			    struct writeback_control *wbc)
 {
-	struct extent_map_tree *tree;
-	tree = &BTRFS_I(mapping->host)->extent_tree;
+	struct extent_io_tree *tree;
+	tree = &BTRFS_I(mapping->host)->io_tree;
 	if (wbc->sync_mode == WB_SYNC_NONE) {
 		u64 num_dirty;
 		u64 start = 0;
@@ -264,18 +269,20 @@ static int btree_writepages(struct address_space *mapping,
 
 int btree_readpage(struct file *file, struct page *page)
 {
-	struct extent_map_tree *tree;
-	tree = &BTRFS_I(page->mapping->host)->extent_tree;
+	struct extent_io_tree *tree;
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
 	return extent_read_full_page(tree, page, btree_get_extent);
 }
 
 static int btree_releasepage(struct page *page, gfp_t unused_gfp_flags)
 {
-	struct extent_map_tree *tree;
+	struct extent_io_tree *tree;
+	struct extent_map_tree *map;
 	int ret;
 
-	tree = &BTRFS_I(page->mapping->host)->extent_tree;
-	ret = try_release_extent_mapping(tree, page);
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
+	map = &BTRFS_I(page->mapping->host)->extent_tree;
+	ret = try_release_extent_mapping(map, tree, page);
 	if (ret == 1) {
 		ClearPagePrivate(page);
 		set_page_private(page, 0);
@@ -286,8 +293,8 @@ static int btree_releasepage(struct page *page, gfp_t unused_gfp_flags)
 
 static void btree_invalidatepage(struct page *page, unsigned long offset)
 {
-	struct extent_map_tree *tree;
-	tree = &BTRFS_I(page->mapping->host)->extent_tree;
+	struct extent_io_tree *tree;
+	tree = &BTRFS_I(page->mapping->host)->io_tree;
 	extent_invalidatepage(tree, page, offset);
 	btree_releasepage(page, GFP_NOFS);
 }
@@ -331,7 +338,7 @@ int readahead_tree_block(struct btrfs_root *root, u64 bytenr, u32 blocksize)
 	buf = btrfs_find_create_tree_block(root, bytenr, blocksize);
 	if (!buf)
 		return 0;
-	read_extent_buffer_pages(&BTRFS_I(btree_inode)->extent_tree,
+	read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree,
 				 buf, 0, 0);
 	free_extent_buffer(buf);
 	return ret;
@@ -342,40 +349,39 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr,
 {
 	struct extent_buffer *buf = NULL;
 	struct inode *btree_inode = root->fs_info->btree_inode;
-	struct extent_map_tree *extent_tree;
+	struct extent_io_tree *io_tree;
 	u64 end;
 	int ret;
 
-	extent_tree = &BTRFS_I(btree_inode)->extent_tree;
+	io_tree = &BTRFS_I(btree_inode)->io_tree;
 
 	buf = btrfs_find_create_tree_block(root, bytenr, blocksize);
 	if (!buf)
 		return NULL;
-	read_extent_buffer_pages(&BTRFS_I(btree_inode)->extent_tree,
-				 buf, 0, 1);
+	read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree, buf, 0, 1);
 
 	if (buf->flags & EXTENT_CSUM)
 		return buf;
 
 	end = buf->start + PAGE_CACHE_SIZE - 1;
-	if (test_range_bit(extent_tree, buf->start, end, EXTENT_CSUM, 1)) {
+	if (test_range_bit(io_tree, buf->start, end, EXTENT_CSUM, 1)) {
 		buf->flags |= EXTENT_CSUM;
 		return buf;
 	}
 
-	lock_extent(extent_tree, buf->start, end, GFP_NOFS);
+	lock_extent(io_tree, buf->start, end, GFP_NOFS);
 
-	if (test_range_bit(extent_tree, buf->start, end, EXTENT_CSUM, 1)) {
+	if (test_range_bit(io_tree, buf->start, end, EXTENT_CSUM, 1)) {
 		buf->flags |= EXTENT_CSUM;
 		goto out_unlock;
 	}
 
 	ret = csum_tree_block(root, buf, 1);
-	set_extent_bits(extent_tree, buf->start, end, EXTENT_CSUM, GFP_NOFS);
+	set_extent_bits(io_tree, buf->start, end, EXTENT_CSUM, GFP_NOFS);
 	buf->flags |= EXTENT_CSUM;
 
 out_unlock:
-	unlock_extent(extent_tree, buf->start, end, GFP_NOFS);
+	unlock_extent(io_tree, buf->start, end, GFP_NOFS);
 	return buf;
 }
 
@@ -385,7 +391,7 @@ int clean_tree_block(struct btrfs_trans_handle *trans, struct btrfs_root *root,
 	struct inode *btree_inode = root->fs_info->btree_inode;
 	if (btrfs_header_generation(buf) ==
 	    root->fs_info->running_transaction->transid)
-		clear_extent_buffer_dirty(&BTRFS_I(btree_inode)->extent_tree,
+		clear_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree,
 					  buf);
 	return 0;
 }
@@ -394,7 +400,7 @@ int wait_on_tree_block_writeback(struct btrfs_root *root,
 				 struct extent_buffer *buf)
 {
 	struct inode *btree_inode = root->fs_info->btree_inode;
-	wait_on_extent_buffer_writeback(&BTRFS_I(btree_inode)->extent_tree,
+	wait_on_extent_buffer_writeback(&BTRFS_I(btree_inode)->io_tree,
 					buf);
 	return 0;
 }
@@ -659,20 +665,23 @@ struct btrfs_root *open_ctree(struct super_block *sb)
 	fs_info->btree_inode->i_nlink = 1;
 	fs_info->btree_inode->i_size = sb->s_bdev->bd_inode->i_size;
 	fs_info->btree_inode->i_mapping->a_ops = &btree_aops;
-	extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree,
+	extent_io_tree_init(&BTRFS_I(fs_info->btree_inode)->io_tree,
 			     fs_info->btree_inode->i_mapping,
 			     GFP_NOFS);
-	BTRFS_I(fs_info->btree_inode)->extent_tree.ops = &btree_extent_map_ops;
+	extent_map_tree_init(&BTRFS_I(fs_info->btree_inode)->extent_tree,
+			     GFP_NOFS);
+
+	BTRFS_I(fs_info->btree_inode)->io_tree.ops = &btree_extent_io_ops;
 
-	extent_map_tree_init(&fs_info->free_space_cache,
+	extent_io_tree_init(&fs_info->free_space_cache,
 			     fs_info->btree_inode->i_mapping, GFP_NOFS);
-	extent_map_tree_init(&fs_info->block_group_cache,
+	extent_io_tree_init(&fs_info->block_group_cache,
 			     fs_info->btree_inode->i_mapping, GFP_NOFS);
-	extent_map_tree_init(&fs_info->pinned_extents,
+	extent_io_tree_init(&fs_info->pinned_extents,
 			     fs_info->btree_inode->i_mapping, GFP_NOFS);
-	extent_map_tree_init(&fs_info->pending_del,
+	extent_io_tree_init(&fs_info->pending_del,
 			     fs_info->btree_inode->i_mapping, GFP_NOFS);
-	extent_map_tree_init(&fs_info->extent_ins,
+	extent_io_tree_init(&fs_info->extent_ins,
 			     fs_info->btree_inode->i_mapping, GFP_NOFS);
 	fs_info->do_barriers = 1;
 	fs_info->closing = 0;
@@ -787,7 +796,7 @@ int write_ctree_super(struct btrfs_trans_handle *trans, struct btrfs_root
 
 	if (!btrfs_test_opt(root, NOBARRIER))
 		blkdev_issue_flush(sb->s_bdev, NULL);
-	set_extent_buffer_dirty(&BTRFS_I(btree_inode)->extent_tree, super);
+	set_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree, super);
 	ret = sync_page_range_nolock(btree_inode, btree_inode->i_mapping,
 				     super->start, super->len);
 	if (!btrfs_test_opt(root, NOBARRIER))
@@ -864,12 +873,12 @@ int close_ctree(struct btrfs_root *root)
 
 	filemap_write_and_wait(fs_info->btree_inode->i_mapping);
 
-	extent_map_tree_empty_lru(&fs_info->free_space_cache);
-	extent_map_tree_empty_lru(&fs_info->block_group_cache);
-	extent_map_tree_empty_lru(&fs_info->pinned_extents);
-	extent_map_tree_empty_lru(&fs_info->pending_del);
-	extent_map_tree_empty_lru(&fs_info->extent_ins);
-	extent_map_tree_empty_lru(&BTRFS_I(fs_info->btree_inode)->extent_tree);
+	extent_io_tree_empty_lru(&fs_info->free_space_cache);
+	extent_io_tree_empty_lru(&fs_info->block_group_cache);
+	extent_io_tree_empty_lru(&fs_info->pinned_extents);
+	extent_io_tree_empty_lru(&fs_info->pending_del);
+	extent_io_tree_empty_lru(&fs_info->extent_ins);
+	extent_io_tree_empty_lru(&BTRFS_I(fs_info->btree_inode)->io_tree);
 
 	truncate_inode_pages(fs_info->btree_inode->i_mapping, 0);
 
@@ -892,13 +901,13 @@ int close_ctree(struct btrfs_root *root)
 int btrfs_buffer_uptodate(struct extent_buffer *buf)
 {
 	struct inode *btree_inode = buf->first_page->mapping->host;
-	return extent_buffer_uptodate(&BTRFS_I(btree_inode)->extent_tree, buf);
+	return extent_buffer_uptodate(&BTRFS_I(btree_inode)->io_tree, buf);
 }
 
 int btrfs_set_buffer_uptodate(struct extent_buffer *buf)
 {
 	struct inode *btree_inode = buf->first_page->mapping->host;
-	return set_extent_buffer_uptodate(&BTRFS_I(btree_inode)->extent_tree,
+	return set_extent_buffer_uptodate(&BTRFS_I(btree_inode)->io_tree,
 					  buf);
 }
 
@@ -914,7 +923,7 @@ void btrfs_mark_buffer_dirty(struct extent_buffer *buf)
 			transid, root->fs_info->generation);
 		WARN_ON(1);
 	}
-	set_extent_buffer_dirty(&BTRFS_I(btree_inode)->extent_tree, buf);
+	set_extent_buffer_dirty(&BTRFS_I(btree_inode)->io_tree, buf);
 }
 
 void btrfs_throttle(struct btrfs_root *root)
@@ -941,7 +950,7 @@ void btrfs_set_buffer_defrag(struct extent_buffer *buf)
 {
 	struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
 	struct inode *btree_inode = root->fs_info->btree_inode;
-	set_extent_bits(&BTRFS_I(btree_inode)->extent_tree, buf->start,
+	set_extent_bits(&BTRFS_I(btree_inode)->io_tree, buf->start,
 			buf->start + buf->len - 1, EXTENT_DEFRAG, GFP_NOFS);
 }
 
@@ -949,7 +958,7 @@ void btrfs_set_buffer_defrag_done(struct extent_buffer *buf)
 {
 	struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
 	struct inode *btree_inode = root->fs_info->btree_inode;
-	set_extent_bits(&BTRFS_I(btree_inode)->extent_tree, buf->start,
+	set_extent_bits(&BTRFS_I(btree_inode)->io_tree, buf->start,
 			buf->start + buf->len - 1, EXTENT_DEFRAG_DONE,
 			GFP_NOFS);
 }
@@ -958,7 +967,7 @@ int btrfs_buffer_defrag(struct extent_buffer *buf)
 {
 	struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
 	struct inode *btree_inode = root->fs_info->btree_inode;
-	return test_range_bit(&BTRFS_I(btree_inode)->extent_tree,
+	return test_range_bit(&BTRFS_I(btree_inode)->io_tree,
 		     buf->start, buf->start + buf->len - 1, EXTENT_DEFRAG, 0);
 }
 
@@ -966,7 +975,7 @@ int btrfs_buffer_defrag_done(struct extent_buffer *buf)
 {
 	struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
 	struct inode *btree_inode = root->fs_info->btree_inode;
-	return test_range_bit(&BTRFS_I(btree_inode)->extent_tree,
+	return test_range_bit(&BTRFS_I(btree_inode)->io_tree,
 		     buf->start, buf->start + buf->len - 1,
 		     EXTENT_DEFRAG_DONE, 0);
 }
@@ -975,7 +984,7 @@ int btrfs_clear_buffer_defrag_done(struct extent_buffer *buf)
 {
 	struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
 	struct inode *btree_inode = root->fs_info->btree_inode;
-	return clear_extent_bits(&BTRFS_I(btree_inode)->extent_tree,
+	return clear_extent_bits(&BTRFS_I(btree_inode)->io_tree,
 		     buf->start, buf->start + buf->len - 1,
 		     EXTENT_DEFRAG_DONE, GFP_NOFS);
 }
@@ -984,7 +993,7 @@ int btrfs_clear_buffer_defrag(struct extent_buffer *buf)
 {
 	struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
 	struct inode *btree_inode = root->fs_info->btree_inode;
-	return clear_extent_bits(&BTRFS_I(btree_inode)->extent_tree,
+	return clear_extent_bits(&BTRFS_I(btree_inode)->io_tree,
 		     buf->start, buf->start + buf->len - 1,
 		     EXTENT_DEFRAG, GFP_NOFS);
 }
@@ -993,10 +1002,10 @@ int btrfs_read_buffer(struct extent_buffer *buf)
 {
 	struct btrfs_root *root = BTRFS_I(buf->first_page->mapping->host)->root;
 	struct inode *btree_inode = root->fs_info->btree_inode;
-	return read_extent_buffer_pages(&BTRFS_I(btree_inode)->extent_tree,
+	return read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree,
 					buf, 0, 1);
 }
 
-static struct extent_map_ops btree_extent_map_ops = {
+static struct extent_io_ops btree_extent_io_ops = {
 	.writepage_io_hook = btree_writepage_io_hook,
 };
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index b69a46691a9..1cf125ab782 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -63,7 +63,7 @@ static int cache_block_group(struct btrfs_root *root,
 	int ret;
 	struct btrfs_key key;
 	struct extent_buffer *leaf;
-	struct extent_map_tree *free_space_cache;
+	struct extent_io_tree *free_space_cache;
 	int slot;
 	u64 last = 0;
 	u64 hole_size;
@@ -158,7 +158,7 @@ struct btrfs_block_group_cache *btrfs_lookup_block_group(struct
 							 btrfs_fs_info *info,
 							 u64 bytenr)
 {
-	struct extent_map_tree *block_group_cache;
+	struct extent_io_tree *block_group_cache;
 	struct btrfs_block_group_cache *block_group = NULL;
 	u64 ptr;
 	u64 start;
@@ -281,7 +281,7 @@ struct btrfs_block_group_cache *btrfs_find_block_group(struct btrfs_root *root,
 						 int data, int owner)
 {
 	struct btrfs_block_group_cache *cache;
-	struct extent_map_tree *block_group_cache;
+	struct extent_io_tree *block_group_cache;
 	struct btrfs_block_group_cache *found_group = NULL;
 	struct btrfs_fs_info *info = root->fs_info;
 	u64 used;
@@ -951,7 +951,7 @@ fail:
 int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans,
 				   struct btrfs_root *root)
 {
-	struct extent_map_tree *block_group_cache;
+	struct extent_io_tree *block_group_cache;
 	struct btrfs_block_group_cache *cache;
 	int ret;
 	int err = 0;
@@ -1107,12 +1107,12 @@ static int update_pinned_extents(struct btrfs_root *root,
 	return 0;
 }
 
-int btrfs_copy_pinned(struct btrfs_root *root, struct extent_map_tree *copy)
+int btrfs_copy_pinned(struct btrfs_root *root, struct extent_io_tree *copy)
 {
 	u64 last = 0;
 	u64 start;
 	u64 end;
-	struct extent_map_tree *pinned_extents = &root->fs_info->pinned_extents;
+	struct extent_io_tree *pinned_extents = &root->fs_info->pinned_extents;
 	int ret;
 
 	while(1) {
@@ -1128,12 +1128,12 @@ int btrfs_copy_pinned(struct btrfs_root *root, struct extent_map_tree *copy)
 
 int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 			       struct btrfs_root *root,
-			       struct extent_map_tree *unpin)
+			       struct extent_io_tree *unpin)
 {
 	u64 start;
 	u64 end;
 	int ret;
-	struct extent_map_tree *free_space_cache;
+	struct extent_io_tree *free_space_cache;
 	free_space_cache = &root->fs_info->free_space_cache;
 
 	while(1) {
@@ -1329,8 +1329,8 @@ static int del_pending_extents(struct btrfs_trans_handle *trans, struct
 	int err = 0;
 	u64 start;
 	u64 end;
-	struct extent_map_tree *pending_del;
-	struct extent_map_tree *pinned_extents;
+	struct extent_io_tree *pending_del;
+	struct extent_io_tree *pinned_extents;
 
 	pending_del = &extent_root->fs_info->pending_del;
 	pinned_extents = &extent_root->fs_info->pinned_extents;
@@ -1802,7 +1802,7 @@ struct extent_buffer *__btrfs_alloc_free_block(struct btrfs_trans_handle *trans,
 
 	set_extent_dirty(&trans->transaction->dirty_pages, buf->start,
 			 buf->start + buf->len - 1, GFP_NOFS);
-	set_extent_bits(&BTRFS_I(root->fs_info->btree_inode)->extent_tree,
+	set_extent_bits(&BTRFS_I(root->fs_info->btree_inode)->io_tree,
 			buf->start, buf->start + buf->len - 1,
 			EXTENT_CSUM, GFP_NOFS);
 	buf->flags |= EXTENT_CSUM;
@@ -2166,7 +2166,7 @@ static int noinline relocate_inode_pages(struct inode *inode, u64 start,
 	unsigned long i;
 	struct page *page;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
-	struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
 	struct file_ra_state *ra;
 
 	ra = kzalloc(sizeof(*ra), GFP_NOFS);
@@ -2195,15 +2195,14 @@ static int noinline relocate_inode_pages(struct inode *inode, u64 start,
 		page_start = (u64)page->index << PAGE_CACHE_SHIFT;
 		page_end = page_start + PAGE_CACHE_SIZE - 1;
 
-		lock_extent(em_tree, page_start, page_end, GFP_NOFS);
+		lock_extent(io_tree, page_start, page_end, GFP_NOFS);
 
 		delalloc_start = page_start;
-		existing_delalloc =
-			count_range_bits(&BTRFS_I(inode)->extent_tree,
-					 &delalloc_start, page_end,
-					 PAGE_CACHE_SIZE, EXTENT_DELALLOC);
+		existing_delalloc = count_range_bits(io_tree,
+					     &delalloc_start, page_end,
+					     PAGE_CACHE_SIZE, EXTENT_DELALLOC);
 
-		set_extent_delalloc(em_tree, page_start,
+		set_extent_delalloc(io_tree, page_start,
 				    page_end, GFP_NOFS);
 
 		spin_lock(&root->fs_info->delalloc_lock);
@@ -2211,7 +2210,7 @@ static int noinline relocate_inode_pages(struct inode *inode, u64 start,
 						 existing_delalloc;
 		spin_unlock(&root->fs_info->delalloc_lock);
 
-		unlock_extent(em_tree, page_start, page_end, GFP_NOFS);
+		unlock_extent(io_tree, page_start, page_end, GFP_NOFS);
 		set_page_dirty(page);
 		unlock_page(page);
 		page_cache_release(page);
@@ -2379,7 +2378,7 @@ int btrfs_shrink_extent_tree(struct btrfs_root *root, u64 new_size)
 	u64 cur_byte;
 	u64 total_found;
 	struct btrfs_fs_info *info = root->fs_info;
-	struct extent_map_tree *block_group_cache;
+	struct extent_io_tree *block_group_cache;
 	struct btrfs_key key;
 	struct btrfs_key found_key;
 	struct extent_buffer *leaf;
@@ -2561,7 +2560,7 @@ int btrfs_grow_extent_tree(struct btrfs_trans_handle *trans,
 	struct btrfs_block_group_cache *cache;
 	struct btrfs_block_group_item *item;
 	struct btrfs_fs_info *info = root->fs_info;
-	struct extent_map_tree *block_group_cache;
+	struct extent_io_tree *block_group_cache;
 	struct btrfs_key key;
 	struct extent_buffer *leaf;
 	int ret;
@@ -2645,7 +2644,7 @@ int btrfs_read_block_groups(struct btrfs_root *root)
 	int bit;
 	struct btrfs_block_group_cache *cache;
 	struct btrfs_fs_info *info = root->fs_info;
-	struct extent_map_tree *block_group_cache;
+	struct extent_io_tree *block_group_cache;
 	struct btrfs_key key;
 	struct btrfs_key found_key;
 	struct extent_buffer *leaf;
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
new file mode 100644
index 00000000000..15cc158a049
--- /dev/null
+++ b/fs/btrfs/extent_io.c
@@ -0,0 +1,3089 @@
+#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 <linux/swap.h>
+#include <linux/version.h>
+#include <linux/writeback.h>
+#include <linux/pagevec.h>
+#include "extent_io.h"
+#include "extent_map.h"
+
+/* temporary define until extent_map moves out of btrfs */
+struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
+				       unsigned long extra_flags,
+				       void (*ctor)(void *, struct kmem_cache *,
+						    unsigned long));
+
+static struct kmem_cache *extent_state_cache;
+static struct kmem_cache *extent_buffer_cache;
+
+static LIST_HEAD(buffers);
+static LIST_HEAD(states);
+
+static spinlock_t state_lock = SPIN_LOCK_UNLOCKED;
+#define BUFFER_LRU_MAX 64
+
+struct tree_entry {
+	u64 start;
+	u64 end;
+	int in_tree;
+	struct rb_node rb_node;
+};
+
+struct extent_page_data {
+	struct bio *bio;
+	struct extent_io_tree *tree;
+	get_extent_t *get_extent;
+};
+
+int __init extent_io_init(void)
+{
+	extent_state_cache = btrfs_cache_create("extent_state",
+					    sizeof(struct extent_state), 0,
+					    NULL);
+	if (!extent_state_cache)
+		return -ENOMEM;
+
+	extent_buffer_cache = btrfs_cache_create("extent_buffers",
+					    sizeof(struct extent_buffer), 0,
+					    NULL);
+	if (!extent_buffer_cache)
+		goto free_state_cache;
+	return 0;
+
+free_state_cache:
+	kmem_cache_destroy(extent_state_cache);
+	return -ENOMEM;
+}
+
+void extent_io_exit(void)
+{
+	struct extent_state *state;
+
+	while (!list_empty(&states)) {
+		state = list_entry(states.next, struct extent_state, list);
+		printk("state leak: start %Lu end %Lu state %lu in tree %d refs %d\n", state->start, state->end, state->state, state->in_tree, atomic_read(&state->refs));
+		list_del(&state->list);
+		kmem_cache_free(extent_state_cache, state);
+
+	}
+
+	if (extent_state_cache)
+		kmem_cache_destroy(extent_state_cache);
+	if (extent_buffer_cache)
+		kmem_cache_destroy(extent_buffer_cache);
+}
+
+void extent_io_tree_init(struct extent_io_tree *tree,
+			  struct address_space *mapping, gfp_t mask)
+{
+	tree->state.rb_node = NULL;
+	tree->ops = NULL;
+	tree->dirty_bytes = 0;
+	rwlock_init(&tree->lock);
+	spin_lock_init(&tree->lru_lock);
+	tree->mapping = mapping;
+	INIT_LIST_HEAD(&tree->buffer_lru);
+	tree->lru_size = 0;
+}
+EXPORT_SYMBOL(extent_io_tree_init);
+
+void extent_io_tree_empty_lru(struct extent_io_tree *tree)
+{
+	struct extent_buffer *eb;
+	while(!list_empty(&tree->buffer_lru)) {
+		eb = list_entry(tree->buffer_lru.next, struct extent_buffer,
+				lru);
+		list_del_init(&eb->lru);
+		free_extent_buffer(eb);
+	}
+}
+EXPORT_SYMBOL(extent_io_tree_empty_lru);
+
+struct extent_state *alloc_extent_state(gfp_t mask)
+{
+	struct extent_state *state;
+	unsigned long flags;
+
+	state = kmem_cache_alloc(extent_state_cache, mask);
+	if (!state || IS_ERR(state))
+		return state;
+	state->state = 0;
+	state->in_tree = 0;
+	state->private = 0;
+
+	spin_lock_irqsave(&state_lock, flags);
+	list_add(&state->list, &states);
+	spin_unlock_irqrestore(&state_lock, flags);
+
+	atomic_set(&state->refs, 1);
+	init_waitqueue_head(&state->wq);
+	return state;
+}
+EXPORT_SYMBOL(alloc_extent_state);
+
+void free_extent_state(struct extent_state *state)
+{
+	unsigned long flags;
+	if (!state)
+		return;
+	if (atomic_dec_and_test(&state->refs)) {
+		WARN_ON(state->in_tree);
+		spin_lock_irqsave(&state_lock, flags);
+		list_del(&state->list);
+		spin_unlock_irqrestore(&state_lock, flags);
+		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 **next_ret)
+{
+	struct rb_node * n = root->rb_node;
+	struct rb_node *prev = NULL;
+	struct rb_node *orig_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) {
+		orig_prev = prev;
+		while(prev && offset > prev_entry->end) {
+			prev = rb_next(prev);
+			prev_entry = rb_entry(prev, struct tree_entry, rb_node);
+		}
+		*prev_ret = prev;
+		prev = orig_prev;
+	}
+
+	if (next_ret) {
+		prev_entry = rb_entry(prev, struct tree_entry, rb_node);
+		while(prev && offset < prev_entry->start) {
+			prev = rb_prev(prev);
+			prev_entry = rb_entry(prev, struct tree_entry, rb_node);
+		}
+		*next_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, NULL);
+	if (!ret)
+		return prev;
+	return ret;
+}
+
+/*
+ * 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_io_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_io_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);
+	}
+	if (bits & EXTENT_DIRTY)
+		tree->dirty_bytes += end - start + 1;
+	state->state |= bits;
+	state->start = start;
+	state->end = end;
+	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' a